diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e4d34d40d80a1947f261c4d62fe47e785aafab4b --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2ad30759d93084b8a8c6e03129e3239fac9a644b4c17c638deb76d82aa8610d5 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..95755d5fee4c9014930f79ee8099abf27f868521 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/config.full.yaml @@ -0,0 +1,219 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_llm_bottom_ratio: 0.0 + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_full_tuning +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..61d8a5e4f967798cb9cea4473f292745daf19053 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/config.yaml @@ -0,0 +1,99 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + eval_num_workers: 8 + include_state: true + latency_curriculum: + enabled: false + obs_image_size: null + per_device_batch_size: 32 + persistent_workers: true + pin_memory: true + prefetch_factor: 4 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning +rl_games: + env_eval: + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 1 + image_size: 224 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: '' + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: flappy +run_id: flappy_fix_latency_2_200ep_full_tuning +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_action_classification_interval: null + eval_interval: 100 + eval_num_batches: 100 + freeze_llm_bottom_ratio: 0.0 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + reload_modules: null + save_interval: 500 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..aa03e6cbf8e375d660528e17fafddb4804b596b7 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.6028500199317932, + 0.3971499800682068, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4890792667865753, + 0.4890792667865753, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 330734, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..0caa254376d5805f972b0b0411e0b5188e3634e3 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7959861159324646, + 0.2040138840675354, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4030573070049286, + 0.4030573070049286, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..1a1a38febf6e9f95ce5bf09dfcbbd840173b1582 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/hydra/.hydra/config.yaml @@ -0,0 +1,217 @@ +framework: + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_llm_bottom_ratio: 0.0 + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_full_tuning +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..beb5cb57e2a7be5650bd5c13c61fbe723add83c1 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/hydra/.hydra/hydra.yaml @@ -0,0 +1,261 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=flappy_fix_latency_2_200ep_full_tuning + - run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=saberrr-zju + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=500 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.eval_action_classification=true + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.freeze_llm_bottom_ratio=0.0 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9,0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=true + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.latency_curriculum.enabled=false + - datasets.vla_data.latency_curriculum.strategy=exclusive + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.num_workers=8 + - datasets.vla_data.eval_num_workers=8 + - datasets.vla_data.prefetch_factor=4 + - datasets.vla_data.persistent_workers=true + - datasets.vla_data.pin_memory=true + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=250 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[2] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.eval_num_workers=8,datasets.vla_data.include_state=true,datasets.vla_data.latency_curriculum.enabled=false,datasets.vla_data.latency_curriculum.strategy=exclusive,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.num_workers=8,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.persistent_workers=true,datasets.vla_data.pin_memory=true,datasets.vla_data.prefetch_factor=4,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=250,rl_games.env_eval.mid_train.latencies=[2],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=flappy_fix_latency_2_200ep_full_tuning,run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_action_classification=true,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.freeze_llm_bottom_ratio=0.0,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=true,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=500,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=saberrr-zju,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/latency-sensitive-bench/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/latency-sensitive-bench/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..a68064bd15651db092fcc05a7ebe0823f33b27b7 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/hydra/.hydra/overrides.yaml @@ -0,0 +1,101 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=flappy_fix_latency_2_200ep_full_tuning +- run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=saberrr-zju +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=500 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.eval_action_classification=true +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.freeze_llm_bottom_ratio=0.0 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9,0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=true +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.latency_curriculum.enabled=false +- datasets.vla_data.latency_curriculum.strategy=exclusive +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.num_workers=8 +- datasets.vla_data.eval_num_workers=8 +- datasets.vla_data.prefetch_factor=4 +- datasets.vla_data.persistent_workers=true +- datasets.vla_data.pin_memory=true +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=250 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[2] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..42f954b32f9c59e610cc46403e714b2fdf9fc85d --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/hydra/train_starvla_hydra.log @@ -0,0 +1,5375 @@ +[2026-06-23 16:40:55,287][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-23 16:40:55,481][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured. +[2026-06-23 16:40:55,481][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub +[2026-06-23 16:40:55,793][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases +[2026-06-23 16:40:55,838][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-23 16:41:06,250][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-23 16:41:06,313][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-23 16:41:06,932][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-23 16:41:06,933][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-23 16:41:06,933][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-23 16:41:06,933][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-23 16:41:06,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-23 16:41:06,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-23 16:41:06,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-23 16:41:06,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-23 16:41:06,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-23 16:41:06,934][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-23 16:41:06,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-23 16:41:06,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-23 16:41:06,935][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-23 16:41:06,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-23 16:41:06,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-23 16:41:06,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-23 16:41:06,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-23 16:41:06,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-23 16:41:06,936][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-23 16:41:06,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-23 16:41:06,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-23 16:41:06,937][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-23 16:41:06,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-23 16:41:06,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-23 16:41:06,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-23 16:41:06,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-23 16:41:06,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-23 16:41:06,938][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-23 16:41:06,939][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-23 16:41:06,939][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-23 16:41:06,939][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-23 16:41:07,000][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-23 16:41:07,047][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/config.full.yaml +[2026-06-23 16:41:07,059][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/config.yaml +[2026-06-23 16:41:07,059][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-23 16:41:10,065][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-23 16:41:10,066][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-23 16:41:10,066][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-23 16:41:10,067][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-23 16:41:10,067][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-23 16:42:07,107][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6955766081809998, 'train/grad_norm_pre_clip': 9.008247375488281, 'timing/data': 0.0007325540063902736, 'timing/model': 1.012440002989024, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-23 16:42:10,763][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6940546631813049, 'train/grad_norm_pre_clip': 10.597851753234863, 'timing/data': 0.0007123159884940833, 'timing/model': 1.0035100709937979, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-23 16:42:14,323][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6910815238952637, 'train/grad_norm_pre_clip': 19.544443130493164, 'timing/data': 0.0006980400066822767, 'timing/model': 0.9984574439877179, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-23 16:42:17,920][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.6912089586257935, 'train/grad_norm_pre_clip': 10.082191467285156, 'timing/data': 0.0007308300118893385, 'timing/model': 1.0017898110090755, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 16:42:21,488][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6869080066680908, 'train/grad_norm_pre_clip': 2.349541187286377, 'timing/data': 0.0006931509997230023, 'timing/model': 0.9906460900092497, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-23 16:42:25,066][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6434117555618286, 'train/grad_norm_pre_clip': 5.964610576629639, 'timing/data': 0.0007026279927231371, 'timing/model': 0.9908367609896231, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-23 16:42:28,658][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.6633473634719849, 'train/grad_norm_pre_clip': 2.351620674133301, 'timing/data': 0.0009224439854733646, 'timing/model': 1.0049841839936562, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 16:42:32,249][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.7092223763465881, 'train/grad_norm_pre_clip': 4.8048906326293945, 'timing/data': 0.0007067060214467347, 'timing/model': 0.9892762879899237, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 16:42:35,828][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.6265909075737, 'train/grad_norm_pre_clip': 9.513448715209961, 'timing/data': 0.0006519049929920584, 'timing/model': 1.0061920550069772, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-23 16:42:39,411][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.7170851230621338, 'train/grad_norm_pre_clip': 7.6946516036987305, 'timing/data': 0.0010075919853989035, 'timing/model': 0.9929410379845649, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-23 16:42:43,027][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.661186933517456, 'train/grad_norm_pre_clip': 7.3249430656433105, 'timing/data': 0.0007218940008897334, 'timing/model': 1.0116388280002866, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-23 16:42:46,633][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.7102130055427551, 'train/grad_norm_pre_clip': 2.1878576278686523, 'timing/data': 0.0006974080170039088, 'timing/model': 1.0041340699826833, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-23 16:42:50,229][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.6610013842582703, 'train/grad_norm_pre_clip': 1.1687136888504028, 'timing/data': 0.0006956360011827201, 'timing/model': 0.9982057849992998, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-23 16:42:54,055][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.6259679198265076, 'train/grad_norm_pre_clip': 1.4213979244232178, 'timing/data': 0.0008402520033996552, 'timing/model': 1.0183490969939157, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-23 16:42:57,669][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.6458240747451782, 'train/grad_norm_pre_clip': 1.9180293083190918, 'timing/data': 0.0006733240152243525, 'timing/model': 0.9951607580005657, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-23 16:43:01,261][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.6492191553115845, 'train/grad_norm_pre_clip': 8.267796516418457, 'timing/data': 0.0006978200108278543, 'timing/model': 1.00502600800246, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 16:43:04,885][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.7097525000572205, 'train/grad_norm_pre_clip': 6.902057647705078, 'timing/data': 0.0013114330067764968, 'timing/model': 1.0078863630187698, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 16:43:08,477][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.6986181735992432, 'train/grad_norm_pre_clip': 7.098935127258301, 'timing/data': 0.0007692810031585395, 'timing/model': 0.9916734129947145, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-23 16:43:12,099][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.681491494178772, 'train/grad_norm_pre_clip': 6.806947708129883, 'timing/data': 0.0011820940126199275, 'timing/model': 1.015390276996186, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-23 16:43:15,728][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.6751313209533691, 'train/grad_norm_pre_clip': 5.720006465911865, 'timing/data': 0.0006974889838602394, 'timing/model': 1.0099181050027255, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-23 16:43:19,344][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.669748842716217, 'train/grad_norm_pre_clip': 1.9009218215942383, 'timing/data': 0.0007455580052919686, 'timing/model': 0.9978513180103619, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-23 16:43:22,982][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.6925359964370728, 'train/grad_norm_pre_clip': 5.0194525718688965, 'timing/data': 0.0008671820105519146, 'timing/model': 1.015195156011032, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 16:43:26,597][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.628494918346405, 'train/grad_norm_pre_clip': 5.692363739013672, 'timing/data': 0.0007149409793782979, 'timing/model': 0.9999454370117746, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 16:43:30,243][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.6143627762794495, 'train/grad_norm_pre_clip': 3.064434766769409, 'timing/data': 0.000859738007420674, 'timing/model': 1.0151698679837864, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-23 16:43:33,860][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.6471208333969116, 'train/grad_norm_pre_clip': 3.2528374195098877, 'timing/data': 0.0009288369910791516, 'timing/model': 1.0020958799868822, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-23 16:43:37,485][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.7297102212905884, 'train/grad_norm_pre_clip': 10.112456321716309, 'timing/data': 0.0008511230116710067, 'timing/model': 1.0100689019891433, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-23 16:43:41,102][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.6949378848075867, 'train/grad_norm_pre_clip': 5.074321746826172, 'timing/data': 0.0011723659990821034, 'timing/model': 0.9987316729966551, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-23 16:43:44,756][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.6311507821083069, 'train/grad_norm_pre_clip': 3.9050066471099854, 'timing/data': 0.001017008995404467, 'timing/model': 1.0152922830020543, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 16:43:48,401][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.6273593306541443, 'train/grad_norm_pre_clip': 5.001396656036377, 'timing/data': 0.0008989909838419408, 'timing/model': 1.0152356179896742, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-23 16:43:52,056][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.6247836947441101, 'train/grad_norm_pre_clip': 11.643320083618164, 'timing/data': 0.0007517790072597563, 'timing/model': 1.0216701379977167, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-23 16:43:55,720][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.641571581363678, 'train/grad_norm_pre_clip': 10.661477088928223, 'timing/data': 0.0007537520141340792, 'timing/model': 1.0158165709872264, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-23 16:43:59,381][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.598578929901123, 'train/grad_norm_pre_clip': 14.80809497833252, 'timing/data': 0.0007238380203489214, 'timing/model': 1.0227420179871842, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-23 16:44:03,061][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.6804113388061523, 'train/grad_norm_pre_clip': 9.86231517791748, 'timing/data': 0.0010577850043773651, 'timing/model': 1.0248052699898835, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-23 16:44:06,717][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.5775644183158875, 'train/grad_norm_pre_clip': 8.470978736877441, 'timing/data': 0.0011971729982178658, 'timing/model': 1.0158221409947146, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-23 16:44:10,382][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.6044731140136719, 'train/grad_norm_pre_clip': 16.492647171020508, 'timing/data': 0.0007869340188335627, 'timing/model': 1.0239748150052037, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01}) +[2026-06-23 16:44:14,057][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.5624058842658997, 'train/grad_norm_pre_clip': 26.270164489746094, 'timing/data': 0.0007938370108604431, 'timing/model': 1.0324651340197306, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01}) +[2026-06-23 16:44:17,746][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.5883805155754089, 'train/grad_norm_pre_clip': 14.634256362915039, 'timing/data': 0.0011306089872960001, 'timing/model': 1.0218849229859188, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.01}) +[2026-06-23 16:44:21,410][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.5149492025375366, 'train/grad_norm_pre_clip': 11.420334815979004, 'timing/data': 0.0007764949987176806, 'timing/model': 1.0276720749970991, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.01}) +[2026-06-23 16:44:25,095][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.4169376492500305, 'train/grad_norm_pre_clip': 15.184076309204102, 'timing/data': 0.000688944011926651, 'timing/model': 1.0133970379829407, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 16:44:28,791][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.4867936968803406, 'train/grad_norm_pre_clip': 21.21910858154297, 'timing/data': 0.0009109640086535364, 'timing/model': 1.0234902380034328, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-23 16:44:32,520][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.3953237235546112, 'train/grad_norm_pre_clip': 14.801687240600586, 'timing/data': 0.000909920985577628, 'timing/model': 1.034131880995119, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-23 16:44:36,200][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.7422051429748535, 'train/grad_norm_pre_clip': 34.15066146850586, 'timing/data': 0.0012321069953031838, 'timing/model': 1.027021562011214, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-23 16:44:39,867][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.5431013703346252, 'train/grad_norm_pre_clip': 46.790592193603516, 'timing/data': 0.0007661549898330122, 'timing/model': 1.0150120110192802, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-23 16:44:43,558][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.40797752141952515, 'train/grad_norm_pre_clip': 42.89387893676758, 'timing/data': 0.0007720460125710815, 'timing/model': 1.033129116985947, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 16:44:47,263][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.46095991134643555, 'train/grad_norm_pre_clip': 35.035892486572266, 'timing/data': 0.0007634310168214142, 'timing/model': 1.0263645480154082, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-23 16:44:50,951][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.3847231864929199, 'train/grad_norm_pre_clip': 52.52548599243164, 'timing/data': 0.0008669419912621379, 'timing/model': 1.0239096709992737, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-23 16:44:54,650][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.35908862948417664, 'train/grad_norm_pre_clip': 33.94683837890625, 'timing/data': 0.0007281650032382458, 'timing/model': 1.0247760510246735, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-23 16:44:58,341][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.3892066180706024, 'train/grad_norm_pre_clip': 34.99935531616211, 'timing/data': 0.0007010159897617996, 'timing/model': 1.0372118600062095, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-23 16:45:02,057][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.6028006672859192, 'train/grad_norm_pre_clip': 39.67591094970703, 'timing/data': 0.001122002984629944, 'timing/model': 1.0369864029926248, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-23 16:45:05,730][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.39375633001327515, 'train/grad_norm_pre_clip': 16.63416290283203, 'timing/data': 0.0007172159967012703, 'timing/model': 1.0210092209745198, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-23 16:45:09,442][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.4974870979785919, 'train/grad_norm_pre_clip': 22.670560836791992, 'timing/data': 0.0007244079897645861, 'timing/model': 1.0344222639978398, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 16:45:13,124][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.4175063967704773, 'train/grad_norm_pre_clip': 50.748931884765625, 'timing/data': 0.0007373730186372995, 'timing/model': 1.0251681249937974, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-23 16:45:16,838][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.49754494428634644, 'train/grad_norm_pre_clip': 30.285385131835938, 'timing/data': 0.0008033050107769668, 'timing/model': 1.020431743003428, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-23 16:45:20,519][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.5458887219429016, 'train/grad_norm_pre_clip': 36.73937225341797, 'timing/data': 0.0007067759870551527, 'timing/model': 1.0118623680027667, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-23 16:45:24,189][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.6049382090568542, 'train/grad_norm_pre_clip': 40.11499786376953, 'timing/data': 0.0007262219733092934, 'timing/model': 1.0208482520247344, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-23 16:45:27,865][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.48763376474380493, 'train/grad_norm_pre_clip': 51.856563568115234, 'timing/data': 0.0007744109898339957, 'timing/model': 1.0098259540100116, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 16:45:31,511][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.47766369581222534, 'train/grad_norm_pre_clip': 33.48275375366211, 'timing/data': 0.0008019609958864748, 'timing/model': 1.0097544720047154, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-23 16:45:35,176][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.37512433528900146, 'train/grad_norm_pre_clip': 53.717926025390625, 'timing/data': 0.0008195650007110089, 'timing/model': 1.0262775930168573, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-23 16:45:38,840][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.44389665126800537, 'train/grad_norm_pre_clip': 30.50655174255371, 'timing/data': 0.0007299690041691065, 'timing/model': 1.0131250099802855, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.02}) +[2026-06-23 16:45:42,504][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.42499077320098877, 'train/grad_norm_pre_clip': 19.995450973510742, 'timing/data': 0.0007169349992182106, 'timing/model': 1.024816582998028, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-23 16:45:46,194][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.3687848448753357, 'train/grad_norm_pre_clip': 23.768993377685547, 'timing/data': 0.0008480269752908498, 'timing/model': 1.0275200990145095, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.02}) +[2026-06-23 16:45:49,888][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.36847490072250366, 'train/grad_norm_pre_clip': 36.25761413574219, 'timing/data': 0.0009583210048731416, 'timing/model': 1.023233848012751, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.02}) +[2026-06-23 16:45:53,582][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.5363596677780151, 'train/grad_norm_pre_clip': 39.687599182128906, 'timing/data': 0.0008337710169143975, 'timing/model': 1.0255281570134684, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.02}) +[2026-06-23 16:45:57,258][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.5149616599082947, 'train/grad_norm_pre_clip': 35.14980697631836, 'timing/data': 0.0007100020011421293, 'timing/model': 1.0263177159940824, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.02}) +[2026-06-23 16:46:00,965][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.3750714957714081, 'train/grad_norm_pre_clip': 53.33540725708008, 'timing/data': 0.0010643159912433475, 'timing/model': 1.0355089090007823, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:46:04,662][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.39276811480522156, 'train/grad_norm_pre_clip': 53.58192443847656, 'timing/data': 0.000869826995767653, 'timing/model': 1.0272742329980247, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-23 16:46:08,358][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.412262499332428, 'train/grad_norm_pre_clip': 47.012603759765625, 'timing/data': 0.0008412640017922968, 'timing/model': 1.0308934699860401, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-23 16:46:12,007][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.5780971050262451, 'train/grad_norm_pre_clip': 70.62079620361328, 'timing/data': 0.0006920390005689114, 'timing/model': 1.0106721649935935, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:46:15,653][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.5518782734870911, 'train/grad_norm_pre_clip': 69.24944305419922, 'timing/data': 0.0008096270030364394, 'timing/model': 1.012459587014746, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-23 16:46:19,367][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.35320472717285156, 'train/grad_norm_pre_clip': 33.203433990478516, 'timing/data': 0.0010577139910310507, 'timing/model': 1.0335112570028286, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-23 16:46:23,081][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.5031068325042725, 'train/grad_norm_pre_clip': 28.43570899963379, 'timing/data': 0.0007349079824052751, 'timing/model': 1.027811483014375, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-23 16:46:26,779][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.4243042469024658, 'train/grad_norm_pre_clip': 33.461448669433594, 'timing/data': 0.0008744460064917803, 'timing/model': 1.0153144820069429, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-23 16:46:30,487][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.3958287537097931, 'train/grad_norm_pre_clip': 16.542917251586914, 'timing/data': 0.000716553971869871, 'timing/model': 1.02995753500727, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-23 16:46:34,196][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.3956775963306427, 'train/grad_norm_pre_clip': 19.33051109313965, 'timing/data': 0.0007250890193972737, 'timing/model': 1.0269071839866228, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-23 16:46:37,901][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.297220915555954, 'train/grad_norm_pre_clip': 27.333070755004883, 'timing/data': 0.000812250014860183, 'timing/model': 1.0166383379837498, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:46:41,583][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.2970806360244751, 'train/grad_norm_pre_clip': 35.721290588378906, 'timing/data': 0.0009640709904488176, 'timing/model': 1.0296466909931041, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-23 16:46:45,275][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.35707375407218933, 'train/grad_norm_pre_clip': 31.25203514099121, 'timing/data': 0.0011230140225961804, 'timing/model': 1.0142861489730421, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-23 16:46:48,961][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.4638465642929077, 'train/grad_norm_pre_clip': 73.01205444335938, 'timing/data': 0.0007361600000876933, 'timing/model': 1.0149667999939993, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:46:52,653][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.5450083017349243, 'train/grad_norm_pre_clip': 66.93305969238281, 'timing/data': 0.0007350480009336025, 'timing/model': 1.0141804800077807, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:46:56,339][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.43312299251556396, 'train/grad_norm_pre_clip': 60.43819808959961, 'timing/data': 0.0008810180006548762, 'timing/model': 1.0169835420092568, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-23 16:47:00,007][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.515146791934967, 'train/grad_norm_pre_clip': 39.7589225769043, 'timing/data': 0.0007355480047408491, 'timing/model': 1.0134163849870674, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.03}) +[2026-06-23 16:47:03,744][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.307670921087265, 'train/grad_norm_pre_clip': 36.59646987915039, 'timing/data': 0.0012572029954753816, 'timing/model': 1.0312156490108464, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.03}) +[2026-06-23 16:47:07,436][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.4410008490085602, 'train/grad_norm_pre_clip': 29.571596145629883, 'timing/data': 0.0009445339965168387, 'timing/model': 1.019319295999594, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.03}) +[2026-06-23 16:47:11,165][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.314146488904953, 'train/grad_norm_pre_clip': 39.241249084472656, 'timing/data': 0.0007407189987134188, 'timing/model': 1.0476529389852658, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.03}) +[2026-06-23 16:47:14,880][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.4222840964794159, 'train/grad_norm_pre_clip': 47.364044189453125, 'timing/data': 0.0008993410156108439, 'timing/model': 1.0341042750224005, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.03}) +[2026-06-23 16:47:18,594][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.33120518922805786, 'train/grad_norm_pre_clip': 87.58101654052734, 'timing/data': 0.0007401170150842518, 'timing/model': 1.027675952995196, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.03}) +[2026-06-23 16:47:22,290][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.38880452513694763, 'train/grad_norm_pre_clip': 47.54763412475586, 'timing/data': 0.0009293470066040754, 'timing/model': 1.0331638830248266, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.03}) +[2026-06-23 16:47:26,002][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.4184238016605377, 'train/grad_norm_pre_clip': 105.25081634521484, 'timing/data': 0.0008669919916428626, 'timing/model': 1.0377966149826534, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:47:29,707][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.9510231614112854, 'train/grad_norm_pre_clip': 172.26904296875, 'timing/data': 0.0008772510045673698, 'timing/model': 1.0331168429984245, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:47:33,392][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.4953397512435913, 'train/grad_norm_pre_clip': 104.27394104003906, 'timing/data': 0.0008684050117153674, 'timing/model': 1.0132715350191575, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-23 16:47:37,063][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.5136736035346985, 'train/grad_norm_pre_clip': 92.9942855834961, 'timing/data': 0.000855329999467358, 'timing/model': 1.0119849720213097, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-23 16:47:40,773][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.5439713597297668, 'train/grad_norm_pre_clip': 78.41514587402344, 'timing/data': 0.0009543829946778715, 'timing/model': 1.0347665789886378, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-23 16:47:44,484][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.27793633937835693, 'train/grad_norm_pre_clip': 61.74953079223633, 'timing/data': 0.0009073560067918152, 'timing/model': 1.0307493710133713, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-23 16:47:48,155][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.24981689453125, 'train/grad_norm_pre_clip': 17.459728240966797, 'timing/data': 0.0008630340162198991, 'timing/model': 1.0234972269972786, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-23 16:47:51,859][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.3665089011192322, 'train/grad_norm_pre_clip': 57.57790756225586, 'timing/data': 0.0013153300096746534, 'timing/model': 1.0196153990109451, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-23 16:47:55,553][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.2831037640571594, 'train/grad_norm_pre_clip': 78.45492553710938, 'timing/data': 0.0007535020122304559, 'timing/model': 1.0321454079821706, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-23 16:47:59,340][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.29778727889060974, 'train/grad_norm_pre_clip': 23.15927505493164, 'timing/data': 0.0009236669866368175, 'timing/model': 1.0209152860043105, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-23 16:48:03,038][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.35577481985092163, 'train/grad_norm_pre_clip': 19.694366455078125, 'timing/data': 0.0007454069855157286, 'timing/model': 1.0321098589920439, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-23 16:48:06,740][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.43967169523239136, 'train/grad_norm_pre_clip': 28.218746185302734, 'timing/data': 0.0007853510032873601, 'timing/model': 1.0378086490090936, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-23 16:48:10,434][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 100 +[2026-06-23 16:50:11,382][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 100 +[2026-06-23 16:50:11,383][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 100 +[2026-06-23 16:50:11,649][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmpo7bvht92/test.c -o /tmp/tmpo7bvht92/test.o +[2026-06-23 16:50:11,683][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmpo7bvht92/test.o -laio -o /tmp/tmpo7bvht92/a.out +[2026-06-23 16:50:12,216][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmpohw6oxha/test.c -o /tmp/tmpohw6oxha/test.o +[2026-06-23 16:50:12,253][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmpohw6oxha/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpohw6oxha/a.out +[2026-06-23 16:50:12,316][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmp9vfu9oc0/test.c -o /tmp/tmp9vfu9oc0/test.o +[2026-06-23 16:50:12,352][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmp9vfu9oc0/test.o -laio -o /tmp/tmp9vfu9oc0/a.out +[2026-06-23 16:51:40,517][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 100 +[2026-06-23 16:51:40,519][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.4440631866455078, 'train/grad_norm_pre_clip': 77.88936614990234, 'eval/action_loss/seconds': 120.94390463398304, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.203623965382576, 'eval/latency_2/loss': 0.203623965382576, 'eval/flappy/loss': 0.203623965382576, 'eval/flappy/latency_2/loss': 0.203623965382576, 'eval/flap_precision': 0.7631935047361299, 'eval/flap_recall': 0.7652645861601085, 'eval/flap_f1': 0.7642276422764227, 'eval/flap_rate_model': 0.20527777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.7642276422764227, 'eval/latency_2/flap_precision': 0.7631935047361299, 'eval/latency_2/flap_recall': 0.7652645861601085, 'eval/latency_2/flap_f1': 0.7642276422764227, 'eval/latency_2/flap_rate_model': 0.20527777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.7642276422764227, 'eval/action_classification/data_wait_seconds': 48.33732604980469, 'eval/action_classification/predict_seconds': 38.81840515136719, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007290459761861712, 'timing/model': 1.0141183480154723, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.04}) +[2026-06-23 16:51:44,876][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.3808729648590088, 'train/grad_norm_pre_clip': 51.793182373046875, 'timing/data': 0.0007621679978910834, 'timing/model': 1.1951116469863337, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.04}) +[2026-06-23 16:51:49,279][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.32825347781181335, 'train/grad_norm_pre_clip': 111.3596420288086, 'timing/data': 0.0007957810012158006, 'timing/model': 1.2012394909979776, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.04}) +[2026-06-23 16:51:53,884][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.43604692816734314, 'train/grad_norm_pre_clip': 108.65287017822266, 'timing/data': 0.0007800209859851748, 'timing/model': 1.446703390975017, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.04}) +[2026-06-23 16:51:58,290][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.37718310952186584, 'train/grad_norm_pre_clip': 57.383697509765625, 'timing/data': 0.0009368599858134985, 'timing/model': 1.1952487480011769, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.04}) +[2026-06-23 16:52:02,682][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.505385160446167, 'train/grad_norm_pre_clip': 143.2822723388672, 'timing/data': 0.0007951489824336022, 'timing/model': 1.1999463590036612, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.04}) +[2026-06-23 16:52:07,128][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.9133795499801636, 'train/grad_norm_pre_clip': 197.5916748046875, 'timing/data': 0.0007330749940592796, 'timing/model': 1.2127840699977241, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.04}) +[2026-06-23 16:52:11,752][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.6999157071113586, 'train/grad_norm_pre_clip': 144.27545166015625, 'timing/data': 0.0008843439863994718, 'timing/model': 1.2234084210067522, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.04}) +[2026-06-23 16:52:16,145][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.22615420818328857, 'train/grad_norm_pre_clip': 47.42346954345703, 'timing/data': 0.0007641519769094884, 'timing/model': 1.193935341987526, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.04}) +[2026-06-23 16:52:20,564][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.6204099655151367, 'train/grad_norm_pre_clip': 149.96829223632812, 'timing/data': 0.0014254839916247874, 'timing/model': 1.2109385080111679, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.04}) +[2026-06-23 16:52:24,989][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.9189550876617432, 'train/grad_norm_pre_clip': 200.66944885253906, 'timing/data': 0.0009144289942923933, 'timing/model': 1.2019130359985866, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.04}) +[2026-06-23 16:52:29,655][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.41239169239997864, 'train/grad_norm_pre_clip': 140.96621704101562, 'timing/data': 0.0008136830001603812, 'timing/model': 1.2343795390042942, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.04}) +[2026-06-23 16:52:34,144][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.3683789372444153, 'train/grad_norm_pre_clip': 59.225711822509766, 'timing/data': 0.0008223590266425163, 'timing/model': 1.2279899969871622, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.04}) +[2026-06-23 16:52:38,585][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.47557148337364197, 'train/grad_norm_pre_clip': 99.33966827392578, 'timing/data': 0.0007710949867032468, 'timing/model': 1.2306628370133694, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.04}) +[2026-06-23 16:52:43,064][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.7703783512115479, 'train/grad_norm_pre_clip': 102.28857421875, 'timing/data': 0.0007872740097809583, 'timing/model': 1.2219832050031982, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.04}) +[2026-06-23 16:52:47,761][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.7943015694618225, 'train/grad_norm_pre_clip': 104.31169128417969, 'timing/data': 0.0011916710063815117, 'timing/model': 1.2040017670078669, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.04}) +[2026-06-23 16:52:52,221][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.4149448871612549, 'train/grad_norm_pre_clip': 38.59092712402344, 'timing/data': 0.0008165990002453327, 'timing/model': 1.2146001090004575, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.04}) +[2026-06-23 16:52:56,689][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.5139908194541931, 'train/grad_norm_pre_clip': 45.366798400878906, 'timing/data': 0.000770993996411562, 'timing/model': 1.2158614210202359, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.05}) +[2026-06-23 16:53:01,352][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.38638779520988464, 'train/grad_norm_pre_clip': 47.60112762451172, 'timing/data': 0.0007611369946971536, 'timing/model': 1.407727166020777, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.05}) +[2026-06-23 16:53:05,808][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.432136207818985, 'train/grad_norm_pre_clip': 17.273391723632812, 'timing/data': 0.0007727969787083566, 'timing/model': 1.2280404370103497, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.05}) +[2026-06-23 16:53:10,293][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.38991281390190125, 'train/grad_norm_pre_clip': 16.92533302307129, 'timing/data': 0.0007740900036878884, 'timing/model': 1.2213011379935779, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.05}) +[2026-06-23 16:53:14,778][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.4638100564479828, 'train/grad_norm_pre_clip': 60.35047149658203, 'timing/data': 0.0009823449945542961, 'timing/model': 1.2285652910068166, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.05}) +[2026-06-23 16:53:19,391][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.3146378993988037, 'train/grad_norm_pre_clip': 102.66238403320312, 'timing/data': 0.0008333289879374206, 'timing/model': 1.2195315140124876, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.05}) +[2026-06-23 16:53:23,881][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.44704943895339966, 'train/grad_norm_pre_clip': 42.75503158569336, 'timing/data': 0.001211458002217114, 'timing/model': 1.2249810239882208, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.05}) +[2026-06-23 16:53:28,338][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.3996535539627075, 'train/grad_norm_pre_clip': 26.70743751525879, 'timing/data': 0.0008014610211830586, 'timing/model': 1.2078252799983602, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.05}) +[2026-06-23 16:53:32,803][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.6060521602630615, 'train/grad_norm_pre_clip': 28.02759552001953, 'timing/data': 0.0007683089934289455, 'timing/model': 1.2314440080081113, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.05}) +[2026-06-23 16:53:37,531][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.40211719274520874, 'train/grad_norm_pre_clip': 16.731056213378906, 'timing/data': 0.0007660849951207638, 'timing/model': 1.4481444480188657, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.05}) +[2026-06-23 16:53:41,996][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.34588387608528137, 'train/grad_norm_pre_clip': 13.975132942199707, 'timing/data': 0.0007980539812706411, 'timing/model': 1.212121204007417, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.05}) +[2026-06-23 16:53:46,450][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.33904290199279785, 'train/grad_norm_pre_clip': 17.121675491333008, 'timing/data': 0.0006948229856789112, 'timing/model': 1.2062581930076703, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.05}) +[2026-06-23 16:53:50,900][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.33817043900489807, 'train/grad_norm_pre_clip': 40.71736145019531, 'timing/data': 0.0008327389950864017, 'timing/model': 1.2212243260000832, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.05}) +[2026-06-23 16:53:55,523][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.22879047691822052, 'train/grad_norm_pre_clip': 29.013086318969727, 'timing/data': 0.0006988009845372289, 'timing/model': 1.3740499459963758, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.05}) +[2026-06-23 16:53:59,982][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.34004563093185425, 'train/grad_norm_pre_clip': 63.55796432495117, 'timing/data': 0.0007060540083330125, 'timing/model': 1.2112860289926175, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.05}) +[2026-06-23 16:54:04,428][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.17153498530387878, 'train/grad_norm_pre_clip': 33.65174865722656, 'timing/data': 0.0007884169754106551, 'timing/model': 1.2084920689812861, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.05}) +[2026-06-23 16:54:08,894][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.3318418562412262, 'train/grad_norm_pre_clip': 105.22959899902344, 'timing/data': 0.0006746260041836649, 'timing/model': 1.2214310959971044, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.05}) +[2026-06-23 16:54:13,379][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.2077626734972, 'train/grad_norm_pre_clip': 93.19091796875, 'timing/data': 0.0008487479935865849, 'timing/model': 1.211129434988834, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.05}) +[2026-06-23 16:54:18,103][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.2199595719575882, 'train/grad_norm_pre_clip': 63.166175842285156, 'timing/data': 0.0007296080002561212, 'timing/model': 1.2216991659952328, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.05}) +[2026-06-23 16:54:22,558][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.4427970349788666, 'train/grad_norm_pre_clip': 65.80109405517578, 'timing/data': 0.0009253599855583161, 'timing/model': 1.209042573987972, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.05}) +[2026-06-23 16:54:27,007][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.2966877818107605, 'train/grad_norm_pre_clip': 69.4874496459961, 'timing/data': 0.0007246890163514763, 'timing/model': 1.2109910559956916, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.05}) +[2026-06-23 16:54:31,460][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.3930548429489136, 'train/grad_norm_pre_clip': 49.98028564453125, 'timing/data': 0.0007511979783885181, 'timing/model': 1.2115793139964808, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.05}) +[2026-06-23 16:54:36,124][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.32688915729522705, 'train/grad_norm_pre_clip': 48.11406326293945, 'timing/data': 0.0007073070155456662, 'timing/model': 1.2092756900237873, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.05}) +[2026-06-23 16:54:40,598][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.25197675824165344, 'train/grad_norm_pre_clip': 33.55968475341797, 'timing/data': 0.0006978090095799416, 'timing/model': 1.2225793709803838, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.05}) +[2026-06-23 16:54:45,091][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.5512173175811768, 'train/grad_norm_pre_clip': 73.98619842529297, 'timing/data': 0.0007208219903986901, 'timing/model': 1.224058976978995, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.05}) +[2026-06-23 16:54:49,745][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.3729226291179657, 'train/grad_norm_pre_clip': 26.451400756835938, 'timing/data': 0.0010968760179821402, 'timing/model': 1.394665931002237, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.05}) +[2026-06-23 16:54:54,227][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.1982630342245102, 'train/grad_norm_pre_clip': 74.50794219970703, 'timing/data': 0.0007476109894923866, 'timing/model': 1.2289327250036877, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.06}) +[2026-06-23 16:54:58,731][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.3475761115550995, 'train/grad_norm_pre_clip': 90.05775451660156, 'timing/data': 0.0007611059991177171, 'timing/model': 1.2241844439995475, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.06}) +[2026-06-23 16:55:03,231][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.18255551159381866, 'train/grad_norm_pre_clip': 22.478025436401367, 'timing/data': 0.0007562369864899665, 'timing/model': 1.2288495049870107, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.06}) +[2026-06-23 16:55:07,896][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.2514230012893677, 'train/grad_norm_pre_clip': 53.69502258300781, 'timing/data': 0.0007501560030505061, 'timing/model': 1.4075034179841168, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.06}) +[2026-06-23 16:55:12,376][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.3311465084552765, 'train/grad_norm_pre_clip': 82.38453674316406, 'timing/data': 0.0008935010118875653, 'timing/model': 1.2153033219801728, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.06}) +[2026-06-23 16:55:16,854][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.40362364053726196, 'train/grad_norm_pre_clip': 25.89834976196289, 'timing/data': 0.000759673974243924, 'timing/model': 1.224160857993411, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.06}) +[2026-06-23 16:55:21,336][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.33779850602149963, 'train/grad_norm_pre_clip': 21.232938766479492, 'timing/data': 0.0008772809815127403, 'timing/model': 1.2253382250200957, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.06}) +[2026-06-23 16:55:26,025][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.27764949202537537, 'train/grad_norm_pre_clip': 25.084980010986328, 'timing/data': 0.0010294020175933838, 'timing/model': 1.4107280139869545, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.06}) +[2026-06-23 16:55:30,527][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.47671249508857727, 'train/grad_norm_pre_clip': 64.64781188964844, 'timing/data': 0.0009330339962616563, 'timing/model': 1.2276102059986442, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.06}) +[2026-06-23 16:55:35,000][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.2831350862979889, 'train/grad_norm_pre_clip': 26.27356719970703, 'timing/data': 0.0007837390003260225, 'timing/model': 1.2244473340106197, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.06}) +[2026-06-23 16:55:39,491][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.30550986528396606, 'train/grad_norm_pre_clip': 45.754146575927734, 'timing/data': 0.0008808470156509429, 'timing/model': 1.2233618299942464, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.06}) +[2026-06-23 16:55:44,153][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.21682143211364746, 'train/grad_norm_pre_clip': 88.7072525024414, 'timing/data': 0.0008160780125763267, 'timing/model': 1.4124282229749952, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.06}) +[2026-06-23 16:55:48,665][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.26002511382102966, 'train/grad_norm_pre_clip': 34.056949615478516, 'timing/data': 0.0014699450111947954, 'timing/model': 1.2281210669898428, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.06}) +[2026-06-23 16:55:53,170][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.5783394575119019, 'train/grad_norm_pre_clip': 128.45263671875, 'timing/data': 0.0007747620111331344, 'timing/model': 1.2310135220177472, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.06}) +[2026-06-23 16:55:57,676][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.5037391781806946, 'train/grad_norm_pre_clip': 147.44024658203125, 'timing/data': 0.0007420509937219322, 'timing/model': 1.2299921560043003, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.06}) +[2026-06-23 16:56:02,368][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.29734519124031067, 'train/grad_norm_pre_clip': 64.06413269042969, 'timing/data': 0.000789048004662618, 'timing/model': 1.4133794180233963, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.06}) +[2026-06-23 16:56:06,876][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.8704833388328552, 'train/grad_norm_pre_clip': 204.256591796875, 'timing/data': 0.0007323730096686631, 'timing/model': 1.229230674012797, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.06}) +[2026-06-23 16:56:11,393][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.43542200326919556, 'train/grad_norm_pre_clip': 192.17481994628906, 'timing/data': 0.0007672679785173386, 'timing/model': 1.2272143689915538, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.06}) +[2026-06-23 16:56:15,895][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.3568609952926636, 'train/grad_norm_pre_clip': 126.0163803100586, 'timing/data': 0.0012933700054418296, 'timing/model': 1.21750742700533, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.06}) +[2026-06-23 16:56:20,583][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.2797548472881317, 'train/grad_norm_pre_clip': 31.35042381286621, 'timing/data': 0.0007299190037883818, 'timing/model': 1.4160869979823474, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.06}) +[2026-06-23 16:56:25,070][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.36396682262420654, 'train/grad_norm_pre_clip': 57.32119369506836, 'timing/data': 0.0008564719755668193, 'timing/model': 1.2268797950237058, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.06}) +[2026-06-23 16:56:29,560][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.3489777147769928, 'train/grad_norm_pre_clip': 65.57801055908203, 'timing/data': 0.0007340970041695982, 'timing/model': 1.2223508100141771, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.06}) +[2026-06-23 16:56:34,073][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.24991877377033234, 'train/grad_norm_pre_clip': 46.95071029663086, 'timing/data': 0.00082711799768731, 'timing/model': 1.2369093500019517, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.06}) +[2026-06-23 16:56:38,737][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.156391441822052, 'train/grad_norm_pre_clip': 43.1460075378418, 'timing/data': 0.000766846991609782, 'timing/model': 1.2356285240093712, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.06}) +[2026-06-23 16:56:43,239][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.2985183000564575, 'train/grad_norm_pre_clip': 38.070960998535156, 'timing/data': 0.0008078529790509492, 'timing/model': 1.2320131260203198, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.06}) +[2026-06-23 16:56:47,732][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.2953943610191345, 'train/grad_norm_pre_clip': 33.80977249145508, 'timing/data': 0.0007636210066266358, 'timing/model': 1.2156696619931608, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.07}) +[2026-06-23 16:56:52,268][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.1954992115497589, 'train/grad_norm_pre_clip': 39.231544494628906, 'timing/data': 0.0007864730141591281, 'timing/model': 1.2221289580047596, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.07}) +[2026-06-23 16:56:56,939][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.16210785508155823, 'train/grad_norm_pre_clip': 76.80321502685547, 'timing/data': 0.0009043909958563745, 'timing/model': 1.2320706380123738, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.07}) +[2026-06-23 16:57:01,434][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.11298692971467972, 'train/grad_norm_pre_clip': 39.9259147644043, 'timing/data': 0.0007350670057348907, 'timing/model': 1.2162708370015025, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.07}) +[2026-06-23 16:57:05,936][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.20347636938095093, 'train/grad_norm_pre_clip': 61.26332473754883, 'timing/data': 0.0009503759793005884, 'timing/model': 1.2237167360144667, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.07}) +[2026-06-23 16:57:10,448][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.19965428113937378, 'train/grad_norm_pre_clip': 65.8676986694336, 'timing/data': 0.0006671620067209005, 'timing/model': 1.236315705027664, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.07}) +[2026-06-23 16:57:15,191][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.16961921751499176, 'train/grad_norm_pre_clip': 91.33513641357422, 'timing/data': 0.000778027024352923, 'timing/model': 1.2294183020130731, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.07}) +[2026-06-23 16:57:19,703][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.2325180470943451, 'train/grad_norm_pre_clip': 57.367122650146484, 'timing/data': 0.0008250440005213022, 'timing/model': 1.2298176920157857, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.07}) +[2026-06-23 16:57:24,183][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.21111318469047546, 'train/grad_norm_pre_clip': 63.8962287902832, 'timing/data': 0.0008107579778879881, 'timing/model': 1.2164284749887884, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.07}) +[2026-06-23 16:57:28,665][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.252226859331131, 'train/grad_norm_pre_clip': 44.36482620239258, 'timing/data': 0.0009842590079642832, 'timing/model': 1.2157359380216803, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.07}) +[2026-06-23 16:57:33,382][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.2547285854816437, 'train/grad_norm_pre_clip': 86.72893524169922, 'timing/data': 0.0008001789974514395, 'timing/model': 1.2282305239932612, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.07}) +[2026-06-23 16:57:37,863][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.232730433344841, 'train/grad_norm_pre_clip': 71.39155578613281, 'timing/data': 0.0008952639764174819, 'timing/model': 1.2193750730075408, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.07}) +[2026-06-23 16:57:42,347][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.3566819727420807, 'train/grad_norm_pre_clip': 86.69146728515625, 'timing/data': 0.0009362100099679083, 'timing/model': 1.231090059998678, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.07}) +[2026-06-23 16:57:46,867][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.42870333790779114, 'train/grad_norm_pre_clip': 177.81668090820312, 'timing/data': 0.0008354629972018301, 'timing/model': 1.2582603699993342, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.07}) +[2026-06-23 16:57:51,534][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.3346244692802429, 'train/grad_norm_pre_clip': 127.13694763183594, 'timing/data': 0.0009524600172881037, 'timing/model': 1.2136004290077835, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.07}) +[2026-06-23 16:57:56,012][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.2812087833881378, 'train/grad_norm_pre_clip': 97.8324203491211, 'timing/data': 0.0007630299951415509, 'timing/model': 1.2190051960060373, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.07}) +[2026-06-23 16:58:00,494][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.4225061535835266, 'train/grad_norm_pre_clip': 67.60334014892578, 'timing/data': 0.0011519090039655566, 'timing/model': 1.2156984980101697, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.07}) +[2026-06-23 16:58:05,015][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.21710367500782013, 'train/grad_norm_pre_clip': 25.579477310180664, 'timing/data': 0.001138043007813394, 'timing/model': 1.2368543859920464, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.07}) +[2026-06-23 16:58:09,772][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.22446630895137787, 'train/grad_norm_pre_clip': 34.23045349121094, 'timing/data': 0.0007807330111972988, 'timing/model': 1.2160418239946011, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.07}) +[2026-06-23 16:58:14,259][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.1525496244430542, 'train/grad_norm_pre_clip': 28.687313079833984, 'timing/data': 0.000764231983339414, 'timing/model': 1.2153089820058085, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.07}) +[2026-06-23 16:58:18,760][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.23381970822811127, 'train/grad_norm_pre_clip': 34.795745849609375, 'timing/data': 0.000788075994933024, 'timing/model': 1.2372863779892214, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.07}) +[2026-06-23 16:58:23,276][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.32980385422706604, 'train/grad_norm_pre_clip': 41.160491943359375, 'timing/data': 0.0011625780025497079, 'timing/model': 1.226758538017748, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.07}) +[2026-06-23 16:58:28,019][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.16066810488700867, 'train/grad_norm_pre_clip': 36.87345886230469, 'timing/data': 0.0007772259996272624, 'timing/model': 1.2279027140175458, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.07}) +[2026-06-23 16:58:32,492][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.31965434551239014, 'train/grad_norm_pre_clip': 98.01133728027344, 'timing/data': 0.0008964760054368526, 'timing/model': 1.2164399059838615, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.07}) +[2026-06-23 16:58:36,989][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.16026148200035095, 'train/grad_norm_pre_clip': 70.9219970703125, 'timing/data': 0.0007697019900660962, 'timing/model': 1.23608023498673, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.07}) +[2026-06-23 16:58:41,462][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.13029584288597107, 'train/grad_norm_pre_clip': 39.743751525878906, 'timing/data': 0.00105837601586245, 'timing/model': 1.21879300501314, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.07}) +[2026-06-23 16:58:46,237][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.1474112868309021, 'train/grad_norm_pre_clip': 29.907529830932617, 'timing/data': 0.0008659599989186972, 'timing/model': 1.2300152810057625, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.08}) +[2026-06-23 16:58:50,745][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.28501781821250916, 'train/grad_norm_pre_clip': 69.17636108398438, 'timing/data': 0.0007394659915007651, 'timing/model': 1.2306097809923813, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.08}) +[2026-06-23 16:58:55,254][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.3484918177127838, 'train/grad_norm_pre_clip': 60.23128128051758, 'timing/data': 0.000744696008041501, 'timing/model': 1.2296525739948265, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.08}) +[2026-06-23 16:58:59,988][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.19569650292396545, 'train/grad_norm_pre_clip': 61.158634185791016, 'timing/data': 0.0007328439969569445, 'timing/model': 1.46502217600937, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.08}) +[2026-06-23 16:59:04,460][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.315799355506897, 'train/grad_norm_pre_clip': 97.61282348632812, 'timing/data': 0.000836535997223109, 'timing/model': 1.2064643740013707, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.08}) +[2026-06-23 16:59:08,903][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.09774062782526016, 'train/grad_norm_pre_clip': 31.11125373840332, 'timing/data': 0.0008092850039247423, 'timing/model': 1.2095456880051643, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.08}) +[2026-06-23 16:59:13,350][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 200 +[2026-06-23 17:00:18,258][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 200 +[2026-06-23 17:00:18,259][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 200 +[2026-06-23 17:00:59,589][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 200 +[2026-06-23 17:00:59,591][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.1827656775712967, 'train/grad_norm_pre_clip': 33.29393005371094, 'eval/action_loss/seconds': 64.90537662201677, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.14114302396774292, 'eval/latency_2/loss': 0.14114302396774292, 'eval/flappy/loss': 0.14114302396774292, 'eval/flappy/latency_2/loss': 0.14114302396774292, 'eval/flap_precision': 0.933786078098472, 'eval/flap_recall': 0.746268656716418, 'eval/flap_f1': 0.8295625942684766, 'eval/flap_rate_model': 0.16361111111111112, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8295625942684766, 'eval/latency_2/flap_precision': 0.933786078098472, 'eval/latency_2/flap_recall': 0.746268656716418, 'eval/latency_2/flap_f1': 0.8295625942684766, 'eval/latency_2/flap_rate_model': 0.16361111111111112, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8295625942684766, 'eval/action_classification/data_wait_seconds': 0.42378801107406616, 'eval/action_classification/predict_seconds': 40.77865219116211, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0008874300110619515, 'timing/model': 1.2140990610059816, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.08}) +[2026-06-23 17:01:04,005][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.3906666934490204, 'train/grad_norm_pre_clip': 96.7743911743164, 'timing/data': 0.0007740499859210104, 'timing/model': 1.209095601021545, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.08}) +[2026-06-23 17:01:08,417][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.17644844949245453, 'train/grad_norm_pre_clip': 34.2800178527832, 'timing/data': 0.0011918120144400746, 'timing/model': 1.206918379990384, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.08}) +[2026-06-23 17:01:12,828][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.26273059844970703, 'train/grad_norm_pre_clip': 50.62892532348633, 'timing/data': 0.0007022380013950169, 'timing/model': 1.2019723519915715, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.08}) +[2026-06-23 17:01:17,417][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.352409690618515, 'train/grad_norm_pre_clip': 55.86783981323242, 'timing/data': 0.0006984710053075105, 'timing/model': 1.3567547540005762, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.08}) +[2026-06-23 17:01:21,799][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.23657463490962982, 'train/grad_norm_pre_clip': 46.096107482910156, 'timing/data': 0.0011056229996029288, 'timing/model': 1.1935656009882223, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.08}) +[2026-06-23 17:01:26,231][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.10315925627946854, 'train/grad_norm_pre_clip': 44.3808479309082, 'timing/data': 0.0007528409769292921, 'timing/model': 1.1979055799893104, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.08}) +[2026-06-23 17:01:30,634][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.24135631322860718, 'train/grad_norm_pre_clip': 65.5259780883789, 'timing/data': 0.0007409090176224709, 'timing/model': 1.2084268909820821, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.08}) +[2026-06-23 17:01:35,336][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.29784920811653137, 'train/grad_norm_pre_clip': 45.72603225708008, 'timing/data': 0.0008917880186345428, 'timing/model': 1.492463177011814, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.08}) +[2026-06-23 17:01:39,771][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.10143624246120453, 'train/grad_norm_pre_clip': 52.9888916015625, 'timing/data': 0.000813512975582853, 'timing/model': 1.212397598021198, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.08}) +[2026-06-23 17:01:44,211][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.14155104756355286, 'train/grad_norm_pre_clip': 65.2840805053711, 'timing/data': 0.0008371869917027652, 'timing/model': 1.215825783991022, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.08}) +[2026-06-23 17:01:48,645][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.22462493181228638, 'train/grad_norm_pre_clip': 76.71429443359375, 'timing/data': 0.000737032009055838, 'timing/model': 1.1962975329952314, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.08}) +[2026-06-23 17:01:53,276][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.0587632842361927, 'train/grad_norm_pre_clip': 35.872718811035156, 'timing/data': 0.0006791240011807531, 'timing/model': 1.2119396750058513, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.08}) +[2026-06-23 17:01:57,670][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.2619052529335022, 'train/grad_norm_pre_clip': 132.94873046875, 'timing/data': 0.0008481880067847669, 'timing/model': 1.1976033270184416, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.08}) +[2026-06-23 17:02:02,078][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.9635704755783081, 'train/grad_norm_pre_clip': 185.7909393310547, 'timing/data': 0.0007425119983963668, 'timing/model': 1.2087707619939465, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.08}) +[2026-06-23 17:02:06,474][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.4381360411643982, 'train/grad_norm_pre_clip': 74.84744262695312, 'timing/data': 0.0006978999881539494, 'timing/model': 1.1978449559828732, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.08}) +[2026-06-23 17:02:11,111][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.3126773536205292, 'train/grad_norm_pre_clip': 67.40477752685547, 'timing/data': 0.0009315109928138554, 'timing/model': 1.224548225000035, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.08}) +[2026-06-23 17:02:15,587][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.39721041917800903, 'train/grad_norm_pre_clip': 104.302734375, 'timing/data': 0.0007888670079410076, 'timing/model': 1.2203364029992372, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.08}) +[2026-06-23 17:02:20,069][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.21531358361244202, 'train/grad_norm_pre_clip': 54.221900939941406, 'timing/data': 0.0009025879844557494, 'timing/model': 1.2212909480149392, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.08}) +[2026-06-23 17:02:24,541][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.24244467914104462, 'train/grad_norm_pre_clip': 34.86100769042969, 'timing/data': 0.0007558359939139336, 'timing/model': 1.2338160100043751, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.08}) +[2026-06-23 17:02:29,239][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.36209604144096375, 'train/grad_norm_pre_clip': 105.24056243896484, 'timing/data': 0.0006876609986647964, 'timing/model': 1.217066995013738, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.09}) +[2026-06-23 17:02:33,730][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.23641008138656616, 'train/grad_norm_pre_clip': 42.68162155151367, 'timing/data': 0.0007478019979316741, 'timing/model': 1.2184017520048656, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.09}) +[2026-06-23 17:02:38,167][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.1518194079399109, 'train/grad_norm_pre_clip': 33.86954879760742, 'timing/data': 0.0006990609981585294, 'timing/model': 1.2209183119994123, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.09}) +[2026-06-23 17:02:42,628][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.17915461957454681, 'train/grad_norm_pre_clip': 51.20677947998047, 'timing/data': 0.0008987000037450343, 'timing/model': 1.2202075099921785, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.09}) +[2026-06-23 17:02:47,360][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.23272770643234253, 'train/grad_norm_pre_clip': 26.499069213867188, 'timing/data': 0.0007204110152088106, 'timing/model': 1.2342875450267456, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.09}) +[2026-06-23 17:02:51,838][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.5508527159690857, 'train/grad_norm_pre_clip': 44.0135383605957, 'timing/data': 0.001142029999755323, 'timing/model': 1.2198568239982706, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.09}) +[2026-06-23 17:02:56,300][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.2363612800836563, 'train/grad_norm_pre_clip': 44.626426696777344, 'timing/data': 0.0007419210160151124, 'timing/model': 1.2227133820124436, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.09}) +[2026-06-23 17:03:00,786][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.263668030500412, 'train/grad_norm_pre_clip': 25.206253051757812, 'timing/data': 0.0007904309895820916, 'timing/model': 1.2135034959937911, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.09}) +[2026-06-23 17:03:05,417][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.22522753477096558, 'train/grad_norm_pre_clip': 23.449384689331055, 'timing/data': 0.000714389985660091, 'timing/model': 1.2203610769938678, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.09}) +[2026-06-23 17:03:09,902][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.22260554134845734, 'train/grad_norm_pre_clip': 38.114158630371094, 'timing/data': 0.0007800710154697299, 'timing/model': 1.2205385160050355, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.09}) +[2026-06-23 17:03:14,359][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.2122923582792282, 'train/grad_norm_pre_clip': 37.51744079589844, 'timing/data': 0.0008840239897836, 'timing/model': 1.221104202006245, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.09}) +[2026-06-23 17:03:18,807][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.3996581435203552, 'train/grad_norm_pre_clip': 50.86064910888672, 'timing/data': 0.0007814540003892034, 'timing/model': 1.2227673540182877, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.09}) +[2026-06-23 17:03:23,489][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.10535892844200134, 'train/grad_norm_pre_clip': 25.201086044311523, 'timing/data': 0.0008156480034813285, 'timing/model': 1.2086119309824426, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.09}) +[2026-06-23 17:03:27,933][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.30676984786987305, 'train/grad_norm_pre_clip': 57.08555221557617, 'timing/data': 0.0007641720003448427, 'timing/model': 1.2196001450065523, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.09}) +[2026-06-23 17:03:32,430][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.19489195942878723, 'train/grad_norm_pre_clip': 26.451812744140625, 'timing/data': 0.0009076769929379225, 'timing/model': 1.2145304690056946, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.09}) +[2026-06-23 17:03:36,856][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.05948755517601967, 'train/grad_norm_pre_clip': 43.98735809326172, 'timing/data': 0.0007628000166732818, 'timing/model': 1.205264148011338, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.09}) +[2026-06-23 17:03:41,479][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.33848437666893005, 'train/grad_norm_pre_clip': 157.8829803466797, 'timing/data': 0.0007868040120229125, 'timing/model': 1.2243362969893496, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.09}) +[2026-06-23 17:03:45,939][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.12149602174758911, 'train/grad_norm_pre_clip': 70.50519561767578, 'timing/data': 0.0008831209852360189, 'timing/model': 1.2053596650075633, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.09}) +[2026-06-23 17:03:50,408][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.1671246886253357, 'train/grad_norm_pre_clip': 123.24462127685547, 'timing/data': 0.0007679090194869787, 'timing/model': 1.2200307339953724, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.09}) +[2026-06-23 17:03:54,859][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.13612781465053558, 'train/grad_norm_pre_clip': 105.66535949707031, 'timing/data': 0.0008594079990871251, 'timing/model': 1.2109905410034116, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.09}) +[2026-06-23 17:03:59,554][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.29339274764060974, 'train/grad_norm_pre_clip': 61.831642150878906, 'timing/data': 0.0010886820091400295, 'timing/model': 1.2065794009831734, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.09}) +[2026-06-23 17:04:03,990][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.2642095983028412, 'train/grad_norm_pre_clip': 46.69796371459961, 'timing/data': 0.0007549750152975321, 'timing/model': 1.2080113980046008, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.09}) +[2026-06-23 17:04:08,428][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.19119447469711304, 'train/grad_norm_pre_clip': 104.51991271972656, 'timing/data': 0.0007559859950561076, 'timing/model': 1.2074150659900624, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.09}) +[2026-06-23 17:04:12,877][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.23757193982601166, 'train/grad_norm_pre_clip': 84.97965240478516, 'timing/data': 0.0009723159892018884, 'timing/model': 1.2127150790183805, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.09}) +[2026-06-23 17:04:17,531][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.13682551681995392, 'train/grad_norm_pre_clip': 54.93162536621094, 'timing/data': 0.0008143050072249025, 'timing/model': 1.203836165019311, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.09}) +[2026-06-23 17:04:21,983][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.335936576128006, 'train/grad_norm_pre_clip': 83.79094696044922, 'timing/data': 0.0007413499988615513, 'timing/model': 1.2192597129906062, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.09}) +[2026-06-23 17:04:26,435][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.1801709681749344, 'train/grad_norm_pre_clip': 38.73196029663086, 'timing/data': 0.00075934300548397, 'timing/model': 1.2080667130067013, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.1}) +[2026-06-23 17:04:30,896][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.12896676361560822, 'train/grad_norm_pre_clip': 74.21105194091797, 'timing/data': 0.0008067500020842999, 'timing/model': 1.2088561830169056, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.1}) +[2026-06-23 17:04:35,597][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.39544054865837097, 'train/grad_norm_pre_clip': 88.7020034790039, 'timing/data': 0.0010006890224758536, 'timing/model': 1.2070311720017344, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.1}) +[2026-06-23 17:04:40,084][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.2695515751838684, 'train/grad_norm_pre_clip': 19.94521141052246, 'timing/data': 0.0009609160188119859, 'timing/model': 1.2244177160027903, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.1}) +[2026-06-23 17:05:58,227][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 250 +[2026-06-23 17:05:58,228][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.18811669945716858, 'train/grad_norm_pre_clip': 42.53536605834961, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 17.12000000000005, 'rl_games_eval/mid_train/mean_length': 157.0, 'rl_games_eval/mid_train/std_reward': 13.225339315117882, 'rl_games_eval/mid_train/std_length': 107.26415990441542, 'rl_games_eval/mid_train/macro_mean_reward': 17.12000000000005, 'rl_games_eval/mid_train/macro_mean_length': 157.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 17.12000000000005, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 157.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 13.225339315117882, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 107.26415990441542, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 17.12000000000005, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0008819689974188805, 'timing/model': 1.2096528559923172, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.1}) +[2026-06-23 17:06:02,838][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.27455148100852966, 'train/grad_norm_pre_clip': 93.08891296386719, 'timing/data': 0.0008589070057496428, 'timing/model': 1.205557890003547, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.1}) +[2026-06-23 17:06:07,217][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.1558171808719635, 'train/grad_norm_pre_clip': 46.494380950927734, 'timing/data': 0.0009694720210973173, 'timing/model': 1.2074017859995365, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.1}) +[2026-06-23 17:06:11,611][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.15209238231182098, 'train/grad_norm_pre_clip': 32.08940505981445, 'timing/data': 0.0007262419967446476, 'timing/model': 1.1893043190066237, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.1}) +[2026-06-23 17:06:16,001][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.3744537830352783, 'train/grad_norm_pre_clip': 105.02273559570312, 'timing/data': 0.0007397469889838248, 'timing/model': 1.2047277959936764, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.1}) +[2026-06-23 17:06:20,607][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.32495367527008057, 'train/grad_norm_pre_clip': 134.9104766845703, 'timing/data': 0.000743162992876023, 'timing/model': 1.21295825499692, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.1}) +[2026-06-23 17:06:25,026][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.11670035123825073, 'train/grad_norm_pre_clip': 61.48733139038086, 'timing/data': 0.0007991869933903217, 'timing/model': 1.1943645989813376, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.1}) +[2026-06-23 17:06:29,464][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.2524726986885071, 'train/grad_norm_pre_clip': 42.87588119506836, 'timing/data': 0.0007364609919022769, 'timing/model': 1.21615751698846, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.1}) +[2026-06-23 17:06:33,926][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.2604812681674957, 'train/grad_norm_pre_clip': 123.1022720336914, 'timing/data': 0.000935148011194542, 'timing/model': 1.2187642030185089, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.1}) +[2026-06-23 17:06:38,646][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.3245529234409332, 'train/grad_norm_pre_clip': 58.405330657958984, 'timing/data': 0.0007497649930883199, 'timing/model': 1.2153663850040175, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.1}) +[2026-06-23 17:06:43,043][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.11258546262979507, 'train/grad_norm_pre_clip': 38.02523422241211, 'timing/data': 0.0007856919837649912, 'timing/model': 1.1945267419796437, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.1}) +[2026-06-23 17:06:47,512][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.1585898995399475, 'train/grad_norm_pre_clip': 30.81256103515625, 'timing/data': 0.0007571590249426663, 'timing/model': 1.2065784689912107, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.1}) +[2026-06-23 17:06:51,965][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.19732558727264404, 'train/grad_norm_pre_clip': 35.72694778442383, 'timing/data': 0.0007737189880572259, 'timing/model': 1.203011409001192, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.1}) +[2026-06-23 17:06:56,589][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.16861234605312347, 'train/grad_norm_pre_clip': 19.620683670043945, 'timing/data': 0.000995068985503167, 'timing/model': 1.2196893649816047, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.1}) +[2026-06-23 17:07:01,057][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.19553826749324799, 'train/grad_norm_pre_clip': 44.449581146240234, 'timing/data': 0.0007690110069233924, 'timing/model': 1.2227875989919994, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.1}) +[2026-06-23 17:07:05,530][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.13816791772842407, 'train/grad_norm_pre_clip': 27.70216941833496, 'timing/data': 0.000996310991467908, 'timing/model': 1.2319711490126792, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.1}) +[2026-06-23 17:07:09,979][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.2210172563791275, 'train/grad_norm_pre_clip': 47.64506149291992, 'timing/data': 0.0008015409985091537, 'timing/model': 1.213707824004814, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.1}) +[2026-06-23 17:07:14,674][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.12340360879898071, 'train/grad_norm_pre_clip': 45.948612213134766, 'timing/data': 0.0012234510213602334, 'timing/model': 1.2310893499816302, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.1}) +[2026-06-23 17:07:19,166][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.1576562225818634, 'train/grad_norm_pre_clip': 66.108642578125, 'timing/data': 0.0007386149954982102, 'timing/model': 1.2220518620160874, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.1}) +[2026-06-23 17:07:23,680][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.19843466579914093, 'train/grad_norm_pre_clip': 45.82712173461914, 'timing/data': 0.0008696260047145188, 'timing/model': 1.2267736469802912, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.1}) +[2026-06-23 17:07:28,201][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.4164774715900421, 'train/grad_norm_pre_clip': 47.3552360534668, 'timing/data': 0.0008549499907530844, 'timing/model': 1.2350442890019622, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.1}) +[2026-06-23 17:07:32,829][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.24783620238304138, 'train/grad_norm_pre_clip': 75.1826171875, 'timing/data': 0.0008980999991763383, 'timing/model': 1.209187854983611, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.1}) +[2026-06-23 17:07:37,336][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.2329346090555191, 'train/grad_norm_pre_clip': 47.45266342163086, 'timing/data': 0.0008623029862064868, 'timing/model': 1.2210025149979629, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.11}) +[2026-06-23 17:07:41,823][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.1895081102848053, 'train/grad_norm_pre_clip': 36.50537109375, 'timing/data': 0.0009699619840830564, 'timing/model': 1.234763551998185, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.11}) +[2026-06-23 17:07:46,415][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.24136224389076233, 'train/grad_norm_pre_clip': 50.67351150512695, 'timing/data': 0.0010602390102576464, 'timing/model': 1.2327527170127723, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.11}) +[2026-06-23 17:07:51,154][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.3224625885486603, 'train/grad_norm_pre_clip': 26.570589065551758, 'timing/data': 0.0008261270122602582, 'timing/model': 1.2163156969763804, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.11}) +[2026-06-23 17:07:55,651][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.23784016072750092, 'train/grad_norm_pre_clip': 44.76746368408203, 'timing/data': 0.0007474710000678897, 'timing/model': 1.2140733530104626, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.11}) +[2026-06-23 17:08:00,145][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.1720752865076065, 'train/grad_norm_pre_clip': 15.850186347961426, 'timing/data': 0.0007174449856393039, 'timing/model': 1.230073797982186, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.11}) +[2026-06-23 17:08:04,607][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.22829802334308624, 'train/grad_norm_pre_clip': 46.87142562866211, 'timing/data': 0.0007835879805497825, 'timing/model': 1.2114322849956807, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.11}) +[2026-06-23 17:08:09,377][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.18508116900920868, 'train/grad_norm_pre_clip': 73.34724426269531, 'timing/data': 0.0007557759818155318, 'timing/model': 1.2432975520205218, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.11}) +[2026-06-23 17:08:13,872][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.13323235511779785, 'train/grad_norm_pre_clip': 89.61407470703125, 'timing/data': 0.0007720970024820417, 'timing/model': 1.2419802680087741, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.11}) +[2026-06-23 17:08:18,356][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.4147857129573822, 'train/grad_norm_pre_clip': 141.61529541015625, 'timing/data': 0.0013070550048723817, 'timing/model': 1.218544217001181, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.11}) +[2026-06-23 17:08:23,059][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.057774681597948074, 'train/grad_norm_pre_clip': 66.00189208984375, 'timing/data': 0.0007769850199110806, 'timing/model': 1.435288370004855, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.11}) +[2026-06-23 17:08:27,553][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.6037164926528931, 'train/grad_norm_pre_clip': 125.18921661376953, 'timing/data': 0.0009491539967712015, 'timing/model': 1.2126267839921638, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.11}) +[2026-06-23 17:08:32,057][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.2361360490322113, 'train/grad_norm_pre_clip': 102.34563446044922, 'timing/data': 0.0007119649962987751, 'timing/model': 1.2343189639796037, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.11}) +[2026-06-23 17:08:36,563][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.20319154858589172, 'train/grad_norm_pre_clip': 90.50755310058594, 'timing/data': 0.0008786830003373325, 'timing/model': 1.2304072270053439, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.11}) +[2026-06-23 17:08:41,296][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.11919906735420227, 'train/grad_norm_pre_clip': 39.94465255737305, 'timing/data': 0.0007265929889399558, 'timing/model': 1.4737173769972287, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.11}) +[2026-06-23 17:08:45,747][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.14089277386665344, 'train/grad_norm_pre_clip': 47.08494567871094, 'timing/data': 0.0007220240077003837, 'timing/model': 1.2104287040128838, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.11}) +[2026-06-23 17:08:50,216][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.2328007072210312, 'train/grad_norm_pre_clip': 38.91851806640625, 'timing/data': 0.0007234570221044123, 'timing/model': 1.2192940239910968, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.11}) +[2026-06-23 17:08:54,689][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.26142650842666626, 'train/grad_norm_pre_clip': 55.45392608642578, 'timing/data': 0.0008450520108453929, 'timing/model': 1.218627733993344, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.11}) +[2026-06-23 17:08:59,387][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.12775450944900513, 'train/grad_norm_pre_clip': 28.782283782958984, 'timing/data': 0.0008686450019013137, 'timing/model': 1.4377535990206525, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.11}) +[2026-06-23 17:09:03,939][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.13563857972621918, 'train/grad_norm_pre_clip': 23.61035919189453, 'timing/data': 0.0008568130142521113, 'timing/model': 1.2422782130015548, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.11}) +[2026-06-23 17:09:08,428][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.30152496695518494, 'train/grad_norm_pre_clip': 25.70777130126953, 'timing/data': 0.0008526850142516196, 'timing/model': 1.238613441993948, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.11}) +[2026-06-23 17:09:12,964][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.18557459115982056, 'train/grad_norm_pre_clip': 21.44434356689453, 'timing/data': 0.0007689000049140304, 'timing/model': 1.2324529199977405, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.11}) +[2026-06-23 17:09:17,485][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.15767750144004822, 'train/grad_norm_pre_clip': 20.606647491455078, 'timing/data': 0.0007793699915055186, 'timing/model': 1.2306802149978466, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.11}) +[2026-06-23 17:09:22,218][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.16292725503444672, 'train/grad_norm_pre_clip': 28.843246459960938, 'timing/data': 0.0007422320195473731, 'timing/model': 1.2313349549949635, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.11}) +[2026-06-23 17:09:26,750][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.11103590577840805, 'train/grad_norm_pre_clip': 58.45705795288086, 'timing/data': 0.001077140012057498, 'timing/model': 1.237755357986316, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.11}) +[2026-06-23 17:09:31,255][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.15526925027370453, 'train/grad_norm_pre_clip': 36.59010696411133, 'timing/data': 0.0007175759819801897, 'timing/model': 1.2358731499989517, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.11}) +[2026-06-23 17:09:35,968][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.1871364265680313, 'train/grad_norm_pre_clip': 38.8037223815918, 'timing/data': 0.000984609010629356, 'timing/model': 1.456152937025763, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.12}) +[2026-06-23 17:09:40,480][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.22125951945781708, 'train/grad_norm_pre_clip': 32.06227111816406, 'timing/data': 0.001275136019103229, 'timing/model': 1.2444135709956754, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.12}) +[2026-06-23 17:09:44,957][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 300 +[2026-06-23 17:10:46,965][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 300 +[2026-06-23 17:10:46,965][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 300 +[2026-06-23 17:11:29,763][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 300 +[2026-06-23 17:11:29,764][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.3580402731895447, 'train/grad_norm_pre_clip': 36.07923126220703, 'eval/action_loss/seconds': 62.004505946009886, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.11116763204336166, 'eval/latency_2/loss': 0.11116763204336166, 'eval/flappy/loss': 0.11116763204336166, 'eval/flappy/latency_2/loss': 0.11116763204336166, 'eval/flap_precision': 0.9570747217806042, 'eval/flap_recall': 0.8168249660786974, 'eval/flap_f1': 0.8814055636896045, 'eval/flap_rate_model': 0.17472222222222222, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8814055636896044, 'eval/latency_2/flap_precision': 0.9570747217806042, 'eval/latency_2/flap_recall': 0.8168249660786974, 'eval/latency_2/flap_f1': 0.8814055636896045, 'eval/latency_2/flap_rate_model': 0.17472222222222222, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8814055636896044, 'eval/action_classification/data_wait_seconds': 0.5048478841781616, 'eval/action_classification/predict_seconds': 42.16154861450195, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.000751118001062423, 'timing/model': 1.2188224169949535, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.12}) +[2026-06-23 17:11:34,454][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.25967171788215637, 'train/grad_norm_pre_clip': 82.63219451904297, 'timing/data': 0.0008416560012847185, 'timing/model': 1.3983906199864578, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.12}) +[2026-06-23 17:11:38,932][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.0816405862569809, 'train/grad_norm_pre_clip': 44.73629379272461, 'timing/data': 0.0007383839983958751, 'timing/model': 1.218564889015397, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.12}) +[2026-06-23 17:11:43,411][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.2096586674451828, 'train/grad_norm_pre_clip': 45.87957000732422, 'timing/data': 0.0010586759890429676, 'timing/model': 1.2155350210086908, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.12}) +[2026-06-23 17:11:47,902][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.25088706612586975, 'train/grad_norm_pre_clip': 71.4186019897461, 'timing/data': 0.0010521440126467496, 'timing/model': 1.2212584060034715, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.12}) +[2026-06-23 17:11:52,595][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.16057413816452026, 'train/grad_norm_pre_clip': 26.55019187927246, 'timing/data': 0.0007563279941678047, 'timing/model': 1.447875033976743, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.12}) +[2026-06-23 17:11:57,073][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.12622296810150146, 'train/grad_norm_pre_clip': 29.008882522583008, 'timing/data': 0.000957489013671875, 'timing/model': 1.214099900011206, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.12}) +[2026-06-23 17:12:01,540][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.0849132388830185, 'train/grad_norm_pre_clip': 32.634246826171875, 'timing/data': 0.0007030390261206776, 'timing/model': 1.2230393179925159, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.12}) +[2026-06-23 17:12:06,024][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.1684553176164627, 'train/grad_norm_pre_clip': 24.331504821777344, 'timing/data': 0.0007285460014827549, 'timing/model': 1.2152573179919273, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.12}) +[2026-06-23 17:12:10,749][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.16078302264213562, 'train/grad_norm_pre_clip': 67.9685287475586, 'timing/data': 0.0007302189769688994, 'timing/model': 1.4666793790238444, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.12}) +[2026-06-23 17:12:15,243][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.2649405300617218, 'train/grad_norm_pre_clip': 32.75465393066406, 'timing/data': 0.0008431270252913237, 'timing/model': 1.215979262982728, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.12}) +[2026-06-23 17:12:19,731][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.03669381141662598, 'train/grad_norm_pre_clip': 84.95328521728516, 'timing/data': 0.0007401670154649764, 'timing/model': 1.2193462940049358, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.12}) +[2026-06-23 17:12:24,233][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.31053391098976135, 'train/grad_norm_pre_clip': 99.0156478881836, 'timing/data': 0.0007877660100348294, 'timing/model': 1.2157698269875254, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.12}) +[2026-06-23 17:12:28,945][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.486344575881958, 'train/grad_norm_pre_clip': 152.18190002441406, 'timing/data': 0.0009853910014498979, 'timing/model': 1.2160201499937102, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.12}) +[2026-06-23 17:12:33,478][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.2354768067598343, 'train/grad_norm_pre_clip': 34.47257614135742, 'timing/data': 0.0008242930052801967, 'timing/model': 1.2384864209743682, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.12}) +[2026-06-23 17:12:38,010][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.21493729948997498, 'train/grad_norm_pre_clip': 68.62651824951172, 'timing/data': 0.0007345870253629982, 'timing/model': 1.2340551230008714, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.12}) +[2026-06-23 17:12:42,561][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.12461688369512558, 'train/grad_norm_pre_clip': 77.26490020751953, 'timing/data': 0.000876259000506252, 'timing/model': 1.2482018019945826, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.12}) +[2026-06-23 17:12:47,279][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.3289453387260437, 'train/grad_norm_pre_clip': 55.95473098754883, 'timing/data': 0.0010142639803234488, 'timing/model': 1.240840771002695, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.12}) +[2026-06-23 17:12:51,778][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.15592114627361298, 'train/grad_norm_pre_clip': 19.566463470458984, 'timing/data': 0.000802832975750789, 'timing/model': 1.2342872919980437, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.12}) +[2026-06-23 17:12:56,337][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.1914350688457489, 'train/grad_norm_pre_clip': 32.78013610839844, 'timing/data': 0.0012868779886048287, 'timing/model': 1.2349808839790057, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.12}) +[2026-06-23 17:13:00,855][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.06021824851632118, 'train/grad_norm_pre_clip': 27.56462287902832, 'timing/data': 0.0007374720007646829, 'timing/model': 1.2371325709973462, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.12}) +[2026-06-23 17:13:05,589][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.10129837691783905, 'train/grad_norm_pre_clip': 62.771060943603516, 'timing/data': 0.0011452159960754216, 'timing/model': 1.2399778680119198, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.12}) +[2026-06-23 17:13:10,099][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.22789020836353302, 'train/grad_norm_pre_clip': 29.97477912902832, 'timing/data': 0.0007207320013549179, 'timing/model': 1.226319642009912, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.12}) +[2026-06-23 17:13:14,613][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.3052186071872711, 'train/grad_norm_pre_clip': 91.1394271850586, 'timing/data': 0.0007388749800156802, 'timing/model': 1.2453221020114142, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.13}) +[2026-06-23 17:13:19,171][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.15806473791599274, 'train/grad_norm_pre_clip': 64.67645263671875, 'timing/data': 0.0011781970097217709, 'timing/model': 1.2467987819982227, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.13}) +[2026-06-23 17:13:24,006][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.05628044530749321, 'train/grad_norm_pre_clip': 67.73151397705078, 'timing/data': 0.0009255500044673681, 'timing/model': 1.248498742992524, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.13}) +[2026-06-23 17:13:28,500][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.3321534991264343, 'train/grad_norm_pre_clip': 67.69503784179688, 'timing/data': 0.0009842980070970953, 'timing/model': 1.2174789480050094, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.13}) +[2026-06-23 17:13:33,057][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.08727338910102844, 'train/grad_norm_pre_clip': 48.574180603027344, 'timing/data': 0.0010340199805796146, 'timing/model': 1.2362845860188827, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.13}) +[2026-06-23 17:13:37,558][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.20450086891651154, 'train/grad_norm_pre_clip': 53.837039947509766, 'timing/data': 0.0007729479984845966, 'timing/model': 1.2400668339978438, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.13}) +[2026-06-23 17:13:42,259][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.23472970724105835, 'train/grad_norm_pre_clip': 71.26066589355469, 'timing/data': 0.0007755529950372875, 'timing/model': 1.2363361610041466, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.13}) +[2026-06-23 17:13:46,788][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.20813165605068207, 'train/grad_norm_pre_clip': 23.45972442626953, 'timing/data': 0.0007828469970263541, 'timing/model': 1.2185271760099567, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.13}) +[2026-06-23 17:13:51,344][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.13628987967967987, 'train/grad_norm_pre_clip': 32.768646240234375, 'timing/data': 0.0007828160014469177, 'timing/model': 1.2395553389796987, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.13}) +[2026-06-23 17:13:55,847][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.2719676196575165, 'train/grad_norm_pre_clip': 63.84312057495117, 'timing/data': 0.0007324639882426709, 'timing/model': 1.2197477919980884, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.13}) +[2026-06-23 17:14:00,586][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.20456406474113464, 'train/grad_norm_pre_clip': 50.50405502319336, 'timing/data': 0.0007662359857931733, 'timing/model': 1.2588731230061967, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.13}) +[2026-06-23 17:14:05,133][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.23648521304130554, 'train/grad_norm_pre_clip': 33.95859909057617, 'timing/data': 0.0008379379869438708, 'timing/model': 1.2428009569994174, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.13}) +[2026-06-23 17:14:09,684][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.19536913931369781, 'train/grad_norm_pre_clip': 36.05656051635742, 'timing/data': 0.0007078280032146722, 'timing/model': 1.238313576002838, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.13}) +[2026-06-23 17:14:14,218][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.1763765513896942, 'train/grad_norm_pre_clip': 73.63017272949219, 'timing/data': 0.0007929249841254205, 'timing/model': 1.2412473169970326, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.13}) +[2026-06-23 17:14:18,891][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.13215069472789764, 'train/grad_norm_pre_clip': 21.328611373901367, 'timing/data': 0.001140808017225936, 'timing/model': 1.227875309996307, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.13}) +[2026-06-23 17:14:23,435][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.08915778249502182, 'train/grad_norm_pre_clip': 75.62506103515625, 'timing/data': 0.0007544340041931719, 'timing/model': 1.239917911007069, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.13}) +[2026-06-23 17:14:27,955][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.14919906854629517, 'train/grad_norm_pre_clip': 117.70530700683594, 'timing/data': 0.000840492983115837, 'timing/model': 1.2408822019933723, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.13}) +[2026-06-23 17:14:32,473][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.27349117398262024, 'train/grad_norm_pre_clip': 46.32973861694336, 'timing/data': 0.0007299779972527176, 'timing/model': 1.2258183789963368, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.13}) +[2026-06-23 17:14:37,169][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.4784266948699951, 'train/grad_norm_pre_clip': 76.57262420654297, 'timing/data': 0.0007299889985006303, 'timing/model': 1.2433649000013247, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.13}) +[2026-06-23 17:14:41,725][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.13325025141239166, 'train/grad_norm_pre_clip': 71.47904968261719, 'timing/data': 0.000932603987166658, 'timing/model': 1.2375398470030632, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.13}) +[2026-06-23 17:14:46,256][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.18820123374462128, 'train/grad_norm_pre_clip': 28.068950653076172, 'timing/data': 0.0010074719903059304, 'timing/model': 1.2287507379951421, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.13}) +[2026-06-23 17:14:50,801][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.1354908049106598, 'train/grad_norm_pre_clip': 27.809484481811523, 'timing/data': 0.0007504670065827668, 'timing/model': 1.2401755879982375, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.13}) +[2026-06-23 17:14:55,467][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.15528759360313416, 'train/grad_norm_pre_clip': 68.45128631591797, 'timing/data': 0.0009899789874907583, 'timing/model': 1.221899759984808, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.13}) +[2026-06-23 17:14:59,986][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.07917796820402145, 'train/grad_norm_pre_clip': 19.5065860748291, 'timing/data': 0.0007046919781714678, 'timing/model': 1.2360919429920614, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.13}) +[2026-06-23 17:15:04,490][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.05268357694149017, 'train/grad_norm_pre_clip': 20.907588958740234, 'timing/data': 0.0009577889868523926, 'timing/model': 1.228219237003941, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.13}) +[2026-06-23 17:15:09,015][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.07936154305934906, 'train/grad_norm_pre_clip': 105.50286102294922, 'timing/data': 0.0008530259947292507, 'timing/model': 1.2379575899976771, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.13}) +[2026-06-23 17:15:13,711][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.05507325381040573, 'train/grad_norm_pre_clip': 60.6822395324707, 'timing/data': 0.0010732229857239872, 'timing/model': 1.2390450699895155, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.14}) +[2026-06-23 17:15:18,215][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.16923174262046814, 'train/grad_norm_pre_clip': 27.586023330688477, 'timing/data': 0.0007279549899976701, 'timing/model': 1.2244590389891528, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.14}) +[2026-06-23 17:15:22,744][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.1309451162815094, 'train/grad_norm_pre_clip': 20.859909057617188, 'timing/data': 0.0007007250096648932, 'timing/model': 1.2385752430127468, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.14}) +[2026-06-23 17:15:27,294][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.08553654700517654, 'train/grad_norm_pre_clip': 59.1580696105957, 'timing/data': 0.00073616998270154, 'timing/model': 1.2392031930212397, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.14}) +[2026-06-23 17:15:31,966][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.3465154469013214, 'train/grad_norm_pre_clip': 28.927953720092773, 'timing/data': 0.0007320529839489609, 'timing/model': 1.2377288379939273, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.14}) +[2026-06-23 17:15:36,470][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.13190220296382904, 'train/grad_norm_pre_clip': 62.00838851928711, 'timing/data': 0.0007132080208975822, 'timing/model': 1.223602848011069, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.14}) +[2026-06-23 17:15:40,986][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.2318161278963089, 'train/grad_norm_pre_clip': 116.9306411743164, 'timing/data': 0.0007196700025815517, 'timing/model': 1.224445806001313, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.14}) +[2026-06-23 17:15:45,499][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.25247886776924133, 'train/grad_norm_pre_clip': 127.34637451171875, 'timing/data': 0.000710052001522854, 'timing/model': 1.2242687300022226, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.14}) +[2026-06-23 17:15:50,161][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.12347936630249023, 'train/grad_norm_pre_clip': 81.88694763183594, 'timing/data': 0.0008927599992603064, 'timing/model': 1.2234509480185807, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.14}) +[2026-06-23 17:15:54,667][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.08726339042186737, 'train/grad_norm_pre_clip': 20.126445770263672, 'timing/data': 0.0006775519868824631, 'timing/model': 1.2238584609876852, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.14}) +[2026-06-23 17:15:59,195][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.3865993022918701, 'train/grad_norm_pre_clip': 77.6472396850586, 'timing/data': 0.0007075580069795251, 'timing/model': 1.2364177170093171, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.14}) +[2026-06-23 17:16:03,748][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.17744290828704834, 'train/grad_norm_pre_clip': 82.72916412353516, 'timing/data': 0.0007392160187009722, 'timing/model': 1.2393317019741517, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.14}) +[2026-06-23 17:16:08,434][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.11263192445039749, 'train/grad_norm_pre_clip': 20.7679386138916, 'timing/data': 0.00086066999938339, 'timing/model': 1.2242833779891953, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.14}) +[2026-06-23 17:16:12,952][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.05145321413874626, 'train/grad_norm_pre_clip': 37.66738510131836, 'timing/data': 0.0007089800201356411, 'timing/model': 1.233714322996093, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.14}) +[2026-06-23 17:16:17,473][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.14260031282901764, 'train/grad_norm_pre_clip': 25.85679817199707, 'timing/data': 0.000744285003747791, 'timing/model': 1.2346260370104574, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.14}) +[2026-06-23 17:16:21,999][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.0822422057390213, 'train/grad_norm_pre_clip': 23.717918395996094, 'timing/data': 0.0009465689945500344, 'timing/model': 1.2311668760085013, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.14}) +[2026-06-23 17:16:26,705][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.02535114996135235, 'train/grad_norm_pre_clip': 25.521669387817383, 'timing/data': 0.0007945279940031469, 'timing/model': 1.2273868720221799, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.14}) +[2026-06-23 17:16:31,249][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.09032930433750153, 'train/grad_norm_pre_clip': 29.138755798339844, 'timing/data': 0.0009624189988244325, 'timing/model': 1.2360390779795125, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.14}) +[2026-06-23 17:16:35,766][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.2342892736196518, 'train/grad_norm_pre_clip': 45.8940315246582, 'timing/data': 0.0008655999845359474, 'timing/model': 1.2311701319995336, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.14}) +[2026-06-23 17:16:40,283][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.11897194385528564, 'train/grad_norm_pre_clip': 39.698184967041016, 'timing/data': 0.00118746398948133, 'timing/model': 1.21990869397996, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.14}) +[2026-06-23 17:16:44,965][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.18478593230247498, 'train/grad_norm_pre_clip': 45.40727996826172, 'timing/data': 0.0011768150143325329, 'timing/model': 1.258772733999649, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.14}) +[2026-06-23 17:16:49,509][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.19508308172225952, 'train/grad_norm_pre_clip': 33.973365783691406, 'timing/data': 0.0007084889803081751, 'timing/model': 1.2428047099965625, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.14}) +[2026-06-23 17:16:54,083][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.09130851179361343, 'train/grad_norm_pre_clip': 18.570697784423828, 'timing/data': 0.0009527099900878966, 'timing/model': 1.2411281920212787, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.14}) +[2026-06-23 17:16:58,623][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.030109498649835587, 'train/grad_norm_pre_clip': 36.66269302368164, 'timing/data': 0.001339514972642064, 'timing/model': 1.2528322079742793, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.14}) +[2026-06-23 17:17:03,313][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.1366015076637268, 'train/grad_norm_pre_clip': 35.63540267944336, 'timing/data': 0.0008079929975792766, 'timing/model': 1.2301669299777132, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.14}) +[2026-06-23 17:17:07,846][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.22835321724414825, 'train/grad_norm_pre_clip': 39.66202926635742, 'timing/data': 0.0009469790093135089, 'timing/model': 1.231195831001969, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.14}) +[2026-06-23 17:17:12,372][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.1775941103696823, 'train/grad_norm_pre_clip': 26.256515502929688, 'timing/data': 0.0007893390138633549, 'timing/model': 1.2273533509869594, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.15}) +[2026-06-23 17:17:16,938][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.258959025144577, 'train/grad_norm_pre_clip': 24.81588363647461, 'timing/data': 0.0010977679921779782, 'timing/model': 1.2496349709981587, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.15}) +[2026-06-23 17:17:21,700][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.3984796404838562, 'train/grad_norm_pre_clip': 39.52216720581055, 'timing/data': 0.0007245790038723499, 'timing/model': 1.2522781049774494, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.15}) +[2026-06-23 17:17:26,242][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.1730666309595108, 'train/grad_norm_pre_clip': 26.016098022460938, 'timing/data': 0.0007632909982930869, 'timing/model': 1.2355792910093442, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.15}) +[2026-06-23 17:17:30,747][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.15784187614917755, 'train/grad_norm_pre_clip': 35.326995849609375, 'timing/data': 0.0007086200057528913, 'timing/model': 1.2255516830191482, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.15}) +[2026-06-23 17:17:35,236][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.1088709831237793, 'train/grad_norm_pre_clip': 21.46316146850586, 'timing/data': 0.0009960310126189142, 'timing/model': 1.2171988899935968, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.15}) +[2026-06-23 17:17:39,970][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.15126101672649384, 'train/grad_norm_pre_clip': 15.419476509094238, 'timing/data': 0.0008085939916782081, 'timing/model': 1.2220234840060584, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.15}) +[2026-06-23 17:17:44,468][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.1909816563129425, 'train/grad_norm_pre_clip': 55.19064712524414, 'timing/data': 0.0007410900143440813, 'timing/model': 1.225946994003607, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.15}) +[2026-06-23 17:17:48,966][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.20108099281787872, 'train/grad_norm_pre_clip': 28.65506935119629, 'timing/data': 0.0007391649996861815, 'timing/model': 1.2246787100157235, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.15}) +[2026-06-23 17:17:53,461][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.6278762221336365, 'train/grad_norm_pre_clip': 182.19642639160156, 'timing/data': 0.0007217829988803715, 'timing/model': 1.2176189360034186, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.15}) +[2026-06-23 17:17:58,235][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.277559369802475, 'train/grad_norm_pre_clip': 111.64838409423828, 'timing/data': 0.0007570589950773865, 'timing/model': 1.2523675010015722, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.15}) +[2026-06-23 17:18:02,774][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.1321154087781906, 'train/grad_norm_pre_clip': 53.40636444091797, 'timing/data': 0.0008895540086086839, 'timing/model': 1.2293984910065774, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.15}) +[2026-06-23 17:18:07,289][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.09160720556974411, 'train/grad_norm_pre_clip': 54.56482696533203, 'timing/data': 0.0008470150060020387, 'timing/model': 1.2248955520044547, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.15}) +[2026-06-23 17:18:11,849][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.16857734322547913, 'train/grad_norm_pre_clip': 59.03981399536133, 'timing/data': 0.0008699279860593379, 'timing/model': 1.251204311993206, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.15}) +[2026-06-23 17:18:16,622][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.04633849114179611, 'train/grad_norm_pre_clip': 57.679805755615234, 'timing/data': 0.0008246439974755049, 'timing/model': 1.2396993150177877, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.15}) +[2026-06-23 17:18:21,186][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.1516576111316681, 'train/grad_norm_pre_clip': 34.40981674194336, 'timing/data': 0.0010742449958343059, 'timing/model': 1.2522033890127204, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.15}) +[2026-06-23 17:18:25,718][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.22991108894348145, 'train/grad_norm_pre_clip': 95.15156555175781, 'timing/data': 0.0007158630178309977, 'timing/model': 1.2408128429960925, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.15}) +[2026-06-23 17:18:30,460][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.1827094852924347, 'train/grad_norm_pre_clip': 45.822689056396484, 'timing/data': 0.0007460280030500144, 'timing/model': 1.4191389459883794, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.15}) +[2026-06-23 17:18:34,993][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.09969839453697205, 'train/grad_norm_pre_clip': 46.574241638183594, 'timing/data': 0.001120700006140396, 'timing/model': 1.2442234039772302, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.15}) +[2026-06-23 17:18:39,549][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.3687555193901062, 'train/grad_norm_pre_clip': 43.51738357543945, 'timing/data': 0.0007367809885181487, 'timing/model': 1.2447018880047835, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.15}) +[2026-06-23 17:18:44,094][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.29254546761512756, 'train/grad_norm_pre_clip': 27.95486068725586, 'timing/data': 0.0007332349778153002, 'timing/model': 1.2394340650062077, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.15}) +[2026-06-23 17:18:48,902][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.12710101902484894, 'train/grad_norm_pre_clip': 41.661705017089844, 'timing/data': 0.0007345569902099669, 'timing/model': 1.495251209998969, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.15}) +[2026-06-23 17:18:53,457][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.09559506922960281, 'train/grad_norm_pre_clip': 45.8469123840332, 'timing/data': 0.0007233670039568096, 'timing/model': 1.24638786399737, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.15}) +[2026-06-23 17:18:57,958][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.12476231157779694, 'train/grad_norm_pre_clip': 19.009836196899414, 'timing/data': 0.0007561970269307494, 'timing/model': 1.22229675098788, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.15}) +[2026-06-23 17:19:02,498][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.09221848845481873, 'train/grad_norm_pre_clip': 16.272308349609375, 'timing/data': 0.0007336650160141289, 'timing/model': 1.2517806589894462, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.15}) +[2026-06-23 17:19:07,151][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 400 +[2026-06-23 17:20:06,906][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 400 +[2026-06-23 17:20:06,906][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 400 +[2026-06-23 17:20:48,941][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 400 +[2026-06-23 17:20:48,942][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.15077197551727295, 'train/grad_norm_pre_clip': 14.041366577148438, 'eval/action_loss/seconds': 59.75090972401085, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.10637923330068588, 'eval/latency_2/loss': 0.10637923330068588, 'eval/flappy/loss': 0.10637923330068588, 'eval/flappy/latency_2/loss': 0.10637923330068588, 'eval/flap_precision': 0.8289623717217788, 'eval/flap_recall': 0.9864314789687924, 'eval/flap_f1': 0.9008674101610904, 'eval/flap_rate_model': 0.2436111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9008674101610904, 'eval/latency_2/flap_precision': 0.8289623717217788, 'eval/latency_2/flap_recall': 0.9864314789687924, 'eval/latency_2/flap_f1': 0.9008674101610904, 'eval/latency_2/flap_rate_model': 0.2436111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9008674101610904, 'eval/action_classification/data_wait_seconds': 0.4188217520713806, 'eval/action_classification/predict_seconds': 41.447662353515625, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007879459881223738, 'timing/model': 1.2261376079986803, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.15}) +[2026-06-23 17:20:53,446][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.363640695810318, 'train/grad_norm_pre_clip': 91.53268432617188, 'timing/data': 0.0007315439870581031, 'timing/model': 1.2251106979965698, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.16}) +[2026-06-23 17:20:57,985][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.05164962634444237, 'train/grad_norm_pre_clip': 46.404666900634766, 'timing/data': 0.00093230398488231, 'timing/model': 1.2380139050073922, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.16}) +[2026-06-23 17:21:02,639][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.1566152274608612, 'train/grad_norm_pre_clip': 43.91949462890625, 'timing/data': 0.0007406999939121306, 'timing/model': 1.3844812960014679, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.16}) +[2026-06-23 17:21:07,139][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.35956069827079773, 'train/grad_norm_pre_clip': 49.55133056640625, 'timing/data': 0.0008027449948713183, 'timing/model': 1.2353144850058015, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.16}) +[2026-06-23 17:21:11,655][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.10371046513319016, 'train/grad_norm_pre_clip': 95.98025512695312, 'timing/data': 0.000968211010331288, 'timing/model': 1.231486506992951, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.16}) +[2026-06-23 17:21:16,170][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.17513668537139893, 'train/grad_norm_pre_clip': 44.39242935180664, 'timing/data': 0.000750777981011197, 'timing/model': 1.232790459995158, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.16}) +[2026-06-23 17:21:20,706][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.28778785467147827, 'train/grad_norm_pre_clip': 118.04730224609375, 'timing/data': 0.0007445170194841921, 'timing/model': 1.2317430859839078, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.16}) +[2026-06-23 17:21:25,436][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.3058197796344757, 'train/grad_norm_pre_clip': 63.18607711791992, 'timing/data': 0.0009325550054199994, 'timing/model': 1.2438973870011978, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.16}) +[2026-06-23 17:21:29,945][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.12081016600131989, 'train/grad_norm_pre_clip': 87.43311309814453, 'timing/data': 0.0009120170143432915, 'timing/model': 1.2364681400067639, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.16}) +[2026-06-23 17:21:34,405][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.10561308264732361, 'train/grad_norm_pre_clip': 21.514881134033203, 'timing/data': 0.0011567889887373894, 'timing/model': 1.2136203540139832, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.16}) +[2026-06-23 17:21:38,880][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.05415361747145653, 'train/grad_norm_pre_clip': 87.340576171875, 'timing/data': 0.0008215400157496333, 'timing/model': 1.216467232996365, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.16}) +[2026-06-23 17:21:43,591][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.5079589486122131, 'train/grad_norm_pre_clip': 96.53119659423828, 'timing/data': 0.000756238994654268, 'timing/model': 1.2307549210090656, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.16}) +[2026-06-23 17:21:48,072][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.27194899320602417, 'train/grad_norm_pre_clip': 60.63041687011719, 'timing/data': 0.0012689570139627904, 'timing/model': 1.2198619379778393, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.16}) +[2026-06-23 17:21:52,577][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.09616280347108841, 'train/grad_norm_pre_clip': 24.189905166625977, 'timing/data': 0.0008562330040149391, 'timing/model': 1.2339792309794575, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.16}) +[2026-06-23 17:21:57,080][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.10429615527391434, 'train/grad_norm_pre_clip': 44.78749465942383, 'timing/data': 0.0007324239995796233, 'timing/model': 1.216920184000628, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.16}) +[2026-06-23 17:22:01,768][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.3364417552947998, 'train/grad_norm_pre_clip': 57.28461837768555, 'timing/data': 0.0008394529868382961, 'timing/model': 1.237114074989222, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.16}) +[2026-06-23 17:22:06,302][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.050605908036231995, 'train/grad_norm_pre_clip': 30.144601821899414, 'timing/data': 0.0008231719839386642, 'timing/model': 1.2358535469975322, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.16}) +[2026-06-23 17:22:10,845][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.17863920331001282, 'train/grad_norm_pre_clip': 22.548110961914062, 'timing/data': 0.0011900310055352747, 'timing/model': 1.2362614049925469, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.16}) +[2026-06-23 17:22:15,354][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.19690966606140137, 'train/grad_norm_pre_clip': 19.412067413330078, 'timing/data': 0.0010917689942289144, 'timing/model': 1.2273421529971529, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.16}) +[2026-06-23 17:22:20,090][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.31078040599823, 'train/grad_norm_pre_clip': 72.34211730957031, 'timing/data': 0.0007484240049961954, 'timing/model': 1.2251690679986496, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.16}) +[2026-06-23 17:22:24,611][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.03915565460920334, 'train/grad_norm_pre_clip': 25.186105728149414, 'timing/data': 0.0007450880075339228, 'timing/model': 1.2231232789927162, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.16}) +[2026-06-23 17:22:29,136][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.06728576123714447, 'train/grad_norm_pre_clip': 20.338550567626953, 'timing/data': 0.0007335259870160371, 'timing/model': 1.238687476987252, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.16}) +[2026-06-23 17:22:33,683][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.179779052734375, 'train/grad_norm_pre_clip': 24.62582778930664, 'timing/data': 0.000730460014892742, 'timing/model': 1.2253763750195503, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.16}) +[2026-06-23 17:22:38,445][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.15241119265556335, 'train/grad_norm_pre_clip': 41.7709846496582, 'timing/data': 0.0010376590071246028, 'timing/model': 1.2242442850256339, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.16}) +[2026-06-23 17:22:42,989][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.0929488018155098, 'train/grad_norm_pre_clip': 40.0440559387207, 'timing/data': 0.0008044579881243408, 'timing/model': 1.2396602499939036, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.16}) +[2026-06-23 17:22:47,502][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.20011232793331146, 'train/grad_norm_pre_clip': 88.3093032836914, 'timing/data': 0.0007510989962611347, 'timing/model': 1.2244318519951776, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.16}) +[2026-06-23 17:22:52,010][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.2618180513381958, 'train/grad_norm_pre_clip': 49.09330749511719, 'timing/data': 0.0009469409997109324, 'timing/model': 1.2241085899877362, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.17}) +[2026-06-23 17:22:56,677][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.15206004679203033, 'train/grad_norm_pre_clip': 54.944175720214844, 'timing/data': 0.0007772079843562096, 'timing/model': 1.2357358069857582, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.17}) +[2026-06-23 17:23:01,214][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.19372238218784332, 'train/grad_norm_pre_clip': 30.512224197387695, 'timing/data': 0.0007896210008766502, 'timing/model': 1.228297866997309, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.17}) +[2026-06-23 17:23:05,779][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.14112326502799988, 'train/grad_norm_pre_clip': 20.184003829956055, 'timing/data': 0.0011066570004913956, 'timing/model': 1.2395807479915675, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.17}) +[2026-06-23 17:23:10,296][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.09693162143230438, 'train/grad_norm_pre_clip': 21.591922760009766, 'timing/data': 0.0007860830228310078, 'timing/model': 1.2298186800035182, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.17}) +[2026-06-23 17:23:15,042][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.19217880070209503, 'train/grad_norm_pre_clip': 41.50733947753906, 'timing/data': 0.0006920500018168241, 'timing/model': 1.220950405986514, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.17}) +[2026-06-23 17:23:19,576][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.11127135902643204, 'train/grad_norm_pre_clip': 47.28422164916992, 'timing/data': 0.0008764610101934522, 'timing/model': 1.2340542540187016, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.17}) +[2026-06-23 17:23:24,106][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.1917174756526947, 'train/grad_norm_pre_clip': 34.54220962524414, 'timing/data': 0.0009077080176211894, 'timing/model': 1.2215565590013284, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.17}) +[2026-06-23 17:23:28,632][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.14801812171936035, 'train/grad_norm_pre_clip': 29.245811462402344, 'timing/data': 0.000723918987205252, 'timing/model': 1.2340660089976154, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.17}) +[2026-06-23 17:23:33,429][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.14474868774414062, 'train/grad_norm_pre_clip': 47.90349197387695, 'timing/data': 0.0008142860024236143, 'timing/model': 1.2541542609978933, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.17}) +[2026-06-23 17:23:37,993][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.21529877185821533, 'train/grad_norm_pre_clip': 41.83320236206055, 'timing/data': 0.0021623889915645123, 'timing/model': 1.2411363039864227, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.17}) +[2026-06-23 17:23:42,564][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.4152395725250244, 'train/grad_norm_pre_clip': 92.77678680419922, 'timing/data': 0.0009315919887740165, 'timing/model': 1.232996948005166, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.17}) +[2026-06-23 17:23:47,297][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.17939087748527527, 'train/grad_norm_pre_clip': 70.50639343261719, 'timing/data': 0.0007266140019055456, 'timing/model': 1.4425525770056993, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.17}) +[2026-06-23 17:23:51,845][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.13200347125530243, 'train/grad_norm_pre_clip': 71.85313415527344, 'timing/data': 0.0007270739879459143, 'timing/model': 1.2391106570139527, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.17}) +[2026-06-23 17:23:56,367][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.12078508734703064, 'train/grad_norm_pre_clip': 20.707717895507812, 'timing/data': 0.0009032299858517945, 'timing/model': 1.2373383969825227, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.17}) +[2026-06-23 17:24:00,926][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.2055135816335678, 'train/grad_norm_pre_clip': 65.26155853271484, 'timing/data': 0.0009977850131690502, 'timing/model': 1.2390685220016167, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.17}) +[2026-06-23 17:24:05,695][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.11850212514400482, 'train/grad_norm_pre_clip': 35.46102523803711, 'timing/data': 0.0011793109879363328, 'timing/model': 1.4686767710081767, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.17}) +[2026-06-23 17:24:10,246][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.08859460800886154, 'train/grad_norm_pre_clip': 68.04341888427734, 'timing/data': 0.0007445669907610863, 'timing/model': 1.2276021430152468, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.17}) +[2026-06-23 17:24:14,777][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.070309117436409, 'train/grad_norm_pre_clip': 33.6411018371582, 'timing/data': 0.0010816499998327345, 'timing/model': 1.2291466270107776, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.17}) +[2026-06-23 17:24:19,337][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.2891486883163452, 'train/grad_norm_pre_clip': 40.00094223022461, 'timing/data': 0.0007658959948457778, 'timing/model': 1.2415147330029868, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.17}) +[2026-06-23 17:24:24,071][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.051650743931531906, 'train/grad_norm_pre_clip': 29.53703498840332, 'timing/data': 0.0007222749991342425, 'timing/model': 1.4378364440053701, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.17}) +[2026-06-23 17:24:28,599][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.05654265359044075, 'train/grad_norm_pre_clip': 23.39647102355957, 'timing/data': 0.000953412993112579, 'timing/model': 1.2182617419748567, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.17}) +[2026-06-23 17:24:33,161][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.34294676780700684, 'train/grad_norm_pre_clip': 32.65401840209961, 'timing/data': 0.0007214340148493648, 'timing/model': 1.2447046439920086, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.17}) +[2026-06-23 17:24:37,701][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.22808508574962616, 'train/grad_norm_pre_clip': 38.19502258300781, 'timing/data': 0.0006996140000410378, 'timing/model': 1.2461347760108765, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.17}) +[2026-06-23 17:24:42,406][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.21698583662509918, 'train/grad_norm_pre_clip': 30.191455841064453, 'timing/data': 0.0007298900163732469, 'timing/model': 1.2385985489818268, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.17}) +[2026-06-23 17:24:46,888][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.18364450335502625, 'train/grad_norm_pre_clip': 42.932212829589844, 'timing/data': 0.0009032609814312309, 'timing/model': 1.217978340020636, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.17}) +[2026-06-23 17:24:51,376][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.057218339294195175, 'train/grad_norm_pre_clip': 39.73429870605469, 'timing/data': 0.0008638779981993139, 'timing/model': 1.221201767999446, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.18}) +[2026-06-23 17:24:55,881][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.15647733211517334, 'train/grad_norm_pre_clip': 85.14615631103516, 'timing/data': 0.0007135900086723268, 'timing/model': 1.2346405549906194, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.18}) +[2026-06-23 17:25:00,634][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.009476627223193645, 'train/grad_norm_pre_clip': 46.865875244140625, 'timing/data': 0.0009368019818793982, 'timing/model': 1.2187180230102967, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.18}) +[2026-06-23 17:25:05,184][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.06884178519248962, 'train/grad_norm_pre_clip': 20.05440902709961, 'timing/data': 0.0009027089981827885, 'timing/model': 1.245650609023869, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.18}) +[2026-06-23 17:25:09,698][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.012740151025354862, 'train/grad_norm_pre_clip': 20.834333419799805, 'timing/data': 0.0007638119859620929, 'timing/model': 1.2275237010035198, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.18}) +[2026-06-23 17:25:14,238][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.0703471228480339, 'train/grad_norm_pre_clip': 32.08735656738281, 'timing/data': 0.0008318089821841568, 'timing/model': 1.2341090050176717, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.18}) +[2026-06-23 17:25:19,016][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.17955967783927917, 'train/grad_norm_pre_clip': 37.869991302490234, 'timing/data': 0.0007200620020739734, 'timing/model': 1.228231015993515, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.18}) +[2026-06-23 17:25:23,536][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.025348592549562454, 'train/grad_norm_pre_clip': 32.042633056640625, 'timing/data': 0.0007148909789975733, 'timing/model': 1.2286778610141482, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.18}) +[2026-06-23 17:25:28,057][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.09744202345609665, 'train/grad_norm_pre_clip': 34.440677642822266, 'timing/data': 0.0007454980222973973, 'timing/model': 1.2285693490121048, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.18}) +[2026-06-23 17:25:32,572][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.10454925149679184, 'train/grad_norm_pre_clip': 44.95658493041992, 'timing/data': 0.0007384649943560362, 'timing/model': 1.2277100589999463, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.18}) +[2026-06-23 17:25:37,334][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.14254529774188995, 'train/grad_norm_pre_clip': 37.21742630004883, 'timing/data': 0.0007257920224219561, 'timing/model': 1.2284081490070093, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.18}) +[2026-06-23 17:25:41,877][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.10509923845529556, 'train/grad_norm_pre_clip': 27.704978942871094, 'timing/data': 0.0007850409892853349, 'timing/model': 1.234750415023882, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.18}) +[2026-06-23 17:25:46,438][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.3437891900539398, 'train/grad_norm_pre_clip': 39.66508483886719, 'timing/data': 0.0010376890131738037, 'timing/model': 1.231470589991659, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.18}) +[2026-06-23 17:25:50,975][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.12347240000963211, 'train/grad_norm_pre_clip': 23.055091857910156, 'timing/data': 0.0007503879896830767, 'timing/model': 1.228152304975083, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.18}) +[2026-06-23 17:25:55,662][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.14509277045726776, 'train/grad_norm_pre_clip': 38.18415451049805, 'timing/data': 0.0007095919863786548, 'timing/model': 1.2250440460047685, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.18}) +[2026-06-23 17:26:00,162][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.09364353865385056, 'train/grad_norm_pre_clip': 37.1590461730957, 'timing/data': 0.0007105940021574497, 'timing/model': 1.2245523870224133, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.18}) +[2026-06-23 17:26:04,668][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.10926925390958786, 'train/grad_norm_pre_clip': 34.56991195678711, 'timing/data': 0.0007227460155263543, 'timing/model': 1.2243547730031423, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.18}) +[2026-06-23 17:26:09,168][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.14799784123897552, 'train/grad_norm_pre_clip': 54.6166877746582, 'timing/data': 0.0007205420115496963, 'timing/model': 1.2242720689973794, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.18}) +[2026-06-23 17:26:13,837][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.08822478353977203, 'train/grad_norm_pre_clip': 45.02275848388672, 'timing/data': 0.0007323739991988987, 'timing/model': 1.2231741689902265, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.18}) +[2026-06-23 17:26:18,348][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.0682467371225357, 'train/grad_norm_pre_clip': 36.01347732543945, 'timing/data': 0.0007199009996838868, 'timing/model': 1.2259694169915747, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.18}) +[2026-06-23 17:26:22,859][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.1818358451128006, 'train/grad_norm_pre_clip': 34.08147048950195, 'timing/data': 0.0008846749842632562, 'timing/model': 1.2265097970084753, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.18}) +[2026-06-23 17:26:27,369][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.12370620667934418, 'train/grad_norm_pre_clip': 30.061250686645508, 'timing/data': 0.000716745009412989, 'timing/model': 1.2263610520167276, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.18}) +[2026-06-23 17:26:32,083][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.15313690900802612, 'train/grad_norm_pre_clip': 32.8553352355957, 'timing/data': 0.00112431900925003, 'timing/model': 1.2421349350188393, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.18}) +[2026-06-23 17:26:36,638][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.1553540676832199, 'train/grad_norm_pre_clip': 47.165077209472656, 'timing/data': 0.0008894540078472346, 'timing/model': 1.2431108900054824, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.18}) +[2026-06-23 17:26:41,167][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.16767403483390808, 'train/grad_norm_pre_clip': 36.33332443237305, 'timing/data': 0.0008608819916844368, 'timing/model': 1.245350811019307, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.18}) +[2026-06-23 17:26:45,693][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.13632279634475708, 'train/grad_norm_pre_clip': 32.15986633300781, 'timing/data': 0.0007059750205371529, 'timing/model': 1.2265679859847296, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.18}) +[2026-06-23 17:26:50,410][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.24427223205566406, 'train/grad_norm_pre_clip': 36.591392517089844, 'timing/data': 0.0007486040121875703, 'timing/model': 1.239724239014322, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.19}) +[2026-06-23 17:26:54,957][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.13662250339984894, 'train/grad_norm_pre_clip': 11.19088363647461, 'timing/data': 0.0007245989982038736, 'timing/model': 1.2356669589935336, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.19}) +[2026-06-23 17:26:59,524][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.12429168820381165, 'train/grad_norm_pre_clip': 24.80702781677246, 'timing/data': 0.001227530010510236, 'timing/model': 1.245633705984801, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.19}) +[2026-06-23 17:27:04,031][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.09644009917974472, 'train/grad_norm_pre_clip': 26.286569595336914, 'timing/data': 0.0007284070015884936, 'timing/model': 1.223082846001489, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.19}) +[2026-06-23 17:27:08,739][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.23193182051181793, 'train/grad_norm_pre_clip': 23.483991622924805, 'timing/data': 0.0007498759950976819, 'timing/model': 1.2418154579936527, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.19}) +[2026-06-23 17:27:13,251][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.2137811779975891, 'train/grad_norm_pre_clip': 24.582895278930664, 'timing/data': 0.0008661909960210323, 'timing/model': 1.224316007021116, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.19}) +[2026-06-23 17:27:17,771][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.22248175740242004, 'train/grad_norm_pre_clip': 20.5640926361084, 'timing/data': 0.0009498760045971721, 'timing/model': 1.2259206349845044, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.19}) +[2026-06-23 17:27:22,288][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.0515349805355072, 'train/grad_norm_pre_clip': 41.33671951293945, 'timing/data': 0.0007274649979081005, 'timing/model': 1.2249391609802842, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.19}) +[2026-06-23 17:27:27,067][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.07292576879262924, 'train/grad_norm_pre_clip': 46.30521774291992, 'timing/data': 0.0009934369882103056, 'timing/model': 1.2251525270112325, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.19}) +[2026-06-23 17:27:31,609][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.23059850931167603, 'train/grad_norm_pre_clip': 101.37481689453125, 'timing/data': 0.0007474620069842786, 'timing/model': 1.2265485289972275, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.19}) +[2026-06-23 17:27:36,158][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.10691443085670471, 'train/grad_norm_pre_clip': 70.14387512207031, 'timing/data': 0.0010606810101307929, 'timing/model': 1.2245354490005411, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.19}) +[2026-06-23 17:27:40,830][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.5916287899017334, 'train/grad_norm_pre_clip': 106.90631866455078, 'timing/data': 0.0009000950085464865, 'timing/model': 1.3899226930225268, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.19}) +[2026-06-23 17:27:45,373][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.1617898941040039, 'train/grad_norm_pre_clip': 67.22351837158203, 'timing/data': 0.000739947980036959, 'timing/model': 1.2276722109818365, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.19}) +[2026-06-23 17:27:49,941][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.22252708673477173, 'train/grad_norm_pre_clip': 69.37837219238281, 'timing/data': 0.0007220949919428676, 'timing/model': 1.2430508430115879, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.19}) +[2026-06-23 17:27:54,521][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.10838047415018082, 'train/grad_norm_pre_clip': 28.07166290283203, 'timing/data': 0.0008932719938457012, 'timing/model': 1.2603885470016394, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.19}) +[2026-06-23 17:27:59,301][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.1856006532907486, 'train/grad_norm_pre_clip': 37.652137756347656, 'timing/data': 0.0007799819868523628, 'timing/model': 1.4621664709993638, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.19}) +[2026-06-23 17:28:03,862][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.09625370800495148, 'train/grad_norm_pre_clip': 104.69457244873047, 'timing/data': 0.0014084940194152296, 'timing/model': 1.2437147349992301, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.19}) +[2026-06-23 17:28:08,414][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.20415988564491272, 'train/grad_norm_pre_clip': 69.4844970703125, 'timing/data': 0.0007248900074046105, 'timing/model': 1.2449917269987054, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.19}) +[2026-06-23 17:28:12,979][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.06915032118558884, 'train/grad_norm_pre_clip': 14.775973320007324, 'timing/data': 0.0007654250075574964, 'timing/model': 1.2410271299886517, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.19}) +[2026-06-23 17:28:17,758][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.18564295768737793, 'train/grad_norm_pre_clip': 59.693328857421875, 'timing/data': 0.0008596999978180975, 'timing/model': 1.2454492669785395, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.19}) +[2026-06-23 17:28:22,318][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.11042813211679459, 'train/grad_norm_pre_clip': 50.513675689697266, 'timing/data': 0.001113187987357378, 'timing/model': 1.236236277007265, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.19}) +[2026-06-23 17:28:26,873][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 500 +[2026-06-23 17:29:25,843][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 500 +[2026-06-23 17:29:25,843][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 500 +[2026-06-23 17:30:08,806][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 500 +[2026-06-23 17:39:35,055][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-23 17:39:35,057][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.20060297846794128, 'train/grad_norm_pre_clip': 110.65890502929688, 'eval/action_loss/seconds': 58.96595409497968, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.08343077450990677, 'eval/latency_2/loss': 0.08343077450990677, 'eval/flappy/loss': 0.08343077450990677, 'eval/flappy/latency_2/loss': 0.08343077450990677, 'eval/flap_precision': 0.8997395833333334, 'eval/flap_recall': 0.937584803256445, 'eval/flap_f1': 0.9182724252491694, 'eval/flap_rate_model': 0.21333333333333335, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9182724252491694, 'eval/latency_2/flap_precision': 0.8997395833333334, 'eval/latency_2/flap_recall': 0.937584803256445, 'eval/latency_2/flap_f1': 0.9182724252491694, 'eval/latency_2/flap_rate_model': 0.21333333333333335, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9182724252491694, 'eval/action_classification/data_wait_seconds': 0.4899207353591919, 'eval/action_classification/predict_seconds': 42.32561111450195, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 146.1799999999996, 'rl_games_eval/mid_train/mean_length': 1197.4, 'rl_games_eval/mid_train/std_reward': 110.17210899315654, 'rl_games_eval/mid_train/std_length': 889.6151077853838, 'rl_games_eval/mid_train/macro_mean_reward': 146.1799999999996, 'rl_games_eval/mid_train/macro_mean_length': 1197.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 146.1799999999996, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1197.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 110.17210899315654, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 889.6151077853838, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 146.1799999999996, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0007403279887512326, 'timing/model': 1.249722100008512, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.19}) +[2026-06-23 17:39:35,058][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_500_state +[2026-06-23 17:39:35,064][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 17:39:43,998][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_500_state/model.safetensors +[2026-06-23 17:39:54,994][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_500_state/optimizer.bin +[2026-06-23 17:39:54,995][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_500_state/sampler.bin +[2026-06-23 17:39:54,995][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_500_state/sampler_1.bin +[2026-06-23 17:39:54,997][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-23 17:39:54,999][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-23 17:39:54,999][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 17:39:55,023][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 17:39:59,264][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.2633635401725769, 'train/grad_norm_pre_clip': 20.423429489135742, 'timing/data': 0.0011353289883118123, 'timing/model': 1.169758631993318, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.19}) +[2026-06-23 17:40:03,688][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.19959892332553864, 'train/grad_norm_pre_clip': 16.452436447143555, 'timing/data': 0.0007703940209466964, 'timing/model': 1.3872953139944002, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.19}) +[2026-06-23 17:40:07,922][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.3438967168331146, 'train/grad_norm_pre_clip': 54.46784973144531, 'timing/data': 0.0007781579915899783, 'timing/model': 1.1656493940099608, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.19}) +[2026-06-23 17:40:12,163][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.03866181895136833, 'train/grad_norm_pre_clip': 26.106657028198242, 'timing/data': 0.0007321930024772882, 'timing/model': 1.1582659679988865, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.2}) +[2026-06-23 17:40:16,402][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.2741039991378784, 'train/grad_norm_pre_clip': 44.578399658203125, 'timing/data': 0.0007262619910761714, 'timing/model': 1.1637969969888218, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.2}) +[2026-06-23 17:40:20,814][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.11148035526275635, 'train/grad_norm_pre_clip': 50.84585189819336, 'timing/data': 0.0007215529913082719, 'timing/model': 1.1512643639871385, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.2}) +[2026-06-23 17:40:25,025][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.4356089234352112, 'train/grad_norm_pre_clip': 96.50424194335938, 'timing/data': 0.0007399170135613531, 'timing/model': 1.1486155570019037, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.2}) +[2026-06-23 17:40:29,296][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.0777459591627121, 'train/grad_norm_pre_clip': 39.59354782104492, 'timing/data': 0.0007790399831719697, 'timing/model': 1.1764938480046112, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.2}) +[2026-06-23 17:40:33,552][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.09290314465761185, 'train/grad_norm_pre_clip': 35.92074203491211, 'timing/data': 0.0007394870044663548, 'timing/model': 1.1764714540040586, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.2}) +[2026-06-23 17:40:38,034][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.10085523873567581, 'train/grad_norm_pre_clip': 35.30851364135742, 'timing/data': 0.0007390950049739331, 'timing/model': 1.4069399129948579, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.2}) +[2026-06-23 17:40:42,317][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.1744534969329834, 'train/grad_norm_pre_clip': 28.79877471923828, 'timing/data': 0.0008501909906044602, 'timing/model': 1.175489065994043, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.2}) +[2026-06-23 17:40:46,578][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.26184797286987305, 'train/grad_norm_pre_clip': 32.191993713378906, 'timing/data': 0.0007534220058005303, 'timing/model': 1.1599281249800697, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.2}) +[2026-06-23 17:40:50,877][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.07610267400741577, 'train/grad_norm_pre_clip': 19.47545623779297, 'timing/data': 0.00101728001027368, 'timing/model': 1.1739753450092394, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.2}) +[2026-06-23 17:40:55,307][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.17806567251682281, 'train/grad_norm_pre_clip': 26.993560791015625, 'timing/data': 0.0008274599967990071, 'timing/model': 1.1611635659937747, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.2}) +[2026-06-23 17:40:59,608][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.2704978287220001, 'train/grad_norm_pre_clip': 56.89675521850586, 'timing/data': 0.0009367710154037923, 'timing/model': 1.184563098999206, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.2}) +[2026-06-23 17:41:03,916][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.09972117841243744, 'train/grad_norm_pre_clip': 86.61781311035156, 'timing/data': 0.0009612360154278576, 'timing/model': 1.1896809060126543, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.2}) +[2026-06-23 17:41:08,248][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.136504665017128, 'train/grad_norm_pre_clip': 56.981048583984375, 'timing/data': 0.0010574450134299695, 'timing/model': 1.1817593940068036, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.2}) +[2026-06-23 17:41:12,827][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.12749932706356049, 'train/grad_norm_pre_clip': 31.23477554321289, 'timing/data': 0.001117225008783862, 'timing/model': 1.4472921220003627, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.2}) +[2026-06-23 17:41:17,119][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.10935832560062408, 'train/grad_norm_pre_clip': 26.133174896240234, 'timing/data': 0.0010972880118060857, 'timing/model': 1.1674672270019073, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.2}) +[2026-06-23 17:41:21,441][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.3042815625667572, 'train/grad_norm_pre_clip': 54.04779815673828, 'timing/data': 0.0007740809815004468, 'timing/model': 1.1905817429942545, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.2}) +[2026-06-23 17:41:25,767][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.1833130419254303, 'train/grad_norm_pre_clip': 61.114810943603516, 'timing/data': 0.0009285359992645681, 'timing/model': 1.1928828249801882, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.2}) +[2026-06-23 17:41:30,274][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.17941640317440033, 'train/grad_norm_pre_clip': 36.682334899902344, 'timing/data': 0.0009407579782418907, 'timing/model': 1.1714613040094264, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.2}) +[2026-06-23 17:41:34,588][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.1387285739183426, 'train/grad_norm_pre_clip': 21.78141975402832, 'timing/data': 0.0007713960076216608, 'timing/model': 1.192761998012429, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.2}) +[2026-06-23 17:41:38,914][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.05459894984960556, 'train/grad_norm_pre_clip': 48.66706085205078, 'timing/data': 0.0007043119985610247, 'timing/model': 1.1984242009930313, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.2}) +[2026-06-23 17:41:43,276][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.0941271111369133, 'train/grad_norm_pre_clip': 22.84107208251953, 'timing/data': 0.001081910013454035, 'timing/model': 1.1954543809988536, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.2}) +[2026-06-23 17:41:47,894][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.20482690632343292, 'train/grad_norm_pre_clip': 28.68912696838379, 'timing/data': 0.0007853720162529498, 'timing/model': 1.4450393930019345, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.2}) +[2026-06-23 17:41:52,212][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.19821365177631378, 'train/grad_norm_pre_clip': 49.990318298339844, 'timing/data': 0.0007432129932567477, 'timing/model': 1.1763858570193406, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.2}) +[2026-06-23 17:41:56,562][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.1449190378189087, 'train/grad_norm_pre_clip': 45.08780288696289, 'timing/data': 0.0011926039878744632, 'timing/model': 1.194016780005768, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.2}) +[2026-06-23 17:42:00,912][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.05199476704001427, 'train/grad_norm_pre_clip': 17.439411163330078, 'timing/data': 0.000803956005256623, 'timing/model': 1.1772433200094383, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.2}) +[2026-06-23 17:42:05,518][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.09358788281679153, 'train/grad_norm_pre_clip': 45.18144989013672, 'timing/data': 0.0008794250024948269, 'timing/model': 1.448721471009776, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.21}) +[2026-06-23 17:42:09,858][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.21821433305740356, 'train/grad_norm_pre_clip': 39.83598709106445, 'timing/data': 0.00122936099069193, 'timing/model': 1.178498096007388, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.21}) +[2026-06-23 17:42:14,247][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.17483286559581757, 'train/grad_norm_pre_clip': 41.79322814941406, 'timing/data': 0.0007534229953307658, 'timing/model': 1.2001792520168237, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.21}) +[2026-06-23 17:42:18,608][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.04411016032099724, 'train/grad_norm_pre_clip': 20.592737197875977, 'timing/data': 0.0007273840019479394, 'timing/model': 1.1842991930025164, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.21}) +[2026-06-23 17:42:23,260][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.14865779876708984, 'train/grad_norm_pre_clip': 19.846961975097656, 'timing/data': 0.0007303800084628165, 'timing/model': 1.4672805320005864, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.21}) +[2026-06-23 17:42:27,681][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.10589384287595749, 'train/grad_norm_pre_clip': 39.1318244934082, 'timing/data': 0.0009836079843807966, 'timing/model': 1.182332957017934, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.21}) +[2026-06-23 17:42:32,012][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.08803591877222061, 'train/grad_norm_pre_clip': 50.801414489746094, 'timing/data': 0.0007688600162509829, 'timing/model': 1.181403267022688, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.21}) +[2026-06-23 17:42:36,363][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.12914921343326569, 'train/grad_norm_pre_clip': 42.70591354370117, 'timing/data': 0.0007473209989257157, 'timing/model': 1.1841827020107303, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.21}) +[2026-06-23 17:42:40,986][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.09793926030397415, 'train/grad_norm_pre_clip': 43.675472259521484, 'timing/data': 0.0007256009848788381, 'timing/model': 1.4442128150258213, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.21}) +[2026-06-23 17:42:45,364][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.16219595074653625, 'train/grad_norm_pre_clip': 25.90774154663086, 'timing/data': 0.0008323379734065384, 'timing/model': 1.2017284049943555, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.21}) +[2026-06-23 17:42:49,716][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.1359439343214035, 'train/grad_norm_pre_clip': 27.799718856811523, 'timing/data': 0.0007403380004689097, 'timing/model': 1.2035785220214166, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.21}) +[2026-06-23 17:42:54,096][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.11977007985115051, 'train/grad_norm_pre_clip': 21.201885223388672, 'timing/data': 0.0007415710133500397, 'timing/model': 1.1856703259982169, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.21}) +[2026-06-23 17:42:58,712][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.04915076121687889, 'train/grad_norm_pre_clip': 19.69484519958496, 'timing/data': 0.0008869300072547048, 'timing/model': 1.440523088996997, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.21}) +[2026-06-23 17:43:03,049][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.1901046484708786, 'train/grad_norm_pre_clip': 50.867977142333984, 'timing/data': 0.000752610998461023, 'timing/model': 1.1855506520078052, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.21}) +[2026-06-23 17:43:07,403][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.08559124916791916, 'train/grad_norm_pre_clip': 47.598323822021484, 'timing/data': 0.0007765749760437757, 'timing/model': 1.1831818929931615, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.21}) +[2026-06-23 17:43:11,772][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.1192428469657898, 'train/grad_norm_pre_clip': 21.32589340209961, 'timing/data': 0.0009262620005756617, 'timing/model': 1.1945935919939075, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.21}) +[2026-06-23 17:43:16,349][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.11468596011400223, 'train/grad_norm_pre_clip': 33.99202346801758, 'timing/data': 0.0009793600183911622, 'timing/model': 1.198943199007772, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.21}) +[2026-06-23 17:43:20,735][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.3131140172481537, 'train/grad_norm_pre_clip': 26.31844139099121, 'timing/data': 0.0010444899962749332, 'timing/model': 1.1850736049818806, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.21}) +[2026-06-23 17:43:25,091][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.11293888837099075, 'train/grad_norm_pre_clip': 25.67525863647461, 'timing/data': 0.0009326139988843352, 'timing/model': 1.184739457006799, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.21}) +[2026-06-23 17:43:29,663][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.07573480159044266, 'train/grad_norm_pre_clip': 62.21942138671875, 'timing/data': 0.0011154910025652498, 'timing/model': 1.1866496859875042, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.21}) +[2026-06-23 17:43:34,303][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.017106957733631134, 'train/grad_norm_pre_clip': 26.759037017822266, 'timing/data': 0.0007596340146847069, 'timing/model': 1.4397732610232197, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.21}) +[2026-06-23 17:43:38,646][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.10682996362447739, 'train/grad_norm_pre_clip': 35.12953186035156, 'timing/data': 0.0007306000043172389, 'timing/model': 1.1835646920080762, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.21}) +[2026-06-23 17:43:43,010][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.016348503530025482, 'train/grad_norm_pre_clip': 13.925311088562012, 'timing/data': 0.0007133180042728782, 'timing/model': 1.1841875220125075, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.21}) +[2026-06-23 17:43:47,360][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.14494860172271729, 'train/grad_norm_pre_clip': 32.813079833984375, 'timing/data': 0.0007548050198238343, 'timing/model': 1.191619545017602, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.21}) +[2026-06-23 17:43:52,021][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.4298950433731079, 'train/grad_norm_pre_clip': 82.76969909667969, 'timing/data': 0.0014891820028424263, 'timing/model': 1.4751063930161763, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.21}) +[2026-06-23 17:43:56,428][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.12171366810798645, 'train/grad_norm_pre_clip': 58.358604431152344, 'timing/data': 0.000772197003243491, 'timing/model': 1.2064694559958298, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.21}) +[2026-06-23 17:44:00,818][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.03713890537619591, 'train/grad_norm_pre_clip': 17.96601104736328, 'timing/data': 0.0007985660049598664, 'timing/model': 1.1879222889838275, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.22}) +[2026-06-23 17:44:05,177][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.21223635971546173, 'train/grad_norm_pre_clip': 73.11827087402344, 'timing/data': 0.000836015009554103, 'timing/model': 1.1906780809804332, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.22}) +[2026-06-23 17:44:09,856][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.25517538189888, 'train/grad_norm_pre_clip': 107.94894409179688, 'timing/data': 0.0007346679922193289, 'timing/model': 1.4696757399942726, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.22}) +[2026-06-23 17:44:14,266][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.5346922874450684, 'train/grad_norm_pre_clip': 133.67506408691406, 'timing/data': 0.0010922890214715153, 'timing/model': 1.2070939149998594, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.22}) +[2026-06-23 17:44:18,653][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.09466048330068588, 'train/grad_norm_pre_clip': 77.84065246582031, 'timing/data': 0.0011039400123991072, 'timing/model': 1.2138918570126407, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.22}) +[2026-06-23 17:44:23,015][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.17439883947372437, 'train/grad_norm_pre_clip': 20.43061065673828, 'timing/data': 0.0007362409960478544, 'timing/model': 1.1948621589981485, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.22}) +[2026-06-23 17:44:27,574][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.06704720854759216, 'train/grad_norm_pre_clip': 44.13756561279297, 'timing/data': 0.0009815630037337542, 'timing/model': 1.1889701990003232, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.22}) +[2026-06-23 17:44:31,981][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.129638209939003, 'train/grad_norm_pre_clip': 21.62440299987793, 'timing/data': 0.0007798419974278659, 'timing/model': 1.2060288630018476, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.22}) +[2026-06-23 17:44:36,389][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.1824207454919815, 'train/grad_norm_pre_clip': 25.359840393066406, 'timing/data': 0.0007353690161835402, 'timing/model': 1.201540676993318, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.22}) +[2026-06-23 17:44:40,782][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.10060854256153107, 'train/grad_norm_pre_clip': 30.90363311767578, 'timing/data': 0.0007179959793575108, 'timing/model': 1.2081496699829586, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.22}) +[2026-06-23 17:44:45,419][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.1508888602256775, 'train/grad_norm_pre_clip': 37.33230972290039, 'timing/data': 0.000913548021344468, 'timing/model': 1.4551872439915314, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.22}) +[2026-06-23 17:44:49,812][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.057875510305166245, 'train/grad_norm_pre_clip': 16.54170799255371, 'timing/data': 0.0007522499945480376, 'timing/model': 1.2084672850032803, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.22}) +[2026-06-23 17:44:54,182][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.06268197298049927, 'train/grad_norm_pre_clip': 25.141504287719727, 'timing/data': 0.0008153369999490678, 'timing/model': 1.1876954670005944, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.22}) +[2026-06-23 17:44:58,564][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.036883432418107986, 'train/grad_norm_pre_clip': 54.144752502441406, 'timing/data': 0.0007491240103263408, 'timing/model': 1.217291128996294, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.22}) +[2026-06-23 17:45:03,189][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.048486411571502686, 'train/grad_norm_pre_clip': 29.942296981811523, 'timing/data': 0.0007188890012912452, 'timing/model': 1.4342254810035229, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.22}) +[2026-06-23 17:45:07,575][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.07246680557727814, 'train/grad_norm_pre_clip': 40.71989822387695, 'timing/data': 0.0008964470180217177, 'timing/model': 1.2131739089963958, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.22}) +[2026-06-23 17:45:11,971][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.06788096576929092, 'train/grad_norm_pre_clip': 32.8009147644043, 'timing/data': 0.0007819849997758865, 'timing/model': 1.1961543350189459, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.22}) +[2026-06-23 17:45:16,346][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.1948123425245285, 'train/grad_norm_pre_clip': 58.38797378540039, 'timing/data': 0.0007245589804369956, 'timing/model': 1.1924124979996122, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.22}) +[2026-06-23 17:45:20,994][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.008014526218175888, 'train/grad_norm_pre_clip': 70.1731185913086, 'timing/data': 0.0010440099867992103, 'timing/model': 1.4725811190146487, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.22}) +[2026-06-23 17:45:25,420][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.03997053578495979, 'train/grad_norm_pre_clip': 52.2229118347168, 'timing/data': 0.0012402920110616833, 'timing/model': 1.2167714230017737, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.22}) +[2026-06-23 17:45:29,799][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.21824179589748383, 'train/grad_norm_pre_clip': 76.23682403564453, 'timing/data': 0.0007975540065672249, 'timing/model': 1.1902902240108233, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.22}) +[2026-06-23 17:45:34,178][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.035704389214515686, 'train/grad_norm_pre_clip': 72.61322021484375, 'timing/data': 0.0007549449801445007, 'timing/model': 1.2081688129983377, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.22}) +[2026-06-23 17:45:38,786][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.14043401181697845, 'train/grad_norm_pre_clip': 61.87274932861328, 'timing/data': 0.0009024869941640645, 'timing/model': 1.4356944159953855, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.22}) +[2026-06-23 17:45:43,141][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.015798987820744514, 'train/grad_norm_pre_clip': 47.366878509521484, 'timing/data': 0.0008760490163695067, 'timing/model': 1.1886630530061666, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.22}) +[2026-06-23 17:45:47,514][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.07967539131641388, 'train/grad_norm_pre_clip': 48.91322708129883, 'timing/data': 0.0007425119983963668, 'timing/model': 1.1997578559967224, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.22}) +[2026-06-23 17:45:51,897][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.0683666542172432, 'train/grad_norm_pre_clip': 36.51310348510742, 'timing/data': 0.0007496659818571061, 'timing/model': 1.2102345470048022, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.22}) +[2026-06-23 17:45:56,526][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.04504294693470001, 'train/grad_norm_pre_clip': 45.16506576538086, 'timing/data': 0.0007960520160850137, 'timing/model': 1.443488744989736, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.23}) +[2026-06-23 17:46:00,886][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.1564619094133377, 'train/grad_norm_pre_clip': 48.66117477416992, 'timing/data': 0.0007426819938700646, 'timing/model': 1.1883670530223753, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.23}) +[2026-06-23 17:46:05,251][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.13028371334075928, 'train/grad_norm_pre_clip': 22.560293197631836, 'timing/data': 0.0007903309888206422, 'timing/model': 1.188943136978196, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.23}) +[2026-06-23 17:46:09,617][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.08941953629255295, 'train/grad_norm_pre_clip': 40.3323860168457, 'timing/data': 0.0007437950116582215, 'timing/model': 1.1921388109913096, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.23}) +[2026-06-23 17:46:14,269][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.10364020615816116, 'train/grad_norm_pre_clip': 26.591238021850586, 'timing/data': 0.00084483198588714, 'timing/model': 1.4614937450096477, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.23}) +[2026-06-23 17:46:18,650][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.07253772020339966, 'train/grad_norm_pre_clip': 45.92023849487305, 'timing/data': 0.0009130270045716316, 'timing/model': 1.1905870320042595, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.23}) +[2026-06-23 17:46:23,031][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.07522089779376984, 'train/grad_norm_pre_clip': 38.25526428222656, 'timing/data': 0.0008786440012045205, 'timing/model': 1.2081568199791946, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.23}) +[2026-06-23 17:46:27,504][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.5007022619247437, 'train/grad_norm_pre_clip': 27.80254364013672, 'timing/data': 0.0008901349792722613, 'timing/model': 1.2072273119993042, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.23}) +[2026-06-23 17:46:32,074][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.2555367946624756, 'train/grad_norm_pre_clip': 40.03048324584961, 'timing/data': 0.0007731589721515775, 'timing/model': 1.2206829560163897, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.23}) +[2026-06-23 17:46:36,515][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.18486645817756653, 'train/grad_norm_pre_clip': 60.564144134521484, 'timing/data': 0.000924307998502627, 'timing/model': 1.2146903719985858, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.23}) +[2026-06-23 17:46:40,889][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.051337648183107376, 'train/grad_norm_pre_clip': 27.799745559692383, 'timing/data': 0.0007557270000688732, 'timing/model': 1.1932910149917006, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.23}) +[2026-06-23 17:46:45,366][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.023743271827697754, 'train/grad_norm_pre_clip': 15.929844856262207, 'timing/data': 0.0008924489957280457, 'timing/model': 1.2759306939842645, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.23}) +[2026-06-23 17:46:50,020][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.2989380955696106, 'train/grad_norm_pre_clip': 33.43193054199219, 'timing/data': 0.000948934000916779, 'timing/model': 1.440165509993676, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.23}) +[2026-06-23 17:46:54,406][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.11756239086389542, 'train/grad_norm_pre_clip': 15.761343002319336, 'timing/data': 0.0010404520144220442, 'timing/model': 1.1925831029948313, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.23}) +[2026-06-23 17:46:58,880][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.013571403920650482, 'train/grad_norm_pre_clip': 17.418174743652344, 'timing/data': 0.0007559369842056185, 'timing/model': 1.2093098140030634, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.23}) +[2026-06-23 17:47:03,305][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.05474679544568062, 'train/grad_norm_pre_clip': 29.721435546875, 'timing/data': 0.0012136730074416846, 'timing/model': 1.2124537829949986, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.23}) +[2026-06-23 17:47:07,997][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.06654761731624603, 'train/grad_norm_pre_clip': 25.663097381591797, 'timing/data': 0.0007529920258093625, 'timing/model': 1.4698914810142014, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.23}) +[2026-06-23 17:47:12,414][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.05665251612663269, 'train/grad_norm_pre_clip': 34.64363098144531, 'timing/data': 0.000809075019787997, 'timing/model': 1.209192886017263, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.23}) +[2026-06-23 17:47:16,818][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 600 +[2026-06-23 17:48:14,224][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 600 +[2026-06-23 17:48:14,225][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 600 +[2026-06-23 17:48:51,285][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 600 +[2026-06-23 17:48:51,287][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.27536657452583313, 'train/grad_norm_pre_clip': 33.117828369140625, 'eval/action_loss/seconds': 57.40341071700095, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.08486880362033844, 'eval/latency_2/loss': 0.08486880362033844, 'eval/flappy/loss': 0.08486880362033844, 'eval/flappy/latency_2/loss': 0.08486880362033844, 'eval/flap_precision': 0.9677914110429447, 'eval/flap_recall': 0.8561736770691994, 'eval/flap_f1': 0.9085673146148308, 'eval/flap_rate_model': 0.1811111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9085673146148308, 'eval/latency_2/flap_precision': 0.9677914110429447, 'eval/latency_2/flap_recall': 0.8561736770691994, 'eval/latency_2/flap_f1': 0.9085673146148308, 'eval/latency_2/flap_rate_model': 0.1811111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9085673146148308, 'eval/action_classification/data_wait_seconds': 0.4701348543167114, 'eval/action_classification/predict_seconds': 36.45417404174805, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007874360017012805, 'timing/model': 1.2192959650128614, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.23}) +[2026-06-23 17:48:55,810][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.061748720705509186, 'train/grad_norm_pre_clip': 39.964927673339844, 'timing/data': 0.0011202399909961969, 'timing/model': 1.3770988379837945, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.23}) +[2026-06-23 17:49:00,169][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.17378270626068115, 'train/grad_norm_pre_clip': 61.95576477050781, 'timing/data': 0.0008145049796439707, 'timing/model': 1.2036874030018225, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.23}) +[2026-06-23 17:49:04,525][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.2318199723958969, 'train/grad_norm_pre_clip': 36.44736099243164, 'timing/data': 0.0010150559828616679, 'timing/model': 1.207977481011767, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.23}) +[2026-06-23 17:49:08,888][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.11767016351222992, 'train/grad_norm_pre_clip': 29.535236358642578, 'timing/data': 0.0007252010109368712, 'timing/model': 1.1995863529737107, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.23}) +[2026-06-23 17:49:13,246][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.21891078352928162, 'train/grad_norm_pre_clip': 55.287899017333984, 'timing/data': 0.0007469010015483946, 'timing/model': 1.1939531120005995, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.23}) +[2026-06-23 17:49:17,808][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.0850750282406807, 'train/grad_norm_pre_clip': 18.35291290283203, 'timing/data': 0.0007481130014639348, 'timing/model': 1.1762262029806152, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.23}) +[2026-06-23 17:49:22,145][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.012490786612033844, 'train/grad_norm_pre_clip': 28.716524124145508, 'timing/data': 0.0007384750060737133, 'timing/model': 1.1786357749952003, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.23}) +[2026-06-23 17:49:26,481][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.1566147804260254, 'train/grad_norm_pre_clip': 33.445350646972656, 'timing/data': 0.000774360989453271, 'timing/model': 1.1989380329905543, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.24}) +[2026-06-23 17:49:30,801][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.09923511743545532, 'train/grad_norm_pre_clip': 49.92293930053711, 'timing/data': 0.0007383049814961851, 'timing/model': 1.182726372993784, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.24}) +[2026-06-23 17:49:35,420][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.15059907734394073, 'train/grad_norm_pre_clip': 35.690311431884766, 'timing/data': 0.0008002289978321642, 'timing/model': 1.2119793820020277, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.24}) +[2026-06-23 17:49:39,801][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.18759053945541382, 'train/grad_norm_pre_clip': 48.756553649902344, 'timing/data': 0.0007298490090761334, 'timing/model': 1.2092169859970454, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.24}) +[2026-06-23 17:49:44,173][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.08129775524139404, 'train/grad_norm_pre_clip': 26.189777374267578, 'timing/data': 0.0007394370040856302, 'timing/model': 1.1989039670152124, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.24}) +[2026-06-23 17:49:48,572][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.1774294525384903, 'train/grad_norm_pre_clip': 21.40412712097168, 'timing/data': 0.0007480920176021755, 'timing/model': 1.2107071690261364, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.24}) +[2026-06-23 17:49:53,119][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.08305366337299347, 'train/grad_norm_pre_clip': 24.408199310302734, 'timing/data': 0.0007565380074083805, 'timing/model': 1.2036323540087324, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.24}) +[2026-06-23 17:49:57,536][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.08716540038585663, 'train/grad_norm_pre_clip': 20.58209991455078, 'timing/data': 0.0007993580074980855, 'timing/model': 1.1931384009949397, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.24}) +[2026-06-23 17:50:01,952][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.14400339126586914, 'train/grad_norm_pre_clip': 33.012935638427734, 'timing/data': 0.0013105729885865003, 'timing/model': 1.211784657993121, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.24}) +[2026-06-23 17:50:06,332][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.22653785347938538, 'train/grad_norm_pre_clip': 35.27894592285156, 'timing/data': 0.0009291580063290894, 'timing/model': 1.2203575599996839, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.24}) +[2026-06-23 17:50:10,953][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.22690515220165253, 'train/grad_norm_pre_clip': 25.29549789428711, 'timing/data': 0.0008384090033359826, 'timing/model': 1.214480057999026, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.24}) +[2026-06-23 17:50:15,327][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.03955567255616188, 'train/grad_norm_pre_clip': 19.543657302856445, 'timing/data': 0.000759193004341796, 'timing/model': 1.1857953180151526, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.24}) +[2026-06-23 17:50:19,720][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.01963481493294239, 'train/grad_norm_pre_clip': 20.61836814880371, 'timing/data': 0.0008844039984978735, 'timing/model': 1.2061173229885753, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.24}) +[2026-06-23 17:50:24,113][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.06436868011951447, 'train/grad_norm_pre_clip': 42.445465087890625, 'timing/data': 0.000827679003123194, 'timing/model': 1.2077659089991357, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.24}) +[2026-06-23 17:50:28,807][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.11894994974136353, 'train/grad_norm_pre_clip': 35.93441390991211, 'timing/data': 0.0007968230056576431, 'timing/model': 1.2175464220053982, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.24}) +[2026-06-23 17:50:33,204][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.08779586106538773, 'train/grad_norm_pre_clip': 29.770469665527344, 'timing/data': 0.0007392769912257791, 'timing/model': 1.2058079889975488, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.24}) +[2026-06-23 17:50:37,579][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.08759547024965286, 'train/grad_norm_pre_clip': 49.57060623168945, 'timing/data': 0.0012122000043746084, 'timing/model': 1.1854100539931096, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.24}) +[2026-06-23 17:50:41,968][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.14285218715667725, 'train/grad_norm_pre_clip': 52.458282470703125, 'timing/data': 0.0009163939976133406, 'timing/model': 1.1858278349973261, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.24}) +[2026-06-23 17:50:46,625][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.29229772090911865, 'train/grad_norm_pre_clip': 28.075679779052734, 'timing/data': 0.0009036599949467927, 'timing/model': 1.1818899270147085, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.24}) +[2026-06-23 17:50:51,004][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.09654116630554199, 'train/grad_norm_pre_clip': 15.097908020019531, 'timing/data': 0.0007920529751572758, 'timing/model': 1.2040624810033478, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.24}) +[2026-06-23 17:50:55,380][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.19127050042152405, 'train/grad_norm_pre_clip': 42.768035888671875, 'timing/data': 0.0010163779952563345, 'timing/model': 1.1902552859974094, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.24}) +[2026-06-23 17:50:59,750][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.0511602908372879, 'train/grad_norm_pre_clip': 28.990205764770508, 'timing/data': 0.0009423520241398364, 'timing/model': 1.1869122969801538, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.24}) +[2026-06-23 17:51:04,385][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.2702849209308624, 'train/grad_norm_pre_clip': 36.32098388671875, 'timing/data': 0.0008545889868400991, 'timing/model': 1.216896728001302, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.24}) +[2026-06-23 17:51:08,806][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.16861677169799805, 'train/grad_norm_pre_clip': 35.52674102783203, 'timing/data': 0.000758481997763738, 'timing/model': 1.20902932799072, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.24}) +[2026-06-23 17:51:13,195][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.16380411386489868, 'train/grad_norm_pre_clip': 32.60343551635742, 'timing/data': 0.0009416699758730829, 'timing/model': 1.1924814770172816, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.24}) +[2026-06-23 17:51:17,603][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.13671064376831055, 'train/grad_norm_pre_clip': 27.52524185180664, 'timing/data': 0.000743464013794437, 'timing/model': 1.1866499409952667, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.24}) +[2026-06-23 17:51:22,108][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.09257907420396805, 'train/grad_norm_pre_clip': 19.427953720092773, 'timing/data': 0.0007653650245629251, 'timing/model': 1.1867071760061663, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.25}) +[2026-06-23 17:51:26,480][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.1859769970178604, 'train/grad_norm_pre_clip': 48.16802215576172, 'timing/data': 0.0008094160002656281, 'timing/model': 1.189584551990265, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.25}) +[2026-06-23 17:51:30,878][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.13511618971824646, 'train/grad_norm_pre_clip': 37.72813415527344, 'timing/data': 0.0007199409883469343, 'timing/model': 1.2145102140202653, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.25}) +[2026-06-23 17:51:35,236][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.05256665125489235, 'train/grad_norm_pre_clip': 46.051517486572266, 'timing/data': 0.0010781119926832616, 'timing/model': 1.186585228977492, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.25}) +[2026-06-23 17:51:39,860][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.022151032462716103, 'train/grad_norm_pre_clip': 19.617645263671875, 'timing/data': 0.0009352589841000736, 'timing/model': 1.1859377209912054, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.25}) +[2026-06-23 17:51:44,233][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.16630445420742035, 'train/grad_norm_pre_clip': 21.067129135131836, 'timing/data': 0.0007350080122705549, 'timing/model': 1.1907468689896632, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.25}) +[2026-06-23 17:51:48,610][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.22750970721244812, 'train/grad_norm_pre_clip': 47.9399299621582, 'timing/data': 0.0007439640176016837, 'timing/model': 1.1894391420064494, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.25}) +[2026-06-23 17:51:53,007][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.06987270712852478, 'train/grad_norm_pre_clip': 67.67477416992188, 'timing/data': 0.0007604949933011085, 'timing/model': 1.201143881015014, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.25}) +[2026-06-23 17:51:57,665][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.05069230869412422, 'train/grad_norm_pre_clip': 23.506206512451172, 'timing/data': 0.0009037299896590412, 'timing/model': 1.2048819280171301, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.25}) +[2026-06-23 17:52:02,074][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.11496365070343018, 'train/grad_norm_pre_clip': 45.23009490966797, 'timing/data': 0.0007092099986039102, 'timing/model': 1.2104673989815637, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.25}) +[2026-06-23 17:52:06,464][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.03073911741375923, 'train/grad_norm_pre_clip': 33.809173583984375, 'timing/data': 0.0007620879914611578, 'timing/model': 1.1878195359895471, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.25}) +[2026-06-23 17:52:10,850][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.15165166556835175, 'train/grad_norm_pre_clip': 23.34939956665039, 'timing/data': 0.0007765349873807281, 'timing/model': 1.207025338022504, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.25}) +[2026-06-23 17:52:15,473][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.02745877578854561, 'train/grad_norm_pre_clip': 15.782829284667969, 'timing/data': 0.0007311209919862449, 'timing/model': 1.188495876005618, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.25}) +[2026-06-23 17:52:19,886][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.1176348328590393, 'train/grad_norm_pre_clip': 43.02891159057617, 'timing/data': 0.0007446770032402128, 'timing/model': 1.2061049960029777, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.25}) +[2026-06-23 17:52:24,308][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.25713032484054565, 'train/grad_norm_pre_clip': 55.8500862121582, 'timing/data': 0.0007660659903194755, 'timing/model': 1.2078832020051777, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.25}) +[2026-06-23 17:52:28,668][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.3592998683452606, 'train/grad_norm_pre_clip': 27.94135284423828, 'timing/data': 0.0007594240014441311, 'timing/model': 1.1895865100086667, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.25}) +[2026-06-23 17:52:33,311][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.19825157523155212, 'train/grad_norm_pre_clip': 70.00422668457031, 'timing/data': 0.001100222987588495, 'timing/model': 1.212052927003242, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.25}) +[2026-06-23 17:52:37,718][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.17610391974449158, 'train/grad_norm_pre_clip': 47.63322830200195, 'timing/data': 0.0007396969886031002, 'timing/model': 1.2038927259854972, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.25}) +[2026-06-23 17:52:42,124][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.14027097821235657, 'train/grad_norm_pre_clip': 22.76811981201172, 'timing/data': 0.0010989010042976588, 'timing/model': 1.19079747601063, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.25}) +[2026-06-23 17:52:46,518][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.14182217419147491, 'train/grad_norm_pre_clip': 32.124290466308594, 'timing/data': 0.000784820003900677, 'timing/model': 1.2070367760024965, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.25}) +[2026-06-23 17:52:51,051][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.06482411175966263, 'train/grad_norm_pre_clip': 25.15305519104004, 'timing/data': 0.0007088099955581129, 'timing/model': 1.2071600230119657, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.25}) +[2026-06-23 17:52:55,456][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.08615148812532425, 'train/grad_norm_pre_clip': 37.56521987915039, 'timing/data': 0.0008826209814287722, 'timing/model': 1.2100906780106016, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.25}) +[2026-06-23 17:52:59,867][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.18451867997646332, 'train/grad_norm_pre_clip': 25.870525360107422, 'timing/data': 0.0007461289933416992, 'timing/model': 1.2121903280203696, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.25}) +[2026-06-23 17:53:04,271][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.061041176319122314, 'train/grad_norm_pre_clip': 22.65059471130371, 'timing/data': 0.0007539629878010601, 'timing/model': 1.2092691189900506, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.25}) +[2026-06-23 17:53:08,899][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.13111117482185364, 'train/grad_norm_pre_clip': 28.565710067749023, 'timing/data': 0.000920270977076143, 'timing/model': 1.1870476439944468, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.25}) +[2026-06-23 17:53:13,329][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.12008361518383026, 'train/grad_norm_pre_clip': 23.65635871887207, 'timing/data': 0.0007980150112416595, 'timing/model': 1.2154371270153206, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.26}) +[2026-06-23 17:53:17,726][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.2536131739616394, 'train/grad_norm_pre_clip': 16.104755401611328, 'timing/data': 0.000775292981415987, 'timing/model': 1.209091199998511, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.26}) +[2026-06-23 17:53:22,154][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.15319746732711792, 'train/grad_norm_pre_clip': 33.5224723815918, 'timing/data': 0.0007328749925363809, 'timing/model': 1.2078541629889514, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.26}) +[2026-06-23 17:53:26,782][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.17025405168533325, 'train/grad_norm_pre_clip': 26.527315139770508, 'timing/data': 0.0009376929956488311, 'timing/model': 1.1873484779789578, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.26}) +[2026-06-23 17:53:31,193][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.13686782121658325, 'train/grad_norm_pre_clip': 32.270286560058594, 'timing/data': 0.0007076069887261838, 'timing/model': 1.2037160439940635, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.26}) +[2026-06-23 17:53:35,591][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.15065665543079376, 'train/grad_norm_pre_clip': 29.594989776611328, 'timing/data': 0.0007234170043375343, 'timing/model': 1.190409875009209, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.26}) +[2026-06-23 17:53:39,995][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.1519913673400879, 'train/grad_norm_pre_clip': 29.72637939453125, 'timing/data': 0.0011321819911245257, 'timing/model': 1.1843087500019465, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.26}) +[2026-06-23 17:53:44,605][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.21258385479450226, 'train/grad_norm_pre_clip': 66.46957397460938, 'timing/data': 0.0007276350224856287, 'timing/model': 1.1836887340177782, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.26}) +[2026-06-23 17:53:48,991][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.06742893159389496, 'train/grad_norm_pre_clip': 53.54324722290039, 'timing/data': 0.0009008249908220023, 'timing/model': 1.2076829239958897, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.26}) +[2026-06-23 17:53:53,375][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.06477441638708115, 'train/grad_norm_pre_clip': 32.2318229675293, 'timing/data': 0.0009258909849449992, 'timing/model': 1.2068206389958505, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.26}) +[2026-06-23 17:53:57,793][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.07429368793964386, 'train/grad_norm_pre_clip': 36.48968505859375, 'timing/data': 0.0007194800127763301, 'timing/model': 1.2046863459981978, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.26}) +[2026-06-23 17:54:02,429][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.14791688323020935, 'train/grad_norm_pre_clip': 31.061532974243164, 'timing/data': 0.001199537015054375, 'timing/model': 1.1906160139769781, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.26}) +[2026-06-23 17:54:06,853][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.12787894904613495, 'train/grad_norm_pre_clip': 58.56958770751953, 'timing/data': 0.0007998679939191788, 'timing/model': 1.2140708869847003, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.26}) +[2026-06-23 17:54:11,232][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.11931732296943665, 'train/grad_norm_pre_clip': 18.772397994995117, 'timing/data': 0.0007836880104150623, 'timing/model': 1.1905672730063088, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.26}) +[2026-06-23 17:54:15,656][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.16664108633995056, 'train/grad_norm_pre_clip': 53.31552505493164, 'timing/data': 0.0007411800033878535, 'timing/model': 1.2093399339937605, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.26}) +[2026-06-23 17:54:20,290][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.11703869700431824, 'train/grad_norm_pre_clip': 28.951168060302734, 'timing/data': 0.0011982540017925203, 'timing/model': 1.189193703001365, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.26}) +[2026-06-23 17:54:24,659][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.21414917707443237, 'train/grad_norm_pre_clip': 23.754064559936523, 'timing/data': 0.0007447359967045486, 'timing/model': 1.2034087329811882, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.26}) +[2026-06-23 17:54:29,028][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.1052146852016449, 'train/grad_norm_pre_clip': 27.45530891418457, 'timing/data': 0.0007637809903826565, 'timing/model': 1.1873873459990136, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.26}) +[2026-06-23 17:54:33,397][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.1395944207906723, 'train/grad_norm_pre_clip': 49.135162353515625, 'timing/data': 0.0007164740236476064, 'timing/model': 1.2095507519843522, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.26}) +[2026-06-23 17:54:38,002][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.10949090868234634, 'train/grad_norm_pre_clip': 19.18385887145996, 'timing/data': 0.0007357900030910969, 'timing/model': 1.2025458670104854, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.26}) +[2026-06-23 17:54:42,398][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.10315749794244766, 'train/grad_norm_pre_clip': 15.525046348571777, 'timing/data': 0.0007376839930657297, 'timing/model': 1.2103656230028719, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.26}) +[2026-06-23 17:54:46,799][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.06393269449472427, 'train/grad_norm_pre_clip': 52.637027740478516, 'timing/data': 0.0007936479814816266, 'timing/model': 1.214284095010953, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.26}) +[2026-06-23 17:54:51,200][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.1143290251493454, 'train/grad_norm_pre_clip': 23.559490203857422, 'timing/data': 0.0007585129933431745, 'timing/model': 1.2086790659814142, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.26}) +[2026-06-23 17:54:55,854][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.11168551445007324, 'train/grad_norm_pre_clip': 28.796354293823242, 'timing/data': 0.0007704849995207042, 'timing/model': 1.2052420899854042, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.26}) +[2026-06-23 17:55:00,226][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.13442784547805786, 'train/grad_norm_pre_clip': 27.826200485229492, 'timing/data': 0.0007299899880308658, 'timing/model': 1.1838455400138628, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.26}) +[2026-06-23 17:55:04,646][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.28083133697509766, 'train/grad_norm_pre_clip': 53.62032699584961, 'timing/data': 0.0007956519839353859, 'timing/model': 1.2139248379971832, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.26}) +[2026-06-23 17:55:08,989][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.2954237461090088, 'train/grad_norm_pre_clip': 38.01065444946289, 'timing/data': 0.000929989997530356, 'timing/model': 1.1854855910059996, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.27}) +[2026-06-23 17:55:13,632][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.11943038552999496, 'train/grad_norm_pre_clip': 34.619056701660156, 'timing/data': 0.0009695839835330844, 'timing/model': 1.1908337560016662, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.27}) +[2026-06-23 17:55:18,000][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.06226785108447075, 'train/grad_norm_pre_clip': 23.71505355834961, 'timing/data': 0.0007820570026524365, 'timing/model': 1.186755604983773, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.27}) +[2026-06-23 17:55:22,416][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.20041592419147491, 'train/grad_norm_pre_clip': 39.93054962158203, 'timing/data': 0.0009077490249183029, 'timing/model': 1.1972128440102097, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.27}) +[2026-06-23 17:55:26,789][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.11481215059757233, 'train/grad_norm_pre_clip': 24.913496017456055, 'timing/data': 0.0007859340112190694, 'timing/model': 1.1862535440013744, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.27}) +[2026-06-23 17:55:31,422][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.014623859897255898, 'train/grad_norm_pre_clip': 17.17814826965332, 'timing/data': 0.0008945339941419661, 'timing/model': 1.186963761021616, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.27}) +[2026-06-23 17:55:35,836][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.11703194677829742, 'train/grad_norm_pre_clip': 24.7958927154541, 'timing/data': 0.0007190699980128556, 'timing/model': 1.2197517360036727, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.27}) +[2026-06-23 17:55:40,215][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.072788767516613, 'train/grad_norm_pre_clip': 29.721006393432617, 'timing/data': 0.0009144019859377295, 'timing/model': 1.193848925002385, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.27}) +[2026-06-23 17:55:44,598][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.2553795576095581, 'train/grad_norm_pre_clip': 23.26210594177246, 'timing/data': 0.0008061109983827919, 'timing/model': 1.1873648749897256, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.27}) +[2026-06-23 17:55:49,309][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.04066423326730728, 'train/grad_norm_pre_clip': 13.922514915466309, 'timing/data': 0.000961378013016656, 'timing/model': 1.188213157991413, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.27}) +[2026-06-23 17:55:53,765][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.024841932579874992, 'train/grad_norm_pre_clip': 54.76626205444336, 'timing/data': 0.0007288579945452511, 'timing/model': 1.2120788680040278, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.27}) +[2026-06-23 17:55:58,187][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.057553730905056, 'train/grad_norm_pre_clip': 21.91141128540039, 'timing/data': 0.0007920849893707782, 'timing/model': 1.2042273679981008, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.27}) +[2026-06-23 17:56:02,532][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.10928155481815338, 'train/grad_norm_pre_clip': 19.6367130279541, 'timing/data': 0.0006976799922995269, 'timing/model': 1.1907779150060378, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.27}) +[2026-06-23 17:56:07,126][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.06446991115808487, 'train/grad_norm_pre_clip': 25.50669288635254, 'timing/data': 0.0008678149897605181, 'timing/model': 1.1897179179941304, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.27}) +[2026-06-23 17:56:11,462][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.26375845074653625, 'train/grad_norm_pre_clip': 42.44404602050781, 'timing/data': 0.0007149320153985173, 'timing/model': 1.18323783800588, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.27}) +[2026-06-23 17:56:15,819][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 700 +[2026-06-23 17:57:12,660][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 700 +[2026-06-23 17:57:12,660][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 700 +[2026-06-23 17:57:49,679][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 700 +[2026-06-23 17:57:49,681][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.03117343969643116, 'train/grad_norm_pre_clip': 29.47344970703125, 'eval/action_loss/seconds': 56.8376743660192, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.0631805956363678, 'eval/latency_2/loss': 0.0631805956363678, 'eval/flappy/loss': 0.0631805956363678, 'eval/flappy/latency_2/loss': 0.0631805956363678, 'eval/flap_precision': 0.9448924731182796, 'eval/flap_recall': 0.9538670284938942, 'eval/flap_f1': 0.949358541525996, 'eval/flap_rate_model': 0.20666666666666667, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.949358541525996, 'eval/latency_2/flap_precision': 0.9448924731182796, 'eval/latency_2/flap_recall': 0.9538670284938942, 'eval/latency_2/flap_f1': 0.949358541525996, 'eval/latency_2/flap_rate_model': 0.20666666666666667, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.949358541525996, 'eval/action_classification/data_wait_seconds': 0.4296343922615051, 'eval/action_classification/predict_seconds': 36.448150634765625, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0008845360134728253, 'timing/model': 1.1944220500008669, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.27}) +[2026-06-23 17:57:54,297][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.07858376950025558, 'train/grad_norm_pre_clip': 21.752429962158203, 'timing/data': 0.0010442209895700216, 'timing/model': 1.440907842013985, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.27}) +[2026-06-23 17:57:58,631][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.17014633119106293, 'train/grad_norm_pre_clip': 51.05705642700195, 'timing/data': 0.0010687070025596768, 'timing/model': 1.1915278030210175, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.27}) +[2026-06-23 17:58:02,972][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.18632772564888, 'train/grad_norm_pre_clip': 54.08123016357422, 'timing/data': 0.0007209739997051656, 'timing/model': 1.1980475499876775, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.27}) +[2026-06-23 17:58:07,325][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.26533517241477966, 'train/grad_norm_pre_clip': 26.31003189086914, 'timing/data': 0.0007656159868929535, 'timing/model': 1.1938127049943432, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.27}) +[2026-06-23 17:58:11,915][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.048752523958683014, 'train/grad_norm_pre_clip': 37.334007263183594, 'timing/data': 0.0007675699889659882, 'timing/model': 1.4430472610110883, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.27}) +[2026-06-23 17:58:16,327][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.16656258702278137, 'train/grad_norm_pre_clip': 35.07780075073242, 'timing/data': 0.0008159989956766367, 'timing/model': 1.2112256240216084, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.27}) +[2026-06-23 17:58:20,722][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.06472399085760117, 'train/grad_norm_pre_clip': 47.74289321899414, 'timing/data': 0.0008074130164459348, 'timing/model': 1.2088197029952426, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.27}) +[2026-06-23 17:58:25,113][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.1318172961473465, 'train/grad_norm_pre_clip': 45.262298583984375, 'timing/data': 0.0009630809945520014, 'timing/model': 1.2002282519824803, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.27}) +[2026-06-23 17:58:29,685][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.13799291849136353, 'train/grad_norm_pre_clip': 28.6846981048584, 'timing/data': 0.0009044219914358109, 'timing/model': 1.176409791019978, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.27}) +[2026-06-23 17:58:34,022][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.2580421268939972, 'train/grad_norm_pre_clip': 58.846893310546875, 'timing/data': 0.0007823969935998321, 'timing/model': 1.194607454992365, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.27}) +[2026-06-23 17:58:38,410][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.22523033618927002, 'train/grad_norm_pre_clip': 62.4090576171875, 'timing/data': 0.000808955985121429, 'timing/model': 1.197783748008078, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.28}) +[2026-06-23 17:58:42,772][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.1492350548505783, 'train/grad_norm_pre_clip': 39.367340087890625, 'timing/data': 0.00076721899677068, 'timing/model': 1.1965822349884547, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.28}) +[2026-06-23 17:58:47,439][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.10390692949295044, 'train/grad_norm_pre_clip': 35.36460494995117, 'timing/data': 0.0010630259930621833, 'timing/model': 1.4599616299965419, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.28}) +[2026-06-23 17:58:51,800][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.0412217453122139, 'train/grad_norm_pre_clip': 13.862380981445312, 'timing/data': 0.0010285220050718635, 'timing/model': 1.1837938920070883, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.28}) +[2026-06-23 17:58:56,220][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.17059093713760376, 'train/grad_norm_pre_clip': 30.23170280456543, 'timing/data': 0.0009428340126760304, 'timing/model': 1.2155214850208722, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.28}) +[2026-06-23 17:59:00,619][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.18811821937561035, 'train/grad_norm_pre_clip': 23.31827163696289, 'timing/data': 0.0007941880030557513, 'timing/model': 1.2075626160076354, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.28}) +[2026-06-23 17:59:05,253][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.18323178589344025, 'train/grad_norm_pre_clip': 15.066482543945312, 'timing/data': 0.0008005000127013773, 'timing/model': 1.2112915799953043, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.28}) +[2026-06-23 17:59:09,666][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.13933292031288147, 'train/grad_norm_pre_clip': 17.879743576049805, 'timing/data': 0.0008371979929506779, 'timing/model': 1.2081084029923659, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.28}) +[2026-06-23 17:59:14,014][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.1310604065656662, 'train/grad_norm_pre_clip': 17.430768966674805, 'timing/data': 0.000976695999270305, 'timing/model': 1.1821062170201913, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.28}) +[2026-06-23 17:59:18,383][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.11889620125293732, 'train/grad_norm_pre_clip': 75.20687103271484, 'timing/data': 0.0009831080096773803, 'timing/model': 1.1843636219855398, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.28}) +[2026-06-23 17:59:23,070][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.07556238025426865, 'train/grad_norm_pre_clip': 29.814712524414062, 'timing/data': 0.0007449980184901506, 'timing/model': 1.2120109679817688, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.28}) +[2026-06-23 17:59:27,460][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.17853426933288574, 'train/grad_norm_pre_clip': 24.153793334960938, 'timing/data': 0.0009404790180269629, 'timing/model': 1.1901015620096587, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.28}) +[2026-06-23 17:59:31,803][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.16909189522266388, 'train/grad_norm_pre_clip': 28.095441818237305, 'timing/data': 0.0009276359924115241, 'timing/model': 1.1827640419942327, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.28}) +[2026-06-23 17:59:36,186][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.08463702350854874, 'train/grad_norm_pre_clip': 35.45331573486328, 'timing/data': 0.0007781790045555681, 'timing/model': 1.2059137509786524, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.28}) +[2026-06-23 17:59:40,767][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.11207935959100723, 'train/grad_norm_pre_clip': 16.808481216430664, 'timing/data': 0.000732133979909122, 'timing/model': 1.1860160760115832, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.28}) +[2026-06-23 17:59:45,106][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.14867465198040009, 'train/grad_norm_pre_clip': 10.92354679107666, 'timing/data': 0.0007939280185382813, 'timing/model': 1.184959512000205, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.28}) +[2026-06-23 17:59:49,528][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.08583386242389679, 'train/grad_norm_pre_clip': 48.68788528442383, 'timing/data': 0.0008236129942815751, 'timing/model': 1.2044652129989117, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.28}) +[2026-06-23 17:59:53,925][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.24825851619243622, 'train/grad_norm_pre_clip': 53.393402099609375, 'timing/data': 0.0008063209825195372, 'timing/model': 1.2082431399903726, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.28}) +[2026-06-23 17:59:58,532][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.059695977717638016, 'train/grad_norm_pre_clip': 22.474781036376953, 'timing/data': 0.001149735995568335, 'timing/model': 1.2134392379957717, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.28}) +[2026-06-23 18:00:02,933][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.026082953438162804, 'train/grad_norm_pre_clip': 15.216172218322754, 'timing/data': 0.0007759949949104339, 'timing/model': 1.2064686589874327, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.28}) +[2026-06-23 18:00:07,336][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.13273115456104279, 'train/grad_norm_pre_clip': 41.34154510498047, 'timing/data': 0.0010013110004365444, 'timing/model': 1.1907197649998125, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.28}) +[2026-06-23 18:00:11,724][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.10843215137720108, 'train/grad_norm_pre_clip': 25.617767333984375, 'timing/data': 0.0009593440045136958, 'timing/model': 1.2165774420136586, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.28}) +[2026-06-23 18:00:16,319][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.07664527744054794, 'train/grad_norm_pre_clip': 55.85232162475586, 'timing/data': 0.0008327199902851135, 'timing/model': 1.2175009700004011, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.28}) +[2026-06-23 18:00:20,734][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.33226490020751953, 'train/grad_norm_pre_clip': 66.65108489990234, 'timing/data': 0.0010375389829277992, 'timing/model': 1.1924282359832432, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.28}) +[2026-06-23 18:00:25,127][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.3595312237739563, 'train/grad_norm_pre_clip': 33.289180755615234, 'timing/data': 0.0009844910237006843, 'timing/model': 1.209074017999228, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.28}) +[2026-06-23 18:00:29,530][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.22577662765979767, 'train/grad_norm_pre_clip': 25.777233123779297, 'timing/data': 0.000858588027767837, 'timing/model': 1.2042064080014825, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.28}) +[2026-06-23 18:00:34,168][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.07145573198795319, 'train/grad_norm_pre_clip': 58.972190856933594, 'timing/data': 0.0011091409833170474, 'timing/model': 1.2095041640277486, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.29}) +[2026-06-23 18:00:38,570][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.2816319465637207, 'train/grad_norm_pre_clip': 62.488224029541016, 'timing/data': 0.0007349899969995022, 'timing/model': 1.1901416220061947, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.29}) +[2026-06-23 18:00:42,967][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.03547133132815361, 'train/grad_norm_pre_clip': 15.320058822631836, 'timing/data': 0.0009490950033068657, 'timing/model': 1.2188516519963741, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.29}) +[2026-06-23 18:00:47,371][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.2405463308095932, 'train/grad_norm_pre_clip': 21.73781967163086, 'timing/data': 0.000775634020101279, 'timing/model': 1.220713776012417, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.29}) +[2026-06-23 18:00:51,987][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.23216070234775543, 'train/grad_norm_pre_clip': 51.5649299621582, 'timing/data': 0.000989018997643143, 'timing/model': 1.2245771799935028, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.29}) +[2026-06-23 18:00:56,391][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.21467773616313934, 'train/grad_norm_pre_clip': 42.165225982666016, 'timing/data': 0.0008140249992720783, 'timing/model': 1.2042981230188161, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.29}) +[2026-06-23 18:01:00,778][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.167795330286026, 'train/grad_norm_pre_clip': 17.257102966308594, 'timing/data': 0.0008084260043688118, 'timing/model': 1.1910390670236666, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.29}) +[2026-06-23 18:01:05,192][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.26605087518692017, 'train/grad_norm_pre_clip': 32.04779052734375, 'timing/data': 0.0007342479948420078, 'timing/model': 1.2199601360189263, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.29}) +[2026-06-23 18:01:09,781][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.10884260386228561, 'train/grad_norm_pre_clip': 28.453746795654297, 'timing/data': 0.0007350290252361447, 'timing/model': 1.1920322570076678, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.29}) +[2026-06-23 18:01:14,185][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.16213953495025635, 'train/grad_norm_pre_clip': 41.89580154418945, 'timing/data': 0.0007392769912257791, 'timing/model': 1.2219652670028154, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.29}) +[2026-06-23 18:01:18,594][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.2029588222503662, 'train/grad_norm_pre_clip': 63.74418258666992, 'timing/data': 0.001011520012980327, 'timing/model': 1.1979027240013238, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.29}) +[2026-06-23 18:01:23,029][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.14536137878894806, 'train/grad_norm_pre_clip': 32.60470199584961, 'timing/data': 0.0009910530061461031, 'timing/model': 1.2124731079966296, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.29}) +[2026-06-23 18:01:27,713][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.1398608535528183, 'train/grad_norm_pre_clip': 21.677223205566406, 'timing/data': 0.0011072870111092925, 'timing/model': 1.211791378998896, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.29}) +[2026-06-23 18:01:55,996][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 750 +[2026-06-23 18:01:55,997][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.06735069304704666, 'train/grad_norm_pre_clip': 32.0640754699707, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/mean_length': 50.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/macro_mean_length': 50.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.8999999999999986, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 50.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 3.8999999999999986, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0008089460025075823, 'timing/model': 1.2098728030105121, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.29}) +[2026-06-23 18:02:00,266][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.2238527238368988, 'train/grad_norm_pre_clip': 39.761409759521484, 'timing/data': 0.0008875710191205144, 'timing/model': 1.1661127700062934, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.29}) +[2026-06-23 18:02:04,592][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.2510870695114136, 'train/grad_norm_pre_clip': 36.734405517578125, 'timing/data': 0.0008099469996523112, 'timing/model': 1.1908607009972911, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.29}) +[2026-06-23 18:02:09,115][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.10853277891874313, 'train/grad_norm_pre_clip': 11.710119247436523, 'timing/data': 0.0007628099992871284, 'timing/model': 1.1969888250168879, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.29}) +[2026-06-23 18:02:13,440][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.1468278169631958, 'train/grad_norm_pre_clip': 62.67567825317383, 'timing/data': 0.0008161890145856887, 'timing/model': 1.1990080010145903, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.29}) +[2026-06-23 18:02:17,786][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.33068037033081055, 'train/grad_norm_pre_clip': 51.8473014831543, 'timing/data': 0.0011957099777646363, 'timing/model': 1.2020897319889627, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.29}) +[2026-06-23 18:02:22,123][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.2926150858402252, 'train/grad_norm_pre_clip': 60.76218032836914, 'timing/data': 0.0010928100091405213, 'timing/model': 1.1777652089949697, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.29}) +[2026-06-23 18:02:26,671][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.11858227849006653, 'train/grad_norm_pre_clip': 29.875579833984375, 'timing/data': 0.0008039459935389459, 'timing/model': 1.2143683200119995, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.29}) +[2026-06-23 18:02:31,004][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.16749432682991028, 'train/grad_norm_pre_clip': 19.03326988220215, 'timing/data': 0.0007877059979364276, 'timing/model': 1.1822541959991213, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.29}) +[2026-06-23 18:02:35,397][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.16635310649871826, 'train/grad_norm_pre_clip': 49.28800582885742, 'timing/data': 0.0008364570094272494, 'timing/model': 1.2022238109784666, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.29}) +[2026-06-23 18:02:39,738][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.0685323178768158, 'train/grad_norm_pre_clip': 14.797757148742676, 'timing/data': 0.0011374930036254227, 'timing/model': 1.1919888900010847, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.29}) +[2026-06-23 18:02:44,403][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.06913939863443375, 'train/grad_norm_pre_clip': 26.578258514404297, 'timing/data': 0.0010025039955507964, 'timing/model': 1.2042897469946183, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.29}) +[2026-06-23 18:02:48,793][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.20187319815158844, 'train/grad_norm_pre_clip': 38.00102233886719, 'timing/data': 0.0008705689979251474, 'timing/model': 1.1848266829911154, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.29}) +[2026-06-23 18:02:53,171][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.11478601396083832, 'train/grad_norm_pre_clip': 17.841394424438477, 'timing/data': 0.0007412199920509011, 'timing/model': 1.1827242890140042, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.3}) +[2026-06-23 18:02:57,527][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.06337573379278183, 'train/grad_norm_pre_clip': 22.069236755371094, 'timing/data': 0.0007594839844387025, 'timing/model': 1.2035782909952104, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.3}) +[2026-06-23 18:03:02,131][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.22151532769203186, 'train/grad_norm_pre_clip': 22.902658462524414, 'timing/data': 0.0009364310244563967, 'timing/model': 1.186193078989163, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.3}) +[2026-06-23 18:03:06,489][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.02865018881857395, 'train/grad_norm_pre_clip': 26.730228424072266, 'timing/data': 0.0007929159910418093, 'timing/model': 1.1828037060040515, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.3}) +[2026-06-23 18:03:10,919][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.0551590770483017, 'train/grad_norm_pre_clip': 19.442829132080078, 'timing/data': 0.0010452730057295412, 'timing/model': 1.210386668011779, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.3}) +[2026-06-23 18:03:15,507][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.1531885415315628, 'train/grad_norm_pre_clip': 21.030729293823242, 'timing/data': 0.0008316780149471015, 'timing/model': 1.3868484899867326, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.3}) +[2026-06-23 18:03:19,901][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.08848307281732559, 'train/grad_norm_pre_clip': 35.19243621826172, 'timing/data': 0.0007569800072815269, 'timing/model': 1.2096358859853353, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.3}) +[2026-06-23 18:03:24,296][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.09338399767875671, 'train/grad_norm_pre_clip': 18.14067840576172, 'timing/data': 0.0008315280138049275, 'timing/model': 1.2136866390064824, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.3}) +[2026-06-23 18:03:28,699][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.21947352588176727, 'train/grad_norm_pre_clip': 45.354278564453125, 'timing/data': 0.0007970439910423011, 'timing/model': 1.2048047010030132, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.3}) +[2026-06-23 18:03:33,237][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.27372440695762634, 'train/grad_norm_pre_clip': 42.7939453125, 'timing/data': 0.0008506820013280958, 'timing/model': 1.188811383995926, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.3}) +[2026-06-23 18:03:37,631][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.12629440426826477, 'train/grad_norm_pre_clip': 39.89128112792969, 'timing/data': 0.000917436002055183, 'timing/model': 1.1874763099767733, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.3}) +[2026-06-23 18:03:42,023][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.07686017453670502, 'train/grad_norm_pre_clip': 23.72174072265625, 'timing/data': 0.0008024730195757002, 'timing/model': 1.2054344060015865, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.3}) +[2026-06-23 18:03:46,422][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.05952592194080353, 'train/grad_norm_pre_clip': 17.262680053710938, 'timing/data': 0.0008871399913914502, 'timing/model': 1.2082650039810687, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.3}) +[2026-06-23 18:03:51,064][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.04422811418771744, 'train/grad_norm_pre_clip': 15.278254508972168, 'timing/data': 0.0007385549833998084, 'timing/model': 1.45775041799061, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.3}) +[2026-06-23 18:03:55,443][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.13738858699798584, 'train/grad_norm_pre_clip': 23.159330368041992, 'timing/data': 0.0007268440094776452, 'timing/model': 1.208144076023018, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.3}) +[2026-06-23 18:03:59,859][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.11716427654027939, 'train/grad_norm_pre_clip': 24.079622268676758, 'timing/data': 0.000837658008094877, 'timing/model': 1.192717237980105, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.3}) +[2026-06-23 18:04:04,277][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.0688805878162384, 'train/grad_norm_pre_clip': 15.8198823928833, 'timing/data': 0.0012741060054395348, 'timing/model': 1.2199640670150984, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.3}) +[2026-06-23 18:04:08,903][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.24821637570858002, 'train/grad_norm_pre_clip': 47.22944641113281, 'timing/data': 0.000801151996711269, 'timing/model': 1.1923294900043402, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.3}) +[2026-06-23 18:04:13,273][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.04738443344831467, 'train/grad_norm_pre_clip': 16.097291946411133, 'timing/data': 0.001064157986547798, 'timing/model': 1.1937604389968328, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.3}) +[2026-06-23 18:04:17,664][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.03280772268772125, 'train/grad_norm_pre_clip': 31.713848114013672, 'timing/data': 0.0007503480010200292, 'timing/model': 1.220860948000336, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.3}) +[2026-06-23 18:04:22,090][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.022986801341176033, 'train/grad_norm_pre_clip': 15.041062355041504, 'timing/data': 0.0009162640199065208, 'timing/model': 1.2192517969815526, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.3}) +[2026-06-23 18:04:26,698][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.1293511837720871, 'train/grad_norm_pre_clip': 31.607675552368164, 'timing/data': 0.00098876902484335, 'timing/model': 1.1900402029859833, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.3}) +[2026-06-23 18:04:31,142][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.06032620742917061, 'train/grad_norm_pre_clip': 18.540769577026367, 'timing/data': 0.0009237379999831319, 'timing/model': 1.21599712697207, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.3}) +[2026-06-23 18:04:35,526][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.17792104184627533, 'train/grad_norm_pre_clip': 46.849666595458984, 'timing/data': 0.0009052440000232309, 'timing/model': 1.1931819510064088, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.3}) +[2026-06-23 18:04:39,881][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.016713736578822136, 'train/grad_norm_pre_clip': 28.949003219604492, 'timing/data': 0.0008502819982822984, 'timing/model': 1.1837013979966287, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.3}) +[2026-06-23 18:04:44,451][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.25999414920806885, 'train/grad_norm_pre_clip': 22.44715118408203, 'timing/data': 0.0011257709993515164, 'timing/model': 1.1855314850108698, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.3}) +[2026-06-23 18:04:48,818][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.15879251062870026, 'train/grad_norm_pre_clip': 29.580713272094727, 'timing/data': 0.0007384249765891582, 'timing/model': 1.2068483809998725, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.31}) +[2026-06-23 18:04:53,188][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.19030018150806427, 'train/grad_norm_pre_clip': 38.41783142089844, 'timing/data': 0.0007180380052886903, 'timing/model': 1.2072751939995214, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.31}) +[2026-06-23 18:04:57,538][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.18066546320915222, 'train/grad_norm_pre_clip': 44.635276794433594, 'timing/data': 0.0009498359868302941, 'timing/model': 1.1835582610219717, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.31}) +[2026-06-23 18:05:02,145][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.17751815915107727, 'train/grad_norm_pre_clip': 20.100017547607422, 'timing/data': 0.0007602050027344376, 'timing/model': 1.2035792100068647, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.31}) +[2026-06-23 18:05:06,566][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.15182310342788696, 'train/grad_norm_pre_clip': 44.5564079284668, 'timing/data': 0.0007414799765683711, 'timing/model': 1.2134445930132642, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.31}) +[2026-06-23 18:05:10,986][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.07290534675121307, 'train/grad_norm_pre_clip': 36.325782775878906, 'timing/data': 0.001132092991610989, 'timing/model': 1.2051167419995181, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.31}) +[2026-06-23 18:05:15,422][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.3674284815788269, 'train/grad_norm_pre_clip': 37.2207145690918, 'timing/data': 0.0007362120086327195, 'timing/model': 1.2053855739941355, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.31}) +[2026-06-23 18:05:19,981][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.11135085672140121, 'train/grad_norm_pre_clip': 31.148550033569336, 'timing/data': 0.0007236870005726814, 'timing/model': 1.186629873001948, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.31}) +[2026-06-23 18:05:24,357][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.13338689506053925, 'train/grad_norm_pre_clip': 14.558337211608887, 'timing/data': 0.00078527198638767, 'timing/model': 1.2112357089936268, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.31}) +[2026-06-23 18:05:28,718][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.12008781731128693, 'train/grad_norm_pre_clip': 22.2110538482666, 'timing/data': 0.0007067770056892186, 'timing/model': 1.1928519670036621, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.31}) +[2026-06-23 18:05:33,139][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.0807156190276146, 'train/grad_norm_pre_clip': 25.883195877075195, 'timing/data': 0.0007216939993668348, 'timing/model': 1.2100074170157313, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.31}) +[2026-06-23 18:05:37,714][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 800 +[2026-06-23 18:06:35,417][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 800 +[2026-06-23 18:06:35,418][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 800 +[2026-06-23 18:07:12,096][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 800 +[2026-06-23 18:07:12,098][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.178335502743721, 'train/grad_norm_pre_clip': 14.98318099975586, 'eval/action_loss/seconds': 57.69959972699871, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.07963237166404724, 'eval/latency_2/loss': 0.07963237166404724, 'eval/flappy/loss': 0.07963237166404724, 'eval/flappy/latency_2/loss': 0.07963237166404724, 'eval/flap_precision': 0.9688427299703264, 'eval/flap_recall': 0.8860244233378561, 'eval/flap_f1': 0.9255846917080085, 'eval/flap_rate_model': 0.18722222222222223, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9255846917080086, 'eval/latency_2/flap_precision': 0.9688427299703264, 'eval/latency_2/flap_recall': 0.8860244233378561, 'eval/latency_2/flap_f1': 0.9255846917080085, 'eval/latency_2/flap_rate_model': 0.18722222222222223, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9255846917080086, 'eval/action_classification/data_wait_seconds': 0.42808541655540466, 'eval/action_classification/predict_seconds': 36.108665466308594, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007028990075923502, 'timing/model': 1.2050139110069722, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.31}) +[2026-06-23 18:07:16,468][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.10234465450048447, 'train/grad_norm_pre_clip': 13.489237785339355, 'timing/data': 0.0007250810158438981, 'timing/model': 1.1989300819986966, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.31}) +[2026-06-23 18:07:20,842][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.3618500828742981, 'train/grad_norm_pre_clip': 56.23187255859375, 'timing/data': 0.0007907120161689818, 'timing/model': 1.2168477609811816, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.31}) +[2026-06-23 18:07:25,414][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.05892365425825119, 'train/grad_norm_pre_clip': 55.50784683227539, 'timing/data': 0.0008282510098069906, 'timing/model': 1.2035772840026766, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.31}) +[2026-06-23 18:07:29,751][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.1800043284893036, 'train/grad_norm_pre_clip': 61.539852142333984, 'timing/data': 0.0011333049915265292, 'timing/model': 1.198210900009144, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.31}) +[2026-06-23 18:07:34,136][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.06474895775318146, 'train/grad_norm_pre_clip': 21.949148178100586, 'timing/data': 0.000729488005163148, 'timing/model': 1.204397092020372, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.31}) +[2026-06-23 18:07:38,476][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.15784096717834473, 'train/grad_norm_pre_clip': 26.175949096679688, 'timing/data': 0.0009710049780551344, 'timing/model': 1.1792164990038145, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.31}) +[2026-06-23 18:07:43,090][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.036451660096645355, 'train/grad_norm_pre_clip': 35.12672805786133, 'timing/data': 0.0008622339810244739, 'timing/model': 1.199624180997489, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.31}) +[2026-06-23 18:07:47,463][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.0913272500038147, 'train/grad_norm_pre_clip': 58.66969680786133, 'timing/data': 0.0007388459926005453, 'timing/model': 1.1827889650012366, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.31}) +[2026-06-23 18:07:51,857][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.12403231114149094, 'train/grad_norm_pre_clip': 41.59238052368164, 'timing/data': 0.0009826270106714219, 'timing/model': 1.197137326002121, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.31}) +[2026-06-23 18:07:56,229][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.32422128319740295, 'train/grad_norm_pre_clip': 32.90689468383789, 'timing/data': 0.0008579760033171624, 'timing/model': 1.1823911319952458, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.31}) +[2026-06-23 18:08:00,802][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.17164671421051025, 'train/grad_norm_pre_clip': 37.744407653808594, 'timing/data': 0.0007230969786178321, 'timing/model': 1.1975912070192862, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.31}) +[2026-06-23 18:08:05,170][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.15907740592956543, 'train/grad_norm_pre_clip': 31.011384963989258, 'timing/data': 0.0009002739971037954, 'timing/model': 1.1871314870077185, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.31}) +[2026-06-23 18:08:09,536][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.10843809694051743, 'train/grad_norm_pre_clip': 86.38610076904297, 'timing/data': 0.0008088650065474212, 'timing/model': 1.2132056849950459, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.31}) +[2026-06-23 18:08:13,949][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.20509663224220276, 'train/grad_norm_pre_clip': 47.90237808227539, 'timing/data': 0.0007883079815655947, 'timing/model': 1.1996546810260043, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.32}) +[2026-06-23 18:08:18,513][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.15178357064723969, 'train/grad_norm_pre_clip': 41.268951416015625, 'timing/data': 0.0011588619963731617, 'timing/model': 1.1999208089837339, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.32}) +[2026-06-23 18:08:22,910][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.21683181822299957, 'train/grad_norm_pre_clip': 40.924468994140625, 'timing/data': 0.0007645230216439813, 'timing/model': 1.2123925700143445, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.32}) +[2026-06-23 18:08:27,297][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.0998731479048729, 'train/grad_norm_pre_clip': 47.4932861328125, 'timing/data': 0.0011361100187059492, 'timing/model': 1.2124594310007524, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.32}) +[2026-06-23 18:08:31,937][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.100395068526268, 'train/grad_norm_pre_clip': 61.08034133911133, 'timing/data': 0.0009557770099490881, 'timing/model': 1.4511014109884854, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.32}) +[2026-06-23 18:08:36,295][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.311281681060791, 'train/grad_norm_pre_clip': 47.14622497558594, 'timing/data': 0.0009242589876521379, 'timing/model': 1.187640944001032, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.32}) +[2026-06-23 18:08:40,661][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.14882177114486694, 'train/grad_norm_pre_clip': 23.40427589416504, 'timing/data': 0.0007528419955633581, 'timing/model': 1.1828839869995136, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.32}) +[2026-06-23 18:08:45,056][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.14450019598007202, 'train/grad_norm_pre_clip': 22.253620147705078, 'timing/data': 0.0009963519987650216, 'timing/model': 1.2019768489990383, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.32}) +[2026-06-23 18:08:49,689][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.059071898460388184, 'train/grad_norm_pre_clip': 24.205228805541992, 'timing/data': 0.0011034700146410614, 'timing/model': 1.4502888020069804, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.32}) +[2026-06-23 18:08:54,101][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.04527370259165764, 'train/grad_norm_pre_clip': 22.64015007019043, 'timing/data': 0.0008880310051608831, 'timing/model': 1.2121127379941754, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.32}) +[2026-06-23 18:08:58,501][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.12507714331150055, 'train/grad_norm_pre_clip': 14.118498802185059, 'timing/data': 0.0008543489966541529, 'timing/model': 1.2057774430140853, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.32}) +[2026-06-23 18:09:02,894][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.06829477846622467, 'train/grad_norm_pre_clip': 36.77553176879883, 'timing/data': 0.0011412099993322045, 'timing/model': 1.2042061960091814, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.32}) +[2026-06-23 18:09:07,445][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.06777764856815338, 'train/grad_norm_pre_clip': 39.2691535949707, 'timing/data': 0.0008073820208664984, 'timing/model': 1.200508437003009, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.32}) +[2026-06-23 18:09:11,835][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.07914523780345917, 'train/grad_norm_pre_clip': 20.7978572845459, 'timing/data': 0.0014361859939526767, 'timing/model': 1.2040954339900054, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.32}) +[2026-06-23 18:09:16,187][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.04515719413757324, 'train/grad_norm_pre_clip': 15.491109848022461, 'timing/data': 0.0010155469935853034, 'timing/model': 1.1866750479966868, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.32}) +[2026-06-23 18:09:20,587][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.03916826471686363, 'train/grad_norm_pre_clip': 33.760948181152344, 'timing/data': 0.0007894599984865636, 'timing/model': 1.1858067699940875, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.32}) +[2026-06-23 18:09:25,225][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.5442689061164856, 'train/grad_norm_pre_clip': 47.42224884033203, 'timing/data': 0.0007643529970664531, 'timing/model': 1.204525139008183, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.32}) +[2026-06-23 18:09:29,644][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.09414955973625183, 'train/grad_norm_pre_clip': 90.35762023925781, 'timing/data': 0.0009034300164785236, 'timing/model': 1.208759593020659, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.32}) +[2026-06-23 18:09:34,089][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.054307423532009125, 'train/grad_norm_pre_clip': 54.28763961791992, 'timing/data': 0.0009196299943141639, 'timing/model': 1.204642619995866, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.32}) +[2026-06-23 18:09:38,647][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.13516226410865784, 'train/grad_norm_pre_clip': 36.567012786865234, 'timing/data': 0.0007444560178555548, 'timing/model': 1.203072472999338, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.32}) +[2026-06-23 18:09:43,278][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.10372617095708847, 'train/grad_norm_pre_clip': 52.03339767456055, 'timing/data': 0.0007428739918395877, 'timing/model': 1.2094723209738731, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.32}) +[2026-06-23 18:09:47,683][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.19430950284004211, 'train/grad_norm_pre_clip': 34.352455139160156, 'timing/data': 0.0009663859964348376, 'timing/model': 1.2116854490013793, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.32}) +[2026-06-23 18:09:52,105][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.12610319256782532, 'train/grad_norm_pre_clip': 25.149864196777344, 'timing/data': 0.0007404990028589964, 'timing/model': 1.2104356449854095, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.32}) +[2026-06-23 18:09:56,541][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.17683593928813934, 'train/grad_norm_pre_clip': 17.547321319580078, 'timing/data': 0.0008005300187505782, 'timing/model': 1.2103053919854574, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.32}) +[2026-06-23 18:10:01,142][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.05175921693444252, 'train/grad_norm_pre_clip': 20.13296890258789, 'timing/data': 0.0007259419944602996, 'timing/model': 1.4079942459939048, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.32}) +[2026-06-23 18:10:05,532][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.08512650430202484, 'train/grad_norm_pre_clip': 14.082232475280762, 'timing/data': 0.0011515680234879255, 'timing/model': 1.1878667270066217, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.32}) +[2026-06-23 18:10:09,917][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.16089336574077606, 'train/grad_norm_pre_clip': 22.165206909179688, 'timing/data': 0.0007228359754662961, 'timing/model': 1.2105318710091524, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.33}) +[2026-06-23 18:10:14,342][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.08426494896411896, 'train/grad_norm_pre_clip': 18.933937072753906, 'timing/data': 0.000904020998859778, 'timing/model': 1.2253515830088872, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.33}) +[2026-06-23 18:10:18,978][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.06281370669603348, 'train/grad_norm_pre_clip': 12.131563186645508, 'timing/data': 0.0007491150172427297, 'timing/model': 1.4425356430001557, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.33}) +[2026-06-23 18:10:23,377][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.09278357028961182, 'train/grad_norm_pre_clip': 19.21449851989746, 'timing/data': 0.00087671002256684, 'timing/model': 1.19192017900059, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.33}) +[2026-06-23 18:10:27,777][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.2115667164325714, 'train/grad_norm_pre_clip': 21.925512313842773, 'timing/data': 0.0007171460019890219, 'timing/model': 1.1918654039909597, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.33}) +[2026-06-23 18:10:32,145][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.21732467412948608, 'train/grad_norm_pre_clip': 19.196928024291992, 'timing/data': 0.0008927609887905419, 'timing/model': 1.1895681839960162, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.33}) +[2026-06-23 18:10:36,712][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.08031835407018661, 'train/grad_norm_pre_clip': 24.378644943237305, 'timing/data': 0.0012438190169632435, 'timing/model': 1.1907138999958988, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.33}) +[2026-06-23 18:10:41,104][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.13417594134807587, 'train/grad_norm_pre_clip': 38.187744140625, 'timing/data': 0.00108979397919029, 'timing/model': 1.2150864400027785, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.33}) +[2026-06-23 18:10:45,491][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.1393166184425354, 'train/grad_norm_pre_clip': 57.7344856262207, 'timing/data': 0.0007644029974471778, 'timing/model': 1.1925069389981218, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.33}) +[2026-06-23 18:10:49,884][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.06503063440322876, 'train/grad_norm_pre_clip': 17.481691360473633, 'timing/data': 0.0009954400011338294, 'timing/model': 1.2172770340112038, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.33}) +[2026-06-23 18:10:54,565][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.043331146240234375, 'train/grad_norm_pre_clip': 25.917572021484375, 'timing/data': 0.0008479279931634665, 'timing/model': 1.2124010380066466, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.33}) +[2026-06-23 18:10:58,972][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.025357577949762344, 'train/grad_norm_pre_clip': 21.687217712402344, 'timing/data': 0.0009079680021386594, 'timing/model': 1.2125427669961937, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.33}) +[2026-06-23 18:11:03,373][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.06397709250450134, 'train/grad_norm_pre_clip': 20.33565330505371, 'timing/data': 0.0007317519921343774, 'timing/model': 1.2171203259786125, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.33}) +[2026-06-23 18:11:07,878][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.19685082137584686, 'train/grad_norm_pre_clip': 43.52436828613281, 'timing/data': 0.0008158879936672747, 'timing/model': 1.2105340590060223, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.33}) +[2026-06-23 18:11:12,510][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.16334985196590424, 'train/grad_norm_pre_clip': 34.52245330810547, 'timing/data': 0.000750406994484365, 'timing/model': 1.1914536859840155, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.33}) +[2026-06-23 18:11:16,921][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.062738336622715, 'train/grad_norm_pre_clip': 23.504446029663086, 'timing/data': 0.0007385349890682846, 'timing/model': 1.2110652849951293, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.33}) +[2026-06-23 18:11:21,357][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.19594475626945496, 'train/grad_norm_pre_clip': 38.136905670166016, 'timing/data': 0.0008609210199210793, 'timing/model': 1.2246872129908297, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.33}) +[2026-06-23 18:11:25,797][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.07689006626605988, 'train/grad_norm_pre_clip': 16.692058563232422, 'timing/data': 0.0009089600061997771, 'timing/model': 1.1925798199954443, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.33}) +[2026-06-23 18:11:30,447][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.015042580664157867, 'train/grad_norm_pre_clip': 22.355974197387695, 'timing/data': 0.0007230859773699194, 'timing/model': 1.2208885450090747, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.33}) +[2026-06-23 18:11:34,823][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.1591518074274063, 'train/grad_norm_pre_clip': 25.679222106933594, 'timing/data': 0.00084547302685678, 'timing/model': 1.2052397860097699, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.33}) +[2026-06-23 18:11:39,201][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.14801068603992462, 'train/grad_norm_pre_clip': 27.413742065429688, 'timing/data': 0.0007523410022258759, 'timing/model': 1.1926710400148295, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.33}) +[2026-06-23 18:11:43,596][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.1436585783958435, 'train/grad_norm_pre_clip': 34.30097579956055, 'timing/data': 0.0008387709967792034, 'timing/model': 1.2211393820180092, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.33}) +[2026-06-23 18:11:48,227][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.24901849031448364, 'train/grad_norm_pre_clip': 54.41978073120117, 'timing/data': 0.0008876310021150857, 'timing/model': 1.2106704270117916, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.33}) +[2026-06-23 18:11:52,595][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.09561378508806229, 'train/grad_norm_pre_clip': 43.35213088989258, 'timing/data': 0.000848569005029276, 'timing/model': 1.1924433620006312, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.33}) +[2026-06-23 18:11:57,006][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.10811091959476471, 'train/grad_norm_pre_clip': 20.47227668762207, 'timing/data': 0.0009902710153255612, 'timing/model': 1.2286411720269825, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.33}) +[2026-06-23 18:12:01,398][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.11094682663679123, 'train/grad_norm_pre_clip': 41.84978485107422, 'timing/data': 0.0007184580026660115, 'timing/model': 1.2165510210033972, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.33}) +[2026-06-23 18:12:06,050][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.06390571594238281, 'train/grad_norm_pre_clip': 21.370702743530273, 'timing/data': 0.0007985660049598664, 'timing/model': 1.2178445699973963, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.34}) +[2026-06-23 18:12:10,462][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.14359022676944733, 'train/grad_norm_pre_clip': 47.173667907714844, 'timing/data': 0.001025415025651455, 'timing/model': 1.2127868570096325, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.34}) +[2026-06-23 18:12:14,878][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.3708154261112213, 'train/grad_norm_pre_clip': 47.64411544799805, 'timing/data': 0.0012548289960250258, 'timing/model': 1.1945241460052785, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.34}) +[2026-06-23 18:12:19,291][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.16788750886917114, 'train/grad_norm_pre_clip': 21.462743759155273, 'timing/data': 0.0012044059985782951, 'timing/model': 1.213543521007523, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.34}) +[2026-06-23 18:12:23,911][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.1096736490726471, 'train/grad_norm_pre_clip': 22.00562858581543, 'timing/data': 0.0009161739726550877, 'timing/model': 1.1912336699897423, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.34}) +[2026-06-23 18:12:28,348][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.09142740815877914, 'train/grad_norm_pre_clip': 17.42424774169922, 'timing/data': 0.0007909620180726051, 'timing/model': 1.216050731978612, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.34}) +[2026-06-23 18:12:32,791][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.06248824670910835, 'train/grad_norm_pre_clip': 30.713048934936523, 'timing/data': 0.000790391000919044, 'timing/model': 1.2123719390074257, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.34}) +[2026-06-23 18:12:37,182][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.39441362023353577, 'train/grad_norm_pre_clip': 23.904048919677734, 'timing/data': 0.001024783996399492, 'timing/model': 1.1952643539989367, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.34}) +[2026-06-23 18:12:41,871][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.17672841250896454, 'train/grad_norm_pre_clip': 11.516510009765625, 'timing/data': 0.0009471910016145557, 'timing/model': 1.2179534820024855, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.34}) +[2026-06-23 18:12:46,259][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.08524182438850403, 'train/grad_norm_pre_clip': 33.74335861206055, 'timing/data': 0.0007840290199965239, 'timing/model': 1.1917654320131987, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.34}) +[2026-06-23 18:12:50,633][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.102454774081707, 'train/grad_norm_pre_clip': 23.246309280395508, 'timing/data': 0.0007846610096748918, 'timing/model': 1.195620775979478, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.34}) +[2026-06-23 18:12:54,999][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.25434258580207825, 'train/grad_norm_pre_clip': 44.36670684814453, 'timing/data': 0.0010031640122178942, 'timing/model': 1.1927704150148202, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.34}) +[2026-06-23 18:12:59,667][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.04219458997249603, 'train/grad_norm_pre_clip': 40.39500045776367, 'timing/data': 0.0008615830156486481, 'timing/model': 1.2185347070044372, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.34}) +[2026-06-23 18:13:04,046][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.10160960257053375, 'train/grad_norm_pre_clip': 35.742218017578125, 'timing/data': 0.00096436298917979, 'timing/model': 1.1894912290154025, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.34}) +[2026-06-23 18:13:08,407][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.3300439119338989, 'train/grad_norm_pre_clip': 41.783416748046875, 'timing/data': 0.0007652550120837986, 'timing/model': 1.1904568120080512, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.34}) +[2026-06-23 18:13:12,797][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.025838181376457214, 'train/grad_norm_pre_clip': 14.77953815460205, 'timing/data': 0.0007328040082938969, 'timing/model': 1.2190256690082606, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.34}) +[2026-06-23 18:13:17,430][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.007363424636423588, 'train/grad_norm_pre_clip': 20.69635581970215, 'timing/data': 0.0007747230120003223, 'timing/model': 1.1910993559868075, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.34}) +[2026-06-23 18:13:21,795][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.055086735635995865, 'train/grad_norm_pre_clip': 18.433082580566406, 'timing/data': 0.001101375004509464, 'timing/model': 1.1921866129850969, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.34}) +[2026-06-23 18:13:26,202][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.05306018888950348, 'train/grad_norm_pre_clip': 15.891120910644531, 'timing/data': 0.001025705976644531, 'timing/model': 1.211341298010666, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.34}) +[2026-06-23 18:13:30,619][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.09708623588085175, 'train/grad_norm_pre_clip': 22.358875274658203, 'timing/data': 0.0009183269867207855, 'timing/model': 1.209535208006855, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.34}) +[2026-06-23 18:13:35,245][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.09737957268953323, 'train/grad_norm_pre_clip': 54.85319900512695, 'timing/data': 0.00076464400626719, 'timing/model': 1.2093475889996625, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.34}) +[2026-06-23 18:13:39,612][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.06816377490758896, 'train/grad_norm_pre_clip': 32.13541030883789, 'timing/data': 0.0007649640028830618, 'timing/model': 1.1912397959968075, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.34}) +[2026-06-23 18:13:43,983][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.22256436944007874, 'train/grad_norm_pre_clip': 43.906795501708984, 'timing/data': 0.0007602959813084453, 'timing/model': 1.1907142430136446, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.34}) +[2026-06-23 18:13:48,390][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.24970680475234985, 'train/grad_norm_pre_clip': 39.1809196472168, 'timing/data': 0.0007592340116389096, 'timing/model': 1.2114195229951292, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.34}) +[2026-06-23 18:13:52,998][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.22442297637462616, 'train/grad_norm_pre_clip': 25.93031883239746, 'timing/data': 0.0009502070024609566, 'timing/model': 1.190359111991711, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.34}) +[2026-06-23 18:13:57,436][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.02229633554816246, 'train/grad_norm_pre_clip': 23.618980407714844, 'timing/data': 0.0007385450007859617, 'timing/model': 1.2209850699873641, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.34}) +[2026-06-23 18:14:01,842][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.15349963307380676, 'train/grad_norm_pre_clip': 17.882850646972656, 'timing/data': 0.0007975539774633944, 'timing/model': 1.2134159170091152, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.35}) +[2026-06-23 18:14:06,296][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.13248087465763092, 'train/grad_norm_pre_clip': 35.3992919921875, 'timing/data': 0.0007942280208226293, 'timing/model': 1.2269331510178745, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.35}) +[2026-06-23 18:14:10,964][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.25068899989128113, 'train/grad_norm_pre_clip': 63.7855224609375, 'timing/data': 0.0007695929962210357, 'timing/model': 1.2109951989841647, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.35}) +[2026-06-23 18:14:15,382][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.3509518802165985, 'train/grad_norm_pre_clip': 56.747833251953125, 'timing/data': 0.0008677549776621163, 'timing/model': 1.212442602001829, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.35}) +[2026-06-23 18:14:19,805][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.09607435762882233, 'train/grad_norm_pre_clip': 38.803340911865234, 'timing/data': 0.0009737099753692746, 'timing/model': 1.2148685450083576, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.35}) +[2026-06-23 18:14:24,259][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.12832121551036835, 'train/grad_norm_pre_clip': 17.47869873046875, 'timing/data': 0.0008014419872779399, 'timing/model': 1.2215540309844073, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.35}) +[2026-06-23 18:14:28,932][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.09364648908376694, 'train/grad_norm_pre_clip': 27.886743545532227, 'timing/data': 0.0007998789951670915, 'timing/model': 1.2282669080013875, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.35}) +[2026-06-23 18:14:33,350][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.055460862815380096, 'train/grad_norm_pre_clip': 13.528491020202637, 'timing/data': 0.0007540340011473745, 'timing/model': 1.213478670018958, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.35}) +[2026-06-23 18:14:37,739][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 900 +[2026-06-23 18:15:35,690][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 900 +[2026-06-23 18:15:35,690][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 900 +[2026-06-23 18:16:12,932][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 900 +[2026-06-23 18:16:12,934][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.21779686212539673, 'train/grad_norm_pre_clip': 25.548593521118164, 'eval/action_loss/seconds': 57.94677456197678, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.06338523328304291, 'eval/latency_2/loss': 0.06338523328304291, 'eval/flappy/loss': 0.06338523328304291, 'eval/flappy/latency_2/loss': 0.06338523328304291, 'eval/flap_precision': 0.9170918367346939, 'eval/flap_recall': 0.9755766621438263, 'eval/flap_f1': 0.94543063773833, 'eval/flap_rate_model': 0.21777777777777776, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.94543063773833, 'eval/latency_2/flap_precision': 0.9170918367346939, 'eval/latency_2/flap_recall': 0.9755766621438263, 'eval/latency_2/flap_f1': 0.94543063773833, 'eval/latency_2/flap_rate_model': 0.21777777777777776, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.94543063773833, 'eval/action_classification/data_wait_seconds': 0.45528653264045715, 'eval/action_classification/predict_seconds': 36.63246154785156, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0009821649873629212, 'timing/model': 1.1899781500105746, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.35}) +[2026-06-23 18:16:17,550][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.15566782653331757, 'train/grad_norm_pre_clip': 23.271812438964844, 'timing/data': 0.000746861012885347, 'timing/model': 1.1937835180142429, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.35}) +[2026-06-23 18:16:21,886][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.05336034297943115, 'train/grad_norm_pre_clip': 13.0243558883667, 'timing/data': 0.000737964001018554, 'timing/model': 1.199992273002863, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.35}) +[2026-06-23 18:16:26,247][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.14841431379318237, 'train/grad_norm_pre_clip': 30.663806915283203, 'timing/data': 0.0009698419889900833, 'timing/model': 1.185057446011342, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.35}) +[2026-06-23 18:16:30,637][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.20549964904785156, 'train/grad_norm_pre_clip': 19.541311264038086, 'timing/data': 0.0010435990116093308, 'timing/model': 1.2000606460205745, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.35}) +[2026-06-23 18:16:35,235][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.05400095507502556, 'train/grad_norm_pre_clip': 18.636932373046875, 'timing/data': 0.0008253659761976451, 'timing/model': 1.2027220340096392, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.35}) +[2026-06-23 18:16:39,628][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.0814242884516716, 'train/grad_norm_pre_clip': 45.57221221923828, 'timing/data': 0.0009228450071532279, 'timing/model': 1.182990992005216, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.35}) +[2026-06-23 18:16:44,006][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.10659593343734741, 'train/grad_norm_pre_clip': 40.47301483154297, 'timing/data': 0.0012175909942016006, 'timing/model': 1.180922516999999, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.35}) +[2026-06-23 18:16:48,387][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.04771801084280014, 'train/grad_norm_pre_clip': 23.047134399414062, 'timing/data': 0.0009702940005809069, 'timing/model': 1.2036924939893652, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.35}) +[2026-06-23 18:16:52,983][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.1426844745874405, 'train/grad_norm_pre_clip': 18.08939552307129, 'timing/data': 0.0007442859932780266, 'timing/model': 1.2050451929972041, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.35}) +[2026-06-23 18:16:57,405][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.22427460551261902, 'train/grad_norm_pre_clip': 37.505489349365234, 'timing/data': 0.0008640570158604532, 'timing/model': 1.1986548900022171, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.35}) +[2026-06-23 18:17:01,813][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.12250042706727982, 'train/grad_norm_pre_clip': 40.0865364074707, 'timing/data': 0.0007622090051881969, 'timing/model': 1.1993699400045443, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.35}) +[2026-06-23 18:17:06,382][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.19567188620567322, 'train/grad_norm_pre_clip': 44.02352523803711, 'timing/data': 0.0008350930002052337, 'timing/model': 1.3770492890034802, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.35}) +[2026-06-23 18:17:10,763][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.09634844958782196, 'train/grad_norm_pre_clip': 40.19411087036133, 'timing/data': 0.0008864080009516329, 'timing/model': 1.2106728500220925, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.35}) +[2026-06-23 18:17:15,154][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.08658958226442337, 'train/grad_norm_pre_clip': 21.45627212524414, 'timing/data': 0.0009474210091866553, 'timing/model': 1.2006601300090551, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.35}) +[2026-06-23 18:17:19,552][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.11278540641069412, 'train/grad_norm_pre_clip': 17.288742065429688, 'timing/data': 0.0007216439989861101, 'timing/model': 1.2023408960085362, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.35}) +[2026-06-23 18:17:23,938][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.10675930976867676, 'train/grad_norm_pre_clip': 43.53036117553711, 'timing/data': 0.0010037360189016908, 'timing/model': 1.2094830769929104, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.35}) +[2026-06-23 18:17:28,518][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.20839005708694458, 'train/grad_norm_pre_clip': 34.6575927734375, 'timing/data': 0.0007457289902959019, 'timing/model': 1.1839137740025762, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.35}) +[2026-06-23 18:17:32,925][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.022327804937958717, 'train/grad_norm_pre_clip': 27.53508949279785, 'timing/data': 0.0008777219918556511, 'timing/model': 1.2050835519912653, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.36}) +[2026-06-23 18:17:37,332][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.15657883882522583, 'train/grad_norm_pre_clip': 47.70960235595703, 'timing/data': 0.0007299989811144769, 'timing/model': 1.209281790012028, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.36}) +[2026-06-23 18:17:41,740][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.38905590772628784, 'train/grad_norm_pre_clip': 35.271968841552734, 'timing/data': 0.0007989369914866984, 'timing/model': 1.2034015500103123, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.36}) +[2026-06-23 18:17:46,376][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.06993653625249863, 'train/grad_norm_pre_clip': 54.7470817565918, 'timing/data': 0.0010327389754820615, 'timing/model': 1.2104157940193545, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.36}) +[2026-06-23 18:17:50,751][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.14787018299102783, 'train/grad_norm_pre_clip': 26.934606552124023, 'timing/data': 0.0010840839822776616, 'timing/model': 1.1884877019911073, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.36}) +[2026-06-23 18:17:55,194][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.19351546466350555, 'train/grad_norm_pre_clip': 19.174619674682617, 'timing/data': 0.0014554110239259899, 'timing/model': 1.219346760015469, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.36}) +[2026-06-23 18:17:59,613][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.05295545980334282, 'train/grad_norm_pre_clip': 58.38401412963867, 'timing/data': 0.0008107290195766836, 'timing/model': 1.208045648992993, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.36}) +[2026-06-23 18:18:04,208][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.12784770131111145, 'train/grad_norm_pre_clip': 53.09135055541992, 'timing/data': 0.0008907269802875817, 'timing/model': 1.2085681640019175, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.36}) +[2026-06-23 18:18:08,613][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.06684134155511856, 'train/grad_norm_pre_clip': 21.721458435058594, 'timing/data': 0.0008062010165303946, 'timing/model': 1.2211057679960504, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.36}) +[2026-06-23 18:18:13,013][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.01466679759323597, 'train/grad_norm_pre_clip': 11.09282112121582, 'timing/data': 0.0007813639822416008, 'timing/model': 1.2055560650187545, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.36}) +[2026-06-23 18:18:17,398][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.20567479729652405, 'train/grad_norm_pre_clip': 52.528106689453125, 'timing/data': 0.000931932998355478, 'timing/model': 1.2112019089981914, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.36}) +[2026-06-23 18:18:21,924][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.11230204254388809, 'train/grad_norm_pre_clip': 37.857669830322266, 'timing/data': 0.0007228759932331741, 'timing/model': 1.1850112069805618, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.36}) +[2026-06-23 18:18:26,322][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.09634891897439957, 'train/grad_norm_pre_clip': 30.184234619140625, 'timing/data': 0.0009609970147721469, 'timing/model': 1.2086502299935091, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.36}) +[2026-06-23 18:18:30,764][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.0949452668428421, 'train/grad_norm_pre_clip': 20.637659072875977, 'timing/data': 0.0007707650074735284, 'timing/model': 1.209338601998752, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.36}) +[2026-06-23 18:18:35,193][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.06990694999694824, 'train/grad_norm_pre_clip': 15.40186595916748, 'timing/data': 0.0007332150125876069, 'timing/model': 1.2128232209943235, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.36}) +[2026-06-23 18:18:39,745][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.04483356699347496, 'train/grad_norm_pre_clip': 10.77716064453125, 'timing/data': 0.000848348019644618, 'timing/model': 1.2245294269814622, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.36}) +[2026-06-23 18:18:44,164][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.1015787199139595, 'train/grad_norm_pre_clip': 22.20735740661621, 'timing/data': 0.0010151560127269477, 'timing/model': 1.2194165760010947, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.36}) +[2026-06-23 18:18:48,585][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.12001662701368332, 'train/grad_norm_pre_clip': 11.623047828674316, 'timing/data': 0.0009442149894312024, 'timing/model': 1.2007315780210774, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.36}) +[2026-06-23 18:18:52,979][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.39082998037338257, 'train/grad_norm_pre_clip': 34.77901077270508, 'timing/data': 0.0007479320047423244, 'timing/model': 1.1894855500140693, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.36}) +[2026-06-23 18:18:57,641][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.10187693685293198, 'train/grad_norm_pre_clip': 14.956791877746582, 'timing/data': 0.0008011210011318326, 'timing/model': 1.2068498349981382, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.36}) +[2026-06-23 18:19:02,078][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.09920672327280045, 'train/grad_norm_pre_clip': 27.72520637512207, 'timing/data': 0.0010111990268342197, 'timing/model': 1.2162348800047766, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.36}) +[2026-06-23 18:19:06,477][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.0025286644231528044, 'train/grad_norm_pre_clip': 27.1263484954834, 'timing/data': 0.0007784590125083923, 'timing/model': 1.2198722930043004, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.36}) +[2026-06-23 18:19:10,895][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.26257821917533875, 'train/grad_norm_pre_clip': 30.326404571533203, 'timing/data': 0.0007197500090114772, 'timing/model': 1.21528338000644, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.36}) +[2026-06-23 18:19:15,512][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.13594458997249603, 'train/grad_norm_pre_clip': 72.40168762207031, 'timing/data': 0.0007434439903590828, 'timing/model': 1.191086673992686, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.36}) +[2026-06-23 18:19:19,936][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.22510650753974915, 'train/grad_norm_pre_clip': 31.174081802368164, 'timing/data': 0.0007227759924717247, 'timing/model': 1.2229611090151593, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.36}) +[2026-06-23 18:19:24,345][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.2665645480155945, 'train/grad_norm_pre_clip': 29.4197998046875, 'timing/data': 0.0007370120147243142, 'timing/model': 1.2123895279946737, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.36}) +[2026-06-23 18:19:28,754][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.11603410542011261, 'train/grad_norm_pre_clip': 33.96945571899414, 'timing/data': 0.001020596973830834, 'timing/model': 1.1906660839740653, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.37}) +[2026-06-23 18:19:33,449][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.30682751536369324, 'train/grad_norm_pre_clip': 26.091806411743164, 'timing/data': 0.000727806007489562, 'timing/model': 1.2153501579887234, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.37}) +[2026-06-23 18:19:37,866][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.12103485316038132, 'train/grad_norm_pre_clip': 19.220077514648438, 'timing/data': 0.001034872984746471, 'timing/model': 1.1881108090165071, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.37}) +[2026-06-23 18:19:42,246][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.06339888274669647, 'train/grad_norm_pre_clip': 15.770895957946777, 'timing/data': 0.0007449369877576828, 'timing/model': 1.2074958490266, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.37}) +[2026-06-23 18:19:46,639][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.19279760122299194, 'train/grad_norm_pre_clip': 27.76167106628418, 'timing/data': 0.0007514290045946836, 'timing/model': 1.1899527120112907, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.37}) +[2026-06-23 18:19:51,254][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.1701222062110901, 'train/grad_norm_pre_clip': 20.889331817626953, 'timing/data': 0.0007233060023281723, 'timing/model': 1.191678139992291, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.37}) +[2026-06-23 18:19:55,668][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.07703253626823425, 'train/grad_norm_pre_clip': 17.587024688720703, 'timing/data': 0.0007528620189987123, 'timing/model': 1.1931752160016913, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.37}) +[2026-06-23 18:20:00,051][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.165481299161911, 'train/grad_norm_pre_clip': 20.14403533935547, 'timing/data': 0.0007381039904430509, 'timing/model': 1.20651696601999, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.37}) +[2026-06-23 18:20:04,436][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.325337290763855, 'train/grad_norm_pre_clip': 78.59917449951172, 'timing/data': 0.0007139590161386877, 'timing/model': 1.2151035749993753, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.37}) +[2026-06-23 18:20:09,096][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.03646203130483627, 'train/grad_norm_pre_clip': 43.47206497192383, 'timing/data': 0.0008521549752913415, 'timing/model': 1.2117032700043637, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.37}) +[2026-06-23 18:20:13,468][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.2272099107503891, 'train/grad_norm_pre_clip': 31.30683708190918, 'timing/data': 0.0007348390063270926, 'timing/model': 1.1947543410060462, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.37}) +[2026-06-23 18:20:17,870][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.08767916262149811, 'train/grad_norm_pre_clip': 40.57831954956055, 'timing/data': 0.0007390160171780735, 'timing/model': 1.1913678809942212, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.37}) +[2026-06-23 18:20:22,297][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.047606054693460464, 'train/grad_norm_pre_clip': 13.079961776733398, 'timing/data': 0.0007429340039379895, 'timing/model': 1.218745073012542, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.37}) +[2026-06-23 18:20:26,945][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.07841348648071289, 'train/grad_norm_pre_clip': 24.605403900146484, 'timing/data': 0.0007158730004448444, 'timing/model': 1.2094057390058879, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.37}) +[2026-06-23 18:20:31,349][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.09066690504550934, 'train/grad_norm_pre_clip': 22.774423599243164, 'timing/data': 0.0006744359852746129, 'timing/model': 1.186901603010483, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.37}) +[2026-06-23 18:20:35,730][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.09823436290025711, 'train/grad_norm_pre_clip': 18.599348068237305, 'timing/data': 0.0007760239823255688, 'timing/model': 1.186823497992009, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.37}) +[2026-06-23 18:20:40,121][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.08645278960466385, 'train/grad_norm_pre_clip': 29.015016555786133, 'timing/data': 0.000877711980137974, 'timing/model': 1.1878411790239625, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.37}) +[2026-06-23 18:20:44,779][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.02921513468027115, 'train/grad_norm_pre_clip': 43.16571044921875, 'timing/data': 0.0007561879756394774, 'timing/model': 1.1872801599965896, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.37}) +[2026-06-23 18:20:49,205][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.022671179845929146, 'train/grad_norm_pre_clip': 30.012794494628906, 'timing/data': 0.0009618379990570247, 'timing/model': 1.2158431540010497, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.37}) +[2026-06-23 18:20:53,592][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.08818475902080536, 'train/grad_norm_pre_clip': 33.029014587402344, 'timing/data': 0.0011273029958829284, 'timing/model': 1.1930886149930302, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.37}) +[2026-06-23 18:20:57,999][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.08452728390693665, 'train/grad_norm_pre_clip': 23.446313858032227, 'timing/data': 0.0008531470084562898, 'timing/model': 1.1888416439760476, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.37}) +[2026-06-23 18:21:02,655][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.24197643995285034, 'train/grad_norm_pre_clip': 48.421539306640625, 'timing/data': 0.000898300000699237, 'timing/model': 1.221693764993688, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.37}) +[2026-06-23 18:21:07,053][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.05615921691060066, 'train/grad_norm_pre_clip': 23.23212432861328, 'timing/data': 0.0007198109815362841, 'timing/model': 1.1909922169870697, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.37}) +[2026-06-23 18:21:11,456][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.2447095513343811, 'train/grad_norm_pre_clip': 43.1674919128418, 'timing/data': 0.0009189789998345077, 'timing/model': 1.2085567619942594, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.37}) +[2026-06-23 18:21:15,834][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.13553807139396667, 'train/grad_norm_pre_clip': 63.13599395751953, 'timing/data': 0.0006902559834998101, 'timing/model': 1.1896153209963813, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.37}) +[2026-06-23 18:21:20,463][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.276804119348526, 'train/grad_norm_pre_clip': 26.394439697265625, 'timing/data': 0.0011028489971067756, 'timing/model': 1.2064386659767479, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.38}) +[2026-06-23 18:21:24,882][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.30219247937202454, 'train/grad_norm_pre_clip': 22.081878662109375, 'timing/data': 0.0008005390118341893, 'timing/model': 1.2057725229824428, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.38}) +[2026-06-23 18:21:29,275][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.020445188507437706, 'train/grad_norm_pre_clip': 19.377662658691406, 'timing/data': 0.0007052239961922169, 'timing/model': 1.2188616750063375, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.38}) +[2026-06-23 18:21:33,652][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.17691855132579803, 'train/grad_norm_pre_clip': 26.427658081054688, 'timing/data': 0.0007937469927128404, 'timing/model': 1.1962426279787906, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.38}) +[2026-06-23 18:21:38,355][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.15182968974113464, 'train/grad_norm_pre_clip': 41.29984664916992, 'timing/data': 0.0008820600050967187, 'timing/model': 1.1984207420027815, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.38}) +[2026-06-23 18:21:42,813][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.04710633307695389, 'train/grad_norm_pre_clip': 23.634370803833008, 'timing/data': 0.0011691809922922403, 'timing/model': 1.2207390739931725, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.38}) +[2026-06-23 18:21:47,237][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.14422005414962769, 'train/grad_norm_pre_clip': 15.373848915100098, 'timing/data': 0.0007282659935299307, 'timing/model': 1.2024882689875085, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.38}) +[2026-06-23 18:21:51,684][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.1927388608455658, 'train/grad_norm_pre_clip': 33.66926193237305, 'timing/data': 0.001351056998828426, 'timing/model': 1.2091264979972038, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.38}) +[2026-06-23 18:21:56,362][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.15824691951274872, 'train/grad_norm_pre_clip': 17.51683235168457, 'timing/data': 0.0007728879863861948, 'timing/model': 1.1915621619846206, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.38}) +[2026-06-23 18:22:00,886][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.06610026210546494, 'train/grad_norm_pre_clip': 20.870681762695312, 'timing/data': 0.0007917039911262691, 'timing/model': 1.2092604839999694, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.38}) +[2026-06-23 18:22:05,314][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.10135161876678467, 'train/grad_norm_pre_clip': 18.539657592773438, 'timing/data': 0.0007547659915871918, 'timing/model': 1.2123284440021962, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.38}) +[2026-06-23 18:22:09,733][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.09673932939767838, 'train/grad_norm_pre_clip': 16.160980224609375, 'timing/data': 0.0007871950219850987, 'timing/model': 1.199886967020575, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.38}) +[2026-06-23 18:22:14,316][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.12758679687976837, 'train/grad_norm_pre_clip': 13.99563217163086, 'timing/data': 0.0009157930035144091, 'timing/model': 1.212396988994442, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.38}) +[2026-06-23 18:22:18,766][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.08833785355091095, 'train/grad_norm_pre_clip': 28.922019958496094, 'timing/data': 0.000774470972828567, 'timing/model': 1.2195877589983866, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.38}) +[2026-06-23 18:22:23,163][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.11863149702548981, 'train/grad_norm_pre_clip': 23.387338638305664, 'timing/data': 0.0007569590234197676, 'timing/model': 1.191715230990667, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.38}) +[2026-06-23 18:22:27,579][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.2134568840265274, 'train/grad_norm_pre_clip': 29.30582618713379, 'timing/data': 0.000912496994715184, 'timing/model': 1.21282601999701, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.38}) +[2026-06-23 18:22:32,201][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.08132848888635635, 'train/grad_norm_pre_clip': 12.246312141418457, 'timing/data': 0.0008973079966381192, 'timing/model': 1.2100816360034514, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.38}) +[2026-06-23 18:22:36,603][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.02623240277171135, 'train/grad_norm_pre_clip': 18.76499366760254, 'timing/data': 0.0009033889800775796, 'timing/model': 1.192211966990726, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.38}) +[2026-06-23 18:22:40,986][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.08670918643474579, 'train/grad_norm_pre_clip': 29.040063858032227, 'timing/data': 0.0008196349954232574, 'timing/model': 1.191553038021084, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.38}) +[2026-06-23 18:22:45,341][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.15236017107963562, 'train/grad_norm_pre_clip': 44.85975646972656, 'timing/data': 0.0007259919948410243, 'timing/model': 1.1882248259789776, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.38}) +[2026-06-23 18:22:49,957][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.1279193013906479, 'train/grad_norm_pre_clip': 19.5108699798584, 'timing/data': 0.0007385649951174855, 'timing/model': 1.1864617060055025, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.38}) +[2026-06-23 18:22:54,357][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.15257884562015533, 'train/grad_norm_pre_clip': 44.11054992675781, 'timing/data': 0.0010956939950119704, 'timing/model': 1.2083820620027836, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.38}) +[2026-06-23 18:22:58,822][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.038116198033094406, 'train/grad_norm_pre_clip': 25.311216354370117, 'timing/data': 0.0007739599968772382, 'timing/model': 1.2080293989856727, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.38}) +[2026-06-23 18:23:03,230][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.04668424278497696, 'train/grad_norm_pre_clip': 38.90363693237305, 'timing/data': 0.0008645679918117821, 'timing/model': 1.1906636409985367, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.38}) +[2026-06-23 18:23:07,909][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.04311446473002434, 'train/grad_norm_pre_clip': 11.117712020874023, 'timing/data': 0.0011005639971699566, 'timing/model': 1.1949538699991535, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.38}) +[2026-06-23 18:23:12,284][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.37721335887908936, 'train/grad_norm_pre_clip': 38.721282958984375, 'timing/data': 0.0007908719999250025, 'timing/model': 1.194875575019978, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.38}) +[2026-06-23 18:23:16,701][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.1338164359331131, 'train/grad_norm_pre_clip': 36.8160285949707, 'timing/data': 0.0007488340197596699, 'timing/model': 1.2164793339907192, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.39}) +[2026-06-23 18:23:21,127][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.3629845380783081, 'train/grad_norm_pre_clip': 30.4223690032959, 'timing/data': 0.0008810579893179238, 'timing/model': 1.214181897987146, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.39}) +[2026-06-23 18:23:25,796][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.26774561405181885, 'train/grad_norm_pre_clip': 24.446155548095703, 'timing/data': 0.0008151569927576929, 'timing/model': 1.2092230120033491, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.39}) +[2026-06-23 18:23:30,186][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.12472226470708847, 'train/grad_norm_pre_clip': 28.109228134155273, 'timing/data': 0.0008648689836263657, 'timing/model': 1.2156802960089408, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.39}) +[2026-06-23 18:23:34,565][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.06844548135995865, 'train/grad_norm_pre_clip': 20.98235511779785, 'timing/data': 0.0009183470101561397, 'timing/model': 1.1918223180109635, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.39}) +[2026-06-23 18:23:38,971][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1000 +[2026-06-23 18:24:36,373][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1000 +[2026-06-23 18:24:36,374][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1000 +[2026-06-23 18:25:13,811][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1000 +[2026-06-23 18:27:05,836][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-23 18:27:05,838][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.06714572757482529, 'train/grad_norm_pre_clip': 20.50981330871582, 'eval/action_loss/seconds': 57.3985320639913, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.07302667945623398, 'eval/latency_2/loss': 0.07302667945623398, 'eval/flappy/loss': 0.07302667945623398, 'eval/flappy/latency_2/loss': 0.07302667945623398, 'eval/flap_precision': 0.9861538461538462, 'eval/flap_recall': 0.8697421981004071, 'eval/flap_f1': 0.9242970439798126, 'eval/flap_rate_model': 0.18055555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9242970439798126, 'eval/latency_2/flap_precision': 0.9861538461538462, 'eval/latency_2/flap_recall': 0.8697421981004071, 'eval/latency_2/flap_f1': 0.9242970439798126, 'eval/latency_2/flap_rate_model': 0.18055555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9242970439798126, 'eval/action_classification/data_wait_seconds': 0.4534274935722351, 'eval/action_classification/predict_seconds': 36.84313201904297, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 27.100000000000158, 'rl_games_eval/mid_train/mean_length': 237.0, 'rl_games_eval/mid_train/std_reward': 19.978988963408707, 'rl_games_eval/mid_train/std_length': 161.18436648757225, 'rl_games_eval/mid_train/macro_mean_reward': 27.100000000000158, 'rl_games_eval/mid_train/macro_mean_length': 237.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 27.100000000000158, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 237.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 19.978988963408707, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 161.18436648757225, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 27.100000000000158, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0008872299804352224, 'timing/model': 1.1868249109829776, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.39}) +[2026-06-23 18:27:05,838][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_1000_state +[2026-06-23 18:27:05,847][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 18:27:19,158][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_1000_state/model.safetensors +[2026-06-23 18:27:30,230][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_1000_state/optimizer.bin +[2026-06-23 18:27:30,230][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_1000_state/sampler.bin +[2026-06-23 18:27:30,230][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-23 18:27:30,231][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-23 18:27:30,412][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-23 18:27:30,413][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 18:27:30,438][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 18:27:34,651][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.217499241232872, 'train/grad_norm_pre_clip': 36.285762786865234, 'timing/data': 0.0007634609937667847, 'timing/model': 1.1598314639995806, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.39}) +[2026-06-23 18:27:38,847][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.15921583771705627, 'train/grad_norm_pre_clip': 38.67391586303711, 'timing/data': 0.000791622995166108, 'timing/model': 1.1434794499946292, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.39}) +[2026-06-23 18:27:43,024][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.09676564484834671, 'train/grad_norm_pre_clip': 18.251049041748047, 'timing/data': 0.0007719269779045135, 'timing/model': 1.1422679300012533, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.39}) +[2026-06-23 18:27:47,443][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.022332172840833664, 'train/grad_norm_pre_clip': 27.19318962097168, 'timing/data': 0.0009540440223645419, 'timing/model': 1.14667691901559, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.39}) +[2026-06-23 18:27:51,683][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.04369889572262764, 'train/grad_norm_pre_clip': 19.567245483398438, 'timing/data': 0.0007384549826383591, 'timing/model': 1.170258685015142, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.39}) +[2026-06-23 18:27:55,931][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.041732534766197205, 'train/grad_norm_pre_clip': 11.488448143005371, 'timing/data': 0.0009690810111351311, 'timing/model': 1.1522708569827955, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.39}) +[2026-06-23 18:28:00,166][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.1261240392923355, 'train/grad_norm_pre_clip': 23.879240036010742, 'timing/data': 0.0007717669941484928, 'timing/model': 1.1698948220000602, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.39}) +[2026-06-23 18:28:04,658][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.1689971685409546, 'train/grad_norm_pre_clip': 33.851444244384766, 'timing/data': 0.0007311109802685678, 'timing/model': 1.158858501992654, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.39}) +[2026-06-23 18:28:08,953][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.006057523190975189, 'train/grad_norm_pre_clip': 12.204486846923828, 'timing/data': 0.0007422720082104206, 'timing/model': 1.1844312980247196, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.39}) +[2026-06-23 18:28:13,193][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.17018422484397888, 'train/grad_norm_pre_clip': 37.48275375366211, 'timing/data': 0.00072743502096273, 'timing/model': 1.1602672160079237, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.39}) +[2026-06-23 18:28:17,501][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.09804690629243851, 'train/grad_norm_pre_clip': 39.74766540527344, 'timing/data': 0.0007490550051443279, 'timing/model': 1.1914675229927525, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.39}) +[2026-06-23 18:28:22,015][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.011622092686593533, 'train/grad_norm_pre_clip': 15.514450073242188, 'timing/data': 0.0007344580080825835, 'timing/model': 1.193793440994341, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.39}) +[2026-06-23 18:28:26,354][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.08514781296253204, 'train/grad_norm_pre_clip': 58.473243713378906, 'timing/data': 0.0007346979982685298, 'timing/model': 1.1872535340080503, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.39}) +[2026-06-23 18:28:30,692][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.07643559575080872, 'train/grad_norm_pre_clip': 51.1649169921875, 'timing/data': 0.0011721960036084056, 'timing/model': 1.1820866930065677, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.39}) +[2026-06-23 18:28:34,997][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.22552070021629333, 'train/grad_norm_pre_clip': 45.097782135009766, 'timing/data': 0.000863364984979853, 'timing/model': 1.1821349020174239, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.39}) +[2026-06-23 18:28:39,460][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.18251517415046692, 'train/grad_norm_pre_clip': 28.387968063354492, 'timing/data': 0.0010324989852961153, 'timing/model': 1.183612182008801, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.39}) +[2026-06-23 18:28:43,776][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.10109750181436539, 'train/grad_norm_pre_clip': 22.490501403808594, 'timing/data': 0.0007342569879256189, 'timing/model': 1.2028114479908254, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.39}) +[2026-06-23 18:28:48,097][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.056053243577480316, 'train/grad_norm_pre_clip': 36.24604415893555, 'timing/data': 0.0008615420083515346, 'timing/model': 1.1733552039950155, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.39}) +[2026-06-23 18:28:52,415][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.11456196010112762, 'train/grad_norm_pre_clip': 27.204570770263672, 'timing/data': 0.0007469620031770319, 'timing/model': 1.1893749629962258, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.39}) +[2026-06-23 18:28:57,021][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.08397321403026581, 'train/grad_norm_pre_clip': 29.73554039001465, 'timing/data': 0.0007620989927090704, 'timing/model': 1.1933347619778942, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.39}) +[2026-06-23 18:29:01,383][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.02868230827152729, 'train/grad_norm_pre_clip': 35.925209045410156, 'timing/data': 0.0007435660227201879, 'timing/model': 1.1955037089937832, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.4}) +[2026-06-23 18:29:05,726][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.11217997223138809, 'train/grad_norm_pre_clip': 22.610252380371094, 'timing/data': 0.0008647390059195459, 'timing/model': 1.1761362550023478, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.4}) +[2026-06-23 18:29:10,039][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.06625884026288986, 'train/grad_norm_pre_clip': 17.625463485717773, 'timing/data': 0.0007990479934960604, 'timing/model': 1.176289294991875, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.4}) +[2026-06-23 18:29:14,651][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.07816574722528458, 'train/grad_norm_pre_clip': 21.657859802246094, 'timing/data': 0.0007582819962408394, 'timing/model': 1.1727254950092174, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.4}) +[2026-06-23 18:29:19,025][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.087473563849926, 'train/grad_norm_pre_clip': 29.229286193847656, 'timing/data': 0.0010965980181936175, 'timing/model': 1.1930931889801286, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.4}) +[2026-06-23 18:29:23,342][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.06864940375089645, 'train/grad_norm_pre_clip': 20.93314552307129, 'timing/data': 0.0007512799929827452, 'timing/model': 1.1795261569786817, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.4}) +[2026-06-23 18:29:27,676][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.051535822451114655, 'train/grad_norm_pre_clip': 18.308218002319336, 'timing/data': 0.000724890996934846, 'timing/model': 1.1973988749959972, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.4}) +[2026-06-23 18:29:32,308][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.26631733775138855, 'train/grad_norm_pre_clip': 24.762258529663086, 'timing/data': 0.0007521210063714534, 'timing/model': 1.2066171719925478, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.4}) +[2026-06-23 18:29:36,660][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.2316833883523941, 'train/grad_norm_pre_clip': 32.262603759765625, 'timing/data': 0.0007437959720846266, 'timing/model': 1.1776840469974559, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.4}) +[2026-06-23 18:29:41,029][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.0273437537252903, 'train/grad_norm_pre_clip': 25.741024017333984, 'timing/data': 0.0007369929808191955, 'timing/model': 1.2020186669833492, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.4}) +[2026-06-23 18:29:45,411][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.10251974314451218, 'train/grad_norm_pre_clip': 29.86065101623535, 'timing/data': 0.0007423530041705817, 'timing/model': 1.2070360040233936, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.4}) +[2026-06-23 18:29:50,033][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.13282857835292816, 'train/grad_norm_pre_clip': 20.560321807861328, 'timing/data': 0.0011391779989935458, 'timing/model': 1.209853431006195, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.4}) +[2026-06-23 18:29:54,388][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.15972717106342316, 'train/grad_norm_pre_clip': 25.165306091308594, 'timing/data': 0.0007554979820270091, 'timing/model': 1.2019518809975125, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.4}) +[2026-06-23 18:29:58,765][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.1441800743341446, 'train/grad_norm_pre_clip': 25.408159255981445, 'timing/data': 0.0012470569927245378, 'timing/model': 1.2082951110205613, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.4}) +[2026-06-23 18:30:03,174][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.033430181443691254, 'train/grad_norm_pre_clip': 10.468267440795898, 'timing/data': 0.0011357710172887892, 'timing/model': 1.1901774230063893, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.4}) +[2026-06-23 18:30:07,814][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.11616712063550949, 'train/grad_norm_pre_clip': 28.76786994934082, 'timing/data': 0.0007668290054425597, 'timing/model': 1.213037184003042, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.4}) +[2026-06-23 18:30:12,237][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.08101096749305725, 'train/grad_norm_pre_clip': 34.509544372558594, 'timing/data': 0.0009018380078487098, 'timing/model': 1.2221206729882397, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.4}) +[2026-06-23 18:30:16,641][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.06090731918811798, 'train/grad_norm_pre_clip': 18.68240737915039, 'timing/data': 0.0007465810049325228, 'timing/model': 1.2086810500186402, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.4}) +[2026-06-23 18:30:21,055][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.07121560722589493, 'train/grad_norm_pre_clip': 24.044464111328125, 'timing/data': 0.0008551520004402846, 'timing/model': 1.2060005279781763, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.4}) +[2026-06-23 18:30:25,718][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.010797671042382717, 'train/grad_norm_pre_clip': 37.44786071777344, 'timing/data': 0.0009645039972383529, 'timing/model': 1.2059126409876626, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.4}) +[2026-06-23 18:30:30,065][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.02033841609954834, 'train/grad_norm_pre_clip': 46.295867919921875, 'timing/data': 0.0008730539993848652, 'timing/model': 1.196188470989, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.4}) +[2026-06-23 18:30:34,438][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.06342735886573792, 'train/grad_norm_pre_clip': 21.259014129638672, 'timing/data': 0.0010240940027870238, 'timing/model': 1.1871330680151004, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.4}) +[2026-06-23 18:30:38,825][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.07415681332349777, 'train/grad_norm_pre_clip': 15.462363243103027, 'timing/data': 0.0009281860257033259, 'timing/model': 1.225253156007966, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.4}) +[2026-06-23 18:30:43,475][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.12902569770812988, 'train/grad_norm_pre_clip': 28.529949188232422, 'timing/data': 0.0012025549949612468, 'timing/model': 1.2132901250151917, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.4}) +[2026-06-23 18:30:47,870][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.0828327015042305, 'train/grad_norm_pre_clip': 35.2402229309082, 'timing/data': 0.0009149120014626533, 'timing/model': 1.1941559500119183, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.4}) +[2026-06-23 18:30:52,232][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.09089041501283646, 'train/grad_norm_pre_clip': 16.272390365600586, 'timing/data': 0.0007123379909899086, 'timing/model': 1.1981485830037855, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.4}) +[2026-06-23 18:30:56,617][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.009762381203472614, 'train/grad_norm_pre_clip': 32.47341537475586, 'timing/data': 0.0009914530091919005, 'timing/model': 1.2136141960218083, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.41}) +[2026-06-23 18:31:01,269][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.10203134268522263, 'train/grad_norm_pre_clip': 22.77645492553711, 'timing/data': 0.001114421000238508, 'timing/model': 1.199694285984151, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.41}) +[2026-06-23 18:31:05,617][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.03144806623458862, 'train/grad_norm_pre_clip': 34.66447830200195, 'timing/data': 0.0007863339851610363, 'timing/model': 1.1832519120071083, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.41}) +[2026-06-23 18:31:10,035][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.010501839220523834, 'train/grad_norm_pre_clip': 17.47718620300293, 'timing/data': 0.0007901919889263809, 'timing/model': 1.1859095959807746, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.41}) +[2026-06-23 18:31:14,436][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.08667275309562683, 'train/grad_norm_pre_clip': 41.5251350402832, 'timing/data': 0.0007489950221497566, 'timing/model': 1.2216383980121464, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.41}) +[2026-06-23 18:31:19,053][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.10889642685651779, 'train/grad_norm_pre_clip': 33.03440475463867, 'timing/data': 0.001088263001292944, 'timing/model': 1.1857236979994923, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.41}) +[2026-06-23 18:31:23,444][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.28017744421958923, 'train/grad_norm_pre_clip': 30.76333999633789, 'timing/data': 0.0007310319924727082, 'timing/model': 1.2049991980020422, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.41}) +[2026-06-23 18:31:27,822][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.01573868654668331, 'train/grad_norm_pre_clip': 18.943628311157227, 'timing/data': 0.0009465710027143359, 'timing/model': 1.2000125430058688, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.41}) +[2026-06-23 18:31:32,179][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.13583435118198395, 'train/grad_norm_pre_clip': 32.63167190551758, 'timing/data': 0.0007267140026669949, 'timing/model': 1.206992902996717, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.41}) +[2026-06-23 18:31:36,852][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.2256401777267456, 'train/grad_norm_pre_clip': 26.198301315307617, 'timing/data': 0.0007986769778653979, 'timing/model': 1.205915658007143, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.41}) +[2026-06-23 18:31:41,247][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.06370477378368378, 'train/grad_norm_pre_clip': 16.960086822509766, 'timing/data': 0.0007752339879516512, 'timing/model': 1.1859107759955805, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.41}) +[2026-06-23 18:31:45,637][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.03196494281291962, 'train/grad_norm_pre_clip': 15.80185604095459, 'timing/data': 0.0007365019991993904, 'timing/model': 1.214699071017094, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.41}) +[2026-06-23 18:31:49,999][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.023807287216186523, 'train/grad_norm_pre_clip': 21.041950225830078, 'timing/data': 0.0009570300171617419, 'timing/model': 1.183984180999687, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.41}) +[2026-06-23 18:31:54,687][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.06003086641430855, 'train/grad_norm_pre_clip': 18.180294036865234, 'timing/data': 0.0008775729802437127, 'timing/model': 1.2311791170213837, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.41}) +[2026-06-23 18:31:59,094][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.135776549577713, 'train/grad_norm_pre_clip': 20.98546600341797, 'timing/data': 0.0009495460253674537, 'timing/model': 1.2130277449905407, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.41}) +[2026-06-23 18:32:03,499][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.18087923526763916, 'train/grad_norm_pre_clip': 21.492910385131836, 'timing/data': 0.0008372180163860321, 'timing/model': 1.2064710879931226, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.41}) +[2026-06-23 18:32:07,902][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.12002288550138474, 'train/grad_norm_pre_clip': 18.40666961669922, 'timing/data': 0.0007951410079840571, 'timing/model': 1.2024829410074744, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.41}) +[2026-06-23 18:32:12,414][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.08392871916294098, 'train/grad_norm_pre_clip': 16.96841812133789, 'timing/data': 0.0007976359920576215, 'timing/model': 1.1896863129804842, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.41}) +[2026-06-23 18:32:16,789][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.05333558842539787, 'train/grad_norm_pre_clip': 14.976828575134277, 'timing/data': 0.0009181880159303546, 'timing/model': 1.1855527779844124, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.41}) +[2026-06-23 18:32:21,203][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.045947834849357605, 'train/grad_norm_pre_clip': 21.920486450195312, 'timing/data': 0.0007812039984855801, 'timing/model': 1.2140320280159358, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.41}) +[2026-06-23 18:32:25,623][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.03180435299873352, 'train/grad_norm_pre_clip': 10.237166404724121, 'timing/data': 0.0009274950134567916, 'timing/model': 1.2054029679857194, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.41}) +[2026-06-23 18:32:30,271][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.02156209945678711, 'train/grad_norm_pre_clip': 31.376136779785156, 'timing/data': 0.0009096529975067824, 'timing/model': 1.191089715983253, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.41}) +[2026-06-23 18:32:34,694][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.2821716070175171, 'train/grad_norm_pre_clip': 38.91450119018555, 'timing/data': 0.0008156090043485165, 'timing/model': 1.2073483350104652, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.41}) +[2026-06-23 18:32:39,113][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.09934070706367493, 'train/grad_norm_pre_clip': 14.655013084411621, 'timing/data': 0.0007600060198456049, 'timing/model': 1.2049326070118695, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.41}) +[2026-06-23 18:32:43,519][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.21831920742988586, 'train/grad_norm_pre_clip': 25.73950958251953, 'timing/data': 0.00110814900835976, 'timing/model': 1.2120405420137104, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.41}) +[2026-06-23 18:32:48,138][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.16541534662246704, 'train/grad_norm_pre_clip': 25.302261352539062, 'timing/data': 0.0007550160225946456, 'timing/model': 1.2049610290268902, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.41}) +[2026-06-23 18:32:52,531][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.06557373702526093, 'train/grad_norm_pre_clip': 18.75786018371582, 'timing/data': 0.000774882995756343, 'timing/model': 1.215224270010367, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.42}) +[2026-06-23 18:32:56,942][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.07823483645915985, 'train/grad_norm_pre_clip': 22.666824340820312, 'timing/data': 0.0007487149850931019, 'timing/model': 1.210343835002277, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.42}) +[2026-06-23 18:33:01,329][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.1444811075925827, 'train/grad_norm_pre_clip': 25.372060775756836, 'timing/data': 0.0007753140234854072, 'timing/model': 1.2076019940141123, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.42}) +[2026-06-23 18:33:05,983][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.05518994480371475, 'train/grad_norm_pre_clip': 24.50056266784668, 'timing/data': 0.0011231469979975373, 'timing/model': 1.2054694700054824, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.42}) +[2026-06-23 18:33:10,412][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.027589503675699234, 'train/grad_norm_pre_clip': 30.650367736816406, 'timing/data': 0.0009657859918661416, 'timing/model': 1.2154078499879688, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.42}) +[2026-06-23 18:33:14,810][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.004978754557669163, 'train/grad_norm_pre_clip': 21.297073364257812, 'timing/data': 0.0007757339917588979, 'timing/model': 1.2054424509988166, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.42}) +[2026-06-23 18:33:19,243][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.16348755359649658, 'train/grad_norm_pre_clip': 14.225892066955566, 'timing/data': 0.0011207520146854222, 'timing/model': 1.2028863040031865, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.42}) +[2026-06-23 18:33:23,886][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.1362563669681549, 'train/grad_norm_pre_clip': 27.1106014251709, 'timing/data': 0.0009475919941905886, 'timing/model': 1.1877004099951591, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.42}) +[2026-06-23 18:33:28,284][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.012672088108956814, 'train/grad_norm_pre_clip': 8.716471672058105, 'timing/data': 0.0009776079969014972, 'timing/model': 1.2026143319963012, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.42}) +[2026-06-23 18:33:32,694][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.12549185752868652, 'train/grad_norm_pre_clip': 27.406808853149414, 'timing/data': 0.0010777130082715303, 'timing/model': 1.209290980012156, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.42}) +[2026-06-23 18:33:37,102][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.21881945431232452, 'train/grad_norm_pre_clip': 19.74464225769043, 'timing/data': 0.001176555990241468, 'timing/model': 1.1872360650158953, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.42}) +[2026-06-23 18:33:41,754][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.0459330677986145, 'train/grad_norm_pre_clip': 23.12355613708496, 'timing/data': 0.0012015320244245231, 'timing/model': 1.2043704520037863, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.42}) +[2026-06-23 18:33:46,194][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.06697936356067657, 'train/grad_norm_pre_clip': 24.28546714782715, 'timing/data': 0.0014685249770991504, 'timing/model': 1.2152665230096318, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.42}) +[2026-06-23 18:33:50,626][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.10663916170597076, 'train/grad_norm_pre_clip': 43.588043212890625, 'timing/data': 0.0010343329922761768, 'timing/model': 1.1979889720096253, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.42}) +[2026-06-23 18:33:54,986][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.13866378366947174, 'train/grad_norm_pre_clip': 26.959196090698242, 'timing/data': 0.0007509490242227912, 'timing/model': 1.1881829100020695, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.42}) +[2026-06-23 18:33:59,641][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.04675043001770973, 'train/grad_norm_pre_clip': 51.22163009643555, 'timing/data': 0.000833992991829291, 'timing/model': 1.2110697049938608, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.42}) +[2026-06-23 18:34:04,063][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.16750192642211914, 'train/grad_norm_pre_clip': 56.5439338684082, 'timing/data': 0.0008139259880408645, 'timing/model': 1.213259284995729, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.42}) +[2026-06-23 18:34:08,485][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.049436021596193314, 'train/grad_norm_pre_clip': 25.913297653198242, 'timing/data': 0.0009705550037324429, 'timing/model': 1.211969941999996, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.42}) +[2026-06-23 18:34:12,836][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.014633824117481709, 'train/grad_norm_pre_clip': 14.966712951660156, 'timing/data': 0.0008728240209165961, 'timing/model': 1.1898181400028989, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.42}) +[2026-06-23 18:34:17,463][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.17199110984802246, 'train/grad_norm_pre_clip': 41.73012161254883, 'timing/data': 0.0007843599887564778, 'timing/model': 1.1870227710169274, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.42}) +[2026-06-23 18:34:21,851][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.08963069319725037, 'train/grad_norm_pre_clip': 42.60709762573242, 'timing/data': 0.00150204700184986, 'timing/model': 1.2188629239972215, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.42}) +[2026-06-23 18:34:26,252][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.06289298832416534, 'train/grad_norm_pre_clip': 44.82420349121094, 'timing/data': 0.0007890989945735782, 'timing/model': 1.2116097419930156, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.42}) +[2026-06-23 18:34:30,616][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.05794433876872063, 'train/grad_norm_pre_clip': 58.2890625, 'timing/data': 0.0007471319986507297, 'timing/model': 1.189888326014625, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.42}) +[2026-06-23 18:34:35,259][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.07121553272008896, 'train/grad_norm_pre_clip': 24.870325088500977, 'timing/data': 0.0007749229844193906, 'timing/model': 1.2175569929822814, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.42}) +[2026-06-23 18:34:39,685][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.0024413960054516792, 'train/grad_norm_pre_clip': 25.685819625854492, 'timing/data': 0.0009508880029898137, 'timing/model': 1.2053553620062303, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.42}) +[2026-06-23 18:34:44,045][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.13608795404434204, 'train/grad_norm_pre_clip': 19.15964698791504, 'timing/data': 0.0007155629864428192, 'timing/model': 1.1861942210234702, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.42}) +[2026-06-23 18:34:48,481][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.09042948484420776, 'train/grad_norm_pre_clip': 20.526214599609375, 'timing/data': 0.0008285920193884522, 'timing/model': 1.2091822719958145, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.43}) +[2026-06-23 18:34:53,126][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1100 +[2026-06-23 18:35:50,714][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1100 +[2026-06-23 18:35:50,715][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1100 +[2026-06-23 18:36:27,880][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1100 +[2026-06-23 18:36:27,881][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.12742087244987488, 'train/grad_norm_pre_clip': 31.015987396240234, 'eval/action_loss/seconds': 57.58490611700108, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.057800184935331345, 'eval/latency_2/loss': 0.057800184935331345, 'eval/flappy/loss': 0.057800184935331345, 'eval/flappy/latency_2/loss': 0.057800184935331345, 'eval/flap_precision': 0.9208173690932312, 'eval/flap_recall': 0.9782903663500678, 'eval/flap_f1': 0.9486842105263158, 'eval/flap_rate_model': 0.2175, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9486842105263158, 'eval/latency_2/flap_precision': 0.9208173690932312, 'eval/latency_2/flap_recall': 0.9782903663500678, 'eval/latency_2/flap_f1': 0.9486842105263158, 'eval/latency_2/flap_rate_model': 0.2175, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9486842105263158, 'eval/action_classification/data_wait_seconds': 0.4672794044017792, 'eval/action_classification/predict_seconds': 36.538612365722656, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0012485290062613785, 'timing/model': 1.204965666984208, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.43}) +[2026-06-23 18:36:32,252][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.05418733134865761, 'train/grad_norm_pre_clip': 21.739521026611328, 'timing/data': 0.0007626700098626316, 'timing/model': 1.1903634369955398, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.43}) +[2026-06-23 18:36:36,569][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.2647554874420166, 'train/grad_norm_pre_clip': 25.186189651489258, 'timing/data': 0.0007255010132212192, 'timing/model': 1.178972817986505, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.43}) +[2026-06-23 18:36:41,163][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.14386409521102905, 'train/grad_norm_pre_clip': 20.240585327148438, 'timing/data': 0.0007519310165662318, 'timing/model': 1.4239428360015154, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.43}) +[2026-06-23 18:36:45,493][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.10537998378276825, 'train/grad_norm_pre_clip': 15.65597152709961, 'timing/data': 0.0008092159987427294, 'timing/model': 1.1826376719982363, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.43}) +[2026-06-23 18:36:49,805][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.04373880848288536, 'train/grad_norm_pre_clip': 18.536218643188477, 'timing/data': 0.000886158988578245, 'timing/model': 1.1773618620063644, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.43}) +[2026-06-23 18:36:54,150][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.01675860956311226, 'train/grad_norm_pre_clip': 16.82752799987793, 'timing/data': 0.0008284819778054953, 'timing/model': 1.1899856370000634, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.43}) +[2026-06-23 18:36:58,847][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.0632982850074768, 'train/grad_norm_pre_clip': 21.701814651489258, 'timing/data': 0.0007322739984374493, 'timing/model': 1.4943026050168555, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.43}) +[2026-06-23 18:37:03,171][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.07741059362888336, 'train/grad_norm_pre_clip': 20.663944244384766, 'timing/data': 0.0008757800096645951, 'timing/model': 1.182999596989248, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.43}) +[2026-06-23 18:37:07,573][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.03807544335722923, 'train/grad_norm_pre_clip': 22.38274574279785, 'timing/data': 0.000748243008274585, 'timing/model': 1.1779432029870804, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.43}) +[2026-06-23 18:37:11,950][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.11629833281040192, 'train/grad_norm_pre_clip': 42.328857421875, 'timing/data': 0.0007260020065587014, 'timing/model': 1.1826751759799663, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.43}) +[2026-06-23 18:37:16,550][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.04270598664879799, 'train/grad_norm_pre_clip': 22.698362350463867, 'timing/data': 0.0007345380145125091, 'timing/model': 1.436058260005666, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.43}) +[2026-06-23 18:37:20,984][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.27048689126968384, 'train/grad_norm_pre_clip': 21.379497528076172, 'timing/data': 0.000943213002756238, 'timing/model': 1.2602061499783304, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.43}) +[2026-06-23 18:37:25,353][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.03620832785964012, 'train/grad_norm_pre_clip': 26.89366912841797, 'timing/data': 0.000956848991336301, 'timing/model': 1.1925065869872924, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.43}) +[2026-06-23 18:37:30,002][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.09541229903697968, 'train/grad_norm_pre_clip': 31.273725509643555, 'timing/data': 0.0011359610070940107, 'timing/model': 1.2082701580075081, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.43}) +[2026-06-23 18:37:34,535][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.09275135397911072, 'train/grad_norm_pre_clip': 37.510986328125, 'timing/data': 0.0009651739965192974, 'timing/model': 1.2000605720095336, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.43}) +[2026-06-23 18:37:38,914][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.036578308790922165, 'train/grad_norm_pre_clip': 16.842546463012695, 'timing/data': 0.0007620590040460229, 'timing/model': 1.20387990900781, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.43}) +[2026-06-23 18:37:43,315][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.07218030095100403, 'train/grad_norm_pre_clip': 27.018468856811523, 'timing/data': 0.0008146360050886869, 'timing/model': 1.2048675190017093, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.43}) +[2026-06-23 18:37:47,696][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.07128667831420898, 'train/grad_norm_pre_clip': 53.15351104736328, 'timing/data': 0.0007147119904402643, 'timing/model': 1.208256149024237, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.43}) +[2026-06-23 18:37:52,336][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.13968290388584137, 'train/grad_norm_pre_clip': 30.632568359375, 'timing/data': 0.0007433139835484326, 'timing/model': 1.4493776019953657, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.43}) +[2026-06-23 18:37:56,696][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.16513776779174805, 'train/grad_norm_pre_clip': 49.76905822753906, 'timing/data': 0.0007834289863239974, 'timing/model': 1.202933195978403, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.43}) +[2026-06-23 18:38:01,066][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.21846146881580353, 'train/grad_norm_pre_clip': 35.21118927001953, 'timing/data': 0.0007551060116384178, 'timing/model': 1.1837679439922795, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.43}) +[2026-06-23 18:38:05,454][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.07301211357116699, 'train/grad_norm_pre_clip': 31.601713180541992, 'timing/data': 0.0012561220210045576, 'timing/model': 1.20676799499779, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.43}) +[2026-06-23 18:38:10,101][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.02686610072851181, 'train/grad_norm_pre_clip': 27.794340133666992, 'timing/data': 0.0007994589977897704, 'timing/model': 1.4415237049979623, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.43}) +[2026-06-23 18:38:14,546][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.13502319157123566, 'train/grad_norm_pre_clip': 47.93840789794922, 'timing/data': 0.0010167110012844205, 'timing/model': 1.2124238410033286, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.44}) +[2026-06-23 18:38:18,926][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.07528121024370193, 'train/grad_norm_pre_clip': 20.226470947265625, 'timing/data': 0.0008398630016017705, 'timing/model': 1.1884298429940827, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.44}) +[2026-06-23 18:38:23,337][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.12188028544187546, 'train/grad_norm_pre_clip': 12.550423622131348, 'timing/data': 0.0007243300206027925, 'timing/model': 1.204685964010423, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.44}) +[2026-06-23 18:38:28,048][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.20407548546791077, 'train/grad_norm_pre_clip': 21.2249755859375, 'timing/data': 0.0008257059962488711, 'timing/model': 1.5053593680204358, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.44}) +[2026-06-23 18:38:32,440][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.06624748557806015, 'train/grad_norm_pre_clip': 14.926589012145996, 'timing/data': 0.000770815007854253, 'timing/model': 1.201815426000394, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.44}) +[2026-06-23 18:38:36,784][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.11375442147254944, 'train/grad_norm_pre_clip': 18.95891571044922, 'timing/data': 0.0008990629867184907, 'timing/model': 1.1875620080099907, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.44}) +[2026-06-23 18:38:41,159][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.09083592891693115, 'train/grad_norm_pre_clip': 19.38263511657715, 'timing/data': 0.0007503670058213174, 'timing/model': 1.1872034989937674, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.44}) +[2026-06-23 18:38:45,743][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.06497722864151001, 'train/grad_norm_pre_clip': 21.092531204223633, 'timing/data': 0.0007937579939607531, 'timing/model': 1.2069287349877413, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.44}) +[2026-06-23 18:38:50,117][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.051917701959609985, 'train/grad_norm_pre_clip': 15.207466125488281, 'timing/data': 0.0007865750230848789, 'timing/model': 1.2087268349714577, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.44}) +[2026-06-23 18:38:54,509][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.09634749591350555, 'train/grad_norm_pre_clip': 13.79627799987793, 'timing/data': 0.0008154469833243638, 'timing/model': 1.2035301709838677, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.44}) +[2026-06-23 18:38:58,846][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.11904668807983398, 'train/grad_norm_pre_clip': 20.350788116455078, 'timing/data': 0.0007756339909974486, 'timing/model': 1.182226617995184, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.44}) +[2026-06-23 18:39:03,476][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.018650755286216736, 'train/grad_norm_pre_clip': 18.397262573242188, 'timing/data': 0.0007583120022900403, 'timing/model': 1.4540274899918586, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.44}) +[2026-06-23 18:39:07,841][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.21032807230949402, 'train/grad_norm_pre_clip': 32.30764389038086, 'timing/data': 0.0007478319748770446, 'timing/model': 1.1858483850082848, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.44}) +[2026-06-23 18:39:12,181][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.031888317316770554, 'train/grad_norm_pre_clip': 16.033483505249023, 'timing/data': 0.0007556170166935772, 'timing/model': 1.1857814670074731, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.44}) +[2026-06-23 18:39:16,538][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.1598745733499527, 'train/grad_norm_pre_clip': 26.866003036499023, 'timing/data': 0.0007906320097390562, 'timing/model': 1.1840779329941142, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.44}) +[2026-06-23 18:39:21,149][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.11622638255357742, 'train/grad_norm_pre_clip': 23.93526840209961, 'timing/data': 0.0008168010099325329, 'timing/model': 1.4382886470120866, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.44}) +[2026-06-23 18:39:25,496][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.06590329110622406, 'train/grad_norm_pre_clip': 9.972962379455566, 'timing/data': 0.0007354300178121775, 'timing/model': 1.1833469620032702, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.44}) +[2026-06-23 18:39:29,857][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.03440102934837341, 'train/grad_norm_pre_clip': 15.378620147705078, 'timing/data': 0.0012041270092595369, 'timing/model': 1.2032462790084537, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.44}) +[2026-06-23 18:39:34,228][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.1025015115737915, 'train/grad_norm_pre_clip': 16.700910568237305, 'timing/data': 0.0007346989877987653, 'timing/model': 1.1842771130031906, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.44}) +[2026-06-23 18:39:38,829][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.17308077216148376, 'train/grad_norm_pre_clip': 31.328264236450195, 'timing/data': 0.0009660260111559182, 'timing/model': 1.4308091670100112, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.44}) +[2026-06-23 18:39:43,179][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.20349636673927307, 'train/grad_norm_pre_clip': 21.21609115600586, 'timing/data': 0.0008735150040592998, 'timing/model': 1.185551430011401, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.44}) +[2026-06-23 18:39:47,549][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.07714755833148956, 'train/grad_norm_pre_clip': 54.00170135498047, 'timing/data': 0.0007340880110859871, 'timing/model': 1.1882971119775902, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.44}) +[2026-06-23 18:39:51,951][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.05436704307794571, 'train/grad_norm_pre_clip': 14.916461944580078, 'timing/data': 0.000723578006727621, 'timing/model': 1.184937315993011, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.44}) +[2026-06-23 18:39:56,567][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.1535792350769043, 'train/grad_norm_pre_clip': 25.69841766357422, 'timing/data': 0.0010535379988141358, 'timing/model': 1.4299756649998017, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.44}) +[2026-06-23 18:40:00,927][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.028226209804415703, 'train/grad_norm_pre_clip': 18.446428298950195, 'timing/data': 0.0008026639989111573, 'timing/model': 1.1859896229871083, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.44}) +[2026-06-23 18:40:05,265][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.049719758331775665, 'train/grad_norm_pre_clip': 32.901084899902344, 'timing/data': 0.0009276750206481665, 'timing/model': 1.1859407989832107, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.44}) +[2026-06-23 18:40:09,641][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.15900671482086182, 'train/grad_norm_pre_clip': 23.582595825195312, 'timing/data': 0.000813554011983797, 'timing/model': 1.1891907830140553, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.45}) +[2026-06-23 18:40:14,236][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.26434850692749023, 'train/grad_norm_pre_clip': 42.392215728759766, 'timing/data': 0.0011688410013448447, 'timing/model': 1.4293639439856634, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.45}) +[2026-06-23 18:40:18,615][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.08168989419937134, 'train/grad_norm_pre_clip': 30.826061248779297, 'timing/data': 0.0007869450200814754, 'timing/model': 1.1961025740019977, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.45}) +[2026-06-23 18:40:22,967][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.020771823823451996, 'train/grad_norm_pre_clip': 11.409799575805664, 'timing/data': 0.000817702995846048, 'timing/model': 1.1856292659940664, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.45}) +[2026-06-23 18:40:27,336][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.15670235455036163, 'train/grad_norm_pre_clip': 18.488536834716797, 'timing/data': 0.0007444660004694015, 'timing/model': 1.2070927750028204, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.45}) +[2026-06-23 18:40:31,988][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.0580279715359211, 'train/grad_norm_pre_clip': 20.397811889648438, 'timing/data': 0.0008518949907738715, 'timing/model': 1.442801197990775, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.45}) +[2026-06-23 18:40:36,378][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.18095166981220245, 'train/grad_norm_pre_clip': 21.32293701171875, 'timing/data': 0.0007393370033241808, 'timing/model': 1.2036247049982194, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.45}) +[2026-06-23 18:40:40,743][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.09852922707796097, 'train/grad_norm_pre_clip': 29.139665603637695, 'timing/data': 0.0007103240059223026, 'timing/model': 1.1826566740055569, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.45}) +[2026-06-23 18:40:45,112][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.18090836703777313, 'train/grad_norm_pre_clip': 39.35422134399414, 'timing/data': 0.0009725989948492497, 'timing/model': 1.212951941008214, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.45}) +[2026-06-23 18:40:49,685][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.05834118276834488, 'train/grad_norm_pre_clip': 67.34262084960938, 'timing/data': 0.0007578409858979285, 'timing/model': 1.1843934479984455, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.45}) +[2026-06-23 18:40:54,055][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.0681835487484932, 'train/grad_norm_pre_clip': 9.190996170043945, 'timing/data': 0.0007258519763126969, 'timing/model': 1.188188760977937, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.45}) +[2026-06-23 18:40:58,489][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.16769075393676758, 'train/grad_norm_pre_clip': 48.54639434814453, 'timing/data': 0.0009991370025090873, 'timing/model': 1.2081037960015237, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.45}) +[2026-06-23 18:41:02,915][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.10617567598819733, 'train/grad_norm_pre_clip': 48.91447067260742, 'timing/data': 0.0009125369833782315, 'timing/model': 1.2126060760056134, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.45}) +[2026-06-23 18:41:07,546][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.19348913431167603, 'train/grad_norm_pre_clip': 32.47055435180664, 'timing/data': 0.0009383040014654398, 'timing/model': 1.4378070370003115, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.45}) +[2026-06-23 18:41:11,919][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.1387646645307541, 'train/grad_norm_pre_clip': 37.369102478027344, 'timing/data': 0.0007420620240736753, 'timing/model': 1.2060333270055708, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.45}) +[2026-06-23 18:41:16,312][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.05016081780195236, 'train/grad_norm_pre_clip': 23.781415939331055, 'timing/data': 0.0006968689849600196, 'timing/model': 1.2074234530155081, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.45}) +[2026-06-23 18:41:20,742][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.2882496118545532, 'train/grad_norm_pre_clip': 51.270538330078125, 'timing/data': 0.0008048079907894135, 'timing/model': 1.2190865419979673, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.45}) +[2026-06-23 18:41:25,325][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.014177880249917507, 'train/grad_norm_pre_clip': 29.554519653320312, 'timing/data': 0.0007382449985016137, 'timing/model': 1.2093154939939268, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.45}) +[2026-06-23 18:41:29,709][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.019815752282738686, 'train/grad_norm_pre_clip': 19.36634635925293, 'timing/data': 0.0008547909965272993, 'timing/model': 1.1934645650035236, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.45}) +[2026-06-23 18:41:34,075][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.2337442934513092, 'train/grad_norm_pre_clip': 23.66027069091797, 'timing/data': 0.0009140800102613866, 'timing/model': 1.1901733260019682, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.45}) +[2026-06-23 18:41:38,485][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.008863084018230438, 'train/grad_norm_pre_clip': 22.02281951904297, 'timing/data': 0.0009866440086625516, 'timing/model': 1.218123477010522, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.45}) +[2026-06-23 18:41:43,170][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.0341489240527153, 'train/grad_norm_pre_clip': 36.44514083862305, 'timing/data': 0.0007628199819009751, 'timing/model': 1.467544931016164, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.45}) +[2026-06-23 18:41:47,568][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.11915073543787003, 'train/grad_norm_pre_clip': 56.74525833129883, 'timing/data': 0.0008744469960220158, 'timing/model': 1.207996737997746, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.45}) +[2026-06-23 18:41:51,968][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.19592349231243134, 'train/grad_norm_pre_clip': 40.43760299682617, 'timing/data': 0.000921433005714789, 'timing/model': 1.2044231870095246, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.45}) +[2026-06-23 18:41:56,336][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.17669150233268738, 'train/grad_norm_pre_clip': 33.72196578979492, 'timing/data': 0.0007565490086562932, 'timing/model': 1.1873778009903617, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.45}) +[2026-06-23 18:42:00,988][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.16497114300727844, 'train/grad_norm_pre_clip': 32.11341094970703, 'timing/data': 0.0007992770115379244, 'timing/model': 1.4418640350049827, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.45}) +[2026-06-23 18:42:05,411][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.021112017333507538, 'train/grad_norm_pre_clip': 12.50101089477539, 'timing/data': 0.0007824459753464907, 'timing/model': 1.212639236007817, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.46}) +[2026-06-23 18:42:09,783][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.13093599677085876, 'train/grad_norm_pre_clip': 35.3582649230957, 'timing/data': 0.000748352991649881, 'timing/model': 1.1879573519981932, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.46}) +[2026-06-23 18:42:14,217][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.15396423637866974, 'train/grad_norm_pre_clip': 16.623777389526367, 'timing/data': 0.0007874150178395212, 'timing/model': 1.2164294170215726, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.46}) +[2026-06-23 18:42:18,823][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.07467430830001831, 'train/grad_norm_pre_clip': 23.948232650756836, 'timing/data': 0.0014551409985870123, 'timing/model': 1.203857911983505, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.46}) +[2026-06-23 18:42:23,199][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.07697999477386475, 'train/grad_norm_pre_clip': 14.045623779296875, 'timing/data': 0.000773019011830911, 'timing/model': 1.1900153160095215, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.46}) +[2026-06-23 18:42:27,596][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.1489800065755844, 'train/grad_norm_pre_clip': 20.72661781311035, 'timing/data': 0.0012030140205752105, 'timing/model': 1.1899552120012231, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.46}) +[2026-06-23 18:42:32,016][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.11216208338737488, 'train/grad_norm_pre_clip': 12.616981506347656, 'timing/data': 0.001061783026671037, 'timing/model': 1.1887842449941672, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.46}) +[2026-06-23 18:42:36,697][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.05497804284095764, 'train/grad_norm_pre_clip': 36.878318786621094, 'timing/data': 0.000772869010688737, 'timing/model': 1.4541034220019355, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.46}) +[2026-06-23 18:42:41,121][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.1121082678437233, 'train/grad_norm_pre_clip': 19.319265365600586, 'timing/data': 0.0010659300023689866, 'timing/model': 1.2082424810214434, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.46}) +[2026-06-23 18:42:45,524][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.00556153804063797, 'train/grad_norm_pre_clip': 21.5661563873291, 'timing/data': 0.0009679389768280089, 'timing/model': 1.208587325992994, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.46}) +[2026-06-23 18:42:49,921][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.024368613958358765, 'train/grad_norm_pre_clip': 17.865848541259766, 'timing/data': 0.0007782389875501394, 'timing/model': 1.2067304599913768, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.46}) +[2026-06-23 18:42:54,567][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.04964882507920265, 'train/grad_norm_pre_clip': 16.294992446899414, 'timing/data': 0.000897849997272715, 'timing/model': 1.458289425005205, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.46}) +[2026-06-23 18:42:58,960][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.096761055290699, 'train/grad_norm_pre_clip': 16.617549896240234, 'timing/data': 0.0007691319915466011, 'timing/model': 1.2091622400039341, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.46}) +[2026-06-23 18:43:03,376][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.06942091882228851, 'train/grad_norm_pre_clip': 16.54776954650879, 'timing/data': 0.0007596540090162307, 'timing/model': 1.207440514001064, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.46}) +[2026-06-23 18:43:07,778][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.10384196788072586, 'train/grad_norm_pre_clip': 40.709373474121094, 'timing/data': 0.0007440059853252023, 'timing/model': 1.2087376790004782, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.46}) +[2026-06-23 18:43:12,429][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.021182071417570114, 'train/grad_norm_pre_clip': 22.059818267822266, 'timing/data': 0.0008886729774530977, 'timing/model': 1.4587282220018096, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.46}) +[2026-06-23 18:43:16,801][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.04249441251158714, 'train/grad_norm_pre_clip': 24.204946517944336, 'timing/data': 0.0009040820004884154, 'timing/model': 1.2030450080055743, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.46}) +[2026-06-23 18:43:21,175][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.0002616054262034595, 'train/grad_norm_pre_clip': 31.774728775024414, 'timing/data': 0.0007483429799322039, 'timing/model': 1.1889692329859827, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.46}) +[2026-06-23 18:43:25,582][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.08014896512031555, 'train/grad_norm_pre_clip': 30.59049415588379, 'timing/data': 0.0007729289936833084, 'timing/model': 1.2026781940076035, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.46}) +[2026-06-23 18:43:30,294][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.0012943276669830084, 'train/grad_norm_pre_clip': 78.9837646484375, 'timing/data': 0.0009435540123376995, 'timing/model': 1.4568894779949915, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.46}) +[2026-06-23 18:43:34,703][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.06274425983428955, 'train/grad_norm_pre_clip': 15.804374694824219, 'timing/data': 0.000761097006034106, 'timing/model': 1.2124907689867541, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.46}) +[2026-06-23 18:43:39,085][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.04109515622258186, 'train/grad_norm_pre_clip': 21.370349884033203, 'timing/data': 0.0008233920088969171, 'timing/model': 1.1876678960106801, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.46}) +[2026-06-23 18:43:43,492][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.22453610599040985, 'train/grad_norm_pre_clip': 28.540740966796875, 'timing/data': 0.0007671490020584315, 'timing/model': 1.2015760159993079, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.46}) +[2026-06-23 18:43:48,131][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.008475011214613914, 'train/grad_norm_pre_clip': 18.76653480529785, 'timing/data': 0.000919710990274325, 'timing/model': 1.4517020549974404, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.46}) +[2026-06-23 18:43:52,543][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1200 +[2026-06-23 18:44:50,458][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1200 +[2026-06-23 18:44:50,459][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1200 +[2026-06-23 18:45:27,615][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1200 +[2026-06-23 18:45:27,617][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.2562161087989807, 'train/grad_norm_pre_clip': 20.764318466186523, 'eval/action_loss/seconds': 57.91206791298464, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.05002394691109657, 'eval/latency_2/loss': 0.05002394691109657, 'eval/flappy/loss': 0.05002394691109657, 'eval/flappy/latency_2/loss': 0.05002394691109657, 'eval/flap_precision': 0.9550408719346049, 'eval/flap_recall': 0.9511533242876526, 'eval/flap_f1': 0.9530931339225016, 'eval/flap_rate_model': 0.2038888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9530931339225016, 'eval/latency_2/flap_precision': 0.9550408719346049, 'eval/latency_2/flap_recall': 0.9511533242876526, 'eval/latency_2/flap_f1': 0.9530931339225016, 'eval/latency_2/flap_rate_model': 0.2038888888888889, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9530931339225016, 'eval/action_classification/data_wait_seconds': 0.5402835607528687, 'eval/action_classification/predict_seconds': 36.46790313720703, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0011865840060636401, 'timing/model': 1.2137620479916222, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.46}) +[2026-06-23 18:45:31,992][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.060933127999305725, 'train/grad_norm_pre_clip': 20.720237731933594, 'timing/data': 0.0009483430185355246, 'timing/model': 1.1850665339734405, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.46}) +[2026-06-23 18:45:36,312][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.09719078242778778, 'train/grad_norm_pre_clip': 20.927963256835938, 'timing/data': 0.0010008099779952317, 'timing/model': 1.1726501980156172, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.47}) +[2026-06-23 18:45:40,872][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.10635851323604584, 'train/grad_norm_pre_clip': 13.380083084106445, 'timing/data': 0.0009349779866170138, 'timing/model': 1.1727141640149057, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.47}) +[2026-06-23 18:45:45,214][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.10000725090503693, 'train/grad_norm_pre_clip': 23.53664207458496, 'timing/data': 0.000775523018091917, 'timing/model': 1.175546047015814, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.47}) +[2026-06-23 18:45:49,550][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.052522577345371246, 'train/grad_norm_pre_clip': 20.411575317382812, 'timing/data': 0.0010302650043740869, 'timing/model': 1.2042958329839166, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.47}) +[2026-06-23 18:45:53,870][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.1048460528254509, 'train/grad_norm_pre_clip': 15.61356258392334, 'timing/data': 0.0009079489973373711, 'timing/model': 1.178390971996123, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.47}) +[2026-06-23 18:45:58,481][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.14398261904716492, 'train/grad_norm_pre_clip': 15.481538772583008, 'timing/data': 0.0009454879909753799, 'timing/model': 1.2078954729950055, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.47}) +[2026-06-23 18:46:02,827][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.046056345105171204, 'train/grad_norm_pre_clip': 16.94978141784668, 'timing/data': 0.00073590999818407, 'timing/model': 1.178888888010988, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.47}) +[2026-06-23 18:46:07,179][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.06814184039831161, 'train/grad_norm_pre_clip': 8.579733848571777, 'timing/data': 0.000841697008581832, 'timing/model': 1.1875452990061603, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.47}) +[2026-06-23 18:46:11,507][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.11093375831842422, 'train/grad_norm_pre_clip': 16.384523391723633, 'timing/data': 0.0008282510098069906, 'timing/model': 1.1809990049805492, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.47}) +[2026-06-23 18:46:16,075][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.025207852944731712, 'train/grad_norm_pre_clip': 22.91398048400879, 'timing/data': 0.0010136939818039536, 'timing/model': 1.182421011995757, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.47}) +[2026-06-23 18:46:20,449][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.15539821982383728, 'train/grad_norm_pre_clip': 33.54602813720703, 'timing/data': 0.0011879660014528781, 'timing/model': 1.1809620729763992, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.47}) +[2026-06-23 18:46:24,823][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.042568646371364594, 'train/grad_norm_pre_clip': 15.363947868347168, 'timing/data': 0.0010988920112140477, 'timing/model': 1.1779341439832933, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.47}) +[2026-06-23 18:46:29,230][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.2612752914428711, 'train/grad_norm_pre_clip': 25.834808349609375, 'timing/data': 0.0011046019790228456, 'timing/model': 1.222470703010913, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.47}) +[2026-06-23 18:46:33,864][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.07019991427659988, 'train/grad_norm_pre_clip': 30.37086296081543, 'timing/data': 0.0007900809869170189, 'timing/model': 1.2177895870117936, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.47}) +[2026-06-23 18:46:38,249][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.06132752075791359, 'train/grad_norm_pre_clip': 38.93896484375, 'timing/data': 0.0009897100098896772, 'timing/model': 1.199778119014809, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.47}) +[2026-06-23 18:46:42,669][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.056745775043964386, 'train/grad_norm_pre_clip': 16.33213233947754, 'timing/data': 0.0007889090047683567, 'timing/model': 1.2160275560163427, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.47}) +[2026-06-23 18:46:47,038][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.042913343757390976, 'train/grad_norm_pre_clip': 12.565045356750488, 'timing/data': 0.0008880309760570526, 'timing/model': 1.185621683980571, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.47}) +[2026-06-23 18:46:51,683][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.03325792774558067, 'train/grad_norm_pre_clip': 8.19234848022461, 'timing/data': 0.0011287459929008037, 'timing/model': 1.2042662400053814, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.47}) +[2026-06-23 18:46:56,100][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.0819796696305275, 'train/grad_norm_pre_clip': 16.815855026245117, 'timing/data': 0.0008508930040989071, 'timing/model': 1.2036473540065344, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.47}) +[2026-06-23 18:47:00,518][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.2274581342935562, 'train/grad_norm_pre_clip': 25.168201446533203, 'timing/data': 0.0007437349995598197, 'timing/model': 1.21483546498348, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.47}) +[2026-06-23 18:47:04,880][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.2766236364841461, 'train/grad_norm_pre_clip': 18.511638641357422, 'timing/data': 0.0007444560178555548, 'timing/model': 1.2002156559901778, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.47}) +[2026-06-23 18:47:09,537][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.05560687184333801, 'train/grad_norm_pre_clip': 13.44081974029541, 'timing/data': 0.0008046069997362792, 'timing/model': 1.2101575230190065, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.47}) +[2026-06-23 18:47:13,911][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.059259045869112015, 'train/grad_norm_pre_clip': 21.070375442504883, 'timing/data': 0.0007376839930657297, 'timing/model': 1.1996710860112216, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.47}) +[2026-06-23 18:47:18,307][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.24842065572738647, 'train/grad_norm_pre_clip': 19.5546875, 'timing/data': 0.0007201609842013568, 'timing/model': 1.2036561610002536, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.47}) +[2026-06-23 18:47:22,934][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.021175960078835487, 'train/grad_norm_pre_clip': 11.392102241516113, 'timing/data': 0.0011095909867435694, 'timing/model': 1.4338370070036035, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.47}) +[2026-06-23 18:47:27,304][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.014429695904254913, 'train/grad_norm_pre_clip': 18.044937133789062, 'timing/data': 0.0007538729987572879, 'timing/model': 1.2193466529715806, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.47}) +[2026-06-23 18:47:31,741][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.07755164057016373, 'train/grad_norm_pre_clip': 27.766845703125, 'timing/data': 0.0007821960025466979, 'timing/model': 1.2298064469941892, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.48}) +[2026-06-23 18:47:36,160][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.023270083591341972, 'train/grad_norm_pre_clip': 17.069204330444336, 'timing/data': 0.0008054990030359477, 'timing/model': 1.2099579730129335, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.48}) +[2026-06-23 18:47:40,691][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.22717754542827606, 'train/grad_norm_pre_clip': 29.08432388305664, 'timing/data': 0.0007387959922198206, 'timing/model': 1.1917700890044216, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.48}) +[2026-06-23 18:47:45,088][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.17251181602478027, 'train/grad_norm_pre_clip': 20.371532440185547, 'timing/data': 0.000777457986259833, 'timing/model': 1.199742345022969, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.48}) +[2026-06-23 18:47:49,468][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.13934819400310516, 'train/grad_norm_pre_clip': 24.297637939453125, 'timing/data': 0.0007195800135377795, 'timing/model': 1.2182281959976535, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.48}) +[2026-06-23 18:47:53,825][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.12667199969291687, 'train/grad_norm_pre_clip': 25.1256160736084, 'timing/data': 0.0007476519967895001, 'timing/model': 1.1870892299921252, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.48}) +[2026-06-23 18:47:58,499][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.04611034691333771, 'train/grad_norm_pre_clip': 20.270883560180664, 'timing/data': 0.0010613019985612482, 'timing/model': 1.4909913869923912, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.48}) +[2026-06-23 18:48:02,893][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.22249113023281097, 'train/grad_norm_pre_clip': 29.35840606689453, 'timing/data': 0.0007218840182758868, 'timing/model': 1.207979836006416, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.48}) +[2026-06-23 18:48:07,238][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.04909596964716911, 'train/grad_norm_pre_clip': 15.664624214172363, 'timing/data': 0.0007098320056684315, 'timing/model': 1.1806193260126747, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.48}) +[2026-06-23 18:48:11,585][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.055122245103120804, 'train/grad_norm_pre_clip': 17.228910446166992, 'timing/data': 0.0011366310063749552, 'timing/model': 1.2005249419889878, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.48}) +[2026-06-23 18:48:16,189][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.05170414224267006, 'train/grad_norm_pre_clip': 17.39824104309082, 'timing/data': 0.0008578760025557131, 'timing/model': 1.434503564989427, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.48}) +[2026-06-23 18:48:20,546][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.04578540101647377, 'train/grad_norm_pre_clip': 13.423332214355469, 'timing/data': 0.0007431539997924119, 'timing/model': 1.1823614409950096, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.48}) +[2026-06-23 18:48:24,927][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.017921002581715584, 'train/grad_norm_pre_clip': 10.575478553771973, 'timing/data': 0.0007289579953067005, 'timing/model': 1.1834323900111485, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.48}) +[2026-06-23 18:48:29,337][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.0940326526761055, 'train/grad_norm_pre_clip': 20.381479263305664, 'timing/data': 0.0008516749949194491, 'timing/model': 1.2008741209865548, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.48}) +[2026-06-23 18:48:33,960][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.10385958105325699, 'train/grad_norm_pre_clip': 18.10199737548828, 'timing/data': 0.0009649140120018274, 'timing/model': 1.440276609006105, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.48}) +[2026-06-23 18:48:38,368][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.02283734828233719, 'train/grad_norm_pre_clip': 20.38454246520996, 'timing/data': 0.0011001039820257574, 'timing/model': 1.199756090005394, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.48}) +[2026-06-23 18:48:42,758][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.07376790046691895, 'train/grad_norm_pre_clip': 17.50701332092285, 'timing/data': 0.0007237380195874721, 'timing/model': 1.2018337400222663, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.48}) +[2026-06-23 18:48:47,111][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.0031144688837230206, 'train/grad_norm_pre_clip': 18.86058807373047, 'timing/data': 0.0008814589818939567, 'timing/model': 1.202099668997107, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.48}) +[2026-06-23 18:48:51,643][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.012656952254474163, 'train/grad_norm_pre_clip': 19.166507720947266, 'timing/data': 0.0008619239961262792, 'timing/model': 1.1824000919878017, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.48}) +[2026-06-23 18:48:56,035][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.00708357896655798, 'train/grad_norm_pre_clip': 12.074703216552734, 'timing/data': 0.0008433289767708629, 'timing/model': 1.1835118859889917, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.48}) +[2026-06-23 18:49:00,372][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.13971558213233948, 'train/grad_norm_pre_clip': 21.910219192504883, 'timing/data': 0.0007399980095215142, 'timing/model': 1.1840163570013829, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.48}) +[2026-06-23 18:49:04,716][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.059653088450431824, 'train/grad_norm_pre_clip': 10.446959495544434, 'timing/data': 0.0007311720110010356, 'timing/model': 1.1860968609980773, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.48}) +[2026-06-23 19:03:04,679][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1250 +[2026-06-23 19:03:04,680][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.04367215186357498, 'train/grad_norm_pre_clip': 27.8978328704834, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 216.78000000000762, 'rl_games_eval/mid_train/mean_length': 1764.4, 'rl_games_eval/mid_train/std_reward': 173.94496141022182, 'rl_games_eval/mid_train/std_length': 1402.9521160752422, 'rl_games_eval/mid_train/macro_mean_reward': 216.78000000000762, 'rl_games_eval/mid_train/macro_mean_length': 1764.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 216.78000000000762, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1764.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 173.94496141022182, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1402.9521160752422, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 216.78000000000762, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0007101119845174253, 'timing/model': 1.437284412997542, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.48}) +[2026-06-23 19:03:08,929][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.012909011915326118, 'train/grad_norm_pre_clip': 40.57133483886719, 'timing/data': 0.0008496700029354542, 'timing/model': 1.1629961099824868, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.48}) +[2026-06-23 19:03:13,116][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.010146954096853733, 'train/grad_norm_pre_clip': 50.368629455566406, 'timing/data': 0.0007485240057576448, 'timing/model': 1.1389417620084714, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.48}) +[2026-06-23 19:03:17,343][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.0062482054345309734, 'train/grad_norm_pre_clip': 37.519989013671875, 'timing/data': 0.0007260919956024736, 'timing/model': 1.1753201030078344, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.48}) +[2026-06-23 19:03:21,741][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.06787267327308655, 'train/grad_norm_pre_clip': 20.88724136352539, 'timing/data': 0.0007733590027783066, 'timing/model': 1.1420245560002513, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.49}) +[2026-06-23 19:03:25,993][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.14041322469711304, 'train/grad_norm_pre_clip': 34.079002380371094, 'timing/data': 0.00079763398389332, 'timing/model': 1.1859419200045522, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.49}) +[2026-06-23 19:03:30,206][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.15129071474075317, 'train/grad_norm_pre_clip': 23.751026153564453, 'timing/data': 0.0007319230062421411, 'timing/model': 1.1496801410103217, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.49}) +[2026-06-23 19:03:34,484][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.0048083835281431675, 'train/grad_norm_pre_clip': 26.593008041381836, 'timing/data': 0.0009825460147112608, 'timing/model': 1.1731673699978273, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.49}) +[2026-06-23 19:03:38,921][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.027941066771745682, 'train/grad_norm_pre_clip': 28.843914031982422, 'timing/data': 0.000803545000962913, 'timing/model': 1.1492220920044929, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.49}) +[2026-06-23 19:03:43,184][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.02998313307762146, 'train/grad_norm_pre_clip': 12.69124698638916, 'timing/data': 0.000982885976554826, 'timing/model': 1.1767919449775945, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.49}) +[2026-06-23 19:03:47,442][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.021863432601094246, 'train/grad_norm_pre_clip': 16.985010147094727, 'timing/data': 0.0010894729930441827, 'timing/model': 1.1853138380101882, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.49}) +[2026-06-23 19:03:51,746][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.08268098533153534, 'train/grad_norm_pre_clip': 16.815223693847656, 'timing/data': 0.0013577189820352942, 'timing/model': 1.1738598239899147, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.49}) +[2026-06-23 19:03:56,273][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.10816922783851624, 'train/grad_norm_pre_clip': 27.783567428588867, 'timing/data': 0.0008063600107561797, 'timing/model': 1.1793029370019212, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.49}) +[2026-06-23 19:04:00,586][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.16145163774490356, 'train/grad_norm_pre_clip': 17.002201080322266, 'timing/data': 0.0008013609913177788, 'timing/model': 1.1884461469890084, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.49}) +[2026-06-23 19:04:04,954][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.079877570271492, 'train/grad_norm_pre_clip': 23.790977478027344, 'timing/data': 0.001094473002012819, 'timing/model': 1.1901982290146407, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.49}) +[2026-06-23 19:04:09,288][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.03310130164027214, 'train/grad_norm_pre_clip': 12.330527305603027, 'timing/data': 0.0007515799952670932, 'timing/model': 1.184169360989472, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.49}) +[2026-06-23 19:04:13,780][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.12645882368087769, 'train/grad_norm_pre_clip': 24.20473861694336, 'timing/data': 0.0011920630058739334, 'timing/model': 1.1670582530205138, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.49}) +[2026-06-23 19:04:18,120][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.05171537399291992, 'train/grad_norm_pre_clip': 12.946322441101074, 'timing/data': 0.0007361799944192171, 'timing/model': 1.188194176997058, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.49}) +[2026-06-23 19:04:22,463][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.058379996567964554, 'train/grad_norm_pre_clip': 12.33963394165039, 'timing/data': 0.0008083129941951483, 'timing/model': 1.1965498609934002, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.49}) +[2026-06-23 19:04:26,781][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.21826742589473724, 'train/grad_norm_pre_clip': 18.213682174682617, 'timing/data': 0.0008837030036374927, 'timing/model': 1.2027051099867094, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.49}) +[2026-06-23 19:04:31,290][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.0887998566031456, 'train/grad_norm_pre_clip': 16.206218719482422, 'timing/data': 0.0008996519900392741, 'timing/model': 1.1940418839803897, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.49}) +[2026-06-23 19:04:35,610][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.10897602140903473, 'train/grad_norm_pre_clip': 48.09528350830078, 'timing/data': 0.000774921994889155, 'timing/model': 1.2009411480103154, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.49}) +[2026-06-23 19:04:39,955][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.11760643124580383, 'train/grad_norm_pre_clip': 27.13218116760254, 'timing/data': 0.0007742809830233455, 'timing/model': 1.1757786990201566, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.49}) +[2026-06-23 19:04:44,314][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.04676835983991623, 'train/grad_norm_pre_clip': 15.66922664642334, 'timing/data': 0.0009616469906177372, 'timing/model': 1.195222816983005, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.49}) +[2026-06-23 19:04:48,839][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.09349865466356277, 'train/grad_norm_pre_clip': 17.24607276916504, 'timing/data': 0.0007805419736541808, 'timing/model': 1.1735301399894524, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.49}) +[2026-06-23 19:04:53,202][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.239755779504776, 'train/grad_norm_pre_clip': 43.42399597167969, 'timing/data': 0.0007426420052070171, 'timing/model': 1.1975760129862465, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.49}) +[2026-06-23 19:04:57,559][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.08534200489521027, 'train/grad_norm_pre_clip': 24.418092727661133, 'timing/data': 0.0007615270151291043, 'timing/model': 1.1949776739929803, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.49}) +[2026-06-23 19:05:01,895][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.05630157142877579, 'train/grad_norm_pre_clip': 27.32000160217285, 'timing/data': 0.0007579300145152956, 'timing/model': 1.1967473169788718, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.49}) +[2026-06-23 19:05:06,442][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.200672447681427, 'train/grad_norm_pre_clip': 40.138771057128906, 'timing/data': 0.0009117959998548031, 'timing/model': 1.1814130510028917, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.49}) +[2026-06-23 19:05:10,807][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.02635100856423378, 'train/grad_norm_pre_clip': 12.411436080932617, 'timing/data': 0.0007821949839126319, 'timing/model': 1.2055170920211822, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.49}) +[2026-06-23 19:05:15,169][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.03080560266971588, 'train/grad_norm_pre_clip': 27.56157684326172, 'timing/data': 0.0009268129942938685, 'timing/model': 1.1982682270172518, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.5}) +[2026-06-23 19:05:19,548][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.037198156118392944, 'train/grad_norm_pre_clip': 18.78083038330078, 'timing/data': 0.0007302899903152138, 'timing/model': 1.207380241015926, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.5}) +[2026-06-23 19:05:24,138][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.34041768312454224, 'train/grad_norm_pre_clip': 26.191022872924805, 'timing/data': 0.000778918998548761, 'timing/model': 1.2079035910137463, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.5}) +[2026-06-23 19:05:28,475][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.0471322126686573, 'train/grad_norm_pre_clip': 24.919923782348633, 'timing/data': 0.0007674680091440678, 'timing/model': 1.1804374039929826, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.5}) +[2026-06-23 19:05:32,824][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.1222100481390953, 'train/grad_norm_pre_clip': 20.832670211791992, 'timing/data': 0.0010468050022609532, 'timing/model': 1.1841632529976778, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.5}) +[2026-06-23 19:05:37,157][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.13774994015693665, 'train/grad_norm_pre_clip': 20.71674919128418, 'timing/data': 0.0007607059960719198, 'timing/model': 1.182681999023771, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.5}) +[2026-06-23 19:05:41,734][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.07305047661066055, 'train/grad_norm_pre_clip': 21.563232421875, 'timing/data': 0.0009219739877153188, 'timing/model': 1.210388817009516, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.5}) +[2026-06-23 19:05:46,118][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.11551360040903091, 'train/grad_norm_pre_clip': 22.356870651245117, 'timing/data': 0.0007514680037274957, 'timing/model': 1.2066414639994036, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.5}) +[2026-06-23 19:05:50,635][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.02182735502719879, 'train/grad_norm_pre_clip': 19.194780349731445, 'timing/data': 0.0007376830035354942, 'timing/model': 1.2030310849950183, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.5}) +[2026-06-23 19:05:54,989][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.10707277804613113, 'train/grad_norm_pre_clip': 18.938756942749023, 'timing/data': 0.0007453880098182708, 'timing/model': 1.200983112998074, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.5}) +[2026-06-23 19:05:59,571][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.06346648931503296, 'train/grad_norm_pre_clip': 27.503190994262695, 'timing/data': 0.0007395359862130135, 'timing/model': 1.1831209140073042, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.5}) +[2026-06-23 19:06:03,928][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.11989538371562958, 'train/grad_norm_pre_clip': 16.773527145385742, 'timing/data': 0.0008788040140643716, 'timing/model': 1.2035851679975167, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.5}) +[2026-06-23 19:06:08,319][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.18436124920845032, 'train/grad_norm_pre_clip': 17.889413833618164, 'timing/data': 0.0012773710186593235, 'timing/model': 1.2018596019770484, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.5}) +[2026-06-23 19:06:12,680][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.05332355946302414, 'train/grad_norm_pre_clip': 18.59658432006836, 'timing/data': 0.0007707040058448911, 'timing/model': 1.203660281986231, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.5}) +[2026-06-23 19:06:17,259][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.15517504513263702, 'train/grad_norm_pre_clip': 32.50115203857422, 'timing/data': 0.0007273439841810614, 'timing/model': 1.1994154510030057, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.5}) +[2026-06-23 19:06:21,640][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.11279032379388809, 'train/grad_norm_pre_clip': 29.42473793029785, 'timing/data': 0.0007703839801251888, 'timing/model': 1.202132945007179, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.5}) +[2026-06-23 19:06:26,025][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.09889697283506393, 'train/grad_norm_pre_clip': 16.586706161499023, 'timing/data': 0.0007207919843494892, 'timing/model': 1.2026050709828269, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.5}) +[2026-06-23 19:06:30,382][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.07948146015405655, 'train/grad_norm_pre_clip': 22.598695755004883, 'timing/data': 0.0007726679905317724, 'timing/model': 1.1849156509852037, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.5}) +[2026-06-23 19:06:34,986][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.06392336636781693, 'train/grad_norm_pre_clip': 19.27067756652832, 'timing/data': 0.0007812439871486276, 'timing/model': 1.1823633969761431, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.5}) +[2026-06-23 19:06:39,312][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.32262274622917175, 'train/grad_norm_pre_clip': 74.97059631347656, 'timing/data': 0.0007304399914573878, 'timing/model': 1.1796664260036778, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.5}) +[2026-06-23 19:06:43,665][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1300 +[2026-06-23 19:07:41,231][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1300 +[2026-06-23 19:07:41,232][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1300 +[2026-06-23 19:08:18,031][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1300 +[2026-06-23 19:08:18,032][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.004659728612750769, 'train/grad_norm_pre_clip': 40.51811981201172, 'eval/action_loss/seconds': 57.56284417700954, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.08848552405834198, 'eval/latency_2/loss': 0.08848552405834198, 'eval/flappy/loss': 0.08848552405834198, 'eval/flappy/latency_2/loss': 0.08848552405834198, 'eval/flap_precision': 0.9777777777777777, 'eval/flap_recall': 0.8955223880597015, 'eval/flap_f1': 0.9348441926345609, 'eval/flap_rate_model': 0.1875, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9348441926345608, 'eval/latency_2/flap_precision': 0.9777777777777777, 'eval/latency_2/flap_recall': 0.8955223880597015, 'eval/latency_2/flap_f1': 0.9348441926345609, 'eval/latency_2/flap_rate_model': 0.1875, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9348441926345608, 'eval/action_classification/data_wait_seconds': 0.4438568651676178, 'eval/action_classification/predict_seconds': 36.218482971191406, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0009132779960054904, 'timing/model': 1.2105456969875377, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.5}) +[2026-06-23 19:08:22,587][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.08266304433345795, 'train/grad_norm_pre_clip': 51.138065338134766, 'timing/data': 0.0010702990111894906, 'timing/model': 1.2003475549863651, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.5}) +[2026-06-23 19:08:26,885][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.0941307470202446, 'train/grad_norm_pre_clip': 46.73504638671875, 'timing/data': 0.0007271440117619932, 'timing/model': 1.1947927910077851, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.5}) +[2026-06-23 19:08:31,211][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.044499434530735016, 'train/grad_norm_pre_clip': 25.571001052856445, 'timing/data': 0.0009421109862159938, 'timing/model': 1.1805639250087552, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.5}) +[2026-06-23 19:08:35,593][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.3499632775783539, 'train/grad_norm_pre_clip': 75.44670104980469, 'timing/data': 0.0007942279917187989, 'timing/model': 1.1895666460040957, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.5}) +[2026-06-23 19:08:40,202][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.26701658964157104, 'train/grad_norm_pre_clip': 71.07550811767578, 'timing/data': 0.0008271979750134051, 'timing/model': 1.172672804998001, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.51}) +[2026-06-23 19:08:44,518][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.7931003570556641, 'train/grad_norm_pre_clip': 79.94441986083984, 'timing/data': 0.0011100410192739218, 'timing/model': 1.1872105360089336, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.51}) +[2026-06-23 19:08:48,809][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.7396341562271118, 'train/grad_norm_pre_clip': 79.19123840332031, 'timing/data': 0.0007966429984662682, 'timing/model': 1.1717898680071812, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.51}) +[2026-06-23 19:08:53,102][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.3608246445655823, 'train/grad_norm_pre_clip': 48.41319274902344, 'timing/data': 0.0007002339989412576, 'timing/model': 1.1722690649912693, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.51}) +[2026-06-23 19:08:57,642][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.0919114351272583, 'train/grad_norm_pre_clip': 15.552421569824219, 'timing/data': 0.000927745015360415, 'timing/model': 1.1803095169889275, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.51}) +[2026-06-23 19:09:01,966][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.1077377200126648, 'train/grad_norm_pre_clip': 12.230993270874023, 'timing/data': 0.0007842389750294387, 'timing/model': 1.178956930001732, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.51}) +[2026-06-23 19:09:06,317][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.16408281028270721, 'train/grad_norm_pre_clip': 23.30965232849121, 'timing/data': 0.001216467993799597, 'timing/model': 1.2116927850001957, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.51}) +[2026-06-23 19:09:10,648][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.15056148171424866, 'train/grad_norm_pre_clip': 41.18618392944336, 'timing/data': 0.0008024929848033935, 'timing/model': 1.1835630779969506, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.51}) +[2026-06-23 19:09:15,203][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.2866562604904175, 'train/grad_norm_pre_clip': 40.90158462524414, 'timing/data': 0.0010376779828220606, 'timing/model': 1.1827757839928381, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.51}) +[2026-06-23 19:09:19,574][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.22551849484443665, 'train/grad_norm_pre_clip': 35.177940368652344, 'timing/data': 0.0009534720156807452, 'timing/model': 1.1805238239758182, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.51}) +[2026-06-23 19:09:23,961][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.14612647891044617, 'train/grad_norm_pre_clip': 12.48241901397705, 'timing/data': 0.0010704690066631883, 'timing/model': 1.2031842069991399, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.51}) +[2026-06-23 19:09:28,328][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.03779461979866028, 'train/grad_norm_pre_clip': 9.252185821533203, 'timing/data': 0.0009391760104335845, 'timing/model': 1.181590417982079, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.51}) +[2026-06-23 19:09:32,873][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.08925239741802216, 'train/grad_norm_pre_clip': 21.493764877319336, 'timing/data': 0.0010357949940953404, 'timing/model': 1.1911409019958228, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.51}) +[2026-06-23 19:09:37,238][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.10260811448097229, 'train/grad_norm_pre_clip': 19.58650016784668, 'timing/data': 0.0009642929944675416, 'timing/model': 1.1909792149963323, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.51}) +[2026-06-23 19:09:41,613][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.03567944094538689, 'train/grad_norm_pre_clip': 14.983292579650879, 'timing/data': 0.0008218079747166485, 'timing/model': 1.2054164400033187, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.51}) +[2026-06-23 19:09:45,985][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.1683540642261505, 'train/grad_norm_pre_clip': 23.114208221435547, 'timing/data': 0.0014575030072592199, 'timing/model': 1.2017320020240732, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.51}) +[2026-06-23 19:09:50,563][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.037107761949300766, 'train/grad_norm_pre_clip': 14.723970413208008, 'timing/data': 0.0008931009797379375, 'timing/model': 1.1829214690078516, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.51}) +[2026-06-23 19:09:54,897][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.03736116737127304, 'train/grad_norm_pre_clip': 28.326435089111328, 'timing/data': 0.0008200749871321023, 'timing/model': 1.1779696680023335, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.51}) +[2026-06-23 19:09:59,219][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.09055068343877792, 'train/grad_norm_pre_clip': 32.93464660644531, 'timing/data': 0.0007245790038723499, 'timing/model': 1.1791777219914366, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.51}) +[2026-06-23 19:10:03,603][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.1295485496520996, 'train/grad_norm_pre_clip': 27.697797775268555, 'timing/data': 0.0006972379924263805, 'timing/model': 1.1972294579900336, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.51}) +[2026-06-23 19:10:08,135][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.053374357521533966, 'train/grad_norm_pre_clip': 19.724212646484375, 'timing/data': 0.0007533619937021285, 'timing/model': 1.1806426540133543, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.51}) +[2026-06-23 19:10:12,502][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.2789645195007324, 'train/grad_norm_pre_clip': 44.999534606933594, 'timing/data': 0.0009352389897685498, 'timing/model': 1.2004927530069835, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.51}) +[2026-06-23 19:10:16,867][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.009782319888472557, 'train/grad_norm_pre_clip': 22.768062591552734, 'timing/data': 0.0007360900053754449, 'timing/model': 1.200227885012282, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.51}) +[2026-06-23 19:10:21,234][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.030367227271199226, 'train/grad_norm_pre_clip': 24.166526794433594, 'timing/data': 0.0007485829992219806, 'timing/model': 1.1900469399988651, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.51}) +[2026-06-23 19:10:25,749][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.004079968202859163, 'train/grad_norm_pre_clip': 23.347625732421875, 'timing/data': 0.0007105539843905717, 'timing/model': 1.1772771010000724, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.51}) +[2026-06-23 19:10:30,074][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.05778785049915314, 'train/grad_norm_pre_clip': 87.71768188476562, 'timing/data': 0.0007234369986690581, 'timing/model': 1.1786890819785185, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.51}) +[2026-06-23 19:10:34,406][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.11314069479703903, 'train/grad_norm_pre_clip': 74.50367736816406, 'timing/data': 0.0007331249944400042, 'timing/model': 1.1800124699948356, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.52}) +[2026-06-23 19:10:38,733][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.4033878743648529, 'train/grad_norm_pre_clip': 86.5914535522461, 'timing/data': 0.0009070970118045807, 'timing/model': 1.1795141399779823, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.52}) +[2026-06-23 19:10:43,259][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.10266619920730591, 'train/grad_norm_pre_clip': 25.79851722717285, 'timing/data': 0.000704563019098714, 'timing/model': 1.1804777430079412, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.52}) +[2026-06-23 19:10:47,610][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.14493906497955322, 'train/grad_norm_pre_clip': 14.215967178344727, 'timing/data': 0.0007993780018296093, 'timing/model': 1.1809378850157373, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.52}) +[2026-06-23 19:10:51,995][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.16614671051502228, 'train/grad_norm_pre_clip': 17.99543571472168, 'timing/data': 0.0007071769796311855, 'timing/model': 1.19728641200345, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.52}) +[2026-06-23 19:10:56,333][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.06403439491987228, 'train/grad_norm_pre_clip': 24.1325626373291, 'timing/data': 0.0008312059799209237, 'timing/model': 1.1863065709767397, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.52}) +[2026-06-23 19:11:00,870][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.07347169518470764, 'train/grad_norm_pre_clip': 13.844681739807129, 'timing/data': 0.0007488240080419928, 'timing/model': 1.1845015159924515, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 0.52}) +[2026-06-23 19:11:05,277][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.021913405507802963, 'train/grad_norm_pre_clip': 17.210412979125977, 'timing/data': 0.0007974639884196222, 'timing/model': 1.2058358089998364, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 0.52}) +[2026-06-23 19:11:09,670][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.025248505175113678, 'train/grad_norm_pre_clip': 18.70183563232422, 'timing/data': 0.0007662659918423742, 'timing/model': 1.2125866569986101, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 0.52}) +[2026-06-23 19:11:14,064][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.07044433057308197, 'train/grad_norm_pre_clip': 18.233015060424805, 'timing/data': 0.0007701540016569197, 'timing/model': 1.1862769830040634, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.52}) +[2026-06-23 19:11:18,602][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.05612131953239441, 'train/grad_norm_pre_clip': 22.4017391204834, 'timing/data': 0.0006729240121785551, 'timing/model': 1.181892341002822, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 0.52}) +[2026-06-23 19:11:22,996][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.17126354575157166, 'train/grad_norm_pre_clip': 27.067594528198242, 'timing/data': 0.0007619079842697829, 'timing/model': 1.1900948319816962, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 0.52}) +[2026-06-23 19:11:27,410][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.1193670779466629, 'train/grad_norm_pre_clip': 11.616713523864746, 'timing/data': 0.000766435987316072, 'timing/model': 1.2016434079851024, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 0.52}) +[2026-06-23 19:11:31,764][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.03471016511321068, 'train/grad_norm_pre_clip': 8.867837905883789, 'timing/data': 0.0007550159934908152, 'timing/model': 1.1832734679919668, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 0.52}) +[2026-06-23 19:11:36,367][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.19214019179344177, 'train/grad_norm_pre_clip': 30.623640060424805, 'timing/data': 0.0007474010053556412, 'timing/model': 1.1835622430080548, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 0.52}) +[2026-06-23 19:11:40,765][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.005636266898363829, 'train/grad_norm_pre_clip': 16.986759185791016, 'timing/data': 0.0007898199837654829, 'timing/model': 1.2041730089986231, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 0.52}) +[2026-06-23 19:11:45,191][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.13491402566432953, 'train/grad_norm_pre_clip': 19.88217544555664, 'timing/data': 0.0008034250058699399, 'timing/model': 1.2260647020011675, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 0.52}) +[2026-06-23 19:11:49,570][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.08599507063627243, 'train/grad_norm_pre_clip': 12.482481956481934, 'timing/data': 0.0008236519934143871, 'timing/model': 1.1962491559970658, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 0.52}) +[2026-06-23 19:11:54,094][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.14992131292819977, 'train/grad_norm_pre_clip': 16.680734634399414, 'timing/data': 0.0007448560209013522, 'timing/model': 1.1835166739765555, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 0.52}) +[2026-06-23 19:11:58,490][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.15393675863742828, 'train/grad_norm_pre_clip': 16.41303062438965, 'timing/data': 0.0008088249887805432, 'timing/model': 1.1858096639916766, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.52}) +[2026-06-23 19:12:02,870][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.20532183349132538, 'train/grad_norm_pre_clip': 18.87181854248047, 'timing/data': 0.0007907309918664396, 'timing/model': 1.1852744859934319, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 0.52}) +[2026-06-23 19:12:07,223][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.07610154896974564, 'train/grad_norm_pre_clip': 28.83858299255371, 'timing/data': 0.001019583985907957, 'timing/model': 1.1878433059901, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 0.52}) +[2026-06-23 19:12:11,895][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.09007301926612854, 'train/grad_norm_pre_clip': 19.218061447143555, 'timing/data': 0.0008118200057651848, 'timing/model': 1.4532371850218624, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 0.52}) +[2026-06-23 19:12:16,278][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.01979829929769039, 'train/grad_norm_pre_clip': 24.056421279907227, 'timing/data': 0.0009825459856074303, 'timing/model': 1.21361942999647, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 0.52}) +[2026-06-23 19:12:20,714][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.059588249772787094, 'train/grad_norm_pre_clip': 20.55489730834961, 'timing/data': 0.000952890986809507, 'timing/model': 1.2089395720104221, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 0.52}) +[2026-06-23 19:12:25,107][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.05808711424469948, 'train/grad_norm_pre_clip': 25.045745849609375, 'timing/data': 0.0007433939899783581, 'timing/model': 1.2081315490067936, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 0.52}) +[2026-06-23 19:12:29,680][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.05595600605010986, 'train/grad_norm_pre_clip': 13.690367698669434, 'timing/data': 0.0011408580176066607, 'timing/model': 1.210467097000219, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 0.53}) +[2026-06-23 19:12:34,096][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.03938458859920502, 'train/grad_norm_pre_clip': 10.856637001037598, 'timing/data': 0.0009200299973599613, 'timing/model': 1.2153220019827131, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 0.53}) +[2026-06-23 19:12:38,476][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.11365532130002975, 'train/grad_norm_pre_clip': 21.273563385009766, 'timing/data': 0.0010061999782919884, 'timing/model': 1.1916263770253863, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 0.53}) +[2026-06-23 19:12:42,884][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.26854750514030457, 'train/grad_norm_pre_clip': 29.737693786621094, 'timing/data': 0.0009297290234826505, 'timing/model': 1.1877088980108965, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.53}) +[2026-06-23 19:12:47,568][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.19244465231895447, 'train/grad_norm_pre_clip': 38.93171310424805, 'timing/data': 0.0009228749840985984, 'timing/model': 1.487341241008835, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 0.53}) +[2026-06-23 19:12:52,014][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.07226632535457611, 'train/grad_norm_pre_clip': 28.03253173828125, 'timing/data': 0.0010355339909438044, 'timing/model': 1.2163139309850521, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 0.53}) +[2026-06-23 19:12:56,439][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.06361057609319687, 'train/grad_norm_pre_clip': 16.60388946533203, 'timing/data': 0.0010182020196225494, 'timing/model': 1.2033259269956034, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 0.53}) +[2026-06-23 19:13:00,804][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.16369324922561646, 'train/grad_norm_pre_clip': 21.223060607910156, 'timing/data': 0.0007795109995640814, 'timing/model': 1.2020359540183563, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 0.53}) +[2026-06-23 19:13:05,363][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.08043009787797928, 'train/grad_norm_pre_clip': 21.734546661376953, 'timing/data': 0.0009694819746073335, 'timing/model': 1.1888198390079197, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 0.53}) +[2026-06-23 19:13:09,751][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.18602155148983002, 'train/grad_norm_pre_clip': 54.796546936035156, 'timing/data': 0.0007438850007019937, 'timing/model': 1.1869274519849569, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 0.53}) +[2026-06-23 19:13:14,139][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.20680779218673706, 'train/grad_norm_pre_clip': 33.51618576049805, 'timing/data': 0.0008806179976090789, 'timing/model': 1.1942008860060014, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 0.53}) +[2026-06-23 19:13:18,510][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.19194531440734863, 'train/grad_norm_pre_clip': 22.920761108398438, 'timing/data': 0.0007784879999235272, 'timing/model': 1.1922203269787133, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 0.53}) +[2026-06-23 19:13:23,192][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.10463111847639084, 'train/grad_norm_pre_clip': 18.65387535095215, 'timing/data': 0.000933535979129374, 'timing/model': 1.4410746610083152, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 0.53}) +[2026-06-23 19:13:27,581][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.1468418389558792, 'train/grad_norm_pre_clip': 21.586376190185547, 'timing/data': 0.0012039850116707385, 'timing/model': 1.186647369991988, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.53}) +[2026-06-23 19:13:31,958][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.09359309077262878, 'train/grad_norm_pre_clip': 15.853798866271973, 'timing/data': 0.0011355189781170338, 'timing/model': 1.2058100489957724, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 0.53}) +[2026-06-23 19:13:36,348][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.08058833330869675, 'train/grad_norm_pre_clip': 19.681171417236328, 'timing/data': 0.0007770369993522763, 'timing/model': 1.1915898759907577, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 0.53}) +[2026-06-23 19:13:40,978][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.09889709204435349, 'train/grad_norm_pre_clip': 18.403059005737305, 'timing/data': 0.0009511079988442361, 'timing/model': 1.1899414109939244, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 0.53}) +[2026-06-23 19:13:45,374][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.005154502112418413, 'train/grad_norm_pre_clip': 11.453402519226074, 'timing/data': 0.0009089110244531184, 'timing/model': 1.1921743290149607, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.53}) +[2026-06-23 19:13:49,795][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.0462007075548172, 'train/grad_norm_pre_clip': 6.737058162689209, 'timing/data': 0.0009715960186440498, 'timing/model': 1.204364531004103, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 0.53}) +[2026-06-23 19:13:54,236][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.05231563746929169, 'train/grad_norm_pre_clip': 22.317913055419922, 'timing/data': 0.000751338986447081, 'timing/model': 1.2168318859767169, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 0.53}) +[2026-06-23 19:13:58,844][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.08059235662221909, 'train/grad_norm_pre_clip': 22.67688751220703, 'timing/data': 0.0008056089864112437, 'timing/model': 1.2146605550078675, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 0.53}) +[2026-06-23 19:14:03,242][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.07506973296403885, 'train/grad_norm_pre_clip': 12.303061485290527, 'timing/data': 0.0009987360099330544, 'timing/model': 1.2099232700129505, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 0.53}) +[2026-06-23 19:14:07,654][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.005473330616950989, 'train/grad_norm_pre_clip': 45.2340087890625, 'timing/data': 0.0007445469964295626, 'timing/model': 1.2037922690215055, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 0.53}) +[2026-06-23 19:14:12,024][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.10299205034971237, 'train/grad_norm_pre_clip': 34.21121597290039, 'timing/data': 0.0008516240050084889, 'timing/model': 1.1881292500183918, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.53}) +[2026-06-23 19:14:16,605][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.18570101261138916, 'train/grad_norm_pre_clip': 22.241079330444336, 'timing/data': 0.0007458189793396741, 'timing/model': 1.2120513560075779, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 0.53}) +[2026-06-23 19:14:21,012][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.035372402518987656, 'train/grad_norm_pre_clip': 37.23603820800781, 'timing/data': 0.0007594730122946203, 'timing/model': 1.19199501699768, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 0.53}) +[2026-06-23 19:14:25,465][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.08879884332418442, 'train/grad_norm_pre_clip': 23.583295822143555, 'timing/data': 0.001130518998252228, 'timing/model': 1.2247354520077351, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 0.54}) +[2026-06-23 19:14:30,109][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.058010634034872055, 'train/grad_norm_pre_clip': 28.82679557800293, 'timing/data': 0.0006915490084793419, 'timing/model': 1.4662067739991471, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 0.54}) +[2026-06-23 19:14:34,553][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.04010009765625, 'train/grad_norm_pre_clip': 26.364063262939453, 'timing/data': 0.0009411500068381429, 'timing/model': 1.2135369559982792, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 0.54}) +[2026-06-23 19:14:38,946][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.08489254862070084, 'train/grad_norm_pre_clip': 44.409603118896484, 'timing/data': 0.000749886006815359, 'timing/model': 1.1911322870000731, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 0.54}) +[2026-06-23 19:14:43,346][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.01400207169353962, 'train/grad_norm_pre_clip': 12.53349781036377, 'timing/data': 0.0008844350231811404, 'timing/model': 1.2063278639980126, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 0.54}) +[2026-06-23 19:14:47,967][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.09408783912658691, 'train/grad_norm_pre_clip': 42.4762077331543, 'timing/data': 0.0007820350001566112, 'timing/model': 1.4385952139855362, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 0.54}) +[2026-06-23 19:14:52,393][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.21913281083106995, 'train/grad_norm_pre_clip': 31.716209411621094, 'timing/data': 0.0007881170022301376, 'timing/model': 1.2104498849948868, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 0.54}) +[2026-06-23 19:14:56,871][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.12991511821746826, 'train/grad_norm_pre_clip': 37.52684783935547, 'timing/data': 0.001154022989794612, 'timing/model': 1.210115999012487, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.54}) +[2026-06-23 19:15:01,308][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.08889395743608475, 'train/grad_norm_pre_clip': 17.497413635253906, 'timing/data': 0.0009546439978294075, 'timing/model': 1.2251413720077835, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 0.54}) +[2026-06-23 19:15:05,875][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.018763801082968712, 'train/grad_norm_pre_clip': 7.6057939529418945, 'timing/data': 0.0007341470045503229, 'timing/model': 1.1878600850177463, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 0.54}) +[2026-06-23 19:15:10,263][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.10762183368206024, 'train/grad_norm_pre_clip': 39.14487075805664, 'timing/data': 0.0008141850121319294, 'timing/model': 1.2173055009916425, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 0.54}) +[2026-06-23 19:15:14,642][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.11055729538202286, 'train/grad_norm_pre_clip': 31.803613662719727, 'timing/data': 0.0009516180143691599, 'timing/model': 1.194192571012536, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 0.54}) +[2026-06-23 19:15:19,046][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.05721241235733032, 'train/grad_norm_pre_clip': 11.364580154418945, 'timing/data': 0.0011604639876168221, 'timing/model': 1.188865166011965, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.54}) +[2026-06-23 19:15:23,667][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.00719560869038105, 'train/grad_norm_pre_clip': 36.52149963378906, 'timing/data': 0.0007332259847316891, 'timing/model': 1.1908142589963973, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 0.54}) +[2026-06-23 19:15:28,046][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.12109753489494324, 'train/grad_norm_pre_clip': 32.96122360229492, 'timing/data': 0.000776414992287755, 'timing/model': 1.1857023319753353, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 0.54}) +[2026-06-23 19:15:32,429][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.1102805957198143, 'train/grad_norm_pre_clip': 20.580318450927734, 'timing/data': 0.000774101004935801, 'timing/model': 1.2041489779949188, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 0.54}) +[2026-06-23 19:15:36,810][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.0765833929181099, 'train/grad_norm_pre_clip': 22.8280029296875, 'timing/data': 0.0007397070003207773, 'timing/model': 1.1917805479897652, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 0.54}) +[2026-06-23 19:15:41,445][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1400 +[2026-06-23 19:16:39,570][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1400 +[2026-06-23 19:16:39,571][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1400 +[2026-06-23 19:17:16,582][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1400 +[2026-06-23 19:17:16,584][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.1714547574520111, 'train/grad_norm_pre_clip': 23.570175170898438, 'eval/action_loss/seconds': 58.12081439798931, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.052332162857055664, 'eval/latency_2/loss': 0.052332162857055664, 'eval/flappy/loss': 0.052332162857055664, 'eval/flappy/latency_2/loss': 0.052332162857055664, 'eval/flap_precision': 0.938562091503268, 'eval/flap_recall': 0.9742198100407056, 'eval/flap_f1': 0.9560585885486019, 'eval/flap_rate_model': 0.2125, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9560585885486019, 'eval/latency_2/flap_precision': 0.938562091503268, 'eval/latency_2/flap_recall': 0.9742198100407056, 'eval/latency_2/flap_f1': 0.9560585885486019, 'eval/latency_2/flap_rate_model': 0.2125, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9560585885486019, 'eval/action_classification/data_wait_seconds': 0.5024208426475525, 'eval/action_classification/predict_seconds': 36.3596076965332, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0008100470004137605, 'timing/model': 1.1938223939796444, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.54}) +[2026-06-23 19:17:20,905][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.049800727516412735, 'train/grad_norm_pre_clip': 24.655363082885742, 'timing/data': 0.0007935070025268942, 'timing/model': 1.168715676991269, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 0.54}) +[2026-06-23 19:17:25,186][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.015662407502532005, 'train/grad_norm_pre_clip': 10.289690017700195, 'timing/data': 0.0007688610057812184, 'timing/model': 1.1690191289817449, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 0.54}) +[2026-06-23 19:17:29,686][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.14739777147769928, 'train/grad_norm_pre_clip': 23.436906814575195, 'timing/data': 0.0008146560285240412, 'timing/model': 1.19019407700398, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 0.54}) +[2026-06-23 19:17:34,014][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.26296648383140564, 'train/grad_norm_pre_clip': 22.692214965820312, 'timing/data': 0.0007458790205419064, 'timing/model': 1.1901238590071443, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 0.54}) +[2026-06-23 19:17:38,384][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.06083811819553375, 'train/grad_norm_pre_clip': 22.193222045898438, 'timing/data': 0.0011282660125289112, 'timing/model': 1.2004549439880066, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 0.54}) +[2026-06-23 19:17:42,708][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.043993912637233734, 'train/grad_norm_pre_clip': 22.117063522338867, 'timing/data': 0.0007500660140067339, 'timing/model': 1.1719822479935829, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 0.54}) +[2026-06-23 19:17:47,320][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.25679031014442444, 'train/grad_norm_pre_clip': 20.58720588684082, 'timing/data': 0.0007268540211953223, 'timing/model': 1.1956641209835652, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 0.54}) +[2026-06-23 19:17:51,658][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.10252903401851654, 'train/grad_norm_pre_clip': 31.234529495239258, 'timing/data': 0.0007298190030269325, 'timing/model': 1.1761174830025993, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 0.54}) +[2026-06-23 19:17:56,012][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.10121000558137894, 'train/grad_norm_pre_clip': 22.902854919433594, 'timing/data': 0.0007288970227818936, 'timing/model': 1.1956317729782313, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 0.55}) +[2026-06-23 19:18:00,342][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.1747732162475586, 'train/grad_norm_pre_clip': 23.74527931213379, 'timing/data': 0.0007552359893452376, 'timing/model': 1.1812134129868355, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.55}) +[2026-06-23 19:18:04,893][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.06246281787753105, 'train/grad_norm_pre_clip': 18.351945877075195, 'timing/data': 0.0007818450103513896, 'timing/model': 1.181985742005054, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 0.55}) +[2026-06-23 19:18:09,219][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.30509617924690247, 'train/grad_norm_pre_clip': 37.630027770996094, 'timing/data': 0.0007270940113812685, 'timing/model': 1.1824416759773158, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 0.55}) +[2026-06-23 19:18:13,563][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.09624619036912918, 'train/grad_norm_pre_clip': 10.426316261291504, 'timing/data': 0.0007731500081717968, 'timing/model': 1.1827740900043864, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 0.55}) +[2026-06-23 19:18:18,067][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.05890241265296936, 'train/grad_norm_pre_clip': 15.025086402893066, 'timing/data': 0.001006400998448953, 'timing/model': 1.355142746004276, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 0.55}) +[2026-06-23 19:18:22,404][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.08614026755094528, 'train/grad_norm_pre_clip': 15.174083709716797, 'timing/data': 0.001042607007548213, 'timing/model': 1.181234988005599, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 0.55}) +[2026-06-23 19:18:26,784][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.2561667859554291, 'train/grad_norm_pre_clip': 36.532772064208984, 'timing/data': 0.0011172649974469095, 'timing/model': 1.202410466998117, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 0.55}) +[2026-06-23 19:18:31,186][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.2601921558380127, 'train/grad_norm_pre_clip': 38.73386764526367, 'timing/data': 0.0007371929823420942, 'timing/model': 1.2018053590145428, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 0.55}) +[2026-06-23 19:18:35,564][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.12421054393053055, 'train/grad_norm_pre_clip': 18.358646392822266, 'timing/data': 0.0007066060206852853, 'timing/model': 1.2032762590097263, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 0.55}) +[2026-06-23 19:18:40,135][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.07876215130090714, 'train/grad_norm_pre_clip': 17.283620834350586, 'timing/data': 0.0009074569970835, 'timing/model': 1.183736412000144, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 0.55}) +[2026-06-23 19:18:44,465][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.09011077135801315, 'train/grad_norm_pre_clip': 19.730871200561523, 'timing/data': 0.0007095620094332844, 'timing/model': 1.1810205740039237, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.55}) +[2026-06-23 19:18:48,845][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.06093570217490196, 'train/grad_norm_pre_clip': 12.640497207641602, 'timing/data': 0.0007350690138991922, 'timing/model': 1.20060338001349, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 0.55}) +[2026-06-23 19:18:53,172][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.059637438505887985, 'train/grad_norm_pre_clip': 23.449464797973633, 'timing/data': 0.0008393320022150874, 'timing/model': 1.1801453180087265, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 0.55}) +[2026-06-23 19:18:57,778][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.12307225912809372, 'train/grad_norm_pre_clip': 16.779430389404297, 'timing/data': 0.0008649080118630081, 'timing/model': 1.1803490640013479, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.55}) +[2026-06-23 19:19:02,109][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.0799216702580452, 'train/grad_norm_pre_clip': 23.136737823486328, 'timing/data': 0.0009774169884622097, 'timing/model': 1.1822832809993997, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 0.55}) +[2026-06-23 19:19:06,500][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.10597351938486099, 'train/grad_norm_pre_clip': 29.51099395751953, 'timing/data': 0.0006886929913889617, 'timing/model': 1.21423252398381, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 0.55}) +[2026-06-23 19:19:10,836][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.12935549020767212, 'train/grad_norm_pre_clip': 29.16423225402832, 'timing/data': 0.0007245290034916252, 'timing/model': 1.18238346700673, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 0.55}) +[2026-06-23 19:19:15,349][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.08941052109003067, 'train/grad_norm_pre_clip': 32.996761322021484, 'timing/data': 0.0008754489826969802, 'timing/model': 1.1848149609868415, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 0.55}) +[2026-06-23 19:19:19,691][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.08602584153413773, 'train/grad_norm_pre_clip': 24.370981216430664, 'timing/data': 0.0007291380024980754, 'timing/model': 1.1847705400141422, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 0.55}) +[2026-06-23 19:19:24,040][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.023468194529414177, 'train/grad_norm_pre_clip': 14.826786041259766, 'timing/data': 0.0007274840027093887, 'timing/model': 1.187148505996447, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 0.55}) +[2026-06-23 19:19:28,447][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.030715826898813248, 'train/grad_norm_pre_clip': 16.10293197631836, 'timing/data': 0.0008003799885045737, 'timing/model': 1.2044670470058918, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.55}) +[2026-06-23 19:19:32,989][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.018869047984480858, 'train/grad_norm_pre_clip': 21.814151763916016, 'timing/data': 0.0008926699811127037, 'timing/model': 1.2083602839848027, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 0.55}) +[2026-06-23 19:19:37,357][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.05914739891886711, 'train/grad_norm_pre_clip': 22.863712310791016, 'timing/data': 0.0007314720132853836, 'timing/model': 1.1838680839864537, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 0.55}) +[2026-06-23 19:19:41,723][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.19897368550300598, 'train/grad_norm_pre_clip': 41.50080871582031, 'timing/data': 0.0008571950020268559, 'timing/model': 1.184701946011046, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 0.55}) +[2026-06-23 19:19:46,141][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.21176761388778687, 'train/grad_norm_pre_clip': 31.7602481842041, 'timing/data': 0.0007473309815395623, 'timing/model': 1.2378026449878234, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 0.55}) +[2026-06-23 19:19:50,725][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.05389595404267311, 'train/grad_norm_pre_clip': 10.290306091308594, 'timing/data': 0.0008094759832601994, 'timing/model': 1.1845468999817967, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 0.56}) +[2026-06-23 19:19:55,118][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.12969304621219635, 'train/grad_norm_pre_clip': 21.399295806884766, 'timing/data': 0.0008542489958927035, 'timing/model': 1.2014612650091294, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 0.56}) +[2026-06-23 19:19:59,457][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.1774076521396637, 'train/grad_norm_pre_clip': 51.32978439331055, 'timing/data': 0.0012630140117835253, 'timing/model': 1.183571920992108, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 0.56}) +[2026-06-23 19:20:03,860][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.14613044261932373, 'train/grad_norm_pre_clip': 54.42533493041992, 'timing/data': 0.0007138199871405959, 'timing/model': 1.2177188459900208, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 0.56}) +[2026-06-23 19:20:08,469][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.12567856907844543, 'train/grad_norm_pre_clip': 19.05182647705078, 'timing/data': 0.0007999990193638951, 'timing/model': 1.1836512880108785, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 0.56}) +[2026-06-23 19:20:12,850][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.13406504690647125, 'train/grad_norm_pre_clip': 25.126569747924805, 'timing/data': 0.000719920004485175, 'timing/model': 1.2039455990015995, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.56}) +[2026-06-23 19:20:17,238][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.09734419733285904, 'train/grad_norm_pre_clip': 19.660865783691406, 'timing/data': 0.0007278559787664562, 'timing/model': 1.2046801270043943, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 0.56}) +[2026-06-23 19:20:21,579][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.04694528877735138, 'train/grad_norm_pre_clip': 18.436079025268555, 'timing/data': 0.0008137139957398176, 'timing/model': 1.1837601410225034, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 0.56}) +[2026-06-23 19:20:26,234][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.18280938267707825, 'train/grad_norm_pre_clip': 38.822052001953125, 'timing/data': 0.0009332449990324676, 'timing/model': 1.2029144449916203, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 0.56}) +[2026-06-23 19:20:30,645][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.1813553273677826, 'train/grad_norm_pre_clip': 54.683956146240234, 'timing/data': 0.0007208219903986901, 'timing/model': 1.2217260260076728, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 0.56}) +[2026-06-23 19:20:35,079][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.08410846441984177, 'train/grad_norm_pre_clip': 16.52042007446289, 'timing/data': 0.0009860220015980303, 'timing/model': 1.2457771509943996, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 0.56}) +[2026-06-23 19:20:39,438][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.15397146344184875, 'train/grad_norm_pre_clip': 22.17242431640625, 'timing/data': 0.0007541130180470645, 'timing/model': 1.1834402209788095, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 0.56}) +[2026-06-23 19:20:44,086][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.07907089591026306, 'train/grad_norm_pre_clip': 21.997135162353516, 'timing/data': 0.0010976180201396346, 'timing/model': 1.2219417459855322, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 0.56}) +[2026-06-23 19:20:48,518][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.11921916902065277, 'train/grad_norm_pre_clip': 22.17070960998535, 'timing/data': 0.0008893939957488328, 'timing/model': 1.1988382629933767, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 0.56}) +[2026-06-23 19:20:52,936][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.14957265555858612, 'train/grad_norm_pre_clip': 34.35021209716797, 'timing/data': 0.0008984809974208474, 'timing/model': 1.2103948649892118, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 0.56}) +[2026-06-23 19:20:57,317][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.14560377597808838, 'train/grad_norm_pre_clip': 18.312705993652344, 'timing/data': 0.0009012770024128258, 'timing/model': 1.2074814500228968, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.56}) +[2026-06-23 19:21:01,993][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.08051933348178864, 'train/grad_norm_pre_clip': 31.030012130737305, 'timing/data': 0.0007870050030760467, 'timing/model': 1.2170320339791942, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 0.56}) +[2026-06-23 19:21:06,491][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.08218914270401001, 'train/grad_norm_pre_clip': 19.457183837890625, 'timing/data': 0.0008618629944976419, 'timing/model': 1.2891563750163186, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 0.56}) +[2026-06-23 19:21:10,883][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.10840678960084915, 'train/grad_norm_pre_clip': 10.391241073608398, 'timing/data': 0.0007575810013804585, 'timing/model': 1.2064218930026982, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 0.56}) +[2026-06-23 19:21:15,292][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.09857837110757828, 'train/grad_norm_pre_clip': 11.482094764709473, 'timing/data': 0.0007431730045937002, 'timing/model': 1.217933350999374, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 0.56}) +[2026-06-23 19:21:19,947][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.07406304776668549, 'train/grad_norm_pre_clip': 15.768796920776367, 'timing/data': 0.0007275850221049041, 'timing/model': 1.2102371259825304, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 0.56}) +[2026-06-23 19:21:24,362][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.11722491681575775, 'train/grad_norm_pre_clip': 12.063989639282227, 'timing/data': 0.0007237780082505196, 'timing/model': 1.2108986100065522, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 0.56}) +[2026-06-23 19:21:28,759][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.10509199649095535, 'train/grad_norm_pre_clip': 13.067909240722656, 'timing/data': 0.0007432040001731366, 'timing/model': 1.2024828549765516, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 0.56}) +[2026-06-23 19:21:33,172][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.11504732817411423, 'train/grad_norm_pre_clip': 17.557607650756836, 'timing/data': 0.0009111939871218055, 'timing/model': 1.2037423230067361, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 0.56}) +[2026-06-23 19:21:37,678][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.026653151959180832, 'train/grad_norm_pre_clip': 19.383459091186523, 'timing/data': 0.000787024007877335, 'timing/model': 1.1855014880129602, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 0.56}) +[2026-06-23 19:21:42,095][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.06301011145114899, 'train/grad_norm_pre_clip': 18.530988693237305, 'timing/data': 0.0012208970147185028, 'timing/model': 1.2056724709982518, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.57}) +[2026-06-23 19:21:46,486][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.061056673526763916, 'train/grad_norm_pre_clip': 14.169707298278809, 'timing/data': 0.0007486339891329408, 'timing/model': 1.2056248640001286, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 0.57}) +[2026-06-23 19:21:50,873][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.12644343078136444, 'train/grad_norm_pre_clip': 19.114675521850586, 'timing/data': 0.0007748130010440946, 'timing/model': 1.1886924170248676, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 0.57}) +[2026-06-23 19:21:55,527][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.08762018382549286, 'train/grad_norm_pre_clip': 21.087387084960938, 'timing/data': 0.001039841998135671, 'timing/model': 1.1927828080079053, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 0.57}) +[2026-06-23 19:21:59,941][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.03406612202525139, 'train/grad_norm_pre_clip': 31.26845932006836, 'timing/data': 0.0012660610082093626, 'timing/model': 1.2156952689983882, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 0.57}) +[2026-06-23 19:22:04,345][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.1045871302485466, 'train/grad_norm_pre_clip': 36.41619873046875, 'timing/data': 0.0007490139978472143, 'timing/model': 1.217720292013837, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.57}) +[2026-06-23 19:22:08,744][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.14215543866157532, 'train/grad_norm_pre_clip': 28.936901092529297, 'timing/data': 0.0007219940016511828, 'timing/model': 1.1841283440007828, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 0.57}) +[2026-06-23 19:22:13,381][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.11702966690063477, 'train/grad_norm_pre_clip': 43.233402252197266, 'timing/data': 0.001223861996550113, 'timing/model': 1.1906238100200426, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 0.57}) +[2026-06-23 19:22:17,742][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.06432101130485535, 'train/grad_norm_pre_clip': 115.73249816894531, 'timing/data': 0.0007786090136505663, 'timing/model': 1.19048125698464, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 0.57}) +[2026-06-23 19:22:22,151][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.028524747118353844, 'train/grad_norm_pre_clip': 42.22857666015625, 'timing/data': 0.0007728879863861948, 'timing/model': 1.1899393550120294, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 0.57}) +[2026-06-23 19:22:26,552][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.17600256204605103, 'train/grad_norm_pre_clip': 34.69729995727539, 'timing/data': 0.000947580992942676, 'timing/model': 1.2086056370171718, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.57}) +[2026-06-23 19:22:31,189][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.09114022552967072, 'train/grad_norm_pre_clip': 38.1156120300293, 'timing/data': 0.00108090799767524, 'timing/model': 1.2056022369943094, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 0.57}) +[2026-06-23 19:22:35,564][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.04185415431857109, 'train/grad_norm_pre_clip': 59.57418441772461, 'timing/data': 0.0007605059945490211, 'timing/model': 1.1899876359966584, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.57}) +[2026-06-23 19:22:39,973][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.11700040847063065, 'train/grad_norm_pre_clip': 42.04003143310547, 'timing/data': 0.0011970530031248927, 'timing/model': 1.2037911590014119, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 0.57}) +[2026-06-23 19:22:44,348][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.14990736544132233, 'train/grad_norm_pre_clip': 109.93294525146484, 'timing/data': 0.0008342619985342026, 'timing/model': 1.190383948996896, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 0.57}) +[2026-06-23 19:22:48,994][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.11057725548744202, 'train/grad_norm_pre_clip': 32.36228942871094, 'timing/data': 0.0009387549944221973, 'timing/model': 1.1875395920069423, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 0.57}) +[2026-06-23 19:22:53,422][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.08594318479299545, 'train/grad_norm_pre_clip': 13.084138870239258, 'timing/data': 0.0009638119954615831, 'timing/model': 1.2141079699795227, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 0.57}) +[2026-06-23 19:22:57,809][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.0518437959253788, 'train/grad_norm_pre_clip': 21.315654754638672, 'timing/data': 0.0007752829988021404, 'timing/model': 1.205725086008897, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 0.57}) +[2026-06-23 19:23:02,188][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.039648767560720444, 'train/grad_norm_pre_clip': 17.767864227294922, 'timing/data': 0.0008965769957285374, 'timing/model': 1.1886364599922672, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 0.57}) +[2026-06-23 19:23:06,800][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.007194643374532461, 'train/grad_norm_pre_clip': 17.844404220581055, 'timing/data': 0.0007520910003222525, 'timing/model': 1.1871875129872933, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 0.57}) +[2026-06-23 19:23:11,190][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.02071041613817215, 'train/grad_norm_pre_clip': 16.955486297607422, 'timing/data': 0.0010588869918137789, 'timing/model': 1.1951074720127508, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.57}) +[2026-06-23 19:23:15,596][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.036377888172864914, 'train/grad_norm_pre_clip': 10.497709274291992, 'timing/data': 0.0011135480017401278, 'timing/model': 1.1908337830100209, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 0.57}) +[2026-06-23 19:23:20,001][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.26719823479652405, 'train/grad_norm_pre_clip': 17.546939849853516, 'timing/data': 0.0009130469989031553, 'timing/model': 1.2134359110204969, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 0.57}) +[2026-06-23 19:23:24,662][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.077177993953228, 'train/grad_norm_pre_clip': 18.147157669067383, 'timing/data': 0.0009398980182595551, 'timing/model': 1.1886915240029339, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 0.57}) +[2026-06-23 19:23:29,063][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.03327033668756485, 'train/grad_norm_pre_clip': 10.820189476013184, 'timing/data': 0.0007614869973622262, 'timing/model': 1.2063733380055055, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 0.57}) +[2026-06-23 19:23:33,449][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.08665525168180466, 'train/grad_norm_pre_clip': 25.187458038330078, 'timing/data': 0.0007455580052919686, 'timing/model': 1.2146654350217432, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 0.57}) +[2026-06-23 19:23:37,834][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.1167282685637474, 'train/grad_norm_pre_clip': 13.459537506103516, 'timing/data': 0.0007431440171785653, 'timing/model': 1.1915102060011122, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 0.58}) +[2026-06-23 19:23:42,497][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.05880798026919365, 'train/grad_norm_pre_clip': 17.234527587890625, 'timing/data': 0.0007480019994545728, 'timing/model': 1.1941121050040238, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 0.58}) +[2026-06-23 19:23:46,883][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.048047829419374466, 'train/grad_norm_pre_clip': 24.246963500976562, 'timing/data': 0.0007954300090204924, 'timing/model': 1.1891534099995624, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 0.58}) +[2026-06-23 19:23:51,314][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.10233449935913086, 'train/grad_norm_pre_clip': 15.583535194396973, 'timing/data': 0.0007324939942918718, 'timing/model': 1.2141130750242155, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 0.58}) +[2026-06-23 19:23:55,728][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.058222707360982895, 'train/grad_norm_pre_clip': 15.064291954040527, 'timing/data': 0.0007388159865513444, 'timing/model': 1.2095358849910554, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.58}) +[2026-06-23 19:24:00,382][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.0418756864964962, 'train/grad_norm_pre_clip': 11.679027557373047, 'timing/data': 0.0007224360015243292, 'timing/model': 1.1915690469904803, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 0.58}) +[2026-06-23 19:24:04,761][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.013387728482484818, 'train/grad_norm_pre_clip': 10.88145637512207, 'timing/data': 0.0007399980095215142, 'timing/model': 1.19120695698075, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 0.58}) +[2026-06-23 19:24:09,155][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.11755863577127457, 'train/grad_norm_pre_clip': 21.833162307739258, 'timing/data': 0.0007675480155739933, 'timing/model': 1.1899162519839592, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 0.58}) +[2026-06-23 19:24:13,556][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.04149062559008598, 'train/grad_norm_pre_clip': 22.91942024230957, 'timing/data': 0.0007742309826426208, 'timing/model': 1.1911843970010523, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 0.58}) +[2026-06-23 19:24:18,201][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.007152872160077095, 'train/grad_norm_pre_clip': 12.237607955932617, 'timing/data': 0.0007531919982284307, 'timing/model': 1.1880346249963623, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 0.58}) +[2026-06-23 19:24:22,601][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.04566863924264908, 'train/grad_norm_pre_clip': 28.369434356689453, 'timing/data': 0.0008955759985838085, 'timing/model': 1.1876301679876633, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 0.58}) +[2026-06-23 19:24:27,014][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.0921451672911644, 'train/grad_norm_pre_clip': 43.944435119628906, 'timing/data': 0.0007580610108561814, 'timing/model': 1.203979337005876, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 0.58}) +[2026-06-23 19:24:31,444][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.05237951874732971, 'train/grad_norm_pre_clip': 34.46061706542969, 'timing/data': 0.001065839984221384, 'timing/model': 1.2077500979939941, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 0.58}) +[2026-06-23 19:24:36,091][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.146906316280365, 'train/grad_norm_pre_clip': 57.220027923583984, 'timing/data': 0.000755226006731391, 'timing/model': 1.208051615016302, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 0.58}) +[2026-06-23 19:24:40,508][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1500 +[2026-06-23 19:25:37,978][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1500 +[2026-06-23 19:25:37,979][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1500 +[2026-06-23 19:26:15,044][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1500 +[2026-06-23 19:52:21,428][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-23 19:52:21,430][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.10882538557052612, 'train/grad_norm_pre_clip': 41.80691146850586, 'eval/action_loss/seconds': 57.466689614986535, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.053324781358242035, 'eval/latency_2/loss': 0.053324781358242035, 'eval/flappy/loss': 0.053324781358242035, 'eval/flappy/latency_2/loss': 0.053324781358242035, 'eval/flap_precision': 0.9577080491132333, 'eval/flap_recall': 0.9525101763907734, 'eval/flap_f1': 0.9551020408163265, 'eval/flap_rate_model': 0.2036111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9551020408163264, 'eval/latency_2/flap_precision': 0.9577080491132333, 'eval/latency_2/flap_recall': 0.9525101763907734, 'eval/latency_2/flap_f1': 0.9551020408163265, 'eval/latency_2/flap_rate_model': 0.2036111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9551020408163264, 'eval/action_classification/data_wait_seconds': 0.4966152310371399, 'eval/action_classification/predict_seconds': 36.4241943359375, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 407.7600000000191, 'rl_games_eval/mid_train/mean_length': 3304.0, 'rl_games_eval/mid_train/std_reward': 73.68000000001311, 'rl_games_eval/mid_train/std_length': 592.0, 'rl_games_eval/mid_train/macro_mean_reward': 407.7600000000191, 'rl_games_eval/mid_train/macro_mean_length': 3304.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 407.7600000000191, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3304.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 73.68000000001311, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 592.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 407.7600000000191, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0007396970177069306, 'timing/model': 1.208833810989745, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.58}) +[2026-06-23 19:52:21,431][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_1500_state +[2026-06-23 19:52:21,439][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 19:52:30,276][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_1500_state/model.safetensors +[2026-06-23 19:52:40,580][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_1500_state/optimizer.bin +[2026-06-23 19:52:40,580][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_1500_state/sampler.bin +[2026-06-23 19:52:40,581][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-23 19:52:40,582][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-23 19:52:40,781][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-23 19:52:40,781][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 19:52:40,806][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 19:52:45,121][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.1507752239704132, 'train/grad_norm_pre_clip': 43.907100677490234, 'timing/data': 0.0007993479957804084, 'timing/model': 1.1445898700039834, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 0.58}) +[2026-06-23 19:52:49,570][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.006194925867021084, 'train/grad_norm_pre_clip': 18.02553367614746, 'timing/data': 0.0007866340165492147, 'timing/model': 1.1579354380082805, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 0.58}) +[2026-06-23 19:52:53,783][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.06821473687887192, 'train/grad_norm_pre_clip': 40.196746826171875, 'timing/data': 0.0008098979888018221, 'timing/model': 1.1480781869904604, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 0.58}) +[2026-06-23 19:52:58,037][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.13277536630630493, 'train/grad_norm_pre_clip': 14.853737831115723, 'timing/data': 0.000867543974891305, 'timing/model': 1.1622105500136968, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 0.58}) +[2026-06-23 19:53:02,267][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.03398566320538521, 'train/grad_norm_pre_clip': 6.686885833740234, 'timing/data': 0.0010493089794181287, 'timing/model': 1.1611762770044152, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 0.58}) +[2026-06-23 19:53:06,630][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.06998326629400253, 'train/grad_norm_pre_clip': 26.17635726928711, 'timing/data': 0.000714430003426969, 'timing/model': 1.158554321009433, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 0.58}) +[2026-06-23 19:53:10,843][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.17771394550800323, 'train/grad_norm_pre_clip': 11.350652694702148, 'timing/data': 0.0010163679835386574, 'timing/model': 1.155183961003786, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 0.58}) +[2026-06-23 19:53:15,060][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.24201111495494843, 'train/grad_norm_pre_clip': 25.1552791595459, 'timing/data': 0.001035834982758388, 'timing/model': 1.1677143929991871, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 0.58}) +[2026-06-23 19:53:19,319][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.0027080823201686144, 'train/grad_norm_pre_clip': 24.144453048706055, 'timing/data': 0.001171996002085507, 'timing/model': 1.1785511740017682, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 0.58}) +[2026-06-23 19:53:24,065][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.20198775827884674, 'train/grad_norm_pre_clip': 46.90700912475586, 'timing/data': 0.0007749420183245093, 'timing/model': 1.1705607530020643, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.58}) +[2026-06-23 19:53:28,381][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.08167200535535812, 'train/grad_norm_pre_clip': 26.756181716918945, 'timing/data': 0.0008266880176961422, 'timing/model': 1.1746804989816155, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 0.58}) +[2026-06-23 19:53:32,645][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.021422769874334335, 'train/grad_norm_pre_clip': 13.772632598876953, 'timing/data': 0.001197112986119464, 'timing/model': 1.1714814360020682, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 0.59}) +[2026-06-23 19:53:36,882][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.20328672230243683, 'train/grad_norm_pre_clip': 28.626197814941406, 'timing/data': 0.0008144349849317223, 'timing/model': 1.1800299170135986, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 0.59}) +[2026-06-23 19:53:41,324][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.21536146104335785, 'train/grad_norm_pre_clip': 24.648157119750977, 'timing/data': 0.0007379840244539082, 'timing/model': 1.170638167008292, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 0.59}) +[2026-06-23 19:53:45,580][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.057978928089141846, 'train/grad_norm_pre_clip': 10.605260848999023, 'timing/data': 0.0007611670007463545, 'timing/model': 1.15139081000234, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 0.59}) +[2026-06-23 19:53:49,857][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.039742983877658844, 'train/grad_norm_pre_clip': 16.445093154907227, 'timing/data': 0.0012285510019864887, 'timing/model': 1.1870817019953392, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 0.59}) +[2026-06-23 19:53:54,120][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.03714603930711746, 'train/grad_norm_pre_clip': 7.758908271789551, 'timing/data': 0.0008346219838131219, 'timing/model': 1.1756922990025487, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 0.59}) +[2026-06-23 19:53:58,563][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.15566134452819824, 'train/grad_norm_pre_clip': 17.542407989501953, 'timing/data': 0.0007530520088039339, 'timing/model': 1.1545826399815269, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 0.59}) +[2026-06-23 19:54:02,796][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.0877869501709938, 'train/grad_norm_pre_clip': 17.328227996826172, 'timing/data': 0.0007473810110241175, 'timing/model': 1.157583677006187, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 0.59}) +[2026-06-23 19:54:07,038][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.2417374700307846, 'train/grad_norm_pre_clip': 34.19319152832031, 'timing/data': 0.0007946690020617098, 'timing/model': 1.157775280997157, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.59}) +[2026-06-23 19:54:11,309][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.12485650926828384, 'train/grad_norm_pre_clip': 19.075557708740234, 'timing/data': 0.0008448509906884283, 'timing/model': 1.1844190999981947, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.59}) +[2026-06-23 19:54:15,793][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.055693965405225754, 'train/grad_norm_pre_clip': 13.69511890411377, 'timing/data': 0.0008155879913829267, 'timing/model': 1.164548121014377, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 0.59}) +[2026-06-23 19:54:20,094][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.03292923793196678, 'train/grad_norm_pre_clip': 30.88509178161621, 'timing/data': 0.0007932160224299878, 'timing/model': 1.1630055389832705, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 0.59}) +[2026-06-23 19:54:24,357][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.04129019379615784, 'train/grad_norm_pre_clip': 17.5892391204834, 'timing/data': 0.0008187829807866365, 'timing/model': 1.1640313559910282, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 0.59}) +[2026-06-23 19:54:28,669][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.13232289254665375, 'train/grad_norm_pre_clip': 29.785404205322266, 'timing/data': 0.0007047629915177822, 'timing/model': 1.1683609579922631, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 0.59}) +[2026-06-23 19:54:33,144][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.03623469918966293, 'train/grad_norm_pre_clip': 28.354711532592773, 'timing/data': 0.0007318229763768613, 'timing/model': 1.168228935013758, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 0.59}) +[2026-06-23 19:54:37,440][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.003205451648682356, 'train/grad_norm_pre_clip': 24.213848114013672, 'timing/data': 0.0008680039900355041, 'timing/model': 1.1866596119944006, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 0.59}) +[2026-06-23 19:54:41,719][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.15300032496452332, 'train/grad_norm_pre_clip': 32.77970886230469, 'timing/data': 0.0010581560200080276, 'timing/model': 1.1684819020156283, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 0.59}) +[2026-06-23 19:54:46,065][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.18024422228336334, 'train/grad_norm_pre_clip': 32.422149658203125, 'timing/data': 0.0007338870200328529, 'timing/model': 1.1953866019903217, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 0.59}) +[2026-06-23 19:54:50,624][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.07460471987724304, 'train/grad_norm_pre_clip': 43.82353973388672, 'timing/data': 0.0009255099867004901, 'timing/model': 1.1708418839843944, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.59}) +[2026-06-23 19:54:54,934][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.01189137902110815, 'train/grad_norm_pre_clip': 22.078346252441406, 'timing/data': 0.0012141140177845955, 'timing/model': 1.1757682290044613, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 0.59}) +[2026-06-23 19:54:59,239][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.04191206023097038, 'train/grad_norm_pre_clip': 25.9392147064209, 'timing/data': 0.0007428929966408759, 'timing/model': 1.174702429008903, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 0.59}) +[2026-06-23 19:55:03,552][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.12717851996421814, 'train/grad_norm_pre_clip': 24.357229232788086, 'timing/data': 0.0007989069854374975, 'timing/model': 1.173381848988356, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 0.59}) +[2026-06-23 19:55:08,079][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.11276934295892715, 'train/grad_norm_pre_clip': 16.138626098632812, 'timing/data': 0.0007548659923486412, 'timing/model': 1.1757382949872408, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 0.59}) +[2026-06-23 19:55:12,391][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.12851855158805847, 'train/grad_norm_pre_clip': 22.959400177001953, 'timing/data': 0.000956178002525121, 'timing/model': 1.1772246419859584, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.59}) +[2026-06-23 19:55:16,735][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.06749344617128372, 'train/grad_norm_pre_clip': 16.17228889465332, 'timing/data': 0.0008729239925742149, 'timing/model': 1.1787974199978635, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 0.59}) +[2026-06-23 19:55:21,059][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.07143793255090714, 'train/grad_norm_pre_clip': 37.14616775512695, 'timing/data': 0.0008475359936710447, 'timing/model': 1.1839064120140392, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 0.59}) +[2026-06-23 19:55:25,628][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.02991051971912384, 'train/grad_norm_pre_clip': 15.262998580932617, 'timing/data': 0.0007951800071168691, 'timing/model': 1.2032196309883147, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 0.6}) +[2026-06-23 19:55:29,986][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.13159240782260895, 'train/grad_norm_pre_clip': 21.54855728149414, 'timing/data': 0.0007755839906167239, 'timing/model': 1.2021649919915944, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 0.6}) +[2026-06-23 19:55:34,342][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.07561888545751572, 'train/grad_norm_pre_clip': 19.09056282043457, 'timing/data': 0.000747982005123049, 'timing/model': 1.1971461950161029, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.6}) +[2026-06-23 19:55:38,663][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.016950810328125954, 'train/grad_norm_pre_clip': 24.513999938964844, 'timing/data': 0.0008063710120040923, 'timing/model': 1.1851441100006923, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 0.6}) +[2026-06-23 19:55:43,196][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.10436411201953888, 'train/grad_norm_pre_clip': 20.20555305480957, 'timing/data': 0.0008027130097616464, 'timing/model': 1.173580755974399, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 0.6}) +[2026-06-23 19:55:47,541][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.13622519373893738, 'train/grad_norm_pre_clip': 12.044739723205566, 'timing/data': 0.0008748870168346912, 'timing/model': 1.1792319719970692, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 0.6}) +[2026-06-23 19:55:51,858][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.04672452062368393, 'train/grad_norm_pre_clip': 10.451322555541992, 'timing/data': 0.001129217998823151, 'timing/model': 1.1772037930204533, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 0.6}) +[2026-06-23 19:55:56,205][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.14006246626377106, 'train/grad_norm_pre_clip': 20.522178649902344, 'timing/data': 0.001302707998547703, 'timing/model': 1.1796902200148907, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 0.6}) +[2026-06-23 19:56:00,782][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.013781096786260605, 'train/grad_norm_pre_clip': 6.803738117218018, 'timing/data': 0.0007301699952222407, 'timing/model': 1.1928298070270102, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 0.6}) +[2026-06-23 19:56:05,077][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.17824450135231018, 'train/grad_norm_pre_clip': 15.29931926727295, 'timing/data': 0.0007377440051641315, 'timing/model': 1.1750550510187168, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 0.6}) +[2026-06-23 19:56:09,408][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.11629903316497803, 'train/grad_norm_pre_clip': 24.224773406982422, 'timing/data': 0.0009177569882012904, 'timing/model': 1.1970827280019876, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 0.6}) +[2026-06-23 19:56:13,751][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.011743756011128426, 'train/grad_norm_pre_clip': 5.980596542358398, 'timing/data': 0.0007842389750294387, 'timing/model': 1.1765622180246282, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 0.6}) +[2026-06-23 19:56:18,378][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.032846011221408844, 'train/grad_norm_pre_clip': 21.985382080078125, 'timing/data': 0.0008931710035540164, 'timing/model': 1.2192148769972846, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.6}) +[2026-06-23 19:56:22,733][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.05884980782866478, 'train/grad_norm_pre_clip': 12.046395301818848, 'timing/data': 0.0008633960096631199, 'timing/model': 1.1896899730199948, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 0.6}) +[2026-06-23 19:56:27,070][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.31513938307762146, 'train/grad_norm_pre_clip': 46.345542907714844, 'timing/data': 0.0009153620048891753, 'timing/model': 1.1772980179812293, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 0.6}) +[2026-06-23 19:56:31,444][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.03460647165775299, 'train/grad_norm_pre_clip': 15.293055534362793, 'timing/data': 0.0007257609977386892, 'timing/model': 1.1969002810074016, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 0.6}) +[2026-06-23 19:56:35,930][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.0558999627828598, 'train/grad_norm_pre_clip': 79.4286117553711, 'timing/data': 0.0008537379908375442, 'timing/model': 1.17809798597591, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 0.6}) +[2026-06-23 19:56:40,230][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.10583440959453583, 'train/grad_norm_pre_clip': 77.46297454833984, 'timing/data': 0.0008319379994645715, 'timing/model': 1.1739972970099188, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 0.6}) +[2026-06-23 19:56:44,548][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.0647142305970192, 'train/grad_norm_pre_clip': 23.798202514648438, 'timing/data': 0.0011284970096312463, 'timing/model': 1.1736797100165859, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 0.6}) +[2026-06-23 19:56:48,891][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.054742734879255295, 'train/grad_norm_pre_clip': 32.972023010253906, 'timing/data': 0.000676911004120484, 'timing/model': 1.193618063989561, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 0.6}) +[2026-06-23 19:56:53,372][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.08744780719280243, 'train/grad_norm_pre_clip': 56.649837493896484, 'timing/data': 0.000768489990150556, 'timing/model': 1.1766215569805354, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 0.6}) +[2026-06-23 19:56:57,704][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.5303600430488586, 'train/grad_norm_pre_clip': 38.758644104003906, 'timing/data': 0.0007787799986544997, 'timing/model': 1.1753098529879935, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 0.6}) +[2026-06-23 19:57:02,029][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.13219434022903442, 'train/grad_norm_pre_clip': 21.864704132080078, 'timing/data': 0.0007330459775403142, 'timing/model': 1.1764349720033351, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.6}) +[2026-06-23 19:57:06,359][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.35024720430374146, 'train/grad_norm_pre_clip': 25.555212020874023, 'timing/data': 0.0007078780035953969, 'timing/model': 1.176419914001599, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 0.6}) +[2026-06-23 19:57:10,951][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.10934588313102722, 'train/grad_norm_pre_clip': 25.024017333984375, 'timing/data': 0.001125851005781442, 'timing/model': 1.1929848429863341, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 0.6}) +[2026-06-23 19:57:15,316][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.12616916000843048, 'train/grad_norm_pre_clip': 20.028844833374023, 'timing/data': 0.0007466099923476577, 'timing/model': 1.1954465140006505, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 0.6}) +[2026-06-23 19:57:19,712][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.103800468146801, 'train/grad_norm_pre_clip': 33.48961639404297, 'timing/data': 0.0007321129960473627, 'timing/model': 1.19424968099338, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 0.61}) +[2026-06-23 19:57:24,026][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.2925487458705902, 'train/grad_norm_pre_clip': 39.53579330444336, 'timing/data': 0.0007251200149767101, 'timing/model': 1.178644637024263, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 0.61}) +[2026-06-23 19:57:28,612][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.19194012880325317, 'train/grad_norm_pre_clip': 23.290679931640625, 'timing/data': 0.0007506969850510359, 'timing/model': 1.1776857739896514, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 0.61}) +[2026-06-23 19:57:32,989][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.12876446545124054, 'train/grad_norm_pre_clip': 11.973245620727539, 'timing/data': 0.0006954749987926334, 'timing/model': 1.2099273409985472, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 0.61}) +[2026-06-23 19:57:37,354][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.017821861431002617, 'train/grad_norm_pre_clip': 20.94638442993164, 'timing/data': 0.0011485319992061704, 'timing/model': 1.1773425599967595, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 0.61}) +[2026-06-23 19:57:41,722][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.1646001935005188, 'train/grad_norm_pre_clip': 23.182592391967773, 'timing/data': 0.0007181580003816634, 'timing/model': 1.2083630579872988, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 0.61}) +[2026-06-23 19:57:46,302][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.14555290341377258, 'train/grad_norm_pre_clip': 11.43651008605957, 'timing/data': 0.0007290680077858269, 'timing/model': 1.179283938021399, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.61}) +[2026-06-23 19:57:50,618][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.13687098026275635, 'train/grad_norm_pre_clip': 12.13378620147705, 'timing/data': 0.0007731090008746833, 'timing/model': 1.1769120339886285, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 0.61}) +[2026-06-23 19:57:55,019][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.13675536215305328, 'train/grad_norm_pre_clip': 14.357986450195312, 'timing/data': 0.0007473510049749166, 'timing/model': 1.1973499610030558, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 0.61}) +[2026-06-23 19:57:59,355][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.06348512321710587, 'train/grad_norm_pre_clip': 17.54783058166504, 'timing/data': 0.00070691600558348, 'timing/model': 1.1773655830183998, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 0.61}) +[2026-06-23 19:58:03,961][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.04199731722474098, 'train/grad_norm_pre_clip': 12.065962791442871, 'timing/data': 0.0011445960262790322, 'timing/model': 1.177137541002594, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 0.61}) +[2026-06-23 19:58:08,298][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.06998632103204727, 'train/grad_norm_pre_clip': 26.086545944213867, 'timing/data': 0.0007261429855134338, 'timing/model': 1.1792037900013383, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 0.61}) +[2026-06-23 19:58:12,640][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.10091713815927505, 'train/grad_norm_pre_clip': 17.598674774169922, 'timing/data': 0.0007395760039798915, 'timing/model': 1.1810389820020646, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 0.61}) +[2026-06-23 19:58:17,003][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.01995532587170601, 'train/grad_norm_pre_clip': 7.454980850219727, 'timing/data': 0.0008338309999089688, 'timing/model': 1.1974956409831066, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 0.61}) +[2026-06-23 19:58:21,584][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.1152711808681488, 'train/grad_norm_pre_clip': 30.82814598083496, 'timing/data': 0.000784108997322619, 'timing/model': 1.1948114290134981, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 0.61}) +[2026-06-23 19:58:25,917][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.08677665144205093, 'train/grad_norm_pre_clip': 27.321443557739258, 'timing/data': 0.0007437950116582215, 'timing/model': 1.1766190330090467, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 0.61}) +[2026-06-23 19:58:30,238][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.10886652767658234, 'train/grad_norm_pre_clip': 34.65351486206055, 'timing/data': 0.0008025130082387477, 'timing/model': 1.1803779710025992, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.61}) +[2026-06-23 19:58:34,554][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.08552059531211853, 'train/grad_norm_pre_clip': 9.345097541809082, 'timing/data': 0.000779961992520839, 'timing/model': 1.1765885869972408, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 0.61}) +[2026-06-23 19:58:39,163][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.012200044468045235, 'train/grad_norm_pre_clip': 13.639100074768066, 'timing/data': 0.0007473709993064404, 'timing/model': 1.2022292299952824, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 0.61}) +[2026-06-23 19:58:43,515][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.1945871114730835, 'train/grad_norm_pre_clip': 38.7166748046875, 'timing/data': 0.000923337007407099, 'timing/model': 1.1826717509829905, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 0.61}) +[2026-06-23 19:58:47,839][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.07284969091415405, 'train/grad_norm_pre_clip': 41.49848937988281, 'timing/data': 0.0007452370191458613, 'timing/model': 1.180309794988716, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 0.61}) +[2026-06-23 19:58:52,156][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.10421225428581238, 'train/grad_norm_pre_clip': 27.44233512878418, 'timing/data': 0.0007831880066078156, 'timing/model': 1.17736183598754, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 0.61}) +[2026-06-23 19:58:56,791][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.12686964869499207, 'train/grad_norm_pre_clip': 38.310306549072266, 'timing/data': 0.001131280994741246, 'timing/model': 1.1988755619968288, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 0.61}) +[2026-06-23 19:59:01,109][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.134687602519989, 'train/grad_norm_pre_clip': 37.694393157958984, 'timing/data': 0.0009946280042640865, 'timing/model': 1.1785593900131062, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 0.61}) +[2026-06-23 19:59:05,455][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.010716984048485756, 'train/grad_norm_pre_clip': 25.890012741088867, 'timing/data': 0.0007491249998565763, 'timing/model': 1.1802782969898544, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 0.61}) +[2026-06-23 19:59:09,817][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.20679059624671936, 'train/grad_norm_pre_clip': 102.07426452636719, 'timing/data': 0.0008242230105679482, 'timing/model': 1.199844522983767, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 0.61}) +[2026-06-23 19:59:14,465][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.12761491537094116, 'train/grad_norm_pre_clip': 67.31021118164062, 'timing/data': 0.0013680779957212508, 'timing/model': 1.1996935140050482, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.62}) +[2026-06-23 19:59:18,813][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.11831623315811157, 'train/grad_norm_pre_clip': 35.32892608642578, 'timing/data': 0.0008753490110393614, 'timing/model': 1.1848152019956615, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 0.62}) +[2026-06-23 19:59:23,161][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.2386673390865326, 'train/grad_norm_pre_clip': 29.296192169189453, 'timing/data': 0.0008069619943853468, 'timing/model': 1.1831242569896858, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 0.62}) +[2026-06-23 19:59:27,527][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.03004135563969612, 'train/grad_norm_pre_clip': 26.871477127075195, 'timing/data': 0.0007651250052731484, 'timing/model': 1.2007591529982165, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 0.62}) +[2026-06-23 19:59:32,171][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.03899025917053223, 'train/grad_norm_pre_clip': 15.489349365234375, 'timing/data': 0.0008449709857814014, 'timing/model': 1.1814670340099838, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 0.62}) +[2026-06-23 19:59:36,542][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.3237439692020416, 'train/grad_norm_pre_clip': 75.53469848632812, 'timing/data': 0.0011908910237252712, 'timing/model': 1.1802678070089314, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 0.62}) +[2026-06-23 19:59:40,940][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.23770001530647278, 'train/grad_norm_pre_clip': 44.155311584472656, 'timing/data': 0.0009368019818793982, 'timing/model': 1.2026411429978907, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 0.62}) +[2026-06-23 19:59:45,296][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.04175911471247673, 'train/grad_norm_pre_clip': 30.43160629272461, 'timing/data': 0.0007485840178560466, 'timing/model': 1.1792156519950368, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 0.62}) +[2026-06-23 19:59:49,983][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.25448915362358093, 'train/grad_norm_pre_clip': 57.58416748046875, 'timing/data': 0.00094291300047189, 'timing/model': 1.2129331459873356, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 0.62}) +[2026-06-23 19:59:54,388][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.10579873621463776, 'train/grad_norm_pre_clip': 22.753238677978516, 'timing/data': 0.0008605509938206524, 'timing/model': 1.208329977991525, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 0.62}) +[2026-06-23 19:59:58,795][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1600 +[2026-06-23 20:00:56,082][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1600 +[2026-06-23 20:00:56,082][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1600 +[2026-06-23 20:01:33,000][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1600 +[2026-06-23 20:01:33,001][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.0642595961689949, 'train/grad_norm_pre_clip': 12.540711402893066, 'eval/action_loss/seconds': 57.282489680015715, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.05723099783062935, 'eval/latency_2/loss': 0.05723099783062935, 'eval/flappy/loss': 0.05723099783062935, 'eval/flappy/latency_2/loss': 0.05723099783062935, 'eval/flap_precision': 0.9028290282902829, 'eval/flap_recall': 0.9959294436906377, 'eval/flap_f1': 0.9470967741935484, 'eval/flap_rate_model': 0.22583333333333333, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9470967741935484, 'eval/latency_2/flap_precision': 0.9028290282902829, 'eval/latency_2/flap_recall': 0.9959294436906377, 'eval/latency_2/flap_f1': 0.9470967741935484, 'eval/latency_2/flap_rate_model': 0.22583333333333333, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9470967741935484, 'eval/action_classification/data_wait_seconds': 0.42137786746025085, 'eval/action_classification/predict_seconds': 36.33797836303711, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0012433480005711317, 'timing/model': 1.2049020710110199, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.62}) +[2026-06-23 20:01:37,590][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.020075088366866112, 'train/grad_norm_pre_clip': 17.604978561401367, 'timing/data': 0.0007764649926684797, 'timing/model': 1.4253641840186901, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 0.62}) +[2026-06-23 20:01:41,890][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.09579220414161682, 'train/grad_norm_pre_clip': 23.167646408081055, 'timing/data': 0.0007894389855209738, 'timing/model': 1.1683344399789348, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 0.62}) +[2026-06-23 20:01:46,190][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.22514446079730988, 'train/grad_norm_pre_clip': 33.27405548095703, 'timing/data': 0.0007491550059057772, 'timing/model': 1.1671836730092764, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 0.62}) +[2026-06-23 20:01:50,477][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.021204138174653053, 'train/grad_norm_pre_clip': 14.19898796081543, 'timing/data': 0.0008478269737679511, 'timing/model': 1.1700131430116016, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 0.62}) +[2026-06-23 20:01:54,975][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.09813442081212997, 'train/grad_norm_pre_clip': 14.460028648376465, 'timing/data': 0.000953151989961043, 'timing/model': 1.1759914540161844, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.62}) +[2026-06-23 20:01:59,331][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.07420843839645386, 'train/grad_norm_pre_clip': 7.362521648406982, 'timing/data': 0.0007444559887517244, 'timing/model': 1.1881875960098114, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 0.62}) +[2026-06-23 20:02:03,626][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.09628712385892868, 'train/grad_norm_pre_clip': 19.038503646850586, 'timing/data': 0.0009062949975486845, 'timing/model': 1.1766541480028536, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 0.62}) +[2026-06-23 20:02:07,984][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.12812237441539764, 'train/grad_norm_pre_clip': 17.493179321289062, 'timing/data': 0.0010073820012621582, 'timing/model': 1.190838797017932, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 0.62}) +[2026-06-23 20:02:12,486][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.04157695919275284, 'train/grad_norm_pre_clip': 15.651131629943848, 'timing/data': 0.0008202859899029136, 'timing/model': 1.1745475140050985, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 0.62}) +[2026-06-23 20:02:16,812][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.10515601187944412, 'train/grad_norm_pre_clip': 22.9978084564209, 'timing/data': 0.0007427030068356544, 'timing/model': 1.1752163599885534, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.62}) +[2026-06-23 20:02:21,161][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.10053163766860962, 'train/grad_norm_pre_clip': 9.223669052124023, 'timing/data': 0.0008066709851846099, 'timing/model': 1.1922236170212273, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 0.62}) +[2026-06-23 20:02:25,540][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.030889339745044708, 'train/grad_norm_pre_clip': 22.516672134399414, 'timing/data': 0.000743425014661625, 'timing/model': 1.1955667359870858, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 0.62}) +[2026-06-23 20:02:30,117][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.1541692614555359, 'train/grad_norm_pre_clip': 33.855125427246094, 'timing/data': 0.0007579509983770549, 'timing/model': 1.1786898899881635, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 0.62}) +[2026-06-23 20:02:34,454][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.048488304018974304, 'train/grad_norm_pre_clip': 30.194089889526367, 'timing/data': 0.0007781280146446079, 'timing/model': 1.1992809009971097, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 0.62}) +[2026-06-23 20:02:38,828][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.07054521143436432, 'train/grad_norm_pre_clip': 17.954463958740234, 'timing/data': 0.0007397479785140604, 'timing/model': 1.1950091260077897, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 0.63}) +[2026-06-23 20:02:43,198][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.41653063893318176, 'train/grad_norm_pre_clip': 30.213380813598633, 'timing/data': 0.0007239780097734183, 'timing/model': 1.1792760349926539, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 0.63}) +[2026-06-23 20:02:47,800][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.1508825272321701, 'train/grad_norm_pre_clip': 39.145938873291016, 'timing/data': 0.001517253986094147, 'timing/model': 1.2094156139937695, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 0.63}) +[2026-06-23 20:02:52,140][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.01192456018179655, 'train/grad_norm_pre_clip': 16.018659591674805, 'timing/data': 0.0009496150014456362, 'timing/model': 1.1922471500001848, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 0.63}) +[2026-06-23 20:02:56,476][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.04461415857076645, 'train/grad_norm_pre_clip': 32.513710021972656, 'timing/data': 0.0007304399914573878, 'timing/model': 1.1944893050240353, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.63}) +[2026-06-23 20:03:00,799][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.005961686838418245, 'train/grad_norm_pre_clip': 11.847609519958496, 'timing/data': 0.0007270139758475125, 'timing/model': 1.1738404479983728, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.63}) +[2026-06-23 20:03:05,318][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.25233662128448486, 'train/grad_norm_pre_clip': 23.017162322998047, 'timing/data': 0.0008122210274450481, 'timing/model': 1.2008556830114685, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 0.63}) +[2026-06-23 20:03:09,664][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.01857740432024002, 'train/grad_norm_pre_clip': 19.193239212036133, 'timing/data': 0.0008129130001179874, 'timing/model': 1.1745633230020758, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 0.63}) +[2026-06-23 20:03:13,973][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.08290896564722061, 'train/grad_norm_pre_clip': 23.122556686401367, 'timing/data': 0.000780823000241071, 'timing/model': 1.175819896016037, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 0.63}) +[2026-06-23 20:03:18,563][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.09743355214595795, 'train/grad_norm_pre_clip': 15.590923309326172, 'timing/data': 0.0007602850091643631, 'timing/model': 1.4455346770118922, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 0.63}) +[2026-06-23 20:03:22,947][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.04557592794299126, 'train/grad_norm_pre_clip': 15.269386291503906, 'timing/data': 0.0007705850002821535, 'timing/model': 1.202673482010141, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 0.63}) +[2026-06-23 20:03:27,325][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.1473110318183899, 'train/grad_norm_pre_clip': 40.108375549316406, 'timing/data': 0.0009340769902337343, 'timing/model': 1.2064843559928704, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 0.63}) +[2026-06-23 20:03:31,675][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.05773220583796501, 'train/grad_norm_pre_clip': 33.77323913574219, 'timing/data': 0.0009139689791481942, 'timing/model': 1.2041646389989182, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 0.63}) +[2026-06-23 20:03:36,300][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.2207978218793869, 'train/grad_norm_pre_clip': 36.31991195678711, 'timing/data': 0.0007923850207589567, 'timing/model': 1.4824807179975323, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 0.63}) +[2026-06-23 20:03:40,644][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.05135231465101242, 'train/grad_norm_pre_clip': 17.30420684814453, 'timing/data': 0.0008453830087091774, 'timing/model': 1.1855606509780046, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 0.63}) +[2026-06-23 20:03:44,988][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.011254272423684597, 'train/grad_norm_pre_clip': 19.190593719482422, 'timing/data': 0.0007903010118752718, 'timing/model': 1.1893907900084741, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.63}) +[2026-06-23 20:03:49,363][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.09458678215742111, 'train/grad_norm_pre_clip': 22.435060501098633, 'timing/data': 0.0009324930142611265, 'timing/model': 1.2068958269956056, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 0.63}) +[2026-06-23 20:03:53,886][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.11437362432479858, 'train/grad_norm_pre_clip': 20.91437530517578, 'timing/data': 0.0009119060123339295, 'timing/model': 1.1824536680069286, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 0.63}) +[2026-06-23 20:03:58,226][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.06822364032268524, 'train/grad_norm_pre_clip': 20.513540267944336, 'timing/data': 0.0007365919882431626, 'timing/model': 1.173931386001641, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 0.63}) +[2026-06-23 20:04:02,565][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.08009988814592361, 'train/grad_norm_pre_clip': 17.345802307128906, 'timing/data': 0.0007307000050786883, 'timing/model': 1.2081573890172876, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 0.63}) +[2026-06-23 20:04:06,899][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.10952319949865341, 'train/grad_norm_pre_clip': 21.03502655029297, 'timing/data': 0.0010318279964849353, 'timing/model': 1.1874742579821032, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 0.63}) +[2026-06-23 20:04:11,432][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.15547123551368713, 'train/grad_norm_pre_clip': 27.993125915527344, 'timing/data': 0.0008833929896354675, 'timing/model': 1.1931596570066176, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 0.63}) +[2026-06-23 20:04:15,804][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.056427013128995895, 'train/grad_norm_pre_clip': 14.89021110534668, 'timing/data': 0.000804926996352151, 'timing/model': 1.2039113299979363, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 0.63}) +[2026-06-23 20:04:20,194][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.1647837907075882, 'train/grad_norm_pre_clip': 24.772907257080078, 'timing/data': 0.0008080240222625434, 'timing/model': 1.2005283269973006, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 0.63}) +[2026-06-23 20:04:24,538][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.03914571553468704, 'train/grad_norm_pre_clip': 25.050518035888672, 'timing/data': 0.0008070619951467961, 'timing/model': 1.203982831997564, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 0.63}) +[2026-06-23 20:04:29,169][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.09568721801042557, 'train/grad_norm_pre_clip': 27.397066116333008, 'timing/data': 0.000750517996493727, 'timing/model': 1.1799333989911247, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.63}) +[2026-06-23 20:04:33,546][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.15290771424770355, 'train/grad_norm_pre_clip': 22.32459831237793, 'timing/data': 0.0007378740119747818, 'timing/model': 1.2012460439873394, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 0.64}) +[2026-06-23 20:04:37,914][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.02953050099313259, 'train/grad_norm_pre_clip': 26.51337432861328, 'timing/data': 0.000735609995899722, 'timing/model': 1.2019726370053831, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 0.64}) +[2026-06-23 20:04:42,276][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.033940162509679794, 'train/grad_norm_pre_clip': 14.433554649353027, 'timing/data': 0.0007266730244737118, 'timing/model': 1.2097492299799342, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 0.64}) +[2026-06-23 20:04:46,846][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.025180868804454803, 'train/grad_norm_pre_clip': 17.08518409729004, 'timing/data': 0.0007821159961167723, 'timing/model': 1.2044737100077327, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 0.64}) +[2026-06-23 20:04:51,235][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.05360803008079529, 'train/grad_norm_pre_clip': 26.050832748413086, 'timing/data': 0.0009542639891151339, 'timing/model': 1.1855544999998529, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 0.64}) +[2026-06-23 20:04:55,606][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.08134070783853531, 'train/grad_norm_pre_clip': 15.309136390686035, 'timing/data': 0.0007954099855851382, 'timing/model': 1.205020911002066, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 0.64}) +[2026-06-23 20:04:59,995][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.08555413782596588, 'train/grad_norm_pre_clip': 27.10015106201172, 'timing/data': 0.0009363910066895187, 'timing/model': 1.2166635619942099, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 0.64}) +[2026-06-23 20:05:04,694][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.09171579033136368, 'train/grad_norm_pre_clip': 17.4175968170166, 'timing/data': 0.0009498359868302941, 'timing/model': 1.5064805189904291, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 0.64}) +[2026-06-23 20:05:09,026][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.1038670763373375, 'train/grad_norm_pre_clip': 13.192293167114258, 'timing/data': 0.000781454989919439, 'timing/model': 1.190915758983465, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 0.64}) +[2026-06-23 20:05:13,403][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.13500261306762695, 'train/grad_norm_pre_clip': 10.914161682128906, 'timing/data': 0.0009395570086780936, 'timing/model': 1.205388350004796, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.64}) +[2026-06-23 20:05:17,776][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.22771760821342468, 'train/grad_norm_pre_clip': 34.698158264160156, 'timing/data': 0.0008015219937078655, 'timing/model': 1.2109480570070446, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 0.64}) +[2026-06-23 20:05:22,164][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.09511435031890869, 'train/grad_norm_pre_clip': 25.47399139404297, 'timing/data': 0.0007786389905959368, 'timing/model': 1.2100427930126898, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 0.64}) +[2026-06-23 20:05:26,789][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.09073740988969803, 'train/grad_norm_pre_clip': 16.898775100708008, 'timing/data': 0.0009382540010847151, 'timing/model': 1.1848760560096707, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 0.64}) +[2026-06-23 20:05:31,158][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.19852833449840546, 'train/grad_norm_pre_clip': 26.99139976501465, 'timing/data': 0.000787816010415554, 'timing/model': 1.2163144759833813, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 0.64}) +[2026-06-23 20:05:35,613][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.1385965496301651, 'train/grad_norm_pre_clip': 23.410614013671875, 'timing/data': 0.0007479630003217608, 'timing/model': 1.1837210009980481, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 0.64}) +[2026-06-23 20:05:40,289][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.015984687954187393, 'train/grad_norm_pre_clip': 18.781444549560547, 'timing/data': 0.0007436749874614179, 'timing/model': 1.4506716420000885, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 0.64}) +[2026-06-23 20:05:44,626][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.18887914717197418, 'train/grad_norm_pre_clip': 39.48277282714844, 'timing/data': 0.0007301389996428043, 'timing/model': 1.1800964009889867, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 0.64}) +[2026-06-23 20:05:49,008][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.17353598773479462, 'train/grad_norm_pre_clip': 24.9895076751709, 'timing/data': 0.0008390009752474725, 'timing/model': 1.2104500970162917, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 0.64}) +[2026-06-23 20:05:53,381][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.01569845713675022, 'train/grad_norm_pre_clip': 14.86678409576416, 'timing/data': 0.0007557170174550265, 'timing/model': 1.1846895400085486, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 0.64}) +[2026-06-23 20:05:58,019][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.054915063083171844, 'train/grad_norm_pre_clip': 26.107019424438477, 'timing/data': 0.0008048480085562915, 'timing/model': 1.4483195449865889, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.64}) +[2026-06-23 20:06:02,393][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.07553543895483017, 'train/grad_norm_pre_clip': 20.596878051757812, 'timing/data': 0.0009269629954360425, 'timing/model': 1.1836067489930429, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 0.64}) +[2026-06-23 20:06:06,784][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.061944399029016495, 'train/grad_norm_pre_clip': 11.398999214172363, 'timing/data': 0.0007609669992234558, 'timing/model': 1.198523141007172, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 0.64}) +[2026-06-23 20:06:11,169][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.014141377061605453, 'train/grad_norm_pre_clip': 11.17308235168457, 'timing/data': 0.000782315997639671, 'timing/model': 1.1895604640012607, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 0.64}) +[2026-06-23 20:06:15,714][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.11990150064229965, 'train/grad_norm_pre_clip': 21.22588539123535, 'timing/data': 0.0007434739964082837, 'timing/model': 1.1801793340127915, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 0.64}) +[2026-06-23 20:06:20,109][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.1525682806968689, 'train/grad_norm_pre_clip': 21.482662200927734, 'timing/data': 0.0007508279813919216, 'timing/model': 1.2130463650100864, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 0.64}) +[2026-06-23 20:06:24,509][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.06910237669944763, 'train/grad_norm_pre_clip': 14.902545928955078, 'timing/data': 0.0007720380090177059, 'timing/model': 1.201343733991962, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 0.64}) +[2026-06-23 20:06:28,868][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.06509961187839508, 'train/grad_norm_pre_clip': 12.904848098754883, 'timing/data': 0.001230814988957718, 'timing/model': 1.1803668409993406, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 0.65}) +[2026-06-23 20:06:33,414][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.11504853516817093, 'train/grad_norm_pre_clip': 19.116985321044922, 'timing/data': 0.0007626889855600893, 'timing/model': 1.209137028985424, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.65}) +[2026-06-23 20:06:37,788][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.10636980831623077, 'train/grad_norm_pre_clip': 16.22121238708496, 'timing/data': 0.0007434549916069955, 'timing/model': 1.182535158994142, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 0.65}) +[2026-06-23 20:06:42,146][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.06283309310674667, 'train/grad_norm_pre_clip': 13.31343936920166, 'timing/data': 0.0009906009945552796, 'timing/model': 1.1840122890134808, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.65}) +[2026-06-23 20:06:46,542][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.16378243267536163, 'train/grad_norm_pre_clip': 15.441485404968262, 'timing/data': 0.0007386459910776466, 'timing/model': 1.2046171049878467, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 0.65}) +[2026-06-23 20:06:51,213][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.016345027834177017, 'train/grad_norm_pre_clip': 15.088648796081543, 'timing/data': 0.0008994829840958118, 'timing/model': 1.216535476007266, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 0.65}) +[2026-06-23 20:06:55,581][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.0051139225251972675, 'train/grad_norm_pre_clip': 14.65872573852539, 'timing/data': 0.0008306660165544599, 'timing/model': 1.2036298699968029, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 0.65}) +[2026-06-23 20:06:59,946][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.002351676346734166, 'train/grad_norm_pre_clip': 8.2640962600708, 'timing/data': 0.0009548649832140654, 'timing/model': 1.1976768369786441, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 0.65}) +[2026-06-23 20:07:04,329][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.08839237689971924, 'train/grad_norm_pre_clip': 25.38119125366211, 'timing/data': 0.000748984981328249, 'timing/model': 1.2027888159791473, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.65}) +[2026-06-23 20:07:08,901][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.05497303977608681, 'train/grad_norm_pre_clip': 45.18671417236328, 'timing/data': 0.0011613370152190328, 'timing/model': 1.196885563986143, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 0.65}) +[2026-06-23 20:07:13,252][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.07582076638936996, 'train/grad_norm_pre_clip': 22.12969207763672, 'timing/data': 0.0007583519909530878, 'timing/model': 1.184445440012496, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 0.65}) +[2026-06-23 20:07:17,619][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.034381404519081116, 'train/grad_norm_pre_clip': 20.317787170410156, 'timing/data': 0.0011159419955220073, 'timing/model': 1.2113652380066924, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 0.65}) +[2026-06-23 20:07:22,013][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.06786004453897476, 'train/grad_norm_pre_clip': 13.481566429138184, 'timing/data': 0.0008273090061265975, 'timing/model': 1.2132141250185668, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 0.65}) +[2026-06-23 20:07:26,596][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.02921348251402378, 'train/grad_norm_pre_clip': 39.349185943603516, 'timing/data': 0.0007827070076018572, 'timing/model': 1.426958611002192, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.65}) +[2026-06-23 20:07:30,943][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.10162369906902313, 'train/grad_norm_pre_clip': 22.294950485229492, 'timing/data': 0.0007894290029071271, 'timing/model': 1.1817325149895623, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 0.65}) +[2026-06-23 20:07:35,292][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.02314230240881443, 'train/grad_norm_pre_clip': 21.561552047729492, 'timing/data': 0.0007775969861540943, 'timing/model': 1.181575194990728, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 0.65}) +[2026-06-23 20:07:39,679][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.08485067635774612, 'train/grad_norm_pre_clip': 27.214645385742188, 'timing/data': 0.0007613180205225945, 'timing/model': 1.2022896129928995, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 0.65}) +[2026-06-23 20:07:44,273][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.0846274271607399, 'train/grad_norm_pre_clip': 16.336000442504883, 'timing/data': 0.000910894013941288, 'timing/model': 1.443343727994943, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 0.65}) +[2026-06-23 20:07:48,626][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.012713804841041565, 'train/grad_norm_pre_clip': 13.150379180908203, 'timing/data': 0.0007525800028815866, 'timing/model': 1.203664533997653, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 0.65}) +[2026-06-23 20:07:52,967][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.06620189547538757, 'train/grad_norm_pre_clip': 16.784624099731445, 'timing/data': 0.0009149109828285873, 'timing/model': 1.1857918590249028, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 0.65}) +[2026-06-23 20:07:57,310][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.047330912202596664, 'train/grad_norm_pre_clip': 19.060033798217773, 'timing/data': 0.0007672680076211691, 'timing/model': 1.1852514899801463, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 0.65}) +[2026-06-23 20:08:01,853][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.04335547238588333, 'train/grad_norm_pre_clip': 26.985902786254883, 'timing/data': 0.00078658401616849, 'timing/model': 1.1844360830145888, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 0.65}) +[2026-06-23 20:08:06,193][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.07381318509578705, 'train/grad_norm_pre_clip': 20.475597381591797, 'timing/data': 0.000732514017727226, 'timing/model': 1.1832299930101726, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 0.65}) +[2026-06-23 20:08:10,534][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.022179171442985535, 'train/grad_norm_pre_clip': 30.161602020263672, 'timing/data': 0.0007872749993111938, 'timing/model': 1.1839490619895514, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.65}) +[2026-06-23 20:08:14,895][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.22994905710220337, 'train/grad_norm_pre_clip': 21.734617233276367, 'timing/data': 0.0011473810009192675, 'timing/model': 1.1840285299986135, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 0.65}) +[2026-06-23 20:08:19,426][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.07882536202669144, 'train/grad_norm_pre_clip': 40.12059020996094, 'timing/data': 0.0007801819883752614, 'timing/model': 1.208437667984981, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 0.65}) +[2026-06-23 20:08:23,814][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.1427249312400818, 'train/grad_norm_pre_clip': 39.796390533447266, 'timing/data': 0.001377024978864938, 'timing/model': 1.2037481489824131, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 0.66}) +[2026-06-23 20:08:28,204][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.13225331902503967, 'train/grad_norm_pre_clip': 21.002859115600586, 'timing/data': 0.0008046170114539564, 'timing/model': 1.2107923289877363, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 0.66}) +[2026-06-23 20:08:32,596][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.1830289661884308, 'train/grad_norm_pre_clip': 18.866220474243164, 'timing/data': 0.000804628012701869, 'timing/model': 1.204789363022428, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 0.66}) +[2026-06-23 20:08:37,196][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.017010275274515152, 'train/grad_norm_pre_clip': 11.80675983428955, 'timing/data': 0.0007961619994603097, 'timing/model': 1.1836220179975498, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 0.66}) +[2026-06-23 20:08:41,535][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.15105538070201874, 'train/grad_norm_pre_clip': 19.152799606323242, 'timing/data': 0.0007976439956109971, 'timing/model': 1.1855456630000845, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 0.66}) +[2026-06-23 20:08:45,881][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.04598129168152809, 'train/grad_norm_pre_clip': 26.32967758178711, 'timing/data': 0.0008292929851450026, 'timing/model': 1.18625840000459, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 0.66}) +[2026-06-23 20:08:50,225][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.02355656772851944, 'train/grad_norm_pre_clip': 17.54159927368164, 'timing/data': 0.0011813940072897822, 'timing/model': 1.187685398006579, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 0.66}) +[2026-06-23 20:08:54,799][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1700 +[2026-06-23 20:09:52,620][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1700 +[2026-06-23 20:09:52,621][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1700 +[2026-06-23 20:10:29,743][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1700 +[2026-06-23 20:10:29,745][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.22857694327831268, 'train/grad_norm_pre_clip': 17.555492401123047, 'eval/action_loss/seconds': 57.818269902985776, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.048348747193813324, 'eval/latency_2/loss': 0.048348747193813324, 'eval/flappy/loss': 0.048348747193813324, 'eval/flappy/latency_2/loss': 0.048348747193813324, 'eval/flap_precision': 0.9509536784741145, 'eval/flap_recall': 0.9470827679782904, 'eval/flap_f1': 0.9490142760027193, 'eval/flap_rate_model': 0.2038888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9490142760027191, 'eval/latency_2/flap_precision': 0.9509536784741145, 'eval/latency_2/flap_recall': 0.9470827679782904, 'eval/latency_2/flap_f1': 0.9490142760027193, 'eval/latency_2/flap_rate_model': 0.2038888888888889, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9490142760027191, 'eval/action_classification/data_wait_seconds': 0.45873934030532837, 'eval/action_classification/predict_seconds': 36.51819610595703, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0009266319975722581, 'timing/model': 1.1856575900164898, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.66}) +[2026-06-23 20:10:34,113][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.026080839335918427, 'train/grad_norm_pre_clip': 10.229087829589844, 'timing/data': 0.001000649994239211, 'timing/model': 1.1817011670209467, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 0.66}) +[2026-06-23 20:10:38,452][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.1886039674282074, 'train/grad_norm_pre_clip': 12.664514541625977, 'timing/data': 0.0009845199820119888, 'timing/model': 1.1894835330022033, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 0.66}) +[2026-06-23 20:10:42,991][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.10444780439138412, 'train/grad_norm_pre_clip': 14.89050579071045, 'timing/data': 0.0011646320053841919, 'timing/model': 1.3745957239880227, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 0.66}) +[2026-06-23 20:10:47,302][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.2444324493408203, 'train/grad_norm_pre_clip': 38.45235824584961, 'timing/data': 0.0008335309976246208, 'timing/model': 1.179594443005044, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 0.66}) +[2026-06-23 20:10:51,676][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.02373194880783558, 'train/grad_norm_pre_clip': 26.26373863220215, 'timing/data': 0.0011317720054648817, 'timing/model': 1.2028871270013042, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 0.66}) +[2026-06-23 20:10:56,020][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.20575393736362457, 'train/grad_norm_pre_clip': 22.380373001098633, 'timing/data': 0.0014387289993464947, 'timing/model': 1.1901020750228781, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 0.66}) +[2026-06-23 20:11:00,362][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.002778183901682496, 'train/grad_norm_pre_clip': 9.884078025817871, 'timing/data': 0.0008415259944740683, 'timing/model': 1.1970605179958511, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 0.66}) +[2026-06-23 20:11:04,881][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.024257471784949303, 'train/grad_norm_pre_clip': 16.418094635009766, 'timing/data': 0.0008791050058789551, 'timing/model': 1.1745340200141072, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 0.66}) +[2026-06-23 20:11:09,216][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.007910256274044514, 'train/grad_norm_pre_clip': 28.522964477539062, 'timing/data': 0.001206559012643993, 'timing/model': 1.1977419469913002, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 0.66}) +[2026-06-23 20:11:13,551][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.04350464791059494, 'train/grad_norm_pre_clip': 39.71146774291992, 'timing/data': 0.0008901459805201739, 'timing/model': 1.1762497710005846, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.66}) +[2026-06-23 20:11:17,885][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.015030928887426853, 'train/grad_norm_pre_clip': 13.50188159942627, 'timing/data': 0.0010087309929076582, 'timing/model': 1.1768411849916447, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 0.66}) +[2026-06-23 20:11:22,488][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.13636934757232666, 'train/grad_norm_pre_clip': 14.906904220581055, 'timing/data': 0.0008829889993648976, 'timing/model': 1.1936769550084136, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 0.66}) +[2026-06-23 20:11:26,817][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.09822674095630646, 'train/grad_norm_pre_clip': 29.043603897094727, 'timing/data': 0.0008059170213527977, 'timing/model': 1.1778934659960214, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 0.66}) +[2026-06-23 20:11:31,145][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.10871899873018265, 'train/grad_norm_pre_clip': 20.90180206298828, 'timing/data': 0.0009042379970196635, 'timing/model': 1.1781822820194066, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 0.66}) +[2026-06-23 20:11:35,489][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.06383910030126572, 'train/grad_norm_pre_clip': 23.45724105834961, 'timing/data': 0.0008660870080348104, 'timing/model': 1.1805188250145875, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 0.66}) +[2026-06-23 20:11:40,054][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.026540467515587807, 'train/grad_norm_pre_clip': 13.258525848388672, 'timing/data': 0.0009150279802270234, 'timing/model': 1.1989475229929667, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 0.66}) +[2026-06-23 20:11:44,397][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.0467335470020771, 'train/grad_norm_pre_clip': 14.18544864654541, 'timing/data': 0.0008904830028768629, 'timing/model': 1.1807799240050372, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.66}) +[2026-06-23 20:11:48,739][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.126163050532341, 'train/grad_norm_pre_clip': 14.195961952209473, 'timing/data': 0.0009461250156164169, 'timing/model': 1.198335680004675, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 0.66}) +[2026-06-23 20:11:53,303][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.0336526557803154, 'train/grad_norm_pre_clip': 9.650330543518066, 'timing/data': 0.0008694840071257204, 'timing/model': 1.3812264290172607, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 0.67}) +[2026-06-23 20:11:57,620][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.15218232572078705, 'train/grad_norm_pre_clip': 23.1081600189209, 'timing/data': 0.0007341039890889078, 'timing/model': 1.1749670789868105, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.67}) +[2026-06-23 20:12:01,941][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.019148146733641624, 'train/grad_norm_pre_clip': 7.879498481750488, 'timing/data': 0.0007006229716353118, 'timing/model': 1.188144558982458, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 0.67}) +[2026-06-23 20:12:06,306][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.10492297261953354, 'train/grad_norm_pre_clip': 20.328462600708008, 'timing/data': 0.0007327119819819927, 'timing/model': 1.1974955509940628, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 0.67}) +[2026-06-23 20:12:10,761][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.21065054833889008, 'train/grad_norm_pre_clip': 15.741569519042969, 'timing/data': 0.0008476240036543459, 'timing/model': 1.1761965019977652, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 0.67}) +[2026-06-23 20:12:15,077][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.009209797717630863, 'train/grad_norm_pre_clip': 18.141338348388672, 'timing/data': 0.0006946109933778644, 'timing/model': 1.1783347369928379, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 0.67}) +[2026-06-23 20:12:19,393][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.12130734324455261, 'train/grad_norm_pre_clip': 22.022045135498047, 'timing/data': 0.0006852739898022264, 'timing/model': 1.1767376310017426, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 0.67}) +[2026-06-23 20:12:23,710][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.09295012801885605, 'train/grad_norm_pre_clip': 16.983413696289062, 'timing/data': 0.0007407169905491173, 'timing/model': 1.1778091870073695, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 0.67}) +[2026-06-23 20:12:28,281][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.06731124222278595, 'train/grad_norm_pre_clip': 16.906164169311523, 'timing/data': 0.0008479840180370957, 'timing/model': 1.4281430850096513, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 0.67}) +[2026-06-23 20:12:32,654][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.07917582988739014, 'train/grad_norm_pre_clip': 30.453279495239258, 'timing/data': 0.0007933640154078603, 'timing/model': 1.2045381820062175, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 0.67}) +[2026-06-23 20:12:37,048][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.3487464189529419, 'train/grad_norm_pre_clip': 40.48380661010742, 'timing/data': 0.0009648799896240234, 'timing/model': 1.2094870009750593, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 0.67}) +[2026-06-23 20:12:41,366][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.04262601584196091, 'train/grad_norm_pre_clip': 8.177931785583496, 'timing/data': 0.0009355060174129903, 'timing/model': 1.1792199109913781, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.67}) +[2026-06-23 20:12:45,926][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.05860370397567749, 'train/grad_norm_pre_clip': 9.61754035949707, 'timing/data': 0.001193672011140734, 'timing/model': 1.390239105996443, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 0.67}) +[2026-06-23 20:12:50,304][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.13039857149124146, 'train/grad_norm_pre_clip': 30.764225006103516, 'timing/data': 0.0011803779925685376, 'timing/model': 1.2073409339936916, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 0.67}) +[2026-06-23 20:12:54,664][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.08179409056901932, 'train/grad_norm_pre_clip': 23.657716751098633, 'timing/data': 0.0007707220211159438, 'timing/model': 1.1971090050064959, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 0.67}) +[2026-06-23 20:12:59,043][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.005918621551245451, 'train/grad_norm_pre_clip': 7.379477500915527, 'timing/data': 0.001159790001111105, 'timing/model': 1.2126988170202821, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 0.67}) +[2026-06-23 20:13:03,624][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.042691271752119064, 'train/grad_norm_pre_clip': 22.504894256591797, 'timing/data': 0.0007349959923885763, 'timing/model': 1.370961745007662, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 0.67}) +[2026-06-23 20:13:08,007][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.09881702065467834, 'train/grad_norm_pre_clip': 36.086875915527344, 'timing/data': 0.0011834540055133402, 'timing/model': 1.1828494110086467, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 0.67}) +[2026-06-23 20:13:12,398][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.06158730387687683, 'train/grad_norm_pre_clip': 24.321014404296875, 'timing/data': 0.0007660939882043749, 'timing/model': 1.2145683849812485, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 0.67}) +[2026-06-23 20:13:16,780][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.02639186568558216, 'train/grad_norm_pre_clip': 19.45815086364746, 'timing/data': 0.001033817999996245, 'timing/model': 1.2288704820093699, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 0.67}) +[2026-06-23 20:13:21,124][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.02755027823150158, 'train/grad_norm_pre_clip': 17.419790267944336, 'timing/data': 0.0008852829923853278, 'timing/model': 1.1839699909905903, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 0.67}) +[2026-06-23 20:13:25,776][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.05503667891025543, 'train/grad_norm_pre_clip': 25.80034637451172, 'timing/data': 0.0007813719857949764, 'timing/model': 1.2133608520089183, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.67}) +[2026-06-23 20:13:30,180][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.1302553415298462, 'train/grad_norm_pre_clip': 17.758277893066406, 'timing/data': 0.0010063269874081016, 'timing/model': 1.20861599498312, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 0.67}) +[2026-06-23 20:13:34,578][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.015671133995056152, 'train/grad_norm_pre_clip': 4.969626426696777, 'timing/data': 0.0007649610051885247, 'timing/model': 1.2266480549878906, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 0.67}) +[2026-06-23 20:13:39,133][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.06899821758270264, 'train/grad_norm_pre_clip': 18.762649536132812, 'timing/data': 0.00097583097522147, 'timing/model': 1.3637931229895912, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 0.67}) +[2026-06-23 20:13:43,488][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.028024714440107346, 'train/grad_norm_pre_clip': 13.289809226989746, 'timing/data': 0.0007483600056730211, 'timing/model': 1.1807928149937652, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 0.67}) +[2026-06-23 20:13:47,857][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.05362042784690857, 'train/grad_norm_pre_clip': 34.78632736206055, 'timing/data': 0.0007808409864082932, 'timing/model': 1.1828040420077741, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.68}) +[2026-06-23 20:13:52,225][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.01871718280017376, 'train/grad_norm_pre_clip': 16.306928634643555, 'timing/data': 0.0011202780005987734, 'timing/model': 1.180443421995733, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 0.68}) +[2026-06-23 20:13:56,652][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.052517589181661606, 'train/grad_norm_pre_clip': 35.42457962036133, 'timing/data': 0.0009222920052707195, 'timing/model': 1.2094976519874763, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 0.68}) +[2026-06-23 20:14:01,282][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.020045027136802673, 'train/grad_norm_pre_clip': 20.590717315673828, 'timing/data': 0.0007693290244787931, 'timing/model': 1.1804118860163726, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 0.68}) +[2026-06-23 20:14:05,673][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.0011397430207580328, 'train/grad_norm_pre_clip': 16.68209457397461, 'timing/data': 0.0009346040023956448, 'timing/model': 1.207655446982244, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 0.68}) +[2026-06-23 20:32:06,917][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1750 +[2026-06-23 20:32:06,919][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.13732321560382843, 'train/grad_norm_pre_clip': 23.361963272094727, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 279.9000000000077, 'rl_games_eval/mid_train/mean_length': 2272.8, 'rl_games_eval/mid_train/std_reward': 142.95424442808172, 'rl_games_eval/mid_train/std_length': 1151.6602624038046, 'rl_games_eval/mid_train/macro_mean_reward': 279.9000000000077, 'rl_games_eval/mid_train/macro_mean_length': 2272.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 279.9000000000077, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2272.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 142.95424442808172, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1151.6602624038046, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 279.9000000000077, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0009219310013577342, 'timing/model': 1.209272095002234, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.68}) +[2026-06-23 20:32:11,100][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.08063311129808426, 'train/grad_norm_pre_clip': 9.037156105041504, 'timing/data': 0.0009165939991362393, 'timing/model': 1.1424873239884619, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 0.68}) +[2026-06-23 20:32:15,286][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.017841987311840057, 'train/grad_norm_pre_clip': 26.383705139160156, 'timing/data': 0.0007700829883106053, 'timing/model': 1.1363948009966407, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 0.68}) +[2026-06-23 20:32:19,501][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.055622272193431854, 'train/grad_norm_pre_clip': 16.408334732055664, 'timing/data': 0.0007619079842697829, 'timing/model': 1.16762115698657, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 0.68}) +[2026-06-23 20:32:23,927][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.2333652377128601, 'train/grad_norm_pre_clip': 24.02450942993164, 'timing/data': 0.0009043710015248507, 'timing/model': 1.1412190490227658, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 0.68}) +[2026-06-23 20:32:28,118][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.07099594920873642, 'train/grad_norm_pre_clip': 11.695965766906738, 'timing/data': 0.000904090004041791, 'timing/model': 1.1423366569797508, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 0.68}) +[2026-06-23 20:32:32,344][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.13833631575107574, 'train/grad_norm_pre_clip': 25.376834869384766, 'timing/data': 0.0007594430062454194, 'timing/model': 1.1474773769732565, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 0.68}) +[2026-06-23 20:32:36,567][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.14584296941757202, 'train/grad_norm_pre_clip': 17.790712356567383, 'timing/data': 0.0007718059932813048, 'timing/model': 1.1480669800075702, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 0.68}) +[2026-06-23 20:32:40,973][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.07179807126522064, 'train/grad_norm_pre_clip': 9.232637405395508, 'timing/data': 0.0007183270063251257, 'timing/model': 1.1484969060111325, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 0.68}) +[2026-06-23 20:32:45,207][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.01886638253927231, 'train/grad_norm_pre_clip': 13.543668746948242, 'timing/data': 0.0007328939973376691, 'timing/model': 1.1642093299888074, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 0.68}) +[2026-06-23 20:32:49,436][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.04736267775297165, 'train/grad_norm_pre_clip': 19.04776382446289, 'timing/data': 0.001106725016143173, 'timing/model': 1.15018247498665, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.68}) +[2026-06-23 20:32:53,668][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.10283339023590088, 'train/grad_norm_pre_clip': 10.3783597946167, 'timing/data': 0.0007423920033033937, 'timing/model': 1.1729281430016272, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 0.68}) +[2026-06-23 20:32:58,086][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.18045148253440857, 'train/grad_norm_pre_clip': 22.790584564208984, 'timing/data': 0.0007762839959468693, 'timing/model': 1.1551851539989002, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 0.68}) +[2026-06-23 20:33:02,378][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.08903247863054276, 'train/grad_norm_pre_clip': 16.067975997924805, 'timing/data': 0.0010183309786953032, 'timing/model': 1.1800003849784844, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 0.68}) +[2026-06-23 20:33:06,649][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.2065613716840744, 'train/grad_norm_pre_clip': 16.632253646850586, 'timing/data': 0.000760465016355738, 'timing/model': 1.1752910880022682, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 0.68}) +[2026-06-23 20:33:10,926][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.032514844089746475, 'train/grad_norm_pre_clip': 21.062829971313477, 'timing/data': 0.0011474999773781747, 'timing/model': 1.1772493580065202, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 0.68}) +[2026-06-23 20:33:15,492][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.02022619917988777, 'train/grad_norm_pre_clip': 13.807095527648926, 'timing/data': 0.0006997730233706534, 'timing/model': 1.1563900810142513, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.68}) +[2026-06-23 20:33:19,807][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.0477529838681221, 'train/grad_norm_pre_clip': 16.52433967590332, 'timing/data': 0.0007611559994984418, 'timing/model': 1.1811782919976395, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 0.68}) +[2026-06-23 20:33:24,125][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.022708678618073463, 'train/grad_norm_pre_clip': 8.908522605895996, 'timing/data': 0.0007888179970905185, 'timing/model': 1.1628033249871805, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 0.68}) +[2026-06-23 20:33:28,392][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.13250742852687836, 'train/grad_norm_pre_clip': 30.247989654541016, 'timing/data': 0.0008054680074565113, 'timing/model': 1.1682886730122846, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 0.68}) +[2026-06-23 20:33:33,011][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.12072711437940598, 'train/grad_norm_pre_clip': 45.38751983642578, 'timing/data': 0.0007294680108316243, 'timing/model': 1.1892126050079241, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.69}) +[2026-06-23 20:33:37,286][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.15095876157283783, 'train/grad_norm_pre_clip': 18.627246856689453, 'timing/data': 0.0007717570115346462, 'timing/model': 1.1687585650070105, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 0.69}) +[2026-06-23 20:33:41,605][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.1518331617116928, 'train/grad_norm_pre_clip': 28.237483978271484, 'timing/data': 0.0007742310117464513, 'timing/model': 1.2018572719825897, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 0.69}) +[2026-06-23 20:33:45,946][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.11002974957227707, 'train/grad_norm_pre_clip': 30.826519012451172, 'timing/data': 0.0007234170043375343, 'timing/model': 1.193400138989091, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 0.69}) +[2026-06-23 20:33:50,538][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.025000015273690224, 'train/grad_norm_pre_clip': 16.271366119384766, 'timing/data': 0.0007573399925604463, 'timing/model': 1.191400206007529, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 0.69}) +[2026-06-23 20:33:54,826][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.06266072392463684, 'train/grad_norm_pre_clip': 16.458660125732422, 'timing/data': 0.0007131169841159135, 'timing/model': 1.171408311027335, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 0.69}) +[2026-06-23 20:33:59,172][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.10995492339134216, 'train/grad_norm_pre_clip': 22.69170379638672, 'timing/data': 0.000750034989323467, 'timing/model': 1.2050740949925967, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 0.69}) +[2026-06-23 20:34:03,547][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.21686626970767975, 'train/grad_norm_pre_clip': 41.78620910644531, 'timing/data': 0.0007080679934006184, 'timing/model': 1.1924062419857364, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 0.69}) +[2026-06-23 20:34:08,183][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.030005335807800293, 'train/grad_norm_pre_clip': 14.706483840942383, 'timing/data': 0.000771925988374278, 'timing/model': 1.2029080579814035, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 0.69}) +[2026-06-23 20:34:12,549][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.09956695139408112, 'train/grad_norm_pre_clip': 17.14395523071289, 'timing/data': 0.0011361389770172536, 'timing/model': 1.2009868110180832, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 0.69}) +[2026-06-23 20:34:16,895][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.06161585450172424, 'train/grad_norm_pre_clip': 19.222536087036133, 'timing/data': 0.00079330601147376, 'timing/model': 1.1743599660112523, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.69}) +[2026-06-23 20:34:21,235][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.008693317882716656, 'train/grad_norm_pre_clip': 20.02968978881836, 'timing/data': 0.0007710439967922866, 'timing/model': 1.194651120022172, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 0.69}) +[2026-06-23 20:34:25,810][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.021148577332496643, 'train/grad_norm_pre_clip': 33.63965606689453, 'timing/data': 0.0007394260028377175, 'timing/model': 1.1751610210048966, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 0.69}) +[2026-06-23 20:34:30,181][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.1163100153207779, 'train/grad_norm_pre_clip': 33.53710174560547, 'timing/data': 0.0008613819954916835, 'timing/model': 1.1944730450049974, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 0.69}) +[2026-06-23 20:34:34,629][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.10886387526988983, 'train/grad_norm_pre_clip': 25.4781494140625, 'timing/data': 0.0007629790052305907, 'timing/model': 1.1981215690029785, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 0.69}) +[2026-06-23 20:34:39,021][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.09452750533819199, 'train/grad_norm_pre_clip': 19.914752960205078, 'timing/data': 0.0007838379824534059, 'timing/model': 1.1967614169989247, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 0.69}) +[2026-06-23 20:34:43,595][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.07340922206640244, 'train/grad_norm_pre_clip': 22.563800811767578, 'timing/data': 0.0008057790109887719, 'timing/model': 1.1801383169949986, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 0.69}) +[2026-06-23 20:34:47,968][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.020002538338303566, 'train/grad_norm_pre_clip': 6.948709964752197, 'timing/data': 0.000772717990912497, 'timing/model': 1.1969578759744763, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 0.69}) +[2026-06-23 20:34:52,354][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.30404940247535706, 'train/grad_norm_pre_clip': 26.47891616821289, 'timing/data': 0.0007418309978675097, 'timing/model': 1.2016709960007574, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 0.69}) +[2026-06-23 20:34:56,729][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.01243826188147068, 'train/grad_norm_pre_clip': 11.626155853271484, 'timing/data': 0.0007948889979161322, 'timing/model': 1.1971243770094588, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 0.69}) +[2026-06-23 20:35:01,294][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.006603355519473553, 'train/grad_norm_pre_clip': 37.22299575805664, 'timing/data': 0.0007749919896014035, 'timing/model': 1.1804283410019707, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.69}) +[2026-06-23 20:35:05,724][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.12210223078727722, 'train/grad_norm_pre_clip': 35.34511947631836, 'timing/data': 0.0007879559998400509, 'timing/model': 1.2092514790128917, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 0.69}) +[2026-06-23 20:35:10,110][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.08472134917974472, 'train/grad_norm_pre_clip': 18.49657440185547, 'timing/data': 0.0009430130012333393, 'timing/model': 1.1963605530036148, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 0.69}) +[2026-06-23 20:35:14,458][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.005535181611776352, 'train/grad_norm_pre_clip': 12.509095191955566, 'timing/data': 0.0007699129928369075, 'timing/model': 1.2041110949940048, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 0.69}) +[2026-06-23 20:35:19,104][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.08281555026769638, 'train/grad_norm_pre_clip': 18.205265045166016, 'timing/data': 0.0007486829999834299, 'timing/model': 1.1993079929961823, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 0.69}) +[2026-06-23 20:35:23,432][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.2237987071275711, 'train/grad_norm_pre_clip': 27.63275909423828, 'timing/data': 0.0009186469833366573, 'timing/model': 1.1805625700217206, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 0.69}) +[2026-06-23 20:35:27,795][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.09142737835645676, 'train/grad_norm_pre_clip': 16.086483001708984, 'timing/data': 0.0007497050100937486, 'timing/model': 1.1802686989831273, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 0.7}) +[2026-06-23 20:35:32,132][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.07227364927530289, 'train/grad_norm_pre_clip': 21.56035041809082, 'timing/data': 0.0009138179884757847, 'timing/model': 1.1816566259949468, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 0.7}) +[2026-06-23 20:35:36,705][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.04879556968808174, 'train/grad_norm_pre_clip': 15.028509140014648, 'timing/data': 0.0007551849994342774, 'timing/model': 1.179996830993332, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 0.7}) +[2026-06-23 20:35:41,088][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.12997683882713318, 'train/grad_norm_pre_clip': 19.90907859802246, 'timing/data': 0.0011098409886471927, 'timing/model': 1.2069955780170858, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 0.7}) +[2026-06-23 20:35:45,414][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1800 +[2026-06-23 20:36:43,283][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1800 +[2026-06-23 20:36:43,283][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1800 +[2026-06-23 20:37:20,149][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1800 +[2026-06-23 20:37:20,151][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.04119928553700447, 'train/grad_norm_pre_clip': 21.37612533569336, 'eval/action_loss/seconds': 57.86516486702021, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.04126366227865219, 'eval/latency_2/loss': 0.04126366227865219, 'eval/flappy/loss': 0.04126366227865219, 'eval/flappy/latency_2/loss': 0.04126366227865219, 'eval/flap_precision': 0.9657064471879286, 'eval/flap_recall': 0.9552238805970149, 'eval/flap_f1': 0.9604365620736699, 'eval/flap_rate_model': 0.2025, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.96043656207367, 'eval/latency_2/flap_precision': 0.9657064471879286, 'eval/latency_2/flap_recall': 0.9552238805970149, 'eval/latency_2/flap_f1': 0.9604365620736699, 'eval/latency_2/flap_rate_model': 0.2025, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.96043656207367, 'eval/action_classification/data_wait_seconds': 0.4572247266769409, 'eval/action_classification/predict_seconds': 36.265140533447266, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007524109969381243, 'timing/model': 1.180592445016373, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.7}) +[2026-06-23 20:37:24,763][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.09694472700357437, 'train/grad_norm_pre_clip': 18.798364639282227, 'timing/data': 0.0008154069946613163, 'timing/model': 1.4503770279989112, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 0.7}) +[2026-06-23 20:37:29,135][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.05182689428329468, 'train/grad_norm_pre_clip': 16.78562355041504, 'timing/data': 0.0009128469973802567, 'timing/model': 1.2043066839978565, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 0.7}) +[2026-06-23 20:37:33,469][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.17821167409420013, 'train/grad_norm_pre_clip': 26.036624908447266, 'timing/data': 0.0009092700202018023, 'timing/model': 1.1692886230011936, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 0.7}) +[2026-06-23 20:37:37,808][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.1300126612186432, 'train/grad_norm_pre_clip': 13.723626136779785, 'timing/data': 0.0011553939839359373, 'timing/model': 1.1989744949969463, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 0.7}) +[2026-06-23 20:37:42,423][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.07397650927305222, 'train/grad_norm_pre_clip': 20.98356056213379, 'timing/data': 0.0010518440103624016, 'timing/model': 1.4476689519942738, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 0.7}) +[2026-06-23 20:37:46,731][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.012824561446905136, 'train/grad_norm_pre_clip': 8.506935119628906, 'timing/data': 0.0007703129958827049, 'timing/model': 1.1729996850190219, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 0.7}) +[2026-06-23 20:37:51,072][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.16384534537792206, 'train/grad_norm_pre_clip': 15.351154327392578, 'timing/data': 0.0007275239913724363, 'timing/model': 1.1706308200082276, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 0.7}) +[2026-06-23 20:37:55,398][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.0786690041422844, 'train/grad_norm_pre_clip': 11.408742904663086, 'timing/data': 0.0009260309743694961, 'timing/model': 1.1994809169846121, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 0.7}) +[2026-06-23 20:37:59,991][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.0966244786977768, 'train/grad_norm_pre_clip': 16.193695068359375, 'timing/data': 0.000757319008698687, 'timing/model': 1.4419484599784482, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 0.7}) +[2026-06-23 20:38:04,364][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.023713955655694008, 'train/grad_norm_pre_clip': 12.245443344116211, 'timing/data': 0.0007599340169690549, 'timing/model': 1.1880486860172823, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.7}) +[2026-06-23 20:38:08,710][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.10236547142267227, 'train/grad_norm_pre_clip': 16.908245086669922, 'timing/data': 0.0010737039847299457, 'timing/model': 1.1788322089996655, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 0.7}) +[2026-06-23 20:38:13,048][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.11601222306489944, 'train/grad_norm_pre_clip': 13.634475708007812, 'timing/data': 0.0008392099989578128, 'timing/model': 1.1835757849912625, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 0.7}) +[2026-06-23 20:38:17,621][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.1932477504014969, 'train/grad_norm_pre_clip': 28.67340850830078, 'timing/data': 0.0007616270158905536, 'timing/model': 1.1974453129805624, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 0.7}) +[2026-06-23 20:38:21,953][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.04623452201485634, 'train/grad_norm_pre_clip': 21.405384063720703, 'timing/data': 0.000799166999058798, 'timing/model': 1.1793571730086114, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 0.7}) +[2026-06-23 20:38:26,294][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.040594637393951416, 'train/grad_norm_pre_clip': 13.022493362426758, 'timing/data': 0.0008613910176791251, 'timing/model': 1.1852095380018, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.7}) +[2026-06-23 20:38:30,629][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.1308228075504303, 'train/grad_norm_pre_clip': 14.077621459960938, 'timing/data': 0.0007563669933006167, 'timing/model': 1.1970000349974725, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 0.7}) +[2026-06-23 20:38:35,200][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.027843128889799118, 'train/grad_norm_pre_clip': 24.576051712036133, 'timing/data': 0.0007402980118058622, 'timing/model': 1.1968152359768283, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 0.7}) +[2026-06-23 20:38:39,581][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.16084562242031097, 'train/grad_norm_pre_clip': 29.962757110595703, 'timing/data': 0.00071106301038526, 'timing/model': 1.2014225799939595, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 0.7}) +[2026-06-23 20:38:43,904][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.1340344250202179, 'train/grad_norm_pre_clip': 22.752477645874023, 'timing/data': 0.0007543429965153337, 'timing/model': 1.182214809989091, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 0.7}) +[2026-06-23 20:38:48,281][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.2857036888599396, 'train/grad_norm_pre_clip': 20.406517028808594, 'timing/data': 0.0007496849866583943, 'timing/model': 1.1877658980083652, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.7}) +[2026-06-23 20:38:52,899][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.05595308542251587, 'train/grad_norm_pre_clip': 7.866039752960205, 'timing/data': 0.0007185879803728312, 'timing/model': 1.4292891990044154, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 0.7}) +[2026-06-23 20:38:57,270][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.015123364515602589, 'train/grad_norm_pre_clip': 15.35604190826416, 'timing/data': 0.0008797259943094105, 'timing/model': 1.181672702979995, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 0.71}) +[2026-06-23 20:39:01,613][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.023561643436551094, 'train/grad_norm_pre_clip': 23.48769187927246, 'timing/data': 0.0007204209978226572, 'timing/model': 1.1819097620027605, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 0.71}) +[2026-06-23 20:39:05,967][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.10776037722826004, 'train/grad_norm_pre_clip': 31.429027557373047, 'timing/data': 0.0007042310026008636, 'timing/model': 1.1792415239906404, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 0.71}) +[2026-06-23 20:39:10,592][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.12768356502056122, 'train/grad_norm_pre_clip': 32.970035552978516, 'timing/data': 0.0011197089916095138, 'timing/model': 1.4518176220008172, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 0.71}) +[2026-06-23 20:39:15,003][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.00863546784967184, 'train/grad_norm_pre_clip': 11.662484169006348, 'timing/data': 0.0008633359975647181, 'timing/model': 1.208049435983412, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 0.71}) +[2026-06-23 20:39:19,349][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.03478867933154106, 'train/grad_norm_pre_clip': 25.40901756286621, 'timing/data': 0.001056431996403262, 'timing/model': 1.1834553240041714, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 0.71}) +[2026-06-23 20:39:23,690][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.0624375119805336, 'train/grad_norm_pre_clip': 14.551013946533203, 'timing/data': 0.0009397960093338042, 'timing/model': 1.1803372539870907, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 0.71}) +[2026-06-23 20:39:28,318][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.11835814267396927, 'train/grad_norm_pre_clip': 21.156208038330078, 'timing/data': 0.0007498660124838352, 'timing/model': 1.4576421020028647, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 0.71}) +[2026-06-23 20:39:32,645][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.06333746016025543, 'train/grad_norm_pre_clip': 11.164313316345215, 'timing/data': 0.0008746560197323561, 'timing/model': 1.1789147869858425, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.71}) +[2026-06-23 20:39:36,978][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.02395196259021759, 'train/grad_norm_pre_clip': 22.68758201599121, 'timing/data': 0.0008938309911172837, 'timing/model': 1.1809339789906517, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 0.71}) +[2026-06-23 20:39:41,346][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.0239751897752285, 'train/grad_norm_pre_clip': 11.544379234313965, 'timing/data': 0.0009064449986908585, 'timing/model': 1.198675783991348, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 0.71}) +[2026-06-23 20:39:46,013][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.16811683773994446, 'train/grad_norm_pre_clip': 20.91007423400879, 'timing/data': 0.0008316469902638346, 'timing/model': 1.4800194059789646, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 0.71}) +[2026-06-23 20:39:50,384][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.08393152058124542, 'train/grad_norm_pre_clip': 16.092859268188477, 'timing/data': 0.001181041996460408, 'timing/model': 1.1976310460013337, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 0.71}) +[2026-06-23 20:39:54,735][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.10573231428861618, 'train/grad_norm_pre_clip': 26.60895538330078, 'timing/data': 0.0010652990022208542, 'timing/model': 1.1933336360089015, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 0.71}) +[2026-06-23 20:39:59,119][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.03222592547535896, 'train/grad_norm_pre_clip': 24.1562442779541, 'timing/data': 0.0007446359959430993, 'timing/model': 1.1971863560029306, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 0.71}) +[2026-06-23 20:40:03,731][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.07342491298913956, 'train/grad_norm_pre_clip': 30.471845626831055, 'timing/data': 0.0007202110136859119, 'timing/model': 1.4528659280040301, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 0.71}) +[2026-06-23 20:40:08,086][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.031057052314281464, 'train/grad_norm_pre_clip': 37.4963264465332, 'timing/data': 0.0009147200034931302, 'timing/model': 1.177867923019221, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 0.71}) +[2026-06-23 20:40:12,471][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.11986088752746582, 'train/grad_norm_pre_clip': 19.578577041625977, 'timing/data': 0.0011343660007696599, 'timing/model': 1.2019728340092115, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 0.71}) +[2026-06-23 20:40:16,871][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.0037492155097424984, 'train/grad_norm_pre_clip': 13.765080451965332, 'timing/data': 0.0009023869934026152, 'timing/model': 1.202976350003155, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.71}) +[2026-06-23 20:40:21,469][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.04911959171295166, 'train/grad_norm_pre_clip': 32.424072265625, 'timing/data': 0.0008810979779809713, 'timing/model': 1.436220317002153, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 0.71}) +[2026-06-23 20:40:25,802][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.20521961152553558, 'train/grad_norm_pre_clip': 33.69384765625, 'timing/data': 0.0007445859955623746, 'timing/model': 1.1825271880079526, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 0.71}) +[2026-06-23 20:40:30,195][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.10795552283525467, 'train/grad_norm_pre_clip': 23.598051071166992, 'timing/data': 0.0009414390078745782, 'timing/model': 1.2063227569742594, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 0.71}) +[2026-06-23 20:40:34,568][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.1573529839515686, 'train/grad_norm_pre_clip': 28.624866485595703, 'timing/data': 0.0009314909984823316, 'timing/model': 1.212767196004279, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 0.71}) +[2026-06-23 20:40:39,172][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.09178205579519272, 'train/grad_norm_pre_clip': 13.766209602355957, 'timing/data': 0.0009635109745431691, 'timing/model': 1.4341413720103446, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 0.71}) +[2026-06-23 20:40:43,558][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.04735441505908966, 'train/grad_norm_pre_clip': 16.040498733520508, 'timing/data': 0.0007754230173304677, 'timing/model': 1.2059230010199826, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 0.71}) +[2026-06-23 20:40:47,926][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.11347523331642151, 'train/grad_norm_pre_clip': 24.65999412536621, 'timing/data': 0.0012136920122429729, 'timing/model': 1.1895457509963308, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 0.71}) +[2026-06-23 20:40:52,319][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.03906377777457237, 'train/grad_norm_pre_clip': 19.269453048706055, 'timing/data': 0.0010617520019877702, 'timing/model': 1.2112607460003346, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 0.72}) +[2026-06-23 20:40:56,951][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.0555763877928257, 'train/grad_norm_pre_clip': 14.946298599243164, 'timing/data': 0.0007710240024607629, 'timing/model': 1.4686553400242701, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 0.72}) +[2026-06-23 20:41:01,322][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.03542565181851387, 'train/grad_norm_pre_clip': 15.797475814819336, 'timing/data': 0.0007310499786399305, 'timing/model': 1.21138091100147, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.72}) +[2026-06-23 20:41:05,657][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.10237813740968704, 'train/grad_norm_pre_clip': 25.81151008605957, 'timing/data': 0.0008859370136633515, 'timing/model': 1.185510700015584, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 0.72}) +[2026-06-23 20:41:10,016][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.012857848778367043, 'train/grad_norm_pre_clip': 7.098416328430176, 'timing/data': 0.0008054680074565113, 'timing/model': 1.1842972690064926, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 0.72}) +[2026-06-23 20:41:14,543][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.03560059145092964, 'train/grad_norm_pre_clip': 28.574560165405273, 'timing/data': 0.0008980199927464128, 'timing/model': 1.1854573030141182, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 0.72}) +[2026-06-23 20:41:18,911][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.008128372952342033, 'train/grad_norm_pre_clip': 25.344663619995117, 'timing/data': 0.0007882459904067218, 'timing/model': 1.1858317160222214, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 0.72}) +[2026-06-23 20:41:23,259][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.21692480146884918, 'train/grad_norm_pre_clip': 14.906720161437988, 'timing/data': 0.0008250840182881802, 'timing/model': 1.1863408479839563, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 0.72}) +[2026-06-23 20:41:27,633][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.036502350121736526, 'train/grad_norm_pre_clip': 16.739017486572266, 'timing/data': 0.0007436939922627062, 'timing/model': 1.1878329159808345, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 0.72}) +[2026-06-23 20:41:32,263][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.20170855522155762, 'train/grad_norm_pre_clip': 30.71051597595215, 'timing/data': 0.0007027489773463458, 'timing/model': 1.446176996018039, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 0.72}) +[2026-06-23 20:41:36,650][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.1135648787021637, 'train/grad_norm_pre_clip': 32.9001579284668, 'timing/data': 0.0013659139804076403, 'timing/model': 1.2006211669940967, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 0.72}) +[2026-06-23 20:41:41,070][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.06971673667430878, 'train/grad_norm_pre_clip': 15.63191032409668, 'timing/data': 0.0008478870149701834, 'timing/model': 1.2111117820022628, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 0.72}) +[2026-06-23 20:41:45,479][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.05066632851958275, 'train/grad_norm_pre_clip': 23.349966049194336, 'timing/data': 0.0007378629816230386, 'timing/model': 1.1988392570056021, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.72}) +[2026-06-23 20:41:50,055][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.20538102090358734, 'train/grad_norm_pre_clip': 25.34427833557129, 'timing/data': 0.0008017719956114888, 'timing/model': 1.2009003559942357, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 0.72}) +[2026-06-23 20:41:54,460][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.13337519764900208, 'train/grad_norm_pre_clip': 13.53366756439209, 'timing/data': 0.0008154659881256521, 'timing/model': 1.2108847159834113, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 0.72}) +[2026-06-23 20:41:58,851][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.05813070759177208, 'train/grad_norm_pre_clip': 15.876471519470215, 'timing/data': 0.0007583019905723631, 'timing/model': 1.2169371799973305, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 0.72}) +[2026-06-23 20:42:03,219][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.09985668957233429, 'train/grad_norm_pre_clip': 33.711219787597656, 'timing/data': 0.0008222089963965118, 'timing/model': 1.188184684026055, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.72}) +[2026-06-23 20:42:07,861][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.21378079056739807, 'train/grad_norm_pre_clip': 49.68842697143555, 'timing/data': 0.0007516490004491061, 'timing/model': 1.4470831500075292, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 0.72}) +[2026-06-23 20:42:12,257][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.06796109676361084, 'train/grad_norm_pre_clip': 26.736770629882812, 'timing/data': 0.0007675880042370409, 'timing/model': 1.2113280890043825, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 0.72}) +[2026-06-23 20:42:16,648][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.02475474588572979, 'train/grad_norm_pre_clip': 7.998260021209717, 'timing/data': 0.0007536219782195985, 'timing/model': 1.2057439420023002, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 0.72}) +[2026-06-23 20:42:21,101][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.049398671835660934, 'train/grad_norm_pre_clip': 25.01644515991211, 'timing/data': 0.0007453670259565115, 'timing/model': 1.2114435840048827, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 0.72}) +[2026-06-23 20:42:25,738][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.10404309630393982, 'train/grad_norm_pre_clip': 25.390514373779297, 'timing/data': 0.0008503909921273589, 'timing/model': 1.460415951005416, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 0.72}) +[2026-06-23 20:42:30,139][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.056113049387931824, 'train/grad_norm_pre_clip': 6.695562839508057, 'timing/data': 0.0007775269914418459, 'timing/model': 1.1984826480038464, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.72}) +[2026-06-23 20:42:34,513][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.06728293001651764, 'train/grad_norm_pre_clip': 14.29774284362793, 'timing/data': 0.0008736839808989316, 'timing/model': 1.2045500300009735, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 0.72}) +[2026-06-23 20:42:38,897][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.020968927070498466, 'train/grad_norm_pre_clip': 12.6643648147583, 'timing/data': 0.0007529910071752965, 'timing/model': 1.2055796940112486, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 0.72}) +[2026-06-23 20:42:43,472][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.16556528210639954, 'train/grad_norm_pre_clip': 21.30341911315918, 'timing/data': 0.000716743990778923, 'timing/model': 1.428757879009936, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 0.72}) +[2026-06-23 20:42:47,794][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.09649297595024109, 'train/grad_norm_pre_clip': 17.64024543762207, 'timing/data': 0.0006867589836474508, 'timing/model': 1.1782827269926202, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 0.73}) +[2026-06-23 20:42:52,121][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.047544579952955246, 'train/grad_norm_pre_clip': 12.615355491638184, 'timing/data': 0.0007097210036590695, 'timing/model': 1.1798336030042265, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 0.73}) +[2026-06-23 20:42:56,446][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.04764757305383682, 'train/grad_norm_pre_clip': 11.838726997375488, 'timing/data': 0.0007511579897254705, 'timing/model': 1.1806355949956924, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 0.73}) +[2026-06-23 20:43:01,016][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.02763497643172741, 'train/grad_norm_pre_clip': 15.713830947875977, 'timing/data': 0.0006997130112722516, 'timing/model': 1.4258095340046566, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 0.73}) +[2026-06-23 20:43:05,354][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.01640697754919529, 'train/grad_norm_pre_clip': 42.470149993896484, 'timing/data': 0.0007128970173653215, 'timing/model': 1.1908056300017051, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 0.73}) +[2026-06-23 20:43:09,715][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.02610052190721035, 'train/grad_norm_pre_clip': 10.55420207977295, 'timing/data': 0.0007476710015907884, 'timing/model': 1.1788740630145185, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 0.73}) +[2026-06-23 20:43:14,128][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.08522216230630875, 'train/grad_norm_pre_clip': 16.202709197998047, 'timing/data': 0.0008017910004127771, 'timing/model': 1.208913411013782, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.73}) +[2026-06-23 20:43:18,735][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.09977147728204727, 'train/grad_norm_pre_clip': 24.007291793823242, 'timing/data': 0.0007464700029231608, 'timing/model': 1.4334438249934465, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 0.73}) +[2026-06-23 20:43:23,066][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.006429257337003946, 'train/grad_norm_pre_clip': 31.060827255249023, 'timing/data': 0.0007889980042818934, 'timing/model': 1.1836354240076616, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 0.73}) +[2026-06-23 20:43:27,422][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.10469479858875275, 'train/grad_norm_pre_clip': 25.368900299072266, 'timing/data': 0.0007572790200356394, 'timing/model': 1.1823773790092673, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 0.73}) +[2026-06-23 20:43:31,799][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.1362157016992569, 'train/grad_norm_pre_clip': 17.198793411254883, 'timing/data': 0.0007873249996919185, 'timing/model': 1.1841470199869946, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 0.73}) +[2026-06-23 20:43:36,433][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.20496316254138947, 'train/grad_norm_pre_clip': 24.47844886779785, 'timing/data': 0.0007378329755738378, 'timing/model': 1.4633243209973443, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.73}) +[2026-06-23 20:43:40,808][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.11935582011938095, 'train/grad_norm_pre_clip': 57.35785675048828, 'timing/data': 0.0009346670121885836, 'timing/model': 1.1986249340116046, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 0.73}) +[2026-06-23 20:43:45,175][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.15844282507896423, 'train/grad_norm_pre_clip': 38.95697784423828, 'timing/data': 0.0008471049950458109, 'timing/model': 1.1817054309940431, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 0.73}) +[2026-06-23 20:43:49,515][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.04160885885357857, 'train/grad_norm_pre_clip': 12.27269172668457, 'timing/data': 0.0007351780077442527, 'timing/model': 1.1787418960011564, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 0.73}) +[2026-06-23 20:43:54,104][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.007594517432153225, 'train/grad_norm_pre_clip': 16.38203239440918, 'timing/data': 0.0008093659998849034, 'timing/model': 1.2173599750094581, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 0.73}) +[2026-06-23 20:43:58,495][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.048635084182024, 'train/grad_norm_pre_clip': 18.004180908203125, 'timing/data': 0.0007446060189977288, 'timing/model': 1.2112447069957852, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.73}) +[2026-06-23 20:44:02,864][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.3115955591201782, 'train/grad_norm_pre_clip': 30.43745994567871, 'timing/data': 0.0007110539881978184, 'timing/model': 1.1922663079749327, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 0.73}) +[2026-06-23 20:44:07,252][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.06619948893785477, 'train/grad_norm_pre_clip': 21.619531631469727, 'timing/data': 0.001137220999225974, 'timing/model': 1.1885044689988717, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 0.73}) +[2026-06-23 20:44:11,858][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.08446706086397171, 'train/grad_norm_pre_clip': 29.749662399291992, 'timing/data': 0.0007289870118256658, 'timing/model': 1.4268707500013988, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 0.73}) +[2026-06-23 20:44:16,243][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.044549040496349335, 'train/grad_norm_pre_clip': 16.328575134277344, 'timing/data': 0.0007928150007501245, 'timing/model': 1.1865851129987277, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 0.73}) +[2026-06-23 20:44:20,581][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.05492144078016281, 'train/grad_norm_pre_clip': 19.27250099182129, 'timing/data': 0.0007177460065577179, 'timing/model': 1.1782922899874393, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 0.73}) +[2026-06-23 20:44:24,956][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.03187226131558418, 'train/grad_norm_pre_clip': 35.049678802490234, 'timing/data': 0.0014634640247095376, 'timing/model': 1.1932649379887152, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 0.73}) +[2026-06-23 20:44:29,592][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.12681682407855988, 'train/grad_norm_pre_clip': 27.054956436157227, 'timing/data': 0.0012154059950262308, 'timing/model': 1.4603925450064708, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 0.73}) +[2026-06-23 20:44:33,927][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.09343583881855011, 'train/grad_norm_pre_clip': 19.31932830810547, 'timing/data': 0.0007632809865754098, 'timing/model': 1.1873559210216627, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 0.73}) +[2026-06-23 20:44:38,274][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.06778636574745178, 'train/grad_norm_pre_clip': 16.597047805786133, 'timing/data': 0.0009185669769067317, 'timing/model': 1.185752338002203, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 0.73}) +[2026-06-23 20:44:42,651][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1900 +[2026-06-23 20:45:40,613][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1900 +[2026-06-23 20:45:40,613][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1900 +[2026-06-23 20:46:17,727][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1900 +[2026-06-23 20:46:17,729][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.13002580404281616, 'train/grad_norm_pre_clip': 9.896649360656738, 'eval/action_loss/seconds': 57.957686943991575, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.0513722263276577, 'eval/latency_2/loss': 0.0513722263276577, 'eval/flappy/loss': 0.0513722263276577, 'eval/flappy/latency_2/loss': 0.0513722263276577, 'eval/flap_precision': 0.91375, 'eval/flap_recall': 0.9918588873812755, 'eval/flap_f1': 0.9512036434612882, 'eval/flap_rate_model': 0.2222222222222222, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9512036434612882, 'eval/latency_2/flap_precision': 0.91375, 'eval/latency_2/flap_recall': 0.9918588873812755, 'eval/latency_2/flap_f1': 0.9512036434612882, 'eval/latency_2/flap_rate_model': 0.2222222222222222, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9512036434612882, 'eval/action_classification/data_wait_seconds': 0.4902823567390442, 'eval/action_classification/predict_seconds': 36.47072982788086, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007711140206083655, 'timing/model': 1.1965610720217228, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.74}) +[2026-06-23 20:46:22,366][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.2028038054704666, 'train/grad_norm_pre_clip': 24.256622314453125, 'timing/data': 0.0008716089942026883, 'timing/model': 1.1955917470040731, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 0.74}) +[2026-06-23 20:46:26,664][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.08435419946908951, 'train/grad_norm_pre_clip': 18.58365249633789, 'timing/data': 0.0012335280189290643, 'timing/model': 1.1671786040242296, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 0.74}) +[2026-06-23 20:46:31,041][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.07704221457242966, 'train/grad_norm_pre_clip': 16.67540740966797, 'timing/data': 0.0011291450064163655, 'timing/model': 1.2099515970039647, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 0.74}) +[2026-06-23 20:46:35,570][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.04564359411597252, 'train/grad_norm_pre_clip': 16.924367904663086, 'timing/data': 0.0012297409994062036, 'timing/model': 1.3794686640030704, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 0.74}) +[2026-06-23 20:46:39,902][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.03084951639175415, 'train/grad_norm_pre_clip': 6.792140483856201, 'timing/data': 0.0007700519927311689, 'timing/model': 1.1710784369788598, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 0.74}) +[2026-06-23 20:46:44,221][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.1498713493347168, 'train/grad_norm_pre_clip': 22.102115631103516, 'timing/data': 0.0008191819943021983, 'timing/model': 1.1745062510017306, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 0.74}) +[2026-06-23 20:46:48,534][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.12810154259204865, 'train/grad_norm_pre_clip': 18.661096572875977, 'timing/data': 0.0007702620059717447, 'timing/model': 1.168187357019633, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 0.74}) +[2026-06-23 20:46:53,066][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.19137631356716156, 'train/grad_norm_pre_clip': 27.35276222229004, 'timing/data': 0.0007284639868885279, 'timing/model': 1.4165126719744876, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 0.74}) +[2026-06-23 20:46:57,400][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.08387955278158188, 'train/grad_norm_pre_clip': 24.18050193786621, 'timing/data': 0.0009981229959521443, 'timing/model': 1.192814053996699, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 0.74}) +[2026-06-23 20:47:01,769][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.03358841687440872, 'train/grad_norm_pre_clip': 17.62577247619629, 'timing/data': 0.0009215619938913733, 'timing/model': 1.173734420008259, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.74}) +[2026-06-23 20:47:06,114][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.013179280795156956, 'train/grad_norm_pre_clip': 14.813332557678223, 'timing/data': 0.0009204100060742348, 'timing/model': 1.199858656997094, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 0.74}) +[2026-06-23 20:47:10,584][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.19668348133563995, 'train/grad_norm_pre_clip': 25.42758560180664, 'timing/data': 0.0010226880258414894, 'timing/model': 1.1759599249926396, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 0.74}) +[2026-06-23 20:47:14,929][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.048015765845775604, 'train/grad_norm_pre_clip': 46.01219940185547, 'timing/data': 0.0008086829911917448, 'timing/model': 1.2046644980146084, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.74}) +[2026-06-23 20:47:19,260][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.007703156676143408, 'train/grad_norm_pre_clip': 33.13629913330078, 'timing/data': 0.0007610750035382807, 'timing/model': 1.1746937140123919, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 0.74}) +[2026-06-23 20:47:23,608][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.05920970439910889, 'train/grad_norm_pre_clip': 15.824480056762695, 'timing/data': 0.0007378619920928031, 'timing/model': 1.1964095020084642, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 0.74}) +[2026-06-23 20:47:28,066][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.11097980290651321, 'train/grad_norm_pre_clip': 36.760231018066406, 'timing/data': 0.0008835109765641391, 'timing/model': 1.174409253988415, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 0.74}) +[2026-06-23 20:47:32,424][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.028086574748158455, 'train/grad_norm_pre_clip': 15.842327117919922, 'timing/data': 0.0012318340013734996, 'timing/model': 1.185720802983269, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 0.74}) +[2026-06-23 20:47:36,816][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.09209983795881271, 'train/grad_norm_pre_clip': 15.303394317626953, 'timing/data': 0.00076498300768435, 'timing/model': 1.1987818719935603, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 0.74}) +[2026-06-23 20:47:41,186][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.016449347138404846, 'train/grad_norm_pre_clip': 22.73777961730957, 'timing/data': 0.0012289000151213259, 'timing/model': 1.1822756670007948, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 0.74}) +[2026-06-23 20:47:45,793][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.0323113389313221, 'train/grad_norm_pre_clip': 13.915719985961914, 'timing/data': 0.0009376920061185956, 'timing/model': 1.177720148989465, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.74}) +[2026-06-23 20:47:50,175][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.027968207374215126, 'train/grad_norm_pre_clip': 11.265618324279785, 'timing/data': 0.0007591219909954816, 'timing/model': 1.197195410990389, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 0.74}) +[2026-06-23 20:47:54,537][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.10798551142215729, 'train/grad_norm_pre_clip': 24.058208465576172, 'timing/data': 0.0007872239802964032, 'timing/model': 1.179197530989768, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 0.74}) +[2026-06-23 20:47:58,910][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.10584837198257446, 'train/grad_norm_pre_clip': 13.473329544067383, 'timing/data': 0.0007207010057754815, 'timing/model': 1.1846143850125372, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 0.74}) +[2026-06-23 20:48:03,457][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.12331046909093857, 'train/grad_norm_pre_clip': 29.181257247924805, 'timing/data': 0.0007334540132433176, 'timing/model': 1.1783104090136476, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 0.74}) +[2026-06-23 20:48:07,857][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.18984799087047577, 'train/grad_norm_pre_clip': 40.6708984375, 'timing/data': 0.0007231849886011332, 'timing/model': 1.200483772001462, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 0.75}) +[2026-06-23 20:48:12,223][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.18776661157608032, 'train/grad_norm_pre_clip': 52.73200607299805, 'timing/data': 0.0007364490011241287, 'timing/model': 1.1976401740103029, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 0.75}) +[2026-06-23 20:48:16,608][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.28262123465538025, 'train/grad_norm_pre_clip': 48.89435577392578, 'timing/data': 0.0007364190241787583, 'timing/model': 1.1991931429947726, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 0.75}) +[2026-06-23 20:48:21,198][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.17287592589855194, 'train/grad_norm_pre_clip': 44.15568923950195, 'timing/data': 0.001231784000992775, 'timing/model': 1.176858966005966, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 0.75}) +[2026-06-23 20:48:25,536][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.1940317302942276, 'train/grad_norm_pre_clip': 25.762269973754883, 'timing/data': 0.0007727970078121871, 'timing/model': 1.1780353600042872, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 0.75}) +[2026-06-23 20:48:29,918][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.009084395132958889, 'train/grad_norm_pre_clip': 21.07313346862793, 'timing/data': 0.0007456770108547062, 'timing/model': 1.2017606610024814, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.75}) +[2026-06-23 20:48:34,331][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.05061398819088936, 'train/grad_norm_pre_clip': 16.774066925048828, 'timing/data': 0.000956515985308215, 'timing/model': 1.2002304370107595, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 0.75}) +[2026-06-23 20:48:38,856][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.06596337258815765, 'train/grad_norm_pre_clip': 31.896564483642578, 'timing/data': 0.0007257300021592528, 'timing/model': 1.1752694460155908, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 0.75}) +[2026-06-23 20:48:43,175][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.09000320732593536, 'train/grad_norm_pre_clip': 23.566341400146484, 'timing/data': 0.000816095998743549, 'timing/model': 1.1791196870035492, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 0.75}) +[2026-06-23 20:48:47,510][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.01790701597929001, 'train/grad_norm_pre_clip': 10.60802173614502, 'timing/data': 0.0008738540054764599, 'timing/model': 1.1761582499893848, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 0.75}) +[2026-06-23 20:48:51,871][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.10967562347650528, 'train/grad_norm_pre_clip': 13.105484008789062, 'timing/data': 0.0007517379999626428, 'timing/model': 1.1948031100037042, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 0.75}) +[2026-06-23 20:48:56,390][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.026000941172242165, 'train/grad_norm_pre_clip': 12.919198036193848, 'timing/data': 0.0007314499816857278, 'timing/model': 1.176583035994554, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 0.75}) +[2026-06-23 20:49:00,792][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.0008516490925103426, 'train/grad_norm_pre_clip': 28.27131462097168, 'timing/data': 0.0008133719966281205, 'timing/model': 1.205464850005228, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 0.75}) +[2026-06-23 20:49:05,168][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.05887418985366821, 'train/grad_norm_pre_clip': 14.558127403259277, 'timing/data': 0.0007680580019950867, 'timing/model': 1.179984399001114, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 0.75}) +[2026-06-23 20:49:09,509][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.04885602369904518, 'train/grad_norm_pre_clip': 23.50641441345215, 'timing/data': 0.0008319759799633175, 'timing/model': 1.1913548150041606, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 0.75}) +[2026-06-23 20:49:14,034][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.10673246532678604, 'train/grad_norm_pre_clip': 12.969347953796387, 'timing/data': 0.0007800099847372621, 'timing/model': 1.1814861320017371, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.75}) +[2026-06-23 20:49:18,367][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.03330320119857788, 'train/grad_norm_pre_clip': 15.085986137390137, 'timing/data': 0.0011049000022467226, 'timing/model': 1.1839129840082023, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 0.75}) +[2026-06-23 20:49:22,704][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.21001741290092468, 'train/grad_norm_pre_clip': 35.91362762451172, 'timing/data': 0.0007937659975141287, 'timing/model': 1.182550469005946, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 0.75}) +[2026-06-23 20:49:27,083][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.07662879675626755, 'train/grad_norm_pre_clip': 13.293420791625977, 'timing/data': 0.000784597999881953, 'timing/model': 1.204990007012384, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 0.75}) +[2026-06-23 20:49:31,677][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.16082863509655, 'train/grad_norm_pre_clip': 43.70292663574219, 'timing/data': 0.0009147589735221118, 'timing/model': 1.2102099980111234, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 0.75}) +[2026-06-23 20:49:36,060][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.1566184163093567, 'train/grad_norm_pre_clip': 26.165443420410156, 'timing/data': 0.0010219979740213603, 'timing/model': 1.210678111994639, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 0.75}) +[2026-06-23 20:49:40,407][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.07843612134456635, 'train/grad_norm_pre_clip': 8.301473617553711, 'timing/data': 0.0007915309979580343, 'timing/model': 1.1792715949995909, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 0.75}) +[2026-06-23 20:49:44,776][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.06020258739590645, 'train/grad_norm_pre_clip': 17.439605712890625, 'timing/data': 0.0007978230132721364, 'timing/model': 1.2016811280045658, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 0.75}) +[2026-06-23 20:49:49,383][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.0877825915813446, 'train/grad_norm_pre_clip': 21.853548049926758, 'timing/data': 0.0011638589785434306, 'timing/model': 1.2065910059900489, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 0.75}) +[2026-06-23 20:49:53,755][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.03760372847318649, 'train/grad_norm_pre_clip': 17.041719436645508, 'timing/data': 0.0007647229940630496, 'timing/model': 1.2013354600057937, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 0.75}) +[2026-06-23 20:49:58,138][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.17535842955112457, 'train/grad_norm_pre_clip': 25.57284927368164, 'timing/data': 0.00111629199818708, 'timing/model': 1.2002557060041, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.75}) +[2026-06-23 20:50:02,513][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.0732445940375328, 'train/grad_norm_pre_clip': 14.585005760192871, 'timing/data': 0.0008109170012176037, 'timing/model': 1.2019232349994127, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 0.76}) +[2026-06-23 20:50:07,075][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.08024633675813675, 'train/grad_norm_pre_clip': 55.99311447143555, 'timing/data': 0.000834030011901632, 'timing/model': 1.1841609539987985, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 0.76}) +[2026-06-23 20:50:11,463][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.0906275063753128, 'train/grad_norm_pre_clip': 21.939533233642578, 'timing/data': 0.0010122400126419961, 'timing/model': 1.1884673150198068, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 0.76}) +[2026-06-23 20:50:15,830][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.03995000571012497, 'train/grad_norm_pre_clip': 20.649003982543945, 'timing/data': 0.0007881349883973598, 'timing/model': 1.1839838650193997, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 0.76}) +[2026-06-23 20:50:20,190][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.04429515451192856, 'train/grad_norm_pre_clip': 18.108163833618164, 'timing/data': 0.0007456570165231824, 'timing/model': 1.1870035309984814, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.76}) +[2026-06-23 20:50:24,785][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.03328753635287285, 'train/grad_norm_pre_clip': 18.842226028442383, 'timing/data': 0.0009181349887512624, 'timing/model': 1.1864043530076742, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 0.76}) +[2026-06-23 20:50:29,123][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.014973741956055164, 'train/grad_norm_pre_clip': 31.43839454650879, 'timing/data': 0.0007780359883327037, 'timing/model': 1.186165731982328, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 0.76}) +[2026-06-23 20:50:33,475][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.08825485408306122, 'train/grad_norm_pre_clip': 17.68943214416504, 'timing/data': 0.0007731680234428495, 'timing/model': 1.183573994989274, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 0.76}) +[2026-06-23 20:50:37,860][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.09884439408779144, 'train/grad_norm_pre_clip': 18.31348991394043, 'timing/data': 0.0007840580074116588, 'timing/model': 1.206177106971154, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 0.76}) +[2026-06-23 20:50:42,423][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.04256478324532509, 'train/grad_norm_pre_clip': 12.36823844909668, 'timing/data': 0.0007835369906388223, 'timing/model': 1.184212171996478, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.76}) +[2026-06-23 20:50:46,849][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.12093175947666168, 'train/grad_norm_pre_clip': 21.121713638305664, 'timing/data': 0.0008510820043738931, 'timing/model': 1.2181032200169284, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 0.76}) +[2026-06-23 20:50:51,266][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.1285122036933899, 'train/grad_norm_pre_clip': 14.427825927734375, 'timing/data': 0.000810157012892887, 'timing/model': 1.2033312380081043, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.76}) +[2026-06-23 20:50:55,612][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.03101578913629055, 'train/grad_norm_pre_clip': 8.517391204833984, 'timing/data': 0.0007485529931727797, 'timing/model': 1.187355732981814, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 0.76}) +[2026-06-23 20:51:00,211][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.005893896799534559, 'train/grad_norm_pre_clip': 9.495841979980469, 'timing/data': 0.0007709839846938848, 'timing/model': 1.2005762280023191, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 0.76}) +[2026-06-23 20:51:04,553][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.044579628854990005, 'train/grad_norm_pre_clip': 14.414680480957031, 'timing/data': 0.0007638699898961931, 'timing/model': 1.1882897560135461, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 0.76}) +[2026-06-23 20:51:08,950][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.01978607475757599, 'train/grad_norm_pre_clip': 14.499435424804688, 'timing/data': 0.0009481410088483244, 'timing/model': 1.189483908994589, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 0.76}) +[2026-06-23 20:51:13,362][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.04055733606219292, 'train/grad_norm_pre_clip': 40.13121032714844, 'timing/data': 0.0007425009971484542, 'timing/model': 1.2175703090033494, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 0.76}) +[2026-06-23 20:51:17,946][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.10639694333076477, 'train/grad_norm_pre_clip': 16.5542049407959, 'timing/data': 0.0008241119794547558, 'timing/model': 1.2234732299984898, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 0.76}) +[2026-06-23 20:51:22,353][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.0322350449860096, 'train/grad_norm_pre_clip': 22.830183029174805, 'timing/data': 0.0008260259928647429, 'timing/model': 1.1911541000008583, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 0.76}) +[2026-06-23 20:51:26,754][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.28624460101127625, 'train/grad_norm_pre_clip': 39.31354522705078, 'timing/data': 0.0010908550175372511, 'timing/model': 1.2088870390143711, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.76}) +[2026-06-23 20:51:31,139][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.23222887516021729, 'train/grad_norm_pre_clip': 50.47673416137695, 'timing/data': 0.0007320919830817729, 'timing/model': 1.2109375440049917, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 0.76}) +[2026-06-23 20:51:35,674][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.025010937824845314, 'train/grad_norm_pre_clip': 16.105876922607422, 'timing/data': 0.001151866978034377, 'timing/model': 1.1813027499883901, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 0.76}) +[2026-06-23 20:51:40,036][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.07519960403442383, 'train/grad_norm_pre_clip': 19.509355545043945, 'timing/data': 0.0007148000004235655, 'timing/model': 1.2131378230114933, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 0.76}) +[2026-06-23 20:51:44,417][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.07537566125392914, 'train/grad_norm_pre_clip': 45.28596878051758, 'timing/data': 0.0007220029947347939, 'timing/model': 1.2145320159906987, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 0.76}) +[2026-06-23 20:51:48,792][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.006378992926329374, 'train/grad_norm_pre_clip': 32.02002716064453, 'timing/data': 0.0007278150005731732, 'timing/model': 1.1826911579992156, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 0.76}) +[2026-06-23 20:51:53,389][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.12268917262554169, 'train/grad_norm_pre_clip': 21.524263381958008, 'timing/data': 0.000923935993341729, 'timing/model': 1.2021959589910693, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 0.76}) +[2026-06-23 20:51:57,753][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.06830257177352905, 'train/grad_norm_pre_clip': 14.05671215057373, 'timing/data': 0.0007351769891101867, 'timing/model': 1.1835453820240218, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 0.77}) +[2026-06-23 20:52:02,134][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.03160586208105087, 'train/grad_norm_pre_clip': 11.242180824279785, 'timing/data': 0.001323253003647551, 'timing/model': 1.2002654789830558, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 0.77}) +[2026-06-23 20:52:06,493][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.11208601295948029, 'train/grad_norm_pre_clip': 30.54549217224121, 'timing/data': 0.0007107520068529993, 'timing/model': 1.185185475012986, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 0.77}) +[2026-06-23 20:52:11,030][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.1098027303814888, 'train/grad_norm_pre_clip': 20.769132614135742, 'timing/data': 0.0008891519973985851, 'timing/model': 1.1831062130222563, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.77}) +[2026-06-23 20:52:15,395][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.0022024752106517553, 'train/grad_norm_pre_clip': 24.54238510131836, 'timing/data': 0.0007363100012298673, 'timing/model': 1.1841389579931274, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 0.77}) +[2026-06-23 20:52:19,759][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.08186689764261246, 'train/grad_norm_pre_clip': 10.174673080444336, 'timing/data': 0.0012093039986211807, 'timing/model': 1.1790898199833464, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 0.77}) +[2026-06-23 20:52:24,105][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.0221906416118145, 'train/grad_norm_pre_clip': 11.728978157043457, 'timing/data': 0.0007450059929396957, 'timing/model': 1.1999093900085427, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 0.77}) +[2026-06-23 20:52:28,715][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.0057721869088709354, 'train/grad_norm_pre_clip': 30.945337295532227, 'timing/data': 0.0007386239885818213, 'timing/model': 1.20668238599319, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 0.77}) +[2026-06-23 20:52:33,094][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.006262202747166157, 'train/grad_norm_pre_clip': 18.491046905517578, 'timing/data': 0.0007562160026282072, 'timing/model': 1.1862925240129698, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 0.77}) +[2026-06-23 20:52:37,475][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.03368426486849785, 'train/grad_norm_pre_clip': 15.2562837600708, 'timing/data': 0.0008966059831436723, 'timing/model': 1.2059851629892364, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 0.77}) +[2026-06-23 20:52:41,893][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.039144400507211685, 'train/grad_norm_pre_clip': 27.68631935119629, 'timing/data': 0.001199315011035651, 'timing/model': 1.2142501460039057, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 0.77}) +[2026-06-23 20:52:46,484][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.017892220988869667, 'train/grad_norm_pre_clip': 32.55342102050781, 'timing/data': 0.000726491998648271, 'timing/model': 1.2076922810229007, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 0.77}) +[2026-06-23 20:52:50,889][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.05831008404493332, 'train/grad_norm_pre_clip': 68.7921142578125, 'timing/data': 0.0010861869959626347, 'timing/model': 1.2115244260057807, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 0.77}) +[2026-06-23 20:52:55,306][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.007623618934303522, 'train/grad_norm_pre_clip': 34.94654846191406, 'timing/data': 0.0007656350207980722, 'timing/model': 1.2124955769977532, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.77}) +[2026-06-23 20:52:59,685][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.02096983790397644, 'train/grad_norm_pre_clip': 14.894975662231445, 'timing/data': 0.0008652179967612028, 'timing/model': 1.2017634010117035, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 0.77}) +[2026-06-23 20:53:04,261][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.0014134717639535666, 'train/grad_norm_pre_clip': 16.944080352783203, 'timing/data': 0.0007575290219392627, 'timing/model': 1.2117360070114955, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 0.77}) +[2026-06-23 20:53:08,638][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.08947916328907013, 'train/grad_norm_pre_clip': 18.778547286987305, 'timing/data': 0.0007960000075399876, 'timing/model': 1.2026957570051309, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 0.77}) +[2026-06-23 20:53:13,001][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.07827621698379517, 'train/grad_norm_pre_clip': 19.226085662841797, 'timing/data': 0.0009955790010280907, 'timing/model': 1.1846787059912458, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 0.77}) +[2026-06-23 20:53:17,407][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.0015132330590859056, 'train/grad_norm_pre_clip': 18.886096954345703, 'timing/data': 0.000779068999690935, 'timing/model': 1.2043025190068875, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 0.77}) +[2026-06-23 20:53:21,991][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.07329456508159637, 'train/grad_norm_pre_clip': 27.983861923217773, 'timing/data': 0.0009500040032435209, 'timing/model': 1.2057075600023381, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 0.77}) +[2026-06-23 20:53:26,387][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.07331260293722153, 'train/grad_norm_pre_clip': 10.895442008972168, 'timing/data': 0.0009052520035766065, 'timing/model': 1.211314993997803, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 0.77}) +[2026-06-23 20:53:30,802][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.057297199964523315, 'train/grad_norm_pre_clip': 12.827778816223145, 'timing/data': 0.0007552250172011554, 'timing/model': 1.2039145010057837, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 0.77}) +[2026-06-23 20:53:35,205][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.018902797251939774, 'train/grad_norm_pre_clip': 12.589068412780762, 'timing/data': 0.0007975639891810715, 'timing/model': 1.2000010199844837, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 0.77}) +[2026-06-23 20:53:39,839][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2000 +[2026-06-23 20:54:37,248][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2000 +[2026-06-23 20:54:37,249][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2000 +[2026-06-23 20:55:14,097][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2000 +[2026-06-23 21:23:36,961][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-06-23 21:23:36,963][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.02421393245458603, 'train/grad_norm_pre_clip': 8.96461009979248, 'eval/action_loss/seconds': 57.40495763198123, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.037003885954618454, 'eval/latency_2/loss': 0.037003885954618454, 'eval/flappy/loss': 0.037003885954618454, 'eval/flappy/latency_2/loss': 0.037003885954618454, 'eval/flap_precision': 0.9486166007905138, 'eval/flap_recall': 0.9769335142469471, 'eval/flap_f1': 0.9625668449197862, 'eval/flap_rate_model': 0.21083333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9625668449197863, 'eval/latency_2/flap_precision': 0.9486166007905138, 'eval/latency_2/flap_recall': 0.9769335142469471, 'eval/latency_2/flap_f1': 0.9625668449197862, 'eval/latency_2/flap_rate_model': 0.21083333333333334, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9625668449197863, 'eval/action_classification/data_wait_seconds': 0.44476625323295593, 'eval/action_classification/predict_seconds': 36.24930953979492, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0009017859993036836, 'timing/model': 1.2391279559815302, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.77}) +[2026-06-23 21:23:36,964][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_2000_state +[2026-06-23 21:23:36,970][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 21:23:45,781][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_2000_state/model.safetensors +[2026-06-23 21:23:56,179][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_2000_state/optimizer.bin +[2026-06-23 21:23:56,179][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_2000_state/sampler.bin +[2026-06-23 21:23:56,179][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-23 21:23:56,181][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-23 21:23:56,379][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-23 21:23:56,380][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 21:23:56,405][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 21:24:00,788][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.08932303637266159, 'train/grad_norm_pre_clip': 24.69186019897461, 'timing/data': 0.0009220429928973317, 'timing/model': 1.1622620479902253, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 0.77}) +[2026-06-23 21:24:05,237][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.12393134087324142, 'train/grad_norm_pre_clip': 21.40703773498535, 'timing/data': 0.0009412080107722431, 'timing/model': 1.393719592015259, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 0.77}) +[2026-06-23 21:24:09,504][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.047275543212890625, 'train/grad_norm_pre_clip': 29.130502700805664, 'timing/data': 0.0008091660274658352, 'timing/model': 1.1698482260107994, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 0.78}) +[2026-06-23 21:24:13,745][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.022553039714694023, 'train/grad_norm_pre_clip': 15.428141593933105, 'timing/data': 0.0009224439854733646, 'timing/model': 1.163496272987686, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 0.78}) +[2026-06-23 21:24:17,993][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.16813281178474426, 'train/grad_norm_pre_clip': 53.6149787902832, 'timing/data': 0.001198754005599767, 'timing/model': 1.1618948350078426, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 0.78}) +[2026-06-23 21:24:22,391][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.07165966182947159, 'train/grad_norm_pre_clip': 21.51095962524414, 'timing/data': 0.0007694920059293509, 'timing/model': 1.162517918011872, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 0.78}) +[2026-06-23 21:24:26,578][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.0040982551872730255, 'train/grad_norm_pre_clip': 6.097590446472168, 'timing/data': 0.0007969120051711798, 'timing/model': 1.146155369002372, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 0.78}) +[2026-06-23 21:24:30,831][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.05448506027460098, 'train/grad_norm_pre_clip': 19.241727828979492, 'timing/data': 0.0010421459737699479, 'timing/model': 1.1624230449961033, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 0.78}) +[2026-06-23 21:24:35,022][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.05961661785840988, 'train/grad_norm_pre_clip': 11.709986686706543, 'timing/data': 0.0007276739925146103, 'timing/model': 1.1460295500000939, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 0.78}) +[2026-06-23 21:24:39,516][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.03401922807097435, 'train/grad_norm_pre_clip': 30.868181228637695, 'timing/data': 0.0007551549933850765, 'timing/model': 1.405629229004262, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.78}) +[2026-06-23 21:24:43,745][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.158502995967865, 'train/grad_norm_pre_clip': 85.33937072753906, 'timing/data': 0.0007215929799713194, 'timing/model': 1.1475710099912249, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.78}) +[2026-06-23 21:24:47,971][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.16010822355747223, 'train/grad_norm_pre_clip': 92.43568420410156, 'timing/data': 0.0009746099822223186, 'timing/model': 1.168258721998427, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 0.78}) +[2026-06-23 21:24:52,242][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.009454659186303616, 'train/grad_norm_pre_clip': 19.811800003051758, 'timing/data': 0.0008909360039979219, 'timing/model': 1.152952564007137, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 0.78}) +[2026-06-23 21:24:56,751][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.02284442074596882, 'train/grad_norm_pre_clip': 14.402581214904785, 'timing/data': 0.0008070710173342377, 'timing/model': 1.411764149001101, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 0.78}) +[2026-06-23 21:25:00,981][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.14835315942764282, 'train/grad_norm_pre_clip': 34.347774505615234, 'timing/data': 0.0007570890011265874, 'timing/model': 1.1563078369945288, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 0.78}) +[2026-06-23 21:25:05,295][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.00764861423522234, 'train/grad_norm_pre_clip': 19.493877410888672, 'timing/data': 0.0011031480098608881, 'timing/model': 1.1933745790156536, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 0.78}) +[2026-06-23 21:25:09,605][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.04167158529162407, 'train/grad_norm_pre_clip': 25.559919357299805, 'timing/data': 0.000892398995347321, 'timing/model': 1.1908852710039355, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 0.78}) +[2026-06-23 21:25:14,166][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.04040401428937912, 'train/grad_norm_pre_clip': 30.532115936279297, 'timing/data': 0.0007631090120412409, 'timing/model': 1.4487993189832196, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 0.78}) +[2026-06-23 21:25:18,476][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.038747869431972504, 'train/grad_norm_pre_clip': 61.9206428527832, 'timing/data': 0.0009189380216412246, 'timing/model': 1.1779859820089769, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 0.78}) +[2026-06-23 21:25:22,820][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.10226844996213913, 'train/grad_norm_pre_clip': 38.21709060668945, 'timing/data': 0.0009301980026066303, 'timing/model': 1.1931547779822722, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.78}) +[2026-06-23 21:25:27,115][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.20129752159118652, 'train/grad_norm_pre_clip': 34.198402404785156, 'timing/data': 0.0007271630165632814, 'timing/model': 1.1878612689906731, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 0.78}) +[2026-06-23 21:25:31,646][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.002533843507990241, 'train/grad_norm_pre_clip': 14.959108352661133, 'timing/data': 0.0008394309843424708, 'timing/model': 1.1871860029932577, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 0.78}) +[2026-06-23 21:25:35,975][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.025164542719721794, 'train/grad_norm_pre_clip': 10.819748878479004, 'timing/data': 0.0010073309822473675, 'timing/model': 1.1698608960141428, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 0.78}) +[2026-06-23 21:25:40,305][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.047957971692085266, 'train/grad_norm_pre_clip': 20.675899505615234, 'timing/data': 0.000893782009370625, 'timing/model': 1.169017898995662, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 0.78}) +[2026-06-23 21:25:44,669][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.10291732847690582, 'train/grad_norm_pre_clip': 16.65419578552246, 'timing/data': 0.0007170939934439957, 'timing/model': 1.1912677770014852, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.78}) +[2026-06-23 21:25:49,199][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.06288700550794601, 'train/grad_norm_pre_clip': 19.778757095336914, 'timing/data': 0.0012006489851046354, 'timing/model': 1.1780795240192674, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 0.78}) +[2026-06-23 21:25:53,510][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.027973566204309464, 'train/grad_norm_pre_clip': 15.936588287353516, 'timing/data': 0.0007355389825534075, 'timing/model': 1.168883033999009, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 0.78}) +[2026-06-23 21:25:57,796][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.12181814014911652, 'train/grad_norm_pre_clip': 26.753263473510742, 'timing/data': 0.0008038750092964619, 'timing/model': 1.169852977996925, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 0.78}) +[2026-06-23 21:26:02,144][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.014992066659033298, 'train/grad_norm_pre_clip': 17.302988052368164, 'timing/data': 0.0008074110082816333, 'timing/model': 1.1821455690078437, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 0.79}) +[2026-06-23 21:26:06,732][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.02196938544511795, 'train/grad_norm_pre_clip': 13.581737518310547, 'timing/data': 0.000974188995314762, 'timing/model': 1.4421719720121473, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.79}) +[2026-06-23 21:26:11,062][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.029890645295381546, 'train/grad_norm_pre_clip': 21.753883361816406, 'timing/data': 0.0007782879984006286, 'timing/model': 1.1918368059850764, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 0.79}) +[2026-06-23 21:26:15,388][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.051210783421993256, 'train/grad_norm_pre_clip': 18.215656280517578, 'timing/data': 0.0009094399865716696, 'timing/model': 1.171555848006392, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 0.79}) +[2026-06-23 21:26:19,727][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.033799052238464355, 'train/grad_norm_pre_clip': 17.283329010009766, 'timing/data': 0.0011749710247386247, 'timing/model': 1.1922312770038843, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 0.79}) +[2026-06-23 21:26:24,311][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.05601970851421356, 'train/grad_norm_pre_clip': 21.201475143432617, 'timing/data': 0.0007677490066271275, 'timing/model': 1.4290042009961326, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 0.79}) +[2026-06-23 21:26:28,651][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.05946841835975647, 'train/grad_norm_pre_clip': 32.901641845703125, 'timing/data': 0.0007266229949891567, 'timing/model': 1.1932343230000697, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 0.79}) +[2026-06-23 21:26:33,014][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.08810451626777649, 'train/grad_norm_pre_clip': 17.058080673217773, 'timing/data': 0.0008060380059760064, 'timing/model': 1.1977115260087885, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 0.79}) +[2026-06-23 21:26:37,394][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.01963246427476406, 'train/grad_norm_pre_clip': 13.650445938110352, 'timing/data': 0.0011526300222612917, 'timing/model': 1.2041590489970986, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 0.79}) +[2026-06-23 21:26:41,908][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.11213170737028122, 'train/grad_norm_pre_clip': 30.373035430908203, 'timing/data': 0.0010016199958045036, 'timing/model': 1.1965416059829295, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 0.79}) +[2026-06-23 21:26:46,237][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.03744930773973465, 'train/grad_norm_pre_clip': 30.45716667175293, 'timing/data': 0.00079008998000063, 'timing/model': 1.2046585640055127, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 0.79}) +[2026-06-23 21:26:50,562][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.08504699915647507, 'train/grad_norm_pre_clip': 34.676734924316406, 'timing/data': 0.0007856820011511445, 'timing/model': 1.180282338988036, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.79}) +[2026-06-23 21:26:54,931][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.017136260867118835, 'train/grad_norm_pre_clip': 16.80272674560547, 'timing/data': 0.0011214420082978904, 'timing/model': 1.1950306490180083, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 0.79}) +[2026-06-23 21:26:59,523][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.06980127841234207, 'train/grad_norm_pre_clip': 15.092545509338379, 'timing/data': 0.0007577200012747198, 'timing/model': 1.446592596010305, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 0.79}) +[2026-06-23 21:27:03,901][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.06630384176969528, 'train/grad_norm_pre_clip': 14.702391624450684, 'timing/data': 0.0009266120032407343, 'timing/model': 1.201778115995694, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 0.79}) +[2026-06-23 21:27:08,216][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.11240752786397934, 'train/grad_norm_pre_clip': 22.03587532043457, 'timing/data': 0.0010174790222663432, 'timing/model': 1.179677134990925, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 0.79}) +[2026-06-23 21:27:12,560][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.09986571222543716, 'train/grad_norm_pre_clip': 12.17270278930664, 'timing/data': 0.001361375005217269, 'timing/model': 1.2059265629795846, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 0.79}) +[2026-06-23 21:27:17,211][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.07535015046596527, 'train/grad_norm_pre_clip': 12.925509452819824, 'timing/data': 0.0008003980037756264, 'timing/model': 1.468427140993299, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 0.79}) +[2026-06-23 21:27:21,587][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.14869531989097595, 'train/grad_norm_pre_clip': 16.100727081298828, 'timing/data': 0.0008270679973065853, 'timing/model': 1.1855052459868602, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 0.79}) +[2026-06-23 21:27:26,220][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.127363383769989, 'train/grad_norm_pre_clip': 15.850122451782227, 'timing/data': 0.0009818940015975386, 'timing/model': 1.4875240770052187, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 0.79}) +[2026-06-23 21:27:30,579][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.08568371087312698, 'train/grad_norm_pre_clip': 12.927651405334473, 'timing/data': 0.000736059999326244, 'timing/model': 1.192652072000783, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 0.79}) +[2026-06-23 21:27:35,145][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.13886740803718567, 'train/grad_norm_pre_clip': 30.001192092895508, 'timing/data': 0.0007446860254276544, 'timing/model': 1.1759487540111877, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.79}) +[2026-06-23 21:27:39,484][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.08106046915054321, 'train/grad_norm_pre_clip': 13.006909370422363, 'timing/data': 0.0007909010164439678, 'timing/model': 1.1982120569737162, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 0.79}) +[2026-06-23 21:27:43,877][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.058928925544023514, 'train/grad_norm_pre_clip': 14.963314056396484, 'timing/data': 0.0007487029943149537, 'timing/model': 1.2021746190148406, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 0.79}) +[2026-06-23 21:27:48,257][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.04472891613841057, 'train/grad_norm_pre_clip': 16.536231994628906, 'timing/data': 0.000800548994448036, 'timing/model': 1.207340202992782, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 0.79}) +[2026-06-23 21:27:52,772][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.03354717046022415, 'train/grad_norm_pre_clip': 13.715242385864258, 'timing/data': 0.0007648229948244989, 'timing/model': 1.1816066699975636, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 0.79}) +[2026-06-23 21:27:57,147][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.1873030811548233, 'train/grad_norm_pre_clip': 29.64091682434082, 'timing/data': 0.0007510979776270688, 'timing/model': 1.2023947080015205, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 0.8}) +[2026-06-23 21:28:01,518][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.051802098751068115, 'train/grad_norm_pre_clip': 16.560272216796875, 'timing/data': 0.0007324239995796233, 'timing/model': 1.2044583319802769, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 0.8}) +[2026-06-23 21:28:05,880][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.08576303720474243, 'train/grad_norm_pre_clip': 15.561984062194824, 'timing/data': 0.0007561169913969934, 'timing/model': 1.2081648310122546, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 0.8}) +[2026-06-23 21:28:10,476][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.11522432416677475, 'train/grad_norm_pre_clip': 10.202289581298828, 'timing/data': 0.0007853209972381592, 'timing/model': 1.1780581079947297, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 0.8}) +[2026-06-23 21:28:14,807][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.05421577766537666, 'train/grad_norm_pre_clip': 19.551549911499023, 'timing/data': 0.0010162980179302394, 'timing/model': 1.1905119080038276, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 0.8}) +[2026-06-23 21:28:19,177][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.07127545773983002, 'train/grad_norm_pre_clip': 22.454051971435547, 'timing/data': 0.000966496008913964, 'timing/model': 1.2092408320168033, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.8}) +[2026-06-23 21:28:23,530][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.04430652782320976, 'train/grad_norm_pre_clip': 9.673724174499512, 'timing/data': 0.0008363550005014986, 'timing/model': 1.179464609012939, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 0.8}) +[2026-06-23 21:28:28,142][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.024921707808971405, 'train/grad_norm_pre_clip': 43.64220428466797, 'timing/data': 0.001512584974989295, 'timing/model': 1.1999429349962156, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 0.8}) +[2026-06-23 21:28:32,499][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.11139757931232452, 'train/grad_norm_pre_clip': 18.85565757751465, 'timing/data': 0.000782354996772483, 'timing/model': 1.180973709007958, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 0.8}) +[2026-06-23 21:28:36,872][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.13105998933315277, 'train/grad_norm_pre_clip': 24.99090003967285, 'timing/data': 0.0008041459950618446, 'timing/model': 1.2059222940006293, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 0.8}) +[2026-06-23 21:28:41,291][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.011547708883881569, 'train/grad_norm_pre_clip': 9.93844223022461, 'timing/data': 0.0007987660064827651, 'timing/model': 1.1979855720128398, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 0.8}) +[2026-06-23 21:28:45,826][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.05271286517381668, 'train/grad_norm_pre_clip': 12.526753425598145, 'timing/data': 0.0007606250001117587, 'timing/model': 1.2028296509815846, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 0.8}) +[2026-06-23 21:28:50,153][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.043130550533533096, 'train/grad_norm_pre_clip': 22.201688766479492, 'timing/data': 0.0007678790134377778, 'timing/model': 1.1786693649773952, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 0.8}) +[2026-06-23 21:28:54,528][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.007033695466816425, 'train/grad_norm_pre_clip': 22.703933715820312, 'timing/data': 0.0011372620065230876, 'timing/model': 1.2003015329828486, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 0.8}) +[2026-06-23 21:28:58,865][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.1365634948015213, 'train/grad_norm_pre_clip': 60.43589782714844, 'timing/data': 0.0008112280047498643, 'timing/model': 1.1977114779874682, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 0.8}) +[2026-06-23 21:29:03,474][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.05905994400382042, 'train/grad_norm_pre_clip': 38.12236404418945, 'timing/data': 0.0007798909791745245, 'timing/model': 1.2005173130019102, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.8}) +[2026-06-23 21:29:07,820][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.2275494933128357, 'train/grad_norm_pre_clip': 54.153160095214844, 'timing/data': 0.0007805620261933655, 'timing/model': 1.1828587490017526, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 0.8}) +[2026-06-23 21:29:12,173][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.03186532109975815, 'train/grad_norm_pre_clip': 14.512951850891113, 'timing/data': 0.0012693650205619633, 'timing/model': 1.1802255340153351, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 0.8}) +[2026-06-23 21:29:16,493][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.10864613205194473, 'train/grad_norm_pre_clip': 16.959823608398438, 'timing/data': 0.0007914830057416111, 'timing/model': 1.1790756410046015, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 0.8}) +[2026-06-23 21:29:21,106][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.02051086723804474, 'train/grad_norm_pre_clip': 15.742915153503418, 'timing/data': 0.0008045459981076419, 'timing/model': 1.1982638339977711, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 0.8}) +[2026-06-23 21:29:25,452][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.05422858148813248, 'train/grad_norm_pre_clip': 56.84444046020508, 'timing/data': 0.0007555469928774983, 'timing/model': 1.1820532940037083, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 0.8}) +[2026-06-23 21:29:29,805][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.08777794986963272, 'train/grad_norm_pre_clip': 39.09709167480469, 'timing/data': 0.0008070809999480844, 'timing/model': 1.1817833149980288, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 0.8}) +[2026-06-23 21:29:34,162][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.0893678143620491, 'train/grad_norm_pre_clip': 26.757062911987305, 'timing/data': 0.0007562469982076436, 'timing/model': 1.1848260189872235, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 0.8}) +[2026-06-23 21:29:38,753][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.03937140479683876, 'train/grad_norm_pre_clip': 28.827795028686523, 'timing/data': 0.0008156870026141405, 'timing/model': 1.4283878120186273, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 0.8}) +[2026-06-23 21:29:43,145][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.09102919697761536, 'train/grad_norm_pre_clip': 18.991029739379883, 'timing/data': 0.0007725479954387993, 'timing/model': 1.2007700670219492, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 0.8}) +[2026-06-23 21:29:47,537][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.12262391299009323, 'train/grad_norm_pre_clip': 24.87493324279785, 'timing/data': 0.0007967920100782067, 'timing/model': 1.1851570919970982, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.8}) +[2026-06-23 21:29:51,867][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.09473064541816711, 'train/grad_norm_pre_clip': 32.9206657409668, 'timing/data': 0.0007659849943593144, 'timing/model': 1.1822563440073282, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 0.81}) +[2026-06-23 21:29:56,265][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.093850277364254, 'train/grad_norm_pre_clip': 23.622087478637695, 'timing/data': 0.0011773760197684169, 'timing/model': 1.2065061469911598, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 0.81}) +[2026-06-23 21:30:00,874][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.011187411844730377, 'train/grad_norm_pre_clip': 9.790811538696289, 'timing/data': 0.0007507370028179139, 'timing/model': 1.1773016919905785, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 0.81}) +[2026-06-23 21:30:05,275][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.08172152936458588, 'train/grad_norm_pre_clip': 19.67337417602539, 'timing/data': 0.0007321829907596111, 'timing/model': 1.2018281359924003, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 0.81}) +[2026-06-23 21:30:09,610][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.016954567283391953, 'train/grad_norm_pre_clip': 16.271560668945312, 'timing/data': 0.0007747510098852217, 'timing/model': 1.1821988799783867, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 0.81}) +[2026-06-23 21:30:13,973][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.04925446957349777, 'train/grad_norm_pre_clip': 26.14649200439453, 'timing/data': 0.0009068159852176905, 'timing/model': 1.1829154060105793, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 0.81}) +[2026-06-23 21:30:18,558][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.06464596837759018, 'train/grad_norm_pre_clip': 19.027353286743164, 'timing/data': 0.0007463189831469208, 'timing/model': 1.2051095200004056, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 0.81}) +[2026-06-23 21:30:22,943][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.057578254491090775, 'train/grad_norm_pre_clip': 15.115096092224121, 'timing/data': 0.0007739409920759499, 'timing/model': 1.1840259080054238, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 0.81}) +[2026-06-23 21:30:27,306][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.08024628460407257, 'train/grad_norm_pre_clip': 27.71599578857422, 'timing/data': 0.0007694420055486262, 'timing/model': 1.2065559439943172, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 0.81}) +[2026-06-23 21:30:31,697][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.04564525559544563, 'train/grad_norm_pre_clip': 16.20354652404785, 'timing/data': 0.000891467003384605, 'timing/model': 1.2032463009818457, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.81}) +[2026-06-23 21:30:36,245][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.029072724282741547, 'train/grad_norm_pre_clip': 34.570457458496094, 'timing/data': 0.0008904959831852466, 'timing/model': 1.1842746500042267, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 0.81}) +[2026-06-23 21:30:40,595][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.12897856533527374, 'train/grad_norm_pre_clip': 34.6136589050293, 'timing/data': 0.000765614997362718, 'timing/model': 1.183394355000928, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 0.81}) +[2026-06-23 21:30:44,957][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.026412712410092354, 'train/grad_norm_pre_clip': 5.774118423461914, 'timing/data': 0.001009505009278655, 'timing/model': 1.1847339700034354, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 0.81}) +[2026-06-23 21:30:49,362][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.07454687356948853, 'train/grad_norm_pre_clip': 12.593852043151855, 'timing/data': 0.0010326990159228444, 'timing/model': 1.2065736609802116, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 0.81}) +[2026-06-23 21:30:53,921][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.08369029313325882, 'train/grad_norm_pre_clip': 14.379997253417969, 'timing/data': 0.0009043109894264489, 'timing/model': 1.1843887940049171, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.81}) +[2026-06-23 21:30:58,292][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.024594655260443687, 'train/grad_norm_pre_clip': 18.224451065063477, 'timing/data': 0.001003484008833766, 'timing/model': 1.191146249009762, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 0.81}) +[2026-06-23 21:31:02,645][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.04750679060816765, 'train/grad_norm_pre_clip': 12.760979652404785, 'timing/data': 0.000753020984120667, 'timing/model': 1.1876547590072732, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 0.81}) +[2026-06-23 21:31:07,008][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.11499355733394623, 'train/grad_norm_pre_clip': 14.975715637207031, 'timing/data': 0.0008093359938357025, 'timing/model': 1.1875950700195972, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 0.81}) +[2026-06-23 21:31:11,573][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.006159376818686724, 'train/grad_norm_pre_clip': 11.857275009155273, 'timing/data': 0.0008299740147776902, 'timing/model': 1.2003939369751606, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 0.81}) +[2026-06-23 21:31:15,975][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2100 +[2026-06-23 21:32:13,458][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2100 +[2026-06-23 21:32:13,459][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2100 +[2026-06-23 21:32:50,348][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2100 +[2026-06-23 21:32:50,350][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.059739310294389725, 'train/grad_norm_pre_clip': 23.193931579589844, 'eval/action_loss/seconds': 57.47965358800138, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.04215485602617264, 'eval/latency_2/loss': 0.04215485602617264, 'eval/flappy/loss': 0.04215485602617264, 'eval/flappy/latency_2/loss': 0.04215485602617264, 'eval/flap_precision': 0.9776536312849162, 'eval/flap_recall': 0.9497964721845319, 'eval/flap_f1': 0.9635237439779767, 'eval/flap_rate_model': 0.1988888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9635237439779767, 'eval/latency_2/flap_precision': 0.9776536312849162, 'eval/latency_2/flap_recall': 0.9497964721845319, 'eval/latency_2/flap_f1': 0.9635237439779767, 'eval/latency_2/flap_rate_model': 0.1988888888888889, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9635237439779767, 'eval/action_classification/data_wait_seconds': 0.46012401580810547, 'eval/action_classification/predict_seconds': 36.28733444213867, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0008146550098899752, 'timing/model': 1.2161769179801922, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.81}) +[2026-06-23 21:32:54,672][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.046076178550720215, 'train/grad_norm_pre_clip': 21.740306854248047, 'timing/data': 0.0009623680089134723, 'timing/model': 1.1766468069981784, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 0.81}) +[2026-06-23 21:32:59,221][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.047371894121170044, 'train/grad_norm_pre_clip': 21.44544792175293, 'timing/data': 0.000754704000428319, 'timing/model': 1.2137425029941369, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 0.81}) +[2026-06-23 21:33:03,568][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.05180366337299347, 'train/grad_norm_pre_clip': 10.588756561279297, 'timing/data': 0.0007828059897292405, 'timing/model': 1.1881846850155853, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 0.81}) +[2026-06-23 21:33:07,878][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.17323870956897736, 'train/grad_norm_pre_clip': 27.754533767700195, 'timing/data': 0.0008808770217001438, 'timing/model': 1.1896017330000177, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 0.81}) +[2026-06-23 21:33:12,231][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.07316608726978302, 'train/grad_norm_pre_clip': 13.881831169128418, 'timing/data': 0.0009273640171159059, 'timing/model': 1.1945590329996776, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 0.81}) +[2026-06-23 21:33:16,815][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.3346995413303375, 'train/grad_norm_pre_clip': 38.077003479003906, 'timing/data': 0.0008166989719029516, 'timing/model': 1.1726109519950114, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 0.82}) +[2026-06-23 21:33:21,136][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.11311952769756317, 'train/grad_norm_pre_clip': 32.552650451660156, 'timing/data': 0.000793456012615934, 'timing/model': 1.2014666959876195, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 0.82}) +[2026-06-23 21:33:25,486][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.03801388666033745, 'train/grad_norm_pre_clip': 28.873075485229492, 'timing/data': 0.0010850650141946971, 'timing/model': 1.1976622580259573, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 0.82}) +[2026-06-23 21:33:29,822][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.04439118877053261, 'train/grad_norm_pre_clip': 28.417333602905273, 'timing/data': 0.0009320109966211021, 'timing/model': 1.1756969350099098, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.82}) +[2026-06-23 21:33:34,409][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.05283621326088905, 'train/grad_norm_pre_clip': 19.99064064025879, 'timing/data': 0.0008315859886351973, 'timing/model': 1.1779010180034675, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.82}) +[2026-06-23 21:33:38,777][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.06923844665288925, 'train/grad_norm_pre_clip': 13.136344909667969, 'timing/data': 0.0007756930135656148, 'timing/model': 1.1922052869922481, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 0.82}) +[2026-06-23 21:33:43,134][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.035143911838531494, 'train/grad_norm_pre_clip': 33.45319366455078, 'timing/data': 0.001236825017258525, 'timing/model': 1.1918075119901914, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 0.82}) +[2026-06-23 21:33:47,521][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.1691400706768036, 'train/grad_norm_pre_clip': 27.20069122314453, 'timing/data': 0.0009059739822987467, 'timing/model': 1.2040872809884604, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 0.82}) +[2026-06-23 21:33:52,107][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.08410973846912384, 'train/grad_norm_pre_clip': 31.435598373413086, 'timing/data': 0.0007490339921787381, 'timing/model': 1.1957393680058885, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 0.82}) +[2026-06-23 21:33:56,459][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.11384187638759613, 'train/grad_norm_pre_clip': 23.69337272644043, 'timing/data': 0.0007634799985680729, 'timing/model': 1.1802219179808162, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 0.82}) +[2026-06-23 21:34:00,849][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.007185727823525667, 'train/grad_norm_pre_clip': 14.114779472351074, 'timing/data': 0.0007336650160141289, 'timing/model': 1.2065500830067322, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 0.82}) +[2026-06-23 21:34:05,181][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.09008927643299103, 'train/grad_norm_pre_clip': 22.32901382446289, 'timing/data': 0.0007600550015922636, 'timing/model': 1.1945783239789307, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 0.82}) +[2026-06-23 21:34:09,767][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.11363217979669571, 'train/grad_norm_pre_clip': 21.65053939819336, 'timing/data': 0.0007269729976542294, 'timing/model': 1.180474956985563, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 0.82}) +[2026-06-23 21:34:14,164][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.11823532730340958, 'train/grad_norm_pre_clip': 32.95261001586914, 'timing/data': 0.0007936659967526793, 'timing/model': 1.195392569003161, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 0.82}) +[2026-06-23 21:34:18,563][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.14185164868831635, 'train/grad_norm_pre_clip': 30.068029403686523, 'timing/data': 0.0007482219953089952, 'timing/model': 1.1974589779856615, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.82}) +[2026-06-23 21:34:22,911][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.08697528392076492, 'train/grad_norm_pre_clip': 10.222959518432617, 'timing/data': 0.0007546439883299172, 'timing/model': 1.1807784600241575, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 0.82}) +[2026-06-23 21:34:27,513][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.040455110371112823, 'train/grad_norm_pre_clip': 8.068220138549805, 'timing/data': 0.0007983649848029017, 'timing/model': 1.1955904469941743, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 0.82}) +[2026-06-23 21:34:31,877][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.14126056432724, 'train/grad_norm_pre_clip': 13.167707443237305, 'timing/data': 0.0007723769813310355, 'timing/model': 1.1795521449821535, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 0.82}) +[2026-06-23 21:34:36,302][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.16229216754436493, 'train/grad_norm_pre_clip': 22.75057029724121, 'timing/data': 0.0008118299883790314, 'timing/model': 1.214843269000994, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 0.82}) +[2026-06-23 21:34:40,712][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.12461912631988525, 'train/grad_norm_pre_clip': 14.67601490020752, 'timing/data': 0.0011438129877205938, 'timing/model': 1.221123332012212, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 0.82}) +[2026-06-23 21:34:45,384][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.04066583886742592, 'train/grad_norm_pre_clip': 8.344694137573242, 'timing/data': 0.0008485169964842498, 'timing/model': 1.2063091110030655, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 0.82}) +[2026-06-23 21:34:49,761][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.08749931305646896, 'train/grad_norm_pre_clip': 17.94698715209961, 'timing/data': 0.0011015449999831617, 'timing/model': 1.2159804810071364, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 0.82}) +[2026-06-23 21:34:54,156][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.11304258555173874, 'train/grad_norm_pre_clip': 19.634017944335938, 'timing/data': 0.0008423060062341392, 'timing/model': 1.1991324029804673, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 0.82}) +[2026-06-23 21:34:58,531][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.041323471814394, 'train/grad_norm_pre_clip': 10.084521293640137, 'timing/data': 0.001242355996510014, 'timing/model': 1.196653733990388, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 0.82}) +[2026-06-23 21:35:03,097][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.07897333800792694, 'train/grad_norm_pre_clip': 8.50018310546875, 'timing/data': 0.0008042260014917701, 'timing/model': 1.1959658340201713, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.82}) +[2026-06-23 21:35:07,468][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.040564779192209244, 'train/grad_norm_pre_clip': 17.128124237060547, 'timing/data': 0.0007721760193817317, 'timing/model': 1.1778910389984958, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 0.82}) +[2026-06-23 21:35:11,843][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.11374854296445847, 'train/grad_norm_pre_clip': 9.767727851867676, 'timing/data': 0.0007609659805893898, 'timing/model': 1.2006635239813477, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 0.83}) +[2026-06-23 21:35:16,235][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.0933891087770462, 'train/grad_norm_pre_clip': 15.457853317260742, 'timing/data': 0.0007759440049994737, 'timing/model': 1.208102347998647, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 0.83}) +[2026-06-23 21:35:20,774][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.03497123718261719, 'train/grad_norm_pre_clip': 28.81736183166504, 'timing/data': 0.0007725369941908866, 'timing/model': 1.1802117790211923, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 0.83}) +[2026-06-23 21:35:25,143][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.22018757462501526, 'train/grad_norm_pre_clip': 41.3795280456543, 'timing/data': 0.0007536229968536645, 'timing/model': 1.2091206099721603, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 0.83}) +[2026-06-23 21:35:29,511][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.035356707870960236, 'train/grad_norm_pre_clip': 17.114450454711914, 'timing/data': 0.0007704330200795084, 'timing/model': 1.1958263790002093, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 0.83}) +[2026-06-23 21:35:33,910][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.06623826920986176, 'train/grad_norm_pre_clip': 17.102638244628906, 'timing/data': 0.0008698370074853301, 'timing/model': 1.2059290450124536, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 0.83}) +[2026-06-23 21:35:38,480][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.005497036036103964, 'train/grad_norm_pre_clip': 6.427712917327881, 'timing/data': 0.0007745609909761697, 'timing/model': 1.1955827589845285, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 0.83}) +[2026-06-23 21:35:42,839][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.13163889944553375, 'train/grad_norm_pre_clip': 23.86178207397461, 'timing/data': 0.0009662759839557111, 'timing/model': 1.199889957002597, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 0.83}) +[2026-06-23 21:35:47,226][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.02639787085354328, 'train/grad_norm_pre_clip': 10.184850692749023, 'timing/data': 0.0008590469951741397, 'timing/model': 1.2008859979978297, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.83}) +[2026-06-23 21:35:51,624][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.035342250019311905, 'train/grad_norm_pre_clip': 19.366281509399414, 'timing/data': 0.0007832669944036752, 'timing/model': 1.2018376759951934, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 0.83}) +[2026-06-23 21:35:56,204][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.011722410097718239, 'train/grad_norm_pre_clip': 5.245752811431885, 'timing/data': 0.000943472987273708, 'timing/model': 1.1983715030073654, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 0.83}) +[2026-06-23 21:36:00,555][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.04053656756877899, 'train/grad_norm_pre_clip': 21.699581146240234, 'timing/data': 0.0011812419979833066, 'timing/model': 1.1809269439836498, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 0.83}) +[2026-06-23 21:36:04,903][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.0654192641377449, 'train/grad_norm_pre_clip': 32.84770965576172, 'timing/data': 0.0008861670212354511, 'timing/model': 1.1987689490197226, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 0.83}) +[2026-06-23 21:36:09,299][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.00018380486289970577, 'train/grad_norm_pre_clip': 14.299358367919922, 'timing/data': 0.0007539440121036023, 'timing/model': 1.2030434460029937, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 0.83}) +[2026-06-23 21:36:13,887][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.16992832720279694, 'train/grad_norm_pre_clip': 60.86748504638672, 'timing/data': 0.0007922940130811185, 'timing/model': 1.2031379909894895, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 0.83}) +[2026-06-23 21:36:18,248][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.10531719774007797, 'train/grad_norm_pre_clip': 35.24530029296875, 'timing/data': 0.0009703320101834834, 'timing/model': 1.1817519089963753, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 0.83}) +[2026-06-23 21:36:22,661][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.007351365406066179, 'train/grad_norm_pre_clip': 11.4852876663208, 'timing/data': 0.000804146024165675, 'timing/model': 1.214447817998007, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 0.83}) +[2026-06-23 21:36:27,035][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.06038396432995796, 'train/grad_norm_pre_clip': 17.02870750427246, 'timing/data': 0.0007823250198271126, 'timing/model': 1.1825049710168969, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 0.83}) +[2026-06-23 21:36:31,586][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.014509277418255806, 'train/grad_norm_pre_clip': 25.881057739257812, 'timing/data': 0.0009801400010474026, 'timing/model': 1.194693830999313, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.83}) +[2026-06-23 21:36:35,961][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.05663083493709564, 'train/grad_norm_pre_clip': 29.718053817749023, 'timing/data': 0.0007155320199672133, 'timing/model': 1.1981126180035062, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 0.83}) +[2026-06-23 21:36:40,374][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.28096139430999756, 'train/grad_norm_pre_clip': 55.45884323120117, 'timing/data': 0.0007537729979958385, 'timing/model': 1.2218354289943818, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 0.83}) +[2026-06-23 21:36:44,766][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.06984778493642807, 'train/grad_norm_pre_clip': 29.106016159057617, 'timing/data': 0.0008314159931614995, 'timing/model': 1.2002520420064684, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 0.83}) +[2026-06-23 21:36:49,330][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.0014921280089765787, 'train/grad_norm_pre_clip': 12.943174362182617, 'timing/data': 0.0007475409947801381, 'timing/model': 1.1973798550025094, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 0.83}) +[2026-06-23 21:36:53,674][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.10759735852479935, 'train/grad_norm_pre_clip': 42.72077178955078, 'timing/data': 0.0007196789956651628, 'timing/model': 1.176601004000986, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 0.83}) +[2026-06-23 21:36:58,057][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.0037563429214060307, 'train/grad_norm_pre_clip': 15.845998764038086, 'timing/data': 0.0007510370051022619, 'timing/model': 1.201202067983104, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 0.83}) +[2026-06-23 21:37:02,411][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.09047195315361023, 'train/grad_norm_pre_clip': 27.21337890625, 'timing/data': 0.0009126969962380826, 'timing/model': 1.1825640320021193, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 0.83}) +[2026-06-23 21:37:06,990][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.021745657548308372, 'train/grad_norm_pre_clip': 21.224145889282227, 'timing/data': 0.0010240219999104738, 'timing/model': 1.2025256239867304, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.84}) +[2026-06-23 21:37:11,366][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.06162102892994881, 'train/grad_norm_pre_clip': 37.4964714050293, 'timing/data': 0.0008037649968173355, 'timing/model': 1.203359785984503, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 0.84}) +[2026-06-23 21:37:15,772][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.01238352432847023, 'train/grad_norm_pre_clip': 30.74224090576172, 'timing/data': 0.0009025269828271121, 'timing/model': 1.2022348269820213, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.84}) +[2026-06-23 21:37:20,174][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.025644108653068542, 'train/grad_norm_pre_clip': 18.663293838500977, 'timing/data': 0.0007439749897457659, 'timing/model': 1.2094584239821415, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 0.84}) +[2026-06-23 21:37:24,743][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.08865100145339966, 'train/grad_norm_pre_clip': 14.814719200134277, 'timing/data': 0.0008415750053245574, 'timing/model': 1.2152226839971263, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 0.84}) +[2026-06-23 21:37:29,125][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.04827270656824112, 'train/grad_norm_pre_clip': 15.159632682800293, 'timing/data': 0.0009688699792604893, 'timing/model': 1.205704699998023, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 0.84}) +[2026-06-23 21:37:33,511][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.024898650124669075, 'train/grad_norm_pre_clip': 21.860403060913086, 'timing/data': 0.0008230209932662547, 'timing/model': 1.2003429530013818, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 0.84}) +[2026-06-23 21:37:37,897][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.030935904011130333, 'train/grad_norm_pre_clip': 26.63655662536621, 'timing/data': 0.000982715020654723, 'timing/model': 1.2030744710064027, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.84}) +[2026-06-23 21:37:42,508][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.11802922189235687, 'train/grad_norm_pre_clip': 25.565452575683594, 'timing/data': 0.0007421720074489713, 'timing/model': 1.1984244589984883, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 0.84}) +[2026-06-23 21:37:46,915][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.22412069141864777, 'train/grad_norm_pre_clip': 17.084781646728516, 'timing/data': 0.0016430660034529865, 'timing/model': 1.2089924750034697, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 0.84}) +[2026-06-23 21:37:51,307][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.0235338993370533, 'train/grad_norm_pre_clip': 17.2119140625, 'timing/data': 0.0007447969983331859, 'timing/model': 1.1842551099834964, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 0.84}) +[2026-06-23 21:37:55,704][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.06821399182081223, 'train/grad_norm_pre_clip': 14.23670482635498, 'timing/data': 0.0008005090057849884, 'timing/model': 1.2099961600033566, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 0.84}) +[2026-06-23 21:38:00,290][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.044935453683137894, 'train/grad_norm_pre_clip': 10.006407737731934, 'timing/data': 0.0007694319938309491, 'timing/model': 1.2093407490174286, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.84}) +[2026-06-23 21:38:04,630][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.015275074169039726, 'train/grad_norm_pre_clip': 9.872038841247559, 'timing/data': 0.0010238719987682998, 'timing/model': 1.1880387419951148, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 0.84}) +[2026-06-23 21:38:08,999][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.08603747189044952, 'train/grad_norm_pre_clip': 25.251699447631836, 'timing/data': 0.0011672370019368827, 'timing/model': 1.1898569419863634, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 0.84}) +[2026-06-23 21:38:13,342][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.10560484975576401, 'train/grad_norm_pre_clip': 27.230594635009766, 'timing/data': 0.0009457069972995669, 'timing/model': 1.179194392025238, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 0.84}) +[2026-06-23 21:38:17,906][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.026480847969651222, 'train/grad_norm_pre_clip': 34.117916107177734, 'timing/data': 0.0010700569837354124, 'timing/model': 1.1814869979862124, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 0.84}) +[2026-06-23 21:38:22,270][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.25352954864501953, 'train/grad_norm_pre_clip': 17.313194274902344, 'timing/data': 0.000750817998778075, 'timing/model': 1.1893658350163605, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 0.84}) +[2026-06-23 21:38:26,628][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.053274042904376984, 'train/grad_norm_pre_clip': 16.737586975097656, 'timing/data': 0.0010185910214204341, 'timing/model': 1.1875859269930515, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 0.84}) +[2026-06-23 21:38:30,971][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.08404435962438583, 'train/grad_norm_pre_clip': 9.20709228515625, 'timing/data': 0.0007949389982968569, 'timing/model': 1.1881852340011392, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 0.84}) +[2026-06-23 21:38:35,557][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.15063638985157013, 'train/grad_norm_pre_clip': 19.728872299194336, 'timing/data': 0.0010400110040791333, 'timing/model': 1.2038706750026904, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 0.84}) +[2026-06-23 21:38:39,914][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.1550782024860382, 'train/grad_norm_pre_clip': 37.744590759277344, 'timing/data': 0.0007977640198078007, 'timing/model': 1.1846476089849602, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 0.84}) +[2026-06-23 21:38:44,299][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.1126822829246521, 'train/grad_norm_pre_clip': 17.775659561157227, 'timing/data': 0.0007648030004929751, 'timing/model': 1.1860403329774272, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.84}) +[2026-06-23 21:38:48,696][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.05967162549495697, 'train/grad_norm_pre_clip': 13.859183311462402, 'timing/data': 0.0009486330091021955, 'timing/model': 1.2161069640133064, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 0.84}) +[2026-06-23 21:38:53,276][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.029564067721366882, 'train/grad_norm_pre_clip': 14.182506561279297, 'timing/data': 0.0007528309943154454, 'timing/model': 1.1994937600102276, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 0.84}) +[2026-06-23 21:38:57,640][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.0544220507144928, 'train/grad_norm_pre_clip': 32.705413818359375, 'timing/data': 0.0011855410120915622, 'timing/model': 1.184924191009486, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 0.84}) +[2026-06-23 21:39:02,009][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.2273070514202118, 'train/grad_norm_pre_clip': 51.62249755859375, 'timing/data': 0.0007616769871674478, 'timing/model': 1.1828262750059366, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 0.85}) +[2026-06-23 21:39:06,380][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.003582162782549858, 'train/grad_norm_pre_clip': 6.428731918334961, 'timing/data': 0.0007969930011313409, 'timing/model': 1.2049882619758137, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 0.85}) +[2026-06-23 21:39:10,893][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.04823575168848038, 'train/grad_norm_pre_clip': 12.440185546875, 'timing/data': 0.0007383530028164387, 'timing/model': 1.1839420970063657, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 0.85}) +[2026-06-23 21:39:15,249][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.012600522488355637, 'train/grad_norm_pre_clip': 46.98365020751953, 'timing/data': 0.0009065050107892603, 'timing/model': 1.2060869530250784, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 0.85}) +[2026-06-23 21:39:19,625][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.1485646367073059, 'train/grad_norm_pre_clip': 40.404422760009766, 'timing/data': 0.0007236370001919568, 'timing/model': 1.193645566003397, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 0.85}) +[2026-06-23 21:39:23,953][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.008825193159282207, 'train/grad_norm_pre_clip': 16.833099365234375, 'timing/data': 0.0008028229931369424, 'timing/model': 1.1827496140031144, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 0.85}) +[2026-06-23 21:39:28,501][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.05054507032036781, 'train/grad_norm_pre_clip': 12.99260139465332, 'timing/data': 0.0007052430009935051, 'timing/model': 1.1978896479995456, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.85}) +[2026-06-23 21:39:32,842][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.010445943102240562, 'train/grad_norm_pre_clip': 17.040557861328125, 'timing/data': 0.0008456229988951236, 'timing/model': 1.1785655349958688, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 0.85}) +[2026-06-23 21:39:37,184][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.1764589548110962, 'train/grad_norm_pre_clip': 39.04062271118164, 'timing/data': 0.0009349769970867783, 'timing/model': 1.1779425730055664, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 0.85}) +[2026-06-23 21:39:41,614][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.008061295375227928, 'train/grad_norm_pre_clip': 8.965154647827148, 'timing/data': 0.0010013900173362345, 'timing/model': 1.2082328790274914, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 0.85}) +[2026-06-23 21:39:46,147][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.1539670079946518, 'train/grad_norm_pre_clip': 40.2650146484375, 'timing/data': 0.0007501670042984188, 'timing/model': 1.2057658409758005, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 0.85}) +[2026-06-23 21:39:50,523][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.052632253617048264, 'train/grad_norm_pre_clip': 26.388521194458008, 'timing/data': 0.0007359299925155938, 'timing/model': 1.2005488019785844, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 0.85}) +[2026-06-23 21:39:54,868][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.05290739983320236, 'train/grad_norm_pre_clip': 20.775148391723633, 'timing/data': 0.0007590129971504211, 'timing/model': 1.1776872129994445, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 0.85}) +[2026-06-23 21:39:59,203][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.034298233687877655, 'train/grad_norm_pre_clip': 19.66947364807129, 'timing/data': 0.0007388450030703098, 'timing/model': 1.1787948900018819, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 0.85}) +[2026-06-23 21:40:03,790][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.02655060961842537, 'train/grad_norm_pre_clip': 38.28220748901367, 'timing/data': 0.0007346580096054822, 'timing/model': 1.2033059069945011, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 0.85}) +[2026-06-23 21:40:08,162][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.07803227007389069, 'train/grad_norm_pre_clip': 14.853958129882812, 'timing/data': 0.0007265220046974719, 'timing/model': 1.1772265619947575, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 0.85}) +[2026-06-23 21:40:12,537][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2200 +[2026-06-23 21:41:10,168][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2200 +[2026-06-23 21:41:10,169][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2200 +[2026-06-23 21:41:46,967][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2200 +[2026-06-23 21:41:46,969][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.016397887840867043, 'train/grad_norm_pre_clip': 14.259856224060059, 'eval/action_loss/seconds': 57.62745813699439, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.04363635554909706, 'eval/latency_2/loss': 0.04363635554909706, 'eval/flappy/loss': 0.04363635554909706, 'eval/flappy/latency_2/loss': 0.04363635554909706, 'eval/flap_precision': 0.9571619812583668, 'eval/flap_recall': 0.9701492537313433, 'eval/flap_f1': 0.963611859838275, 'eval/flap_rate_model': 0.2075, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9636118598382749, 'eval/latency_2/flap_precision': 0.9571619812583668, 'eval/latency_2/flap_recall': 0.9701492537313433, 'eval/latency_2/flap_f1': 0.963611859838275, 'eval/latency_2/flap_rate_model': 0.2075, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9636118598382749, 'eval/action_classification/data_wait_seconds': 0.46809810400009155, 'eval/action_classification/predict_seconds': 36.18558883666992, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007265220046974719, 'timing/model': 1.2035318950074725, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.85}) +[2026-06-23 21:41:51,560][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.002376462100073695, 'train/grad_norm_pre_clip': 17.085506439208984, 'timing/data': 0.0011708240199368447, 'timing/model': 1.1884168529941235, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 0.85}) +[2026-06-23 21:41:55,881][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.08407796174287796, 'train/grad_norm_pre_clip': 20.985519409179688, 'timing/data': 0.0007549340080004185, 'timing/model': 1.1756955689925235, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 0.85}) +[2026-06-23 21:42:00,184][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.023289930075407028, 'train/grad_norm_pre_clip': 31.01095962524414, 'timing/data': 0.0008300440094899386, 'timing/model': 1.1926253189740237, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 0.85}) +[2026-06-23 21:42:04,498][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.07804223895072937, 'train/grad_norm_pre_clip': 28.294736862182617, 'timing/data': 0.0008747870160732418, 'timing/model': 1.1756423809856642, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 0.85}) +[2026-06-23 21:42:09,090][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.0854337066411972, 'train/grad_norm_pre_clip': 29.097200393676758, 'timing/data': 0.0007659850234631449, 'timing/model': 1.1989199289819226, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 0.85}) +[2026-06-23 21:42:13,429][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.21694546937942505, 'train/grad_norm_pre_clip': 33.319732666015625, 'timing/data': 0.0008927299932111055, 'timing/model': 1.1711831229913514, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 0.85}) +[2026-06-23 21:42:17,762][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.08045978099107742, 'train/grad_norm_pre_clip': 16.042905807495117, 'timing/data': 0.0011932149936910719, 'timing/model': 1.1899166450020857, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.85}) +[2026-06-23 21:42:22,103][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.04678714647889137, 'train/grad_norm_pre_clip': 9.580365180969238, 'timing/data': 0.0012628640106413513, 'timing/model': 1.1930923129839357, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 0.85}) +[2026-06-23 21:42:26,662][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.014173699542880058, 'train/grad_norm_pre_clip': 15.346870422363281, 'timing/data': 0.0009776270017027855, 'timing/model': 1.1734735950012691, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 0.85}) +[2026-06-23 21:42:30,980][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.0700850561261177, 'train/grad_norm_pre_clip': 16.827545166015625, 'timing/data': 0.0007456580060534179, 'timing/model': 1.1771162649965845, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.86}) +[2026-06-23 21:42:35,360][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.0041306656785309315, 'train/grad_norm_pre_clip': 32.761837005615234, 'timing/data': 0.0007589219894725829, 'timing/model': 1.1984104290022515, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 0.86}) +[2026-06-23 21:42:39,695][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.04524477571249008, 'train/grad_norm_pre_clip': 12.964241981506348, 'timing/data': 0.0008526250021532178, 'timing/model': 1.1864024329988752, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 0.86}) +[2026-06-23 21:42:44,253][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.06646998226642609, 'train/grad_norm_pre_clip': 15.025900840759277, 'timing/data': 0.0007538219797424972, 'timing/model': 1.1785973020014353, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 0.86}) +[2026-06-23 21:42:48,601][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.07051870226860046, 'train/grad_norm_pre_clip': 17.316608428955078, 'timing/data': 0.0008156569965649396, 'timing/model': 1.181027503014775, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 0.86}) +[2026-06-23 21:42:52,950][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.014347221702337265, 'train/grad_norm_pre_clip': 19.26802635192871, 'timing/data': 0.0007513280143029988, 'timing/model': 1.2041386119963136, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 0.86}) +[2026-06-23 21:42:57,305][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.0833875760436058, 'train/grad_norm_pre_clip': 21.729516983032227, 'timing/data': 0.0007407589873764664, 'timing/model': 1.1857532039866783, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 0.86}) +[2026-06-23 21:43:01,849][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.08098620921373367, 'train/grad_norm_pre_clip': 11.76948356628418, 'timing/data': 0.0008538179972674698, 'timing/model': 1.1993582809809595, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 0.86}) +[2026-06-23 21:43:06,205][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.211617112159729, 'train/grad_norm_pre_clip': 20.957063674926758, 'timing/data': 0.0009034189861267805, 'timing/model': 1.1823044929769821, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 0.86}) +[2026-06-23 21:43:10,588][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.1480434536933899, 'train/grad_norm_pre_clip': 32.71577072143555, 'timing/data': 0.0012537860020529479, 'timing/model': 1.1795001510181464, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 0.86}) +[2026-06-23 21:43:14,963][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.09243471920490265, 'train/grad_norm_pre_clip': 21.498863220214844, 'timing/data': 0.0007510879950132221, 'timing/model': 1.202022133016726, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.86}) +[2026-06-23 21:43:19,542][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.027398955076932907, 'train/grad_norm_pre_clip': 9.141735076904297, 'timing/data': 0.0009536020224913955, 'timing/model': 1.2054855430033058, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 0.86}) +[2026-06-23 21:43:23,923][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.01812300644814968, 'train/grad_norm_pre_clip': 33.18841552734375, 'timing/data': 0.0009182959911413491, 'timing/model': 1.1973635759786703, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 0.86}) +[2026-06-23 21:43:28,294][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.020382175222039223, 'train/grad_norm_pre_clip': 31.315593719482422, 'timing/data': 0.0009225649992004037, 'timing/model': 1.1972942989959847, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 0.86}) +[2026-06-23 21:43:32,641][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.09822888672351837, 'train/grad_norm_pre_clip': 33.9195556640625, 'timing/data': 0.0007319819997064769, 'timing/model': 1.1824012020078953, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 0.86}) +[2026-06-23 21:43:37,207][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.02168426476418972, 'train/grad_norm_pre_clip': 11.023164749145508, 'timing/data': 0.0008914569916669279, 'timing/model': 1.2046025109884795, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 0.86}) +[2026-06-23 21:43:41,583][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.06408814340829849, 'train/grad_norm_pre_clip': 28.89116859436035, 'timing/data': 0.0012266770063433796, 'timing/model': 1.1959318120207172, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 0.86}) +[2026-06-23 21:43:45,952][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.031298596411943436, 'train/grad_norm_pre_clip': 26.65977668762207, 'timing/data': 0.0010362649918533862, 'timing/model': 1.1799053910071962, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 0.86}) +[2026-06-23 21:43:50,338][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.03952553868293762, 'train/grad_norm_pre_clip': 36.86857223510742, 'timing/data': 0.0012101050233468413, 'timing/model': 1.197050589020364, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 0.86}) +[2026-06-23 21:43:54,906][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.12301833927631378, 'train/grad_norm_pre_clip': 20.892560958862305, 'timing/data': 0.0011583400191739202, 'timing/model': 1.2055970299988985, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 0.86}) +[2026-06-23 21:43:59,273][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.0972035750746727, 'train/grad_norm_pre_clip': 27.22750473022461, 'timing/data': 0.0008065300062298775, 'timing/model': 1.2000825020077173, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.86}) +[2026-06-23 21:44:03,658][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.01173455361276865, 'train/grad_norm_pre_clip': 18.94214630126953, 'timing/data': 0.0008333400182891637, 'timing/model': 1.196344224998029, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 0.86}) +[2026-06-23 21:44:08,019][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.048951923847198486, 'train/grad_norm_pre_clip': 25.733163833618164, 'timing/data': 0.0009020560246426612, 'timing/model': 1.2009294960007537, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 0.86}) +[2026-06-23 21:44:12,571][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.006551236379891634, 'train/grad_norm_pre_clip': 21.830480575561523, 'timing/data': 0.0008407139976043254, 'timing/model': 1.181044347002171, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 0.86}) +[2026-06-23 21:44:16,963][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.005495534278452396, 'train/grad_norm_pre_clip': 31.294567108154297, 'timing/data': 0.0007650429906789213, 'timing/model': 1.1999868150160182, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 0.86}) +[2026-06-23 21:44:21,316][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.033919502049684525, 'train/grad_norm_pre_clip': 18.456470489501953, 'timing/data': 0.0007771160162519664, 'timing/model': 1.204751758981729, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.86}) +[2026-06-23 21:44:25,673][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.11663621664047241, 'train/grad_norm_pre_clip': 22.549427032470703, 'timing/data': 0.0007735289982520044, 'timing/model': 1.2041085310047492, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 0.87}) +[2026-06-23 21:44:30,232][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.24698227643966675, 'train/grad_norm_pre_clip': 34.74433898925781, 'timing/data': 0.000738634989829734, 'timing/model': 1.1850514199759346, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 0.87}) +[2026-06-23 21:44:34,617][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.020308632403612137, 'train/grad_norm_pre_clip': 27.03754234313965, 'timing/data': 0.0010936599865090102, 'timing/model': 1.2148718380194623, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 0.87}) +[2026-06-23 21:44:38,999][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.10032925754785538, 'train/grad_norm_pre_clip': 9.665714263916016, 'timing/data': 0.0008906360017135739, 'timing/model': 1.205734866001876, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 0.87}) +[2026-06-23 21:44:43,389][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.00892537459731102, 'train/grad_norm_pre_clip': 16.85392189025879, 'timing/data': 0.0007767450006213039, 'timing/model': 1.2010387300106231, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.87}) +[2026-06-23 21:44:47,949][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.049187205731868744, 'train/grad_norm_pre_clip': 18.512027740478516, 'timing/data': 0.0007599939999636263, 'timing/model': 1.1837611370137893, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 0.87}) +[2026-06-23 21:44:52,357][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.08186125010251999, 'train/grad_norm_pre_clip': 19.486440658569336, 'timing/data': 0.0009665050019975752, 'timing/model': 1.1988086389901582, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 0.87}) +[2026-06-23 21:44:56,709][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.08288835734128952, 'train/grad_norm_pre_clip': 12.377976417541504, 'timing/data': 0.0008153169765137136, 'timing/model': 1.201280006003799, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 0.87}) +[2026-06-23 21:45:01,103][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.005514513701200485, 'train/grad_norm_pre_clip': 16.962718963623047, 'timing/data': 0.0007457270112354308, 'timing/model': 1.1969741900102235, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 0.87}) +[2026-06-23 21:45:05,685][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.01384993176907301, 'train/grad_norm_pre_clip': 18.271724700927734, 'timing/data': 0.0007833470008336008, 'timing/model': 1.2102529440016951, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 0.87}) +[2026-06-23 21:45:10,082][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.009883840568363667, 'train/grad_norm_pre_clip': 13.556699752807617, 'timing/data': 0.0010948429990094155, 'timing/model': 1.2036673969996627, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 0.87}) +[2026-06-23 21:45:14,465][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.0042338608764112, 'train/grad_norm_pre_clip': 11.44837760925293, 'timing/data': 0.0008901440014597028, 'timing/model': 1.1835905629850458, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 0.87}) +[2026-06-23 21:45:18,832][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.05499578267335892, 'train/grad_norm_pre_clip': 28.1114501953125, 'timing/data': 0.0007960509974509478, 'timing/model': 1.2012863669951912, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 0.87}) +[2026-06-23 21:45:23,427][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.04920274764299393, 'train/grad_norm_pre_clip': 18.181154251098633, 'timing/data': 0.0007893379952292889, 'timing/model': 1.2131889199954458, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 0.87}) +[2026-06-23 22:13:47,540][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2250 +[2026-06-23 22:13:47,542][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.04077224060893059, 'train/grad_norm_pre_clip': 28.556764602661133, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 443.8400000000255, 'rl_games_eval/mid_train/mean_length': 3594.6, 'rl_games_eval/mid_train/std_reward': 1.5200000000002547, 'rl_games_eval/mid_train/std_length': 10.8, 'rl_games_eval/mid_train/macro_mean_reward': 443.8400000000255, 'rl_games_eval/mid_train/macro_mean_length': 3594.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 443.8400000000255, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3594.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 1.5200000000002547, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 10.8, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 443.8400000000255, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0008265879878308624, 'timing/model': 1.1855680769949686, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.87}) +[2026-06-23 22:13:51,720][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.05140264704823494, 'train/grad_norm_pre_clip': 9.76323127746582, 'timing/data': 0.0007332239765673876, 'timing/model': 1.1548806200153194, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 0.87}) +[2026-06-23 22:13:56,141][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.09675275534391403, 'train/grad_norm_pre_clip': 22.369508743286133, 'timing/data': 0.001028630998916924, 'timing/model': 1.3683132879959885, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 0.87}) +[2026-06-23 22:14:00,362][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.03611355274915695, 'train/grad_norm_pre_clip': 26.06854248046875, 'timing/data': 0.0008102869905997068, 'timing/model': 1.1404402339830995, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 0.87}) +[2026-06-23 22:14:04,548][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.12361206859350204, 'train/grad_norm_pre_clip': 16.842199325561523, 'timing/data': 0.000981702993158251, 'timing/model': 1.1574609149829485, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 0.87}) +[2026-06-23 22:14:08,779][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.07054881751537323, 'train/grad_norm_pre_clip': 16.56187629699707, 'timing/data': 0.0007333249959629029, 'timing/model': 1.1541630100109614, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 0.87}) +[2026-06-23 22:14:13,237][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.09048223495483398, 'train/grad_norm_pre_clip': 13.98999309539795, 'timing/data': 0.0014495290233753622, 'timing/model': 1.1638436170178466, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.87}) +[2026-06-23 22:14:17,463][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.0014952042838558555, 'train/grad_norm_pre_clip': 14.780314445495605, 'timing/data': 0.0010083139932248741, 'timing/model': 1.1655948020052165, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 0.87}) +[2026-06-23 22:14:21,683][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.09969722479581833, 'train/grad_norm_pre_clip': 19.831871032714844, 'timing/data': 0.0007141490059439093, 'timing/model': 1.1725700510141905, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 0.87}) +[2026-06-23 22:14:25,948][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.042328279465436935, 'train/grad_norm_pre_clip': 13.816295623779297, 'timing/data': 0.0008378280035685748, 'timing/model': 1.1838954690028913, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 0.87}) +[2026-06-23 22:14:30,384][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.017768854275345802, 'train/grad_norm_pre_clip': 26.806154251098633, 'timing/data': 0.0007735100225545466, 'timing/model': 1.1680146110011265, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.87}) +[2026-06-23 22:14:34,671][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.026710283011198044, 'train/grad_norm_pre_clip': 24.208698272705078, 'timing/data': 0.0009935359994415194, 'timing/model': 1.1785986270115245, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 0.88}) +[2026-06-23 22:14:38,952][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.05213825777173042, 'train/grad_norm_pre_clip': 12.529362678527832, 'timing/data': 0.0009388050239067525, 'timing/model': 1.187398267997196, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 0.88}) +[2026-06-23 22:14:43,219][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.13887065649032593, 'train/grad_norm_pre_clip': 18.131847381591797, 'timing/data': 0.0007449459808412939, 'timing/model': 1.1763063200050965, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 0.88}) +[2026-06-23 22:14:47,748][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.009052553214132786, 'train/grad_norm_pre_clip': 74.42525482177734, 'timing/data': 0.0007686610042583197, 'timing/model': 1.1584259980008937, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 0.88}) +[2026-06-23 22:14:52,068][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.06757762283086777, 'train/grad_norm_pre_clip': 28.547496795654297, 'timing/data': 0.0007786790083628148, 'timing/model': 1.183500801998889, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 0.88}) +[2026-06-23 22:14:56,357][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.021036429330706596, 'train/grad_norm_pre_clip': 15.276654243469238, 'timing/data': 0.0008841429953463376, 'timing/model': 1.18280414500623, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 0.88}) +[2026-06-23 22:15:00,654][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.003702709451317787, 'train/grad_norm_pre_clip': 27.46970558166504, 'timing/data': 0.0007342070166487247, 'timing/model': 1.1899465809983667, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 0.88}) +[2026-06-23 22:15:05,132][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.033891815692186356, 'train/grad_norm_pre_clip': 12.2672119140625, 'timing/data': 0.0008277290035039186, 'timing/model': 1.1621264119748957, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 0.88}) +[2026-06-23 22:15:09,419][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.06727463752031326, 'train/grad_norm_pre_clip': 16.20804786682129, 'timing/data': 0.0007573399925604463, 'timing/model': 1.1669904699956533, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 0.88}) +[2026-06-23 22:15:13,721][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.0516328290104866, 'train/grad_norm_pre_clip': 25.607698440551758, 'timing/data': 0.0007429730030708015, 'timing/model': 1.1861091070168186, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.88}) +[2026-06-23 22:15:18,026][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.07360722869634628, 'train/grad_norm_pre_clip': 25.297576904296875, 'timing/data': 0.0010366660135332495, 'timing/model': 1.1846324279904366, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 0.88}) +[2026-06-23 22:15:22,572][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.004338547121733427, 'train/grad_norm_pre_clip': 7.669564247131348, 'timing/data': 0.0010541479859966785, 'timing/model': 1.407860803999938, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 0.88}) +[2026-06-23 22:15:26,875][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.04440612345933914, 'train/grad_norm_pre_clip': 16.485443115234375, 'timing/data': 0.0009129579993896186, 'timing/model': 1.19425794298877, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 0.88}) +[2026-06-23 22:15:31,194][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.056596022099256516, 'train/grad_norm_pre_clip': 23.85136604309082, 'timing/data': 0.000927704997593537, 'timing/model': 1.163838320004288, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 0.88}) +[2026-06-23 22:15:35,535][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.031567610800266266, 'train/grad_norm_pre_clip': 13.95592975616455, 'timing/data': 0.0006945040076971054, 'timing/model': 1.1689534519973677, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 0.88}) +[2026-06-23 22:15:40,137][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.013397830538451672, 'train/grad_norm_pre_clip': 14.29448127746582, 'timing/data': 0.0007557569770142436, 'timing/model': 1.43762958800653, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 0.88}) +[2026-06-23 22:15:44,474][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.046423543244600296, 'train/grad_norm_pre_clip': 22.59450340270996, 'timing/data': 0.0007798009901307523, 'timing/model': 1.1719440360029694, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 0.88}) +[2026-06-23 22:15:48,847][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.0866696685552597, 'train/grad_norm_pre_clip': 21.964174270629883, 'timing/data': 0.0008408130088355392, 'timing/model': 1.1888682690041605, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 0.88}) +[2026-06-23 22:15:53,160][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.015419448725879192, 'train/grad_norm_pre_clip': 18.804555892944336, 'timing/data': 0.0008019920205697417, 'timing/model': 1.1731302560074255, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 0.88}) +[2026-06-23 22:15:57,695][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.0028446803335100412, 'train/grad_norm_pre_clip': 23.208518981933594, 'timing/data': 0.0010510020074434578, 'timing/model': 1.1736793509917334, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.88}) +[2026-06-23 22:16:01,999][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.001974363811314106, 'train/grad_norm_pre_clip': 17.804773330688477, 'timing/data': 0.0009735789790283889, 'timing/model': 1.1758466039900668, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 0.88}) +[2026-06-23 22:16:06,335][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.03069237992167473, 'train/grad_norm_pre_clip': 12.972201347351074, 'timing/data': 0.0007683000003453344, 'timing/model': 1.20392634102609, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 0.88}) +[2026-06-23 22:16:10,697][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.03768296167254448, 'train/grad_norm_pre_clip': 15.885308265686035, 'timing/data': 0.0014554900117218494, 'timing/model': 1.190889208985027, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 0.88}) +[2026-06-23 22:16:15,302][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.2690124213695526, 'train/grad_norm_pre_clip': 20.336339950561523, 'timing/data': 0.000819534994661808, 'timing/model': 1.193932432011934, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 0.88}) +[2026-06-23 22:16:19,685][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.06116531044244766, 'train/grad_norm_pre_clip': 21.924169540405273, 'timing/data': 0.0008668310183566064, 'timing/model': 1.1889359870110638, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 0.88}) +[2026-06-23 22:16:23,972][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.02384161576628685, 'train/grad_norm_pre_clip': 34.34016799926758, 'timing/data': 0.0008027039875742048, 'timing/model': 1.1713427369832061, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 0.88}) +[2026-06-23 22:16:28,332][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.1334649622440338, 'train/grad_norm_pre_clip': 24.73215103149414, 'timing/data': 0.0007542830135207623, 'timing/model': 1.1915504169883206, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 0.89}) +[2026-06-23 22:16:32,887][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.0021817623637616634, 'train/grad_norm_pre_clip': 19.757097244262695, 'timing/data': 0.0007638520037289709, 'timing/model': 1.1772740360174794, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 0.89}) +[2026-06-23 22:16:37,217][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.02229655534029007, 'train/grad_norm_pre_clip': 16.158531188964844, 'timing/data': 0.0007359999872278422, 'timing/model': 1.198064683994744, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 0.89}) +[2026-06-23 22:16:41,548][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.0056636882945895195, 'train/grad_norm_pre_clip': 24.33272361755371, 'timing/data': 0.0007456080056726933, 'timing/model': 1.2027107480098493, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.89}) +[2026-06-23 22:16:45,874][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.10263274610042572, 'train/grad_norm_pre_clip': 47.32311248779297, 'timing/data': 0.0007343569886870682, 'timing/model': 1.193260474014096, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 0.89}) +[2026-06-23 22:16:50,391][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.026663826778531075, 'train/grad_norm_pre_clip': 26.599496841430664, 'timing/data': 0.0008469250169582665, 'timing/model': 1.3749962149886414, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 0.89}) +[2026-06-23 22:16:54,707][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.09849989414215088, 'train/grad_norm_pre_clip': 16.845317840576172, 'timing/data': 0.0006865079922135919, 'timing/model': 1.173194782983046, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 0.89}) +[2026-06-23 22:16:59,022][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.12912800908088684, 'train/grad_norm_pre_clip': 14.692988395690918, 'timing/data': 0.0011331639834679663, 'timing/model': 1.1707744910090696, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 0.89}) +[2026-06-23 22:17:03,339][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.06595905870199203, 'train/grad_norm_pre_clip': 42.150428771972656, 'timing/data': 0.0006969179958105087, 'timing/model': 1.1919760479941033, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 0.89}) +[2026-06-23 22:17:07,918][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.054135873913764954, 'train/grad_norm_pre_clip': 24.997926712036133, 'timing/data': 0.0007405889919027686, 'timing/model': 1.4561651590047404, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 0.89}) +[2026-06-23 22:17:12,211][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.11047831922769547, 'train/grad_norm_pre_clip': 26.509355545043945, 'timing/data': 0.0006989320099819452, 'timing/model': 1.172040448000189, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 0.89}) +[2026-06-23 22:17:16,532][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.02629300206899643, 'train/grad_norm_pre_clip': 17.281208038330078, 'timing/data': 0.0007202010019682348, 'timing/model': 1.1741258000256494, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 0.89}) +[2026-06-23 22:17:20,844][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.03988491743803024, 'train/grad_norm_pre_clip': 15.228930473327637, 'timing/data': 0.0007428720127791166, 'timing/model': 1.1777474219852593, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 0.89}) +[2026-06-23 22:17:25,320][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2300 +[2026-06-23 22:18:23,092][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2300 +[2026-06-23 22:18:23,093][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2300 +[2026-06-23 22:18:59,646][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2300 +[2026-06-23 22:18:59,648][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.09797623008489609, 'train/grad_norm_pre_clip': 16.478548049926758, 'eval/action_loss/seconds': 57.76855767497909, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.04589366540312767, 'eval/latency_2/loss': 0.04589366540312767, 'eval/flappy/loss': 0.04589366540312767, 'eval/flappy/latency_2/loss': 0.04589366540312767, 'eval/flap_precision': 0.9404145077720207, 'eval/flap_recall': 0.9850746268656716, 'eval/flap_f1': 0.9622266401590457, 'eval/flap_rate_model': 0.21444444444444444, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9622266401590456, 'eval/latency_2/flap_precision': 0.9404145077720207, 'eval/latency_2/flap_recall': 0.9850746268656716, 'eval/latency_2/flap_f1': 0.9622266401590457, 'eval/latency_2/flap_rate_model': 0.21444444444444444, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9622266401590456, 'eval/action_classification/data_wait_seconds': 0.4027922451496124, 'eval/action_classification/predict_seconds': 36.00027084350586, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007712350052315742, 'timing/model': 1.1769249320204835, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.89}) +[2026-06-23 22:19:03,930][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.06747761368751526, 'train/grad_norm_pre_clip': 35.85171127319336, 'timing/data': 0.0009818249964155257, 'timing/model': 1.1664313360233791, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 0.89}) +[2026-06-23 22:19:08,201][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.19831451773643494, 'train/grad_norm_pre_clip': 29.922409057617188, 'timing/data': 0.0007515189936384559, 'timing/model': 1.1655530220014043, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 0.89}) +[2026-06-23 22:19:12,749][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.08519759774208069, 'train/grad_norm_pre_clip': 23.008054733276367, 'timing/data': 0.000753772008465603, 'timing/model': 1.437690095975995, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 0.89}) +[2026-06-23 22:19:17,089][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.07370409369468689, 'train/grad_norm_pre_clip': 18.82891845703125, 'timing/data': 0.0007495049794670194, 'timing/model': 1.1904416609904729, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 0.89}) +[2026-06-23 22:19:21,381][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.0405125617980957, 'train/grad_norm_pre_clip': 16.84934425354004, 'timing/data': 0.0007711649814154953, 'timing/model': 1.1744479070184752, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.89}) +[2026-06-23 22:19:25,679][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.06067284941673279, 'train/grad_norm_pre_clip': 20.45274543762207, 'timing/data': 0.0008151259971782565, 'timing/model': 1.1713433199911378, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 0.89}) +[2026-06-23 22:19:30,175][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.07130338996648788, 'train/grad_norm_pre_clip': 8.58742618560791, 'timing/data': 0.0007486929825972766, 'timing/model': 1.1914271509740502, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 0.89}) +[2026-06-23 22:19:34,494][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.020747827365994453, 'train/grad_norm_pre_clip': 12.302770614624023, 'timing/data': 0.000710011983755976, 'timing/model': 1.1991315159830265, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 0.89}) +[2026-06-23 22:19:38,789][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.020267456769943237, 'train/grad_norm_pre_clip': 14.628199577331543, 'timing/data': 0.0009283660037908703, 'timing/model': 1.176291141018737, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 0.89}) +[2026-06-23 22:19:43,156][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.08263152837753296, 'train/grad_norm_pre_clip': 29.871606826782227, 'timing/data': 0.0007375620189122856, 'timing/model': 1.1942570179817267, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.89}) +[2026-06-23 22:19:47,705][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.04300513118505478, 'train/grad_norm_pre_clip': 10.368644714355469, 'timing/data': 0.0008109089976642281, 'timing/model': 1.1727886679873336, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 0.89}) +[2026-06-23 22:19:52,025][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.1573166847229004, 'train/grad_norm_pre_clip': 21.065021514892578, 'timing/data': 0.0008763289952185005, 'timing/model': 1.1746985659992788, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 0.89}) +[2026-06-23 22:19:56,369][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.06279049813747406, 'train/grad_norm_pre_clip': 14.999343872070312, 'timing/data': 0.0007838080055080354, 'timing/model': 1.197121609991882, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 0.9}) +[2026-06-23 22:20:00,778][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.02734849601984024, 'train/grad_norm_pre_clip': 8.074372291564941, 'timing/data': 0.000979078991804272, 'timing/model': 1.177253836998716, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 0.9}) +[2026-06-23 22:20:05,383][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.06421887129545212, 'train/grad_norm_pre_clip': 9.576605796813965, 'timing/data': 0.0007649740146007389, 'timing/model': 1.1935640849987976, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 0.9}) +[2026-06-23 22:20:09,715][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.0450802817940712, 'train/grad_norm_pre_clip': 24.56539535522461, 'timing/data': 0.0009400470007676631, 'timing/model': 1.1785086739982944, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 0.9}) +[2026-06-23 22:20:14,073][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.11899036169052124, 'train/grad_norm_pre_clip': 17.551755905151367, 'timing/data': 0.0007506870024371892, 'timing/model': 1.1935563290026039, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 0.9}) +[2026-06-23 22:20:18,428][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.1041446328163147, 'train/grad_norm_pre_clip': 14.85885238647461, 'timing/data': 0.0007397070003207773, 'timing/model': 1.1777583770162892, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 0.9}) +[2026-06-23 22:20:22,980][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.0043561807833611965, 'train/grad_norm_pre_clip': 13.121593475341797, 'timing/data': 0.0007190190081018955, 'timing/model': 1.1772636719979346, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 0.9}) +[2026-06-23 22:20:27,317][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.0037871128879487514, 'train/grad_norm_pre_clip': 12.545662879943848, 'timing/data': 0.0007385450007859617, 'timing/model': 1.1944404410023708, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.9}) +[2026-06-23 22:20:31,670][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.035719700157642365, 'train/grad_norm_pre_clip': 28.775461196899414, 'timing/data': 0.0010274579981341958, 'timing/model': 1.1772323329932988, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 0.9}) +[2026-06-23 22:20:36,007][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.22839592397212982, 'train/grad_norm_pre_clip': 30.926485061645508, 'timing/data': 0.0011232560209464282, 'timing/model': 1.1998798819840886, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 0.9}) +[2026-06-23 22:20:40,517][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.038085926324129105, 'train/grad_norm_pre_clip': 17.181106567382812, 'timing/data': 0.0009619280172046274, 'timing/model': 1.1801983829936944, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 0.9}) +[2026-06-23 22:20:44,877][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.03598397225141525, 'train/grad_norm_pre_clip': 11.869112968444824, 'timing/data': 0.0008456930227112025, 'timing/model': 1.1815612110076472, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 0.9}) +[2026-06-23 22:20:49,233][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.008841478265821934, 'train/grad_norm_pre_clip': 19.612993240356445, 'timing/data': 0.0008444800041615963, 'timing/model': 1.1950625579920597, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 0.9}) +[2026-06-23 22:20:53,846][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.2713528275489807, 'train/grad_norm_pre_clip': 38.27825927734375, 'timing/data': 0.0008862270042300224, 'timing/model': 1.460313475021394, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 0.9}) +[2026-06-23 22:20:58,210][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.05778077244758606, 'train/grad_norm_pre_clip': 28.12218475341797, 'timing/data': 0.000720181007636711, 'timing/model': 1.1768618149799295, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 0.9}) +[2026-06-23 22:21:02,581][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.05771970376372337, 'train/grad_norm_pre_clip': 15.766244888305664, 'timing/data': 0.000748913997085765, 'timing/model': 1.200789538997924, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 0.9}) +[2026-06-23 22:21:06,938][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.03009508177638054, 'train/grad_norm_pre_clip': 9.401091575622559, 'timing/data': 0.0007598840165883303, 'timing/model': 1.197718534996966, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 0.9}) +[2026-06-23 22:21:11,586][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.08234698325395584, 'train/grad_norm_pre_clip': 64.19760131835938, 'timing/data': 0.0010294019884895533, 'timing/model': 1.4621752729872242, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.9}) +[2026-06-23 22:21:15,922][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.2519263029098511, 'train/grad_norm_pre_clip': 40.03667449951172, 'timing/data': 0.0007985550037119538, 'timing/model': 1.2036306069931015, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 0.9}) +[2026-06-23 22:21:20,236][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.24702298641204834, 'train/grad_norm_pre_clip': 22.19877052307129, 'timing/data': 0.0007937769987620413, 'timing/model': 1.1776179510052316, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 0.9}) +[2026-06-23 22:21:24,546][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.1010592132806778, 'train/grad_norm_pre_clip': 15.582982063293457, 'timing/data': 0.0008684149943292141, 'timing/model': 1.177247736020945, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 0.9}) +[2026-06-23 22:21:29,054][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.22450441122055054, 'train/grad_norm_pre_clip': 34.07858657836914, 'timing/data': 0.0008190130174625665, 'timing/model': 1.1730582239979412, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 0.9}) +[2026-06-23 22:21:33,380][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.028950028121471405, 'train/grad_norm_pre_clip': 22.896543502807617, 'timing/data': 0.0007760940061416477, 'timing/model': 1.174933790025534, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 0.9}) +[2026-06-23 22:21:37,704][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.006642629392445087, 'train/grad_norm_pre_clip': 18.74972152709961, 'timing/data': 0.001163380016805604, 'timing/model': 1.1745155160024296, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 0.9}) +[2026-06-23 22:21:42,041][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.054916296154260635, 'train/grad_norm_pre_clip': 11.426156044006348, 'timing/data': 0.0007197099912445992, 'timing/model': 1.1753111360012554, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 0.9}) +[2026-06-23 22:21:46,522][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.02990572527050972, 'train/grad_norm_pre_clip': 13.69816780090332, 'timing/data': 0.0007301690056920052, 'timing/model': 1.1768051050021313, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 0.9}) +[2026-06-23 22:21:50,838][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.0086055938154459, 'train/grad_norm_pre_clip': 16.775272369384766, 'timing/data': 0.0007685199961997569, 'timing/model': 1.1760335089929868, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 0.91}) +[2026-06-23 22:21:55,171][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.1376822292804718, 'train/grad_norm_pre_clip': 16.300437927246094, 'timing/data': 0.001084744988474995, 'timing/model': 1.1959114300261717, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.91}) +[2026-06-23 22:21:59,515][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.12596380710601807, 'train/grad_norm_pre_clip': 27.228811264038086, 'timing/data': 0.0009994069987442344, 'timing/model': 1.194189758010907, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 0.91}) +[2026-06-23 22:22:04,157][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.04619687795639038, 'train/grad_norm_pre_clip': 13.715316772460938, 'timing/data': 0.0007596930081490427, 'timing/model': 1.1985492229869124, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 0.91}) +[2026-06-23 22:22:08,554][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.005976464599370956, 'train/grad_norm_pre_clip': 6.99585485458374, 'timing/data': 0.0011606240004766732, 'timing/model': 1.208224320987938, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 0.91}) +[2026-06-23 22:22:12,919][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.12265517562627792, 'train/grad_norm_pre_clip': 18.24962043762207, 'timing/data': 0.000809896009741351, 'timing/model': 1.2063529419829138, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 0.91}) +[2026-06-23 22:22:17,294][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.0442827083170414, 'train/grad_norm_pre_clip': 17.323856353759766, 'timing/data': 0.0007443749927915633, 'timing/model': 1.2029883549839724, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 0.91}) +[2026-06-23 22:22:21,827][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.04927336424589157, 'train/grad_norm_pre_clip': 15.899195671081543, 'timing/data': 0.000951438007177785, 'timing/model': 1.1863567119871732, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 0.91}) +[2026-06-23 22:22:26,180][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.014475779607892036, 'train/grad_norm_pre_clip': 14.645796775817871, 'timing/data': 0.0007741809822618961, 'timing/model': 1.1856953789829277, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 0.91}) +[2026-06-23 22:22:30,521][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.011039311066269875, 'train/grad_norm_pre_clip': 12.114422798156738, 'timing/data': 0.0007144499977584928, 'timing/model': 1.1809458429925144, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 0.91}) +[2026-06-23 22:22:34,841][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.046374548226594925, 'train/grad_norm_pre_clip': 14.684782028198242, 'timing/data': 0.0007245790038723499, 'timing/model': 1.1781320429872721, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 0.91}) +[2026-06-23 22:22:39,361][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.07566726207733154, 'train/grad_norm_pre_clip': 18.545047760009766, 'timing/data': 0.001156496990006417, 'timing/model': 1.3788251660007518, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.91}) +[2026-06-23 22:22:43,689][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.18421052396297455, 'train/grad_norm_pre_clip': 21.387096405029297, 'timing/data': 0.0007178370142355561, 'timing/model': 1.175748998997733, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 0.91}) +[2026-06-23 22:22:48,015][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.02050500363111496, 'train/grad_norm_pre_clip': 15.715446472167969, 'timing/data': 0.0009502159955445677, 'timing/model': 1.1799561330117285, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 0.91}) +[2026-06-23 22:22:52,329][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.06924306601285934, 'train/grad_norm_pre_clip': 19.890432357788086, 'timing/data': 0.0008390600269194692, 'timing/model': 1.1767489250050858, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 0.91}) +[2026-06-23 22:22:56,649][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.011369607411324978, 'train/grad_norm_pre_clip': 6.8445515632629395, 'timing/data': 0.000851482996949926, 'timing/model': 1.1781484719831496, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.91}) +[2026-06-23 22:23:01,206][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.030921272933483124, 'train/grad_norm_pre_clip': 8.905142784118652, 'timing/data': 0.0007549850270152092, 'timing/model': 1.1764281920040958, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 0.91}) +[2026-06-23 22:23:05,556][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.037433117628097534, 'train/grad_norm_pre_clip': 36.47753143310547, 'timing/data': 0.0007173860212787986, 'timing/model': 1.2018375339976046, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 0.91}) +[2026-06-23 22:23:09,898][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.035543620586395264, 'train/grad_norm_pre_clip': 9.830617904663086, 'timing/data': 0.0006969079840928316, 'timing/model': 1.1795503000030294, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 0.91}) +[2026-06-23 22:23:14,294][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.07158719003200531, 'train/grad_norm_pre_clip': 11.178709983825684, 'timing/data': 0.0010804059857036918, 'timing/model': 1.2045267119829077, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 0.91}) +[2026-06-23 22:23:18,835][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.03559121489524841, 'train/grad_norm_pre_clip': 32.41499328613281, 'timing/data': 0.0009839480044320226, 'timing/model': 1.1834429730079137, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 0.91}) +[2026-06-23 22:23:23,177][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.023758677765727043, 'train/grad_norm_pre_clip': 21.427921295166016, 'timing/data': 0.000999256008071825, 'timing/model': 1.184201623982517, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.91}) +[2026-06-23 22:23:27,566][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.05880620330572128, 'train/grad_norm_pre_clip': 21.604246139526367, 'timing/data': 0.0010505009850021452, 'timing/model': 1.1824779079761356, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 0.91}) +[2026-06-23 22:23:31,953][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.04054156318306923, 'train/grad_norm_pre_clip': 25.685884475708008, 'timing/data': 0.0011159220011904836, 'timing/model': 1.2055782130046282, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 0.91}) +[2026-06-23 22:23:36,515][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.008853999897837639, 'train/grad_norm_pre_clip': 23.046096801757812, 'timing/data': 0.0007509179995395243, 'timing/model': 1.1770686130039394, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 0.91}) +[2026-06-23 22:23:40,832][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.014751800335943699, 'train/grad_norm_pre_clip': 22.62723159790039, 'timing/data': 0.0008055380021687597, 'timing/model': 1.1763942270190455, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 0.91}) +[2026-06-23 22:23:45,209][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.04084458202123642, 'train/grad_norm_pre_clip': 48.469940185546875, 'timing/data': 0.0016305829922202975, 'timing/model': 1.2115830179827753, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 0.92}) +[2026-06-23 22:23:49,566][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.46538010239601135, 'train/grad_norm_pre_clip': 65.37976837158203, 'timing/data': 0.0009792800119612366, 'timing/model': 1.1811093180149328, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 0.92}) +[2026-06-23 22:23:54,131][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.012004412710666656, 'train/grad_norm_pre_clip': 46.116573333740234, 'timing/data': 0.000900844024727121, 'timing/model': 1.2000411900226027, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 0.92}) +[2026-06-23 22:23:58,532][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.11321510374546051, 'train/grad_norm_pre_clip': 16.650041580200195, 'timing/data': 0.000818864005850628, 'timing/model': 1.203245140000945, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 0.92}) +[2026-06-23 22:24:02,895][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.012954373843967915, 'train/grad_norm_pre_clip': 17.300321578979492, 'timing/data': 0.00076879101106897, 'timing/model': 1.1823482060281094, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 0.92}) +[2026-06-23 22:24:07,229][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.12241493165493011, 'train/grad_norm_pre_clip': 17.762847900390625, 'timing/data': 0.0010170800087507814, 'timing/model': 1.1825510410126299, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.92}) +[2026-06-23 22:24:11,744][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.04967061057686806, 'train/grad_norm_pre_clip': 23.180395126342773, 'timing/data': 0.0009584009821992368, 'timing/model': 1.1802830699889455, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 0.92}) +[2026-06-23 22:24:16,096][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.032052744179964066, 'train/grad_norm_pre_clip': 8.755980491638184, 'timing/data': 0.0007840289908926934, 'timing/model': 1.180816225009039, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 0.92}) +[2026-06-23 22:24:20,416][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.001520570833235979, 'train/grad_norm_pre_clip': 11.81139087677002, 'timing/data': 0.0009207109978888184, 'timing/model': 1.178329237998696, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 0.92}) +[2026-06-23 22:24:24,823][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.18495087325572968, 'train/grad_norm_pre_clip': 35.342830657958984, 'timing/data': 0.001373136998154223, 'timing/model': 1.206412980973255, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 0.92}) +[2026-06-23 22:24:29,409][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.07593560963869095, 'train/grad_norm_pre_clip': 14.8453369140625, 'timing/data': 0.0009664160024840385, 'timing/model': 1.2032683010038454, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.92}) +[2026-06-23 22:24:33,740][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.07174839079380035, 'train/grad_norm_pre_clip': 16.008073806762695, 'timing/data': 0.0010784620244521648, 'timing/model': 1.1798358210071456, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 0.92}) +[2026-06-23 22:24:38,146][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.007776608224958181, 'train/grad_norm_pre_clip': 17.1478214263916, 'timing/data': 0.0008885220158845186, 'timing/model': 1.2180136470124125, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 0.92}) +[2026-06-23 22:24:42,543][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.03806060925126076, 'train/grad_norm_pre_clip': 17.675235748291016, 'timing/data': 0.0007985949923750013, 'timing/model': 1.2123363050050102, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 0.92}) +[2026-06-23 22:24:47,145][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.09416162222623825, 'train/grad_norm_pre_clip': 20.964941024780273, 'timing/data': 0.0009672870219219476, 'timing/model': 1.207828635990154, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 0.92}) +[2026-06-23 22:24:51,554][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.023226572200655937, 'train/grad_norm_pre_clip': 9.17768669128418, 'timing/data': 0.0007711159996688366, 'timing/model': 1.1989942609798163, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.92}) +[2026-06-23 22:24:55,918][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.03550463169813156, 'train/grad_norm_pre_clip': 17.1729679107666, 'timing/data': 0.0009578899771440774, 'timing/model': 1.2146857350016944, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 0.92}) +[2026-06-23 22:25:00,283][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.11623404175043106, 'train/grad_norm_pre_clip': 24.098098754882812, 'timing/data': 0.0007523810199927539, 'timing/model': 1.2090054689906538, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 0.92}) +[2026-06-23 22:25:04,833][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.04429760202765465, 'train/grad_norm_pre_clip': 16.408966064453125, 'timing/data': 0.0009846489992924035, 'timing/model': 1.2092135139973834, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 0.92}) +[2026-06-23 22:25:09,186][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.023948680609464645, 'train/grad_norm_pre_clip': 17.229217529296875, 'timing/data': 0.0007898600015323609, 'timing/model': 1.2070638919831254, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 0.92}) +[2026-06-23 22:25:13,559][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.12248078733682632, 'train/grad_norm_pre_clip': 24.975595474243164, 'timing/data': 0.0008696469885762781, 'timing/model': 1.2011963480035774, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 0.92}) +[2026-06-23 22:25:17,956][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.045075323432683945, 'train/grad_norm_pre_clip': 12.220450401306152, 'timing/data': 0.0007318319985643029, 'timing/model': 1.2024424209957942, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 0.92}) +[2026-06-23 22:25:22,551][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.08344905823469162, 'train/grad_norm_pre_clip': 14.345844268798828, 'timing/data': 0.0007395870052278042, 'timing/model': 1.2026318610005546, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 0.92}) +[2026-06-23 22:25:26,934][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.017072411254048347, 'train/grad_norm_pre_clip': 27.944690704345703, 'timing/data': 0.0007610759930685163, 'timing/model': 1.1994747770077083, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 0.92}) +[2026-06-23 22:25:31,361][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.04172096028923988, 'train/grad_norm_pre_clip': 32.94963455200195, 'timing/data': 0.0009499150037299842, 'timing/model': 1.212840944994241, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 0.92}) +[2026-06-23 22:25:35,702][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.03649705648422241, 'train/grad_norm_pre_clip': 25.200817108154297, 'timing/data': 0.0007613670022692531, 'timing/model': 1.1815655140089802, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.92}) +[2026-06-23 22:25:40,265][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.11888454109430313, 'train/grad_norm_pre_clip': 33.23856735229492, 'timing/data': 0.0008914170030038804, 'timing/model': 1.1950631240033545, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 0.93}) +[2026-06-23 22:25:44,604][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.04767141118645668, 'train/grad_norm_pre_clip': 15.090752601623535, 'timing/data': 0.0009541129984427243, 'timing/model': 1.1975241010077298, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 0.93}) +[2026-06-23 22:25:48,944][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.005763302557170391, 'train/grad_norm_pre_clip': 7.740446090698242, 'timing/data': 0.0010244930163025856, 'timing/model': 1.176661060017068, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 0.93}) +[2026-06-23 22:25:53,300][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.20743359625339508, 'train/grad_norm_pre_clip': 25.578723907470703, 'timing/data': 0.0007927750120870769, 'timing/model': 1.1939184069924522, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 0.93}) +[2026-06-23 22:25:57,842][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.06282214820384979, 'train/grad_norm_pre_clip': 9.702357292175293, 'timing/data': 0.0007514489989262074, 'timing/model': 1.2085444339900278, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 0.93}) +[2026-06-23 22:26:02,179][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.0594492144882679, 'train/grad_norm_pre_clip': 22.84112548828125, 'timing/data': 0.001032858999678865, 'timing/model': 1.1989719860139303, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 0.93}) +[2026-06-23 22:26:06,533][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.2714962363243103, 'train/grad_norm_pre_clip': 39.08442306518555, 'timing/data': 0.0008535179949831218, 'timing/model': 1.2011827300011646, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 0.93}) +[2026-06-23 22:26:10,912][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.020372090861201286, 'train/grad_norm_pre_clip': 11.621172904968262, 'timing/data': 0.000742301985155791, 'timing/model': 1.1832739890087396, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 0.93}) +[2026-06-23 22:26:15,447][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.020955707877874374, 'train/grad_norm_pre_clip': 47.308326721191406, 'timing/data': 0.0008696959994267672, 'timing/model': 1.1967746239970438, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 0.93}) +[2026-06-23 22:26:19,838][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2400 +[2026-06-23 22:27:17,508][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2400 +[2026-06-23 22:27:17,509][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2400 +[2026-06-23 22:27:54,303][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2400 +[2026-06-23 22:27:54,305][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.04028303548693657, 'train/grad_norm_pre_clip': 23.244565963745117, 'eval/action_loss/seconds': 57.66522149401135, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.040567733347415924, 'eval/latency_2/loss': 0.040567733347415924, 'eval/flappy/loss': 0.040567733347415924, 'eval/flappy/latency_2/loss': 0.040567733347415924, 'eval/flap_precision': 0.9765193370165746, 'eval/flap_recall': 0.9592944369063772, 'eval/flap_f1': 0.9678302532511978, 'eval/flap_rate_model': 0.2011111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9678302532511978, 'eval/latency_2/flap_precision': 0.9765193370165746, 'eval/latency_2/flap_recall': 0.9592944369063772, 'eval/latency_2/flap_f1': 0.9678302532511978, 'eval/latency_2/flap_rate_model': 0.2011111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9678302532511978, 'eval/action_classification/data_wait_seconds': 0.4407706558704376, 'eval/action_classification/predict_seconds': 36.20994567871094, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007492839940823615, 'timing/model': 1.1801735890039708, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.93}) +[2026-06-23 22:27:58,648][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.00606890581548214, 'train/grad_norm_pre_clip': 32.47909164428711, 'timing/data': 0.0010216970113106072, 'timing/model': 1.1811679220118094, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 0.93}) +[2026-06-23 22:28:03,152][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.009357534348964691, 'train/grad_norm_pre_clip': 26.150968551635742, 'timing/data': 0.0011166729964315891, 'timing/model': 1.1725892960093915, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 0.93}) +[2026-06-23 22:28:07,512][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.032590173184871674, 'train/grad_norm_pre_clip': 22.395601272583008, 'timing/data': 0.0011469190067145973, 'timing/model': 1.190168384986464, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.93}) +[2026-06-23 22:28:11,862][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.09065103530883789, 'train/grad_norm_pre_clip': 23.703258514404297, 'timing/data': 0.0007845689833629876, 'timing/model': 1.1907686269842088, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 0.93}) +[2026-06-23 22:28:16,128][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.009173532947897911, 'train/grad_norm_pre_clip': 12.851680755615234, 'timing/data': 0.0007777870050631464, 'timing/model': 1.1664318690018263, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 0.93}) +[2026-06-23 22:28:20,583][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.05440308153629303, 'train/grad_norm_pre_clip': 17.0981388092041, 'timing/data': 0.0007164730050135404, 'timing/model': 1.1650582739966922, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 0.93}) +[2026-06-23 22:28:24,859][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.08488421887159348, 'train/grad_norm_pre_clip': 16.953336715698242, 'timing/data': 0.0007133070030249655, 'timing/model': 1.1686060609936249, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 0.93}) +[2026-06-23 22:28:29,141][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.07245588302612305, 'train/grad_norm_pre_clip': 28.184494018554688, 'timing/data': 0.0007017659954726696, 'timing/model': 1.167255839012796, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 0.93}) +[2026-06-23 22:28:33,460][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.011392680928111076, 'train/grad_norm_pre_clip': 28.86526107788086, 'timing/data': 0.0007332540117204189, 'timing/model': 1.1687195049889851, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 0.93}) +[2026-06-23 22:28:38,045][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.07076835632324219, 'train/grad_norm_pre_clip': 20.31533432006836, 'timing/data': 0.0012043549795635045, 'timing/model': 1.1883905899885576, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.93}) +[2026-06-23 22:28:42,407][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.028578342869877815, 'train/grad_norm_pre_clip': 8.304854393005371, 'timing/data': 0.0010613909980747849, 'timing/model': 1.1886146860197186, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 0.93}) +[2026-06-23 22:28:46,759][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.023518191650509834, 'train/grad_norm_pre_clip': 18.65767478942871, 'timing/data': 0.0007189080060925335, 'timing/model': 1.1937028700194787, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 0.93}) +[2026-06-23 22:28:51,125][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.04809704050421715, 'train/grad_norm_pre_clip': 12.634366035461426, 'timing/data': 0.0008868990116752684, 'timing/model': 1.197767120989738, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 0.93}) +[2026-06-23 22:28:55,609][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.019661985337734222, 'train/grad_norm_pre_clip': 11.282622337341309, 'timing/data': 0.0007104830001480877, 'timing/model': 1.1736165350012016, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 0.93}) +[2026-06-23 22:28:59,927][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.04073058068752289, 'train/grad_norm_pre_clip': 25.253416061401367, 'timing/data': 0.0007131070015020669, 'timing/model': 1.1739376409968827, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 0.93}) +[2026-06-23 22:29:04,273][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.08013255894184113, 'train/grad_norm_pre_clip': 12.12894058227539, 'timing/data': 0.0007100319780874997, 'timing/model': 1.1774828229972627, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 0.94}) +[2026-06-23 22:29:08,594][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.1177472248673439, 'train/grad_norm_pre_clip': 47.79076385498047, 'timing/data': 0.0007250500202644616, 'timing/model': 1.1931840330071282, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 0.94}) +[2026-06-23 22:29:13,119][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.012085535563528538, 'train/grad_norm_pre_clip': 31.551393508911133, 'timing/data': 0.0007204709982033819, 'timing/model': 1.3963214289979078, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 0.94}) +[2026-06-23 22:29:17,438][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.12187362462282181, 'train/grad_norm_pre_clip': 17.94578742980957, 'timing/data': 0.0011308300017844886, 'timing/model': 1.1736759079794865, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 0.94}) +[2026-06-23 22:29:21,740][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.04726371914148331, 'train/grad_norm_pre_clip': 32.736812591552734, 'timing/data': 0.0007114850159268826, 'timing/model': 1.1735798729932867, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.94}) +[2026-06-23 22:29:26,088][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.04419916868209839, 'train/grad_norm_pre_clip': 12.51292896270752, 'timing/data': 0.0008599590219091624, 'timing/model': 1.19609292698442, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 0.94}) +[2026-06-23 22:29:30,424][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.15213525295257568, 'train/grad_norm_pre_clip': 25.80250358581543, 'timing/data': 0.000797522981883958, 'timing/model': 1.1862092599913012, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 0.94}) +[2026-06-23 22:29:35,086][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.12970668077468872, 'train/grad_norm_pre_clip': 24.861352920532227, 'timing/data': 0.0011543520085979253, 'timing/model': 1.2045810440031346, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 0.94}) +[2026-06-23 22:29:39,448][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.03524200990796089, 'train/grad_norm_pre_clip': 27.38448715209961, 'timing/data': 0.0008076429949142039, 'timing/model': 1.2082868430006783, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 0.94}) +[2026-06-23 22:29:43,844][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.005628828424960375, 'train/grad_norm_pre_clip': 23.02545928955078, 'timing/data': 0.0007752019737381488, 'timing/model': 1.2009837329969741, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 0.94}) +[2026-06-23 22:29:48,210][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.17678514122962952, 'train/grad_norm_pre_clip': 24.466175079345703, 'timing/data': 0.0009790589974727482, 'timing/model': 1.1771448830259033, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 0.94}) +[2026-06-23 22:29:52,746][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.021917439997196198, 'train/grad_norm_pre_clip': 23.572298049926758, 'timing/data': 0.000728325016098097, 'timing/model': 1.1793654370121658, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 0.94}) +[2026-06-23 22:29:57,140][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.0609743669629097, 'train/grad_norm_pre_clip': 15.408295631408691, 'timing/data': 0.0010081820073537529, 'timing/model': 1.209953230980318, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 0.94}) +[2026-06-23 22:30:01,522][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.039280496537685394, 'train/grad_norm_pre_clip': 19.681354522705078, 'timing/data': 0.0007571190071757883, 'timing/model': 1.1798534740228206, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 0.94}) +[2026-06-23 22:30:05,895][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.2974622845649719, 'train/grad_norm_pre_clip': 34.63199234008789, 'timing/data': 0.0011046510189771652, 'timing/model': 1.1799844579945784, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 0.94}) +[2026-06-23 22:30:10,441][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.0804905965924263, 'train/grad_norm_pre_clip': 18.459463119506836, 'timing/data': 0.000721824006177485, 'timing/model': 1.1811332229990512, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 0.94}) +[2026-06-23 22:30:14,811][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.050270870327949524, 'train/grad_norm_pre_clip': 11.076415061950684, 'timing/data': 0.0009457569976802915, 'timing/model': 1.1987878220097627, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 0.94}) +[2026-06-23 22:30:19,192][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.010199256241321564, 'train/grad_norm_pre_clip': 4.533092021942139, 'timing/data': 0.0012782219855580479, 'timing/model': 1.206622439000057, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 0.94}) +[2026-06-23 22:30:23,572][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.08580497652292252, 'train/grad_norm_pre_clip': 34.890350341796875, 'timing/data': 0.0007481520005967468, 'timing/model': 1.2002554979990236, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 0.94}) +[2026-06-23 22:30:28,160][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.011645056307315826, 'train/grad_norm_pre_clip': 12.035541534423828, 'timing/data': 0.0006965859793126583, 'timing/model': 1.2061606599891093, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 0.94}) +[2026-06-23 22:30:32,505][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.07653435319662094, 'train/grad_norm_pre_clip': 21.00504493713379, 'timing/data': 0.0008010599995031953, 'timing/model': 1.1860247809963766, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 0.94}) +[2026-06-23 22:30:36,872][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.031201565638184547, 'train/grad_norm_pre_clip': 12.307108879089355, 'timing/data': 0.000750706996768713, 'timing/model': 1.2056797250115778, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 0.94}) +[2026-06-23 22:30:41,250][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.05149363726377487, 'train/grad_norm_pre_clip': 7.421854496002197, 'timing/data': 0.0008993619994726032, 'timing/model': 1.2118711770162918, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 0.94}) +[2026-06-23 22:30:45,802][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.1438337117433548, 'train/grad_norm_pre_clip': 17.568809509277344, 'timing/data': 0.0007437139865942299, 'timing/model': 1.1966507979959715, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 0.94}) +[2026-06-23 22:30:50,185][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.22412779927253723, 'train/grad_norm_pre_clip': 23.87088394165039, 'timing/data': 0.0009853910014498979, 'timing/model': 1.2069378370069899, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 0.94}) +[2026-06-23 22:30:54,534][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.039134684950113297, 'train/grad_norm_pre_clip': 15.138760566711426, 'timing/data': 0.00105068099219352, 'timing/model': 1.2068878159916494, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 0.94}) +[2026-06-23 22:30:58,909][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.010152819566428661, 'train/grad_norm_pre_clip': 20.933805465698242, 'timing/data': 0.0007970430015120655, 'timing/model': 1.1766686670016497, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 0.95}) +[2026-06-23 22:31:03,441][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.005692878272384405, 'train/grad_norm_pre_clip': 23.22721290588379, 'timing/data': 0.0008010400051716715, 'timing/model': 1.1777294600033201, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 0.95}) +[2026-06-23 22:31:07,754][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.011442881077528, 'train/grad_norm_pre_clip': 19.55117416381836, 'timing/data': 0.0012214670132379979, 'timing/model': 1.1771901040046941, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 0.95}) +[2026-06-23 22:31:12,072][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.10074248909950256, 'train/grad_norm_pre_clip': 20.238340377807617, 'timing/data': 0.0008013100014068186, 'timing/model': 1.1791211939998902, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.95}) +[2026-06-23 22:31:16,388][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.12705950438976288, 'train/grad_norm_pre_clip': 23.777381896972656, 'timing/data': 0.0007854310097172856, 'timing/model': 1.1786609959963243, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 0.95}) +[2026-06-23 22:31:20,930][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.020645754411816597, 'train/grad_norm_pre_clip': 22.596895217895508, 'timing/data': 0.0007243879954330623, 'timing/model': 1.1955560019996483, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 0.95}) +[2026-06-23 22:31:25,285][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.09873366355895996, 'train/grad_norm_pre_clip': 20.296905517578125, 'timing/data': 0.0009322829719167203, 'timing/model': 1.1792335479985923, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 0.95}) +[2026-06-23 22:31:29,611][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.037974968552589417, 'train/grad_norm_pre_clip': 26.46582794189453, 'timing/data': 0.0008990009955596179, 'timing/model': 1.1750578150094952, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 0.95}) +[2026-06-23 22:31:33,931][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.33925583958625793, 'train/grad_norm_pre_clip': 31.17089080810547, 'timing/data': 0.0008848040015436709, 'timing/model': 1.1769032060110476, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 0.95}) +[2026-06-23 22:31:38,475][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.023821420967578888, 'train/grad_norm_pre_clip': 19.50628089904785, 'timing/data': 0.0007766649941913784, 'timing/model': 1.2028535929857753, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 0.95}) +[2026-06-23 22:31:42,870][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.01794230379164219, 'train/grad_norm_pre_clip': 36.140052795410156, 'timing/data': 0.0007340559968724847, 'timing/model': 1.200274859991623, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.95}) +[2026-06-23 22:31:47,214][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.005577710457146168, 'train/grad_norm_pre_clip': 16.087631225585938, 'timing/data': 0.0007198390085250139, 'timing/model': 1.2001549699925818, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 0.95}) +[2026-06-23 22:31:51,582][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.03242657333612442, 'train/grad_norm_pre_clip': 28.963958740234375, 'timing/data': 0.0007067460101097822, 'timing/model': 1.1999334830034059, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 0.95}) +[2026-06-23 22:31:56,100][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.00787892285734415, 'train/grad_norm_pre_clip': 44.3555908203125, 'timing/data': 0.0007904999947641045, 'timing/model': 1.1771162289951462, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 0.95}) +[2026-06-23 22:32:00,420][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.05652507022023201, 'train/grad_norm_pre_clip': 19.557302474975586, 'timing/data': 0.0008989810012280941, 'timing/model': 1.1779608900251333, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 0.95}) +[2026-06-23 22:32:04,735][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.02974681556224823, 'train/grad_norm_pre_clip': 17.465181350708008, 'timing/data': 0.0008101070125121623, 'timing/model': 1.1788235970016103, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 0.95}) +[2026-06-23 22:32:09,055][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.12202524393796921, 'train/grad_norm_pre_clip': 30.333833694458008, 'timing/data': 0.0007536020129919052, 'timing/model': 1.180392867012415, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 0.95}) +[2026-06-23 22:32:13,656][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.09430084377527237, 'train/grad_norm_pre_clip': 27.019132614135742, 'timing/data': 0.0009377920068800449, 'timing/model': 1.177593447995605, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 0.95}) +[2026-06-23 22:32:17,997][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.11651671677827835, 'train/grad_norm_pre_clip': 60.97294616699219, 'timing/data': 0.0009655340109020472, 'timing/model': 1.1809927700087428, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 0.95}) +[2026-06-23 22:32:22,368][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.02487066388130188, 'train/grad_norm_pre_clip': 12.265893936157227, 'timing/data': 0.0008906449947971851, 'timing/model': 1.1846530539914966, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 0.95}) +[2026-06-23 22:32:26,732][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.04472928121685982, 'train/grad_norm_pre_clip': 22.30615997314453, 'timing/data': 0.0010560819937381893, 'timing/model': 1.1961340679845307, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 0.95}) +[2026-06-23 22:32:31,263][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.05343623831868172, 'train/grad_norm_pre_clip': 21.711883544921875, 'timing/data': 0.0007341860036831349, 'timing/model': 1.1788067460001912, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 0.95}) +[2026-06-23 22:32:35,603][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.03818714991211891, 'train/grad_norm_pre_clip': 10.01867389678955, 'timing/data': 0.0008545690216124058, 'timing/model': 1.1784727009944618, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 0.95}) +[2026-06-23 22:32:39,972][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.05638474225997925, 'train/grad_norm_pre_clip': 16.657276153564453, 'timing/data': 0.0007286760082934052, 'timing/model': 1.1815331160032656, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 0.95}) +[2026-06-23 22:32:44,338][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.042583223432302475, 'train/grad_norm_pre_clip': 14.657242774963379, 'timing/data': 0.0007170440221671015, 'timing/model': 1.1986405450152233, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 0.95}) +[2026-06-23 22:32:48,862][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.016568074002861977, 'train/grad_norm_pre_clip': 5.748142719268799, 'timing/data': 0.0007456679886672646, 'timing/model': 1.2070583299791906, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 0.95}) +[2026-06-23 22:32:53,203][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.0050095790065824986, 'train/grad_norm_pre_clip': 7.153530120849609, 'timing/data': 0.000747050013160333, 'timing/model': 1.1994492020166945, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 0.96}) +[2026-06-23 22:32:57,530][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.07790657132863998, 'train/grad_norm_pre_clip': 23.00210952758789, 'timing/data': 0.0007351780077442527, 'timing/model': 1.1826989269757178, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 0.96}) +[2026-06-23 22:33:01,885][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.126663476228714, 'train/grad_norm_pre_clip': 25.865571975708008, 'timing/data': 0.0007262919971253723, 'timing/model': 1.194165182998404, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 0.96}) +[2026-06-23 22:33:06,437][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.2884162962436676, 'train/grad_norm_pre_clip': 29.752063751220703, 'timing/data': 0.0007333250250667334, 'timing/model': 1.2009798149811104, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 0.96}) +[2026-06-23 22:33:10,758][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.04417281597852707, 'train/grad_norm_pre_clip': 17.46556282043457, 'timing/data': 0.000726491998648271, 'timing/model': 1.1793290349887684, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 0.96}) +[2026-06-23 22:33:15,095][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.0027763734105974436, 'train/grad_norm_pre_clip': 18.230669021606445, 'timing/data': 0.0008466540020890534, 'timing/model': 1.1779716889723204, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 0.96}) +[2026-06-23 22:33:19,457][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.02195245772600174, 'train/grad_norm_pre_clip': 16.9536190032959, 'timing/data': 0.0007214730139821768, 'timing/model': 1.200107698008651, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 0.96}) +[2026-06-23 22:33:23,994][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.059713080525398254, 'train/grad_norm_pre_clip': 8.824627876281738, 'timing/data': 0.000731953012291342, 'timing/model': 1.1952862410107628, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 0.96}) +[2026-06-23 22:33:28,390][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.03142669424414635, 'train/grad_norm_pre_clip': 10.803976058959961, 'timing/data': 0.000882680993527174, 'timing/model': 1.209001362993149, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 0.96}) +[2026-06-23 22:33:32,753][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.08277755230665207, 'train/grad_norm_pre_clip': 25.229644775390625, 'timing/data': 0.0007093909953255206, 'timing/model': 1.218381815997418, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 0.96}) +[2026-06-23 22:33:37,097][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.08098799735307693, 'train/grad_norm_pre_clip': 40.384117126464844, 'timing/data': 0.0007519999926444143, 'timing/model': 1.1778100279916544, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 0.96}) +[2026-06-23 22:33:41,665][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.03244126960635185, 'train/grad_norm_pre_clip': 13.696638107299805, 'timing/data': 0.0009390150080434978, 'timing/model': 1.2021254349965602, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 0.96}) +[2026-06-23 22:33:46,015][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.12118496745824814, 'train/grad_norm_pre_clip': 21.53145408630371, 'timing/data': 0.000733744993340224, 'timing/model': 1.179592442989815, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.96}) +[2026-06-23 22:33:50,397][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.13846459984779358, 'train/grad_norm_pre_clip': 13.284679412841797, 'timing/data': 0.0007712850056122988, 'timing/model': 1.1993087130249478, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 0.96}) +[2026-06-23 22:33:54,795][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.09263554960489273, 'train/grad_norm_pre_clip': 20.636320114135742, 'timing/data': 0.000746558973332867, 'timing/model': 1.2086099080042914, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 0.96}) +[2026-06-23 22:33:59,365][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.2967854142189026, 'train/grad_norm_pre_clip': 20.809194564819336, 'timing/data': 0.0007227649912238121, 'timing/model': 1.1971182379929814, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 0.96}) +[2026-06-23 22:34:03,709][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.013720395043492317, 'train/grad_norm_pre_clip': 13.273853302001953, 'timing/data': 0.0007702229777351022, 'timing/model': 1.1795144549978431, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 0.96}) +[2026-06-23 22:34:08,048][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.025563227012753487, 'train/grad_norm_pre_clip': 13.565452575683594, 'timing/data': 0.0007054329907987267, 'timing/model': 1.1970634709869046, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 0.96}) +[2026-06-23 22:34:12,417][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.0167769193649292, 'train/grad_norm_pre_clip': 37.24742126464844, 'timing/data': 0.0007259109988808632, 'timing/model': 1.1975286809902173, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 0.96}) +[2026-06-23 22:34:16,910][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.05300280079245567, 'train/grad_norm_pre_clip': 25.840072631835938, 'timing/data': 0.000927743996726349, 'timing/model': 1.1764141320018098, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 0.96}) +[2026-06-23 22:34:21,269][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.1382574886083603, 'train/grad_norm_pre_clip': 19.140138626098633, 'timing/data': 0.0008337209874298424, 'timing/model': 1.1972949820046779, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 0.96}) +[2026-06-23 22:34:25,663][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.03473988547921181, 'train/grad_norm_pre_clip': 24.96372413635254, 'timing/data': 0.0007179669919423759, 'timing/model': 1.2547878269979265, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 0.96}) +[2026-06-23 22:34:29,983][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.01869754120707512, 'train/grad_norm_pre_clip': 17.604143142700195, 'timing/data': 0.0009081180032808334, 'timing/model': 1.1786810039775446, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 0.96}) +[2026-06-23 22:34:34,523][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.11342956870794296, 'train/grad_norm_pre_clip': 23.241779327392578, 'timing/data': 0.0007013659924268723, 'timing/model': 1.2006488330080174, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 0.96}) +[2026-06-23 22:34:38,847][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.015887679532170296, 'train/grad_norm_pre_clip': 12.371994972229004, 'timing/data': 0.0011256199795752764, 'timing/model': 1.1795731449965388, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 0.96}) +[2026-06-23 22:34:43,271][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.05656927451491356, 'train/grad_norm_pre_clip': 11.433856010437012, 'timing/data': 0.0008317969914060086, 'timing/model': 1.1992706390155945, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 0.96}) +[2026-06-23 22:34:47,664][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.1431848555803299, 'train/grad_norm_pre_clip': 30.60028648376465, 'timing/data': 0.0011007339926436543, 'timing/model': 1.216224163013976, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 0.97}) +[2026-06-23 22:34:52,187][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.07430080324411392, 'train/grad_norm_pre_clip': 22.510234832763672, 'timing/data': 0.0008194539987016469, 'timing/model': 1.1825057769892737, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 0.97}) +[2026-06-23 22:34:56,549][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.057105615735054016, 'train/grad_norm_pre_clip': 14.175740242004395, 'timing/data': 0.0008118900004774332, 'timing/model': 1.2103715029952582, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 0.97}) +[2026-06-23 22:35:00,928][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.022313782945275307, 'train/grad_norm_pre_clip': 28.660682678222656, 'timing/data': 0.0007855609874241054, 'timing/model': 1.1984879590163473, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 0.97}) +[2026-06-23 22:35:05,300][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.008820735849440098, 'train/grad_norm_pre_clip': 17.083362579345703, 'timing/data': 0.0007831069815438241, 'timing/model': 1.2029645610018633, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 0.97}) +[2026-06-23 22:35:09,882][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.0760180652141571, 'train/grad_norm_pre_clip': 33.31233215332031, 'timing/data': 0.0010871790000237525, 'timing/model': 1.200493776006624, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 0.97}) +[2026-06-23 22:35:14,223][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2500 +[2026-06-23 22:36:11,762][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2500 +[2026-06-23 22:36:11,763][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2500 +[2026-06-23 22:36:48,760][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2500 +[2026-06-23 22:57:27,306][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-06-23 22:57:27,308][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.03133014217019081, 'train/grad_norm_pre_clip': 22.227060317993164, 'eval/action_loss/seconds': 57.535837267991155, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03589066118001938, 'eval/latency_2/loss': 0.03589066118001938, 'eval/flappy/loss': 0.03589066118001938, 'eval/flappy/latency_2/loss': 0.03589066118001938, 'eval/flap_precision': 0.9739726027397261, 'eval/flap_recall': 0.9647218453188603, 'eval/flap_f1': 0.9693251533742332, 'eval/flap_rate_model': 0.20277777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9693251533742332, 'eval/latency_2/flap_precision': 0.9739726027397261, 'eval/latency_2/flap_recall': 0.9647218453188603, 'eval/latency_2/flap_f1': 0.9693251533742332, 'eval/latency_2/flap_rate_model': 0.20277777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9693251533742332, 'eval/action_classification/data_wait_seconds': 0.4666627049446106, 'eval/action_classification/predict_seconds': 36.38654708862305, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 322.98000000001014, 'rl_games_eval/mid_train/mean_length': 2622.6, 'rl_games_eval/mid_train/std_reward': 121.9776602497484, 'rl_games_eval/mid_train/std_length': 981.7129111914542, 'rl_games_eval/mid_train/macro_mean_reward': 322.98000000001014, 'rl_games_eval/mid_train/macro_mean_length': 2622.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 322.98000000001014, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2622.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 121.9776602497484, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 981.7129111914542, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 322.98000000001014, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0007177770021371543, 'timing/model': 1.1984421469969675, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 0.97}) +[2026-06-23 22:57:27,308][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_2500_state +[2026-06-23 22:57:27,316][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 22:57:36,122][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_2500_state/model.safetensors +[2026-06-23 22:57:46,401][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_2500_state/optimizer.bin +[2026-06-23 22:57:46,402][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_2500_state/sampler.bin +[2026-06-23 22:57:46,402][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-23 22:57:46,403][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-23 22:57:46,600][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-23 22:57:46,600][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 22:57:46,624][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 22:57:50,936][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.0447256974875927, 'train/grad_norm_pre_clip': 9.369562149047852, 'timing/data': 0.0008030030003283173, 'timing/model': 1.1638714440050535, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 0.97}) +[2026-06-23 22:57:55,158][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.012374081648886204, 'train/grad_norm_pre_clip': 11.145659446716309, 'timing/data': 0.001435984013369307, 'timing/model': 1.1560899059986696, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 0.97}) +[2026-06-23 22:57:59,595][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.003978493623435497, 'train/grad_norm_pre_clip': 10.510002136230469, 'timing/data': 0.0011991659994237125, 'timing/model': 1.4094690080091823, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 0.97}) +[2026-06-23 22:58:03,797][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.030882004648447037, 'train/grad_norm_pre_clip': 8.060349464416504, 'timing/data': 0.0009852910006884485, 'timing/model': 1.1487835760053713, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 0.97}) +[2026-06-23 22:58:08,019][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.030949164181947708, 'train/grad_norm_pre_clip': 13.250825881958008, 'timing/data': 0.0009358600073028356, 'timing/model': 1.1645449080097023, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 0.97}) +[2026-06-23 22:58:12,228][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.02135157585144043, 'train/grad_norm_pre_clip': 15.546477317810059, 'timing/data': 0.0009774659993126988, 'timing/model': 1.14716020601918, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 0.97}) +[2026-06-23 22:58:16,616][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.003312589367851615, 'train/grad_norm_pre_clip': 9.235148429870605, 'timing/data': 0.0008093259821180254, 'timing/model': 1.3560727580043022, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 0.97}) +[2026-06-23 22:58:20,789][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.01677628420293331, 'train/grad_norm_pre_clip': 10.040392875671387, 'timing/data': 0.0010419359896332026, 'timing/model': 1.1402045039867517, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 0.97}) +[2026-06-23 22:58:24,969][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.02602413482964039, 'train/grad_norm_pre_clip': 67.38603210449219, 'timing/data': 0.000988576008239761, 'timing/model': 1.1445273919962347, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 0.97}) +[2026-06-23 22:58:29,225][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.012563677504658699, 'train/grad_norm_pre_clip': 8.848822593688965, 'timing/data': 0.0011798400082625449, 'timing/model': 1.1744720010028686, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 0.97}) +[2026-06-23 22:58:33,711][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.005038794595748186, 'train/grad_norm_pre_clip': 15.96608829498291, 'timing/data': 0.0007593030168209225, 'timing/model': 1.4254665379994549, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 0.97}) +[2026-06-23 22:58:37,929][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.045300036668777466, 'train/grad_norm_pre_clip': 29.082353591918945, 'timing/data': 0.0007943780219648033, 'timing/model': 1.1496736509725451, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 0.97}) +[2026-06-23 22:58:42,151][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.05689913034439087, 'train/grad_norm_pre_clip': 18.454120635986328, 'timing/data': 0.0008129820053000003, 'timing/model': 1.1498416110116523, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 0.97}) +[2026-06-23 22:58:46,384][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.03731656074523926, 'train/grad_norm_pre_clip': 81.14325714111328, 'timing/data': 0.0007269530033227056, 'timing/model': 1.1531160109734628, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 0.97}) +[2026-06-23 22:58:50,815][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.019571248441934586, 'train/grad_norm_pre_clip': 11.251995086669922, 'timing/data': 0.0007772959943395108, 'timing/model': 1.1520302949938923, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.97}) +[2026-06-23 22:58:55,057][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.07523544877767563, 'train/grad_norm_pre_clip': 22.31773567199707, 'timing/data': 0.0008012810139916837, 'timing/model': 1.157212080026511, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 0.97}) +[2026-06-23 22:58:59,344][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.11778267472982407, 'train/grad_norm_pre_clip': 32.566375732421875, 'timing/data': 0.0007467000104952604, 'timing/model': 1.1764277149923146, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 0.97}) +[2026-06-23 22:59:03,629][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.13620539009571075, 'train/grad_norm_pre_clip': 30.483448028564453, 'timing/data': 0.0007728280033916235, 'timing/model': 1.158966301009059, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 0.97}) +[2026-06-23 22:59:08,187][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.1264645755290985, 'train/grad_norm_pre_clip': 22.15754508972168, 'timing/data': 0.0007172650075517595, 'timing/model': 1.179231786984019, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 0.97}) +[2026-06-23 22:59:12,473][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.1328856199979782, 'train/grad_norm_pre_clip': 16.138500213623047, 'timing/data': 0.0007801020110491663, 'timing/model': 1.1573080079979263, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 0.98}) +[2026-06-23 22:59:16,722][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.026596607640385628, 'train/grad_norm_pre_clip': 17.331727981567383, 'timing/data': 0.0007340669981203973, 'timing/model': 1.1632407800061628, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 0.98}) +[2026-06-23 22:59:20,981][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.2814375162124634, 'train/grad_norm_pre_clip': 48.089359283447266, 'timing/data': 0.0007836779986973852, 'timing/model': 1.1650971300259698, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 0.98}) +[2026-06-23 22:59:25,435][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.10144665837287903, 'train/grad_norm_pre_clip': 45.13241195678711, 'timing/data': 0.0007119350193534046, 'timing/model': 1.1609033120039385, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 0.98}) +[2026-06-23 22:59:29,701][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.0025072137359529734, 'train/grad_norm_pre_clip': 8.235868453979492, 'timing/data': 0.0009468499920330942, 'timing/model': 1.1663522899907548, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 0.98}) +[2026-06-23 22:59:34,020][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.0007774842670187354, 'train/grad_norm_pre_clip': 11.573258399963379, 'timing/data': 0.0007399180030915886, 'timing/model': 1.1891883840144146, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 0.98}) +[2026-06-23 22:59:38,321][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.046930186450481415, 'train/grad_norm_pre_clip': 20.110206604003906, 'timing/data': 0.0009343769925180823, 'timing/model': 1.1683662310242653, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 0.98}) +[2026-06-23 22:59:42,853][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.10996644198894501, 'train/grad_norm_pre_clip': 13.450510025024414, 'timing/data': 0.0007148399890866131, 'timing/model': 1.40287709000404, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 0.98}) +[2026-06-23 22:59:47,162][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.02668795734643936, 'train/grad_norm_pre_clip': 21.367351531982422, 'timing/data': 0.0008546700119040906, 'timing/model': 1.1701005450158846, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 0.98}) +[2026-06-23 22:59:51,433][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.07418874651193619, 'train/grad_norm_pre_clip': 13.673446655273438, 'timing/data': 0.0007642119890078902, 'timing/model': 1.1659764450159855, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 0.98}) +[2026-06-23 22:59:55,716][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.012865033000707626, 'train/grad_norm_pre_clip': 50.364810943603516, 'timing/data': 0.00075975400977768, 'timing/model': 1.169262878014706, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 0.98}) +[2026-06-23 22:59:59,999][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.13411590456962585, 'train/grad_norm_pre_clip': 16.179697036743164, 'timing/data': 0.0007469200063496828, 'timing/model': 1.1710917570162565, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 0.98}) +[2026-06-23 23:00:04,519][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.034029558300971985, 'train/grad_norm_pre_clip': 23.32394027709961, 'timing/data': 0.0007329740037675947, 'timing/model': 1.1670238210062962, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 0.98}) +[2026-06-23 23:00:08,834][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.20693618059158325, 'train/grad_norm_pre_clip': 49.064815521240234, 'timing/data': 0.001154292986029759, 'timing/model': 1.1952912310080137, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 0.98}) +[2026-06-23 23:00:13,131][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.01974223554134369, 'train/grad_norm_pre_clip': 16.90686798095703, 'timing/data': 0.0007247999892570078, 'timing/model': 1.1755848489992786, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 0.98}) +[2026-06-23 23:00:17,457][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.02548561990261078, 'train/grad_norm_pre_clip': 16.589763641357422, 'timing/data': 0.000707246974343434, 'timing/model': 1.173442861996591, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 0.98}) +[2026-06-23 23:00:21,953][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.06813126802444458, 'train/grad_norm_pre_clip': 23.952835083007812, 'timing/data': 0.0007635410001967102, 'timing/model': 1.1736844669794664, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 0.98}) +[2026-06-23 23:00:26,301][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.11814586818218231, 'train/grad_norm_pre_clip': 19.837238311767578, 'timing/data': 0.0007434329891111702, 'timing/model': 1.1949267480231356, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 0.98}) +[2026-06-23 23:00:30,626][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.04139421880245209, 'train/grad_norm_pre_clip': 13.569015502929688, 'timing/data': 0.0007264019805006683, 'timing/model': 1.1964083060156554, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 0.98}) +[2026-06-23 23:00:34,936][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.0404864065349102, 'train/grad_norm_pre_clip': 14.977079391479492, 'timing/data': 0.0007172359910327941, 'timing/model': 1.1727901849953923, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 0.98}) +[2026-06-23 23:00:39,409][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.20050159096717834, 'train/grad_norm_pre_clip': 32.634765625, 'timing/data': 0.0007300090219359845, 'timing/model': 1.173285450990079, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 0.98}) +[2026-06-23 23:00:43,745][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.001592070795595646, 'train/grad_norm_pre_clip': 12.62240982055664, 'timing/data': 0.0007449959812220186, 'timing/model': 1.1744086960097775, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 0.98}) +[2026-06-23 23:00:48,051][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.05353826284408569, 'train/grad_norm_pre_clip': 17.57498550415039, 'timing/data': 0.0007422009948641062, 'timing/model': 1.1787741720036138, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 0.98}) +[2026-06-23 23:00:52,363][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.017364080995321274, 'train/grad_norm_pre_clip': 27.492172241210938, 'timing/data': 0.0007548340072389692, 'timing/model': 1.1766014169843402, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 0.98}) +[2026-06-23 23:00:56,859][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.0987694039940834, 'train/grad_norm_pre_clip': 21.572561264038086, 'timing/data': 0.0007248590118251741, 'timing/model': 1.176732110005105, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 0.98}) +[2026-06-23 23:01:01,204][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.04862549901008606, 'train/grad_norm_pre_clip': 12.767401695251465, 'timing/data': 0.0007404180068988353, 'timing/model': 1.1781755469855852, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 0.98}) +[2026-06-23 23:01:05,517][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.023727254942059517, 'train/grad_norm_pre_clip': 18.92019271850586, 'timing/data': 0.0007702430011704564, 'timing/model': 1.1763151990016922, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 0.99}) +[2026-06-23 23:01:09,827][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.002433216432109475, 'train/grad_norm_pre_clip': 16.447275161743164, 'timing/data': 0.0007527100096922368, 'timing/model': 1.1750142350210808, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 0.99}) +[2026-06-23 23:01:14,406][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.13737690448760986, 'train/grad_norm_pre_clip': 13.911125183105469, 'timing/data': 0.0007178460073191673, 'timing/model': 1.1931192390038632, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 0.99}) +[2026-06-23 23:01:18,739][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.06300486624240875, 'train/grad_norm_pre_clip': 12.013856887817383, 'timing/data': 0.0007026679813861847, 'timing/model': 1.198750583018409, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 0.99}) +[2026-06-23 23:01:23,086][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.023763742297887802, 'train/grad_norm_pre_clip': 26.204084396362305, 'timing/data': 0.0007233469805214554, 'timing/model': 1.1959748349909205, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.99}) +[2026-06-23 23:01:27,456][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.01959681324660778, 'train/grad_norm_pre_clip': 10.618575096130371, 'timing/data': 0.0008608299831394106, 'timing/model': 1.1947049680165946, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 0.99}) +[2026-06-23 23:01:32,016][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.003919024020433426, 'train/grad_norm_pre_clip': 8.135414123535156, 'timing/data': 0.0008737340103834867, 'timing/model': 1.1757989559846465, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 0.99}) +[2026-06-23 23:01:36,325][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.07387980818748474, 'train/grad_norm_pre_clip': 14.353074073791504, 'timing/data': 0.000738675007596612, 'timing/model': 1.1763565759756602, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 0.99}) +[2026-06-23 23:01:40,686][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.06466616690158844, 'train/grad_norm_pre_clip': 22.23557472229004, 'timing/data': 0.0007580810051877052, 'timing/model': 1.2019930249953177, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 0.99}) +[2026-06-23 23:01:44,996][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.009844550862908363, 'train/grad_norm_pre_clip': 21.896333694458008, 'timing/data': 0.0007537320198025554, 'timing/model': 1.1750522260263097, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 0.99}) +[2026-06-23 23:01:49,583][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.020815107971429825, 'train/grad_norm_pre_clip': 9.619682312011719, 'timing/data': 0.000791361992014572, 'timing/model': 1.195863629982341, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 0.99}) +[2026-06-23 23:01:53,887][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.10213343799114227, 'train/grad_norm_pre_clip': 20.052444458007812, 'timing/data': 0.0008920480031520128, 'timing/model': 1.1747692930221092, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 0.99}) +[2026-06-23 23:01:58,225][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.006737763062119484, 'train/grad_norm_pre_clip': 18.966598510742188, 'timing/data': 0.0007096189947333187, 'timing/model': 1.173633611004334, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 0.99}) +[2026-06-23 23:02:02,580][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.001043084543198347, 'train/grad_norm_pre_clip': 25.68525505065918, 'timing/data': 0.000755524990381673, 'timing/model': 1.1762704249995295, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 0.99}) +[2026-06-23 23:02:07,155][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.06879129260778427, 'train/grad_norm_pre_clip': 23.02522850036621, 'timing/data': 0.0007276629912666976, 'timing/model': 1.1724163189937826, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 0.99}) +[2026-06-23 23:02:11,535][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.010891750454902649, 'train/grad_norm_pre_clip': 28.78890609741211, 'timing/data': 0.001163859007647261, 'timing/model': 1.2044021069887094, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 0.99}) +[2026-06-23 23:02:15,842][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.017490088939666748, 'train/grad_norm_pre_clip': 16.62830352783203, 'timing/data': 0.0007888859836384654, 'timing/model': 1.1751517610100564, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 0.99}) +[2026-06-23 23:02:20,150][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.016152435913681984, 'train/grad_norm_pre_clip': 15.574090003967285, 'timing/data': 0.0009955380228348076, 'timing/model': 1.1769384260114748, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 0.99}) +[2026-06-23 23:02:24,733][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.09262366592884064, 'train/grad_norm_pre_clip': 40.112693786621094, 'timing/data': 0.0007506760011892766, 'timing/model': 1.175635207007872, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 0.99}) +[2026-06-23 23:02:29,100][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.00984159391373396, 'train/grad_norm_pre_clip': 53.06480026245117, 'timing/data': 0.0007744300237391144, 'timing/model': 1.199976865987992, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 0.99}) +[2026-06-23 23:02:33,428][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.05148056894540787, 'train/grad_norm_pre_clip': 26.48113250732422, 'timing/data': 0.0008818780188448727, 'timing/model': 1.1801045199972577, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 0.99}) +[2026-06-23 23:02:37,759][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.028296910226345062, 'train/grad_norm_pre_clip': 11.90985107421875, 'timing/data': 0.00079297402407974, 'timing/model': 1.1826504250057042, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 0.99}) +[2026-06-23 23:02:42,383][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.026990478858351707, 'train/grad_norm_pre_clip': 25.1964111328125, 'timing/data': 0.001158618979388848, 'timing/model': 1.192706044006627, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 0.99}) +[2026-06-23 23:02:46,742][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.04855374991893768, 'train/grad_norm_pre_clip': 23.32942008972168, 'timing/data': 0.0007427310047205538, 'timing/model': 1.1832894879917149, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 0.99}) +[2026-06-23 23:02:51,083][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.18788832426071167, 'train/grad_norm_pre_clip': 18.217124938964844, 'timing/data': 0.0011101489944849163, 'timing/model': 1.1780652890156489, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 0.99}) +[2026-06-23 23:02:55,417][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.10807633399963379, 'train/grad_norm_pre_clip': 35.82942199707031, 'timing/data': 0.0007981130038388073, 'timing/model': 1.1838002110016532, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 1.0}) +[2026-06-23 23:03:00,009][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.10726189613342285, 'train/grad_norm_pre_clip': 27.207622528076172, 'timing/data': 0.0007578390068374574, 'timing/model': 1.181703105015913, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 1.0}) +[2026-06-23 23:03:04,376][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.05662539601325989, 'train/grad_norm_pre_clip': 16.805940628051758, 'timing/data': 0.0007816930010449141, 'timing/model': 1.2032992170134094, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 1.0}) +[2026-06-23 23:03:08,812][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.04867422580718994, 'train/grad_norm_pre_clip': 13.812382698059082, 'timing/data': 0.0009673949971329421, 'timing/model': 1.2028234140016139, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 1.0}) +[2026-06-23 23:03:13,162][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.0016863533528521657, 'train/grad_norm_pre_clip': 27.188501358032227, 'timing/data': 0.0009392139909323305, 'timing/model': 1.1992623269907199, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 1.0}) +[2026-06-23 23:03:17,753][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.12865234911441803, 'train/grad_norm_pre_clip': 30.735349655151367, 'timing/data': 0.0008600680157542229, 'timing/model': 1.1846632520027924, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 1.0}) +[2026-06-23 23:03:22,109][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.003671416314318776, 'train/grad_norm_pre_clip': 17.846084594726562, 'timing/data': 0.0008073399949353188, 'timing/model': 1.1884342480043415, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 1.0}) +[2026-06-23 23:03:26,479][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.14811378717422485, 'train/grad_norm_pre_clip': 36.30809020996094, 'timing/data': 0.0008480550022795796, 'timing/model': 1.1840028210135642, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 1.0}) +[2026-06-23 23:03:30,850][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.02754487656056881, 'train/grad_norm_pre_clip': 14.034634590148926, 'timing/data': 0.0008298119937535375, 'timing/model': 1.210673333000159, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 1.0}) +[2026-06-23 23:03:35,619][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.015246791765093803, 'train/grad_norm_pre_clip': 15.552042007446289, 'timing/data': 0.0008487269806209952, 'timing/model': 1.1856220919871703, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 1.0}) +[2026-06-23 23:03:39,971][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.0716383084654808, 'train/grad_norm_pre_clip': 11.48076343536377, 'timing/data': 0.0007842679915484041, 'timing/model': 1.1901547230081633, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 1.0}) +[2026-06-23 23:03:44,332][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.10108871012926102, 'train/grad_norm_pre_clip': 13.176446914672852, 'timing/data': 0.0007950380095280707, 'timing/model': 1.1914415520150214, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 1.0}) +[2026-06-23 23:03:48,693][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.061858922243118286, 'train/grad_norm_pre_clip': 12.639913558959961, 'timing/data': 0.0008367550035472959, 'timing/model': 1.1887635329912882, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 1.0}) +[2026-06-23 23:03:53,229][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.02322075515985489, 'train/grad_norm_pre_clip': 11.49392032623291, 'timing/data': 0.0004821899929083884, 'timing/model': 0.6675534090027213, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 1.0}) +[2026-06-23 23:03:59,685][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.015244178473949432, 'train/grad_norm_pre_clip': 8.310150146484375, 'timing/data': 0.0007506450056098402, 'timing/model': 1.3004014090111014, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 1.0}) +[2026-06-23 23:04:04,107][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.04805935546755791, 'train/grad_norm_pre_clip': 7.345643043518066, 'timing/data': 0.001145504997111857, 'timing/model': 1.1750895669974852, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 1.0}) +[2026-06-23 23:04:08,480][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.04539353400468826, 'train/grad_norm_pre_clip': 10.685515403747559, 'timing/data': 0.0007168029842432588, 'timing/model': 1.1817159749916755, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 1.0}) +[2026-06-23 23:04:12,993][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.05554649233818054, 'train/grad_norm_pre_clip': 11.221553802490234, 'timing/data': 0.000760875002015382, 'timing/model': 1.1753224339918233, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 1.0}) +[2026-06-23 23:04:17,315][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.038414862006902695, 'train/grad_norm_pre_clip': 10.478857040405273, 'timing/data': 0.0006972980045247823, 'timing/model': 1.1728537890012376, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 1.0}) +[2026-06-23 23:04:21,652][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.03473174571990967, 'train/grad_norm_pre_clip': 19.641490936279297, 'timing/data': 0.0007445249939337373, 'timing/model': 1.1792896540137008, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 1.0}) +[2026-06-23 23:04:25,969][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.005450350698083639, 'train/grad_norm_pre_clip': 9.389591217041016, 'timing/data': 0.0008542170107830316, 'timing/model': 1.1810953799868003, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 1.0}) +[2026-06-23 23:04:30,576][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.01811916194856167, 'train/grad_norm_pre_clip': 17.222347259521484, 'timing/data': 0.0007591909961774945, 'timing/model': 1.2003991039819084, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 1.0}) +[2026-06-23 23:04:34,936][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.0851319283246994, 'train/grad_norm_pre_clip': 33.47776794433594, 'timing/data': 0.0007041000062599778, 'timing/model': 1.2070428820152301, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 1.0}) +[2026-06-23 23:04:39,279][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.08032095432281494, 'train/grad_norm_pre_clip': 20.083492279052734, 'timing/data': 0.0008409430156461895, 'timing/model': 1.1798456329852343, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 1.0}) +[2026-06-23 23:04:43,617][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.00032978624221868813, 'train/grad_norm_pre_clip': 10.457096099853516, 'timing/data': 0.0008798539929557592, 'timing/model': 1.1792104830092285, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 1.0}) +[2026-06-23 23:04:48,201][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.0904177576303482, 'train/grad_norm_pre_clip': 19.966108322143555, 'timing/data': 0.0007183059933595359, 'timing/model': 1.175357621978037, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 1.0}) +[2026-06-23 23:04:52,568][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.011795168742537498, 'train/grad_norm_pre_clip': 21.478248596191406, 'timing/data': 0.000708968989783898, 'timing/model': 1.214419905008981, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 1.01}) +[2026-06-23 23:04:56,888][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.003622835734859109, 'train/grad_norm_pre_clip': 32.91015625, 'timing/data': 0.0008773999870754778, 'timing/model': 1.1794805380050093, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 1.01}) +[2026-06-23 23:05:01,213][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.008241543546319008, 'train/grad_norm_pre_clip': 26.592863082885742, 'timing/data': 0.0009992339764721692, 'timing/model': 1.1811936429876368, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.01}) +[2026-06-23 23:05:05,817][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2600 +[2026-06-23 23:06:03,308][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2600 +[2026-06-23 23:06:03,309][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2600 +[2026-06-23 23:06:40,307][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2600 +[2026-06-23 23:06:40,309][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.0047523258253932, 'train/grad_norm_pre_clip': 15.220561981201172, 'eval/action_loss/seconds': 57.487905113986926, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03940059244632721, 'eval/latency_2/loss': 0.03940059244632721, 'eval/flappy/loss': 0.03940059244632721, 'eval/flappy/latency_2/loss': 0.03940059244632721, 'eval/flap_precision': 0.9602649006622517, 'eval/flap_recall': 0.9837177747625508, 'eval/flap_f1': 0.9718498659517426, 'eval/flap_rate_model': 0.20972222222222223, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9718498659517425, 'eval/latency_2/flap_precision': 0.9602649006622517, 'eval/latency_2/flap_recall': 0.9837177747625508, 'eval/latency_2/flap_f1': 0.9718498659517426, 'eval/latency_2/flap_rate_model': 0.20972222222222223, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9718498659517425, 'eval/action_classification/data_wait_seconds': 0.4483644664287567, 'eval/action_classification/predict_seconds': 36.4000129699707, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007233060023281723, 'timing/model': 1.1938961390114855, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 1.01}) +[2026-06-23 23:06:44,616][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.11102147400379181, 'train/grad_norm_pre_clip': 27.147676467895508, 'timing/data': 0.0006855559768155217, 'timing/model': 1.1652311570069287, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 1.01}) +[2026-06-23 23:06:48,940][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.07052402943372726, 'train/grad_norm_pre_clip': 23.470230102539062, 'timing/data': 0.0007752520032227039, 'timing/model': 1.1694848909974098, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 1.01}) +[2026-06-23 23:06:53,491][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.025707857683300972, 'train/grad_norm_pre_clip': 30.051843643188477, 'timing/data': 0.0007127260032575577, 'timing/model': 1.4419979699887335, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 1.01}) +[2026-06-23 23:06:57,800][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.08882233500480652, 'train/grad_norm_pre_clip': 18.25328826904297, 'timing/data': 0.0007037689792923629, 'timing/model': 1.1858457420021296, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 1.01}) +[2026-06-23 23:07:02,139][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.0790056437253952, 'train/grad_norm_pre_clip': 30.608469009399414, 'timing/data': 0.0009069439838640392, 'timing/model': 1.1972596080158837, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 1.01}) +[2026-06-23 23:07:06,507][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.05710209533572197, 'train/grad_norm_pre_clip': 25.89024543762207, 'timing/data': 0.0008821289811749011, 'timing/model': 1.1657446720055304, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 1.01}) +[2026-06-23 23:07:11,097][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.02746538259088993, 'train/grad_norm_pre_clip': 16.42694091796875, 'timing/data': 0.0007628890161868185, 'timing/model': 1.4529219079995528, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 1.01}) +[2026-06-23 23:07:15,391][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.0076108574867248535, 'train/grad_norm_pre_clip': 17.955915451049805, 'timing/data': 0.0007008340035099536, 'timing/model': 1.1699124879960436, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 1.01}) +[2026-06-23 23:07:19,699][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.020107615739107132, 'train/grad_norm_pre_clip': 11.756301879882812, 'timing/data': 0.0008609300130046904, 'timing/model': 1.171460785990348, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 1.01}) +[2026-06-23 23:07:23,998][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.07399294525384903, 'train/grad_norm_pre_clip': 9.073837280273438, 'timing/data': 0.0007000430196058005, 'timing/model': 1.1728394110105, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 1.01}) +[2026-06-23 23:07:28,607][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.05707751587033272, 'train/grad_norm_pre_clip': 10.003363609313965, 'timing/data': 0.0009536609868519008, 'timing/model': 1.4615394979773555, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 1.01}) +[2026-06-23 23:07:32,977][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.09514769166707993, 'train/grad_norm_pre_clip': 13.777612686157227, 'timing/data': 0.000723765988368541, 'timing/model': 1.191974813991692, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 1.01}) +[2026-06-23 23:07:37,331][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.06367655843496323, 'train/grad_norm_pre_clip': 14.31295108795166, 'timing/data': 0.001402399007929489, 'timing/model': 1.1970446160121355, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 1.01}) +[2026-06-23 23:07:41,667][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.09652374684810638, 'train/grad_norm_pre_clip': 20.553321838378906, 'timing/data': 0.0007860620098654181, 'timing/model': 1.190913233003812, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 1.01}) +[2026-06-23 23:07:46,280][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.09959643334150314, 'train/grad_norm_pre_clip': 16.988008499145508, 'timing/data': 0.0009252679883502424, 'timing/model': 1.4572104280232452, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 1.01}) +[2026-06-23 23:07:50,613][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.08245140314102173, 'train/grad_norm_pre_clip': 15.13252067565918, 'timing/data': 0.001279543008422479, 'timing/model': 1.2061199519957881, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 1.01}) +[2026-06-23 23:07:54,958][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.024291232228279114, 'train/grad_norm_pre_clip': 8.355382919311523, 'timing/data': 0.0007709829951636493, 'timing/model': 1.1819571999949403, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 1.01}) +[2026-06-23 23:07:59,329][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.12164714187383652, 'train/grad_norm_pre_clip': 19.1491641998291, 'timing/data': 0.0007820540049578995, 'timing/model': 1.1946441880136263, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 1.01}) +[2026-06-23 23:08:03,925][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.1083325743675232, 'train/grad_norm_pre_clip': 21.498109817504883, 'timing/data': 0.0009465269977226853, 'timing/model': 1.4361396290187258, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 1.01}) +[2026-06-23 23:08:08,320][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.03757905960083008, 'train/grad_norm_pre_clip': 16.01251792907715, 'timing/data': 0.00119980497402139, 'timing/model': 1.2081563679967076, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 1.01}) +[2026-06-23 23:08:12,657][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.1136632114648819, 'train/grad_norm_pre_clip': 23.831003189086914, 'timing/data': 0.0007298170239664614, 'timing/model': 1.179032123007346, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 1.01}) +[2026-06-23 23:08:17,012][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.06140889227390289, 'train/grad_norm_pre_clip': 24.209552764892578, 'timing/data': 0.0009095699933823198, 'timing/model': 1.2084985969995614, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 1.01}) +[2026-06-23 23:08:21,570][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.027929849922657013, 'train/grad_norm_pre_clip': 12.180514335632324, 'timing/data': 0.0007155210187193006, 'timing/model': 1.422602520993678, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 1.02}) +[2026-06-23 23:08:25,913][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.05485054850578308, 'train/grad_norm_pre_clip': 18.508804321289062, 'timing/data': 0.0007838680176064372, 'timing/model': 1.2042212939995807, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 1.02}) +[2026-06-23 23:08:30,297][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.009832412004470825, 'train/grad_norm_pre_clip': 16.86052131652832, 'timing/data': 0.0007963000098243356, 'timing/model': 1.2065795579983387, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 1.02}) +[2026-06-23 23:08:34,669][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.09213309735059738, 'train/grad_norm_pre_clip': 14.795654296875, 'timing/data': 0.0008955039957072586, 'timing/model': 1.198055413988186, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 1.02}) +[2026-06-23 23:08:39,269][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.00223755044862628, 'train/grad_norm_pre_clip': 28.601478576660156, 'timing/data': 0.000752129009924829, 'timing/model': 1.444351946003735, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 1.02}) +[2026-06-23 23:08:43,590][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.019306689500808716, 'train/grad_norm_pre_clip': 18.084802627563477, 'timing/data': 0.000834672013297677, 'timing/model': 1.1784968709980603, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 1.02}) +[2026-06-23 23:08:47,950][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.06712611764669418, 'train/grad_norm_pre_clip': 16.350357055664062, 'timing/data': 0.0012169370020274073, 'timing/model': 1.2015988019993529, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 1.02}) +[2026-06-23 23:08:52,270][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.037094101309776306, 'train/grad_norm_pre_clip': 19.210538864135742, 'timing/data': 0.0007323529862333089, 'timing/model': 1.1814891120011453, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 1.02}) +[2026-06-23 23:08:56,838][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.024485671892762184, 'train/grad_norm_pre_clip': 17.879249572753906, 'timing/data': 0.0007162329857237637, 'timing/model': 1.4267869250033982, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 1.02}) +[2026-06-23 23:09:01,179][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.04318026453256607, 'train/grad_norm_pre_clip': 10.219118118286133, 'timing/data': 0.000752018007915467, 'timing/model': 1.1780946500075515, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 1.02}) +[2026-06-23 23:09:05,498][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.009175388142466545, 'train/grad_norm_pre_clip': 11.701066017150879, 'timing/data': 0.0009239669889211655, 'timing/model': 1.179720129002817, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 1.02}) +[2026-06-23 23:09:09,819][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.05590549483895302, 'train/grad_norm_pre_clip': 27.39080238342285, 'timing/data': 0.0007175449864007533, 'timing/model': 1.1800040300004184, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 1.02}) +[2026-06-23 23:09:14,445][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.03309544175863266, 'train/grad_norm_pre_clip': 15.871260643005371, 'timing/data': 0.0007163429982028902, 'timing/model': 1.4691705459845252, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 1.02}) +[2026-06-23 23:09:18,768][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.1504083126783371, 'train/grad_norm_pre_clip': 25.9599552154541, 'timing/data': 0.000797903019702062, 'timing/model': 1.1776176079874858, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 1.02}) +[2026-06-23 23:09:23,089][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.13814134895801544, 'train/grad_norm_pre_clip': 27.6708927154541, 'timing/data': 0.0007263509905897081, 'timing/model': 1.1799118019989692, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 1.02}) +[2026-06-23 23:09:27,435][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.043801311403512955, 'train/grad_norm_pre_clip': 26.299488067626953, 'timing/data': 0.0007240469858516008, 'timing/model': 1.182382021972444, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 1.02}) +[2026-06-23 23:09:32,015][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.052082743495702744, 'train/grad_norm_pre_clip': 44.80977249145508, 'timing/data': 0.0007145900162868202, 'timing/model': 1.4376732620003168, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 1.02}) +[2026-06-23 23:09:36,405][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.09001946449279785, 'train/grad_norm_pre_clip': 30.429386138916016, 'timing/data': 0.0008273679995909333, 'timing/model': 1.1950733849953394, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 1.02}) +[2026-06-23 23:09:40,784][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.011028725653886795, 'train/grad_norm_pre_clip': 12.087410926818848, 'timing/data': 0.0007765539921820164, 'timing/model': 1.1979545029753353, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 1.02}) +[2026-06-23 23:09:45,131][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.0396931916475296, 'train/grad_norm_pre_clip': 18.640867233276367, 'timing/data': 0.0008941810228861868, 'timing/model': 1.179321687988704, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 1.02}) +[2026-06-23 23:09:49,723][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.006467917468398809, 'train/grad_norm_pre_clip': 13.38383960723877, 'timing/data': 0.0007682490104343742, 'timing/model': 1.4433341110125184, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 1.02}) +[2026-06-23 23:09:54,100][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.004532954655587673, 'train/grad_norm_pre_clip': 8.340502738952637, 'timing/data': 0.000775371998315677, 'timing/model': 1.1944108799798414, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 1.02}) +[2026-06-23 23:09:58,493][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.12016550451517105, 'train/grad_norm_pre_clip': 13.518776893615723, 'timing/data': 0.0007221940031740814, 'timing/model': 1.2037515169940889, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 1.02}) +[2026-06-23 23:10:02,860][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.026945708319544792, 'train/grad_norm_pre_clip': 15.240705490112305, 'timing/data': 0.0010184810089413077, 'timing/model': 1.1990933410124853, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 1.02}) +[2026-06-23 23:10:07,479][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.12420233339071274, 'train/grad_norm_pre_clip': 23.085010528564453, 'timing/data': 0.000734976987587288, 'timing/model': 1.449554242019076, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 1.02}) +[2026-06-23 23:10:11,840][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.023693563416600227, 'train/grad_norm_pre_clip': 15.178093910217285, 'timing/data': 0.0011389440041966736, 'timing/model': 1.2031008769990876, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.02}) +[2026-06-23 23:10:16,212][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.14262962341308594, 'train/grad_norm_pre_clip': 25.711139678955078, 'timing/data': 0.0008137930126395077, 'timing/model': 1.210015173011925, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 1.03}) +[2026-06-23 23:10:20,549][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.01582442969083786, 'train/grad_norm_pre_clip': 15.095620155334473, 'timing/data': 0.0013489509874489158, 'timing/model': 1.181263469014084, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 1.03}) +[2026-06-23 23:10:25,201][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.07247292250394821, 'train/grad_norm_pre_clip': 19.225923538208008, 'timing/data': 0.0007316620030906051, 'timing/model': 1.4565407169866376, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 1.03}) +[2026-06-23 23:10:29,556][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.029954155907034874, 'train/grad_norm_pre_clip': 21.71365737915039, 'timing/data': 0.0013453740102704614, 'timing/model': 1.1958473369886633, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 1.03}) +[2026-06-23 23:10:33,899][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.21382060647010803, 'train/grad_norm_pre_clip': 51.9078483581543, 'timing/data': 0.0007439540058840066, 'timing/model': 1.1793396349821705, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 1.03}) +[2026-06-23 23:10:38,258][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.008612483739852905, 'train/grad_norm_pre_clip': 10.849956512451172, 'timing/data': 0.0007966419798322022, 'timing/model': 1.21493694500532, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 1.03}) +[2026-06-23 23:10:42,799][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.4557894170284271, 'train/grad_norm_pre_clip': 27.16408348083496, 'timing/data': 0.0007738590065855533, 'timing/model': 1.178712850989541, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 1.03}) +[2026-06-23 23:10:47,127][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.023684073239564896, 'train/grad_norm_pre_clip': 12.077807426452637, 'timing/data': 0.0007473299920093268, 'timing/model': 1.182769938983256, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 1.03}) +[2026-06-23 23:10:51,510][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.18522492051124573, 'train/grad_norm_pre_clip': 17.453628540039062, 'timing/data': 0.0007491740107070655, 'timing/model': 1.2055983680184, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 1.03}) +[2026-06-23 23:10:55,865][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.11301159113645554, 'train/grad_norm_pre_clip': 19.151803970336914, 'timing/data': 0.0007696019893046468, 'timing/model': 1.1847197730094194, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 1.03}) +[2026-06-23 23:11:00,461][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.015261203981935978, 'train/grad_norm_pre_clip': 8.870523452758789, 'timing/data': 0.0007422709895763546, 'timing/model': 1.4298026570177171, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 1.03}) +[2026-06-23 23:11:04,806][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.10620523244142532, 'train/grad_norm_pre_clip': 18.712024688720703, 'timing/data': 0.0007230050105135888, 'timing/model': 1.181496431992855, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 1.03}) +[2026-06-23 23:11:09,130][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.03922775760293007, 'train/grad_norm_pre_clip': 11.868535041809082, 'timing/data': 0.0013089569984003901, 'timing/model': 1.1802411790122278, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 1.03}) +[2026-06-23 23:11:13,475][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.1257859617471695, 'train/grad_norm_pre_clip': 18.183528900146484, 'timing/data': 0.0006980990001466125, 'timing/model': 1.1804510119836777, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 1.03}) +[2026-06-23 23:11:18,055][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.069208063185215, 'train/grad_norm_pre_clip': 12.220686912536621, 'timing/data': 0.0007178859959822148, 'timing/model': 1.4275636999809649, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 1.03}) +[2026-06-23 23:11:22,435][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.14851948618888855, 'train/grad_norm_pre_clip': 16.81157875061035, 'timing/data': 0.0006963669729884714, 'timing/model': 1.2018402190005872, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 1.03}) +[2026-06-23 23:11:26,772][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.05783333629369736, 'train/grad_norm_pre_clip': 15.775309562683105, 'timing/data': 0.000898748985491693, 'timing/model': 1.1850707420089748, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 1.03}) +[2026-06-23 23:11:31,142][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.08952043205499649, 'train/grad_norm_pre_clip': 19.08807373046875, 'timing/data': 0.0011035780189558864, 'timing/model': 1.1900148600107059, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 1.03}) +[2026-06-23 23:11:35,773][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.09080957621335983, 'train/grad_norm_pre_clip': 22.221477508544922, 'timing/data': 0.0007837270095478743, 'timing/model': 1.446284915000433, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 1.03}) +[2026-06-23 23:11:40,135][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.03827865421772003, 'train/grad_norm_pre_clip': 9.880010604858398, 'timing/data': 0.0009727769938763231, 'timing/model': 1.1825613850087393, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 1.03}) +[2026-06-23 23:11:44,482][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.08025643974542618, 'train/grad_norm_pre_clip': 19.378385543823242, 'timing/data': 0.000717414019163698, 'timing/model': 1.1810925570025574, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 1.03}) +[2026-06-23 23:11:48,810][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.011442241258919239, 'train/grad_norm_pre_clip': 18.27587127685547, 'timing/data': 0.0006950539827812463, 'timing/model': 1.1819498160039075, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 1.03}) +[2026-06-23 23:11:53,315][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.22221599519252777, 'train/grad_norm_pre_clip': 31.98554801940918, 'timing/data': 0.00076980100129731, 'timing/model': 1.1795088330109138, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 1.03}) +[2026-06-23 23:11:57,677][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.041648026555776596, 'train/grad_norm_pre_clip': 14.479744911193848, 'timing/data': 0.0007444040093105286, 'timing/model': 1.180352657014737, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 1.03}) +[2026-06-23 23:12:02,001][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.1812349408864975, 'train/grad_norm_pre_clip': 17.070270538330078, 'timing/data': 0.0008803759992588311, 'timing/model': 1.1805676500080153, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 1.03}) +[2026-06-23 23:12:06,388][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.0112822987139225, 'train/grad_norm_pre_clip': 16.05745506286621, 'timing/data': 0.0008291519770864397, 'timing/model': 1.211095820995979, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 1.03}) +[2026-06-23 23:12:10,991][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.03418923541903496, 'train/grad_norm_pre_clip': 15.266485214233398, 'timing/data': 0.0007433630235027522, 'timing/model': 1.439065885002492, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 1.04}) +[2026-06-23 23:12:15,333][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.13520190119743347, 'train/grad_norm_pre_clip': 26.042631149291992, 'timing/data': 0.0009924930054694414, 'timing/model': 1.1826828430057503, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 1.04}) +[2026-06-23 23:12:19,693][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.035888683050870895, 'train/grad_norm_pre_clip': 15.936646461486816, 'timing/data': 0.0007292369846254587, 'timing/model': 1.1838806699961424, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 1.04}) +[2026-06-23 23:12:24,041][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.011946921236813068, 'train/grad_norm_pre_clip': 25.43635368347168, 'timing/data': 0.0007326730119530112, 'timing/model': 1.1817333890066948, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 1.04}) +[2026-06-23 23:12:28,631][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.0070888204500079155, 'train/grad_norm_pre_clip': 20.89891242980957, 'timing/data': 0.0007728369964752346, 'timing/model': 1.4369174310122617, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 1.04}) +[2026-06-23 23:12:32,986][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.23971706628799438, 'train/grad_norm_pre_clip': 37.08039474487305, 'timing/data': 0.000729366991436109, 'timing/model': 1.181686673022341, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 1.04}) +[2026-06-23 23:12:37,374][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.06707220524549484, 'train/grad_norm_pre_clip': 13.179824829101562, 'timing/data': 0.000727012986317277, 'timing/model': 1.1869968860119116, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 1.04}) +[2026-06-23 23:12:41,697][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.014434728771448135, 'train/grad_norm_pre_clip': 14.968676567077637, 'timing/data': 0.000689805019646883, 'timing/model': 1.1772997399966698, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 1.04}) +[2026-06-23 23:12:46,265][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.08337151259183884, 'train/grad_norm_pre_clip': 14.786924362182617, 'timing/data': 0.0008612209931015968, 'timing/model': 1.4286397770047188, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 1.04}) +[2026-06-23 23:12:50,600][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.04893273115158081, 'train/grad_norm_pre_clip': 9.631242752075195, 'timing/data': 0.0006833419902250171, 'timing/model': 1.197553229983896, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 1.04}) +[2026-06-23 23:12:54,914][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.031689826399087906, 'train/grad_norm_pre_clip': 9.459815979003906, 'timing/data': 0.0007008350221440196, 'timing/model': 1.1771781099960208, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 1.04}) +[2026-06-23 23:12:59,228][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.008633507415652275, 'train/grad_norm_pre_clip': 13.038884162902832, 'timing/data': 0.0007288960041478276, 'timing/model': 1.1773823820112739, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 1.04}) +[2026-06-23 23:13:03,844][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.15822115540504456, 'train/grad_norm_pre_clip': 17.55829620361328, 'timing/data': 0.0007479310152120888, 'timing/model': 1.461273486027494, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 1.04}) +[2026-06-23 23:13:08,168][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.03504401817917824, 'train/grad_norm_pre_clip': 14.951835632324219, 'timing/data': 0.000820694986032322, 'timing/model': 1.1738378849986475, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 1.04}) +[2026-06-23 23:13:12,572][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.02829478494822979, 'train/grad_norm_pre_clip': 22.88994789123535, 'timing/data': 0.0007327630009967834, 'timing/model': 1.1952210350136738, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 1.04}) +[2026-06-23 23:13:16,898][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.06193334609270096, 'train/grad_norm_pre_clip': 15.642930030822754, 'timing/data': 0.0007152009929995984, 'timing/model': 1.1744034890143666, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 1.04}) +[2026-06-23 23:13:21,505][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.023382846266031265, 'train/grad_norm_pre_clip': 18.063947677612305, 'timing/data': 0.0007196789956651628, 'timing/model': 1.4494541680032853, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 1.04}) +[2026-06-23 23:13:25,855][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.03334829956293106, 'train/grad_norm_pre_clip': 13.755332946777344, 'timing/data': 0.0006947130023036152, 'timing/model': 1.174588980997214, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 1.04}) +[2026-06-23 23:13:30,166][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.14000500738620758, 'train/grad_norm_pre_clip': 15.598738670349121, 'timing/data': 0.000696375995175913, 'timing/model': 1.176901184982853, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 1.04}) +[2026-06-23 23:13:34,478][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.01118536852300167, 'train/grad_norm_pre_clip': 23.553173065185547, 'timing/data': 0.0007098419882822782, 'timing/model': 1.1770025879959576, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 1.04}) +[2026-06-23 23:13:39,044][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.0023198884446173906, 'train/grad_norm_pre_clip': 13.723464965820312, 'timing/data': 0.000682630023220554, 'timing/model': 1.4245928099844605, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 1.04}) +[2026-06-23 23:13:43,409][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.010405142791569233, 'train/grad_norm_pre_clip': 19.59428596496582, 'timing/data': 0.0006716909992974252, 'timing/model': 1.1767709080013447, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 1.04}) +[2026-06-23 23:13:47,724][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.023468898609280586, 'train/grad_norm_pre_clip': 17.294574737548828, 'timing/data': 0.0006919380102772266, 'timing/model': 1.176780860987492, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.04}) +[2026-06-23 23:13:52,066][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.10777927935123444, 'train/grad_norm_pre_clip': 20.350299835205078, 'timing/data': 0.0007506969850510359, 'timing/model': 1.17792376599391, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 1.04}) +[2026-06-23 23:13:56,633][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.029754742980003357, 'train/grad_norm_pre_clip': 20.64003562927246, 'timing/data': 0.000744654971640557, 'timing/model': 1.430316171987215, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 1.04}) +[2026-06-23 23:14:00,949][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2700 +[2026-06-23 23:14:58,856][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2700 +[2026-06-23 23:14:58,856][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2700 +[2026-06-23 23:15:35,548][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2700 +[2026-06-23 23:15:35,550][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.037529412657022476, 'train/grad_norm_pre_clip': 28.751049041748047, 'eval/action_loss/seconds': 57.903708055004245, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03873303532600403, 'eval/latency_2/loss': 0.03873303532600403, 'eval/flappy/loss': 0.03873303532600403, 'eval/flappy/latency_2/loss': 0.03873303532600403, 'eval/flap_precision': 0.9861687413554634, 'eval/flap_recall': 0.9674355495251018, 'eval/flap_f1': 0.9767123287671233, 'eval/flap_rate_model': 0.20083333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9767123287671232, 'eval/latency_2/flap_precision': 0.9861687413554634, 'eval/latency_2/flap_recall': 0.9674355495251018, 'eval/latency_2/flap_f1': 0.9767123287671233, 'eval/latency_2/flap_rate_model': 0.20083333333333334, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9767123287671232, 'eval/action_classification/data_wait_seconds': 0.4356580972671509, 'eval/action_classification/predict_seconds': 36.114662170410156, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0008678629819769412, 'timing/model': 1.1799779089924414, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 1.04}) +[2026-06-23 23:15:39,892][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.030487043783068657, 'train/grad_norm_pre_clip': 30.655298233032227, 'timing/data': 0.0007359999872278422, 'timing/model': 1.184996135008987, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 1.05}) +[2026-06-23 23:15:44,384][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.035107847303152084, 'train/grad_norm_pre_clip': 11.435433387756348, 'timing/data': 0.0009465689945500344, 'timing/model': 1.3725161500042304, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 1.05}) +[2026-06-23 23:15:48,718][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.12138616293668747, 'train/grad_norm_pre_clip': 15.85456657409668, 'timing/data': 0.000766024983022362, 'timing/model': 1.184239723981591, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 1.05}) +[2026-06-23 23:15:53,056][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.014613989740610123, 'train/grad_norm_pre_clip': 29.40629005432129, 'timing/data': 0.0007297980191651732, 'timing/model': 1.204160724999383, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 1.05}) +[2026-06-23 23:15:57,366][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.03162870556116104, 'train/grad_norm_pre_clip': 7.580227375030518, 'timing/data': 0.0007527100096922368, 'timing/model': 1.179151322983671, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 1.05}) +[2026-06-23 23:16:01,970][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.015500565990805626, 'train/grad_norm_pre_clip': 11.465278625488281, 'timing/data': 0.000710913009243086, 'timing/model': 1.460817945015151, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 1.05}) +[2026-06-23 23:16:06,308][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.21425847709178925, 'train/grad_norm_pre_clip': 29.624826431274414, 'timing/data': 0.0008743650105316192, 'timing/model': 1.192252679989906, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 1.05}) +[2026-06-23 23:16:10,641][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.004729652311652899, 'train/grad_norm_pre_clip': 10.363273620605469, 'timing/data': 0.0007297880074474961, 'timing/model': 1.1912954450235702, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 1.05}) +[2026-06-23 23:16:14,993][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.1423366665840149, 'train/grad_norm_pre_clip': 21.86721420288086, 'timing/data': 0.0007411889964714646, 'timing/model': 1.2160018850117922, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 1.05}) +[2026-06-23 23:16:19,484][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.04768387973308563, 'train/grad_norm_pre_clip': 24.139263153076172, 'timing/data': 0.0007877449970692396, 'timing/model': 1.189596183015965, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 1.05}) +[2026-06-23 23:16:23,848][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.005865626968443394, 'train/grad_norm_pre_clip': 18.385026931762695, 'timing/data': 0.0007475920137949288, 'timing/model': 1.1999603399890475, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 1.05}) +[2026-06-23 23:16:28,220][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.06800888478755951, 'train/grad_norm_pre_clip': 30.000394821166992, 'timing/data': 0.0007833870186004788, 'timing/model': 1.1954465559974778, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 1.05}) +[2026-06-23 23:16:32,567][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.02692042663693428, 'train/grad_norm_pre_clip': 40.372222900390625, 'timing/data': 0.000724587996955961, 'timing/model': 1.2028149699908681, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 1.05}) +[2026-06-23 23:16:37,114][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.02784397266805172, 'train/grad_norm_pre_clip': 7.899736404418945, 'timing/data': 0.0006782029813621193, 'timing/model': 1.1982513740076683, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 1.05}) +[2026-06-23 23:16:41,480][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.13336063921451569, 'train/grad_norm_pre_clip': 30.654664993286133, 'timing/data': 0.0007773760007694364, 'timing/model': 1.175751755014062, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 1.05}) +[2026-06-23 23:16:45,805][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.008996592834591866, 'train/grad_norm_pre_clip': 11.517470359802246, 'timing/data': 0.0008855460037011653, 'timing/model': 1.1729940340155736, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 1.05}) +[2026-06-23 23:16:50,153][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.013771838508546352, 'train/grad_norm_pre_clip': 16.108701705932617, 'timing/data': 0.0007264410087373108, 'timing/model': 1.196356772998115, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 1.05}) +[2026-06-23 23:16:54,699][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.09609355032444, 'train/grad_norm_pre_clip': 13.40746021270752, 'timing/data': 0.0006855659885331988, 'timing/model': 1.171953727985965, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 1.05}) +[2026-06-23 23:16:59,027][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.09149851649999619, 'train/grad_norm_pre_clip': 23.896650314331055, 'timing/data': 0.0007271229987964034, 'timing/model': 1.202656236011535, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 1.05}) +[2026-06-23 23:17:03,385][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.0002616811834741384, 'train/grad_norm_pre_clip': 40.202552795410156, 'timing/data': 0.0008359139901585877, 'timing/model': 1.1934496219910216, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 1.05}) +[2026-06-23 23:17:07,710][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.04136774316430092, 'train/grad_norm_pre_clip': 29.867507934570312, 'timing/data': 0.0008163980091921985, 'timing/model': 1.199439027986955, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 1.05}) +[2026-06-23 23:17:12,280][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.002336266450583935, 'train/grad_norm_pre_clip': 12.447395324707031, 'timing/data': 0.0009224539971910417, 'timing/model': 1.205582280003, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 1.05}) +[2026-06-23 23:17:16,641][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.06963743269443512, 'train/grad_norm_pre_clip': 31.332544326782227, 'timing/data': 0.000767296995036304, 'timing/model': 1.1961443379987031, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 1.05}) +[2026-06-23 23:17:21,019][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.09452688694000244, 'train/grad_norm_pre_clip': 19.672327041625977, 'timing/data': 0.0011191369849257171, 'timing/model': 1.2038803510076832, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 1.05}) +[2026-06-23 23:17:25,388][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.0599045492708683, 'train/grad_norm_pre_clip': 21.756669998168945, 'timing/data': 0.0007772259996272624, 'timing/model': 1.178445230005309, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 1.05}) +[2026-06-23 23:17:29,959][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.0053415619768202305, 'train/grad_norm_pre_clip': 19.241422653198242, 'timing/data': 0.0010290509962942451, 'timing/model': 1.19801878000726, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 1.06}) +[2026-06-23 23:17:34,296][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.06453092396259308, 'train/grad_norm_pre_clip': 25.413848876953125, 'timing/data': 0.0007172049954533577, 'timing/model': 1.196773381991079, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 1.06}) +[2026-06-23 23:17:38,740][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.0246721338480711, 'train/grad_norm_pre_clip': 22.637174606323242, 'timing/data': 0.0008901639957912266, 'timing/model': 1.1754545329895336, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 1.06}) +[2026-06-23 23:17:43,339][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.010489445179700851, 'train/grad_norm_pre_clip': 17.958866119384766, 'timing/data': 0.0007157830114010721, 'timing/model': 1.4457184189814143, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 1.06}) +[2026-06-23 23:17:47,694][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.009321099147200584, 'train/grad_norm_pre_clip': 17.652236938476562, 'timing/data': 0.0007391549879685044, 'timing/model': 1.1977103609824553, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 1.06}) +[2026-06-23 23:17:52,020][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.029004819691181183, 'train/grad_norm_pre_clip': 22.386953353881836, 'timing/data': 0.0011695899884216487, 'timing/model': 1.1725068669766188, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 1.06}) +[2026-06-23 23:17:56,351][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.11108148097991943, 'train/grad_norm_pre_clip': 12.603555679321289, 'timing/data': 0.0007444350048899651, 'timing/model': 1.1926657779840752, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 1.06}) +[2026-06-23 23:18:00,923][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.036779724061489105, 'train/grad_norm_pre_clip': 12.187446594238281, 'timing/data': 0.0007123060058802366, 'timing/model': 1.4200031090003904, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 1.06}) +[2026-06-23 23:18:05,227][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.01760958507657051, 'train/grad_norm_pre_clip': 11.896072387695312, 'timing/data': 0.0007218430109787732, 'timing/model': 1.1760484570113476, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 1.06}) +[2026-06-23 23:18:09,537][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.05151337757706642, 'train/grad_norm_pre_clip': 28.38982582092285, 'timing/data': 0.0007649729959666729, 'timing/model': 1.1738968350109644, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 1.06}) +[2026-06-23 23:18:13,856][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.011246534995734692, 'train/grad_norm_pre_clip': 22.278745651245117, 'timing/data': 0.0007365299970842898, 'timing/model': 1.1939279399812222, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 1.06}) +[2026-06-23 23:18:18,320][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.0805940181016922, 'train/grad_norm_pre_clip': 26.867198944091797, 'timing/data': 0.0007484529924113303, 'timing/model': 1.1729241740249563, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 1.06}) +[2026-06-23 23:18:22,645][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.09431178122758865, 'train/grad_norm_pre_clip': 20.053512573242188, 'timing/data': 0.0007142690010368824, 'timing/model': 1.1753795319818892, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 1.06}) +[2026-06-23 23:18:26,949][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.0919596254825592, 'train/grad_norm_pre_clip': 24.484264373779297, 'timing/data': 0.0007296779949683696, 'timing/model': 1.1757533549971413, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 1.06}) +[2026-06-23 23:18:31,249][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.18965503573417664, 'train/grad_norm_pre_clip': 47.64356231689453, 'timing/data': 0.0006976689910516143, 'timing/model': 1.1726288659847341, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 1.06}) +[2026-06-23 23:18:35,811][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.12717606127262115, 'train/grad_norm_pre_clip': 52.14263153076172, 'timing/data': 0.0007075770117808133, 'timing/model': 1.1711785689985845, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 1.06}) +[2026-06-23 23:18:40,125][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.05539407953619957, 'train/grad_norm_pre_clip': 16.825590133666992, 'timing/data': 0.001029261009534821, 'timing/model': 1.1917917990067508, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 1.06}) +[2026-06-23 23:18:44,424][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.0210508294403553, 'train/grad_norm_pre_clip': 20.52130126953125, 'timing/data': 0.0007327040075324476, 'timing/model': 1.17361724798684, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 1.06}) +[2026-06-23 23:18:48,786][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.0398525670170784, 'train/grad_norm_pre_clip': 18.35246467590332, 'timing/data': 0.0008605399925727397, 'timing/model': 1.1929766430112068, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 1.06}) +[2026-06-23 23:18:53,327][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.05253695324063301, 'train/grad_norm_pre_clip': 14.240663528442383, 'timing/data': 0.0007197889790404588, 'timing/model': 1.1798444260202814, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 1.06}) +[2026-06-23 23:18:57,723][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.058582890778779984, 'train/grad_norm_pre_clip': 28.921995162963867, 'timing/data': 0.0007379830058198422, 'timing/model': 1.181751833006274, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.06}) +[2026-06-23 23:19:02,085][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.06926306337118149, 'train/grad_norm_pre_clip': 16.516387939453125, 'timing/data': 0.0007389949751086533, 'timing/model': 1.202770372008672, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 1.06}) +[2026-06-23 23:19:06,441][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.025444339960813522, 'train/grad_norm_pre_clip': 22.73854637145996, 'timing/data': 0.0007350579835474491, 'timing/model': 1.1936099830199964, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 1.06}) +[2026-06-23 23:19:11,031][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.03705696016550064, 'train/grad_norm_pre_clip': 19.55887794494629, 'timing/data': 0.0006780329858884215, 'timing/model': 1.4127956920128781, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 1.06}) +[2026-06-23 23:28:56,511][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2750 +[2026-06-23 23:28:56,512][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.03088313899934292, 'train/grad_norm_pre_clip': 13.109919548034668, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 150.84000000000373, 'rl_games_eval/mid_train/mean_length': 1234.6, 'rl_games_eval/mid_train/std_reward': 168.12832717898195, 'rl_games_eval/mid_train/std_length': 1354.5628962879503, 'rl_games_eval/mid_train/macro_mean_reward': 150.84000000000373, 'rl_games_eval/mid_train/macro_mean_length': 1234.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 150.84000000000373, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1234.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 168.12832717898195, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1354.5628962879503, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 150.84000000000373, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006917669961694628, 'timing/model': 1.2027353949961253, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 1.06}) +[2026-06-23 23:29:00,647][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.06853064149618149, 'train/grad_norm_pre_clip': 12.756759643554688, 'timing/data': 0.00088626699289307, 'timing/model': 1.13406347198179, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 1.06}) +[2026-06-23 23:29:04,812][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.0855683758854866, 'train/grad_norm_pre_clip': 22.098825454711914, 'timing/data': 0.0007448160031344742, 'timing/model': 1.1379774839733727, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 1.07}) +[2026-06-23 23:29:08,995][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.0008811222505755723, 'train/grad_norm_pre_clip': 36.90247344970703, 'timing/data': 0.0009608050168026239, 'timing/model': 1.1380636030226015, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 1.07}) +[2026-06-23 23:29:13,392][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.00816336739808321, 'train/grad_norm_pre_clip': 28.386735916137695, 'timing/data': 0.0007624989957548678, 'timing/model': 1.375994427013211, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 1.07}) +[2026-06-23 23:29:17,614][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.003091066377237439, 'train/grad_norm_pre_clip': 9.89516544342041, 'timing/data': 0.0007572989852633327, 'timing/model': 1.1480865520134103, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 1.07}) +[2026-06-23 23:29:21,843][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.008632178418338299, 'train/grad_norm_pre_clip': 9.051544189453125, 'timing/data': 0.0011880450183525681, 'timing/model': 1.1450770020019263, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 1.07}) +[2026-06-23 23:29:26,077][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.17274966835975647, 'train/grad_norm_pre_clip': 29.663022994995117, 'timing/data': 0.0006926089990884066, 'timing/model': 1.1692125020199455, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 1.07}) +[2026-06-23 23:29:30,520][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.037129130214452744, 'train/grad_norm_pre_clip': 30.01473617553711, 'timing/data': 0.0006734640046488494, 'timing/model': 1.3961931859957986, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 1.07}) +[2026-06-23 23:29:34,720][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.04821683093905449, 'train/grad_norm_pre_clip': 33.76632308959961, 'timing/data': 0.0007304089958779514, 'timing/model': 1.1655017870070878, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 1.07}) +[2026-06-23 23:29:38,965][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.0028428088407963514, 'train/grad_norm_pre_clip': 23.975536346435547, 'timing/data': 0.000875055993674323, 'timing/model': 1.1850408249883913, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 1.07}) +[2026-06-23 23:29:43,241][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.06675010174512863, 'train/grad_norm_pre_clip': 39.122398376464844, 'timing/data': 0.000788186996942386, 'timing/model': 1.1697453380038496, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 1.07}) +[2026-06-23 23:29:47,730][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.02832137607038021, 'train/grad_norm_pre_clip': 32.22513198852539, 'timing/data': 0.0007430730038322508, 'timing/model': 1.379289130010875, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 1.07}) +[2026-06-23 23:29:51,986][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.1194429099559784, 'train/grad_norm_pre_clip': 15.349574089050293, 'timing/data': 0.0007323530153371394, 'timing/model': 1.1590210779977497, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 1.07}) +[2026-06-23 23:29:56,274][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.0859399288892746, 'train/grad_norm_pre_clip': 11.903634071350098, 'timing/data': 0.0007528600108344108, 'timing/model': 1.1858479189977515, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 1.07}) +[2026-06-23 23:30:00,546][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.018185138702392578, 'train/grad_norm_pre_clip': 8.192124366760254, 'timing/data': 0.0007488440023735166, 'timing/model': 1.1649520760111045, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 1.07}) +[2026-06-23 23:30:05,008][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.06819319725036621, 'train/grad_norm_pre_clip': 24.8187198638916, 'timing/data': 0.0007002539932727814, 'timing/model': 1.3798064980073832, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 1.07}) +[2026-06-23 23:30:09,304][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.012637979350984097, 'train/grad_norm_pre_clip': 14.34106731414795, 'timing/data': 0.0006978690216783434, 'timing/model': 1.187989728001412, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 1.07}) +[2026-06-23 23:30:13,646][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.08970833569765091, 'train/grad_norm_pre_clip': 30.685747146606445, 'timing/data': 0.0009992359846364707, 'timing/model': 1.1862188070081174, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 1.07}) +[2026-06-23 23:30:17,951][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.15101197361946106, 'train/grad_norm_pre_clip': 20.69034767150879, 'timing/data': 0.0009032989910338074, 'timing/model': 1.185493086988572, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 1.07}) +[2026-06-23 23:30:22,530][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.10756655782461166, 'train/grad_norm_pre_clip': 42.292476654052734, 'timing/data': 0.000712817010935396, 'timing/model': 1.19849372599856, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 1.07}) +[2026-06-23 23:30:26,852][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.03737201914191246, 'train/grad_norm_pre_clip': 6.169186115264893, 'timing/data': 0.0007917930197436363, 'timing/model': 1.1920562330051325, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 1.07}) +[2026-06-23 23:30:31,180][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.0017539194086566567, 'train/grad_norm_pre_clip': 20.752695083618164, 'timing/data': 0.0012022509763482958, 'timing/model': 1.1831961660063826, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 1.07}) +[2026-06-23 23:30:35,517][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.030963994562625885, 'train/grad_norm_pre_clip': 18.37852668762207, 'timing/data': 0.0007574189803563058, 'timing/model': 1.18670356500661, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 1.07}) +[2026-06-23 23:30:40,114][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.0501939058303833, 'train/grad_norm_pre_clip': 12.9575834274292, 'timing/data': 0.0007232249772641808, 'timing/model': 1.189598794997437, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 1.07}) +[2026-06-23 23:30:44,439][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.019713524729013443, 'train/grad_norm_pre_clip': 16.383087158203125, 'timing/data': 0.000946839980315417, 'timing/model': 1.1864413429866545, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 1.07}) +[2026-06-23 23:30:48,790][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.055760592222213745, 'train/grad_norm_pre_clip': 20.464502334594727, 'timing/data': 0.0007379539893008769, 'timing/model': 1.1897907109814696, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 1.07}) +[2026-06-23 23:30:53,119][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.003782650688663125, 'train/grad_norm_pre_clip': 8.771808624267578, 'timing/data': 0.0009211119904648513, 'timing/model': 1.1948551179957576, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 1.07}) +[2026-06-23 23:30:57,733][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.09384482353925705, 'train/grad_norm_pre_clip': 49.82036209106445, 'timing/data': 0.0007182169938459992, 'timing/model': 1.1876975249906536, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 1.08}) +[2026-06-23 23:31:02,082][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.036568913608789444, 'train/grad_norm_pre_clip': 25.59889793395996, 'timing/data': 0.0008020410023164004, 'timing/model': 1.1759252570045646, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 1.08}) +[2026-06-23 23:31:06,443][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.04547440633177757, 'train/grad_norm_pre_clip': 13.995985984802246, 'timing/data': 0.0007181669934652746, 'timing/model': 1.193083285994362, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 1.08}) +[2026-06-23 23:31:10,775][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.02147749252617359, 'train/grad_norm_pre_clip': 22.478870391845703, 'timing/data': 0.0010740150173660368, 'timing/model': 1.1898816319881007, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 1.08}) +[2026-06-23 23:31:15,256][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.003482674714177847, 'train/grad_norm_pre_clip': 27.33905792236328, 'timing/data': 0.0007669170154258609, 'timing/model': 1.193945420003729, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 1.08}) +[2026-06-23 23:31:19,606][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.03934009373188019, 'train/grad_norm_pre_clip': 30.942480087280273, 'timing/data': 0.0011137079854961485, 'timing/model': 1.1751667780044954, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 1.08}) +[2026-06-23 23:31:23,969][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.0780886635184288, 'train/grad_norm_pre_clip': 31.727684020996094, 'timing/data': 0.000749815022572875, 'timing/model': 1.1969543099985458, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 1.08}) +[2026-06-23 23:31:28,346][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.07800203561782837, 'train/grad_norm_pre_clip': 20.692562103271484, 'timing/data': 0.0009166929812636226, 'timing/model': 1.1974628299940377, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 1.08}) +[2026-06-23 23:31:32,924][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.06168314442038536, 'train/grad_norm_pre_clip': 27.115144729614258, 'timing/data': 0.0007342970056924969, 'timing/model': 1.1982675069884863, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 1.08}) +[2026-06-23 23:31:37,288][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.2642425000667572, 'train/grad_norm_pre_clip': 32.674373626708984, 'timing/data': 0.0009911310044117272, 'timing/model': 1.1826576390012633, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 1.08}) +[2026-06-23 23:31:41,684][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.10990197956562042, 'train/grad_norm_pre_clip': 18.142253875732422, 'timing/data': 0.0007546440174337476, 'timing/model': 1.205916700011585, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 1.08}) +[2026-06-23 23:31:46,229][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.058117836713790894, 'train/grad_norm_pre_clip': 21.621685028076172, 'timing/data': 0.0008350230054929852, 'timing/model': 1.3996592510084156, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 1.08}) +[2026-06-23 23:31:50,580][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.03483910486102104, 'train/grad_norm_pre_clip': 16.11138916015625, 'timing/data': 0.0007402780174743384, 'timing/model': 1.1927215330069885, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 1.08}) +[2026-06-23 23:31:54,981][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.057969074696302414, 'train/grad_norm_pre_clip': 28.839181900024414, 'timing/data': 0.0009071259992197156, 'timing/model': 1.2151901640172582, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 1.08}) +[2026-06-23 23:31:59,355][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.041125476360321045, 'train/grad_norm_pre_clip': 15.60731029510498, 'timing/data': 0.0007585409912280738, 'timing/model': 1.202980887988815, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 1.08}) +[2026-06-23 23:32:03,920][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.00084548513405025, 'train/grad_norm_pre_clip': 19.69657325744629, 'timing/data': 0.0009775060170795768, 'timing/model': 1.4339449150138535, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 1.08}) +[2026-06-23 23:32:08,269][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.0034305015578866005, 'train/grad_norm_pre_clip': 11.1450777053833, 'timing/data': 0.0007753320096526295, 'timing/model': 1.1975136190012563, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 1.08}) +[2026-06-23 23:32:12,660][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.011326958425343037, 'train/grad_norm_pre_clip': 26.849929809570312, 'timing/data': 0.0008336800092365593, 'timing/model': 1.2025270810117945, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.08}) +[2026-06-23 23:32:17,000][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.015058486722409725, 'train/grad_norm_pre_clip': 13.613360404968262, 'timing/data': 0.0007495149911846966, 'timing/model': 1.1953622049768455, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 1.08}) +[2026-06-23 23:32:21,535][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.07283077389001846, 'train/grad_norm_pre_clip': 122.46463012695312, 'timing/data': 0.0007228050089906901, 'timing/model': 1.1947406170074828, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 1.08}) +[2026-06-23 23:32:25,917][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.19274942576885223, 'train/grad_norm_pre_clip': 12.463549613952637, 'timing/data': 0.0012083229958079755, 'timing/model': 1.2058937120018527, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 1.08}) +[2026-06-23 23:32:30,290][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.09783399105072021, 'train/grad_norm_pre_clip': 13.323148727416992, 'timing/data': 0.0006897439889144152, 'timing/model': 1.1916457700135652, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 1.08}) +[2026-06-23 23:32:34,677][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2800 +[2026-06-23 23:33:32,474][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2800 +[2026-06-23 23:33:32,475][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2800 +[2026-06-23 23:34:09,707][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2800 +[2026-06-23 23:34:09,708][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.0646701455116272, 'train/grad_norm_pre_clip': 16.7973575592041, 'eval/action_loss/seconds': 57.792789785016794, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03219126537442207, 'eval/latency_2/loss': 0.03219126537442207, 'eval/flappy/loss': 0.03219126537442207, 'eval/flappy/latency_2/loss': 0.03219126537442207, 'eval/flap_precision': 0.958994708994709, 'eval/flap_recall': 0.9837177747625508, 'eval/flap_f1': 0.9711989283322169, 'eval/flap_rate_model': 0.21, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9711989283322169, 'eval/latency_2/flap_precision': 0.958994708994709, 'eval/latency_2/flap_recall': 0.9837177747625508, 'eval/latency_2/flap_f1': 0.9711989283322169, 'eval/latency_2/flap_rate_model': 0.21, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9711989283322169, 'eval/action_classification/data_wait_seconds': 0.4943831264972687, 'eval/action_classification/predict_seconds': 36.58259201049805, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007455569866579026, 'timing/model': 1.1997765819833148, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 1.08}) +[2026-06-23 23:34:14,044][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.016193576157093048, 'train/grad_norm_pre_clip': 18.65980339050293, 'timing/data': 0.0007423810020554811, 'timing/model': 1.1870943320100196, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 1.08}) +[2026-06-23 23:34:18,371][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.020888052880764008, 'train/grad_norm_pre_clip': 7.179208755493164, 'timing/data': 0.0009012950176838785, 'timing/model': 1.16608325499692, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 1.08}) +[2026-06-23 23:34:22,686][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.05567404255270958, 'train/grad_norm_pre_clip': 18.098291397094727, 'timing/data': 0.0007515790057368577, 'timing/model': 1.1842407610092778, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 1.08}) +[2026-06-23 23:34:27,207][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.15888676047325134, 'train/grad_norm_pre_clip': 33.71308135986328, 'timing/data': 0.0007089999853633344, 'timing/model': 1.35613492099219, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 1.09}) +[2026-06-23 23:34:31,510][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.03520768880844116, 'train/grad_norm_pre_clip': 16.78628158569336, 'timing/data': 0.0007852509734220803, 'timing/model': 1.1880585349863395, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 1.09}) +[2026-06-23 23:34:35,851][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.07097639888525009, 'train/grad_norm_pre_clip': 10.87745475769043, 'timing/data': 0.001495292002800852, 'timing/model': 1.1848839109879918, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 1.09}) +[2026-06-23 23:34:40,181][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.003309166058897972, 'train/grad_norm_pre_clip': 24.07179069519043, 'timing/data': 0.0010436279990244657, 'timing/model': 1.1879382740007713, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 1.09}) +[2026-06-23 23:34:44,664][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.1790492683649063, 'train/grad_norm_pre_clip': 26.21755027770996, 'timing/data': 0.0007483320077881217, 'timing/model': 1.3608559949789196, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 1.09}) +[2026-06-23 23:34:49,022][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.11946184188127518, 'train/grad_norm_pre_clip': 16.810728073120117, 'timing/data': 0.0007856909942347556, 'timing/model': 1.1892633620009292, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 1.09}) +[2026-06-23 23:34:53,347][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.08842611312866211, 'train/grad_norm_pre_clip': 18.401674270629883, 'timing/data': 0.000791963015217334, 'timing/model': 1.1735147180152126, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 1.09}) +[2026-06-23 23:34:57,739][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.059067267924547195, 'train/grad_norm_pre_clip': 12.76722526550293, 'timing/data': 0.000975011003902182, 'timing/model': 1.205909004987916, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 1.09}) +[2026-06-23 23:35:02,247][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.05278825759887695, 'train/grad_norm_pre_clip': 22.66594886779785, 'timing/data': 0.0009716550121083856, 'timing/model': 1.3544669229886495, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 1.09}) +[2026-06-23 23:35:06,586][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.052843671292066574, 'train/grad_norm_pre_clip': 11.429536819458008, 'timing/data': 0.0007658950053155422, 'timing/model': 1.18623466900317, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 1.09}) +[2026-06-23 23:35:10,966][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.072209432721138, 'train/grad_norm_pre_clip': 15.942059516906738, 'timing/data': 0.0008254449930973351, 'timing/model': 1.1960284929955378, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 1.09}) +[2026-06-23 23:35:15,301][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.04107121750712395, 'train/grad_norm_pre_clip': 19.89253807067871, 'timing/data': 0.0009681280062068254, 'timing/model': 1.195452628977364, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 1.09}) +[2026-06-23 23:35:19,824][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.004244396463036537, 'train/grad_norm_pre_clip': 17.280006408691406, 'timing/data': 0.000930920010432601, 'timing/model': 1.3434433880029246, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 1.09}) +[2026-06-23 23:35:24,176][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.021657288074493408, 'train/grad_norm_pre_clip': 13.938050270080566, 'timing/data': 0.0009221840009558946, 'timing/model': 1.1896133310219739, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 1.09}) +[2026-06-23 23:35:28,536][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.011430240236222744, 'train/grad_norm_pre_clip': 14.73184585571289, 'timing/data': 0.0009016760159283876, 'timing/model': 1.205643727997085, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 1.09}) +[2026-06-23 23:35:32,925][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.009655570611357689, 'train/grad_norm_pre_clip': 17.45926284790039, 'timing/data': 0.0009389749902766198, 'timing/model': 1.199901400017552, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 1.09}) +[2026-06-23 23:35:37,309][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.017634376883506775, 'train/grad_norm_pre_clip': 20.585283279418945, 'timing/data': 0.000733033986762166, 'timing/model': 1.1992129080172163, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 1.09}) +[2026-06-23 23:35:41,933][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.023490119725465775, 'train/grad_norm_pre_clip': 18.058012008666992, 'timing/data': 0.0009932649845723063, 'timing/model': 1.1766564749996178, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 1.09}) +[2026-06-23 23:35:46,305][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.08377058804035187, 'train/grad_norm_pre_clip': 23.58582878112793, 'timing/data': 0.0009064349869731814, 'timing/model': 1.1997575650166254, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 1.09}) +[2026-06-23 23:35:50,631][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.14249765872955322, 'train/grad_norm_pre_clip': 40.786800384521484, 'timing/data': 0.0008122700091917068, 'timing/model': 1.1779730080161244, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 1.09}) +[2026-06-23 23:35:54,980][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.15093238651752472, 'train/grad_norm_pre_clip': 13.963691711425781, 'timing/data': 0.0010575339838396758, 'timing/model': 1.175084793008864, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 1.09}) +[2026-06-23 23:35:59,561][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.11436513811349869, 'train/grad_norm_pre_clip': 15.232447624206543, 'timing/data': 0.0007440340123139322, 'timing/model': 1.1956221449945588, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 1.09}) +[2026-06-23 23:36:03,919][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.006705730687826872, 'train/grad_norm_pre_clip': 40.04585266113281, 'timing/data': 0.0007675779925193638, 'timing/model': 1.193849155009957, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 1.09}) +[2026-06-23 23:36:08,239][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.010556835681200027, 'train/grad_norm_pre_clip': 11.80423355102539, 'timing/data': 0.0009885180043056607, 'timing/model': 1.192533204011852, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 1.09}) +[2026-06-23 23:36:12,543][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.008392240852117538, 'train/grad_norm_pre_clip': 11.502361297607422, 'timing/data': 0.0007478030165657401, 'timing/model': 1.1762754019873682, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 1.09}) +[2026-06-23 23:36:17,145][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.039411187171936035, 'train/grad_norm_pre_clip': 20.563831329345703, 'timing/data': 0.0010129930160474032, 'timing/model': 1.2035898600006476, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 1.09}) +[2026-06-23 23:36:21,478][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.07167404890060425, 'train/grad_norm_pre_clip': 13.532500267028809, 'timing/data': 0.0007481040083803236, 'timing/model': 1.2023970120062586, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 1.1}) +[2026-06-23 23:36:25,843][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.03473081439733505, 'train/grad_norm_pre_clip': 22.825700759887695, 'timing/data': 0.0007757340208627284, 'timing/model': 1.1963088439952116, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 1.1}) +[2026-06-23 23:36:30,167][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.08135082572698593, 'train/grad_norm_pre_clip': 30.4973087310791, 'timing/data': 0.0007337360002566129, 'timing/model': 1.1786957069998607, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 1.1}) +[2026-06-23 23:36:34,772][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.08735862374305725, 'train/grad_norm_pre_clip': 20.426382064819336, 'timing/data': 0.0009027089981827885, 'timing/model': 1.206157185981283, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 1.1}) +[2026-06-23 23:36:39,096][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.09398229420185089, 'train/grad_norm_pre_clip': 17.071969985961914, 'timing/data': 0.000920162012334913, 'timing/model': 1.1842822399921715, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 1.1}) +[2026-06-23 23:36:43,461][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.1341993659734726, 'train/grad_norm_pre_clip': 17.916412353515625, 'timing/data': 0.0007839889731258154, 'timing/model': 1.2001338860136457, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 1.1}) +[2026-06-23 23:36:47,840][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.003183517837896943, 'train/grad_norm_pre_clip': 29.196130752563477, 'timing/data': 0.000735980982426554, 'timing/model': 1.2082752459973563, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 1.1}) +[2026-06-23 23:36:52,488][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.004899975843727589, 'train/grad_norm_pre_clip': 8.804427146911621, 'timing/data': 0.0009385150042362511, 'timing/model': 1.1963747710105963, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 1.1}) +[2026-06-23 23:36:56,817][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.011765961535274982, 'train/grad_norm_pre_clip': 14.634450912475586, 'timing/data': 0.000729449006030336, 'timing/model': 1.1726412099960726, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 1.1}) +[2026-06-23 23:37:01,152][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.0427074059844017, 'train/grad_norm_pre_clip': 19.363065719604492, 'timing/data': 0.0008469049935229123, 'timing/model': 1.2021076059900224, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 1.1}) +[2026-06-23 23:37:05,513][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.07789460569620132, 'train/grad_norm_pre_clip': 15.974742889404297, 'timing/data': 0.0009612379944883287, 'timing/model': 1.197523809998529, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 1.1}) +[2026-06-23 23:37:10,147][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.09608075767755508, 'train/grad_norm_pre_clip': 25.240055084228516, 'timing/data': 0.0007563580002170056, 'timing/model': 1.2119349210115615, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 1.1}) +[2026-06-23 23:37:14,522][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.046573709696531296, 'train/grad_norm_pre_clip': 19.103410720825195, 'timing/data': 0.0007765759946778417, 'timing/model': 1.205497799004661, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 1.1}) +[2026-06-23 23:37:18,882][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.07615379244089127, 'train/grad_norm_pre_clip': 23.00135612487793, 'timing/data': 0.0007378340233117342, 'timing/model': 1.1973921320168301, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 1.1}) +[2026-06-23 23:37:23,277][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.09534259140491486, 'train/grad_norm_pre_clip': 35.99036407470703, 'timing/data': 0.0007265229942277074, 'timing/model': 1.2102016230055597, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.1}) +[2026-06-23 23:37:27,880][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.0007339420262724161, 'train/grad_norm_pre_clip': 9.121000289916992, 'timing/data': 0.0007336960115935653, 'timing/model': 1.182878507010173, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 1.1}) +[2026-06-23 23:37:32,215][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.04395916312932968, 'train/grad_norm_pre_clip': 11.52991771697998, 'timing/data': 0.0007842800114303827, 'timing/model': 1.178469976002816, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 1.1}) +[2026-06-23 23:37:36,566][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.03856165334582329, 'train/grad_norm_pre_clip': 11.485150337219238, 'timing/data': 0.0007482630026061088, 'timing/model': 1.1793778930150438, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 1.1}) +[2026-06-23 23:37:40,955][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.0032235796097666025, 'train/grad_norm_pre_clip': 11.8080415725708, 'timing/data': 0.0008852269966155291, 'timing/model': 1.178593552991515, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 1.1}) +[2026-06-23 23:37:45,568][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.0020485674031078815, 'train/grad_norm_pre_clip': 7.858608722686768, 'timing/data': 0.0010330700024496764, 'timing/model': 1.1969034720095806, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 1.1}) +[2026-06-23 23:37:49,900][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.006407745182514191, 'train/grad_norm_pre_clip': 5.697267532348633, 'timing/data': 0.0008306659874506295, 'timing/model': 1.1865805269917473, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 1.1}) +[2026-06-23 23:37:54,239][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.003840846475213766, 'train/grad_norm_pre_clip': 17.778100967407227, 'timing/data': 0.0007735099934507161, 'timing/model': 1.1984855059999973, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 1.1}) +[2026-06-23 23:37:58,604][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.004322123248130083, 'train/grad_norm_pre_clip': 27.048107147216797, 'timing/data': 0.0007240790000651032, 'timing/model': 1.2030099389958195, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 1.1}) +[2026-06-23 23:38:03,193][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.0345112644135952, 'train/grad_norm_pre_clip': 9.453338623046875, 'timing/data': 0.0007824859931133687, 'timing/model': 1.206889669992961, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 1.1}) +[2026-06-23 23:38:07,561][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.16216996312141418, 'train/grad_norm_pre_clip': 39.40882110595703, 'timing/data': 0.0007720669964328408, 'timing/model': 1.1968621629930567, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 1.1}) +[2026-06-23 23:38:11,961][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.047122154384851456, 'train/grad_norm_pre_clip': 18.5854549407959, 'timing/data': 0.0009691109880805016, 'timing/model': 1.2016050409874879, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 1.1}) +[2026-06-23 23:38:16,326][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.0008448347798548639, 'train/grad_norm_pre_clip': 28.055192947387695, 'timing/data': 0.0007645830046385527, 'timing/model': 1.197184588003438, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 1.11}) +[2026-06-23 23:38:20,908][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.031749479472637177, 'train/grad_norm_pre_clip': 17.802310943603516, 'timing/data': 0.0007738609856460243, 'timing/model': 1.1789992700214498, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 1.11}) +[2026-06-23 23:38:25,243][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.1915758103132248, 'train/grad_norm_pre_clip': 34.65480041503906, 'timing/data': 0.0007093520252965391, 'timing/model': 1.2015908330213279, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 1.11}) +[2026-06-23 23:38:29,577][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.043024398386478424, 'train/grad_norm_pre_clip': 29.09280776977539, 'timing/data': 0.0007591429748572409, 'timing/model': 1.1922700429859105, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 1.11}) +[2026-06-23 23:38:33,938][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.1315104365348816, 'train/grad_norm_pre_clip': 16.578861236572266, 'timing/data': 0.0007382750045508146, 'timing/model': 1.1829933449917007, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 1.11}) +[2026-06-23 23:38:38,552][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.018152326345443726, 'train/grad_norm_pre_clip': 13.795068740844727, 'timing/data': 0.0011692109983414412, 'timing/model': 1.1802242350240704, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 1.11}) +[2026-06-23 23:38:42,990][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.17748236656188965, 'train/grad_norm_pre_clip': 55.9691047668457, 'timing/data': 0.0007470200071111321, 'timing/model': 1.198532720998628, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 1.11}) +[2026-06-23 23:38:47,376][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.009512640535831451, 'train/grad_norm_pre_clip': 37.58330535888672, 'timing/data': 0.000876740989042446, 'timing/model': 1.1941605989995878, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 1.11}) +[2026-06-23 23:38:51,903][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.04064978286623955, 'train/grad_norm_pre_clip': 30.138120651245117, 'timing/data': 0.0007819260063115507, 'timing/model': 1.3734238329925574, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 1.11}) +[2026-06-23 23:38:56,277][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.09555677324533463, 'train/grad_norm_pre_clip': 37.52516174316406, 'timing/data': 0.0007802929903846234, 'timing/model': 1.2102047090011183, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 1.11}) +[2026-06-23 23:39:00,658][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.029834285378456116, 'train/grad_norm_pre_clip': 23.610511779785156, 'timing/data': 0.0007481630018446594, 'timing/model': 1.1831083099823445, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 1.11}) +[2026-06-23 23:39:05,038][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.0827912837266922, 'train/grad_norm_pre_clip': 15.303155899047852, 'timing/data': 0.0008338209881912917, 'timing/model': 1.1874150810181163, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 1.11}) +[2026-06-23 23:39:09,455][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.02773885987699032, 'train/grad_norm_pre_clip': 17.391935348510742, 'timing/data': 0.0007606059953104705, 'timing/model': 1.2113210040261038, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 1.11}) +[2026-06-23 23:39:14,037][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.06839330494403839, 'train/grad_norm_pre_clip': 9.04824447631836, 'timing/data': 0.0007536020129919052, 'timing/model': 1.1813819509989116, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 1.11}) +[2026-06-23 23:39:18,421][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.03450535610318184, 'train/grad_norm_pre_clip': 45.572731018066406, 'timing/data': 0.0007570600137114525, 'timing/model': 1.2012695249868557, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 1.11}) +[2026-06-23 23:39:22,819][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.05658555403351784, 'train/grad_norm_pre_clip': 13.965170860290527, 'timing/data': 0.0007581920071970671, 'timing/model': 1.1958034420094918, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 1.11}) +[2026-06-23 23:39:27,194][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.12352388352155685, 'train/grad_norm_pre_clip': 38.33041763305664, 'timing/data': 0.0007488639967050403, 'timing/model': 1.1823909999802709, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 1.11}) +[2026-06-23 23:39:31,815][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.16831427812576294, 'train/grad_norm_pre_clip': 13.084726333618164, 'timing/data': 0.0009627190011087805, 'timing/model': 1.2011050840083044, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 1.11}) +[2026-06-23 23:39:36,202][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.007965436205267906, 'train/grad_norm_pre_clip': 23.682966232299805, 'timing/data': 0.00076150699169375, 'timing/model': 1.2048891399754211, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 1.11}) +[2026-06-23 23:39:40,552][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.0023479866795241833, 'train/grad_norm_pre_clip': 20.42154312133789, 'timing/data': 0.0007262119906954467, 'timing/model': 1.1854782400187105, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 1.11}) +[2026-06-23 23:39:44,930][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.19006983935832977, 'train/grad_norm_pre_clip': 22.76202964782715, 'timing/data': 0.0007359810115303844, 'timing/model': 1.2063340840104502, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 1.11}) +[2026-06-23 23:39:49,554][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.11528903990983963, 'train/grad_norm_pre_clip': 27.112930297851562, 'timing/data': 0.0009732990001793951, 'timing/model': 1.2064646359940525, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 1.11}) +[2026-06-23 23:39:53,924][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.03876174986362457, 'train/grad_norm_pre_clip': 11.218399047851562, 'timing/data': 0.000738525006454438, 'timing/model': 1.18080847300007, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 1.11}) +[2026-06-23 23:39:58,259][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.08170749992132187, 'train/grad_norm_pre_clip': 28.28809356689453, 'timing/data': 0.0007552460010629147, 'timing/model': 1.1892571719945408, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 1.11}) +[2026-06-23 23:40:02,621][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.04789464548230171, 'train/grad_norm_pre_clip': 14.269783020019531, 'timing/data': 0.0007822859915904701, 'timing/model': 1.1807958029967267, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 1.11}) +[2026-06-23 23:40:07,230][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.007642496842890978, 'train/grad_norm_pre_clip': 26.869468688964844, 'timing/data': 0.001183287997264415, 'timing/model': 1.183003626007121, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 1.11}) +[2026-06-23 23:40:11,560][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.13538755476474762, 'train/grad_norm_pre_clip': 15.156200408935547, 'timing/data': 0.0008141050057020038, 'timing/model': 1.1835355410003103, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 1.12}) +[2026-06-23 23:40:15,914][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.05014079809188843, 'train/grad_norm_pre_clip': 23.159427642822266, 'timing/data': 0.0009165040100924671, 'timing/model': 1.2018464099965058, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 1.12}) +[2026-06-23 23:40:20,245][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.06468082964420319, 'train/grad_norm_pre_clip': 74.59567260742188, 'timing/data': 0.0007509179995395243, 'timing/model': 1.185055484995246, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 1.12}) +[2026-06-23 23:40:24,873][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.13265521824359894, 'train/grad_norm_pre_clip': 55.60972213745117, 'timing/data': 0.0009288269793614745, 'timing/model': 1.1834843730030116, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 1.12}) +[2026-06-23 23:40:29,213][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.012489724904298782, 'train/grad_norm_pre_clip': 21.515121459960938, 'timing/data': 0.000781795009970665, 'timing/model': 1.1956156289961655, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 1.12}) +[2026-06-23 23:40:33,605][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.04076281562447548, 'train/grad_norm_pre_clip': 21.30623435974121, 'timing/data': 0.0009360710100736469, 'timing/model': 1.2019397910044063, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 1.12}) +[2026-06-23 23:40:37,960][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.05136389285326004, 'train/grad_norm_pre_clip': 14.047322273254395, 'timing/data': 0.0008397620113100857, 'timing/model': 1.1769762609910686, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 1.12}) +[2026-06-23 23:40:42,596][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.05400052294135094, 'train/grad_norm_pre_clip': 20.6138973236084, 'timing/data': 0.0008475070062559098, 'timing/model': 1.200527425011387, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 1.12}) +[2026-06-23 23:40:46,940][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.04827612638473511, 'train/grad_norm_pre_clip': 24.051660537719727, 'timing/data': 0.0011370519932825118, 'timing/model': 1.1808563070080709, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 1.12}) +[2026-06-23 23:40:51,290][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.055905621498823166, 'train/grad_norm_pre_clip': 11.935836791992188, 'timing/data': 0.0008378089987672865, 'timing/model': 1.1787300689902622, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 1.12}) +[2026-06-23 23:40:55,630][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.016487590968608856, 'train/grad_norm_pre_clip': 6.754199028015137, 'timing/data': 0.000742192001780495, 'timing/model': 1.1820301820116583, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 1.12}) +[2026-06-23 23:41:00,243][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.09846081584692001, 'train/grad_norm_pre_clip': 22.77556610107422, 'timing/data': 0.0008722819911781698, 'timing/model': 1.1787763999891467, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.12}) +[2026-06-23 23:41:04,579][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.03487517312169075, 'train/grad_norm_pre_clip': 10.902586936950684, 'timing/data': 0.0007704150048084557, 'timing/model': 1.1787413080164697, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 1.12}) +[2026-06-23 23:41:08,962][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.019615206867456436, 'train/grad_norm_pre_clip': 26.613901138305664, 'timing/data': 0.0008363560191355646, 'timing/model': 1.1788663629849907, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 1.12}) +[2026-06-23 23:41:13,341][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.12720707058906555, 'train/grad_norm_pre_clip': 32.84320831298828, 'timing/data': 0.0008001789974514395, 'timing/model': 1.1981288490060251, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 1.12}) +[2026-06-23 23:41:17,974][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.2524692118167877, 'train/grad_norm_pre_clip': 23.29045867919922, 'timing/data': 0.0007778379949741066, 'timing/model': 1.193646717001684, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 1.12}) +[2026-06-23 23:41:22,346][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.059591975063085556, 'train/grad_norm_pre_clip': 14.640033721923828, 'timing/data': 0.00122320101945661, 'timing/model': 1.1991284730029292, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 1.12}) +[2026-06-23 23:41:26,699][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.04513275995850563, 'train/grad_norm_pre_clip': 13.657795906066895, 'timing/data': 0.000927703978959471, 'timing/model': 1.1781093339959625, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 1.12}) +[2026-06-23 23:41:31,046][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2900 +[2026-06-23 23:42:28,552][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2900 +[2026-06-23 23:42:28,553][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2900 +[2026-06-23 23:43:05,512][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2900 +[2026-06-23 23:43:05,514][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.07045476138591766, 'train/grad_norm_pre_clip': 19.85099220275879, 'eval/action_loss/seconds': 57.503418887994485, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.035509753972291946, 'eval/latency_2/loss': 0.035509753972291946, 'eval/flappy/loss': 0.035509753972291946, 'eval/flappy/latency_2/loss': 0.035509753972291946, 'eval/flap_precision': 0.9716216216216216, 'eval/flap_recall': 0.9755766621438263, 'eval/flap_f1': 0.973595125253893, 'eval/flap_rate_model': 0.20555555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9735951252538931, 'eval/latency_2/flap_precision': 0.9716216216216216, 'eval/latency_2/flap_recall': 0.9755766621438263, 'eval/latency_2/flap_f1': 0.973595125253893, 'eval/latency_2/flap_rate_model': 0.20555555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9735951252538931, 'eval/action_classification/data_wait_seconds': 0.5143282413482666, 'eval/action_classification/predict_seconds': 36.3020133972168, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007223649881780148, 'timing/model': 1.2088864350225776, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 1.12}) +[2026-06-23 23:43:09,803][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.23463910818099976, 'train/grad_norm_pre_clip': 24.850482940673828, 'timing/data': 0.0007004939834587276, 'timing/model': 1.1834558400150854, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 1.12}) +[2026-06-23 23:43:14,143][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.026545777916908264, 'train/grad_norm_pre_clip': 9.80161190032959, 'timing/data': 0.0008875000057742, 'timing/model': 1.194911256985506, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 1.12}) +[2026-06-23 23:43:18,453][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.030839379876852036, 'train/grad_norm_pre_clip': 21.173202514648438, 'timing/data': 0.0007720669964328408, 'timing/model': 1.1988487479975447, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 1.12}) +[2026-06-23 23:43:23,040][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.004396697506308556, 'train/grad_norm_pre_clip': 24.146286010742188, 'timing/data': 0.000751379004213959, 'timing/model': 1.191452342987759, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 1.12}) +[2026-06-23 23:43:27,384][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.009269081987440586, 'train/grad_norm_pre_clip': 11.914984703063965, 'timing/data': 0.000908579007955268, 'timing/model': 1.1831977599940728, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 1.12}) +[2026-06-23 23:43:31,665][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.06573998183012009, 'train/grad_norm_pre_clip': 14.28960132598877, 'timing/data': 0.0007409800018649548, 'timing/model': 1.1702665869961493, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 1.12}) +[2026-06-23 23:43:36,171][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.1525258719921112, 'train/grad_norm_pre_clip': 29.80264663696289, 'timing/data': 0.0009400379785802215, 'timing/model': 1.3471967320074327, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 1.13}) +[2026-06-23 23:43:40,508][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.06934456527233124, 'train/grad_norm_pre_clip': 7.01720666885376, 'timing/data': 0.0007264130108524114, 'timing/model': 1.194250482978532, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 1.13}) +[2026-06-23 23:43:44,832][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.022378820925951004, 'train/grad_norm_pre_clip': 8.414591789245605, 'timing/data': 0.0007527910056523979, 'timing/model': 1.1887507329811342, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 1.13}) +[2026-06-23 23:43:49,178][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.07053018361330032, 'train/grad_norm_pre_clip': 8.260092735290527, 'timing/data': 0.0011654740083031356, 'timing/model': 1.2042723239865154, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 1.13}) +[2026-06-23 23:43:53,788][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.02082211896777153, 'train/grad_norm_pre_clip': 6.09151554107666, 'timing/data': 0.0008967870089691132, 'timing/model': 1.446573495020857, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 1.13}) +[2026-06-23 23:43:58,114][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.04833483695983887, 'train/grad_norm_pre_clip': 8.978948593139648, 'timing/data': 0.0008124420128297061, 'timing/model': 1.1706923109886702, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 1.13}) +[2026-06-23 23:44:02,473][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.05482657253742218, 'train/grad_norm_pre_clip': 19.516328811645508, 'timing/data': 0.0010134129843208939, 'timing/model': 1.1936224660021253, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 1.13}) +[2026-06-23 23:44:06,810][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.24653293192386627, 'train/grad_norm_pre_clip': 23.115739822387695, 'timing/data': 0.0007365709752775729, 'timing/model': 1.1732546679850202, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 1.13}) +[2026-06-23 23:44:11,317][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.030114976689219475, 'train/grad_norm_pre_clip': 23.5718936920166, 'timing/data': 0.0009491739911027253, 'timing/model': 1.2000076069962233, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 1.13}) +[2026-06-23 23:44:15,684][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.04488205909729004, 'train/grad_norm_pre_clip': 18.683141708374023, 'timing/data': 0.0007880569901317358, 'timing/model': 1.1925534650217742, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 1.13}) +[2026-06-23 23:44:20,064][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.07563220709562302, 'train/grad_norm_pre_clip': 10.417045593261719, 'timing/data': 0.0007572299800813198, 'timing/model': 1.1983919909980614, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 1.13}) +[2026-06-23 23:44:24,389][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.003887537866830826, 'train/grad_norm_pre_clip': 27.004682540893555, 'timing/data': 0.0007906309911049902, 'timing/model': 1.1936281109810807, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 1.13}) +[2026-06-23 23:44:28,878][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.16012856364250183, 'train/grad_norm_pre_clip': 41.2148323059082, 'timing/data': 0.0007300699944607913, 'timing/model': 1.1914350830193143, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 1.13}) +[2026-06-23 23:44:33,216][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.27144065499305725, 'train/grad_norm_pre_clip': 44.956790924072266, 'timing/data': 0.0017302880005445331, 'timing/model': 1.174493208003696, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 1.13}) +[2026-06-23 23:44:37,556][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.1111079752445221, 'train/grad_norm_pre_clip': 38.38017272949219, 'timing/data': 0.0008531569910701364, 'timing/model': 1.1912180119834375, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 1.13}) +[2026-06-23 23:44:41,934][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.02389906719326973, 'train/grad_norm_pre_clip': 20.29553985595703, 'timing/data': 0.0007325139886233956, 'timing/model': 1.1827500479994342, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 1.13}) +[2026-06-23 23:44:46,549][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.0067640733905136585, 'train/grad_norm_pre_clip': 23.610260009765625, 'timing/data': 0.0007289070053957403, 'timing/model': 1.1753559979842976, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 1.13}) +[2026-06-23 23:44:50,894][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.04099294915795326, 'train/grad_norm_pre_clip': 17.535432815551758, 'timing/data': 0.000722124008461833, 'timing/model': 1.204545207991032, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 1.13}) +[2026-06-23 23:44:55,262][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.09398017078638077, 'train/grad_norm_pre_clip': 42.07568359375, 'timing/data': 0.0011984349985141307, 'timing/model': 1.2021210330130998, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 1.13}) +[2026-06-23 23:44:59,624][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.03705789893865585, 'train/grad_norm_pre_clip': 14.959601402282715, 'timing/data': 0.0007369720260612667, 'timing/model': 1.1961408170172945, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 1.13}) +[2026-06-23 23:45:04,134][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.11728669703006744, 'train/grad_norm_pre_clip': 42.8980598449707, 'timing/data': 0.0007037209870759398, 'timing/model': 1.196227562992135, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 1.13}) +[2026-06-23 23:45:08,454][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.016095034778118134, 'train/grad_norm_pre_clip': 17.768848419189453, 'timing/data': 0.0007303099846467376, 'timing/model': 1.192465075000655, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 1.13}) +[2026-06-23 23:45:12,799][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.17623400688171387, 'train/grad_norm_pre_clip': 40.06190872192383, 'timing/data': 0.0007693619991187006, 'timing/model': 1.19674999898416, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 1.13}) +[2026-06-23 23:45:17,126][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.015039270743727684, 'train/grad_norm_pre_clip': 35.31812286376953, 'timing/data': 0.0009290369926020503, 'timing/model': 1.1791001159872394, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 1.13}) +[2026-06-23 23:45:21,668][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.0066756377927958965, 'train/grad_norm_pre_clip': 16.24372100830078, 'timing/data': 0.0007784379995428026, 'timing/model': 1.3982267790124752, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 1.13}) +[2026-06-23 23:45:26,036][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.03542306274175644, 'train/grad_norm_pre_clip': 39.95475387573242, 'timing/data': 0.000851744000101462, 'timing/model': 1.2034279479994439, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 1.13}) +[2026-06-23 23:45:30,354][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.050535283982753754, 'train/grad_norm_pre_clip': 28.183942794799805, 'timing/data': 0.0007300989818759263, 'timing/model': 1.1789306789869443, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 1.14}) +[2026-06-23 23:45:34,712][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.07437319308519363, 'train/grad_norm_pre_clip': 17.35198402404785, 'timing/data': 0.0007748320058453828, 'timing/model': 1.1941582179861143, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 1.14}) +[2026-06-23 23:45:39,292][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.008165508508682251, 'train/grad_norm_pre_clip': 19.879112243652344, 'timing/data': 0.0007355289999395609, 'timing/model': 1.4288163110031746, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 1.14}) +[2026-06-23 23:45:43,591][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.04419520124793053, 'train/grad_norm_pre_clip': 21.076610565185547, 'timing/data': 0.0008627540082670748, 'timing/model': 1.1738001760095358, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 1.14}) +[2026-06-23 23:45:47,941][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.11081285029649734, 'train/grad_norm_pre_clip': 22.820825576782227, 'timing/data': 0.0007096719928085804, 'timing/model': 1.1915217420028057, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 1.14}) +[2026-06-23 23:45:52,300][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.062277816236019135, 'train/grad_norm_pre_clip': 45.17502975463867, 'timing/data': 0.0007419509929604828, 'timing/model': 1.1955753189977258, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 1.14}) +[2026-06-23 23:45:56,819][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.00021990742243360728, 'train/grad_norm_pre_clip': 20.368087768554688, 'timing/data': 0.0007295980176422745, 'timing/model': 1.17829652299406, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 1.14}) +[2026-06-23 23:46:01,177][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.03479563444852829, 'train/grad_norm_pre_clip': 43.56298065185547, 'timing/data': 0.0011248279770370573, 'timing/model': 1.1762740609992761, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 1.14}) +[2026-06-23 23:46:05,505][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.06552689522504807, 'train/grad_norm_pre_clip': 35.45443344116211, 'timing/data': 0.0007597729854751378, 'timing/model': 1.1941314290161245, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 1.14}) +[2026-06-23 23:46:09,822][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.008906776085495949, 'train/grad_norm_pre_clip': 7.4312639236450195, 'timing/data': 0.0007173559861257672, 'timing/model': 1.1753038599854335, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.14}) +[2026-06-23 23:46:14,442][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.08729396015405655, 'train/grad_norm_pre_clip': 29.04643440246582, 'timing/data': 0.0007437840104103088, 'timing/model': 1.1976574159925804, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 1.14}) +[2026-06-23 23:46:18,809][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.05684056505560875, 'train/grad_norm_pre_clip': 18.21965789794922, 'timing/data': 0.0008264770149253309, 'timing/model': 1.1807302609959152, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 1.14}) +[2026-06-23 23:46:23,171][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.04340134933590889, 'train/grad_norm_pre_clip': 15.37053108215332, 'timing/data': 0.0007176570070441812, 'timing/model': 1.1935391890001483, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 1.14}) +[2026-06-23 23:46:27,507][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.0009525719215162098, 'train/grad_norm_pre_clip': 13.405172348022461, 'timing/data': 0.0008092159987427294, 'timing/model': 1.1997606820077635, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 1.14}) +[2026-06-23 23:46:31,995][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.017910463735461235, 'train/grad_norm_pre_clip': 11.5781888961792, 'timing/data': 0.0007219639956019819, 'timing/model': 1.1769945159903727, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 1.14}) +[2026-06-23 23:46:36,386][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.03324904665350914, 'train/grad_norm_pre_clip': 12.20832347869873, 'timing/data': 0.0009791300108190626, 'timing/model': 1.1973642359953374, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 1.14}) +[2026-06-23 23:46:40,751][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.007563543040305376, 'train/grad_norm_pre_clip': 14.391843795776367, 'timing/data': 0.0010881010093726218, 'timing/model': 1.1826457410061266, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 1.14}) +[2026-06-23 23:46:45,130][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.12107493728399277, 'train/grad_norm_pre_clip': 17.018217086791992, 'timing/data': 0.0011227840150240809, 'timing/model': 1.1957582299946807, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 1.14}) +[2026-06-23 23:46:49,684][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.03252319619059563, 'train/grad_norm_pre_clip': 26.861948013305664, 'timing/data': 0.0007700829883106053, 'timing/model': 1.4111225160013419, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 1.14}) +[2026-06-23 23:46:54,073][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.052218034863471985, 'train/grad_norm_pre_clip': 75.98238372802734, 'timing/data': 0.0007924239907879382, 'timing/model': 1.1975073749781586, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 1.14}) +[2026-06-23 23:46:58,414][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.09007063508033752, 'train/grad_norm_pre_clip': 22.446107864379883, 'timing/data': 0.0007751019729766995, 'timing/model': 1.1827273670060094, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 1.14}) +[2026-06-23 23:47:02,747][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.07120969891548157, 'train/grad_norm_pre_clip': 35.58308029174805, 'timing/data': 0.0011693710112012923, 'timing/model': 1.1801298240025062, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 1.14}) +[2026-06-23 23:47:07,075][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.021241534501314163, 'train/grad_norm_pre_clip': 62.11742401123047, 'timing/data': 0.0008070010226219893, 'timing/model': 1.18307026399998, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 1.14}) +[2026-06-23 23:47:11,702][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.0034741442650556564, 'train/grad_norm_pre_clip': 10.724212646484375, 'timing/data': 0.0010292219812981784, 'timing/model': 1.2110469009785447, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 1.14}) +[2026-06-23 23:47:16,051][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.03387438878417015, 'train/grad_norm_pre_clip': 34.606868743896484, 'timing/data': 0.0009054129768628627, 'timing/model': 1.1987301670014858, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 1.14}) +[2026-06-23 23:47:20,414][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.05520637705922127, 'train/grad_norm_pre_clip': 28.488279342651367, 'timing/data': 0.0009319620148744434, 'timing/model': 1.187477608997142, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 1.14}) +[2026-06-23 23:47:24,759][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.03835199773311615, 'train/grad_norm_pre_clip': 20.222856521606445, 'timing/data': 0.001133213983848691, 'timing/model': 1.1779535239911638, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 1.15}) +[2026-06-23 23:47:29,266][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.023857198655605316, 'train/grad_norm_pre_clip': 11.51993465423584, 'timing/data': 0.0008063300047069788, 'timing/model': 1.1845509429986123, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 1.15}) +[2026-06-23 23:47:33,634][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.0686945840716362, 'train/grad_norm_pre_clip': 15.975959777832031, 'timing/data': 0.0008671119867358357, 'timing/model': 1.1974763059988618, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 1.15}) +[2026-06-23 23:47:38,008][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.0089913634583354, 'train/grad_norm_pre_clip': 19.021873474121094, 'timing/data': 0.0007444249931722879, 'timing/model': 1.1798268859856762, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 1.15}) +[2026-06-23 23:47:42,348][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.07193329185247421, 'train/grad_norm_pre_clip': 15.662131309509277, 'timing/data': 0.0007809829839970917, 'timing/model': 1.1934150839806534, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 1.15}) +[2026-06-23 23:47:46,877][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.0641014575958252, 'train/grad_norm_pre_clip': 23.696754455566406, 'timing/data': 0.0008554600062780082, 'timing/model': 1.1727306560205761, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 1.15}) +[2026-06-23 23:47:51,231][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.1861727386713028, 'train/grad_norm_pre_clip': 170.19259643554688, 'timing/data': 0.0007879459881223738, 'timing/model': 1.2139894690189976, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 1.15}) +[2026-06-23 23:47:55,572][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.16918963193893433, 'train/grad_norm_pre_clip': 26.49717140197754, 'timing/data': 0.0007235969824250787, 'timing/model': 1.201520372997038, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 1.15}) +[2026-06-23 23:47:59,950][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.05769976228475571, 'train/grad_norm_pre_clip': 19.225061416625977, 'timing/data': 0.000806982017820701, 'timing/model': 1.2143182999861892, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 1.15}) +[2026-06-23 23:48:04,490][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.18753576278686523, 'train/grad_norm_pre_clip': 26.679100036621094, 'timing/data': 0.0008879010274540633, 'timing/model': 1.1998687029990833, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 1.15}) +[2026-06-23 23:48:08,824][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.04266177862882614, 'train/grad_norm_pre_clip': 26.327835083007812, 'timing/data': 0.0008417850185651332, 'timing/model': 1.1907253639947157, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 1.15}) +[2026-06-23 23:48:13,209][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.13526885211467743, 'train/grad_norm_pre_clip': 29.61785125732422, 'timing/data': 0.0008098859980236739, 'timing/model': 1.1823097279993817, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 1.15}) +[2026-06-23 23:48:17,582][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.06910441815853119, 'train/grad_norm_pre_clip': 23.5207462310791, 'timing/data': 0.0008787840197328478, 'timing/model': 1.2066319019941147, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 1.15}) +[2026-06-23 23:48:22,162][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.06857284158468246, 'train/grad_norm_pre_clip': 17.675540924072266, 'timing/data': 0.0007760640000924468, 'timing/model': 1.2016840059950482, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 1.15}) +[2026-06-23 23:48:26,518][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.041715845465660095, 'train/grad_norm_pre_clip': 20.32308006286621, 'timing/data': 0.0011708539968822151, 'timing/model': 1.2076821069931611, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 1.15}) +[2026-06-23 23:48:30,889][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.030076295137405396, 'train/grad_norm_pre_clip': 274.5321044921875, 'timing/data': 0.0007778680010233074, 'timing/model': 1.1980142579996027, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 1.15}) +[2026-06-23 23:48:35,289][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.08337106555700302, 'train/grad_norm_pre_clip': 24.352027893066406, 'timing/data': 0.0007565380074083805, 'timing/model': 1.2041491419950034, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 1.15}) +[2026-06-23 23:48:39,879][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.035541653633117676, 'train/grad_norm_pre_clip': 29.745960235595703, 'timing/data': 0.0007132079917937517, 'timing/model': 1.211722064006608, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 1.15}) +[2026-06-23 23:48:44,260][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.002207301091402769, 'train/grad_norm_pre_clip': 11.60098648071289, 'timing/data': 0.0008822700183372945, 'timing/model': 1.1843229179794434, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 1.15}) +[2026-06-23 23:48:48,637][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.07987631857395172, 'train/grad_norm_pre_clip': 9.693807601928711, 'timing/data': 0.0010685150045901537, 'timing/model': 1.1962196090025827, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 1.15}) +[2026-06-23 23:48:53,021][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.11165841668844223, 'train/grad_norm_pre_clip': 25.07680892944336, 'timing/data': 0.0008388899732381105, 'timing/model': 1.183997640007874, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 1.15}) +[2026-06-23 23:48:57,530][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.04899405688047409, 'train/grad_norm_pre_clip': 11.376321792602539, 'timing/data': 0.0007989869918674231, 'timing/model': 1.1788894580095075, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 1.15}) +[2026-06-23 23:49:01,871][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.013124262914061546, 'train/grad_norm_pre_clip': 15.87865924835205, 'timing/data': 0.0007989760197233409, 'timing/model': 1.1784826740040444, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 1.15}) +[2026-06-23 23:49:06,235][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.035039499402046204, 'train/grad_norm_pre_clip': 7.534911632537842, 'timing/data': 0.0008053990022744983, 'timing/model': 1.2071125270158518, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 1.15}) +[2026-06-23 23:49:10,551][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.02239282988011837, 'train/grad_norm_pre_clip': 13.603732109069824, 'timing/data': 0.0007735090039204806, 'timing/model': 1.1767930560163222, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 1.15}) +[2026-06-23 23:49:15,080][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.04628629609942436, 'train/grad_norm_pre_clip': 10.059425354003906, 'timing/data': 0.0008061100088525563, 'timing/model': 1.1776624500053003, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 1.15}) +[2026-06-23 23:49:19,448][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.19318610429763794, 'train/grad_norm_pre_clip': 23.430360794067383, 'timing/data': 0.0009298590011894703, 'timing/model': 1.2001699899847154, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 1.16}) +[2026-06-23 23:49:23,796][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.01805824041366577, 'train/grad_norm_pre_clip': 11.18497371673584, 'timing/data': 0.0008500200056005269, 'timing/model': 1.2070027730078436, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 1.16}) +[2026-06-23 23:49:28,182][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.023516297340393066, 'train/grad_norm_pre_clip': 10.617494583129883, 'timing/data': 0.0007289680070243776, 'timing/model': 1.2027041569817811, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 1.16}) +[2026-06-23 23:49:32,771][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.1689848154783249, 'train/grad_norm_pre_clip': 29.68988609313965, 'timing/data': 0.0007663560099899769, 'timing/model': 1.2016717859951314, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 1.16}) +[2026-06-23 23:49:37,180][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.010911654680967331, 'train/grad_norm_pre_clip': 22.674667358398438, 'timing/data': 0.001255259005120024, 'timing/model': 1.2006286660034675, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 1.16}) +[2026-06-23 23:49:41,550][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.02253628894686699, 'train/grad_norm_pre_clip': 9.01916217803955, 'timing/data': 0.001382935995934531, 'timing/model': 1.1970507810183335, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 1.16}) +[2026-06-23 23:49:45,933][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.04642968997359276, 'train/grad_norm_pre_clip': 22.511343002319336, 'timing/data': 0.0007571190071757883, 'timing/model': 1.1978634510014672, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.16}) +[2026-06-23 23:49:50,524][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.005393541418015957, 'train/grad_norm_pre_clip': 8.49216365814209, 'timing/data': 0.0007512180018238723, 'timing/model': 1.199486057012109, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 1.16}) +[2026-06-23 23:49:54,871][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.07947816699743271, 'train/grad_norm_pre_clip': 17.494705200195312, 'timing/data': 0.0007348379876930267, 'timing/model': 1.1930980400065891, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 1.16}) +[2026-06-23 23:49:59,218][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.06869290769100189, 'train/grad_norm_pre_clip': 15.939957618713379, 'timing/data': 0.0009792799828574061, 'timing/model': 1.2040701330115553, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 1.16}) +[2026-06-23 23:50:03,568][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.0003427452757023275, 'train/grad_norm_pre_clip': 20.70586585998535, 'timing/data': 0.0010082029912155122, 'timing/model': 1.1786500460002571, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 1.16}) +[2026-06-23 23:50:08,122][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.08436375856399536, 'train/grad_norm_pre_clip': 20.50330924987793, 'timing/data': 0.0008408440044149756, 'timing/model': 1.2095623769855592, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 1.16}) +[2026-06-23 23:50:12,530][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.047148317098617554, 'train/grad_norm_pre_clip': 14.1055908203125, 'timing/data': 0.0007132980099413544, 'timing/model': 1.2072818700107746, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 1.16}) +[2026-06-23 23:50:16,864][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.008254123851656914, 'train/grad_norm_pre_clip': 28.409395217895508, 'timing/data': 0.0007694320229347795, 'timing/model': 1.178023511020001, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 1.16}) +[2026-06-23 23:50:21,222][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.01964355818927288, 'train/grad_norm_pre_clip': 12.338091850280762, 'timing/data': 0.0008078029786702245, 'timing/model': 1.177846670994768, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 1.16}) +[2026-06-23 23:50:25,769][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3000 +[2026-06-23 23:51:23,564][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3000 +[2026-06-23 23:51:23,565][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3000 +[2026-06-23 23:52:00,391][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3000 +[2026-06-24 00:18:41,735][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3000 +[2026-06-24 00:18:41,737][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.08934852480888367, 'train/grad_norm_pre_clip': 27.819875717163086, 'eval/action_loss/seconds': 57.792127979017096, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.037826403975486755, 'eval/latency_2/loss': 0.037826403975486755, 'eval/flappy/loss': 0.037826403975486755, 'eval/flappy/latency_2/loss': 0.037826403975486755, 'eval/flap_precision': 0.9679572763684913, 'eval/flap_recall': 0.9837177747625508, 'eval/flap_f1': 0.9757738896366083, 'eval/flap_rate_model': 0.20805555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9757738896366082, 'eval/latency_2/flap_precision': 0.9679572763684913, 'eval/latency_2/flap_recall': 0.9837177747625508, 'eval/latency_2/flap_f1': 0.9757738896366083, 'eval/latency_2/flap_rate_model': 0.20805555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9757738896366082, 'eval/action_classification/data_wait_seconds': 0.41095682978630066, 'eval/action_classification/predict_seconds': 36.26315689086914, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 419.8600000000212, 'rl_games_eval/mid_train/mean_length': 3401.6, 'rl_games_eval/mid_train/std_reward': 49.48000000000879, 'rl_games_eval/mid_train/std_length': 396.8, 'rl_games_eval/mid_train/macro_mean_reward': 419.8600000000212, 'rl_games_eval/mid_train/macro_mean_length': 3401.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 419.8600000000212, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3401.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 49.48000000000879, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 396.8, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 419.8600000000212, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0007757740095257759, 'timing/model': 1.198075449996395, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 1.16}) +[2026-06-24 00:18:41,737][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_3000_state +[2026-06-24 00:18:41,744][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 00:18:50,677][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_3000_state/model.safetensors +[2026-06-24 00:19:00,954][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_3000_state/optimizer.bin +[2026-06-24 00:19:00,955][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_3000_state/sampler.bin +[2026-06-24 00:19:00,955][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-24 00:19:00,956][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-24 00:19:01,159][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-24 00:19:01,159][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 00:19:01,186][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 00:19:05,353][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.015523037873208523, 'train/grad_norm_pre_clip': 24.358421325683594, 'timing/data': 0.0011677470174618065, 'timing/model': 1.140579141996568, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 1.16}) +[2026-06-24 00:19:09,760][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.07013946026563644, 'train/grad_norm_pre_clip': 24.168594360351562, 'timing/data': 0.000877751997904852, 'timing/model': 1.3608458330272697, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 1.16}) +[2026-06-24 00:19:14,006][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.10755327343940735, 'train/grad_norm_pre_clip': 30.00652313232422, 'timing/data': 0.0009909910149872303, 'timing/model': 1.1741440189944115, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 1.16}) +[2026-06-24 00:19:18,242][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.1187925636768341, 'train/grad_norm_pre_clip': 34.75794982910156, 'timing/data': 0.0007189990137703717, 'timing/model': 1.1575738620012999, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 1.16}) +[2026-06-24 00:19:22,481][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.0019671483896672726, 'train/grad_norm_pre_clip': 24.89169692993164, 'timing/data': 0.0007281959988176823, 'timing/model': 1.1619879880163353, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 1.16}) +[2026-06-24 00:19:26,866][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.12943653762340546, 'train/grad_norm_pre_clip': 17.924039840698242, 'timing/data': 0.0008811689913272858, 'timing/model': 1.3266823299927637, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 1.16}) +[2026-06-24 00:19:31,059][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.03280692175030708, 'train/grad_norm_pre_clip': 10.28811264038086, 'timing/data': 0.0008605199982412159, 'timing/model': 1.1468749540217686, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 1.16}) +[2026-06-24 00:19:35,247][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.009554484859108925, 'train/grad_norm_pre_clip': 64.5108642578125, 'timing/data': 0.0007491340220440179, 'timing/model': 1.1455060850130394, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 1.16}) +[2026-06-24 00:19:39,488][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.0030064028687775135, 'train/grad_norm_pre_clip': 19.77370834350586, 'timing/data': 0.0007392459956463426, 'timing/model': 1.1489925260248128, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 1.16}) +[2026-06-24 00:19:43,695][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.018477775156497955, 'train/grad_norm_pre_clip': 32.949893951416016, 'timing/data': 0.0007203109853435308, 'timing/model': 1.1509468570002355, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 1.16}) +[2026-06-24 00:19:48,205][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.0159770380705595, 'train/grad_norm_pre_clip': 14.876688957214355, 'timing/data': 0.0008913360070437193, 'timing/model': 1.1690202510217205, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 1.17}) +[2026-06-24 00:19:52,430][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.0545150451362133, 'train/grad_norm_pre_clip': 28.960289001464844, 'timing/data': 0.000738415023079142, 'timing/model': 1.1736284250218887, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 1.17}) +[2026-06-24 00:19:56,665][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.016130652278661728, 'train/grad_norm_pre_clip': 12.20071792602539, 'timing/data': 0.0008887920121196657, 'timing/model': 1.1536852470017038, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 1.17}) +[2026-06-24 00:20:00,945][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.014185221865773201, 'train/grad_norm_pre_clip': 16.653186798095703, 'timing/data': 0.0007799809973221272, 'timing/model': 1.172897173993988, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 1.17}) +[2026-06-24 00:20:05,461][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.0012897514970973134, 'train/grad_norm_pre_clip': 10.380524635314941, 'timing/data': 0.0011114039807580411, 'timing/model': 1.1746310279995669, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 1.17}) +[2026-06-24 00:20:09,735][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.015015351586043835, 'train/grad_norm_pre_clip': 16.275726318359375, 'timing/data': 0.0007771560049150139, 'timing/model': 1.1570498129876796, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 1.17}) +[2026-06-24 00:20:14,069][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.015651661902666092, 'train/grad_norm_pre_clip': 21.050151824951172, 'timing/data': 0.0008752779976930469, 'timing/model': 1.1996280040184502, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 1.17}) +[2026-06-24 00:20:18,316][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.033701762557029724, 'train/grad_norm_pre_clip': 19.436513900756836, 'timing/data': 0.000872292002895847, 'timing/model': 1.1601197150012013, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 1.17}) +[2026-06-24 00:20:22,820][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.0857883021235466, 'train/grad_norm_pre_clip': 20.985807418823242, 'timing/data': 0.0007819239981472492, 'timing/model': 1.1630312649940606, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 1.17}) +[2026-06-24 00:20:27,081][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.11429765820503235, 'train/grad_norm_pre_clip': 11.384190559387207, 'timing/data': 0.0007272639777511358, 'timing/model': 1.1623469199985266, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 1.17}) +[2026-06-24 00:20:31,411][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.05737140774726868, 'train/grad_norm_pre_clip': 15.111574172973633, 'timing/data': 0.0007426429947372526, 'timing/model': 1.183773819997441, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 1.17}) +[2026-06-24 00:20:35,881][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.11150528490543365, 'train/grad_norm_pre_clip': 24.07190704345703, 'timing/data': 0.0007165040005929768, 'timing/model': 1.3350446590047795, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 1.17}) +[2026-06-24 00:20:40,160][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.015775758773088455, 'train/grad_norm_pre_clip': 15.525188446044922, 'timing/data': 0.0007567880093120039, 'timing/model': 1.165635705983732, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 1.17}) +[2026-06-24 00:20:44,425][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.05164526402950287, 'train/grad_norm_pre_clip': 11.527167320251465, 'timing/data': 0.0008486380102112889, 'timing/model': 1.1643808580120094, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 1.17}) +[2026-06-24 00:20:48,727][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.007316830568015575, 'train/grad_norm_pre_clip': 8.895577430725098, 'timing/data': 0.0007090589788276702, 'timing/model': 1.1916447730036452, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 1.17}) +[2026-06-24 00:20:53,052][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.15319105982780457, 'train/grad_norm_pre_clip': 20.745956420898438, 'timing/data': 0.0007080490177031606, 'timing/model': 1.1888197630178183, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 1.17}) +[2026-06-24 00:20:57,589][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.028369933366775513, 'train/grad_norm_pre_clip': 8.82020378112793, 'timing/data': 0.0007543329847976565, 'timing/model': 1.1679947650118265, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 1.17}) +[2026-06-24 00:21:01,924][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.05055353417992592, 'train/grad_norm_pre_clip': 26.537694931030273, 'timing/data': 0.0011361090000718832, 'timing/model': 1.1868277530011255, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 1.17}) +[2026-06-24 00:21:06,227][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.01941860094666481, 'train/grad_norm_pre_clip': 15.477828025817871, 'timing/data': 0.0007058439950924367, 'timing/model': 1.1683434089936782, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 1.17}) +[2026-06-24 00:21:10,533][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.1611558049917221, 'train/grad_norm_pre_clip': 25.528833389282227, 'timing/data': 0.0007295569812413305, 'timing/model': 1.1712035739910789, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 1.17}) +[2026-06-24 00:21:15,081][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.001672790851444006, 'train/grad_norm_pre_clip': 13.605412483215332, 'timing/data': 0.0011231259850319475, 'timing/model': 1.171581011993112, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 1.17}) +[2026-06-24 00:21:19,379][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.011550291441380978, 'train/grad_norm_pre_clip': 23.72014045715332, 'timing/data': 0.0007375429850071669, 'timing/model': 1.176722300995607, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 1.17}) +[2026-06-24 00:21:23,725][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.1266910433769226, 'train/grad_norm_pre_clip': 26.647628784179688, 'timing/data': 0.0006949540111236274, 'timing/model': 1.2062445890042, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 1.17}) +[2026-06-24 00:21:28,120][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.1487792581319809, 'train/grad_norm_pre_clip': 29.718379974365234, 'timing/data': 0.0007810730021446943, 'timing/model': 1.2287853560119402, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 1.17}) +[2026-06-24 00:21:32,673][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.08389922976493835, 'train/grad_norm_pre_clip': 17.298707962036133, 'timing/data': 0.0008652289980091155, 'timing/model': 1.176364298997214, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 1.17}) +[2026-06-24 00:21:37,017][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.032583534717559814, 'train/grad_norm_pre_clip': 35.373008728027344, 'timing/data': 0.0007088000129442662, 'timing/model': 1.1739751339773647, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 1.17}) +[2026-06-24 00:21:41,353][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.015335366129875183, 'train/grad_norm_pre_clip': 12.098418235778809, 'timing/data': 0.001008222985547036, 'timing/model': 1.1912360480055213, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 1.18}) +[2026-06-24 00:21:45,700][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.017344774678349495, 'train/grad_norm_pre_clip': 29.68389129638672, 'timing/data': 0.0011686299985740334, 'timing/model': 1.199832129990682, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 1.18}) +[2026-06-24 00:21:50,304][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.0010572226019576192, 'train/grad_norm_pre_clip': 24.326805114746094, 'timing/data': 0.0008112789946608245, 'timing/model': 1.1920281909988262, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 1.18}) +[2026-06-24 00:21:54,647][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.011032148264348507, 'train/grad_norm_pre_clip': 19.14126205444336, 'timing/data': 0.0007226350135169923, 'timing/model': 1.1916536590142641, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.18}) +[2026-06-24 00:21:58,965][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.002351420233026147, 'train/grad_norm_pre_clip': 20.146059036254883, 'timing/data': 0.0007846800144761801, 'timing/model': 1.1728269529994577, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 1.18}) +[2026-06-24 00:22:03,306][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.012115809135138988, 'train/grad_norm_pre_clip': 22.082862854003906, 'timing/data': 0.0006999829784035683, 'timing/model': 1.171463753998978, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 1.18}) +[2026-06-24 00:22:07,863][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.009911363013088703, 'train/grad_norm_pre_clip': 11.589903831481934, 'timing/data': 0.0009191879944410175, 'timing/model': 1.1747520790086128, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 1.18}) +[2026-06-24 00:22:12,164][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.0002367819397477433, 'train/grad_norm_pre_clip': 24.840627670288086, 'timing/data': 0.0011576389952097088, 'timing/model': 1.172237132996088, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 1.18}) +[2026-06-24 00:22:16,514][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.002645652275532484, 'train/grad_norm_pre_clip': 14.230709075927734, 'timing/data': 0.0007140889938455075, 'timing/model': 1.1764604260097258, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 1.18}) +[2026-06-24 00:22:20,908][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.07304012775421143, 'train/grad_norm_pre_clip': 27.031625747680664, 'timing/data': 0.0007909510168246925, 'timing/model': 1.2287144550064113, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 1.18}) +[2026-06-24 00:22:25,491][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.008993711322546005, 'train/grad_norm_pre_clip': 13.55244255065918, 'timing/data': 0.0007683699950575829, 'timing/model': 1.1745385059912223, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 1.18}) +[2026-06-24 00:22:29,841][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.08560268580913544, 'train/grad_norm_pre_clip': 14.832008361816406, 'timing/data': 0.0007868339889682829, 'timing/model': 1.1746063299942762, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 1.18}) +[2026-06-24 00:22:34,151][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.07109735161066055, 'train/grad_norm_pre_clip': 34.86151885986328, 'timing/data': 0.0007203010027296841, 'timing/model': 1.1758344100089744, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 1.18}) +[2026-06-24 00:22:38,458][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.02140645682811737, 'train/grad_norm_pre_clip': 23.843507766723633, 'timing/data': 0.0007748019997961819, 'timing/model': 1.1743531730026007, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 1.18}) +[2026-06-24 00:22:43,038][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.01485428400337696, 'train/grad_norm_pre_clip': 6.164544582366943, 'timing/data': 0.0007647820166312158, 'timing/model': 1.1755178339954, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 1.18}) +[2026-06-24 00:22:47,342][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.054144006222486496, 'train/grad_norm_pre_clip': 42.54850769042969, 'timing/data': 0.000778317975345999, 'timing/model': 1.1713915900036227, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 1.18}) +[2026-06-24 00:22:51,644][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.04052763059735298, 'train/grad_norm_pre_clip': 20.65508270263672, 'timing/data': 0.0008193340036086738, 'timing/model': 1.1766667669871822, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 1.18}) +[2026-06-24 00:22:55,981][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.018398059532046318, 'train/grad_norm_pre_clip': 13.31601619720459, 'timing/data': 0.0008103470026981086, 'timing/model': 1.17361505900044, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 1.18}) +[2026-06-24 00:23:00,574][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.03147554397583008, 'train/grad_norm_pre_clip': 20.94887351989746, 'timing/data': 0.000927332992432639, 'timing/model': 1.191724559001159, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 1.18}) +[2026-06-24 00:23:04,910][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.021522818133234978, 'train/grad_norm_pre_clip': 17.378585815429688, 'timing/data': 0.00099261399009265, 'timing/model': 1.1786133429850452, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 1.18}) +[2026-06-24 00:23:09,253][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.0896587073802948, 'train/grad_norm_pre_clip': 16.809236526489258, 'timing/data': 0.0007441749912686646, 'timing/model': 1.2005460579821374, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 1.18}) +[2026-06-24 00:23:13,628][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.010128958150744438, 'train/grad_norm_pre_clip': 9.559237480163574, 'timing/data': 0.0007673979853279889, 'timing/model': 1.1970844919851515, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 1.18}) +[2026-06-24 00:23:18,213][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.12119404226541519, 'train/grad_norm_pre_clip': 27.437047958374023, 'timing/data': 0.0007924340025056154, 'timing/model': 1.1947039639926516, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 1.18}) +[2026-06-24 00:23:22,580][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.24124228954315186, 'train/grad_norm_pre_clip': 42.059226989746094, 'timing/data': 0.0007996769854798913, 'timing/model': 1.1930897010024637, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 1.18}) +[2026-06-24 00:23:26,925][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.07815618813037872, 'train/grad_norm_pre_clip': 17.3803768157959, 'timing/data': 0.0009668459824752063, 'timing/model': 1.200439070991706, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 1.18}) +[2026-06-24 00:23:31,273][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.02119748294353485, 'train/grad_norm_pre_clip': 16.565244674682617, 'timing/data': 0.0007853609859012067, 'timing/model': 1.1960861370025668, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 1.19}) +[2026-06-24 00:23:35,852][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.024950528517365456, 'train/grad_norm_pre_clip': 11.451335906982422, 'timing/data': 0.0010869279794860631, 'timing/model': 1.17621413100278, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 1.19}) +[2026-06-24 00:23:40,200][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.045524775981903076, 'train/grad_norm_pre_clip': 10.917428016662598, 'timing/data': 0.0013496740139089525, 'timing/model': 1.1932446679857094, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 1.19}) +[2026-06-24 00:23:44,527][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.004415492061525583, 'train/grad_norm_pre_clip': 15.204987525939941, 'timing/data': 0.0010796449787449092, 'timing/model': 1.19191237600171, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 1.19}) +[2026-06-24 00:23:48,886][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.006933055818080902, 'train/grad_norm_pre_clip': 18.887535095214844, 'timing/data': 0.000779741007136181, 'timing/model': 1.194628093013307, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 1.19}) +[2026-06-24 00:23:53,493][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.0597323477268219, 'train/grad_norm_pre_clip': 11.717126846313477, 'timing/data': 0.0008977279940154403, 'timing/model': 1.198340923001524, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 1.19}) +[2026-06-24 00:23:57,855][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.09185869246721268, 'train/grad_norm_pre_clip': 12.513039588928223, 'timing/data': 0.0009603739890735596, 'timing/model': 1.1958211279998068, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 1.19}) +[2026-06-24 00:24:02,215][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.034697044640779495, 'train/grad_norm_pre_clip': 19.188674926757812, 'timing/data': 0.0007260010170284659, 'timing/model': 1.1992932030116208, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 1.19}) +[2026-06-24 00:24:06,602][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.026160016655921936, 'train/grad_norm_pre_clip': 21.605918884277344, 'timing/data': 0.0012601079943124205, 'timing/model': 1.1974635719961952, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 1.19}) +[2026-06-24 00:24:11,248][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.00512456800788641, 'train/grad_norm_pre_clip': 30.825592041015625, 'timing/data': 0.001113558013457805, 'timing/model': 1.2052418340172153, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 1.19}) +[2026-06-24 00:24:15,609][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.10993136465549469, 'train/grad_norm_pre_clip': 13.5964994430542, 'timing/data': 0.000972206995356828, 'timing/model': 1.1778357959992718, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 1.19}) +[2026-06-24 00:24:19,991][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.03722762688994408, 'train/grad_norm_pre_clip': 12.3018159866333, 'timing/data': 0.0009569579851813614, 'timing/model': 1.2122693790006451, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 1.19}) +[2026-06-24 00:24:24,397][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.0770820677280426, 'train/grad_norm_pre_clip': 16.34563636779785, 'timing/data': 0.0010711989889387041, 'timing/model': 1.2136160069785547, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 1.19}) +[2026-06-24 00:24:28,968][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.0004439659824129194, 'train/grad_norm_pre_clip': 12.056475639343262, 'timing/data': 0.001017049013171345, 'timing/model': 1.1834339899942279, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 1.19}) +[2026-06-24 00:24:33,340][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.027560219168663025, 'train/grad_norm_pre_clip': 9.35746955871582, 'timing/data': 0.0010578350047580898, 'timing/model': 1.1845845759962685, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 1.19}) +[2026-06-24 00:24:37,676][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.020258497446775436, 'train/grad_norm_pre_clip': 30.300235748291016, 'timing/data': 0.0009230559808202088, 'timing/model': 1.185816703015007, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 1.19}) +[2026-06-24 00:24:42,029][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.0082652997225523, 'train/grad_norm_pre_clip': 13.342032432556152, 'timing/data': 0.0009760830143932253, 'timing/model': 1.1794108169851825, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 1.19}) +[2026-06-24 00:24:46,589][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.0032974169589579105, 'train/grad_norm_pre_clip': 19.16938591003418, 'timing/data': 0.0007153110054787248, 'timing/model': 1.1748616299882997, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 1.19}) +[2026-06-24 00:24:50,946][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.041717298328876495, 'train/grad_norm_pre_clip': 21.115549087524414, 'timing/data': 0.0006904460024088621, 'timing/model': 1.1959527910221368, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 1.19}) +[2026-06-24 00:24:55,260][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.0022369129583239555, 'train/grad_norm_pre_clip': 11.531083106994629, 'timing/data': 0.0007191089971456677, 'timing/model': 1.1784191340266261, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 1.19}) +[2026-06-24 00:24:59,633][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.033580195158720016, 'train/grad_norm_pre_clip': 25.907939910888672, 'timing/data': 0.0007731390069238842, 'timing/model': 1.20037665500422, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 1.19}) +[2026-06-24 00:25:04,221][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.000972266832832247, 'train/grad_norm_pre_clip': 28.59853744506836, 'timing/data': 0.0008173399837687612, 'timing/model': 1.1797953169734683, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 1.19}) +[2026-06-24 00:25:08,586][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.005002574995160103, 'train/grad_norm_pre_clip': 15.964019775390625, 'timing/data': 0.0009777270024642348, 'timing/model': 1.1966117680130992, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 1.19}) +[2026-06-24 00:25:12,937][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.03221849724650383, 'train/grad_norm_pre_clip': 25.510265350341797, 'timing/data': 0.0008981100108940154, 'timing/model': 1.20278388200677, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 1.19}) +[2026-06-24 00:25:17,333][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.05505463853478432, 'train/grad_norm_pre_clip': 15.186254501342773, 'timing/data': 0.0009309400047641248, 'timing/model': 1.19812779899803, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 1.19}) +[2026-06-24 00:25:21,943][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.11717889457941055, 'train/grad_norm_pre_clip': 61.90165710449219, 'timing/data': 0.0008578049892093986, 'timing/model': 1.190088264993392, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 1.19}) +[2026-06-24 00:25:26,293][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.06338360160589218, 'train/grad_norm_pre_clip': 21.91900062561035, 'timing/data': 0.0008730429981369525, 'timing/model': 1.18105871102307, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 1.2}) +[2026-06-24 00:25:30,621][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.15988178551197052, 'train/grad_norm_pre_clip': 30.459455490112305, 'timing/data': 0.0007058040064293891, 'timing/model': 1.1761256459867582, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.2}) +[2026-06-24 00:25:34,971][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.05255896598100662, 'train/grad_norm_pre_clip': 21.515562057495117, 'timing/data': 0.0006689259898848832, 'timing/model': 1.1775918449857272, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 1.2}) +[2026-06-24 00:25:39,566][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.03696238994598389, 'train/grad_norm_pre_clip': 15.095795631408691, 'timing/data': 0.0006991919944994152, 'timing/model': 1.1761765999835916, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 1.2}) +[2026-06-24 00:25:43,875][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.004989751148968935, 'train/grad_norm_pre_clip': 5.587748050689697, 'timing/data': 0.0007692810031585395, 'timing/model': 1.17584963599802, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 1.2}) +[2026-06-24 00:25:48,221][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.009718981571495533, 'train/grad_norm_pre_clip': 28.100393295288086, 'timing/data': 0.0007645320147275925, 'timing/model': 1.1740545800130349, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 1.2}) +[2026-06-24 00:25:52,548][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.01771860010921955, 'train/grad_norm_pre_clip': 9.653454780578613, 'timing/data': 0.0009079670126084238, 'timing/model': 1.1738925499957986, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 1.2}) +[2026-06-24 00:25:57,173][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.019380420446395874, 'train/grad_norm_pre_clip': 20.673463821411133, 'timing/data': 0.000850510987220332, 'timing/model': 1.2034981009783223, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 1.2}) +[2026-06-24 00:26:01,518][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.14585645496845245, 'train/grad_norm_pre_clip': 20.258974075317383, 'timing/data': 0.0010639259999152273, 'timing/model': 1.1940811819804367, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 1.2}) +[2026-06-24 00:26:05,832][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.04132002592086792, 'train/grad_norm_pre_clip': 18.549652099609375, 'timing/data': 0.0008247140212915838, 'timing/model': 1.1756702039856464, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 1.2}) +[2026-06-24 00:26:10,171][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.07920681685209274, 'train/grad_norm_pre_clip': 12.398423194885254, 'timing/data': 0.0010504399833735079, 'timing/model': 1.1965660729911178, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 1.2}) +[2026-06-24 00:26:14,759][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.0036933794617652893, 'train/grad_norm_pre_clip': 12.913125991821289, 'timing/data': 0.0007595829956699163, 'timing/model': 1.2032273810182232, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 1.2}) +[2026-06-24 00:26:19,135][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3100 +[2026-06-24 00:27:16,614][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3100 +[2026-06-24 00:27:16,615][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3100 +[2026-06-24 00:27:53,780][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3100 +[2026-06-24 00:27:53,782][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.0026975006330758333, 'train/grad_norm_pre_clip': 21.46514320373535, 'eval/action_loss/seconds': 57.476808230014285, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.039178814738988876, 'eval/latency_2/loss': 0.039178814738988876, 'eval/flappy/loss': 0.039178814738988876, 'eval/flappy/latency_2/loss': 0.039178814738988876, 'eval/flap_precision': 0.9796747967479674, 'eval/flap_recall': 0.9810040705563093, 'eval/flap_f1': 0.9803389830508474, 'eval/flap_rate_model': 0.205, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9803389830508475, 'eval/latency_2/flap_precision': 0.9796747967479674, 'eval/latency_2/flap_recall': 0.9810040705563093, 'eval/latency_2/flap_f1': 0.9803389830508474, 'eval/latency_2/flap_rate_model': 0.205, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9803389830508475, 'eval/action_classification/data_wait_seconds': 0.47556188702583313, 'eval/action_classification/predict_seconds': 36.38422393798828, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0006951740069780499, 'timing/model': 1.1964755069930106, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 1.2}) +[2026-06-24 00:27:58,112][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.003525768406689167, 'train/grad_norm_pre_clip': 17.08720588684082, 'timing/data': 0.0008131519716698676, 'timing/model': 1.1957795399939641, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 1.2}) +[2026-06-24 00:28:02,626][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.047063156962394714, 'train/grad_norm_pre_clip': 9.692826271057129, 'timing/data': 0.0007711760117672384, 'timing/model': 1.189490388991544, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 1.2}) +[2026-06-24 00:28:06,950][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.031405724585056305, 'train/grad_norm_pre_clip': 14.657217025756836, 'timing/data': 0.0007269629859365523, 'timing/model': 1.179809622000903, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 1.2}) +[2026-06-24 00:28:11,261][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.08902374655008316, 'train/grad_norm_pre_clip': 30.2736873626709, 'timing/data': 0.0007248189940582961, 'timing/model': 1.196273903013207, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 1.2}) +[2026-06-24 00:28:15,863][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.1417427808046341, 'train/grad_norm_pre_clip': 59.3048095703125, 'timing/data': 0.000756758003262803, 'timing/model': 1.461733144999016, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 1.2}) +[2026-06-24 00:28:20,216][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.0018925830954685807, 'train/grad_norm_pre_clip': 49.24376678466797, 'timing/data': 0.0009188269905280322, 'timing/model': 1.1921723450068384, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 1.2}) +[2026-06-24 00:28:24,525][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.0025804745964705944, 'train/grad_norm_pre_clip': 11.319890975952148, 'timing/data': 0.0007808129885233939, 'timing/model': 1.171039428008953, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 1.2}) +[2026-06-24 00:28:28,831][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.019466081634163857, 'train/grad_norm_pre_clip': 36.74444580078125, 'timing/data': 0.000729147985111922, 'timing/model': 1.1707649599993601, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 1.2}) +[2026-06-24 00:28:33,376][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.0006235141190700233, 'train/grad_norm_pre_clip': 27.273143768310547, 'timing/data': 0.0006944530177861452, 'timing/model': 1.4159205180185381, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 1.2}) +[2026-06-24 00:28:37,672][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.1117570623755455, 'train/grad_norm_pre_clip': 40.54066848754883, 'timing/data': 0.0006893640093039721, 'timing/model': 1.16656359800254, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 1.2}) +[2026-06-24 00:28:41,993][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.016644755378365517, 'train/grad_norm_pre_clip': 12.626908302307129, 'timing/data': 0.0008311760029755533, 'timing/model': 1.1919673459779005, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 1.2}) +[2026-06-24 00:28:46,329][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.023437578231096268, 'train/grad_norm_pre_clip': 15.677133560180664, 'timing/data': 0.0008558919944334775, 'timing/model': 1.1986439320025966, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 1.2}) +[2026-06-24 00:28:50,795][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.008856598287820816, 'train/grad_norm_pre_clip': 29.396425247192383, 'timing/data': 0.0007449160038959235, 'timing/model': 1.1873408069950528, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 1.2}) +[2026-06-24 00:28:55,133][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.04667374864220619, 'train/grad_norm_pre_clip': 21.307891845703125, 'timing/data': 0.0007335659756790847, 'timing/model': 1.1930443150049541, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 1.21}) +[2026-06-24 00:28:59,487][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.2835586965084076, 'train/grad_norm_pre_clip': 43.280025482177734, 'timing/data': 0.0007421310001518577, 'timing/model': 1.191317254997557, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 1.21}) +[2026-06-24 00:29:03,823][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.053132206201553345, 'train/grad_norm_pre_clip': 33.868412017822266, 'timing/data': 0.0007491839933209121, 'timing/model': 1.186506777012255, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 1.21}) +[2026-06-24 00:29:08,459][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.016884606331586838, 'train/grad_norm_pre_clip': 11.71910285949707, 'timing/data': 0.0009477410058025271, 'timing/model': 1.2010070079995785, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 1.21}) +[2026-06-24 00:29:12,828][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.04216964542865753, 'train/grad_norm_pre_clip': 15.025444030761719, 'timing/data': 0.0007274939853232354, 'timing/model': 1.2053502929920796, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 1.21}) +[2026-06-24 00:29:17,172][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.05272170156240463, 'train/grad_norm_pre_clip': 24.00542640686035, 'timing/data': 0.0009163630020339042, 'timing/model': 1.2067031619953923, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 1.21}) +[2026-06-24 00:29:21,490][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.016301268711686134, 'train/grad_norm_pre_clip': 37.597930908203125, 'timing/data': 0.0007505670073442161, 'timing/model': 1.1797083049896173, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 1.21}) +[2026-06-24 00:29:26,010][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.030134417116642, 'train/grad_norm_pre_clip': 5.670576095581055, 'timing/data': 0.0007492850127164274, 'timing/model': 1.1818207379837986, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 1.21}) +[2026-06-24 00:29:30,392][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.028749113902449608, 'train/grad_norm_pre_clip': 14.549534797668457, 'timing/data': 0.0007089799910318106, 'timing/model': 1.1989065269881394, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 1.21}) +[2026-06-24 00:29:34,751][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.022660762071609497, 'train/grad_norm_pre_clip': 28.215211868286133, 'timing/data': 0.0011127559992019087, 'timing/model': 1.1787239050026983, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 1.21}) +[2026-06-24 00:29:39,091][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.036440346390008926, 'train/grad_norm_pre_clip': 18.962718963623047, 'timing/data': 0.0006945729837752879, 'timing/model': 1.1775843299983535, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 1.21}) +[2026-06-24 00:29:43,677][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.08604465425014496, 'train/grad_norm_pre_clip': 22.425373077392578, 'timing/data': 0.0008058490057010204, 'timing/model': 1.4067226120096166, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 1.21}) +[2026-06-24 00:29:48,010][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.019198058173060417, 'train/grad_norm_pre_clip': 3.9495902061462402, 'timing/data': 0.0007252400100696832, 'timing/model': 1.1774153980077244, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 1.21}) +[2026-06-24 00:29:52,587][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.02067410573363304, 'train/grad_norm_pre_clip': 24.108999252319336, 'timing/data': 0.000894641998456791, 'timing/model': 1.1886273539857939, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 1.21}) +[2026-06-24 00:29:56,963][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.03069472312927246, 'train/grad_norm_pre_clip': 25.842660903930664, 'timing/data': 0.0009743499977048486, 'timing/model': 1.1779234369751066, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 1.21}) +[2026-06-24 00:30:01,322][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.024018632248044014, 'train/grad_norm_pre_clip': 14.883807182312012, 'timing/data': 0.0009967410005629063, 'timing/model': 1.1926247899828013, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 1.21}) +[2026-06-24 00:30:05,960][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.0010872200364246964, 'train/grad_norm_pre_clip': 24.009294509887695, 'timing/data': 0.0007344069890677929, 'timing/model': 1.1798799819953274, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 1.21}) +[2026-06-24 00:30:10,281][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.0564340278506279, 'train/grad_norm_pre_clip': 25.718116760253906, 'timing/data': 0.0009565679938532412, 'timing/model': 1.182738503004657, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 1.21}) +[2026-06-24 00:30:14,646][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.0021174002904444933, 'train/grad_norm_pre_clip': 4.468417167663574, 'timing/data': 0.0007513980090152472, 'timing/model': 1.198316968017025, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 1.21}) +[2026-06-24 00:30:19,183][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.0027944250032305717, 'train/grad_norm_pre_clip': 21.452924728393555, 'timing/data': 0.0011167539923917502, 'timing/model': 1.4031287830148358, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 1.21}) +[2026-06-24 00:30:23,559][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.0004497448680922389, 'train/grad_norm_pre_clip': 24.259138107299805, 'timing/data': 0.0007160629902500659, 'timing/model': 1.1988558120210655, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 1.21}) +[2026-06-24 00:30:27,892][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.005370986647903919, 'train/grad_norm_pre_clip': 27.346223831176758, 'timing/data': 0.0008273889834526926, 'timing/model': 1.1974157220101915, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 1.21}) +[2026-06-24 00:30:32,242][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.003924654331058264, 'train/grad_norm_pre_clip': 28.194339752197266, 'timing/data': 0.0010485670063644648, 'timing/model': 1.181223413994303, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 1.21}) +[2026-06-24 00:30:36,843][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.07996802031993866, 'train/grad_norm_pre_clip': 30.278308868408203, 'timing/data': 0.0008190230000764132, 'timing/model': 1.4367590300098527, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 1.21}) +[2026-06-24 00:30:41,179][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.19582322239875793, 'train/grad_norm_pre_clip': 60.603363037109375, 'timing/data': 0.0007814540003892034, 'timing/model': 1.199942931998521, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.21}) +[2026-06-24 00:30:45,584][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.10717849433422089, 'train/grad_norm_pre_clip': 32.20999526977539, 'timing/data': 0.0008125310123432428, 'timing/model': 1.2086102559987921, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 1.21}) +[2026-06-24 00:30:49,919][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.023253358900547028, 'train/grad_norm_pre_clip': 15.581958770751953, 'timing/data': 0.0007780469895806164, 'timing/model': 1.1951601339969784, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 1.22}) +[2026-06-24 00:30:54,505][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.06974854320287704, 'train/grad_norm_pre_clip': 16.857593536376953, 'timing/data': 0.0009323930134996772, 'timing/model': 1.4075241039972752, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 1.22}) +[2026-06-24 00:30:58,833][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.018501099199056625, 'train/grad_norm_pre_clip': 20.964555740356445, 'timing/data': 0.0007263619918376207, 'timing/model': 1.1767491069913376, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 1.22}) +[2026-06-24 00:31:03,251][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.10674656927585602, 'train/grad_norm_pre_clip': 31.924089431762695, 'timing/data': 0.0008159370045177639, 'timing/model': 1.2039047010184731, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 1.22}) +[2026-06-24 00:31:07,584][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.020341617986559868, 'train/grad_norm_pre_clip': 7.016384601593018, 'timing/data': 0.001064395997673273, 'timing/model': 1.1750979560019914, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 1.22}) +[2026-06-24 00:31:12,164][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.18806764483451843, 'train/grad_norm_pre_clip': 31.1175594329834, 'timing/data': 0.0007580409874208272, 'timing/model': 1.4294858610082883, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 1.22}) +[2026-06-24 00:31:16,499][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.006996730342507362, 'train/grad_norm_pre_clip': 20.586244583129883, 'timing/data': 0.0007128369761630893, 'timing/model': 1.1713178219797555, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 1.22}) +[2026-06-24 00:31:20,866][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.0376117043197155, 'train/grad_norm_pre_clip': 17.105791091918945, 'timing/data': 0.000870197982294485, 'timing/model': 1.1975348519918043, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 1.22}) +[2026-06-24 00:31:25,205][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.03576115891337395, 'train/grad_norm_pre_clip': 13.299372673034668, 'timing/data': 0.0006958259909879416, 'timing/model': 1.1857509189867415, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 1.22}) +[2026-06-24 00:31:29,753][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.0031325609888881445, 'train/grad_norm_pre_clip': 25.574541091918945, 'timing/data': 0.0007464399968739599, 'timing/model': 1.2037672069855034, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 1.22}) +[2026-06-24 00:31:34,068][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.006929201539605856, 'train/grad_norm_pre_clip': 11.983891487121582, 'timing/data': 0.0007196090009529144, 'timing/model': 1.1716318609833252, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 1.22}) +[2026-06-24 00:31:38,373][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.014551020227372646, 'train/grad_norm_pre_clip': 21.066316604614258, 'timing/data': 0.0008592680096626282, 'timing/model': 1.1703647100075614, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 1.22}) +[2026-06-24 00:31:42,696][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.012105852365493774, 'train/grad_norm_pre_clip': 9.076358795166016, 'timing/data': 0.0007220740080811083, 'timing/model': 1.1725943199999165, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 1.22}) +[2026-06-24 00:31:47,354][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.008126775734126568, 'train/grad_norm_pre_clip': 14.760727882385254, 'timing/data': 0.0009193090081680566, 'timing/model': 1.213656768988585, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 1.22}) +[2026-06-24 00:31:51,739][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.06545118242502213, 'train/grad_norm_pre_clip': 29.783884048461914, 'timing/data': 0.001043689000653103, 'timing/model': 1.1802664359856863, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 1.22}) +[2026-06-24 00:31:56,137][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.0032601237762719393, 'train/grad_norm_pre_clip': 15.379146575927734, 'timing/data': 0.000882530992385, 'timing/model': 1.2008811249979772, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 1.22}) +[2026-06-24 00:32:00,523][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.027462543919682503, 'train/grad_norm_pre_clip': 17.154478073120117, 'timing/data': 0.00101645800168626, 'timing/model': 1.1800836470210925, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 1.22}) +[2026-06-24 00:32:05,101][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.021207697689533234, 'train/grad_norm_pre_clip': 30.427139282226562, 'timing/data': 0.0007306899933610111, 'timing/model': 1.1844004440063145, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 1.22}) +[2026-06-24 00:32:09,420][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.0010802170727401972, 'train/grad_norm_pre_clip': 12.103201866149902, 'timing/data': 0.0007240679988171905, 'timing/model': 1.179084862000309, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 1.22}) +[2026-06-24 00:32:13,765][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.0034090091940015554, 'train/grad_norm_pre_clip': 32.86270523071289, 'timing/data': 0.0008428579894825816, 'timing/model': 1.2109567610023078, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 1.22}) +[2026-06-24 00:32:18,162][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.07937361299991608, 'train/grad_norm_pre_clip': 28.12924575805664, 'timing/data': 0.0007711249927524477, 'timing/model': 1.209343380003702, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 1.22}) +[2026-06-24 00:32:22,685][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.00024888216285035014, 'train/grad_norm_pre_clip': 27.428003311157227, 'timing/data': 0.0007521699881181121, 'timing/model': 1.1992263170250226, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 1.22}) +[2026-06-24 00:32:27,057][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.11166836321353912, 'train/grad_norm_pre_clip': 21.92215347290039, 'timing/data': 0.0011326829844620079, 'timing/model': 1.2034034350072034, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 1.22}) +[2026-06-24 00:32:31,401][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.023956410586833954, 'train/grad_norm_pre_clip': 19.904752731323242, 'timing/data': 0.0007395469874609262, 'timing/model': 1.2039386539836414, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 1.22}) +[2026-06-24 00:32:35,780][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.018683070316910744, 'train/grad_norm_pre_clip': 13.931319236755371, 'timing/data': 0.000743835000321269, 'timing/model': 1.1830155949865002, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 1.22}) +[2026-06-24 00:32:40,378][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.027689877897500992, 'train/grad_norm_pre_clip': 16.390975952148438, 'timing/data': 0.0007504370005335659, 'timing/model': 1.1999326219956856, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 1.22}) +[2026-06-24 00:32:44,758][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.02409248612821102, 'train/grad_norm_pre_clip': 29.222394943237305, 'timing/data': 0.0010360940068494529, 'timing/model': 1.2029763649916276, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 1.23}) +[2026-06-24 00:32:49,120][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.053326189517974854, 'train/grad_norm_pre_clip': 15.494111061096191, 'timing/data': 0.0007655739900656044, 'timing/model': 1.200773766002385, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 1.23}) +[2026-06-24 00:32:53,697][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.053087446838617325, 'train/grad_norm_pre_clip': 21.5554141998291, 'timing/data': 0.0007597740041092038, 'timing/model': 1.3967068569909316, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 1.23}) +[2026-06-24 00:32:58,090][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.02513897605240345, 'train/grad_norm_pre_clip': 14.628934860229492, 'timing/data': 0.0008233610133174807, 'timing/model': 1.2064508959883824, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 1.23}) +[2026-06-24 00:33:02,462][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.024852707982063293, 'train/grad_norm_pre_clip': 11.504500389099121, 'timing/data': 0.0008804370008874685, 'timing/model': 1.2038072109862696, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 1.23}) +[2026-06-24 00:33:06,835][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.020179979503154755, 'train/grad_norm_pre_clip': 11.031659126281738, 'timing/data': 0.0007347670034505427, 'timing/model': 1.1808939850016031, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 1.23}) +[2026-06-24 00:33:11,469][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.03836728259921074, 'train/grad_norm_pre_clip': 8.798479080200195, 'timing/data': 0.0007698319968767464, 'timing/model': 1.4528057879942935, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 1.23}) +[2026-06-24 00:33:15,820][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.13095176219940186, 'train/grad_norm_pre_clip': 21.367183685302734, 'timing/data': 0.000932152004679665, 'timing/model': 1.1811671900213696, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 1.23}) +[2026-06-24 00:33:20,210][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.027510706335306168, 'train/grad_norm_pre_clip': 21.81736183166504, 'timing/data': 0.0008143850136548281, 'timing/model': 1.2000187810044736, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 1.23}) +[2026-06-24 00:33:24,633][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.1366535723209381, 'train/grad_norm_pre_clip': 20.61780548095703, 'timing/data': 0.0010688350012060255, 'timing/model': 1.2210049470013473, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 1.23}) +[2026-06-24 00:33:29,253][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.0037916575092822313, 'train/grad_norm_pre_clip': 14.599287986755371, 'timing/data': 0.0008028640004340559, 'timing/model': 1.4548461790254805, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 1.23}) +[2026-06-24 00:33:33,601][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.00012793247879017144, 'train/grad_norm_pre_clip': 34.36430358886719, 'timing/data': 0.0007314909889828414, 'timing/model': 1.2023322059831116, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 1.23}) +[2026-06-24 00:33:37,993][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.0331537164747715, 'train/grad_norm_pre_clip': 24.644771575927734, 'timing/data': 0.0007466500101145357, 'timing/model': 1.1977032430004328, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 1.23}) +[2026-06-24 00:33:42,349][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.0031872671097517014, 'train/grad_norm_pre_clip': 7.92644739151001, 'timing/data': 0.0007657549867872149, 'timing/model': 1.1971260769932996, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 1.23}) +[2026-06-24 00:33:46,913][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.0004255248641129583, 'train/grad_norm_pre_clip': 7.580263614654541, 'timing/data': 0.0007473209989257157, 'timing/model': 1.2099578749912325, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 1.23}) +[2026-06-24 00:33:51,295][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.031375668942928314, 'train/grad_norm_pre_clip': 28.58487892150879, 'timing/data': 0.0007188880117610097, 'timing/model': 1.1971902749792207, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 1.23}) +[2026-06-24 00:33:55,684][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.01558440551161766, 'train/grad_norm_pre_clip': 13.035265922546387, 'timing/data': 0.0007746710034552962, 'timing/model': 1.2012816749920603, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 1.23}) +[2026-06-24 00:34:00,053][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.0021435508970171213, 'train/grad_norm_pre_clip': 15.928486824035645, 'timing/data': 0.0007258309924509376, 'timing/model': 1.1985872259829193, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 1.23}) +[2026-06-24 00:34:04,718][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.44188427925109863, 'train/grad_norm_pre_clip': 21.69431495666504, 'timing/data': 0.0011697020090650767, 'timing/model': 1.472491862019524, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 1.23}) +[2026-06-24 00:34:09,082][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.005391260143369436, 'train/grad_norm_pre_clip': 11.445794105529785, 'timing/data': 0.000740838993806392, 'timing/model': 1.180883855005959, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 1.23}) +[2026-06-24 00:34:13,457][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.06781432777643204, 'train/grad_norm_pre_clip': 24.987714767456055, 'timing/data': 0.0008732230053283274, 'timing/model': 1.1817162740044296, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 1.23}) +[2026-06-24 00:34:17,821][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.03496432676911354, 'train/grad_norm_pre_clip': 17.925874710083008, 'timing/data': 0.0007199409883469343, 'timing/model': 1.1800398259947542, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.23}) +[2026-06-24 00:34:22,455][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.017031880095601082, 'train/grad_norm_pre_clip': 17.87512969970703, 'timing/data': 0.0007286769978236407, 'timing/model': 1.466885503003141, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 1.23}) +[2026-06-24 00:34:26,836][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.07454938441514969, 'train/grad_norm_pre_clip': 36.189884185791016, 'timing/data': 0.0007509280112572014, 'timing/model': 1.180784232012229, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 1.23}) +[2026-06-24 00:34:31,205][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.0021619065664708614, 'train/grad_norm_pre_clip': 16.872621536254883, 'timing/data': 0.000756196997826919, 'timing/model': 1.2016631189908367, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 1.23}) +[2026-06-24 00:34:35,564][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.00040746593731455505, 'train/grad_norm_pre_clip': 17.273406982421875, 'timing/data': 0.0007201200060080737, 'timing/model': 1.1977854239812586, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 1.23}) +[2026-06-24 00:34:40,226][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.10383985191583633, 'train/grad_norm_pre_clip': 13.361467361450195, 'timing/data': 0.0007419210160151124, 'timing/model': 1.4684293850150425, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 1.24}) +[2026-06-24 00:34:44,605][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.002160074654966593, 'train/grad_norm_pre_clip': 14.312870025634766, 'timing/data': 0.0008672729891259223, 'timing/model': 1.1815652250079438, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 1.24}) +[2026-06-24 00:34:48,978][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.1191168874502182, 'train/grad_norm_pre_clip': 30.954376220703125, 'timing/data': 0.0007644020079169422, 'timing/model': 1.2054150820185896, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 1.24}) +[2026-06-24 00:34:53,327][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.044020432978868484, 'train/grad_norm_pre_clip': 9.517159461975098, 'timing/data': 0.0007268729968927801, 'timing/model': 1.1831813209864777, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 1.24}) +[2026-06-24 00:34:57,872][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.0018674066523090005, 'train/grad_norm_pre_clip': 9.188467025756836, 'timing/data': 0.0007452970021404326, 'timing/model': 1.2036442499957047, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 1.24}) +[2026-06-24 00:35:02,250][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.026118967682123184, 'train/grad_norm_pre_clip': 9.714964866638184, 'timing/data': 0.0008453420014120638, 'timing/model': 1.2161880060157273, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 1.24}) +[2026-06-24 00:35:06,622][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.02748466283082962, 'train/grad_norm_pre_clip': 7.028343677520752, 'timing/data': 0.0008157870033755898, 'timing/model': 1.2019004269968718, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 1.24}) +[2026-06-24 00:35:11,018][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.006498648785054684, 'train/grad_norm_pre_clip': 13.640853881835938, 'timing/data': 0.0011758430046029389, 'timing/model': 1.2030866899876855, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 1.24}) +[2026-06-24 00:35:15,619][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3200 +[2026-06-24 00:36:13,134][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3200 +[2026-06-24 00:36:13,134][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3200 +[2026-06-24 00:36:50,001][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3200 +[2026-06-24 00:36:50,003][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.06714343279600143, 'train/grad_norm_pre_clip': 11.40026569366455, 'eval/action_loss/seconds': 57.51165027698153, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03627384826540947, 'eval/latency_2/loss': 0.03627384826540947, 'eval/flappy/loss': 0.03627384826540947, 'eval/flappy/latency_2/loss': 0.03627384826540947, 'eval/flap_precision': 0.9718875502008032, 'eval/flap_recall': 0.9850746268656716, 'eval/flap_f1': 0.9784366576819408, 'eval/flap_rate_model': 0.2075, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9784366576819407, 'eval/latency_2/flap_precision': 0.9718875502008032, 'eval/latency_2/flap_recall': 0.9850746268656716, 'eval/latency_2/flap_f1': 0.9784366576819408, 'eval/latency_2/flap_rate_model': 0.2075, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9784366576819407, 'eval/action_classification/data_wait_seconds': 0.4564359784126282, 'eval/action_classification/predict_seconds': 36.263729095458984, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007948790153022856, 'timing/model': 1.4248777469911147, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 1.24}) +[2026-06-24 00:36:54,328][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.013647889718413353, 'train/grad_norm_pre_clip': 27.93599510192871, 'timing/data': 0.0009191879944410175, 'timing/model': 1.1636464569892269, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 1.24}) +[2026-06-24 00:36:58,619][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.2069806009531021, 'train/grad_norm_pre_clip': 19.3216495513916, 'timing/data': 0.0009680189832579345, 'timing/model': 1.1681155140104238, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 1.24}) +[2026-06-24 00:37:02,926][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.01598864048719406, 'train/grad_norm_pre_clip': 20.01154327392578, 'timing/data': 0.0007550750160589814, 'timing/model': 1.1708178269909695, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 1.24}) +[2026-06-24 00:37:07,479][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.0015266903210431337, 'train/grad_norm_pre_clip': 17.681610107421875, 'timing/data': 0.0007525399851147085, 'timing/model': 1.1916714070248418, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 1.24}) +[2026-06-24 00:37:11,798][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.14912515878677368, 'train/grad_norm_pre_clip': 28.029937744140625, 'timing/data': 0.0008069309988059103, 'timing/model': 1.1904429380083457, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 1.24}) +[2026-06-24 00:37:16,132][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.01991375908255577, 'train/grad_norm_pre_clip': 18.028738021850586, 'timing/data': 0.0007726370240561664, 'timing/model': 1.1952373359818012, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 1.24}) +[2026-06-24 00:37:20,474][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.10903898626565933, 'train/grad_norm_pre_clip': 26.002199172973633, 'timing/data': 0.0007512880256399512, 'timing/model': 1.1717893210006878, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 1.24}) +[2026-06-24 00:37:25,077][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.018434951081871986, 'train/grad_norm_pre_clip': 11.800413131713867, 'timing/data': 0.0007327240018639714, 'timing/model': 1.1882519589853473, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 1.24}) +[2026-06-24 00:37:29,390][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.006752203218638897, 'train/grad_norm_pre_clip': 6.157533168792725, 'timing/data': 0.0008177509880624712, 'timing/model': 1.1745923910057172, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 1.24}) +[2026-06-24 00:37:33,718][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.018767042085528374, 'train/grad_norm_pre_clip': 29.444149017333984, 'timing/data': 0.0012197739852126688, 'timing/model': 1.1756136280018836, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 1.24}) +[2026-06-24 00:37:38,030][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.054664529860019684, 'train/grad_norm_pre_clip': 13.551910400390625, 'timing/data': 0.0007576100178994238, 'timing/model': 1.175985945010325, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 1.24}) +[2026-06-24 00:37:42,620][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.0036502662114799023, 'train/grad_norm_pre_clip': 20.036157608032227, 'timing/data': 0.0010752870002761483, 'timing/model': 1.1734068999940064, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 1.24}) +[2026-06-24 00:37:46,968][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.04788586497306824, 'train/grad_norm_pre_clip': 13.259215354919434, 'timing/data': 0.0009454760001972318, 'timing/model': 1.1853843209974002, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 1.24}) +[2026-06-24 00:37:51,301][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.05767523869872093, 'train/grad_norm_pre_clip': 26.63585662841797, 'timing/data': 0.0007772249809931964, 'timing/model': 1.180635465978412, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 1.24}) +[2026-06-24 00:37:55,823][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.15365993976593018, 'train/grad_norm_pre_clip': 10.010062217712402, 'timing/data': 0.000997062015812844, 'timing/model': 1.3768727690039668, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 1.24}) +[2026-06-24 00:38:00,175][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.08923830837011337, 'train/grad_norm_pre_clip': 26.11518096923828, 'timing/data': 0.0011751710262615234, 'timing/model': 1.1792189389816485, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 1.24}) +[2026-06-24 00:38:04,558][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.016306763514876366, 'train/grad_norm_pre_clip': 21.7042293548584, 'timing/data': 0.0007572790200356394, 'timing/model': 1.1973809570190497, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 1.25}) +[2026-06-24 00:38:08,871][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.023339591920375824, 'train/grad_norm_pre_clip': 19.646835327148438, 'timing/data': 0.0010173000046052039, 'timing/model': 1.176158144022338, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 1.25}) +[2026-06-24 00:38:13,414][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.0228022038936615, 'train/grad_norm_pre_clip': 9.821989059448242, 'timing/data': 0.0007198699750006199, 'timing/model': 1.3691360439988784, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 1.25}) +[2026-06-24 00:38:17,771][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.014013507403433323, 'train/grad_norm_pre_clip': 8.71529483795166, 'timing/data': 0.0007590320019517094, 'timing/model': 1.171923469984904, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 1.25}) +[2026-06-24 00:38:22,070][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.0010977713391184807, 'train/grad_norm_pre_clip': 8.849652290344238, 'timing/data': 0.0007800119929015636, 'timing/model': 1.1725010270019993, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 1.25}) +[2026-06-24 00:38:26,405][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.004762967582792044, 'train/grad_norm_pre_clip': 8.506129264831543, 'timing/data': 0.0009093299740925431, 'timing/model': 1.176658096985193, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 1.25}) +[2026-06-24 00:38:31,005][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.020546797662973404, 'train/grad_norm_pre_clip': 8.621234893798828, 'timing/data': 0.0007518799975514412, 'timing/model': 1.4409775730164256, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 1.25}) +[2026-06-24 00:38:35,381][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.262230783700943, 'train/grad_norm_pre_clip': 23.88031578063965, 'timing/data': 0.0009168840188067406, 'timing/model': 1.2014876939938404, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 1.25}) +[2026-06-24 00:38:39,738][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.014078452251851559, 'train/grad_norm_pre_clip': 13.410937309265137, 'timing/data': 0.0008642169996164739, 'timing/model': 1.1779984240129124, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 1.25}) +[2026-06-24 00:38:44,072][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.014731780625879765, 'train/grad_norm_pre_clip': 41.16429138183594, 'timing/data': 0.0008144449966493994, 'timing/model': 1.1791212589887436, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 1.25}) +[2026-06-24 00:38:48,580][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.03940116986632347, 'train/grad_norm_pre_clip': 19.445838928222656, 'timing/data': 0.0009306300198659301, 'timing/model': 1.1776012199989054, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 1.25}) +[2026-06-24 00:38:52,949][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.001993864309042692, 'train/grad_norm_pre_clip': 35.89453887939453, 'timing/data': 0.0007391560066025704, 'timing/model': 1.2075703020091169, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 1.25}) +[2026-06-24 00:38:57,284][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.0497596375644207, 'train/grad_norm_pre_clip': 20.008026123046875, 'timing/data': 0.0008287309901788831, 'timing/model': 1.1782348999986425, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 1.25}) +[2026-06-24 00:39:01,619][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.002142411656677723, 'train/grad_norm_pre_clip': 15.913602828979492, 'timing/data': 0.0009274230105802417, 'timing/model': 1.1783395639795344, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 1.25}) +[2026-06-24 00:39:06,132][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.0032010844442993402, 'train/grad_norm_pre_clip': 7.606532096862793, 'timing/data': 0.0007785679772496223, 'timing/model': 1.1969752570148557, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 1.25}) +[2026-06-24 00:39:10,466][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.02567887306213379, 'train/grad_norm_pre_clip': 21.56012725830078, 'timing/data': 0.0007417910092044622, 'timing/model': 1.178382854006486, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 1.25}) +[2026-06-24 00:39:14,826][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.26589345932006836, 'train/grad_norm_pre_clip': 25.451187133789062, 'timing/data': 0.0010281700233463198, 'timing/model': 1.1773325240064878, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 1.25}) +[2026-06-24 00:39:19,192][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.12008427828550339, 'train/grad_norm_pre_clip': 20.007823944091797, 'timing/data': 0.0009194590093102306, 'timing/model': 1.1987683789920993, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 1.25}) +[2026-06-24 00:39:23,793][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.05312192440032959, 'train/grad_norm_pre_clip': 29.295337677001953, 'timing/data': 0.001121792010962963, 'timing/model': 1.196464183012722, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 1.25}) +[2026-06-24 00:39:28,128][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.02677324041724205, 'train/grad_norm_pre_clip': 16.460416793823242, 'timing/data': 0.0007639320101588964, 'timing/model': 1.1861577499948908, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.25}) +[2026-06-24 00:39:32,493][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.06397101283073425, 'train/grad_norm_pre_clip': 35.3563346862793, 'timing/data': 0.0007951289881020784, 'timing/model': 1.201675962016452, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 1.25}) +[2026-06-24 00:39:36,838][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.0004523549578152597, 'train/grad_norm_pre_clip': 19.357297897338867, 'timing/data': 0.0009345459984615445, 'timing/model': 1.1803490570164286, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 1.25}) +[2026-06-24 00:39:41,448][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.04337408393621445, 'train/grad_norm_pre_clip': 17.017698287963867, 'timing/data': 0.0007068359991535544, 'timing/model': 1.195820654014824, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 1.25}) +[2026-06-24 00:39:45,843][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.26790568232536316, 'train/grad_norm_pre_clip': 30.405014038085938, 'timing/data': 0.0007551950111519545, 'timing/model': 1.1895738010061905, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 1.25}) +[2026-06-24 00:39:50,210][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.011355908587574959, 'train/grad_norm_pre_clip': 14.530922889709473, 'timing/data': 0.0011942259734496474, 'timing/model': 1.1994124299963005, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 1.25}) +[2026-06-24 00:39:54,574][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.02515164576470852, 'train/grad_norm_pre_clip': 30.968976974487305, 'timing/data': 0.0008017809886951, 'timing/model': 1.1822448199964128, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 1.25}) +[2026-06-24 00:39:59,119][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.025581035763025284, 'train/grad_norm_pre_clip': 32.02925109863281, 'timing/data': 0.0007674579974263906, 'timing/model': 1.1785543709993362, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 1.26}) +[2026-06-24 00:40:03,496][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.08310152590274811, 'train/grad_norm_pre_clip': 20.372838973999023, 'timing/data': 0.0008762490178924054, 'timing/model': 1.214533789025154, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 1.26}) +[2026-06-24 00:40:07,853][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.10005789995193481, 'train/grad_norm_pre_clip': 21.856239318847656, 'timing/data': 0.0007460589986294508, 'timing/model': 1.2016806610045023, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 1.26}) +[2026-06-24 00:40:12,226][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.020812567323446274, 'train/grad_norm_pre_clip': 10.475245475769043, 'timing/data': 0.0008917079831007868, 'timing/model': 1.2061268760007806, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 1.26}) +[2026-06-24 00:40:16,837][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.044665444642305374, 'train/grad_norm_pre_clip': 25.44791603088379, 'timing/data': 0.0007988959841895849, 'timing/model': 1.1871743879746646, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 1.26}) +[2026-06-24 00:40:21,194][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.11319718509912491, 'train/grad_norm_pre_clip': 41.81232452392578, 'timing/data': 0.0009723870025482029, 'timing/model': 1.1811536939931102, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 1.26}) +[2026-06-24 00:40:25,579][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.03301234170794487, 'train/grad_norm_pre_clip': 35.51349639892578, 'timing/data': 0.0007554359908681363, 'timing/model': 1.207340369001031, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 1.26}) +[2026-06-24 01:08:44,159][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3250 +[2026-06-24 01:08:44,160][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.15121158957481384, 'train/grad_norm_pre_clip': 20.849185943603516, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0007422909839078784, 'timing/model': 1.183469473005971, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 1.26}) +[2026-06-24 01:08:48,394][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.021249467507004738, 'train/grad_norm_pre_clip': 25.7614803314209, 'timing/data': 0.0011488529853522778, 'timing/model': 1.1534524619928561, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 1.26}) +[2026-06-24 01:08:52,605][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.0006244035321287811, 'train/grad_norm_pre_clip': 9.700746536254883, 'timing/data': 0.0011783270165324211, 'timing/model': 1.1455518529983237, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 1.26}) +[2026-06-24 01:08:56,814][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.0016982279485091567, 'train/grad_norm_pre_clip': 16.25862693786621, 'timing/data': 0.001029481994919479, 'timing/model': 1.164270399021916, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 1.26}) +[2026-06-24 01:09:01,227][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.09810391813516617, 'train/grad_norm_pre_clip': 9.976780891418457, 'timing/data': 0.0009645620011724532, 'timing/model': 1.1531050199992023, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 1.26}) +[2026-06-24 01:09:05,389][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.08071327209472656, 'train/grad_norm_pre_clip': 21.5866641998291, 'timing/data': 0.0007158929947763681, 'timing/model': 1.1410129469877575, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 1.26}) +[2026-06-24 01:09:09,559][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.024636385962367058, 'train/grad_norm_pre_clip': 18.529701232910156, 'timing/data': 0.0006941629981156439, 'timing/model': 1.141101841989439, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 1.26}) +[2026-06-24 01:09:13,756][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.09719887375831604, 'train/grad_norm_pre_clip': 35.01066970825195, 'timing/data': 0.0007016559829935431, 'timing/model': 1.1632556249969639, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 1.26}) +[2026-06-24 01:09:18,193][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.045411739498376846, 'train/grad_norm_pre_clip': 11.148581504821777, 'timing/data': 0.0007672270003240556, 'timing/model': 1.1476226590166334, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 1.26}) +[2026-06-24 01:09:22,436][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.07181352376937866, 'train/grad_norm_pre_clip': 9.108566284179688, 'timing/data': 0.0011031980102416128, 'timing/model': 1.1493768699874636, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 1.26}) +[2026-06-24 01:09:26,708][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.06380798667669296, 'train/grad_norm_pre_clip': 20.98131561279297, 'timing/data': 0.0008637059945613146, 'timing/model': 1.18046808699728, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 1.26}) +[2026-06-24 01:09:30,943][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.07267681509256363, 'train/grad_norm_pre_clip': 27.74504280090332, 'timing/data': 0.0007292680093087256, 'timing/model': 1.1487168409803417, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 1.26}) +[2026-06-24 01:09:35,376][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.14855661988258362, 'train/grad_norm_pre_clip': 29.160137176513672, 'timing/data': 0.0007441950147040188, 'timing/model': 1.1833315789990593, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 1.26}) +[2026-06-24 01:09:39,641][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.009333834052085876, 'train/grad_norm_pre_clip': 19.596824645996094, 'timing/data': 0.0009419500129297376, 'timing/model': 1.1758133669791277, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 1.26}) +[2026-06-24 01:09:43,921][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.002389272442087531, 'train/grad_norm_pre_clip': 19.76752281188965, 'timing/data': 0.0007552749884780496, 'timing/model': 1.179734041012125, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 1.26}) +[2026-06-24 01:09:48,206][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.020128034055233, 'train/grad_norm_pre_clip': 28.382165908813477, 'timing/data': 0.0008910559990908951, 'timing/model': 1.1857006159843877, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 1.26}) +[2026-06-24 01:09:52,698][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.059555038809776306, 'train/grad_norm_pre_clip': 15.259676933288574, 'timing/data': 0.0007999379886314273, 'timing/model': 1.1546113319927827, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 1.26}) +[2026-06-24 01:09:56,972][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.027986466884613037, 'train/grad_norm_pre_clip': 27.82759666442871, 'timing/data': 0.0007268719782587141, 'timing/model': 1.1716769839986227, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 1.26}) +[2026-06-24 01:10:01,261][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.13870033621788025, 'train/grad_norm_pre_clip': 35.527339935302734, 'timing/data': 0.0007703739975113422, 'timing/model': 1.1783762919949368, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 1.26}) +[2026-06-24 01:10:05,586][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.17270483076572418, 'train/grad_norm_pre_clip': 22.56192398071289, 'timing/data': 0.0012179310142528266, 'timing/model': 1.1868745849933475, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 1.27}) +[2026-06-24 01:10:10,101][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.004419003613293171, 'train/grad_norm_pre_clip': 13.113420486450195, 'timing/data': 0.000758291018428281, 'timing/model': 1.192369307973422, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 1.27}) +[2026-06-24 01:10:14,412][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.0541461817920208, 'train/grad_norm_pre_clip': 33.114013671875, 'timing/data': 0.0008864690025802702, 'timing/model': 1.1876384149945807, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 1.27}) +[2026-06-24 01:10:18,744][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.00584467314183712, 'train/grad_norm_pre_clip': 18.781286239624023, 'timing/data': 0.0007459679909516126, 'timing/model': 1.186938131984789, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 1.27}) +[2026-06-24 01:10:23,021][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.005169643554836512, 'train/grad_norm_pre_clip': 12.241260528564453, 'timing/data': 0.000774210988311097, 'timing/model': 1.1688495900016278, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 1.27}) +[2026-06-24 01:10:27,521][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.002043074229732156, 'train/grad_norm_pre_clip': 29.56244659423828, 'timing/data': 0.0007535619952250272, 'timing/model': 1.1924365010054316, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 1.27}) +[2026-06-24 01:10:31,856][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.05845377594232559, 'train/grad_norm_pre_clip': 21.286165237426758, 'timing/data': 0.0011173250095453113, 'timing/model': 1.1974408980167937, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 1.27}) +[2026-06-24 01:10:36,201][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.0017912500770762563, 'train/grad_norm_pre_clip': 24.758270263671875, 'timing/data': 0.0007705329917371273, 'timing/model': 1.1890266999835148, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 1.27}) +[2026-06-24 01:10:40,531][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.12183327972888947, 'train/grad_norm_pre_clip': 24.352354049682617, 'timing/data': 0.0007875650189816952, 'timing/model': 1.1919163700076751, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 1.27}) +[2026-06-24 01:10:45,004][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.002930735470727086, 'train/grad_norm_pre_clip': 9.11931037902832, 'timing/data': 0.0007092699815984815, 'timing/model': 1.1718384019914083, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 1.27}) +[2026-06-24 01:10:49,301][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.002966518048197031, 'train/grad_norm_pre_clip': 11.535065650939941, 'timing/data': 0.0006971580151002854, 'timing/model': 1.1684158080024645, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 1.27}) +[2026-06-24 01:10:53,611][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.004255876410752535, 'train/grad_norm_pre_clip': 23.561246871948242, 'timing/data': 0.0007178860250860453, 'timing/model': 1.1988808790047187, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 1.27}) +[2026-06-24 01:10:57,911][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.07326416671276093, 'train/grad_norm_pre_clip': 13.170969009399414, 'timing/data': 0.0007726979965809733, 'timing/model': 1.1780547669914085, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 1.27}) +[2026-06-24 01:11:02,438][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.017240319401025772, 'train/grad_norm_pre_clip': 33.07052993774414, 'timing/data': 0.000906784989638254, 'timing/model': 1.1913341940089595, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 1.27}) +[2026-06-24 01:11:06,762][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.043702900409698486, 'train/grad_norm_pre_clip': 20.55927848815918, 'timing/data': 0.0010187030129600316, 'timing/model': 1.1757974170031957, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 1.27}) +[2026-06-24 01:11:11,116][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.0020279858727008104, 'train/grad_norm_pre_clip': 14.827926635742188, 'timing/data': 0.0011845290136989206, 'timing/model': 1.2013932300033048, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 1.27}) +[2026-06-24 01:11:15,439][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.010697000660002232, 'train/grad_norm_pre_clip': 10.391907691955566, 'timing/data': 0.0007531019800808281, 'timing/model': 1.1752803309937008, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.27}) +[2026-06-24 01:11:19,927][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.058707959949970245, 'train/grad_norm_pre_clip': 15.771135330200195, 'timing/data': 0.0007399969908874482, 'timing/model': 1.174873869982548, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 1.27}) +[2026-06-24 01:11:24,266][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.031536512076854706, 'train/grad_norm_pre_clip': 11.709582328796387, 'timing/data': 0.0007148300064727664, 'timing/model': 1.17842046302394, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 1.27}) +[2026-06-24 01:11:28,595][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.028626754879951477, 'train/grad_norm_pre_clip': 18.695280075073242, 'timing/data': 0.0007428019889630377, 'timing/model': 1.1752270529977977, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 1.27}) +[2026-06-24 01:11:32,928][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.0005991256330162287, 'train/grad_norm_pre_clip': 19.990379333496094, 'timing/data': 0.001059077010722831, 'timing/model': 1.1816291609720793, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 1.27}) +[2026-06-24 01:11:37,469][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.008684652857482433, 'train/grad_norm_pre_clip': 19.982200622558594, 'timing/data': 0.0007146610005293041, 'timing/model': 1.1986516540055163, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 1.27}) +[2026-06-24 01:11:41,826][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.0065940325148403645, 'train/grad_norm_pre_clip': 9.751472473144531, 'timing/data': 0.0007205709989648312, 'timing/model': 1.1961435600242112, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 1.27}) +[2026-06-24 01:11:46,188][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.026416072621941566, 'train/grad_norm_pre_clip': 13.312921524047852, 'timing/data': 0.0007214329962152988, 'timing/model': 1.1809380439808592, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 1.27}) +[2026-06-24 01:11:50,560][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.042611513286828995, 'train/grad_norm_pre_clip': 12.291422843933105, 'timing/data': 0.0011151799990329891, 'timing/model': 1.1941395659814589, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 1.27}) +[2026-06-24 01:11:55,129][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.013868352398276329, 'train/grad_norm_pre_clip': 16.38570785522461, 'timing/data': 0.000763299991376698, 'timing/model': 1.1964146310056094, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 1.27}) +[2026-06-24 01:11:59,463][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.05826660990715027, 'train/grad_norm_pre_clip': 22.684865951538086, 'timing/data': 0.0009493939869571477, 'timing/model': 1.1797451199963689, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 1.28}) +[2026-06-24 01:12:03,833][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.021274061873555183, 'train/grad_norm_pre_clip': 124.66569519042969, 'timing/data': 0.0007585010025650263, 'timing/model': 1.1976185849925969, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 1.28}) +[2026-06-24 01:12:08,211][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.040474895387887955, 'train/grad_norm_pre_clip': 23.86488914489746, 'timing/data': 0.0008836329798214138, 'timing/model': 1.1806576960079838, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 1.28}) +[2026-06-24 01:12:12,771][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.0017943548737093806, 'train/grad_norm_pre_clip': 13.004478454589844, 'timing/data': 0.0012469339999370277, 'timing/model': 1.1987189879873767, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 1.28}) +[2026-06-24 01:12:17,108][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.017199259251356125, 'train/grad_norm_pre_clip': 27.029146194458008, 'timing/data': 0.0007296979892998934, 'timing/model': 1.1983736800029874, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 1.28}) +[2026-06-24 01:12:21,451][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3300 +[2026-06-24 01:13:19,280][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3300 +[2026-06-24 01:13:19,281][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3300 +[2026-06-24 01:13:56,054][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3300 +[2026-06-24 01:13:56,056][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.017141621559858322, 'train/grad_norm_pre_clip': 23.90233039855957, 'eval/action_loss/seconds': 57.82536408599117, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03747190162539482, 'eval/latency_2/loss': 0.03747190162539482, 'eval/flappy/loss': 0.03747190162539482, 'eval/flappy/latency_2/loss': 0.03747190162539482, 'eval/flap_precision': 0.9758064516129032, 'eval/flap_recall': 0.9850746268656716, 'eval/flap_f1': 0.9804186360567184, 'eval/flap_rate_model': 0.20666666666666667, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9804186360567184, 'eval/latency_2/flap_precision': 0.9758064516129032, 'eval/latency_2/flap_recall': 0.9850746268656716, 'eval/latency_2/flap_f1': 0.9804186360567184, 'eval/latency_2/flap_rate_model': 0.20666666666666667, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9804186360567184, 'eval/action_classification/data_wait_seconds': 0.5243767499923706, 'eval/action_classification/predict_seconds': 36.10379409790039, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0008137740078382194, 'timing/model': 1.1975282980129123, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 1.28}) +[2026-06-24 01:14:00,645][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.16639508306980133, 'train/grad_norm_pre_clip': 20.716739654541016, 'timing/data': 0.0007375730201601982, 'timing/model': 1.1836156909994315, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 1.28}) +[2026-06-24 01:14:04,977][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.07096189260482788, 'train/grad_norm_pre_clip': 17.948896408081055, 'timing/data': 0.0007815840071998537, 'timing/model': 1.1982095059938729, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 1.28}) +[2026-06-24 01:14:09,284][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.10840221494436264, 'train/grad_norm_pre_clip': 18.75126838684082, 'timing/data': 0.0007507379923481494, 'timing/model': 1.1938133729854599, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 1.28}) +[2026-06-24 01:14:13,603][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.08810795098543167, 'train/grad_norm_pre_clip': 14.729326248168945, 'timing/data': 0.0007609460153616965, 'timing/model': 1.171436005999567, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 1.28}) +[2026-06-24 01:14:18,197][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.0378362312912941, 'train/grad_norm_pre_clip': 16.8057861328125, 'timing/data': 0.0007617670053150505, 'timing/model': 1.196240076009417, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 1.28}) +[2026-06-24 01:14:22,541][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.025396987795829773, 'train/grad_norm_pre_clip': 23.81924057006836, 'timing/data': 0.000742070988053456, 'timing/model': 1.1785502809798345, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 1.28}) +[2026-06-24 01:14:26,863][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.057768549770116806, 'train/grad_norm_pre_clip': 28.933561325073242, 'timing/data': 0.0009022369922604412, 'timing/model': 1.1841804940195289, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 1.28}) +[2026-06-24 01:14:31,158][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.00687560997903347, 'train/grad_norm_pre_clip': 6.800647258758545, 'timing/data': 0.0007325639890041202, 'timing/model': 1.183230819995515, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 1.28}) +[2026-06-24 01:14:35,689][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.04683830961585045, 'train/grad_norm_pre_clip': 18.746639251708984, 'timing/data': 0.0007528509886469692, 'timing/model': 1.1948587340011727, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 1.28}) +[2026-06-24 01:14:40,050][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.2027588039636612, 'train/grad_norm_pre_clip': 29.158100128173828, 'timing/data': 0.0007732590020168573, 'timing/model': 1.198875064990716, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 1.28}) +[2026-06-24 01:14:44,391][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.08887403458356857, 'train/grad_norm_pre_clip': 39.39430618286133, 'timing/data': 0.000787364988354966, 'timing/model': 1.176387564977631, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 1.28}) +[2026-06-24 01:14:48,699][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.05837720260024071, 'train/grad_norm_pre_clip': 22.218088150024414, 'timing/data': 0.001150786003563553, 'timing/model': 1.1731402349832933, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 1.28}) +[2026-06-24 01:14:53,267][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.012103370390832424, 'train/grad_norm_pre_clip': 28.98082160949707, 'timing/data': 0.0008906250004656613, 'timing/model': 1.1939705229888204, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 1.28}) +[2026-06-24 01:14:57,610][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.011694871820509434, 'train/grad_norm_pre_clip': 7.542563438415527, 'timing/data': 0.000981963996309787, 'timing/model': 1.1899219820043072, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 1.28}) +[2026-06-24 01:15:01,922][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.12240082025527954, 'train/grad_norm_pre_clip': 34.15372085571289, 'timing/data': 0.001163629989605397, 'timing/model': 1.1795744739938527, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 1.28}) +[2026-06-24 01:15:06,439][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.14294151961803436, 'train/grad_norm_pre_clip': 25.620256423950195, 'timing/data': 0.0006841339927632362, 'timing/model': 1.3642592940013856, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 1.28}) +[2026-06-24 01:15:10,821][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.011960308998823166, 'train/grad_norm_pre_clip': 17.720169067382812, 'timing/data': 0.0008494700014125556, 'timing/model': 1.2015743920055684, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 1.28}) +[2026-06-24 01:15:15,172][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.054291822016239166, 'train/grad_norm_pre_clip': 41.02561950683594, 'timing/data': 0.0010277690016664565, 'timing/model': 1.176106716011418, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 1.28}) +[2026-06-24 01:15:19,509][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.08036952465772629, 'train/grad_norm_pre_clip': 39.33657455444336, 'timing/data': 0.001088822027668357, 'timing/model': 1.1770862159901299, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 1.28}) +[2026-06-24 01:15:24,129][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.08123912662267685, 'train/grad_norm_pre_clip': 35.398521423339844, 'timing/data': 0.0008133630035445094, 'timing/model': 1.4529282370058354, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 1.28}) +[2026-06-24 01:15:28,507][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.018194274976849556, 'train/grad_norm_pre_clip': 25.883407592773438, 'timing/data': 0.0009429619822185487, 'timing/model': 1.1949403749895282, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 1.29}) +[2026-06-24 01:15:32,889][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.0033863175194710493, 'train/grad_norm_pre_clip': 20.45429801940918, 'timing/data': 0.0008082639833446592, 'timing/model': 1.1864823270007037, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 1.29}) +[2026-06-24 01:15:37,260][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.0038189685437828302, 'train/grad_norm_pre_clip': 5.132765293121338, 'timing/data': 0.0011074460053350776, 'timing/model': 1.195578332990408, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 1.29}) +[2026-06-24 01:15:41,787][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.05499771237373352, 'train/grad_norm_pre_clip': 30.99477767944336, 'timing/data': 0.0011696410074364394, 'timing/model': 1.2022125509975012, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 1.29}) +[2026-06-24 01:15:46,162][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.05930693447589874, 'train/grad_norm_pre_clip': 11.767388343811035, 'timing/data': 0.0011695910070557147, 'timing/model': 1.1900776800175663, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 1.29}) +[2026-06-24 01:15:50,545][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.019683126360177994, 'train/grad_norm_pre_clip': 4.924351692199707, 'timing/data': 0.0007465900271199644, 'timing/model': 1.2025808109901845, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 1.29}) +[2026-06-24 01:15:54,925][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.0635809376835823, 'train/grad_norm_pre_clip': 34.88102340698242, 'timing/data': 0.0008934710058383644, 'timing/model': 1.1796301849826705, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 1.29}) +[2026-06-24 01:15:59,525][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.1487560123205185, 'train/grad_norm_pre_clip': 32.672828674316406, 'timing/data': 0.0008819899812806398, 'timing/model': 1.4315896689950023, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 1.29}) +[2026-06-24 01:16:03,894][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.08370278030633926, 'train/grad_norm_pre_clip': 32.454219818115234, 'timing/data': 0.0007600339886266738, 'timing/model': 1.1917976870026905, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 1.29}) +[2026-06-24 01:16:08,292][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.03134852647781372, 'train/grad_norm_pre_clip': 16.677087783813477, 'timing/data': 0.000719360017683357, 'timing/model': 1.2040213530126493, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 1.29}) +[2026-06-24 01:16:12,639][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.023160388693213463, 'train/grad_norm_pre_clip': 4.053889274597168, 'timing/data': 0.000920720980502665, 'timing/model': 1.1789442379958928, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 1.29}) +[2026-06-24 01:16:17,264][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.0025027072988450527, 'train/grad_norm_pre_clip': 30.361886978149414, 'timing/data': 0.000774962012656033, 'timing/model': 1.4659188889781944, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 1.29}) +[2026-06-24 01:16:21,610][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.10151969641447067, 'train/grad_norm_pre_clip': 29.731096267700195, 'timing/data': 0.0009697620116639882, 'timing/model': 1.1898016589984763, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 1.29}) +[2026-06-24 01:16:26,013][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.027945125475525856, 'train/grad_norm_pre_clip': 29.791990280151367, 'timing/data': 0.0014506599982269108, 'timing/model': 1.2037165999936406, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.29}) +[2026-06-24 01:16:30,348][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.054770808666944504, 'train/grad_norm_pre_clip': 30.971471786499023, 'timing/data': 0.0009681489900685847, 'timing/model': 1.1890261659864336, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 1.29}) +[2026-06-24 01:16:34,942][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.05256042256951332, 'train/grad_norm_pre_clip': 12.44420051574707, 'timing/data': 0.0007414399879053235, 'timing/model': 1.4355310829996597, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 1.29}) +[2026-06-24 01:16:39,246][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.029262077063322067, 'train/grad_norm_pre_clip': 10.185709953308105, 'timing/data': 0.0007505160174332559, 'timing/model': 1.1762242230179254, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 1.29}) +[2026-06-24 01:16:43,609][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.030883006751537323, 'train/grad_norm_pre_clip': 13.132049560546875, 'timing/data': 0.0008028540178202093, 'timing/model': 1.1979035490076058, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 1.29}) +[2026-06-24 01:16:47,968][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.2109798789024353, 'train/grad_norm_pre_clip': 29.70962142944336, 'timing/data': 0.0007518589845858514, 'timing/model': 1.2007218360085972, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 1.29}) +[2026-06-24 01:16:52,611][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.014919550158083439, 'train/grad_norm_pre_clip': 27.058961868286133, 'timing/data': 0.0007038399926386774, 'timing/model': 1.46561769201071, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 1.29}) +[2026-06-24 01:16:56,945][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.04960785061120987, 'train/grad_norm_pre_clip': 14.668519020080566, 'timing/data': 0.0009560270118527114, 'timing/model': 1.1749757850193419, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 1.29}) +[2026-06-24 01:17:01,286][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.04547085985541344, 'train/grad_norm_pre_clip': 25.940187454223633, 'timing/data': 0.0009325729915872216, 'timing/model': 1.1803816739993636, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 1.29}) +[2026-06-24 01:17:05,617][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.15327808260917664, 'train/grad_norm_pre_clip': 33.116085052490234, 'timing/data': 0.00099474802846089, 'timing/model': 1.1819047380122356, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 1.29}) +[2026-06-24 01:17:10,208][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.07475981116294861, 'train/grad_norm_pre_clip': 33.62483596801758, 'timing/data': 0.0008965760061983019, 'timing/model': 1.4333143350086175, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 1.29}) +[2026-06-24 01:17:14,589][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.010205655358731747, 'train/grad_norm_pre_clip': 23.37532615661621, 'timing/data': 0.0007752719975542277, 'timing/model': 1.2160828380146995, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 1.29}) +[2026-06-24 01:17:18,903][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.045552708208560944, 'train/grad_norm_pre_clip': 11.397428512573242, 'timing/data': 0.0009144099894911051, 'timing/model': 1.1788098200049717, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 1.29}) +[2026-06-24 01:17:23,287][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.039019010961055756, 'train/grad_norm_pre_clip': 18.468997955322266, 'timing/data': 0.0007895889866631478, 'timing/model': 1.1806070900056511, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 1.3}) +[2026-06-24 01:17:27,923][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.001633070525713265, 'train/grad_norm_pre_clip': 11.72889518737793, 'timing/data': 0.00076390101457946, 'timing/model': 1.4413818409957457, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 1.3}) +[2026-06-24 01:17:32,306][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.14495868980884552, 'train/grad_norm_pre_clip': 28.48371124267578, 'timing/data': 0.0009505660273134708, 'timing/model': 1.2021610460069496, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 1.3}) +[2026-06-24 01:17:36,670][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.0013628914020955563, 'train/grad_norm_pre_clip': 40.632530212402344, 'timing/data': 0.0009100220049731433, 'timing/model': 1.180764721997548, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 1.3}) +[2026-06-24 01:17:41,039][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.000805573770776391, 'train/grad_norm_pre_clip': 68.961669921875, 'timing/data': 0.0007584220147691667, 'timing/model': 1.2085914869967382, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 1.3}) +[2026-06-24 01:17:45,669][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.0633154883980751, 'train/grad_norm_pre_clip': 37.23176193237305, 'timing/data': 0.0009061640012077987, 'timing/model': 1.4571889890066814, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 1.3}) +[2026-06-24 01:17:50,046][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.058576468378305435, 'train/grad_norm_pre_clip': 22.059226989746094, 'timing/data': 0.000738003000151366, 'timing/model': 1.202815886004828, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 1.3}) +[2026-06-24 01:17:54,446][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.112004853785038, 'train/grad_norm_pre_clip': 15.57087230682373, 'timing/data': 0.0007728580094408244, 'timing/model': 1.2058915679808706, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 1.3}) +[2026-06-24 01:17:58,806][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.06885646283626556, 'train/grad_norm_pre_clip': 20.975618362426758, 'timing/data': 0.0008944230212364346, 'timing/model': 1.2093960329948459, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 1.3}) +[2026-06-24 01:18:03,450][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.1481432318687439, 'train/grad_norm_pre_clip': 23.34925079345703, 'timing/data': 0.0007683100120630115, 'timing/model': 1.4631731259869412, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 1.3}) +[2026-06-24 01:18:07,863][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.03131791949272156, 'train/grad_norm_pre_clip': 92.11874389648438, 'timing/data': 0.0012591760023497045, 'timing/model': 1.2125546589959413, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 1.3}) +[2026-06-24 01:18:12,232][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.04831651225686073, 'train/grad_norm_pre_clip': 13.6361722946167, 'timing/data': 0.00101483499747701, 'timing/model': 1.183451615012018, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 1.3}) +[2026-06-24 01:18:16,573][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.06974036246538162, 'train/grad_norm_pre_clip': 25.25617218017578, 'timing/data': 0.0009736479842104018, 'timing/model': 1.1832927619980182, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 1.3}) +[2026-06-24 01:18:21,196][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.013990447856485844, 'train/grad_norm_pre_clip': 21.068166732788086, 'timing/data': 0.0007789600058458745, 'timing/model': 1.457807711994974, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 1.3}) +[2026-06-24 01:18:25,520][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.04185366630554199, 'train/grad_norm_pre_clip': 19.111541748046875, 'timing/data': 0.000974931987002492, 'timing/model': 1.182434156013187, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 1.3}) +[2026-06-24 01:18:29,897][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.0670386329293251, 'train/grad_norm_pre_clip': 24.057817459106445, 'timing/data': 0.0013485719973687083, 'timing/model': 1.2054420369968284, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 1.3}) +[2026-06-24 01:18:34,276][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.021762743592262268, 'train/grad_norm_pre_clip': 13.80569076538086, 'timing/data': 0.0013857710000593215, 'timing/model': 1.20224528099061, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 1.3}) +[2026-06-24 01:18:38,880][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.061870478093624115, 'train/grad_norm_pre_clip': 11.135522842407227, 'timing/data': 0.000981543998932466, 'timing/model': 1.210493510006927, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 1.3}) +[2026-06-24 01:18:43,291][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.016915082931518555, 'train/grad_norm_pre_clip': 30.96303939819336, 'timing/data': 0.0009357989765703678, 'timing/model': 1.2179400399909355, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 1.3}) +[2026-06-24 01:18:47,696][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.0009187971008941531, 'train/grad_norm_pre_clip': 20.1886043548584, 'timing/data': 0.001049637998221442, 'timing/model': 1.184202553005889, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 1.3}) +[2026-06-24 01:18:52,073][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.023814944550395012, 'train/grad_norm_pre_clip': 28.030467987060547, 'timing/data': 0.0009985229989979416, 'timing/model': 1.2007527999812737, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 1.3}) +[2026-06-24 01:18:56,637][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.0028321577701717615, 'train/grad_norm_pre_clip': 84.70384979248047, 'timing/data': 0.001167856011306867, 'timing/model': 1.1823365690070204, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 1.3}) +[2026-06-24 01:19:01,044][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.05411365255713463, 'train/grad_norm_pre_clip': 36.0932731628418, 'timing/data': 0.0007895170128904283, 'timing/model': 1.2149442590016406, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 1.3}) +[2026-06-24 01:19:05,428][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.002429502084851265, 'train/grad_norm_pre_clip': 17.73602867126465, 'timing/data': 0.0010861960181500763, 'timing/model': 1.2093348389898892, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 1.3}) +[2026-06-24 01:19:09,825][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.015118548646569252, 'train/grad_norm_pre_clip': 15.97969913482666, 'timing/data': 0.0010616600047796965, 'timing/model': 1.2093596490158234, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 1.3}) +[2026-06-24 01:19:14,359][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.049423955380916595, 'train/grad_norm_pre_clip': 16.538692474365234, 'timing/data': 0.0007320820004679263, 'timing/model': 1.1815496750059538, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 1.31}) +[2026-06-24 01:19:18,749][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.08110323548316956, 'train/grad_norm_pre_clip': 20.98525619506836, 'timing/data': 0.000813332007965073, 'timing/model': 1.2020067839766853, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 1.31}) +[2026-06-24 01:19:23,142][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.026682516559958458, 'train/grad_norm_pre_clip': 21.586944580078125, 'timing/data': 0.0008397299970965832, 'timing/model': 1.2080433730152436, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 1.31}) +[2026-06-24 01:19:27,552][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.11178550124168396, 'train/grad_norm_pre_clip': 31.83406639099121, 'timing/data': 0.0012467220076359808, 'timing/model': 1.2165458450035658, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 1.31}) +[2026-06-24 01:19:32,181][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.022420519962906837, 'train/grad_norm_pre_clip': 4.965476036071777, 'timing/data': 0.0007540229998994619, 'timing/model': 1.4398637780104764, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 1.31}) +[2026-06-24 01:19:36,557][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.03110405243933201, 'train/grad_norm_pre_clip': 12.763699531555176, 'timing/data': 0.000756686000386253, 'timing/model': 1.1969309909909498, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 1.31}) +[2026-06-24 01:19:40,899][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.015787174925208092, 'train/grad_norm_pre_clip': 14.638053894042969, 'timing/data': 0.0011348649859428406, 'timing/model': 1.2042215040128212, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 1.31}) +[2026-06-24 01:19:45,257][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.07070788741111755, 'train/grad_norm_pre_clip': 18.731040954589844, 'timing/data': 0.0007302890007849783, 'timing/model': 1.1984593799861614, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 1.31}) +[2026-06-24 01:19:49,893][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.00651582982391119, 'train/grad_norm_pre_clip': 16.58828353881836, 'timing/data': 0.0007322109886445105, 'timing/model': 1.4529894260049332, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 1.31}) +[2026-06-24 01:19:54,247][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.027008138597011566, 'train/grad_norm_pre_clip': 13.692663192749023, 'timing/data': 0.0007685589953325689, 'timing/model': 1.1993947440059856, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 1.31}) +[2026-06-24 01:19:58,613][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.026242587715387344, 'train/grad_norm_pre_clip': 21.4947452545166, 'timing/data': 0.0007321419834624976, 'timing/model': 1.2072150759922806, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 1.31}) +[2026-06-24 01:20:02,944][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.10607259720563889, 'train/grad_norm_pre_clip': 23.767480850219727, 'timing/data': 0.0010370440140832216, 'timing/model': 1.1801452109939419, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.31}) +[2026-06-24 01:20:07,528][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.09923068434000015, 'train/grad_norm_pre_clip': 19.60264015197754, 'timing/data': 0.0007755719998385757, 'timing/model': 1.4226572310144547, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 1.31}) +[2026-06-24 01:20:11,917][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.04326312243938446, 'train/grad_norm_pre_clip': 24.3557071685791, 'timing/data': 0.0007827949884813279, 'timing/model': 1.2062890350061934, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 1.31}) +[2026-06-24 01:20:16,264][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.03523877263069153, 'train/grad_norm_pre_clip': 11.760826110839844, 'timing/data': 0.0009055419941432774, 'timing/model': 1.1796150220034178, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 1.31}) +[2026-06-24 01:20:20,589][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.013267965987324715, 'train/grad_norm_pre_clip': 19.519001007080078, 'timing/data': 0.0007858999888412654, 'timing/model': 1.1823948660166934, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 1.31}) +[2026-06-24 01:20:25,153][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.04552102088928223, 'train/grad_norm_pre_clip': 24.83978271484375, 'timing/data': 0.0010656380036380142, 'timing/model': 1.4191648029955104, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 1.31}) +[2026-06-24 01:20:29,513][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.04970868304371834, 'train/grad_norm_pre_clip': 15.555058479309082, 'timing/data': 0.0007393049891106784, 'timing/model': 1.1948125479975715, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 1.31}) +[2026-06-24 01:20:33,881][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.044298022985458374, 'train/grad_norm_pre_clip': 12.767075538635254, 'timing/data': 0.0007707930053584278, 'timing/model': 1.2087914790026844, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 1.31}) +[2026-06-24 01:20:38,205][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.0018013154622167349, 'train/grad_norm_pre_clip': 20.01180076599121, 'timing/data': 0.0007097499910742044, 'timing/model': 1.1831151339865755, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 1.31}) +[2026-06-24 01:20:42,708][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.011363686062395573, 'train/grad_norm_pre_clip': 12.212987899780273, 'timing/data': 0.0010784210171550512, 'timing/model': 1.1815038220083807, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 1.31}) +[2026-06-24 01:20:47,039][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.15390586853027344, 'train/grad_norm_pre_clip': 30.21601104736328, 'timing/data': 0.000889552989974618, 'timing/model': 1.182907045003958, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 1.31}) +[2026-06-24 01:20:51,429][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.14265874028205872, 'train/grad_norm_pre_clip': 22.115510940551758, 'timing/data': 0.0009376610105391592, 'timing/model': 1.18027772201458, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 1.31}) +[2026-06-24 01:20:55,819][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.04606455937027931, 'train/grad_norm_pre_clip': 34.61965560913086, 'timing/data': 0.0008924379944801331, 'timing/model': 1.1898988979810383, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 1.31}) +[2026-06-24 01:21:00,393][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.012237987481057644, 'train/grad_norm_pre_clip': 36.928314208984375, 'timing/data': 0.0007329430081881583, 'timing/model': 1.431043521995889, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 1.31}) +[2026-06-24 01:21:04,713][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.1546984612941742, 'train/grad_norm_pre_clip': 27.928504943847656, 'timing/data': 0.0007394449785351753, 'timing/model': 1.1773935119854286, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 1.31}) +[2026-06-24 01:21:09,039][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.08001401275396347, 'train/grad_norm_pre_clip': 17.993314743041992, 'timing/data': 0.0007486719987355173, 'timing/model': 1.1844110899837688, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 1.32}) +[2026-06-24 01:21:13,379][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.07488624006509781, 'train/grad_norm_pre_clip': 30.9085636138916, 'timing/data': 0.0007801300089340657, 'timing/model': 1.195654221024597, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 1.32}) +[2026-06-24 01:21:17,964][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3400 +[2026-06-24 01:22:15,100][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3400 +[2026-06-24 01:22:15,100][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3400 +[2026-06-24 01:22:51,898][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3400 +[2026-06-24 01:22:51,899][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.02276327833533287, 'train/grad_norm_pre_clip': 15.039518356323242, 'eval/action_loss/seconds': 57.13188619198627, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.040595512837171555, 'eval/latency_2/loss': 0.040595512837171555, 'eval/flappy/loss': 0.040595512837171555, 'eval/flappy/latency_2/loss': 0.040595512837171555, 'eval/flap_precision': 0.9779310344827586, 'eval/flap_recall': 0.9620081411126187, 'eval/flap_f1': 0.969904240766074, 'eval/flap_rate_model': 0.2013888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.969904240766074, 'eval/latency_2/flap_precision': 0.9779310344827586, 'eval/latency_2/flap_recall': 0.9620081411126187, 'eval/latency_2/flap_f1': 0.969904240766074, 'eval/latency_2/flap_rate_model': 0.2013888888888889, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.969904240766074, 'eval/action_classification/data_wait_seconds': 0.4916054606437683, 'eval/action_classification/predict_seconds': 36.1617546081543, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007822539773769677, 'timing/model': 1.4479977199807763, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 1.32}) +[2026-06-24 01:22:56,217][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.02211633324623108, 'train/grad_norm_pre_clip': 35.15227508544922, 'timing/data': 0.001008342020213604, 'timing/model': 1.1677514170005452, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 1.32}) +[2026-06-24 01:23:00,507][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.044335320591926575, 'train/grad_norm_pre_clip': 10.268446922302246, 'timing/data': 0.0007206410227809101, 'timing/model': 1.1676742630079389, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 1.32}) +[2026-06-24 01:23:04,784][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.10926442593336105, 'train/grad_norm_pre_clip': 23.724586486816406, 'timing/data': 0.0007991250022314489, 'timing/model': 1.1698846939834766, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 1.32}) +[2026-06-24 01:23:09,373][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.03224663436412811, 'train/grad_norm_pre_clip': 16.054237365722656, 'timing/data': 0.0013431599945761263, 'timing/model': 1.1869123469805345, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 1.32}) +[2026-06-24 01:23:13,705][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.01723322458565235, 'train/grad_norm_pre_clip': 12.519041061401367, 'timing/data': 0.0012055670085828751, 'timing/model': 1.1697203499788884, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 1.32}) +[2026-06-24 01:23:18,068][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.06530554592609406, 'train/grad_norm_pre_clip': 18.052701950073242, 'timing/data': 0.0007541420054621994, 'timing/model': 1.193484070012346, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 1.32}) +[2026-06-24 01:23:22,398][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.0005024150596000254, 'train/grad_norm_pre_clip': 23.4545841217041, 'timing/data': 0.0007356290007010102, 'timing/model': 1.1919440949859563, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 1.32}) +[2026-06-24 01:23:26,946][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.02747686579823494, 'train/grad_norm_pre_clip': 19.24066162109375, 'timing/data': 0.0007660539995413274, 'timing/model': 1.1898622500011697, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 1.32}) +[2026-06-24 01:23:31,318][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.004192643333226442, 'train/grad_norm_pre_clip': 9.054824829101562, 'timing/data': 0.0010468130058143288, 'timing/model': 1.193828883988317, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 1.32}) +[2026-06-24 01:23:35,652][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.00863043311983347, 'train/grad_norm_pre_clip': 17.44403076171875, 'timing/data': 0.0007779260049574077, 'timing/model': 1.1737436710100155, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 1.32}) +[2026-06-24 01:23:39,991][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.02595546282827854, 'train/grad_norm_pre_clip': 13.40492057800293, 'timing/data': 0.0009528899972792715, 'timing/model': 1.1748966579907574, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 1.32}) +[2026-06-24 01:23:44,529][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.012215977534651756, 'train/grad_norm_pre_clip': 7.368767261505127, 'timing/data': 0.0009458870044909418, 'timing/model': 1.1718293130106758, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 1.32}) +[2026-06-24 01:23:48,883][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.03590093553066254, 'train/grad_norm_pre_clip': 39.970375061035156, 'timing/data': 0.0007913610024843365, 'timing/model': 1.1921598250046372, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 1.32}) +[2026-06-24 01:23:53,235][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.035078566521406174, 'train/grad_norm_pre_clip': 23.621313095092773, 'timing/data': 0.0015604520158376545, 'timing/model': 1.1986091040016618, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 1.32}) +[2026-06-24 01:23:57,646][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.07945896685123444, 'train/grad_norm_pre_clip': 24.41421127319336, 'timing/data': 0.001399074011715129, 'timing/model': 1.2055384280101862, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 1.32}) +[2026-06-24 01:24:02,183][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.029466072097420692, 'train/grad_norm_pre_clip': 18.14878273010254, 'timing/data': 0.0007606750004924834, 'timing/model': 1.1753520220227074, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 1.32}) +[2026-06-24 01:24:06,539][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.07942608743906021, 'train/grad_norm_pre_clip': 28.405925750732422, 'timing/data': 0.0009215619938913733, 'timing/model': 1.1951944830070715, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 1.32}) +[2026-06-24 01:24:10,876][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.017766794189810753, 'train/grad_norm_pre_clip': 40.39720916748047, 'timing/data': 0.0008814680040813982, 'timing/model': 1.181298147974303, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 1.32}) +[2026-06-24 01:24:15,220][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.022326281294226646, 'train/grad_norm_pre_clip': 11.046847343444824, 'timing/data': 0.0010685829911381006, 'timing/model': 1.1764485699823126, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 1.32}) +[2026-06-24 01:24:19,803][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.04095950722694397, 'train/grad_norm_pre_clip': 14.578449249267578, 'timing/data': 0.0011479509994387627, 'timing/model': 1.2088919930101838, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 1.32}) +[2026-06-24 01:24:24,152][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.0897071585059166, 'train/grad_norm_pre_clip': 16.725976943969727, 'timing/data': 0.0007927039987407625, 'timing/model': 1.1966934929951094, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 1.32}) +[2026-06-24 01:24:28,525][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.0009934166446328163, 'train/grad_norm_pre_clip': 8.9695463180542, 'timing/data': 0.000963800004683435, 'timing/model': 1.1938295520085376, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 1.32}) +[2026-06-24 01:24:32,884][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.020352011546492577, 'train/grad_norm_pre_clip': 13.476377487182617, 'timing/data': 0.0007967309793457389, 'timing/model': 1.1762986149988137, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 1.32}) +[2026-06-24 01:24:37,445][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.1036258339881897, 'train/grad_norm_pre_clip': 30.771568298339844, 'timing/data': 0.000748802995076403, 'timing/model': 1.193892865005182, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 1.33}) +[2026-06-24 01:24:41,803][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.01043859962373972, 'train/grad_norm_pre_clip': 8.593565940856934, 'timing/data': 0.0009541919862385839, 'timing/model': 1.1962397070019506, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 1.33}) +[2026-06-24 01:24:46,136][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.0008057248778641224, 'train/grad_norm_pre_clip': 10.161428451538086, 'timing/data': 0.000907155015738681, 'timing/model': 1.2048272230022121, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 1.33}) +[2026-06-24 01:24:50,485][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.09200411289930344, 'train/grad_norm_pre_clip': 26.853174209594727, 'timing/data': 0.0007550639857072383, 'timing/model': 1.1955592349986546, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 1.33}) +[2026-06-24 01:24:55,070][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.19303743541240692, 'train/grad_norm_pre_clip': 57.874267578125, 'timing/data': 0.000752238993300125, 'timing/model': 1.191710735001834, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 1.33}) +[2026-06-24 01:24:59,436][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.14055806398391724, 'train/grad_norm_pre_clip': 40.590553283691406, 'timing/data': 0.0008010290039237589, 'timing/model': 1.1991144050261937, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 1.33}) +[2026-06-24 01:25:03,820][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.05745236575603485, 'train/grad_norm_pre_clip': 24.32218360900879, 'timing/data': 0.0007878050091676414, 'timing/model': 1.209999604994664, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 1.33}) +[2026-06-24 01:25:08,159][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.03720501810312271, 'train/grad_norm_pre_clip': 17.900177001953125, 'timing/data': 0.0010012389975599945, 'timing/model': 1.1822700010088738, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 1.33}) +[2026-06-24 01:25:12,705][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.005162084940820932, 'train/grad_norm_pre_clip': 39.633384704589844, 'timing/data': 0.0007392450061161071, 'timing/model': 1.1821222259895876, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.33}) +[2026-06-24 01:25:17,053][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.040222086012363434, 'train/grad_norm_pre_clip': 21.725234985351562, 'timing/data': 0.0007546629931312054, 'timing/model': 1.1779928659962025, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 1.33}) +[2026-06-24 01:25:21,371][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.000432007567724213, 'train/grad_norm_pre_clip': 21.594802856445312, 'timing/data': 0.0007524000247940421, 'timing/model': 1.1774990909907501, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 1.33}) +[2026-06-24 01:25:25,737][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.00208491669036448, 'train/grad_norm_pre_clip': 10.422460556030273, 'timing/data': 0.001137440005550161, 'timing/model': 1.1773272710270248, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 1.33}) +[2026-06-24 01:25:30,286][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.02111418917775154, 'train/grad_norm_pre_clip': 34.652587890625, 'timing/data': 0.0007577290234621614, 'timing/model': 1.1765634379989933, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 1.33}) +[2026-06-24 01:25:34,614][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.07456166297197342, 'train/grad_norm_pre_clip': 44.60139846801758, 'timing/data': 0.0010406720102764666, 'timing/model': 1.175743580010021, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 1.33}) +[2026-06-24 01:25:38,965][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.017858289182186127, 'train/grad_norm_pre_clip': 16.079519271850586, 'timing/data': 0.0007464789960067719, 'timing/model': 1.1757953919877764, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 1.33}) +[2026-06-24 01:25:43,295][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.0517122745513916, 'train/grad_norm_pre_clip': 17.609403610229492, 'timing/data': 0.0008761280041653663, 'timing/model': 1.1762748039909638, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 1.33}) +[2026-06-24 01:25:47,863][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.16629891097545624, 'train/grad_norm_pre_clip': 48.00094223022461, 'timing/data': 0.0009057130082510412, 'timing/model': 1.1814441779861227, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 1.33}) +[2026-06-24 01:25:52,246][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.015648243948817253, 'train/grad_norm_pre_clip': 8.785806655883789, 'timing/data': 0.0010201840195804834, 'timing/model': 1.2167290420038626, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 1.33}) +[2026-06-24 01:25:56,613][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.0014301117043942213, 'train/grad_norm_pre_clip': 18.7346134185791, 'timing/data': 0.000911644019652158, 'timing/model': 1.180117281008279, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 1.33}) +[2026-06-24 01:26:00,997][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.055285412818193436, 'train/grad_norm_pre_clip': 15.941932678222656, 'timing/data': 0.0008087139867711812, 'timing/model': 1.1871164229814894, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 1.33}) +[2026-06-24 01:26:05,576][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.07814981043338776, 'train/grad_norm_pre_clip': 10.409744262695312, 'timing/data': 0.0007793389959260821, 'timing/model': 1.2071851049840916, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 1.33}) +[2026-06-24 01:26:09,903][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.033566828817129135, 'train/grad_norm_pre_clip': 16.49399185180664, 'timing/data': 0.0008771400025580078, 'timing/model': 1.1789072550018318, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 1.33}) +[2026-06-24 01:26:14,279][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.0580010861158371, 'train/grad_norm_pre_clip': 26.753128051757812, 'timing/data': 0.0011784259986598045, 'timing/model': 1.1976251720043365, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 1.33}) +[2026-06-24 01:26:18,645][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.07998474687337875, 'train/grad_norm_pre_clip': 22.883859634399414, 'timing/data': 0.0010288910125382245, 'timing/model': 1.1988277380005457, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 1.33}) +[2026-06-24 01:26:23,273][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.022837849333882332, 'train/grad_norm_pre_clip': 9.5673246383667, 'timing/data': 0.0011586089967750013, 'timing/model': 1.2016005659825169, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 1.33}) +[2026-06-24 01:26:27,606][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.07740273326635361, 'train/grad_norm_pre_clip': 37.56025695800781, 'timing/data': 0.0008548390178475529, 'timing/model': 1.1913022130029276, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 1.33}) +[2026-06-24 01:26:31,920][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.04222799837589264, 'train/grad_norm_pre_clip': 49.3828010559082, 'timing/data': 0.0008051769982557744, 'timing/model': 1.1750071230053436, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 1.34}) +[2026-06-24 01:26:36,304][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.01991015300154686, 'train/grad_norm_pre_clip': 30.078445434570312, 'timing/data': 0.0013478700129780918, 'timing/model': 1.2038192030158825, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 1.34}) +[2026-06-24 01:26:40,989][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.24581469595432281, 'train/grad_norm_pre_clip': 30.559844970703125, 'timing/data': 0.0008384879911318421, 'timing/model': 1.2219666549935937, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 1.34}) +[2026-06-24 01:26:45,339][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.008075060322880745, 'train/grad_norm_pre_clip': 33.155845642089844, 'timing/data': 0.000847895978949964, 'timing/model': 1.1829094450222328, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 1.34}) +[2026-06-24 01:26:49,724][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.011986682191491127, 'train/grad_norm_pre_clip': 30.97865867614746, 'timing/data': 0.000864786998135969, 'timing/model': 1.2090794249961618, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 1.34}) +[2026-06-24 01:26:54,313][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.002350527560338378, 'train/grad_norm_pre_clip': 3.478193521499634, 'timing/data': 0.0009124350035563111, 'timing/model': 1.3913135069888085, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 1.34}) +[2026-06-24 01:26:58,706][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.002861150074750185, 'train/grad_norm_pre_clip': 39.53871536254883, 'timing/data': 0.0010570329905021936, 'timing/model': 1.213643397000851, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 1.34}) +[2026-06-24 01:27:03,093][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.22407542169094086, 'train/grad_norm_pre_clip': 16.1398983001709, 'timing/data': 0.0010638749809004366, 'timing/model': 1.205777836003108, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 1.34}) +[2026-06-24 01:27:07,479][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.06831955164670944, 'train/grad_norm_pre_clip': 10.5658540725708, 'timing/data': 0.0008016409992706031, 'timing/model': 1.2029226180166006, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 1.34}) +[2026-06-24 01:27:12,099][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.08832523226737976, 'train/grad_norm_pre_clip': 25.523521423339844, 'timing/data': 0.0007892769936006516, 'timing/model': 1.4537016510148533, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 1.34}) +[2026-06-24 01:27:16,454][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.06401494145393372, 'train/grad_norm_pre_clip': 21.434274673461914, 'timing/data': 0.0009503949841018766, 'timing/model': 1.201958168996498, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 1.34}) +[2026-06-24 01:27:20,808][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.07996149361133575, 'train/grad_norm_pre_clip': 24.954471588134766, 'timing/data': 0.0009035280090756714, 'timing/model': 1.195627131004585, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 1.34}) +[2026-06-24 01:27:25,161][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.001986841205507517, 'train/grad_norm_pre_clip': 53.27443313598633, 'timing/data': 0.0010043249931186438, 'timing/model': 1.212784356001066, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 1.34}) +[2026-06-24 01:27:29,712][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.018725808709859848, 'train/grad_norm_pre_clip': 11.716660499572754, 'timing/data': 0.0008142139995470643, 'timing/model': 1.1991811670013703, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 1.34}) +[2026-06-24 01:27:34,076][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.03370972350239754, 'train/grad_norm_pre_clip': 22.200952529907227, 'timing/data': 0.0007521590159740299, 'timing/model': 1.2049491450015921, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 1.34}) +[2026-06-24 01:27:38,456][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.0027802824042737484, 'train/grad_norm_pre_clip': 14.080924034118652, 'timing/data': 0.0007394549902528524, 'timing/model': 1.204591159010306, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 1.34}) +[2026-06-24 01:27:42,800][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.00014464558626059443, 'train/grad_norm_pre_clip': 9.527920722961426, 'timing/data': 0.0014045030111446977, 'timing/model': 1.181488353991881, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 1.34}) +[2026-06-24 01:27:47,340][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.026042787358164787, 'train/grad_norm_pre_clip': 9.695094108581543, 'timing/data': 0.0008797349873930216, 'timing/model': 1.1776168890064582, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 1.34}) +[2026-06-24 01:27:51,695][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.0031154064927250147, 'train/grad_norm_pre_clip': 7.9461541175842285, 'timing/data': 0.0008686840010341257, 'timing/model': 1.175516974995844, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 1.34}) +[2026-06-24 01:27:56,033][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.05390962213277817, 'train/grad_norm_pre_clip': 10.096074104309082, 'timing/data': 0.000751157000195235, 'timing/model': 1.1794965309964027, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 1.34}) +[2026-06-24 01:28:00,354][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.06184041127562523, 'train/grad_norm_pre_clip': 17.316396713256836, 'timing/data': 0.0007194289937615395, 'timing/model': 1.1761896880052518, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 1.34}) +[2026-06-24 01:28:04,959][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.03446564823389053, 'train/grad_norm_pre_clip': 17.484466552734375, 'timing/data': 0.0008524649892933667, 'timing/model': 1.19410756797879, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 1.34}) +[2026-06-24 01:28:09,333][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.051764920353889465, 'train/grad_norm_pre_clip': 37.35606002807617, 'timing/data': 0.0007436130254063755, 'timing/model': 1.1973200470092706, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 1.34}) +[2026-06-24 01:28:13,679][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.13323907554149628, 'train/grad_norm_pre_clip': 21.699310302734375, 'timing/data': 0.001403853006195277, 'timing/model': 1.1814470039971638, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 1.34}) +[2026-06-24 01:28:18,035][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.11595948785543442, 'train/grad_norm_pre_clip': 27.16575050354004, 'timing/data': 0.0008147449989337474, 'timing/model': 1.207565249002073, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 1.34}) +[2026-06-24 01:28:22,587][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.0024756325874477625, 'train/grad_norm_pre_clip': 16.863094329833984, 'timing/data': 0.0009041099983733147, 'timing/model': 1.182259533990873, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 1.34}) +[2026-06-24 01:28:26,940][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.14688082039356232, 'train/grad_norm_pre_clip': 66.30226135253906, 'timing/data': 0.0010093540186062455, 'timing/model': 1.1842120150104165, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 1.35}) +[2026-06-24 01:28:31,274][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.15594153106212616, 'train/grad_norm_pre_clip': 30.300554275512695, 'timing/data': 0.0007524389948230237, 'timing/model': 1.1834242350014392, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 1.35}) +[2026-06-24 01:28:35,640][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.189017191529274, 'train/grad_norm_pre_clip': 20.5226993560791, 'timing/data': 0.00123541199718602, 'timing/model': 1.2097907990100794, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 1.35}) +[2026-06-24 01:28:40,237][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.11955582350492477, 'train/grad_norm_pre_clip': 40.93525695800781, 'timing/data': 0.0012809259933419526, 'timing/model': 1.1988902910088655, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 1.35}) +[2026-06-24 01:28:44,629][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.12261299043893814, 'train/grad_norm_pre_clip': 39.21137619018555, 'timing/data': 0.0007358090078923851, 'timing/model': 1.1982765419816133, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 1.35}) +[2026-06-24 01:28:48,995][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.09808999300003052, 'train/grad_norm_pre_clip': 20.394018173217773, 'timing/data': 0.0007508159906137735, 'timing/model': 1.1834223899932113, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.35}) +[2026-06-24 01:28:53,621][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.06157830357551575, 'train/grad_norm_pre_clip': 30.172094345092773, 'timing/data': 0.0011397649941500276, 'timing/model': 1.4370635240047704, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 1.35}) +[2026-06-24 01:28:57,997][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.06792335212230682, 'train/grad_norm_pre_clip': 22.063573837280273, 'timing/data': 0.0010335189872421324, 'timing/model': 1.2094361390045378, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 1.35}) +[2026-06-24 01:29:02,373][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.051972903311252594, 'train/grad_norm_pre_clip': 13.857856750488281, 'timing/data': 0.0011576180113479495, 'timing/model': 1.2239176909788512, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 1.35}) +[2026-06-24 01:29:06,763][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.017982693389058113, 'train/grad_norm_pre_clip': 13.35103702545166, 'timing/data': 0.0007960500079207122, 'timing/model': 1.1988839670084417, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 1.35}) +[2026-06-24 01:29:11,407][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.0017486357828602195, 'train/grad_norm_pre_clip': 7.544336318969727, 'timing/data': 0.0008627130009699613, 'timing/model': 1.4561618130246643, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 1.35}) +[2026-06-24 01:29:15,754][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.0181216299533844, 'train/grad_norm_pre_clip': 15.792572021484375, 'timing/data': 0.0008001980022527277, 'timing/model': 1.2018575989932287, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 1.35}) +[2026-06-24 01:29:20,133][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.06256955116987228, 'train/grad_norm_pre_clip': 14.90345573425293, 'timing/data': 0.0012195129820611328, 'timing/model': 1.210860783001408, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 1.35}) +[2026-06-24 01:29:24,528][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.0034777410328388214, 'train/grad_norm_pre_clip': 7.570724010467529, 'timing/data': 0.0009183869988191873, 'timing/model': 1.20276895101415, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 1.35}) +[2026-06-24 01:29:29,061][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.0948798805475235, 'train/grad_norm_pre_clip': 27.967113494873047, 'timing/data': 0.0007185669965110719, 'timing/model': 1.1938731250120327, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 1.35}) +[2026-06-24 01:29:33,395][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.02875616028904915, 'train/grad_norm_pre_clip': 19.038496017456055, 'timing/data': 0.0007361300231423229, 'timing/model': 1.1895637290144805, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 1.35}) +[2026-06-24 01:29:37,749][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.03566606342792511, 'train/grad_norm_pre_clip': 33.51213836669922, 'timing/data': 0.0007599729869980365, 'timing/model': 1.193083553982433, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 1.35}) +[2026-06-24 01:29:42,079][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.0863000676035881, 'train/grad_norm_pre_clip': 19.768909454345703, 'timing/data': 0.0007186169968917966, 'timing/model': 1.1925446140230633, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 1.35}) +[2026-06-24 01:29:46,650][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.045572828501462936, 'train/grad_norm_pre_clip': 15.641800880432129, 'timing/data': 0.0009033890091814101, 'timing/model': 1.1879086840199307, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 1.35}) +[2026-06-24 01:29:50,970][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.005295734386891127, 'train/grad_norm_pre_clip': 14.088942527770996, 'timing/data': 0.000808703014627099, 'timing/model': 1.1754062850086484, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 1.35}) +[2026-06-24 01:29:55,340][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.03720230236649513, 'train/grad_norm_pre_clip': 33.0748405456543, 'timing/data': 0.000745107012335211, 'timing/model': 1.1957525110046845, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 1.35}) +[2026-06-24 01:29:59,668][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.017244389280676842, 'train/grad_norm_pre_clip': 19.079788208007812, 'timing/data': 0.000808814016636461, 'timing/model': 1.176693832007004, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 1.35}) +[2026-06-24 01:30:04,160][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.068244569003582, 'train/grad_norm_pre_clip': 22.800884246826172, 'timing/data': 0.0007285659958142787, 'timing/model': 1.1902108840004075, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 1.35}) +[2026-06-24 01:30:08,515][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.04765806719660759, 'train/grad_norm_pre_clip': 12.35290241241455, 'timing/data': 0.0007110339938662946, 'timing/model': 1.2030157500121277, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 1.35}) +[2026-06-24 01:30:12,836][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3500 +[2026-06-24 01:31:10,463][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3500 +[2026-06-24 01:31:10,464][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3500 +[2026-06-24 01:31:47,100][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3500 +[2026-06-24 01:49:05,712][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3500 +[2026-06-24 01:49:05,714][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.007609704975038767, 'train/grad_norm_pre_clip': 9.482215881347656, 'eval/action_loss/seconds': 57.624594603985315, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03379644453525543, 'eval/latency_2/loss': 0.03379644453525543, 'eval/flappy/loss': 0.03379644453525543, 'eval/flappy/latency_2/loss': 0.03379644453525543, 'eval/flap_precision': 0.9808743169398907, 'eval/flap_recall': 0.9742198100407056, 'eval/flap_f1': 0.9775357385976856, 'eval/flap_rate_model': 0.20333333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9775357385976856, 'eval/latency_2/flap_precision': 0.9808743169398907, 'eval/latency_2/flap_recall': 0.9742198100407056, 'eval/latency_2/flap_f1': 0.9775357385976856, 'eval/latency_2/flap_rate_model': 0.20333333333333334, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9775357385976856, 'eval/action_classification/data_wait_seconds': 0.4099850654602051, 'eval/action_classification/predict_seconds': 36.08104705810547, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 271.7000000000154, 'rl_games_eval/mid_train/mean_length': 2206.6, 'rl_games_eval/mid_train/std_reward': 211.8258152350778, 'rl_games_eval/mid_train/std_length': 1707.0776900891185, 'rl_games_eval/mid_train/macro_mean_reward': 271.7000000000154, 'rl_games_eval/mid_train/macro_mean_length': 2206.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 271.7000000000154, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2206.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 211.8258152350778, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1707.0776900891185, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 271.7000000000154, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0007088890124578029, 'timing/model': 1.1750910129921976, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 1.35}) +[2026-06-24 01:49:05,714][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_3500_state +[2026-06-24 01:49:05,722][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 01:49:14,496][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_3500_state/model.safetensors +[2026-06-24 01:49:24,796][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_3500_state/optimizer.bin +[2026-06-24 01:49:24,797][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_3500_state/sampler.bin +[2026-06-24 01:49:24,797][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-24 01:49:24,798][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-24 01:49:25,151][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-24 01:49:25,152][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 01:49:25,176][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 01:49:29,366][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.13246819376945496, 'train/grad_norm_pre_clip': 33.70743179321289, 'timing/data': 0.0010501000215299428, 'timing/model': 1.136742934002541, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 1.35}) +[2026-06-24 01:49:33,563][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.034002501517534256, 'train/grad_norm_pre_clip': 19.306867599487305, 'timing/data': 0.0008037850202526897, 'timing/model': 1.1584162339859176, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 1.36}) +[2026-06-24 01:49:37,725][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.026353701949119568, 'train/grad_norm_pre_clip': 11.853011131286621, 'timing/data': 0.0009793089993763715, 'timing/model': 1.1384570409718435, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 1.36}) +[2026-06-24 01:49:42,122][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.1560746431350708, 'train/grad_norm_pre_clip': 21.618144989013672, 'timing/data': 0.0007922140066511929, 'timing/model': 1.1573198300029617, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 1.36}) +[2026-06-24 01:49:46,384][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.06967028230428696, 'train/grad_norm_pre_clip': 14.847835540771484, 'timing/data': 0.001297818002058193, 'timing/model': 1.169690984999761, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 1.36}) +[2026-06-24 01:49:50,609][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.07412534952163696, 'train/grad_norm_pre_clip': 14.419571876525879, 'timing/data': 0.000992904999293387, 'timing/model': 1.1689445069932844, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 1.36}) +[2026-06-24 01:49:54,850][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.03596958518028259, 'train/grad_norm_pre_clip': 11.174461364746094, 'timing/data': 0.0011875640193466097, 'timing/model': 1.1618906549992971, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 1.36}) +[2026-06-24 01:49:59,357][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.003986186347901821, 'train/grad_norm_pre_clip': 6.8502044677734375, 'timing/data': 0.0010918070038314909, 'timing/model': 1.1633914779813495, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 1.36}) +[2026-06-24 01:50:03,618][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.021212944760918617, 'train/grad_norm_pre_clip': 36.2883415222168, 'timing/data': 0.0010298730048816651, 'timing/model': 1.1749156400037464, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 1.36}) +[2026-06-24 01:50:07,837][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.002531561767682433, 'train/grad_norm_pre_clip': 26.912458419799805, 'timing/data': 0.0009844079904723912, 'timing/model': 1.1523936900193803, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 1.36}) +[2026-06-24 01:50:12,030][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.024045929312705994, 'train/grad_norm_pre_clip': 12.162607192993164, 'timing/data': 0.0007506569963879883, 'timing/model': 1.14902111000265, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 1.36}) +[2026-06-24 01:50:16,467][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.06588941067457199, 'train/grad_norm_pre_clip': 28.55860710144043, 'timing/data': 0.000884664012119174, 'timing/model': 1.1742854859912768, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 1.36}) +[2026-06-24 01:50:20,677][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.0068822866305708885, 'train/grad_norm_pre_clip': 8.915262222290039, 'timing/data': 0.000729899009456858, 'timing/model': 1.147098619985627, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 1.36}) +[2026-06-24 01:50:24,979][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.1509854644536972, 'train/grad_norm_pre_clip': 22.64656639099121, 'timing/data': 0.0008873190090525895, 'timing/model': 1.1762978140031919, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 1.36}) +[2026-06-24 01:50:29,199][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.005852961912751198, 'train/grad_norm_pre_clip': 20.45478630065918, 'timing/data': 0.0007327340135816485, 'timing/model': 1.154263456002809, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 1.36}) +[2026-06-24 01:50:33,722][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.013424841687083244, 'train/grad_norm_pre_clip': 5.478344917297363, 'timing/data': 0.0009222439839504659, 'timing/model': 1.1802391069941223, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 1.36}) +[2026-06-24 01:50:38,032][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.022156083956360817, 'train/grad_norm_pre_clip': 4.008530616760254, 'timing/data': 0.001144384004874155, 'timing/model': 1.1878589240077417, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 1.36}) +[2026-06-24 01:50:42,338][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.11999252438545227, 'train/grad_norm_pre_clip': 39.66465759277344, 'timing/data': 0.0009978529997169971, 'timing/model': 1.1804198210011236, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 1.36}) +[2026-06-24 01:50:46,608][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.005255171563476324, 'train/grad_norm_pre_clip': 11.613774299621582, 'timing/data': 0.0007375120185315609, 'timing/model': 1.1876400809851475, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 1.36}) +[2026-06-24 01:50:51,333][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.1855917125940323, 'train/grad_norm_pre_clip': 37.34355545043945, 'timing/data': 0.000701015000231564, 'timing/model': 1.1567039680085145, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 1.36}) +[2026-06-24 01:50:55,681][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.051913436502218246, 'train/grad_norm_pre_clip': 23.23512840270996, 'timing/data': 0.0007501759973820299, 'timing/model': 1.160500932019204, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 1.36}) +[2026-06-24 01:50:59,945][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.05367591977119446, 'train/grad_norm_pre_clip': 29.413522720336914, 'timing/data': 0.0007494839956052601, 'timing/model': 1.1586634890118148, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 1.36}) +[2026-06-24 01:51:04,246][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.0009045129991136491, 'train/grad_norm_pre_clip': 19.94710350036621, 'timing/data': 0.0009426420147065073, 'timing/model': 1.1636502119945362, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 1.36}) +[2026-06-24 01:51:08,768][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.05394894629716873, 'train/grad_norm_pre_clip': 45.796142578125, 'timing/data': 0.0008668319787830114, 'timing/model': 1.1815616130188573, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 1.36}) +[2026-06-24 01:51:13,073][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.07091843336820602, 'train/grad_norm_pre_clip': 26.545799255371094, 'timing/data': 0.0006902349996380508, 'timing/model': 1.186758785013808, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 1.36}) +[2026-06-24 01:51:17,383][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.10828137397766113, 'train/grad_norm_pre_clip': 31.704666137695312, 'timing/data': 0.001374079001834616, 'timing/model': 1.1849309490062296, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 1.36}) +[2026-06-24 01:51:21,680][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.1305294930934906, 'train/grad_norm_pre_clip': 33.30742263793945, 'timing/data': 0.0007554260082542896, 'timing/model': 1.163173912995262, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 1.37}) +[2026-06-24 01:51:26,257][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.02369813248515129, 'train/grad_norm_pre_clip': 24.0931396484375, 'timing/data': 0.0010796150017995387, 'timing/model': 1.1647901899996214, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 1.37}) +[2026-06-24 01:51:30,535][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.16677561402320862, 'train/grad_norm_pre_clip': 19.710439682006836, 'timing/data': 0.0007714749954175204, 'timing/model': 1.1824315210105851, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 1.37}) +[2026-06-24 01:51:34,864][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.007632619235664606, 'train/grad_norm_pre_clip': 15.638508796691895, 'timing/data': 0.0007594029884785414, 'timing/model': 1.1668009650020394, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.37}) +[2026-06-24 01:51:39,145][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.004489037208259106, 'train/grad_norm_pre_clip': 14.713634490966797, 'timing/data': 0.000717525981599465, 'timing/model': 1.1648363360145595, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 1.37}) +[2026-06-24 01:51:43,723][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.026311738416552544, 'train/grad_norm_pre_clip': 7.242516040802002, 'timing/data': 0.000969069980783388, 'timing/model': 1.1716135460010264, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 1.37}) +[2026-06-24 01:51:48,083][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.2186616212129593, 'train/grad_norm_pre_clip': 20.649967193603516, 'timing/data': 0.0011725759832188487, 'timing/model': 1.194487765984377, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 1.37}) +[2026-06-24 01:51:52,421][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.07343005388975143, 'train/grad_norm_pre_clip': 12.763272285461426, 'timing/data': 0.0010158169898204505, 'timing/model': 1.191150621016277, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 1.37}) +[2026-06-24 01:51:56,737][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.0042364331893622875, 'train/grad_norm_pre_clip': 20.3657169342041, 'timing/data': 0.0009772049961611629, 'timing/model': 1.1800042490067426, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 1.37}) +[2026-06-24 01:52:01,313][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.09940896928310394, 'train/grad_norm_pre_clip': 41.58150863647461, 'timing/data': 0.00082603600458242, 'timing/model': 1.1738557310018223, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 1.37}) +[2026-06-24 01:52:05,608][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.03604208678007126, 'train/grad_norm_pre_clip': 22.2799015045166, 'timing/data': 0.000817069987533614, 'timing/model': 1.1665479599905666, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 1.37}) +[2026-06-24 01:52:09,978][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.10764074325561523, 'train/grad_norm_pre_clip': 26.405378341674805, 'timing/data': 0.0008471350010950118, 'timing/model': 1.202180309977848, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 1.37}) +[2026-06-24 01:52:14,303][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.06195088475942612, 'train/grad_norm_pre_clip': 32.27879333496094, 'timing/data': 0.0008237019937951118, 'timing/model': 1.190249458013568, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 1.37}) +[2026-06-24 01:52:18,872][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.030594559386372566, 'train/grad_norm_pre_clip': 5.116219520568848, 'timing/data': 0.0007078979979269207, 'timing/model': 1.1677811689733062, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 1.37}) +[2026-06-24 01:52:23,240][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.002095356583595276, 'train/grad_norm_pre_clip': 20.48299217224121, 'timing/data': 0.0009986450022552162, 'timing/model': 1.2010566350072622, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 1.37}) +[2026-06-24 01:52:27,569][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.031713083386421204, 'train/grad_norm_pre_clip': 20.087566375732422, 'timing/data': 0.0008235910208895802, 'timing/model': 1.1812704800104257, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 1.37}) +[2026-06-24 01:52:31,893][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.027355263009667397, 'train/grad_norm_pre_clip': 8.837692260742188, 'timing/data': 0.0007871940033510327, 'timing/model': 1.1806894479959738, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 1.37}) +[2026-06-24 01:52:36,493][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.09189789742231369, 'train/grad_norm_pre_clip': 11.875800132751465, 'timing/data': 0.001239779987372458, 'timing/model': 1.1788409230066463, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 1.37}) +[2026-06-24 01:52:40,823][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.03402448445558548, 'train/grad_norm_pre_clip': 26.814355850219727, 'timing/data': 0.0007786680071149021, 'timing/model': 1.1986635049979668, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 1.37}) +[2026-06-24 01:52:45,185][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.00519957160577178, 'train/grad_norm_pre_clip': 12.741559028625488, 'timing/data': 0.0009301490208599716, 'timing/model': 1.1821724079782143, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 1.37}) +[2026-06-24 01:52:49,564][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.06137751042842865, 'train/grad_norm_pre_clip': 23.7607421875, 'timing/data': 0.0013096600014250726, 'timing/model': 1.2032620909740217, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 1.37}) +[2026-06-24 01:52:54,172][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.08548478037118912, 'train/grad_norm_pre_clip': 39.35403060913086, 'timing/data': 0.0007629689935129136, 'timing/model': 1.1788400239893235, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 1.37}) +[2026-06-24 01:52:58,503][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.22454459965229034, 'train/grad_norm_pre_clip': 39.8105354309082, 'timing/data': 0.000926250999327749, 'timing/model': 1.19087669099099, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 1.37}) +[2026-06-24 01:53:02,896][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.0054592350497841835, 'train/grad_norm_pre_clip': 14.308164596557617, 'timing/data': 0.0007964010001160204, 'timing/model': 1.198373381019337, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 1.37}) +[2026-06-24 01:53:07,269][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.10109470039606094, 'train/grad_norm_pre_clip': 28.68818473815918, 'timing/data': 0.0007198489911388606, 'timing/model': 1.2036903149855789, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 1.37}) +[2026-06-24 01:53:11,926][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.01255493238568306, 'train/grad_norm_pre_clip': 21.18999671936035, 'timing/data': 0.0007104729884304106, 'timing/model': 1.2007086779922247, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 1.37}) +[2026-06-24 01:53:16,314][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.2742767333984375, 'train/grad_norm_pre_clip': 28.83484649658203, 'timing/data': 0.000760574999731034, 'timing/model': 1.2102993979933672, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 1.38}) +[2026-06-24 01:53:20,680][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.12056025117635727, 'train/grad_norm_pre_clip': 12.932044982910156, 'timing/data': 0.0014068689779378474, 'timing/model': 1.1947263709735125, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 1.38}) +[2026-06-24 01:53:24,994][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.011409702710807323, 'train/grad_norm_pre_clip': 13.226170539855957, 'timing/data': 0.0008437390206381679, 'timing/model': 1.1816654650028795, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 1.38}) +[2026-06-24 01:53:29,597][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.053608424961566925, 'train/grad_norm_pre_clip': 19.111600875854492, 'timing/data': 0.0010146540007553995, 'timing/model': 1.1970853619859554, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 1.38}) +[2026-06-24 01:53:33,937][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.12619994580745697, 'train/grad_norm_pre_clip': 36.87538528442383, 'timing/data': 0.001057102985214442, 'timing/model': 1.176534639991587, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 1.38}) +[2026-06-24 01:53:38,318][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.001023144694045186, 'train/grad_norm_pre_clip': 31.36720085144043, 'timing/data': 0.0007547340064775199, 'timing/model': 1.1996708910155576, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 1.38}) +[2026-06-24 01:53:42,655][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.014669129624962807, 'train/grad_norm_pre_clip': 29.469514846801758, 'timing/data': 0.0009389849728904665, 'timing/model': 1.1771709980093874, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 1.38}) +[2026-06-24 01:53:47,222][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.20242801308631897, 'train/grad_norm_pre_clip': 37.5135498046875, 'timing/data': 0.0008000569941941649, 'timing/model': 1.1903460180037655, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 1.38}) +[2026-06-24 01:53:51,620][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.08326660841703415, 'train/grad_norm_pre_clip': 17.131879806518555, 'timing/data': 0.000720181007636711, 'timing/model': 1.2389608079975005, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 1.38}) +[2026-06-24 01:53:55,959][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.059259120374917984, 'train/grad_norm_pre_clip': 32.1644172668457, 'timing/data': 0.0007605549762956798, 'timing/model': 1.1753834410046693, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 1.38}) +[2026-06-24 01:54:00,266][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.026307513937354088, 'train/grad_norm_pre_clip': 27.1704158782959, 'timing/data': 0.0007204309804365039, 'timing/model': 1.1765671590110287, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 1.38}) +[2026-06-24 01:54:04,834][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.03810363635420799, 'train/grad_norm_pre_clip': 29.97567367553711, 'timing/data': 0.0008995120006147772, 'timing/model': 1.1768889049999416, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 1.38}) +[2026-06-24 01:54:09,194][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.035359643399715424, 'train/grad_norm_pre_clip': 20.122909545898438, 'timing/data': 0.0009141289920080453, 'timing/model': 1.214182339987019, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 1.38}) +[2026-06-24 01:54:13,509][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.009269758127629757, 'train/grad_norm_pre_clip': 10.767876625061035, 'timing/data': 0.0007946779951453209, 'timing/model': 1.1779486050072592, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 1.38}) +[2026-06-24 01:54:17,857][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.05892137065529823, 'train/grad_norm_pre_clip': 21.353282928466797, 'timing/data': 0.0007573890034109354, 'timing/model': 1.2084174629999325, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 1.38}) +[2026-06-24 01:54:22,461][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.10855869203805923, 'train/grad_norm_pre_clip': 46.482940673828125, 'timing/data': 0.0007445359951816499, 'timing/model': 1.19751022499986, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 1.38}) +[2026-06-24 01:54:26,806][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.018550299108028412, 'train/grad_norm_pre_clip': 25.05008888244629, 'timing/data': 0.0009078769944608212, 'timing/model': 1.1936259610229172, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 1.38}) +[2026-06-24 01:54:31,147][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.08543413132429123, 'train/grad_norm_pre_clip': 21.750375747680664, 'timing/data': 0.0007561570091638714, 'timing/model': 1.2054674380051438, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 1.38}) +[2026-06-24 01:54:35,519][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.08557002246379852, 'train/grad_norm_pre_clip': 33.919410705566406, 'timing/data': 0.0007303000020328909, 'timing/model': 1.1798412080097478, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 1.38}) +[2026-06-24 01:54:40,126][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.025975489988923073, 'train/grad_norm_pre_clip': 20.6298770904541, 'timing/data': 0.0007534210162702948, 'timing/model': 1.1966445409925655, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 1.38}) +[2026-06-24 01:54:44,483][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.004508456215262413, 'train/grad_norm_pre_clip': 7.147286415100098, 'timing/data': 0.0007350470114033669, 'timing/model': 1.1793047900137026, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 1.38}) +[2026-06-24 01:54:48,852][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.037060707807540894, 'train/grad_norm_pre_clip': 21.116518020629883, 'timing/data': 0.0009647520200815052, 'timing/model': 1.1731700289819855, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 1.38}) +[2026-06-24 01:54:53,191][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.014438110403716564, 'train/grad_norm_pre_clip': 14.05453872680664, 'timing/data': 0.000870126998052001, 'timing/model': 1.1997139859886374, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 1.38}) +[2026-06-24 01:54:57,772][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.024429263547062874, 'train/grad_norm_pre_clip': 20.474138259887695, 'timing/data': 0.0007414399879053235, 'timing/model': 1.1953915619815234, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 1.38}) +[2026-06-24 01:55:02,127][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.01732337847352028, 'train/grad_norm_pre_clip': 7.194736480712891, 'timing/data': 0.000728256010916084, 'timing/model': 1.1963487019820604, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 1.38}) +[2026-06-24 01:55:06,500][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.13047783076763153, 'train/grad_norm_pre_clip': 17.671506881713867, 'timing/data': 0.000848417985253036, 'timing/model': 1.1965481220104266, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 1.38}) +[2026-06-24 01:55:10,903][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.002357073128223419, 'train/grad_norm_pre_clip': 6.626297473907471, 'timing/data': 0.0008210770029108971, 'timing/model': 1.207756629009964, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.39}) +[2026-06-24 01:55:15,519][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.17837777733802795, 'train/grad_norm_pre_clip': 35.929988861083984, 'timing/data': 0.0007405579963233322, 'timing/model': 1.1764368119766004, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 1.39}) +[2026-06-24 01:55:19,895][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.028822939842939377, 'train/grad_norm_pre_clip': 18.613018035888672, 'timing/data': 0.0008917379891499877, 'timing/model': 1.2061920709966216, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 1.39}) +[2026-06-24 01:55:24,241][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.03039124794304371, 'train/grad_norm_pre_clip': 21.20090675354004, 'timing/data': 0.0009034280083142221, 'timing/model': 1.179821395984618, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 1.39}) +[2026-06-24 01:55:28,614][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.009383650496602058, 'train/grad_norm_pre_clip': 22.6347599029541, 'timing/data': 0.000921322003705427, 'timing/model': 1.1892811129800975, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 1.39}) +[2026-06-24 01:55:33,226][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.030809340998530388, 'train/grad_norm_pre_clip': 43.071388244628906, 'timing/data': 0.0007291459769476205, 'timing/model': 1.1795785700087436, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 1.39}) +[2026-06-24 01:55:37,594][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.23295378684997559, 'train/grad_norm_pre_clip': 25.125017166137695, 'timing/data': 0.0008142339938785881, 'timing/model': 1.1977052479924168, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 1.39}) +[2026-06-24 01:55:41,971][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.059507422149181366, 'train/grad_norm_pre_clip': 19.573888778686523, 'timing/data': 0.0007638510141987354, 'timing/model': 1.198674329003552, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 1.39}) +[2026-06-24 01:55:46,314][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.010198910720646381, 'train/grad_norm_pre_clip': 33.284603118896484, 'timing/data': 0.0007894480077084154, 'timing/model': 1.1800468389992602, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 1.39}) +[2026-06-24 01:55:50,902][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.03945469856262207, 'train/grad_norm_pre_clip': 37.65431594848633, 'timing/data': 0.0007389749807771295, 'timing/model': 1.175222527992446, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 1.39}) +[2026-06-24 01:55:55,291][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.03761081025004387, 'train/grad_norm_pre_clip': 15.376032829284668, 'timing/data': 0.0008211969980038702, 'timing/model': 1.1950848319975194, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 1.39}) +[2026-06-24 01:55:59,676][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.020673906430602074, 'train/grad_norm_pre_clip': 21.238428115844727, 'timing/data': 0.0008035039936657995, 'timing/model': 1.2067183540202677, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 1.39}) +[2026-06-24 01:56:04,005][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.0558200441300869, 'train/grad_norm_pre_clip': 20.42165756225586, 'timing/data': 0.0008647570211905986, 'timing/model': 1.1793742720037699, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 1.39}) +[2026-06-24 01:56:08,583][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.227979376912117, 'train/grad_norm_pre_clip': 25.28934097290039, 'timing/data': 0.0007388649974018335, 'timing/model': 1.1768577629991341, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 1.39}) +[2026-06-24 01:56:12,960][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.0005235397256910801, 'train/grad_norm_pre_clip': 18.50967025756836, 'timing/data': 0.0007193989877123386, 'timing/model': 1.1986679260153323, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 1.39}) +[2026-06-24 01:56:17,297][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.07157247513532639, 'train/grad_norm_pre_clip': 31.477617263793945, 'timing/data': 0.000784859003033489, 'timing/model': 1.1769202810246497, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 1.39}) +[2026-06-24 01:56:21,665][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.03161536157131195, 'train/grad_norm_pre_clip': 26.226089477539062, 'timing/data': 0.0009165230148937553, 'timing/model': 1.2049338340002578, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 1.39}) +[2026-06-24 01:56:26,293][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.01908988691866398, 'train/grad_norm_pre_clip': 22.22498321533203, 'timing/data': 0.0008830009901430458, 'timing/model': 1.2010391810035799, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 1.39}) +[2026-06-24 01:56:30,646][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.004818398505449295, 'train/grad_norm_pre_clip': 26.702064514160156, 'timing/data': 0.0008939509862102568, 'timing/model': 1.1933333560009487, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 1.39}) +[2026-06-24 01:56:35,006][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.033464860171079636, 'train/grad_norm_pre_clip': 15.584081649780273, 'timing/data': 0.0009417399996891618, 'timing/model': 1.1832257639907766, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 1.39}) +[2026-06-24 01:56:39,380][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.05284057557582855, 'train/grad_norm_pre_clip': 31.007139205932617, 'timing/data': 0.0007257710094563663, 'timing/model': 1.1968230950005818, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 1.39}) +[2026-06-24 01:56:43,981][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3600 +[2026-06-24 01:57:41,144][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3600 +[2026-06-24 01:57:41,144][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3600 +[2026-06-24 01:58:18,042][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3600 +[2026-06-24 01:58:18,044][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.02917483076453209, 'train/grad_norm_pre_clip': 18.66169548034668, 'eval/action_loss/seconds': 57.15961900001275, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03282982110977173, 'eval/latency_2/loss': 0.03282982110977173, 'eval/flappy/loss': 0.03282982110977173, 'eval/flappy/latency_2/loss': 0.03282982110977173, 'eval/flap_precision': 0.9809523809523809, 'eval/flap_recall': 0.9782903663500678, 'eval/flap_f1': 0.9796195652173914, 'eval/flap_rate_model': 0.20416666666666666, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9796195652173914, 'eval/latency_2/flap_precision': 0.9809523809523809, 'eval/latency_2/flap_recall': 0.9782903663500678, 'eval/latency_2/flap_f1': 0.9796195652173914, 'eval/latency_2/flap_rate_model': 0.20416666666666666, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9796195652173914, 'eval/action_classification/data_wait_seconds': 0.472475528717041, 'eval/action_classification/predict_seconds': 36.28090286254883, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0008539170084986836, 'timing/model': 1.1836852160049602, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 1.39}) +[2026-06-24 01:58:22,380][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.08303247392177582, 'train/grad_norm_pre_clip': 25.021459579467773, 'timing/data': 0.0007709439960308373, 'timing/model': 1.1843828870041762, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 1.39}) +[2026-06-24 01:58:26,671][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.022673916071653366, 'train/grad_norm_pre_clip': 14.947072982788086, 'timing/data': 0.001000717980787158, 'timing/model': 1.165596381004434, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 1.39}) +[2026-06-24 01:58:31,184][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.010183911770582199, 'train/grad_norm_pre_clip': 24.309524536132812, 'timing/data': 0.0008161469886545092, 'timing/model': 1.1842085459793452, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 1.39}) +[2026-06-24 01:58:35,507][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.03250822052359581, 'train/grad_norm_pre_clip': 27.298616409301758, 'timing/data': 0.0007997879874892533, 'timing/model': 1.1873317069839686, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 1.39}) +[2026-06-24 01:58:39,787][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.009058765135705471, 'train/grad_norm_pre_clip': 40.86678695678711, 'timing/data': 0.0009663950186222792, 'timing/model': 1.170919156022137, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 1.4}) +[2026-06-24 01:58:44,123][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.02483944222331047, 'train/grad_norm_pre_clip': 22.76937484741211, 'timing/data': 0.001329506019828841, 'timing/model': 1.1931156330101658, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 1.4}) +[2026-06-24 01:58:48,611][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.03759997710585594, 'train/grad_norm_pre_clip': 13.413440704345703, 'timing/data': 0.0007427620002999902, 'timing/model': 1.1748843730019871, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 1.4}) +[2026-06-24 01:58:52,952][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.022493712604045868, 'train/grad_norm_pre_clip': 30.794389724731445, 'timing/data': 0.0008261259936261922, 'timing/model': 1.189021842990769, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 1.4}) +[2026-06-24 01:58:57,302][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.05409197881817818, 'train/grad_norm_pre_clip': 45.46723175048828, 'timing/data': 0.0011131670034956187, 'timing/model': 1.18203892299789, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 1.4}) +[2026-06-24 01:59:01,621][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.05144103616476059, 'train/grad_norm_pre_clip': 12.64804458618164, 'timing/data': 0.0007989160076249391, 'timing/model': 1.1715321039955597, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 1.4}) +[2026-06-24 01:59:06,188][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.06556735187768936, 'train/grad_norm_pre_clip': 18.592439651489258, 'timing/data': 0.0008340200001839548, 'timing/model': 1.2049648779793642, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 1.4}) +[2026-06-24 01:59:10,546][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.0761476457118988, 'train/grad_norm_pre_clip': 36.86762619018555, 'timing/data': 0.0008138430130202323, 'timing/model': 1.1928583330009133, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 1.4}) +[2026-06-24 01:59:14,874][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.02287592552602291, 'train/grad_norm_pre_clip': 19.28548240661621, 'timing/data': 0.0007518189959228039, 'timing/model': 1.1902349770243745, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 1.4}) +[2026-06-24 01:59:19,239][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.0006204322562552989, 'train/grad_norm_pre_clip': 12.276982307434082, 'timing/data': 0.0012072609970346093, 'timing/model': 1.1937583679973613, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 1.4}) +[2026-06-24 01:59:23,754][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.08090764284133911, 'train/grad_norm_pre_clip': 33.27047348022461, 'timing/data': 0.0007564880070276558, 'timing/model': 1.169764761987608, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 1.4}) +[2026-06-24 01:59:28,061][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.08855845779180527, 'train/grad_norm_pre_clip': 7.166436672210693, 'timing/data': 0.0007658349932171404, 'timing/model': 1.1716622189851478, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 1.4}) +[2026-06-24 01:59:32,373][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.005797889083623886, 'train/grad_norm_pre_clip': 5.144364356994629, 'timing/data': 0.0008783419907558709, 'timing/model': 1.1912806629843544, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 1.4}) +[2026-06-24 01:59:36,674][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.041509684175252914, 'train/grad_norm_pre_clip': 14.070406913757324, 'timing/data': 0.0007495349855162203, 'timing/model': 1.172054244991159, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 1.4}) +[2026-06-24 01:59:41,163][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.0040069688111543655, 'train/grad_norm_pre_clip': 13.868799209594727, 'timing/data': 0.0007588519947603345, 'timing/model': 1.1728617269836832, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 1.4}) +[2026-06-24 01:59:45,501][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.052433520555496216, 'train/grad_norm_pre_clip': 21.278759002685547, 'timing/data': 0.001006729988148436, 'timing/model': 1.1804851819761097, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 1.4}) +[2026-06-24 01:59:49,838][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.029242027550935745, 'train/grad_norm_pre_clip': 10.892288208007812, 'timing/data': 0.0009938760194927454, 'timing/model': 1.2000471810169984, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 1.4}) +[2026-06-24 01:59:54,220][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.015081468969583511, 'train/grad_norm_pre_clip': 13.042621612548828, 'timing/data': 0.001094912993721664, 'timing/model': 1.202135900995927, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 1.4}) +[2026-06-24 01:59:58,834][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.031465932726860046, 'train/grad_norm_pre_clip': 23.9267635345459, 'timing/data': 0.000786011980380863, 'timing/model': 1.200446139991982, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 1.4}) +[2026-06-24 02:00:03,208][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.0004369405214674771, 'train/grad_norm_pre_clip': 7.06600284576416, 'timing/data': 0.0008271379920188338, 'timing/model': 1.1926470720209181, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 1.4}) +[2026-06-24 02:00:07,526][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.0023282053880393505, 'train/grad_norm_pre_clip': 21.373262405395508, 'timing/data': 0.0007521499937865883, 'timing/model': 1.1749036939872894, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 1.4}) +[2026-06-24 02:00:11,878][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.0015872754156589508, 'train/grad_norm_pre_clip': 19.44548988342285, 'timing/data': 0.0008661000174470246, 'timing/model': 1.2057897930208128, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 1.4}) +[2026-06-24 02:00:16,402][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.001426688046194613, 'train/grad_norm_pre_clip': 17.259401321411133, 'timing/data': 0.0007813230040483177, 'timing/model': 1.1928483060037252, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 1.4}) +[2026-06-24 02:00:20,718][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.08568722009658813, 'train/grad_norm_pre_clip': 11.645159721374512, 'timing/data': 0.0009133769781328738, 'timing/model': 1.1747423379856627, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.4}) +[2026-06-24 02:00:25,087][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.0008545693126507103, 'train/grad_norm_pre_clip': 22.536033630371094, 'timing/data': 0.0008859559893608093, 'timing/model': 1.2018230569956359, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 1.4}) +[2026-06-24 02:00:29,468][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.008265824988484383, 'train/grad_norm_pre_clip': 4.618918418884277, 'timing/data': 0.0007719059940427542, 'timing/model': 1.1978927169984672, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 1.4}) +[2026-06-24 02:00:34,011][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.0029434149619191885, 'train/grad_norm_pre_clip': 6.601515769958496, 'timing/data': 0.0011157809931319207, 'timing/model': 1.1954599739983678, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 1.41}) +[2026-06-24 02:00:38,360][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.05433778092265129, 'train/grad_norm_pre_clip': 23.32209014892578, 'timing/data': 0.0008929090108722448, 'timing/model': 1.1951739159994759, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 1.41}) +[2026-06-24 02:00:42,689][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.02249492146074772, 'train/grad_norm_pre_clip': 19.75772476196289, 'timing/data': 0.0010535159963183105, 'timing/model': 1.1955172710004263, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 1.41}) +[2026-06-24 02:00:47,056][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.0025062134955078363, 'train/grad_norm_pre_clip': 13.675413131713867, 'timing/data': 0.0008207159989979118, 'timing/model': 1.2012986519839615, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 1.41}) +[2026-06-24 02:00:51,591][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.033870089799165726, 'train/grad_norm_pre_clip': 21.649272918701172, 'timing/data': 0.0007602249970659614, 'timing/model': 1.197605059016496, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 1.41}) +[2026-06-24 02:00:55,918][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.06420130282640457, 'train/grad_norm_pre_clip': 22.03546905517578, 'timing/data': 0.0009212220029439777, 'timing/model': 1.1949904699868057, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 1.41}) +[2026-06-24 02:01:00,287][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.12016652524471283, 'train/grad_norm_pre_clip': 21.956159591674805, 'timing/data': 0.0007967920100782067, 'timing/model': 1.1995024149946403, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 1.41}) +[2026-06-24 02:01:04,665][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.0026085684075951576, 'train/grad_norm_pre_clip': 25.950963973999023, 'timing/data': 0.0008686950022820383, 'timing/model': 1.2010468999797013, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 1.41}) +[2026-06-24 02:01:09,221][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.0335199274122715, 'train/grad_norm_pre_clip': 18.968202590942383, 'timing/data': 0.0009016659751068801, 'timing/model': 1.1769126649887767, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 1.41}) +[2026-06-24 02:01:13,604][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.10131499916315079, 'train/grad_norm_pre_clip': 18.779525756835938, 'timing/data': 0.0009207009861711413, 'timing/model': 1.2029214549984317, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 1.41}) +[2026-06-24 02:01:17,963][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.005424375645816326, 'train/grad_norm_pre_clip': 6.903451442718506, 'timing/data': 0.001234991999808699, 'timing/model': 1.183797745994525, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 1.41}) +[2026-06-24 02:01:22,309][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.016304412856698036, 'train/grad_norm_pre_clip': 40.700435638427734, 'timing/data': 0.0008013410260900855, 'timing/model': 1.179117519990541, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 1.41}) +[2026-06-24 02:01:26,868][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.002691345289349556, 'train/grad_norm_pre_clip': 10.605960845947266, 'timing/data': 0.0008289710094686598, 'timing/model': 1.1996360560005996, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 1.41}) +[2026-06-24 02:01:31,245][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.0036481439601629972, 'train/grad_norm_pre_clip': 16.267850875854492, 'timing/data': 0.0007390659884549677, 'timing/model': 1.1950610259955283, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 1.41}) +[2026-06-24 02:01:35,606][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.0058455863036215305, 'train/grad_norm_pre_clip': 22.820018768310547, 'timing/data': 0.0007663150026928633, 'timing/model': 1.198440459993435, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 1.41}) +[2026-06-24 02:01:39,968][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.03725389391183853, 'train/grad_norm_pre_clip': 16.498004913330078, 'timing/data': 0.001283351011807099, 'timing/model': 1.1985573359997943, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 1.41}) +[2026-06-24 02:01:44,541][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.04463091120123863, 'train/grad_norm_pre_clip': 26.40308952331543, 'timing/data': 0.0008841840026434511, 'timing/model': 1.1955839529982768, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 1.41}) +[2026-06-24 02:01:48,927][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.03776984289288521, 'train/grad_norm_pre_clip': 17.326845169067383, 'timing/data': 0.0008584059833083302, 'timing/model': 1.2017679589916952, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 1.41}) +[2026-06-24 02:01:53,266][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.09847553074359894, 'train/grad_norm_pre_clip': 45.4071159362793, 'timing/data': 0.0009667470003478229, 'timing/model': 1.1991723729879595, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 1.41}) +[2026-06-24 02:01:57,631][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.006860681809484959, 'train/grad_norm_pre_clip': 15.541893005371094, 'timing/data': 0.0007578999793622643, 'timing/model': 1.1997960369917564, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 1.41}) +[2026-06-24 02:02:02,156][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.006030304357409477, 'train/grad_norm_pre_clip': 8.603470802307129, 'timing/data': 0.0007566779968328774, 'timing/model': 1.1964182369993068, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 1.41}) +[2026-06-24 02:02:06,519][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 8.593309757998213e-05, 'train/grad_norm_pre_clip': 32.020538330078125, 'timing/data': 0.0007333550020121038, 'timing/model': 1.204887993982993, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 1.41}) +[2026-06-24 02:02:10,874][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.26598671078681946, 'train/grad_norm_pre_clip': 42.397830963134766, 'timing/data': 0.000878272985573858, 'timing/model': 1.2015536249964498, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 1.41}) +[2026-06-24 02:02:15,271][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.004749649204313755, 'train/grad_norm_pre_clip': 19.33953285217285, 'timing/data': 0.0007384350174106658, 'timing/model': 1.205252048006514, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 1.41}) +[2026-06-24 02:02:19,842][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.12924045324325562, 'train/grad_norm_pre_clip': 18.864118576049805, 'timing/data': 0.0007574789924547076, 'timing/model': 1.2041085560049396, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 1.41}) +[2026-06-24 02:02:24,214][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.088015116751194, 'train/grad_norm_pre_clip': 25.995389938354492, 'timing/data': 0.0007389660167973489, 'timing/model': 1.198439903993858, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 1.41}) +[2026-06-24 02:02:28,560][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.09840817749500275, 'train/grad_norm_pre_clip': 38.998355865478516, 'timing/data': 0.000912345974938944, 'timing/model': 1.1976345669827424, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 1.42}) +[2026-06-24 02:02:32,917][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.037393249571323395, 'train/grad_norm_pre_clip': 12.258386611938477, 'timing/data': 0.00071404900518246, 'timing/model': 1.2090672999911476, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 1.42}) +[2026-06-24 02:02:37,517][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.07848964631557465, 'train/grad_norm_pre_clip': 19.28205108642578, 'timing/data': 0.0007430230034515262, 'timing/model': 1.1970099140016828, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 1.42}) +[2026-06-24 02:02:41,881][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.00045805974514223635, 'train/grad_norm_pre_clip': 22.685951232910156, 'timing/data': 0.0009634800080675632, 'timing/model': 1.1800349050026853, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 1.42}) +[2026-06-24 02:02:46,270][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.007054541725665331, 'train/grad_norm_pre_clip': 25.830507278442383, 'timing/data': 0.0008337309991475195, 'timing/model': 1.2013090079999529, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 1.42}) +[2026-06-24 02:02:50,592][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 4.4095162593293935e-05, 'train/grad_norm_pre_clip': 29.62303352355957, 'timing/data': 0.0009331339970231056, 'timing/model': 1.1811849499936216, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 1.42}) +[2026-06-24 02:02:55,132][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.06927450746297836, 'train/grad_norm_pre_clip': 21.466081619262695, 'timing/data': 0.0009078579896595329, 'timing/model': 1.1993945620197337, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 1.42}) +[2026-06-24 02:02:59,483][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.0504927821457386, 'train/grad_norm_pre_clip': 18.383377075195312, 'timing/data': 0.0007780870073474944, 'timing/model': 1.179279851989122, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 1.42}) +[2026-06-24 02:03:03,922][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.0002017551159951836, 'train/grad_norm_pre_clip': 11.382535934448242, 'timing/data': 0.0007742900052107871, 'timing/model': 1.1815737409924623, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 1.42}) +[2026-06-24 02:03:08,348][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.039338286966085434, 'train/grad_norm_pre_clip': 38.752925872802734, 'timing/data': 0.0013059829943813384, 'timing/model': 1.206529414979741, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 1.42}) +[2026-06-24 02:03:12,895][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.09763635694980621, 'train/grad_norm_pre_clip': 40.72734832763672, 'timing/data': 0.0008073709905147552, 'timing/model': 1.1934302059817128, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 1.42}) +[2026-06-24 02:03:17,215][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.025341976433992386, 'train/grad_norm_pre_clip': 8.439109802246094, 'timing/data': 0.0011301179765723646, 'timing/model': 1.1774908789957408, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 1.42}) +[2026-06-24 02:03:21,543][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.0048367055132985115, 'train/grad_norm_pre_clip': 22.524799346923828, 'timing/data': 0.000767707999330014, 'timing/model': 1.1875695099879522, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 1.42}) +[2026-06-24 02:03:25,933][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.06340765953063965, 'train/grad_norm_pre_clip': 17.956729888916016, 'timing/data': 0.0009217019833158702, 'timing/model': 1.196896410023328, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 1.42}) +[2026-06-24 02:03:30,508][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.1128193661570549, 'train/grad_norm_pre_clip': 75.21343994140625, 'timing/data': 0.0008468939922749996, 'timing/model': 1.2045011199952569, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 1.42}) +[2026-06-24 02:03:34,864][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.02083800733089447, 'train/grad_norm_pre_clip': 11.005274772644043, 'timing/data': 0.0008873699989635497, 'timing/model': 1.1805109810084105, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 1.42}) +[2026-06-24 02:03:39,229][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.014196246862411499, 'train/grad_norm_pre_clip': 42.87784194946289, 'timing/data': 0.0008957750105764717, 'timing/model': 1.1814325529849157, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 1.42}) +[2026-06-24 02:03:43,606][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.0034766518510878086, 'train/grad_norm_pre_clip': 15.994373321533203, 'timing/data': 0.0008991809736471623, 'timing/model': 1.1990903450059704, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 1.42}) +[2026-06-24 02:03:48,134][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.03225637227296829, 'train/grad_norm_pre_clip': 15.692058563232422, 'timing/data': 0.0007748019997961819, 'timing/model': 1.1802327079931274, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 1.42}) +[2026-06-24 02:03:52,525][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.1308853179216385, 'train/grad_norm_pre_clip': 49.801666259765625, 'timing/data': 0.0009708630095701665, 'timing/model': 1.2038402309990488, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 1.42}) +[2026-06-24 02:03:56,845][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.014699351973831654, 'train/grad_norm_pre_clip': 25.844600677490234, 'timing/data': 0.0007901189965195954, 'timing/model': 1.1817743670253549, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.42}) +[2026-06-24 02:04:01,199][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.014067107811570168, 'train/grad_norm_pre_clip': 16.201854705810547, 'timing/data': 0.0007672469946555793, 'timing/model': 1.1788585300091654, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 1.42}) +[2026-06-24 02:04:05,710][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.03456953912973404, 'train/grad_norm_pre_clip': 24.552356719970703, 'timing/data': 0.0010103370004799217, 'timing/model': 1.1798521770106163, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 1.42}) +[2026-06-24 02:04:10,062][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.0412086546421051, 'train/grad_norm_pre_clip': 21.235546112060547, 'timing/data': 0.0012756869837176055, 'timing/model': 1.1773695499869063, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 1.42}) +[2026-06-24 02:04:14,396][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.00015147004160098732, 'train/grad_norm_pre_clip': 16.194032669067383, 'timing/data': 0.0007821049948688596, 'timing/model': 1.180599147977773, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 1.42}) +[2026-06-24 02:04:18,767][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.08223367482423782, 'train/grad_norm_pre_clip': 34.73432159423828, 'timing/data': 0.0012269069848116487, 'timing/model': 1.2067709380062297, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 1.43}) +[2026-06-24 02:04:23,318][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.01613934524357319, 'train/grad_norm_pre_clip': 8.401251792907715, 'timing/data': 0.001069997000740841, 'timing/model': 1.1837985389865935, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 1.43}) +[2026-06-24 02:04:27,716][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.027446772903203964, 'train/grad_norm_pre_clip': 35.975929260253906, 'timing/data': 0.0007622879929840565, 'timing/model': 1.200243570987368, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 1.43}) +[2026-06-24 02:04:32,112][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.009037219919264317, 'train/grad_norm_pre_clip': 18.423871994018555, 'timing/data': 0.001442854991182685, 'timing/model': 1.1971689909987617, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 1.43}) +[2026-06-24 02:04:36,461][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.18766061961650848, 'train/grad_norm_pre_clip': 26.717029571533203, 'timing/data': 0.0009131069818977267, 'timing/model': 1.1997043659794144, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 1.43}) +[2026-06-24 02:04:41,038][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.04701344668865204, 'train/grad_norm_pre_clip': 18.592912673950195, 'timing/data': 0.0009184069931507111, 'timing/model': 1.1935863140097354, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 1.43}) +[2026-06-24 02:04:45,379][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.05988315865397453, 'train/grad_norm_pre_clip': 12.72914981842041, 'timing/data': 0.0009201400098390877, 'timing/model': 1.2045102349948138, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 1.43}) +[2026-06-24 02:04:49,771][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.01494593545794487, 'train/grad_norm_pre_clip': 14.517014503479004, 'timing/data': 0.00099153199698776, 'timing/model': 1.2079537769895978, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 1.43}) +[2026-06-24 02:04:54,183][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.0023412033915519714, 'train/grad_norm_pre_clip': 16.390987396240234, 'timing/data': 0.0010421749902889132, 'timing/model': 1.2024602080055047, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 1.43}) +[2026-06-24 02:04:58,772][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.002281692111864686, 'train/grad_norm_pre_clip': 6.245022296905518, 'timing/data': 0.0007952489831950516, 'timing/model': 1.1891650760080665, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 1.43}) +[2026-06-24 02:05:03,125][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.042377226054668427, 'train/grad_norm_pre_clip': 20.436147689819336, 'timing/data': 0.0007667360187042505, 'timing/model': 1.1780559940089006, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 1.43}) +[2026-06-24 02:05:07,481][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.012370645999908447, 'train/grad_norm_pre_clip': 29.810192108154297, 'timing/data': 0.0007331340166274458, 'timing/model': 1.1754038850194775, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 1.43}) +[2026-06-24 02:05:11,804][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.01933574490249157, 'train/grad_norm_pre_clip': 10.512969970703125, 'timing/data': 0.0007474409940186888, 'timing/model': 1.1737690239970107, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 1.43}) +[2026-06-24 02:05:16,294][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.00028134824242442846, 'train/grad_norm_pre_clip': 11.867384910583496, 'timing/data': 0.0007263919978868216, 'timing/model': 1.1748415659822058, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 1.43}) +[2026-06-24 02:05:20,629][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.04857778549194336, 'train/grad_norm_pre_clip': 10.806350708007812, 'timing/data': 0.0007479219930246472, 'timing/model': 1.1974509470164776, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 1.43}) +[2026-06-24 02:05:24,964][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.05733254551887512, 'train/grad_norm_pre_clip': 28.240947723388672, 'timing/data': 0.0007582810067106038, 'timing/model': 1.1780948969826568, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 1.43}) +[2026-06-24 02:05:29,286][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.04208371043205261, 'train/grad_norm_pre_clip': 36.22946548461914, 'timing/data': 0.0008943130087573081, 'timing/model': 1.1768948109820485, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 1.43}) +[2026-06-24 02:05:33,805][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.09511200338602066, 'train/grad_norm_pre_clip': 16.135669708251953, 'timing/data': 0.0007460979977622628, 'timing/model': 1.178992005996406, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 1.43}) +[2026-06-24 02:05:38,158][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3700 +[2026-06-24 02:06:36,156][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3700 +[2026-06-24 02:06:36,156][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3700 +[2026-06-24 02:07:13,124][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3700 +[2026-06-24 02:07:13,126][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.03939039260149002, 'train/grad_norm_pre_clip': 16.990400314331055, 'eval/action_loss/seconds': 57.99421588599216, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03410209342837334, 'eval/latency_2/loss': 0.03410209342837334, 'eval/flappy/loss': 0.03410209342837334, 'eval/flappy/latency_2/loss': 0.03410209342837334, 'eval/flap_precision': 0.9769647696476965, 'eval/flap_recall': 0.9782903663500678, 'eval/flap_f1': 0.9776271186440678, 'eval/flap_rate_model': 0.205, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9776271186440678, 'eval/latency_2/flap_precision': 0.9769647696476965, 'eval/latency_2/flap_recall': 0.9782903663500678, 'eval/latency_2/flap_f1': 0.9776271186440678, 'eval/latency_2/flap_rate_model': 0.205, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9776271186440678, 'eval/action_classification/data_wait_seconds': 0.49561622738838196, 'eval/action_classification/predict_seconds': 36.326133728027344, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0011597319971770048, 'timing/model': 1.2070906769949943, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 1.43}) +[2026-06-24 02:07:17,439][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.05263672396540642, 'train/grad_norm_pre_clip': 13.38469123840332, 'timing/data': 0.0007773560064379126, 'timing/model': 1.1825560989964288, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 1.43}) +[2026-06-24 02:07:21,900][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.0006014194223098457, 'train/grad_norm_pre_clip': 29.82028579711914, 'timing/data': 0.0007825350039638579, 'timing/model': 1.1622679170104675, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 1.43}) +[2026-06-24 02:07:26,203][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.026896607130765915, 'train/grad_norm_pre_clip': 17.1911563873291, 'timing/data': 0.0007719670247752219, 'timing/model': 1.1643176839861553, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 1.43}) +[2026-06-24 02:07:30,479][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.01693626493215561, 'train/grad_norm_pre_clip': 53.99302291870117, 'timing/data': 0.0007334149850066751, 'timing/model': 1.1847639980260283, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 1.43}) +[2026-06-24 02:07:34,805][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.06200516223907471, 'train/grad_norm_pre_clip': 21.19631576538086, 'timing/data': 0.0007488529954571277, 'timing/model': 1.1899050860083662, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 1.43}) +[2026-06-24 02:07:39,330][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.0021745075937360525, 'train/grad_norm_pre_clip': 14.765817642211914, 'timing/data': 0.0007481719949282706, 'timing/model': 1.383021401008591, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 1.43}) +[2026-06-24 02:07:43,626][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.015027852728962898, 'train/grad_norm_pre_clip': 38.10810470581055, 'timing/data': 0.0008100169943645597, 'timing/model': 1.1887124839995522, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 1.43}) +[2026-06-24 02:07:47,935][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.011321023106575012, 'train/grad_norm_pre_clip': 26.146146774291992, 'timing/data': 0.0007648630125913769, 'timing/model': 1.1968972760078032, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 1.44}) +[2026-06-24 02:07:52,270][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.014827181585133076, 'train/grad_norm_pre_clip': 22.713258743286133, 'timing/data': 0.0007637009839527309, 'timing/model': 1.1921174250019249, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 1.44}) +[2026-06-24 02:07:56,622][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.03421769291162491, 'train/grad_norm_pre_clip': 13.387535095214844, 'timing/data': 0.0009332149929832667, 'timing/model': 1.1918513139826246, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 1.44}) +[2026-06-24 02:08:01,157][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.10566829144954681, 'train/grad_norm_pre_clip': 23.413681030273438, 'timing/data': 0.0008118599944282323, 'timing/model': 1.1740903619793244, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 1.44}) +[2026-06-24 02:08:05,481][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.01199174765497446, 'train/grad_norm_pre_clip': 10.004024505615234, 'timing/data': 0.0007681290153414011, 'timing/model': 1.203956487996038, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 1.44}) +[2026-06-24 02:08:09,793][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.05545312538743019, 'train/grad_norm_pre_clip': 9.015718460083008, 'timing/data': 0.0007346470083575696, 'timing/model': 1.1729146589932498, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 1.44}) +[2026-06-24 02:08:14,115][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.004734279587864876, 'train/grad_norm_pre_clip': 26.628578186035156, 'timing/data': 0.0008621229790151119, 'timing/model': 1.2012212960107718, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 1.44}) +[2026-06-24 02:08:18,650][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.03591746836900711, 'train/grad_norm_pre_clip': 29.04302978515625, 'timing/data': 0.0007665969897061586, 'timing/model': 1.1780177280015778, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 1.44}) +[2026-06-24 02:08:23,027][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.02656923234462738, 'train/grad_norm_pre_clip': 32.4975471496582, 'timing/data': 0.0008961860148701817, 'timing/model': 1.2036369589914102, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 1.44}) +[2026-06-24 02:08:27,375][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.0322556234896183, 'train/grad_norm_pre_clip': 20.438283920288086, 'timing/data': 0.0008047260052990168, 'timing/model': 1.191110056010075, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 1.44}) +[2026-06-24 02:08:31,713][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.03988562896847725, 'train/grad_norm_pre_clip': 14.145489692687988, 'timing/data': 0.0010522340016905218, 'timing/model': 1.1965490140137263, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 1.44}) +[2026-06-24 02:08:36,250][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.0005730384727939963, 'train/grad_norm_pre_clip': 13.861648559570312, 'timing/data': 0.0007576399948447943, 'timing/model': 1.1776534660020843, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 1.44}) +[2026-06-24 02:08:40,588][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.09265879541635513, 'train/grad_norm_pre_clip': 21.932771682739258, 'timing/data': 0.0007266319880727679, 'timing/model': 1.1910235570103396, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 1.44}) +[2026-06-24 02:08:44,902][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.0004758716677315533, 'train/grad_norm_pre_clip': 7.53953218460083, 'timing/data': 0.0008954349905252457, 'timing/model': 1.1711375990125816, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 1.44}) +[2026-06-24 02:08:49,247][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.13717995584011078, 'train/grad_norm_pre_clip': 28.205398559570312, 'timing/data': 0.000935107993427664, 'timing/model': 1.1991966670029797, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 1.44}) +[2026-06-24 02:08:53,795][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.08676060289144516, 'train/grad_norm_pre_clip': 25.46676254272461, 'timing/data': 0.0007387840014416724, 'timing/model': 1.1753449419920798, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 1.44}) +[2026-06-24 02:08:58,167][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.035443004220724106, 'train/grad_norm_pre_clip': 26.62818145751953, 'timing/data': 0.0007553450122941285, 'timing/model': 1.1997838520037476, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 1.44}) +[2026-06-24 02:09:02,532][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.008634144440293312, 'train/grad_norm_pre_clip': 6.126532554626465, 'timing/data': 0.0007594730122946203, 'timing/model': 1.1930785749864299, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 1.44}) +[2026-06-24 02:09:06,912][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.006822244729846716, 'train/grad_norm_pre_clip': 6.177259922027588, 'timing/data': 0.0007355179986916482, 'timing/model': 1.1965629720070865, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 1.44}) +[2026-06-24 02:09:11,479][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.20043420791625977, 'train/grad_norm_pre_clip': 22.780210494995117, 'timing/data': 0.0010864869982469827, 'timing/model': 1.1779102229920682, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 1.44}) +[2026-06-24 02:09:15,832][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.07097640633583069, 'train/grad_norm_pre_clip': 27.15899658203125, 'timing/data': 0.0008889420132618397, 'timing/model': 1.1952427139913198, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 1.44}) +[2026-06-24 02:09:20,161][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.004802716430276632, 'train/grad_norm_pre_clip': 15.403054237365723, 'timing/data': 0.0010993620089720935, 'timing/model': 1.1755731849989388, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 1.44}) +[2026-06-24 02:09:24,518][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.034780003130435944, 'train/grad_norm_pre_clip': 16.248422622680664, 'timing/data': 0.001090814999770373, 'timing/model': 1.2046837469970342, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 1.44}) +[2026-06-24 02:09:29,043][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.09282425791025162, 'train/grad_norm_pre_clip': 29.545711517333984, 'timing/data': 0.0008477660012431443, 'timing/model': 1.2007650670129806, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 1.44}) +[2026-06-24 02:09:33,396][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 4.3413998355390504e-05, 'train/grad_norm_pre_clip': 14.201674461364746, 'timing/data': 0.0007264119922183454, 'timing/model': 1.1786405729944818, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 1.44}) +[2026-06-24 02:09:37,713][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.0023779699113219976, 'train/grad_norm_pre_clip': 39.60563278198242, 'timing/data': 0.0007593030168209225, 'timing/model': 1.1791688689845614, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 1.44}) +[2026-06-24 02:09:42,057][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.03358267992734909, 'train/grad_norm_pre_clip': 9.403266906738281, 'timing/data': 0.0009045809856615961, 'timing/model': 1.1962805229995865, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 1.45}) +[2026-06-24 02:09:46,638][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.033893484622240067, 'train/grad_norm_pre_clip': 24.210935592651367, 'timing/data': 0.0008915470098145306, 'timing/model': 1.4390055680123623, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 1.45}) +[2026-06-24 02:09:50,984][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.016251249238848686, 'train/grad_norm_pre_clip': 25.047456741333008, 'timing/data': 0.000736280984710902, 'timing/model': 1.1974713030213024, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 1.45}) +[2026-06-24 02:09:55,328][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.1616034060716629, 'train/grad_norm_pre_clip': 39.16204071044922, 'timing/data': 0.0007541629893239588, 'timing/model': 1.1991057529812679, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 1.45}) +[2026-06-24 02:09:59,682][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.09890482574701309, 'train/grad_norm_pre_clip': 23.137928009033203, 'timing/data': 0.0007771149976179004, 'timing/model': 1.1800135610101279, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 1.45}) +[2026-06-24 02:10:04,235][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.03291754797101021, 'train/grad_norm_pre_clip': 52.421077728271484, 'timing/data': 0.0009227940172422677, 'timing/model': 1.1911091599904466, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 1.45}) +[2026-06-24 02:10:08,656][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.18161992728710175, 'train/grad_norm_pre_clip': 91.4293212890625, 'timing/data': 0.0008854459738358855, 'timing/model': 1.1762466320069507, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 1.45}) +[2026-06-24 02:10:13,033][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.02993445098400116, 'train/grad_norm_pre_clip': 29.41005516052246, 'timing/data': 0.0007582810067106038, 'timing/model': 1.1775526459969115, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 1.45}) +[2026-06-24 02:10:17,384][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.08079130202531815, 'train/grad_norm_pre_clip': 31.519302368164062, 'timing/data': 0.000915410986635834, 'timing/model': 1.183209025010001, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 1.45}) +[2026-06-24 02:10:21,989][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.07040218263864517, 'train/grad_norm_pre_clip': 35.36624526977539, 'timing/data': 0.0007450970006175339, 'timing/model': 1.1743868369958363, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 1.45}) +[2026-06-24 02:10:26,307][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.004458373878151178, 'train/grad_norm_pre_clip': 23.18622589111328, 'timing/data': 0.0008882010006345809, 'timing/model': 1.1799059539916925, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 1.45}) +[2026-06-24 02:10:30,682][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.004941656719893217, 'train/grad_norm_pre_clip': 6.619206428527832, 'timing/data': 0.0007622379926033318, 'timing/model': 1.2066233329824172, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 1.45}) +[2026-06-24 02:10:35,025][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.0824526771903038, 'train/grad_norm_pre_clip': 21.148115158081055, 'timing/data': 0.0008614120015408844, 'timing/model': 1.1775884120143019, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 1.45}) +[2026-06-24 02:10:39,572][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.06346439570188522, 'train/grad_norm_pre_clip': 12.447059631347656, 'timing/data': 0.0008215779962483793, 'timing/model': 1.1795527920185123, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 1.45}) +[2026-06-24 02:10:43,961][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.00851600430905819, 'train/grad_norm_pre_clip': 17.80571746826172, 'timing/data': 0.0008994109812192619, 'timing/model': 1.197754888999043, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 1.45}) +[2026-06-24 02:10:48,290][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.003708064556121826, 'train/grad_norm_pre_clip': 7.222545146942139, 'timing/data': 0.0007623979763593525, 'timing/model': 1.1746092989924364, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 1.45}) +[2026-06-24 02:33:55,037][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3750 +[2026-06-24 02:33:55,038][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.030664850026369095, 'train/grad_norm_pre_clip': 29.388776779174805, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 362.0600000000205, 'rl_games_eval/mid_train/mean_length': 2935.2, 'rl_games_eval/mid_train/std_reward': 165.08000000001022, 'rl_games_eval/mid_train/std_length': 1329.6000000000001, 'rl_games_eval/mid_train/macro_mean_reward': 362.0600000000205, 'rl_games_eval/mid_train/macro_mean_length': 2935.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 362.0600000000205, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2935.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 165.08000000001022, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1329.6000000000001, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 362.0600000000205, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0008151460206136107, 'timing/model': 1.1751956029911526, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 1.45}) +[2026-06-24 02:33:59,243][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.002715512178838253, 'train/grad_norm_pre_clip': 6.342140197753906, 'timing/data': 0.0008053870114963502, 'timing/model': 1.169793221022701, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 1.45}) +[2026-06-24 02:34:03,679][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.0023511021863669157, 'train/grad_norm_pre_clip': 36.889549255371094, 'timing/data': 0.0012128309754189104, 'timing/model': 1.1523001710011158, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 1.45}) +[2026-06-24 02:34:07,871][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.22217538952827454, 'train/grad_norm_pre_clip': 43.768123626708984, 'timing/data': 0.0008423760009463876, 'timing/model': 1.1370323310256936, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 1.45}) +[2026-06-24 02:34:12,048][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.002220451133325696, 'train/grad_norm_pre_clip': 9.159771919250488, 'timing/data': 0.0007692610088270158, 'timing/model': 1.1578680289967451, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 1.45}) +[2026-06-24 02:34:16,282][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.025442661717534065, 'train/grad_norm_pre_clip': 15.519099235534668, 'timing/data': 0.0012291999883018434, 'timing/model': 1.1568101870070677, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 1.45}) +[2026-06-24 02:34:20,825][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.030071040615439415, 'train/grad_norm_pre_clip': 9.20363998413086, 'timing/data': 0.0007491240103263408, 'timing/model': 1.1659989329928067, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 1.45}) +[2026-06-24 02:34:25,079][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.10769374668598175, 'train/grad_norm_pre_clip': 47.36500549316406, 'timing/data': 0.0008074719808064401, 'timing/model': 1.1617351890017744, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 1.45}) +[2026-06-24 02:34:29,315][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.020878713577985764, 'train/grad_norm_pre_clip': 15.319428443908691, 'timing/data': 0.0007293870148714632, 'timing/model': 1.1659746719815303, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 1.45}) +[2026-06-24 02:34:33,574][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.11744273453950882, 'train/grad_norm_pre_clip': 45.42271041870117, 'timing/data': 0.0007523199892602861, 'timing/model': 1.1798524399928283, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 1.45}) +[2026-06-24 02:34:38,074][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.09244517982006073, 'train/grad_norm_pre_clip': 45.593658447265625, 'timing/data': 0.0011866719869431108, 'timing/model': 1.151352576998761, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 1.46}) +[2026-06-24 02:34:42,402][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.045046791434288025, 'train/grad_norm_pre_clip': 22.479816436767578, 'timing/data': 0.0008978479891084135, 'timing/model': 1.2175908729841467, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 1.46}) +[2026-06-24 02:34:46,683][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.0946139544248581, 'train/grad_norm_pre_clip': 34.24683380126953, 'timing/data': 0.0008318060135934502, 'timing/model': 1.179926893004449, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 1.46}) +[2026-06-24 02:34:51,128][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.011477030813694, 'train/grad_norm_pre_clip': 10.0048828125, 'timing/data': 0.0011894269846379757, 'timing/model': 1.3572205590025987, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 1.46}) +[2026-06-24 02:34:55,402][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.016354244202375412, 'train/grad_norm_pre_clip': 18.391340255737305, 'timing/data': 0.0009078470175154507, 'timing/model': 1.1762605250114575, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 1.46}) +[2026-06-24 02:34:59,662][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.20375585556030273, 'train/grad_norm_pre_clip': 27.211843490600586, 'timing/data': 0.0008215779962483793, 'timing/model': 1.1815921160159633, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 1.46}) +[2026-06-24 02:35:03,937][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.10274165868759155, 'train/grad_norm_pre_clip': 41.67021560668945, 'timing/data': 0.0007146499992813915, 'timing/model': 1.172960002993932, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 1.46}) +[2026-06-24 02:35:08,198][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.0013140032533556223, 'train/grad_norm_pre_clip': 28.235689163208008, 'timing/data': 0.0009982140036299825, 'timing/model': 1.1804788330045994, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 1.46}) +[2026-06-24 02:35:12,697][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.032159481197595596, 'train/grad_norm_pre_clip': 23.963706970214844, 'timing/data': 0.0009541530162096024, 'timing/model': 1.167272821010556, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 1.46}) +[2026-06-24 02:35:16,999][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.005340085364878178, 'train/grad_norm_pre_clip': 24.079769134521484, 'timing/data': 0.0014096939994487911, 'timing/model': 1.1648900620057248, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 1.46}) +[2026-06-24 02:35:21,319][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.009471057914197445, 'train/grad_norm_pre_clip': 10.465507507324219, 'timing/data': 0.0008922580163925886, 'timing/model': 1.1878348070022184, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 1.46}) +[2026-06-24 02:35:25,632][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.011736741289496422, 'train/grad_norm_pre_clip': 6.847435474395752, 'timing/data': 0.001274974987609312, 'timing/model': 1.1829456280102022, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 1.46}) +[2026-06-24 02:35:30,156][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.01505554560571909, 'train/grad_norm_pre_clip': 5.827544212341309, 'timing/data': 0.0007323829922825098, 'timing/model': 1.1686315090046264, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 1.46}) +[2026-06-24 02:35:34,501][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.007750908378511667, 'train/grad_norm_pre_clip': 8.103687286376953, 'timing/data': 0.0008150649955496192, 'timing/model': 1.1955160819925368, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 1.46}) +[2026-06-24 02:35:38,840][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.02277260646224022, 'train/grad_norm_pre_clip': 16.331270217895508, 'timing/data': 0.0006994220020715147, 'timing/model': 1.1603721560095437, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 1.46}) +[2026-06-24 02:35:43,182][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.005860840901732445, 'train/grad_norm_pre_clip': 5.7793684005737305, 'timing/data': 0.0007995659834705293, 'timing/model': 1.1848088560218457, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 1.46}) +[2026-06-24 02:35:47,657][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.0047504170797765255, 'train/grad_norm_pre_clip': 14.788222312927246, 'timing/data': 0.000684194004861638, 'timing/model': 1.1668473610188812, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 1.46}) +[2026-06-24 02:35:51,951][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 7.538333011325449e-05, 'train/grad_norm_pre_clip': 18.63978385925293, 'timing/data': 0.0009294569899793714, 'timing/model': 1.1662093240011018, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 1.46}) +[2026-06-24 02:35:56,283][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.043077725917100906, 'train/grad_norm_pre_clip': 11.683059692382812, 'timing/data': 0.0009679379872977734, 'timing/model': 1.1812129899917636, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 1.46}) +[2026-06-24 02:36:00,622][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.0001944254181580618, 'train/grad_norm_pre_clip': 20.498939514160156, 'timing/data': 0.0010660889965947717, 'timing/model': 1.1915661759849172, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 1.46}) +[2026-06-24 02:36:05,115][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.016239548102021217, 'train/grad_norm_pre_clip': 14.035261154174805, 'timing/data': 0.0010891720012295991, 'timing/model': 1.1746411640197039, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 1.46}) +[2026-06-24 02:36:09,472][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.0010075672762468457, 'train/grad_norm_pre_clip': 14.117764472961426, 'timing/data': 0.0007760430162306875, 'timing/model': 1.188618725980632, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 1.46}) +[2026-06-24 02:36:13,824][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.1358984410762787, 'train/grad_norm_pre_clip': 27.693397521972656, 'timing/data': 0.0008048570016399026, 'timing/model': 1.1930524170165882, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 1.46}) +[2026-06-24 02:36:18,150][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.019050277769565582, 'train/grad_norm_pre_clip': 22.24311065673828, 'timing/data': 0.0008917580125853419, 'timing/model': 1.1984476659854408, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 1.46}) +[2026-06-24 02:36:22,757][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.16589319705963135, 'train/grad_norm_pre_clip': 33.006744384765625, 'timing/data': 0.0012448699853848666, 'timing/model': 1.2066752149839886, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 1.46}) +[2026-06-24 02:36:27,097][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.022971484810113907, 'train/grad_norm_pre_clip': 14.296838760375977, 'timing/data': 0.0010221789998468012, 'timing/model': 1.1752257129992358, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 1.46}) +[2026-06-24 02:36:31,415][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.08018377423286438, 'train/grad_norm_pre_clip': 30.40867042541504, 'timing/data': 0.0008153159869834781, 'timing/model': 1.172452822997002, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 1.47}) +[2026-06-24 02:36:35,729][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.0959821417927742, 'train/grad_norm_pre_clip': 21.207408905029297, 'timing/data': 0.0007410689722746611, 'timing/model': 1.17289527200046, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 1.47}) +[2026-06-24 02:36:40,303][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.003134397091343999, 'train/grad_norm_pre_clip': 24.120363235473633, 'timing/data': 0.0007341260206885636, 'timing/model': 1.1697617850149982, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 1.47}) +[2026-06-24 02:36:44,613][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.021995102986693382, 'train/grad_norm_pre_clip': 29.35720443725586, 'timing/data': 0.0007183570123743266, 'timing/model': 1.1913488229911309, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 1.47}) +[2026-06-24 02:36:48,960][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.013319042511284351, 'train/grad_norm_pre_clip': 29.9553165435791, 'timing/data': 0.0008682339976076037, 'timing/model': 1.174284714012174, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 1.47}) +[2026-06-24 02:36:53,463][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.006906614173203707, 'train/grad_norm_pre_clip': 11.176759719848633, 'timing/data': 0.0011896270152647048, 'timing/model': 1.200311883003451, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 1.47}) +[2026-06-24 02:36:58,018][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.04767659679055214, 'train/grad_norm_pre_clip': 26.98278045654297, 'timing/data': 0.0007033990113995969, 'timing/model': 1.1836941590008792, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 1.47}) +[2026-06-24 02:37:02,404][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.0537271723151207, 'train/grad_norm_pre_clip': 24.209020614624023, 'timing/data': 0.0008857560169417411, 'timing/model': 1.1940312359947711, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 1.47}) +[2026-06-24 02:37:06,708][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.04343421757221222, 'train/grad_norm_pre_clip': 21.220117568969727, 'timing/data': 0.0008917380182538182, 'timing/model': 1.1752953170216642, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 1.47}) +[2026-06-24 02:37:11,013][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.00034032121766358614, 'train/grad_norm_pre_clip': 34.750301361083984, 'timing/data': 0.0007101119845174253, 'timing/model': 1.1736222979961894, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 1.47}) +[2026-06-24 02:37:15,626][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.1085527241230011, 'train/grad_norm_pre_clip': 26.011127471923828, 'timing/data': 0.0007150509918574244, 'timing/model': 1.2002254929975607, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 1.47}) +[2026-06-24 02:37:19,986][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.01051775086671114, 'train/grad_norm_pre_clip': 15.145868301391602, 'timing/data': 0.0011505750007927418, 'timing/model': 1.1951886000169907, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 1.47}) +[2026-06-24 02:37:24,329][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.07783934473991394, 'train/grad_norm_pre_clip': 23.68999671936035, 'timing/data': 0.0010528750135563314, 'timing/model': 1.174795800994616, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 1.47}) +[2026-06-24 02:37:28,626][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.014559502713382244, 'train/grad_norm_pre_clip': 24.29437828063965, 'timing/data': 0.0008597479900345206, 'timing/model': 1.1728655869956128, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 1.47}) +[2026-06-24 02:37:33,231][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3800 +[2026-06-24 02:38:30,576][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3800 +[2026-06-24 02:38:30,576][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3800 +[2026-06-24 02:39:07,675][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3800 +[2026-06-24 02:39:07,677][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.03288216516375542, 'train/grad_norm_pre_clip': 18.256851196289062, 'eval/action_loss/seconds': 57.34151407299214, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03298976272344589, 'eval/latency_2/loss': 0.03298976272344589, 'eval/flappy/loss': 0.03298976272344589, 'eval/flappy/latency_2/loss': 0.03298976272344589, 'eval/flap_precision': 0.977027027027027, 'eval/flap_recall': 0.9810040705563093, 'eval/flap_f1': 0.9790115098171971, 'eval/flap_rate_model': 0.20555555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9790115098171971, 'eval/latency_2/flap_precision': 0.977027027027027, 'eval/latency_2/flap_recall': 0.9810040705563093, 'eval/latency_2/flap_f1': 0.9790115098171971, 'eval/latency_2/flap_rate_model': 0.20555555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9790115098171971, 'eval/action_classification/data_wait_seconds': 0.38655778765678406, 'eval/action_classification/predict_seconds': 36.56913757324219, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007422909839078784, 'timing/model': 1.1786886549962219, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 1.47}) +[2026-06-24 02:39:11,991][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.09283164143562317, 'train/grad_norm_pre_clip': 32.200748443603516, 'timing/data': 0.000752610998461023, 'timing/model': 1.1657548540097196, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 1.47}) +[2026-06-24 02:39:16,329][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.022546015679836273, 'train/grad_norm_pre_clip': 14.691162109375, 'timing/data': 0.0007327139901462942, 'timing/model': 1.200693727005273, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 1.47}) +[2026-06-24 02:39:20,874][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.09148288518190384, 'train/grad_norm_pre_clip': 18.659997940063477, 'timing/data': 0.0009751619945745915, 'timing/model': 1.4354826050112024, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 1.47}) +[2026-06-24 02:39:25,210][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.025049369782209396, 'train/grad_norm_pre_clip': 11.896379470825195, 'timing/data': 0.0008158670098055154, 'timing/model': 1.1810364380071405, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 1.47}) +[2026-06-24 02:39:29,544][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.01974424161016941, 'train/grad_norm_pre_clip': 20.09958267211914, 'timing/data': 0.0007419909816235304, 'timing/model': 1.184879987995373, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 1.47}) +[2026-06-24 02:39:33,851][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.19206535816192627, 'train/grad_norm_pre_clip': 23.0251522064209, 'timing/data': 0.0009141990158241242, 'timing/model': 1.1724732009752188, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 1.47}) +[2026-06-24 02:39:38,370][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.002345846965909004, 'train/grad_norm_pre_clip': 6.634838104248047, 'timing/data': 0.0007254499942064285, 'timing/model': 1.3774139930028468, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 1.47}) +[2026-06-24 02:39:42,669][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.009779812768101692, 'train/grad_norm_pre_clip': 16.329648971557617, 'timing/data': 0.001118386018788442, 'timing/model': 1.1897111170110293, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 1.47}) +[2026-06-24 02:39:46,964][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.2705358862876892, 'train/grad_norm_pre_clip': 26.324506759643555, 'timing/data': 0.0008025130082387477, 'timing/model': 1.1730035439832136, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 1.47}) +[2026-06-24 02:39:51,314][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.002444655867293477, 'train/grad_norm_pre_clip': 10.591414451599121, 'timing/data': 0.0007631500193383545, 'timing/model': 1.1890707559941802, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 1.47}) +[2026-06-24 02:39:55,844][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.006622631102800369, 'train/grad_norm_pre_clip': 25.311229705810547, 'timing/data': 0.0009645920072216541, 'timing/model': 1.3692502979829442, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 1.47}) +[2026-06-24 02:40:00,188][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.022062210366129875, 'train/grad_norm_pre_clip': 13.357226371765137, 'timing/data': 0.001164231012808159, 'timing/model': 1.1937716419924982, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 1.48}) +[2026-06-24 02:40:04,531][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.00877675972878933, 'train/grad_norm_pre_clip': 8.71363353729248, 'timing/data': 0.0010908749827649444, 'timing/model': 1.1951555199921131, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 1.48}) +[2026-06-24 02:40:08,870][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.06282374262809753, 'train/grad_norm_pre_clip': 14.713839530944824, 'timing/data': 0.001198754005599767, 'timing/model': 1.1906452990078833, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 1.48}) +[2026-06-24 02:40:13,407][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.08907396346330643, 'train/grad_norm_pre_clip': 30.928085327148438, 'timing/data': 0.0007656450034119189, 'timing/model': 1.3910491220012773, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 1.48}) +[2026-06-24 02:40:17,768][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.046147532761096954, 'train/grad_norm_pre_clip': 13.69970703125, 'timing/data': 0.0007945279940031469, 'timing/model': 1.1926730169798248, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 1.48}) +[2026-06-24 02:40:22,105][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.11441174894571304, 'train/grad_norm_pre_clip': 32.822601318359375, 'timing/data': 0.0007307000050786883, 'timing/model': 1.1742426929995418, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 1.48}) +[2026-06-24 02:40:26,444][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.049942076206207275, 'train/grad_norm_pre_clip': 26.733135223388672, 'timing/data': 0.0007907109975349158, 'timing/model': 1.197773521009367, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 1.48}) +[2026-06-24 02:40:30,946][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.1147555336356163, 'train/grad_norm_pre_clip': 25.363710403442383, 'timing/data': 0.0013849890092387795, 'timing/model': 1.3650080510124099, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 1.48}) +[2026-06-24 02:40:35,260][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.09646549820899963, 'train/grad_norm_pre_clip': 21.900663375854492, 'timing/data': 0.000782095012255013, 'timing/model': 1.1720533210027497, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 1.48}) +[2026-06-24 02:40:39,580][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.0406470000743866, 'train/grad_norm_pre_clip': 19.18943977355957, 'timing/data': 0.001090664998628199, 'timing/model': 1.1916684469906613, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 1.48}) +[2026-06-24 02:40:43,967][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.0050780558958649635, 'train/grad_norm_pre_clip': 19.502464294433594, 'timing/data': 0.0010337700077798218, 'timing/model': 1.201671446004184, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 1.48}) +[2026-06-24 02:40:48,329][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.07268506288528442, 'train/grad_norm_pre_clip': 13.771617889404297, 'timing/data': 0.00096791802207008, 'timing/model': 1.1975946169986855, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 1.48}) +[2026-06-24 02:40:52,963][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.00794356968253851, 'train/grad_norm_pre_clip': 14.326921463012695, 'timing/data': 0.0008728529792279005, 'timing/model': 1.20676224300405, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 1.48}) +[2026-06-24 02:40:57,334][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.00196445663459599, 'train/grad_norm_pre_clip': 16.1807918548584, 'timing/data': 0.0009461680019740015, 'timing/model': 1.2023836160078645, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 1.48}) +[2026-06-24 02:41:01,699][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.08042710274457932, 'train/grad_norm_pre_clip': 12.914575576782227, 'timing/data': 0.0007308100175578147, 'timing/model': 1.1940672030032147, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 1.48}) +[2026-06-24 02:41:06,036][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.004610881209373474, 'train/grad_norm_pre_clip': 8.36405086517334, 'timing/data': 0.0009440639987587929, 'timing/model': 1.176724107004702, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 1.48}) +[2026-06-24 02:41:10,628][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.0988401472568512, 'train/grad_norm_pre_clip': 17.555728912353516, 'timing/data': 0.0007461189816240221, 'timing/model': 1.1748213930113707, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 1.48}) +[2026-06-24 02:41:14,936][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.0013313834788277745, 'train/grad_norm_pre_clip': 24.16573143005371, 'timing/data': 0.0008371859730686992, 'timing/model': 1.177854606998153, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 1.48}) +[2026-06-24 02:41:19,279][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.003531316528096795, 'train/grad_norm_pre_clip': 13.185982704162598, 'timing/data': 0.0007900090131442994, 'timing/model': 1.177589479018934, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 1.48}) +[2026-06-24 02:41:23,622][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.0012419559061527252, 'train/grad_norm_pre_clip': 10.432611465454102, 'timing/data': 0.000784820003900677, 'timing/model': 1.209953976009274, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 1.48}) +[2026-06-24 02:41:28,250][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.02497866563498974, 'train/grad_norm_pre_clip': 8.750884056091309, 'timing/data': 0.0009597729949746281, 'timing/model': 1.1930367760069203, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 1.48}) +[2026-06-24 02:41:32,625][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.004598589148372412, 'train/grad_norm_pre_clip': 5.186339855194092, 'timing/data': 0.0008995919779408723, 'timing/model': 1.1953411040012725, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 1.48}) +[2026-06-24 02:41:37,008][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.05063091591000557, 'train/grad_norm_pre_clip': 27.80306625366211, 'timing/data': 0.0007432530110236257, 'timing/model': 1.2042097979865503, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 1.48}) +[2026-06-24 02:41:41,391][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.0007229415932670236, 'train/grad_norm_pre_clip': 11.568225860595703, 'timing/data': 0.0007569380104541779, 'timing/model': 1.1987888650037348, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 1.48}) +[2026-06-24 02:41:46,005][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.008139243349432945, 'train/grad_norm_pre_clip': 30.158390045166016, 'timing/data': 0.0007578700024168938, 'timing/model': 1.197643599996809, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 1.48}) +[2026-06-24 02:41:50,372][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.025731349363923073, 'train/grad_norm_pre_clip': 17.339574813842773, 'timing/data': 0.0007273230003193021, 'timing/model': 1.2016700899985153, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 1.48}) +[2026-06-24 02:41:54,712][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.016050711274147034, 'train/grad_norm_pre_clip': 15.344849586486816, 'timing/data': 0.000773909006966278, 'timing/model': 1.2097963319974951, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 1.49}) +[2026-06-24 02:41:59,272][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.0878080427646637, 'train/grad_norm_pre_clip': 33.028690338134766, 'timing/data': 0.0008153759990818799, 'timing/model': 1.3806063699885271, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 1.49}) +[2026-06-24 02:42:03,639][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.1895511895418167, 'train/grad_norm_pre_clip': 23.489370346069336, 'timing/data': 0.0007985649863258004, 'timing/model': 1.2013343420112506, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 1.49}) +[2026-06-24 02:42:07,983][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.0005974415107630193, 'train/grad_norm_pre_clip': 8.672247886657715, 'timing/data': 0.0008536969835404307, 'timing/model': 1.1924769100151025, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 1.49}) +[2026-06-24 02:42:12,333][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.025087211281061172, 'train/grad_norm_pre_clip': 21.364330291748047, 'timing/data': 0.0008901249966584146, 'timing/model': 1.1774831539951265, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 1.49}) +[2026-06-24 02:42:16,865][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.0012916577979922295, 'train/grad_norm_pre_clip': 12.38404655456543, 'timing/data': 0.0009924540063366294, 'timing/model': 1.3730801200144924, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 1.49}) +[2026-06-24 02:42:21,202][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.03720809891819954, 'train/grad_norm_pre_clip': 24.976350784301758, 'timing/data': 0.0009475009865127504, 'timing/model': 1.1825913320062682, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 1.49}) +[2026-06-24 02:42:25,521][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.0048654889687895775, 'train/grad_norm_pre_clip': 16.526880264282227, 'timing/data': 0.000804736017016694, 'timing/model': 1.1766284569748677, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 1.49}) +[2026-06-24 02:42:29,916][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.03513071686029434, 'train/grad_norm_pre_clip': 21.4703369140625, 'timing/data': 0.0008107270114123821, 'timing/model': 1.1967841709847562, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 1.49}) +[2026-06-24 02:42:34,247][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.13053159415721893, 'train/grad_norm_pre_clip': 20.9636173248291, 'timing/data': 0.0009330230241175741, 'timing/model': 1.1886816750047728, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 1.49}) +[2026-06-24 02:42:38,840][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.02201039530336857, 'train/grad_norm_pre_clip': 15.684587478637695, 'timing/data': 0.000869786978000775, 'timing/model': 1.1982067610078957, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 1.49}) +[2026-06-24 02:42:43,241][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.0042387619614601135, 'train/grad_norm_pre_clip': 30.924543380737305, 'timing/data': 0.000866841000970453, 'timing/model': 1.20412391301943, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 1.49}) +[2026-06-24 02:42:47,580][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.05466453358530998, 'train/grad_norm_pre_clip': 19.618371963500977, 'timing/data': 0.0009285449923481792, 'timing/model': 1.180106715997681, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 1.49}) +[2026-06-24 02:42:51,946][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.01915760338306427, 'train/grad_norm_pre_clip': 14.280752182006836, 'timing/data': 0.0007170249882619828, 'timing/model': 1.179387279000366, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 1.49}) +[2026-06-24 02:42:56,577][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.015329333953559399, 'train/grad_norm_pre_clip': 22.24686050415039, 'timing/data': 0.0008012109901756048, 'timing/model': 1.1798082570021506, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 1.49}) +[2026-06-24 02:43:00,964][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.10670052468776703, 'train/grad_norm_pre_clip': 34.11524963378906, 'timing/data': 0.000768900994444266, 'timing/model': 1.1861440619977657, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 1.49}) +[2026-06-24 02:43:05,316][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.024107616394758224, 'train/grad_norm_pre_clip': 23.35841941833496, 'timing/data': 0.0007161929970607162, 'timing/model': 1.189525108988164, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 1.49}) +[2026-06-24 02:43:09,844][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.015790345147252083, 'train/grad_norm_pre_clip': 40.80274963378906, 'timing/data': 0.0007318310090340674, 'timing/model': 1.3581536780111492, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 1.49}) +[2026-06-24 02:43:14,234][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.047167256474494934, 'train/grad_norm_pre_clip': 30.8739070892334, 'timing/data': 0.0007949789869599044, 'timing/model': 1.203849982994143, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 1.49}) +[2026-06-24 02:43:18,586][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.00198151683434844, 'train/grad_norm_pre_clip': 18.445093154907227, 'timing/data': 0.0007534119940828532, 'timing/model': 1.2020809949899558, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 1.49}) +[2026-06-24 02:43:22,947][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.05698968842625618, 'train/grad_norm_pre_clip': 9.78845500946045, 'timing/data': 0.0009108019876293838, 'timing/model': 1.2024888100277167, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 1.49}) +[2026-06-24 02:43:27,263][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.08186329901218414, 'train/grad_norm_pre_clip': 18.977752685546875, 'timing/data': 0.0007771160162519664, 'timing/model': 1.1778575290227309, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 1.49}) +[2026-06-24 02:43:31,836][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.011173991486430168, 'train/grad_norm_pre_clip': 10.039275169372559, 'timing/data': 0.0009471589874010533, 'timing/model': 1.1793168360018171, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 1.49}) +[2026-06-24 02:43:36,168][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.0032105231657624245, 'train/grad_norm_pre_clip': 40.068580627441406, 'timing/data': 0.0007456069870386273, 'timing/model': 1.2004949800029863, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 1.49}) +[2026-06-24 02:43:40,567][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.02211174927651882, 'train/grad_norm_pre_clip': 13.651163101196289, 'timing/data': 0.001051953004207462, 'timing/model': 1.2110728310071863, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 1.49}) +[2026-06-24 02:43:44,902][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.007883410900831223, 'train/grad_norm_pre_clip': 11.608731269836426, 'timing/data': 0.0007337050046771765, 'timing/model': 1.172716484987177, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 1.5}) +[2026-06-24 02:43:49,459][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.0380692183971405, 'train/grad_norm_pre_clip': 27.054595947265625, 'timing/data': 0.0007738190179225057, 'timing/model': 1.1981500599940773, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 1.5}) +[2026-06-24 02:43:53,804][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.016289616003632545, 'train/grad_norm_pre_clip': 11.816744804382324, 'timing/data': 0.00103113500517793, 'timing/model': 1.1780485670024063, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 1.5}) +[2026-06-24 02:43:58,165][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.2802744507789612, 'train/grad_norm_pre_clip': 42.130767822265625, 'timing/data': 0.0008735840092413127, 'timing/model': 1.1986928239930421, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 1.5}) +[2026-06-24 02:44:02,483][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.08734866976737976, 'train/grad_norm_pre_clip': 35.85324478149414, 'timing/data': 0.0009280540107283741, 'timing/model': 1.1810280719946604, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 1.5}) +[2026-06-24 02:44:07,080][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.0727493092417717, 'train/grad_norm_pre_clip': 25.63077163696289, 'timing/data': 0.0008128409972414374, 'timing/model': 1.1786863270099275, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 1.5}) +[2026-06-24 02:44:11,466][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.06354300677776337, 'train/grad_norm_pre_clip': 29.13044548034668, 'timing/data': 0.0012226190010551363, 'timing/model': 1.1855634920066223, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 1.5}) +[2026-06-24 02:44:15,825][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.009466643445193768, 'train/grad_norm_pre_clip': 15.945540428161621, 'timing/data': 0.0007556460041087121, 'timing/model': 1.1815844720113091, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 1.5}) +[2026-06-24 02:44:20,157][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.02407786436378956, 'train/grad_norm_pre_clip': 11.208584785461426, 'timing/data': 0.001075526000931859, 'timing/model': 1.183751636010129, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 1.5}) +[2026-06-24 02:44:24,740][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.0014242364559322596, 'train/grad_norm_pre_clip': 17.915985107421875, 'timing/data': 0.0007638510141987354, 'timing/model': 1.180875844991533, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 1.5}) +[2026-06-24 02:44:29,100][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.1434987187385559, 'train/grad_norm_pre_clip': 21.77292823791504, 'timing/data': 0.0007935660250950605, 'timing/model': 1.2070714269939344, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 1.5}) +[2026-06-24 02:44:33,468][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.03266412764787674, 'train/grad_norm_pre_clip': 20.233503341674805, 'timing/data': 0.0007239370024763048, 'timing/model': 1.2002306100039277, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 1.5}) +[2026-06-24 02:44:37,800][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.08283936977386475, 'train/grad_norm_pre_clip': 19.92652702331543, 'timing/data': 0.0007453360012732446, 'timing/model': 1.1911316920013633, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 1.5}) +[2026-06-24 02:44:42,371][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.051859352737665176, 'train/grad_norm_pre_clip': 12.507346153259277, 'timing/data': 0.0007424520154017955, 'timing/model': 1.1778322150057647, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 1.5}) +[2026-06-24 02:44:46,690][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.004017355386167765, 'train/grad_norm_pre_clip': 23.543193817138672, 'timing/data': 0.0008162879967130721, 'timing/model': 1.1764388809970114, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 1.5}) +[2026-06-24 02:44:51,057][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.03615434467792511, 'train/grad_norm_pre_clip': 9.72330379486084, 'timing/data': 0.0007640420226380229, 'timing/model': 1.1957576900022104, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 1.5}) +[2026-06-24 02:44:55,460][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.031199226155877113, 'train/grad_norm_pre_clip': 25.216083526611328, 'timing/data': 0.0008595680119469762, 'timing/model': 1.207604135008296, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 1.5}) +[2026-06-24 02:45:00,057][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.05155741423368454, 'train/grad_norm_pre_clip': 13.959866523742676, 'timing/data': 0.0007209919858723879, 'timing/model': 1.1784270360076334, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 1.5}) +[2026-06-24 02:45:04,396][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.007554768584668636, 'train/grad_norm_pre_clip': 20.260021209716797, 'timing/data': 0.001063344010617584, 'timing/model': 1.1777624690148514, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 1.5}) +[2026-06-24 02:45:08,775][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.04003143310546875, 'train/grad_norm_pre_clip': 37.59623718261719, 'timing/data': 0.0007422409835271537, 'timing/model': 1.193649988999823, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 1.5}) +[2026-06-24 02:45:13,140][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.0538661926984787, 'train/grad_norm_pre_clip': 16.619976043701172, 'timing/data': 0.0009438039851374924, 'timing/model': 1.2063324589980766, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 1.5}) +[2026-06-24 02:45:17,767][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.003712683916091919, 'train/grad_norm_pre_clip': 9.223685264587402, 'timing/data': 0.0007752520032227039, 'timing/model': 1.1825535509851761, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 1.5}) +[2026-06-24 02:45:22,116][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.0596301406621933, 'train/grad_norm_pre_clip': 15.78922176361084, 'timing/data': 0.0009081779862754047, 'timing/model': 1.1783890899969265, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 1.5}) +[2026-06-24 02:45:26,521][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.026278061792254448, 'train/grad_norm_pre_clip': 19.474990844726562, 'timing/data': 0.0010860869952011853, 'timing/model': 1.2115527240093797, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 1.5}) +[2026-06-24 02:45:30,856][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.037719979882240295, 'train/grad_norm_pre_clip': 15.041756629943848, 'timing/data': 0.0007235159864649177, 'timing/model': 1.1821778700104915, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 1.5}) +[2026-06-24 02:45:35,476][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.03242453187704086, 'train/grad_norm_pre_clip': 10.513148307800293, 'timing/data': 0.0008973280200734735, 'timing/model': 1.1922372730041388, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 1.5}) +[2026-06-24 02:45:39,821][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.0104008624330163, 'train/grad_norm_pre_clip': 30.619388580322266, 'timing/data': 0.0007374619890470058, 'timing/model': 1.1919797180162277, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 1.51}) +[2026-06-24 02:45:44,189][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.04947986826300621, 'train/grad_norm_pre_clip': 18.957618713378906, 'timing/data': 0.00073251299909316, 'timing/model': 1.199306594004156, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 1.51}) +[2026-06-24 02:45:48,575][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.17375770211219788, 'train/grad_norm_pre_clip': 33.151283264160156, 'timing/data': 0.0007309000066015869, 'timing/model': 1.2009284810046665, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 1.51}) +[2026-06-24 02:45:53,204][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.13556845486164093, 'train/grad_norm_pre_clip': 23.41790771484375, 'timing/data': 0.0009103019838221371, 'timing/model': 1.2056084369833115, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 1.51}) +[2026-06-24 02:45:57,551][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.019760500639677048, 'train/grad_norm_pre_clip': 10.722270965576172, 'timing/data': 0.0007128670113161206, 'timing/model': 1.2015859259990975, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 1.51}) +[2026-06-24 02:46:01,872][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.04905538633465767, 'train/grad_norm_pre_clip': 15.750568389892578, 'timing/data': 0.0007109139987733215, 'timing/model': 1.1737365699955262, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 1.51}) +[2026-06-24 02:46:06,212][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.02849791757762432, 'train/grad_norm_pre_clip': 25.578031539916992, 'timing/data': 0.000903528998605907, 'timing/model': 1.1798422390129417, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 1.51}) +[2026-06-24 02:46:10,809][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.03408455848693848, 'train/grad_norm_pre_clip': 40.369083404541016, 'timing/data': 0.0007751630037091672, 'timing/model': 1.1828402990067843, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 1.51}) +[2026-06-24 02:46:15,169][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.10827063769102097, 'train/grad_norm_pre_clip': 32.15349197387695, 'timing/data': 0.000744654971640557, 'timing/model': 1.2012480739795137, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 1.51}) +[2026-06-24 02:46:19,512][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.0632769837975502, 'train/grad_norm_pre_clip': 20.110986709594727, 'timing/data': 0.0009073970140889287, 'timing/model': 1.1820154859742615, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 1.51}) +[2026-06-24 02:46:23,867][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.07823511958122253, 'train/grad_norm_pre_clip': 17.44378089904785, 'timing/data': 0.0010403319902252406, 'timing/model': 1.1955976269964594, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 1.51}) +[2026-06-24 02:46:28,492][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3900 +[2026-06-24 02:47:26,849][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3900 +[2026-06-24 02:47:26,850][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3900 +[2026-06-24 02:48:03,698][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3900 +[2026-06-24 02:48:03,700][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.010107331909239292, 'train/grad_norm_pre_clip': 15.853951454162598, 'eval/action_loss/seconds': 58.35393905098317, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.031353969126939774, 'eval/latency_2/loss': 0.031353969126939774, 'eval/flappy/loss': 0.031353969126939774, 'eval/flappy/latency_2/loss': 0.031353969126939774, 'eval/flap_precision': 0.977088948787062, 'eval/flap_recall': 0.9837177747625508, 'eval/flap_f1': 0.9803921568627451, 'eval/flap_rate_model': 0.2061111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9803921568627451, 'eval/latency_2/flap_precision': 0.977088948787062, 'eval/latency_2/flap_recall': 0.9837177747625508, 'eval/latency_2/flap_f1': 0.9803921568627451, 'eval/latency_2/flap_rate_model': 0.2061111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9803921568627451, 'eval/action_classification/data_wait_seconds': 0.4260096549987793, 'eval/action_classification/predict_seconds': 36.27813720703125, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007661950075998902, 'timing/model': 1.1985197849862743, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 1.51}) +[2026-06-24 02:48:08,032][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.014180617406964302, 'train/grad_norm_pre_clip': 29.261049270629883, 'timing/data': 0.0009187670075334609, 'timing/model': 1.179261259996565, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 1.51}) +[2026-06-24 02:48:12,305][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.05450786277651787, 'train/grad_norm_pre_clip': 9.647968292236328, 'timing/data': 0.0009117439913097769, 'timing/model': 1.1667646449932363, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 1.51}) +[2026-06-24 02:48:16,796][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.0005447002477012575, 'train/grad_norm_pre_clip': 9.977834701538086, 'timing/data': 0.0007613460184074938, 'timing/model': 1.165476474998286, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 1.51}) +[2026-06-24 02:48:21,100][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.10190904140472412, 'train/grad_norm_pre_clip': 13.98311996459961, 'timing/data': 0.0007434629951603711, 'timing/model': 1.171035495994147, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 1.51}) +[2026-06-24 02:48:25,466][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.0031585460528731346, 'train/grad_norm_pre_clip': 22.244436264038086, 'timing/data': 0.0009202799992635846, 'timing/model': 1.198693068989087, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 1.51}) +[2026-06-24 02:48:29,804][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.008778383024036884, 'train/grad_norm_pre_clip': 17.577892303466797, 'timing/data': 0.0008674430137034506, 'timing/model': 1.1951611450058408, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 1.51}) +[2026-06-24 02:48:34,342][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.06964389234781265, 'train/grad_norm_pre_clip': 20.137910842895508, 'timing/data': 0.000800328009063378, 'timing/model': 1.1671763369813561, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 1.51}) +[2026-06-24 02:48:38,696][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.04043078050017357, 'train/grad_norm_pre_clip': 36.62043380737305, 'timing/data': 0.0009962899785023183, 'timing/model': 1.1925958060019184, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 1.51}) +[2026-06-24 02:48:43,041][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.0278902817517519, 'train/grad_norm_pre_clip': 24.14735984802246, 'timing/data': 0.000896434998139739, 'timing/model': 1.1764459530240856, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 1.51}) +[2026-06-24 02:48:47,553][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.13773901760578156, 'train/grad_norm_pre_clip': 28.485641479492188, 'timing/data': 0.0007978539797477424, 'timing/model': 1.3604143669945188, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 1.51}) +[2026-06-24 02:48:51,832][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.0026200213469564915, 'train/grad_norm_pre_clip': 20.428194046020508, 'timing/data': 0.0009295470081269741, 'timing/model': 1.1687213830009568, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 1.51}) +[2026-06-24 02:48:56,132][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.012368366122245789, 'train/grad_norm_pre_clip': 30.132526397705078, 'timing/data': 0.0007346979982685298, 'timing/model': 1.1893005710153375, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 1.51}) +[2026-06-24 02:49:00,416][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.00440894765779376, 'train/grad_norm_pre_clip': 15.733521461486816, 'timing/data': 0.0007040710188448429, 'timing/model': 1.1706846820015926, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 1.51}) +[2026-06-24 02:49:04,851][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.027917593717575073, 'train/grad_norm_pre_clip': 18.515893936157227, 'timing/data': 0.0009385040029883385, 'timing/model': 1.32329977399786, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 1.51}) +[2026-06-24 02:49:09,160][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.006692386232316494, 'train/grad_norm_pre_clip': 22.96933937072754, 'timing/data': 0.0009481020097155124, 'timing/model': 1.1905428869940806, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 1.52}) +[2026-06-24 02:49:13,525][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.146441251039505, 'train/grad_norm_pre_clip': 37.88939666748047, 'timing/data': 0.0010015000007115304, 'timing/model': 1.1994902460137382, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 1.52}) +[2026-06-24 02:49:17,861][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.02379249408841133, 'train/grad_norm_pre_clip': 42.91447830200195, 'timing/data': 0.0011086080048698932, 'timing/model': 1.176242489978904, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 1.52}) +[2026-06-24 02:49:22,438][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.021921778097748756, 'train/grad_norm_pre_clip': 15.220245361328125, 'timing/data': 0.0008182309975381941, 'timing/model': 1.4059016609971877, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 1.52}) +[2026-06-24 02:49:26,815][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.18703120946884155, 'train/grad_norm_pre_clip': 37.2784423828125, 'timing/data': 0.000792964972788468, 'timing/model': 1.1955882190086413, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 1.52}) +[2026-06-24 02:49:31,162][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.006770775653421879, 'train/grad_norm_pre_clip': 14.466012954711914, 'timing/data': 0.0007303589954972267, 'timing/model': 1.1778035259922035, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 1.52}) +[2026-06-24 02:49:35,505][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.014635389670729637, 'train/grad_norm_pre_clip': 9.523773193359375, 'timing/data': 0.0007721059955656528, 'timing/model': 1.2052594159904402, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 1.52}) +[2026-06-24 02:49:40,022][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.0008388269925490022, 'train/grad_norm_pre_clip': 19.771915435791016, 'timing/data': 0.0008920879918150604, 'timing/model': 1.386762074980652, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 1.52}) +[2026-06-24 02:49:44,371][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.01159700844436884, 'train/grad_norm_pre_clip': 14.410435676574707, 'timing/data': 0.0008243230113293976, 'timing/model': 1.198114869010169, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 1.52}) +[2026-06-24 02:49:48,678][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.00044105181586928666, 'train/grad_norm_pre_clip': 23.055557250976562, 'timing/data': 0.0009338450036011636, 'timing/model': 1.1765546399983577, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 1.52}) +[2026-06-24 02:49:53,013][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.0039019673131406307, 'train/grad_norm_pre_clip': 30.095754623413086, 'timing/data': 0.0009357889939565212, 'timing/model': 1.1875673149770591, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 1.52}) +[2026-06-24 02:49:57,568][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.08461090177297592, 'train/grad_norm_pre_clip': 33.92229461669922, 'timing/data': 0.0011132969812024385, 'timing/model': 1.2092613209970295, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 1.52}) +[2026-06-24 02:50:01,945][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.07478272169828415, 'train/grad_norm_pre_clip': 15.01167106628418, 'timing/data': 0.001071790000423789, 'timing/model': 1.19041940299212, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 1.52}) +[2026-06-24 02:50:06,269][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.03060217760503292, 'train/grad_norm_pre_clip': 15.98059368133545, 'timing/data': 0.0007605940045323223, 'timing/model': 1.1837632279784884, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 1.52}) +[2026-06-24 02:50:10,609][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.1585051417350769, 'train/grad_norm_pre_clip': 20.912885665893555, 'timing/data': 0.0007393559790216386, 'timing/model': 1.1943247689923737, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 1.52}) +[2026-06-24 02:50:15,172][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 5.923494609305635e-05, 'train/grad_norm_pre_clip': 20.229961395263672, 'timing/data': 0.0007589919841848314, 'timing/model': 1.1970051809912547, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 1.52}) +[2026-06-24 02:50:19,474][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.06122128665447235, 'train/grad_norm_pre_clip': 19.04430389404297, 'timing/data': 0.0008555410022381693, 'timing/model': 1.1747602869872935, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 1.52}) +[2026-06-24 02:50:23,780][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.05336663872003555, 'train/grad_norm_pre_clip': 24.88968276977539, 'timing/data': 0.0008066299778874964, 'timing/model': 1.174550970987184, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 1.52}) +[2026-06-24 02:50:28,127][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.03457917645573616, 'train/grad_norm_pre_clip': 14.275238037109375, 'timing/data': 0.0008023919945117086, 'timing/model': 1.1941943890124094, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 1.52}) +[2026-06-24 02:50:32,822][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.05560670793056488, 'train/grad_norm_pre_clip': 29.284557342529297, 'timing/data': 0.0007633499917574227, 'timing/model': 1.198705863993382, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 1.52}) +[2026-06-24 02:50:37,241][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.0524350181221962, 'train/grad_norm_pre_clip': 14.115029335021973, 'timing/data': 0.0012817580136470497, 'timing/model': 1.1834974889934529, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 1.52}) +[2026-06-24 02:50:41,595][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.027525372803211212, 'train/grad_norm_pre_clip': 23.27043914794922, 'timing/data': 0.0013245870068203658, 'timing/model': 1.1780372219800483, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 1.52}) +[2026-06-24 02:50:45,959][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.0048308223485946655, 'train/grad_norm_pre_clip': 14.88261604309082, 'timing/data': 0.000957860000198707, 'timing/model': 1.2017779610177968, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 1.52}) +[2026-06-24 02:50:50,508][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.01348647940903902, 'train/grad_norm_pre_clip': 9.107475280761719, 'timing/data': 0.0008602089947089553, 'timing/model': 1.2104356349736918, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 1.52}) +[2026-06-24 02:50:54,858][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.002176502486690879, 'train/grad_norm_pre_clip': 24.446537017822266, 'timing/data': 0.0007953490130603313, 'timing/model': 1.2072406120132655, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 1.52}) +[2026-06-24 02:50:59,203][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.13311737775802612, 'train/grad_norm_pre_clip': 77.53926849365234, 'timing/data': 0.001014804991427809, 'timing/model': 1.194676040991908, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 1.52}) +[2026-06-24 02:51:03,546][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.022076871246099472, 'train/grad_norm_pre_clip': 42.6320915222168, 'timing/data': 0.0008945419976953417, 'timing/model': 1.197570499003632, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 1.53}) +[2026-06-24 02:51:08,087][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.0012784452410414815, 'train/grad_norm_pre_clip': 12.667777061462402, 'timing/data': 0.0010092839947901666, 'timing/model': 1.2005719340231735, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 1.53}) +[2026-06-24 02:51:12,464][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.004755983594805002, 'train/grad_norm_pre_clip': 23.533231735229492, 'timing/data': 0.0007524700195062906, 'timing/model': 1.1987626619811635, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 1.53}) +[2026-06-24 02:51:16,830][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.148262158036232, 'train/grad_norm_pre_clip': 25.381969451904297, 'timing/data': 0.0010017409804277122, 'timing/model': 1.1965167129819747, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 1.53}) +[2026-06-24 02:51:21,190][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.10240692645311356, 'train/grad_norm_pre_clip': 31.260597229003906, 'timing/data': 0.0007914320158306509, 'timing/model': 1.1948865620070137, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 1.53}) +[2026-06-24 02:51:25,691][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.012758408673107624, 'train/grad_norm_pre_clip': 9.353256225585938, 'timing/data': 0.0008092450152616948, 'timing/model': 1.18207307101693, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 1.53}) +[2026-06-24 02:51:30,080][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.005151193123310804, 'train/grad_norm_pre_clip': 16.911388397216797, 'timing/data': 0.0007651430205442011, 'timing/model': 1.2057424280210398, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 1.53}) +[2026-06-24 02:51:34,426][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.07207732647657394, 'train/grad_norm_pre_clip': 10.230172157287598, 'timing/data': 0.0008064099820330739, 'timing/model': 1.1984751830168534, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 1.53}) +[2026-06-24 02:51:38,775][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.10333403199911118, 'train/grad_norm_pre_clip': 14.03415298461914, 'timing/data': 0.0007820640166755766, 'timing/model': 1.1848426890210249, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 1.53}) +[2026-06-24 02:51:43,356][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.01017655711621046, 'train/grad_norm_pre_clip': 20.644481658935547, 'timing/data': 0.0008635260164737701, 'timing/model': 1.1994872630166356, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 1.53}) +[2026-06-24 02:51:47,698][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.0464705228805542, 'train/grad_norm_pre_clip': 12.698883056640625, 'timing/data': 0.0008712590206414461, 'timing/model': 1.1777004750038031, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 1.53}) +[2026-06-24 02:51:52,032][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.1494932323694229, 'train/grad_norm_pre_clip': 24.985410690307617, 'timing/data': 0.000771664985222742, 'timing/model': 1.1769916260091122, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 1.53}) +[2026-06-24 02:51:56,365][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.04637607932090759, 'train/grad_norm_pre_clip': 29.70716667175293, 'timing/data': 0.0008053990022744983, 'timing/model': 1.1805290110060014, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 1.53}) +[2026-06-24 02:52:00,882][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.035379450768232346, 'train/grad_norm_pre_clip': 37.985050201416016, 'timing/data': 0.0007075670000631362, 'timing/model': 1.1974185529979877, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 1.53}) +[2026-06-24 02:52:05,207][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.038224924355745316, 'train/grad_norm_pre_clip': 10.161834716796875, 'timing/data': 0.0011388139973860234, 'timing/model': 1.1812535299977753, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 1.53}) +[2026-06-24 02:52:09,584][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.07871054112911224, 'train/grad_norm_pre_clip': 27.9144229888916, 'timing/data': 0.0008809769933577627, 'timing/model': 1.1962087170104496, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 1.53}) +[2026-06-24 02:52:13,936][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.054779279977083206, 'train/grad_norm_pre_clip': 14.019683837890625, 'timing/data': 0.0007393260020762682, 'timing/model': 1.2045568999892566, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 1.53}) +[2026-06-24 02:52:18,415][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.0031317195389419794, 'train/grad_norm_pre_clip': 26.366554260253906, 'timing/data': 0.0009545829962007701, 'timing/model': 1.1791770349955186, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 1.53}) +[2026-06-24 02:52:22,771][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.005399319808930159, 'train/grad_norm_pre_clip': 6.034742832183838, 'timing/data': 0.0008933809876907617, 'timing/model': 1.1770234470022842, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 1.53}) +[2026-06-24 02:52:27,094][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.02126758173108101, 'train/grad_norm_pre_clip': 37.01520538330078, 'timing/data': 0.0008684739877935499, 'timing/model': 1.1820734060020186, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 1.53}) +[2026-06-24 02:52:31,443][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.009330471977591515, 'train/grad_norm_pre_clip': 8.334288597106934, 'timing/data': 0.0010286300093866885, 'timing/model': 1.193427291000262, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 1.53}) +[2026-06-24 02:52:35,935][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.005695946514606476, 'train/grad_norm_pre_clip': 2.947692632675171, 'timing/data': 0.0007380940078292042, 'timing/model': 1.1973579630139284, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 1.53}) +[2026-06-24 02:52:40,328][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.0017081336118280888, 'train/grad_norm_pre_clip': 31.3153133392334, 'timing/data': 0.0007668359903618693, 'timing/model': 1.1983881659980398, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 1.53}) +[2026-06-24 02:52:44,663][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.04172789677977562, 'train/grad_norm_pre_clip': 23.312211990356445, 'timing/data': 0.0007809230010025203, 'timing/model': 1.1826911179814488, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 1.53}) +[2026-06-24 02:52:48,994][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.08180197328329086, 'train/grad_norm_pre_clip': 17.505067825317383, 'timing/data': 0.0007911409775260836, 'timing/model': 1.1776215930003673, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 1.53}) +[2026-06-24 02:52:53,509][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.002777869813144207, 'train/grad_norm_pre_clip': 17.138259887695312, 'timing/data': 0.0008641460153739899, 'timing/model': 1.1972688690002542, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 1.53}) +[2026-06-24 02:52:57,922][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.012475836090743542, 'train/grad_norm_pre_clip': 26.855186462402344, 'timing/data': 0.0009600940102245659, 'timing/model': 1.179279847012367, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 1.54}) +[2026-06-24 02:53:02,249][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.015700524672865868, 'train/grad_norm_pre_clip': 20.521289825439453, 'timing/data': 0.0009130070102401078, 'timing/model': 1.185346654994646, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 1.54}) +[2026-06-24 02:53:06,623][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.006947225891053677, 'train/grad_norm_pre_clip': 14.29838752746582, 'timing/data': 0.0008137229888234288, 'timing/model': 1.178967650019331, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 1.54}) +[2026-06-24 02:53:11,130][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.027794932946562767, 'train/grad_norm_pre_clip': 29.280839920043945, 'timing/data': 0.0009424710005987436, 'timing/model': 1.1868031159974635, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 1.54}) +[2026-06-24 02:53:15,478][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.0069312057457864285, 'train/grad_norm_pre_clip': 11.361745834350586, 'timing/data': 0.0010321870213374496, 'timing/model': 1.2030875490163453, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 1.54}) +[2026-06-24 02:53:19,809][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.010314724408090115, 'train/grad_norm_pre_clip': 21.176965713500977, 'timing/data': 0.0010388790105935186, 'timing/model': 1.1750826640054584, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 1.54}) +[2026-06-24 02:53:24,161][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.01842786930501461, 'train/grad_norm_pre_clip': 33.09881591796875, 'timing/data': 0.0007437739986926317, 'timing/model': 1.1944059419911355, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 1.54}) +[2026-06-24 02:53:28,620][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.08824682980775833, 'train/grad_norm_pre_clip': 27.662160873413086, 'timing/data': 0.0007342869939748198, 'timing/model': 1.174521086999448, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 1.54}) +[2026-06-24 02:53:32,976][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.008657609112560749, 'train/grad_norm_pre_clip': 5.53433084487915, 'timing/data': 0.0012486070045270026, 'timing/model': 1.190996332996292, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 1.54}) +[2026-06-24 02:53:37,351][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.027223162353038788, 'train/grad_norm_pre_clip': 14.38971996307373, 'timing/data': 0.0007905499951448292, 'timing/model': 1.1809317610168364, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 1.54}) +[2026-06-24 02:53:41,742][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.01608465611934662, 'train/grad_norm_pre_clip': 29.33681869506836, 'timing/data': 0.0009509169904049486, 'timing/model': 1.2030191349913366, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 1.54}) +[2026-06-24 02:53:46,234][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.0960266962647438, 'train/grad_norm_pre_clip': 33.737728118896484, 'timing/data': 0.0008907149895094335, 'timing/model': 1.1804447419999633, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 1.54}) +[2026-06-24 02:53:50,597][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.0031864012125879526, 'train/grad_norm_pre_clip': 11.507537841796875, 'timing/data': 0.0007649030012544245, 'timing/model': 1.195102389989188, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 1.54}) +[2026-06-24 02:53:54,974][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.0021657964680343866, 'train/grad_norm_pre_clip': 47.25676727294922, 'timing/data': 0.0007285160245373845, 'timing/model': 1.2045581580023281, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 1.54}) +[2026-06-24 02:53:59,329][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.036632239818573, 'train/grad_norm_pre_clip': 18.661340713500977, 'timing/data': 0.0007684089941903949, 'timing/model': 1.1778276869736146, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 1.54}) +[2026-06-24 02:54:03,823][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.016910478472709656, 'train/grad_norm_pre_clip': 9.804363250732422, 'timing/data': 0.000725410005543381, 'timing/model': 1.174189926998224, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 1.54}) +[2026-06-24 02:54:08,149][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.01130734197795391, 'train/grad_norm_pre_clip': 19.933441162109375, 'timing/data': 0.000728244980564341, 'timing/model': 1.1762859500013292, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 1.54}) +[2026-06-24 02:54:12,479][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.03538104146718979, 'train/grad_norm_pre_clip': 32.93759536743164, 'timing/data': 0.0007771060045342892, 'timing/model': 1.1752362109837122, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 1.54}) +[2026-06-24 02:54:16,794][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.04983300343155861, 'train/grad_norm_pre_clip': 8.012347221374512, 'timing/data': 0.0007637010130565614, 'timing/model': 1.1801539160078391, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 1.54}) +[2026-06-24 02:54:21,398][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.0008036089711822569, 'train/grad_norm_pre_clip': 11.242713928222656, 'timing/data': 0.0009245779947377741, 'timing/model': 1.191352795023704, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 1.54}) +[2026-06-24 02:54:25,744][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.2240801900625229, 'train/grad_norm_pre_clip': 49.26792526245117, 'timing/data': 0.0008844039984978735, 'timing/model': 1.1776947239995934, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 1.54}) +[2026-06-24 02:54:30,109][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.06022217869758606, 'train/grad_norm_pre_clip': 18.759963989257812, 'timing/data': 0.0007512579904869199, 'timing/model': 1.1814220999949612, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 1.54}) +[2026-06-24 02:54:34,475][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.03298269957304001, 'train/grad_norm_pre_clip': 42.50987243652344, 'timing/data': 0.0011101400014013052, 'timing/model': 1.1991969459922984, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 1.54}) +[2026-06-24 02:54:39,105][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.0571167953312397, 'train/grad_norm_pre_clip': 16.934478759765625, 'timing/data': 0.0008493489876855165, 'timing/model': 1.1779396170168184, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 1.54}) +[2026-06-24 02:54:43,461][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.11223095655441284, 'train/grad_norm_pre_clip': 29.803165435791016, 'timing/data': 0.0007902600045781583, 'timing/model': 1.2109858650073875, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 1.54}) +[2026-06-24 02:54:47,828][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.24132311344146729, 'train/grad_norm_pre_clip': 46.298213958740234, 'timing/data': 0.0011915710056200624, 'timing/model': 1.2059594900056254, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 1.54}) +[2026-06-24 02:54:52,433][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.03478207439184189, 'train/grad_norm_pre_clip': 33.83082962036133, 'timing/data': 0.0011096599919255823, 'timing/model': 1.4603529279993381, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 1.55}) +[2026-06-24 02:54:56,762][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.07402301579713821, 'train/grad_norm_pre_clip': 26.700841903686523, 'timing/data': 0.0011752010032068938, 'timing/model': 1.1791762610082515, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 1.55}) +[2026-06-24 02:55:01,159][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.04267982766032219, 'train/grad_norm_pre_clip': 11.003800392150879, 'timing/data': 0.000977735995547846, 'timing/model': 1.2118380310130306, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 1.55}) +[2026-06-24 02:55:05,530][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.08109228312969208, 'train/grad_norm_pre_clip': 40.55198669433594, 'timing/data': 0.0007695919775869697, 'timing/model': 1.195167414000025, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 1.55}) +[2026-06-24 02:55:10,182][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.05264779552817345, 'train/grad_norm_pre_clip': 23.283058166503906, 'timing/data': 0.0008332300058100373, 'timing/model': 1.4713696910184808, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 1.55}) +[2026-06-24 02:55:14,546][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.0008633268880657852, 'train/grad_norm_pre_clip': 35.50640106201172, 'timing/data': 0.0009024269820656627, 'timing/model': 1.2006886739982292, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 1.55}) +[2026-06-24 02:55:18,890][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.0343334823846817, 'train/grad_norm_pre_clip': 23.11211395263672, 'timing/data': 0.0007991059974301606, 'timing/model': 1.1801972690154798, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 1.55}) +[2026-06-24 02:55:23,255][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4000 +[2026-06-24 02:56:21,204][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4000 +[2026-06-24 02:56:21,205][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4000 +[2026-06-24 02:56:58,569][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4000 +[2026-06-24 03:25:18,473][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000 +[2026-06-24 03:25:18,474][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.12603768706321716, 'train/grad_norm_pre_clip': 27.27924156188965, 'eval/action_loss/seconds': 57.94654825900216, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03325643017888069, 'eval/latency_2/loss': 0.03325643017888069, 'eval/flappy/loss': 0.03325643017888069, 'eval/flappy/latency_2/loss': 0.03325643017888069, 'eval/flap_precision': 0.984993178717599, 'eval/flap_recall': 0.9796472184531886, 'eval/flap_f1': 0.9823129251700681, 'eval/flap_rate_model': 0.2036111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9823129251700681, 'eval/latency_2/flap_precision': 0.984993178717599, 'eval/latency_2/flap_recall': 0.9796472184531886, 'eval/latency_2/flap_f1': 0.9823129251700681, 'eval/latency_2/flap_rate_model': 0.2036111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9823129251700681, 'eval/action_classification/data_wait_seconds': 0.5035779476165771, 'eval/action_classification/predict_seconds': 36.71969985961914, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0010682540014386177, 'timing/model': 1.1788611609954387, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 1.55}) +[2026-06-24 03:25:18,475][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_4000_state +[2026-06-24 03:25:18,482][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 03:25:27,255][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_4000_state/model.safetensors +[2026-06-24 03:25:37,728][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_4000_state/optimizer.bin +[2026-06-24 03:25:37,729][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_4000_state/sampler.bin +[2026-06-24 03:25:37,729][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-24 03:25:37,730][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-24 03:25:37,922][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-24 03:25:37,923][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 03:25:37,947][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 03:25:42,119][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.1405641734600067, 'train/grad_norm_pre_clip': 80.53997802734375, 'timing/data': 0.0007188080053310841, 'timing/model': 1.1455850579950493, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 1.55}) +[2026-06-24 03:25:46,310][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.028572292998433113, 'train/grad_norm_pre_clip': 14.535884857177734, 'timing/data': 0.0007933259767014533, 'timing/model': 1.1635053060017526, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 1.55}) +[2026-06-24 03:25:50,501][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.015513311140239239, 'train/grad_norm_pre_clip': 48.21610641479492, 'timing/data': 0.0008008999866433442, 'timing/model': 1.1540968030167278, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 1.55}) +[2026-06-24 03:25:54,879][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.004469217266887426, 'train/grad_norm_pre_clip': 23.908946990966797, 'timing/data': 0.0007422920025419444, 'timing/model': 1.136715100001311, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 1.55}) +[2026-06-24 03:25:59,053][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.05578841641545296, 'train/grad_norm_pre_clip': 14.538043975830078, 'timing/data': 0.0008927999879233539, 'timing/model': 1.1372426049783826, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 1.55}) +[2026-06-24 03:26:03,243][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.004944877699017525, 'train/grad_norm_pre_clip': 14.697663307189941, 'timing/data': 0.0006849750061519444, 'timing/model': 1.1378194309945684, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 1.55}) +[2026-06-24 03:26:07,599][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.22691094875335693, 'train/grad_norm_pre_clip': 37.838890075683594, 'timing/data': 0.0008362350054085255, 'timing/model': 1.331149876001291, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 1.55}) +[2026-06-24 03:26:11,841][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.021089961752295494, 'train/grad_norm_pre_clip': 78.26300048828125, 'timing/data': 0.0007935559842735529, 'timing/model': 1.1618835380068049, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 1.55}) +[2026-06-24 03:26:16,052][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.029027163982391357, 'train/grad_norm_pre_clip': 22.01356315612793, 'timing/data': 0.0007209320028778166, 'timing/model': 1.148738662013784, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 1.55}) +[2026-06-24 03:26:20,272][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.01294513326138258, 'train/grad_norm_pre_clip': 12.343764305114746, 'timing/data': 0.0007838580058887601, 'timing/model': 1.1498595229932107, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 1.55}) +[2026-06-24 03:26:24,534][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.13476590812206268, 'train/grad_norm_pre_clip': 21.879222869873047, 'timing/data': 0.0007150209858082235, 'timing/model': 1.1692873149877414, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 1.55}) +[2026-06-24 03:26:28,969][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.17727462947368622, 'train/grad_norm_pre_clip': 24.168163299560547, 'timing/data': 0.0007968819991219789, 'timing/model': 1.1644333880103659, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 1.55}) +[2026-06-24 03:26:33,214][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.009170936420559883, 'train/grad_norm_pre_clip': 13.542177200317383, 'timing/data': 0.0007343269826378673, 'timing/model': 1.1576479310169816, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 1.55}) +[2026-06-24 03:26:37,463][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.0022479891777038574, 'train/grad_norm_pre_clip': 25.542396545410156, 'timing/data': 0.0009703630057629198, 'timing/model': 1.1556019910203759, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 1.55}) +[2026-06-24 03:26:41,710][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.039914414286613464, 'train/grad_norm_pre_clip': 21.052858352661133, 'timing/data': 0.001002892997348681, 'timing/model': 1.158111357013695, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 1.55}) +[2026-06-24 03:26:46,231][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.014323013834655285, 'train/grad_norm_pre_clip': 15.473077774047852, 'timing/data': 0.0009316009818576276, 'timing/model': 1.1510050980141386, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 1.55}) +[2026-06-24 03:26:50,525][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.019020041450858116, 'train/grad_norm_pre_clip': 14.246369361877441, 'timing/data': 0.0008860570087563246, 'timing/model': 1.1558220860024448, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 1.55}) +[2026-06-24 03:26:54,821][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.08277887850999832, 'train/grad_norm_pre_clip': 25.70223617553711, 'timing/data': 0.000823982001747936, 'timing/model': 1.1757579270051792, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 1.56}) +[2026-06-24 03:26:59,110][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.03415155038237572, 'train/grad_norm_pre_clip': 12.743575096130371, 'timing/data': 0.000813702994491905, 'timing/model': 1.1686911809956655, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 1.56}) +[2026-06-24 03:27:03,537][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.011904583312571049, 'train/grad_norm_pre_clip': 12.291852951049805, 'timing/data': 0.0008923689892981201, 'timing/model': 1.1576790869876277, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 1.56}) +[2026-06-24 03:27:07,785][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.291568398475647, 'train/grad_norm_pre_clip': 40.367008209228516, 'timing/data': 0.0011579889978747815, 'timing/model': 1.1603258159884717, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 1.56}) +[2026-06-24 03:27:12,042][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.06259206682443619, 'train/grad_norm_pre_clip': 33.440284729003906, 'timing/data': 0.0009125249926000834, 'timing/model': 1.1633167609979864, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 1.56}) +[2026-06-24 03:27:16,298][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.04936904460191727, 'train/grad_norm_pre_clip': 17.86897087097168, 'timing/data': 0.0007426119991578162, 'timing/model': 1.1620160379970912, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 1.56}) +[2026-06-24 03:27:20,752][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.11201166361570358, 'train/grad_norm_pre_clip': 20.47222137451172, 'timing/data': 0.0007301290170289576, 'timing/model': 1.182888028997695, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 1.56}) +[2026-06-24 03:27:25,055][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.0692044347524643, 'train/grad_norm_pre_clip': 16.62190818786621, 'timing/data': 0.00073938601417467, 'timing/model': 1.1842297080147546, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 1.56}) +[2026-06-24 03:27:29,385][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.039859332144260406, 'train/grad_norm_pre_clip': 32.20927429199219, 'timing/data': 0.0011546130117494613, 'timing/model': 1.187461577996146, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 1.56}) +[2026-06-24 03:27:33,727][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.01092025451362133, 'train/grad_norm_pre_clip': 20.2519474029541, 'timing/data': 0.0011247980000916868, 'timing/model': 1.1761153349943925, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 1.56}) +[2026-06-24 03:27:38,292][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.017049983143806458, 'train/grad_norm_pre_clip': 23.534902572631836, 'timing/data': 0.0008309439872391522, 'timing/model': 1.1739257499866653, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 1.56}) +[2026-06-24 03:27:42,590][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.07141823321580887, 'train/grad_norm_pre_clip': 23.77570915222168, 'timing/data': 0.001001811004243791, 'timing/model': 1.1640493439917918, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 1.56}) +[2026-06-24 03:27:46,890][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.05146539956331253, 'train/grad_norm_pre_clip': 23.10832977294922, 'timing/data': 0.0009487929928582162, 'timing/model': 1.1960338939970825, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 1.56}) +[2026-06-24 03:27:51,213][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.021723754703998566, 'train/grad_norm_pre_clip': 23.946287155151367, 'timing/data': 0.0007948880083858967, 'timing/model': 1.1915295719809365, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 1.56}) +[2026-06-24 03:27:55,847][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.0295525211840868, 'train/grad_norm_pre_clip': 5.685862064361572, 'timing/data': 0.0007311700028367341, 'timing/model': 1.1956890170113184, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 1.56}) +[2026-06-24 03:28:00,185][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.05818809196352959, 'train/grad_norm_pre_clip': 8.765402793884277, 'timing/data': 0.0008006999851204455, 'timing/model': 1.198473850992741, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 1.56}) +[2026-06-24 03:28:04,510][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.030643874779343605, 'train/grad_norm_pre_clip': 16.101924896240234, 'timing/data': 0.000810678000561893, 'timing/model': 1.1699449430161621, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 1.56}) +[2026-06-24 03:28:08,821][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.049586113542318344, 'train/grad_norm_pre_clip': 40.74636459350586, 'timing/data': 0.0007612159824930131, 'timing/model': 1.1732361940084957, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 1.56}) +[2026-06-24 03:28:13,379][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.03973982483148575, 'train/grad_norm_pre_clip': 32.384986877441406, 'timing/data': 0.0007234060030896217, 'timing/model': 1.167714151990367, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 1.56}) +[2026-06-24 03:28:17,663][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.034175753593444824, 'train/grad_norm_pre_clip': 17.849864959716797, 'timing/data': 0.0010215479996986687, 'timing/model': 1.1712283839879092, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 1.56}) +[2026-06-24 03:28:21,997][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.04439954832196236, 'train/grad_norm_pre_clip': 39.56355285644531, 'timing/data': 0.0012483259779401124, 'timing/model': 1.193147929996485, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 1.56}) +[2026-06-24 03:28:26,334][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.08060447126626968, 'train/grad_norm_pre_clip': 19.37196922302246, 'timing/data': 0.0007850909896660596, 'timing/model': 1.1778631019988097, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 1.56}) +[2026-06-24 03:28:30,948][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.10916963219642639, 'train/grad_norm_pre_clip': 18.690187454223633, 'timing/data': 0.0009052219975274056, 'timing/model': 1.1920037459931336, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 1.56}) +[2026-06-24 03:28:35,297][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.020883547142148018, 'train/grad_norm_pre_clip': 15.192164421081543, 'timing/data': 0.0007813939882908016, 'timing/model': 1.1943474559811875, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 1.56}) +[2026-06-24 03:28:39,637][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.04665582627058029, 'train/grad_norm_pre_clip': 20.615108489990234, 'timing/data': 0.0009763939888216555, 'timing/model': 1.1700267330161296, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 1.56}) +[2026-06-24 03:28:43,991][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.029121724888682365, 'train/grad_norm_pre_clip': 9.672730445861816, 'timing/data': 0.0010462529899086803, 'timing/model': 1.172180843015667, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 1.56}) +[2026-06-24 03:28:48,593][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.05765116959810257, 'train/grad_norm_pre_clip': 26.06879425048828, 'timing/data': 0.0011590520152822137, 'timing/model': 1.2066373900161125, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 1.57}) +[2026-06-24 03:28:52,927][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.05111997574567795, 'train/grad_norm_pre_clip': 8.221571922302246, 'timing/data': 0.0007500050123780966, 'timing/model': 1.1724053570069373, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 1.57}) +[2026-06-24 03:28:57,258][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.0639190822839737, 'train/grad_norm_pre_clip': 29.179201126098633, 'timing/data': 0.001405607006745413, 'timing/model': 1.1752403450082056, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 1.57}) +[2026-06-24 03:29:01,620][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.004553081933408976, 'train/grad_norm_pre_clip': 19.030471801757812, 'timing/data': 0.0009652530134189874, 'timing/model': 1.1899485379981343, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 1.57}) +[2026-06-24 03:29:06,163][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.02041260153055191, 'train/grad_norm_pre_clip': 26.030479431152344, 'timing/data': 0.0007081080111674964, 'timing/model': 1.1715627299854532, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 1.57}) +[2026-06-24 03:29:10,497][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.010767454281449318, 'train/grad_norm_pre_clip': 12.155021667480469, 'timing/data': 0.0007188279996626079, 'timing/model': 1.1910261269949842, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 1.57}) +[2026-06-24 03:29:14,877][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.05115086957812309, 'train/grad_norm_pre_clip': 16.923885345458984, 'timing/data': 0.0011631280067376792, 'timing/model': 1.1998373660026118, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 1.57}) +[2026-06-24 03:29:19,254][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.05254637449979782, 'train/grad_norm_pre_clip': 24.228906631469727, 'timing/data': 0.0009044409962370992, 'timing/model': 1.1949740600248333, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 1.57}) +[2026-06-24 03:29:23,865][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.032797884196043015, 'train/grad_norm_pre_clip': 23.954607009887695, 'timing/data': 0.0007340360025409609, 'timing/model': 1.1734773860080168, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 1.57}) +[2026-06-24 03:29:28,197][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.0006176796741783619, 'train/grad_norm_pre_clip': 17.110477447509766, 'timing/data': 0.0013043299841228873, 'timing/model': 1.1933282999962103, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 1.57}) +[2026-06-24 03:29:32,532][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.09481502324342728, 'train/grad_norm_pre_clip': 20.33437156677246, 'timing/data': 0.0007198700041044503, 'timing/model': 1.1889610910147894, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 1.57}) +[2026-06-24 03:29:37,146][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.11305837333202362, 'train/grad_norm_pre_clip': 44.072628021240234, 'timing/data': 0.0007111140002962202, 'timing/model': 1.1713082779897377, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 1.57}) +[2026-06-24 03:29:41,736][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.020949093624949455, 'train/grad_norm_pre_clip': 14.577826499938965, 'timing/data': 0.0012260249932296574, 'timing/model': 1.1909843790053856, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 1.57}) +[2026-06-24 03:29:46,054][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.0130185317248106, 'train/grad_norm_pre_clip': 5.0808234214782715, 'timing/data': 0.0007938370108604431, 'timing/model': 1.176598752004793, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 1.57}) +[2026-06-24 03:29:50,390][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.09469862282276154, 'train/grad_norm_pre_clip': 29.948753356933594, 'timing/data': 0.0010272279905620962, 'timing/model': 1.2001726739981677, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 1.57}) +[2026-06-24 03:29:54,705][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.013966362923383713, 'train/grad_norm_pre_clip': 21.414522171020508, 'timing/data': 0.0007309500069823116, 'timing/model': 1.1708759980101604, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 1.57}) +[2026-06-24 03:29:59,344][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.06233371049165726, 'train/grad_norm_pre_clip': 15.682129859924316, 'timing/data': 0.0014091830234974623, 'timing/model': 1.2104634850111324, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 1.57}) +[2026-06-24 03:30:03,683][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.025974074378609657, 'train/grad_norm_pre_clip': 11.035592079162598, 'timing/data': 0.000981021992629394, 'timing/model': 1.203898547013523, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 1.57}) +[2026-06-24 03:30:08,050][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.08989708870649338, 'train/grad_norm_pre_clip': 18.367090225219727, 'timing/data': 0.0007991369930095971, 'timing/model': 1.2059565579984337, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 1.57}) +[2026-06-24 03:30:12,421][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.018253112211823463, 'train/grad_norm_pre_clip': 22.078943252563477, 'timing/data': 0.0011964500008616596, 'timing/model': 1.212303002976114, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 1.57}) +[2026-06-24 03:30:17,014][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.0035556170623749495, 'train/grad_norm_pre_clip': 18.76633644104004, 'timing/data': 0.0011280739854555577, 'timing/model': 1.2008113019983284, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 1.57}) +[2026-06-24 03:30:21,375][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.057314183562994, 'train/grad_norm_pre_clip': 17.611238479614258, 'timing/data': 0.0010954239987768233, 'timing/model': 1.2157398630224634, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 1.57}) +[2026-06-24 03:30:25,767][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.10832139849662781, 'train/grad_norm_pre_clip': 29.31863021850586, 'timing/data': 0.0011654529953375459, 'timing/model': 1.1989938140031882, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 1.57}) +[2026-06-24 03:30:30,127][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.00451959203928709, 'train/grad_norm_pre_clip': 14.655730247497559, 'timing/data': 0.0008461230027023703, 'timing/model': 1.2083237110055052, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 1.57}) +[2026-06-24 03:30:34,680][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.009883669205009937, 'train/grad_norm_pre_clip': 9.45064640045166, 'timing/data': 0.0009204100060742348, 'timing/model': 1.1815889280114789, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 1.57}) +[2026-06-24 03:30:39,020][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.0030917758122086525, 'train/grad_norm_pre_clip': 19.132966995239258, 'timing/data': 0.0007956100162118673, 'timing/model': 1.1932565239840187, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 1.57}) +[2026-06-24 03:30:43,410][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.09507564455270767, 'train/grad_norm_pre_clip': 9.850967407226562, 'timing/data': 0.0011776559986174107, 'timing/model': 1.2012889540055767, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 1.58}) +[2026-06-24 03:30:47,777][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.1749284565448761, 'train/grad_norm_pre_clip': 46.26805114746094, 'timing/data': 0.0007436340092681348, 'timing/model': 1.2044073950091843, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 1.58}) +[2026-06-24 03:30:52,391][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.014657219871878624, 'train/grad_norm_pre_clip': 11.322589874267578, 'timing/data': 0.0009138479945249856, 'timing/model': 1.1735051029827446, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 1.58}) +[2026-06-24 03:30:56,692][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.0029136724770069122, 'train/grad_norm_pre_clip': 23.035968780517578, 'timing/data': 0.0007791300013195723, 'timing/model': 1.1739927750022616, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 1.58}) +[2026-06-24 03:31:01,016][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.01466979831457138, 'train/grad_norm_pre_clip': 35.54052734375, 'timing/data': 0.0007669570040889084, 'timing/model': 1.2027067620074376, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 1.58}) +[2026-06-24 03:31:05,391][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.016062624752521515, 'train/grad_norm_pre_clip': 10.047595977783203, 'timing/data': 0.0011851800081785768, 'timing/model': 1.2013065659848507, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 1.58}) +[2026-06-24 03:31:09,879][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.0020447091665118933, 'train/grad_norm_pre_clip': 13.202337265014648, 'timing/data': 0.0007913019799161702, 'timing/model': 1.1797421959927306, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 1.58}) +[2026-06-24 03:31:14,190][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.03546890243887901, 'train/grad_norm_pre_clip': 16.330421447753906, 'timing/data': 0.0012086729984730482, 'timing/model': 1.1737960090104025, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 1.58}) +[2026-06-24 03:31:18,500][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.003347584279254079, 'train/grad_norm_pre_clip': 18.07229995727539, 'timing/data': 0.0007621679978910834, 'timing/model': 1.1774806159955915, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 1.58}) +[2026-06-24 03:31:22,834][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.004462929908186197, 'train/grad_norm_pre_clip': 10.879097938537598, 'timing/data': 0.0009004340099636465, 'timing/model': 1.1810408839955926, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 1.58}) +[2026-06-24 03:31:27,458][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.0042930105701088905, 'train/grad_norm_pre_clip': 11.149575233459473, 'timing/data': 0.000760986004024744, 'timing/model': 1.1806221999868285, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 1.58}) +[2026-06-24 03:31:31,799][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.0024765939451754093, 'train/grad_norm_pre_clip': 8.564652442932129, 'timing/data': 0.0009394259832333773, 'timing/model': 1.1801551780081354, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 1.58}) +[2026-06-24 03:31:36,146][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.02170623652637005, 'train/grad_norm_pre_clip': 33.471832275390625, 'timing/data': 0.0010790140077006072, 'timing/model': 1.1833739040012006, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 1.58}) +[2026-06-24 03:31:40,462][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.11448299884796143, 'train/grad_norm_pre_clip': 41.11199951171875, 'timing/data': 0.0008495489892084152, 'timing/model': 1.181251161993714, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 1.58}) +[2026-06-24 03:31:45,096][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.006608688738197088, 'train/grad_norm_pre_clip': 5.905020713806152, 'timing/data': 0.00079894601367414, 'timing/model': 1.20469281001715, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 1.58}) +[2026-06-24 03:31:49,441][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.16614161431789398, 'train/grad_norm_pre_clip': 30.46872901916504, 'timing/data': 0.0008238420123234391, 'timing/model': 1.1786986360093579, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 1.58}) +[2026-06-24 03:31:53,823][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.001991728786379099, 'train/grad_norm_pre_clip': 22.33578872680664, 'timing/data': 0.0012714490003418177, 'timing/model': 1.2068907200009562, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 1.58}) +[2026-06-24 03:31:58,178][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.12487907707691193, 'train/grad_norm_pre_clip': 39.6142463684082, 'timing/data': 0.000746339006582275, 'timing/model': 1.1796285930031445, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 1.58}) +[2026-06-24 03:32:02,819][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.06411167979240417, 'train/grad_norm_pre_clip': 16.182966232299805, 'timing/data': 0.000878353021107614, 'timing/model': 1.201194627006771, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 1.58}) +[2026-06-24 03:32:07,177][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.05400950461626053, 'train/grad_norm_pre_clip': 22.51885223388672, 'timing/data': 0.0010707390028983355, 'timing/model': 1.1777297259832267, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 1.58}) +[2026-06-24 03:32:11,539][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.020638294517993927, 'train/grad_norm_pre_clip': 8.18901252746582, 'timing/data': 0.0012095549900550395, 'timing/model': 1.1936305210110731, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 1.58}) +[2026-06-24 03:32:15,855][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.13179737329483032, 'train/grad_norm_pre_clip': 16.595827102661133, 'timing/data': 0.0007942480151541531, 'timing/model': 1.1779669729876332, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 1.58}) +[2026-06-24 03:32:20,463][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.15381301939487457, 'train/grad_norm_pre_clip': 39.173614501953125, 'timing/data': 0.000916232995223254, 'timing/model': 1.194077758002095, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 1.58}) +[2026-06-24 03:32:24,832][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.0002577630802989006, 'train/grad_norm_pre_clip': 25.101425170898438, 'timing/data': 0.0007668769976589829, 'timing/model': 1.1937404840136878, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 1.58}) +[2026-06-24 03:32:29,194][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.05162953585386276, 'train/grad_norm_pre_clip': 60.49314498901367, 'timing/data': 0.001471147988922894, 'timing/model': 1.1948408569733147, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 1.58}) +[2026-06-24 03:32:33,536][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.04505455866456032, 'train/grad_norm_pre_clip': 84.97935485839844, 'timing/data': 0.0007693709922023118, 'timing/model': 1.1941394509922247, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 1.58}) +[2026-06-24 03:32:38,114][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.010646376758813858, 'train/grad_norm_pre_clip': 10.711058616638184, 'timing/data': 0.0009331540204584599, 'timing/model': 1.1771209590078797, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 1.59}) +[2026-06-24 03:32:42,473][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.09285228699445724, 'train/grad_norm_pre_clip': 41.16032028198242, 'timing/data': 0.0007699719863012433, 'timing/model': 1.1796446219959762, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 1.59}) +[2026-06-24 03:32:46,843][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.12149174511432648, 'train/grad_norm_pre_clip': 18.070850372314453, 'timing/data': 0.0014162760053295642, 'timing/model': 1.1924202439840883, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 1.59}) +[2026-06-24 03:32:51,189][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.029842182993888855, 'train/grad_norm_pre_clip': 8.691309928894043, 'timing/data': 0.0008302839996758848, 'timing/model': 1.1933081039751414, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 1.59}) +[2026-06-24 03:32:55,816][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4100 +[2026-06-24 03:33:53,277][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4100 +[2026-06-24 03:33:53,277][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4100 +[2026-06-24 03:34:29,938][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4100 +[2026-06-24 03:34:29,940][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.0034439254086464643, 'train/grad_norm_pre_clip': 17.631980895996094, 'eval/action_loss/seconds': 57.45676895001088, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.0322321318089962, 'eval/latency_2/loss': 0.0322321318089962, 'eval/flappy/loss': 0.0322321318089962, 'eval/flappy/latency_2/loss': 0.0322321318089962, 'eval/flap_precision': 0.977027027027027, 'eval/flap_recall': 0.9810040705563093, 'eval/flap_f1': 0.9790115098171971, 'eval/flap_rate_model': 0.20555555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9790115098171971, 'eval/latency_2/flap_precision': 0.977027027027027, 'eval/latency_2/flap_recall': 0.9810040705563093, 'eval/latency_2/flap_f1': 0.9790115098171971, 'eval/latency_2/flap_rate_model': 0.20555555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9790115098171971, 'eval/action_classification/data_wait_seconds': 0.4115184545516968, 'eval/action_classification/predict_seconds': 36.10773849487305, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0008788029954303056, 'timing/model': 1.177133140998194, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 1.59}) +[2026-06-24 03:34:34,286][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.017132433131337166, 'train/grad_norm_pre_clip': 5.466944694519043, 'timing/data': 0.001419512991560623, 'timing/model': 1.1804199909965973, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 1.59}) +[2026-06-24 03:34:38,591][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.005502372048795223, 'train/grad_norm_pre_clip': 6.8134284019470215, 'timing/data': 0.0009002739971037954, 'timing/model': 1.176597668993054, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 1.59}) +[2026-06-24 03:34:43,126][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.025573410093784332, 'train/grad_norm_pre_clip': 17.333030700683594, 'timing/data': 0.0007334639958571643, 'timing/model': 1.4317993570002727, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 1.59}) +[2026-06-24 03:34:47,390][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.12509216368198395, 'train/grad_norm_pre_clip': 24.86848258972168, 'timing/data': 0.0007321129960473627, 'timing/model': 1.1797022740065586, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 1.59}) +[2026-06-24 03:34:51,698][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.03084826096892357, 'train/grad_norm_pre_clip': 15.342951774597168, 'timing/data': 0.0007583010010421276, 'timing/model': 1.185988568002358, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 1.59}) +[2026-06-24 03:34:55,984][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.004910096991807222, 'train/grad_norm_pre_clip': 16.844099044799805, 'timing/data': 0.0007561869861092418, 'timing/model': 1.1828283600043505, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 1.59}) +[2026-06-24 03:35:00,447][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.039727162569761276, 'train/grad_norm_pre_clip': 27.030916213989258, 'timing/data': 0.0009261610102839768, 'timing/model': 1.1664924200158566, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 1.59}) +[2026-06-24 03:35:04,721][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.03422454372048378, 'train/grad_norm_pre_clip': 33.75528335571289, 'timing/data': 0.0007658539980184287, 'timing/model': 1.1680197110108566, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 1.59}) +[2026-06-24 03:35:09,051][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.03597250580787659, 'train/grad_norm_pre_clip': 19.46318817138672, 'timing/data': 0.0007614959904458374, 'timing/model': 1.173327166994568, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 1.59}) +[2026-06-24 03:35:13,363][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.010877395048737526, 'train/grad_norm_pre_clip': 21.2683048248291, 'timing/data': 0.0009402369905728847, 'timing/model': 1.1852593299990986, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 1.59}) +[2026-06-24 03:35:17,898][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.00010061842476716265, 'train/grad_norm_pre_clip': 9.782881736755371, 'timing/data': 0.0008854049956426024, 'timing/model': 1.1882560130034108, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 1.59}) +[2026-06-24 03:35:22,221][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.1052151694893837, 'train/grad_norm_pre_clip': 35.62447738647461, 'timing/data': 0.0007735580147709697, 'timing/model': 1.1876101619855035, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 1.59}) +[2026-06-24 03:35:26,551][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.015458540059626102, 'train/grad_norm_pre_clip': 34.097755432128906, 'timing/data': 0.0008106260211206973, 'timing/model': 1.188534852990415, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 1.59}) +[2026-06-24 03:35:30,854][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.0020982339046895504, 'train/grad_norm_pre_clip': 34.19649887084961, 'timing/data': 0.00091840501409024, 'timing/model': 1.1878424290043768, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 1.59}) +[2026-06-24 03:35:35,399][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.00035308190854266286, 'train/grad_norm_pre_clip': 22.895105361938477, 'timing/data': 0.0009665739780757576, 'timing/model': 1.2000778389920015, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 1.59}) +[2026-06-24 03:35:39,762][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.03375778719782829, 'train/grad_norm_pre_clip': 28.946657180786133, 'timing/data': 0.0007388840022031218, 'timing/model': 1.1930982090125326, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 1.59}) +[2026-06-24 03:35:44,119][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.0427427738904953, 'train/grad_norm_pre_clip': 32.992713928222656, 'timing/data': 0.0007870430126786232, 'timing/model': 1.1729022439976688, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 1.59}) +[2026-06-24 03:35:48,476][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.06295513361692429, 'train/grad_norm_pre_clip': 26.944849014282227, 'timing/data': 0.0007753519748803228, 'timing/model': 1.1932660870079417, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 1.59}) +[2026-06-24 03:35:52,972][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.1909741461277008, 'train/grad_norm_pre_clip': 48.806888580322266, 'timing/data': 0.0007207800226751715, 'timing/model': 1.1684022739937063, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 1.59}) +[2026-06-24 03:35:57,298][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.019308293238282204, 'train/grad_norm_pre_clip': 35.09532928466797, 'timing/data': 0.00080118901678361, 'timing/model': 1.1905037550022826, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 1.59}) +[2026-06-24 03:36:01,607][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.02797549217939377, 'train/grad_norm_pre_clip': 23.674144744873047, 'timing/data': 0.0009629680134821683, 'timing/model': 1.1888560569786932, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 1.59}) +[2026-06-24 03:36:06,190][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.0010392977856099606, 'train/grad_norm_pre_clip': 14.950743675231934, 'timing/data': 0.0007053420122247189, 'timing/model': 1.4160894679953344, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 1.6}) +[2026-06-24 03:36:10,511][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.029440339654684067, 'train/grad_norm_pre_clip': 18.3173885345459, 'timing/data': 0.0008320260094478726, 'timing/model': 1.1699733469868079, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 1.6}) +[2026-06-24 03:36:14,892][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.06428255140781403, 'train/grad_norm_pre_clip': 32.32884216308594, 'timing/data': 0.0007814319978933781, 'timing/model': 1.1949006150243804, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 1.6}) +[2026-06-24 03:36:19,190][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.0008058203966356814, 'train/grad_norm_pre_clip': 23.678789138793945, 'timing/data': 0.0008795440080575645, 'timing/model': 1.1747948529955465, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 1.6}) +[2026-06-24 03:36:23,753][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.05864044278860092, 'train/grad_norm_pre_clip': 29.592632293701172, 'timing/data': 0.0007857400050852448, 'timing/model': 1.4083622060134076, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 1.6}) +[2026-06-24 03:36:28,076][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.14800067245960236, 'train/grad_norm_pre_clip': 30.84417152404785, 'timing/data': 0.0007684490119572729, 'timing/model': 1.1730544759775512, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 1.6}) +[2026-06-24 03:36:32,434][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.007038129959255457, 'train/grad_norm_pre_clip': 16.56138801574707, 'timing/data': 0.0008210550004150718, 'timing/model': 1.2038090220012236, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 1.6}) +[2026-06-24 03:36:36,808][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.1032920852303505, 'train/grad_norm_pre_clip': 22.10076141357422, 'timing/data': 0.0010899920016527176, 'timing/model': 1.207779845019104, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 1.6}) +[2026-06-24 03:36:41,392][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.001200644881464541, 'train/grad_norm_pre_clip': 8.773006439208984, 'timing/data': 0.0012218960036989301, 'timing/model': 1.4217949740123004, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 1.6}) +[2026-06-24 03:36:45,705][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.04874240234494209, 'train/grad_norm_pre_clip': 22.328054428100586, 'timing/data': 0.0008080920088104904, 'timing/model': 1.1727038279932458, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 1.6}) +[2026-06-24 03:36:50,071][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.047888707369565964, 'train/grad_norm_pre_clip': 19.298524856567383, 'timing/data': 0.0007845289946999401, 'timing/model': 1.1978829620056786, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 1.6}) +[2026-06-24 03:36:54,450][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.003520800732076168, 'train/grad_norm_pre_clip': 10.57680892944336, 'timing/data': 0.0007607839943375438, 'timing/model': 1.203512267995393, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 1.6}) +[2026-06-24 03:36:58,980][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.021938608959317207, 'train/grad_norm_pre_clip': 15.729687690734863, 'timing/data': 0.0007397549925372005, 'timing/model': 1.1715721949876752, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 1.6}) +[2026-06-24 03:37:03,295][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.005950799211859703, 'train/grad_norm_pre_clip': 4.674271583557129, 'timing/data': 0.0007477600011043251, 'timing/model': 1.192799673997797, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 1.6}) +[2026-06-24 03:37:07,609][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.11711833626031876, 'train/grad_norm_pre_clip': 33.47632598876953, 'timing/data': 0.000754483015043661, 'timing/model': 1.194083156005945, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 1.6}) +[2026-06-24 03:37:11,963][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.0605686716735363, 'train/grad_norm_pre_clip': 19.499897003173828, 'timing/data': 0.0009748300071805716, 'timing/model': 1.1989876620064024, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 1.6}) +[2026-06-24 03:37:16,593][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.030670125037431717, 'train/grad_norm_pre_clip': 31.73748016357422, 'timing/data': 0.0007293960079550743, 'timing/model': 1.1932709400134627, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 1.6}) +[2026-06-24 03:37:20,910][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.06780000776052475, 'train/grad_norm_pre_clip': 39.68748474121094, 'timing/data': 0.0007493129814974964, 'timing/model': 1.1739714930008631, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 1.6}) +[2026-06-24 03:37:25,215][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.10327199846506119, 'train/grad_norm_pre_clip': 17.46014976501465, 'timing/data': 0.0008946410089265555, 'timing/model': 1.178225597977871, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 1.6}) +[2026-06-24 03:37:29,602][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.0017745961667969823, 'train/grad_norm_pre_clip': 17.488880157470703, 'timing/data': 0.0009657340124249458, 'timing/model': 1.1961517129966523, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 1.6}) +[2026-06-24 03:37:34,169][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.0024879388511180878, 'train/grad_norm_pre_clip': 23.891233444213867, 'timing/data': 0.0007792990072630346, 'timing/model': 1.1845699130208232, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 1.6}) +[2026-06-24 03:37:38,538][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.1837466061115265, 'train/grad_norm_pre_clip': 21.719675064086914, 'timing/data': 0.0007689499761909246, 'timing/model': 1.174250491982093, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 1.6}) +[2026-06-24 03:37:42,923][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.015416771173477173, 'train/grad_norm_pre_clip': 21.3798828125, 'timing/data': 0.0009274329931940883, 'timing/model': 1.2070152029918972, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 1.6}) +[2026-06-24 03:37:47,275][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.0775776356458664, 'train/grad_norm_pre_clip': 19.095081329345703, 'timing/data': 0.0010379960003774613, 'timing/model': 1.1780271460011136, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 1.6}) +[2026-06-24 03:37:51,797][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.06305775046348572, 'train/grad_norm_pre_clip': 27.128799438476562, 'timing/data': 0.001170690986327827, 'timing/model': 1.203888393996749, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 1.6}) +[2026-06-24 03:37:56,111][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.026722298935055733, 'train/grad_norm_pre_clip': 28.843671798706055, 'timing/data': 0.0009163719951175153, 'timing/model': 1.1781505450198893, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 1.6}) +[2026-06-24 03:38:00,520][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.01265264768153429, 'train/grad_norm_pre_clip': 47.34078598022461, 'timing/data': 0.0009164320072159171, 'timing/model': 1.21541985499789, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 1.61}) +[2026-06-24 03:38:04,887][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.1043948382139206, 'train/grad_norm_pre_clip': 17.9013729095459, 'timing/data': 0.0007541230006609112, 'timing/model': 1.1924688720027916, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 1.61}) +[2026-06-24 03:38:09,473][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.07458660006523132, 'train/grad_norm_pre_clip': 26.90926742553711, 'timing/data': 0.0009912399982567877, 'timing/model': 1.1744326839980204, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 1.61}) +[2026-06-24 03:38:13,808][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.035391900688409805, 'train/grad_norm_pre_clip': 10.397375106811523, 'timing/data': 0.0008158069977071136, 'timing/model': 1.2001602839736734, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 1.61}) +[2026-06-24 03:38:18,127][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.009734058752655983, 'train/grad_norm_pre_clip': 11.575400352478027, 'timing/data': 0.0008096849778667092, 'timing/model': 1.1904465200204868, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 1.61}) +[2026-06-24 03:38:22,641][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.023869987577199936, 'train/grad_norm_pre_clip': 15.787798881530762, 'timing/data': 0.0007523389940615743, 'timing/model': 1.3752289099793416, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 1.61}) +[2026-06-24 03:38:26,991][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.061927326023578644, 'train/grad_norm_pre_clip': 15.739195823669434, 'timing/data': 0.0008966460009105504, 'timing/model': 1.1932950890040956, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 1.61}) +[2026-06-24 03:38:31,335][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.06187121197581291, 'train/grad_norm_pre_clip': 17.046361923217773, 'timing/data': 0.0008145550091285259, 'timing/model': 1.1935600290016737, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 1.61}) +[2026-06-24 03:38:35,697][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.12398964911699295, 'train/grad_norm_pre_clip': 26.6728515625, 'timing/data': 0.0007850900001358241, 'timing/model': 1.1963433980126865, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 1.61}) +[2026-06-24 03:38:40,269][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.09316734969615936, 'train/grad_norm_pre_clip': 37.0561408996582, 'timing/data': 0.0007618560048285872, 'timing/model': 1.4375940619793255, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 1.61}) +[2026-06-24 03:38:44,620][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.021901370957493782, 'train/grad_norm_pre_clip': 7.1613688468933105, 'timing/data': 0.000757208006689325, 'timing/model': 1.1931294749956578, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 1.61}) +[2026-06-24 03:38:48,980][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.019255077466368675, 'train/grad_norm_pre_clip': 26.763425827026367, 'timing/data': 0.0007684680167585611, 'timing/model': 1.1994830939802341, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 1.61}) +[2026-06-24 03:38:53,392][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.07132977992296219, 'train/grad_norm_pre_clip': 21.79521369934082, 'timing/data': 0.0011439319932833314, 'timing/model': 1.2070709539984819, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 1.61}) +[2026-06-24 03:38:57,881][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.0007547186687588692, 'train/grad_norm_pre_clip': 16.428285598754883, 'timing/data': 0.0008086230081971735, 'timing/model': 1.1843700499739498, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 1.61}) +[2026-06-24 03:39:02,228][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.05487672984600067, 'train/grad_norm_pre_clip': 33.318538665771484, 'timing/data': 0.0007767050119582564, 'timing/model': 1.1937975450127851, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 1.61}) +[2026-06-24 03:39:06,597][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.004065339453518391, 'train/grad_norm_pre_clip': 4.951099872589111, 'timing/data': 0.0007560959784314036, 'timing/model': 1.1966447820013855, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 1.61}) +[2026-06-24 03:39:10,957][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.032905213534832, 'train/grad_norm_pre_clip': 22.453548431396484, 'timing/data': 0.000979798991465941, 'timing/model': 1.200342198018916, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 1.61}) +[2026-06-24 03:39:15,477][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.06568058580160141, 'train/grad_norm_pre_clip': 14.769730567932129, 'timing/data': 0.0007881160126999021, 'timing/model': 1.1945985030033626, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 1.61}) +[2026-06-24 03:39:19,845][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.008836070075631142, 'train/grad_norm_pre_clip': 14.330662727355957, 'timing/data': 0.0009025369945447892, 'timing/model': 1.199063156003831, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 1.61}) +[2026-06-24 03:39:24,192][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.03077121078968048, 'train/grad_norm_pre_clip': 28.23346710205078, 'timing/data': 0.0007926929974928498, 'timing/model': 1.175226171995746, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 1.61}) +[2026-06-24 03:39:28,550][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.0034492199774831533, 'train/grad_norm_pre_clip': 13.816596031188965, 'timing/data': 0.0011202990135643631, 'timing/model': 1.1788566450122744, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 1.61}) +[2026-06-24 03:39:33,137][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.04963713139295578, 'train/grad_norm_pre_clip': 26.24720001220703, 'timing/data': 0.0009315299976151437, 'timing/model': 1.2051239630091004, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 1.61}) +[2026-06-24 03:39:37,519][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.07422786206007004, 'train/grad_norm_pre_clip': 34.540733337402344, 'timing/data': 0.0008162370068021119, 'timing/model': 1.1811598950007465, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 1.61}) +[2026-06-24 03:39:41,905][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.011863208375871181, 'train/grad_norm_pre_clip': 12.559070587158203, 'timing/data': 0.0009573280112817883, 'timing/model': 1.202298635005718, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 1.61}) +[2026-06-24 03:39:46,279][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.02938997745513916, 'train/grad_norm_pre_clip': 20.491512298583984, 'timing/data': 0.0009916409908328205, 'timing/model': 1.1933405529998709, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 1.61}) +[2026-06-24 03:39:50,849][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.11456382274627686, 'train/grad_norm_pre_clip': 19.310720443725586, 'timing/data': 0.0007426220108754933, 'timing/model': 1.1964881639869418, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 1.62}) +[2026-06-24 03:39:55,229][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.004002970643341541, 'train/grad_norm_pre_clip': 36.946510314941406, 'timing/data': 0.0007378029986284673, 'timing/model': 1.198649263009429, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 1.62}) +[2026-06-24 03:39:59,605][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.010669618844985962, 'train/grad_norm_pre_clip': 8.188214302062988, 'timing/data': 0.001098659005947411, 'timing/model': 1.1968620420084335, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 1.62}) +[2026-06-24 03:40:03,978][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.0747503712773323, 'train/grad_norm_pre_clip': 18.412870407104492, 'timing/data': 0.0007795790152158588, 'timing/model': 1.2030532310018316, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 1.62}) +[2026-06-24 03:40:08,523][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.00565352663397789, 'train/grad_norm_pre_clip': 54.11915588378906, 'timing/data': 0.000867400987772271, 'timing/model': 1.1799704690056387, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 1.62}) +[2026-06-24 03:40:12,863][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.043607838451862335, 'train/grad_norm_pre_clip': 29.062423706054688, 'timing/data': 0.0008709380053915083, 'timing/model': 1.1825613009859808, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 1.62}) +[2026-06-24 03:40:17,219][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.07696673274040222, 'train/grad_norm_pre_clip': 23.20452880859375, 'timing/data': 0.0008657279831822962, 'timing/model': 1.1944747870147694, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 1.62}) +[2026-06-24 03:40:21,881][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.2937828004360199, 'train/grad_norm_pre_clip': 33.85383224487305, 'timing/data': 0.000860858999658376, 'timing/model': 1.4990271270216908, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 1.62}) +[2026-06-24 03:40:26,239][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.0042945328168570995, 'train/grad_norm_pre_clip': 27.688352584838867, 'timing/data': 0.001195438002469018, 'timing/model': 1.1904799269977957, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 1.62}) +[2026-06-24 03:40:30,552][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.01256694458425045, 'train/grad_norm_pre_clip': 26.29458236694336, 'timing/data': 0.0007466289971489459, 'timing/model': 1.173654946993338, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 1.62}) +[2026-06-24 03:40:34,884][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.01938331127166748, 'train/grad_norm_pre_clip': 6.371468544006348, 'timing/data': 0.0008476850052829832, 'timing/model': 1.1796684990113135, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 1.62}) +[2026-06-24 03:40:39,483][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.01991862989962101, 'train/grad_norm_pre_clip': 20.48552131652832, 'timing/data': 0.0010363539913669229, 'timing/model': 1.4416645070014056, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 1.62}) +[2026-06-24 03:40:43,822][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.07567897439002991, 'train/grad_norm_pre_clip': 36.704986572265625, 'timing/data': 0.0008681030012667179, 'timing/model': 1.1767662960046437, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 1.62}) +[2026-06-24 03:40:48,156][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.011828717775642872, 'train/grad_norm_pre_clip': 5.732144832611084, 'timing/data': 0.0007384239870589226, 'timing/model': 1.1816053990041837, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 1.62}) +[2026-06-24 03:40:52,520][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.06353452056646347, 'train/grad_norm_pre_clip': 22.777128219604492, 'timing/data': 0.0007615259964950383, 'timing/model': 1.180364034982631, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 1.62}) +[2026-06-24 03:40:57,009][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.0015338794328272343, 'train/grad_norm_pre_clip': 11.4849271774292, 'timing/data': 0.0007455369923263788, 'timing/model': 1.1781542619864922, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 1.62}) +[2026-06-24 03:41:01,370][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.04332337528467178, 'train/grad_norm_pre_clip': 30.22160530090332, 'timing/data': 0.0007280939898919314, 'timing/model': 1.199791863997234, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 1.62}) +[2026-06-24 03:41:05,736][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.011147474870085716, 'train/grad_norm_pre_clip': 15.57629108428955, 'timing/data': 0.0007304789905901998, 'timing/model': 1.1985023090092, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 1.62}) +[2026-06-24 03:41:10,122][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.005256061442196369, 'train/grad_norm_pre_clip': 14.382481575012207, 'timing/data': 0.0009136770095210522, 'timing/model': 1.2019230339792557, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 1.62}) +[2026-06-24 03:41:14,737][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.05805911496281624, 'train/grad_norm_pre_clip': 15.855223655700684, 'timing/data': 0.0007437329913955182, 'timing/model': 1.1756060059997253, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 1.62}) +[2026-06-24 03:41:19,083][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.007591496221721172, 'train/grad_norm_pre_clip': 21.314096450805664, 'timing/data': 0.0007303080055862665, 'timing/model': 1.1941480910172686, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 1.62}) +[2026-06-24 03:41:23,423][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.06350240856409073, 'train/grad_norm_pre_clip': 24.552854537963867, 'timing/data': 0.0007701219874434173, 'timing/model': 1.1803837119950913, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 1.62}) +[2026-06-24 03:41:27,749][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.08630440384149551, 'train/grad_norm_pre_clip': 19.162078857421875, 'timing/data': 0.0010404110071249306, 'timing/model': 1.185890979017131, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 1.62}) +[2026-06-24 03:41:32,285][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.016706248745322227, 'train/grad_norm_pre_clip': 19.215726852416992, 'timing/data': 0.0008930590120144188, 'timing/model': 1.1769891080039088, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 1.62}) +[2026-06-24 03:41:36,626][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.01938582956790924, 'train/grad_norm_pre_clip': 22.356874465942383, 'timing/data': 0.0007766140042804182, 'timing/model': 1.187256294011604, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 1.62}) +[2026-06-24 03:41:40,986][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.013811741024255753, 'train/grad_norm_pre_clip': 32.47018051147461, 'timing/data': 0.0009949870000127703, 'timing/model': 1.1822876360092778, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 1.62}) +[2026-06-24 03:41:45,340][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.00011384407844161615, 'train/grad_norm_pre_clip': 23.290164947509766, 'timing/data': 0.0008005679992493242, 'timing/model': 1.198904988996219, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 1.63}) +[2026-06-24 03:41:49,987][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4200 +[2026-06-24 03:42:47,498][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4200 +[2026-06-24 03:42:47,499][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4200 +[2026-06-24 03:43:24,716][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4200 +[2026-06-24 03:43:24,717][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.2412569373846054, 'train/grad_norm_pre_clip': 22.546606063842773, 'eval/action_loss/seconds': 57.50790765302372, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03220947086811066, 'eval/latency_2/loss': 0.03220947086811066, 'eval/flappy/loss': 0.03220947086811066, 'eval/flappy/latency_2/loss': 0.03220947086811066, 'eval/flap_precision': 0.9783783783783784, 'eval/flap_recall': 0.9823609226594301, 'eval/flap_f1': 0.980365605958023, 'eval/flap_rate_model': 0.20555555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9803656059580231, 'eval/latency_2/flap_precision': 0.9783783783783784, 'eval/latency_2/flap_recall': 0.9823609226594301, 'eval/latency_2/flap_f1': 0.980365605958023, 'eval/latency_2/flap_rate_model': 0.20555555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9803656059580231, 'eval/action_classification/data_wait_seconds': 0.5437554121017456, 'eval/action_classification/predict_seconds': 36.52952194213867, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.000998553994577378, 'timing/model': 1.4036496399785392, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 1.63}) +[2026-06-24 03:43:29,034][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.15276500582695007, 'train/grad_norm_pre_clip': 34.634090423583984, 'timing/data': 0.0007939560164231807, 'timing/model': 1.1593492469983175, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 1.63}) +[2026-06-24 03:43:33,361][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.016199788078665733, 'train/grad_norm_pre_clip': 5.999310493469238, 'timing/data': 0.000710953027009964, 'timing/model': 1.1936154940049164, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 1.63}) +[2026-06-24 03:43:37,677][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.03341624140739441, 'train/grad_norm_pre_clip': 28.598955154418945, 'timing/data': 0.000787694996688515, 'timing/model': 1.1867921680095606, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 1.63}) +[2026-06-24 03:43:42,293][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.008522629737854004, 'train/grad_norm_pre_clip': 11.3225679397583, 'timing/data': 0.0012662480003200471, 'timing/model': 1.1718868780008052, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 1.63}) +[2026-06-24 03:43:46,622][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.05925429239869118, 'train/grad_norm_pre_clip': 21.146377563476562, 'timing/data': 0.0011195880069863051, 'timing/model': 1.1954218359896913, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 1.63}) +[2026-06-24 03:43:50,956][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.0030214006546884775, 'train/grad_norm_pre_clip': 10.263518333435059, 'timing/data': 0.0009819930128287524, 'timing/model': 1.1981160969880875, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 1.63}) +[2026-06-24 03:43:55,474][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.002829865086823702, 'train/grad_norm_pre_clip': 15.125937461853027, 'timing/data': 0.0012802339915651828, 'timing/model': 1.3607156240032054, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 1.63}) +[2026-06-24 03:43:59,824][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.042255811393260956, 'train/grad_norm_pre_clip': 16.46690559387207, 'timing/data': 0.001157127000624314, 'timing/model': 1.1915023430192377, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 1.63}) +[2026-06-24 03:44:04,135][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.029754934832453728, 'train/grad_norm_pre_clip': 28.30124282836914, 'timing/data': 0.0008435779891442508, 'timing/model': 1.1708855440083425, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 1.63}) +[2026-06-24 03:44:08,449][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.014029049314558506, 'train/grad_norm_pre_clip': 9.729278564453125, 'timing/data': 0.0008019410015549511, 'timing/model': 1.194124076020671, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 1.63}) +[2026-06-24 03:44:13,023][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.02636372298002243, 'train/grad_norm_pre_clip': 13.725381851196289, 'timing/data': 0.001036173984175548, 'timing/model': 1.4156312449777033, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 1.63}) +[2026-06-24 03:44:17,370][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.2238377183675766, 'train/grad_norm_pre_clip': 36.5704345703125, 'timing/data': 0.0007867030217312276, 'timing/model': 1.177874987013638, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 1.63}) +[2026-06-24 03:44:21,707][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.023454705253243446, 'train/grad_norm_pre_clip': 12.991888999938965, 'timing/data': 0.0007498449995182455, 'timing/model': 1.175587777979672, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 1.63}) +[2026-06-24 03:44:26,022][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.0005205611232668161, 'train/grad_norm_pre_clip': 14.649178504943848, 'timing/data': 0.0009231750154867768, 'timing/model': 1.1719978540204465, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 1.63}) +[2026-06-24 03:44:30,534][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.05902174115180969, 'train/grad_norm_pre_clip': 46.33497619628906, 'timing/data': 0.0007517989724874496, 'timing/model': 1.1770922869909555, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 1.63}) +[2026-06-24 03:44:34,883][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.024286184459924698, 'train/grad_norm_pre_clip': 16.389217376708984, 'timing/data': 0.0007770750089548528, 'timing/model': 1.1879637489910237, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 1.63}) +[2026-06-24 03:44:39,237][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.017368219792842865, 'train/grad_norm_pre_clip': 15.640789985656738, 'timing/data': 0.0008948020113166422, 'timing/model': 1.2031597180175595, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 1.63}) +[2026-06-24 03:44:43,581][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.010126376524567604, 'train/grad_norm_pre_clip': 10.106757164001465, 'timing/data': 0.0009771650074981153, 'timing/model': 1.1929210940143093, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 1.63}) +[2026-06-24 03:44:48,065][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.0678190365433693, 'train/grad_norm_pre_clip': 23.633861541748047, 'timing/data': 0.0009836659883148968, 'timing/model': 1.1731048820074648, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 1.63}) +[2026-06-24 03:44:52,413][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.09396172314882278, 'train/grad_norm_pre_clip': 30.1273250579834, 'timing/data': 0.0007972519961185753, 'timing/model': 1.19845649600029, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 1.63}) +[2026-06-24 03:44:56,757][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.01688995398581028, 'train/grad_norm_pre_clip': 18.247591018676758, 'timing/data': 0.0008367859991267323, 'timing/model': 1.1784512549929786, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 1.63}) +[2026-06-24 03:45:01,087][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.2086523473262787, 'train/grad_norm_pre_clip': 25.980632781982422, 'timing/data': 0.000896696001291275, 'timing/model': 1.1974011570273433, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 1.63}) +[2026-06-24 03:45:05,698][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.0024862191639840603, 'train/grad_norm_pre_clip': 21.64976692199707, 'timing/data': 0.0007611860055476427, 'timing/model': 1.1959765709761996, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 1.63}) +[2026-06-24 03:45:10,062][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.02615015022456646, 'train/grad_norm_pre_clip': 37.26338577270508, 'timing/data': 0.0007832669944036752, 'timing/model': 1.2039058789960109, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 1.63}) +[2026-06-24 03:45:14,422][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.03863916173577309, 'train/grad_norm_pre_clip': 14.102983474731445, 'timing/data': 0.0013200179964769632, 'timing/model': 1.2007404679898173, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 1.64}) +[2026-06-24 03:45:18,761][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.0033777826465666294, 'train/grad_norm_pre_clip': 55.15913391113281, 'timing/data': 0.0007853300194256008, 'timing/model': 1.199720619013533, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 1.64}) +[2026-06-24 03:45:23,299][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.006469503976404667, 'train/grad_norm_pre_clip': 48.58278274536133, 'timing/data': 0.0011064330174122006, 'timing/model': 1.2088989250187296, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 1.64}) +[2026-06-24 03:45:27,674][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.014730287715792656, 'train/grad_norm_pre_clip': 14.516424179077148, 'timing/data': 0.0007982150127645582, 'timing/model': 1.178595042991219, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 1.64}) +[2026-06-24 03:45:32,030][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.002964104525744915, 'train/grad_norm_pre_clip': 24.53028678894043, 'timing/data': 0.0007761639717500657, 'timing/model': 1.177872412983561, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 1.64}) +[2026-06-24 03:45:36,446][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.06931713223457336, 'train/grad_norm_pre_clip': 27.669536590576172, 'timing/data': 0.0008011499885469675, 'timing/model': 1.1763083790137898, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 1.64}) +[2026-06-24 03:45:40,975][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.06337645649909973, 'train/grad_norm_pre_clip': 29.75138282775879, 'timing/data': 0.0007538619975093752, 'timing/model': 1.3915730649896432, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 1.64}) +[2026-06-24 03:45:45,382][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.06825929135084152, 'train/grad_norm_pre_clip': 21.00553321838379, 'timing/data': 0.0010036029852926731, 'timing/model': 1.1987391139846295, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 1.64}) +[2026-06-24 03:45:49,726][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.0835605040192604, 'train/grad_norm_pre_clip': 28.41121482849121, 'timing/data': 0.0007425310031976551, 'timing/model': 1.1796180040109903, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 1.64}) +[2026-06-24 03:45:54,067][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.126798614859581, 'train/grad_norm_pre_clip': 22.819849014282227, 'timing/data': 0.0008859559893608093, 'timing/model': 1.1829526069923304, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 1.64}) +[2026-06-24 03:45:58,405][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.0992816910147667, 'train/grad_norm_pre_clip': 18.406864166259766, 'timing/data': 0.0007933549932204187, 'timing/model': 1.2005153160134796, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 1.64}) +[2026-06-24 03:46:03,010][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.042871374636888504, 'train/grad_norm_pre_clip': 16.968223571777344, 'timing/data': 0.0007654929941054434, 'timing/model': 1.1821553619811311, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 1.64}) +[2026-06-24 03:46:07,377][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.03713946416974068, 'train/grad_norm_pre_clip': 20.081247329711914, 'timing/data': 0.000933383998926729, 'timing/model': 1.180181850999361, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 1.64}) +[2026-06-24 03:46:11,737][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.10715404897928238, 'train/grad_norm_pre_clip': 27.117328643798828, 'timing/data': 0.0008278889872599393, 'timing/model': 1.1943014030111954, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 1.64}) +[2026-06-24 03:46:16,084][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.031662724912166595, 'train/grad_norm_pre_clip': 19.47884750366211, 'timing/data': 0.0011239659797865897, 'timing/model': 1.181387506017927, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 1.64}) +[2026-06-24 03:46:20,646][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.020686831325292587, 'train/grad_norm_pre_clip': 22.434162139892578, 'timing/data': 0.0008147949993144721, 'timing/model': 1.1977667009923607, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 1.64}) +[2026-06-24 03:46:24,981][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.04425591602921486, 'train/grad_norm_pre_clip': 4.802817344665527, 'timing/data': 0.0007685000018682331, 'timing/model': 1.1783248680003453, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 1.64}) +[2026-06-24 03:46:29,299][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.09881556779146194, 'train/grad_norm_pre_clip': 24.98341941833496, 'timing/data': 0.0008091240015346557, 'timing/model': 1.177112172008492, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 1.64}) +[2026-06-24 03:46:33,646][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.0019327119225636125, 'train/grad_norm_pre_clip': 6.457113742828369, 'timing/data': 0.0009214720048476011, 'timing/model': 1.194251451001037, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 1.64}) +[2026-06-24 03:46:38,219][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.0031033207196742296, 'train/grad_norm_pre_clip': 26.431798934936523, 'timing/data': 0.0008462530095130205, 'timing/model': 1.203916644997662, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 1.64}) +[2026-06-24 03:46:42,543][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.03572801873087883, 'train/grad_norm_pre_clip': 16.854686737060547, 'timing/data': 0.0007900590135250241, 'timing/model': 1.17602079801145, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 1.64}) +[2026-06-24 03:46:46,871][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.03363931551575661, 'train/grad_norm_pre_clip': 7.822227478027344, 'timing/data': 0.0007471400022041053, 'timing/model': 1.1780002480081748, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 1.64}) +[2026-06-24 03:46:51,246][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.010547392070293427, 'train/grad_norm_pre_clip': 21.618371963500977, 'timing/data': 0.0008160170109476894, 'timing/model': 1.1795899770222604, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 1.64}) +[2026-06-24 03:46:55,756][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.06262867152690887, 'train/grad_norm_pre_clip': 21.144634246826172, 'timing/data': 0.0009103120246436447, 'timing/model': 1.1807675069721881, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 1.64}) +[2026-06-24 03:47:00,099][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.04167795553803444, 'train/grad_norm_pre_clip': 26.032751083374023, 'timing/data': 0.0010894819861277938, 'timing/model': 1.179313333996106, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 1.64}) +[2026-06-24 04:15:18,913][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4250 +[2026-06-24 04:15:18,914][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.09074152261018753, 'train/grad_norm_pre_clip': 51.732826232910156, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000804515992058441, 'timing/model': 1.1777294630010147, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 1.64}) +[2026-06-24 04:15:23,316][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.006300510838627815, 'train/grad_norm_pre_clip': 13.199694633483887, 'timing/data': 0.0011997370165772736, 'timing/model': 1.3638808710093144, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 1.65}) +[2026-06-24 04:15:27,505][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.08625128120183945, 'train/grad_norm_pre_clip': 27.32451057434082, 'timing/data': 0.0007383740157820284, 'timing/model': 1.1613280010060407, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 1.65}) +[2026-06-24 04:15:31,711][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.10362550616264343, 'train/grad_norm_pre_clip': 19.89406394958496, 'timing/data': 0.0007458179898094386, 'timing/model': 1.1367774849932175, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 1.65}) +[2026-06-24 04:15:35,890][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.04175712168216705, 'train/grad_norm_pre_clip': 11.453075408935547, 'timing/data': 0.0007367310172412544, 'timing/model': 1.1457029430021066, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 1.65}) +[2026-06-24 04:15:40,315][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.05988285690546036, 'train/grad_norm_pre_clip': 12.184410095214844, 'timing/data': 0.0007563880062662065, 'timing/model': 1.3578598809835967, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 1.65}) +[2026-06-24 04:15:44,544][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.061866480857133865, 'train/grad_norm_pre_clip': 21.036096572875977, 'timing/data': 0.0007493350130971521, 'timing/model': 1.172643321013311, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 1.65}) +[2026-06-24 04:15:48,801][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.0005647132056765258, 'train/grad_norm_pre_clip': 5.535937786102295, 'timing/data': 0.0009555759897921234, 'timing/model': 1.171304046001751, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 1.65}) +[2026-06-24 04:15:53,015][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.00010547771671554074, 'train/grad_norm_pre_clip': 6.849693775177002, 'timing/data': 0.0010312660015188158, 'timing/model': 1.1523943899956066, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 1.65}) +[2026-06-24 04:15:57,476][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.005979676730930805, 'train/grad_norm_pre_clip': 7.693903923034668, 'timing/data': 0.0007603040139656514, 'timing/model': 1.3688564900076017, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 1.65}) +[2026-06-24 04:16:01,757][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.01566099375486374, 'train/grad_norm_pre_clip': 16.10127067565918, 'timing/data': 0.0009315409988630563, 'timing/model': 1.1883570060017519, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 1.65}) +[2026-06-24 04:16:06,013][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.042108312249183655, 'train/grad_norm_pre_clip': 21.273120880126953, 'timing/data': 0.0007577100186608732, 'timing/model': 1.1542443130165339, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 1.65}) +[2026-06-24 04:16:10,294][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.0370483472943306, 'train/grad_norm_pre_clip': 15.62078857421875, 'timing/data': 0.0007921529759187251, 'timing/model': 1.1752511019876692, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 1.65}) +[2026-06-24 04:16:14,747][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.024289177730679512, 'train/grad_norm_pre_clip': 13.808000564575195, 'timing/data': 0.0008179310243576765, 'timing/model': 1.3777083579916507, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 1.65}) +[2026-06-24 04:16:19,058][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.020276160910725594, 'train/grad_norm_pre_clip': 16.506418228149414, 'timing/data': 0.0010537060152273625, 'timing/model': 1.1826321739936247, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 1.65}) +[2026-06-24 04:16:23,351][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.08400417119264603, 'train/grad_norm_pre_clip': 27.83517837524414, 'timing/data': 0.0011170429934281856, 'timing/model': 1.1876021150092129, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 1.65}) +[2026-06-24 04:16:27,629][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.0013103308156132698, 'train/grad_norm_pre_clip': 15.420072555541992, 'timing/data': 0.0013178250228520483, 'timing/model': 1.1614389080205, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 1.65}) +[2026-06-24 04:16:32,124][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.008037645369768143, 'train/grad_norm_pre_clip': 13.628933906555176, 'timing/data': 0.0012074410042259842, 'timing/model': 1.3697979380085599, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 1.65}) +[2026-06-24 04:16:36,471][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.023718317970633507, 'train/grad_norm_pre_clip': 20.487651824951172, 'timing/data': 0.0009065550111699849, 'timing/model': 1.194418224011315, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 1.65}) +[2026-06-24 04:16:40,785][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.0003699831140693277, 'train/grad_norm_pre_clip': 17.721519470214844, 'timing/data': 0.0007403279887512326, 'timing/model': 1.1735580640088301, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 1.65}) +[2026-06-24 04:16:45,049][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.007135601248592138, 'train/grad_norm_pre_clip': 11.574012756347656, 'timing/data': 0.0007269429916050285, 'timing/model': 1.1635983930027578, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 1.65}) +[2026-06-24 04:16:49,525][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.032767899334430695, 'train/grad_norm_pre_clip': 19.0957088470459, 'timing/data': 0.0007586819992866367, 'timing/model': 1.357719838997582, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 1.65}) +[2026-06-24 04:16:53,788][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.008703787811100483, 'train/grad_norm_pre_clip': 19.666778564453125, 'timing/data': 0.0007285649771802127, 'timing/model': 1.17085391501314, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 1.65}) +[2026-06-24 04:16:58,075][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.04073440283536911, 'train/grad_norm_pre_clip': 21.133983612060547, 'timing/data': 0.0007695620006415993, 'timing/model': 1.1645341990224551, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 1.65}) +[2026-06-24 04:17:02,367][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.0023817946203052998, 'train/grad_norm_pre_clip': 53.20254135131836, 'timing/data': 0.0007626189908478409, 'timing/model': 1.1675512210058514, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 1.65}) +[2026-06-24 04:17:06,922][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.27496373653411865, 'train/grad_norm_pre_clip': 35.20798110961914, 'timing/data': 0.001079153997125104, 'timing/model': 1.4317044949857518, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 1.65}) +[2026-06-24 04:17:11,251][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.09845640510320663, 'train/grad_norm_pre_clip': 28.434972763061523, 'timing/data': 0.0009359990071970969, 'timing/model': 1.1821790429821704, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 1.65}) +[2026-06-24 04:17:15,569][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.08362365514039993, 'train/grad_norm_pre_clip': 25.363317489624023, 'timing/data': 0.0007921830110717565, 'timing/model': 1.1667480540054385, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 1.66}) +[2026-06-24 04:17:19,916][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.004511493723839521, 'train/grad_norm_pre_clip': 15.293975830078125, 'timing/data': 0.0007796200225129724, 'timing/model': 1.1926535810052883, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 1.66}) +[2026-06-24 04:17:24,412][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.10043452680110931, 'train/grad_norm_pre_clip': 23.332107543945312, 'timing/data': 0.0009925439953804016, 'timing/model': 1.1865579790028278, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 1.66}) +[2026-06-24 04:17:28,743][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.02771845832467079, 'train/grad_norm_pre_clip': 14.282092094421387, 'timing/data': 0.0008929899777285755, 'timing/model': 1.1889289279934019, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 1.66}) +[2026-06-24 04:17:33,049][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.018402989953756332, 'train/grad_norm_pre_clip': 19.128543853759766, 'timing/data': 0.0007611860055476427, 'timing/model': 1.1706287690030877, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 1.66}) +[2026-06-24 04:17:37,360][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.06084655597805977, 'train/grad_norm_pre_clip': 16.031482696533203, 'timing/data': 0.0008775419846642762, 'timing/model': 1.1925161330145784, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 1.66}) +[2026-06-24 04:17:41,872][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.0001352238468825817, 'train/grad_norm_pre_clip': 11.302695274353027, 'timing/data': 0.0007142789836507291, 'timing/model': 1.1879166739818174, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 1.66}) +[2026-06-24 04:17:46,216][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.11352260410785675, 'train/grad_norm_pre_clip': 16.25159454345703, 'timing/data': 0.0007221049745567143, 'timing/model': 1.1890192300197668, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 1.66}) +[2026-06-24 04:17:50,520][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.019407639279961586, 'train/grad_norm_pre_clip': 7.413219451904297, 'timing/data': 0.000728134997189045, 'timing/model': 1.168245762004517, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 1.66}) +[2026-06-24 04:17:54,859][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.21469779312610626, 'train/grad_norm_pre_clip': 19.147598266601562, 'timing/data': 0.00070367002626881, 'timing/model': 1.1891517530020792, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 1.66}) +[2026-06-24 04:17:59,435][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.05926008149981499, 'train/grad_norm_pre_clip': 15.97609806060791, 'timing/data': 0.0007351179956458509, 'timing/model': 1.1691788140160497, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 1.66}) +[2026-06-24 04:18:03,778][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.026646988466382027, 'train/grad_norm_pre_clip': 21.377248764038086, 'timing/data': 0.0006905659975018352, 'timing/model': 1.192692522017751, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 1.66}) +[2026-06-24 04:18:08,091][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.0020387149415910244, 'train/grad_norm_pre_clip': 16.903440475463867, 'timing/data': 0.0007527100096922368, 'timing/model': 1.1990098920068704, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 1.66}) +[2026-06-24 04:18:12,454][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.08252478390932083, 'train/grad_norm_pre_clip': 15.897869110107422, 'timing/data': 0.0006998330063652247, 'timing/model': 1.196488261979539, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 1.66}) +[2026-06-24 04:18:16,945][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.00018849373736884445, 'train/grad_norm_pre_clip': 43.79194259643555, 'timing/data': 0.0007101719966158271, 'timing/model': 1.174736977001885, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 1.66}) +[2026-06-24 04:18:21,276][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.02142820693552494, 'train/grad_norm_pre_clip': 14.115812301635742, 'timing/data': 0.0010516119946260005, 'timing/model': 1.1702926959842443, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 1.66}) +[2026-06-24 04:18:25,588][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.07101564854383469, 'train/grad_norm_pre_clip': 22.18011474609375, 'timing/data': 0.0007302089943550527, 'timing/model': 1.1715324550168589, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 1.66}) +[2026-06-24 04:18:29,945][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.005672950763255358, 'train/grad_norm_pre_clip': 8.061273574829102, 'timing/data': 0.0010020910121966153, 'timing/model': 1.1952888530213386, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 1.66}) +[2026-06-24 04:18:34,435][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.00021624828514177352, 'train/grad_norm_pre_clip': 14.585658073425293, 'timing/data': 0.0007197599916253239, 'timing/model': 1.1739474959904328, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 1.66}) +[2026-06-24 04:18:38,746][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.07403121143579483, 'train/grad_norm_pre_clip': 21.807300567626953, 'timing/data': 0.0007182869885582477, 'timing/model': 1.1724593349790666, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 1.66}) +[2026-06-24 04:18:43,129][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.027117108926177025, 'train/grad_norm_pre_clip': 8.170537948608398, 'timing/data': 0.0012603880022652447, 'timing/model': 1.1963666310184635, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 1.66}) +[2026-06-24 04:18:47,761][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.06193368509411812, 'train/grad_norm_pre_clip': 14.283675193786621, 'timing/data': 0.0007369110244326293, 'timing/model': 1.4563123859988991, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 1.66}) +[2026-06-24 04:18:52,102][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.09037667512893677, 'train/grad_norm_pre_clip': 57.99392318725586, 'timing/data': 0.0007900990021880716, 'timing/model': 1.1758823160198517, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 1.66}) +[2026-06-24 04:18:56,436][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4300 +[2026-06-24 04:19:54,424][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4300 +[2026-06-24 04:19:54,424][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4300 +[2026-06-24 04:20:31,268][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4300 +[2026-06-24 04:20:31,270][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.08317693322896957, 'train/grad_norm_pre_clip': 49.10861587524414, 'eval/action_loss/seconds': 57.98432890398544, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.0325341671705246, 'eval/latency_2/loss': 0.0325341671705246, 'eval/flappy/loss': 0.0325341671705246, 'eval/flappy/latency_2/loss': 0.0325341671705246, 'eval/flap_precision': 0.9863387978142076, 'eval/flap_recall': 0.9796472184531886, 'eval/flap_f1': 0.9829816201497616, 'eval/flap_rate_model': 0.20333333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9829816201497618, 'eval/latency_2/flap_precision': 0.9863387978142076, 'eval/latency_2/flap_recall': 0.9796472184531886, 'eval/latency_2/flap_f1': 0.9829816201497616, 'eval/latency_2/flap_rate_model': 0.20333333333333334, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9829816201497618, 'eval/action_classification/data_wait_seconds': 0.44942474365234375, 'eval/action_classification/predict_seconds': 36.24757766723633, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0009354680078104138, 'timing/model': 1.2002147269959096, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 1.66}) +[2026-06-24 04:20:35,594][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.008035490289330482, 'train/grad_norm_pre_clip': 19.822528839111328, 'timing/data': 0.0010024229995906353, 'timing/model': 1.1643870590196457, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 1.66}) +[2026-06-24 04:20:40,148][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.01561144832521677, 'train/grad_norm_pre_clip': 19.837291717529297, 'timing/data': 0.0007452259887941182, 'timing/model': 1.168689907004591, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 1.66}) +[2026-06-24 04:20:44,468][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.009492420591413975, 'train/grad_norm_pre_clip': 11.189717292785645, 'timing/data': 0.0008177910058293492, 'timing/model': 1.1875329369795509, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 1.67}) +[2026-06-24 04:20:48,774][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.12466590106487274, 'train/grad_norm_pre_clip': 20.433059692382812, 'timing/data': 0.0008190030057448894, 'timing/model': 1.1673631659941748, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 1.67}) +[2026-06-24 04:20:53,050][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.0019719174597412348, 'train/grad_norm_pre_clip': 7.1500139236450195, 'timing/data': 0.0009586910018697381, 'timing/model': 1.1716155899921432, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 1.67}) +[2026-06-24 04:20:57,612][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.04010610654950142, 'train/grad_norm_pre_clip': 15.850970268249512, 'timing/data': 0.0007556659984402359, 'timing/model': 1.1919417589961085, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 1.67}) +[2026-06-24 04:21:01,934][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.04454483091831207, 'train/grad_norm_pre_clip': 13.840553283691406, 'timing/data': 0.0007662349962629378, 'timing/model': 1.183821637998335, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 1.67}) +[2026-06-24 04:21:06,271][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.0006592519930563867, 'train/grad_norm_pre_clip': 27.986499786376953, 'timing/data': 0.001161616004537791, 'timing/model': 1.1981322139909025, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 1.67}) +[2026-06-24 04:21:10,590][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.02516624704003334, 'train/grad_norm_pre_clip': 20.929269790649414, 'timing/data': 0.000918998004635796, 'timing/model': 1.1864036670012865, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 1.67}) +[2026-06-24 04:21:15,129][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.13127873837947845, 'train/grad_norm_pre_clip': 19.70698356628418, 'timing/data': 0.0008418750076089054, 'timing/model': 1.1645730760064907, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 1.67}) +[2026-06-24 04:21:19,416][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.014562033116817474, 'train/grad_norm_pre_clip': 10.280963897705078, 'timing/data': 0.0007266519824042916, 'timing/model': 1.1657136119902134, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 1.67}) +[2026-06-24 04:21:23,755][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.054397501051425934, 'train/grad_norm_pre_clip': 26.084186553955078, 'timing/data': 0.0007246290042530745, 'timing/model': 1.1868212560075335, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 1.67}) +[2026-06-24 04:21:28,032][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.11552296578884125, 'train/grad_norm_pre_clip': 20.05722427368164, 'timing/data': 0.0007415600121021271, 'timing/model': 1.1672086430189665, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 1.67}) +[2026-06-24 04:21:32,646][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.03128911182284355, 'train/grad_norm_pre_clip': 14.44568920135498, 'timing/data': 0.0007215529913082719, 'timing/model': 1.185819353995612, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 1.67}) +[2026-06-24 04:21:36,948][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.02722003683447838, 'train/grad_norm_pre_clip': 9.966907501220703, 'timing/data': 0.0007416099833790213, 'timing/model': 1.1694778640230652, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 1.67}) +[2026-06-24 04:21:41,258][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.08584539592266083, 'train/grad_norm_pre_clip': 23.280990600585938, 'timing/data': 0.0010877999884542078, 'timing/model': 1.1932285600050818, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 1.67}) +[2026-06-24 04:21:45,779][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.06686065346002579, 'train/grad_norm_pre_clip': 37.311344146728516, 'timing/data': 0.001054819003911689, 'timing/model': 1.3461870930041187, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 1.67}) +[2026-06-24 04:21:50,073][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.0027005530428141356, 'train/grad_norm_pre_clip': 12.415716171264648, 'timing/data': 0.0007154210179578513, 'timing/model': 1.1707658620143775, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 1.67}) +[2026-06-24 04:21:54,363][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.006793893873691559, 'train/grad_norm_pre_clip': 14.723490715026855, 'timing/data': 0.0007735999824944884, 'timing/model': 1.1701745890022721, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 1.67}) +[2026-06-24 04:21:58,771][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.0259807538241148, 'train/grad_norm_pre_clip': 28.88852310180664, 'timing/data': 0.0007381440082099289, 'timing/model': 1.1975241229811218, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 1.67}) +[2026-06-24 04:22:03,301][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.01956798881292343, 'train/grad_norm_pre_clip': 7.975196838378906, 'timing/data': 0.0008318369800690562, 'timing/model': 1.3537401119829156, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 1.67}) +[2026-06-24 04:22:07,600][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.002190639730542898, 'train/grad_norm_pre_clip': 23.892780303955078, 'timing/data': 0.0007269830093719065, 'timing/model': 1.1734646749973763, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 1.67}) +[2026-06-24 04:22:11,924][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.036354586482048035, 'train/grad_norm_pre_clip': 19.453588485717773, 'timing/data': 0.0007482529908884317, 'timing/model': 1.172655328991823, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 1.67}) +[2026-06-24 04:22:16,245][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.06204667314887047, 'train/grad_norm_pre_clip': 24.256547927856445, 'timing/data': 0.0007385240169242024, 'timing/model': 1.1733605510089546, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 1.67}) +[2026-06-24 04:22:20,716][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.07297533750534058, 'train/grad_norm_pre_clip': 31.73257827758789, 'timing/data': 0.000836535997223109, 'timing/model': 1.3394971069938038, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 1.67}) +[2026-06-24 04:22:25,037][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.012530378997325897, 'train/grad_norm_pre_clip': 6.798244953155518, 'timing/data': 0.0007008450047578663, 'timing/model': 1.1760024210088886, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 1.67}) +[2026-06-24 04:22:29,356][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.015772150829434395, 'train/grad_norm_pre_clip': 17.137744903564453, 'timing/data': 0.000840792985400185, 'timing/model': 1.1737319650128484, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 1.67}) +[2026-06-24 04:22:33,663][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.0076160491444170475, 'train/grad_norm_pre_clip': 19.2775936126709, 'timing/data': 0.0008293420250993222, 'timing/model': 1.1754289790114854, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 1.68}) +[2026-06-24 04:22:37,978][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.04599125683307648, 'train/grad_norm_pre_clip': 30.33482551574707, 'timing/data': 0.0008129920170176774, 'timing/model': 1.1715910490020178, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 1.68}) +[2026-06-24 04:22:42,539][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.0048285312950611115, 'train/grad_norm_pre_clip': 9.25721549987793, 'timing/data': 0.0009118739981204271, 'timing/model': 1.1781312499952037, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 1.68}) +[2026-06-24 04:22:46,853][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.24296927452087402, 'train/grad_norm_pre_clip': 42.528621673583984, 'timing/data': 0.0007934150053188205, 'timing/model': 1.1765317250101361, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 1.68}) +[2026-06-24 04:22:51,164][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.004986896179616451, 'train/grad_norm_pre_clip': 13.411079406738281, 'timing/data': 0.0007726480253040791, 'timing/model': 1.1760426109831315, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 1.68}) +[2026-06-24 04:22:55,484][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 6.514137203339487e-05, 'train/grad_norm_pre_clip': 23.106050491333008, 'timing/data': 0.0007572890026494861, 'timing/model': 1.1899231229908764, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 1.68}) +[2026-06-24 04:23:00,083][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.05289556086063385, 'train/grad_norm_pre_clip': 25.765853881835938, 'timing/data': 0.0009214929887093604, 'timing/model': 1.1808424470073078, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 1.68}) +[2026-06-24 04:23:04,381][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.06655807793140411, 'train/grad_norm_pre_clip': 13.049503326416016, 'timing/data': 0.000817630992969498, 'timing/model': 1.1711145949957427, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 1.68}) +[2026-06-24 04:23:08,688][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.0010409087408334017, 'train/grad_norm_pre_clip': 27.72279930114746, 'timing/data': 0.0008164280152413994, 'timing/model': 1.173419022001326, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 1.68}) +[2026-06-24 04:23:12,990][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.032897572964429855, 'train/grad_norm_pre_clip': 19.93363380432129, 'timing/data': 0.0007598130032420158, 'timing/model': 1.1750016850128304, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 1.68}) +[2026-06-24 04:23:17,582][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.21496661007404327, 'train/grad_norm_pre_clip': 42.242977142333984, 'timing/data': 0.0007960000075399876, 'timing/model': 1.2028261460072827, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 1.68}) +[2026-06-24 04:23:21,879][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.07218948006629944, 'train/grad_norm_pre_clip': 8.746065139770508, 'timing/data': 0.0007396859873551875, 'timing/model': 1.1725339149998035, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 1.68}) +[2026-06-24 04:23:26,198][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.005978482309728861, 'train/grad_norm_pre_clip': 18.29396629333496, 'timing/data': 0.0007002440106589347, 'timing/model': 1.1732154630008154, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 1.68}) +[2026-06-24 04:23:30,495][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.03354688361287117, 'train/grad_norm_pre_clip': 14.35456657409668, 'timing/data': 0.0008285610238090158, 'timing/model': 1.1706035079841968, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 1.68}) +[2026-06-24 04:23:35,032][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.01937250606715679, 'train/grad_norm_pre_clip': 18.421932220458984, 'timing/data': 0.0008249040110968053, 'timing/model': 1.170955858018715, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 1.68}) +[2026-06-24 04:23:39,326][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.003063623793423176, 'train/grad_norm_pre_clip': 26.736412048339844, 'timing/data': 0.0007633699860889465, 'timing/model': 1.1715820849931333, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 1.68}) +[2026-06-24 04:23:43,684][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.02098645083606243, 'train/grad_norm_pre_clip': 40.4177131652832, 'timing/data': 0.0007482630026061088, 'timing/model': 1.1940534969908185, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 1.68}) +[2026-06-24 04:23:47,989][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.02343132346868515, 'train/grad_norm_pre_clip': 15.568376541137695, 'timing/data': 0.0008520940027665347, 'timing/model': 1.1722711069742218, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 1.68}) +[2026-06-24 04:23:52,546][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.09775073826313019, 'train/grad_norm_pre_clip': 31.78768539428711, 'timing/data': 0.0007375520071946084, 'timing/model': 1.173968270974001, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 1.68}) +[2026-06-24 04:23:56,880][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.01067644264549017, 'train/grad_norm_pre_clip': 13.411554336547852, 'timing/data': 0.0007782379980199039, 'timing/model': 1.1776097679976374, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 1.68}) +[2026-06-24 04:24:01,199][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.15256434679031372, 'train/grad_norm_pre_clip': 12.22899055480957, 'timing/data': 0.0007496750040445477, 'timing/model': 1.1782172599923797, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 1.68}) +[2026-06-24 04:24:05,556][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.04402393847703934, 'train/grad_norm_pre_clip': 52.87839889526367, 'timing/data': 0.000815386010799557, 'timing/model': 1.1941345340164844, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 1.68}) +[2026-06-24 04:24:10,128][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.0014769445406273007, 'train/grad_norm_pre_clip': 18.734943389892578, 'timing/data': 0.0007380039896816015, 'timing/model': 1.1772908570128493, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 1.68}) +[2026-06-24 04:24:14,446][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.06959562748670578, 'train/grad_norm_pre_clip': 14.369709968566895, 'timing/data': 0.0007922239892650396, 'timing/model': 1.1745423990068957, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 1.68}) +[2026-06-24 04:24:18,762][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.009754798375070095, 'train/grad_norm_pre_clip': 11.69993782043457, 'timing/data': 0.0007245790038723499, 'timing/model': 1.1797898619843181, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 1.68}) +[2026-06-24 04:24:23,094][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.053435318171978, 'train/grad_norm_pre_clip': 42.74296569824219, 'timing/data': 0.0008150050125550479, 'timing/model': 1.1779228829836939, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 1.68}) +[2026-06-24 04:24:27,695][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.01171261165291071, 'train/grad_norm_pre_clip': 35.34117126464844, 'timing/data': 0.0007241380226332694, 'timing/model': 1.178785796975717, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 1.69}) +[2026-06-24 04:24:32,101][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.07161761820316315, 'train/grad_norm_pre_clip': 10.805155754089355, 'timing/data': 0.0009356189984828234, 'timing/model': 1.1759945800004061, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 1.69}) +[2026-06-24 04:24:36,422][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.027487117797136307, 'train/grad_norm_pre_clip': 19.343059539794922, 'timing/data': 0.0007237069949042052, 'timing/model': 1.1810067089973018, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 1.69}) +[2026-06-24 04:24:40,747][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.05637563392519951, 'train/grad_norm_pre_clip': 26.346590042114258, 'timing/data': 0.0013110619911458343, 'timing/model': 1.17842292500427, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 1.69}) +[2026-06-24 04:24:45,340][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.00044299091678112745, 'train/grad_norm_pre_clip': 2.8362576961517334, 'timing/data': 0.0008452610054519027, 'timing/model': 1.1753182990069035, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 1.69}) +[2026-06-24 04:24:49,674][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.012815202586352825, 'train/grad_norm_pre_clip': 17.01005744934082, 'timing/data': 0.0008652579854242504, 'timing/model': 1.199699649005197, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 1.69}) +[2026-06-24 04:24:54,002][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.00967087596654892, 'train/grad_norm_pre_clip': 15.476792335510254, 'timing/data': 0.000721522985259071, 'timing/model': 1.1763593909854535, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 1.69}) +[2026-06-24 04:24:58,308][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.02409457415342331, 'train/grad_norm_pre_clip': 22.68243980407715, 'timing/data': 0.0011084580037277192, 'timing/model': 1.1740623579826206, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 1.69}) +[2026-06-24 04:25:02,864][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.029189150780439377, 'train/grad_norm_pre_clip': 13.963748931884766, 'timing/data': 0.0007028680120129138, 'timing/model': 1.1834481359983329, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 1.69}) +[2026-06-24 04:25:07,175][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.04687502235174179, 'train/grad_norm_pre_clip': 17.268526077270508, 'timing/data': 0.0007399869791697711, 'timing/model': 1.1783722690015566, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 1.69}) +[2026-06-24 04:25:11,502][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.06704965978860855, 'train/grad_norm_pre_clip': 33.286537170410156, 'timing/data': 0.0009086579957511276, 'timing/model': 1.1732853910070844, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 1.69}) +[2026-06-24 04:25:15,826][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.00028857641154900193, 'train/grad_norm_pre_clip': 14.316339492797852, 'timing/data': 0.0008462029800284654, 'timing/model': 1.1955498599854764, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 1.69}) +[2026-06-24 04:25:20,385][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.0044692461378872395, 'train/grad_norm_pre_clip': 20.106887817382812, 'timing/data': 0.0007375130080617964, 'timing/model': 1.1742611530062277, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 1.69}) +[2026-06-24 04:25:24,695][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.032591961324214935, 'train/grad_norm_pre_clip': 44.35580062866211, 'timing/data': 0.0010137030039913952, 'timing/model': 1.1752222969953436, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 1.69}) +[2026-06-24 04:25:29,080][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.04835377261042595, 'train/grad_norm_pre_clip': 12.233881950378418, 'timing/data': 0.0009030379878822714, 'timing/model': 1.2055323999957182, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 1.69}) +[2026-06-24 04:25:33,450][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.07156631350517273, 'train/grad_norm_pre_clip': 24.691991806030273, 'timing/data': 0.0007822449842933565, 'timing/model': 1.1942833789798897, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 1.69}) +[2026-06-24 04:25:38,070][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.00837165117263794, 'train/grad_norm_pre_clip': 16.52979278564453, 'timing/data': 0.0007496849866583943, 'timing/model': 1.1995291200000793, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 1.69}) +[2026-06-24 04:25:42,405][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.005693893413990736, 'train/grad_norm_pre_clip': 27.989612579345703, 'timing/data': 0.0007818239973857999, 'timing/model': 1.1767033310025, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 1.69}) +[2026-06-24 04:25:46,778][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.035295743495225906, 'train/grad_norm_pre_clip': 26.28139877319336, 'timing/data': 0.0008227599901147187, 'timing/model': 1.2094655749970116, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 1.69}) +[2026-06-24 04:25:51,157][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.04223991930484772, 'train/grad_norm_pre_clip': 16.652008056640625, 'timing/data': 0.0007851509726606309, 'timing/model': 1.2087455449800473, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 1.69}) +[2026-06-24 04:25:55,757][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 8.329014235641807e-05, 'train/grad_norm_pre_clip': 21.988218307495117, 'timing/data': 0.0006897739949636161, 'timing/model': 1.188492529996438, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 1.69}) +[2026-06-24 04:26:00,057][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.08620377629995346, 'train/grad_norm_pre_clip': 15.275162696838379, 'timing/data': 0.0007234270160552114, 'timing/model': 1.1729231270146556, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 1.69}) +[2026-06-24 04:26:04,360][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.03884461149573326, 'train/grad_norm_pre_clip': 22.01731300354004, 'timing/data': 0.0009416199754923582, 'timing/model': 1.1750511579739396, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 1.69}) +[2026-06-24 04:26:08,718][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.0011830850271508098, 'train/grad_norm_pre_clip': 31.61991310119629, 'timing/data': 0.0008374669996555895, 'timing/model': 1.1810072099906392, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 1.69}) +[2026-06-24 04:26:13,320][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.05980714038014412, 'train/grad_norm_pre_clip': 11.795831680297852, 'timing/data': 0.0008456530049443245, 'timing/model': 1.203786611004034, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 1.69}) +[2026-06-24 04:26:17,680][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.04524562507867813, 'train/grad_norm_pre_clip': 28.88092613220215, 'timing/data': 0.0008421860111411661, 'timing/model': 1.206415257009212, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 1.69}) +[2026-06-24 04:26:22,057][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.15000613033771515, 'train/grad_norm_pre_clip': 45.1279411315918, 'timing/data': 0.0011468689772300422, 'timing/model': 1.1940068330150098, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 1.7}) +[2026-06-24 04:26:26,388][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.10457149147987366, 'train/grad_norm_pre_clip': 32.145240783691406, 'timing/data': 0.0009279740042984486, 'timing/model': 1.1921866710181348, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 1.7}) +[2026-06-24 04:26:30,961][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.054322049021720886, 'train/grad_norm_pre_clip': 11.747135162353516, 'timing/data': 0.0007345270132645965, 'timing/model': 1.1757442450034432, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 1.7}) +[2026-06-24 04:26:35,322][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.06805872917175293, 'train/grad_norm_pre_clip': 21.001779556274414, 'timing/data': 0.0013534000026993454, 'timing/model': 1.1908927799959201, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 1.7}) +[2026-06-24 04:26:39,658][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.0013361758319661021, 'train/grad_norm_pre_clip': 22.442485809326172, 'timing/data': 0.0007355589768849313, 'timing/model': 1.2028169370023534, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 1.7}) +[2026-06-24 04:26:43,978][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.09793028980493546, 'train/grad_norm_pre_clip': 25.066558837890625, 'timing/data': 0.0011413989996071905, 'timing/model': 1.1773286609968636, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 1.7}) +[2026-06-24 04:26:48,586][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.07797912508249283, 'train/grad_norm_pre_clip': 23.676830291748047, 'timing/data': 0.000739647017326206, 'timing/model': 1.1784563240071293, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 1.7}) +[2026-06-24 04:26:52,907][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.03454170376062393, 'train/grad_norm_pre_clip': 19.773075103759766, 'timing/data': 0.0009829759947024286, 'timing/model': 1.1829956489964388, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 1.7}) +[2026-06-24 04:26:57,278][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.08917290717363358, 'train/grad_norm_pre_clip': 24.500286102294922, 'timing/data': 0.000747159996535629, 'timing/model': 1.1827490339928772, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 1.7}) +[2026-06-24 04:27:01,643][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.023608144372701645, 'train/grad_norm_pre_clip': 20.380762100219727, 'timing/data': 0.000734976987587288, 'timing/model': 1.1985892549855635, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 1.7}) +[2026-06-24 04:27:06,254][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.00166526238899678, 'train/grad_norm_pre_clip': 16.44821548461914, 'timing/data': 0.0010243720025755465, 'timing/model': 1.1961187309934758, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 1.7}) +[2026-06-24 04:27:10,595][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.1387973576784134, 'train/grad_norm_pre_clip': 12.503667831420898, 'timing/data': 0.0007386239885818213, 'timing/model': 1.1947865900001489, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 1.7}) +[2026-06-24 04:27:14,927][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.00918150320649147, 'train/grad_norm_pre_clip': 8.593865394592285, 'timing/data': 0.0007586809806525707, 'timing/model': 1.1798715259938035, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 1.7}) +[2026-06-24 04:27:19,256][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.05880693718791008, 'train/grad_norm_pre_clip': 26.519800186157227, 'timing/data': 0.0010115689947269857, 'timing/model': 1.1856642070051748, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 1.7}) +[2026-06-24 04:27:23,863][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.000275027850875631, 'train/grad_norm_pre_clip': 28.77968978881836, 'timing/data': 0.0007429529796354473, 'timing/model': 1.196693821984809, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 1.7}) +[2026-06-24 04:27:28,219][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.006762899458408356, 'train/grad_norm_pre_clip': 98.48322296142578, 'timing/data': 0.0008127109904307872, 'timing/model': 1.1992085670062806, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 1.7}) +[2026-06-24 04:27:32,573][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.04259519279003143, 'train/grad_norm_pre_clip': 15.792339324951172, 'timing/data': 0.0007390750106424093, 'timing/model': 1.1780978589958977, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 1.7}) +[2026-06-24 04:27:36,914][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.14700494706630707, 'train/grad_norm_pre_clip': 38.6764030456543, 'timing/data': 0.0008283199858851731, 'timing/model': 1.1837235809944104, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 1.7}) +[2026-06-24 04:27:41,476][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.013064338825643063, 'train/grad_norm_pre_clip': 28.860309600830078, 'timing/data': 0.0007133680046536028, 'timing/model': 1.1764102700108197, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 1.7}) +[2026-06-24 04:27:45,841][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.014945438131690025, 'train/grad_norm_pre_clip': 20.35426139831543, 'timing/data': 0.000745477998862043, 'timing/model': 1.1985982390178833, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 1.7}) +[2026-06-24 04:27:50,217][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4400 +[2026-06-24 04:28:47,663][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4400 +[2026-06-24 04:28:47,664][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4400 +[2026-06-24 04:29:24,391][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4400 +[2026-06-24 04:29:24,393][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.031427521258592606, 'train/grad_norm_pre_clip': 14.549500465393066, 'eval/action_loss/seconds': 57.44260287901852, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03322717547416687, 'eval/latency_2/loss': 0.03322717547416687, 'eval/flappy/loss': 0.03322717547416687, 'eval/flappy/latency_2/loss': 0.03322717547416687, 'eval/flap_precision': 0.9796472184531886, 'eval/flap_recall': 0.9796472184531886, 'eval/flap_f1': 0.9796472184531886, 'eval/flap_rate_model': 0.20472222222222222, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9796472184531887, 'eval/latency_2/flap_precision': 0.9796472184531886, 'eval/latency_2/flap_recall': 0.9796472184531886, 'eval/latency_2/flap_f1': 0.9796472184531886, 'eval/latency_2/flap_rate_model': 0.20472222222222222, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9796472184531887, 'eval/action_classification/data_wait_seconds': 0.4388006031513214, 'eval/action_classification/predict_seconds': 36.137752532958984, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0009901990124490112, 'timing/model': 1.1937324699829333, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 1.7}) +[2026-06-24 04:29:28,973][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.009043711237609386, 'train/grad_norm_pre_clip': 26.3856201171875, 'timing/data': 0.0007552350289188325, 'timing/model': 1.1650977450190112, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 1.7}) +[2026-06-24 04:29:33,317][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.0876208171248436, 'train/grad_norm_pre_clip': 30.496780395507812, 'timing/data': 0.0008634849800728261, 'timing/model': 1.190961985004833, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 1.7}) +[2026-06-24 04:29:37,604][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.018868647515773773, 'train/grad_norm_pre_clip': 22.42026710510254, 'timing/data': 0.00113163999048993, 'timing/model': 1.1607157989928965, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 1.7}) +[2026-06-24 04:29:41,909][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.05086323246359825, 'train/grad_norm_pre_clip': 23.866239547729492, 'timing/data': 0.0007622780103702098, 'timing/model': 1.1897178960207384, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 1.7}) +[2026-06-24 04:29:46,508][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.030657151713967323, 'train/grad_norm_pre_clip': 15.9049654006958, 'timing/data': 0.0008059389947447926, 'timing/model': 1.1960626460204367, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 1.7}) +[2026-06-24 04:29:50,853][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.0006855707033537328, 'train/grad_norm_pre_clip': 18.18819808959961, 'timing/data': 0.0007800110033713281, 'timing/model': 1.1895222530001774, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 1.71}) +[2026-06-24 04:29:55,159][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.014435065910220146, 'train/grad_norm_pre_clip': 39.4917106628418, 'timing/data': 0.0007460179913323373, 'timing/model': 1.1723983899864834, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 1.71}) +[2026-06-24 04:29:59,630][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.027598153799772263, 'train/grad_norm_pre_clip': 16.741485595703125, 'timing/data': 0.0011768440017476678, 'timing/model': 1.3308956779947039, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 1.71}) +[2026-06-24 04:30:03,964][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.053403206169605255, 'train/grad_norm_pre_clip': 30.620807647705078, 'timing/data': 0.000875987985637039, 'timing/model': 1.191035328025464, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 1.71}) +[2026-06-24 04:30:08,309][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.004087145905941725, 'train/grad_norm_pre_clip': 11.547735214233398, 'timing/data': 0.0008071210177149624, 'timing/model': 1.1900047740200534, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 1.71}) +[2026-06-24 04:30:12,692][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.002500618575140834, 'train/grad_norm_pre_clip': 29.20327377319336, 'timing/data': 0.0008076019876170903, 'timing/model': 1.2083367290033493, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 1.71}) +[2026-06-24 04:30:17,130][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.01598348096013069, 'train/grad_norm_pre_clip': 21.13593292236328, 'timing/data': 0.0007620380201842636, 'timing/model': 1.3276319980213884, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 1.71}) +[2026-06-24 04:30:21,474][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.0210940632969141, 'train/grad_norm_pre_clip': 35.74956512451172, 'timing/data': 0.0011741900234483182, 'timing/model': 1.1815814430010505, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 1.71}) +[2026-06-24 04:30:25,802][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.010044874623417854, 'train/grad_norm_pre_clip': 23.224414825439453, 'timing/data': 0.0008271470142062753, 'timing/model': 1.1738572329923045, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 1.71}) +[2026-06-24 04:30:30,104][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.008788321167230606, 'train/grad_norm_pre_clip': 33.884212493896484, 'timing/data': 0.0007313310052268207, 'timing/model': 1.1768443410110194, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 1.71}) +[2026-06-24 04:30:34,564][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.012901420705020428, 'train/grad_norm_pre_clip': 4.526101589202881, 'timing/data': 0.0009613459988031536, 'timing/model': 1.325758685998153, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 1.71}) +[2026-06-24 04:30:38,868][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.07907096296548843, 'train/grad_norm_pre_clip': 16.0436954498291, 'timing/data': 0.0009312700130976737, 'timing/model': 1.1779912700003479, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 1.71}) +[2026-06-24 04:30:43,185][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.031135156750679016, 'train/grad_norm_pre_clip': 8.612767219543457, 'timing/data': 0.0007890170090831816, 'timing/model': 1.1737821440037806, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 1.71}) +[2026-06-24 04:30:47,504][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.05130115896463394, 'train/grad_norm_pre_clip': 15.258307456970215, 'timing/data': 0.0009343659912701696, 'timing/model': 1.1927544510108419, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 1.71}) +[2026-06-24 04:30:52,018][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.02782423049211502, 'train/grad_norm_pre_clip': 16.250850677490234, 'timing/data': 0.000750817998778075, 'timing/model': 1.3876278480165638, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 1.71}) +[2026-06-24 04:30:56,329][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.03401368111371994, 'train/grad_norm_pre_clip': 20.02281951904297, 'timing/data': 0.001017158996546641, 'timing/model': 1.1744209329772275, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 1.71}) +[2026-06-24 04:31:00,635][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.006490429397672415, 'train/grad_norm_pre_clip': 33.05904769897461, 'timing/data': 0.0007674380030948669, 'timing/model': 1.175703541986877, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 1.71}) +[2026-06-24 04:31:04,952][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.07392942160367966, 'train/grad_norm_pre_clip': 21.84827995300293, 'timing/data': 0.0007605649880133569, 'timing/model': 1.176535528007662, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 1.71}) +[2026-06-24 04:31:09,529][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.005793812219053507, 'train/grad_norm_pre_clip': 10.220181465148926, 'timing/data': 0.0007981139933690429, 'timing/model': 1.4084263289987575, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 1.71}) +[2026-06-24 04:31:13,872][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.012961017899215221, 'train/grad_norm_pre_clip': 14.647579193115234, 'timing/data': 0.0007217239763122052, 'timing/model': 1.1923853479966056, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 1.71}) +[2026-06-24 04:31:18,209][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.1998080462217331, 'train/grad_norm_pre_clip': 27.162424087524414, 'timing/data': 0.0012274369946680963, 'timing/model': 1.1822924930020235, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 1.71}) +[2026-06-24 04:31:22,554][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.021848032251000404, 'train/grad_norm_pre_clip': 11.743123054504395, 'timing/data': 0.0009538620070088655, 'timing/model': 1.1934170040185563, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 1.71}) +[2026-06-24 04:31:27,100][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.042861711233854294, 'train/grad_norm_pre_clip': 20.80057144165039, 'timing/data': 0.0009359489777125418, 'timing/model': 1.4092911260086112, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 1.71}) +[2026-06-24 04:31:31,452][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.03597227483987808, 'train/grad_norm_pre_clip': 9.474933624267578, 'timing/data': 0.0011453759798314422, 'timing/model': 1.1775670629867818, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 1.71}) +[2026-06-24 04:31:35,789][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.07689051330089569, 'train/grad_norm_pre_clip': 29.702518463134766, 'timing/data': 0.000996531016426161, 'timing/model': 1.1793983259995002, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 1.71}) +[2026-06-24 04:31:40,128][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.003597671166062355, 'train/grad_norm_pre_clip': 15.349474906921387, 'timing/data': 0.0008540679991710931, 'timing/model': 1.1804159060120583, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 1.71}) +[2026-06-24 04:31:44,669][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.0029780834447592497, 'train/grad_norm_pre_clip': 8.777514457702637, 'timing/data': 0.0007694920059293509, 'timing/model': 1.4022121870075352, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 1.72}) +[2026-06-24 04:31:49,004][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.0038073984906077385, 'train/grad_norm_pre_clip': 12.932744026184082, 'timing/data': 0.0012988790113013238, 'timing/model': 1.1952686330187134, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 1.72}) +[2026-06-24 04:31:53,355][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.022761378437280655, 'train/grad_norm_pre_clip': 8.957488059997559, 'timing/data': 0.001228570006787777, 'timing/model': 1.1843915590143297, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 1.72}) +[2026-06-24 04:31:57,718][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.05955817550420761, 'train/grad_norm_pre_clip': 34.83535385131836, 'timing/data': 0.0008248839876614511, 'timing/model': 1.1768168960115872, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 1.72}) +[2026-06-24 04:32:02,303][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.026550084352493286, 'train/grad_norm_pre_clip': 6.846768379211426, 'timing/data': 0.0012264960096217692, 'timing/model': 1.203228978003608, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 1.72}) +[2026-06-24 04:32:06,660][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.006908337585628033, 'train/grad_norm_pre_clip': 16.22787857055664, 'timing/data': 0.0008271280094049871, 'timing/model': 1.1919365369831212, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 1.72}) +[2026-06-24 04:32:11,010][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.005114450119435787, 'train/grad_norm_pre_clip': 31.18655776977539, 'timing/data': 0.0009471100056543946, 'timing/model': 1.180937928991625, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 1.72}) +[2026-06-24 04:32:15,360][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.12267209589481354, 'train/grad_norm_pre_clip': 29.292940139770508, 'timing/data': 0.0008004279807209969, 'timing/model': 1.176594643999124, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 1.72}) +[2026-06-24 04:32:19,893][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.002923191525042057, 'train/grad_norm_pre_clip': 27.106477737426758, 'timing/data': 0.0007523300009779632, 'timing/model': 1.173576518020127, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 1.72}) +[2026-06-24 04:32:24,297][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.054880108684301376, 'train/grad_norm_pre_clip': 18.653470993041992, 'timing/data': 0.0007375620189122856, 'timing/model': 1.2138558579899836, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 1.72}) +[2026-06-24 04:32:28,661][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.021821336820721626, 'train/grad_norm_pre_clip': 35.36569595336914, 'timing/data': 0.0008062600099947304, 'timing/model': 1.1955272990162484, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 1.72}) +[2026-06-24 04:32:32,984][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.0005171264056116343, 'train/grad_norm_pre_clip': 14.336536407470703, 'timing/data': 0.0007392459956463426, 'timing/model': 1.178328766982304, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 1.72}) +[2026-06-24 04:32:37,558][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.011977486312389374, 'train/grad_norm_pre_clip': 22.530929565429688, 'timing/data': 0.0007814040000084788, 'timing/model': 1.1990614270034712, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 1.72}) +[2026-06-24 04:32:41,900][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.10118861496448517, 'train/grad_norm_pre_clip': 39.48542404174805, 'timing/data': 0.0008194639813154936, 'timing/model': 1.1915413849928882, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 1.72}) +[2026-06-24 04:32:46,233][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.06428119540214539, 'train/grad_norm_pre_clip': 12.037636756896973, 'timing/data': 0.0007671370112802833, 'timing/model': 1.1923897319938987, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 1.72}) +[2026-06-24 04:32:50,618][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.06597738713026047, 'train/grad_norm_pre_clip': 22.74833106994629, 'timing/data': 0.000938132987357676, 'timing/model': 1.196193199022673, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 1.72}) +[2026-06-24 04:32:55,194][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.03243567794561386, 'train/grad_norm_pre_clip': 13.552241325378418, 'timing/data': 0.0007944579992908984, 'timing/model': 1.1793623460107483, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 1.72}) +[2026-06-24 04:32:59,553][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.2463269978761673, 'train/grad_norm_pre_clip': 36.712852478027344, 'timing/data': 0.0010531950101722032, 'timing/model': 1.2009815729979891, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 1.72}) +[2026-06-24 04:33:03,927][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.0009536137804389, 'train/grad_norm_pre_clip': 31.194318771362305, 'timing/data': 0.000756608002120629, 'timing/model': 1.2056473919947166, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 1.72}) +[2026-06-24 04:33:08,278][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.04245968908071518, 'train/grad_norm_pre_clip': 14.096434593200684, 'timing/data': 0.0007461479981429875, 'timing/model': 1.1938690939859953, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 1.72}) +[2026-06-24 04:33:12,850][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.0045998552814126015, 'train/grad_norm_pre_clip': 13.584508895874023, 'timing/data': 0.0007464889786206186, 'timing/model': 1.1756973239826038, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 1.72}) +[2026-06-24 04:33:17,183][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.0772472396492958, 'train/grad_norm_pre_clip': 25.029695510864258, 'timing/data': 0.0007358999864663929, 'timing/model': 1.176325425010873, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 1.72}) +[2026-06-24 04:33:21,549][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.02797170728445053, 'train/grad_norm_pre_clip': 8.389561653137207, 'timing/data': 0.0007683100120630115, 'timing/model': 1.1734802389983088, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 1.72}) +[2026-06-24 04:33:25,896][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.005743090994656086, 'train/grad_norm_pre_clip': 18.32369041442871, 'timing/data': 0.0007323729805648327, 'timing/model': 1.1781631390040275, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 1.72}) +[2026-06-24 04:33:30,435][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.02268429473042488, 'train/grad_norm_pre_clip': 8.414949417114258, 'timing/data': 0.0007157619984354824, 'timing/model': 1.193893088027835, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 1.72}) +[2026-06-24 04:33:34,773][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.09223005175590515, 'train/grad_norm_pre_clip': 26.210472106933594, 'timing/data': 0.0007225550070870668, 'timing/model': 1.1730960780114401, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 1.72}) +[2026-06-24 04:33:39,136][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.022306542843580246, 'train/grad_norm_pre_clip': 12.954286575317383, 'timing/data': 0.0006964269850868732, 'timing/model': 1.175528060994111, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 1.73}) +[2026-06-24 04:33:43,441][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.0013156436616554856, 'train/grad_norm_pre_clip': 11.66292667388916, 'timing/data': 0.001054257998475805, 'timing/model': 1.1745694190030918, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 1.73}) +[2026-06-24 04:33:47,977][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.0017961289267987013, 'train/grad_norm_pre_clip': 7.7056803703308105, 'timing/data': 0.0007190280011855066, 'timing/model': 1.1730165609915275, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 1.73}) +[2026-06-24 04:33:52,328][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.0052611310966312885, 'train/grad_norm_pre_clip': 15.460693359375, 'timing/data': 0.0007417299784719944, 'timing/model': 1.2050151219882537, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 1.73}) +[2026-06-24 04:33:56,678][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.027052003890275955, 'train/grad_norm_pre_clip': 22.693191528320312, 'timing/data': 0.0008720210171304643, 'timing/model': 1.1956696080160327, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 1.73}) +[2026-06-24 04:34:01,019][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.026238154619932175, 'train/grad_norm_pre_clip': 15.81302547454834, 'timing/data': 0.0008767599938437343, 'timing/model': 1.1757390399870928, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 1.73}) +[2026-06-24 04:34:05,618][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.0234568789601326, 'train/grad_norm_pre_clip': 30.732736587524414, 'timing/data': 0.0007668370089959353, 'timing/model': 1.1953109549940564, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 1.73}) +[2026-06-24 04:34:09,992][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.024207938462495804, 'train/grad_norm_pre_clip': 10.182002067565918, 'timing/data': 0.0007562369864899665, 'timing/model': 1.1764234129805118, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 1.73}) +[2026-06-24 04:34:14,339][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 5.549597699427977e-05, 'train/grad_norm_pre_clip': 13.348649978637695, 'timing/data': 0.0007239670085255057, 'timing/model': 1.179816601012135, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 1.73}) +[2026-06-24 04:34:18,639][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.09644948691129684, 'train/grad_norm_pre_clip': 22.54691505432129, 'timing/data': 0.0007420209876727313, 'timing/model': 1.1725348710024264, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 1.73}) +[2026-06-24 04:34:23,167][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.03568578138947487, 'train/grad_norm_pre_clip': 6.857285976409912, 'timing/data': 0.0007400979811791331, 'timing/model': 1.1740666400000919, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 1.73}) +[2026-06-24 04:34:27,513][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.21174581348896027, 'train/grad_norm_pre_clip': 43.349674224853516, 'timing/data': 0.0007410389953292906, 'timing/model': 1.1961543829820585, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 1.73}) +[2026-06-24 04:34:31,839][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.03626367822289467, 'train/grad_norm_pre_clip': 27.593326568603516, 'timing/data': 0.0008798560011200607, 'timing/model': 1.1954336810158566, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 1.73}) +[2026-06-24 04:34:36,145][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.012523986399173737, 'train/grad_norm_pre_clip': 13.334844589233398, 'timing/data': 0.0007458369946107268, 'timing/model': 1.1733730069827288, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 1.73}) +[2026-06-24 04:34:40,725][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.026918524876236916, 'train/grad_norm_pre_clip': 10.97962474822998, 'timing/data': 0.0007162239926401526, 'timing/model': 1.1941398599883541, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 1.73}) +[2026-06-24 04:34:45,103][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.009778479114174843, 'train/grad_norm_pre_clip': 10.58901596069336, 'timing/data': 0.0009480110020376742, 'timing/model': 1.2072383380145766, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 1.73}) +[2026-06-24 04:34:49,487][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.12330204248428345, 'train/grad_norm_pre_clip': 18.821334838867188, 'timing/data': 0.0007353989931289107, 'timing/model': 1.1952290829794947, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 1.73}) +[2026-06-24 04:34:53,855][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.045667678117752075, 'train/grad_norm_pre_clip': 29.762981414794922, 'timing/data': 0.0008905050053726882, 'timing/model': 1.1858018880011514, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 1.73}) +[2026-06-24 04:34:58,410][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.0323505774140358, 'train/grad_norm_pre_clip': 15.65649700164795, 'timing/data': 0.0007952690066304058, 'timing/model': 1.1803827959811315, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 1.73}) +[2026-06-24 04:35:02,768][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.09349489212036133, 'train/grad_norm_pre_clip': 31.924179077148438, 'timing/data': 0.0008114289958029985, 'timing/model': 1.2006715259922203, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 1.73}) +[2026-06-24 04:35:07,103][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.01033859234303236, 'train/grad_norm_pre_clip': 21.843509674072266, 'timing/data': 0.0007525700202677399, 'timing/model': 1.1793787410133518, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 1.73}) +[2026-06-24 04:35:11,472][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.001452074502594769, 'train/grad_norm_pre_clip': 12.027112007141113, 'timing/data': 0.001041774987243116, 'timing/model': 1.1892002940003294, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 1.73}) +[2026-06-24 04:35:16,010][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.0006248935824260116, 'train/grad_norm_pre_clip': 19.058073043823242, 'timing/data': 0.0007469500123988837, 'timing/model': 1.1776992889936082, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 1.73}) +[2026-06-24 04:35:20,370][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.10617513954639435, 'train/grad_norm_pre_clip': 36.244544982910156, 'timing/data': 0.000920530001167208, 'timing/model': 1.20738206198439, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 1.73}) +[2026-06-24 04:35:24,694][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.03285658732056618, 'train/grad_norm_pre_clip': 22.942808151245117, 'timing/data': 0.0010289000056218356, 'timing/model': 1.1890741609968245, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 1.73}) +[2026-06-24 04:35:29,034][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.005124710965901613, 'train/grad_norm_pre_clip': 20.7181396484375, 'timing/data': 0.0007768239884171635, 'timing/model': 1.1780268330185208, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 1.74}) +[2026-06-24 04:35:33,579][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.054090116173028946, 'train/grad_norm_pre_clip': 14.534751892089844, 'timing/data': 0.0011367500119376928, 'timing/model': 1.1758770819869824, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 1.74}) +[2026-06-24 04:35:37,926][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.21638205647468567, 'train/grad_norm_pre_clip': 24.000865936279297, 'timing/data': 0.0007442249916493893, 'timing/model': 1.1783407219918445, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 1.74}) +[2026-06-24 04:35:42,285][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.005461234599351883, 'train/grad_norm_pre_clip': 6.024768352508545, 'timing/data': 0.0007422109774779528, 'timing/model': 1.1978190150111914, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 1.74}) +[2026-06-24 04:35:46,622][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.018743785098195076, 'train/grad_norm_pre_clip': 13.170758247375488, 'timing/data': 0.000711022992618382, 'timing/model': 1.1788966399908531, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 1.74}) +[2026-06-24 04:35:51,177][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.012809442356228828, 'train/grad_norm_pre_clip': 14.901330947875977, 'timing/data': 0.0007816540019121021, 'timing/model': 1.1962291069794446, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 1.74}) +[2026-06-24 04:35:55,520][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.0009544843924231827, 'train/grad_norm_pre_clip': 2.042537212371826, 'timing/data': 0.000880847976077348, 'timing/model': 1.1787777290155645, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 1.74}) +[2026-06-24 04:35:59,852][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.06293332576751709, 'train/grad_norm_pre_clip': 16.300840377807617, 'timing/data': 0.0007172140176407993, 'timing/model': 1.176899087993661, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 1.74}) +[2026-06-24 04:36:04,254][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.029155028983950615, 'train/grad_norm_pre_clip': 22.716154098510742, 'timing/data': 0.0008862969989422709, 'timing/model': 1.2010841069859453, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 1.74}) +[2026-06-24 04:36:08,836][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.13315889239311218, 'train/grad_norm_pre_clip': 30.12326431274414, 'timing/data': 0.000782725983299315, 'timing/model': 1.173765599989565, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 1.74}) +[2026-06-24 04:36:13,206][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.011993927881121635, 'train/grad_norm_pre_clip': 23.192270278930664, 'timing/data': 0.0007457079773303121, 'timing/model': 1.1933671410079114, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 1.74}) +[2026-06-24 04:36:17,541][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.003919059876352549, 'train/grad_norm_pre_clip': 21.398866653442383, 'timing/data': 0.0008772910223342478, 'timing/model': 1.1817674619960599, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 1.74}) +[2026-06-24 04:36:21,901][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.015429642051458359, 'train/grad_norm_pre_clip': 12.547135353088379, 'timing/data': 0.0007583610131405294, 'timing/model': 1.1804276970215142, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 1.74}) +[2026-06-24 04:36:26,500][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.02836853265762329, 'train/grad_norm_pre_clip': 16.005727767944336, 'timing/data': 0.0008223589975386858, 'timing/model': 1.1984435780032072, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 1.74}) +[2026-06-24 04:36:30,868][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.020497653633356094, 'train/grad_norm_pre_clip': 24.119159698486328, 'timing/data': 0.0007346570200752467, 'timing/model': 1.1983958589844406, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 1.74}) +[2026-06-24 04:36:35,214][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.0004175806534476578, 'train/grad_norm_pre_clip': 34.6970100402832, 'timing/data': 0.0008756280003581196, 'timing/model': 1.183028450992424, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 1.74}) +[2026-06-24 04:36:39,557][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.02407381311058998, 'train/grad_norm_pre_clip': 23.73998260498047, 'timing/data': 0.0007342259923461825, 'timing/model': 1.1777612510195468, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 1.74}) +[2026-06-24 04:36:44,165][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4500 +[2026-06-24 04:37:42,627][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4500 +[2026-06-24 04:37:42,627][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4500 +[2026-06-24 04:38:19,303][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4500 +[2026-06-24 05:06:37,945][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4500 +[2026-06-24 05:06:37,947][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.1246032640337944, 'train/grad_norm_pre_clip': 27.248302459716797, 'eval/action_loss/seconds': 58.45901919098105, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03292079642415047, 'eval/latency_2/loss': 0.03292079642415047, 'eval/flappy/loss': 0.03292079642415047, 'eval/flappy/latency_2/loss': 0.03292079642415047, 'eval/flap_precision': 0.9823369565217391, 'eval/flap_recall': 0.9810040705563093, 'eval/flap_f1': 0.9816700610997963, 'eval/flap_rate_model': 0.20444444444444446, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9816700610997962, 'eval/latency_2/flap_precision': 0.9823369565217391, 'eval/latency_2/flap_recall': 0.9810040705563093, 'eval/latency_2/flap_f1': 0.9816700610997963, 'eval/latency_2/flap_rate_model': 0.20444444444444446, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9816700610997962, 'eval/action_classification/data_wait_seconds': 0.43500620126724243, 'eval/action_classification/predict_seconds': 36.0964469909668, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0011801109940279275, 'timing/model': 1.2083402600255795, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 1.74}) +[2026-06-24 05:06:37,947][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_4500_state +[2026-06-24 05:06:37,957][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 05:06:46,652][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_4500_state/model.safetensors +[2026-06-24 05:06:56,907][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_4500_state/optimizer.bin +[2026-06-24 05:06:56,908][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_4500_state/sampler.bin +[2026-06-24 05:06:56,908][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-24 05:06:56,910][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-24 05:06:57,113][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-24 05:06:57,114][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 05:06:57,138][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 05:07:01,361][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.013245494104921818, 'train/grad_norm_pre_clip': 8.67041015625, 'timing/data': 0.000999365991447121, 'timing/model': 1.1507913910027128, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 1.74}) +[2026-06-24 05:07:05,765][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.02693638950586319, 'train/grad_norm_pre_clip': 17.885967254638672, 'timing/data': 0.0009155909938272089, 'timing/model': 1.3585466890071984, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 1.74}) +[2026-06-24 05:07:09,984][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.019694078713655472, 'train/grad_norm_pre_clip': 20.427814483642578, 'timing/data': 0.0007419400208164006, 'timing/model': 1.1553945539926644, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 1.74}) +[2026-06-24 05:07:14,220][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.006468924228101969, 'train/grad_norm_pre_clip': 3.540649175643921, 'timing/data': 0.0007818650046829134, 'timing/model': 1.1737199100025464, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 1.74}) +[2026-06-24 05:07:18,431][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.003679074114188552, 'train/grad_norm_pre_clip': 19.458887100219727, 'timing/data': 0.0007587319996673614, 'timing/model': 1.1570906970009673, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 1.74}) +[2026-06-24 05:07:22,879][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.008071989752352238, 'train/grad_norm_pre_clip': 24.293315887451172, 'timing/data': 0.0007613360066898167, 'timing/model': 1.3667366369918454, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 1.74}) +[2026-06-24 05:07:27,081][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.04017334431409836, 'train/grad_norm_pre_clip': 20.049747467041016, 'timing/data': 0.0007873139984440058, 'timing/model': 1.1517181289964356, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 1.74}) +[2026-06-24 05:07:31,286][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.004070853348821402, 'train/grad_norm_pre_clip': 7.447310924530029, 'timing/data': 0.0007594740018248558, 'timing/model': 1.1438299849978648, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 1.74}) +[2026-06-24 05:07:35,506][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.015159227885305882, 'train/grad_norm_pre_clip': 6.523331165313721, 'timing/data': 0.0009709840232972056, 'timing/model': 1.1494625109771732, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 1.75}) +[2026-06-24 05:07:40,006][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.04901975020766258, 'train/grad_norm_pre_clip': 27.630693435668945, 'timing/data': 0.0009345260041300207, 'timing/model': 1.4156537420058157, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 1.75}) +[2026-06-24 05:07:44,280][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.24640046060085297, 'train/grad_norm_pre_clip': 25.173871994018555, 'timing/data': 0.000724287994671613, 'timing/model': 1.166704725008458, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 1.75}) +[2026-06-24 05:07:48,562][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.17997609078884125, 'train/grad_norm_pre_clip': 19.44219398498535, 'timing/data': 0.0008323479851242155, 'timing/model': 1.1799553360033315, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 1.75}) +[2026-06-24 05:07:52,784][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.014346251264214516, 'train/grad_norm_pre_clip': 48.480194091796875, 'timing/data': 0.0009599240147508681, 'timing/model': 1.1535749370232224, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 1.75}) +[2026-06-24 05:07:57,271][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.00029109243769198656, 'train/grad_norm_pre_clip': 26.863155364990234, 'timing/data': 0.0007447960088029504, 'timing/model': 1.1705179980199318, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 1.75}) +[2026-06-24 05:08:01,506][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.006628144532442093, 'train/grad_norm_pre_clip': 10.66025447845459, 'timing/data': 0.0008646070200484246, 'timing/model': 1.1491611399978865, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 1.75}) +[2026-06-24 05:08:05,791][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.014988546259701252, 'train/grad_norm_pre_clip': 17.06587791442871, 'timing/data': 0.0008818389906082302, 'timing/model': 1.1878438180137891, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 1.75}) +[2026-06-24 05:08:10,065][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.01699383556842804, 'train/grad_norm_pre_clip': 31.12008285522461, 'timing/data': 0.0010645570000633597, 'timing/model': 1.1584717109799385, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 1.75}) +[2026-06-24 05:08:14,535][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.001023723278194666, 'train/grad_norm_pre_clip': 13.526409149169922, 'timing/data': 0.0013059120101388544, 'timing/model': 1.1658079749904573, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 1.75}) +[2026-06-24 05:08:18,839][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.10561592876911163, 'train/grad_norm_pre_clip': 23.624874114990234, 'timing/data': 0.0011823149980045855, 'timing/model': 1.1802398379950318, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 1.75}) +[2026-06-24 05:08:23,157][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.009459198452532291, 'train/grad_norm_pre_clip': 11.747217178344727, 'timing/data': 0.0008979700214695185, 'timing/model': 1.1795510849915445, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 1.75}) +[2026-06-24 05:08:27,423][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 8.66391128511168e-05, 'train/grad_norm_pre_clip': 11.73436164855957, 'timing/data': 0.0007531919982284307, 'timing/model': 1.1640922829974443, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 1.75}) +[2026-06-24 05:08:31,932][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.021542591974139214, 'train/grad_norm_pre_clip': 10.786463737487793, 'timing/data': 0.0007795710116624832, 'timing/model': 1.1650973299983889, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 1.75}) +[2026-06-24 05:08:36,251][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.033318690955638885, 'train/grad_norm_pre_clip': 11.59689998626709, 'timing/data': 0.000739686016459018, 'timing/model': 1.1869269399903715, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 1.75}) +[2026-06-24 05:08:40,581][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.09317520260810852, 'train/grad_norm_pre_clip': 30.28508758544922, 'timing/data': 0.0008965370070654899, 'timing/model': 1.1845542970113456, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 1.75}) +[2026-06-24 05:08:44,921][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.016488460823893547, 'train/grad_norm_pre_clip': 11.354612350463867, 'timing/data': 0.0007384539931081235, 'timing/model': 1.1880113519728184, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 1.75}) +[2026-06-24 05:08:49,408][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.057022880762815475, 'train/grad_norm_pre_clip': 26.71934700012207, 'timing/data': 0.0009042399760801345, 'timing/model': 1.181605567020597, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 1.75}) +[2026-06-24 05:08:53,736][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.0353274941444397, 'train/grad_norm_pre_clip': 11.884343147277832, 'timing/data': 0.0010910860146395862, 'timing/model': 1.1834278340102173, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 1.75}) +[2026-06-24 05:08:58,055][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.08238093554973602, 'train/grad_norm_pre_clip': 35.23503494262695, 'timing/data': 0.0009307499858550727, 'timing/model': 1.1672378109942656, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 1.75}) +[2026-06-24 05:09:02,372][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.006781184114515781, 'train/grad_norm_pre_clip': 17.735633850097656, 'timing/data': 0.0007610059983562678, 'timing/model': 1.1910032379964832, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 1.75}) +[2026-06-24 05:09:06,866][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.04810759425163269, 'train/grad_norm_pre_clip': 54.31410598754883, 'timing/data': 0.0009040000149980187, 'timing/model': 1.18881643700297, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 1.75}) +[2026-06-24 05:09:11,202][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.017077479511499405, 'train/grad_norm_pre_clip': 23.411405563354492, 'timing/data': 0.0012050060031469911, 'timing/model': 1.1962976090144366, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 1.75}) +[2026-06-24 05:09:15,549][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.02831077389419079, 'train/grad_norm_pre_clip': 22.047771453857422, 'timing/data': 0.0007741499866824597, 'timing/model': 1.188107216003118, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 1.75}) +[2026-06-24 05:09:20,122][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.04271373897790909, 'train/grad_norm_pre_clip': 14.902704238891602, 'timing/data': 0.000928416004171595, 'timing/model': 1.4445383029815275, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 1.75}) +[2026-06-24 05:09:24,443][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.040276072919368744, 'train/grad_norm_pre_clip': 69.14224243164062, 'timing/data': 0.0008259960159193724, 'timing/model': 1.16921731599723, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 1.75}) +[2026-06-24 05:09:28,793][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.030156578868627548, 'train/grad_norm_pre_clip': 12.173995018005371, 'timing/data': 0.0007999980007298291, 'timing/model': 1.1912931620026939, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 1.76}) +[2026-06-24 05:09:33,121][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.10883372277021408, 'train/grad_norm_pre_clip': 32.53308868408203, 'timing/data': 0.0007668070029467344, 'timing/model': 1.1938170630019158, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 1.76}) +[2026-06-24 05:09:37,703][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.02876669354736805, 'train/grad_norm_pre_clip': 26.62898826599121, 'timing/data': 0.000756237015593797, 'timing/model': 1.4345894900034182, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 1.76}) +[2026-06-24 05:09:42,069][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.0065772645175457, 'train/grad_norm_pre_clip': 20.328569412231445, 'timing/data': 0.0007835169963072985, 'timing/model': 1.189211384014925, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 1.76}) +[2026-06-24 05:09:46,387][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.044053640216588974, 'train/grad_norm_pre_clip': 15.017399787902832, 'timing/data': 0.0009404080046806484, 'timing/model': 1.1930454669927713, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 1.76}) +[2026-06-24 05:09:50,743][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.01725207082927227, 'train/grad_norm_pre_clip': 7.467731475830078, 'timing/data': 0.0007818540034350008, 'timing/model': 1.193186706979759, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 1.76}) +[2026-06-24 05:09:55,274][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.02626689523458481, 'train/grad_norm_pre_clip': 11.191631317138672, 'timing/data': 0.0008015209750737995, 'timing/model': 1.2010383730230387, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 1.76}) +[2026-06-24 05:09:59,622][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.03869038447737694, 'train/grad_norm_pre_clip': 27.26906967163086, 'timing/data': 0.0007188479939941317, 'timing/model': 1.1728649720025714, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 1.76}) +[2026-06-24 05:10:04,118][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.014691399410367012, 'train/grad_norm_pre_clip': 4.573509216308594, 'timing/data': 0.000708188017597422, 'timing/model': 1.3519361020007636, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 1.76}) +[2026-06-24 05:10:08,429][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.009966852143406868, 'train/grad_norm_pre_clip': 16.49612808227539, 'timing/data': 0.0007006440137047321, 'timing/model': 1.190613400976872, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 1.76}) +[2026-06-24 05:10:13,021][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.019187036901712418, 'train/grad_norm_pre_clip': 14.512784004211426, 'timing/data': 0.0007403579948004335, 'timing/model': 1.1880958400142845, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 1.76}) +[2026-06-24 05:10:17,311][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.027779055759310722, 'train/grad_norm_pre_clip': 23.986413955688477, 'timing/data': 0.0007107729907147586, 'timing/model': 1.1711844149976969, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 1.76}) +[2026-06-24 05:10:21,645][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.04396665468811989, 'train/grad_norm_pre_clip': 15.268111228942871, 'timing/data': 0.000972286012256518, 'timing/model': 1.211101602006238, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 1.76}) +[2026-06-24 05:10:26,048][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.0012164070503786206, 'train/grad_norm_pre_clip': 14.448956489562988, 'timing/data': 0.0009020870202220976, 'timing/model': 1.2087580219958909, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 1.76}) +[2026-06-24 05:10:30,635][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.010495798662304878, 'train/grad_norm_pre_clip': 18.292163848876953, 'timing/data': 0.001053585991030559, 'timing/model': 1.1781484200037085, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 1.76}) +[2026-06-24 05:10:34,998][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.008856397122144699, 'train/grad_norm_pre_clip': 17.0120906829834, 'timing/data': 0.0007838289893697947, 'timing/model': 1.1939707819838077, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 1.76}) +[2026-06-24 05:10:39,366][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.006569295190274715, 'train/grad_norm_pre_clip': 36.45941925048828, 'timing/data': 0.0007431530102621764, 'timing/model': 1.1950850090070162, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 1.76}) +[2026-06-24 05:10:43,679][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.0724867656826973, 'train/grad_norm_pre_clip': 28.755027770996094, 'timing/data': 0.0007959000067785382, 'timing/model': 1.1765301680134144, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 1.76}) +[2026-06-24 05:10:48,225][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.004735878203064203, 'train/grad_norm_pre_clip': 13.538732528686523, 'timing/data': 0.0007709549972787499, 'timing/model': 1.1784344669722486, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 1.76}) +[2026-06-24 05:10:52,593][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.07208415120840073, 'train/grad_norm_pre_clip': 22.5367488861084, 'timing/data': 0.0008275979780592024, 'timing/model': 1.1929259799944703, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 1.76}) +[2026-06-24 05:10:56,937][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.09156791120767593, 'train/grad_norm_pre_clip': 22.532909393310547, 'timing/data': 0.000770323007600382, 'timing/model': 1.1981710899854079, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 1.76}) +[2026-06-24 05:11:01,457][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.004679033998399973, 'train/grad_norm_pre_clip': 7.575997352600098, 'timing/data': 0.0007533509924542159, 'timing/model': 1.364650735980831, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 1.76}) +[2026-06-24 05:11:05,824][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.0019364039180800319, 'train/grad_norm_pre_clip': 16.8582706451416, 'timing/data': 0.0007278549892362207, 'timing/model': 1.20086221100064, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 1.76}) +[2026-06-24 05:11:10,186][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.1703210026025772, 'train/grad_norm_pre_clip': 48.16782760620117, 'timing/data': 0.0008164480095729232, 'timing/model': 1.2021327490219846, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 1.76}) +[2026-06-24 05:11:14,512][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.003815336851403117, 'train/grad_norm_pre_clip': 40.06424331665039, 'timing/data': 0.0007622980047017336, 'timing/model': 1.1887036569823977, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 1.76}) +[2026-06-24 05:11:19,057][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.10721512883901596, 'train/grad_norm_pre_clip': 42.28105163574219, 'timing/data': 0.0008984089945442975, 'timing/model': 1.2062260099919513, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 1.76}) +[2026-06-24 05:11:23,412][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.007075097877532244, 'train/grad_norm_pre_clip': 17.859004974365234, 'timing/data': 0.0007626689912285656, 'timing/model': 1.2052330750157125, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 1.77}) +[2026-06-24 05:11:27,762][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.024804379791021347, 'train/grad_norm_pre_clip': 12.638501167297363, 'timing/data': 0.001154061988927424, 'timing/model': 1.1715326129924506, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 1.77}) +[2026-06-24 05:11:32,099][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.03519720211625099, 'train/grad_norm_pre_clip': 41.80350112915039, 'timing/data': 0.0007448259857483208, 'timing/model': 1.1923689670220483, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 1.77}) +[2026-06-24 05:11:36,688][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.006037975661456585, 'train/grad_norm_pre_clip': 15.029914855957031, 'timing/data': 0.0008573339728172868, 'timing/model': 1.4156703829939943, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 1.77}) +[2026-06-24 05:11:40,976][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.0037559629417955875, 'train/grad_norm_pre_clip': 28.494964599609375, 'timing/data': 0.0007122759998310357, 'timing/model': 1.1744324690080248, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 1.77}) +[2026-06-24 05:11:45,286][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.08446043729782104, 'train/grad_norm_pre_clip': 27.679479598999023, 'timing/data': 0.0008069909818004817, 'timing/model': 1.1760401189967524, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 1.77}) +[2026-06-24 05:11:49,593][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.006165676284581423, 'train/grad_norm_pre_clip': 5.513627529144287, 'timing/data': 0.0009235460020136088, 'timing/model': 1.1809545180003624, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 1.77}) +[2026-06-24 05:11:54,077][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.002944754436612129, 'train/grad_norm_pre_clip': 19.892608642578125, 'timing/data': 0.0009333950001746416, 'timing/model': 1.354986476013437, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 1.77}) +[2026-06-24 05:11:58,437][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.0004495481261983514, 'train/grad_norm_pre_clip': 21.35234260559082, 'timing/data': 0.000731540989363566, 'timing/model': 1.190637546009384, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 1.77}) +[2026-06-24 05:12:02,760][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.0036014120560139418, 'train/grad_norm_pre_clip': 25.003524780273438, 'timing/data': 0.0008594670216552913, 'timing/model': 1.171180388017092, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 1.77}) +[2026-06-24 05:12:07,055][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.09260175377130508, 'train/grad_norm_pre_clip': 32.26644515991211, 'timing/data': 0.000755595974624157, 'timing/model': 1.1748041230021045, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 1.77}) +[2026-06-24 05:12:11,410][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.010155534371733665, 'train/grad_norm_pre_clip': 31.30152130126953, 'timing/data': 0.0009316409996245056, 'timing/model': 1.1967053150001448, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 1.77}) +[2026-06-24 05:12:15,946][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.06993751227855682, 'train/grad_norm_pre_clip': 34.634315490722656, 'timing/data': 0.0012362839770503342, 'timing/model': 1.1835806969902478, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 1.77}) +[2026-06-24 05:12:20,315][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.0021046127658337355, 'train/grad_norm_pre_clip': 15.74329662322998, 'timing/data': 0.0007995569903869182, 'timing/model': 1.1958895979914814, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 1.77}) +[2026-06-24 05:12:24,646][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.004350464791059494, 'train/grad_norm_pre_clip': 22.288827896118164, 'timing/data': 0.000773840001784265, 'timing/model': 1.1781991540046874, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 1.77}) +[2026-06-24 05:12:28,995][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 2.048766145890113e-05, 'train/grad_norm_pre_clip': 11.262163162231445, 'timing/data': 0.0010182709957007319, 'timing/model': 1.1935247479996178, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 1.77}) +[2026-06-24 05:12:33,602][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.012962274253368378, 'train/grad_norm_pre_clip': 10.111654281616211, 'timing/data': 0.0011439640074968338, 'timing/model': 1.1964808990014717, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 1.77}) +[2026-06-24 05:12:37,914][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.016527405008673668, 'train/grad_norm_pre_clip': 12.315345764160156, 'timing/data': 0.0009628190018702298, 'timing/model': 1.172372828994412, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 1.77}) +[2026-06-24 05:12:42,218][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 4.952240487909876e-05, 'train/grad_norm_pre_clip': 25.026172637939453, 'timing/data': 0.000768078985856846, 'timing/model': 1.1778228179900907, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 1.77}) +[2026-06-24 05:12:46,550][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.0383869931101799, 'train/grad_norm_pre_clip': 18.074817657470703, 'timing/data': 0.0007817650039214641, 'timing/model': 1.173278892005328, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 1.77}) +[2026-06-24 05:12:51,010][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.012612910941243172, 'train/grad_norm_pre_clip': 8.482450485229492, 'timing/data': 0.0007418410095851868, 'timing/model': 1.1717077899957076, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 1.77}) +[2026-06-24 05:12:55,374][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.03361262008547783, 'train/grad_norm_pre_clip': 22.51852798461914, 'timing/data': 0.000780672999098897, 'timing/model': 1.1914577069983352, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 1.77}) +[2026-06-24 05:12:59,678][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.0042500970885157585, 'train/grad_norm_pre_clip': 6.543262004852295, 'timing/data': 0.0009146599913947284, 'timing/model': 1.1782717760070227, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 1.77}) +[2026-06-24 05:13:04,057][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.004544923081994057, 'train/grad_norm_pre_clip': 7.972733020782471, 'timing/data': 0.0011147100012749434, 'timing/model': 1.2044629040174186, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 1.77}) +[2026-06-24 05:13:08,574][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.08155964314937592, 'train/grad_norm_pre_clip': 30.38934326171875, 'timing/data': 0.0007555850024800748, 'timing/model': 1.1796461140038446, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 1.77}) +[2026-06-24 05:13:13,317][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.0065023149363696575, 'train/grad_norm_pre_clip': 26.924028396606445, 'timing/data': 0.0013201989931985736, 'timing/model': 1.238854313007323, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 1.77}) +[2026-06-24 05:13:17,678][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.006989043205976486, 'train/grad_norm_pre_clip': 22.393299102783203, 'timing/data': 0.001358239009277895, 'timing/model': 1.2075183979759458, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 1.78}) +[2026-06-24 05:13:22,024][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.0004317440907470882, 'train/grad_norm_pre_clip': 23.192903518676758, 'timing/data': 0.000997803988866508, 'timing/model': 1.1989667890011333, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 1.78}) +[2026-06-24 05:13:26,598][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.07167549431324005, 'train/grad_norm_pre_clip': 58.12541198730469, 'timing/data': 0.0009353779896628112, 'timing/model': 1.1736503050196916, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 1.78}) +[2026-06-24 05:13:30,906][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.010094642639160156, 'train/grad_norm_pre_clip': 13.261951446533203, 'timing/data': 0.0007265020103659481, 'timing/model': 1.1684225159988273, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 1.78}) +[2026-06-24 05:13:35,226][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.0004849569813814014, 'train/grad_norm_pre_clip': 2.103485584259033, 'timing/data': 0.0007274340023286641, 'timing/model': 1.1812763740017544, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 1.78}) +[2026-06-24 05:13:39,545][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.0782959908246994, 'train/grad_norm_pre_clip': 31.136768341064453, 'timing/data': 0.0007606149883940816, 'timing/model': 1.1738607040024363, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 1.78}) +[2026-06-24 05:13:44,185][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.005157524719834328, 'train/grad_norm_pre_clip': 24.320234298706055, 'timing/data': 0.000764642987633124, 'timing/model': 1.1945672889996786, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 1.78}) +[2026-06-24 05:13:48,507][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.03541816025972366, 'train/grad_norm_pre_clip': 30.88089370727539, 'timing/data': 0.0007338160066865385, 'timing/model': 1.1930075069831219, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 1.78}) +[2026-06-24 05:13:52,904][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.029112819582223892, 'train/grad_norm_pre_clip': 21.315673828125, 'timing/data': 0.0009199199848808348, 'timing/model': 1.2005093870102428, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 1.78}) +[2026-06-24 05:13:57,278][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.02687893807888031, 'train/grad_norm_pre_clip': 29.301006317138672, 'timing/data': 0.0007468200055882335, 'timing/model': 1.1928286059992388, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 1.78}) +[2026-06-24 05:14:01,862][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.005080741830170155, 'train/grad_norm_pre_clip': 19.040355682373047, 'timing/data': 0.000701316021149978, 'timing/model': 1.1729382979974616, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 1.78}) +[2026-06-24 05:14:06,176][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.011897382326424122, 'train/grad_norm_pre_clip': 3.6210031509399414, 'timing/data': 0.0007186080038081855, 'timing/model': 1.1708574229851365, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 1.78}) +[2026-06-24 05:14:10,513][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.004178092814981937, 'train/grad_norm_pre_clip': 27.630840301513672, 'timing/data': 0.0009212119912263006, 'timing/model': 1.1758037919935305, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 1.78}) +[2026-06-24 05:14:15,002][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4600 +[2026-06-24 05:15:12,980][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4600 +[2026-06-24 05:15:12,981][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4600 +[2026-06-24 05:15:49,915][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4600 +[2026-06-24 05:15:49,917][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.001722001819871366, 'train/grad_norm_pre_clip': 9.70289421081543, 'eval/action_loss/seconds': 57.97486956199282, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03398251160979271, 'eval/latency_2/loss': 0.03398251160979271, 'eval/flappy/loss': 0.03398251160979271, 'eval/flappy/latency_2/loss': 0.03398251160979271, 'eval/flap_precision': 0.9757085020242915, 'eval/flap_recall': 0.9810040705563093, 'eval/flap_f1': 0.9783491204330177, 'eval/flap_rate_model': 0.20583333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9783491204330177, 'eval/latency_2/flap_precision': 0.9757085020242915, 'eval/latency_2/flap_recall': 0.9810040705563093, 'eval/latency_2/flap_f1': 0.9783491204330177, 'eval/latency_2/flap_rate_model': 0.20583333333333334, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9783491204330177, 'eval/action_classification/data_wait_seconds': 0.509686291217804, 'eval/action_classification/predict_seconds': 36.27323532104492, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007395870052278042, 'timing/model': 1.3621159670001362, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 1.78}) +[2026-06-24 05:15:54,278][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.019199104979634285, 'train/grad_norm_pre_clip': 23.016170501708984, 'timing/data': 0.0008144750026986003, 'timing/model': 1.1611053300148342, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 1.78}) +[2026-06-24 05:15:58,590][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.0033040812704712152, 'train/grad_norm_pre_clip': 20.72344207763672, 'timing/data': 0.000737702997867018, 'timing/model': 1.187887831008993, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 1.78}) +[2026-06-24 05:16:02,893][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.007636074908077717, 'train/grad_norm_pre_clip': 43.508705139160156, 'timing/data': 0.0008314849983435124, 'timing/model': 1.1663323969987687, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 1.78}) +[2026-06-24 05:16:07,433][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.010489095002412796, 'train/grad_norm_pre_clip': 15.919086456298828, 'timing/data': 0.001163368986453861, 'timing/model': 1.194996013015043, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 1.78}) +[2026-06-24 05:16:11,765][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.0037513042334467173, 'train/grad_norm_pre_clip': 17.284114837646484, 'timing/data': 0.0007503160159103572, 'timing/model': 1.186371618998237, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 1.78}) +[2026-06-24 05:16:16,093][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.010365519672632217, 'train/grad_norm_pre_clip': 10.120052337646484, 'timing/data': 0.0009870539943221956, 'timing/model': 1.1845532489824109, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 1.78}) +[2026-06-24 05:16:20,393][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.001760957413353026, 'train/grad_norm_pre_clip': 23.677045822143555, 'timing/data': 0.0007608960149809718, 'timing/model': 1.1629950299975462, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 1.78}) +[2026-06-24 05:16:24,846][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.015797007828950882, 'train/grad_norm_pre_clip': 25.692569732666016, 'timing/data': 0.0007233070209622383, 'timing/model': 1.1631815950095188, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 1.78}) +[2026-06-24 05:16:29,172][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.0395091213285923, 'train/grad_norm_pre_clip': 25.079872131347656, 'timing/data': 0.0011906290019396693, 'timing/model': 1.1670429579971824, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 1.78}) +[2026-06-24 05:16:33,501][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.061931025236845016, 'train/grad_norm_pre_clip': 38.76272201538086, 'timing/data': 0.0007476010068785399, 'timing/model': 1.189023848011857, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 1.78}) +[2026-06-24 05:16:37,774][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.00318111851811409, 'train/grad_norm_pre_clip': 41.69851303100586, 'timing/data': 0.0007301279983948916, 'timing/model': 1.167749623011332, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 1.78}) +[2026-06-24 05:16:42,250][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.023575015366077423, 'train/grad_norm_pre_clip': 15.384905815124512, 'timing/data': 0.0009303989936597645, 'timing/model': 1.166684065014124, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 1.78}) +[2026-06-24 05:16:46,551][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.002946146996691823, 'train/grad_norm_pre_clip': 17.167348861694336, 'timing/data': 0.000707888015313074, 'timing/model': 1.1663430350017734, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 1.79}) +[2026-06-24 05:16:50,850][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.006222818046808243, 'train/grad_norm_pre_clip': 15.334529876708984, 'timing/data': 0.0007055339810904115, 'timing/model': 1.1934522889787331, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 1.79}) +[2026-06-24 05:16:55,131][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.0303681418299675, 'train/grad_norm_pre_clip': 19.666759490966797, 'timing/data': 0.0006903850007802248, 'timing/model': 1.168686784018064, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 1.79}) +[2026-06-24 05:16:59,614][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.031301431357860565, 'train/grad_norm_pre_clip': 22.663471221923828, 'timing/data': 0.0007450770062860101, 'timing/model': 1.194617000001017, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 1.79}) +[2026-06-24 05:17:03,916][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.008187390863895416, 'train/grad_norm_pre_clip': 25.256654739379883, 'timing/data': 0.0006889530050102621, 'timing/model': 1.1677812010166235, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 1.79}) +[2026-06-24 05:17:08,202][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.02650841698050499, 'train/grad_norm_pre_clip': 30.908720016479492, 'timing/data': 0.0006971079856157303, 'timing/model': 1.1687253850104753, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 1.79}) +[2026-06-24 05:17:12,512][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.001095118117518723, 'train/grad_norm_pre_clip': 56.00522232055664, 'timing/data': 0.0007061150099616498, 'timing/model': 1.1893111160025, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 1.79}) +[2026-06-24 05:17:17,046][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.08122752606868744, 'train/grad_norm_pre_clip': 32.215721130371094, 'timing/data': 0.000708970008417964, 'timing/model': 1.201422782993177, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 1.79}) +[2026-06-24 05:17:21,355][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.000258035579463467, 'train/grad_norm_pre_clip': 14.203972816467285, 'timing/data': 0.0006950639944989234, 'timing/model': 1.1741927560069598, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 1.79}) +[2026-06-24 05:17:25,643][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.1012231782078743, 'train/grad_norm_pre_clip': 27.242298126220703, 'timing/data': 0.000700303993653506, 'timing/model': 1.1714980769902468, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 1.79}) +[2026-06-24 05:17:29,993][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.0054398272186517715, 'train/grad_norm_pre_clip': 23.327211380004883, 'timing/data': 0.00071098399348557, 'timing/model': 1.1934382610197645, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 1.79}) +[2026-06-24 05:17:34,547][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.05493352189660072, 'train/grad_norm_pre_clip': 41.066829681396484, 'timing/data': 0.0007605049759149551, 'timing/model': 1.1971017699979711, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 1.79}) +[2026-06-24 05:17:38,895][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 9.782954293768853e-05, 'train/grad_norm_pre_clip': 20.440662384033203, 'timing/data': 0.0007711449870839715, 'timing/model': 1.195709296996938, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 1.79}) +[2026-06-24 05:17:43,260][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.013488000258803368, 'train/grad_norm_pre_clip': 18.480653762817383, 'timing/data': 0.0008718600147403777, 'timing/model': 1.2027874590130523, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 1.79}) +[2026-06-24 05:17:47,614][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.04156588390469551, 'train/grad_norm_pre_clip': 13.599300384521484, 'timing/data': 0.0008791930158622563, 'timing/model': 1.1897069750120863, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 1.79}) +[2026-06-24 05:17:52,174][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.005375117529183626, 'train/grad_norm_pre_clip': 34.63539123535156, 'timing/data': 0.0007012849964667112, 'timing/model': 1.1922523950051982, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 1.79}) +[2026-06-24 05:17:56,494][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.015057369135320187, 'train/grad_norm_pre_clip': 14.674198150634766, 'timing/data': 0.0007136280182749033, 'timing/model': 1.2033487930020783, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 1.79}) +[2026-06-24 05:18:00,835][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.022261112928390503, 'train/grad_norm_pre_clip': 17.653484344482422, 'timing/data': 0.0007287460030056536, 'timing/model': 1.1723543800180778, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 1.79}) +[2026-06-24 05:18:05,202][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.1099795252084732, 'train/grad_norm_pre_clip': 36.89604949951172, 'timing/data': 0.001190647977637127, 'timing/model': 1.1974067490082234, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 1.79}) +[2026-06-24 05:18:09,754][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.03552259877324104, 'train/grad_norm_pre_clip': 44.05513000488281, 'timing/data': 0.000782504997914657, 'timing/model': 1.176695292990189, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 1.79}) +[2026-06-24 05:18:14,093][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.16398267447948456, 'train/grad_norm_pre_clip': 26.918575286865234, 'timing/data': 0.0009790680196601897, 'timing/model': 1.197723882010905, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 1.79}) +[2026-06-24 05:18:18,436][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.08749493211507797, 'train/grad_norm_pre_clip': 32.943416595458984, 'timing/data': 0.0008706270018592477, 'timing/model': 1.1944033089966979, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 1.79}) +[2026-06-24 05:18:22,797][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.008929859846830368, 'train/grad_norm_pre_clip': 101.5933609008789, 'timing/data': 0.0009749699966050684, 'timing/model': 1.1965135940117761, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 1.79}) +[2026-06-24 05:18:27,334][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.010109897702932358, 'train/grad_norm_pre_clip': 14.15727710723877, 'timing/data': 0.0009977019799407572, 'timing/model': 1.1762475699943025, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 1.79}) +[2026-06-24 05:18:31,688][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.01700938493013382, 'train/grad_norm_pre_clip': 26.561630249023438, 'timing/data': 0.0009987140074372292, 'timing/model': 1.1777436020202003, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 1.79}) +[2026-06-24 05:18:36,056][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.0007614136557094753, 'train/grad_norm_pre_clip': 18.49137306213379, 'timing/data': 0.0008838720095809549, 'timing/model': 1.1958713539934251, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 1.79}) +[2026-06-24 05:18:40,413][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.01136077381670475, 'train/grad_norm_pre_clip': 21.831357955932617, 'timing/data': 0.0009479500004090369, 'timing/model': 1.2014939440123271, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 1.8}) +[2026-06-24 05:18:44,936][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.03765634074807167, 'train/grad_norm_pre_clip': 21.918909072875977, 'timing/data': 0.0009072549873962998, 'timing/model': 1.1767051279894076, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 1.8}) +[2026-06-24 05:18:49,256][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.0015706949634477496, 'train/grad_norm_pre_clip': 12.19992446899414, 'timing/data': 0.0007601539837196469, 'timing/model': 1.1759232920012437, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 1.8}) +[2026-06-24 05:18:53,601][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.037026554346084595, 'train/grad_norm_pre_clip': 6.8491291999816895, 'timing/data': 0.0008216470014303923, 'timing/model': 1.189813099015737, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 1.8}) +[2026-06-24 05:18:57,946][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.07134919613599777, 'train/grad_norm_pre_clip': 39.36620330810547, 'timing/data': 0.0010014490108005702, 'timing/model': 1.20400515399524, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 1.8}) +[2026-06-24 05:19:02,493][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.03383117541670799, 'train/grad_norm_pre_clip': 36.46687316894531, 'timing/data': 0.0011905089777428657, 'timing/model': 1.1943151179875713, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 1.8}) +[2026-06-24 05:19:06,883][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.03479929640889168, 'train/grad_norm_pre_clip': 9.983577728271484, 'timing/data': 0.0007925539975985885, 'timing/model': 1.1979147840174846, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 1.8}) +[2026-06-24 05:19:11,205][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.007191972341388464, 'train/grad_norm_pre_clip': 26.87934112548828, 'timing/data': 0.0008130820060614496, 'timing/model': 1.1757010710134637, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 1.8}) +[2026-06-24 05:19:15,549][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.033765994012355804, 'train/grad_norm_pre_clip': 51.82628631591797, 'timing/data': 0.0010095949983224273, 'timing/model': 1.1953160899865907, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 1.8}) +[2026-06-24 05:19:20,109][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.028409242630004883, 'train/grad_norm_pre_clip': 29.75827407836914, 'timing/data': 0.0008234609849750996, 'timing/model': 1.1786749179882463, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 1.8}) +[2026-06-24 05:19:24,513][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.0035036206245422363, 'train/grad_norm_pre_clip': 6.640524864196777, 'timing/data': 0.0008138720004353672, 'timing/model': 1.2221465519978665, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 1.8}) +[2026-06-24 05:19:28,826][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.021653801202774048, 'train/grad_norm_pre_clip': 11.417250633239746, 'timing/data': 0.0007142790127545595, 'timing/model': 1.1737445979961194, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 1.8}) +[2026-06-24 05:19:33,162][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.029927032068371773, 'train/grad_norm_pre_clip': 42.82243347167969, 'timing/data': 0.0007261509890668094, 'timing/model': 1.1694366709853057, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 1.8}) +[2026-06-24 05:19:37,705][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.005822960287332535, 'train/grad_norm_pre_clip': 18.97837257385254, 'timing/data': 0.0008040849934332073, 'timing/model': 1.1726673129887786, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 1.8}) +[2026-06-24 05:19:42,018][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.033653970807790756, 'train/grad_norm_pre_clip': 21.383670806884766, 'timing/data': 0.001153290009824559, 'timing/model': 1.1736448359733913, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 1.8}) +[2026-06-24 05:19:46,333][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.001177388010546565, 'train/grad_norm_pre_clip': 8.445116996765137, 'timing/data': 0.0008186320192180574, 'timing/model': 1.1729984380071983, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 1.8}) +[2026-06-24 05:19:50,690][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.04716091603040695, 'train/grad_norm_pre_clip': 13.317928314208984, 'timing/data': 0.0007390249811578542, 'timing/model': 1.2059194100147579, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 1.8}) +[2026-06-24 05:19:55,245][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.01995517499744892, 'train/grad_norm_pre_clip': 9.31605052947998, 'timing/data': 0.0009548829984851182, 'timing/model': 1.1966659799800254, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 1.8}) +[2026-06-24 05:19:59,572][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.13137589395046234, 'train/grad_norm_pre_clip': 19.41554069519043, 'timing/data': 0.0007336749986279756, 'timing/model': 1.1769999800017104, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 1.8}) +[2026-06-24 05:20:03,870][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.0022694526705890894, 'train/grad_norm_pre_clip': 9.02139663696289, 'timing/data': 0.0006768289895262569, 'timing/model': 1.1694002919830382, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 1.8}) +[2026-06-24 05:20:08,217][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.02052776701748371, 'train/grad_norm_pre_clip': 12.206923484802246, 'timing/data': 0.0011287640081718564, 'timing/model': 1.1718669740075711, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 1.8}) +[2026-06-24 05:20:12,701][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.10941734910011292, 'train/grad_norm_pre_clip': 36.76958465576172, 'timing/data': 0.0007771050150040537, 'timing/model': 1.17012978400453, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 1.8}) +[2026-06-24 05:20:17,014][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.0555107481777668, 'train/grad_norm_pre_clip': 30.619461059570312, 'timing/data': 0.0007868229877203703, 'timing/model': 1.1737458079878706, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 1.8}) +[2026-06-24 05:20:21,374][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.0011792221339419484, 'train/grad_norm_pre_clip': 14.772061347961426, 'timing/data': 0.001425262016709894, 'timing/model': 1.194238629017491, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 1.8}) +[2026-06-24 05:20:25,743][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.01610635779798031, 'train/grad_norm_pre_clip': 28.669336318969727, 'timing/data': 0.0010079609928652644, 'timing/model': 1.199268333002692, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 1.8}) +[2026-06-24 05:20:30,260][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.052337970584630966, 'train/grad_norm_pre_clip': 19.1143798828125, 'timing/data': 0.0008052070043049753, 'timing/model': 1.2065534350113012, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 1.81}) +[2026-06-24 05:20:34,584][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.011549131013453007, 'train/grad_norm_pre_clip': 10.461419105529785, 'timing/data': 0.0007471200078725815, 'timing/model': 1.193392135988688, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 1.81}) +[2026-06-24 05:20:38,951][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.11289609223604202, 'train/grad_norm_pre_clip': 24.37415885925293, 'timing/data': 0.0009790780022740364, 'timing/model': 1.1903023759950884, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 1.81}) +[2026-06-24 05:20:43,318][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.0012855938402935863, 'train/grad_norm_pre_clip': 14.050997734069824, 'timing/data': 0.0007340959855355322, 'timing/model': 1.1942145979846828, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 1.81}) +[2026-06-24 05:20:47,853][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.04407863691449165, 'train/grad_norm_pre_clip': 10.078935623168945, 'timing/data': 0.0008962049905676395, 'timing/model': 1.1957875789958052, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 1.81}) +[2026-06-24 05:20:52,175][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.0017870324663817883, 'train/grad_norm_pre_clip': 22.502099990844727, 'timing/data': 0.0007114840263966471, 'timing/model': 1.174443718016846, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 1.81}) +[2026-06-24 05:20:56,526][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.03753505274653435, 'train/grad_norm_pre_clip': 16.312362670898438, 'timing/data': 0.0011750499834306538, 'timing/model': 1.189025702973595, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 1.81}) +[2026-06-24 05:21:00,838][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.0012188672553747892, 'train/grad_norm_pre_clip': 11.020618438720703, 'timing/data': 0.0007427919772453606, 'timing/model': 1.177349948004121, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 1.81}) +[2026-06-24 05:21:05,410][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.03887820243835449, 'train/grad_norm_pre_clip': 28.860647201538086, 'timing/data': 0.0011348059924785048, 'timing/model': 1.1855483539984562, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 1.81}) +[2026-06-24 05:21:09,756][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.005765929352492094, 'train/grad_norm_pre_clip': 33.055206298828125, 'timing/data': 0.0007372319814749062, 'timing/model': 1.195598221005639, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 1.81}) +[2026-06-24 05:21:14,119][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.03511834889650345, 'train/grad_norm_pre_clip': 22.16359519958496, 'timing/data': 0.0008273079874925315, 'timing/model': 1.1918630590080284, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 1.81}) +[2026-06-24 05:21:18,431][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.04502556100487709, 'train/grad_norm_pre_clip': 24.26964569091797, 'timing/data': 0.0007380740134976804, 'timing/model': 1.1834168860223144, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 1.81}) +[2026-06-24 05:21:22,971][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.05206071212887764, 'train/grad_norm_pre_clip': 27.92671775817871, 'timing/data': 0.000770473008742556, 'timing/model': 1.1898787760001142, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 1.81}) +[2026-06-24 05:21:27,319][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.002856561914086342, 'train/grad_norm_pre_clip': 2.522547721862793, 'timing/data': 0.0009231750154867768, 'timing/model': 1.177057552995393, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 1.81}) +[2026-06-24 05:21:31,651][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.053242966532707214, 'train/grad_norm_pre_clip': 29.901113510131836, 'timing/data': 0.0009673259919509292, 'timing/model': 1.1963551079970784, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 1.81}) +[2026-06-24 05:21:36,010][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.08003386110067368, 'train/grad_norm_pre_clip': 23.408872604370117, 'timing/data': 0.000982413999736309, 'timing/model': 1.1750791020167526, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 1.81}) +[2026-06-24 05:21:40,566][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.01336224377155304, 'train/grad_norm_pre_clip': 21.478168487548828, 'timing/data': 0.0008050670148804784, 'timing/model': 1.1910055309999734, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 1.81}) +[2026-06-24 05:21:44,893][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.013544389046728611, 'train/grad_norm_pre_clip': 12.031024932861328, 'timing/data': 0.0011300469923298806, 'timing/model': 1.1973585689847823, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 1.81}) +[2026-06-24 05:21:49,242][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.17398381233215332, 'train/grad_norm_pre_clip': 34.644901275634766, 'timing/data': 0.0008194939873646945, 'timing/model': 1.1966190590173937, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 1.81}) +[2026-06-24 05:21:53,572][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.0007388427620753646, 'train/grad_norm_pre_clip': 3.877206802368164, 'timing/data': 0.0008760379860177636, 'timing/model': 1.1826206719852053, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 1.81}) +[2026-06-24 05:21:58,130][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.018115371465682983, 'train/grad_norm_pre_clip': 5.888449192047119, 'timing/data': 0.0007684900192543864, 'timing/model': 1.1950933019979857, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 1.81}) +[2026-06-24 05:22:02,452][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.020643968135118484, 'train/grad_norm_pre_clip': 28.138731002807617, 'timing/data': 0.0008470439934171736, 'timing/model': 1.180286643008003, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 1.81}) +[2026-06-24 05:22:06,824][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.04091845080256462, 'train/grad_norm_pre_clip': 28.04443359375, 'timing/data': 0.0009623780206311494, 'timing/model': 1.2160825869941618, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 1.81}) +[2026-06-24 05:22:11,140][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.018647193908691406, 'train/grad_norm_pre_clip': 19.36749267578125, 'timing/data': 0.0008774699817877263, 'timing/model': 1.1732633120263927, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 1.81}) +[2026-06-24 05:22:15,676][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.0029892418533563614, 'train/grad_norm_pre_clip': 18.792747497558594, 'timing/data': 0.0008982889994513243, 'timing/model': 1.19737466800143, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 1.81}) +[2026-06-24 05:22:19,973][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.0002988986379932612, 'train/grad_norm_pre_clip': 16.279048919677734, 'timing/data': 0.0007537819910794497, 'timing/model': 1.1769976590003353, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 1.81}) +[2026-06-24 05:22:24,335][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.006671267095953226, 'train/grad_norm_pre_clip': 8.697405815124512, 'timing/data': 0.0009507859940640628, 'timing/model': 1.1933403679868206, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 1.82}) +[2026-06-24 05:22:28,663][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.04207342490553856, 'train/grad_norm_pre_clip': 33.0857048034668, 'timing/data': 0.0008134219970088452, 'timing/model': 1.174991495005088, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 1.82}) +[2026-06-24 05:22:33,219][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.005883082281798124, 'train/grad_norm_pre_clip': 42.32795333862305, 'timing/data': 0.0007652229978702962, 'timing/model': 1.1972363039967604, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 1.82}) +[2026-06-24 05:22:37,566][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.020624881610274315, 'train/grad_norm_pre_clip': 50.030704498291016, 'timing/data': 0.0011118529946543276, 'timing/model': 1.1916010309942067, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 1.82}) +[2026-06-24 05:22:41,916][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.0018076772103086114, 'train/grad_norm_pre_clip': 13.570959091186523, 'timing/data': 0.0009334839996881783, 'timing/model': 1.1957486559986137, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 1.82}) +[2026-06-24 05:22:46,260][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.05449794977903366, 'train/grad_norm_pre_clip': 16.093891143798828, 'timing/data': 0.0009293269831687212, 'timing/model': 1.174863808992086, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 1.82}) +[2026-06-24 05:22:50,771][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.017344653606414795, 'train/grad_norm_pre_clip': 15.250938415527344, 'timing/data': 0.0007408990059047937, 'timing/model': 1.1961521320044994, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 1.82}) +[2026-06-24 05:22:55,112][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.05867713317275047, 'train/grad_norm_pre_clip': 35.273380279541016, 'timing/data': 0.0009071950044017285, 'timing/model': 1.1942013329826295, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 1.82}) +[2026-06-24 05:22:59,476][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.023010121658444405, 'train/grad_norm_pre_clip': 17.34503173828125, 'timing/data': 0.0007784180052112788, 'timing/model': 1.2087882889900357, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 1.82}) +[2026-06-24 05:23:03,813][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.08779304474592209, 'train/grad_norm_pre_clip': 44.46645736694336, 'timing/data': 0.0007706840115133673, 'timing/model': 1.1782505339942873, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 1.82}) +[2026-06-24 05:23:08,334][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4700 +[2026-06-24 05:24:05,857][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4700 +[2026-06-24 05:24:05,858][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4700 +[2026-06-24 05:24:42,749][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4700 +[2026-06-24 05:24:42,751][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.002826534677296877, 'train/grad_norm_pre_clip': 23.770713806152344, 'eval/action_loss/seconds': 57.519694719987456, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.034517042338848114, 'eval/latency_2/loss': 0.034517042338848114, 'eval/flappy/loss': 0.034517042338848114, 'eval/flappy/latency_2/loss': 0.034517042338848114, 'eval/flap_precision': 0.977027027027027, 'eval/flap_recall': 0.9810040705563093, 'eval/flap_f1': 0.9790115098171971, 'eval/flap_rate_model': 0.20555555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9790115098171971, 'eval/latency_2/flap_precision': 0.977027027027027, 'eval/latency_2/flap_recall': 0.9810040705563093, 'eval/latency_2/flap_f1': 0.9790115098171971, 'eval/latency_2/flap_rate_model': 0.20555555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9790115098171971, 'eval/action_classification/data_wait_seconds': 0.4294879138469696, 'eval/action_classification/predict_seconds': 36.31730270385742, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.000899981998372823, 'timing/model': 1.1781369200034533, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 1.82}) +[2026-06-24 05:24:47,029][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.035306815057992935, 'train/grad_norm_pre_clip': 34.31053161621094, 'timing/data': 0.0007518990023527294, 'timing/model': 1.1572914460266475, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 1.82}) +[2026-06-24 05:24:51,291][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.005143280141055584, 'train/grad_norm_pre_clip': 15.074942588806152, 'timing/data': 0.0007342459866777062, 'timing/model': 1.1592657269793563, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 1.82}) +[2026-06-24 05:24:55,813][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.15857566893100739, 'train/grad_norm_pre_clip': 41.23337936401367, 'timing/data': 0.0008000980014912784, 'timing/model': 1.1786781339906156, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 1.82}) +[2026-06-24 05:25:00,110][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.07736647129058838, 'train/grad_norm_pre_clip': 43.30769729614258, 'timing/data': 0.0007813030097167939, 'timing/model': 1.1840084140130784, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 1.82}) +[2026-06-24 05:25:04,402][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.03549494966864586, 'train/grad_norm_pre_clip': 13.957901000976562, 'timing/data': 0.0008628640207462013, 'timing/model': 1.1984433489851654, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 1.82}) +[2026-06-24 05:25:08,684][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.0029580912087112665, 'train/grad_norm_pre_clip': 3.1435039043426514, 'timing/data': 0.0008912560006137937, 'timing/model': 1.167550603975542, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 1.82}) +[2026-06-24 05:25:13,229][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.105593241751194, 'train/grad_norm_pre_clip': 34.37538528442383, 'timing/data': 0.0008029120217543095, 'timing/model': 1.1779039399989415, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 1.82}) +[2026-06-24 05:25:17,599][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.16715136170387268, 'train/grad_norm_pre_clip': 52.314735412597656, 'timing/data': 0.0010735740070231259, 'timing/model': 1.2074219640053343, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 1.82}) +[2026-06-24 05:25:21,876][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.11661579459905624, 'train/grad_norm_pre_clip': 36.67383575439453, 'timing/data': 0.0007502960215788335, 'timing/model': 1.1737423810118344, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 1.82}) +[2026-06-24 05:25:26,182][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.006052803713828325, 'train/grad_norm_pre_clip': 11.942505836486816, 'timing/data': 0.0009644820238463581, 'timing/model': 1.2016532970010303, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 1.82}) +[2026-06-24 05:25:30,771][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.042568083852529526, 'train/grad_norm_pre_clip': 25.381500244140625, 'timing/data': 0.00082248798571527, 'timing/model': 1.1826701179961674, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 1.82}) +[2026-06-24 05:25:35,135][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.07165683060884476, 'train/grad_norm_pre_clip': 27.578088760375977, 'timing/data': 0.0007996979984454811, 'timing/model': 1.1941407379927114, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 1.82}) +[2026-06-24 05:25:39,464][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.0037038957234472036, 'train/grad_norm_pre_clip': 29.575775146484375, 'timing/data': 0.0007624480058439076, 'timing/model': 1.1948188839887735, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 1.82}) +[2026-06-24 05:25:43,809][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.05575604364275932, 'train/grad_norm_pre_clip': 13.461275100708008, 'timing/data': 0.0007876950257923454, 'timing/model': 1.189216258993838, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 1.82}) +[2026-06-24 05:25:48,384][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.002378419041633606, 'train/grad_norm_pre_clip': 11.344095230102539, 'timing/data': 0.0007490340212825686, 'timing/model': 1.1697232899896335, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 1.82}) +[2026-06-24 05:25:52,716][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.10409844666719437, 'train/grad_norm_pre_clip': 12.715384483337402, 'timing/data': 0.0008033339981921017, 'timing/model': 1.1949468920065556, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 1.83}) +[2026-06-24 05:25:57,042][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.06867528706789017, 'train/grad_norm_pre_clip': 39.19667053222656, 'timing/data': 0.0008165189938154072, 'timing/model': 1.16735228698235, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 1.83}) +[2026-06-24 05:26:01,358][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.04621046781539917, 'train/grad_norm_pre_clip': 16.6280574798584, 'timing/data': 0.0007406479853671044, 'timing/model': 1.1984050160099287, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 1.83}) +[2026-06-24 05:26:05,864][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.1250656247138977, 'train/grad_norm_pre_clip': 21.700332641601562, 'timing/data': 0.0007530009897891432, 'timing/model': 1.167336444981629, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 1.83}) +[2026-06-24 05:26:10,191][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.017219359055161476, 'train/grad_norm_pre_clip': 12.439807891845703, 'timing/data': 0.0008956240199040622, 'timing/model': 1.1916167839954142, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 1.83}) +[2026-06-24 05:26:14,524][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.01268225722014904, 'train/grad_norm_pre_clip': 23.710325241088867, 'timing/data': 0.0007187369919847697, 'timing/model': 1.190817470982438, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 1.83}) +[2026-06-24 05:26:18,830][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 4.7659777919761837e-05, 'train/grad_norm_pre_clip': 4.577931880950928, 'timing/data': 0.0007394459971692413, 'timing/model': 1.1707103360095061, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 1.83}) +[2026-06-24 05:26:23,406][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.015291481278836727, 'train/grad_norm_pre_clip': 8.899542808532715, 'timing/data': 0.0007297580013982952, 'timing/model': 1.1888712139916606, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 1.83}) +[2026-06-24 05:26:27,741][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.006984109058976173, 'train/grad_norm_pre_clip': 4.308996677398682, 'timing/data': 0.0007628590101376176, 'timing/model': 1.182089520007139, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 1.83}) +[2026-06-24 05:26:32,123][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.0017799679189920425, 'train/grad_norm_pre_clip': 19.57269859313965, 'timing/data': 0.001109979988541454, 'timing/model': 1.2009166079806164, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 1.83}) +[2026-06-24 05:26:36,498][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.010520972311496735, 'train/grad_norm_pre_clip': 9.370657920837402, 'timing/data': 0.0009987549856305122, 'timing/model': 1.1771050299867056, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 1.83}) +[2026-06-24 05:26:41,067][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.004372101277112961, 'train/grad_norm_pre_clip': 3.1922898292541504, 'timing/data': 0.0007427020173054188, 'timing/model': 1.183839823002927, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 1.83}) +[2026-06-24 05:26:45,409][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.030599480494856834, 'train/grad_norm_pre_clip': 15.964567184448242, 'timing/data': 0.0009411289938725531, 'timing/model': 1.1972330950084142, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 1.83}) +[2026-06-24 05:26:49,810][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.0002530580386519432, 'train/grad_norm_pre_clip': 17.0626220703125, 'timing/data': 0.0013116930203977972, 'timing/model': 1.1916719159926288, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 1.83}) +[2026-06-24 05:26:54,227][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.011754818260669708, 'train/grad_norm_pre_clip': 11.3275146484375, 'timing/data': 0.0008328489784616977, 'timing/model': 1.1934027159877587, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 1.83}) +[2026-06-24 05:26:58,820][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.020643679425120354, 'train/grad_norm_pre_clip': 8.659502029418945, 'timing/data': 0.0011881249956786633, 'timing/model': 1.1927708419971168, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 1.83}) +[2026-06-24 05:27:03,141][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.005373993888497353, 'train/grad_norm_pre_clip': 9.42734432220459, 'timing/data': 0.0009271219896618277, 'timing/model': 1.1746098799922038, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 1.83}) +[2026-06-24 05:27:07,475][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.03854038938879967, 'train/grad_norm_pre_clip': 23.17731285095215, 'timing/data': 0.000776564993429929, 'timing/model': 1.1787493790034205, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 1.83}) +[2026-06-24 05:27:11,846][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.00048624389455653727, 'train/grad_norm_pre_clip': 17.333221435546875, 'timing/data': 0.0009441040165256709, 'timing/model': 1.1916386699886061, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 1.83}) +[2026-06-24 05:27:16,433][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.001393150188960135, 'train/grad_norm_pre_clip': 3.600010633468628, 'timing/data': 0.0009608440159354359, 'timing/model': 1.1976529019884765, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 1.83}) +[2026-06-24 05:27:20,815][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.04394657909870148, 'train/grad_norm_pre_clip': 54.90045928955078, 'timing/data': 0.0008125210006255656, 'timing/model': 1.2035033919964917, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 1.83}) +[2026-06-24 05:27:25,163][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.010971571318805218, 'train/grad_norm_pre_clip': 20.295883178710938, 'timing/data': 0.0007698120025452226, 'timing/model': 1.1874190050002653, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 1.83}) +[2026-06-24 05:27:29,511][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.0001101763264159672, 'train/grad_norm_pre_clip': 69.50643157958984, 'timing/data': 0.0008227190119214356, 'timing/model': 1.1763759000168648, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 1.83}) +[2026-06-24 05:27:34,073][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.0009132565464824438, 'train/grad_norm_pre_clip': 26.554039001464844, 'timing/data': 0.0009181849891319871, 'timing/model': 1.1738783000037074, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 1.83}) +[2026-06-24 05:27:38,408][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.005431806202977896, 'train/grad_norm_pre_clip': 10.121869087219238, 'timing/data': 0.0007869330001994967, 'timing/model': 1.1771342769789044, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 1.83}) +[2026-06-24 05:27:42,719][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.04923133924603462, 'train/grad_norm_pre_clip': 20.370424270629883, 'timing/data': 0.0009124450152739882, 'timing/model': 1.1906858010042924, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 1.83}) +[2026-06-24 05:27:47,033][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.0027280887588858604, 'train/grad_norm_pre_clip': 22.187114715576172, 'timing/data': 0.0007687000033911318, 'timing/model': 1.1893059249850921, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 1.84}) +[2026-06-24 05:27:51,571][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.04303360357880592, 'train/grad_norm_pre_clip': 36.790950775146484, 'timing/data': 0.0007750119839329273, 'timing/model': 1.179649360012263, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 1.84}) +[2026-06-24 05:27:55,929][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.005626547150313854, 'train/grad_norm_pre_clip': 13.311570167541504, 'timing/data': 0.0008831110026221722, 'timing/model': 1.1885397040168755, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 1.84}) +[2026-06-24 05:28:00,279][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.08324655890464783, 'train/grad_norm_pre_clip': 27.51802635192871, 'timing/data': 0.0007896890165284276, 'timing/model': 1.2062450209923554, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 1.84}) +[2026-06-24 05:28:04,599][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.1339988261461258, 'train/grad_norm_pre_clip': 14.665376663208008, 'timing/data': 0.0008903149864636362, 'timing/model': 1.1730830000014976, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 1.84}) +[2026-06-24 05:28:09,206][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.015466065146028996, 'train/grad_norm_pre_clip': 27.939603805541992, 'timing/data': 0.0007205310103017837, 'timing/model': 1.198753709992161, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 1.84}) +[2026-06-24 05:28:13,554][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 3.311934051453136e-05, 'train/grad_norm_pre_clip': 37.47370910644531, 'timing/data': 0.0008714099822100252, 'timing/model': 1.2127589910232928, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 1.84}) +[2026-06-24 05:28:17,899][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.011952336877584457, 'train/grad_norm_pre_clip': 28.84505844116211, 'timing/data': 0.0007506159890908748, 'timing/model': 1.1946236260118894, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 1.84}) +[2026-06-24 05:56:39,291][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4750 +[2026-06-24 05:56:39,293][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.013664384372532368, 'train/grad_norm_pre_clip': 15.092140197753906, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0008571539947297424, 'timing/model': 1.1948086810007226, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 1.84}) +[2026-06-24 05:56:43,502][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.21853381395339966, 'train/grad_norm_pre_clip': 24.31388282775879, 'timing/data': 0.0008848939905874431, 'timing/model': 1.1605141990003176, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 1.84}) +[2026-06-24 05:56:47,697][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.002674797549843788, 'train/grad_norm_pre_clip': 25.511510848999023, 'timing/data': 0.0009374819928780198, 'timing/model': 1.1631367880036123, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 1.84}) +[2026-06-24 05:56:51,895][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.020699189975857735, 'train/grad_norm_pre_clip': 18.132654190063477, 'timing/data': 0.0008266860095318407, 'timing/model': 1.1420039600052405, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 1.84}) +[2026-06-24 05:56:56,362][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.02912609465420246, 'train/grad_norm_pre_clip': 12.263442993164062, 'timing/data': 0.0007645820151083171, 'timing/model': 1.157020526996348, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 1.84}) +[2026-06-24 05:57:00,538][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.0021921724546700716, 'train/grad_norm_pre_clip': 11.535222053527832, 'timing/data': 0.0012366839800961316, 'timing/model': 1.1373576540208887, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 1.84}) +[2026-06-24 05:57:04,762][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.030549652874469757, 'train/grad_norm_pre_clip': 34.4039192199707, 'timing/data': 0.0008155159885063767, 'timing/model': 1.1603457590099424, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 1.84}) +[2026-06-24 05:57:08,952][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.0023053735494613647, 'train/grad_norm_pre_clip': 11.455693244934082, 'timing/data': 0.0007460479973815382, 'timing/model': 1.1436242670170031, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 1.84}) +[2026-06-24 05:57:13,379][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.027753274887800217, 'train/grad_norm_pre_clip': 49.614925384521484, 'timing/data': 0.0010094140016008168, 'timing/model': 1.1447680160054006, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 1.84}) +[2026-06-24 05:57:17,583][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.07900656014680862, 'train/grad_norm_pre_clip': 22.07685661315918, 'timing/data': 0.000776494009187445, 'timing/model': 1.1460994399967603, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 1.84}) +[2026-06-24 05:57:21,796][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.02674654684960842, 'train/grad_norm_pre_clip': 13.53726863861084, 'timing/data': 0.0007678379770368338, 'timing/model': 1.146937063022051, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 1.84}) +[2026-06-24 05:57:26,013][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.030842851847410202, 'train/grad_norm_pre_clip': 6.2501444816589355, 'timing/data': 0.000820204004412517, 'timing/model': 1.1717153250065167, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 1.84}) +[2026-06-24 05:57:30,432][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.06375443935394287, 'train/grad_norm_pre_clip': 19.061119079589844, 'timing/data': 0.001343059993814677, 'timing/model': 1.1647620009898674, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 1.84}) +[2026-06-24 05:57:34,688][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.02257792465388775, 'train/grad_norm_pre_clip': 23.166288375854492, 'timing/data': 0.0013577869976870716, 'timing/model': 1.1746566589863505, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 1.84}) +[2026-06-24 05:57:38,916][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.061299681663513184, 'train/grad_norm_pre_clip': 12.663034439086914, 'timing/data': 0.0008114179945550859, 'timing/model': 1.1610517209919635, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 1.84}) +[2026-06-24 05:57:43,199][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.0361824631690979, 'train/grad_norm_pre_clip': 40.343448638916016, 'timing/data': 0.0007745000184513628, 'timing/model': 1.195613490010146, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 1.84}) +[2026-06-24 05:57:47,693][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.01633036881685257, 'train/grad_norm_pre_clip': 20.50360679626465, 'timing/data': 0.0007338049763347954, 'timing/model': 1.393548470980022, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 1.84}) +[2026-06-24 05:57:51,900][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.007950332947075367, 'train/grad_norm_pre_clip': 26.15077018737793, 'timing/data': 0.0008061980188358575, 'timing/model': 1.1503341970092151, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 1.84}) +[2026-06-24 05:57:56,118][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.07507472485303879, 'train/grad_norm_pre_clip': 34.79844665527344, 'timing/data': 0.0007950089930091053, 'timing/model': 1.1539947930141352, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 1.85}) +[2026-06-24 05:58:00,447][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.00033202662598341703, 'train/grad_norm_pre_clip': 8.767788887023926, 'timing/data': 0.000794536987086758, 'timing/model': 1.1995334030070808, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 1.85}) +[2026-06-24 05:58:04,954][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.020584329962730408, 'train/grad_norm_pre_clip': 12.9634428024292, 'timing/data': 0.0008422649989370257, 'timing/model': 1.1647438239888288, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 1.85}) +[2026-06-24 05:58:09,195][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.003974811639636755, 'train/grad_norm_pre_clip': 25.563400268554688, 'timing/data': 0.0007618670060764998, 'timing/model': 1.1591294819954783, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 1.85}) +[2026-06-24 05:58:13,457][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.1494380384683609, 'train/grad_norm_pre_clip': 34.6744270324707, 'timing/data': 0.0009631079738028347, 'timing/model': 1.1681005829887, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 1.85}) +[2026-06-24 05:58:17,733][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.0006806752062402666, 'train/grad_norm_pre_clip': 8.37314224243164, 'timing/data': 0.00075960299000144, 'timing/model': 1.1619955959904473, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 1.85}) +[2026-06-24 05:58:22,239][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.056091178208589554, 'train/grad_norm_pre_clip': 27.389053344726562, 'timing/data': 0.0007644209836144, 'timing/model': 1.1923110139905475, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 1.85}) +[2026-06-24 05:58:26,571][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.007231628987938166, 'train/grad_norm_pre_clip': 18.735055923461914, 'timing/data': 0.000674555980367586, 'timing/model': 1.1795074559922796, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 1.85}) +[2026-06-24 05:58:30,890][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.040316738188266754, 'train/grad_norm_pre_clip': 24.725088119506836, 'timing/data': 0.0007399269961751997, 'timing/model': 1.1817307470191736, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 1.85}) +[2026-06-24 05:58:35,142][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.011828449554741383, 'train/grad_norm_pre_clip': 7.972330570220947, 'timing/data': 0.0007056329923216254, 'timing/model': 1.1617438230023254, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 1.85}) +[2026-06-24 05:58:39,679][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.15698979794979095, 'train/grad_norm_pre_clip': 31.806964874267578, 'timing/data': 0.0008607000054325908, 'timing/model': 1.1610528320015874, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 1.85}) +[2026-06-24 05:58:43,929][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.010501103475689888, 'train/grad_norm_pre_clip': 22.72092628479004, 'timing/data': 0.0007057029870338738, 'timing/model': 1.1607612780062482, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 1.85}) +[2026-06-24 05:58:48,204][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.01857512630522251, 'train/grad_norm_pre_clip': 25.478321075439453, 'timing/data': 0.0007473599980585277, 'timing/model': 1.16390286598471, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 1.85}) +[2026-06-24 05:58:52,462][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.011225908063352108, 'train/grad_norm_pre_clip': 27.39713478088379, 'timing/data': 0.0007125959964469075, 'timing/model': 1.1648726209823508, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 1.85}) +[2026-06-24 05:58:56,948][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.0210859477519989, 'train/grad_norm_pre_clip': 26.773038864135742, 'timing/data': 0.0009055820119101554, 'timing/model': 1.1866911700053606, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 1.85}) +[2026-06-24 05:59:01,228][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.002046783221885562, 'train/grad_norm_pre_clip': 2.176060199737549, 'timing/data': 0.000731951993657276, 'timing/model': 1.1841771219915245, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 1.85}) +[2026-06-24 05:59:05,493][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.0030992222018539906, 'train/grad_norm_pre_clip': 15.462559700012207, 'timing/data': 0.00088029500329867, 'timing/model': 1.1633279240049887, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 1.85}) +[2026-06-24 05:59:09,806][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.005899352021515369, 'train/grad_norm_pre_clip': 28.3972225189209, 'timing/data': 0.0007608250016346574, 'timing/model': 1.1891310409992002, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 1.85}) +[2026-06-24 05:59:14,307][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.06626089662313461, 'train/grad_norm_pre_clip': 33.21260070800781, 'timing/data': 0.0009733480110298842, 'timing/model': 1.1699007819988765, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 1.85}) +[2026-06-24 05:59:18,601][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.0019684345461428165, 'train/grad_norm_pre_clip': 19.73883819580078, 'timing/data': 0.0008537769899703562, 'timing/model': 1.1669408819871023, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 1.85}) +[2026-06-24 05:59:22,873][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.0030808113515377045, 'train/grad_norm_pre_clip': 7.196026802062988, 'timing/data': 0.0008752860012464225, 'timing/model': 1.1649464520160109, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 1.85}) +[2026-06-24 05:59:27,212][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.13565611839294434, 'train/grad_norm_pre_clip': 33.78559875488281, 'timing/data': 0.0007463079818990082, 'timing/model': 1.1787811960093677, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 1.85}) +[2026-06-24 05:59:31,750][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.06103729084134102, 'train/grad_norm_pre_clip': 26.60939598083496, 'timing/data': 0.0008666210051160306, 'timing/model': 1.1928391530236695, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 1.85}) +[2026-06-24 05:59:36,068][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.03586431220173836, 'train/grad_norm_pre_clip': 16.30610466003418, 'timing/data': 0.0007178859959822148, 'timing/model': 1.1777500720054377, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 1.85}) +[2026-06-24 05:59:40,420][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.005175963509827852, 'train/grad_norm_pre_clip': 9.787869453430176, 'timing/data': 0.000877559999935329, 'timing/model': 1.1922566069988534, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 1.85}) +[2026-06-24 05:59:44,742][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.002891132840886712, 'train/grad_norm_pre_clip': 4.780599117279053, 'timing/data': 0.0008733230060897768, 'timing/model': 1.1701399680168834, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 1.85}) +[2026-06-24 05:59:49,261][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.04518377408385277, 'train/grad_norm_pre_clip': 30.244447708129883, 'timing/data': 0.0007687000033911318, 'timing/model': 1.1880005059938412, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 1.86}) +[2026-06-24 05:59:53,564][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.00017573825607541949, 'train/grad_norm_pre_clip': 21.169235229492188, 'timing/data': 0.0007585209968965501, 'timing/model': 1.1705633290112019, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 1.86}) +[2026-06-24 05:59:57,869][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.0052437796257436275, 'train/grad_norm_pre_clip': 9.144735336303711, 'timing/data': 0.0007334640249609947, 'timing/model': 1.194325768010458, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 1.86}) +[2026-06-24 06:00:02,154][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.025226309895515442, 'train/grad_norm_pre_clip': 27.335289001464844, 'timing/data': 0.0007635000220034271, 'timing/model': 1.171137038996676, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 1.86}) +[2026-06-24 06:00:06,705][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.07902384549379349, 'train/grad_norm_pre_clip': 21.03037452697754, 'timing/data': 0.0007770549855194986, 'timing/model': 1.1715831370092928, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 1.86}) +[2026-06-24 06:00:11,038][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.0009699785150587559, 'train/grad_norm_pre_clip': 11.6293306350708, 'timing/data': 0.0007555149786639959, 'timing/model': 1.1726884529925883, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 1.86}) +[2026-06-24 06:00:15,327][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4800 +[2026-06-24 06:01:13,041][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4800 +[2026-06-24 06:01:13,041][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4800 +[2026-06-24 06:01:49,784][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4800 +[2026-06-24 06:01:49,786][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.06688099354505539, 'train/grad_norm_pre_clip': 31.6540470123291, 'eval/action_loss/seconds': 57.71005081498879, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.0349661223590374, 'eval/latency_2/loss': 0.0349661223590374, 'eval/flappy/loss': 0.0349661223590374, 'eval/flappy/latency_2/loss': 0.0349661223590374, 'eval/flap_precision': 0.9757085020242915, 'eval/flap_recall': 0.9810040705563093, 'eval/flap_f1': 0.9783491204330177, 'eval/flap_rate_model': 0.20583333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9783491204330177, 'eval/latency_2/flap_precision': 0.9757085020242915, 'eval/latency_2/flap_recall': 0.9810040705563093, 'eval/latency_2/flap_f1': 0.9783491204330177, 'eval/latency_2/flap_rate_model': 0.20583333333333334, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9783491204330177, 'eval/action_classification/data_wait_seconds': 0.4601963758468628, 'eval/action_classification/predict_seconds': 36.13789749145508, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007297279953490943, 'timing/model': 1.1713822039891966, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 1.86}) +[2026-06-24 06:01:54,337][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.028659280389547348, 'train/grad_norm_pre_clip': 15.404535293579102, 'timing/data': 0.001201748993480578, 'timing/model': 1.1772007520194165, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 1.86}) +[2026-06-24 06:01:58,642][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.07986512780189514, 'train/grad_norm_pre_clip': 20.882877349853516, 'timing/data': 0.0007511269941460341, 'timing/model': 1.1879473039880395, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 1.86}) +[2026-06-24 06:02:02,970][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.0820603147149086, 'train/grad_norm_pre_clip': 36.00775146484375, 'timing/data': 0.0007096909976098686, 'timing/model': 1.1925189190078527, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 1.86}) +[2026-06-24 06:02:07,294][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.04451686888933182, 'train/grad_norm_pre_clip': 26.070594787597656, 'timing/data': 0.0008190629887394607, 'timing/model': 1.188143722974928, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 1.86}) +[2026-06-24 06:02:11,916][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.023623475804924965, 'train/grad_norm_pre_clip': 19.423742294311523, 'timing/data': 0.0009151500125881284, 'timing/model': 1.1815064980182797, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 1.86}) +[2026-06-24 06:02:16,251][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.008150347508490086, 'train/grad_norm_pre_clip': 32.98566436767578, 'timing/data': 0.0007574790215585381, 'timing/model': 1.1911487000179477, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 1.86}) +[2026-06-24 06:02:20,591][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.001015559653751552, 'train/grad_norm_pre_clip': 15.660316467285156, 'timing/data': 0.0007301779987756163, 'timing/model': 1.185826693981653, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 1.86}) +[2026-06-24 06:02:24,909][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.04044613987207413, 'train/grad_norm_pre_clip': 21.29203224182129, 'timing/data': 0.0007374020060524344, 'timing/model': 1.1911647870147135, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 1.86}) +[2026-06-24 06:02:29,391][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.017186444252729416, 'train/grad_norm_pre_clip': 19.226795196533203, 'timing/data': 0.0009304979757871479, 'timing/model': 1.1744232210039627, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 1.86}) +[2026-06-24 06:02:33,725][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.05043257400393486, 'train/grad_norm_pre_clip': 14.097602844238281, 'timing/data': 0.0008297129825223237, 'timing/model': 1.1804115539998747, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 1.86}) +[2026-06-24 06:02:38,023][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.005299891345202923, 'train/grad_norm_pre_clip': 14.506902694702148, 'timing/data': 0.0007510569994337857, 'timing/model': 1.1692666540038772, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 1.86}) +[2026-06-24 06:02:42,352][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.011345133185386658, 'train/grad_norm_pre_clip': 29.555805206298828, 'timing/data': 0.0009041299927048385, 'timing/model': 1.1915668940055184, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 1.86}) +[2026-06-24 06:02:46,890][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.0851648598909378, 'train/grad_norm_pre_clip': 39.201927185058594, 'timing/data': 0.0007424209907185286, 'timing/model': 1.1717905020050239, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 1.86}) +[2026-06-24 06:02:51,208][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.0763343945145607, 'train/grad_norm_pre_clip': 30.33086395263672, 'timing/data': 0.000898409984074533, 'timing/model': 1.1714172259962652, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 1.86}) +[2026-06-24 06:02:55,540][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.0022614779882133007, 'train/grad_norm_pre_clip': 2.1283812522888184, 'timing/data': 0.0007744199829176068, 'timing/model': 1.192320866015507, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 1.86}) +[2026-06-24 06:03:00,076][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.001996884122490883, 'train/grad_norm_pre_clip': 31.056541442871094, 'timing/data': 0.0009443440067116171, 'timing/model': 1.3608465409779456, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 1.86}) +[2026-06-24 06:03:04,378][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.0007312948582693934, 'train/grad_norm_pre_clip': 29.796154022216797, 'timing/data': 0.0007320120057556778, 'timing/model': 1.169279737980105, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 1.86}) +[2026-06-24 06:03:08,725][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.00033228195388801396, 'train/grad_norm_pre_clip': 12.717612266540527, 'timing/data': 0.0013810500095132738, 'timing/model': 1.1727313860028517, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 1.86}) +[2026-06-24 06:03:13,055][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.013999865390360355, 'train/grad_norm_pre_clip': 23.609193801879883, 'timing/data': 0.0008108580077532679, 'timing/model': 1.1767750779981725, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 1.87}) +[2026-06-24 06:03:17,574][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.05762643367052078, 'train/grad_norm_pre_clip': 21.387514114379883, 'timing/data': 0.0007798009901307523, 'timing/model': 1.3607279569841921, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 1.87}) +[2026-06-24 06:03:21,937][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.08296862244606018, 'train/grad_norm_pre_clip': 47.72813034057617, 'timing/data': 0.001124587986851111, 'timing/model': 1.1820967720123008, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 1.87}) +[2026-06-24 06:03:26,259][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.02992556244134903, 'train/grad_norm_pre_clip': 33.34235382080078, 'timing/data': 0.0011177950073033571, 'timing/model': 1.1756537559849676, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 1.87}) +[2026-06-24 06:03:30,596][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.0636969730257988, 'train/grad_norm_pre_clip': 13.003690719604492, 'timing/data': 0.0009466990013606846, 'timing/model': 1.197117169998819, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 1.87}) +[2026-06-24 06:03:35,170][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.0007910293061286211, 'train/grad_norm_pre_clip': 21.54524803161621, 'timing/data': 0.001115099003072828, 'timing/model': 1.4255689480050933, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 1.87}) +[2026-06-24 06:03:39,509][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.09115245938301086, 'train/grad_norm_pre_clip': 40.391265869140625, 'timing/data': 0.001053024985594675, 'timing/model': 1.1752330890158191, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 1.87}) +[2026-06-24 06:03:43,841][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.044842641800642014, 'train/grad_norm_pre_clip': 30.216014862060547, 'timing/data': 0.0009289150184486061, 'timing/model': 1.1755499260034412, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 1.87}) +[2026-06-24 06:03:48,206][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.015310674905776978, 'train/grad_norm_pre_clip': 16.920923233032227, 'timing/data': 0.0012413329968694597, 'timing/model': 1.2063553250045516, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 1.87}) +[2026-06-24 06:03:52,708][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.001069308491423726, 'train/grad_norm_pre_clip': 42.17741394042969, 'timing/data': 0.000802232010755688, 'timing/model': 1.1790490350103937, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 1.87}) +[2026-06-24 06:03:57,080][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.00334117840975523, 'train/grad_norm_pre_clip': 16.347673416137695, 'timing/data': 0.0011098999821115285, 'timing/model': 1.1908184950007126, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 1.87}) +[2026-06-24 06:04:01,431][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.0012035652762278914, 'train/grad_norm_pre_clip': 27.392303466796875, 'timing/data': 0.0007625780126545578, 'timing/model': 1.1789490370138083, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 1.87}) +[2026-06-24 06:04:05,784][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 2.1553119950112887e-05, 'train/grad_norm_pre_clip': 13.432125091552734, 'timing/data': 0.001036323985317722, 'timing/model': 1.2152399260085076, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 1.87}) +[2026-06-24 06:04:10,270][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.11835826188325882, 'train/grad_norm_pre_clip': 25.81093978881836, 'timing/data': 0.0008428770233877003, 'timing/model': 1.1767276799946558, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 1.87}) +[2026-06-24 06:04:14,608][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.06526345014572144, 'train/grad_norm_pre_clip': 32.26690673828125, 'timing/data': 0.0009043099998962134, 'timing/model': 1.1987892079923768, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 1.87}) +[2026-06-24 06:04:18,916][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.016414044424891472, 'train/grad_norm_pre_clip': 33.40376281738281, 'timing/data': 0.0007725870236754417, 'timing/model': 1.1789386980235577, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 1.87}) +[2026-06-24 06:04:23,290][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.008597430773079395, 'train/grad_norm_pre_clip': 3.1534669399261475, 'timing/data': 0.0011931330082006752, 'timing/model': 1.2040075020049699, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 1.87}) +[2026-06-24 06:04:27,870][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.034461118280887604, 'train/grad_norm_pre_clip': 20.223007202148438, 'timing/data': 0.0007584109844174236, 'timing/model': 1.2009260439954232, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 1.87}) +[2026-06-24 06:04:32,208][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.014533214271068573, 'train/grad_norm_pre_clip': 35.85798645019531, 'timing/data': 0.0009902189776767045, 'timing/model': 1.1828942330030259, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 1.87}) +[2026-06-24 06:04:36,526][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.055477965623140335, 'train/grad_norm_pre_clip': 22.022050857543945, 'timing/data': 0.0009518779988866299, 'timing/model': 1.173917195003014, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 1.87}) +[2026-06-24 06:04:40,899][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.0349225290119648, 'train/grad_norm_pre_clip': 9.595025062561035, 'timing/data': 0.0009228540002368391, 'timing/model': 1.1779698519967496, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 1.87}) +[2026-06-24 06:04:45,430][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.11160249263048172, 'train/grad_norm_pre_clip': 20.348312377929688, 'timing/data': 0.0007572990143671632, 'timing/model': 1.1742334250011481, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 1.87}) +[2026-06-24 06:04:49,817][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 8.271507249446586e-05, 'train/grad_norm_pre_clip': 27.475337982177734, 'timing/data': 0.0011939049873035401, 'timing/model': 1.2095734089962207, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 1.87}) +[2026-06-24 06:04:54,185][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.016293570399284363, 'train/grad_norm_pre_clip': 15.619420051574707, 'timing/data': 0.0008931199845392257, 'timing/model': 1.1996950770262629, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 1.87}) +[2026-06-24 06:04:58,545][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.007673181127756834, 'train/grad_norm_pre_clip': 4.62089204788208, 'timing/data': 0.0008249439997598529, 'timing/model': 1.1900569580029696, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 1.87}) +[2026-06-24 06:05:03,106][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.0009620141936466098, 'train/grad_norm_pre_clip': 38.56911849975586, 'timing/data': 0.0012902930029667914, 'timing/model': 1.18482719201711, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 1.87}) +[2026-06-24 06:05:07,423][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.014453100971877575, 'train/grad_norm_pre_clip': 15.12520980834961, 'timing/data': 0.000750445993617177, 'timing/model': 1.1769129970052745, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 1.88}) +[2026-06-24 06:05:11,773][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.0518697090446949, 'train/grad_norm_pre_clip': 17.99846649169922, 'timing/data': 0.0009475199913140386, 'timing/model': 1.1793374069966376, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 1.88}) +[2026-06-24 06:05:16,390][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.10498204827308655, 'train/grad_norm_pre_clip': 51.5968132019043, 'timing/data': 0.0007718660053797066, 'timing/model': 1.456285068008583, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 1.88}) +[2026-06-24 06:05:20,748][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.011907999403774738, 'train/grad_norm_pre_clip': 22.39225959777832, 'timing/data': 0.0007613469788338989, 'timing/model': 1.2006080680002924, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 1.88}) +[2026-06-24 06:05:25,095][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.016087960451841354, 'train/grad_norm_pre_clip': 28.741262435913086, 'timing/data': 0.0011612950183916837, 'timing/model': 1.2161911519942805, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 1.88}) +[2026-06-24 06:05:29,445][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.001475856639444828, 'train/grad_norm_pre_clip': 23.368576049804688, 'timing/data': 0.0008786229882389307, 'timing/model': 1.188587341021048, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 1.88}) +[2026-06-24 06:05:33,982][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.0015302728861570358, 'train/grad_norm_pre_clip': 32.67742156982422, 'timing/data': 0.0009457669802941382, 'timing/model': 1.4090713070181664, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 1.88}) +[2026-06-24 06:05:38,397][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.01025690883398056, 'train/grad_norm_pre_clip': 10.93116569519043, 'timing/data': 0.000692378991516307, 'timing/model': 1.1737057599821128, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 1.88}) +[2026-06-24 06:05:42,693][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.022405454888939857, 'train/grad_norm_pre_clip': 31.127573013305664, 'timing/data': 0.0007317109848372638, 'timing/model': 1.1710053739952855, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 1.88}) +[2026-06-24 06:05:47,027][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.0029805616941303015, 'train/grad_norm_pre_clip': 23.150033950805664, 'timing/data': 0.0006864769966341555, 'timing/model': 1.1922026579850353, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 1.88}) +[2026-06-24 06:05:51,522][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.031533487141132355, 'train/grad_norm_pre_clip': 33.019798278808594, 'timing/data': 0.0007259609992615879, 'timing/model': 1.1703620009939186, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 1.88}) +[2026-06-24 06:05:55,837][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.00014143279986456037, 'train/grad_norm_pre_clip': 4.875706672668457, 'timing/data': 0.0007314709946513176, 'timing/model': 1.1720911070005968, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 1.88}) +[2026-06-24 06:06:00,147][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.011543665081262589, 'train/grad_norm_pre_clip': 14.839568138122559, 'timing/data': 0.000874085002578795, 'timing/model': 1.1701034780126065, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 1.88}) +[2026-06-24 06:06:04,488][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.04693359509110451, 'train/grad_norm_pre_clip': 16.9613094329834, 'timing/data': 0.0006865979812573642, 'timing/model': 1.172351001005154, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 1.88}) +[2026-06-24 06:06:08,955][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.0483672134578228, 'train/grad_norm_pre_clip': 19.64995574951172, 'timing/data': 0.0007442549976985902, 'timing/model': 1.1732514380128123, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 1.88}) +[2026-06-24 06:06:13,255][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.0003263794060330838, 'train/grad_norm_pre_clip': 10.004817962646484, 'timing/data': 0.0008151860092766583, 'timing/model': 1.172201449982822, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 1.88}) +[2026-06-24 06:06:17,555][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.0073850080370903015, 'train/grad_norm_pre_clip': 43.53512191772461, 'timing/data': 0.0008333990117534995, 'timing/model': 1.1733758310147095, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 1.88}) +[2026-06-24 06:06:21,854][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.05953836441040039, 'train/grad_norm_pre_clip': 32.47390365600586, 'timing/data': 0.0007432429993059486, 'timing/model': 1.1728298350062687, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 1.88}) +[2026-06-24 06:06:26,451][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.0016186705324798822, 'train/grad_norm_pre_clip': 50.91355895996094, 'timing/data': 0.0007912919973023236, 'timing/model': 1.1712454009975772, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 1.88}) +[2026-06-24 06:06:30,786][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.004644285421818495, 'train/grad_norm_pre_clip': 22.698583602905273, 'timing/data': 0.0007280150020960718, 'timing/model': 1.1727249959949404, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 1.88}) +[2026-06-24 06:06:35,094][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.0005596928531304002, 'train/grad_norm_pre_clip': 3.9476232528686523, 'timing/data': 0.0006958949961699545, 'timing/model': 1.1710746900062077, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 1.88}) +[2026-06-24 06:06:39,410][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.024989651516079903, 'train/grad_norm_pre_clip': 7.295424938201904, 'timing/data': 0.0007147500000428408, 'timing/model': 1.1909070330148097, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 1.88}) +[2026-06-24 06:06:43,976][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.00046053968253545463, 'train/grad_norm_pre_clip': 10.827295303344727, 'timing/data': 0.0007636410009581596, 'timing/model': 1.1958750710182358, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 1.88}) +[2026-06-24 06:06:48,316][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.3097989559173584, 'train/grad_norm_pre_clip': 26.22357940673828, 'timing/data': 0.0010208860039710999, 'timing/model': 1.1933868970081676, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 1.88}) +[2026-06-24 06:06:52,647][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.03369734808802605, 'train/grad_norm_pre_clip': 33.55426025390625, 'timing/data': 0.0007819539750926197, 'timing/model': 1.1959875829925295, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 1.88}) +[2026-06-24 06:06:56,965][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.012518076226115227, 'train/grad_norm_pre_clip': 8.16376781463623, 'timing/data': 0.0007977740024216473, 'timing/model': 1.1776574190007523, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 1.88}) +[2026-06-24 06:07:01,554][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.07821191102266312, 'train/grad_norm_pre_clip': 17.225622177124023, 'timing/data': 0.0008949530019890517, 'timing/model': 1.3957131370261777, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 1.89}) +[2026-06-24 06:07:05,921][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.00014733488205820322, 'train/grad_norm_pre_clip': 37.513954162597656, 'timing/data': 0.0007462379871867597, 'timing/model': 1.194811296008993, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 1.89}) +[2026-06-24 06:07:10,220][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.008806921541690826, 'train/grad_norm_pre_clip': 18.64548683166504, 'timing/data': 0.0008974570082500577, 'timing/model': 1.175351687008515, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 1.89}) +[2026-06-24 06:07:14,560][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.000832980265840888, 'train/grad_norm_pre_clip': 4.45206356048584, 'timing/data': 0.0007255399832502007, 'timing/model': 1.1947611570067238, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 1.89}) +[2026-06-24 06:07:19,109][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.0939159095287323, 'train/grad_norm_pre_clip': 27.32729148864746, 'timing/data': 0.0007500460196752101, 'timing/model': 1.4235776620043907, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 1.89}) +[2026-06-24 06:07:23,439][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.05400144308805466, 'train/grad_norm_pre_clip': 43.66419982910156, 'timing/data': 0.0007445359951816499, 'timing/model': 1.1754397539771162, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 1.89}) +[2026-06-24 06:07:27,794][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.0463637113571167, 'train/grad_norm_pre_clip': 11.54664421081543, 'timing/data': 0.0008464340062346309, 'timing/model': 1.1833854019932915, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 1.89}) +[2026-06-24 06:07:32,120][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.04593179374933243, 'train/grad_norm_pre_clip': 18.22715187072754, 'timing/data': 0.0007634599751327187, 'timing/model': 1.1751269619853701, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 1.89}) +[2026-06-24 06:07:36,688][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.007865812629461288, 'train/grad_norm_pre_clip': 12.409146308898926, 'timing/data': 0.0007625680009368807, 'timing/model': 1.1924805250018835, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 1.89}) +[2026-06-24 06:07:41,083][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.05409079045057297, 'train/grad_norm_pre_clip': 26.818359375, 'timing/data': 0.0009104920027311891, 'timing/model': 1.199604344990803, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 1.89}) +[2026-06-24 06:07:45,449][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.012717340141534805, 'train/grad_norm_pre_clip': 10.00999641418457, 'timing/data': 0.0010971570154652, 'timing/model': 1.176938150980277, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 1.89}) +[2026-06-24 06:07:49,810][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.024447353556752205, 'train/grad_norm_pre_clip': 7.568822860717773, 'timing/data': 0.0008781119831837714, 'timing/model': 1.1908725609828252, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 1.89}) +[2026-06-24 06:07:54,420][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.005019398406147957, 'train/grad_norm_pre_clip': 19.978450775146484, 'timing/data': 0.0008684129861649126, 'timing/model': 1.175048233999405, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 1.89}) +[2026-06-24 06:07:58,783][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.03335976228117943, 'train/grad_norm_pre_clip': 32.56483459472656, 'timing/data': 0.000944856001297012, 'timing/model': 1.1753800890001003, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 1.89}) +[2026-06-24 06:08:03,169][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.042388901114463806, 'train/grad_norm_pre_clip': 27.33376693725586, 'timing/data': 0.000772636994952336, 'timing/model': 1.2306822680111509, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 1.89}) +[2026-06-24 06:08:07,522][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.07381566613912582, 'train/grad_norm_pre_clip': 18.032564163208008, 'timing/data': 0.000992091983789578, 'timing/model': 1.1987387069966644, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 1.89}) +[2026-06-24 06:08:12,110][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.17237448692321777, 'train/grad_norm_pre_clip': 24.036590576171875, 'timing/data': 0.0010516030015423894, 'timing/model': 1.2037739400111604, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 1.89}) +[2026-06-24 06:08:16,468][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.11607769876718521, 'train/grad_norm_pre_clip': 28.300945281982422, 'timing/data': 0.0010235210065729916, 'timing/model': 1.2062411879887804, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 1.89}) +[2026-06-24 06:08:20,817][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.0017313775606453419, 'train/grad_norm_pre_clip': 20.40225601196289, 'timing/data': 0.0007451460114680231, 'timing/model': 1.178037456003949, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 1.89}) +[2026-06-24 06:08:25,146][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.05209914594888687, 'train/grad_norm_pre_clip': 27.210412979125977, 'timing/data': 0.0009762030094861984, 'timing/model': 1.198489714995958, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 1.89}) +[2026-06-24 06:08:29,760][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.04955170676112175, 'train/grad_norm_pre_clip': 19.964021682739258, 'timing/data': 0.0007414990104734898, 'timing/model': 1.1772871539869811, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 1.89}) +[2026-06-24 06:08:34,108][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.0031838673166930676, 'train/grad_norm_pre_clip': 19.71262550354004, 'timing/data': 0.0007783070032019168, 'timing/model': 1.1940703750005923, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 1.89}) +[2026-06-24 06:08:38,431][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.0010043693473562598, 'train/grad_norm_pre_clip': 30.944839477539062, 'timing/data': 0.0009022559970617294, 'timing/model': 1.179651804995956, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 1.89}) +[2026-06-24 06:08:42,765][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.10151507705450058, 'train/grad_norm_pre_clip': 43.57032775878906, 'timing/data': 0.001055759988958016, 'timing/model': 1.199952773982659, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 1.89}) +[2026-06-24 06:08:47,386][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.04817815124988556, 'train/grad_norm_pre_clip': 20.55344581604004, 'timing/data': 0.0008116300159599632, 'timing/model': 1.1773919440165628, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 1.89}) +[2026-06-24 06:08:51,747][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.07948972284793854, 'train/grad_norm_pre_clip': 26.485204696655273, 'timing/data': 0.0009792789933271706, 'timing/model': 1.1795772129844408, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 1.89}) +[2026-06-24 06:08:56,108][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.011188028380274773, 'train/grad_norm_pre_clip': 28.959470748901367, 'timing/data': 0.0007666849996894598, 'timing/model': 1.1841959649755154, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 1.9}) +[2026-06-24 06:09:00,453][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.0209055133163929, 'train/grad_norm_pre_clip': 10.581134796142578, 'timing/data': 0.0009778360254131258, 'timing/model': 1.1755677400215063, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 1.9}) +[2026-06-24 06:09:05,049][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.021953685209155083, 'train/grad_norm_pre_clip': 20.66411590576172, 'timing/data': 0.0012633829901460558, 'timing/model': 1.1770311440050136, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 1.9}) +[2026-06-24 06:09:09,400][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4900 +[2026-06-24 06:10:07,377][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4900 +[2026-06-24 06:10:07,378][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4900 +[2026-06-24 06:10:44,269][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4900 +[2026-06-24 06:10:44,271][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.04858360439538956, 'train/grad_norm_pre_clip': 31.950653076171875, 'eval/action_loss/seconds': 57.974410243012244, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03337794914841652, 'eval/latency_2/loss': 0.03337794914841652, 'eval/flappy/loss': 0.03337794914841652, 'eval/flappy/latency_2/loss': 0.03337794914841652, 'eval/flap_precision': 0.9770580296896086, 'eval/flap_recall': 0.9823609226594301, 'eval/flap_f1': 0.9797023004059541, 'eval/flap_rate_model': 0.20583333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9797023004059541, 'eval/latency_2/flap_precision': 0.9770580296896086, 'eval/latency_2/flap_recall': 0.9823609226594301, 'eval/latency_2/flap_f1': 0.9797023004059541, 'eval/latency_2/flap_rate_model': 0.20583333333333334, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9797023004059541, 'eval/action_classification/data_wait_seconds': 0.5158244967460632, 'eval/action_classification/predict_seconds': 36.23029708862305, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0007740599976386875, 'timing/model': 1.1941539639956318, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 1.9}) +[2026-06-24 06:10:48,560][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.0500117763876915, 'train/grad_norm_pre_clip': 22.347692489624023, 'timing/data': 0.0007563270046375692, 'timing/model': 1.1574555700062774, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 1.9}) +[2026-06-24 06:10:53,091][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.04268312826752663, 'train/grad_norm_pre_clip': 12.091509819030762, 'timing/data': 0.001081317022908479, 'timing/model': 1.1562916230177507, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 1.9}) +[2026-06-24 06:10:57,371][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.0004040118074044585, 'train/grad_norm_pre_clip': 20.327157974243164, 'timing/data': 0.0007846590015105903, 'timing/model': 1.1643243510043249, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 1.9}) +[2026-06-24 06:11:01,651][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.001155743026174605, 'train/grad_norm_pre_clip': 29.282451629638672, 'timing/data': 0.0007227150199469179, 'timing/model': 1.1636308710149024, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 1.9}) +[2026-06-24 06:11:05,962][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.016386527568101883, 'train/grad_norm_pre_clip': 30.845182418823242, 'timing/data': 0.0007788279908709228, 'timing/model': 1.1697383639984764, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 1.9}) +[2026-06-24 06:11:10,523][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.12337730079889297, 'train/grad_norm_pre_clip': 31.122787475585938, 'timing/data': 0.0007510180003009737, 'timing/model': 1.186523678014055, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 1.9}) +[2026-06-24 06:11:14,812][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.04006262123584747, 'train/grad_norm_pre_clip': 19.521671295166016, 'timing/data': 0.0007348170038312674, 'timing/model': 1.1798762010002974, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 1.9}) +[2026-06-24 06:11:19,107][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.030393250286579132, 'train/grad_norm_pre_clip': 18.101280212402344, 'timing/data': 0.0010192829940933734, 'timing/model': 1.1682322149863467, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 1.9}) +[2026-06-24 06:11:23,436][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.006917781662195921, 'train/grad_norm_pre_clip': 13.61711311340332, 'timing/data': 0.0007644420256838202, 'timing/model': 1.1968239690177143, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 1.9}) +[2026-06-24 06:11:27,973][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.15800204873085022, 'train/grad_norm_pre_clip': 32.840450286865234, 'timing/data': 0.0007410489779431373, 'timing/model': 1.4154690320137888, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 1.9}) +[2026-06-24 06:11:32,317][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.002836491446942091, 'train/grad_norm_pre_clip': 3.3197062015533447, 'timing/data': 0.0008701370097696781, 'timing/model': 1.1892867779824883, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 1.9}) +[2026-06-24 06:11:36,657][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.036181289702653885, 'train/grad_norm_pre_clip': 15.915701866149902, 'timing/data': 0.0009100009920075536, 'timing/model': 1.1969860710087232, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 1.9}) +[2026-06-24 06:11:41,000][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.013533160090446472, 'train/grad_norm_pre_clip': 13.147199630737305, 'timing/data': 0.001071108999894932, 'timing/model': 1.1965253620001022, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 1.9}) +[2026-06-24 06:11:45,566][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.035970307886600494, 'train/grad_norm_pre_clip': 14.035944938659668, 'timing/data': 0.0007493550074286759, 'timing/model': 1.428043613996124, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 1.9}) +[2026-06-24 06:11:49,904][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.06914235651493073, 'train/grad_norm_pre_clip': 31.671342849731445, 'timing/data': 0.0007138290093280375, 'timing/model': 1.1785450720053632, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 1.9}) +[2026-06-24 06:11:54,276][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.164352685213089, 'train/grad_norm_pre_clip': 25.899803161621094, 'timing/data': 0.0008757580071687698, 'timing/model': 1.1739362610096578, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 1.9}) +[2026-06-24 06:11:58,574][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.007860478945076466, 'train/grad_norm_pre_clip': 7.151544094085693, 'timing/data': 0.0008688140078447759, 'timing/model': 1.174563450011192, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 1.9}) +[2026-06-24 06:12:03,113][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.05560709536075592, 'train/grad_norm_pre_clip': 22.8051815032959, 'timing/data': 0.0007507269911002368, 'timing/model': 1.1743128880043514, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 1.9}) +[2026-06-24 06:12:07,405][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.0026930873282253742, 'train/grad_norm_pre_clip': 25.50653648376465, 'timing/data': 0.0006962369952816516, 'timing/model': 1.1688833399966825, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 1.9}) +[2026-06-24 06:12:11,685][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.00014182539598550647, 'train/grad_norm_pre_clip': 6.5996623039245605, 'timing/data': 0.0007241669809445739, 'timing/model': 1.1700913929962553, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 1.9}) +[2026-06-24 06:12:15,994][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.0011058654636144638, 'train/grad_norm_pre_clip': 13.508755683898926, 'timing/data': 0.001077800989151001, 'timing/model': 1.17079276998993, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 1.9}) +[2026-06-24 06:12:20,466][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.015905819833278656, 'train/grad_norm_pre_clip': 14.244929313659668, 'timing/data': 0.0008672619878780097, 'timing/model': 1.1716798190027475, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 1.9}) +[2026-06-24 06:12:24,772][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.004620111081749201, 'train/grad_norm_pre_clip': 14.136945724487305, 'timing/data': 0.0006955449935048819, 'timing/model': 1.1884800919797271, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 1.91}) +[2026-06-24 06:12:29,060][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.10441417247056961, 'train/grad_norm_pre_clip': 24.201318740844727, 'timing/data': 0.0007268029730767012, 'timing/model': 1.1694599309994373, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 1.91}) +[2026-06-24 06:12:33,379][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.10932272672653198, 'train/grad_norm_pre_clip': 24.04328155517578, 'timing/data': 0.0007049629930406809, 'timing/model': 1.1740719209774397, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 1.91}) +[2026-06-24 06:12:37,953][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.015674587339162827, 'train/grad_norm_pre_clip': 11.038488388061523, 'timing/data': 0.000747050013160333, 'timing/model': 1.173670490999939, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 1.91}) +[2026-06-24 06:12:42,273][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 6.847678741905838e-05, 'train/grad_norm_pre_clip': 15.904312133789062, 'timing/data': 0.0007299679855350405, 'timing/model': 1.1755759319930803, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 1.91}) +[2026-06-24 06:12:46,593][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.02249126508831978, 'train/grad_norm_pre_clip': 18.538206100463867, 'timing/data': 0.0007948980201035738, 'timing/model': 1.1741195220092777, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 1.91}) +[2026-06-24 06:12:50,896][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.05048888176679611, 'train/grad_norm_pre_clip': 29.456981658935547, 'timing/data': 0.0008685739885549992, 'timing/model': 1.174844823020976, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 1.91}) +[2026-06-24 06:12:55,393][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.0029289827216416597, 'train/grad_norm_pre_clip': 23.60518455505371, 'timing/data': 0.0009618460026104003, 'timing/model': 1.176131769985659, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 1.91}) +[2026-06-24 06:12:59,713][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.0013011499540880322, 'train/grad_norm_pre_clip': 21.188831329345703, 'timing/data': 0.000768269004765898, 'timing/model': 1.1741400829923805, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 1.91}) +[2026-06-24 06:13:04,016][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.001006977166980505, 'train/grad_norm_pre_clip': 6.905961990356445, 'timing/data': 0.0007503859815187752, 'timing/model': 1.1748147800099105, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 1.91}) +[2026-06-24 06:13:08,319][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.043674733489751816, 'train/grad_norm_pre_clip': 35.52195358276367, 'timing/data': 0.0007518780184909701, 'timing/model': 1.1762990830175113, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 1.91}) +[2026-06-24 06:13:12,841][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.08501440286636353, 'train/grad_norm_pre_clip': 23.92796516418457, 'timing/data': 0.0007648129831068218, 'timing/model': 1.3900345119764097, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 1.91}) +[2026-06-24 06:13:17,137][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.11187779903411865, 'train/grad_norm_pre_clip': 33.663143157958984, 'timing/data': 0.0007363700133282691, 'timing/model': 1.172725945012644, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 1.91}) +[2026-06-24 06:13:21,447][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.001093515777029097, 'train/grad_norm_pre_clip': 19.431365966796875, 'timing/data': 0.00108554502367042, 'timing/model': 1.1870094389887527, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 1.91}) +[2026-06-24 06:13:25,749][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.00960718747228384, 'train/grad_norm_pre_clip': 30.393903732299805, 'timing/data': 0.0009565270156599581, 'timing/model': 1.1741881969792303, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 1.91}) +[2026-06-24 06:13:30,098][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.0012641025241464376, 'train/grad_norm_pre_clip': 10.672369003295898, 'timing/data': 0.0007722369919065386, 'timing/model': 1.2198542100086343, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 1.91}) +[2026-06-24 06:13:34,667][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.0029272353276610374, 'train/grad_norm_pre_clip': 9.158171653747559, 'timing/data': 0.000890263996552676, 'timing/model': 1.175209546025144, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 1.91}) +[2026-06-24 06:13:38,990][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.01359118428081274, 'train/grad_norm_pre_clip': 34.53956604003906, 'timing/data': 0.0009377829846926033, 'timing/model': 1.1818836550228298, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 1.91}) +[2026-06-24 06:13:43,301][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.0008027168223634362, 'train/grad_norm_pre_clip': 36.55594253540039, 'timing/data': 0.000782095012255013, 'timing/model': 1.1779361740045715, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 1.91}) +[2026-06-24 06:13:47,633][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.01867973990738392, 'train/grad_norm_pre_clip': 19.897863388061523, 'timing/data': 0.0008673420234117657, 'timing/model': 1.1774639530049171, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 1.91}) +[2026-06-24 06:13:52,158][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.00276371231302619, 'train/grad_norm_pre_clip': 9.965863227844238, 'timing/data': 0.0007586920110043138, 'timing/model': 1.1739094279764686, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 1.91}) +[2026-06-24 06:13:56,521][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.003637004178017378, 'train/grad_norm_pre_clip': 22.71837043762207, 'timing/data': 0.000731761014321819, 'timing/model': 1.2025113709969446, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 1.91}) +[2026-06-24 06:14:00,838][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.043013401329517365, 'train/grad_norm_pre_clip': 23.14684295654297, 'timing/data': 0.0008611810044385493, 'timing/model': 1.171186460996978, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 1.91}) +[2026-06-24 06:14:05,136][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.004688640125095844, 'train/grad_norm_pre_clip': 5.715506553649902, 'timing/data': 0.0007329639920499176, 'timing/model': 1.1793692659994122, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 1.91}) +[2026-06-24 06:14:09,675][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.00039626494981348515, 'train/grad_norm_pre_clip': 18.726747512817383, 'timing/data': 0.0007473709993064404, 'timing/model': 1.174524948000908, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 1.91}) +[2026-06-24 06:14:14,004][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.0025048484094440937, 'train/grad_norm_pre_clip': 20.432292938232422, 'timing/data': 0.0007321719895116985, 'timing/model': 1.2048395239980891, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 1.91}) +[2026-06-24 06:14:18,339][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.06537161022424698, 'train/grad_norm_pre_clip': 18.11490249633789, 'timing/data': 0.0009398659749422222, 'timing/model': 1.196227077016374, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 1.92}) +[2026-06-24 06:14:22,714][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.0688621923327446, 'train/grad_norm_pre_clip': 19.271291732788086, 'timing/data': 0.0007378330046776682, 'timing/model': 1.1954580479941797, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 1.92}) +[2026-06-24 06:14:27,239][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.030527183786034584, 'train/grad_norm_pre_clip': 20.182504653930664, 'timing/data': 0.0007459979970008135, 'timing/model': 1.1763297049910761, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 1.92}) +[2026-06-24 06:14:31,569][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.042570047080516815, 'train/grad_norm_pre_clip': 4.5767822265625, 'timing/data': 0.0007559460063930601, 'timing/model': 1.2014988180017099, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 1.92}) +[2026-06-24 06:14:35,881][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.0029299743473529816, 'train/grad_norm_pre_clip': 30.97782325744629, 'timing/data': 0.0007260410056915134, 'timing/model': 1.1757615779933985, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 1.92}) +[2026-06-24 06:14:40,217][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.008560173213481903, 'train/grad_norm_pre_clip': 4.791842937469482, 'timing/data': 0.0009788080060388893, 'timing/model': 1.1768256940122228, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 1.92}) +[2026-06-24 06:14:44,687][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.017728062346577644, 'train/grad_norm_pre_clip': 27.907405853271484, 'timing/data': 0.0007685100135859102, 'timing/model': 1.1929564610181842, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 1.92}) +[2026-06-24 06:14:49,049][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.0035495487973093987, 'train/grad_norm_pre_clip': 2.6728410720825195, 'timing/data': 0.000886157009517774, 'timing/model': 1.2040936749835964, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 1.92}) +[2026-06-24 06:14:53,355][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.00010443462088005617, 'train/grad_norm_pre_clip': 17.490327835083008, 'timing/data': 0.000715471018338576, 'timing/model': 1.1758876839885488, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 1.92}) +[2026-06-24 06:14:57,661][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.13947810232639313, 'train/grad_norm_pre_clip': 23.063474655151367, 'timing/data': 0.0007354890112765133, 'timing/model': 1.1765924769861158, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 1.92}) +[2026-06-24 06:15:02,226][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.04447508230805397, 'train/grad_norm_pre_clip': 31.252517700195312, 'timing/data': 0.0009391949861310422, 'timing/model': 1.4289800169935916, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 1.92}) +[2026-06-24 06:15:06,558][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.019619975239038467, 'train/grad_norm_pre_clip': 37.17922592163086, 'timing/data': 0.0007741299923509359, 'timing/model': 1.1760164549923502, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 1.92}) +[2026-06-24 06:15:10,892][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 8.053072087932378e-05, 'train/grad_norm_pre_clip': 10.794464111328125, 'timing/data': 0.0007606949948240072, 'timing/model': 1.1760484539845493, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 1.92}) +[2026-06-24 06:15:15,195][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.0003409348428249359, 'train/grad_norm_pre_clip': 30.19152069091797, 'timing/data': 0.0007340260199271142, 'timing/model': 1.174753001978388, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 1.92}) +[2026-06-24 06:15:19,738][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.022717585787177086, 'train/grad_norm_pre_clip': 18.14191246032715, 'timing/data': 0.0007584510021843016, 'timing/model': 1.192560276016593, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 1.92}) +[2026-06-24 06:15:24,114][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.029980376362800598, 'train/grad_norm_pre_clip': 17.974472045898438, 'timing/data': 0.0009306589781772345, 'timing/model': 1.203049881994957, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 1.92}) +[2026-06-24 06:15:28,409][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.006067655980587006, 'train/grad_norm_pre_clip': 28.960838317871094, 'timing/data': 0.0007356290007010102, 'timing/model': 1.1724729929992463, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 1.92}) +[2026-06-24 06:15:32,701][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.022588755935430527, 'train/grad_norm_pre_clip': 16.287946701049805, 'timing/data': 0.0008040250104386359, 'timing/model': 1.1730099659762345, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 1.92}) +[2026-06-24 06:15:37,241][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.0130592230707407, 'train/grad_norm_pre_clip': 6.193830966949463, 'timing/data': 0.0007676469977013767, 'timing/model': 1.1720857189793605, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 1.92}) +[2026-06-24 06:15:41,556][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.0031336199026554823, 'train/grad_norm_pre_clip': 8.618205070495605, 'timing/data': 0.0008281299960799515, 'timing/model': 1.1717254650138784, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 1.92}) +[2026-06-24 06:15:45,878][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.01982051692903042, 'train/grad_norm_pre_clip': 32.11376190185547, 'timing/data': 0.0007256410026457161, 'timing/model': 1.1707351659715641, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 1.92}) +[2026-06-24 06:15:50,171][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.007475213147699833, 'train/grad_norm_pre_clip': 14.494460105895996, 'timing/data': 0.0008176200208254158, 'timing/model': 1.171960379986558, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 1.92}) +[2026-06-24 06:15:54,660][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.018863284960389137, 'train/grad_norm_pre_clip': 8.256208419799805, 'timing/data': 0.0008168889908120036, 'timing/model': 1.1711550820036791, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 1.92}) +[2026-06-24 06:15:58,952][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.03258487954735756, 'train/grad_norm_pre_clip': 11.408492088317871, 'timing/data': 0.0007981739763636142, 'timing/model': 1.1725820490100887, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 1.92}) +[2026-06-24 06:16:03,254][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.0009668246493674815, 'train/grad_norm_pre_clip': 24.176025390625, 'timing/data': 0.0007098009809851646, 'timing/model': 1.1748457709909417, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 1.92}) +[2026-06-24 06:16:07,567][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.018454331904649734, 'train/grad_norm_pre_clip': 5.787186145782471, 'timing/data': 0.0008199339790735394, 'timing/model': 1.1724578010034747, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 1.93}) +[2026-06-24 06:16:12,112][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.013132993131875992, 'train/grad_norm_pre_clip': 34.72787857055664, 'timing/data': 0.0007910209824331105, 'timing/model': 1.197317772981478, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 1.93}) +[2026-06-24 06:16:16,451][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.011654580011963844, 'train/grad_norm_pre_clip': 26.055227279663086, 'timing/data': 0.0007828659727238119, 'timing/model': 1.1816440410038922, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 1.93}) +[2026-06-24 06:16:20,801][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.0021515213884413242, 'train/grad_norm_pre_clip': 12.057758331298828, 'timing/data': 0.0008648680231999606, 'timing/model': 1.1999465499829967, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 1.93}) +[2026-06-24 06:16:25,169][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.07883293181657791, 'train/grad_norm_pre_clip': 36.96187973022461, 'timing/data': 0.0010069400013890117, 'timing/model': 1.1811283499991987, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 1.93}) +[2026-06-24 06:16:29,704][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.017764775082468987, 'train/grad_norm_pre_clip': 6.825270175933838, 'timing/data': 0.0007679989794269204, 'timing/model': 1.1799243170244154, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 1.93}) +[2026-06-24 06:16:34,078][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.004559751134365797, 'train/grad_norm_pre_clip': 5.172698497772217, 'timing/data': 0.0013271210191305727, 'timing/model': 1.2041873859998304, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 1.93}) +[2026-06-24 06:16:38,450][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.1162281483411789, 'train/grad_norm_pre_clip': 28.113189697265625, 'timing/data': 0.001072942017344758, 'timing/model': 1.1938912170007825, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 1.93}) +[2026-06-24 06:16:42,769][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.0013261366402730346, 'train/grad_norm_pre_clip': 4.0489420890808105, 'timing/data': 0.0007581599929835647, 'timing/model': 1.1725125369848683, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 1.93}) +[2026-06-24 06:16:47,296][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.0020920054521411657, 'train/grad_norm_pre_clip': 19.27211570739746, 'timing/data': 0.0007389649981632829, 'timing/model': 1.1731619490019511, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 1.93}) +[2026-06-24 06:16:51,602][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.006930322386324406, 'train/grad_norm_pre_clip': 25.819433212280273, 'timing/data': 0.0007288769993465394, 'timing/model': 1.1745506739825942, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 1.93}) +[2026-06-24 06:16:55,932][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.00889955461025238, 'train/grad_norm_pre_clip': 12.82775592803955, 'timing/data': 0.0007612270128447562, 'timing/model': 1.178114446986001, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 1.93}) +[2026-06-24 06:17:00,272][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.005175824277102947, 'train/grad_norm_pre_clip': 17.953733444213867, 'timing/data': 0.0010955139878205955, 'timing/model': 1.2052575660054572, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 1.93}) +[2026-06-24 06:17:04,767][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.016793332993984222, 'train/grad_norm_pre_clip': 17.635465621948242, 'timing/data': 0.000719620002200827, 'timing/model': 1.1766501940146554, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 1.93}) +[2026-06-24 06:17:09,132][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.02327745594084263, 'train/grad_norm_pre_clip': 36.0714111328125, 'timing/data': 0.0007121359813027084, 'timing/model': 1.2125209639780223, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 1.93}) +[2026-06-24 06:17:13,464][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.012736817821860313, 'train/grad_norm_pre_clip': 13.556472778320312, 'timing/data': 0.00073210199479945, 'timing/model': 1.174267073016381, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 1.93}) +[2026-06-24 06:17:17,810][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.09088990092277527, 'train/grad_norm_pre_clip': 7.522594928741455, 'timing/data': 0.0007128459983505309, 'timing/model': 1.1726403490174562, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 1.93}) +[2026-06-24 06:17:22,336][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.018438823521137238, 'train/grad_norm_pre_clip': 21.8249568939209, 'timing/data': 0.0006942229811102152, 'timing/model': 1.1920443220005836, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 1.93}) +[2026-06-24 06:17:26,673][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.07844652235507965, 'train/grad_norm_pre_clip': 26.581947326660156, 'timing/data': 0.0007411289843730628, 'timing/model': 1.1933872730005533, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 1.93}) +[2026-06-24 06:17:30,991][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.016763584688305855, 'train/grad_norm_pre_clip': 16.302528381347656, 'timing/data': 0.001198563986690715, 'timing/model': 1.1738863520149607, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 1.93}) +[2026-06-24 06:17:35,368][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.025779569521546364, 'train/grad_norm_pre_clip': 13.262130737304688, 'timing/data': 0.0010245330049656332, 'timing/model': 1.2287645220058039, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 1.93}) +[2026-06-24 06:17:39,908][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.014402205124497414, 'train/grad_norm_pre_clip': 9.893143653869629, 'timing/data': 0.0010892319842241704, 'timing/model': 1.1836625220021233, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 1.93}) +[2026-06-24 06:17:44,258][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.0008174720569513738, 'train/grad_norm_pre_clip': 17.143901824951172, 'timing/data': 0.0007276239921338856, 'timing/model': 1.1918506170040928, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 1.93}) +[2026-06-24 06:17:48,739][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.014164522290229797, 'train/grad_norm_pre_clip': 35.76704406738281, 'timing/data': 0.0010954430035781115, 'timing/model': 1.1993008090066724, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 1.93}) +[2026-06-24 06:17:53,057][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.007586408406496048, 'train/grad_norm_pre_clip': 12.311202049255371, 'timing/data': 0.0008183719764929265, 'timing/model': 1.1927996009762865, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 1.93}) +[2026-06-24 06:17:57,625][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.014120745472609997, 'train/grad_norm_pre_clip': 37.48908615112305, 'timing/data': 0.0007327939965762198, 'timing/model': 1.194164874003036, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 1.93}) +[2026-06-24 06:18:01,955][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 5000 +[2026-06-24 06:18:59,929][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 5000 +[2026-06-24 06:18:59,930][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 5000 +[2026-06-24 06:19:36,811][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 5000 +[2026-06-24 06:47:51,811][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 5000 +[2026-06-24 06:47:51,813][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.03742458298802376, 'train/grad_norm_pre_clip': 16.293434143066406, 'eval/action_loss/seconds': 57.97137111000484, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03338811546564102, 'eval/latency_2/loss': 0.03338811546564102, 'eval/flappy/loss': 0.03338811546564102, 'eval/flappy/latency_2/loss': 0.03338811546564102, 'eval/flap_precision': 0.9810040705563093, 'eval/flap_recall': 0.9810040705563093, 'eval/flap_f1': 0.9810040705563093, 'eval/flap_rate_model': 0.20472222222222222, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9810040705563094, 'eval/latency_2/flap_precision': 0.9810040705563093, 'eval/latency_2/flap_recall': 0.9810040705563093, 'eval/latency_2/flap_f1': 0.9810040705563093, 'eval/latency_2/flap_rate_model': 0.20472222222222222, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9810040705563094, 'eval/action_classification/data_wait_seconds': 0.4992876648902893, 'eval/action_classification/predict_seconds': 36.236846923828125, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0007372620166279376, 'timing/model': 1.194568220002111, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 1.94}) +[2026-06-24 06:47:51,813][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_5000_state +[2026-06-24 06:47:51,819][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 06:48:00,546][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_5000_state/model.safetensors +[2026-06-24 06:48:10,841][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_5000_state/optimizer.bin +[2026-06-24 06:48:10,842][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_5000_state/sampler.bin +[2026-06-24 06:48:10,842][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-24 06:48:10,843][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-24 06:48:11,046][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-24 06:48:11,047][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 06:48:11,072][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 06:48:11,073][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning/checkpoints +[2026-06-24 06:48:12,777][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..a0e21985e6b7ed5b88fddcc5f74f6ec45063b603 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning/summary.jsonl @@ -0,0 +1,10 @@ +{"steps": 500} +{"steps": 1000} +{"steps": 1500} +{"steps": 2000} +{"steps": 2500} +{"steps": 3000} +{"steps": 3500} +{"steps": 4000} +{"steps": 4500} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..cc5c658bdd179872f539ebef2a71d31dfb641ffe --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff343b81e73667493a90c8c0696872ab7f3ed1bf55fae59cf8b57f9785b040eb +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..014715f73627aa060626b1777386b58830faaf34 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/config.full.yaml @@ -0,0 +1,219 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_llm_bottom_ratio: 0.0 + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_full_tuning_corrected +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..6b8a09793b3da600d5811d123b302c8dcd49ab2a --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/config.yaml @@ -0,0 +1,99 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + eval_num_workers: 8 + include_state: true + latency_curriculum: + enabled: false + obs_image_size: null + per_device_batch_size: 32 + persistent_workers: true + pin_memory: true + prefetch_factor: 4 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected +rl_games: + env_eval: + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 1 + image_size: 224 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: '' + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: flappy +run_id: flappy_fix_latency_2_200ep_full_tuning_corrected +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_action_classification_interval: null + eval_interval: 100 + eval_num_batches: 100 + freeze_llm_bottom_ratio: 0.0 + freeze_modules: '' + gradient_accumulation_steps: 4 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + reload_modules: null + save_interval: 500 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..aa03e6cbf8e375d660528e17fafddb4804b596b7 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.6028500199317932, + 0.3971499800682068, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4890792667865753, + 0.4890792667865753, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 330734, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..0caa254376d5805f972b0b0411e0b5188e3634e3 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7959861159324646, + 0.2040138840675354, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4030573070049286, + 0.4030573070049286, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c828fd36466d1abf1d9bf91958ab43d6632b4a48 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra/.hydra/config.yaml @@ -0,0 +1,217 @@ +framework: + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_llm_bottom_ratio: 0.0 + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_full_tuning_corrected +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..893b25cd32d349fca1b3192c7222c33ecb27f139 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra/.hydra/hydra.yaml @@ -0,0 +1,266 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=flappy_fix_latency_2_200ep_full_tuning_corrected + - run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=saberrr-zju + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=500 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.eval_action_classification=true + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=4 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.freeze_llm_bottom_ratio=0.0 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9,0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=true + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.num_workers=8 + - datasets.vla_data.eval_num_workers=8 + - datasets.vla_data.prefetch_factor=4 + - datasets.vla_data.persistent_workers=true + - datasets.vla_data.pin_memory=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.latency_curriculum.enabled=false + - datasets.vla_data.latency_curriculum.strategy=exclusive + - datasets.vla_data.per_device_batch_size=32 + - datasets.vla_data.num_workers=8 + - datasets.vla_data.eval_num_workers=8 + - datasets.vla_data.prefetch_factor=4 + - datasets.vla_data.persistent_workers=true + - datasets.vla_data.pin_memory=true + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=250 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[2] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.eval_num_workers=8,datasets.vla_data.eval_num_workers=8,datasets.vla_data.include_state=true,datasets.vla_data.latency_curriculum.enabled=false,datasets.vla_data.latency_curriculum.strategy=exclusive,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.num_workers=8,datasets.vla_data.num_workers=8,datasets.vla_data.per_device_batch_size=32,datasets.vla_data.persistent_workers=true,datasets.vla_data.persistent_workers=true,datasets.vla_data.pin_memory=true,datasets.vla_data.pin_memory=true,datasets.vla_data.prefetch_factor=4,datasets.vla_data.prefetch_factor=4,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=250,rl_games.env_eval.mid_train.latencies=[2],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=flappy_fix_latency_2_200ep_full_tuning_corrected,run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_action_classification=true,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.freeze_llm_bottom_ratio=0.0,trainer.gradient_accumulation_steps=4,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=true,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=500,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=saberrr-zju,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/latency-sensitive-bench/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/latency-sensitive-bench/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..1c3965d90df6f2361be093d0bbf3415751d8ff5c --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra/.hydra/overrides.yaml @@ -0,0 +1,106 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=flappy_fix_latency_2_200ep_full_tuning_corrected +- run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=saberrr-zju +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=500 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.eval_action_classification=true +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=4 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.freeze_llm_bottom_ratio=0.0 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9,0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=true +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.num_workers=8 +- datasets.vla_data.eval_num_workers=8 +- datasets.vla_data.prefetch_factor=4 +- datasets.vla_data.persistent_workers=true +- datasets.vla_data.pin_memory=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.latency_curriculum.enabled=false +- datasets.vla_data.latency_curriculum.strategy=exclusive +- datasets.vla_data.per_device_batch_size=32 +- datasets.vla_data.num_workers=8 +- datasets.vla_data.eval_num_workers=8 +- datasets.vla_data.prefetch_factor=4 +- datasets.vla_data.persistent_workers=true +- datasets.vla_data.pin_memory=true +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=250 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[2] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..e4454034ddabf3b62c912cf05d5a932f21591e14 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/hydra/train_starvla_hydra.log @@ -0,0 +1,5375 @@ +[2026-06-23 20:07:59,735][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-23 20:08:00,212][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured. +[2026-06-23 20:08:00,212][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub +[2026-06-23 20:08:00,740][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases +[2026-06-23 20:08:00,755][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-23 20:08:07,007][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-23 20:08:07,033][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-23 20:08:07,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-23 20:08:07,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-23 20:08:07,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-23 20:08:07,745][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-23 20:08:07,746][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-23 20:08:07,747][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-23 20:08:07,747][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-23 20:08:07,747][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-23 20:08:07,747][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-23 20:08:07,747][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-23 20:08:07,747][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-23 20:08:07,747][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-23 20:08:07,747][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-23 20:08:07,747][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-23 20:08:07,775][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-23 20:08:07,796][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/config.full.yaml +[2026-06-23 20:08:07,802][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/config.yaml +[2026-06-23 20:08:07,802][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-23 20:08:11,194][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-23 20:08:11,194][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-23 20:08:11,194][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-23 20:08:11,195][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 4 +[2026-06-23 20:08:11,195][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-23 20:08:44,150][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6955766081809998, 'train/grad_norm_pre_clip': 9.008363723754883, 'timing/data': 0.0018448230111971498, 'timing/model': 1.4110667859204113, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-23 20:08:48,628][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6940546631813049, 'train/grad_norm_pre_clip': 10.597947120666504, 'timing/data': 0.0023384690284729004, 'timing/model': 1.3005128420190886, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-23 20:08:52,378][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6910155415534973, 'train/grad_norm_pre_clip': 19.52017593383789, 'timing/data': 0.00032999494578689337, 'timing/model': 0.9148532070685178, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-23 20:08:55,648][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.6910101175308228, 'train/grad_norm_pre_clip': 10.094944953918457, 'timing/data': 0.00034633802715688944, 'timing/model': 0.9146089019486681, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 20:08:58,932][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6867585182189941, 'train/grad_norm_pre_clip': 2.331069231033325, 'timing/data': 0.000426487997174263, 'timing/model': 0.9248203459428623, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-23 20:09:02,238][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6435124278068542, 'train/grad_norm_pre_clip': 5.967395782470703, 'timing/data': 0.0003561810590326786, 'timing/model': 0.9220742909237742, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-23 20:09:05,517][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.6633611917495728, 'train/grad_norm_pre_clip': 2.403151512145996, 'timing/data': 0.0004190569743514061, 'timing/model': 0.9185438529821113, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 20:09:08,816][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.7092061638832092, 'train/grad_norm_pre_clip': 4.848846912384033, 'timing/data': 0.00048340100329369307, 'timing/model': 0.9190183780156076, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 20:09:12,110][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.6263697147369385, 'train/grad_norm_pre_clip': 9.571824073791504, 'timing/data': 0.0003807030152529478, 'timing/model': 0.9180673650698736, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-23 20:09:15,414][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.7175413370132446, 'train/grad_norm_pre_clip': 7.7875261306762695, 'timing/data': 0.0003328899620100856, 'timing/model': 0.9243548070080578, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-23 20:09:18,736][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.661257803440094, 'train/grad_norm_pre_clip': 7.572865009307861, 'timing/data': 0.0003449249779805541, 'timing/model': 0.9268981089117005, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-23 20:09:22,063][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.7106396555900574, 'train/grad_norm_pre_clip': 2.2584056854248047, 'timing/data': 0.0005580210126936436, 'timing/model': 0.9272025629179552, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-23 20:09:25,377][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.6610202193260193, 'train/grad_norm_pre_clip': 0.9520538449287415, 'timing/data': 0.00041341304313391447, 'timing/model': 0.9270680480403826, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-23 20:09:28,706][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.6252720355987549, 'train/grad_norm_pre_clip': 1.3168033361434937, 'timing/data': 0.0003883560420945287, 'timing/model': 0.9313233910361305, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-23 20:09:32,033][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.645853579044342, 'train/grad_norm_pre_clip': 1.9443702697753906, 'timing/data': 0.0005175760015845299, 'timing/model': 0.9291187999770045, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-23 20:09:35,367][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.6505611538887024, 'train/grad_norm_pre_clip': 8.498483657836914, 'timing/data': 0.0003842839505523443, 'timing/model': 0.9348705060547218, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 20:09:38,701][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.7084778547286987, 'train/grad_norm_pre_clip': 6.529586315155029, 'timing/data': 0.0003700779052451253, 'timing/model': 0.929216620977968, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 20:09:42,037][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.6972705125808716, 'train/grad_norm_pre_clip': 6.547842025756836, 'timing/data': 0.00034416699782013893, 'timing/model': 0.9343867390416563, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-23 20:09:45,388][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.6818976998329163, 'train/grad_norm_pre_clip': 7.42368221282959, 'timing/data': 0.00035916094202548265, 'timing/model': 0.9299372150562704, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-23 20:09:48,725][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.6767580509185791, 'train/grad_norm_pre_clip': 6.4386725425720215, 'timing/data': 0.00035021803341805935, 'timing/model': 0.9324022580403835, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-23 20:09:52,072][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.6722210645675659, 'train/grad_norm_pre_clip': 2.4892091751098633, 'timing/data': 0.00037259492091834545, 'timing/model': 0.9316031839698553, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-23 20:09:55,434][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.6905072331428528, 'train/grad_norm_pre_clip': 4.4482340812683105, 'timing/data': 0.0005448349984362721, 'timing/model': 0.936449496075511, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 20:09:58,778][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.6312748789787292, 'train/grad_norm_pre_clip': 6.9079437255859375, 'timing/data': 0.00034326594322919846, 'timing/model': 0.9362892269855365, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 20:10:02,125][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.616506814956665, 'train/grad_norm_pre_clip': 3.486485719680786, 'timing/data': 0.00035144505091011524, 'timing/model': 0.9305500199552625, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-23 20:10:05,489][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.646053671836853, 'train/grad_norm_pre_clip': 3.044862985610962, 'timing/data': 0.0005928280297666788, 'timing/model': 0.9439190810080618, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-23 20:10:08,839][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.7300118803977966, 'train/grad_norm_pre_clip': 10.480050086975098, 'timing/data': 0.0005440759705379605, 'timing/model': 0.9283541790209711, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-23 20:10:12,200][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.6955938935279846, 'train/grad_norm_pre_clip': 5.146539688110352, 'timing/data': 0.0003396719694137573, 'timing/model': 0.9453223529271781, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-23 20:10:15,616][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.6328834295272827, 'train/grad_norm_pre_clip': 3.7448127269744873, 'timing/data': 0.0003481869352981448, 'timing/model': 0.9425152419134974, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 20:10:18,973][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.626349151134491, 'train/grad_norm_pre_clip': 3.3999388217926025, 'timing/data': 0.00034218200016766787, 'timing/model': 0.9364657138939947, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-23 20:10:22,327][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.6251775026321411, 'train/grad_norm_pre_clip': 12.122554779052734, 'timing/data': 0.0003475010162219405, 'timing/model': 0.9416208729380742, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-23 20:10:25,694][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.6428605914115906, 'train/grad_norm_pre_clip': 10.773923873901367, 'timing/data': 0.0004863329231739044, 'timing/model': 0.937128658988513, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-23 20:10:29,060][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.5967385768890381, 'train/grad_norm_pre_clip': 14.738883972167969, 'timing/data': 0.0005655810236930847, 'timing/model': 0.9427372410427779, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-23 20:10:32,443][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.6824991106987, 'train/grad_norm_pre_clip': 7.51481294631958, 'timing/data': 0.0004432810237631202, 'timing/model': 0.9438236859859899, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-23 20:10:35,819][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.5725750923156738, 'train/grad_norm_pre_clip': 7.358329772949219, 'timing/data': 0.00035839900374412537, 'timing/model': 0.9423141649458557, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-23 20:10:39,178][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.6033614873886108, 'train/grad_norm_pre_clip': 10.639487266540527, 'timing/data': 0.0003525179345160723, 'timing/model': 0.938332595047541, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01}) +[2026-06-23 20:10:42,546][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.553669273853302, 'train/grad_norm_pre_clip': 16.055919647216797, 'timing/data': 0.0006051890086382627, 'timing/model': 0.9377856700448319, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01}) +[2026-06-23 20:10:45,925][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.579288899898529, 'train/grad_norm_pre_clip': 9.740020751953125, 'timing/data': 0.0004034299636259675, 'timing/model': 0.9497506649931893, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.01}) +[2026-06-23 20:10:49,297][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.5100615620613098, 'train/grad_norm_pre_clip': 23.090478897094727, 'timing/data': 0.0003463439643383026, 'timing/model': 0.9406928590033203, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.01}) +[2026-06-23 20:10:52,663][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.40742164850234985, 'train/grad_norm_pre_clip': 17.77889633178711, 'timing/data': 0.00039958092384040356, 'timing/model': 0.9401195369428024, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 20:10:56,027][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.4881500005722046, 'train/grad_norm_pre_clip': 24.935428619384766, 'timing/data': 0.0003542759222909808, 'timing/model': 0.9402533440152183, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-23 20:10:59,402][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.40425434708595276, 'train/grad_norm_pre_clip': 29.365100860595703, 'timing/data': 0.00039348809514194727, 'timing/model': 0.943205043906346, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-23 20:11:02,791][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.7606134414672852, 'train/grad_norm_pre_clip': 41.130123138427734, 'timing/data': 0.00035528908483684063, 'timing/model': 0.9443660830147564, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-23 20:11:06,173][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.55709308385849, 'train/grad_norm_pre_clip': 74.8887939453125, 'timing/data': 0.0003571660490706563, 'timing/model': 0.9473563000792637, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-23 20:11:09,544][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.4197885990142822, 'train/grad_norm_pre_clip': 49.15243911743164, 'timing/data': 0.00035591796040534973, 'timing/model': 0.9429746679961681, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 20:11:12,928][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.4689052700996399, 'train/grad_norm_pre_clip': 42.909244537353516, 'timing/data': 0.00041714299004524946, 'timing/model': 0.9458871649112552, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-23 20:11:16,308][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.36388397216796875, 'train/grad_norm_pre_clip': 35.477272033691406, 'timing/data': 0.0003842030419036746, 'timing/model': 0.9414409829769284, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-23 20:11:19,693][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.3723832070827484, 'train/grad_norm_pre_clip': 45.42378616333008, 'timing/data': 0.00035342597402632236, 'timing/model': 0.9468858010368422, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-23 20:11:23,063][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.4054621756076813, 'train/grad_norm_pre_clip': 59.61148452758789, 'timing/data': 0.0003520429600030184, 'timing/model': 0.9388514270540327, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-23 20:11:26,448][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.6396889686584473, 'train/grad_norm_pre_clip': 51.33597946166992, 'timing/data': 0.00034737796522676945, 'timing/model': 0.9410411659628153, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-23 20:11:29,835][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.43375101685523987, 'train/grad_norm_pre_clip': 27.361133575439453, 'timing/data': 0.00035182794090360403, 'timing/model': 0.9512028270401061, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-23 20:11:33,218][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.479132741689682, 'train/grad_norm_pre_clip': 26.50603675842285, 'timing/data': 0.0003527509979903698, 'timing/model': 0.9400700979167596, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 20:11:36,616][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.44209566712379456, 'train/grad_norm_pre_clip': 29.108762741088867, 'timing/data': 0.0005763940280303359, 'timing/model': 0.9431347089121118, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-23 20:11:40,002][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.662663459777832, 'train/grad_norm_pre_clip': 51.685462951660156, 'timing/data': 0.0004954829346388578, 'timing/model': 0.9406846050405875, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-23 20:11:43,379][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.5830445885658264, 'train/grad_norm_pre_clip': 64.96551513671875, 'timing/data': 0.00035347696393728256, 'timing/model': 0.9407774930587038, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-23 20:11:46,764][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.489638090133667, 'train/grad_norm_pre_clip': 15.888803482055664, 'timing/data': 0.0003467060159891844, 'timing/model': 0.9453139450633898, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-23 20:11:50,149][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.48821282386779785, 'train/grad_norm_pre_clip': 36.208858489990234, 'timing/data': 0.0005535179516300559, 'timing/model': 0.9447341039776802, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 20:11:53,528][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.473039448261261, 'train/grad_norm_pre_clip': 40.37211990356445, 'timing/data': 0.000349742011167109, 'timing/model': 0.9448430460179225, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-23 20:11:56,924][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.31441524624824524, 'train/grad_norm_pre_clip': 19.408172607421875, 'timing/data': 0.00035334506537765265, 'timing/model': 0.9471752280369401, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-23 20:12:00,308][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.4460476040840149, 'train/grad_norm_pre_clip': 32.82054901123047, 'timing/data': 0.0005391649901866913, 'timing/model': 0.9417015770450234, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.02}) +[2026-06-23 20:12:03,693][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.43438103795051575, 'train/grad_norm_pre_clip': 21.824119567871094, 'timing/data': 0.0004368009977042675, 'timing/model': 0.9420566000044346, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-23 20:12:07,065][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.3817169964313507, 'train/grad_norm_pre_clip': 27.057790756225586, 'timing/data': 0.00035237998235970736, 'timing/model': 0.9446287400787696, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.02}) +[2026-06-23 20:12:10,460][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.3615635931491852, 'train/grad_norm_pre_clip': 30.77548599243164, 'timing/data': 0.0005469858879223466, 'timing/model': 0.9496624229941517, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.02}) +[2026-06-23 20:12:13,843][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.5856062769889832, 'train/grad_norm_pre_clip': 36.64398193359375, 'timing/data': 0.0003495980054140091, 'timing/model': 0.9409094699658453, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.02}) +[2026-06-23 20:12:17,222][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.5469145774841309, 'train/grad_norm_pre_clip': 40.43613052368164, 'timing/data': 0.00036752899177372456, 'timing/model': 0.945961297955364, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.02}) +[2026-06-23 20:12:20,591][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.4590156078338623, 'train/grad_norm_pre_clip': 30.59754180908203, 'timing/data': 0.00036165700294077396, 'timing/model': 0.9366439940640703, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-23 20:12:23,973][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.44283613562583923, 'train/grad_norm_pre_clip': 15.535613059997559, 'timing/data': 0.0005156659753993154, 'timing/model': 0.9451187240192667, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-23 20:12:27,364][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.42610445618629456, 'train/grad_norm_pre_clip': 50.296661376953125, 'timing/data': 0.0005418619839474559, 'timing/model': 0.9446293300716206, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-23 20:12:30,752][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.4162967801094055, 'train/grad_norm_pre_clip': 35.13462448120117, 'timing/data': 0.00035463401582092047, 'timing/model': 0.9431858450407162, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 20:12:34,136][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.42447707056999207, 'train/grad_norm_pre_clip': 23.776657104492188, 'timing/data': 0.000347528955899179, 'timing/model': 0.9490840149810538, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-23 20:12:37,519][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.366558313369751, 'train/grad_norm_pre_clip': 72.50698852539062, 'timing/data': 0.0003560279728844762, 'timing/model': 0.9471014849841595, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-23 20:12:40,927][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.48139601945877075, 'train/grad_norm_pre_clip': 22.175167083740234, 'timing/data': 0.0003584859659895301, 'timing/model': 0.9547035479918122, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-23 20:12:44,322][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.46373653411865234, 'train/grad_norm_pre_clip': 28.52238655090332, 'timing/data': 0.00034670508466660976, 'timing/model': 0.9442122289910913, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-23 20:12:47,700][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.4004960060119629, 'train/grad_norm_pre_clip': 22.478389739990234, 'timing/data': 0.00035210105124861, 'timing/model': 0.938598082982935, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-23 20:12:51,083][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.42899245023727417, 'train/grad_norm_pre_clip': 20.524864196777344, 'timing/data': 0.0003558200551196933, 'timing/model': 0.9427383840084076, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-23 20:12:54,475][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.28956344723701477, 'train/grad_norm_pre_clip': 44.07809066772461, 'timing/data': 0.00036027294117957354, 'timing/model': 0.9493905799463391, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-23 20:12:57,861][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.2823810875415802, 'train/grad_norm_pre_clip': 39.93820571899414, 'timing/data': 0.0003499810118228197, 'timing/model': 0.9485847490141168, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-23 20:13:01,266][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.2929403781890869, 'train/grad_norm_pre_clip': 39.77562713623047, 'timing/data': 0.0003527189837768674, 'timing/model': 0.9491776350187138, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-23 20:13:04,662][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.4035662114620209, 'train/grad_norm_pre_clip': 63.45466232299805, 'timing/data': 0.00035459594801068306, 'timing/model': 0.9468367110239342, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 20:13:08,070][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.5014765858650208, 'train/grad_norm_pre_clip': 72.50887298583984, 'timing/data': 0.00034746702294796705, 'timing/model': 0.949408293934539, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.03}) +[2026-06-23 20:13:11,460][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.3923778235912323, 'train/grad_norm_pre_clip': 40.50788116455078, 'timing/data': 0.000487504992634058, 'timing/model': 0.943970849039033, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-23 20:13:14,843][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.5144274830818176, 'train/grad_norm_pre_clip': 46.239952087402344, 'timing/data': 0.000511980033479631, 'timing/model': 0.9437918029725552, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.03}) +[2026-06-23 20:13:18,244][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.30668017268180847, 'train/grad_norm_pre_clip': 26.059139251708984, 'timing/data': 0.00041308498475700617, 'timing/model': 0.9431469450937584, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.03}) +[2026-06-23 20:13:21,624][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.42795005440711975, 'train/grad_norm_pre_clip': 29.536727905273438, 'timing/data': 0.0003440320724621415, 'timing/model': 0.9393651220016181, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.03}) +[2026-06-23 20:13:25,016][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.2764431834220886, 'train/grad_norm_pre_clip': 36.768428802490234, 'timing/data': 0.00034882198087871075, 'timing/model': 0.9464594670571387, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.03}) +[2026-06-23 20:13:28,397][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.4730098843574524, 'train/grad_norm_pre_clip': 93.91354370117188, 'timing/data': 0.0003616380272433162, 'timing/model': 0.9435732470592484, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.03}) +[2026-06-23 20:13:31,870][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.38929033279418945, 'train/grad_norm_pre_clip': 175.88442993164062, 'timing/data': 0.0004962489474564791, 'timing/model': 0.9444008549908176, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.03}) +[2026-06-23 20:13:35,256][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.5386733412742615, 'train/grad_norm_pre_clip': 103.80098724365234, 'timing/data': 0.0004987450083717704, 'timing/model': 0.9482432530494407, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.03}) +[2026-06-23 20:13:38,652][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.4039824903011322, 'train/grad_norm_pre_clip': 111.81175231933594, 'timing/data': 0.0005641540046781301, 'timing/model': 0.9430133090354502, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 20:13:42,050][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.7549414038658142, 'train/grad_norm_pre_clip': 192.878173828125, 'timing/data': 0.000444715959019959, 'timing/model': 0.947662458056584, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.03}) +[2026-06-23 20:13:45,444][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.5846444368362427, 'train/grad_norm_pre_clip': 126.41598510742188, 'timing/data': 0.00034973397850990295, 'timing/model': 0.9491207979153842, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-23 20:13:48,836][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.49650025367736816, 'train/grad_norm_pre_clip': 29.343591690063477, 'timing/data': 0.0006199229974299669, 'timing/model': 0.9458157729823142, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-23 20:13:52,224][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.46727892756462097, 'train/grad_norm_pre_clip': 42.981353759765625, 'timing/data': 0.00040073704440146685, 'timing/model': 0.9458677039947361, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-23 20:13:55,624][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.2429542988538742, 'train/grad_norm_pre_clip': 63.025333404541016, 'timing/data': 0.00035464903339743614, 'timing/model': 0.950690416037105, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-23 20:13:59,006][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.25015881657600403, 'train/grad_norm_pre_clip': 99.94174194335938, 'timing/data': 0.00034835096448659897, 'timing/model': 0.9387450600042939, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-23 20:14:02,404][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.3315739035606384, 'train/grad_norm_pre_clip': 51.514766693115234, 'timing/data': 0.00034597504418343306, 'timing/model': 0.9418022619793192, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-23 20:14:05,796][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.2689044773578644, 'train/grad_norm_pre_clip': 63.914710998535156, 'timing/data': 0.0003559610340744257, 'timing/model': 0.9457831080071628, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-23 20:14:09,192][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.3307334780693054, 'train/grad_norm_pre_clip': 28.49803924560547, 'timing/data': 0.0007278920384123921, 'timing/model': 0.9491683570668101, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-23 20:14:12,590][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.39553919434547424, 'train/grad_norm_pre_clip': 41.79979705810547, 'timing/data': 0.0004942420637235045, 'timing/model': 0.9397467550588772, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-23 20:14:15,968][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.4445195496082306, 'train/grad_norm_pre_clip': 23.279945373535156, 'timing/data': 0.0004384079948067665, 'timing/model': 0.9409837269922718, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-23 20:14:19,348][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 100 +[2026-06-23 20:15:50,163][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 100 +[2026-06-23 20:15:50,165][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 100 +[2026-06-23 20:15:50,566][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmpyt7pkhtv/test.c -o /tmp/tmpyt7pkhtv/test.o +[2026-06-23 20:15:50,585][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmpyt7pkhtv/test.o -laio -o /tmp/tmpyt7pkhtv/a.out +[2026-06-23 20:15:51,711][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmp8l2s447s/test.c -o /tmp/tmp8l2s447s/test.o +[2026-06-23 20:15:51,732][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmp8l2s447s/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp8l2s447s/a.out +[2026-06-23 20:15:51,804][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmp3jjd2us7/test.c -o /tmp/tmp3jjd2us7/test.o +[2026-06-23 20:15:51,825][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmp3jjd2us7/test.o -laio -o /tmp/tmp3jjd2us7/a.out +[2026-06-23 20:16:55,547][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 100 +[2026-06-23 20:16:55,548][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.37257489562034607, 'train/grad_norm_pre_clip': 46.717262268066406, 'eval/action_loss/seconds': 90.81139854202047, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.19921018183231354, 'eval/latency_2/loss': 0.19921018183231354, 'eval/flappy/loss': 0.19921018183231354, 'eval/flappy/latency_2/loss': 0.19921018183231354, 'eval/flap_precision': 0.7, 'eval/flap_recall': 0.9402985074626866, 'eval/flap_f1': 0.8025477707006369, 'eval/flap_rate_model': 0.275, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.802547770700637, 'eval/latency_2/flap_precision': 0.7, 'eval/latency_2/flap_recall': 0.9402985074626866, 'eval/latency_2/flap_f1': 0.8025477707006369, 'eval/latency_2/flap_rate_model': 0.275, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.802547770700637, 'eval/action_classification/data_wait_seconds': 30.821514129638672, 'eval/action_classification/predict_seconds': 31.283281326293945, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00035555195063352585, 'timing/model': 0.9427600289927796, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.04}) +[2026-06-23 20:16:59,588][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.32850104570388794, 'train/grad_norm_pre_clip': 19.51473045349121, 'timing/data': 0.00034626899287104607, 'timing/model': 1.1063019150169566, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.04}) +[2026-06-23 20:17:03,632][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.2810368239879608, 'train/grad_norm_pre_clip': 94.04691314697266, 'timing/data': 0.0004123040707781911, 'timing/model': 1.1049627580214292, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.04}) +[2026-06-23 20:17:07,827][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.36402398347854614, 'train/grad_norm_pre_clip': 31.041545867919922, 'timing/data': 0.0003474080003798008, 'timing/model': 1.113674870110117, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.04}) +[2026-06-23 20:17:11,906][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.6612356305122375, 'train/grad_norm_pre_clip': 100.33650207519531, 'timing/data': 0.0003520520403981209, 'timing/model': 1.120767634944059, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.04}) +[2026-06-23 20:17:15,983][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.33541712164878845, 'train/grad_norm_pre_clip': 46.989280700683594, 'timing/data': 0.00036870199255645275, 'timing/model': 1.1180592159507796, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.04}) +[2026-06-23 20:17:20,051][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.3093269169330597, 'train/grad_norm_pre_clip': 49.182552337646484, 'timing/data': 0.0003463970497250557, 'timing/model': 1.1154277580790222, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.04}) +[2026-06-23 20:17:24,197][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.26966243982315063, 'train/grad_norm_pre_clip': 61.70561599731445, 'timing/data': 0.0003613330191001296, 'timing/model': 1.1193252389784902, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.04}) +[2026-06-23 20:17:28,268][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.2236482799053192, 'train/grad_norm_pre_clip': 50.20478439331055, 'timing/data': 0.0005474709905683994, 'timing/model': 1.1151078549446538, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.04}) +[2026-06-23 20:17:32,347][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.561630368232727, 'train/grad_norm_pre_clip': 19.145009994506836, 'timing/data': 0.0004305529873818159, 'timing/model': 1.113431604928337, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.04}) +[2026-06-23 20:17:36,440][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.40475621819496155, 'train/grad_norm_pre_clip': 57.627113342285156, 'timing/data': 0.0004010669654235244, 'timing/model': 1.1248055479954928, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.04}) +[2026-06-23 20:17:40,641][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.22702424228191376, 'train/grad_norm_pre_clip': 77.22782897949219, 'timing/data': 0.000508354976773262, 'timing/model': 1.1202042680233717, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.04}) +[2026-06-23 20:17:44,729][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.21596622467041016, 'train/grad_norm_pre_clip': 28.19300651550293, 'timing/data': 0.00035733007825911045, 'timing/model': 1.127740086056292, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.04}) +[2026-06-23 20:17:48,824][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.2500944435596466, 'train/grad_norm_pre_clip': 80.48506164550781, 'timing/data': 0.00035143899731338024, 'timing/model': 1.1207384360022843, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.04}) +[2026-06-23 20:17:52,920][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.40700504183769226, 'train/grad_norm_pre_clip': 55.48951721191406, 'timing/data': 0.00036068400368094444, 'timing/model': 1.1191547460621223, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.04}) +[2026-06-23 20:17:57,109][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.5141428112983704, 'train/grad_norm_pre_clip': 50.176597595214844, 'timing/data': 0.0005543199367821217, 'timing/model': 1.123132681939751, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.04}) +[2026-06-23 20:18:01,202][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.5757929086685181, 'train/grad_norm_pre_clip': 115.55048370361328, 'timing/data': 0.00035112909972667694, 'timing/model': 1.1202222660649568, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.04}) +[2026-06-23 20:18:05,298][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.35068660974502563, 'train/grad_norm_pre_clip': 39.621925354003906, 'timing/data': 0.0005907469894737005, 'timing/model': 1.1201310120522976, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.05}) +[2026-06-23 20:18:09,402][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.3796595633029938, 'train/grad_norm_pre_clip': 137.57476806640625, 'timing/data': 0.0005893179913982749, 'timing/model': 1.1242764319758862, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.05}) +[2026-06-23 20:18:13,695][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.5358713269233704, 'train/grad_norm_pre_clip': 175.18450927734375, 'timing/data': 0.0005308899562805891, 'timing/model': 1.1277421629056334, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.05}) +[2026-06-23 20:18:17,788][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.24991625547409058, 'train/grad_norm_pre_clip': 24.627195358276367, 'timing/data': 0.00043116905726492405, 'timing/model': 1.1200692530255765, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.05}) +[2026-06-23 20:18:21,903][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.47352471947669983, 'train/grad_norm_pre_clip': 130.7603759765625, 'timing/data': 0.00040980393532663584, 'timing/model': 1.128354882937856, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.05}) +[2026-06-23 20:18:26,001][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.19773423671722412, 'train/grad_norm_pre_clip': 86.70475006103516, 'timing/data': 0.00035368907265365124, 'timing/model': 1.123766585951671, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.05}) +[2026-06-23 20:18:30,185][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.37954795360565186, 'train/grad_norm_pre_clip': 50.34275817871094, 'timing/data': 0.0004164481069892645, 'timing/model': 1.116649407078512, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.05}) +[2026-06-23 20:18:34,275][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.6270363330841064, 'train/grad_norm_pre_clip': 183.70123291015625, 'timing/data': 0.0003526529762893915, 'timing/model': 1.118642724934034, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.05}) +[2026-06-23 20:18:38,382][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.8118773102760315, 'train/grad_norm_pre_clip': 194.867431640625, 'timing/data': 0.00034535606391727924, 'timing/model': 1.1297432830324396, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.05}) +[2026-06-23 20:18:42,486][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.2659298777580261, 'train/grad_norm_pre_clip': 98.34194946289062, 'timing/data': 0.0004470720887184143, 'timing/model': 1.1238090810365975, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.05}) +[2026-06-23 20:18:46,660][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.29208800196647644, 'train/grad_norm_pre_clip': 76.36223602294922, 'timing/data': 0.00034118304029107094, 'timing/model': 1.117788647999987, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.05}) +[2026-06-23 20:18:50,763][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.47800004482269287, 'train/grad_norm_pre_clip': 81.00700378417969, 'timing/data': 0.0003497250145301223, 'timing/model': 1.1212072360794991, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.05}) +[2026-06-23 20:18:54,868][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.31173375248908997, 'train/grad_norm_pre_clip': 41.7660026550293, 'timing/data': 0.00035765301436185837, 'timing/model': 1.123749496997334, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.05}) +[2026-06-23 20:18:58,969][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.44282016158103943, 'train/grad_norm_pre_clip': 79.66120910644531, 'timing/data': 0.00036531395744532347, 'timing/model': 1.1277393320342526, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.05}) +[2026-06-23 20:19:03,227][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.3345784544944763, 'train/grad_norm_pre_clip': 74.70051574707031, 'timing/data': 0.0003582719946280122, 'timing/model': 1.1324991750298068, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.05}) +[2026-06-23 20:19:07,330][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.2061009556055069, 'train/grad_norm_pre_clip': 18.825336456298828, 'timing/data': 0.0003699880326166749, 'timing/model': 1.1285989310126752, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.05}) +[2026-06-23 20:19:11,450][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.18029049038887024, 'train/grad_norm_pre_clip': 44.234405517578125, 'timing/data': 0.0003663250245153904, 'timing/model': 1.1215470490278676, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.05}) +[2026-06-23 20:19:15,650][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.22553865611553192, 'train/grad_norm_pre_clip': 44.16264343261719, 'timing/data': 0.00035503890831023455, 'timing/model': 1.2141161870677024, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.05}) +[2026-06-23 20:19:19,757][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.1836361140012741, 'train/grad_norm_pre_clip': 35.4384765625, 'timing/data': 0.00036018399987369776, 'timing/model': 1.1308865030296147, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.05}) +[2026-06-23 20:19:23,871][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.2588985562324524, 'train/grad_norm_pre_clip': 89.70099639892578, 'timing/data': 0.00034897797740995884, 'timing/model': 1.1274032490327954, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.05}) +[2026-06-23 20:19:27,981][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.3239535093307495, 'train/grad_norm_pre_clip': 69.6938705444336, 'timing/data': 0.00035088800359517336, 'timing/model': 1.124681015033275, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.05}) +[2026-06-23 20:19:32,256][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.4604720175266266, 'train/grad_norm_pre_clip': 65.5543441772461, 'timing/data': 0.00040940498001873493, 'timing/model': 1.2940884149866179, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.05}) +[2026-06-23 20:19:36,375][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.573506236076355, 'train/grad_norm_pre_clip': 91.92971801757812, 'timing/data': 0.000465265940874815, 'timing/model': 1.1278753470396623, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.05}) +[2026-06-23 20:19:40,485][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.369773805141449, 'train/grad_norm_pre_clip': 33.82538604736328, 'timing/data': 0.0005897009978070855, 'timing/model': 1.127744491910562, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.05}) +[2026-06-23 20:19:44,592][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.41352981328964233, 'train/grad_norm_pre_clip': 79.28775024414062, 'timing/data': 0.0003511341055855155, 'timing/model': 1.1244330300251022, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.05}) +[2026-06-23 20:19:48,853][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.2846516966819763, 'train/grad_norm_pre_clip': 70.52642822265625, 'timing/data': 0.0007127750432118773, 'timing/model': 1.1223920349730179, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.05}) +[2026-06-23 20:19:52,965][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.25009170174598694, 'train/grad_norm_pre_clip': 27.26383399963379, 'timing/data': 0.0005428469739854336, 'timing/model': 1.1257175869541243, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.06}) +[2026-06-23 20:19:57,092][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.29327452182769775, 'train/grad_norm_pre_clip': 38.50587844848633, 'timing/data': 0.0003518160665407777, 'timing/model': 1.1301121689612046, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.06}) +[2026-06-23 20:20:01,212][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.1713048368692398, 'train/grad_norm_pre_clip': 40.918418884277344, 'timing/data': 0.0004764420446008444, 'timing/model': 1.1252151429653168, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.06}) +[2026-06-23 20:20:05,447][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.2988749146461487, 'train/grad_norm_pre_clip': 62.0107307434082, 'timing/data': 0.00035451503936201334, 'timing/model': 1.126214773976244, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.06}) +[2026-06-23 20:20:09,572][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.25634464621543884, 'train/grad_norm_pre_clip': 46.8270378112793, 'timing/data': 0.000380938989110291, 'timing/model': 1.1271480630384758, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.06}) +[2026-06-23 20:20:13,687][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.5602251887321472, 'train/grad_norm_pre_clip': 67.46614837646484, 'timing/data': 0.0005813699681311846, 'timing/model': 1.1267385029932484, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.06}) +[2026-06-23 20:20:17,808][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.39297130703926086, 'train/grad_norm_pre_clip': 26.130687713623047, 'timing/data': 0.0003458870342001319, 'timing/model': 1.1266913389554247, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.06}) +[2026-06-23 20:20:22,099][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.26816028356552124, 'train/grad_norm_pre_clip': 34.37169647216797, 'timing/data': 0.0003772570053115487, 'timing/model': 1.1270103709539399, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.06}) +[2026-06-23 20:20:26,215][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.41111910343170166, 'train/grad_norm_pre_clip': 31.855712890625, 'timing/data': 0.00035006203688681126, 'timing/model': 1.1316824889509007, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.06}) +[2026-06-23 20:20:30,341][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.1900714784860611, 'train/grad_norm_pre_clip': 26.1322078704834, 'timing/data': 0.00039303209632635117, 'timing/model': 1.1275550669524819, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.06}) +[2026-06-23 20:20:34,455][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.2228228747844696, 'train/grad_norm_pre_clip': 26.43803596496582, 'timing/data': 0.0004232199862599373, 'timing/model': 1.1256102410843596, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.06}) +[2026-06-23 20:20:38,730][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.1698543280363083, 'train/grad_norm_pre_clip': 57.277610778808594, 'timing/data': 0.00043873500544577837, 'timing/model': 1.1240005310392007, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.06}) +[2026-06-23 20:20:42,839][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.2658528983592987, 'train/grad_norm_pre_clip': 41.28001022338867, 'timing/data': 0.0003527899971231818, 'timing/model': 1.1236535189673305, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.06}) +[2026-06-23 20:20:46,963][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.5928064584732056, 'train/grad_norm_pre_clip': 130.22817993164062, 'timing/data': 0.00035099301021546125, 'timing/model': 1.1263490789569914, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.06}) +[2026-06-23 20:20:51,223][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.28568097949028015, 'train/grad_norm_pre_clip': 117.11321258544922, 'timing/data': 0.000403617974370718, 'timing/model': 1.266185271088034, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.06}) +[2026-06-23 20:20:55,334][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.2697288990020752, 'train/grad_norm_pre_clip': 30.305938720703125, 'timing/data': 0.00035760889295488596, 'timing/model': 1.1277550898957998, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.06}) +[2026-06-23 20:20:59,450][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.7314614653587341, 'train/grad_norm_pre_clip': 233.92727661132812, 'timing/data': 0.00039916893001645803, 'timing/model': 1.128744108020328, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.06}) +[2026-06-23 20:21:03,556][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.2929401695728302, 'train/grad_norm_pre_clip': 145.86544799804688, 'timing/data': 0.0003598439507186413, 'timing/model': 1.1215753419091925, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.06}) +[2026-06-23 20:21:07,776][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.3072323203086853, 'train/grad_norm_pre_clip': 102.50733947753906, 'timing/data': 0.00046702497638761997, 'timing/model': 1.1191214040154591, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.06}) +[2026-06-23 20:21:11,893][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.28338029980659485, 'train/grad_norm_pre_clip': 113.30157470703125, 'timing/data': 0.00045413593761622906, 'timing/model': 1.129396133008413, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.06}) +[2026-06-23 20:21:16,005][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.34276777505874634, 'train/grad_norm_pre_clip': 67.20472717285156, 'timing/data': 0.00035099603701382875, 'timing/model': 1.1262855720706284, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.06}) +[2026-06-23 20:21:20,141][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.11901143193244934, 'train/grad_norm_pre_clip': 100.4243392944336, 'timing/data': 0.0005998719716444612, 'timing/model': 1.1245080300141126, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.06}) +[2026-06-23 20:21:24,350][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.3607392907142639, 'train/grad_norm_pre_clip': 154.9024658203125, 'timing/data': 0.00035033898893743753, 'timing/model': 1.1320021849824116, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.06}) +[2026-06-23 20:21:28,478][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.21133527159690857, 'train/grad_norm_pre_clip': 60.05131530761719, 'timing/data': 0.0004099179059267044, 'timing/model': 1.1352803689660504, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.06}) +[2026-06-23 20:21:32,594][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.5078884959220886, 'train/grad_norm_pre_clip': 157.9891815185547, 'timing/data': 0.0003972639096900821, 'timing/model': 1.1260567380813882, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.06}) +[2026-06-23 20:21:36,709][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.5538874268531799, 'train/grad_norm_pre_clip': 122.19956970214844, 'timing/data': 0.00034559902269393206, 'timing/model': 1.125231481040828, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.07}) +[2026-06-23 20:21:40,889][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.4135374426841736, 'train/grad_norm_pre_clip': 128.7189178466797, 'timing/data': 0.00046016008127480745, 'timing/model': 1.1252272530691698, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.07}) +[2026-06-23 20:21:45,020][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.2030971199274063, 'train/grad_norm_pre_clip': 32.278045654296875, 'timing/data': 0.00035666499752551317, 'timing/model': 1.1332012600032613, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.07}) +[2026-06-23 20:21:49,129][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.20075900852680206, 'train/grad_norm_pre_clip': 33.753299713134766, 'timing/data': 0.0003450519870966673, 'timing/model': 1.1242036449257284, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.07}) +[2026-06-23 20:21:53,251][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.2532629072666168, 'train/grad_norm_pre_clip': 49.54935073852539, 'timing/data': 0.00035749992821365595, 'timing/model': 1.1295563320163637, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.07}) +[2026-06-23 20:21:57,487][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.2857641279697418, 'train/grad_norm_pre_clip': 55.17461395263672, 'timing/data': 0.00035152502823621035, 'timing/model': 1.1236529600573704, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.07}) +[2026-06-23 20:22:01,593][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.28197821974754333, 'train/grad_norm_pre_clip': 93.86528778076172, 'timing/data': 0.0003510289825499058, 'timing/model': 1.1249439551029354, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.07}) +[2026-06-23 20:22:05,697][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.22838512063026428, 'train/grad_norm_pre_clip': 56.56633758544922, 'timing/data': 0.00044773099943995476, 'timing/model': 1.1199332269607112, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.07}) +[2026-06-23 20:22:09,875][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.3094012141227722, 'train/grad_norm_pre_clip': 36.113468170166016, 'timing/data': 0.0003528059460222721, 'timing/model': 1.1939338319934905, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.07}) +[2026-06-23 20:22:13,998][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.33059075474739075, 'train/grad_norm_pre_clip': 38.76346969604492, 'timing/data': 0.00039973296225070953, 'timing/model': 1.1329292689915746, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.07}) +[2026-06-23 20:22:18,129][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.3072786331176758, 'train/grad_norm_pre_clip': 68.32160949707031, 'timing/data': 0.00037489505484700203, 'timing/model': 1.1268210830166936, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.07}) +[2026-06-23 20:22:22,250][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.3648681938648224, 'train/grad_norm_pre_clip': 34.820858001708984, 'timing/data': 0.0003484799526631832, 'timing/model': 1.1272697530221194, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.07}) +[2026-06-23 20:22:26,371][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.3574464023113251, 'train/grad_norm_pre_clip': 76.7548599243164, 'timing/data': 0.0003538189921528101, 'timing/model': 1.1256497639697045, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.07}) +[2026-06-23 20:22:30,656][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.2545410394668579, 'train/grad_norm_pre_clip': 84.93345642089844, 'timing/data': 0.0003903269534930587, 'timing/model': 1.1228752629831433, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.07}) +[2026-06-23 20:22:34,783][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.30197808146476746, 'train/grad_norm_pre_clip': 87.78944396972656, 'timing/data': 0.00034726294688880444, 'timing/model': 1.130251943017356, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.07}) +[2026-06-23 20:22:38,915][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.4068143665790558, 'train/grad_norm_pre_clip': 89.17667388916016, 'timing/data': 0.0006687920540571213, 'timing/model': 1.1345301389228553, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.07}) +[2026-06-23 20:22:43,046][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.30138054490089417, 'train/grad_norm_pre_clip': 24.263629913330078, 'timing/data': 0.00037049304228276014, 'timing/model': 1.1273992939386517, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.07}) +[2026-06-23 20:22:47,338][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.2934608459472656, 'train/grad_norm_pre_clip': 131.35391235351562, 'timing/data': 0.0005825059488415718, 'timing/model': 1.133789154002443, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.07}) +[2026-06-23 20:22:51,458][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.31583040952682495, 'train/grad_norm_pre_clip': 112.30970764160156, 'timing/data': 0.00034649588633328676, 'timing/model': 1.1279503609985113, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.07}) +[2026-06-23 20:22:55,587][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.27405521273612976, 'train/grad_norm_pre_clip': 48.56566619873047, 'timing/data': 0.000546284019947052, 'timing/model': 1.1316102319397032, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.07}) +[2026-06-23 20:22:59,707][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.30795472860336304, 'train/grad_norm_pre_clip': 77.7193374633789, 'timing/data': 0.000406423001550138, 'timing/model': 1.1322909810114652, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.07}) +[2026-06-23 20:23:03,994][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.5416857004165649, 'train/grad_norm_pre_clip': 153.986572265625, 'timing/data': 0.00042304396629333496, 'timing/model': 1.1236884091049433, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.07}) +[2026-06-23 20:23:08,133][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.15298886597156525, 'train/grad_norm_pre_clip': 62.15427017211914, 'timing/data': 0.00035405601374804974, 'timing/model': 1.142335525015369, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.07}) +[2026-06-23 20:23:12,247][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.3134416937828064, 'train/grad_norm_pre_clip': 33.92634963989258, 'timing/data': 0.0005689830286428332, 'timing/model': 1.123540191911161, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.07}) +[2026-06-23 20:23:16,379][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.6379945278167725, 'train/grad_norm_pre_clip': 170.1682891845703, 'timing/data': 0.0003548199310898781, 'timing/model': 1.1268005849560723, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.07}) +[2026-06-23 20:23:20,681][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.8386731147766113, 'train/grad_norm_pre_clip': 160.64495849609375, 'timing/data': 0.00034972105640918016, 'timing/model': 1.131918139057234, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.07}) +[2026-06-23 20:23:24,814][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.6501047015190125, 'train/grad_norm_pre_clip': 162.97320556640625, 'timing/data': 0.00040558900218456984, 'timing/model': 1.129873396945186, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.08}) +[2026-06-23 20:23:28,948][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.5476861000061035, 'train/grad_norm_pre_clip': 103.30157470703125, 'timing/data': 0.0005550719797611237, 'timing/model': 1.1270461200037971, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.08}) +[2026-06-23 20:23:33,077][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.27911317348480225, 'train/grad_norm_pre_clip': 38.204917907714844, 'timing/data': 0.00035446800757199526, 'timing/model': 1.1335976549889892, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.08}) +[2026-06-23 20:23:37,376][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.3301638662815094, 'train/grad_norm_pre_clip': 70.55760955810547, 'timing/data': 0.00035579397808760405, 'timing/model': 1.1304044529097155, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.08}) +[2026-06-23 20:23:41,504][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.3792855739593506, 'train/grad_norm_pre_clip': 85.411376953125, 'timing/data': 0.000546499970369041, 'timing/model': 1.1247284879209474, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.08}) +[2026-06-23 20:23:45,638][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.13586221635341644, 'train/grad_norm_pre_clip': 31.480724334716797, 'timing/data': 0.000404415070079267, 'timing/model': 1.1292149500222877, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.08}) +[2026-06-23 20:23:49,762][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 200 +[2026-06-23 20:24:53,906][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 200 +[2026-06-23 20:24:53,908][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 200 +[2026-06-23 20:25:34,866][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 200 +[2026-06-23 20:25:34,866][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.2800493538379669, 'train/grad_norm_pre_clip': 23.08055305480957, 'eval/action_loss/seconds': 64.14108683704399, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.27237528562545776, 'eval/latency_2/loss': 0.27237528562545776, 'eval/flappy/loss': 0.27237528562545776, 'eval/flappy/latency_2/loss': 0.27237528562545776, 'eval/flap_precision': 0.9847328244274809, 'eval/flap_recall': 0.35006784260515605, 'eval/flap_f1': 0.5165165165165165, 'eval/flap_rate_model': 0.07277777777777777, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.5165165165165165, 'eval/latency_2/flap_precision': 0.9847328244274809, 'eval/latency_2/flap_recall': 0.35006784260515605, 'eval/latency_2/flap_f1': 0.5165165165165165, 'eval/latency_2/flap_rate_model': 0.07277777777777777, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.5165165165165165, 'eval/action_classification/data_wait_seconds': 0.46275243163108826, 'eval/action_classification/predict_seconds': 40.38499069213867, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00035245902836322784, 'timing/model': 1.127173653920181, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.08}) +[2026-06-23 20:25:38,954][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.7062431573867798, 'train/grad_norm_pre_clip': 144.5651397705078, 'timing/data': 0.00035219592973589897, 'timing/model': 1.1218151249922812, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.08}) +[2026-06-23 20:25:43,050][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.6431127786636353, 'train/grad_norm_pre_clip': 174.6202850341797, 'timing/data': 0.0003521360922604799, 'timing/model': 1.122626741998829, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.08}) +[2026-06-23 20:25:47,160][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.988197386264801, 'train/grad_norm_pre_clip': 170.57278442382812, 'timing/data': 0.000352495932020247, 'timing/model': 1.1271019480191171, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.08}) +[2026-06-23 20:25:51,456][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.3182257115840912, 'train/grad_norm_pre_clip': 93.99514770507812, 'timing/data': 0.0005022959085181355, 'timing/model': 1.1276140450499952, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.08}) +[2026-06-23 20:25:55,566][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.19826436042785645, 'train/grad_norm_pre_clip': 59.956336975097656, 'timing/data': 0.0003512660041451454, 'timing/model': 1.1206704439828172, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.08}) +[2026-06-23 20:25:59,671][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.4524358808994293, 'train/grad_norm_pre_clip': 97.26909637451172, 'timing/data': 0.00035737198777496815, 'timing/model': 1.1259799220133573, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.08}) +[2026-06-23 20:26:03,774][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.1875344216823578, 'train/grad_norm_pre_clip': 59.29323959350586, 'timing/data': 0.0003479919396340847, 'timing/model': 1.1253289310261607, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.08}) +[2026-06-23 20:26:08,039][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.5715363621711731, 'train/grad_norm_pre_clip': 91.06034088134766, 'timing/data': 0.0003615119494497776, 'timing/model': 1.122050283011049, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.08}) +[2026-06-23 20:26:12,160][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.22162848711013794, 'train/grad_norm_pre_clip': 19.969600677490234, 'timing/data': 0.0003492289688438177, 'timing/model': 1.125348923029378, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.08}) +[2026-06-23 20:26:16,275][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.1889094114303589, 'train/grad_norm_pre_clip': 20.608200073242188, 'timing/data': 0.0003485720371827483, 'timing/model': 1.1339816380059347, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.08}) +[2026-06-23 20:26:20,380][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.32254451513290405, 'train/grad_norm_pre_clip': 30.920333862304688, 'timing/data': 0.0003682089736685157, 'timing/model': 1.1260755769908428, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.08}) +[2026-06-23 20:26:24,661][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.18906761705875397, 'train/grad_norm_pre_clip': 43.533790588378906, 'timing/data': 0.000373351969756186, 'timing/model': 1.1245640360284597, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.08}) +[2026-06-23 20:26:28,767][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.2296595722436905, 'train/grad_norm_pre_clip': 50.403873443603516, 'timing/data': 0.00034044194035232067, 'timing/model': 1.1244469149969518, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.08}) +[2026-06-23 20:26:32,883][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.2397885024547577, 'train/grad_norm_pre_clip': 32.90492248535156, 'timing/data': 0.0005410050507634878, 'timing/model': 1.128931322018616, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.08}) +[2026-06-23 20:26:36,999][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.5286388993263245, 'train/grad_norm_pre_clip': 67.47161865234375, 'timing/data': 0.0005475940415635705, 'timing/model': 1.126243119011633, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.08}) +[2026-06-23 20:26:41,296][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.26582086086273193, 'train/grad_norm_pre_clip': 32.91556930541992, 'timing/data': 0.0003419111017137766, 'timing/model': 1.1320268149720505, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.08}) +[2026-06-23 20:26:45,411][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.16442829370498657, 'train/grad_norm_pre_clip': 34.68979263305664, 'timing/data': 0.0005535189993679523, 'timing/model': 1.1258465310093015, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.08}) +[2026-06-23 20:26:49,529][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.2944953441619873, 'train/grad_norm_pre_clip': 52.244667053222656, 'timing/data': 0.0005734369624406099, 'timing/model': 1.1255725979572162, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.08}) +[2026-06-23 20:26:53,655][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.26325029134750366, 'train/grad_norm_pre_clip': 56.67008590698242, 'timing/data': 0.0005618460709229112, 'timing/model': 1.1286033369833604, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.08}) +[2026-06-23 20:26:57,932][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.41543999314308167, 'train/grad_norm_pre_clip': 37.03889083862305, 'timing/data': 0.0003995179431512952, 'timing/model': 1.1253285360289738, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.09}) +[2026-06-23 20:27:02,055][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.19006787240505219, 'train/grad_norm_pre_clip': 34.238014221191406, 'timing/data': 0.0003522350452840328, 'timing/model': 1.128678857930936, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.09}) +[2026-06-23 20:27:06,181][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.2773010730743408, 'train/grad_norm_pre_clip': 129.47157287597656, 'timing/data': 0.00036022800486534834, 'timing/model': 1.1275741999270394, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.09}) +[2026-06-23 20:27:10,311][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.20710299909114838, 'train/grad_norm_pre_clip': 109.6666030883789, 'timing/data': 0.000354715040884912, 'timing/model': 1.1266180949751288, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.09}) +[2026-06-23 20:27:14,601][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.33837825059890747, 'train/grad_norm_pre_clip': 84.08086395263672, 'timing/data': 0.00045047199819236994, 'timing/model': 1.1288268599892035, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.09}) +[2026-06-23 20:27:18,728][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.6298401951789856, 'train/grad_norm_pre_clip': 55.89963912963867, 'timing/data': 0.00034930696710944176, 'timing/model': 1.1316559090046212, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.09}) +[2026-06-23 20:27:22,844][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.30152615904808044, 'train/grad_norm_pre_clip': 71.7845230102539, 'timing/data': 0.0004898989573121071, 'timing/model': 1.1229927010135725, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.09}) +[2026-06-23 20:27:26,951][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.2864922285079956, 'train/grad_norm_pre_clip': 55.81119918823242, 'timing/data': 0.0005321219796314836, 'timing/model': 1.1241164520615712, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.09}) +[2026-06-23 20:27:31,234][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.2345747947692871, 'train/grad_norm_pre_clip': 30.617753982543945, 'timing/data': 0.0003871229710057378, 'timing/model': 1.126900215051137, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.09}) +[2026-06-23 20:27:35,363][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.2784197926521301, 'train/grad_norm_pre_clip': 33.179359436035156, 'timing/data': 0.0003516640281304717, 'timing/model': 1.127552828984335, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.09}) +[2026-06-23 20:27:39,493][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.2242790013551712, 'train/grad_norm_pre_clip': 17.2963924407959, 'timing/data': 0.0003522790502756834, 'timing/model': 1.131918595987372, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.09}) +[2026-06-23 20:27:43,615][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.29121777415275574, 'train/grad_norm_pre_clip': 26.639524459838867, 'timing/data': 0.0004117859061807394, 'timing/model': 1.1247072130208835, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.09}) +[2026-06-23 20:27:47,904][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.14285844564437866, 'train/grad_norm_pre_clip': 19.41916847229004, 'timing/data': 0.0003683050163090229, 'timing/model': 1.1309266899479553, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.09}) +[2026-06-23 20:27:52,044][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.31273603439331055, 'train/grad_norm_pre_clip': 37.256595611572266, 'timing/data': 0.00035129603929817677, 'timing/model': 1.1310409500729293, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.09}) +[2026-06-23 20:27:56,152][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.11262214183807373, 'train/grad_norm_pre_clip': 30.121337890625, 'timing/data': 0.0004019920015707612, 'timing/model': 1.1268025829922408, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.09}) +[2026-06-23 20:28:00,269][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.10597402602434158, 'train/grad_norm_pre_clip': 42.11054992675781, 'timing/data': 0.00035702099557965994, 'timing/model': 1.1251225239830092, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.09}) +[2026-06-23 20:28:04,570][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.19043348729610443, 'train/grad_norm_pre_clip': 62.428714752197266, 'timing/data': 0.0003669849829748273, 'timing/model': 1.1291721679735929, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.09}) +[2026-06-23 20:28:08,687][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.09663686901330948, 'train/grad_norm_pre_clip': 37.87587356567383, 'timing/data': 0.00037909706588834524, 'timing/model': 1.125584063003771, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.09}) +[2026-06-23 20:28:12,810][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.11678659915924072, 'train/grad_norm_pre_clip': 112.69361877441406, 'timing/data': 0.0006272589089348912, 'timing/model': 1.1261001189704984, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.09}) +[2026-06-23 20:28:16,929][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.12427365779876709, 'train/grad_norm_pre_clip': 90.7528305053711, 'timing/data': 0.00035462493542581797, 'timing/model': 1.1305462529417127, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.09}) +[2026-06-23 20:28:21,228][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.21501392126083374, 'train/grad_norm_pre_clip': 65.23171997070312, 'timing/data': 0.0003673770697787404, 'timing/model': 1.127431320026517, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.09}) +[2026-06-23 20:28:25,361][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.40993374586105347, 'train/grad_norm_pre_clip': 57.80292510986328, 'timing/data': 0.0003787870518863201, 'timing/model': 1.1406994730932638, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.09}) +[2026-06-23 20:28:29,491][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.21622510254383087, 'train/grad_norm_pre_clip': 62.99529266357422, 'timing/data': 0.00040702999103814363, 'timing/model': 1.1275989060522988, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.09}) +[2026-06-23 20:28:33,626][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.19875337183475494, 'train/grad_norm_pre_clip': 53.5904655456543, 'timing/data': 0.0004103629617020488, 'timing/model': 1.1320373240159824, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.09}) +[2026-06-23 20:28:37,909][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.17859822511672974, 'train/grad_norm_pre_clip': 72.56196594238281, 'timing/data': 0.00036172603722661734, 'timing/model': 1.126801113015972, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.09}) +[2026-06-23 20:28:42,042][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.3671848177909851, 'train/grad_norm_pre_clip': 80.02042388916016, 'timing/data': 0.0005526440218091011, 'timing/model': 1.13939874607604, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.09}) +[2026-06-23 20:28:46,170][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.27910906076431274, 'train/grad_norm_pre_clip': 48.09642028808594, 'timing/data': 0.00039809104055166245, 'timing/model': 1.131502166041173, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.1}) +[2026-06-23 20:28:50,304][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.16458727419376373, 'train/grad_norm_pre_clip': 58.07870864868164, 'timing/data': 0.0003472008975222707, 'timing/model': 1.129033045959659, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.1}) +[2026-06-23 20:28:54,606][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.38169917464256287, 'train/grad_norm_pre_clip': 27.95610237121582, 'timing/data': 0.0005642219912260771, 'timing/model': 1.1269656280055642, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.1}) +[2026-06-23 20:28:58,731][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.4540318548679352, 'train/grad_norm_pre_clip': 48.78678512573242, 'timing/data': 0.0006598400650545955, 'timing/model': 1.133852110011503, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.1}) +[2026-06-23 20:30:50,902][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 250 +[2026-06-23 20:30:50,903][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.2131061851978302, 'train/grad_norm_pre_clip': 29.329059600830078, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 65.11999999999905, 'rl_games_eval/mid_train/mean_length': 543.4, 'rl_games_eval/mid_train/std_reward': 57.274371231815906, 'rl_games_eval/mid_train/std_length': 461.5935874771225, 'rl_games_eval/mid_train/macro_mean_reward': 65.11999999999905, 'rl_games_eval/mid_train/macro_mean_length': 543.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 65.11999999999905, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 543.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 57.274371231815906, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 461.5935874771225, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 65.11999999999905, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0004015270387753844, 'timing/model': 1.129089477006346, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.1}) +[2026-06-23 20:30:54,947][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.21067526936531067, 'train/grad_norm_pre_clip': 25.31087303161621, 'timing/data': 0.0003613629378378391, 'timing/model': 1.1181842039804906, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.1}) +[2026-06-23 20:30:59,001][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.15101729333400726, 'train/grad_norm_pre_clip': 106.92424774169922, 'timing/data': 0.00038590305484831333, 'timing/model': 1.1126250310335308, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.1}) +[2026-06-23 20:31:03,044][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.32504454255104065, 'train/grad_norm_pre_clip': 71.8292236328125, 'timing/data': 0.0005535910604521632, 'timing/model': 1.1089893480530009, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.1}) +[2026-06-23 20:31:07,301][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.25670450925827026, 'train/grad_norm_pre_clip': 64.85868835449219, 'timing/data': 0.0003901129821315408, 'timing/model': 1.3125256260391325, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.1}) +[2026-06-23 20:31:11,365][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.1297292411327362, 'train/grad_norm_pre_clip': 22.415170669555664, 'timing/data': 0.00035451294388622046, 'timing/model': 1.1148846748983487, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.1}) +[2026-06-23 20:31:15,442][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.2653700113296509, 'train/grad_norm_pre_clip': 58.83150100708008, 'timing/data': 0.00042695202864706516, 'timing/model': 1.114022733992897, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.1}) +[2026-06-23 20:31:19,527][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.419002890586853, 'train/grad_norm_pre_clip': 70.03931427001953, 'timing/data': 0.0003643839154392481, 'timing/model': 1.1210984470089898, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.1}) +[2026-06-23 20:31:23,680][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.18121235072612762, 'train/grad_norm_pre_clip': 66.70417022705078, 'timing/data': 0.0003516760189086199, 'timing/model': 1.1878814470255747, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.1}) +[2026-06-23 20:31:27,785][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.2536780834197998, 'train/grad_norm_pre_clip': 74.03173065185547, 'timing/data': 0.00037097407039254904, 'timing/model': 1.126389343990013, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.1}) +[2026-06-23 20:31:31,866][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.4451294243335724, 'train/grad_norm_pre_clip': 114.32258605957031, 'timing/data': 0.0004138839431107044, 'timing/model': 1.1137407349888235, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.1}) +[2026-06-23 20:31:35,961][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.1554436832666397, 'train/grad_norm_pre_clip': 59.966583251953125, 'timing/data': 0.0003530159592628479, 'timing/model': 1.1227542250417173, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.1}) +[2026-06-23 20:31:40,236][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.24125422537326813, 'train/grad_norm_pre_clip': 37.232357025146484, 'timing/data': 0.00035635195672512054, 'timing/model': 1.2967920950613916, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.1}) +[2026-06-23 20:31:44,346][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.14272838830947876, 'train/grad_norm_pre_clip': 58.65894317626953, 'timing/data': 0.0003820399288088083, 'timing/model': 1.1314485589973629, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.1}) +[2026-06-23 20:31:48,458][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.280413955450058, 'train/grad_norm_pre_clip': 87.87183380126953, 'timing/data': 0.0003815729869529605, 'timing/model': 1.1233816869789734, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.1}) +[2026-06-23 20:31:52,569][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.26488229632377625, 'train/grad_norm_pre_clip': 74.24261474609375, 'timing/data': 0.0003517330624163151, 'timing/model': 1.128246548003517, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.1}) +[2026-06-23 20:31:56,788][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.27487948536872864, 'train/grad_norm_pre_clip': 45.578712463378906, 'timing/data': 0.00039914599619805813, 'timing/model': 1.2378939019981772, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.1}) +[2026-06-23 20:32:00,887][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.10019069910049438, 'train/grad_norm_pre_clip': 66.33349609375, 'timing/data': 0.00037132599391043186, 'timing/model': 1.1292163230245933, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.1}) +[2026-06-23 20:32:05,004][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.13846322894096375, 'train/grad_norm_pre_clip': 45.13682556152344, 'timing/data': 0.0005703730275854468, 'timing/model': 1.1277699309866875, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.1}) +[2026-06-23 20:32:09,102][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.28358718752861023, 'train/grad_norm_pre_clip': 78.75151062011719, 'timing/data': 0.00035231595393270254, 'timing/model': 1.1303840740583837, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.1}) +[2026-06-23 20:32:13,350][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.3059503734111786, 'train/grad_norm_pre_clip': 25.45359230041504, 'timing/data': 0.0003428930649533868, 'timing/model': 1.1343338350998238, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.1}) +[2026-06-23 20:32:17,475][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.26304370164871216, 'train/grad_norm_pre_clip': 25.223106384277344, 'timing/data': 0.0003735290374606848, 'timing/model': 1.1393677860032767, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.1}) +[2026-06-23 20:32:21,626][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.08733934164047241, 'train/grad_norm_pre_clip': 37.453739166259766, 'timing/data': 0.00035501294769346714, 'timing/model': 1.1493828790262341, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.11}) +[2026-06-23 20:32:25,742][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.08293379098176956, 'train/grad_norm_pre_clip': 20.55497169494629, 'timing/data': 0.00035155401565134525, 'timing/model': 1.1294432680588216, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.11}) +[2026-06-23 20:32:30,019][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.28310441970825195, 'train/grad_norm_pre_clip': 60.394718170166016, 'timing/data': 0.0004803420742973685, 'timing/model': 1.1248833859572187, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.11}) +[2026-06-23 20:32:34,133][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.23947176337242126, 'train/grad_norm_pre_clip': 67.0181884765625, 'timing/data': 0.00040921790059655905, 'timing/model': 1.1243744299281389, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.11}) +[2026-06-23 20:32:38,263][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.16785003244876862, 'train/grad_norm_pre_clip': 38.577327728271484, 'timing/data': 0.0003818920813500881, 'timing/model': 1.1383043410023674, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.11}) +[2026-06-23 20:32:42,373][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.40396082401275635, 'train/grad_norm_pre_clip': 59.760921478271484, 'timing/data': 0.0004975689807906747, 'timing/model': 1.1249313179869205, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.11}) +[2026-06-23 20:32:46,668][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.1959286630153656, 'train/grad_norm_pre_clip': 53.2907829284668, 'timing/data': 0.0005479790270328522, 'timing/model': 1.1263319699792191, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.11}) +[2026-06-23 20:32:50,782][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.1193016842007637, 'train/grad_norm_pre_clip': 89.21876525878906, 'timing/data': 0.000586574082262814, 'timing/model': 1.1239859210327268, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.11}) +[2026-06-23 20:32:54,900][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.07115823775529861, 'train/grad_norm_pre_clip': 42.866111755371094, 'timing/data': 0.00034584791865199804, 'timing/model': 1.127767784986645, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.11}) +[2026-06-23 20:32:59,028][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.2823631465435028, 'train/grad_norm_pre_clip': 196.05313110351562, 'timing/data': 0.0003480789018794894, 'timing/model': 1.129025807022117, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.11}) +[2026-06-23 20:33:03,223][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.06007971987128258, 'train/grad_norm_pre_clip': 81.50601196289062, 'timing/data': 0.0003645059186965227, 'timing/model': 1.1283132729586214, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.11}) +[2026-06-23 20:33:07,340][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.2663848102092743, 'train/grad_norm_pre_clip': 46.638023376464844, 'timing/data': 0.00038621609564870596, 'timing/model': 1.1265344109851867, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.11}) +[2026-06-23 20:33:11,472][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.23123101890087128, 'train/grad_norm_pre_clip': 35.66318130493164, 'timing/data': 0.0003528969828039408, 'timing/model': 1.1306351100793108, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.11}) +[2026-06-23 20:33:15,590][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.18196992576122284, 'train/grad_norm_pre_clip': 59.80370330810547, 'timing/data': 0.0004066979745402932, 'timing/model': 1.1324471361003816, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.11}) +[2026-06-23 20:33:19,794][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.16952930390834808, 'train/grad_norm_pre_clip': 38.43527603149414, 'timing/data': 0.0003621629439294338, 'timing/model': 1.1315707369940355, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.11}) +[2026-06-23 20:33:23,911][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.12371965497732162, 'train/grad_norm_pre_clip': 27.08990478515625, 'timing/data': 0.00036559696309268475, 'timing/model': 1.1355889990227297, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.11}) +[2026-06-23 20:33:28,040][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.2553326189517975, 'train/grad_norm_pre_clip': 39.393062591552734, 'timing/data': 0.0004031009739264846, 'timing/model': 1.1310507250018418, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.11}) +[2026-06-23 20:33:32,157][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.31810569763183594, 'train/grad_norm_pre_clip': 59.43268585205078, 'timing/data': 0.0004921630024909973, 'timing/model': 1.122874307911843, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.11}) +[2026-06-23 20:33:36,413][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.10167569667100906, 'train/grad_norm_pre_clip': 24.82212257385254, 'timing/data': 0.0003627799451351166, 'timing/model': 1.1287762869615108, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.11}) +[2026-06-23 20:33:40,549][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.04164982587099075, 'train/grad_norm_pre_clip': 19.166555404663086, 'timing/data': 0.00041444599628448486, 'timing/model': 1.1323202309431508, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.11}) +[2026-06-23 20:33:44,671][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.41902807354927063, 'train/grad_norm_pre_clip': 47.34340286254883, 'timing/data': 0.00035263190511614084, 'timing/model': 1.1315034799044952, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.11}) +[2026-06-23 20:33:48,890][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.2473817765712738, 'train/grad_norm_pre_clip': 46.95383071899414, 'timing/data': 0.0003557739546522498, 'timing/model': 1.2249458460137248, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.11}) +[2026-06-23 20:33:53,028][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.17725937068462372, 'train/grad_norm_pre_clip': 119.59535217285156, 'timing/data': 0.00040819100104272366, 'timing/model': 1.1373072280548513, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.11}) +[2026-06-23 20:33:57,161][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.21744778752326965, 'train/grad_norm_pre_clip': 84.18854522705078, 'timing/data': 0.0006858679698780179, 'timing/model': 1.1283560170559213, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.11}) +[2026-06-23 20:34:01,281][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.1060023084282875, 'train/grad_norm_pre_clip': 130.13397216796875, 'timing/data': 0.00035857304465025663, 'timing/model': 1.1289741719374433, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.11}) +[2026-06-23 20:34:05,577][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.13527674973011017, 'train/grad_norm_pre_clip': 44.47145080566406, 'timing/data': 0.00034950696863234043, 'timing/model': 1.2962031950009987, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.11}) +[2026-06-23 20:34:09,712][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.20244459807872772, 'train/grad_norm_pre_clip': 66.55460357666016, 'timing/data': 0.0003709739539772272, 'timing/model': 1.1333161550574005, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.12}) +[2026-06-23 20:34:13,840][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.217476949095726, 'train/grad_norm_pre_clip': 27.585248947143555, 'timing/data': 0.00035331607796251774, 'timing/model': 1.1321988819399849, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.12}) +[2026-06-23 20:34:17,972][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 300 +[2026-06-23 20:35:22,560][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 300 +[2026-06-23 20:35:22,562][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 300 +[2026-06-23 20:36:03,407][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 300 +[2026-06-23 20:36:03,407][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.25050970911979675, 'train/grad_norm_pre_clip': 42.61213302612305, 'eval/action_loss/seconds': 64.58596529799979, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.1153707280755043, 'eval/latency_2/loss': 0.1153707280755043, 'eval/flappy/loss': 0.1153707280755043, 'eval/flappy/latency_2/loss': 0.1153707280755043, 'eval/flap_precision': 0.9645868465430016, 'eval/flap_recall': 0.7761194029850746, 'eval/flap_f1': 0.8601503759398497, 'eval/flap_rate_model': 0.16472222222222221, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8601503759398497, 'eval/latency_2/flap_precision': 0.9645868465430016, 'eval/latency_2/flap_recall': 0.7761194029850746, 'eval/latency_2/flap_f1': 0.8601503759398497, 'eval/latency_2/flap_rate_model': 0.16472222222222221, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8601503759398497, 'eval/action_classification/data_wait_seconds': 0.6005173921585083, 'eval/action_classification/predict_seconds': 40.119014739990234, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00035338406451046467, 'timing/model': 1.127501411945559, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.12}) +[2026-06-23 20:36:07,515][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.2932148277759552, 'train/grad_norm_pre_clip': 79.324462890625, 'timing/data': 0.000362978084012866, 'timing/model': 1.1171387730864808, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.12}) +[2026-06-23 20:36:11,620][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.09960908442735672, 'train/grad_norm_pre_clip': 38.450828552246094, 'timing/data': 0.00034978101029992104, 'timing/model': 1.119397334055975, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.12}) +[2026-06-23 20:36:15,702][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.13350901007652283, 'train/grad_norm_pre_clip': 30.688335418701172, 'timing/data': 0.0003680059453472495, 'timing/model': 1.1215814049355686, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.12}) +[2026-06-23 20:36:19,884][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.3633204698562622, 'train/grad_norm_pre_clip': 66.8553695678711, 'timing/data': 0.0003958079032599926, 'timing/model': 1.123960891040042, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.12}) +[2026-06-23 20:36:23,989][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.12216972559690475, 'train/grad_norm_pre_clip': 26.301477432250977, 'timing/data': 0.00040497793816030025, 'timing/model': 1.1247478530276567, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.12}) +[2026-06-23 20:36:28,107][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.14301133155822754, 'train/grad_norm_pre_clip': 24.13564682006836, 'timing/data': 0.00034734595101326704, 'timing/model': 1.126935635926202, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.12}) +[2026-06-23 20:36:32,214][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.11068998277187347, 'train/grad_norm_pre_clip': 40.149253845214844, 'timing/data': 0.0005357079207897186, 'timing/model': 1.1253111800178885, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.12}) +[2026-06-23 20:36:36,385][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.124247707426548, 'train/grad_norm_pre_clip': 42.2041130065918, 'timing/data': 0.0005352400476112962, 'timing/model': 1.1211048400728032, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.12}) +[2026-06-23 20:36:40,495][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.2775283455848694, 'train/grad_norm_pre_clip': 64.89017486572266, 'timing/data': 0.00034849089570343494, 'timing/model': 1.1306326909689233, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.12}) +[2026-06-23 20:36:44,612][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.2765294015407562, 'train/grad_norm_pre_clip': 62.979496002197266, 'timing/data': 0.0003509910311549902, 'timing/model': 1.1262982069747522, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.12}) +[2026-06-23 20:36:48,727][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.12307634204626083, 'train/grad_norm_pre_clip': 72.21236419677734, 'timing/data': 0.0005257099401205778, 'timing/model': 1.125651011010632, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.12}) +[2026-06-23 20:36:52,980][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.27064061164855957, 'train/grad_norm_pre_clip': 50.942054748535156, 'timing/data': 0.0003621849464252591, 'timing/model': 1.1202009710250422, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.12}) +[2026-06-23 20:36:57,107][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.5986982583999634, 'train/grad_norm_pre_clip': 161.53150939941406, 'timing/data': 0.00042010797187685966, 'timing/model': 1.1250032329699025, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.12}) +[2026-06-23 20:37:01,224][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.14063510298728943, 'train/grad_norm_pre_clip': 55.96810531616211, 'timing/data': 0.00039823504630476236, 'timing/model': 1.1265894640237093, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.12}) +[2026-06-23 20:37:05,337][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.10974491387605667, 'train/grad_norm_pre_clip': 30.071325302124023, 'timing/data': 0.000367230037227273, 'timing/model': 1.1235101519851014, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.12}) +[2026-06-23 20:37:09,627][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.10446661710739136, 'train/grad_norm_pre_clip': 70.73169708251953, 'timing/data': 0.0003640849608927965, 'timing/model': 1.1287728670286015, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.12}) +[2026-06-23 20:37:13,748][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.12526637315750122, 'train/grad_norm_pre_clip': 22.5356388092041, 'timing/data': 0.0004190730396658182, 'timing/model': 1.1283178819576278, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.12}) +[2026-06-23 20:37:17,861][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.17693176865577698, 'train/grad_norm_pre_clip': 16.345657348632812, 'timing/data': 0.00040416792035102844, 'timing/model': 1.1250572031131014, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.12}) +[2026-06-23 20:37:21,980][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.24760481715202332, 'train/grad_norm_pre_clip': 67.84626007080078, 'timing/data': 0.0003847799962386489, 'timing/model': 1.1285049460129812, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.12}) +[2026-06-23 20:37:26,222][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.1173352375626564, 'train/grad_norm_pre_clip': 42.89469909667969, 'timing/data': 0.000699323951266706, 'timing/model': 1.1244432379025966, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.12}) +[2026-06-23 20:37:30,346][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.1337863951921463, 'train/grad_norm_pre_clip': 33.48772048950195, 'timing/data': 0.0004027469549328089, 'timing/model': 1.1286898489343002, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.12}) +[2026-06-23 20:37:34,467][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.23600660264492035, 'train/grad_norm_pre_clip': 40.15849304199219, 'timing/data': 0.0005587519844993949, 'timing/model': 1.1317654160084203, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.12}) +[2026-06-23 20:37:38,587][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.28351959586143494, 'train/grad_norm_pre_clip': 32.299400329589844, 'timing/data': 0.0005543529987335205, 'timing/model': 1.1243922169087455, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.13}) +[2026-06-23 20:37:42,789][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.1245456337928772, 'train/grad_norm_pre_clip': 42.882625579833984, 'timing/data': 0.00039927102625370026, 'timing/model': 1.1280555130215362, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.13}) +[2026-06-23 20:37:46,912][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.11816851794719696, 'train/grad_norm_pre_clip': 42.5664176940918, 'timing/data': 0.0003565630177035928, 'timing/model': 1.1271378099918365, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.13}) +[2026-06-23 20:37:51,041][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.30148082971572876, 'train/grad_norm_pre_clip': 22.908245086669922, 'timing/data': 0.00046486896462738514, 'timing/model': 1.130600219941698, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.13}) +[2026-06-23 20:37:55,172][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.06924901157617569, 'train/grad_norm_pre_clip': 43.95295715332031, 'timing/data': 0.0005572170484811068, 'timing/model': 1.1242957280483097, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.13}) +[2026-06-23 20:37:59,459][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.17078790068626404, 'train/grad_norm_pre_clip': 20.355932235717773, 'timing/data': 0.0003574700094759464, 'timing/model': 1.12488170794677, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.13}) +[2026-06-23 20:38:03,587][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.08196773380041122, 'train/grad_norm_pre_clip': 31.62921142578125, 'timing/data': 0.00042410101741552353, 'timing/model': 1.1304487090092152, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.13}) +[2026-06-23 20:38:07,713][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.19785000383853912, 'train/grad_norm_pre_clip': 41.075645446777344, 'timing/data': 0.00047263200394809246, 'timing/model': 1.1272458860184997, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.13}) +[2026-06-23 20:38:12,005][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.20949599146842957, 'train/grad_norm_pre_clip': 86.95894622802734, 'timing/data': 0.0005765989189967513, 'timing/model': 1.2747601229930297, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.13}) +[2026-06-23 20:38:16,153][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.20118416845798492, 'train/grad_norm_pre_clip': 79.43038940429688, 'timing/data': 0.00040903699118644, 'timing/model': 1.1287768529728055, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.13}) +[2026-06-23 20:38:20,278][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.15013059973716736, 'train/grad_norm_pre_clip': 48.05869674682617, 'timing/data': 0.00036907801404595375, 'timing/model': 1.1275701760314405, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.13}) +[2026-06-23 20:38:24,414][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.21836957335472107, 'train/grad_norm_pre_clip': 60.967132568359375, 'timing/data': 0.0005040529649704695, 'timing/model': 1.1333142779767513, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.13}) +[2026-06-23 20:38:28,650][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.30395689606666565, 'train/grad_norm_pre_clip': 115.17039489746094, 'timing/data': 0.000557646038942039, 'timing/model': 1.1222019789274782, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.13}) +[2026-06-23 20:38:32,791][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.26176753640174866, 'train/grad_norm_pre_clip': 60.01038360595703, 'timing/data': 0.000457984977401793, 'timing/model': 1.136133143096231, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.13}) +[2026-06-23 20:38:36,932][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.1650533229112625, 'train/grad_norm_pre_clip': 40.25881576538086, 'timing/data': 0.00037465093191713095, 'timing/model': 1.1327974050072953, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.13}) +[2026-06-23 20:38:41,048][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.08207879215478897, 'train/grad_norm_pre_clip': 63.56121063232422, 'timing/data': 0.00035592401400208473, 'timing/model': 1.1270723439520225, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.13}) +[2026-06-23 20:38:45,246][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.10636048763990402, 'train/grad_norm_pre_clip': 37.64301681518555, 'timing/data': 0.00036132102832198143, 'timing/model': 1.1229641569079831, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.13}) +[2026-06-23 20:38:49,377][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.22960573434829712, 'train/grad_norm_pre_clip': 47.059261322021484, 'timing/data': 0.0003623400116339326, 'timing/model': 1.1354710489977151, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.13}) +[2026-06-23 20:38:53,512][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.23552197217941284, 'train/grad_norm_pre_clip': 47.20492935180664, 'timing/data': 0.00040228094439953566, 'timing/model': 1.129151177010499, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.13}) +[2026-06-23 20:38:57,643][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.16904667019844055, 'train/grad_norm_pre_clip': 32.28645324707031, 'timing/data': 0.0004727169871330261, 'timing/model': 1.1334631730569527, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.13}) +[2026-06-23 20:39:01,853][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.2133401334285736, 'train/grad_norm_pre_clip': 34.11574172973633, 'timing/data': 0.0004191519692540169, 'timing/model': 1.1270316060399637, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.13}) +[2026-06-23 20:39:05,979][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.10671981424093246, 'train/grad_norm_pre_clip': 43.01801681518555, 'timing/data': 0.0006278170039877295, 'timing/model': 1.1296842459123582, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.13}) +[2026-06-23 20:39:10,100][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.1703350991010666, 'train/grad_norm_pre_clip': 29.647491455078125, 'timing/data': 0.0003562450874596834, 'timing/model': 1.1278101260541007, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.13}) +[2026-06-23 20:39:14,229][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.04337034374475479, 'train/grad_norm_pre_clip': 24.45907211303711, 'timing/data': 0.0005814959295094013, 'timing/model': 1.1305835719686002, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.13}) +[2026-06-23 20:39:18,469][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.1471378207206726, 'train/grad_norm_pre_clip': 76.76334381103516, 'timing/data': 0.00039588601794093847, 'timing/model': 1.1320603960193694, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.13}) +[2026-06-23 20:39:22,597][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.31351351737976074, 'train/grad_norm_pre_clip': 102.3641357421875, 'timing/data': 0.00036012998316437006, 'timing/model': 1.1284341879654676, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.13}) +[2026-06-23 20:39:26,724][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.31938111782073975, 'train/grad_norm_pre_clip': 83.8314208984375, 'timing/data': 0.00040486198849976063, 'timing/model': 1.1258457620861009, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.14}) +[2026-06-23 20:39:30,920][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.31026214361190796, 'train/grad_norm_pre_clip': 36.24256896972656, 'timing/data': 0.00035577104426920414, 'timing/model': 1.2036564560839906, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.14}) +[2026-06-23 20:39:35,046][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.18929317593574524, 'train/grad_norm_pre_clip': 103.0802230834961, 'timing/data': 0.00035580003168433905, 'timing/model': 1.1299540679901838, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.14}) +[2026-06-23 20:39:39,196][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.1335943341255188, 'train/grad_norm_pre_clip': 54.07836151123047, 'timing/data': 0.00035349803511053324, 'timing/model': 1.12956914503593, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.14}) +[2026-06-23 20:39:43,327][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.3149557113647461, 'train/grad_norm_pre_clip': 49.393470764160156, 'timing/data': 0.00042863795533776283, 'timing/model': 1.1283157779835165, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.14}) +[2026-06-23 20:39:47,448][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.14426815509796143, 'train/grad_norm_pre_clip': 28.82135581970215, 'timing/data': 0.00040971196722239256, 'timing/model': 1.1284706080332398, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.14}) +[2026-06-23 20:39:51,706][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.05692717060446739, 'train/grad_norm_pre_clip': 21.628734588623047, 'timing/data': 0.000465641962364316, 'timing/model': 1.1336249930318445, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.14}) +[2026-06-23 20:39:55,835][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.17147105932235718, 'train/grad_norm_pre_clip': 32.5994987487793, 'timing/data': 0.00036341091617941856, 'timing/model': 1.1275882920017466, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.14}) +[2026-06-23 20:39:59,969][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.08717785775661469, 'train/grad_norm_pre_clip': 28.444873809814453, 'timing/data': 0.0004440059419721365, 'timing/model': 1.1342679080553353, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.14}) +[2026-06-23 20:40:04,113][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.11749960482120514, 'train/grad_norm_pre_clip': 21.713571548461914, 'timing/data': 0.0003651130245998502, 'timing/model': 1.135010399040766, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.14}) +[2026-06-23 20:40:08,394][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.2642005383968353, 'train/grad_norm_pre_clip': 56.45569610595703, 'timing/data': 0.00035584403667598963, 'timing/model': 1.1280348569853231, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.14}) +[2026-06-23 20:40:12,534][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.12073714286088943, 'train/grad_norm_pre_clip': 50.90327072143555, 'timing/data': 0.0006521029863506556, 'timing/model': 1.1257831359980628, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.14}) +[2026-06-23 20:40:16,662][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.1207534521818161, 'train/grad_norm_pre_clip': 29.545883178710938, 'timing/data': 0.0003523590276017785, 'timing/model': 1.133919037063606, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.14}) +[2026-06-23 20:40:20,789][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.03971575200557709, 'train/grad_norm_pre_clip': 30.940214157104492, 'timing/data': 0.00035397999454289675, 'timing/model': 1.1261308450484648, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.14}) +[2026-06-23 20:40:24,998][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.1519000381231308, 'train/grad_norm_pre_clip': 47.07450866699219, 'timing/data': 0.00042032997589558363, 'timing/model': 1.128373623942025, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.14}) +[2026-06-23 20:40:29,141][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.05338655784726143, 'train/grad_norm_pre_clip': 27.594120025634766, 'timing/data': 0.0006560749607160687, 'timing/model': 1.1420645089820027, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.14}) +[2026-06-23 20:40:33,272][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.05173322558403015, 'train/grad_norm_pre_clip': 36.98247528076172, 'timing/data': 0.0005306529346853495, 'timing/model': 1.1283858390524983, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.14}) +[2026-06-23 20:40:37,404][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.12245678156614304, 'train/grad_norm_pre_clip': 22.967496871948242, 'timing/data': 0.0005625910125672817, 'timing/model': 1.1352093330351636, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.14}) +[2026-06-23 20:40:41,678][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.1254633069038391, 'train/grad_norm_pre_clip': 45.08272933959961, 'timing/data': 0.0005874560447409749, 'timing/model': 1.1266381500754505, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.14}) +[2026-06-23 20:40:45,814][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.20681264996528625, 'train/grad_norm_pre_clip': 29.12498664855957, 'timing/data': 0.00035938399378210306, 'timing/model': 1.1277548599755391, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.14}) +[2026-06-23 20:40:50,006][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.09170566499233246, 'train/grad_norm_pre_clip': 32.313777923583984, 'timing/data': 0.0004154510097578168, 'timing/model': 1.128811185946688, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.14}) +[2026-06-23 20:40:54,130][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.18820108473300934, 'train/grad_norm_pre_clip': 52.22438049316406, 'timing/data': 0.0003533789422363043, 'timing/model': 1.1321900279726833, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.14}) +[2026-06-23 20:40:58,364][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.1598302274942398, 'train/grad_norm_pre_clip': 42.44390106201172, 'timing/data': 0.0003539019962772727, 'timing/model': 1.1266321539878845, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.14}) +[2026-06-23 20:41:02,497][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.05772321671247482, 'train/grad_norm_pre_clip': 18.333154678344727, 'timing/data': 0.00035447406116873026, 'timing/model': 1.1304403799585998, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.14}) +[2026-06-23 20:41:06,627][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.11594817787408829, 'train/grad_norm_pre_clip': 20.49954605102539, 'timing/data': 0.00035581004340201616, 'timing/model': 1.133328833966516, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.14}) +[2026-06-23 20:41:10,764][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.17030024528503418, 'train/grad_norm_pre_clip': 24.514999389648438, 'timing/data': 0.0003646330442279577, 'timing/model': 1.130763030028902, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.14}) +[2026-06-23 20:41:15,060][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.23096974194049835, 'train/grad_norm_pre_clip': 30.368572235107422, 'timing/data': 0.00043102295603603125, 'timing/model': 1.126034136977978, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.15}) +[2026-06-23 20:41:19,176][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.16243472695350647, 'train/grad_norm_pre_clip': 19.97602081298828, 'timing/data': 0.0003641319926828146, 'timing/model': 1.1221400300273672, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.15}) +[2026-06-23 20:41:23,318][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.3306930959224701, 'train/grad_norm_pre_clip': 56.30221176147461, 'timing/data': 0.0006369788898155093, 'timing/model': 1.1428931900300086, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.15}) +[2026-06-23 20:41:27,523][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.12469286471605301, 'train/grad_norm_pre_clip': 41.3355598449707, 'timing/data': 0.000444684992544353, 'timing/model': 1.2052860849071294, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.15}) +[2026-06-23 20:41:31,659][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.1522098332643509, 'train/grad_norm_pre_clip': 21.08353614807129, 'timing/data': 0.00035729899536818266, 'timing/model': 1.1368626379407942, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.15}) +[2026-06-23 20:41:35,783][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.11246755719184875, 'train/grad_norm_pre_clip': 56.40003967285156, 'timing/data': 0.0003836170071735978, 'timing/model': 1.1282443819800392, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.15}) +[2026-06-23 20:41:39,921][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.20783516764640808, 'train/grad_norm_pre_clip': 49.18694305419922, 'timing/data': 0.00035936501808464527, 'timing/model': 1.1227573809446767, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.15}) +[2026-06-23 20:41:44,130][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.2663959860801697, 'train/grad_norm_pre_clip': 42.671974182128906, 'timing/data': 0.0004920709179714322, 'timing/model': 1.214738092967309, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.15}) +[2026-06-23 20:41:48,269][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.24684786796569824, 'train/grad_norm_pre_clip': 46.56484603881836, 'timing/data': 0.0004924439126625657, 'timing/model': 1.1316281540784985, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.15}) +[2026-06-23 20:41:52,392][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.27737659215927124, 'train/grad_norm_pre_clip': 87.35403442382812, 'timing/data': 0.0003568320535123348, 'timing/model': 1.1311280420050025, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.15}) +[2026-06-23 20:41:56,520][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.2021373063325882, 'train/grad_norm_pre_clip': 48.52881622314453, 'timing/data': 0.0006128930253908038, 'timing/model': 1.1307367640547454, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.15}) +[2026-06-23 20:42:00,791][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.21893949806690216, 'train/grad_norm_pre_clip': 24.301679611206055, 'timing/data': 0.00035845290403813124, 'timing/model': 1.2625528629869223, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.15}) +[2026-06-23 20:42:04,937][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.14470021426677704, 'train/grad_norm_pre_clip': 25.380502700805664, 'timing/data': 0.000536541105248034, 'timing/model': 1.1302440980216488, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.15}) +[2026-06-23 20:42:09,064][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.1913958489894867, 'train/grad_norm_pre_clip': 32.91065979003906, 'timing/data': 0.0003653171006590128, 'timing/model': 1.128431451972574, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.15}) +[2026-06-23 20:42:13,209][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.12361707538366318, 'train/grad_norm_pre_clip': 19.74186897277832, 'timing/data': 0.00040077592711895704, 'timing/model': 1.134666771045886, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.15}) +[2026-06-23 20:42:17,455][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.1340429186820984, 'train/grad_norm_pre_clip': 25.603084564208984, 'timing/data': 0.000356901902705431, 'timing/model': 1.1289688140386716, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.15}) +[2026-06-23 20:42:21,579][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.243142768740654, 'train/grad_norm_pre_clip': 76.81407165527344, 'timing/data': 0.00041957502253353596, 'timing/model': 1.1292886309092864, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.15}) +[2026-06-23 20:42:25,723][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.20276464521884918, 'train/grad_norm_pre_clip': 36.7901725769043, 'timing/data': 0.00035929190926253796, 'timing/model': 1.1283951880177483, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.15}) +[2026-06-23 20:42:29,859][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.16982920467853546, 'train/grad_norm_pre_clip': 55.40066146850586, 'timing/data': 0.00034606701228767633, 'timing/model': 1.1362439570948482, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.15}) +[2026-06-23 20:42:34,091][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.17371788620948792, 'train/grad_norm_pre_clip': 30.836017608642578, 'timing/data': 0.0003912170650437474, 'timing/model': 1.1293113980209455, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.15}) +[2026-06-23 20:42:38,217][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.14786604046821594, 'train/grad_norm_pre_clip': 23.552919387817383, 'timing/data': 0.0003570250701159239, 'timing/model': 1.1272159609943628, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.15}) +[2026-06-23 20:42:42,363][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.15339183807373047, 'train/grad_norm_pre_clip': 44.265838623046875, 'timing/data': 0.00040320109110325575, 'timing/model': 1.1328090189490467, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.15}) +[2026-06-23 20:42:46,504][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.2208099216222763, 'train/grad_norm_pre_clip': 58.209449768066406, 'timing/data': 0.000520537025295198, 'timing/model': 1.1379163569072261, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.15}) +[2026-06-23 20:42:50,699][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.11253701895475388, 'train/grad_norm_pre_clip': 19.40338134765625, 'timing/data': 0.00037972594145685434, 'timing/model': 1.1275672370102257, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.15}) +[2026-06-23 20:42:54,815][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.05411001294851303, 'train/grad_norm_pre_clip': 24.15155029296875, 'timing/data': 0.0003588589606806636, 'timing/model': 1.123560157022439, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.15}) +[2026-06-23 20:42:58,945][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 400 +[2026-06-23 20:44:04,046][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 400 +[2026-06-23 20:44:04,047][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 400 +[2026-06-23 20:44:44,010][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 400 +[2026-06-23 20:44:44,010][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.1828538030385971, 'train/grad_norm_pre_clip': 16.658769607543945, 'eval/action_loss/seconds': 65.09851406491362, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.08767598867416382, 'eval/latency_2/loss': 0.08767598867416382, 'eval/flappy/loss': 0.08767598867416382, 'eval/flappy/latency_2/loss': 0.08767598867416382, 'eval/flap_precision': 0.8792029887920298, 'eval/flap_recall': 0.9579375848032564, 'eval/flap_f1': 0.9168831168831169, 'eval/flap_rate_model': 0.22305555555555556, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9168831168831169, 'eval/latency_2/flap_precision': 0.8792029887920298, 'eval/latency_2/flap_recall': 0.9579375848032564, 'eval/latency_2/flap_f1': 0.9168831168831169, 'eval/latency_2/flap_rate_model': 0.22305555555555556, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9168831168831169, 'eval/action_classification/data_wait_seconds': 0.536372184753418, 'eval/action_classification/predict_seconds': 39.25762176513672, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.000352094997651875, 'timing/model': 1.131984414998442, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.15}) +[2026-06-23 20:44:48,194][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.2717674970626831, 'train/grad_norm_pre_clip': 119.8575439453125, 'timing/data': 0.000609629089012742, 'timing/model': 1.116273691994138, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.16}) +[2026-06-23 20:44:52,284][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.04086970537900925, 'train/grad_norm_pre_clip': 37.302024841308594, 'timing/data': 0.00041264097671955824, 'timing/model': 1.1178347750101238, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.16}) +[2026-06-23 20:44:56,372][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.15409207344055176, 'train/grad_norm_pre_clip': 44.71241760253906, 'timing/data': 0.0003487570211291313, 'timing/model': 1.1191140299197286, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.16}) +[2026-06-23 20:45:00,468][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.1318567395210266, 'train/grad_norm_pre_clip': 24.408708572387695, 'timing/data': 0.00038361106999218464, 'timing/model': 1.1202213349752128, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.16}) +[2026-06-23 20:45:04,701][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.09550533443689346, 'train/grad_norm_pre_clip': 33.27190399169922, 'timing/data': 0.0005721490597352386, 'timing/model': 1.127540760091506, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.16}) +[2026-06-23 20:45:08,801][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.1587483137845993, 'train/grad_norm_pre_clip': 33.793006896972656, 'timing/data': 0.00044733507093042135, 'timing/model': 1.118154428084381, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.16}) +[2026-06-23 20:45:12,910][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.3096863925457001, 'train/grad_norm_pre_clip': 164.8882598876953, 'timing/data': 0.00035394495353102684, 'timing/model': 1.1232839080039412, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.16}) +[2026-06-23 20:45:17,118][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.32462751865386963, 'train/grad_norm_pre_clip': 89.27064514160156, 'timing/data': 0.0003683950053527951, 'timing/model': 1.2112643820000812, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.16}) +[2026-06-23 20:45:21,241][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.12062717229127884, 'train/grad_norm_pre_clip': 97.1833267211914, 'timing/data': 0.0005184320034459233, 'timing/model': 1.1262137240264565, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.16}) +[2026-06-23 20:45:25,348][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.1502898931503296, 'train/grad_norm_pre_clip': 32.762237548828125, 'timing/data': 0.00036440102849155664, 'timing/model': 1.1233288020594046, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.16}) +[2026-06-23 20:45:29,461][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.089418925344944, 'train/grad_norm_pre_clip': 87.90116882324219, 'timing/data': 0.0005549639463424683, 'timing/model': 1.1242990700993687, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.16}) +[2026-06-23 20:45:33,577][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.6013839840888977, 'train/grad_norm_pre_clip': 99.22581481933594, 'timing/data': 0.00041192397475242615, 'timing/model': 1.1268229479901493, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.16}) +[2026-06-23 20:45:37,817][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.24318109452724457, 'train/grad_norm_pre_clip': 70.07859802246094, 'timing/data': 0.00036239204928278923, 'timing/model': 1.1325577750103548, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.16}) +[2026-06-23 20:45:41,922][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.20601171255111694, 'train/grad_norm_pre_clip': 29.1513729095459, 'timing/data': 0.00035528000444173813, 'timing/model': 1.1266291729407385, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.16}) +[2026-06-23 20:45:46,039][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.10409533977508545, 'train/grad_norm_pre_clip': 22.598915100097656, 'timing/data': 0.0004108030116185546, 'timing/model': 1.1264141129795462, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.16}) +[2026-06-23 20:45:50,156][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.2135763168334961, 'train/grad_norm_pre_clip': 26.17344856262207, 'timing/data': 0.00035023398231714964, 'timing/model': 1.1329105210024863, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.16}) +[2026-06-23 20:45:54,429][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.07156460732221603, 'train/grad_norm_pre_clip': 15.545415878295898, 'timing/data': 0.00036404095590114594, 'timing/model': 1.1268163800705224, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.16}) +[2026-06-23 20:45:58,553][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.14882627129554749, 'train/grad_norm_pre_clip': 21.01097869873047, 'timing/data': 0.0005948740290477872, 'timing/model': 1.1260167019208893, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.16}) +[2026-06-23 20:46:02,698][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.19898563623428345, 'train/grad_norm_pre_clip': 29.374473571777344, 'timing/data': 0.0004716250114142895, 'timing/model': 1.1361871120752767, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.16}) +[2026-06-23 20:46:06,831][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.19098125398159027, 'train/grad_norm_pre_clip': 66.12100982666016, 'timing/data': 0.0003518239827826619, 'timing/model': 1.1297451748978347, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.16}) +[2026-06-23 20:46:11,039][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.05610092356801033, 'train/grad_norm_pre_clip': 28.235328674316406, 'timing/data': 0.00046061701141297817, 'timing/model': 1.1320508739445359, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.16}) +[2026-06-23 20:46:15,168][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.11859170347452164, 'train/grad_norm_pre_clip': 51.473934173583984, 'timing/data': 0.0004128210712224245, 'timing/model': 1.1257646779995412, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.16}) +[2026-06-23 20:46:19,299][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.23411740362644196, 'train/grad_norm_pre_clip': 34.067569732666016, 'timing/data': 0.0003551250556483865, 'timing/model': 1.1369317920180038, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.16}) +[2026-06-23 20:46:23,434][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.030569549649953842, 'train/grad_norm_pre_clip': 38.89982223510742, 'timing/data': 0.00035131804179400206, 'timing/model': 1.1278380929725245, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.16}) +[2026-06-23 20:46:27,719][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.1128525510430336, 'train/grad_norm_pre_clip': 59.58738708496094, 'timing/data': 0.0005268779350444674, 'timing/model': 1.131173935951665, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.16}) +[2026-06-23 20:46:31,850][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.10106238722801208, 'train/grad_norm_pre_clip': 26.984006881713867, 'timing/data': 0.0003525209613144398, 'timing/model': 1.1319611660437658, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.16}) +[2026-06-23 20:46:35,986][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.15175727009773254, 'train/grad_norm_pre_clip': 28.132078170776367, 'timing/data': 0.0005195740377530456, 'timing/model': 1.124144027940929, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.17}) +[2026-06-23 20:46:40,125][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.06703626364469528, 'train/grad_norm_pre_clip': 22.01805305480957, 'timing/data': 0.0005939590046182275, 'timing/model': 1.131484901998192, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.17}) +[2026-06-23 20:46:44,372][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.13970443606376648, 'train/grad_norm_pre_clip': 20.122772216796875, 'timing/data': 0.00040229002479463816, 'timing/model': 1.1386573340278119, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.17}) +[2026-06-23 20:46:48,510][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.2605784237384796, 'train/grad_norm_pre_clip': 23.281747817993164, 'timing/data': 0.00038149196188896894, 'timing/model': 1.1350620469311252, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.17}) +[2026-06-23 20:46:52,659][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.031308289617300034, 'train/grad_norm_pre_clip': 20.22089958190918, 'timing/data': 0.0003602949436753988, 'timing/model': 1.1341675160219893, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.17}) +[2026-06-23 20:46:56,797][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.14901210367679596, 'train/grad_norm_pre_clip': 45.45485305786133, 'timing/data': 0.0005078969988971949, 'timing/model': 1.125192602048628, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.17}) +[2026-06-23 20:47:01,084][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.15210552513599396, 'train/grad_norm_pre_clip': 66.70039367675781, 'timing/data': 0.00048375397454947233, 'timing/model': 1.1248836650047451, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.17}) +[2026-06-23 20:47:05,208][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.3255891501903534, 'train/grad_norm_pre_clip': 52.09638214111328, 'timing/data': 0.00041824299842119217, 'timing/model': 1.127183983917348, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.17}) +[2026-06-23 20:47:09,337][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.08927995711565018, 'train/grad_norm_pre_clip': 35.035682678222656, 'timing/data': 0.00037137605249881744, 'timing/model': 1.131133683025837, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.17}) +[2026-06-23 20:47:13,546][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.14163126051425934, 'train/grad_norm_pre_clip': 54.33782958984375, 'timing/data': 0.0004131470341235399, 'timing/model': 1.2085794889135286, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.17}) +[2026-06-23 20:47:17,685][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.15613317489624023, 'train/grad_norm_pre_clip': 48.193382263183594, 'timing/data': 0.0004252060316503048, 'timing/model': 1.135419175028801, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.17}) +[2026-06-23 20:47:21,829][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.4911196529865265, 'train/grad_norm_pre_clip': 80.84441375732422, 'timing/data': 0.0005715610459446907, 'timing/model': 1.1378477000398561, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.17}) +[2026-06-23 20:47:25,964][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.11663869768381119, 'train/grad_norm_pre_clip': 67.68266296386719, 'timing/data': 0.00035652995575219393, 'timing/model': 1.1311258400091901, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.17}) +[2026-06-23 20:47:30,166][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.08150514215230942, 'train/grad_norm_pre_clip': 55.28195571899414, 'timing/data': 0.0003638860071077943, 'timing/model': 1.1990841770311818, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.17}) +[2026-06-23 20:47:34,306][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.10037749260663986, 'train/grad_norm_pre_clip': 39.959991455078125, 'timing/data': 0.0003544379724189639, 'timing/model': 1.1311702050734311, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.17}) +[2026-06-23 20:47:38,427][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.1668601930141449, 'train/grad_norm_pre_clip': 71.42546081542969, 'timing/data': 0.0004853989230468869, 'timing/model': 1.1321712599601597, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.17}) +[2026-06-23 20:47:42,557][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.15624624490737915, 'train/grad_norm_pre_clip': 37.87490463256836, 'timing/data': 0.0004004359943792224, 'timing/model': 1.1294526149285957, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.17}) +[2026-06-23 20:47:46,823][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.13868746161460876, 'train/grad_norm_pre_clip': 52.440494537353516, 'timing/data': 0.0004109180299565196, 'timing/model': 1.267572515993379, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.17}) +[2026-06-23 20:47:50,953][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.10866221785545349, 'train/grad_norm_pre_clip': 50.892295837402344, 'timing/data': 0.0005042770644649863, 'timing/model': 1.1313841639785096, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.17}) +[2026-06-23 20:47:55,094][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.2944834530353546, 'train/grad_norm_pre_clip': 42.90663528442383, 'timing/data': 0.0003599710762500763, 'timing/model': 1.1294968270231038, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.17}) +[2026-06-23 20:47:59,226][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.04707997292280197, 'train/grad_norm_pre_clip': 58.99006652832031, 'timing/data': 0.0003541440237313509, 'timing/model': 1.1321863109478727, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.17}) +[2026-06-23 20:48:03,476][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.05815720185637474, 'train/grad_norm_pre_clip': 25.26665687561035, 'timing/data': 0.0004049669951200485, 'timing/model': 1.1295616260031238, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.17}) +[2026-06-23 20:48:07,620][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.25951525568962097, 'train/grad_norm_pre_clip': 45.6373176574707, 'timing/data': 0.00036259402986615896, 'timing/model': 1.1363601320190355, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.17}) +[2026-06-23 20:48:11,744][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.5114255547523499, 'train/grad_norm_pre_clip': 84.0094223022461, 'timing/data': 0.0004389369860291481, 'timing/model': 1.1281426709610969, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.17}) +[2026-06-23 20:48:15,877][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.18838681280612946, 'train/grad_norm_pre_clip': 56.58799362182617, 'timing/data': 0.00037383707240223885, 'timing/model': 1.1303062459919602, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.17}) +[2026-06-23 20:48:20,084][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.29902592301368713, 'train/grad_norm_pre_clip': 56.138126373291016, 'timing/data': 0.0004280760185793042, 'timing/model': 1.1278455749852583, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.17}) +[2026-06-23 20:48:24,211][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.10733355581760406, 'train/grad_norm_pre_clip': 30.100614547729492, 'timing/data': 0.0004123880062252283, 'timing/model': 1.1311350599862635, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.18}) +[2026-06-23 20:48:28,362][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.2512759864330292, 'train/grad_norm_pre_clip': 59.396636962890625, 'timing/data': 0.00034713291097432375, 'timing/model': 1.1299907530192286, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.18}) +[2026-06-23 20:48:32,492][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.018693018704652786, 'train/grad_norm_pre_clip': 70.01131439208984, 'timing/data': 0.00040661601815372705, 'timing/model': 1.1299144020304084, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.18}) +[2026-06-23 20:48:36,702][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.07017842680215836, 'train/grad_norm_pre_clip': 32.945465087890625, 'timing/data': 0.0005426700226962566, 'timing/model': 1.1334330060053617, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.18}) +[2026-06-23 20:48:40,833][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.027799366042017937, 'train/grad_norm_pre_clip': 13.430392265319824, 'timing/data': 0.0003512349212542176, 'timing/model': 1.1318287659669295, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.18}) +[2026-06-23 20:48:44,971][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.08771729469299316, 'train/grad_norm_pre_clip': 26.9968318939209, 'timing/data': 0.00038717908319085836, 'timing/model': 1.128169603063725, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.18}) +[2026-06-23 20:48:49,108][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.2539886236190796, 'train/grad_norm_pre_clip': 39.86442184448242, 'timing/data': 0.00035803194623440504, 'timing/model': 1.1364699359983206, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.18}) +[2026-06-23 20:48:53,425][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.02574741095304489, 'train/grad_norm_pre_clip': 25.96384620666504, 'timing/data': 0.00041237997356802225, 'timing/model': 1.134879891993478, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.18}) +[2026-06-23 20:48:57,561][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.09104980528354645, 'train/grad_norm_pre_clip': 37.916419982910156, 'timing/data': 0.0003545039799064398, 'timing/model': 1.1276500169187784, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.18}) +[2026-06-23 20:49:01,698][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.14690053462982178, 'train/grad_norm_pre_clip': 42.615821838378906, 'timing/data': 0.00036510301288217306, 'timing/model': 1.1300520999357104, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.18}) +[2026-06-23 20:49:05,929][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.05025589093565941, 'train/grad_norm_pre_clip': 59.700782775878906, 'timing/data': 0.0005643009208142757, 'timing/model': 1.2271399250021204, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.18}) +[2026-06-23 20:49:10,081][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.06732476502656937, 'train/grad_norm_pre_clip': 40.35162353515625, 'timing/data': 0.0003570959670469165, 'timing/model': 1.128928564954549, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.18}) +[2026-06-23 20:49:14,212][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.28654763102531433, 'train/grad_norm_pre_clip': 52.007781982421875, 'timing/data': 0.00036108598578721285, 'timing/model': 1.131420650985092, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.18}) +[2026-06-23 20:49:18,353][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.06712964922189713, 'train/grad_norm_pre_clip': 17.118141174316406, 'timing/data': 0.00035847804974764585, 'timing/model': 1.130384855903685, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.18}) +[2026-06-23 20:49:22,486][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.04180994629859924, 'train/grad_norm_pre_clip': 76.56194305419922, 'timing/data': 0.0006873509846627712, 'timing/model': 1.1313898810185492, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.18}) +[2026-06-23 20:49:26,724][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.19984501600265503, 'train/grad_norm_pre_clip': 52.00544357299805, 'timing/data': 0.0003504599444568157, 'timing/model': 1.1252127490006387, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.18}) +[2026-06-23 20:49:30,859][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.13000592589378357, 'train/grad_norm_pre_clip': 44.933937072753906, 'timing/data': 0.0003517339937388897, 'timing/model': 1.1361744109308347, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.18}) +[2026-06-23 20:49:35,003][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.2706450819969177, 'train/grad_norm_pre_clip': 95.44025421142578, 'timing/data': 0.00044031499419361353, 'timing/model': 1.1296206479892135, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.18}) +[2026-06-23 20:49:39,132][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.6046071648597717, 'train/grad_norm_pre_clip': 116.82437133789062, 'timing/data': 0.0003575430018827319, 'timing/model': 1.1278213820187375, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.18}) +[2026-06-23 20:49:43,407][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.053755082190036774, 'train/grad_norm_pre_clip': 83.69864654541016, 'timing/data': 0.0003594810841605067, 'timing/model': 1.1245352809783071, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.18}) +[2026-06-23 20:49:47,547][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.1886143833398819, 'train/grad_norm_pre_clip': 54.81440734863281, 'timing/data': 0.000547884963452816, 'timing/model': 1.1353080939734355, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.18}) +[2026-06-23 20:49:51,687][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.0959378033876419, 'train/grad_norm_pre_clip': 17.769052505493164, 'timing/data': 0.0003569939872249961, 'timing/model': 1.133147977059707, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.18}) +[2026-06-23 20:49:55,819][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.5254260897636414, 'train/grad_norm_pre_clip': 110.56480407714844, 'timing/data': 0.00035791401751339436, 'timing/model': 1.1259825041051954, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.18}) +[2026-06-23 20:50:00,040][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.41161343455314636, 'train/grad_norm_pre_clip': 120.32328796386719, 'timing/data': 0.0005753160221502185, 'timing/model': 1.1275879309978336, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.18}) +[2026-06-23 20:50:04,210][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.23170240223407745, 'train/grad_norm_pre_clip': 101.04166412353516, 'timing/data': 0.0003625450190156698, 'timing/model': 1.1388094349531457, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.18}) +[2026-06-23 20:50:08,343][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.31145423650741577, 'train/grad_norm_pre_clip': 63.13083267211914, 'timing/data': 0.0005242309998720884, 'timing/model': 1.1347036749357358, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.18}) +[2026-06-23 20:50:12,488][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.20827913284301758, 'train/grad_norm_pre_clip': 21.135446548461914, 'timing/data': 0.00035131291951984167, 'timing/model': 1.1326776728965342, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.19}) +[2026-06-23 20:50:16,791][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.12264934182167053, 'train/grad_norm_pre_clip': 39.12602615356445, 'timing/data': 0.00035244098398834467, 'timing/model': 1.1327085200464353, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.19}) +[2026-06-23 20:50:20,918][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.31592756509780884, 'train/grad_norm_pre_clip': 61.53075408935547, 'timing/data': 0.000363002996891737, 'timing/model': 1.128423824091442, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.19}) +[2026-06-23 20:50:25,053][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.11807616800069809, 'train/grad_norm_pre_clip': 43.60236740112305, 'timing/data': 0.0006897499551996589, 'timing/model': 1.1373972300207242, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.19}) +[2026-06-23 20:50:29,185][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.24892431497573853, 'train/grad_norm_pre_clip': 29.24650001525879, 'timing/data': 0.00036574609111994505, 'timing/model': 1.131075656041503, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.19}) +[2026-06-23 20:50:33,435][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.25016549229621887, 'train/grad_norm_pre_clip': 18.364370346069336, 'timing/data': 0.0003849589265882969, 'timing/model': 1.1293063609628007, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.19}) +[2026-06-23 20:50:37,573][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.12167870998382568, 'train/grad_norm_pre_clip': 39.13639450073242, 'timing/data': 0.0004207598976790905, 'timing/model': 1.1350267110392451, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.19}) +[2026-06-23 20:50:41,704][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.12756498157978058, 'train/grad_norm_pre_clip': 15.587169647216797, 'timing/data': 0.0004974609473720193, 'timing/model': 1.129745653946884, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.19}) +[2026-06-23 20:50:45,846][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.06744445115327835, 'train/grad_norm_pre_clip': 15.894875526428223, 'timing/data': 0.0003630639985203743, 'timing/model': 1.1300888790283352, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.19}) +[2026-06-23 20:50:50,158][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.17475086450576782, 'train/grad_norm_pre_clip': 16.675941467285156, 'timing/data': 0.00036476494278758764, 'timing/model': 1.142222161986865, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.19}) +[2026-06-23 20:50:54,284][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.09863345324993134, 'train/grad_norm_pre_clip': 21.40070152282715, 'timing/data': 0.0007452130084857345, 'timing/model': 1.1310415659099817, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.19}) +[2026-06-23 20:50:58,428][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.6781346797943115, 'train/grad_norm_pre_clip': 118.74201965332031, 'timing/data': 0.00035759899765253067, 'timing/model': 1.1340040310751647, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.19}) +[2026-06-23 20:51:02,650][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.039289627224206924, 'train/grad_norm_pre_clip': 26.73447608947754, 'timing/data': 0.0003555789589881897, 'timing/model': 1.2176519619533792, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.19}) +[2026-06-23 20:51:06,781][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.048739247024059296, 'train/grad_norm_pre_clip': 46.33918762207031, 'timing/data': 0.0003952740225940943, 'timing/model': 1.125989965046756, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.19}) +[2026-06-23 20:51:10,916][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.01620100997388363, 'train/grad_norm_pre_clip': 49.323524475097656, 'timing/data': 0.000545488903298974, 'timing/model': 1.1348606389947236, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.19}) +[2026-06-23 20:51:15,070][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.06870252639055252, 'train/grad_norm_pre_clip': 44.694698333740234, 'timing/data': 0.00036310695577412844, 'timing/model': 1.1316461840178818, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.19}) +[2026-06-23 20:51:19,293][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.1633308231830597, 'train/grad_norm_pre_clip': 85.77391815185547, 'timing/data': 0.00035022106021642685, 'timing/model': 1.2145339390262961, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.19}) +[2026-06-23 20:51:23,436][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.20894832909107208, 'train/grad_norm_pre_clip': 56.252479553222656, 'timing/data': 0.0003593279980123043, 'timing/model': 1.1341714049922302, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.19}) +[2026-06-23 20:51:27,573][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.21932075917720795, 'train/grad_norm_pre_clip': 80.10330963134766, 'timing/data': 0.0003582850331440568, 'timing/model': 1.132742848014459, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.19}) +[2026-06-23 20:51:31,707][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.6346814036369324, 'train/grad_norm_pre_clip': 175.70751953125, 'timing/data': 0.0003628210397437215, 'timing/model': 1.137288011959754, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.19}) +[2026-06-23 20:51:35,992][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.15984126925468445, 'train/grad_norm_pre_clip': 85.0604248046875, 'timing/data': 0.0003650959115475416, 'timing/model': 1.2811039839871228, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.19}) +[2026-06-23 20:51:40,123][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 500 +[2026-06-23 20:52:45,036][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 500 +[2026-06-23 20:52:45,037][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 500 +[2026-06-23 20:53:24,881][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 500 +[2026-06-23 20:58:11,694][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-23 20:58:11,694][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.23901230096817017, 'train/grad_norm_pre_clip': 67.2583999633789, 'eval/action_loss/seconds': 64.91077491396572, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.10603951662778854, 'eval/latency_2/loss': 0.10603951662778854, 'eval/flappy/loss': 0.10603951662778854, 'eval/flappy/latency_2/loss': 0.10603951662778854, 'eval/flap_precision': 0.8220436280137773, 'eval/flap_recall': 0.9715061058344641, 'eval/flap_f1': 0.890547263681592, 'eval/flap_rate_model': 0.24194444444444443, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8905472636815921, 'eval/latency_2/flap_precision': 0.8220436280137773, 'eval/latency_2/flap_recall': 0.9715061058344641, 'eval/latency_2/flap_f1': 0.890547263681592, 'eval/latency_2/flap_rate_model': 0.24194444444444443, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8905472636815921, 'eval/action_classification/data_wait_seconds': 0.6215303540229797, 'eval/action_classification/predict_seconds': 39.05349349975586, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 176.28000000000358, 'rl_games_eval/mid_train/mean_length': 1440.4, 'rl_games_eval/mid_train/std_reward': 163.1896246702057, 'rl_games_eval/mid_train/std_length': 1314.9364395285422, 'rl_games_eval/mid_train/macro_mean_reward': 176.28000000000358, 'rl_games_eval/mid_train/macro_mean_length': 1440.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 176.28000000000358, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1440.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 163.1896246702057, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1314.9364395285422, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 176.28000000000358, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0004942959640175104, 'timing/model': 1.1315121739171445, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.19}) +[2026-06-23 20:58:11,695][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_500_state +[2026-06-23 20:58:11,698][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 20:58:15,498][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_500_state/model.safetensors +[2026-06-23 20:58:23,404][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_500_state/optimizer.bin +[2026-06-23 20:58:23,404][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_500_state/sampler.bin +[2026-06-23 20:58:23,404][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_500_state/sampler_1.bin +[2026-06-23 20:58:23,405][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-23 20:58:23,406][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-23 20:58:23,406][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 20:58:23,427][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 20:58:27,399][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.40028128027915955, 'train/grad_norm_pre_clip': 49.93726348876953, 'timing/data': 0.00045031204354017973, 'timing/model': 1.0837136750342324, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.19}) +[2026-06-23 20:58:31,384][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.24406465888023376, 'train/grad_norm_pre_clip': 34.9091911315918, 'timing/data': 0.0003617920447140932, 'timing/model': 1.0947318060789257, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.19}) +[2026-06-23 20:58:35,487][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.21617606282234192, 'train/grad_norm_pre_clip': 36.009971618652344, 'timing/data': 0.0003610439598560333, 'timing/model': 1.0991485239937901, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.19}) +[2026-06-23 20:58:39,484][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.15661150217056274, 'train/grad_norm_pre_clip': 27.82618522644043, 'timing/data': 0.00036935496609658003, 'timing/model': 1.098136065993458, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.2}) +[2026-06-23 20:58:43,487][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.2406773865222931, 'train/grad_norm_pre_clip': 36.52437973022461, 'timing/data': 0.0004099459620192647, 'timing/model': 1.1008091160329059, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.2}) +[2026-06-23 20:58:47,574][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.22615233063697815, 'train/grad_norm_pre_clip': 22.094432830810547, 'timing/data': 0.00036027899477630854, 'timing/model': 1.1829441479640082, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.2}) +[2026-06-23 20:58:51,574][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.22761750221252441, 'train/grad_norm_pre_clip': 24.12047004699707, 'timing/data': 0.00038169603794813156, 'timing/model': 1.0949218770256266, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.2}) +[2026-06-23 20:58:55,592][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.12954673171043396, 'train/grad_norm_pre_clip': 25.513044357299805, 'timing/data': 0.00035379501059651375, 'timing/model': 1.1065782689256594, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.2}) +[2026-06-23 20:58:59,633][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.12340385466814041, 'train/grad_norm_pre_clip': 28.067384719848633, 'timing/data': 0.0004073890158906579, 'timing/model': 1.1089674989925697, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.2}) +[2026-06-23 20:59:03,672][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.12617634236812592, 'train/grad_norm_pre_clip': 16.824565887451172, 'timing/data': 0.0003526200307533145, 'timing/model': 1.1053259170148522, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.2}) +[2026-06-23 20:59:07,821][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.2854630947113037, 'train/grad_norm_pre_clip': 32.733646392822266, 'timing/data': 0.0003605140373110771, 'timing/model': 1.1100697530200705, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.2}) +[2026-06-23 20:59:11,866][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.36385098099708557, 'train/grad_norm_pre_clip': 24.172903060913086, 'timing/data': 0.00036248203832656145, 'timing/model': 1.1133530410006642, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.2}) +[2026-06-23 20:59:15,909][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.03989941254258156, 'train/grad_norm_pre_clip': 17.782760620117188, 'timing/data': 0.0003565929364413023, 'timing/model': 1.1101691230433062, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.2}) +[2026-06-23 20:59:19,949][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.09605175256729126, 'train/grad_norm_pre_clip': 62.307228088378906, 'timing/data': 0.00036065804306417704, 'timing/model': 1.1088024759665132, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.2}) +[2026-06-23 20:59:24,152][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.08684764057397842, 'train/grad_norm_pre_clip': 34.20775604248047, 'timing/data': 0.0003685259725898504, 'timing/model': 1.1087690179701895, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.2}) +[2026-06-23 20:59:28,226][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.1069527193903923, 'train/grad_norm_pre_clip': 88.39015197753906, 'timing/data': 0.00036782503593713045, 'timing/model': 1.1271003839792684, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.2}) +[2026-06-23 20:59:32,273][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.3132309317588806, 'train/grad_norm_pre_clip': 49.95296859741211, 'timing/data': 0.0003593879519030452, 'timing/model': 1.111167281982489, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.2}) +[2026-06-23 20:59:36,337][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.18278318643569946, 'train/grad_norm_pre_clip': 36.099395751953125, 'timing/data': 0.0004619050305336714, 'timing/model': 1.1113190030446276, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.2}) +[2026-06-23 20:59:40,469][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.11680782586336136, 'train/grad_norm_pre_clip': 34.977561950683594, 'timing/data': 0.00040089094545692205, 'timing/model': 1.1113242009887472, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.2}) +[2026-06-23 20:59:44,530][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.23613935708999634, 'train/grad_norm_pre_clip': 66.87001037597656, 'timing/data': 0.00046485697384923697, 'timing/model': 1.1161206000251696, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.2}) +[2026-06-23 20:59:48,591][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.14908018708229065, 'train/grad_norm_pre_clip': 26.761749267578125, 'timing/data': 0.00036421604454517365, 'timing/model': 1.1155444130999967, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.2}) +[2026-06-23 20:59:52,671][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.10581044107675552, 'train/grad_norm_pre_clip': 52.13792037963867, 'timing/data': 0.0003823420265689492, 'timing/model': 1.1171110810246319, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.2}) +[2026-06-23 20:59:56,844][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.10331253707408905, 'train/grad_norm_pre_clip': 32.27022171020508, 'timing/data': 0.0006003599846735597, 'timing/model': 1.1193610989721492, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.2}) +[2026-06-23 21:00:00,903][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.10905948281288147, 'train/grad_norm_pre_clip': 42.17349624633789, 'timing/data': 0.0003653409657999873, 'timing/model': 1.1163191390223801, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.2}) +[2026-06-23 21:00:04,970][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.05782090499997139, 'train/grad_norm_pre_clip': 18.61447525024414, 'timing/data': 0.0004021869972348213, 'timing/model': 1.1156690020579845, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.2}) +[2026-06-23 21:00:09,026][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.15743069350719452, 'train/grad_norm_pre_clip': 37.76884460449219, 'timing/data': 0.0003524480853229761, 'timing/model': 1.1152453280519694, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.2}) +[2026-06-23 21:00:13,262][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.0960870012640953, 'train/grad_norm_pre_clip': 20.659385681152344, 'timing/data': 0.0003651359584182501, 'timing/model': 1.1169998669065535, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.2}) +[2026-06-23 21:00:17,342][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.06537480652332306, 'train/grad_norm_pre_clip': 21.914459228515625, 'timing/data': 0.0005438840016722679, 'timing/model': 1.1149950879625976, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.2}) +[2026-06-23 21:00:21,423][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.07322414219379425, 'train/grad_norm_pre_clip': 35.99262237548828, 'timing/data': 0.0003558120224624872, 'timing/model': 1.1173117299331352, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.2}) +[2026-06-23 21:00:25,498][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.04723126441240311, 'train/grad_norm_pre_clip': 35.71505355834961, 'timing/data': 0.0006124059436842799, 'timing/model': 1.11323988402728, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.21}) +[2026-06-23 21:00:29,766][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.22993436455726624, 'train/grad_norm_pre_clip': 28.047382354736328, 'timing/data': 0.00035762006882578135, 'timing/model': 1.1221327729290351, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.21}) +[2026-06-23 21:00:33,850][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.09371661394834518, 'train/grad_norm_pre_clip': 23.215585708618164, 'timing/data': 0.000356729025952518, 'timing/model': 1.115686809993349, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.21}) +[2026-06-23 21:00:37,940][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.013778367079794407, 'train/grad_norm_pre_clip': 25.570465087890625, 'timing/data': 0.000409758067689836, 'timing/model': 1.1171925700036809, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.21}) +[2026-06-23 21:00:42,014][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.16528597474098206, 'train/grad_norm_pre_clip': 29.849899291992188, 'timing/data': 0.000399696989916265, 'timing/model': 1.1162552909227088, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.21}) +[2026-06-23 21:00:46,254][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.06080136075615883, 'train/grad_norm_pre_clip': 33.09616470336914, 'timing/data': 0.00040722195990383625, 'timing/model': 1.1174551870208234, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.21}) +[2026-06-23 21:00:50,332][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.1502382755279541, 'train/grad_norm_pre_clip': 50.92902374267578, 'timing/data': 0.00035285402555018663, 'timing/model': 1.118649901007302, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.21}) +[2026-06-23 21:00:54,410][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.1541249006986618, 'train/grad_norm_pre_clip': 37.9269905090332, 'timing/data': 0.0003602239303290844, 'timing/model': 1.1146264809649438, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.21}) +[2026-06-23 21:00:58,492][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.10843223333358765, 'train/grad_norm_pre_clip': 43.30190658569336, 'timing/data': 0.0006647480186074972, 'timing/model': 1.1137486998923123, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.21}) +[2026-06-23 21:01:02,733][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.06555256247520447, 'train/grad_norm_pre_clip': 17.08819007873535, 'timing/data': 0.000354256946593523, 'timing/model': 1.1172425849363208, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.21}) +[2026-06-23 21:01:06,810][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.11305392533540726, 'train/grad_norm_pre_clip': 21.177358627319336, 'timing/data': 0.00044562702532857656, 'timing/model': 1.1174225769937038, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.21}) +[2026-06-23 21:01:10,876][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.1766410917043686, 'train/grad_norm_pre_clip': 29.11690330505371, 'timing/data': 0.0003633740125223994, 'timing/model': 1.1186165029648691, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.21}) +[2026-06-23 21:01:14,962][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.12808342278003693, 'train/grad_norm_pre_clip': 16.154701232910156, 'timing/data': 0.00041500397492200136, 'timing/model': 1.11552788305562, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.21}) +[2026-06-23 21:01:19,195][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.07702304422855377, 'train/grad_norm_pre_clip': 17.695375442504883, 'timing/data': 0.0004048190312460065, 'timing/model': 1.1160155179677531, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.21}) +[2026-06-23 21:01:23,262][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.11005144566297531, 'train/grad_norm_pre_clip': 22.209781646728516, 'timing/data': 0.0003533479757606983, 'timing/model': 1.117599436081946, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.21}) +[2026-06-23 21:01:27,349][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.09870771318674088, 'train/grad_norm_pre_clip': 22.78434181213379, 'timing/data': 0.0006245090626180172, 'timing/model': 1.1169842750532553, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.21}) +[2026-06-23 21:01:31,424][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.17915493249893188, 'train/grad_norm_pre_clip': 37.85655212402344, 'timing/data': 0.0003623530501499772, 'timing/model': 1.1151843219995499, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.21}) +[2026-06-23 21:01:35,683][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.22862771153450012, 'train/grad_norm_pre_clip': 33.32646560668945, 'timing/data': 0.0006380280246958137, 'timing/model': 1.11462196696084, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.21}) +[2026-06-23 21:01:39,751][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.0529952272772789, 'train/grad_norm_pre_clip': 31.245445251464844, 'timing/data': 0.0003541590413078666, 'timing/model': 1.1113301169825718, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.21}) +[2026-06-23 21:01:43,830][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.1243225708603859, 'train/grad_norm_pre_clip': 31.27577018737793, 'timing/data': 0.0005779159255325794, 'timing/model': 1.11133864603471, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.21}) +[2026-06-23 21:01:47,906][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.007578815333545208, 'train/grad_norm_pre_clip': 22.41445541381836, 'timing/data': 0.0003788850735872984, 'timing/model': 1.1121043969178572, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.21}) +[2026-06-23 21:01:52,131][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.12242589145898819, 'train/grad_norm_pre_clip': 27.449050903320312, 'timing/data': 0.000493630999699235, 'timing/model': 1.1112478700233623, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.21}) +[2026-06-23 21:01:56,211][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.015638966113328934, 'train/grad_norm_pre_clip': 38.1269645690918, 'timing/data': 0.0005228329682722688, 'timing/model': 1.1155905979685485, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.21}) +[2026-06-23 21:02:00,292][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.12157651036977768, 'train/grad_norm_pre_clip': 37.687225341796875, 'timing/data': 0.0003629460697993636, 'timing/model': 1.1151023679412901, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.21}) +[2026-06-23 21:02:04,372][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.40592366456985474, 'train/grad_norm_pre_clip': 71.74056243896484, 'timing/data': 0.0005728710675612092, 'timing/model': 1.1164219590136781, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.21}) +[2026-06-23 21:02:08,656][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.17928217351436615, 'train/grad_norm_pre_clip': 81.56438446044922, 'timing/data': 0.00047332397662103176, 'timing/model': 1.1299884559120983, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.21}) +[2026-06-23 21:02:12,739][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.039881207048892975, 'train/grad_norm_pre_clip': 46.98532485961914, 'timing/data': 0.00036232301499694586, 'timing/model': 1.1180710030021146, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.22}) +[2026-06-23 21:02:16,812][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.08943930268287659, 'train/grad_norm_pre_clip': 21.428665161132812, 'timing/data': 0.0003968779928982258, 'timing/model': 1.1150550859747455, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.22}) +[2026-06-23 21:02:20,893][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.19474713504314423, 'train/grad_norm_pre_clip': 46.46170425415039, 'timing/data': 0.000362035003490746, 'timing/model': 1.1109583990182728, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.22}) +[2026-06-23 21:02:25,129][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.2942870557308197, 'train/grad_norm_pre_clip': 33.42234420776367, 'timing/data': 0.0003572100540623069, 'timing/model': 1.1123146719764918, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.22}) +[2026-06-23 21:02:29,207][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.18010737001895905, 'train/grad_norm_pre_clip': 62.31926727294922, 'timing/data': 0.0003592920256778598, 'timing/model': 1.1141347660450265, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.22}) +[2026-06-23 21:02:33,288][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.14970014989376068, 'train/grad_norm_pre_clip': 17.009979248046875, 'timing/data': 0.00036020390689373016, 'timing/model': 1.1205939520150423, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.22}) +[2026-06-23 21:02:37,373][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.06734993308782578, 'train/grad_norm_pre_clip': 34.729156494140625, 'timing/data': 0.00036072602961212397, 'timing/model': 1.1179804740240797, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.22}) +[2026-06-23 21:02:41,651][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.12187443673610687, 'train/grad_norm_pre_clip': 15.898150444030762, 'timing/data': 0.000517592066898942, 'timing/model': 1.1243474069051445, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.22}) +[2026-06-23 21:02:45,739][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.14567705988883972, 'train/grad_norm_pre_clip': 16.597490310668945, 'timing/data': 0.000385618070140481, 'timing/model': 1.1207931839162484, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.22}) +[2026-06-23 21:02:49,821][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.04559721052646637, 'train/grad_norm_pre_clip': 25.74232292175293, 'timing/data': 0.0006177820032462478, 'timing/model': 1.121807782095857, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.22}) +[2026-06-23 21:02:53,906][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.06594735383987427, 'train/grad_norm_pre_clip': 24.298992156982422, 'timing/data': 0.00036231300327926874, 'timing/model': 1.1168093120213598, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.22}) +[2026-06-23 21:02:58,154][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.09512900561094284, 'train/grad_norm_pre_clip': 21.669097900390625, 'timing/data': 0.00048533594235777855, 'timing/model': 1.1235268390737474, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.22}) +[2026-06-23 21:03:02,235][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.14971300959587097, 'train/grad_norm_pre_clip': 44.760986328125, 'timing/data': 0.00040083599742501974, 'timing/model': 1.11664629005827, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.22}) +[2026-06-23 21:03:06,319][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.11376939713954926, 'train/grad_norm_pre_clip': 43.423553466796875, 'timing/data': 0.00036718801129609346, 'timing/model': 1.1212642809841782, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.22}) +[2026-06-23 21:03:10,417][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.12061337381601334, 'train/grad_norm_pre_clip': 42.92486572265625, 'timing/data': 0.00036175502464175224, 'timing/model': 1.1267635530093685, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.22}) +[2026-06-23 21:03:14,701][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.056849200278520584, 'train/grad_norm_pre_clip': 48.04195785522461, 'timing/data': 0.0003667300334200263, 'timing/model': 1.1221768439281732, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.22}) +[2026-06-23 21:03:18,798][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.1146198958158493, 'train/grad_norm_pre_clip': 28.464622497558594, 'timing/data': 0.0006455649854615331, 'timing/model': 1.1212873009499162, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.22}) +[2026-06-23 21:03:22,896][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.08935346454381943, 'train/grad_norm_pre_clip': 27.85870933532715, 'timing/data': 0.0003731389297172427, 'timing/model': 1.1224418290657923, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.22}) +[2026-06-23 21:03:27,002][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.062074895948171616, 'train/grad_norm_pre_clip': 28.866880416870117, 'timing/data': 0.0003551200497895479, 'timing/model': 1.1298665259964764, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.22}) +[2026-06-23 21:03:31,266][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.06143094599246979, 'train/grad_norm_pre_clip': 46.30282974243164, 'timing/data': 0.0004015360027551651, 'timing/model': 1.116626178030856, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.22}) +[2026-06-23 21:03:35,354][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.1962508261203766, 'train/grad_norm_pre_clip': 27.157194137573242, 'timing/data': 0.00035657407715916634, 'timing/model': 1.118264924036339, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.22}) +[2026-06-23 21:03:39,435][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.035953205078840256, 'train/grad_norm_pre_clip': 59.54686737060547, 'timing/data': 0.0004217979731038213, 'timing/model': 1.11465965595562, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.22}) +[2026-06-23 21:03:43,530][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.09797888994216919, 'train/grad_norm_pre_clip': 22.533662796020508, 'timing/data': 0.0003596110036596656, 'timing/model': 1.1182332360185683, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.22}) +[2026-06-23 21:03:47,763][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.05626959353685379, 'train/grad_norm_pre_clip': 37.183349609375, 'timing/data': 0.0003508670488372445, 'timing/model': 1.112547260010615, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.22}) +[2026-06-23 21:03:51,851][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.06311400234699249, 'train/grad_norm_pre_clip': 23.12558364868164, 'timing/data': 0.0003956869477406144, 'timing/model': 1.1297833389835432, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.22}) +[2026-06-23 21:03:55,939][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.17296528816223145, 'train/grad_norm_pre_clip': 25.230716705322266, 'timing/data': 0.0003608099650591612, 'timing/model': 1.1212392200250179, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.22}) +[2026-06-23 21:04:00,018][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.15047712624073029, 'train/grad_norm_pre_clip': 54.87871170043945, 'timing/data': 0.00035774800926446915, 'timing/model': 1.1244813209632412, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.23}) +[2026-06-23 21:04:04,283][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.2320331633090973, 'train/grad_norm_pre_clip': 27.82530403137207, 'timing/data': 0.0003795779775828123, 'timing/model': 1.127582799992524, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.23}) +[2026-06-23 21:04:08,371][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.09165063500404358, 'train/grad_norm_pre_clip': 18.68029022216797, 'timing/data': 0.0003817479591816664, 'timing/model': 1.1233025370165706, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.23}) +[2026-06-23 21:04:12,445][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.16089864075183868, 'train/grad_norm_pre_clip': 35.32095718383789, 'timing/data': 0.0003829570487141609, 'timing/model': 1.1174761999864131, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.23}) +[2026-06-23 21:04:16,532][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.05397573858499527, 'train/grad_norm_pre_clip': 12.551618576049805, 'timing/data': 0.0003608099650591612, 'timing/model': 1.1204553920542821, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.23}) +[2026-06-23 21:04:20,780][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.10726109892129898, 'train/grad_norm_pre_clip': 39.13087463378906, 'timing/data': 0.0003544529899954796, 'timing/model': 1.1153422879287973, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.23}) +[2026-06-23 21:04:24,863][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.05531051382422447, 'train/grad_norm_pre_clip': 29.165145874023438, 'timing/data': 0.00034543301444500685, 'timing/model': 1.1176786069991067, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.23}) +[2026-06-23 21:04:28,962][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.4232886731624603, 'train/grad_norm_pre_clip': 22.727848052978516, 'timing/data': 0.000359719037078321, 'timing/model': 1.1302907130448148, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.23}) +[2026-06-23 21:04:33,045][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.2941794693470001, 'train/grad_norm_pre_clip': 50.19121551513672, 'timing/data': 0.0003686969866976142, 'timing/model': 1.1195339519763365, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.23}) +[2026-06-23 21:04:37,313][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.11628121137619019, 'train/grad_norm_pre_clip': 30.905794143676758, 'timing/data': 0.00036249507684260607, 'timing/model': 1.114932224038057, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.23}) +[2026-06-23 21:04:41,394][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.04270675778388977, 'train/grad_norm_pre_clip': 22.459569931030273, 'timing/data': 0.0006846690084785223, 'timing/model': 1.115297892014496, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.23}) +[2026-06-23 21:04:45,488][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.01480846293270588, 'train/grad_norm_pre_clip': 19.83638572692871, 'timing/data': 0.0003589258994907141, 'timing/model': 1.1156618199311197, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.23}) +[2026-06-23 21:04:49,561][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.27591145038604736, 'train/grad_norm_pre_clip': 56.0109977722168, 'timing/data': 0.0003747820155695081, 'timing/model': 1.1173147149384022, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.23}) +[2026-06-23 21:04:53,818][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.07883592694997787, 'train/grad_norm_pre_clip': 25.52490997314453, 'timing/data': 0.00035752891562879086, 'timing/model': 1.1242896319599822, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.23}) +[2026-06-23 21:04:57,892][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.01477678120136261, 'train/grad_norm_pre_clip': 19.199377059936523, 'timing/data': 0.0004914510063827038, 'timing/model': 1.1128216949291527, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.23}) +[2026-06-23 21:05:01,974][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.049353666603565216, 'train/grad_norm_pre_clip': 27.11502456665039, 'timing/data': 0.00036990002263337374, 'timing/model': 1.1129079420352355, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.23}) +[2026-06-23 21:05:06,044][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.07002493739128113, 'train/grad_norm_pre_clip': 24.73933219909668, 'timing/data': 0.00041720899753272533, 'timing/model': 1.1139109480427578, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.23}) +[2026-06-23 21:05:10,311][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.09925175458192825, 'train/grad_norm_pre_clip': 16.857519149780273, 'timing/data': 0.0005837930366396904, 'timing/model': 1.1204766390146688, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.23}) +[2026-06-23 21:05:14,399][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 600 +[2026-06-23 21:06:19,265][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 600 +[2026-06-23 21:06:19,267][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 600 +[2026-06-23 21:06:57,760][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 600 +[2026-06-23 21:06:57,761][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.3608516752719879, 'train/grad_norm_pre_clip': 22.472084045410156, 'eval/action_loss/seconds': 64.8637658850057, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.08843491226434708, 'eval/latency_2/loss': 0.08843491226434708, 'eval/flappy/loss': 0.08843491226434708, 'eval/flappy/latency_2/loss': 0.08843491226434708, 'eval/flap_precision': 0.9842271293375394, 'eval/flap_recall': 0.8466757123473542, 'eval/flap_f1': 0.9102844638949672, 'eval/flap_rate_model': 0.1761111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9102844638949672, 'eval/latency_2/flap_precision': 0.9842271293375394, 'eval/latency_2/flap_recall': 0.8466757123473542, 'eval/latency_2/flap_f1': 0.9102844638949672, 'eval/latency_2/flap_rate_model': 0.1761111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9102844638949672, 'eval/action_classification/data_wait_seconds': 0.4645358920097351, 'eval/action_classification/predict_seconds': 37.88266372680664, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0005014949711039662, 'timing/model': 1.1247714250348508, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.23}) +[2026-06-23 21:07:01,819][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.0324273481965065, 'train/grad_norm_pre_clip': 49.53618240356445, 'timing/data': 0.0005606099730357528, 'timing/model': 1.1127097370335832, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.23}) +[2026-06-23 21:07:06,032][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.16583427786827087, 'train/grad_norm_pre_clip': 63.842655181884766, 'timing/data': 0.00036702200304716825, 'timing/model': 1.2541908930288628, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.23}) +[2026-06-23 21:07:10,083][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.15789727866649628, 'train/grad_norm_pre_clip': 23.797922134399414, 'timing/data': 0.0005295039154589176, 'timing/model': 1.1112983439816162, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.23}) +[2026-06-23 21:07:14,130][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.12043201178312302, 'train/grad_norm_pre_clip': 41.33850860595703, 'timing/data': 0.00046661996748298407, 'timing/model': 1.1062639540759847, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.23}) +[2026-06-23 21:07:18,184][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.18808290362358093, 'train/grad_norm_pre_clip': 54.571937561035156, 'timing/data': 0.0005360880168154836, 'timing/model': 1.1119145760312676, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.23}) +[2026-06-23 21:07:22,390][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.11160396784543991, 'train/grad_norm_pre_clip': 26.20629119873047, 'timing/data': 0.000364303938113153, 'timing/model': 1.2600025099236518, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.23}) +[2026-06-23 21:07:26,445][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.0169750414788723, 'train/grad_norm_pre_clip': 21.207433700561523, 'timing/data': 0.00037953106220811605, 'timing/model': 1.1115939390147105, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.23}) +[2026-06-23 21:07:30,509][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.2783987522125244, 'train/grad_norm_pre_clip': 18.30347442626953, 'timing/data': 0.00036263593938201666, 'timing/model': 1.1145537389675155, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.24}) +[2026-06-23 21:07:34,593][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.10035760700702667, 'train/grad_norm_pre_clip': 22.11907196044922, 'timing/data': 0.0003544120118021965, 'timing/model': 1.1240257419412956, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.24}) +[2026-06-23 21:07:38,794][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.11488232761621475, 'train/grad_norm_pre_clip': 29.709270477294922, 'timing/data': 0.00039308995474129915, 'timing/model': 1.2566870889859274, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.24}) +[2026-06-23 21:07:42,868][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.13352420926094055, 'train/grad_norm_pre_clip': 25.668081283569336, 'timing/data': 0.00063244195189327, 'timing/model': 1.1135900489753112, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.24}) +[2026-06-23 21:07:46,934][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.03185906261205673, 'train/grad_norm_pre_clip': 14.95849323272705, 'timing/data': 0.0003557959571480751, 'timing/model': 1.1162651700433344, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.24}) +[2026-06-23 21:07:51,007][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.1730116903781891, 'train/grad_norm_pre_clip': 14.728484153747559, 'timing/data': 0.0004183750133961439, 'timing/model': 1.1200281620258465, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.24}) +[2026-06-23 21:07:55,224][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.08594075590372086, 'train/grad_norm_pre_clip': 13.590415954589844, 'timing/data': 0.00035906198900192976, 'timing/model': 1.2619346890132874, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.24}) +[2026-06-23 21:07:59,293][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.14685387909412384, 'train/grad_norm_pre_clip': 22.953704833984375, 'timing/data': 0.00037698401138186455, 'timing/model': 1.1199027060065418, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.24}) +[2026-06-23 21:08:03,370][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.10207515209913254, 'train/grad_norm_pre_clip': 22.45347785949707, 'timing/data': 0.00036031706258654594, 'timing/model': 1.1211542499950156, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.24}) +[2026-06-23 21:08:07,450][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.2531605064868927, 'train/grad_norm_pre_clip': 38.45705032348633, 'timing/data': 0.00035975396167486906, 'timing/model': 1.10954040591605, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.24}) +[2026-06-23 21:08:11,678][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.17174306511878967, 'train/grad_norm_pre_clip': 15.796299934387207, 'timing/data': 0.00040497409645467997, 'timing/model': 1.2702212899457663, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.24}) +[2026-06-23 21:08:15,760][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.01653633452951908, 'train/grad_norm_pre_clip': 12.062536239624023, 'timing/data': 0.0005914469948038459, 'timing/model': 1.1187428960110992, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.24}) +[2026-06-23 21:08:19,831][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.048473045229911804, 'train/grad_norm_pre_clip': 25.686214447021484, 'timing/data': 0.0003569300752133131, 'timing/model': 1.111760608968325, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.24}) +[2026-06-23 21:08:23,922][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.07974247634410858, 'train/grad_norm_pre_clip': 20.348812103271484, 'timing/data': 0.0004933120217174292, 'timing/model': 1.1197990969521925, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.24}) +[2026-06-23 21:08:28,143][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.051483046263456345, 'train/grad_norm_pre_clip': 36.8732795715332, 'timing/data': 0.0003673100145533681, 'timing/model': 1.2682045940309763, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.24}) +[2026-06-23 21:08:32,214][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.06925313174724579, 'train/grad_norm_pre_clip': 37.2921028137207, 'timing/data': 0.00044630502816289663, 'timing/model': 1.1136097679845989, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.24}) +[2026-06-23 21:08:36,279][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.13763698935508728, 'train/grad_norm_pre_clip': 47.86955261230469, 'timing/data': 0.00048247305676341057, 'timing/model': 1.1140390869695693, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.24}) +[2026-06-23 21:08:40,357][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.17164158821105957, 'train/grad_norm_pre_clip': 58.1229133605957, 'timing/data': 0.00036153197288513184, 'timing/model': 1.1164839810226113, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.24}) +[2026-06-23 21:08:44,592][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.27126842737197876, 'train/grad_norm_pre_clip': 32.15928649902344, 'timing/data': 0.00040341203566640615, 'timing/model': 1.2763115159468725, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.24}) +[2026-06-23 21:08:48,670][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.057730671018362045, 'train/grad_norm_pre_clip': 9.616212844848633, 'timing/data': 0.0003893940011039376, 'timing/model': 1.1172826300607994, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.24}) +[2026-06-23 21:08:52,749][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.13514022529125214, 'train/grad_norm_pre_clip': 22.0289306640625, 'timing/data': 0.00035456300247460604, 'timing/model': 1.1171225359430537, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.24}) +[2026-06-23 21:08:56,829][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.035341668874025345, 'train/grad_norm_pre_clip': 20.9381103515625, 'timing/data': 0.000538803986273706, 'timing/model': 1.120620802976191, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.24}) +[2026-06-23 21:09:01,057][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.1426284909248352, 'train/grad_norm_pre_clip': 28.046585083007812, 'timing/data': 0.00036623794585466385, 'timing/model': 1.263786467956379, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.24}) +[2026-06-23 21:09:05,143][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.26361149549484253, 'train/grad_norm_pre_clip': 33.35639190673828, 'timing/data': 0.00041656801477074623, 'timing/model': 1.1205766850616783, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.24}) +[2026-06-23 21:09:09,222][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.17926421761512756, 'train/grad_norm_pre_clip': 21.668697357177734, 'timing/data': 0.0003585210070014, 'timing/model': 1.1178844289388508, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.24}) +[2026-06-23 21:09:13,314][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.08346012234687805, 'train/grad_norm_pre_clip': 24.893192291259766, 'timing/data': 0.0003615879686549306, 'timing/model': 1.1158336609369144, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.24}) +[2026-06-23 21:09:17,556][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.0740736871957779, 'train/grad_norm_pre_clip': 14.237764358520508, 'timing/data': 0.0003675309708341956, 'timing/model': 1.27018907899037, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.25}) +[2026-06-23 21:09:21,638][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.14903691411018372, 'train/grad_norm_pre_clip': 44.98003387451172, 'timing/data': 0.00035143597051501274, 'timing/model': 1.11891232104972, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.25}) +[2026-06-23 21:09:25,735][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.050203319638967514, 'train/grad_norm_pre_clip': 14.162044525146484, 'timing/data': 0.00049351598136127, 'timing/model': 1.1191722099902108, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.25}) +[2026-06-23 21:09:29,828][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.033166225999593735, 'train/grad_norm_pre_clip': 24.399513244628906, 'timing/data': 0.0005642930045723915, 'timing/model': 1.1166364150121808, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.25}) +[2026-06-23 21:09:34,087][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.03108135797083378, 'train/grad_norm_pre_clip': 18.681886672973633, 'timing/data': 0.0003550780238583684, 'timing/model': 1.2848301670746878, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.25}) +[2026-06-23 21:09:38,169][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.07707677781581879, 'train/grad_norm_pre_clip': 22.64965057373047, 'timing/data': 0.00035643402952700853, 'timing/model': 1.1213581890333444, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.25}) +[2026-06-23 21:09:42,258][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.3947980999946594, 'train/grad_norm_pre_clip': 57.04728698730469, 'timing/data': 0.00036612001713365316, 'timing/model': 1.1194488959154114, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.25}) +[2026-06-23 21:09:46,341][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.2643398344516754, 'train/grad_norm_pre_clip': 81.29840087890625, 'timing/data': 0.00035595393273979425, 'timing/model': 1.1157071379711851, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.25}) +[2026-06-23 21:09:50,554][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.049502283334732056, 'train/grad_norm_pre_clip': 31.722318649291992, 'timing/data': 0.00037216697819530964, 'timing/model': 1.2581869860878214, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.25}) +[2026-06-23 21:09:54,631][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.16662658751010895, 'train/grad_norm_pre_clip': 84.56075286865234, 'timing/data': 0.00040268199518322945, 'timing/model': 1.1175488909939304, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.25}) +[2026-06-23 21:09:58,710][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.09058866649866104, 'train/grad_norm_pre_clip': 68.32987213134766, 'timing/data': 0.0005187990609556437, 'timing/model': 1.1206583878956735, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.25}) +[2026-06-23 21:10:02,785][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.16101519763469696, 'train/grad_norm_pre_clip': 46.9800910949707, 'timing/data': 0.0005238220328465104, 'timing/model': 1.1162266000173986, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.25}) +[2026-06-23 21:10:07,012][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.007181116379797459, 'train/grad_norm_pre_clip': 55.42694091796875, 'timing/data': 0.0004356940044090152, 'timing/model': 1.263951917993836, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.25}) +[2026-06-23 21:10:11,099][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.19331882894039154, 'train/grad_norm_pre_clip': 37.87423324584961, 'timing/data': 0.0005317389732226729, 'timing/model': 1.1197096490068361, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.25}) +[2026-06-23 21:10:15,178][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.33401253819465637, 'train/grad_norm_pre_clip': 78.54518127441406, 'timing/data': 0.00043676106724888086, 'timing/model': 1.1200651279650629, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.25}) +[2026-06-23 21:10:19,253][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.31341955065727234, 'train/grad_norm_pre_clip': 62.93475341796875, 'timing/data': 0.0005505499430000782, 'timing/model': 1.1173853719374165, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.25}) +[2026-06-23 21:10:23,479][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.1869206428527832, 'train/grad_norm_pre_clip': 45.10823440551758, 'timing/data': 0.0005420229863375425, 'timing/model': 1.2625454029766843, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.25}) +[2026-06-23 21:10:27,556][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.2273087203502655, 'train/grad_norm_pre_clip': 59.567020416259766, 'timing/data': 0.00035234796814620495, 'timing/model': 1.116233455017209, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.25}) +[2026-06-23 21:10:31,661][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.16069042682647705, 'train/grad_norm_pre_clip': 21.309755325317383, 'timing/data': 0.00041817896999418736, 'timing/model': 1.1167282710084692, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.25}) +[2026-06-23 21:10:35,726][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.11504888534545898, 'train/grad_norm_pre_clip': 18.61397933959961, 'timing/data': 0.0004091039299964905, 'timing/model': 1.1128740599378943, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.25}) +[2026-06-23 21:10:39,947][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.057155486196279526, 'train/grad_norm_pre_clip': 22.18907356262207, 'timing/data': 0.000401182915084064, 'timing/model': 1.259779985062778, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.25}) +[2026-06-23 21:10:44,024][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.10368208587169647, 'train/grad_norm_pre_clip': 26.630321502685547, 'timing/data': 0.0005379380891099572, 'timing/model': 1.1187455679755658, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.25}) +[2026-06-23 21:10:48,104][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.21977859735488892, 'train/grad_norm_pre_clip': 17.56780242919922, 'timing/data': 0.00044552807230502367, 'timing/model': 1.1192504840437323, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.25}) +[2026-06-23 21:10:52,188][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.08246954530477524, 'train/grad_norm_pre_clip': 25.291967391967773, 'timing/data': 0.00042739405762404203, 'timing/model': 1.118683662964031, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.25}) +[2026-06-23 21:10:56,420][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.186737522482872, 'train/grad_norm_pre_clip': 19.382862091064453, 'timing/data': 0.0003659309586510062, 'timing/model': 1.2725461319787428, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.25}) +[2026-06-23 21:11:00,516][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.09803009033203125, 'train/grad_norm_pre_clip': 23.14008903503418, 'timing/data': 0.000356137054041028, 'timing/model': 1.12211464100983, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.26}) +[2026-06-23 21:11:04,619][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.21635277569293976, 'train/grad_norm_pre_clip': 15.628314018249512, 'timing/data': 0.00037632801104336977, 'timing/model': 1.122721015010029, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.26}) +[2026-06-23 21:11:08,705][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.1590254008769989, 'train/grad_norm_pre_clip': 20.65252685546875, 'timing/data': 0.0003657640190795064, 'timing/model': 1.120166023960337, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.26}) +[2026-06-23 21:11:12,944][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.12430977821350098, 'train/grad_norm_pre_clip': 22.636131286621094, 'timing/data': 0.0004994369810447097, 'timing/model': 1.2740733940154314, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.26}) +[2026-06-23 21:11:17,014][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.1975121945142746, 'train/grad_norm_pre_clip': 32.737937927246094, 'timing/data': 0.0006207360420376062, 'timing/model': 1.122359668952413, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.26}) +[2026-06-23 21:11:21,103][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.1363990604877472, 'train/grad_norm_pre_clip': 29.33282470703125, 'timing/data': 0.00035619991831481457, 'timing/model': 1.1179302770178765, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.26}) +[2026-06-23 21:11:25,186][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.2148112952709198, 'train/grad_norm_pre_clip': 36.87641525268555, 'timing/data': 0.00036031194031238556, 'timing/model': 1.123114412999712, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.26}) +[2026-06-23 21:11:29,429][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.10894394665956497, 'train/grad_norm_pre_clip': 62.423824310302734, 'timing/data': 0.00035443296656012535, 'timing/model': 1.2736143310321495, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.26}) +[2026-06-23 21:11:33,511][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.21598920226097107, 'train/grad_norm_pre_clip': 82.62919616699219, 'timing/data': 0.00038752693217247725, 'timing/model': 1.1183006070787087, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.26}) +[2026-06-23 21:11:37,606][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.07759294658899307, 'train/grad_norm_pre_clip': 63.758934020996094, 'timing/data': 0.0003542680060490966, 'timing/model': 1.1183677320368588, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.26}) +[2026-06-23 21:11:41,753][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.042856499552726746, 'train/grad_norm_pre_clip': 84.00342559814453, 'timing/data': 0.00035472295712679625, 'timing/model': 1.1195336669916287, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.26}) +[2026-06-23 21:11:45,981][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.17560043931007385, 'train/grad_norm_pre_clip': 26.174427032470703, 'timing/data': 0.00039852794725447893, 'timing/model': 1.2576127840438858, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.26}) +[2026-06-23 21:11:50,062][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.1294572651386261, 'train/grad_norm_pre_clip': 97.85044860839844, 'timing/data': 0.0005280290497466922, 'timing/model': 1.1278867570217699, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.26}) +[2026-06-23 21:11:54,141][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.3627004623413086, 'train/grad_norm_pre_clip': 67.7953109741211, 'timing/data': 0.0004100729711353779, 'timing/model': 1.1155101179610938, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.26}) +[2026-06-23 21:11:58,224][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.11339126527309418, 'train/grad_norm_pre_clip': 72.42840576171875, 'timing/data': 0.0004713769303634763, 'timing/model': 1.1182419969700277, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.26}) +[2026-06-23 21:12:02,467][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.13921548426151276, 'train/grad_norm_pre_clip': 37.997188568115234, 'timing/data': 0.0004043889930471778, 'timing/model': 1.2707082129782066, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.26}) +[2026-06-23 21:12:06,555][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.25752317905426025, 'train/grad_norm_pre_clip': 43.3434944152832, 'timing/data': 0.0005538580007851124, 'timing/model': 1.117846614914015, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.26}) +[2026-06-23 21:12:10,640][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.20834799110889435, 'train/grad_norm_pre_clip': 42.11079406738281, 'timing/data': 0.0003716200590133667, 'timing/model': 1.119416982983239, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.26}) +[2026-06-23 21:12:14,726][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.14031797647476196, 'train/grad_norm_pre_clip': 30.421903610229492, 'timing/data': 0.0003688220167532563, 'timing/model': 1.1169291029218584, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.26}) +[2026-06-23 21:12:18,958][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.13263358175754547, 'train/grad_norm_pre_clip': 42.477508544921875, 'timing/data': 0.00037916505243629217, 'timing/model': 1.2648445069789886, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.26}) +[2026-06-23 21:12:23,043][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.08766569942235947, 'train/grad_norm_pre_clip': 22.09904670715332, 'timing/data': 0.00035645696334540844, 'timing/model': 1.1206856910139322, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.26}) +[2026-06-23 21:12:27,133][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.10316821187734604, 'train/grad_norm_pre_clip': 30.301437377929688, 'timing/data': 0.0004245940363034606, 'timing/model': 1.1236749209929258, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.26}) +[2026-06-23 21:12:31,221][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.09640450775623322, 'train/grad_norm_pre_clip': 23.67848014831543, 'timing/data': 0.00042557495180517435, 'timing/model': 1.1188505500322208, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.26}) +[2026-06-23 21:12:35,469][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.11177803575992584, 'train/grad_norm_pre_clip': 32.043609619140625, 'timing/data': 0.000523892929777503, 'timing/model': 1.2849272299790755, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.26}) +[2026-06-23 21:12:39,557][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.13690508902072906, 'train/grad_norm_pre_clip': 18.16518211364746, 'timing/data': 0.0003566059749573469, 'timing/model': 1.1228505339240655, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.26}) +[2026-06-23 21:12:43,652][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.23708388209342957, 'train/grad_norm_pre_clip': 24.781816482543945, 'timing/data': 0.0003543290076777339, 'timing/model': 1.124308196012862, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.26}) +[2026-06-23 21:12:47,738][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.2574341893196106, 'train/grad_norm_pre_clip': 35.726593017578125, 'timing/data': 0.0003815310774371028, 'timing/model': 1.122661042958498, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.27}) +[2026-06-23 21:12:51,970][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.1436365693807602, 'train/grad_norm_pre_clip': 34.31428527832031, 'timing/data': 0.0003555330913513899, 'timing/model': 1.27524682902731, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.27}) +[2026-06-23 21:12:56,062][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.08570992201566696, 'train/grad_norm_pre_clip': 19.952295303344727, 'timing/data': 0.0003558649914339185, 'timing/model': 1.1270897960057482, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.27}) +[2026-06-23 21:13:00,148][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.17340022325515747, 'train/grad_norm_pre_clip': 21.86652183532715, 'timing/data': 0.0003541409969329834, 'timing/model': 1.1184302370529622, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.27}) +[2026-06-23 21:13:04,239][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.06801296770572662, 'train/grad_norm_pre_clip': 33.664180755615234, 'timing/data': 0.0005663150222972035, 'timing/model': 1.117116182926111, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.27}) +[2026-06-23 21:13:08,482][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.03532600402832031, 'train/grad_norm_pre_clip': 26.800212860107422, 'timing/data': 0.0003670849837362766, 'timing/model': 1.2796869419980794, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.27}) +[2026-06-23 21:13:12,561][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.0801476240158081, 'train/grad_norm_pre_clip': 15.804120063781738, 'timing/data': 0.0006685450207442045, 'timing/model': 1.11472686403431, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.27}) +[2026-06-23 21:13:16,646][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.09852450340986252, 'train/grad_norm_pre_clip': 20.339988708496094, 'timing/data': 0.00040033902041614056, 'timing/model': 1.118473566020839, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.27}) +[2026-06-23 21:13:20,730][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.19383692741394043, 'train/grad_norm_pre_clip': 23.329565048217773, 'timing/data': 0.00048393593169748783, 'timing/model': 1.1231104360194877, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.27}) +[2026-06-23 21:13:24,973][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.060841552913188934, 'train/grad_norm_pre_clip': 14.185185432434082, 'timing/data': 0.00041692506056278944, 'timing/model': 1.2743110730079934, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.27}) +[2026-06-23 21:13:29,047][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.030159717425704002, 'train/grad_norm_pre_clip': 28.80990219116211, 'timing/data': 0.00036171404644846916, 'timing/model': 1.1133670180570334, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.27}) +[2026-06-23 21:13:33,129][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.0896187573671341, 'train/grad_norm_pre_clip': 16.01837158203125, 'timing/data': 0.00040406198240816593, 'timing/model': 1.116392762050964, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.27}) +[2026-06-23 21:13:37,217][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.11718624830245972, 'train/grad_norm_pre_clip': 14.658674240112305, 'timing/data': 0.0003836769610643387, 'timing/model': 1.1185832679038867, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.27}) +[2026-06-23 21:13:41,444][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.053824227303266525, 'train/grad_norm_pre_clip': 16.484222412109375, 'timing/data': 0.0003564979415386915, 'timing/model': 1.2636764390626922, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.27}) +[2026-06-23 21:13:45,517][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.15748852491378784, 'train/grad_norm_pre_clip': 21.417221069335938, 'timing/data': 0.0003935860004276037, 'timing/model': 1.118493961985223, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.27}) +[2026-06-23 21:13:49,601][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 700 +[2026-06-23 21:14:54,005][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 700 +[2026-06-23 21:14:54,007][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 700 +[2026-06-23 21:15:33,013][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 700 +[2026-06-23 21:15:33,014][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.05600608512759209, 'train/grad_norm_pre_clip': 20.043476104736328, 'eval/action_loss/seconds': 64.38750179903582, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.06332497298717499, 'eval/latency_2/loss': 0.06332497298717499, 'eval/flappy/loss': 0.06332497298717499, 'eval/flappy/latency_2/loss': 0.06332497298717499, 'eval/flap_precision': 0.9643874643874644, 'eval/flap_recall': 0.9185888738127544, 'eval/flap_f1': 0.9409312022237665, 'eval/flap_rate_model': 0.195, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9409312022237665, 'eval/latency_2/flap_precision': 0.9643874643874644, 'eval/latency_2/flap_recall': 0.9185888738127544, 'eval/latency_2/flap_f1': 0.9409312022237665, 'eval/latency_2/flap_rate_model': 0.195, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9409312022237665, 'eval/action_classification/data_wait_seconds': 0.4751415550708771, 'eval/action_classification/predict_seconds': 38.40077590942383, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0003664690302684903, 'timing/model': 1.1204655119217932, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.27}) +[2026-06-23 21:15:37,060][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.0813276544213295, 'train/grad_norm_pre_clip': 21.165834426879883, 'timing/data': 0.0003853470552712679, 'timing/model': 1.1107347749639302, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.27}) +[2026-06-23 21:15:41,240][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.13784395158290863, 'train/grad_norm_pre_clip': 45.737430572509766, 'timing/data': 0.0003594830632209778, 'timing/model': 1.1104879319900647, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.27}) +[2026-06-23 21:15:45,296][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.1426921933889389, 'train/grad_norm_pre_clip': 41.21345138549805, 'timing/data': 0.0003692179452627897, 'timing/model': 1.106931207003072, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.27}) +[2026-06-23 21:15:49,360][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.19110995531082153, 'train/grad_norm_pre_clip': 20.8824462890625, 'timing/data': 0.0005051000043749809, 'timing/model': 1.1155528570525348, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.27}) +[2026-06-23 21:15:53,421][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.08129043877124786, 'train/grad_norm_pre_clip': 44.57023620605469, 'timing/data': 0.0005474679637700319, 'timing/model': 1.1152117080055177, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.27}) +[2026-06-23 21:15:57,605][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.11985788494348526, 'train/grad_norm_pre_clip': 30.525999069213867, 'timing/data': 0.0004862330388277769, 'timing/model': 1.106693857931532, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.27}) +[2026-06-23 21:16:01,658][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.04818448796868324, 'train/grad_norm_pre_clip': 47.49264907836914, 'timing/data': 0.0004058720078319311, 'timing/model': 1.1072206780081615, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.27}) +[2026-06-23 21:16:05,718][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.06428521126508713, 'train/grad_norm_pre_clip': 25.045547485351562, 'timing/data': 0.0005629090592265129, 'timing/model': 1.111575128044933, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.27}) +[2026-06-23 21:16:09,798][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.2782411575317383, 'train/grad_norm_pre_clip': 66.01677703857422, 'timing/data': 0.0003913069376721978, 'timing/model': 1.1263797039864585, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.27}) +[2026-06-23 21:16:14,017][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.3255504071712494, 'train/grad_norm_pre_clip': 79.68611145019531, 'timing/data': 0.00035553902853280306, 'timing/model': 1.1233016460901126, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.27}) +[2026-06-23 21:16:18,077][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.2284349948167801, 'train/grad_norm_pre_clip': 34.50064468383789, 'timing/data': 0.0004100920632481575, 'timing/model': 1.115418556961231, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.28}) +[2026-06-23 21:16:22,124][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.22439029812812805, 'train/grad_norm_pre_clip': 72.39397430419922, 'timing/data': 0.0004405659856274724, 'timing/model': 1.106254318030551, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.28}) +[2026-06-23 21:16:26,185][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.12224338948726654, 'train/grad_norm_pre_clip': 65.81438446044922, 'timing/data': 0.0005017051007598639, 'timing/model': 1.1110620719846338, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.28}) +[2026-06-23 21:16:30,377][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.028357017785310745, 'train/grad_norm_pre_clip': 28.18808364868164, 'timing/data': 0.00034624896943569183, 'timing/model': 1.1136205700458959, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.28}) +[2026-06-23 21:16:34,441][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.09612570703029633, 'train/grad_norm_pre_clip': 33.47924041748047, 'timing/data': 0.0005377589259296656, 'timing/model': 1.1137977490434423, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.28}) +[2026-06-23 21:16:38,526][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.16806575655937195, 'train/grad_norm_pre_clip': 20.252286911010742, 'timing/data': 0.00040286104194819927, 'timing/model': 1.1176362170372158, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.28}) +[2026-06-23 21:16:42,596][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.24573107063770294, 'train/grad_norm_pre_clip': 32.858001708984375, 'timing/data': 0.0003591689746826887, 'timing/model': 1.1195740950061008, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.28}) +[2026-06-23 21:16:46,794][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.34042298793792725, 'train/grad_norm_pre_clip': 43.704315185546875, 'timing/data': 0.0005577750271186233, 'timing/model': 1.1136902190046385, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.28}) +[2026-06-23 21:16:50,897][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.17721185088157654, 'train/grad_norm_pre_clip': 44.67662048339844, 'timing/data': 0.0003535440191626549, 'timing/model': 1.1178825929528102, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.28}) +[2026-06-23 21:16:54,972][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.12763188779354095, 'train/grad_norm_pre_clip': 12.467872619628906, 'timing/data': 0.0003686490235850215, 'timing/model': 1.1114500530529767, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.28}) +[2026-06-23 21:16:59,047][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.08595096319913864, 'train/grad_norm_pre_clip': 38.90372848510742, 'timing/data': 0.0004894249141216278, 'timing/model': 1.1109305339632556, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.28}) +[2026-06-23 21:17:03,275][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.15638615190982819, 'train/grad_norm_pre_clip': 35.50385284423828, 'timing/data': 0.0004564429400488734, 'timing/model': 1.1327958969632164, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.28}) +[2026-06-23 21:17:07,362][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.08260177820920944, 'train/grad_norm_pre_clip': 13.056702613830566, 'timing/data': 0.00034862698521465063, 'timing/model': 1.1214047949761152, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.28}) +[2026-06-23 21:17:11,427][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.04364565759897232, 'train/grad_norm_pre_clip': 16.544570922851562, 'timing/data': 0.000592909986153245, 'timing/model': 1.112355470075272, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.28}) +[2026-06-23 21:17:15,501][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.143598735332489, 'train/grad_norm_pre_clip': 30.18397331237793, 'timing/data': 0.00039686798118054867, 'timing/model': 1.1154663190245628, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.28}) +[2026-06-23 21:17:19,702][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.12817542254924774, 'train/grad_norm_pre_clip': 26.177873611450195, 'timing/data': 0.0004159320378676057, 'timing/model': 1.109422664041631, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.28}) +[2026-06-23 21:17:23,759][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.059836361557245255, 'train/grad_norm_pre_clip': 28.748291015625, 'timing/data': 0.00047701294533908367, 'timing/model': 1.1154074090300128, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.28}) +[2026-06-23 21:17:27,821][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.28349366784095764, 'train/grad_norm_pre_clip': 35.650909423828125, 'timing/data': 0.00040209293365478516, 'timing/model': 1.11569689004682, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.28}) +[2026-06-23 21:17:31,899][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.14466026425361633, 'train/grad_norm_pre_clip': 23.62440299987793, 'timing/data': 0.000420138007029891, 'timing/model': 1.1162765909684822, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.28}) +[2026-06-23 21:17:36,098][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.06633973121643066, 'train/grad_norm_pre_clip': 15.039525985717773, 'timing/data': 0.00042202090844511986, 'timing/model': 1.1138985219877213, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.28}) +[2026-06-23 21:17:40,192][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.10033834725618362, 'train/grad_norm_pre_clip': 42.09907531738281, 'timing/data': 0.0004196680383756757, 'timing/model': 1.1284516870509833, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.28}) +[2026-06-23 21:17:44,269][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.1455647200345993, 'train/grad_norm_pre_clip': 28.12782096862793, 'timing/data': 0.000358375022187829, 'timing/model': 1.1112803749274462, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.28}) +[2026-06-23 21:17:48,349][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.037608299404382706, 'train/grad_norm_pre_clip': 24.8173828125, 'timing/data': 0.00036301102954894304, 'timing/model': 1.1238840820733458, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.28}) +[2026-06-23 21:17:52,592][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.14491476118564606, 'train/grad_norm_pre_clip': 25.051372528076172, 'timing/data': 0.00037085695657879114, 'timing/model': 1.1194270889973268, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.28}) +[2026-06-23 21:17:56,657][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.2332383543252945, 'train/grad_norm_pre_clip': 32.88100051879883, 'timing/data': 0.00036192405968904495, 'timing/model': 1.11531094298698, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.28}) +[2026-06-23 21:18:00,728][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.1385822743177414, 'train/grad_norm_pre_clip': 26.746047973632812, 'timing/data': 0.0004480009665712714, 'timing/model': 1.1156617179512978, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.28}) +[2026-06-23 21:18:04,798][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.08515314757823944, 'train/grad_norm_pre_clip': 22.344724655151367, 'timing/data': 0.0004492140142247081, 'timing/model': 1.1172463539987803, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.29}) +[2026-06-23 21:18:09,008][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.1369168907403946, 'train/grad_norm_pre_clip': 28.72076988220215, 'timing/data': 0.0003592859720811248, 'timing/model': 1.1184892939636484, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.29}) +[2026-06-23 21:18:13,085][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.025304168462753296, 'train/grad_norm_pre_clip': 15.666321754455566, 'timing/data': 0.0005040880059823394, 'timing/model': 1.1154239830793813, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.29}) +[2026-06-23 21:18:17,163][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.22940099239349365, 'train/grad_norm_pre_clip': 25.592815399169922, 'timing/data': 0.00036527495831251144, 'timing/model': 1.1212435259949416, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.29}) +[2026-06-23 21:18:21,257][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.12315693497657776, 'train/grad_norm_pre_clip': 21.396329879760742, 'timing/data': 0.00036591803655028343, 'timing/model': 1.121930249966681, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.29}) +[2026-06-23 21:18:25,463][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.05973829701542854, 'train/grad_norm_pre_clip': 15.338165283203125, 'timing/data': 0.0004116169875487685, 'timing/model': 1.1165895530721173, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.29}) +[2026-06-23 21:18:29,546][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.11291560530662537, 'train/grad_norm_pre_clip': 13.859066009521484, 'timing/data': 0.00036630197428166866, 'timing/model': 1.1180232670158148, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.29}) +[2026-06-23 21:18:33,628][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.18351763486862183, 'train/grad_norm_pre_clip': 32.59656524658203, 'timing/data': 0.00037232600152492523, 'timing/model': 1.1204353889916092, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.29}) +[2026-06-23 21:18:37,720][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.14103055000305176, 'train/grad_norm_pre_clip': 31.997095108032227, 'timing/data': 0.00035633600782603025, 'timing/model': 1.122930439072661, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.29}) +[2026-06-23 21:18:41,932][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.12827400863170624, 'train/grad_norm_pre_clip': 43.91136169433594, 'timing/data': 0.0003674400504678488, 'timing/model': 1.116978585952893, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.29}) +[2026-06-23 21:18:46,015][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.28434833884239197, 'train/grad_norm_pre_clip': 61.83452606201172, 'timing/data': 0.00045443803537636995, 'timing/model': 1.1230099849635735, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.29}) +[2026-06-23 21:18:50,114][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.11980099976062775, 'train/grad_norm_pre_clip': 31.8765869140625, 'timing/data': 0.00036593805998563766, 'timing/model': 1.1147040580399334, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.29}) +[2026-06-23 21:18:54,186][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.08941613882780075, 'train/grad_norm_pre_clip': 31.76951789855957, 'timing/data': 0.00036479905247688293, 'timing/model': 1.1177684180438519, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.29}) +[2026-06-23 21:21:23,017][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 750 +[2026-06-23 21:21:23,018][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.0323515348136425, 'train/grad_norm_pre_clip': 21.821861267089844, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 87.69999999999818, 'rl_games_eval/mid_train/mean_length': 726.0, 'rl_games_eval/mid_train/std_reward': 100.46123630535064, 'rl_games_eval/mid_train/std_length': 810.7465695271242, 'rl_games_eval/mid_train/macro_mean_reward': 87.69999999999818, 'rl_games_eval/mid_train/macro_mean_length': 726.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 87.69999999999818, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 726.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 100.46123630535064, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 810.7465695271242, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 87.69999999999818, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005440600216388702, 'timing/model': 1.1182703770464286, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.29}) +[2026-06-23 21:21:27,000][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.09692730754613876, 'train/grad_norm_pre_clip': 22.62065315246582, 'timing/data': 0.0005227669607847929, 'timing/model': 1.0948915949556977, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.29}) +[2026-06-23 21:21:31,130][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.23512977361679077, 'train/grad_norm_pre_clip': 18.21346092224121, 'timing/data': 0.00045080005656927824, 'timing/model': 1.0903761819936335, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.29}) +[2026-06-23 21:21:35,134][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.10928867012262344, 'train/grad_norm_pre_clip': 19.044811248779297, 'timing/data': 0.00036108505446463823, 'timing/model': 1.096686008037068, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.29}) +[2026-06-23 21:21:39,136][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.12784963846206665, 'train/grad_norm_pre_clip': 28.98200798034668, 'timing/data': 0.0005778010236099362, 'timing/model': 1.0997280180454254, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.29}) +[2026-06-23 21:21:43,156][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.16838687658309937, 'train/grad_norm_pre_clip': 24.644746780395508, 'timing/data': 0.0003530579851940274, 'timing/model': 1.1057777899550274, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.29}) +[2026-06-23 21:21:47,282][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.17566819489002228, 'train/grad_norm_pre_clip': 48.574703216552734, 'timing/data': 0.0005923339631408453, 'timing/model': 1.1076530430000275, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.29}) +[2026-06-23 21:21:51,316][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.05111783742904663, 'train/grad_norm_pre_clip': 12.923944473266602, 'timing/data': 0.00035772297997027636, 'timing/model': 1.103942877962254, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.29}) +[2026-06-23 21:21:55,350][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.15659856796264648, 'train/grad_norm_pre_clip': 11.357356071472168, 'timing/data': 0.00035508396103978157, 'timing/model': 1.1035119090229273, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.29}) +[2026-06-23 21:21:59,385][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.09965457022190094, 'train/grad_norm_pre_clip': 32.79310607910156, 'timing/data': 0.0005733040161430836, 'timing/model': 1.105535086011514, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.29}) +[2026-06-23 21:22:03,503][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.21067117154598236, 'train/grad_norm_pre_clip': 58.9128303527832, 'timing/data': 0.00035542401019483805, 'timing/model': 1.1169295859290287, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.29}) +[2026-06-23 21:22:07,573][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.07900513708591461, 'train/grad_norm_pre_clip': 42.00687026977539, 'timing/data': 0.0003610689891502261, 'timing/model': 1.1136474780505523, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.29}) +[2026-06-23 21:22:11,621][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.2479284703731537, 'train/grad_norm_pre_clip': 45.83767318725586, 'timing/data': 0.00040650798473507166, 'timing/model': 1.1087949849897996, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.29}) +[2026-06-23 21:22:15,669][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.2164534330368042, 'train/grad_norm_pre_clip': 30.94993782043457, 'timing/data': 0.0005206300411373377, 'timing/model': 1.1082408999791369, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.3}) +[2026-06-23 21:22:19,883][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.05029422044754028, 'train/grad_norm_pre_clip': 23.48860740661621, 'timing/data': 0.0003890369553118944, 'timing/model': 1.1154373519821092, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.3}) +[2026-06-23 21:22:23,948][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.12605416774749756, 'train/grad_norm_pre_clip': 25.061420440673828, 'timing/data': 0.0003557479940354824, 'timing/model': 1.1154383160173893, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.3}) +[2026-06-23 21:22:28,020][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.04250604659318924, 'train/grad_norm_pre_clip': 29.806304931640625, 'timing/data': 0.00034187803976237774, 'timing/model': 1.114246275043115, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.3}) +[2026-06-23 21:22:32,091][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.04317657649517059, 'train/grad_norm_pre_clip': 22.654436111450195, 'timing/data': 0.0005361780058592558, 'timing/model': 1.1163003400433809, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.3}) +[2026-06-23 21:22:36,310][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.17804814875125885, 'train/grad_norm_pre_clip': 28.20511817932129, 'timing/data': 0.00042210298124700785, 'timing/model': 1.1117478769738227, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.3}) +[2026-06-23 21:22:40,385][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.13304777443408966, 'train/grad_norm_pre_clip': 58.615787506103516, 'timing/data': 0.0003690639277920127, 'timing/model': 1.1147602590499446, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.3}) +[2026-06-23 21:22:44,452][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.10648563504219055, 'train/grad_norm_pre_clip': 23.979339599609375, 'timing/data': 0.00036071299109607935, 'timing/model': 1.1160020460374653, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.3}) +[2026-06-23 21:22:48,511][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.17240706086158752, 'train/grad_norm_pre_clip': 23.46242904663086, 'timing/data': 0.0003549540415406227, 'timing/model': 1.1154848580481485, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.3}) +[2026-06-23 21:22:52,748][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.16569139063358307, 'train/grad_norm_pre_clip': 21.548362731933594, 'timing/data': 0.0003626170800998807, 'timing/model': 1.118519180919975, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.3}) +[2026-06-23 21:22:56,826][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.0647345632314682, 'train/grad_norm_pre_clip': 25.452939987182617, 'timing/data': 0.0004172900225967169, 'timing/model': 1.1155392089858651, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.3}) +[2026-06-23 21:23:00,894][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.05183856189250946, 'train/grad_norm_pre_clip': 18.963239669799805, 'timing/data': 0.00040587596595287323, 'timing/model': 1.1144217849941924, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.3}) +[2026-06-23 21:23:04,958][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.05396208539605141, 'train/grad_norm_pre_clip': 13.49911880493164, 'timing/data': 0.000578596955165267, 'timing/model': 1.1106550869299099, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.3}) +[2026-06-23 21:23:09,218][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.08350029587745667, 'train/grad_norm_pre_clip': 16.76358985900879, 'timing/data': 0.00040433695539832115, 'timing/model': 1.1211522900266573, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.3}) +[2026-06-23 21:23:13,295][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.15939262509346008, 'train/grad_norm_pre_clip': 20.269126892089844, 'timing/data': 0.000386658008210361, 'timing/model': 1.1155515379505232, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.3}) +[2026-06-23 21:23:17,366][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.10976573824882507, 'train/grad_norm_pre_clip': 20.891746520996094, 'timing/data': 0.0003603609511628747, 'timing/model': 1.1132420310750604, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.3}) +[2026-06-23 21:23:21,447][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.06894730031490326, 'train/grad_norm_pre_clip': 19.480926513671875, 'timing/data': 0.0003544529899954796, 'timing/model': 1.1194248269312084, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.3}) +[2026-06-23 21:23:25,688][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.19363848865032196, 'train/grad_norm_pre_clip': 21.64622688293457, 'timing/data': 0.00035904196556657553, 'timing/model': 1.1152289710007608, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.3}) +[2026-06-23 21:23:29,769][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.053510092198848724, 'train/grad_norm_pre_clip': 16.198570251464844, 'timing/data': 0.00036127399653196335, 'timing/model': 1.1161360919941217, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.3}) +[2026-06-23 21:23:33,836][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.03863944858312607, 'train/grad_norm_pre_clip': 32.3560905456543, 'timing/data': 0.000465711927972734, 'timing/model': 1.1215326869860291, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.3}) +[2026-06-23 21:23:37,917][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.043697938323020935, 'train/grad_norm_pre_clip': 19.5744686126709, 'timing/data': 0.0006478980649262667, 'timing/model': 1.116232114029117, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.3}) +[2026-06-23 21:23:42,170][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.04126780852675438, 'train/grad_norm_pre_clip': 28.734355926513672, 'timing/data': 0.00036527100019156933, 'timing/model': 1.1157266950467601, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.3}) +[2026-06-23 21:23:46,254][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.021479487419128418, 'train/grad_norm_pre_clip': 23.378578186035156, 'timing/data': 0.00035105610731989145, 'timing/model': 1.128355437074788, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.3}) +[2026-06-23 21:23:50,328][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.26597434282302856, 'train/grad_norm_pre_clip': 47.99857711791992, 'timing/data': 0.0003524220082908869, 'timing/model': 1.1125490239355713, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.3}) +[2026-06-23 21:23:54,422][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.0212081428617239, 'train/grad_norm_pre_clip': 21.907695770263672, 'timing/data': 0.0003613999579101801, 'timing/model': 1.1250811850186437, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.3}) +[2026-06-23 21:23:58,660][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.1217949241399765, 'train/grad_norm_pre_clip': 19.948205947875977, 'timing/data': 0.00046932604163885117, 'timing/model': 1.1163756729802117, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.3}) +[2026-06-23 21:24:02,724][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.14720550179481506, 'train/grad_norm_pre_clip': 23.336488723754883, 'timing/data': 0.00037344498559832573, 'timing/model': 1.1127988189691678, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.31}) +[2026-06-23 21:24:06,799][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.1741901934146881, 'train/grad_norm_pre_clip': 28.122934341430664, 'timing/data': 0.00042780698277056217, 'timing/model': 1.1139693250879645, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.31}) +[2026-06-23 21:24:10,866][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.30585944652557373, 'train/grad_norm_pre_clip': 49.84463119506836, 'timing/data': 0.0003668509889394045, 'timing/model': 1.1129086189903319, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.31}) +[2026-06-23 21:24:15,106][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.10468801856040955, 'train/grad_norm_pre_clip': 29.0189151763916, 'timing/data': 0.0003574719885364175, 'timing/model': 1.1204043499892578, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.31}) +[2026-06-23 21:24:19,193][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.11826321482658386, 'train/grad_norm_pre_clip': 19.23021125793457, 'timing/data': 0.00035679806023836136, 'timing/model': 1.11773862503469, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.31}) +[2026-06-23 21:24:23,269][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.09402178972959518, 'train/grad_norm_pre_clip': 33.54611587524414, 'timing/data': 0.0003632839070633054, 'timing/model': 1.1126288949744776, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.31}) +[2026-06-23 21:24:27,352][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.3944260776042938, 'train/grad_norm_pre_clip': 29.719343185424805, 'timing/data': 0.0003519289894029498, 'timing/model': 1.1235696779331192, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.31}) +[2026-06-23 21:24:31,583][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.24855400621891022, 'train/grad_norm_pre_clip': 47.025917053222656, 'timing/data': 0.0003597199684008956, 'timing/model': 1.1167030669748783, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.31}) +[2026-06-23 21:24:35,669][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.1268414556980133, 'train/grad_norm_pre_clip': 19.008743286132812, 'timing/data': 0.0003574620932340622, 'timing/model': 1.1148732569999993, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.31}) +[2026-06-23 21:24:39,745][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.07008468359708786, 'train/grad_norm_pre_clip': 24.498416900634766, 'timing/data': 0.0003575769951567054, 'timing/model': 1.115374529035762, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.31}) +[2026-06-23 21:24:43,829][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.1336754411458969, 'train/grad_norm_pre_clip': 32.9456672668457, 'timing/data': 0.0003884979523718357, 'timing/model': 1.11907415301539, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.31}) +[2026-06-23 21:24:48,070][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 800 +[2026-06-23 21:25:52,963][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 800 +[2026-06-23 21:25:52,965][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 800 +[2026-06-23 21:26:31,750][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 800 +[2026-06-23 21:26:31,750][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.21590453386306763, 'train/grad_norm_pre_clip': 43.239036560058594, 'eval/action_loss/seconds': 64.89114139205776, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.06844933331012726, 'eval/latency_2/loss': 0.06844933331012726, 'eval/flappy/loss': 0.06844933331012726, 'eval/flappy/latency_2/loss': 0.06844933331012726, 'eval/flap_precision': 0.9053030303030303, 'eval/flap_recall': 0.9728629579375848, 'eval/flap_f1': 0.9378678875081752, 'eval/flap_rate_model': 0.22, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9378678875081753, 'eval/latency_2/flap_precision': 0.9053030303030303, 'eval/latency_2/flap_recall': 0.9728629579375848, 'eval/latency_2/flap_f1': 0.9378678875081752, 'eval/latency_2/flap_rate_model': 0.22, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9378678875081753, 'eval/action_classification/data_wait_seconds': 0.5555477738380432, 'eval/action_classification/predict_seconds': 38.1077880859375, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00036374793853610754, 'timing/model': 1.119645874015987, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.31}) +[2026-06-23 21:26:35,799][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.19944928586483002, 'train/grad_norm_pre_clip': 43.796756744384766, 'timing/data': 0.0003707669675350189, 'timing/model': 1.1112522089388222, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.31}) +[2026-06-23 21:26:39,856][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.19941087067127228, 'train/grad_norm_pre_clip': 27.28602409362793, 'timing/data': 0.0003942610928788781, 'timing/model': 1.1128854469861835, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.31}) +[2026-06-23 21:26:44,071][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.04518279433250427, 'train/grad_norm_pre_clip': 24.256671905517578, 'timing/data': 0.0003522150218486786, 'timing/model': 1.2860345049994066, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.31}) +[2026-06-23 21:26:48,140][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.1473826915025711, 'train/grad_norm_pre_clip': 34.762306213378906, 'timing/data': 0.0004764110781252384, 'timing/model': 1.1178492560284212, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.31}) +[2026-06-23 21:26:52,212][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.08504340797662735, 'train/grad_norm_pre_clip': 16.396656036376953, 'timing/data': 0.0003666949924081564, 'timing/model': 1.118222885997966, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.31}) +[2026-06-23 21:26:56,259][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.23979540169239044, 'train/grad_norm_pre_clip': 16.470178604125977, 'timing/data': 0.0003619120689108968, 'timing/model': 1.1069156759185717, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.31}) +[2026-06-23 21:27:00,399][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.033721767365932465, 'train/grad_norm_pre_clip': 9.443514823913574, 'timing/data': 0.0003641740186139941, 'timing/model': 1.1115554900607094, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.31}) +[2026-06-23 21:27:04,474][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.10946251451969147, 'train/grad_norm_pre_clip': 33.03946304321289, 'timing/data': 0.00036752503365278244, 'timing/model': 1.113431500038132, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.31}) +[2026-06-23 21:27:08,538][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.0985584408044815, 'train/grad_norm_pre_clip': 24.67981719970703, 'timing/data': 0.0005016420036554337, 'timing/model': 1.1201501979958266, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.31}) +[2026-06-23 21:27:12,606][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.31533190608024597, 'train/grad_norm_pre_clip': 31.23242950439453, 'timing/data': 0.00042581395246088505, 'timing/model': 1.1190672230441123, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.31}) +[2026-06-23 21:27:16,818][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.08785783499479294, 'train/grad_norm_pre_clip': 20.388839721679688, 'timing/data': 0.0004881990607827902, 'timing/model': 1.1084340329980478, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.31}) +[2026-06-23 21:27:20,884][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.1648346185684204, 'train/grad_norm_pre_clip': 40.9165153503418, 'timing/data': 0.00046614406164735556, 'timing/model': 1.1145326730329543, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.31}) +[2026-06-23 21:27:24,948][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.12016459554433823, 'train/grad_norm_pre_clip': 95.03155517578125, 'timing/data': 0.0003673600731417537, 'timing/model': 1.1152996210148558, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.31}) +[2026-06-23 21:27:29,013][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.30344635248184204, 'train/grad_norm_pre_clip': 63.922786712646484, 'timing/data': 0.00035307707730680704, 'timing/model': 1.1142139290459454, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.32}) +[2026-06-23 21:27:33,197][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.13700883090496063, 'train/grad_norm_pre_clip': 24.40789222717285, 'timing/data': 0.0004363870248198509, 'timing/model': 1.1172884270781651, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.32}) +[2026-06-23 21:27:37,273][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.15522712469100952, 'train/grad_norm_pre_clip': 24.54055404663086, 'timing/data': 0.0004691650392487645, 'timing/model': 1.1258811140432954, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.32}) +[2026-06-23 21:27:41,359][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.06929896771907806, 'train/grad_norm_pre_clip': 36.64274597167969, 'timing/data': 0.0003508679801598191, 'timing/model': 1.1251581920078024, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.32}) +[2026-06-23 21:27:45,432][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.07992222160100937, 'train/grad_norm_pre_clip': 65.76371002197266, 'timing/data': 0.0003703390248119831, 'timing/model': 1.1175762380007654, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.32}) +[2026-06-23 21:27:49,589][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.281794011592865, 'train/grad_norm_pre_clip': 47.494834899902344, 'timing/data': 0.0003571590641513467, 'timing/model': 1.1228389689931646, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.32}) +[2026-06-23 21:27:53,655][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.10840365290641785, 'train/grad_norm_pre_clip': 26.950969696044922, 'timing/data': 0.00036555901169776917, 'timing/model': 1.1151064910227433, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.32}) +[2026-06-23 21:27:57,729][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.07835324853658676, 'train/grad_norm_pre_clip': 14.298471450805664, 'timing/data': 0.00046125997323542833, 'timing/model': 1.111556129064411, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.32}) +[2026-06-23 21:28:01,800][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.03738480061292648, 'train/grad_norm_pre_clip': 23.876358032226562, 'timing/data': 0.00036977697163820267, 'timing/model': 1.116092020063661, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.32}) +[2026-06-23 21:28:06,051][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.07158143818378448, 'train/grad_norm_pre_clip': 22.68868064880371, 'timing/data': 0.000354108982719481, 'timing/model': 1.1199403789360076, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.32}) +[2026-06-23 21:28:10,123][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.08526867628097534, 'train/grad_norm_pre_clip': 22.75237464904785, 'timing/data': 0.000355798052623868, 'timing/model': 1.1153581589460373, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.32}) +[2026-06-23 21:28:14,200][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.08719991892576218, 'train/grad_norm_pre_clip': 19.009307861328125, 'timing/data': 0.0003581730416044593, 'timing/model': 1.1160071620251983, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.32}) +[2026-06-23 21:28:18,279][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.04076177626848221, 'train/grad_norm_pre_clip': 19.909671783447266, 'timing/data': 0.0006364180007949471, 'timing/model': 1.1181450709700584, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.32}) +[2026-06-23 21:28:22,467][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.13598477840423584, 'train/grad_norm_pre_clip': 16.232925415039062, 'timing/data': 0.0003502450417727232, 'timing/model': 1.119935801019892, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.32}) +[2026-06-23 21:28:26,555][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.04228102043271065, 'train/grad_norm_pre_clip': 15.408293724060059, 'timing/data': 0.0004381019389256835, 'timing/model': 1.1254036140162498, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.32}) +[2026-06-23 21:28:30,645][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.048379480838775635, 'train/grad_norm_pre_clip': 25.475133895874023, 'timing/data': 0.000360752921551466, 'timing/model': 1.1142820989480242, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.32}) +[2026-06-23 21:28:34,719][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.3085858225822449, 'train/grad_norm_pre_clip': 18.993404388427734, 'timing/data': 0.0003595480229705572, 'timing/model': 1.1173886600881815, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.32}) +[2026-06-23 21:28:38,982][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.10155098140239716, 'train/grad_norm_pre_clip': 40.2519416809082, 'timing/data': 0.0004907329566776752, 'timing/model': 1.1210804930888116, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.32}) +[2026-06-23 21:28:43,056][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.049281176179647446, 'train/grad_norm_pre_clip': 35.19843292236328, 'timing/data': 0.00038841005880385637, 'timing/model': 1.1128090299898759, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.32}) +[2026-06-23 21:28:47,160][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.1104060560464859, 'train/grad_norm_pre_clip': 16.698543548583984, 'timing/data': 0.00040165300015360117, 'timing/model': 1.1227842150256038, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.32}) +[2026-06-23 21:28:51,254][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.03452431783080101, 'train/grad_norm_pre_clip': 32.4173469543457, 'timing/data': 0.0004938709316775203, 'timing/model': 1.122076796949841, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.32}) +[2026-06-23 21:28:55,416][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.12787503004074097, 'train/grad_norm_pre_clip': 28.54263687133789, 'timing/data': 0.00035032699815928936, 'timing/model': 1.1196535540511832, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.32}) +[2026-06-23 21:28:59,520][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.06296578794717789, 'train/grad_norm_pre_clip': 16.5653133392334, 'timing/data': 0.0003492190735414624, 'timing/model': 1.1271985620260239, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.32}) +[2026-06-23 21:29:03,628][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.161708801984787, 'train/grad_norm_pre_clip': 37.45539093017578, 'timing/data': 0.00036227796226739883, 'timing/model': 1.1172449620207772, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.32}) +[2026-06-23 21:29:07,749][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.09093926101922989, 'train/grad_norm_pre_clip': 36.938777923583984, 'timing/data': 0.0005564609309658408, 'timing/model': 1.1321813680697232, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.32}) +[2026-06-23 21:29:12,020][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.06389427930116653, 'train/grad_norm_pre_clip': 12.996861457824707, 'timing/data': 0.00035331700928509235, 'timing/model': 1.119922215002589, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.32}) +[2026-06-23 21:29:16,099][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.0784604474902153, 'train/grad_norm_pre_clip': 13.647987365722656, 'timing/data': 0.00044685998000204563, 'timing/model': 1.1195861289743334, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.33}) +[2026-06-23 21:29:20,180][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.08557158708572388, 'train/grad_norm_pre_clip': 15.51416015625, 'timing/data': 0.0005677969893440604, 'timing/model': 1.1183414380066097, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.33}) +[2026-06-23 21:29:24,269][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.15140685439109802, 'train/grad_norm_pre_clip': 27.680795669555664, 'timing/data': 0.00037630507722496986, 'timing/model': 1.122801229939796, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.33}) +[2026-06-23 21:29:28,539][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.04473055526614189, 'train/grad_norm_pre_clip': 24.775724411010742, 'timing/data': 0.0004150369204580784, 'timing/model': 1.1229799799621105, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.33}) +[2026-06-23 21:29:32,632][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.3284929692745209, 'train/grad_norm_pre_clip': 21.201313018798828, 'timing/data': 0.0003473060205578804, 'timing/model': 1.1248412779532373, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.33}) +[2026-06-23 21:29:36,746][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.1843487173318863, 'train/grad_norm_pre_clip': 27.204116821289062, 'timing/data': 0.00042200495954602957, 'timing/model': 1.1189291779883206, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.33}) +[2026-06-23 21:29:40,829][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.07192577421665192, 'train/grad_norm_pre_clip': 38.56631851196289, 'timing/data': 0.00036517903208732605, 'timing/model': 1.1195858060382307, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.33}) +[2026-06-23 21:29:45,088][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.13172340393066406, 'train/grad_norm_pre_clip': 32.00899124145508, 'timing/data': 0.0005769890267401934, 'timing/model': 1.1231508719502017, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.33}) +[2026-06-23 21:29:49,160][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.1951964944601059, 'train/grad_norm_pre_clip': 30.1698055267334, 'timing/data': 0.00035704101901501417, 'timing/model': 1.1172429909929633, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.33}) +[2026-06-23 21:29:53,248][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.10832805186510086, 'train/grad_norm_pre_clip': 18.260656356811523, 'timing/data': 0.0005677089793607593, 'timing/model': 1.1211626259610057, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.33}) +[2026-06-23 21:29:57,321][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.024648740887641907, 'train/grad_norm_pre_clip': 14.764435768127441, 'timing/data': 0.000354065909050405, 'timing/model': 1.1175463799154386, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.33}) +[2026-06-23 21:30:01,591][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.047047208994627, 'train/grad_norm_pre_clip': 22.33460235595703, 'timing/data': 0.00037408689968287945, 'timing/model': 1.1256851219804958, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.33}) +[2026-06-23 21:30:05,675][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.10436581075191498, 'train/grad_norm_pre_clip': 19.454795837402344, 'timing/data': 0.0003581199562177062, 'timing/model': 1.1179922710871324, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.33}) +[2026-06-23 21:30:09,755][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.22694121301174164, 'train/grad_norm_pre_clip': 43.860897064208984, 'timing/data': 0.0003585659433156252, 'timing/model': 1.1167329349555075, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.33}) +[2026-06-23 21:30:13,835][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.279452383518219, 'train/grad_norm_pre_clip': 29.303701400756836, 'timing/data': 0.0005779450293630362, 'timing/model': 1.117063247016631, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.33}) +[2026-06-23 21:30:18,103][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.16231697797775269, 'train/grad_norm_pre_clip': 17.900392532348633, 'timing/data': 0.0005366220138967037, 'timing/model': 1.1141719740116969, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.33}) +[2026-06-23 21:30:22,188][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.2977563738822937, 'train/grad_norm_pre_clip': 30.871566772460938, 'timing/data': 0.00040774804074317217, 'timing/model': 1.1191181869944558, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.33}) +[2026-06-23 21:30:26,303][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.06676246970891953, 'train/grad_norm_pre_clip': 13.135407447814941, 'timing/data': 0.00042037898674607277, 'timing/model': 1.1297324190381914, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.33}) +[2026-06-23 21:30:30,395][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.014340484514832497, 'train/grad_norm_pre_clip': 17.867536544799805, 'timing/data': 0.0005040570395067334, 'timing/model': 1.1227900240337476, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.33}) +[2026-06-23 21:30:34,657][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.20765230059623718, 'train/grad_norm_pre_clip': 19.363792419433594, 'timing/data': 0.000357144046574831, 'timing/model': 1.1310443780384958, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.33}) +[2026-06-23 21:30:38,801][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.12223552912473679, 'train/grad_norm_pre_clip': 11.002180099487305, 'timing/data': 0.00043702207040041685, 'timing/model': 1.1117097999667749, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.33}) +[2026-06-23 21:30:42,883][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.08356456458568573, 'train/grad_norm_pre_clip': 19.720752716064453, 'timing/data': 0.0003627410624176264, 'timing/model': 1.114477826980874, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.33}) +[2026-06-23 21:30:46,970][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.2165961116552353, 'train/grad_norm_pre_clip': 21.973955154418945, 'timing/data': 0.0005804859101772308, 'timing/model': 1.1166188720380887, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.33}) +[2026-06-23 21:30:51,226][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.07676145434379578, 'train/grad_norm_pre_clip': 34.649932861328125, 'timing/data': 0.0005410840967670083, 'timing/model': 1.1219014619709924, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.33}) +[2026-06-23 21:30:55,303][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.2219313383102417, 'train/grad_norm_pre_clip': 29.608753204345703, 'timing/data': 0.00041906407568603754, 'timing/model': 1.1165009090909734, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.33}) +[2026-06-23 21:30:59,387][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.16272595524787903, 'train/grad_norm_pre_clip': 41.3064079284668, 'timing/data': 0.0004325410118326545, 'timing/model': 1.1174525719834492, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.33}) +[2026-06-23 21:31:03,464][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.06828595697879791, 'train/grad_norm_pre_clip': 24.5130615234375, 'timing/data': 0.0003533439012244344, 'timing/model': 1.1174866990186274, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.34}) +[2026-06-23 21:31:07,708][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.06797219067811966, 'train/grad_norm_pre_clip': 31.777984619140625, 'timing/data': 0.0003722760593518615, 'timing/model': 1.119351853034459, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.34}) +[2026-06-23 21:31:11,797][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.1569720059633255, 'train/grad_norm_pre_clip': 30.961130142211914, 'timing/data': 0.00036447797901928425, 'timing/model': 1.1140374400420114, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.34}) +[2026-06-23 21:31:15,898][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.1788039356470108, 'train/grad_norm_pre_clip': 20.532312393188477, 'timing/data': 0.00037826201878488064, 'timing/model': 1.12142403004691, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.34}) +[2026-06-23 21:31:19,998][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.12052109092473984, 'train/grad_norm_pre_clip': 22.760787963867188, 'timing/data': 0.0003646840341389179, 'timing/model': 1.119234342011623, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.34}) +[2026-06-23 21:31:24,272][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.1744372397661209, 'train/grad_norm_pre_clip': 18.467731475830078, 'timing/data': 0.00040906399954110384, 'timing/model': 1.1284333929652348, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.34}) +[2026-06-23 21:31:28,361][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.0529327318072319, 'train/grad_norm_pre_clip': 34.99187469482422, 'timing/data': 0.00036195595748722553, 'timing/model': 1.1179342638934031, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.34}) +[2026-06-23 21:31:32,451][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.3525736927986145, 'train/grad_norm_pre_clip': 28.153207778930664, 'timing/data': 0.00035363505594432354, 'timing/model': 1.1146623729728162, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.34}) +[2026-06-23 21:31:36,530][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.14443586766719818, 'train/grad_norm_pre_clip': 15.902573585510254, 'timing/data': 0.00040072703268378973, 'timing/model': 1.117346896091476, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.34}) +[2026-06-23 21:31:40,780][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.15699690580368042, 'train/grad_norm_pre_clip': 31.796987533569336, 'timing/data': 0.00046594999730587006, 'timing/model': 1.1178812989965081, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.34}) +[2026-06-23 21:31:44,859][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.11842187494039536, 'train/grad_norm_pre_clip': 22.44683837890625, 'timing/data': 0.0003561970079317689, 'timing/model': 1.1166510799666867, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.34}) +[2026-06-23 21:31:48,949][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.20049992203712463, 'train/grad_norm_pre_clip': 22.32326316833496, 'timing/data': 0.0005239320453256369, 'timing/model': 1.118789535947144, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.34}) +[2026-06-23 21:31:53,067][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.06355537474155426, 'train/grad_norm_pre_clip': 11.890192031860352, 'timing/data': 0.0003546640509739518, 'timing/model': 1.1381804100237787, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.34}) +[2026-06-23 21:31:57,330][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.21895207464694977, 'train/grad_norm_pre_clip': 52.08197021484375, 'timing/data': 0.00035861204378306866, 'timing/model': 1.1196497339988127, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.34}) +[2026-06-23 21:32:01,410][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.40206360816955566, 'train/grad_norm_pre_clip': 58.86423873901367, 'timing/data': 0.0004464880330488086, 'timing/model': 1.1165106919361278, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.34}) +[2026-06-23 21:32:05,492][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.06160414218902588, 'train/grad_norm_pre_clip': 28.362333297729492, 'timing/data': 0.0003563310019671917, 'timing/model': 1.1183930659899488, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.34}) +[2026-06-23 21:32:09,582][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.0063490732572972775, 'train/grad_norm_pre_clip': 16.46392822265625, 'timing/data': 0.00043535500299185514, 'timing/model': 1.1237225560471416, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.34}) +[2026-06-23 21:32:13,841][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.05567271634936333, 'train/grad_norm_pre_clip': 22.209754943847656, 'timing/data': 0.0003923039184883237, 'timing/model': 1.1157725810771808, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.34}) +[2026-06-23 21:32:17,942][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.09085695445537567, 'train/grad_norm_pre_clip': 29.896631240844727, 'timing/data': 0.0004123810213059187, 'timing/model': 1.128579327953048, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.34}) +[2026-06-23 21:32:22,033][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.12996743619441986, 'train/grad_norm_pre_clip': 31.948102951049805, 'timing/data': 0.0003712590551003814, 'timing/model': 1.1201731039909646, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.34}) +[2026-06-23 21:32:26,111][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.18720529973506927, 'train/grad_norm_pre_clip': 67.42176055908203, 'timing/data': 0.00044042093213647604, 'timing/model': 1.1144450409337878, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.34}) +[2026-06-23 21:32:30,364][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.1748172789812088, 'train/grad_norm_pre_clip': 23.280227661132812, 'timing/data': 0.0004192260093986988, 'timing/model': 1.1131637379294261, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.34}) +[2026-06-23 21:32:34,462][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.18097153306007385, 'train/grad_norm_pre_clip': 39.16524124145508, 'timing/data': 0.00042036199010908604, 'timing/model': 1.1253357130335644, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.34}) +[2026-06-23 21:32:38,544][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.23553907871246338, 'train/grad_norm_pre_clip': 36.90042495727539, 'timing/data': 0.00036331999581307173, 'timing/model': 1.1250071950489655, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.34}) +[2026-06-23 21:32:42,632][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.33420616388320923, 'train/grad_norm_pre_clip': 32.598934173583984, 'timing/data': 0.00037674407940357924, 'timing/model': 1.1204510719981045, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.34}) +[2026-06-23 21:32:46,883][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.03593241423368454, 'train/grad_norm_pre_clip': 19.788349151611328, 'timing/data': 0.0005046760197728872, 'timing/model': 1.1212894329801202, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.34}) +[2026-06-23 21:32:50,976][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.18847697973251343, 'train/grad_norm_pre_clip': 33.71156311035156, 'timing/data': 0.0003635920584201813, 'timing/model': 1.121817915001884, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.35}) +[2026-06-23 21:32:55,059][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.10748973488807678, 'train/grad_norm_pre_clip': 35.003963470458984, 'timing/data': 0.0005909550236538053, 'timing/model': 1.1203725739615038, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.35}) +[2026-06-23 21:32:59,167][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.25022149085998535, 'train/grad_norm_pre_clip': 49.81299591064453, 'timing/data': 0.0005453240592032671, 'timing/model': 1.1141917660133913, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.35}) +[2026-06-23 21:33:03,426][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.4618166983127594, 'train/grad_norm_pre_clip': 72.83071899414062, 'timing/data': 0.00037368293851614, 'timing/model': 1.121641656034626, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.35}) +[2026-06-23 21:33:07,515][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.07716765254735947, 'train/grad_norm_pre_clip': 22.080217361450195, 'timing/data': 0.0003563690697774291, 'timing/model': 1.1165709800552577, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.35}) +[2026-06-23 21:33:11,603][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.08744220435619354, 'train/grad_norm_pre_clip': 23.587684631347656, 'timing/data': 0.0003561900230124593, 'timing/model': 1.1236676949774846, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.35}) +[2026-06-23 21:33:15,704][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.13695839047431946, 'train/grad_norm_pre_clip': 41.087467193603516, 'timing/data': 0.00035848096013069153, 'timing/model': 1.122406080015935, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.35}) +[2026-06-23 21:33:19,960][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.05983920022845268, 'train/grad_norm_pre_clip': 35.137184143066406, 'timing/data': 0.0003967060474678874, 'timing/model': 1.1190760909812525, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.35}) +[2026-06-23 21:33:24,048][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 900 +[2026-06-23 21:34:28,028][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 900 +[2026-06-23 21:34:28,030][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 900 +[2026-06-23 21:35:06,819][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 900 +[2026-06-23 21:35:06,820][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.3438423275947571, 'train/grad_norm_pre_clip': 56.70214080810547, 'eval/action_loss/seconds': 63.97667888295837, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.09818260371685028, 'eval/latency_2/loss': 0.09818260371685028, 'eval/flappy/loss': 0.09818260371685028, 'eval/flappy/latency_2/loss': 0.09818260371685028, 'eval/flap_precision': 0.9805194805194806, 'eval/flap_recall': 0.819538670284939, 'eval/flap_f1': 0.8928307464892831, 'eval/flap_rate_model': 0.1711111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8928307464892832, 'eval/latency_2/flap_precision': 0.9805194805194806, 'eval/latency_2/flap_recall': 0.819538670284939, 'eval/latency_2/flap_f1': 0.8928307464892831, 'eval/latency_2/flap_rate_model': 0.1711111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8928307464892832, 'eval/action_classification/data_wait_seconds': 0.5339869260787964, 'eval/action_classification/predict_seconds': 38.11329650878906, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00042985298205167055, 'timing/model': 1.1209072289057076, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.35}) +[2026-06-23 21:35:10,885][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.08046415448188782, 'train/grad_norm_pre_clip': 37.66436004638672, 'timing/data': 0.0003602519864216447, 'timing/model': 1.123650711029768, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.35}) +[2026-06-23 21:35:15,063][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.050898775458335876, 'train/grad_norm_pre_clip': 19.874866485595703, 'timing/data': 0.00035551097244024277, 'timing/model': 1.1124313269974664, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.35}) +[2026-06-23 21:35:19,104][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.10963401943445206, 'train/grad_norm_pre_clip': 17.336395263671875, 'timing/data': 0.0003630230203270912, 'timing/model': 1.1140908179804683, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.35}) +[2026-06-23 21:35:23,169][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.24228087067604065, 'train/grad_norm_pre_clip': 31.858280181884766, 'timing/data': 0.0003652769373729825, 'timing/model': 1.1073352888925, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.35}) +[2026-06-23 21:35:27,228][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.1070253923535347, 'train/grad_norm_pre_clip': 40.53727722167969, 'timing/data': 0.0003524450585246086, 'timing/model': 1.1144459480419755, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.35}) +[2026-06-23 21:35:31,421][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.20974229276180267, 'train/grad_norm_pre_clip': 45.0401496887207, 'timing/data': 0.0005483570275828242, 'timing/model': 1.1144396140007302, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.35}) +[2026-06-23 21:35:35,488][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.17484986782073975, 'train/grad_norm_pre_clip': 17.286907196044922, 'timing/data': 0.0003583739744499326, 'timing/model': 1.114920041989535, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.35}) +[2026-06-23 21:35:39,552][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.09606503695249557, 'train/grad_norm_pre_clip': 15.527628898620605, 'timing/data': 0.00040074100252240896, 'timing/model': 1.1122128339484334, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.35}) +[2026-06-23 21:35:43,618][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.1105007529258728, 'train/grad_norm_pre_clip': 36.11272430419922, 'timing/data': 0.00035639305133372545, 'timing/model': 1.1183615629561245, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.35}) +[2026-06-23 21:35:47,763][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.1470775455236435, 'train/grad_norm_pre_clip': 46.06756591796875, 'timing/data': 0.00040900695603340864, 'timing/model': 1.1063154510920867, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.35}) +[2026-06-23 21:35:51,826][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.03661555051803589, 'train/grad_norm_pre_clip': 24.333507537841797, 'timing/data': 0.00037847994826734066, 'timing/model': 1.1110733059467748, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.35}) +[2026-06-23 21:35:55,880][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.1621338278055191, 'train/grad_norm_pre_clip': 19.57077407836914, 'timing/data': 0.0005087710451334715, 'timing/model': 1.1111312019638717, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.35}) +[2026-06-23 21:35:59,951][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.10116654634475708, 'train/grad_norm_pre_clip': 24.500019073486328, 'timing/data': 0.0003558599855750799, 'timing/model': 1.11377208691556, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.35}) +[2026-06-23 21:36:04,189][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.0456845797598362, 'train/grad_norm_pre_clip': 23.143465042114258, 'timing/data': 0.0004411611007526517, 'timing/model': 1.1207102069165558, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.35}) +[2026-06-23 21:36:08,272][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.16377028822898865, 'train/grad_norm_pre_clip': 34.45195388793945, 'timing/data': 0.0006060120649635792, 'timing/model': 1.120919800014235, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.35}) +[2026-06-23 21:36:12,337][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.09193958342075348, 'train/grad_norm_pre_clip': 20.19761085510254, 'timing/data': 0.00036167993675917387, 'timing/model': 1.1122984290122986, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.35}) +[2026-06-23 21:36:16,479][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.19150002300739288, 'train/grad_norm_pre_clip': 19.770444869995117, 'timing/data': 0.0003591339336708188, 'timing/model': 1.1912264730781317, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.35}) +[2026-06-23 21:36:20,570][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.034298233687877655, 'train/grad_norm_pre_clip': 21.918411254882812, 'timing/data': 0.00035971798934042454, 'timing/model': 1.1202502440428361, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.36}) +[2026-06-23 21:36:24,658][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.11277350783348083, 'train/grad_norm_pre_clip': 15.617376327514648, 'timing/data': 0.0003929600352421403, 'timing/model': 1.1159392088884488, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.36}) +[2026-06-23 21:36:28,748][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.3489128351211548, 'train/grad_norm_pre_clip': 134.554931640625, 'timing/data': 0.00036181602627038956, 'timing/model': 1.1176783599657938, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.36}) +[2026-06-23 21:36:32,897][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.17121949791908264, 'train/grad_norm_pre_clip': 23.590673446655273, 'timing/data': 0.00040184007957577705, 'timing/model': 1.1883546020835638, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.36}) +[2026-06-23 21:36:36,968][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.11645116657018661, 'train/grad_norm_pre_clip': 18.813695907592773, 'timing/data': 0.0005462639965116978, 'timing/model': 1.1215623460011557, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.36}) +[2026-06-23 21:36:41,043][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.13180512189865112, 'train/grad_norm_pre_clip': 26.73101806640625, 'timing/data': 0.00035482796374708414, 'timing/model': 1.1181170790223405, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.36}) +[2026-06-23 21:36:45,129][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.05079216510057449, 'train/grad_norm_pre_clip': 16.194580078125, 'timing/data': 0.000362263061106205, 'timing/model': 1.116046064067632, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.36}) +[2026-06-23 21:36:49,362][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.06015211343765259, 'train/grad_norm_pre_clip': 38.99453353881836, 'timing/data': 0.0003591870190575719, 'timing/model': 1.257600731914863, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.36}) +[2026-06-23 21:36:53,440][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.05624764412641525, 'train/grad_norm_pre_clip': 15.01926040649414, 'timing/data': 0.00041678594425320625, 'timing/model': 1.1147125059505925, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.36}) +[2026-06-23 21:36:57,514][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.007457265164703131, 'train/grad_norm_pre_clip': 11.929953575134277, 'timing/data': 0.0004377990262582898, 'timing/model': 1.1152384449960664, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.36}) +[2026-06-23 21:37:01,584][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.19829395413398743, 'train/grad_norm_pre_clip': 46.48550796508789, 'timing/data': 0.00037413800600916147, 'timing/model': 1.1211239910917357, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.36}) +[2026-06-23 21:37:05,753][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.05345239117741585, 'train/grad_norm_pre_clip': 16.693435668945312, 'timing/data': 0.000495839980430901, 'timing/model': 1.1145498390542343, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.36}) +[2026-06-23 21:37:09,825][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.15842771530151367, 'train/grad_norm_pre_clip': 35.26827621459961, 'timing/data': 0.00041598593816161156, 'timing/model': 1.113811337039806, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.36}) +[2026-06-23 21:37:13,898][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.09215767681598663, 'train/grad_norm_pre_clip': 29.45355987548828, 'timing/data': 0.0005086550954729319, 'timing/model': 1.1191211689729244, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.36}) +[2026-06-23 21:37:17,964][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.0672474354505539, 'train/grad_norm_pre_clip': 20.889305114746094, 'timing/data': 0.000351940980181098, 'timing/model': 1.1137914340943098, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.36}) +[2026-06-23 21:37:22,107][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.038556020706892014, 'train/grad_norm_pre_clip': 22.043901443481445, 'timing/data': 0.00036594097036868334, 'timing/model': 1.1176679310156032, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.36}) +[2026-06-23 21:37:26,197][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.06706489622592926, 'train/grad_norm_pre_clip': 20.303213119506836, 'timing/data': 0.00044706289190799, 'timing/model': 1.1229111920110881, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.36}) +[2026-06-23 21:37:30,281][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.07196167856454849, 'train/grad_norm_pre_clip': 21.037355422973633, 'timing/data': 0.0005067900056019425, 'timing/model': 1.117122569005005, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.36}) +[2026-06-23 21:37:34,367][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.2750978469848633, 'train/grad_norm_pre_clip': 34.78562927246094, 'timing/data': 0.0004068210255354643, 'timing/model': 1.1180994680617005, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.36}) +[2026-06-23 21:37:38,612][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.10005546361207962, 'train/grad_norm_pre_clip': 22.653724670410156, 'timing/data': 0.0005456280196085572, 'timing/model': 1.115526361973025, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.36}) +[2026-06-23 21:37:42,679][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.08525493741035461, 'train/grad_norm_pre_clip': 14.974932670593262, 'timing/data': 0.0004430270055308938, 'timing/model': 1.109807139961049, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.36}) +[2026-06-23 21:37:46,763][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.014181900769472122, 'train/grad_norm_pre_clip': 13.937284469604492, 'timing/data': 0.00035548594314605, 'timing/model': 1.1194361420348287, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.36}) +[2026-06-23 21:37:50,829][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.19264380633831024, 'train/grad_norm_pre_clip': 31.11275291442871, 'timing/data': 0.000506295938976109, 'timing/model': 1.1096690089907497, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.36}) +[2026-06-23 21:37:54,981][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.15525445342063904, 'train/grad_norm_pre_clip': 36.459659576416016, 'timing/data': 0.0005065050208941102, 'timing/model': 1.1153877329779789, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.36}) +[2026-06-23 21:37:59,075][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.19027578830718994, 'train/grad_norm_pre_clip': 18.157636642456055, 'timing/data': 0.00040093809366226196, 'timing/model': 1.1230885769473389, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.36}) +[2026-06-23 21:38:03,140][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.15059654414653778, 'train/grad_norm_pre_clip': 37.611778259277344, 'timing/data': 0.00041529792360961437, 'timing/model': 1.110832046950236, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.36}) +[2026-06-23 21:38:07,230][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.12714983522891998, 'train/grad_norm_pre_clip': 38.65380859375, 'timing/data': 0.0003609259147197008, 'timing/model': 1.1202337479917333, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.37}) +[2026-06-23 21:38:11,444][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.2832503616809845, 'train/grad_norm_pre_clip': 45.70787048339844, 'timing/data': 0.00034898705780506134, 'timing/model': 1.231696823029779, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.37}) +[2026-06-23 21:38:15,531][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.19961020350456238, 'train/grad_norm_pre_clip': 31.08843421936035, 'timing/data': 0.00036951201036572456, 'timing/model': 1.1185987680219114, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.37}) +[2026-06-23 21:38:19,618][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.06567788124084473, 'train/grad_norm_pre_clip': 12.04995346069336, 'timing/data': 0.0003575580194592476, 'timing/model': 1.1247219739016145, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.37}) +[2026-06-23 21:38:23,710][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.21584856510162354, 'train/grad_norm_pre_clip': 44.7130241394043, 'timing/data': 0.00036176491994410753, 'timing/model': 1.1194482300197706, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.37}) +[2026-06-23 21:38:27,793][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.19619883596897125, 'train/grad_norm_pre_clip': 28.020694732666016, 'timing/data': 0.0006791070336475968, 'timing/model': 1.1195784880546853, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.37}) +[2026-06-23 21:38:32,038][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.23623840510845184, 'train/grad_norm_pre_clip': 43.63749313354492, 'timing/data': 0.00042749103158712387, 'timing/model': 1.1169702149927616, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.37}) +[2026-06-23 21:38:36,117][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.1406583935022354, 'train/grad_norm_pre_clip': 21.41267967224121, 'timing/data': 0.0004025240195915103, 'timing/model': 1.1160263599595055, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.37}) +[2026-06-23 21:38:40,202][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.27303576469421387, 'train/grad_norm_pre_clip': 56.32874298095703, 'timing/data': 0.0007266390603035688, 'timing/model': 1.1203984740423039, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.37}) +[2026-06-23 21:38:44,285][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.04120341315865517, 'train/grad_norm_pre_clip': 28.409255981445312, 'timing/data': 0.0003838749835267663, 'timing/model': 1.1177806700579822, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.37}) +[2026-06-23 21:38:48,452][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.23352299630641937, 'train/grad_norm_pre_clip': 24.96613311767578, 'timing/data': 0.0004700769204646349, 'timing/model': 1.1211632220074534, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.37}) +[2026-06-23 21:38:52,541][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.11975885927677155, 'train/grad_norm_pre_clip': 13.132952690124512, 'timing/data': 0.0005996399559080601, 'timing/model': 1.119460187968798, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.37}) +[2026-06-23 21:38:56,634][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.05520516633987427, 'train/grad_norm_pre_clip': 7.808830261230469, 'timing/data': 0.0003600320778787136, 'timing/model': 1.12110611400567, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.37}) +[2026-06-23 21:39:00,725][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.050989869982004166, 'train/grad_norm_pre_clip': 17.79706573486328, 'timing/data': 0.000365929096005857, 'timing/model': 1.1205764949554577, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.37}) +[2026-06-23 21:39:04,892][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.13362687826156616, 'train/grad_norm_pre_clip': 23.883773803710938, 'timing/data': 0.0004103659885004163, 'timing/model': 1.1114198309369385, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.37}) +[2026-06-23 21:39:08,977][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.05546991527080536, 'train/grad_norm_pre_clip': 11.826760292053223, 'timing/data': 0.00034925201907753944, 'timing/model': 1.120350155979395, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.37}) +[2026-06-23 21:39:13,065][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.08323640376329422, 'train/grad_norm_pre_clip': 13.468769073486328, 'timing/data': 0.0005921790143474936, 'timing/model': 1.120417378959246, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.37}) +[2026-06-23 21:39:17,167][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.05538681149482727, 'train/grad_norm_pre_clip': 42.61539077758789, 'timing/data': 0.0007215549703687429, 'timing/model': 1.1161801029229537, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.37}) +[2026-06-23 21:39:21,326][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.021168863400816917, 'train/grad_norm_pre_clip': 26.19795036315918, 'timing/data': 0.00046369200572371483, 'timing/model': 1.1252932080533355, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.37}) +[2026-06-23 21:39:25,404][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.14515376091003418, 'train/grad_norm_pre_clip': 54.739192962646484, 'timing/data': 0.0003604469820857048, 'timing/model': 1.1220391920069233, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.37}) +[2026-06-23 21:39:29,504][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.3773694634437561, 'train/grad_norm_pre_clip': 56.70539855957031, 'timing/data': 0.0005566120380535722, 'timing/model': 1.120397480088286, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.37}) +[2026-06-23 21:39:33,584][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.24999754130840302, 'train/grad_norm_pre_clip': 21.886621475219727, 'timing/data': 0.00035488803405314684, 'timing/model': 1.1197727640392259, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.37}) +[2026-06-23 21:39:37,802][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.09618514776229858, 'train/grad_norm_pre_clip': 26.799692153930664, 'timing/data': 0.0003759219544008374, 'timing/model': 1.1178059260128066, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.37}) +[2026-06-23 21:39:41,882][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.31439051032066345, 'train/grad_norm_pre_clip': 110.1139144897461, 'timing/data': 0.0003769379109144211, 'timing/model': 1.1142087150365114, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.37}) +[2026-06-23 21:39:45,964][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.08838022500276566, 'train/grad_norm_pre_clip': 91.42625427246094, 'timing/data': 0.00045443198177963495, 'timing/model': 1.1155182890361175, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.37}) +[2026-06-23 21:39:50,144][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.32334455847740173, 'train/grad_norm_pre_clip': 65.2490463256836, 'timing/data': 0.00035774894058704376, 'timing/model': 1.21466026105918, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.38}) +[2026-06-23 21:39:54,238][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.2066989243030548, 'train/grad_norm_pre_clip': 27.179271697998047, 'timing/data': 0.0003587929531931877, 'timing/model': 1.124174713040702, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.38}) +[2026-06-23 21:39:58,325][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.016804039478302002, 'train/grad_norm_pre_clip': 10.205957412719727, 'timing/data': 0.0004268230404704809, 'timing/model': 1.1153755789855495, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.38}) +[2026-06-23 21:40:02,412][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.21459409594535828, 'train/grad_norm_pre_clip': 32.35478591918945, 'timing/data': 0.0007077120244503021, 'timing/model': 1.117481421912089, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.38}) +[2026-06-23 21:40:06,683][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.2438715249300003, 'train/grad_norm_pre_clip': 62.0057258605957, 'timing/data': 0.0003604240482673049, 'timing/model': 1.3061399910366163, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.38}) +[2026-06-23 21:40:10,768][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.04386785253882408, 'train/grad_norm_pre_clip': 38.108367919921875, 'timing/data': 0.0003578879404813051, 'timing/model': 1.1244879859732464, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.38}) +[2026-06-23 21:40:14,861][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.122707799077034, 'train/grad_norm_pre_clip': 41.370094299316406, 'timing/data': 0.0003584519727155566, 'timing/model': 1.1204312751069665, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.38}) +[2026-06-23 21:40:18,942][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.16630268096923828, 'train/grad_norm_pre_clip': 15.962681770324707, 'timing/data': 0.00035440002102404833, 'timing/model': 1.1187752779806033, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.38}) +[2026-06-23 21:40:23,117][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.13321122527122498, 'train/grad_norm_pre_clip': 19.51210594177246, 'timing/data': 0.00040857307612895966, 'timing/model': 1.1332402609987184, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.38}) +[2026-06-23 21:40:27,229][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.03121226839721203, 'train/grad_norm_pre_clip': 54.70654296875, 'timing/data': 0.0006698679644614458, 'timing/model': 1.1231023679720238, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.38}) +[2026-06-23 21:40:31,330][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.17071139812469482, 'train/grad_norm_pre_clip': 48.084510803222656, 'timing/data': 0.00036261999048292637, 'timing/model': 1.1279941090615466, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.38}) +[2026-06-23 21:40:35,437][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.20822475850582123, 'train/grad_norm_pre_clip': 35.4159049987793, 'timing/data': 0.00036133104003965855, 'timing/model': 1.1303753750398755, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.38}) +[2026-06-23 21:40:39,604][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.2134954035282135, 'train/grad_norm_pre_clip': 24.878337860107422, 'timing/data': 0.0003587030805647373, 'timing/model': 1.1176501919981092, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.38}) +[2026-06-23 21:40:43,687][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.07598164677619934, 'train/grad_norm_pre_clip': 16.084463119506836, 'timing/data': 0.000373288057744503, 'timing/model': 1.1266095009632409, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.38}) +[2026-06-23 21:40:47,775][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.15917836129665375, 'train/grad_norm_pre_clip': 33.52766036987305, 'timing/data': 0.0003598569892346859, 'timing/model': 1.1164988589007407, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.38}) +[2026-06-23 21:40:51,869][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.3406461477279663, 'train/grad_norm_pre_clip': 49.53049087524414, 'timing/data': 0.00036502291914075613, 'timing/model': 1.1155924000777304, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.38}) +[2026-06-23 21:40:56,098][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.245185986161232, 'train/grad_norm_pre_clip': 65.41899108886719, 'timing/data': 0.00035947293508797884, 'timing/model': 1.120126574067399, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.38}) +[2026-06-23 21:41:00,188][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.0645643025636673, 'train/grad_norm_pre_clip': 44.40004348754883, 'timing/data': 0.0003651679726317525, 'timing/model': 1.125941007048823, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.38}) +[2026-06-23 21:41:04,272][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.0631791427731514, 'train/grad_norm_pre_clip': 33.09280014038086, 'timing/data': 0.0006620469503104687, 'timing/model': 1.1248203020077199, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.38}) +[2026-06-23 21:41:08,361][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.047268614172935486, 'train/grad_norm_pre_clip': 15.574324607849121, 'timing/data': 0.00039700395427644253, 'timing/model': 1.1204022360034287, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.38}) +[2026-06-23 21:41:12,566][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.12380000203847885, 'train/grad_norm_pre_clip': 11.615114212036133, 'timing/data': 0.0003555150469765067, 'timing/model': 1.1167844829615206, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.38}) +[2026-06-23 21:41:16,650][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.07718458771705627, 'train/grad_norm_pre_clip': 30.54167366027832, 'timing/data': 0.0003821499412879348, 'timing/model': 1.113540305988863, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.38}) +[2026-06-23 21:41:20,734][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.09137022495269775, 'train/grad_norm_pre_clip': 17.80759620666504, 'timing/data': 0.0004549719160422683, 'timing/model': 1.1149142440408468, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.38}) +[2026-06-23 21:41:24,818][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.051968757063150406, 'train/grad_norm_pre_clip': 15.642178535461426, 'timing/data': 0.0003806289751082659, 'timing/model': 1.1175103010609746, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.38}) +[2026-06-23 21:41:28,958][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.05447456240653992, 'train/grad_norm_pre_clip': 12.40864086151123, 'timing/data': 0.00055334297940135, 'timing/model': 1.114353068987839, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.38}) +[2026-06-23 21:41:33,060][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.22720833122730255, 'train/grad_norm_pre_clip': 15.16303539276123, 'timing/data': 0.0005707789678126574, 'timing/model': 1.1195585109526291, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.38}) +[2026-06-23 21:41:37,146][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.005579467862844467, 'train/grad_norm_pre_clip': 15.209918975830078, 'timing/data': 0.0005614770343527198, 'timing/model': 1.1212320940103382, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.39}) +[2026-06-23 21:41:41,236][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.332994669675827, 'train/grad_norm_pre_clip': 38.452880859375, 'timing/data': 0.0003671350423246622, 'timing/model': 1.120504159014672, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.39}) +[2026-06-23 21:41:45,491][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.12275740504264832, 'train/grad_norm_pre_clip': 27.232349395751953, 'timing/data': 0.0003501339815557003, 'timing/model': 1.1169245700584725, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.39}) +[2026-06-23 21:41:49,589][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.08728101849555969, 'train/grad_norm_pre_clip': 17.613601684570312, 'timing/data': 0.0003598120529204607, 'timing/model': 1.1204285499406978, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.39}) +[2026-06-23 21:41:53,656][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.023736009374260902, 'train/grad_norm_pre_clip': 19.994211196899414, 'timing/data': 0.0004698849515989423, 'timing/model': 1.1153889630222693, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.39}) +[2026-06-23 21:41:57,740][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1000 +[2026-06-23 21:43:02,032][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1000 +[2026-06-23 21:43:02,033][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1000 +[2026-06-23 21:43:41,349][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1000 +[2026-06-23 21:45:52,615][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-23 21:45:52,616][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.03792152553796768, 'train/grad_norm_pre_clip': 17.683387756347656, 'eval/action_loss/seconds': 64.28988871898036, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.08519867062568665, 'eval/latency_2/loss': 0.08519867062568665, 'eval/flappy/loss': 0.08519867062568665, 'eval/flappy/latency_2/loss': 0.08519867062568665, 'eval/flap_precision': 0.9714714714714715, 'eval/flap_recall': 0.8778833107191316, 'eval/flap_f1': 0.9223093371347113, 'eval/flap_rate_model': 0.185, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9223093371347113, 'eval/latency_2/flap_precision': 0.9714714714714715, 'eval/latency_2/flap_recall': 0.8778833107191316, 'eval/latency_2/flap_f1': 0.9223093371347113, 'eval/latency_2/flap_rate_model': 0.185, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9223093371347113, 'eval/action_classification/data_wait_seconds': 0.5794416666030884, 'eval/action_classification/predict_seconds': 38.60301208496094, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 79.39999999999868, 'rl_games_eval/mid_train/mean_length': 659.2, 'rl_games_eval/mid_train/std_reward': 62.80047770518764, 'rl_games_eval/mid_train/std_length': 506.935656666603, 'rl_games_eval/mid_train/macro_mean_reward': 79.39999999999868, 'rl_games_eval/mid_train/macro_mean_length': 659.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 79.39999999999868, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 659.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 62.80047770518764, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 506.935656666603, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 79.39999999999868, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000400181976146996, 'timing/model': 1.1159947559935972, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.39}) +[2026-06-23 21:45:52,616][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_1000_state +[2026-06-23 21:45:52,619][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 21:45:56,336][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_1000_state/model.safetensors +[2026-06-23 21:46:04,159][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_1000_state/optimizer.bin +[2026-06-23 21:46:04,160][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_1000_state/sampler.bin +[2026-06-23 21:46:04,160][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-23 21:46:04,161][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-23 21:46:04,227][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-23 21:46:04,228][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 21:46:04,242][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 21:46:08,227][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.43481627106666565, 'train/grad_norm_pre_clip': 67.5407485961914, 'timing/data': 0.000560745014809072, 'timing/model': 1.0949301139917225, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.39}) +[2026-06-23 21:46:12,387][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.15579186379909515, 'train/grad_norm_pre_clip': 63.83019256591797, 'timing/data': 0.00036308798007667065, 'timing/model': 1.2676324299536645, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.39}) +[2026-06-23 21:46:16,382][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.1459283083677292, 'train/grad_norm_pre_clip': 27.42361831665039, 'timing/data': 0.0003711929311975837, 'timing/model': 1.094453300931491, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.39}) +[2026-06-23 21:46:20,389][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.04184505343437195, 'train/grad_norm_pre_clip': 29.53816032409668, 'timing/data': 0.0003542749909684062, 'timing/model': 1.1019302069908008, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.39}) +[2026-06-23 21:46:24,394][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.14082768559455872, 'train/grad_norm_pre_clip': 110.18242645263672, 'timing/data': 0.0003642560914158821, 'timing/model': 1.1013730539707467, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.39}) +[2026-06-23 21:46:28,472][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.11251110583543777, 'train/grad_norm_pre_clip': 39.01042938232422, 'timing/data': 0.0003683529794216156, 'timing/model': 1.098952138912864, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.39}) +[2026-06-23 21:46:32,506][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.17760510742664337, 'train/grad_norm_pre_clip': 52.26469802856445, 'timing/data': 0.0004024350782856345, 'timing/model': 1.1111641880124807, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.39}) +[2026-06-23 21:46:36,523][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.22592824697494507, 'train/grad_norm_pre_clip': 22.65310287475586, 'timing/data': 0.00045725307427346706, 'timing/model': 1.0992934399982914, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.39}) +[2026-06-23 21:46:40,554][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.03754175826907158, 'train/grad_norm_pre_clip': 22.435949325561523, 'timing/data': 0.00036407995503395796, 'timing/model': 1.1086148370523006, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.39}) +[2026-06-23 21:46:44,744][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.06383916735649109, 'train/grad_norm_pre_clip': 26.404293060302734, 'timing/data': 0.00036774598993360996, 'timing/model': 1.1047483910806477, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.39}) +[2026-06-23 21:46:48,787][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.20666687190532684, 'train/grad_norm_pre_clip': 44.740943908691406, 'timing/data': 0.0003733679186552763, 'timing/model': 1.1111805189866573, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.39}) +[2026-06-23 21:46:52,826][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.06574610620737076, 'train/grad_norm_pre_clip': 29.15411949157715, 'timing/data': 0.0004316960694268346, 'timing/model': 1.1078996079741046, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.39}) +[2026-06-23 21:46:56,863][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.054646849632263184, 'train/grad_norm_pre_clip': 17.798551559448242, 'timing/data': 0.000609206035733223, 'timing/model': 1.105134516954422, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.39}) +[2026-06-23 21:47:01,016][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.09499365091323853, 'train/grad_norm_pre_clip': 16.372516632080078, 'timing/data': 0.0004211569903418422, 'timing/model': 1.1142578379949555, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.39}) +[2026-06-23 21:47:05,067][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.1296270191669464, 'train/grad_norm_pre_clip': 16.21897315979004, 'timing/data': 0.0005056229420006275, 'timing/model': 1.109794836025685, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.39}) +[2026-06-23 21:47:09,122][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.14554007351398468, 'train/grad_norm_pre_clip': 17.187990188598633, 'timing/data': 0.0005110170459374785, 'timing/model': 1.1087405630387366, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.39}) +[2026-06-23 21:47:13,181][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.21824827790260315, 'train/grad_norm_pre_clip': 27.316997528076172, 'timing/data': 0.0003640389768406749, 'timing/model': 1.110594004043378, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.39}) +[2026-06-23 21:47:17,332][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.07135996222496033, 'train/grad_norm_pre_clip': 14.050567626953125, 'timing/data': 0.00036949804052710533, 'timing/model': 1.1149988840334117, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.39}) +[2026-06-23 21:47:21,397][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.0883917436003685, 'train/grad_norm_pre_clip': 13.315156936645508, 'timing/data': 0.0003504169872030616, 'timing/model': 1.1228727410780266, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.39}) +[2026-06-23 21:47:25,466][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.03689800202846527, 'train/grad_norm_pre_clip': 29.19791030883789, 'timing/data': 0.0005302340723574162, 'timing/model': 1.1164668779820204, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.39}) +[2026-06-23 21:47:29,535][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.03038034401834011, 'train/grad_norm_pre_clip': 29.76072120666504, 'timing/data': 0.00036435993388295174, 'timing/model': 1.1175711510004476, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.4}) +[2026-06-23 21:47:33,767][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.054641999304294586, 'train/grad_norm_pre_clip': 15.981554985046387, 'timing/data': 0.00037932698614895344, 'timing/model': 1.118196501978673, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.4}) +[2026-06-23 21:47:37,834][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.07265661656856537, 'train/grad_norm_pre_clip': 24.296741485595703, 'timing/data': 0.00036095106042921543, 'timing/model': 1.1230588409816846, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.4}) +[2026-06-23 21:47:41,908][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.1334802508354187, 'train/grad_norm_pre_clip': 30.283239364624023, 'timing/data': 0.0003862050361931324, 'timing/model': 1.1162717620609328, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.4}) +[2026-06-23 21:47:45,971][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.06312111765146255, 'train/grad_norm_pre_clip': 22.194791793823242, 'timing/data': 0.0006372099742293358, 'timing/model': 1.1160590969957411, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.4}) +[2026-06-23 21:47:50,131][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.165919691324234, 'train/grad_norm_pre_clip': 25.424715042114258, 'timing/data': 0.00040835398249328136, 'timing/model': 1.1071644440526143, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.4}) +[2026-06-23 21:47:54,206][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.123149074614048, 'train/grad_norm_pre_clip': 20.69639778137207, 'timing/data': 0.0003592139109969139, 'timing/model': 1.1138159660622478, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.4}) +[2026-06-23 21:47:58,298][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.1998097002506256, 'train/grad_norm_pre_clip': 61.20532989501953, 'timing/data': 0.0003622800577431917, 'timing/model': 1.1236240810249, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.4}) +[2026-06-23 21:48:02,374][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.2715706527233124, 'train/grad_norm_pre_clip': 31.412403106689453, 'timing/data': 0.0004147919826209545, 'timing/model': 1.1193088409490883, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.4}) +[2026-06-23 21:48:06,634][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.052489183843135834, 'train/grad_norm_pre_clip': 14.287529945373535, 'timing/data': 0.00041364890057593584, 'timing/model': 1.1177865179488435, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.4}) +[2026-06-23 21:48:10,728][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.09719748795032501, 'train/grad_norm_pre_clip': 21.830488204956055, 'timing/data': 0.0004915000172331929, 'timing/model': 1.1180308780167252, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.4}) +[2026-06-23 21:48:14,801][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.22100241482257843, 'train/grad_norm_pre_clip': 14.780938148498535, 'timing/data': 0.0005732449935749173, 'timing/model': 1.1176615660078824, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.4}) +[2026-06-23 21:48:18,878][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.08865787833929062, 'train/grad_norm_pre_clip': 12.182799339294434, 'timing/data': 0.0003669880097731948, 'timing/model': 1.1182215629378334, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.4}) +[2026-06-23 21:48:23,037][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.1035318449139595, 'train/grad_norm_pre_clip': 12.384745597839355, 'timing/data': 0.0003781520063057542, 'timing/model': 1.1177660450339317, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.4}) +[2026-06-23 21:48:27,113][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.016686351969838142, 'train/grad_norm_pre_clip': 15.524441719055176, 'timing/data': 0.0003641020739451051, 'timing/model': 1.1210751979378983, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.4}) +[2026-06-23 21:48:31,360][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.10655449330806732, 'train/grad_norm_pre_clip': 15.578835487365723, 'timing/data': 0.0006559190806001425, 'timing/model': 1.2819341219728813, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.4}) +[2026-06-23 21:48:35,437][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.04940013214945793, 'train/grad_norm_pre_clip': 13.86376953125, 'timing/data': 0.0003565669758245349, 'timing/model': 1.116683056927286, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.4}) +[2026-06-23 21:48:39,600][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.03573981299996376, 'train/grad_norm_pre_clip': 21.698999404907227, 'timing/data': 0.0003772119525820017, 'timing/model': 1.1194865889847279, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.4}) +[2026-06-23 21:48:43,687][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.091420479118824, 'train/grad_norm_pre_clip': 17.963457107543945, 'timing/data': 0.00035625207237899303, 'timing/model': 1.1224112840136513, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.4}) +[2026-06-23 21:48:47,771][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.014421990141272545, 'train/grad_norm_pre_clip': 200.91683959960938, 'timing/data': 0.0003528560046106577, 'timing/model': 1.1184515570057556, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.4}) +[2026-06-23 21:48:51,858][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.006199591793119907, 'train/grad_norm_pre_clip': 22.524024963378906, 'timing/data': 0.0004324520705267787, 'timing/model': 1.1195333999348804, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.4}) +[2026-06-23 21:48:56,077][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.057805970311164856, 'train/grad_norm_pre_clip': 32.45030212402344, 'timing/data': 0.00036734191235154867, 'timing/model': 1.1174111049622297, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.4}) +[2026-06-23 21:49:00,165][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.04284697398543358, 'train/grad_norm_pre_clip': 23.563648223876953, 'timing/data': 0.0004307309864088893, 'timing/model': 1.121658495045267, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.4}) +[2026-06-23 21:49:04,233][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.13130322098731995, 'train/grad_norm_pre_clip': 27.636127471923828, 'timing/data': 0.0005580330034717917, 'timing/model': 1.1174717739922926, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.4}) +[2026-06-23 21:49:08,416][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.10688358545303345, 'train/grad_norm_pre_clip': 21.881301879882812, 'timing/data': 0.0005508930189535022, 'timing/model': 1.2268350119702518, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.4}) +[2026-06-23 21:49:12,497][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.028611283749341965, 'train/grad_norm_pre_clip': 12.77402114868164, 'timing/data': 0.0004202940035611391, 'timing/model': 1.1181209239875898, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.4}) +[2026-06-23 21:49:16,575][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.015946010127663612, 'train/grad_norm_pre_clip': 15.282888412475586, 'timing/data': 0.00035582599230110645, 'timing/model': 1.1201403080485761, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.41}) +[2026-06-23 21:49:20,649][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.14478303492069244, 'train/grad_norm_pre_clip': 30.352304458618164, 'timing/data': 0.0003747319569811225, 'timing/model': 1.1164336620131508, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.41}) +[2026-06-23 21:49:24,898][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.09091944992542267, 'train/grad_norm_pre_clip': 41.08366012573242, 'timing/data': 0.0005476119695231318, 'timing/model': 1.277067897026427, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.41}) +[2026-06-23 21:49:28,974][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.020237848162651062, 'train/grad_norm_pre_clip': 27.404863357543945, 'timing/data': 0.0004174530040472746, 'timing/model': 1.1185617370065302, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.41}) +[2026-06-23 21:49:33,058][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.09279924631118774, 'train/grad_norm_pre_clip': 39.263980865478516, 'timing/data': 0.00036222394555807114, 'timing/model': 1.1205183550482616, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.41}) +[2026-06-23 21:49:37,130][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.1552709937095642, 'train/grad_norm_pre_clip': 40.59526443481445, 'timing/data': 0.00035394506994634867, 'timing/model': 1.1234296871116385, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.41}) +[2026-06-23 21:49:41,303][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.28823789954185486, 'train/grad_norm_pre_clip': 34.018280029296875, 'timing/data': 0.0005939089460298419, 'timing/model': 1.1167716089403257, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.41}) +[2026-06-23 21:49:45,388][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.008212218061089516, 'train/grad_norm_pre_clip': 28.66427230834961, 'timing/data': 0.00046877702698111534, 'timing/model': 1.1171021279878914, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.41}) +[2026-06-23 21:49:49,478][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.06911299377679825, 'train/grad_norm_pre_clip': 23.818737030029297, 'timing/data': 0.0003677599597722292, 'timing/model': 1.11989104305394, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.41}) +[2026-06-23 21:49:53,613][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.21593157947063446, 'train/grad_norm_pre_clip': 51.53030776977539, 'timing/data': 0.00035581900738179684, 'timing/model': 1.1173449299531057, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.41}) +[2026-06-23 21:49:57,798][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.015385390259325504, 'train/grad_norm_pre_clip': 50.56857681274414, 'timing/data': 0.00036615203134715557, 'timing/model': 1.1238678409717977, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.41}) +[2026-06-23 21:50:01,871][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.06102082505822182, 'train/grad_norm_pre_clip': 27.303871154785156, 'timing/data': 0.0004156130598857999, 'timing/model': 1.1206750610144809, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.41}) +[2026-06-23 21:50:05,968][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.024684125557541847, 'train/grad_norm_pre_clip': 9.788069725036621, 'timing/data': 0.00035393110010772943, 'timing/model': 1.122441068990156, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.41}) +[2026-06-23 21:50:10,050][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.07034631073474884, 'train/grad_norm_pre_clip': 24.295772552490234, 'timing/data': 0.00037561997305601835, 'timing/model': 1.1205386130604893, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.41}) +[2026-06-23 21:50:14,275][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.1571260392665863, 'train/grad_norm_pre_clip': 13.11108684539795, 'timing/data': 0.0004093180177733302, 'timing/model': 1.1214069409761578, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.41}) +[2026-06-23 21:50:18,362][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.11706878989934921, 'train/grad_norm_pre_clip': 19.4599609375, 'timing/data': 0.0004139620577916503, 'timing/model': 1.1228183729108423, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.41}) +[2026-06-23 21:50:22,435][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.10999169945716858, 'train/grad_norm_pre_clip': 14.146157264709473, 'timing/data': 0.00036198797170072794, 'timing/model': 1.1177450970280915, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.41}) +[2026-06-23 21:50:26,527][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.07203474640846252, 'train/grad_norm_pre_clip': 25.058374404907227, 'timing/data': 0.0003603229997679591, 'timing/model': 1.1225149480160326, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.41}) +[2026-06-23 21:50:30,726][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.0650523230433464, 'train/grad_norm_pre_clip': 14.473308563232422, 'timing/data': 0.0003585569793358445, 'timing/model': 1.1223195970524102, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.41}) +[2026-06-23 21:50:34,820][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.06023482605814934, 'train/grad_norm_pre_clip': 19.787864685058594, 'timing/data': 0.0003614820307120681, 'timing/model': 1.116921867011115, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.41}) +[2026-06-23 21:50:38,920][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.026349293068051338, 'train/grad_norm_pre_clip': 12.372855186462402, 'timing/data': 0.00035830610431730747, 'timing/model': 1.1360248039709404, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.41}) +[2026-06-23 21:50:43,007][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.06463172286748886, 'train/grad_norm_pre_clip': 37.96784210205078, 'timing/data': 0.0005116830579936504, 'timing/model': 1.1194250429980457, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.41}) +[2026-06-23 21:50:47,167][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.19860738515853882, 'train/grad_norm_pre_clip': 30.073867797851562, 'timing/data': 0.0005673749838024378, 'timing/model': 1.1222544990014285, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.41}) +[2026-06-23 21:50:51,259][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.07159774005413055, 'train/grad_norm_pre_clip': 17.797374725341797, 'timing/data': 0.0003653520252555609, 'timing/model': 1.1206483729183674, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.41}) +[2026-06-23 21:50:55,335][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.2065148651599884, 'train/grad_norm_pre_clip': 23.375408172607422, 'timing/data': 0.0005653500556945801, 'timing/model': 1.1176572650438175, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.41}) +[2026-06-23 21:50:59,411][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.12321029603481293, 'train/grad_norm_pre_clip': 13.574862480163574, 'timing/data': 0.00046244903933256865, 'timing/model': 1.1138951600296423, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.41}) +[2026-06-23 21:51:03,662][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.07424167543649673, 'train/grad_norm_pre_clip': 31.554031372070312, 'timing/data': 0.00036343804094940424, 'timing/model': 1.1175391229335219, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.42}) +[2026-06-23 21:51:07,743][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.04778975993394852, 'train/grad_norm_pre_clip': 20.297985076904297, 'timing/data': 0.0003619389608502388, 'timing/model': 1.119287946028635, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.42}) +[2026-06-23 21:51:11,819][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.11023147404193878, 'train/grad_norm_pre_clip': 31.395347595214844, 'timing/data': 0.000371718080714345, 'timing/model': 1.1164197570178658, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.42}) +[2026-06-23 21:51:15,898][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.04510752111673355, 'train/grad_norm_pre_clip': 27.5653076171875, 'timing/data': 0.00036221800837665796, 'timing/model': 1.1211580659728497, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.42}) +[2026-06-23 21:51:20,115][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.015200725756585598, 'train/grad_norm_pre_clip': 48.3954963684082, 'timing/data': 0.00036670302506536245, 'timing/model': 1.1284110189881176, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.42}) +[2026-06-23 21:51:24,216][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.054472774267196655, 'train/grad_norm_pre_clip': 42.98311233520508, 'timing/data': 0.00036620604805648327, 'timing/model': 1.1177285220474005, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.42}) +[2026-06-23 21:51:28,305][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.17211824655532837, 'train/grad_norm_pre_clip': 30.561365127563477, 'timing/data': 0.0003568199463188648, 'timing/model': 1.115507717942819, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.42}) +[2026-06-23 21:51:32,386][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.18143603205680847, 'train/grad_norm_pre_clip': 26.63549041748047, 'timing/data': 0.0006449909415096045, 'timing/model': 1.1209093899233267, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.42}) +[2026-06-23 21:51:36,651][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.02699476294219494, 'train/grad_norm_pre_clip': 16.78807830810547, 'timing/data': 0.0003744929563254118, 'timing/model': 1.1238805609755218, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.42}) +[2026-06-23 21:51:40,754][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.09097957611083984, 'train/grad_norm_pre_clip': 19.35086441040039, 'timing/data': 0.0003651370061561465, 'timing/model': 1.1195837550330907, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.42}) +[2026-06-23 21:51:44,837][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.15396243333816528, 'train/grad_norm_pre_clip': 21.261877059936523, 'timing/data': 0.00037254299968481064, 'timing/model': 1.1163642390165478, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.42}) +[2026-06-23 21:51:48,920][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.09944140911102295, 'train/grad_norm_pre_clip': 21.85231590270996, 'timing/data': 0.00035281002055853605, 'timing/model': 1.1205160160316154, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.42}) +[2026-06-23 21:51:53,076][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.0985829085111618, 'train/grad_norm_pre_clip': 20.260997772216797, 'timing/data': 0.0003579530166462064, 'timing/model': 1.1240881909616292, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.42}) +[2026-06-23 21:51:57,183][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.09527131170034409, 'train/grad_norm_pre_clip': 10.922268867492676, 'timing/data': 0.00036663003265857697, 'timing/model': 1.1245647630421445, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.42}) +[2026-06-23 21:52:01,258][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.21167302131652832, 'train/grad_norm_pre_clip': 39.217594146728516, 'timing/data': 0.00035694404505193233, 'timing/model': 1.1201279329834506, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.42}) +[2026-06-23 21:52:05,330][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.033056799322366714, 'train/grad_norm_pre_clip': 28.48603630065918, 'timing/data': 0.000567721901461482, 'timing/model': 1.113516186014749, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.42}) +[2026-06-23 21:52:09,480][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.08494285494089127, 'train/grad_norm_pre_clip': 33.840606689453125, 'timing/data': 0.000440418953076005, 'timing/model': 1.11847695405595, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.42}) +[2026-06-23 21:52:13,553][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.02826867438852787, 'train/grad_norm_pre_clip': 11.530122756958008, 'timing/data': 0.0003550749970600009, 'timing/model': 1.1180908919777721, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.42}) +[2026-06-23 21:52:17,625][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.05508365109562874, 'train/grad_norm_pre_clip': 15.857698440551758, 'timing/data': 0.0005527279572561383, 'timing/model': 1.1160659310407937, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.42}) +[2026-06-23 21:52:21,712][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.1629151850938797, 'train/grad_norm_pre_clip': 45.219451904296875, 'timing/data': 0.0004490229766815901, 'timing/model': 1.1186391399241984, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.42}) +[2026-06-23 21:52:25,935][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.08504900336265564, 'train/grad_norm_pre_clip': 14.252618789672852, 'timing/data': 0.0006471140077337623, 'timing/model': 1.1175140409031883, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.42}) +[2026-06-23 21:52:30,028][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.043357934802770615, 'train/grad_norm_pre_clip': 22.601512908935547, 'timing/data': 0.00039541604928672314, 'timing/model': 1.1181522599654272, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.42}) +[2026-06-23 21:52:34,113][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.1679476499557495, 'train/grad_norm_pre_clip': 28.18531036376953, 'timing/data': 0.0005042229313403368, 'timing/model': 1.1185539060970768, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.42}) +[2026-06-23 21:52:38,299][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.10262696444988251, 'train/grad_norm_pre_clip': 32.61524200439453, 'timing/data': 0.0005994950188323855, 'timing/model': 1.2145709190517664, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.42}) +[2026-06-23 21:52:42,385][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.005047110375016928, 'train/grad_norm_pre_clip': 33.15467834472656, 'timing/data': 0.0003635069588199258, 'timing/model': 1.1171721590217203, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.42}) +[2026-06-23 21:52:46,495][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.17255331575870514, 'train/grad_norm_pre_clip': 24.671937942504883, 'timing/data': 0.00036384095437824726, 'timing/model': 1.1222552160033956, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.42}) +[2026-06-23 21:52:50,580][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.18249191343784332, 'train/grad_norm_pre_clip': 19.032276153564453, 'timing/data': 0.00037206895649433136, 'timing/model': 1.1165090190479532, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.43}) +[2026-06-23 21:52:54,840][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1100 +[2026-06-23 21:53:58,361][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1100 +[2026-06-23 21:53:58,363][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1100 +[2026-06-23 21:54:38,004][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1100 +[2026-06-23 21:54:38,005][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.0876130685210228, 'train/grad_norm_pre_clip': 29.44956398010254, 'eval/action_loss/seconds': 63.51771677297074, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.05736162140965462, 'eval/latency_2/loss': 0.05736162140965462, 'eval/flappy/loss': 0.05736162140965462, 'eval/flappy/latency_2/loss': 0.05736162140965462, 'eval/flap_precision': 0.9078282828282829, 'eval/flap_recall': 0.9755766621438263, 'eval/flap_f1': 0.9404839764551995, 'eval/flap_rate_model': 0.22, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9404839764551995, 'eval/latency_2/flap_precision': 0.9078282828282829, 'eval/latency_2/flap_recall': 0.9755766621438263, 'eval/latency_2/flap_f1': 0.9404839764551995, 'eval/latency_2/flap_rate_model': 0.22, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9404839764551995, 'eval/action_classification/data_wait_seconds': 0.3793719410896301, 'eval/action_classification/predict_seconds': 39.10052490234375, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0004059700295329094, 'timing/model': 1.2979374580318108, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.43}) +[2026-06-23 21:54:42,040][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.05024182051420212, 'train/grad_norm_pre_clip': 15.395800590515137, 'timing/data': 0.00035805103834718466, 'timing/model': 1.1090049130143598, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.43}) +[2026-06-23 21:54:46,068][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.23349016904830933, 'train/grad_norm_pre_clip': 23.04471206665039, 'timing/data': 0.0005622510798275471, 'timing/model': 1.1015842279884964, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.43}) +[2026-06-23 21:54:50,262][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.07444654405117035, 'train/grad_norm_pre_clip': 10.057430267333984, 'timing/data': 0.00036229402758181095, 'timing/model': 1.2443950719898567, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.43}) +[2026-06-23 21:54:54,313][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.10329161584377289, 'train/grad_norm_pre_clip': 6.870567798614502, 'timing/data': 0.0003657200140878558, 'timing/model': 1.113784527988173, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.43}) +[2026-06-23 21:54:58,364][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.023677930235862732, 'train/grad_norm_pre_clip': 19.64169692993164, 'timing/data': 0.0003663100069388747, 'timing/model': 1.1139992059906945, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.43}) +[2026-06-23 21:55:02,414][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.01148796547204256, 'train/grad_norm_pre_clip': 21.68754005432129, 'timing/data': 0.00035631703212857246, 'timing/model': 1.1134758489206433, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.43}) +[2026-06-23 21:55:06,564][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.025272663682699203, 'train/grad_norm_pre_clip': 24.08869743347168, 'timing/data': 0.0003668790450319648, 'timing/model': 1.11277682101354, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.43}) +[2026-06-23 21:55:10,626][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.03212204575538635, 'train/grad_norm_pre_clip': 22.701786041259766, 'timing/data': 0.0004330060910433531, 'timing/model': 1.112879463005811, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.43}) +[2026-06-23 21:55:14,688][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.032207343727350235, 'train/grad_norm_pre_clip': 19.979833602905273, 'timing/data': 0.0006013719830662012, 'timing/model': 1.1150668960763142, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.43}) +[2026-06-23 21:55:18,754][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.11414874345064163, 'train/grad_norm_pre_clip': 33.30836486816406, 'timing/data': 0.00041707197669893503, 'timing/model': 1.1137107190443203, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.43}) +[2026-06-23 21:55:22,892][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.025081299245357513, 'train/grad_norm_pre_clip': 12.862478256225586, 'timing/data': 0.0005590819055214524, 'timing/model': 1.1142680239863694, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.43}) +[2026-06-23 21:55:26,952][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.4136967658996582, 'train/grad_norm_pre_clip': 28.102561950683594, 'timing/data': 0.0003667899873107672, 'timing/model': 1.1135085249552503, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.43}) +[2026-06-23 21:55:31,026][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.05028199404478073, 'train/grad_norm_pre_clip': 30.56224250793457, 'timing/data': 0.0003557999152690172, 'timing/model': 1.1141871470026672, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.43}) +[2026-06-23 21:55:35,085][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.09494507312774658, 'train/grad_norm_pre_clip': 19.986282348632812, 'timing/data': 0.0004238580586388707, 'timing/model': 1.1151466970331967, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.43}) +[2026-06-23 21:55:39,337][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.17067976295948029, 'train/grad_norm_pre_clip': 46.0485725402832, 'timing/data': 0.0005665919743478298, 'timing/model': 1.1228917880216613, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.43}) +[2026-06-23 21:55:43,408][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.0806736871600151, 'train/grad_norm_pre_clip': 24.236127853393555, 'timing/data': 0.00035290291998535395, 'timing/model': 1.1136708030244336, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.43}) +[2026-06-23 21:55:47,491][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.08511863648891449, 'train/grad_norm_pre_clip': 18.500598907470703, 'timing/data': 0.000358636025339365, 'timing/model': 1.1111478139646351, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.43}) +[2026-06-23 21:55:51,580][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.03320576995611191, 'train/grad_norm_pre_clip': 19.879980087280273, 'timing/data': 0.00035812705755233765, 'timing/model': 1.1162351339589804, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.43}) +[2026-06-23 21:55:55,776][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.15031012892723083, 'train/grad_norm_pre_clip': 25.47946548461914, 'timing/data': 0.0003692680038511753, 'timing/model': 1.1097463140031323, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.43}) +[2026-06-23 21:55:59,854][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.15121008455753326, 'train/grad_norm_pre_clip': 23.093814849853516, 'timing/data': 0.00036278797779232264, 'timing/model': 1.1182033428922296, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.43}) +[2026-06-23 21:56:03,929][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.13282397389411926, 'train/grad_norm_pre_clip': 25.2175350189209, 'timing/data': 0.000362824066542089, 'timing/model': 1.1172442199895158, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.43}) +[2026-06-23 21:56:08,004][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.0977933257818222, 'train/grad_norm_pre_clip': 24.871623992919922, 'timing/data': 0.00037619308568537235, 'timing/model': 1.116673220996745, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.43}) +[2026-06-23 21:56:12,259][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.015179860405623913, 'train/grad_norm_pre_clip': 16.857501983642578, 'timing/data': 0.00040015997365117073, 'timing/model': 1.2233542780159041, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.43}) +[2026-06-23 21:56:16,338][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.03023809939622879, 'train/grad_norm_pre_clip': 23.259252548217773, 'timing/data': 0.00036180007737129927, 'timing/model': 1.1191507959738374, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.44}) +[2026-06-23 21:56:20,411][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.03492370992898941, 'train/grad_norm_pre_clip': 7.861490726470947, 'timing/data': 0.0003689948935061693, 'timing/model': 1.1098005609819666, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.44}) +[2026-06-23 21:56:24,491][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.08238651603460312, 'train/grad_norm_pre_clip': 12.337509155273438, 'timing/data': 0.00036049901973456144, 'timing/model': 1.117783286026679, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.44}) +[2026-06-23 21:56:28,563][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.19271846115589142, 'train/grad_norm_pre_clip': 20.513578414916992, 'timing/data': 0.00035643496084958315, 'timing/model': 1.1142987050116062, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.44}) +[2026-06-23 21:56:32,818][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.03306068852543831, 'train/grad_norm_pre_clip': 16.58061981201172, 'timing/data': 0.0003566750092431903, 'timing/model': 1.1185420269612223, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.44}) +[2026-06-23 21:56:36,900][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.058356691151857376, 'train/grad_norm_pre_clip': 29.900699615478516, 'timing/data': 0.00036246899981051683, 'timing/model': 1.1119188980665058, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.44}) +[2026-06-23 21:56:40,976][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.07451499253511429, 'train/grad_norm_pre_clip': 13.390839576721191, 'timing/data': 0.00044561701361089945, 'timing/model': 1.1207329900935292, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.44}) +[2026-06-23 21:56:45,061][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.04791156202554703, 'train/grad_norm_pre_clip': 14.603876113891602, 'timing/data': 0.00036169204395264387, 'timing/model': 1.1198177760234103, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.44}) +[2026-06-23 21:56:49,228][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.16199946403503418, 'train/grad_norm_pre_clip': 43.83102798461914, 'timing/data': 0.00037297792732715607, 'timing/model': 1.1150084189139307, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.44}) +[2026-06-23 21:56:53,312][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.1329929679632187, 'train/grad_norm_pre_clip': 23.551977157592773, 'timing/data': 0.00035037705674767494, 'timing/model': 1.117753553087823, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.44}) +[2026-06-23 21:56:57,392][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.15191112458705902, 'train/grad_norm_pre_clip': 16.882102966308594, 'timing/data': 0.0005122889997437596, 'timing/model': 1.1170287060085684, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.44}) +[2026-06-23 21:57:01,464][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.010169689543545246, 'train/grad_norm_pre_clip': 30.780561447143555, 'timing/data': 0.000358566059730947, 'timing/model': 1.1173001759452745, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.44}) +[2026-06-23 21:57:05,613][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.08953831344842911, 'train/grad_norm_pre_clip': 16.24527359008789, 'timing/data': 0.00042988499626517296, 'timing/model': 1.1177576950285584, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.44}) +[2026-06-23 21:57:09,702][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.07178517431020737, 'train/grad_norm_pre_clip': 21.523221969604492, 'timing/data': 0.0003617219626903534, 'timing/model': 1.1213371810736135, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.44}) +[2026-06-23 21:57:13,776][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.16810157895088196, 'train/grad_norm_pre_clip': 32.52886199951172, 'timing/data': 0.00037493009585887194, 'timing/model': 1.1121788180898875, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.44}) +[2026-06-23 21:57:17,858][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.12865696847438812, 'train/grad_norm_pre_clip': 36.469078063964844, 'timing/data': 0.000396400922909379, 'timing/model': 1.1177843799814582, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.44}) +[2026-06-23 21:57:22,008][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.14212791621685028, 'train/grad_norm_pre_clip': 18.194799423217773, 'timing/data': 0.00037231098394840956, 'timing/model': 1.119331095018424, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.44}) +[2026-06-23 21:57:26,093][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.04512621462345123, 'train/grad_norm_pre_clip': 27.240745544433594, 'timing/data': 0.00059407704975456, 'timing/model': 1.1179433970246464, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.44}) +[2026-06-23 21:57:30,174][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.0978679433465004, 'train/grad_norm_pre_clip': 24.682546615600586, 'timing/data': 0.00035751296672970057, 'timing/model': 1.1198056410066783, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.44}) +[2026-06-23 21:57:34,257][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.22055383026599884, 'train/grad_norm_pre_clip': 23.600343704223633, 'timing/data': 0.00035679806023836136, 'timing/model': 1.1218634030083194, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.44}) +[2026-06-23 21:57:38,467][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.18100115656852722, 'train/grad_norm_pre_clip': 20.00300407409668, 'timing/data': 0.0005422759568318725, 'timing/model': 1.1166953060310334, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.44}) +[2026-06-23 21:57:42,595][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.12396112084388733, 'train/grad_norm_pre_clip': 40.17900466918945, 'timing/data': 0.0005869329907000065, 'timing/model': 1.1194983429741114, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.44}) +[2026-06-23 21:57:46,685][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.08725611865520477, 'train/grad_norm_pre_clip': 30.197214126586914, 'timing/data': 0.0005848150467500091, 'timing/model': 1.1173797050723806, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.44}) +[2026-06-23 21:57:50,867][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.1225917786359787, 'train/grad_norm_pre_clip': 21.768686294555664, 'timing/data': 0.0004926020046696067, 'timing/model': 1.2125586989568546, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.44}) +[2026-06-23 21:57:54,944][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.030077798292040825, 'train/grad_norm_pre_clip': 11.413055419921875, 'timing/data': 0.00036295398604124784, 'timing/model': 1.118637294974178, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.44}) +[2026-06-23 21:57:59,026][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.06338631361722946, 'train/grad_norm_pre_clip': 22.53559112548828, 'timing/data': 0.00036006001755595207, 'timing/model': 1.1196347499499097, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.44}) +[2026-06-23 21:58:03,112][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.19834338128566742, 'train/grad_norm_pre_clip': 37.580535888671875, 'timing/data': 0.0003766670124605298, 'timing/model': 1.1245026619872078, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.45}) +[2026-06-23 21:58:07,363][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.11632636189460754, 'train/grad_norm_pre_clip': 13.602190017700195, 'timing/data': 0.0003617359325289726, 'timing/model': 1.2988475420279428, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.45}) +[2026-06-23 21:58:11,439][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.054843831807374954, 'train/grad_norm_pre_clip': 20.235763549804688, 'timing/data': 0.0003777090460062027, 'timing/model': 1.1213326569413766, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.45}) +[2026-06-23 21:58:15,529][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.018218528479337692, 'train/grad_norm_pre_clip': 7.961071014404297, 'timing/data': 0.00035382003989070654, 'timing/model': 1.1173600839683786, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.45}) +[2026-06-23 21:58:19,612][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.21747945249080658, 'train/grad_norm_pre_clip': 24.598703384399414, 'timing/data': 0.0003676280612125993, 'timing/model': 1.1176640369230881, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.45}) +[2026-06-23 21:58:23,796][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.1506480723619461, 'train/grad_norm_pre_clip': 50.60526657104492, 'timing/data': 0.0003607269609346986, 'timing/model': 1.1184320190222934, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.45}) +[2026-06-23 21:58:27,880][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.40698733925819397, 'train/grad_norm_pre_clip': 52.537452697753906, 'timing/data': 0.0007079459028318524, 'timing/model': 1.1195785839809105, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.45}) +[2026-06-23 21:58:31,968][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.1049242615699768, 'train/grad_norm_pre_clip': 40.881317138671875, 'timing/data': 0.00037974503356963396, 'timing/model': 1.1193218929693103, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.45}) +[2026-06-23 21:58:36,064][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.11780236661434174, 'train/grad_norm_pre_clip': 25.05290412902832, 'timing/data': 0.000409828033298254, 'timing/model': 1.1189435560954735, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.45}) +[2026-06-23 21:58:40,241][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.052396852523088455, 'train/grad_norm_pre_clip': 39.45039367675781, 'timing/data': 0.0004059989005327225, 'timing/model': 1.124283420969732, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.45}) +[2026-06-23 21:58:44,327][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.084819495677948, 'train/grad_norm_pre_clip': 23.211151123046875, 'timing/data': 0.0004404329229146242, 'timing/model': 1.117331099929288, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.45}) +[2026-06-23 21:58:48,415][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.0911729484796524, 'train/grad_norm_pre_clip': 21.71382713317871, 'timing/data': 0.00044382293708622456, 'timing/model': 1.1162028680555522, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.45}) +[2026-06-23 21:58:52,489][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.09696868807077408, 'train/grad_norm_pre_clip': 18.323091506958008, 'timing/data': 0.0003747789887711406, 'timing/model': 1.113433600985445, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.45}) +[2026-06-23 21:58:56,705][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.13673174381256104, 'train/grad_norm_pre_clip': 16.640392303466797, 'timing/data': 0.0003618180053308606, 'timing/model': 1.1211928949924186, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.45}) +[2026-06-23 21:59:00,773][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.23279382288455963, 'train/grad_norm_pre_clip': 21.17119789123535, 'timing/data': 0.000368052045814693, 'timing/model': 1.1149577329633757, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.45}) +[2026-06-23 21:59:04,870][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.05899745970964432, 'train/grad_norm_pre_clip': 15.54216480255127, 'timing/data': 0.0004363220650702715, 'timing/model': 1.1242147190496325, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.45}) +[2026-06-23 21:59:08,959][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.1461266130208969, 'train/grad_norm_pre_clip': 18.267597198486328, 'timing/data': 0.00036480696871876717, 'timing/model': 1.1253598870243877, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.45}) +[2026-06-23 21:59:13,160][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.03350174054503441, 'train/grad_norm_pre_clip': 12.663472175598145, 'timing/data': 0.0003639019560068846, 'timing/model': 1.1219110700767487, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.45}) +[2026-06-23 21:59:17,255][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.03446228802204132, 'train/grad_norm_pre_clip': 13.87336540222168, 'timing/data': 0.00034751195926219225, 'timing/model': 1.1231725270627066, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.45}) +[2026-06-23 21:59:21,329][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.14253464341163635, 'train/grad_norm_pre_clip': 20.202608108520508, 'timing/data': 0.0006608800031244755, 'timing/model': 1.1158337439410388, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.45}) +[2026-06-23 21:59:25,412][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.009976062923669815, 'train/grad_norm_pre_clip': 11.701468467712402, 'timing/data': 0.0005699570756405592, 'timing/model': 1.1215405890252441, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.45}) +[2026-06-23 21:59:29,601][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.07294141501188278, 'train/grad_norm_pre_clip': 9.820480346679688, 'timing/data': 0.0005665509961545467, 'timing/model': 1.13296335702762, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.45}) +[2026-06-23 21:59:33,694][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.08320482820272446, 'train/grad_norm_pre_clip': 25.90988540649414, 'timing/data': 0.0004346009809523821, 'timing/model': 1.1162662180140615, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.45}) +[2026-06-23 21:59:37,792][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.29003462195396423, 'train/grad_norm_pre_clip': 17.529741287231445, 'timing/data': 0.0004074530443176627, 'timing/model': 1.134484141948633, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.45}) +[2026-06-23 21:59:41,879][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.13491465151309967, 'train/grad_norm_pre_clip': 16.112125396728516, 'timing/data': 0.0006807669997215271, 'timing/model': 1.119521555956453, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.45}) +[2026-06-23 21:59:46,151][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.21449622511863708, 'train/grad_norm_pre_clip': 38.15149688720703, 'timing/data': 0.000649568042717874, 'timing/model': 1.1216060420265421, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.45}) +[2026-06-23 21:59:50,237][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.016069971024990082, 'train/grad_norm_pre_clip': 16.071199417114258, 'timing/data': 0.0003710719756782055, 'timing/model': 1.1186898481100798, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.46}) +[2026-06-23 21:59:54,328][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.13499003648757935, 'train/grad_norm_pre_clip': 48.58368682861328, 'timing/data': 0.0004438789328560233, 'timing/model': 1.1190884809475392, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.46}) +[2026-06-23 21:59:58,415][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.12836256623268127, 'train/grad_norm_pre_clip': 21.703235626220703, 'timing/data': 0.000525765004567802, 'timing/model': 1.1172424010001123, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.46}) +[2026-06-23 22:00:02,595][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.01802890934050083, 'train/grad_norm_pre_clip': 19.063037872314453, 'timing/data': 0.0004842189373448491, 'timing/model': 1.1147170680342242, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.46}) +[2026-06-23 22:00:06,681][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.027893420308828354, 'train/grad_norm_pre_clip': 24.60691261291504, 'timing/data': 0.000355428084731102, 'timing/model': 1.114856980042532, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.46}) +[2026-06-23 22:00:10,786][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.10522757470607758, 'train/grad_norm_pre_clip': 42.497314453125, 'timing/data': 0.0005833170143887401, 'timing/model': 1.115905515034683, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.46}) +[2026-06-23 22:00:14,894][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.14050981402397156, 'train/grad_norm_pre_clip': 23.50737953186035, 'timing/data': 0.0004727350315079093, 'timing/model': 1.116368665941991, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.46}) +[2026-06-23 22:00:19,145][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.09932244569063187, 'train/grad_norm_pre_clip': 34.15007019042969, 'timing/data': 0.00035894999746233225, 'timing/model': 1.117709034937434, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.46}) +[2026-06-23 22:00:23,225][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.13023683428764343, 'train/grad_norm_pre_clip': 17.954835891723633, 'timing/data': 0.00047011394053697586, 'timing/model': 1.1200973299564794, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.46}) +[2026-06-23 22:00:27,304][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.002612100215628743, 'train/grad_norm_pre_clip': 17.019786834716797, 'timing/data': 0.0005118140252307057, 'timing/model': 1.1168341560987756, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.46}) +[2026-06-23 22:00:31,399][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.02246748097240925, 'train/grad_norm_pre_clip': 15.207077980041504, 'timing/data': 0.0005661959294229746, 'timing/model': 1.1167845060117543, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.46}) +[2026-06-23 22:00:35,565][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.02861020527780056, 'train/grad_norm_pre_clip': 19.348543167114258, 'timing/data': 0.00043403508607298136, 'timing/model': 1.1227589109912515, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.46}) +[2026-06-23 22:00:39,649][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.08742105215787888, 'train/grad_norm_pre_clip': 19.4227352142334, 'timing/data': 0.00041018007323145866, 'timing/model': 1.1234408530872315, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.46}) +[2026-06-23 22:00:43,747][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.029530217871069908, 'train/grad_norm_pre_clip': 15.938645362854004, 'timing/data': 0.00036457309033721685, 'timing/model': 1.1212252830155194, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.46}) +[2026-06-23 22:00:47,844][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.17242549359798431, 'train/grad_norm_pre_clip': 25.29244041442871, 'timing/data': 0.00037022808101028204, 'timing/model': 1.1289051589556038, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.46}) +[2026-06-23 22:00:52,117][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.03525136038661003, 'train/grad_norm_pre_clip': 16.138456344604492, 'timing/data': 0.00036154501140117645, 'timing/model': 1.1323311999440193, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.46}) +[2026-06-23 22:00:56,209][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.13242797553539276, 'train/grad_norm_pre_clip': 18.977481842041016, 'timing/data': 0.0003529610112309456, 'timing/model': 1.118323688977398, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.46}) +[2026-06-23 22:01:00,294][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.0047088488936424255, 'train/grad_norm_pre_clip': 17.316274642944336, 'timing/data': 0.0003633189480751753, 'timing/model': 1.1232183669926599, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.46}) +[2026-06-23 22:01:04,397][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.08430152386426926, 'train/grad_norm_pre_clip': 11.861180305480957, 'timing/data': 0.0003640239592641592, 'timing/model': 1.1300117620266974, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.46}) +[2026-06-23 22:01:08,657][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.011923401616513729, 'train/grad_norm_pre_clip': 62.84867477416992, 'timing/data': 0.0003694209735840559, 'timing/model': 1.1251659380504861, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.46}) +[2026-06-23 22:01:12,750][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.031592514365911484, 'train/grad_norm_pre_clip': 23.406099319458008, 'timing/data': 0.0003561230842024088, 'timing/model': 1.124803913058713, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.46}) +[2026-06-23 22:01:16,836][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.03431493043899536, 'train/grad_norm_pre_clip': 11.903083801269531, 'timing/data': 0.000523335998877883, 'timing/model': 1.1173983750632033, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.46}) +[2026-06-23 22:01:20,941][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.2368870973587036, 'train/grad_norm_pre_clip': 13.802583694458008, 'timing/data': 0.0003863680176436901, 'timing/model': 1.121932215988636, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.46}) +[2026-06-23 22:01:25,209][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.017255935817956924, 'train/grad_norm_pre_clip': 18.940706253051758, 'timing/data': 0.00036667799577116966, 'timing/model': 1.1214305920293555, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.46}) +[2026-06-23 22:01:29,287][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1200 +[2026-06-23 22:02:33,541][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1200 +[2026-06-23 22:02:33,543][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1200 +[2026-06-23 22:03:12,391][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1200 +[2026-06-23 22:03:12,391][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.30503904819488525, 'train/grad_norm_pre_clip': 19.72089195251465, 'eval/action_loss/seconds': 64.25217128102668, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.05069531500339508, 'eval/latency_2/loss': 0.05069531500339508, 'eval/flappy/loss': 0.05069531500339508, 'eval/flappy/latency_2/loss': 0.05069531500339508, 'eval/flap_precision': 0.925974025974026, 'eval/flap_recall': 0.9674355495251018, 'eval/flap_f1': 0.9462508294625084, 'eval/flap_rate_model': 0.21388888888888888, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9462508294625084, 'eval/latency_2/flap_precision': 0.925974025974026, 'eval/latency_2/flap_recall': 0.9674355495251018, 'eval/latency_2/flap_f1': 0.9462508294625084, 'eval/latency_2/flap_rate_model': 0.21388888888888888, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9462508294625084, 'eval/action_classification/data_wait_seconds': 0.6128148436546326, 'eval/action_classification/predict_seconds': 38.10570526123047, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0003687069984152913, 'timing/model': 1.1147493739845231, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.46}) +[2026-06-23 22:03:16,451][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.11917980015277863, 'train/grad_norm_pre_clip': 25.236928939819336, 'timing/data': 0.0006217779591679573, 'timing/model': 1.1190030970610678, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.46}) +[2026-06-23 22:03:20,668][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.053398266434669495, 'train/grad_norm_pre_clip': 24.285717010498047, 'timing/data': 0.000495724962092936, 'timing/model': 1.1073071740102023, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.47}) +[2026-06-23 22:03:24,713][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.08998974412679672, 'train/grad_norm_pre_clip': 19.003406524658203, 'timing/data': 0.0006553990533575416, 'timing/model': 1.1096265779342502, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.47}) +[2026-06-23 22:03:28,767][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.051096536219120026, 'train/grad_norm_pre_clip': 20.106637954711914, 'timing/data': 0.0003701030509546399, 'timing/model': 1.111914558103308, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.47}) +[2026-06-23 22:03:32,819][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.03720194101333618, 'train/grad_norm_pre_clip': 16.41162872314453, 'timing/data': 0.0003623299999162555, 'timing/model': 1.1062987260520458, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.47}) +[2026-06-23 22:03:36,948][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.06501608341932297, 'train/grad_norm_pre_clip': 13.553522109985352, 'timing/data': 0.0005553620867431164, 'timing/model': 1.1090863220160827, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.47}) +[2026-06-23 22:03:41,005][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.10767414420843124, 'train/grad_norm_pre_clip': 13.203559875488281, 'timing/data': 0.0003586940001696348, 'timing/model': 1.1090534690301865, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.47}) +[2026-06-23 22:03:45,062][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.06388667970895767, 'train/grad_norm_pre_clip': 13.162202835083008, 'timing/data': 0.00036666402593255043, 'timing/model': 1.115300651988946, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.47}) +[2026-06-23 22:03:49,114][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.06283692270517349, 'train/grad_norm_pre_clip': 17.60586929321289, 'timing/data': 0.00036588101647794247, 'timing/model': 1.11082137003541, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.47}) +[2026-06-23 22:03:53,296][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.006078443955630064, 'train/grad_norm_pre_clip': 17.08805274963379, 'timing/data': 0.0003686390118673444, 'timing/model': 1.2359667050186545, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.47}) +[2026-06-23 22:03:57,359][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.026490967720746994, 'train/grad_norm_pre_clip': 19.663116455078125, 'timing/data': 0.0005958039546385407, 'timing/model': 1.1158635469619185, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.47}) +[2026-06-23 22:04:01,436][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.2692101299762726, 'train/grad_norm_pre_clip': 27.743173599243164, 'timing/data': 0.0005853159818798304, 'timing/model': 1.1123255250276998, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.47}) +[2026-06-23 22:04:05,501][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.08920913189649582, 'train/grad_norm_pre_clip': 13.24665355682373, 'timing/data': 0.0003897401038557291, 'timing/model': 1.1115645660320297, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.47}) +[2026-06-23 22:04:09,570][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.30770301818847656, 'train/grad_norm_pre_clip': 29.06041145324707, 'timing/data': 0.00041783007327467203, 'timing/model': 1.1118683549575508, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.47}) +[2026-06-23 22:04:13,816][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.03004540503025055, 'train/grad_norm_pre_clip': 21.02855110168457, 'timing/data': 0.0004281650763005018, 'timing/model': 1.116788907092996, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.47}) +[2026-06-23 22:04:17,892][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.1256619095802307, 'train/grad_norm_pre_clip': 44.936912536621094, 'timing/data': 0.0003494739066809416, 'timing/model': 1.1208919700002298, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.47}) +[2026-06-23 22:04:21,967][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.09557090699672699, 'train/grad_norm_pre_clip': 24.474872589111328, 'timing/data': 0.00039076595567166805, 'timing/model': 1.120915214996785, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.47}) +[2026-06-23 22:04:26,039][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.01951863244175911, 'train/grad_norm_pre_clip': 13.422348022460938, 'timing/data': 0.00036738195922225714, 'timing/model': 1.1184036269551143, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.47}) +[2026-06-23 22:04:30,191][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.05839909240603447, 'train/grad_norm_pre_clip': 12.756818771362305, 'timing/data': 0.00037430692464113235, 'timing/model': 1.119274153956212, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.47}) +[2026-06-23 22:04:34,272][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.0709281638264656, 'train/grad_norm_pre_clip': 17.387165069580078, 'timing/data': 0.0003560570767149329, 'timing/model': 1.1194912659702823, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.47}) +[2026-06-23 22:04:38,349][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.20388007164001465, 'train/grad_norm_pre_clip': 23.39524269104004, 'timing/data': 0.0003844270249828696, 'timing/model': 1.1187363010831177, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.47}) +[2026-06-23 22:04:42,433][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.2069919854402542, 'train/grad_norm_pre_clip': 20.76415252685547, 'timing/data': 0.0005758339539170265, 'timing/model': 1.117347399936989, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.47}) +[2026-06-23 22:04:46,606][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.020116081461310387, 'train/grad_norm_pre_clip': 12.046676635742188, 'timing/data': 0.00035593402571976185, 'timing/model': 1.1188078609993681, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.47}) +[2026-06-23 22:04:50,698][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.025600537657737732, 'train/grad_norm_pre_clip': 13.067694664001465, 'timing/data': 0.00038680003490298986, 'timing/model': 1.1228870550403371, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.47}) +[2026-06-23 22:04:54,780][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.18998271226882935, 'train/grad_norm_pre_clip': 14.50818920135498, 'timing/data': 0.00035912496969103813, 'timing/model': 1.1231002690037712, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.47}) +[2026-06-23 22:04:58,873][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.05461933836340904, 'train/grad_norm_pre_clip': 25.52117156982422, 'timing/data': 0.00038831098936498165, 'timing/model': 1.1219710280420259, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.47}) +[2026-06-23 22:05:03,058][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.031765230000019073, 'train/grad_norm_pre_clip': 23.727378845214844, 'timing/data': 0.0003578870091587305, 'timing/model': 1.1402212029788643, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.47}) +[2026-06-23 22:05:07,142][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.05009327083826065, 'train/grad_norm_pre_clip': 8.918249130249023, 'timing/data': 0.0003582249628379941, 'timing/model': 1.1165842770133168, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.48}) +[2026-06-23 22:05:11,220][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.013960647396743298, 'train/grad_norm_pre_clip': 15.408418655395508, 'timing/data': 0.00035674101673066616, 'timing/model': 1.1182049530325457, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.48}) +[2026-06-23 22:05:15,300][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.14740411937236786, 'train/grad_norm_pre_clip': 20.549165725708008, 'timing/data': 0.00036080891732126474, 'timing/model': 1.117072465014644, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.48}) +[2026-06-23 22:05:19,514][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.2485141009092331, 'train/grad_norm_pre_clip': 34.919342041015625, 'timing/data': 0.00036917603574693203, 'timing/model': 1.1127357359509915, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.48}) +[2026-06-23 22:05:23,606][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.11634218692779541, 'train/grad_norm_pre_clip': 39.14570236206055, 'timing/data': 0.0003579789772629738, 'timing/model': 1.1161311559844762, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.48}) +[2026-06-23 22:05:27,699][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.13815359771251678, 'train/grad_norm_pre_clip': 15.296375274658203, 'timing/data': 0.0003814700758084655, 'timing/model': 1.113635980989784, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.48}) +[2026-06-23 22:05:31,893][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.026952184736728668, 'train/grad_norm_pre_clip': 16.493955612182617, 'timing/data': 0.00036408810410648584, 'timing/model': 1.2183508669259027, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.48}) +[2026-06-23 22:05:35,991][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.2750934958457947, 'train/grad_norm_pre_clip': 26.81142807006836, 'timing/data': 0.0003575430018827319, 'timing/model': 1.1216021350119263, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.48}) +[2026-06-23 22:05:40,085][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.027259979397058487, 'train/grad_norm_pre_clip': 15.244046211242676, 'timing/data': 0.0004751369124278426, 'timing/model': 1.1193948129657656, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.48}) +[2026-06-23 22:05:44,163][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.09594423323869705, 'train/grad_norm_pre_clip': 12.94588565826416, 'timing/data': 0.0004152629990130663, 'timing/model': 1.1179615330183879, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.48}) +[2026-06-23 22:05:48,420][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.029135847464203835, 'train/grad_norm_pre_clip': 17.605018615722656, 'timing/data': 0.0003678679931908846, 'timing/model': 1.2871537120081484, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.48}) +[2026-06-23 22:05:52,508][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.048952169716358185, 'train/grad_norm_pre_clip': 10.476229667663574, 'timing/data': 0.000353038078173995, 'timing/model': 1.1268355610081926, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.48}) +[2026-06-23 22:05:56,593][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.006753101944923401, 'train/grad_norm_pre_clip': 20.57826805114746, 'timing/data': 0.00037982501089572906, 'timing/model': 1.1186319310218096, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.48}) +[2026-06-23 22:06:00,665][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.050654131919145584, 'train/grad_norm_pre_clip': 29.99920082092285, 'timing/data': 0.00036480394192039967, 'timing/model': 1.1128384330077097, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.48}) +[2026-06-23 22:06:04,824][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.08847087621688843, 'train/grad_norm_pre_clip': 22.276376724243164, 'timing/data': 0.00035857909824699163, 'timing/model': 1.1232824260368943, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.48}) +[2026-06-23 22:06:08,917][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.021751217544078827, 'train/grad_norm_pre_clip': 15.75235652923584, 'timing/data': 0.00035872403532266617, 'timing/model': 1.1183564260136336, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.48}) +[2026-06-23 22:06:13,009][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.07737082988023758, 'train/grad_norm_pre_clip': 21.3171443939209, 'timing/data': 0.0005727410316467285, 'timing/model': 1.117990077007562, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.48}) +[2026-06-23 22:06:17,108][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.0029386423993855715, 'train/grad_norm_pre_clip': 10.34414005279541, 'timing/data': 0.00035372795537114143, 'timing/model': 1.1213201220380142, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.48}) +[2026-06-23 22:06:21,270][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.015113642439246178, 'train/grad_norm_pre_clip': 27.23615074157715, 'timing/data': 0.0003956200089305639, 'timing/model': 1.1214233549544588, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.48}) +[2026-06-23 22:06:25,352][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.008791456930339336, 'train/grad_norm_pre_clip': 7.8555521965026855, 'timing/data': 0.00038492202293127775, 'timing/model': 1.1179054429521784, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.48}) +[2026-06-23 22:06:29,425][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.17364124953746796, 'train/grad_norm_pre_clip': 26.705373764038086, 'timing/data': 0.0003971810219809413, 'timing/model': 1.116078397957608, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.48}) +[2026-06-23 22:06:33,507][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.024412747472524643, 'train/grad_norm_pre_clip': 14.528791427612305, 'timing/data': 0.0004149180604144931, 'timing/model': 1.1197330800350755, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.48}) +[2026-06-23 22:15:41,484][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1250 +[2026-06-23 22:15:41,484][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.11420279741287231, 'train/grad_norm_pre_clip': 27.632488250732422, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 335.94000000001523, 'rl_games_eval/mid_train/mean_length': 2724.8, 'rl_games_eval/mid_train/std_reward': 154.00615052654007, 'rl_games_eval/mid_train/std_length': 1240.415317544894, 'rl_games_eval/mid_train/macro_mean_reward': 335.94000000001523, 'rl_games_eval/mid_train/macro_mean_length': 2724.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 335.94000000001523, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2724.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 154.00615052654007, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1240.415317544894, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 335.94000000001523, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00037120399065315723, 'timing/model': 1.1202000830089673, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.48}) +[2026-06-23 22:15:45,552][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.0029317860025912523, 'train/grad_norm_pre_clip': 30.14170265197754, 'timing/data': 0.0004349959781393409, 'timing/model': 1.1866984429070726, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.48}) +[2026-06-23 22:15:49,554][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.012359580025076866, 'train/grad_norm_pre_clip': 26.934593200683594, 'timing/data': 0.00038989505264908075, 'timing/model': 1.0994763319613412, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.48}) +[2026-06-23 22:15:53,551][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.05487823113799095, 'train/grad_norm_pre_clip': 32.971710205078125, 'timing/data': 0.00040996901225298643, 'timing/model': 1.095873925019987, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.48}) +[2026-06-23 22:15:57,561][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.04417990520596504, 'train/grad_norm_pre_clip': 14.118819236755371, 'timing/data': 0.00037557200994342566, 'timing/model': 1.0954620669363067, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.49}) +[2026-06-23 22:16:01,757][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.21511626243591309, 'train/grad_norm_pre_clip': 26.556013107299805, 'timing/data': 0.00038760306779295206, 'timing/model': 1.2789945229887962, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.49}) +[2026-06-23 22:16:05,783][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.278306782245636, 'train/grad_norm_pre_clip': 35.05079650878906, 'timing/data': 0.00038784893695265055, 'timing/model': 1.1140002690954134, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.49}) +[2026-06-23 22:16:09,796][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.004885226953774691, 'train/grad_norm_pre_clip': 15.155752182006836, 'timing/data': 0.0004259650595486164, 'timing/model': 1.104091736022383, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.49}) +[2026-06-23 22:16:13,849][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.06716398149728775, 'train/grad_norm_pre_clip': 28.790258407592773, 'timing/data': 0.00046962592750787735, 'timing/model': 1.105229189968668, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.49}) +[2026-06-23 22:16:17,986][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.004964605905115604, 'train/grad_norm_pre_clip': 17.770620346069336, 'timing/data': 0.0003587530227378011, 'timing/model': 1.1077207599300891, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.49}) +[2026-06-23 22:16:22,031][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.013081728480756283, 'train/grad_norm_pre_clip': 25.23536491394043, 'timing/data': 0.000656403019092977, 'timing/model': 1.110752667998895, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.49}) +[2026-06-23 22:16:26,077][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.04742495343089104, 'train/grad_norm_pre_clip': 13.120247840881348, 'timing/data': 0.0003593780566006899, 'timing/model': 1.1082219620002434, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.49}) +[2026-06-23 22:16:30,115][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.11620481312274933, 'train/grad_norm_pre_clip': 27.631088256835938, 'timing/data': 0.000526564079336822, 'timing/model': 1.1090232690330595, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.49}) +[2026-06-23 22:16:34,258][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.17941619455814362, 'train/grad_norm_pre_clip': 16.803720474243164, 'timing/data': 0.0006268579745665193, 'timing/model': 1.1075345380231738, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.49}) +[2026-06-23 22:16:38,309][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.04074995592236519, 'train/grad_norm_pre_clip': 14.175700187683105, 'timing/data': 0.00043917400762438774, 'timing/model': 1.1114974319934845, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.49}) +[2026-06-23 22:16:42,360][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.036185234785079956, 'train/grad_norm_pre_clip': 13.12295913696289, 'timing/data': 0.00035897898487746716, 'timing/model': 1.1093777159694582, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.49}) +[2026-06-23 22:16:46,422][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.1032780110836029, 'train/grad_norm_pre_clip': 14.58000659942627, 'timing/data': 0.00041360303293913603, 'timing/model': 1.1135653690434992, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.49}) +[2026-06-23 22:16:50,628][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.053398627787828445, 'train/grad_norm_pre_clip': 7.269418239593506, 'timing/data': 0.0003657069755718112, 'timing/model': 1.1124905130127445, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.49}) +[2026-06-23 22:16:54,679][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.0502898171544075, 'train/grad_norm_pre_clip': 9.501860618591309, 'timing/data': 0.0006671909941360354, 'timing/model': 1.1100011540111154, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.49}) +[2026-06-23 22:16:58,721][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.2177901417016983, 'train/grad_norm_pre_clip': 22.990049362182617, 'timing/data': 0.0003697610227391124, 'timing/model': 1.108666761079803, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.49}) +[2026-06-23 22:17:02,776][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.08953636884689331, 'train/grad_norm_pre_clip': 69.98344421386719, 'timing/data': 0.000430906075052917, 'timing/model': 1.114888811018318, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.49}) +[2026-06-23 22:17:06,953][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.25632330775260925, 'train/grad_norm_pre_clip': 47.299190521240234, 'timing/data': 0.0005098390392959118, 'timing/model': 1.115614564041607, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.49}) +[2026-06-23 22:17:11,027][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.17177484929561615, 'train/grad_norm_pre_clip': 39.02400588989258, 'timing/data': 0.00036888394970446825, 'timing/model': 1.1187252650270239, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.49}) +[2026-06-23 22:17:15,088][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.02543441765010357, 'train/grad_norm_pre_clip': 13.122600555419922, 'timing/data': 0.00038464099634438753, 'timing/model': 1.1152808029437438, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.49}) +[2026-06-23 22:17:19,166][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.13895563781261444, 'train/grad_norm_pre_clip': 27.602188110351562, 'timing/data': 0.000367077998816967, 'timing/model': 1.1194653420243412, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.49}) +[2026-06-23 22:17:23,302][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.41284146904945374, 'train/grad_norm_pre_clip': 48.9460334777832, 'timing/data': 0.00035418395418673754, 'timing/model': 1.1127549690427259, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.49}) +[2026-06-23 22:17:27,376][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.0742572546005249, 'train/grad_norm_pre_clip': 27.957263946533203, 'timing/data': 0.0004800989991053939, 'timing/model': 1.1158147560199723, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.49}) +[2026-06-23 22:17:31,443][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.1111564114689827, 'train/grad_norm_pre_clip': 35.37831115722656, 'timing/data': 0.0003559320466592908, 'timing/model': 1.1150149569148198, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.49}) +[2026-06-23 22:17:35,513][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.0821341946721077, 'train/grad_norm_pre_clip': 36.653038024902344, 'timing/data': 0.00035986595321446657, 'timing/model': 1.1115613440051675, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.49}) +[2026-06-23 22:17:39,755][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.1509311944246292, 'train/grad_norm_pre_clip': 21.761178970336914, 'timing/data': 0.00038641097489744425, 'timing/model': 1.1158602359937504, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.49}) +[2026-06-23 22:17:43,814][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.038506533950567245, 'train/grad_norm_pre_clip': 21.733102798461914, 'timing/data': 0.00038317497819662094, 'timing/model': 1.1138185679446906, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.5}) +[2026-06-23 22:17:47,892][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.05783186852931976, 'train/grad_norm_pre_clip': 15.49117374420166, 'timing/data': 0.0003908309154212475, 'timing/model': 1.1148226929362863, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.5}) +[2026-06-23 22:17:51,962][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.17916102707386017, 'train/grad_norm_pre_clip': 17.537601470947266, 'timing/data': 0.00038367300294339657, 'timing/model': 1.11797347094398, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.5}) +[2026-06-23 22:17:56,127][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.06434562802314758, 'train/grad_norm_pre_clip': 22.603853225708008, 'timing/data': 0.0003561490448191762, 'timing/model': 1.1115202159853652, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.5}) +[2026-06-23 22:18:00,201][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.08356612175703049, 'train/grad_norm_pre_clip': 15.666226387023926, 'timing/data': 0.00039599102456122637, 'timing/model': 1.1167481939774007, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.5}) +[2026-06-23 22:18:04,301][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.14513958990573883, 'train/grad_norm_pre_clip': 17.04233169555664, 'timing/data': 0.000401680008508265, 'timing/model': 1.1173083670437336, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.5}) +[2026-06-23 22:18:08,409][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.08070565015077591, 'train/grad_norm_pre_clip': 25.961990356445312, 'timing/data': 0.0003587600076571107, 'timing/model': 1.1134075899608433, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.5}) +[2026-06-23 22:18:12,678][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.06524330377578735, 'train/grad_norm_pre_clip': 19.154891967773438, 'timing/data': 0.00036154710687696934, 'timing/model': 1.1188188149826601, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.5}) +[2026-06-23 22:18:16,749][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.01634771190583706, 'train/grad_norm_pre_clip': 22.51342010498047, 'timing/data': 0.0003673030296340585, 'timing/model': 1.1129172150976956, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.5}) +[2026-06-23 22:18:20,821][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.06012078747153282, 'train/grad_norm_pre_clip': 25.21710205078125, 'timing/data': 0.0004026399692520499, 'timing/model': 1.1164625079836696, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.5}) +[2026-06-23 22:18:24,909][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.09595032036304474, 'train/grad_norm_pre_clip': 39.041343688964844, 'timing/data': 0.000414157984778285, 'timing/model': 1.1176006799796596, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.5}) +[2026-06-23 22:18:29,052][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.10155507922172546, 'train/grad_norm_pre_clip': 21.45193862915039, 'timing/data': 0.0003912290558218956, 'timing/model': 1.1201884730253369, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.5}) +[2026-06-23 22:18:33,148][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.2912159264087677, 'train/grad_norm_pre_clip': 32.86992645263672, 'timing/data': 0.0003621620126068592, 'timing/model': 1.118860858026892, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.5}) +[2026-06-23 22:18:37,226][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.029495319351553917, 'train/grad_norm_pre_clip': 16.10871696472168, 'timing/data': 0.0003534810384735465, 'timing/model': 1.117838463978842, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.5}) +[2026-06-23 22:18:41,308][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.20600980520248413, 'train/grad_norm_pre_clip': 39.56159210205078, 'timing/data': 0.0006280210800468922, 'timing/model': 1.1185819830279797, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.5}) +[2026-06-23 22:18:45,482][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.13138173520565033, 'train/grad_norm_pre_clip': 56.480796813964844, 'timing/data': 0.0003600900527089834, 'timing/model': 1.119653531932272, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.5}) +[2026-06-23 22:18:49,571][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.23529383540153503, 'train/grad_norm_pre_clip': 32.949214935302734, 'timing/data': 0.0003569719847291708, 'timing/model': 1.116704044979997, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.5}) +[2026-06-23 22:18:53,663][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.08753132820129395, 'train/grad_norm_pre_clip': 27.570907592773438, 'timing/data': 0.00037892197724431753, 'timing/model': 1.117857855046168, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.5}) +[2026-06-23 22:18:57,738][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.031040117144584656, 'train/grad_norm_pre_clip': 29.491193771362305, 'timing/data': 0.0004949299618601799, 'timing/model': 1.1197088069748133, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.5}) +[2026-06-23 22:19:01,962][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.17458459734916687, 'train/grad_norm_pre_clip': 38.77707290649414, 'timing/data': 0.0003733669873327017, 'timing/model': 1.1357085059862584, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.5}) +[2026-06-23 22:19:06,040][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1300 +[2026-06-23 22:20:10,191][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1300 +[2026-06-23 22:20:10,193][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1300 +[2026-06-23 22:20:49,778][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1300 +[2026-06-23 22:20:49,779][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.08583391457796097, 'train/grad_norm_pre_clip': 24.55936050415039, 'eval/action_loss/seconds': 64.14930169202853, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.08410519361495972, 'eval/latency_2/loss': 0.08410519361495972, 'eval/flappy/loss': 0.08410519361495972, 'eval/flappy/latency_2/loss': 0.08410519361495972, 'eval/flap_precision': 0.9833080424886191, 'eval/flap_recall': 0.8792401628222524, 'eval/flap_f1': 0.9283667621776505, 'eval/flap_rate_model': 0.18305555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9283667621776505, 'eval/latency_2/flap_precision': 0.9833080424886191, 'eval/latency_2/flap_recall': 0.8792401628222524, 'eval/latency_2/flap_f1': 0.9283667621776505, 'eval/latency_2/flap_rate_model': 0.18305555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9283667621776505, 'eval/action_classification/data_wait_seconds': 0.7234644293785095, 'eval/action_classification/predict_seconds': 38.74134063720703, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0005947670433670282, 'timing/model': 1.1153885769890621, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.5}) +[2026-06-23 22:20:53,818][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.055886268615722656, 'train/grad_norm_pre_clip': 51.61875534057617, 'timing/data': 0.0004086289554834366, 'timing/model': 1.1078937130514532, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.5}) +[2026-06-23 22:20:57,926][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.06952770054340363, 'train/grad_norm_pre_clip': 40.630088806152344, 'timing/data': 0.00036173604894429445, 'timing/model': 1.1097804849268869, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.5}) +[2026-06-23 22:21:01,976][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.1431676596403122, 'train/grad_norm_pre_clip': 33.490806579589844, 'timing/data': 0.00044880807399749756, 'timing/model': 1.10877014498692, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.5}) +[2026-06-23 22:21:06,018][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.17998544871807098, 'train/grad_norm_pre_clip': 18.92594337463379, 'timing/data': 0.0003532510017976165, 'timing/model': 1.1065336680039763, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.5}) +[2026-06-23 22:21:10,071][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.050502631813287735, 'train/grad_norm_pre_clip': 18.268198013305664, 'timing/data': 0.0004443629877641797, 'timing/model': 1.1128069310216233, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.51}) +[2026-06-23 22:21:14,249][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.4713486433029175, 'train/grad_norm_pre_clip': 30.55019760131836, 'timing/data': 0.00043090106919407845, 'timing/model': 1.1147410970879719, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.51}) +[2026-06-23 22:21:18,329][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.4111008644104004, 'train/grad_norm_pre_clip': 38.6566276550293, 'timing/data': 0.00035448302514851093, 'timing/model': 1.109617430018261, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.51}) +[2026-06-23 22:21:22,383][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.3128657042980194, 'train/grad_norm_pre_clip': 19.27130699157715, 'timing/data': 0.0004862700589001179, 'timing/model': 1.108633903088048, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.51}) +[2026-06-23 22:21:26,536][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.0836930200457573, 'train/grad_norm_pre_clip': 11.263960838317871, 'timing/data': 0.0005248290253803134, 'timing/model': 1.2086730890441686, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.51}) +[2026-06-23 22:21:30,604][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.10100232064723969, 'train/grad_norm_pre_clip': 12.540185928344727, 'timing/data': 0.0003599079791456461, 'timing/model': 1.1182859789114445, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.51}) +[2026-06-23 22:21:34,673][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.12386667728424072, 'train/grad_norm_pre_clip': 18.67166519165039, 'timing/data': 0.00036646996159106493, 'timing/model': 1.1189424850745127, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.51}) +[2026-06-23 22:21:38,737][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.10019605606794357, 'train/grad_norm_pre_clip': 30.767751693725586, 'timing/data': 0.0005887990118935704, 'timing/model': 1.1133023239672184, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.51}) +[2026-06-23 22:21:42,969][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.24971698224544525, 'train/grad_norm_pre_clip': 31.96185874938965, 'timing/data': 0.00035690306685864925, 'timing/model': 1.2868240440730006, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.51}) +[2026-06-23 22:21:47,038][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.2228626310825348, 'train/grad_norm_pre_clip': 36.72109603881836, 'timing/data': 0.0005213950062170625, 'timing/model': 1.1096839889651164, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.51}) +[2026-06-23 22:21:51,100][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.08600939065217972, 'train/grad_norm_pre_clip': 13.744172096252441, 'timing/data': 0.0003522129263728857, 'timing/model': 1.113905780017376, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.51}) +[2026-06-23 22:21:55,178][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.024235982447862625, 'train/grad_norm_pre_clip': 12.03548812866211, 'timing/data': 0.00035417499020695686, 'timing/model': 1.1184593719663098, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.51}) +[2026-06-23 22:21:59,326][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.12437771260738373, 'train/grad_norm_pre_clip': 20.48457145690918, 'timing/data': 0.00036017398815602064, 'timing/model': 1.1209198200376704, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.51}) +[2026-06-23 22:22:03,402][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.07817061990499496, 'train/grad_norm_pre_clip': 19.718847274780273, 'timing/data': 0.0003908890066668391, 'timing/model': 1.1127709619468078, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.51}) +[2026-06-23 22:22:07,468][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.020558657124638557, 'train/grad_norm_pre_clip': 16.039297103881836, 'timing/data': 0.0003598190378397703, 'timing/model': 1.1151777049526572, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.51}) +[2026-06-23 22:22:11,554][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.2276439666748047, 'train/grad_norm_pre_clip': 25.527217864990234, 'timing/data': 0.00036798708606511354, 'timing/model': 1.1258002260001376, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.51}) +[2026-06-23 22:22:15,696][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.016304779797792435, 'train/grad_norm_pre_clip': 6.3569512367248535, 'timing/data': 0.0003627869300544262, 'timing/model': 1.1204651109874249, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.51}) +[2026-06-23 22:22:19,768][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.048479318618774414, 'train/grad_norm_pre_clip': 30.093700408935547, 'timing/data': 0.0003822990693151951, 'timing/model': 1.1151179010048509, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.51}) +[2026-06-23 22:22:23,859][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.16956883668899536, 'train/grad_norm_pre_clip': 37.204872131347656, 'timing/data': 0.0006777290254831314, 'timing/model': 1.1126940019894391, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.51}) +[2026-06-23 22:22:27,934][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.11825458705425262, 'train/grad_norm_pre_clip': 25.877758026123047, 'timing/data': 0.00035497010685503483, 'timing/model': 1.1233107330044731, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.51}) +[2026-06-23 22:22:32,161][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.08184020221233368, 'train/grad_norm_pre_clip': 33.328372955322266, 'timing/data': 0.0003706389106810093, 'timing/model': 1.1185463099973276, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.51}) +[2026-06-23 22:22:36,233][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.12283942848443985, 'train/grad_norm_pre_clip': 18.260791778564453, 'timing/data': 0.00036633701529353857, 'timing/model': 1.1135071499738842, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.51}) +[2026-06-23 22:22:40,319][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.008673179894685745, 'train/grad_norm_pre_clip': 16.47347068786621, 'timing/data': 0.0005042280536144972, 'timing/model': 1.115561163984239, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.51}) +[2026-06-23 22:22:44,409][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.04967886582016945, 'train/grad_norm_pre_clip': 23.300235748291016, 'timing/data': 0.0003589040134102106, 'timing/model': 1.1227806580718607, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.51}) +[2026-06-23 22:22:48,591][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.00862775556743145, 'train/grad_norm_pre_clip': 27.673076629638672, 'timing/data': 0.0005509150214493275, 'timing/model': 1.1089885199908167, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.51}) +[2026-06-23 22:22:52,683][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.037413448095321655, 'train/grad_norm_pre_clip': 23.514652252197266, 'timing/data': 0.0003547749947756529, 'timing/model': 1.1235119079938158, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.51}) +[2026-06-23 22:22:56,763][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.05978090316057205, 'train/grad_norm_pre_clip': 32.088924407958984, 'timing/data': 0.00040910905227065086, 'timing/model': 1.121261523105204, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.52}) +[2026-06-23 22:23:00,845][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.22660042345523834, 'train/grad_norm_pre_clip': 22.228673934936523, 'timing/data': 0.0003749040188267827, 'timing/model': 1.1169666870264336, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.52}) +[2026-06-23 22:23:05,017][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.14935249090194702, 'train/grad_norm_pre_clip': 20.562089920043945, 'timing/data': 0.0004098089411854744, 'timing/model': 1.122639462002553, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.52}) +[2026-06-23 22:23:09,099][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.1586541086435318, 'train/grad_norm_pre_clip': 19.038284301757812, 'timing/data': 0.00035963591653853655, 'timing/model': 1.119726094068028, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.52}) +[2026-06-23 22:23:13,176][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.17961330711841583, 'train/grad_norm_pre_clip': 18.003787994384766, 'timing/data': 0.00041505496483296156, 'timing/model': 1.1182020049309358, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.52}) +[2026-06-23 22:23:17,255][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.06027384102344513, 'train/grad_norm_pre_clip': 14.487354278564453, 'timing/data': 0.0003690100274980068, 'timing/model': 1.1205938589992002, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.52}) +[2026-06-23 22:23:21,526][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.09472979605197906, 'train/grad_norm_pre_clip': 14.573391914367676, 'timing/data': 0.00044088903814554214, 'timing/model': 1.1187424870440736, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 0.52}) +[2026-06-23 22:23:25,604][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.03232365846633911, 'train/grad_norm_pre_clip': 23.763038635253906, 'timing/data': 0.00035834498703479767, 'timing/model': 1.117883043945767, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 0.52}) +[2026-06-23 22:23:29,669][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.019387386739253998, 'train/grad_norm_pre_clip': 12.499652862548828, 'timing/data': 0.0006079239537939429, 'timing/model': 1.1177995090838522, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 0.52}) +[2026-06-23 22:23:33,733][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.06596757471561432, 'train/grad_norm_pre_clip': 15.197545051574707, 'timing/data': 0.00040588504634797573, 'timing/model': 1.1173601240152493, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.52}) +[2026-06-23 22:23:37,922][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.02885271981358528, 'train/grad_norm_pre_clip': 10.426702499389648, 'timing/data': 0.0005524030420929193, 'timing/model': 1.1164628649130464, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 0.52}) +[2026-06-23 22:23:42,000][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.1815653294324875, 'train/grad_norm_pre_clip': 17.79156494140625, 'timing/data': 0.0003765879664570093, 'timing/model': 1.115906681981869, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 0.52}) +[2026-06-23 22:23:46,073][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.10500938445329666, 'train/grad_norm_pre_clip': 13.439833641052246, 'timing/data': 0.00035190500784665346, 'timing/model': 1.1153617550153285, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 0.52}) +[2026-06-23 22:23:50,156][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.06399019807577133, 'train/grad_norm_pre_clip': 11.571456909179688, 'timing/data': 0.00035777094308286905, 'timing/model': 1.1193290159571916, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 0.52}) +[2026-06-23 22:23:54,394][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.11781045794487, 'train/grad_norm_pre_clip': 17.94465446472168, 'timing/data': 0.00036052800714969635, 'timing/model': 1.1187298049917445, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 0.52}) +[2026-06-23 22:23:58,487][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.024956556037068367, 'train/grad_norm_pre_clip': 31.484407424926758, 'timing/data': 0.0004346870118752122, 'timing/model': 1.134446848067455, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 0.52}) +[2026-06-23 22:24:02,575][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.13185158371925354, 'train/grad_norm_pre_clip': 38.481136322021484, 'timing/data': 0.0003676199121400714, 'timing/model': 1.1170259549980983, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 0.52}) +[2026-06-23 22:24:06,675][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.13010995090007782, 'train/grad_norm_pre_clip': 15.87720775604248, 'timing/data': 0.0003535090945661068, 'timing/model': 1.1275555320316926, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 0.52}) +[2026-06-23 22:24:10,836][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.12370855361223221, 'train/grad_norm_pre_clip': 17.167888641357422, 'timing/data': 0.00038688292261213064, 'timing/model': 1.1175471020396799, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 0.52}) +[2026-06-23 22:24:14,938][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.14656612277030945, 'train/grad_norm_pre_clip': 15.806265830993652, 'timing/data': 0.0003635829780250788, 'timing/model': 1.1141141740372404, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.52}) +[2026-06-23 22:24:19,018][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.18262092769145966, 'train/grad_norm_pre_clip': 21.505481719970703, 'timing/data': 0.00041395402513444424, 'timing/model': 1.1156755849951878, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 0.52}) +[2026-06-23 22:24:23,097][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.05056577920913696, 'train/grad_norm_pre_clip': 14.164944648742676, 'timing/data': 0.0003646609839051962, 'timing/model': 1.1218957579694688, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 0.52}) +[2026-06-23 22:24:27,259][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.07458604127168655, 'train/grad_norm_pre_clip': 26.425884246826172, 'timing/data': 0.0004527790006250143, 'timing/model': 1.120447597000748, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 0.52}) +[2026-06-23 22:24:31,341][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.005746278911828995, 'train/grad_norm_pre_clip': 26.45728874206543, 'timing/data': 0.00040914001874625683, 'timing/model': 1.1204032630193979, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 0.52}) +[2026-06-23 22:24:35,427][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.10170573741197586, 'train/grad_norm_pre_clip': 24.425716400146484, 'timing/data': 0.0004454209702089429, 'timing/model': 1.1220562580274418, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 0.52}) +[2026-06-23 22:24:39,493][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.10696208477020264, 'train/grad_norm_pre_clip': 42.60371398925781, 'timing/data': 0.00036372803151607513, 'timing/model': 1.1193220529239625, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 0.52}) +[2026-06-23 22:24:43,714][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.05181140452623367, 'train/grad_norm_pre_clip': 42.66434097290039, 'timing/data': 0.0004430109402164817, 'timing/model': 1.118871980928816, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 0.53}) +[2026-06-23 22:24:47,792][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.045031603425741196, 'train/grad_norm_pre_clip': 13.441070556640625, 'timing/data': 0.0003636500332504511, 'timing/model': 1.1182741139782593, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 0.53}) +[2026-06-23 22:24:51,873][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.23434306681156158, 'train/grad_norm_pre_clip': 27.898712158203125, 'timing/data': 0.0003653810126706958, 'timing/model': 1.110578343970701, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 0.53}) +[2026-06-23 22:24:56,087][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.12516996264457703, 'train/grad_norm_pre_clip': 22.618831634521484, 'timing/data': 0.0003622230142354965, 'timing/model': 1.2253927360288799, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.53}) +[2026-06-23 22:25:00,173][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.21655912697315216, 'train/grad_norm_pre_clip': 30.604036331176758, 'timing/data': 0.0005977940745651722, 'timing/model': 1.119868258945644, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 0.53}) +[2026-06-23 22:25:04,269][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.03585243970155716, 'train/grad_norm_pre_clip': 9.817828178405762, 'timing/data': 0.0003596280002966523, 'timing/model': 1.1225772339385003, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 0.53}) +[2026-06-23 22:25:08,350][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.06477866321802139, 'train/grad_norm_pre_clip': 11.230778694152832, 'timing/data': 0.0003600749187171459, 'timing/model': 1.1168570349691436, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 0.53}) +[2026-06-23 22:25:12,617][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.14770671725273132, 'train/grad_norm_pre_clip': 22.57048988342285, 'timing/data': 0.00037358992267400026, 'timing/model': 1.2991526799742132, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 0.53}) +[2026-06-23 22:25:16,694][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.08834841102361679, 'train/grad_norm_pre_clip': 19.349653244018555, 'timing/data': 0.0004149191081523895, 'timing/model': 1.1173396499361843, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 0.53}) +[2026-06-23 22:25:20,757][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.13827098906040192, 'train/grad_norm_pre_clip': 37.17589569091797, 'timing/data': 0.0004497499903663993, 'timing/model': 1.1109611910069361, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 0.53}) +[2026-06-23 22:25:24,833][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.09169196337461472, 'train/grad_norm_pre_clip': 13.073925971984863, 'timing/data': 0.00037509400863200426, 'timing/model': 1.1196798669407144, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 0.53}) +[2026-06-23 22:25:28,979][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.21126706898212433, 'train/grad_norm_pre_clip': 20.26578712463379, 'timing/data': 0.000367900007404387, 'timing/model': 1.1199949500150979, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 0.53}) +[2026-06-23 22:25:33,066][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.12090418487787247, 'train/grad_norm_pre_clip': 29.92544937133789, 'timing/data': 0.0003625149838626385, 'timing/model': 1.1218438039068133, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 0.53}) +[2026-06-23 22:25:37,161][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.1740422248840332, 'train/grad_norm_pre_clip': 19.57490348815918, 'timing/data': 0.00039146991912275553, 'timing/model': 1.1207521199248731, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.53}) +[2026-06-23 22:25:41,255][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.07999880611896515, 'train/grad_norm_pre_clip': 13.843509674072266, 'timing/data': 0.0006430379580706358, 'timing/model': 1.11884747503791, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 0.53}) +[2026-06-23 22:25:45,398][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.036979738622903824, 'train/grad_norm_pre_clip': 11.927301406860352, 'timing/data': 0.0003590240376070142, 'timing/model': 1.120307911070995, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 0.53}) +[2026-06-23 22:25:49,489][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.09161181002855301, 'train/grad_norm_pre_clip': 10.123749732971191, 'timing/data': 0.0003745310241356492, 'timing/model': 1.1305569250835106, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 0.53}) +[2026-06-23 22:25:53,573][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.0025516950991004705, 'train/grad_norm_pre_clip': 34.748207092285156, 'timing/data': 0.00035261898301541805, 'timing/model': 1.1178465420380235, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.53}) +[2026-06-23 22:25:57,648][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.06877747923135757, 'train/grad_norm_pre_clip': 25.388805389404297, 'timing/data': 0.0006137139862403274, 'timing/model': 1.11555705498904, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 0.53}) +[2026-06-23 22:26:01,865][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.03338013216853142, 'train/grad_norm_pre_clip': 15.232531547546387, 'timing/data': 0.0003630280261859298, 'timing/model': 1.1168679080437869, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 0.53}) +[2026-06-23 22:26:05,956][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.11063404381275177, 'train/grad_norm_pre_clip': 19.494937896728516, 'timing/data': 0.0003593460423871875, 'timing/model': 1.1255640729796141, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 0.53}) +[2026-06-23 22:26:10,044][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.09509080648422241, 'train/grad_norm_pre_clip': 10.953530311584473, 'timing/data': 0.0003688660217449069, 'timing/model': 1.1179238629993051, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 0.53}) +[2026-06-23 22:26:14,128][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.003581586293876171, 'train/grad_norm_pre_clip': 26.817773818969727, 'timing/data': 0.0003576419549062848, 'timing/model': 1.117980121052824, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 0.53}) +[2026-06-23 22:26:18,330][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.04538344964385033, 'train/grad_norm_pre_clip': 21.902000427246094, 'timing/data': 0.0003715159837156534, 'timing/model': 1.121376431081444, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.53}) +[2026-06-23 22:26:22,415][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.19131319224834442, 'train/grad_norm_pre_clip': 43.528194427490234, 'timing/data': 0.00035975396167486906, 'timing/model': 1.1222762999823317, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 0.53}) +[2026-06-23 22:26:26,508][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.0388469435274601, 'train/grad_norm_pre_clip': 44.18762969970703, 'timing/data': 0.0005253680283203721, 'timing/model': 1.1315299989655614, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 0.53}) +[2026-06-23 22:26:30,591][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.14624235033988953, 'train/grad_norm_pre_clip': 27.09775733947754, 'timing/data': 0.00041620596311986446, 'timing/model': 1.1224237179849297, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 0.54}) +[2026-06-23 22:26:34,765][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.1418832391500473, 'train/grad_norm_pre_clip': 29.38335418701172, 'timing/data': 0.0004192920168861747, 'timing/model': 1.1204091400140896, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 0.54}) +[2026-06-23 22:26:38,851][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.06976388394832611, 'train/grad_norm_pre_clip': 77.8400650024414, 'timing/data': 0.00035735091660171747, 'timing/model': 1.120285882963799, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 0.54}) +[2026-06-23 22:26:42,934][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.0708186998963356, 'train/grad_norm_pre_clip': 20.31844139099121, 'timing/data': 0.00036457006353884935, 'timing/model': 1.1213910509832203, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 0.54}) +[2026-06-23 22:26:47,139][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.02025527134537697, 'train/grad_norm_pre_clip': 19.65947723388672, 'timing/data': 0.000530505902133882, 'timing/model': 1.1167455801041797, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 0.54}) +[2026-06-23 22:26:51,400][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.036436405032873154, 'train/grad_norm_pre_clip': 38.64805221557617, 'timing/data': 0.0003739199601113796, 'timing/model': 1.1190131419571117, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 0.54}) +[2026-06-23 22:26:55,487][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.14344356954097748, 'train/grad_norm_pre_clip': 37.83388137817383, 'timing/data': 0.0003625419922173023, 'timing/model': 1.1226040920009837, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 0.54}) +[2026-06-23 22:26:59,565][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.08645578473806381, 'train/grad_norm_pre_clip': 21.15052604675293, 'timing/data': 0.00035667105112224817, 'timing/model': 1.1153419890906662, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.54}) +[2026-06-23 22:27:03,639][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.04702701047062874, 'train/grad_norm_pre_clip': 27.516977310180664, 'timing/data': 0.000534247956238687, 'timing/model': 1.1141333179548383, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 0.54}) +[2026-06-23 22:27:07,816][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.04836506024003029, 'train/grad_norm_pre_clip': 15.100030899047852, 'timing/data': 0.0005454180063679814, 'timing/model': 1.1132774879224598, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 0.54}) +[2026-06-23 22:27:11,894][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.13846296072006226, 'train/grad_norm_pre_clip': 40.983882904052734, 'timing/data': 0.0003644479438662529, 'timing/model': 1.1172415749169886, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 0.54}) +[2026-06-23 22:27:15,973][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.11531504988670349, 'train/grad_norm_pre_clip': 23.14206886291504, 'timing/data': 0.0003616360481828451, 'timing/model': 1.1138565649744123, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 0.54}) +[2026-06-23 22:27:20,066][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.09958516061306, 'train/grad_norm_pre_clip': 35.01972198486328, 'timing/data': 0.0003569279797375202, 'timing/model': 1.1244871249655262, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.54}) +[2026-06-23 22:27:24,350][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.04365084692835808, 'train/grad_norm_pre_clip': 30.4660587310791, 'timing/data': 0.0003628670237958431, 'timing/model': 1.1220814029220492, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 0.54}) +[2026-06-23 22:27:28,424][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.10429994016885757, 'train/grad_norm_pre_clip': 18.53023338317871, 'timing/data': 0.0003589589614421129, 'timing/model': 1.1151224200148135, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 0.54}) +[2026-06-23 22:27:32,518][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.09411382675170898, 'train/grad_norm_pre_clip': 13.784775733947754, 'timing/data': 0.0003593940054997802, 'timing/model': 1.1200891820481047, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 0.54}) +[2026-06-23 22:27:36,611][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.13409392535686493, 'train/grad_norm_pre_clip': 20.613805770874023, 'timing/data': 0.0003723310073837638, 'timing/model': 1.1249597129644826, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 0.54}) +[2026-06-23 22:27:40,781][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1400 +[2026-06-23 22:28:44,467][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1400 +[2026-06-23 22:28:44,469][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1400 +[2026-06-23 22:29:23,798][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1400 +[2026-06-23 22:29:23,799][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.2098878026008606, 'train/grad_norm_pre_clip': 23.57852554321289, 'eval/action_loss/seconds': 63.68407002103049, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.04882030189037323, 'eval/latency_2/loss': 0.04882030189037323, 'eval/flappy/loss': 0.04882030189037323, 'eval/flappy/latency_2/loss': 0.04882030189037323, 'eval/flap_precision': 0.9502018842530283, 'eval/flap_recall': 0.9579375848032564, 'eval/flap_f1': 0.954054054054054, 'eval/flap_rate_model': 0.2063888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.954054054054054, 'eval/latency_2/flap_precision': 0.9502018842530283, 'eval/latency_2/flap_recall': 0.9579375848032564, 'eval/latency_2/flap_f1': 0.954054054054054, 'eval/latency_2/flap_rate_model': 0.2063888888888889, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.954054054054054, 'eval/action_classification/data_wait_seconds': 0.5852212309837341, 'eval/action_classification/predict_seconds': 38.622779846191406, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0005586049519479275, 'timing/model': 1.1169900179374963, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.54}) +[2026-06-23 22:29:27,850][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.0726243257522583, 'train/grad_norm_pre_clip': 7.823905944824219, 'timing/data': 0.00040382298175245523, 'timing/model': 1.103853594046086, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 0.54}) +[2026-06-23 22:29:31,892][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.022299515083432198, 'train/grad_norm_pre_clip': 12.09054183959961, 'timing/data': 0.0004117030184715986, 'timing/model': 1.108223892049864, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 0.54}) +[2026-06-23 22:29:36,028][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.1294347643852234, 'train/grad_norm_pre_clip': 19.391061782836914, 'timing/data': 0.0003641800722107291, 'timing/model': 1.110920381033793, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 0.54}) +[2026-06-23 22:29:40,087][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.18097703158855438, 'train/grad_norm_pre_clip': 17.82852554321289, 'timing/data': 0.00036866997834295034, 'timing/model': 1.1122505409875885, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 0.54}) +[2026-06-23 22:29:44,154][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.08149638026952744, 'train/grad_norm_pre_clip': 40.971378326416016, 'timing/data': 0.00047369208186864853, 'timing/model': 1.115782596054487, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 0.54}) +[2026-06-23 22:29:48,221][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.0456508994102478, 'train/grad_norm_pre_clip': 14.44931411743164, 'timing/data': 0.0004545829724520445, 'timing/model': 1.11419194098562, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 0.54}) +[2026-06-23 22:29:52,479][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.1913028359413147, 'train/grad_norm_pre_clip': 12.731358528137207, 'timing/data': 0.00035732192918658257, 'timing/model': 1.120150379021652, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 0.54}) +[2026-06-23 22:29:56,546][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.17164739966392517, 'train/grad_norm_pre_clip': 21.823760986328125, 'timing/data': 0.00035447406116873026, 'timing/model': 1.109673670027405, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 0.54}) +[2026-06-23 22:30:00,596][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.09020529687404633, 'train/grad_norm_pre_clip': 27.217348098754883, 'timing/data': 0.00034629704896360636, 'timing/model': 1.106369029963389, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 0.55}) +[2026-06-23 22:30:04,643][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.07247229665517807, 'train/grad_norm_pre_clip': 21.99785614013672, 'timing/data': 0.0005217499565333128, 'timing/model': 1.1110944210086018, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.55}) +[2026-06-23 22:30:08,856][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.05608927831053734, 'train/grad_norm_pre_clip': 19.355417251586914, 'timing/data': 0.0004805739736184478, 'timing/model': 1.1064516380429268, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 0.55}) +[2026-06-23 22:30:12,918][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.25920626521110535, 'train/grad_norm_pre_clip': 29.610851287841797, 'timing/data': 0.00035134097561240196, 'timing/model': 1.1121947600040585, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 0.55}) +[2026-06-23 22:30:16,982][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.14988915622234344, 'train/grad_norm_pre_clip': 20.052255630493164, 'timing/data': 0.00035528303124010563, 'timing/model': 1.1143953390419483, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 0.55}) +[2026-06-23 22:30:21,050][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.03212312236428261, 'train/grad_norm_pre_clip': 17.04654312133789, 'timing/data': 0.0003607580438256264, 'timing/model': 1.117237314931117, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 0.55}) +[2026-06-23 22:30:25,277][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.06255700439214706, 'train/grad_norm_pre_clip': 19.385208129882812, 'timing/data': 0.0003512779949232936, 'timing/model': 1.1112787739839405, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 0.55}) +[2026-06-23 22:30:29,343][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.18358615040779114, 'train/grad_norm_pre_clip': 26.512575149536133, 'timing/data': 0.0004374260315671563, 'timing/model': 1.118873780942522, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 0.55}) +[2026-06-23 22:30:33,407][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.21133480966091156, 'train/grad_norm_pre_clip': 21.487943649291992, 'timing/data': 0.00037157093174755573, 'timing/model': 1.1129293469712138, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 0.55}) +[2026-06-23 22:30:37,485][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.08361221104860306, 'train/grad_norm_pre_clip': 14.176271438598633, 'timing/data': 0.0006084189517423511, 'timing/model': 1.11943558906205, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 0.55}) +[2026-06-23 22:30:41,750][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.09248161315917969, 'train/grad_norm_pre_clip': 15.639357566833496, 'timing/data': 0.00041728606447577477, 'timing/model': 1.1124961510067806, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 0.55}) +[2026-06-23 22:30:45,822][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.07562604546546936, 'train/grad_norm_pre_clip': 11.047600746154785, 'timing/data': 0.00036397192161530256, 'timing/model': 1.1177342280279845, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.55}) +[2026-06-23 22:30:49,904][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.0741330087184906, 'train/grad_norm_pre_clip': 16.619314193725586, 'timing/data': 0.0003569439286366105, 'timing/model': 1.114445305080153, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 0.55}) +[2026-06-23 22:30:53,983][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.0859428271651268, 'train/grad_norm_pre_clip': 15.221034049987793, 'timing/data': 0.00044758105650544167, 'timing/model': 1.1207140759797767, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 0.55}) +[2026-06-23 22:30:58,262][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.10950319468975067, 'train/grad_norm_pre_clip': 12.028231620788574, 'timing/data': 0.00036310404539108276, 'timing/model': 1.1208882839418948, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.55}) +[2026-06-23 22:31:02,346][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.06024516746401787, 'train/grad_norm_pre_clip': 13.300426483154297, 'timing/data': 0.00036669301334768534, 'timing/model': 1.1218310300027952, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 0.55}) +[2026-06-23 22:31:06,442][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.11335832625627518, 'train/grad_norm_pre_clip': 21.809358596801758, 'timing/data': 0.0004043090157210827, 'timing/model': 1.124489277950488, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 0.55}) +[2026-06-23 22:31:10,519][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.12145021557807922, 'train/grad_norm_pre_clip': 29.72394561767578, 'timing/data': 0.0003591339336708188, 'timing/model': 1.1179928629426286, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 0.55}) +[2026-06-23 22:31:14,758][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.1440735012292862, 'train/grad_norm_pre_clip': 23.29362678527832, 'timing/data': 0.00038204400334507227, 'timing/model': 1.1139353819889948, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 0.55}) +[2026-06-23 22:31:18,820][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.06289318203926086, 'train/grad_norm_pre_clip': 11.080718994140625, 'timing/data': 0.00035665498580783606, 'timing/model': 1.1118059610016644, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 0.55}) +[2026-06-23 22:31:22,896][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.0582822822034359, 'train/grad_norm_pre_clip': 11.065051078796387, 'timing/data': 0.0003681840607896447, 'timing/model': 1.1133612269768491, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 0.55}) +[2026-06-23 22:31:26,985][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.049794815480709076, 'train/grad_norm_pre_clip': 13.414912223815918, 'timing/data': 0.0003600060008466244, 'timing/model': 1.1222244830569252, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.55}) +[2026-06-23 22:31:31,223][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.02305126190185547, 'train/grad_norm_pre_clip': 25.3914737701416, 'timing/data': 0.0004967550048604608, 'timing/model': 1.1139958429848775, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 0.55}) +[2026-06-23 22:31:35,289][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.04467932507395744, 'train/grad_norm_pre_clip': 11.30885124206543, 'timing/data': 0.0003729729214683175, 'timing/model': 1.1136860639089718, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 0.55}) +[2026-06-23 22:31:39,388][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.1900203377008438, 'train/grad_norm_pre_clip': 23.910200119018555, 'timing/data': 0.00036144896876066923, 'timing/model': 1.115552722942084, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 0.55}) +[2026-06-23 22:31:43,468][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.10168574750423431, 'train/grad_norm_pre_clip': 17.21433448791504, 'timing/data': 0.00035895302426069975, 'timing/model': 1.1162805170752108, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 0.55}) +[2026-06-23 22:31:47,716][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.10231620073318481, 'train/grad_norm_pre_clip': 17.690465927124023, 'timing/data': 0.0004961859667673707, 'timing/model': 1.1212792920414358, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 0.56}) +[2026-06-23 22:31:51,796][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.1668805629014969, 'train/grad_norm_pre_clip': 49.744056701660156, 'timing/data': 0.0003590069245547056, 'timing/model': 1.1207754119532183, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 0.56}) +[2026-06-23 22:31:55,880][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.27484971284866333, 'train/grad_norm_pre_clip': 37.58456802368164, 'timing/data': 0.0005635010311380029, 'timing/model': 1.1182974959956482, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 0.56}) +[2026-06-23 22:31:59,958][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.0932014137506485, 'train/grad_norm_pre_clip': 39.38059997558594, 'timing/data': 0.0004901420325040817, 'timing/model': 1.1156226149760187, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 0.56}) +[2026-06-23 22:32:04,195][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.40687933564186096, 'train/grad_norm_pre_clip': 35.047813415527344, 'timing/data': 0.00037404196336865425, 'timing/model': 1.118106909096241, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 0.56}) +[2026-06-23 22:32:08,266][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.1954551786184311, 'train/grad_norm_pre_clip': 42.98733901977539, 'timing/data': 0.0005679799942299724, 'timing/model': 1.1185443010181189, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.56}) +[2026-06-23 22:32:12,351][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.0695493072271347, 'train/grad_norm_pre_clip': 24.331453323364258, 'timing/data': 0.000515194027684629, 'timing/model': 1.118042714893818, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 0.56}) +[2026-06-23 22:32:16,433][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.026748061180114746, 'train/grad_norm_pre_clip': 25.805191040039062, 'timing/data': 0.0005616809939965606, 'timing/model': 1.1156684729503468, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 0.56}) +[2026-06-23 22:32:20,665][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.12758108973503113, 'train/grad_norm_pre_clip': 44.72025680541992, 'timing/data': 0.00036228704266250134, 'timing/model': 1.113852899055928, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 0.56}) +[2026-06-23 22:32:24,727][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.03589729592204094, 'train/grad_norm_pre_clip': 20.131567001342773, 'timing/data': 0.0006039360305294394, 'timing/model': 1.1122791119851172, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 0.56}) +[2026-06-23 22:32:28,797][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.15305635333061218, 'train/grad_norm_pre_clip': 17.582569122314453, 'timing/data': 0.0005919040413573384, 'timing/model': 1.1136185410432518, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 0.56}) +[2026-06-23 22:32:32,882][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.14670757949352264, 'train/grad_norm_pre_clip': 23.094648361206055, 'timing/data': 0.0004072710871696472, 'timing/model': 1.1174998370697722, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 0.56}) +[2026-06-23 22:32:37,125][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.06925829499959946, 'train/grad_norm_pre_clip': 21.43735694885254, 'timing/data': 0.0005105230957269669, 'timing/model': 1.111680343048647, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 0.56}) +[2026-06-23 22:32:41,202][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.07932072132825851, 'train/grad_norm_pre_clip': 16.15740203857422, 'timing/data': 0.00035905197728425264, 'timing/model': 1.117161049041897, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 0.56}) +[2026-06-23 22:32:45,275][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.1803690493106842, 'train/grad_norm_pre_clip': 20.888465881347656, 'timing/data': 0.00035505404230207205, 'timing/model': 1.1209687190130353, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 0.56}) +[2026-06-23 22:32:49,371][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.07145398110151291, 'train/grad_norm_pre_clip': 27.788190841674805, 'timing/data': 0.0003670899895951152, 'timing/model': 1.1254997629439458, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.56}) +[2026-06-23 22:32:53,644][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.13444289565086365, 'train/grad_norm_pre_clip': 31.275850296020508, 'timing/data': 0.0005806840490549803, 'timing/model': 1.124765192042105, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 0.56}) +[2026-06-23 22:32:57,736][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.08355247974395752, 'train/grad_norm_pre_clip': 23.200138092041016, 'timing/data': 0.0003588239196687937, 'timing/model': 1.1207792910281569, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 0.56}) +[2026-06-23 22:33:01,814][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.124612957239151, 'train/grad_norm_pre_clip': 17.957611083984375, 'timing/data': 0.00036910700146108866, 'timing/model': 1.120288830017671, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 0.56}) +[2026-06-23 22:33:05,896][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.06388989835977554, 'train/grad_norm_pre_clip': 8.356343269348145, 'timing/data': 0.00038872798904776573, 'timing/model': 1.1169549518963322, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 0.56}) +[2026-06-23 22:33:10,162][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.06193501129746437, 'train/grad_norm_pre_clip': 11.600156784057617, 'timing/data': 0.0005080279661342502, 'timing/model': 1.1166283970233053, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 0.56}) +[2026-06-23 22:33:14,238][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.1700638234615326, 'train/grad_norm_pre_clip': 18.15006446838379, 'timing/data': 0.0004331920063123107, 'timing/model': 1.115199530031532, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 0.56}) +[2026-06-23 22:33:18,317][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.05520275980234146, 'train/grad_norm_pre_clip': 9.596818923950195, 'timing/data': 0.0003538120072335005, 'timing/model': 1.1145242520142347, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 0.56}) +[2026-06-23 22:33:22,401][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.14894616603851318, 'train/grad_norm_pre_clip': 19.63787269592285, 'timing/data': 0.0003651930019259453, 'timing/model': 1.119552206946537, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 0.56}) +[2026-06-23 22:33:26,666][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.035735875368118286, 'train/grad_norm_pre_clip': 15.192259788513184, 'timing/data': 0.00037444394547492266, 'timing/model': 1.1220093310112134, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 0.56}) +[2026-06-23 22:33:30,763][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.17224545776844025, 'train/grad_norm_pre_clip': 31.12551498413086, 'timing/data': 0.0003647370031103492, 'timing/model': 1.1147226250031963, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.57}) +[2026-06-23 22:33:34,831][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.06037285551428795, 'train/grad_norm_pre_clip': 9.068646430969238, 'timing/data': 0.0003605389501899481, 'timing/model': 1.1175045790150762, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 0.57}) +[2026-06-23 22:33:38,911][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.0883631631731987, 'train/grad_norm_pre_clip': 24.04493522644043, 'timing/data': 0.0003557440359145403, 'timing/model': 1.1197768930578604, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 0.57}) +[2026-06-23 22:33:43,156][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.18169881403446198, 'train/grad_norm_pre_clip': 17.150802612304688, 'timing/data': 0.00041427602991461754, 'timing/model': 1.120523755904287, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 0.57}) +[2026-06-23 22:33:47,239][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.08443087339401245, 'train/grad_norm_pre_clip': 24.834482192993164, 'timing/data': 0.000643986975774169, 'timing/model': 1.1208610830362886, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 0.57}) +[2026-06-23 22:33:51,327][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.024615708738565445, 'train/grad_norm_pre_clip': 16.0758113861084, 'timing/data': 0.0006210290594026446, 'timing/model': 1.1213914040708914, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.57}) +[2026-06-23 22:33:55,407][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.28083154559135437, 'train/grad_norm_pre_clip': 41.93736267089844, 'timing/data': 0.000550503027625382, 'timing/model': 1.115979859023355, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 0.57}) +[2026-06-23 22:33:59,699][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.10713448375463486, 'train/grad_norm_pre_clip': 18.340425491333008, 'timing/data': 0.00036194396670907736, 'timing/model': 1.1224529830506071, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 0.57}) +[2026-06-23 22:34:03,815][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.18773823976516724, 'train/grad_norm_pre_clip': 20.250642776489258, 'timing/data': 0.0003539969911798835, 'timing/model': 1.128973692888394, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 0.57}) +[2026-06-23 22:34:07,891][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.019523683935403824, 'train/grad_norm_pre_clip': 13.622426986694336, 'timing/data': 0.00044776604045182467, 'timing/model': 1.1186617870116606, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 0.57}) +[2026-06-23 22:34:11,988][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.1079297810792923, 'train/grad_norm_pre_clip': 17.802104949951172, 'timing/data': 0.0003623249940574169, 'timing/model': 1.1264935770304874, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.57}) +[2026-06-23 22:34:16,245][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.13769449293613434, 'train/grad_norm_pre_clip': 33.280059814453125, 'timing/data': 0.00041570805478841066, 'timing/model': 1.1213619229383767, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 0.57}) +[2026-06-23 22:34:20,333][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.02627428062260151, 'train/grad_norm_pre_clip': 27.1601619720459, 'timing/data': 0.0003636769251897931, 'timing/model': 1.12045946798753, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.57}) +[2026-06-23 22:34:24,434][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.0819828063249588, 'train/grad_norm_pre_clip': 21.628576278686523, 'timing/data': 0.0005385130643844604, 'timing/model': 1.1122752560768276, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 0.57}) +[2026-06-23 22:34:28,517][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.06495342403650284, 'train/grad_norm_pre_clip': 61.34267807006836, 'timing/data': 0.00037088198587298393, 'timing/model': 1.1198653690516949, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 0.57}) +[2026-06-23 22:34:32,777][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.284565806388855, 'train/grad_norm_pre_clip': 54.83528518676758, 'timing/data': 0.00035968003794550896, 'timing/model': 1.1146997690666467, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 0.57}) +[2026-06-23 22:34:36,849][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.29407092928886414, 'train/grad_norm_pre_clip': 40.23829650878906, 'timing/data': 0.0004134930204600096, 'timing/model': 1.1163939989637583, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 0.57}) +[2026-06-23 22:34:40,918][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.08022081106901169, 'train/grad_norm_pre_clip': 41.98326873779297, 'timing/data': 0.00036887999158352613, 'timing/model': 1.1118693850003183, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 0.57}) +[2026-06-23 22:34:44,994][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.09853328764438629, 'train/grad_norm_pre_clip': 25.4239501953125, 'timing/data': 0.00035949796438217163, 'timing/model': 1.1100078149465844, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 0.57}) +[2026-06-23 22:34:49,232][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.007320990785956383, 'train/grad_norm_pre_clip': 16.438600540161133, 'timing/data': 0.0003585430094972253, 'timing/model': 1.1187187409959733, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 0.57}) +[2026-06-23 22:34:53,308][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.037903476506471634, 'train/grad_norm_pre_clip': 11.806702613830566, 'timing/data': 0.0003569630207493901, 'timing/model': 1.1156603230629116, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.57}) +[2026-06-23 22:34:57,378][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.07175164669752121, 'train/grad_norm_pre_clip': 23.90496826171875, 'timing/data': 0.0003485620254650712, 'timing/model': 1.118162986007519, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 0.57}) +[2026-06-23 22:35:01,453][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.48021554946899414, 'train/grad_norm_pre_clip': 54.67338943481445, 'timing/data': 0.000578387058340013, 'timing/model': 1.1185709270648658, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 0.57}) +[2026-06-23 22:35:05,704][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.23202627897262573, 'train/grad_norm_pre_clip': 39.52178955078125, 'timing/data': 0.0003722480032593012, 'timing/model': 1.118962597916834, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 0.57}) +[2026-06-23 22:35:09,786][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.06424778699874878, 'train/grad_norm_pre_clip': 17.65081024169922, 'timing/data': 0.0003586140228435397, 'timing/model': 1.11759807204362, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 0.57}) +[2026-06-23 22:35:13,868][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.05890095606446266, 'train/grad_norm_pre_clip': 25.99987030029297, 'timing/data': 0.0003592959837988019, 'timing/model': 1.1246216190047562, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 0.57}) +[2026-06-23 22:35:17,952][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.08294083178043365, 'train/grad_norm_pre_clip': 14.910636901855469, 'timing/data': 0.0004087509587407112, 'timing/model': 1.1163811800070107, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 0.58}) +[2026-06-23 22:35:22,192][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.09220172464847565, 'train/grad_norm_pre_clip': 12.622681617736816, 'timing/data': 0.00041542307008057833, 'timing/model': 1.117954891989939, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 0.58}) +[2026-06-23 22:35:26,272][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.08140958100557327, 'train/grad_norm_pre_clip': 21.505855560302734, 'timing/data': 0.0003672110615298152, 'timing/model': 1.116670379997231, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 0.58}) +[2026-06-23 22:35:30,374][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.11141163855791092, 'train/grad_norm_pre_clip': 24.614456176757812, 'timing/data': 0.00036673492286354303, 'timing/model': 1.1267388329142705, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 0.58}) +[2026-06-23 22:35:34,490][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.053420599550008774, 'train/grad_norm_pre_clip': 16.180334091186523, 'timing/data': 0.0003611890133470297, 'timing/model': 1.133856002939865, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.58}) +[2026-06-23 22:35:38,752][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.06481103599071503, 'train/grad_norm_pre_clip': 10.983450889587402, 'timing/data': 0.00036943797022104263, 'timing/model': 1.1186636660713702, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 0.58}) +[2026-06-23 22:35:42,838][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.029030852019786835, 'train/grad_norm_pre_clip': 11.058233261108398, 'timing/data': 0.0004305969923734665, 'timing/model': 1.1152092309202999, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 0.58}) +[2026-06-23 22:35:46,921][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.11111616343259811, 'train/grad_norm_pre_clip': 12.465083122253418, 'timing/data': 0.0005217120051383972, 'timing/model': 1.1197696609888226, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 0.58}) +[2026-06-23 22:35:51,015][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.043840017169713974, 'train/grad_norm_pre_clip': 25.034744262695312, 'timing/data': 0.0004079240607097745, 'timing/model': 1.1281289889011532, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 0.58}) +[2026-06-23 22:35:55,275][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.027369754388928413, 'train/grad_norm_pre_clip': 11.376875877380371, 'timing/data': 0.00038303190376609564, 'timing/model': 1.1168559469515458, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 0.58}) +[2026-06-23 22:35:59,373][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.03137470409274101, 'train/grad_norm_pre_clip': 13.101268768310547, 'timing/data': 0.00036854296922683716, 'timing/model': 1.124716418911703, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 0.58}) +[2026-06-23 22:36:03,496][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.09389153122901917, 'train/grad_norm_pre_clip': 28.485614776611328, 'timing/data': 0.00038305798079818487, 'timing/model': 1.1271334590855986, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 0.58}) +[2026-06-23 22:36:07,577][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.04502406343817711, 'train/grad_norm_pre_clip': 18.224014282226562, 'timing/data': 0.0003657860215753317, 'timing/model': 1.1123698910232633, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 0.58}) +[2026-06-23 22:36:11,831][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.24081425368785858, 'train/grad_norm_pre_clip': 37.90496826171875, 'timing/data': 0.0004238070687279105, 'timing/model': 1.1209871199680492, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 0.58}) +[2026-06-23 22:36:15,911][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1500 +[2026-06-23 22:37:20,846][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1500 +[2026-06-23 22:37:20,847][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1500 +[2026-06-23 22:37:59,196][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1500 +[2026-06-23 22:45:53,150][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-23 22:45:53,151][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.09245400875806808, 'train/grad_norm_pre_clip': 19.752559661865234, 'eval/action_loss/seconds': 64.932689085952, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.049052029848098755, 'eval/latency_2/loss': 0.049052029848098755, 'eval/flappy/loss': 0.049052029848098755, 'eval/flappy/latency_2/loss': 0.049052029848098755, 'eval/flap_precision': 0.9223803363518758, 'eval/flap_recall': 0.9674355495251018, 'eval/flap_f1': 0.9443708609271523, 'eval/flap_rate_model': 0.21472222222222223, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9443708609271523, 'eval/latency_2/flap_precision': 0.9223803363518758, 'eval/latency_2/flap_recall': 0.9674355495251018, 'eval/latency_2/flap_f1': 0.9443708609271523, 'eval/latency_2/flap_rate_model': 0.21472222222222223, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9443708609271523, 'eval/action_classification/data_wait_seconds': 0.5387763977050781, 'eval/action_classification/predict_seconds': 37.67637252807617, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 292.5400000000077, 'rl_games_eval/mid_train/mean_length': 2375.8, 'rl_games_eval/mid_train/std_reward': 145.61193082987145, 'rl_games_eval/mid_train/std_length': 1173.9889948376858, 'rl_games_eval/mid_train/macro_mean_reward': 292.5400000000077, 'rl_games_eval/mid_train/macro_mean_length': 2375.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 292.5400000000077, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2375.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 145.61193082987145, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1173.9889948376858, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 292.5400000000077, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003613049630075693, 'timing/model': 1.1241748849861324, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.58}) +[2026-06-23 22:45:53,151][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_1500_state +[2026-06-23 22:45:53,154][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 22:45:56,838][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_1500_state/model.safetensors +[2026-06-23 22:46:04,580][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_1500_state/optimizer.bin +[2026-06-23 22:46:04,581][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_1500_state/sampler.bin +[2026-06-23 22:46:04,581][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-23 22:46:04,581][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-23 22:46:04,626][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-23 22:46:04,627][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 22:46:04,638][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 22:46:08,683][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.164766326546669, 'train/grad_norm_pre_clip': 41.933143615722656, 'timing/data': 0.00040097092278301716, 'timing/model': 1.089017010992393, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 0.58}) +[2026-06-23 22:46:12,656][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.021013738587498665, 'train/grad_norm_pre_clip': 10.819998741149902, 'timing/data': 0.0004047940019518137, 'timing/model': 1.0916223160456866, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 0.58}) +[2026-06-23 22:46:16,637][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.22879737615585327, 'train/grad_norm_pre_clip': 78.60848236083984, 'timing/data': 0.00039253407157957554, 'timing/model': 1.0945467960555106, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 0.58}) +[2026-06-23 22:46:20,621][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.25992390513420105, 'train/grad_norm_pre_clip': 54.34804153442383, 'timing/data': 0.000374219031073153, 'timing/model': 1.0962845269823447, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 0.58}) +[2026-06-23 22:46:24,696][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.02654985710978508, 'train/grad_norm_pre_clip': 10.91562557220459, 'timing/data': 0.00037152902223169804, 'timing/model': 1.0984430310782045, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 0.58}) +[2026-06-23 22:46:28,712][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.050563324242830276, 'train/grad_norm_pre_clip': 19.905019760131836, 'timing/data': 0.0003555850125849247, 'timing/model': 1.1006607479648665, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 0.58}) +[2026-06-23 22:46:32,734][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.20577563345432281, 'train/grad_norm_pre_clip': 16.31821060180664, 'timing/data': 0.0005893540801480412, 'timing/model': 1.1008315669605508, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 0.58}) +[2026-06-23 22:46:36,745][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.22745724022388458, 'train/grad_norm_pre_clip': 35.751487731933594, 'timing/data': 0.00036653305869549513, 'timing/model': 1.0999826970510185, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 0.58}) +[2026-06-23 22:46:40,903][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.0036236948799341917, 'train/grad_norm_pre_clip': 23.892820358276367, 'timing/data': 0.00039336795452982187, 'timing/model': 1.105348202981986, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 0.58}) +[2026-06-23 22:46:44,944][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.15284165740013123, 'train/grad_norm_pre_clip': 40.27194595336914, 'timing/data': 0.0003535069990903139, 'timing/model': 1.1115008660126477, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.58}) +[2026-06-23 22:46:48,997][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.08148445934057236, 'train/grad_norm_pre_clip': 40.87366485595703, 'timing/data': 0.0005562200676649809, 'timing/model': 1.1115796960657462, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 0.58}) +[2026-06-23 22:46:53,117][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.0559646375477314, 'train/grad_norm_pre_clip': 8.436147689819336, 'timing/data': 0.0004079049685969949, 'timing/model': 1.1796652299817652, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 0.59}) +[2026-06-23 22:46:57,155][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.19558176398277283, 'train/grad_norm_pre_clip': 44.33225631713867, 'timing/data': 0.0003723279805853963, 'timing/model': 1.106776015018113, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 0.59}) +[2026-06-23 22:47:01,198][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.20319987833499908, 'train/grad_norm_pre_clip': 19.492942810058594, 'timing/data': 0.0004364399937912822, 'timing/model': 1.108436367008835, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 0.59}) +[2026-06-23 22:47:05,245][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.05807331204414368, 'train/grad_norm_pre_clip': 9.577679634094238, 'timing/data': 0.000366268097423017, 'timing/model': 1.112606661976315, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 0.59}) +[2026-06-23 22:47:09,299][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.046843089163303375, 'train/grad_norm_pre_clip': 12.44154167175293, 'timing/data': 0.00036928895860910416, 'timing/model': 1.1131716979434714, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 0.59}) +[2026-06-23 22:47:13,495][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.02897333912551403, 'train/grad_norm_pre_clip': 11.956047058105469, 'timing/data': 0.0004202589625492692, 'timing/model': 1.1133697980549186, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 0.59}) +[2026-06-23 22:47:17,552][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.0477016419172287, 'train/grad_norm_pre_clip': 8.796697616577148, 'timing/data': 0.0006460709264501929, 'timing/model': 1.108091126079671, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 0.59}) +[2026-06-23 22:47:21,606][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.056785255670547485, 'train/grad_norm_pre_clip': 11.3796968460083, 'timing/data': 0.00036058295518159866, 'timing/model': 1.117039934033528, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 0.59}) +[2026-06-23 22:47:25,664][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.12185033410787582, 'train/grad_norm_pre_clip': 14.359976768493652, 'timing/data': 0.0004252099897712469, 'timing/model': 1.1076645610155538, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.59}) +[2026-06-23 22:47:29,875][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.12554019689559937, 'train/grad_norm_pre_clip': 16.463760375976562, 'timing/data': 0.0005623109173029661, 'timing/model': 1.1090796240605414, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.59}) +[2026-06-23 22:47:33,939][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.183460995554924, 'train/grad_norm_pre_clip': 25.136241912841797, 'timing/data': 0.0003532100236043334, 'timing/model': 1.1129661209415644, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 0.59}) +[2026-06-23 22:47:37,995][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.02374705858528614, 'train/grad_norm_pre_clip': 17.87709617614746, 'timing/data': 0.0004401060286909342, 'timing/model': 1.111043204087764, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 0.59}) +[2026-06-23 22:47:42,043][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.017155135050415993, 'train/grad_norm_pre_clip': 20.24223518371582, 'timing/data': 0.00036535703111439943, 'timing/model': 1.1086868970887735, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 0.59}) +[2026-06-23 22:47:46,183][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.08267662674188614, 'train/grad_norm_pre_clip': 15.953085899353027, 'timing/data': 0.0005530070047825575, 'timing/model': 1.1125283519504592, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 0.59}) +[2026-06-23 22:47:50,260][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.016189785674214363, 'train/grad_norm_pre_clip': 18.34346580505371, 'timing/data': 0.0005279489560052752, 'timing/model': 1.1200720110209659, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 0.59}) +[2026-06-23 22:47:54,327][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.005870708730071783, 'train/grad_norm_pre_clip': 15.448395729064941, 'timing/data': 0.0003668570425361395, 'timing/model': 1.1140324430307373, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 0.59}) +[2026-06-23 22:47:58,395][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.050190672278404236, 'train/grad_norm_pre_clip': 15.230147361755371, 'timing/data': 0.00035246205516159534, 'timing/model': 1.113940624985844, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 0.59}) +[2026-06-23 22:48:02,547][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.2759883403778076, 'train/grad_norm_pre_clip': 32.405757904052734, 'timing/data': 0.0005467688897624612, 'timing/model': 1.1204104499192908, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 0.59}) +[2026-06-23 22:48:06,610][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.03683257848024368, 'train/grad_norm_pre_clip': 37.16054153442383, 'timing/data': 0.00043526000808924437, 'timing/model': 1.1167137320153415, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.59}) +[2026-06-23 22:48:10,671][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.0028419338632375, 'train/grad_norm_pre_clip': 39.6429443359375, 'timing/data': 0.0004081861115992069, 'timing/model': 1.1121620249468833, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 0.59}) +[2026-06-23 22:48:14,805][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.11469987779855728, 'train/grad_norm_pre_clip': 15.529345512390137, 'timing/data': 0.0003453269600868225, 'timing/model': 1.1237968259956688, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 0.59}) +[2026-06-23 22:48:19,045][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.13386358320713043, 'train/grad_norm_pre_clip': 26.363590240478516, 'timing/data': 0.0004578439984470606, 'timing/model': 1.113338717026636, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 0.59}) +[2026-06-23 22:48:23,109][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.10595006495714188, 'train/grad_norm_pre_clip': 28.05901527404785, 'timing/data': 0.00035572401247918606, 'timing/model': 1.1166420469526201, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 0.59}) +[2026-06-23 22:48:27,177][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.08031364530324936, 'train/grad_norm_pre_clip': 20.02008628845215, 'timing/data': 0.0004042170476168394, 'timing/model': 1.1166354810120538, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.59}) +[2026-06-23 22:48:31,256][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.0903119295835495, 'train/grad_norm_pre_clip': 13.295083999633789, 'timing/data': 0.0003510640235617757, 'timing/model': 1.1171730789355934, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 0.59}) +[2026-06-23 22:48:35,488][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.06263981759548187, 'train/grad_norm_pre_clip': 31.828275680541992, 'timing/data': 0.0003581159980967641, 'timing/model': 1.1249962690053508, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 0.59}) +[2026-06-23 22:48:39,555][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.0266132690012455, 'train/grad_norm_pre_clip': 29.8350887298584, 'timing/data': 0.000358329969458282, 'timing/model': 1.1111216250574216, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 0.6}) +[2026-06-23 22:48:43,627][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.06959867477416992, 'train/grad_norm_pre_clip': 33.87309265136719, 'timing/data': 0.0003570390399545431, 'timing/model': 1.115103763062507, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 0.6}) +[2026-06-23 22:48:47,697][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.1096649020910263, 'train/grad_norm_pre_clip': 20.214855194091797, 'timing/data': 0.0003630920546129346, 'timing/model': 1.115474677993916, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.6}) +[2026-06-23 22:48:51,936][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.008627704344689846, 'train/grad_norm_pre_clip': 15.19896411895752, 'timing/data': 0.000367108965292573, 'timing/model': 1.1139823329867795, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 0.6}) +[2026-06-23 22:48:56,003][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.0962267592549324, 'train/grad_norm_pre_clip': 17.53040313720703, 'timing/data': 0.00035640306305140257, 'timing/model': 1.1161398069234565, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 0.6}) +[2026-06-23 22:49:00,090][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.1463773399591446, 'train/grad_norm_pre_clip': 16.39984130859375, 'timing/data': 0.00042985694017261267, 'timing/model': 1.123915954027325, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 0.6}) +[2026-06-23 22:49:04,171][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.03518844395875931, 'train/grad_norm_pre_clip': 12.119720458984375, 'timing/data': 0.00041898107156157494, 'timing/model': 1.1160016359062865, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 0.6}) +[2026-06-23 22:49:08,402][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.13237975537776947, 'train/grad_norm_pre_clip': 15.554434776306152, 'timing/data': 0.0004641859559342265, 'timing/model': 1.1082935240119696, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 0.6}) +[2026-06-23 22:49:12,486][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.04167554900050163, 'train/grad_norm_pre_clip': 14.89467716217041, 'timing/data': 0.0005111440550535917, 'timing/model': 1.118390463059768, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 0.6}) +[2026-06-23 22:49:16,561][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.16282473504543304, 'train/grad_norm_pre_clip': 11.818501472473145, 'timing/data': 0.0004942750092595816, 'timing/model': 1.1118598590837792, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 0.6}) +[2026-06-23 22:49:20,644][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.1130334809422493, 'train/grad_norm_pre_clip': 22.657421112060547, 'timing/data': 0.00036967499181628227, 'timing/model': 1.1192392479861155, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 0.6}) +[2026-06-23 22:49:24,878][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.019389282912015915, 'train/grad_norm_pre_clip': 8.266104698181152, 'timing/data': 0.0006015960825607181, 'timing/model': 1.1175750999245793, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 0.6}) +[2026-06-23 22:49:28,958][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.025222735479474068, 'train/grad_norm_pre_clip': 18.455297470092773, 'timing/data': 0.0003560739569365978, 'timing/model': 1.1164612260181457, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.6}) +[2026-06-23 22:49:33,042][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.08734271675348282, 'train/grad_norm_pre_clip': 15.060934066772461, 'timing/data': 0.0005841240053996444, 'timing/model': 1.1197115570539609, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 0.6}) +[2026-06-23 22:49:37,121][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.06528947502374649, 'train/grad_norm_pre_clip': 6.906284809112549, 'timing/data': 0.00037650903686881065, 'timing/model': 1.1242581490660086, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 0.6}) +[2026-06-23 22:49:41,363][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.05548069626092911, 'train/grad_norm_pre_clip': 11.757831573486328, 'timing/data': 0.00040585105307400227, 'timing/model': 1.118504124111496, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 0.6}) +[2026-06-23 22:49:45,441][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.14893196523189545, 'train/grad_norm_pre_clip': 49.428077697753906, 'timing/data': 0.0003617069451138377, 'timing/model': 1.116495501017198, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 0.6}) +[2026-06-23 22:49:49,528][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.09381666779518127, 'train/grad_norm_pre_clip': 74.53114318847656, 'timing/data': 0.00039402302354574203, 'timing/model': 1.1175546849844977, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 0.6}) +[2026-06-23 22:49:53,618][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.07542296499013901, 'train/grad_norm_pre_clip': 19.944469451904297, 'timing/data': 0.000357283977791667, 'timing/model': 1.1279585140291601, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 0.6}) +[2026-06-23 22:49:57,867][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.05135028064250946, 'train/grad_norm_pre_clip': 17.419618606567383, 'timing/data': 0.00041131197940558195, 'timing/model': 1.1173698999918997, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 0.6}) +[2026-06-23 22:50:01,943][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.07563240826129913, 'train/grad_norm_pre_clip': 13.515146255493164, 'timing/data': 0.00040005799382925034, 'timing/model': 1.1166594439418986, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 0.6}) +[2026-06-23 22:50:06,025][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.2925734221935272, 'train/grad_norm_pre_clip': 28.70603370666504, 'timing/data': 0.0003954690182581544, 'timing/model': 1.1116843089694157, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 0.6}) +[2026-06-23 22:50:10,102][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.03427902236580849, 'train/grad_norm_pre_clip': 16.198055267333984, 'timing/data': 0.00047019205521792173, 'timing/model': 1.1165790819795802, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.6}) +[2026-06-23 22:50:14,350][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.15150386095046997, 'train/grad_norm_pre_clip': 18.00794219970703, 'timing/data': 0.0004314369289204478, 'timing/model': 1.1154688679380342, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 0.6}) +[2026-06-23 22:50:18,429][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.10211500525474548, 'train/grad_norm_pre_clip': 24.408342361450195, 'timing/data': 0.0004115119809284806, 'timing/model': 1.114338209037669, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 0.6}) +[2026-06-23 22:50:22,513][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.0859595239162445, 'train/grad_norm_pre_clip': 23.7133731842041, 'timing/data': 0.00036700896453112364, 'timing/model': 1.1182842770358548, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 0.6}) +[2026-06-23 22:50:26,607][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.25527212023735046, 'train/grad_norm_pre_clip': 43.24734115600586, 'timing/data': 0.00041879399213939905, 'timing/model': 1.1206825571134686, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 0.61}) +[2026-06-23 22:50:30,879][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.3437558710575104, 'train/grad_norm_pre_clip': 25.733367919921875, 'timing/data': 0.0003701929235830903, 'timing/model': 1.1229312680661678, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 0.61}) +[2026-06-23 22:50:34,966][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.12215688824653625, 'train/grad_norm_pre_clip': 18.748573303222656, 'timing/data': 0.0003625820390880108, 'timing/model': 1.127224451978691, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 0.61}) +[2026-06-23 22:50:39,050][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.1796448528766632, 'train/grad_norm_pre_clip': 16.42403793334961, 'timing/data': 0.0003636759938672185, 'timing/model': 1.1209385299589485, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 0.61}) +[2026-06-23 22:50:43,145][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.010480504482984543, 'train/grad_norm_pre_clip': 17.259666442871094, 'timing/data': 0.0004843330243602395, 'timing/model': 1.121025540982373, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 0.61}) +[2026-06-23 22:50:47,396][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.12312851846218109, 'train/grad_norm_pre_clip': 19.96398162841797, 'timing/data': 0.0006261760136112571, 'timing/model': 1.1214440839830786, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 0.61}) +[2026-06-23 22:50:51,473][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.17357608675956726, 'train/grad_norm_pre_clip': 13.319755554199219, 'timing/data': 0.00037355406675487757, 'timing/model': 1.114462316967547, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.61}) +[2026-06-23 22:50:55,558][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.1581655591726303, 'train/grad_norm_pre_clip': 11.481673240661621, 'timing/data': 0.0005643740296363831, 'timing/model': 1.1210480789886788, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 0.61}) +[2026-06-23 22:50:59,644][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.2137160301208496, 'train/grad_norm_pre_clip': 15.55734920501709, 'timing/data': 0.00035862100776284933, 'timing/model': 1.1209030579775572, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 0.61}) +[2026-06-23 22:51:03,904][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.059918370097875595, 'train/grad_norm_pre_clip': 18.854639053344727, 'timing/data': 0.00036151602398604155, 'timing/model': 1.1309487789403647, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 0.61}) +[2026-06-23 22:51:07,967][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.04567168280482292, 'train/grad_norm_pre_clip': 7.741876602172852, 'timing/data': 0.0004925569519400597, 'timing/model': 1.116259939968586, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 0.61}) +[2026-06-23 22:51:12,050][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.05125753581523895, 'train/grad_norm_pre_clip': 21.833599090576172, 'timing/data': 0.000406423001550138, 'timing/model': 1.1159181940602139, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 0.61}) +[2026-06-23 22:51:16,123][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.10888144373893738, 'train/grad_norm_pre_clip': 14.010078430175781, 'timing/data': 0.0006112219998613, 'timing/model': 1.1170372479828075, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 0.61}) +[2026-06-23 22:51:20,390][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.014074581675231457, 'train/grad_norm_pre_clip': 15.503541946411133, 'timing/data': 0.00038729608058929443, 'timing/model': 1.1175051790196449, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 0.61}) +[2026-06-23 22:51:24,468][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.07158991694450378, 'train/grad_norm_pre_clip': 18.874574661254883, 'timing/data': 0.0003606959944590926, 'timing/model': 1.1194819830125198, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 0.61}) +[2026-06-23 22:51:28,557][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.1023821234703064, 'train/grad_norm_pre_clip': 16.191984176635742, 'timing/data': 0.00036232906859368086, 'timing/model': 1.114445999963209, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 0.61}) +[2026-06-23 22:51:32,636][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.09531661868095398, 'train/grad_norm_pre_clip': 22.290233612060547, 'timing/data': 0.0003601410426199436, 'timing/model': 1.1201902580214664, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.61}) +[2026-06-23 22:51:36,884][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.16126014292240143, 'train/grad_norm_pre_clip': 16.302642822265625, 'timing/data': 0.00035989005118608475, 'timing/model': 1.1217518179910257, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 0.61}) +[2026-06-23 22:51:40,957][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.011396612972021103, 'train/grad_norm_pre_clip': 23.583337783813477, 'timing/data': 0.0003611380234360695, 'timing/model': 1.1168175059137866, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 0.61}) +[2026-06-23 22:51:45,054][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.10304243117570877, 'train/grad_norm_pre_clip': 22.197772979736328, 'timing/data': 0.0003571840934455395, 'timing/model': 1.1206423540133983, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 0.61}) +[2026-06-23 22:51:49,141][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.05034884810447693, 'train/grad_norm_pre_clip': 30.57074546813965, 'timing/data': 0.0003555750008672476, 'timing/model': 1.1179040500428528, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 0.61}) +[2026-06-23 22:51:53,407][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.11075996607542038, 'train/grad_norm_pre_clip': 18.941818237304688, 'timing/data': 0.0003657009219750762, 'timing/model': 1.122437478043139, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 0.61}) +[2026-06-23 22:51:57,496][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.09573906660079956, 'train/grad_norm_pre_clip': 28.008373260498047, 'timing/data': 0.0005475629586726427, 'timing/model': 1.1247629519784823, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 0.61}) +[2026-06-23 22:52:01,579][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.14392037689685822, 'train/grad_norm_pre_clip': 27.63394546508789, 'timing/data': 0.00035303004551678896, 'timing/model': 1.1181603480363265, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 0.61}) +[2026-06-23 22:52:05,656][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.008588760159909725, 'train/grad_norm_pre_clip': 18.47829246520996, 'timing/data': 0.00035137601662427187, 'timing/model': 1.1139482780126855, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 0.61}) +[2026-06-23 22:52:09,922][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.181100994348526, 'train/grad_norm_pre_clip': 29.528560638427734, 'timing/data': 0.0003571760607883334, 'timing/model': 1.1161956050200388, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 0.61}) +[2026-06-23 22:52:14,003][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.06874935328960419, 'train/grad_norm_pre_clip': 32.17352294921875, 'timing/data': 0.00040614104364067316, 'timing/model': 1.1239760589087382, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.62}) +[2026-06-23 22:52:18,082][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.05048786476254463, 'train/grad_norm_pre_clip': 22.21080207824707, 'timing/data': 0.0003569879336282611, 'timing/model': 1.116613126010634, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 0.62}) +[2026-06-23 22:52:22,166][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.16058161854743958, 'train/grad_norm_pre_clip': 17.30284309387207, 'timing/data': 0.0006205450044944882, 'timing/model': 1.1229294160148129, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 0.62}) +[2026-06-23 22:52:26,433][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.01564028486609459, 'train/grad_norm_pre_clip': 15.52724838256836, 'timing/data': 0.0007231839699670672, 'timing/model': 1.118179709999822, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 0.62}) +[2026-06-23 22:52:30,515][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.020946983247995377, 'train/grad_norm_pre_clip': 18.410303115844727, 'timing/data': 0.0003601700300350785, 'timing/model': 1.1161091319518164, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 0.62}) +[2026-06-23 22:52:34,619][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.28766393661499023, 'train/grad_norm_pre_clip': 76.78941345214844, 'timing/data': 0.00035468104761093855, 'timing/model': 1.1193221709690988, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 0.62}) +[2026-06-23 22:52:38,698][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.34966763854026794, 'train/grad_norm_pre_clip': 35.43019485473633, 'timing/data': 0.00036522303707897663, 'timing/model': 1.120812947046943, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 0.62}) +[2026-06-23 22:52:42,948][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.040035463869571686, 'train/grad_norm_pre_clip': 43.5249137878418, 'timing/data': 0.00036422896664589643, 'timing/model': 1.1181319740135223, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 0.62}) +[2026-06-23 22:52:47,038][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.30521297454833984, 'train/grad_norm_pre_clip': 47.02304458618164, 'timing/data': 0.00035926804412156343, 'timing/model': 1.1165160249220207, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 0.62}) +[2026-06-23 22:52:51,133][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.08551830053329468, 'train/grad_norm_pre_clip': 23.281766891479492, 'timing/data': 0.0005863099358975887, 'timing/model': 1.1183881480246782, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 0.62}) +[2026-06-23 22:52:55,235][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1600 +[2026-06-23 22:53:59,766][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1600 +[2026-06-23 22:53:59,767][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1600 +[2026-06-23 22:54:38,804][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1600 +[2026-06-23 22:54:38,805][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.02720184437930584, 'train/grad_norm_pre_clip': 14.830772399902344, 'eval/action_loss/seconds': 64.52864858601242, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.05783766135573387, 'eval/latency_2/loss': 0.05783766135573387, 'eval/flappy/loss': 0.05783766135573387, 'eval/flappy/latency_2/loss': 0.05783766135573387, 'eval/flap_precision': 0.9103362391033624, 'eval/flap_recall': 0.9918588873812755, 'eval/flap_f1': 0.9493506493506494, 'eval/flap_rate_model': 0.22305555555555556, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9493506493506495, 'eval/latency_2/flap_precision': 0.9103362391033624, 'eval/latency_2/flap_recall': 0.9918588873812755, 'eval/latency_2/flap_f1': 0.9493506493506494, 'eval/latency_2/flap_rate_model': 0.22305555555555556, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9493506493506495, 'eval/action_classification/data_wait_seconds': 0.5794430375099182, 'eval/action_classification/predict_seconds': 38.314632415771484, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0005468559684231877, 'timing/model': 1.120247429003939, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.62}) +[2026-06-23 22:54:42,849][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.026762979105114937, 'train/grad_norm_pre_clip': 17.18074607849121, 'timing/data': 0.0003689780132845044, 'timing/model': 1.1018571270396933, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 0.62}) +[2026-06-23 22:54:46,890][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.1314813792705536, 'train/grad_norm_pre_clip': 27.685739517211914, 'timing/data': 0.0003666279371827841, 'timing/model': 1.102946744998917, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 0.62}) +[2026-06-23 22:54:50,933][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.1503261774778366, 'train/grad_norm_pre_clip': 38.503990173339844, 'timing/data': 0.0003760410472750664, 'timing/model': 1.1119723620358855, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 0.62}) +[2026-06-23 22:54:55,083][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.01255618967115879, 'train/grad_norm_pre_clip': 14.897497177124023, 'timing/data': 0.00039789697621017694, 'timing/model': 1.1079191099852324, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 0.62}) +[2026-06-23 22:54:59,142][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.07823700457811356, 'train/grad_norm_pre_clip': 18.792617797851562, 'timing/data': 0.0003765369765460491, 'timing/model': 1.1193417969625443, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.62}) +[2026-06-23 22:55:03,217][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.11034572869539261, 'train/grad_norm_pre_clip': 11.288847923278809, 'timing/data': 0.0005685789510607719, 'timing/model': 1.1122189909219742, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 0.62}) +[2026-06-23 22:55:07,284][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.099570631980896, 'train/grad_norm_pre_clip': 10.490921020507812, 'timing/data': 0.0003610330168157816, 'timing/model': 1.1131039640167728, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 0.62}) +[2026-06-23 22:55:11,410][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.1172667145729065, 'train/grad_norm_pre_clip': 22.972312927246094, 'timing/data': 0.0004033150617033243, 'timing/model': 1.1147866959217936, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 0.62}) +[2026-06-23 22:55:15,480][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.04904796928167343, 'train/grad_norm_pre_clip': 23.18777847290039, 'timing/data': 0.00037306302692741156, 'timing/model': 1.1109501640312374, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 0.62}) +[2026-06-23 22:55:19,541][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.17166019976139069, 'train/grad_norm_pre_clip': 29.91934585571289, 'timing/data': 0.000382463913410902, 'timing/model': 1.116342360037379, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.62}) +[2026-06-23 22:55:23,612][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.09580035507678986, 'train/grad_norm_pre_clip': 12.517457008361816, 'timing/data': 0.00035076297353953123, 'timing/model': 1.1164955599233508, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 0.62}) +[2026-06-23 22:55:27,832][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.04567338526248932, 'train/grad_norm_pre_clip': 14.61083984375, 'timing/data': 0.0003531560068950057, 'timing/model': 1.1131502860225737, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 0.62}) +[2026-06-23 22:55:31,899][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.1385582536458969, 'train/grad_norm_pre_clip': 28.388042449951172, 'timing/data': 0.0003587530227378011, 'timing/model': 1.1182917080586776, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 0.62}) +[2026-06-23 22:55:35,962][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.030730877071619034, 'train/grad_norm_pre_clip': 34.514312744140625, 'timing/data': 0.0004484750097617507, 'timing/model': 1.112423978978768, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 0.62}) +[2026-06-23 22:55:40,020][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.005785889457911253, 'train/grad_norm_pre_clip': 6.846790790557861, 'timing/data': 0.00043597491458058357, 'timing/model': 1.1088083159411326, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 0.63}) +[2026-06-23 22:55:44,169][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.38168156147003174, 'train/grad_norm_pre_clip': 30.129348754882812, 'timing/data': 0.00041283597238361835, 'timing/model': 1.113431895035319, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 0.63}) +[2026-06-23 22:55:48,235][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.10305305570363998, 'train/grad_norm_pre_clip': 20.368120193481445, 'timing/data': 0.000561427092179656, 'timing/model': 1.1122711340431124, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 0.63}) +[2026-06-23 22:55:52,319][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.008523543365299702, 'train/grad_norm_pre_clip': 10.944707870483398, 'timing/data': 0.00035139103420078754, 'timing/model': 1.1179079950088635, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 0.63}) +[2026-06-23 22:55:56,395][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.08614382892847061, 'train/grad_norm_pre_clip': 19.884822845458984, 'timing/data': 0.00036006805021315813, 'timing/model': 1.1173338550142944, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.63}) +[2026-06-23 22:56:00,584][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.01850636675953865, 'train/grad_norm_pre_clip': 8.284770965576172, 'timing/data': 0.0005625459598377347, 'timing/model': 1.2361230689566582, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.63}) +[2026-06-23 22:56:04,654][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.10386856645345688, 'train/grad_norm_pre_clip': 16.60293960571289, 'timing/data': 0.0003978069871664047, 'timing/model': 1.111250524991192, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 0.63}) +[2026-06-23 22:56:08,727][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.01308047492057085, 'train/grad_norm_pre_clip': 14.197405815124512, 'timing/data': 0.00036186794750392437, 'timing/model': 1.1167219260241836, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 0.63}) +[2026-06-23 22:56:12,805][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.07192602008581161, 'train/grad_norm_pre_clip': 21.59737205505371, 'timing/data': 0.0004088260466232896, 'timing/model': 1.1186299490509555, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 0.63}) +[2026-06-23 22:56:16,881][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.10553281009197235, 'train/grad_norm_pre_clip': 28.922780990600586, 'timing/data': 0.00048812199383974075, 'timing/model': 1.1214360280428082, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 0.63}) +[2026-06-23 22:56:21,131][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.03213103488087654, 'train/grad_norm_pre_clip': 12.896690368652344, 'timing/data': 0.00035467196721583605, 'timing/model': 1.1220809650840238, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 0.63}) +[2026-06-23 22:56:25,202][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.019997844472527504, 'train/grad_norm_pre_clip': 24.043264389038086, 'timing/data': 0.00040460494346916676, 'timing/model': 1.1175641890149564, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 0.63}) +[2026-06-23 22:56:29,274][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.03277403488755226, 'train/grad_norm_pre_clip': 18.73143196105957, 'timing/data': 0.0003600739873945713, 'timing/model': 1.1146804100135341, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 0.63}) +[2026-06-23 22:56:33,345][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.07486031949520111, 'train/grad_norm_pre_clip': 21.4331111907959, 'timing/data': 0.00036962295416742563, 'timing/model': 1.1187166549498215, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 0.63}) +[2026-06-23 22:56:37,489][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.03473573923110962, 'train/grad_norm_pre_clip': 27.902080535888672, 'timing/data': 0.000387703999876976, 'timing/model': 1.1190837559988722, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 0.63}) +[2026-06-23 22:56:41,556][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.0043651568703353405, 'train/grad_norm_pre_clip': 40.403202056884766, 'timing/data': 0.0003623529337346554, 'timing/model': 1.1144726279890165, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.63}) +[2026-06-23 22:56:45,639][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.11196613311767578, 'train/grad_norm_pre_clip': 45.86876678466797, 'timing/data': 0.0005787840345874429, 'timing/model': 1.1188687820686027, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 0.63}) +[2026-06-23 22:56:49,722][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.03810717537999153, 'train/grad_norm_pre_clip': 14.613779067993164, 'timing/data': 0.0003590319538488984, 'timing/model': 1.1167274550534785, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 0.63}) +[2026-06-23 22:56:53,876][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.0808386355638504, 'train/grad_norm_pre_clip': 34.77361297607422, 'timing/data': 0.000414686044678092, 'timing/model': 1.119087483966723, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 0.63}) +[2026-06-23 22:56:57,958][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.07398767024278641, 'train/grad_norm_pre_clip': 14.489739418029785, 'timing/data': 0.0005413009785115719, 'timing/model': 1.1218118459219113, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 0.63}) +[2026-06-23 22:57:02,040][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.16346830129623413, 'train/grad_norm_pre_clip': 30.124696731567383, 'timing/data': 0.0003631958970800042, 'timing/model': 1.1249429680174217, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 0.63}) +[2026-06-23 22:57:06,131][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.27477654814720154, 'train/grad_norm_pre_clip': 61.74761962890625, 'timing/data': 0.00036392698530107737, 'timing/model': 1.1333936390001327, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 0.63}) +[2026-06-23 22:57:10,276][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.09050391614437103, 'train/grad_norm_pre_clip': 51.00386047363281, 'timing/data': 0.000493893981911242, 'timing/model': 1.1230495589552447, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 0.63}) +[2026-06-23 22:57:14,358][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.08653823286294937, 'train/grad_norm_pre_clip': 18.419267654418945, 'timing/data': 0.00036949990317225456, 'timing/model': 1.1166629280196503, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 0.63}) +[2026-06-23 22:57:18,436][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.010061620734632015, 'train/grad_norm_pre_clip': 24.645475387573242, 'timing/data': 0.0003769960021600127, 'timing/model': 1.1184000229695812, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 0.63}) +[2026-06-23 22:57:22,531][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.12537486851215363, 'train/grad_norm_pre_clip': 19.962963104248047, 'timing/data': 0.0003609530394896865, 'timing/model': 1.1252278200117871, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.63}) +[2026-06-23 22:57:26,757][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.1615847796201706, 'train/grad_norm_pre_clip': 29.37822914123535, 'timing/data': 0.0004399190656840801, 'timing/model': 1.1202737449202687, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 0.64}) +[2026-06-23 22:57:30,837][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.059213049709796906, 'train/grad_norm_pre_clip': 20.337507247924805, 'timing/data': 0.0003654269967228174, 'timing/model': 1.1151754269376397, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 0.64}) +[2026-06-23 22:57:34,911][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.0680258721113205, 'train/grad_norm_pre_clip': 22.536314010620117, 'timing/data': 0.0004987299907952547, 'timing/model': 1.1182338739745319, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 0.64}) +[2026-06-23 22:57:39,080][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.07228074967861176, 'train/grad_norm_pre_clip': 22.431968688964844, 'timing/data': 0.00042170099914073944, 'timing/model': 1.212802444002591, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 0.64}) +[2026-06-23 22:57:43,184][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.06653174757957458, 'train/grad_norm_pre_clip': 18.981903076171875, 'timing/data': 0.00042002799455076456, 'timing/model': 1.114651134936139, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 0.64}) +[2026-06-23 22:57:47,276][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.04939621686935425, 'train/grad_norm_pre_clip': 10.603422164916992, 'timing/data': 0.00036598602309823036, 'timing/model': 1.1265616659075022, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 0.64}) +[2026-06-23 22:57:51,375][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.1022161915898323, 'train/grad_norm_pre_clip': 35.241249084472656, 'timing/data': 0.0005395180778577924, 'timing/model': 1.1283514449605718, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 0.64}) +[2026-06-23 22:57:55,641][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.16083544492721558, 'train/grad_norm_pre_clip': 21.10431671142578, 'timing/data': 0.0004987079882994294, 'timing/model': 1.2865236399229616, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 0.64}) +[2026-06-23 22:57:59,726][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.13423198461532593, 'train/grad_norm_pre_clip': 20.77638816833496, 'timing/data': 0.00036955298855900764, 'timing/model': 1.12418709800113, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 0.64}) +[2026-06-23 22:58:03,821][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.12151263654232025, 'train/grad_norm_pre_clip': 14.854460716247559, 'timing/data': 0.00035287102218717337, 'timing/model': 1.1190968879964203, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.64}) +[2026-06-23 22:58:07,901][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.13868211209774017, 'train/grad_norm_pre_clip': 14.956506729125977, 'timing/data': 0.0003609310369938612, 'timing/model': 1.11727789696306, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 0.64}) +[2026-06-23 22:58:12,049][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.09346804022789001, 'train/grad_norm_pre_clip': 14.15284538269043, 'timing/data': 0.000437091919593513, 'timing/model': 1.1115321919787675, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 0.64}) +[2026-06-23 22:58:16,134][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.14458996057510376, 'train/grad_norm_pre_clip': 22.814401626586914, 'timing/data': 0.0005577999399974942, 'timing/model': 1.1236131599871442, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 0.64}) +[2026-06-23 22:58:20,211][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.25994232296943665, 'train/grad_norm_pre_clip': 18.873729705810547, 'timing/data': 0.0004867149982601404, 'timing/model': 1.1169103119755164, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 0.64}) +[2026-06-23 22:58:24,319][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.13077786564826965, 'train/grad_norm_pre_clip': 23.363618850708008, 'timing/data': 0.00034843606408685446, 'timing/model': 1.1251891989959404, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 0.64}) +[2026-06-23 22:58:28,474][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.02596534602344036, 'train/grad_norm_pre_clip': 12.525067329406738, 'timing/data': 0.0005767890252172947, 'timing/model': 1.121127498918213, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 0.64}) +[2026-06-23 22:58:32,559][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.13742059469223022, 'train/grad_norm_pre_clip': 25.316070556640625, 'timing/data': 0.0006299460073933005, 'timing/model': 1.1171985779656097, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 0.64}) +[2026-06-23 22:58:36,639][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.1497061848640442, 'train/grad_norm_pre_clip': 22.96340560913086, 'timing/data': 0.0003874580143019557, 'timing/model': 1.1140470440732315, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 0.64}) +[2026-06-23 22:58:40,719][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.02123044617474079, 'train/grad_norm_pre_clip': 20.605411529541016, 'timing/data': 0.0003692810423672199, 'timing/model': 1.1168958559865132, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 0.64}) +[2026-06-23 22:58:45,131][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.06539100408554077, 'train/grad_norm_pre_clip': 14.081669807434082, 'timing/data': 0.0005708610406145453, 'timing/model': 1.1140154720051214, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.64}) +[2026-06-23 22:58:49,207][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.05496488884091377, 'train/grad_norm_pre_clip': 24.00727653503418, 'timing/data': 0.00036721397191286087, 'timing/model': 1.122186240972951, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 0.64}) +[2026-06-23 22:58:53,296][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.05424346774816513, 'train/grad_norm_pre_clip': 7.394989490509033, 'timing/data': 0.00037358992267400026, 'timing/model': 1.1126713369740173, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 0.64}) +[2026-06-23 22:58:57,388][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.016297388821840286, 'train/grad_norm_pre_clip': 10.4334077835083, 'timing/data': 0.000375532079488039, 'timing/model': 1.1157531429780647, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 0.64}) +[2026-06-23 22:59:01,564][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.167608842253685, 'train/grad_norm_pre_clip': 28.704105377197266, 'timing/data': 0.00038712797686457634, 'timing/model': 1.112542079994455, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 0.64}) +[2026-06-23 22:59:05,651][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.16118375957012177, 'train/grad_norm_pre_clip': 24.33211898803711, 'timing/data': 0.00037955096922814846, 'timing/model': 1.1172110300976783, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 0.64}) +[2026-06-23 22:59:09,731][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.10598606616258621, 'train/grad_norm_pre_clip': 24.78411865234375, 'timing/data': 0.00035639700945466757, 'timing/model': 1.117766413022764, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 0.64}) +[2026-06-23 22:59:13,804][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.019878366962075233, 'train/grad_norm_pre_clip': 31.657333374023438, 'timing/data': 0.00036535400431603193, 'timing/model': 1.120849491097033, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 0.65}) +[2026-06-23 22:59:17,957][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.11957602202892303, 'train/grad_norm_pre_clip': 18.279701232910156, 'timing/data': 0.00035939901135861874, 'timing/model': 1.1130064179887995, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.65}) +[2026-06-23 22:59:22,032][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.17273969948291779, 'train/grad_norm_pre_clip': 20.168113708496094, 'timing/data': 0.00036506401374936104, 'timing/model': 1.1167481750017032, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 0.65}) +[2026-06-23 22:59:26,119][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.08982113003730774, 'train/grad_norm_pre_clip': 19.726768493652344, 'timing/data': 0.0003567771054804325, 'timing/model': 1.121156155015342, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.65}) +[2026-06-23 22:59:30,194][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.20320530235767365, 'train/grad_norm_pre_clip': 31.89963722229004, 'timing/data': 0.0003571449778974056, 'timing/model': 1.115302971098572, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 0.65}) +[2026-06-23 22:59:34,435][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.0838882178068161, 'train/grad_norm_pre_clip': 65.7184066772461, 'timing/data': 0.0003605480305850506, 'timing/model': 1.1184650160139427, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 0.65}) +[2026-06-23 22:59:38,517][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.060967572033405304, 'train/grad_norm_pre_clip': 24.26557731628418, 'timing/data': 0.0005743650253862143, 'timing/model': 1.1236527969595045, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 0.65}) +[2026-06-23 22:59:42,596][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.01126090157777071, 'train/grad_norm_pre_clip': 12.710588455200195, 'timing/data': 0.00046928192023187876, 'timing/model': 1.1204183839727193, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 0.65}) +[2026-06-23 22:59:46,677][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.05465788021683693, 'train/grad_norm_pre_clip': 21.154932022094727, 'timing/data': 0.00035640690475702286, 'timing/model': 1.119638102944009, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.65}) +[2026-06-23 22:59:50,857][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.041034433990716934, 'train/grad_norm_pre_clip': 30.141935348510742, 'timing/data': 0.00036891200579702854, 'timing/model': 1.116961985011585, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 0.65}) +[2026-06-23 22:59:55,050][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.10089774429798126, 'train/grad_norm_pre_clip': 27.534082412719727, 'timing/data': 0.00044981599785387516, 'timing/model': 1.1187807880342007, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 0.65}) +[2026-06-23 22:59:59,128][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.061562519520521164, 'train/grad_norm_pre_clip': 36.19068145751953, 'timing/data': 0.0005095689557492733, 'timing/model': 1.1175430270377547, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 0.65}) +[2026-06-23 23:00:03,207][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.1363753229379654, 'train/grad_norm_pre_clip': 38.49713134765625, 'timing/data': 0.000357098993845284, 'timing/model': 1.1182066060137004, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 0.65}) +[2026-06-23 23:00:07,460][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.06327089667320251, 'train/grad_norm_pre_clip': 44.66177749633789, 'timing/data': 0.00035632296930998564, 'timing/model': 1.1203525019809604, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.65}) +[2026-06-23 23:00:11,549][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.508140504360199, 'train/grad_norm_pre_clip': 44.82054901123047, 'timing/data': 0.000355868018232286, 'timing/model': 1.1157964279409498, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 0.65}) +[2026-06-23 23:00:15,632][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.06471468508243561, 'train/grad_norm_pre_clip': 36.016231536865234, 'timing/data': 0.0005394089967012405, 'timing/model': 1.1210169489495456, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 0.65}) +[2026-06-23 23:00:19,716][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.050943899899721146, 'train/grad_norm_pre_clip': 16.11017417907715, 'timing/data': 0.0004570190794765949, 'timing/model': 1.118714916985482, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 0.65}) +[2026-06-23 23:00:23,869][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.15813104808330536, 'train/grad_norm_pre_clip': 21.071857452392578, 'timing/data': 0.00035562505945563316, 'timing/model': 1.1203544089803472, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 0.65}) +[2026-06-23 23:00:27,944][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.02233397401869297, 'train/grad_norm_pre_clip': 27.290849685668945, 'timing/data': 0.0004404800711199641, 'timing/model': 1.1129353169817477, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 0.65}) +[2026-06-23 23:00:32,023][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.05373372510075569, 'train/grad_norm_pre_clip': 29.572608947753906, 'timing/data': 0.0003785319859161973, 'timing/model': 1.111473684082739, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 0.65}) +[2026-06-23 23:00:36,112][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.07063683867454529, 'train/grad_norm_pre_clip': 41.558895111083984, 'timing/data': 0.00036122091114521027, 'timing/model': 1.1199682959122583, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 0.65}) +[2026-06-23 23:00:40,307][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.03311936557292938, 'train/grad_norm_pre_clip': 21.13382339477539, 'timing/data': 0.00035917398054152727, 'timing/model': 1.1222767169820145, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 0.65}) +[2026-06-23 23:00:44,415][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.04664527252316475, 'train/grad_norm_pre_clip': 15.916322708129883, 'timing/data': 0.0003842900041490793, 'timing/model': 1.1199215990491211, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 0.65}) +[2026-06-23 23:00:48,504][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.032216478139162064, 'train/grad_norm_pre_clip': 13.71793270111084, 'timing/data': 0.0003625549143180251, 'timing/model': 1.1257854000432417, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.65}) +[2026-06-23 23:00:52,596][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.19498322904109955, 'train/grad_norm_pre_clip': 20.71706771850586, 'timing/data': 0.00035481806844472885, 'timing/model': 1.1261610330548137, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 0.65}) +[2026-06-23 23:00:56,825][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.19741591811180115, 'train/grad_norm_pre_clip': 13.706948280334473, 'timing/data': 0.0003585440572351217, 'timing/model': 1.1184830920537934, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 0.65}) +[2026-06-23 23:01:00,913][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.08049408346414566, 'train/grad_norm_pre_clip': 12.866165161132812, 'timing/data': 0.0003508409718051553, 'timing/model': 1.1211013830034062, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 0.66}) +[2026-06-23 23:01:04,998][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.10476992279291153, 'train/grad_norm_pre_clip': 19.101930618286133, 'timing/data': 0.0003768220776692033, 'timing/model': 1.1187076029600576, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 0.66}) +[2026-06-23 23:01:09,171][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.18600136041641235, 'train/grad_norm_pre_clip': 15.471941947937012, 'timing/data': 0.000462919007986784, 'timing/model': 1.2126724200788885, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 0.66}) +[2026-06-23 23:01:13,256][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.01755860447883606, 'train/grad_norm_pre_clip': 10.169889450073242, 'timing/data': 0.000353478011675179, 'timing/model': 1.1199018160114065, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 0.66}) +[2026-06-23 23:01:17,346][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.07107318192720413, 'train/grad_norm_pre_clip': 10.522502899169922, 'timing/data': 0.00036209297832101583, 'timing/model': 1.1161426129983738, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 0.66}) +[2026-06-23 23:01:21,426][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.054683931171894073, 'train/grad_norm_pre_clip': 11.356220245361328, 'timing/data': 0.0003719009691849351, 'timing/model': 1.1211032059509307, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 0.66}) +[2026-06-23 23:01:25,677][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.004996204283088446, 'train/grad_norm_pre_clip': 10.946039199829102, 'timing/data': 0.0003517469158396125, 'timing/model': 1.2876783430110663, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 0.66}) +[2026-06-23 23:01:29,761][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1700 +[2026-06-23 23:02:35,270][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1700 +[2026-06-23 23:02:35,271][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1700 +[2026-06-23 23:03:13,537][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1700 +[2026-06-23 23:03:13,538][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.1309143751859665, 'train/grad_norm_pre_clip': 15.07371711730957, 'eval/action_loss/seconds': 65.50738838792313, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.04274208843708038, 'eval/latency_2/loss': 0.04274208843708038, 'eval/flappy/loss': 0.04274208843708038, 'eval/flappy/latency_2/loss': 0.04274208843708038, 'eval/flap_precision': 0.9534574468085106, 'eval/flap_recall': 0.9728629579375848, 'eval/flap_f1': 0.9630624580255206, 'eval/flap_rate_model': 0.2088888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9630624580255205, 'eval/latency_2/flap_precision': 0.9534574468085106, 'eval/latency_2/flap_recall': 0.9728629579375848, 'eval/latency_2/flap_f1': 0.9630624580255206, 'eval/latency_2/flap_rate_model': 0.2088888888888889, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9630624580255205, 'eval/action_classification/data_wait_seconds': 0.6466353535652161, 'eval/action_classification/predict_seconds': 37.453243255615234, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0003554760478436947, 'timing/model': 1.1239470910513774, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.66}) +[2026-06-23 23:03:17,590][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.04075793921947479, 'train/grad_norm_pre_clip': 10.257584571838379, 'timing/data': 0.0003675080370157957, 'timing/model': 1.1110516450135037, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 0.66}) +[2026-06-23 23:03:21,625][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.30483123660087585, 'train/grad_norm_pre_clip': 26.51255226135254, 'timing/data': 0.0003715059719979763, 'timing/model': 1.1040593769866973, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 0.66}) +[2026-06-23 23:03:25,825][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.08920905739068985, 'train/grad_norm_pre_clip': 33.68415832519531, 'timing/data': 0.0005628840299323201, 'timing/model': 1.1030808289069682, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 0.66}) +[2026-06-23 23:03:29,876][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.06319157779216766, 'train/grad_norm_pre_clip': 11.440438270568848, 'timing/data': 0.00036201102193444967, 'timing/model': 1.1094300240511075, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 0.66}) +[2026-06-23 23:03:33,928][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.015001729130744934, 'train/grad_norm_pre_clip': 16.206436157226562, 'timing/data': 0.0005736590828746557, 'timing/model': 1.107118007959798, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 0.66}) +[2026-06-23 23:03:37,980][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.17291687428951263, 'train/grad_norm_pre_clip': 26.994821548461914, 'timing/data': 0.0003594110021367669, 'timing/model': 1.1085642910329625, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 0.66}) +[2026-06-23 23:03:42,109][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.00043948268285021186, 'train/grad_norm_pre_clip': 15.064749717712402, 'timing/data': 0.0003985520452260971, 'timing/model': 1.112152150948532, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 0.66}) +[2026-06-23 23:03:46,198][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.07516075670719147, 'train/grad_norm_pre_clip': 17.456106185913086, 'timing/data': 0.0006554400315508246, 'timing/model': 1.1175923279952258, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 0.66}) +[2026-06-23 23:03:50,259][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.027226647362113, 'train/grad_norm_pre_clip': 31.456466674804688, 'timing/data': 0.0005833549657836556, 'timing/model': 1.1112815460655838, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 0.66}) +[2026-06-23 23:03:54,311][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.007596943061798811, 'train/grad_norm_pre_clip': 16.536277770996094, 'timing/data': 0.0003685819683596492, 'timing/model': 1.1071238110307604, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.66}) +[2026-06-23 23:03:58,441][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.0030176471918821335, 'train/grad_norm_pre_clip': 5.446876525878906, 'timing/data': 0.0005543039878830314, 'timing/model': 1.1118663299130276, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 0.66}) +[2026-06-23 23:04:02,511][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.06805988401174545, 'train/grad_norm_pre_clip': 14.47889518737793, 'timing/data': 0.00039085804019123316, 'timing/model': 1.11468570609577, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 0.66}) +[2026-06-23 23:04:06,582][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.06410349905490875, 'train/grad_norm_pre_clip': 16.013526916503906, 'timing/data': 0.0005947829922661185, 'timing/model': 1.1219611950218678, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 0.66}) +[2026-06-23 23:04:10,653][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.10745259374380112, 'train/grad_norm_pre_clip': 23.360008239746094, 'timing/data': 0.0005040980177000165, 'timing/model': 1.115958530921489, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 0.66}) +[2026-06-23 23:04:14,793][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.06152806058526039, 'train/grad_norm_pre_clip': 13.48486042022705, 'timing/data': 0.00043772789649665356, 'timing/model': 1.1067252260399982, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 0.66}) +[2026-06-23 23:04:18,853][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.012938331812620163, 'train/grad_norm_pre_clip': 14.544946670532227, 'timing/data': 0.0003588119288906455, 'timing/model': 1.1183856230927631, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 0.66}) +[2026-06-23 23:04:22,930][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.045602139085531235, 'train/grad_norm_pre_clip': 25.41266441345215, 'timing/data': 0.0003662810195237398, 'timing/model': 1.1195295880315825, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.66}) +[2026-06-23 23:04:26,999][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.08630481362342834, 'train/grad_norm_pre_clip': 13.894867897033691, 'timing/data': 0.0005634500412270427, 'timing/model': 1.112589744036086, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 0.66}) +[2026-06-23 23:04:31,207][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.09546168148517609, 'train/grad_norm_pre_clip': 17.22764778137207, 'timing/data': 0.0005497699603438377, 'timing/model': 1.1184900010703132, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 0.67}) +[2026-06-23 23:04:35,282][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.21150702238082886, 'train/grad_norm_pre_clip': 30.62303924560547, 'timing/data': 0.0004945909604430199, 'timing/model': 1.1141931490274146, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.67}) +[2026-06-23 23:04:39,349][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.00930874701589346, 'train/grad_norm_pre_clip': 7.2632293701171875, 'timing/data': 0.00036711397115141153, 'timing/model': 1.1118948329240084, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 0.67}) +[2026-06-23 23:04:43,514][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.10963377356529236, 'train/grad_norm_pre_clip': 16.857149124145508, 'timing/data': 0.00034718005917966366, 'timing/model': 1.2121742790332064, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 0.67}) +[2026-06-23 23:04:47,585][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.27394846081733704, 'train/grad_norm_pre_clip': 12.567781448364258, 'timing/data': 0.0003633709857240319, 'timing/model': 1.1155580600025132, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 0.67}) +[2026-06-23 23:04:51,661][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.004250594414770603, 'train/grad_norm_pre_clip': 18.67324447631836, 'timing/data': 0.0003544719656929374, 'timing/model': 1.1161614549346268, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 0.67}) +[2026-06-23 23:04:55,737][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.08285428583621979, 'train/grad_norm_pre_clip': 20.798084259033203, 'timing/data': 0.00035233201924711466, 'timing/model': 1.1157308120746166, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 0.67}) +[2026-06-23 23:04:59,994][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.1603137105703354, 'train/grad_norm_pre_clip': 23.7745304107666, 'timing/data': 0.0004140299279242754, 'timing/model': 1.286358368000947, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 0.67}) +[2026-06-23 23:05:04,064][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.024662571027874947, 'train/grad_norm_pre_clip': 19.400732040405273, 'timing/data': 0.0003574410220608115, 'timing/model': 1.11722502799239, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 0.67}) +[2026-06-23 23:05:08,141][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.0607491135597229, 'train/grad_norm_pre_clip': 17.460676193237305, 'timing/data': 0.0003679649671539664, 'timing/model': 1.117561364080757, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 0.67}) +[2026-06-23 23:05:12,219][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.2025616466999054, 'train/grad_norm_pre_clip': 28.394960403442383, 'timing/data': 0.00036306993570178747, 'timing/model': 1.1200422879774123, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 0.67}) +[2026-06-23 23:05:16,376][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.05146666243672371, 'train/grad_norm_pre_clip': 8.762083053588867, 'timing/data': 0.00036495306994765997, 'timing/model': 1.1193069679429755, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.67}) +[2026-06-23 23:05:20,453][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.03477821499109268, 'train/grad_norm_pre_clip': 17.554222106933594, 'timing/data': 0.0003819590201601386, 'timing/model': 1.1150095670018345, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 0.67}) +[2026-06-23 23:05:24,527][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.12855875492095947, 'train/grad_norm_pre_clip': 12.549132347106934, 'timing/data': 0.0004080230137333274, 'timing/model': 1.1155145580414683, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 0.67}) +[2026-06-23 23:05:28,609][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.17035774886608124, 'train/grad_norm_pre_clip': 20.881616592407227, 'timing/data': 0.0005584059981629252, 'timing/model': 1.1249855939531699, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 0.67}) +[2026-06-23 23:05:32,757][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.02294168435037136, 'train/grad_norm_pre_clip': 11.21800708770752, 'timing/data': 0.0003571189008653164, 'timing/model': 1.1213148779934272, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 0.67}) +[2026-06-23 23:05:36,840][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.031594015657901764, 'train/grad_norm_pre_clip': 20.739776611328125, 'timing/data': 0.00037017103750258684, 'timing/model': 1.1203923320863396, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 0.67}) +[2026-06-23 23:05:40,917][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.07909021526575089, 'train/grad_norm_pre_clip': 20.963499069213867, 'timing/data': 0.0003739059902727604, 'timing/model': 1.1184820280177519, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 0.67}) +[2026-06-23 23:05:44,998][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.033483728766441345, 'train/grad_norm_pre_clip': 6.493364334106445, 'timing/data': 0.00036894704680889845, 'timing/model': 1.1177791119553149, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 0.67}) +[2026-06-23 23:05:49,223][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.03192603960633278, 'train/grad_norm_pre_clip': 13.855649948120117, 'timing/data': 0.00039432698395103216, 'timing/model': 1.1257821000181139, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 0.67}) +[2026-06-23 23:05:53,293][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.04203314334154129, 'train/grad_norm_pre_clip': 23.514076232910156, 'timing/data': 0.00039644294884055853, 'timing/model': 1.121281735948287, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 0.67}) +[2026-06-23 23:05:57,367][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.07331299036741257, 'train/grad_norm_pre_clip': 32.978912353515625, 'timing/data': 0.00036443397402763367, 'timing/model': 1.1163953499635682, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.67}) +[2026-06-23 23:06:01,477][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.1753728687763214, 'train/grad_norm_pre_clip': 24.420454025268555, 'timing/data': 0.0006411919603124261, 'timing/model': 1.1197738199261948, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 0.67}) +[2026-06-23 23:06:05,663][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.017466822639107704, 'train/grad_norm_pre_clip': 8.791112899780273, 'timing/data': 0.0005964600713923573, 'timing/model': 1.1201126630185172, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 0.67}) +[2026-06-23 23:06:09,748][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.11908086389303207, 'train/grad_norm_pre_clip': 17.445276260375977, 'timing/data': 0.0003784680739045143, 'timing/model': 1.118110770941712, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 0.67}) +[2026-06-23 23:06:13,855][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.031711310148239136, 'train/grad_norm_pre_clip': 12.263083457946777, 'timing/data': 0.0003898750292137265, 'timing/model': 1.1298576759872958, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 0.67}) +[2026-06-23 23:06:17,945][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.13129974901676178, 'train/grad_norm_pre_clip': 57.208499908447266, 'timing/data': 0.0003798360703513026, 'timing/model': 1.1205690380884334, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.68}) +[2026-06-23 23:06:22,103][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.010800113901495934, 'train/grad_norm_pre_clip': 32.93397903442383, 'timing/data': 0.0004189320607110858, 'timing/model': 1.1167845559539273, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 0.68}) +[2026-06-23 23:06:26,189][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.03086894564330578, 'train/grad_norm_pre_clip': 51.85578918457031, 'timing/data': 0.00037646200507879257, 'timing/model': 1.1183313799556345, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 0.68}) +[2026-06-23 23:06:30,278][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.011223402805626392, 'train/grad_norm_pre_clip': 34.867919921875, 'timing/data': 0.00036137993447482586, 'timing/model': 1.1254051740979776, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 0.68}) +[2026-06-23 23:06:34,355][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.01892496459186077, 'train/grad_norm_pre_clip': 13.407048225402832, 'timing/data': 0.00035396998282521963, 'timing/model': 1.1193791970144957, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 0.68}) +[2026-06-23 23:17:10,442][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1750 +[2026-06-23 23:17:10,443][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.22100168466567993, 'train/grad_norm_pre_clip': 28.660320281982422, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 390.9600000000196, 'rl_games_eval/mid_train/mean_length': 3168.4, 'rl_games_eval/mid_train/std_reward': 107.2800000000121, 'rl_games_eval/mid_train/std_length': 863.2, 'rl_games_eval/mid_train/macro_mean_reward': 390.9600000000196, 'rl_games_eval/mid_train/macro_mean_length': 3168.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 390.9600000000196, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3168.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 107.2800000000121, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 863.2, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 390.9600000000196, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00035369605757296085, 'timing/model': 1.122046433039941, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.68}) +[2026-06-23 23:17:14,418][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.09272979199886322, 'train/grad_norm_pre_clip': 16.999252319335938, 'timing/data': 0.0003693470498546958, 'timing/model': 1.092329730046913, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 0.68}) +[2026-06-23 23:17:18,421][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.004912045784294605, 'train/grad_norm_pre_clip': 28.89531135559082, 'timing/data': 0.00036635110154747963, 'timing/model': 1.0986440069973469, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 0.68}) +[2026-06-23 23:17:22,421][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.141437366604805, 'train/grad_norm_pre_clip': 26.953636169433594, 'timing/data': 0.0003542960621416569, 'timing/model': 1.0968459030846134, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 0.68}) +[2026-06-23 23:17:26,499][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.13233667612075806, 'train/grad_norm_pre_clip': 16.36835479736328, 'timing/data': 0.0003593979636207223, 'timing/model': 1.1012028170516714, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 0.68}) +[2026-06-23 23:17:30,504][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.05600462853908539, 'train/grad_norm_pre_clip': 9.430804252624512, 'timing/data': 0.0006511589745059609, 'timing/model': 1.0941405489575118, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 0.68}) +[2026-06-23 23:17:34,520][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.07287687808275223, 'train/grad_norm_pre_clip': 48.11496353149414, 'timing/data': 0.00038468395359814167, 'timing/model': 1.103956816950813, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 0.68}) +[2026-06-23 23:17:38,547][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.11348003894090652, 'train/grad_norm_pre_clip': 11.727144241333008, 'timing/data': 0.0003582669887691736, 'timing/model': 1.1089276060229167, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 0.68}) +[2026-06-23 23:17:42,747][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.06429049372673035, 'train/grad_norm_pre_clip': 30.336868286132812, 'timing/data': 0.0003583841025829315, 'timing/model': 1.1094753500074148, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 0.68}) +[2026-06-23 23:17:46,791][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.025229910388588905, 'train/grad_norm_pre_clip': 15.609892845153809, 'timing/data': 0.00044740899465978146, 'timing/model': 1.1043296799762174, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 0.68}) +[2026-06-23 23:17:50,822][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.13885942101478577, 'train/grad_norm_pre_clip': 31.492610931396484, 'timing/data': 0.0003736600046977401, 'timing/model': 1.1126239839941263, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.68}) +[2026-06-23 23:17:54,886][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.1004737913608551, 'train/grad_norm_pre_clip': 14.921635627746582, 'timing/data': 0.00037030107341706753, 'timing/model': 1.1199306939961389, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 0.68}) +[2026-06-23 23:17:59,025][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.12389742583036423, 'train/grad_norm_pre_clip': 16.295909881591797, 'timing/data': 0.0003739030798897147, 'timing/model': 1.1040860200300813, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 0.68}) +[2026-06-23 23:18:03,071][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.04416489973664284, 'train/grad_norm_pre_clip': 12.746602058410645, 'timing/data': 0.0005637370049953461, 'timing/model': 1.1074697260046378, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 0.68}) +[2026-06-23 23:18:07,116][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.21700698137283325, 'train/grad_norm_pre_clip': 16.126184463500977, 'timing/data': 0.0004204559372738004, 'timing/model': 1.1142969289794564, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 0.68}) +[2026-06-23 23:18:11,174][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.08121449500322342, 'train/grad_norm_pre_clip': 19.50775909423828, 'timing/data': 0.0003765519941225648, 'timing/model': 1.1191217390587553, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 0.68}) +[2026-06-23 23:18:15,388][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.03472275286912918, 'train/grad_norm_pre_clip': 8.241013526916504, 'timing/data': 0.00035884801764041185, 'timing/model': 1.1097276409855112, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.68}) +[2026-06-23 23:18:19,449][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.046377163380384445, 'train/grad_norm_pre_clip': 9.549508094787598, 'timing/data': 0.00041211198549717665, 'timing/model': 1.1149405480828136, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 0.68}) +[2026-06-23 23:18:23,514][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.03713328763842583, 'train/grad_norm_pre_clip': 8.76929759979248, 'timing/data': 0.00035844394005835056, 'timing/model': 1.1141747230431065, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 0.68}) +[2026-06-23 23:18:27,564][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.18140767514705658, 'train/grad_norm_pre_clip': 32.10642623901367, 'timing/data': 0.0003855800023302436, 'timing/model': 1.112809679005295, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 0.68}) +[2026-06-23 23:18:31,689][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.16859443485736847, 'train/grad_norm_pre_clip': 45.368499755859375, 'timing/data': 0.0003674500621855259, 'timing/model': 1.1106288339942694, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.69}) +[2026-06-23 23:18:35,741][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.2550381124019623, 'train/grad_norm_pre_clip': 19.303787231445312, 'timing/data': 0.00036333303432911634, 'timing/model': 1.1155145489610732, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 0.69}) +[2026-06-23 23:18:39,801][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.08400288224220276, 'train/grad_norm_pre_clip': 23.659521102905273, 'timing/data': 0.0003628690028563142, 'timing/model': 1.1138506800634786, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 0.69}) +[2026-06-23 23:18:43,874][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.13560451567173004, 'train/grad_norm_pre_clip': 23.16943359375, 'timing/data': 0.00041931401938199997, 'timing/model': 1.1173387169837952, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 0.69}) +[2026-06-23 23:18:48,025][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.016444940119981766, 'train/grad_norm_pre_clip': 14.909693717956543, 'timing/data': 0.00047008797992020845, 'timing/model': 1.1166397989727557, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 0.69}) +[2026-06-23 23:18:52,091][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.019577423110604286, 'train/grad_norm_pre_clip': 12.725896835327148, 'timing/data': 0.00036431802436709404, 'timing/model': 1.1169422200182453, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 0.69}) +[2026-06-23 23:18:56,163][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.06756173819303513, 'train/grad_norm_pre_clip': 25.69260597229004, 'timing/data': 0.0003619259223341942, 'timing/model': 1.118271206971258, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 0.69}) +[2026-06-23 23:19:00,227][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.21257098019123077, 'train/grad_norm_pre_clip': 41.05086898803711, 'timing/data': 0.00035149697214365005, 'timing/model': 1.110522369039245, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 0.69}) +[2026-06-23 23:19:04,440][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.024525320157408714, 'train/grad_norm_pre_clip': 16.41227912902832, 'timing/data': 0.0005455900682136416, 'timing/model': 1.1123572250362486, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 0.69}) +[2026-06-23 23:19:08,518][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.09552601724863052, 'train/grad_norm_pre_clip': 25.927438735961914, 'timing/data': 0.0005890839966014028, 'timing/model': 1.1221227480564266, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 0.69}) +[2026-06-23 23:19:12,582][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.06822926551103592, 'train/grad_norm_pre_clip': 15.39463996887207, 'timing/data': 0.0003897689748555422, 'timing/model': 1.1149473920231685, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.69}) +[2026-06-23 23:19:16,732][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.0013435494620352983, 'train/grad_norm_pre_clip': 11.926506042480469, 'timing/data': 0.00036546005867421627, 'timing/model': 1.20014467893634, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 0.69}) +[2026-06-23 23:19:20,790][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.018717408180236816, 'train/grad_norm_pre_clip': 20.363500595092773, 'timing/data': 0.0003606990212574601, 'timing/model': 1.1129980209516361, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 0.69}) +[2026-06-23 23:19:24,869][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.10483884811401367, 'train/grad_norm_pre_clip': 20.06212043762207, 'timing/data': 0.00036064605228602886, 'timing/model': 1.132571559981443, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 0.69}) +[2026-06-23 23:19:28,936][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.1025705561041832, 'train/grad_norm_pre_clip': 21.960176467895508, 'timing/data': 0.00036896904930472374, 'timing/model': 1.1168392130639404, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 0.69}) +[2026-06-23 23:19:33,177][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.06305635720491409, 'train/grad_norm_pre_clip': 31.369596481323242, 'timing/data': 0.0006045519839972258, 'timing/model': 1.286573110963218, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 0.69}) +[2026-06-23 23:19:37,251][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.12025659531354904, 'train/grad_norm_pre_clip': 22.35036849975586, 'timing/data': 0.000356161966919899, 'timing/model': 1.1146488429512829, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 0.69}) +[2026-06-23 23:19:41,332][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.04941563308238983, 'train/grad_norm_pre_clip': 13.09990119934082, 'timing/data': 0.0003615219611674547, 'timing/model': 1.1224719299934804, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 0.69}) +[2026-06-23 23:19:45,403][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.14972835779190063, 'train/grad_norm_pre_clip': 19.741756439208984, 'timing/data': 0.000468116020783782, 'timing/model': 1.1134147390257567, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 0.69}) +[2026-06-23 23:19:49,576][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.012581352144479752, 'train/grad_norm_pre_clip': 13.83105182647705, 'timing/data': 0.000372943002730608, 'timing/model': 1.1158703760011122, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 0.69}) +[2026-06-23 23:19:53,652][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.004371309652924538, 'train/grad_norm_pre_clip': 22.51964569091797, 'timing/data': 0.0003856249386444688, 'timing/model': 1.1132772880373523, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.69}) +[2026-06-23 23:19:57,711][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.09778344631195068, 'train/grad_norm_pre_clip': 24.927200317382812, 'timing/data': 0.0003703580005094409, 'timing/model': 1.111087279045023, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 0.69}) +[2026-06-23 23:20:01,791][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.07003170251846313, 'train/grad_norm_pre_clip': 18.920976638793945, 'timing/data': 0.00036221102345734835, 'timing/model': 1.1158813260262832, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 0.69}) +[2026-06-23 23:20:05,951][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.0037463479675352573, 'train/grad_norm_pre_clip': 16.374135971069336, 'timing/data': 0.0003622419899329543, 'timing/model': 1.114430304034613, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 0.69}) +[2026-06-23 23:20:10,013][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.0396587997674942, 'train/grad_norm_pre_clip': 16.86939811706543, 'timing/data': 0.00036379904486238956, 'timing/model': 1.1087649510009214, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 0.69}) +[2026-06-23 23:20:14,081][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.3053492605686188, 'train/grad_norm_pre_clip': 47.576663970947266, 'timing/data': 0.0003624830860644579, 'timing/model': 1.1123587749898434, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 0.69}) +[2026-06-23 23:20:18,165][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.03559493646025658, 'train/grad_norm_pre_clip': 19.032485961914062, 'timing/data': 0.000529056997038424, 'timing/model': 1.115002844016999, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 0.7}) +[2026-06-23 23:20:22,372][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.12293276190757751, 'train/grad_norm_pre_clip': 42.11884689331055, 'timing/data': 0.000545207061804831, 'timing/model': 1.116367395967245, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 0.7}) +[2026-06-23 23:20:26,443][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.0753655806183815, 'train/grad_norm_pre_clip': 12.837979316711426, 'timing/data': 0.000594818964600563, 'timing/model': 1.116282168077305, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 0.7}) +[2026-06-23 23:20:30,523][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.12035023421049118, 'train/grad_norm_pre_clip': 20.35437774658203, 'timing/data': 0.0003667469136416912, 'timing/model': 1.1188847350422293, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 0.7}) +[2026-06-23 23:20:34,610][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1800 +[2026-06-23 23:21:39,048][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1800 +[2026-06-23 23:21:39,050][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1800 +[2026-06-23 23:22:18,374][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1800 +[2026-06-23 23:22:18,374][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.017391055822372437, 'train/grad_norm_pre_clip': 18.987796783447266, 'eval/action_loss/seconds': 64.43543332698755, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.04430323466658592, 'eval/latency_2/loss': 0.04430323466658592, 'eval/flappy/loss': 0.04430323466658592, 'eval/flappy/latency_2/loss': 0.04430323466658592, 'eval/flap_precision': 0.9750346740638003, 'eval/flap_recall': 0.9538670284938942, 'eval/flap_f1': 0.9643347050754458, 'eval/flap_rate_model': 0.20027777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9643347050754458, 'eval/latency_2/flap_precision': 0.9750346740638003, 'eval/latency_2/flap_recall': 0.9538670284938942, 'eval/latency_2/flap_f1': 0.9643347050754458, 'eval/latency_2/flap_rate_model': 0.20027777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9643347050754458, 'eval/action_classification/data_wait_seconds': 0.600601851940155, 'eval/action_classification/predict_seconds': 38.52553939819336, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0003513010451570153, 'timing/model': 1.1179416759405285, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.7}) +[2026-06-23 23:22:22,414][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.07898640632629395, 'train/grad_norm_pre_clip': 14.848435401916504, 'timing/data': 0.0004213779466226697, 'timing/model': 1.1054136840393767, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 0.7}) +[2026-06-23 23:22:26,460][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.1865783929824829, 'train/grad_norm_pre_clip': 19.112024307250977, 'timing/data': 0.0005422120448201895, 'timing/model': 1.111572019988671, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 0.7}) +[2026-06-23 23:22:30,507][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.16029763221740723, 'train/grad_norm_pre_clip': 18.03660011291504, 'timing/data': 0.0003918509464710951, 'timing/model': 1.105758107965812, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 0.7}) +[2026-06-23 23:22:34,659][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.11030692607164383, 'train/grad_norm_pre_clip': 11.246541023254395, 'timing/data': 0.000360771082341671, 'timing/model': 1.1084587840596214, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 0.7}) +[2026-06-23 23:22:38,701][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.09429069608449936, 'train/grad_norm_pre_clip': 15.51087474822998, 'timing/data': 0.00038109603337943554, 'timing/model': 1.1076696299714968, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 0.7}) +[2026-06-23 23:22:42,751][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.059832796454429626, 'train/grad_norm_pre_clip': 11.8660249710083, 'timing/data': 0.000402132049202919, 'timing/model': 1.1116646699374542, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 0.7}) +[2026-06-23 23:22:46,802][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.1720145046710968, 'train/grad_norm_pre_clip': 10.595757484436035, 'timing/data': 0.0003583169309422374, 'timing/model': 1.1120200719451532, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 0.7}) +[2026-06-23 23:22:50,973][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.07787173986434937, 'train/grad_norm_pre_clip': 9.402292251586914, 'timing/data': 0.0003681649686768651, 'timing/model': 1.1133760679513216, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 0.7}) +[2026-06-23 23:22:55,036][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.10834033787250519, 'train/grad_norm_pre_clip': 13.256784439086914, 'timing/data': 0.0006507739890366793, 'timing/model': 1.1178483130643144, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 0.7}) +[2026-06-23 23:22:59,114][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.02915201149880886, 'train/grad_norm_pre_clip': 13.266822814941406, 'timing/data': 0.0004424790386110544, 'timing/model': 1.1115497540449724, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.7}) +[2026-06-23 23:23:03,170][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.08599641174077988, 'train/grad_norm_pre_clip': 14.702197074890137, 'timing/data': 0.00045277096796780825, 'timing/model': 1.1136558370199054, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 0.7}) +[2026-06-23 23:23:07,334][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.05535091832280159, 'train/grad_norm_pre_clip': 11.53857421875, 'timing/data': 0.00035576499067246914, 'timing/model': 1.1074647839413956, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 0.7}) +[2026-06-23 23:23:11,388][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.16115395724773407, 'train/grad_norm_pre_clip': 19.839448928833008, 'timing/data': 0.00034947902895510197, 'timing/model': 1.115203929017298, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 0.7}) +[2026-06-23 23:23:15,443][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.05000467970967293, 'train/grad_norm_pre_clip': 12.346702575683594, 'timing/data': 0.0004862429341301322, 'timing/model': 1.1169386829715222, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 0.7}) +[2026-06-23 23:23:19,513][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.07053213566541672, 'train/grad_norm_pre_clip': 13.221456527709961, 'timing/data': 0.0004955809563398361, 'timing/model': 1.1137964370427653, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.7}) +[2026-06-23 23:23:23,680][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.09616810083389282, 'train/grad_norm_pre_clip': 14.046079635620117, 'timing/data': 0.0004117699572816491, 'timing/model': 1.1136577680008486, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 0.7}) +[2026-06-23 23:23:27,749][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.01625397987663746, 'train/grad_norm_pre_clip': 13.433459281921387, 'timing/data': 0.0003529589157551527, 'timing/model': 1.1184214099776, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 0.7}) +[2026-06-23 23:23:31,820][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.1243845522403717, 'train/grad_norm_pre_clip': 38.967777252197266, 'timing/data': 0.00037459598388522863, 'timing/model': 1.115733788930811, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 0.7}) +[2026-06-23 23:23:35,890][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.11493860185146332, 'train/grad_norm_pre_clip': 34.6091423034668, 'timing/data': 0.00035658804699778557, 'timing/model': 1.1205072859302163, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 0.7}) +[2026-06-23 23:23:40,084][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.4086849093437195, 'train/grad_norm_pre_clip': 34.07448959350586, 'timing/data': 0.0004030220443382859, 'timing/model': 1.1189916039584205, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.7}) +[2026-06-23 23:23:44,159][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.06516215950250626, 'train/grad_norm_pre_clip': 18.544593811035156, 'timing/data': 0.00040776701644062996, 'timing/model': 1.1155624239472672, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 0.7}) +[2026-06-23 23:23:48,243][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.015151020139455795, 'train/grad_norm_pre_clip': 11.644096374511719, 'timing/data': 0.00035506696440279484, 'timing/model': 1.1201895950362086, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 0.71}) +[2026-06-23 23:23:52,323][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.025500375777482986, 'train/grad_norm_pre_clip': 28.839527130126953, 'timing/data': 0.0003576120361685753, 'timing/model': 1.113925356999971, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 0.71}) +[2026-06-23 23:23:56,521][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.1003669798374176, 'train/grad_norm_pre_clip': 33.70091247558594, 'timing/data': 0.0003591240383684635, 'timing/model': 1.1331201870925725, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 0.71}) +[2026-06-23 23:24:00,588][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.08270175755023956, 'train/grad_norm_pre_clip': 49.587738037109375, 'timing/data': 0.000556898070499301, 'timing/model': 1.1162275939714164, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 0.71}) +[2026-06-23 23:24:04,666][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.008478525094687939, 'train/grad_norm_pre_clip': 16.44565773010254, 'timing/data': 0.0003672749735414982, 'timing/model': 1.1212062649428844, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 0.71}) +[2026-06-23 23:24:08,743][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.05119222030043602, 'train/grad_norm_pre_clip': 11.513583183288574, 'timing/data': 0.0006085080094635487, 'timing/model': 1.1169275069842115, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 0.71}) +[2026-06-23 23:24:12,929][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.05699086934328079, 'train/grad_norm_pre_clip': 16.97110939025879, 'timing/data': 0.0004597919760271907, 'timing/model': 1.1168686819728464, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 0.71}) +[2026-06-23 23:24:17,001][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.10555616766214371, 'train/grad_norm_pre_clip': 27.191783905029297, 'timing/data': 0.0004272849764674902, 'timing/model': 1.1146749990293756, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 0.71}) +[2026-06-23 23:24:21,092][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.06348249316215515, 'train/grad_norm_pre_clip': 15.370162963867188, 'timing/data': 0.00041088799480348825, 'timing/model': 1.1201385699678212, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.71}) +[2026-06-23 23:24:25,182][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.04913388565182686, 'train/grad_norm_pre_clip': 24.1018009185791, 'timing/data': 0.0004572200123220682, 'timing/model': 1.1209087929455563, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 0.71}) +[2026-06-23 23:24:29,382][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.03939080238342285, 'train/grad_norm_pre_clip': 22.42995262145996, 'timing/data': 0.0003680730005726218, 'timing/model': 1.1204294810304418, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 0.71}) +[2026-06-23 23:24:33,459][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.2658843696117401, 'train/grad_norm_pre_clip': 32.442291259765625, 'timing/data': 0.00045458204112946987, 'timing/model': 1.1192732119234279, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 0.71}) +[2026-06-23 23:24:37,535][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.06715851277112961, 'train/grad_norm_pre_clip': 8.49242877960205, 'timing/data': 0.00041425798553973436, 'timing/model': 1.1185156360734254, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 0.71}) +[2026-06-23 23:24:41,624][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.05211026966571808, 'train/grad_norm_pre_clip': 9.74462890625, 'timing/data': 0.00036054698284715414, 'timing/model': 1.1196048200363293, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 0.71}) +[2026-06-23 23:24:45,842][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.053637951612472534, 'train/grad_norm_pre_clip': 9.952259063720703, 'timing/data': 0.00036124594043940306, 'timing/model': 1.1160826369887218, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 0.71}) +[2026-06-23 23:24:49,908][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.06538651883602142, 'train/grad_norm_pre_clip': 61.7421875, 'timing/data': 0.00048562500160187483, 'timing/model': 1.114905183087103, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 0.71}) +[2026-06-23 23:24:53,992][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.01969727873802185, 'train/grad_norm_pre_clip': 57.682064056396484, 'timing/data': 0.0003517699660733342, 'timing/model': 1.123059666948393, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 0.71}) +[2026-06-23 23:24:58,069][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.18106907606124878, 'train/grad_norm_pre_clip': 69.61904907226562, 'timing/data': 0.0003563979407772422, 'timing/model': 1.118198781972751, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 0.71}) +[2026-06-23 23:25:02,267][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.010953693650662899, 'train/grad_norm_pre_clip': 12.757230758666992, 'timing/data': 0.0005200390005484223, 'timing/model': 1.1149188720155507, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.71}) +[2026-06-23 23:25:06,340][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.026614665985107422, 'train/grad_norm_pre_clip': 15.583514213562012, 'timing/data': 0.0005730080883949995, 'timing/model': 1.1165468199178576, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 0.71}) +[2026-06-23 23:25:10,429][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.31036338210105896, 'train/grad_norm_pre_clip': 21.898500442504883, 'timing/data': 0.00036115606781095266, 'timing/model': 1.1180599349318072, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 0.71}) +[2026-06-23 23:25:14,517][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.11892294138669968, 'train/grad_norm_pre_clip': 26.85012435913086, 'timing/data': 0.0003714290214702487, 'timing/model': 1.119732809951529, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 0.71}) +[2026-06-23 23:25:18,741][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.11820156127214432, 'train/grad_norm_pre_clip': 21.15742301940918, 'timing/data': 0.00044464704114943743, 'timing/model': 1.1226007021032274, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 0.71}) +[2026-06-23 23:25:22,821][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.11192522197961807, 'train/grad_norm_pre_clip': 23.035425186157227, 'timing/data': 0.0004104999825358391, 'timing/model': 1.1151622590841725, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 0.71}) +[2026-06-23 23:25:26,903][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.08676286041736603, 'train/grad_norm_pre_clip': 27.074176788330078, 'timing/data': 0.0003580329939723015, 'timing/model': 1.1162988109281287, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 0.71}) +[2026-06-23 23:25:30,992][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.024927357211709023, 'train/grad_norm_pre_clip': 15.693843841552734, 'timing/data': 0.0003994959406554699, 'timing/model': 1.1200734600424767, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 0.71}) +[2026-06-23 23:25:35,190][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.051200367510318756, 'train/grad_norm_pre_clip': 14.88094425201416, 'timing/data': 0.00036277901381254196, 'timing/model': 1.1166017079958692, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 0.72}) +[2026-06-23 23:25:39,267][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.04038576781749725, 'train/grad_norm_pre_clip': 9.468557357788086, 'timing/data': 0.0003705270355567336, 'timing/model': 1.1121812459314242, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 0.72}) +[2026-06-23 23:25:43,338][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.023295341059565544, 'train/grad_norm_pre_clip': 13.365743637084961, 'timing/data': 0.0003840388962998986, 'timing/model': 1.1178532210178673, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.72}) +[2026-06-23 23:25:47,412][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.03196725249290466, 'train/grad_norm_pre_clip': 17.702938079833984, 'timing/data': 0.00035619805566966534, 'timing/model': 1.1175119710387662, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 0.72}) +[2026-06-23 23:25:51,607][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.054133668541908264, 'train/grad_norm_pre_clip': 15.628893852233887, 'timing/data': 0.00040667306166142225, 'timing/model': 1.1202468849951401, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 0.72}) +[2026-06-23 23:25:55,689][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.10523010045289993, 'train/grad_norm_pre_clip': 38.74969482421875, 'timing/data': 0.0006604880327358842, 'timing/model': 1.1229541219072416, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 0.72}) +[2026-06-23 23:25:59,797][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.010194296017289162, 'train/grad_norm_pre_clip': 19.74608039855957, 'timing/data': 0.00035747792571783066, 'timing/model': 1.1350384230026975, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 0.72}) +[2026-06-23 23:26:03,891][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.13543029129505157, 'train/grad_norm_pre_clip': 17.46622657775879, 'timing/data': 0.0006774710491299629, 'timing/model': 1.1212255450664088, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 0.72}) +[2026-06-23 23:26:08,080][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.026386002078652382, 'train/grad_norm_pre_clip': 12.593789100646973, 'timing/data': 0.00037345499731600285, 'timing/model': 1.115791745018214, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 0.72}) +[2026-06-23 23:26:12,170][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.08946391940116882, 'train/grad_norm_pre_clip': 18.254924774169922, 'timing/data': 0.000371381058357656, 'timing/model': 1.1230776610318571, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 0.72}) +[2026-06-23 23:26:16,242][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.061124593019485474, 'train/grad_norm_pre_clip': 21.03325653076172, 'timing/data': 0.00037256500218063593, 'timing/model': 1.121998783084564, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 0.72}) +[2026-06-23 23:26:20,326][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.008700000122189522, 'train/grad_norm_pre_clip': 11.612076759338379, 'timing/data': 0.0003684139810502529, 'timing/model': 1.1212467330042273, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 0.72}) +[2026-06-23 23:26:24,522][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.09017818421125412, 'train/grad_norm_pre_clip': 28.094717025756836, 'timing/data': 0.0004209670005366206, 'timing/model': 1.1256998690078035, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.72}) +[2026-06-23 23:26:28,598][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.09969016164541245, 'train/grad_norm_pre_clip': 23.631006240844727, 'timing/data': 0.0003646330442279577, 'timing/model': 1.1213402359280735, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 0.72}) +[2026-06-23 23:26:32,674][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.16236867010593414, 'train/grad_norm_pre_clip': 17.283145904541016, 'timing/data': 0.00041807303205132484, 'timing/model': 1.1103280389215797, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 0.72}) +[2026-06-23 23:26:36,744][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.07560501992702484, 'train/grad_norm_pre_clip': 18.175291061401367, 'timing/data': 0.0003541830228641629, 'timing/model': 1.108929498004727, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 0.72}) +[2026-06-23 23:26:40,935][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.15120692551136017, 'train/grad_norm_pre_clip': 16.858823776245117, 'timing/data': 0.00040367094334214926, 'timing/model': 1.1197393919574097, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.72}) +[2026-06-23 23:26:45,017][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.2205870896577835, 'train/grad_norm_pre_clip': 20.94038963317871, 'timing/data': 0.000646985019557178, 'timing/model': 1.1184687740169466, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 0.72}) +[2026-06-23 23:26:49,096][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.030242830514907837, 'train/grad_norm_pre_clip': 11.558262825012207, 'timing/data': 0.00035881903022527695, 'timing/model': 1.1196750509552658, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 0.72}) +[2026-06-23 23:26:53,307][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.02791525237262249, 'train/grad_norm_pre_clip': 7.030252456665039, 'timing/data': 0.00039044604636728764, 'timing/model': 1.1190683619352058, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 0.72}) +[2026-06-23 23:26:57,498][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.09299136698246002, 'train/grad_norm_pre_clip': 37.80860137939453, 'timing/data': 0.0003556810552254319, 'timing/model': 1.120470465044491, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 0.72}) +[2026-06-23 23:27:01,595][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.06742481142282486, 'train/grad_norm_pre_clip': 110.29652404785156, 'timing/data': 0.00036772002931684256, 'timing/model': 1.1295669219689444, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 0.72}) +[2026-06-23 23:27:05,668][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.05964227393269539, 'train/grad_norm_pre_clip': 10.123917579650879, 'timing/data': 0.0003617280162870884, 'timing/model': 1.1196007950929925, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.72}) +[2026-06-23 23:27:09,755][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.08512119948863983, 'train/grad_norm_pre_clip': 17.905908584594727, 'timing/data': 0.00037732196506112814, 'timing/model': 1.1178026719717309, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 0.72}) +[2026-06-23 23:27:13,941][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.010936704464256763, 'train/grad_norm_pre_clip': 6.376033306121826, 'timing/data': 0.00041487906128168106, 'timing/model': 1.1192970059346408, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 0.72}) +[2026-06-23 23:27:18,038][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.17879292368888855, 'train/grad_norm_pre_clip': 27.835390090942383, 'timing/data': 0.0003542400663718581, 'timing/model': 1.1194750450085849, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 0.72}) +[2026-06-23 23:27:22,132][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.08873546868562698, 'train/grad_norm_pre_clip': 32.71031188964844, 'timing/data': 0.0004536540945991874, 'timing/model': 1.1202956209890544, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 0.73}) +[2026-06-23 23:27:26,206][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.07357285916805267, 'train/grad_norm_pre_clip': 26.124099731445312, 'timing/data': 0.0003734179772436619, 'timing/model': 1.117337097064592, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 0.73}) +[2026-06-23 23:27:30,403][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.011777348816394806, 'train/grad_norm_pre_clip': 8.676688194274902, 'timing/data': 0.00037432205863296986, 'timing/model': 1.1162564320256934, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 0.73}) +[2026-06-23 23:27:34,498][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.0229506716132164, 'train/grad_norm_pre_clip': 27.6549129486084, 'timing/data': 0.0003596689784899354, 'timing/model': 1.122384057031013, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 0.73}) +[2026-06-23 23:27:38,568][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.05334726721048355, 'train/grad_norm_pre_clip': 19.577152252197266, 'timing/data': 0.0005962461000308394, 'timing/model': 1.1168266109889373, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 0.73}) +[2026-06-23 23:27:42,659][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.08634215593338013, 'train/grad_norm_pre_clip': 15.84728717803955, 'timing/data': 0.0003633259329944849, 'timing/model': 1.1280179899185896, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 0.73}) +[2026-06-23 23:27:46,865][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.0976490005850792, 'train/grad_norm_pre_clip': 24.504714965820312, 'timing/data': 0.0003822330618277192, 'timing/model': 1.119465806055814, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.73}) +[2026-06-23 23:27:50,950][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.09802191704511642, 'train/grad_norm_pre_clip': 24.55975341796875, 'timing/data': 0.00036267703399062157, 'timing/model': 1.1240541589213535, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 0.73}) +[2026-06-23 23:27:55,040][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.008586760610342026, 'train/grad_norm_pre_clip': 19.809497833251953, 'timing/data': 0.0004770389059558511, 'timing/model': 1.1212874100310728, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 0.73}) +[2026-06-23 23:27:59,129][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.07198788225650787, 'train/grad_norm_pre_clip': 19.8121337890625, 'timing/data': 0.0005548310000449419, 'timing/model': 1.122444384964183, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 0.73}) +[2026-06-23 23:28:03,322][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.1413910835981369, 'train/grad_norm_pre_clip': 24.289451599121094, 'timing/data': 0.0003679220099002123, 'timing/model': 1.122030935017392, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 0.73}) +[2026-06-23 23:28:07,413][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.16141347587108612, 'train/grad_norm_pre_clip': 23.03995704650879, 'timing/data': 0.0003710939781740308, 'timing/model': 1.1172897379146889, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.73}) +[2026-06-23 23:28:11,490][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.06583390384912491, 'train/grad_norm_pre_clip': 37.65653991699219, 'timing/data': 0.0003798459656536579, 'timing/model': 1.118757908931002, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 0.73}) +[2026-06-23 23:28:15,569][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.14738880097866058, 'train/grad_norm_pre_clip': 20.785249710083008, 'timing/data': 0.00036478007677942514, 'timing/model': 1.1185673769796267, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 0.73}) +[2026-06-23 23:28:19,780][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.08636375516653061, 'train/grad_norm_pre_clip': 14.420158386230469, 'timing/data': 0.000429593026638031, 'timing/model': 1.122701216954738, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 0.73}) +[2026-06-23 23:28:23,862][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.053401101380586624, 'train/grad_norm_pre_clip': 24.745075225830078, 'timing/data': 0.0006497690919786692, 'timing/model': 1.116366858012043, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 0.73}) +[2026-06-23 23:28:27,950][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.05375457927584648, 'train/grad_norm_pre_clip': 18.03411865234375, 'timing/data': 0.0005915999645367265, 'timing/model': 1.117655465961434, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.73}) +[2026-06-23 23:28:32,043][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.23630814254283905, 'train/grad_norm_pre_clip': 20.52592658996582, 'timing/data': 0.00037758296821266413, 'timing/model': 1.1187722100876272, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 0.73}) +[2026-06-23 23:28:36,243][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.10295264422893524, 'train/grad_norm_pre_clip': 19.75786018371582, 'timing/data': 0.00040985795203596354, 'timing/model': 1.1164111039834097, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 0.73}) +[2026-06-23 23:28:40,336][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.10486079007387161, 'train/grad_norm_pre_clip': 15.640107154846191, 'timing/data': 0.00035633600782603025, 'timing/model': 1.1262348250020295, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 0.73}) +[2026-06-23 23:28:44,428][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.03256295248866081, 'train/grad_norm_pre_clip': 12.382769584655762, 'timing/data': 0.0003565700026229024, 'timing/model': 1.1261839530197904, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 0.73}) +[2026-06-23 23:28:48,509][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.04269770532846451, 'train/grad_norm_pre_clip': 19.82544708251953, 'timing/data': 0.00034919893369078636, 'timing/model': 1.1234669469995424, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 0.73}) +[2026-06-23 23:28:52,722][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.06519590318202972, 'train/grad_norm_pre_clip': 31.255212783813477, 'timing/data': 0.00035750202368944883, 'timing/model': 1.1183535610325634, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 0.73}) +[2026-06-23 23:28:56,800][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.16701768338680267, 'train/grad_norm_pre_clip': 23.721759796142578, 'timing/data': 0.000368863926269114, 'timing/model': 1.111390564008616, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 0.73}) +[2026-06-23 23:29:00,915][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.10708755999803543, 'train/grad_norm_pre_clip': 19.681516647338867, 'timing/data': 0.00037690403405576944, 'timing/model': 1.1210786500014365, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 0.73}) +[2026-06-23 23:29:05,009][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.07118531316518784, 'train/grad_norm_pre_clip': 10.420910835266113, 'timing/data': 0.00035917700733989477, 'timing/model': 1.124582485994324, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 0.73}) +[2026-06-23 23:29:09,224][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1900 +[2026-06-23 23:30:13,198][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1900 +[2026-06-23 23:30:13,200][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1900 +[2026-06-23 23:30:51,956][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1900 +[2026-06-23 23:30:51,957][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.22798368334770203, 'train/grad_norm_pre_clip': 10.413114547729492, 'eval/action_loss/seconds': 63.971872642985545, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.04923257231712341, 'eval/latency_2/loss': 0.04923257231712341, 'eval/flappy/loss': 0.04923257231712341, 'eval/flappy/latency_2/loss': 0.04923257231712341, 'eval/flap_precision': 0.9259259259259259, 'eval/flap_recall': 0.9837177747625508, 'eval/flap_f1': 0.9539473684210527, 'eval/flap_rate_model': 0.2175, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9539473684210528, 'eval/latency_2/flap_precision': 0.9259259259259259, 'eval/latency_2/flap_recall': 0.9837177747625508, 'eval/latency_2/flap_f1': 0.9539473684210527, 'eval/latency_2/flap_rate_model': 0.2175, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9539473684210528, 'eval/action_classification/data_wait_seconds': 0.46637189388275146, 'eval/action_classification/predict_seconds': 38.13883590698242, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00035708595532923937, 'timing/model': 1.115750313969329, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.74}) +[2026-06-23 23:30:56,000][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.1712074875831604, 'train/grad_norm_pre_clip': 20.44530487060547, 'timing/data': 0.00038877897895872593, 'timing/model': 1.1096563159953803, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 0.74}) +[2026-06-23 23:31:00,053][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.10916414111852646, 'train/grad_norm_pre_clip': 21.70901107788086, 'timing/data': 0.00041015096940100193, 'timing/model': 1.112355884979479, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 0.74}) +[2026-06-23 23:31:04,294][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.1202850267291069, 'train/grad_norm_pre_clip': 25.49904441833496, 'timing/data': 0.0003638259368017316, 'timing/model': 1.1125433669658378, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 0.74}) +[2026-06-23 23:31:08,367][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.04687849059700966, 'train/grad_norm_pre_clip': 20.947927474975586, 'timing/data': 0.0003785570152103901, 'timing/model': 1.1189848489593714, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 0.74}) +[2026-06-23 23:31:12,448][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.06069106236100197, 'train/grad_norm_pre_clip': 10.380331039428711, 'timing/data': 0.0004403049824759364, 'timing/model': 1.1137966060778126, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 0.74}) +[2026-06-23 23:31:16,505][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.13702379167079926, 'train/grad_norm_pre_clip': 28.6395263671875, 'timing/data': 0.0006015279795974493, 'timing/model': 1.1109210579888895, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 0.74}) +[2026-06-23 23:31:20,637][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.10773219168186188, 'train/grad_norm_pre_clip': 11.533015251159668, 'timing/data': 0.00035904801916331053, 'timing/model': 1.1146484829951078, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 0.74}) +[2026-06-23 23:31:24,703][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.19788385927677155, 'train/grad_norm_pre_clip': 26.192028045654297, 'timing/data': 0.0003682069946080446, 'timing/model': 1.1107245639432222, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 0.74}) +[2026-06-23 23:31:28,769][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.13081857562065125, 'train/grad_norm_pre_clip': 31.061038970947266, 'timing/data': 0.0003497221041470766, 'timing/model': 1.1140113569563255, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 0.74}) +[2026-06-23 23:31:32,830][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.017891323193907738, 'train/grad_norm_pre_clip': 16.677791595458984, 'timing/data': 0.0003535919822752476, 'timing/model': 1.1122219410026446, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.74}) +[2026-06-23 23:31:37,043][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.07522590458393097, 'train/grad_norm_pre_clip': 17.521032333374023, 'timing/data': 0.00035542098339647055, 'timing/model': 1.1123804999515414, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 0.74}) +[2026-06-23 23:31:41,093][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.12139760702848434, 'train/grad_norm_pre_clip': 14.987526893615723, 'timing/data': 0.0006065280176699162, 'timing/model': 1.111858656979166, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 0.74}) +[2026-06-23 23:31:45,163][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.036856699734926224, 'train/grad_norm_pre_clip': 26.382604598999023, 'timing/data': 0.0007830540416762233, 'timing/model': 1.1132169150514528, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.74}) +[2026-06-23 23:31:49,236][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.02430913969874382, 'train/grad_norm_pre_clip': 13.237529754638672, 'timing/data': 0.0005527170142158866, 'timing/model': 1.117962924996391, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 0.74}) +[2026-06-23 23:31:53,428][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.06533610075712204, 'train/grad_norm_pre_clip': 15.050711631774902, 'timing/data': 0.00045760103967040777, 'timing/model': 1.116366452886723, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 0.74}) +[2026-06-23 23:31:57,510][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.06656131893396378, 'train/grad_norm_pre_clip': 18.545270919799805, 'timing/data': 0.0003786690067499876, 'timing/model': 1.115890054963529, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 0.74}) +[2026-06-23 23:32:01,586][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.021989470347762108, 'train/grad_norm_pre_clip': 18.125940322875977, 'timing/data': 0.0003628470003604889, 'timing/model': 1.1152159119956195, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 0.74}) +[2026-06-23 23:32:05,742][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.11872315406799316, 'train/grad_norm_pre_clip': 14.991861343383789, 'timing/data': 0.00036059506237506866, 'timing/model': 1.1871591969393194, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 0.74}) +[2026-06-23 23:32:09,802][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.02771899476647377, 'train/grad_norm_pre_clip': 18.51935386657715, 'timing/data': 0.00041912298183888197, 'timing/model': 1.1124082619789988, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 0.74}) +[2026-06-23 23:32:13,881][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.02802092209458351, 'train/grad_norm_pre_clip': 10.790125846862793, 'timing/data': 0.00041963194962590933, 'timing/model': 1.1166283219354227, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.74}) +[2026-06-23 23:32:17,964][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.029357684776186943, 'train/grad_norm_pre_clip': 13.255938529968262, 'timing/data': 0.0006011209916323423, 'timing/model': 1.1219924190081656, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 0.74}) +[2026-06-23 23:32:22,034][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.0827421024441719, 'train/grad_norm_pre_clip': 29.814990997314453, 'timing/data': 0.00040313892532140017, 'timing/model': 1.1172621899750084, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 0.74}) +[2026-06-23 23:32:26,244][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.13257016241550446, 'train/grad_norm_pre_clip': 15.936501502990723, 'timing/data': 0.00035296601708978415, 'timing/model': 1.1137158210622147, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 0.74}) +[2026-06-23 23:32:30,356][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.11394205689430237, 'train/grad_norm_pre_clip': 21.0078125, 'timing/data': 0.0003583139041438699, 'timing/model': 1.1196392129641026, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 0.74}) +[2026-06-23 23:32:34,426][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.043246883898973465, 'train/grad_norm_pre_clip': 10.497109413146973, 'timing/data': 0.00035549793392419815, 'timing/model': 1.1156330519588664, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 0.75}) +[2026-06-23 23:32:38,530][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.15058638155460358, 'train/grad_norm_pre_clip': 27.726606369018555, 'timing/data': 0.0005669939564540982, 'timing/model': 1.1203596759587526, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 0.75}) +[2026-06-23 23:32:42,756][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.24955327808856964, 'train/grad_norm_pre_clip': 12.53242301940918, 'timing/data': 0.000699406024068594, 'timing/model': 1.1162665890296921, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 0.75}) +[2026-06-23 23:32:46,829][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.14604881405830383, 'train/grad_norm_pre_clip': 29.653461456298828, 'timing/data': 0.00035501900129020214, 'timing/model': 1.1164333989145234, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 0.75}) +[2026-06-23 23:32:50,911][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.29415929317474365, 'train/grad_norm_pre_clip': 46.33835220336914, 'timing/data': 0.0003644099924713373, 'timing/model': 1.1170155829750001, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 0.75}) +[2026-06-23 23:32:54,992][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.0496179535984993, 'train/grad_norm_pre_clip': 21.91216468811035, 'timing/data': 0.00036188692320138216, 'timing/model': 1.1194108719937503, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.75}) +[2026-06-23 23:32:59,152][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.09670878201723099, 'train/grad_norm_pre_clip': 28.56630516052246, 'timing/data': 0.00038486707489937544, 'timing/model': 1.1209612370003015, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 0.75}) +[2026-06-23 23:33:03,258][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.04969489574432373, 'train/grad_norm_pre_clip': 30.161958694458008, 'timing/data': 0.0005782829830422997, 'timing/model': 1.1190255749970675, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 0.75}) +[2026-06-23 23:33:07,337][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.0776328518986702, 'train/grad_norm_pre_clip': 28.888782501220703, 'timing/data': 0.00040663208346813917, 'timing/model': 1.1184944169363007, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 0.75}) +[2026-06-23 23:33:11,418][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.010448676533997059, 'train/grad_norm_pre_clip': 13.46660041809082, 'timing/data': 0.000409752014093101, 'timing/model': 1.116822800016962, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 0.75}) +[2026-06-23 23:33:15,582][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.2159392535686493, 'train/grad_norm_pre_clip': 32.808475494384766, 'timing/data': 0.0003579069161787629, 'timing/model': 1.1197605449706316, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 0.75}) +[2026-06-23 23:33:19,662][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.01327755767852068, 'train/grad_norm_pre_clip': 14.129505157470703, 'timing/data': 0.00036196596920490265, 'timing/model': 1.1186726230662316, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 0.75}) +[2026-06-23 23:33:23,746][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.002465300727635622, 'train/grad_norm_pre_clip': 37.998191833496094, 'timing/data': 0.0004007200477644801, 'timing/model': 1.1220267119351774, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 0.75}) +[2026-06-23 23:33:27,834][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.12017236649990082, 'train/grad_norm_pre_clip': 19.072872161865234, 'timing/data': 0.00040198094211518764, 'timing/model': 1.1211506049148738, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 0.75}) +[2026-06-23 23:33:32,071][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.07232619822025299, 'train/grad_norm_pre_clip': 28.103843688964844, 'timing/data': 0.00040779600385576487, 'timing/model': 1.1185474259546027, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 0.75}) +[2026-06-23 23:33:36,161][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.052972763776779175, 'train/grad_norm_pre_clip': 6.324748516082764, 'timing/data': 0.00036554294638335705, 'timing/model': 1.1176003370201215, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.75}) +[2026-06-23 23:33:40,238][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.05561794340610504, 'train/grad_norm_pre_clip': 26.328853607177734, 'timing/data': 0.000360064092092216, 'timing/model': 1.1162534039467573, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 0.75}) +[2026-06-23 23:33:44,323][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.21528246998786926, 'train/grad_norm_pre_clip': 38.060733795166016, 'timing/data': 0.0006507060024887323, 'timing/model': 1.118940411019139, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 0.75}) +[2026-06-23 23:33:48,570][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.1873399019241333, 'train/grad_norm_pre_clip': 35.909934997558594, 'timing/data': 0.0004908209666609764, 'timing/model': 1.1184048230061308, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 0.75}) +[2026-06-23 23:33:52,665][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.23395484685897827, 'train/grad_norm_pre_clip': 39.929866790771484, 'timing/data': 0.00043548003304749727, 'timing/model': 1.118785101105459, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 0.75}) +[2026-06-23 23:33:56,731][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.12601584196090698, 'train/grad_norm_pre_clip': 24.13181495666504, 'timing/data': 0.00035542307887226343, 'timing/model': 1.1123448569560423, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 0.75}) +[2026-06-23 23:34:00,813][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.054477449506521225, 'train/grad_norm_pre_clip': 10.97374153137207, 'timing/data': 0.00035427301190793514, 'timing/model': 1.1166089529870078, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 0.75}) +[2026-06-23 23:34:05,065][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.09372544288635254, 'train/grad_norm_pre_clip': 19.668684005737305, 'timing/data': 0.000496201915666461, 'timing/model': 1.118434865027666, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 0.75}) +[2026-06-23 23:34:09,159][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.14176782965660095, 'train/grad_norm_pre_clip': 20.124958038330078, 'timing/data': 0.0003636159235611558, 'timing/model': 1.124243592028506, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 0.75}) +[2026-06-23 23:34:13,249][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.03649060055613518, 'train/grad_norm_pre_clip': 17.23359489440918, 'timing/data': 0.00035358197055757046, 'timing/model': 1.1203166249906644, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 0.75}) +[2026-06-23 23:34:17,346][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.12074004113674164, 'train/grad_norm_pre_clip': 17.094738006591797, 'timing/data': 0.00036238902248442173, 'timing/model': 1.1305913929827511, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.75}) +[2026-06-23 23:34:21,612][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.23416511714458466, 'train/grad_norm_pre_clip': 31.11760902404785, 'timing/data': 0.00036725797690451145, 'timing/model': 1.1171123220119625, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 0.76}) +[2026-06-23 23:34:25,687][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.07683360576629639, 'train/grad_norm_pre_clip': 15.4642915725708, 'timing/data': 0.0003537459997460246, 'timing/model': 1.1191535020479932, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 0.76}) +[2026-06-23 23:34:29,760][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.042805980890989304, 'train/grad_norm_pre_clip': 10.127545356750488, 'timing/data': 0.00037582602817565203, 'timing/model': 1.1179726589471102, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 0.76}) +[2026-06-23 23:34:33,848][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.07195614278316498, 'train/grad_norm_pre_clip': 31.95976448059082, 'timing/data': 0.0004099590005353093, 'timing/model': 1.1172612580703571, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 0.76}) +[2026-06-23 23:34:38,114][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.0743468627333641, 'train/grad_norm_pre_clip': 18.893415451049805, 'timing/data': 0.0003971339901909232, 'timing/model': 1.1183393499813974, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.76}) +[2026-06-23 23:34:42,199][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.026905063539743423, 'train/grad_norm_pre_clip': 14.732824325561523, 'timing/data': 0.0005157130071893334, 'timing/model': 1.1188246809178963, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 0.76}) +[2026-06-23 23:34:46,297][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.03180261701345444, 'train/grad_norm_pre_clip': 29.503582000732422, 'timing/data': 0.00035630003549158573, 'timing/model': 1.1216163740027696, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 0.76}) +[2026-06-23 23:34:50,398][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.11026711016893387, 'train/grad_norm_pre_clip': 12.045039176940918, 'timing/data': 0.0003679700894281268, 'timing/model': 1.1236373720457777, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 0.76}) +[2026-06-23 23:34:54,669][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.09712538868188858, 'train/grad_norm_pre_clip': 11.734078407287598, 'timing/data': 0.00036841596011072397, 'timing/model': 1.125541614019312, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 0.76}) +[2026-06-23 23:34:58,742][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.05384589359164238, 'train/grad_norm_pre_clip': 13.88210391998291, 'timing/data': 0.0003625329118221998, 'timing/model': 1.1146799140842631, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.76}) +[2026-06-23 23:35:02,834][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.06750426441431046, 'train/grad_norm_pre_clip': 13.464140892028809, 'timing/data': 0.0003958070883527398, 'timing/model': 1.1196302650496364, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 0.76}) +[2026-06-23 23:35:06,938][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.2930747866630554, 'train/grad_norm_pre_clip': 13.668736457824707, 'timing/data': 0.0003550340188667178, 'timing/model': 1.1267179279820994, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.76}) +[2026-06-23 23:35:11,204][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.034685708582401276, 'train/grad_norm_pre_clip': 8.043410301208496, 'timing/data': 0.0003673830069601536, 'timing/model': 1.1218310469994321, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 0.76}) +[2026-06-23 23:35:15,281][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.01343019213527441, 'train/grad_norm_pre_clip': 8.481142044067383, 'timing/data': 0.00037447200156748295, 'timing/model': 1.1183789229253307, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 0.76}) +[2026-06-23 23:35:19,362][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.04969941824674606, 'train/grad_norm_pre_clip': 23.98236656188965, 'timing/data': 0.00046885397750884295, 'timing/model': 1.1170766319846734, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 0.76}) +[2026-06-23 23:35:23,447][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.018035752698779106, 'train/grad_norm_pre_clip': 35.11062240600586, 'timing/data': 0.0004386790096759796, 'timing/model': 1.1161065939813852, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 0.76}) +[2026-06-23 23:35:27,716][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.024179158732295036, 'train/grad_norm_pre_clip': 45.258827209472656, 'timing/data': 0.0003772379131987691, 'timing/model': 1.1311920640291646, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 0.76}) +[2026-06-23 23:35:31,806][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.08596748858690262, 'train/grad_norm_pre_clip': 23.39432144165039, 'timing/data': 0.00035718607250601053, 'timing/model': 1.1173714400501922, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 0.76}) +[2026-06-23 23:35:35,877][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.014992432668805122, 'train/grad_norm_pre_clip': 15.57253646850586, 'timing/data': 0.0006463580066338181, 'timing/model': 1.1119239369872957, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 0.76}) +[2026-06-23 23:35:39,967][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.19922050833702087, 'train/grad_norm_pre_clip': 24.312742233276367, 'timing/data': 0.0004113810136914253, 'timing/model': 1.1222281149821356, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.76}) +[2026-06-23 23:35:44,224][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.17903390526771545, 'train/grad_norm_pre_clip': 28.121870040893555, 'timing/data': 0.00036179996095597744, 'timing/model': 1.1183252449845895, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 0.76}) +[2026-06-23 23:35:48,322][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.06758357584476471, 'train/grad_norm_pre_clip': 10.331204414367676, 'timing/data': 0.00035545206628739834, 'timing/model': 1.115674672066234, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 0.76}) +[2026-06-23 23:35:52,404][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.09024695307016373, 'train/grad_norm_pre_clip': 17.978473663330078, 'timing/data': 0.0005533939693123102, 'timing/model': 1.1205880609340966, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 0.76}) +[2026-06-23 23:35:56,497][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.03597696125507355, 'train/grad_norm_pre_clip': 17.190778732299805, 'timing/data': 0.0005954219959676266, 'timing/model': 1.1212095079245046, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 0.76}) +[2026-06-23 23:36:00,735][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.016051078215241432, 'train/grad_norm_pre_clip': 22.620384216308594, 'timing/data': 0.00037389888893812895, 'timing/model': 1.1094600859796628, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 0.76}) +[2026-06-23 23:36:04,803][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.1133139431476593, 'train/grad_norm_pre_clip': 10.005048751831055, 'timing/data': 0.0004145549610257149, 'timing/model': 1.1139814279740676, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 0.76}) +[2026-06-23 23:36:08,892][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.02220284938812256, 'train/grad_norm_pre_clip': 9.530096054077148, 'timing/data': 0.0003593940054997802, 'timing/model': 1.1212617370765656, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 0.77}) +[2026-06-23 23:36:12,981][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.037243813276290894, 'train/grad_norm_pre_clip': 8.53650188446045, 'timing/data': 0.0005063330754637718, 'timing/model': 1.1200476179365069, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 0.77}) +[2026-06-23 23:36:17,230][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.14458486437797546, 'train/grad_norm_pre_clip': 25.8226261138916, 'timing/data': 0.00037260795943439007, 'timing/model': 1.1191777910571545, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 0.77}) +[2026-06-23 23:36:21,321][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.139018714427948, 'train/grad_norm_pre_clip': 18.43044662475586, 'timing/data': 0.00036153197288513184, 'timing/model': 1.1256959249731153, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.77}) +[2026-06-23 23:36:25,411][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.031564123928546906, 'train/grad_norm_pre_clip': 12.691039085388184, 'timing/data': 0.00036098400596529245, 'timing/model': 1.1213272879831493, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 0.77}) +[2026-06-23 23:36:29,499][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.08668956160545349, 'train/grad_norm_pre_clip': 16.94854164123535, 'timing/data': 0.00038108101580291986, 'timing/model': 1.118861711001955, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 0.77}) +[2026-06-23 23:36:33,758][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.019613545387983322, 'train/grad_norm_pre_clip': 7.307260036468506, 'timing/data': 0.0003583610523492098, 'timing/model': 1.1194887730525807, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 0.77}) +[2026-06-23 23:36:37,840][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.0053649721667170525, 'train/grad_norm_pre_clip': 10.262706756591797, 'timing/data': 0.0005795870674774051, 'timing/model': 1.1160002590622753, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 0.77}) +[2026-06-23 23:36:41,926][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.005020995624363422, 'train/grad_norm_pre_clip': 26.216413497924805, 'timing/data': 0.0006813529180362821, 'timing/model': 1.116830263985321, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 0.77}) +[2026-06-23 23:36:46,023][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.01749359630048275, 'train/grad_norm_pre_clip': 15.404892921447754, 'timing/data': 0.00036255898885428905, 'timing/model': 1.1175309729296714, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 0.77}) +[2026-06-23 23:36:50,283][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.28082725405693054, 'train/grad_norm_pre_clip': 42.352298736572266, 'timing/data': 0.00042847509030252695, 'timing/model': 1.1269496550085023, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 0.77}) +[2026-06-23 23:36:54,363][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.0326799638569355, 'train/grad_norm_pre_clip': 18.395950317382812, 'timing/data': 0.00036329193972051144, 'timing/model': 1.125149153987877, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 0.77}) +[2026-06-23 23:36:58,459][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.09732186049222946, 'train/grad_norm_pre_clip': 24.865633010864258, 'timing/data': 0.0003542080521583557, 'timing/model': 1.1200908089522272, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 0.77}) +[2026-06-23 23:37:02,548][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.005361029412597418, 'train/grad_norm_pre_clip': 17.999710083007812, 'timing/data': 0.0003707719733938575, 'timing/model': 1.1121190859703347, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.77}) +[2026-06-23 23:37:06,811][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.02415415272116661, 'train/grad_norm_pre_clip': 10.994059562683105, 'timing/data': 0.00035635801032185555, 'timing/model': 1.120399669976905, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 0.77}) +[2026-06-23 23:37:10,898][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.0012034166138619184, 'train/grad_norm_pre_clip': 21.281652450561523, 'timing/data': 0.00035982392728328705, 'timing/model': 1.120487879961729, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 0.77}) +[2026-06-23 23:37:14,997][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.11296901851892471, 'train/grad_norm_pre_clip': 37.78837203979492, 'timing/data': 0.000362965976819396, 'timing/model': 1.1209332710132003, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 0.77}) +[2026-06-23 23:37:19,089][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.029195407405495644, 'train/grad_norm_pre_clip': 10.706243515014648, 'timing/data': 0.00035603006836026907, 'timing/model': 1.1262953240657225, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 0.77}) +[2026-06-23 23:37:23,329][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.0015118656447157264, 'train/grad_norm_pre_clip': 15.345464706420898, 'timing/data': 0.00046286103315651417, 'timing/model': 1.120363711961545, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 0.77}) +[2026-06-23 23:37:27,442][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.05227034166455269, 'train/grad_norm_pre_clip': 13.251045227050781, 'timing/data': 0.00036185397766530514, 'timing/model': 1.1204655420733616, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 0.77}) +[2026-06-23 23:37:31,533][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.04600334167480469, 'train/grad_norm_pre_clip': 19.545408248901367, 'timing/data': 0.00036248506512492895, 'timing/model': 1.1182089939247817, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 0.77}) +[2026-06-23 23:37:35,628][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.05765946954488754, 'train/grad_norm_pre_clip': 14.195491790771484, 'timing/data': 0.0003669440047815442, 'timing/model': 1.1201010379008949, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 0.77}) +[2026-06-23 23:37:39,899][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.01274776365607977, 'train/grad_norm_pre_clip': 15.599068641662598, 'timing/data': 0.00035658199340105057, 'timing/model': 1.1219430950004607, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 0.77}) +[2026-06-23 23:37:43,977][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2000 +[2026-06-23 23:38:48,174][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2000 +[2026-06-23 23:38:48,175][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2000 +[2026-06-23 23:39:27,322][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2000 +[2026-06-23 23:47:04,445][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-06-23 23:47:04,446][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.03269375488162041, 'train/grad_norm_pre_clip': 15.108436584472656, 'eval/action_loss/seconds': 64.19457897800021, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.037909697741270065, 'eval/latency_2/loss': 0.037909697741270065, 'eval/flappy/loss': 0.037909697741270065, 'eval/flappy/latency_2/loss': 0.037909697741270065, 'eval/flap_precision': 0.963265306122449, 'eval/flap_recall': 0.9606512890094979, 'eval/flap_f1': 0.9619565217391304, 'eval/flap_rate_model': 0.20416666666666666, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9619565217391305, 'eval/latency_2/flap_precision': 0.963265306122449, 'eval/latency_2/flap_recall': 0.9606512890094979, 'eval/latency_2/flap_f1': 0.9619565217391304, 'eval/latency_2/flap_rate_model': 0.20416666666666666, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9619565217391305, 'eval/action_classification/data_wait_seconds': 0.5336084961891174, 'eval/action_classification/predict_seconds': 38.48283386230469, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 282.3200000000154, 'rl_games_eval/mid_train/mean_length': 2293.0, 'rl_games_eval/mid_train/std_reward': 199.98049304870963, 'rl_games_eval/mid_train/std_length': 1610.6818431956076, 'rl_games_eval/mid_train/macro_mean_reward': 282.3200000000154, 'rl_games_eval/mid_train/macro_mean_length': 2293.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 282.3200000000154, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2293.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 199.98049304870963, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1610.6818431956076, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 282.3200000000154, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0004087379202246666, 'timing/model': 1.1201748689636588, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.77}) +[2026-06-23 23:47:04,447][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_2000_state +[2026-06-23 23:47:04,452][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 23:47:08,126][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_2000_state/model.safetensors +[2026-06-23 23:47:15,905][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_2000_state/optimizer.bin +[2026-06-23 23:47:15,905][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_2000_state/sampler.bin +[2026-06-23 23:47:15,906][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-23 23:47:15,906][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-23 23:47:15,939][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-23 23:47:15,940][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 23:47:15,951][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 23:47:20,055][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.14476408064365387, 'train/grad_norm_pre_clip': 26.343746185302734, 'timing/data': 0.00035994104109704494, 'timing/model': 1.0860633630072698, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 0.77}) +[2026-06-23 23:47:24,020][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.07169406116008759, 'train/grad_norm_pre_clip': 19.294694900512695, 'timing/data': 0.0004286780022084713, 'timing/model': 1.0867385619785637, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 0.77}) +[2026-06-23 23:47:27,987][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.03897062689065933, 'train/grad_norm_pre_clip': 14.492270469665527, 'timing/data': 0.0003656740300357342, 'timing/model': 1.0925077570136636, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 0.78}) +[2026-06-23 23:47:31,974][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.008395045064389706, 'train/grad_norm_pre_clip': 12.522571563720703, 'timing/data': 0.0003601060016080737, 'timing/model': 1.0982489460147917, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 0.78}) +[2026-06-23 23:47:36,095][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.13374313712120056, 'train/grad_norm_pre_clip': 27.021881103515625, 'timing/data': 0.0003981409827247262, 'timing/model': 1.0949294159654528, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 0.78}) +[2026-06-23 23:47:40,105][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.07047984004020691, 'train/grad_norm_pre_clip': 14.003748893737793, 'timing/data': 0.0006869420176371932, 'timing/model': 1.1004287720425054, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 0.78}) +[2026-06-23 23:47:44,113][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.10409567505121231, 'train/grad_norm_pre_clip': 17.01014518737793, 'timing/data': 0.0003910450031980872, 'timing/model': 1.1095094690099359, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 0.78}) +[2026-06-23 23:47:48,136][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.18184919655323029, 'train/grad_norm_pre_clip': 22.724763870239258, 'timing/data': 0.0003752519842237234, 'timing/model': 1.1001131649827585, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 0.78}) +[2026-06-23 23:47:52,258][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.1356016844511032, 'train/grad_norm_pre_clip': 15.818207740783691, 'timing/data': 0.0006536260480061173, 'timing/model': 1.1050332490121946, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 0.78}) +[2026-06-23 23:47:56,286][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.03956814110279083, 'train/grad_norm_pre_clip': 19.06093978881836, 'timing/data': 0.0004529319703578949, 'timing/model': 1.1038206379162148, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.78}) +[2026-06-23 23:48:00,307][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.06565997004508972, 'train/grad_norm_pre_clip': 20.39771270751953, 'timing/data': 0.0007857559248805046, 'timing/model': 1.1055194590007886, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.78}) +[2026-06-23 23:48:04,351][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.029057063162326813, 'train/grad_norm_pre_clip': 19.546201705932617, 'timing/data': 0.0004109949804842472, 'timing/model': 1.1092633260414004, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 0.78}) +[2026-06-23 23:48:08,552][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.01328169647604227, 'train/grad_norm_pre_clip': 11.787269592285156, 'timing/data': 0.00037700694520026445, 'timing/model': 1.1060010140063241, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 0.78}) +[2026-06-23 23:48:12,592][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.03626113757491112, 'train/grad_norm_pre_clip': 7.9410223960876465, 'timing/data': 0.0007166489958763123, 'timing/model': 1.1129732190165669, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 0.78}) +[2026-06-23 23:48:16,651][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.08662153780460358, 'train/grad_norm_pre_clip': 23.030303955078125, 'timing/data': 0.0003800239646807313, 'timing/model': 1.1140859469305724, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 0.78}) +[2026-06-23 23:48:20,793][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.009433391503989697, 'train/grad_norm_pre_clip': 11.469311714172363, 'timing/data': 0.0003698329674080014, 'timing/model': 1.1897834410192445, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 0.78}) +[2026-06-23 23:48:24,836][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.04237479716539383, 'train/grad_norm_pre_clip': 13.445502281188965, 'timing/data': 0.00040823896415531635, 'timing/model': 1.1102784909307957, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 0.78}) +[2026-06-23 23:48:28,892][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.04127488285303116, 'train/grad_norm_pre_clip': 22.348432540893555, 'timing/data': 0.0004080309299752116, 'timing/model': 1.116167793981731, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 0.78}) +[2026-06-23 23:48:32,943][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.010848929174244404, 'train/grad_norm_pre_clip': 15.449475288391113, 'timing/data': 0.0005747789982706308, 'timing/model': 1.113964464981109, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 0.78}) +[2026-06-23 23:48:37,079][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.06895327568054199, 'train/grad_norm_pre_clip': 19.069576263427734, 'timing/data': 0.0003978670574724674, 'timing/model': 1.1931447410024703, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.78}) +[2026-06-23 23:48:41,148][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.058236971497535706, 'train/grad_norm_pre_clip': 13.187688827514648, 'timing/data': 0.000363302999176085, 'timing/model': 1.11031448491849, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 0.78}) +[2026-06-23 23:48:45,209][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.010312720201909542, 'train/grad_norm_pre_clip': 19.471179962158203, 'timing/data': 0.00040244904812425375, 'timing/model': 1.1138822759967297, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 0.78}) +[2026-06-23 23:48:49,270][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.03438723459839821, 'train/grad_norm_pre_clip': 14.336604118347168, 'timing/data': 0.00040624302346259356, 'timing/model': 1.1130695780739188, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 0.78}) +[2026-06-23 23:48:53,454][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.05674969032406807, 'train/grad_norm_pre_clip': 20.996774673461914, 'timing/data': 0.00036697695031762123, 'timing/model': 1.2348413689760491, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 0.78}) +[2026-06-23 23:48:57,513][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.08573265373706818, 'train/grad_norm_pre_clip': 18.61167335510254, 'timing/data': 0.00038522202521562576, 'timing/model': 1.1125702619319782, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.78}) +[2026-06-23 23:49:01,583][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.11970154941082001, 'train/grad_norm_pre_clip': 15.248608589172363, 'timing/data': 0.0003591689746826887, 'timing/model': 1.1198114010039717, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 0.78}) +[2026-06-23 23:49:05,678][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.023455072194337845, 'train/grad_norm_pre_clip': 16.92934799194336, 'timing/data': 0.00048473000060766935, 'timing/model': 1.1334551880136132, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 0.78}) +[2026-06-23 23:49:09,833][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.09669594466686249, 'train/grad_norm_pre_clip': 16.042495727539062, 'timing/data': 0.000374104012735188, 'timing/model': 1.1155596930766478, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 0.78}) +[2026-06-23 23:49:13,910][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.0027898799162358046, 'train/grad_norm_pre_clip': 18.33483123779297, 'timing/data': 0.0006090350216254592, 'timing/model': 1.1175911410246044, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 0.79}) +[2026-06-23 23:49:18,001][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.018393877893686295, 'train/grad_norm_pre_clip': 14.577836036682129, 'timing/data': 0.0003837760305032134, 'timing/model': 1.1180415559792891, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.79}) +[2026-06-23 23:49:22,093][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.034080810844898224, 'train/grad_norm_pre_clip': 15.290904998779297, 'timing/data': 0.0003584639634937048, 'timing/model': 1.1166314110159874, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 0.79}) +[2026-06-23 23:49:26,230][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.1057802140712738, 'train/grad_norm_pre_clip': 19.66816520690918, 'timing/data': 0.0005188769428059459, 'timing/model': 1.1077497079968452, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 0.79}) +[2026-06-23 23:49:30,315][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.05684494227170944, 'train/grad_norm_pre_clip': 19.455490112304688, 'timing/data': 0.0005124310264363885, 'timing/model': 1.1182662040228024, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 0.79}) +[2026-06-23 23:49:34,384][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.0432000495493412, 'train/grad_norm_pre_clip': 23.292221069335938, 'timing/data': 0.00040615396574139595, 'timing/model': 1.1166920829564333, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 0.79}) +[2026-06-23 23:49:38,465][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.0889909416437149, 'train/grad_norm_pre_clip': 31.36867332458496, 'timing/data': 0.0003614000743255019, 'timing/model': 1.116844987962395, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 0.79}) +[2026-06-23 23:49:42,727][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.0436127744615078, 'train/grad_norm_pre_clip': 34.592620849609375, 'timing/data': 0.0003582190256565809, 'timing/model': 1.1179259630152956, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 0.79}) +[2026-06-23 23:49:46,795][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.02800736390054226, 'train/grad_norm_pre_clip': 14.291120529174805, 'timing/data': 0.0003684449475258589, 'timing/model': 1.1161135219736025, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 0.79}) +[2026-06-23 23:49:50,871][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.06958100944757462, 'train/grad_norm_pre_clip': 29.275360107421875, 'timing/data': 0.000415010959841311, 'timing/model': 1.1152268450241536, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 0.79}) +[2026-06-23 23:49:54,956][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.011028283275663853, 'train/grad_norm_pre_clip': 27.911184310913086, 'timing/data': 0.0006589529803022742, 'timing/model': 1.1188699300400913, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 0.79}) +[2026-06-23 23:49:59,125][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.07302330434322357, 'train/grad_norm_pre_clip': 40.91161346435547, 'timing/data': 0.0005381510127335787, 'timing/model': 1.119225948001258, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.79}) +[2026-06-23 23:50:03,198][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.04059799015522003, 'train/grad_norm_pre_clip': 19.921274185180664, 'timing/data': 0.0005076860543340445, 'timing/model': 1.1166808559792116, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 0.79}) +[2026-06-23 23:50:07,295][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.12211708724498749, 'train/grad_norm_pre_clip': 31.97447967529297, 'timing/data': 0.0005388699937611818, 'timing/model': 1.1178019620710984, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 0.79}) +[2026-06-23 23:50:11,378][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.08578993380069733, 'train/grad_norm_pre_clip': 17.341793060302734, 'timing/data': 0.0003621489740908146, 'timing/model': 1.1206435200292617, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 0.79}) +[2026-06-23 23:50:15,606][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.12665104866027832, 'train/grad_norm_pre_clip': 24.899415969848633, 'timing/data': 0.00035523506812751293, 'timing/model': 1.25993357389234, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 0.79}) +[2026-06-23 23:50:19,701][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.06347465515136719, 'train/grad_norm_pre_clip': 7.182992458343506, 'timing/data': 0.00035289302468299866, 'timing/model': 1.1231195979053155, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 0.79}) +[2026-06-23 23:50:23,785][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.07846683263778687, 'train/grad_norm_pre_clip': 17.509931564331055, 'timing/data': 0.00037674803752452135, 'timing/model': 1.1235088750254363, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 0.79}) +[2026-06-23 23:50:27,878][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.16667939722537994, 'train/grad_norm_pre_clip': 21.90150260925293, 'timing/data': 0.00035816500894725323, 'timing/model': 1.1152258159127086, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 0.79}) +[2026-06-23 23:50:31,949][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.19193723797798157, 'train/grad_norm_pre_clip': 30.283016204833984, 'timing/data': 0.0004075499018654227, 'timing/model': 1.1166997510008514, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 0.79}) +[2026-06-23 23:50:36,199][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.05372167006134987, 'train/grad_norm_pre_clip': 13.320154190063477, 'timing/data': 0.0003533110721036792, 'timing/model': 1.1179049389902502, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 0.79}) +[2026-06-23 23:50:40,271][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.15156379342079163, 'train/grad_norm_pre_clip': 25.134294509887695, 'timing/data': 0.00036128098145127296, 'timing/model': 1.1128246270818636, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.79}) +[2026-06-23 23:50:44,347][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.07332897186279297, 'train/grad_norm_pre_clip': 29.884532928466797, 'timing/data': 0.0003748390590772033, 'timing/model': 1.1163813089951873, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 0.79}) +[2026-06-23 23:50:48,426][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.05059976130723953, 'train/grad_norm_pre_clip': 11.623736381530762, 'timing/data': 0.00036549801006913185, 'timing/model': 1.1172796419123188, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 0.79}) +[2026-06-23 23:50:52,594][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.03867103159427643, 'train/grad_norm_pre_clip': 9.301589012145996, 'timing/data': 0.00036396796349436045, 'timing/model': 1.1161835209932178, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 0.79}) +[2026-06-23 23:50:56,670][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.044654831290245056, 'train/grad_norm_pre_clip': 17.6628475189209, 'timing/data': 0.00036451907362788916, 'timing/model': 1.1159710380015895, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 0.79}) +[2026-06-23 23:51:00,768][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.04814351722598076, 'train/grad_norm_pre_clip': 15.054364204406738, 'timing/data': 0.00035202992148697376, 'timing/model': 1.119567792979069, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 0.8}) +[2026-06-23 23:51:04,852][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.051331985741853714, 'train/grad_norm_pre_clip': 10.917691230773926, 'timing/data': 0.0003553749993443489, 'timing/model': 1.1218865149421617, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 0.8}) +[2026-06-23 23:51:08,998][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.030855339020490646, 'train/grad_norm_pre_clip': 11.693154335021973, 'timing/data': 0.00035877106711268425, 'timing/model': 1.1131235720822588, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 0.8}) +[2026-06-23 23:51:13,075][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.11058122664690018, 'train/grad_norm_pre_clip': 11.163331985473633, 'timing/data': 0.00036062905564904213, 'timing/model': 1.1193238840205595, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 0.8}) +[2026-06-23 23:51:17,164][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.03718144819140434, 'train/grad_norm_pre_clip': 12.554378509521484, 'timing/data': 0.0004083529347553849, 'timing/model': 1.1149152819998562, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 0.8}) +[2026-06-23 23:51:21,259][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.00910214800387621, 'train/grad_norm_pre_clip': 28.49660301208496, 'timing/data': 0.000397228985093534, 'timing/model': 1.1203570029465482, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.8}) +[2026-06-23 23:51:25,437][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.040259700268507004, 'train/grad_norm_pre_clip': 8.291983604431152, 'timing/data': 0.0006003719754517078, 'timing/model': 1.117631145985797, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 0.8}) +[2026-06-23 23:51:29,517][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.02817363105714321, 'train/grad_norm_pre_clip': 38.650211334228516, 'timing/data': 0.0005669889505952597, 'timing/model': 1.114674089010805, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 0.8}) +[2026-06-23 23:51:33,580][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.040078166872262955, 'train/grad_norm_pre_clip': 29.876611709594727, 'timing/data': 0.0003661330556496978, 'timing/model': 1.1136055720271543, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 0.8}) +[2026-06-23 23:51:37,657][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.10108589380979538, 'train/grad_norm_pre_clip': 17.986143112182617, 'timing/data': 0.00036620907485485077, 'timing/model': 1.1163480659015477, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 0.8}) +[2026-06-23 23:51:41,872][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.03374508395791054, 'train/grad_norm_pre_clip': 12.861553192138672, 'timing/data': 0.0003597900504246354, 'timing/model': 1.1170502939494327, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 0.8}) +[2026-06-23 23:51:45,958][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.12936432659626007, 'train/grad_norm_pre_clip': 15.663054466247559, 'timing/data': 0.0003613870358094573, 'timing/model': 1.1167837450047955, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 0.8}) +[2026-06-23 23:51:50,035][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.03197569027543068, 'train/grad_norm_pre_clip': 22.60576057434082, 'timing/data': 0.0004160859389230609, 'timing/model': 1.1168125050608069, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 0.8}) +[2026-06-23 23:51:54,207][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.0011863065883517265, 'train/grad_norm_pre_clip': 22.256567001342773, 'timing/data': 0.0004382220795378089, 'timing/model': 1.2118531190790236, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 0.8}) +[2026-06-23 23:51:58,306][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.04778547212481499, 'train/grad_norm_pre_clip': 39.27156066894531, 'timing/data': 0.0003565250663086772, 'timing/model': 1.127081885933876, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 0.8}) +[2026-06-23 23:52:02,396][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.008406204171478748, 'train/grad_norm_pre_clip': 13.15639591217041, 'timing/data': 0.000357851036824286, 'timing/model': 1.1262077660067007, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.8}) +[2026-06-23 23:52:06,484][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.19596846401691437, 'train/grad_norm_pre_clip': 42.04587173461914, 'timing/data': 0.0006173660513013601, 'timing/model': 1.1185493229422718, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 0.8}) +[2026-06-23 23:52:10,750][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.03296046704053879, 'train/grad_norm_pre_clip': 8.4777250289917, 'timing/data': 0.0003741559339687228, 'timing/model': 1.2949337100144476, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 0.8}) +[2026-06-23 23:52:14,828][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.17307329177856445, 'train/grad_norm_pre_clip': 16.945573806762695, 'timing/data': 0.0005925849545747042, 'timing/model': 1.122246632934548, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 0.8}) +[2026-06-23 23:52:18,902][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.0097205750644207, 'train/grad_norm_pre_clip': 9.355907440185547, 'timing/data': 0.0004273800877854228, 'timing/model': 1.1104935329640284, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 0.8}) +[2026-06-23 23:52:22,982][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.08538862317800522, 'train/grad_norm_pre_clip': 20.92513084411621, 'timing/data': 0.0003727290313690901, 'timing/model': 1.117504319990985, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 0.8}) +[2026-06-23 23:52:27,131][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.01830623485147953, 'train/grad_norm_pre_clip': 28.92987632751465, 'timing/data': 0.0004899919731542468, 'timing/model': 1.1138627670006827, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 0.8}) +[2026-06-23 23:52:31,209][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.1742601990699768, 'train/grad_norm_pre_clip': 42.04276657104492, 'timing/data': 0.00035881903022527695, 'timing/model': 1.1221158730331808, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 0.8}) +[2026-06-23 23:52:35,288][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.2259332537651062, 'train/grad_norm_pre_clip': 23.61094856262207, 'timing/data': 0.0006957320729270577, 'timing/model': 1.1171699590049684, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 0.8}) +[2026-06-23 23:52:39,376][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.08504519611597061, 'train/grad_norm_pre_clip': 31.598846435546875, 'timing/data': 0.0004410790279507637, 'timing/model': 1.1194599739974365, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 0.8}) +[2026-06-23 23:52:43,523][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.09862207621335983, 'train/grad_norm_pre_clip': 18.424428939819336, 'timing/data': 0.0006366840098053217, 'timing/model': 1.1185238540638238, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.8}) +[2026-06-23 23:52:47,592][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.16551843285560608, 'train/grad_norm_pre_clip': 28.686269760131836, 'timing/data': 0.00040588201954960823, 'timing/model': 1.1180833149701357, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 0.81}) +[2026-06-23 23:52:51,681][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.10098082572221756, 'train/grad_norm_pre_clip': 31.313812255859375, 'timing/data': 0.0003627950791269541, 'timing/model': 1.1196247029583901, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 0.81}) +[2026-06-23 23:52:55,752][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.009451411664485931, 'train/grad_norm_pre_clip': 9.904003143310547, 'timing/data': 0.0003723950358107686, 'timing/model': 1.1138537210645154, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 0.81}) +[2026-06-23 23:52:59,996][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.044145580381155014, 'train/grad_norm_pre_clip': 25.94855308532715, 'timing/data': 0.0003681499511003494, 'timing/model': 1.1266300659626722, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 0.81}) +[2026-06-23 23:53:04,083][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.024084564298391342, 'train/grad_norm_pre_clip': 17.549123764038086, 'timing/data': 0.0004479560302570462, 'timing/model': 1.122600435046479, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 0.81}) +[2026-06-23 23:53:08,166][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.05560972914099693, 'train/grad_norm_pre_clip': 14.877974510192871, 'timing/data': 0.0004310909425839782, 'timing/model': 1.1186408289941028, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 0.81}) +[2026-06-23 23:53:12,253][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.04544743150472641, 'train/grad_norm_pre_clip': 9.938704490661621, 'timing/data': 0.000497444998472929, 'timing/model': 1.1227043620310724, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 0.81}) +[2026-06-23 23:53:16,450][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.07487513870000839, 'train/grad_norm_pre_clip': 11.85090160369873, 'timing/data': 0.0003705619601532817, 'timing/model': 1.1150999699020758, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 0.81}) +[2026-06-23 23:53:20,591][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.07036878913640976, 'train/grad_norm_pre_clip': 28.56248664855957, 'timing/data': 0.00039175001438707113, 'timing/model': 1.1190322110196576, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 0.81}) +[2026-06-23 23:53:24,674][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.03473832830786705, 'train/grad_norm_pre_clip': 11.877132415771484, 'timing/data': 0.0005683520575985312, 'timing/model': 1.1136832159245387, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.81}) +[2026-06-23 23:53:28,755][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.03193284571170807, 'train/grad_norm_pre_clip': 16.078371047973633, 'timing/data': 0.0006389090558513999, 'timing/model': 1.1201306990114972, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 0.81}) +[2026-06-23 23:53:32,923][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.06094061955809593, 'train/grad_norm_pre_clip': 12.953554153442383, 'timing/data': 0.00041431793943047523, 'timing/model': 1.1188616548897699, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 0.81}) +[2026-06-23 23:53:37,001][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.04676593095064163, 'train/grad_norm_pre_clip': 6.800539493560791, 'timing/data': 0.0005040810210630298, 'timing/model': 1.1148479119874537, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 0.81}) +[2026-06-23 23:53:41,080][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.07982027530670166, 'train/grad_norm_pre_clip': 10.412236213684082, 'timing/data': 0.00035566010046750307, 'timing/model': 1.123515463899821, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 0.81}) +[2026-06-23 23:53:45,159][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.06270904093980789, 'train/grad_norm_pre_clip': 15.36064624786377, 'timing/data': 0.00042638008017092943, 'timing/model': 1.1132991310441867, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.81}) +[2026-06-23 23:53:49,411][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.027773646637797356, 'train/grad_norm_pre_clip': 19.663278579711914, 'timing/data': 0.0005263700149953365, 'timing/model': 1.1193860929924995, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 0.81}) +[2026-06-23 23:53:53,513][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.03367450833320618, 'train/grad_norm_pre_clip': 16.405210494995117, 'timing/data': 0.00036543491296470165, 'timing/model': 1.124453364056535, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 0.81}) +[2026-06-23 23:53:57,586][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.14634650945663452, 'train/grad_norm_pre_clip': 15.518682479858398, 'timing/data': 0.0003638289636000991, 'timing/model': 1.1172686540521681, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 0.81}) +[2026-06-23 23:54:01,691][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.004167534876614809, 'train/grad_norm_pre_clip': 9.976692199707031, 'timing/data': 0.0005347860278561711, 'timing/model': 1.1478897469351068, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 0.81}) +[2026-06-23 23:54:05,873][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2100 +[2026-06-23 23:55:10,225][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2100 +[2026-06-23 23:55:10,226][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2100 +[2026-06-23 23:55:49,489][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2100 +[2026-06-23 23:55:49,490][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.04460465535521507, 'train/grad_norm_pre_clip': 22.909870147705078, 'eval/action_loss/seconds': 64.34937722794712, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.04746502265334129, 'eval/latency_2/loss': 0.04746502265334129, 'eval/flappy/loss': 0.04746502265334129, 'eval/flappy/latency_2/loss': 0.04746502265334129, 'eval/flap_precision': 0.9830028328611898, 'eval/flap_recall': 0.9416553595658074, 'eval/flap_f1': 0.961884961884962, 'eval/flap_rate_model': 0.19611111111111112, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.961884961884962, 'eval/latency_2/flap_precision': 0.9830028328611898, 'eval/latency_2/flap_recall': 0.9416553595658074, 'eval/latency_2/flap_f1': 0.961884961884962, 'eval/latency_2/flap_rate_model': 0.19611111111111112, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.961884961884962, 'eval/action_classification/data_wait_seconds': 0.6494102478027344, 'eval/action_classification/predict_seconds': 38.467254638671875, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0003627139376476407, 'timing/model': 1.1182719350326806, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.81}) +[2026-06-23 23:55:53,531][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.01704317517578602, 'train/grad_norm_pre_clip': 41.02616882324219, 'timing/data': 0.00037299597170203924, 'timing/model': 1.1106016730191186, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 0.81}) +[2026-06-23 23:55:57,570][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.08068549633026123, 'train/grad_norm_pre_clip': 36.50276184082031, 'timing/data': 0.0005505218869075179, 'timing/model': 1.1080878620268777, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 0.81}) +[2026-06-23 23:56:01,697][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.04632310941815376, 'train/grad_norm_pre_clip': 13.096511840820312, 'timing/data': 0.00035795895382761955, 'timing/model': 1.1115897810086608, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 0.81}) +[2026-06-23 23:56:05,752][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.03497808799147606, 'train/grad_norm_pre_clip': 30.358713150024414, 'timing/data': 0.0005641649477183819, 'timing/model': 1.1102324270177633, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 0.81}) +[2026-06-23 23:56:09,851][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.010716954246163368, 'train/grad_norm_pre_clip': 8.916716575622559, 'timing/data': 0.0005868659354746342, 'timing/model': 1.1108680269680917, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 0.81}) +[2026-06-23 23:56:13,910][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.12210942804813385, 'train/grad_norm_pre_clip': 23.338092803955078, 'timing/data': 0.0003727119183167815, 'timing/model': 1.112181629985571, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 0.82}) +[2026-06-23 23:56:18,135][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.14576290547847748, 'train/grad_norm_pre_clip': 24.14995002746582, 'timing/data': 0.0003798099933192134, 'timing/model': 1.1197222099872306, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 0.82}) +[2026-06-23 23:56:22,197][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.04782280698418617, 'train/grad_norm_pre_clip': 29.936542510986328, 'timing/data': 0.0003599530318751931, 'timing/model': 1.112209461047314, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 0.82}) +[2026-06-23 23:56:26,258][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.046138275414705276, 'train/grad_norm_pre_clip': 26.65447235107422, 'timing/data': 0.0003557279706001282, 'timing/model': 1.1168027309468016, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.82}) +[2026-06-23 23:56:30,321][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.058366358280181885, 'train/grad_norm_pre_clip': 21.829444885253906, 'timing/data': 0.0003491709940135479, 'timing/model': 1.1144453630549833, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.82}) +[2026-06-23 23:56:34,506][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.0773545354604721, 'train/grad_norm_pre_clip': 20.15205955505371, 'timing/data': 0.00041461701039224863, 'timing/model': 1.1150186189915985, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 0.82}) +[2026-06-23 23:56:38,571][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.010232028551399708, 'train/grad_norm_pre_clip': 20.8566837310791, 'timing/data': 0.0003569909604266286, 'timing/model': 1.1107543610269204, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 0.82}) +[2026-06-23 23:56:42,639][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.09786143153905869, 'train/grad_norm_pre_clip': 22.1043643951416, 'timing/data': 0.00037081900518387556, 'timing/model': 1.1122476229211316, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 0.82}) +[2026-06-23 23:56:46,713][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.15500633418560028, 'train/grad_norm_pre_clip': 25.52642250061035, 'timing/data': 0.00044385192450135946, 'timing/model': 1.113952764077112, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 0.82}) +[2026-06-23 23:56:50,903][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.17570900917053223, 'train/grad_norm_pre_clip': 31.114402770996094, 'timing/data': 0.0003636750625446439, 'timing/model': 1.1153790929820389, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 0.82}) +[2026-06-23 23:56:54,965][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.005856722593307495, 'train/grad_norm_pre_clip': 4.793573379516602, 'timing/data': 0.0003625829704105854, 'timing/model': 1.1081120159942657, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 0.82}) +[2026-06-23 23:56:59,019][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.067881740629673, 'train/grad_norm_pre_clip': 26.26775550842285, 'timing/data': 0.0003670169971883297, 'timing/model': 1.1098794330609962, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 0.82}) +[2026-06-23 23:57:03,105][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.08085590600967407, 'train/grad_norm_pre_clip': 24.317575454711914, 'timing/data': 0.00036940001882612705, 'timing/model': 1.1208512340672314, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 0.82}) +[2026-06-23 23:57:07,296][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.1723909080028534, 'train/grad_norm_pre_clip': 24.444379806518555, 'timing/data': 0.0003620370989665389, 'timing/model': 1.1212013940094039, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 0.82}) +[2026-06-23 23:57:11,375][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.07337860763072968, 'train/grad_norm_pre_clip': 31.44379234313965, 'timing/data': 0.00043915899004787207, 'timing/model': 1.1178374210139737, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.82}) +[2026-06-23 23:57:15,457][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.12144898623228073, 'train/grad_norm_pre_clip': 15.785577774047852, 'timing/data': 0.0004029610427096486, 'timing/model': 1.121318644960411, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 0.82}) +[2026-06-23 23:57:19,535][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.044457968324422836, 'train/grad_norm_pre_clip': 13.509060859680176, 'timing/data': 0.0003587199607864022, 'timing/model': 1.1182319999206811, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 0.82}) +[2026-06-23 23:57:23,778][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.1436401754617691, 'train/grad_norm_pre_clip': 20.004161834716797, 'timing/data': 0.0003668220015242696, 'timing/model': 1.1144862150540575, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 0.82}) +[2026-06-23 23:57:27,856][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.14559830725193024, 'train/grad_norm_pre_clip': 18.524944305419922, 'timing/data': 0.0003590669948607683, 'timing/model': 1.1154284019721672, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 0.82}) +[2026-06-23 23:57:31,938][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.06080363318324089, 'train/grad_norm_pre_clip': 10.19937515258789, 'timing/data': 0.000546691007912159, 'timing/model': 1.1168472568970174, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 0.82}) +[2026-06-23 23:57:36,096][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.06803954392671585, 'train/grad_norm_pre_clip': 24.570539474487305, 'timing/data': 0.00038816104643046856, 'timing/model': 1.1816587620414793, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 0.82}) +[2026-06-23 23:57:40,177][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.05671270564198494, 'train/grad_norm_pre_clip': 23.850141525268555, 'timing/data': 0.0003757800441235304, 'timing/model': 1.1194307249970734, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 0.82}) +[2026-06-23 23:57:44,247][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.07201515138149261, 'train/grad_norm_pre_clip': 16.014982223510742, 'timing/data': 0.00037389202043414116, 'timing/model': 1.116856366978027, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 0.82}) +[2026-06-23 23:57:48,310][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.025265438482165337, 'train/grad_norm_pre_clip': 17.055580139160156, 'timing/data': 0.0003622579388320446, 'timing/model': 1.1197598479920998, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 0.82}) +[2026-06-23 23:57:52,494][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.06274847686290741, 'train/grad_norm_pre_clip': 7.542771816253662, 'timing/data': 0.00035448605194687843, 'timing/model': 1.2036849170690402, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.82}) +[2026-06-23 23:57:56,583][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.05348759517073631, 'train/grad_norm_pre_clip': 21.332653045654297, 'timing/data': 0.0003644030075520277, 'timing/model': 1.1214737030677497, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 0.82}) +[2026-06-23 23:58:00,674][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.13624270260334015, 'train/grad_norm_pre_clip': 12.9533109664917, 'timing/data': 0.00037226290442049503, 'timing/model': 1.1179901730502024, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 0.83}) +[2026-06-23 23:58:04,766][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.08653705567121506, 'train/grad_norm_pre_clip': 12.717988967895508, 'timing/data': 0.0005472599295899272, 'timing/model': 1.1173574270214885, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 0.83}) +[2026-06-23 23:58:08,983][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.03306161239743233, 'train/grad_norm_pre_clip': 17.133365631103516, 'timing/data': 0.00040819798596203327, 'timing/model': 1.2549357670359313, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 0.83}) +[2026-06-23 23:58:13,055][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.19062310457229614, 'train/grad_norm_pre_clip': 25.339065551757812, 'timing/data': 0.00043449702206999063, 'timing/model': 1.1190381520427763, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 0.83}) +[2026-06-23 23:58:17,130][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.04692256823182106, 'train/grad_norm_pre_clip': 12.82935619354248, 'timing/data': 0.0004209729377180338, 'timing/model': 1.1170982889598235, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 0.83}) +[2026-06-23 23:58:21,211][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.07990382611751556, 'train/grad_norm_pre_clip': 13.528403282165527, 'timing/data': 0.0005860300734639168, 'timing/model': 1.119462335947901, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 0.83}) +[2026-06-23 23:58:25,410][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.005654704291373491, 'train/grad_norm_pre_clip': 7.099123954772949, 'timing/data': 0.0005484790308400989, 'timing/model': 1.118159574107267, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 0.83}) +[2026-06-23 23:58:29,495][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.06627623736858368, 'train/grad_norm_pre_clip': 10.695780754089355, 'timing/data': 0.0003613209119066596, 'timing/model': 1.1225574070122093, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 0.83}) +[2026-06-23 23:58:33,581][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.02043386921286583, 'train/grad_norm_pre_clip': 10.926441192626953, 'timing/data': 0.00035892496816813946, 'timing/model': 1.1188891850179061, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.83}) +[2026-06-23 23:58:37,661][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.04346968233585358, 'train/grad_norm_pre_clip': 19.105283737182617, 'timing/data': 0.000381608959287405, 'timing/model': 1.1120473879855126, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 0.83}) +[2026-06-23 23:58:41,815][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.016688652336597443, 'train/grad_norm_pre_clip': 8.239680290222168, 'timing/data': 0.00037977006286382675, 'timing/model': 1.1154000089736655, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 0.83}) +[2026-06-23 23:58:45,908][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.03560793772339821, 'train/grad_norm_pre_clip': 29.943151473999023, 'timing/data': 0.00035663601011037827, 'timing/model': 1.1233260659500957, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 0.83}) +[2026-06-23 23:58:49,996][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.09976073354482651, 'train/grad_norm_pre_clip': 24.37032699584961, 'timing/data': 0.00040706503205001354, 'timing/model': 1.118991793016903, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 0.83}) +[2026-06-23 23:58:54,079][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.002811768325045705, 'train/grad_norm_pre_clip': 21.883272171020508, 'timing/data': 0.0003673010505735874, 'timing/model': 1.1179007749306038, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 0.83}) +[2026-06-23 23:58:58,323][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.062047798186540604, 'train/grad_norm_pre_clip': 34.30535125732422, 'timing/data': 0.00035666709300130606, 'timing/model': 1.1158920710440725, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 0.83}) +[2026-06-23 23:59:02,406][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.09083764255046844, 'train/grad_norm_pre_clip': 21.538745880126953, 'timing/data': 0.00035312504041939974, 'timing/model': 1.1175780920311809, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 0.83}) +[2026-06-23 23:59:06,485][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.014454122632741928, 'train/grad_norm_pre_clip': 14.722904205322266, 'timing/data': 0.0005130450008437037, 'timing/model': 1.1230437519261613, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 0.83}) +[2026-06-23 23:59:10,581][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.06863149255514145, 'train/grad_norm_pre_clip': 14.53812026977539, 'timing/data': 0.00044779200106859207, 'timing/model': 1.1212980339769274, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 0.83}) +[2026-06-23 23:59:14,730][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.004627343267202377, 'train/grad_norm_pre_clip': 16.33635902404785, 'timing/data': 0.0003583589568734169, 'timing/model': 1.1149529221002012, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.83}) +[2026-06-23 23:59:18,815][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.012319306842982769, 'train/grad_norm_pre_clip': 10.403440475463867, 'timing/data': 0.0006152170244604349, 'timing/model': 1.12865613901522, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 0.83}) +[2026-06-23 23:59:22,924][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.09600629657506943, 'train/grad_norm_pre_clip': 52.870521545410156, 'timing/data': 0.00037996808532625437, 'timing/model': 1.141839658957906, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 0.83}) +[2026-06-23 23:59:27,032][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.11316504329442978, 'train/grad_norm_pre_clip': 40.35440444946289, 'timing/data': 0.00037028000224381685, 'timing/model': 1.117902601021342, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 0.83}) +[2026-06-23 23:59:31,237][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.0007654227665625513, 'train/grad_norm_pre_clip': 12.469703674316406, 'timing/data': 0.0006159819895401597, 'timing/model': 1.2383868909673765, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 0.83}) +[2026-06-23 23:59:35,325][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.010436340235173702, 'train/grad_norm_pre_clip': 18.75841522216797, 'timing/data': 0.0004026100505143404, 'timing/model': 1.1268602790078148, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 0.83}) +[2026-06-23 23:59:39,411][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.012214526534080505, 'train/grad_norm_pre_clip': 18.528345108032227, 'timing/data': 0.0003624800592660904, 'timing/model': 1.1197853119811043, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 0.83}) +[2026-06-23 23:59:43,498][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.09100337326526642, 'train/grad_norm_pre_clip': 23.089078903198242, 'timing/data': 0.0005212660180404782, 'timing/model': 1.1151795589830726, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 0.83}) +[2026-06-23 23:59:47,576][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.030605072155594826, 'train/grad_norm_pre_clip': 22.0626220703125, 'timing/data': 0.0003564830403774977, 'timing/model': 1.1164870550855994, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.84}) +[2026-06-23 23:59:51,817][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.05358244106173515, 'train/grad_norm_pre_clip': 27.248886108398438, 'timing/data': 0.00035889900755137205, 'timing/model': 1.1182709349086508, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 0.84}) +[2026-06-23 23:59:55,892][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.012262528762221336, 'train/grad_norm_pre_clip': 13.990102767944336, 'timing/data': 0.0003599909832701087, 'timing/model': 1.116695391945541, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.84}) +[2026-06-23 23:59:59,969][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.043895915150642395, 'train/grad_norm_pre_clip': 15.475247383117676, 'timing/data': 0.00035556009970605373, 'timing/model': 1.1176221939967945, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 0.84}) +[2026-06-24 00:00:04,065][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.08776956051588058, 'train/grad_norm_pre_clip': 15.928848266601562, 'timing/data': 0.0006075209239497781, 'timing/model': 1.1214830069802701, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 0.84}) +[2026-06-24 00:00:08,201][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.03441496193408966, 'train/grad_norm_pre_clip': 10.821868896484375, 'timing/data': 0.0004928299458697438, 'timing/model': 1.1161262330133468, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 0.84}) +[2026-06-24 00:00:12,280][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.010155417956411839, 'train/grad_norm_pre_clip': 19.446504592895508, 'timing/data': 0.0003959729801863432, 'timing/model': 1.117541051004082, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 0.84}) +[2026-06-24 00:00:16,360][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.0461047999560833, 'train/grad_norm_pre_clip': 27.001203536987305, 'timing/data': 0.00038673391100019217, 'timing/model': 1.1169769979314879, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.84}) +[2026-06-24 00:00:20,467][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.04435257613658905, 'train/grad_norm_pre_clip': 21.643375396728516, 'timing/data': 0.0003717039944604039, 'timing/model': 1.1318758289562538, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 0.84}) +[2026-06-24 00:00:24,623][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.19707490503787994, 'train/grad_norm_pre_clip': 18.463773727416992, 'timing/data': 0.0003524509957060218, 'timing/model': 1.1216794929932803, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 0.84}) +[2026-06-24 00:00:28,710][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.03331119567155838, 'train/grad_norm_pre_clip': 15.601195335388184, 'timing/data': 0.00035788805689662695, 'timing/model': 1.1209140879800543, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 0.84}) +[2026-06-24 00:00:32,791][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.04570801928639412, 'train/grad_norm_pre_clip': 17.60089683532715, 'timing/data': 0.0003724420676007867, 'timing/model': 1.1205938559724018, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 0.84}) +[2026-06-24 00:00:36,872][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.051784999668598175, 'train/grad_norm_pre_clip': 15.047981262207031, 'timing/data': 0.0003846969921141863, 'timing/model': 1.1185718259075657, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.84}) +[2026-06-24 00:00:41,020][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.02862381935119629, 'train/grad_norm_pre_clip': 14.61699104309082, 'timing/data': 0.0003932280233129859, 'timing/model': 1.1198169259587303, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 0.84}) +[2026-06-24 00:00:45,100][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.01759214513003826, 'train/grad_norm_pre_clip': 12.032021522521973, 'timing/data': 0.00036020507104694843, 'timing/model': 1.1144375819712877, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 0.84}) +[2026-06-24 00:00:49,170][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.06921610236167908, 'train/grad_norm_pre_clip': 39.61039733886719, 'timing/data': 0.00035971496254205704, 'timing/model': 1.1145778050413355, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 0.84}) +[2026-06-24 00:00:53,252][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.040724121034145355, 'train/grad_norm_pre_clip': 22.135948181152344, 'timing/data': 0.00036558404099196196, 'timing/model': 1.1175714009441435, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 0.84}) +[2026-06-24 00:00:57,471][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.21711131930351257, 'train/grad_norm_pre_clip': 26.005233764648438, 'timing/data': 0.0004136189818382263, 'timing/model': 1.1156049120472744, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 0.84}) +[2026-06-24 00:01:01,561][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.08142535388469696, 'train/grad_norm_pre_clip': 24.551584243774414, 'timing/data': 0.00051265605725348, 'timing/model': 1.1191571749513969, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 0.84}) +[2026-06-24 00:01:05,637][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.09081602841615677, 'train/grad_norm_pre_clip': 10.696577072143555, 'timing/data': 0.00036027596797794104, 'timing/model': 1.119081255979836, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 0.84}) +[2026-06-24 00:01:09,812][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.07374976575374603, 'train/grad_norm_pre_clip': 12.349163055419922, 'timing/data': 0.0005620010197162628, 'timing/model': 1.2189346998929977, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 0.84}) +[2026-06-24 00:01:13,895][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.11479589343070984, 'train/grad_norm_pre_clip': 29.303844451904297, 'timing/data': 0.0004208610625937581, 'timing/model': 1.1175194589886814, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 0.84}) +[2026-06-24 00:01:17,983][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.09469844400882721, 'train/grad_norm_pre_clip': 19.047481536865234, 'timing/data': 0.0003567900275811553, 'timing/model': 1.1208918340271339, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.84}) +[2026-06-24 00:01:22,055][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.07943145185709, 'train/grad_norm_pre_clip': 19.096111297607422, 'timing/data': 0.00037248292937874794, 'timing/model': 1.113540651043877, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 0.84}) +[2026-06-24 00:01:26,293][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.008403284475207329, 'train/grad_norm_pre_clip': 7.821732044219971, 'timing/data': 0.0005412918981164694, 'timing/model': 1.278075703070499, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 0.84}) +[2026-06-24 00:01:30,394][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.02073827013373375, 'train/grad_norm_pre_clip': 26.81435203552246, 'timing/data': 0.00035983603447675705, 'timing/model': 1.1175449609290808, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 0.84}) +[2026-06-24 00:01:34,494][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.08630216866731644, 'train/grad_norm_pre_clip': 22.80323028564453, 'timing/data': 0.000359712983481586, 'timing/model': 1.128344565979205, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 0.85}) +[2026-06-24 00:01:38,578][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.00022715024533681571, 'train/grad_norm_pre_clip': 4.051970481872559, 'timing/data': 0.0005456359358504415, 'timing/model': 1.1204439320135862, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 0.85}) +[2026-06-24 00:01:42,727][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.045756302773952484, 'train/grad_norm_pre_clip': 26.610633850097656, 'timing/data': 0.00037767202593386173, 'timing/model': 1.1204865230247378, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 0.85}) +[2026-06-24 00:01:46,791][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.017841041088104248, 'train/grad_norm_pre_clip': 54.539268493652344, 'timing/data': 0.00042958406265825033, 'timing/model': 1.117244228022173, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 0.85}) +[2026-06-24 00:01:50,864][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.23817825317382812, 'train/grad_norm_pre_clip': 59.71589660644531, 'timing/data': 0.0003860039869323373, 'timing/model': 1.1181045681005344, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 0.85}) +[2026-06-24 00:01:54,948][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.01963363215327263, 'train/grad_norm_pre_clip': 20.719402313232422, 'timing/data': 0.0005974380765110254, 'timing/model': 1.116723719984293, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 0.85}) +[2026-06-24 00:01:59,105][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.018777644261717796, 'train/grad_norm_pre_clip': 11.438153266906738, 'timing/data': 0.0005462790140882134, 'timing/model': 1.1156702189473435, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.85}) +[2026-06-24 00:02:03,182][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.011165606789290905, 'train/grad_norm_pre_clip': 13.629314422607422, 'timing/data': 0.000364279025234282, 'timing/model': 1.1134147399570793, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 0.85}) +[2026-06-24 00:02:07,311][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.024977345019578934, 'train/grad_norm_pre_clip': 17.86803436279297, 'timing/data': 0.0005216199206188321, 'timing/model': 1.1408616000553593, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 0.85}) +[2026-06-24 00:02:11,388][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.011624917387962341, 'train/grad_norm_pre_clip': 7.645961284637451, 'timing/data': 0.0004566510906443, 'timing/model': 1.116886842995882, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 0.85}) +[2026-06-24 00:02:15,614][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.2971418499946594, 'train/grad_norm_pre_clip': 47.56477737426758, 'timing/data': 0.0003669379511848092, 'timing/model': 1.1173479799181223, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 0.85}) +[2026-06-24 00:02:19,687][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.020996931940317154, 'train/grad_norm_pre_clip': 22.06863784790039, 'timing/data': 0.00036364910192787647, 'timing/model': 1.1168982409872115, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 0.85}) +[2026-06-24 00:02:23,759][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.06340175122022629, 'train/grad_norm_pre_clip': 31.449462890625, 'timing/data': 0.00035597593523561954, 'timing/model': 1.1163744089426473, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 0.85}) +[2026-06-24 00:02:27,838][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.0249547827988863, 'train/grad_norm_pre_clip': 18.969791412353516, 'timing/data': 0.0003562360070645809, 'timing/model': 1.1177862570621073, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 0.85}) +[2026-06-24 00:02:32,033][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.02176911011338234, 'train/grad_norm_pre_clip': 31.993318557739258, 'timing/data': 0.00036329205613583326, 'timing/model': 1.1175691490061581, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 0.85}) +[2026-06-24 00:02:36,116][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.032733045518398285, 'train/grad_norm_pre_clip': 16.263648986816406, 'timing/data': 0.0003567070234566927, 'timing/model': 1.119693397078663, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 0.85}) +[2026-06-24 00:02:40,208][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2200 +[2026-06-24 00:03:44,437][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2200 +[2026-06-24 00:03:44,438][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2200 +[2026-06-24 00:04:23,201][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2200 +[2026-06-24 00:04:23,202][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.024127092212438583, 'train/grad_norm_pre_clip': 22.389123916625977, 'eval/action_loss/seconds': 64.22563804802485, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.040863942354917526, 'eval/latency_2/loss': 0.040863942354917526, 'eval/flappy/loss': 0.040863942354917526, 'eval/flappy/latency_2/loss': 0.040863942354917526, 'eval/flap_precision': 0.9571619812583668, 'eval/flap_recall': 0.9701492537313433, 'eval/flap_f1': 0.963611859838275, 'eval/flap_rate_model': 0.2075, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9636118598382749, 'eval/latency_2/flap_precision': 0.9571619812583668, 'eval/latency_2/flap_recall': 0.9701492537313433, 'eval/latency_2/flap_f1': 0.963611859838275, 'eval/latency_2/flap_rate_model': 0.2075, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9636118598382749, 'eval/action_classification/data_wait_seconds': 0.512933611869812, 'eval/action_classification/predict_seconds': 38.08607482910156, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0003577430034056306, 'timing/model': 1.1261143230367452, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.85}) +[2026-06-24 00:04:27,401][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.0001908511039800942, 'train/grad_norm_pre_clip': 23.35528564453125, 'timing/data': 0.0005704399663954973, 'timing/model': 1.1049538479419425, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 0.85}) +[2026-06-24 00:04:31,432][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.08891287446022034, 'train/grad_norm_pre_clip': 22.257205963134766, 'timing/data': 0.0003737880615517497, 'timing/model': 1.102851960924454, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 0.85}) +[2026-06-24 00:04:35,473][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.02901449427008629, 'train/grad_norm_pre_clip': 26.666234970092773, 'timing/data': 0.00036570499651134014, 'timing/model': 1.1076869230018929, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 0.85}) +[2026-06-24 00:04:39,516][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.020597519353032112, 'train/grad_norm_pre_clip': 17.270648956298828, 'timing/data': 0.0003691919846460223, 'timing/model': 1.115566558088176, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 0.85}) +[2026-06-24 00:04:43,641][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.15038733184337616, 'train/grad_norm_pre_clip': 20.217924118041992, 'timing/data': 0.00038958608638495207, 'timing/model': 1.1134452470578253, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 0.85}) +[2026-06-24 00:04:47,699][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.18167512118816376, 'train/grad_norm_pre_clip': 26.833402633666992, 'timing/data': 0.0003845819737762213, 'timing/model': 1.1078109729569405, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 0.85}) +[2026-06-24 00:04:51,741][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.04226437211036682, 'train/grad_norm_pre_clip': 9.56397819519043, 'timing/data': 0.00039026804734021425, 'timing/model': 1.105580459930934, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.85}) +[2026-06-24 00:04:55,797][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.029350729659199715, 'train/grad_norm_pre_clip': 16.320533752441406, 'timing/data': 0.0003962509799748659, 'timing/model': 1.107696773018688, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 0.85}) +[2026-06-24 00:04:59,981][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.020201558247208595, 'train/grad_norm_pre_clip': 6.726449489593506, 'timing/data': 0.00037228409200906754, 'timing/model': 1.235426697996445, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 0.85}) +[2026-06-24 00:05:04,059][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.06745970249176025, 'train/grad_norm_pre_clip': 16.087873458862305, 'timing/data': 0.0004347990034148097, 'timing/model': 1.1106904059415683, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.86}) +[2026-06-24 00:05:08,120][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.005584646947681904, 'train/grad_norm_pre_clip': 23.516183853149414, 'timing/data': 0.00042187399230897427, 'timing/model': 1.112869129050523, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 0.86}) +[2026-06-24 00:05:12,181][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.022728463634848595, 'train/grad_norm_pre_clip': 9.077463150024414, 'timing/data': 0.0003540640464052558, 'timing/model': 1.1109558059833944, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 0.86}) +[2026-06-24 00:05:16,375][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.06259894371032715, 'train/grad_norm_pre_clip': 13.238089561462402, 'timing/data': 0.00042507401667535305, 'timing/model': 1.1115759080275893, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 0.86}) +[2026-06-24 00:05:20,434][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.0668708011507988, 'train/grad_norm_pre_clip': 10.210054397583008, 'timing/data': 0.00036383408587425947, 'timing/model': 1.11315193597693, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 0.86}) +[2026-06-24 00:05:24,512][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.004384730011224747, 'train/grad_norm_pre_clip': 14.680005073547363, 'timing/data': 0.0003689100267365575, 'timing/model': 1.1201871500816196, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 0.86}) +[2026-06-24 00:05:28,592][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.09163251519203186, 'train/grad_norm_pre_clip': 23.251388549804688, 'timing/data': 0.00036340905353426933, 'timing/model': 1.1152336230734363, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 0.86}) +[2026-06-24 00:05:32,841][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.07760743051767349, 'train/grad_norm_pre_clip': 22.03514289855957, 'timing/data': 0.0006822490831837058, 'timing/model': 1.1127181550255045, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 0.86}) +[2026-06-24 00:05:36,909][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.35430634021759033, 'train/grad_norm_pre_clip': 23.217330932617188, 'timing/data': 0.00037464790511876345, 'timing/model': 1.1168287560576573, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 0.86}) +[2026-06-24 00:05:40,986][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.13444404304027557, 'train/grad_norm_pre_clip': 41.359859466552734, 'timing/data': 0.00037288398016244173, 'timing/model': 1.115486558061093, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 0.86}) +[2026-06-24 00:05:45,064][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.06964869052171707, 'train/grad_norm_pre_clip': 23.473411560058594, 'timing/data': 0.0003622389631345868, 'timing/model': 1.1188055750681087, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.86}) +[2026-06-24 00:05:49,316][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.010461739264428616, 'train/grad_norm_pre_clip': 12.054106712341309, 'timing/data': 0.0003591650165617466, 'timing/model': 1.1140315419761464, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 0.86}) +[2026-06-24 00:05:53,398][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.0305252093821764, 'train/grad_norm_pre_clip': 20.0397891998291, 'timing/data': 0.00035520107485353947, 'timing/model': 1.1191305309766904, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 0.86}) +[2026-06-24 00:05:57,485][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.009148279204964638, 'train/grad_norm_pre_clip': 23.57965660095215, 'timing/data': 0.00037759903352707624, 'timing/model': 1.1176416970556602, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 0.86}) +[2026-06-24 00:06:01,558][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.08861195296049118, 'train/grad_norm_pre_clip': 21.629199981689453, 'timing/data': 0.0004109250148758292, 'timing/model': 1.113772033015266, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 0.86}) +[2026-06-24 00:06:05,715][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.0040232776664197445, 'train/grad_norm_pre_clip': 13.156747817993164, 'timing/data': 0.00035971496254205704, 'timing/model': 1.1219436540268362, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 0.86}) +[2026-06-24 00:06:09,779][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.09933746606111526, 'train/grad_norm_pre_clip': 42.395992279052734, 'timing/data': 0.00037900498136878014, 'timing/model': 1.1110363990301266, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 0.86}) +[2026-06-24 00:06:13,863][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.010802188888192177, 'train/grad_norm_pre_clip': 10.726447105407715, 'timing/data': 0.00036949501372873783, 'timing/model': 1.117357361014001, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 0.86}) +[2026-06-24 00:06:17,943][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.04520312696695328, 'train/grad_norm_pre_clip': 18.566347122192383, 'timing/data': 0.00043845700565725565, 'timing/model': 1.115830031922087, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 0.86}) +[2026-06-24 00:06:22,086][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.07486703246831894, 'train/grad_norm_pre_clip': 15.187666893005371, 'timing/data': 0.000369398039765656, 'timing/model': 1.1164906889898703, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 0.86}) +[2026-06-24 00:06:26,166][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.09968554973602295, 'train/grad_norm_pre_clip': 28.8748779296875, 'timing/data': 0.0004952340386807919, 'timing/model': 1.1172126670135185, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.86}) +[2026-06-24 00:06:30,235][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.005145659204572439, 'train/grad_norm_pre_clip': 18.912933349609375, 'timing/data': 0.0005277430173009634, 'timing/model': 1.1131437229923904, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 0.86}) +[2026-06-24 00:06:34,318][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.13909807801246643, 'train/grad_norm_pre_clip': 23.31987953186035, 'timing/data': 0.0003534560091793537, 'timing/model': 1.1169658459257334, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 0.86}) +[2026-06-24 00:06:38,539][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.005014652386307716, 'train/grad_norm_pre_clip': 28.6543025970459, 'timing/data': 0.01612465304788202, 'timing/model': 1.1216548350639641, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 0.86}) +[2026-06-24 00:06:42,607][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.011121677234768867, 'train/grad_norm_pre_clip': 12.241059303283691, 'timing/data': 0.0004662929568439722, 'timing/model': 1.117684875964187, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 0.86}) +[2026-06-24 00:06:46,685][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.03965311497449875, 'train/grad_norm_pre_clip': 19.50518226623535, 'timing/data': 0.0003613079898059368, 'timing/model': 1.1146286339499056, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.86}) +[2026-06-24 00:06:50,850][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.19089555740356445, 'train/grad_norm_pre_clip': 18.329727172851562, 'timing/data': 0.0003610749263316393, 'timing/model': 1.2065387780312449, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 0.87}) +[2026-06-24 00:06:54,929][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.190940260887146, 'train/grad_norm_pre_clip': 25.676267623901367, 'timing/data': 0.0005161849549040198, 'timing/model': 1.1196911688894033, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 0.87}) +[2026-06-24 00:06:59,011][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.026640206575393677, 'train/grad_norm_pre_clip': 17.998849868774414, 'timing/data': 0.00039012404158711433, 'timing/model': 1.1218056390061975, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 0.87}) +[2026-06-24 00:07:03,109][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.19165465235710144, 'train/grad_norm_pre_clip': 14.226378440856934, 'timing/data': 0.0003689229488372803, 'timing/model': 1.1154247249942273, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 0.87}) +[2026-06-24 00:07:07,368][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.013452237471938133, 'train/grad_norm_pre_clip': 17.079349517822266, 'timing/data': 0.0003634849563241005, 'timing/model': 1.291300688055344, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.87}) +[2026-06-24 00:07:11,461][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.05374746397137642, 'train/grad_norm_pre_clip': 18.056262969970703, 'timing/data': 0.00040975294541567564, 'timing/model': 1.1144106859574094, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 0.87}) +[2026-06-24 00:07:15,538][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.07496566325426102, 'train/grad_norm_pre_clip': 10.833368301391602, 'timing/data': 0.0003923169570043683, 'timing/model': 1.1152239519869909, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 0.87}) +[2026-06-24 00:07:19,618][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.11404260993003845, 'train/grad_norm_pre_clip': 15.835061073303223, 'timing/data': 0.00036534201353788376, 'timing/model': 1.1184667779598385, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 0.87}) +[2026-06-24 00:07:23,771][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.030389724299311638, 'train/grad_norm_pre_clip': 18.297151565551758, 'timing/data': 0.000672138063237071, 'timing/model': 1.1170033609960228, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 0.87}) +[2026-06-24 00:07:27,837][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.02232442796230316, 'train/grad_norm_pre_clip': 18.745874404907227, 'timing/data': 0.00036051205825060606, 'timing/model': 1.1158868569182232, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 0.87}) +[2026-06-24 00:07:31,916][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.008605960756540298, 'train/grad_norm_pre_clip': 12.275444030761719, 'timing/data': 0.00041531200986355543, 'timing/model': 1.113859535078518, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 0.87}) +[2026-06-24 00:07:35,993][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.01172939594835043, 'train/grad_norm_pre_clip': 7.521949291229248, 'timing/data': 0.0005044189747422934, 'timing/model': 1.1167067569913343, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 0.87}) +[2026-06-24 00:07:40,155][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.13279519975185394, 'train/grad_norm_pre_clip': 25.003475189208984, 'timing/data': 0.0005586940096691251, 'timing/model': 1.1192223880207166, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 0.87}) +[2026-06-24 00:07:44,240][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.0706755667924881, 'train/grad_norm_pre_clip': 18.453535079956055, 'timing/data': 0.0003659010399132967, 'timing/model': 1.1231451430357993, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 0.87}) +[2026-06-24 00:19:46,125][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2250 +[2026-06-24 00:19:46,126][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.03335427865386009, 'train/grad_norm_pre_clip': 18.56218147277832, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00046825502067804337, 'timing/model': 1.118663166067563, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.87}) +[2026-06-24 00:19:50,279][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.03971916809678078, 'train/grad_norm_pre_clip': 9.949982643127441, 'timing/data': 0.0005577959818765521, 'timing/model': 1.2786728659411892, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 0.87}) +[2026-06-24 00:19:54,267][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.09932506829500198, 'train/grad_norm_pre_clip': 45.572566986083984, 'timing/data': 0.00037006800994277, 'timing/model': 1.093875790014863, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 0.87}) +[2026-06-24 00:19:58,270][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.038572948426008224, 'train/grad_norm_pre_clip': 35.708900451660156, 'timing/data': 0.00037551193963736296, 'timing/model': 1.1053995679831132, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 0.87}) +[2026-06-24 00:20:02,270][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.11942234635353088, 'train/grad_norm_pre_clip': 18.69135093688965, 'timing/data': 0.00038370105903595686, 'timing/model': 1.0942264350596815, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 0.87}) +[2026-06-24 00:20:06,458][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.06790687143802643, 'train/grad_norm_pre_clip': 12.069197654724121, 'timing/data': 0.00047942192759364843, 'timing/model': 1.2790034929057583, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 0.87}) +[2026-06-24 00:20:10,467][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.050853632390499115, 'train/grad_norm_pre_clip': 8.979759216308594, 'timing/data': 0.0005413329927250743, 'timing/model': 1.098975314060226, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.87}) +[2026-06-24 00:20:14,491][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.0006265522097237408, 'train/grad_norm_pre_clip': 17.264339447021484, 'timing/data': 0.0003565800143405795, 'timing/model': 1.1053486129967496, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 0.87}) +[2026-06-24 00:20:18,525][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.0744497999548912, 'train/grad_norm_pre_clip': 14.975801467895508, 'timing/data': 0.00038330606184899807, 'timing/model': 1.1050833530025557, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 0.87}) +[2026-06-24 00:20:22,731][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.027742533013224602, 'train/grad_norm_pre_clip': 14.13391399383545, 'timing/data': 0.000361097976565361, 'timing/model': 1.2790891589829698, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 0.87}) +[2026-06-24 00:20:26,770][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.02419228106737137, 'train/grad_norm_pre_clip': 28.163488388061523, 'timing/data': 0.0003569020191207528, 'timing/model': 1.1107811779947951, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.87}) +[2026-06-24 00:20:30,803][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.0068258182145655155, 'train/grad_norm_pre_clip': 53.17850875854492, 'timing/data': 0.000471967039629817, 'timing/model': 1.1046592670027167, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 0.88}) +[2026-06-24 00:20:34,855][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.1280534267425537, 'train/grad_norm_pre_clip': 15.321423530578613, 'timing/data': 0.00036167295183986425, 'timing/model': 1.1153528649592772, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 0.88}) +[2026-06-24 00:20:39,081][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.17074517905712128, 'train/grad_norm_pre_clip': 29.13004493713379, 'timing/data': 0.00038543599657714367, 'timing/model': 1.2828785580350086, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 0.88}) +[2026-06-24 00:20:43,133][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.012928491458296776, 'train/grad_norm_pre_clip': 26.472976684570312, 'timing/data': 0.00035757594741880894, 'timing/model': 1.1065244659548625, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 0.88}) +[2026-06-24 00:20:47,186][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.04683947563171387, 'train/grad_norm_pre_clip': 24.815895080566406, 'timing/data': 0.0003643139498308301, 'timing/model': 1.1176890569040552, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 0.88}) +[2026-06-24 00:20:51,221][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.030133865773677826, 'train/grad_norm_pre_clip': 17.053070068359375, 'timing/data': 0.0004429910331964493, 'timing/model': 1.1034443769603968, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 0.88}) +[2026-06-24 00:20:55,469][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.010497186332941055, 'train/grad_norm_pre_clip': 31.0350284576416, 'timing/data': 0.00035397999454289675, 'timing/model': 1.2998639619909227, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 0.88}) +[2026-06-24 00:20:59,535][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.015562127344310284, 'train/grad_norm_pre_clip': 17.65563201904297, 'timing/data': 0.00040871300734579563, 'timing/model': 1.1132183698937297, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 0.88}) +[2026-06-24 00:21:03,584][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.09178213775157928, 'train/grad_norm_pre_clip': 16.275163650512695, 'timing/data': 0.00036955007817596197, 'timing/model': 1.106556665035896, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 0.88}) +[2026-06-24 00:21:07,638][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.0523335300385952, 'train/grad_norm_pre_clip': 29.994003295898438, 'timing/data': 0.0003639340866357088, 'timing/model': 1.1156896649627015, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.88}) +[2026-06-24 00:21:11,871][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.1540290266275406, 'train/grad_norm_pre_clip': 27.101409912109375, 'timing/data': 0.00039680395275354385, 'timing/model': 1.2735265149967745, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 0.88}) +[2026-06-24 00:21:15,930][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.005313705187290907, 'train/grad_norm_pre_clip': 4.771474838256836, 'timing/data': 0.00036209600511938334, 'timing/model': 1.1191229050746188, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 0.88}) +[2026-06-24 00:21:20,004][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.008381899446249008, 'train/grad_norm_pre_clip': 29.834362030029297, 'timing/data': 0.00039379496593028307, 'timing/model': 1.1171952459262684, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 0.88}) +[2026-06-24 00:21:24,077][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.024697542190551758, 'train/grad_norm_pre_clip': 19.449373245239258, 'timing/data': 0.0003821030259132385, 'timing/model': 1.1129876390332356, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 0.88}) +[2026-06-24 00:21:28,325][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.048655956983566284, 'train/grad_norm_pre_clip': 20.312969207763672, 'timing/data': 0.000410759006626904, 'timing/model': 1.2973276219563559, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 0.88}) +[2026-06-24 00:21:32,406][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.018782280385494232, 'train/grad_norm_pre_clip': 11.630515098571777, 'timing/data': 0.0003494609845802188, 'timing/model': 1.1130671019200236, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 0.88}) +[2026-06-24 00:21:36,476][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.13460473716259003, 'train/grad_norm_pre_clip': 27.87285041809082, 'timing/data': 0.00035456009209156036, 'timing/model': 1.1188533800886944, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 0.88}) +[2026-06-24 00:21:40,552][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.12211930751800537, 'train/grad_norm_pre_clip': 17.572336196899414, 'timing/data': 0.0003608450060710311, 'timing/model': 1.1154999409336597, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 0.88}) +[2026-06-24 00:21:44,823][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.015892187133431435, 'train/grad_norm_pre_clip': 20.52035140991211, 'timing/data': 0.0007015129085630178, 'timing/model': 1.2966060009784997, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 0.88}) +[2026-06-24 00:21:48,900][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.010419058613479137, 'train/grad_norm_pre_clip': 13.601082801818848, 'timing/data': 0.0003629400162026286, 'timing/model': 1.1143014789558947, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.88}) +[2026-06-24 00:21:52,957][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.004777930676937103, 'train/grad_norm_pre_clip': 11.240260124206543, 'timing/data': 0.0004418249009177089, 'timing/model': 1.1074343450600281, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 0.88}) +[2026-06-24 00:21:57,028][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.08806821703910828, 'train/grad_norm_pre_clip': 13.966313362121582, 'timing/data': 0.00036398705560714006, 'timing/model': 1.115538576967083, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 0.88}) +[2026-06-24 00:22:01,271][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.03091863915324211, 'train/grad_norm_pre_clip': 16.28265953063965, 'timing/data': 0.00036259694024920464, 'timing/model': 1.2920362889999524, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 0.88}) +[2026-06-24 00:22:05,331][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.28512686491012573, 'train/grad_norm_pre_clip': 26.314434051513672, 'timing/data': 0.00035749899689108133, 'timing/model': 1.1154502850258723, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 0.88}) +[2026-06-24 00:22:09,412][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.0422893762588501, 'train/grad_norm_pre_clip': 54.263877868652344, 'timing/data': 0.0003603260265663266, 'timing/model': 1.1087606929941103, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 0.88}) +[2026-06-24 00:22:13,491][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.050223808735609055, 'train/grad_norm_pre_clip': 16.137710571289062, 'timing/data': 0.00035773299168795347, 'timing/model': 1.1214947029948235, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 0.88}) +[2026-06-24 00:22:17,754][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.08610483258962631, 'train/grad_norm_pre_clip': 13.8500394821167, 'timing/data': 0.0005304550286382437, 'timing/model': 1.3059762719785795, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 0.89}) +[2026-06-24 00:22:21,827][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.007069080136716366, 'train/grad_norm_pre_clip': 15.07836627960205, 'timing/data': 0.000512728001922369, 'timing/model': 1.118906326009892, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 0.89}) +[2026-06-24 00:22:25,907][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.056076448410749435, 'train/grad_norm_pre_clip': 18.028820037841797, 'timing/data': 0.0003507839282974601, 'timing/model': 1.1187746829818934, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 0.89}) +[2026-06-24 00:22:29,983][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.007067068479955196, 'train/grad_norm_pre_clip': 10.373858451843262, 'timing/data': 0.00035383901558816433, 'timing/model': 1.113566213985905, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.89}) +[2026-06-24 00:22:34,227][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.08032315969467163, 'train/grad_norm_pre_clip': 26.806194305419922, 'timing/data': 0.00047074095346033573, 'timing/model': 1.285948875010945, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 0.89}) +[2026-06-24 00:22:38,300][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.020330378785729408, 'train/grad_norm_pre_clip': 18.60496711730957, 'timing/data': 0.0007759289583191276, 'timing/model': 1.1199140929384157, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 0.89}) +[2026-06-24 00:22:42,380][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.05370993912220001, 'train/grad_norm_pre_clip': 8.400951385498047, 'timing/data': 0.0003623650409281254, 'timing/model': 1.1163794379681349, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 0.89}) +[2026-06-24 00:22:46,457][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.13298290967941284, 'train/grad_norm_pre_clip': 14.00564193725586, 'timing/data': 0.00040275498759001493, 'timing/model': 1.113881630008109, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 0.89}) +[2026-06-24 00:22:50,687][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.0689694955945015, 'train/grad_norm_pre_clip': 28.494394302368164, 'timing/data': 0.00040991592686623335, 'timing/model': 1.2826153899077326, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 0.89}) +[2026-06-24 00:22:54,762][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.06570172309875488, 'train/grad_norm_pre_clip': 19.47361946105957, 'timing/data': 0.00041289301589131355, 'timing/model': 1.1208686580648646, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 0.89}) +[2026-06-24 00:22:58,835][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.08852536976337433, 'train/grad_norm_pre_clip': 21.936519622802734, 'timing/data': 0.00042080204002559185, 'timing/model': 1.1118491960223764, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 0.89}) +[2026-06-24 00:23:02,909][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.01624220237135887, 'train/grad_norm_pre_clip': 15.734454154968262, 'timing/data': 0.0003608230035752058, 'timing/model': 1.113211222924292, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 0.89}) +[2026-06-24 00:23:07,196][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.014329701662063599, 'train/grad_norm_pre_clip': 6.4243597984313965, 'timing/data': 0.00040831195656210184, 'timing/model': 1.3297251060139388, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 0.89}) +[2026-06-24 00:23:11,294][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2300 +[2026-06-24 00:24:15,252][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2300 +[2026-06-24 00:24:15,254][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2300 +[2026-06-24 00:24:53,770][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2300 +[2026-06-24 00:24:53,771][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.08630406111478806, 'train/grad_norm_pre_clip': 15.30576229095459, 'eval/action_loss/seconds': 63.95537904405501, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.04114396870136261, 'eval/latency_2/loss': 0.04114396870136261, 'eval/flappy/loss': 0.04114396870136261, 'eval/flappy/latency_2/loss': 0.04114396870136261, 'eval/flap_precision': 0.9382239382239382, 'eval/flap_recall': 0.989145183175034, 'eval/flap_f1': 0.9630118890356671, 'eval/flap_rate_model': 0.21583333333333332, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9630118890356671, 'eval/latency_2/flap_precision': 0.9382239382239382, 'eval/latency_2/flap_recall': 0.989145183175034, 'eval/latency_2/flap_f1': 0.9630118890356671, 'eval/latency_2/flap_rate_model': 0.21583333333333332, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9630118890356671, 'eval/action_classification/data_wait_seconds': 0.5921973586082458, 'eval/action_classification/predict_seconds': 37.764434814453125, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00041910901200026274, 'timing/model': 1.1176866659661755, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.89}) +[2026-06-24 00:24:57,802][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.09041819721460342, 'train/grad_norm_pre_clip': 47.12946319580078, 'timing/data': 0.00036790897138416767, 'timing/model': 1.1084631859557703, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 0.89}) +[2026-06-24 00:25:01,938][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.1973997801542282, 'train/grad_norm_pre_clip': 22.895265579223633, 'timing/data': 0.0004328869981691241, 'timing/model': 1.1922641500132158, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 0.89}) +[2026-06-24 00:25:05,977][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.08888520300388336, 'train/grad_norm_pre_clip': 19.987728118896484, 'timing/data': 0.00043366500176489353, 'timing/model': 1.106586575973779, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 0.89}) +[2026-06-24 00:25:10,026][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.06428484618663788, 'train/grad_norm_pre_clip': 18.589599609375, 'timing/data': 0.00037872104439884424, 'timing/model': 1.1117000000085682, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 0.89}) +[2026-06-24 00:25:14,080][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.02762093022465706, 'train/grad_norm_pre_clip': 20.0701847076416, 'timing/data': 0.0003721240209415555, 'timing/model': 1.111873244983144, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.89}) +[2026-06-24 00:25:18,127][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.07679221779108047, 'train/grad_norm_pre_clip': 26.760669708251953, 'timing/data': 0.0004964389372617006, 'timing/model': 1.1065994889941067, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 0.89}) +[2026-06-24 00:25:22,310][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.055470190942287445, 'train/grad_norm_pre_clip': 9.347061157226562, 'timing/data': 0.0005759230116382241, 'timing/model': 1.1101335319690406, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 0.89}) +[2026-06-24 00:25:26,359][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.008574399165809155, 'train/grad_norm_pre_clip': 18.67491340637207, 'timing/data': 0.00036043499130755663, 'timing/model': 1.1116030339617282, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 0.89}) +[2026-06-24 00:25:30,416][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.0232371985912323, 'train/grad_norm_pre_clip': 17.505386352539062, 'timing/data': 0.0003630489809438586, 'timing/model': 1.1132790619740263, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 0.89}) +[2026-06-24 00:25:34,477][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.10469212383031845, 'train/grad_norm_pre_clip': 35.317710876464844, 'timing/data': 0.0005031459731981158, 'timing/model': 1.1107392529956996, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.89}) +[2026-06-24 00:25:38,696][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.009489506483078003, 'train/grad_norm_pre_clip': 8.35031795501709, 'timing/data': 0.00038927001878619194, 'timing/model': 1.1193964440608397, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 0.89}) +[2026-06-24 00:25:42,765][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.20848965644836426, 'train/grad_norm_pre_clip': 27.432464599609375, 'timing/data': 0.0004420280456542969, 'timing/model': 1.1167290260782465, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 0.89}) +[2026-06-24 00:25:46,825][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.037318576127290726, 'train/grad_norm_pre_clip': 26.65298843383789, 'timing/data': 0.0006123670609667897, 'timing/model': 1.1108490279875696, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 0.9}) +[2026-06-24 00:25:50,892][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.022446438670158386, 'train/grad_norm_pre_clip': 10.209331512451172, 'timing/data': 0.00037495000287890434, 'timing/model': 1.1053013990167528, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 0.9}) +[2026-06-24 00:25:55,027][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.04481693357229233, 'train/grad_norm_pre_clip': 8.245269775390625, 'timing/data': 0.0005742140347138047, 'timing/model': 1.1145252189598978, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 0.9}) +[2026-06-24 00:25:59,106][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.03035975620150566, 'train/grad_norm_pre_clip': 20.33800506591797, 'timing/data': 0.0006258180364966393, 'timing/model': 1.1112193570006639, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 0.9}) +[2026-06-24 00:26:03,178][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.08677367866039276, 'train/grad_norm_pre_clip': 22.69194984436035, 'timing/data': 0.00036070903297513723, 'timing/model': 1.1170375189976767, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 0.9}) +[2026-06-24 00:26:07,251][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.09371035546064377, 'train/grad_norm_pre_clip': 17.890165328979492, 'timing/data': 0.0004896939499303699, 'timing/model': 1.1196575850481167, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 0.9}) +[2026-06-24 00:26:11,402][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.011026558466255665, 'train/grad_norm_pre_clip': 9.423990249633789, 'timing/data': 0.0006632430013269186, 'timing/model': 1.1179721449734643, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 0.9}) +[2026-06-24 00:26:15,477][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.0034260121174156666, 'train/grad_norm_pre_clip': 13.693079948425293, 'timing/data': 0.0003603219520300627, 'timing/model': 1.1184662260347977, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.9}) +[2026-06-24 00:26:19,550][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.04853430017828941, 'train/grad_norm_pre_clip': 18.467546463012695, 'timing/data': 0.00040583196096122265, 'timing/model': 1.1213609649566934, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 0.9}) +[2026-06-24 00:26:23,625][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.2028597891330719, 'train/grad_norm_pre_clip': 25.415470123291016, 'timing/data': 0.00043128698598593473, 'timing/model': 1.1198694229824468, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 0.9}) +[2026-06-24 00:26:27,848][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.019201606512069702, 'train/grad_norm_pre_clip': 11.28763484954834, 'timing/data': 0.00041479989886283875, 'timing/model': 1.1129722279729322, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 0.9}) +[2026-06-24 00:26:31,914][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.03474460914731026, 'train/grad_norm_pre_clip': 26.65143394470215, 'timing/data': 0.0003535000141710043, 'timing/model': 1.1132314939750358, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 0.9}) +[2026-06-24 00:26:35,984][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.013026445172727108, 'train/grad_norm_pre_clip': 19.421215057373047, 'timing/data': 0.00040655501652508974, 'timing/model': 1.1127365489955992, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 0.9}) +[2026-06-24 00:26:40,062][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.22390060126781464, 'train/grad_norm_pre_clip': 25.166542053222656, 'timing/data': 0.0003941149916499853, 'timing/model': 1.1146779579576105, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 0.9}) +[2026-06-24 00:26:44,316][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.06835057586431503, 'train/grad_norm_pre_clip': 20.136960983276367, 'timing/data': 0.0005991869838908315, 'timing/model': 1.1183030059328303, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 0.9}) +[2026-06-24 00:26:48,402][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.04378392547369003, 'train/grad_norm_pre_clip': 9.648883819580078, 'timing/data': 0.0003571310080587864, 'timing/model': 1.1174600130179897, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 0.9}) +[2026-06-24 00:26:52,471][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.05532805994153023, 'train/grad_norm_pre_clip': 9.173797607421875, 'timing/data': 0.00036101602017879486, 'timing/model': 1.1148848769953474, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 0.9}) +[2026-06-24 00:26:56,547][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.09869226068258286, 'train/grad_norm_pre_clip': 26.534751892089844, 'timing/data': 0.0004059090279042721, 'timing/model': 1.1176995610585436, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.9}) +[2026-06-24 00:27:00,792][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.059242382645606995, 'train/grad_norm_pre_clip': 12.995805740356445, 'timing/data': 0.0003602879587560892, 'timing/model': 1.1206466750008985, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 0.9}) +[2026-06-24 00:27:04,871][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.16840465366840363, 'train/grad_norm_pre_clip': 18.871606826782227, 'timing/data': 0.0004512619925662875, 'timing/model': 1.1188350400188938, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 0.9}) +[2026-06-24 00:27:08,932][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.11629590392112732, 'train/grad_norm_pre_clip': 21.399105072021484, 'timing/data': 0.00043852999806404114, 'timing/model': 1.1132382520008832, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 0.9}) +[2026-06-24 00:27:12,999][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.20879197120666504, 'train/grad_norm_pre_clip': 27.779685974121094, 'timing/data': 0.00035952392499893904, 'timing/model': 1.1168152639875188, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 0.9}) +[2026-06-24 00:27:17,235][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.043370671570301056, 'train/grad_norm_pre_clip': 20.731698989868164, 'timing/data': 0.0003884430043399334, 'timing/model': 1.1177321739960462, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 0.9}) +[2026-06-24 00:27:21,343][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.0038245529867708683, 'train/grad_norm_pre_clip': 17.583236694335938, 'timing/data': 0.0003695320338010788, 'timing/model': 1.1211285879835486, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 0.9}) +[2026-06-24 00:27:25,419][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.06666620075702667, 'train/grad_norm_pre_clip': 12.56222915649414, 'timing/data': 0.00036860303953289986, 'timing/model': 1.117253922042437, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 0.9}) +[2026-06-24 00:27:29,510][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.01816936396062374, 'train/grad_norm_pre_clip': 38.96868133544922, 'timing/data': 0.0005447190487757325, 'timing/model': 1.1266631470061839, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 0.9}) +[2026-06-24 00:27:33,754][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.027483617886900902, 'train/grad_norm_pre_clip': 26.359373092651367, 'timing/data': 0.00036761502269655466, 'timing/model': 1.117142215021886, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 0.91}) +[2026-06-24 00:27:37,844][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.06054547056555748, 'train/grad_norm_pre_clip': 15.913302421569824, 'timing/data': 0.0005225769709795713, 'timing/model': 1.118827524012886, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.91}) +[2026-06-24 00:27:41,927][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.17055831849575043, 'train/grad_norm_pre_clip': 28.66922950744629, 'timing/data': 0.0003582129720598459, 'timing/model': 1.122779174009338, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 0.91}) +[2026-06-24 00:27:46,003][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.04488704353570938, 'train/grad_norm_pre_clip': 11.027241706848145, 'timing/data': 0.00036719697527587414, 'timing/model': 1.117008846020326, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 0.91}) +[2026-06-24 00:27:50,253][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.003824361599981785, 'train/grad_norm_pre_clip': 21.70699691772461, 'timing/data': 0.000557276071049273, 'timing/model': 1.116312014986761, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 0.91}) +[2026-06-24 00:27:54,351][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.10038040578365326, 'train/grad_norm_pre_clip': 22.164518356323242, 'timing/data': 0.0003539719618856907, 'timing/model': 1.1351873399689794, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 0.91}) +[2026-06-24 00:27:58,436][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.06000467389822006, 'train/grad_norm_pre_clip': 20.638208389282227, 'timing/data': 0.0003741809632629156, 'timing/model': 1.116534507018514, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 0.91}) +[2026-06-24 00:28:02,509][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.020614488050341606, 'train/grad_norm_pre_clip': 19.394365310668945, 'timing/data': 0.0003861580044031143, 'timing/model': 1.1150659639388323, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 0.91}) +[2026-06-24 00:28:06,756][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.03664981201291084, 'train/grad_norm_pre_clip': 20.434412002563477, 'timing/data': 0.0005998030537739396, 'timing/model': 1.1183345339959487, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 0.91}) +[2026-06-24 00:28:10,838][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.006218723487108946, 'train/grad_norm_pre_clip': 11.067962646484375, 'timing/data': 0.0003575519658625126, 'timing/model': 1.1178102320991457, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 0.91}) +[2026-06-24 00:28:14,933][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.03958093002438545, 'train/grad_norm_pre_clip': 11.99131965637207, 'timing/data': 0.0006678939098492265, 'timing/model': 1.1252490259939805, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 0.91}) +[2026-06-24 00:28:19,011][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.06634129583835602, 'train/grad_norm_pre_clip': 18.90435028076172, 'timing/data': 0.0003687350545078516, 'timing/model': 1.116915947990492, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.91}) +[2026-06-24 00:28:23,254][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.1666601002216339, 'train/grad_norm_pre_clip': 13.44843864440918, 'timing/data': 0.0003667559940367937, 'timing/model': 1.1164892149390653, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 0.91}) +[2026-06-24 00:28:27,345][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.0199083611369133, 'train/grad_norm_pre_clip': 12.242700576782227, 'timing/data': 0.0003557989839464426, 'timing/model': 1.1244223269168288, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 0.91}) +[2026-06-24 00:28:31,424][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.026852145791053772, 'train/grad_norm_pre_clip': 12.158287048339844, 'timing/data': 0.0005741739878430963, 'timing/model': 1.1217653759522364, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 0.91}) +[2026-06-24 00:28:35,523][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.007718243636190891, 'train/grad_norm_pre_clip': 10.638673782348633, 'timing/data': 0.0003587630344554782, 'timing/model': 1.1237714790040627, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.91}) +[2026-06-24 00:28:39,786][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.03878495469689369, 'train/grad_norm_pre_clip': 15.13074016571045, 'timing/data': 0.0006146120140329003, 'timing/model': 1.1206437379587442, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 0.91}) +[2026-06-24 00:28:43,880][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.032324571162462234, 'train/grad_norm_pre_clip': 33.14030456542969, 'timing/data': 0.00035424798261374235, 'timing/model': 1.1205778159201145, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 0.91}) +[2026-06-24 00:28:47,954][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.007206920068711042, 'train/grad_norm_pre_clip': 16.035734176635742, 'timing/data': 0.00040807202458381653, 'timing/model': 1.1171183619881049, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 0.91}) +[2026-06-24 00:28:52,036][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.06032760441303253, 'train/grad_norm_pre_clip': 13.351818084716797, 'timing/data': 0.0006715409690514207, 'timing/model': 1.1164270309964195, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 0.91}) +[2026-06-24 00:28:56,273][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.020160458981990814, 'train/grad_norm_pre_clip': 31.36840057373047, 'timing/data': 0.0007154429331421852, 'timing/model': 1.1155641729710624, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 0.91}) +[2026-06-24 00:29:00,353][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.046035006642341614, 'train/grad_norm_pre_clip': 13.463142395019531, 'timing/data': 0.0003570610424503684, 'timing/model': 1.1181273180991411, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.91}) +[2026-06-24 00:29:04,440][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.03202681615948677, 'train/grad_norm_pre_clip': 15.230449676513672, 'timing/data': 0.0003535220166668296, 'timing/model': 1.1341276339953765, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 0.91}) +[2026-06-24 00:29:08,518][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.003429867560043931, 'train/grad_norm_pre_clip': 8.85730266571045, 'timing/data': 0.0003574499860405922, 'timing/model': 1.119153478066437, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 0.91}) +[2026-06-24 00:29:12,764][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.008933713659644127, 'train/grad_norm_pre_clip': 22.01996421813965, 'timing/data': 0.0003610270796343684, 'timing/model': 1.1154869170859456, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 0.91}) +[2026-06-24 00:29:16,860][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.018040750175714493, 'train/grad_norm_pre_clip': 19.21109390258789, 'timing/data': 0.00036721001379191875, 'timing/model': 1.1234915690729395, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 0.91}) +[2026-06-24 00:29:20,938][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.026011386886239052, 'train/grad_norm_pre_clip': 42.93153762817383, 'timing/data': 0.0004227019380778074, 'timing/model': 1.1166648750659078, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 0.92}) +[2026-06-24 00:29:25,038][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.3779682219028473, 'train/grad_norm_pre_clip': 58.057926177978516, 'timing/data': 0.00041887001134455204, 'timing/model': 1.1160091169876978, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 0.92}) +[2026-06-24 00:29:29,300][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.0044508446007966995, 'train/grad_norm_pre_clip': 47.45809555053711, 'timing/data': 0.0003646000986918807, 'timing/model': 1.123178077978082, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 0.92}) +[2026-06-24 00:29:33,405][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.08283715695142746, 'train/grad_norm_pre_clip': 10.092034339904785, 'timing/data': 0.00037137209437787533, 'timing/model': 1.123477328917943, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 0.92}) +[2026-06-24 00:29:37,496][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.01183127611875534, 'train/grad_norm_pre_clip': 14.808063507080078, 'timing/data': 0.0003592349821701646, 'timing/model': 1.1209279189351946, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 0.92}) +[2026-06-24 00:29:41,575][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.04401854798197746, 'train/grad_norm_pre_clip': 12.837571144104004, 'timing/data': 0.00044578302185982466, 'timing/model': 1.1129412800073624, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.92}) +[2026-06-24 00:29:45,822][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.03684700280427933, 'train/grad_norm_pre_clip': 34.865013122558594, 'timing/data': 0.0006842139409855008, 'timing/model': 1.1200087670003995, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 0.92}) +[2026-06-24 00:29:49,916][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.011288237757980824, 'train/grad_norm_pre_clip': 5.972642421722412, 'timing/data': 0.00035753194242715836, 'timing/model': 1.1182152449619025, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 0.92}) +[2026-06-24 00:29:53,998][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.00018923355673905462, 'train/grad_norm_pre_clip': 9.631277084350586, 'timing/data': 0.000377955031581223, 'timing/model': 1.117342344019562, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 0.92}) +[2026-06-24 00:29:58,102][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.23076231777668, 'train/grad_norm_pre_clip': 38.02536392211914, 'timing/data': 0.0003640709910541773, 'timing/model': 1.1206972170621157, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 0.92}) +[2026-06-24 00:30:02,372][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.15223081409931183, 'train/grad_norm_pre_clip': 23.46472930908203, 'timing/data': 0.0004848010139539838, 'timing/model': 1.119779287953861, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.92}) +[2026-06-24 00:30:06,490][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.12542276084423065, 'train/grad_norm_pre_clip': 20.177215576171875, 'timing/data': 0.00039905100129544735, 'timing/model': 1.1351015089312568, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 0.92}) +[2026-06-24 00:30:10,577][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.011721035465598106, 'train/grad_norm_pre_clip': 20.554275512695312, 'timing/data': 0.0004091169685125351, 'timing/model': 1.118329571094364, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 0.92}) +[2026-06-24 00:30:14,665][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.03928517550230026, 'train/grad_norm_pre_clip': 23.93287467956543, 'timing/data': 0.0003528599627315998, 'timing/model': 1.120927395997569, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 0.92}) +[2026-06-24 00:30:18,943][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.0524270199239254, 'train/grad_norm_pre_clip': 20.363079071044922, 'timing/data': 0.00039295502938330173, 'timing/model': 1.1214207800803706, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 0.92}) +[2026-06-24 00:30:23,011][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.023191850632429123, 'train/grad_norm_pre_clip': 7.035923957824707, 'timing/data': 0.0003614160232245922, 'timing/model': 1.118061018991284, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.92}) +[2026-06-24 00:30:27,100][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.013357569463551044, 'train/grad_norm_pre_clip': 10.866442680358887, 'timing/data': 0.0004217319656163454, 'timing/model': 1.1150287239579484, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 0.92}) +[2026-06-24 00:30:31,181][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.12321010231971741, 'train/grad_norm_pre_clip': 38.86370849609375, 'timing/data': 0.00035774894058704376, 'timing/model': 1.1229383270256221, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 0.92}) +[2026-06-24 00:30:35,464][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.04073130711913109, 'train/grad_norm_pre_clip': 27.016172409057617, 'timing/data': 0.00034500600304454565, 'timing/model': 1.124589886981994, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 0.92}) +[2026-06-24 00:30:39,565][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.02279118075966835, 'train/grad_norm_pre_clip': 22.720396041870117, 'timing/data': 0.0003836719552055001, 'timing/model': 1.1175927240401506, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 0.92}) +[2026-06-24 00:30:43,645][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.18486569821834564, 'train/grad_norm_pre_clip': 46.70463180541992, 'timing/data': 0.0005051419138908386, 'timing/model': 1.1191690519917756, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 0.92}) +[2026-06-24 00:30:47,743][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.07216492295265198, 'train/grad_norm_pre_clip': 24.867389678955078, 'timing/data': 0.00036667496897280216, 'timing/model': 1.1182996119605377, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 0.92}) +[2026-06-24 00:30:51,984][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.049446720629930496, 'train/grad_norm_pre_clip': 14.030372619628906, 'timing/data': 0.000418122042901814, 'timing/model': 1.1183288149768487, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 0.92}) +[2026-06-24 00:30:56,070][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.025739697739481926, 'train/grad_norm_pre_clip': 11.398069381713867, 'timing/data': 0.0004056410398334265, 'timing/model': 1.1216198530746624, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 0.92}) +[2026-06-24 00:31:00,178][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.010593006387352943, 'train/grad_norm_pre_clip': 20.131423950195312, 'timing/data': 0.0003947969526052475, 'timing/model': 1.1370359610300511, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 0.92}) +[2026-06-24 00:31:04,296][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.02226685732603073, 'train/grad_norm_pre_clip': 10.983184814453125, 'timing/data': 0.00035642494913190603, 'timing/model': 1.1321733359945938, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.92}) +[2026-06-24 00:31:08,576][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.039379652589559555, 'train/grad_norm_pre_clip': 14.168042182922363, 'timing/data': 0.0003571601118892431, 'timing/model': 1.1178194939857349, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 0.93}) +[2026-06-24 00:31:12,665][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.029049720615148544, 'train/grad_norm_pre_clip': 12.663207054138184, 'timing/data': 0.0006097860168665648, 'timing/model': 1.1207614729646593, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 0.93}) +[2026-06-24 00:31:16,748][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.005156922619789839, 'train/grad_norm_pre_clip': 4.811360836029053, 'timing/data': 0.00037792196962982416, 'timing/model': 1.1190130889881402, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 0.93}) +[2026-06-24 00:31:20,830][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.08926154673099518, 'train/grad_norm_pre_clip': 20.28500747680664, 'timing/data': 0.0005481750704348087, 'timing/model': 1.1157105109887198, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 0.93}) +[2026-06-24 00:31:25,076][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.043229155242443085, 'train/grad_norm_pre_clip': 16.62803077697754, 'timing/data': 0.00041637197136878967, 'timing/model': 1.1157529259799048, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 0.93}) +[2026-06-24 00:31:29,160][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.047297026962041855, 'train/grad_norm_pre_clip': 11.001896858215332, 'timing/data': 0.0004215369699522853, 'timing/model': 1.1186915610451251, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 0.93}) +[2026-06-24 00:31:33,237][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.15083885192871094, 'train/grad_norm_pre_clip': 18.18773651123047, 'timing/data': 0.0004354619886726141, 'timing/model': 1.1173256870824844, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 0.93}) +[2026-06-24 00:31:37,323][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.03374344855546951, 'train/grad_norm_pre_clip': 12.916260719299316, 'timing/data': 0.00037209794390946627, 'timing/model': 1.1185112319653854, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 0.93}) +[2026-06-24 00:31:41,570][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.00903910119086504, 'train/grad_norm_pre_clip': 37.254310607910156, 'timing/data': 0.0007567419670522213, 'timing/model': 1.1148281760979444, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 0.93}) +[2026-06-24 00:31:45,653][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2400 +[2026-06-24 00:32:50,276][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2400 +[2026-06-24 00:32:50,278][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2400 +[2026-06-24 00:33:30,067][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2400 +[2026-06-24 00:33:30,068][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.012034366838634014, 'train/grad_norm_pre_clip': 27.755905151367188, 'eval/action_loss/seconds': 64.61974311200902, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.049981068819761276, 'eval/latency_2/loss': 0.049981068819761276, 'eval/flappy/loss': 0.049981068819761276, 'eval/flappy/latency_2/loss': 0.049981068819761276, 'eval/flap_precision': 0.9815602836879432, 'eval/flap_recall': 0.9389416553595658, 'eval/flap_f1': 0.9597780859916782, 'eval/flap_rate_model': 0.19583333333333333, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9597780859916784, 'eval/latency_2/flap_precision': 0.9815602836879432, 'eval/latency_2/flap_recall': 0.9389416553595658, 'eval/latency_2/flap_f1': 0.9597780859916782, 'eval/latency_2/flap_rate_model': 0.19583333333333333, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9597780859916784, 'eval/action_classification/data_wait_seconds': 0.5913307070732117, 'eval/action_classification/predict_seconds': 39.027137756347656, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0006742510013282299, 'timing/model': 1.1154059689724818, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.93}) +[2026-06-24 00:33:34,120][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.014008289203047752, 'train/grad_norm_pre_clip': 31.84903335571289, 'timing/data': 0.00045173289254307747, 'timing/model': 1.1075756340287626, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 0.93}) +[2026-06-24 00:33:38,287][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.16619554162025452, 'train/grad_norm_pre_clip': 50.071983337402344, 'timing/data': 0.00048200401943176985, 'timing/model': 1.2402113259304315, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 0.93}) +[2026-06-24 00:33:42,326][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.05883665010333061, 'train/grad_norm_pre_clip': 22.55030059814453, 'timing/data': 0.0003753759665414691, 'timing/model': 1.1120609630597755, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.93}) +[2026-06-24 00:33:46,360][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.22544865310192108, 'train/grad_norm_pre_clip': 26.69218635559082, 'timing/data': 0.0004958410281687975, 'timing/model': 1.1024192039621994, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 0.93}) +[2026-06-24 00:33:50,402][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.007072328124195337, 'train/grad_norm_pre_clip': 23.680965423583984, 'timing/data': 0.00035742600448429585, 'timing/model': 1.1017361610429361, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 0.93}) +[2026-06-24 00:33:54,590][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.09239757061004639, 'train/grad_norm_pre_clip': 14.47533893585205, 'timing/data': 0.0003581959754228592, 'timing/model': 1.254073953954503, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 0.93}) +[2026-06-24 00:33:58,652][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.06254366040229797, 'train/grad_norm_pre_clip': 14.343083381652832, 'timing/data': 0.0003589170519262552, 'timing/model': 1.1130730779841542, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 0.93}) +[2026-06-24 00:34:02,716][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.056853752583265305, 'train/grad_norm_pre_clip': 30.20734405517578, 'timing/data': 0.00043639400973916054, 'timing/model': 1.109922599978745, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 0.93}) +[2026-06-24 00:34:06,767][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.02377837710082531, 'train/grad_norm_pre_clip': 42.662742614746094, 'timing/data': 0.0003728440497070551, 'timing/model': 1.111152933910489, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 0.93}) +[2026-06-24 00:34:10,965][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.24991576373577118, 'train/grad_norm_pre_clip': 47.75468826293945, 'timing/data': 0.0004575470229610801, 'timing/model': 1.2489933191100135, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.93}) +[2026-06-24 00:34:15,019][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.042327068746089935, 'train/grad_norm_pre_clip': 26.151519775390625, 'timing/data': 0.00044103292748332024, 'timing/model': 1.1121023639570922, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 0.93}) +[2026-06-24 00:34:19,079][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.02048773504793644, 'train/grad_norm_pre_clip': 19.013744354248047, 'timing/data': 0.00034740299452096224, 'timing/model': 1.111229502945207, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 0.93}) +[2026-06-24 00:34:23,158][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.06336060166358948, 'train/grad_norm_pre_clip': 31.945463180541992, 'timing/data': 0.0004914359888061881, 'timing/model': 1.1253774130018428, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 0.93}) +[2026-06-24 00:34:27,368][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.011095995083451271, 'train/grad_norm_pre_clip': 5.098740100860596, 'timing/data': 0.0005834900075569749, 'timing/model': 1.2514356520259753, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 0.93}) +[2026-06-24 00:34:31,445][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.045402299612760544, 'train/grad_norm_pre_clip': 27.66442108154297, 'timing/data': 0.0003505790373310447, 'timing/model': 1.1151459739776328, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 0.93}) +[2026-06-24 00:34:35,506][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.06336550414562225, 'train/grad_norm_pre_clip': 12.236674308776855, 'timing/data': 0.000361428945325315, 'timing/model': 1.112309188931249, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 0.94}) +[2026-06-24 00:34:39,566][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.08935248106718063, 'train/grad_norm_pre_clip': 21.39071273803711, 'timing/data': 0.00041813694406300783, 'timing/model': 1.1066389069892466, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 0.94}) +[2026-06-24 00:34:43,770][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.010100201703608036, 'train/grad_norm_pre_clip': 17.31313133239746, 'timing/data': 0.00040022493340075016, 'timing/model': 1.2417935889679939, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 0.94}) +[2026-06-24 00:34:47,840][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.1588994711637497, 'train/grad_norm_pre_clip': 22.218013763427734, 'timing/data': 0.0003604050725698471, 'timing/model': 1.1184507650323212, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 0.94}) +[2026-06-24 00:34:51,913][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.1561489850282669, 'train/grad_norm_pre_clip': 64.69532012939453, 'timing/data': 0.0007015109295025468, 'timing/model': 1.114476904971525, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.94}) +[2026-06-24 00:34:55,983][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.09650114923715591, 'train/grad_norm_pre_clip': 13.275469779968262, 'timing/data': 0.0003583240322768688, 'timing/model': 1.1127712589222938, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 0.94}) +[2026-06-24 00:35:00,186][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.16324013471603394, 'train/grad_norm_pre_clip': 28.86724853515625, 'timing/data': 0.00035774102434515953, 'timing/model': 1.239654240082018, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 0.94}) +[2026-06-24 00:35:04,249][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.1547439694404602, 'train/grad_norm_pre_clip': 26.66185760498047, 'timing/data': 0.0006312839686870575, 'timing/model': 1.1165064909728244, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 0.94}) +[2026-06-24 00:35:08,322][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.03578360006213188, 'train/grad_norm_pre_clip': 17.535364151000977, 'timing/data': 0.0004808970261365175, 'timing/model': 1.1190054760081694, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 0.94}) +[2026-06-24 00:35:12,396][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.014856047928333282, 'train/grad_norm_pre_clip': 25.45201873779297, 'timing/data': 0.00035743298940360546, 'timing/model': 1.1148367719724774, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 0.94}) +[2026-06-24 00:35:16,608][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.18017493188381195, 'train/grad_norm_pre_clip': 22.525976181030273, 'timing/data': 0.00039629999082535505, 'timing/model': 1.254143537953496, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 0.94}) +[2026-06-24 00:35:20,699][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.019913332536816597, 'train/grad_norm_pre_clip': 42.65913009643555, 'timing/data': 0.00044104596599936485, 'timing/model': 1.12366543605458, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 0.94}) +[2026-06-24 00:35:24,781][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.05793234705924988, 'train/grad_norm_pre_clip': 25.601268768310547, 'timing/data': 0.00041363993659615517, 'timing/model': 1.120797289069742, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 0.94}) +[2026-06-24 00:35:28,865][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.06324528902769089, 'train/grad_norm_pre_clip': 24.910917282104492, 'timing/data': 0.00041760201565921307, 'timing/model': 1.1158055110136047, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 0.94}) +[2026-06-24 00:35:33,059][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.34137603640556335, 'train/grad_norm_pre_clip': 28.56296730041504, 'timing/data': 0.00035927502904087305, 'timing/model': 1.244306487031281, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 0.94}) +[2026-06-24 00:35:37,133][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.03434472531080246, 'train/grad_norm_pre_clip': 14.672977447509766, 'timing/data': 0.0003495409619063139, 'timing/model': 1.1160294419387355, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 0.94}) +[2026-06-24 00:35:41,208][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.05048695579171181, 'train/grad_norm_pre_clip': 20.58547592163086, 'timing/data': 0.00041175202932208776, 'timing/model': 1.1172881970414892, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 0.94}) +[2026-06-24 00:35:45,277][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.002457187045365572, 'train/grad_norm_pre_clip': 4.627943992614746, 'timing/data': 0.00040639599319547415, 'timing/model': 1.110330799012445, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 0.94}) +[2026-06-24 00:35:49,488][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.00818118266761303, 'train/grad_norm_pre_clip': 11.797304153442383, 'timing/data': 0.0003588020335882902, 'timing/model': 1.2573221329366788, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 0.94}) +[2026-06-24 00:35:53,563][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.008688687346875668, 'train/grad_norm_pre_clip': 13.793841361999512, 'timing/data': 0.00035993405617773533, 'timing/model': 1.1158563690260053, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 0.94}) +[2026-06-24 00:35:57,644][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.09069298952817917, 'train/grad_norm_pre_clip': 22.21669578552246, 'timing/data': 0.00044606998562812805, 'timing/model': 1.113090759026818, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 0.94}) +[2026-06-24 00:36:01,722][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.04339177906513214, 'train/grad_norm_pre_clip': 18.160512924194336, 'timing/data': 0.00048225000500679016, 'timing/model': 1.1177944189403206, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 0.94}) +[2026-06-24 00:36:05,939][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.0287595447152853, 'train/grad_norm_pre_clip': 6.771091461181641, 'timing/data': 0.00036220101173967123, 'timing/model': 1.2479483550414443, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 0.94}) +[2026-06-24 00:36:10,021][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.22551019489765167, 'train/grad_norm_pre_clip': 26.416379928588867, 'timing/data': 0.00038303202018141747, 'timing/model': 1.115151472040452, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 0.94}) +[2026-06-24 00:36:14,107][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.23211872577667236, 'train/grad_norm_pre_clip': 21.43119239807129, 'timing/data': 0.0003630109131336212, 'timing/model': 1.11690558004193, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 0.94}) +[2026-06-24 00:36:18,197][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.14091706275939941, 'train/grad_norm_pre_clip': 16.962238311767578, 'timing/data': 0.0003899350995197892, 'timing/model': 1.116955196019262, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 0.94}) +[2026-06-24 00:36:22,407][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.00703962380066514, 'train/grad_norm_pre_clip': 16.939146041870117, 'timing/data': 0.0003826699685305357, 'timing/model': 1.2455117959761992, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 0.95}) +[2026-06-24 00:36:26,486][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.0007601705729030073, 'train/grad_norm_pre_clip': 15.536721229553223, 'timing/data': 0.00037003098987042904, 'timing/model': 1.112346883979626, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 0.95}) +[2026-06-24 00:36:30,573][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.02570442482829094, 'train/grad_norm_pre_clip': 23.92345428466797, 'timing/data': 0.00036801001988351345, 'timing/model': 1.115993448998779, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 0.95}) +[2026-06-24 00:36:34,661][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.029012050479650497, 'train/grad_norm_pre_clip': 16.993818283081055, 'timing/data': 0.0005190339870750904, 'timing/model': 1.1158953489502892, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.95}) +[2026-06-24 00:36:38,889][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.0724930390715599, 'train/grad_norm_pre_clip': 18.253162384033203, 'timing/data': 0.0003571530105546117, 'timing/model': 1.2623330239439383, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 0.95}) +[2026-06-24 00:36:42,965][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.034710559993982315, 'train/grad_norm_pre_clip': 25.43553924560547, 'timing/data': 0.0003974869614467025, 'timing/model': 1.1196499589132145, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 0.95}) +[2026-06-24 00:36:47,067][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.10523583739995956, 'train/grad_norm_pre_clip': 21.8159236907959, 'timing/data': 0.0003671220038086176, 'timing/model': 1.1242416439345106, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 0.95}) +[2026-06-24 00:36:51,145][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.08202949166297913, 'train/grad_norm_pre_clip': 28.091089248657227, 'timing/data': 0.00035787594970315695, 'timing/model': 1.1176427140599117, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 0.95}) +[2026-06-24 00:36:55,355][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.20509567856788635, 'train/grad_norm_pre_clip': 39.91463851928711, 'timing/data': 0.0006095350254327059, 'timing/model': 1.2551239259773865, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 0.95}) +[2026-06-24 00:36:59,442][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.05474081262946129, 'train/grad_norm_pre_clip': 17.3597469329834, 'timing/data': 0.00038807408418506384, 'timing/model': 1.11793861805927, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 0.95}) +[2026-06-24 00:37:03,526][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.038948193192481995, 'train/grad_norm_pre_clip': 46.80727767944336, 'timing/data': 0.00036293105222284794, 'timing/model': 1.1174576529301703, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.95}) +[2026-06-24 00:37:07,603][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.010063783265650272, 'train/grad_norm_pre_clip': 14.781242370605469, 'timing/data': 0.0005138718988746405, 'timing/model': 1.1194507899926975, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 0.95}) +[2026-06-24 00:37:11,832][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.018489887937903404, 'train/grad_norm_pre_clip': 20.51239013671875, 'timing/data': 0.0006020850269123912, 'timing/model': 1.2495175810763612, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 0.95}) +[2026-06-24 00:37:15,910][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.013785446994006634, 'train/grad_norm_pre_clip': 73.13717651367188, 'timing/data': 0.000389978988096118, 'timing/model': 1.1175014259060845, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 0.95}) +[2026-06-24 00:37:19,984][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.16990646719932556, 'train/grad_norm_pre_clip': 30.59354591369629, 'timing/data': 0.0003818600671365857, 'timing/model': 1.1124804749852046, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 0.95}) +[2026-06-24 00:37:24,068][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.009587548673152924, 'train/grad_norm_pre_clip': 15.70011043548584, 'timing/data': 0.00035947293508797884, 'timing/model': 1.1223683269927278, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 0.95}) +[2026-06-24 00:37:28,273][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.06614533066749573, 'train/grad_norm_pre_clip': 16.139995574951172, 'timing/data': 0.0004014569567516446, 'timing/model': 1.2393399509601295, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 0.95}) +[2026-06-24 00:37:32,345][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.09327100962400436, 'train/grad_norm_pre_clip': 18.811819076538086, 'timing/data': 0.000405161059461534, 'timing/model': 1.1177199699450284, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 0.95}) +[2026-06-24 00:37:36,427][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.1402748078107834, 'train/grad_norm_pre_clip': 37.79045486450195, 'timing/data': 0.00046673999167978764, 'timing/model': 1.1168440299807116, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 0.95}) +[2026-06-24 00:37:40,505][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.012559383176267147, 'train/grad_norm_pre_clip': 13.851993560791016, 'timing/data': 0.0003984340000897646, 'timing/model': 1.1171764470636845, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 0.95}) +[2026-06-24 00:37:44,730][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.05640721321105957, 'train/grad_norm_pre_clip': 18.642353057861328, 'timing/data': 0.00039253104478120804, 'timing/model': 1.2605776209384203, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 0.95}) +[2026-06-24 00:37:48,815][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.0292437095195055, 'train/grad_norm_pre_clip': 26.237295150756836, 'timing/data': 0.00035306892823427916, 'timing/model': 1.124617004999891, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 0.95}) +[2026-06-24 00:37:52,910][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.05085569992661476, 'train/grad_norm_pre_clip': 20.459854125976562, 'timing/data': 0.00038895895704627037, 'timing/model': 1.1212803310481831, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 0.95}) +[2026-06-24 00:37:56,998][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.04612049087882042, 'train/grad_norm_pre_clip': 27.153905868530273, 'timing/data': 0.00036809209268540144, 'timing/model': 1.1195583769585937, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 0.95}) +[2026-06-24 00:38:01,204][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.06186918541789055, 'train/grad_norm_pre_clip': 10.386312484741211, 'timing/data': 0.0003662360832095146, 'timing/model': 1.249249187996611, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 0.95}) +[2026-06-24 00:38:05,291][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.011081387288868427, 'train/grad_norm_pre_clip': 7.560017108917236, 'timing/data': 0.00036429602187126875, 'timing/model': 1.1188376130303368, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 0.95}) +[2026-06-24 00:38:09,404][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.005475153680890799, 'train/grad_norm_pre_clip': 21.077096939086914, 'timing/data': 0.0003685510018840432, 'timing/model': 1.122584852972068, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 0.96}) +[2026-06-24 00:38:13,493][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.05824221670627594, 'train/grad_norm_pre_clip': 14.482100486755371, 'timing/data': 0.00043526303488761187, 'timing/model': 1.1221112840576097, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 0.96}) +[2026-06-24 00:38:17,722][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.18192622065544128, 'train/grad_norm_pre_clip': 24.288230895996094, 'timing/data': 0.0003641380462795496, 'timing/model': 1.2578879130305722, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 0.96}) +[2026-06-24 00:38:21,805][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.2910251021385193, 'train/grad_norm_pre_clip': 20.603445053100586, 'timing/data': 0.0003676170017570257, 'timing/model': 1.1195290490286425, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 0.96}) +[2026-06-24 00:38:25,892][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.038089025765657425, 'train/grad_norm_pre_clip': 11.217008590698242, 'timing/data': 0.0003671190934255719, 'timing/model': 1.1172527930466458, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 0.96}) +[2026-06-24 00:38:29,984][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.02667122706770897, 'train/grad_norm_pre_clip': 18.89425277709961, 'timing/data': 0.0003507389919832349, 'timing/model': 1.1261912870686501, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 0.96}) +[2026-06-24 00:38:34,192][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.0573311485350132, 'train/grad_norm_pre_clip': 24.738264083862305, 'timing/data': 0.0003561209887266159, 'timing/model': 1.249240527045913, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 0.96}) +[2026-06-24 00:38:38,282][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.04896054044365883, 'train/grad_norm_pre_clip': 14.676140785217285, 'timing/data': 0.0003657459747046232, 'timing/model': 1.1177003390621394, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 0.96}) +[2026-06-24 00:38:42,360][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.046805862337350845, 'train/grad_norm_pre_clip': 9.796138763427734, 'timing/data': 0.0003589480184018612, 'timing/model': 1.1133160460740328, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 0.96}) +[2026-06-24 00:38:46,460][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.025198066607117653, 'train/grad_norm_pre_clip': 27.457683563232422, 'timing/data': 0.0003639289643615484, 'timing/model': 1.123911997070536, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 0.96}) +[2026-06-24 00:38:50,667][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.05718988552689552, 'train/grad_norm_pre_clip': 28.201364517211914, 'timing/data': 0.00035546498838812113, 'timing/model': 1.242665247991681, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 0.96}) +[2026-06-24 00:38:54,750][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.0380825474858284, 'train/grad_norm_pre_clip': 14.882953643798828, 'timing/data': 0.00037599203642457724, 'timing/model': 1.124121151980944, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 0.96}) +[2026-06-24 00:38:58,842][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.27188220620155334, 'train/grad_norm_pre_clip': 15.90699291229248, 'timing/data': 0.000359299941919744, 'timing/model': 1.1215539670083672, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.96}) +[2026-06-24 00:39:02,940][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.09456758946180344, 'train/grad_norm_pre_clip': 12.728650093078613, 'timing/data': 0.00036168110091239214, 'timing/model': 1.1210343170678243, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 0.96}) +[2026-06-24 00:39:07,166][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.06696504354476929, 'train/grad_norm_pre_clip': 14.17053508758545, 'timing/data': 0.0006080560851842165, 'timing/model': 1.2602004029322416, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 0.96}) +[2026-06-24 00:39:11,262][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.2852311432361603, 'train/grad_norm_pre_clip': 19.11962127685547, 'timing/data': 0.00036804797127842903, 'timing/model': 1.1280460560228676, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 0.96}) +[2026-06-24 00:39:15,345][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.010279702953994274, 'train/grad_norm_pre_clip': 11.756083488464355, 'timing/data': 0.0005465400172397494, 'timing/model': 1.1167062079766765, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 0.96}) +[2026-06-24 00:39:19,437][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.045807577669620514, 'train/grad_norm_pre_clip': 12.983983993530273, 'timing/data': 0.0003696409985423088, 'timing/model': 1.123645593994297, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 0.96}) +[2026-06-24 00:39:23,669][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.015603159554302692, 'train/grad_norm_pre_clip': 22.37541961669922, 'timing/data': 0.0006262270035222173, 'timing/model': 1.2574063359061256, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 0.96}) +[2026-06-24 00:39:27,750][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.03718424215912819, 'train/grad_norm_pre_clip': 27.402997970581055, 'timing/data': 0.00036662700586020947, 'timing/model': 1.115781668922864, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 0.96}) +[2026-06-24 00:39:31,825][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.10345682501792908, 'train/grad_norm_pre_clip': 13.536705017089844, 'timing/data': 0.00042448495514690876, 'timing/model': 1.1191080340649933, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 0.96}) +[2026-06-24 00:39:35,911][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.058342818170785904, 'train/grad_norm_pre_clip': 24.308664321899414, 'timing/data': 0.00037791591603308916, 'timing/model': 1.1209940270055085, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 0.96}) +[2026-06-24 00:39:40,150][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.023121479898691177, 'train/grad_norm_pre_clip': 11.04141902923584, 'timing/data': 0.00035478908102959394, 'timing/model': 1.2616077880375087, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 0.96}) +[2026-06-24 00:39:44,248][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.052914269268512726, 'train/grad_norm_pre_clip': 24.60881233215332, 'timing/data': 0.0003606539685279131, 'timing/model': 1.1244152419967577, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 0.96}) +[2026-06-24 00:39:48,341][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.03175525367259979, 'train/grad_norm_pre_clip': 18.839162826538086, 'timing/data': 0.0003626879770308733, 'timing/model': 1.1272365159820765, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 0.96}) +[2026-06-24 00:39:52,426][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.05369342863559723, 'train/grad_norm_pre_clip': 8.436270713806152, 'timing/data': 0.0006380460690706968, 'timing/model': 1.122846772079356, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 0.96}) +[2026-06-24 00:39:56,649][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.0699186846613884, 'train/grad_norm_pre_clip': 29.801054000854492, 'timing/data': 0.0003574879374355078, 'timing/model': 1.2523385870736092, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 0.97}) +[2026-06-24 00:40:00,723][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.06874195486307144, 'train/grad_norm_pre_clip': 13.717340469360352, 'timing/data': 0.00035164097789675, 'timing/model': 1.1109413679223508, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 0.97}) +[2026-06-24 00:40:04,811][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.047669123858213425, 'train/grad_norm_pre_clip': 14.271900177001953, 'timing/data': 0.0003655190812423825, 'timing/model': 1.1225465330062434, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 0.97}) +[2026-06-24 00:40:08,881][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.01413167268037796, 'train/grad_norm_pre_clip': 29.000625610351562, 'timing/data': 0.0003660069778561592, 'timing/model': 1.1130584389902651, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 0.97}) +[2026-06-24 00:40:13,099][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.0059530180878937244, 'train/grad_norm_pre_clip': 15.031973838806152, 'timing/data': 0.00045043393038213253, 'timing/model': 1.2506235729670152, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 0.97}) +[2026-06-24 00:40:17,178][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.08100409060716629, 'train/grad_norm_pre_clip': 28.23740577697754, 'timing/data': 0.00040912104304879904, 'timing/model': 1.1144552968908101, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 0.97}) +[2026-06-24 00:40:21,283][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2500 +[2026-06-24 00:41:25,161][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2500 +[2026-06-24 00:41:25,163][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2500 +[2026-06-24 00:42:04,342][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2500 +[2026-06-24 00:52:02,078][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-06-24 00:52:02,079][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.01760539971292019, 'train/grad_norm_pre_clip': 24.632722854614258, 'eval/action_loss/seconds': 63.87637579801958, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03314562141895294, 'eval/latency_2/loss': 0.03314562141895294, 'eval/flappy/loss': 0.03314562141895294, 'eval/flappy/latency_2/loss': 0.03314562141895294, 'eval/flap_precision': 0.9611780455153949, 'eval/flap_recall': 0.9742198100407056, 'eval/flap_f1': 0.9676549865229109, 'eval/flap_rate_model': 0.2075, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9676549865229109, 'eval/latency_2/flap_precision': 0.9611780455153949, 'eval/latency_2/flap_recall': 0.9742198100407056, 'eval/latency_2/flap_f1': 0.9676549865229109, 'eval/latency_2/flap_rate_model': 0.2075, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9676549865229109, 'eval/action_classification/data_wait_seconds': 0.5642914175987244, 'eval/action_classification/predict_seconds': 38.4476318359375, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 370.4600000000205, 'rl_games_eval/mid_train/mean_length': 3003.0, 'rl_games_eval/mid_train/std_reward': 148.28000000001026, 'rl_games_eval/mid_train/std_length': 1194.0, 'rl_games_eval/mid_train/macro_mean_reward': 370.4600000000205, 'rl_games_eval/mid_train/macro_mean_length': 3003.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 370.4600000000205, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3003.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 148.28000000001026, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1194.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 370.4600000000205, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005343719385564327, 'timing/model': 1.1353543830337003, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 0.97}) +[2026-06-24 00:52:02,079][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_2500_state +[2026-06-24 00:52:02,082][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 00:52:05,791][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_2500_state/model.safetensors +[2026-06-24 00:52:13,613][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_2500_state/optimizer.bin +[2026-06-24 00:52:13,614][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_2500_state/sampler.bin +[2026-06-24 00:52:13,614][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-24 00:52:13,615][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-24 00:52:13,690][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-24 00:52:13,690][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 00:52:13,721][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 00:52:17,692][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.030684521421790123, 'train/grad_norm_pre_clip': 10.597513198852539, 'timing/data': 0.0006166519597172737, 'timing/model': 1.0869570981012657, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 0.97}) +[2026-06-24 00:52:21,753][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.014844408258795738, 'train/grad_norm_pre_clip': 19.546310424804688, 'timing/data': 0.0005825520493090153, 'timing/model': 1.179463693057187, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 0.97}) +[2026-06-24 00:52:25,752][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.004347232636064291, 'train/grad_norm_pre_clip': 12.507896423339844, 'timing/data': 0.0003798799589276314, 'timing/model': 1.0972334939287975, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 0.97}) +[2026-06-24 00:52:29,764][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.0360507033765316, 'train/grad_norm_pre_clip': 12.273452758789062, 'timing/data': 0.00044724508188664913, 'timing/model': 1.1027936909813434, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 0.97}) +[2026-06-24 00:52:33,767][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.049531273543834686, 'train/grad_norm_pre_clip': 22.76377296447754, 'timing/data': 0.0003985350485891104, 'timing/model': 1.1035105569753796, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 0.97}) +[2026-06-24 00:52:37,940][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.010226551443338394, 'train/grad_norm_pre_clip': 9.742147445678711, 'timing/data': 0.0003659300273284316, 'timing/model': 1.2627440840005875, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 0.97}) +[2026-06-24 00:52:41,962][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.0064832232892513275, 'train/grad_norm_pre_clip': 9.989343643188477, 'timing/data': 0.0003979619359597564, 'timing/model': 1.1070434410357848, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 0.97}) +[2026-06-24 00:52:45,985][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.00992558617144823, 'train/grad_norm_pre_clip': 10.347855567932129, 'timing/data': 0.0003777319798246026, 'timing/model': 1.1028971929335967, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 0.97}) +[2026-06-24 00:52:50,008][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.01416830811649561, 'train/grad_norm_pre_clip': 38.53166198730469, 'timing/data': 0.0003689840668812394, 'timing/model': 1.106474008061923, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 0.97}) +[2026-06-24 00:52:54,171][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.011977944523096085, 'train/grad_norm_pre_clip': 11.120657920837402, 'timing/data': 0.0003753859782591462, 'timing/model': 1.2214951469795778, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 0.97}) +[2026-06-24 00:52:58,197][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.06727071851491928, 'train/grad_norm_pre_clip': 29.362802505493164, 'timing/data': 0.0005483209388330579, 'timing/model': 1.1029819910181686, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 0.97}) +[2026-06-24 00:53:02,235][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.050193049013614655, 'train/grad_norm_pre_clip': 8.274991035461426, 'timing/data': 0.0003558059688657522, 'timing/model': 1.1119292239891365, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 0.97}) +[2026-06-24 00:53:06,275][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.03738129511475563, 'train/grad_norm_pre_clip': 11.051239967346191, 'timing/data': 0.0003545960644260049, 'timing/model': 1.1088310308987275, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 0.97}) +[2026-06-24 00:53:10,457][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.041809190064668655, 'train/grad_norm_pre_clip': 45.415836334228516, 'timing/data': 0.0005826129345223308, 'timing/model': 1.1057663300307468, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 0.97}) +[2026-06-24 00:53:14,504][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.006253850180655718, 'train/grad_norm_pre_clip': 12.987618446350098, 'timing/data': 0.0003610580461099744, 'timing/model': 1.1044385900022462, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.97}) +[2026-06-24 00:53:18,558][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.0549662709236145, 'train/grad_norm_pre_clip': 19.440214157104492, 'timing/data': 0.0003637170884758234, 'timing/model': 1.1178191408980638, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 0.97}) +[2026-06-24 00:53:22,607][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.09171812236309052, 'train/grad_norm_pre_clip': 24.975830078125, 'timing/data': 0.00048556493129581213, 'timing/model': 1.1080850450089201, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 0.97}) +[2026-06-24 00:53:26,819][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.2537638545036316, 'train/grad_norm_pre_clip': 30.830543518066406, 'timing/data': 0.00035725696943700314, 'timing/model': 1.1081831639166921, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 0.97}) +[2026-06-24 00:53:30,879][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.07868773490190506, 'train/grad_norm_pre_clip': 19.403642654418945, 'timing/data': 0.00035623996518552303, 'timing/model': 1.1169156259857118, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 0.97}) +[2026-06-24 00:53:34,941][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.11831198632717133, 'train/grad_norm_pre_clip': 18.516956329345703, 'timing/data': 0.0003761210246011615, 'timing/model': 1.1142961010336876, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 0.98}) +[2026-06-24 00:53:39,006][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.015016291290521622, 'train/grad_norm_pre_clip': 13.425049781799316, 'timing/data': 0.00035225495230406523, 'timing/model': 1.1122155119664967, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 0.98}) +[2026-06-24 00:53:43,230][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.11817595362663269, 'train/grad_norm_pre_clip': 47.09444046020508, 'timing/data': 0.0003617439651861787, 'timing/model': 1.112947239074856, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 0.98}) +[2026-06-24 00:53:47,300][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.012524489313364029, 'train/grad_norm_pre_clip': 26.73996925354004, 'timing/data': 0.00046413205564022064, 'timing/model': 1.1147450929274783, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 0.98}) +[2026-06-24 00:53:51,354][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.013743274845182896, 'train/grad_norm_pre_clip': 14.152787208557129, 'timing/data': 0.0003563819918781519, 'timing/model': 1.1162162700202316, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 0.98}) +[2026-06-24 00:53:55,425][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.0032071706373244524, 'train/grad_norm_pre_clip': 9.946743965148926, 'timing/data': 0.0003697369247674942, 'timing/model': 1.1179745950503275, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 0.98}) +[2026-06-24 00:53:59,578][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.10061861574649811, 'train/grad_norm_pre_clip': 28.27320098876953, 'timing/data': 0.00042188900988548994, 'timing/model': 1.1130620109615847, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 0.98}) +[2026-06-24 00:54:03,655][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.1409001350402832, 'train/grad_norm_pre_clip': 15.419400215148926, 'timing/data': 0.000389202032238245, 'timing/model': 1.1318133710883558, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 0.98}) +[2026-06-24 00:54:07,713][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.0028364963363856077, 'train/grad_norm_pre_clip': 20.63422203063965, 'timing/data': 0.000568319926969707, 'timing/model': 1.110976657946594, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 0.98}) +[2026-06-24 00:54:11,781][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.03331417217850685, 'train/grad_norm_pre_clip': 18.890682220458984, 'timing/data': 0.00047967000864446163, 'timing/model': 1.111055830027908, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 0.98}) +[2026-06-24 00:54:15,916][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.012922841124236584, 'train/grad_norm_pre_clip': 31.212533950805664, 'timing/data': 0.0003684170078486204, 'timing/model': 1.1156721169827506, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 0.98}) +[2026-06-24 00:54:19,986][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.15299198031425476, 'train/grad_norm_pre_clip': 15.508179664611816, 'timing/data': 0.00035733100958168507, 'timing/model': 1.1171095460886136, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 0.98}) +[2026-06-24 00:54:24,062][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.014944188296794891, 'train/grad_norm_pre_clip': 13.725298881530762, 'timing/data': 0.0003915809793397784, 'timing/model': 1.1187375620938838, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 0.98}) +[2026-06-24 00:54:28,134][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.14174924790859222, 'train/grad_norm_pre_clip': 36.176361083984375, 'timing/data': 0.0003638459602370858, 'timing/model': 1.1185378580121323, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 0.98}) +[2026-06-24 00:54:32,346][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.005551890004426241, 'train/grad_norm_pre_clip': 14.805060386657715, 'timing/data': 0.00036364910192787647, 'timing/model': 1.114143589977175, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 0.98}) +[2026-06-24 00:54:36,417][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.02578617073595524, 'train/grad_norm_pre_clip': 20.520824432373047, 'timing/data': 0.0003578420728445053, 'timing/model': 1.1174593420000747, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 0.98}) +[2026-06-24 00:54:40,478][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.03173625469207764, 'train/grad_norm_pre_clip': 13.229641914367676, 'timing/data': 0.0003658219939097762, 'timing/model': 1.1111963510047644, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 0.98}) +[2026-06-24 00:54:44,650][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.1776871681213379, 'train/grad_norm_pre_clip': 22.58831214904785, 'timing/data': 0.00036643503699451685, 'timing/model': 1.207308211014606, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 0.98}) +[2026-06-24 00:54:48,741][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.044553641229867935, 'train/grad_norm_pre_clip': 9.803329467773438, 'timing/data': 0.00038276996929198503, 'timing/model': 1.1179969570366666, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 0.98}) +[2026-06-24 00:54:52,810][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.06582813709974289, 'train/grad_norm_pre_clip': 28.82115364074707, 'timing/data': 0.00035999808460474014, 'timing/model': 1.1153758550062776, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 0.98}) +[2026-06-24 00:54:56,884][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.2412164956331253, 'train/grad_norm_pre_clip': 34.029972076416016, 'timing/data': 0.0003592269495129585, 'timing/model': 1.1147964210249484, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 0.98}) +[2026-06-24 00:55:01,146][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.0004222889256197959, 'train/grad_norm_pre_clip': 14.41515827178955, 'timing/data': 0.00036482501309365034, 'timing/model': 1.296082470100373, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 0.98}) +[2026-06-24 00:55:05,223][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.05505084618926048, 'train/grad_norm_pre_clip': 22.28191375732422, 'timing/data': 0.0005881380056962371, 'timing/model': 1.1150289189536124, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 0.98}) +[2026-06-24 00:55:09,318][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.012076730839908123, 'train/grad_norm_pre_clip': 24.53533172607422, 'timing/data': 0.00037408608477562666, 'timing/model': 1.1172041969839483, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 0.98}) +[2026-06-24 00:55:13,395][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.09021754562854767, 'train/grad_norm_pre_clip': 17.935049057006836, 'timing/data': 0.00037028698716312647, 'timing/model': 1.1222256829496473, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 0.98}) +[2026-06-24 00:55:17,551][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.04640893265604973, 'train/grad_norm_pre_clip': 8.154424667358398, 'timing/data': 0.0005940059199929237, 'timing/model': 1.1176374899223447, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 0.98}) +[2026-06-24 00:55:21,632][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.03478942811489105, 'train/grad_norm_pre_clip': 24.23698616027832, 'timing/data': 0.0004023549845442176, 'timing/model': 1.117755284998566, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 0.99}) +[2026-06-24 00:55:25,705][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.0006424880702979863, 'train/grad_norm_pre_clip': 11.161370277404785, 'timing/data': 0.0003651670413091779, 'timing/model': 1.1177401520544663, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 0.99}) +[2026-06-24 00:55:29,795][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.1405758410692215, 'train/grad_norm_pre_clip': 16.460004806518555, 'timing/data': 0.00035669899079948664, 'timing/model': 1.1240440629189834, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 0.99}) +[2026-06-24 00:55:33,942][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.04870326817035675, 'train/grad_norm_pre_clip': 7.35233211517334, 'timing/data': 0.0004325900226831436, 'timing/model': 1.1132543629501015, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 0.99}) +[2026-06-24 00:55:38,014][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.04287349805235863, 'train/grad_norm_pre_clip': 21.529237747192383, 'timing/data': 0.0004067739937454462, 'timing/model': 1.1153151979669929, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.99}) +[2026-06-24 00:55:42,085][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.004854774568229914, 'train/grad_norm_pre_clip': 5.79697847366333, 'timing/data': 0.000365586020052433, 'timing/model': 1.1145192019175738, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 0.99}) +[2026-06-24 00:55:46,173][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.020558206364512444, 'train/grad_norm_pre_clip': 7.932310581207275, 'timing/data': 0.00035762402694672346, 'timing/model': 1.1196478460915387, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 0.99}) +[2026-06-24 00:55:50,387][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.09613429754972458, 'train/grad_norm_pre_clip': 12.470925331115723, 'timing/data': 0.00035333295818418264, 'timing/model': 1.1115506190108135, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 0.99}) +[2026-06-24 00:55:54,465][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.05650956183671951, 'train/grad_norm_pre_clip': 15.149679183959961, 'timing/data': 0.0004086639964953065, 'timing/model': 1.115955870016478, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 0.99}) +[2026-06-24 00:55:58,559][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.010580303147435188, 'train/grad_norm_pre_clip': 29.037324905395508, 'timing/data': 0.00039395701605826616, 'timing/model': 1.120453083072789, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 0.99}) +[2026-06-24 00:56:02,637][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.020770523697137833, 'train/grad_norm_pre_clip': 9.238486289978027, 'timing/data': 0.0003597229951992631, 'timing/model': 1.121127933030948, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 0.99}) +[2026-06-24 00:56:06,837][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.03412870317697525, 'train/grad_norm_pre_clip': 11.630464553833008, 'timing/data': 0.00035471899900585413, 'timing/model': 1.1202988629229367, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 0.99}) +[2026-06-24 00:56:10,916][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.0058373948559165, 'train/grad_norm_pre_clip': 10.73546028137207, 'timing/data': 0.00039035396184772253, 'timing/model': 1.1187227569753304, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 0.99}) +[2026-06-24 00:56:15,009][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.02219739742577076, 'train/grad_norm_pre_clip': 33.073829650878906, 'timing/data': 0.00036103103775531054, 'timing/model': 1.1190173890208825, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 0.99}) +[2026-06-24 00:56:19,096][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.009334838017821312, 'train/grad_norm_pre_clip': 16.504579544067383, 'timing/data': 0.00039408798329532146, 'timing/model': 1.1181455019395798, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 0.99}) +[2026-06-24 00:56:23,244][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.023661980405449867, 'train/grad_norm_pre_clip': 31.01173210144043, 'timing/data': 0.00036331708543002605, 'timing/model': 1.1175468859728426, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 0.99}) +[2026-06-24 00:56:27,323][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.03344335779547691, 'train/grad_norm_pre_clip': 24.735719680786133, 'timing/data': 0.0005642890464514494, 'timing/model': 1.117463396047242, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 0.99}) +[2026-06-24 00:56:31,395][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.005624564830213785, 'train/grad_norm_pre_clip': 31.54224967956543, 'timing/data': 0.0004157059593126178, 'timing/model': 1.114313306985423, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 0.99}) +[2026-06-24 00:56:35,467][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.10835479944944382, 'train/grad_norm_pre_clip': 27.40037727355957, 'timing/data': 0.0004868359537795186, 'timing/model': 1.1223240570398048, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 0.99}) +[2026-06-24 00:56:39,722][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.004752857610583305, 'train/grad_norm_pre_clip': 22.37105941772461, 'timing/data': 0.0004142209654673934, 'timing/model': 1.1141126659931615, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 0.99}) +[2026-06-24 00:56:43,798][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.02013237215578556, 'train/grad_norm_pre_clip': 26.939529418945312, 'timing/data': 0.0003623310476541519, 'timing/model': 1.1162579109659418, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 0.99}) +[2026-06-24 00:56:47,883][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.0036484147422015667, 'train/grad_norm_pre_clip': 14.068187713623047, 'timing/data': 0.00035147706512361765, 'timing/model': 1.118649333016947, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 0.99}) +[2026-06-24 00:56:51,975][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.0020460658706724644, 'train/grad_norm_pre_clip': 15.884098052978516, 'timing/data': 0.00036795204505324364, 'timing/model': 1.1214412580011413, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 0.99}) +[2026-06-24 00:56:56,237][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.017622200772166252, 'train/grad_norm_pre_clip': 23.84404754638672, 'timing/data': 0.0003553719725459814, 'timing/model': 1.134613291011192, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 0.99}) +[2026-06-24 00:57:00,342][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.101849764585495, 'train/grad_norm_pre_clip': 13.407991409301758, 'timing/data': 0.0003626110265031457, 'timing/model': 1.1225477850530297, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 0.99}) +[2026-06-24 00:57:04,433][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.10485853254795074, 'train/grad_norm_pre_clip': 40.6751594543457, 'timing/data': 0.00042447203304618597, 'timing/model': 1.1204356029629707, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 1.0}) +[2026-06-24 00:57:08,512][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.1188027486205101, 'train/grad_norm_pre_clip': 26.7348575592041, 'timing/data': 0.0004231471102684736, 'timing/model': 1.1205675509991124, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 1.0}) +[2026-06-24 00:57:12,761][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.09559845179319382, 'train/grad_norm_pre_clip': 25.21363639831543, 'timing/data': 0.00037252099718898535, 'timing/model': 1.1203167430358008, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 1.0}) +[2026-06-24 00:57:16,835][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.05325919762253761, 'train/grad_norm_pre_clip': 16.80082130432129, 'timing/data': 0.0003746909787878394, 'timing/model': 1.1142149739898741, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 1.0}) +[2026-06-24 00:57:20,912][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.0021898008417338133, 'train/grad_norm_pre_clip': 31.322147369384766, 'timing/data': 0.0003602689830586314, 'timing/model': 1.114874054910615, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 1.0}) +[2026-06-24 00:57:24,987][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.20579634606838226, 'train/grad_norm_pre_clip': 35.11927795410156, 'timing/data': 0.00044404598884284496, 'timing/model': 1.1033119020285085, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 1.0}) +[2026-06-24 00:57:29,024][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.0028263337444514036, 'train/grad_norm_pre_clip': 16.12093162536621, 'timing/data': 0.00038126390427351, 'timing/model': 1.1040023409295827, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 1.0}) +[2026-06-24 00:57:33,038][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.13530592620372772, 'train/grad_norm_pre_clip': 34.8784065246582, 'timing/data': 0.000438741990365088, 'timing/model': 1.1011027980130166, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 1.0}) +[2026-06-24 00:57:37,056][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.01148450281471014, 'train/grad_norm_pre_clip': 15.333788871765137, 'timing/data': 0.0005105950403958559, 'timing/model': 1.1030740049900487, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 1.0}) +[2026-06-24 00:57:41,080][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.023967504501342773, 'train/grad_norm_pre_clip': 9.970354080200195, 'timing/data': 0.0005016339709982276, 'timing/model': 1.1032760910457, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 1.0}) +[2026-06-24 00:57:45,214][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.058096785098314285, 'train/grad_norm_pre_clip': 14.097037315368652, 'timing/data': 0.0003954049898311496, 'timing/model': 1.1020648749545217, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 1.0}) +[2026-06-24 00:57:49,236][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.056131456047296524, 'train/grad_norm_pre_clip': 27.766279220581055, 'timing/data': 0.00036736601032316685, 'timing/model': 1.1021245039300993, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 1.0}) +[2026-06-24 00:57:53,253][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.04443570598959923, 'train/grad_norm_pre_clip': 16.983295440673828, 'timing/data': 0.0003648389829322696, 'timing/model': 1.1030450130347162, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 1.0}) +[2026-06-24 00:57:56,725][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.04771675169467926, 'train/grad_norm_pre_clip': 15.450471878051758, 'timing/data': 0.00025220599491149187, 'timing/model': 0.555250364006497, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 1.0}) +[2026-06-24 00:58:06,082][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.024874422699213028, 'train/grad_norm_pre_clip': 13.457356452941895, 'timing/data': 0.001502164057455957, 'timing/model': 1.7427679160609841, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 1.0}) +[2026-06-24 00:58:11,158][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.036054328083992004, 'train/grad_norm_pre_clip': 10.342787742614746, 'timing/data': 0.0016051200218498707, 'timing/model': 1.2091946570435539, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 1.0}) +[2026-06-24 00:58:15,236][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.047559063881635666, 'train/grad_norm_pre_clip': 13.739526748657227, 'timing/data': 0.00034448597580194473, 'timing/model': 1.113382774987258, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 1.0}) +[2026-06-24 00:58:19,311][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.0452449806034565, 'train/grad_norm_pre_clip': 20.747777938842773, 'timing/data': 0.0003697139909490943, 'timing/model': 1.117705435026437, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 1.0}) +[2026-06-24 00:58:23,520][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.02496793493628502, 'train/grad_norm_pre_clip': 8.427306175231934, 'timing/data': 0.0004322649911046028, 'timing/model': 1.1140115750022233, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 1.0}) +[2026-06-24 00:58:27,597][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.061167068779468536, 'train/grad_norm_pre_clip': 18.11009407043457, 'timing/data': 0.0003401930443942547, 'timing/model': 1.114246545941569, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 1.0}) +[2026-06-24 00:58:31,675][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.003910473547875881, 'train/grad_norm_pre_clip': 11.661253929138184, 'timing/data': 0.0003319979878142476, 'timing/model': 1.1181177530670539, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 1.0}) +[2026-06-24 00:58:35,824][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.028211088851094246, 'train/grad_norm_pre_clip': 10.336670875549316, 'timing/data': 0.00036656204611063004, 'timing/model': 1.188027019961737, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 1.0}) +[2026-06-24 00:58:39,886][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.07164692878723145, 'train/grad_norm_pre_clip': 26.061851501464844, 'timing/data': 0.0003386229509487748, 'timing/model': 1.1124604979995638, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 1.0}) +[2026-06-24 00:58:43,952][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.08161723613739014, 'train/grad_norm_pre_clip': 29.415063858032227, 'timing/data': 0.0003609179984778166, 'timing/model': 1.1116762239253148, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 1.0}) +[2026-06-24 00:58:48,039][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.0007379193557426333, 'train/grad_norm_pre_clip': 16.92443084716797, 'timing/data': 0.00034811708610504866, 'timing/model': 1.123275795020163, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 1.0}) +[2026-06-24 00:58:52,116][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.011575326323509216, 'train/grad_norm_pre_clip': 11.891342163085938, 'timing/data': 0.000655472045764327, 'timing/model': 1.1204651460284367, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 1.0}) +[2026-06-24 00:58:56,317][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.007684438489377499, 'train/grad_norm_pre_clip': 20.32952308654785, 'timing/data': 0.0004020779160782695, 'timing/model': 1.1163255639839917, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 1.01}) +[2026-06-24 00:59:00,385][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.004124811384826899, 'train/grad_norm_pre_clip': 23.447580337524414, 'timing/data': 0.0004131239838898182, 'timing/model': 1.114209005027078, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 1.01}) +[2026-06-24 00:59:04,460][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.017252609133720398, 'train/grad_norm_pre_clip': 13.141400337219238, 'timing/data': 0.00034044496715068817, 'timing/model': 1.1196492851013318, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.01}) +[2026-06-24 00:59:08,586][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2600 +[2026-06-24 01:00:13,013][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2600 +[2026-06-24 01:00:13,015][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2600 +[2026-06-24 01:00:52,164][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2600 +[2026-06-24 01:00:52,164][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.005533845163881779, 'train/grad_norm_pre_clip': 12.279487609863281, 'eval/action_loss/seconds': 64.42578664701432, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03448816388845444, 'eval/latency_2/loss': 0.03448816388845444, 'eval/flappy/loss': 0.03448816388845444, 'eval/flappy/latency_2/loss': 0.03448816388845444, 'eval/flap_precision': 0.9587217043941412, 'eval/flap_recall': 0.9769335142469471, 'eval/flap_f1': 0.967741935483871, 'eval/flap_rate_model': 0.2086111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9677419354838711, 'eval/latency_2/flap_precision': 0.9587217043941412, 'eval/latency_2/flap_recall': 0.9769335142469471, 'eval/latency_2/flap_f1': 0.967741935483871, 'eval/latency_2/flap_rate_model': 0.2086111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9677419354838711, 'eval/action_classification/data_wait_seconds': 0.5184364318847656, 'eval/action_classification/predict_seconds': 38.450782775878906, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0005389299476519227, 'timing/model': 1.1337281829910353, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 1.01}) +[2026-06-24 01:00:56,211][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.05722351744771004, 'train/grad_norm_pre_clip': 34.93680191040039, 'timing/data': 0.0005939699476584792, 'timing/model': 1.1219558790326118, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 1.01}) +[2026-06-24 01:01:00,249][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.05665524676442146, 'train/grad_norm_pre_clip': 22.247211456298828, 'timing/data': 0.0003442770102992654, 'timing/model': 1.1045064850477502, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 1.01}) +[2026-06-24 01:01:04,292][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.028883617371320724, 'train/grad_norm_pre_clip': 31.77842903137207, 'timing/data': 0.0003563240170478821, 'timing/model': 1.1061629190808162, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 1.01}) +[2026-06-24 01:01:08,473][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.06606971472501755, 'train/grad_norm_pre_clip': 13.402463912963867, 'timing/data': 0.0005462169647216797, 'timing/model': 1.1084924710448831, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 1.01}) +[2026-06-24 01:01:12,536][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.09632778912782669, 'train/grad_norm_pre_clip': 22.382680892944336, 'timing/data': 0.00034697994124144316, 'timing/model': 1.1201281539397314, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 1.01}) +[2026-06-24 01:01:16,599][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.02562907338142395, 'train/grad_norm_pre_clip': 22.086271286010742, 'timing/data': 0.0003534459974616766, 'timing/model': 1.1184501450043172, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 1.01}) +[2026-06-24 01:01:20,684][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.05961732193827629, 'train/grad_norm_pre_clip': 24.33222198486328, 'timing/data': 0.0003349150065332651, 'timing/model': 1.1275230019818991, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 1.01}) +[2026-06-24 01:01:24,918][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.02910684235394001, 'train/grad_norm_pre_clip': 29.69812774658203, 'timing/data': 0.0003412399673834443, 'timing/model': 1.1101770639652386, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 1.01}) +[2026-06-24 01:01:29,009][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.006209703162312508, 'train/grad_norm_pre_clip': 20.88037872314453, 'timing/data': 0.00035263702739030123, 'timing/model': 1.1114817009074613, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 1.01}) +[2026-06-24 01:01:33,066][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.03521663695573807, 'train/grad_norm_pre_clip': 11.493988037109375, 'timing/data': 0.0003976779989898205, 'timing/model': 1.1164458650164306, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 1.01}) +[2026-06-24 01:01:37,122][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.04957728832960129, 'train/grad_norm_pre_clip': 10.840059280395508, 'timing/data': 0.0005165379261597991, 'timing/model': 1.1115035259863362, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 1.01}) +[2026-06-24 01:01:41,346][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.06537164747714996, 'train/grad_norm_pre_clip': 17.6405086517334, 'timing/data': 0.0004278329433873296, 'timing/model': 1.1140800609719008, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 1.01}) +[2026-06-24 01:01:45,399][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.06411218643188477, 'train/grad_norm_pre_clip': 16.9986515045166, 'timing/data': 0.0004123279359191656, 'timing/model': 1.1105045800795779, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 1.01}) +[2026-06-24 01:01:49,460][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.03133172541856766, 'train/grad_norm_pre_clip': 24.738035202026367, 'timing/data': 0.00036084908060729504, 'timing/model': 1.1131592589663342, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 1.01}) +[2026-06-24 01:01:53,607][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.11241552233695984, 'train/grad_norm_pre_clip': 26.761537551879883, 'timing/data': 0.0003722610417753458, 'timing/model': 1.201827839948237, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 1.01}) +[2026-06-24 01:01:57,692][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.09246501326560974, 'train/grad_norm_pre_clip': 25.31365966796875, 'timing/data': 0.00039556703995913267, 'timing/model': 1.1141931270249188, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 1.01}) +[2026-06-24 01:02:01,764][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.025370873510837555, 'train/grad_norm_pre_clip': 10.47726058959961, 'timing/data': 0.0003975980216637254, 'timing/model': 1.1110337839927524, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 1.01}) +[2026-06-24 01:02:05,830][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.09811680018901825, 'train/grad_norm_pre_clip': 15.806171417236328, 'timing/data': 0.0003593970322981477, 'timing/model': 1.1144935679621994, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 1.01}) +[2026-06-24 01:02:09,953][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.03265484794974327, 'train/grad_norm_pre_clip': 11.731252670288086, 'timing/data': 0.00039962795563042164, 'timing/model': 1.182603755965829, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 1.01}) +[2026-06-24 01:02:14,014][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.023234739899635315, 'train/grad_norm_pre_clip': 22.207910537719727, 'timing/data': 0.0004762099124491215, 'timing/model': 1.1109491409733891, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 1.01}) +[2026-06-24 01:02:18,091][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.14943251013755798, 'train/grad_norm_pre_clip': 19.64861297607422, 'timing/data': 0.0005717669846490026, 'timing/model': 1.1167678770143539, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 1.01}) +[2026-06-24 01:02:22,157][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.02761635184288025, 'train/grad_norm_pre_clip': 31.777406692504883, 'timing/data': 0.00043068896047770977, 'timing/model': 1.1096981529844925, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 1.01}) +[2026-06-24 01:02:26,358][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.07707368582487106, 'train/grad_norm_pre_clip': 15.128042221069336, 'timing/data': 0.00035801506601274014, 'timing/model': 1.2469730289885774, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 1.02}) +[2026-06-24 01:02:30,430][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.06591618061065674, 'train/grad_norm_pre_clip': 24.55669403076172, 'timing/data': 0.00035716197453439236, 'timing/model': 1.1190378339961171, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 1.02}) +[2026-06-24 01:02:34,523][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.006403052248060703, 'train/grad_norm_pre_clip': 13.870630264282227, 'timing/data': 0.0005421019159257412, 'timing/model': 1.1252933669602498, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 1.02}) +[2026-06-24 01:02:38,585][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.10358791798353195, 'train/grad_norm_pre_clip': 32.1469612121582, 'timing/data': 0.00046658702194690704, 'timing/model': 1.1152616509934887, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 1.02}) +[2026-06-24 01:02:42,774][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.0019977842457592487, 'train/grad_norm_pre_clip': 35.25979995727539, 'timing/data': 0.0005453560734167695, 'timing/model': 1.1164858300471678, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 1.02}) +[2026-06-24 01:02:46,844][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.010219148360192776, 'train/grad_norm_pre_clip': 18.905668258666992, 'timing/data': 0.0003775280201807618, 'timing/model': 1.1161419429117814, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 1.02}) +[2026-06-24 01:02:50,917][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.020926525816321373, 'train/grad_norm_pre_clip': 8.88024616241455, 'timing/data': 0.0003598639741539955, 'timing/model': 1.1131013230187818, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 1.02}) +[2026-06-24 01:02:54,997][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.11297017335891724, 'train/grad_norm_pre_clip': 24.047557830810547, 'timing/data': 0.00041086203418672085, 'timing/model': 1.1205924240639433, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 1.02}) +[2026-06-24 01:02:59,261][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.01217638049274683, 'train/grad_norm_pre_clip': 25.33684539794922, 'timing/data': 0.00047964195255190134, 'timing/model': 1.1229517180472612, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 1.02}) +[2026-06-24 01:03:03,324][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.06107085943222046, 'train/grad_norm_pre_clip': 13.639299392700195, 'timing/data': 0.0006018229760229588, 'timing/model': 1.1157627799548209, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 1.02}) +[2026-06-24 01:03:07,395][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.01753445714712143, 'train/grad_norm_pre_clip': 17.43202018737793, 'timing/data': 0.0003487759968265891, 'timing/model': 1.112836241023615, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 1.02}) +[2026-06-24 01:03:11,467][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.05255687236785889, 'train/grad_norm_pre_clip': 29.35973358154297, 'timing/data': 0.0005485860165208578, 'timing/model': 1.1171444030478597, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 1.02}) +[2026-06-24 01:03:15,723][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.054762523621320724, 'train/grad_norm_pre_clip': 18.863245010375977, 'timing/data': 0.0003521289909258485, 'timing/model': 1.1132056129863486, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 1.02}) +[2026-06-24 01:03:19,812][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.15626460313796997, 'train/grad_norm_pre_clip': 23.49655532836914, 'timing/data': 0.0003486230270937085, 'timing/model': 1.1126768470276147, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 1.02}) +[2026-06-24 01:03:23,900][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.06124628335237503, 'train/grad_norm_pre_clip': 20.355567932128906, 'timing/data': 0.00040950800757855177, 'timing/model': 1.1233878060011193, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 1.02}) +[2026-06-24 01:03:27,971][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.07514378428459167, 'train/grad_norm_pre_clip': 21.76032066345215, 'timing/data': 0.0003977799788117409, 'timing/model': 1.112772056949325, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 1.02}) +[2026-06-24 01:03:32,113][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.11238846927881241, 'train/grad_norm_pre_clip': 40.04779052734375, 'timing/data': 0.00041959795635193586, 'timing/model': 1.10919633298181, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 1.02}) +[2026-06-24 01:03:36,181][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.12062432616949081, 'train/grad_norm_pre_clip': 29.100967407226562, 'timing/data': 0.0003631958970800042, 'timing/model': 1.1152130879927427, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 1.02}) +[2026-06-24 01:03:40,257][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.012336505576968193, 'train/grad_norm_pre_clip': 10.04883098602295, 'timing/data': 0.0005588809726759791, 'timing/model': 1.1187251540832222, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 1.02}) +[2026-06-24 01:03:44,338][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.027385808527469635, 'train/grad_norm_pre_clip': 14.17053508758545, 'timing/data': 0.00035775406286120415, 'timing/model': 1.1191302658990026, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 1.02}) +[2026-06-24 01:03:48,552][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.012173546478152275, 'train/grad_norm_pre_clip': 9.766692161560059, 'timing/data': 0.0003509810194373131, 'timing/model': 1.1161492499522865, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 1.02}) +[2026-06-24 01:03:52,637][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.0030062764417380095, 'train/grad_norm_pre_clip': 9.05623722076416, 'timing/data': 0.0003538640448823571, 'timing/model': 1.1190294000552967, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 1.02}) +[2026-06-24 01:03:56,716][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.06971336156129837, 'train/grad_norm_pre_clip': 11.533136367797852, 'timing/data': 0.0004913440207019448, 'timing/model': 1.1171206090366468, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 1.02}) +[2026-06-24 01:04:00,790][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.030441151931881905, 'train/grad_norm_pre_clip': 16.33209991455078, 'timing/data': 0.0006168470717966557, 'timing/model': 1.1143983110086992, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 1.02}) +[2026-06-24 01:04:04,992][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.06590807437896729, 'train/grad_norm_pre_clip': 13.76793098449707, 'timing/data': 0.000355785945430398, 'timing/model': 1.120633199927397, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 1.02}) +[2026-06-24 01:04:09,089][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.03800757974386215, 'train/grad_norm_pre_clip': 23.65610694885254, 'timing/data': 0.00039029098115861416, 'timing/model': 1.123511562938802, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.02}) +[2026-06-24 01:04:13,176][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.12279053777456284, 'train/grad_norm_pre_clip': 13.144217491149902, 'timing/data': 0.00035397696774452925, 'timing/model': 1.1129985420266166, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 1.03}) +[2026-06-24 01:04:17,375][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.023194855079054832, 'train/grad_norm_pre_clip': 16.093629837036133, 'timing/data': 0.0003600030904635787, 'timing/model': 1.226487535983324, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 1.03}) +[2026-06-24 01:04:21,472][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.02210664004087448, 'train/grad_norm_pre_clip': 8.058226585388184, 'timing/data': 0.00034768390469253063, 'timing/model': 1.1208338740980253, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 1.03}) +[2026-06-24 01:04:25,541][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.015039641410112381, 'train/grad_norm_pre_clip': 18.580774307250977, 'timing/data': 0.0004282519221305847, 'timing/model': 1.1164477479178458, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 1.03}) +[2026-06-24 01:04:29,628][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.14259210228919983, 'train/grad_norm_pre_clip': 47.079246520996094, 'timing/data': 0.0003530409885570407, 'timing/model': 1.1182600969914347, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 1.03}) +[2026-06-24 01:04:33,704][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.012358342297375202, 'train/grad_norm_pre_clip': 12.885903358459473, 'timing/data': 0.0006201560609042645, 'timing/model': 1.1156188059831038, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 1.03}) +[2026-06-24 01:04:37,908][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.2595868706703186, 'train/grad_norm_pre_clip': 26.763980865478516, 'timing/data': 0.00034932210110127926, 'timing/model': 1.1163467019796371, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 1.03}) +[2026-06-24 01:04:42,023][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.02098633348941803, 'train/grad_norm_pre_clip': 17.41442108154297, 'timing/data': 0.00035716197453439236, 'timing/model': 1.124415613943711, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 1.03}) +[2026-06-24 01:04:46,090][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.18608589470386505, 'train/grad_norm_pre_clip': 24.189922332763672, 'timing/data': 0.00036606297362595797, 'timing/model': 1.1093296250328422, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 1.03}) +[2026-06-24 01:04:50,164][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.13232949376106262, 'train/grad_norm_pre_clip': 18.131948471069336, 'timing/data': 0.0006303261034190655, 'timing/model': 1.1183818919816986, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 1.03}) +[2026-06-24 01:04:54,407][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.017306411638855934, 'train/grad_norm_pre_clip': 16.684476852416992, 'timing/data': 0.0003584860824048519, 'timing/model': 1.1269234059145674, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 1.03}) +[2026-06-24 01:04:58,497][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.11757060885429382, 'train/grad_norm_pre_clip': 22.09564208984375, 'timing/data': 0.00035379803739488125, 'timing/model': 1.1172976520610973, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 1.03}) +[2026-06-24 01:05:02,570][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.03957551717758179, 'train/grad_norm_pre_clip': 15.66084098815918, 'timing/data': 0.0004177780356258154, 'timing/model': 1.117529634037055, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 1.03}) +[2026-06-24 01:05:06,644][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.19360758364200592, 'train/grad_norm_pre_clip': 23.54336929321289, 'timing/data': 0.0003694209735840559, 'timing/model': 1.1203704290091991, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 1.03}) +[2026-06-24 01:05:10,783][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.13437233865261078, 'train/grad_norm_pre_clip': 14.460314750671387, 'timing/data': 0.00039058097172528505, 'timing/model': 1.1138354999711737, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 1.03}) +[2026-06-24 01:05:14,861][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.10322117805480957, 'train/grad_norm_pre_clip': 15.54909610748291, 'timing/data': 0.000349098932929337, 'timing/model': 1.119315215968527, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 1.03}) +[2026-06-24 01:05:18,937][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.06905563920736313, 'train/grad_norm_pre_clip': 15.557831764221191, 'timing/data': 0.0003669330617412925, 'timing/model': 1.1165186410071328, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 1.03}) +[2026-06-24 01:05:23,017][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.040883347392082214, 'train/grad_norm_pre_clip': 18.372642517089844, 'timing/data': 0.00036565400660037994, 'timing/model': 1.1188560459995642, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 1.03}) +[2026-06-24 01:05:27,156][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.10499744117259979, 'train/grad_norm_pre_clip': 34.844078063964844, 'timing/data': 0.0004140199162065983, 'timing/model': 1.1163762409705669, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 1.03}) +[2026-06-24 01:05:31,240][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.032191477715969086, 'train/grad_norm_pre_clip': 7.17968225479126, 'timing/data': 0.0006272780010476708, 'timing/model': 1.1180670990142971, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 1.03}) +[2026-06-24 01:05:35,311][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.05901022255420685, 'train/grad_norm_pre_clip': 21.844768524169922, 'timing/data': 0.00035824801307171583, 'timing/model': 1.1175203749444336, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 1.03}) +[2026-06-24 01:05:39,403][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.0148581238463521, 'train/grad_norm_pre_clip': 15.19015121459961, 'timing/data': 0.0003548200475051999, 'timing/model': 1.118315606028773, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 1.03}) +[2026-06-24 01:05:43,633][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.07869882136583328, 'train/grad_norm_pre_clip': 23.411401748657227, 'timing/data': 0.0003581549972295761, 'timing/model': 1.116416109027341, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 1.03}) +[2026-06-24 01:05:47,707][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.10471431910991669, 'train/grad_norm_pre_clip': 16.615182876586914, 'timing/data': 0.0005040550604462624, 'timing/model': 1.1165211430052295, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 1.03}) +[2026-06-24 01:05:51,805][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.11227700114250183, 'train/grad_norm_pre_clip': 15.972935676574707, 'timing/data': 0.0003513399278745055, 'timing/model': 1.118575138039887, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 1.03}) +[2026-06-24 01:05:55,913][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.004751995205879211, 'train/grad_norm_pre_clip': 18.15581703186035, 'timing/data': 0.00035589304752647877, 'timing/model': 1.1252457529772073, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 1.03}) +[2026-06-24 01:06:00,171][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.027145911008119583, 'train/grad_norm_pre_clip': 11.457498550415039, 'timing/data': 0.00035823602229356766, 'timing/model': 1.1279425619868562, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 1.04}) +[2026-06-24 01:06:04,270][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.2719733417034149, 'train/grad_norm_pre_clip': 26.701807022094727, 'timing/data': 0.0003745750291272998, 'timing/model': 1.1239319340093061, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 1.04}) +[2026-06-24 01:06:08,354][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.022799570113420486, 'train/grad_norm_pre_clip': 17.23868179321289, 'timing/data': 0.0005643079057335854, 'timing/model': 1.1218577469699085, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 1.04}) +[2026-06-24 01:06:12,438][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.01719042658805847, 'train/grad_norm_pre_clip': 14.823307991027832, 'timing/data': 0.0006354700308293104, 'timing/model': 1.117549804970622, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 1.04}) +[2026-06-24 01:06:16,693][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.008456873707473278, 'train/grad_norm_pre_clip': 19.47016716003418, 'timing/data': 0.0004325699992477894, 'timing/model': 1.119130081962794, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 1.04}) +[2026-06-24 01:06:20,775][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.11506767570972443, 'train/grad_norm_pre_clip': 24.416080474853516, 'timing/data': 0.0005446560680866241, 'timing/model': 1.1127474108943716, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 1.04}) +[2026-06-24 01:06:24,849][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.040775079280138016, 'train/grad_norm_pre_clip': 14.960982322692871, 'timing/data': 0.0005117590771988034, 'timing/model': 1.1134591129375622, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 1.04}) +[2026-06-24 01:06:28,926][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.021163851022720337, 'train/grad_norm_pre_clip': 17.73440170288086, 'timing/data': 0.0003801310667768121, 'timing/model': 1.1151610799133778, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 1.04}) +[2026-06-24 01:06:33,173][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.0716751292347908, 'train/grad_norm_pre_clip': 14.067842483520508, 'timing/data': 0.0004894220037385821, 'timing/model': 1.1177943899529055, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 1.04}) +[2026-06-24 01:06:37,265][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.08361335098743439, 'train/grad_norm_pre_clip': 11.802484512329102, 'timing/data': 0.0003544209757819772, 'timing/model': 1.1242040720535442, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 1.04}) +[2026-06-24 01:06:41,349][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.026021374389529228, 'train/grad_norm_pre_clip': 8.35560131072998, 'timing/data': 0.0003525980282574892, 'timing/model': 1.1233021470252424, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 1.04}) +[2026-06-24 01:06:45,425][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.004828234668821096, 'train/grad_norm_pre_clip': 15.036321640014648, 'timing/data': 0.00035545905120670795, 'timing/model': 1.11676322191488, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 1.04}) +[2026-06-24 01:06:49,684][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.11309374123811722, 'train/grad_norm_pre_clip': 18.447834014892578, 'timing/data': 0.00039973692037165165, 'timing/model': 1.1172754539875314, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 1.04}) +[2026-06-24 01:06:53,760][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.07418351620435715, 'train/grad_norm_pre_clip': 17.86300277709961, 'timing/data': 0.0003579279873520136, 'timing/model': 1.1181767829693854, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 1.04}) +[2026-06-24 01:06:57,837][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.011214175261557102, 'train/grad_norm_pre_clip': 22.758541107177734, 'timing/data': 0.00041380792390555143, 'timing/model': 1.1128708929754794, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 1.04}) +[2026-06-24 01:07:01,930][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.0480397567152977, 'train/grad_norm_pre_clip': 17.650726318359375, 'timing/data': 0.0005620178999379277, 'timing/model': 1.1210379200056195, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 1.04}) +[2026-06-24 01:07:06,196][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.07050016522407532, 'train/grad_norm_pre_clip': 12.085271835327148, 'timing/data': 0.0003523360937833786, 'timing/model': 1.1201117429882288, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 1.04}) +[2026-06-24 01:07:10,305][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.014011151157319546, 'train/grad_norm_pre_clip': 12.711782455444336, 'timing/data': 0.0003504889318719506, 'timing/model': 1.1352735350374132, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 1.04}) +[2026-06-24 01:07:14,375][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.19130249321460724, 'train/grad_norm_pre_clip': 15.064787864685059, 'timing/data': 0.0003940409515053034, 'timing/model': 1.1119813550030813, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 1.04}) +[2026-06-24 01:07:18,456][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.005623797886073589, 'train/grad_norm_pre_clip': 28.207368850708008, 'timing/data': 0.00039769895374774933, 'timing/model': 1.1190781319746748, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 1.04}) +[2026-06-24 01:07:22,695][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.0012489191722124815, 'train/grad_norm_pre_clip': 17.406749725341797, 'timing/data': 0.00036292197182774544, 'timing/model': 1.1145417499355972, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 1.04}) +[2026-06-24 01:07:26,775][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.00748257152736187, 'train/grad_norm_pre_clip': 9.941411018371582, 'timing/data': 0.00041196902748197317, 'timing/model': 1.1292536490364, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 1.04}) +[2026-06-24 01:07:30,864][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.013864465057849884, 'train/grad_norm_pre_clip': 19.793380737304688, 'timing/data': 0.0003518969751894474, 'timing/model': 1.1185313418973237, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.04}) +[2026-06-24 01:07:34,951][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.029034404084086418, 'train/grad_norm_pre_clip': 17.663414001464844, 'timing/data': 0.0003453940153121948, 'timing/model': 1.1211770829977468, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 1.04}) +[2026-06-24 01:07:39,202][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.032931312918663025, 'train/grad_norm_pre_clip': 25.24732208251953, 'timing/data': 0.00041820399928838015, 'timing/model': 1.117932719993405, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 1.04}) +[2026-06-24 01:07:43,283][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2700 +[2026-06-24 01:08:46,606][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2700 +[2026-06-24 01:08:46,608][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2700 +[2026-06-24 01:09:25,992][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2700 +[2026-06-24 01:09:25,993][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.04854331538081169, 'train/grad_norm_pre_clip': 18.806415557861328, 'eval/action_loss/seconds': 63.32089786906727, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03655243292450905, 'eval/latency_2/loss': 0.03655243292450905, 'eval/flappy/loss': 0.03655243292450905, 'eval/flappy/latency_2/loss': 0.03655243292450905, 'eval/flap_precision': 0.9781121751025992, 'eval/flap_recall': 0.9701492537313433, 'eval/flap_f1': 0.9741144414168937, 'eval/flap_rate_model': 0.20305555555555554, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9741144414168937, 'eval/latency_2/flap_precision': 0.9781121751025992, 'eval/latency_2/flap_recall': 0.9701492537313433, 'eval/latency_2/flap_f1': 0.9741144414168937, 'eval/latency_2/flap_rate_model': 0.20305555555555554, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9741144414168937, 'eval/action_classification/data_wait_seconds': 0.4731844663619995, 'eval/action_classification/predict_seconds': 38.7834587097168, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.000552144949324429, 'timing/model': 1.1116101760417223, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 1.04}) +[2026-06-24 01:09:30,026][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.052453406155109406, 'train/grad_norm_pre_clip': 34.08510208129883, 'timing/data': 0.0003900789888575673, 'timing/model': 1.1058320649899542, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 1.05}) +[2026-06-24 01:09:34,255][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.056339409202337265, 'train/grad_norm_pre_clip': 8.855311393737793, 'timing/data': 0.000363157014362514, 'timing/model': 1.2928358289645985, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 1.05}) +[2026-06-24 01:09:38,295][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.14497976005077362, 'train/grad_norm_pre_clip': 24.334056854248047, 'timing/data': 0.00043024204205721617, 'timing/model': 1.1085078320465982, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 1.05}) +[2026-06-24 01:09:42,345][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.016334403306245804, 'train/grad_norm_pre_clip': 22.89148712158203, 'timing/data': 0.0006826519966125488, 'timing/model': 1.1127110869856551, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 1.05}) +[2026-06-24 01:09:46,397][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.029576221480965614, 'train/grad_norm_pre_clip': 10.467534065246582, 'timing/data': 0.0003453929675742984, 'timing/model': 1.1092578070238233, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 1.05}) +[2026-06-24 01:09:50,518][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.05855579674243927, 'train/grad_norm_pre_clip': 20.145252227783203, 'timing/data': 0.0006046020425856113, 'timing/model': 1.1116851749829948, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 1.05}) +[2026-06-24 01:09:54,566][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.16917233169078827, 'train/grad_norm_pre_clip': 20.624221801757812, 'timing/data': 0.0006732880137860775, 'timing/model': 1.1108306689420715, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 1.05}) +[2026-06-24 01:09:58,623][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.0022735551465302706, 'train/grad_norm_pre_clip': 14.774285316467285, 'timing/data': 0.00046369305346161127, 'timing/model': 1.1130852949572727, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 1.05}) +[2026-06-24 01:10:02,680][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.163628950715065, 'train/grad_norm_pre_clip': 26.567041397094727, 'timing/data': 0.0003731769975274801, 'timing/model': 1.1177488549146801, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 1.05}) +[2026-06-24 01:10:06,895][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.013786407187581062, 'train/grad_norm_pre_clip': 22.348703384399414, 'timing/data': 0.00035063200630247593, 'timing/model': 1.1179705230752006, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 1.05}) +[2026-06-24 01:10:10,949][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.005697460845112801, 'train/grad_norm_pre_clip': 15.617815017700195, 'timing/data': 0.0003791309427469969, 'timing/model': 1.109192012110725, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 1.05}) +[2026-06-24 01:10:15,007][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.058030810207128525, 'train/grad_norm_pre_clip': 23.08544158935547, 'timing/data': 0.0005832030437886715, 'timing/model': 1.1090470099588856, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 1.05}) +[2026-06-24 01:10:19,059][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.018355753272771835, 'train/grad_norm_pre_clip': 30.885177612304688, 'timing/data': 0.0006744180573150516, 'timing/model': 1.1121572880074382, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 1.05}) +[2026-06-24 01:10:23,234][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.024369297549128532, 'train/grad_norm_pre_clip': 7.506808757781982, 'timing/data': 0.0004610030446201563, 'timing/model': 1.1136582150356844, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 1.05}) +[2026-06-24 01:10:27,297][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.09450434893369675, 'train/grad_norm_pre_clip': 18.873172760009766, 'timing/data': 0.00040020805317908525, 'timing/model': 1.1135793270077556, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 1.05}) +[2026-06-24 01:10:31,373][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.006572023965418339, 'train/grad_norm_pre_clip': 15.008024215698242, 'timing/data': 0.00035352795384824276, 'timing/model': 1.1171449550893158, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 1.05}) +[2026-06-24 01:10:35,443][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.010220831260085106, 'train/grad_norm_pre_clip': 15.425202369689941, 'timing/data': 0.0005524719599634409, 'timing/model': 1.1124003730947152, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 1.05}) +[2026-06-24 01:10:39,578][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.07586657255887985, 'train/grad_norm_pre_clip': 7.748843669891357, 'timing/data': 0.0003515079151839018, 'timing/model': 1.1171873200219125, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 1.05}) +[2026-06-24 01:10:43,653][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.0982639342546463, 'train/grad_norm_pre_clip': 29.940303802490234, 'timing/data': 0.0003432569792494178, 'timing/model': 1.1145893060602248, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 1.05}) +[2026-06-24 01:10:47,725][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.0001492941373726353, 'train/grad_norm_pre_clip': 40.24325180053711, 'timing/data': 0.0003808559849858284, 'timing/model': 1.1108949090121314, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 1.05}) +[2026-06-24 01:10:51,799][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.051366522908210754, 'train/grad_norm_pre_clip': 51.10357666015625, 'timing/data': 0.0003467510687187314, 'timing/model': 1.1140093998983502, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 1.05}) +[2026-06-24 01:10:56,043][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.00557594234123826, 'train/grad_norm_pre_clip': 20.14579963684082, 'timing/data': 0.00034694396890699863, 'timing/model': 1.1118648499250412, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 1.05}) +[2026-06-24 01:11:00,129][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.06451187282800674, 'train/grad_norm_pre_clip': 37.338050842285156, 'timing/data': 0.0004507809644564986, 'timing/model': 1.1148207649821416, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 1.05}) +[2026-06-24 01:11:04,227][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.02114064246416092, 'train/grad_norm_pre_clip': 7.914138317108154, 'timing/data': 0.0003643740201368928, 'timing/model': 1.1201738870004192, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 1.05}) +[2026-06-24 01:11:08,301][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.1027534231543541, 'train/grad_norm_pre_clip': 29.440153121948242, 'timing/data': 0.0004348469665274024, 'timing/model': 1.111330026993528, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 1.05}) +[2026-06-24 01:11:12,488][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.009505825117230415, 'train/grad_norm_pre_clip': 21.028532028198242, 'timing/data': 0.0003384340088814497, 'timing/model': 1.1170241619693115, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 1.06}) +[2026-06-24 01:11:16,573][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.15140201151371002, 'train/grad_norm_pre_clip': 33.99908447265625, 'timing/data': 0.00036726705729961395, 'timing/model': 1.1170748069416732, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 1.06}) +[2026-06-24 01:11:20,652][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.019791243597865105, 'train/grad_norm_pre_clip': 36.219112396240234, 'timing/data': 0.00039595295675098896, 'timing/model': 1.1166506530717015, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 1.06}) +[2026-06-24 01:11:24,728][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.01551884040236473, 'train/grad_norm_pre_clip': 16.175817489624023, 'timing/data': 0.00040444498881697655, 'timing/model': 1.115513570024632, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 1.06}) +[2026-06-24 01:11:28,978][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.004020642954856157, 'train/grad_norm_pre_clip': 14.588250160217285, 'timing/data': 0.00035161792766302824, 'timing/model': 1.1213751489995047, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 1.06}) +[2026-06-24 01:11:33,053][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.014768436551094055, 'train/grad_norm_pre_clip': 23.425275802612305, 'timing/data': 0.00037094298750162125, 'timing/model': 1.1127202099887654, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 1.06}) +[2026-06-24 01:11:37,132][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.13740874826908112, 'train/grad_norm_pre_clip': 15.887713432312012, 'timing/data': 0.0005460949614644051, 'timing/model': 1.1202126720454544, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 1.06}) +[2026-06-24 01:11:41,227][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.03894910588860512, 'train/grad_norm_pre_clip': 10.093517303466797, 'timing/data': 0.00034701800905168056, 'timing/model': 1.1243207060033455, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 1.06}) +[2026-06-24 01:11:45,376][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.015752555802464485, 'train/grad_norm_pre_clip': 12.373282432556152, 'timing/data': 0.0004081679508090019, 'timing/model': 1.1123240899760276, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 1.06}) +[2026-06-24 01:11:49,448][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.028193242847919464, 'train/grad_norm_pre_clip': 28.08275604248047, 'timing/data': 0.00035533204209059477, 'timing/model': 1.115982659975998, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 1.06}) +[2026-06-24 01:11:53,521][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.011337259784340858, 'train/grad_norm_pre_clip': 21.378711700439453, 'timing/data': 0.00040881894528865814, 'timing/model': 1.113397364038974, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 1.06}) +[2026-06-24 01:11:57,600][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.08772251754999161, 'train/grad_norm_pre_clip': 31.61469268798828, 'timing/data': 0.0003576310118660331, 'timing/model': 1.1204100819304585, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 1.06}) +[2026-06-24 01:12:01,766][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.12228688597679138, 'train/grad_norm_pre_clip': 18.754106521606445, 'timing/data': 0.00038863695226609707, 'timing/model': 1.1187487279530615, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 1.06}) +[2026-06-24 01:12:05,834][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.03371593356132507, 'train/grad_norm_pre_clip': 18.57883071899414, 'timing/data': 0.0003550619585439563, 'timing/model': 1.1116970460861921, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 1.06}) +[2026-06-24 01:12:09,913][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.2297905534505844, 'train/grad_norm_pre_clip': 39.30171585083008, 'timing/data': 0.0005010430468246341, 'timing/model': 1.1171660230029374, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 1.06}) +[2026-06-24 01:12:13,996][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.05646785721182823, 'train/grad_norm_pre_clip': 22.069210052490234, 'timing/data': 0.0003550240071490407, 'timing/model': 1.1208973940229043, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 1.06}) +[2026-06-24 01:12:18,227][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.057367876172065735, 'train/grad_norm_pre_clip': 13.480098724365234, 'timing/data': 0.00034755305387079716, 'timing/model': 1.1165660640690476, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 1.06}) +[2026-06-24 01:12:22,307][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.03767608106136322, 'train/grad_norm_pre_clip': 7.682125091552734, 'timing/data': 0.0003576489398255944, 'timing/model': 1.12078905897215, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 1.06}) +[2026-06-24 01:12:26,404][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.01540494617074728, 'train/grad_norm_pre_clip': 19.059267044067383, 'timing/data': 0.0004433171125128865, 'timing/model': 1.129431613953784, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 1.06}) +[2026-06-24 01:12:30,582][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.04834330081939697, 'train/grad_norm_pre_clip': 18.5126953125, 'timing/data': 0.0003690259763970971, 'timing/model': 1.2153692089486867, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 1.06}) +[2026-06-24 01:12:34,656][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.0786202922463417, 'train/grad_norm_pre_clip': 47.00071334838867, 'timing/data': 0.0003706440329551697, 'timing/model': 1.118283531977795, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.06}) +[2026-06-24 01:12:38,744][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.07298854738473892, 'train/grad_norm_pre_clip': 15.06491756439209, 'timing/data': 0.0004201059928163886, 'timing/model': 1.1243350200820714, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 1.06}) +[2026-06-24 01:12:42,830][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.03609559312462807, 'train/grad_norm_pre_clip': 22.190248489379883, 'timing/data': 0.0004226560704410076, 'timing/model': 1.1187866610707715, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 1.06}) +[2026-06-24 01:12:47,072][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.05437146872282028, 'train/grad_norm_pre_clip': 13.237711906433105, 'timing/data': 0.0003605770180001855, 'timing/model': 1.2831067439401522, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 1.06}) +[2026-06-24 01:18:45,800][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2750 +[2026-06-24 01:18:45,800][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.010207831859588623, 'train/grad_norm_pre_clip': 7.416253089904785, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 218.3800000000104, 'rl_games_eval/mid_train/mean_length': 1778.2, 'rl_games_eval/mid_train/std_reward': 184.83810646077453, 'rl_games_eval/mid_train/std_length': 1488.5559982748382, 'rl_games_eval/mid_train/macro_mean_reward': 218.3800000000104, 'rl_games_eval/mid_train/macro_mean_length': 1778.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 218.3800000000104, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1778.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 184.83810646077453, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1488.5559982748382, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 218.3800000000104, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00034934794530272484, 'timing/model': 1.1126879879739136, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 1.06}) +[2026-06-24 01:18:49,781][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.13475118577480316, 'train/grad_norm_pre_clip': 16.672012329101562, 'timing/data': 0.0003555190050974488, 'timing/model': 1.0886894509894773, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 1.06}) +[2026-06-24 01:18:53,771][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.09239203482866287, 'train/grad_norm_pre_clip': 14.756850242614746, 'timing/data': 0.00035089091397821903, 'timing/model': 1.0994365960359573, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 1.07}) +[2026-06-24 01:18:57,927][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.003968706354498863, 'train/grad_norm_pre_clip': 22.52182960510254, 'timing/data': 0.0003511830000206828, 'timing/model': 1.1015701020369306, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 1.07}) +[2026-06-24 01:19:01,936][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.00287142232991755, 'train/grad_norm_pre_clip': 21.375585556030273, 'timing/data': 0.00040675397031009197, 'timing/model': 1.0971251600421965, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 1.07}) +[2026-06-24 01:19:05,946][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.005637888330966234, 'train/grad_norm_pre_clip': 20.6519832611084, 'timing/data': 0.00035464600659906864, 'timing/model': 1.1050360589288175, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 1.07}) +[2026-06-24 01:19:09,966][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.0019256924279034138, 'train/grad_norm_pre_clip': 7.160612106323242, 'timing/data': 0.00035237905103713274, 'timing/model': 1.0976786660030484, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 1.07}) +[2026-06-24 01:19:14,166][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.14234700798988342, 'train/grad_norm_pre_clip': 22.327573776245117, 'timing/data': 0.00035099696833640337, 'timing/model': 1.1027388840448111, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 1.07}) +[2026-06-24 01:19:18,200][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.04076038673520088, 'train/grad_norm_pre_clip': 26.65785026550293, 'timing/data': 0.0004855350125581026, 'timing/model': 1.10813932295423, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 1.07}) +[2026-06-24 01:19:22,243][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.02824624627828598, 'train/grad_norm_pre_clip': 21.643239974975586, 'timing/data': 0.0005737339379265904, 'timing/model': 1.1043305079219863, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 1.07}) +[2026-06-24 01:19:26,282][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.006105605512857437, 'train/grad_norm_pre_clip': 17.81879997253418, 'timing/data': 0.000349557027220726, 'timing/model': 1.1066506520146504, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 1.07}) +[2026-06-24 01:19:30,471][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.05228154733777046, 'train/grad_norm_pre_clip': 26.50360107421875, 'timing/data': 0.0003489729715511203, 'timing/model': 1.1067765770712867, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 1.07}) +[2026-06-24 01:19:34,515][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.024376526474952698, 'train/grad_norm_pre_clip': 19.50404167175293, 'timing/data': 0.0006732179317623377, 'timing/model': 1.1082137579796836, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 1.07}) +[2026-06-24 01:19:38,568][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.16569925844669342, 'train/grad_norm_pre_clip': 19.497711181640625, 'timing/data': 0.00034787505865097046, 'timing/model': 1.1118862670846283, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 1.07}) +[2026-06-24 01:19:42,624][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.10478239506483078, 'train/grad_norm_pre_clip': 10.392949104309082, 'timing/data': 0.0004075350007042289, 'timing/model': 1.1108863119734451, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 1.07}) +[2026-06-24 01:19:46,837][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.03933338820934296, 'train/grad_norm_pre_clip': 9.189871788024902, 'timing/data': 0.0004016419406980276, 'timing/model': 1.1138286949135363, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 1.07}) +[2026-06-24 01:19:50,892][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.07062143087387085, 'train/grad_norm_pre_clip': 24.192096710205078, 'timing/data': 0.0004269318887963891, 'timing/model': 1.1151018589735031, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 1.07}) +[2026-06-24 01:19:54,947][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.058768924325704575, 'train/grad_norm_pre_clip': 17.70035743713379, 'timing/data': 0.0004168190062046051, 'timing/model': 1.1075309759471565, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 1.07}) +[2026-06-24 01:19:58,993][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.08582982420921326, 'train/grad_norm_pre_clip': 19.82524871826172, 'timing/data': 0.0004449679981917143, 'timing/model': 1.111185853020288, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 1.07}) +[2026-06-24 01:20:03,208][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.11008108407258987, 'train/grad_norm_pre_clip': 24.190322875976562, 'timing/data': 0.0004220729460939765, 'timing/model': 1.1120042100083083, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 1.07}) +[2026-06-24 01:20:07,278][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.031342968344688416, 'train/grad_norm_pre_clip': 7.882561683654785, 'timing/data': 0.00035829900298267603, 'timing/model': 1.114981401944533, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 1.07}) +[2026-06-24 01:20:11,343][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.042969465255737305, 'train/grad_norm_pre_clip': 8.596179962158203, 'timing/data': 0.0006202720105648041, 'timing/model': 1.1198733659693971, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 1.07}) +[2026-06-24 01:20:15,406][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.004479622934013605, 'train/grad_norm_pre_clip': 12.225425720214844, 'timing/data': 0.000588688999414444, 'timing/model': 1.1086310959653929, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 1.07}) +[2026-06-24 01:20:19,642][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.021631842479109764, 'train/grad_norm_pre_clip': 9.931974411010742, 'timing/data': 0.00036440708208829165, 'timing/model': 1.1128851789981127, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 1.07}) +[2026-06-24 01:20:23,717][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.037033263593912125, 'train/grad_norm_pre_clip': 17.768129348754883, 'timing/data': 0.00035794288851320744, 'timing/model': 1.1154200880555436, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 1.07}) +[2026-06-24 01:20:27,771][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.01890525035560131, 'train/grad_norm_pre_clip': 15.01350212097168, 'timing/data': 0.0003582120407372713, 'timing/model': 1.1169948419556022, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 1.07}) +[2026-06-24 01:20:31,827][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.07323439419269562, 'train/grad_norm_pre_clip': 21.744735717773438, 'timing/data': 0.0005097989924252033, 'timing/model': 1.1094960040645674, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 1.07}) +[2026-06-24 01:20:36,052][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.007265483494848013, 'train/grad_norm_pre_clip': 11.562832832336426, 'timing/data': 0.0004918540362268686, 'timing/model': 1.1113116519991308, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 1.07}) +[2026-06-24 01:20:40,124][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.025816423818469048, 'train/grad_norm_pre_clip': 35.88203811645508, 'timing/data': 0.00036630802787840366, 'timing/model': 1.1217026780359447, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 1.08}) +[2026-06-24 01:20:44,196][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.04837750270962715, 'train/grad_norm_pre_clip': 19.535200119018555, 'timing/data': 0.0004110320005565882, 'timing/model': 1.1219284581020474, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 1.08}) +[2026-06-24 01:20:48,289][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.03675859421491623, 'train/grad_norm_pre_clip': 29.677661895751953, 'timing/data': 0.0004391989205032587, 'timing/model': 1.1212606300832704, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 1.08}) +[2026-06-24 01:20:52,528][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.023331988602876663, 'train/grad_norm_pre_clip': 14.783024787902832, 'timing/data': 0.0003957569133490324, 'timing/model': 1.1154473370406777, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 1.08}) +[2026-06-24 01:20:56,610][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.10973072797060013, 'train/grad_norm_pre_clip': 32.2097053527832, 'timing/data': 0.0003913340624421835, 'timing/model': 1.1154676609439775, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 1.08}) +[2026-06-24 01:21:00,679][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.0919232964515686, 'train/grad_norm_pre_clip': 39.62494659423828, 'timing/data': 0.0005259440513327718, 'timing/model': 1.1144132319604978, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 1.08}) +[2026-06-24 01:21:04,751][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.0569460429251194, 'train/grad_norm_pre_clip': 21.898090362548828, 'timing/data': 0.0005654729902744293, 'timing/model': 1.1182828800519928, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 1.08}) +[2026-06-24 01:21:08,981][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.0820116475224495, 'train/grad_norm_pre_clip': 21.406780242919922, 'timing/data': 0.0006563049973919988, 'timing/model': 1.1076982939848676, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 1.08}) +[2026-06-24 01:21:13,059][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.07182270288467407, 'train/grad_norm_pre_clip': 24.48296356201172, 'timing/data': 0.000359566998668015, 'timing/model': 1.1214627549052238, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 1.08}) +[2026-06-24 01:21:17,117][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.19418971240520477, 'train/grad_norm_pre_clip': 31.723661422729492, 'timing/data': 0.0003608010010793805, 'timing/model': 1.1111477159429342, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 1.08}) +[2026-06-24 01:21:21,192][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.08808495104312897, 'train/grad_norm_pre_clip': 34.73140335083008, 'timing/data': 0.00043585000094026327, 'timing/model': 1.1188045250019059, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 1.08}) +[2026-06-24 01:21:25,429][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.0867842435836792, 'train/grad_norm_pre_clip': 18.802099227905273, 'timing/data': 0.0005442359251901507, 'timing/model': 1.115030478918925, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 1.08}) +[2026-06-24 01:21:29,510][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.05016206577420235, 'train/grad_norm_pre_clip': 25.270959854125977, 'timing/data': 0.0003970060497522354, 'timing/model': 1.1210530659882352, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 1.08}) +[2026-06-24 01:21:33,592][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.035674311220645905, 'train/grad_norm_pre_clip': 23.444082260131836, 'timing/data': 0.00041855790186673403, 'timing/model': 1.1181181960273534, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 1.08}) +[2026-06-24 01:21:37,680][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.11225336790084839, 'train/grad_norm_pre_clip': 23.806665420532227, 'timing/data': 0.00035296299029141665, 'timing/model': 1.115465166978538, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 1.08}) +[2026-06-24 01:21:41,921][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.0017240983434021473, 'train/grad_norm_pre_clip': 17.423009872436523, 'timing/data': 0.00036011403426527977, 'timing/model': 1.1146924910135567, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 1.08}) +[2026-06-24 01:21:45,988][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.0020602222066372633, 'train/grad_norm_pre_clip': 16.23787498474121, 'timing/data': 0.00035572005435824394, 'timing/model': 1.1144242119044065, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 1.08}) +[2026-06-24 01:21:50,061][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.015165840275585651, 'train/grad_norm_pre_clip': 22.379154205322266, 'timing/data': 0.00038861704524606466, 'timing/model': 1.1138206589967012, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.08}) +[2026-06-24 01:21:54,131][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.0039804065600037575, 'train/grad_norm_pre_clip': 9.833308219909668, 'timing/data': 0.0005274099530652165, 'timing/model': 1.1154684569919482, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 1.08}) +[2026-06-24 01:21:58,365][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.118329718708992, 'train/grad_norm_pre_clip': 21.452590942382812, 'timing/data': 0.00036161195021122694, 'timing/model': 1.1144374189898372, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 1.08}) +[2026-06-24 01:22:02,460][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.15963862836360931, 'train/grad_norm_pre_clip': 25.83636474609375, 'timing/data': 0.0003585759550333023, 'timing/model': 1.124610936967656, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 1.08}) +[2026-06-24 01:22:06,546][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.049883123487234116, 'train/grad_norm_pre_clip': 9.238574028015137, 'timing/data': 0.00035720004234462976, 'timing/model': 1.1214410609100014, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 1.08}) +[2026-06-24 01:22:10,612][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2800 +[2026-06-24 01:23:14,820][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2800 +[2026-06-24 01:23:14,822][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2800 +[2026-06-24 01:23:54,004][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2800 +[2026-06-24 01:23:54,004][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.009847461245954037, 'train/grad_norm_pre_clip': 22.09250259399414, 'eval/action_loss/seconds': 64.20635826897342, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.030670640990138054, 'eval/latency_2/loss': 0.030670640990138054, 'eval/flappy/loss': 0.030670640990138054, 'eval/flappy/latency_2/loss': 0.030670640990138054, 'eval/flap_precision': 0.9602122015915119, 'eval/flap_recall': 0.9823609226594301, 'eval/flap_f1': 0.9711602951039571, 'eval/flap_rate_model': 0.20944444444444443, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9711602951039571, 'eval/latency_2/flap_precision': 0.9602122015915119, 'eval/latency_2/flap_recall': 0.9823609226594301, 'eval/latency_2/flap_f1': 0.9711602951039571, 'eval/latency_2/flap_rate_model': 0.20944444444444443, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9711602951039571, 'eval/action_classification/data_wait_seconds': 0.6141614317893982, 'eval/action_classification/predict_seconds': 38.41828918457031, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0003575680311769247, 'timing/model': 1.1164987430674955, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 1.08}) +[2026-06-24 01:23:58,048][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.008099308237433434, 'train/grad_norm_pre_clip': 7.174896240234375, 'timing/data': 0.0003593110013753176, 'timing/model': 1.1112102250335738, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 1.08}) +[2026-06-24 01:24:02,093][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.020945483818650246, 'train/grad_norm_pre_clip': 12.66117000579834, 'timing/data': 0.00042317097540944815, 'timing/model': 1.1107597249792889, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 1.08}) +[2026-06-24 01:24:06,156][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.07851886749267578, 'train/grad_norm_pre_clip': 16.21261978149414, 'timing/data': 0.0003578129690140486, 'timing/model': 1.1141560400137678, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 1.08}) +[2026-06-24 01:24:10,334][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.12287040054798126, 'train/grad_norm_pre_clip': 25.07232666015625, 'timing/data': 0.00048480601981282234, 'timing/model': 1.1115804839646444, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 1.09}) +[2026-06-24 01:24:14,386][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.019446080550551414, 'train/grad_norm_pre_clip': 12.343490600585938, 'timing/data': 0.00035934196785092354, 'timing/model': 1.1086387450341135, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 1.09}) +[2026-06-24 01:24:18,448][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.09169147163629532, 'train/grad_norm_pre_clip': 10.228852272033691, 'timing/data': 0.0004022399662062526, 'timing/model': 1.1116562180686742, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 1.09}) +[2026-06-24 01:24:22,510][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.0018417423125356436, 'train/grad_norm_pre_clip': 14.678850173950195, 'timing/data': 0.0003537100274115801, 'timing/model': 1.1118038210552186, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 1.09}) +[2026-06-24 01:24:26,679][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.20129242539405823, 'train/grad_norm_pre_clip': 24.021562576293945, 'timing/data': 0.00046157906763255596, 'timing/model': 1.1093037080718204, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 1.09}) +[2026-06-24 01:24:30,729][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.17117390036582947, 'train/grad_norm_pre_clip': 19.23513412475586, 'timing/data': 0.0004500810755416751, 'timing/model': 1.1120429609436542, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 1.09}) +[2026-06-24 01:24:34,790][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.03809173405170441, 'train/grad_norm_pre_clip': 15.128585815429688, 'timing/data': 0.0003441410372033715, 'timing/model': 1.1139929379569367, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 1.09}) +[2026-06-24 01:24:38,840][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.03240885213017464, 'train/grad_norm_pre_clip': 10.144137382507324, 'timing/data': 0.00036033603828400373, 'timing/model': 1.1117586430627853, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 1.09}) +[2026-06-24 01:24:43,031][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.020814264193177223, 'train/grad_norm_pre_clip': 11.124658584594727, 'timing/data': 0.0003671670565381646, 'timing/model': 1.109314135974273, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 1.09}) +[2026-06-24 01:24:47,082][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.08584019541740417, 'train/grad_norm_pre_clip': 18.511211395263672, 'timing/data': 0.0003884619800373912, 'timing/model': 1.1108145929174498, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 1.09}) +[2026-06-24 01:24:51,148][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.04633794724941254, 'train/grad_norm_pre_clip': 16.283334732055664, 'timing/data': 0.0003893970279023051, 'timing/model': 1.1142329910071567, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 1.09}) +[2026-06-24 01:24:55,227][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.040893226861953735, 'train/grad_norm_pre_clip': 26.22186279296875, 'timing/data': 0.00042591895908117294, 'timing/model': 1.113677266985178, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 1.09}) +[2026-06-24 01:24:59,415][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.0034785254392772913, 'train/grad_norm_pre_clip': 10.913162231445312, 'timing/data': 0.00040040002204477787, 'timing/model': 1.1163639470469207, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 1.09}) +[2026-06-24 01:25:03,494][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.031509075313806534, 'train/grad_norm_pre_clip': 14.79261302947998, 'timing/data': 0.0006042489549145103, 'timing/model': 1.118590253056027, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 1.09}) +[2026-06-24 01:25:07,579][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.01161065325140953, 'train/grad_norm_pre_clip': 23.845256805419922, 'timing/data': 0.0005737129831686616, 'timing/model': 1.114884780952707, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 1.09}) +[2026-06-24 01:25:11,664][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.007182038389146328, 'train/grad_norm_pre_clip': 11.528079986572266, 'timing/data': 0.0003632840234786272, 'timing/model': 1.1143962750211358, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 1.09}) +[2026-06-24 01:25:15,854][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.01536274328827858, 'train/grad_norm_pre_clip': 14.654052734375, 'timing/data': 0.0003526590298861265, 'timing/model': 1.1163141980068758, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 1.09}) +[2026-06-24 01:25:19,903][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.03533901274204254, 'train/grad_norm_pre_clip': 18.646942138671875, 'timing/data': 0.00043742498382925987, 'timing/model': 1.113475737045519, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 1.09}) +[2026-06-24 01:25:23,970][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.06569650024175644, 'train/grad_norm_pre_clip': 15.378332138061523, 'timing/data': 0.00035217893309891224, 'timing/model': 1.1171918959589675, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 1.09}) +[2026-06-24 01:25:28,027][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.2207147181034088, 'train/grad_norm_pre_clip': 25.205093383789062, 'timing/data': 0.00036162196192890406, 'timing/model': 1.1146228859433904, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 1.09}) +[2026-06-24 01:25:32,198][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.1362333744764328, 'train/grad_norm_pre_clip': 15.18553352355957, 'timing/data': 0.00035924999974668026, 'timing/model': 1.1076634830096737, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 1.09}) +[2026-06-24 01:25:36,253][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.05413997918367386, 'train/grad_norm_pre_clip': 23.624229431152344, 'timing/data': 0.00035430106800049543, 'timing/model': 1.1148562710732222, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 1.09}) +[2026-06-24 01:25:40,334][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.0070869880728423595, 'train/grad_norm_pre_clip': 29.16754722595215, 'timing/data': 0.000402687001042068, 'timing/model': 1.121962334960699, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 1.09}) +[2026-06-24 01:25:44,411][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.00129506119992584, 'train/grad_norm_pre_clip': 12.207611083984375, 'timing/data': 0.0003546531079337001, 'timing/model': 1.1188170439563692, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 1.09}) +[2026-06-24 01:25:48,620][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.003900013631209731, 'train/grad_norm_pre_clip': 21.512754440307617, 'timing/data': 0.00035774405114352703, 'timing/model': 1.1242717610439286, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 1.09}) +[2026-06-24 01:25:52,695][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.03181268647313118, 'train/grad_norm_pre_clip': 14.283941268920898, 'timing/data': 0.000620140926912427, 'timing/model': 1.122221803991124, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 1.09}) +[2026-06-24 01:25:56,761][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.10844637453556061, 'train/grad_norm_pre_clip': 19.515090942382812, 'timing/data': 0.0005834420444443822, 'timing/model': 1.1141469390131533, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 1.1}) +[2026-06-24 01:26:00,841][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.013475777581334114, 'train/grad_norm_pre_clip': 20.27321434020996, 'timing/data': 0.0003613709704950452, 'timing/model': 1.1184030889999121, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 1.1}) +[2026-06-24 01:26:05,025][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.0863422304391861, 'train/grad_norm_pre_clip': 36.29338455200195, 'timing/data': 0.00039047596510499716, 'timing/model': 1.1119394110282883, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 1.1}) +[2026-06-24 01:26:09,091][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.11788289994001389, 'train/grad_norm_pre_clip': 19.4512996673584, 'timing/data': 0.0003604890080168843, 'timing/model': 1.1136426329612732, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 1.1}) +[2026-06-24 01:26:13,173][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.08403119444847107, 'train/grad_norm_pre_clip': 16.313325881958008, 'timing/data': 0.000408743042498827, 'timing/model': 1.1179920010035858, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 1.1}) +[2026-06-24 01:26:17,240][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.10633068531751633, 'train/grad_norm_pre_clip': 18.549976348876953, 'timing/data': 0.00041167205199599266, 'timing/model': 1.1148382850224152, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 1.1}) +[2026-06-24 01:26:21,442][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.004696324001997709, 'train/grad_norm_pre_clip': 17.911516189575195, 'timing/data': 0.00035399303305894136, 'timing/model': 1.113639177987352, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 1.1}) +[2026-06-24 01:26:25,505][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.00906873308122158, 'train/grad_norm_pre_clip': 9.742329597473145, 'timing/data': 0.0005696490406990051, 'timing/model': 1.1094246529974043, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 1.1}) +[2026-06-24 01:26:29,584][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.038912709802389145, 'train/grad_norm_pre_clip': 26.309362411499023, 'timing/data': 0.0003586430102586746, 'timing/model': 1.1175967589952052, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 1.1}) +[2026-06-24 01:26:33,669][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.0857369676232338, 'train/grad_norm_pre_clip': 24.65328598022461, 'timing/data': 0.000565935974009335, 'timing/model': 1.1180039279861376, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 1.1}) +[2026-06-24 01:26:37,869][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.14025112986564636, 'train/grad_norm_pre_clip': 25.79119300842285, 'timing/data': 0.0004064540844410658, 'timing/model': 1.1170028549386188, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 1.1}) +[2026-06-24 01:26:41,950][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.0795803964138031, 'train/grad_norm_pre_clip': 17.085107803344727, 'timing/data': 0.0004892430733889341, 'timing/model': 1.1180468329694122, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 1.1}) +[2026-06-24 01:26:46,021][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.04062886908650398, 'train/grad_norm_pre_clip': 17.78469467163086, 'timing/data': 0.0004031800199300051, 'timing/model': 1.11218367703259, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 1.1}) +[2026-06-24 01:26:50,109][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.06720007956027985, 'train/grad_norm_pre_clip': 17.222755432128906, 'timing/data': 0.0004293980309739709, 'timing/model': 1.119921088917181, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 1.1}) +[2026-06-24 01:26:54,303][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.14999328553676605, 'train/grad_norm_pre_clip': 38.36079406738281, 'timing/data': 0.00041873904410749674, 'timing/model': 1.118754607043229, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.1}) +[2026-06-24 01:26:58,385][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.005907730665057898, 'train/grad_norm_pre_clip': 14.999201774597168, 'timing/data': 0.0004791009705513716, 'timing/model': 1.115324626909569, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 1.1}) +[2026-06-24 01:27:02,462][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.05918053165078163, 'train/grad_norm_pre_clip': 10.571462631225586, 'timing/data': 0.00036644504871219397, 'timing/model': 1.1162593170301989, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 1.1}) +[2026-06-24 01:27:06,535][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.048060446977615356, 'train/grad_norm_pre_clip': 9.572860717773438, 'timing/data': 0.00039572594687342644, 'timing/model': 1.1145835929783061, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 1.1}) +[2026-06-24 01:27:10,734][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.005623610224574804, 'train/grad_norm_pre_clip': 14.768616676330566, 'timing/data': 0.00036632001865655184, 'timing/model': 1.1117358210030943, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 1.1}) +[2026-06-24 01:27:14,797][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.004065928049385548, 'train/grad_norm_pre_clip': 11.645356178283691, 'timing/data': 0.0004506839904934168, 'timing/model': 1.1142186539946124, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 1.1}) +[2026-06-24 01:27:18,880][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.01959412917494774, 'train/grad_norm_pre_clip': 7.674234867095947, 'timing/data': 0.0003541110781952739, 'timing/model': 1.1211989360162988, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 1.1}) +[2026-06-24 01:27:22,969][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.00889076292514801, 'train/grad_norm_pre_clip': 16.115821838378906, 'timing/data': 0.0004045169334858656, 'timing/model': 1.1180528070544824, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 1.1}) +[2026-06-24 01:27:27,163][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.016915474086999893, 'train/grad_norm_pre_clip': 20.436918258666992, 'timing/data': 0.000365586020052433, 'timing/model': 1.117692533065565, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 1.1}) +[2026-06-24 01:27:31,245][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.053042054176330566, 'train/grad_norm_pre_clip': 11.964727401733398, 'timing/data': 0.0005321359494701028, 'timing/model': 1.123923446983099, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 1.1}) +[2026-06-24 01:27:35,322][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.11885680258274078, 'train/grad_norm_pre_clip': 21.014801025390625, 'timing/data': 0.00043640495277941227, 'timing/model': 1.1188740320503712, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 1.1}) +[2026-06-24 01:27:39,409][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.04407639801502228, 'train/grad_norm_pre_clip': 11.742431640625, 'timing/data': 0.00036012998316437006, 'timing/model': 1.1201005949405953, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 1.1}) +[2026-06-24 01:27:43,602][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.012330656871199608, 'train/grad_norm_pre_clip': 13.160200119018555, 'timing/data': 0.00036190799437463284, 'timing/model': 1.1237359980586916, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 1.11}) +[2026-06-24 01:27:47,694][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.030350496992468834, 'train/grad_norm_pre_clip': 19.436235427856445, 'timing/data': 0.0005769369890913367, 'timing/model': 1.1147579170065, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 1.11}) +[2026-06-24 01:27:51,772][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.10278528928756714, 'train/grad_norm_pre_clip': 23.533714294433594, 'timing/data': 0.00035914801992475986, 'timing/model': 1.117175460094586, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 1.11}) +[2026-06-24 01:27:55,872][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.04970478639006615, 'train/grad_norm_pre_clip': 19.421398162841797, 'timing/data': 0.00037502602208405733, 'timing/model': 1.121009671012871, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 1.11}) +[2026-06-24 01:28:00,075][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.07965947687625885, 'train/grad_norm_pre_clip': 9.905631065368652, 'timing/data': 0.00035117100924253464, 'timing/model': 1.1236839649500325, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 1.11}) +[2026-06-24 01:28:04,146][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.05959053337574005, 'train/grad_norm_pre_clip': 14.213318824768066, 'timing/data': 0.0004897889448329806, 'timing/model': 1.1120743269566447, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 1.11}) +[2026-06-24 01:28:08,228][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.05183711647987366, 'train/grad_norm_pre_clip': 26.928762435913086, 'timing/data': 0.0003566939849406481, 'timing/model': 1.1254919919883832, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 1.11}) +[2026-06-24 01:28:12,315][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.0022332852240651846, 'train/grad_norm_pre_clip': 21.258419036865234, 'timing/data': 0.0003678339999169111, 'timing/model': 1.1185032119974494, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 1.11}) +[2026-06-24 01:28:16,544][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.03492244705557823, 'train/grad_norm_pre_clip': 30.387813568115234, 'timing/data': 0.0003547440283000469, 'timing/model': 1.1229266669834033, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 1.11}) +[2026-06-24 01:28:20,635][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.11873901635408401, 'train/grad_norm_pre_clip': 32.007076263427734, 'timing/data': 0.000354211893863976, 'timing/model': 1.119647434912622, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 1.11}) +[2026-06-24 01:28:24,734][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.017408080399036407, 'train/grad_norm_pre_clip': 21.431041717529297, 'timing/data': 0.0005913510685786605, 'timing/model': 1.1357071989914402, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 1.11}) +[2026-06-24 01:28:28,826][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.08978558331727982, 'train/grad_norm_pre_clip': 15.240032196044922, 'timing/data': 0.0005601149750873446, 'timing/model': 1.1215497680241242, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 1.11}) +[2026-06-24 01:28:33,032][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.04073944315314293, 'train/grad_norm_pre_clip': 19.350461959838867, 'timing/data': 0.000396719085983932, 'timing/model': 1.1236877560149878, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 1.11}) +[2026-06-24 01:28:37,102][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.06144064664840698, 'train/grad_norm_pre_clip': 7.077216625213623, 'timing/data': 0.0003547270316630602, 'timing/model': 1.1127023010049015, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 1.11}) +[2026-06-24 01:28:41,174][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.027020685374736786, 'train/grad_norm_pre_clip': 29.46666145324707, 'timing/data': 0.00035761401522904634, 'timing/model': 1.113214318989776, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 1.11}) +[2026-06-24 01:28:45,237][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.06873967498540878, 'train/grad_norm_pre_clip': 22.223859786987305, 'timing/data': 0.00043918704614043236, 'timing/model': 1.1135011849692091, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 1.11}) +[2026-06-24 01:28:49,430][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.10675019770860672, 'train/grad_norm_pre_clip': 20.225452423095703, 'timing/data': 0.0003951729740947485, 'timing/model': 1.1216394390212372, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 1.11}) +[2026-06-24 01:28:53,541][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.178518608212471, 'train/grad_norm_pre_clip': 20.25040054321289, 'timing/data': 0.0003548719687387347, 'timing/model': 1.120039050001651, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 1.11}) +[2026-06-24 01:28:57,645][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.01973022148013115, 'train/grad_norm_pre_clip': 27.962806701660156, 'timing/data': 0.0003503289772197604, 'timing/model': 1.1173167730448768, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 1.11}) +[2026-06-24 01:29:01,731][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.002139912685379386, 'train/grad_norm_pre_clip': 27.84046173095703, 'timing/data': 0.0006275629857555032, 'timing/model': 1.121157219982706, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 1.11}) +[2026-06-24 01:29:05,964][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.15049326419830322, 'train/grad_norm_pre_clip': 26.240299224853516, 'timing/data': 0.0003587139071896672, 'timing/model': 1.1248562019318342, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 1.11}) +[2026-06-24 01:29:10,042][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.16219520568847656, 'train/grad_norm_pre_clip': 40.05175018310547, 'timing/data': 0.00034952606074512005, 'timing/model': 1.1117098230170086, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 1.11}) +[2026-06-24 01:29:14,137][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.03182589262723923, 'train/grad_norm_pre_clip': 18.767492294311523, 'timing/data': 0.0004165129503235221, 'timing/model': 1.119448869023472, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 1.11}) +[2026-06-24 01:29:18,212][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.05410725623369217, 'train/grad_norm_pre_clip': 37.41807556152344, 'timing/data': 0.0003803099971264601, 'timing/model': 1.116958691040054, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 1.11}) +[2026-06-24 01:29:22,414][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.049105413258075714, 'train/grad_norm_pre_clip': 25.79490089416504, 'timing/data': 0.0004017180763185024, 'timing/model': 1.12051275302656, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 1.11}) +[2026-06-24 01:29:26,497][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.021457064896821976, 'train/grad_norm_pre_clip': 19.255321502685547, 'timing/data': 0.00035733694676309824, 'timing/model': 1.1167840249836445, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 1.11}) +[2026-06-24 01:29:30,576][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.069950170814991, 'train/grad_norm_pre_clip': 13.213480949401855, 'timing/data': 0.0005540410056710243, 'timing/model': 1.1184935439378023, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 1.12}) +[2026-06-24 01:29:34,661][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.015560843981802464, 'train/grad_norm_pre_clip': 17.40383529663086, 'timing/data': 0.00035583297722041607, 'timing/model': 1.1188510140636936, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 1.12}) +[2026-06-24 01:29:38,858][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.057851649820804596, 'train/grad_norm_pre_clip': 10.507379531860352, 'timing/data': 0.000660802936181426, 'timing/model': 1.1161079190205783, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 1.12}) +[2026-06-24 01:29:42,950][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.16256973147392273, 'train/grad_norm_pre_clip': 46.4424934387207, 'timing/data': 0.00035809597466140985, 'timing/model': 1.1172475689090788, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 1.12}) +[2026-06-24 01:29:47,040][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.018230516463518143, 'train/grad_norm_pre_clip': 11.876922607421875, 'timing/data': 0.0006355600198730826, 'timing/model': 1.1248677920084447, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 1.12}) +[2026-06-24 01:29:51,128][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.017047535628080368, 'train/grad_norm_pre_clip': 10.356621742248535, 'timing/data': 0.0005059809191152453, 'timing/model': 1.1167040270520374, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 1.12}) +[2026-06-24 01:29:55,323][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.06568602472543716, 'train/grad_norm_pre_clip': 14.458162307739258, 'timing/data': 0.0003786809975281358, 'timing/model': 1.1227724469499663, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 1.12}) +[2026-06-24 01:29:59,397][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.035601936280727386, 'train/grad_norm_pre_clip': 16.03571319580078, 'timing/data': 0.0004156780196353793, 'timing/model': 1.117786205955781, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 1.12}) +[2026-06-24 01:30:03,479][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.04245860129594803, 'train/grad_norm_pre_clip': 17.066516876220703, 'timing/data': 0.0004907549591735005, 'timing/model': 1.1173781789839268, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 1.12}) +[2026-06-24 01:30:07,551][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.04891134053468704, 'train/grad_norm_pre_clip': 12.732861518859863, 'timing/data': 0.0006450759246945381, 'timing/model': 1.1149137699976563, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 1.12}) +[2026-06-24 01:30:11,778][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.010048935189843178, 'train/grad_norm_pre_clip': 6.046656131744385, 'timing/data': 0.00035478209611028433, 'timing/model': 1.1173623049398884, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 1.12}) +[2026-06-24 01:30:15,866][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.09703068435192108, 'train/grad_norm_pre_clip': 15.645744323730469, 'timing/data': 0.0003557188902050257, 'timing/model': 1.121359052020125, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.12}) +[2026-06-24 01:30:19,950][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.08412336558103561, 'train/grad_norm_pre_clip': 16.417896270751953, 'timing/data': 0.00035423890221863985, 'timing/model': 1.1184200290590525, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 1.12}) +[2026-06-24 01:30:24,031][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.02104843594133854, 'train/grad_norm_pre_clip': 23.780467987060547, 'timing/data': 0.0003628591075539589, 'timing/model': 1.1182411090703681, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 1.12}) +[2026-06-24 01:30:28,236][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.10232643038034439, 'train/grad_norm_pre_clip': 36.449440002441406, 'timing/data': 0.00036351103335618973, 'timing/model': 1.124014142085798, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 1.12}) +[2026-06-24 01:30:32,328][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.3518396019935608, 'train/grad_norm_pre_clip': 20.890321731567383, 'timing/data': 0.0003487060312181711, 'timing/model': 1.1187532889889553, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 1.12}) +[2026-06-24 01:30:36,420][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.07873111963272095, 'train/grad_norm_pre_clip': 24.802047729492188, 'timing/data': 0.0003556960728019476, 'timing/model': 1.1193706210469827, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 1.12}) +[2026-06-24 01:30:40,513][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.022502435371279716, 'train/grad_norm_pre_clip': 11.203152656555176, 'timing/data': 0.000717189977876842, 'timing/model': 1.1155647409614176, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 1.12}) +[2026-06-24 01:30:44,693][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2900 +[2026-06-24 01:31:48,571][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2900 +[2026-06-24 01:31:48,573][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2900 +[2026-06-24 01:32:27,832][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2900 +[2026-06-24 01:32:27,832][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.03045785240828991, 'train/grad_norm_pre_clip': 31.253313064575195, 'eval/action_loss/seconds': 63.87539778696373, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.032146576792001724, 'eval/latency_2/loss': 0.032146576792001724, 'eval/flappy/loss': 0.032146576792001724, 'eval/flappy/latency_2/loss': 0.032146576792001724, 'eval/flap_precision': 0.9740791268758526, 'eval/flap_recall': 0.9687924016282226, 'eval/flap_f1': 0.9714285714285714, 'eval/flap_rate_model': 0.2036111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9714285714285713, 'eval/latency_2/flap_precision': 0.9740791268758526, 'eval/latency_2/flap_recall': 0.9687924016282226, 'eval/latency_2/flap_f1': 0.9714285714285714, 'eval/latency_2/flap_rate_model': 0.2036111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9714285714285713, 'eval/action_classification/data_wait_seconds': 0.5181130170822144, 'eval/action_classification/predict_seconds': 38.611331939697266, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0003845549654215574, 'timing/model': 1.1106212539598346, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 1.12}) +[2026-06-24 01:32:31,890][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.2471236139535904, 'train/grad_norm_pre_clip': 29.17397689819336, 'timing/data': 0.000362787046469748, 'timing/model': 1.103900264017284, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 1.12}) +[2026-06-24 01:32:35,934][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.04145398363471031, 'train/grad_norm_pre_clip': 15.36570930480957, 'timing/data': 0.0004009749973192811, 'timing/model': 1.107548633008264, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 1.12}) +[2026-06-24 01:32:40,045][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.017662005499005318, 'train/grad_norm_pre_clip': 12.909170150756836, 'timing/data': 0.00039884098805487156, 'timing/model': 1.1090652720304206, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 1.12}) +[2026-06-24 01:32:44,086][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.009032710455358028, 'train/grad_norm_pre_clip': 23.700382232666016, 'timing/data': 0.0003583229845389724, 'timing/model': 1.112687885062769, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 1.12}) +[2026-06-24 01:32:48,142][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.007958885282278061, 'train/grad_norm_pre_clip': 15.573614120483398, 'timing/data': 0.0003701600944623351, 'timing/model': 1.1082082350039855, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 1.12}) +[2026-06-24 01:32:52,195][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.0621112696826458, 'train/grad_norm_pre_clip': 24.91822052001953, 'timing/data': 0.00041814905125647783, 'timing/model': 1.109023683005944, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 1.12}) +[2026-06-24 01:32:56,443][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.1613282859325409, 'train/grad_norm_pre_clip': 27.900781631469727, 'timing/data': 0.0003638459602370858, 'timing/model': 1.112797679961659, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 1.13}) +[2026-06-24 01:33:00,530][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.08582698553800583, 'train/grad_norm_pre_clip': 10.039373397827148, 'timing/data': 0.00037615804467350245, 'timing/model': 1.1149804309243336, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 1.13}) +[2026-06-24 01:33:04,602][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.030588336288928986, 'train/grad_norm_pre_clip': 9.630271911621094, 'timing/data': 0.0005323679652065039, 'timing/model': 1.114881594083272, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 1.13}) +[2026-06-24 01:33:08,663][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.04732329025864601, 'train/grad_norm_pre_clip': 9.740204811096191, 'timing/data': 0.0003674000035971403, 'timing/model': 1.1069222789956257, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 1.13}) +[2026-06-24 01:33:12,837][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.015163416042923927, 'train/grad_norm_pre_clip': 3.6543362140655518, 'timing/data': 0.00035295903217047453, 'timing/model': 1.1194794189650565, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 1.13}) +[2026-06-24 01:33:16,892][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.04029318690299988, 'train/grad_norm_pre_clip': 18.459047317504883, 'timing/data': 0.00036622199695557356, 'timing/model': 1.1071327360114083, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 1.13}) +[2026-06-24 01:33:20,947][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.09182848781347275, 'train/grad_norm_pre_clip': 21.923885345458984, 'timing/data': 0.0003492949763312936, 'timing/model': 1.107571569038555, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 1.13}) +[2026-06-24 01:33:25,022][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.12626880407333374, 'train/grad_norm_pre_clip': 16.22307014465332, 'timing/data': 0.0007036529714241624, 'timing/model': 1.1119714810047299, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 1.13}) +[2026-06-24 01:33:29,207][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.03160072863101959, 'train/grad_norm_pre_clip': 24.831241607666016, 'timing/data': 0.0005091039929538965, 'timing/model': 1.2318649219814688, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 1.13}) +[2026-06-24 01:33:33,292][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.041776008903980255, 'train/grad_norm_pre_clip': 16.522741317749023, 'timing/data': 0.0005633329274132848, 'timing/model': 1.1219706090632826, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 1.13}) +[2026-06-24 01:33:37,369][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.06472675502300262, 'train/grad_norm_pre_clip': 13.65841293334961, 'timing/data': 0.0004385310458019376, 'timing/model': 1.115191366057843, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 1.13}) +[2026-06-24 01:33:41,446][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.047214262187480927, 'train/grad_norm_pre_clip': 17.455427169799805, 'timing/data': 0.0003607110120356083, 'timing/model': 1.1206578819546849, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 1.13}) +[2026-06-24 01:33:45,527][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.03511536866426468, 'train/grad_norm_pre_clip': 19.92238426208496, 'timing/data': 0.0003519250312820077, 'timing/model': 1.1216028020717204, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 1.13}) +[2026-06-24 01:33:49,756][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.2937353551387787, 'train/grad_norm_pre_clip': 26.803316116333008, 'timing/data': 0.00035203108564019203, 'timing/model': 1.1152966350782663, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 1.13}) +[2026-06-24 01:33:53,831][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.05560697987675667, 'train/grad_norm_pre_clip': 34.62681579589844, 'timing/data': 0.0005666990764439106, 'timing/model': 1.1173399650724605, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 1.13}) +[2026-06-24 01:33:57,895][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.0080448929220438, 'train/grad_norm_pre_clip': 11.341195106506348, 'timing/data': 0.000543853035196662, 'timing/model': 1.1168261349666864, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 1.13}) +[2026-06-24 01:34:01,985][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.004247256089001894, 'train/grad_norm_pre_clip': 14.262492179870605, 'timing/data': 0.0006564410869032145, 'timing/model': 1.1119467259850353, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 1.13}) +[2026-06-24 01:34:06,156][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.02388792112469673, 'train/grad_norm_pre_clip': 13.296371459960938, 'timing/data': 0.00036355399060994387, 'timing/model': 1.1202520410297439, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 1.13}) +[2026-06-24 01:34:10,227][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.029380004853010178, 'train/grad_norm_pre_clip': 22.185596466064453, 'timing/data': 0.0005475250072777271, 'timing/model': 1.1159327430650592, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 1.13}) +[2026-06-24 01:34:14,302][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.024202672764658928, 'train/grad_norm_pre_clip': 13.359153747558594, 'timing/data': 0.00037004798650741577, 'timing/model': 1.114755769027397, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 1.13}) +[2026-06-24 01:34:18,382][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.14389947056770325, 'train/grad_norm_pre_clip': 25.623022079467773, 'timing/data': 0.00040811498183757067, 'timing/model': 1.1149773080833256, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 1.13}) +[2026-06-24 01:34:22,536][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.01759692095220089, 'train/grad_norm_pre_clip': 18.70334815979004, 'timing/data': 0.00037489994429051876, 'timing/model': 1.1183044170029461, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 1.13}) +[2026-06-24 01:34:26,607][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.09827854484319687, 'train/grad_norm_pre_clip': 22.698373794555664, 'timing/data': 0.00035820703487843275, 'timing/model': 1.116205916972831, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 1.13}) +[2026-06-24 01:34:30,691][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.004035107791423798, 'train/grad_norm_pre_clip': 20.810062408447266, 'timing/data': 0.0003636570181697607, 'timing/model': 1.11772472190205, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 1.13}) +[2026-06-24 01:34:34,777][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.006058403290808201, 'train/grad_norm_pre_clip': 41.12813949584961, 'timing/data': 0.00040184997487813234, 'timing/model': 1.1217801179736853, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 1.13}) +[2026-06-24 01:34:38,935][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.018944507464766502, 'train/grad_norm_pre_clip': 37.76845169067383, 'timing/data': 0.00035397999454289675, 'timing/model': 1.1250940499594435, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 1.13}) +[2026-06-24 01:34:43,013][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.021957574412226677, 'train/grad_norm_pre_clip': 21.11359977722168, 'timing/data': 0.00035951100289821625, 'timing/model': 1.1160517160315067, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 1.14}) +[2026-06-24 01:34:47,109][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.033519353717565536, 'train/grad_norm_pre_clip': 10.165365219116211, 'timing/data': 0.00035398302134126425, 'timing/model': 1.1178789519472048, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 1.14}) +[2026-06-24 01:34:51,183][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.011643904261291027, 'train/grad_norm_pre_clip': 37.46367645263672, 'timing/data': 0.00044080906081944704, 'timing/model': 1.1145998070714995, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 1.14}) +[2026-06-24 01:34:55,399][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.07193674892187119, 'train/grad_norm_pre_clip': 19.09476661682129, 'timing/data': 0.00035403610672801733, 'timing/model': 1.1177918900502846, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 1.14}) +[2026-06-24 01:34:59,469][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.2749430239200592, 'train/grad_norm_pre_clip': 19.779619216918945, 'timing/data': 0.000355300959199667, 'timing/model': 1.1107177330413833, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 1.14}) +[2026-06-24 01:35:03,537][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.18299958109855652, 'train/grad_norm_pre_clip': 45.0386848449707, 'timing/data': 0.0005706229712814093, 'timing/model': 1.1146800819551572, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 1.14}) +[2026-06-24 01:35:07,706][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.005660037975758314, 'train/grad_norm_pre_clip': 15.29606819152832, 'timing/data': 0.00044370500836521387, 'timing/model': 1.2078145110281184, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 1.14}) +[2026-06-24 01:35:11,789][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.009342529810965061, 'train/grad_norm_pre_clip': 16.949445724487305, 'timing/data': 0.0003497930010780692, 'timing/model': 1.116072959965095, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 1.14}) +[2026-06-24 01:35:15,870][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.1202407255768776, 'train/grad_norm_pre_clip': 19.251262664794922, 'timing/data': 0.0003493690164759755, 'timing/model': 1.1118506139609963, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 1.14}) +[2026-06-24 01:35:19,954][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.006916638929396868, 'train/grad_norm_pre_clip': 11.936132431030273, 'timing/data': 0.00035581307020038366, 'timing/model': 1.1157346019754186, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.14}) +[2026-06-24 01:35:24,221][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.08384912461042404, 'train/grad_norm_pre_clip': 30.764490127563477, 'timing/data': 0.000348128960467875, 'timing/model': 1.296620315988548, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 1.14}) +[2026-06-24 01:35:28,310][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.05228489637374878, 'train/grad_norm_pre_clip': 11.394042015075684, 'timing/data': 0.0003578390460461378, 'timing/model': 1.1240612079855055, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 1.14}) +[2026-06-24 01:35:32,401][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.09383852034807205, 'train/grad_norm_pre_clip': 22.283472061157227, 'timing/data': 0.0006834600353613496, 'timing/model': 1.1216296550119296, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 1.14}) +[2026-06-24 01:35:36,498][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.0012620158959180117, 'train/grad_norm_pre_clip': 9.600809097290039, 'timing/data': 0.00041199405677616596, 'timing/model': 1.1215770690469071, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 1.14}) +[2026-06-24 01:35:40,678][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.041585154831409454, 'train/grad_norm_pre_clip': 11.550233840942383, 'timing/data': 0.00036540103610605, 'timing/model': 1.1198257900541648, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 1.14}) +[2026-06-24 01:35:44,773][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.027856940403580666, 'train/grad_norm_pre_clip': 8.543937683105469, 'timing/data': 0.00041147705633193254, 'timing/model': 1.1255641420139, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 1.14}) +[2026-06-24 01:35:48,850][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.006269888021051884, 'train/grad_norm_pre_clip': 9.807000160217285, 'timing/data': 0.00035950704477727413, 'timing/model': 1.1224121560808271, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 1.14}) +[2026-06-24 01:35:52,942][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.10640083253383636, 'train/grad_norm_pre_clip': 14.548396110534668, 'timing/data': 0.0006236369954422116, 'timing/model': 1.1154923069989309, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 1.14}) +[2026-06-24 01:35:57,103][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.01531418226659298, 'train/grad_norm_pre_clip': 20.069377899169922, 'timing/data': 0.00041485298424959183, 'timing/model': 1.1203250519465655, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 1.14}) +[2026-06-24 01:36:01,200][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.049803879112005234, 'train/grad_norm_pre_clip': 15.50947093963623, 'timing/data': 0.00044593901839107275, 'timing/model': 1.1275096050230786, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 1.14}) +[2026-06-24 01:36:05,280][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.07727042585611343, 'train/grad_norm_pre_clip': 28.91500473022461, 'timing/data': 0.0005491569172590971, 'timing/model': 1.1175132979406044, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 1.14}) +[2026-06-24 01:36:09,355][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.09885081648826599, 'train/grad_norm_pre_clip': 35.37409973144531, 'timing/data': 0.00035502994433045387, 'timing/model': 1.118008096003905, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 1.14}) +[2026-06-24 01:36:13,592][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.027889903634786606, 'train/grad_norm_pre_clip': 20.627412796020508, 'timing/data': 0.0003576090093702078, 'timing/model': 1.1263454270083457, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 1.14}) +[2026-06-24 01:36:17,685][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.0022942577488720417, 'train/grad_norm_pre_clip': 12.779576301574707, 'timing/data': 0.00043485802598297596, 'timing/model': 1.1186222470132634, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 1.14}) +[2026-06-24 01:36:21,759][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.03484710678458214, 'train/grad_norm_pre_clip': 9.616130828857422, 'timing/data': 0.00035167590249329805, 'timing/model': 1.1132743780035526, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 1.14}) +[2026-06-24 01:36:25,845][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.012598482891917229, 'train/grad_norm_pre_clip': 13.869065284729004, 'timing/data': 0.00055718794465065, 'timing/model': 1.1194376960629597, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 1.14}) +[2026-06-24 01:36:30,035][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.031207576394081116, 'train/grad_norm_pre_clip': 23.091411590576172, 'timing/data': 0.000435751979239285, 'timing/model': 1.1137352120131254, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 1.15}) +[2026-06-24 01:36:34,120][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.013120868243277073, 'train/grad_norm_pre_clip': 12.475457191467285, 'timing/data': 0.0004088809946551919, 'timing/model': 1.1228116810088977, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 1.15}) +[2026-06-24 01:36:38,201][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.07190608978271484, 'train/grad_norm_pre_clip': 18.645689010620117, 'timing/data': 0.0004923529922962189, 'timing/model': 1.1150954730110243, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 1.15}) +[2026-06-24 01:36:42,286][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.025451337918639183, 'train/grad_norm_pre_clip': 20.2729549407959, 'timing/data': 0.00047988398000597954, 'timing/model': 1.1102691889973357, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 1.15}) +[2026-06-24 01:36:46,455][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.08806466311216354, 'train/grad_norm_pre_clip': 38.498497009277344, 'timing/data': 0.0003576889866963029, 'timing/model': 1.1229394499678165, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 1.15}) +[2026-06-24 01:36:50,544][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.03254515677690506, 'train/grad_norm_pre_clip': 9.81824016571045, 'timing/data': 0.0004417860182002187, 'timing/model': 1.1174502320354804, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 1.15}) +[2026-06-24 01:36:54,637][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.08952730894088745, 'train/grad_norm_pre_clip': 24.674610137939453, 'timing/data': 0.0004065609537065029, 'timing/model': 1.1237695629242808, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 1.15}) +[2026-06-24 01:36:58,717][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.13817450404167175, 'train/grad_norm_pre_clip': 27.815078735351562, 'timing/data': 0.00038984697312116623, 'timing/model': 1.1163322100182995, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 1.15}) +[2026-06-24 01:37:02,967][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.026653531938791275, 'train/grad_norm_pre_clip': 14.305753707885742, 'timing/data': 0.0003600220661610365, 'timing/model': 1.1223298630211502, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 1.15}) +[2026-06-24 01:37:07,049][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.10932657122612, 'train/grad_norm_pre_clip': 29.447500228881836, 'timing/data': 0.0003504629712551832, 'timing/model': 1.1171608119038865, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 1.15}) +[2026-06-24 01:37:11,130][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.05020241439342499, 'train/grad_norm_pre_clip': 33.22010040283203, 'timing/data': 0.00042642303742468357, 'timing/model': 1.1138986919540912, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 1.15}) +[2026-06-24 01:37:15,230][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.08157403022050858, 'train/grad_norm_pre_clip': 19.378921508789062, 'timing/data': 0.0003583739744499326, 'timing/model': 1.1373806230258197, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 1.15}) +[2026-06-24 01:37:19,436][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.1328877955675125, 'train/grad_norm_pre_clip': 22.52189064025879, 'timing/data': 0.00035205797757953405, 'timing/model': 1.1219153729034588, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 1.15}) +[2026-06-24 01:37:23,525][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.051062487065792084, 'train/grad_norm_pre_clip': 22.28352928161621, 'timing/data': 0.0003588589606806636, 'timing/model': 1.1211241679266095, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 1.15}) +[2026-06-24 01:37:27,645][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.04976639151573181, 'train/grad_norm_pre_clip': 38.09539794921875, 'timing/data': 0.000506593962199986, 'timing/model': 1.1260430649854243, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 1.15}) +[2026-06-24 01:37:31,734][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.036084745079278946, 'train/grad_norm_pre_clip': 19.48965835571289, 'timing/data': 0.00035807699896395206, 'timing/model': 1.1220188080333173, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 1.15}) +[2026-06-24 01:37:35,982][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.04343675449490547, 'train/grad_norm_pre_clip': 13.08873176574707, 'timing/data': 0.00040318502578884363, 'timing/model': 1.1194829809246585, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 1.15}) +[2026-06-24 01:37:40,059][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.042502012103796005, 'train/grad_norm_pre_clip': 21.738540649414062, 'timing/data': 0.00036691303830593824, 'timing/model': 1.111349945073016, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 1.15}) +[2026-06-24 01:37:44,145][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.017281245440244675, 'train/grad_norm_pre_clip': 26.061614990234375, 'timing/data': 0.0004051739815622568, 'timing/model': 1.1208460159832612, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 1.15}) +[2026-06-24 01:37:48,232][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.13922354578971863, 'train/grad_norm_pre_clip': 28.127220153808594, 'timing/data': 0.0005523399449884892, 'timing/model': 1.1167620200430974, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 1.15}) +[2026-06-24 01:37:52,387][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.07977701723575592, 'train/grad_norm_pre_clip': 35.8489875793457, 'timing/data': 0.0004036279860883951, 'timing/model': 1.1183315250091255, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 1.15}) +[2026-06-24 01:37:56,471][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.012890802696347237, 'train/grad_norm_pre_clip': 16.61538314819336, 'timing/data': 0.00038525101263076067, 'timing/model': 1.1169973829528317, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 1.15}) +[2026-06-24 01:38:00,548][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.03834812343120575, 'train/grad_norm_pre_clip': 21.03556251525879, 'timing/data': 0.000369207002222538, 'timing/model': 1.116325228009373, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 1.15}) +[2026-06-24 01:38:04,629][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.03835196793079376, 'train/grad_norm_pre_clip': 18.482343673706055, 'timing/data': 0.000349026988260448, 'timing/model': 1.1139553469838575, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 1.15}) +[2026-06-24 01:38:08,767][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.021648231893777847, 'train/grad_norm_pre_clip': 15.64893913269043, 'timing/data': 0.00035993207711726427, 'timing/model': 1.1115041919983923, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 1.15}) +[2026-06-24 01:38:12,844][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.05772660672664642, 'train/grad_norm_pre_clip': 14.033039093017578, 'timing/data': 0.0006246030097827315, 'timing/model': 1.1147445440292358, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 1.15}) +[2026-06-24 01:38:16,924][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.20913425087928772, 'train/grad_norm_pre_clip': 25.40507698059082, 'timing/data': 0.00035307102371007204, 'timing/model': 1.1167068949434906, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 1.16}) +[2026-06-24 01:38:21,010][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.021140174940228462, 'train/grad_norm_pre_clip': 9.544821739196777, 'timing/data': 0.0006994449067860842, 'timing/model': 1.1215976929524913, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 1.16}) +[2026-06-24 01:38:25,271][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.06723345816135406, 'train/grad_norm_pre_clip': 17.412195205688477, 'timing/data': 0.0005221719620749354, 'timing/model': 1.1179654150037095, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 1.16}) +[2026-06-24 01:38:29,368][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.10401955246925354, 'train/grad_norm_pre_clip': 23.13971710205078, 'timing/data': 0.0006767699960619211, 'timing/model': 1.122706368099898, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 1.16}) +[2026-06-24 01:38:33,458][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.009101491421461105, 'train/grad_norm_pre_clip': 16.83561134338379, 'timing/data': 0.0004198249662294984, 'timing/model': 1.1158737940713763, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 1.16}) +[2026-06-24 01:38:37,649][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.021929148584604263, 'train/grad_norm_pre_clip': 12.774409294128418, 'timing/data': 0.0003545680083334446, 'timing/model': 1.2234894280554727, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 1.16}) +[2026-06-24 01:38:41,734][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.026264671236276627, 'train/grad_norm_pre_clip': 22.740306854248047, 'timing/data': 0.0003462139284238219, 'timing/model': 1.1266759709687904, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.16}) +[2026-06-24 01:38:45,828][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.007872134447097778, 'train/grad_norm_pre_clip': 8.273605346679688, 'timing/data': 0.0006101300241425633, 'timing/model': 1.1237024600850418, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 1.16}) +[2026-06-24 01:38:49,900][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.049486223608255386, 'train/grad_norm_pre_clip': 21.244516372680664, 'timing/data': 0.00036119495052844286, 'timing/model': 1.1142263589426875, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 1.16}) +[2026-06-24 01:38:54,170][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.05608672276139259, 'train/grad_norm_pre_clip': 17.601192474365234, 'timing/data': 0.0004323510220274329, 'timing/model': 1.3008611700497568, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 1.16}) +[2026-06-24 01:38:58,260][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.0004907645634375513, 'train/grad_norm_pre_clip': 23.12922477722168, 'timing/data': 0.00036248296964913607, 'timing/model': 1.1233213539235294, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 1.16}) +[2026-06-24 01:39:02,359][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.08096294850111008, 'train/grad_norm_pre_clip': 13.131622314453125, 'timing/data': 0.00039251602720469236, 'timing/model': 1.1232469629030675, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 1.16}) +[2026-06-24 01:39:06,469][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.023443423211574554, 'train/grad_norm_pre_clip': 19.89235496520996, 'timing/data': 0.0003572050482034683, 'timing/model': 1.133108134032227, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 1.16}) +[2026-06-24 01:39:10,634][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.008006691001355648, 'train/grad_norm_pre_clip': 33.839508056640625, 'timing/data': 0.0005679529858753085, 'timing/model': 1.1181907888967544, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 1.16}) +[2026-06-24 01:39:14,734][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.07910044491291046, 'train/grad_norm_pre_clip': 12.40247917175293, 'timing/data': 0.00038335693534463644, 'timing/model': 1.1216717779170722, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 1.16}) +[2026-06-24 01:39:18,827][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3000 +[2026-06-24 01:40:22,786][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3000 +[2026-06-24 01:40:22,788][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3000 +[2026-06-24 01:41:02,145][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3000 +[2026-06-24 01:49:35,879][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3000 +[2026-06-24 01:49:35,879][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.16756045818328857, 'train/grad_norm_pre_clip': 30.09198760986328, 'eval/action_loss/seconds': 63.955877874046564, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.029622018337249756, 'eval/latency_2/loss': 0.029622018337249756, 'eval/flappy/loss': 0.029622018337249756, 'eval/flappy/latency_2/loss': 0.029622018337249756, 'eval/flap_precision': 0.9741847826086957, 'eval/flap_recall': 0.9728629579375848, 'eval/flap_f1': 0.9735234215885947, 'eval/flap_rate_model': 0.20444444444444446, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9735234215885947, 'eval/latency_2/flap_precision': 0.9741847826086957, 'eval/latency_2/flap_recall': 0.9728629579375848, 'eval/latency_2/flap_f1': 0.9735234215885947, 'eval/latency_2/flap_rate_model': 0.20444444444444446, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9735234215885947, 'eval/action_classification/data_wait_seconds': 0.6193892955780029, 'eval/action_classification/predict_seconds': 38.58260726928711, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 317.72000000001435, 'rl_games_eval/mid_train/mean_length': 2578.6, 'rl_games_eval/mid_train/std_reward': 162.01944821534275, 'rl_games_eval/mid_train/std_length': 1304.6114517357264, 'rl_games_eval/mid_train/macro_mean_reward': 317.72000000001435, 'rl_games_eval/mid_train/macro_mean_length': 2578.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 317.72000000001435, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2578.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 162.01944821534275, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1304.6114517357264, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 317.72000000001435, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003547919914126396, 'timing/model': 1.1252611150266603, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 1.16}) +[2026-06-24 01:49:35,880][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_3000_state +[2026-06-24 01:49:35,882][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 01:49:39,588][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_3000_state/model.safetensors +[2026-06-24 01:49:47,300][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_3000_state/optimizer.bin +[2026-06-24 01:49:47,301][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_3000_state/sampler.bin +[2026-06-24 01:49:47,301][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-24 01:49:47,302][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-24 01:49:47,380][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-24 01:49:47,380][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 01:49:47,393][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 01:49:51,349][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.019894124940037727, 'train/grad_norm_pre_clip': 14.700844764709473, 'timing/data': 0.00043004192411899567, 'timing/model': 1.0865206390153617, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 1.16}) +[2026-06-24 01:49:55,318][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.04377266392111778, 'train/grad_norm_pre_clip': 30.839275360107422, 'timing/data': 0.0004810020327568054, 'timing/model': 1.087888985988684, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 1.16}) +[2026-06-24 01:49:59,384][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.036934517323970795, 'train/grad_norm_pre_clip': 5.863966941833496, 'timing/data': 0.0004140048986300826, 'timing/model': 1.0925971580436453, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 1.16}) +[2026-06-24 01:50:03,377][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.023114658892154694, 'train/grad_norm_pre_clip': 16.415904998779297, 'timing/data': 0.0004901539068669081, 'timing/model': 1.1041149250231683, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 1.16}) +[2026-06-24 01:50:07,378][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.0008783834055066109, 'train/grad_norm_pre_clip': 21.105798721313477, 'timing/data': 0.0005581820150837302, 'timing/model': 1.0985606079921126, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 1.16}) +[2026-06-24 01:50:11,373][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.1529291868209839, 'train/grad_norm_pre_clip': 31.783401489257812, 'timing/data': 0.0004092729650437832, 'timing/model': 1.0990533729782328, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 1.16}) +[2026-06-24 01:50:15,462][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.018958566710352898, 'train/grad_norm_pre_clip': 13.308609962463379, 'timing/data': 0.0005030670436099172, 'timing/model': 1.1010726400418207, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 1.16}) +[2026-06-24 01:50:19,489][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.05952376499772072, 'train/grad_norm_pre_clip': 41.158050537109375, 'timing/data': 0.00041397393215447664, 'timing/model': 1.1126646699849516, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 1.16}) +[2026-06-24 01:50:23,516][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.006341330707073212, 'train/grad_norm_pre_clip': 20.712020874023438, 'timing/data': 0.000599185936152935, 'timing/model': 1.102013297029771, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 1.16}) +[2026-06-24 01:50:27,563][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.014385031536221504, 'train/grad_norm_pre_clip': 20.059953689575195, 'timing/data': 0.00035920203663408756, 'timing/model': 1.1132743700873107, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 1.16}) +[2026-06-24 01:50:31,665][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.023849355056881905, 'train/grad_norm_pre_clip': 10.788305282592773, 'timing/data': 0.0003591360291466117, 'timing/model': 1.1056691720150411, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 1.17}) +[2026-06-24 01:50:35,712][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.06602182239294052, 'train/grad_norm_pre_clip': 16.812969207763672, 'timing/data': 0.0003504339838400483, 'timing/model': 1.109141501947306, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 1.17}) +[2026-06-24 01:50:39,756][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.011584734544157982, 'train/grad_norm_pre_clip': 9.346086502075195, 'timing/data': 0.0003453639801591635, 'timing/model': 1.110882233013399, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 1.17}) +[2026-06-24 01:50:43,810][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.01709258370101452, 'train/grad_norm_pre_clip': 11.109234809875488, 'timing/data': 0.0005635099951177835, 'timing/model': 1.1108656850410625, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 1.17}) +[2026-06-24 01:50:47,937][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.0014713103882968426, 'train/grad_norm_pre_clip': 6.898327827453613, 'timing/data': 0.00042813702020794153, 'timing/model': 1.1097800449933857, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 1.17}) +[2026-06-24 01:50:51,984][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.028437532484531403, 'train/grad_norm_pre_clip': 11.328924179077148, 'timing/data': 0.00035910599399358034, 'timing/model': 1.1162896821042523, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 1.17}) +[2026-06-24 01:50:56,051][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.01408635824918747, 'train/grad_norm_pre_clip': 9.631213188171387, 'timing/data': 0.00040030095260590315, 'timing/model': 1.1136577440192923, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 1.17}) +[2026-06-24 01:51:00,131][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.026935117319226265, 'train/grad_norm_pre_clip': 11.797781944274902, 'timing/data': 0.0003466210328042507, 'timing/model': 1.1319326830562204, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 1.17}) +[2026-06-24 01:51:04,321][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.05536423251032829, 'train/grad_norm_pre_clip': 33.98118591308594, 'timing/data': 0.0004859629552811384, 'timing/model': 1.1124749389709905, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 1.17}) +[2026-06-24 01:51:08,389][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.055039286613464355, 'train/grad_norm_pre_clip': 25.24601936340332, 'timing/data': 0.0005391070153564215, 'timing/model': 1.1118516500573605, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 1.17}) +[2026-06-24 01:51:12,445][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.06279084086418152, 'train/grad_norm_pre_clip': 19.806535720825195, 'timing/data': 0.00048188993241637945, 'timing/model': 1.109103249036707, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 1.17}) +[2026-06-24 01:51:16,509][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.12460969388484955, 'train/grad_norm_pre_clip': 27.71184539794922, 'timing/data': 0.0003945640055462718, 'timing/model': 1.1108089019544423, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 1.17}) +[2026-06-24 01:51:20,651][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.024799102917313576, 'train/grad_norm_pre_clip': 20.07672691345215, 'timing/data': 0.0005547220353037119, 'timing/model': 1.1119441810296848, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 1.17}) +[2026-06-24 01:51:24,730][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.08656298369169235, 'train/grad_norm_pre_clip': 16.76603889465332, 'timing/data': 0.00040704908315092325, 'timing/model': 1.1161908849608153, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 1.17}) +[2026-06-24 01:51:28,799][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.007997690699994564, 'train/grad_norm_pre_clip': 9.777294158935547, 'timing/data': 0.0003737310180440545, 'timing/model': 1.117019159020856, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 1.17}) +[2026-06-24 01:51:32,870][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.1195809617638588, 'train/grad_norm_pre_clip': 26.700599670410156, 'timing/data': 0.0005068109603598714, 'timing/model': 1.109850803972222, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 1.17}) +[2026-06-24 01:51:37,013][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.014076126739382744, 'train/grad_norm_pre_clip': 10.40320873260498, 'timing/data': 0.000364616047590971, 'timing/model': 1.1150144869461656, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 1.17}) +[2026-06-24 01:51:41,087][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.01675996743142605, 'train/grad_norm_pre_clip': 23.524816513061523, 'timing/data': 0.0003565809456631541, 'timing/model': 1.115954456035979, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 1.17}) +[2026-06-24 01:51:45,148][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.005150625016540289, 'train/grad_norm_pre_clip': 18.708967208862305, 'timing/data': 0.0005633770488202572, 'timing/model': 1.1102503458969295, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 1.17}) +[2026-06-24 01:51:49,253][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.25711771845817566, 'train/grad_norm_pre_clip': 23.00063133239746, 'timing/data': 0.0005672289989888668, 'timing/model': 1.1250810381025076, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 1.17}) +[2026-06-24 01:51:53,425][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.0027817583177238703, 'train/grad_norm_pre_clip': 28.455421447753906, 'timing/data': 0.00044301000889390707, 'timing/model': 1.1142103070160374, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 1.17}) +[2026-06-24 01:51:57,501][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.0060901581309735775, 'train/grad_norm_pre_clip': 12.163178443908691, 'timing/data': 0.00035624008160084486, 'timing/model': 1.116893416037783, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 1.17}) +[2026-06-24 01:52:01,575][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.2566267251968384, 'train/grad_norm_pre_clip': 31.16077995300293, 'timing/data': 0.000354045070707798, 'timing/model': 1.1156046510441229, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 1.17}) +[2026-06-24 01:52:05,659][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.13176673650741577, 'train/grad_norm_pre_clip': 26.063222885131836, 'timing/data': 0.00035932997707277536, 'timing/model': 1.115194009966217, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 1.17}) +[2026-06-24 01:52:09,828][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.07813209295272827, 'train/grad_norm_pre_clip': 15.622536659240723, 'timing/data': 0.0003618620103225112, 'timing/model': 1.1133625829825178, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 1.17}) +[2026-06-24 01:52:13,902][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.0325176864862442, 'train/grad_norm_pre_clip': 46.755897521972656, 'timing/data': 0.00035555707290768623, 'timing/model': 1.1172277390724048, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 1.17}) +[2026-06-24 01:52:17,966][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.008353456854820251, 'train/grad_norm_pre_clip': 24.108304977416992, 'timing/data': 0.00035962206311523914, 'timing/model': 1.1141594810178503, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 1.18}) +[2026-06-24 01:52:22,052][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.016588648781180382, 'train/grad_norm_pre_clip': 39.137916564941406, 'timing/data': 0.00036642199847847223, 'timing/model': 1.11797939799726, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 1.18}) +[2026-06-24 01:52:26,215][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.00209215865470469, 'train/grad_norm_pre_clip': 16.62154769897461, 'timing/data': 0.00034624303225427866, 'timing/model': 1.1169004259863868, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 1.18}) +[2026-06-24 01:52:30,313][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.047333650290966034, 'train/grad_norm_pre_clip': 21.736101150512695, 'timing/data': 0.00036331498995423317, 'timing/model': 1.1242064479738474, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.18}) +[2026-06-24 01:52:34,389][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.0073117041029036045, 'train/grad_norm_pre_clip': 25.17168617248535, 'timing/data': 0.0004021000349894166, 'timing/model': 1.1196078300708905, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 1.18}) +[2026-06-24 01:52:38,471][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.021977193653583527, 'train/grad_norm_pre_clip': 28.473163604736328, 'timing/data': 0.0005428049480542541, 'timing/model': 1.1191044880542904, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 1.18}) +[2026-06-24 01:52:42,615][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.04034774750471115, 'train/grad_norm_pre_clip': 8.79056453704834, 'timing/data': 0.0005286569939926267, 'timing/model': 1.1170418560504913, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 1.18}) +[2026-06-24 01:52:46,685][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.001525231171399355, 'train/grad_norm_pre_clip': 28.023681640625, 'timing/data': 0.0003453519893810153, 'timing/model': 1.118801370030269, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 1.18}) +[2026-06-24 01:52:50,773][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.012887043878436089, 'train/grad_norm_pre_clip': 16.748729705810547, 'timing/data': 0.00040498096495866776, 'timing/model': 1.1187925309641287, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 1.18}) +[2026-06-24 01:52:54,857][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.14446216821670532, 'train/grad_norm_pre_clip': 25.76835823059082, 'timing/data': 0.00036081194411963224, 'timing/model': 1.1255929309409112, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 1.18}) +[2026-06-24 01:52:59,036][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.008078513666987419, 'train/grad_norm_pre_clip': 10.503095626831055, 'timing/data': 0.00037174299359321594, 'timing/model': 1.1164317290531471, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 1.18}) +[2026-06-24 01:53:03,122][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.06875676661729813, 'train/grad_norm_pre_clip': 12.027549743652344, 'timing/data': 0.00035704695619642735, 'timing/model': 1.121136169997044, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 1.18}) +[2026-06-24 01:53:07,195][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.11238621920347214, 'train/grad_norm_pre_clip': 20.457611083984375, 'timing/data': 0.0003554190043359995, 'timing/model': 1.113543945015408, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 1.18}) +[2026-06-24 01:53:11,281][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.016329502686858177, 'train/grad_norm_pre_clip': 14.799107551574707, 'timing/data': 0.0004986790008842945, 'timing/model': 1.1210269370349124, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 1.18}) +[2026-06-24 01:53:15,437][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.033420201390981674, 'train/grad_norm_pre_clip': 19.833463668823242, 'timing/data': 0.0003473380347713828, 'timing/model': 1.116000184090808, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 1.18}) +[2026-06-24 01:53:19,509][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.07020340114831924, 'train/grad_norm_pre_clip': 27.492294311523438, 'timing/data': 0.0003590720007196069, 'timing/model': 1.1187243179883808, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 1.18}) +[2026-06-24 01:53:23,592][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.04602660611271858, 'train/grad_norm_pre_clip': 22.599166870117188, 'timing/data': 0.00039350998122245073, 'timing/model': 1.113734768005088, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 1.18}) +[2026-06-24 01:53:27,657][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.02692709118127823, 'train/grad_norm_pre_clip': 11.531014442443848, 'timing/data': 0.0004125339910387993, 'timing/model': 1.1162600390380248, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 1.18}) +[2026-06-24 01:53:31,798][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.044798947870731354, 'train/grad_norm_pre_clip': 38.1293830871582, 'timing/data': 0.0005067290039733052, 'timing/model': 1.1206648190272972, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 1.18}) +[2026-06-24 01:53:35,890][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.0509357675909996, 'train/grad_norm_pre_clip': 17.60343360900879, 'timing/data': 0.0004277440020814538, 'timing/model': 1.1302118679741398, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 1.18}) +[2026-06-24 01:53:40,041][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.13667502999305725, 'train/grad_norm_pre_clip': 14.901933670043945, 'timing/data': 0.0004400040488690138, 'timing/model': 1.1202506710542366, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 1.18}) +[2026-06-24 01:53:44,121][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.022630637511610985, 'train/grad_norm_pre_clip': 11.514461517333984, 'timing/data': 0.00048059807159006596, 'timing/model': 1.1150120829697698, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 1.18}) +[2026-06-24 01:53:48,301][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.1495496779680252, 'train/grad_norm_pre_clip': 14.23401927947998, 'timing/data': 0.0004060049541294575, 'timing/model': 1.1194086549803615, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 1.18}) +[2026-06-24 01:53:52,369][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.10581613332033157, 'train/grad_norm_pre_clip': 29.99921417236328, 'timing/data': 0.00048681499902158976, 'timing/model': 1.117826160043478, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 1.18}) +[2026-06-24 01:53:56,459][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.04782378301024437, 'train/grad_norm_pre_clip': 16.20087432861328, 'timing/data': 0.00035876804031431675, 'timing/model': 1.1200693460414186, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 1.18}) +[2026-06-24 01:54:00,530][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.03256736695766449, 'train/grad_norm_pre_clip': 11.426074028015137, 'timing/data': 0.0005202079191803932, 'timing/model': 1.116871219011955, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 1.19}) +[2026-06-24 01:54:04,691][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.06817328184843063, 'train/grad_norm_pre_clip': 8.490462303161621, 'timing/data': 0.00035269593354314566, 'timing/model': 1.1165722300065681, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 1.19}) +[2026-06-24 01:54:08,776][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.01894087716937065, 'train/grad_norm_pre_clip': 13.222140312194824, 'timing/data': 0.0004026679089292884, 'timing/model': 1.1223533790325746, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 1.19}) +[2026-06-24 01:54:12,865][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.006175104528665543, 'train/grad_norm_pre_clip': 12.314959526062012, 'timing/data': 0.0003523279447108507, 'timing/model': 1.1165877939201891, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 1.19}) +[2026-06-24 01:54:16,927][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.007980103604495525, 'train/grad_norm_pre_clip': 18.97209358215332, 'timing/data': 0.00042696099262684584, 'timing/model': 1.1088214439805597, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 1.19}) +[2026-06-24 01:54:21,115][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.05477983132004738, 'train/grad_norm_pre_clip': 10.313603401184082, 'timing/data': 0.0003575680311769247, 'timing/model': 1.1190833949949592, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 1.19}) +[2026-06-24 01:54:25,196][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.07920379936695099, 'train/grad_norm_pre_clip': 13.347026824951172, 'timing/data': 0.0005555349634960294, 'timing/model': 1.1262144850334153, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 1.19}) +[2026-06-24 01:54:29,282][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.03226438909769058, 'train/grad_norm_pre_clip': 16.599063873291016, 'timing/data': 0.0003643441013991833, 'timing/model': 1.1179187149973586, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 1.19}) +[2026-06-24 01:54:33,370][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.007987085729837418, 'train/grad_norm_pre_clip': 17.44162368774414, 'timing/data': 0.00035690097138285637, 'timing/model': 1.1223894100403413, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 1.19}) +[2026-06-24 01:54:37,530][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.03279826045036316, 'train/grad_norm_pre_clip': 22.59406280517578, 'timing/data': 0.0003546440275385976, 'timing/model': 1.1171249729814008, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 1.19}) +[2026-06-24 01:54:41,610][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.10653544962406158, 'train/grad_norm_pre_clip': 22.84964942932129, 'timing/data': 0.0003564530052244663, 'timing/model': 1.1177428879309446, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 1.19}) +[2026-06-24 01:54:45,698][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.041471973061561584, 'train/grad_norm_pre_clip': 10.262795448303223, 'timing/data': 0.00035921495873481035, 'timing/model': 1.1227325630607083, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 1.19}) +[2026-06-24 01:54:49,777][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.13140356540679932, 'train/grad_norm_pre_clip': 13.582356452941895, 'timing/data': 0.0005198399303480983, 'timing/model': 1.118188341963105, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 1.19}) +[2026-06-24 01:54:53,936][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.0014473324408754706, 'train/grad_norm_pre_clip': 17.36752700805664, 'timing/data': 0.00041397602763026953, 'timing/model': 1.1189217100618407, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 1.19}) +[2026-06-24 01:54:58,022][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.023278387263417244, 'train/grad_norm_pre_clip': 10.152570724487305, 'timing/data': 0.000352884060703218, 'timing/model': 1.11805352603551, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 1.19}) +[2026-06-24 01:55:02,110][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.059079188853502274, 'train/grad_norm_pre_clip': 34.113033294677734, 'timing/data': 0.0003608259139582515, 'timing/model': 1.1157199770677835, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 1.19}) +[2026-06-24 01:55:06,202][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.013770656660199165, 'train/grad_norm_pre_clip': 13.775599479675293, 'timing/data': 0.0006281889509409666, 'timing/model': 1.125441447016783, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 1.19}) +[2026-06-24 01:55:10,361][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.002160855568945408, 'train/grad_norm_pre_clip': 26.04633903503418, 'timing/data': 0.0003418790875002742, 'timing/model': 1.1248370669782162, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 1.19}) +[2026-06-24 01:55:14,458][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.049557484686374664, 'train/grad_norm_pre_clip': 27.590389251708984, 'timing/data': 0.00039025803562253714, 'timing/model': 1.1252631110837683, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 1.19}) +[2026-06-24 01:55:18,541][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.002750289160758257, 'train/grad_norm_pre_clip': 9.081250190734863, 'timing/data': 0.00037790299393236637, 'timing/model': 1.1206929830368608, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 1.19}) +[2026-06-24 01:55:22,632][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.03607400879263878, 'train/grad_norm_pre_clip': 34.84473419189453, 'timing/data': 0.0006763609126210213, 'timing/model': 1.1286564599722624, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 1.19}) +[2026-06-24 01:55:26,813][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.001276549301110208, 'train/grad_norm_pre_clip': 34.84456253051758, 'timing/data': 0.0005690960679203272, 'timing/model': 1.1223354840185493, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 1.19}) +[2026-06-24 01:55:30,901][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.011590479873120785, 'train/grad_norm_pre_clip': 18.359970092773438, 'timing/data': 0.000387480016797781, 'timing/model': 1.1171799449948594, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 1.19}) +[2026-06-24 01:55:34,982][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.021697742864489555, 'train/grad_norm_pre_clip': 22.97265625, 'timing/data': 0.0004256489919498563, 'timing/model': 1.1179021979914978, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 1.19}) +[2026-06-24 01:55:39,070][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.08392336219549179, 'train/grad_norm_pre_clip': 22.606489181518555, 'timing/data': 0.0005054549546912313, 'timing/model': 1.121885106083937, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 1.19}) +[2026-06-24 01:55:43,241][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.08812329918146133, 'train/grad_norm_pre_clip': 20.220640182495117, 'timing/data': 0.0003605239326134324, 'timing/model': 1.122207374079153, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 1.19}) +[2026-06-24 01:55:47,333][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.015324163250625134, 'train/grad_norm_pre_clip': 14.991392135620117, 'timing/data': 0.00037498597521334887, 'timing/model': 1.1190400019986555, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 1.2}) +[2026-06-24 01:55:51,419][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.18041324615478516, 'train/grad_norm_pre_clip': 27.097318649291992, 'timing/data': 0.0003560950281098485, 'timing/model': 1.1257408739766106, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.2}) +[2026-06-24 01:55:55,508][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.05067753791809082, 'train/grad_norm_pre_clip': 19.591093063354492, 'timing/data': 0.00036331196315586567, 'timing/model': 1.124372588004917, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 1.2}) +[2026-06-24 01:55:59,653][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.04149181768298149, 'train/grad_norm_pre_clip': 14.586545944213867, 'timing/data': 0.00036533898673951626, 'timing/model': 1.1121591359842569, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 1.2}) +[2026-06-24 01:56:03,738][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.005170675925910473, 'train/grad_norm_pre_clip': 10.240882873535156, 'timing/data': 0.00035971193574368954, 'timing/model': 1.1166917859809473, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 1.2}) +[2026-06-24 01:56:07,821][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.011430385522544384, 'train/grad_norm_pre_clip': 28.542129516601562, 'timing/data': 0.00035705603659152985, 'timing/model': 1.1191427720477805, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 1.2}) +[2026-06-24 01:56:11,895][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.027865037322044373, 'train/grad_norm_pre_clip': 4.412681579589844, 'timing/data': 0.0003705149283632636, 'timing/model': 1.1199519290821627, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 1.2}) +[2026-06-24 01:56:16,067][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.07546701282262802, 'train/grad_norm_pre_clip': 18.7169246673584, 'timing/data': 0.00036127003841102123, 'timing/model': 1.1177067949902266, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 1.2}) +[2026-06-24 01:56:20,154][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.09145954996347427, 'train/grad_norm_pre_clip': 14.80075740814209, 'timing/data': 0.0003688859287649393, 'timing/model': 1.1175106619484723, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 1.2}) +[2026-06-24 01:56:24,254][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.03935634717345238, 'train/grad_norm_pre_clip': 14.026656150817871, 'timing/data': 0.0006144549697637558, 'timing/model': 1.1278888100059703, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 1.2}) +[2026-06-24 01:56:28,339][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.051675837486982346, 'train/grad_norm_pre_clip': 6.4020843505859375, 'timing/data': 0.00044370407704263926, 'timing/model': 1.1206985120661557, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 1.2}) +[2026-06-24 01:56:32,522][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.006434638053178787, 'train/grad_norm_pre_clip': 18.159908294677734, 'timing/data': 0.0003623299999162555, 'timing/model': 1.1233906300039962, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 1.2}) +[2026-06-24 01:56:36,597][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3100 +[2026-06-24 01:57:40,923][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3100 +[2026-06-24 01:57:40,924][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3100 +[2026-06-24 01:58:19,281][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3100 +[2026-06-24 01:58:19,282][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.026768261566758156, 'train/grad_norm_pre_clip': 28.135984420776367, 'eval/action_loss/seconds': 64.3239372529788, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.031013181433081627, 'eval/latency_2/loss': 0.031013181433081627, 'eval/flappy/loss': 0.031013181433081627, 'eval/flappy/latency_2/loss': 0.031013181433081627, 'eval/flap_precision': 0.9666666666666667, 'eval/flap_recall': 0.9837177747625508, 'eval/flap_f1': 0.9751176866173503, 'eval/flap_rate_model': 0.20833333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9751176866173504, 'eval/latency_2/flap_precision': 0.9666666666666667, 'eval/latency_2/flap_recall': 0.9837177747625508, 'eval/latency_2/flap_f1': 0.9751176866173503, 'eval/latency_2/flap_rate_model': 0.20833333333333334, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9751176866173504, 'eval/action_classification/data_wait_seconds': 0.4432821273803711, 'eval/action_classification/predict_seconds': 37.78398895263672, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00039930595085024834, 'timing/model': 1.1223704600706697, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 1.2}) +[2026-06-24 01:58:23,318][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.01038372702896595, 'train/grad_norm_pre_clip': 11.018229484558105, 'timing/data': 0.0006184730445966125, 'timing/model': 1.10638321202714, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 1.2}) +[2026-06-24 01:58:27,440][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.05536193773150444, 'train/grad_norm_pre_clip': 24.04286003112793, 'timing/data': 0.0003513640258461237, 'timing/model': 1.1075512620154768, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 1.2}) +[2026-06-24 01:58:31,480][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.01632634364068508, 'train/grad_norm_pre_clip': 14.339864730834961, 'timing/data': 0.0003567709354683757, 'timing/model': 1.104258370003663, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 1.2}) +[2026-06-24 01:58:35,542][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.026430832222104073, 'train/grad_norm_pre_clip': 20.127538681030273, 'timing/data': 0.0005879739765077829, 'timing/model': 1.1185451649362221, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 1.2}) +[2026-06-24 01:58:39,598][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.06615911424160004, 'train/grad_norm_pre_clip': 22.775602340698242, 'timing/data': 0.0004950129659846425, 'timing/model': 1.1164397590328008, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 1.2}) +[2026-06-24 01:58:43,788][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.01226255763322115, 'train/grad_norm_pre_clip': 43.78158187866211, 'timing/data': 0.0005489559844136238, 'timing/model': 1.109634091029875, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 1.2}) +[2026-06-24 01:58:47,842][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.012575293891131878, 'train/grad_norm_pre_clip': 6.732469081878662, 'timing/data': 0.000352125964127481, 'timing/model': 1.1148920599371195, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 1.2}) +[2026-06-24 01:58:51,919][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.02326446957886219, 'train/grad_norm_pre_clip': 23.852840423583984, 'timing/data': 0.0003631910076364875, 'timing/model': 1.1135475520277396, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 1.2}) +[2026-06-24 01:58:56,011][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.0018115367274731398, 'train/grad_norm_pre_clip': 19.305055618286133, 'timing/data': 0.0006432710215449333, 'timing/model': 1.1162864400539547, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 1.2}) +[2026-06-24 01:59:00,184][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.19412709772586823, 'train/grad_norm_pre_clip': 46.557796478271484, 'timing/data': 0.0003492860123515129, 'timing/model': 1.1144800590118393, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 1.2}) +[2026-06-24 01:59:04,251][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.04127277806401253, 'train/grad_norm_pre_clip': 19.131412506103516, 'timing/data': 0.00034749298356473446, 'timing/model': 1.1158017580164596, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 1.2}) +[2026-06-24 01:59:08,333][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.031513866037130356, 'train/grad_norm_pre_clip': 19.182594299316406, 'timing/data': 0.0005975270178169012, 'timing/model': 1.1200097650289536, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 1.2}) +[2026-06-24 01:59:12,412][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.006591789424419403, 'train/grad_norm_pre_clip': 25.262727737426758, 'timing/data': 0.0004905390087515116, 'timing/model': 1.1131311160279438, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 1.2}) +[2026-06-24 01:59:16,563][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.020978683605790138, 'train/grad_norm_pre_clip': 17.7379207611084, 'timing/data': 0.0003574059810489416, 'timing/model': 1.1165238809771836, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 1.21}) +[2026-06-24 01:59:20,644][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.28990858793258667, 'train/grad_norm_pre_clip': 27.757415771484375, 'timing/data': 0.0004286270122975111, 'timing/model': 1.1187179909320548, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 1.21}) +[2026-06-24 01:59:24,710][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.03412806615233421, 'train/grad_norm_pre_clip': 38.221561431884766, 'timing/data': 0.0003573999274522066, 'timing/model': 1.1164682690287009, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 1.21}) +[2026-06-24 01:59:28,789][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.017682867124676704, 'train/grad_norm_pre_clip': 17.273984909057617, 'timing/data': 0.0004088300047442317, 'timing/model': 1.114357857964933, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 1.21}) +[2026-06-24 01:59:33,031][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.03055289015173912, 'train/grad_norm_pre_clip': 16.739818572998047, 'timing/data': 0.0003498750738799572, 'timing/model': 1.1178799760527909, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 1.21}) +[2026-06-24 01:59:37,114][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.02786969393491745, 'train/grad_norm_pre_clip': 19.002281188964844, 'timing/data': 0.00036677310708910227, 'timing/model': 1.1171927149407566, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 1.21}) +[2026-06-24 01:59:41,184][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.0076044523157179356, 'train/grad_norm_pre_clip': 32.267822265625, 'timing/data': 0.00035539502277970314, 'timing/model': 1.1214998420327902, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 1.21}) +[2026-06-24 01:59:45,250][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.02945755608379841, 'train/grad_norm_pre_clip': 9.299129486083984, 'timing/data': 0.0007401210023090243, 'timing/model': 1.1150146080181003, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 1.21}) +[2026-06-24 01:59:49,430][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.009320813231170177, 'train/grad_norm_pre_clip': 30.49091148376465, 'timing/data': 0.0003580399788916111, 'timing/model': 1.1137362220324576, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 1.21}) +[2026-06-24 01:59:53,512][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.021430684253573418, 'train/grad_norm_pre_clip': 28.474563598632812, 'timing/data': 0.0007591290632262826, 'timing/model': 1.1168729029595852, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 1.21}) +[2026-06-24 01:59:57,596][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.0494341216981411, 'train/grad_norm_pre_clip': 71.83181762695312, 'timing/data': 0.0003598160110414028, 'timing/model': 1.1198677819920704, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 1.21}) +[2026-06-24 02:00:01,688][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.11420822143554688, 'train/grad_norm_pre_clip': 22.121797561645508, 'timing/data': 0.00034358096309006214, 'timing/model': 1.116716630058363, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 1.21}) +[2026-06-24 02:00:05,935][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.023249411955475807, 'train/grad_norm_pre_clip': 6.205899238586426, 'timing/data': 0.00035107205621898174, 'timing/model': 1.1190690839430317, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 1.21}) +[2026-06-24 02:00:10,014][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.03550836443901062, 'train/grad_norm_pre_clip': 23.537439346313477, 'timing/data': 0.0003582020290195942, 'timing/model': 1.1143756919773296, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 1.21}) +[2026-06-24 02:00:14,101][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.1497795134782791, 'train/grad_norm_pre_clip': 15.688328742980957, 'timing/data': 0.00034854409750550985, 'timing/model': 1.123641602927819, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 1.21}) +[2026-06-24 02:00:18,180][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.022453974932432175, 'train/grad_norm_pre_clip': 21.876619338989258, 'timing/data': 0.0003534100251272321, 'timing/model': 1.1166853450704366, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 1.21}) +[2026-06-24 02:00:22,329][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.003541596233844757, 'train/grad_norm_pre_clip': 26.48924446105957, 'timing/data': 0.00040052796248346567, 'timing/model': 1.1207883770111948, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 1.21}) +[2026-06-24 02:00:26,421][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.03197275102138519, 'train/grad_norm_pre_clip': 21.378944396972656, 'timing/data': 0.00037789694033563137, 'timing/model': 1.1147747719660401, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 1.21}) +[2026-06-24 02:00:30,495][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.003252065973356366, 'train/grad_norm_pre_clip': 8.961088180541992, 'timing/data': 0.0003993869759142399, 'timing/model': 1.1166024709818885, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 1.21}) +[2026-06-24 02:00:34,578][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.02301488257944584, 'train/grad_norm_pre_clip': 21.974979400634766, 'timing/data': 0.0003574270522221923, 'timing/model': 1.1179785049753264, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 1.21}) +[2026-06-24 02:00:38,735][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.0009419643320143223, 'train/grad_norm_pre_clip': 23.845666885375977, 'timing/data': 0.00046095007564872503, 'timing/model': 1.118085422902368, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 1.21}) +[2026-06-24 02:00:42,796][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.032466962933540344, 'train/grad_norm_pre_clip': 27.406267166137695, 'timing/data': 0.0004048130940645933, 'timing/model': 1.1156731200171635, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 1.21}) +[2026-06-24 02:00:46,866][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.0022177102509886026, 'train/grad_norm_pre_clip': 22.640066146850586, 'timing/data': 0.0003572771092876792, 'timing/model': 1.1182292989687994, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 1.21}) +[2026-06-24 02:00:50,933][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.0900612473487854, 'train/grad_norm_pre_clip': 24.6994686126709, 'timing/data': 0.0005026910221204162, 'timing/model': 1.1134923059726134, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 1.21}) +[2026-06-24 02:00:55,149][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.20392107963562012, 'train/grad_norm_pre_clip': 28.513587951660156, 'timing/data': 0.000375378062017262, 'timing/model': 1.1140371030196548, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.21}) +[2026-06-24 02:00:59,218][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.04309244081377983, 'train/grad_norm_pre_clip': 17.638792037963867, 'timing/data': 0.00040393706876784563, 'timing/model': 1.1169699489837512, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 1.21}) +[2026-06-24 02:01:03,297][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.02622867375612259, 'train/grad_norm_pre_clip': 13.004849433898926, 'timing/data': 0.00041012302972376347, 'timing/model': 1.1183302729623392, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 1.22}) +[2026-06-24 02:01:07,460][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.09256426244974136, 'train/grad_norm_pre_clip': 18.708215713500977, 'timing/data': 0.0005230100359767675, 'timing/model': 1.2074105530045927, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 1.22}) +[2026-06-24 02:01:11,547][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.06342469155788422, 'train/grad_norm_pre_clip': 30.66924285888672, 'timing/data': 0.0004105969564989209, 'timing/model': 1.1170684930402786, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 1.22}) +[2026-06-24 02:01:15,627][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.14827758073806763, 'train/grad_norm_pre_clip': 34.62984848022461, 'timing/data': 0.000544237089343369, 'timing/model': 1.1136427499586716, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 1.22}) +[2026-06-24 02:01:19,718][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.02645450085401535, 'train/grad_norm_pre_clip': 10.408751487731934, 'timing/data': 0.000615105964243412, 'timing/model': 1.125011133030057, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 1.22}) +[2026-06-24 02:01:23,984][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.24457134306430817, 'train/grad_norm_pre_clip': 28.685564041137695, 'timing/data': 0.0003659740323200822, 'timing/model': 1.2954147269483656, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 1.22}) +[2026-06-24 02:01:28,072][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.007538686506450176, 'train/grad_norm_pre_clip': 19.25459861755371, 'timing/data': 0.000363870058208704, 'timing/model': 1.1174548029666767, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 1.22}) +[2026-06-24 02:01:32,151][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.05764057859778404, 'train/grad_norm_pre_clip': 25.368427276611328, 'timing/data': 0.0003809909103438258, 'timing/model': 1.115235322038643, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 1.22}) +[2026-06-24 02:01:36,239][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.049759332090616226, 'train/grad_norm_pre_clip': 20.966995239257812, 'timing/data': 0.00047116598580032587, 'timing/model': 1.1191762000089511, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 1.22}) +[2026-06-24 02:01:40,435][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.004493752494454384, 'train/grad_norm_pre_clip': 19.112302780151367, 'timing/data': 0.0003541320329532027, 'timing/model': 1.1214999429648742, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 1.22}) +[2026-06-24 02:01:44,524][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.006124714855104685, 'train/grad_norm_pre_clip': 10.077447891235352, 'timing/data': 0.0005514910444617271, 'timing/model': 1.1155998040921986, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 1.22}) +[2026-06-24 02:01:48,600][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.020214134827256203, 'train/grad_norm_pre_clip': 15.224451065063477, 'timing/data': 0.0003586839884519577, 'timing/model': 1.1147677219705656, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 1.22}) +[2026-06-24 02:01:52,681][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.01600176841020584, 'train/grad_norm_pre_clip': 10.267911911010742, 'timing/data': 0.0005916099762544036, 'timing/model': 1.1172208590433002, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 1.22}) +[2026-06-24 02:01:56,843][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.0118193244561553, 'train/grad_norm_pre_clip': 3.8537986278533936, 'timing/data': 0.000360167003236711, 'timing/model': 1.115760262007825, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 1.22}) +[2026-06-24 02:02:00,926][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.05775626003742218, 'train/grad_norm_pre_clip': 23.53236961364746, 'timing/data': 0.0004013829166069627, 'timing/model': 1.118786092964001, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 1.22}) +[2026-06-24 02:02:05,008][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.00461856834590435, 'train/grad_norm_pre_clip': 18.467851638793945, 'timing/data': 0.00035711703822016716, 'timing/model': 1.117744195042178, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 1.22}) +[2026-06-24 02:02:09,095][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.03216246888041496, 'train/grad_norm_pre_clip': 14.314626693725586, 'timing/data': 0.0004801750183105469, 'timing/model': 1.1246144759934396, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 1.22}) +[2026-06-24 02:02:13,336][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.051981765776872635, 'train/grad_norm_pre_clip': 22.853717803955078, 'timing/data': 0.0005540229612961411, 'timing/model': 1.1286724620731547, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 1.22}) +[2026-06-24 02:02:17,433][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.0021636488381773233, 'train/grad_norm_pre_clip': 19.42749786376953, 'timing/data': 0.0003528660163283348, 'timing/model': 1.123028562986292, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 1.22}) +[2026-06-24 02:02:21,542][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.032104820013046265, 'train/grad_norm_pre_clip': 22.94339942932129, 'timing/data': 0.0005956549430266023, 'timing/model': 1.1309726309264079, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 1.22}) +[2026-06-24 02:02:25,630][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.0809725746512413, 'train/grad_norm_pre_clip': 19.546085357666016, 'timing/data': 0.0006393290823325515, 'timing/model': 1.1134859090670943, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 1.22}) +[2026-06-24 02:02:29,806][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.00031405503978021443, 'train/grad_norm_pre_clip': 24.955724716186523, 'timing/data': 0.0003706510178744793, 'timing/model': 1.115369000006467, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 1.22}) +[2026-06-24 02:02:33,907][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.10007806122303009, 'train/grad_norm_pre_clip': 13.503235816955566, 'timing/data': 0.0004011050332337618, 'timing/model': 1.119288417045027, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 1.22}) +[2026-06-24 02:02:37,981][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.011373051442205906, 'train/grad_norm_pre_clip': 16.512298583984375, 'timing/data': 0.00040845200419425964, 'timing/model': 1.115827179979533, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 1.22}) +[2026-06-24 02:02:42,064][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.009801537729799747, 'train/grad_norm_pre_clip': 9.601031303405762, 'timing/data': 0.0003562140045687556, 'timing/model': 1.1242015529423952, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 1.22}) +[2026-06-24 02:02:46,251][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.023084312677383423, 'train/grad_norm_pre_clip': 15.8624267578125, 'timing/data': 0.00040015694685280323, 'timing/model': 1.1198257979704067, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 1.22}) +[2026-06-24 02:02:50,346][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.0573793426156044, 'train/grad_norm_pre_clip': 30.565765380859375, 'timing/data': 0.000371305039152503, 'timing/model': 1.1218678010627627, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 1.23}) +[2026-06-24 02:02:54,439][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.05027008056640625, 'train/grad_norm_pre_clip': 14.758615493774414, 'timing/data': 0.00039947300683707, 'timing/model': 1.118823018972762, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 1.23}) +[2026-06-24 02:02:58,541][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.028257640078663826, 'train/grad_norm_pre_clip': 20.839576721191406, 'timing/data': 0.0003755060024559498, 'timing/model': 1.1338614200940356, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 1.23}) +[2026-06-24 02:03:02,802][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.01714320480823517, 'train/grad_norm_pre_clip': 15.636146545410156, 'timing/data': 0.00036425201687961817, 'timing/model': 1.1185318910283968, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 1.23}) +[2026-06-24 02:03:06,893][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.019219592213630676, 'train/grad_norm_pre_clip': 14.25930118560791, 'timing/data': 0.0003561991034075618, 'timing/model': 1.1187840850325301, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 1.23}) +[2026-06-24 02:03:10,976][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.01383969560265541, 'train/grad_norm_pre_clip': 8.339981079101562, 'timing/data': 0.00034847098868340254, 'timing/model': 1.1171955070458353, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 1.23}) +[2026-06-24 02:03:15,080][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.03964829072356224, 'train/grad_norm_pre_clip': 7.4650163650512695, 'timing/data': 0.0005842960672453046, 'timing/model': 1.1167046630289406, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 1.23}) +[2026-06-24 02:03:19,280][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.14961515367031097, 'train/grad_norm_pre_clip': 15.995278358459473, 'timing/data': 0.00035564706195145845, 'timing/model': 1.1172667269129306, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 1.23}) +[2026-06-24 02:03:23,361][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.027484972029924393, 'train/grad_norm_pre_clip': 19.5795955657959, 'timing/data': 0.0003741700202226639, 'timing/model': 1.1196865239180624, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 1.23}) +[2026-06-24 02:03:27,444][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.09029766917228699, 'train/grad_norm_pre_clip': 31.248491287231445, 'timing/data': 0.0004155669594183564, 'timing/model': 1.1218069200403988, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 1.23}) +[2026-06-24 02:03:31,521][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.0006396415992639959, 'train/grad_norm_pre_clip': 22.534147262573242, 'timing/data': 0.0005350840510800481, 'timing/model': 1.1200611350359395, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 1.23}) +[2026-06-24 02:03:35,788][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 6.726299034198746e-05, 'train/grad_norm_pre_clip': 20.366111755371094, 'timing/data': 0.0004073890158906579, 'timing/model': 1.116003519971855, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 1.23}) +[2026-06-24 02:03:39,864][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.0417511872947216, 'train/grad_norm_pre_clip': 22.937965393066406, 'timing/data': 0.00044746696949005127, 'timing/model': 1.1203435879433528, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 1.23}) +[2026-06-24 02:03:43,937][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.0012222526129335165, 'train/grad_norm_pre_clip': 13.221565246582031, 'timing/data': 0.00037699402309954166, 'timing/model': 1.11787945299875, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 1.23}) +[2026-06-24 02:03:48,019][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.0012992129195481539, 'train/grad_norm_pre_clip': 7.216323375701904, 'timing/data': 0.0003555630100890994, 'timing/model': 1.1219672110164538, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 1.23}) +[2026-06-24 02:03:52,159][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.022993436083197594, 'train/grad_norm_pre_clip': 18.26171112060547, 'timing/data': 0.00035533797927200794, 'timing/model': 1.118963337969035, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 1.23}) +[2026-06-24 02:03:56,251][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.004133734852075577, 'train/grad_norm_pre_clip': 6.869687557220459, 'timing/data': 0.0007236009696498513, 'timing/model': 1.1185469620395452, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 1.23}) +[2026-06-24 02:04:00,329][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.0037641809321939945, 'train/grad_norm_pre_clip': 11.650113105773926, 'timing/data': 0.00037463405169546604, 'timing/model': 1.1165018840692937, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 1.23}) +[2026-06-24 02:04:04,419][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.3937707841396332, 'train/grad_norm_pre_clip': 22.62809181213379, 'timing/data': 0.00039021496195346117, 'timing/model': 1.1224662269232795, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 1.23}) +[2026-06-24 02:04:08,577][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.0014398865168914199, 'train/grad_norm_pre_clip': 20.712705612182617, 'timing/data': 0.0004078859928995371, 'timing/model': 1.1158174698939547, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 1.23}) +[2026-06-24 02:04:12,665][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.08378437161445618, 'train/grad_norm_pre_clip': 35.176692962646484, 'timing/data': 0.0003685530973598361, 'timing/model': 1.1156552790198475, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 1.23}) +[2026-06-24 02:04:16,787][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.028121313080191612, 'train/grad_norm_pre_clip': 17.987131118774414, 'timing/data': 0.0005698049208149314, 'timing/model': 1.1337238459382206, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.23}) +[2026-06-24 02:04:20,854][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.03998304158449173, 'train/grad_norm_pre_clip': 26.307750701904297, 'timing/data': 0.0003576310118660331, 'timing/model': 1.118428822956048, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 1.23}) +[2026-06-24 02:04:25,083][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.021942047402262688, 'train/grad_norm_pre_clip': 17.620100021362305, 'timing/data': 0.00039950001519173384, 'timing/model': 1.1163476549554616, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 1.23}) +[2026-06-24 02:04:29,165][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.0011176875559613109, 'train/grad_norm_pre_clip': 26.950908660888672, 'timing/data': 0.0003651160513982177, 'timing/model': 1.1216540220193565, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 1.23}) +[2026-06-24 02:04:33,242][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.0010915887542068958, 'train/grad_norm_pre_clip': 20.139541625976562, 'timing/data': 0.0003543250495567918, 'timing/model': 1.1161291030002758, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 1.23}) +[2026-06-24 02:04:37,408][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.09075330942869186, 'train/grad_norm_pre_clip': 13.637796401977539, 'timing/data': 0.0006296869833022356, 'timing/model': 1.2080458489945158, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 1.24}) +[2026-06-24 02:04:41,499][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.0037475423887372017, 'train/grad_norm_pre_clip': 17.824642181396484, 'timing/data': 0.00035769899841398, 'timing/model': 1.1225436719832942, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 1.24}) +[2026-06-24 02:04:45,569][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.04745526984333992, 'train/grad_norm_pre_clip': 25.794519424438477, 'timing/data': 0.0005268839886412024, 'timing/model': 1.114031589939259, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 1.24}) +[2026-06-24 02:04:49,671][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.03455321118235588, 'train/grad_norm_pre_clip': 6.350103378295898, 'timing/data': 0.0003773830831050873, 'timing/model': 1.1141251529334113, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 1.24}) +[2026-06-24 02:04:53,934][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.006673868279904127, 'train/grad_norm_pre_clip': 11.536399841308594, 'timing/data': 0.0003783690044656396, 'timing/model': 1.2935059210285544, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 1.24}) +[2026-06-24 02:04:58,018][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.01521935872733593, 'train/grad_norm_pre_clip': 7.504316329956055, 'timing/data': 0.00039855693466961384, 'timing/model': 1.1197918670950457, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 1.24}) +[2026-06-24 02:05:02,102][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.036671239882707596, 'train/grad_norm_pre_clip': 9.946516036987305, 'timing/data': 0.00035596999805420637, 'timing/model': 1.1187445049872622, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 1.24}) +[2026-06-24 02:05:06,178][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.006181579548865557, 'train/grad_norm_pre_clip': 18.37105369567871, 'timing/data': 0.00036886497400701046, 'timing/model': 1.1145017250673845, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 1.24}) +[2026-06-24 02:05:10,336][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3200 +[2026-06-24 02:06:15,423][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3200 +[2026-06-24 02:06:15,424][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3200 +[2026-06-24 02:06:54,259][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3200 +[2026-06-24 02:06:54,276][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.06689499318599701, 'train/grad_norm_pre_clip': 16.619707107543945, 'eval/action_loss/seconds': 65.08453193900641, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.030553536489605904, 'eval/latency_2/loss': 0.030553536489605904, 'eval/flappy/loss': 0.030553536489605904, 'eval/flappy/latency_2/loss': 0.030553536489605904, 'eval/flap_precision': 0.965194109772423, 'eval/flap_recall': 0.9782903663500678, 'eval/flap_f1': 0.9716981132075471, 'eval/flap_rate_model': 0.2075, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.971698113207547, 'eval/latency_2/flap_precision': 0.965194109772423, 'eval/latency_2/flap_recall': 0.9782903663500678, 'eval/latency_2/flap_f1': 0.9716981132075471, 'eval/latency_2/flap_rate_model': 0.2075, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.971698113207547, 'eval/action_classification/data_wait_seconds': 0.4467058777809143, 'eval/action_classification/predict_seconds': 38.22898864746094, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00041091104503721, 'timing/model': 1.1198669919976965, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 1.24}) +[2026-06-24 02:06:58,314][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.02331998385488987, 'train/grad_norm_pre_clip': 31.212299346923828, 'timing/data': 0.00045854097697883844, 'timing/model': 1.107523337006569, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 1.24}) +[2026-06-24 02:07:02,367][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.20559218525886536, 'train/grad_norm_pre_clip': 20.367198944091797, 'timing/data': 0.00035763904452323914, 'timing/model': 1.112460667034611, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 1.24}) +[2026-06-24 02:07:06,591][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.007524379063397646, 'train/grad_norm_pre_clip': 15.504192352294922, 'timing/data': 0.0003675789339467883, 'timing/model': 1.285333703039214, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 1.24}) +[2026-06-24 02:07:10,644][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.0003190033312421292, 'train/grad_norm_pre_clip': 17.143220901489258, 'timing/data': 0.0005583500023931265, 'timing/model': 1.1114394998876378, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 1.24}) +[2026-06-24 02:07:14,689][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.10041419416666031, 'train/grad_norm_pre_clip': 23.9158935546875, 'timing/data': 0.00046504195779561996, 'timing/model': 1.106243339017965, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 1.24}) +[2026-06-24 02:07:18,750][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.0038147650193423033, 'train/grad_norm_pre_clip': 15.83444881439209, 'timing/data': 0.0003601911012083292, 'timing/model': 1.1140438580187038, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 1.24}) +[2026-06-24 02:07:22,869][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.06688933819532394, 'train/grad_norm_pre_clip': 17.080495834350586, 'timing/data': 0.0003498179139569402, 'timing/model': 1.110198040958494, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 1.24}) +[2026-06-24 02:07:26,925][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.011987916193902493, 'train/grad_norm_pre_clip': 6.856788158416748, 'timing/data': 0.0003647500416263938, 'timing/model': 1.1082863690098748, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 1.24}) +[2026-06-24 02:07:30,999][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.005265859887003899, 'train/grad_norm_pre_clip': 10.670145034790039, 'timing/data': 0.0004772869870066643, 'timing/model': 1.117677356931381, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 1.24}) +[2026-06-24 02:07:35,069][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.018780101090669632, 'train/grad_norm_pre_clip': 31.077478408813477, 'timing/data': 0.00036265398375689983, 'timing/model': 1.110517928027548, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 1.24}) +[2026-06-24 02:07:39,280][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.04519215226173401, 'train/grad_norm_pre_clip': 12.151246070861816, 'timing/data': 0.0005482200067490339, 'timing/model': 1.1103789450135082, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 1.24}) +[2026-06-24 02:07:43,348][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.009854061529040337, 'train/grad_norm_pre_clip': 17.276622772216797, 'timing/data': 0.0003607909893617034, 'timing/model': 1.11506995302625, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 1.24}) +[2026-06-24 02:07:47,403][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.04522336274385452, 'train/grad_norm_pre_clip': 17.247474670410156, 'timing/data': 0.000585959991440177, 'timing/model': 1.1150129629531875, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 1.24}) +[2026-06-24 02:07:51,468][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.056545041501522064, 'train/grad_norm_pre_clip': 23.09305191040039, 'timing/data': 0.00036070693749934435, 'timing/model': 1.11250382498838, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 1.24}) +[2026-06-24 02:07:55,623][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.09270945936441422, 'train/grad_norm_pre_clip': 10.282766342163086, 'timing/data': 0.0004980569938197732, 'timing/model': 1.1225450639612973, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 1.24}) +[2026-06-24 02:07:59,703][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.01123792864382267, 'train/grad_norm_pre_clip': 6.475396156311035, 'timing/data': 0.00040890194941312075, 'timing/model': 1.1129559089895338, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 1.24}) +[2026-06-24 02:08:03,770][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.017703860998153687, 'train/grad_norm_pre_clip': 17.160470962524414, 'timing/data': 0.0003558810567483306, 'timing/model': 1.1154103860026225, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 1.25}) +[2026-06-24 02:08:07,854][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.03436347097158432, 'train/grad_norm_pre_clip': 18.0446720123291, 'timing/data': 0.0003517180448397994, 'timing/model': 1.1157872669864446, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 1.25}) +[2026-06-24 02:08:12,045][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.020030826330184937, 'train/grad_norm_pre_clip': 7.45188045501709, 'timing/data': 0.0003681799862533808, 'timing/model': 1.2360813280101866, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 1.25}) +[2026-06-24 02:08:16,132][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.005257372744381428, 'train/grad_norm_pre_clip': 5.495614051818848, 'timing/data': 0.0004966580308973789, 'timing/model': 1.1254512260202318, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 1.25}) +[2026-06-24 02:08:20,220][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.0013507112162187696, 'train/grad_norm_pre_clip': 4.103696346282959, 'timing/data': 0.0003583320649340749, 'timing/model': 1.1219267819542438, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 1.25}) +[2026-06-24 02:08:24,298][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.004148044623434544, 'train/grad_norm_pre_clip': 10.67673110961914, 'timing/data': 0.00036012509372085333, 'timing/model': 1.1162479940103367, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 1.25}) +[2026-06-24 02:08:28,374][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.0045577725395560265, 'train/grad_norm_pre_clip': 7.677354335784912, 'timing/data': 0.0005775300087407231, 'timing/model': 1.1166139770066366, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 1.25}) +[2026-06-24 02:08:32,612][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.24352604150772095, 'train/grad_norm_pre_clip': 25.157411575317383, 'timing/data': 0.0004039650084450841, 'timing/model': 1.1189395260298625, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 1.25}) +[2026-06-24 02:08:36,695][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.005894571542739868, 'train/grad_norm_pre_clip': 16.432106018066406, 'timing/data': 0.00039692199788987637, 'timing/model': 1.1153383300406858, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 1.25}) +[2026-06-24 02:08:40,773][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.013896476477384567, 'train/grad_norm_pre_clip': 19.015914916992188, 'timing/data': 0.0004865429364144802, 'timing/model': 1.116900488967076, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 1.25}) +[2026-06-24 02:08:44,844][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.02989186905324459, 'train/grad_norm_pre_clip': 13.244441986083984, 'timing/data': 0.0005625690100714564, 'timing/model': 1.113025028957054, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 1.25}) +[2026-06-24 02:08:49,003][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.007848210632801056, 'train/grad_norm_pre_clip': 28.215681076049805, 'timing/data': 0.0003749079769477248, 'timing/model': 1.1197237949818373, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 1.25}) +[2026-06-24 02:08:53,075][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.07069490104913712, 'train/grad_norm_pre_clip': 36.76563262939453, 'timing/data': 0.00035673403181135654, 'timing/model': 1.114625301095657, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 1.25}) +[2026-06-24 02:08:57,169][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.003306196304038167, 'train/grad_norm_pre_clip': 14.366414070129395, 'timing/data': 0.0006648990092799067, 'timing/model': 1.1178546420997009, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 1.25}) +[2026-06-24 02:09:01,259][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.0042955465614795685, 'train/grad_norm_pre_clip': 8.709491729736328, 'timing/data': 0.00036393990740180016, 'timing/model': 1.1236987609881908, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 1.25}) +[2026-06-24 02:09:05,448][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.017420150339603424, 'train/grad_norm_pre_clip': 27.830209732055664, 'timing/data': 0.0003584359074011445, 'timing/model': 1.1171381489839405, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 1.25}) +[2026-06-24 02:09:09,534][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.1498204916715622, 'train/grad_norm_pre_clip': 24.642736434936523, 'timing/data': 0.0005618380382657051, 'timing/model': 1.1179606360383332, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 1.25}) +[2026-06-24 02:09:13,618][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.1552269458770752, 'train/grad_norm_pre_clip': 26.657968521118164, 'timing/data': 0.0003598540788516402, 'timing/model': 1.1239541930845007, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 1.25}) +[2026-06-24 02:09:17,734][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.06368956714868546, 'train/grad_norm_pre_clip': 20.904621124267578, 'timing/data': 0.00036658591125160456, 'timing/model': 1.1217515070457011, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 1.25}) +[2026-06-24 02:09:21,895][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.02846999280154705, 'train/grad_norm_pre_clip': 21.719287872314453, 'timing/data': 0.0005276940064504743, 'timing/model': 1.1217757479753345, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.25}) +[2026-06-24 02:09:25,978][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.166142076253891, 'train/grad_norm_pre_clip': 45.58378219604492, 'timing/data': 0.0005061270203441381, 'timing/model': 1.1188743370585144, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 1.25}) +[2026-06-24 02:09:30,074][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.0012012553634122014, 'train/grad_norm_pre_clip': 19.953609466552734, 'timing/data': 0.0004035669844597578, 'timing/model': 1.1206125750904903, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 1.25}) +[2026-06-24 02:09:34,167][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.010309201665222645, 'train/grad_norm_pre_clip': 19.13909912109375, 'timing/data': 0.0004029860720038414, 'timing/model': 1.1220183100085706, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 1.25}) +[2026-06-24 02:09:38,405][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.18910685181617737, 'train/grad_norm_pre_clip': 25.934528350830078, 'timing/data': 0.0005624250043183565, 'timing/model': 1.133866035961546, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 1.25}) +[2026-06-24 02:09:42,480][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.0024794875644147396, 'train/grad_norm_pre_clip': 25.69425392150879, 'timing/data': 0.0004971289308741689, 'timing/model': 1.1103866500779986, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 1.25}) +[2026-06-24 02:09:46,576][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.026971867308020592, 'train/grad_norm_pre_clip': 24.068553924560547, 'timing/data': 0.00035775196738541126, 'timing/model': 1.123866681009531, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 1.25}) +[2026-06-24 02:09:50,738][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.0038736993446946144, 'train/grad_norm_pre_clip': 31.565885543823242, 'timing/data': 0.0003641510847955942, 'timing/model': 1.205448373919353, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 1.26}) +[2026-06-24 02:09:54,821][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.06275510042905807, 'train/grad_norm_pre_clip': 15.714081764221191, 'timing/data': 0.00044119206722825766, 'timing/model': 1.1147053319728002, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 1.26}) +[2026-06-24 02:09:58,907][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.06890511512756348, 'train/grad_norm_pre_clip': 32.43186569213867, 'timing/data': 0.000356322037987411, 'timing/model': 1.117611502064392, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 1.26}) +[2026-06-24 02:10:02,986][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.006408484652638435, 'train/grad_norm_pre_clip': 5.450077056884766, 'timing/data': 0.0005195289850234985, 'timing/model': 1.113852035952732, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 1.26}) +[2026-06-24 02:10:07,245][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.024420777335762978, 'train/grad_norm_pre_clip': 29.477176666259766, 'timing/data': 0.0003776899538934231, 'timing/model': 1.2949015470221639, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 1.26}) +[2026-06-24 02:10:11,333][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.02100967988371849, 'train/grad_norm_pre_clip': 9.63676643371582, 'timing/data': 0.0004781220341101289, 'timing/model': 1.1236904809484258, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 1.26}) +[2026-06-24 02:10:15,420][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.009770544245839119, 'train/grad_norm_pre_clip': 31.399572372436523, 'timing/data': 0.00040197104681283236, 'timing/model': 1.1222835869994015, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 1.26}) +[2026-06-24 02:20:12,295][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3250 +[2026-06-24 02:20:12,295][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.09064384549856186, 'train/grad_norm_pre_clip': 11.308013916015625, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 365.76000000002057, 'rl_games_eval/mid_train/mean_length': 2965.0, 'rl_games_eval/mid_train/std_reward': 157.6800000000101, 'rl_games_eval/mid_train/std_length': 1270.0, 'rl_games_eval/mid_train/macro_mean_reward': 365.76000000002057, 'rl_games_eval/mid_train/macro_mean_length': 2965.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 365.76000000002057, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2965.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 157.6800000000101, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1270.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 365.76000000002057, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000561184948310256, 'timing/model': 1.1188317240448669, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 1.26}) +[2026-06-24 02:20:16,267][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.020629609003663063, 'train/grad_norm_pre_clip': 32.96615219116211, 'timing/data': 0.0004044430097565055, 'timing/model': 1.0924699789611623, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 1.26}) +[2026-06-24 02:20:20,381][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.0005896895308978856, 'train/grad_norm_pre_clip': 16.51595115661621, 'timing/data': 0.0004884370137006044, 'timing/model': 1.0936993919312954, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 1.26}) +[2026-06-24 02:20:24,395][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.0019835135899484158, 'train/grad_norm_pre_clip': 12.559195518493652, 'timing/data': 0.0003847250482067466, 'timing/model': 1.1031411779113114, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 1.26}) +[2026-06-24 02:20:28,416][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.08070699125528336, 'train/grad_norm_pre_clip': 10.658939361572266, 'timing/data': 0.0005342199001461267, 'timing/model': 1.0981666300212964, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 1.26}) +[2026-06-24 02:20:32,449][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.033382806926965714, 'train/grad_norm_pre_clip': 29.51443862915039, 'timing/data': 0.00035286496859043837, 'timing/model': 1.0980887720361352, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 1.26}) +[2026-06-24 02:20:36,539][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.027814799919724464, 'train/grad_norm_pre_clip': 21.797693252563477, 'timing/data': 0.0003604890080168843, 'timing/model': 1.1017343769781291, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 1.26}) +[2026-06-24 02:20:40,577][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.0875503420829773, 'train/grad_norm_pre_clip': 34.001651763916016, 'timing/data': 0.0003506889333948493, 'timing/model': 1.1063329599564895, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 1.26}) +[2026-06-24 02:20:44,602][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.038024529814720154, 'train/grad_norm_pre_clip': 17.842763900756836, 'timing/data': 0.00040507095400244, 'timing/model': 1.1065812199376523, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 1.26}) +[2026-06-24 02:20:48,622][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.08050024509429932, 'train/grad_norm_pre_clip': 13.59378719329834, 'timing/data': 0.00035180896520614624, 'timing/model': 1.1052746450295672, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 1.26}) +[2026-06-24 02:20:52,777][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.030101431533694267, 'train/grad_norm_pre_clip': 14.99311351776123, 'timing/data': 0.00035813392605632544, 'timing/model': 1.2214067300083116, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 1.26}) +[2026-06-24 02:20:56,833][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.07829506695270538, 'train/grad_norm_pre_clip': 38.08403778076172, 'timing/data': 0.0003701320383697748, 'timing/model': 1.104348479071632, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 1.26}) +[2026-06-24 02:21:00,882][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.10594046115875244, 'train/grad_norm_pre_clip': 23.71065330505371, 'timing/data': 0.0005680889589712024, 'timing/model': 1.1252891439944506, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 1.26}) +[2026-06-24 02:21:04,941][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.009007786400616169, 'train/grad_norm_pre_clip': 14.276803970336914, 'timing/data': 0.0003601150820031762, 'timing/model': 1.1079872470581904, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 1.26}) +[2026-06-24 02:21:09,102][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.0011549799237400293, 'train/grad_norm_pre_clip': 20.726295471191406, 'timing/data': 0.0005570299690589309, 'timing/model': 1.1093678050674498, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 1.26}) +[2026-06-24 02:21:13,154][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.004122868180274963, 'train/grad_norm_pre_clip': 26.52764129638672, 'timing/data': 0.0003556140000000596, 'timing/model': 1.1111210450762883, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 1.26}) +[2026-06-24 02:21:17,228][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.00834241509437561, 'train/grad_norm_pre_clip': 16.87564468383789, 'timing/data': 0.0003595859743654728, 'timing/model': 1.1143969330005348, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 1.26}) +[2026-06-24 02:21:21,283][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.0348932221531868, 'train/grad_norm_pre_clip': 25.879636764526367, 'timing/data': 0.0003593889996409416, 'timing/model': 1.1062819900689647, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 1.26}) +[2026-06-24 02:21:25,516][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.11132899671792984, 'train/grad_norm_pre_clip': 39.04790496826172, 'timing/data': 0.000565857975743711, 'timing/model': 1.114256508066319, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 1.26}) +[2026-06-24 02:21:29,614][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.12119904160499573, 'train/grad_norm_pre_clip': 18.356197357177734, 'timing/data': 0.0003523359773680568, 'timing/model': 1.1221930850297213, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 1.27}) +[2026-06-24 02:21:33,698][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.006628119386732578, 'train/grad_norm_pre_clip': 17.304285049438477, 'timing/data': 0.0003565539373084903, 'timing/model': 1.130635587964207, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 1.27}) +[2026-06-24 02:21:37,765][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.030222579836845398, 'train/grad_norm_pre_clip': 33.276004791259766, 'timing/data': 0.00039676798041909933, 'timing/model': 1.1073494569864124, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 1.27}) +[2026-06-24 02:21:42,005][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.010238008573651314, 'train/grad_norm_pre_clip': 34.23520278930664, 'timing/data': 0.0004987969296053052, 'timing/model': 1.1118428960908204, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 1.27}) +[2026-06-24 02:21:46,088][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.0005419043591246009, 'train/grad_norm_pre_clip': 28.88156509399414, 'timing/data': 0.0005996699910610914, 'timing/model': 1.1138300050515682, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 1.27}) +[2026-06-24 02:21:50,147][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.00244584446772933, 'train/grad_norm_pre_clip': 27.628206253051758, 'timing/data': 0.000408330000936985, 'timing/model': 1.1136377870570868, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 1.27}) +[2026-06-24 02:21:54,215][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.04717981442809105, 'train/grad_norm_pre_clip': 23.105812072753906, 'timing/data': 0.0003975389990955591, 'timing/model': 1.113921336014755, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 1.27}) +[2026-06-24 02:21:58,352][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.0009434860548935831, 'train/grad_norm_pre_clip': 20.789701461791992, 'timing/data': 0.00047855405136942863, 'timing/model': 1.1134261130355299, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 1.27}) +[2026-06-24 02:22:02,408][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.11124598979949951, 'train/grad_norm_pre_clip': 28.452991485595703, 'timing/data': 0.00043894792906939983, 'timing/model': 1.1121649370761588, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 1.27}) +[2026-06-24 02:22:06,473][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.0008285068324767053, 'train/grad_norm_pre_clip': 17.768896102905273, 'timing/data': 0.0003808110486716032, 'timing/model': 1.1141518000513315, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 1.27}) +[2026-06-24 02:22:10,533][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.001281167147681117, 'train/grad_norm_pre_clip': 11.053560256958008, 'timing/data': 0.0004907470429316163, 'timing/model': 1.1079282979480922, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 1.27}) +[2026-06-24 02:22:14,670][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.005993630271404982, 'train/grad_norm_pre_clip': 17.185548782348633, 'timing/data': 0.0003570739645510912, 'timing/model': 1.1144642310682684, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 1.27}) +[2026-06-24 02:22:18,744][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.05819282308220863, 'train/grad_norm_pre_clip': 10.070423126220703, 'timing/data': 0.0004161949036642909, 'timing/model': 1.115934869972989, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 1.27}) +[2026-06-24 02:22:22,823][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.012679395265877247, 'train/grad_norm_pre_clip': 10.624462127685547, 'timing/data': 0.00036394200287759304, 'timing/model': 1.1181157040409744, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 1.27}) +[2026-06-24 02:22:26,895][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.0508028082549572, 'train/grad_norm_pre_clip': 21.341522216796875, 'timing/data': 0.0003685419214889407, 'timing/model': 1.1095397099852562, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 1.27}) +[2026-06-24 02:22:31,106][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.0035328499507158995, 'train/grad_norm_pre_clip': 28.0221004486084, 'timing/data': 0.0004964809631928802, 'timing/model': 1.1187427039258182, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 1.27}) +[2026-06-24 02:22:35,171][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.012445801869034767, 'train/grad_norm_pre_clip': 18.871034622192383, 'timing/data': 0.00047877593897283077, 'timing/model': 1.115204727044329, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.27}) +[2026-06-24 02:22:39,244][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.03435041755437851, 'train/grad_norm_pre_clip': 12.42392349243164, 'timing/data': 0.00040772499050945044, 'timing/model': 1.1172023329418153, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 1.27}) +[2026-06-24 02:22:43,419][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.023888682946562767, 'train/grad_norm_pre_clip': 14.187965393066406, 'timing/data': 0.000373012968339026, 'timing/model': 1.2117381570860744, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 1.27}) +[2026-06-24 02:22:47,486][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.023140553385019302, 'train/grad_norm_pre_clip': 31.730241775512695, 'timing/data': 0.00035870401188731194, 'timing/model': 1.1168531531002373, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 1.27}) +[2026-06-24 02:22:51,566][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.0009193063015118241, 'train/grad_norm_pre_clip': 10.907503128051758, 'timing/data': 0.0005651829997077584, 'timing/model': 1.1181631179060787, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 1.27}) +[2026-06-24 02:22:55,648][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.007756971288472414, 'train/grad_norm_pre_clip': 10.618739128112793, 'timing/data': 0.0005700249457731843, 'timing/model': 1.118601203081198, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 1.27}) +[2026-06-24 02:22:59,924][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.019645771011710167, 'train/grad_norm_pre_clip': 11.151589393615723, 'timing/data': 0.00036879698745906353, 'timing/model': 1.2939141530077904, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 1.27}) +[2026-06-24 02:23:03,992][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.02465701662003994, 'train/grad_norm_pre_clip': 11.256892204284668, 'timing/data': 0.00037228199653327465, 'timing/model': 1.1143919919850305, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 1.27}) +[2026-06-24 02:23:08,072][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.07514060288667679, 'train/grad_norm_pre_clip': 19.927202224731445, 'timing/data': 0.0004109189612790942, 'timing/model': 1.1163481129333377, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 1.27}) +[2026-06-24 02:23:12,147][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.056547217071056366, 'train/grad_norm_pre_clip': 36.29474639892578, 'timing/data': 0.0006302099209278822, 'timing/model': 1.1148534240201116, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 1.27}) +[2026-06-24 02:23:16,305][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.029993928968906403, 'train/grad_norm_pre_clip': 32.83795928955078, 'timing/data': 0.00035605800803750753, 'timing/model': 1.1309303890448064, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 1.28}) +[2026-06-24 02:23:20,384][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.012671238742768764, 'train/grad_norm_pre_clip': 14.417695045471191, 'timing/data': 0.0005570990033447742, 'timing/model': 1.118037200998515, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 1.28}) +[2026-06-24 02:23:24,461][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.04140252247452736, 'train/grad_norm_pre_clip': 13.298052787780762, 'timing/data': 0.00035609689075499773, 'timing/model': 1.1229418340371922, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 1.28}) +[2026-06-24 02:23:28,548][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.002313197124749422, 'train/grad_norm_pre_clip': 9.720593452453613, 'timing/data': 0.00036802003160119057, 'timing/model': 1.1195851219817996, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 1.28}) +[2026-06-24 02:23:32,721][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.013454819098114967, 'train/grad_norm_pre_clip': 21.28371238708496, 'timing/data': 0.0003696880303323269, 'timing/model': 1.1271852349163964, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 1.28}) +[2026-06-24 02:23:36,798][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3300 +[2026-06-24 02:24:40,627][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3300 +[2026-06-24 02:24:40,629][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3300 +[2026-06-24 02:25:19,641][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3300 +[2026-06-24 02:25:19,642][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.05948204547166824, 'train/grad_norm_pre_clip': 24.638404846191406, 'eval/action_loss/seconds': 63.82633839105256, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.0331478975713253, 'eval/latency_2/loss': 0.0331478975713253, 'eval/flappy/loss': 0.0331478975713253, 'eval/flappy/latency_2/loss': 0.0331478975713253, 'eval/flap_precision': 0.9847009735744089, 'eval/flap_recall': 0.9606512890094979, 'eval/flap_f1': 0.9725274725274725, 'eval/flap_rate_model': 0.19972222222222222, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9725274725274725, 'eval/latency_2/flap_precision': 0.9847009735744089, 'eval/latency_2/flap_recall': 0.9606512890094979, 'eval/latency_2/flap_f1': 0.9725274725274725, 'eval/latency_2/flap_rate_model': 0.19972222222222222, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9725274725274725, 'eval/action_classification/data_wait_seconds': 0.6374927163124084, 'eval/action_classification/predict_seconds': 38.2050666809082, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0005941279232501984, 'timing/model': 1.1190377159509808, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 1.28}) +[2026-06-24 02:25:23,687][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.1673911064863205, 'train/grad_norm_pre_clip': 19.368053436279297, 'timing/data': 0.0003623568918555975, 'timing/model': 1.1063439489807934, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 1.28}) +[2026-06-24 02:25:27,854][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.10264437645673752, 'train/grad_norm_pre_clip': 29.736186981201172, 'timing/data': 0.0003629239508882165, 'timing/model': 1.2366978370118886, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 1.28}) +[2026-06-24 02:25:31,900][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.09222685545682907, 'train/grad_norm_pre_clip': 15.121485710144043, 'timing/data': 0.0005550700007006526, 'timing/model': 1.1053279590560123, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 1.28}) +[2026-06-24 02:25:35,963][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.07452231645584106, 'train/grad_norm_pre_clip': 24.28680992126465, 'timing/data': 0.0005634230328723788, 'timing/model': 1.1159916369942948, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 1.28}) +[2026-06-24 02:25:40,021][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.03478757664561272, 'train/grad_norm_pre_clip': 17.558652877807617, 'timing/data': 0.0005487750750035048, 'timing/model': 1.1159686640603468, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 1.28}) +[2026-06-24 02:25:44,210][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.014282229356467724, 'train/grad_norm_pre_clip': 14.297669410705566, 'timing/data': 0.0003579379990696907, 'timing/model': 1.2434942489489913, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 1.28}) +[2026-06-24 02:25:48,271][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.015269383788108826, 'train/grad_norm_pre_clip': 14.119903564453125, 'timing/data': 0.0004158490337431431, 'timing/model': 1.1198525229701772, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 1.28}) +[2026-06-24 02:25:52,338][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.001714805606752634, 'train/grad_norm_pre_clip': 6.087702751159668, 'timing/data': 0.0004124509869143367, 'timing/model': 1.1195447510108352, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 1.28}) +[2026-06-24 02:25:56,399][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.03531108796596527, 'train/grad_norm_pre_clip': 11.422626495361328, 'timing/data': 0.0004416810115799308, 'timing/model': 1.1101462839869782, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 1.28}) +[2026-06-24 02:26:00,610][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.10609576851129532, 'train/grad_norm_pre_clip': 25.66771125793457, 'timing/data': 0.0003566330997273326, 'timing/model': 1.2593989949673414, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 1.28}) +[2026-06-24 02:26:04,673][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.03563106805086136, 'train/grad_norm_pre_clip': 47.387516021728516, 'timing/data': 0.0003580650081858039, 'timing/model': 1.115350695909001, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 1.28}) +[2026-06-24 02:26:08,729][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.054276008158922195, 'train/grad_norm_pre_clip': 32.63076400756836, 'timing/data': 0.00034941197372972965, 'timing/model': 1.1120908489683643, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 1.28}) +[2026-06-24 02:26:12,793][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.012621487490832806, 'train/grad_norm_pre_clip': 23.707439422607422, 'timing/data': 0.00047343107871711254, 'timing/model': 1.115832061972469, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 1.28}) +[2026-06-24 02:26:16,979][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.017574220895767212, 'train/grad_norm_pre_clip': 18.11484718322754, 'timing/data': 0.0003619130002334714, 'timing/model': 1.2416222960455343, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 1.28}) +[2026-06-24 02:26:21,037][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.06503525376319885, 'train/grad_norm_pre_clip': 74.43539428710938, 'timing/data': 0.00035121990367770195, 'timing/model': 1.11488876293879, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 1.28}) +[2026-06-24 02:26:25,125][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.13916319608688354, 'train/grad_norm_pre_clip': 33.29709243774414, 'timing/data': 0.0006358670070767403, 'timing/model': 1.1139271699357778, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 1.28}) +[2026-06-24 02:26:29,208][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.009188632480800152, 'train/grad_norm_pre_clip': 15.557242393493652, 'timing/data': 0.00035582808777689934, 'timing/model': 1.1159401630284265, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 1.28}) +[2026-06-24 02:26:33,440][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.03251153975725174, 'train/grad_norm_pre_clip': 40.25558090209961, 'timing/data': 0.00041331793181598186, 'timing/model': 1.2611206519650295, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 1.28}) +[2026-06-24 02:26:37,520][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.15687081217765808, 'train/grad_norm_pre_clip': 37.17007064819336, 'timing/data': 0.00035446800757199526, 'timing/model': 1.1108398759970441, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 1.28}) +[2026-06-24 02:26:41,608][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.07028596103191376, 'train/grad_norm_pre_clip': 43.893795013427734, 'timing/data': 0.00036221405025571585, 'timing/model': 1.1181039119837806, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 1.28}) +[2026-06-24 02:26:45,679][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.010078726336359978, 'train/grad_norm_pre_clip': 15.467705726623535, 'timing/data': 0.00038087600842118263, 'timing/model': 1.1125154150649905, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 1.29}) +[2026-06-24 02:26:49,884][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.002193884691223502, 'train/grad_norm_pre_clip': 20.047245025634766, 'timing/data': 0.000395973096601665, 'timing/model': 1.2448313389904797, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 1.29}) +[2026-06-24 02:26:53,955][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.004695496521890163, 'train/grad_norm_pre_clip': 9.020471572875977, 'timing/data': 0.00040176790207624435, 'timing/model': 1.1136575320269912, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 1.29}) +[2026-06-24 02:26:58,025][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.06423193961381912, 'train/grad_norm_pre_clip': 41.54540252685547, 'timing/data': 0.0004565659910440445, 'timing/model': 1.1115213739685714, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 1.29}) +[2026-06-24 02:27:02,098][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.03855094313621521, 'train/grad_norm_pre_clip': 10.24685287475586, 'timing/data': 0.00034607306588441133, 'timing/model': 1.1110014690784737, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 1.29}) +[2026-06-24 02:27:06,321][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.020685717463493347, 'train/grad_norm_pre_clip': 4.240241050720215, 'timing/data': 0.00035131501499563456, 'timing/model': 1.2596829250687733, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 1.29}) +[2026-06-24 02:27:10,397][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.04292809218168259, 'train/grad_norm_pre_clip': 25.467876434326172, 'timing/data': 0.000624743988737464, 'timing/model': 1.1160344990203157, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 1.29}) +[2026-06-24 02:27:14,476][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.1839873343706131, 'train/grad_norm_pre_clip': 30.24808692932129, 'timing/data': 0.00035062304232269526, 'timing/model': 1.1222115140408278, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 1.29}) +[2026-06-24 02:27:18,550][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.04483942314982414, 'train/grad_norm_pre_clip': 21.274263381958008, 'timing/data': 0.0003623079974204302, 'timing/model': 1.116257086978294, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 1.29}) +[2026-06-24 02:27:22,867][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.02135089971125126, 'train/grad_norm_pre_clip': 26.889970779418945, 'timing/data': 0.00040092202834784985, 'timing/model': 1.2410751919960603, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 1.29}) +[2026-06-24 02:27:26,936][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.028727702796459198, 'train/grad_norm_pre_clip': 6.2277421951293945, 'timing/data': 0.00040554895531386137, 'timing/model': 1.1118460670113564, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 1.29}) +[2026-06-24 02:27:31,012][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.004297580104321241, 'train/grad_norm_pre_clip': 34.63545227050781, 'timing/data': 0.0005435830680653453, 'timing/model': 1.1194207400549203, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 1.29}) +[2026-06-24 02:27:35,087][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.07843092828989029, 'train/grad_norm_pre_clip': 21.389530181884766, 'timing/data': 0.00042122998274862766, 'timing/model': 1.1170998780289665, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 1.29}) +[2026-06-24 02:27:39,312][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.020927274599671364, 'train/grad_norm_pre_clip': 22.841815948486328, 'timing/data': 0.0005436099600046873, 'timing/model': 1.257784478017129, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.29}) +[2026-06-24 02:27:43,405][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.11232834309339523, 'train/grad_norm_pre_clip': 23.02091407775879, 'timing/data': 0.00038364995270967484, 'timing/model': 1.1202774179400876, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 1.29}) +[2026-06-24 02:27:47,495][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.08233796805143356, 'train/grad_norm_pre_clip': 17.530458450317383, 'timing/data': 0.0005745250964537263, 'timing/model': 1.1162562899990007, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 1.29}) +[2026-06-24 02:27:51,563][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.008662148378789425, 'train/grad_norm_pre_clip': 10.359603881835938, 'timing/data': 0.0005480539985001087, 'timing/model': 1.1164120209868997, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 1.29}) +[2026-06-24 02:27:55,770][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.013399006798863411, 'train/grad_norm_pre_clip': 22.8638973236084, 'timing/data': 0.0004155910573899746, 'timing/model': 1.253594939946197, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 1.29}) +[2026-06-24 02:27:59,849][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.08096878230571747, 'train/grad_norm_pre_clip': 24.91901206970215, 'timing/data': 0.00041322095785290003, 'timing/model': 1.1168750469805673, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 1.29}) +[2026-06-24 02:28:03,919][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.0132685462012887, 'train/grad_norm_pre_clip': 32.53263473510742, 'timing/data': 0.000349772977642715, 'timing/model': 1.1208404540084302, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 1.29}) +[2026-06-24 02:28:08,004][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.024675510823726654, 'train/grad_norm_pre_clip': 7.372433662414551, 'timing/data': 0.0004735729889944196, 'timing/model': 1.1227128900354728, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 1.29}) +[2026-06-24 02:28:12,237][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.08525006473064423, 'train/grad_norm_pre_clip': 17.485990524291992, 'timing/data': 0.00035682006273418665, 'timing/model': 1.2657615140778944, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 1.29}) +[2026-06-24 02:28:16,343][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.1699596494436264, 'train/grad_norm_pre_clip': 20.47402000427246, 'timing/data': 0.0003587029641494155, 'timing/model': 1.1323109209770337, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 1.29}) +[2026-06-24 02:28:20,441][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.03423939272761345, 'train/grad_norm_pre_clip': 20.927295684814453, 'timing/data': 0.00040410307701677084, 'timing/model': 1.1248762720497325, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 1.29}) +[2026-06-24 02:28:24,526][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.044390641152858734, 'train/grad_norm_pre_clip': 24.35755729675293, 'timing/data': 0.0005881449906155467, 'timing/model': 1.1148871240438893, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 1.29}) +[2026-06-24 02:28:28,744][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.11143076419830322, 'train/grad_norm_pre_clip': 27.145227432250977, 'timing/data': 0.0003588469699025154, 'timing/model': 1.2376886419951916, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 1.29}) +[2026-06-24 02:28:32,821][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.03675162047147751, 'train/grad_norm_pre_clip': 18.091075897216797, 'timing/data': 0.000517055974341929, 'timing/model': 1.1138223570305854, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 1.3}) +[2026-06-24 02:28:36,908][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.005145749542862177, 'train/grad_norm_pre_clip': 11.72629165649414, 'timing/data': 0.00041036889888346195, 'timing/model': 1.1192461150931194, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 1.3}) +[2026-06-24 02:28:40,995][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.1835150271654129, 'train/grad_norm_pre_clip': 19.490541458129883, 'timing/data': 0.0003564070211723447, 'timing/model': 1.115533014992252, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 1.3}) +[2026-06-24 02:28:45,219][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.0015368497697636485, 'train/grad_norm_pre_clip': 35.05433654785156, 'timing/data': 0.00035512191243469715, 'timing/model': 1.2613458120031282, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 1.3}) +[2026-06-24 02:28:49,328][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.0023807836696505547, 'train/grad_norm_pre_clip': 20.183032989501953, 'timing/data': 0.0004080849466845393, 'timing/model': 1.1207120170583948, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 1.3}) +[2026-06-24 02:28:53,450][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.016775885596871376, 'train/grad_norm_pre_clip': 36.7075080871582, 'timing/data': 0.0003539810422807932, 'timing/model': 1.1202219849219546, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 1.3}) +[2026-06-24 02:28:57,522][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.026610329747200012, 'train/grad_norm_pre_clip': 19.929969787597656, 'timing/data': 0.00040902302134782076, 'timing/model': 1.115830936934799, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 1.3}) +[2026-06-24 02:29:01,765][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.19859084486961365, 'train/grad_norm_pre_clip': 28.972244262695312, 'timing/data': 0.0004501320654526353, 'timing/model': 1.2640259710606188, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 1.3}) +[2026-06-24 02:29:05,847][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.06792835891246796, 'train/grad_norm_pre_clip': 13.142489433288574, 'timing/data': 0.00041135295759886503, 'timing/model': 1.1190250189974904, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 1.3}) +[2026-06-24 02:29:09,937][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.21748381853103638, 'train/grad_norm_pre_clip': 32.40802001953125, 'timing/data': 0.00035800004843622446, 'timing/model': 1.1163940529804677, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 1.3}) +[2026-06-24 02:29:14,020][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.026096967980265617, 'train/grad_norm_pre_clip': 37.0827751159668, 'timing/data': 0.0004429289838299155, 'timing/model': 1.1112413330702111, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 1.3}) +[2026-06-24 02:29:18,249][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.08567474037408829, 'train/grad_norm_pre_clip': 18.919174194335938, 'timing/data': 0.0003638160414993763, 'timing/model': 1.2568679199321195, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 1.3}) +[2026-06-24 02:29:22,337][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.09392675012350082, 'train/grad_norm_pre_clip': 20.5899658203125, 'timing/data': 0.0005859379889443517, 'timing/model': 1.1211780899902806, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 1.3}) +[2026-06-24 02:29:26,425][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.01254679262638092, 'train/grad_norm_pre_clip': 12.682524681091309, 'timing/data': 0.000403471989557147, 'timing/model': 1.1244828149210662, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 1.3}) +[2026-06-24 02:29:30,516][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.028818417340517044, 'train/grad_norm_pre_clip': 14.981389999389648, 'timing/data': 0.0003633899614214897, 'timing/model': 1.1218971610069275, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 1.3}) +[2026-06-24 02:29:34,737][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.065348319709301, 'train/grad_norm_pre_clip': 35.71084213256836, 'timing/data': 0.0004109150031581521, 'timing/model': 1.257319359923713, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 1.3}) +[2026-06-24 02:29:38,829][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.011862118728458881, 'train/grad_norm_pre_clip': 16.092042922973633, 'timing/data': 0.0004080199869349599, 'timing/model': 1.1250254219630733, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 1.3}) +[2026-06-24 02:29:42,910][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.06374502927064896, 'train/grad_norm_pre_clip': 21.8360595703125, 'timing/data': 0.0003696089843288064, 'timing/model': 1.1178716870490462, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 1.3}) +[2026-06-24 02:29:46,991][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.004624345805495977, 'train/grad_norm_pre_clip': 21.53018569946289, 'timing/data': 0.0006780689582228661, 'timing/model': 1.1153397569432855, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 1.3}) +[2026-06-24 02:29:51,202][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.001122273039072752, 'train/grad_norm_pre_clip': 14.67009449005127, 'timing/data': 0.00035742693580687046, 'timing/model': 1.239835134940222, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 1.3}) +[2026-06-24 02:29:55,277][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.03065325692296028, 'train/grad_norm_pre_clip': 19.092098236083984, 'timing/data': 0.0005068769678473473, 'timing/model': 1.1147601050324738, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 1.3}) +[2026-06-24 02:29:59,500][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.010210401378571987, 'train/grad_norm_pre_clip': 23.393661499023438, 'timing/data': 0.0005021160468459129, 'timing/model': 1.1163415049668401, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 1.3}) +[2026-06-24 02:30:03,588][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.0790606141090393, 'train/grad_norm_pre_clip': 43.936805725097656, 'timing/data': 0.0003615079913288355, 'timing/model': 1.1210364300059155, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 1.3}) +[2026-06-24 02:30:07,827][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.0038558475207537413, 'train/grad_norm_pre_clip': 14.504273414611816, 'timing/data': 0.0003440899308770895, 'timing/model': 1.2774216630496085, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 1.3}) +[2026-06-24 02:30:11,917][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.0015426689060404897, 'train/grad_norm_pre_clip': 11.168440818786621, 'timing/data': 0.0005389429861679673, 'timing/model': 1.1169123819563538, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 1.3}) +[2026-06-24 02:30:15,989][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.015962813049554825, 'train/grad_norm_pre_clip': 13.61529541015625, 'timing/data': 0.0004115210613235831, 'timing/model': 1.1116931210272014, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 1.31}) +[2026-06-24 02:30:20,082][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.11002814769744873, 'train/grad_norm_pre_clip': 22.735490798950195, 'timing/data': 0.0003608199767768383, 'timing/model': 1.1186970109120011, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 1.31}) +[2026-06-24 02:30:24,305][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.06519708037376404, 'train/grad_norm_pre_clip': 25.565839767456055, 'timing/data': 0.00040463602636009455, 'timing/model': 1.252900991938077, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 1.31}) +[2026-06-24 02:30:28,394][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.08338551968336105, 'train/grad_norm_pre_clip': 35.5483283996582, 'timing/data': 0.000358884921297431, 'timing/model': 1.1223436730215326, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 1.31}) +[2026-06-24 02:30:32,479][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.027347125113010406, 'train/grad_norm_pre_clip': 9.835932731628418, 'timing/data': 0.0005394140025600791, 'timing/model': 1.116241839947179, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 1.31}) +[2026-06-24 02:30:36,566][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.05868487060070038, 'train/grad_norm_pre_clip': 14.930007934570312, 'timing/data': 0.0003593149594962597, 'timing/model': 1.119924520025961, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 1.31}) +[2026-06-24 02:30:40,779][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.022908225655555725, 'train/grad_norm_pre_clip': 19.404603958129883, 'timing/data': 0.0005747420946136117, 'timing/model': 1.2457148079993203, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 1.31}) +[2026-06-24 02:30:44,861][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.07792263478040695, 'train/grad_norm_pre_clip': 15.801420211791992, 'timing/data': 0.00044018204789608717, 'timing/model': 1.1169747200328857, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 1.31}) +[2026-06-24 02:30:48,946][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.04519474506378174, 'train/grad_norm_pre_clip': 18.77066421508789, 'timing/data': 0.0005768790142610669, 'timing/model': 1.1183650610037148, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 1.31}) +[2026-06-24 02:30:53,036][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.0839337706565857, 'train/grad_norm_pre_clip': 18.84620475769043, 'timing/data': 0.0004055460449308157, 'timing/model': 1.11767264502123, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 1.31}) +[2026-06-24 02:30:57,252][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.029803991317749023, 'train/grad_norm_pre_clip': 20.951974868774414, 'timing/data': 0.00037043599877506495, 'timing/model': 1.2520020280499011, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 1.31}) +[2026-06-24 02:31:01,341][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.10151100903749466, 'train/grad_norm_pre_clip': 16.3950138092041, 'timing/data': 0.0004128669388592243, 'timing/model': 1.123177130939439, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.31}) +[2026-06-24 02:31:05,445][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.1049007996916771, 'train/grad_norm_pre_clip': 18.15057373046875, 'timing/data': 0.0004183780401945114, 'timing/model': 1.1252699859905988, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 1.31}) +[2026-06-24 02:31:09,528][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.04217970371246338, 'train/grad_norm_pre_clip': 31.55573081970215, 'timing/data': 0.000373797956854105, 'timing/model': 1.1172514889622107, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 1.31}) +[2026-06-24 02:31:13,757][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.0427132248878479, 'train/grad_norm_pre_clip': 21.40786361694336, 'timing/data': 0.00041237997356802225, 'timing/model': 1.264892021077685, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 1.31}) +[2026-06-24 02:31:17,848][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.01120653934776783, 'train/grad_norm_pre_clip': 31.29853630065918, 'timing/data': 0.0004225069424137473, 'timing/model': 1.1161795780062675, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 1.31}) +[2026-06-24 02:31:21,932][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.04477845877408981, 'train/grad_norm_pre_clip': 20.477890014648438, 'timing/data': 0.0004176120273768902, 'timing/model': 1.1266774568939582, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 1.31}) +[2026-06-24 02:31:26,007][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.02512417361140251, 'train/grad_norm_pre_clip': 16.402894973754883, 'timing/data': 0.0005040260730311275, 'timing/model': 1.1166927230078727, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 1.31}) +[2026-06-24 02:31:30,258][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.07900330424308777, 'train/grad_norm_pre_clip': 11.683398246765137, 'timing/data': 0.0003620279021561146, 'timing/model': 1.2633849849225953, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 1.31}) +[2026-06-24 02:31:34,342][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.0015738036017864943, 'train/grad_norm_pre_clip': 17.196941375732422, 'timing/data': 0.00035365193616598845, 'timing/model': 1.1204121230402961, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 1.31}) +[2026-06-24 02:31:38,435][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.033816155046224594, 'train/grad_norm_pre_clip': 20.672040939331055, 'timing/data': 0.0003630190622061491, 'timing/model': 1.1218144060112536, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 1.31}) +[2026-06-24 02:31:42,519][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.0868540108203888, 'train/grad_norm_pre_clip': 21.215368270874023, 'timing/data': 0.0003654520260170102, 'timing/model': 1.1147819149773568, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 1.31}) +[2026-06-24 02:31:46,732][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.18296362459659576, 'train/grad_norm_pre_clip': 28.1383113861084, 'timing/data': 0.00036052404902875423, 'timing/model': 1.2504459650954232, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 1.31}) +[2026-06-24 02:31:50,812][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.16981108486652374, 'train/grad_norm_pre_clip': 57.36556625366211, 'timing/data': 0.000363666913472116, 'timing/model': 1.1178373820148408, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 1.31}) +[2026-06-24 02:31:54,890][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.006157307885587215, 'train/grad_norm_pre_clip': 41.40680694580078, 'timing/data': 0.00038182607386261225, 'timing/model': 1.1179815098876134, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 1.31}) +[2026-06-24 02:31:58,982][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.08378299325704575, 'train/grad_norm_pre_clip': 29.66012954711914, 'timing/data': 0.00035556789953261614, 'timing/model': 1.1233682360034436, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 1.31}) +[2026-06-24 02:32:03,207][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.04946126043796539, 'train/grad_norm_pre_clip': 13.470795631408691, 'timing/data': 0.0004176730290055275, 'timing/model': 1.260241854004562, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 1.32}) +[2026-06-24 02:32:07,292][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.05057620257139206, 'train/grad_norm_pre_clip': 13.060802459716797, 'timing/data': 0.0005088060861453414, 'timing/model': 1.1250949939712882, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 1.32}) +[2026-06-24 02:32:11,393][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3400 +[2026-06-24 02:33:16,498][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3400 +[2026-06-24 02:33:16,500][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3400 +[2026-06-24 02:33:55,573][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3400 +[2026-06-24 02:33:55,574][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.07481826096773148, 'train/grad_norm_pre_clip': 22.49630355834961, 'eval/action_loss/seconds': 65.10290482302662, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.030265461653470993, 'eval/latency_2/loss': 0.030265461653470993, 'eval/flappy/loss': 0.030265461653470993, 'eval/flappy/latency_2/loss': 0.030265461653470993, 'eval/flap_precision': 0.9754098360655737, 'eval/flap_recall': 0.9687924016282226, 'eval/flap_f1': 0.9720898570456092, 'eval/flap_rate_model': 0.20333333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9720898570456091, 'eval/latency_2/flap_precision': 0.9754098360655737, 'eval/latency_2/flap_recall': 0.9687924016282226, 'eval/latency_2/flap_f1': 0.9720898570456092, 'eval/latency_2/flap_rate_model': 0.20333333333333334, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9720898570456091, 'eval/action_classification/data_wait_seconds': 0.6362774968147278, 'eval/action_classification/predict_seconds': 38.30116653442383, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00038998096715658903, 'timing/model': 1.1211683029541746, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 1.32}) +[2026-06-24 02:33:59,621][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.012104866094887257, 'train/grad_norm_pre_clip': 14.039169311523438, 'timing/data': 0.0003696240019053221, 'timing/model': 1.111409401986748, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 1.32}) +[2026-06-24 02:34:03,853][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.047882888466119766, 'train/grad_norm_pre_clip': 19.918479919433594, 'timing/data': 0.0004118740325793624, 'timing/model': 1.1093372469767928, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 1.32}) +[2026-06-24 02:34:07,898][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.19461016356945038, 'train/grad_norm_pre_clip': 26.15835189819336, 'timing/data': 0.00035473599564284086, 'timing/model': 1.1085515570593998, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 1.32}) +[2026-06-24 02:34:11,952][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.03394303470849991, 'train/grad_norm_pre_clip': 11.364214897155762, 'timing/data': 0.00035198801197111607, 'timing/model': 1.1174518769839779, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 1.32}) +[2026-06-24 02:34:16,090][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.018870238214731216, 'train/grad_norm_pre_clip': 15.974834442138672, 'timing/data': 0.0003578390460461378, 'timing/model': 1.1100582820363343, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 1.32}) +[2026-06-24 02:34:20,152][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.026436954736709595, 'train/grad_norm_pre_clip': 11.313467025756836, 'timing/data': 0.00037203298415988684, 'timing/model': 1.1123281019972637, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 1.32}) +[2026-06-24 02:34:24,217][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.0023697020951658487, 'train/grad_norm_pre_clip': 14.043641090393066, 'timing/data': 0.0003609099658206105, 'timing/model': 1.1148275280138478, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 1.32}) +[2026-06-24 02:34:28,278][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.03609441965818405, 'train/grad_norm_pre_clip': 20.099363327026367, 'timing/data': 0.0005829369183629751, 'timing/model': 1.1155160799389705, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 1.32}) +[2026-06-24 02:34:32,508][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.001929264166392386, 'train/grad_norm_pre_clip': 113.01182556152344, 'timing/data': 0.0003596539609134197, 'timing/model': 1.2789582510013133, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 1.32}) +[2026-06-24 02:34:36,580][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.02630162611603737, 'train/grad_norm_pre_clip': 19.573030471801758, 'timing/data': 0.0005310049746185541, 'timing/model': 1.1123069329187274, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 1.32}) +[2026-06-24 02:34:40,655][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.06751126050949097, 'train/grad_norm_pre_clip': 16.04836082458496, 'timing/data': 0.0004323340253904462, 'timing/model': 1.111136187100783, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 1.32}) +[2026-06-24 02:34:44,726][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.005197523627430201, 'train/grad_norm_pre_clip': 5.058985710144043, 'timing/data': 0.00036384991835802794, 'timing/model': 1.1187418620102108, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 1.32}) +[2026-06-24 02:34:48,888][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.01471608504652977, 'train/grad_norm_pre_clip': 31.28301429748535, 'timing/data': 0.0003550969995558262, 'timing/model': 1.1978272360283881, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 1.32}) +[2026-06-24 02:34:52,961][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.04880977421998978, 'train/grad_norm_pre_clip': 8.745183944702148, 'timing/data': 0.00034989090636372566, 'timing/model': 1.1175060640089214, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 1.32}) +[2026-06-24 02:34:57,036][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.04661790281534195, 'train/grad_norm_pre_clip': 20.28885269165039, 'timing/data': 0.0003467480419203639, 'timing/model': 1.1153677179245278, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 1.32}) +[2026-06-24 02:35:01,108][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.021583199501037598, 'train/grad_norm_pre_clip': 15.642772674560547, 'timing/data': 0.0003749299794435501, 'timing/model': 1.1165686760796234, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 1.32}) +[2026-06-24 02:35:05,256][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.06663268059492111, 'train/grad_norm_pre_clip': 16.811595916748047, 'timing/data': 0.00037011702079325914, 'timing/model': 1.192051779013127, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 1.32}) +[2026-06-24 02:35:09,316][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.010876757092773914, 'train/grad_norm_pre_clip': 21.01120948791504, 'timing/data': 0.00036701199132949114, 'timing/model': 1.1173373489873484, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 1.32}) +[2026-06-24 02:35:13,386][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.034048471599817276, 'train/grad_norm_pre_clip': 13.182319641113281, 'timing/data': 0.00036157702561467886, 'timing/model': 1.1152854770189151, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 1.32}) +[2026-06-24 02:35:17,479][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.029054714366793633, 'train/grad_norm_pre_clip': 20.035449981689453, 'timing/data': 0.0003544959472492337, 'timing/model': 1.11897954903543, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 1.32}) +[2026-06-24 02:35:21,628][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.10255328565835953, 'train/grad_norm_pre_clip': 27.109996795654297, 'timing/data': 0.0004007460083812475, 'timing/model': 1.192208581021987, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 1.32}) +[2026-06-24 02:35:25,703][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.0010315561667084694, 'train/grad_norm_pre_clip': 10.078752517700195, 'timing/data': 0.0003683939576148987, 'timing/model': 1.114676343044266, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 1.32}) +[2026-06-24 02:35:29,782][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.024307649582624435, 'train/grad_norm_pre_clip': 17.545297622680664, 'timing/data': 0.00036263500805944204, 'timing/model': 1.1168986010598019, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 1.32}) +[2026-06-24 02:35:33,864][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.039754536002874374, 'train/grad_norm_pre_clip': 38.039093017578125, 'timing/data': 0.00037810695357620716, 'timing/model': 1.114681908977218, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 1.33}) +[2026-06-24 02:35:38,097][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.0074101099744439125, 'train/grad_norm_pre_clip': 9.566845893859863, 'timing/data': 0.0004210949409753084, 'timing/model': 1.2621249079238623, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 1.33}) +[2026-06-24 02:35:42,175][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.0028691997285932302, 'train/grad_norm_pre_clip': 17.170574188232422, 'timing/data': 0.00036166002973914146, 'timing/model': 1.1133419019170105, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 1.33}) +[2026-06-24 02:35:46,257][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.08115871995687485, 'train/grad_norm_pre_clip': 30.015274047851562, 'timing/data': 0.00037259189411997795, 'timing/model': 1.1159406750230119, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 1.33}) +[2026-06-24 02:35:50,327][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.12225154787302017, 'train/grad_norm_pre_clip': 24.631792068481445, 'timing/data': 0.0003736040089279413, 'timing/model': 1.116281104972586, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 1.33}) +[2026-06-24 02:35:54,513][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.11625803261995316, 'train/grad_norm_pre_clip': 18.56930923461914, 'timing/data': 0.00035990693140774965, 'timing/model': 1.1138974580680951, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 1.33}) +[2026-06-24 02:35:58,597][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.0713142454624176, 'train/grad_norm_pre_clip': 16.775924682617188, 'timing/data': 0.00035410490818321705, 'timing/model': 1.119861170067452, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 1.33}) +[2026-06-24 02:36:02,682][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.02925829030573368, 'train/grad_norm_pre_clip': 14.02630615234375, 'timing/data': 0.0003568211104720831, 'timing/model': 1.1200476589147002, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 1.33}) +[2026-06-24 02:36:06,760][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.011456980369985104, 'train/grad_norm_pre_clip': 31.89492416381836, 'timing/data': 0.00036173907574266195, 'timing/model': 1.1207245650002733, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.33}) +[2026-06-24 02:36:11,008][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.010412778705358505, 'train/grad_norm_pre_clip': 13.871686935424805, 'timing/data': 0.0003693989710882306, 'timing/model': 1.116793806082569, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 1.33}) +[2026-06-24 02:36:15,096][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.0004505309625528753, 'train/grad_norm_pre_clip': 22.591535568237305, 'timing/data': 0.00035620504058897495, 'timing/model': 1.1251208579633385, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 1.33}) +[2026-06-24 02:36:19,179][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.0040380703285336494, 'train/grad_norm_pre_clip': 7.775461673736572, 'timing/data': 0.00035055296029895544, 'timing/model': 1.1202158969826996, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 1.33}) +[2026-06-24 02:36:23,253][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.027908075600862503, 'train/grad_norm_pre_clip': 22.902149200439453, 'timing/data': 0.00038287602365016937, 'timing/model': 1.118432002956979, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 1.33}) +[2026-06-24 02:36:27,505][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.048309069126844406, 'train/grad_norm_pre_clip': 38.785037994384766, 'timing/data': 0.00035622401628643274, 'timing/model': 1.1124571620021015, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 1.33}) +[2026-06-24 02:36:31,604][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.006957726553082466, 'train/grad_norm_pre_clip': 19.245708465576172, 'timing/data': 0.00035778700839728117, 'timing/model': 1.1194341100053862, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 1.33}) +[2026-06-24 02:36:35,688][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.0673707127571106, 'train/grad_norm_pre_clip': 25.385513305664062, 'timing/data': 0.00036070297937840223, 'timing/model': 1.114354621968232, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 1.33}) +[2026-06-24 02:36:39,770][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.174184650182724, 'train/grad_norm_pre_clip': 34.33598709106445, 'timing/data': 0.00038559793028980494, 'timing/model': 1.1152049669763073, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 1.33}) +[2026-06-24 02:36:43,926][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.014725111424922943, 'train/grad_norm_pre_clip': 7.133340835571289, 'timing/data': 0.00035937305074185133, 'timing/model': 1.1181461149826646, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 1.33}) +[2026-06-24 02:36:48,031][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.0006885715411044657, 'train/grad_norm_pre_clip': 16.58332061767578, 'timing/data': 0.00036156096030026674, 'timing/model': 1.119524402078241, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 1.33}) +[2026-06-24 02:36:52,126][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.06483042985200882, 'train/grad_norm_pre_clip': 19.135662078857422, 'timing/data': 0.0004872690187767148, 'timing/model': 1.1224045370472595, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 1.33}) +[2026-06-24 02:36:56,213][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.06534881144762039, 'train/grad_norm_pre_clip': 13.8764009475708, 'timing/data': 0.0005604269681498408, 'timing/model': 1.1200019509997219, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 1.33}) +[2026-06-24 02:37:00,373][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.02273690141737461, 'train/grad_norm_pre_clip': 17.834365844726562, 'timing/data': 0.00044765102211385965, 'timing/model': 1.1202076100744307, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 1.33}) +[2026-06-24 02:37:04,469][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.12291201949119568, 'train/grad_norm_pre_clip': 41.360923767089844, 'timing/data': 0.0003472650423645973, 'timing/model': 1.1225154340500012, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 1.33}) +[2026-06-24 02:37:08,567][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.0620821975171566, 'train/grad_norm_pre_clip': 16.82757568359375, 'timing/data': 0.0003726129652932286, 'timing/model': 1.126764836953953, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 1.33}) +[2026-06-24 02:37:12,647][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.017736168578267097, 'train/grad_norm_pre_clip': 15.872361183166504, 'timing/data': 0.0006039880681782961, 'timing/model': 1.1205877119209617, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 1.33}) +[2026-06-24 02:37:16,878][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.07417736947536469, 'train/grad_norm_pre_clip': 54.02700424194336, 'timing/data': 0.0005557530093938112, 'timing/model': 1.1280255150049925, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 1.33}) +[2026-06-24 02:37:20,992][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.0278757493942976, 'train/grad_norm_pre_clip': 42.09497833251953, 'timing/data': 0.00043380504939705133, 'timing/model': 1.1227722810581326, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 1.34}) +[2026-06-24 02:37:25,081][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.023487122729420662, 'train/grad_norm_pre_clip': 27.918197631835938, 'timing/data': 0.0004007419338449836, 'timing/model': 1.1207245420664549, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 1.34}) +[2026-06-24 02:37:29,244][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.3900449573993683, 'train/grad_norm_pre_clip': 32.48454666137695, 'timing/data': 0.00040273298509418964, 'timing/model': 1.191563419997692, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 1.34}) +[2026-06-24 02:37:33,332][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.01641695201396942, 'train/grad_norm_pre_clip': 48.15644454956055, 'timing/data': 0.00035466498229652643, 'timing/model': 1.1166704409988597, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 1.34}) +[2026-06-24 02:37:37,430][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.008591018617153168, 'train/grad_norm_pre_clip': 29.980144500732422, 'timing/data': 0.0005749179981648922, 'timing/model': 1.1257529850117862, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 1.34}) +[2026-06-24 02:37:41,523][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.029106836766004562, 'train/grad_norm_pre_clip': 12.79476261138916, 'timing/data': 0.00037420110311359167, 'timing/model': 1.1146673220209777, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 1.34}) +[2026-06-24 02:37:45,616][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.008019907400012016, 'train/grad_norm_pre_clip': 30.453197479248047, 'timing/data': 0.0003564909566193819, 'timing/model': 1.1235660569509491, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 1.34}) +[2026-06-24 02:37:49,841][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.3097960650920868, 'train/grad_norm_pre_clip': 22.987937927246094, 'timing/data': 0.0003562190104275942, 'timing/model': 1.1262819049879909, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 1.34}) +[2026-06-24 02:37:53,921][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.07035767287015915, 'train/grad_norm_pre_clip': 10.601460456848145, 'timing/data': 0.0003581640776246786, 'timing/model': 1.1180746890604496, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 1.34}) +[2026-06-24 02:37:58,013][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.07135206460952759, 'train/grad_norm_pre_clip': 18.25925636291504, 'timing/data': 0.00042669696267694235, 'timing/model': 1.1217257040552795, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 1.34}) +[2026-06-24 02:38:02,092][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.08565528690814972, 'train/grad_norm_pre_clip': 15.45407772064209, 'timing/data': 0.0006252189632505178, 'timing/model': 1.1147775530116633, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 1.34}) +[2026-06-24 02:38:06,331][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.1410517543554306, 'train/grad_norm_pre_clip': 32.529808044433594, 'timing/data': 0.00036532606463879347, 'timing/model': 1.1258640949381515, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 1.34}) +[2026-06-24 02:38:10,417][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.0015094743575900793, 'train/grad_norm_pre_clip': 34.4945068359375, 'timing/data': 0.0003638539928942919, 'timing/model': 1.1207775649381801, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 1.34}) +[2026-06-24 02:38:14,503][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.007565408945083618, 'train/grad_norm_pre_clip': 4.569927215576172, 'timing/data': 0.0003595850430428982, 'timing/model': 1.1192757729440928, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 1.34}) +[2026-06-24 02:38:18,594][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.036613937467336655, 'train/grad_norm_pre_clip': 18.92719078063965, 'timing/data': 0.0004111670423299074, 'timing/model': 1.1222217300673947, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 1.34}) +[2026-06-24 02:38:22,746][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.0025456957519054413, 'train/grad_norm_pre_clip': 9.081815719604492, 'timing/data': 0.0003997380845248699, 'timing/model': 1.1191126368939877, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 1.34}) +[2026-06-24 02:38:26,837][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.0003606972168199718, 'train/grad_norm_pre_clip': 13.657537460327148, 'timing/data': 0.00039670791011303663, 'timing/model': 1.1190452500013635, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 1.34}) +[2026-06-24 02:38:30,939][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.01599971204996109, 'train/grad_norm_pre_clip': 10.411482810974121, 'timing/data': 0.00035945000126957893, 'timing/model': 1.1229892160044983, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 1.34}) +[2026-06-24 02:38:35,026][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.006661622319370508, 'train/grad_norm_pre_clip': 5.3241987228393555, 'timing/data': 0.0003620149800553918, 'timing/model': 1.1177519570337608, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 1.34}) +[2026-06-24 02:38:39,161][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.05039150267839432, 'train/grad_norm_pre_clip': 15.298935890197754, 'timing/data': 0.0003564649960026145, 'timing/model': 1.1151629600208253, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 1.34}) +[2026-06-24 02:38:43,255][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.09112044423818588, 'train/grad_norm_pre_clip': 14.330229759216309, 'timing/data': 0.0003551510162651539, 'timing/model': 1.1208688409533352, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 1.34}) +[2026-06-24 02:38:47,344][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.05689563602209091, 'train/grad_norm_pre_clip': 41.44959259033203, 'timing/data': 0.0004974699113518, 'timing/model': 1.1185802340041846, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 1.34}) +[2026-06-24 02:38:51,450][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.09955589473247528, 'train/grad_norm_pre_clip': 33.058265686035156, 'timing/data': 0.0003733229823410511, 'timing/model': 1.1201856209663674, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 1.34}) +[2026-06-24 02:38:55,706][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.0502733513712883, 'train/grad_norm_pre_clip': 33.44040298461914, 'timing/data': 0.00035915698390454054, 'timing/model': 1.1173356430372223, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 1.34}) +[2026-06-24 02:38:59,801][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.1588324010372162, 'train/grad_norm_pre_clip': 20.40410804748535, 'timing/data': 0.0005807429552078247, 'timing/model': 1.125319250044413, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 1.34}) +[2026-06-24 02:39:03,886][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.0326816588640213, 'train/grad_norm_pre_clip': 17.833370208740234, 'timing/data': 0.00036128005012869835, 'timing/model': 1.1197949199704453, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 1.34}) +[2026-06-24 02:39:07,977][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.19385075569152832, 'train/grad_norm_pre_clip': 40.27579879760742, 'timing/data': 0.0005106040043756366, 'timing/model': 1.1209500670665875, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 1.35}) +[2026-06-24 02:39:12,225][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.16236479580402374, 'train/grad_norm_pre_clip': 37.16197967529297, 'timing/data': 0.0004939170321449637, 'timing/model': 1.1183421689784154, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 1.35}) +[2026-06-24 02:39:16,310][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.1846836358308792, 'train/grad_norm_pre_clip': 23.839704513549805, 'timing/data': 0.00035127997398376465, 'timing/model': 1.1171336090192199, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 1.35}) +[2026-06-24 02:39:20,392][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.25348812341690063, 'train/grad_norm_pre_clip': 36.652591705322266, 'timing/data': 0.00035455403849482536, 'timing/model': 1.1202296960400417, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 1.35}) +[2026-06-24 02:39:24,495][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.06214720383286476, 'train/grad_norm_pre_clip': 20.21107292175293, 'timing/data': 0.00041305599734187126, 'timing/model': 1.115978152025491, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 1.35}) +[2026-06-24 02:39:28,764][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.04129371419548988, 'train/grad_norm_pre_clip': 19.55219268798828, 'timing/data': 0.0005589439533650875, 'timing/model': 1.1210336510557681, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.35}) +[2026-06-24 02:39:32,854][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.18977394700050354, 'train/grad_norm_pre_clip': 31.112525939941406, 'timing/data': 0.0003683279501274228, 'timing/model': 1.1308425910538062, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 1.35}) +[2026-06-24 02:39:36,946][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.06017300859093666, 'train/grad_norm_pre_clip': 19.578659057617188, 'timing/data': 0.00035100197419524193, 'timing/model': 1.1197363229002804, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 1.35}) +[2026-06-24 02:39:41,037][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.040296636521816254, 'train/grad_norm_pre_clip': 24.50489044189453, 'timing/data': 0.00035074108745902777, 'timing/model': 1.1181864549871534, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 1.35}) +[2026-06-24 02:39:45,298][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.01406183186918497, 'train/grad_norm_pre_clip': 19.378841400146484, 'timing/data': 0.0003625730751082301, 'timing/model': 1.1211916889296845, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 1.35}) +[2026-06-24 02:39:49,391][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.0017564166337251663, 'train/grad_norm_pre_clip': 6.828467845916748, 'timing/data': 0.00043794605880975723, 'timing/model': 1.1226971501018852, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 1.35}) +[2026-06-24 02:39:53,471][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.024537639692425728, 'train/grad_norm_pre_clip': 23.045387268066406, 'timing/data': 0.0003929509548470378, 'timing/model': 1.120857222005725, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 1.35}) +[2026-06-24 02:39:57,564][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.08554169535636902, 'train/grad_norm_pre_clip': 19.919208526611328, 'timing/data': 0.00036373105831444263, 'timing/model': 1.1187548480229452, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 1.35}) +[2026-06-24 02:40:01,807][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.006458983290940523, 'train/grad_norm_pre_clip': 3.531095266342163, 'timing/data': 0.00036068400368094444, 'timing/model': 1.1210489500081167, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 1.35}) +[2026-06-24 02:40:05,888][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.030948881059885025, 'train/grad_norm_pre_clip': 21.00765609741211, 'timing/data': 0.00042784796096384525, 'timing/model': 1.116986520937644, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 1.35}) +[2026-06-24 02:40:09,978][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.015349042601883411, 'train/grad_norm_pre_clip': 21.70895004272461, 'timing/data': 0.0004435880109667778, 'timing/model': 1.11965953593608, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 1.35}) +[2026-06-24 02:40:14,072][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.032496511936187744, 'train/grad_norm_pre_clip': 33.62971115112305, 'timing/data': 0.00041745707858353853, 'timing/model': 1.1179052999941632, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 1.35}) +[2026-06-24 02:40:18,332][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.10984955728054047, 'train/grad_norm_pre_clip': 21.05175018310547, 'timing/data': 0.0004178490489721298, 'timing/model': 1.125877258949913, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 1.35}) +[2026-06-24 02:40:22,437][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.04201672971248627, 'train/grad_norm_pre_clip': 14.535505294799805, 'timing/data': 0.00036941899452358484, 'timing/model': 1.1288983720587566, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 1.35}) +[2026-06-24 02:40:26,522][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.015138248912990093, 'train/grad_norm_pre_clip': 21.646570205688477, 'timing/data': 0.0004042740911245346, 'timing/model': 1.1186833400279284, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 1.35}) +[2026-06-24 02:40:30,617][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.03303316980600357, 'train/grad_norm_pre_clip': 23.69615936279297, 'timing/data': 0.0003603999502956867, 'timing/model': 1.137079919106327, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 1.35}) +[2026-06-24 02:40:34,876][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.0015353545313701034, 'train/grad_norm_pre_clip': 17.870527267456055, 'timing/data': 0.0004521770169958472, 'timing/model': 1.1175084300339222, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 1.35}) +[2026-06-24 02:40:38,954][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.056782130151987076, 'train/grad_norm_pre_clip': 20.586965560913086, 'timing/data': 0.0006065199850127101, 'timing/model': 1.1194954459788278, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 1.35}) +[2026-06-24 02:40:43,031][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.06809879094362259, 'train/grad_norm_pre_clip': 17.63377571105957, 'timing/data': 0.0004684000741690397, 'timing/model': 1.1145019839750603, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 1.35}) +[2026-06-24 02:40:47,104][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3500 +[2026-06-24 02:41:52,492][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3500 +[2026-06-24 02:41:52,494][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3500 +[2026-06-24 02:42:30,676][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3500 +[2026-06-24 02:53:28,877][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3500 +[2026-06-24 02:53:28,878][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.013893869705498219, 'train/grad_norm_pre_clip': 18.14067840576172, 'eval/action_loss/seconds': 65.38561936200131, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.03276028484106064, 'eval/latency_2/loss': 0.03276028484106064, 'eval/flappy/loss': 0.03276028484106064, 'eval/flappy/latency_2/loss': 0.03276028484106064, 'eval/flap_precision': 0.9832402234636871, 'eval/flap_recall': 0.9552238805970149, 'eval/flap_f1': 0.969029593943565, 'eval/flap_rate_model': 0.1988888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.969029593943565, 'eval/latency_2/flap_precision': 0.9832402234636871, 'eval/latency_2/flap_recall': 0.9552238805970149, 'eval/latency_2/flap_f1': 0.969029593943565, 'eval/latency_2/flap_rate_model': 0.1988888888888889, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.969029593943565, 'eval/action_classification/data_wait_seconds': 0.61031174659729, 'eval/action_classification/predict_seconds': 37.44184112548828, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 408.66000000001924, 'rl_games_eval/mid_train/mean_length': 3311.2, 'rl_games_eval/mid_train/std_reward': 71.8800000000128, 'rl_games_eval/mid_train/std_length': 577.6, 'rl_games_eval/mid_train/macro_mean_reward': 408.66000000001924, 'rl_games_eval/mid_train/macro_mean_length': 3311.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 408.66000000001924, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3311.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 71.8800000000128, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 577.6, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 408.66000000001924, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003573920112103224, 'timing/model': 1.118084448040463, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 1.35}) +[2026-06-24 02:53:28,878][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_3500_state +[2026-06-24 02:53:28,881][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 02:53:32,672][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_3500_state/model.safetensors +[2026-06-24 02:53:40,523][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_3500_state/optimizer.bin +[2026-06-24 02:53:40,523][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_3500_state/sampler.bin +[2026-06-24 02:53:40,524][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-24 02:53:40,525][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-24 02:53:40,603][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-24 02:53:40,603][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 02:53:40,616][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 02:53:44,564][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.11615049093961716, 'train/grad_norm_pre_clip': 15.250418663024902, 'timing/data': 0.0003599140327423811, 'timing/model': 1.0806412380188704, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 1.35}) +[2026-06-24 02:53:48,527][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.019469985738396645, 'train/grad_norm_pre_clip': 17.652687072753906, 'timing/data': 0.00035865698009729385, 'timing/model': 1.0902504989644513, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 1.36}) +[2026-06-24 02:53:52,512][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.07014346867799759, 'train/grad_norm_pre_clip': 21.4705810546875, 'timing/data': 0.0003494920674711466, 'timing/model': 1.0926398950396106, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 1.36}) +[2026-06-24 02:53:56,596][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.21163766086101532, 'train/grad_norm_pre_clip': 29.332094192504883, 'timing/data': 0.00035794905852526426, 'timing/model': 1.095031277043745, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 1.36}) +[2026-06-24 02:54:00,584][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.12684662640094757, 'train/grad_norm_pre_clip': 18.533918380737305, 'timing/data': 0.00035742903128266335, 'timing/model': 1.0974261730443686, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 1.36}) +[2026-06-24 02:54:04,584][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.0858108252286911, 'train/grad_norm_pre_clip': 14.652793884277344, 'timing/data': 0.0004004549700766802, 'timing/model': 1.094552065944299, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 1.36}) +[2026-06-24 02:54:08,596][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.027843626216053963, 'train/grad_norm_pre_clip': 13.916047096252441, 'timing/data': 0.00044951599556952715, 'timing/model': 1.1016975529491901, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 1.36}) +[2026-06-24 02:54:12,705][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.024858633056282997, 'train/grad_norm_pre_clip': 16.884918212890625, 'timing/data': 0.00041295692790299654, 'timing/model': 1.1080278989393264, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 1.36}) +[2026-06-24 02:54:16,736][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.037860263139009476, 'train/grad_norm_pre_clip': 39.152523040771484, 'timing/data': 0.0003514980198815465, 'timing/model': 1.1029111800016835, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 1.36}) +[2026-06-24 02:54:20,759][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.004315491765737534, 'train/grad_norm_pre_clip': 13.250541687011719, 'timing/data': 0.00034785596653819084, 'timing/model': 1.1040246119955555, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 1.36}) +[2026-06-24 02:54:24,797][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.01821354776620865, 'train/grad_norm_pre_clip': 5.986318111419678, 'timing/data': 0.00035063596442341805, 'timing/model': 1.1054982079658657, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 1.36}) +[2026-06-24 02:54:28,990][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.071345254778862, 'train/grad_norm_pre_clip': 29.725597381591797, 'timing/data': 0.0005723579088225961, 'timing/model': 1.106484368094243, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 1.36}) +[2026-06-24 02:54:33,027][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.013422002084553242, 'train/grad_norm_pre_clip': 9.321313858032227, 'timing/data': 0.00038596102967858315, 'timing/model': 1.1064397380687296, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 1.36}) +[2026-06-24 02:54:37,070][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.08319686353206635, 'train/grad_norm_pre_clip': 29.83013916015625, 'timing/data': 0.000396591960452497, 'timing/model': 1.1082775359973311, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 1.36}) +[2026-06-24 02:54:41,125][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.013230492360889912, 'train/grad_norm_pre_clip': 17.8743839263916, 'timing/data': 0.00034784700255841017, 'timing/model': 1.1173701609950513, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 1.36}) +[2026-06-24 02:54:45,260][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.01429132092744112, 'train/grad_norm_pre_clip': 6.9562668800354, 'timing/data': 0.0005726750241592526, 'timing/model': 1.1056593969697133, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 1.36}) +[2026-06-24 02:54:49,312][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.02358975261449814, 'train/grad_norm_pre_clip': 9.846080780029297, 'timing/data': 0.00036570196971297264, 'timing/model': 1.1101437769830227, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 1.36}) +[2026-06-24 02:54:53,363][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.1236882135272026, 'train/grad_norm_pre_clip': 22.092472076416016, 'timing/data': 0.0003472109092399478, 'timing/model': 1.1061202900018543, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 1.36}) +[2026-06-24 02:54:57,413][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.004427282605320215, 'train/grad_norm_pre_clip': 7.974771976470947, 'timing/data': 0.0003519420279189944, 'timing/model': 1.1107910010032356, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 1.36}) +[2026-06-24 02:55:01,600][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.05368461087346077, 'train/grad_norm_pre_clip': 14.94856071472168, 'timing/data': 0.00034768402110785246, 'timing/model': 1.2318783659720793, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 1.36}) +[2026-06-24 02:55:05,665][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.030342712998390198, 'train/grad_norm_pre_clip': 19.015745162963867, 'timing/data': 0.00035149697214365005, 'timing/model': 1.1083366880193353, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 1.36}) +[2026-06-24 02:55:09,726][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.0745556503534317, 'train/grad_norm_pre_clip': 27.633955001831055, 'timing/data': 0.0003450009971857071, 'timing/model': 1.108568444964476, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 1.36}) +[2026-06-24 02:55:13,789][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.0004010230186395347, 'train/grad_norm_pre_clip': 11.966218948364258, 'timing/data': 0.00034710997715592384, 'timing/model': 1.1136851540068164, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 1.36}) +[2026-06-24 02:55:17,857][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.05305115878582001, 'train/grad_norm_pre_clip': 47.54526138305664, 'timing/data': 0.000357068027369678, 'timing/model': 1.1139581319876015, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 1.36}) +[2026-06-24 02:55:22,098][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.041133761405944824, 'train/grad_norm_pre_clip': 21.81556510925293, 'timing/data': 0.00039771199226379395, 'timing/model': 1.1150836129672825, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 1.36}) +[2026-06-24 02:55:26,160][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.07206538319587708, 'train/grad_norm_pre_clip': 22.034931182861328, 'timing/data': 0.0006020690780133009, 'timing/model': 1.1133358930237591, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 1.36}) +[2026-06-24 02:55:30,232][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.05797315016388893, 'train/grad_norm_pre_clip': 26.572757720947266, 'timing/data': 0.00040537596214562654, 'timing/model': 1.1156626710435376, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 1.37}) +[2026-06-24 02:55:34,313][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.02095072530210018, 'train/grad_norm_pre_clip': 24.420333862304688, 'timing/data': 0.00035702798049896955, 'timing/model': 1.112673080060631, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 1.37}) +[2026-06-24 02:55:38,450][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.13030003011226654, 'train/grad_norm_pre_clip': 14.19308090209961, 'timing/data': 0.00035186298191547394, 'timing/model': 1.115375841036439, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 1.37}) +[2026-06-24 02:55:42,527][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.007784911897033453, 'train/grad_norm_pre_clip': 23.351871490478516, 'timing/data': 0.0003494899719953537, 'timing/model': 1.1200228170491755, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.37}) +[2026-06-24 02:55:46,593][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.0019344482570886612, 'train/grad_norm_pre_clip': 9.6094970703125, 'timing/data': 0.0003956030122935772, 'timing/model': 1.112610241980292, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 1.37}) +[2026-06-24 02:55:50,667][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.03755809739232063, 'train/grad_norm_pre_clip': 8.947896957397461, 'timing/data': 0.0003566909581422806, 'timing/model': 1.1155858249403536, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 1.37}) +[2026-06-24 02:55:54,820][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.14605332911014557, 'train/grad_norm_pre_clip': 24.033119201660156, 'timing/data': 0.0003455500118434429, 'timing/model': 1.1196462260559201, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 1.37}) +[2026-06-24 02:55:58,895][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.05515880882740021, 'train/grad_norm_pre_clip': 18.34111976623535, 'timing/data': 0.0003438289277255535, 'timing/model': 1.1148456070804968, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 1.37}) +[2026-06-24 02:56:02,970][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.0005820911610499024, 'train/grad_norm_pre_clip': 13.842070579528809, 'timing/data': 0.0003453620010986924, 'timing/model': 1.115339134936221, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 1.37}) +[2026-06-24 02:56:07,054][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.07572988420724869, 'train/grad_norm_pre_clip': 28.470788955688477, 'timing/data': 0.000390655011869967, 'timing/model': 1.115821473998949, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 1.37}) +[2026-06-24 02:56:11,203][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.028809739276766777, 'train/grad_norm_pre_clip': 12.026511192321777, 'timing/data': 0.0006075049750506878, 'timing/model': 1.1131750450003892, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 1.37}) +[2026-06-24 02:56:15,268][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.11994507908821106, 'train/grad_norm_pre_clip': 46.34066390991211, 'timing/data': 0.000353395938873291, 'timing/model': 1.1125690449262038, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 1.37}) +[2026-06-24 02:56:19,324][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.05910369008779526, 'train/grad_norm_pre_clip': 19.462329864501953, 'timing/data': 0.0004049309063702822, 'timing/model': 1.1130812180927023, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 1.37}) +[2026-06-24 02:56:23,392][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.033401601016521454, 'train/grad_norm_pre_clip': 4.036365985870361, 'timing/data': 0.0003511341055855155, 'timing/model': 1.1103425489272922, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 1.37}) +[2026-06-24 02:56:27,597][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.0011275712167844176, 'train/grad_norm_pre_clip': 57.451908111572266, 'timing/data': 0.0006008949130773544, 'timing/model': 1.1148633039556444, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 1.37}) +[2026-06-24 02:56:31,670][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.04038291051983833, 'train/grad_norm_pre_clip': 15.623544692993164, 'timing/data': 0.0004101820522919297, 'timing/model': 1.1095241870498285, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 1.37}) +[2026-06-24 02:56:35,763][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.03548235073685646, 'train/grad_norm_pre_clip': 13.327258110046387, 'timing/data': 0.000579355051741004, 'timing/model': 1.1220686220331118, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 1.37}) +[2026-06-24 02:56:39,949][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.051495328545570374, 'train/grad_norm_pre_clip': 9.138191223144531, 'timing/data': 0.00035059498623013496, 'timing/model': 1.2155912619782612, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 1.37}) +[2026-06-24 02:56:44,033][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.018906736746430397, 'train/grad_norm_pre_clip': 21.94700050354004, 'timing/data': 0.0003643200034275651, 'timing/model': 1.1186030480312183, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 1.37}) +[2026-06-24 02:56:48,106][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.004496510606259108, 'train/grad_norm_pre_clip': 16.667613983154297, 'timing/data': 0.0005740640917792916, 'timing/model': 1.1206540330313146, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 1.37}) +[2026-06-24 02:56:52,174][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.04020099341869354, 'train/grad_norm_pre_clip': 34.15822219848633, 'timing/data': 0.0004067549016326666, 'timing/model': 1.117968280101195, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 1.37}) +[2026-06-24 02:56:56,432][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.09463588893413544, 'train/grad_norm_pre_clip': 21.526744842529297, 'timing/data': 0.0005067760357633233, 'timing/model': 1.2916253129951656, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 1.37}) +[2026-06-24 02:57:00,514][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.15223023295402527, 'train/grad_norm_pre_clip': 40.005863189697266, 'timing/data': 0.0003561010817065835, 'timing/model': 1.123313186923042, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 1.37}) +[2026-06-24 02:57:04,592][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.003202814143151045, 'train/grad_norm_pre_clip': 16.229143142700195, 'timing/data': 0.0003941210452467203, 'timing/model': 1.1154501559212804, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 1.37}) +[2026-06-24 02:57:08,664][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.09270249307155609, 'train/grad_norm_pre_clip': 22.698299407958984, 'timing/data': 0.0003631219733506441, 'timing/model': 1.109434206970036, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 1.37}) +[2026-06-24 02:57:12,829][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.02085314691066742, 'train/grad_norm_pre_clip': 22.332876205444336, 'timing/data': 0.0003472269745543599, 'timing/model': 1.1172021710081026, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 1.37}) +[2026-06-24 02:57:16,915][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.2306119054555893, 'train/grad_norm_pre_clip': 30.091718673706055, 'timing/data': 0.0003522430779412389, 'timing/model': 1.121042498969473, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 1.38}) +[2026-06-24 02:57:20,990][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.1177426278591156, 'train/grad_norm_pre_clip': 9.520084381103516, 'timing/data': 0.000354524003341794, 'timing/model': 1.1166172380326316, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 1.38}) +[2026-06-24 02:57:25,057][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.006865556351840496, 'train/grad_norm_pre_clip': 22.80599021911621, 'timing/data': 0.00035149697214365005, 'timing/model': 1.111558930017054, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 1.38}) +[2026-06-24 02:57:29,207][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.022445689886808395, 'train/grad_norm_pre_clip': 11.405411720275879, 'timing/data': 0.00036405096761882305, 'timing/model': 1.1137171169975773, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 1.38}) +[2026-06-24 02:57:33,285][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.15380600094795227, 'train/grad_norm_pre_clip': 38.341705322265625, 'timing/data': 0.0003523719497025013, 'timing/model': 1.1148381190141663, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 1.38}) +[2026-06-24 02:57:37,367][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.0001529673463664949, 'train/grad_norm_pre_clip': 22.36725425720215, 'timing/data': 0.00035978597588837147, 'timing/model': 1.121217904961668, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 1.38}) +[2026-06-24 02:57:41,474][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.01636693812906742, 'train/grad_norm_pre_clip': 51.14204025268555, 'timing/data': 0.00034516199957579374, 'timing/model': 1.116157492971979, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 1.38}) +[2026-06-24 02:57:45,691][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.12599600851535797, 'train/grad_norm_pre_clip': 39.83305358886719, 'timing/data': 0.0003796570235863328, 'timing/model': 1.1175811480497941, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 1.38}) +[2026-06-24 02:57:49,766][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.12283314019441605, 'train/grad_norm_pre_clip': 20.6842098236084, 'timing/data': 0.00038532994221895933, 'timing/model': 1.1200131480582058, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 1.38}) +[2026-06-24 02:57:53,837][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.06288713216781616, 'train/grad_norm_pre_clip': 30.625659942626953, 'timing/data': 0.00035271397791802883, 'timing/model': 1.1149107259698212, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 1.38}) +[2026-06-24 02:57:57,916][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.09032108634710312, 'train/grad_norm_pre_clip': 26.808982849121094, 'timing/data': 0.00035811797715723515, 'timing/model': 1.117860326077789, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 1.38}) +[2026-06-24 02:58:02,107][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.038489434868097305, 'train/grad_norm_pre_clip': 7.904973030090332, 'timing/data': 0.0003657410852611065, 'timing/model': 1.1160382869420573, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 1.38}) +[2026-06-24 02:58:06,179][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.06700114905834198, 'train/grad_norm_pre_clip': 17.53632354736328, 'timing/data': 0.00036298995837569237, 'timing/model': 1.1140536320162937, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 1.38}) +[2026-06-24 02:58:10,258][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.013396165333688259, 'train/grad_norm_pre_clip': 8.45472526550293, 'timing/data': 0.0003877950366586447, 'timing/model': 1.1179363670526072, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 1.38}) +[2026-06-24 02:58:14,346][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.07152846455574036, 'train/grad_norm_pre_clip': 22.57135009765625, 'timing/data': 0.00035942799877375364, 'timing/model': 1.1182298299390823, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 1.38}) +[2026-06-24 02:58:18,505][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.14663365483283997, 'train/grad_norm_pre_clip': 34.28596115112305, 'timing/data': 0.0003826400497928262, 'timing/model': 1.1154592599486932, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 1.38}) +[2026-06-24 02:58:22,586][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.04063943028450012, 'train/grad_norm_pre_clip': 17.828386306762695, 'timing/data': 0.0003487799549475312, 'timing/model': 1.1182173639535904, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 1.38}) +[2026-06-24 02:58:26,682][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.1160646304488182, 'train/grad_norm_pre_clip': 28.501754760742188, 'timing/data': 0.0003460630541667342, 'timing/model': 1.1214726188918576, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 1.38}) +[2026-06-24 02:58:30,766][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.05975070968270302, 'train/grad_norm_pre_clip': 26.051776885986328, 'timing/data': 0.000356200966052711, 'timing/model': 1.1170240530045703, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 1.38}) +[2026-06-24 02:58:34,998][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.015170742757618427, 'train/grad_norm_pre_clip': 9.309843063354492, 'timing/data': 0.00036043499130755663, 'timing/model': 1.112360249971971, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 1.38}) +[2026-06-24 02:58:39,076][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.007613595575094223, 'train/grad_norm_pre_clip': 3.702467441558838, 'timing/data': 0.00035130500327795744, 'timing/model': 1.1159820799948648, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 1.38}) +[2026-06-24 02:58:43,166][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.09203220903873444, 'train/grad_norm_pre_clip': 21.562580108642578, 'timing/data': 0.00038744707126170397, 'timing/model': 1.1211593779735267, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 1.38}) +[2026-06-24 02:58:47,248][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.0074773263186216354, 'train/grad_norm_pre_clip': 19.236745834350586, 'timing/data': 0.0003517989534884691, 'timing/model': 1.116591481026262, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 1.38}) +[2026-06-24 02:58:51,427][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.012219519354403019, 'train/grad_norm_pre_clip': 11.03529167175293, 'timing/data': 0.00035043794196099043, 'timing/model': 1.1222505200421438, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 1.38}) +[2026-06-24 02:58:55,503][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.022686593234539032, 'train/grad_norm_pre_clip': 18.342636108398438, 'timing/data': 0.0003545749932527542, 'timing/model': 1.1160891589242965, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 1.38}) +[2026-06-24 02:58:59,577][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.07410342991352081, 'train/grad_norm_pre_clip': 10.699141502380371, 'timing/data': 0.0004009460099041462, 'timing/model': 1.1138033280149102, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 1.38}) +[2026-06-24 02:59:03,661][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.005404261872172356, 'train/grad_norm_pre_clip': 9.478318214416504, 'timing/data': 0.0003537740558385849, 'timing/model': 1.1179722669767216, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.39}) +[2026-06-24 02:59:07,920][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.08574814349412918, 'train/grad_norm_pre_clip': 39.33244323730469, 'timing/data': 0.0004915270255878568, 'timing/model': 1.1220717260148376, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 1.39}) +[2026-06-24 02:59:12,008][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.03938416764140129, 'train/grad_norm_pre_clip': 24.61591339111328, 'timing/data': 0.000400564051233232, 'timing/model': 1.1152305490104482, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 1.39}) +[2026-06-24 02:59:16,097][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.02404518611729145, 'train/grad_norm_pre_clip': 23.057086944580078, 'timing/data': 0.00040944607462733984, 'timing/model': 1.1198861169395968, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 1.39}) +[2026-06-24 02:59:20,191][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.013395468704402447, 'train/grad_norm_pre_clip': 18.871065139770508, 'timing/data': 0.0004481810610741377, 'timing/model': 1.1263243870344013, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 1.39}) +[2026-06-24 02:59:24,349][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.027046173810958862, 'train/grad_norm_pre_clip': 33.582603454589844, 'timing/data': 0.0003560769837349653, 'timing/model': 1.1176487680058926, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 1.39}) +[2026-06-24 02:59:28,441][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.1607055515050888, 'train/grad_norm_pre_clip': 27.599803924560547, 'timing/data': 0.00035147799644619226, 'timing/model': 1.1243530169595033, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 1.39}) +[2026-06-24 02:59:32,527][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.031204216182231903, 'train/grad_norm_pre_clip': 20.054752349853516, 'timing/data': 0.0003555470611900091, 'timing/model': 1.118079188046977, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 1.39}) +[2026-06-24 02:59:36,601][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.004174390807747841, 'train/grad_norm_pre_clip': 22.911706924438477, 'timing/data': 0.0003953069681301713, 'timing/model': 1.11752578697633, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 1.39}) +[2026-06-24 02:59:40,747][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.01911165937781334, 'train/grad_norm_pre_clip': 25.731386184692383, 'timing/data': 0.0003526899963617325, 'timing/model': 1.1190695420373231, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 1.39}) +[2026-06-24 02:59:44,822][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.005315830931067467, 'train/grad_norm_pre_clip': 27.514888763427734, 'timing/data': 0.0003515490097925067, 'timing/model': 1.1155101800104603, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 1.39}) +[2026-06-24 02:59:48,902][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.012611648067831993, 'train/grad_norm_pre_clip': 4.161252498626709, 'timing/data': 0.00034484697971493006, 'timing/model': 1.117722684983164, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 1.39}) +[2026-06-24 02:59:52,987][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.023898251354694366, 'train/grad_norm_pre_clip': 24.579364776611328, 'timing/data': 0.0003484180197119713, 'timing/model': 1.1132480070227757, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 1.39}) +[2026-06-24 02:59:57,218][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.06844791024923325, 'train/grad_norm_pre_clip': 18.854955673217773, 'timing/data': 0.00036595796700567007, 'timing/model': 1.121149840997532, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 1.39}) +[2026-06-24 03:00:01,301][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.00011396170884836465, 'train/grad_norm_pre_clip': 18.153894424438477, 'timing/data': 0.00034397607669234276, 'timing/model': 1.1168515790486708, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 1.39}) +[2026-06-24 03:00:05,392][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.022978251799941063, 'train/grad_norm_pre_clip': 43.63777542114258, 'timing/data': 0.0003468090435490012, 'timing/model': 1.1190206019673496, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 1.39}) +[2026-06-24 03:00:09,567][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.04265856370329857, 'train/grad_norm_pre_clip': 43.72889709472656, 'timing/data': 0.0003607819089666009, 'timing/model': 1.2091889450093731, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 1.39}) +[2026-06-24 03:00:13,642][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.012987825088202953, 'train/grad_norm_pre_clip': 23.01546859741211, 'timing/data': 0.0003668579738587141, 'timing/model': 1.1181644860189408, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 1.39}) +[2026-06-24 03:00:17,725][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.0023849480785429478, 'train/grad_norm_pre_clip': 25.54311752319336, 'timing/data': 0.000352279981598258, 'timing/model': 1.1165483379736543, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 1.39}) +[2026-06-24 03:00:21,810][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.015469835139811039, 'train/grad_norm_pre_clip': 22.26547622680664, 'timing/data': 0.00035534193739295006, 'timing/model': 1.1195078289601952, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 1.39}) +[2026-06-24 03:00:26,065][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.0380239337682724, 'train/grad_norm_pre_clip': 24.167842864990234, 'timing/data': 0.00035169697366654873, 'timing/model': 1.2919503189623356, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 1.39}) +[2026-06-24 03:00:30,144][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3600 +[2026-06-24 03:01:33,867][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3600 +[2026-06-24 03:01:33,868][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3600 +[2026-06-24 03:02:12,951][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3600 +[2026-06-24 03:02:12,952][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.01572752743959427, 'train/grad_norm_pre_clip': 5.603480815887451, 'eval/action_loss/seconds': 63.72011778107844, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.030185962095856667, 'eval/latency_2/loss': 0.030185962095856667, 'eval/flappy/loss': 0.030185962095856667, 'eval/flappy/latency_2/loss': 0.030185962095856667, 'eval/flap_precision': 0.9795640326975477, 'eval/flap_recall': 0.9755766621438263, 'eval/flap_f1': 0.9775662814411965, 'eval/flap_rate_model': 0.2038888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9775662814411965, 'eval/latency_2/flap_precision': 0.9795640326975477, 'eval/latency_2/flap_recall': 0.9755766621438263, 'eval/latency_2/flap_f1': 0.9775662814411965, 'eval/latency_2/flap_rate_model': 0.2038888888888889, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9775662814411965, 'eval/action_classification/data_wait_seconds': 0.4741913974285126, 'eval/action_classification/predict_seconds': 38.47163391113281, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00035067799035459757, 'timing/model': 1.1191767619457096, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 1.39}) +[2026-06-24 03:02:17,004][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.15133658051490784, 'train/grad_norm_pre_clip': 23.932344436645508, 'timing/data': 0.0004014789592474699, 'timing/model': 1.1067120949737728, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 1.39}) +[2026-06-24 03:02:21,131][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.0426972322165966, 'train/grad_norm_pre_clip': 10.627330780029297, 'timing/data': 0.00034760404378175735, 'timing/model': 1.1915118579054251, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 1.39}) +[2026-06-24 03:02:25,200][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.022809697315096855, 'train/grad_norm_pre_clip': 16.747119903564453, 'timing/data': 0.0005414190236479044, 'timing/model': 1.1116338570136577, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 1.39}) +[2026-06-24 03:02:29,262][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.008982852101325989, 'train/grad_norm_pre_clip': 15.55689811706543, 'timing/data': 0.0003497679717838764, 'timing/model': 1.1122736539691687, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 1.39}) +[2026-06-24 03:02:33,311][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.011346935294568539, 'train/grad_norm_pre_clip': 32.77642059326172, 'timing/data': 0.0005509669426828623, 'timing/model': 1.109307463048026, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 1.4}) +[2026-06-24 03:02:37,442][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.018630946055054665, 'train/grad_norm_pre_clip': 28.769855499267578, 'timing/data': 0.0003564499784260988, 'timing/model': 1.1819786940468475, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 1.4}) +[2026-06-24 03:02:41,507][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.05773055553436279, 'train/grad_norm_pre_clip': 16.30651092529297, 'timing/data': 0.0003545109648257494, 'timing/model': 1.119060960947536, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 1.4}) +[2026-06-24 03:02:45,563][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.04831596463918686, 'train/grad_norm_pre_clip': 65.71023559570312, 'timing/data': 0.00035695405676960945, 'timing/model': 1.11122923810035, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 1.4}) +[2026-06-24 03:02:49,619][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.063692606985569, 'train/grad_norm_pre_clip': 38.599510192871094, 'timing/data': 0.00039932201616466045, 'timing/model': 1.1174165710108355, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 1.4}) +[2026-06-24 03:02:53,801][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.03914293274283409, 'train/grad_norm_pre_clip': 12.176591873168945, 'timing/data': 0.0004021379863843322, 'timing/model': 1.2339194579981267, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 1.4}) +[2026-06-24 03:02:57,929][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.08718423545360565, 'train/grad_norm_pre_clip': 29.50286865234375, 'timing/data': 0.0004090940346941352, 'timing/model': 1.114159837947227, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 1.4}) +[2026-06-24 03:03:02,024][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.1975240260362625, 'train/grad_norm_pre_clip': 35.9005126953125, 'timing/data': 0.0005356919718906283, 'timing/model': 1.1132056560600176, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 1.4}) +[2026-06-24 03:03:06,095][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.05267000570893288, 'train/grad_norm_pre_clip': 24.594806671142578, 'timing/data': 0.0005439560627564788, 'timing/model': 1.112601184984669, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 1.4}) +[2026-06-24 03:03:10,280][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.00035330100217834115, 'train/grad_norm_pre_clip': 20.2779483795166, 'timing/data': 0.0003488429356366396, 'timing/model': 1.1157994760433212, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 1.4}) +[2026-06-24 03:03:14,362][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.0697382465004921, 'train/grad_norm_pre_clip': 28.8869686126709, 'timing/data': 0.000354639021679759, 'timing/model': 1.1233412439469248, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 1.4}) +[2026-06-24 03:03:18,459][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.08357887715101242, 'train/grad_norm_pre_clip': 9.417020797729492, 'timing/data': 0.00035184796433895826, 'timing/model': 1.1194285369710997, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 1.4}) +[2026-06-24 03:03:22,538][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.004115832969546318, 'train/grad_norm_pre_clip': 5.969912528991699, 'timing/data': 0.0003582729259505868, 'timing/model': 1.1178799729095772, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 1.4}) +[2026-06-24 03:03:26,693][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.12986092269420624, 'train/grad_norm_pre_clip': 19.084753036499023, 'timing/data': 0.00040330609772354364, 'timing/model': 1.1149577539181337, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 1.4}) +[2026-06-24 03:03:30,769][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.0004030584532301873, 'train/grad_norm_pre_clip': 13.016239166259766, 'timing/data': 0.0003467859933152795, 'timing/model': 1.1145646260119975, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 1.4}) +[2026-06-24 03:03:34,847][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.023740269243717194, 'train/grad_norm_pre_clip': 19.93320655822754, 'timing/data': 0.0005458039231598377, 'timing/model': 1.126392327947542, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 1.4}) +[2026-06-24 03:03:38,939][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.04033348336815834, 'train/grad_norm_pre_clip': 4.549422740936279, 'timing/data': 0.0003904589684680104, 'timing/model': 1.1208493200829253, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 1.4}) +[2026-06-24 03:03:43,182][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.017070049419999123, 'train/grad_norm_pre_clip': 14.468148231506348, 'timing/data': 0.0004844670183956623, 'timing/model': 1.113245996995829, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 1.4}) +[2026-06-24 03:03:47,260][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.004597853869199753, 'train/grad_norm_pre_clip': 20.00626564025879, 'timing/data': 0.0003584859659895301, 'timing/model': 1.120339444023557, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 1.4}) +[2026-06-24 03:03:51,334][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.000271715281996876, 'train/grad_norm_pre_clip': 13.168025016784668, 'timing/data': 0.000504649942740798, 'timing/model': 1.1170386859448627, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 1.4}) +[2026-06-24 03:03:55,535][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.0005909762112423778, 'train/grad_norm_pre_clip': 26.1365966796875, 'timing/data': 0.0003512679832056165, 'timing/model': 1.1185189590323716, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 1.4}) +[2026-06-24 03:03:59,691][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.0018486033659428358, 'train/grad_norm_pre_clip': 10.764305114746094, 'timing/data': 0.0003579129697754979, 'timing/model': 1.1133022489957511, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 1.4}) +[2026-06-24 03:04:03,764][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.006256729364395142, 'train/grad_norm_pre_clip': 21.640159606933594, 'timing/data': 0.0003511219983920455, 'timing/model': 1.1133520080475137, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 1.4}) +[2026-06-24 03:04:07,864][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.06404364854097366, 'train/grad_norm_pre_clip': 14.863165855407715, 'timing/data': 0.000567088951356709, 'timing/model': 1.1209249260136858, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.4}) +[2026-06-24 03:04:11,954][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.0017529422184452415, 'train/grad_norm_pre_clip': 25.608896255493164, 'timing/data': 0.0003451160155236721, 'timing/model': 1.1172016080236062, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 1.4}) +[2026-06-24 03:04:16,175][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.008970318362116814, 'train/grad_norm_pre_clip': 8.15117073059082, 'timing/data': 0.0004179510287940502, 'timing/model': 1.2398596120765433, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 1.4}) +[2026-06-24 03:04:20,264][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.00039976683910936117, 'train/grad_norm_pre_clip': 19.136871337890625, 'timing/data': 0.00036454806104302406, 'timing/model': 1.122691879980266, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 1.41}) +[2026-06-24 03:04:24,348][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.06315789371728897, 'train/grad_norm_pre_clip': 27.86458969116211, 'timing/data': 0.0005604380276054144, 'timing/model': 1.1183477189624682, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 1.41}) +[2026-06-24 03:04:28,446][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.01612965576350689, 'train/grad_norm_pre_clip': 16.458919525146484, 'timing/data': 0.0004074180033057928, 'timing/model': 1.1354857799597085, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 1.41}) +[2026-06-24 03:04:32,530][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.003952539060264826, 'train/grad_norm_pre_clip': 16.0434627532959, 'timing/data': 0.0003380129346624017, 'timing/model': 1.119821373024024, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 1.41}) +[2026-06-24 03:04:36,782][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.184931680560112, 'train/grad_norm_pre_clip': 55.2106819152832, 'timing/data': 0.00036313000600785017, 'timing/model': 1.1212755659362301, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 1.41}) +[2026-06-24 03:04:40,849][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.09321029484272003, 'train/grad_norm_pre_clip': 19.67523956298828, 'timing/data': 0.00034786900505423546, 'timing/model': 1.1119575479533523, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 1.41}) +[2026-06-24 03:04:44,923][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.12300471216440201, 'train/grad_norm_pre_clip': 30.0094051361084, 'timing/data': 0.00036685599479824305, 'timing/model': 1.1115501900203526, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 1.41}) +[2026-06-24 03:04:49,011][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.0026196304243057966, 'train/grad_norm_pre_clip': 14.297938346862793, 'timing/data': 0.00034968997351825237, 'timing/model': 1.1216225700918585, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 1.41}) +[2026-06-24 03:04:53,156][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.002781305694952607, 'train/grad_norm_pre_clip': 8.817262649536133, 'timing/data': 0.00034934200812131166, 'timing/model': 1.1168368969811127, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 1.41}) +[2026-06-24 03:04:57,242][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.068141870200634, 'train/grad_norm_pre_clip': 16.93289566040039, 'timing/data': 0.0004994820337742567, 'timing/model': 1.123298292979598, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 1.41}) +[2026-06-24 03:05:01,313][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.0032766927033662796, 'train/grad_norm_pre_clip': 10.050291061401367, 'timing/data': 0.00036365806590765715, 'timing/model': 1.1109225990949199, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 1.41}) +[2026-06-24 03:05:05,400][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.030952295288443565, 'train/grad_norm_pre_clip': 28.063222885131836, 'timing/data': 0.0003571180859580636, 'timing/model': 1.1197010030737147, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 1.41}) +[2026-06-24 03:05:09,557][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.009402298368513584, 'train/grad_norm_pre_clip': 5.51192569732666, 'timing/data': 0.00037074496503919363, 'timing/model': 1.1196453570155427, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 1.41}) +[2026-06-24 03:05:13,645][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.009070364758372307, 'train/grad_norm_pre_clip': 26.970684051513672, 'timing/data': 0.0003529270179569721, 'timing/model': 1.1166833499446511, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 1.41}) +[2026-06-24 03:05:17,723][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.010402059182524681, 'train/grad_norm_pre_clip': 6.749579429626465, 'timing/data': 0.0003625869285315275, 'timing/model': 1.1181878170464188, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 1.41}) +[2026-06-24 03:05:21,803][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.011050231754779816, 'train/grad_norm_pre_clip': 14.574100494384766, 'timing/data': 0.0003611990250647068, 'timing/model': 1.117030835011974, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 1.41}) +[2026-06-24 03:05:25,987][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.06827469915151596, 'train/grad_norm_pre_clip': 25.070415496826172, 'timing/data': 0.00036564096808433533, 'timing/model': 1.136064681923017, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 1.41}) +[2026-06-24 03:05:30,073][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.027990180999040604, 'train/grad_norm_pre_clip': 20.361265182495117, 'timing/data': 0.0003672470338642597, 'timing/model': 1.1153587810695171, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 1.41}) +[2026-06-24 03:05:34,177][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.14488783478736877, 'train/grad_norm_pre_clip': 25.586942672729492, 'timing/data': 0.0004961499944329262, 'timing/model': 1.1360988380620256, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 1.41}) +[2026-06-24 03:05:38,262][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.008382399566471577, 'train/grad_norm_pre_clip': 16.093786239624023, 'timing/data': 0.0003545560175552964, 'timing/model': 1.1231823459966108, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 1.41}) +[2026-06-24 03:05:42,488][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.009244933724403381, 'train/grad_norm_pre_clip': 21.026033401489258, 'timing/data': 0.000356098054908216, 'timing/model': 1.1165105539839715, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 1.41}) +[2026-06-24 03:05:46,575][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.00019763891759794205, 'train/grad_norm_pre_clip': 31.37716293334961, 'timing/data': 0.00036063394509255886, 'timing/model': 1.1243311249418184, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 1.41}) +[2026-06-24 03:05:50,658][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.1752261370420456, 'train/grad_norm_pre_clip': 41.19040298461914, 'timing/data': 0.00048615504056215286, 'timing/model': 1.117537969024852, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 1.41}) +[2026-06-24 03:05:54,828][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.002064427360892296, 'train/grad_norm_pre_clip': 26.60581398010254, 'timing/data': 0.0004176539368927479, 'timing/model': 1.2062052409164608, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 1.41}) +[2026-06-24 03:05:58,904][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.10562029480934143, 'train/grad_norm_pre_clip': 21.25066566467285, 'timing/data': 0.00036237307358533144, 'timing/model': 1.1207156940363348, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 1.41}) +[2026-06-24 03:06:02,989][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.013882177881896496, 'train/grad_norm_pre_clip': 28.041324615478516, 'timing/data': 0.0006092769326642156, 'timing/model': 1.1236656220862642, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 1.41}) +[2026-06-24 03:06:07,071][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.0570998452603817, 'train/grad_norm_pre_clip': 22.02277183532715, 'timing/data': 0.00035452295560389757, 'timing/model': 1.1201981210615486, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 1.42}) +[2026-06-24 03:06:11,323][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.026348672807216644, 'train/grad_norm_pre_clip': 7.704351902008057, 'timing/data': 0.0003633060259744525, 'timing/model': 1.282938656047918, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 1.42}) +[2026-06-24 03:06:15,400][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.03198869153857231, 'train/grad_norm_pre_clip': 13.341814994812012, 'timing/data': 0.0004305649781599641, 'timing/model': 1.1154858960071579, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 1.42}) +[2026-06-24 03:06:19,489][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.002629712224006653, 'train/grad_norm_pre_clip': 30.60719108581543, 'timing/data': 0.00035519502125680447, 'timing/model': 1.1176940840668976, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 1.42}) +[2026-06-24 03:06:23,577][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.0028242904227226973, 'train/grad_norm_pre_clip': 25.48430633544922, 'timing/data': 0.00036289298441261053, 'timing/model': 1.1205370710231364, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 1.42}) +[2026-06-24 03:06:27,729][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 6.236443732632324e-05, 'train/grad_norm_pre_clip': 27.942590713500977, 'timing/data': 0.0005394400795921683, 'timing/model': 1.1200931679923087, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 1.42}) +[2026-06-24 03:06:31,811][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.07865607738494873, 'train/grad_norm_pre_clip': 29.41535758972168, 'timing/data': 0.0003538760356605053, 'timing/model': 1.1190645570168272, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 1.42}) +[2026-06-24 03:06:35,881][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.013650205917656422, 'train/grad_norm_pre_clip': 16.46569061279297, 'timing/data': 0.000474989996291697, 'timing/model': 1.113606909988448, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 1.42}) +[2026-06-24 03:06:39,979][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.007170350756496191, 'train/grad_norm_pre_clip': 26.9742488861084, 'timing/data': 0.00036586797796189785, 'timing/model': 1.1191432219929993, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 1.42}) +[2026-06-24 03:06:44,133][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.11834845691919327, 'train/grad_norm_pre_clip': 36.133583068847656, 'timing/data': 0.000419907970353961, 'timing/model': 1.1149855859111995, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 1.42}) +[2026-06-24 03:06:48,213][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.05579357594251633, 'train/grad_norm_pre_clip': 34.568824768066406, 'timing/data': 0.0003696760395541787, 'timing/model': 1.1161989059764892, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 1.42}) +[2026-06-24 03:06:52,290][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.022536227479577065, 'train/grad_norm_pre_clip': 23.602420806884766, 'timing/data': 0.00037952302955091, 'timing/model': 1.114236488007009, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 1.42}) +[2026-06-24 03:06:56,380][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.018575232475996017, 'train/grad_norm_pre_clip': 25.8582820892334, 'timing/data': 0.00040536897722631693, 'timing/model': 1.1237384059932083, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 1.42}) +[2026-06-24 03:07:00,604][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.04575534537434578, 'train/grad_norm_pre_clip': 19.249317169189453, 'timing/data': 0.0003619429189711809, 'timing/model': 1.1203457249794155, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 1.42}) +[2026-06-24 03:07:04,702][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.07344907522201538, 'train/grad_norm_pre_clip': 35.53656005859375, 'timing/data': 0.0003661059308797121, 'timing/model': 1.1236217920668423, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 1.42}) +[2026-06-24 03:07:08,788][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.014577222056686878, 'train/grad_norm_pre_clip': 8.758033752441406, 'timing/data': 0.0005814259639009833, 'timing/model': 1.1180362519808114, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 1.42}) +[2026-06-24 03:07:12,876][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.024377556517720222, 'train/grad_norm_pre_clip': 30.274938583374023, 'timing/data': 0.0003560830373317003, 'timing/model': 1.1239176049130037, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 1.42}) +[2026-06-24 03:07:17,070][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.0030812015756964684, 'train/grad_norm_pre_clip': 14.715523719787598, 'timing/data': 0.0003671799786388874, 'timing/model': 1.121338116005063, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 1.42}) +[2026-06-24 03:07:21,154][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.02101827785372734, 'train/grad_norm_pre_clip': 9.645812034606934, 'timing/data': 0.0003577430034056306, 'timing/model': 1.1206479780375957, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 1.42}) +[2026-06-24 03:07:25,230][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.10548388212919235, 'train/grad_norm_pre_clip': 36.0174446105957, 'timing/data': 0.0004203609423711896, 'timing/model': 1.1180575239704922, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 1.42}) +[2026-06-24 03:07:29,317][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.01049548014998436, 'train/grad_norm_pre_clip': 20.946531295776367, 'timing/data': 0.0005634080152958632, 'timing/model': 1.1150547449942678, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.42}) +[2026-06-24 03:07:33,475][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.0646701455116272, 'train/grad_norm_pre_clip': 34.77546691894531, 'timing/data': 0.0003751430194824934, 'timing/model': 1.1153846960514784, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 1.42}) +[2026-06-24 03:07:37,547][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.013195118866860867, 'train/grad_norm_pre_clip': 16.330020904541016, 'timing/data': 0.0005960100097581744, 'timing/model': 1.113986762939021, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 1.42}) +[2026-06-24 03:07:41,627][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.046937841922044754, 'train/grad_norm_pre_clip': 27.058399200439453, 'timing/data': 0.0003685590345412493, 'timing/model': 1.1177138219354674, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 1.42}) +[2026-06-24 03:07:45,707][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.00011424720287322998, 'train/grad_norm_pre_clip': 13.131046295166016, 'timing/data': 0.00045389006845653057, 'timing/model': 1.119416355038993, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 1.42}) +[2026-06-24 03:07:49,980][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.1846458464860916, 'train/grad_norm_pre_clip': 54.28355407714844, 'timing/data': 0.0005500330589711666, 'timing/model': 1.1216038600541651, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 1.43}) +[2026-06-24 03:07:54,061][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.006271205376833677, 'train/grad_norm_pre_clip': 10.062500953674316, 'timing/data': 0.00035543402191251516, 'timing/model': 1.1216290399897844, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 1.43}) +[2026-06-24 03:07:58,146][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.010810879059135914, 'train/grad_norm_pre_clip': 44.533138275146484, 'timing/data': 0.0007642840500921011, 'timing/model': 1.1184109899913892, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 1.43}) +[2026-06-24 03:08:02,236][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.016281260177493095, 'train/grad_norm_pre_clip': 15.537397384643555, 'timing/data': 0.0003773169592022896, 'timing/model': 1.1204723209375516, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 1.43}) +[2026-06-24 03:08:06,422][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.1324385404586792, 'train/grad_norm_pre_clip': 26.500507354736328, 'timing/data': 0.0003677560016512871, 'timing/model': 1.1198233530158177, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 1.43}) +[2026-06-24 03:08:10,514][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.04571974277496338, 'train/grad_norm_pre_clip': 20.685836791992188, 'timing/data': 0.0003626019461080432, 'timing/model': 1.1202268180204555, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 1.43}) +[2026-06-24 03:08:14,596][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.06505031883716583, 'train/grad_norm_pre_clip': 18.717376708984375, 'timing/data': 0.00037093099672347307, 'timing/model': 1.117409759084694, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 1.43}) +[2026-06-24 03:08:18,676][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.008517703041434288, 'train/grad_norm_pre_clip': 13.740080833435059, 'timing/data': 0.0003609039122238755, 'timing/model': 1.1189569340785965, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 1.43}) +[2026-06-24 03:08:22,947][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.00801176205277443, 'train/grad_norm_pre_clip': 25.84517478942871, 'timing/data': 0.00035815604496747255, 'timing/model': 1.1215072760824114, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 1.43}) +[2026-06-24 03:08:27,049][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.0029599296394735575, 'train/grad_norm_pre_clip': 4.115393161773682, 'timing/data': 0.00042225897777825594, 'timing/model': 1.1185285039246082, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 1.43}) +[2026-06-24 03:08:31,156][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.014727726578712463, 'train/grad_norm_pre_clip': 31.53789710998535, 'timing/data': 0.00036025105509907007, 'timing/model': 1.1294335360871628, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 1.43}) +[2026-06-24 03:08:35,255][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.011954939924180508, 'train/grad_norm_pre_clip': 22.61467742919922, 'timing/data': 0.000364540028385818, 'timing/model': 1.1203297500032932, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 1.43}) +[2026-06-24 03:08:39,415][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.06685855984687805, 'train/grad_norm_pre_clip': 23.915145874023438, 'timing/data': 0.0005025840364396572, 'timing/model': 1.1219698829809204, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 1.43}) +[2026-06-24 03:08:43,503][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.000594614481087774, 'train/grad_norm_pre_clip': 11.463738441467285, 'timing/data': 0.0004032059805467725, 'timing/model': 1.127906194073148, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 1.43}) +[2026-06-24 03:08:47,584][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.08777658641338348, 'train/grad_norm_pre_clip': 27.007644653320312, 'timing/data': 0.0003843529848381877, 'timing/model': 1.1174709630431607, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 1.43}) +[2026-06-24 03:08:51,666][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.02495717629790306, 'train/grad_norm_pre_clip': 21.87326431274414, 'timing/data': 0.0005623629549518228, 'timing/model': 1.1220296169631183, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 1.43}) +[2026-06-24 03:08:55,830][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.016648750752210617, 'train/grad_norm_pre_clip': 21.90843391418457, 'timing/data': 0.0003608420956879854, 'timing/model': 1.122988290968351, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 1.43}) +[2026-06-24 03:08:59,923][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.14965549111366272, 'train/grad_norm_pre_clip': 24.831256866455078, 'timing/data': 0.0005960880080237985, 'timing/model': 1.1174573129974306, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 1.43}) +[2026-06-24 03:09:04,012][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3700 +[2026-06-24 03:10:08,660][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3700 +[2026-06-24 03:10:08,662][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3700 +[2026-06-24 03:10:47,288][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3700 +[2026-06-24 03:10:47,289][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.05240115895867348, 'train/grad_norm_pre_clip': 17.537513732910156, 'eval/action_loss/seconds': 64.64615532697644, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.030633969232439995, 'eval/latency_2/loss': 0.030633969232439995, 'eval/flappy/loss': 0.030633969232439995, 'eval/flappy/latency_2/loss': 0.030633969232439995, 'eval/flap_precision': 0.9794520547945206, 'eval/flap_recall': 0.9701492537313433, 'eval/flap_f1': 0.9747784594410361, 'eval/flap_rate_model': 0.20277777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.974778459441036, 'eval/latency_2/flap_precision': 0.9794520547945206, 'eval/latency_2/flap_recall': 0.9701492537313433, 'eval/latency_2/flap_f1': 0.9747784594410361, 'eval/latency_2/flap_rate_model': 0.20277777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.974778459441036, 'eval/action_classification/data_wait_seconds': 0.5496376156806946, 'eval/action_classification/predict_seconds': 37.92745590209961, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0004839950706809759, 'timing/model': 1.1206534850643948, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 1.43}) +[2026-06-24 03:10:51,420][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.0623079352080822, 'train/grad_norm_pre_clip': 13.503945350646973, 'timing/data': 0.0003669189754873514, 'timing/model': 1.1135127570014447, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 1.43}) +[2026-06-24 03:10:55,483][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.0002442820987198502, 'train/grad_norm_pre_clip': 19.233665466308594, 'timing/data': 0.0004160220269113779, 'timing/model': 1.1126073880586773, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 1.43}) +[2026-06-24 03:10:59,525][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.02490629442036152, 'train/grad_norm_pre_clip': 10.80527400970459, 'timing/data': 0.00035670597571879625, 'timing/model': 1.1091696420917287, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 1.43}) +[2026-06-24 03:11:03,582][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.0212194062769413, 'train/grad_norm_pre_clip': 26.12688446044922, 'timing/data': 0.0003743330016732216, 'timing/model': 1.1112144680228084, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 1.43}) +[2026-06-24 03:11:07,797][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.041545264422893524, 'train/grad_norm_pre_clip': 62.22557830810547, 'timing/data': 0.0003593090223148465, 'timing/model': 1.1108518870314583, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 1.43}) +[2026-06-24 03:11:11,860][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.0033736906480044127, 'train/grad_norm_pre_clip': 11.047714233398438, 'timing/data': 0.0003850549692288041, 'timing/model': 1.1090174040291458, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 1.43}) +[2026-06-24 03:11:15,926][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.028650235384702682, 'train/grad_norm_pre_clip': 18.5085506439209, 'timing/data': 0.0005702840862795711, 'timing/model': 1.1148255299776793, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 1.43}) +[2026-06-24 03:11:19,992][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.026630163192749023, 'train/grad_norm_pre_clip': 29.111936569213867, 'timing/data': 0.0005369429709389806, 'timing/model': 1.115722340065986, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 1.44}) +[2026-06-24 03:11:24,224][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.004200630355626345, 'train/grad_norm_pre_clip': 3.644878387451172, 'timing/data': 0.0004184790886938572, 'timing/model': 1.1164118730230257, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 1.44}) +[2026-06-24 03:11:28,290][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.026849204674363136, 'train/grad_norm_pre_clip': 10.789076805114746, 'timing/data': 0.0005606639897450805, 'timing/model': 1.1130325929261744, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 1.44}) +[2026-06-24 03:11:32,363][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.09281551837921143, 'train/grad_norm_pre_clip': 12.671792984008789, 'timing/data': 0.0003505790373310447, 'timing/model': 1.1137044969946146, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 1.44}) +[2026-06-24 03:11:36,449][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.008780097588896751, 'train/grad_norm_pre_clip': 9.643071174621582, 'timing/data': 0.00042034091893583536, 'timing/model': 1.1220425589708611, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 1.44}) +[2026-06-24 03:11:40,717][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.02988283708691597, 'train/grad_norm_pre_clip': 24.231121063232422, 'timing/data': 0.0003521139733493328, 'timing/model': 1.1219374999636784, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 1.44}) +[2026-06-24 03:11:44,778][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.00042507945909164846, 'train/grad_norm_pre_clip': 31.380573272705078, 'timing/data': 0.00035841099452227354, 'timing/model': 1.1163305609952658, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 1.44}) +[2026-06-24 03:11:48,852][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.09431027621030807, 'train/grad_norm_pre_clip': 38.93404006958008, 'timing/data': 0.00040353590156883, 'timing/model': 1.1148174969712272, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 1.44}) +[2026-06-24 03:11:52,927][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.05362638831138611, 'train/grad_norm_pre_clip': 24.60711097717285, 'timing/data': 0.00041910295840352774, 'timing/model': 1.1163223970215768, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 1.44}) +[2026-06-24 03:11:57,195][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.015550175681710243, 'train/grad_norm_pre_clip': 14.886640548706055, 'timing/data': 0.0005552360089495778, 'timing/model': 1.1166252749972045, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 1.44}) +[2026-06-24 03:12:01,262][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.04406588897109032, 'train/grad_norm_pre_clip': 12.412213325500488, 'timing/data': 0.000375372008420527, 'timing/model': 1.1102721369825304, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 1.44}) +[2026-06-24 03:12:05,352][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.00025220520910806954, 'train/grad_norm_pre_clip': 11.480646133422852, 'timing/data': 0.0006618889747187495, 'timing/model': 1.1223927680402994, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 1.44}) +[2026-06-24 03:12:09,458][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.10149820148944855, 'train/grad_norm_pre_clip': 17.83000373840332, 'timing/data': 0.0003578420728445053, 'timing/model': 1.1337127020815387, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 1.44}) +[2026-06-24 03:12:13,700][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.0006680303486064076, 'train/grad_norm_pre_clip': 5.167395114898682, 'timing/data': 0.00041762494947761297, 'timing/model': 1.1174821680178866, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 1.44}) +[2026-06-24 03:12:17,759][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.24013791978359222, 'train/grad_norm_pre_clip': 46.52028274536133, 'timing/data': 0.0003567429957911372, 'timing/model': 1.1158290778985247, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 1.44}) +[2026-06-24 03:12:21,831][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.08004609495401382, 'train/grad_norm_pre_clip': 23.95465850830078, 'timing/data': 0.0004182240227237344, 'timing/model': 1.116507831029594, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 1.44}) +[2026-06-24 03:12:25,910][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.04282096400856972, 'train/grad_norm_pre_clip': 36.536983489990234, 'timing/data': 0.00038962496910244226, 'timing/model': 1.1155516229337081, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 1.44}) +[2026-06-24 03:12:30,170][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.0033489835914224386, 'train/grad_norm_pre_clip': 8.066596984863281, 'timing/data': 0.0004732890520244837, 'timing/model': 1.1199984949780628, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 1.44}) +[2026-06-24 03:12:34,253][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.0020332373678684235, 'train/grad_norm_pre_clip': 2.7652649879455566, 'timing/data': 0.0004090180154889822, 'timing/model': 1.1134029449895024, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 1.44}) +[2026-06-24 03:12:38,326][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.13951484858989716, 'train/grad_norm_pre_clip': 26.433258056640625, 'timing/data': 0.0004003179492428899, 'timing/model': 1.1186128080589697, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 1.44}) +[2026-06-24 03:12:42,417][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.07030309736728668, 'train/grad_norm_pre_clip': 28.72003173828125, 'timing/data': 0.0005460490938276052, 'timing/model': 1.1190513350302354, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 1.44}) +[2026-06-24 03:12:46,668][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.005307109095156193, 'train/grad_norm_pre_clip': 18.90526580810547, 'timing/data': 0.000527380034327507, 'timing/model': 1.1173019770067185, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 1.44}) +[2026-06-24 03:12:50,751][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.022474970668554306, 'train/grad_norm_pre_clip': 15.933778762817383, 'timing/data': 0.0004086539847776294, 'timing/model': 1.1169397609774023, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 1.44}) +[2026-06-24 03:12:54,831][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.12154312431812286, 'train/grad_norm_pre_clip': 34.51877212524414, 'timing/data': 0.00040654302574694157, 'timing/model': 1.1179949480574578, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 1.44}) +[2026-06-24 03:12:58,908][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 4.075278775417246e-05, 'train/grad_norm_pre_clip': 8.260601997375488, 'timing/data': 0.00036264408845454454, 'timing/model': 1.113416766980663, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 1.44}) +[2026-06-24 03:13:03,160][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.017042705789208412, 'train/grad_norm_pre_clip': 28.7139892578125, 'timing/data': 0.00036608101800084114, 'timing/model': 1.1163309059338644, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 1.44}) +[2026-06-24 03:13:07,255][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.08913186192512512, 'train/grad_norm_pre_clip': 26.837467193603516, 'timing/data': 0.0003603990189731121, 'timing/model': 1.122437778976746, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 1.45}) +[2026-06-24 03:13:11,339][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.08788596838712692, 'train/grad_norm_pre_clip': 25.28292465209961, 'timing/data': 0.0003634559689089656, 'timing/model': 1.1186509639956057, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 1.45}) +[2026-06-24 03:13:15,440][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.012442074716091156, 'train/grad_norm_pre_clip': 30.76041030883789, 'timing/data': 0.000566495000384748, 'timing/model': 1.1175346430391073, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 1.45}) +[2026-06-24 03:13:19,713][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.1465711146593094, 'train/grad_norm_pre_clip': 44.5914421081543, 'timing/data': 0.0003609240520745516, 'timing/model': 1.1162840790348127, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 1.45}) +[2026-06-24 03:13:23,810][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.10044611245393753, 'train/grad_norm_pre_clip': 18.797199249267578, 'timing/data': 0.0003549850080162287, 'timing/model': 1.1216825599549338, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 1.45}) +[2026-06-24 03:13:27,913][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.046031489968299866, 'train/grad_norm_pre_clip': 50.87084197998047, 'timing/data': 0.0006964210188016295, 'timing/model': 1.1338216740405187, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 1.45}) +[2026-06-24 03:13:32,004][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.18718405067920685, 'train/grad_norm_pre_clip': 39.95752716064453, 'timing/data': 0.0005744110094383359, 'timing/model': 1.1250073179835454, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 1.45}) +[2026-06-24 03:13:36,240][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.058593735098838806, 'train/grad_norm_pre_clip': 25.89323616027832, 'timing/data': 0.00048284605145454407, 'timing/model': 1.1184696321142837, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 1.45}) +[2026-06-24 03:13:40,307][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.07909876853227615, 'train/grad_norm_pre_clip': 21.03647804260254, 'timing/data': 0.00041925592813640833, 'timing/model': 1.1092502489918843, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 1.45}) +[2026-06-24 03:13:44,394][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.042559292167425156, 'train/grad_norm_pre_clip': 20.9619197845459, 'timing/data': 0.00036355305928736925, 'timing/model': 1.1174236229853705, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 1.45}) +[2026-06-24 03:13:48,486][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.0005140674766153097, 'train/grad_norm_pre_clip': 7.364171981811523, 'timing/data': 0.00037055392749607563, 'timing/model': 1.1155571850249544, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 1.45}) +[2026-06-24 03:13:52,745][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.002489432692527771, 'train/grad_norm_pre_clip': 16.811569213867188, 'timing/data': 0.0005604990292340517, 'timing/model': 1.1257737050764263, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 1.45}) +[2026-06-24 03:13:56,830][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.05555951967835426, 'train/grad_norm_pre_clip': 24.603416442871094, 'timing/data': 0.0003538610180839896, 'timing/model': 1.1178427500417456, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 1.45}) +[2026-06-24 03:14:00,904][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.01281232014298439, 'train/grad_norm_pre_clip': 22.282459259033203, 'timing/data': 0.00037264195270836353, 'timing/model': 1.1148026869632304, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 1.45}) +[2026-06-24 03:14:04,983][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.016994653269648552, 'train/grad_norm_pre_clip': 22.083070755004883, 'timing/data': 0.0003562320489436388, 'timing/model': 1.1243238929891959, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 1.45}) +[2026-06-24 03:14:09,243][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.0009456850821152329, 'train/grad_norm_pre_clip': 18.799171447753906, 'timing/data': 0.00036147201899439096, 'timing/model': 1.1191529460484162, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 1.45}) +[2026-06-24 03:26:06,591][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3750 +[2026-06-24 03:26:06,592][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.005235244520008564, 'train/grad_norm_pre_clip': 11.604633331298828, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 442.26000000002523, 'rl_games_eval/mid_train/mean_length': 3582.4, 'rl_games_eval/mid_train/std_reward': 4.680000000000791, 'rl_games_eval/mid_train/std_length': 35.2, 'rl_games_eval/mid_train/macro_mean_reward': 442.26000000002523, 'rl_games_eval/mid_train/macro_mean_length': 3582.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 442.26000000002523, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3582.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 4.680000000000791, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 35.2, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 442.26000000002523, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003546440275385976, 'timing/model': 1.1150155420182273, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 1.45}) +[2026-06-24 03:26:10,692][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.0010166780557483435, 'train/grad_norm_pre_clip': 4.254634857177734, 'timing/data': 0.0004404130158945918, 'timing/model': 1.0984635209897533, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 1.45}) +[2026-06-24 03:26:14,687][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.0019153464818373322, 'train/grad_norm_pre_clip': 32.292884826660156, 'timing/data': 0.00042832200415432453, 'timing/model': 1.098827934009023, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 1.45}) +[2026-06-24 03:26:18,847][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.14443786442279816, 'train/grad_norm_pre_clip': 57.912471771240234, 'timing/data': 0.0005742869107052684, 'timing/model': 1.0969542720122263, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 1.45}) +[2026-06-24 03:26:22,844][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 7.995174382813275e-05, 'train/grad_norm_pre_clip': 12.97541332244873, 'timing/data': 0.00039426598232239485, 'timing/model': 1.0965758570237085, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 1.45}) +[2026-06-24 03:26:26,858][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.029285568743944168, 'train/grad_norm_pre_clip': 20.286827087402344, 'timing/data': 0.0006039399886503816, 'timing/model': 1.103059503948316, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 1.45}) +[2026-06-24 03:26:30,954][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.005723176524043083, 'train/grad_norm_pre_clip': 13.450605392456055, 'timing/data': 0.0003709419397637248, 'timing/model': 1.183805375942029, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 1.45}) +[2026-06-24 03:26:34,969][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.0854533240199089, 'train/grad_norm_pre_clip': 24.300954818725586, 'timing/data': 0.00041633809451013803, 'timing/model': 1.098872039001435, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 1.45}) +[2026-06-24 03:26:39,032][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.030147874727845192, 'train/grad_norm_pre_clip': 19.449628829956055, 'timing/data': 0.00041897594928741455, 'timing/model': 1.1070644609862939, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 1.45}) +[2026-06-24 03:26:43,066][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.12008475512266159, 'train/grad_norm_pre_clip': 32.345008850097656, 'timing/data': 0.0003742149565368891, 'timing/model': 1.1058861509663984, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 1.45}) +[2026-06-24 03:26:47,206][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.03372644633054733, 'train/grad_norm_pre_clip': 36.45631408691406, 'timing/data': 0.0003785340813919902, 'timing/model': 1.2024726209929213, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 1.46}) +[2026-06-24 03:26:51,254][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.08249025046825409, 'train/grad_norm_pre_clip': 23.61971092224121, 'timing/data': 0.00035899493377655745, 'timing/model': 1.1083308099769056, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 1.46}) +[2026-06-24 03:26:55,306][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.03471467271447182, 'train/grad_norm_pre_clip': 30.650474548339844, 'timing/data': 0.0004277029074728489, 'timing/model': 1.1116238019894809, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 1.46}) +[2026-06-24 03:26:59,351][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.0036521751899272203, 'train/grad_norm_pre_clip': 6.3461809158325195, 'timing/data': 0.00036998698487877846, 'timing/model': 1.1066041380399838, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 1.46}) +[2026-06-24 03:27:03,552][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.01658380590379238, 'train/grad_norm_pre_clip': 16.185914993286133, 'timing/data': 0.0003591109998524189, 'timing/model': 1.2684403760358691, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 1.46}) +[2026-06-24 03:27:07,591][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.1534952074289322, 'train/grad_norm_pre_clip': 21.354455947875977, 'timing/data': 0.00040634709876030684, 'timing/model': 1.1045399069553241, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 1.46}) +[2026-06-24 03:27:11,645][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.07536769658327103, 'train/grad_norm_pre_clip': 46.585838317871094, 'timing/data': 0.0005242039915174246, 'timing/model': 1.11774923291523, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 1.46}) +[2026-06-24 03:27:15,703][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.001193094765767455, 'train/grad_norm_pre_clip': 24.730106353759766, 'timing/data': 0.0004092089366167784, 'timing/model': 1.1113571440801024, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 1.46}) +[2026-06-24 03:27:19,867][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.03839436545968056, 'train/grad_norm_pre_clip': 20.00436019897461, 'timing/data': 0.0004341740859672427, 'timing/model': 1.112313254037872, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 1.46}) +[2026-06-24 03:27:23,914][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.003740583313629031, 'train/grad_norm_pre_clip': 30.62305450439453, 'timing/data': 0.00036114989779889584, 'timing/model': 1.1110529840225354, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 1.46}) +[2026-06-24 03:27:27,974][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.022268321365118027, 'train/grad_norm_pre_clip': 15.423303604125977, 'timing/data': 0.0003921570023521781, 'timing/model': 1.114096584962681, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 1.46}) +[2026-06-24 03:27:32,051][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.012479267083108425, 'train/grad_norm_pre_clip': 9.483183860778809, 'timing/data': 0.00041253899689763784, 'timing/model': 1.115812962059863, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 1.46}) +[2026-06-24 03:27:36,282][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.03113522380590439, 'train/grad_norm_pre_clip': 13.647602081298828, 'timing/data': 0.0003594079753383994, 'timing/model': 1.1142690220149234, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 1.46}) +[2026-06-24 03:27:40,350][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.0062376768328249454, 'train/grad_norm_pre_clip': 14.337146759033203, 'timing/data': 0.0003650330472737551, 'timing/model': 1.122192715993151, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 1.46}) +[2026-06-24 03:27:44,418][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.03863228112459183, 'train/grad_norm_pre_clip': 17.888015747070312, 'timing/data': 0.000365662039257586, 'timing/model': 1.117207731003873, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 1.46}) +[2026-06-24 03:27:48,490][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.013082240708172321, 'train/grad_norm_pre_clip': 10.176694869995117, 'timing/data': 0.0003570480039343238, 'timing/model': 1.1126388770062476, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 1.46}) +[2026-06-24 03:27:52,731][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.00552254356443882, 'train/grad_norm_pre_clip': 10.74983024597168, 'timing/data': 0.000400142977014184, 'timing/model': 1.1209693260025233, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 1.46}) +[2026-06-24 03:27:56,820][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.0003599153715185821, 'train/grad_norm_pre_clip': 29.215103149414062, 'timing/data': 0.0004260160494595766, 'timing/model': 1.114080018014647, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 1.46}) +[2026-06-24 03:28:00,883][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.02203875221312046, 'train/grad_norm_pre_clip': 11.515386581420898, 'timing/data': 0.0004084310494363308, 'timing/model': 1.1144806679803878, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 1.46}) +[2026-06-24 03:28:04,952][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.00014942216512281448, 'train/grad_norm_pre_clip': 35.00682830810547, 'timing/data': 0.0003630860010161996, 'timing/model': 1.122460808022879, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 1.46}) +[2026-06-24 03:28:09,107][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.028637392446398735, 'train/grad_norm_pre_clip': 28.251117706298828, 'timing/data': 0.0003623509546741843, 'timing/model': 1.1190063949907199, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 1.46}) +[2026-06-24 03:28:13,189][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.004655270371586084, 'train/grad_norm_pre_clip': 16.646291732788086, 'timing/data': 0.00035175506491214037, 'timing/model': 1.117938622017391, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 1.46}) +[2026-06-24 03:28:17,265][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.12642215192317963, 'train/grad_norm_pre_clip': 19.167598724365234, 'timing/data': 0.0003564719809219241, 'timing/model': 1.12219667294994, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 1.46}) +[2026-06-24 03:28:21,334][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.009161046706140041, 'train/grad_norm_pre_clip': 20.084701538085938, 'timing/data': 0.00041661993600428104, 'timing/model': 1.116605959017761, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 1.46}) +[2026-06-24 03:28:25,468][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.21585755050182343, 'train/grad_norm_pre_clip': 33.36912536621094, 'timing/data': 0.00041578488890081644, 'timing/model': 1.118587402976118, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 1.46}) +[2026-06-24 03:28:29,556][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.015955431386828423, 'train/grad_norm_pre_clip': 9.043950080871582, 'timing/data': 0.00041451503057032824, 'timing/model': 1.121583387022838, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 1.46}) +[2026-06-24 03:28:33,626][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.07268480956554413, 'train/grad_norm_pre_clip': 30.0577392578125, 'timing/data': 0.0004129830049350858, 'timing/model': 1.1157056749798357, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 1.47}) +[2026-06-24 03:28:37,686][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.046647291630506516, 'train/grad_norm_pre_clip': 25.576313018798828, 'timing/data': 0.0003573269350454211, 'timing/model': 1.1146651469171047, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 1.47}) +[2026-06-24 03:28:41,883][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.0040421392768621445, 'train/grad_norm_pre_clip': 6.206515312194824, 'timing/data': 0.0003580119227990508, 'timing/model': 1.1152950290124863, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 1.47}) +[2026-06-24 03:28:45,972][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.015186714939773083, 'train/grad_norm_pre_clip': 27.97136878967285, 'timing/data': 0.00035816803574562073, 'timing/model': 1.1165066190296784, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 1.47}) +[2026-06-24 03:28:50,055][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.02539176493883133, 'train/grad_norm_pre_clip': 28.32240104675293, 'timing/data': 0.0003593149594962597, 'timing/model': 1.1187186100287363, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 1.47}) +[2026-06-24 03:28:54,223][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.03620850294828415, 'train/grad_norm_pre_clip': 13.707125663757324, 'timing/data': 0.0004467569524422288, 'timing/model': 1.1986962920054793, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 1.47}) +[2026-06-24 03:28:58,307][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.03451063856482506, 'train/grad_norm_pre_clip': 27.628238677978516, 'timing/data': 0.00035541097167879343, 'timing/model': 1.1190486440900713, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 1.47}) +[2026-06-24 03:29:02,418][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.07017914205789566, 'train/grad_norm_pre_clip': 23.41860008239746, 'timing/data': 0.0003651370061561465, 'timing/model': 1.1341540609719232, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 1.47}) +[2026-06-24 03:29:06,509][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.010921016335487366, 'train/grad_norm_pre_clip': 15.164328575134277, 'timing/data': 0.00035277102142572403, 'timing/model': 1.1310319720068946, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 1.47}) +[2026-06-24 03:29:10,594][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.0003136967134196311, 'train/grad_norm_pre_clip': 31.569154739379883, 'timing/data': 0.0003579499898478389, 'timing/model': 1.120482126949355, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 1.47}) +[2026-06-24 03:29:14,790][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.14723815023899078, 'train/grad_norm_pre_clip': 25.971176147460938, 'timing/data': 0.00039273803122341633, 'timing/model': 1.117596877971664, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 1.47}) +[2026-06-24 03:29:18,869][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.028585193678736687, 'train/grad_norm_pre_clip': 17.425477981567383, 'timing/data': 0.00035612506326287985, 'timing/model': 1.1171171029563993, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 1.47}) +[2026-06-24 03:29:22,959][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.0858774483203888, 'train/grad_norm_pre_clip': 12.35836124420166, 'timing/data': 0.0005028479499742389, 'timing/model': 1.1207834789529443, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 1.47}) +[2026-06-24 03:29:27,059][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.015564988367259502, 'train/grad_norm_pre_clip': 18.684709548950195, 'timing/data': 0.00036101893056184053, 'timing/model': 1.1234479009872302, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 1.47}) +[2026-06-24 03:29:31,327][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3800 +[2026-06-24 03:30:36,280][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3800 +[2026-06-24 03:30:36,282][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3800 +[2026-06-24 03:31:16,107][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3800 +[2026-06-24 03:31:16,107][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.01687607355415821, 'train/grad_norm_pre_clip': 13.281524658203125, 'eval/action_loss/seconds': 64.9506325260736, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.028942303732037544, 'eval/latency_2/loss': 0.028942303732037544, 'eval/flappy/loss': 0.028942303732037544, 'eval/flappy/latency_2/loss': 0.028942303732037544, 'eval/flap_precision': 0.9794801641586868, 'eval/flap_recall': 0.9715061058344641, 'eval/flap_f1': 0.9754768392370573, 'eval/flap_rate_model': 0.20305555555555554, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9754768392370572, 'eval/latency_2/flap_precision': 0.9794801641586868, 'eval/latency_2/flap_recall': 0.9715061058344641, 'eval/latency_2/flap_f1': 0.9754768392370573, 'eval/latency_2/flap_rate_model': 0.20305555555555554, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9754768392370572, 'eval/action_classification/data_wait_seconds': 0.4975792467594147, 'eval/action_classification/predict_seconds': 39.185943603515625, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0003632160369306803, 'timing/model': 1.1388512500561774, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 1.47}) +[2026-06-24 03:31:20,151][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.12252999097108841, 'train/grad_norm_pre_clip': 22.348867416381836, 'timing/data': 0.0004988859873265028, 'timing/model': 1.1063089040108025, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 1.47}) +[2026-06-24 03:31:24,196][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.003753731958568096, 'train/grad_norm_pre_clip': 10.6633939743042, 'timing/data': 0.0003674350446090102, 'timing/model': 1.1112300159875304, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 1.47}) +[2026-06-24 03:31:28,397][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.13232067227363586, 'train/grad_norm_pre_clip': 26.7879695892334, 'timing/data': 0.0005599489668384194, 'timing/model': 1.1073805389460176, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 1.47}) +[2026-06-24 03:31:32,446][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.03458992391824722, 'train/grad_norm_pre_clip': 16.5736141204834, 'timing/data': 0.00037350901402533054, 'timing/model': 1.1121958040166646, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 1.47}) +[2026-06-24 03:31:36,496][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.06409679353237152, 'train/grad_norm_pre_clip': 30.136337280273438, 'timing/data': 0.00046003900934010744, 'timing/model': 1.1114107000175864, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 1.47}) +[2026-06-24 03:31:40,546][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.10811064392328262, 'train/grad_norm_pre_clip': 18.82172393798828, 'timing/data': 0.0005029699532315135, 'timing/model': 1.1115234200842679, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 1.47}) +[2026-06-24 03:31:44,668][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.002450882224366069, 'train/grad_norm_pre_clip': 7.172558784484863, 'timing/data': 0.0003723920090124011, 'timing/model': 1.1118595460429788, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 1.47}) +[2026-06-24 03:31:48,725][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.004715245682746172, 'train/grad_norm_pre_clip': 16.941226959228516, 'timing/data': 0.00034956401214003563, 'timing/model': 1.1140189890284091, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 1.47}) +[2026-06-24 03:31:52,786][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.14432735741138458, 'train/grad_norm_pre_clip': 16.1868839263916, 'timing/data': 0.000409828033298254, 'timing/model': 1.1165944039821625, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 1.47}) +[2026-06-24 03:31:56,841][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.0076354192569851875, 'train/grad_norm_pre_clip': 5.195040702819824, 'timing/data': 0.00040829996578395367, 'timing/model': 1.1116205430589616, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 1.47}) +[2026-06-24 03:32:01,090][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.00466611422598362, 'train/grad_norm_pre_clip': 27.75554084777832, 'timing/data': 0.0006276160711422563, 'timing/model': 1.2902110500726849, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 1.47}) +[2026-06-24 03:32:05,149][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.01763133332133293, 'train/grad_norm_pre_clip': 9.9334077835083, 'timing/data': 0.0004118260694667697, 'timing/model': 1.1156062530353665, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 1.48}) +[2026-06-24 03:32:09,225][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.002400945872068405, 'train/grad_norm_pre_clip': 22.041057586669922, 'timing/data': 0.00036092998925596476, 'timing/model': 1.118454284965992, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 1.48}) +[2026-06-24 03:32:13,293][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.13513420522212982, 'train/grad_norm_pre_clip': 35.934383392333984, 'timing/data': 0.00037038198206573725, 'timing/model': 1.1153876319294795, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 1.48}) +[2026-06-24 03:32:17,475][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.10868775844573975, 'train/grad_norm_pre_clip': 40.576499938964844, 'timing/data': 0.00035566603764891624, 'timing/model': 1.1128283030120656, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 1.48}) +[2026-06-24 03:32:21,544][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.0404898039996624, 'train/grad_norm_pre_clip': 14.43895435333252, 'timing/data': 0.0003607521066442132, 'timing/model': 1.109751175972633, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 1.48}) +[2026-06-24 03:32:25,611][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.06483239680528641, 'train/grad_norm_pre_clip': 21.871068954467773, 'timing/data': 0.00035248103085905313, 'timing/model': 1.1163779150228947, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 1.48}) +[2026-06-24 03:32:29,682][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.03407859802246094, 'train/grad_norm_pre_clip': 25.35822105407715, 'timing/data': 0.0003598479088395834, 'timing/model': 1.1230653920210898, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 1.48}) +[2026-06-24 03:32:33,942][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.061981070786714554, 'train/grad_norm_pre_clip': 34.77385330200195, 'timing/data': 0.00035895302426069975, 'timing/model': 1.1141588219907135, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 1.48}) +[2026-06-24 03:32:38,015][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.11670789122581482, 'train/grad_norm_pre_clip': 17.61224365234375, 'timing/data': 0.00038771796971559525, 'timing/model': 1.117374317953363, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 1.48}) +[2026-06-24 03:32:42,080][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.011832579039037228, 'train/grad_norm_pre_clip': 16.930103302001953, 'timing/data': 0.00036027294117957354, 'timing/model': 1.114247870980762, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 1.48}) +[2026-06-24 03:32:46,148][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.0070254746824502945, 'train/grad_norm_pre_clip': 13.775014877319336, 'timing/data': 0.00037052202969789505, 'timing/model': 1.1137336110696197, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 1.48}) +[2026-06-24 03:32:50,391][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.06331662833690643, 'train/grad_norm_pre_clip': 16.419326782226562, 'timing/data': 0.0005359579809010029, 'timing/model': 1.1205297460546717, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 1.48}) +[2026-06-24 03:32:54,470][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.01613651029765606, 'train/grad_norm_pre_clip': 14.410736083984375, 'timing/data': 0.0005982270231470466, 'timing/model': 1.1131650320021436, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 1.48}) +[2026-06-24 03:32:58,555][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.004516218788921833, 'train/grad_norm_pre_clip': 25.356000900268555, 'timing/data': 0.0005622169701382518, 'timing/model': 1.117598628043197, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 1.48}) +[2026-06-24 03:33:02,630][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.1345788687467575, 'train/grad_norm_pre_clip': 28.348155975341797, 'timing/data': 0.00040796399116516113, 'timing/model': 1.1202029079431668, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 1.48}) +[2026-06-24 03:33:06,808][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.011982700787484646, 'train/grad_norm_pre_clip': 21.776376724243164, 'timing/data': 0.00035564496647566557, 'timing/model': 1.1337114210473374, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 1.48}) +[2026-06-24 03:33:10,889][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.11980672925710678, 'train/grad_norm_pre_clip': 16.210947036743164, 'timing/data': 0.00041279394645243883, 'timing/model': 1.1212964650476351, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 1.48}) +[2026-06-24 03:33:14,972][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.0005847515421919525, 'train/grad_norm_pre_clip': 35.792781829833984, 'timing/data': 0.0005103580188006163, 'timing/model': 1.1135981120169163, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 1.48}) +[2026-06-24 03:33:19,058][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.004347275476902723, 'train/grad_norm_pre_clip': 6.940730094909668, 'timing/data': 0.0003564460203051567, 'timing/model': 1.1157861519604921, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 1.48}) +[2026-06-24 03:33:23,213][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.0017785043455660343, 'train/grad_norm_pre_clip': 14.731812477111816, 'timing/data': 0.00037925492506474257, 'timing/model': 1.119937048992142, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 1.48}) +[2026-06-24 03:33:27,297][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.021446507424116135, 'train/grad_norm_pre_clip': 16.636322021484375, 'timing/data': 0.0003644989337772131, 'timing/model': 1.1222005020827055, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 1.48}) +[2026-06-24 03:33:31,382][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.0033200487960129976, 'train/grad_norm_pre_clip': 16.45853042602539, 'timing/data': 0.0005917110247537494, 'timing/model': 1.1210778509266675, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 1.48}) +[2026-06-24 03:33:35,465][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.01545831747353077, 'train/grad_norm_pre_clip': 14.384744644165039, 'timing/data': 0.00039914099033921957, 'timing/model': 1.1184097749646753, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 1.48}) +[2026-06-24 03:33:39,691][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.0015132147818803787, 'train/grad_norm_pre_clip': 12.581587791442871, 'timing/data': 0.0003621230134740472, 'timing/model': 1.12137324095238, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 1.48}) +[2026-06-24 03:33:43,771][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.003377362387254834, 'train/grad_norm_pre_clip': 25.432125091552734, 'timing/data': 0.00035932206083089113, 'timing/model': 1.1149412720697, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 1.48}) +[2026-06-24 03:33:47,843][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.01606040820479393, 'train/grad_norm_pre_clip': 18.818954467773438, 'timing/data': 0.0004917420446872711, 'timing/model': 1.1097665920387954, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 1.48}) +[2026-06-24 03:33:52,030][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.01889631897211075, 'train/grad_norm_pre_clip': 6.675446510314941, 'timing/data': 0.0003605220699682832, 'timing/model': 1.229984421050176, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 1.49}) +[2026-06-24 03:33:56,120][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.15880633890628815, 'train/grad_norm_pre_clip': 23.710908889770508, 'timing/data': 0.00035179604310542345, 'timing/model': 1.1275100480997935, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 1.49}) +[2026-06-24 03:34:00,196][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.249729722738266, 'train/grad_norm_pre_clip': 20.82064437866211, 'timing/data': 0.00035522901453077793, 'timing/model': 1.1116437099408358, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 1.49}) +[2026-06-24 03:34:04,285][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.0026269189547747374, 'train/grad_norm_pre_clip': 15.517973899841309, 'timing/data': 0.00040345697198063135, 'timing/model': 1.1164534469135106, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 1.49}) +[2026-06-24 03:34:08,551][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.021684937179088593, 'train/grad_norm_pre_clip': 22.57662582397461, 'timing/data': 0.00037517701275646687, 'timing/model': 1.2864205510122702, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 1.49}) +[2026-06-24 03:34:12,655][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.0005669089150615036, 'train/grad_norm_pre_clip': 6.175751686096191, 'timing/data': 0.0003566490486264229, 'timing/model': 1.1256319120293483, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 1.49}) +[2026-06-24 03:34:16,760][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.03687717393040657, 'train/grad_norm_pre_clip': 21.317583084106445, 'timing/data': 0.00036820408422499895, 'timing/model': 1.1214473189320415, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 1.49}) +[2026-06-24 03:34:20,827][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.0036882045678794384, 'train/grad_norm_pre_clip': 4.100338459014893, 'timing/data': 0.0006175140151754022, 'timing/model': 1.1155750210164115, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 1.49}) +[2026-06-24 03:34:24,983][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.0268528051674366, 'train/grad_norm_pre_clip': 19.42243003845215, 'timing/data': 0.00035674404352903366, 'timing/model': 1.1163951570633799, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 1.49}) +[2026-06-24 03:34:29,061][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.1293371617794037, 'train/grad_norm_pre_clip': 23.663084030151367, 'timing/data': 0.00036805495619773865, 'timing/model': 1.1144252750091255, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 1.49}) +[2026-06-24 03:34:33,141][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.014256632886826992, 'train/grad_norm_pre_clip': 29.99574089050293, 'timing/data': 0.00043281703256070614, 'timing/model': 1.1234461120329797, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 1.49}) +[2026-06-24 03:34:37,234][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.0037290744949132204, 'train/grad_norm_pre_clip': 26.313011169433594, 'timing/data': 0.0003635529428720474, 'timing/model': 1.1205069739371538, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 1.49}) +[2026-06-24 03:34:41,405][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.035013873130083084, 'train/grad_norm_pre_clip': 18.64126205444336, 'timing/data': 0.0003640789072960615, 'timing/model': 1.1194548771018162, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 1.49}) +[2026-06-24 03:34:45,485][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.03183296322822571, 'train/grad_norm_pre_clip': 17.12933349609375, 'timing/data': 0.0005990229547023773, 'timing/model': 1.1167441370198503, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 1.49}) +[2026-06-24 03:34:49,567][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.004421241115778685, 'train/grad_norm_pre_clip': 20.146100997924805, 'timing/data': 0.0003643650561571121, 'timing/model': 1.113888725056313, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 1.49}) +[2026-06-24 03:34:53,657][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.09394574165344238, 'train/grad_norm_pre_clip': 38.7335205078125, 'timing/data': 0.0003741550026461482, 'timing/model': 1.1212939309189096, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 1.49}) +[2026-06-24 03:34:57,864][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.03687717393040657, 'train/grad_norm_pre_clip': 27.3712158203125, 'timing/data': 0.00045695703011006117, 'timing/model': 1.1111058510141447, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 1.49}) +[2026-06-24 03:35:01,938][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.0699278935790062, 'train/grad_norm_pre_clip': 34.13215255737305, 'timing/data': 0.00040829204954206944, 'timing/model': 1.1186397180426866, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 1.49}) +[2026-06-24 03:35:06,039][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.043382372707128525, 'train/grad_norm_pre_clip': 48.946136474609375, 'timing/data': 0.0003553659189492464, 'timing/model': 1.1190684699686244, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 1.49}) +[2026-06-24 03:35:10,118][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.0013432225678116083, 'train/grad_norm_pre_clip': 23.434600830078125, 'timing/data': 0.0003585979575291276, 'timing/model': 1.117323816055432, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 1.49}) +[2026-06-24 03:35:14,307][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.08520086854696274, 'train/grad_norm_pre_clip': 26.54242515563965, 'timing/data': 0.0003606729442253709, 'timing/model': 1.1164896769914776, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 1.49}) +[2026-06-24 03:35:18,416][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.0730479285120964, 'train/grad_norm_pre_clip': 28.352209091186523, 'timing/data': 0.00036030402407050133, 'timing/model': 1.1242790310643613, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 1.49}) +[2026-06-24 03:35:22,533][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.006497590336948633, 'train/grad_norm_pre_clip': 21.449663162231445, 'timing/data': 0.0005692000268027186, 'timing/model': 1.124192066024989, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 1.49}) +[2026-06-24 03:35:26,620][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.011803542263805866, 'train/grad_norm_pre_clip': 39.40095520019531, 'timing/data': 0.00039942795410752296, 'timing/model': 1.1130552630638704, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 1.49}) +[2026-06-24 03:35:30,768][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.004395797848701477, 'train/grad_norm_pre_clip': 10.683609008789062, 'timing/data': 0.0003561939811334014, 'timing/model': 1.1153162609552965, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 1.49}) +[2026-06-24 03:35:34,859][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.002449916210025549, 'train/grad_norm_pre_clip': 15.826204299926758, 'timing/data': 0.0004836439620703459, 'timing/model': 1.116524587967433, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 1.5}) +[2026-06-24 03:35:38,938][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.07473542541265488, 'train/grad_norm_pre_clip': 33.0639533996582, 'timing/data': 0.00045299401972442865, 'timing/model': 1.1116172530455515, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 1.5}) +[2026-06-24 03:35:43,041][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.028438091278076172, 'train/grad_norm_pre_clip': 14.136845588684082, 'timing/data': 0.0003592579159885645, 'timing/model': 1.1334596070228145, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 1.5}) +[2026-06-24 03:35:47,311][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.2329981029033661, 'train/grad_norm_pre_clip': 33.854007720947266, 'timing/data': 0.0003616330213844776, 'timing/model': 1.1267181050498039, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 1.5}) +[2026-06-24 03:35:51,396][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.05105282738804817, 'train/grad_norm_pre_clip': 28.815698623657227, 'timing/data': 0.0003492149990051985, 'timing/model': 1.1168990969890729, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 1.5}) +[2026-06-24 03:35:55,504][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.06679687649011612, 'train/grad_norm_pre_clip': 23.435869216918945, 'timing/data': 0.0003697880310937762, 'timing/model': 1.126140251988545, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 1.5}) +[2026-06-24 03:35:59,621][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.024439336732029915, 'train/grad_norm_pre_clip': 17.379873275756836, 'timing/data': 0.0006234339671209455, 'timing/model': 1.1244732399936765, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 1.5}) +[2026-06-24 03:36:03,815][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.013793988153338432, 'train/grad_norm_pre_clip': 22.64024543762207, 'timing/data': 0.0005198830040171742, 'timing/model': 1.1204702439717948, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 1.5}) +[2026-06-24 03:36:07,892][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.05917216092348099, 'train/grad_norm_pre_clip': 17.676292419433594, 'timing/data': 0.00046760495752096176, 'timing/model': 1.1173205969389528, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 1.5}) +[2026-06-24 03:36:11,975][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.0007575807394459844, 'train/grad_norm_pre_clip': 19.611318588256836, 'timing/data': 0.0005319759948179126, 'timing/model': 1.119846583926119, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 1.5}) +[2026-06-24 03:36:16,066][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.07728590816259384, 'train/grad_norm_pre_clip': 22.01682472229004, 'timing/data': 0.00035992497578263283, 'timing/model': 1.1190540120005608, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 1.5}) +[2026-06-24 03:36:20,316][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.04275225102901459, 'train/grad_norm_pre_clip': 20.321075439453125, 'timing/data': 0.0006225699326023459, 'timing/model': 1.1230309039819986, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 1.5}) +[2026-06-24 03:36:24,415][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.03826053440570831, 'train/grad_norm_pre_clip': 21.468690872192383, 'timing/data': 0.0003553900169208646, 'timing/model': 1.1194721719948575, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 1.5}) +[2026-06-24 03:36:28,506][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.049102529883384705, 'train/grad_norm_pre_clip': 16.269350051879883, 'timing/data': 0.0005495870718732476, 'timing/model': 1.1229312039213255, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 1.5}) +[2026-06-24 03:36:32,604][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.0013141445815563202, 'train/grad_norm_pre_clip': 20.727048873901367, 'timing/data': 0.00045355106703937054, 'timing/model': 1.1199882170185447, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 1.5}) +[2026-06-24 03:36:36,772][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.056567925959825516, 'train/grad_norm_pre_clip': 15.142271041870117, 'timing/data': 0.0007222769781947136, 'timing/model': 1.1221129760378972, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 1.5}) +[2026-06-24 03:36:40,860][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.008484513498842716, 'train/grad_norm_pre_clip': 19.19687843322754, 'timing/data': 0.0003497389843687415, 'timing/model': 1.122379946988076, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 1.5}) +[2026-06-24 03:36:44,946][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.07435508817434311, 'train/grad_norm_pre_clip': 15.460653305053711, 'timing/data': 0.0004053299780935049, 'timing/model': 1.1170054440153763, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 1.5}) +[2026-06-24 03:36:49,021][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.0009568608948029578, 'train/grad_norm_pre_clip': 18.602014541625977, 'timing/data': 0.0004319149302318692, 'timing/model': 1.1186371259391308, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 1.5}) +[2026-06-24 03:36:53,170][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.0028201567474752665, 'train/grad_norm_pre_clip': 34.317325592041016, 'timing/data': 0.0004960399819537997, 'timing/model': 1.1148427589796484, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 1.5}) +[2026-06-24 03:36:57,237][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.0745067447423935, 'train/grad_norm_pre_clip': 30.100711822509766, 'timing/data': 0.0005299979820847511, 'timing/model': 1.113712124992162, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 1.5}) +[2026-06-24 03:37:01,346][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.0034341812133789062, 'train/grad_norm_pre_clip': 17.32464599609375, 'timing/data': 0.00036042602732777596, 'timing/model': 1.1229831849923357, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 1.5}) +[2026-06-24 03:37:05,434][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.09732179343700409, 'train/grad_norm_pre_clip': 15.987594604492188, 'timing/data': 0.00036367797292768955, 'timing/model': 1.1206546190660447, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 1.5}) +[2026-06-24 03:37:09,681][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.026907024905085564, 'train/grad_norm_pre_clip': 15.689434051513672, 'timing/data': 0.0005305709782987833, 'timing/model': 1.1265346070285887, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 1.5}) +[2026-06-24 03:37:13,760][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.012902511283755302, 'train/grad_norm_pre_clip': 15.96612548828125, 'timing/data': 0.00036355911288410425, 'timing/model': 1.118043606984429, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 1.5}) +[2026-06-24 03:37:17,877][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.04175311699509621, 'train/grad_norm_pre_clip': 36.1413459777832, 'timing/data': 0.00040252809412777424, 'timing/model': 1.1315685029840097, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 1.5}) +[2026-06-24 03:37:22,073][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.0064202118664979935, 'train/grad_norm_pre_clip': 15.608589172363281, 'timing/data': 0.00035579397808760405, 'timing/model': 1.2188839780865237, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 1.51}) +[2026-06-24 03:37:26,170][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.06192529574036598, 'train/grad_norm_pre_clip': 16.117544174194336, 'timing/data': 0.00035222002770751715, 'timing/model': 1.1195106930099428, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 1.51}) +[2026-06-24 03:37:30,269][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.2052915245294571, 'train/grad_norm_pre_clip': 50.63276672363281, 'timing/data': 0.00044031301513314247, 'timing/model': 1.1261000799713656, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 1.51}) +[2026-06-24 03:37:34,355][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.07049688696861267, 'train/grad_norm_pre_clip': 12.902878761291504, 'timing/data': 0.0005063250428065658, 'timing/model': 1.119443215080537, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 1.51}) +[2026-06-24 03:37:38,634][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.02111578732728958, 'train/grad_norm_pre_clip': 6.823574542999268, 'timing/data': 0.0005819449434056878, 'timing/model': 1.3053999019321054, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 1.51}) +[2026-06-24 03:37:42,744][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.0812680795788765, 'train/grad_norm_pre_clip': 28.01065444946289, 'timing/data': 0.00038631202187389135, 'timing/model': 1.1179458420956507, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 1.51}) +[2026-06-24 03:37:46,837][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.024688826873898506, 'train/grad_norm_pre_clip': 40.766090393066406, 'timing/data': 0.0003653269959613681, 'timing/model': 1.1171471920097247, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 1.51}) +[2026-06-24 03:37:50,917][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.043426088988780975, 'train/grad_norm_pre_clip': 44.267845153808594, 'timing/data': 0.0003778629470616579, 'timing/model': 1.1191852639894933, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 1.51}) +[2026-06-24 03:37:55,075][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.05163270980119705, 'train/grad_norm_pre_clip': 24.164339065551758, 'timing/data': 0.00039733294397592545, 'timing/model': 1.113020714954473, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 1.51}) +[2026-06-24 03:37:59,178][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.04290970042347908, 'train/grad_norm_pre_clip': 27.88775062561035, 'timing/data': 0.0003671731101348996, 'timing/model': 1.1202522970270365, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 1.51}) +[2026-06-24 03:38:03,287][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.1259741187095642, 'train/grad_norm_pre_clip': 29.381210327148438, 'timing/data': 0.0005376269109547138, 'timing/model': 1.1447184230200946, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 1.51}) +[2026-06-24 03:38:07,402][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3900 +[2026-06-24 03:39:11,308][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3900 +[2026-06-24 03:39:11,310][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3900 +[2026-06-24 03:39:50,977][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3900 +[2026-06-24 03:39:50,977][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.003556522075086832, 'train/grad_norm_pre_clip': 19.64521026611328, 'eval/action_loss/seconds': 63.904092960990965, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.029388070106506348, 'eval/latency_2/loss': 0.029388070106506348, 'eval/flappy/loss': 0.029388070106506348, 'eval/flappy/latency_2/loss': 0.029388070106506348, 'eval/flap_precision': 0.9756427604871448, 'eval/flap_recall': 0.9782903663500678, 'eval/flap_f1': 0.9769647696476965, 'eval/flap_rate_model': 0.20527777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9769647696476965, 'eval/latency_2/flap_precision': 0.9756427604871448, 'eval/latency_2/flap_recall': 0.9782903663500678, 'eval/latency_2/flap_f1': 0.9769647696476965, 'eval/latency_2/flap_rate_model': 0.20527777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9769647696476965, 'eval/action_classification/data_wait_seconds': 0.48079708218574524, 'eval/action_classification/predict_seconds': 39.05429458618164, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.000522630987688899, 'timing/model': 1.1416234150528908, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 1.51}) +[2026-06-24 03:39:55,011][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.0280179250985384, 'train/grad_norm_pre_clip': 29.680458068847656, 'timing/data': 0.0006127760279923677, 'timing/model': 1.1085653160698712, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 1.51}) +[2026-06-24 03:39:59,055][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.06539089232683182, 'train/grad_norm_pre_clip': 10.342758178710938, 'timing/data': 0.0003693690523505211, 'timing/model': 1.1081823860295117, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 1.51}) +[2026-06-24 03:40:03,110][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.000509446079377085, 'train/grad_norm_pre_clip': 8.08102035522461, 'timing/data': 0.00036372989416122437, 'timing/model': 1.1092388569377363, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 1.51}) +[2026-06-24 03:40:07,324][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.10206447541713715, 'train/grad_norm_pre_clip': 17.85914421081543, 'timing/data': 0.0004952979506924748, 'timing/model': 1.1147530949674547, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 1.51}) +[2026-06-24 03:40:11,362][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.0029326272197067738, 'train/grad_norm_pre_clip': 9.300667762756348, 'timing/data': 0.0004128379514440894, 'timing/model': 1.1122193320188671, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 1.51}) +[2026-06-24 03:40:15,422][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.011176700703799725, 'train/grad_norm_pre_clip': 17.793733596801758, 'timing/data': 0.0006262250244617462, 'timing/model': 1.1144404100487009, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 1.51}) +[2026-06-24 03:40:19,505][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.07197482138872147, 'train/grad_norm_pre_clip': 19.947004318237305, 'timing/data': 0.0005749550182372332, 'timing/model': 1.1241333950310946, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 1.51}) +[2026-06-24 03:40:23,671][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.06598366796970367, 'train/grad_norm_pre_clip': 32.13277816772461, 'timing/data': 0.00036144594196230173, 'timing/model': 1.109421042026952, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 1.51}) +[2026-06-24 03:40:27,724][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.023482318967580795, 'train/grad_norm_pre_clip': 24.995168685913086, 'timing/data': 0.0005061690462753177, 'timing/model': 1.1134489090181887, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 1.51}) +[2026-06-24 03:40:31,814][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.05975230410695076, 'train/grad_norm_pre_clip': 14.506651878356934, 'timing/data': 0.0003543769707903266, 'timing/model': 1.1234143530018628, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 1.51}) +[2026-06-24 03:40:35,874][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.004085181746631861, 'train/grad_norm_pre_clip': 18.90069580078125, 'timing/data': 0.00036383699625730515, 'timing/model': 1.1111506429733709, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 1.51}) +[2026-06-24 03:40:40,015][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.011968317441642284, 'train/grad_norm_pre_clip': 19.30040168762207, 'timing/data': 0.000402871984988451, 'timing/model': 1.1131330280331895, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 1.51}) +[2026-06-24 03:40:44,077][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.006875000428408384, 'train/grad_norm_pre_clip': 14.76726245880127, 'timing/data': 0.0006483630277216434, 'timing/model': 1.114281577989459, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 1.51}) +[2026-06-24 03:40:48,161][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.009169071912765503, 'train/grad_norm_pre_clip': 25.554058074951172, 'timing/data': 0.000359158031642437, 'timing/model': 1.1159752620151266, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 1.51}) +[2026-06-24 03:40:52,224][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.05935123562812805, 'train/grad_norm_pre_clip': 30.84278678894043, 'timing/data': 0.00036653189454227686, 'timing/model': 1.1115536839934066, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 1.52}) +[2026-06-24 03:40:56,465][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.11676494777202606, 'train/grad_norm_pre_clip': 42.78562927246094, 'timing/data': 0.00037507398519665003, 'timing/model': 1.1137046510120854, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 1.52}) +[2026-06-24 03:41:00,531][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.01694868691265583, 'train/grad_norm_pre_clip': 32.947391510009766, 'timing/data': 0.0005258589517325163, 'timing/model': 1.115592930931598, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 1.52}) +[2026-06-24 03:41:04,621][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.021520622074604034, 'train/grad_norm_pre_clip': 11.689806938171387, 'timing/data': 0.00036047096364200115, 'timing/model': 1.1129390900023282, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 1.52}) +[2026-06-24 03:41:08,692][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.10365678369998932, 'train/grad_norm_pre_clip': 20.27387237548828, 'timing/data': 0.0004394860006868839, 'timing/model': 1.1173121790634468, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 1.52}) +[2026-06-24 03:41:12,868][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.010705841705203056, 'train/grad_norm_pre_clip': 17.06093406677246, 'timing/data': 0.0005521740531548858, 'timing/model': 1.1187475869664922, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 1.52}) +[2026-06-24 03:41:16,943][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.019614944234490395, 'train/grad_norm_pre_clip': 21.712261199951172, 'timing/data': 0.00035648199263960123, 'timing/model': 1.1154390699230134, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 1.52}) +[2026-06-24 03:41:21,025][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.0019040507031604648, 'train/grad_norm_pre_clip': 18.35866355895996, 'timing/data': 0.00035671796649694443, 'timing/model': 1.1217140549561009, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 1.52}) +[2026-06-24 03:41:25,113][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.00899556651711464, 'train/grad_norm_pre_clip': 19.464990615844727, 'timing/data': 0.00039092195220291615, 'timing/model': 1.11904580402188, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 1.52}) +[2026-06-24 03:41:29,377][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.00028193212347105145, 'train/grad_norm_pre_clip': 19.41495704650879, 'timing/data': 0.00036001100670546293, 'timing/model': 1.1199636049568653, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 1.52}) +[2026-06-24 03:41:33,487][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.003364283125847578, 'train/grad_norm_pre_clip': 20.895524978637695, 'timing/data': 0.0003671059384942055, 'timing/model': 1.1213646889664233, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 1.52}) +[2026-06-24 03:41:37,580][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.05399797484278679, 'train/grad_norm_pre_clip': 23.26288604736328, 'timing/data': 0.0003520400496199727, 'timing/model': 1.1207395748933777, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 1.52}) +[2026-06-24 03:41:41,667][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.08262099325656891, 'train/grad_norm_pre_clip': 14.827648162841797, 'timing/data': 0.0003694269107654691, 'timing/model': 1.1148585840128362, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 1.52}) +[2026-06-24 03:41:45,813][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.007315368391573429, 'train/grad_norm_pre_clip': 10.894851684570312, 'timing/data': 0.00035740004386752844, 'timing/model': 1.1174713590880856, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 1.52}) +[2026-06-24 03:41:49,885][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.12526899576187134, 'train/grad_norm_pre_clip': 17.06509017944336, 'timing/data': 0.0006210589781403542, 'timing/model': 1.1162691770587116, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 1.52}) +[2026-06-24 03:41:53,962][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.0002870449097827077, 'train/grad_norm_pre_clip': 27.605417251586914, 'timing/data': 0.00040475092828273773, 'timing/model': 1.1178222160087898, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 1.52}) +[2026-06-24 03:41:58,025][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.04525614529848099, 'train/grad_norm_pre_clip': 51.08313751220703, 'timing/data': 0.0005349429557099938, 'timing/model': 1.1130240049678832, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 1.52}) +[2026-06-24 03:42:02,194][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.018793988972902298, 'train/grad_norm_pre_clip': 26.45223617553711, 'timing/data': 0.0003882769960910082, 'timing/model': 1.1162616369547322, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 1.52}) +[2026-06-24 03:42:06,272][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.029114626348018646, 'train/grad_norm_pre_clip': 15.442532539367676, 'timing/data': 0.00045377796050161123, 'timing/model': 1.1143786560278386, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 1.52}) +[2026-06-24 03:42:10,393][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.0180787593126297, 'train/grad_norm_pre_clip': 23.029369354248047, 'timing/data': 0.00035332900006324053, 'timing/model': 1.1322039340157062, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 1.52}) +[2026-06-24 03:42:14,485][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.03785213083028793, 'train/grad_norm_pre_clip': 18.841632843017578, 'timing/data': 0.0003727140137925744, 'timing/model': 1.1277047790354118, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 1.52}) +[2026-06-24 03:42:18,762][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.038667745888233185, 'train/grad_norm_pre_clip': 35.89118194580078, 'timing/data': 0.0003503799671307206, 'timing/model': 1.128440759028308, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 1.52}) +[2026-06-24 03:42:22,841][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.0027864109724760056, 'train/grad_norm_pre_clip': 21.44377326965332, 'timing/data': 0.00036003405693918467, 'timing/model': 1.1200046730227768, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 1.52}) +[2026-06-24 03:42:26,930][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.015292325988411903, 'train/grad_norm_pre_clip': 25.925521850585938, 'timing/data': 0.0003632879815995693, 'timing/model': 1.12851079297252, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 1.52}) +[2026-06-24 03:42:31,098][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.0030456141103059053, 'train/grad_norm_pre_clip': 17.644838333129883, 'timing/data': 0.0003566460218280554, 'timing/model': 1.2040668190456927, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 1.52}) +[2026-06-24 03:42:35,186][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.04837565869092941, 'train/grad_norm_pre_clip': 45.28384780883789, 'timing/data': 0.0004696889081969857, 'timing/model': 1.1178222720045596, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 1.52}) +[2026-06-24 03:42:39,257][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.06246548146009445, 'train/grad_norm_pre_clip': 41.69915771484375, 'timing/data': 0.000578935956582427, 'timing/model': 1.1163961689453572, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 1.53}) +[2026-06-24 03:42:43,348][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.00746991578489542, 'train/grad_norm_pre_clip': 5.4418044090271, 'timing/data': 0.0006258049979805946, 'timing/model': 1.1144232379738241, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 1.53}) +[2026-06-24 03:42:47,631][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.008888701908290386, 'train/grad_norm_pre_clip': 10.845901489257812, 'timing/data': 0.00035215297248214483, 'timing/model': 1.2847854580031708, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 1.53}) +[2026-06-24 03:42:51,724][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.1728440374135971, 'train/grad_norm_pre_clip': 28.5792236328125, 'timing/data': 0.0003717980580404401, 'timing/model': 1.1189081189222634, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 1.53}) +[2026-06-24 03:42:55,813][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.11845973879098892, 'train/grad_norm_pre_clip': 48.697052001953125, 'timing/data': 0.00035801203921437263, 'timing/model': 1.1192600639769807, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 1.53}) +[2026-06-24 03:42:59,903][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.013121324591338634, 'train/grad_norm_pre_clip': 5.57623291015625, 'timing/data': 0.0004281869623810053, 'timing/model': 1.1201339330291376, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 1.53}) +[2026-06-24 03:43:04,070][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.011507516726851463, 'train/grad_norm_pre_clip': 10.773776054382324, 'timing/data': 0.00036293803714215755, 'timing/model': 1.1237253160215914, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 1.53}) +[2026-06-24 03:43:08,145][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.05703675374388695, 'train/grad_norm_pre_clip': 10.97659683227539, 'timing/data': 0.0003594299778342247, 'timing/model': 1.1174032590351999, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 1.53}) +[2026-06-24 03:43:12,225][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.05959837883710861, 'train/grad_norm_pre_clip': 5.873991966247559, 'timing/data': 0.0004105860134586692, 'timing/model': 1.1225148669909686, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 1.53}) +[2026-06-24 03:43:16,331][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.012735276482999325, 'train/grad_norm_pre_clip': 25.478422164916992, 'timing/data': 0.00035898794885724783, 'timing/model': 1.1249452299671248, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 1.53}) +[2026-06-24 03:43:20,497][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.033224400132894516, 'train/grad_norm_pre_clip': 9.00043773651123, 'timing/data': 0.000369283021427691, 'timing/model': 1.119977081893012, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 1.53}) +[2026-06-24 03:43:24,578][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.08048270642757416, 'train/grad_norm_pre_clip': 24.444658279418945, 'timing/data': 0.00041720899753272533, 'timing/model': 1.1140786000760272, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 1.53}) +[2026-06-24 03:43:28,669][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.06959414482116699, 'train/grad_norm_pre_clip': 34.20033645629883, 'timing/data': 0.0004126110579818487, 'timing/model': 1.1213764859130606, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 1.53}) +[2026-06-24 03:43:32,760][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.015431978739798069, 'train/grad_norm_pre_clip': 24.649574279785156, 'timing/data': 0.0004109849687665701, 'timing/model': 1.1187709040241316, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 1.53}) +[2026-06-24 03:43:36,973][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.04205304756760597, 'train/grad_norm_pre_clip': 13.638511657714844, 'timing/data': 0.0003570449771359563, 'timing/model': 1.1138286229688674, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 1.53}) +[2026-06-24 03:43:41,090][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.006046912632882595, 'train/grad_norm_pre_clip': 5.048474311828613, 'timing/data': 0.0005763010121881962, 'timing/model': 1.1364470509579405, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 1.53}) +[2026-06-24 03:43:45,188][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.032804589718580246, 'train/grad_norm_pre_clip': 9.326355934143066, 'timing/data': 0.00036571803502738476, 'timing/model': 1.1161896680714563, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 1.53}) +[2026-06-24 03:43:49,286][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.0017356503522023559, 'train/grad_norm_pre_clip': 21.2401123046875, 'timing/data': 0.0006431579822674394, 'timing/model': 1.11895536002703, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 1.53}) +[2026-06-24 03:43:53,469][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.0066639576107263565, 'train/grad_norm_pre_clip': 3.579277992248535, 'timing/data': 0.0003609509440138936, 'timing/model': 1.1199998570373282, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 1.53}) +[2026-06-24 03:43:57,569][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.10110176354646683, 'train/grad_norm_pre_clip': 38.20895767211914, 'timing/data': 0.0004654850345104933, 'timing/model': 1.124409868963994, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 1.53}) +[2026-06-24 03:44:01,663][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.05766631290316582, 'train/grad_norm_pre_clip': 27.4658260345459, 'timing/data': 0.0004067419795319438, 'timing/model': 1.1195589560084045, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 1.53}) +[2026-06-24 03:44:05,769][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.06967077404260635, 'train/grad_norm_pre_clip': 19.419761657714844, 'timing/data': 0.00036823900882154703, 'timing/model': 1.1278461179463193, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 1.53}) +[2026-06-24 03:44:09,938][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.00034732482163235545, 'train/grad_norm_pre_clip': 26.176685333251953, 'timing/data': 0.00038645800668746233, 'timing/model': 1.1225286059780046, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 1.53}) +[2026-06-24 03:44:14,030][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.037116359919309616, 'train/grad_norm_pre_clip': 22.72480010986328, 'timing/data': 0.0003585349768400192, 'timing/model': 1.1181371379643679, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 1.53}) +[2026-06-24 03:44:18,144][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.07072614133358002, 'train/grad_norm_pre_clip': 14.856778144836426, 'timing/data': 0.0003584889927878976, 'timing/model': 1.1254751810338348, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 1.53}) +[2026-06-24 03:44:22,239][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.004129039589315653, 'train/grad_norm_pre_clip': 17.836793899536133, 'timing/data': 0.00036742095835506916, 'timing/model': 1.1175880059599876, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 1.53}) +[2026-06-24 03:44:26,498][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.01905815675854683, 'train/grad_norm_pre_clip': 15.26959228515625, 'timing/data': 0.00038508104626089334, 'timing/model': 1.1205351729877293, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 1.54}) +[2026-06-24 03:44:30,571][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.04126182571053505, 'train/grad_norm_pre_clip': 33.28630065917969, 'timing/data': 0.0003537869779393077, 'timing/model': 1.117687518009916, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 1.54}) +[2026-06-24 03:44:34,655][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.012388373725116253, 'train/grad_norm_pre_clip': 27.00088882446289, 'timing/data': 0.000361658982001245, 'timing/model': 1.1206118679838255, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 1.54}) +[2026-06-24 03:44:38,736][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.01574607938528061, 'train/grad_norm_pre_clip': 14.421833992004395, 'timing/data': 0.000481785973533988, 'timing/model': 1.1236384329386055, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 1.54}) +[2026-06-24 03:44:42,926][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.005104831885546446, 'train/grad_norm_pre_clip': 17.78069305419922, 'timing/data': 0.0005670939572155476, 'timing/model': 1.1186177399940789, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 1.54}) +[2026-06-24 03:44:47,011][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.012380006723105907, 'train/grad_norm_pre_clip': 25.06366539001465, 'timing/data': 0.00036830094177275896, 'timing/model': 1.1172717749141157, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 1.54}) +[2026-06-24 03:44:51,108][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.012550052255392075, 'train/grad_norm_pre_clip': 25.971843719482422, 'timing/data': 0.00036065210588276386, 'timing/model': 1.1270435489714146, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 1.54}) +[2026-06-24 03:44:55,194][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.04170043393969536, 'train/grad_norm_pre_clip': 18.954383850097656, 'timing/data': 0.00035826407838612795, 'timing/model': 1.1175834300229326, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 1.54}) +[2026-06-24 03:44:59,443][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.009174025617539883, 'train/grad_norm_pre_clip': 18.618114471435547, 'timing/data': 0.00035506405401974916, 'timing/model': 1.117605453939177, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 1.54}) +[2026-06-24 03:45:03,531][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.015734463930130005, 'train/grad_norm_pre_clip': 8.622397422790527, 'timing/data': 0.00035770994145423174, 'timing/model': 1.1184701999882236, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 1.54}) +[2026-06-24 03:45:07,642][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.005487491842359304, 'train/grad_norm_pre_clip': 37.272029876708984, 'timing/data': 0.00036580103915184736, 'timing/model': 1.1215336179593578, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 1.54}) +[2026-06-24 03:45:11,727][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.0831136628985405, 'train/grad_norm_pre_clip': 31.464397430419922, 'timing/data': 0.0003875099355354905, 'timing/model': 1.120021544978954, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 1.54}) +[2026-06-24 03:45:15,891][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.009193134494125843, 'train/grad_norm_pre_clip': 17.262840270996094, 'timing/data': 0.000364686013199389, 'timing/model': 1.1181718159932643, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 1.54}) +[2026-06-24 03:45:19,979][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.0032370167318731546, 'train/grad_norm_pre_clip': 43.03721618652344, 'timing/data': 0.0005424319533631206, 'timing/model': 1.1194291759748012, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 1.54}) +[2026-06-24 03:45:24,057][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.011835738085210323, 'train/grad_norm_pre_clip': 16.007986068725586, 'timing/data': 0.0003542910562828183, 'timing/model': 1.1188082940643653, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 1.54}) +[2026-06-24 03:45:28,152][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.012601646594703197, 'train/grad_norm_pre_clip': 11.851264953613281, 'timing/data': 0.000376226962544024, 'timing/model': 1.1192700680112466, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 1.54}) +[2026-06-24 03:45:32,335][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.06931019574403763, 'train/grad_norm_pre_clip': 19.17685890197754, 'timing/data': 0.0003867719788104296, 'timing/model': 1.125061703962274, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 1.54}) +[2026-06-24 03:45:36,422][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.10384424030780792, 'train/grad_norm_pre_clip': 27.1653995513916, 'timing/data': 0.00037554698064923286, 'timing/model': 1.1202923249220476, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 1.54}) +[2026-06-24 03:45:40,521][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.047539759427309036, 'train/grad_norm_pre_clip': 18.49786949157715, 'timing/data': 0.0003799460828304291, 'timing/model': 1.124356257962063, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 1.54}) +[2026-06-24 03:45:44,614][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.0022749139461666346, 'train/grad_norm_pre_clip': 17.956762313842773, 'timing/data': 0.0003607309190556407, 'timing/model': 1.1226000040769577, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 1.54}) +[2026-06-24 03:45:48,864][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.13094541430473328, 'train/grad_norm_pre_clip': 44.33655548095703, 'timing/data': 0.0003778589889407158, 'timing/model': 1.1259353349450976, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 1.54}) +[2026-06-24 03:45:52,950][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.1377905309200287, 'train/grad_norm_pre_clip': 21.497228622436523, 'timing/data': 0.0005442720139399171, 'timing/model': 1.1201416889671236, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 1.54}) +[2026-06-24 03:45:57,038][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.029260631650686264, 'train/grad_norm_pre_clip': 26.1199951171875, 'timing/data': 0.00035593705251812935, 'timing/model': 1.1161312538897619, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 1.54}) +[2026-06-24 03:46:01,195][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.04600394889712334, 'train/grad_norm_pre_clip': 21.31272315979004, 'timing/data': 0.00036558799911290407, 'timing/model': 1.1924132580170408, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 1.54}) +[2026-06-24 03:46:05,300][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.059965964406728745, 'train/grad_norm_pre_clip': 30.544710159301758, 'timing/data': 0.00036352104507386684, 'timing/model': 1.1246410249732435, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 1.54}) +[2026-06-24 03:46:09,392][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.1501070261001587, 'train/grad_norm_pre_clip': 33.142948150634766, 'timing/data': 0.00044405495282262564, 'timing/model': 1.12082591408398, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 1.54}) +[2026-06-24 03:46:13,478][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.07458585500717163, 'train/grad_norm_pre_clip': 26.587120056152344, 'timing/data': 0.0003745279973372817, 'timing/model': 1.116486697923392, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 1.55}) +[2026-06-24 03:46:17,743][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.08180070668458939, 'train/grad_norm_pre_clip': 55.21650314331055, 'timing/data': 0.0003790640039369464, 'timing/model': 1.2938392770010978, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 1.55}) +[2026-06-24 03:46:21,834][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.0618932880461216, 'train/grad_norm_pre_clip': 14.434589385986328, 'timing/data': 0.0005173539975658059, 'timing/model': 1.12394379300531, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 1.55}) +[2026-06-24 03:46:25,926][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.04866338148713112, 'train/grad_norm_pre_clip': 30.7738094329834, 'timing/data': 0.0007356069982051849, 'timing/model': 1.1218329028924927, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 1.55}) +[2026-06-24 03:46:30,008][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.1037609651684761, 'train/grad_norm_pre_clip': 33.48751449584961, 'timing/data': 0.00036963599268347025, 'timing/model': 1.1173326800344512, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 1.55}) +[2026-06-24 03:46:34,157][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.006885130889713764, 'train/grad_norm_pre_clip': 22.040830612182617, 'timing/data': 0.0003676259657368064, 'timing/model': 1.1202752930112183, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 1.55}) +[2026-06-24 03:46:38,245][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.03611796721816063, 'train/grad_norm_pre_clip': 21.13633918762207, 'timing/data': 0.00037594104651361704, 'timing/model': 1.1239096249919385, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 1.55}) +[2026-06-24 03:46:42,326][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4000 +[2026-06-24 03:47:46,589][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4000 +[2026-06-24 03:47:46,591][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4000 +[2026-06-24 03:48:25,190][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4000 +[2026-06-24 04:00:17,832][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000 +[2026-06-24 04:00:17,832][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.0640961229801178, 'train/grad_norm_pre_clip': 24.18770980834961, 'eval/action_loss/seconds': 64.26074431999587, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.030285008251667023, 'eval/latency_2/loss': 0.030285008251667023, 'eval/flappy/loss': 0.030285008251667023, 'eval/flappy/latency_2/loss': 0.030285008251667023, 'eval/flap_precision': 0.9794238683127572, 'eval/flap_recall': 0.9687924016282226, 'eval/flap_f1': 0.9740791268758526, 'eval/flap_rate_model': 0.2025, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9740791268758525, 'eval/latency_2/flap_precision': 0.9794238683127572, 'eval/latency_2/flap_recall': 0.9687924016282226, 'eval/latency_2/flap_f1': 0.9740791268758526, 'eval/latency_2/flap_rate_model': 0.2025, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9740791268758525, 'eval/action_classification/data_wait_seconds': 0.400161474943161, 'eval/action_classification/predict_seconds': 38.069427490234375, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 442.26000000002523, 'rl_games_eval/mid_train/mean_length': 3582.4, 'rl_games_eval/mid_train/std_reward': 4.680000000000791, 'rl_games_eval/mid_train/std_length': 35.2, 'rl_games_eval/mid_train/macro_mean_reward': 442.26000000002523, 'rl_games_eval/mid_train/macro_mean_length': 3582.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 442.26000000002523, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3582.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 4.680000000000791, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 35.2, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 442.26000000002523, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00036342302337288857, 'timing/model': 1.1181843219092116, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 1.55}) +[2026-06-24 04:00:17,833][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_4000_state +[2026-06-24 04:00:17,835][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 04:00:21,495][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_4000_state/model.safetensors +[2026-06-24 04:00:29,373][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_4000_state/optimizer.bin +[2026-06-24 04:00:29,373][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_4000_state/sampler.bin +[2026-06-24 04:00:29,373][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-24 04:00:29,374][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-24 04:00:29,481][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-24 04:00:29,481][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 04:00:29,497][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 04:00:33,618][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.08908436447381973, 'train/grad_norm_pre_clip': 40.680908203125, 'timing/data': 0.0003812320064753294, 'timing/model': 1.0918527329340577, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 1.55}) +[2026-06-24 04:00:37,603][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.01575128175318241, 'train/grad_norm_pre_clip': 15.082857131958008, 'timing/data': 0.0003681870875880122, 'timing/model': 1.1029840538976714, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 1.55}) +[2026-06-24 04:00:41,586][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.005571066867560148, 'train/grad_norm_pre_clip': 24.474891662597656, 'timing/data': 0.0005543559091165662, 'timing/model': 1.0959960849722847, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 1.55}) +[2026-06-24 04:00:45,577][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.007915223948657513, 'train/grad_norm_pre_clip': 24.99778938293457, 'timing/data': 0.0003651269944384694, 'timing/model': 1.0972417839802802, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 1.55}) +[2026-06-24 04:00:49,754][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.11327740550041199, 'train/grad_norm_pre_clip': 32.17508316040039, 'timing/data': 0.00036171404644846916, 'timing/model': 1.1016441049287096, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 1.55}) +[2026-06-24 04:00:53,767][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.017765680328011513, 'train/grad_norm_pre_clip': 12.512443542480469, 'timing/data': 0.0003708869917318225, 'timing/model': 1.1000213369261473, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 1.55}) +[2026-06-24 04:00:57,792][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.23987708985805511, 'train/grad_norm_pre_clip': 52.34549331665039, 'timing/data': 0.0003964200150221586, 'timing/model': 1.1044377719517797, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 1.55}) +[2026-06-24 04:01:01,811][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.012217730283737183, 'train/grad_norm_pre_clip': 12.40996265411377, 'timing/data': 0.0004062069347128272, 'timing/model': 1.099609182914719, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 1.55}) +[2026-06-24 04:01:05,994][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.06828076392412186, 'train/grad_norm_pre_clip': 29.172428131103516, 'timing/data': 0.0003651080187410116, 'timing/model': 1.1026636419119313, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 1.55}) +[2026-06-24 04:01:10,039][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.030220460146665573, 'train/grad_norm_pre_clip': 25.221736907958984, 'timing/data': 0.00036119495052844286, 'timing/model': 1.1078702399972826, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 1.55}) +[2026-06-24 04:01:14,087][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.11052016168832779, 'train/grad_norm_pre_clip': 20.486919403076172, 'timing/data': 0.0006168069085106254, 'timing/model': 1.112804337986745, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 1.55}) +[2026-06-24 04:01:18,111][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.14150026440620422, 'train/grad_norm_pre_clip': 19.1033878326416, 'timing/data': 0.0004144989652559161, 'timing/model': 1.1017329060705379, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 1.55}) +[2026-06-24 04:01:22,305][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.02684422768652439, 'train/grad_norm_pre_clip': 14.36817741394043, 'timing/data': 0.000554532976821065, 'timing/model': 1.1095980659592897, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 1.55}) +[2026-06-24 04:01:26,352][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.003008692292496562, 'train/grad_norm_pre_clip': 23.55408477783203, 'timing/data': 0.0005449149757623672, 'timing/model': 1.1085342310834676, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 1.55}) +[2026-06-24 04:01:30,390][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.024072561413049698, 'train/grad_norm_pre_clip': 17.572715759277344, 'timing/data': 0.0005239469464868307, 'timing/model': 1.102212199009955, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 1.55}) +[2026-06-24 04:01:34,429][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.008811191655695438, 'train/grad_norm_pre_clip': 14.322434425354004, 'timing/data': 0.0003547490341588855, 'timing/model': 1.1080543709686026, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 1.55}) +[2026-06-24 04:01:38,647][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.027606653049588203, 'train/grad_norm_pre_clip': 21.410953521728516, 'timing/data': 0.000363194034434855, 'timing/model': 1.1110828999662772, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 1.55}) +[2026-06-24 04:01:42,712][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.07689972966909409, 'train/grad_norm_pre_clip': 13.925810813903809, 'timing/data': 0.0005736079765483737, 'timing/model': 1.1121215299936011, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 1.56}) +[2026-06-24 04:01:46,807][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.04685664176940918, 'train/grad_norm_pre_clip': 14.981867790222168, 'timing/data': 0.0003578130854293704, 'timing/model': 1.1047102389857173, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 1.56}) +[2026-06-24 04:01:50,887][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.0208799596875906, 'train/grad_norm_pre_clip': 18.561874389648438, 'timing/data': 0.0005032080225646496, 'timing/model': 1.1099186309147626, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 1.56}) +[2026-06-24 04:01:55,084][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.17082726955413818, 'train/grad_norm_pre_clip': 35.63618850708008, 'timing/data': 0.0005772340809926391, 'timing/model': 1.1118582589551806, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 1.56}) +[2026-06-24 04:01:59,150][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.043035585433244705, 'train/grad_norm_pre_clip': 32.12538146972656, 'timing/data': 0.0004196219379082322, 'timing/model': 1.116592023987323, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 1.56}) +[2026-06-24 04:02:03,211][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.05344562232494354, 'train/grad_norm_pre_clip': 15.503989219665527, 'timing/data': 0.00041175202932208776, 'timing/model': 1.1141158499522135, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 1.56}) +[2026-06-24 04:02:07,285][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.10762955993413925, 'train/grad_norm_pre_clip': 24.78558349609375, 'timing/data': 0.0003931389655917883, 'timing/model': 1.1189691359177232, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 1.56}) +[2026-06-24 04:02:11,515][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.038325898349285126, 'train/grad_norm_pre_clip': 17.805152893066406, 'timing/data': 0.0003566729137673974, 'timing/model': 1.1193005930399522, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 1.56}) +[2026-06-24 04:02:15,586][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.046470072120428085, 'train/grad_norm_pre_clip': 15.826595306396484, 'timing/data': 0.00037131598219275475, 'timing/model': 1.1163566099712625, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 1.56}) +[2026-06-24 04:02:19,659][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.023673858493566513, 'train/grad_norm_pre_clip': 27.75486946105957, 'timing/data': 0.000357692944817245, 'timing/model': 1.1148364819819108, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 1.56}) +[2026-06-24 04:02:23,755][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.02073889784514904, 'train/grad_norm_pre_clip': 15.956870079040527, 'timing/data': 0.0004222070565447211, 'timing/model': 1.1325707850046456, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 1.56}) +[2026-06-24 04:02:27,985][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.02234770357608795, 'train/grad_norm_pre_clip': 11.689811706542969, 'timing/data': 0.00036260392516851425, 'timing/model': 1.1167131150141358, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 1.56}) +[2026-06-24 04:02:32,065][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.06112118810415268, 'train/grad_norm_pre_clip': 30.74883270263672, 'timing/data': 0.0003691150341182947, 'timing/model': 1.1205207289895043, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 1.56}) +[2026-06-24 04:02:36,149][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.009631291963160038, 'train/grad_norm_pre_clip': 17.780384063720703, 'timing/data': 0.00036625307984650135, 'timing/model': 1.1169337530154735, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 1.56}) +[2026-06-24 04:02:40,238][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.04167165607213974, 'train/grad_norm_pre_clip': 25.097047805786133, 'timing/data': 0.00040679890662431717, 'timing/model': 1.1235361519502476, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 1.56}) +[2026-06-24 04:02:44,444][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.07461906969547272, 'train/grad_norm_pre_clip': 17.814748764038086, 'timing/data': 0.000421590986661613, 'timing/model': 1.1180197600042447, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 1.56}) +[2026-06-24 04:02:48,520][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.07634333521127701, 'train/grad_norm_pre_clip': 27.47156524658203, 'timing/data': 0.0003648430574685335, 'timing/model': 1.1160622640745714, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 1.56}) +[2026-06-24 04:02:52,588][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.0328117236495018, 'train/grad_norm_pre_clip': 22.39570426940918, 'timing/data': 0.00035630795173346996, 'timing/model': 1.1128013570560142, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 1.56}) +[2026-06-24 04:02:56,655][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.019807962700724602, 'train/grad_norm_pre_clip': 23.863052368164062, 'timing/data': 0.000492447055876255, 'timing/model': 1.1102333549642935, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 1.56}) +[2026-06-24 04:03:00,871][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.03360087424516678, 'train/grad_norm_pre_clip': 16.77335548400879, 'timing/data': 0.0003651789156720042, 'timing/model': 1.11542759090662, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 1.56}) +[2026-06-24 04:03:04,955][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.06289217621088028, 'train/grad_norm_pre_clip': 15.757468223571777, 'timing/data': 0.00043931708205491304, 'timing/model': 1.1196536649949849, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 1.56}) +[2026-06-24 04:03:09,035][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.04618893191218376, 'train/grad_norm_pre_clip': 10.431412696838379, 'timing/data': 0.0003647028934210539, 'timing/model': 1.1154286400415003, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 1.56}) +[2026-06-24 04:03:13,112][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.14136295020580292, 'train/grad_norm_pre_clip': 42.85039138793945, 'timing/data': 0.0005130530335009098, 'timing/model': 1.1138194150989875, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 1.56}) +[2026-06-24 04:03:17,327][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.01948731020092964, 'train/grad_norm_pre_clip': 14.955641746520996, 'timing/data': 0.000409191008657217, 'timing/model': 1.108418775955215, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 1.56}) +[2026-06-24 04:03:21,392][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.044250547885894775, 'train/grad_norm_pre_clip': 16.70942497253418, 'timing/data': 0.0005405810661613941, 'timing/model': 1.1159005439840257, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 1.56}) +[2026-06-24 04:03:25,480][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.020408926531672478, 'train/grad_norm_pre_clip': 11.415324211120605, 'timing/data': 0.00045618496369570494, 'timing/model': 1.1195627989945933, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 1.56}) +[2026-06-24 04:03:29,563][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.08370112627744675, 'train/grad_norm_pre_clip': 26.96906089782715, 'timing/data': 0.0005677179433405399, 'timing/model': 1.1185870920307934, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 1.57}) +[2026-06-24 04:03:33,779][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.039790548384189606, 'train/grad_norm_pre_clip': 5.554074764251709, 'timing/data': 0.0004418479511514306, 'timing/model': 1.1116618139203638, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 1.57}) +[2026-06-24 04:03:37,855][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.05876928195357323, 'train/grad_norm_pre_clip': 24.23935890197754, 'timing/data': 0.0006610939744859934, 'timing/model': 1.1194157599238679, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 1.57}) +[2026-06-24 04:03:41,934][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.006143748294562101, 'train/grad_norm_pre_clip': 31.283222198486328, 'timing/data': 0.00043709902092814445, 'timing/model': 1.1173450840869918, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 1.57}) +[2026-06-24 04:03:46,015][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.04335121810436249, 'train/grad_norm_pre_clip': 18.6651611328125, 'timing/data': 0.00043438898865133524, 'timing/model': 1.117726195952855, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 1.57}) +[2026-06-24 04:03:50,252][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.018985973671078682, 'train/grad_norm_pre_clip': 13.909029006958008, 'timing/data': 0.00040742894634604454, 'timing/model': 1.118006260949187, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 1.57}) +[2026-06-24 04:03:54,332][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.03426358103752136, 'train/grad_norm_pre_clip': 19.209590911865234, 'timing/data': 0.00036036595702171326, 'timing/model': 1.1172997070243582, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 1.57}) +[2026-06-24 04:03:58,403][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.047334447503089905, 'train/grad_norm_pre_clip': 15.44164752960205, 'timing/data': 0.0004190249601379037, 'timing/model': 1.1160127049079165, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 1.57}) +[2026-06-24 04:04:02,479][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.06762636452913284, 'train/grad_norm_pre_clip': 15.68825912475586, 'timing/data': 0.00040506001096218824, 'timing/model': 1.114571409067139, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 1.57}) +[2026-06-24 04:04:06,700][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.0016661194385960698, 'train/grad_norm_pre_clip': 12.674206733703613, 'timing/data': 0.0004052789881825447, 'timing/model': 1.1139185320353135, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 1.57}) +[2026-06-24 04:04:10,778][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.06820304691791534, 'train/grad_norm_pre_clip': 20.559249877929688, 'timing/data': 0.00035883497912436724, 'timing/model': 1.1217586239799857, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 1.57}) +[2026-06-24 04:04:14,855][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.08310788869857788, 'train/grad_norm_pre_clip': 14.397541046142578, 'timing/data': 0.00038441503420472145, 'timing/model': 1.119550575967878, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 1.57}) +[2026-06-24 04:04:18,935][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.011184166185557842, 'train/grad_norm_pre_clip': 12.651021003723145, 'timing/data': 0.0004811449907720089, 'timing/model': 1.1153033270966262, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 1.57}) +[2026-06-24 04:04:23,157][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.016247352585196495, 'train/grad_norm_pre_clip': 7.2386908531188965, 'timing/data': 0.0003548659151419997, 'timing/model': 1.115167016047053, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 1.57}) +[2026-06-24 04:04:27,227][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.057635679841041565, 'train/grad_norm_pre_clip': 28.4528751373291, 'timing/data': 0.00036261300556361675, 'timing/model': 1.1156986569985747, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 1.57}) +[2026-06-24 04:04:31,295][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.0166127011179924, 'train/grad_norm_pre_clip': 18.794048309326172, 'timing/data': 0.00034777692053467035, 'timing/model': 1.1107281709555537, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 1.57}) +[2026-06-24 04:04:35,370][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.06125737354159355, 'train/grad_norm_pre_clip': 14.61489486694336, 'timing/data': 0.0003679980291053653, 'timing/model': 1.1154908810276538, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 1.57}) +[2026-06-24 04:04:39,592][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.020397793501615524, 'train/grad_norm_pre_clip': 11.133341789245605, 'timing/data': 0.0004674500087276101, 'timing/model': 1.1129193879896775, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 1.57}) +[2026-06-24 04:04:43,677][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.10755449533462524, 'train/grad_norm_pre_clip': 20.418888092041016, 'timing/data': 0.0003605580423027277, 'timing/model': 1.1148400800302625, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 1.57}) +[2026-06-24 04:04:47,756][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.02338140830397606, 'train/grad_norm_pre_clip': 23.163379669189453, 'timing/data': 0.0003557150484994054, 'timing/model': 1.117747921962291, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 1.57}) +[2026-06-24 04:04:51,841][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.012790479697287083, 'train/grad_norm_pre_clip': 20.775442123413086, 'timing/data': 0.0005598739953711629, 'timing/model': 1.1210480220615864, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 1.57}) +[2026-06-24 04:04:56,079][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.05517830699682236, 'train/grad_norm_pre_clip': 15.645936012268066, 'timing/data': 0.0003651640145108104, 'timing/model': 1.1273617669939995, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 1.57}) +[2026-06-24 04:05:00,176][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.0865577980875969, 'train/grad_norm_pre_clip': 18.584020614624023, 'timing/data': 0.0004098860081285238, 'timing/model': 1.122425558976829, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 1.57}) +[2026-06-24 04:05:04,234][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.050937049090862274, 'train/grad_norm_pre_clip': 27.263460159301758, 'timing/data': 0.0004189800238236785, 'timing/model': 1.1118715699994937, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 1.57}) +[2026-06-24 04:05:08,309][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.03136948496103287, 'train/grad_norm_pre_clip': 17.506732940673828, 'timing/data': 0.00041743810288608074, 'timing/model': 1.1138602489372715, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 1.57}) +[2026-06-24 04:05:12,522][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.007865614257752895, 'train/grad_norm_pre_clip': 13.9771728515625, 'timing/data': 0.0005160609725862741, 'timing/model': 1.1091626069974154, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 1.57}) +[2026-06-24 04:05:16,605][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.08596573024988174, 'train/grad_norm_pre_clip': 24.046737670898438, 'timing/data': 0.00041311897803097963, 'timing/model': 1.1192528869723901, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 1.58}) +[2026-06-24 04:05:20,687][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.21210229396820068, 'train/grad_norm_pre_clip': 46.163116455078125, 'timing/data': 0.0003732119221240282, 'timing/model': 1.1211194850038737, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 1.58}) +[2026-06-24 04:05:24,770][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.0030806483700871468, 'train/grad_norm_pre_clip': 4.378309726715088, 'timing/data': 0.0003690009471029043, 'timing/model': 1.118039295077324, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 1.58}) +[2026-06-24 04:05:28,996][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.0033085569739341736, 'train/grad_norm_pre_clip': 29.504329681396484, 'timing/data': 0.0004889760166406631, 'timing/model': 1.1219477989943698, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 1.58}) +[2026-06-24 04:05:33,102][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.002876301296055317, 'train/grad_norm_pre_clip': 52.19462966918945, 'timing/data': 0.0003524799831211567, 'timing/model': 1.1249126819893718, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 1.58}) +[2026-06-24 04:05:37,187][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.003284790087491274, 'train/grad_norm_pre_clip': 16.006507873535156, 'timing/data': 0.00040169793646782637, 'timing/model': 1.1201033890247345, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 1.58}) +[2026-06-24 04:05:41,256][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.0014003729447722435, 'train/grad_norm_pre_clip': 11.136666297912598, 'timing/data': 0.00036725401878356934, 'timing/model': 1.1136347319697961, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 1.58}) +[2026-06-24 04:05:45,482][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.01550509687513113, 'train/grad_norm_pre_clip': 8.913671493530273, 'timing/data': 0.0003650600556284189, 'timing/model': 1.1195644479012117, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 1.58}) +[2026-06-24 04:05:49,570][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.0020852715242654085, 'train/grad_norm_pre_clip': 28.72484016418457, 'timing/data': 0.00042861897964030504, 'timing/model': 1.1181026779813692, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 1.58}) +[2026-06-24 04:05:53,653][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.004948948509991169, 'train/grad_norm_pre_clip': 9.448973655700684, 'timing/data': 0.0003630609717220068, 'timing/model': 1.1233755679568276, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 1.58}) +[2026-06-24 04:05:57,740][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.004261688329279423, 'train/grad_norm_pre_clip': 12.671663284301758, 'timing/data': 0.0003947940422222018, 'timing/model': 1.1242931300075725, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 1.58}) +[2026-06-24 04:06:01,976][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.0075484514236450195, 'train/grad_norm_pre_clip': 24.14752769470215, 'timing/data': 0.0005661159520968795, 'timing/model': 1.119373460067436, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 1.58}) +[2026-06-24 04:06:06,059][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.00804350059479475, 'train/grad_norm_pre_clip': 33.02888107299805, 'timing/data': 0.000522851012647152, 'timing/model': 1.1160544030135497, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 1.58}) +[2026-06-24 04:06:10,137][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.0953230932354927, 'train/grad_norm_pre_clip': 34.248077392578125, 'timing/data': 0.0005019039381295443, 'timing/model': 1.1211196050280705, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 1.58}) +[2026-06-24 04:06:14,236][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.009011346846818924, 'train/grad_norm_pre_clip': 7.64128303527832, 'timing/data': 0.0003568730317056179, 'timing/model': 1.1258255020948127, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 1.58}) +[2026-06-24 04:06:18,451][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.1438569277524948, 'train/grad_norm_pre_clip': 55.289695739746094, 'timing/data': 0.0005700809415429831, 'timing/model': 1.119715049979277, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 1.58}) +[2026-06-24 04:06:22,536][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.006007271818816662, 'train/grad_norm_pre_clip': 17.874038696289062, 'timing/data': 0.0003575329901650548, 'timing/model': 1.1226632049074396, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 1.58}) +[2026-06-24 04:06:26,613][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.1027461364865303, 'train/grad_norm_pre_clip': 35.734188079833984, 'timing/data': 0.0005800939397886395, 'timing/model': 1.1205526479752734, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 1.58}) +[2026-06-24 04:06:30,702][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.11687290668487549, 'train/grad_norm_pre_clip': 20.469158172607422, 'timing/data': 0.00041389698162674904, 'timing/model': 1.123409113031812, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 1.58}) +[2026-06-24 04:06:34,921][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.12564492225646973, 'train/grad_norm_pre_clip': 20.07680320739746, 'timing/data': 0.0005702289054170251, 'timing/model': 1.1107780730817467, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 1.58}) +[2026-06-24 04:06:39,020][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.0367196686565876, 'train/grad_norm_pre_clip': 26.432283401489258, 'timing/data': 0.00036726798862218857, 'timing/model': 1.1321384609909728, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 1.58}) +[2026-06-24 04:06:43,106][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.05727537348866463, 'train/grad_norm_pre_clip': 14.568913459777832, 'timing/data': 0.0006660319631919265, 'timing/model': 1.120685491943732, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 1.58}) +[2026-06-24 04:06:47,196][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.17203737795352936, 'train/grad_norm_pre_clip': 40.38627624511719, 'timing/data': 0.0003631738945841789, 'timing/model': 1.120061764959246, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 1.58}) +[2026-06-24 04:06:51,416][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.0005234475247561932, 'train/grad_norm_pre_clip': 23.404523849487305, 'timing/data': 0.0004968050634488463, 'timing/model': 1.1213600219925866, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 1.58}) +[2026-06-24 04:06:55,509][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.04460788518190384, 'train/grad_norm_pre_clip': 56.042503356933594, 'timing/data': 0.00037288200110197067, 'timing/model': 1.1228578899754211, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 1.58}) +[2026-06-24 04:06:59,593][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.04895232617855072, 'train/grad_norm_pre_clip': 61.694522857666016, 'timing/data': 0.000362226041033864, 'timing/model': 1.1140114800073206, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 1.58}) +[2026-06-24 04:07:03,677][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.009671998210251331, 'train/grad_norm_pre_clip': 5.922125339508057, 'timing/data': 0.000359299941919744, 'timing/model': 1.1184896680060774, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 1.59}) +[2026-06-24 04:07:07,908][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.10203684866428375, 'train/grad_norm_pre_clip': 45.0050048828125, 'timing/data': 0.0004195190267637372, 'timing/model': 1.1166732290294021, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 1.59}) +[2026-06-24 04:07:11,982][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.05943725258111954, 'train/grad_norm_pre_clip': 21.13991928100586, 'timing/data': 0.0003899770090356469, 'timing/model': 1.117034996044822, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 1.59}) +[2026-06-24 04:07:16,068][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.04274853318929672, 'train/grad_norm_pre_clip': 15.046873092651367, 'timing/data': 0.0003605450037866831, 'timing/model': 1.1197343769017607, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 1.59}) +[2026-06-24 04:07:20,148][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4100 +[2026-06-24 04:08:24,056][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4100 +[2026-06-24 04:08:24,057][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4100 +[2026-06-24 04:09:03,001][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4100 +[2026-06-24 04:09:03,002][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.0024231760762631893, 'train/grad_norm_pre_clip': 24.160478591918945, 'eval/action_loss/seconds': 63.90575000597164, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.028697602450847626, 'eval/latency_2/loss': 0.028697602450847626, 'eval/flappy/loss': 0.028697602450847626, 'eval/flappy/latency_2/loss': 0.028697602450847626, 'eval/flap_precision': 0.9795081967213115, 'eval/flap_recall': 0.9728629579375848, 'eval/flap_f1': 0.9761742682096666, 'eval/flap_rate_model': 0.20333333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9761742682096666, 'eval/latency_2/flap_precision': 0.9795081967213115, 'eval/latency_2/flap_recall': 0.9728629579375848, 'eval/latency_2/flap_f1': 0.9761742682096666, 'eval/latency_2/flap_rate_model': 0.20333333333333334, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9761742682096666, 'eval/action_classification/data_wait_seconds': 0.5428847074508667, 'eval/action_classification/predict_seconds': 38.25126266479492, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00037847505882382393, 'timing/model': 1.1159494429593906, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 1.59}) +[2026-06-24 04:09:07,062][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.01039858814328909, 'train/grad_norm_pre_clip': 16.202726364135742, 'timing/data': 0.0003629019483923912, 'timing/model': 1.1131488259416074, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 1.59}) +[2026-06-24 04:09:11,113][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.002062673680484295, 'train/grad_norm_pre_clip': 3.681825637817383, 'timing/data': 0.00036217295564711094, 'timing/model': 1.1096398899098858, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 1.59}) +[2026-06-24 04:09:15,169][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.016251791268587112, 'train/grad_norm_pre_clip': 18.395477294921875, 'timing/data': 0.0005686189979314804, 'timing/model': 1.1069169839611277, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 1.59}) +[2026-06-24 04:09:19,330][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.0611998625099659, 'train/grad_norm_pre_clip': 22.71445083618164, 'timing/data': 0.000489416066557169, 'timing/model': 1.1122573700267822, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 1.59}) +[2026-06-24 04:09:23,377][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.05391504615545273, 'train/grad_norm_pre_clip': 18.1420841217041, 'timing/data': 0.000365164945833385, 'timing/model': 1.1079872369300574, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 1.59}) +[2026-06-24 04:09:27,433][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.024708569049835205, 'train/grad_norm_pre_clip': 29.891002655029297, 'timing/data': 0.0004312089877203107, 'timing/model': 1.1092610399937257, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 1.59}) +[2026-06-24 04:09:31,575][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.047465428709983826, 'train/grad_norm_pre_clip': 36.44675064086914, 'timing/data': 0.00035735091660171747, 'timing/model': 1.1851991229923442, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 1.59}) +[2026-06-24 04:09:35,643][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.03280850127339363, 'train/grad_norm_pre_clip': 27.210500717163086, 'timing/data': 0.00037396897096186876, 'timing/model': 1.1083400539355353, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 1.59}) +[2026-06-24 04:09:39,710][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.08150657266378403, 'train/grad_norm_pre_clip': 52.30813217163086, 'timing/data': 0.00038248696364462376, 'timing/model': 1.1174646739382297, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 1.59}) +[2026-06-24 04:09:43,788][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.01779058203101158, 'train/grad_norm_pre_clip': 16.119531631469727, 'timing/data': 0.00035559607204049826, 'timing/model': 1.122619642992504, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 1.59}) +[2026-06-24 04:09:47,857][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.00025737384567037225, 'train/grad_norm_pre_clip': 7.7598185539245605, 'timing/data': 0.00046084693167358637, 'timing/model': 1.11503611295484, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 1.59}) +[2026-06-24 04:09:52,043][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.07448413968086243, 'train/grad_norm_pre_clip': 35.06833267211914, 'timing/data': 0.00038797606248408556, 'timing/model': 1.1172981290146708, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 1.59}) +[2026-06-24 04:09:56,107][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.02718566171824932, 'train/grad_norm_pre_clip': 18.305084228515625, 'timing/data': 0.00035948201548308134, 'timing/model': 1.1152281160466373, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 1.59}) +[2026-06-24 04:10:00,184][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.0005169014330022037, 'train/grad_norm_pre_clip': 24.599510192871094, 'timing/data': 0.0003609569976106286, 'timing/model': 1.1207477919524536, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 1.59}) +[2026-06-24 04:10:04,259][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 9.744919225340709e-05, 'train/grad_norm_pre_clip': 42.66706466674805, 'timing/data': 0.0005205569323152304, 'timing/model': 1.119966231053695, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 1.59}) +[2026-06-24 04:10:08,468][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.013428152538836002, 'train/grad_norm_pre_clip': 15.167489051818848, 'timing/data': 0.00035138195380568504, 'timing/model': 1.1101491139270365, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 1.59}) +[2026-06-24 04:10:12,547][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.05053224787116051, 'train/grad_norm_pre_clip': 34.33079147338867, 'timing/data': 0.000358636025339365, 'timing/model': 1.1160292549757287, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 1.59}) +[2026-06-24 04:10:16,623][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.06449597328901291, 'train/grad_norm_pre_clip': 40.08365249633789, 'timing/data': 0.00037023599725216627, 'timing/model': 1.1144256710540503, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 1.59}) +[2026-06-24 04:10:20,695][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.17757302522659302, 'train/grad_norm_pre_clip': 39.8671989440918, 'timing/data': 0.0005088479956611991, 'timing/model': 1.1126928119920194, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 1.59}) +[2026-06-24 04:10:24,824][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.025388820096850395, 'train/grad_norm_pre_clip': 27.644195556640625, 'timing/data': 0.0003911820240318775, 'timing/model': 1.1174284649314359, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 1.59}) +[2026-06-24 04:10:28,901][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.008165141567587852, 'train/grad_norm_pre_clip': 31.927719116210938, 'timing/data': 0.000368358101695776, 'timing/model': 1.11782331392169, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 1.59}) +[2026-06-24 04:10:32,982][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.010456598363816738, 'train/grad_norm_pre_clip': 25.49140739440918, 'timing/data': 0.0005644200136885047, 'timing/model': 1.115113486070186, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 1.6}) +[2026-06-24 04:10:37,060][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.07181254029273987, 'train/grad_norm_pre_clip': 20.5407657623291, 'timing/data': 0.0004934860626235604, 'timing/model': 1.1161235891049728, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 1.6}) +[2026-06-24 04:10:41,242][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.0678393691778183, 'train/grad_norm_pre_clip': 25.000926971435547, 'timing/data': 0.0003684000112116337, 'timing/model': 1.1373795859981328, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 1.6}) +[2026-06-24 04:10:45,357][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.005342116579413414, 'train/grad_norm_pre_clip': 35.511077880859375, 'timing/data': 0.00038198393303900957, 'timing/model': 1.1377957839285955, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 1.6}) +[2026-06-24 04:10:49,451][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.01792878471314907, 'train/grad_norm_pre_clip': 23.92854118347168, 'timing/data': 0.0005780550418421626, 'timing/model': 1.1213164060609415, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 1.6}) +[2026-06-24 04:10:53,538][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.17837931215763092, 'train/grad_norm_pre_clip': 34.99864196777344, 'timing/data': 0.0003710649907588959, 'timing/model': 1.1173822269774973, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 1.6}) +[2026-06-24 04:10:57,795][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.004251166712492704, 'train/grad_norm_pre_clip': 9.879632949829102, 'timing/data': 0.0004068890120834112, 'timing/model': 1.115115814958699, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 1.6}) +[2026-06-24 04:11:01,876][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.1345823109149933, 'train/grad_norm_pre_clip': 40.86085510253906, 'timing/data': 0.0003575809532776475, 'timing/model': 1.118579224916175, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 1.6}) +[2026-06-24 04:11:05,972][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.002615823410451412, 'train/grad_norm_pre_clip': 8.95590877532959, 'timing/data': 0.0003645110409706831, 'timing/model': 1.1214483549119905, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 1.6}) +[2026-06-24 04:11:10,069][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.06713292747735977, 'train/grad_norm_pre_clip': 54.62250900268555, 'timing/data': 0.00035935803316533566, 'timing/model': 1.1282413660082966, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 1.6}) +[2026-06-24 04:11:14,348][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.01086561381816864, 'train/grad_norm_pre_clip': 20.08608055114746, 'timing/data': 0.00036504992749542, 'timing/model': 1.1144326170906425, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 1.6}) +[2026-06-24 04:11:18,441][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.016331180930137634, 'train/grad_norm_pre_clip': 10.500072479248047, 'timing/data': 0.00036333699245005846, 'timing/model': 1.1175352421123534, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 1.6}) +[2026-06-24 04:11:22,514][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.010890173725783825, 'train/grad_norm_pre_clip': 21.905778884887695, 'timing/data': 0.00043895095586776733, 'timing/model': 1.113720126915723, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 1.6}) +[2026-06-24 04:11:26,607][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.00955730490386486, 'train/grad_norm_pre_clip': 22.788759231567383, 'timing/data': 0.00038173189386725426, 'timing/model': 1.1216292750323191, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 1.6}) +[2026-06-24 04:11:30,875][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.0958399549126625, 'train/grad_norm_pre_clip': 16.512224197387695, 'timing/data': 0.0003601780626922846, 'timing/model': 1.1312790140509605, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 1.6}) +[2026-06-24 04:11:34,949][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.0304064080119133, 'train/grad_norm_pre_clip': 16.10493278503418, 'timing/data': 0.0005802110536023974, 'timing/model': 1.1181994150392711, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 1.6}) +[2026-06-24 04:11:39,067][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.060436882078647614, 'train/grad_norm_pre_clip': 27.42681121826172, 'timing/data': 0.00036931096110492945, 'timing/model': 1.1249831271125004, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 1.6}) +[2026-06-24 04:11:43,153][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.031761534512043, 'train/grad_norm_pre_clip': 24.661985397338867, 'timing/data': 0.0004841579357162118, 'timing/model': 1.118840478011407, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 1.6}) +[2026-06-24 04:11:47,412][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.10886283218860626, 'train/grad_norm_pre_clip': 24.931528091430664, 'timing/data': 0.0003708499716594815, 'timing/model': 1.1210062491009012, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 1.6}) +[2026-06-24 04:11:51,505][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.0003812833165284246, 'train/grad_norm_pre_clip': 26.461912155151367, 'timing/data': 0.0003671760205179453, 'timing/model': 1.1216334709897637, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 1.6}) +[2026-06-24 04:11:55,597][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.00270421477034688, 'train/grad_norm_pre_clip': 19.418193817138672, 'timing/data': 0.0005517679965123534, 'timing/model': 1.1176924599567428, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 1.6}) +[2026-06-24 04:11:59,674][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.20641258358955383, 'train/grad_norm_pre_clip': 27.063230514526367, 'timing/data': 0.0003671899903565645, 'timing/model': 1.1147554739145562, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 1.6}) +[2026-06-24 04:12:03,929][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.008587244898080826, 'train/grad_norm_pre_clip': 17.38144874572754, 'timing/data': 0.0004149950109422207, 'timing/model': 1.1183457180159166, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 1.6}) +[2026-06-24 04:12:08,014][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.031001729890704155, 'train/grad_norm_pre_clip': 24.509597778320312, 'timing/data': 0.00036279100459069014, 'timing/model': 1.1211387959774584, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 1.6}) +[2026-06-24 04:12:12,116][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.0918741226196289, 'train/grad_norm_pre_clip': 30.702482223510742, 'timing/data': 0.0006096939323469996, 'timing/model': 1.1181048769503832, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 1.6}) +[2026-06-24 04:12:16,209][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.014176095835864544, 'train/grad_norm_pre_clip': 25.31468963623047, 'timing/data': 0.00035989400930702686, 'timing/model': 1.1223732630023733, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 1.6}) +[2026-06-24 04:12:20,446][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.017310084775090218, 'train/grad_norm_pre_clip': 37.36682891845703, 'timing/data': 0.00045866798609495163, 'timing/model': 1.1190327709773555, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 1.61}) +[2026-06-24 04:12:24,521][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.13599710166454315, 'train/grad_norm_pre_clip': 20.1595516204834, 'timing/data': 0.0004141660174354911, 'timing/model': 1.1184678639983758, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 1.61}) +[2026-06-24 04:12:28,607][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.051633793860673904, 'train/grad_norm_pre_clip': 22.645112991333008, 'timing/data': 0.0003614759771153331, 'timing/model': 1.11792506207712, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 1.61}) +[2026-06-24 04:12:32,695][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.11660046130418777, 'train/grad_norm_pre_clip': 25.41668701171875, 'timing/data': 0.0005781910149380565, 'timing/model': 1.1107413599966094, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 1.61}) +[2026-06-24 04:12:36,959][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.0008336093160323799, 'train/grad_norm_pre_clip': 12.03696060180664, 'timing/data': 0.0003641500370576978, 'timing/model': 1.1179380919784307, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 1.61}) +[2026-06-24 04:12:41,039][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.04298411309719086, 'train/grad_norm_pre_clip': 16.2320613861084, 'timing/data': 0.0005118270637467504, 'timing/model': 1.116556093096733, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 1.61}) +[2026-06-24 04:12:45,122][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.031221358105540276, 'train/grad_norm_pre_clip': 9.9855318069458, 'timing/data': 0.0004532219609245658, 'timing/model': 1.1203836400527507, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 1.61}) +[2026-06-24 04:12:49,205][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.04525991529226303, 'train/grad_norm_pre_clip': 11.738834381103516, 'timing/data': 0.0003816980170086026, 'timing/model': 1.1177494609728456, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 1.61}) +[2026-06-24 04:12:53,453][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.10441476106643677, 'train/grad_norm_pre_clip': 24.004169464111328, 'timing/data': 0.0005397649947553873, 'timing/model': 1.1186767079634592, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 1.61}) +[2026-06-24 04:12:57,532][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.11510685831308365, 'train/grad_norm_pre_clip': 24.890214920043945, 'timing/data': 0.0004972969181835651, 'timing/model': 1.1167447259649634, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 1.61}) +[2026-06-24 04:13:01,633][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.02185857854783535, 'train/grad_norm_pre_clip': 19.08261489868164, 'timing/data': 0.00036629592068493366, 'timing/model': 1.1199398019816726, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 1.61}) +[2026-06-24 04:13:05,728][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.01837185211479664, 'train/grad_norm_pre_clip': 36.47894287109375, 'timing/data': 0.0005307700484991074, 'timing/model': 1.1191324200481176, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 1.61}) +[2026-06-24 04:13:09,986][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.2035679817199707, 'train/grad_norm_pre_clip': 19.687522888183594, 'timing/data': 0.0003986790543422103, 'timing/model': 1.1148643699707463, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 1.61}) +[2026-06-24 04:13:14,082][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.00035485622356645763, 'train/grad_norm_pre_clip': 30.916095733642578, 'timing/data': 0.0005748410476371646, 'timing/model': 1.1284900419414043, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 1.61}) +[2026-06-24 04:13:18,164][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.0689254179596901, 'train/grad_norm_pre_clip': 23.38039207458496, 'timing/data': 0.0005255379946902394, 'timing/model': 1.1204124740324914, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 1.61}) +[2026-06-24 04:13:22,252][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.010242755524814129, 'train/grad_norm_pre_clip': 5.437612533569336, 'timing/data': 0.000405161059461534, 'timing/model': 1.1240892149508, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 1.61}) +[2026-06-24 04:13:26,530][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.006697887554764748, 'train/grad_norm_pre_clip': 13.151165962219238, 'timing/data': 0.00040103495121002197, 'timing/model': 1.1194389279698953, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 1.61}) +[2026-06-24 04:13:30,632][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.06409084796905518, 'train/grad_norm_pre_clip': 10.879511833190918, 'timing/data': 0.0003548369277268648, 'timing/model': 1.1241697201039642, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 1.61}) +[2026-06-24 04:13:34,720][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.01043514534831047, 'train/grad_norm_pre_clip': 19.79920196533203, 'timing/data': 0.00035718397703021765, 'timing/model': 1.121645299019292, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 1.61}) +[2026-06-24 04:13:38,962][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.030608125030994415, 'train/grad_norm_pre_clip': 20.133699417114258, 'timing/data': 0.00035886105615645647, 'timing/model': 1.1195012290263548, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 1.61}) +[2026-06-24 04:13:43,236][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.0023464050609618425, 'train/grad_norm_pre_clip': 11.327220916748047, 'timing/data': 0.0003721240209415555, 'timing/model': 1.119102479889989, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 1.61}) +[2026-06-24 04:13:47,343][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.04535945504903793, 'train/grad_norm_pre_clip': 16.523927688598633, 'timing/data': 0.00035773904528468847, 'timing/model': 1.12443641689606, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 1.61}) +[2026-06-24 04:13:51,434][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.04043972119688988, 'train/grad_norm_pre_clip': 25.5633602142334, 'timing/data': 0.0003609220730140805, 'timing/model': 1.1216155110159889, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 1.61}) +[2026-06-24 04:13:55,516][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.01301634218543768, 'train/grad_norm_pre_clip': 6.362481594085693, 'timing/data': 0.00044582190457731485, 'timing/model': 1.118786255014129, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 1.61}) +[2026-06-24 04:13:59,770][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.025467155501246452, 'train/grad_norm_pre_clip': 18.532880783081055, 'timing/data': 0.0005559389246627688, 'timing/model': 1.1198667930439115, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 1.61}) +[2026-06-24 04:14:03,850][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.16648872196674347, 'train/grad_norm_pre_clip': 27.177459716796875, 'timing/data': 0.0004129259614273906, 'timing/model': 1.117443801020272, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 1.62}) +[2026-06-24 04:14:07,940][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.0023145200684666634, 'train/grad_norm_pre_clip': 23.9267635345459, 'timing/data': 0.00039406795985996723, 'timing/model': 1.120938281994313, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 1.62}) +[2026-06-24 04:14:12,021][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.011495467275381088, 'train/grad_norm_pre_clip': 7.8505072593688965, 'timing/data': 0.00035847199615091085, 'timing/model': 1.1173950840020552, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 1.62}) +[2026-06-24 04:14:16,292][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.10478803515434265, 'train/grad_norm_pre_clip': 29.754756927490234, 'timing/data': 0.00036219600588083267, 'timing/model': 1.1225563059560955, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 1.62}) +[2026-06-24 04:14:20,383][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.022683115676045418, 'train/grad_norm_pre_clip': 27.3771915435791, 'timing/data': 0.0003787620225921273, 'timing/model': 1.1191789779113606, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 1.62}) +[2026-06-24 04:14:24,472][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.03635980933904648, 'train/grad_norm_pre_clip': 20.557767868041992, 'timing/data': 0.0004089880967512727, 'timing/model': 1.1243189310189337, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 1.62}) +[2026-06-24 04:14:28,558][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.07879063487052917, 'train/grad_norm_pre_clip': 20.958829879760742, 'timing/data': 0.00036003999412059784, 'timing/model': 1.1173959959996864, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 1.62}) +[2026-06-24 04:14:32,817][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.21910639107227325, 'train/grad_norm_pre_clip': 46.06531524658203, 'timing/data': 0.00035680097062140703, 'timing/model': 1.11543085006997, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 1.62}) +[2026-06-24 04:14:36,903][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.005700479261577129, 'train/grad_norm_pre_clip': 17.430295944213867, 'timing/data': 0.0003633190644904971, 'timing/model': 1.1199455249588937, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 1.62}) +[2026-06-24 04:14:40,984][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.07226806879043579, 'train/grad_norm_pre_clip': 38.510921478271484, 'timing/data': 0.0003596539609134197, 'timing/model': 1.117310410947539, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 1.62}) +[2026-06-24 04:14:45,069][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.023916179314255714, 'train/grad_norm_pre_clip': 6.653261661529541, 'timing/data': 0.000384898972697556, 'timing/model': 1.121951380977407, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 1.62}) +[2026-06-24 04:14:49,313][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.008416173048317432, 'train/grad_norm_pre_clip': 13.895123481750488, 'timing/data': 0.00041403796058148146, 'timing/model': 1.114615355967544, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 1.62}) +[2026-06-24 04:14:53,396][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.050534237176179886, 'train/grad_norm_pre_clip': 38.84651184082031, 'timing/data': 0.00036120100412517786, 'timing/model': 1.1184735799906775, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 1.62}) +[2026-06-24 04:14:57,481][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.012595221400260925, 'train/grad_norm_pre_clip': 15.038066864013672, 'timing/data': 0.00040017394348978996, 'timing/model': 1.11926669604145, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 1.62}) +[2026-06-24 04:15:01,576][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.02167489007115364, 'train/grad_norm_pre_clip': 14.635568618774414, 'timing/data': 0.0003608139231801033, 'timing/model': 1.1146578619955108, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 1.62}) +[2026-06-24 04:15:05,835][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.0011748827528208494, 'train/grad_norm_pre_clip': 9.951253890991211, 'timing/data': 0.00037905992940068245, 'timing/model': 1.1196209759218618, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 1.62}) +[2026-06-24 04:15:09,916][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.020681107416749, 'train/grad_norm_pre_clip': 19.41312599182129, 'timing/data': 0.00037004495970904827, 'timing/model': 1.1163213370600715, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 1.62}) +[2026-06-24 04:15:13,999][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.04226530343294144, 'train/grad_norm_pre_clip': 14.926005363464355, 'timing/data': 0.00035051198210567236, 'timing/model': 1.1201566490344703, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 1.62}) +[2026-06-24 04:15:18,094][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.0318056121468544, 'train/grad_norm_pre_clip': 16.96636390686035, 'timing/data': 0.0003571889828890562, 'timing/model': 1.1198630490107462, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 1.62}) +[2026-06-24 04:15:22,348][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.06526859849691391, 'train/grad_norm_pre_clip': 12.608030319213867, 'timing/data': 0.0003707469440996647, 'timing/model': 1.1181525860447437, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 1.62}) +[2026-06-24 04:15:26,451][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.0023215431720018387, 'train/grad_norm_pre_clip': 28.90848159790039, 'timing/data': 0.00041809806134551764, 'timing/model': 1.1157543259905651, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 1.62}) +[2026-06-24 04:15:30,542][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.11759941279888153, 'train/grad_norm_pre_clip': 18.734375, 'timing/data': 0.0003580519696697593, 'timing/model': 1.116079714964144, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 1.62}) +[2026-06-24 04:15:34,625][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.07970095425844193, 'train/grad_norm_pre_clip': 20.021265029907227, 'timing/data': 0.0005784030072391033, 'timing/model': 1.1162041529314592, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 1.62}) +[2026-06-24 04:15:38,886][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.005981771741062403, 'train/grad_norm_pre_clip': 14.570528984069824, 'timing/data': 0.00040882406756281853, 'timing/model': 1.1218607380287722, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 1.62}) +[2026-06-24 04:15:42,972][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.038803763687610626, 'train/grad_norm_pre_clip': 17.60425567626953, 'timing/data': 0.0004025619709864259, 'timing/model': 1.117983889998868, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 1.62}) +[2026-06-24 04:15:47,076][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.008346130140125751, 'train/grad_norm_pre_clip': 25.50192642211914, 'timing/data': 0.00041150394827127457, 'timing/model': 1.1221654190449044, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 1.62}) +[2026-06-24 04:15:51,161][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.0001331216626567766, 'train/grad_norm_pre_clip': 3.4418864250183105, 'timing/data': 0.00041780900210142136, 'timing/model': 1.1193051119334996, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 1.63}) +[2026-06-24 04:15:55,417][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4200 +[2026-06-24 04:16:59,632][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4200 +[2026-06-24 04:16:59,633][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4200 +[2026-06-24 04:17:38,462][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4200 +[2026-06-24 04:17:38,463][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.2176257073879242, 'train/grad_norm_pre_clip': 34.47507858276367, 'eval/action_loss/seconds': 64.21282277198043, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.027385331690311432, 'eval/latency_2/loss': 0.027385331690311432, 'eval/flappy/loss': 0.027385331690311432, 'eval/flappy/latency_2/loss': 0.027385331690311432, 'eval/flap_precision': 0.9783491204330176, 'eval/flap_recall': 0.9810040705563093, 'eval/flap_f1': 0.9796747967479675, 'eval/flap_rate_model': 0.20527777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9796747967479675, 'eval/latency_2/flap_precision': 0.9783491204330176, 'eval/latency_2/flap_recall': 0.9810040705563093, 'eval/latency_2/flap_f1': 0.9796747967479675, 'eval/latency_2/flap_rate_model': 0.20527777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9796747967479675, 'eval/action_classification/data_wait_seconds': 0.6074236035346985, 'eval/action_classification/predict_seconds': 38.086849212646484, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00042395596392452717, 'timing/model': 1.1207999429898337, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 1.63}) +[2026-06-24 04:17:42,513][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.09476924687623978, 'train/grad_norm_pre_clip': 30.764816284179688, 'timing/data': 0.0003688590368255973, 'timing/model': 1.1049112069886178, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 1.63}) +[2026-06-24 04:17:46,561][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.030041389167308807, 'train/grad_norm_pre_clip': 13.571877479553223, 'timing/data': 0.00036843493580818176, 'timing/model': 1.109611852094531, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 1.63}) +[2026-06-24 04:17:50,766][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.03559132665395737, 'train/grad_norm_pre_clip': 27.350135803222656, 'timing/data': 0.00041343795601278543, 'timing/model': 1.264838643022813, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 1.63}) +[2026-06-24 04:17:54,825][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.008699924685060978, 'train/grad_norm_pre_clip': 5.060422420501709, 'timing/data': 0.0003641160437837243, 'timing/model': 1.1033814460970461, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 1.63}) +[2026-06-24 04:17:58,872][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.05235617980360985, 'train/grad_norm_pre_clip': 11.672865867614746, 'timing/data': 0.0003700359957292676, 'timing/model': 1.1119529419811442, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 1.63}) +[2026-06-24 04:18:02,921][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.007098184432834387, 'train/grad_norm_pre_clip': 12.054121971130371, 'timing/data': 0.0003932960098609328, 'timing/model': 1.108054682961665, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 1.63}) +[2026-06-24 04:18:07,133][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.01152065210044384, 'train/grad_norm_pre_clip': 9.958354949951172, 'timing/data': 0.0005062619457021356, 'timing/model': 1.265087469946593, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 1.63}) +[2026-06-24 04:18:11,187][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.05689632147550583, 'train/grad_norm_pre_clip': 16.16817855834961, 'timing/data': 0.0004203669959679246, 'timing/model': 1.1121638480108231, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 1.63}) +[2026-06-24 04:18:15,248][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.009532383643090725, 'train/grad_norm_pre_clip': 16.342199325561523, 'timing/data': 0.00036229495890438557, 'timing/model': 1.117069466970861, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 1.63}) +[2026-06-24 04:18:19,311][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.01197998970746994, 'train/grad_norm_pre_clip': 11.930130958557129, 'timing/data': 0.0005002290708944201, 'timing/model': 1.110872886958532, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 1.63}) +[2026-06-24 04:18:23,519][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.006318387575447559, 'train/grad_norm_pre_clip': 18.698932647705078, 'timing/data': 0.00037029001396149397, 'timing/model': 1.2637874039355665, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 1.63}) +[2026-06-24 04:18:27,579][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.1900315135717392, 'train/grad_norm_pre_clip': 31.97429084777832, 'timing/data': 0.00035860901698470116, 'timing/model': 1.1100692000472918, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 1.63}) +[2026-06-24 04:18:31,649][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.01988542638719082, 'train/grad_norm_pre_clip': 12.039337158203125, 'timing/data': 0.0003561620833352208, 'timing/model': 1.1160104719456285, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 1.63}) +[2026-06-24 04:18:35,721][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.00029397872276604176, 'train/grad_norm_pre_clip': 11.90398120880127, 'timing/data': 0.0005678490269929171, 'timing/model': 1.1166168649215251, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 1.63}) +[2026-06-24 04:18:39,938][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.042807530611753464, 'train/grad_norm_pre_clip': 28.26906967163086, 'timing/data': 0.00036866904702037573, 'timing/model': 1.2590989200398326, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 1.63}) +[2026-06-24 04:18:44,011][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.008449717424809933, 'train/grad_norm_pre_clip': 14.523563385009766, 'timing/data': 0.000502872047945857, 'timing/model': 1.1122104119276628, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 1.63}) +[2026-06-24 04:18:48,065][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.015633346512913704, 'train/grad_norm_pre_clip': 19.485328674316406, 'timing/data': 0.0003587440587580204, 'timing/model': 1.1090192289557308, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 1.63}) +[2026-06-24 04:18:52,150][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.007993689738214016, 'train/grad_norm_pre_clip': 21.682445526123047, 'timing/data': 0.00035272492095828056, 'timing/model': 1.1214435290312394, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 1.63}) +[2026-06-24 04:18:56,364][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.06159934028983116, 'train/grad_norm_pre_clip': 25.08650016784668, 'timing/data': 0.0003559570759534836, 'timing/model': 1.26613003492821, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 1.63}) +[2026-06-24 04:19:00,448][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.1034979447722435, 'train/grad_norm_pre_clip': 30.47283363342285, 'timing/data': 0.00043935689609497786, 'timing/model': 1.1181610939092934, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 1.63}) +[2026-06-24 04:19:04,516][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.023649301379919052, 'train/grad_norm_pre_clip': 24.768146514892578, 'timing/data': 0.0005439589731395245, 'timing/model': 1.111970296013169, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 1.63}) +[2026-06-24 04:19:08,603][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.07691426575183868, 'train/grad_norm_pre_clip': 24.703920364379883, 'timing/data': 0.000517925014719367, 'timing/model': 1.1154070380143821, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 1.63}) +[2026-06-24 04:19:12,820][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.0014451055321842432, 'train/grad_norm_pre_clip': 39.64520263671875, 'timing/data': 0.000521691981703043, 'timing/model': 1.264024903997779, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 1.63}) +[2026-06-24 04:19:16,897][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.02357523888349533, 'train/grad_norm_pre_clip': 28.41555404663086, 'timing/data': 0.00042019004467874765, 'timing/model': 1.1203831989550963, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 1.63}) +[2026-06-24 04:19:20,982][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.04338793084025383, 'train/grad_norm_pre_clip': 21.494794845581055, 'timing/data': 0.00036983995232731104, 'timing/model': 1.1161609010305256, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 1.64}) +[2026-06-24 04:19:25,055][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.0025591510348021984, 'train/grad_norm_pre_clip': 40.0236701965332, 'timing/data': 0.0005669109523296356, 'timing/model': 1.1192545980447903, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 1.64}) +[2026-06-24 04:19:29,296][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.010936633683741093, 'train/grad_norm_pre_clip': 43.39482879638672, 'timing/data': 0.00055196403991431, 'timing/model': 1.2644505510106683, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 1.64}) +[2026-06-24 04:19:33,382][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.014775834046304226, 'train/grad_norm_pre_clip': 23.698848724365234, 'timing/data': 0.0003584859659895301, 'timing/model': 1.1233947980217636, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 1.64}) +[2026-06-24 04:19:37,471][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.0025216059293597937, 'train/grad_norm_pre_clip': 22.695077896118164, 'timing/data': 0.0003686329582706094, 'timing/model': 1.1179828490130603, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 1.64}) +[2026-06-24 04:19:41,560][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.037995897233486176, 'train/grad_norm_pre_clip': 26.61016273498535, 'timing/data': 0.00036846904549747705, 'timing/model': 1.1255004239501432, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 1.64}) +[2026-06-24 04:19:45,790][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.0616803802549839, 'train/grad_norm_pre_clip': 44.61798858642578, 'timing/data': 0.0003752299817278981, 'timing/model': 1.2634487580507994, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 1.64}) +[2026-06-24 04:19:49,873][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.0879918709397316, 'train/grad_norm_pre_clip': 23.961515426635742, 'timing/data': 0.0003636570181697607, 'timing/model': 1.1193799859611318, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 1.64}) +[2026-06-24 04:19:53,955][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.1646171510219574, 'train/grad_norm_pre_clip': 36.17053985595703, 'timing/data': 0.0003685529809445143, 'timing/model': 1.1156283860327676, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 1.64}) +[2026-06-24 04:19:58,046][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.049084074795246124, 'train/grad_norm_pre_clip': 18.872472763061523, 'timing/data': 0.0005530680064111948, 'timing/model': 1.1211606319993734, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 1.64}) +[2026-06-24 04:20:02,296][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.057905565947294235, 'train/grad_norm_pre_clip': 20.49757957458496, 'timing/data': 0.00038406893145293, 'timing/model': 1.288307593902573, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 1.64}) +[2026-06-24 04:20:06,366][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.06469348073005676, 'train/grad_norm_pre_clip': 28.27606201171875, 'timing/data': 0.00038478709757328033, 'timing/model': 1.1170997510198504, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 1.64}) +[2026-06-24 04:20:10,451][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.02999451570212841, 'train/grad_norm_pre_clip': 25.53385353088379, 'timing/data': 0.00037832697853446007, 'timing/model': 1.1196558689698577, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 1.64}) +[2026-06-24 04:20:14,531][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.22396448254585266, 'train/grad_norm_pre_clip': 24.042699813842773, 'timing/data': 0.000428464962169528, 'timing/model': 1.1169743579812348, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 1.64}) +[2026-06-24 04:20:18,756][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.026350952684879303, 'train/grad_norm_pre_clip': 20.435199737548828, 'timing/data': 0.00036082102451473475, 'timing/model': 1.2611853838898242, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 1.64}) +[2026-06-24 04:20:22,836][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.008433716371655464, 'train/grad_norm_pre_clip': 12.71057415008545, 'timing/data': 0.0006185219390317798, 'timing/model': 1.1152354270452634, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 1.64}) +[2026-06-24 04:20:26,913][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.07810315489768982, 'train/grad_norm_pre_clip': 7.128443717956543, 'timing/data': 0.0003825780004262924, 'timing/model': 1.1199991189641878, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 1.64}) +[2026-06-24 04:20:31,001][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.06291724741458893, 'train/grad_norm_pre_clip': 29.231464385986328, 'timing/data': 0.00038455601315945387, 'timing/model': 1.1221919289091602, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 1.64}) +[2026-06-24 04:20:35,235][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.0033184923231601715, 'train/grad_norm_pre_clip': 6.179738521575928, 'timing/data': 0.00048720999620854855, 'timing/model': 1.2743752689566463, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 1.64}) +[2026-06-24 04:20:39,351][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.00447586877271533, 'train/grad_norm_pre_clip': 26.002349853515625, 'timing/data': 0.00039549602661281824, 'timing/model': 1.1373184979893267, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 1.64}) +[2026-06-24 04:20:43,440][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.022214049473404884, 'train/grad_norm_pre_clip': 19.880794525146484, 'timing/data': 0.00035796500742435455, 'timing/model': 1.1227777779567987, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 1.64}) +[2026-06-24 04:20:47,534][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.035610128194093704, 'train/grad_norm_pre_clip': 6.6030707359313965, 'timing/data': 0.00041901692748069763, 'timing/model': 1.1235895979916677, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 1.64}) +[2026-06-24 04:20:51,785][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.056143928319215775, 'train/grad_norm_pre_clip': 29.188819885253906, 'timing/data': 0.0003682549577206373, 'timing/model': 1.2849632250145078, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 1.64}) +[2026-06-24 04:20:55,854][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.06300990283489227, 'train/grad_norm_pre_clip': 11.026774406433105, 'timing/data': 0.0003565809456631541, 'timing/model': 1.114948033937253, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 1.64}) +[2026-06-24 04:20:59,954][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.058610741049051285, 'train/grad_norm_pre_clip': 21.940824508666992, 'timing/data': 0.0005723170470446348, 'timing/model': 1.1303011949639767, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 1.64}) +[2026-06-24 04:33:01,253][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4250 +[2026-06-24 04:33:01,253][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.054708048701286316, 'train/grad_norm_pre_clip': 20.65642738342285, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00036095594987273216, 'timing/model': 1.120808999054134, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 1.64}) +[2026-06-24 04:33:05,308][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.010551601648330688, 'train/grad_norm_pre_clip': 12.207573890686035, 'timing/data': 0.0005419100634753704, 'timing/model': 1.0860348319401965, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 1.65}) +[2026-06-24 04:33:09,306][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.03391188010573387, 'train/grad_norm_pre_clip': 21.869863510131836, 'timing/data': 0.0003580300835892558, 'timing/model': 1.1032351590692997, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 1.65}) +[2026-06-24 04:33:13,305][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.02848169021308422, 'train/grad_norm_pre_clip': 16.3701114654541, 'timing/data': 0.0004421790363267064, 'timing/model': 1.0907874290132895, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 1.65}) +[2026-06-24 04:33:17,299][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.10530048608779907, 'train/grad_norm_pre_clip': 16.971603393554688, 'timing/data': 0.00036383699625730515, 'timing/model': 1.097363890032284, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 1.65}) +[2026-06-24 04:33:21,486][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.062497131526470184, 'train/grad_norm_pre_clip': 15.782190322875977, 'timing/data': 0.00047393806744366884, 'timing/model': 1.1046120090177283, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 1.65}) +[2026-06-24 04:33:25,516][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.08937350660562515, 'train/grad_norm_pre_clip': 30.531328201293945, 'timing/data': 0.0003635319881141186, 'timing/model': 1.105421349988319, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 1.65}) +[2026-06-24 04:33:29,533][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.00022604476544074714, 'train/grad_norm_pre_clip': 15.998982429504395, 'timing/data': 0.0003822770668193698, 'timing/model': 1.1013522610301152, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 1.65}) +[2026-06-24 04:33:33,558][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 6.833311636000872e-05, 'train/grad_norm_pre_clip': 6.495619773864746, 'timing/data': 0.00035727303475141525, 'timing/model': 1.0994702649768442, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 1.65}) +[2026-06-24 04:33:37,679][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.0034290843177586794, 'train/grad_norm_pre_clip': 5.718573093414307, 'timing/data': 0.00039688905235379934, 'timing/model': 1.1014381360728294, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 1.65}) +[2026-06-24 04:33:41,722][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.0030455479864031076, 'train/grad_norm_pre_clip': 19.79194450378418, 'timing/data': 0.0003745469730347395, 'timing/model': 1.1089180980343372, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 1.65}) +[2026-06-24 04:33:45,770][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.023763064295053482, 'train/grad_norm_pre_clip': 24.028072357177734, 'timing/data': 0.0003771580522879958, 'timing/model': 1.109402138972655, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 1.65}) +[2026-06-24 04:33:49,809][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.02056807652115822, 'train/grad_norm_pre_clip': 17.504196166992188, 'timing/data': 0.00036899594124406576, 'timing/model': 1.1128671900369227, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 1.65}) +[2026-06-24 04:33:53,962][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.037791911512613297, 'train/grad_norm_pre_clip': 12.9659423828125, 'timing/data': 0.00036282697692513466, 'timing/model': 1.2156275060260668, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 1.65}) +[2026-06-24 04:33:58,024][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.12576837837696075, 'train/grad_norm_pre_clip': 56.774391174316406, 'timing/data': 0.0003582499921321869, 'timing/model': 1.1102788030402735, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 1.65}) +[2026-06-24 04:34:02,088][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.1389765739440918, 'train/grad_norm_pre_clip': 23.72599983215332, 'timing/data': 0.00037336593959480524, 'timing/model': 1.1173516240669414, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 1.65}) +[2026-06-24 04:34:06,137][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.004128546454012394, 'train/grad_norm_pre_clip': 15.733734130859375, 'timing/data': 0.00040604290552437305, 'timing/model': 1.1154550579376519, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 1.65}) +[2026-06-24 04:34:10,206][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.020431073382496834, 'train/grad_norm_pre_clip': 16.239809036254883, 'timing/data': 0.000396782997995615, 'timing/model': 1.1136762710520998, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 1.65}) +[2026-06-24 04:34:14,433][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.03211120143532753, 'train/grad_norm_pre_clip': 23.723388671875, 'timing/data': 0.0004384819185361266, 'timing/model': 1.113792320014909, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 1.65}) +[2026-06-24 04:34:18,487][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.0006391574861481786, 'train/grad_norm_pre_clip': 17.545568466186523, 'timing/data': 0.00039827602449804544, 'timing/model': 1.1131632400210947, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 1.65}) +[2026-06-24 04:34:22,552][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.005255966447293758, 'train/grad_norm_pre_clip': 21.053197860717773, 'timing/data': 0.0005223649786785245, 'timing/model': 1.1119545300025493, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 1.65}) +[2026-06-24 04:34:26,622][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.03405890241265297, 'train/grad_norm_pre_clip': 16.377140045166016, 'timing/data': 0.0005410459125414491, 'timing/model': 1.120100473985076, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 1.65}) +[2026-06-24 04:34:30,763][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.003584738587960601, 'train/grad_norm_pre_clip': 19.539709091186523, 'timing/data': 0.0004670859780162573, 'timing/model': 1.1154963949229568, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 1.65}) +[2026-06-24 04:34:34,819][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.03038409724831581, 'train/grad_norm_pre_clip': 27.070283889770508, 'timing/data': 0.0003662100061774254, 'timing/model': 1.1130703559610993, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 1.65}) +[2026-06-24 04:34:38,875][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.0030451861675828695, 'train/grad_norm_pre_clip': 49.283973693847656, 'timing/data': 0.0003506309585645795, 'timing/model': 1.1138304059859365, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 1.65}) +[2026-06-24 04:34:42,941][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.1870056539773941, 'train/grad_norm_pre_clip': 32.07585906982422, 'timing/data': 0.0004115599440410733, 'timing/model': 1.1206814460456371, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 1.65}) +[2026-06-24 04:34:47,083][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.04600600525736809, 'train/grad_norm_pre_clip': 21.760181427001953, 'timing/data': 0.00040774105582386255, 'timing/model': 1.119619901990518, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 1.65}) +[2026-06-24 04:34:51,149][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.14401111006736755, 'train/grad_norm_pre_clip': 25.793872833251953, 'timing/data': 0.0003544510109350085, 'timing/model': 1.117507519084029, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 1.66}) +[2026-06-24 04:34:55,225][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.018046751618385315, 'train/grad_norm_pre_clip': 10.097522735595703, 'timing/data': 0.00036176794674247503, 'timing/model': 1.1173868180485442, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 1.66}) +[2026-06-24 04:34:59,300][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.09164103120565414, 'train/grad_norm_pre_clip': 17.648887634277344, 'timing/data': 0.0004119820659980178, 'timing/model': 1.1221642170567065, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 1.66}) +[2026-06-24 04:35:03,444][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.025057855993509293, 'train/grad_norm_pre_clip': 11.957317352294922, 'timing/data': 0.0005489250179380178, 'timing/model': 1.1118556989822537, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 1.66}) +[2026-06-24 04:35:07,515][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.012859820388257504, 'train/grad_norm_pre_clip': 13.75633716583252, 'timing/data': 0.0003741900436580181, 'timing/model': 1.1106271130265668, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 1.66}) +[2026-06-24 04:35:11,583][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.06269218772649765, 'train/grad_norm_pre_clip': 24.739990234375, 'timing/data': 0.00036388495936989784, 'timing/model': 1.1107425920199603, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 1.66}) +[2026-06-24 04:35:15,650][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.0014150220667943358, 'train/grad_norm_pre_clip': 7.510869026184082, 'timing/data': 0.00035649898927658796, 'timing/model': 1.1170913309324533, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 1.66}) +[2026-06-24 04:35:19,876][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.059457022696733475, 'train/grad_norm_pre_clip': 15.809104919433594, 'timing/data': 0.00035965105053037405, 'timing/model': 1.1129748478997499, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 1.66}) +[2026-06-24 04:35:23,952][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.017545269802212715, 'train/grad_norm_pre_clip': 6.9980387687683105, 'timing/data': 0.0003649359568953514, 'timing/model': 1.1174358240095899, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 1.66}) +[2026-06-24 04:35:28,028][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.27325576543807983, 'train/grad_norm_pre_clip': 16.171016693115234, 'timing/data': 0.0004513090243563056, 'timing/model': 1.1172291850671172, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 1.66}) +[2026-06-24 04:35:32,192][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.06385789066553116, 'train/grad_norm_pre_clip': 14.215970993041992, 'timing/data': 0.00038868794217705727, 'timing/model': 1.2050370540237054, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 1.66}) +[2026-06-24 04:35:36,272][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.011382176540791988, 'train/grad_norm_pre_clip': 23.13266944885254, 'timing/data': 0.0004179709358140826, 'timing/model': 1.1176048499764875, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 1.66}) +[2026-06-24 04:35:40,349][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.011752775870263577, 'train/grad_norm_pre_clip': 21.17224884033203, 'timing/data': 0.0003609810955822468, 'timing/model': 1.1142174609703943, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 1.66}) +[2026-06-24 04:35:44,430][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.09020838141441345, 'train/grad_norm_pre_clip': 19.82316017150879, 'timing/data': 0.0003657368943095207, 'timing/model': 1.1178533480269834, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 1.66}) +[2026-06-24 04:35:48,685][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 5.1251026889076456e-05, 'train/grad_norm_pre_clip': 12.522286415100098, 'timing/data': 0.00041359802708029747, 'timing/model': 1.2849228790728375, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 1.66}) +[2026-06-24 04:35:52,758][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.02309752069413662, 'train/grad_norm_pre_clip': 17.27446174621582, 'timing/data': 0.0004167769802734256, 'timing/model': 1.1124987479997799, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 1.66}) +[2026-06-24 04:35:56,849][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.06799224764108658, 'train/grad_norm_pre_clip': 20.28803253173828, 'timing/data': 0.00036109203938394785, 'timing/model': 1.121168349054642, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 1.66}) +[2026-06-24 04:36:00,938][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.00195487798191607, 'train/grad_norm_pre_clip': 10.640268325805664, 'timing/data': 0.00035762996412813663, 'timing/model': 1.1212457088986412, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 1.66}) +[2026-06-24 04:36:05,102][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.00021472203661687672, 'train/grad_norm_pre_clip': 10.062705039978027, 'timing/data': 0.0003735750215128064, 'timing/model': 1.113215633900836, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 1.66}) +[2026-06-24 04:36:09,179][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.03045198880136013, 'train/grad_norm_pre_clip': 15.77969741821289, 'timing/data': 0.0005927779711782932, 'timing/model': 1.1153437111061066, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 1.66}) +[2026-06-24 04:36:13,295][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.022288018837571144, 'train/grad_norm_pre_clip': 6.914859294891357, 'timing/data': 0.0003561399644240737, 'timing/model': 1.1304013259941712, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 1.66}) +[2026-06-24 04:36:17,405][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.08549866080284119, 'train/grad_norm_pre_clip': 20.154027938842773, 'timing/data': 0.00036926299799233675, 'timing/model': 1.1213877049740404, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 1.66}) +[2026-06-24 04:36:21,564][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.020424919202923775, 'train/grad_norm_pre_clip': 12.986699104309082, 'timing/data': 0.0006037410348653793, 'timing/model': 1.117890355992131, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 1.66}) +[2026-06-24 04:36:25,648][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4300 +[2026-06-24 04:37:30,360][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4300 +[2026-06-24 04:37:30,362][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4300 +[2026-06-24 04:38:08,979][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4300 +[2026-06-24 04:38:08,980][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.022102709859609604, 'train/grad_norm_pre_clip': 27.150606155395508, 'eval/action_loss/seconds': 64.70970716199372, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.028781823813915253, 'eval/latency_2/loss': 0.028781823813915253, 'eval/flappy/loss': 0.028781823813915253, 'eval/flappy/latency_2/loss': 0.028781823813915253, 'eval/flap_precision': 0.9807956104252401, 'eval/flap_recall': 0.9701492537313433, 'eval/flap_f1': 0.975443383356071, 'eval/flap_rate_model': 0.2025, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.975443383356071, 'eval/latency_2/flap_precision': 0.9807956104252401, 'eval/latency_2/flap_recall': 0.9701492537313433, 'eval/latency_2/flap_f1': 0.975443383356071, 'eval/latency_2/flap_rate_model': 0.2025, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.975443383356071, 'eval/action_classification/data_wait_seconds': 0.6275601983070374, 'eval/action_classification/predict_seconds': 37.81379699707031, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0005205769557505846, 'timing/model': 1.1146927190711722, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 1.66}) +[2026-06-24 04:38:13,030][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.01017700508236885, 'train/grad_norm_pre_clip': 16.708375930786133, 'timing/data': 0.00036996102426201105, 'timing/model': 1.1132349640829489, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 1.66}) +[2026-06-24 04:38:17,268][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.017581136897206306, 'train/grad_norm_pre_clip': 10.540599822998047, 'timing/data': 0.00043171399738639593, 'timing/model': 1.29025270906277, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 1.66}) +[2026-06-24 04:38:21,325][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.019606096670031548, 'train/grad_norm_pre_clip': 13.610655784606934, 'timing/data': 0.0003617439651861787, 'timing/model': 1.1143411389784887, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 1.67}) +[2026-06-24 04:38:25,385][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.1329842358827591, 'train/grad_norm_pre_clip': 27.706104278564453, 'timing/data': 0.00036703108344227076, 'timing/model': 1.1114006099523976, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 1.67}) +[2026-06-24 04:38:29,451][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.000361641839845106, 'train/grad_norm_pre_clip': 5.901327133178711, 'timing/data': 0.0004263310693204403, 'timing/model': 1.1208707810146734, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 1.67}) +[2026-06-24 04:38:33,573][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.027977336198091507, 'train/grad_norm_pre_clip': 10.625442504882812, 'timing/data': 0.00036610395181924105, 'timing/model': 1.1141141980187967, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 1.67}) +[2026-06-24 04:38:37,622][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.06346146762371063, 'train/grad_norm_pre_clip': 15.430376052856445, 'timing/data': 0.00037506897933781147, 'timing/model': 1.1093631139956415, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 1.67}) +[2026-06-24 04:38:41,694][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.000925472762901336, 'train/grad_norm_pre_clip': 10.634713172912598, 'timing/data': 0.0007042560027912259, 'timing/model': 1.115148322074674, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 1.67}) +[2026-06-24 04:38:45,760][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.010696075856685638, 'train/grad_norm_pre_clip': 22.39896011352539, 'timing/data': 0.0004986749263480306, 'timing/model': 1.1142206190852448, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 1.67}) +[2026-06-24 04:38:49,972][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.213201105594635, 'train/grad_norm_pre_clip': 29.028091430664062, 'timing/data': 0.00036162498872727156, 'timing/model': 1.107434876030311, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 1.67}) +[2026-06-24 04:38:54,037][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.01620776765048504, 'train/grad_norm_pre_clip': 13.772359848022461, 'timing/data': 0.00036629103124141693, 'timing/model': 1.1111970619531348, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 1.67}) +[2026-06-24 04:38:58,109][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.022678904235363007, 'train/grad_norm_pre_clip': 23.058902740478516, 'timing/data': 0.0003680590307340026, 'timing/model': 1.1168701719725505, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 1.67}) +[2026-06-24 04:39:02,198][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.06256004422903061, 'train/grad_norm_pre_clip': 10.648210525512695, 'timing/data': 0.0005974540254101157, 'timing/model': 1.1143461099127308, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 1.67}) +[2026-06-24 04:39:06,367][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.00864157173782587, 'train/grad_norm_pre_clip': 9.35807991027832, 'timing/data': 0.0003624520031735301, 'timing/model': 1.121985181933269, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 1.67}) +[2026-06-24 04:39:10,462][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.032288577407598495, 'train/grad_norm_pre_clip': 12.684901237487793, 'timing/data': 0.0004148579901084304, 'timing/model': 1.1152305119903758, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 1.67}) +[2026-06-24 04:39:14,561][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.13590776920318604, 'train/grad_norm_pre_clip': 24.432353973388672, 'timing/data': 0.00036103406455367804, 'timing/model': 1.1172221819870174, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 1.67}) +[2026-06-24 04:39:18,650][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.11734321713447571, 'train/grad_norm_pre_clip': 31.433732986450195, 'timing/data': 0.00036392093170434237, 'timing/model': 1.12205405801069, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 1.67}) +[2026-06-24 04:39:22,840][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.0027691200375556946, 'train/grad_norm_pre_clip': 4.8309760093688965, 'timing/data': 0.0003660980146378279, 'timing/model': 1.2370832180604339, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 1.67}) +[2026-06-24 04:39:26,921][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.0037709029857069254, 'train/grad_norm_pre_clip': 39.07835006713867, 'timing/data': 0.0005049799801781774, 'timing/model': 1.1233012500451878, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 1.67}) +[2026-06-24 04:39:31,012][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.04236742854118347, 'train/grad_norm_pre_clip': 42.78434753417969, 'timing/data': 0.00035661295987665653, 'timing/model': 1.1184865609975532, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 1.67}) +[2026-06-24 04:39:35,086][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.023360934108495712, 'train/grad_norm_pre_clip': 15.024096488952637, 'timing/data': 0.00037115602754056454, 'timing/model': 1.1210435959510505, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 1.67}) +[2026-06-24 04:39:39,159][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.0034685349091887474, 'train/grad_norm_pre_clip': 14.264897346496582, 'timing/data': 0.0006695499178022146, 'timing/model': 1.1147169839823619, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 1.67}) +[2026-06-24 04:39:43,402][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.02668393775820732, 'train/grad_norm_pre_clip': 12.676078796386719, 'timing/data': 0.00036677997559309006, 'timing/model': 1.120601403992623, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 1.67}) +[2026-06-24 04:39:47,478][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.03291911259293556, 'train/grad_norm_pre_clip': 33.07832336425781, 'timing/data': 0.00035002699587494135, 'timing/model': 1.1198075980646536, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 1.67}) +[2026-06-24 04:39:51,554][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.07148131728172302, 'train/grad_norm_pre_clip': 26.539806365966797, 'timing/data': 0.0003668570425361395, 'timing/model': 1.115967761958018, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 1.67}) +[2026-06-24 04:39:55,647][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.012166806496679783, 'train/grad_norm_pre_clip': 9.390120506286621, 'timing/data': 0.0003762840060517192, 'timing/model': 1.1216208830010146, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 1.67}) +[2026-06-24 04:39:59,806][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.017473604530096054, 'train/grad_norm_pre_clip': 13.417681694030762, 'timing/data': 0.0003677440108731389, 'timing/model': 1.1194683249341324, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 1.67}) +[2026-06-24 04:40:03,894][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.0056240130215883255, 'train/grad_norm_pre_clip': 20.213878631591797, 'timing/data': 0.00041715498082339764, 'timing/model': 1.118113094009459, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 1.68}) +[2026-06-24 04:40:07,982][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.09222687035799026, 'train/grad_norm_pre_clip': 39.17188262939453, 'timing/data': 0.0003975949948653579, 'timing/model': 1.1190232411026955, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 1.68}) +[2026-06-24 04:40:12,067][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.0035479802172631025, 'train/grad_norm_pre_clip': 7.564785480499268, 'timing/data': 0.00043225602712482214, 'timing/model': 1.1150915959151462, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 1.68}) +[2026-06-24 04:40:16,216][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.1903344988822937, 'train/grad_norm_pre_clip': 34.35285568237305, 'timing/data': 0.00040626502595841885, 'timing/model': 1.1161711079766974, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 1.68}) +[2026-06-24 04:40:20,300][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.0013968986459076405, 'train/grad_norm_pre_clip': 11.56815242767334, 'timing/data': 0.00035742903128266335, 'timing/model': 1.118916342034936, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 1.68}) +[2026-06-24 04:40:24,392][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.0014279019087553024, 'train/grad_norm_pre_clip': 14.38928508758545, 'timing/data': 0.00035701494198292494, 'timing/model': 1.125381472054869, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 1.68}) +[2026-06-24 04:40:28,488][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.13696718215942383, 'train/grad_norm_pre_clip': 32.4339599609375, 'timing/data': 0.0003556649899110198, 'timing/model': 1.1212980200070888, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 1.68}) +[2026-06-24 04:40:32,669][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.03865911811590195, 'train/grad_norm_pre_clip': 19.05767059326172, 'timing/data': 0.00039940199349075556, 'timing/model': 1.1177054430590943, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 1.68}) +[2026-06-24 04:40:36,755][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.0021417124662548304, 'train/grad_norm_pre_clip': 20.302021026611328, 'timing/data': 0.00035678502172231674, 'timing/model': 1.1224785329541191, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 1.68}) +[2026-06-24 04:40:40,848][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.10364123433828354, 'train/grad_norm_pre_clip': 36.14192581176758, 'timing/data': 0.0006485549965873361, 'timing/model': 1.1207025739131495, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 1.68}) +[2026-06-24 04:40:44,926][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.15173251926898956, 'train/grad_norm_pre_clip': 61.819915771484375, 'timing/data': 0.0004112300230190158, 'timing/model': 1.1151810640003532, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 1.68}) +[2026-06-24 04:40:49,151][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.050475217401981354, 'train/grad_norm_pre_clip': 18.43449592590332, 'timing/data': 0.0005614219699054956, 'timing/model': 1.1207479640142992, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 1.68}) +[2026-06-24 04:40:53,246][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.005692051257938147, 'train/grad_norm_pre_clip': 15.637374877929688, 'timing/data': 0.0003695678897202015, 'timing/model': 1.1285513489274308, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 1.68}) +[2026-06-24 04:40:57,339][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.0563013032078743, 'train/grad_norm_pre_clip': 20.857255935668945, 'timing/data': 0.0005379890790209174, 'timing/model': 1.117090504965745, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 1.68}) +[2026-06-24 04:41:01,525][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.009876315481960773, 'train/grad_norm_pre_clip': 36.113807678222656, 'timing/data': 0.0004195920191705227, 'timing/model': 1.2204276890261099, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 1.68}) +[2026-06-24 04:41:05,608][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.003674409817904234, 'train/grad_norm_pre_clip': 23.727279663085938, 'timing/data': 0.0004519399954006076, 'timing/model': 1.1201974159339443, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 1.68}) +[2026-06-24 04:41:09,695][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.016109252348542213, 'train/grad_norm_pre_clip': 33.143882751464844, 'timing/data': 0.00042538694106042385, 'timing/model': 1.117935837013647, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 1.68}) +[2026-06-24 04:41:13,784][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.04321467876434326, 'train/grad_norm_pre_clip': 21.595359802246094, 'timing/data': 0.0003564979415386915, 'timing/model': 1.1238005900522694, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 1.68}) +[2026-06-24 04:41:18,034][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.07637467980384827, 'train/grad_norm_pre_clip': 37.719966888427734, 'timing/data': 0.0004084069514647126, 'timing/model': 1.2865856540156528, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 1.68}) +[2026-06-24 04:41:22,122][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.039451491087675095, 'train/grad_norm_pre_clip': 24.84610366821289, 'timing/data': 0.0003653489984571934, 'timing/model': 1.1227679629810154, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 1.68}) +[2026-06-24 04:41:26,201][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.1888887733221054, 'train/grad_norm_pre_clip': 17.089399337768555, 'timing/data': 0.00040417397394776344, 'timing/model': 1.1174738339614123, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 1.68}) +[2026-06-24 04:41:30,288][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.07503633946180344, 'train/grad_norm_pre_clip': 43.60108184814453, 'timing/data': 0.00036770408041775227, 'timing/model': 1.1210844269953668, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 1.68}) +[2026-06-24 04:41:34,448][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.0010865043150261045, 'train/grad_norm_pre_clip': 11.2208251953125, 'timing/data': 0.0004399099852889776, 'timing/model': 1.1178979409160092, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 1.68}) +[2026-06-24 04:41:38,537][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.018178272992372513, 'train/grad_norm_pre_clip': 10.25041675567627, 'timing/data': 0.00041126098949462175, 'timing/model': 1.1183616829803213, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 1.68}) +[2026-06-24 04:41:42,610][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.012733062729239464, 'train/grad_norm_pre_clip': 14.716503143310547, 'timing/data': 0.0003579159965738654, 'timing/model': 1.115801225998439, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 1.68}) +[2026-06-24 04:41:46,705][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.049782801419496536, 'train/grad_norm_pre_clip': 42.79799270629883, 'timing/data': 0.000416580936871469, 'timing/model': 1.1235926590161398, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 1.68}) +[2026-06-24 04:41:50,855][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.011477493681013584, 'train/grad_norm_pre_clip': 21.611841201782227, 'timing/data': 0.0003647910198196769, 'timing/model': 1.1188930160133168, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 1.69}) +[2026-06-24 04:41:54,946][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.10980816930532455, 'train/grad_norm_pre_clip': 14.567922592163086, 'timing/data': 0.0004219099646434188, 'timing/model': 1.1221802911022678, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 1.69}) +[2026-06-24 04:41:59,033][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.02197592705488205, 'train/grad_norm_pre_clip': 20.318178176879883, 'timing/data': 0.00035895395558327436, 'timing/model': 1.121392852975987, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 1.69}) +[2026-06-24 04:42:03,120][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.10739874094724655, 'train/grad_norm_pre_clip': 29.716028213500977, 'timing/data': 0.0003616110188886523, 'timing/model': 1.1188149719964713, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 1.69}) +[2026-06-24 04:42:07,354][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.0002662047336343676, 'train/grad_norm_pre_clip': 3.426103353500366, 'timing/data': 0.0003703129477798939, 'timing/model': 1.1183252059854567, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 1.69}) +[2026-06-24 04:42:11,445][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.021480131894350052, 'train/grad_norm_pre_clip': 39.61098098754883, 'timing/data': 0.0004002740606665611, 'timing/model': 1.1135923030087724, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 1.69}) +[2026-06-24 04:42:15,532][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.012888823635876179, 'train/grad_norm_pre_clip': 11.901617050170898, 'timing/data': 0.0003478709841147065, 'timing/model': 1.124388389987871, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 1.69}) +[2026-06-24 04:42:19,622][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.0308409221470356, 'train/grad_norm_pre_clip': 25.885114669799805, 'timing/data': 0.0003635711036622524, 'timing/model': 1.119621476973407, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 1.69}) +[2026-06-24 04:42:23,808][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.036412328481674194, 'train/grad_norm_pre_clip': 13.536550521850586, 'timing/data': 0.0003681778907775879, 'timing/model': 1.1266549490392208, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 1.69}) +[2026-06-24 04:42:27,900][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.059813667088747025, 'train/grad_norm_pre_clip': 17.550235748291016, 'timing/data': 0.00039844599086791277, 'timing/model': 1.1197924979496747, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 1.69}) +[2026-06-24 04:42:31,992][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.07995275408029556, 'train/grad_norm_pre_clip': 21.445720672607422, 'timing/data': 0.0003576379967853427, 'timing/model': 1.119656216003932, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 1.69}) +[2026-06-24 04:42:36,091][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.0016406099312007427, 'train/grad_norm_pre_clip': 12.462158203125, 'timing/data': 0.0003644060343503952, 'timing/model': 1.1227163919247687, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 1.69}) +[2026-06-24 04:42:40,259][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.002432290930300951, 'train/grad_norm_pre_clip': 31.23702621459961, 'timing/data': 0.00040693406481295824, 'timing/model': 1.1199072540039197, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 1.69}) +[2026-06-24 04:42:44,354][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.04989870637655258, 'train/grad_norm_pre_clip': 24.04676628112793, 'timing/data': 0.0003870909567922354, 'timing/model': 1.1220100210048258, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 1.69}) +[2026-06-24 04:42:48,446][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.01210132334381342, 'train/grad_norm_pre_clip': 10.049027442932129, 'timing/data': 0.0005034060450270772, 'timing/model': 1.1205838801106438, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 1.69}) +[2026-06-24 04:42:52,532][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.06915147602558136, 'train/grad_norm_pre_clip': 46.44155502319336, 'timing/data': 0.0003619040362536907, 'timing/model': 1.1222056259866804, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 1.69}) +[2026-06-24 04:42:56,785][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.005903737153857946, 'train/grad_norm_pre_clip': 15.53039836883545, 'timing/data': 0.000659208046272397, 'timing/model': 1.1214510150020942, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 1.69}) +[2026-06-24 04:43:00,880][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.003674848470836878, 'train/grad_norm_pre_clip': 20.46495819091797, 'timing/data': 0.00036664202343672514, 'timing/model': 1.1230194790987298, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 1.69}) +[2026-06-24 04:43:04,971][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.04457651823759079, 'train/grad_norm_pre_clip': 20.730934143066406, 'timing/data': 0.0007305499166250229, 'timing/model': 1.1199001419590786, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 1.69}) +[2026-06-24 04:43:09,075][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.019069384783506393, 'train/grad_norm_pre_clip': 13.744559288024902, 'timing/data': 0.00035200000274926424, 'timing/model': 1.1243379270890728, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 1.69}) +[2026-06-24 04:43:13,265][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.0009319747914560139, 'train/grad_norm_pre_clip': 12.5660400390625, 'timing/data': 0.00035562890116125345, 'timing/model': 1.1186859490117058, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 1.69}) +[2026-06-24 04:43:17,353][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.09018898010253906, 'train/grad_norm_pre_clip': 13.485125541687012, 'timing/data': 0.0003609659615904093, 'timing/model': 1.1146429229993373, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 1.69}) +[2026-06-24 04:43:21,450][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.015802592039108276, 'train/grad_norm_pre_clip': 11.671732902526855, 'timing/data': 0.0005644209450110793, 'timing/model': 1.124082454945892, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 1.69}) +[2026-06-24 04:43:25,545][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 7.230374467326328e-05, 'train/grad_norm_pre_clip': 19.43157196044922, 'timing/data': 0.0005600280128419399, 'timing/model': 1.1175759170437232, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 1.69}) +[2026-06-24 04:43:29,803][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.055413782596588135, 'train/grad_norm_pre_clip': 16.648345947265625, 'timing/data': 0.0005310679553076625, 'timing/model': 1.1215069289319217, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 1.69}) +[2026-06-24 04:43:33,893][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.023491309955716133, 'train/grad_norm_pre_clip': 27.314884185791016, 'timing/data': 0.0005419100634753704, 'timing/model': 1.1187589560868219, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 1.69}) +[2026-06-24 04:43:37,978][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.1629764288663864, 'train/grad_norm_pre_clip': 35.93833923339844, 'timing/data': 0.00037764094304293394, 'timing/model': 1.1141980270622298, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 1.7}) +[2026-06-24 04:43:42,074][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.09326222538948059, 'train/grad_norm_pre_clip': 29.459707260131836, 'timing/data': 0.0004108899738639593, 'timing/model': 1.1190569309983402, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 1.7}) +[2026-06-24 04:43:46,233][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.06667376309633255, 'train/grad_norm_pre_clip': 14.691975593566895, 'timing/data': 0.00037042598705738783, 'timing/model': 1.123409411055036, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 1.7}) +[2026-06-24 04:43:50,321][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.05527583882212639, 'train/grad_norm_pre_clip': 12.838075637817383, 'timing/data': 0.00042173697147518396, 'timing/model': 1.1158245529513806, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 1.7}) +[2026-06-24 04:43:54,413][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.002793187042698264, 'train/grad_norm_pre_clip': 18.46718406677246, 'timing/data': 0.0003932249965146184, 'timing/model': 1.1222019699634984, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 1.7}) +[2026-06-24 04:43:58,501][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.06260242313146591, 'train/grad_norm_pre_clip': 16.105804443359375, 'timing/data': 0.00036248296964913607, 'timing/model': 1.120988684007898, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 1.7}) +[2026-06-24 04:44:02,669][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.027612030506134033, 'train/grad_norm_pre_clip': 12.928868293762207, 'timing/data': 0.0004185170400887728, 'timing/model': 1.1143369590863585, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 1.7}) +[2026-06-24 04:44:06,750][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.03465067222714424, 'train/grad_norm_pre_clip': 19.931015014648438, 'timing/data': 0.0003711000317707658, 'timing/model': 1.1188331139273942, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 1.7}) +[2026-06-24 04:44:10,848][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.12448132783174515, 'train/grad_norm_pre_clip': 24.798402786254883, 'timing/data': 0.00036334001924842596, 'timing/model': 1.1206451489124447, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 1.7}) +[2026-06-24 04:44:14,934][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.046337734907865524, 'train/grad_norm_pre_clip': 25.85491180419922, 'timing/data': 0.0005291800480335951, 'timing/model': 1.1181839379714802, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 1.7}) +[2026-06-24 04:44:19,168][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.00936264917254448, 'train/grad_norm_pre_clip': 27.93299102783203, 'timing/data': 0.000721889897249639, 'timing/model': 1.1225249689305201, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 1.7}) +[2026-06-24 04:44:23,263][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.14842656254768372, 'train/grad_norm_pre_clip': 12.773381233215332, 'timing/data': 0.0005952260689809918, 'timing/model': 1.1219643860822544, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 1.7}) +[2026-06-24 04:44:27,361][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.014198952354490757, 'train/grad_norm_pre_clip': 11.975628852844238, 'timing/data': 0.0003673670580610633, 'timing/model': 1.124721825006418, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 1.7}) +[2026-06-24 04:44:31,563][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.050884101539850235, 'train/grad_norm_pre_clip': 43.271034240722656, 'timing/data': 0.0004199399845674634, 'timing/model': 1.2246851150412112, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 1.7}) +[2026-06-24 04:44:35,659][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.000325382367009297, 'train/grad_norm_pre_clip': 21.308143615722656, 'timing/data': 0.00040079106111079454, 'timing/model': 1.1209591639926657, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 1.7}) +[2026-06-24 04:44:39,756][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.01282939687371254, 'train/grad_norm_pre_clip': 25.904773712158203, 'timing/data': 0.00035984907299280167, 'timing/model': 1.1199923149542883, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 1.7}) +[2026-06-24 04:44:43,859][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.006414329633116722, 'train/grad_norm_pre_clip': 12.28621768951416, 'timing/data': 0.00036703108344227076, 'timing/model': 1.120282090967521, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 1.7}) +[2026-06-24 04:44:48,121][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.08629035949707031, 'train/grad_norm_pre_clip': 34.1129150390625, 'timing/data': 0.0007114739855751395, 'timing/model': 1.2929731119656935, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 1.7}) +[2026-06-24 04:44:52,223][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.016025574877858162, 'train/grad_norm_pre_clip': 16.77480125427246, 'timing/data': 0.0004087249981239438, 'timing/model': 1.12491982197389, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 1.7}) +[2026-06-24 04:44:56,336][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.02683035284280777, 'train/grad_norm_pre_clip': 17.483640670776367, 'timing/data': 0.00035973801277577877, 'timing/model': 1.125846398063004, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 1.7}) +[2026-06-24 04:45:00,426][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4400 +[2026-06-24 04:46:05,159][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4400 +[2026-06-24 04:46:05,161][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4400 +[2026-06-24 04:46:44,062][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4400 +[2026-06-24 04:46:44,063][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.018754146993160248, 'train/grad_norm_pre_clip': 16.48545265197754, 'eval/action_loss/seconds': 64.73009885696229, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.02813352271914482, 'eval/latency_2/loss': 0.02813352271914482, 'eval/flappy/loss': 0.02813352271914482, 'eval/flappy/latency_2/loss': 0.02813352271914482, 'eval/flap_precision': 0.9834710743801653, 'eval/flap_recall': 0.9687924016282226, 'eval/flap_f1': 0.9760765550239235, 'eval/flap_rate_model': 0.20166666666666666, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9760765550239235, 'eval/latency_2/flap_precision': 0.9834710743801653, 'eval/latency_2/flap_recall': 0.9687924016282226, 'eval/latency_2/flap_f1': 0.9760765550239235, 'eval/latency_2/flap_rate_model': 0.20166666666666666, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9760765550239235, 'eval/action_classification/data_wait_seconds': 0.5749540328979492, 'eval/action_classification/predict_seconds': 38.1900520324707, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.000364903942681849, 'timing/model': 1.1200890050968155, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 1.7}) +[2026-06-24 04:46:48,117][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.0009027146734297276, 'train/grad_norm_pre_clip': 30.899620056152344, 'timing/data': 0.0008264710195362568, 'timing/model': 1.110467174090445, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 1.7}) +[2026-06-24 04:46:52,167][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.06984462589025497, 'train/grad_norm_pre_clip': 33.338478088378906, 'timing/data': 0.00036776706110686064, 'timing/model': 1.1122145910048857, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 1.7}) +[2026-06-24 04:46:56,222][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.02618670091032982, 'train/grad_norm_pre_clip': 23.064029693603516, 'timing/data': 0.0003594239242374897, 'timing/model': 1.1093959159916267, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 1.7}) +[2026-06-24 04:47:00,423][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.12599384784698486, 'train/grad_norm_pre_clip': 29.457435607910156, 'timing/data': 0.00036384700797498226, 'timing/model': 1.253290815046057, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 1.7}) +[2026-06-24 04:47:04,472][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.025539876893162727, 'train/grad_norm_pre_clip': 16.111371994018555, 'timing/data': 0.000372758018784225, 'timing/model': 1.1115312459878623, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 1.7}) +[2026-06-24 04:47:08,526][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.0007907651597633958, 'train/grad_norm_pre_clip': 13.611371994018555, 'timing/data': 0.0003640870563685894, 'timing/model': 1.1088716989615932, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 1.71}) +[2026-06-24 04:47:12,596][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.015424779616296291, 'train/grad_norm_pre_clip': 27.729150772094727, 'timing/data': 0.0003599299816414714, 'timing/model': 1.1129306650254875, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 1.71}) +[2026-06-24 04:47:16,785][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.026536166667938232, 'train/grad_norm_pre_clip': 26.528072357177734, 'timing/data': 0.000506141921505332, 'timing/model': 1.118995816912502, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 1.71}) +[2026-06-24 04:47:20,850][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.06809671968221664, 'train/grad_norm_pre_clip': 46.595458984375, 'timing/data': 0.00036321894731372595, 'timing/model': 1.1110024749068543, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 1.71}) +[2026-06-24 04:47:24,936][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.002400748897343874, 'train/grad_norm_pre_clip': 6.016417503356934, 'timing/data': 0.00035656499676406384, 'timing/model': 1.1166013800539076, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 1.71}) +[2026-06-24 04:47:29,004][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.0020037044305354357, 'train/grad_norm_pre_clip': 21.896955490112305, 'timing/data': 0.0003599849296733737, 'timing/model': 1.1151208990486339, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 1.71}) +[2026-06-24 04:47:33,260][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.008866325952112675, 'train/grad_norm_pre_clip': 9.737470626831055, 'timing/data': 0.00036485400050878525, 'timing/model': 1.1180997610790655, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 1.71}) +[2026-06-24 04:47:37,333][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.025655368342995644, 'train/grad_norm_pre_clip': 29.810596466064453, 'timing/data': 0.0003664020914584398, 'timing/model': 1.1160054940264672, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 1.71}) +[2026-06-24 04:47:41,406][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.008418533951044083, 'train/grad_norm_pre_clip': 19.5924072265625, 'timing/data': 0.0005509390030056238, 'timing/model': 1.1187751760007814, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 1.71}) +[2026-06-24 04:47:45,494][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.002937784418463707, 'train/grad_norm_pre_clip': 24.76791763305664, 'timing/data': 0.0003794759977608919, 'timing/model': 1.1215163870947435, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 1.71}) +[2026-06-24 04:47:49,739][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.010491983965039253, 'train/grad_norm_pre_clip': 3.0511059761047363, 'timing/data': 0.0005640980089083314, 'timing/model': 1.1206513380166143, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 1.71}) +[2026-06-24 04:47:53,798][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.11195902526378632, 'train/grad_norm_pre_clip': 18.016029357910156, 'timing/data': 0.0003590470878407359, 'timing/model': 1.1138163590803742, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 1.71}) +[2026-06-24 04:47:57,868][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.04344607889652252, 'train/grad_norm_pre_clip': 11.068765640258789, 'timing/data': 0.0005625349003821611, 'timing/model': 1.1156729210633785, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 1.71}) +[2026-06-24 04:48:01,943][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.10075793415307999, 'train/grad_norm_pre_clip': 21.456605911254883, 'timing/data': 0.00042017397936433554, 'timing/model': 1.1178523580310866, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 1.71}) +[2026-06-24 04:48:06,080][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.011074775829911232, 'train/grad_norm_pre_clip': 15.742679595947266, 'timing/data': 0.0003938280278816819, 'timing/model': 1.1116699100239202, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 1.71}) +[2026-06-24 04:48:10,159][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.029049856588244438, 'train/grad_norm_pre_clip': 26.39999771118164, 'timing/data': 0.0005575140239670873, 'timing/model': 1.1154705529334024, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 1.71}) +[2026-06-24 04:48:14,257][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.00983453169465065, 'train/grad_norm_pre_clip': 33.249176025390625, 'timing/data': 0.000360458972863853, 'timing/model': 1.1248500320361927, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 1.71}) +[2026-06-24 04:48:18,348][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.04980522766709328, 'train/grad_norm_pre_clip': 19.657114028930664, 'timing/data': 0.0005377430934458971, 'timing/model': 1.1199251789366826, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 1.71}) +[2026-06-24 04:48:22,512][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.005807444918900728, 'train/grad_norm_pre_clip': 6.686994552612305, 'timing/data': 0.00035874301102012396, 'timing/model': 1.1133808799786493, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 1.71}) +[2026-06-24 04:48:26,606][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.008876532316207886, 'train/grad_norm_pre_clip': 10.66755199432373, 'timing/data': 0.00050027493853122, 'timing/model': 1.1189039889723063, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 1.71}) +[2026-06-24 04:48:30,690][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.21951784193515778, 'train/grad_norm_pre_clip': 20.028398513793945, 'timing/data': 0.0003624339587986469, 'timing/model': 1.116227968945168, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 1.71}) +[2026-06-24 04:48:34,781][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.014966989867389202, 'train/grad_norm_pre_clip': 6.69243049621582, 'timing/data': 0.0003551190020516515, 'timing/model': 1.1194656329462305, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 1.71}) +[2026-06-24 04:48:38,994][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.05900520458817482, 'train/grad_norm_pre_clip': 23.917999267578125, 'timing/data': 0.00035401294007897377, 'timing/model': 1.1219471469521523, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 1.71}) +[2026-06-24 04:48:43,078][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.0286655742675066, 'train/grad_norm_pre_clip': 13.050115585327148, 'timing/data': 0.0005001510726287961, 'timing/model': 1.112958164070733, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 1.71}) +[2026-06-24 04:48:47,149][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.045819755643606186, 'train/grad_norm_pre_clip': 35.64932632446289, 'timing/data': 0.00043853698298335075, 'timing/model': 1.1136922700097784, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 1.71}) +[2026-06-24 04:48:51,306][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.0009736232459545135, 'train/grad_norm_pre_clip': 11.705488204956055, 'timing/data': 0.0004045770037919283, 'timing/model': 1.195698861964047, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 1.71}) +[2026-06-24 04:48:55,393][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.036403775215148926, 'train/grad_norm_pre_clip': 14.388510704040527, 'timing/data': 0.000429899082519114, 'timing/model': 1.1189897579606622, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 1.72}) +[2026-06-24 04:48:59,473][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.015573051758110523, 'train/grad_norm_pre_clip': 8.408632278442383, 'timing/data': 0.0006897388957440853, 'timing/model': 1.1171699349069968, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 1.72}) +[2026-06-24 04:49:03,563][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.02123013325035572, 'train/grad_norm_pre_clip': 7.094874858856201, 'timing/data': 0.00038478802889585495, 'timing/model': 1.1134926360100508, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 1.72}) +[2026-06-24 04:49:07,662][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.07079607993364334, 'train/grad_norm_pre_clip': 40.91596603393555, 'timing/data': 0.0004126960411667824, 'timing/model': 1.129788788035512, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 1.72}) +[2026-06-24 04:49:11,871][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.031224552541971207, 'train/grad_norm_pre_clip': 7.003820896148682, 'timing/data': 0.0006733109476044774, 'timing/model': 1.1189973509171978, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 1.72}) +[2026-06-24 04:49:15,956][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.016616905108094215, 'train/grad_norm_pre_clip': 22.165237426757812, 'timing/data': 0.00036651501432061195, 'timing/model': 1.1215057669905946, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 1.72}) +[2026-06-24 04:49:20,042][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.005067531950771809, 'train/grad_norm_pre_clip': 16.055540084838867, 'timing/data': 0.0005945010343566537, 'timing/model': 1.1200283010257408, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 1.72}) +[2026-06-24 04:49:24,121][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.10875238478183746, 'train/grad_norm_pre_clip': 27.122514724731445, 'timing/data': 0.00046991591807454824, 'timing/model': 1.1167493389220908, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 1.72}) +[2026-06-24 04:49:28,348][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.010103680193424225, 'train/grad_norm_pre_clip': 35.33155059814453, 'timing/data': 0.0003959150053560734, 'timing/model': 1.1193611500784755, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 1.72}) +[2026-06-24 04:49:32,477][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.04054116830229759, 'train/grad_norm_pre_clip': 16.236032485961914, 'timing/data': 0.00036643107887357473, 'timing/model': 1.126908466918394, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 1.72}) +[2026-06-24 04:49:36,579][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.030255883932113647, 'train/grad_norm_pre_clip': 26.726274490356445, 'timing/data': 0.0004623780259862542, 'timing/model': 1.1224593839142472, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 1.72}) +[2026-06-24 04:49:40,665][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.0002613414835650474, 'train/grad_norm_pre_clip': 15.899124145507812, 'timing/data': 0.00036510906647890806, 'timing/model': 1.1207812760258093, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 1.72}) +[2026-06-24 04:49:44,829][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.011269673705101013, 'train/grad_norm_pre_clip': 26.71230125427246, 'timing/data': 0.0005574519746005535, 'timing/model': 1.1243256379384547, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 1.72}) +[2026-06-24 04:49:48,921][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.05382469668984413, 'train/grad_norm_pre_clip': 23.712934494018555, 'timing/data': 0.0003733079647645354, 'timing/model': 1.1143253560876474, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 1.72}) +[2026-06-24 04:49:53,018][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.077568419277668, 'train/grad_norm_pre_clip': 14.006183624267578, 'timing/data': 0.00038303888868540525, 'timing/model': 1.1198551049456, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 1.72}) +[2026-06-24 04:49:57,095][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.06236789748072624, 'train/grad_norm_pre_clip': 18.797388076782227, 'timing/data': 0.00037717202212661505, 'timing/model': 1.115188185009174, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 1.72}) +[2026-06-24 04:50:01,248][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.015422137454152107, 'train/grad_norm_pre_clip': 19.667587280273438, 'timing/data': 0.00037745595909655094, 'timing/model': 1.1227524359710515, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 1.72}) +[2026-06-24 04:50:05,338][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.20439131557941437, 'train/grad_norm_pre_clip': 41.31951141357422, 'timing/data': 0.00036366796121001244, 'timing/model': 1.1269336889963597, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 1.72}) +[2026-06-24 04:50:09,427][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.004100424237549305, 'train/grad_norm_pre_clip': 14.424787521362305, 'timing/data': 0.000559723936021328, 'timing/model': 1.1188492679502815, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 1.72}) +[2026-06-24 04:50:13,517][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.021657640114426613, 'train/grad_norm_pre_clip': 11.385135650634766, 'timing/data': 0.00036603701300919056, 'timing/model': 1.116560759022832, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 1.72}) +[2026-06-24 04:50:17,768][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.00263748480938375, 'train/grad_norm_pre_clip': 34.208595275878906, 'timing/data': 0.0005484060384333134, 'timing/model': 1.1201917720027268, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 1.72}) +[2026-06-24 04:50:21,883][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.0991596132516861, 'train/grad_norm_pre_clip': 25.693084716796875, 'timing/data': 0.00036061194259673357, 'timing/model': 1.1361454100115225, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 1.72}) +[2026-06-24 04:50:25,978][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.014519969001412392, 'train/grad_norm_pre_clip': 9.470686912536621, 'timing/data': 0.00035712600219994783, 'timing/model': 1.1213287289720029, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 1.72}) +[2026-06-24 04:50:30,082][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.024801768362522125, 'train/grad_norm_pre_clip': 15.350757598876953, 'timing/data': 0.0003600440686568618, 'timing/model': 1.1288473480381072, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 1.72}) +[2026-06-24 04:50:34,336][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.02718663588166237, 'train/grad_norm_pre_clip': 9.425765037536621, 'timing/data': 0.0003588759573176503, 'timing/model': 1.1177569119026884, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 1.72}) +[2026-06-24 04:50:38,427][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.06883355230093002, 'train/grad_norm_pre_clip': 15.456009864807129, 'timing/data': 0.00036923191510140896, 'timing/model': 1.1167832110077143, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 1.72}) +[2026-06-24 04:50:42,518][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.03504876047372818, 'train/grad_norm_pre_clip': 18.375404357910156, 'timing/data': 0.0005441230023279786, 'timing/model': 1.1204487149370834, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 1.73}) +[2026-06-24 04:50:46,606][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.001587510691024363, 'train/grad_norm_pre_clip': 22.46999740600586, 'timing/data': 0.00048798706848174334, 'timing/model': 1.1179458550177515, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 1.73}) +[2026-06-24 04:50:50,863][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.002964078215882182, 'train/grad_norm_pre_clip': 8.308050155639648, 'timing/data': 0.000457008951343596, 'timing/model': 1.1210047119529918, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 1.73}) +[2026-06-24 04:50:54,978][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.007539681158959866, 'train/grad_norm_pre_clip': 21.4327449798584, 'timing/data': 0.00035858398769050837, 'timing/model': 1.1256301219109446, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 1.73}) +[2026-06-24 04:50:59,080][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.03470316901803017, 'train/grad_norm_pre_clip': 28.861867904663086, 'timing/data': 0.00041111395694315434, 'timing/model': 1.1235596359474584, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 1.73}) +[2026-06-24 04:51:03,179][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.07607445865869522, 'train/grad_norm_pre_clip': 22.58594512939453, 'timing/data': 0.00035563099663704634, 'timing/model': 1.1260536659974605, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 1.73}) +[2026-06-24 04:51:07,444][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.020253267139196396, 'train/grad_norm_pre_clip': 18.505569458007812, 'timing/data': 0.0004065579269081354, 'timing/model': 1.1181854419410229, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 1.73}) +[2026-06-24 04:51:11,524][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.011242436245083809, 'train/grad_norm_pre_clip': 7.149700164794922, 'timing/data': 0.0003609759733080864, 'timing/model': 1.117304106010124, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 1.73}) +[2026-06-24 04:51:15,612][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.0005338651826605201, 'train/grad_norm_pre_clip': 17.77037811279297, 'timing/data': 0.00036268297117203474, 'timing/model': 1.1175822829827666, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 1.73}) +[2026-06-24 04:51:19,698][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.11032146215438843, 'train/grad_norm_pre_clip': 26.89913558959961, 'timing/data': 0.0003652020823210478, 'timing/model': 1.1204845969332382, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 1.73}) +[2026-06-24 04:51:23,980][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.04681074246764183, 'train/grad_norm_pre_clip': 14.208379745483398, 'timing/data': 0.0003715530037879944, 'timing/model': 1.1231606259243563, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 1.73}) +[2026-06-24 04:51:28,069][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.18686044216156006, 'train/grad_norm_pre_clip': 24.13605308532715, 'timing/data': 0.0003689670702442527, 'timing/model': 1.1189118899637833, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 1.73}) +[2026-06-24 04:51:32,166][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.025798005983233452, 'train/grad_norm_pre_clip': 15.181304931640625, 'timing/data': 0.0003661270020529628, 'timing/model': 1.120511573040858, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 1.73}) +[2026-06-24 04:51:36,260][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.014217210933566093, 'train/grad_norm_pre_clip': 20.780805587768555, 'timing/data': 0.0005500820698216558, 'timing/model': 1.1210994139546528, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 1.73}) +[2026-06-24 04:51:40,512][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.02854613959789276, 'train/grad_norm_pre_clip': 11.823039054870605, 'timing/data': 0.00038493110332638025, 'timing/model': 1.115283778985031, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 1.73}) +[2026-06-24 04:51:44,604][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.019091704860329628, 'train/grad_norm_pre_clip': 11.681608200073242, 'timing/data': 0.0003625639947131276, 'timing/model': 1.1175138510297984, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 1.73}) +[2026-06-24 04:51:48,702][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.07824164628982544, 'train/grad_norm_pre_clip': 14.386919975280762, 'timing/data': 0.0004200390540063381, 'timing/model': 1.123456089058891, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 1.73}) +[2026-06-24 04:51:52,781][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.0343884713947773, 'train/grad_norm_pre_clip': 28.103073120117188, 'timing/data': 0.0004093300085514784, 'timing/model': 1.1218577380059287, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 1.73}) +[2026-06-24 04:51:57,049][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.0535908043384552, 'train/grad_norm_pre_clip': 23.763261795043945, 'timing/data': 0.0003731789765879512, 'timing/model': 1.1207971840631217, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 1.73}) +[2026-06-24 04:52:01,155][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.10918430984020233, 'train/grad_norm_pre_clip': 29.581066131591797, 'timing/data': 0.0004791710525751114, 'timing/model': 1.1303933860035613, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 1.73}) +[2026-06-24 04:52:05,239][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.03086099959909916, 'train/grad_norm_pre_clip': 26.985898971557617, 'timing/data': 0.00036463700234889984, 'timing/model': 1.1180735619273037, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 1.73}) +[2026-06-24 04:52:09,321][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.0030588326044380665, 'train/grad_norm_pre_clip': 23.493337631225586, 'timing/data': 0.0003665459807962179, 'timing/model': 1.119252154021524, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 1.73}) +[2026-06-24 04:52:13,586][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.0006664691609330475, 'train/grad_norm_pre_clip': 28.456499099731445, 'timing/data': 0.0004000440239906311, 'timing/model': 1.1214548759162426, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 1.73}) +[2026-06-24 04:52:17,677][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.08645595610141754, 'train/grad_norm_pre_clip': 35.2656364440918, 'timing/data': 0.00037412403617054224, 'timing/model': 1.1222292869351804, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 1.73}) +[2026-06-24 04:52:21,766][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.09990102797746658, 'train/grad_norm_pre_clip': 36.97501754760742, 'timing/data': 0.00037315208464860916, 'timing/model': 1.1221105681033805, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 1.73}) +[2026-06-24 04:52:25,863][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.009563235566020012, 'train/grad_norm_pre_clip': 14.668876647949219, 'timing/data': 0.0006026050541549921, 'timing/model': 1.1185305579565465, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 1.74}) +[2026-06-24 04:52:30,115][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.07073746621608734, 'train/grad_norm_pre_clip': 21.38780975341797, 'timing/data': 0.00041431200224906206, 'timing/model': 1.1209482479607686, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 1.74}) +[2026-06-24 04:52:34,194][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.18457823991775513, 'train/grad_norm_pre_clip': 22.54988670349121, 'timing/data': 0.00036172906402498484, 'timing/model': 1.1165538339409977, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 1.74}) +[2026-06-24 04:52:38,290][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.009315351955592632, 'train/grad_norm_pre_clip': 6.513003349304199, 'timing/data': 0.00037139898631721735, 'timing/model': 1.1253618269693106, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 1.74}) +[2026-06-24 04:52:42,373][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.03176001459360123, 'train/grad_norm_pre_clip': 54.82990264892578, 'timing/data': 0.0004177771043032408, 'timing/model': 1.116266967030242, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 1.74}) +[2026-06-24 04:52:46,625][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.028894569724798203, 'train/grad_norm_pre_clip': 18.04026222229004, 'timing/data': 0.000357098993845284, 'timing/model': 1.1232689570169896, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 1.74}) +[2026-06-24 04:52:50,718][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.00017204183677677065, 'train/grad_norm_pre_clip': 3.6221184730529785, 'timing/data': 0.00037334603257477283, 'timing/model': 1.1207273539621383, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 1.74}) +[2026-06-24 04:52:54,801][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.055104538798332214, 'train/grad_norm_pre_clip': 14.093255996704102, 'timing/data': 0.000390241970308125, 'timing/model': 1.1241446590283886, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 1.74}) +[2026-06-24 04:52:58,892][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.02307608351111412, 'train/grad_norm_pre_clip': 20.602895736694336, 'timing/data': 0.0005326749524101615, 'timing/model': 1.1228206909727305, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 1.74}) +[2026-06-24 04:53:03,150][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.08403370529413223, 'train/grad_norm_pre_clip': 25.82345962524414, 'timing/data': 0.0003713020123541355, 'timing/model': 1.116807951941155, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 1.74}) +[2026-06-24 04:53:07,240][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.06250894069671631, 'train/grad_norm_pre_clip': 27.8316650390625, 'timing/data': 0.00036754796747118235, 'timing/model': 1.1128976779291406, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 1.74}) +[2026-06-24 04:53:11,326][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.005423466209322214, 'train/grad_norm_pre_clip': 7.896816730499268, 'timing/data': 0.0003888549981638789, 'timing/model': 1.1173984149936587, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 1.74}) +[2026-06-24 04:53:15,426][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.016924802213907242, 'train/grad_norm_pre_clip': 22.361644744873047, 'timing/data': 0.00041132501792162657, 'timing/model': 1.1195252690231428, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 1.74}) +[2026-06-24 04:53:19,683][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.034075766801834106, 'train/grad_norm_pre_clip': 26.97414207458496, 'timing/data': 0.0003708330914378166, 'timing/model': 1.1206289740512148, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 1.74}) +[2026-06-24 04:53:23,769][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.025861015543341637, 'train/grad_norm_pre_clip': 28.309480667114258, 'timing/data': 0.0003755539655685425, 'timing/model': 1.1205277410335839, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 1.74}) +[2026-06-24 04:53:27,860][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.0009196478058584034, 'train/grad_norm_pre_clip': 20.196109771728516, 'timing/data': 0.000361626036465168, 'timing/model': 1.1263224410358816, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 1.74}) +[2026-06-24 04:53:31,946][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.006715117488056421, 'train/grad_norm_pre_clip': 15.46180534362793, 'timing/data': 0.00044243806041777134, 'timing/model': 1.1195543399080634, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 1.74}) +[2026-06-24 04:53:36,192][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4500 +[2026-06-24 04:54:39,840][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4500 +[2026-06-24 04:54:39,841][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4500 +[2026-06-24 04:55:19,388][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4500 +[2026-06-24 05:05:10,114][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4500 +[2026-06-24 05:05:10,114][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.10930173844099045, 'train/grad_norm_pre_clip': 47.3811149597168, 'eval/action_loss/seconds': 63.644837424042635, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.029454614967107773, 'eval/latency_2/loss': 0.029454614967107773, 'eval/flappy/loss': 0.029454614967107773, 'eval/flappy/latency_2/loss': 0.029454614967107773, 'eval/flap_precision': 0.9834938101788171, 'eval/flap_recall': 0.9701492537313433, 'eval/flap_f1': 0.9767759562841531, 'eval/flap_rate_model': 0.20194444444444445, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9767759562841531, 'eval/latency_2/flap_precision': 0.9834938101788171, 'eval/latency_2/flap_recall': 0.9701492537313433, 'eval/latency_2/flap_f1': 0.9767759562841531, 'eval/latency_2/flap_rate_model': 0.20194444444444445, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9767759562841531, 'eval/action_classification/data_wait_seconds': 0.4766004681587219, 'eval/action_classification/predict_seconds': 38.932899475097656, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 365.76000000002057, 'rl_games_eval/mid_train/mean_length': 2965.0, 'rl_games_eval/mid_train/std_reward': 157.6800000000101, 'rl_games_eval/mid_train/std_length': 1270.0, 'rl_games_eval/mid_train/macro_mean_reward': 365.76000000002057, 'rl_games_eval/mid_train/macro_mean_length': 2965.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 365.76000000002057, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2965.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 157.6800000000101, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1270.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 365.76000000002057, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0004191729240119457, 'timing/model': 1.1213855959940702, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 1.74}) +[2026-06-24 05:05:10,114][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_4500_state +[2026-06-24 05:05:10,117][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 05:05:13,791][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_4500_state/model.safetensors +[2026-06-24 05:05:21,609][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_4500_state/optimizer.bin +[2026-06-24 05:05:21,609][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_4500_state/sampler.bin +[2026-06-24 05:05:21,610][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-24 05:05:21,611][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-24 05:05:21,699][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-24 05:05:21,699][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 05:05:21,714][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 05:05:25,680][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.013902059756219387, 'train/grad_norm_pre_clip': 9.035720825195312, 'timing/data': 0.00036435003858059645, 'timing/model': 1.088528270018287, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 1.74}) +[2026-06-24 05:05:29,660][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.049751609563827515, 'train/grad_norm_pre_clip': 14.792757987976074, 'timing/data': 0.0003614160232245922, 'timing/model': 1.0953816380351782, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 1.74}) +[2026-06-24 05:05:33,643][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.03411458805203438, 'train/grad_norm_pre_clip': 22.336441040039062, 'timing/data': 0.0003771550254896283, 'timing/model': 1.087756845052354, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 1.74}) +[2026-06-24 05:05:37,789][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.007773935329169035, 'train/grad_norm_pre_clip': 3.360231876373291, 'timing/data': 0.00035389396362006664, 'timing/model': 1.0966914090095088, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 1.74}) +[2026-06-24 05:05:41,791][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.008132601156830788, 'train/grad_norm_pre_clip': 10.31941032409668, 'timing/data': 0.00036400905810296535, 'timing/model': 1.099104352062568, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 1.74}) +[2026-06-24 05:05:45,804][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.017038773745298386, 'train/grad_norm_pre_clip': 7.964000225067139, 'timing/data': 0.000387403997592628, 'timing/model': 1.106205676915124, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 1.74}) +[2026-06-24 05:05:49,822][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.07981562614440918, 'train/grad_norm_pre_clip': 29.103914260864258, 'timing/data': 0.0003597869072109461, 'timing/model': 1.1010764319216833, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 1.74}) +[2026-06-24 05:05:53,962][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.005729558877646923, 'train/grad_norm_pre_clip': 10.872925758361816, 'timing/data': 0.00037507794331759214, 'timing/model': 1.1026122419862077, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 1.74}) +[2026-06-24 05:05:57,982][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.0018882889999076724, 'train/grad_norm_pre_clip': 4.3023858070373535, 'timing/data': 0.00037004402838647366, 'timing/model': 1.0999407349154353, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 1.75}) +[2026-06-24 05:06:02,020][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.027039621025323868, 'train/grad_norm_pre_clip': 17.4522647857666, 'timing/data': 0.0003616069443523884, 'timing/model': 1.1063817880349234, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 1.75}) +[2026-06-24 05:06:06,138][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.22297528386116028, 'train/grad_norm_pre_clip': 22.645479202270508, 'timing/data': 0.0003499910235404968, 'timing/model': 1.1937900809571147, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 1.75}) +[2026-06-24 05:06:10,210][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.17275767028331757, 'train/grad_norm_pre_clip': 26.092609405517578, 'timing/data': 0.0003564079524949193, 'timing/model': 1.1171359299914911, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 1.75}) +[2026-06-24 05:06:14,258][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.026254922151565552, 'train/grad_norm_pre_clip': 21.96013069152832, 'timing/data': 0.0007053279550746083, 'timing/model': 1.1011167090618983, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 1.75}) +[2026-06-24 05:06:18,310][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.0003459133149590343, 'train/grad_norm_pre_clip': 21.335416793823242, 'timing/data': 0.0004912971053272486, 'timing/model': 1.1114340389613062, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 1.75}) +[2026-06-24 05:06:22,364][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.011480051092803478, 'train/grad_norm_pre_clip': 18.17464828491211, 'timing/data': 0.00036910490598529577, 'timing/model': 1.1170040650758892, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 1.75}) +[2026-06-24 05:06:26,535][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.0016615010099485517, 'train/grad_norm_pre_clip': 22.268766403198242, 'timing/data': 0.0005231579998508096, 'timing/model': 1.1078081110026687, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 1.75}) +[2026-06-24 05:06:30,592][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.014394311234354973, 'train/grad_norm_pre_clip': 25.374832153320312, 'timing/data': 0.0003649329300969839, 'timing/model': 1.1113031749846414, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 1.75}) +[2026-06-24 05:06:34,645][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.0015078875003382564, 'train/grad_norm_pre_clip': 10.131958961486816, 'timing/data': 0.0003695610212162137, 'timing/model': 1.1131869649980217, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 1.75}) +[2026-06-24 05:06:38,699][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.0740533098578453, 'train/grad_norm_pre_clip': 20.024585723876953, 'timing/data': 0.0003545429790392518, 'timing/model': 1.1168010979890823, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 1.75}) +[2026-06-24 05:06:42,921][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.0055577741004526615, 'train/grad_norm_pre_clip': 7.222906112670898, 'timing/data': 0.00039602897595614195, 'timing/model': 1.1151851039612666, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 1.75}) +[2026-06-24 05:06:46,999][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 3.8050209695938975e-05, 'train/grad_norm_pre_clip': 13.733930587768555, 'timing/data': 0.0003573480062186718, 'timing/model': 1.1166085290024057, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 1.75}) +[2026-06-24 05:06:51,072][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.015078176744282246, 'train/grad_norm_pre_clip': 8.835684776306152, 'timing/data': 0.00036373198963701725, 'timing/model': 1.1203356640180573, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 1.75}) +[2026-06-24 05:06:55,144][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.04418187960982323, 'train/grad_norm_pre_clip': 23.869243621826172, 'timing/data': 0.0005022289697080851, 'timing/model': 1.1114189849467948, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 1.75}) +[2026-06-24 05:06:59,290][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.041245199739933014, 'train/grad_norm_pre_clip': 16.96775245666504, 'timing/data': 0.000520800007507205, 'timing/model': 1.1176015849923715, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 1.75}) +[2026-06-24 05:07:03,360][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.0119488425552845, 'train/grad_norm_pre_clip': 19.13226890563965, 'timing/data': 0.0003656910266727209, 'timing/model': 1.1175203440943733, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 1.75}) +[2026-06-24 05:07:07,427][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.07208719849586487, 'train/grad_norm_pre_clip': 30.798933029174805, 'timing/data': 0.0003513199044391513, 'timing/model': 1.117050496977754, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 1.75}) +[2026-06-24 05:07:11,507][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.06470876187086105, 'train/grad_norm_pre_clip': 25.66954803466797, 'timing/data': 0.0003777539823204279, 'timing/model': 1.1114261830225587, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 1.75}) +[2026-06-24 05:07:15,671][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.11264237761497498, 'train/grad_norm_pre_clip': 41.09688186645508, 'timing/data': 0.000348287052474916, 'timing/model': 1.1228364059934393, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 1.75}) +[2026-06-24 05:07:19,737][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.036390867084264755, 'train/grad_norm_pre_clip': 13.007012367248535, 'timing/data': 0.0003564140060916543, 'timing/model': 1.1134053590940312, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 1.75}) +[2026-06-24 05:07:23,814][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.042515140026807785, 'train/grad_norm_pre_clip': 9.714727401733398, 'timing/data': 0.0006005000323057175, 'timing/model': 1.1213594850851223, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 1.75}) +[2026-06-24 05:07:27,890][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.017017614096403122, 'train/grad_norm_pre_clip': 29.584758758544922, 'timing/data': 0.00043219001963734627, 'timing/model': 1.118606490897946, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 1.75}) +[2026-06-24 05:07:32,137][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.16225063800811768, 'train/grad_norm_pre_clip': 31.09927749633789, 'timing/data': 0.0006263749673962593, 'timing/model': 1.116251477971673, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 1.75}) +[2026-06-24 05:07:36,220][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.02035713382065296, 'train/grad_norm_pre_clip': 16.733476638793945, 'timing/data': 0.0003612689906731248, 'timing/model': 1.1288168539758772, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 1.75}) +[2026-06-24 05:07:40,291][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.05739948898553848, 'train/grad_norm_pre_clip': 24.884017944335938, 'timing/data': 0.0004150710301473737, 'timing/model': 1.1117567690089345, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 1.75}) +[2026-06-24 05:07:44,385][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.018738482147455215, 'train/grad_norm_pre_clip': 10.082818984985352, 'timing/data': 0.00035963498521596193, 'timing/model': 1.1164703340036795, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 1.76}) +[2026-06-24 05:07:48,628][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.03332675248384476, 'train/grad_norm_pre_clip': 35.470489501953125, 'timing/data': 0.0005091129569336772, 'timing/model': 1.1174997839843854, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 1.76}) +[2026-06-24 05:07:52,699][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.01597221940755844, 'train/grad_norm_pre_clip': 24.72678565979004, 'timing/data': 0.00037182506639510393, 'timing/model': 1.1191304080421105, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 1.76}) +[2026-06-24 05:07:56,777][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.01791597343981266, 'train/grad_norm_pre_clip': 22.061843872070312, 'timing/data': 0.00036978593561798334, 'timing/model': 1.1194524849997833, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 1.76}) +[2026-06-24 05:08:00,854][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.040150128304958344, 'train/grad_norm_pre_clip': 15.355097770690918, 'timing/data': 0.0006070180097594857, 'timing/model': 1.1132119100075215, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 1.76}) +[2026-06-24 05:08:05,134][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.01782337576150894, 'train/grad_norm_pre_clip': 6.021035671234131, 'timing/data': 0.00036884495057165623, 'timing/model': 1.1237309629796073, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 1.76}) +[2026-06-24 05:08:09,220][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.055681463330984116, 'train/grad_norm_pre_clip': 21.952299118041992, 'timing/data': 0.0004138509975746274, 'timing/model': 1.116404709056951, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 1.76}) +[2026-06-24 05:08:13,291][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.006098070647567511, 'train/grad_norm_pre_clip': 29.369029998779297, 'timing/data': 0.0003616239409893751, 'timing/model': 1.118122979067266, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 1.76}) +[2026-06-24 05:08:17,368][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.022661680355668068, 'train/grad_norm_pre_clip': 7.149109840393066, 'timing/data': 0.0003747519804164767, 'timing/model': 1.1162517240736634, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 1.76}) +[2026-06-24 05:08:21,608][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.007029516156762838, 'train/grad_norm_pre_clip': 20.8083438873291, 'timing/data': 0.0003611749270930886, 'timing/model': 1.1198383109876886, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 1.76}) +[2026-06-24 05:08:25,690][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.04720551520586014, 'train/grad_norm_pre_clip': 15.33216667175293, 'timing/data': 0.00036929501220583916, 'timing/model': 1.1173705169931054, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 1.76}) +[2026-06-24 05:08:29,763][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.08751211315393448, 'train/grad_norm_pre_clip': 16.407793045043945, 'timing/data': 0.00048677308950573206, 'timing/model': 1.1138693268876523, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 1.76}) +[2026-06-24 05:08:33,836][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.03918463736772537, 'train/grad_norm_pre_clip': 11.342253684997559, 'timing/data': 0.0005058519309386611, 'timing/model': 1.1140655119670555, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 1.76}) +[2026-06-24 05:08:38,089][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.0013935166643932462, 'train/grad_norm_pre_clip': 21.20330047607422, 'timing/data': 0.0005684460047632456, 'timing/model': 1.1222713940078393, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 1.76}) +[2026-06-24 05:08:42,164][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.008924277499318123, 'train/grad_norm_pre_clip': 33.64960861206055, 'timing/data': 0.0004796970169991255, 'timing/model': 1.1152868120698258, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 1.76}) +[2026-06-24 05:08:46,245][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.004029818344861269, 'train/grad_norm_pre_clip': 27.218679428100586, 'timing/data': 0.00037190807051956654, 'timing/model': 1.1226033819839358, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 1.76}) +[2026-06-24 05:08:50,352][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.0093463696539402, 'train/grad_norm_pre_clip': 33.48259353637695, 'timing/data': 0.0004006529925391078, 'timing/model': 1.1174960780190304, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 1.76}) +[2026-06-24 05:08:54,598][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.04613635316491127, 'train/grad_norm_pre_clip': 21.983657836914062, 'timing/data': 0.0006075019482523203, 'timing/model': 1.1182388679590076, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 1.76}) +[2026-06-24 05:08:58,685][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.0016690499614924192, 'train/grad_norm_pre_clip': 9.386323928833008, 'timing/data': 0.0003771199844777584, 'timing/model': 1.1170171879930422, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 1.76}) +[2026-06-24 05:09:02,777][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.050379179418087006, 'train/grad_norm_pre_clip': 13.882854461669922, 'timing/data': 0.0003582310164347291, 'timing/model': 1.1230015309993178, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 1.76}) +[2026-06-24 05:09:06,874][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.09022894501686096, 'train/grad_norm_pre_clip': 20.995691299438477, 'timing/data': 0.00036181602627038956, 'timing/model': 1.12125753890723, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 1.76}) +[2026-06-24 05:09:11,145][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.019860517233610153, 'train/grad_norm_pre_clip': 10.227760314941406, 'timing/data': 0.0005267859669402242, 'timing/model': 1.1235562260262668, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 1.76}) +[2026-06-24 05:09:15,246][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.011554650962352753, 'train/grad_norm_pre_clip': 8.302556037902832, 'timing/data': 0.0004904159577563405, 'timing/model': 1.1288531390018761, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 1.76}) +[2026-06-24 05:09:19,331][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.21921545267105103, 'train/grad_norm_pre_clip': 47.72391891479492, 'timing/data': 0.0003702149260789156, 'timing/model': 1.1178658369462937, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 1.76}) +[2026-06-24 05:09:23,427][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.013220609165728092, 'train/grad_norm_pre_clip': 27.39602279663086, 'timing/data': 0.0003593970322981477, 'timing/model': 1.1214219400426373, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 1.76}) +[2026-06-24 05:09:27,676][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.08893497288227081, 'train/grad_norm_pre_clip': 36.41389083862305, 'timing/data': 0.00035956199280917645, 'timing/model': 1.1218881380045786, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 1.76}) +[2026-06-24 05:09:31,748][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.011634387075901031, 'train/grad_norm_pre_clip': 20.167076110839844, 'timing/data': 0.0003633289597928524, 'timing/model': 1.119073319947347, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 1.77}) +[2026-06-24 05:09:35,843][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.03691500797867775, 'train/grad_norm_pre_clip': 10.109759330749512, 'timing/data': 0.0003725129645317793, 'timing/model': 1.1222388909664005, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 1.77}) +[2026-06-24 05:09:39,932][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.016366085037589073, 'train/grad_norm_pre_clip': 27.92098045349121, 'timing/data': 0.0003877859562635422, 'timing/model': 1.1174131149891764, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 1.77}) +[2026-06-24 05:09:44,181][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.004782346077263355, 'train/grad_norm_pre_clip': 13.250417709350586, 'timing/data': 0.0003549120156094432, 'timing/model': 1.1168409589445218, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 1.77}) +[2026-06-24 05:09:48,288][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.0025869400706142187, 'train/grad_norm_pre_clip': 11.174485206604004, 'timing/data': 0.0005230559036135674, 'timing/model': 1.1347062001004815, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 1.77}) +[2026-06-24 05:09:52,414][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.07792338728904724, 'train/grad_norm_pre_clip': 16.872785568237305, 'timing/data': 0.00035114202182739973, 'timing/model': 1.1214899020269513, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 1.77}) +[2026-06-24 05:09:56,520][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.004093544092029333, 'train/grad_norm_pre_clip': 5.5174174308776855, 'timing/data': 0.00042494700755923986, 'timing/model': 1.1195443760370836, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 1.77}) +[2026-06-24 05:10:00,770][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.0007027898682281375, 'train/grad_norm_pre_clip': 7.0948004722595215, 'timing/data': 0.00039684795774519444, 'timing/model': 1.1175231010420248, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 1.77}) +[2026-06-24 05:10:04,853][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.000400432210881263, 'train/grad_norm_pre_clip': 26.447017669677734, 'timing/data': 0.00035626895260065794, 'timing/model': 1.1197402330581099, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 1.77}) +[2026-06-24 05:10:08,951][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.0022914670407772064, 'train/grad_norm_pre_clip': 16.1554012298584, 'timing/data': 0.0004988459404557943, 'timing/model': 1.1176649309927598, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 1.77}) +[2026-06-24 05:10:13,043][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.06788763403892517, 'train/grad_norm_pre_clip': 56.335506439208984, 'timing/data': 0.0004167939769104123, 'timing/model': 1.1148038669489324, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 1.77}) +[2026-06-24 05:10:17,300][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.008199358358979225, 'train/grad_norm_pre_clip': 19.825178146362305, 'timing/data': 0.0003492160467430949, 'timing/model': 1.1188423780258745, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 1.77}) +[2026-06-24 05:10:21,406][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.010131015442311764, 'train/grad_norm_pre_clip': 21.28974723815918, 'timing/data': 0.0006095990538597107, 'timing/model': 1.1191542249871418, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 1.77}) +[2026-06-24 05:10:25,512][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.003502539824694395, 'train/grad_norm_pre_clip': 17.932035446166992, 'timing/data': 0.00036795204505324364, 'timing/model': 1.1249263880308717, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 1.77}) +[2026-06-24 05:10:29,588][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.013738548383116722, 'train/grad_norm_pre_clip': 24.02812957763672, 'timing/data': 0.0003523150226101279, 'timing/model': 1.111713125021197, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 1.77}) +[2026-06-24 05:10:33,835][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 7.204533631011145e-06, 'train/grad_norm_pre_clip': 12.194095611572266, 'timing/data': 0.0005010140594094992, 'timing/model': 1.1216039409628138, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 1.77}) +[2026-06-24 05:10:37,926][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.006158710923045874, 'train/grad_norm_pre_clip': 17.63603973388672, 'timing/data': 0.00043533893767744303, 'timing/model': 1.120578240021132, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 1.77}) +[2026-06-24 05:10:42,040][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.00994290690869093, 'train/grad_norm_pre_clip': 7.171884536743164, 'timing/data': 0.0005909230094403028, 'timing/model': 1.134607812971808, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 1.77}) +[2026-06-24 05:10:46,146][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 4.6800825657555833e-05, 'train/grad_norm_pre_clip': 21.376771926879883, 'timing/data': 0.0003651590086519718, 'timing/model': 1.124507145024836, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 1.77}) +[2026-06-24 05:10:50,418][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.040113791823387146, 'train/grad_norm_pre_clip': 12.485020637512207, 'timing/data': 0.00036372896283864975, 'timing/model': 1.1198443330358714, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 1.77}) +[2026-06-24 05:10:54,508][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.02470102719962597, 'train/grad_norm_pre_clip': 19.66802406311035, 'timing/data': 0.00035564706195145845, 'timing/model': 1.1189549750415608, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 1.77}) +[2026-06-24 05:10:58,598][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.027562573552131653, 'train/grad_norm_pre_clip': 20.47121810913086, 'timing/data': 0.0003482080064713955, 'timing/model': 1.1202666540630162, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 1.77}) +[2026-06-24 05:11:02,698][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.01129245012998581, 'train/grad_norm_pre_clip': 14.363718032836914, 'timing/data': 0.0003613040316849947, 'timing/model': 1.1179771849419922, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 1.77}) +[2026-06-24 05:11:06,969][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.007335597183555365, 'train/grad_norm_pre_clip': 9.25637149810791, 'timing/data': 0.0006131390109658241, 'timing/model': 1.1215011790627614, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 1.77}) +[2026-06-24 05:11:11,082][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.10611024498939514, 'train/grad_norm_pre_clip': 35.12090301513672, 'timing/data': 0.0003572240239009261, 'timing/model': 1.138631785986945, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 1.77}) +[2026-06-24 05:11:15,180][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.012859249487519264, 'train/grad_norm_pre_clip': 30.32709312438965, 'timing/data': 0.0005109910853207111, 'timing/model': 1.1212693989509717, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 1.77}) +[2026-06-24 05:11:19,278][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.08014561235904694, 'train/grad_norm_pre_clip': 26.213281631469727, 'timing/data': 0.0003606859827414155, 'timing/model': 1.128340750001371, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 1.78}) +[2026-06-24 05:11:23,526][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.003447095165029168, 'train/grad_norm_pre_clip': 31.286212921142578, 'timing/data': 0.0003669540164992213, 'timing/model': 1.118170722038485, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 1.78}) +[2026-06-24 05:11:27,644][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.07249988615512848, 'train/grad_norm_pre_clip': 46.816009521484375, 'timing/data': 0.0004071949515491724, 'timing/model': 1.1216884639579803, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 1.78}) +[2026-06-24 05:11:31,729][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.006531415972858667, 'train/grad_norm_pre_clip': 7.415622711181641, 'timing/data': 0.00036124803591519594, 'timing/model': 1.118941031047143, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 1.78}) +[2026-06-24 05:11:35,834][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.0015472746454179287, 'train/grad_norm_pre_clip': 7.306827068328857, 'timing/data': 0.00035405007656663656, 'timing/model': 1.1189626530976966, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 1.78}) +[2026-06-24 05:11:40,105][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.15484853088855743, 'train/grad_norm_pre_clip': 35.279109954833984, 'timing/data': 0.00045382301323115826, 'timing/model': 1.1231095530092716, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 1.78}) +[2026-06-24 05:11:44,200][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.013556089252233505, 'train/grad_norm_pre_clip': 27.202096939086914, 'timing/data': 0.000360906939022243, 'timing/model': 1.124670173972845, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 1.78}) +[2026-06-24 05:11:48,285][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.04059673100709915, 'train/grad_norm_pre_clip': 19.742271423339844, 'timing/data': 0.0003667909186333418, 'timing/model': 1.1170947550563142, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 1.78}) +[2026-06-24 05:11:52,379][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.1288818120956421, 'train/grad_norm_pre_clip': 26.347431182861328, 'timing/data': 0.00044276402331888676, 'timing/model': 1.1192577900364995, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 1.78}) +[2026-06-24 05:11:56,664][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.016542872413992882, 'train/grad_norm_pre_clip': 20.638124465942383, 'timing/data': 0.0003770790062844753, 'timing/model': 1.1215140640269965, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 1.78}) +[2026-06-24 05:12:00,768][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.0006076183635741472, 'train/grad_norm_pre_clip': 8.30634593963623, 'timing/data': 0.000630503986030817, 'timing/model': 1.119474072009325, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 1.78}) +[2026-06-24 05:12:04,862][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.02281283400952816, 'train/grad_norm_pre_clip': 12.77459716796875, 'timing/data': 0.0004169010790064931, 'timing/model': 1.1230089619057253, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 1.78}) +[2026-06-24 05:12:08,944][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.001306117745116353, 'train/grad_norm_pre_clip': 34.867671966552734, 'timing/data': 0.00042316701728850603, 'timing/model': 1.115876876981929, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 1.78}) +[2026-06-24 05:12:13,186][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4600 +[2026-06-24 05:13:17,685][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4600 +[2026-06-24 05:13:17,686][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4600 +[2026-06-24 05:13:56,323][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4600 +[2026-06-24 05:13:56,324][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.00872328132390976, 'train/grad_norm_pre_clip': 13.603157043457031, 'eval/action_loss/seconds': 64.49674939305987, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.029960401356220245, 'eval/latency_2/loss': 0.029960401356220245, 'eval/flappy/loss': 0.029960401356220245, 'eval/flappy/latency_2/loss': 0.029960401356220245, 'eval/flap_precision': 0.9743243243243244, 'eval/flap_recall': 0.9782903663500678, 'eval/flap_f1': 0.976303317535545, 'eval/flap_rate_model': 0.20555555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.976303317535545, 'eval/latency_2/flap_precision': 0.9743243243243244, 'eval/latency_2/flap_recall': 0.9782903663500678, 'eval/latency_2/flap_f1': 0.976303317535545, 'eval/latency_2/flap_rate_model': 0.20555555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.976303317535545, 'eval/action_classification/data_wait_seconds': 0.5722213387489319, 'eval/action_classification/predict_seconds': 37.88551330566406, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0003627990372478962, 'timing/model': 1.1238031210377812, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 1.78}) +[2026-06-24 05:14:00,374][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.02192937210202217, 'train/grad_norm_pre_clip': 34.3045768737793, 'timing/data': 0.0003601850476115942, 'timing/model': 1.1085102380020544, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 1.78}) +[2026-06-24 05:14:04,449][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.008803004398941994, 'train/grad_norm_pre_clip': 26.793439865112305, 'timing/data': 0.0006674680626019835, 'timing/model': 1.1164316420909017, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 1.78}) +[2026-06-24 05:14:08,569][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.008656634949147701, 'train/grad_norm_pre_clip': 38.635616302490234, 'timing/data': 0.0006092219846323133, 'timing/model': 1.1051622109953314, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 1.78}) +[2026-06-24 05:14:12,633][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.01584874466061592, 'train/grad_norm_pre_clip': 28.863521575927734, 'timing/data': 0.0005786730907857418, 'timing/model': 1.1135966789443046, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 1.78}) +[2026-06-24 05:14:16,680][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.009246844798326492, 'train/grad_norm_pre_clip': 20.109546661376953, 'timing/data': 0.00035707198549062014, 'timing/model': 1.1190152199706063, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 1.78}) +[2026-06-24 05:14:20,752][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.004831913858652115, 'train/grad_norm_pre_clip': 8.175305366516113, 'timing/data': 0.0003658699570223689, 'timing/model': 1.1192068150267005, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 1.78}) +[2026-06-24 05:14:24,972][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.002397169591858983, 'train/grad_norm_pre_clip': 19.009445190429688, 'timing/data': 0.0003741389373317361, 'timing/model': 1.2758811230305582, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 1.78}) +[2026-06-24 05:14:29,040][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.10981250554323196, 'train/grad_norm_pre_clip': 35.75008010864258, 'timing/data': 0.00036835006903856993, 'timing/model': 1.110744925099425, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 1.78}) +[2026-06-24 05:14:33,114][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.05025969445705414, 'train/grad_norm_pre_clip': 30.10868263244629, 'timing/data': 0.0005254819989204407, 'timing/model': 1.1204217139165848, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 1.78}) +[2026-06-24 05:14:37,185][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.008044728077948093, 'train/grad_norm_pre_clip': 13.284549713134766, 'timing/data': 0.00041428511030972004, 'timing/model': 1.1133185980143026, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 1.78}) +[2026-06-24 05:14:41,331][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.005209203343838453, 'train/grad_norm_pre_clip': 21.136672973632812, 'timing/data': 0.0005145840113982558, 'timing/model': 1.1914063970325515, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 1.78}) +[2026-06-24 05:14:45,393][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.044412244111299515, 'train/grad_norm_pre_clip': 25.323556900024414, 'timing/data': 0.00036410195752978325, 'timing/model': 1.1175364289665595, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 1.78}) +[2026-06-24 05:14:49,464][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.0030051339417696, 'train/grad_norm_pre_clip': 14.029197692871094, 'timing/data': 0.00035853590816259384, 'timing/model': 1.1113561230013147, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 1.79}) +[2026-06-24 05:14:53,541][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.008887900039553642, 'train/grad_norm_pre_clip': 15.866708755493164, 'timing/data': 0.00036318995989859104, 'timing/model': 1.1205184110440314, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 1.79}) +[2026-06-24 05:14:57,694][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.010591249912977219, 'train/grad_norm_pre_clip': 10.048977851867676, 'timing/data': 0.0003555399598553777, 'timing/model': 1.1936408700421453, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 1.79}) +[2026-06-24 05:15:01,798][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.02783609926700592, 'train/grad_norm_pre_clip': 24.814008712768555, 'timing/data': 0.0003576850285753608, 'timing/model': 1.1183860199525952, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 1.79}) +[2026-06-24 05:15:05,893][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.002502177143469453, 'train/grad_norm_pre_clip': 8.838488578796387, 'timing/data': 0.00058412691578269, 'timing/model': 1.1211342830210924, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 1.79}) +[2026-06-24 05:15:09,983][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.024792347103357315, 'train/grad_norm_pre_clip': 23.653696060180664, 'timing/data': 0.00036907196044921875, 'timing/model': 1.1239067539572716, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 1.79}) +[2026-06-24 05:15:14,139][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.0006501275929622352, 'train/grad_norm_pre_clip': 21.88311195373535, 'timing/data': 0.0003614560700953007, 'timing/model': 1.1830205889418721, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 1.79}) +[2026-06-24 05:15:18,234][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.08806796371936798, 'train/grad_norm_pre_clip': 29.7852783203125, 'timing/data': 0.0004038539482280612, 'timing/model': 1.1243557899724692, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 1.79}) +[2026-06-24 05:15:22,313][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.002441330347210169, 'train/grad_norm_pre_clip': 18.607561111450195, 'timing/data': 0.0004120669327676296, 'timing/model': 1.1141897030174732, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 1.79}) +[2026-06-24 05:15:26,396][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.08888808637857437, 'train/grad_norm_pre_clip': 42.80856704711914, 'timing/data': 0.0004112400347366929, 'timing/model': 1.1178381990175694, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 1.79}) +[2026-06-24 05:15:30,624][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.005161059554666281, 'train/grad_norm_pre_clip': 35.09489440917969, 'timing/data': 0.0005483749555423856, 'timing/model': 1.270862493896857, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 1.79}) +[2026-06-24 05:15:34,705][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.06595855951309204, 'train/grad_norm_pre_clip': 32.66788864135742, 'timing/data': 0.00035670294892042875, 'timing/model': 1.1179945530602708, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 1.79}) +[2026-06-24 05:15:38,776][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 9.655638132244349e-05, 'train/grad_norm_pre_clip': 19.2808837890625, 'timing/data': 0.0003742979606613517, 'timing/model': 1.1173628319520503, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 1.79}) +[2026-06-24 05:15:42,849][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.014878746122121811, 'train/grad_norm_pre_clip': 16.935701370239258, 'timing/data': 0.0006372829666361213, 'timing/model': 1.1100159250199795, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 1.79}) +[2026-06-24 05:15:47,060][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.04733229801058769, 'train/grad_norm_pre_clip': 17.089696884155273, 'timing/data': 0.00037673197221010923, 'timing/model': 1.1217964430106804, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 1.79}) +[2026-06-24 05:15:51,154][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.06226012483239174, 'train/grad_norm_pre_clip': 39.915313720703125, 'timing/data': 0.0004248900804668665, 'timing/model': 1.1132616339018568, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 1.79}) +[2026-06-24 05:15:55,237][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.010813475586473942, 'train/grad_norm_pre_clip': 16.57219123840332, 'timing/data': 0.00036513805389404297, 'timing/model': 1.1174711950588971, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 1.79}) +[2026-06-24 05:15:59,318][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.012751072645187378, 'train/grad_norm_pre_clip': 23.823204040527344, 'timing/data': 0.0003599900519475341, 'timing/model': 1.1206224510679021, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 1.79}) +[2026-06-24 05:16:03,568][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.14593182504177094, 'train/grad_norm_pre_clip': 33.97019577026367, 'timing/data': 0.000569172902032733, 'timing/model': 1.114556216984056, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 1.79}) +[2026-06-24 05:16:07,653][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.020437538623809814, 'train/grad_norm_pre_clip': 36.0382080078125, 'timing/data': 0.00035872997250407934, 'timing/model': 1.120981484069489, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 1.79}) +[2026-06-24 05:16:11,748][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.10190453380346298, 'train/grad_norm_pre_clip': 19.446260452270508, 'timing/data': 0.00041817198507487774, 'timing/model': 1.1141778170131147, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 1.79}) +[2026-06-24 05:16:15,830][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.06390415877103806, 'train/grad_norm_pre_clip': 28.8576602935791, 'timing/data': 0.00035288603976368904, 'timing/model': 1.1199493410531431, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 1.79}) +[2026-06-24 05:16:20,097][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.055218543857336044, 'train/grad_norm_pre_clip': 21.743438720703125, 'timing/data': 0.0004073729505762458, 'timing/model': 1.1218000040389597, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 1.79}) +[2026-06-24 05:16:24,173][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.012995725497603416, 'train/grad_norm_pre_clip': 11.638671875, 'timing/data': 0.0005127020413056016, 'timing/model': 1.1153171369805932, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 1.79}) +[2026-06-24 05:16:28,245][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.01705726608633995, 'train/grad_norm_pre_clip': 10.793766975402832, 'timing/data': 0.0004139239899814129, 'timing/model': 1.1198708009906113, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 1.79}) +[2026-06-24 05:16:32,323][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.0019114748574793339, 'train/grad_norm_pre_clip': 29.8366756439209, 'timing/data': 0.0004190090112388134, 'timing/model': 1.120984980952926, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 1.79}) +[2026-06-24 05:16:36,482][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.004325926769524813, 'train/grad_norm_pre_clip': 17.43848991394043, 'timing/data': 0.00040850997902452946, 'timing/model': 1.1184976260410622, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 1.8}) +[2026-06-24 05:16:40,572][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.040980905294418335, 'train/grad_norm_pre_clip': 18.9267520904541, 'timing/data': 0.0005562870064750314, 'timing/model': 1.1267674480332062, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 1.8}) +[2026-06-24 05:16:44,659][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.014572393149137497, 'train/grad_norm_pre_clip': 18.895483016967773, 'timing/data': 0.00036481593269854784, 'timing/model': 1.1191163799958304, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 1.8}) +[2026-06-24 05:16:48,761][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.05185430496931076, 'train/grad_norm_pre_clip': 9.846970558166504, 'timing/data': 0.00044895708560943604, 'timing/model': 1.1308326030848548, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 1.8}) +[2026-06-24 05:16:52,933][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.13690459728240967, 'train/grad_norm_pre_clip': 44.67045211791992, 'timing/data': 0.00041578500531613827, 'timing/model': 1.1268940039444715, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 1.8}) +[2026-06-24 05:16:57,021][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.005236613564193249, 'train/grad_norm_pre_clip': 38.37567901611328, 'timing/data': 0.00035826500970870256, 'timing/model': 1.1203452229965478, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 1.8}) +[2026-06-24 05:17:01,101][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.02134518325328827, 'train/grad_norm_pre_clip': 11.39962387084961, 'timing/data': 0.00036027596797794104, 'timing/model': 1.1236401071073487, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 1.8}) +[2026-06-24 05:17:05,199][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.007449810393154621, 'train/grad_norm_pre_clip': 29.44560432434082, 'timing/data': 0.0003641609800979495, 'timing/model': 1.12123893504031, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 1.8}) +[2026-06-24 05:17:09,415][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.01318119652569294, 'train/grad_norm_pre_clip': 27.813982009887695, 'timing/data': 0.00035935803316533566, 'timing/model': 1.121462860959582, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 1.8}) +[2026-06-24 05:17:13,512][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.05816437304019928, 'train/grad_norm_pre_clip': 23.858848571777344, 'timing/data': 0.00035721808671951294, 'timing/model': 1.1203082649735734, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 1.8}) +[2026-06-24 05:17:17,593][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.006522969342768192, 'train/grad_norm_pre_clip': 14.97314453125, 'timing/data': 0.0003641319926828146, 'timing/model': 1.1166896370705217, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 1.8}) +[2026-06-24 05:17:21,762][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.017233526334166527, 'train/grad_norm_pre_clip': 11.718626022338867, 'timing/data': 0.00036382093094289303, 'timing/model': 1.196515467017889, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 1.8}) +[2026-06-24 05:17:25,842][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.028684264048933983, 'train/grad_norm_pre_clip': 24.56369972229004, 'timing/data': 0.00036230706609785557, 'timing/model': 1.123254581936635, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 1.8}) +[2026-06-24 05:17:29,924][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.0020147121977061033, 'train/grad_norm_pre_clip': 29.943252563476562, 'timing/data': 0.0003639949718490243, 'timing/model': 1.1165416759904474, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 1.8}) +[2026-06-24 05:17:34,010][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.0279763862490654, 'train/grad_norm_pre_clip': 18.35126304626465, 'timing/data': 0.0003741809632629156, 'timing/model': 1.121887910994701, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 1.8}) +[2026-06-24 05:17:38,110][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.0005830904701724648, 'train/grad_norm_pre_clip': 2.335904121398926, 'timing/data': 0.00038277602288872004, 'timing/model': 1.1227830519201234, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 1.8}) +[2026-06-24 05:17:42,311][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.05748642235994339, 'train/grad_norm_pre_clip': 15.101892471313477, 'timing/data': 0.0003672949969768524, 'timing/model': 1.1175704050110653, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 1.8}) +[2026-06-24 05:17:46,407][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.00209412002004683, 'train/grad_norm_pre_clip': 5.837771892547607, 'timing/data': 0.0004593069897964597, 'timing/model': 1.1218481769319624, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 1.8}) +[2026-06-24 05:17:50,477][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.12393379956483841, 'train/grad_norm_pre_clip': 20.59303855895996, 'timing/data': 0.00043593498412519693, 'timing/model': 1.1164584879297763, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 1.8}) +[2026-06-24 05:17:54,564][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.004284251946955919, 'train/grad_norm_pre_clip': 5.354214668273926, 'timing/data': 0.0005084149306640029, 'timing/model': 1.121388251078315, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 1.8}) +[2026-06-24 05:17:58,794][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.06801819056272507, 'train/grad_norm_pre_clip': 23.047340393066406, 'timing/data': 0.00041591899935156107, 'timing/model': 1.1173176100710407, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 1.8}) +[2026-06-24 05:18:02,875][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.01864180527627468, 'train/grad_norm_pre_clip': 31.007884979248047, 'timing/data': 0.0005723980721086264, 'timing/model': 1.1108505809679627, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 1.8}) +[2026-06-24 05:18:06,972][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.04240332171320915, 'train/grad_norm_pre_clip': 30.376026153564453, 'timing/data': 0.0003610339481383562, 'timing/model': 1.1187402090290561, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 1.8}) +[2026-06-24 05:18:11,056][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.002256283536553383, 'train/grad_norm_pre_clip': 24.580856323242188, 'timing/data': 0.0003645520191639662, 'timing/model': 1.1174051729030907, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 1.8}) +[2026-06-24 05:18:15,205][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.020301014184951782, 'train/grad_norm_pre_clip': 16.024580001831055, 'timing/data': 0.0003616160247474909, 'timing/model': 1.119018214987591, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 1.8}) +[2026-06-24 05:18:19,293][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.06041436642408371, 'train/grad_norm_pre_clip': 20.174165725708008, 'timing/data': 0.000372821930795908, 'timing/model': 1.1223651670152321, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 1.81}) +[2026-06-24 05:18:23,381][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.019191119819879532, 'train/grad_norm_pre_clip': 11.792902946472168, 'timing/data': 0.00036215700674802065, 'timing/model': 1.1192869059741497, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 1.81}) +[2026-06-24 05:18:27,475][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.10592906922101974, 'train/grad_norm_pre_clip': 25.038827896118164, 'timing/data': 0.00041542109102010727, 'timing/model': 1.115659083938226, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 1.81}) +[2026-06-24 05:18:31,635][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.0010844890493899584, 'train/grad_norm_pre_clip': 14.73925495147705, 'timing/data': 0.00040923396591097116, 'timing/model': 1.1201320979744196, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 1.81}) +[2026-06-24 05:18:35,723][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.03886021301150322, 'train/grad_norm_pre_clip': 11.742958068847656, 'timing/data': 0.0004187509184703231, 'timing/model': 1.1204543599160388, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 1.81}) +[2026-06-24 05:18:39,814][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.0033410675823688507, 'train/grad_norm_pre_clip': 22.513824462890625, 'timing/data': 0.0004005820956081152, 'timing/model': 1.125305443070829, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 1.81}) +[2026-06-24 05:18:43,913][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.02571137621998787, 'train/grad_norm_pre_clip': 19.30986976623535, 'timing/data': 0.00041660794522613287, 'timing/model': 1.1211100440705195, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 1.81}) +[2026-06-24 05:18:48,167][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.0004041938518639654, 'train/grad_norm_pre_clip': 13.102178573608398, 'timing/data': 0.00041247194167226553, 'timing/model': 1.1235787780024111, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 1.81}) +[2026-06-24 05:18:52,252][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.05077676475048065, 'train/grad_norm_pre_clip': 32.65134811401367, 'timing/data': 0.00036659406032413244, 'timing/model': 1.118115799035877, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 1.81}) +[2026-06-24 05:18:56,320][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.01361511368304491, 'train/grad_norm_pre_clip': 31.386131286621094, 'timing/data': 0.00037232309114187956, 'timing/model': 1.113195679965429, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 1.81}) +[2026-06-24 05:19:00,407][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.040973957628011703, 'train/grad_norm_pre_clip': 22.1761531829834, 'timing/data': 0.00040646304842084646, 'timing/model': 1.1206238080048934, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 1.81}) +[2026-06-24 05:19:04,656][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.03355610370635986, 'train/grad_norm_pre_clip': 27.966472625732422, 'timing/data': 0.000369054963812232, 'timing/model': 1.1166402429807931, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 1.81}) +[2026-06-24 05:19:08,732][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.05828390643000603, 'train/grad_norm_pre_clip': 26.877099990844727, 'timing/data': 0.00036478100810199976, 'timing/model': 1.1181478979997337, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 1.81}) +[2026-06-24 05:19:12,816][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.004502784926444292, 'train/grad_norm_pre_clip': 1.787083625793457, 'timing/data': 0.0003657159395515919, 'timing/model': 1.1218875619815663, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 1.81}) +[2026-06-24 05:19:16,913][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.05096808820962906, 'train/grad_norm_pre_clip': 15.104537010192871, 'timing/data': 0.00036854890640825033, 'timing/model': 1.1283626989461482, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 1.81}) +[2026-06-24 05:19:21,188][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.05750156566500664, 'train/grad_norm_pre_clip': 20.514314651489258, 'timing/data': 0.00035979796666651964, 'timing/model': 1.1119906819658354, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 1.81}) +[2026-06-24 05:19:25,287][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.009458404965698719, 'train/grad_norm_pre_clip': 12.370588302612305, 'timing/data': 0.0003667599521577358, 'timing/model': 1.1178870130097494, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 1.81}) +[2026-06-24 05:19:29,367][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.017914101481437683, 'train/grad_norm_pre_clip': 7.3016228675842285, 'timing/data': 0.00048595794942229986, 'timing/model': 1.1133301899535581, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 1.81}) +[2026-06-24 05:19:33,467][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.09944616258144379, 'train/grad_norm_pre_clip': 18.243114471435547, 'timing/data': 0.0003570510307326913, 'timing/model': 1.1231509681092575, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 1.81}) +[2026-06-24 05:19:37,715][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.0004690334026236087, 'train/grad_norm_pre_clip': 17.299259185791016, 'timing/data': 0.00041240197606384754, 'timing/model': 1.1228683929657564, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 1.81}) +[2026-06-24 05:19:41,793][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.03288152068853378, 'train/grad_norm_pre_clip': 12.373701095581055, 'timing/data': 0.0004951599985361099, 'timing/model': 1.1236808620160446, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 1.81}) +[2026-06-24 05:19:45,898][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.0066224923357367516, 'train/grad_norm_pre_clip': 29.639633178710938, 'timing/data': 0.0003690510056912899, 'timing/model': 1.124014948029071, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 1.81}) +[2026-06-24 05:19:49,999][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.030854428187012672, 'train/grad_norm_pre_clip': 23.903812408447266, 'timing/data': 0.0004494759486988187, 'timing/model': 1.1189795170212165, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 1.81}) +[2026-06-24 05:19:54,280][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.0257913488894701, 'train/grad_norm_pre_clip': 24.63402557373047, 'timing/data': 0.0003666570410132408, 'timing/model': 1.1220559609355405, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 1.81}) +[2026-06-24 05:19:58,371][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.0014907920267432928, 'train/grad_norm_pre_clip': 34.99838638305664, 'timing/data': 0.0005072180647403002, 'timing/model': 1.1221655319677666, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 1.81}) +[2026-06-24 05:20:02,462][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.0003134586731903255, 'train/grad_norm_pre_clip': 12.043928146362305, 'timing/data': 0.00036728999111801386, 'timing/model': 1.1207863689633086, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 1.81}) +[2026-06-24 05:20:06,554][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.0030318936333060265, 'train/grad_norm_pre_clip': 3.0998151302337646, 'timing/data': 0.00045033893547952175, 'timing/model': 1.1150875620078295, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 1.82}) +[2026-06-24 05:20:10,799][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.045549970120191574, 'train/grad_norm_pre_clip': 27.243465423583984, 'timing/data': 0.0003637690097093582, 'timing/model': 1.1150322280591354, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 1.82}) +[2026-06-24 05:20:14,896][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.004164915531873703, 'train/grad_norm_pre_clip': 17.80193328857422, 'timing/data': 0.00035509804729372263, 'timing/model': 1.124290001927875, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 1.82}) +[2026-06-24 05:20:18,982][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.02766185812652111, 'train/grad_norm_pre_clip': 39.92137145996094, 'timing/data': 0.00040179805364459753, 'timing/model': 1.1230210709618405, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 1.82}) +[2026-06-24 05:20:23,091][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.002991204150021076, 'train/grad_norm_pre_clip': 5.126522541046143, 'timing/data': 0.0003642119700089097, 'timing/model': 1.121855977922678, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 1.82}) +[2026-06-24 05:20:27,367][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.056017834693193436, 'train/grad_norm_pre_clip': 25.429651260375977, 'timing/data': 0.00045687006786465645, 'timing/model': 1.1243498109979555, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 1.82}) +[2026-06-24 05:20:31,445][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.018346795812249184, 'train/grad_norm_pre_clip': 18.249126434326172, 'timing/data': 0.0005403669783845544, 'timing/model': 1.1190009930869564, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 1.82}) +[2026-06-24 05:20:35,534][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.03404879570007324, 'train/grad_norm_pre_clip': 20.585718154907227, 'timing/data': 0.000359260942786932, 'timing/model': 1.1214061899809167, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 1.82}) +[2026-06-24 05:20:39,621][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.06613300740718842, 'train/grad_norm_pre_clip': 13.901434898376465, 'timing/data': 0.0003624140517786145, 'timing/model': 1.1205445230007172, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 1.82}) +[2026-06-24 05:20:43,882][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.07348808646202087, 'train/grad_norm_pre_clip': 32.43156051635742, 'timing/data': 0.0005120469722896814, 'timing/model': 1.1138533279299736, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 1.82}) +[2026-06-24 05:20:47,970][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4700 +[2026-06-24 05:21:52,749][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4700 +[2026-06-24 05:21:52,750][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4700 +[2026-06-24 05:22:31,669][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4700 +[2026-06-24 05:22:31,669][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.003460182808339596, 'train/grad_norm_pre_clip': 4.387932300567627, 'eval/action_loss/seconds': 64.77691013296135, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.029249053448438644, 'eval/latency_2/loss': 0.029249053448438644, 'eval/flappy/loss': 0.029249053448438644, 'eval/flappy/latency_2/loss': 0.029249053448438644, 'eval/flap_precision': 0.9768707482993197, 'eval/flap_recall': 0.9742198100407056, 'eval/flap_f1': 0.9755434782608695, 'eval/flap_rate_model': 0.20416666666666666, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9755434782608695, 'eval/latency_2/flap_precision': 0.9768707482993197, 'eval/latency_2/flap_recall': 0.9742198100407056, 'eval/latency_2/flap_f1': 0.9755434782608695, 'eval/latency_2/flap_rate_model': 0.20416666666666666, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9755434782608695, 'eval/action_classification/data_wait_seconds': 0.5281139612197876, 'eval/action_classification/predict_seconds': 38.2531852722168, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.0003682009410113096, 'timing/model': 1.1181589970365167, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 1.82}) +[2026-06-24 05:22:35,730][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.056537825614213943, 'train/grad_norm_pre_clip': 27.181604385375977, 'timing/data': 0.00035666103940457106, 'timing/model': 1.1144257290288806, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 1.82}) +[2026-06-24 05:22:39,914][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.005916215945035219, 'train/grad_norm_pre_clip': 41.09698486328125, 'timing/data': 0.0003550560213625431, 'timing/model': 1.2466331999748945, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 1.82}) +[2026-06-24 05:22:43,966][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.13042527437210083, 'train/grad_norm_pre_clip': 29.591257095336914, 'timing/data': 0.0003742449916899204, 'timing/model': 1.1110355149721727, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 1.82}) +[2026-06-24 05:22:48,019][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.04143521934747696, 'train/grad_norm_pre_clip': 29.100112915039062, 'timing/data': 0.0004488199483603239, 'timing/model': 1.1119161989772692, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 1.82}) +[2026-06-24 05:22:52,074][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.02884419821202755, 'train/grad_norm_pre_clip': 7.7877984046936035, 'timing/data': 0.00036724808160215616, 'timing/model': 1.1170383419375867, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 1.82}) +[2026-06-24 05:22:56,273][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.002783533651381731, 'train/grad_norm_pre_clip': 13.14959716796875, 'timing/data': 0.0005704729119315743, 'timing/model': 1.1126405290560797, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 1.82}) +[2026-06-24 05:23:00,354][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.1355416625738144, 'train/grad_norm_pre_clip': 28.14103126525879, 'timing/data': 0.0003599269548431039, 'timing/model': 1.112110078916885, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 1.82}) +[2026-06-24 05:23:04,424][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.20604859292507172, 'train/grad_norm_pre_clip': 43.17439651489258, 'timing/data': 0.0004429880063980818, 'timing/model': 1.115259998012334, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 1.82}) +[2026-06-24 05:23:08,501][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.03230016306042671, 'train/grad_norm_pre_clip': 16.71013641357422, 'timing/data': 0.0004543380346149206, 'timing/model': 1.1142931100912392, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 1.82}) +[2026-06-24 05:23:12,753][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.004738875664770603, 'train/grad_norm_pre_clip': 11.786128044128418, 'timing/data': 0.00036389101296663284, 'timing/model': 1.1157820009393618, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 1.82}) +[2026-06-24 05:23:16,815][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.013025090098381042, 'train/grad_norm_pre_clip': 27.719858169555664, 'timing/data': 0.00035695196129381657, 'timing/model': 1.1132829959969968, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 1.82}) +[2026-06-24 05:23:20,887][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.030234867706894875, 'train/grad_norm_pre_clip': 16.830188751220703, 'timing/data': 0.0005515110678970814, 'timing/model': 1.1159261640859768, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 1.82}) +[2026-06-24 05:23:24,959][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.0010021337075158954, 'train/grad_norm_pre_clip': 11.512089729309082, 'timing/data': 0.00040680496022105217, 'timing/model': 1.1169647499918938, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 1.82}) +[2026-06-24 05:23:29,208][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.051553577184677124, 'train/grad_norm_pre_clip': 13.331927299499512, 'timing/data': 0.00039737706538289785, 'timing/model': 1.1153433140134439, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 1.82}) +[2026-06-24 05:23:33,278][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.003851833287626505, 'train/grad_norm_pre_clip': 11.83229923248291, 'timing/data': 0.00036677997559309006, 'timing/model': 1.1149722000118345, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 1.82}) +[2026-06-24 05:23:37,339][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.11293778568506241, 'train/grad_norm_pre_clip': 20.903261184692383, 'timing/data': 0.0003919330192729831, 'timing/model': 1.1107788750668988, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 1.83}) +[2026-06-24 05:23:41,422][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.0657864362001419, 'train/grad_norm_pre_clip': 25.386127471923828, 'timing/data': 0.000532147940248251, 'timing/model': 1.1164850790519267, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 1.83}) +[2026-06-24 05:23:45,591][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.022006405517458916, 'train/grad_norm_pre_clip': 16.324758529663086, 'timing/data': 0.00047006900422275066, 'timing/model': 1.1157496490050107, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 1.83}) +[2026-06-24 05:23:49,659][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.07295387983322144, 'train/grad_norm_pre_clip': 18.825183868408203, 'timing/data': 0.00036705099046230316, 'timing/model': 1.115703584975563, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 1.83}) +[2026-06-24 05:23:53,745][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.017752865329384804, 'train/grad_norm_pre_clip': 16.282943725585938, 'timing/data': 0.0003665779950097203, 'timing/model': 1.1213349199388176, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 1.83}) +[2026-06-24 05:23:57,822][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.042237360030412674, 'train/grad_norm_pre_clip': 24.116905212402344, 'timing/data': 0.00037721393164247274, 'timing/model': 1.1162907510297373, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 1.83}) +[2026-06-24 05:24:01,982][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 4.1911040170816705e-05, 'train/grad_norm_pre_clip': 6.874950408935547, 'timing/data': 0.00036652095150202513, 'timing/model': 1.1204451080411673, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 1.83}) +[2026-06-24 05:24:06,193][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.0018743977416306734, 'train/grad_norm_pre_clip': 5.713672637939453, 'timing/data': 0.0003618299961090088, 'timing/model': 1.1290767430327833, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 1.83}) +[2026-06-24 05:24:10,282][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.005148088093847036, 'train/grad_norm_pre_clip': 13.320972442626953, 'timing/data': 0.0003545719664543867, 'timing/model': 1.12220253306441, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 1.83}) +[2026-06-24 05:24:14,367][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.007329849526286125, 'train/grad_norm_pre_clip': 17.744810104370117, 'timing/data': 0.00038423703517764807, 'timing/model': 1.1165757320122793, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 1.83}) +[2026-06-24 05:24:18,575][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.00936210062354803, 'train/grad_norm_pre_clip': 9.147673606872559, 'timing/data': 0.00036385597195476294, 'timing/model': 1.1286282429937273, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 1.83}) +[2026-06-24 05:24:22,663][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.04567749798297882, 'train/grad_norm_pre_clip': 18.796470642089844, 'timing/data': 0.00037191901355981827, 'timing/model': 1.1157053590286523, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 1.83}) +[2026-06-24 05:24:26,744][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.04262004792690277, 'train/grad_norm_pre_clip': 24.514713287353516, 'timing/data': 0.0003884279867634177, 'timing/model': 1.1187696980778128, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 1.83}) +[2026-06-24 05:24:30,892][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.0029091276228427887, 'train/grad_norm_pre_clip': 28.45078468322754, 'timing/data': 0.0005479829851537943, 'timing/model': 1.1895875870250165, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 1.83}) +[2026-06-24 05:24:34,984][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.002392694354057312, 'train/grad_norm_pre_clip': 18.42921257019043, 'timing/data': 0.0004136810312047601, 'timing/model': 1.1234488079790026, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 1.83}) +[2026-06-24 05:24:39,066][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.03114171326160431, 'train/grad_norm_pre_clip': 18.683862686157227, 'timing/data': 0.0003613739972934127, 'timing/model': 1.1162040659692138, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 1.83}) +[2026-06-24 05:24:43,155][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.007570263929665089, 'train/grad_norm_pre_clip': 6.920366287231445, 'timing/data': 0.0004278620472177863, 'timing/model': 1.1158340389374644, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 1.83}) +[2026-06-24 05:24:47,245][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.08145666122436523, 'train/grad_norm_pre_clip': 24.481861114501953, 'timing/data': 0.0003766899462789297, 'timing/model': 1.1215103999711573, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 1.83}) +[2026-06-24 05:24:51,435][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.0021464081946760416, 'train/grad_norm_pre_clip': 34.93962478637695, 'timing/data': 0.0003645949764177203, 'timing/model': 1.1161090380046517, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 1.83}) +[2026-06-24 05:24:55,524][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.003781234845519066, 'train/grad_norm_pre_clip': 2.3335275650024414, 'timing/data': 0.0004947959678247571, 'timing/model': 1.118812846019864, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 1.83}) +[2026-06-24 05:24:59,606][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.07551486790180206, 'train/grad_norm_pre_clip': 56.8117790222168, 'timing/data': 0.00036370602902024984, 'timing/model': 1.1175328870303929, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 1.83}) +[2026-06-24 05:25:03,698][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.013318482786417007, 'train/grad_norm_pre_clip': 42.08414077758789, 'timing/data': 0.00037914898712188005, 'timing/model': 1.1282196409301832, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 1.83}) +[2026-06-24 05:25:07,949][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.00019961230282206088, 'train/grad_norm_pre_clip': 14.619441032409668, 'timing/data': 0.00040515000000596046, 'timing/model': 1.1322994309011847, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 1.83}) +[2026-06-24 05:25:12,043][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 5.270669862511568e-05, 'train/grad_norm_pre_clip': 14.326316833496094, 'timing/data': 0.0005621039308607578, 'timing/model': 1.1307855499908328, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 1.83}) +[2026-06-24 05:25:16,130][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.0014497478259727359, 'train/grad_norm_pre_clip': 14.43348217010498, 'timing/data': 0.00042057002428919077, 'timing/model': 1.1193165270378813, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 1.83}) +[2026-06-24 05:25:20,212][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.052819374948740005, 'train/grad_norm_pre_clip': 19.619298934936523, 'timing/data': 0.0005404669791460037, 'timing/model': 1.1146851590601727, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 1.83}) +[2026-06-24 05:25:24,381][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.0029099585954099894, 'train/grad_norm_pre_clip': 9.273359298706055, 'timing/data': 0.0003606770187616348, 'timing/model': 1.1193303520558402, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 1.84}) +[2026-06-24 05:25:28,475][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.007355997804552317, 'train/grad_norm_pre_clip': 19.37898826599121, 'timing/data': 0.00036985694896429777, 'timing/model': 1.1216635670280084, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 1.84}) +[2026-06-24 05:25:32,561][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.0032841290812939405, 'train/grad_norm_pre_clip': 6.02230167388916, 'timing/data': 0.00036624795757234097, 'timing/model': 1.1216036420082673, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 1.84}) +[2026-06-24 05:25:36,648][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.06749243289232254, 'train/grad_norm_pre_clip': 23.759469985961914, 'timing/data': 0.0003616480389609933, 'timing/model': 1.1174572840100154, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 1.84}) +[2026-06-24 05:25:40,805][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.08390135318040848, 'train/grad_norm_pre_clip': 17.522615432739258, 'timing/data': 0.0003724439302459359, 'timing/model': 1.1208578520454466, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 1.84}) +[2026-06-24 05:25:44,900][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.013733495026826859, 'train/grad_norm_pre_clip': 18.147563934326172, 'timing/data': 0.0005566130857914686, 'timing/model': 1.1186376980040222, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 1.84}) +[2026-06-24 05:25:48,996][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 7.000101322773844e-05, 'train/grad_norm_pre_clip': 20.660890579223633, 'timing/data': 0.0004190190229564905, 'timing/model': 1.1216849579941481, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 1.84}) +[2026-06-24 05:25:53,100][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.020496971905231476, 'train/grad_norm_pre_clip': 23.90478515625, 'timing/data': 0.0004191549960523844, 'timing/model': 1.1229552420554683, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 1.84}) +[2026-06-24 05:37:54,376][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4750 +[2026-06-24 05:37:54,376][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.015692347660660744, 'train/grad_norm_pre_clip': 16.986682891845703, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00041306507773697376, 'timing/model': 1.1185080349678174, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 1.84}) +[2026-06-24 05:37:58,380][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.12153829634189606, 'train/grad_norm_pre_clip': 49.930084228515625, 'timing/data': 0.0003521160688251257, 'timing/model': 1.089549619005993, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 1.84}) +[2026-06-24 05:38:02,403][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.002866197843104601, 'train/grad_norm_pre_clip': 30.562023162841797, 'timing/data': 0.00036093289963901043, 'timing/model': 1.0983095379779115, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 1.84}) +[2026-06-24 05:38:06,426][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.012100503779947758, 'train/grad_norm_pre_clip': 6.1333513259887695, 'timing/data': 0.00035262899473309517, 'timing/model': 1.1137693269411102, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 1.84}) +[2026-06-24 05:38:10,541][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.017071982845664024, 'train/grad_norm_pre_clip': 14.648017883300781, 'timing/data': 0.0003623530501499772, 'timing/model': 1.104202002985403, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 1.84}) +[2026-06-24 05:38:14,564][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.003936955705285072, 'train/grad_norm_pre_clip': 14.467228889465332, 'timing/data': 0.0004368969239294529, 'timing/model': 1.0999414229299873, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 1.84}) +[2026-06-24 05:38:18,589][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.0010783423203974962, 'train/grad_norm_pre_clip': 31.419023513793945, 'timing/data': 0.0003479940351098776, 'timing/model': 1.106135439942591, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 1.84}) +[2026-06-24 05:38:22,612][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.0010242622811347246, 'train/grad_norm_pre_clip': 11.852361679077148, 'timing/data': 0.0004029839765280485, 'timing/model': 1.1060658229980618, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 1.84}) +[2026-06-24 05:38:26,823][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.028282394632697105, 'train/grad_norm_pre_clip': 14.588481903076172, 'timing/data': 0.0003524830099195242, 'timing/model': 1.1101601349655539, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 1.84}) +[2026-06-24 05:38:30,863][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.09822341799736023, 'train/grad_norm_pre_clip': 15.218849182128906, 'timing/data': 0.00035466498229652643, 'timing/model': 1.1078447609907016, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 1.84}) +[2026-06-24 05:38:34,908][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.04142263904213905, 'train/grad_norm_pre_clip': 17.076074600219727, 'timing/data': 0.0005767570110037923, 'timing/model': 1.1074889200972393, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 1.84}) +[2026-06-24 05:38:39,031][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.04245207458734512, 'train/grad_norm_pre_clip': 7.987057209014893, 'timing/data': 0.000502450973726809, 'timing/model': 1.1871249499963596, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 1.84}) +[2026-06-24 05:38:43,075][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.029611438512802124, 'train/grad_norm_pre_clip': 15.419122695922852, 'timing/data': 0.00035642797593027353, 'timing/model': 1.106938096927479, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 1.84}) +[2026-06-24 05:38:47,124][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.0083463741466403, 'train/grad_norm_pre_clip': 22.27657699584961, 'timing/data': 0.00035826792009174824, 'timing/model': 1.1065958250546828, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 1.84}) +[2026-06-24 05:38:51,196][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.08900826424360275, 'train/grad_norm_pre_clip': 14.861773490905762, 'timing/data': 0.0003771450137719512, 'timing/model': 1.1078249380225316, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 1.84}) +[2026-06-24 05:38:55,339][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.040342461317777634, 'train/grad_norm_pre_clip': 24.77259063720703, 'timing/data': 0.00040254800114780664, 'timing/model': 1.1964107130188495, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 1.84}) +[2026-06-24 05:38:59,407][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.0028839833103120327, 'train/grad_norm_pre_clip': 13.664329528808594, 'timing/data': 0.00035182107239961624, 'timing/model': 1.1117868630681187, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 1.84}) +[2026-06-24 05:39:03,465][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.0007858440512791276, 'train/grad_norm_pre_clip': 16.725000381469727, 'timing/data': 0.0005863710539415479, 'timing/model': 1.1087213499704376, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 1.84}) +[2026-06-24 05:39:07,531][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.022029802203178406, 'train/grad_norm_pre_clip': 12.322138786315918, 'timing/data': 0.00035996397491544485, 'timing/model': 1.1164970599347726, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 1.85}) +[2026-06-24 05:39:11,717][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 1.613285712664947e-05, 'train/grad_norm_pre_clip': 7.341118335723877, 'timing/data': 0.00035286799538880587, 'timing/model': 1.2299757290165871, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 1.85}) +[2026-06-24 05:39:15,773][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.024760577827692032, 'train/grad_norm_pre_clip': 8.955034255981445, 'timing/data': 0.00034804106689989567, 'timing/model': 1.1148603749461472, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 1.85}) +[2026-06-24 05:39:19,839][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.0015935655683279037, 'train/grad_norm_pre_clip': 19.587682723999023, 'timing/data': 0.0003505860222503543, 'timing/model': 1.1144817830063403, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 1.85}) +[2026-06-24 05:39:23,907][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.08719679713249207, 'train/grad_norm_pre_clip': 22.336833953857422, 'timing/data': 0.00035596697125583887, 'timing/model': 1.1130642680218443, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 1.85}) +[2026-06-24 05:39:28,079][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.001460264902561903, 'train/grad_norm_pre_clip': 25.431055068969727, 'timing/data': 0.0005609029904007912, 'timing/model': 1.1130156790604815, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 1.85}) +[2026-06-24 05:39:32,152][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.014937842264771461, 'train/grad_norm_pre_clip': 12.154814720153809, 'timing/data': 0.0004038159968331456, 'timing/model': 1.1154269059188664, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 1.85}) +[2026-06-24 05:39:36,222][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.011945088393986225, 'train/grad_norm_pre_clip': 17.235408782958984, 'timing/data': 0.00036833202466368675, 'timing/model': 1.11629321903456, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 1.85}) +[2026-06-24 05:39:40,297][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.03414192423224449, 'train/grad_norm_pre_clip': 28.580421447753906, 'timing/data': 0.00036085909232497215, 'timing/model': 1.1166760339401662, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 1.85}) +[2026-06-24 05:39:44,439][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.018793538212776184, 'train/grad_norm_pre_clip': 8.970344543457031, 'timing/data': 0.00035114004276692867, 'timing/model': 1.1141018150374293, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 1.85}) +[2026-06-24 05:39:48,514][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.20064301788806915, 'train/grad_norm_pre_clip': 44.3497428894043, 'timing/data': 0.00036583899054676294, 'timing/model': 1.120415079058148, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 1.85}) +[2026-06-24 05:39:52,588][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.01993461325764656, 'train/grad_norm_pre_clip': 16.46208381652832, 'timing/data': 0.000350888934917748, 'timing/model': 1.1181186069734395, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 1.85}) +[2026-06-24 05:39:56,659][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.026864729821681976, 'train/grad_norm_pre_clip': 22.002025604248047, 'timing/data': 0.0003553719725459814, 'timing/model': 1.1149397681001574, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 1.85}) +[2026-06-24 05:40:00,890][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.0647851973772049, 'train/grad_norm_pre_clip': 29.27069091796875, 'timing/data': 0.0004979459336027503, 'timing/model': 1.1124539580196142, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 1.85}) +[2026-06-24 05:40:04,958][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.015112034045159817, 'train/grad_norm_pre_clip': 18.93158721923828, 'timing/data': 0.0003625990357249975, 'timing/model': 1.1138873060699552, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 1.85}) +[2026-06-24 05:40:09,025][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.0009539257735013962, 'train/grad_norm_pre_clip': 4.192388534545898, 'timing/data': 0.00036562199238687754, 'timing/model': 1.1125903989886865, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 1.85}) +[2026-06-24 05:40:13,114][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.0022159735672175884, 'train/grad_norm_pre_clip': 25.836688995361328, 'timing/data': 0.00035623100120574236, 'timing/model': 1.1231397350784391, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 1.85}) +[2026-06-24 05:40:17,274][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 6.352366472128779e-05, 'train/grad_norm_pre_clip': 23.257352828979492, 'timing/data': 0.00036019296385347843, 'timing/model': 1.1157501799752936, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 1.85}) +[2026-06-24 05:40:21,347][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.017765892669558525, 'train/grad_norm_pre_clip': 27.70449447631836, 'timing/data': 0.00035868806298822165, 'timing/model': 1.1151430760510266, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 1.85}) +[2026-06-24 05:40:25,439][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.0095804613083601, 'train/grad_norm_pre_clip': 23.0991268157959, 'timing/data': 0.0003631599247455597, 'timing/model': 1.1149470799136907, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 1.85}) +[2026-06-24 05:40:29,525][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.0004459827032405883, 'train/grad_norm_pre_clip': 11.504209518432617, 'timing/data': 0.00036020600236952305, 'timing/model': 1.1166327190585434, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 1.85}) +[2026-06-24 05:40:33,703][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.09317426383495331, 'train/grad_norm_pre_clip': 24.25385093688965, 'timing/data': 0.0003596249734982848, 'timing/model': 1.221795469056815, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 1.85}) +[2026-06-24 05:40:37,776][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.07218987494707108, 'train/grad_norm_pre_clip': 31.06062126159668, 'timing/data': 0.00035226205363869667, 'timing/model': 1.1151431569596753, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 1.85}) +[2026-06-24 05:40:41,871][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.003881602780893445, 'train/grad_norm_pre_clip': 2.5640015602111816, 'timing/data': 0.0003562610363587737, 'timing/model': 1.120577138965018, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 1.85}) +[2026-06-24 05:40:45,948][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.00336033059284091, 'train/grad_norm_pre_clip': 17.334674835205078, 'timing/data': 0.00035373296122998, 'timing/model': 1.116043052985333, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 1.85}) +[2026-06-24 05:40:50,059][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.004254971630871296, 'train/grad_norm_pre_clip': 2.08150053024292, 'timing/data': 0.0005482989363372326, 'timing/model': 1.1345599739579484, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 1.85}) +[2026-06-24 05:40:54,304][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.03417952358722687, 'train/grad_norm_pre_clip': 26.276290893554688, 'timing/data': 0.0005556099349632859, 'timing/model': 1.1213583580683917, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 1.86}) +[2026-06-24 05:40:58,384][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.0007533767493441701, 'train/grad_norm_pre_clip': 21.720829010009766, 'timing/data': 0.00035628897603601217, 'timing/model': 1.11766777606681, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 1.86}) +[2026-06-24 05:41:02,487][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.0039953081868588924, 'train/grad_norm_pre_clip': 6.208797454833984, 'timing/data': 0.00043202994856983423, 'timing/model': 1.1182320290245116, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 1.86}) +[2026-06-24 05:41:06,565][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.011633580550551414, 'train/grad_norm_pre_clip': 7.130894184112549, 'timing/data': 0.00035646301694214344, 'timing/model': 1.1095819220645353, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 1.86}) +[2026-06-24 05:41:10,714][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.0639687329530716, 'train/grad_norm_pre_clip': 18.826858520507812, 'timing/data': 0.0003549599787220359, 'timing/model': 1.1162814520066604, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 1.86}) +[2026-06-24 05:41:14,804][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.0001403874484822154, 'train/grad_norm_pre_clip': 20.87868309020996, 'timing/data': 0.0003506989451125264, 'timing/model': 1.117674711975269, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 1.86}) +[2026-06-24 05:41:18,882][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4800 +[2026-06-24 05:42:23,201][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4800 +[2026-06-24 05:42:23,203][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4800 +[2026-06-24 05:43:02,054][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4800 +[2026-06-24 05:43:02,054][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.13675741851329803, 'train/grad_norm_pre_clip': 36.59797286987305, 'eval/action_loss/seconds': 64.3160212939838, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.02991185151040554, 'eval/latency_2/loss': 0.02991185151040554, 'eval/flappy/loss': 0.02991185151040554, 'eval/flappy/latency_2/loss': 0.02991185151040554, 'eval/flap_precision': 0.9781718963165075, 'eval/flap_recall': 0.9728629579375848, 'eval/flap_f1': 0.9755102040816326, 'eval/flap_rate_model': 0.2036111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9755102040816325, 'eval/latency_2/flap_precision': 0.9781718963165075, 'eval/latency_2/flap_recall': 0.9728629579375848, 'eval/latency_2/flap_f1': 0.9755102040816326, 'eval/latency_2/flap_rate_model': 0.2036111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9755102040816325, 'eval/action_classification/data_wait_seconds': 0.5301563739776611, 'eval/action_classification/predict_seconds': 38.17436981201172, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00035898294299840927, 'timing/model': 1.1170793080236763, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 1.86}) +[2026-06-24 05:43:06,260][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.027099180966615677, 'train/grad_norm_pre_clip': 29.860248565673828, 'timing/data': 0.00036071205977350473, 'timing/model': 1.106827863957733, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 1.86}) +[2026-06-24 05:43:10,296][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.15595434606075287, 'train/grad_norm_pre_clip': 23.74388313293457, 'timing/data': 0.00035465601831674576, 'timing/model': 1.1100920849712566, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 1.86}) +[2026-06-24 05:43:14,340][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.1395723670721054, 'train/grad_norm_pre_clip': 48.585662841796875, 'timing/data': 0.00041860598139464855, 'timing/model': 1.1034464759286493, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 1.86}) +[2026-06-24 05:43:18,391][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.04398123547434807, 'train/grad_norm_pre_clip': 27.275043487548828, 'timing/data': 0.00035354297142475843, 'timing/model': 1.1113980290247127, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 1.86}) +[2026-06-24 05:43:22,595][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.011766846291720867, 'train/grad_norm_pre_clip': 20.62480926513672, 'timing/data': 0.00040654896292835474, 'timing/model': 1.1117963900323957, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 1.86}) +[2026-06-24 05:43:26,653][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.004114129114896059, 'train/grad_norm_pre_clip': 16.241348266601562, 'timing/data': 0.00037971395067870617, 'timing/model': 1.1130871929926798, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 1.86}) +[2026-06-24 05:43:30,725][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.0007989797741174698, 'train/grad_norm_pre_clip': 11.462312698364258, 'timing/data': 0.00035721389576792717, 'timing/model': 1.1117051930632442, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 1.86}) +[2026-06-24 05:43:34,793][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.04597347229719162, 'train/grad_norm_pre_clip': 20.691226959228516, 'timing/data': 0.00035638699773699045, 'timing/model': 1.1165649980539456, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 1.86}) +[2026-06-24 05:43:39,023][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.005732922349125147, 'train/grad_norm_pre_clip': 28.10946273803711, 'timing/data': 0.00035692600067704916, 'timing/model': 1.110576112056151, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 1.86}) +[2026-06-24 05:43:43,091][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.0445319227874279, 'train/grad_norm_pre_clip': 34.29214859008789, 'timing/data': 0.0004083719104528427, 'timing/model': 1.118622476933524, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 1.86}) +[2026-06-24 05:43:47,176][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.010591407306492329, 'train/grad_norm_pre_clip': 8.421309471130371, 'timing/data': 0.00038712192326784134, 'timing/model': 1.11798387800809, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 1.86}) +[2026-06-24 05:43:51,247][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.02129829302430153, 'train/grad_norm_pre_clip': 11.558631896972656, 'timing/data': 0.0003656430635601282, 'timing/model': 1.1161059889709577, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 1.86}) +[2026-06-24 05:43:55,487][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.021100636571645737, 'train/grad_norm_pre_clip': 26.725584030151367, 'timing/data': 0.0005214989650994539, 'timing/model': 1.1193637599935755, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 1.86}) +[2026-06-24 05:43:59,541][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.129225954413414, 'train/grad_norm_pre_clip': 32.9281120300293, 'timing/data': 0.00036233803257346153, 'timing/model': 1.1105763399973512, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 1.86}) +[2026-06-24 05:44:03,612][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.0023007073905318975, 'train/grad_norm_pre_clip': 2.7592127323150635, 'timing/data': 0.00035861192736774683, 'timing/model': 1.1137533689616248, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 1.86}) +[2026-06-24 05:44:07,689][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.029170699417591095, 'train/grad_norm_pre_clip': 30.404735565185547, 'timing/data': 0.0005159380380064249, 'timing/model': 1.1148778959177434, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 1.86}) +[2026-06-24 05:44:11,923][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 2.7908256015507504e-05, 'train/grad_norm_pre_clip': 30.1907901763916, 'timing/data': 0.00037316710222512484, 'timing/model': 1.1133829179452732, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 1.86}) +[2026-06-24 05:44:16,003][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.0005743394722230732, 'train/grad_norm_pre_clip': 15.959417343139648, 'timing/data': 0.000402610981836915, 'timing/model': 1.1202244830783457, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 1.86}) +[2026-06-24 05:44:20,079][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.026030095294117928, 'train/grad_norm_pre_clip': 16.35767364501953, 'timing/data': 0.0003603389486670494, 'timing/model': 1.1156122819520533, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 1.87}) +[2026-06-24 05:44:24,170][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.04016142338514328, 'train/grad_norm_pre_clip': 25.25495147705078, 'timing/data': 0.0004031750140711665, 'timing/model': 1.119258987950161, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 1.87}) +[2026-06-24 05:44:28,426][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.1359567940235138, 'train/grad_norm_pre_clip': 45.91856384277344, 'timing/data': 0.00043159699998795986, 'timing/model': 1.1207927470095456, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 1.87}) +[2026-06-24 05:44:32,509][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.05783013999462128, 'train/grad_norm_pre_clip': 30.736072540283203, 'timing/data': 0.0003632740117609501, 'timing/model': 1.1208050389541313, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 1.87}) +[2026-06-24 05:44:36,601][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.15191909670829773, 'train/grad_norm_pre_clip': 27.753692626953125, 'timing/data': 0.0003599519841372967, 'timing/model': 1.1275466929655522, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 1.87}) +[2026-06-24 05:44:40,680][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.003938550129532814, 'train/grad_norm_pre_clip': 19.56538200378418, 'timing/data': 0.0003798420075327158, 'timing/model': 1.1233572700293735, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 1.87}) +[2026-06-24 05:44:44,923][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.019634133204817772, 'train/grad_norm_pre_clip': 33.06846237182617, 'timing/data': 0.00035621202550828457, 'timing/model': 1.116568465018645, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 1.87}) +[2026-06-24 05:44:49,008][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.03794801980257034, 'train/grad_norm_pre_clip': 17.80096435546875, 'timing/data': 0.0003533120034262538, 'timing/model': 1.116889922064729, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 1.87}) +[2026-06-24 05:44:53,105][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.0198366716504097, 'train/grad_norm_pre_clip': 17.72233772277832, 'timing/data': 0.00035286799538880587, 'timing/model': 1.133277181070298, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 1.87}) +[2026-06-24 05:44:57,187][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.0018501603044569492, 'train/grad_norm_pre_clip': 11.843732833862305, 'timing/data': 0.0003618529299274087, 'timing/model': 1.116804405930452, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 1.87}) +[2026-06-24 05:45:01,422][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.0017055660719051957, 'train/grad_norm_pre_clip': 10.397313117980957, 'timing/data': 0.00037973292637616396, 'timing/model': 1.1194857540540397, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 1.87}) +[2026-06-24 05:45:05,502][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.00011843500396935269, 'train/grad_norm_pre_clip': 41.01326370239258, 'timing/data': 0.0003492670366540551, 'timing/model': 1.1149460340384394, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 1.87}) +[2026-06-24 05:45:09,585][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.00012605694064404815, 'train/grad_norm_pre_clip': 13.999983787536621, 'timing/data': 0.0003601720090955496, 'timing/model': 1.1162022789940238, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 1.87}) +[2026-06-24 05:45:13,673][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.1027606949210167, 'train/grad_norm_pre_clip': 25.70985221862793, 'timing/data': 0.00041535997297614813, 'timing/model': 1.1207734390627593, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 1.87}) +[2026-06-24 05:45:17,916][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.054779957979917526, 'train/grad_norm_pre_clip': 22.50309181213379, 'timing/data': 0.0003557100426405668, 'timing/model': 1.1193759490270168, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 1.87}) +[2026-06-24 05:45:21,997][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.02260199934244156, 'train/grad_norm_pre_clip': 33.86421585083008, 'timing/data': 0.0003678749781101942, 'timing/model': 1.116904585971497, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 1.87}) +[2026-06-24 05:45:26,083][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.0068626501597464085, 'train/grad_norm_pre_clip': 3.0935332775115967, 'timing/data': 0.00034708098974078894, 'timing/model': 1.120526658021845, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 1.87}) +[2026-06-24 05:45:30,172][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.02267102338373661, 'train/grad_norm_pre_clip': 32.17760467529297, 'timing/data': 0.0003544189967215061, 'timing/model': 1.1166134619852528, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 1.87}) +[2026-06-24 05:45:34,414][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.008318513631820679, 'train/grad_norm_pre_clip': 39.92420196533203, 'timing/data': 0.0003666030243039131, 'timing/model': 1.1152642609085888, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 1.87}) +[2026-06-24 05:45:38,499][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.08407581597566605, 'train/grad_norm_pre_clip': 23.901039123535156, 'timing/data': 0.0003987629897892475, 'timing/model': 1.1221680040471256, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 1.87}) +[2026-06-24 05:45:42,593][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.042985133826732635, 'train/grad_norm_pre_clip': 16.379491806030273, 'timing/data': 0.0004217179957777262, 'timing/model': 1.1268634939333424, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 1.87}) +[2026-06-24 05:45:46,697][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.07633700221776962, 'train/grad_norm_pre_clip': 20.123241424560547, 'timing/data': 0.00036008504685014486, 'timing/model': 1.1209975900128484, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 1.87}) +[2026-06-24 05:45:50,949][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 3.92559559259098e-05, 'train/grad_norm_pre_clip': 37.32425308227539, 'timing/data': 0.0004452930297702551, 'timing/model': 1.115550561924465, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 1.87}) +[2026-06-24 05:45:55,027][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.03999512642621994, 'train/grad_norm_pre_clip': 18.806201934814453, 'timing/data': 0.0004659349797293544, 'timing/model': 1.1165552220772952, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 1.87}) +[2026-06-24 05:45:59,105][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.010537104681134224, 'train/grad_norm_pre_clip': 18.540613174438477, 'timing/data': 0.00035917595960199833, 'timing/model': 1.1193944440456107, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 1.87}) +[2026-06-24 05:46:03,181][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.0015402494464069605, 'train/grad_norm_pre_clip': 26.589370727539062, 'timing/data': 0.00035811495035886765, 'timing/model': 1.1144996569491923, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 1.87}) +[2026-06-24 05:46:07,448][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.047202251851558685, 'train/grad_norm_pre_clip': 24.36125946044922, 'timing/data': 0.0005347019759938121, 'timing/model': 1.121952346060425, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 1.88}) +[2026-06-24 05:46:11,542][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.03732696548104286, 'train/grad_norm_pre_clip': 17.956457138061523, 'timing/data': 0.0003998529864475131, 'timing/model': 1.1231077819829807, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 1.88}) +[2026-06-24 05:46:15,627][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.14981597661972046, 'train/grad_norm_pre_clip': 41.54750061035156, 'timing/data': 0.000348587054759264, 'timing/model': 1.125617589103058, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 1.88}) +[2026-06-24 05:46:19,722][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.0026195726823061705, 'train/grad_norm_pre_clip': 23.59650421142578, 'timing/data': 0.0005508189788088202, 'timing/model': 1.128335700952448, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 1.88}) +[2026-06-24 05:46:24,010][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.009489675983786583, 'train/grad_norm_pre_clip': 28.645687103271484, 'timing/data': 0.00035095796920359135, 'timing/model': 1.1397497720317915, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 1.88}) +[2026-06-24 05:46:28,106][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.014077425934374332, 'train/grad_norm_pre_clip': 26.25751495361328, 'timing/data': 0.00036023801658302546, 'timing/model': 1.1192340870620683, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 1.88}) +[2026-06-24 05:46:32,225][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.0020544487051665783, 'train/grad_norm_pre_clip': 44.405540466308594, 'timing/data': 0.00039205304346978664, 'timing/model': 1.114191955071874, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 1.88}) +[2026-06-24 05:46:36,306][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.01764787547290325, 'train/grad_norm_pre_clip': 20.306507110595703, 'timing/data': 0.00035055691841989756, 'timing/model': 1.1197374239563942, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 1.88}) +[2026-06-24 05:46:40,572][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.022844018414616585, 'train/grad_norm_pre_clip': 22.783109664916992, 'timing/data': 0.00035390607081353664, 'timing/model': 1.1155144489603117, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 1.88}) +[2026-06-24 05:46:44,663][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.003949106205254793, 'train/grad_norm_pre_clip': 18.924955368041992, 'timing/data': 0.0004402149934321642, 'timing/model': 1.1223958799382672, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 1.88}) +[2026-06-24 05:46:48,756][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.08740705996751785, 'train/grad_norm_pre_clip': 39.40330123901367, 'timing/data': 0.0003553549759089947, 'timing/model': 1.121154158958234, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 1.88}) +[2026-06-24 05:46:52,845][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.0001576196518726647, 'train/grad_norm_pre_clip': 4.254107475280762, 'timing/data': 0.0003501559840515256, 'timing/model': 1.125004603061825, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 1.88}) +[2026-06-24 05:46:57,093][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.0026287706568837166, 'train/grad_norm_pre_clip': 14.271331787109375, 'timing/data': 0.0003661529626697302, 'timing/model': 1.119600867968984, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 1.88}) +[2026-06-24 05:47:01,191][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.036932095885276794, 'train/grad_norm_pre_clip': 32.137237548828125, 'timing/data': 0.0003532459959387779, 'timing/model': 1.1229944260558113, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 1.88}) +[2026-06-24 05:47:05,274][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.0595555454492569, 'train/grad_norm_pre_clip': 18.98600196838379, 'timing/data': 0.0003912990214303136, 'timing/model': 1.118157202959992, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 1.88}) +[2026-06-24 05:47:09,366][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.000669217377435416, 'train/grad_norm_pre_clip': 18.041555404663086, 'timing/data': 0.0003559169126674533, 'timing/model': 1.1189909890526906, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 1.88}) +[2026-06-24 05:47:13,625][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.03783903643488884, 'train/grad_norm_pre_clip': 38.79640579223633, 'timing/data': 0.0003507299115881324, 'timing/model': 1.1170258380006999, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 1.88}) +[2026-06-24 05:47:17,717][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.08538247644901276, 'train/grad_norm_pre_clip': 26.480649948120117, 'timing/data': 0.0003514629788696766, 'timing/model': 1.1215511889895424, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 1.88}) +[2026-06-24 05:47:21,806][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.004867157898843288, 'train/grad_norm_pre_clip': 30.87672996520996, 'timing/data': 0.0003848759224638343, 'timing/model': 1.1206577030243352, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 1.88}) +[2026-06-24 05:47:25,903][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.002951511414721608, 'train/grad_norm_pre_clip': 29.27756690979004, 'timing/data': 0.0003640400245785713, 'timing/model': 1.1233094059862196, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 1.88}) +[2026-06-24 05:47:30,172][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.002622236032038927, 'train/grad_norm_pre_clip': 18.441781997680664, 'timing/data': 0.0005618049763143063, 'timing/model': 1.1282831450225785, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 1.88}) +[2026-06-24 05:47:34,259][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.015950266271829605, 'train/grad_norm_pre_clip': 15.435419082641602, 'timing/data': 0.00035403191577643156, 'timing/model': 1.1197025430155918, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 1.88}) +[2026-06-24 05:47:38,346][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.0014721691841259599, 'train/grad_norm_pre_clip': 8.36197280883789, 'timing/data': 0.0003602970391511917, 'timing/model': 1.1194616249995306, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 1.88}) +[2026-06-24 05:47:42,440][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.2555905878543854, 'train/grad_norm_pre_clip': 26.73618507385254, 'timing/data': 0.0004111030139029026, 'timing/model': 1.120228820014745, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 1.88}) +[2026-06-24 05:47:46,686][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.021865474060177803, 'train/grad_norm_pre_clip': 26.21762466430664, 'timing/data': 0.00041291199158877134, 'timing/model': 1.1141748179215938, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 1.88}) +[2026-06-24 05:47:50,768][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.009080392308533192, 'train/grad_norm_pre_clip': 4.460000514984131, 'timing/data': 0.00036368495784699917, 'timing/model': 1.1207288010045886, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 1.88}) +[2026-06-24 05:47:54,855][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.01980421133339405, 'train/grad_norm_pre_clip': 11.312081336975098, 'timing/data': 0.00036007200833410025, 'timing/model': 1.120000898023136, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 1.89}) +[2026-06-24 05:47:58,944][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.00045442962436936796, 'train/grad_norm_pre_clip': 18.655216217041016, 'timing/data': 0.0005320700583979487, 'timing/model': 1.1159283119486645, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 1.89}) +[2026-06-24 05:48:03,200][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.0034091579727828503, 'train/grad_norm_pre_clip': 11.215447425842285, 'timing/data': 0.00035788107197731733, 'timing/model': 1.1188552329549566, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 1.89}) +[2026-06-24 05:48:07,287][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.001684769755229354, 'train/grad_norm_pre_clip': 5.234816074371338, 'timing/data': 0.0003519130405038595, 'timing/model': 1.1186257400549948, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 1.89}) +[2026-06-24 05:48:11,385][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.05860371142625809, 'train/grad_norm_pre_clip': 25.688913345336914, 'timing/data': 0.00036114302929490805, 'timing/model': 1.1268916509579867, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 1.89}) +[2026-06-24 05:48:15,480][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.006545697338879108, 'train/grad_norm_pre_clip': 23.751131057739258, 'timing/data': 0.00035493692848831415, 'timing/model': 1.1179602020420134, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 1.89}) +[2026-06-24 05:48:19,723][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.02772051841020584, 'train/grad_norm_pre_clip': 13.149481773376465, 'timing/data': 0.0003510549431666732, 'timing/model': 1.1225056090625003, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 1.89}) +[2026-06-24 05:48:23,807][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.015862448140978813, 'train/grad_norm_pre_clip': 13.294599533081055, 'timing/data': 0.0003554470604285598, 'timing/model': 1.1175084409769624, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 1.89}) +[2026-06-24 05:48:27,876][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.04197698086500168, 'train/grad_norm_pre_clip': 12.201007843017578, 'timing/data': 0.00035758200101554394, 'timing/model': 1.116113045020029, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 1.89}) +[2026-06-24 05:48:31,964][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.06812900304794312, 'train/grad_norm_pre_clip': 25.540857315063477, 'timing/data': 0.00036076095420867205, 'timing/model': 1.1192678819643334, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 1.89}) +[2026-06-24 05:48:36,221][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.02073330618441105, 'train/grad_norm_pre_clip': 12.141687393188477, 'timing/data': 0.0003480369923636317, 'timing/model': 1.1210270000156015, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 1.89}) +[2026-06-24 05:48:40,321][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.011901592835783958, 'train/grad_norm_pre_clip': 5.09000301361084, 'timing/data': 0.0003518130397424102, 'timing/model': 1.1292373349424452, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 1.89}) +[2026-06-24 05:48:44,405][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.002489826176315546, 'train/grad_norm_pre_clip': 20.63030242919922, 'timing/data': 0.0003460219595581293, 'timing/model': 1.1173751859460026, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 1.89}) +[2026-06-24 05:48:48,505][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.04387400299310684, 'train/grad_norm_pre_clip': 17.078996658325195, 'timing/data': 0.00035174202639609575, 'timing/model': 1.1261536598904058, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 1.89}) +[2026-06-24 05:48:52,772][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.03880329057574272, 'train/grad_norm_pre_clip': 17.143997192382812, 'timing/data': 0.0003557039890438318, 'timing/model': 1.118903356953524, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 1.89}) +[2026-06-24 05:48:56,857][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.027561664581298828, 'train/grad_norm_pre_clip': 22.865421295166016, 'timing/data': 0.0003570190165191889, 'timing/model': 1.1211449289694428, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 1.89}) +[2026-06-24 05:49:00,932][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.16378086805343628, 'train/grad_norm_pre_clip': 17.11615753173828, 'timing/data': 0.0003569849068298936, 'timing/model': 1.1145580799784511, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 1.89}) +[2026-06-24 05:49:05,018][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.09966479986906052, 'train/grad_norm_pre_clip': 21.50958824157715, 'timing/data': 0.00035949493758380413, 'timing/model': 1.1181675429688767, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 1.89}) +[2026-06-24 05:49:09,257][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.006117966026067734, 'train/grad_norm_pre_clip': 22.997926712036133, 'timing/data': 0.00036363897379487753, 'timing/model': 1.1197376329218969, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 1.89}) +[2026-06-24 05:49:13,343][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.04284321889281273, 'train/grad_norm_pre_clip': 41.43364334106445, 'timing/data': 0.000356174074113369, 'timing/model': 1.11934792494867, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 1.89}) +[2026-06-24 05:49:17,429][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.04991244152188301, 'train/grad_norm_pre_clip': 13.624652862548828, 'timing/data': 0.0005910949548706412, 'timing/model': 1.119625766063109, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 1.89}) +[2026-06-24 05:49:21,511][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.0040297433733940125, 'train/grad_norm_pre_clip': 29.46101188659668, 'timing/data': 0.0003598150797188282, 'timing/model': 1.1181809339905158, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 1.89}) +[2026-06-24 05:49:25,756][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.00037249154411256313, 'train/grad_norm_pre_clip': 22.112258911132812, 'timing/data': 0.000402878038585186, 'timing/model': 1.1182545910123736, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 1.89}) +[2026-06-24 05:49:29,831][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.0762760192155838, 'train/grad_norm_pre_clip': 33.44036865234375, 'timing/data': 0.00044431304559111595, 'timing/model': 1.1152491389075294, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 1.89}) +[2026-06-24 05:49:33,912][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.028317397460341454, 'train/grad_norm_pre_clip': 12.663393020629883, 'timing/data': 0.00035148998722434044, 'timing/model': 1.119373879977502, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 1.89}) +[2026-06-24 05:49:38,011][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.11349478363990784, 'train/grad_norm_pre_clip': 34.32272720336914, 'timing/data': 0.00035517197102308273, 'timing/model': 1.1203943699365482, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 1.89}) +[2026-06-24 05:49:42,272][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.024534067139029503, 'train/grad_norm_pre_clip': 27.751806259155273, 'timing/data': 0.0005038299132138491, 'timing/model': 1.1259434749372303, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 1.9}) +[2026-06-24 05:49:46,357][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.019391780719161034, 'train/grad_norm_pre_clip': 8.616425514221191, 'timing/data': 0.0003966810181736946, 'timing/model': 1.1147968170698732, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 1.9}) +[2026-06-24 05:49:50,452][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.01915440522134304, 'train/grad_norm_pre_clip': 18.1138916015625, 'timing/data': 0.0005470319883897901, 'timing/model': 1.1211136190686375, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 1.9}) +[2026-06-24 05:49:54,544][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4900 +[2026-06-24 05:50:58,652][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4900 +[2026-06-24 05:50:58,654][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4900 +[2026-06-24 05:51:37,877][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4900 +[2026-06-24 05:51:37,877][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.05920865014195442, 'train/grad_norm_pre_clip': 13.941646575927734, 'eval/action_loss/seconds': 64.10560095892288, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.02909514680504799, 'eval/latency_2/loss': 0.02909514680504799, 'eval/flappy/loss': 0.02909514680504799, 'eval/flappy/latency_2/loss': 0.02909514680504799, 'eval/flap_precision': 0.9743589743589743, 'eval/flap_recall': 0.9796472184531886, 'eval/flap_f1': 0.9769959404600813, 'eval/flap_rate_model': 0.20583333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9769959404600813, 'eval/latency_2/flap_precision': 0.9743589743589743, 'eval/latency_2/flap_recall': 0.9796472184531886, 'eval/latency_2/flap_f1': 0.9769959404600813, 'eval/latency_2/flap_rate_model': 0.20583333333333334, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9769959404600813, 'eval/action_classification/data_wait_seconds': 0.4908117651939392, 'eval/action_classification/predict_seconds': 38.56869125366211, 'eval/action_classification/frames': 3600.0, 'timing/data': 0.00035197800025343895, 'timing/model': 1.12237670191098, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 1.9}) +[2026-06-24 05:51:41,928][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.05463603883981705, 'train/grad_norm_pre_clip': 26.70692253112793, 'timing/data': 0.0003558590542525053, 'timing/model': 1.1109477439895272, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 1.9}) +[2026-06-24 05:51:45,984][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.025592926889657974, 'train/grad_norm_pre_clip': 22.582839965820312, 'timing/data': 0.00044471106957644224, 'timing/model': 1.1053088830085471, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 1.9}) +[2026-06-24 05:51:50,037][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.007286816835403442, 'train/grad_norm_pre_clip': 12.76254940032959, 'timing/data': 0.00035852903965860605, 'timing/model': 1.113152028992772, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 1.9}) +[2026-06-24 05:51:54,181][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.00045276209129951894, 'train/grad_norm_pre_clip': 11.350006103515625, 'timing/data': 0.0005551179638132453, 'timing/model': 1.1330387849593535, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 1.9}) +[2026-06-24 05:51:58,260][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.010871587321162224, 'train/grad_norm_pre_clip': 21.680606842041016, 'timing/data': 0.0003628639969974756, 'timing/model': 1.1301353849703446, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 1.9}) +[2026-06-24 05:52:02,330][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.06179288402199745, 'train/grad_norm_pre_clip': 17.350244522094727, 'timing/data': 0.00038495007902383804, 'timing/model': 1.1110274830134585, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 1.9}) +[2026-06-24 05:52:06,394][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.0074889203533530235, 'train/grad_norm_pre_clip': 18.77280044555664, 'timing/data': 0.00036778603680431843, 'timing/model': 1.118749251938425, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 1.9}) +[2026-06-24 05:52:10,597][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.012648635543882847, 'train/grad_norm_pre_clip': 39.58411407470703, 'timing/data': 0.00035723799373954535, 'timing/model': 1.1178564729634672, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 1.9}) +[2026-06-24 05:52:14,672][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.0060020810924470425, 'train/grad_norm_pre_clip': 5.862514019012451, 'timing/data': 0.0005086879245936871, 'timing/model': 1.1188898279797286, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 1.9}) +[2026-06-24 05:52:18,757][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.18228310346603394, 'train/grad_norm_pre_clip': 32.625431060791016, 'timing/data': 0.0005884760757908225, 'timing/model': 1.1156878750771284, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 1.9}) +[2026-06-24 05:52:22,822][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.008415622636675835, 'train/grad_norm_pre_clip': 7.947200298309326, 'timing/data': 0.0006093299016356468, 'timing/model': 1.1109999190084636, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 1.9}) +[2026-06-24 05:52:26,986][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.02229086123406887, 'train/grad_norm_pre_clip': 21.366954803466797, 'timing/data': 0.0008269709069281816, 'timing/model': 1.123602437088266, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 1.9}) +[2026-06-24 05:52:31,071][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.016862032935023308, 'train/grad_norm_pre_clip': 12.244478225708008, 'timing/data': 0.0003703220281749964, 'timing/model': 1.1168645189609379, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 1.9}) +[2026-06-24 05:52:35,150][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.03182274475693703, 'train/grad_norm_pre_clip': 14.17347240447998, 'timing/data': 0.00044027098920196295, 'timing/model': 1.1191178479930386, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 1.9}) +[2026-06-24 05:52:39,220][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.07133623957633972, 'train/grad_norm_pre_clip': 21.35981559753418, 'timing/data': 0.0006593569414690137, 'timing/model': 1.1125512489816174, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 1.9}) +[2026-06-24 05:52:43,472][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.14857813715934753, 'train/grad_norm_pre_clip': 20.095226287841797, 'timing/data': 0.0003613419830799103, 'timing/model': 1.1143201239174232, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 1.9}) +[2026-06-24 05:52:47,571][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.010455936193466187, 'train/grad_norm_pre_clip': 7.783551216125488, 'timing/data': 0.0003700329689309001, 'timing/model': 1.1359990750206634, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 1.9}) +[2026-06-24 05:52:51,675][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.05790762975811958, 'train/grad_norm_pre_clip': 19.171005249023438, 'timing/data': 0.00037314905785024166, 'timing/model': 1.1328498020302504, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 1.9}) +[2026-06-24 05:52:55,826][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.0009107232908718288, 'train/grad_norm_pre_clip': 25.091344833374023, 'timing/data': 0.0005351710133254528, 'timing/model': 1.1866968550020829, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 1.9}) +[2026-06-24 05:52:59,911][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.00032476766500622034, 'train/grad_norm_pre_clip': 9.49267864227295, 'timing/data': 0.000424689962528646, 'timing/model': 1.1171798900468275, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 1.9}) +[2026-06-24 05:53:03,984][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.0048040421679615974, 'train/grad_norm_pre_clip': 16.02898406982422, 'timing/data': 0.0003583329962566495, 'timing/model': 1.1178174139931798, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 1.9}) +[2026-06-24 05:53:08,061][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.014026499353349209, 'train/grad_norm_pre_clip': 24.761198043823242, 'timing/data': 0.00036629498936235905, 'timing/model': 1.1169315930455923, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 1.9}) +[2026-06-24 05:53:12,234][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.011677511967718601, 'train/grad_norm_pre_clip': 20.1680908203125, 'timing/data': 0.0005713070277124643, 'timing/model': 1.2089195899898186, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 1.91}) +[2026-06-24 05:53:16,319][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.1539890468120575, 'train/grad_norm_pre_clip': 24.099306106567383, 'timing/data': 0.00037153507582843304, 'timing/model': 1.1167538879672065, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 1.91}) +[2026-06-24 05:53:20,415][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.15489411354064941, 'train/grad_norm_pre_clip': 43.575626373291016, 'timing/data': 0.0005576899275183678, 'timing/model': 1.1193933159811422, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 1.91}) +[2026-06-24 05:53:24,507][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.013446681201457977, 'train/grad_norm_pre_clip': 19.20219612121582, 'timing/data': 0.0005856940988451242, 'timing/model': 1.1221194810932502, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 1.91}) +[2026-06-24 05:53:28,700][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.00012736243661493063, 'train/grad_norm_pre_clip': 24.76288414001465, 'timing/data': 0.000495581072755158, 'timing/model': 1.2338309170445427, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 1.91}) +[2026-06-24 05:53:32,799][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.02659786492586136, 'train/grad_norm_pre_clip': 22.42241859436035, 'timing/data': 0.0005692261038348079, 'timing/model': 1.118148458073847, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 1.91}) +[2026-06-24 05:53:36,907][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.04662875086069107, 'train/grad_norm_pre_clip': 32.32857894897461, 'timing/data': 0.00038286601193249226, 'timing/model': 1.121241262066178, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 1.91}) +[2026-06-24 05:53:40,987][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.0011065445141866803, 'train/grad_norm_pre_clip': 19.717853546142578, 'timing/data': 0.0003842390142381191, 'timing/model': 1.1147497189231217, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 1.91}) +[2026-06-24 05:53:45,154][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.00042000258690677583, 'train/grad_norm_pre_clip': 36.767547607421875, 'timing/data': 0.0004772969987243414, 'timing/model': 1.1173072049859911, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 1.91}) +[2026-06-24 05:53:49,234][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.0011361857177689672, 'train/grad_norm_pre_clip': 11.189529418945312, 'timing/data': 0.0005849699955433607, 'timing/model': 1.1130647509125993, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 1.91}) +[2026-06-24 05:53:53,322][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.09205742180347443, 'train/grad_norm_pre_clip': 32.192230224609375, 'timing/data': 0.0003873879322782159, 'timing/model': 1.120129197021015, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 1.91}) +[2026-06-24 05:53:57,406][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.10626648366451263, 'train/grad_norm_pre_clip': 24.265888214111328, 'timing/data': 0.00040300097316503525, 'timing/model': 1.1170625920640305, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 1.91}) +[2026-06-24 05:54:01,571][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.1290435492992401, 'train/grad_norm_pre_clip': 45.918704986572266, 'timing/data': 0.0003572020214051008, 'timing/model': 1.1187976200599223, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 1.91}) +[2026-06-24 05:54:05,658][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.0005486644804477692, 'train/grad_norm_pre_clip': 8.833427429199219, 'timing/data': 0.00035976991057395935, 'timing/model': 1.1197922630235553, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 1.91}) +[2026-06-24 05:54:09,735][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.003339172573760152, 'train/grad_norm_pre_clip': 28.999279022216797, 'timing/data': 0.00036432594060897827, 'timing/model': 1.113125570002012, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 1.91}) +[2026-06-24 05:54:13,836][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.0018421427812427282, 'train/grad_norm_pre_clip': 10.6901273727417, 'timing/data': 0.00036508298944681883, 'timing/model': 1.1242352089611813, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 1.91}) +[2026-06-24 05:54:18,094][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.035811927169561386, 'train/grad_norm_pre_clip': 19.59864616394043, 'timing/data': 0.0004350891103968024, 'timing/model': 1.1202122130198404, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 1.91}) +[2026-06-24 05:54:22,177][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.014241219498217106, 'train/grad_norm_pre_clip': 27.72710418701172, 'timing/data': 0.0004186209989711642, 'timing/model': 1.1154348889831454, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 1.91}) +[2026-06-24 05:54:26,254][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.0015599892940372229, 'train/grad_norm_pre_clip': 36.517398834228516, 'timing/data': 0.0003622639924287796, 'timing/model': 1.1168542680097744, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 1.91}) +[2026-06-24 05:54:30,343][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.003827177919447422, 'train/grad_norm_pre_clip': 18.43114471435547, 'timing/data': 0.00036722607910633087, 'timing/model': 1.124952486017719, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 1.91}) +[2026-06-24 05:54:34,514][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.0012046352494508028, 'train/grad_norm_pre_clip': 5.201541423797607, 'timing/data': 0.0003821438876911998, 'timing/model': 1.1165401239413768, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 1.91}) +[2026-06-24 05:54:38,598][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.006894988939166069, 'train/grad_norm_pre_clip': 18.495620727539062, 'timing/data': 0.00036996195558458567, 'timing/model': 1.1154680399922654, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 1.91}) +[2026-06-24 05:54:42,691][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.011218425817787647, 'train/grad_norm_pre_clip': 10.251157760620117, 'timing/data': 0.0003642020747065544, 'timing/model': 1.121735317981802, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 1.91}) +[2026-06-24 05:54:46,792][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.0036615408025681973, 'train/grad_norm_pre_clip': 5.440357685089111, 'timing/data': 0.00036933901719748974, 'timing/model': 1.1235614930046722, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 1.91}) +[2026-06-24 05:54:51,005][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.0006969909882172942, 'train/grad_norm_pre_clip': 30.783510208129883, 'timing/data': 0.0006420500576496124, 'timing/model': 1.242369705112651, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 1.91}) +[2026-06-24 05:54:55,224][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.009939833544194698, 'train/grad_norm_pre_clip': 22.31331443786621, 'timing/data': 0.00036518299020826817, 'timing/model': 1.256752229994163, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 1.91}) +[2026-06-24 05:54:59,313][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.05206858366727829, 'train/grad_norm_pre_clip': 20.60292625427246, 'timing/data': 0.0003597670001909137, 'timing/model': 1.124109691940248, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 1.92}) +[2026-06-24 05:55:03,410][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.05176263675093651, 'train/grad_norm_pre_clip': 20.490734100341797, 'timing/data': 0.0005678500747308135, 'timing/model': 1.1191242040367797, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 1.92}) +[2026-06-24 05:55:07,492][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.07282473146915436, 'train/grad_norm_pre_clip': 23.090543746948242, 'timing/data': 0.000566011993214488, 'timing/model': 1.1149076910223812, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 1.92}) +[2026-06-24 05:55:11,739][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.035246558487415314, 'train/grad_norm_pre_clip': 4.702872276306152, 'timing/data': 0.00046067300718277693, 'timing/model': 1.1187125950818881, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 1.92}) +[2026-06-24 05:55:15,827][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.010175132192671299, 'train/grad_norm_pre_clip': 29.7049503326416, 'timing/data': 0.00036301405634731054, 'timing/model': 1.1176033150404692, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 1.92}) +[2026-06-24 05:55:19,933][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.006538345478475094, 'train/grad_norm_pre_clip': 8.026252746582031, 'timing/data': 0.00040587596595287323, 'timing/model': 1.1149434869876131, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 1.92}) +[2026-06-24 05:55:24,020][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.03259255737066269, 'train/grad_norm_pre_clip': 27.759933471679688, 'timing/data': 0.0005717629101127386, 'timing/model': 1.1167272420134395, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 1.92}) +[2026-06-24 05:55:28,207][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.004242174327373505, 'train/grad_norm_pre_clip': 1.7732338905334473, 'timing/data': 0.0003628759877756238, 'timing/model': 1.1164305669954047, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 1.92}) +[2026-06-24 05:55:32,288][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 1.9239483663113788e-05, 'train/grad_norm_pre_clip': 14.950013160705566, 'timing/data': 0.00036848196759819984, 'timing/model': 1.1190944439731538, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 1.92}) +[2026-06-24 05:55:36,375][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.1103358045220375, 'train/grad_norm_pre_clip': 20.56961441040039, 'timing/data': 0.00043821707367897034, 'timing/model': 1.1183887860970572, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 1.92}) +[2026-06-24 05:55:40,463][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.012263639830052853, 'train/grad_norm_pre_clip': 28.98077964782715, 'timing/data': 0.0003890650114044547, 'timing/model': 1.1189340120181441, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 1.92}) +[2026-06-24 05:55:44,634][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.0335548035800457, 'train/grad_norm_pre_clip': 29.013107299804688, 'timing/data': 0.00036302395164966583, 'timing/model': 1.1212298799073324, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 1.92}) +[2026-06-24 05:55:48,721][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.0030993272084742785, 'train/grad_norm_pre_clip': 10.129969596862793, 'timing/data': 0.0005345200188457966, 'timing/model': 1.11731444101315, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 1.92}) +[2026-06-24 05:55:52,828][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.0003513596602715552, 'train/grad_norm_pre_clip': 22.023155212402344, 'timing/data': 0.00036103592719882727, 'timing/model': 1.1326358589576557, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 1.92}) +[2026-06-24 05:55:56,921][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.037201929837465286, 'train/grad_norm_pre_clip': 16.10170555114746, 'timing/data': 0.0004176009679213166, 'timing/model': 1.123779876041226, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 1.92}) +[2026-06-24 05:56:01,088][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.03607993200421333, 'train/grad_norm_pre_clip': 23.740148544311523, 'timing/data': 0.000372081995010376, 'timing/model': 1.1315351229859516, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 1.92}) +[2026-06-24 05:56:05,192][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.012037034146487713, 'train/grad_norm_pre_clip': 20.671171188354492, 'timing/data': 0.0005572239169850945, 'timing/model': 1.127972749993205, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 1.92}) +[2026-06-24 05:56:09,285][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.012390808202326298, 'train/grad_norm_pre_clip': 16.14959716796875, 'timing/data': 0.000360121950507164, 'timing/model': 1.1205399630125612, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 1.92}) +[2026-06-24 05:56:13,370][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.015273930504918098, 'train/grad_norm_pre_clip': 7.996413230895996, 'timing/data': 0.0003619389608502388, 'timing/model': 1.119856578996405, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 1.92}) +[2026-06-24 05:56:17,620][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.004421710968017578, 'train/grad_norm_pre_clip': 16.972492218017578, 'timing/data': 0.0003824209561571479, 'timing/model': 1.1224815879249945, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 1.92}) +[2026-06-24 05:56:21,720][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.05663900449872017, 'train/grad_norm_pre_clip': 33.36014175415039, 'timing/data': 0.00035934499464929104, 'timing/model': 1.1243946430040523, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 1.92}) +[2026-06-24 05:56:25,807][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.004224100150167942, 'train/grad_norm_pre_clip': 40.249141693115234, 'timing/data': 0.0004134660121053457, 'timing/model': 1.1211115920450538, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 1.92}) +[2026-06-24 05:56:29,984][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.024562932550907135, 'train/grad_norm_pre_clip': 8.215703010559082, 'timing/data': 0.0004354659467935562, 'timing/model': 1.2072770609520376, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 1.92}) +[2026-06-24 05:56:34,065][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.028717011213302612, 'train/grad_norm_pre_clip': 18.186548233032227, 'timing/data': 0.000364757957868278, 'timing/model': 1.1233159470139071, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 1.92}) +[2026-06-24 05:56:38,153][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.0004436907183844596, 'train/grad_norm_pre_clip': 23.703954696655273, 'timing/data': 0.0003624618984758854, 'timing/model': 1.118550557992421, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 1.92}) +[2026-06-24 05:56:42,244][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.01120759267359972, 'train/grad_norm_pre_clip': 18.055828094482422, 'timing/data': 0.0004374029813334346, 'timing/model': 1.1151627600193024, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 1.93}) +[2026-06-24 05:56:46,487][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.00417207321152091, 'train/grad_norm_pre_clip': 14.176519393920898, 'timing/data': 0.00037006800994277, 'timing/model': 1.2829372850246727, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 1.93}) +[2026-06-24 05:56:50,589][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.02189994789659977, 'train/grad_norm_pre_clip': 26.372886657714844, 'timing/data': 0.000368200009688735, 'timing/model': 1.1248959919903427, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 1.93}) +[2026-06-24 05:56:54,685][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.0007091129082255065, 'train/grad_norm_pre_clip': 8.867104530334473, 'timing/data': 0.0003635359462350607, 'timing/model': 1.1201745840953663, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 1.93}) +[2026-06-24 05:56:58,780][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.01321610901504755, 'train/grad_norm_pre_clip': 17.65951156616211, 'timing/data': 0.0003753039054572582, 'timing/model': 1.1189769400516525, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 1.93}) +[2026-06-24 05:57:02,928][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.01166817918419838, 'train/grad_norm_pre_clip': 3.187502384185791, 'timing/data': 0.000368754961527884, 'timing/model': 1.1204867550404742, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 1.93}) +[2026-06-24 05:57:07,006][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.0013960676733404398, 'train/grad_norm_pre_clip': 4.245360374450684, 'timing/data': 0.00035853590816259384, 'timing/model': 1.1133509729988873, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 1.93}) +[2026-06-24 05:57:11,085][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.19120480120182037, 'train/grad_norm_pre_clip': 22.356243133544922, 'timing/data': 0.0004483980592340231, 'timing/model': 1.1155973919667304, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 1.93}) +[2026-06-24 05:57:15,170][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.0009814273798838258, 'train/grad_norm_pre_clip': 4.100912094116211, 'timing/data': 0.0004525900585576892, 'timing/model': 1.1180395809933543, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 1.93}) +[2026-06-24 05:57:19,337][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.002857351442798972, 'train/grad_norm_pre_clip': 25.1568546295166, 'timing/data': 0.0004284520400688052, 'timing/model': 1.1195882760221139, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 1.93}) +[2026-06-24 05:57:23,420][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.009243707172572613, 'train/grad_norm_pre_clip': 25.500883102416992, 'timing/data': 0.00041663297452032566, 'timing/model': 1.1213935560081154, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 1.93}) +[2026-06-24 05:57:27,505][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.004989526700228453, 'train/grad_norm_pre_clip': 11.749225616455078, 'timing/data': 0.0005525239976122975, 'timing/model': 1.1174823580076918, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 1.93}) +[2026-06-24 05:57:31,593][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.003769232425838709, 'train/grad_norm_pre_clip': 17.959348678588867, 'timing/data': 0.00036148400977253914, 'timing/model': 1.125289116986096, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 1.93}) +[2026-06-24 05:57:35,817][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.016103943809866905, 'train/grad_norm_pre_clip': 10.785380363464355, 'timing/data': 0.00037375802639871836, 'timing/model': 1.1167362870182842, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 1.93}) +[2026-06-24 05:57:39,906][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.043846625834703445, 'train/grad_norm_pre_clip': 43.53014373779297, 'timing/data': 0.00039129098877310753, 'timing/model': 1.1234443120192736, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 1.93}) +[2026-06-24 05:57:43,984][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.009889628738164902, 'train/grad_norm_pre_clip': 15.401172637939453, 'timing/data': 0.0003670619335025549, 'timing/model': 1.1144926650449634, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 1.93}) +[2026-06-24 05:57:48,071][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.07546600699424744, 'train/grad_norm_pre_clip': 5.138674736022949, 'timing/data': 0.0005689439130946994, 'timing/model': 1.121734775020741, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 1.93}) +[2026-06-24 05:57:52,264][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.027308618649840355, 'train/grad_norm_pre_clip': 21.46331787109375, 'timing/data': 0.00035960704553872347, 'timing/model': 1.11731820309069, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 1.93}) +[2026-06-24 05:57:56,356][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.10044548660516739, 'train/grad_norm_pre_clip': 29.296903610229492, 'timing/data': 0.0004414429422467947, 'timing/model': 1.1229196639033034, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 1.93}) +[2026-06-24 05:58:00,444][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.003917714115232229, 'train/grad_norm_pre_clip': 9.679804801940918, 'timing/data': 0.0003627779660746455, 'timing/model': 1.118888908997178, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 1.93}) +[2026-06-24 05:58:04,544][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.10525117814540863, 'train/grad_norm_pre_clip': 27.67472267150879, 'timing/data': 0.00037831999361515045, 'timing/model': 1.123075942043215, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 1.93}) +[2026-06-24 05:58:08,701][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.01120692864060402, 'train/grad_norm_pre_clip': 9.395792007446289, 'timing/data': 0.0005630230298265815, 'timing/model': 1.1203095678938553, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 1.93}) +[2026-06-24 05:58:12,809][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.0008330281707458198, 'train/grad_norm_pre_clip': 26.562744140625, 'timing/data': 0.0004280840512365103, 'timing/model': 1.1227431279839948, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 1.93}) +[2026-06-24 05:58:16,913][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.004105333238840103, 'train/grad_norm_pre_clip': 28.635622024536133, 'timing/data': 0.0003736709477379918, 'timing/model': 1.1255706750089303, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 1.93}) +[2026-06-24 05:58:21,003][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.007894174195826054, 'train/grad_norm_pre_clip': 11.506386756896973, 'timing/data': 0.0005672179395332932, 'timing/model': 1.1196753300027922, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 1.93}) +[2026-06-24 05:58:25,261][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.028995228931307793, 'train/grad_norm_pre_clip': 30.49713897705078, 'timing/data': 0.00035879400093108416, 'timing/model': 1.118450182955712, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 1.93}) +[2026-06-24 05:58:29,351][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 5000 +[2026-06-24 05:59:33,859][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 5000 +[2026-06-24 05:59:33,860][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 5000 +[2026-06-24 06:00:12,985][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 5000 +[2026-06-24 06:12:07,270][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 5000 +[2026-06-24 06:12:07,270][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.009611685760319233, 'train/grad_norm_pre_clip': 2.83268141746521, 'eval/action_loss/seconds': 64.50504500290845, 'eval/action_loss/samples': 3200.0, 'eval/loss': 0.02939782105386257, 'eval/latency_2/loss': 0.02939782105386257, 'eval/flappy/loss': 0.02939782105386257, 'eval/flappy/latency_2/loss': 0.02939782105386257, 'eval/flap_precision': 0.9794520547945206, 'eval/flap_recall': 0.9701492537313433, 'eval/flap_f1': 0.9747784594410361, 'eval/flap_rate_model': 0.20277777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.974778459441036, 'eval/latency_2/flap_precision': 0.9794520547945206, 'eval/latency_2/flap_recall': 0.9701492537313433, 'eval/latency_2/flap_f1': 0.9747784594410361, 'eval/latency_2/flap_rate_model': 0.20277777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.974778459441036, 'eval/action_classification/data_wait_seconds': 0.6217380166053772, 'eval/action_classification/predict_seconds': 38.347225189208984, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 442.26000000002523, 'rl_games_eval/mid_train/mean_length': 3582.4, 'rl_games_eval/mid_train/std_reward': 4.680000000000791, 'rl_games_eval/mid_train/std_length': 35.2, 'rl_games_eval/mid_train/macro_mean_reward': 442.26000000002523, 'rl_games_eval/mid_train/macro_mean_length': 3582.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 442.26000000002523, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3582.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 4.680000000000791, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 35.2, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 442.26000000002523, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0004460080526769161, 'timing/model': 1.115945472032763, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 1.94}) +[2026-06-24 06:12:07,270][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_5000_state +[2026-06-24 06:12:07,273][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 06:12:10,948][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_5000_state/model.safetensors +[2026-06-24 06:12:18,717][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_5000_state/optimizer.bin +[2026-06-24 06:12:18,718][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_5000_state/sampler.bin +[2026-06-24 06:12:18,718][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-24 06:12:18,719][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-24 06:12:18,808][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-24 06:12:18,808][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 06:12:18,820][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 06:12:18,820][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected/checkpoints +[2026-06-24 06:12:21,652][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..a0e21985e6b7ed5b88fddcc5f74f6ec45063b603 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected/summary.jsonl @@ -0,0 +1,10 @@ +{"steps": 500} +{"steps": 1000} +{"steps": 1500} +{"steps": 2000} +{"steps": 2500} +{"steps": 3000} +{"steps": 3500} +{"steps": 4000} +{"steps": 4500} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_3000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_3000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3b8dbfbb90e7c22cea6e866b540dfc827f901805 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_3000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dee60fc99cf35e2e6d5908129093c8c66c2cafb62d3b87a236d45acd86522d59 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..a849162c2ba06d3b475af76a9c6157113b1c3e82 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/config.full.yaml @@ -0,0 +1,224 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 3000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 250 + eval_num_batches: 50 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 1 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: data/flappy_fix_latency_2_200ep + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: none + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..17e60dfaf8f7917c3f1712979fb62c2977b8d440 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/config.yaml @@ -0,0 +1,103 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + eval_num_workers: 8 + include_state: true + latency_curriculum: + enabled: false + obs_image_size: null + per_device_batch_size: 32 + persistent_workers: true + pin_memory: true + prefetch_factor: 4 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps +rl_games: + env_eval: + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 1 + image_size: 224 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: '' + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: flappy +run_id: flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_action_classification_interval: null + eval_interval: 250 + eval_num_batches: 50 + freeze_llm_layers: [] + freeze_modules: '' + freeze_tied_embedding: false + freeze_vit: false + gradient_accumulation_steps: 1 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 3000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + profile_timing: + enabled: false + reload_modules: null + save_interval: 500 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..aa03e6cbf8e375d660528e17fafddb4804b596b7 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.6028500199317932, + 0.3971499800682068, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4890792667865753, + 0.4890792667865753, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 330734, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..0caa254376d5805f972b0b0411e0b5188e3634e3 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7959861159324646, + 0.2040138840675354, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4030573070049286, + 0.4030573070049286, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..bbc82ee3e1d788a4aa3fd2151d4e45b6c7a36392 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra/.hydra/config.yaml @@ -0,0 +1,222 @@ +framework: + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 3000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 250 + eval_num_batches: 50 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 1 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: data/flappy_fix_latency_2_200ep + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: none + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps +output_dir: null +config_yaml: null +is_debug: false +version_id: 0.21 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..103e5757697618d2b1c36d2f750ead97b0c9137e --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra/.hydra/hydra.yaml @@ -0,0 +1,341 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - ++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++framework.qwenvl.attn_implementation=flash_attention_2 + - ++framework.qwenvl.enable_gradient_checkpointing=true + - ++framework.action_model.state_dim=7 + - ++framework.action_model.loss_type=discrete_ce + - ++framework.action_model.action_horizon=1 + - ++framework.action_model.future_action_window_size=0 + - ++framework.action_model.past_action_window_size=0 + - ++framework.action_model.action_dim=7 + - ++framework.action_model.action_env_dim=2 + - ++framework.name=QwenOFT + - ++datasets.vla_data.dataset_py=lerobot_datasets + - ++datasets.vla_data.include_state=true + - ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games + - ++datasets.vla_data.data_mix=flappy_train + - ++datasets.vla_data.eval_data_mix=null + - ++datasets.vla_data.custom_mixtures_path=null + - ++datasets.vla_data.action_type=discrete + - ++datasets.vla_data.sequential_step_sampling=false + - ++datasets.vla_data.eval_sequential_step_sampling=null + - ++datasets.vla_data.num_workers=8 + - ++datasets.vla_data.eval_num_workers=8 + - ++datasets.vla_data.prefetch_factor=4 + - ++datasets.vla_data.persistent_workers=true + - ++datasets.vla_data.pin_memory=true + - ++datasets.vla_data.shuffle=true + - ++datasets.vla_data.action_balance.enabled=false + - ++datasets.vla_data.action_balance.strategy=balanced_epoch + - ++datasets.vla_data.action_balance.action_key=action_id + - ++datasets.vla_data.action_balance.target_flap_fraction=0.3 + - ++datasets.vla_data.action_balance.noop_id=0 + - ++datasets.vla_data.action_balance.flap_id=1 + - ++datasets.vla_data.latency_curriculum.enabled=false + - ++datasets.vla_data.latency_curriculum.strategy=exclusive + - ++datasets.vla_data.latency_curriculum.latencies=null + - ++datasets.vla_data.latency_curriculum.phase_steps=null + - ++datasets.vla_data.per_device_batch_size=32 + - ++datasets.vla_data.load_all_data_for_training=true + - ++datasets.vla_data.num_obs_frames=1 + - ++datasets.vla_data.image_mode=single + - ++datasets.vla_data.stitch_grid=[2,2] + - ++datasets.vla_data.obs_image_size=null + - ++datasets.vla_data.video_backend=torchvision_av + - ++dataset.source_hf= + - ++dataset.config_name=null + - ++dataset.source_subdir=null + - ++dataset.converted_name=flappy_train + - ++dataset.single_source_hf= + - ++dataset.mixed_source_hf= + - ++dataset.single_converted_name=flappy_train + - ++dataset.mixed_converted_name=flappy_mixed_latency_train + - ++dataset.single_latency_filter=null + - ++dataset.mixed_latency_filter=null + - ++dataset.force_download=false + - ++dataset.setup_force=false + - ++dataset.skip_verification=false + - ++dataset.verify_rows=200 + - ++dataset.max_episodes=null + - ++dataset.episodes_per_latency=null + - ++dataset.latency_filter=null + - ++dataset.debug_subset.enabled=false + - ++dataset.debug_subset.max_episodes=5 + - ++dataset.debug_subset.suffix=debug + - ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct + - ++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + - ++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1 + - ++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt + - ++trainer.max_train_steps=3000 + - ++trainer.num_warmup_steps=100 + - ++trainer.save_interval=500 + - ++trainer.eval_interval=250 + - ++trainer.eval_num_batches=50 + - ++trainer.per_latency_eval_num_batches=null + - ++trainer.eval_action_classification=true + - ++trainer.eval_action_classification_interval=null + - ++trainer.cc_f1_tolerance=1 + - ++trainer.learning_rate.base=2e-05 + - ++trainer.learning_rate.qwen_vl_interface=1e-05 + - ++trainer.learning_rate.action_model=0.0001 + - ++trainer.lr_scheduler_type=cosine_with_min_lr + - ++trainer.scheduler_specific_kwargs.min_lr=1e-06 + - ++trainer.freeze_modules= + - ++trainer.freeze_vit=false + - ++trainer.freeze_tied_embedding=false + - ++trainer.freeze_llm_layers=[] + - ++trainer.loss_scale.vla=1.0 + - ++trainer.loss_scale.vlm=0.1 + - ++trainer.max_grad_norm=1.0 + - ++trainer.weight_decay=0.0 + - ++trainer.logging_frequency=1 + - ++trainer.profile_timing.enabled=false + - ++trainer.profile_timing.log_interval=10 + - ++trainer.gradient_clipping=1.0 + - ++trainer.gradient_accumulation_steps=1 + - ++trainer.distributed_backend=none + - ++trainer.is_resume=false + - ++trainer.pretrained_checkpoint=null + - ++trainer.resume_step=0 + - ++trainer.reload_modules=null + - ++trainer.optimizer.name=AdamW + - ++trainer.optimizer.betas=[0.9,0.95] + - ++trainer.optimizer.eps=1e-08 + - ++trainer.optimizer.weight_decay=1e-08 + - ++trainer.optimizer.fused=true + - ++trainer.save_format=pt + - ++workspace_dir=WORKSPACE_DIR + - ++run_root_dir=results/Checkpoints + - ++seed=42 + - ++wandb_entity=saberrr-zju + - ++wandb_project=starVLA_rl_games + - ++auth.env_file=null + - ++auth.hf_token_env=HF_TOKEN + - ++auth.wandb_api_key_env=WANDB_API_KEY + - ++paths.run_root_dir=results/Checkpoints + - ++paths.dataset_local_dir=data/flappy_fix_latency_2_200ep + - ++paths.dataset_cache_dir=null + - ++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml + - ++rl_games.model_alias=openvla + - ++rl_games.env_eval.image_size=224 + - ++rl_games.env_eval.frameskip=1 + - ++rl_games.env_eval.seed=42 + - ++rl_games.env_eval.fixed_episode_seeds=true + - ++rl_games.env_eval.latency_seed_stride=0 + - ++rl_games.env_eval.task_seed_stride=0 + - ++rl_games.env_eval.task_description= + - ++rl_games.env_eval.enabled=true + - ++rl_games.env_eval.distributed_mode=none + - ++rl_games.env_eval.vectorized.enabled=false + - ++rl_games.env_eval.vectorized.batch_size=1 + - ++rl_games.env_eval.latency.prompt_map_path=null + - ++rl_games.env_eval.latency.mode=single + - ++rl_games.env_eval.latency.values=[0] + - ++rl_games.env_eval.mid_train.enabled=true + - ++rl_games.env_eval.mid_train.interval_steps=250 + - ++rl_games.env_eval.mid_train.latencies=[2] + - ++rl_games.env_eval.mid_train.num_episodes=5 + - ++rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - ++rl_games.env_eval.post_train.enabled=false + - ++rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + - ++rl_games.env_eval.post_train.num_episodes=5 + - ++rl_games.env_eval.post_train.max_steps_per_episode=3600 + - ++rl_games.task=flappy + - ++rl_games.initialization_mode=bridge + - ++rl_games.action_carrier=bridge + - ++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct + - ++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action + - ++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct + - ++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct + - ++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + - ++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++checkpoint.load=none + - ++checkpoint.hf_repo_id=null + - ++checkpoint.save_best_model=false + - ++checkpoint.save_pt_file=false + - ++checkpoint.local.keep_last_n=1 + - ++checkpoint.sync.enabled=false + - ++checkpoint.sync.repo_id=null + - ++checkpoint.sync.keep_last_n=0 + - ++checkpoint.sync.sync_every_n_checkpoints=1 + - ++checkpoint.sync.resume_policy=local_latest + - ++run_id=flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps + - ++output_dir=null + - ++config_yaml=null + - ++is_debug=false + - ++version_id=0.21 + - ++run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints + - ++trainer.is_resume=false + - ++trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - ++trainer.resume_step=0 + - ++datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + - ++datasets.vla_data.data_mix=flappy_train__bridge + - ++datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - ++framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + job: + name: train_starvla_hydra + chdir: false + override_dirname: ++auth.env_file=null,++auth.hf_token_env=HF_TOKEN,++auth.wandb_api_key_env=WANDB_API_KEY,++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action,++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action,++checkpoint.hf_repo_id=null,++checkpoint.load=none,++checkpoint.local.keep_last_n=1,++checkpoint.save_best_model=false,++checkpoint.save_pt_file=false,++checkpoint.sync.enabled=false,++checkpoint.sync.keep_last_n=0,++checkpoint.sync.repo_id=null,++checkpoint.sync.resume_policy=local_latest,++checkpoint.sync.sync_every_n_checkpoints=1,++config_yaml=null,++dataset.config_name=null,++dataset.converted_name=flappy_train,++dataset.debug_subset.enabled=false,++dataset.debug_subset.max_episodes=5,++dataset.debug_subset.suffix=debug,++dataset.episodes_per_latency=null,++dataset.force_download=false,++dataset.latency_filter=null,++dataset.max_episodes=null,++dataset.mixed_converted_name=flappy_mixed_latency_train,++dataset.mixed_latency_filter=null,++dataset.mixed_source_hf=,++dataset.setup_force=false,++dataset.single_converted_name=flappy_train,++dataset.single_latency_filter=null,++dataset.single_source_hf=,++dataset.skip_verification=false,++dataset.source_hf=,++dataset.source_subdir=null,++dataset.verify_rows=200,++datasets.vla_data.action_balance.action_key=action_id,++datasets.vla_data.action_balance.enabled=false,++datasets.vla_data.action_balance.flap_id=1,++datasets.vla_data.action_balance.noop_id=0,++datasets.vla_data.action_balance.strategy=balanced_epoch,++datasets.vla_data.action_balance.target_flap_fraction=0.3,++datasets.vla_data.action_type=discrete,++datasets.vla_data.custom_mixtures_path=null,++datasets.vla_data.data_mix=flappy_train,++datasets.vla_data.data_mix=flappy_train__bridge,++datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep,++datasets.vla_data.data_root_dir=playground/Datasets/rl_games,++datasets.vla_data.dataset_py=lerobot_datasets,++datasets.vla_data.eval_data_mix=flappy_train__bridge__val,++datasets.vla_data.eval_data_mix=null,++datasets.vla_data.eval_num_workers=8,++datasets.vla_data.eval_sequential_step_sampling=null,++datasets.vla_data.image_mode=single,++datasets.vla_data.include_state=true,++datasets.vla_data.latency_curriculum.enabled=false,++datasets.vla_data.latency_curriculum.latencies=null,++datasets.vla_data.latency_curriculum.phase_steps=null,++datasets.vla_data.latency_curriculum.strategy=exclusive,++datasets.vla_data.load_all_data_for_training=true,++datasets.vla_data.num_obs_frames=1,++datasets.vla_data.num_workers=8,++datasets.vla_data.obs_image_size=null,++datasets.vla_data.per_device_batch_size=32,++datasets.vla_data.persistent_workers=true,++datasets.vla_data.pin_memory=true,++datasets.vla_data.prefetch_factor=4,++datasets.vla_data.sequential_step_sampling=false,++datasets.vla_data.shuffle=true,++datasets.vla_data.stitch_grid=[2,2],++datasets.vla_data.video_backend=torchvision_av,++framework.action_model.action_dim=7,++framework.action_model.action_env_dim=2,++framework.action_model.action_horizon=1,++framework.action_model.future_action_window_size=0,++framework.action_model.loss_type=discrete_ce,++framework.action_model.past_action_window_size=0,++framework.action_model.state_dim=7,++framework.name=QwenOFT,++framework.qwenvl.attn_implementation=flash_attention_2,++framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.enable_gradient_checkpointing=true,++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt,++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1,++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1,++is_debug=false,++output_dir=null,++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml,++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++paths.dataset_cache_dir=null,++paths.dataset_local_dir=data/flappy_fix_latency_2_200ep,++paths.run_root_dir=results/Checkpoints,++rl_games.action_carrier=bridge,++rl_games.env_eval.distributed_mode=none,++rl_games.env_eval.enabled=true,++rl_games.env_eval.fixed_episode_seeds=true,++rl_games.env_eval.frameskip=1,++rl_games.env_eval.image_size=224,++rl_games.env_eval.latency.mode=single,++rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json,++rl_games.env_eval.latency.prompt_map_path=null,++rl_games.env_eval.latency.values=[0],++rl_games.env_eval.latency_seed_stride=0,++rl_games.env_eval.mid_train.enabled=true,++rl_games.env_eval.mid_train.interval_steps=250,++rl_games.env_eval.mid_train.latencies=[2],++rl_games.env_eval.mid_train.max_steps_per_episode=3600,++rl_games.env_eval.mid_train.num_episodes=5,++rl_games.env_eval.post_train.enabled=false,++rl_games.env_eval.post_train.latencies=[0,1,2,3,4],++rl_games.env_eval.post_train.max_steps_per_episode=3600,++rl_games.env_eval.post_train.num_episodes=5,++rl_games.env_eval.seed=42,++rl_games.env_eval.task_description=,++rl_games.env_eval.task_seed_stride=0,++rl_games.env_eval.vectorized.batch_size=1,++rl_games.env_eval.vectorized.enabled=false,++rl_games.initialization_mode=bridge,++rl_games.model_alias=openvla,++rl_games.task=flappy,++run_id=flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps,++run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints,++run_root_dir=results/Checkpoints,++seed=42,++trainer.cc_f1_tolerance=1,++trainer.distributed_backend=none,++trainer.eval_action_classification=true,++trainer.eval_action_classification_interval=null,++trainer.eval_interval=250,++trainer.eval_num_batches=50,++trainer.freeze_llm_layers=[],++trainer.freeze_modules=,++trainer.freeze_tied_embedding=false,++trainer.freeze_vit=false,++trainer.gradient_accumulation_steps=1,++trainer.gradient_clipping=1.0,++trainer.is_resume=false,++trainer.is_resume=false,++trainer.learning_rate.action_model=0.0001,++trainer.learning_rate.base=2e-05,++trainer.learning_rate.qwen_vl_interface=1e-05,++trainer.logging_frequency=1,++trainer.loss_scale.vla=1.0,++trainer.loss_scale.vlm=0.1,++trainer.lr_scheduler_type=cosine_with_min_lr,++trainer.max_grad_norm=1.0,++trainer.max_train_steps=3000,++trainer.num_warmup_steps=100,++trainer.optimizer.betas=[0.9,0.95],++trainer.optimizer.eps=1e-08,++trainer.optimizer.fused=true,++trainer.optimizer.name=AdamW,++trainer.optimizer.weight_decay=1e-08,++trainer.per_latency_eval_num_batches=null,++trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,++trainer.pretrained_checkpoint=null,++trainer.profile_timing.enabled=false,++trainer.profile_timing.log_interval=10,++trainer.reload_modules=null,++trainer.resume_step=0,++trainer.resume_step=0,++trainer.save_format=pt,++trainer.save_interval=500,++trainer.scheduler_specific_kwargs.min_lr=1e-06,++trainer.weight_decay=0.0,++version_id=0.21,++wandb_entity=saberrr-zju,++wandb_project=starVLA_rl_games,++workspace_dir=WORKSPACE_DIR,env=flappy,init=bridge,mode=single,model=openvla + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/latency-sensitive-bench/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/latency-sensitive-bench/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..595353b2e4440cedb7d0835bcb8ebfe28bdf363a --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra/.hydra/overrides.yaml @@ -0,0 +1,181 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- ++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++framework.qwenvl.attn_implementation=flash_attention_2 +- ++framework.qwenvl.enable_gradient_checkpointing=true +- ++framework.action_model.state_dim=7 +- ++framework.action_model.loss_type=discrete_ce +- ++framework.action_model.action_horizon=1 +- ++framework.action_model.future_action_window_size=0 +- ++framework.action_model.past_action_window_size=0 +- ++framework.action_model.action_dim=7 +- ++framework.action_model.action_env_dim=2 +- ++framework.name=QwenOFT +- ++datasets.vla_data.dataset_py=lerobot_datasets +- ++datasets.vla_data.include_state=true +- ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games +- ++datasets.vla_data.data_mix=flappy_train +- ++datasets.vla_data.eval_data_mix=null +- ++datasets.vla_data.custom_mixtures_path=null +- ++datasets.vla_data.action_type=discrete +- ++datasets.vla_data.sequential_step_sampling=false +- ++datasets.vla_data.eval_sequential_step_sampling=null +- ++datasets.vla_data.num_workers=8 +- ++datasets.vla_data.eval_num_workers=8 +- ++datasets.vla_data.prefetch_factor=4 +- ++datasets.vla_data.persistent_workers=true +- ++datasets.vla_data.pin_memory=true +- ++datasets.vla_data.shuffle=true +- ++datasets.vla_data.action_balance.enabled=false +- ++datasets.vla_data.action_balance.strategy=balanced_epoch +- ++datasets.vla_data.action_balance.action_key=action_id +- ++datasets.vla_data.action_balance.target_flap_fraction=0.3 +- ++datasets.vla_data.action_balance.noop_id=0 +- ++datasets.vla_data.action_balance.flap_id=1 +- ++datasets.vla_data.latency_curriculum.enabled=false +- ++datasets.vla_data.latency_curriculum.strategy=exclusive +- ++datasets.vla_data.latency_curriculum.latencies=null +- ++datasets.vla_data.latency_curriculum.phase_steps=null +- ++datasets.vla_data.per_device_batch_size=32 +- ++datasets.vla_data.load_all_data_for_training=true +- ++datasets.vla_data.num_obs_frames=1 +- ++datasets.vla_data.image_mode=single +- ++datasets.vla_data.stitch_grid=[2,2] +- ++datasets.vla_data.obs_image_size=null +- ++datasets.vla_data.video_backend=torchvision_av +- ++dataset.source_hf= +- ++dataset.config_name=null +- ++dataset.source_subdir=null +- ++dataset.converted_name=flappy_train +- ++dataset.single_source_hf= +- ++dataset.mixed_source_hf= +- ++dataset.single_converted_name=flappy_train +- ++dataset.mixed_converted_name=flappy_mixed_latency_train +- ++dataset.single_latency_filter=null +- ++dataset.mixed_latency_filter=null +- ++dataset.force_download=false +- ++dataset.setup_force=false +- ++dataset.skip_verification=false +- ++dataset.verify_rows=200 +- ++dataset.max_episodes=null +- ++dataset.episodes_per_latency=null +- ++dataset.latency_filter=null +- ++dataset.debug_subset.enabled=false +- ++dataset.debug_subset.max_episodes=5 +- ++dataset.debug_subset.suffix=debug +- ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct +- ++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 +- ++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1 +- ++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt +- ++trainer.max_train_steps=3000 +- ++trainer.num_warmup_steps=100 +- ++trainer.save_interval=500 +- ++trainer.eval_interval=250 +- ++trainer.eval_num_batches=50 +- ++trainer.per_latency_eval_num_batches=null +- ++trainer.eval_action_classification=true +- ++trainer.eval_action_classification_interval=null +- ++trainer.cc_f1_tolerance=1 +- ++trainer.learning_rate.base=2e-05 +- ++trainer.learning_rate.qwen_vl_interface=1e-05 +- ++trainer.learning_rate.action_model=0.0001 +- ++trainer.lr_scheduler_type=cosine_with_min_lr +- ++trainer.scheduler_specific_kwargs.min_lr=1e-06 +- ++trainer.freeze_modules= +- ++trainer.freeze_vit=false +- ++trainer.freeze_tied_embedding=false +- ++trainer.freeze_llm_layers=[] +- ++trainer.loss_scale.vla=1.0 +- ++trainer.loss_scale.vlm=0.1 +- ++trainer.max_grad_norm=1.0 +- ++trainer.weight_decay=0.0 +- ++trainer.logging_frequency=1 +- ++trainer.profile_timing.enabled=false +- ++trainer.profile_timing.log_interval=10 +- ++trainer.gradient_clipping=1.0 +- ++trainer.gradient_accumulation_steps=1 +- ++trainer.distributed_backend=none +- ++trainer.is_resume=false +- ++trainer.pretrained_checkpoint=null +- ++trainer.resume_step=0 +- ++trainer.reload_modules=null +- ++trainer.optimizer.name=AdamW +- ++trainer.optimizer.betas=[0.9,0.95] +- ++trainer.optimizer.eps=1e-08 +- ++trainer.optimizer.weight_decay=1e-08 +- ++trainer.optimizer.fused=true +- ++trainer.save_format=pt +- ++workspace_dir=WORKSPACE_DIR +- ++run_root_dir=results/Checkpoints +- ++seed=42 +- ++wandb_entity=saberrr-zju +- ++wandb_project=starVLA_rl_games +- ++auth.env_file=null +- ++auth.hf_token_env=HF_TOKEN +- ++auth.wandb_api_key_env=WANDB_API_KEY +- ++paths.run_root_dir=results/Checkpoints +- ++paths.dataset_local_dir=data/flappy_fix_latency_2_200ep +- ++paths.dataset_cache_dir=null +- ++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml +- ++rl_games.model_alias=openvla +- ++rl_games.env_eval.image_size=224 +- ++rl_games.env_eval.frameskip=1 +- ++rl_games.env_eval.seed=42 +- ++rl_games.env_eval.fixed_episode_seeds=true +- ++rl_games.env_eval.latency_seed_stride=0 +- ++rl_games.env_eval.task_seed_stride=0 +- ++rl_games.env_eval.task_description= +- ++rl_games.env_eval.enabled=true +- ++rl_games.env_eval.distributed_mode=none +- ++rl_games.env_eval.vectorized.enabled=false +- ++rl_games.env_eval.vectorized.batch_size=1 +- ++rl_games.env_eval.latency.prompt_map_path=null +- ++rl_games.env_eval.latency.mode=single +- ++rl_games.env_eval.latency.values=[0] +- ++rl_games.env_eval.mid_train.enabled=true +- ++rl_games.env_eval.mid_train.interval_steps=250 +- ++rl_games.env_eval.mid_train.latencies=[2] +- ++rl_games.env_eval.mid_train.num_episodes=5 +- ++rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- ++rl_games.env_eval.post_train.enabled=false +- ++rl_games.env_eval.post_train.latencies=[0,1,2,3,4] +- ++rl_games.env_eval.post_train.num_episodes=5 +- ++rl_games.env_eval.post_train.max_steps_per_episode=3600 +- ++rl_games.task=flappy +- ++rl_games.initialization_mode=bridge +- ++rl_games.action_carrier=bridge +- ++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct +- ++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action +- ++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct +- ++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct +- ++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action +- ++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++checkpoint.load=none +- ++checkpoint.hf_repo_id=null +- ++checkpoint.save_best_model=false +- ++checkpoint.save_pt_file=false +- ++checkpoint.local.keep_last_n=1 +- ++checkpoint.sync.enabled=false +- ++checkpoint.sync.repo_id=null +- ++checkpoint.sync.keep_last_n=0 +- ++checkpoint.sync.sync_every_n_checkpoints=1 +- ++checkpoint.sync.resume_policy=local_latest +- ++run_id=flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps +- ++output_dir=null +- ++config_yaml=null +- ++is_debug=false +- ++version_id=0.21 +- ++run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints +- ++trainer.is_resume=false +- ++trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- ++trainer.resume_step=0 +- ++datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep +- ++datasets.vla_data.data_mix=flappy_train__bridge +- ++datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- ++framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..26b63cc9b960d6b36df1d9462f6d32ff3b9e9939 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/hydra/train_starvla_hydra.log @@ -0,0 +1,3176 @@ +[2026-06-24 09:52:26,590][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21' +[2026-06-24 09:52:26,590][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-24 09:52:26,799][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured. +[2026-06-24 09:52:26,800][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub +[2026-06-24 09:52:27,158][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases +[2026-06-24 09:52:27,165][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-24 09:52:31,309][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-24 09:52:31,337][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-24 09:52:31,752][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-24 09:52:31,752][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-24 09:52:31,752][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-24 09:52:31,752][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-24 09:52:31,752][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-24 09:52:31,752][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-24 09:52:31,752][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-24 09:52:31,752][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-24 09:52:31,752][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-24 09:52:31,752][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-24 09:52:31,753][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-24 09:52:31,775][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-24 09:52:31,793][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/config.full.yaml +[2026-06-24 09:52:31,798][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/config.yaml +[2026-06-24 09:52:31,798][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-24 09:52:33,654][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-24 09:52:33,654][starVLA.training.train_starvla][INFO] - Total optimization steps = 3000 +[2026-06-24 09:52:33,654][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-24 09:52:33,655][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 1 +[2026-06-24 09:52:33,655][starVLA.training.train_starvla][INFO] - Total batch size = 32 +[2026-06-24 09:52:57,018][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6939594745635986, 'train/grad_norm_pre_clip': 13.036249160766602, 'timing/data': 21.6081434590742, 'timing/model': 1.7519401400350034, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-24 09:52:58,005][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6941108703613281, 'train/grad_norm_pre_clip': 0.8917124271392822, 'timing/data': 0.0007200508844107389, 'timing/model': 0.9829894760623574, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-24 09:52:58,977][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.693626880645752, 'train/grad_norm_pre_clip': 0.6961520910263062, 'timing/data': 0.0007698540575802326, 'timing/model': 0.9685630919411778, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-24 09:52:59,968][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.6873018145561218, 'train/grad_norm_pre_clip': 21.74880027770996, 'timing/data': 0.0006221281364560127, 'timing/model': 0.9876434300094843, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-24 09:53:00,965][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6787329316139221, 'train/grad_norm_pre_clip': 17.295516967773438, 'timing/data': 0.0007390871178358793, 'timing/model': 0.9939895179122686, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-24 09:53:01,955][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6819617748260498, 'train/grad_norm_pre_clip': 5.762785911560059, 'timing/data': 0.0006899749860167503, 'timing/model': 0.985204789089039, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-24 09:53:02,939][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.6692277789115906, 'train/grad_norm_pre_clip': 6.404450416564941, 'timing/data': 0.0007938491180539131, 'timing/model': 0.9806370919104666, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-24 09:53:03,910][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.7126042246818542, 'train/grad_norm_pre_clip': 9.927355766296387, 'timing/data': 0.0007975560147315264, 'timing/model': 0.9673372518736869, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-24 09:53:04,865][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.6452586054801941, 'train/grad_norm_pre_clip': 4.096685409545898, 'timing/data': 0.0008764038793742657, 'timing/model': 0.9521770710125566, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-24 09:53:05,754][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.5682671666145325, 'train/grad_norm_pre_clip': 14.037934303283691, 'timing/data': 0.0006648269481956959, 'timing/model': 0.8866893709637225, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-24 09:53:06,632][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.5729108452796936, 'train/grad_norm_pre_clip': 7.4510393142700195, 'timing/data': 0.0003734300844371319, 'timing/model': 0.8757860260084271, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-24 09:53:07,513][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.7107690572738647, 'train/grad_norm_pre_clip': 12.83685302734375, 'timing/data': 0.0004013630095869303, 'timing/model': 0.8790814499370754, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-24 09:53:08,394][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.6272303462028503, 'train/grad_norm_pre_clip': 6.087057590484619, 'timing/data': 0.00035012606531381607, 'timing/model': 0.879151863977313, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.0}) +[2026-06-24 09:53:09,275][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.70478755235672, 'train/grad_norm_pre_clip': 13.786855697631836, 'timing/data': 0.000372899929061532, 'timing/model': 0.8785705850459635, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.0}) +[2026-06-24 09:53:10,153][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.7058838605880737, 'train/grad_norm_pre_clip': 14.282752990722656, 'timing/data': 0.0003542539197951555, 'timing/model': 0.8769641318358481, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.0}) +[2026-06-24 09:53:11,036][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.8076281547546387, 'train/grad_norm_pre_clip': 23.47836685180664, 'timing/data': 0.00035766116343438625, 'timing/model': 0.8807803730014712, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.0}) +[2026-06-24 09:53:11,915][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.7597971558570862, 'train/grad_norm_pre_clip': 18.755596160888672, 'timing/data': 0.0003742820117622614, 'timing/model': 0.8774185469374061, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.0}) +[2026-06-24 09:53:12,797][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.7093701958656311, 'train/grad_norm_pre_clip': 12.940914154052734, 'timing/data': 0.00043264194391667843, 'timing/model': 0.8797216699458659, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.0}) +[2026-06-24 09:53:13,679][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.7388319969177246, 'train/grad_norm_pre_clip': 15.635051727294922, 'timing/data': 0.00033975811675190926, 'timing/model': 0.8796347079332918, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.0}) +[2026-06-24 09:53:14,560][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.6900069713592529, 'train/grad_norm_pre_clip': 9.162191390991211, 'timing/data': 0.0004366990178823471, 'timing/model': 0.8795015800278634, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.0}) +[2026-06-24 09:53:15,441][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.6817629933357239, 'train/grad_norm_pre_clip': 2.0569324493408203, 'timing/data': 0.0004237948451191187, 'timing/model': 0.8793987280223519, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.0}) +[2026-06-24 09:53:16,326][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.6888807415962219, 'train/grad_norm_pre_clip': 6.275496482849121, 'timing/data': 0.00036809989251196384, 'timing/model': 0.8822531460318714, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.0}) +[2026-06-24 09:53:17,208][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.7178082466125488, 'train/grad_norm_pre_clip': 20.757293701171875, 'timing/data': 0.00036829104647040367, 'timing/model': 0.8803248079493642, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.0}) +[2026-06-24 09:53:18,090][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.7582570910453796, 'train/grad_norm_pre_clip': 30.095714569091797, 'timing/data': 0.000394430011510849, 'timing/model': 0.8802235098555684, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.0}) +[2026-06-24 09:53:18,976][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.7334418296813965, 'train/grad_norm_pre_clip': 18.106157302856445, 'timing/data': 0.00040074181742966175, 'timing/model': 0.8842236858326942, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.0}) +[2026-06-24 09:53:19,860][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.7519400715827942, 'train/grad_norm_pre_clip': 29.54555892944336, 'timing/data': 0.00036185886710882187, 'timing/model': 0.8819743769709021, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.0}) +[2026-06-24 09:53:20,745][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.6689049601554871, 'train/grad_norm_pre_clip': 9.29909896850586, 'timing/data': 0.0003525521606206894, 'timing/model': 0.883178697898984, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.0}) +[2026-06-24 09:53:21,631][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.7007376551628113, 'train/grad_norm_pre_clip': 16.10403823852539, 'timing/data': 0.00039564212784171104, 'timing/model': 0.8835066452156752, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.0}) +[2026-06-24 09:53:22,518][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.6792851686477661, 'train/grad_norm_pre_clip': 12.979602813720703, 'timing/data': 0.00037763803265988827, 'timing/model': 0.8847691551782191, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.0}) +[2026-06-24 09:53:23,403][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.6764053106307983, 'train/grad_norm_pre_clip': 2.382495403289795, 'timing/data': 0.0003691120073199272, 'timing/model': 0.8834294530097395, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.0}) +[2026-06-24 09:53:24,289][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.66363525390625, 'train/grad_norm_pre_clip': 2.656139612197876, 'timing/data': 0.00040823593735694885, 'timing/model': 0.8835951560176909, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.0}) +[2026-06-24 09:53:25,175][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.7074633836746216, 'train/grad_norm_pre_clip': 11.665531158447266, 'timing/data': 0.0003635419998317957, 'timing/model': 0.8842186760157347, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.0}) +[2026-06-24 09:53:26,061][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.7352895140647888, 'train/grad_norm_pre_clip': 18.077970504760742, 'timing/data': 0.0003716670908033848, 'timing/model': 0.8841000550892204, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.0}) +[2026-06-24 09:53:26,950][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.7638437747955322, 'train/grad_norm_pre_clip': 23.73177719116211, 'timing/data': 0.0005726339295506477, 'timing/model': 0.886544442968443, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.0}) +[2026-06-24 09:53:27,836][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.6713013052940369, 'train/grad_norm_pre_clip': 8.707956314086914, 'timing/data': 0.000363972969353199, 'timing/model': 0.8843371740076691, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.0}) +[2026-06-24 09:53:28,724][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.6284361481666565, 'train/grad_norm_pre_clip': 14.144904136657715, 'timing/data': 0.0003423921298235655, 'timing/model': 0.8859256950672716, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.0}) +[2026-06-24 09:53:29,609][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.6388333439826965, 'train/grad_norm_pre_clip': 7.6708292961120605, 'timing/data': 0.0003788699395954609, 'timing/model': 0.8833177529741079, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.0}) +[2026-06-24 09:53:30,497][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.6500242948532104, 'train/grad_norm_pre_clip': 14.447676658630371, 'timing/data': 0.00035422504879534245, 'timing/model': 0.8860631750430912, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.0}) +[2026-06-24 09:53:31,385][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.6664023995399475, 'train/grad_norm_pre_clip': 12.364534378051758, 'timing/data': 0.00032736407592892647, 'timing/model': 0.8859815241303295, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.0}) +[2026-06-24 09:53:32,451][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.6236662268638611, 'train/grad_norm_pre_clip': 7.563601493835449, 'timing/data': 0.00038168602623045444, 'timing/model': 1.064403021009639, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.0}) +[2026-06-24 09:53:33,336][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.6538398861885071, 'train/grad_norm_pre_clip': 16.466968536376953, 'timing/data': 0.0004396250005811453, 'timing/model': 0.882875989889726, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.0}) +[2026-06-24 09:53:34,239][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.6036526560783386, 'train/grad_norm_pre_clip': 14.318831443786621, 'timing/data': 0.00037749786861240864, 'timing/model': 0.9008849589154124, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.0}) +[2026-06-24 09:53:35,137][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.6212102770805359, 'train/grad_norm_pre_clip': 12.682866096496582, 'timing/data': 0.0005104879382997751, 'timing/model': 0.8957578109111637, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.0}) +[2026-06-24 09:53:36,029][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.576941728591919, 'train/grad_norm_pre_clip': 24.260452270507812, 'timing/data': 0.0003915349952876568, 'timing/model': 0.8896682481281459, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.0}) +[2026-06-24 09:53:36,918][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.5593615770339966, 'train/grad_norm_pre_clip': 15.063883781433105, 'timing/data': 0.0004152290057390928, 'timing/model': 0.887379014166072, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.0}) +[2026-06-24 09:53:37,814][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.459546834230423, 'train/grad_norm_pre_clip': 16.24410629272461, 'timing/data': 0.00039237597957253456, 'timing/model': 0.8943510870449245, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.0}) +[2026-06-24 09:53:38,706][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.4610981345176697, 'train/grad_norm_pre_clip': 26.930625915527344, 'timing/data': 0.00036791013553738594, 'timing/model': 0.8895875499583781, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.0}) +[2026-06-24 09:53:39,594][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.5981304049491882, 'train/grad_norm_pre_clip': 58.13402557373047, 'timing/data': 0.0003714070189744234, 'timing/model': 0.8862098548561335, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.0}) +[2026-06-24 09:53:40,481][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.5042813420295715, 'train/grad_norm_pre_clip': 40.19293212890625, 'timing/data': 0.0003892101813107729, 'timing/model': 0.8846905750688165, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.0}) +[2026-06-24 09:53:41,368][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.4889160990715027, 'train/grad_norm_pre_clip': 87.2235107421875, 'timing/data': 0.00036353315226733685, 'timing/model': 0.8852434840518981, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.0}) +[2026-06-24 09:53:42,262][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.3466253876686096, 'train/grad_norm_pre_clip': 16.624855041503906, 'timing/data': 0.0005813299212604761, 'timing/model': 0.8911638129502535, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.0}) +[2026-06-24 09:53:43,153][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.6329150199890137, 'train/grad_norm_pre_clip': 50.28030776977539, 'timing/data': 0.0004316701088100672, 'timing/model': 0.8897506028879434, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.01}) +[2026-06-24 09:53:44,043][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.5235038995742798, 'train/grad_norm_pre_clip': 27.809349060058594, 'timing/data': 0.00039158505387604237, 'timing/model': 0.8872425591107458, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:53:44,934][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.506528377532959, 'train/grad_norm_pre_clip': 23.770658493041992, 'timing/data': 0.0003632109146565199, 'timing/model': 0.8896205718629062, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.01}) +[2026-06-24 09:53:45,823][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.5942642688751221, 'train/grad_norm_pre_clip': 53.892723083496094, 'timing/data': 0.0003657061606645584, 'timing/model': 0.8872305690310895, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:53:46,715][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.4244113564491272, 'train/grad_norm_pre_clip': 20.039392471313477, 'timing/data': 0.00035564694553613663, 'timing/model': 0.8898356980644166, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.01}) +[2026-06-24 09:53:47,614][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.5558047890663147, 'train/grad_norm_pre_clip': 70.341064453125, 'timing/data': 0.0003910230007022619, 'timing/model': 0.8971621349919587, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.01}) +[2026-06-24 09:53:48,505][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.6173275709152222, 'train/grad_norm_pre_clip': 79.11314392089844, 'timing/data': 0.0004505449905991554, 'timing/model': 0.8890150310471654, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.01}) +[2026-06-24 09:53:49,398][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.4896269738674164, 'train/grad_norm_pre_clip': 27.861570358276367, 'timing/data': 0.0003839298151433468, 'timing/model': 0.8904188158921897, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.01}) +[2026-06-24 09:53:50,294][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.5780382752418518, 'train/grad_norm_pre_clip': 62.58814239501953, 'timing/data': 0.0004235748201608658, 'timing/model': 0.8941061759833246, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.01}) +[2026-06-24 09:53:51,186][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.5819230675697327, 'train/grad_norm_pre_clip': 60.976802825927734, 'timing/data': 0.0004352061077952385, 'timing/model': 0.8905189870856702, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.01}) +[2026-06-24 09:53:52,074][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.4856613874435425, 'train/grad_norm_pre_clip': 30.995336532592773, 'timing/data': 0.000432430999353528, 'timing/model': 0.8861946249380708, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.01}) +[2026-06-24 09:53:52,966][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.5136336088180542, 'train/grad_norm_pre_clip': 22.83701515197754, 'timing/data': 0.00039868801832199097, 'timing/model': 0.8891457030549645, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.01}) +[2026-06-24 09:53:53,859][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.509414553642273, 'train/grad_norm_pre_clip': 56.50351333618164, 'timing/data': 0.0003604970406740904, 'timing/model': 0.8910720588173717, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:53:54,751][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.7079429626464844, 'train/grad_norm_pre_clip': 77.07089233398438, 'timing/data': 0.0005277199670672417, 'timing/model': 0.8903984560165554, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:53:55,642][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.6739847660064697, 'train/grad_norm_pre_clip': 74.1055679321289, 'timing/data': 0.00039510102942585945, 'timing/model': 0.8887704028747976, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.01}) +[2026-06-24 09:53:56,532][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.3883773982524872, 'train/grad_norm_pre_clip': 21.879281997680664, 'timing/data': 0.0003989089746028185, 'timing/model': 0.8877021400257945, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.01}) +[2026-06-24 09:53:57,424][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.5404253602027893, 'train/grad_norm_pre_clip': 55.6778450012207, 'timing/data': 0.0004340650048106909, 'timing/model': 0.8907363538164645, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:53:58,318][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.6502044200897217, 'train/grad_norm_pre_clip': 62.374385833740234, 'timing/data': 0.0003641829825937748, 'timing/model': 0.8919054688885808, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.01}) +[2026-06-24 09:53:59,211][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.4965032637119293, 'train/grad_norm_pre_clip': 52.692832946777344, 'timing/data': 0.0003622190561145544, 'timing/model': 0.8904432961717248, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.01}) +[2026-06-24 09:54:00,105][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.34631502628326416, 'train/grad_norm_pre_clip': 32.5958366394043, 'timing/data': 0.0003909929655492306, 'timing/model': 0.8926501187961549, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.01}) +[2026-06-24 09:54:00,998][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.4907495081424713, 'train/grad_norm_pre_clip': 59.14219284057617, 'timing/data': 0.00036217900924384594, 'timing/model': 0.8910236980300397, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.01}) +[2026-06-24 09:54:01,897][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.4446811378002167, 'train/grad_norm_pre_clip': 77.10199737548828, 'timing/data': 0.0004621571861207485, 'timing/model': 0.8971651019528508, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.01}) +[2026-06-24 09:54:02,791][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.35636088252067566, 'train/grad_norm_pre_clip': 43.33030319213867, 'timing/data': 0.00040744501166045666, 'timing/model': 0.8914691768586636, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.01}) +[2026-06-24 09:54:03,686][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.3796464204788208, 'train/grad_norm_pre_clip': 43.58475875854492, 'timing/data': 0.0003814848605543375, 'timing/model': 0.8930863221175969, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:54:04,580][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.5249289274215698, 'train/grad_norm_pre_clip': 93.34419250488281, 'timing/data': 0.0006132400594651699, 'timing/model': 0.891853230074048, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.01}) +[2026-06-24 09:54:05,475][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.7541741728782654, 'train/grad_norm_pre_clip': 73.46907806396484, 'timing/data': 0.00045223813503980637, 'timing/model': 0.8920368349645287, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.01}) +[2026-06-24 09:54:06,369][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.6550895571708679, 'train/grad_norm_pre_clip': 58.25751876831055, 'timing/data': 0.0003747830633074045, 'timing/model': 0.8918935579713434, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:54:07,262][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.3275550603866577, 'train/grad_norm_pre_clip': 24.820343017578125, 'timing/data': 0.0003960628528147936, 'timing/model': 0.8912063199095428, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:54:08,159][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.4469996392726898, 'train/grad_norm_pre_clip': 26.598976135253906, 'timing/data': 0.0003620490897446871, 'timing/model': 0.8953235868830234, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.01}) +[2026-06-24 09:54:09,056][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.5711224675178528, 'train/grad_norm_pre_clip': 65.46298217773438, 'timing/data': 0.00037883082404732704, 'timing/model': 0.895279694814235, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.01}) +[2026-06-24 09:54:09,952][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.5109358429908752, 'train/grad_norm_pre_clip': 35.20279312133789, 'timing/data': 0.00039804703556001186, 'timing/model': 0.8935616631060839, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.01}) +[2026-06-24 09:54:10,848][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.42901378870010376, 'train/grad_norm_pre_clip': 23.54269790649414, 'timing/data': 0.00037016510032117367, 'timing/model': 0.8935411360580474, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.01}) +[2026-06-24 09:54:11,747][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.5506227612495422, 'train/grad_norm_pre_clip': 31.420446395874023, 'timing/data': 0.0004976040218025446, 'timing/model': 0.8976921760477126, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.01}) +[2026-06-24 09:54:12,639][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.4399058222770691, 'train/grad_norm_pre_clip': 51.52230453491211, 'timing/data': 0.0004106599371880293, 'timing/model': 0.8896397580392659, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.01}) +[2026-06-24 09:54:13,534][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.5236071944236755, 'train/grad_norm_pre_clip': 31.2828369140625, 'timing/data': 0.0003834199160337448, 'timing/model': 0.8930890518240631, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.01}) +[2026-06-24 09:54:14,431][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.4339183568954468, 'train/grad_norm_pre_clip': 48.81953811645508, 'timing/data': 0.000394569942727685, 'timing/model': 0.8948473620694131, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.01}) +[2026-06-24 09:54:15,329][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.666771650314331, 'train/grad_norm_pre_clip': 54.11692428588867, 'timing/data': 0.000431339954957366, 'timing/model': 0.8967338518705219, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:54:16,229][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.4894556701183319, 'train/grad_norm_pre_clip': 39.80066680908203, 'timing/data': 0.0003936178982257843, 'timing/model': 0.8976028140168637, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:54:17,124][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.519032895565033, 'train/grad_norm_pre_clip': 28.39556121826172, 'timing/data': 0.0003719879314303398, 'timing/model': 0.8935329800006002, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.01}) +[2026-06-24 09:54:18,022][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.5562658905982971, 'train/grad_norm_pre_clip': 80.10022735595703, 'timing/data': 0.0003763060085475445, 'timing/model': 0.8960790170822293, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.01}) +[2026-06-24 09:54:18,920][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.3579498529434204, 'train/grad_norm_pre_clip': 22.25022315979004, 'timing/data': 0.00047861807979643345, 'timing/model': 0.895246968138963, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:54:19,817][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.38411691784858704, 'train/grad_norm_pre_clip': 22.820964813232422, 'timing/data': 0.00040529994294047356, 'timing/model': 0.8950364838819951, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:54:20,713][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.3156818747520447, 'train/grad_norm_pre_clip': 24.133831024169922, 'timing/data': 0.00039872806519269943, 'timing/model': 0.8943780809640884, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.01}) +[2026-06-24 09:54:21,609][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.735027015209198, 'train/grad_norm_pre_clip': 140.83079528808594, 'timing/data': 0.000428744126111269, 'timing/model': 0.8943891529925168, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.01}) +[2026-06-24 09:54:22,508][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.29244720935821533, 'train/grad_norm_pre_clip': 127.88106536865234, 'timing/data': 0.0004424802027642727, 'timing/model': 0.8968364458996803, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.01}) +[2026-06-24 09:54:23,404][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.4023382365703583, 'train/grad_norm_pre_clip': 40.13397979736328, 'timing/data': 0.00038273795507848263, 'timing/model': 0.8933041989803314, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.01}) +[2026-06-24 09:54:24,299][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.632114589214325, 'train/grad_norm_pre_clip': 60.704505920410156, 'timing/data': 0.00036533595994114876, 'timing/model': 0.89345050486736, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.01}) +[2026-06-24 09:54:25,193][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.4290980398654938, 'train/grad_norm_pre_clip': 57.928016662597656, 'timing/data': 0.0005190540105104446, 'timing/model': 0.8924991509411484, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.01}) +[2026-06-24 09:54:26,091][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.46365758776664734, 'train/grad_norm_pre_clip': 38.5427131652832, 'timing/data': 0.00043078791350126266, 'timing/model': 0.8957649199292064, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.01}) +[2026-06-24 09:54:26,989][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.4609396755695343, 'train/grad_norm_pre_clip': 48.85068893432617, 'timing/data': 0.00040942803025245667, 'timing/model': 0.8957479300443083, 'learning_rate/qwen_vl_interface': 9.999997212805213e-06, 'learning_rate/action_model': 9.999997212805213e-05, 'epoch': 0.01}) +[2026-06-24 09:54:27,885][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.5518694519996643, 'train/grad_norm_pre_clip': 81.58341217041016, 'timing/data': 0.00039264699444174767, 'timing/model': 0.893753268988803, 'learning_rate/qwen_vl_interface': 9.999988851224124e-06, 'learning_rate/action_model': 9.999988851224123e-05, 'epoch': 0.01}) +[2026-06-24 09:54:28,780][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.526074230670929, 'train/grad_norm_pre_clip': 67.55854034423828, 'timing/data': 0.0004180539399385452, 'timing/model': 0.893077532062307, 'learning_rate/qwen_vl_interface': 9.999974915266542e-06, 'learning_rate/action_model': 9.999974915266542e-05, 'epoch': 0.01}) +[2026-06-24 09:54:29,678][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.37653446197509766, 'train/grad_norm_pre_clip': 28.557109832763672, 'timing/data': 0.00039268704131245613, 'timing/model': 0.8953849428799003, 'learning_rate/qwen_vl_interface': 9.999955404948825e-06, 'learning_rate/action_model': 9.999955404948824e-05, 'epoch': 0.01}) +[2026-06-24 09:54:30,574][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.563261091709137, 'train/grad_norm_pre_clip': 81.14250183105469, 'timing/data': 0.00039130495861172676, 'timing/model': 0.8949517011642456, 'learning_rate/qwen_vl_interface': 9.999930320293868e-06, 'learning_rate/action_model': 9.999930320293867e-05, 'epoch': 0.01}) +[2026-06-24 09:54:31,470][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.40565750002861023, 'train/grad_norm_pre_clip': 41.6209716796875, 'timing/data': 0.0003638719208538532, 'timing/model': 0.8939782162196934, 'learning_rate/qwen_vl_interface': 9.999899661331109e-06, 'learning_rate/action_model': 9.99989966133111e-05, 'epoch': 0.01}) +[2026-06-24 09:54:32,368][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.6000891327857971, 'train/grad_norm_pre_clip': 76.63470458984375, 'timing/data': 0.0004965220578014851, 'timing/model': 0.896032891003415, 'learning_rate/qwen_vl_interface': 9.999863428096531e-06, 'learning_rate/action_model': 9.99986342809653e-05, 'epoch': 0.01}) +[2026-06-24 09:54:33,269][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.5079785585403442, 'train/grad_norm_pre_clip': 71.37297058105469, 'timing/data': 0.0003688719589263201, 'timing/model': 0.8985844301059842, 'learning_rate/qwen_vl_interface': 9.999821620632651e-06, 'learning_rate/action_model': 9.99982162063265e-05, 'epoch': 0.01}) +[2026-06-24 09:54:34,164][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.3377031087875366, 'train/grad_norm_pre_clip': 37.92936325073242, 'timing/data': 0.00041252397932112217, 'timing/model': 0.8934471630491316, 'learning_rate/qwen_vl_interface': 9.999774238988534e-06, 'learning_rate/action_model': 9.999774238988534e-05, 'epoch': 0.01}) +[2026-06-24 09:54:35,061][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.4306776523590088, 'train/grad_norm_pre_clip': 57.78202819824219, 'timing/data': 0.0003867759369313717, 'timing/model': 0.8944843499921262, 'learning_rate/qwen_vl_interface': 9.999721283219787e-06, 'learning_rate/action_model': 9.999721283219787e-05, 'epoch': 0.01}) +[2026-06-24 09:54:35,961][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.36128929257392883, 'train/grad_norm_pre_clip': 22.095691680908203, 'timing/data': 0.0007646139711141586, 'timing/model': 0.8972469139844179, 'learning_rate/qwen_vl_interface': 9.999662753388557e-06, 'learning_rate/action_model': 9.999662753388557e-05, 'epoch': 0.01}) +[2026-06-24 09:54:36,855][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.3483334481716156, 'train/grad_norm_pre_clip': 48.117530822753906, 'timing/data': 0.0004233440849930048, 'timing/model': 0.8922806470654905, 'learning_rate/qwen_vl_interface': 9.999598649563527e-06, 'learning_rate/action_model': 9.999598649563527e-05, 'epoch': 0.01}) +[2026-06-24 09:54:37,752][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.5213726162910461, 'train/grad_norm_pre_clip': 56.786895751953125, 'timing/data': 0.00034085894003510475, 'timing/model': 0.8952859069686383, 'learning_rate/qwen_vl_interface': 9.999528971819933e-06, 'learning_rate/action_model': 9.999528971819932e-05, 'epoch': 0.01}) +[2026-06-24 09:54:38,644][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.34734997153282166, 'train/grad_norm_pre_clip': 48.519901275634766, 'timing/data': 0.000403095968067646, 'timing/model': 0.8908439248334616, 'learning_rate/qwen_vl_interface': 9.999453720239541e-06, 'learning_rate/action_model': 9.99945372023954e-05, 'epoch': 0.01}) +[2026-06-24 09:54:39,544][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.5425747632980347, 'train/grad_norm_pre_clip': 35.63370132446289, 'timing/data': 0.0004504951648414135, 'timing/model': 0.8973409559112042, 'learning_rate/qwen_vl_interface': 9.999372894910664e-06, 'learning_rate/action_model': 9.999372894910665e-05, 'epoch': 0.01}) +[2026-06-24 09:54:40,443][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.2975809872150421, 'train/grad_norm_pre_clip': 40.60603713989258, 'timing/data': 0.0004984249826520681, 'timing/model': 0.8969193049706519, 'learning_rate/qwen_vl_interface': 9.999286495928156e-06, 'learning_rate/action_model': 9.999286495928157e-05, 'epoch': 0.01}) +[2026-06-24 09:54:41,343][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.46737605333328247, 'train/grad_norm_pre_clip': 54.77776336669922, 'timing/data': 0.0003905319608747959, 'timing/model': 0.8988496288657188, 'learning_rate/qwen_vl_interface': 9.999194523393413e-06, 'learning_rate/action_model': 9.999194523393412e-05, 'epoch': 0.01}) +[2026-06-24 09:54:42,245][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.40599724650382996, 'train/grad_norm_pre_clip': 57.275146484375, 'timing/data': 0.00041299499571323395, 'timing/model': 0.899474652018398, 'learning_rate/qwen_vl_interface': 9.999096977414362e-06, 'learning_rate/action_model': 9.999096977414362e-05, 'epoch': 0.01}) +[2026-06-24 09:54:43,148][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.38418665528297424, 'train/grad_norm_pre_clip': 21.815988540649414, 'timing/data': 0.00042009796015918255, 'timing/model': 0.9011756158433855, 'learning_rate/qwen_vl_interface': 9.998993858105487e-06, 'learning_rate/action_model': 9.998993858105487e-05, 'epoch': 0.01}) +[2026-06-24 09:54:44,049][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.34552136063575745, 'train/grad_norm_pre_clip': 27.1495304107666, 'timing/data': 0.0003699741791933775, 'timing/model': 0.8996078160125762, 'learning_rate/qwen_vl_interface': 9.9988851655878e-06, 'learning_rate/action_model': 9.9988851655878e-05, 'epoch': 0.01}) +[2026-06-24 09:54:44,952][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.3038314878940582, 'train/grad_norm_pre_clip': 35.06125259399414, 'timing/data': 0.00036547589115798473, 'timing/model': 0.9008750538341701, 'learning_rate/qwen_vl_interface': 9.99877089998886e-06, 'learning_rate/action_model': 9.99877089998886e-05, 'epoch': 0.01}) +[2026-06-24 09:54:45,850][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.40902358293533325, 'train/grad_norm_pre_clip': 189.63949584960938, 'timing/data': 0.0003877580165863037, 'timing/model': 0.8955374411307275, 'learning_rate/qwen_vl_interface': 9.998651061442763e-06, 'learning_rate/action_model': 9.998651061442761e-05, 'epoch': 0.01}) +[2026-06-24 09:54:46,751][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.38016965985298157, 'train/grad_norm_pre_clip': 68.13301849365234, 'timing/data': 0.0004657341632992029, 'timing/model': 0.8994959688279778, 'learning_rate/qwen_vl_interface': 9.998525650090144e-06, 'learning_rate/action_model': 9.998525650090145e-05, 'epoch': 0.01}) +[2026-06-24 09:54:47,648][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.45658084750175476, 'train/grad_norm_pre_clip': 113.41878509521484, 'timing/data': 0.0003660670481622219, 'timing/model': 0.8954282188788056, 'learning_rate/qwen_vl_interface': 9.998394666078184e-06, 'learning_rate/action_model': 9.998394666078185e-05, 'epoch': 0.01}) +[2026-06-24 09:54:48,551][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.4974755644798279, 'train/grad_norm_pre_clip': 88.14956665039062, 'timing/data': 0.00038332887925207615, 'timing/model': 0.9007344758138061, 'learning_rate/qwen_vl_interface': 9.9982581095606e-06, 'learning_rate/action_model': 9.998258109560599e-05, 'epoch': 0.01}) +[2026-06-24 09:54:49,449][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.3456396758556366, 'train/grad_norm_pre_clip': 47.487701416015625, 'timing/data': 0.0005210880190134048, 'timing/model': 0.896162117831409, 'learning_rate/qwen_vl_interface': 9.998115980697646e-06, 'learning_rate/action_model': 9.998115980697645e-05, 'epoch': 0.01}) +[2026-06-24 09:54:50,353][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.4652459919452667, 'train/grad_norm_pre_clip': 98.20100402832031, 'timing/data': 0.0003780289553105831, 'timing/model': 0.9013823228888214, 'learning_rate/qwen_vl_interface': 9.997968279656119e-06, 'learning_rate/action_model': 9.997968279656118e-05, 'epoch': 0.01}) +[2026-06-24 09:54:51,252][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.4257047474384308, 'train/grad_norm_pre_clip': 39.92689895629883, 'timing/data': 0.00038429093547165394, 'timing/model': 0.8971339329145849, 'learning_rate/qwen_vl_interface': 9.997815006609354e-06, 'learning_rate/action_model': 9.997815006609353e-05, 'epoch': 0.01}) +[2026-06-24 09:54:52,151][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.41969144344329834, 'train/grad_norm_pre_clip': 47.83689498901367, 'timing/data': 0.0003868560306727886, 'timing/model': 0.8971782280132174, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.01}) +[2026-06-24 09:54:53,049][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.39784160256385803, 'train/grad_norm_pre_clip': 73.16365814208984, 'timing/data': 0.0003722978290170431, 'timing/model': 0.895933813881129, 'learning_rate/qwen_vl_interface': 9.997491745226147e-06, 'learning_rate/action_model': 9.997491745226148e-05, 'epoch': 0.01}) +[2026-06-24 09:54:53,951][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.2754652798175812, 'train/grad_norm_pre_clip': 18.36534881591797, 'timing/data': 0.00034305290319025517, 'timing/model': 0.9005329238716513, 'learning_rate/qwen_vl_interface': 9.997321757269074e-06, 'learning_rate/action_model': 9.997321757269074e-05, 'epoch': 0.01}) +[2026-06-24 09:54:54,848][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.452891081571579, 'train/grad_norm_pre_clip': 45.99250411987305, 'timing/data': 0.00040043098852038383, 'timing/model': 0.8949168880935758, 'learning_rate/qwen_vl_interface': 9.997146198065492e-06, 'learning_rate/action_model': 9.997146198065491e-05, 'epoch': 0.01}) +[2026-06-24 09:54:55,748][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.3419792056083679, 'train/grad_norm_pre_clip': 26.18607521057129, 'timing/data': 0.00039483094587922096, 'timing/model': 0.8986042679753155, 'learning_rate/qwen_vl_interface': 9.99696506782143e-06, 'learning_rate/action_model': 9.99696506782143e-05, 'epoch': 0.01}) +[2026-06-24 09:54:56,651][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.4002980589866638, 'train/grad_norm_pre_clip': 36.0904426574707, 'timing/data': 0.0003629110287874937, 'timing/model': 0.9006337579339743, 'learning_rate/qwen_vl_interface': 9.996778366749458e-06, 'learning_rate/action_model': 9.996778366749457e-05, 'epoch': 0.01}) +[2026-06-24 09:54:57,552][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.3103441894054413, 'train/grad_norm_pre_clip': 33.583839416503906, 'timing/data': 0.0004147780127823353, 'timing/model': 0.8989663401152939, 'learning_rate/qwen_vl_interface': 9.996586095068676e-06, 'learning_rate/action_model': 9.996586095068676e-05, 'epoch': 0.01}) +[2026-06-24 09:54:58,450][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.461232990026474, 'train/grad_norm_pre_clip': 67.18123626708984, 'timing/data': 0.00038312911055982113, 'timing/model': 0.8963342520873994, 'learning_rate/qwen_vl_interface': 9.996388253004727e-06, 'learning_rate/action_model': 9.996388253004726e-05, 'epoch': 0.01}) +[2026-06-24 09:54:59,352][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.3360665440559387, 'train/grad_norm_pre_clip': 54.235984802246094, 'timing/data': 0.0003953310661017895, 'timing/model': 0.9004679890349507, 'learning_rate/qwen_vl_interface': 9.996184840789791e-06, 'learning_rate/action_model': 9.99618484078979e-05, 'epoch': 0.01}) +[2026-06-24 09:55:00,248][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.23502272367477417, 'train/grad_norm_pre_clip': 31.861217498779297, 'timing/data': 0.0003782100975513458, 'timing/model': 0.8940777210518718, 'learning_rate/qwen_vl_interface': 9.995975858662581e-06, 'learning_rate/action_model': 9.995975858662582e-05, 'epoch': 0.01}) +[2026-06-24 09:55:01,146][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.7102358937263489, 'train/grad_norm_pre_clip': 112.23011016845703, 'timing/data': 0.0004611250478774309, 'timing/model': 0.8952702700626105, 'learning_rate/qwen_vl_interface': 9.995761306868353e-06, 'learning_rate/action_model': 9.995761306868353e-05, 'epoch': 0.01}) +[2026-06-24 09:55:02,044][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.503882884979248, 'train/grad_norm_pre_clip': 43.26385498046875, 'timing/data': 0.00037432205863296986, 'timing/model': 0.8966382280923426, 'learning_rate/qwen_vl_interface': 9.99554118565889e-06, 'learning_rate/action_model': 9.99554118565889e-05, 'epoch': 0.01}) +[2026-06-24 09:55:02,945][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.3518446087837219, 'train/grad_norm_pre_clip': 91.90032196044922, 'timing/data': 0.0003661171067506075, 'timing/model': 0.8988550081849098, 'learning_rate/qwen_vl_interface': 9.995315495292521e-06, 'learning_rate/action_model': 9.995315495292522e-05, 'epoch': 0.01}) +[2026-06-24 09:55:03,847][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.50778728723526, 'train/grad_norm_pre_clip': 57.61894989013672, 'timing/data': 0.0003440950531512499, 'timing/model': 0.8998254709877074, 'learning_rate/qwen_vl_interface': 9.995084236034104e-06, 'learning_rate/action_model': 9.995084236034104e-05, 'epoch': 0.01}) +[2026-06-24 09:55:04,745][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.354646772146225, 'train/grad_norm_pre_clip': 33.24296188354492, 'timing/data': 0.00041532889008522034, 'timing/model': 0.8961475810501724, 'learning_rate/qwen_vl_interface': 9.994847408155038e-06, 'learning_rate/action_model': 9.994847408155038e-05, 'epoch': 0.01}) +[2026-06-24 09:55:05,642][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.4737973213195801, 'train/grad_norm_pre_clip': 38.322025299072266, 'timing/data': 0.0005639779847115278, 'timing/model': 0.894530416931957, 'learning_rate/qwen_vl_interface': 9.994605011933249e-06, 'learning_rate/action_model': 9.99460501193325e-05, 'epoch': 0.01}) +[2026-06-24 09:55:06,542][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.33056601881980896, 'train/grad_norm_pre_clip': 29.16061782836914, 'timing/data': 0.0003938591107726097, 'timing/model': 0.8985926010645926, 'learning_rate/qwen_vl_interface': 9.994357047653206e-06, 'learning_rate/action_model': 9.994357047653205e-05, 'epoch': 0.01}) +[2026-06-24 09:55:07,444][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.4743707776069641, 'train/grad_norm_pre_clip': 35.65903854370117, 'timing/data': 0.0005489399190992117, 'timing/model': 0.8991560889407992, 'learning_rate/qwen_vl_interface': 9.994103515605907e-06, 'learning_rate/action_model': 9.994103515605906e-05, 'epoch': 0.01}) +[2026-06-24 09:55:08,344][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.36680713295936584, 'train/grad_norm_pre_clip': 30.206829071044922, 'timing/data': 0.00039580301381647587, 'timing/model': 0.8989749699831009, 'learning_rate/qwen_vl_interface': 9.993844416088887e-06, 'learning_rate/action_model': 9.993844416088887e-05, 'epoch': 0.01}) +[2026-06-24 09:55:09,244][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.3688896596431732, 'train/grad_norm_pre_clip': 33.15042495727539, 'timing/data': 0.0003659969661384821, 'timing/model': 0.8978077219799161, 'learning_rate/qwen_vl_interface': 9.993579749406213e-06, 'learning_rate/action_model': 9.993579749406212e-05, 'epoch': 0.01}) +[2026-06-24 09:55:10,151][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.637444257736206, 'train/grad_norm_pre_clip': 69.51104736328125, 'timing/data': 0.00038334913551807404, 'timing/model': 0.9050246139522642, 'learning_rate/qwen_vl_interface': 9.993309515868485e-06, 'learning_rate/action_model': 9.993309515868484e-05, 'epoch': 0.01}) +[2026-06-24 09:55:11,046][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.4979938566684723, 'train/grad_norm_pre_clip': 62.33148193359375, 'timing/data': 0.0003603759687393904, 'timing/model': 0.8932186129968613, 'learning_rate/qwen_vl_interface': 9.993033715792838e-06, 'learning_rate/action_model': 9.993033715792838e-05, 'epoch': 0.01}) +[2026-06-24 09:55:11,950][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.6359231472015381, 'train/grad_norm_pre_clip': 103.82044219970703, 'timing/data': 0.00041633215732872486, 'timing/model': 0.9017928619869053, 'learning_rate/qwen_vl_interface': 9.992752349502942e-06, 'learning_rate/action_model': 9.992752349502942e-05, 'epoch': 0.01}) +[2026-06-24 09:55:12,847][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.3324025571346283, 'train/grad_norm_pre_clip': 44.22201156616211, 'timing/data': 0.0003728000447154045, 'timing/model': 0.8951008869335055, 'learning_rate/qwen_vl_interface': 9.992465417328992e-06, 'learning_rate/action_model': 9.99246541732899e-05, 'epoch': 0.01}) +[2026-06-24 09:55:13,749][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.464153528213501, 'train/grad_norm_pre_clip': 69.57495880126953, 'timing/data': 0.00040252506732940674, 'timing/model': 0.8999375540297478, 'learning_rate/qwen_vl_interface': 9.992172919607719e-06, 'learning_rate/action_model': 9.992172919607719e-05, 'epoch': 0.01}) +[2026-06-24 09:55:14,647][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.2974143624305725, 'train/grad_norm_pre_clip': 50.674171447753906, 'timing/data': 0.00032995897345244884, 'timing/model': 0.8961045918986201, 'learning_rate/qwen_vl_interface': 9.991874856682388e-06, 'learning_rate/action_model': 9.991874856682387e-05, 'epoch': 0.01}) +[2026-06-24 09:55:15,546][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.41020774841308594, 'train/grad_norm_pre_clip': 27.221874237060547, 'timing/data': 0.0004127349238842726, 'timing/model': 0.8967019750270993, 'learning_rate/qwen_vl_interface': 9.991571228902788e-06, 'learning_rate/action_model': 9.991571228902788e-05, 'epoch': 0.01}) +[2026-06-24 09:55:16,446][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.45152950286865234, 'train/grad_norm_pre_clip': 87.26190948486328, 'timing/data': 0.0004490718711167574, 'timing/model': 0.8988080881536007, 'learning_rate/qwen_vl_interface': 9.991262036625249e-06, 'learning_rate/action_model': 9.991262036625248e-05, 'epoch': 0.02}) +[2026-06-24 09:55:17,345][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.587723433971405, 'train/grad_norm_pre_clip': 94.6122817993164, 'timing/data': 0.00039886892773211, 'timing/model': 0.8968028160743415, 'learning_rate/qwen_vl_interface': 9.990947280212622e-06, 'learning_rate/action_model': 9.990947280212621e-05, 'epoch': 0.02}) +[2026-06-24 09:55:18,242][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.703546941280365, 'train/grad_norm_pre_clip': 132.07913208007812, 'timing/data': 0.0004236949607729912, 'timing/model': 0.895040380069986, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.02}) +[2026-06-24 09:55:19,142][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.3046627342700958, 'train/grad_norm_pre_clip': 35.46016311645508, 'timing/data': 0.0003776291850954294, 'timing/model': 0.8977195089682937, 'learning_rate/qwen_vl_interface': 9.99030107646617e-06, 'learning_rate/action_model': 9.99030107646617e-05, 'epoch': 0.02}) +[2026-06-24 09:55:20,041][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.3669448494911194, 'train/grad_norm_pre_clip': 36.71761703491211, 'timing/data': 0.00038913008756935596, 'timing/model': 0.8975504029076546, 'learning_rate/qwen_vl_interface': 9.9899696298907e-06, 'learning_rate/action_model': 9.989969629890701e-05, 'epoch': 0.02}) +[2026-06-24 09:55:20,945][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.29556480050086975, 'train/grad_norm_pre_clip': 33.72175598144531, 'timing/data': 0.00033750408329069614, 'timing/model': 0.9014683251734823, 'learning_rate/qwen_vl_interface': 9.989632620696858e-06, 'learning_rate/action_model': 9.989632620696858e-05, 'epoch': 0.02}) +[2026-06-24 09:55:21,844][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.3257277309894562, 'train/grad_norm_pre_clip': 108.35149383544922, 'timing/data': 0.0003737509250640869, 'timing/model': 0.8978111438918859, 'learning_rate/qwen_vl_interface': 9.989290049280135e-06, 'learning_rate/action_model': 9.989290049280135e-05, 'epoch': 0.02}) +[2026-06-24 09:55:22,747][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.6478782296180725, 'train/grad_norm_pre_clip': 147.7804412841797, 'timing/data': 0.0004004421643912792, 'timing/model': 0.9010738569777459, 'learning_rate/qwen_vl_interface': 9.988941916042562e-06, 'learning_rate/action_model': 9.988941916042561e-05, 'epoch': 0.02}) +[2026-06-24 09:55:23,650][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.5815719366073608, 'train/grad_norm_pre_clip': 128.48890686035156, 'timing/data': 0.00043025705963373184, 'timing/model': 0.900270088808611, 'learning_rate/qwen_vl_interface': 9.988588221392693e-06, 'learning_rate/action_model': 9.988588221392692e-05, 'epoch': 0.02}) +[2026-06-24 09:55:24,556][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.27143219113349915, 'train/grad_norm_pre_clip': 52.44590377807617, 'timing/data': 0.00043597794137895107, 'timing/model': 0.9040928739123046, 'learning_rate/qwen_vl_interface': 9.988228965745608e-06, 'learning_rate/action_model': 9.988228965745608e-05, 'epoch': 0.02}) +[2026-06-24 09:55:25,459][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.5090686082839966, 'train/grad_norm_pre_clip': 56.5571174621582, 'timing/data': 0.00047919899225234985, 'timing/model': 0.90155074512586, 'learning_rate/qwen_vl_interface': 9.98786414952291e-06, 'learning_rate/action_model': 9.987864149522911e-05, 'epoch': 0.02}) +[2026-06-24 09:55:26,360][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.51191246509552, 'train/grad_norm_pre_clip': 88.86316680908203, 'timing/data': 0.0004673968069255352, 'timing/model': 0.8986591100692749, 'learning_rate/qwen_vl_interface': 9.987493773152736e-06, 'learning_rate/action_model': 9.987493773152736e-05, 'epoch': 0.02}) +[2026-06-24 09:55:27,262][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.7373365759849548, 'train/grad_norm_pre_clip': 59.19108963012695, 'timing/data': 0.00039820699021220207, 'timing/model': 0.8997073578648269, 'learning_rate/qwen_vl_interface': 9.987117837069742e-06, 'learning_rate/action_model': 9.987117837069741e-05, 'epoch': 0.02}) +[2026-06-24 09:55:28,161][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.3715791702270508, 'train/grad_norm_pre_clip': 19.360607147216797, 'timing/data': 0.0004080049693584442, 'timing/model': 0.8966613451484591, 'learning_rate/qwen_vl_interface': 9.98673634171511e-06, 'learning_rate/action_model': 9.986736341715108e-05, 'epoch': 0.02}) +[2026-06-24 09:55:29,061][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.28677064180374146, 'train/grad_norm_pre_clip': 32.456268310546875, 'timing/data': 0.0003833589144051075, 'timing/model': 0.8974197290372103, 'learning_rate/qwen_vl_interface': 9.986349287536544e-06, 'learning_rate/action_model': 9.986349287536543e-05, 'epoch': 0.02}) +[2026-06-24 09:55:29,959][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.3878892958164215, 'train/grad_norm_pre_clip': 31.31952476501465, 'timing/data': 0.0004244060255587101, 'timing/model': 0.8969105749856681, 'learning_rate/qwen_vl_interface': 9.985956674988274e-06, 'learning_rate/action_model': 9.985956674988274e-05, 'epoch': 0.02}) +[2026-06-24 09:55:30,858][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.2944701611995697, 'train/grad_norm_pre_clip': 15.623950958251953, 'timing/data': 0.0003891908563673496, 'timing/model': 0.897048142971471, 'learning_rate/qwen_vl_interface': 9.985558504531057e-06, 'learning_rate/action_model': 9.985558504531056e-05, 'epoch': 0.02}) +[2026-06-24 09:55:31,763][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.310181200504303, 'train/grad_norm_pre_clip': 50.43757629394531, 'timing/data': 0.0004735891707241535, 'timing/model': 0.9024980908725411, 'learning_rate/qwen_vl_interface': 9.985154776632163e-06, 'learning_rate/action_model': 9.985154776632162e-05, 'epoch': 0.02}) +[2026-06-24 09:55:32,665][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.21717919409275055, 'train/grad_norm_pre_clip': 34.35302734375, 'timing/data': 0.00041097099892795086, 'timing/model': 0.9001767958980054, 'learning_rate/qwen_vl_interface': 9.98474549176539e-06, 'learning_rate/action_model': 9.984745491765389e-05, 'epoch': 0.02}) +[2026-06-24 09:55:33,563][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.48591890931129456, 'train/grad_norm_pre_clip': 50.04340744018555, 'timing/data': 0.0004197079688310623, 'timing/model': 0.8961846269667149, 'learning_rate/qwen_vl_interface': 9.98433065041106e-06, 'learning_rate/action_model': 9.984330650411058e-05, 'epoch': 0.02}) +[2026-06-24 09:55:34,470][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.41034501791000366, 'train/grad_norm_pre_clip': 64.86449432373047, 'timing/data': 0.0004131849855184555, 'timing/model': 0.9049383827950805, 'learning_rate/qwen_vl_interface': 9.983910253056008e-06, 'learning_rate/action_model': 9.983910253056007e-05, 'epoch': 0.02}) +[2026-06-24 09:55:35,370][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.3770124316215515, 'train/grad_norm_pre_clip': 59.75997543334961, 'timing/data': 0.0004300670698285103, 'timing/model': 0.8976341979578137, 'learning_rate/qwen_vl_interface': 9.983484300193597e-06, 'learning_rate/action_model': 9.983484300193596e-05, 'epoch': 0.02}) +[2026-06-24 09:55:36,271][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.38705208897590637, 'train/grad_norm_pre_clip': 58.40968322753906, 'timing/data': 0.00038893893361091614, 'timing/model': 0.8991577278356999, 'learning_rate/qwen_vl_interface': 9.983052792323704e-06, 'learning_rate/action_model': 9.983052792323704e-05, 'epoch': 0.02}) +[2026-06-24 09:55:37,172][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.39205649495124817, 'train/grad_norm_pre_clip': 57.56016159057617, 'timing/data': 0.0005104078445583582, 'timing/model': 0.8990242588333786, 'learning_rate/qwen_vl_interface': 9.982615729952732e-06, 'learning_rate/action_model': 9.982615729952732e-05, 'epoch': 0.02}) +[2026-06-24 09:55:38,072][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.25311461091041565, 'train/grad_norm_pre_clip': 35.4373779296875, 'timing/data': 0.0004065129905939102, 'timing/model': 0.8984006810933352, 'learning_rate/qwen_vl_interface': 9.982173113593591e-06, 'learning_rate/action_model': 9.982173113593591e-05, 'epoch': 0.02}) +[2026-06-24 09:55:38,974][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.5109442472457886, 'train/grad_norm_pre_clip': 123.10496520996094, 'timing/data': 0.0003919650334864855, 'timing/model': 0.9002431679982692, 'learning_rate/qwen_vl_interface': 9.981724943765723e-06, 'learning_rate/action_model': 9.981724943765723e-05, 'epoch': 0.02}) +[2026-06-24 09:55:39,877][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.3877299726009369, 'train/grad_norm_pre_clip': 91.20851135253906, 'timing/data': 0.0004034070298075676, 'timing/model': 0.900513666914776, 'learning_rate/qwen_vl_interface': 9.981271220995076e-06, 'learning_rate/action_model': 9.981271220995075e-05, 'epoch': 0.02}) +[2026-06-24 09:55:40,777][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.5067577362060547, 'train/grad_norm_pre_clip': 49.03539276123047, 'timing/data': 0.0004391439724713564, 'timing/model': 0.8978369061369449, 'learning_rate/qwen_vl_interface': 9.980811945814119e-06, 'learning_rate/action_model': 9.980811945814119e-05, 'epoch': 0.02}) +[2026-06-24 09:55:41,677][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.3370828926563263, 'train/grad_norm_pre_clip': 59.11333084106445, 'timing/data': 0.0003290679305791855, 'timing/model': 0.8978484990075231, 'learning_rate/qwen_vl_interface': 9.980347118761838e-06, 'learning_rate/action_model': 9.980347118761838e-05, 'epoch': 0.02}) +[2026-06-24 09:55:42,578][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.3595336079597473, 'train/grad_norm_pre_clip': 25.604503631591797, 'timing/data': 0.0005352338775992393, 'timing/model': 0.8991821119561791, 'learning_rate/qwen_vl_interface': 9.979876740383735e-06, 'learning_rate/action_model': 9.979876740383734e-05, 'epoch': 0.02}) +[2026-06-24 09:55:43,484][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.32165923714637756, 'train/grad_norm_pre_clip': 19.10057830810547, 'timing/data': 0.0003951210528612137, 'timing/model': 0.90424491208978, 'learning_rate/qwen_vl_interface': 9.979400811231822e-06, 'learning_rate/action_model': 9.979400811231822e-05, 'epoch': 0.02}) +[2026-06-24 09:55:44,393][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.35018911957740784, 'train/grad_norm_pre_clip': 69.55350494384766, 'timing/data': 0.0003630609717220068, 'timing/model': 0.9067864120006561, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.02}) +[2026-06-24 09:55:45,295][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.24030187726020813, 'train/grad_norm_pre_clip': 18.554807662963867, 'timing/data': 0.00040212483145296574, 'timing/model': 0.900589644908905, 'learning_rate/qwen_vl_interface': 9.978432302847203e-06, 'learning_rate/action_model': 9.978432302847203e-05, 'epoch': 0.02}) +[2026-06-24 09:55:46,204][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.32169657945632935, 'train/grad_norm_pre_clip': 27.192184448242188, 'timing/data': 0.0004246770404279232, 'timing/model': 0.9067433748859912, 'learning_rate/qwen_vl_interface': 9.977939724751097e-06, 'learning_rate/action_model': 9.977939724751096e-05, 'epoch': 0.02}) +[2026-06-24 09:55:47,108][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.2085094004869461, 'train/grad_norm_pre_clip': 22.20807456970215, 'timing/data': 0.0003789819311350584, 'timing/model': 0.9019717720802873, 'learning_rate/qwen_vl_interface': 9.977441598154375e-06, 'learning_rate/action_model': 9.977441598154375e-05, 'epoch': 0.02}) +[2026-06-24 09:55:48,009][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.8745041489601135, 'train/grad_norm_pre_clip': 163.97251892089844, 'timing/data': 0.0004578488878905773, 'timing/model': 0.8991194118279964, 'learning_rate/qwen_vl_interface': 9.976937923641621e-06, 'learning_rate/action_model': 9.976937923641621e-05, 'epoch': 0.02}) +[2026-06-24 09:55:48,911][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.5520622134208679, 'train/grad_norm_pre_clip': 152.93511962890625, 'timing/data': 0.00044625718146562576, 'timing/model': 0.8998698401264846, 'learning_rate/qwen_vl_interface': 9.976428701803925e-06, 'learning_rate/action_model': 9.976428701803925e-05, 'epoch': 0.02}) +[2026-06-24 09:55:49,814][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.37832462787628174, 'train/grad_norm_pre_clip': 69.38711547851562, 'timing/data': 0.00038337893784046173, 'timing/model': 0.9007610629778355, 'learning_rate/qwen_vl_interface': 9.975913933238885e-06, 'learning_rate/action_model': 9.975913933238884e-05, 'epoch': 0.02}) +[2026-06-24 09:55:50,719][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.275567889213562, 'train/grad_norm_pre_clip': 32.82420349121094, 'timing/data': 0.0003865649923682213, 'timing/model': 0.9031705160159618, 'learning_rate/qwen_vl_interface': 9.975393618550611e-06, 'learning_rate/action_model': 9.975393618550612e-05, 'epoch': 0.02}) +[2026-06-24 09:55:51,619][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.34255704283714294, 'train/grad_norm_pre_clip': 52.28248977661133, 'timing/data': 0.0003684009425342083, 'timing/model': 0.8981869760900736, 'learning_rate/qwen_vl_interface': 9.974867758349722e-06, 'learning_rate/action_model': 9.974867758349721e-05, 'epoch': 0.02}) +[2026-06-24 09:55:52,523][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.5495254993438721, 'train/grad_norm_pre_clip': 98.39877319335938, 'timing/data': 0.00040983897633850574, 'timing/model': 0.9017815520055592, 'learning_rate/qwen_vl_interface': 9.974336353253346e-06, 'learning_rate/action_model': 9.974336353253345e-05, 'epoch': 0.02}) +[2026-06-24 09:55:53,424][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.4638960063457489, 'train/grad_norm_pre_clip': 53.225852966308594, 'timing/data': 0.00040711392648518085, 'timing/model': 0.8993315459229052, 'learning_rate/qwen_vl_interface': 9.973799403885113e-06, 'learning_rate/action_model': 9.973799403885112e-05, 'epoch': 0.02}) +[2026-06-24 09:55:54,326][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.36716827750205994, 'train/grad_norm_pre_clip': 31.374210357666016, 'timing/data': 0.0004744911566376686, 'timing/model': 0.9003275150898844, 'learning_rate/qwen_vl_interface': 9.973256910875163e-06, 'learning_rate/action_model': 9.973256910875163e-05, 'epoch': 0.02}) +[2026-06-24 09:55:55,230][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.39648038148880005, 'train/grad_norm_pre_clip': 81.85078430175781, 'timing/data': 0.0003943301271647215, 'timing/model': 0.9024129689205438, 'learning_rate/qwen_vl_interface': 9.972708874860145e-06, 'learning_rate/action_model': 9.972708874860145e-05, 'epoch': 0.02}) +[2026-06-24 09:55:56,135][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.24716432392597198, 'train/grad_norm_pre_clip': 44.2139778137207, 'timing/data': 0.00041154189966619015, 'timing/model': 0.9029298110399395, 'learning_rate/qwen_vl_interface': 9.97215529648321e-06, 'learning_rate/action_model': 9.972155296483209e-05, 'epoch': 0.02}) +[2026-06-24 09:55:57,045][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.2829354405403137, 'train/grad_norm_pre_clip': 30.66565704345703, 'timing/data': 0.0005821720696985722, 'timing/model': 0.9070431201253086, 'learning_rate/qwen_vl_interface': 9.971596176394008e-06, 'learning_rate/action_model': 9.971596176394008e-05, 'epoch': 0.02}) +[2026-06-24 09:55:57,945][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.3345046639442444, 'train/grad_norm_pre_clip': 50.15434265136719, 'timing/data': 0.00047405902296304703, 'timing/model': 0.8987968279980123, 'learning_rate/qwen_vl_interface': 9.971031515248702e-06, 'learning_rate/action_model': 9.971031515248701e-05, 'epoch': 0.02}) +[2026-06-24 09:55:58,847][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.3556475043296814, 'train/grad_norm_pre_clip': 52.702491760253906, 'timing/data': 0.0004302470479160547, 'timing/model': 0.900077611906454, 'learning_rate/qwen_vl_interface': 9.970461313709952e-06, 'learning_rate/action_model': 9.970461313709952e-05, 'epoch': 0.02}) +[2026-06-24 09:55:59,751][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.46507537364959717, 'train/grad_norm_pre_clip': 57.11085510253906, 'timing/data': 0.00037965294905006886, 'timing/model': 0.9012160392012447, 'learning_rate/qwen_vl_interface': 9.96988557244692e-06, 'learning_rate/action_model': 9.969885572446921e-05, 'epoch': 0.02}) +[2026-06-24 09:56:00,656][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.27651548385620117, 'train/grad_norm_pre_clip': 26.340682983398438, 'timing/data': 0.00038783811032772064, 'timing/model': 0.9034057890530676, 'learning_rate/qwen_vl_interface': 9.969304292135274e-06, 'learning_rate/action_model': 9.969304292135275e-05, 'epoch': 0.02}) +[2026-06-24 09:56:01,559][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.20208877325057983, 'train/grad_norm_pre_clip': 26.529687881469727, 'timing/data': 0.00044409302063286304, 'timing/model': 0.9007954620756209, 'learning_rate/qwen_vl_interface': 9.968717473457175e-06, 'learning_rate/action_model': 9.968717473457175e-05, 'epoch': 0.02}) +[2026-06-24 09:56:02,459][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.2793302834033966, 'train/grad_norm_pre_clip': 35.04384994506836, 'timing/data': 0.0003974058199673891, 'timing/model': 0.8983282439876348, 'learning_rate/qwen_vl_interface': 9.96812511710129e-06, 'learning_rate/action_model': 9.96812511710129e-05, 'epoch': 0.02}) +[2026-06-24 09:56:03,364][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.2829339802265167, 'train/grad_norm_pre_clip': 48.27212905883789, 'timing/data': 0.000507972901687026, 'timing/model': 0.9030581298284233, 'learning_rate/qwen_vl_interface': 9.967527223762782e-06, 'learning_rate/action_model': 9.967527223762781e-05, 'epoch': 0.02}) +[2026-06-24 09:56:04,270][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.2046533226966858, 'train/grad_norm_pre_clip': 46.51673889160156, 'timing/data': 0.0004417991731315851, 'timing/model': 0.903719421941787, 'learning_rate/qwen_vl_interface': 9.966923794143308e-06, 'learning_rate/action_model': 9.966923794143307e-05, 'epoch': 0.02}) +[2026-06-24 09:56:05,171][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.24333305656909943, 'train/grad_norm_pre_clip': 51.58393859863281, 'timing/data': 0.00043373415246605873, 'timing/model': 0.8987747649662197, 'learning_rate/qwen_vl_interface': 9.966314828951033e-06, 'learning_rate/action_model': 9.966314828951032e-05, 'epoch': 0.02}) +[2026-06-24 09:56:06,073][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.24087978899478912, 'train/grad_norm_pre_clip': 57.56400680541992, 'timing/data': 0.00042164092883467674, 'timing/model': 0.9007078630384058, 'learning_rate/qwen_vl_interface': 9.965700328900607e-06, 'learning_rate/action_model': 9.965700328900606e-05, 'epoch': 0.02}) +[2026-06-24 09:56:06,975][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.5389391183853149, 'train/grad_norm_pre_clip': 114.47088623046875, 'timing/data': 0.0004210399929434061, 'timing/model': 0.8990737670101225, 'learning_rate/qwen_vl_interface': 9.96508029471318e-06, 'learning_rate/action_model': 9.96508029471318e-05, 'epoch': 0.02}) +[2026-06-24 09:56:07,877][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.5752019882202148, 'train/grad_norm_pre_clip': 78.66875457763672, 'timing/data': 0.0004015041049569845, 'timing/model': 0.9003961500711739, 'learning_rate/qwen_vl_interface': 9.964454727116399e-06, 'learning_rate/action_model': 9.964454727116398e-05, 'epoch': 0.02}) +[2026-06-24 09:56:08,779][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.4606521725654602, 'train/grad_norm_pre_clip': 98.82682037353516, 'timing/data': 0.00038988213054835796, 'timing/model': 0.8998048319481313, 'learning_rate/qwen_vl_interface': 9.9638236268444e-06, 'learning_rate/action_model': 9.9638236268444e-05, 'epoch': 0.02}) +[2026-06-24 09:56:09,683][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.22273513674736023, 'train/grad_norm_pre_clip': 28.819154739379883, 'timing/data': 0.0005679160822182894, 'timing/model': 0.9017366680782288, 'learning_rate/qwen_vl_interface': 9.963186994637816e-06, 'learning_rate/action_model': 9.963186994637815e-05, 'epoch': 0.02}) +[2026-06-24 09:56:10,585][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.37927454710006714, 'train/grad_norm_pre_clip': 65.59954833984375, 'timing/data': 0.0005214789416640997, 'timing/model': 0.9001188729889691, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.02}) +[2026-06-24 09:56:11,489][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.5685517191886902, 'train/grad_norm_pre_clip': 114.05762481689453, 'timing/data': 0.0004856809973716736, 'timing/model': 0.9022261979989707, 'learning_rate/qwen_vl_interface': 9.961897137415878e-06, 'learning_rate/action_model': 9.961897137415877e-05, 'epoch': 0.02}) +[2026-06-24 09:56:12,394][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.5970918536186218, 'train/grad_norm_pre_clip': 118.10638427734375, 'timing/data': 0.00042004790157079697, 'timing/model': 0.9027160990517586, 'learning_rate/qwen_vl_interface': 9.96124391391424e-06, 'learning_rate/action_model': 9.96124391391424e-05, 'epoch': 0.02}) +[2026-06-24 09:56:13,299][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.2880432605743408, 'train/grad_norm_pre_clip': 27.031980514526367, 'timing/data': 0.0004926437977701426, 'timing/model': 0.9030973049812019, 'learning_rate/qwen_vl_interface': 9.960585161505456e-06, 'learning_rate/action_model': 9.960585161505456e-05, 'epoch': 0.02}) +[2026-06-24 09:56:14,205][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.3955491781234741, 'train/grad_norm_pre_clip': 52.72625732421875, 'timing/data': 0.0004516979679465294, 'timing/model': 0.9041789851617068, 'learning_rate/qwen_vl_interface': 9.959920880962606e-06, 'learning_rate/action_model': 9.959920880962605e-05, 'epoch': 0.02}) +[2026-06-24 09:56:15,110][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.3967575132846832, 'train/grad_norm_pre_clip': 42.27153015136719, 'timing/data': 0.00038340012542903423, 'timing/model': 0.9036383419297636, 'learning_rate/qwen_vl_interface': 9.95925107306526e-06, 'learning_rate/action_model': 9.959251073065259e-05, 'epoch': 0.02}) +[2026-06-24 09:56:16,014][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.23247554898262024, 'train/grad_norm_pre_clip': 31.306446075439453, 'timing/data': 0.00041415682062506676, 'timing/model': 0.9013345709536225, 'learning_rate/qwen_vl_interface': 9.958575738599474e-06, 'learning_rate/action_model': 9.958575738599474e-05, 'epoch': 0.02}) +[2026-06-24 09:56:16,919][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.22119702398777008, 'train/grad_norm_pre_clip': 23.225826263427734, 'timing/data': 0.00040829600766301155, 'timing/model': 0.9031354698818177, 'learning_rate/qwen_vl_interface': 9.957894878357791e-06, 'learning_rate/action_model': 9.957894878357791e-05, 'epoch': 0.02}) +[2026-06-24 09:56:17,824][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.26628702878952026, 'train/grad_norm_pre_clip': 16.612606048583984, 'timing/data': 0.00040577189065515995, 'timing/model': 0.9032859951257706, 'learning_rate/qwen_vl_interface': 9.957208493139242e-06, 'learning_rate/action_model': 9.957208493139242e-05, 'epoch': 0.02}) +[2026-06-24 09:56:18,727][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.30699852108955383, 'train/grad_norm_pre_clip': 36.54936599731445, 'timing/data': 0.00047198589891195297, 'timing/model': 0.9003468200098723, 'learning_rate/qwen_vl_interface': 9.956516583749333e-06, 'learning_rate/action_model': 9.956516583749333e-05, 'epoch': 0.02}) +[2026-06-24 09:56:19,636][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.34388744831085205, 'train/grad_norm_pre_clip': 53.25886535644531, 'timing/data': 0.0003891510423272848, 'timing/model': 0.9070068481378257, 'learning_rate/qwen_vl_interface': 9.955819151000063e-06, 'learning_rate/action_model': 9.955819151000062e-05, 'epoch': 0.02}) +[2026-06-24 09:56:20,547][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.3872661292552948, 'train/grad_norm_pre_clip': 42.60511016845703, 'timing/data': 0.000573284924030304, 'timing/model': 0.9081189059652388, 'learning_rate/qwen_vl_interface': 9.955116195709901e-06, 'learning_rate/action_model': 9.955116195709901e-05, 'epoch': 0.02}) +[2026-06-24 09:56:21,445][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.38413193821907043, 'train/grad_norm_pre_clip': 46.4089241027832, 'timing/data': 0.00043277209624648094, 'timing/model': 0.8966753939166665, 'learning_rate/qwen_vl_interface': 9.954407718703813e-06, 'learning_rate/action_model': 9.954407718703813e-05, 'epoch': 0.02}) +[2026-06-24 09:56:22,351][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.4514157772064209, 'train/grad_norm_pre_clip': 59.602046966552734, 'timing/data': 0.00046949111856520176, 'timing/model': 0.9035472499672323, 'learning_rate/qwen_vl_interface': 9.953693720813232e-06, 'learning_rate/action_model': 9.953693720813231e-05, 'epoch': 0.02}) +[2026-06-24 09:56:23,254][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.48162078857421875, 'train/grad_norm_pre_clip': 91.81879425048828, 'timing/data': 0.0004390839021652937, 'timing/model': 0.901622629025951, 'learning_rate/qwen_vl_interface': 9.952974202876075e-06, 'learning_rate/action_model': 9.952974202876074e-05, 'epoch': 0.02}) +[2026-06-24 09:56:24,160][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.4199938178062439, 'train/grad_norm_pre_clip': 49.26321792602539, 'timing/data': 0.00039391894824802876, 'timing/model': 0.9038078698795289, 'learning_rate/qwen_vl_interface': 9.952249165736735e-06, 'learning_rate/action_model': 9.952249165736734e-05, 'epoch': 0.02}) +[2026-06-24 09:56:25,066][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.18981227278709412, 'train/grad_norm_pre_clip': 45.1427116394043, 'timing/data': 0.00038661505095660686, 'timing/model': 0.9044377030804753, 'learning_rate/qwen_vl_interface': 9.951518610246083e-06, 'learning_rate/action_model': 9.951518610246083e-05, 'epoch': 0.02}) +[2026-06-24 09:56:25,967][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.3240637481212616, 'train/grad_norm_pre_clip': 41.013858795166016, 'timing/data': 0.00043018697760999203, 'timing/model': 0.8984478048514575, 'learning_rate/qwen_vl_interface': 9.950782537261469e-06, 'learning_rate/action_model': 9.950782537261469e-05, 'epoch': 0.02}) +[2026-06-24 09:56:26,870][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.27036964893341064, 'train/grad_norm_pre_clip': 53.050838470458984, 'timing/data': 0.0003992491401731968, 'timing/model': 0.9017008398659527, 'learning_rate/qwen_vl_interface': 9.950040947646716e-06, 'learning_rate/action_model': 9.950040947646716e-05, 'epoch': 0.02}) +[2026-06-24 09:56:27,774][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.3577616810798645, 'train/grad_norm_pre_clip': 35.0496940612793, 'timing/data': 0.00044017587788403034, 'timing/model': 0.9012466378044337, 'learning_rate/qwen_vl_interface': 9.949293842272116e-06, 'learning_rate/action_model': 9.949293842272117e-05, 'epoch': 0.02}) +[2026-06-24 09:56:28,675][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.2527315020561218, 'train/grad_norm_pre_clip': 28.127613067626953, 'timing/data': 0.0004992368631064892, 'timing/model': 0.899605690035969, 'learning_rate/qwen_vl_interface': 9.948541222014446e-06, 'learning_rate/action_model': 9.948541222014445e-05, 'epoch': 0.02}) +[2026-06-24 09:56:29,579][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.16570599377155304, 'train/grad_norm_pre_clip': 24.344175338745117, 'timing/data': 0.0003771069459617138, 'timing/model': 0.901501287939027, 'learning_rate/qwen_vl_interface': 9.94778308775694e-06, 'learning_rate/action_model': 9.947783087756941e-05, 'epoch': 0.02}) +[2026-06-24 09:56:30,481][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.4894246459007263, 'train/grad_norm_pre_clip': 70.63421630859375, 'timing/data': 0.00041075004264712334, 'timing/model': 0.8998657800257206, 'learning_rate/qwen_vl_interface': 9.947019440389318e-06, 'learning_rate/action_model': 9.947019440389317e-05, 'epoch': 0.02}) +[2026-06-24 09:56:31,385][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.30101311206817627, 'train/grad_norm_pre_clip': 68.39543914794922, 'timing/data': 0.00038119591772556305, 'timing/model': 0.9024497400969267, 'learning_rate/qwen_vl_interface': 9.946250280807756e-06, 'learning_rate/action_model': 9.946250280807757e-05, 'epoch': 0.02}) +[2026-06-24 09:56:32,287][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.2840438187122345, 'train/grad_norm_pre_clip': 58.78504943847656, 'timing/data': 0.00040840613655745983, 'timing/model': 0.899436237057671, 'learning_rate/qwen_vl_interface': 9.945475609914913e-06, 'learning_rate/action_model': 9.945475609914912e-05, 'epoch': 0.02}) +[2026-06-24 09:56:33,186][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.22867387533187866, 'train/grad_norm_pre_clip': 39.520042419433594, 'timing/data': 0.0004690710920840502, 'timing/model': 0.8979592761024833, 'learning_rate/qwen_vl_interface': 9.9446954286199e-06, 'learning_rate/action_model': 9.944695428619901e-05, 'epoch': 0.02}) +[2026-06-24 09:56:34,088][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.2553798258304596, 'train/grad_norm_pre_clip': 37.87474822998047, 'timing/data': 0.00034048897214233875, 'timing/model': 0.9000141930300742, 'learning_rate/qwen_vl_interface': 9.943909737838311e-06, 'learning_rate/action_model': 9.94390973783831e-05, 'epoch': 0.02}) +[2026-06-24 09:56:34,996][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.2256515920162201, 'train/grad_norm_pre_clip': 49.44743347167969, 'timing/data': 0.0003968949895352125, 'timing/model': 0.9056661101058125, 'learning_rate/qwen_vl_interface': 9.943118538492193e-06, 'learning_rate/action_model': 9.943118538492192e-05, 'epoch': 0.02}) +[2026-06-24 09:56:35,904][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.27372294664382935, 'train/grad_norm_pre_clip': 68.10601806640625, 'timing/data': 0.00039696390740573406, 'timing/model': 0.9058922349940985, 'learning_rate/qwen_vl_interface': 9.942321831510066e-06, 'learning_rate/action_model': 9.942321831510066e-05, 'epoch': 0.02}) +[2026-06-24 09:56:36,808][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.29857349395751953, 'train/grad_norm_pre_clip': 59.003665924072266, 'timing/data': 0.0003896821290254593, 'timing/model': 0.9025627479422837, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.02}) +[2026-06-24 09:56:37,727][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.4902925491333008, 'train/grad_norm_pre_clip': 119.04678344726562, 'timing/data': 0.00037517398595809937, 'timing/model': 0.9161281839478761, 'learning_rate/qwen_vl_interface': 9.940711898384158e-06, 'learning_rate/action_model': 9.940711898384157e-05, 'epoch': 0.02}) +[2026-06-24 09:56:38,639][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.30427253246307373, 'train/grad_norm_pre_clip': 90.17292785644531, 'timing/data': 0.0004043090157210827, 'timing/model': 0.9104350919369608, 'learning_rate/qwen_vl_interface': 9.939898674129723e-06, 'learning_rate/action_model': 9.939898674129724e-05, 'epoch': 0.02}) +[2026-06-24 09:56:39,544][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.3964598774909973, 'train/grad_norm_pre_clip': 76.09246063232422, 'timing/data': 0.00043876306153833866, 'timing/model': 0.9029146621469408, 'learning_rate/qwen_vl_interface': 9.93907994601797e-06, 'learning_rate/action_model': 9.939079946017969e-05, 'epoch': 0.02}) +[2026-06-24 09:56:40,444][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.25657230615615845, 'train/grad_norm_pre_clip': 45.525814056396484, 'timing/data': 0.00043967505916953087, 'timing/model': 0.8983834499958903, 'learning_rate/qwen_vl_interface': 9.93825571500972e-06, 'learning_rate/action_model': 9.93825571500972e-05, 'epoch': 0.02}) +[2026-06-24 09:56:41,346][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 250 +[2026-06-24 09:57:26,701][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 250 +[2026-06-24 09:57:26,702][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 250 +[2026-06-24 09:57:27,438][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmp7dox9mu7/test.c -o /tmp/tmp7dox9mu7/test.o +[2026-06-24 09:57:27,453][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmp7dox9mu7/test.o -laio -o /tmp/tmp7dox9mu7/a.out +[2026-06-24 09:57:28,218][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmp8k9wz1g9/test.c -o /tmp/tmp8k9wz1g9/test.o +[2026-06-24 09:57:28,234][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmp8k9wz1g9/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp8k9wz1g9/a.out +[2026-06-24 09:57:28,274][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmpvma_zz6e/test.c -o /tmp/tmpvma_zz6e/test.o +[2026-06-24 09:57:28,289][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmpvma_zz6e/test.o -laio -o /tmp/tmpvma_zz6e/a.out +[2026-06-24 09:58:22,540][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 250 +[2026-06-24 09:59:15,328][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 250 +[2026-06-24 09:59:15,329][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.228384867310524, 'train/grad_norm_pre_clip': 44.95318603515625, 'eval/action_loss/seconds': 45.352913056034595, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.13920949399471283, 'eval/latency_2/loss': 0.13920949399471283, 'eval/flappy/loss': 0.13920949399471283, 'eval/flappy/latency_2/loss': 0.13920949399471283, 'eval/flap_precision': 0.7982779827798278, 'eval/flap_recall': 0.8805970149253731, 'eval/flap_f1': 0.8374193548387097, 'eval/flap_rate_model': 0.22583333333333333, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8374193548387096, 'eval/latency_2/flap_precision': 0.7982779827798278, 'eval/latency_2/flap_recall': 0.8805970149253731, 'eval/latency_2/flap_f1': 0.8374193548387097, 'eval/latency_2/flap_rate_model': 0.22583333333333333, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8374193548387096, 'eval/action_classification/data_wait_seconds': 24.01776695251465, 'eval/action_classification/predict_seconds': 29.72704315185547, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 37.10000000000025, 'rl_games_eval/mid_train/mean_length': 319.0, 'rl_games_eval/mid_train/std_reward': 23.528960878033065, 'rl_games_eval/mid_train/std_length': 189.75668631170814, 'rl_games_eval/mid_train/macro_mean_reward': 37.10000000000025, 'rl_games_eval/mid_train/macro_mean_length': 319.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 37.10000000000025, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 319.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 23.528960878033065, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 189.75668631170814, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 37.10000000000025, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00042891385965049267, 'timing/model': 0.9004790128674358, 'learning_rate/qwen_vl_interface': 9.937425982072251e-06, 'learning_rate/action_model': 9.93742598207225e-05, 'epoch': 0.02}) +[2026-06-24 09:59:16,397][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.38031360507011414, 'train/grad_norm_pre_clip': 53.03569412231445, 'timing/data': 0.0003661068622022867, 'timing/model': 1.0656476218719035, 'learning_rate/qwen_vl_interface': 9.936590748179303e-06, 'learning_rate/action_model': 9.936590748179303e-05, 'epoch': 0.02}) +[2026-06-24 09:59:17,459][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.2968517541885376, 'train/grad_norm_pre_clip': 59.23333740234375, 'timing/data': 0.0004440341144800186, 'timing/model': 1.0607876530848444, 'learning_rate/qwen_vl_interface': 9.93575001431107e-06, 'learning_rate/action_model': 9.935750014311069e-05, 'epoch': 0.02}) +[2026-06-24 09:59:18,519][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.24002757668495178, 'train/grad_norm_pre_clip': 44.24106979370117, 'timing/data': 0.0005225311033427715, 'timing/model': 1.0582386509049684, 'learning_rate/qwen_vl_interface': 9.9349037814542e-06, 'learning_rate/action_model': 9.934903781454198e-05, 'epoch': 0.02}) +[2026-06-24 09:59:19,584][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.29067420959472656, 'train/grad_norm_pre_clip': 41.24409484863281, 'timing/data': 0.00038076494820415974, 'timing/model': 1.0624689711257815, 'learning_rate/qwen_vl_interface': 9.934052050601793e-06, 'learning_rate/action_model': 9.934052050601792e-05, 'epoch': 0.02}) +[2026-06-24 09:59:20,649][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.1921808421611786, 'train/grad_norm_pre_clip': 35.45622253417969, 'timing/data': 0.00042506796307861805, 'timing/model': 1.0631079499144107, 'learning_rate/qwen_vl_interface': 9.933194822753405e-06, 'learning_rate/action_model': 9.933194822753405e-05, 'epoch': 0.02}) +[2026-06-24 09:59:21,715][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.3327981233596802, 'train/grad_norm_pre_clip': 63.3570442199707, 'timing/data': 0.0003807549364864826, 'timing/model': 1.0641724490560591, 'learning_rate/qwen_vl_interface': 9.932332098915038e-06, 'learning_rate/action_model': 9.932332098915036e-05, 'epoch': 0.02}) +[2026-06-24 09:59:22,776][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.2580786645412445, 'train/grad_norm_pre_clip': 27.480337142944336, 'timing/data': 0.000377067131921649, 'timing/model': 1.0600984040647745, 'learning_rate/qwen_vl_interface': 9.931463880099149e-06, 'learning_rate/action_model': 9.931463880099148e-05, 'epoch': 0.02}) +[2026-06-24 09:59:23,845][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.34643375873565674, 'train/grad_norm_pre_clip': 37.12769317626953, 'timing/data': 0.00046184612438082695, 'timing/model': 1.066294953925535, 'learning_rate/qwen_vl_interface': 9.930590167324639e-06, 'learning_rate/action_model': 9.930590167324639e-05, 'epoch': 0.02}) +[2026-06-24 09:59:24,915][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.27377989888191223, 'train/grad_norm_pre_clip': 29.265748977661133, 'timing/data': 0.0003802841529250145, 'timing/model': 1.068537593120709, 'learning_rate/qwen_vl_interface': 9.929710961616859e-06, 'learning_rate/action_model': 9.929710961616858e-05, 'epoch': 0.03}) +[2026-06-24 09:59:25,984][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.29953181743621826, 'train/grad_norm_pre_clip': 74.20492553710938, 'timing/data': 0.00038582412526011467, 'timing/model': 1.0673847009893507, 'learning_rate/qwen_vl_interface': 9.928826264007606e-06, 'learning_rate/action_model': 9.928826264007606e-05, 'epoch': 0.03}) +[2026-06-24 09:59:27,050][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.33217886090278625, 'train/grad_norm_pre_clip': 75.89678955078125, 'timing/data': 0.0004409679677337408, 'timing/model': 1.0638486160896719, 'learning_rate/qwen_vl_interface': 9.927936075535122e-06, 'learning_rate/action_model': 9.927936075535122e-05, 'epoch': 0.03}) +[2026-06-24 09:59:28,116][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.31835800409317017, 'train/grad_norm_pre_clip': 52.995079040527344, 'timing/data': 0.0004374510608613491, 'timing/model': 1.0647298200055957, 'learning_rate/qwen_vl_interface': 9.927040397244093e-06, 'learning_rate/action_model': 9.927040397244092e-05, 'epoch': 0.03}) +[2026-06-24 09:59:29,491][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.3943665325641632, 'train/grad_norm_pre_clip': 90.2513427734375, 'timing/data': 0.00039701489731669426, 'timing/model': 1.3730784989893436, 'learning_rate/qwen_vl_interface': 9.92613923018565e-06, 'learning_rate/action_model': 9.92613923018565e-05, 'epoch': 0.03}) +[2026-06-24 09:59:30,560][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.480156272649765, 'train/grad_norm_pre_clip': 71.17950439453125, 'timing/data': 0.00043475604616105556, 'timing/model': 1.0667282820213586, 'learning_rate/qwen_vl_interface': 9.925232575417357e-06, 'learning_rate/action_model': 9.925232575417357e-05, 'epoch': 0.03}) +[2026-06-24 09:59:31,629][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.2154974639415741, 'train/grad_norm_pre_clip': 47.35001754760742, 'timing/data': 0.000430447980761528, 'timing/model': 1.0671994781587273, 'learning_rate/qwen_vl_interface': 9.924320434003228e-06, 'learning_rate/action_model': 9.924320434003228e-05, 'epoch': 0.03}) +[2026-06-24 09:59:32,694][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.24943816661834717, 'train/grad_norm_pre_clip': 19.37992286682129, 'timing/data': 0.0004956100601702929, 'timing/model': 1.062995880143717, 'learning_rate/qwen_vl_interface': 9.92340280701371e-06, 'learning_rate/action_model': 9.92340280701371e-05, 'epoch': 0.03}) +[2026-06-24 09:59:33,763][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.3640967309474945, 'train/grad_norm_pre_clip': 28.782161712646484, 'timing/data': 0.0003861950244754553, 'timing/model': 1.067259673960507, 'learning_rate/qwen_vl_interface': 9.922479695525692e-06, 'learning_rate/action_model': 9.92247969552569e-05, 'epoch': 0.03}) +[2026-06-24 09:59:34,834][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.2557719051837921, 'train/grad_norm_pre_clip': 34.15088653564453, 'timing/data': 0.00041217287071049213, 'timing/model': 1.068979673087597, 'learning_rate/qwen_vl_interface': 9.921551100622492e-06, 'learning_rate/action_model': 9.921551100622493e-05, 'epoch': 0.03}) +[2026-06-24 09:59:35,902][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.2925030589103699, 'train/grad_norm_pre_clip': 40.181461334228516, 'timing/data': 0.0003933680709451437, 'timing/model': 1.0659642049577087, 'learning_rate/qwen_vl_interface': 9.920617023393872e-06, 'learning_rate/action_model': 9.920617023393872e-05, 'epoch': 0.03}) +[2026-06-24 09:59:36,971][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.2516004145145416, 'train/grad_norm_pre_clip': 28.002239227294922, 'timing/data': 0.0003888588398694992, 'timing/model': 1.0671424260362983, 'learning_rate/qwen_vl_interface': 9.919677464936024e-06, 'learning_rate/action_model': 9.919677464936023e-05, 'epoch': 0.03}) +[2026-06-24 09:59:38,043][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.28003689646720886, 'train/grad_norm_pre_clip': 57.0592041015625, 'timing/data': 0.0003871761728078127, 'timing/model': 1.070597381098196, 'learning_rate/qwen_vl_interface': 9.918732426351569e-06, 'learning_rate/action_model': 9.918732426351568e-05, 'epoch': 0.03}) +[2026-06-24 09:59:39,111][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.35139989852905273, 'train/grad_norm_pre_clip': 40.347896575927734, 'timing/data': 0.0004451051354408264, 'timing/model': 1.0660050041042268, 'learning_rate/qwen_vl_interface': 9.917781908749563e-06, 'learning_rate/action_model': 9.917781908749563e-05, 'epoch': 0.03}) +[2026-06-24 09:59:40,181][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.43395376205444336, 'train/grad_norm_pre_clip': 98.50914764404297, 'timing/data': 0.00043623801320791245, 'timing/model': 1.0677063381299376, 'learning_rate/qwen_vl_interface': 9.916825913245494e-06, 'learning_rate/action_model': 9.916825913245494e-05, 'epoch': 0.03}) +[2026-06-24 09:59:41,252][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.3940022587776184, 'train/grad_norm_pre_clip': 52.9099235534668, 'timing/data': 0.0005439708475023508, 'timing/model': 1.069143646163866, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.03}) +[2026-06-24 09:59:42,323][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.14676204323768616, 'train/grad_norm_pre_clip': 24.076709747314453, 'timing/data': 0.00037841987796127796, 'timing/model': 1.068964321166277, 'learning_rate/qwen_vl_interface': 9.914897493025241e-06, 'learning_rate/action_model': 9.914897493025241e-05, 'epoch': 0.03}) +[2026-06-24 09:59:43,396][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.350082665681839, 'train/grad_norm_pre_clip': 69.67552185058594, 'timing/data': 0.0003533728886395693, 'timing/model': 1.071286167949438, 'learning_rate/qwen_vl_interface': 9.91392507057217e-06, 'learning_rate/action_model': 9.913925070572169e-05, 'epoch': 0.03}) +[2026-06-24 09:59:44,767][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.19017720222473145, 'train/grad_norm_pre_clip': 87.85594177246094, 'timing/data': 0.0003958628512918949, 'timing/model': 1.3695140378549695, 'learning_rate/qwen_vl_interface': 9.912947174743246e-06, 'learning_rate/action_model': 9.912947174743245e-05, 'epoch': 0.03}) +[2026-06-24 09:59:45,836][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.08597615361213684, 'train/grad_norm_pre_clip': 31.667890548706055, 'timing/data': 0.0004016540478914976, 'timing/model': 1.0666939630173147, 'learning_rate/qwen_vl_interface': 9.91196380668609e-06, 'learning_rate/action_model': 9.911963806686089e-05, 'epoch': 0.03}) +[2026-06-24 09:59:46,905][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.12220784276723862, 'train/grad_norm_pre_clip': 17.438556671142578, 'timing/data': 0.0003871771041303873, 'timing/model': 1.0673199489247054, 'learning_rate/qwen_vl_interface': 9.910974967554732e-06, 'learning_rate/action_model': 9.910974967554732e-05, 'epoch': 0.03}) +[2026-06-24 09:59:47,973][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.13111332058906555, 'train/grad_norm_pre_clip': 27.80535888671875, 'timing/data': 0.000426469836384058, 'timing/model': 1.0660554259084165, 'learning_rate/qwen_vl_interface': 9.909980658509636e-06, 'learning_rate/action_model': 9.909980658509636e-05, 'epoch': 0.03}) +[2026-06-24 09:59:49,042][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.05577186122536659, 'train/grad_norm_pre_clip': 32.053749084472656, 'timing/data': 0.000399670097976923, 'timing/model': 1.0675886448007077, 'learning_rate/qwen_vl_interface': 9.908980880717679e-06, 'learning_rate/action_model': 9.908980880717679e-05, 'epoch': 0.03}) +[2026-06-24 09:59:50,110][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.2205519676208496, 'train/grad_norm_pre_clip': 78.35625457763672, 'timing/data': 0.00044418382458388805, 'timing/model': 1.0663648690097034, 'learning_rate/qwen_vl_interface': 9.907975635352152e-06, 'learning_rate/action_model': 9.907975635352151e-05, 'epoch': 0.03}) +[2026-06-24 09:59:51,182][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.6389585137367249, 'train/grad_norm_pre_clip': 184.397705078125, 'timing/data': 0.0004138569347560406, 'timing/model': 1.0696521801874042, 'learning_rate/qwen_vl_interface': 9.906964923592767e-06, 'learning_rate/action_model': 9.906964923592767e-05, 'epoch': 0.03}) +[2026-06-24 09:59:52,252][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.5976763963699341, 'train/grad_norm_pre_clip': 202.81019592285156, 'timing/data': 0.0004255289677530527, 'timing/model': 1.0677481971215457, 'learning_rate/qwen_vl_interface': 9.905948746625654e-06, 'learning_rate/action_model': 9.905948746625653e-05, 'epoch': 0.03}) +[2026-06-24 09:59:53,326][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.29773396253585815, 'train/grad_norm_pre_clip': 95.58735656738281, 'timing/data': 0.0003886499907821417, 'timing/model': 1.0725589361973107, 'learning_rate/qwen_vl_interface': 9.904927105643348e-06, 'learning_rate/action_model': 9.904927105643348e-05, 'epoch': 0.03}) +[2026-06-24 09:59:54,402][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.1873682588338852, 'train/grad_norm_pre_clip': 70.88856506347656, 'timing/data': 0.00036575598642230034, 'timing/model': 1.0739435949362814, 'learning_rate/qwen_vl_interface': 9.903900001844809e-06, 'learning_rate/action_model': 9.903900001844808e-05, 'epoch': 0.03}) +[2026-06-24 09:59:55,473][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.4023756682872772, 'train/grad_norm_pre_clip': 77.3388671875, 'timing/data': 0.00035693985410034657, 'timing/model': 1.068941097939387, 'learning_rate/qwen_vl_interface': 9.902867436435394e-06, 'learning_rate/action_model': 9.902867436435394e-05, 'epoch': 0.03}) +[2026-06-24 09:59:56,545][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.2665039002895355, 'train/grad_norm_pre_clip': 113.1108169555664, 'timing/data': 0.00036155804991722107, 'timing/model': 1.0707171130925417, 'learning_rate/qwen_vl_interface': 9.901829410626877e-06, 'learning_rate/action_model': 9.901829410626878e-05, 'epoch': 0.03}) +[2026-06-24 09:59:57,615][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.2680025100708008, 'train/grad_norm_pre_clip': 164.63291931152344, 'timing/data': 0.0004443940706551075, 'timing/model': 1.0679670919198543, 'learning_rate/qwen_vl_interface': 9.900785925637443e-06, 'learning_rate/action_model': 9.900785925637442e-05, 'epoch': 0.03}) +[2026-06-24 09:59:58,690][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.4362066388130188, 'train/grad_norm_pre_clip': 161.64776611328125, 'timing/data': 0.00038628396578133106, 'timing/model': 1.0733704729937017, 'learning_rate/qwen_vl_interface': 9.899736982691675e-06, 'learning_rate/action_model': 9.899736982691676e-05, 'epoch': 0.03}) +[2026-06-24 10:00:00,061][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.8526694178581238, 'train/grad_norm_pre_clip': 240.09901428222656, 'timing/data': 0.0004324109759181738, 'timing/model': 1.3688416529912502, 'learning_rate/qwen_vl_interface': 9.898682583020568e-06, 'learning_rate/action_model': 9.898682583020568e-05, 'epoch': 0.03}) +[2026-06-24 10:00:01,136][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.2834736704826355, 'train/grad_norm_pre_clip': 89.1247329711914, 'timing/data': 0.0007890309207141399, 'timing/model': 1.0724964758846909, 'learning_rate/qwen_vl_interface': 9.89762272786152e-06, 'learning_rate/action_model': 9.89762272786152e-05, 'epoch': 0.03}) +[2026-06-24 10:00:02,209][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.09564881771802902, 'train/grad_norm_pre_clip': 17.818944931030273, 'timing/data': 0.00041017006151378155, 'timing/model': 1.0713471290655434, 'learning_rate/qwen_vl_interface': 9.896557418458328e-06, 'learning_rate/action_model': 9.896557418458329e-05, 'epoch': 0.03}) +[2026-06-24 10:00:03,286][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.1762971132993698, 'train/grad_norm_pre_clip': 24.745159149169922, 'timing/data': 0.0004212509375065565, 'timing/model': 1.0748878861777484, 'learning_rate/qwen_vl_interface': 9.895486656061196e-06, 'learning_rate/action_model': 9.895486656061195e-05, 'epoch': 0.03}) +[2026-06-24 10:00:04,359][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.6047973036766052, 'train/grad_norm_pre_clip': 95.2187271118164, 'timing/data': 0.00042075011879205704, 'timing/model': 1.0714317609090358, 'learning_rate/qwen_vl_interface': 9.894410441926716e-06, 'learning_rate/action_model': 9.894410441926716e-05, 'epoch': 0.03}) +[2026-06-24 10:00:05,430][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.35267478227615356, 'train/grad_norm_pre_clip': 59.03715515136719, 'timing/data': 0.0003794820513576269, 'timing/model': 1.0696885699871927, 'learning_rate/qwen_vl_interface': 9.893328777317891e-06, 'learning_rate/action_model': 9.893328777317891e-05, 'epoch': 0.03}) +[2026-06-24 10:00:06,500][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.2635987401008606, 'train/grad_norm_pre_clip': 64.00041198730469, 'timing/data': 0.00036073685623705387, 'timing/model': 1.067571328021586, 'learning_rate/qwen_vl_interface': 9.892241663504113e-06, 'learning_rate/action_model': 9.892241663504112e-05, 'epoch': 0.03}) +[2026-06-24 10:00:07,573][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.22894519567489624, 'train/grad_norm_pre_clip': 29.269771575927734, 'timing/data': 0.0006095939315855503, 'timing/model': 1.0705022991169244, 'learning_rate/qwen_vl_interface': 9.89114910176117e-06, 'learning_rate/action_model': 9.891149101761169e-05, 'epoch': 0.03}) +[2026-06-24 10:00:08,644][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.28941279649734497, 'train/grad_norm_pre_clip': 50.068870544433594, 'timing/data': 0.00041938689537346363, 'timing/model': 1.0697667100466788, 'learning_rate/qwen_vl_interface': 9.890051093371245e-06, 'learning_rate/action_model': 9.890051093371246e-05, 'epoch': 0.03}) +[2026-06-24 10:00:09,716][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.46186402440071106, 'train/grad_norm_pre_clip': 124.43146514892578, 'timing/data': 0.0003824278246611357, 'timing/model': 1.0699179938528687, 'learning_rate/qwen_vl_interface': 9.888947639622911e-06, 'learning_rate/action_model': 9.888947639622911e-05, 'epoch': 0.03}) +[2026-06-24 10:00:10,789][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.30398645997047424, 'train/grad_norm_pre_clip': 52.131744384765625, 'timing/data': 0.00038680597208440304, 'timing/model': 1.07112002116628, 'learning_rate/qwen_vl_interface': 9.887838741811136e-06, 'learning_rate/action_model': 9.887838741811136e-05, 'epoch': 0.03}) +[2026-06-24 10:00:11,864][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.1479874849319458, 'train/grad_norm_pre_clip': 24.3096923828125, 'timing/data': 0.00036763004027307034, 'timing/model': 1.072687072912231, 'learning_rate/qwen_vl_interface': 9.886724401237268e-06, 'learning_rate/action_model': 9.886724401237269e-05, 'epoch': 0.03}) +[2026-06-24 10:00:12,937][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.1329536885023117, 'train/grad_norm_pre_clip': 32.543792724609375, 'timing/data': 0.00039167492650449276, 'timing/model': 1.0715493478346616, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.03}) +[2026-06-24 10:00:14,010][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.15375488996505737, 'train/grad_norm_pre_clip': 52.56813430786133, 'timing/data': 0.0004017339088022709, 'timing/model': 1.0713229938410223, 'learning_rate/qwen_vl_interface': 9.88447939704061e-06, 'learning_rate/action_model': 9.884479397040611e-05, 'epoch': 0.03}) +[2026-06-24 10:00:15,385][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.150898277759552, 'train/grad_norm_pre_clip': 35.96080017089844, 'timing/data': 0.0003938991576433182, 'timing/model': 1.373536200961098, 'learning_rate/qwen_vl_interface': 9.883348736052458e-06, 'learning_rate/action_model': 9.883348736052458e-05, 'epoch': 0.03}) +[2026-06-24 10:00:16,459][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.19632872939109802, 'train/grad_norm_pre_clip': 22.948165893554688, 'timing/data': 0.00037282006815075874, 'timing/model': 1.071685086004436, 'learning_rate/qwen_vl_interface': 9.882212637571487e-06, 'learning_rate/action_model': 9.882212637571486e-05, 'epoch': 0.03}) +[2026-06-24 10:00:17,535][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.4843946099281311, 'train/grad_norm_pre_clip': 68.41839599609375, 'timing/data': 0.0005226610228419304, 'timing/model': 1.0738426859024912, 'learning_rate/qwen_vl_interface': 9.881071102930971e-06, 'learning_rate/action_model': 9.881071102930971e-05, 'epoch': 0.03}) +[2026-06-24 10:00:18,610][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.4177517294883728, 'train/grad_norm_pre_clip': 137.34371948242188, 'timing/data': 0.0004056207835674286, 'timing/model': 1.0729807489551604, 'learning_rate/qwen_vl_interface': 9.879924133470567e-06, 'learning_rate/action_model': 9.879924133470567e-05, 'epoch': 0.03}) +[2026-06-24 10:00:19,688][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.5744532942771912, 'train/grad_norm_pre_clip': 193.6615447998047, 'timing/data': 0.0004145978018641472, 'timing/model': 1.0762087479233742, 'learning_rate/qwen_vl_interface': 9.878771730536305e-06, 'learning_rate/action_model': 9.878771730536304e-05, 'epoch': 0.03}) +[2026-06-24 10:00:20,763][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.4929313063621521, 'train/grad_norm_pre_clip': 157.2581787109375, 'timing/data': 0.0003862751182168722, 'timing/model': 1.0732264330144972, 'learning_rate/qwen_vl_interface': 9.877613895480597e-06, 'learning_rate/action_model': 9.877613895480596e-05, 'epoch': 0.03}) +[2026-06-24 10:00:21,840][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.33124807476997375, 'train/grad_norm_pre_clip': 84.3766098022461, 'timing/data': 0.00038305786438286304, 'timing/model': 1.0751016919966787, 'learning_rate/qwen_vl_interface': 9.876450629662224e-06, 'learning_rate/action_model': 9.876450629662223e-05, 'epoch': 0.03}) +[2026-06-24 10:00:22,915][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.43260619044303894, 'train/grad_norm_pre_clip': 63.27634811401367, 'timing/data': 0.0003879878204315901, 'timing/model': 1.0736739439889789, 'learning_rate/qwen_vl_interface': 9.875281934446344e-06, 'learning_rate/action_model': 9.875281934446344e-05, 'epoch': 0.03}) +[2026-06-24 10:00:23,990][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.37856337428092957, 'train/grad_norm_pre_clip': 48.21653366088867, 'timing/data': 0.00040197395719587803, 'timing/model': 1.0731717231683433, 'learning_rate/qwen_vl_interface': 9.874107811204487e-06, 'learning_rate/action_model': 9.874107811204486e-05, 'epoch': 0.03}) +[2026-06-24 10:00:25,070][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.15339958667755127, 'train/grad_norm_pre_clip': 44.96100997924805, 'timing/data': 0.00037960195913910866, 'timing/model': 1.0781297371722758, 'learning_rate/qwen_vl_interface': 9.872928261314551e-06, 'learning_rate/action_model': 9.872928261314551e-05, 'epoch': 0.03}) +[2026-06-24 10:00:26,145][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.26226791739463806, 'train/grad_norm_pre_clip': 36.64921951293945, 'timing/data': 0.000401112949475646, 'timing/model': 1.0725941201671958, 'learning_rate/qwen_vl_interface': 9.871743286160803e-06, 'learning_rate/action_model': 9.871743286160803e-05, 'epoch': 0.03}) +[2026-06-24 10:00:27,224][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.29040229320526123, 'train/grad_norm_pre_clip': 35.96274185180664, 'timing/data': 0.0004335939884185791, 'timing/model': 1.0768650961108506, 'learning_rate/qwen_vl_interface': 9.870552887133881e-06, 'learning_rate/action_model': 9.870552887133881e-05, 'epoch': 0.03}) +[2026-06-24 10:00:28,301][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.32224041223526, 'train/grad_norm_pre_clip': 31.923439025878906, 'timing/data': 0.0004521778319031, 'timing/model': 1.074935446958989, 'learning_rate/qwen_vl_interface': 9.869357065630779e-06, 'learning_rate/action_model': 9.869357065630778e-05, 'epoch': 0.03}) +[2026-06-24 10:00:29,376][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.28513675928115845, 'train/grad_norm_pre_clip': 20.812997817993164, 'timing/data': 0.0005157280247658491, 'timing/model': 1.0733464150689542, 'learning_rate/qwen_vl_interface': 9.868155823054864e-06, 'learning_rate/action_model': 9.868155823054865e-05, 'epoch': 0.03}) +[2026-06-24 10:00:30,752][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.25218868255615234, 'train/grad_norm_pre_clip': 40.04468536376953, 'timing/data': 0.00038848910480737686, 'timing/model': 1.37419225811027, 'learning_rate/qwen_vl_interface': 9.86694916081586e-06, 'learning_rate/action_model': 9.866949160815859e-05, 'epoch': 0.03}) +[2026-06-24 10:00:31,826][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.2800222635269165, 'train/grad_norm_pre_clip': 27.22870445251465, 'timing/data': 0.0004128941800445318, 'timing/model': 1.071802169084549, 'learning_rate/qwen_vl_interface': 9.865737080329853e-06, 'learning_rate/action_model': 9.865737080329852e-05, 'epoch': 0.03}) +[2026-06-24 10:00:32,901][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.21312473714351654, 'train/grad_norm_pre_clip': 21.310359954833984, 'timing/data': 0.00046936096623539925, 'timing/model': 1.0734185331966728, 'learning_rate/qwen_vl_interface': 9.864519583019285e-06, 'learning_rate/action_model': 9.864519583019284e-05, 'epoch': 0.03}) +[2026-06-24 10:00:33,976][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.23297692835330963, 'train/grad_norm_pre_clip': 33.472923278808594, 'timing/data': 0.0005284820217639208, 'timing/model': 1.0721835771109909, 'learning_rate/qwen_vl_interface': 9.863296670312958e-06, 'learning_rate/action_model': 9.863296670312959e-05, 'epoch': 0.03}) +[2026-06-24 10:00:35,053][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.32181316614151, 'train/grad_norm_pre_clip': 71.08187103271484, 'timing/data': 0.00040584197267889977, 'timing/model': 1.0754874178674072, 'learning_rate/qwen_vl_interface': 9.86206834364603e-06, 'learning_rate/action_model': 9.86206834364603e-05, 'epoch': 0.03}) +[2026-06-24 10:00:36,126][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.1815161406993866, 'train/grad_norm_pre_clip': 59.79185104370117, 'timing/data': 0.00038458104245364666, 'timing/model': 1.07162368693389, 'learning_rate/qwen_vl_interface': 9.860834604460008e-06, 'learning_rate/action_model': 9.860834604460008e-05, 'epoch': 0.03}) +[2026-06-24 10:00:37,205][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.08325278759002686, 'train/grad_norm_pre_clip': 27.27272605895996, 'timing/data': 0.00045192777179181576, 'timing/model': 1.0771284690126777, 'learning_rate/qwen_vl_interface': 9.859595454202755e-06, 'learning_rate/action_model': 9.859595454202755e-05, 'epoch': 0.03}) +[2026-06-24 10:00:38,283][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.2916503846645355, 'train/grad_norm_pre_clip': 73.29124450683594, 'timing/data': 0.0004462671931833029, 'timing/model': 1.075280822115019, 'learning_rate/qwen_vl_interface': 9.858350894328484e-06, 'learning_rate/action_model': 9.858350894328483e-05, 'epoch': 0.03}) +[2026-06-24 10:00:39,356][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.6467506289482117, 'train/grad_norm_pre_clip': 140.12782287597656, 'timing/data': 0.00038506300188601017, 'timing/model': 1.0720411729998887, 'learning_rate/qwen_vl_interface': 9.857100926297753e-06, 'learning_rate/action_model': 9.857100926297753e-05, 'epoch': 0.03}) +[2026-06-24 10:00:40,434][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.11134811490774155, 'train/grad_norm_pre_clip': 64.52397155761719, 'timing/data': 0.0004947790876030922, 'timing/model': 1.0758764720521867, 'learning_rate/qwen_vl_interface': 9.855845551577468e-06, 'learning_rate/action_model': 9.855845551577468e-05, 'epoch': 0.03}) +[2026-06-24 10:00:41,515][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.16281946003437042, 'train/grad_norm_pre_clip': 43.03215408325195, 'timing/data': 0.00041994801722466946, 'timing/model': 1.078831127146259, 'learning_rate/qwen_vl_interface': 9.854584771640884e-06, 'learning_rate/action_model': 9.854584771640884e-05, 'epoch': 0.03}) +[2026-06-24 10:00:42,591][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.19380204379558563, 'train/grad_norm_pre_clip': 97.35527038574219, 'timing/data': 0.0005965190939605236, 'timing/model': 1.0739749849308282, 'learning_rate/qwen_vl_interface': 9.853318587967599e-06, 'learning_rate/action_model': 9.853318587967598e-05, 'epoch': 0.03}) +[2026-06-24 10:00:43,668][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.44454216957092285, 'train/grad_norm_pre_clip': 47.793949127197266, 'timing/data': 0.0004157999064773321, 'timing/model': 1.074330833973363, 'learning_rate/qwen_vl_interface': 9.852047002043543e-06, 'learning_rate/action_model': 9.852047002043543e-05, 'epoch': 0.03}) +[2026-06-24 10:00:44,741][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.4453010857105255, 'train/grad_norm_pre_clip': 64.1220703125, 'timing/data': 0.0004915220197290182, 'timing/model': 1.071229533990845, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.03}) +[2026-06-24 10:00:46,122][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.3292216658592224, 'train/grad_norm_pre_clip': 136.3730926513672, 'timing/data': 0.0009029849898070097, 'timing/model': 1.3784787789918482, 'learning_rate/qwen_vl_interface': 9.849487629418578e-06, 'learning_rate/action_model': 9.849487629418578e-05, 'epoch': 0.03}) +[2026-06-24 10:00:47,196][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.0954401046037674, 'train/grad_norm_pre_clip': 40.77801513671875, 'timing/data': 0.0004023248329758644, 'timing/model': 1.0723620718345046, 'learning_rate/qwen_vl_interface': 9.848199845721234e-06, 'learning_rate/action_model': 9.848199845721234e-05, 'epoch': 0.03}) +[2026-06-24 10:00:48,273][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.1831478327512741, 'train/grad_norm_pre_clip': 46.94286346435547, 'timing/data': 0.000635361997410655, 'timing/model': 1.0751195568591356, 'learning_rate/qwen_vl_interface': 9.846906665780253e-06, 'learning_rate/action_model': 9.846906665780252e-05, 'epoch': 0.03}) +[2026-06-24 10:00:49,344][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.28686219453811646, 'train/grad_norm_pre_clip': 76.00499725341797, 'timing/data': 0.00038178684189915657, 'timing/model': 1.06919118296355, 'learning_rate/qwen_vl_interface': 9.845608091113252e-06, 'learning_rate/action_model': 9.845608091113251e-05, 'epoch': 0.03}) +[2026-06-24 10:00:50,418][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.1372811496257782, 'train/grad_norm_pre_clip': 30.734766006469727, 'timing/data': 0.00042544794268906116, 'timing/model': 1.0716725999955088, 'learning_rate/qwen_vl_interface': 9.844304123244179e-06, 'learning_rate/action_model': 9.84430412324418e-05, 'epoch': 0.03}) +[2026-06-24 10:00:51,490][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.23898814618587494, 'train/grad_norm_pre_clip': 52.64915466308594, 'timing/data': 0.00038563390262424946, 'timing/model': 1.0700986781157553, 'learning_rate/qwen_vl_interface': 9.842994763703317e-06, 'learning_rate/action_model': 9.842994763703318e-05, 'epoch': 0.03}) +[2026-06-24 10:00:52,562][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.20101231336593628, 'train/grad_norm_pre_clip': 41.53354263305664, 'timing/data': 0.0003924458287656307, 'timing/model': 1.070506194140762, 'learning_rate/qwen_vl_interface': 9.84168001402727e-06, 'learning_rate/action_model': 9.84168001402727e-05, 'epoch': 0.03}) +[2026-06-24 10:00:53,640][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.31600233912467957, 'train/grad_norm_pre_clip': 55.6480827331543, 'timing/data': 0.00037291902117431164, 'timing/model': 1.075826568994671, 'learning_rate/qwen_vl_interface': 9.840359875758973e-06, 'learning_rate/action_model': 9.840359875758972e-05, 'epoch': 0.03}) +[2026-06-24 10:00:54,713][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.17089194059371948, 'train/grad_norm_pre_clip': 24.762338638305664, 'timing/data': 0.00043672998435795307, 'timing/model': 1.0715033570304513, 'learning_rate/qwen_vl_interface': 9.839034350447678e-06, 'learning_rate/action_model': 9.839034350447677e-05, 'epoch': 0.03}) +[2026-06-24 10:00:55,786][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.3196423649787903, 'train/grad_norm_pre_clip': 47.293861389160156, 'timing/data': 0.00039032287895679474, 'timing/model': 1.0716265491209924, 'learning_rate/qwen_vl_interface': 9.837703439648962e-06, 'learning_rate/action_model': 9.837703439648961e-05, 'epoch': 0.03}) +[2026-06-24 10:00:56,859][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.24018347263336182, 'train/grad_norm_pre_clip': 43.97160720825195, 'timing/data': 0.00033262395299971104, 'timing/model': 1.0707075949758291, 'learning_rate/qwen_vl_interface': 9.836367144924728e-06, 'learning_rate/action_model': 9.836367144924727e-05, 'epoch': 0.03}) +[2026-06-24 10:00:57,932][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.11999794095754623, 'train/grad_norm_pre_clip': 25.588905334472656, 'timing/data': 0.0006320259999483824, 'timing/model': 1.0716273218858987, 'learning_rate/qwen_vl_interface': 9.835025467843188e-06, 'learning_rate/action_model': 9.835025467843189e-05, 'epoch': 0.03}) +[2026-06-24 10:00:59,008][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.22208251059055328, 'train/grad_norm_pre_clip': 56.04624938964844, 'timing/data': 0.0003412200603634119, 'timing/model': 1.0742963710799813, 'learning_rate/qwen_vl_interface': 9.833678409978875e-06, 'learning_rate/action_model': 9.833678409978875e-05, 'epoch': 0.03}) +[2026-06-24 10:01:00,085][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.11394424736499786, 'train/grad_norm_pre_clip': 26.180559158325195, 'timing/data': 0.0003846019972115755, 'timing/model': 1.0745986488182098, 'learning_rate/qwen_vl_interface': 9.83232597291264e-06, 'learning_rate/action_model': 9.83232597291264e-05, 'epoch': 0.03}) +[2026-06-24 10:01:01,458][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.2933788001537323, 'train/grad_norm_pre_clip': 66.15625762939453, 'timing/data': 0.0004627089947462082, 'timing/model': 1.371613509953022, 'learning_rate/qwen_vl_interface': 9.83096815823164e-06, 'learning_rate/action_model': 9.83096815823164e-05, 'epoch': 0.03}) +[2026-06-24 10:01:02,532][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.32286345958709717, 'train/grad_norm_pre_clip': 61.3225212097168, 'timing/data': 0.00039366818964481354, 'timing/model': 1.0717924449127167, 'learning_rate/qwen_vl_interface': 9.829604967529346e-06, 'learning_rate/action_model': 9.829604967529347e-05, 'epoch': 0.03}) +[2026-06-24 10:01:03,607][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.23748622834682465, 'train/grad_norm_pre_clip': 71.51290893554688, 'timing/data': 0.0003578609321266413, 'timing/model': 1.0733783119358122, 'learning_rate/qwen_vl_interface': 9.828236402405541e-06, 'learning_rate/action_model': 9.828236402405541e-05, 'epoch': 0.03}) +[2026-06-24 10:01:04,685][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.146296888589859, 'train/grad_norm_pre_clip': 26.749814987182617, 'timing/data': 0.0004204388242214918, 'timing/model': 1.076120839919895, 'learning_rate/qwen_vl_interface': 9.826862464466309e-06, 'learning_rate/action_model': 9.826862464466309e-05, 'epoch': 0.03}) +[2026-06-24 10:01:05,764][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.17232050001621246, 'train/grad_norm_pre_clip': 33.477420806884766, 'timing/data': 0.0004493540618568659, 'timing/model': 1.0771183429751545, 'learning_rate/qwen_vl_interface': 9.825483155324046e-06, 'learning_rate/action_model': 9.825483155324046e-05, 'epoch': 0.03}) +[2026-06-24 10:01:06,842][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.3486977815628052, 'train/grad_norm_pre_clip': 48.28125, 'timing/data': 0.000417834147810936, 'timing/model': 1.075747560011223, 'learning_rate/qwen_vl_interface': 9.824098476597446e-06, 'learning_rate/action_model': 9.824098476597445e-05, 'epoch': 0.03}) +[2026-06-24 10:01:07,919][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.19913071393966675, 'train/grad_norm_pre_clip': 53.19306182861328, 'timing/data': 0.0003698938526213169, 'timing/model': 1.0755382690113038, 'learning_rate/qwen_vl_interface': 9.822708429911507e-06, 'learning_rate/action_model': 9.822708429911506e-05, 'epoch': 0.03}) +[2026-06-24 10:01:08,997][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.17584475874900818, 'train/grad_norm_pre_clip': 60.8397102355957, 'timing/data': 0.00044082687236368656, 'timing/model': 1.075926247984171, 'learning_rate/qwen_vl_interface': 9.821313016897524e-06, 'learning_rate/action_model': 9.821313016897524e-05, 'epoch': 0.03}) +[2026-06-24 10:01:10,074][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.33278781175613403, 'train/grad_norm_pre_clip': 43.540348052978516, 'timing/data': 0.00038624508306384087, 'timing/model': 1.0750808829907328, 'learning_rate/qwen_vl_interface': 9.819912239193097e-06, 'learning_rate/action_model': 9.819912239193097e-05, 'epoch': 0.03}) +[2026-06-24 10:01:11,151][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.306073397397995, 'train/grad_norm_pre_clip': 67.48468017578125, 'timing/data': 0.0004005718510597944, 'timing/model': 1.0749882601667196, 'learning_rate/qwen_vl_interface': 9.81850609844211e-06, 'learning_rate/action_model': 9.818506098442111e-05, 'epoch': 0.03}) +[2026-06-24 10:01:12,231][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.38425344228744507, 'train/grad_norm_pre_clip': 48.70390319824219, 'timing/data': 0.0003802729770541191, 'timing/model': 1.0777449042070657, 'learning_rate/qwen_vl_interface': 9.817094596294755e-06, 'learning_rate/action_model': 9.817094596294756e-05, 'epoch': 0.03}) +[2026-06-24 10:01:13,304][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.3689090311527252, 'train/grad_norm_pre_clip': 48.32225799560547, 'timing/data': 0.0004391339607536793, 'timing/model': 1.0710915881209075, 'learning_rate/qwen_vl_interface': 9.815677734407507e-06, 'learning_rate/action_model': 9.815677734407505e-05, 'epoch': 0.03}) +[2026-06-24 10:01:14,378][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.15965315699577332, 'train/grad_norm_pre_clip': 48.53922653198242, 'timing/data': 0.0004396250005811453, 'timing/model': 1.0729336959775537, 'learning_rate/qwen_vl_interface': 9.814255514443125e-06, 'learning_rate/action_model': 9.814255514443125e-05, 'epoch': 0.03}) +[2026-06-24 10:01:15,456][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.3852848708629608, 'train/grad_norm_pre_clip': 42.55001449584961, 'timing/data': 0.0003780401311814785, 'timing/model': 1.0754954139702022, 'learning_rate/qwen_vl_interface': 9.812827938070669e-06, 'learning_rate/action_model': 9.812827938070669e-05, 'epoch': 0.03}) +[2026-06-24 10:01:16,836][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.26416918635368347, 'train/grad_norm_pre_clip': 63.9891242980957, 'timing/data': 0.0004141470417380333, 'timing/model': 1.3785253271926194, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.03}) +[2026-06-24 10:01:17,912][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.2138974517583847, 'train/grad_norm_pre_clip': 27.107776641845703, 'timing/data': 0.00040374789386987686, 'timing/model': 1.0738967258948833, 'learning_rate/qwen_vl_interface': 9.809956722809181e-06, 'learning_rate/action_model': 9.809956722809182e-05, 'epoch': 0.04}) +[2026-06-24 10:01:18,987][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.22783824801445007, 'train/grad_norm_pre_clip': 28.865280151367188, 'timing/data': 0.0004897089675068855, 'timing/model': 1.0732290640007704, 'learning_rate/qwen_vl_interface': 9.80851308728968e-06, 'learning_rate/action_model': 9.80851308728968e-05, 'epoch': 0.04}) +[2026-06-24 10:01:20,063][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.3370087146759033, 'train/grad_norm_pre_clip': 58.478981018066406, 'timing/data': 0.00038176681846380234, 'timing/model': 1.0744347469881177, 'learning_rate/qwen_vl_interface': 9.80706410210116e-06, 'learning_rate/action_model': 9.80706410210116e-05, 'epoch': 0.04}) +[2026-06-24 10:01:21,141][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.32023322582244873, 'train/grad_norm_pre_clip': 46.17318344116211, 'timing/data': 0.00039146491326391697, 'timing/model': 1.0760236401110888, 'learning_rate/qwen_vl_interface': 9.805609768944092e-06, 'learning_rate/action_model': 9.805609768944091e-05, 'epoch': 0.04}) +[2026-06-24 10:01:22,214][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.4883434474468231, 'train/grad_norm_pre_clip': 65.24457550048828, 'timing/data': 0.0006744861602783203, 'timing/model': 1.07070045103319, 'learning_rate/qwen_vl_interface': 9.80415008952521e-06, 'learning_rate/action_model': 9.80415008952521e-05, 'epoch': 0.04}) +[2026-06-24 10:01:23,287][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.26192590594291687, 'train/grad_norm_pre_clip': 68.02877807617188, 'timing/data': 0.00041898712515830994, 'timing/model': 1.0712099180091172, 'learning_rate/qwen_vl_interface': 9.802685065557534e-06, 'learning_rate/action_model': 9.802685065557533e-05, 'epoch': 0.04}) +[2026-06-24 10:01:24,361][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.3748086094856262, 'train/grad_norm_pre_clip': 55.890724182128906, 'timing/data': 0.0004205589648336172, 'timing/model': 1.0721377599984407, 'learning_rate/qwen_vl_interface': 9.801214698760348e-06, 'learning_rate/action_model': 9.801214698760348e-05, 'epoch': 0.04}) +[2026-06-24 10:01:25,440][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.3204740881919861, 'train/grad_norm_pre_clip': 71.68401336669922, 'timing/data': 0.0003731301985681057, 'timing/model': 1.0769963790662587, 'learning_rate/qwen_vl_interface': 9.79973899085921e-06, 'learning_rate/action_model': 9.799738990859211e-05, 'epoch': 0.04}) +[2026-06-24 10:01:26,511][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.18657848238945007, 'train/grad_norm_pre_clip': 31.44207000732422, 'timing/data': 0.0003977359738200903, 'timing/model': 1.069732337957248, 'learning_rate/qwen_vl_interface': 9.79825794358595e-06, 'learning_rate/action_model': 9.79825794358595e-05, 'epoch': 0.04}) +[2026-06-24 10:01:27,585][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.18311773240566254, 'train/grad_norm_pre_clip': 22.326969146728516, 'timing/data': 0.000409719068557024, 'timing/model': 1.0715813781134784, 'learning_rate/qwen_vl_interface': 9.796771558678654e-06, 'learning_rate/action_model': 9.796771558678654e-05, 'epoch': 0.04}) +[2026-06-24 10:01:28,660][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.14775878190994263, 'train/grad_norm_pre_clip': 26.36616325378418, 'timing/data': 0.0004592819605022669, 'timing/model': 1.0729809049516916, 'learning_rate/qwen_vl_interface': 9.79527983788168e-06, 'learning_rate/action_model': 9.795279837881679e-05, 'epoch': 0.04}) +[2026-06-24 10:01:29,736][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.1305801421403885, 'train/grad_norm_pre_clip': 38.88177490234375, 'timing/data': 0.0005223511252552271, 'timing/model': 1.0742259628605098, 'learning_rate/qwen_vl_interface': 9.793782782945645e-06, 'learning_rate/action_model': 9.793782782945644e-05, 'epoch': 0.04}) +[2026-06-24 10:01:30,811][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.4570077359676361, 'train/grad_norm_pre_clip': 88.57142639160156, 'timing/data': 0.00038986210711300373, 'timing/model': 1.0727561449166387, 'learning_rate/qwen_vl_interface': 9.792280395627426e-06, 'learning_rate/action_model': 9.792280395627425e-05, 'epoch': 0.04}) +[2026-06-24 10:01:32,184][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.33146315813064575, 'train/grad_norm_pre_clip': 62.671695709228516, 'timing/data': 0.00038188579492270947, 'timing/model': 1.3715884529519826, 'learning_rate/qwen_vl_interface': 9.790772677690158e-06, 'learning_rate/action_model': 9.790772677690157e-05, 'epoch': 0.04}) +[2026-06-24 10:01:33,258][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.09720749408006668, 'train/grad_norm_pre_clip': 44.513423919677734, 'timing/data': 0.0004184648860245943, 'timing/model': 1.072138638002798, 'learning_rate/qwen_vl_interface': 9.78925963090323e-06, 'learning_rate/action_model': 9.78925963090323e-05, 'epoch': 0.04}) +[2026-06-24 10:01:34,336][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.030021991580724716, 'train/grad_norm_pre_clip': 10.33508014678955, 'timing/data': 0.00046792905777692795, 'timing/model': 1.0755787149537355, 'learning_rate/qwen_vl_interface': 9.78774125704229e-06, 'learning_rate/action_model': 9.78774125704229e-05, 'epoch': 0.04}) +[2026-06-24 10:01:35,409][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.16822952032089233, 'train/grad_norm_pre_clip': 114.74715423583984, 'timing/data': 0.00038235681131482124, 'timing/model': 1.0719336750917137, 'learning_rate/qwen_vl_interface': 9.786217557889234e-06, 'learning_rate/action_model': 9.786217557889235e-05, 'epoch': 0.04}) +[2026-06-24 10:01:36,489][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.10009200870990753, 'train/grad_norm_pre_clip': 112.95818328857422, 'timing/data': 0.00037631718441843987, 'timing/model': 1.0778172970749438, 'learning_rate/qwen_vl_interface': 9.784688535232208e-06, 'learning_rate/action_model': 9.784688535232208e-05, 'epoch': 0.04}) +[2026-06-24 10:01:37,561][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.3174501359462738, 'train/grad_norm_pre_clip': 132.23541259765625, 'timing/data': 0.00038938107900321484, 'timing/model': 1.0700096259824932, 'learning_rate/qwen_vl_interface': 9.783154190865604e-06, 'learning_rate/action_model': 9.783154190865604e-05, 'epoch': 0.04}) +[2026-06-24 10:01:38,633][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.284913569688797, 'train/grad_norm_pre_clip': 157.467041015625, 'timing/data': 0.0003886090125888586, 'timing/model': 1.0702597561758012, 'learning_rate/qwen_vl_interface': 9.781614526590061e-06, 'learning_rate/action_model': 9.78161452659006e-05, 'epoch': 0.04}) +[2026-06-24 10:01:39,708][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.30317187309265137, 'train/grad_norm_pre_clip': 113.60564422607422, 'timing/data': 0.00037655699998140335, 'timing/model': 1.0732917268760502, 'learning_rate/qwen_vl_interface': 9.78006954421246e-06, 'learning_rate/action_model': 9.78006954421246e-05, 'epoch': 0.04}) +[2026-06-24 10:01:40,784][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.25296124815940857, 'train/grad_norm_pre_clip': 65.83750915527344, 'timing/data': 0.00042255292646586895, 'timing/model': 1.0744553611148149, 'learning_rate/qwen_vl_interface': 9.778519245545924e-06, 'learning_rate/action_model': 9.778519245545924e-05, 'epoch': 0.04}) +[2026-06-24 10:01:41,863][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.3054247796535492, 'train/grad_norm_pre_clip': 55.00859451293945, 'timing/data': 0.00042359484359622, 'timing/model': 1.0773418189492077, 'learning_rate/qwen_vl_interface': 9.776963632409817e-06, 'learning_rate/action_model': 9.776963632409817e-05, 'epoch': 0.04}) +[2026-06-24 10:01:42,940][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.4096536636352539, 'train/grad_norm_pre_clip': 54.913265228271484, 'timing/data': 0.0003735411446541548, 'timing/model': 1.0749110779725015, 'learning_rate/qwen_vl_interface': 9.775402706629737e-06, 'learning_rate/action_model': 9.775402706629736e-05, 'epoch': 0.04}) +[2026-06-24 10:01:44,016][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.07324501872062683, 'train/grad_norm_pre_clip': 61.98291015625, 'timing/data': 0.0004930358845740557, 'timing/model': 1.0735999778844416, 'learning_rate/qwen_vl_interface': 9.773836470037513e-06, 'learning_rate/action_model': 9.773836470037513e-05, 'epoch': 0.04}) +[2026-06-24 10:01:45,091][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.3654317259788513, 'train/grad_norm_pre_clip': 579.6516723632812, 'timing/data': 0.0003860939759761095, 'timing/model': 1.073108206037432, 'learning_rate/qwen_vl_interface': 9.772264924471218e-06, 'learning_rate/action_model': 9.772264924471217e-05, 'epoch': 0.04}) +[2026-06-24 10:01:46,166][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 1.5546149015426636, 'train/grad_norm_pre_clip': 267.3955078125, 'timing/data': 0.00038350000977516174, 'timing/model': 1.0730268040206283, 'learning_rate/qwen_vl_interface': 9.770688071775141e-06, 'learning_rate/action_model': 9.770688071775142e-05, 'epoch': 0.04}) +[2026-06-24 10:01:47,540][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.8609936237335205, 'train/grad_norm_pre_clip': 195.87013244628906, 'timing/data': 0.00039699487388134, 'timing/model': 1.3724804210942239, 'learning_rate/qwen_vl_interface': 9.769105913799812e-06, 'learning_rate/action_model': 9.769105913799812e-05, 'epoch': 0.04}) +[2026-06-24 10:01:48,618][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.3936489522457123, 'train/grad_norm_pre_clip': 125.70556640625, 'timing/data': 0.00048095197416841984, 'timing/model': 1.0753295470494777, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.04}) +[2026-06-24 10:01:49,691][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.32157352566719055, 'train/grad_norm_pre_clip': 70.2935562133789, 'timing/data': 0.0004190760664641857, 'timing/model': 1.0718682520091534, 'learning_rate/qwen_vl_interface': 9.765925689444618e-06, 'learning_rate/action_model': 9.765925689444618e-05, 'epoch': 0.04}) +[2026-06-24 10:01:50,780][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.15941348671913147, 'train/grad_norm_pre_clip': 34.74949645996094, 'timing/data': 0.0005434809718281031, 'timing/model': 1.086324263131246, 'learning_rate/qwen_vl_interface': 9.764327626796925e-06, 'learning_rate/action_model': 9.764327626796924e-05, 'epoch': 0.04}) +[2026-06-24 10:01:51,867][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.2190636396408081, 'train/grad_norm_pre_clip': 50.03363037109375, 'timing/data': 0.0006310339085757732, 'timing/model': 1.0851957460399717, 'learning_rate/qwen_vl_interface': 9.762724266334312e-06, 'learning_rate/action_model': 9.762724266334312e-05, 'epoch': 0.04}) +[2026-06-24 10:01:52,945][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.4670243561267853, 'train/grad_norm_pre_clip': 115.53001403808594, 'timing/data': 0.0004600828979164362, 'timing/model': 1.0758777689188719, 'learning_rate/qwen_vl_interface': 9.761115609938415e-06, 'learning_rate/action_model': 9.761115609938415e-05, 'epoch': 0.04}) +[2026-06-24 10:01:54,023][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.19932043552398682, 'train/grad_norm_pre_clip': 51.155426025390625, 'timing/data': 0.00037897098809480667, 'timing/model': 1.0760862110182643, 'learning_rate/qwen_vl_interface': 9.759501659497081e-06, 'learning_rate/action_model': 9.75950165949708e-05, 'epoch': 0.04}) +[2026-06-24 10:01:55,099][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.4057006537914276, 'train/grad_norm_pre_clip': 42.81650924682617, 'timing/data': 0.0004333727993071079, 'timing/model': 1.0743479798547924, 'learning_rate/qwen_vl_interface': 9.757882416904371e-06, 'learning_rate/action_model': 9.757882416904371e-05, 'epoch': 0.04}) +[2026-06-24 10:01:56,174][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.21113187074661255, 'train/grad_norm_pre_clip': 31.176877975463867, 'timing/data': 0.0004626780282706022, 'timing/model': 1.072787141194567, 'learning_rate/qwen_vl_interface': 9.756257884060558e-06, 'learning_rate/action_model': 9.756257884060557e-05, 'epoch': 0.04}) +[2026-06-24 10:01:57,252][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.16772976517677307, 'train/grad_norm_pre_clip': 22.464393615722656, 'timing/data': 0.00046724709682166576, 'timing/model': 1.0761401050258428, 'learning_rate/qwen_vl_interface': 9.75462806287212e-06, 'learning_rate/action_model': 9.754628062872119e-05, 'epoch': 0.04}) +[2026-06-24 10:01:58,331][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.43530601263046265, 'train/grad_norm_pre_clip': 57.500038146972656, 'timing/data': 0.00038717593997716904, 'timing/model': 1.0773871061392128, 'learning_rate/qwen_vl_interface': 9.75299295525174e-06, 'learning_rate/action_model': 9.752992955251739e-05, 'epoch': 0.04}) +[2026-06-24 10:01:59,407][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.20867185294628143, 'train/grad_norm_pre_clip': 31.13644790649414, 'timing/data': 0.0003810750786215067, 'timing/model': 1.0742510419804603, 'learning_rate/qwen_vl_interface': 9.751352563118313e-06, 'learning_rate/action_model': 9.751352563118312e-05, 'epoch': 0.04}) +[2026-06-24 10:02:00,485][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.20683546364307404, 'train/grad_norm_pre_clip': 24.84394645690918, 'timing/data': 0.00037708808667957783, 'timing/model': 1.0752932680770755, 'learning_rate/qwen_vl_interface': 9.749706888396928e-06, 'learning_rate/action_model': 9.749706888396928e-05, 'epoch': 0.04}) +[2026-06-24 10:02:01,564][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.2393510341644287, 'train/grad_norm_pre_clip': 48.02507781982422, 'timing/data': 0.0003920861054211855, 'timing/model': 1.0776929419953376, 'learning_rate/qwen_vl_interface': 9.748055933018876e-06, 'learning_rate/action_model': 9.748055933018875e-05, 'epoch': 0.04}) +[2026-06-24 10:02:02,949][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.26168423891067505, 'train/grad_norm_pre_clip': 27.669021606445312, 'timing/data': 0.00041305506601929665, 'timing/model': 1.3827626309357584, 'learning_rate/qwen_vl_interface': 9.746399698921646e-06, 'learning_rate/action_model': 9.746399698921646e-05, 'epoch': 0.04}) +[2026-06-24 10:02:04,021][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.14935080707073212, 'train/grad_norm_pre_clip': 17.854801177978516, 'timing/data': 0.0004904801025986671, 'timing/model': 1.0703596009407192, 'learning_rate/qwen_vl_interface': 9.74473818804892e-06, 'learning_rate/action_model': 9.74473818804892e-05, 'epoch': 0.04}) +[2026-06-24 10:02:05,097][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.381437748670578, 'train/grad_norm_pre_clip': 36.5604133605957, 'timing/data': 0.0006614620797336102, 'timing/model': 1.0735390791669488, 'learning_rate/qwen_vl_interface': 9.743071402350576e-06, 'learning_rate/action_model': 9.743071402350575e-05, 'epoch': 0.04}) +[2026-06-24 10:02:06,171][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.11559345573186874, 'train/grad_norm_pre_clip': 21.196529388427734, 'timing/data': 0.00038131512701511383, 'timing/model': 1.0720210100989789, 'learning_rate/qwen_vl_interface': 9.741399343782674e-06, 'learning_rate/action_model': 9.741399343782675e-05, 'epoch': 0.04}) +[2026-06-24 10:02:07,246][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.14099043607711792, 'train/grad_norm_pre_clip': 39.6451301574707, 'timing/data': 0.0004138860385864973, 'timing/model': 1.0739714419469237, 'learning_rate/qwen_vl_interface': 9.739722014307475e-06, 'learning_rate/action_model': 9.739722014307475e-05, 'epoch': 0.04}) +[2026-06-24 10:02:08,323][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.285842627286911, 'train/grad_norm_pre_clip': 60.455909729003906, 'timing/data': 0.00043369410559535027, 'timing/model': 1.074489625170827, 'learning_rate/qwen_vl_interface': 9.738039415893415e-06, 'learning_rate/action_model': 9.738039415893414e-05, 'epoch': 0.04}) +[2026-06-24 10:02:09,401][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.21525155007839203, 'train/grad_norm_pre_clip': 26.806468963623047, 'timing/data': 0.0004073851741850376, 'timing/model': 1.0759309909772128, 'learning_rate/qwen_vl_interface': 9.736351550515117e-06, 'learning_rate/action_model': 9.736351550515116e-05, 'epoch': 0.04}) +[2026-06-24 10:02:10,479][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.31606391072273254, 'train/grad_norm_pre_clip': 77.48626708984375, 'timing/data': 0.0003977769520133734, 'timing/model': 1.0763798339758068, 'learning_rate/qwen_vl_interface': 9.734658420153384e-06, 'learning_rate/action_model': 9.734658420153384e-05, 'epoch': 0.04}) +[2026-06-24 10:02:11,555][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.292688250541687, 'train/grad_norm_pre_clip': 44.85165023803711, 'timing/data': 0.0005748891271650791, 'timing/model': 1.074208338977769, 'learning_rate/qwen_vl_interface': 9.732960026795202e-06, 'learning_rate/action_model': 9.732960026795202e-05, 'epoch': 0.04}) +[2026-06-24 10:02:12,632][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.35260099172592163, 'train/grad_norm_pre_clip': 49.49451446533203, 'timing/data': 0.0005427191499620676, 'timing/model': 1.0753280820790678, 'learning_rate/qwen_vl_interface': 9.731256372433725e-06, 'learning_rate/action_model': 9.731256372433725e-05, 'epoch': 0.04}) +[2026-06-24 10:02:13,710][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.3170284926891327, 'train/grad_norm_pre_clip': 66.53267669677734, 'timing/data': 0.000397656112909317, 'timing/model': 1.0753825439605862, 'learning_rate/qwen_vl_interface': 9.729547459068292e-06, 'learning_rate/action_model': 9.729547459068292e-05, 'epoch': 0.04}) +[2026-06-24 10:02:14,785][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.5836085081100464, 'train/grad_norm_pre_clip': 216.2882537841797, 'timing/data': 0.0005471270997077227, 'timing/model': 1.073350772028789, 'learning_rate/qwen_vl_interface': 9.727833288704406e-06, 'learning_rate/action_model': 9.727833288704406e-05, 'epoch': 0.04}) +[2026-06-24 10:02:15,861][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.3660382926464081, 'train/grad_norm_pre_clip': 198.8134002685547, 'timing/data': 0.0003879680298268795, 'timing/model': 1.0745348047930747, 'learning_rate/qwen_vl_interface': 9.726113863353741e-06, 'learning_rate/action_model': 9.72611386335374e-05, 'epoch': 0.04}) +[2026-06-24 10:02:16,931][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.16143636405467987, 'train/grad_norm_pre_clip': 72.78053283691406, 'timing/data': 0.0004465680103749037, 'timing/model': 1.0680591620039195, 'learning_rate/qwen_vl_interface': 9.724389185034138e-06, 'learning_rate/action_model': 9.724389185034137e-05, 'epoch': 0.04}) +[2026-06-24 10:02:18,317][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.37396591901779175, 'train/grad_norm_pre_clip': 57.56764602661133, 'timing/data': 0.000497954897582531, 'timing/model': 1.3837114358320832, 'learning_rate/qwen_vl_interface': 9.722659255769602e-06, 'learning_rate/action_model': 9.722659255769602e-05, 'epoch': 0.04}) +[2026-06-24 10:02:19,391][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.31377890706062317, 'train/grad_norm_pre_clip': 88.50439453125, 'timing/data': 0.0003985380753874779, 'timing/model': 1.0720680279191583, 'learning_rate/qwen_vl_interface': 9.720924077590304e-06, 'learning_rate/action_model': 9.720924077590304e-05, 'epoch': 0.04}) +[2026-06-24 10:02:20,466][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.03823750093579292, 'train/grad_norm_pre_clip': 21.488357543945312, 'timing/data': 0.00043223099783062935, 'timing/model': 1.073813053080812, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.04}) +[2026-06-24 10:02:21,538][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.4243224263191223, 'train/grad_norm_pre_clip': 54.73341751098633, 'timing/data': 0.00038712588138878345, 'timing/model': 1.0703442338854074, 'learning_rate/qwen_vl_interface': 9.71743798263889e-06, 'learning_rate/action_model': 9.717437982638889e-05, 'epoch': 0.04}) +[2026-06-24 10:02:22,612][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.16555912792682648, 'train/grad_norm_pre_clip': 41.592979431152344, 'timing/data': 0.0003962430637329817, 'timing/model': 1.0723034909460694, 'learning_rate/qwen_vl_interface': 9.7156870699579e-06, 'learning_rate/action_model': 9.7156870699579e-05, 'epoch': 0.04}) +[2026-06-24 10:02:23,688][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.4546872675418854, 'train/grad_norm_pre_clip': 81.50517272949219, 'timing/data': 0.0003739120438694954, 'timing/model': 1.0735237628687173, 'learning_rate/qwen_vl_interface': 9.713930916544393e-06, 'learning_rate/action_model': 9.713930916544393e-05, 'epoch': 0.04}) +[2026-06-24 10:02:24,764][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.2502526342868805, 'train/grad_norm_pre_clip': 81.61373901367188, 'timing/data': 0.0003987080417573452, 'timing/model': 1.0744142339099199, 'learning_rate/qwen_vl_interface': 9.712169524459318e-06, 'learning_rate/action_model': 9.712169524459317e-05, 'epoch': 0.04}) +[2026-06-24 10:02:25,837][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.1787702590227127, 'train/grad_norm_pre_clip': 46.50164031982422, 'timing/data': 0.0004254381638020277, 'timing/model': 1.07120082504116, 'learning_rate/qwen_vl_interface': 9.710402895769764e-06, 'learning_rate/action_model': 9.710402895769763e-05, 'epoch': 0.04}) +[2026-06-24 10:02:26,913][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.2563175857067108, 'train/grad_norm_pre_clip': 61.139610290527344, 'timing/data': 0.0003886090125888586, 'timing/model': 1.0736728748306632, 'learning_rate/qwen_vl_interface': 9.708631032548966e-06, 'learning_rate/action_model': 9.708631032548966e-05, 'epoch': 0.04}) +[2026-06-24 10:02:27,987][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.20576268434524536, 'train/grad_norm_pre_clip': 50.86822509765625, 'timing/data': 0.0004198080860078335, 'timing/model': 1.0729733631014824, 'learning_rate/qwen_vl_interface': 9.706853936876306e-06, 'learning_rate/action_model': 9.706853936876306e-05, 'epoch': 0.04}) +[2026-06-24 10:02:29,065][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.18323630094528198, 'train/grad_norm_pre_clip': 40.50469970703125, 'timing/data': 0.00037903105840086937, 'timing/model': 1.0760024078190327, 'learning_rate/qwen_vl_interface': 9.705071610837307e-06, 'learning_rate/action_model': 9.705071610837307e-05, 'epoch': 0.04}) +[2026-06-24 10:02:30,142][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.240624338388443, 'train/grad_norm_pre_clip': 38.22750473022461, 'timing/data': 0.0003922157920897007, 'timing/model': 1.0747094119433314, 'learning_rate/qwen_vl_interface': 9.703284056523624e-06, 'learning_rate/action_model': 9.703284056523623e-05, 'epoch': 0.04}) +[2026-06-24 10:02:31,214][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.3779412508010864, 'train/grad_norm_pre_clip': 40.9487419128418, 'timing/data': 0.00043741101399064064, 'timing/model': 1.0697241770103574, 'learning_rate/qwen_vl_interface': 9.701491276033054e-06, 'learning_rate/action_model': 9.701491276033054e-05, 'epoch': 0.04}) +[2026-06-24 10:02:32,294][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.21238172054290771, 'train/grad_norm_pre_clip': 37.98444366455078, 'timing/data': 0.0005380199290812016, 'timing/model': 1.0782440369948745, 'learning_rate/qwen_vl_interface': 9.699693271469523e-06, 'learning_rate/action_model': 9.699693271469524e-05, 'epoch': 0.04}) +[2026-06-24 10:02:33,674][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.11489671468734741, 'train/grad_norm_pre_clip': 35.20125961303711, 'timing/data': 0.00041653099469840527, 'timing/model': 1.3783490750938654, 'learning_rate/qwen_vl_interface': 9.697890044943094e-06, 'learning_rate/action_model': 9.697890044943093e-05, 'epoch': 0.04}) +[2026-06-24 10:02:34,751][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.10347456485033035, 'train/grad_norm_pre_clip': 26.349321365356445, 'timing/data': 0.0004396950826048851, 'timing/model': 1.0750440740957856, 'learning_rate/qwen_vl_interface': 9.696081598569947e-06, 'learning_rate/action_model': 9.696081598569945e-05, 'epoch': 0.04}) +[2026-06-24 10:02:35,830][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.14005309343338013, 'train/grad_norm_pre_clip': 18.855478286743164, 'timing/data': 0.00039222603663802147, 'timing/model': 1.0774721310008317, 'learning_rate/qwen_vl_interface': 9.694267934472398e-06, 'learning_rate/action_model': 9.694267934472397e-05, 'epoch': 0.04}) +[2026-06-24 10:02:36,906][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.14419525861740112, 'train/grad_norm_pre_clip': 24.27069854736328, 'timing/data': 0.0004354671109467745, 'timing/model': 1.0734534601215273, 'learning_rate/qwen_vl_interface': 9.692449054778884e-06, 'learning_rate/action_model': 9.692449054778884e-05, 'epoch': 0.04}) +[2026-06-24 10:02:37,983][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.18999379873275757, 'train/grad_norm_pre_clip': 37.88302993774414, 'timing/data': 0.00042998697608709335, 'timing/model': 1.0759496348910034, 'learning_rate/qwen_vl_interface': 9.69062496162396e-06, 'learning_rate/action_model': 9.69062496162396e-05, 'epoch': 0.04}) +[2026-06-24 10:02:39,058][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.576897382736206, 'train/grad_norm_pre_clip': 62.36164474487305, 'timing/data': 0.00040905806235969067, 'timing/model': 1.0729974750429392, 'learning_rate/qwen_vl_interface': 9.6887956571483e-06, 'learning_rate/action_model': 9.6887956571483e-05, 'epoch': 0.04}) +[2026-06-24 10:02:40,137][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.4000563621520996, 'train/grad_norm_pre_clip': 75.5433349609375, 'timing/data': 0.00042684096843004227, 'timing/model': 1.0766421360895038, 'learning_rate/qwen_vl_interface': 9.686961143498698e-06, 'learning_rate/action_model': 9.686961143498699e-05, 'epoch': 0.04}) +[2026-06-24 10:02:41,213][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.265837162733078, 'train/grad_norm_pre_clip': 55.824729919433594, 'timing/data': 0.0004730178043246269, 'timing/model': 1.0746600369457155, 'learning_rate/qwen_vl_interface': 9.685121422828057e-06, 'learning_rate/action_model': 9.685121422828057e-05, 'epoch': 0.04}) +[2026-06-24 10:02:42,295][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.23632310330867767, 'train/grad_norm_pre_clip': 35.08740997314453, 'timing/data': 0.0004284239839762449, 'timing/model': 1.0800617448985577, 'learning_rate/qwen_vl_interface': 9.683276497295388e-06, 'learning_rate/action_model': 9.683276497295389e-05, 'epoch': 0.04}) +[2026-06-24 10:02:43,372][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.4844667315483093, 'train/grad_norm_pre_clip': 100.7676773071289, 'timing/data': 0.00043827295303344727, 'timing/model': 1.0748933660797775, 'learning_rate/qwen_vl_interface': 9.681426369065817e-06, 'learning_rate/action_model': 9.681426369065817e-05, 'epoch': 0.04}) +[2026-06-24 10:02:44,451][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.11760013550519943, 'train/grad_norm_pre_clip': 27.7775936126709, 'timing/data': 0.0004292160738259554, 'timing/model': 1.0769473710097373, 'learning_rate/qwen_vl_interface': 9.679571040310574e-06, 'learning_rate/action_model': 9.679571040310574e-05, 'epoch': 0.04}) +[2026-06-24 10:02:45,528][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.3219447135925293, 'train/grad_norm_pre_clip': 43.993614196777344, 'timing/data': 0.00040632300078868866, 'timing/model': 1.0753748209681362, 'learning_rate/qwen_vl_interface': 9.677710513206988e-06, 'learning_rate/action_model': 9.677710513206986e-05, 'epoch': 0.04}) +[2026-06-24 10:02:46,605][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.10177425295114517, 'train/grad_norm_pre_clip': 18.43940544128418, 'timing/data': 0.0003912351094186306, 'timing/model': 1.0754150869324803, 'learning_rate/qwen_vl_interface': 9.67584478993849e-06, 'learning_rate/action_model': 9.67584478993849e-05, 'epoch': 0.04}) +[2026-06-24 10:02:47,683][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.11644145101308823, 'train/grad_norm_pre_clip': 22.235576629638672, 'timing/data': 0.00045489403419196606, 'timing/model': 1.0764876811299473, 'learning_rate/qwen_vl_interface': 9.673973872694614e-06, 'learning_rate/action_model': 9.673973872694613e-05, 'epoch': 0.04}) +[2026-06-24 10:02:49,069][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.2638869881629944, 'train/grad_norm_pre_clip': 54.86669158935547, 'timing/data': 0.000502352137118578, 'timing/model': 1.3841371710877866, 'learning_rate/qwen_vl_interface': 9.672097763670982e-06, 'learning_rate/action_model': 9.672097763670981e-05, 'epoch': 0.04}) +[2026-06-24 10:02:50,145][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.22208145260810852, 'train/grad_norm_pre_clip': 28.944547653198242, 'timing/data': 0.00037974282167851925, 'timing/model': 1.0732668379787356, 'learning_rate/qwen_vl_interface': 9.670216465069314e-06, 'learning_rate/action_model': 9.670216465069314e-05, 'epoch': 0.04}) +[2026-06-24 10:02:51,220][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.22753268480300903, 'train/grad_norm_pre_clip': 41.33480453491211, 'timing/data': 0.00040547107346355915, 'timing/model': 1.0731093131471425, 'learning_rate/qwen_vl_interface': 9.668329979097421e-06, 'learning_rate/action_model': 9.66832997909742e-05, 'epoch': 0.04}) +[2026-06-24 10:02:52,293][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.15542243421077728, 'train/grad_norm_pre_clip': 25.316659927368164, 'timing/data': 0.0004950189031660557, 'timing/model': 1.0706906539853662, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.04}) +[2026-06-24 10:02:53,370][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.32486382126808167, 'train/grad_norm_pre_clip': 66.75608825683594, 'timing/data': 0.0004000910557806492, 'timing/model': 1.0756451629567891, 'learning_rate/qwen_vl_interface': 9.66454145390462e-06, 'learning_rate/action_model': 9.66454145390462e-05, 'epoch': 0.04}) +[2026-06-24 10:02:54,445][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.2399894744157791, 'train/grad_norm_pre_clip': 61.088443756103516, 'timing/data': 0.000507652061060071, 'timing/model': 1.0729696520138532, 'learning_rate/qwen_vl_interface': 9.66263941912976e-06, 'learning_rate/action_model': 9.66263941912976e-05, 'epoch': 0.04}) +[2026-06-24 10:02:55,525][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.18842293322086334, 'train/grad_norm_pre_clip': 45.50716781616211, 'timing/data': 0.0003945911303162575, 'timing/model': 1.0782072530128062, 'learning_rate/qwen_vl_interface': 9.660732205876758e-06, 'learning_rate/action_model': 9.660732205876759e-05, 'epoch': 0.04}) +[2026-06-24 10:02:56,607][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.23547743260860443, 'train/grad_norm_pre_clip': 72.7551040649414, 'timing/data': 0.0004050300922244787, 'timing/model': 1.0750785910058767, 'learning_rate/qwen_vl_interface': 9.658819816383836e-06, 'learning_rate/action_model': 9.658819816383836e-05, 'epoch': 0.04}) +[2026-06-24 10:02:57,684][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.32624149322509766, 'train/grad_norm_pre_clip': 76.12277221679688, 'timing/data': 0.00041733309626579285, 'timing/model': 1.07511239615269, 'learning_rate/qwen_vl_interface': 9.656902252895288e-06, 'learning_rate/action_model': 9.656902252895287e-05, 'epoch': 0.04}) +[2026-06-24 10:02:58,760][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.4793781340122223, 'train/grad_norm_pre_clip': 111.29285430908203, 'timing/data': 0.00040835700929164886, 'timing/model': 1.0738412810023874, 'learning_rate/qwen_vl_interface': 9.654979517661483e-06, 'learning_rate/action_model': 9.654979517661482e-05, 'epoch': 0.04}) +[2026-06-24 10:02:59,833][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.21685153245925903, 'train/grad_norm_pre_clip': 57.65022659301758, 'timing/data': 0.0005458749365061522, 'timing/model': 1.0712279670406133, 'learning_rate/qwen_vl_interface': 9.653051612938855e-06, 'learning_rate/action_model': 9.653051612938855e-05, 'epoch': 0.04}) +[2026-06-24 10:03:00,910][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.2931291460990906, 'train/grad_norm_pre_clip': 49.75661087036133, 'timing/data': 0.0004615259822458029, 'timing/model': 1.0751409721560776, 'learning_rate/qwen_vl_interface': 9.651118540989912e-06, 'learning_rate/action_model': 9.651118540989912e-05, 'epoch': 0.04}) +[2026-06-24 10:03:01,984][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.21319371461868286, 'train/grad_norm_pre_clip': 44.395076751708984, 'timing/data': 0.00038975104689598083, 'timing/model': 1.0722297988831997, 'learning_rate/qwen_vl_interface': 9.649180304083218e-06, 'learning_rate/action_model': 9.649180304083219e-05, 'epoch': 0.04}) +[2026-06-24 10:03:03,063][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.40720516443252563, 'train/grad_norm_pre_clip': 55.33012390136719, 'timing/data': 0.00043469597585499287, 'timing/model': 1.0765825288835913, 'learning_rate/qwen_vl_interface': 9.647236904493404e-06, 'learning_rate/action_model': 9.647236904493404e-05, 'epoch': 0.04}) +[2026-06-24 10:03:04,442][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.311416894197464, 'train/grad_norm_pre_clip': 56.60221481323242, 'timing/data': 0.00044040591455996037, 'timing/model': 1.3773266619537026, 'learning_rate/qwen_vl_interface': 9.645288344501159e-06, 'learning_rate/action_model': 9.64528834450116e-05, 'epoch': 0.04}) +[2026-06-24 10:03:05,517][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.43327173590660095, 'train/grad_norm_pre_clip': 81.3143310546875, 'timing/data': 0.0003850420471280813, 'timing/model': 1.073051864048466, 'learning_rate/qwen_vl_interface': 9.643334626393224e-06, 'learning_rate/action_model': 9.643334626393223e-05, 'epoch': 0.04}) +[2026-06-24 10:03:06,590][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.21595826745033264, 'train/grad_norm_pre_clip': 54.262901306152344, 'timing/data': 0.00038084504194557667, 'timing/model': 1.0709499400109053, 'learning_rate/qwen_vl_interface': 9.641375752462395e-06, 'learning_rate/action_model': 9.641375752462395e-05, 'epoch': 0.04}) +[2026-06-24 10:03:07,663][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.19982922077178955, 'train/grad_norm_pre_clip': 29.55890655517578, 'timing/data': 0.00038463203236460686, 'timing/model': 1.0709755490534008, 'learning_rate/qwen_vl_interface': 9.639411725007522e-06, 'learning_rate/action_model': 9.639411725007522e-05, 'epoch': 0.04}) +[2026-06-24 10:03:08,737][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.21300658583641052, 'train/grad_norm_pre_clip': 22.342727661132812, 'timing/data': 0.0003641340881586075, 'timing/model': 1.072323088068515, 'learning_rate/qwen_vl_interface': 9.637442546333502e-06, 'learning_rate/action_model': 9.637442546333501e-05, 'epoch': 0.04}) +[2026-06-24 10:03:09,828][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.28541329503059387, 'train/grad_norm_pre_clip': 36.36554718017578, 'timing/data': 0.00040620286017656326, 'timing/model': 1.0894175011198968, 'learning_rate/qwen_vl_interface': 9.63546821875127e-06, 'learning_rate/action_model': 9.63546821875127e-05, 'epoch': 0.04}) +[2026-06-24 10:03:10,903][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.2901400327682495, 'train/grad_norm_pre_clip': 19.991470336914062, 'timing/data': 0.0004932060837745667, 'timing/model': 1.0734523390419781, 'learning_rate/qwen_vl_interface': 9.633488744577816e-06, 'learning_rate/action_model': 9.633488744577816e-05, 'epoch': 0.04}) +[2026-06-24 10:03:11,978][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.3501591384410858, 'train/grad_norm_pre_clip': 57.59210968017578, 'timing/data': 0.0003791919443756342, 'timing/model': 1.0730599930975586, 'learning_rate/qwen_vl_interface': 9.63150412613616e-06, 'learning_rate/action_model': 9.63150412613616e-05, 'epoch': 0.05}) +[2026-06-24 10:03:13,051][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.19431644678115845, 'train/grad_norm_pre_clip': 27.296289443969727, 'timing/data': 0.00039857812225818634, 'timing/model': 1.070718307979405, 'learning_rate/qwen_vl_interface': 9.629514365755364e-06, 'learning_rate/action_model': 9.629514365755364e-05, 'epoch': 0.05}) +[2026-06-24 10:03:14,126][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.16372747719287872, 'train/grad_norm_pre_clip': 20.109081268310547, 'timing/data': 0.00039942911826074123, 'timing/model': 1.073555904906243, 'learning_rate/qwen_vl_interface': 9.62751946577052e-06, 'learning_rate/action_model': 9.627519465770519e-05, 'epoch': 0.05}) +[2026-06-24 10:03:15,203][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.18852108716964722, 'train/grad_norm_pre_clip': 26.467763900756836, 'timing/data': 0.0003924060147255659, 'timing/model': 1.0747648540418595, 'learning_rate/qwen_vl_interface': 9.625519428522757e-06, 'learning_rate/action_model': 9.625519428522756e-05, 'epoch': 0.05}) +[2026-06-24 10:03:16,275][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.44353896379470825, 'train/grad_norm_pre_clip': 84.08732604980469, 'timing/data': 0.0003911438398063183, 'timing/model': 1.0706630169879645, 'learning_rate/qwen_vl_interface': 9.623514256359228e-06, 'learning_rate/action_model': 9.623514256359227e-05, 'epoch': 0.05}) +[2026-06-24 10:03:17,350][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.12641704082489014, 'train/grad_norm_pre_clip': 29.006486892700195, 'timing/data': 0.00038804812356829643, 'timing/model': 1.0731967131141573, 'learning_rate/qwen_vl_interface': 9.621503951633113e-06, 'learning_rate/action_model': 9.621503951633112e-05, 'epoch': 0.05}) +[2026-06-24 10:03:18,429][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.1153336688876152, 'train/grad_norm_pre_clip': 28.22978401184082, 'timing/data': 0.0003939990419894457, 'timing/model': 1.0768806780688465, 'learning_rate/qwen_vl_interface': 9.61948851670362e-06, 'learning_rate/action_model': 9.619488516703621e-05, 'epoch': 0.05}) +[2026-06-24 10:03:19,805][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.2929256856441498, 'train/grad_norm_pre_clip': 63.649967193603516, 'timing/data': 0.00042899511754512787, 'timing/model': 1.3743729700800031, 'learning_rate/qwen_vl_interface': 9.617467953935973e-06, 'learning_rate/action_model': 9.617467953935974e-05, 'epoch': 0.05}) +[2026-06-24 10:03:20,884][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.12723509967327118, 'train/grad_norm_pre_clip': 23.25933074951172, 'timing/data': 0.0005576459225267172, 'timing/model': 1.0773658391553909, 'learning_rate/qwen_vl_interface': 9.615442265701416e-06, 'learning_rate/action_model': 9.615442265701415e-05, 'epoch': 0.05}) +[2026-06-24 10:03:21,957][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.5485457181930542, 'train/grad_norm_pre_clip': 281.71539306640625, 'timing/data': 0.00042765215039253235, 'timing/model': 1.0715420299675316, 'learning_rate/qwen_vl_interface': 9.613411454377205e-06, 'learning_rate/action_model': 9.613411454377205e-05, 'epoch': 0.05}) +[2026-06-24 10:03:23,032][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.25258785486221313, 'train/grad_norm_pre_clip': 119.30128479003906, 'timing/data': 0.0003385350573807955, 'timing/model': 1.0730255951639265, 'learning_rate/qwen_vl_interface': 9.61137552234661e-06, 'learning_rate/action_model': 9.61137552234661e-05, 'epoch': 0.05}) +[2026-06-24 10:03:24,106][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.8467080593109131, 'train/grad_norm_pre_clip': 263.40887451171875, 'timing/data': 0.0003850830253213644, 'timing/model': 1.0719913351349533, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.05}) +[2026-06-24 10:03:25,183][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.3375527858734131, 'train/grad_norm_pre_clip': 71.1274185180664, 'timing/data': 0.00042212195694446564, 'timing/model': 1.0750504361931235, 'learning_rate/qwen_vl_interface': 9.607288305729396e-06, 'learning_rate/action_model': 9.607288305729395e-05, 'epoch': 0.05}) +[2026-06-24 10:03:26,262][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.2084011286497116, 'train/grad_norm_pre_clip': 69.75301361083984, 'timing/data': 0.0003960828762501478, 'timing/model': 1.0767322331666946, 'learning_rate/qwen_vl_interface': 9.605237025939353e-06, 'learning_rate/action_model': 9.605237025939353e-05, 'epoch': 0.05}) +[2026-06-24 10:03:27,337][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.210784450173378, 'train/grad_norm_pre_clip': 67.11756896972656, 'timing/data': 0.0005687370430678129, 'timing/model': 1.0732983420602977, 'learning_rate/qwen_vl_interface': 9.603180635036074e-06, 'learning_rate/action_model': 9.603180635036074e-05, 'epoch': 0.05}) +[2026-06-24 10:03:28,416][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.12411060184240341, 'train/grad_norm_pre_clip': 30.430307388305664, 'timing/data': 0.0004290060605853796, 'timing/model': 1.0764704239554703, 'learning_rate/qwen_vl_interface': 9.601119135432847e-06, 'learning_rate/action_model': 9.601119135432847e-05, 'epoch': 0.05}) +[2026-06-24 10:03:29,491][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.20904621481895447, 'train/grad_norm_pre_clip': 62.08809280395508, 'timing/data': 0.00045039597898721695, 'timing/model': 1.0735441618598998, 'learning_rate/qwen_vl_interface': 9.59905252954896e-06, 'learning_rate/action_model': 9.599052529548961e-05, 'epoch': 0.05}) +[2026-06-24 10:03:30,569][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.1410025656223297, 'train/grad_norm_pre_clip': 39.42490768432617, 'timing/data': 0.00043245102278888226, 'timing/model': 1.0757589789573103, 'learning_rate/qwen_vl_interface': 9.59698081980969e-06, 'learning_rate/action_model': 9.59698081980969e-05, 'epoch': 0.05}) +[2026-06-24 10:03:31,646][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.14588357508182526, 'train/grad_norm_pre_clip': 32.375244140625, 'timing/data': 0.0004014729056507349, 'timing/model': 1.0757253367919475, 'learning_rate/qwen_vl_interface': 9.594904008646298e-06, 'learning_rate/action_model': 9.594904008646298e-05, 'epoch': 0.05}) +[2026-06-24 10:03:32,721][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.11213722079992294, 'train/grad_norm_pre_clip': 155.00363159179688, 'timing/data': 0.0004552940372377634, 'timing/model': 1.0731670651584864, 'learning_rate/qwen_vl_interface': 9.592822098496041e-06, 'learning_rate/action_model': 9.592822098496041e-05, 'epoch': 0.05}) +[2026-06-24 10:03:33,797][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.1907264143228531, 'train/grad_norm_pre_clip': 49.516807556152344, 'timing/data': 0.0005476679652929306, 'timing/model': 1.073886577039957, 'learning_rate/qwen_vl_interface': 9.59073509180216e-06, 'learning_rate/action_model': 9.590735091802159e-05, 'epoch': 0.05}) +[2026-06-24 10:03:35,176][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.423275351524353, 'train/grad_norm_pre_clip': 112.99724578857422, 'timing/data': 0.00037709809839725494, 'timing/model': 1.3765956528950483, 'learning_rate/qwen_vl_interface': 9.588642991013867e-06, 'learning_rate/action_model': 9.588642991013867e-05, 'epoch': 0.05}) +[2026-06-24 10:03:36,251][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.35005664825439453, 'train/grad_norm_pre_clip': 70.4297866821289, 'timing/data': 0.0004510669969022274, 'timing/model': 1.0731090400367975, 'learning_rate/qwen_vl_interface': 9.586545798586364e-06, 'learning_rate/action_model': 9.586545798586363e-05, 'epoch': 0.05}) +[2026-06-24 10:03:37,326][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.05216482654213905, 'train/grad_norm_pre_clip': 13.229483604431152, 'timing/data': 0.00041941716335713863, 'timing/model': 1.0733524661045521, 'learning_rate/qwen_vl_interface': 9.58444351698082e-06, 'learning_rate/action_model': 9.58444351698082e-05, 'epoch': 0.05}) +[2026-06-24 10:03:38,405][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.1959216594696045, 'train/grad_norm_pre_clip': 29.601839065551758, 'timing/data': 0.0004046198446303606, 'timing/model': 1.0774519408587366, 'learning_rate/qwen_vl_interface': 9.582336148664382e-06, 'learning_rate/action_model': 9.582336148664382e-05, 'epoch': 0.05}) +[2026-06-24 10:03:39,484][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.49005597829818726, 'train/grad_norm_pre_clip': 83.70924377441406, 'timing/data': 0.00044629815965890884, 'timing/model': 1.076524080010131, 'learning_rate/qwen_vl_interface': 9.580223696110163e-06, 'learning_rate/action_model': 9.580223696110164e-05, 'epoch': 0.05}) +[2026-06-24 10:03:40,565][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.580333411693573, 'train/grad_norm_pre_clip': 158.41525268554688, 'timing/data': 0.0004490229766815901, 'timing/model': 1.0797498349566013, 'learning_rate/qwen_vl_interface': 9.578106161797246e-06, 'learning_rate/action_model': 9.578106161797244e-05, 'epoch': 0.05}) +[2026-06-24 10:03:41,644][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.4046647548675537, 'train/grad_norm_pre_clip': 38.70856857299805, 'timing/data': 0.00047737592831254005, 'timing/model': 1.0765900949481875, 'learning_rate/qwen_vl_interface': 9.575983548210668e-06, 'learning_rate/action_model': 9.575983548210668e-05, 'epoch': 0.05}) +[2026-06-24 10:03:42,716][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.17927615344524384, 'train/grad_norm_pre_clip': 67.69525146484375, 'timing/data': 0.00037691788747906685, 'timing/model': 1.070560310035944, 'learning_rate/qwen_vl_interface': 9.573855857841444e-06, 'learning_rate/action_model': 9.573855857841442e-05, 'epoch': 0.05}) +[2026-06-24 10:03:43,791][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.11465156078338623, 'train/grad_norm_pre_clip': 18.721879959106445, 'timing/data': 0.000664958031848073, 'timing/model': 1.0723948939703405, 'learning_rate/qwen_vl_interface': 9.57172309318653e-06, 'learning_rate/action_model': 9.571723093186528e-05, 'epoch': 0.05}) +[2026-06-24 10:03:44,868][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.08863908797502518, 'train/grad_norm_pre_clip': 20.04886817932129, 'timing/data': 0.0003961131442338228, 'timing/model': 1.0756200461182743, 'learning_rate/qwen_vl_interface': 9.569585256748847e-06, 'learning_rate/action_model': 9.569585256748847e-05, 'epoch': 0.05}) +[2026-06-24 10:03:45,945][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.2930963933467865, 'train/grad_norm_pre_clip': 92.87459564208984, 'timing/data': 0.0003525521606206894, 'timing/model': 1.0746319028548896, 'learning_rate/qwen_vl_interface': 9.567442351037268e-06, 'learning_rate/action_model': 9.567442351037266e-05, 'epoch': 0.05}) +[2026-06-24 10:03:47,017][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.2261328250169754, 'train/grad_norm_pre_clip': 39.862876892089844, 'timing/data': 0.0003870758228003979, 'timing/model': 1.0700439929496497, 'learning_rate/qwen_vl_interface': 9.565294378566605e-06, 'learning_rate/action_model': 9.565294378566605e-05, 'epoch': 0.05}) +[2026-06-24 10:03:48,093][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.25522327423095703, 'train/grad_norm_pre_clip': 45.51546096801758, 'timing/data': 0.0003913838881999254, 'timing/model': 1.0744058298878372, 'learning_rate/qwen_vl_interface': 9.56314134185763e-06, 'learning_rate/action_model': 9.563141341857629e-05, 'epoch': 0.05}) +[2026-06-24 10:03:49,171][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 500 +[2026-06-24 10:04:15,502][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 500 +[2026-06-24 10:04:15,503][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 500 +[2026-06-24 10:04:48,741][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 500 +[2026-06-24 10:06:30,804][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-24 10:06:30,805][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.5550442934036255, 'train/grad_norm_pre_clip': 44.039947509765625, 'eval/action_loss/seconds': 26.329614877002314, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.10662183165550232, 'eval/latency_2/loss': 0.10662183165550232, 'eval/flappy/loss': 0.10662183165550232, 'eval/flappy/latency_2/loss': 0.10662183165550232, 'eval/flap_precision': 0.8473557692307693, 'eval/flap_recall': 0.9565807327001357, 'eval/flap_f1': 0.898661567877629, 'eval/flap_rate_model': 0.2311111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8986615678776289, 'eval/latency_2/flap_precision': 0.8473557692307693, 'eval/latency_2/flap_recall': 0.9565807327001357, 'eval/latency_2/flap_f1': 0.898661567877629, 'eval/latency_2/flap_rate_model': 0.2311111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8986615678776289, 'eval/action_classification/data_wait_seconds': 0.26301464438438416, 'eval/action_classification/predict_seconds': 32.19293975830078, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 73.41999999999845, 'rl_games_eval/mid_train/mean_length': 612.0, 'rl_games_eval/mid_train/std_reward': 90.13413116017405, 'rl_games_eval/mid_train/std_length': 727.8139872247579, 'rl_games_eval/mid_train/macro_mean_reward': 73.41999999999845, 'rl_games_eval/mid_train/macro_mean_length': 612.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 73.41999999999845, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 612.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 90.13413116017405, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 727.8139872247579, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 73.41999999999845, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003845309838652611, 'timing/model': 1.076553330058232, 'learning_rate/qwen_vl_interface': 9.560983243437048e-06, 'learning_rate/action_model': 9.560983243437047e-05, 'epoch': 0.05}) +[2026-06-24 10:06:30,806][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_500_state +[2026-06-24 10:06:30,809][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 10:06:33,596][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_500_state/model.safetensors +[2026-06-24 10:06:39,359][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_500_state/optimizer.bin +[2026-06-24 10:06:39,360][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_500_state/sampler.bin +[2026-06-24 10:06:39,360][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_500_state/sampler_1.bin +[2026-06-24 10:06:39,360][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-24 10:06:39,361][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-24 10:06:39,362][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 10:06:39,371][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 10:06:40,424][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.1776381880044937, 'train/grad_norm_pre_clip': 34.46935272216797, 'timing/data': 0.00044644810259342194, 'timing/model': 1.0507132650818676, 'learning_rate/qwen_vl_interface': 9.558820085837506e-06, 'learning_rate/action_model': 9.558820085837506e-05, 'epoch': 0.05}) +[2026-06-24 10:06:41,482][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.2091018408536911, 'train/grad_norm_pre_clip': 42.5357666015625, 'timing/data': 0.0004937069024890661, 'timing/model': 1.0559216209221631, 'learning_rate/qwen_vl_interface': 9.556651871597596e-06, 'learning_rate/action_model': 9.556651871597596e-05, 'epoch': 0.05}) +[2026-06-24 10:06:42,544][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.3162565529346466, 'train/grad_norm_pre_clip': 52.55730056762695, 'timing/data': 0.0004260498099029064, 'timing/model': 1.0601775890681893, 'learning_rate/qwen_vl_interface': 9.554478603261835e-06, 'learning_rate/action_model': 9.554478603261835e-05, 'epoch': 0.05}) +[2026-06-24 10:06:43,607][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.22292199730873108, 'train/grad_norm_pre_clip': 37.73289108276367, 'timing/data': 0.000491501996293664, 'timing/model': 1.0612600401509553, 'learning_rate/qwen_vl_interface': 9.552300283380673e-06, 'learning_rate/action_model': 9.552300283380672e-05, 'epoch': 0.05}) +[2026-06-24 10:06:44,672][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.301725298166275, 'train/grad_norm_pre_clip': 40.45988082885742, 'timing/data': 0.00040844688192009926, 'timing/model': 1.0627411608584225, 'learning_rate/qwen_vl_interface': 9.55011691451049e-06, 'learning_rate/action_model': 9.55011691451049e-05, 'epoch': 0.05}) +[2026-06-24 10:06:45,727][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.18306203186511993, 'train/grad_norm_pre_clip': 19.29051971435547, 'timing/data': 0.0004104301333427429, 'timing/model': 1.0532456489745528, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.05}) +[2026-06-24 10:06:47,091][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.1826203465461731, 'train/grad_norm_pre_clip': 65.03531646728516, 'timing/data': 0.00041743298061192036, 'timing/model': 1.3621833119541407, 'learning_rate/qwen_vl_interface': 9.545735040058209e-06, 'learning_rate/action_model': 9.545735040058208e-05, 'epoch': 0.05}) +[2026-06-24 10:06:48,151][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.2121025025844574, 'train/grad_norm_pre_clip': 176.65121459960938, 'timing/data': 0.00039743585512042046, 'timing/model': 1.058378032874316, 'learning_rate/qwen_vl_interface': 9.543536539618481e-06, 'learning_rate/action_model': 9.543536539618482e-05, 'epoch': 0.05}) +[2026-06-24 10:06:49,211][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.1807328313589096, 'train/grad_norm_pre_clip': 23.96444320678711, 'timing/data': 0.0004978841170668602, 'timing/model': 1.0579110770486295, 'learning_rate/qwen_vl_interface': 9.541333000474477e-06, 'learning_rate/action_model': 9.541333000474477e-05, 'epoch': 0.05}) +[2026-06-24 10:06:50,266][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.32585006952285767, 'train/grad_norm_pre_clip': 129.96360778808594, 'timing/data': 0.0004355870187282562, 'timing/model': 1.0529375630430877, 'learning_rate/qwen_vl_interface': 9.53912442521217e-06, 'learning_rate/action_model': 9.53912442521217e-05, 'epoch': 0.05}) +[2026-06-24 10:06:51,322][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.22408278286457062, 'train/grad_norm_pre_clip': 30.90377426147461, 'timing/data': 0.0004343949258327484, 'timing/model': 1.0546718081459403, 'learning_rate/qwen_vl_interface': 9.536910816423444e-06, 'learning_rate/action_model': 9.536910816423445e-05, 'epoch': 0.05}) +[2026-06-24 10:06:52,378][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.23740150034427643, 'train/grad_norm_pre_clip': 21.5386905670166, 'timing/data': 0.00038171582855284214, 'timing/model': 1.0540846770163625, 'learning_rate/qwen_vl_interface': 9.5346921767061e-06, 'learning_rate/action_model': 9.534692176706099e-05, 'epoch': 0.05}) +[2026-06-24 10:06:53,440][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.16052532196044922, 'train/grad_norm_pre_clip': 35.92375946044922, 'timing/data': 0.0003959729801863432, 'timing/model': 1.0599370321724564, 'learning_rate/qwen_vl_interface': 9.532468508663823e-06, 'learning_rate/action_model': 9.532468508663823e-05, 'epoch': 0.05}) +[2026-06-24 10:06:54,497][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.34071654081344604, 'train/grad_norm_pre_clip': 64.76881408691406, 'timing/data': 0.0004332331009209156, 'timing/model': 1.0548857611138374, 'learning_rate/qwen_vl_interface': 9.530239814906223e-06, 'learning_rate/action_model': 9.530239814906223e-05, 'epoch': 0.05}) +[2026-06-24 10:06:55,559][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.45684704184532166, 'train/grad_norm_pre_clip': 83.83180236816406, 'timing/data': 0.0003957529552280903, 'timing/model': 1.060670980019495, 'learning_rate/qwen_vl_interface': 9.528006098048789e-06, 'learning_rate/action_model': 9.528006098048788e-05, 'epoch': 0.05}) +[2026-06-24 10:06:56,617][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.12657353281974792, 'train/grad_norm_pre_clip': 27.378820419311523, 'timing/data': 0.0003896709531545639, 'timing/model': 1.0559934610500932, 'learning_rate/qwen_vl_interface': 9.525767360712913e-06, 'learning_rate/action_model': 9.525767360712914e-05, 'epoch': 0.05}) +[2026-06-24 10:06:57,677][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.1614772081375122, 'train/grad_norm_pre_clip': 51.89202117919922, 'timing/data': 0.00039932900108397007, 'timing/model': 1.058554467977956, 'learning_rate/qwen_vl_interface': 9.523523605525882e-06, 'learning_rate/action_model': 9.523523605525883e-05, 'epoch': 0.05}) +[2026-06-24 10:06:58,733][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.16501407325267792, 'train/grad_norm_pre_clip': 26.708940505981445, 'timing/data': 0.00038258801214396954, 'timing/model': 1.054001212818548, 'learning_rate/qwen_vl_interface': 9.521274835120866e-06, 'learning_rate/action_model': 9.521274835120866e-05, 'epoch': 0.05}) +[2026-06-24 10:06:59,793][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.31431424617767334, 'train/grad_norm_pre_clip': 72.52680969238281, 'timing/data': 0.0003764471039175987, 'timing/model': 1.0579324718564749, 'learning_rate/qwen_vl_interface': 9.519021052136922e-06, 'learning_rate/action_model': 9.51902105213692e-05, 'epoch': 0.05}) +[2026-06-24 10:07:00,853][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.23806534707546234, 'train/grad_norm_pre_clip': 65.93468475341797, 'timing/data': 0.00039011193439364433, 'timing/model': 1.0579110521357507, 'learning_rate/qwen_vl_interface': 9.516762259218986e-06, 'learning_rate/action_model': 9.516762259218986e-05, 'epoch': 0.05}) +[2026-06-24 10:07:01,920][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.42733269929885864, 'train/grad_norm_pre_clip': 88.88713836669922, 'timing/data': 0.00046846899203956127, 'timing/model': 1.0658219999168068, 'learning_rate/qwen_vl_interface': 9.514498459017882e-06, 'learning_rate/action_model': 9.514498459017882e-05, 'epoch': 0.05}) +[2026-06-24 10:07:02,976][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.507050633430481, 'train/grad_norm_pre_clip': 85.94985961914062, 'timing/data': 0.00041101197712123394, 'timing/model': 1.0536557459272444, 'learning_rate/qwen_vl_interface': 9.512229654190306e-06, 'learning_rate/action_model': 9.512229654190305e-05, 'epoch': 0.05}) +[2026-06-24 10:07:04,342][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.21544848382472992, 'train/grad_norm_pre_clip': 56.37261199951172, 'timing/data': 0.00040888809598982334, 'timing/model': 1.3642492578364909, 'learning_rate/qwen_vl_interface': 9.509955847398824e-06, 'learning_rate/action_model': 9.509955847398824e-05, 'epoch': 0.05}) +[2026-06-24 10:07:05,397][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.12295748293399811, 'train/grad_norm_pre_clip': 41.181026458740234, 'timing/data': 0.0003784908913075924, 'timing/model': 1.0531456500757486, 'learning_rate/qwen_vl_interface': 9.507677041311876e-06, 'learning_rate/action_model': 9.507677041311875e-05, 'epoch': 0.05}) +[2026-06-24 10:07:06,455][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.7463458776473999, 'train/grad_norm_pre_clip': 166.09347534179688, 'timing/data': 0.00039756600745022297, 'timing/model': 1.0559516260400414, 'learning_rate/qwen_vl_interface': 9.505393238603768e-06, 'learning_rate/action_model': 9.505393238603768e-05, 'epoch': 0.05}) +[2026-06-24 10:07:07,515][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.2829006612300873, 'train/grad_norm_pre_clip': 115.75658416748047, 'timing/data': 0.0003850329667329788, 'timing/model': 1.0586589381564409, 'learning_rate/qwen_vl_interface': 9.50310444195467e-06, 'learning_rate/action_model': 9.50310444195467e-05, 'epoch': 0.05}) +[2026-06-24 10:07:08,575][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.614465057849884, 'train/grad_norm_pre_clip': 191.35897827148438, 'timing/data': 0.0003936891444027424, 'timing/model': 1.057606552960351, 'learning_rate/qwen_vl_interface': 9.500810654050612e-06, 'learning_rate/action_model': 9.500810654050611e-05, 'epoch': 0.05}) +[2026-06-24 10:07:09,639][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.6791144013404846, 'train/grad_norm_pre_clip': 186.89659118652344, 'timing/data': 0.000403356971219182, 'timing/model': 1.0623182870913297, 'learning_rate/qwen_vl_interface': 9.498511877583482e-06, 'learning_rate/action_model': 9.498511877583482e-05, 'epoch': 0.05}) +[2026-06-24 10:07:10,704][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.6441145539283752, 'train/grad_norm_pre_clip': 154.20816040039062, 'timing/data': 0.0004389041569083929, 'timing/model': 1.06271638488397, 'learning_rate/qwen_vl_interface': 9.496208115251022e-06, 'learning_rate/action_model': 9.496208115251022e-05, 'epoch': 0.05}) +[2026-06-24 10:07:11,767][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.3420972526073456, 'train/grad_norm_pre_clip': 66.19940185546875, 'timing/data': 0.00045000389218330383, 'timing/model': 1.061527973972261, 'learning_rate/qwen_vl_interface': 9.493899369756825e-06, 'learning_rate/action_model': 9.493899369756826e-05, 'epoch': 0.05}) +[2026-06-24 10:07:12,831][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.3199766278266907, 'train/grad_norm_pre_clip': 41.300933837890625, 'timing/data': 0.0003825169987976551, 'timing/model': 1.061850680038333, 'learning_rate/qwen_vl_interface': 9.491585643810334e-06, 'learning_rate/action_model': 9.491585643810333e-05, 'epoch': 0.05}) +[2026-06-24 10:07:13,892][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.1249222680926323, 'train/grad_norm_pre_clip': 11.55813980102539, 'timing/data': 0.00044616800732910633, 'timing/model': 1.0599580691196024, 'learning_rate/qwen_vl_interface': 9.489266940126833e-06, 'learning_rate/action_model': 9.489266940126833e-05, 'epoch': 0.05}) +[2026-06-24 10:07:14,959][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.2111118584871292, 'train/grad_norm_pre_clip': 19.61069679260254, 'timing/data': 0.00037054484710097313, 'timing/model': 1.0646291358862072, 'learning_rate/qwen_vl_interface': 9.486943261427453e-06, 'learning_rate/action_model': 9.486943261427453e-05, 'epoch': 0.05}) +[2026-06-24 10:07:16,023][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.22266343235969543, 'train/grad_norm_pre_clip': 41.16796112060547, 'timing/data': 0.00039023207500576973, 'timing/model': 1.0622521850746125, 'learning_rate/qwen_vl_interface': 9.484614610439156e-06, 'learning_rate/action_model': 9.484614610439156e-05, 'epoch': 0.05}) +[2026-06-24 10:07:17,088][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.11199264973402023, 'train/grad_norm_pre_clip': 15.378739356994629, 'timing/data': 0.00044751004315912724, 'timing/model': 1.0634055000264198, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.05}) +[2026-06-24 10:07:18,151][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.1938299685716629, 'train/grad_norm_pre_clip': 41.60159683227539, 'timing/data': 0.0004167631268501282, 'timing/model': 1.0609224699437618, 'learning_rate/qwen_vl_interface': 9.47994240253286e-06, 'learning_rate/action_model': 9.47994240253286e-05, 'epoch': 0.05}) +[2026-06-24 10:07:19,216][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.10863593965768814, 'train/grad_norm_pre_clip': 44.98411560058594, 'timing/data': 0.00037731812335550785, 'timing/model': 1.0634070141240954, 'learning_rate/qwen_vl_interface': 9.477598851097957e-06, 'learning_rate/action_model': 9.477598851097956e-05, 'epoch': 0.05}) +[2026-06-24 10:07:20,285][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.4877708852291107, 'train/grad_norm_pre_clip': 75.24994659423828, 'timing/data': 0.0003832499496638775, 'timing/model': 1.0671949840616435, 'learning_rate/qwen_vl_interface': 9.475250338340323e-06, 'learning_rate/action_model': 9.475250338340323e-05, 'epoch': 0.05}) +[2026-06-24 10:07:21,649][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.2248491495847702, 'train/grad_norm_pre_clip': 33.86885452270508, 'timing/data': 0.0004008319228887558, 'timing/model': 1.361683791037649, 'learning_rate/qwen_vl_interface': 9.472896867016072e-06, 'learning_rate/action_model': 9.472896867016072e-05, 'epoch': 0.05}) +[2026-06-24 10:07:22,711][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.11187130212783813, 'train/grad_norm_pre_clip': 19.91325569152832, 'timing/data': 0.0004129749722778797, 'timing/model': 1.0602158450055867, 'learning_rate/qwen_vl_interface': 9.470538439887134e-06, 'learning_rate/action_model': 9.470538439887134e-05, 'epoch': 0.05}) +[2026-06-24 10:07:23,774][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.21296650171279907, 'train/grad_norm_pre_clip': 63.28336715698242, 'timing/data': 0.0004276528488844633, 'timing/model': 1.0613215921912342, 'learning_rate/qwen_vl_interface': 9.468175059721252e-06, 'learning_rate/action_model': 9.468175059721253e-05, 'epoch': 0.05}) +[2026-06-24 10:07:24,838][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.15393108129501343, 'train/grad_norm_pre_clip': 30.28472328186035, 'timing/data': 0.00041219405829906464, 'timing/model': 1.0623765119817108, 'learning_rate/qwen_vl_interface': 9.465806729291986e-06, 'learning_rate/action_model': 9.465806729291987e-05, 'epoch': 0.05}) +[2026-06-24 10:07:25,900][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.22350476682186127, 'train/grad_norm_pre_clip': 50.152015686035156, 'timing/data': 0.0003730999305844307, 'timing/model': 1.0600731489248574, 'learning_rate/qwen_vl_interface': 9.463433451378704e-06, 'learning_rate/action_model': 9.463433451378703e-05, 'epoch': 0.05}) +[2026-06-24 10:07:26,964][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.04795489087700844, 'train/grad_norm_pre_clip': 20.174728393554688, 'timing/data': 0.0004022950306534767, 'timing/model': 1.0618470890913159, 'learning_rate/qwen_vl_interface': 9.461055228766579e-06, 'learning_rate/action_model': 9.461055228766579e-05, 'epoch': 0.05}) +[2026-06-24 10:07:28,028][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.2141267955303192, 'train/grad_norm_pre_clip': 33.961341857910156, 'timing/data': 0.00043262215331196785, 'timing/model': 1.0615912089124322, 'learning_rate/qwen_vl_interface': 9.458672064246587e-06, 'learning_rate/action_model': 9.458672064246586e-05, 'epoch': 0.05}) +[2026-06-24 10:07:29,092][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.18057522177696228, 'train/grad_norm_pre_clip': 60.29799270629883, 'timing/data': 0.00036387285217642784, 'timing/model': 1.0622131540440023, 'learning_rate/qwen_vl_interface': 9.456283960615505e-06, 'learning_rate/action_model': 9.456283960615505e-05, 'epoch': 0.05}) +[2026-06-24 10:07:30,155][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.1077435091137886, 'train/grad_norm_pre_clip': 51.467002868652344, 'timing/data': 0.0003987681120634079, 'timing/model': 1.0618363691028208, 'learning_rate/qwen_vl_interface': 9.453890920675908e-06, 'learning_rate/action_model': 9.453890920675907e-05, 'epoch': 0.05}) +[2026-06-24 10:07:31,215][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.30987757444381714, 'train/grad_norm_pre_clip': 89.59874725341797, 'timing/data': 0.00040712393820285797, 'timing/model': 1.0577912968583405, 'learning_rate/qwen_vl_interface': 9.451492947236158e-06, 'learning_rate/action_model': 9.451492947236157e-05, 'epoch': 0.05}) +[2026-06-24 10:07:32,280][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.24653449654579163, 'train/grad_norm_pre_clip': 76.57417297363281, 'timing/data': 0.0003873170353472233, 'timing/model': 1.0632349450606853, 'learning_rate/qwen_vl_interface': 9.449090043110411e-06, 'learning_rate/action_model': 9.449090043110411e-05, 'epoch': 0.05}) +[2026-06-24 10:07:33,345][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.1301729381084442, 'train/grad_norm_pre_clip': 68.21092224121094, 'timing/data': 0.0004075851757079363, 'timing/model': 1.063143634935841, 'learning_rate/qwen_vl_interface': 9.44668221111861e-06, 'learning_rate/action_model': 9.44668221111861e-05, 'epoch': 0.05}) +[2026-06-24 10:07:34,409][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.15455080568790436, 'train/grad_norm_pre_clip': 46.466129302978516, 'timing/data': 0.000371056143194437, 'timing/model': 1.0624837060458958, 'learning_rate/qwen_vl_interface': 9.44426945408648e-06, 'learning_rate/action_model': 9.444269454086479e-05, 'epoch': 0.05}) +[2026-06-24 10:07:35,471][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.14584311842918396, 'train/grad_norm_pre_clip': 36.94351577758789, 'timing/data': 0.00045249913819134235, 'timing/model': 1.0598902700003237, 'learning_rate/qwen_vl_interface': 9.441851774845526e-06, 'learning_rate/action_model': 9.441851774845525e-05, 'epoch': 0.05}) +[2026-06-24 10:07:36,534][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.04124195873737335, 'train/grad_norm_pre_clip': 23.196908950805664, 'timing/data': 0.00039067305624485016, 'timing/model': 1.0610312318895012, 'learning_rate/qwen_vl_interface': 9.439429176233027e-06, 'learning_rate/action_model': 9.439429176233027e-05, 'epoch': 0.05}) +[2026-06-24 10:07:37,600][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.09432148188352585, 'train/grad_norm_pre_clip': 19.379220962524414, 'timing/data': 0.00041685206815600395, 'timing/model': 1.0638287120964378, 'learning_rate/qwen_vl_interface': 9.437001661092038e-06, 'learning_rate/action_model': 9.437001661092038e-05, 'epoch': 0.05}) +[2026-06-24 10:07:38,970][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.2722317576408386, 'train/grad_norm_pre_clip': 76.92420959472656, 'timing/data': 0.00045434292405843735, 'timing/model': 1.368025611853227, 'learning_rate/qwen_vl_interface': 9.434569232271387e-06, 'learning_rate/action_model': 9.434569232271388e-05, 'epoch': 0.05}) +[2026-06-24 10:07:40,035][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.2688870131969452, 'train/grad_norm_pre_clip': 157.7654266357422, 'timing/data': 0.0003920961171388626, 'timing/model': 1.063810087973252, 'learning_rate/qwen_vl_interface': 9.432131892625662e-06, 'learning_rate/action_model': 9.432131892625661e-05, 'epoch': 0.05}) +[2026-06-24 10:07:41,099][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.17421263456344604, 'train/grad_norm_pre_clip': 33.73931121826172, 'timing/data': 0.0004041979555040598, 'timing/model': 1.0618740739300847, 'learning_rate/qwen_vl_interface': 9.429689645015215e-06, 'learning_rate/action_model': 9.429689645015215e-05, 'epoch': 0.05}) +[2026-06-24 10:07:42,165][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.15540564060211182, 'train/grad_norm_pre_clip': 43.92359924316406, 'timing/data': 0.0005184740293771029, 'timing/model': 1.0637550759129226, 'learning_rate/qwen_vl_interface': 9.427242492306164e-06, 'learning_rate/action_model': 9.427242492306163e-05, 'epoch': 0.05}) +[2026-06-24 10:07:43,230][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.07741344720125198, 'train/grad_norm_pre_clip': 26.981502532958984, 'timing/data': 0.00043055787682533264, 'timing/model': 1.063024725066498, 'learning_rate/qwen_vl_interface': 9.424790437370374e-06, 'learning_rate/action_model': 9.424790437370374e-05, 'epoch': 0.05}) +[2026-06-24 10:07:44,295][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.2398398369550705, 'train/grad_norm_pre_clip': 48.89938735961914, 'timing/data': 0.0003960728645324707, 'timing/model': 1.0631414339877665, 'learning_rate/qwen_vl_interface': 9.42233348308547e-06, 'learning_rate/action_model': 9.42233348308547e-05, 'epoch': 0.05}) +[2026-06-24 10:07:45,358][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.36820289492607117, 'train/grad_norm_pre_clip': 75.39678192138672, 'timing/data': 0.0004006619565188885, 'timing/model': 1.0609725639224052, 'learning_rate/qwen_vl_interface': 9.419871632334827e-06, 'learning_rate/action_model': 9.419871632334826e-05, 'epoch': 0.05}) +[2026-06-24 10:07:46,422][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.2936805188655853, 'train/grad_norm_pre_clip': 75.30001831054688, 'timing/data': 0.0003762568812817335, 'timing/model': 1.0620282760355622, 'learning_rate/qwen_vl_interface': 9.417404888007563e-06, 'learning_rate/action_model': 9.417404888007563e-05, 'epoch': 0.05}) +[2026-06-24 10:07:47,488][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.09069859236478806, 'train/grad_norm_pre_clip': 28.83869171142578, 'timing/data': 0.0006154249422252178, 'timing/model': 1.063919065054506, 'learning_rate/qwen_vl_interface': 9.414933252998537e-06, 'learning_rate/action_model': 9.414933252998537e-05, 'epoch': 0.05}) +[2026-06-24 10:07:48,552][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.41797757148742676, 'train/grad_norm_pre_clip': 48.85035705566406, 'timing/data': 0.00041714287362992764, 'timing/model': 1.0620847519021481, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.05}) +[2026-06-24 10:07:49,618][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.09812340140342712, 'train/grad_norm_pre_clip': 16.938066482543945, 'timing/data': 0.0003829479683190584, 'timing/model': 1.0645586161408573, 'learning_rate/qwen_vl_interface': 9.409975322543348e-06, 'learning_rate/action_model': 9.409975322543348e-05, 'epoch': 0.05}) +[2026-06-24 10:07:50,681][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.11945411562919617, 'train/grad_norm_pre_clip': 28.684043884277344, 'timing/data': 0.0003834988456219435, 'timing/model': 1.061177761061117, 'learning_rate/qwen_vl_interface': 9.407489032915593e-06, 'learning_rate/action_model': 9.407489032915592e-05, 'epoch': 0.05}) +[2026-06-24 10:07:51,747][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.32551324367523193, 'train/grad_norm_pre_clip': 61.35264587402344, 'timing/data': 0.0004574188496917486, 'timing/model': 1.0640782050322741, 'learning_rate/qwen_vl_interface': 9.404997864242884e-06, 'learning_rate/action_model': 9.404997864242882e-05, 'epoch': 0.05}) +[2026-06-24 10:07:52,812][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.14438524842262268, 'train/grad_norm_pre_clip': 26.95313262939453, 'timing/data': 0.00038803904317319393, 'timing/model': 1.062656647991389, 'learning_rate/qwen_vl_interface': 9.402501819448748e-06, 'learning_rate/action_model': 9.402501819448748e-05, 'epoch': 0.05}) +[2026-06-24 10:07:53,874][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.22483238577842712, 'train/grad_norm_pre_clip': 38.327999114990234, 'timing/data': 0.0005276799201965332, 'timing/model': 1.060461597982794, 'learning_rate/qwen_vl_interface': 9.400000901462436e-06, 'learning_rate/action_model': 9.400000901462434e-05, 'epoch': 0.06}) +[2026-06-24 10:07:54,937][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.21327100694179535, 'train/grad_norm_pre_clip': 33.33664321899414, 'timing/data': 0.0004021150525659323, 'timing/model': 1.0613502159249038, 'learning_rate/qwen_vl_interface': 9.39749511321891e-06, 'learning_rate/action_model': 9.397495113218909e-05, 'epoch': 0.06}) +[2026-06-24 10:07:56,300][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.28472572565078735, 'train/grad_norm_pre_clip': 44.31187057495117, 'timing/data': 0.00037430296652019024, 'timing/model': 1.3610125570558012, 'learning_rate/qwen_vl_interface': 9.394984457658854e-06, 'learning_rate/action_model': 9.394984457658855e-05, 'epoch': 0.06}) +[2026-06-24 10:07:57,363][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.15743008255958557, 'train/grad_norm_pre_clip': 29.146862030029297, 'timing/data': 0.0005795171018689871, 'timing/model': 1.060906263999641, 'learning_rate/qwen_vl_interface': 9.392468937728662e-06, 'learning_rate/action_model': 9.392468937728661e-05, 'epoch': 0.06}) +[2026-06-24 10:07:58,430][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.15702055394649506, 'train/grad_norm_pre_clip': 30.019670486450195, 'timing/data': 0.0003737818915396929, 'timing/model': 1.0646944851614535, 'learning_rate/qwen_vl_interface': 9.389948556380436e-06, 'learning_rate/action_model': 9.389948556380435e-05, 'epoch': 0.06}) +[2026-06-24 10:07:59,497][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.1783248633146286, 'train/grad_norm_pre_clip': 25.65818977355957, 'timing/data': 0.000468709971755743, 'timing/model': 1.0657714961562306, 'learning_rate/qwen_vl_interface': 9.387423316571985e-06, 'learning_rate/action_model': 9.387423316571985e-05, 'epoch': 0.06}) +[2026-06-24 10:08:00,565][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.24446114897727966, 'train/grad_norm_pre_clip': 41.83126449584961, 'timing/data': 0.00043904385529458523, 'timing/model': 1.0652563800103962, 'learning_rate/qwen_vl_interface': 9.384893221266818e-06, 'learning_rate/action_model': 9.384893221266818e-05, 'epoch': 0.06}) +[2026-06-24 10:08:01,630][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.32030999660491943, 'train/grad_norm_pre_clip': 53.20219039916992, 'timing/data': 0.00038153701461851597, 'timing/model': 1.0634276478085667, 'learning_rate/qwen_vl_interface': 9.382358273434143e-06, 'learning_rate/action_model': 9.382358273434143e-05, 'epoch': 0.06}) +[2026-06-24 10:08:02,693][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.16763631999492645, 'train/grad_norm_pre_clip': 88.0619125366211, 'timing/data': 0.00038659502752125263, 'timing/model': 1.0612541779410094, 'learning_rate/qwen_vl_interface': 9.379818476048861e-06, 'learning_rate/action_model': 9.379818476048862e-05, 'epoch': 0.06}) +[2026-06-24 10:08:03,755][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.27293500304222107, 'train/grad_norm_pre_clip': 36.366676330566406, 'timing/data': 0.00041403714567422867, 'timing/model': 1.0609112351667136, 'learning_rate/qwen_vl_interface': 9.377273832091571e-06, 'learning_rate/action_model': 9.37727383209157e-05, 'epoch': 0.06}) +[2026-06-24 10:08:04,821][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.3170420825481415, 'train/grad_norm_pre_clip': 41.88395309448242, 'timing/data': 0.0004138059448450804, 'timing/model': 1.0634508118964732, 'learning_rate/qwen_vl_interface': 9.374724344548547e-06, 'learning_rate/action_model': 9.374724344548546e-05, 'epoch': 0.06}) +[2026-06-24 10:08:05,883][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.14801537990570068, 'train/grad_norm_pre_clip': 31.52916145324707, 'timing/data': 0.0006231390871107578, 'timing/model': 1.0599543410353363, 'learning_rate/qwen_vl_interface': 9.372170016411757e-06, 'learning_rate/action_model': 9.372170016411757e-05, 'epoch': 0.06}) +[2026-06-24 10:08:06,947][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.13875198364257812, 'train/grad_norm_pre_clip': 20.036846160888672, 'timing/data': 0.0004018740728497505, 'timing/model': 1.0626912370789796, 'learning_rate/qwen_vl_interface': 9.369610850678851e-06, 'learning_rate/action_model': 9.369610850678851e-05, 'epoch': 0.06}) +[2026-06-24 10:08:08,012][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.27291473746299744, 'train/grad_norm_pre_clip': 39.56003952026367, 'timing/data': 0.00046750809997320175, 'timing/model': 1.0627879488747567, 'learning_rate/qwen_vl_interface': 9.367046850353148e-06, 'learning_rate/action_model': 9.367046850353148e-05, 'epoch': 0.06}) +[2026-06-24 10:08:09,076][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.21961914002895355, 'train/grad_norm_pre_clip': 43.86101150512695, 'timing/data': 0.0004073651507496834, 'timing/model': 1.0617136829532683, 'learning_rate/qwen_vl_interface': 9.364478018443649e-06, 'learning_rate/action_model': 9.364478018443649e-05, 'epoch': 0.06}) +[2026-06-24 10:08:10,140][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.09019356220960617, 'train/grad_norm_pre_clip': 23.63960838317871, 'timing/data': 0.0003862848971039057, 'timing/model': 1.0624964630696923, 'learning_rate/qwen_vl_interface': 9.361904357965018e-06, 'learning_rate/action_model': 9.361904357965018e-05, 'epoch': 0.06}) +[2026-06-24 10:08:11,206][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.20240157842636108, 'train/grad_norm_pre_clip': 45.42714309692383, 'timing/data': 0.00044065690599381924, 'timing/model': 1.064467272022739, 'learning_rate/qwen_vl_interface': 9.359325871937592e-06, 'learning_rate/action_model': 9.359325871937592e-05, 'epoch': 0.06}) +[2026-06-24 10:08:12,270][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.06696554273366928, 'train/grad_norm_pre_clip': 26.43190574645996, 'timing/data': 0.0005039558745920658, 'timing/model': 1.061893611913547, 'learning_rate/qwen_vl_interface': 9.356742563387366e-06, 'learning_rate/action_model': 9.356742563387365e-05, 'epoch': 0.06}) +[2026-06-24 10:08:13,636][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.15903760492801666, 'train/grad_norm_pre_clip': 40.066078186035156, 'timing/data': 0.000377157935872674, 'timing/model': 1.363821045961231, 'learning_rate/qwen_vl_interface': 9.354154435345995e-06, 'learning_rate/action_model': 9.354154435345995e-05, 'epoch': 0.06}) +[2026-06-24 10:08:14,699][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.2982211709022522, 'train/grad_norm_pre_clip': 62.78519058227539, 'timing/data': 0.000387377105653286, 'timing/model': 1.0606614700518548, 'learning_rate/qwen_vl_interface': 9.351561490850794e-06, 'learning_rate/action_model': 9.351561490850794e-05, 'epoch': 0.06}) +[2026-06-24 10:08:15,763][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.18933722376823425, 'train/grad_norm_pre_clip': 52.944725036621094, 'timing/data': 0.00046529294922947884, 'timing/model': 1.0625164119992405, 'learning_rate/qwen_vl_interface': 9.348963732944728e-06, 'learning_rate/action_model': 9.348963732944727e-05, 'epoch': 0.06}) +[2026-06-24 10:08:16,828][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.07144907861948013, 'train/grad_norm_pre_clip': 23.304025650024414, 'timing/data': 0.0003839700948446989, 'timing/model': 1.0632049751002342, 'learning_rate/qwen_vl_interface': 9.346361164676411e-06, 'learning_rate/action_model': 9.34636116467641e-05, 'epoch': 0.06}) +[2026-06-24 10:08:17,890][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.1843666136264801, 'train/grad_norm_pre_clip': 83.04466247558594, 'timing/data': 0.0003518308512866497, 'timing/model': 1.0602022090461105, 'learning_rate/qwen_vl_interface': 9.343753789100098e-06, 'learning_rate/action_model': 9.343753789100098e-05, 'epoch': 0.06}) +[2026-06-24 10:08:18,959][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.5835630297660828, 'train/grad_norm_pre_clip': 95.54852294921875, 'timing/data': 0.0005752900615334511, 'timing/model': 1.0668669189326465, 'learning_rate/qwen_vl_interface': 9.341141609275693e-06, 'learning_rate/action_model': 9.341141609275692e-05, 'epoch': 0.06}) +[2026-06-24 10:08:20,027][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.06898948550224304, 'train/grad_norm_pre_clip': 18.595502853393555, 'timing/data': 0.00039119389839470387, 'timing/model': 1.0668691031169146, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.06}) +[2026-06-24 10:08:21,093][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.23775257170200348, 'train/grad_norm_pre_clip': 46.17556381225586, 'timing/data': 0.0004035781603306532, 'timing/model': 1.063533291919157, 'learning_rate/qwen_vl_interface': 9.335902849150393e-06, 'learning_rate/action_model': 9.335902849150391e-05, 'epoch': 0.06}) +[2026-06-24 10:08:22,157][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.33174970746040344, 'train/grad_norm_pre_clip': 99.01009368896484, 'timing/data': 0.00035970588214695454, 'timing/model': 1.0630896801594645, 'learning_rate/qwen_vl_interface': 9.333276274997471e-06, 'learning_rate/action_model': 9.33327627499747e-05, 'epoch': 0.06}) +[2026-06-24 10:08:23,218][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.7811017036437988, 'train/grad_norm_pre_clip': 176.54635620117188, 'timing/data': 0.00039797695353627205, 'timing/model': 1.0589624040294439, 'learning_rate/qwen_vl_interface': 9.330644908892408e-06, 'learning_rate/action_model': 9.330644908892408e-05, 'epoch': 0.06}) +[2026-06-24 10:08:24,285][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.1523120403289795, 'train/grad_norm_pre_clip': 35.6802978515625, 'timing/data': 0.0003901419695466757, 'timing/model': 1.0645470169838518, 'learning_rate/qwen_vl_interface': 9.328008753923248e-06, 'learning_rate/action_model': 9.328008753923248e-05, 'epoch': 0.06}) +[2026-06-24 10:08:25,350][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.2585241198539734, 'train/grad_norm_pre_clip': 43.50493621826172, 'timing/data': 0.0003808948677033186, 'timing/model': 1.0638229071628302, 'learning_rate/qwen_vl_interface': 9.325367813183675e-06, 'learning_rate/action_model': 9.325367813183675e-05, 'epoch': 0.06}) +[2026-06-24 10:08:26,413][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.2671082615852356, 'train/grad_norm_pre_clip': 55.35210418701172, 'timing/data': 0.00037482287734746933, 'timing/model': 1.0612725410610437, 'learning_rate/qwen_vl_interface': 9.322722089772973e-06, 'learning_rate/action_model': 9.322722089772972e-05, 'epoch': 0.06}) +[2026-06-24 10:08:27,481][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.24923932552337646, 'train/grad_norm_pre_clip': 49.1739387512207, 'timing/data': 0.0005977419205009937, 'timing/model': 1.0653812231030315, 'learning_rate/qwen_vl_interface': 9.320071586796048e-06, 'learning_rate/action_model': 9.320071586796048e-05, 'epoch': 0.06}) +[2026-06-24 10:08:28,548][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.22613754868507385, 'train/grad_norm_pre_clip': 46.164363861083984, 'timing/data': 0.0004633490461856127, 'timing/model': 1.0651980901602656, 'learning_rate/qwen_vl_interface': 9.317416307363417e-06, 'learning_rate/action_model': 9.317416307363417e-05, 'epoch': 0.06}) +[2026-06-24 10:08:29,609][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.0983349084854126, 'train/grad_norm_pre_clip': 28.84413719177246, 'timing/data': 0.00039725610986351967, 'timing/model': 1.059209831058979, 'learning_rate/qwen_vl_interface': 9.314756254591193e-06, 'learning_rate/action_model': 9.314756254591192e-05, 'epoch': 0.06}) +[2026-06-24 10:08:30,969][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.1921151578426361, 'train/grad_norm_pre_clip': 41.70667266845703, 'timing/data': 0.00039944006130099297, 'timing/model': 1.3582251900807023, 'learning_rate/qwen_vl_interface': 9.312091431601097e-06, 'learning_rate/action_model': 9.312091431601097e-05, 'epoch': 0.06}) +[2026-06-24 10:08:32,037][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.26161688566207886, 'train/grad_norm_pre_clip': 37.298583984375, 'timing/data': 0.0003931571263819933, 'timing/model': 1.065660729072988, 'learning_rate/qwen_vl_interface': 9.309421841520447e-06, 'learning_rate/action_model': 9.309421841520447e-05, 'epoch': 0.06}) +[2026-06-24 10:08:33,104][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.1404542773962021, 'train/grad_norm_pre_clip': 21.150419235229492, 'timing/data': 0.00037721777334809303, 'timing/model': 1.0660298520233482, 'learning_rate/qwen_vl_interface': 9.30674748748216e-06, 'learning_rate/action_model': 9.306747487482159e-05, 'epoch': 0.06}) +[2026-06-24 10:08:34,171][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.19101309776306152, 'train/grad_norm_pre_clip': 36.83077621459961, 'timing/data': 0.00037216884084045887, 'timing/model': 1.0646078330464661, 'learning_rate/qwen_vl_interface': 9.304068372624734e-06, 'learning_rate/action_model': 9.304068372624734e-05, 'epoch': 0.06}) +[2026-06-24 10:08:35,240][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.17530709505081177, 'train/grad_norm_pre_clip': 27.012327194213867, 'timing/data': 0.0004565769340842962, 'timing/model': 1.0668214887846261, 'learning_rate/qwen_vl_interface': 9.301384500092262e-06, 'learning_rate/action_model': 9.301384500092261e-05, 'epoch': 0.06}) +[2026-06-24 10:08:36,308][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.257949560880661, 'train/grad_norm_pre_clip': 72.76786804199219, 'timing/data': 0.00038982206024229527, 'timing/model': 1.0664363661780953, 'learning_rate/qwen_vl_interface': 9.298695873034419e-06, 'learning_rate/action_model': 9.298695873034417e-05, 'epoch': 0.06}) +[2026-06-24 10:08:37,373][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.23204396665096283, 'train/grad_norm_pre_clip': 76.11475372314453, 'timing/data': 0.00047641387209296227, 'timing/model': 1.0629044680390507, 'learning_rate/qwen_vl_interface': 9.296002494606454e-06, 'learning_rate/action_model': 9.296002494606453e-05, 'epoch': 0.06}) +[2026-06-24 10:08:38,438][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.31398841738700867, 'train/grad_norm_pre_clip': 31.620624542236328, 'timing/data': 0.000462027033790946, 'timing/model': 1.0636175570543855, 'learning_rate/qwen_vl_interface': 9.293304367969202e-06, 'learning_rate/action_model': 9.293304367969202e-05, 'epoch': 0.06}) +[2026-06-24 10:08:39,504][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.33152201771736145, 'train/grad_norm_pre_clip': 44.96388244628906, 'timing/data': 0.0004522288218140602, 'timing/model': 1.0637940580490977, 'learning_rate/qwen_vl_interface': 9.290601496289064e-06, 'learning_rate/action_model': 9.290601496289064e-05, 'epoch': 0.06}) +[2026-06-24 10:08:40,570][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.18279294669628143, 'train/grad_norm_pre_clip': 30.28721809387207, 'timing/data': 0.00039909896440804005, 'timing/model': 1.0642699210438877, 'learning_rate/qwen_vl_interface': 9.287893882738005e-06, 'learning_rate/action_model': 9.287893882738005e-05, 'epoch': 0.06}) +[2026-06-24 10:08:41,633][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.22496025264263153, 'train/grad_norm_pre_clip': 42.6112174987793, 'timing/data': 0.00039045303128659725, 'timing/model': 1.0608062311075628, 'learning_rate/qwen_vl_interface': 9.285181530493567e-06, 'learning_rate/action_model': 9.285181530493567e-05, 'epoch': 0.06}) +[2026-06-24 10:08:42,698][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.14376193284988403, 'train/grad_norm_pre_clip': 46.35854721069336, 'timing/data': 0.0004072540905326605, 'timing/model': 1.063504314981401, 'learning_rate/qwen_vl_interface': 9.282464442738846e-06, 'learning_rate/action_model': 9.282464442738845e-05, 'epoch': 0.06}) +[2026-06-24 10:08:43,769][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.18664194643497467, 'train/grad_norm_pre_clip': 29.112594604492188, 'timing/data': 0.00039439997635781765, 'timing/model': 1.0687674039509147, 'learning_rate/qwen_vl_interface': 9.27974262266249e-06, 'learning_rate/action_model': 9.279742622662489e-05, 'epoch': 0.06}) +[2026-06-24 10:08:44,836][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.18541018664836884, 'train/grad_norm_pre_clip': 25.325077056884766, 'timing/data': 0.0003873370587825775, 'timing/model': 1.0657343310303986, 'learning_rate/qwen_vl_interface': 9.27701607345871e-06, 'learning_rate/action_model': 9.27701607345871e-05, 'epoch': 0.06}) +[2026-06-24 10:08:45,903][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.20749692618846893, 'train/grad_norm_pre_clip': 42.97318649291992, 'timing/data': 0.0004487019032239914, 'timing/model': 1.06544225406833, 'learning_rate/qwen_vl_interface': 9.274284798327267e-06, 'learning_rate/action_model': 9.274284798327266e-05, 'epoch': 0.06}) +[2026-06-24 10:08:46,969][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.19515405595302582, 'train/grad_norm_pre_clip': 48.187469482421875, 'timing/data': 0.00041416799649596214, 'timing/model': 1.0642090290784836, 'learning_rate/qwen_vl_interface': 9.27154880047346e-06, 'learning_rate/action_model': 9.27154880047346e-05, 'epoch': 0.06}) +[2026-06-24 10:08:48,336][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.09017804265022278, 'train/grad_norm_pre_clip': 23.20121955871582, 'timing/data': 0.0003745430149137974, 'timing/model': 1.3649736531078815, 'learning_rate/qwen_vl_interface': 9.268808083108134e-06, 'learning_rate/action_model': 9.268808083108133e-05, 'epoch': 0.06}) +[2026-06-24 10:08:49,403][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.19133001565933228, 'train/grad_norm_pre_clip': 47.387351989746094, 'timing/data': 0.00038512307219207287, 'timing/model': 1.0657201958820224, 'learning_rate/qwen_vl_interface': 9.266062649447678e-06, 'learning_rate/action_model': 9.266062649447678e-05, 'epoch': 0.06}) +[2026-06-24 10:08:50,470][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.16795825958251953, 'train/grad_norm_pre_clip': 33.35250473022461, 'timing/data': 0.0003755649086087942, 'timing/model': 1.0643241370562464, 'learning_rate/qwen_vl_interface': 9.263312502714006e-06, 'learning_rate/action_model': 9.263312502714005e-05, 'epoch': 0.06}) +[2026-06-24 10:08:51,534][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.2668618857860565, 'train/grad_norm_pre_clip': 49.475250244140625, 'timing/data': 0.0003885990008711815, 'timing/model': 1.0630569679196924, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.06}) +[2026-06-24 10:08:52,606][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.11946109682321548, 'train/grad_norm_pre_clip': 28.722824096679688, 'timing/data': 0.0004330030642449856, 'timing/model': 1.0691798611078411, 'learning_rate/qwen_vl_interface': 9.257798082942355e-06, 'learning_rate/action_model': 9.257798082942354e-05, 'epoch': 0.06}) +[2026-06-24 10:08:53,675][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.30026310682296753, 'train/grad_norm_pre_clip': 49.4389762878418, 'timing/data': 0.00039474107325077057, 'timing/model': 1.0671651491429657, 'learning_rate/qwen_vl_interface': 9.25503381637585e-06, 'learning_rate/action_model': 9.25503381637585e-05, 'epoch': 0.06}) +[2026-06-24 10:08:54,740][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.21491998434066772, 'train/grad_norm_pre_clip': 28.74881362915039, 'timing/data': 0.00038163596764206886, 'timing/model': 1.0633603481110185, 'learning_rate/qwen_vl_interface': 9.252264849679083e-06, 'learning_rate/action_model': 9.252264849679083e-05, 'epoch': 0.06}) +[2026-06-24 10:08:55,808][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.06560239940881729, 'train/grad_norm_pre_clip': 13.8109769821167, 'timing/data': 0.0004743398167192936, 'timing/model': 1.066481296904385, 'learning_rate/qwen_vl_interface': 9.24949118610159e-06, 'learning_rate/action_model': 9.249491186101591e-05, 'epoch': 0.06}) +[2026-06-24 10:08:56,876][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.1657687872648239, 'train/grad_norm_pre_clip': 27.961410522460938, 'timing/data': 0.0004298770800232887, 'timing/model': 1.0664215050637722, 'learning_rate/qwen_vl_interface': 9.24671282889842e-06, 'learning_rate/action_model': 9.24671282889842e-05, 'epoch': 0.06}) +[2026-06-24 10:08:57,941][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.1653752624988556, 'train/grad_norm_pre_clip': 27.371627807617188, 'timing/data': 0.0005261478945612907, 'timing/model': 1.0622866249177605, 'learning_rate/qwen_vl_interface': 9.243929781330127e-06, 'learning_rate/action_model': 9.243929781330127e-05, 'epoch': 0.06}) +[2026-06-24 10:08:59,010][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.14319752156734467, 'train/grad_norm_pre_clip': 33.81909942626953, 'timing/data': 0.0007120969239622355, 'timing/model': 1.066687525017187, 'learning_rate/qwen_vl_interface': 9.241142046662776e-06, 'learning_rate/action_model': 9.241142046662776e-05, 'epoch': 0.06}) +[2026-06-24 10:09:00,074][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.10133001208305359, 'train/grad_norm_pre_clip': 20.92100715637207, 'timing/data': 0.0004099400248378515, 'timing/model': 1.062848879955709, 'learning_rate/qwen_vl_interface': 9.238349628167926e-06, 'learning_rate/action_model': 9.238349628167926e-05, 'epoch': 0.06}) +[2026-06-24 10:09:01,144][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.1240348070859909, 'train/grad_norm_pre_clip': 31.137304306030273, 'timing/data': 0.00046894093975424767, 'timing/model': 1.0677144629880786, 'learning_rate/qwen_vl_interface': 9.235552529122636e-06, 'learning_rate/action_model': 9.235552529122636e-05, 'epoch': 0.06}) +[2026-06-24 10:09:02,213][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.24426813423633575, 'train/grad_norm_pre_clip': 34.038516998291016, 'timing/data': 0.00042116991244256496, 'timing/model': 1.067321346141398, 'learning_rate/qwen_vl_interface': 9.23275075280946e-06, 'learning_rate/action_model': 9.23275075280946e-05, 'epoch': 0.06}) +[2026-06-24 10:09:03,285][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.340302437543869, 'train/grad_norm_pre_clip': 34.25285720825195, 'timing/data': 0.0003837600816041231, 'timing/model': 1.0699761181604117, 'learning_rate/qwen_vl_interface': 9.229944302516436e-06, 'learning_rate/action_model': 9.229944302516437e-05, 'epoch': 0.06}) +[2026-06-24 10:09:04,352][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.11045211553573608, 'train/grad_norm_pre_clip': 57.530029296875, 'timing/data': 0.0004448050167411566, 'timing/model': 1.0654825549572706, 'learning_rate/qwen_vl_interface': 9.227133181537091e-06, 'learning_rate/action_model': 9.227133181537091e-05, 'epoch': 0.06}) +[2026-06-24 10:09:05,729][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.18021640181541443, 'train/grad_norm_pre_clip': 34.902740478515625, 'timing/data': 0.00036148889921605587, 'timing/model': 1.3753234660252929, 'learning_rate/qwen_vl_interface': 9.224317393170431e-06, 'learning_rate/action_model': 9.22431739317043e-05, 'epoch': 0.06}) +[2026-06-24 10:09:06,799][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.24651870131492615, 'train/grad_norm_pre_clip': 43.9653434753418, 'timing/data': 0.0006160959601402283, 'timing/model': 1.0670392368920147, 'learning_rate/qwen_vl_interface': 9.221496940720943e-06, 'learning_rate/action_model': 9.221496940720942e-05, 'epoch': 0.06}) +[2026-06-24 10:09:07,864][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.08213920146226883, 'train/grad_norm_pre_clip': 19.375810623168945, 'timing/data': 0.0005770430434495211, 'timing/model': 1.0629581271205097, 'learning_rate/qwen_vl_interface': 9.218671827498581e-06, 'learning_rate/action_model': 9.218671827498581e-05, 'epoch': 0.06}) +[2026-06-24 10:09:08,933][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.11236470937728882, 'train/grad_norm_pre_clip': 26.426538467407227, 'timing/data': 0.0007774990517646074, 'timing/model': 1.0672124230768532, 'learning_rate/qwen_vl_interface': 9.215842056818776e-06, 'learning_rate/action_model': 9.215842056818776e-05, 'epoch': 0.06}) +[2026-06-24 10:09:10,002][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.16110892593860626, 'train/grad_norm_pre_clip': 35.580223083496094, 'timing/data': 0.0005870119202882051, 'timing/model': 1.0666610777843744, 'learning_rate/qwen_vl_interface': 9.213007632002421e-06, 'learning_rate/action_model': 9.213007632002421e-05, 'epoch': 0.06}) +[2026-06-24 10:09:11,073][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.09650582075119019, 'train/grad_norm_pre_clip': 28.78752899169922, 'timing/data': 0.00042534782551229, 'timing/model': 1.06878571677953, 'learning_rate/qwen_vl_interface': 9.21016855637587e-06, 'learning_rate/action_model': 9.21016855637587e-05, 'epoch': 0.06}) +[2026-06-24 10:09:12,141][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.20532675087451935, 'train/grad_norm_pre_clip': 61.58097839355469, 'timing/data': 0.00042768288403749466, 'timing/model': 1.06624043895863, 'learning_rate/qwen_vl_interface': 9.207324833270937e-06, 'learning_rate/action_model': 9.207324833270936e-05, 'epoch': 0.06}) +[2026-06-24 10:09:13,211][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.20373331010341644, 'train/grad_norm_pre_clip': 26.988271713256836, 'timing/data': 0.0006140829063951969, 'timing/model': 1.0677406860049814, 'learning_rate/qwen_vl_interface': 9.20447646602489e-06, 'learning_rate/action_model': 9.20447646602489e-05, 'epoch': 0.06}) +[2026-06-24 10:09:14,279][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.3085838854312897, 'train/grad_norm_pre_clip': 57.40180587768555, 'timing/data': 0.0004387130029499531, 'timing/model': 1.06634549703449, 'learning_rate/qwen_vl_interface': 9.201623457980447e-06, 'learning_rate/action_model': 9.201623457980446e-05, 'epoch': 0.06}) +[2026-06-24 10:09:15,344][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.2200743854045868, 'train/grad_norm_pre_clip': 43.39204406738281, 'timing/data': 0.00038735708221793175, 'timing/model': 1.0633624990005046, 'learning_rate/qwen_vl_interface': 9.19876581248577e-06, 'learning_rate/action_model': 9.198765812485768e-05, 'epoch': 0.06}) +[2026-06-24 10:09:16,412][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.15842153131961823, 'train/grad_norm_pre_clip': 37.11257553100586, 'timing/data': 0.000418785959482193, 'timing/model': 1.0659577599726617, 'learning_rate/qwen_vl_interface': 9.195903532894466e-06, 'learning_rate/action_model': 9.195903532894465e-05, 'epoch': 0.06}) +[2026-06-24 10:09:17,479][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.2729610800743103, 'train/grad_norm_pre_clip': 50.7436637878418, 'timing/data': 0.00041643204167485237, 'timing/model': 1.0649377962108701, 'learning_rate/qwen_vl_interface': 9.19303662256558e-06, 'learning_rate/action_model': 9.193036622565581e-05, 'epoch': 0.06}) +[2026-06-24 10:09:18,545][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.14245255291461945, 'train/grad_norm_pre_clip': 25.711841583251953, 'timing/data': 0.0003879680298268795, 'timing/model': 1.0638404870405793, 'learning_rate/qwen_vl_interface': 9.190165084863592e-06, 'learning_rate/action_model': 9.190165084863592e-05, 'epoch': 0.06}) +[2026-06-24 10:09:19,614][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.13498124480247498, 'train/grad_norm_pre_clip': 34.781734466552734, 'timing/data': 0.00044675893150269985, 'timing/model': 1.0678223809227347, 'learning_rate/qwen_vl_interface': 9.18728892315841e-06, 'learning_rate/action_model': 9.187288923158411e-05, 'epoch': 0.06}) +[2026-06-24 10:09:20,684][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.25770512223243713, 'train/grad_norm_pre_clip': 51.815189361572266, 'timing/data': 0.00042966706678271294, 'timing/model': 1.0675152749754488, 'learning_rate/qwen_vl_interface': 9.18440814082537e-06, 'learning_rate/action_model': 9.18440814082537e-05, 'epoch': 0.06}) +[2026-06-24 10:09:21,751][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.1848422884941101, 'train/grad_norm_pre_clip': 45.36585235595703, 'timing/data': 0.00039211614057421684, 'timing/model': 1.0657793970312923, 'learning_rate/qwen_vl_interface': 9.18152274124523e-06, 'learning_rate/action_model': 9.181522741245231e-05, 'epoch': 0.06}) +[2026-06-24 10:09:23,117][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.08924239873886108, 'train/grad_norm_pre_clip': 22.788381576538086, 'timing/data': 0.000374282943084836, 'timing/model': 1.364243004936725, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.06}) +[2026-06-24 10:09:24,185][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.12324056029319763, 'train/grad_norm_pre_clip': 35.81110763549805, 'timing/data': 0.0004524989053606987, 'timing/model': 1.0659348401241004, 'learning_rate/qwen_vl_interface': 9.175738103893778e-06, 'learning_rate/action_model': 9.175738103893777e-05, 'epoch': 0.06}) +[2026-06-24 10:09:25,253][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.09759645164012909, 'train/grad_norm_pre_clip': 35.92542266845703, 'timing/data': 0.00040386803448200226, 'timing/model': 1.066621950129047, 'learning_rate/qwen_vl_interface': 9.172838872911058e-06, 'learning_rate/action_model': 9.172838872911059e-05, 'epoch': 0.06}) +[2026-06-24 10:09:26,322][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.4461829364299774, 'train/grad_norm_pre_clip': 85.34402465820312, 'timing/data': 0.00037279981188476086, 'timing/model': 1.0665035878773779, 'learning_rate/qwen_vl_interface': 9.169935038258419e-06, 'learning_rate/action_model': 9.169935038258418e-05, 'epoch': 0.06}) +[2026-06-24 10:09:27,390][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.2968544661998749, 'train/grad_norm_pre_clip': 36.04384231567383, 'timing/data': 0.00043003703467547894, 'timing/model': 1.0663764090277255, 'learning_rate/qwen_vl_interface': 9.167026603343672e-06, 'learning_rate/action_model': 9.167026603343672e-05, 'epoch': 0.06}) +[2026-06-24 10:09:28,457][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.2876479923725128, 'train/grad_norm_pre_clip': 67.93130493164062, 'timing/data': 0.00041045108810067177, 'timing/model': 1.064622808014974, 'learning_rate/qwen_vl_interface': 9.164113571580028e-06, 'learning_rate/action_model': 9.164113571580029e-05, 'epoch': 0.06}) +[2026-06-24 10:09:29,527][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.18663819134235382, 'train/grad_norm_pre_clip': 39.199493408203125, 'timing/data': 0.00047906883992254734, 'timing/model': 1.0682480428367853, 'learning_rate/qwen_vl_interface': 9.161195946386096e-06, 'learning_rate/action_model': 9.161195946386096e-05, 'epoch': 0.06}) +[2026-06-24 10:09:30,593][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.28017714619636536, 'train/grad_norm_pre_clip': 82.1026382446289, 'timing/data': 0.00039397901855409145, 'timing/model': 1.0646011279895902, 'learning_rate/qwen_vl_interface': 9.158273731185863e-06, 'learning_rate/action_model': 9.158273731185864e-05, 'epoch': 0.06}) +[2026-06-24 10:09:31,663][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.07553254812955856, 'train/grad_norm_pre_clip': 22.9671573638916, 'timing/data': 0.0003875067923218012, 'timing/model': 1.0672418449539691, 'learning_rate/qwen_vl_interface': 9.15534692940872e-06, 'learning_rate/action_model': 9.155346929408719e-05, 'epoch': 0.06}) +[2026-06-24 10:09:32,730][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.24574430286884308, 'train/grad_norm_pre_clip': 30.65032958984375, 'timing/data': 0.0004087470006197691, 'timing/model': 1.0656753340736032, 'learning_rate/qwen_vl_interface': 9.152415544489426e-06, 'learning_rate/action_model': 9.152415544489426e-05, 'epoch': 0.06}) +[2026-06-24 10:09:33,796][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.36030638217926025, 'train/grad_norm_pre_clip': 45.351863861083984, 'timing/data': 0.0004342959728091955, 'timing/model': 1.064443413168192, 'learning_rate/qwen_vl_interface': 9.149479579868127e-06, 'learning_rate/action_model': 9.149479579868127e-05, 'epoch': 0.06}) +[2026-06-24 10:09:34,862][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.25190919637680054, 'train/grad_norm_pre_clip': 30.33871841430664, 'timing/data': 0.00044429395347833633, 'timing/model': 1.0635987578425556, 'learning_rate/qwen_vl_interface': 9.146539038990339e-06, 'learning_rate/action_model': 9.14653903899034e-05, 'epoch': 0.06}) +[2026-06-24 10:09:35,929][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.14364975690841675, 'train/grad_norm_pre_clip': 26.202392578125, 'timing/data': 0.0004757130518555641, 'timing/model': 1.065335578052327, 'learning_rate/qwen_vl_interface': 9.143593925306953e-06, 'learning_rate/action_model': 9.143593925306952e-05, 'epoch': 0.06}) +[2026-06-24 10:09:36,994][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.06063689664006233, 'train/grad_norm_pre_clip': 24.44887924194336, 'timing/data': 0.0003886700142174959, 'timing/model': 1.063040821114555, 'learning_rate/qwen_vl_interface': 9.140644242274224e-06, 'learning_rate/action_model': 9.140644242274224e-05, 'epoch': 0.06}) +[2026-06-24 10:09:38,057][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.06933901458978653, 'train/grad_norm_pre_clip': 15.703929901123047, 'timing/data': 0.000384502112865448, 'timing/model': 1.0613101930357516, 'learning_rate/qwen_vl_interface': 9.137689993353766e-06, 'learning_rate/action_model': 9.137689993353766e-05, 'epoch': 0.06}) +[2026-06-24 10:09:39,124][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.24644234776496887, 'train/grad_norm_pre_clip': 29.938215255737305, 'timing/data': 0.0004336440470069647, 'timing/model': 1.0642261661123484, 'learning_rate/qwen_vl_interface': 9.13473118201256e-06, 'learning_rate/action_model': 9.134731182012559e-05, 'epoch': 0.06}) +[2026-06-24 10:09:40,484][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.14664892852306366, 'train/grad_norm_pre_clip': 35.32201385498047, 'timing/data': 0.0004512770101428032, 'timing/model': 1.358597124926746, 'learning_rate/qwen_vl_interface': 9.131767811722931e-06, 'learning_rate/action_model': 9.131767811722931e-05, 'epoch': 0.06}) +[2026-06-24 10:09:41,550][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.21732895076274872, 'train/grad_norm_pre_clip': 26.33037757873535, 'timing/data': 0.00039857812225818634, 'timing/model': 1.0639204930048436, 'learning_rate/qwen_vl_interface': 9.128799885962562e-06, 'learning_rate/action_model': 9.128799885962561e-05, 'epoch': 0.06}) +[2026-06-24 10:09:42,615][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.2553374171257019, 'train/grad_norm_pre_clip': 57.32278060913086, 'timing/data': 0.0004013730213046074, 'timing/model': 1.0638930730056018, 'learning_rate/qwen_vl_interface': 9.125827408214477e-06, 'learning_rate/action_model': 9.125827408214477e-05, 'epoch': 0.06}) +[2026-06-24 10:09:43,688][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.15118394792079926, 'train/grad_norm_pre_clip': 33.55316162109375, 'timing/data': 0.00038285902701318264, 'timing/model': 1.0709286879282445, 'learning_rate/qwen_vl_interface': 9.122850381967046e-06, 'learning_rate/action_model': 9.122850381967046e-05, 'epoch': 0.06}) +[2026-06-24 10:09:44,754][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.09657801687717438, 'train/grad_norm_pre_clip': 41.21065139770508, 'timing/data': 0.00038034399040043354, 'timing/model': 1.0635788538493216, 'learning_rate/qwen_vl_interface': 9.119868810713975e-06, 'learning_rate/action_model': 9.119868810713974e-05, 'epoch': 0.06}) +[2026-06-24 10:09:45,822][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.24346429109573364, 'train/grad_norm_pre_clip': 48.86281967163086, 'timing/data': 0.0004015939775854349, 'timing/model': 1.0664605528581887, 'learning_rate/qwen_vl_interface': 9.116882697954303e-06, 'learning_rate/action_model': 9.116882697954302e-05, 'epoch': 0.07}) +[2026-06-24 10:09:46,890][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.24942399561405182, 'train/grad_norm_pre_clip': 64.73307037353516, 'timing/data': 0.00039468100294470787, 'timing/model': 1.065520809032023, 'learning_rate/qwen_vl_interface': 9.113892047192399e-06, 'learning_rate/action_model': 9.113892047192399e-05, 'epoch': 0.07}) +[2026-06-24 10:09:47,958][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.15768173336982727, 'train/grad_norm_pre_clip': 37.9062614440918, 'timing/data': 0.00038118590600788593, 'timing/model': 1.0660912501625717, 'learning_rate/qwen_vl_interface': 9.110896861937961e-06, 'learning_rate/action_model': 9.11089686193796e-05, 'epoch': 0.07}) +[2026-06-24 10:09:49,024][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.28515902161598206, 'train/grad_norm_pre_clip': 58.17974090576172, 'timing/data': 0.0003911340609192848, 'timing/model': 1.064569303067401, 'learning_rate/qwen_vl_interface': 9.107897145706004e-06, 'learning_rate/action_model': 9.107897145706004e-05, 'epoch': 0.07}) +[2026-06-24 10:09:50,094][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.1392817199230194, 'train/grad_norm_pre_clip': 42.5040397644043, 'timing/data': 0.000521508976817131, 'timing/model': 1.0678216889500618, 'learning_rate/qwen_vl_interface': 9.104892902016862e-06, 'learning_rate/action_model': 9.104892902016861e-05, 'epoch': 0.07}) +[2026-06-24 10:09:51,214][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.19988073408603668, 'train/grad_norm_pre_clip': 54.73093032836914, 'timing/data': 0.00038494309410452843, 'timing/model': 1.068116030888632, 'learning_rate/qwen_vl_interface': 9.101884134396181e-06, 'learning_rate/action_model': 9.10188413439618e-05, 'epoch': 0.07}) +[2026-06-24 10:09:52,291][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.2590196132659912, 'train/grad_norm_pre_clip': 72.86620330810547, 'timing/data': 0.00046425103209912777, 'timing/model': 1.0751046780496836, 'learning_rate/qwen_vl_interface': 9.098870846374921e-06, 'learning_rate/action_model': 9.09887084637492e-05, 'epoch': 0.07}) +[2026-06-24 10:09:53,357][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.14078550040721893, 'train/grad_norm_pre_clip': 42.7447395324707, 'timing/data': 0.000409228028729558, 'timing/model': 1.0642323920037597, 'learning_rate/qwen_vl_interface': 9.09585304148934e-06, 'learning_rate/action_model': 9.09585304148934e-05, 'epoch': 0.07}) +[2026-06-24 10:09:54,422][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.09823599457740784, 'train/grad_norm_pre_clip': 51.462100982666016, 'timing/data': 0.00043277209624648094, 'timing/model': 1.0626451231073588, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.07}) +[2026-06-24 10:09:55,491][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.1017477884888649, 'train/grad_norm_pre_clip': 37.499725341796875, 'timing/data': 0.00042288308031857014, 'timing/model': 1.0672257009427994, 'learning_rate/qwen_vl_interface': 9.089803895296763e-06, 'learning_rate/action_model': 9.089803895296764e-05, 'epoch': 0.07}) +[2026-06-24 10:09:56,561][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.2863571047782898, 'train/grad_norm_pre_clip': 45.661376953125, 'timing/data': 0.0004025660455226898, 'timing/model': 1.0685540148988366, 'learning_rate/qwen_vl_interface': 9.086772561088779e-06, 'learning_rate/action_model': 9.086772561088779e-05, 'epoch': 0.07}) +[2026-06-24 10:09:57,937][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.2128467708826065, 'train/grad_norm_pre_clip': 50.38974380493164, 'timing/data': 0.00043856306001544, 'timing/model': 1.3737603710033, 'learning_rate/qwen_vl_interface': 9.083736724214485e-06, 'learning_rate/action_model': 9.083736724214484e-05, 'epoch': 0.07}) +[2026-06-24 10:09:59,005][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.1012718677520752, 'train/grad_norm_pre_clip': 27.632736206054688, 'timing/data': 0.0004617059603333473, 'timing/model': 1.0662191328592598, 'learning_rate/qwen_vl_interface': 9.080696388236608e-06, 'learning_rate/action_model': 9.080696388236606e-05, 'epoch': 0.07}) +[2026-06-24 10:10:00,075][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.1386660635471344, 'train/grad_norm_pre_clip': 40.51942443847656, 'timing/data': 0.00040259608067572117, 'timing/model': 1.0679066800512373, 'learning_rate/qwen_vl_interface': 9.077651556723149e-06, 'learning_rate/action_model': 9.077651556723148e-05, 'epoch': 0.07}) +[2026-06-24 10:10:01,143][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.1669730246067047, 'train/grad_norm_pre_clip': 28.16607666015625, 'timing/data': 0.00047040311619639397, 'timing/model': 1.065410204930231, 'learning_rate/qwen_vl_interface': 9.074602233247388e-06, 'learning_rate/action_model': 9.074602233247388e-05, 'epoch': 0.07}) +[2026-06-24 10:10:02,212][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.1521940380334854, 'train/grad_norm_pre_clip': 27.450401306152344, 'timing/data': 0.0005126220639795065, 'timing/model': 1.0673926859162748, 'learning_rate/qwen_vl_interface': 9.071548421387877e-06, 'learning_rate/action_model': 9.071548421387878e-05, 'epoch': 0.07}) +[2026-06-24 10:10:03,282][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.16341954469680786, 'train/grad_norm_pre_clip': 42.23155975341797, 'timing/data': 0.0005103279836475849, 'timing/model': 1.068294829921797, 'learning_rate/qwen_vl_interface': 9.068490124728435e-06, 'learning_rate/action_model': 9.068490124728435e-05, 'epoch': 0.07}) +[2026-06-24 10:10:04,346][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.1703982800245285, 'train/grad_norm_pre_clip': 36.952049255371094, 'timing/data': 0.00044554704800248146, 'timing/model': 1.0618229329120368, 'learning_rate/qwen_vl_interface': 9.06542734685814e-06, 'learning_rate/action_model': 9.06542734685814e-05, 'epoch': 0.07}) +[2026-06-24 10:10:05,412][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.13385850191116333, 'train/grad_norm_pre_clip': 44.2890739440918, 'timing/data': 0.000358393182978034, 'timing/model': 1.065083181951195, 'learning_rate/qwen_vl_interface': 9.062360091371337e-06, 'learning_rate/action_model': 9.062360091371336e-05, 'epoch': 0.07}) +[2026-06-24 10:10:06,478][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.15505386888980865, 'train/grad_norm_pre_clip': 23.794614791870117, 'timing/data': 0.00042989710345864296, 'timing/model': 1.0639214920811355, 'learning_rate/qwen_vl_interface': 9.059288361867619e-06, 'learning_rate/action_model': 9.059288361867618e-05, 'epoch': 0.07}) +[2026-06-24 10:10:07,541][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.16843578219413757, 'train/grad_norm_pre_clip': 35.041080474853516, 'timing/data': 0.0003800331614911556, 'timing/model': 1.0615822419058532, 'learning_rate/qwen_vl_interface': 9.05621216195183e-06, 'learning_rate/action_model': 9.05621216195183e-05, 'epoch': 0.07}) +[2026-06-24 10:10:08,608][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.2547789216041565, 'train/grad_norm_pre_clip': 47.299774169921875, 'timing/data': 0.0004011730197817087, 'timing/model': 1.0648096199147403, 'learning_rate/qwen_vl_interface': 9.053131495234063e-06, 'learning_rate/action_model': 9.053131495234063e-05, 'epoch': 0.07}) +[2026-06-24 10:10:09,675][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.10887983441352844, 'train/grad_norm_pre_clip': 23.302627563476562, 'timing/data': 0.00038820807822048664, 'timing/model': 1.0654489600565284, 'learning_rate/qwen_vl_interface': 9.050046365329653e-06, 'learning_rate/action_model': 9.050046365329653e-05, 'epoch': 0.07}) +[2026-06-24 10:10:10,742][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.22592779994010925, 'train/grad_norm_pre_clip': 32.170570373535156, 'timing/data': 0.0003683308605104685, 'timing/model': 1.0644045490771532, 'learning_rate/qwen_vl_interface': 9.046956775859172e-06, 'learning_rate/action_model': 9.04695677585917e-05, 'epoch': 0.07}) +[2026-06-24 10:10:11,809][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.20019564032554626, 'train/grad_norm_pre_clip': 35.13528823852539, 'timing/data': 0.00039011193439364433, 'timing/model': 1.065295052016154, 'learning_rate/qwen_vl_interface': 9.043862730448423e-06, 'learning_rate/action_model': 9.043862730448423e-05, 'epoch': 0.07}) +[2026-06-24 10:10:12,868][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.4001776874065399, 'train/grad_norm_pre_clip': 62.18913269042969, 'timing/data': 0.0003363010473549366, 'timing/model': 1.0583390560932457, 'learning_rate/qwen_vl_interface': 9.040764232728443e-06, 'learning_rate/action_model': 9.040764232728442e-05, 'epoch': 0.07}) +[2026-06-24 10:10:13,930][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.25020354986190796, 'train/grad_norm_pre_clip': 36.35923767089844, 'timing/data': 0.0003756950609385967, 'timing/model': 1.0603126001078635, 'learning_rate/qwen_vl_interface': 9.037661286335491e-06, 'learning_rate/action_model': 9.037661286335492e-05, 'epoch': 0.07}) +[2026-06-24 10:10:15,285][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.10059253871440887, 'train/grad_norm_pre_clip': 18.041330337524414, 'timing/data': 0.0003083879128098488, 'timing/model': 1.3527242939453572, 'learning_rate/qwen_vl_interface': 9.034553894911047e-06, 'learning_rate/action_model': 9.034553894911047e-05, 'epoch': 0.07}) +[2026-06-24 10:10:16,347][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.28666579723358154, 'train/grad_norm_pre_clip': 69.4823226928711, 'timing/data': 0.0003107539378106594, 'timing/model': 1.0610100699122995, 'learning_rate/qwen_vl_interface': 9.031442062101809e-06, 'learning_rate/action_model': 9.031442062101808e-05, 'epoch': 0.07}) +[2026-06-24 10:10:17,409][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.2420467585325241, 'train/grad_norm_pre_clip': 36.31398391723633, 'timing/data': 0.0003559580072760582, 'timing/model': 1.059639978921041, 'learning_rate/qwen_vl_interface': 9.028325791559685e-06, 'learning_rate/action_model': 9.028325791559684e-05, 'epoch': 0.07}) +[2026-06-24 10:10:18,469][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.4255056381225586, 'train/grad_norm_pre_clip': 59.24941635131836, 'timing/data': 0.0003187190741300583, 'timing/model': 1.0586019910406321, 'learning_rate/qwen_vl_interface': 9.025205086941791e-06, 'learning_rate/action_model': 9.025205086941792e-05, 'epoch': 0.07}) +[2026-06-24 10:10:19,533][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.11308732628822327, 'train/grad_norm_pre_clip': 31.69129753112793, 'timing/data': 0.00032125297002494335, 'timing/model': 1.0623743820469826, 'learning_rate/qwen_vl_interface': 9.02207995191045e-06, 'learning_rate/action_model': 9.022079951910449e-05, 'epoch': 0.07}) +[2026-06-24 10:10:20,596][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.09460118412971497, 'train/grad_norm_pre_clip': 34.218231201171875, 'timing/data': 0.00035115890204906464, 'timing/model': 1.061588275944814, 'learning_rate/qwen_vl_interface': 9.018950390133182e-06, 'learning_rate/action_model': 9.018950390133181e-05, 'epoch': 0.07}) +[2026-06-24 10:10:21,660][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.1476486474275589, 'train/grad_norm_pre_clip': 24.86903190612793, 'timing/data': 0.00031404895707964897, 'timing/model': 1.0628082470502704, 'learning_rate/qwen_vl_interface': 9.015816405282698e-06, 'learning_rate/action_model': 9.015816405282699e-05, 'epoch': 0.07}) +[2026-06-24 10:10:22,728][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.18193240463733673, 'train/grad_norm_pre_clip': 61.56758117675781, 'timing/data': 0.0003517910372465849, 'timing/model': 1.0658633809071034, 'learning_rate/qwen_vl_interface': 9.012678001036911e-06, 'learning_rate/action_model': 9.01267800103691e-05, 'epoch': 0.07}) +[2026-06-24 10:10:23,806][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.22658373415470123, 'train/grad_norm_pre_clip': 36.16624450683594, 'timing/data': 0.0005089049227535725, 'timing/model': 1.0762225850485265, 'learning_rate/qwen_vl_interface': 9.009535181078907e-06, 'learning_rate/action_model': 9.009535181078906e-05, 'epoch': 0.07}) +[2026-06-24 10:10:24,869][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.28464123606681824, 'train/grad_norm_pre_clip': 37.3398323059082, 'timing/data': 0.0004869841504842043, 'timing/model': 1.0608667028136551, 'learning_rate/qwen_vl_interface': 9.006387949096962e-06, 'learning_rate/action_model': 9.006387949096962e-05, 'epoch': 0.07}) +[2026-06-24 10:10:25,946][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.1451157033443451, 'train/grad_norm_pre_clip': 59.17252731323242, 'timing/data': 0.00029602600261569023, 'timing/model': 1.0759874638170004, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.07}) +[2026-06-24 10:10:27,011][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.22083047032356262, 'train/grad_norm_pre_clip': 43.38853073120117, 'timing/data': 0.00032982882112264633, 'timing/model': 1.0633220518939197, 'learning_rate/qwen_vl_interface': 9.000080263840233e-06, 'learning_rate/action_model': 9.000080263840233e-05, 'epoch': 0.07}) +[2026-06-24 10:10:28,080][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.289118230342865, 'train/grad_norm_pre_clip': 37.2775993347168, 'timing/data': 0.0002948329783976078, 'timing/model': 1.0665517651941627, 'learning_rate/qwen_vl_interface': 8.996919817967868e-06, 'learning_rate/action_model': 8.996919817967869e-05, 'epoch': 0.07}) +[2026-06-24 10:10:29,144][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.24977324903011322, 'train/grad_norm_pre_clip': 28.82154083251953, 'timing/data': 0.00040074181742966175, 'timing/model': 1.062631927896291, 'learning_rate/qwen_vl_interface': 8.993754974876395e-06, 'learning_rate/action_model': 8.993754974876394e-05, 'epoch': 0.07}) +[2026-06-24 10:10:30,210][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.12320127338171005, 'train/grad_norm_pre_clip': 25.170066833496094, 'timing/data': 0.00031226594001054764, 'timing/model': 1.0643643999937922, 'learning_rate/qwen_vl_interface': 8.990585738279931e-06, 'learning_rate/action_model': 8.990585738279932e-05, 'epoch': 0.07}) +[2026-06-24 10:10:31,272][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.11479922384023666, 'train/grad_norm_pre_clip': 48.87566375732422, 'timing/data': 0.0003430440556257963, 'timing/model': 1.0601833721157163, 'learning_rate/qwen_vl_interface': 8.987412111897754e-06, 'learning_rate/action_model': 8.987412111897753e-05, 'epoch': 0.07}) +[2026-06-24 10:10:32,619][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.24683073163032532, 'train/grad_norm_pre_clip': 77.69639587402344, 'timing/data': 0.0003076069988310337, 'timing/model': 1.345308090094477, 'learning_rate/qwen_vl_interface': 8.984234099454291e-06, 'learning_rate/action_model': 8.98423409945429e-05, 'epoch': 0.07}) +[2026-06-24 10:10:33,683][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.2057877779006958, 'train/grad_norm_pre_clip': 33.88911437988281, 'timing/data': 0.00030178693123161793, 'timing/model': 1.0621977138798684, 'learning_rate/qwen_vl_interface': 8.981051704679115e-06, 'learning_rate/action_model': 8.981051704679115e-05, 'epoch': 0.07}) +[2026-06-24 10:10:34,747][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.10183598846197128, 'train/grad_norm_pre_clip': 16.72803497314453, 'timing/data': 0.00035074795596301556, 'timing/model': 1.063144121086225, 'learning_rate/qwen_vl_interface': 8.977864931306943e-06, 'learning_rate/action_model': 8.977864931306942e-05, 'epoch': 0.07}) +[2026-06-24 10:10:35,811][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.1473056972026825, 'train/grad_norm_pre_clip': 40.55564880371094, 'timing/data': 0.0003368719480931759, 'timing/model': 1.062173037789762, 'learning_rate/qwen_vl_interface': 8.974673783077635e-06, 'learning_rate/action_model': 8.974673783077634e-05, 'epoch': 0.07}) +[2026-06-24 10:10:36,877][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.0468454472720623, 'train/grad_norm_pre_clip': 17.137855529785156, 'timing/data': 0.00033200299367308617, 'timing/model': 1.0638860929757357, 'learning_rate/qwen_vl_interface': 8.971478263736177e-06, 'learning_rate/action_model': 8.971478263736177e-05, 'epoch': 0.07}) +[2026-06-24 10:10:37,939][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.39352938532829285, 'train/grad_norm_pre_clip': 87.28341674804688, 'timing/data': 0.0003179160412400961, 'timing/model': 1.061253973050043, 'learning_rate/qwen_vl_interface': 8.96827837703269e-06, 'learning_rate/action_model': 8.968278377032689e-05, 'epoch': 0.07}) +[2026-06-24 10:10:39,005][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.054753731936216354, 'train/grad_norm_pre_clip': 22.277936935424805, 'timing/data': 0.0003006539773195982, 'timing/model': 1.0635126640554518, 'learning_rate/qwen_vl_interface': 8.965074126722421e-06, 'learning_rate/action_model': 8.965074126722421e-05, 'epoch': 0.07}) +[2026-06-24 10:10:40,070][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.10656610876321793, 'train/grad_norm_pre_clip': 29.185287475585938, 'timing/data': 0.00032381806522607803, 'timing/model': 1.0643509370274842, 'learning_rate/qwen_vl_interface': 8.961865516565735e-06, 'learning_rate/action_model': 8.961865516565734e-05, 'epoch': 0.07}) +[2026-06-24 10:10:41,132][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.063119076192379, 'train/grad_norm_pre_clip': 15.383298873901367, 'timing/data': 0.0005112390499562025, 'timing/model': 1.0592546109110117, 'learning_rate/qwen_vl_interface': 8.958652550328115e-06, 'learning_rate/action_model': 8.958652550328115e-05, 'epoch': 0.07}) +[2026-06-24 10:10:42,199][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.15479676425457, 'train/grad_norm_pre_clip': 39.12721252441406, 'timing/data': 0.00029421201907098293, 'timing/model': 1.0655186690855771, 'learning_rate/qwen_vl_interface': 8.955435231780157e-06, 'learning_rate/action_model': 8.955435231780158e-05, 'epoch': 0.07}) +[2026-06-24 10:10:43,270][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.25386226177215576, 'train/grad_norm_pre_clip': 59.4669303894043, 'timing/data': 0.00031702499836683273, 'timing/model': 1.0700228142086416, 'learning_rate/qwen_vl_interface': 8.952213564697561e-06, 'learning_rate/action_model': 8.95221356469756e-05, 'epoch': 0.07}) +[2026-06-24 10:10:44,334][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.27599072456359863, 'train/grad_norm_pre_clip': 135.5463409423828, 'timing/data': 0.0005061498377472162, 'timing/model': 1.061702135950327, 'learning_rate/qwen_vl_interface': 8.948987552861133e-06, 'learning_rate/action_model': 8.948987552861133e-05, 'epoch': 0.07}) +[2026-06-24 10:10:45,401][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.029896734282374382, 'train/grad_norm_pre_clip': 27.98242950439453, 'timing/data': 0.00039903796277940273, 'timing/model': 1.0655048340559006, 'learning_rate/qwen_vl_interface': 8.94575720005678e-06, 'learning_rate/action_model': 8.94575720005678e-05, 'epoch': 0.07}) +[2026-06-24 10:10:46,465][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.20896655321121216, 'train/grad_norm_pre_clip': 68.34446716308594, 'timing/data': 0.0003810350317507982, 'timing/model': 1.0621714279986918, 'learning_rate/qwen_vl_interface': 8.9425225100755e-06, 'learning_rate/action_model': 8.942522510075499e-05, 'epoch': 0.07}) +[2026-06-24 10:10:47,540][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.03750667721033096, 'train/grad_norm_pre_clip': 35.61151123046875, 'timing/data': 0.0003177160397171974, 'timing/model': 1.073547218926251, 'learning_rate/qwen_vl_interface': 8.939283486713381e-06, 'learning_rate/action_model': 8.939283486713381e-05, 'epoch': 0.07}) +[2026-06-24 10:10:48,605][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.16261619329452515, 'train/grad_norm_pre_clip': 55.236995697021484, 'timing/data': 0.0003646539989858866, 'timing/model': 1.0631728079169989, 'learning_rate/qwen_vl_interface': 8.936040133771598e-06, 'learning_rate/action_model': 8.936040133771597e-05, 'epoch': 0.07}) +[2026-06-24 10:10:49,953][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.07444463670253754, 'train/grad_norm_pre_clip': 64.65838623046875, 'timing/data': 0.00033120205625891685, 'timing/model': 1.346721368143335, 'learning_rate/qwen_vl_interface': 8.932792455056404e-06, 'learning_rate/action_model': 8.932792455056403e-05, 'epoch': 0.07}) +[2026-06-24 10:10:51,019][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.19257120788097382, 'train/grad_norm_pre_clip': 80.7317886352539, 'timing/data': 0.0003294888883829117, 'timing/model': 1.0642036229837686, 'learning_rate/qwen_vl_interface': 8.929540454379135e-06, 'learning_rate/action_model': 8.929540454379135e-05, 'epoch': 0.07}) +[2026-06-24 10:10:52,083][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.13839091360569, 'train/grad_norm_pre_clip': 102.32997131347656, 'timing/data': 0.00034374487586319447, 'timing/model': 1.0625548588577658, 'learning_rate/qwen_vl_interface': 8.926284135556192e-06, 'learning_rate/action_model': 8.926284135556192e-05, 'epoch': 0.07}) +[2026-06-24 10:10:53,149][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.27397996187210083, 'train/grad_norm_pre_clip': 107.33868408203125, 'timing/data': 0.000296636950224638, 'timing/model': 1.0643412719946355, 'learning_rate/qwen_vl_interface': 8.923023502409046e-06, 'learning_rate/action_model': 8.923023502409046e-05, 'epoch': 0.07}) +[2026-06-24 10:10:54,217][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.2936801612377167, 'train/grad_norm_pre_clip': 103.24356842041016, 'timing/data': 0.00029556499794125557, 'timing/model': 1.0660431361757219, 'learning_rate/qwen_vl_interface': 8.919758558764238e-06, 'learning_rate/action_model': 8.919758558764238e-05, 'epoch': 0.07}) +[2026-06-24 10:10:55,280][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.22415301203727722, 'train/grad_norm_pre_clip': 86.55689239501953, 'timing/data': 0.0003177060279995203, 'timing/model': 1.0610191570594907, 'learning_rate/qwen_vl_interface': 8.916489308453351e-06, 'learning_rate/action_model': 8.916489308453352e-05, 'epoch': 0.07}) +[2026-06-24 10:10:56,343][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.3528769016265869, 'train/grad_norm_pre_clip': 62.19960021972656, 'timing/data': 0.000297127990052104, 'timing/model': 1.0621860779356211, 'learning_rate/qwen_vl_interface': 8.913215755313042e-06, 'learning_rate/action_model': 8.913215755313041e-05, 'epoch': 0.07}) +[2026-06-24 10:10:57,406][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.03717152774333954, 'train/grad_norm_pre_clip': 22.04212188720703, 'timing/data': 0.0004815240390598774, 'timing/model': 1.0605686220806092, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.07}) +[2026-06-24 10:10:58,469][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.09964341670274734, 'train/grad_norm_pre_clip': 87.7959213256836, 'timing/data': 0.00031035207211971283, 'timing/model': 1.061532462015748, 'learning_rate/qwen_vl_interface': 8.906655755915972e-06, 'learning_rate/action_model': 8.906655755915971e-05, 'epoch': 0.07}) +[2026-06-24 10:10:59,534][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.08019126206636429, 'train/grad_norm_pre_clip': 24.60810089111328, 'timing/data': 0.0003027678467333317, 'timing/model': 1.0635796138085425, 'learning_rate/qwen_vl_interface': 8.903369317357738e-06, 'learning_rate/action_model': 8.903369317357739e-05, 'epoch': 0.07}) +[2026-06-24 10:11:00,598][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.14415867626667023, 'train/grad_norm_pre_clip': 53.23588562011719, 'timing/data': 0.00030619511380791664, 'timing/model': 1.0621554718818516, 'learning_rate/qwen_vl_interface': 8.900078591367119e-06, 'learning_rate/action_model': 8.900078591367119e-05, 'epoch': 0.07}) +[2026-06-24 10:11:01,668][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.30043312907218933, 'train/grad_norm_pre_clip': 35.22005081176758, 'timing/data': 0.0003051131498068571, 'timing/model': 1.0688974759541452, 'learning_rate/qwen_vl_interface': 8.896783581805963e-06, 'learning_rate/action_model': 8.896783581805962e-05, 'epoch': 0.07}) +[2026-06-24 10:11:02,734][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.17504248023033142, 'train/grad_norm_pre_clip': 56.10230255126953, 'timing/data': 0.00047487090341746807, 'timing/model': 1.0642402151133865, 'learning_rate/qwen_vl_interface': 8.89348429254115e-06, 'learning_rate/action_model': 8.893484292541149e-05, 'epoch': 0.07}) +[2026-06-24 10:11:03,798][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.1724931001663208, 'train/grad_norm_pre_clip': 58.57585525512695, 'timing/data': 0.000302337110042572, 'timing/model': 1.0622899050358683, 'learning_rate/qwen_vl_interface': 8.890180727444577e-06, 'learning_rate/action_model': 8.890180727444577e-05, 'epoch': 0.07}) +[2026-06-24 10:11:04,860][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.08222856372594833, 'train/grad_norm_pre_clip': 30.869821548461914, 'timing/data': 0.000333316158503294, 'timing/model': 1.0595788159407675, 'learning_rate/qwen_vl_interface': 8.886872890393163e-06, 'learning_rate/action_model': 8.886872890393163e-05, 'epoch': 0.07}) +[2026-06-24 10:11:05,921][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.30181416869163513, 'train/grad_norm_pre_clip': 56.442501068115234, 'timing/data': 0.0003017059061676264, 'timing/model': 1.060053087072447, 'learning_rate/qwen_vl_interface': 8.88356078526884e-06, 'learning_rate/action_model': 8.88356078526884e-05, 'epoch': 0.07}) +[2026-06-24 10:11:07,271][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.1661365032196045, 'train/grad_norm_pre_clip': 40.69758605957031, 'timing/data': 0.00038310885429382324, 'timing/model': 1.3481877050362527, 'learning_rate/qwen_vl_interface': 8.880244415958545e-06, 'learning_rate/action_model': 8.880244415958543e-05, 'epoch': 0.07}) +[2026-06-24 10:11:08,337][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.09680221974849701, 'train/grad_norm_pre_clip': 16.347089767456055, 'timing/data': 0.0003254308830946684, 'timing/model': 1.0643145160283893, 'learning_rate/qwen_vl_interface': 8.876923786354223e-06, 'learning_rate/action_model': 8.876923786354223e-05, 'epoch': 0.07}) +[2026-06-24 10:11:09,398][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.15179364383220673, 'train/grad_norm_pre_clip': 39.82396697998047, 'timing/data': 0.00034214300103485584, 'timing/model': 1.0591667839325964, 'learning_rate/qwen_vl_interface': 8.873598900352821e-06, 'learning_rate/action_model': 8.873598900352821e-05, 'epoch': 0.07}) +[2026-06-24 10:11:10,463][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 750 +[2026-06-24 10:11:35,335][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 750 +[2026-06-24 10:11:35,336][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 750 +[2026-06-24 10:12:06,343][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 750 +[2026-06-24 10:13:40,524][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 750 +[2026-06-24 10:13:40,524][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.08943832665681839, 'train/grad_norm_pre_clip': 22.441404342651367, 'eval/action_loss/seconds': 24.870539451017976, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.08851198852062225, 'eval/latency_2/loss': 0.08851198852062225, 'eval/flappy/loss': 0.08851198852062225, 'eval/flappy/latency_2/loss': 0.08851198852062225, 'eval/flap_precision': 0.9171122994652406, 'eval/flap_recall': 0.9308005427408412, 'eval/flap_f1': 0.9239057239057239, 'eval/flap_rate_model': 0.20777777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9239057239057238, 'eval/latency_2/flap_precision': 0.9171122994652406, 'eval/latency_2/flap_recall': 0.9308005427408412, 'eval/latency_2/flap_f1': 0.9239057239057239, 'eval/latency_2/flap_rate_model': 0.20777777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9239057239057238, 'eval/action_classification/data_wait_seconds': 0.3319501280784607, 'eval/action_classification/predict_seconds': 30.58116340637207, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 72.7599999999988, 'rl_games_eval/mid_train/mean_length': 607.2, 'rl_games_eval/mid_train/std_reward': 66.34210126307238, 'rl_games_eval/mid_train/std_length': 535.8131763964003, 'rl_games_eval/mid_train/macro_mean_reward': 72.7599999999988, 'rl_games_eval/mid_train/macro_mean_length': 607.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 72.7599999999988, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 607.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 66.34210126307238, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 535.8131763964003, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 72.7599999999988, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000367951113730669, 'timing/model': 1.0638448749668896, 'learning_rate/qwen_vl_interface': 8.870269761856274e-06, 'learning_rate/action_model': 8.870269761856274e-05, 'epoch': 0.07}) +[2026-06-24 10:13:41,575][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.051522016525268555, 'train/grad_norm_pre_clip': 16.031997680664062, 'timing/data': 0.0005487599410116673, 'timing/model': 1.0486091480124742, 'learning_rate/qwen_vl_interface': 8.866936374771513e-06, 'learning_rate/action_model': 8.866936374771512e-05, 'epoch': 0.07}) +[2026-06-24 10:13:42,624][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.08812564611434937, 'train/grad_norm_pre_clip': 29.707443237304688, 'timing/data': 0.0003071569371968508, 'timing/model': 1.0476562201511115, 'learning_rate/qwen_vl_interface': 8.863598743010454e-06, 'learning_rate/action_model': 8.863598743010454e-05, 'epoch': 0.07}) +[2026-06-24 10:13:43,674][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.12583675980567932, 'train/grad_norm_pre_clip': 23.828908920288086, 'timing/data': 0.00030132592655718327, 'timing/model': 1.0480733229778707, 'learning_rate/qwen_vl_interface': 8.860256870489992e-06, 'learning_rate/action_model': 8.860256870489991e-05, 'epoch': 0.07}) +[2026-06-24 10:13:44,972][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.2555057406425476, 'train/grad_norm_pre_clip': 68.65760803222656, 'timing/data': 0.00030518206767737865, 'timing/model': 1.2962955159600824, 'learning_rate/qwen_vl_interface': 8.856910761132003e-06, 'learning_rate/action_model': 8.856910761132003e-05, 'epoch': 0.07}) +[2026-06-24 10:13:46,023][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.21663330495357513, 'train/grad_norm_pre_clip': 327.901611328125, 'timing/data': 0.0003440349828451872, 'timing/model': 1.0500270500779152, 'learning_rate/qwen_vl_interface': 8.85356041886333e-06, 'learning_rate/action_model': 8.853560418863329e-05, 'epoch': 0.07}) +[2026-06-24 10:13:47,076][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.13595348596572876, 'train/grad_norm_pre_clip': 37.54692077636719, 'timing/data': 0.00030425190925598145, 'timing/model': 1.050879481015727, 'learning_rate/qwen_vl_interface': 8.850205847615788e-06, 'learning_rate/action_model': 8.850205847615788e-05, 'epoch': 0.07}) +[2026-06-24 10:13:48,128][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.22573190927505493, 'train/grad_norm_pre_clip': 40.93901443481445, 'timing/data': 0.0003166839014738798, 'timing/model': 1.0499559380114079, 'learning_rate/qwen_vl_interface': 8.846847051326155e-06, 'learning_rate/action_model': 8.846847051326154e-05, 'epoch': 0.07}) +[2026-06-24 10:13:49,181][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.323434054851532, 'train/grad_norm_pre_clip': 56.9083251953125, 'timing/data': 0.00036791106685996056, 'timing/model': 1.0521717958617955, 'learning_rate/qwen_vl_interface': 8.84348403393616e-06, 'learning_rate/action_model': 8.84348403393616e-05, 'epoch': 0.07}) +[2026-06-24 10:13:50,236][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.19166012108325958, 'train/grad_norm_pre_clip': 66.45032501220703, 'timing/data': 0.00032766489312052727, 'timing/model': 1.0533858451526612, 'learning_rate/qwen_vl_interface': 8.840116799392496e-06, 'learning_rate/action_model': 8.840116799392496e-05, 'epoch': 0.07}) +[2026-06-24 10:13:51,292][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.06615529954433441, 'train/grad_norm_pre_clip': 54.5138053894043, 'timing/data': 0.0003379140980541706, 'timing/model': 1.0540315881371498, 'learning_rate/qwen_vl_interface': 8.8367453516468e-06, 'learning_rate/action_model': 8.8367453516468e-05, 'epoch': 0.07}) +[2026-06-24 10:13:52,344][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.18186774849891663, 'train/grad_norm_pre_clip': 38.81146240234375, 'timing/data': 0.0004805619828402996, 'timing/model': 1.0500509061384946, 'learning_rate/qwen_vl_interface': 8.83336969465565e-06, 'learning_rate/action_model': 8.833369694655649e-05, 'epoch': 0.07}) +[2026-06-24 10:13:53,398][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.37983763217926025, 'train/grad_norm_pre_clip': 76.7624740600586, 'timing/data': 0.00031569204293191433, 'timing/model': 1.0527723338454962, 'learning_rate/qwen_vl_interface': 8.829989832380572e-06, 'learning_rate/action_model': 8.829989832380571e-05, 'epoch': 0.07}) +[2026-06-24 10:13:54,454][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.1560828983783722, 'train/grad_norm_pre_clip': 45.464744567871094, 'timing/data': 0.00033471803180873394, 'timing/model': 1.053649501176551, 'learning_rate/qwen_vl_interface': 8.826605768788021e-06, 'learning_rate/action_model': 8.826605768788021e-05, 'epoch': 0.07}) +[2026-06-24 10:13:55,510][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.24383395910263062, 'train/grad_norm_pre_clip': 46.35108947753906, 'timing/data': 0.00032544019632041454, 'timing/model': 1.0551232569850981, 'learning_rate/qwen_vl_interface': 8.823217507849383e-06, 'learning_rate/action_model': 8.823217507849382e-05, 'epoch': 0.07}) +[2026-06-24 10:13:56,562][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.24363161623477936, 'train/grad_norm_pre_clip': 37.95401382446289, 'timing/data': 0.0003765369765460491, 'timing/model': 1.0502610420808196, 'learning_rate/qwen_vl_interface': 8.819825053540973e-06, 'learning_rate/action_model': 8.819825053540972e-05, 'epoch': 0.07}) +[2026-06-24 10:13:57,619][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.14977794885635376, 'train/grad_norm_pre_clip': 30.24323844909668, 'timing/data': 0.00029814010486006737, 'timing/model': 1.0547798029147089, 'learning_rate/qwen_vl_interface': 8.816428409844023e-06, 'learning_rate/action_model': 8.816428409844023e-05, 'epoch': 0.07}) +[2026-06-24 10:13:58,673][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.4773661494255066, 'train/grad_norm_pre_clip': 75.65673065185547, 'timing/data': 0.00032369792461395264, 'timing/model': 1.0528257158584893, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.07}) +[2026-06-24 10:13:59,728][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.2761763036251068, 'train/grad_norm_pre_clip': 54.193115234375, 'timing/data': 0.0003058239817619324, 'timing/model': 1.0533263550605625, 'learning_rate/qwen_vl_interface': 8.809622570234018e-06, 'learning_rate/action_model': 8.809622570234019e-05, 'epoch': 0.07}) +[2026-06-24 10:14:00,784][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.12594814598560333, 'train/grad_norm_pre_clip': 53.08004379272461, 'timing/data': 0.000335498945787549, 'timing/model': 1.0546359329018742, 'learning_rate/qwen_vl_interface': 8.806213382307997e-06, 'learning_rate/action_model': 8.806213382307996e-05, 'epoch': 0.07}) +[2026-06-24 10:14:01,839][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.1236313059926033, 'train/grad_norm_pre_clip': 57.7183723449707, 'timing/data': 0.00031244708225131035, 'timing/model': 1.0534965449478477, 'learning_rate/qwen_vl_interface': 8.80280002096749e-06, 'learning_rate/action_model': 8.802800020967489e-05, 'epoch': 0.07}) +[2026-06-24 10:14:03,108][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.26093682646751404, 'train/grad_norm_pre_clip': 34.805782318115234, 'timing/data': 0.0003015659749507904, 'timing/model': 1.2672190149314702, 'learning_rate/qwen_vl_interface': 8.79938249021827e-06, 'learning_rate/action_model': 8.79938249021827e-05, 'epoch': 0.07}) +[2026-06-24 10:14:04,169][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.07271162420511246, 'train/grad_norm_pre_clip': 13.4285888671875, 'timing/data': 0.00030040391720831394, 'timing/model': 1.0591287571005523, 'learning_rate/qwen_vl_interface': 8.795960794070995e-06, 'learning_rate/action_model': 8.795960794070996e-05, 'epoch': 0.07}) +[2026-06-24 10:14:05,226][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.2662340998649597, 'train/grad_norm_pre_clip': 54.2889518737793, 'timing/data': 0.00032665301114320755, 'timing/model': 1.0552886789664626, 'learning_rate/qwen_vl_interface': 8.792534936541222e-06, 'learning_rate/action_model': 8.79253493654122e-05, 'epoch': 0.07}) +[2026-06-24 10:14:06,284][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.18880292773246765, 'train/grad_norm_pre_clip': 58.44517135620117, 'timing/data': 0.0003638230264186859, 'timing/model': 1.056173259858042, 'learning_rate/qwen_vl_interface': 8.78910492164938e-06, 'learning_rate/action_model': 8.78910492164938e-05, 'epoch': 0.07}) +[2026-06-24 10:14:07,342][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.2216208428144455, 'train/grad_norm_pre_clip': 41.22932434082031, 'timing/data': 0.0003142401110380888, 'timing/model': 1.0564927300438285, 'learning_rate/qwen_vl_interface': 8.785670753420784e-06, 'learning_rate/action_model': 8.785670753420785e-05, 'epoch': 0.07}) +[2026-06-24 10:14:08,400][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.10001818835735321, 'train/grad_norm_pre_clip': 14.60571575164795, 'timing/data': 0.0003110240213572979, 'timing/model': 1.056327270111069, 'learning_rate/qwen_vl_interface': 8.782232435885624e-06, 'learning_rate/action_model': 8.782232435885625e-05, 'epoch': 0.08}) +[2026-06-24 10:14:09,458][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.15299229323863983, 'train/grad_norm_pre_clip': 27.789358139038086, 'timing/data': 0.0003286469727754593, 'timing/model': 1.0568330690730363, 'learning_rate/qwen_vl_interface': 8.778789973078954e-06, 'learning_rate/action_model': 8.778789973078955e-05, 'epoch': 0.08}) +[2026-06-24 10:14:10,515][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.09594019502401352, 'train/grad_norm_pre_clip': 22.164051055908203, 'timing/data': 0.0003110740799456835, 'timing/model': 1.0552166448906064, 'learning_rate/qwen_vl_interface': 8.7753433690407e-06, 'learning_rate/action_model': 8.775343369040699e-05, 'epoch': 0.08}) +[2026-06-24 10:14:11,572][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.29433709383010864, 'train/grad_norm_pre_clip': 61.91766357421875, 'timing/data': 0.00031518214382231236, 'timing/model': 1.0553997089155018, 'learning_rate/qwen_vl_interface': 8.77189262781564e-06, 'learning_rate/action_model': 8.771892627815639e-05, 'epoch': 0.08}) +[2026-06-24 10:14:12,630][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.1664838343858719, 'train/grad_norm_pre_clip': 45.8039665222168, 'timing/data': 0.000305041903629899, 'timing/model': 1.056486097862944, 'learning_rate/qwen_vl_interface': 8.768437753453411e-06, 'learning_rate/action_model': 8.768437753453411e-05, 'epoch': 0.08}) +[2026-06-24 10:14:13,688][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.45358702540397644, 'train/grad_norm_pre_clip': 57.684974670410156, 'timing/data': 0.00034844386391341686, 'timing/model': 1.0564709298778325, 'learning_rate/qwen_vl_interface': 8.764978750008504e-06, 'learning_rate/action_model': 8.764978750008504e-05, 'epoch': 0.08}) +[2026-06-24 10:14:14,749][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.06747402995824814, 'train/grad_norm_pre_clip': 37.08496856689453, 'timing/data': 0.0003596050664782524, 'timing/model': 1.0593663051258773, 'learning_rate/qwen_vl_interface': 8.76151562154025e-06, 'learning_rate/action_model': 8.761515621540248e-05, 'epoch': 0.08}) +[2026-06-24 10:14:15,810][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.21286079287528992, 'train/grad_norm_pre_clip': 352.9805908203125, 'timing/data': 0.00030142604373395443, 'timing/model': 1.058857660042122, 'learning_rate/qwen_vl_interface': 8.75804837211282e-06, 'learning_rate/action_model': 8.75804837211282e-05, 'epoch': 0.08}) +[2026-06-24 10:14:16,868][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.2913268506526947, 'train/grad_norm_pre_clip': 64.83881378173828, 'timing/data': 0.0003029489889740944, 'timing/model': 1.0570705158170313, 'learning_rate/qwen_vl_interface': 8.754577005795231e-06, 'learning_rate/action_model': 8.75457700579523e-05, 'epoch': 0.08}) +[2026-06-24 10:14:17,926][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.15460576117038727, 'train/grad_norm_pre_clip': 53.25919723510742, 'timing/data': 0.0003353399224579334, 'timing/model': 1.05592779093422, 'learning_rate/qwen_vl_interface': 8.751101526661318e-06, 'learning_rate/action_model': 8.751101526661318e-05, 'epoch': 0.08}) +[2026-06-24 10:14:18,985][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.10868865996599197, 'train/grad_norm_pre_clip': 87.70791625976562, 'timing/data': 0.00031549204140901566, 'timing/model': 1.057950858026743, 'learning_rate/qwen_vl_interface': 8.747621938789755e-06, 'learning_rate/action_model': 8.747621938789754e-05, 'epoch': 0.08}) +[2026-06-24 10:14:20,256][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.38118913769721985, 'train/grad_norm_pre_clip': 58.57842254638672, 'timing/data': 0.00030320906080305576, 'timing/model': 1.2691591100301594, 'learning_rate/qwen_vl_interface': 8.74413824626403e-06, 'learning_rate/action_model': 8.74413824626403e-05, 'epoch': 0.08}) +[2026-06-24 10:14:21,316][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.1836681067943573, 'train/grad_norm_pre_clip': 102.1000747680664, 'timing/data': 0.0003375930245965719, 'timing/model': 1.058026660932228, 'learning_rate/qwen_vl_interface': 8.74065045317245e-06, 'learning_rate/action_model': 8.740650453172449e-05, 'epoch': 0.08}) +[2026-06-24 10:14:22,374][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.11331448704004288, 'train/grad_norm_pre_clip': 38.045989990234375, 'timing/data': 0.0003491649404168129, 'timing/model': 1.0571358199231327, 'learning_rate/qwen_vl_interface': 8.737158563608132e-06, 'learning_rate/action_model': 8.737158563608132e-05, 'epoch': 0.08}) +[2026-06-24 10:14:23,436][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.13189969956874847, 'train/grad_norm_pre_clip': 50.30278778076172, 'timing/data': 0.00048636295832693577, 'timing/model': 1.0599893059115857, 'learning_rate/qwen_vl_interface': 8.733662581669007e-06, 'learning_rate/action_model': 8.733662581669006e-05, 'epoch': 0.08}) +[2026-06-24 10:14:24,497][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.14707046747207642, 'train/grad_norm_pre_clip': 67.14712524414062, 'timing/data': 0.00034178211353719234, 'timing/model': 1.0594091371167451, 'learning_rate/qwen_vl_interface': 8.730162511457803e-06, 'learning_rate/action_model': 8.730162511457803e-05, 'epoch': 0.08}) +[2026-06-24 10:14:25,559][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.11895078420639038, 'train/grad_norm_pre_clip': 37.72611618041992, 'timing/data': 0.0003048719372600317, 'timing/model': 1.0603371881879866, 'learning_rate/qwen_vl_interface': 8.726658357082047e-06, 'learning_rate/action_model': 8.726658357082047e-05, 'epoch': 0.08}) +[2026-06-24 10:14:26,615][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.19344447553157806, 'train/grad_norm_pre_clip': 73.97100067138672, 'timing/data': 0.00031462102197110653, 'timing/model': 1.0537430020049214, 'learning_rate/qwen_vl_interface': 8.723150122654058e-06, 'learning_rate/action_model': 8.723150122654058e-05, 'epoch': 0.08}) +[2026-06-24 10:14:27,674][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.18746882677078247, 'train/grad_norm_pre_clip': 60.34880065917969, 'timing/data': 0.0003318928647786379, 'timing/model': 1.0576265710406005, 'learning_rate/qwen_vl_interface': 8.719637812290949e-06, 'learning_rate/action_model': 8.71963781229095e-05, 'epoch': 0.08}) +[2026-06-24 10:14:28,737][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.10907687246799469, 'train/grad_norm_pre_clip': 39.6787109375, 'timing/data': 0.0004847492091357708, 'timing/model': 1.0610333259683102, 'learning_rate/qwen_vl_interface': 8.716121430114606e-06, 'learning_rate/action_model': 8.716121430114607e-05, 'epoch': 0.08}) +[2026-06-24 10:14:29,799][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.06827172636985779, 'train/grad_norm_pre_clip': 28.118667602539062, 'timing/data': 0.00030314899049699306, 'timing/model': 1.0608150858897716, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.08}) +[2026-06-24 10:14:30,859][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.188245490193367, 'train/grad_norm_pre_clip': 37.439910888671875, 'timing/data': 0.0003248997963964939, 'timing/model': 1.0579750949982554, 'learning_rate/qwen_vl_interface': 8.709076466833686e-06, 'learning_rate/action_model': 8.709076466833686e-05, 'epoch': 0.08}) +[2026-06-24 10:14:31,924][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.3325124680995941, 'train/grad_norm_pre_clip': 126.4345932006836, 'timing/data': 0.00033830502070486546, 'timing/model': 1.0632366118952632, 'learning_rate/qwen_vl_interface': 8.705547893996764e-06, 'learning_rate/action_model': 8.705547893996763e-05, 'epoch': 0.08}) +[2026-06-24 10:14:32,986][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.44151079654693604, 'train/grad_norm_pre_clip': 76.25984191894531, 'timing/data': 0.0003306299913674593, 'timing/model': 1.0605418239720166, 'learning_rate/qwen_vl_interface': 8.702015265881914e-06, 'learning_rate/action_model': 8.702015265881913e-05, 'epoch': 0.08}) +[2026-06-24 10:14:34,047][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.375356525182724, 'train/grad_norm_pre_clip': 76.80479431152344, 'timing/data': 0.0003120661713182926, 'timing/model': 1.0597412819042802, 'learning_rate/qwen_vl_interface': 8.698478586634871e-06, 'learning_rate/action_model': 8.698478586634871e-05, 'epoch': 0.08}) +[2026-06-24 10:14:35,109][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.19395585358142853, 'train/grad_norm_pre_clip': 29.648271560668945, 'timing/data': 0.00030730688013136387, 'timing/model': 1.059901142027229, 'learning_rate/qwen_vl_interface': 8.69493786040613e-06, 'learning_rate/action_model': 8.694937860406131e-05, 'epoch': 0.08}) +[2026-06-24 10:14:36,172][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.151228129863739, 'train/grad_norm_pre_clip': 23.384403228759766, 'timing/data': 0.00032525090500712395, 'timing/model': 1.0616914120037109, 'learning_rate/qwen_vl_interface': 8.691393091350927e-06, 'learning_rate/action_model': 8.691393091350926e-05, 'epoch': 0.08}) +[2026-06-24 10:14:37,438][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.13079774379730225, 'train/grad_norm_pre_clip': 30.319110870361328, 'timing/data': 0.0003211330622434616, 'timing/model': 1.264853912172839, 'learning_rate/qwen_vl_interface': 8.687844283629244e-06, 'learning_rate/action_model': 8.687844283629243e-05, 'epoch': 0.08}) +[2026-06-24 10:14:38,502][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.38606542348861694, 'train/grad_norm_pre_clip': 66.38695526123047, 'timing/data': 0.0003117749001830816, 'timing/model': 1.0622597699984908, 'learning_rate/qwen_vl_interface': 8.684291441405807e-06, 'learning_rate/action_model': 8.684291441405807e-05, 'epoch': 0.08}) +[2026-06-24 10:14:39,565][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.4093259871006012, 'train/grad_norm_pre_clip': 58.3482780456543, 'timing/data': 0.0003423430025577545, 'timing/model': 1.0614386389497668, 'learning_rate/qwen_vl_interface': 8.680734568850076e-06, 'learning_rate/action_model': 8.680734568850076e-05, 'epoch': 0.08}) +[2026-06-24 10:14:40,626][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.14886574447155, 'train/grad_norm_pre_clip': 18.952251434326172, 'timing/data': 0.00032784510403871536, 'timing/model': 1.0595346649643034, 'learning_rate/qwen_vl_interface': 8.677173670136233e-06, 'learning_rate/action_model': 8.677173670136232e-05, 'epoch': 0.08}) +[2026-06-24 10:14:41,691][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.14903771877288818, 'train/grad_norm_pre_clip': 17.09650230407715, 'timing/data': 0.0006199139170348644, 'timing/model': 1.0628178778570145, 'learning_rate/qwen_vl_interface': 8.673608749443198e-06, 'learning_rate/action_model': 8.673608749443198e-05, 'epoch': 0.08}) +[2026-06-24 10:14:42,755][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.2240857034921646, 'train/grad_norm_pre_clip': 42.953857421875, 'timing/data': 0.0003174759913235903, 'timing/model': 1.0616377021651715, 'learning_rate/qwen_vl_interface': 8.670039810954602e-06, 'learning_rate/action_model': 8.670039810954601e-05, 'epoch': 0.08}) +[2026-06-24 10:14:43,817][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.26572030782699585, 'train/grad_norm_pre_clip': 36.76679992675781, 'timing/data': 0.0002978791017085314, 'timing/model': 1.0607484439387918, 'learning_rate/qwen_vl_interface': 8.66646685885879e-06, 'learning_rate/action_model': 8.666466858858789e-05, 'epoch': 0.08}) +[2026-06-24 10:14:44,880][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.13806922733783722, 'train/grad_norm_pre_clip': 26.714120864868164, 'timing/data': 0.0003161430358886719, 'timing/model': 1.061455181101337, 'learning_rate/qwen_vl_interface': 8.662889897348823e-06, 'learning_rate/action_model': 8.662889897348823e-05, 'epoch': 0.08}) +[2026-06-24 10:14:45,943][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.12659025192260742, 'train/grad_norm_pre_clip': 40.72016525268555, 'timing/data': 0.0003450580406934023, 'timing/model': 1.0617461970541626, 'learning_rate/qwen_vl_interface': 8.659308930622464e-06, 'learning_rate/action_model': 8.659308930622462e-05, 'epoch': 0.08}) +[2026-06-24 10:14:47,007][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.24280878901481628, 'train/grad_norm_pre_clip': 55.919822692871094, 'timing/data': 0.00030305911786854267, 'timing/model': 1.061960289021954, 'learning_rate/qwen_vl_interface': 8.655723962882176e-06, 'learning_rate/action_model': 8.655723962882175e-05, 'epoch': 0.08}) +[2026-06-24 10:14:48,070][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.18426506221294403, 'train/grad_norm_pre_clip': 35.402767181396484, 'timing/data': 0.000492024002596736, 'timing/model': 1.0616922758053988, 'learning_rate/qwen_vl_interface': 8.65213499833512e-06, 'learning_rate/action_model': 8.65213499833512e-05, 'epoch': 0.08}) +[2026-06-24 10:14:49,135][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.1274479329586029, 'train/grad_norm_pre_clip': 51.796730041503906, 'timing/data': 0.00031248596496880054, 'timing/model': 1.0632990321610123, 'learning_rate/qwen_vl_interface': 8.648542041193144e-06, 'learning_rate/action_model': 8.648542041193143e-05, 'epoch': 0.08}) +[2026-06-24 10:14:50,199][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.14811824262142181, 'train/grad_norm_pre_clip': 32.51539993286133, 'timing/data': 0.000311434967443347, 'timing/model': 1.0624287188984454, 'learning_rate/qwen_vl_interface': 8.644945095672785e-06, 'learning_rate/action_model': 8.644945095672785e-05, 'epoch': 0.08}) +[2026-06-24 10:14:51,260][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.17166391015052795, 'train/grad_norm_pre_clip': 31.791645050048828, 'timing/data': 0.0003113741986453533, 'timing/model': 1.0598212040495127, 'learning_rate/qwen_vl_interface': 8.64134416599526e-06, 'learning_rate/action_model': 8.641344165995259e-05, 'epoch': 0.08}) +[2026-06-24 10:14:52,321][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.18631590902805328, 'train/grad_norm_pre_clip': 67.67465209960938, 'timing/data': 0.00033289496786892414, 'timing/model': 1.0594308921135962, 'learning_rate/qwen_vl_interface': 8.637739256386457e-06, 'learning_rate/action_model': 8.637739256386457e-05, 'epoch': 0.08}) +[2026-06-24 10:14:53,384][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.17938579618930817, 'train/grad_norm_pre_clip': 53.556270599365234, 'timing/data': 0.0004864728543907404, 'timing/model': 1.061111977091059, 'learning_rate/qwen_vl_interface': 8.63413037107694e-06, 'learning_rate/action_model': 8.634130371076939e-05, 'epoch': 0.08}) +[2026-06-24 10:14:54,716][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.22611868381500244, 'train/grad_norm_pre_clip': 56.144371032714844, 'timing/data': 0.00030408985912799835, 'timing/model': 1.3304024678654969, 'learning_rate/qwen_vl_interface': 8.630517514301937e-06, 'learning_rate/action_model': 8.630517514301937e-05, 'epoch': 0.08}) +[2026-06-24 10:14:55,778][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.1742810159921646, 'train/grad_norm_pre_clip': 24.13681983947754, 'timing/data': 0.00031211599707603455, 'timing/model': 1.0605609640479088, 'learning_rate/qwen_vl_interface': 8.626900690301337e-06, 'learning_rate/action_model': 8.626900690301337e-05, 'epoch': 0.08}) +[2026-06-24 10:14:56,841][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.1349552869796753, 'train/grad_norm_pre_clip': 27.337696075439453, 'timing/data': 0.0004062431398779154, 'timing/model': 1.0616586438845843, 'learning_rate/qwen_vl_interface': 8.623279903319683e-06, 'learning_rate/action_model': 8.623279903319682e-05, 'epoch': 0.08}) +[2026-06-24 10:14:57,907][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.16381259262561798, 'train/grad_norm_pre_clip': 24.37071990966797, 'timing/data': 0.00031334790401160717, 'timing/model': 1.0637357330415398, 'learning_rate/qwen_vl_interface': 8.619655157606172e-06, 'learning_rate/action_model': 8.619655157606172e-05, 'epoch': 0.08}) +[2026-06-24 10:14:58,973][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.1489780694246292, 'train/grad_norm_pre_clip': 49.343162536621094, 'timing/data': 0.0003995099104940891, 'timing/model': 1.0647371129598469, 'learning_rate/qwen_vl_interface': 8.616026457414644e-06, 'learning_rate/action_model': 8.616026457414643e-05, 'epoch': 0.08}) +[2026-06-24 10:15:00,034][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.11312779784202576, 'train/grad_norm_pre_clip': 28.94881248474121, 'timing/data': 0.0003233170136809349, 'timing/model': 1.0592636892106384, 'learning_rate/qwen_vl_interface': 8.612393807003581e-06, 'learning_rate/action_model': 8.612393807003581e-05, 'epoch': 0.08}) +[2026-06-24 10:15:01,097][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.10366374999284744, 'train/grad_norm_pre_clip': 41.80065155029297, 'timing/data': 0.000312156043946743, 'timing/model': 1.0611631439533085, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.08}) +[2026-06-24 10:15:02,162][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.14423468708992004, 'train/grad_norm_pre_clip': 32.32460021972656, 'timing/data': 0.0003152319695800543, 'timing/model': 1.06348983081989, 'learning_rate/qwen_vl_interface': 8.605116672579951e-06, 'learning_rate/action_model': 8.605116672579951e-05, 'epoch': 0.08}) +[2026-06-24 10:15:03,225][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.07676377892494202, 'train/grad_norm_pre_clip': 17.410518646240234, 'timing/data': 0.00032398803159594536, 'timing/model': 1.0613808738999069, 'learning_rate/qwen_vl_interface': 8.601472197107508e-06, 'learning_rate/action_model': 8.601472197107508e-05, 'epoch': 0.08}) +[2026-06-24 10:15:04,289][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.11848894506692886, 'train/grad_norm_pre_clip': 40.46501541137695, 'timing/data': 0.00030616484582424164, 'timing/model': 1.062976849032566, 'learning_rate/qwen_vl_interface': 8.597823788495765e-06, 'learning_rate/action_model': 8.597823788495765e-05, 'epoch': 0.08}) +[2026-06-24 10:15:05,351][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.12372318655252457, 'train/grad_norm_pre_clip': 89.81388854980469, 'timing/data': 0.00029562506824731827, 'timing/model': 1.0604346580803394, 'learning_rate/qwen_vl_interface': 8.59417145102633e-06, 'learning_rate/action_model': 8.59417145102633e-05, 'epoch': 0.08}) +[2026-06-24 10:15:06,416][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.21683411300182343, 'train/grad_norm_pre_clip': 46.84357452392578, 'timing/data': 0.00031626387499272823, 'timing/model': 1.0631146379746497, 'learning_rate/qwen_vl_interface': 8.59051518898543e-06, 'learning_rate/action_model': 8.59051518898543e-05, 'epoch': 0.08}) +[2026-06-24 10:15:07,480][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.046559274196624756, 'train/grad_norm_pre_clip': 10.492793083190918, 'timing/data': 0.000310664065182209, 'timing/model': 1.062306981999427, 'learning_rate/qwen_vl_interface': 8.586855006663889e-06, 'learning_rate/action_model': 8.586855006663888e-05, 'epoch': 0.08}) +[2026-06-24 10:15:08,542][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.11696471273899078, 'train/grad_norm_pre_clip': 30.75287437438965, 'timing/data': 0.0003272038884460926, 'timing/model': 1.060770939802751, 'learning_rate/qwen_vl_interface': 8.583190908357137e-06, 'learning_rate/action_model': 8.583190908357136e-05, 'epoch': 0.08}) +[2026-06-24 10:15:09,603][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.164609894156456, 'train/grad_norm_pre_clip': 62.02176284790039, 'timing/data': 0.0003086701035499573, 'timing/model': 1.059696001932025, 'learning_rate/qwen_vl_interface': 8.579522898365192e-06, 'learning_rate/action_model': 8.579522898365192e-05, 'epoch': 0.08}) +[2026-06-24 10:15:10,879][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.049306128174066544, 'train/grad_norm_pre_clip': 34.45387268066406, 'timing/data': 0.00032926793210208416, 'timing/model': 1.273541154107079, 'learning_rate/qwen_vl_interface': 8.575850980992673e-06, 'learning_rate/action_model': 8.575850980992674e-05, 'epoch': 0.08}) +[2026-06-24 10:15:11,945][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.18870781362056732, 'train/grad_norm_pre_clip': 76.99629974365234, 'timing/data': 0.0003001729492098093, 'timing/model': 1.0645303858909756, 'learning_rate/qwen_vl_interface': 8.572175160548779e-06, 'learning_rate/action_model': 8.572175160548778e-05, 'epoch': 0.08}) +[2026-06-24 10:15:13,006][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.07133529335260391, 'train/grad_norm_pre_clip': 44.793312072753906, 'timing/data': 0.00030260789208114147, 'timing/model': 1.0600331148598343, 'learning_rate/qwen_vl_interface': 8.56849544134729e-06, 'learning_rate/action_model': 8.56849544134729e-05, 'epoch': 0.08}) +[2026-06-24 10:15:14,082][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.11374937742948532, 'train/grad_norm_pre_clip': 105.26190185546875, 'timing/data': 0.00036154803819954395, 'timing/model': 1.0744579799938947, 'learning_rate/qwen_vl_interface': 8.56481182770656e-06, 'learning_rate/action_model': 8.564811827706561e-05, 'epoch': 0.08}) +[2026-06-24 10:15:15,279][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.080748051404953, 'train/grad_norm_pre_clip': 25.05689239501953, 'timing/data': 0.0004837680608034134, 'timing/model': 1.1946441770996898, 'learning_rate/qwen_vl_interface': 8.561124323949515e-06, 'learning_rate/action_model': 8.561124323949516e-05, 'epoch': 0.08}) +[2026-06-24 10:15:16,364][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.13319872319698334, 'train/grad_norm_pre_clip': 31.608713150024414, 'timing/data': 0.00040833605453372, 'timing/model': 1.0835478131193668, 'learning_rate/qwen_vl_interface': 8.557432934403648e-06, 'learning_rate/action_model': 8.557432934403648e-05, 'epoch': 0.08}) +[2026-06-24 10:15:17,426][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.07861213386058807, 'train/grad_norm_pre_clip': 61.62672805786133, 'timing/data': 0.0003204911481589079, 'timing/model': 1.0600067761261016, 'learning_rate/qwen_vl_interface': 8.553737663401009e-06, 'learning_rate/action_model': 8.553737663401009e-05, 'epoch': 0.08}) +[2026-06-24 10:15:18,486][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.02895076386630535, 'train/grad_norm_pre_clip': 43.17729949951172, 'timing/data': 0.0003063851036131382, 'timing/model': 1.0588065129704773, 'learning_rate/qwen_vl_interface': 8.550038515278202e-06, 'learning_rate/action_model': 8.550038515278203e-05, 'epoch': 0.08}) +[2026-06-24 10:15:19,546][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.5757519006729126, 'train/grad_norm_pre_clip': 83.89476013183594, 'timing/data': 0.0002930699847638607, 'timing/model': 1.0587624511681497, 'learning_rate/qwen_vl_interface': 8.54633549437639e-06, 'learning_rate/action_model': 8.546335494376388e-05, 'epoch': 0.08}) +[2026-06-24 10:15:20,609][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.011378168128430843, 'train/grad_norm_pre_clip': 8.974638938903809, 'timing/data': 0.0003684619441628456, 'timing/model': 1.061352713033557, 'learning_rate/qwen_vl_interface': 8.542628605041265e-06, 'learning_rate/action_model': 8.542628605041265e-05, 'epoch': 0.08}) +[2026-06-24 10:15:21,670][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.08621129393577576, 'train/grad_norm_pre_clip': 34.800682067871094, 'timing/data': 0.0003152817953377962, 'timing/model': 1.0588794411160052, 'learning_rate/qwen_vl_interface': 8.538917851623075e-06, 'learning_rate/action_model': 8.538917851623075e-05, 'epoch': 0.08}) +[2026-06-24 10:15:22,732][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.1503904163837433, 'train/grad_norm_pre_clip': 55.44606399536133, 'timing/data': 0.00030753808096051216, 'timing/model': 1.0608158861286938, 'learning_rate/qwen_vl_interface': 8.535203238476594e-06, 'learning_rate/action_model': 8.535203238476594e-05, 'epoch': 0.08}) +[2026-06-24 10:15:23,796][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.0037513766437768936, 'train/grad_norm_pre_clip': 4.751160144805908, 'timing/data': 0.00032347789965569973, 'timing/model': 1.0623420809861273, 'learning_rate/qwen_vl_interface': 8.531484769961129e-06, 'learning_rate/action_model': 8.531484769961127e-05, 'epoch': 0.08}) +[2026-06-24 10:15:24,861][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.8807456493377686, 'train/grad_norm_pre_clip': 162.7905731201172, 'timing/data': 0.00030956114642322063, 'timing/model': 1.0632784280460328, 'learning_rate/qwen_vl_interface': 8.527762450440506e-06, 'learning_rate/action_model': 8.527762450440505e-05, 'epoch': 0.08}) +[2026-06-24 10:15:25,926][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.0458582378923893, 'train/grad_norm_pre_clip': 36.48650360107422, 'timing/data': 0.000300985062494874, 'timing/model': 1.0631457089912146, 'learning_rate/qwen_vl_interface': 8.524036284283077e-06, 'learning_rate/action_model': 8.524036284283076e-05, 'epoch': 0.08}) +[2026-06-24 10:15:26,991][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.10984019190073013, 'train/grad_norm_pre_clip': 71.8459243774414, 'timing/data': 0.0003333559725433588, 'timing/model': 1.0629500020295382, 'learning_rate/qwen_vl_interface': 8.520306275861702e-06, 'learning_rate/action_model': 8.520306275861702e-05, 'epoch': 0.08}) +[2026-06-24 10:15:28,334][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.3323223888874054, 'train/grad_norm_pre_clip': 135.13865661621094, 'timing/data': 0.00032982905395329, 'timing/model': 1.341740685980767, 'learning_rate/qwen_vl_interface': 8.516572429553761e-06, 'learning_rate/action_model': 8.516572429553761e-05, 'epoch': 0.08}) +[2026-06-24 10:15:29,398][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.0550568662583828, 'train/grad_norm_pre_clip': 32.433719635009766, 'timing/data': 0.00031041307374835014, 'timing/model': 1.0622583848889917, 'learning_rate/qwen_vl_interface': 8.512834749741124e-06, 'learning_rate/action_model': 8.512834749741125e-05, 'epoch': 0.08}) +[2026-06-24 10:15:30,461][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.14306049048900604, 'train/grad_norm_pre_clip': 83.16844177246094, 'timing/data': 0.0003273340407758951, 'timing/model': 1.061468621948734, 'learning_rate/qwen_vl_interface': 8.509093240810171e-06, 'learning_rate/action_model': 8.509093240810171e-05, 'epoch': 0.08}) +[2026-06-24 10:15:31,523][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.08446238189935684, 'train/grad_norm_pre_clip': 42.129302978515625, 'timing/data': 0.00032660295255482197, 'timing/model': 1.060329353902489, 'learning_rate/qwen_vl_interface': 8.505347907151766e-06, 'learning_rate/action_model': 8.505347907151766e-05, 'epoch': 0.08}) +[2026-06-24 10:15:32,588][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.24673354625701904, 'train/grad_norm_pre_clip': 78.3917007446289, 'timing/data': 0.0003029690124094486, 'timing/model': 1.063702124170959, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.08}) +[2026-06-24 10:15:33,653][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.09967058151960373, 'train/grad_norm_pre_clip': 25.147201538085938, 'timing/data': 0.0003084891941398382, 'timing/model': 1.0635138910729438, 'learning_rate/qwen_vl_interface': 8.497845783238527e-06, 'learning_rate/action_model': 8.497845783238526e-05, 'epoch': 0.08}) +[2026-06-24 10:15:34,718][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.38407257199287415, 'train/grad_norm_pre_clip': 99.07283782958984, 'timing/data': 0.0003128570970147848, 'timing/model': 1.0633990159258246, 'learning_rate/qwen_vl_interface': 8.494089001787852e-06, 'learning_rate/action_model': 8.49408900178785e-05, 'epoch': 0.08}) +[2026-06-24 10:15:35,784][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.0248581413179636, 'train/grad_norm_pre_clip': 13.386994361877441, 'timing/data': 0.00032398896291852, 'timing/model': 1.0644514421001077, 'learning_rate/qwen_vl_interface': 8.490328413218034e-06, 'learning_rate/action_model': 8.490328413218033e-05, 'epoch': 0.08}) +[2026-06-24 10:15:36,848][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.18657593429088593, 'train/grad_norm_pre_clip': 64.8883056640625, 'timing/data': 0.00033900607377290726, 'timing/model': 1.0621525370515883, 'learning_rate/qwen_vl_interface': 8.48656402194234e-06, 'learning_rate/action_model': 8.48656402194234e-05, 'epoch': 0.08}) +[2026-06-24 10:15:37,914][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.3693256080150604, 'train/grad_norm_pre_clip': 217.79962158203125, 'timing/data': 0.00031569204293191433, 'timing/model': 1.0641079668421298, 'learning_rate/qwen_vl_interface': 8.482795832378489e-06, 'learning_rate/action_model': 8.482795832378489e-05, 'epoch': 0.08}) +[2026-06-24 10:15:38,977][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.3854674994945526, 'train/grad_norm_pre_clip': 46.718238830566406, 'timing/data': 0.0003203512169420719, 'timing/model': 1.0620731578674167, 'learning_rate/qwen_vl_interface': 8.479023848948664e-06, 'learning_rate/action_model': 8.479023848948663e-05, 'epoch': 0.08}) +[2026-06-24 10:15:40,040][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.370169460773468, 'train/grad_norm_pre_clip': 51.769386291503906, 'timing/data': 0.0003322530537843704, 'timing/model': 1.0609350111335516, 'learning_rate/qwen_vl_interface': 8.475248076079494e-06, 'learning_rate/action_model': 8.475248076079493e-05, 'epoch': 0.08}) +[2026-06-24 10:15:41,103][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.2229093462228775, 'train/grad_norm_pre_clip': 66.94625854492188, 'timing/data': 0.00029886094853281975, 'timing/model': 1.0616190559230745, 'learning_rate/qwen_vl_interface': 8.471468518202061e-06, 'learning_rate/action_model': 8.471468518202062e-05, 'epoch': 0.08}) +[2026-06-24 10:15:42,167][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.2992011308670044, 'train/grad_norm_pre_clip': 63.11823654174805, 'timing/data': 0.0003108428791165352, 'timing/model': 1.06247455207631, 'learning_rate/qwen_vl_interface': 8.467685179751888e-06, 'learning_rate/action_model': 8.467685179751889e-05, 'epoch': 0.08}) +[2026-06-24 10:15:43,231][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.15050067007541656, 'train/grad_norm_pre_clip': 53.69741439819336, 'timing/data': 0.000302107073366642, 'timing/model': 1.0625652519520372, 'learning_rate/qwen_vl_interface': 8.463898065168931e-06, 'learning_rate/action_model': 8.463898065168931e-05, 'epoch': 0.08}) +[2026-06-24 10:15:44,502][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.1270400881767273, 'train/grad_norm_pre_clip': 21.574758529663086, 'timing/data': 0.0003169549163430929, 'timing/model': 1.2694094739854336, 'learning_rate/qwen_vl_interface': 8.460107178897581e-06, 'learning_rate/action_model': 8.460107178897582e-05, 'epoch': 0.08}) +[2026-06-24 10:15:45,565][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.18171954154968262, 'train/grad_norm_pre_clip': 24.035762786865234, 'timing/data': 0.000365586020052433, 'timing/model': 1.0612775150220841, 'learning_rate/qwen_vl_interface': 8.456312525386655e-06, 'learning_rate/action_model': 8.456312525386655e-05, 'epoch': 0.08}) +[2026-06-24 10:15:46,626][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.16411711275577545, 'train/grad_norm_pre_clip': 25.017335891723633, 'timing/data': 0.0003069161903113127, 'timing/model': 1.0594482528977096, 'learning_rate/qwen_vl_interface': 8.45251410908939e-06, 'learning_rate/action_model': 8.452514109089389e-05, 'epoch': 0.08}) +[2026-06-24 10:15:47,687][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.1248600110411644, 'train/grad_norm_pre_clip': 16.759153366088867, 'timing/data': 0.0002959459088742733, 'timing/model': 1.0597237111069262, 'learning_rate/qwen_vl_interface': 8.448711934463435e-06, 'learning_rate/action_model': 8.448711934463435e-05, 'epoch': 0.08}) +[2026-06-24 10:15:48,751][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.1428261250257492, 'train/grad_norm_pre_clip': 15.49489688873291, 'timing/data': 0.0006090330425649881, 'timing/model': 1.0614521750248969, 'learning_rate/qwen_vl_interface': 8.444906005970858e-06, 'learning_rate/action_model': 8.444906005970858e-05, 'epoch': 0.08}) +[2026-06-24 10:15:49,816][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.179653599858284, 'train/grad_norm_pre_clip': 26.970951080322266, 'timing/data': 0.00045589497312903404, 'timing/model': 1.0635551610030234, 'learning_rate/qwen_vl_interface': 8.44109632807813e-06, 'learning_rate/action_model': 8.44109632807813e-05, 'epoch': 0.08}) +[2026-06-24 10:15:50,880][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.0835391953587532, 'train/grad_norm_pre_clip': 27.270841598510742, 'timing/data': 0.00030622491613030434, 'timing/model': 1.062025350984186, 'learning_rate/qwen_vl_interface': 8.437282905256114e-06, 'learning_rate/action_model': 8.437282905256114e-05, 'epoch': 0.08}) +[2026-06-24 10:15:51,944][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.12947481870651245, 'train/grad_norm_pre_clip': 56.52000045776367, 'timing/data': 0.00037962314672768116, 'timing/model': 1.062832745956257, 'learning_rate/qwen_vl_interface': 8.43346574198008e-06, 'learning_rate/action_model': 8.43346574198008e-05, 'epoch': 0.08}) +[2026-06-24 10:15:53,007][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.23370099067687988, 'train/grad_norm_pre_clip': 47.32488250732422, 'timing/data': 0.0003091611433774233, 'timing/model': 1.0614108871668577, 'learning_rate/qwen_vl_interface': 8.429644842729678e-06, 'learning_rate/action_model': 8.429644842729678e-05, 'epoch': 0.08}) +[2026-06-24 10:15:54,073][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.17608043551445007, 'train/grad_norm_pre_clip': 19.937042236328125, 'timing/data': 0.00033001997508108616, 'timing/model': 1.0636788138654083, 'learning_rate/qwen_vl_interface': 8.425820211988948e-06, 'learning_rate/action_model': 8.425820211988947e-05, 'epoch': 0.08}) +[2026-06-24 10:15:55,137][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.09154535830020905, 'train/grad_norm_pre_clip': 51.910789489746094, 'timing/data': 0.000320221995934844, 'timing/model': 1.0625362989958376, 'learning_rate/qwen_vl_interface': 8.421991854246304e-06, 'learning_rate/action_model': 8.421991854246304e-05, 'epoch': 0.08}) +[2026-06-24 10:15:56,201][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.23924903571605682, 'train/grad_norm_pre_clip': 47.11777877807617, 'timing/data': 0.0003057238645851612, 'timing/model': 1.062343327794224, 'learning_rate/qwen_vl_interface': 8.418159773994543e-06, 'learning_rate/action_model': 8.418159773994543e-05, 'epoch': 0.08}) +[2026-06-24 10:15:57,265][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.03712304309010506, 'train/grad_norm_pre_clip': 10.483441352844238, 'timing/data': 0.0002921589184552431, 'timing/model': 1.0627272580750287, 'learning_rate/qwen_vl_interface': 8.41432397573082e-06, 'learning_rate/action_model': 8.414323975730819e-05, 'epoch': 0.08}) +[2026-06-24 10:15:58,329][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.22466599941253662, 'train/grad_norm_pre_clip': 66.8770980834961, 'timing/data': 0.00032054190523922443, 'timing/model': 1.0627532969228923, 'learning_rate/qwen_vl_interface': 8.410484463956657e-06, 'learning_rate/action_model': 8.410484463956656e-05, 'epoch': 0.08}) +[2026-06-24 10:15:59,394][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.26211971044540405, 'train/grad_norm_pre_clip': 45.965118408203125, 'timing/data': 0.00033947709016501904, 'timing/model': 1.0632562411483377, 'learning_rate/qwen_vl_interface': 8.406641243177937e-06, 'learning_rate/action_model': 8.406641243177938e-05, 'epoch': 0.09}) +[2026-06-24 10:16:00,671][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.11125951260328293, 'train/grad_norm_pre_clip': 89.88579559326172, 'timing/data': 0.00034257303923368454, 'timing/model': 1.2753544361330569, 'learning_rate/qwen_vl_interface': 8.402794317904896e-06, 'learning_rate/action_model': 8.402794317904897e-05, 'epoch': 0.09}) +[2026-06-24 10:16:01,732][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.008680219762027264, 'train/grad_norm_pre_clip': 4.202518463134766, 'timing/data': 0.00030592503026127815, 'timing/model': 1.0595336041878909, 'learning_rate/qwen_vl_interface': 8.398943692652114e-06, 'learning_rate/action_model': 8.398943692652113e-05, 'epoch': 0.09}) +[2026-06-24 10:16:02,797][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.2169225960969925, 'train/grad_norm_pre_clip': 91.19377899169922, 'timing/data': 0.00029775919392704964, 'timing/model': 1.0628950640093535, 'learning_rate/qwen_vl_interface': 8.39508937193851e-06, 'learning_rate/action_model': 8.39508937193851e-05, 'epoch': 0.09}) +[2026-06-24 10:16:03,861][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.007144713308662176, 'train/grad_norm_pre_clip': 6.565900802612305, 'timing/data': 0.000299111008644104, 'timing/model': 1.0630814128089696, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.09}) +[2026-06-24 10:16:04,927][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.084080770611763, 'train/grad_norm_pre_clip': 42.735164642333984, 'timing/data': 0.0003076579887419939, 'timing/model': 1.0644183231052011, 'learning_rate/qwen_vl_interface': 8.387369662226219e-06, 'learning_rate/action_model': 8.387369662226218e-05, 'epoch': 0.09}) +[2026-06-24 10:16:05,991][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.2328852415084839, 'train/grad_norm_pre_clip': 90.30184936523438, 'timing/data': 0.0003508280497044325, 'timing/model': 1.062406016048044, 'learning_rate/qwen_vl_interface': 8.383504282287038e-06, 'learning_rate/action_model': 8.383504282287038e-05, 'epoch': 0.09}) +[2026-06-24 10:16:07,060][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.2874833047389984, 'train/grad_norm_pre_clip': 136.18064880371094, 'timing/data': 0.0003241079393774271, 'timing/model': 1.0666433500591666, 'learning_rate/qwen_vl_interface': 8.379635225006048e-06, 'learning_rate/action_model': 8.379635225006047e-05, 'epoch': 0.09}) +[2026-06-24 10:16:08,124][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.13592450320720673, 'train/grad_norm_pre_clip': 59.22885513305664, 'timing/data': 0.0003058339934796095, 'timing/model': 1.0624786929693073, 'learning_rate/qwen_vl_interface': 8.3757624949238e-06, 'learning_rate/action_model': 8.3757624949238e-05, 'epoch': 0.09}) +[2026-06-24 10:16:09,188][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.04955105856060982, 'train/grad_norm_pre_clip': 29.306827545166016, 'timing/data': 0.00038541294634342194, 'timing/model': 1.0629549380391836, 'learning_rate/qwen_vl_interface': 8.371886096585162e-06, 'learning_rate/action_model': 8.37188609658516e-05, 'epoch': 0.09}) +[2026-06-24 10:16:10,252][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.06596389412879944, 'train/grad_norm_pre_clip': 37.48518753051758, 'timing/data': 0.00029134703800082207, 'timing/model': 1.0623497019987553, 'learning_rate/qwen_vl_interface': 8.368006034539298e-06, 'learning_rate/action_model': 8.368006034539299e-05, 'epoch': 0.09}) +[2026-06-24 10:16:11,316][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.06888587772846222, 'train/grad_norm_pre_clip': 58.59724807739258, 'timing/data': 0.0004823352210223675, 'timing/model': 1.061393867013976, 'learning_rate/qwen_vl_interface': 8.364122313339681e-06, 'learning_rate/action_model': 8.364122313339682e-05, 'epoch': 0.09}) +[2026-06-24 10:16:12,381][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.5759149789810181, 'train/grad_norm_pre_clip': 155.87916564941406, 'timing/data': 0.0003692731261253357, 'timing/model': 1.064078897004947, 'learning_rate/qwen_vl_interface': 8.360234937544072e-06, 'learning_rate/action_model': 8.360234937544073e-05, 'epoch': 0.09}) +[2026-06-24 10:16:13,444][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.013774189166724682, 'train/grad_norm_pre_clip': 25.754667282104492, 'timing/data': 0.00029849004931747913, 'timing/model': 1.061223667813465, 'learning_rate/qwen_vl_interface': 8.356343911714529e-06, 'learning_rate/action_model': 8.356343911714529e-05, 'epoch': 0.09}) +[2026-06-24 10:16:14,505][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.2606010138988495, 'train/grad_norm_pre_clip': 142.05450439453125, 'timing/data': 0.00032755499705672264, 'timing/model': 1.0593542899005115, 'learning_rate/qwen_vl_interface': 8.352449240417377e-06, 'learning_rate/action_model': 8.352449240417377e-05, 'epoch': 0.09}) +[2026-06-24 10:16:15,569][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.37540191411972046, 'train/grad_norm_pre_clip': 48.756874084472656, 'timing/data': 0.0003650849685072899, 'timing/model': 1.0622550349216908, 'learning_rate/qwen_vl_interface': 8.348550928223239e-06, 'learning_rate/action_model': 8.348550928223237e-05, 'epoch': 0.09}) +[2026-06-24 10:16:16,841][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.2756403088569641, 'train/grad_norm_pre_clip': 81.74715423583984, 'timing/data': 0.0005253760609775782, 'timing/model': 1.2697689039632678, 'learning_rate/qwen_vl_interface': 8.344648979706997e-06, 'learning_rate/action_model': 8.344648979706995e-05, 'epoch': 0.09}) +[2026-06-24 10:16:17,903][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.16816169023513794, 'train/grad_norm_pre_clip': 122.85664367675781, 'timing/data': 0.00030790804885327816, 'timing/model': 1.0604996599722654, 'learning_rate/qwen_vl_interface': 8.340743399447804e-06, 'learning_rate/action_model': 8.340743399447803e-05, 'epoch': 0.09}) +[2026-06-24 10:16:18,966][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.05694505572319031, 'train/grad_norm_pre_clip': 42.90341567993164, 'timing/data': 0.00047187507152557373, 'timing/model': 1.06081299809739, 'learning_rate/qwen_vl_interface': 8.336834192029079e-06, 'learning_rate/action_model': 8.336834192029078e-05, 'epoch': 0.09}) +[2026-06-24 10:16:20,027][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.3054865896701813, 'train/grad_norm_pre_clip': 41.97569274902344, 'timing/data': 0.0003183470107614994, 'timing/model': 1.0598418950103223, 'learning_rate/qwen_vl_interface': 8.332921362038494e-06, 'learning_rate/action_model': 8.332921362038493e-05, 'epoch': 0.09}) +[2026-06-24 10:16:21,088][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.08415117114782333, 'train/grad_norm_pre_clip': 29.048219680786133, 'timing/data': 0.00035411491990089417, 'timing/model': 1.059192717075348, 'learning_rate/qwen_vl_interface': 8.329004914067972e-06, 'learning_rate/action_model': 8.329004914067972e-05, 'epoch': 0.09}) +[2026-06-24 10:16:22,150][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.7049506306648254, 'train/grad_norm_pre_clip': 104.74539184570312, 'timing/data': 0.0003378449473530054, 'timing/model': 1.0602614630479366, 'learning_rate/qwen_vl_interface': 8.325084852713683e-06, 'learning_rate/action_model': 8.325084852713681e-05, 'epoch': 0.09}) +[2026-06-24 10:16:23,213][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.13915804028511047, 'train/grad_norm_pre_clip': 72.49513244628906, 'timing/data': 0.00031710509210824966, 'timing/model': 1.0610760119743645, 'learning_rate/qwen_vl_interface': 8.321161182576037e-06, 'learning_rate/action_model': 8.321161182576036e-05, 'epoch': 0.09}) +[2026-06-24 10:16:24,278][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.3138374388217926, 'train/grad_norm_pre_clip': 55.143577575683594, 'timing/data': 0.00030090496875345707, 'timing/model': 1.0640126049984246, 'learning_rate/qwen_vl_interface': 8.31723390825968e-06, 'learning_rate/action_model': 8.31723390825968e-05, 'epoch': 0.09}) +[2026-06-24 10:16:25,343][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.21830910444259644, 'train/grad_norm_pre_clip': 67.23614501953125, 'timing/data': 0.0003115150611847639, 'timing/model': 1.0627788191195577, 'learning_rate/qwen_vl_interface': 8.313303034373488e-06, 'learning_rate/action_model': 8.313303034373488e-05, 'epoch': 0.09}) +[2026-06-24 10:16:26,403][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.24844475090503693, 'train/grad_norm_pre_clip': 64.01409912109375, 'timing/data': 0.00033068982884287834, 'timing/model': 1.0591596860904247, 'learning_rate/qwen_vl_interface': 8.30936856553056e-06, 'learning_rate/action_model': 8.309368565530559e-05, 'epoch': 0.09}) +[2026-06-24 10:16:27,467][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.0677105262875557, 'train/grad_norm_pre_clip': 27.522932052612305, 'timing/data': 0.0004974841140210629, 'timing/model': 1.0621499880217016, 'learning_rate/qwen_vl_interface': 8.305430506348212e-06, 'learning_rate/action_model': 8.305430506348212e-05, 'epoch': 0.09}) +[2026-06-24 10:16:28,545][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.13113020360469818, 'train/grad_norm_pre_clip': 42.557979583740234, 'timing/data': 0.0003096119035035372, 'timing/model': 1.0609022569842637, 'learning_rate/qwen_vl_interface': 8.301488861447978e-06, 'learning_rate/action_model': 8.301488861447978e-05, 'epoch': 0.09}) +[2026-06-24 10:16:29,618][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.09104542434215546, 'train/grad_norm_pre_clip': 58.16322708129883, 'timing/data': 0.00040325685404241085, 'timing/model': 1.0713847719598562, 'learning_rate/qwen_vl_interface': 8.2975436354556e-06, 'learning_rate/action_model': 8.297543635455599e-05, 'epoch': 0.09}) +[2026-06-24 10:16:30,683][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.12244459986686707, 'train/grad_norm_pre_clip': 25.209325790405273, 'timing/data': 0.0003004239406436682, 'timing/model': 1.0627462489064783, 'learning_rate/qwen_vl_interface': 8.293594833001015e-06, 'learning_rate/action_model': 8.293594833001015e-05, 'epoch': 0.09}) +[2026-06-24 10:16:31,748][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.14943405985832214, 'train/grad_norm_pre_clip': 45.70598220825195, 'timing/data': 0.00030609406530857086, 'timing/model': 1.0635374439880252, 'learning_rate/qwen_vl_interface': 8.289642458718366e-06, 'learning_rate/action_model': 8.289642458718366e-05, 'epoch': 0.09}) +[2026-06-24 10:16:33,108][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.1325484663248062, 'train/grad_norm_pre_clip': 29.485427856445312, 'timing/data': 0.0003804850857704878, 'timing/model': 1.357859592186287, 'learning_rate/qwen_vl_interface': 8.285686517245986e-06, 'learning_rate/action_model': 8.285686517245986e-05, 'epoch': 0.09}) +[2026-06-24 10:16:34,171][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.27947932481765747, 'train/grad_norm_pre_clip': 59.437923431396484, 'timing/data': 0.00030288915149867535, 'timing/model': 1.0622346480377018, 'learning_rate/qwen_vl_interface': 8.28172701322639e-06, 'learning_rate/action_model': 8.281727013226389e-05, 'epoch': 0.09}) +[2026-06-24 10:16:35,235][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.15745337307453156, 'train/grad_norm_pre_clip': 26.867197036743164, 'timing/data': 0.0003511288668960333, 'timing/model': 1.0624370488803834, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.09}) +[2026-06-24 10:16:36,300][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.2169535756111145, 'train/grad_norm_pre_clip': 36.22825241088867, 'timing/data': 0.00047647394239902496, 'timing/model': 1.063006327021867, 'learning_rate/qwen_vl_interface': 8.273797336136521e-06, 'learning_rate/action_model': 8.27379733613652e-05, 'epoch': 0.09}) +[2026-06-24 10:16:37,364][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.2169192135334015, 'train/grad_norm_pre_clip': 20.974428176879883, 'timing/data': 0.0003328649327158928, 'timing/model': 1.0619769550394267, 'learning_rate/qwen_vl_interface': 8.269827172372166e-06, 'learning_rate/action_model': 8.269827172372165e-05, 'epoch': 0.09}) +[2026-06-24 10:16:38,429][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.2275654822587967, 'train/grad_norm_pre_clip': 44.83912658691406, 'timing/data': 0.00030992203392088413, 'timing/model': 1.0632811831310391, 'learning_rate/qwen_vl_interface': 8.26585346467242e-06, 'learning_rate/action_model': 8.26585346467242e-05, 'epoch': 0.09}) +[2026-06-24 10:16:39,493][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.21705739200115204, 'train/grad_norm_pre_clip': 32.670047760009766, 'timing/data': 0.00031056301668286324, 'timing/model': 1.0625501819886267, 'learning_rate/qwen_vl_interface': 8.261876217700651e-06, 'learning_rate/action_model': 8.26187621770065e-05, 'epoch': 0.09}) +[2026-06-24 10:16:40,555][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.22009523212909698, 'train/grad_norm_pre_clip': 34.497135162353516, 'timing/data': 0.00032232492230832577, 'timing/model': 1.061133212177083, 'learning_rate/qwen_vl_interface': 8.25789543612438e-06, 'learning_rate/action_model': 8.25789543612438e-05, 'epoch': 0.09}) +[2026-06-24 10:16:41,621][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.13626666367053986, 'train/grad_norm_pre_clip': 27.70075225830078, 'timing/data': 0.0005201660096645355, 'timing/model': 1.063575335079804, 'learning_rate/qwen_vl_interface': 8.253911124615273e-06, 'learning_rate/action_model': 8.253911124615273e-05, 'epoch': 0.09}) +[2026-06-24 10:16:42,686][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.31877002120018005, 'train/grad_norm_pre_clip': 44.0223503112793, 'timing/data': 0.0003026081249117851, 'timing/model': 1.0636213519610465, 'learning_rate/qwen_vl_interface': 8.249923287849146e-06, 'learning_rate/action_model': 8.249923287849146e-05, 'epoch': 0.09}) +[2026-06-24 10:16:43,751][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.1577579379081726, 'train/grad_norm_pre_clip': 16.59577751159668, 'timing/data': 0.00029863021336495876, 'timing/model': 1.063483933918178, 'learning_rate/qwen_vl_interface': 8.245931930505945e-06, 'learning_rate/action_model': 8.245931930505945e-05, 'epoch': 0.09}) +[2026-06-24 10:16:44,817][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.18640831112861633, 'train/grad_norm_pre_clip': 52.11782455444336, 'timing/data': 0.0004895881284028292, 'timing/model': 1.0644357409328222, 'learning_rate/qwen_vl_interface': 8.24193705726975e-06, 'learning_rate/action_model': 8.24193705726975e-05, 'epoch': 0.09}) +[2026-06-24 10:16:45,883][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.2906821072101593, 'train/grad_norm_pre_clip': 45.929443359375, 'timing/data': 0.00030590384267270565, 'timing/model': 1.064134876942262, 'learning_rate/qwen_vl_interface': 8.237938672828768e-06, 'learning_rate/action_model': 8.237938672828768e-05, 'epoch': 0.09}) +[2026-06-24 10:16:46,950][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.0669138953089714, 'train/grad_norm_pre_clip': 11.60538387298584, 'timing/data': 0.0003245789557695389, 'timing/model': 1.0650064530782402, 'learning_rate/qwen_vl_interface': 8.233936781875325e-06, 'learning_rate/action_model': 8.233936781875325e-05, 'epoch': 0.09}) +[2026-06-24 10:16:48,225][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.2780455946922302, 'train/grad_norm_pre_clip': 49.8986701965332, 'timing/data': 0.0003184480592608452, 'timing/model': 1.2736798599362373, 'learning_rate/qwen_vl_interface': 8.229931389105864e-06, 'learning_rate/action_model': 8.229931389105863e-05, 'epoch': 0.09}) +[2026-06-24 10:16:49,291][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.1627645194530487, 'train/grad_norm_pre_clip': 30.87386703491211, 'timing/data': 0.00031740590929985046, 'timing/model': 1.0642712430562824, 'learning_rate/qwen_vl_interface': 8.225922499220938e-06, 'learning_rate/action_model': 8.225922499220937e-05, 'epoch': 0.09}) +[2026-06-24 10:16:50,356][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.1589358150959015, 'train/grad_norm_pre_clip': 34.410091400146484, 'timing/data': 0.00035008788108825684, 'timing/model': 1.063673621043563, 'learning_rate/qwen_vl_interface': 8.2219101169252e-06, 'learning_rate/action_model': 8.221910116925199e-05, 'epoch': 0.09}) +[2026-06-24 10:16:51,422][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.049017444252967834, 'train/grad_norm_pre_clip': 14.73244571685791, 'timing/data': 0.0003288271836936474, 'timing/model': 1.0646326618734747, 'learning_rate/qwen_vl_interface': 8.217894246927402e-06, 'learning_rate/action_model': 8.217894246927402e-05, 'epoch': 0.09}) +[2026-06-24 10:16:52,485][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.11074361205101013, 'train/grad_norm_pre_clip': 30.314380645751953, 'timing/data': 0.0003765861038118601, 'timing/model': 1.0605133101344109, 'learning_rate/qwen_vl_interface': 8.213874893940395e-06, 'learning_rate/action_model': 8.213874893940395e-05, 'epoch': 0.09}) +[2026-06-24 10:16:53,549][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.11793962121009827, 'train/grad_norm_pre_clip': 49.42915725708008, 'timing/data': 0.0003558581229299307, 'timing/model': 1.0623609879985452, 'learning_rate/qwen_vl_interface': 8.209852062681115e-06, 'learning_rate/action_model': 8.209852062681115e-05, 'epoch': 0.09}) +[2026-06-24 10:16:54,613][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.15288260579109192, 'train/grad_norm_pre_clip': 53.08221435546875, 'timing/data': 0.00035800086334347725, 'timing/model': 1.0629171018954366, 'learning_rate/qwen_vl_interface': 8.205825757870573e-06, 'learning_rate/action_model': 8.205825757870574e-05, 'epoch': 0.09}) +[2026-06-24 10:16:55,679][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.42520377039909363, 'train/grad_norm_pre_clip': 66.06859588623047, 'timing/data': 0.0003351890482008457, 'timing/model': 1.0637085780035704, 'learning_rate/qwen_vl_interface': 8.201795984233868e-06, 'learning_rate/action_model': 8.201795984233868e-05, 'epoch': 0.09}) +[2026-06-24 10:16:56,743][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.1456698328256607, 'train/grad_norm_pre_clip': 39.38420104980469, 'timing/data': 0.00032987911254167557, 'timing/model': 1.0631551290862262, 'learning_rate/qwen_vl_interface': 8.19776274650016e-06, 'learning_rate/action_model': 8.197762746500158e-05, 'epoch': 0.09}) +[2026-06-24 10:16:57,806][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.16980025172233582, 'train/grad_norm_pre_clip': 56.24619674682617, 'timing/data': 0.00033312500454485416, 'timing/model': 1.0615762160159647, 'learning_rate/qwen_vl_interface': 8.193726049402675e-06, 'learning_rate/action_model': 8.193726049402675e-05, 'epoch': 0.09}) +[2026-06-24 10:16:58,870][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.17827633023262024, 'train/grad_norm_pre_clip': 50.19819259643555, 'timing/data': 0.0002989410422742367, 'timing/model': 1.0617405839730054, 'learning_rate/qwen_vl_interface': 8.18968589767871e-06, 'learning_rate/action_model': 8.189685897678711e-05, 'epoch': 0.09}) +[2026-06-24 10:16:59,935][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.26465368270874023, 'train/grad_norm_pre_clip': 55.368934631347656, 'timing/data': 0.0003504080232232809, 'timing/model': 1.0633947390597314, 'learning_rate/qwen_vl_interface': 8.185642296069604e-06, 'learning_rate/action_model': 8.185642296069603e-05, 'epoch': 0.09}) +[2026-06-24 10:17:01,000][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.048860352486371994, 'train/grad_norm_pre_clip': 21.777807235717773, 'timing/data': 0.0003418009728193283, 'timing/model': 1.063601857982576, 'learning_rate/qwen_vl_interface': 8.181595249320749e-06, 'learning_rate/action_model': 8.181595249320748e-05, 'epoch': 0.09}) +[2026-06-24 10:17:02,066][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.10025768727064133, 'train/grad_norm_pre_clip': 47.31846618652344, 'timing/data': 0.000300845131278038, 'timing/model': 1.0639519260730594, 'learning_rate/qwen_vl_interface': 8.177544762181575e-06, 'learning_rate/action_model': 8.177544762181576e-05, 'epoch': 0.09}) +[2026-06-24 10:17:03,130][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.33114326000213623, 'train/grad_norm_pre_clip': 86.37238311767578, 'timing/data': 0.0003128370735794306, 'timing/model': 1.062727847835049, 'learning_rate/qwen_vl_interface': 8.173490839405565e-06, 'learning_rate/action_model': 8.173490839405564e-05, 'epoch': 0.09}) +[2026-06-24 10:17:04,196][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.16665388643741608, 'train/grad_norm_pre_clip': 65.09516143798828, 'timing/data': 0.00032468000426888466, 'timing/model': 1.0649709678255022, 'learning_rate/qwen_vl_interface': 8.169433485750214e-06, 'learning_rate/action_model': 8.169433485750213e-05, 'epoch': 0.09}) +[2026-06-24 10:17:05,263][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.027528082951903343, 'train/grad_norm_pre_clip': 15.944815635681152, 'timing/data': 0.0003157730679959059, 'timing/model': 1.0651584200095385, 'learning_rate/qwen_vl_interface': 8.165372705977054e-06, 'learning_rate/action_model': 8.165372705977053e-05, 'epoch': 0.09}) +[2026-06-24 10:17:06,329][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.060218095779418945, 'train/grad_norm_pre_clip': 27.25568199157715, 'timing/data': 0.00031698495149612427, 'timing/model': 1.064377293922007, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.09}) +[2026-06-24 10:17:07,679][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.07362233847379684, 'train/grad_norm_pre_clip': 33.78809356689453, 'timing/data': 0.0003492860123515129, 'timing/model': 1.3477634070441127, 'learning_rate/qwen_vl_interface': 8.157240887143535e-06, 'learning_rate/action_model': 8.157240887143535e-05, 'epoch': 0.09}) +[2026-06-24 10:17:08,744][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.17118346691131592, 'train/grad_norm_pre_clip': 82.54734802246094, 'timing/data': 0.0003003031015396118, 'timing/model': 1.0634942471515387, 'learning_rate/qwen_vl_interface': 8.153169857626316e-06, 'learning_rate/action_model': 8.153169857626315e-05, 'epoch': 0.09}) +[2026-06-24 10:17:09,809][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.1145024374127388, 'train/grad_norm_pre_clip': 31.44441795349121, 'timing/data': 0.0003140191547572613, 'timing/model': 1.0635420070029795, 'learning_rate/qwen_vl_interface': 8.149095421077563e-06, 'learning_rate/action_model': 8.149095421077563e-05, 'epoch': 0.09}) +[2026-06-24 10:17:10,873][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.3188505470752716, 'train/grad_norm_pre_clip': 107.07937622070312, 'timing/data': 0.0003113949205726385, 'timing/model': 1.0631584769580513, 'learning_rate/qwen_vl_interface': 8.145017582278853e-06, 'learning_rate/action_model': 8.145017582278853e-05, 'epoch': 0.09}) +[2026-06-24 10:17:11,939][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.11546020954847336, 'train/grad_norm_pre_clip': 33.98527145385742, 'timing/data': 0.00047934986650943756, 'timing/model': 1.0641572019085288, 'learning_rate/qwen_vl_interface': 8.14093634601576e-06, 'learning_rate/action_model': 8.14093634601576e-05, 'epoch': 0.09}) +[2026-06-24 10:17:13,005][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.12568624317646027, 'train/grad_norm_pre_clip': 41.56804656982422, 'timing/data': 0.0003139888867735863, 'timing/model': 1.0637243390083313, 'learning_rate/qwen_vl_interface': 8.13685171707784e-06, 'learning_rate/action_model': 8.13685171707784e-05, 'epoch': 0.09}) +[2026-06-24 10:17:14,069][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.07310239225625992, 'train/grad_norm_pre_clip': 50.44050598144531, 'timing/data': 0.0003138990141451359, 'timing/model': 1.0627827220596373, 'learning_rate/qwen_vl_interface': 8.13276370025863e-06, 'learning_rate/action_model': 8.132763700258632e-05, 'epoch': 0.09}) +[2026-06-24 10:17:15,135][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.18108807504177094, 'train/grad_norm_pre_clip': 65.66712188720703, 'timing/data': 0.0003449579235166311, 'timing/model': 1.0644187128636986, 'learning_rate/qwen_vl_interface': 8.128672300355651e-06, 'learning_rate/action_model': 8.128672300355651e-05, 'epoch': 0.09}) +[2026-06-24 10:17:16,197][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.31759291887283325, 'train/grad_norm_pre_clip': 163.19068908691406, 'timing/data': 0.0003224448300898075, 'timing/model': 1.0601787748746574, 'learning_rate/qwen_vl_interface': 8.124577522170386e-06, 'learning_rate/action_model': 8.124577522170386e-05, 'epoch': 0.09}) +[2026-06-24 10:17:17,262][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.06659432500600815, 'train/grad_norm_pre_clip': 47.91151428222656, 'timing/data': 0.00030761794187128544, 'timing/model': 1.063954541925341, 'learning_rate/qwen_vl_interface': 8.120479370508283e-06, 'learning_rate/action_model': 8.120479370508284e-05, 'epoch': 0.09}) +[2026-06-24 10:17:18,329][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.12332349270582199, 'train/grad_norm_pre_clip': 57.63363265991211, 'timing/data': 0.0002992311492562294, 'timing/model': 1.0656991261057556, 'learning_rate/qwen_vl_interface': 8.116377850178756e-06, 'learning_rate/action_model': 8.116377850178755e-05, 'epoch': 0.09}) +[2026-06-24 10:17:19,395][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.1504794955253601, 'train/grad_norm_pre_clip': 48.054412841796875, 'timing/data': 0.00031590298749506474, 'timing/model': 1.0635663329157978, 'learning_rate/qwen_vl_interface': 8.112272965995162e-06, 'learning_rate/action_model': 8.112272965995163e-05, 'epoch': 0.09}) +[2026-06-24 10:17:20,462][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.005552154034376144, 'train/grad_norm_pre_clip': 4.606973648071289, 'timing/data': 0.0003130671102553606, 'timing/model': 1.065730856033042, 'learning_rate/qwen_vl_interface': 8.108164722774818e-06, 'learning_rate/action_model': 8.108164722774817e-05, 'epoch': 0.09}) +[2026-06-24 10:17:21,527][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.0927542969584465, 'train/grad_norm_pre_clip': 43.08588409423828, 'timing/data': 0.00034235301427543163, 'timing/model': 1.0635271300561726, 'learning_rate/qwen_vl_interface': 8.10405312533897e-06, 'learning_rate/action_model': 8.104053125338969e-05, 'epoch': 0.09}) +[2026-06-24 10:17:22,590][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.5528140664100647, 'train/grad_norm_pre_clip': 123.15127563476562, 'timing/data': 0.000325710978358984, 'timing/model': 1.0610833528917283, 'learning_rate/qwen_vl_interface': 8.09993817851281e-06, 'learning_rate/action_model': 8.099938178512809e-05, 'epoch': 0.09}) +[2026-06-24 10:17:23,652][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.44516921043395996, 'train/grad_norm_pre_clip': 176.33917236328125, 'timing/data': 0.00030474201776087284, 'timing/model': 1.0608762248884887, 'learning_rate/qwen_vl_interface': 8.095819887125453e-06, 'learning_rate/action_model': 8.095819887125453e-05, 'epoch': 0.09}) +[2026-06-24 10:17:24,715][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.19457225501537323, 'train/grad_norm_pre_clip': 115.49915313720703, 'timing/data': 0.00037488387897610664, 'timing/model': 1.0606825309805572, 'learning_rate/qwen_vl_interface': 8.09169825600995e-06, 'learning_rate/action_model': 8.091698256009951e-05, 'epoch': 0.09}) +[2026-06-24 10:17:26,022][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.2886485457420349, 'train/grad_norm_pre_clip': 143.31983947753906, 'timing/data': 0.0005543099250644445, 'timing/model': 1.304922026116401, 'learning_rate/qwen_vl_interface': 8.08757329000326e-06, 'learning_rate/action_model': 8.08757329000326e-05, 'epoch': 0.09}) +[2026-06-24 10:17:27,086][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.22365330159664154, 'train/grad_norm_pre_clip': 101.40535736083984, 'timing/data': 0.00030024396255612373, 'timing/model': 1.0625077469740063, 'learning_rate/qwen_vl_interface': 8.083444993946263e-06, 'learning_rate/action_model': 8.083444993946263e-05, 'epoch': 0.09}) +[2026-06-24 10:17:28,150][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.35401973128318787, 'train/grad_norm_pre_clip': 84.4202880859375, 'timing/data': 0.0003478729631751776, 'timing/model': 1.062113197054714, 'learning_rate/qwen_vl_interface': 8.079313372683744e-06, 'learning_rate/action_model': 8.079313372683743e-05, 'epoch': 0.09}) +[2026-06-24 10:17:29,212][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.08291987329721451, 'train/grad_norm_pre_clip': 31.44228744506836, 'timing/data': 0.0002914669457823038, 'timing/model': 1.0606785039417446, 'learning_rate/qwen_vl_interface': 8.075178431064386e-06, 'learning_rate/action_model': 8.075178431064387e-05, 'epoch': 0.09}) +[2026-06-24 10:17:30,277][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.115488700568676, 'train/grad_norm_pre_clip': 63.78663635253906, 'timing/data': 0.000327324029058218, 'timing/model': 1.0634673789609224, 'learning_rate/qwen_vl_interface': 8.07104017394078e-06, 'learning_rate/action_model': 8.071040173940779e-05, 'epoch': 0.09}) +[2026-06-24 10:17:31,342][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.27938687801361084, 'train/grad_norm_pre_clip': 109.42051696777344, 'timing/data': 0.0003320130053907633, 'timing/model': 1.0633648559451103, 'learning_rate/qwen_vl_interface': 8.066898606169398e-06, 'learning_rate/action_model': 8.066898606169399e-05, 'epoch': 0.09}) +[2026-06-24 10:17:32,407][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.3018537163734436, 'train/grad_norm_pre_clip': 49.71356201171875, 'timing/data': 0.0003431139048188925, 'timing/model': 1.0632455919403583, 'learning_rate/qwen_vl_interface': 8.062753732610602e-06, 'learning_rate/action_model': 8.062753732610601e-05, 'epoch': 0.09}) +[2026-06-24 10:17:33,471][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.23211532831192017, 'train/grad_norm_pre_clip': 45.22897720336914, 'timing/data': 0.0003054130356758833, 'timing/model': 1.0625437358394265, 'learning_rate/qwen_vl_interface': 8.058605558128628e-06, 'learning_rate/action_model': 8.058605558128628e-05, 'epoch': 0.09}) +[2026-06-24 10:17:34,535][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.240408256649971, 'train/grad_norm_pre_clip': 44.797996520996094, 'timing/data': 0.00030933995731174946, 'timing/model': 1.0629508199635893, 'learning_rate/qwen_vl_interface': 8.054454087591594e-06, 'learning_rate/action_model': 8.054454087591595e-05, 'epoch': 0.09}) +[2026-06-24 10:17:35,599][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.1865759938955307, 'train/grad_norm_pre_clip': 52.5938720703125, 'timing/data': 0.0003157129976898432, 'timing/model': 1.0622129049152136, 'learning_rate/qwen_vl_interface': 8.05029932587148e-06, 'learning_rate/action_model': 8.05029932587148e-05, 'epoch': 0.09}) +[2026-06-24 10:17:36,665][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.05755564570426941, 'train/grad_norm_pre_clip': 33.41376495361328, 'timing/data': 0.00032774615101516247, 'timing/model': 1.064949621912092, 'learning_rate/qwen_vl_interface': 8.04614127784413e-06, 'learning_rate/action_model': 8.04614127784413e-05, 'epoch': 0.09}) +[2026-06-24 10:17:37,729][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.09980735927820206, 'train/grad_norm_pre_clip': 32.096065521240234, 'timing/data': 0.0003143399953842163, 'timing/model': 1.061800140189007, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.09}) +[2026-06-24 10:17:38,794][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.2596190571784973, 'train/grad_norm_pre_clip': 47.62991714477539, 'timing/data': 0.00032482994720339775, 'timing/model': 1.0642890019807965, 'learning_rate/qwen_vl_interface': 8.037815342390383e-06, 'learning_rate/action_model': 8.037815342390383e-05, 'epoch': 0.09}) +[2026-06-24 10:17:39,859][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.13360626995563507, 'train/grad_norm_pre_clip': 37.25353240966797, 'timing/data': 0.0003055729903280735, 'timing/model': 1.0625286779832095, 'learning_rate/qwen_vl_interface': 8.033647464734932e-06, 'learning_rate/action_model': 8.033647464734932e-05, 'epoch': 0.09}) +[2026-06-24 10:17:40,923][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.2994464039802551, 'train/grad_norm_pre_clip': 53.41593933105469, 'timing/data': 0.00030131591483950615, 'timing/model': 1.0629148329608142, 'learning_rate/qwen_vl_interface': 8.029476320314133e-06, 'learning_rate/action_model': 8.029476320314132e-05, 'epoch': 0.09}) +[2026-06-24 10:17:41,988][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.1694440096616745, 'train/grad_norm_pre_clip': 27.789819717407227, 'timing/data': 0.0003195800818502903, 'timing/model': 1.0633530660998076, 'learning_rate/qwen_vl_interface': 8.025301914023055e-06, 'learning_rate/action_model': 8.025301914023054e-05, 'epoch': 0.09}) +[2026-06-24 10:17:43,051][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.08404498547315598, 'train/grad_norm_pre_clip': 18.519542694091797, 'timing/data': 0.00031723594292998314, 'timing/model': 1.061645400011912, 'learning_rate/qwen_vl_interface': 8.021124250760598e-06, 'learning_rate/action_model': 8.021124250760598e-05, 'epoch': 0.09}) +[2026-06-24 10:17:44,316][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.0863904058933258, 'train/grad_norm_pre_clip': 18.34076690673828, 'timing/data': 0.00031704502180218697, 'timing/model': 1.26345826103352, 'learning_rate/qwen_vl_interface': 8.016943335429481e-06, 'learning_rate/action_model': 8.01694333542948e-05, 'epoch': 0.09}) +[2026-06-24 10:17:45,379][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.18377970159053802, 'train/grad_norm_pre_clip': 36.73999786376953, 'timing/data': 0.0003059541340917349, 'timing/model': 1.0615568237844855, 'learning_rate/qwen_vl_interface': 8.01275917293624e-06, 'learning_rate/action_model': 8.012759172936241e-05, 'epoch': 0.09}) +[2026-06-24 10:17:46,443][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.12702669203281403, 'train/grad_norm_pre_clip': 19.19252586364746, 'timing/data': 0.0003413709346204996, 'timing/model': 1.061973406933248, 'learning_rate/qwen_vl_interface': 8.008571768191224e-06, 'learning_rate/action_model': 8.008571768191225e-05, 'epoch': 0.09}) +[2026-06-24 10:17:47,506][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.0879230946302414, 'train/grad_norm_pre_clip': 45.27671432495117, 'timing/data': 0.000296145910397172, 'timing/model': 1.0621801239904016, 'learning_rate/qwen_vl_interface': 8.004381126108588e-06, 'learning_rate/action_model': 8.004381126108588e-05, 'epoch': 0.09}) +[2026-06-24 10:17:48,571][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.3823074400424957, 'train/grad_norm_pre_clip': 49.81785583496094, 'timing/data': 0.00030954089015722275, 'timing/model': 1.062583161983639, 'learning_rate/qwen_vl_interface': 8.000187251606283e-06, 'learning_rate/action_model': 8.000187251606283e-05, 'epoch': 0.09}) +[2026-06-24 10:17:49,634][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.120541512966156, 'train/grad_norm_pre_clip': 38.32649612426758, 'timing/data': 0.0003745928406715393, 'timing/model': 1.062273500021547, 'learning_rate/qwen_vl_interface': 7.99599014960605e-06, 'learning_rate/action_model': 7.99599014960605e-05, 'epoch': 0.09}) +[2026-06-24 10:17:50,698][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.18280625343322754, 'train/grad_norm_pre_clip': 32.07265853881836, 'timing/data': 0.0002966970205307007, 'timing/model': 1.0618694419972599, 'learning_rate/qwen_vl_interface': 7.991789825033427e-06, 'learning_rate/action_model': 7.991789825033426e-05, 'epoch': 0.1}) +[2026-06-24 10:17:51,762][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.2093326300382614, 'train/grad_norm_pre_clip': 50.74298858642578, 'timing/data': 0.0003497060388326645, 'timing/model': 1.062523860950023, 'learning_rate/qwen_vl_interface': 7.987586282817728e-06, 'learning_rate/action_model': 7.987586282817728e-05, 'epoch': 0.1}) +[2026-06-24 10:17:52,822][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.12558665871620178, 'train/grad_norm_pre_clip': 22.96592140197754, 'timing/data': 0.0003177858889102936, 'timing/model': 1.058522520121187, 'learning_rate/qwen_vl_interface': 7.98337952789204e-06, 'learning_rate/action_model': 7.98337952789204e-05, 'epoch': 0.1}) +[2026-06-24 10:17:53,888][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.07977495342493057, 'train/grad_norm_pre_clip': 18.675519943237305, 'timing/data': 0.0002968569751828909, 'timing/model': 1.0642842140514404, 'learning_rate/qwen_vl_interface': 7.979169565193229e-06, 'learning_rate/action_model': 7.979169565193228e-05, 'epoch': 0.1}) +[2026-06-24 10:17:54,948][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.26384487748146057, 'train/grad_norm_pre_clip': 53.1866569519043, 'timing/data': 0.00031307805329561234, 'timing/model': 1.0586909458506852, 'learning_rate/qwen_vl_interface': 7.974956399661919e-06, 'learning_rate/action_model': 7.974956399661918e-05, 'epoch': 0.1}) +[2026-06-24 10:17:56,012][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.06995823979377747, 'train/grad_norm_pre_clip': 31.13324737548828, 'timing/data': 0.0002965368330478668, 'timing/model': 1.0622910240199417, 'learning_rate/qwen_vl_interface': 7.970740036242493e-06, 'learning_rate/action_model': 7.970740036242492e-05, 'epoch': 0.1}) +[2026-06-24 10:17:57,079][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.1041158139705658, 'train/grad_norm_pre_clip': 24.219480514526367, 'timing/data': 0.00030286796391010284, 'timing/model': 1.0657252999953926, 'learning_rate/qwen_vl_interface': 7.96652047988309e-06, 'learning_rate/action_model': 7.966520479883089e-05, 'epoch': 0.1}) +[2026-06-24 10:17:58,145][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.11659885942935944, 'train/grad_norm_pre_clip': 39.88356399536133, 'timing/data': 0.0003480231389403343, 'timing/model': 1.0638481369242072, 'learning_rate/qwen_vl_interface': 7.962297735535595e-06, 'learning_rate/action_model': 7.962297735535595e-05, 'epoch': 0.1}) +[2026-06-24 10:17:59,212][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.36150071024894714, 'train/grad_norm_pre_clip': 63.789974212646484, 'timing/data': 0.00032051210291683674, 'timing/model': 1.065478908130899, 'learning_rate/qwen_vl_interface': 7.958071808155632e-06, 'learning_rate/action_model': 7.958071808155631e-05, 'epoch': 0.1}) +[2026-06-24 10:18:00,277][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.4111848771572113, 'train/grad_norm_pre_clip': 78.46732330322266, 'timing/data': 0.0003015559632331133, 'timing/model': 1.0632170620374382, 'learning_rate/qwen_vl_interface': 7.953842702702563e-06, 'learning_rate/action_model': 7.953842702702563e-05, 'epoch': 0.1}) +[2026-06-24 10:18:01,549][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.4121588468551636, 'train/grad_norm_pre_clip': 67.44120025634766, 'timing/data': 0.000310912961140275, 'timing/model': 1.2708673710003495, 'learning_rate/qwen_vl_interface': 7.949610424139478e-06, 'learning_rate/action_model': 7.949610424139477e-05, 'epoch': 0.1}) +[2026-06-24 10:18:02,613][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.6060383319854736, 'train/grad_norm_pre_clip': 98.60985565185547, 'timing/data': 0.00032771495170891285, 'timing/model': 1.062361255986616, 'learning_rate/qwen_vl_interface': 7.945374977433194e-06, 'learning_rate/action_model': 7.945374977433195e-05, 'epoch': 0.1}) +[2026-06-24 10:18:03,673][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.25823312997817993, 'train/grad_norm_pre_clip': 87.1255111694336, 'timing/data': 0.0003034300170838833, 'timing/model': 1.0584626279305667, 'learning_rate/qwen_vl_interface': 7.941136367554241e-06, 'learning_rate/action_model': 7.94113636755424e-05, 'epoch': 0.1}) +[2026-06-24 10:18:04,742][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.16117124259471893, 'train/grad_norm_pre_clip': 41.91661071777344, 'timing/data': 0.0003506280481815338, 'timing/model': 1.067266693804413, 'learning_rate/qwen_vl_interface': 7.936894599476864e-06, 'learning_rate/action_model': 7.936894599476863e-05, 'epoch': 0.1}) +[2026-06-24 10:18:05,808][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.3591064214706421, 'train/grad_norm_pre_clip': 61.435577392578125, 'timing/data': 0.00031286710873246193, 'timing/model': 1.064390165032819, 'learning_rate/qwen_vl_interface': 7.932649678179016e-06, 'learning_rate/action_model': 7.932649678179016e-05, 'epoch': 0.1}) +[2026-06-24 10:18:06,876][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.15490111708641052, 'train/grad_norm_pre_clip': 34.49748611450195, 'timing/data': 0.00030308892019093037, 'timing/model': 1.066401218995452, 'learning_rate/qwen_vl_interface': 7.928401608642346e-06, 'learning_rate/action_model': 7.928401608642346e-05, 'epoch': 0.1}) +[2026-06-24 10:18:07,942][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.2517758309841156, 'train/grad_norm_pre_clip': 49.29954147338867, 'timing/data': 0.00030143605545163155, 'timing/model': 1.0639860569499433, 'learning_rate/qwen_vl_interface': 7.924150395852204e-06, 'learning_rate/action_model': 7.924150395852203e-05, 'epoch': 0.1}) +[2026-06-24 10:18:09,007][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.27964165806770325, 'train/grad_norm_pre_clip': 54.97224426269531, 'timing/data': 0.00033208285458385944, 'timing/model': 1.0643158350139856, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.1}) +[2026-06-24 10:18:10,072][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1000 +[2026-06-24 10:18:35,629][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1000 +[2026-06-24 10:18:35,630][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1000 +[2026-06-24 10:19:06,643][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1000 +[2026-06-24 10:19:15,512][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-24 10:19:15,512][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.4443323314189911, 'train/grad_norm_pre_clip': 108.65338134765625, 'eval/action_loss/seconds': 25.55507951392792, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.14840750396251678, 'eval/latency_2/loss': 0.14840750396251678, 'eval/flappy/loss': 0.14840750396251678, 'eval/flappy/latency_2/loss': 0.14840750396251678, 'eval/flap_precision': 0.9922779922779923, 'eval/flap_recall': 0.6974219810040706, 'eval/flap_f1': 0.8191235059760955, 'eval/flap_rate_model': 0.1438888888888889, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8191235059760955, 'eval/latency_2/flap_precision': 0.9922779922779923, 'eval/latency_2/flap_recall': 0.6974219810040706, 'eval/latency_2/flap_f1': 0.8191235059760955, 'eval/latency_2/flap_rate_model': 0.1438888888888889, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8191235059760955, 'eval/action_classification/data_wait_seconds': 0.3127109408378601, 'eval/action_classification/predict_seconds': 30.59832000732422, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4.799999999999995, 'rl_games_eval/mid_train/mean_length': 57.2, 'rl_games_eval/mid_train/std_reward': 1.7999999999999952, 'rl_games_eval/mid_train/std_length': 14.4, 'rl_games_eval/mid_train/macro_mean_reward': 4.799999999999995, 'rl_games_eval/mid_train/macro_mean_length': 57.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 4.799999999999995, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 57.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 1.7999999999999952, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 14.4, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 4.799999999999995, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000332132913172245, 'timing/model': 1.0637841580901295, 'learning_rate/qwen_vl_interface': 7.915638560471316e-06, 'learning_rate/action_model': 7.915638560471317e-05, 'epoch': 0.1}) +[2026-06-24 10:19:15,513][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_1000_state +[2026-06-24 10:19:15,515][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 10:19:18,097][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_1000_state/model.safetensors +[2026-06-24 10:19:23,826][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_1000_state/optimizer.bin +[2026-06-24 10:19:23,827][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_1000_state/sampler.bin +[2026-06-24 10:19:23,827][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-24 10:19:23,828][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-24 10:19:23,873][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-24 10:19:23,873][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 10:19:23,883][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 10:19:24,971][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.16118909418582916, 'train/grad_norm_pre_clip': 31.698150634765625, 'timing/data': 0.0004111519083380699, 'timing/model': 1.0862960091326386, 'learning_rate/qwen_vl_interface': 7.911377947869686e-06, 'learning_rate/action_model': 7.911377947869685e-05, 'epoch': 0.1}) +[2026-06-24 10:19:26,023][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.20969834923744202, 'train/grad_norm_pre_clip': 37.33831787109375, 'timing/data': 0.000495188869535923, 'timing/model': 1.0500002519693226, 'learning_rate/qwen_vl_interface': 7.907114211992796e-06, 'learning_rate/action_model': 7.907114211992796e-05, 'epoch': 0.1}) +[2026-06-24 10:19:27,077][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.2078624963760376, 'train/grad_norm_pre_clip': 19.189773559570312, 'timing/data': 0.0003507779911160469, 'timing/model': 1.0519905579276383, 'learning_rate/qwen_vl_interface': 7.902847357844376e-06, 'learning_rate/action_model': 7.902847357844376e-05, 'epoch': 0.1}) +[2026-06-24 10:19:28,428][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.2075006365776062, 'train/grad_norm_pre_clip': 36.7506217956543, 'timing/data': 0.0003835300449281931, 'timing/model': 1.3494620830751956, 'learning_rate/qwen_vl_interface': 7.89857739043182e-06, 'learning_rate/action_model': 7.89857739043182e-05, 'epoch': 0.1}) +[2026-06-24 10:19:29,481][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.1609117090702057, 'train/grad_norm_pre_clip': 17.964216232299805, 'timing/data': 0.0003075569402426481, 'timing/model': 1.0514674470759928, 'learning_rate/qwen_vl_interface': 7.894304314766169e-06, 'learning_rate/action_model': 7.894304314766169e-05, 'epoch': 0.1}) +[2026-06-24 10:19:30,535][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.13942918181419373, 'train/grad_norm_pre_clip': 12.827739715576172, 'timing/data': 0.00031484104692935944, 'timing/model': 1.0527150791604072, 'learning_rate/qwen_vl_interface': 7.890028135862119e-06, 'learning_rate/action_model': 7.890028135862118e-05, 'epoch': 0.1}) +[2026-06-24 10:19:31,590][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.10603281110525131, 'train/grad_norm_pre_clip': 17.609525680541992, 'timing/data': 0.00033210404217243195, 'timing/model': 1.0531511178705841, 'learning_rate/qwen_vl_interface': 7.885748858738e-06, 'learning_rate/action_model': 7.885748858738001e-05, 'epoch': 0.1}) +[2026-06-24 10:19:32,640][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.09159577637910843, 'train/grad_norm_pre_clip': 28.98831558227539, 'timing/data': 0.00030500185675919056, 'timing/model': 1.0486957849934697, 'learning_rate/qwen_vl_interface': 7.881466488415787e-06, 'learning_rate/action_model': 7.881466488415786e-05, 'epoch': 0.1}) +[2026-06-24 10:19:33,694][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.2000354826450348, 'train/grad_norm_pre_clip': 26.13629722595215, 'timing/data': 0.00032467907294631004, 'timing/model': 1.0522597848903388, 'learning_rate/qwen_vl_interface': 7.877181029921076e-06, 'learning_rate/action_model': 7.877181029921074e-05, 'epoch': 0.1}) +[2026-06-24 10:19:34,750][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.22205600142478943, 'train/grad_norm_pre_clip': 41.997962951660156, 'timing/data': 0.00032376800663769245, 'timing/model': 1.0543918469920754, 'learning_rate/qwen_vl_interface': 7.872892488283091e-06, 'learning_rate/action_model': 7.872892488283091e-05, 'epoch': 0.1}) +[2026-06-24 10:19:35,806][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.20540358126163483, 'train/grad_norm_pre_clip': 48.02382278442383, 'timing/data': 0.000294252997264266, 'timing/model': 1.0545585891231894, 'learning_rate/qwen_vl_interface': 7.86860086853468e-06, 'learning_rate/action_model': 7.86860086853468e-05, 'epoch': 0.1}) +[2026-06-24 10:19:36,863][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.11288813501596451, 'train/grad_norm_pre_clip': 48.40034866333008, 'timing/data': 0.00030181603506207466, 'timing/model': 1.0549713540822268, 'learning_rate/qwen_vl_interface': 7.864306175712292e-06, 'learning_rate/action_model': 7.864306175712292e-05, 'epoch': 0.1}) +[2026-06-24 10:19:37,919][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.06430872529745102, 'train/grad_norm_pre_clip': 28.727434158325195, 'timing/data': 0.0004379618912935257, 'timing/model': 1.05475441692397, 'learning_rate/qwen_vl_interface': 7.860008414855992e-06, 'learning_rate/action_model': 7.86000841485599e-05, 'epoch': 0.1}) +[2026-06-24 10:19:38,977][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.39592570066452026, 'train/grad_norm_pre_clip': 131.49783325195312, 'timing/data': 0.00030046398751437664, 'timing/model': 1.0563723240047693, 'learning_rate/qwen_vl_interface': 7.855707591009437e-06, 'learning_rate/action_model': 7.855707591009437e-05, 'epoch': 0.1}) +[2026-06-24 10:19:40,035][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.05992431193590164, 'train/grad_norm_pre_clip': 26.412479400634766, 'timing/data': 0.0003133779391646385, 'timing/model': 1.0564741659909487, 'learning_rate/qwen_vl_interface': 7.85140370921989e-06, 'learning_rate/action_model': 7.851403709219889e-05, 'epoch': 0.1}) +[2026-06-24 10:19:41,095][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.132041335105896, 'train/grad_norm_pre_clip': 43.034751892089844, 'timing/data': 0.00031648389995098114, 'timing/model': 1.057828287128359, 'learning_rate/qwen_vl_interface': 7.847096774538188e-06, 'learning_rate/action_model': 7.847096774538189e-05, 'epoch': 0.1}) +[2026-06-24 10:19:42,152][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.19883641600608826, 'train/grad_norm_pre_clip': 71.14249420166016, 'timing/data': 0.0003780289553105831, 'timing/model': 1.0557117748539895, 'learning_rate/qwen_vl_interface': 7.842786792018768e-06, 'learning_rate/action_model': 7.842786792018766e-05, 'epoch': 0.1}) +[2026-06-24 10:19:43,508][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.288099467754364, 'train/grad_norm_pre_clip': 160.4937744140625, 'timing/data': 0.00030532293021678925, 'timing/model': 1.3544183380436152, 'learning_rate/qwen_vl_interface': 7.838473766719627e-06, 'learning_rate/action_model': 7.838473766719626e-05, 'epoch': 0.1}) +[2026-06-24 10:19:44,566][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.28588470816612244, 'train/grad_norm_pre_clip': 138.37771606445312, 'timing/data': 0.0003336761146783829, 'timing/model': 1.0562443840317428, 'learning_rate/qwen_vl_interface': 7.834157703702343e-06, 'learning_rate/action_model': 7.834157703702343e-05, 'epoch': 0.1}) +[2026-06-24 10:19:45,625][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.22757776081562042, 'train/grad_norm_pre_clip': 121.57514953613281, 'timing/data': 0.0003054130356758833, 'timing/model': 1.057464695069939, 'learning_rate/qwen_vl_interface': 7.829838608032057e-06, 'learning_rate/action_model': 7.829838608032056e-05, 'epoch': 0.1}) +[2026-06-24 10:19:46,680][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.15912611782550812, 'train/grad_norm_pre_clip': 18.159900665283203, 'timing/data': 0.0003105930518358946, 'timing/model': 1.0536958610173315, 'learning_rate/qwen_vl_interface': 7.825516484777468e-06, 'learning_rate/action_model': 7.825516484777467e-05, 'epoch': 0.1}) +[2026-06-24 10:19:47,740][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.04973159357905388, 'train/grad_norm_pre_clip': 15.153806686401367, 'timing/data': 0.00034189200960099697, 'timing/model': 1.0576999769546092, 'learning_rate/qwen_vl_interface': 7.821191339010827e-06, 'learning_rate/action_model': 7.821191339010827e-05, 'epoch': 0.1}) +[2026-06-24 10:19:48,801][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.11193568259477615, 'train/grad_norm_pre_clip': 26.04732322692871, 'timing/data': 0.0003266229759901762, 'timing/model': 1.0591704880353063, 'learning_rate/qwen_vl_interface': 7.816863175807934e-06, 'learning_rate/action_model': 7.816863175807935e-05, 'epoch': 0.1}) +[2026-06-24 10:19:49,861][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.2226627767086029, 'train/grad_norm_pre_clip': 32.67723083496094, 'timing/data': 0.00030537298880517483, 'timing/model': 1.0585363570135087, 'learning_rate/qwen_vl_interface': 7.812532000248134e-06, 'learning_rate/action_model': 7.812532000248134e-05, 'epoch': 0.1}) +[2026-06-24 10:19:50,923][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.27256545424461365, 'train/grad_norm_pre_clip': 56.20002365112305, 'timing/data': 0.00030860002152621746, 'timing/model': 1.060397329973057, 'learning_rate/qwen_vl_interface': 7.808197817414297e-06, 'learning_rate/action_model': 7.808197817414295e-05, 'epoch': 0.1}) +[2026-06-24 10:19:51,982][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.16624751687049866, 'train/grad_norm_pre_clip': 142.80274963378906, 'timing/data': 0.0003065250348299742, 'timing/model': 1.0577106680721045, 'learning_rate/qwen_vl_interface': 7.803860632392829e-06, 'learning_rate/action_model': 7.803860632392828e-05, 'epoch': 0.1}) +[2026-06-24 10:19:53,042][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.19970999658107758, 'train/grad_norm_pre_clip': 34.3243522644043, 'timing/data': 0.00031543197110295296, 'timing/model': 1.0588377329986542, 'learning_rate/qwen_vl_interface': 7.799520450273662e-06, 'learning_rate/action_model': 7.79952045027366e-05, 'epoch': 0.1}) +[2026-06-24 10:19:54,101][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.34703686833381653, 'train/grad_norm_pre_clip': 106.68963623046875, 'timing/data': 0.0003211430739611387, 'timing/model': 1.057134314905852, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.1}) +[2026-06-24 10:19:55,161][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.12503324449062347, 'train/grad_norm_pre_clip': 32.18389129638672, 'timing/data': 0.0003096810542047024, 'timing/model': 1.0588559180032462, 'learning_rate/qwen_vl_interface': 7.790831115119509e-06, 'learning_rate/action_model': 7.790831115119509e-05, 'epoch': 0.1}) +[2026-06-24 10:19:56,221][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.1595529466867447, 'train/grad_norm_pre_clip': 37.166465759277344, 'timing/data': 0.00035712006501853466, 'timing/model': 1.0581996659748256, 'learning_rate/qwen_vl_interface': 7.786481972281948e-06, 'learning_rate/action_model': 7.786481972281949e-05, 'epoch': 0.1}) +[2026-06-24 10:19:57,282][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.2572673261165619, 'train/grad_norm_pre_clip': 39.72801208496094, 'timing/data': 0.00034054904244840145, 'timing/model': 1.0588313520420343, 'learning_rate/qwen_vl_interface': 7.78212985274151e-06, 'learning_rate/action_model': 7.782129852741508e-05, 'epoch': 0.1}) +[2026-06-24 10:19:58,550][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.2069701850414276, 'train/grad_norm_pre_clip': 39.01542282104492, 'timing/data': 0.0002951039932668209, 'timing/model': 1.266454349970445, 'learning_rate/qwen_vl_interface': 7.777774761605649e-06, 'learning_rate/action_model': 7.777774761605649e-05, 'epoch': 0.1}) +[2026-06-24 10:19:59,611][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.18810692429542542, 'train/grad_norm_pre_clip': 48.713348388671875, 'timing/data': 0.00032265600748360157, 'timing/model': 1.0595996940974146, 'learning_rate/qwen_vl_interface': 7.77341670398531e-06, 'learning_rate/action_model': 7.773416703985308e-05, 'epoch': 0.1}) +[2026-06-24 10:20:00,674][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.36447852849960327, 'train/grad_norm_pre_clip': 96.44677734375, 'timing/data': 0.0004971430171281099, 'timing/model': 1.0612659910693765, 'learning_rate/qwen_vl_interface': 7.769055684994913e-06, 'learning_rate/action_model': 7.769055684994912e-05, 'epoch': 0.1}) +[2026-06-24 10:20:01,737][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.10737968236207962, 'train/grad_norm_pre_clip': 14.35232925415039, 'timing/data': 0.00034418608993291855, 'timing/model': 1.0612872811034322, 'learning_rate/qwen_vl_interface': 7.764691709752362e-06, 'learning_rate/action_model': 7.764691709752361e-05, 'epoch': 0.1}) +[2026-06-24 10:20:02,795][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.17151321470737457, 'train/grad_norm_pre_clip': 25.9808292388916, 'timing/data': 0.0004941769875586033, 'timing/model': 1.0567458991426975, 'learning_rate/qwen_vl_interface': 7.760324783379017e-06, 'learning_rate/action_model': 7.760324783379017e-05, 'epoch': 0.1}) +[2026-06-24 10:20:03,855][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.09645180404186249, 'train/grad_norm_pre_clip': 19.4659366607666, 'timing/data': 0.0005378490313887596, 'timing/model': 1.0577760830055922, 'learning_rate/qwen_vl_interface': 7.755954910999716e-06, 'learning_rate/action_model': 7.755954910999716e-05, 'epoch': 0.1}) +[2026-06-24 10:20:04,918][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.09255953878164291, 'train/grad_norm_pre_clip': 32.863712310791016, 'timing/data': 0.00032381806522607803, 'timing/model': 1.0609908560290933, 'learning_rate/qwen_vl_interface': 7.751582097742746e-06, 'learning_rate/action_model': 7.751582097742746e-05, 'epoch': 0.1}) +[2026-06-24 10:20:05,981][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.21484164893627167, 'train/grad_norm_pre_clip': 39.28376388549805, 'timing/data': 0.00028556608594954014, 'timing/model': 1.0613448799122125, 'learning_rate/qwen_vl_interface': 7.747206348739847e-06, 'learning_rate/action_model': 7.747206348739848e-05, 'epoch': 0.1}) +[2026-06-24 10:20:07,041][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.05471326783299446, 'train/grad_norm_pre_clip': 13.547318458557129, 'timing/data': 0.00030003394931554794, 'timing/model': 1.0590193551033735, 'learning_rate/qwen_vl_interface': 7.742827669126205e-06, 'learning_rate/action_model': 7.742827669126203e-05, 'epoch': 0.1}) +[2026-06-24 10:20:08,102][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.3417782783508301, 'train/grad_norm_pre_clip': 37.980350494384766, 'timing/data': 0.00030361907556653023, 'timing/model': 1.05898447916843, 'learning_rate/qwen_vl_interface': 7.73844606404044e-06, 'learning_rate/action_model': 7.73844606404044e-05, 'epoch': 0.1}) +[2026-06-24 10:20:09,165][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.07255401462316513, 'train/grad_norm_pre_clip': 13.011667251586914, 'timing/data': 0.00031478097662329674, 'timing/model': 1.0613895740825683, 'learning_rate/qwen_vl_interface': 7.734061538624615e-06, 'learning_rate/action_model': 7.734061538624614e-05, 'epoch': 0.1}) +[2026-06-24 10:20:10,227][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.128037691116333, 'train/grad_norm_pre_clip': 33.411991119384766, 'timing/data': 0.0003063150215893984, 'timing/model': 1.0604728139005601, 'learning_rate/qwen_vl_interface': 7.729674098024211e-06, 'learning_rate/action_model': 7.729674098024212e-05, 'epoch': 0.1}) +[2026-06-24 10:20:11,288][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.0902867540717125, 'train/grad_norm_pre_clip': 41.06058883666992, 'timing/data': 0.00035832193680107594, 'timing/model': 1.0599316868465394, 'learning_rate/qwen_vl_interface': 7.725283747388136e-06, 'learning_rate/action_model': 7.725283747388135e-05, 'epoch': 0.1}) +[2026-06-24 10:20:12,352][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.20792986452579498, 'train/grad_norm_pre_clip': 59.18809509277344, 'timing/data': 0.0004220518749207258, 'timing/model': 1.061986766057089, 'learning_rate/qwen_vl_interface': 7.72089049186871e-06, 'learning_rate/action_model': 7.72089049186871e-05, 'epoch': 0.1}) +[2026-06-24 10:20:13,416][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.15697930753231049, 'train/grad_norm_pre_clip': 48.72318649291992, 'timing/data': 0.000294883968308568, 'timing/model': 1.0622074590064585, 'learning_rate/qwen_vl_interface': 7.716494336621662e-06, 'learning_rate/action_model': 7.716494336621663e-05, 'epoch': 0.1}) +[2026-06-24 10:20:14,479][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.11201391369104385, 'train/grad_norm_pre_clip': 26.274457931518555, 'timing/data': 0.0003306809812784195, 'timing/model': 1.0617315869312733, 'learning_rate/qwen_vl_interface': 7.71209528680613e-06, 'learning_rate/action_model': 7.71209528680613e-05, 'epoch': 0.1}) +[2026-06-24 10:20:15,544][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.24430522322654724, 'train/grad_norm_pre_clip': 35.29084014892578, 'timing/data': 0.0003120249602943659, 'timing/model': 1.0631055650301278, 'learning_rate/qwen_vl_interface': 7.707693347584639e-06, 'learning_rate/action_model': 7.707693347584639e-05, 'epoch': 0.1}) +[2026-06-24 10:20:16,608][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.03713749349117279, 'train/grad_norm_pre_clip': 27.500844955444336, 'timing/data': 0.00030777789652347565, 'timing/model': 1.0632519300561398, 'learning_rate/qwen_vl_interface': 7.703288524123114e-06, 'learning_rate/action_model': 7.703288524123114e-05, 'epoch': 0.1}) +[2026-06-24 10:20:17,961][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.037998855113983154, 'train/grad_norm_pre_clip': 13.935029029846191, 'timing/data': 0.0003197400365024805, 'timing/model': 1.351520051015541, 'learning_rate/qwen_vl_interface': 7.698880821590854e-06, 'learning_rate/action_model': 7.698880821590853e-05, 'epoch': 0.1}) +[2026-06-24 10:20:19,024][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.18267741799354553, 'train/grad_norm_pre_clip': 30.020469665527344, 'timing/data': 0.0003204511012881994, 'timing/model': 1.0611344559583813, 'learning_rate/qwen_vl_interface': 7.694470245160551e-06, 'learning_rate/action_model': 7.69447024516055e-05, 'epoch': 0.1}) +[2026-06-24 10:20:20,088][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.12467347085475922, 'train/grad_norm_pre_clip': 65.9233627319336, 'timing/data': 0.00030453200452029705, 'timing/model': 1.0620353671256453, 'learning_rate/qwen_vl_interface': 7.690056800008254e-06, 'learning_rate/action_model': 7.690056800008254e-05, 'epoch': 0.1}) +[2026-06-24 10:20:21,151][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.02995440922677517, 'train/grad_norm_pre_clip': 14.338712692260742, 'timing/data': 0.00030772690661251545, 'timing/model': 1.0620794699061662, 'learning_rate/qwen_vl_interface': 7.685640491313395e-06, 'learning_rate/action_model': 7.685640491313394e-05, 'epoch': 0.1}) +[2026-06-24 10:20:22,215][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.09791053831577301, 'train/grad_norm_pre_clip': 32.27986526489258, 'timing/data': 0.0003318230155855417, 'timing/model': 1.0620556159410626, 'learning_rate/qwen_vl_interface': 7.681221324258755e-06, 'learning_rate/action_model': 7.681221324258755e-05, 'epoch': 0.1}) +[2026-06-24 10:20:23,283][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.09678319096565247, 'train/grad_norm_pre_clip': 41.141998291015625, 'timing/data': 0.0003103818744421005, 'timing/model': 1.0665400219149888, 'learning_rate/qwen_vl_interface': 7.67679930403047e-06, 'learning_rate/action_model': 7.67679930403047e-05, 'epoch': 0.1}) +[2026-06-24 10:20:24,348][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.03989047557115555, 'train/grad_norm_pre_clip': 38.21359634399414, 'timing/data': 0.00033304491080343723, 'timing/model': 1.062531858915463, 'learning_rate/qwen_vl_interface': 7.672374435818032e-06, 'learning_rate/action_model': 7.672374435818031e-05, 'epoch': 0.1}) +[2026-06-24 10:20:25,413][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.14993612468242645, 'train/grad_norm_pre_clip': 39.833251953125, 'timing/data': 0.0003381140995770693, 'timing/model': 1.063430316047743, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.1}) +[2026-06-24 10:20:26,478][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.1828516572713852, 'train/grad_norm_pre_clip': 66.32858276367188, 'timing/data': 0.000310912961140275, 'timing/model': 1.063881813082844, 'learning_rate/qwen_vl_interface': 7.663516176215338e-06, 'learning_rate/action_model': 7.663516176215337e-05, 'epoch': 0.1}) +[2026-06-24 10:20:27,544][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.04984675720334053, 'train/grad_norm_pre_clip': 14.413284301757812, 'timing/data': 0.0003302700351923704, 'timing/model': 1.0638305880129337, 'learning_rate/qwen_vl_interface': 7.659082795220747e-06, 'learning_rate/action_model': 7.659082795220746e-05, 'epoch': 0.1}) +[2026-06-24 10:20:28,608][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.07381533086299896, 'train/grad_norm_pre_clip': 46.72551345825195, 'timing/data': 0.0003067459911108017, 'timing/model': 1.062929887091741, 'learning_rate/qwen_vl_interface': 7.654646587033308e-06, 'learning_rate/action_model': 7.654646587033307e-05, 'epoch': 0.1}) +[2026-06-24 10:20:29,674][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.00856661144644022, 'train/grad_norm_pre_clip': 4.925883769989014, 'timing/data': 0.0003544050268828869, 'timing/model': 1.063963467022404, 'learning_rate/qwen_vl_interface': 7.650207556859162e-06, 'learning_rate/action_model': 7.650207556859161e-05, 'epoch': 0.1}) +[2026-06-24 10:20:30,742][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.036775581538677216, 'train/grad_norm_pre_clip': 22.22266387939453, 'timing/data': 0.00031153508462011814, 'timing/model': 1.0664535900577903, 'learning_rate/qwen_vl_interface': 7.645765709907756e-06, 'learning_rate/action_model': 7.645765709907756e-05, 'epoch': 0.1}) +[2026-06-24 10:20:31,807][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.0656435638666153, 'train/grad_norm_pre_clip': 94.085205078125, 'timing/data': 0.0003234769683331251, 'timing/model': 1.0640377060044557, 'learning_rate/qwen_vl_interface': 7.641321051391845e-06, 'learning_rate/action_model': 7.641321051391846e-05, 'epoch': 0.1}) +[2026-06-24 10:20:32,872][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.14760944247245789, 'train/grad_norm_pre_clip': 42.93208694458008, 'timing/data': 0.00034322403371334076, 'timing/model': 1.062764175934717, 'learning_rate/qwen_vl_interface': 7.636873586527484e-06, 'learning_rate/action_model': 7.636873586527484e-05, 'epoch': 0.1}) +[2026-06-24 10:20:33,935][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.17411354184150696, 'train/grad_norm_pre_clip': 58.932647705078125, 'timing/data': 0.0003392768558114767, 'timing/model': 1.0614251329097897, 'learning_rate/qwen_vl_interface': 7.632423320534023e-06, 'learning_rate/action_model': 7.632423320534022e-05, 'epoch': 0.1}) +[2026-06-24 10:20:34,999][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.20999659597873688, 'train/grad_norm_pre_clip': 73.58500671386719, 'timing/data': 0.00030853901989758015, 'timing/model': 1.0624763960950077, 'learning_rate/qwen_vl_interface': 7.627970258634093e-06, 'learning_rate/action_model': 7.627970258634093e-05, 'epoch': 0.1}) +[2026-06-24 10:20:36,308][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.2568255066871643, 'train/grad_norm_pre_clip': 67.42247009277344, 'timing/data': 0.00029997294768691063, 'timing/model': 1.30710813193582, 'learning_rate/qwen_vl_interface': 7.623514406053613e-06, 'learning_rate/action_model': 7.623514406053613e-05, 'epoch': 0.1}) +[2026-06-24 10:20:37,371][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.004770426545292139, 'train/grad_norm_pre_clip': 5.539901256561279, 'timing/data': 0.00033269403502345085, 'timing/model': 1.0618730830028653, 'learning_rate/qwen_vl_interface': 7.619055768021772e-06, 'learning_rate/action_model': 7.619055768021772e-05, 'epoch': 0.1}) +[2026-06-24 10:20:38,436][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.46547284722328186, 'train/grad_norm_pre_clip': 59.77103042602539, 'timing/data': 0.00032512005418539047, 'timing/model': 1.0632286090403795, 'learning_rate/qwen_vl_interface': 7.6145943497710325e-06, 'learning_rate/action_model': 7.614594349771032e-05, 'epoch': 0.1}) +[2026-06-24 10:20:39,499][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.1362985223531723, 'train/grad_norm_pre_clip': 57.05868148803711, 'timing/data': 0.00033010891638696194, 'timing/model': 1.0617306558415294, 'learning_rate/qwen_vl_interface': 7.610130156537114e-06, 'learning_rate/action_model': 7.610130156537113e-05, 'epoch': 0.1}) +[2026-06-24 10:20:40,563][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.13673987984657288, 'train/grad_norm_pre_clip': 43.51449966430664, 'timing/data': 0.0003287368454039097, 'timing/model': 1.0619474539998919, 'learning_rate/qwen_vl_interface': 7.605663193558998e-06, 'learning_rate/action_model': 7.605663193558999e-05, 'epoch': 0.1}) +[2026-06-24 10:20:41,629][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.01279602013528347, 'train/grad_norm_pre_clip': 13.89091968536377, 'timing/data': 0.00032669282518327236, 'timing/model': 1.0641977980267256, 'learning_rate/qwen_vl_interface': 7.601193466078915e-06, 'learning_rate/action_model': 7.601193466078914e-05, 'epoch': 0.1}) +[2026-06-24 10:20:42,692][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.07142048329114914, 'train/grad_norm_pre_clip': 60.24633026123047, 'timing/data': 0.00032110302709043026, 'timing/model': 1.0619711889885366, 'learning_rate/qwen_vl_interface': 7.596720979342339e-06, 'learning_rate/action_model': 7.596720979342338e-05, 'epoch': 0.1}) +[2026-06-24 10:20:43,760][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.08806321024894714, 'train/grad_norm_pre_clip': 33.88711166381836, 'timing/data': 0.0003223950043320656, 'timing/model': 1.0661336118355393, 'learning_rate/qwen_vl_interface': 7.592245738597981e-06, 'learning_rate/action_model': 7.592245738597981e-05, 'epoch': 0.1}) +[2026-06-24 10:20:44,825][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.028216496109962463, 'train/grad_norm_pre_clip': 16.181127548217773, 'timing/data': 0.0003293680492788553, 'timing/model': 1.0635039259213954, 'learning_rate/qwen_vl_interface': 7.587767749097785e-06, 'learning_rate/action_model': 7.587767749097785e-05, 'epoch': 0.1}) +[2026-06-24 10:20:45,887][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.07027383893728256, 'train/grad_norm_pre_clip': 25.66974449157715, 'timing/data': 0.00030981097370386124, 'timing/model': 1.060368019156158, 'learning_rate/qwen_vl_interface': 7.583287016096923e-06, 'learning_rate/action_model': 7.583287016096923e-05, 'epoch': 0.1}) +[2026-06-24 10:20:46,953][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.12542755901813507, 'train/grad_norm_pre_clip': 33.82460021972656, 'timing/data': 0.00032135401852428913, 'timing/model': 1.0649673629086465, 'learning_rate/qwen_vl_interface': 7.578803544853786e-06, 'learning_rate/action_model': 7.578803544853785e-05, 'epoch': 0.1}) +[2026-06-24 10:20:48,013][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.2450992614030838, 'train/grad_norm_pre_clip': 52.96609115600586, 'timing/data': 0.0003051930107176304, 'timing/model': 1.057854101061821, 'learning_rate/qwen_vl_interface': 7.574317340629972e-06, 'learning_rate/action_model': 7.574317340629972e-05, 'epoch': 0.1}) +[2026-06-24 10:20:49,082][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.16533340513706207, 'train/grad_norm_pre_clip': 197.74099731445312, 'timing/data': 0.0003124261274933815, 'timing/model': 1.067114703124389, 'learning_rate/qwen_vl_interface': 7.569828408690295e-06, 'learning_rate/action_model': 7.569828408690295e-05, 'epoch': 0.1}) +[2026-06-24 10:20:50,146][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.28773602843284607, 'train/grad_norm_pre_clip': 63.6342658996582, 'timing/data': 0.00046130595728754997, 'timing/model': 1.062598466174677, 'learning_rate/qwen_vl_interface': 7.565336754302766e-06, 'learning_rate/action_model': 7.565336754302766e-05, 'epoch': 0.1}) +[2026-06-24 10:20:51,212][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.15091335773468018, 'train/grad_norm_pre_clip': 46.75183868408203, 'timing/data': 0.00034331390634179115, 'timing/model': 1.0644842758774757, 'learning_rate/qwen_vl_interface': 7.5608423827385936e-06, 'learning_rate/action_model': 7.560842382738593e-05, 'epoch': 0.1}) +[2026-06-24 10:20:52,278][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.17765690386295319, 'train/grad_norm_pre_clip': 35.02250671386719, 'timing/data': 0.00033975811675190926, 'timing/model': 1.064143736846745, 'learning_rate/qwen_vl_interface': 7.556345299272172e-06, 'learning_rate/action_model': 7.556345299272171e-05, 'epoch': 0.1}) +[2026-06-24 10:20:53,341][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.09279347211122513, 'train/grad_norm_pre_clip': 47.54882049560547, 'timing/data': 0.0003223151434212923, 'timing/model': 1.0617792690172791, 'learning_rate/qwen_vl_interface': 7.551845509181076e-06, 'learning_rate/action_model': 7.551845509181076e-05, 'epoch': 0.1}) +[2026-06-24 10:20:54,610][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.3197467625141144, 'train/grad_norm_pre_clip': 35.81618118286133, 'timing/data': 0.00038147601298987865, 'timing/model': 1.267493115970865, 'learning_rate/qwen_vl_interface': 7.547343017746063e-06, 'learning_rate/action_model': 7.547343017746063e-05, 'epoch': 0.1}) +[2026-06-24 10:20:55,674][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.115455761551857, 'train/grad_norm_pre_clip': 21.237680435180664, 'timing/data': 0.0004892479628324509, 'timing/model': 1.062143734190613, 'learning_rate/qwen_vl_interface': 7.542837830251058e-06, 'learning_rate/action_model': 7.542837830251058e-05, 'epoch': 0.1}) +[2026-06-24 10:20:56,736][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.12019068747758865, 'train/grad_norm_pre_clip': 50.41575241088867, 'timing/data': 0.00036650802940130234, 'timing/model': 1.0603179361205548, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.11}) +[2026-06-24 10:20:57,801][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.11739780753850937, 'train/grad_norm_pre_clip': 34.03205871582031, 'timing/data': 0.000344857107847929, 'timing/model': 1.0631711019668728, 'learning_rate/qwen_vl_interface': 7.53381938823258e-06, 'learning_rate/action_model': 7.53381938823258e-05, 'epoch': 0.11}) +[2026-06-24 10:20:58,866][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.10849064588546753, 'train/grad_norm_pre_clip': 47.33754348754883, 'timing/data': 0.0003036700654774904, 'timing/model': 1.0638035689480603, 'learning_rate/qwen_vl_interface': 7.529306144292752e-06, 'learning_rate/action_model': 7.529306144292751e-05, 'epoch': 0.11}) +[2026-06-24 10:20:59,931][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.04933462664484978, 'train/grad_norm_pre_clip': 32.15186309814453, 'timing/data': 0.0003165940288454294, 'timing/model': 1.0626585599966347, 'learning_rate/qwen_vl_interface': 7.524790225460206e-06, 'learning_rate/action_model': 7.524790225460205e-05, 'epoch': 0.11}) +[2026-06-24 10:21:00,995][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.1045595183968544, 'train/grad_norm_pre_clip': 65.47058868408203, 'timing/data': 0.0002981899306178093, 'timing/model': 1.063199314987287, 'learning_rate/qwen_vl_interface': 7.520271637034624e-06, 'learning_rate/action_model': 7.520271637034624e-05, 'epoch': 0.11}) +[2026-06-24 10:21:02,061][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.21708865463733673, 'train/grad_norm_pre_clip': 49.463050842285156, 'timing/data': 0.00030449102632701397, 'timing/model': 1.0637706669513136, 'learning_rate/qwen_vl_interface': 7.515750384318824e-06, 'learning_rate/action_model': 7.515750384318824e-05, 'epoch': 0.11}) +[2026-06-24 10:21:03,124][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.24490979313850403, 'train/grad_norm_pre_clip': 47.87176513671875, 'timing/data': 0.00033841608092188835, 'timing/model': 1.061494184890762, 'learning_rate/qwen_vl_interface': 7.5112264726187445e-06, 'learning_rate/action_model': 7.511226472618745e-05, 'epoch': 0.11}) +[2026-06-24 10:21:04,185][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.17900025844573975, 'train/grad_norm_pre_clip': 79.95515441894531, 'timing/data': 0.0002990711946040392, 'timing/model': 1.0596756010781974, 'learning_rate/qwen_vl_interface': 7.5066999072434485e-06, 'learning_rate/action_model': 7.506699907243449e-05, 'epoch': 0.11}) +[2026-06-24 10:21:05,248][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.038433726876974106, 'train/grad_norm_pre_clip': 9.262776374816895, 'timing/data': 0.00031648389995098114, 'timing/model': 1.061736810952425, 'learning_rate/qwen_vl_interface': 7.502170693505116e-06, 'learning_rate/action_model': 7.502170693505115e-05, 'epoch': 0.11}) +[2026-06-24 10:21:06,314][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.1305200308561325, 'train/grad_norm_pre_clip': 34.8382453918457, 'timing/data': 0.000329198082908988, 'timing/model': 1.0638751341030002, 'learning_rate/qwen_vl_interface': 7.497638836719027e-06, 'learning_rate/action_model': 7.497638836719026e-05, 'epoch': 0.11}) +[2026-06-24 10:21:07,378][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.31832757592201233, 'train/grad_norm_pre_clip': 57.18406677246094, 'timing/data': 0.00031345896422863007, 'timing/model': 1.0626769049558789, 'learning_rate/qwen_vl_interface': 7.493104342203571e-06, 'learning_rate/action_model': 7.49310434220357e-05, 'epoch': 0.11}) +[2026-06-24 10:21:08,446][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.1637818068265915, 'train/grad_norm_pre_clip': 39.169227600097656, 'timing/data': 0.0003044821787625551, 'timing/model': 1.0665263310074806, 'learning_rate/qwen_vl_interface': 7.488567215280228e-06, 'learning_rate/action_model': 7.488567215280227e-05, 'epoch': 0.11}) +[2026-06-24 10:21:09,511][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.14453944563865662, 'train/grad_norm_pre_clip': 42.8239860534668, 'timing/data': 0.00032538105733692646, 'timing/model': 1.0638435648288578, 'learning_rate/qwen_vl_interface': 7.4840274612735716e-06, 'learning_rate/action_model': 7.484027461273572e-05, 'epoch': 0.11}) +[2026-06-24 10:21:10,578][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.0826081782579422, 'train/grad_norm_pre_clip': 25.261289596557617, 'timing/data': 0.00031296699307858944, 'timing/model': 1.0651035499759018, 'learning_rate/qwen_vl_interface': 7.4794850855112514e-06, 'learning_rate/action_model': 7.479485085511251e-05, 'epoch': 0.11}) +[2026-06-24 10:21:11,851][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.27023980021476746, 'train/grad_norm_pre_clip': 49.154117584228516, 'timing/data': 0.0003050919622182846, 'timing/model': 1.2717450179625303, 'learning_rate/qwen_vl_interface': 7.4749400933240035e-06, 'learning_rate/action_model': 7.474940093324003e-05, 'epoch': 0.11}) +[2026-06-24 10:21:12,917][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.10354423522949219, 'train/grad_norm_pre_clip': 25.182085037231445, 'timing/data': 0.000308929942548275, 'timing/model': 1.0635199879761785, 'learning_rate/qwen_vl_interface': 7.470392490045628e-06, 'learning_rate/action_model': 7.470392490045628e-05, 'epoch': 0.11}) +[2026-06-24 10:21:13,981][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.29855722188949585, 'train/grad_norm_pre_clip': 66.9685287475586, 'timing/data': 0.0003916958812624216, 'timing/model': 1.0631754810456187, 'learning_rate/qwen_vl_interface': 7.4658422810129885e-06, 'learning_rate/action_model': 7.465842281012987e-05, 'epoch': 0.11}) +[2026-06-24 10:21:15,048][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.08912910521030426, 'train/grad_norm_pre_clip': 54.0862922668457, 'timing/data': 0.00029023410752415657, 'timing/model': 1.0652712951414287, 'learning_rate/qwen_vl_interface': 7.46128947156601e-06, 'learning_rate/action_model': 7.46128947156601e-05, 'epoch': 0.11}) +[2026-06-24 10:21:16,113][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.15403568744659424, 'train/grad_norm_pre_clip': 33.87409210205078, 'timing/data': 0.0003134279977530241, 'timing/model': 1.0632038549520075, 'learning_rate/qwen_vl_interface': 7.456734067047669e-06, 'learning_rate/action_model': 7.456734067047668e-05, 'epoch': 0.11}) +[2026-06-24 10:21:17,179][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.023572800680994987, 'train/grad_norm_pre_clip': 7.175469875335693, 'timing/data': 0.00034988694824278355, 'timing/model': 1.063851109938696, 'learning_rate/qwen_vl_interface': 7.452176072803985e-06, 'learning_rate/action_model': 7.452176072803985e-05, 'epoch': 0.11}) +[2026-06-24 10:21:18,244][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.20882093906402588, 'train/grad_norm_pre_clip': 64.73126983642578, 'timing/data': 0.00033087097108364105, 'timing/model': 1.0638315330725163, 'learning_rate/qwen_vl_interface': 7.447615494184018e-06, 'learning_rate/action_model': 7.447615494184018e-05, 'epoch': 0.11}) +[2026-06-24 10:21:19,309][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.1580328792333603, 'train/grad_norm_pre_clip': 36.16532897949219, 'timing/data': 0.00029093585908412933, 'timing/model': 1.0635159008670598, 'learning_rate/qwen_vl_interface': 7.443052336539862e-06, 'learning_rate/action_model': 7.443052336539862e-05, 'epoch': 0.11}) +[2026-06-24 10:21:20,373][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.3602004647254944, 'train/grad_norm_pre_clip': 61.27655029296875, 'timing/data': 0.0003201907966285944, 'timing/model': 1.0619836139958352, 'learning_rate/qwen_vl_interface': 7.438486605226637e-06, 'learning_rate/action_model': 7.438486605226637e-05, 'epoch': 0.11}) +[2026-06-24 10:21:21,435][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.05445870757102966, 'train/grad_norm_pre_clip': 28.07060432434082, 'timing/data': 0.0005021430552005768, 'timing/model': 1.0604147699195892, 'learning_rate/qwen_vl_interface': 7.4339183056024826e-06, 'learning_rate/action_model': 7.433918305602481e-05, 'epoch': 0.11}) +[2026-06-24 10:21:22,497][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.25457456707954407, 'train/grad_norm_pre_clip': 71.43284606933594, 'timing/data': 0.00030829920433461666, 'timing/model': 1.0602970789186656, 'learning_rate/qwen_vl_interface': 7.429347443028551e-06, 'learning_rate/action_model': 7.429347443028552e-05, 'epoch': 0.11}) +[2026-06-24 10:21:23,558][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.07363099604845047, 'train/grad_norm_pre_clip': 51.270530700683594, 'timing/data': 0.00034609995782375336, 'timing/model': 1.059722871053964, 'learning_rate/qwen_vl_interface': 7.424774022869008e-06, 'learning_rate/action_model': 7.424774022869008e-05, 'epoch': 0.11}) +[2026-06-24 10:21:24,621][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.045575570315122604, 'train/grad_norm_pre_clip': 27.621294021606445, 'timing/data': 0.00030756695196032524, 'timing/model': 1.0612742330413312, 'learning_rate/qwen_vl_interface': 7.420198050491013e-06, 'learning_rate/action_model': 7.420198050491012e-05, 'epoch': 0.11}) +[2026-06-24 10:21:25,688][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.06413257122039795, 'train/grad_norm_pre_clip': 86.73223114013672, 'timing/data': 0.00030903005972504616, 'timing/model': 1.0655763989780098, 'learning_rate/qwen_vl_interface': 7.415619531264726e-06, 'learning_rate/action_model': 7.415619531264725e-05, 'epoch': 0.11}) +[2026-06-24 10:21:26,751][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.3406834304332733, 'train/grad_norm_pre_clip': 49.38893508911133, 'timing/data': 0.00031234603375196457, 'timing/model': 1.0614115819334984, 'learning_rate/qwen_vl_interface': 7.4110384705632946e-06, 'learning_rate/action_model': 7.411038470563294e-05, 'epoch': 0.11}) +[2026-06-24 10:21:27,819][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.16611121594905853, 'train/grad_norm_pre_clip': 62.91531753540039, 'timing/data': 0.0003041911404579878, 'timing/model': 1.0666596021037549, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.11}) +[2026-06-24 10:21:29,080][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.16677068173885345, 'train/grad_norm_pre_clip': 41.241615295410156, 'timing/data': 0.0003252709284424782, 'timing/model': 1.2589259478263557, 'learning_rate/qwen_vl_interface': 7.401868746242495e-06, 'learning_rate/action_model': 7.401868746242494e-05, 'epoch': 0.11}) +[2026-06-24 10:21:30,145][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.21706874668598175, 'train/grad_norm_pre_clip': 53.2899284362793, 'timing/data': 0.0006069990340620279, 'timing/model': 1.0636756119783968, 'learning_rate/qwen_vl_interface': 7.397280093384306e-06, 'learning_rate/action_model': 7.397280093384305e-05, 'epoch': 0.11}) +[2026-06-24 10:21:31,211][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.11229602247476578, 'train/grad_norm_pre_clip': 31.877605438232422, 'timing/data': 0.0005147659685462713, 'timing/model': 1.063542861957103, 'learning_rate/qwen_vl_interface': 7.392688920573325e-06, 'learning_rate/action_model': 7.392688920573325e-05, 'epoch': 0.11}) +[2026-06-24 10:21:32,276][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.19298055768013, 'train/grad_norm_pre_clip': 50.687591552734375, 'timing/data': 0.000350557966157794, 'timing/model': 1.0634551369585097, 'learning_rate/qwen_vl_interface': 7.38809523319755e-06, 'learning_rate/action_model': 7.388095233197549e-05, 'epoch': 0.11}) +[2026-06-24 10:21:33,341][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.09900320321321487, 'train/grad_norm_pre_clip': 36.20541000366211, 'timing/data': 0.00031663384288549423, 'timing/model': 1.062862915918231, 'learning_rate/qwen_vl_interface': 7.383499036647928e-06, 'learning_rate/action_model': 7.383499036647927e-05, 'epoch': 0.11}) +[2026-06-24 10:21:34,408][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.024633321911096573, 'train/grad_norm_pre_clip': 6.171119689941406, 'timing/data': 0.0006361240521073341, 'timing/model': 1.0649835870135576, 'learning_rate/qwen_vl_interface': 7.378900336318347e-06, 'learning_rate/action_model': 7.378900336318348e-05, 'epoch': 0.11}) +[2026-06-24 10:21:35,473][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.021433450281620026, 'train/grad_norm_pre_clip': 6.963322162628174, 'timing/data': 0.00032263598404824734, 'timing/model': 1.0635711851064116, 'learning_rate/qwen_vl_interface': 7.374299137605645e-06, 'learning_rate/action_model': 7.374299137605644e-05, 'epoch': 0.11}) +[2026-06-24 10:21:36,542][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.2983401119709015, 'train/grad_norm_pre_clip': 55.6798210144043, 'timing/data': 0.0003406889736652374, 'timing/model': 1.0678748830687255, 'learning_rate/qwen_vl_interface': 7.369695445909581e-06, 'learning_rate/action_model': 7.36969544590958e-05, 'epoch': 0.11}) +[2026-06-24 10:21:37,605][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.11836860328912735, 'train/grad_norm_pre_clip': 33.54908752441406, 'timing/data': 0.00034761312417685986, 'timing/model': 1.0612201830372214, 'learning_rate/qwen_vl_interface': 7.365089266632844e-06, 'learning_rate/action_model': 7.365089266632844e-05, 'epoch': 0.11}) +[2026-06-24 10:21:38,666][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.279371976852417, 'train/grad_norm_pre_clip': 81.88375091552734, 'timing/data': 0.0004547040443867445, 'timing/model': 1.0590400211513042, 'learning_rate/qwen_vl_interface': 7.360480605181043e-06, 'learning_rate/action_model': 7.360480605181043e-05, 'epoch': 0.11}) +[2026-06-24 10:21:39,729][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.1232999786734581, 'train/grad_norm_pre_clip': 33.383174896240234, 'timing/data': 0.0003501768223941326, 'timing/model': 1.06148229399696, 'learning_rate/qwen_vl_interface': 7.355869466962697e-06, 'learning_rate/action_model': 7.355869466962697e-05, 'epoch': 0.11}) +[2026-06-24 10:21:40,795][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.3757582902908325, 'train/grad_norm_pre_clip': 92.95552062988281, 'timing/data': 0.0003609668929129839, 'timing/model': 1.063765280181542, 'learning_rate/qwen_vl_interface': 7.3512558573892376e-06, 'learning_rate/action_model': 7.351255857389238e-05, 'epoch': 0.11}) +[2026-06-24 10:21:41,862][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.12100958079099655, 'train/grad_norm_pre_clip': 50.81910705566406, 'timing/data': 0.0003097511362284422, 'timing/model': 1.0659837538842112, 'learning_rate/qwen_vl_interface': 7.346639781874987e-06, 'learning_rate/action_model': 7.346639781874987e-05, 'epoch': 0.11}) +[2026-06-24 10:21:42,927][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.4670866131782532, 'train/grad_norm_pre_clip': 59.46274948120117, 'timing/data': 0.0003311911132186651, 'timing/model': 1.0629198630340397, 'learning_rate/qwen_vl_interface': 7.342021245837172e-06, 'learning_rate/action_model': 7.342021245837172e-05, 'epoch': 0.11}) +[2026-06-24 10:21:43,992][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.06486216932535172, 'train/grad_norm_pre_clip': 20.10399055480957, 'timing/data': 0.0003302099648863077, 'timing/model': 1.063850220059976, 'learning_rate/qwen_vl_interface': 7.337400254695897e-06, 'learning_rate/action_model': 7.337400254695897e-05, 'epoch': 0.11}) +[2026-06-24 10:21:45,058][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.2023458182811737, 'train/grad_norm_pre_clip': 168.7707977294922, 'timing/data': 0.000304951798170805, 'timing/model': 1.0638893127907068, 'learning_rate/qwen_vl_interface': 7.332776813874158e-06, 'learning_rate/action_model': 7.332776813874158e-05, 'epoch': 0.11}) +[2026-06-24 10:21:46,384][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.19941528141498566, 'train/grad_norm_pre_clip': 77.3907241821289, 'timing/data': 0.0003068558871746063, 'timing/model': 1.3246596129611135, 'learning_rate/qwen_vl_interface': 7.328150928797817e-06, 'learning_rate/action_model': 7.328150928797817e-05, 'epoch': 0.11}) +[2026-06-24 10:21:47,511][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.3067938983440399, 'train/grad_norm_pre_clip': 89.08020782470703, 'timing/data': 0.00032923812977969646, 'timing/model': 1.057657094905153, 'learning_rate/qwen_vl_interface': 7.323522604895609e-06, 'learning_rate/action_model': 7.323522604895608e-05, 'epoch': 0.11}) +[2026-06-24 10:21:48,582][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.6387829780578613, 'train/grad_norm_pre_clip': 102.69136047363281, 'timing/data': 0.0005714120343327522, 'timing/model': 1.0693892871495336, 'learning_rate/qwen_vl_interface': 7.318891847599127e-06, 'learning_rate/action_model': 7.318891847599126e-05, 'epoch': 0.11}) +[2026-06-24 10:21:49,647][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.24703659117221832, 'train/grad_norm_pre_clip': 76.73797607421875, 'timing/data': 0.0002956148236989975, 'timing/model': 1.063039899803698, 'learning_rate/qwen_vl_interface': 7.314258662342826e-06, 'learning_rate/action_model': 7.314258662342826e-05, 'epoch': 0.11}) +[2026-06-24 10:21:50,711][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.04250118136405945, 'train/grad_norm_pre_clip': 23.6944580078125, 'timing/data': 0.0003033189568668604, 'timing/model': 1.0620589079335332, 'learning_rate/qwen_vl_interface': 7.309623054564006e-06, 'learning_rate/action_model': 7.309623054564005e-05, 'epoch': 0.11}) +[2026-06-24 10:21:51,776][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.053616710007190704, 'train/grad_norm_pre_clip': 13.725454330444336, 'timing/data': 0.0005651311948895454, 'timing/model': 1.0631266820710152, 'learning_rate/qwen_vl_interface': 7.304985029702809e-06, 'learning_rate/action_model': 7.304985029702808e-05, 'epoch': 0.11}) +[2026-06-24 10:21:52,842][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.06833578646183014, 'train/grad_norm_pre_clip': 19.65363121032715, 'timing/data': 0.00030666496604681015, 'timing/model': 1.0647822499740869, 'learning_rate/qwen_vl_interface': 7.3003445932022195e-06, 'learning_rate/action_model': 7.30034459320222e-05, 'epoch': 0.11}) +[2026-06-24 10:21:53,907][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.16459082067012787, 'train/grad_norm_pre_clip': 37.26160430908203, 'timing/data': 0.0003171060234308243, 'timing/model': 1.0636003219988197, 'learning_rate/qwen_vl_interface': 7.295701750508045e-06, 'learning_rate/action_model': 7.295701750508045e-05, 'epoch': 0.11}) +[2026-06-24 10:21:54,973][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.17368610203266144, 'train/grad_norm_pre_clip': 35.608985900878906, 'timing/data': 0.0003759248647838831, 'timing/model': 1.0640703940298408, 'learning_rate/qwen_vl_interface': 7.2910565070689225e-06, 'learning_rate/action_model': 7.291056507068922e-05, 'epoch': 0.11}) +[2026-06-24 10:21:56,037][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.11973831802606583, 'train/grad_norm_pre_clip': 62.31425476074219, 'timing/data': 0.0003308909945189953, 'timing/model': 1.0624318679329008, 'learning_rate/qwen_vl_interface': 7.2864088683363e-06, 'learning_rate/action_model': 7.286408868336299e-05, 'epoch': 0.11}) +[2026-06-24 10:21:57,104][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.06716716289520264, 'train/grad_norm_pre_clip': 23.404815673828125, 'timing/data': 0.00032489001750946045, 'timing/model': 1.0654179430566728, 'learning_rate/qwen_vl_interface': 7.281758839764444e-06, 'learning_rate/action_model': 7.281758839764443e-05, 'epoch': 0.11}) +[2026-06-24 10:21:58,171][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.18493713438510895, 'train/grad_norm_pre_clip': 52.86148452758789, 'timing/data': 0.0005643491167575121, 'timing/model': 1.0644664780702442, 'learning_rate/qwen_vl_interface': 7.277106426810422e-06, 'learning_rate/action_model': 7.277106426810421e-05, 'epoch': 0.11}) +[2026-06-24 10:21:59,241][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.14670048654079437, 'train/grad_norm_pre_clip': 38.71767044067383, 'timing/data': 0.00035314285196363926, 'timing/model': 1.0689641300123185, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.11}) +[2026-06-24 10:22:00,309][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.096866674721241, 'train/grad_norm_pre_clip': 46.34370422363281, 'timing/data': 0.00034027802757918835, 'timing/model': 1.0662982258945704, 'learning_rate/qwen_vl_interface': 7.267794469598128e-06, 'learning_rate/action_model': 7.267794469598128e-05, 'epoch': 0.11}) +[2026-06-24 10:22:01,375][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.03551531955599785, 'train/grad_norm_pre_clip': 26.975719451904297, 'timing/data': 0.0003301100805401802, 'timing/model': 1.0639611999504268, 'learning_rate/qwen_vl_interface': 7.2631349362679585e-06, 'learning_rate/action_model': 7.263134936267958e-05, 'epoch': 0.11}) +[2026-06-24 10:22:02,661][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.06433047354221344, 'train/grad_norm_pre_clip': 22.467899322509766, 'timing/data': 0.0003372929058969021, 'timing/model': 1.284364779945463, 'learning_rate/qwen_vl_interface': 7.25847304041181e-06, 'learning_rate/action_model': 7.25847304041181e-05, 'epoch': 0.11}) +[2026-06-24 10:22:03,726][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.1861090064048767, 'train/grad_norm_pre_clip': 53.19921875, 'timing/data': 0.00031986087560653687, 'timing/model': 1.0640301590319723, 'learning_rate/qwen_vl_interface': 7.253808787500677e-06, 'learning_rate/action_model': 7.253808787500677e-05, 'epoch': 0.11}) +[2026-06-24 10:22:04,791][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.2596059739589691, 'train/grad_norm_pre_clip': 60.44114303588867, 'timing/data': 0.0003698039799928665, 'timing/model': 1.0631681110244244, 'learning_rate/qwen_vl_interface': 7.2491421830083196e-06, 'learning_rate/action_model': 7.24914218300832e-05, 'epoch': 0.11}) +[2026-06-24 10:22:05,858][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.11450562626123428, 'train/grad_norm_pre_clip': 45.3682746887207, 'timing/data': 0.00033391593024134636, 'timing/model': 1.0655724538955837, 'learning_rate/qwen_vl_interface': 7.2444732324112575e-06, 'learning_rate/action_model': 7.244473232411258e-05, 'epoch': 0.11}) +[2026-06-24 10:22:06,925][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.13954176008701324, 'train/grad_norm_pre_clip': 34.52975082397461, 'timing/data': 0.000331762945279479, 'timing/model': 1.0656492679845542, 'learning_rate/qwen_vl_interface': 7.239801941188766e-06, 'learning_rate/action_model': 7.239801941188766e-05, 'epoch': 0.11}) +[2026-06-24 10:22:07,990][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.22196902334690094, 'train/grad_norm_pre_clip': 49.93816375732422, 'timing/data': 0.0003150210250169039, 'timing/model': 1.063447925960645, 'learning_rate/qwen_vl_interface': 7.235128314822865e-06, 'learning_rate/action_model': 7.235128314822864e-05, 'epoch': 0.11}) +[2026-06-24 10:22:09,059][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.17528986930847168, 'train/grad_norm_pre_clip': 47.001502990722656, 'timing/data': 0.00033704214729368687, 'timing/model': 1.0675226340536028, 'learning_rate/qwen_vl_interface': 7.2304523587983165e-06, 'learning_rate/action_model': 7.230452358798317e-05, 'epoch': 0.11}) +[2026-06-24 10:22:10,125][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.13693049550056458, 'train/grad_norm_pre_clip': 58.572242736816406, 'timing/data': 0.0004985257983207703, 'timing/model': 1.0641437121666968, 'learning_rate/qwen_vl_interface': 7.225774078602613e-06, 'learning_rate/action_model': 7.225774078602613e-05, 'epoch': 0.11}) +[2026-06-24 10:22:11,192][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.1608458161354065, 'train/grad_norm_pre_clip': 51.170326232910156, 'timing/data': 0.0003017769195139408, 'timing/model': 1.0646869919728488, 'learning_rate/qwen_vl_interface': 7.221093479725977e-06, 'learning_rate/action_model': 7.221093479725977e-05, 'epoch': 0.11}) +[2026-06-24 10:22:12,257][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.2931407392024994, 'train/grad_norm_pre_clip': 72.10586547851562, 'timing/data': 0.0004025059752166271, 'timing/model': 1.0638621640391648, 'learning_rate/qwen_vl_interface': 7.216410567661355e-06, 'learning_rate/action_model': 7.216410567661355e-05, 'epoch': 0.11}) +[2026-06-24 10:22:13,319][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.151829332113266, 'train/grad_norm_pre_clip': 25.254796981811523, 'timing/data': 0.0002971179783344269, 'timing/model': 1.0605606569442898, 'learning_rate/qwen_vl_interface': 7.2117253479044004e-06, 'learning_rate/action_model': 7.2117253479044e-05, 'epoch': 0.11}) +[2026-06-24 10:22:14,385][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.09721089899539948, 'train/grad_norm_pre_clip': 13.88914680480957, 'timing/data': 0.0002993620000779629, 'timing/model': 1.0638947451952845, 'learning_rate/qwen_vl_interface': 7.207037825953485e-06, 'learning_rate/action_model': 7.207037825953484e-05, 'epoch': 0.11}) +[2026-06-24 10:22:15,451][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.2298922836780548, 'train/grad_norm_pre_clip': 49.69179916381836, 'timing/data': 0.00034332508221268654, 'timing/model': 1.064741875976324, 'learning_rate/qwen_vl_interface': 7.202348007309672e-06, 'learning_rate/action_model': 7.202348007309671e-05, 'epoch': 0.11}) +[2026-06-24 10:22:16,517][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.14686717092990875, 'train/grad_norm_pre_clip': 79.72686767578125, 'timing/data': 0.0003249300643801689, 'timing/model': 1.0635796540882438, 'learning_rate/qwen_vl_interface': 7.197655897476727e-06, 'learning_rate/action_model': 7.197655897476727e-05, 'epoch': 0.11}) +[2026-06-24 10:22:17,579][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.12126843631267548, 'train/grad_norm_pre_clip': 22.460622787475586, 'timing/data': 0.00031048315577208996, 'timing/model': 1.0612044960726053, 'learning_rate/qwen_vl_interface': 7.1929615019611024e-06, 'learning_rate/action_model': 7.192961501961101e-05, 'epoch': 0.11}) +[2026-06-24 10:22:18,645][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.07561817765235901, 'train/grad_norm_pre_clip': 26.104230880737305, 'timing/data': 0.00034285406582057476, 'timing/model': 1.0637770851608366, 'learning_rate/qwen_vl_interface': 7.1882648262719335e-06, 'learning_rate/action_model': 7.188264826271933e-05, 'epoch': 0.11}) +[2026-06-24 10:22:20,002][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.15993931889533997, 'train/grad_norm_pre_clip': 26.65869140625, 'timing/data': 0.0003123458009213209, 'timing/model': 1.3556314799934626, 'learning_rate/qwen_vl_interface': 7.183565875921029e-06, 'learning_rate/action_model': 7.183565875921029e-05, 'epoch': 0.11}) +[2026-06-24 10:22:21,066][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.022585583850741386, 'train/grad_norm_pre_clip': 6.138864994049072, 'timing/data': 0.0003034300170838833, 'timing/model': 1.0629061500076205, 'learning_rate/qwen_vl_interface': 7.178864656422869e-06, 'learning_rate/action_model': 7.17886465642287e-05, 'epoch': 0.11}) +[2026-06-24 10:22:22,130][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.12185055017471313, 'train/grad_norm_pre_clip': 33.74848175048828, 'timing/data': 0.0003112850245088339, 'timing/model': 1.0620662639848888, 'learning_rate/qwen_vl_interface': 7.174161173294599e-06, 'learning_rate/action_model': 7.174161173294599e-05, 'epoch': 0.11}) +[2026-06-24 10:22:23,198][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.11516813188791275, 'train/grad_norm_pre_clip': 44.577880859375, 'timing/data': 0.00033170287497341633, 'timing/model': 1.0662847820203751, 'learning_rate/qwen_vl_interface': 7.169455432056015e-06, 'learning_rate/action_model': 7.169455432056015e-05, 'epoch': 0.11}) +[2026-06-24 10:22:24,265][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.1473553627729416, 'train/grad_norm_pre_clip': 41.65970993041992, 'timing/data': 0.0003093911800533533, 'timing/model': 1.065740681020543, 'learning_rate/qwen_vl_interface': 7.16474743822957e-06, 'learning_rate/action_model': 7.16474743822957e-05, 'epoch': 0.11}) +[2026-06-24 10:22:25,330][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.18274128437042236, 'train/grad_norm_pre_clip': 50.616329193115234, 'timing/data': 0.0004988559521734715, 'timing/model': 1.0627901928965002, 'learning_rate/qwen_vl_interface': 7.160037197340354e-06, 'learning_rate/action_model': 7.160037197340354e-05, 'epoch': 0.11}) +[2026-06-24 10:22:26,394][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.1637745052576065, 'train/grad_norm_pre_clip': 39.92726516723633, 'timing/data': 0.00040251598693430424, 'timing/model': 1.0623814561404288, 'learning_rate/qwen_vl_interface': 7.155324714916098e-06, 'learning_rate/action_model': 7.155324714916097e-05, 'epoch': 0.11}) +[2026-06-24 10:22:27,460][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.1259589046239853, 'train/grad_norm_pre_clip': 24.391237258911133, 'timing/data': 0.00030554295517504215, 'timing/model': 1.0643135621212423, 'learning_rate/qwen_vl_interface': 7.1506099964871626e-06, 'learning_rate/action_model': 7.150609996487162e-05, 'epoch': 0.11}) +[2026-06-24 10:22:28,525][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.060336459428071976, 'train/grad_norm_pre_clip': 28.125633239746094, 'timing/data': 0.0002967570908367634, 'timing/model': 1.063724417006597, 'learning_rate/qwen_vl_interface': 7.145893047586532e-06, 'learning_rate/action_model': 7.145893047586532e-05, 'epoch': 0.11}) +[2026-06-24 10:22:29,588][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.3632025718688965, 'train/grad_norm_pre_clip': 67.8666000366211, 'timing/data': 0.0005467559676617384, 'timing/model': 1.0604673428460956, 'learning_rate/qwen_vl_interface': 7.14117387374981e-06, 'learning_rate/action_model': 7.14117387374981e-05, 'epoch': 0.11}) +[2026-06-24 10:22:30,652][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.0747208520770073, 'train/grad_norm_pre_clip': 35.188175201416016, 'timing/data': 0.00029534404166042805, 'timing/model': 1.0631294299382716, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.11}) +[2026-06-24 10:22:31,717][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.027854515239596367, 'train/grad_norm_pre_clip': 12.729538917541504, 'timing/data': 0.00045476295053958893, 'timing/model': 1.0632406889926642, 'learning_rate/qwen_vl_interface': 7.131728873423547e-06, 'learning_rate/action_model': 7.131728873423546e-05, 'epoch': 0.11}) +[2026-06-24 10:22:32,783][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.3004216253757477, 'train/grad_norm_pre_clip': 46.043704986572266, 'timing/data': 0.00033852504566311836, 'timing/model': 1.0641090681310743, 'learning_rate/qwen_vl_interface': 7.1270030580182396e-06, 'learning_rate/action_model': 7.127003058018239e-05, 'epoch': 0.11}) +[2026-06-24 10:22:33,848][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.08657939732074738, 'train/grad_norm_pre_clip': 32.91278076171875, 'timing/data': 0.00030248798429965973, 'timing/model': 1.0639097450766712, 'learning_rate/qwen_vl_interface': 7.122275039845296e-06, 'learning_rate/action_model': 7.122275039845295e-05, 'epoch': 0.11}) +[2026-06-24 10:22:34,910][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.13934609293937683, 'train/grad_norm_pre_clip': 62.147705078125, 'timing/data': 0.00030848896130919456, 'timing/model': 1.0603399639949203, 'learning_rate/qwen_vl_interface': 7.117544824453305e-06, 'learning_rate/action_model': 7.117544824453305e-05, 'epoch': 0.11}) +[2026-06-24 10:22:36,185][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.08014357089996338, 'train/grad_norm_pre_clip': 34.878108978271484, 'timing/data': 0.000506439944729209, 'timing/model': 1.2728593461215496, 'learning_rate/qwen_vl_interface': 7.112812417393442e-06, 'learning_rate/action_model': 7.112812417393441e-05, 'epoch': 0.11}) +[2026-06-24 10:22:37,251][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.22692514955997467, 'train/grad_norm_pre_clip': 69.60074615478516, 'timing/data': 0.00030507310293614864, 'timing/model': 1.0643096049316227, 'learning_rate/qwen_vl_interface': 7.108077824219448e-06, 'learning_rate/action_model': 7.108077824219447e-05, 'epoch': 0.11}) +[2026-06-24 10:22:38,318][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.049881771206855774, 'train/grad_norm_pre_clip': 48.44228744506836, 'timing/data': 0.0003238481003791094, 'timing/model': 1.0645462600514293, 'learning_rate/qwen_vl_interface': 7.103341050487635e-06, 'learning_rate/action_model': 7.103341050487636e-05, 'epoch': 0.11}) +[2026-06-24 10:22:39,381][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.22498458623886108, 'train/grad_norm_pre_clip': 89.47332000732422, 'timing/data': 0.0003440459258854389, 'timing/model': 1.0614729919470847, 'learning_rate/qwen_vl_interface': 7.098602101756866e-06, 'learning_rate/action_model': 7.098602101756865e-05, 'epoch': 0.11}) +[2026-06-24 10:22:40,447][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.24704758822917938, 'train/grad_norm_pre_clip': 45.146488189697266, 'timing/data': 0.00031303707510232925, 'timing/model': 1.0650074060540646, 'learning_rate/qwen_vl_interface': 7.093860983588564e-06, 'learning_rate/action_model': 7.093860983588564e-05, 'epoch': 0.11}) +[2026-06-24 10:22:41,513][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.30076608061790466, 'train/grad_norm_pre_clip': 116.89390563964844, 'timing/data': 0.0003047219943255186, 'timing/model': 1.0643499521538615, 'learning_rate/qwen_vl_interface': 7.089117701546697e-06, 'learning_rate/action_model': 7.089117701546696e-05, 'epoch': 0.11}) +[2026-06-24 10:22:42,578][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.03983115404844284, 'train/grad_norm_pre_clip': 11.643918991088867, 'timing/data': 0.00031130388379096985, 'timing/model': 1.0632471120916307, 'learning_rate/qwen_vl_interface': 7.084372261197767e-06, 'learning_rate/action_model': 7.084372261197767e-05, 'epoch': 0.11}) +[2026-06-24 10:22:43,641][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.12040228396654129, 'train/grad_norm_pre_clip': 39.947235107421875, 'timing/data': 0.0002989410422742367, 'timing/model': 1.0608689279761165, 'learning_rate/qwen_vl_interface': 7.079624668110814e-06, 'learning_rate/action_model': 7.079624668110813e-05, 'epoch': 0.11}) +[2026-06-24 10:22:44,706][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.01897323876619339, 'train/grad_norm_pre_clip': 17.688236236572266, 'timing/data': 0.0003232969902455807, 'timing/model': 1.0642086770385504, 'learning_rate/qwen_vl_interface': 7.074874927857403e-06, 'learning_rate/action_model': 7.074874927857404e-05, 'epoch': 0.11}) +[2026-06-24 10:22:45,772][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.23917558789253235, 'train/grad_norm_pre_clip': 76.5403060913086, 'timing/data': 0.00030260789208114147, 'timing/model': 1.0636335171293467, 'learning_rate/qwen_vl_interface': 7.070123046011621e-06, 'learning_rate/action_model': 7.070123046011621e-05, 'epoch': 0.11}) +[2026-06-24 10:22:46,833][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.11957372725009918, 'train/grad_norm_pre_clip': 130.54002380371094, 'timing/data': 0.0003861950244754553, 'timing/model': 1.0592899438925087, 'learning_rate/qwen_vl_interface': 7.065369028150061e-06, 'learning_rate/action_model': 7.065369028150061e-05, 'epoch': 0.11}) +[2026-06-24 10:22:47,898][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.03776870295405388, 'train/grad_norm_pre_clip': 18.716266632080078, 'timing/data': 0.000322083942592144, 'timing/model': 1.0635122088715434, 'learning_rate/qwen_vl_interface': 7.060612879851833e-06, 'learning_rate/action_model': 7.060612879851833e-05, 'epoch': 0.12}) +[2026-06-24 10:22:48,964][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.051384713500738144, 'train/grad_norm_pre_clip': 45.010719299316406, 'timing/data': 0.00036793085746467113, 'timing/model': 1.0642520692199469, 'learning_rate/qwen_vl_interface': 7.0558546066985374e-06, 'learning_rate/action_model': 7.055854606698538e-05, 'epoch': 0.12}) +[2026-06-24 10:22:50,028][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.1667151302099228, 'train/grad_norm_pre_clip': 45.54911422729492, 'timing/data': 0.00037825992330908775, 'timing/model': 1.062137468950823, 'learning_rate/qwen_vl_interface': 7.0510942142742774e-06, 'learning_rate/action_model': 7.051094214274277e-05, 'epoch': 0.12}) +[2026-06-24 10:22:51,093][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.25201964378356934, 'train/grad_norm_pre_clip': 67.1224136352539, 'timing/data': 0.00032146298326551914, 'timing/model': 1.0640657090116292, 'learning_rate/qwen_vl_interface': 7.046331708165633e-06, 'learning_rate/action_model': 7.046331708165633e-05, 'epoch': 0.12}) +[2026-06-24 10:22:52,363][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.06767501682043076, 'train/grad_norm_pre_clip': 40.369937896728516, 'timing/data': 0.00028904317878186703, 'timing/model': 1.268174845026806, 'learning_rate/qwen_vl_interface': 7.041567093961677e-06, 'learning_rate/action_model': 7.041567093961677e-05, 'epoch': 0.12}) +[2026-06-24 10:22:53,427][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.0524250827729702, 'train/grad_norm_pre_clip': 26.508594512939453, 'timing/data': 0.0003081990871578455, 'timing/model': 1.06206766795367, 'learning_rate/qwen_vl_interface': 7.036800377253944e-06, 'learning_rate/action_model': 7.036800377253944e-05, 'epoch': 0.12}) +[2026-06-24 10:22:54,493][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.04759678244590759, 'train/grad_norm_pre_clip': 17.959535598754883, 'timing/data': 0.0003574909642338753, 'timing/model': 1.0648805380333215, 'learning_rate/qwen_vl_interface': 7.0320315636364445e-06, 'learning_rate/action_model': 7.032031563636445e-05, 'epoch': 0.12}) +[2026-06-24 10:22:55,554][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.22095558047294617, 'train/grad_norm_pre_clip': 121.17565155029297, 'timing/data': 0.0002968371845781803, 'timing/model': 1.0594828468747437, 'learning_rate/qwen_vl_interface': 7.027260658705646e-06, 'learning_rate/action_model': 7.027260658705646e-05, 'epoch': 0.12}) +[2026-06-24 10:22:56,616][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.10180613398551941, 'train/grad_norm_pre_clip': 200.21780395507812, 'timing/data': 0.00030997185967862606, 'timing/model': 1.060519042192027, 'learning_rate/qwen_vl_interface': 7.022487668060471e-06, 'learning_rate/action_model': 7.022487668060471e-05, 'epoch': 0.12}) +[2026-06-24 10:22:57,681][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.032137900590896606, 'train/grad_norm_pre_clip': 26.283470153808594, 'timing/data': 0.00032084202393889427, 'timing/model': 1.0625441828742623, 'learning_rate/qwen_vl_interface': 7.017712597302292e-06, 'learning_rate/action_model': 7.017712597302292e-05, 'epoch': 0.12}) +[2026-06-24 10:22:58,746][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.06700251996517181, 'train/grad_norm_pre_clip': 49.90074920654297, 'timing/data': 0.0003044409677386284, 'timing/model': 1.064263128908351, 'learning_rate/qwen_vl_interface': 7.012935452034916e-06, 'learning_rate/action_model': 7.012935452034916e-05, 'epoch': 0.12}) +[2026-06-24 10:22:59,812][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.4435659945011139, 'train/grad_norm_pre_clip': 88.7457275390625, 'timing/data': 0.00032878690399229527, 'timing/model': 1.063619752181694, 'learning_rate/qwen_vl_interface': 7.008156237864595e-06, 'learning_rate/action_model': 7.008156237864595e-05, 'epoch': 0.12}) +[2026-06-24 10:23:00,878][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.27402418851852417, 'train/grad_norm_pre_clip': 84.26158905029297, 'timing/data': 0.00029770913533866405, 'timing/model': 1.0644799468573183, 'learning_rate/qwen_vl_interface': 7.0033749603999964e-06, 'learning_rate/action_model': 7.003374960399996e-05, 'epoch': 0.12}) +[2026-06-24 10:23:01,941][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.3256036639213562, 'train/grad_norm_pre_clip': 143.28367614746094, 'timing/data': 0.0003086188808083534, 'timing/model': 1.061508288839832, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.12}) +[2026-06-24 10:23:03,005][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.15833218395709991, 'train/grad_norm_pre_clip': 46.17659378051758, 'timing/data': 0.000325710978358984, 'timing/model': 1.063089487142861, 'learning_rate/qwen_vl_interface': 6.993806238034776e-06, 'learning_rate/action_model': 6.993806238034776e-05, 'epoch': 0.12}) +[2026-06-24 10:23:04,069][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.0662267729640007, 'train/grad_norm_pre_clip': 42.480438232421875, 'timing/data': 0.00032014097087085247, 'timing/model': 1.0623183799907565, 'learning_rate/qwen_vl_interface': 6.989018804363582e-06, 'learning_rate/action_model': 6.989018804363582e-05, 'epoch': 0.12}) +[2026-06-24 10:23:05,133][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.04472751542925835, 'train/grad_norm_pre_clip': 29.99551010131836, 'timing/data': 0.00032225484028458595, 'timing/model': 1.0623946429695934, 'learning_rate/qwen_vl_interface': 6.9842293298569584e-06, 'learning_rate/action_model': 6.984229329856958e-05, 'epoch': 0.12}) +[2026-06-24 10:23:06,197][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.2655828893184662, 'train/grad_norm_pre_clip': 62.48603057861328, 'timing/data': 0.0003287871368229389, 'timing/model': 1.0621014030184597, 'learning_rate/qwen_vl_interface': 6.9794378201356185e-06, 'learning_rate/action_model': 6.979437820135618e-05, 'epoch': 0.12}) +[2026-06-24 10:23:07,262][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.14232009649276733, 'train/grad_norm_pre_clip': 53.92837142944336, 'timing/data': 0.00038861879147589207, 'timing/model': 1.0625721369870007, 'learning_rate/qwen_vl_interface': 6.974644280822668e-06, 'learning_rate/action_model': 6.974644280822667e-05, 'epoch': 0.12}) +[2026-06-24 10:23:08,524][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.017218973487615585, 'train/grad_norm_pre_clip': 34.513885498046875, 'timing/data': 0.0003049219958484173, 'timing/model': 1.2609916829969734, 'learning_rate/qwen_vl_interface': 6.96984871754359e-06, 'learning_rate/action_model': 6.969848717543591e-05, 'epoch': 0.12}) +[2026-06-24 10:23:09,587][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.1712019145488739, 'train/grad_norm_pre_clip': 41.92211151123047, 'timing/data': 0.0003062549512833357, 'timing/model': 1.0610008968506008, 'learning_rate/qwen_vl_interface': 6.965051135926248e-06, 'learning_rate/action_model': 6.965051135926248e-05, 'epoch': 0.12}) +[2026-06-24 10:23:10,651][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.11589530110359192, 'train/grad_norm_pre_clip': 58.84962463378906, 'timing/data': 0.00030959094874560833, 'timing/model': 1.0628319741226733, 'learning_rate/qwen_vl_interface': 6.960251541600871e-06, 'learning_rate/action_model': 6.96025154160087e-05, 'epoch': 0.12}) +[2026-06-24 10:23:11,716][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.13623182475566864, 'train/grad_norm_pre_clip': 45.80366897583008, 'timing/data': 0.00029715592972934246, 'timing/model': 1.062818334903568, 'learning_rate/qwen_vl_interface': 6.955449940200048e-06, 'learning_rate/action_model': 6.955449940200047e-05, 'epoch': 0.12}) +[2026-06-24 10:23:12,776][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.09116677194833755, 'train/grad_norm_pre_clip': 20.012706756591797, 'timing/data': 0.0003043899778276682, 'timing/model': 1.0583733078092337, 'learning_rate/qwen_vl_interface': 6.950646337358726e-06, 'learning_rate/action_model': 6.950646337358726e-05, 'epoch': 0.12}) +[2026-06-24 10:23:13,834][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.1324007660150528, 'train/grad_norm_pre_clip': 34.9624137878418, 'timing/data': 0.0003252490423619747, 'timing/model': 1.0570198700297624, 'learning_rate/qwen_vl_interface': 6.945840738714203e-06, 'learning_rate/action_model': 6.945840738714202e-05, 'epoch': 0.12}) +[2026-06-24 10:23:14,893][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.09521934390068054, 'train/grad_norm_pre_clip': 14.69978141784668, 'timing/data': 0.00029438105411827564, 'timing/model': 1.0576619999483228, 'learning_rate/qwen_vl_interface': 6.9410331499061135e-06, 'learning_rate/action_model': 6.941033149906113e-05, 'epoch': 0.12}) +[2026-06-24 10:23:15,955][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.0846480131149292, 'train/grad_norm_pre_clip': 62.57449722290039, 'timing/data': 0.0003254490438848734, 'timing/model': 1.0597995519638062, 'learning_rate/qwen_vl_interface': 6.936223576576432e-06, 'learning_rate/action_model': 6.936223576576431e-05, 'epoch': 0.12}) +[2026-06-24 10:23:17,020][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.1106434091925621, 'train/grad_norm_pre_clip': 22.735586166381836, 'timing/data': 0.0003997890744358301, 'timing/model': 1.0634003540035337, 'learning_rate/qwen_vl_interface': 6.931412024369461e-06, 'learning_rate/action_model': 6.931412024369461e-05, 'epoch': 0.12}) +[2026-06-24 10:23:18,083][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.2105759084224701, 'train/grad_norm_pre_clip': 61.906105041503906, 'timing/data': 0.0003317110240459442, 'timing/model': 1.0611762860789895, 'learning_rate/qwen_vl_interface': 6.926598498931826e-06, 'learning_rate/action_model': 6.926598498931825e-05, 'epoch': 0.12}) +[2026-06-24 10:23:19,143][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.1702200025320053, 'train/grad_norm_pre_clip': 54.48820495605469, 'timing/data': 0.0003125250805169344, 'timing/model': 1.0584034889470786, 'learning_rate/qwen_vl_interface': 6.921783005912466e-06, 'learning_rate/action_model': 6.921783005912465e-05, 'epoch': 0.12}) +[2026-06-24 10:23:20,209][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.08835703879594803, 'train/grad_norm_pre_clip': 40.807472229003906, 'timing/data': 0.00032855477184057236, 'timing/model': 1.064285364933312, 'learning_rate/qwen_vl_interface': 6.916965550962631e-06, 'learning_rate/action_model': 6.91696555096263e-05, 'epoch': 0.12}) +[2026-06-24 10:23:21,271][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.06806343793869019, 'train/grad_norm_pre_clip': 21.12367057800293, 'timing/data': 0.0003056118730455637, 'timing/model': 1.0604549369309098, 'learning_rate/qwen_vl_interface': 6.912146139735871e-06, 'learning_rate/action_model': 6.91214613973587e-05, 'epoch': 0.12}) +[2026-06-24 10:23:22,334][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.09645793586969376, 'train/grad_norm_pre_clip': 28.39336395263672, 'timing/data': 0.00032535893842577934, 'timing/model': 1.0614028309937567, 'learning_rate/qwen_vl_interface': 6.907324777888035e-06, 'learning_rate/action_model': 6.907324777888035e-05, 'epoch': 0.12}) +[2026-06-24 10:23:23,396][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.18742665648460388, 'train/grad_norm_pre_clip': 38.19467544555664, 'timing/data': 0.0003398468252271414, 'timing/model': 1.060564637882635, 'learning_rate/qwen_vl_interface': 6.90250147107726e-06, 'learning_rate/action_model': 6.90250147107726e-05, 'epoch': 0.12}) +[2026-06-24 10:23:24,720][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.3273370862007141, 'train/grad_norm_pre_clip': 47.552093505859375, 'timing/data': 0.00031655305065214634, 'timing/model': 1.3229118711315095, 'learning_rate/qwen_vl_interface': 6.897676224963965e-06, 'learning_rate/action_model': 6.897676224963965e-05, 'epoch': 0.12}) +[2026-06-24 10:23:25,785][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.16709524393081665, 'train/grad_norm_pre_clip': 57.56129837036133, 'timing/data': 0.00035791099071502686, 'timing/model': 1.0628765050787479, 'learning_rate/qwen_vl_interface': 6.892849045210843e-06, 'learning_rate/action_model': 6.892849045210842e-05, 'epoch': 0.12}) +[2026-06-24 10:23:26,850][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.11330439150333405, 'train/grad_norm_pre_clip': 27.38495445251465, 'timing/data': 0.0003346570301800966, 'timing/model': 1.0633504998404533, 'learning_rate/qwen_vl_interface': 6.88801993748286e-06, 'learning_rate/action_model': 6.888019937482861e-05, 'epoch': 0.12}) +[2026-06-24 10:23:27,915][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.012859449721872807, 'train/grad_norm_pre_clip': 6.547183990478516, 'timing/data': 0.0003203700762242079, 'timing/model': 1.0639301240444183, 'learning_rate/qwen_vl_interface': 6.883188907447244e-06, 'learning_rate/action_model': 6.883188907447243e-05, 'epoch': 0.12}) +[2026-06-24 10:23:28,980][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.13715492188930511, 'train/grad_norm_pre_clip': 34.476280212402344, 'timing/data': 0.0003239070065319538, 'timing/model': 1.0631804768927395, 'learning_rate/qwen_vl_interface': 6.878355960773476e-06, 'learning_rate/action_model': 6.878355960773475e-05, 'epoch': 0.12}) +[2026-06-24 10:23:30,043][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.07097353786230087, 'train/grad_norm_pre_clip': 16.484424591064453, 'timing/data': 0.0003183758817613125, 'timing/model': 1.0616179071366787, 'learning_rate/qwen_vl_interface': 6.8735211031332894e-06, 'learning_rate/action_model': 6.87352110313329e-05, 'epoch': 0.12}) +[2026-06-24 10:23:31,107][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.08359509706497192, 'train/grad_norm_pre_clip': 56.814754486083984, 'timing/data': 0.00034169992431998253, 'timing/model': 1.0622518120799214, 'learning_rate/qwen_vl_interface': 6.86868434020066e-06, 'learning_rate/action_model': 6.86868434020066e-05, 'epoch': 0.12}) +[2026-06-24 10:23:32,170][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.06998691707849503, 'train/grad_norm_pre_clip': 25.6931209564209, 'timing/data': 0.00029754708521068096, 'timing/model': 1.0614505088888109, 'learning_rate/qwen_vl_interface': 6.863845677651793e-06, 'learning_rate/action_model': 6.863845677651794e-05, 'epoch': 0.12}) +[2026-06-24 10:23:33,238][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.168937548995018, 'train/grad_norm_pre_clip': 60.01033020019531, 'timing/data': 0.0002979480195790529, 'timing/model': 1.0664163478650153, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.12}) +[2026-06-24 10:23:34,304][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.144382044672966, 'train/grad_norm_pre_clip': 64.17528533935547, 'timing/data': 0.0005061191041022539, 'timing/model': 1.064463417045772, 'learning_rate/qwen_vl_interface': 6.854162676421344e-06, 'learning_rate/action_model': 6.854162676421344e-05, 'epoch': 0.12}) +[2026-06-24 10:23:35,369][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.12104296684265137, 'train/grad_norm_pre_clip': 41.040733337402344, 'timing/data': 0.00031809485517442226, 'timing/model': 1.0628632868174464, 'learning_rate/qwen_vl_interface': 6.8493183491033035e-06, 'learning_rate/action_model': 6.849318349103303e-05, 'epoch': 0.12}) +[2026-06-24 10:23:36,437][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.11965899914503098, 'train/grad_norm_pre_clip': 92.74297332763672, 'timing/data': 0.00034727994352579117, 'timing/model': 1.066782264970243, 'learning_rate/qwen_vl_interface': 6.844472144896097e-06, 'learning_rate/action_model': 6.844472144896097e-05, 'epoch': 0.12}) +[2026-06-24 10:23:37,502][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.1635233610868454, 'train/grad_norm_pre_clip': 37.70993423461914, 'timing/data': 0.00031237490475177765, 'timing/model': 1.0637580670882016, 'learning_rate/qwen_vl_interface': 6.839624069487016e-06, 'learning_rate/action_model': 6.839624069487015e-05, 'epoch': 0.12}) +[2026-06-24 10:23:38,562][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.12924323976039886, 'train/grad_norm_pre_clip': 31.9254207611084, 'timing/data': 0.00049233203753829, 'timing/model': 1.057732918066904, 'learning_rate/qwen_vl_interface': 6.834774128565548e-06, 'learning_rate/action_model': 6.834774128565547e-05, 'epoch': 0.12}) +[2026-06-24 10:23:39,825][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.06604903191328049, 'train/grad_norm_pre_clip': 33.279972076416016, 'timing/data': 0.0003141779452562332, 'timing/model': 1.2615210071671754, 'learning_rate/qwen_vl_interface': 6.829922327823368e-06, 'learning_rate/action_model': 6.829922327823368e-05, 'epoch': 0.12}) +[2026-06-24 10:23:40,887][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.04880910366773605, 'train/grad_norm_pre_clip': 23.01466178894043, 'timing/data': 0.00031923805363476276, 'timing/model': 1.059717962052673, 'learning_rate/qwen_vl_interface': 6.825068672954337e-06, 'learning_rate/action_model': 6.825068672954337e-05, 'epoch': 0.12}) +[2026-06-24 10:23:41,950][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.03896230086684227, 'train/grad_norm_pre_clip': 14.723174095153809, 'timing/data': 0.0003422810696065426, 'timing/model': 1.0619119598995894, 'learning_rate/qwen_vl_interface': 6.820213169654484e-06, 'learning_rate/action_model': 6.820213169654483e-05, 'epoch': 0.12}) +[2026-06-24 10:23:43,010][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.28796085715293884, 'train/grad_norm_pre_clip': 59.44646072387695, 'timing/data': 0.00031686387956142426, 'timing/model': 1.0585572321433574, 'learning_rate/qwen_vl_interface': 6.815355823622017e-06, 'learning_rate/action_model': 6.815355823622017e-05, 'epoch': 0.12}) +[2026-06-24 10:23:44,077][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.2056054174900055, 'train/grad_norm_pre_clip': 58.44924545288086, 'timing/data': 0.00031311693601310253, 'timing/model': 1.0655648279935122, 'learning_rate/qwen_vl_interface': 6.8104966405573e-06, 'learning_rate/action_model': 6.8104966405573e-05, 'epoch': 0.12}) +[2026-06-24 10:23:45,143][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.16847223043441772, 'train/grad_norm_pre_clip': 91.85881042480469, 'timing/data': 0.0003274839837104082, 'timing/model': 1.0638896969612688, 'learning_rate/qwen_vl_interface': 6.805635626162859e-06, 'learning_rate/action_model': 6.805635626162858e-05, 'epoch': 0.12}) +[2026-06-24 10:23:46,207][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.33016008138656616, 'train/grad_norm_pre_clip': 72.2508316040039, 'timing/data': 0.0003419301938265562, 'timing/model': 1.0626196539960802, 'learning_rate/qwen_vl_interface': 6.800772786143361e-06, 'learning_rate/action_model': 6.80077278614336e-05, 'epoch': 0.12}) +[2026-06-24 10:23:47,273][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.026988035067915916, 'train/grad_norm_pre_clip': 17.1772518157959, 'timing/data': 0.00031241588294506073, 'timing/model': 1.0643301461823285, 'learning_rate/qwen_vl_interface': 6.795908126205617e-06, 'learning_rate/action_model': 6.795908126205617e-05, 'epoch': 0.12}) +[2026-06-24 10:23:48,338][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.18661701679229736, 'train/grad_norm_pre_clip': 48.70821762084961, 'timing/data': 0.00031395815312862396, 'timing/model': 1.0628024109173566, 'learning_rate/qwen_vl_interface': 6.791041652058584e-06, 'learning_rate/action_model': 6.791041652058584e-05, 'epoch': 0.12}) +[2026-06-24 10:23:49,400][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.17692101001739502, 'train/grad_norm_pre_clip': 64.5954360961914, 'timing/data': 0.00030731619335711, 'timing/model': 1.0608936429489404, 'learning_rate/qwen_vl_interface': 6.786173369413333e-06, 'learning_rate/action_model': 6.786173369413333e-05, 'epoch': 0.12}) +[2026-06-24 10:23:50,465][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.13449525833129883, 'train/grad_norm_pre_clip': 86.8367919921875, 'timing/data': 0.00032873591408133507, 'timing/model': 1.0627120858989656, 'learning_rate/qwen_vl_interface': 6.781303283983074e-06, 'learning_rate/action_model': 6.781303283983073e-05, 'epoch': 0.12}) +[2026-06-24 10:23:51,528][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.13525447249412537, 'train/grad_norm_pre_clip': 23.261043548583984, 'timing/data': 0.00031580193899571896, 'timing/model': 1.0612522379960865, 'learning_rate/qwen_vl_interface': 6.776431401483112e-06, 'learning_rate/action_model': 6.776431401483112e-05, 'epoch': 0.12}) +[2026-06-24 10:23:52,591][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.10190370678901672, 'train/grad_norm_pre_clip': 34.21584701538086, 'timing/data': 0.00030335900373756886, 'timing/model': 1.0615306068211794, 'learning_rate/qwen_vl_interface': 6.771557727630882e-06, 'learning_rate/action_model': 6.771557727630881e-05, 'epoch': 0.12}) +[2026-06-24 10:23:53,654][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1250 +[2026-06-24 10:24:18,630][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1250 +[2026-06-24 10:24:18,631][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1250 +[2026-06-24 10:24:50,308][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1250 +[2026-06-24 10:24:59,192][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1250 +[2026-06-24 10:24:59,192][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.7457146644592285, 'train/grad_norm_pre_clip': 181.12818908691406, 'eval/action_loss/seconds': 24.974058381980285, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.17219701409339905, 'eval/latency_2/loss': 0.17219701409339905, 'eval/flappy/loss': 0.17219701409339905, 'eval/flappy/latency_2/loss': 0.17219701409339905, 'eval/flap_precision': 0.9848993288590604, 'eval/flap_recall': 0.7964721845318861, 'eval/flap_f1': 0.8807201800450112, 'eval/flap_rate_model': 0.16555555555555557, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8807201800450112, 'eval/latency_2/flap_precision': 0.9848993288590604, 'eval/latency_2/flap_recall': 0.7964721845318861, 'eval/latency_2/flap_f1': 0.8807201800450112, 'eval/latency_2/flap_rate_model': 0.16555555555555557, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8807201800450112, 'eval/action_classification/data_wait_seconds': 0.2921600639820099, 'eval/action_classification/predict_seconds': 31.298933029174805, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4.799999999999995, 'rl_games_eval/mid_train/mean_length': 57.2, 'rl_games_eval/mid_train/std_reward': 1.7999999999999952, 'rl_games_eval/mid_train/std_length': 14.4, 'rl_games_eval/mid_train/macro_mean_reward': 4.799999999999995, 'rl_games_eval/mid_train/macro_mean_length': 57.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 4.799999999999995, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 57.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 1.7999999999999952, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 14.4, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 4.799999999999995, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003066249191761017, 'timing/model': 1.0623907949775457, 'learning_rate/qwen_vl_interface': 6.766682268145905e-06, 'learning_rate/action_model': 6.766682268145905e-05, 'epoch': 0.12}) +[2026-06-24 10:25:00,464][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.1978682428598404, 'train/grad_norm_pre_clip': 57.53173065185547, 'timing/data': 0.0005737149622291327, 'timing/model': 1.2698047740850598, 'learning_rate/qwen_vl_interface': 6.761805028749809e-06, 'learning_rate/action_model': 6.761805028749809e-05, 'epoch': 0.12}) +[2026-06-24 10:25:01,521][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.5066512227058411, 'train/grad_norm_pre_clip': 52.49973678588867, 'timing/data': 0.000504304887726903, 'timing/model': 1.0545365901198238, 'learning_rate/qwen_vl_interface': 6.756926015166305e-06, 'learning_rate/action_model': 6.756926015166305e-05, 'epoch': 0.12}) +[2026-06-24 10:25:02,576][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.2539096474647522, 'train/grad_norm_pre_clip': 85.820068359375, 'timing/data': 0.0002986188046634197, 'timing/model': 1.0534392700064927, 'learning_rate/qwen_vl_interface': 6.752045233121185e-06, 'learning_rate/action_model': 6.752045233121185e-05, 'epoch': 0.12}) +[2026-06-24 10:25:03,638][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.036747150123119354, 'train/grad_norm_pre_clip': 20.507173538208008, 'timing/data': 0.000320380087941885, 'timing/model': 1.0612991338130087, 'learning_rate/qwen_vl_interface': 6.747162688342322e-06, 'learning_rate/action_model': 6.747162688342323e-05, 'epoch': 0.12}) +[2026-06-24 10:25:04,697][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.22578983008861542, 'train/grad_norm_pre_clip': 136.7456817626953, 'timing/data': 0.0004969618748873472, 'timing/model': 1.0565836150199175, 'learning_rate/qwen_vl_interface': 6.742278386559651e-06, 'learning_rate/action_model': 6.74227838655965e-05, 'epoch': 0.12}) +[2026-06-24 10:25:05,754][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.09965456277132034, 'train/grad_norm_pre_clip': 40.159542083740234, 'timing/data': 0.0002957349643111229, 'timing/model': 1.0558540530037135, 'learning_rate/qwen_vl_interface': 6.737392333505177e-06, 'learning_rate/action_model': 6.737392333505178e-05, 'epoch': 0.12}) +[2026-06-24 10:25:06,812][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.12547647953033447, 'train/grad_norm_pre_clip': 47.785282135009766, 'timing/data': 0.0003038691356778145, 'timing/model': 1.0555706080049276, 'learning_rate/qwen_vl_interface': 6.732504534912948e-06, 'learning_rate/action_model': 6.732504534912948e-05, 'epoch': 0.12}) +[2026-06-24 10:25:07,872][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.10872197151184082, 'train/grad_norm_pre_clip': 56.64246368408203, 'timing/data': 0.0003159318584948778, 'timing/model': 1.0588882809970528, 'learning_rate/qwen_vl_interface': 6.727614996519074e-06, 'learning_rate/action_model': 6.727614996519074e-05, 'epoch': 0.12}) +[2026-06-24 10:25:08,933][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.30334338545799255, 'train/grad_norm_pre_clip': 52.33290481567383, 'timing/data': 0.0003129469696432352, 'timing/model': 1.0591413311194628, 'learning_rate/qwen_vl_interface': 6.722723724061697e-06, 'learning_rate/action_model': 6.722723724061697e-05, 'epoch': 0.12}) +[2026-06-24 10:25:09,994][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.18184854090213776, 'train/grad_norm_pre_clip': 62.83100891113281, 'timing/data': 0.0003057930152863264, 'timing/model': 1.060111965984106, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.12}) +[2026-06-24 10:25:11,054][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.297574907541275, 'train/grad_norm_pre_clip': 66.71765899658203, 'timing/data': 0.0003094498533755541, 'timing/model': 1.0578740988858044, 'learning_rate/qwen_vl_interface': 6.712935999919194e-06, 'learning_rate/action_model': 6.712935999919193e-05, 'epoch': 0.12}) +[2026-06-24 10:25:12,113][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.2928188443183899, 'train/grad_norm_pre_clip': 51.54384231567383, 'timing/data': 0.0003122549969702959, 'timing/model': 1.0571789320092648, 'learning_rate/qwen_vl_interface': 6.708039559720503e-06, 'learning_rate/action_model': 6.708039559720502e-05, 'epoch': 0.12}) +[2026-06-24 10:25:13,172][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.3982798457145691, 'train/grad_norm_pre_clip': 44.34181594848633, 'timing/data': 0.00031200400553643703, 'timing/model': 1.057755057932809, 'learning_rate/qwen_vl_interface': 6.7031414084311776e-06, 'learning_rate/action_model': 6.703141408431178e-05, 'epoch': 0.12}) +[2026-06-24 10:25:14,233][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.08688574284315109, 'train/grad_norm_pre_clip': 14.290319442749023, 'timing/data': 0.0003332940395921469, 'timing/model': 1.0587581328582019, 'learning_rate/qwen_vl_interface': 6.698241551799469e-06, 'learning_rate/action_model': 6.698241551799469e-05, 'epoch': 0.12}) +[2026-06-24 10:25:15,293][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.13280104100704193, 'train/grad_norm_pre_clip': 31.463428497314453, 'timing/data': 0.0003082370385527611, 'timing/model': 1.0584838050417602, 'learning_rate/qwen_vl_interface': 6.693339995575634e-06, 'learning_rate/action_model': 6.693339995575634e-05, 'epoch': 0.12}) +[2026-06-24 10:25:16,566][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.09320858865976334, 'train/grad_norm_pre_clip': 116.06898498535156, 'timing/data': 0.0003169141709804535, 'timing/model': 1.2715890780091286, 'learning_rate/qwen_vl_interface': 6.688436745511918e-06, 'learning_rate/action_model': 6.688436745511918e-05, 'epoch': 0.12}) +[2026-06-24 10:25:17,626][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.22089886665344238, 'train/grad_norm_pre_clip': 37.84934616088867, 'timing/data': 0.0003212930168956518, 'timing/model': 1.058100489899516, 'learning_rate/qwen_vl_interface': 6.683531807362562e-06, 'learning_rate/action_model': 6.683531807362562e-05, 'epoch': 0.12}) +[2026-06-24 10:25:18,685][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.2365630716085434, 'train/grad_norm_pre_clip': 48.39691925048828, 'timing/data': 0.0003209109418094158, 'timing/model': 1.0584338509943336, 'learning_rate/qwen_vl_interface': 6.6786251868837836e-06, 'learning_rate/action_model': 6.678625186883783e-05, 'epoch': 0.12}) +[2026-06-24 10:25:19,741][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.12681005895137787, 'train/grad_norm_pre_clip': 17.0203914642334, 'timing/data': 0.0003383241128176451, 'timing/model': 1.0540691979695112, 'learning_rate/qwen_vl_interface': 6.6737168898337726e-06, 'learning_rate/action_model': 6.673716889833772e-05, 'epoch': 0.12}) +[2026-06-24 10:25:20,801][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.1152987852692604, 'train/grad_norm_pre_clip': 31.876924514770508, 'timing/data': 0.0003096300642937422, 'timing/model': 1.0582349898759276, 'learning_rate/qwen_vl_interface': 6.668806921972691e-06, 'learning_rate/action_model': 6.66880692197269e-05, 'epoch': 0.12}) +[2026-06-24 10:25:21,862][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.07793685793876648, 'train/grad_norm_pre_clip': 18.480093002319336, 'timing/data': 0.0005315260495990515, 'timing/model': 1.059127698885277, 'learning_rate/qwen_vl_interface': 6.6638952890626584e-06, 'learning_rate/action_model': 6.663895289062659e-05, 'epoch': 0.12}) +[2026-06-24 10:25:22,924][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.08980546146631241, 'train/grad_norm_pre_clip': 26.21658706665039, 'timing/data': 0.000528200063854456, 'timing/model': 1.0603058231063187, 'learning_rate/qwen_vl_interface': 6.65898199686775e-06, 'learning_rate/action_model': 6.65898199686775e-05, 'epoch': 0.12}) +[2026-06-24 10:25:23,987][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.1405181884765625, 'train/grad_norm_pre_clip': 25.545623779296875, 'timing/data': 0.00031465012580156326, 'timing/model': 1.0614692291710526, 'learning_rate/qwen_vl_interface': 6.654067051153989e-06, 'learning_rate/action_model': 6.65406705115399e-05, 'epoch': 0.12}) +[2026-06-24 10:25:25,049][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.0660017728805542, 'train/grad_norm_pre_clip': 12.574048042297363, 'timing/data': 0.00033592013642191887, 'timing/model': 1.0607833790127188, 'learning_rate/qwen_vl_interface': 6.649150457689337e-06, 'learning_rate/action_model': 6.649150457689337e-05, 'epoch': 0.12}) +[2026-06-24 10:25:26,112][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.17476709187030792, 'train/grad_norm_pre_clip': 42.35693359375, 'timing/data': 0.00032697292044758797, 'timing/model': 1.0605339480098337, 'learning_rate/qwen_vl_interface': 6.644232222243691e-06, 'learning_rate/action_model': 6.644232222243691e-05, 'epoch': 0.12}) +[2026-06-24 10:25:27,169][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.10441058874130249, 'train/grad_norm_pre_clip': 28.695308685302734, 'timing/data': 0.0002995319664478302, 'timing/model': 1.0562633508816361, 'learning_rate/qwen_vl_interface': 6.639312350588874e-06, 'learning_rate/action_model': 6.639312350588873e-05, 'epoch': 0.12}) +[2026-06-24 10:25:28,229][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.23040999472141266, 'train/grad_norm_pre_clip': 43.89414978027344, 'timing/data': 0.0003080780152231455, 'timing/model': 1.058108964934945, 'learning_rate/qwen_vl_interface': 6.6343908484986265e-06, 'learning_rate/action_model': 6.634390848498627e-05, 'epoch': 0.12}) +[2026-06-24 10:25:29,292][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.14021481573581696, 'train/grad_norm_pre_clip': 28.783641815185547, 'timing/data': 0.0003586420789361, 'timing/model': 1.0606673448346555, 'learning_rate/qwen_vl_interface': 6.629467721748611e-06, 'learning_rate/action_model': 6.629467721748611e-05, 'epoch': 0.12}) +[2026-06-24 10:25:30,353][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.10957515239715576, 'train/grad_norm_pre_clip': 52.92667007446289, 'timing/data': 0.0003026269841939211, 'timing/model': 1.0597070208750665, 'learning_rate/qwen_vl_interface': 6.624542976116386e-06, 'learning_rate/action_model': 6.624542976116386e-05, 'epoch': 0.12}) +[2026-06-24 10:25:31,415][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.0074676270596683025, 'train/grad_norm_pre_clip': 1.918051838874817, 'timing/data': 0.0003455879632383585, 'timing/model': 1.0604921290650964, 'learning_rate/qwen_vl_interface': 6.619616617381418e-06, 'learning_rate/action_model': 6.619616617381417e-05, 'epoch': 0.12}) +[2026-06-24 10:25:32,681][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.056167129427194595, 'train/grad_norm_pre_clip': 19.66083335876465, 'timing/data': 0.00030329800210893154, 'timing/model': 1.264235493959859, 'learning_rate/qwen_vl_interface': 6.614688651325061e-06, 'learning_rate/action_model': 6.61468865132506e-05, 'epoch': 0.12}) +[2026-06-24 10:25:33,744][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.2888912856578827, 'train/grad_norm_pre_clip': 77.08547973632812, 'timing/data': 0.0003163630608469248, 'timing/model': 1.0618140690494329, 'learning_rate/qwen_vl_interface': 6.609759083730558e-06, 'learning_rate/action_model': 6.609759083730558e-05, 'epoch': 0.12}) +[2026-06-24 10:25:34,806][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.22747394442558289, 'train/grad_norm_pre_clip': 101.96500396728516, 'timing/data': 0.0003245980478823185, 'timing/model': 1.0603304128162563, 'learning_rate/qwen_vl_interface': 6.604827920383032e-06, 'learning_rate/action_model': 6.604827920383032e-05, 'epoch': 0.12}) +[2026-06-24 10:25:35,867][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.04592418298125267, 'train/grad_norm_pre_clip': 35.2733154296875, 'timing/data': 0.0003302281256765127, 'timing/model': 1.0595266530290246, 'learning_rate/qwen_vl_interface': 6.599895167069478e-06, 'learning_rate/action_model': 6.599895167069478e-05, 'epoch': 0.12}) +[2026-06-24 10:25:36,929][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.21758916974067688, 'train/grad_norm_pre_clip': 30.67609214782715, 'timing/data': 0.0003039089497178793, 'timing/model': 1.0600116879213601, 'learning_rate/qwen_vl_interface': 6.594960829578758e-06, 'learning_rate/action_model': 6.594960829578758e-05, 'epoch': 0.12}) +[2026-06-24 10:25:37,992][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.040557049214839935, 'train/grad_norm_pre_clip': 61.42488098144531, 'timing/data': 0.00035408418625593185, 'timing/model': 1.0618507200852036, 'learning_rate/qwen_vl_interface': 6.590024913701589e-06, 'learning_rate/action_model': 6.590024913701588e-05, 'epoch': 0.12}) +[2026-06-24 10:25:39,054][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.45376840233802795, 'train/grad_norm_pre_clip': 68.88421630859375, 'timing/data': 0.00029893009923398495, 'timing/model': 1.0598975459579378, 'learning_rate/qwen_vl_interface': 6.585087425230545e-06, 'learning_rate/action_model': 6.585087425230545e-05, 'epoch': 0.12}) +[2026-06-24 10:25:40,118][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.20980200171470642, 'train/grad_norm_pre_clip': 57.19684982299805, 'timing/data': 0.0003050919622182846, 'timing/model': 1.0630226070061326, 'learning_rate/qwen_vl_interface': 6.5801483699600436e-06, 'learning_rate/action_model': 6.580148369960044e-05, 'epoch': 0.12}) +[2026-06-24 10:25:41,182][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.0379009023308754, 'train/grad_norm_pre_clip': 42.89436340332031, 'timing/data': 0.00030789710581302643, 'timing/model': 1.0624040348920971, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.12}) +[2026-06-24 10:25:42,246][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.32951176166534424, 'train/grad_norm_pre_clip': 60.506553649902344, 'timing/data': 0.00032019009813666344, 'timing/model': 1.0625599320046604, 'learning_rate/qwen_vl_interface': 6.570265582207523e-06, 'learning_rate/action_model': 6.570265582207524e-05, 'epoch': 0.12}) +[2026-06-24 10:25:43,309][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.3562627136707306, 'train/grad_norm_pre_clip': 85.46044158935547, 'timing/data': 0.0003020158037543297, 'timing/model': 1.0612823138944805, 'learning_rate/qwen_vl_interface': 6.565321861323509e-06, 'learning_rate/action_model': 6.565321861323509e-05, 'epoch': 0.12}) +[2026-06-24 10:25:44,370][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.32345688343048096, 'train/grad_norm_pre_clip': 57.88521194458008, 'timing/data': 0.000304460059851408, 'timing/model': 1.0592689770273864, 'learning_rate/qwen_vl_interface': 6.560376596836024e-06, 'learning_rate/action_model': 6.560376596836025e-05, 'epoch': 0.13}) +[2026-06-24 10:25:45,435][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.10208774358034134, 'train/grad_norm_pre_clip': 55.5968017578125, 'timing/data': 0.00032949703745543957, 'timing/model': 1.063501314027235, 'learning_rate/qwen_vl_interface': 6.555429794548619e-06, 'learning_rate/action_model': 6.555429794548618e-05, 'epoch': 0.13}) +[2026-06-24 10:25:46,500][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.2757396101951599, 'train/grad_norm_pre_clip': 54.19830322265625, 'timing/data': 0.00032661203294992447, 'timing/model': 1.0627288718242198, 'learning_rate/qwen_vl_interface': 6.550481460266635e-06, 'learning_rate/action_model': 6.550481460266636e-05, 'epoch': 0.13}) +[2026-06-24 10:25:47,563][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.027083467692136765, 'train/grad_norm_pre_clip': 29.513755798339844, 'timing/data': 0.0003010139334946871, 'timing/model': 1.0620647829491645, 'learning_rate/qwen_vl_interface': 6.545531599797222e-06, 'learning_rate/action_model': 6.545531599797222e-05, 'epoch': 0.13}) +[2026-06-24 10:25:48,928][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.14869453012943268, 'train/grad_norm_pre_clip': 54.26979446411133, 'timing/data': 0.0003104819916188717, 'timing/model': 1.3630718840286136, 'learning_rate/qwen_vl_interface': 6.540580218949315e-06, 'learning_rate/action_model': 6.540580218949315e-05, 'epoch': 0.13}) +[2026-06-24 10:25:49,991][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.05095502361655235, 'train/grad_norm_pre_clip': 37.95356369018555, 'timing/data': 0.0003385939635336399, 'timing/model': 1.0616171159781516, 'learning_rate/qwen_vl_interface': 6.5356273235336365e-06, 'learning_rate/action_model': 6.535627323533636e-05, 'epoch': 0.13}) +[2026-06-24 10:25:51,057][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.09135300666093826, 'train/grad_norm_pre_clip': 56.656742095947266, 'timing/data': 0.0003125860821455717, 'timing/model': 1.0641765489708632, 'learning_rate/qwen_vl_interface': 6.530672919362684e-06, 'learning_rate/action_model': 6.530672919362684e-05, 'epoch': 0.13}) +[2026-06-24 10:25:52,121][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.016518885269761086, 'train/grad_norm_pre_clip': 13.285226821899414, 'timing/data': 0.0003558569587767124, 'timing/model': 1.0628356919623911, 'learning_rate/qwen_vl_interface': 6.525717012250727e-06, 'learning_rate/action_model': 6.525717012250727e-05, 'epoch': 0.13}) +[2026-06-24 10:25:53,184][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.021889425814151764, 'train/grad_norm_pre_clip': 10.774574279785156, 'timing/data': 0.0002922180574387312, 'timing/model': 1.0613372488878667, 'learning_rate/qwen_vl_interface': 6.520759608013801e-06, 'learning_rate/action_model': 6.520759608013801e-05, 'epoch': 0.13}) +[2026-06-24 10:25:54,249][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.032992955297231674, 'train/grad_norm_pre_clip': 23.032344818115234, 'timing/data': 0.00048588006757199764, 'timing/model': 1.0625103630591184, 'learning_rate/qwen_vl_interface': 6.515800712469691e-06, 'learning_rate/action_model': 6.515800712469691e-05, 'epoch': 0.13}) +[2026-06-24 10:25:55,314][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.2771914005279541, 'train/grad_norm_pre_clip': 68.8310546875, 'timing/data': 0.0003553261049091816, 'timing/model': 1.0634196531027555, 'learning_rate/qwen_vl_interface': 6.510840331437941e-06, 'learning_rate/action_model': 6.510840331437941e-05, 'epoch': 0.13}) +[2026-06-24 10:25:56,378][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.10862018167972565, 'train/grad_norm_pre_clip': 42.34593963623047, 'timing/data': 0.000325819943100214, 'timing/model': 1.0628690661396831, 'learning_rate/qwen_vl_interface': 6.505878470739835e-06, 'learning_rate/action_model': 6.505878470739834e-05, 'epoch': 0.13}) +[2026-06-24 10:25:57,444][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.12857456505298615, 'train/grad_norm_pre_clip': 51.90873718261719, 'timing/data': 0.0005155960097908974, 'timing/model': 1.0632180359680206, 'learning_rate/qwen_vl_interface': 6.500915136198389e-06, 'learning_rate/action_model': 6.500915136198389e-05, 'epoch': 0.13}) +[2026-06-24 10:25:58,508][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.08231744915246964, 'train/grad_norm_pre_clip': 18.877487182617188, 'timing/data': 0.00031992001459002495, 'timing/model': 1.0631927840877324, 'learning_rate/qwen_vl_interface': 6.495950333638356e-06, 'learning_rate/action_model': 6.495950333638356e-05, 'epoch': 0.13}) +[2026-06-24 10:25:59,574][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.04615725949406624, 'train/grad_norm_pre_clip': 27.18706512451172, 'timing/data': 0.000496220076456666, 'timing/model': 1.0635625028517097, 'learning_rate/qwen_vl_interface': 6.490984068886206e-06, 'learning_rate/action_model': 6.490984068886206e-05, 'epoch': 0.13}) +[2026-06-24 10:26:00,639][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.0549539215862751, 'train/grad_norm_pre_clip': 28.71396827697754, 'timing/data': 0.00030983099713921547, 'timing/model': 1.0633885320276022, 'learning_rate/qwen_vl_interface': 6.4860163477701295e-06, 'learning_rate/action_model': 6.486016347770129e-05, 'epoch': 0.13}) +[2026-06-24 10:26:01,701][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.0919056087732315, 'train/grad_norm_pre_clip': 31.999025344848633, 'timing/data': 0.00033067003823816776, 'timing/model': 1.0609026988968253, 'learning_rate/qwen_vl_interface': 6.481047176120023e-06, 'learning_rate/action_model': 6.481047176120023e-05, 'epoch': 0.13}) +[2026-06-24 10:26:02,762][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.24247176945209503, 'train/grad_norm_pre_clip': 53.60896301269531, 'timing/data': 0.0003226648550480604, 'timing/model': 1.059403825784102, 'learning_rate/qwen_vl_interface': 6.476076559767488e-06, 'learning_rate/action_model': 6.476076559767488e-05, 'epoch': 0.13}) +[2026-06-24 10:26:04,024][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.3126288056373596, 'train/grad_norm_pre_clip': 71.69470977783203, 'timing/data': 0.0003038998693227768, 'timing/model': 1.2599109280854464, 'learning_rate/qwen_vl_interface': 6.4711045045458145e-06, 'learning_rate/action_model': 6.471104504545814e-05, 'epoch': 0.13}) +[2026-06-24 10:26:05,085][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.03299576789140701, 'train/grad_norm_pre_clip': 13.856829643249512, 'timing/data': 0.0004950480069965124, 'timing/model': 1.0597163629718125, 'learning_rate/qwen_vl_interface': 6.4661310162899916e-06, 'learning_rate/action_model': 6.466131016289992e-05, 'epoch': 0.13}) +[2026-06-24 10:26:06,151][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.08084538578987122, 'train/grad_norm_pre_clip': 16.75046157836914, 'timing/data': 0.0004207189194858074, 'timing/model': 1.0643722140230238, 'learning_rate/qwen_vl_interface': 6.461156100836681e-06, 'learning_rate/action_model': 6.46115610083668e-05, 'epoch': 0.13}) +[2026-06-24 10:26:07,215][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.045819371938705444, 'train/grad_norm_pre_clip': 29.78399658203125, 'timing/data': 0.0003071350511163473, 'timing/model': 1.061866614036262, 'learning_rate/qwen_vl_interface': 6.456179764024226e-06, 'learning_rate/action_model': 6.456179764024226e-05, 'epoch': 0.13}) +[2026-06-24 10:26:08,282][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.04022424668073654, 'train/grad_norm_pre_clip': 19.41714096069336, 'timing/data': 0.0004489920102059841, 'timing/model': 1.0652614519931376, 'learning_rate/qwen_vl_interface': 6.451202011692636e-06, 'learning_rate/action_model': 6.451202011692635e-05, 'epoch': 0.13}) +[2026-06-24 10:26:09,346][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.022793807089328766, 'train/grad_norm_pre_clip': 13.69103717803955, 'timing/data': 0.00044782902114093304, 'timing/model': 1.062797810183838, 'learning_rate/qwen_vl_interface': 6.4462228496835755e-06, 'learning_rate/action_model': 6.446222849683576e-05, 'epoch': 0.13}) +[2026-06-24 10:26:10,411][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.05320551618933678, 'train/grad_norm_pre_clip': 25.632604598999023, 'timing/data': 0.0004780960734933615, 'timing/model': 1.0623618480749428, 'learning_rate/qwen_vl_interface': 6.441242283840372e-06, 'learning_rate/action_model': 6.441242283840371e-05, 'epoch': 0.13}) +[2026-06-24 10:26:11,471][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.22148005664348602, 'train/grad_norm_pre_clip': 75.69453430175781, 'timing/data': 0.00030480092391371727, 'timing/model': 1.0585196930915117, 'learning_rate/qwen_vl_interface': 6.436260320007997e-06, 'learning_rate/action_model': 6.436260320007996e-05, 'epoch': 0.13}) +[2026-06-24 10:26:12,537][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.007893756031990051, 'train/grad_norm_pre_clip': 3.669816493988037, 'timing/data': 0.00034899404272437096, 'timing/model': 1.0644762020092458, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.13}) +[2026-06-24 10:26:13,599][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.11472196877002716, 'train/grad_norm_pre_clip': 42.81927490234375, 'timing/data': 0.000321993138641119, 'timing/model': 1.0604919190518558, 'learning_rate/qwen_vl_interface': 6.426292221763811e-06, 'learning_rate/action_model': 6.426292221763811e-05, 'epoch': 0.13}) +[2026-06-24 10:26:14,665][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.2254471629858017, 'train/grad_norm_pre_clip': 28.775968551635742, 'timing/data': 0.00031084311194717884, 'timing/model': 1.0638471839483827, 'learning_rate/qwen_vl_interface': 6.421306099050121e-06, 'learning_rate/action_model': 6.42130609905012e-05, 'epoch': 0.13}) +[2026-06-24 10:26:15,728][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.21805568039417267, 'train/grad_norm_pre_clip': 88.42410278320312, 'timing/data': 0.00032773404382169247, 'timing/model': 1.0619047980289906, 'learning_rate/qwen_vl_interface': 6.41631860174348e-06, 'learning_rate/action_model': 6.41631860174348e-05, 'epoch': 0.13}) +[2026-06-24 10:26:16,795][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.14710892736911774, 'train/grad_norm_pre_clip': 55.794185638427734, 'timing/data': 0.00030287704430520535, 'timing/model': 1.0658913149964064, 'learning_rate/qwen_vl_interface': 6.411329735696997e-06, 'learning_rate/action_model': 6.411329735696997e-05, 'epoch': 0.13}) +[2026-06-24 10:26:17,861][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.02501324936747551, 'train/grad_norm_pre_clip': 10.859238624572754, 'timing/data': 0.00030270800925791264, 'timing/model': 1.0635913501027972, 'learning_rate/qwen_vl_interface': 6.406339506765383e-06, 'learning_rate/action_model': 6.406339506765383e-05, 'epoch': 0.13}) +[2026-06-24 10:26:18,922][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.08972066640853882, 'train/grad_norm_pre_clip': 27.26002311706543, 'timing/data': 0.00030494085513055325, 'timing/model': 1.0595487479586154, 'learning_rate/qwen_vl_interface': 6.40134792080495e-06, 'learning_rate/action_model': 6.40134792080495e-05, 'epoch': 0.13}) +[2026-06-24 10:26:19,985][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.07087327539920807, 'train/grad_norm_pre_clip': 34.59877014160156, 'timing/data': 0.0003200299106538296, 'timing/model': 1.061223743017763, 'learning_rate/qwen_vl_interface': 6.396354983673603e-06, 'learning_rate/action_model': 6.396354983673602e-05, 'epoch': 0.13}) +[2026-06-24 10:26:21,046][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.1697862446308136, 'train/grad_norm_pre_clip': 35.67049026489258, 'timing/data': 0.0005924201104789972, 'timing/model': 1.059368351008743, 'learning_rate/qwen_vl_interface': 6.391360701230829e-06, 'learning_rate/action_model': 6.391360701230828e-05, 'epoch': 0.13}) +[2026-06-24 10:26:22,108][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.09199775010347366, 'train/grad_norm_pre_clip': 35.67299270629883, 'timing/data': 0.00031783594749867916, 'timing/model': 1.0610065769869834, 'learning_rate/qwen_vl_interface': 6.386365079337701e-06, 'learning_rate/action_model': 6.3863650793377e-05, 'epoch': 0.13}) +[2026-06-24 10:26:23,460][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.23791958391666412, 'train/grad_norm_pre_clip': 69.59768676757812, 'timing/data': 0.0003244990948587656, 'timing/model': 1.3499779000412673, 'learning_rate/qwen_vl_interface': 6.381368123856854e-06, 'learning_rate/action_model': 6.381368123856855e-05, 'epoch': 0.13}) +[2026-06-24 10:26:24,529][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.2884424328804016, 'train/grad_norm_pre_clip': 37.219993591308594, 'timing/data': 0.0003143490757793188, 'timing/model': 1.0672508119605482, 'learning_rate/qwen_vl_interface': 6.376369840652499e-06, 'learning_rate/action_model': 6.376369840652499e-05, 'epoch': 0.13}) +[2026-06-24 10:26:25,593][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.10662003606557846, 'train/grad_norm_pre_clip': 43.54839324951172, 'timing/data': 0.00031779520213603973, 'timing/model': 1.0624870490282774, 'learning_rate/qwen_vl_interface': 6.371370235590396e-06, 'learning_rate/action_model': 6.371370235590396e-05, 'epoch': 0.13}) +[2026-06-24 10:26:26,656][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.07527343928813934, 'train/grad_norm_pre_clip': 33.29834747314453, 'timing/data': 0.00038312794640660286, 'timing/model': 1.0616291158366948, 'learning_rate/qwen_vl_interface': 6.3663693145378625e-06, 'learning_rate/action_model': 6.366369314537862e-05, 'epoch': 0.13}) +[2026-06-24 10:26:27,718][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.06715976446866989, 'train/grad_norm_pre_clip': 45.09370422363281, 'timing/data': 0.00031195487827062607, 'timing/model': 1.0601362641900778, 'learning_rate/qwen_vl_interface': 6.361367083363756e-06, 'learning_rate/action_model': 6.361367083363756e-05, 'epoch': 0.13}) +[2026-06-24 10:26:28,782][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.15472620725631714, 'train/grad_norm_pre_clip': 67.57488250732422, 'timing/data': 0.00030588405206799507, 'timing/model': 1.0620803020428866, 'learning_rate/qwen_vl_interface': 6.356363547938473e-06, 'learning_rate/action_model': 6.356363547938473e-05, 'epoch': 0.13}) +[2026-06-24 10:26:29,844][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.16442961990833282, 'train/grad_norm_pre_clip': 39.96732711791992, 'timing/data': 0.0004192059859633446, 'timing/model': 1.060555849922821, 'learning_rate/qwen_vl_interface': 6.351358714133944e-06, 'learning_rate/action_model': 6.351358714133944e-05, 'epoch': 0.13}) +[2026-06-24 10:26:30,907][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.021352989599108696, 'train/grad_norm_pre_clip': 10.648201942443848, 'timing/data': 0.00033614994026720524, 'timing/model': 1.0618288519326597, 'learning_rate/qwen_vl_interface': 6.3463525878236176e-06, 'learning_rate/action_model': 6.346352587823618e-05, 'epoch': 0.13}) +[2026-06-24 10:26:31,973][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.1212308257818222, 'train/grad_norm_pre_clip': 42.5496711730957, 'timing/data': 0.0003097411245107651, 'timing/model': 1.0644796448759735, 'learning_rate/qwen_vl_interface': 6.341345174882464e-06, 'learning_rate/action_model': 6.341345174882464e-05, 'epoch': 0.13}) +[2026-06-24 10:26:33,037][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.14011548459529877, 'train/grad_norm_pre_clip': 34.868099212646484, 'timing/data': 0.0005184120964258909, 'timing/model': 1.061962532112375, 'learning_rate/qwen_vl_interface': 6.336336481186957e-06, 'learning_rate/action_model': 6.336336481186958e-05, 'epoch': 0.13}) +[2026-06-24 10:26:34,101][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.04643264040350914, 'train/grad_norm_pre_clip': 14.461501121520996, 'timing/data': 0.0003036488778889179, 'timing/model': 1.062353610061109, 'learning_rate/qwen_vl_interface': 6.3313265126150834e-06, 'learning_rate/action_model': 6.331326512615083e-05, 'epoch': 0.13}) +[2026-06-24 10:26:35,167][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.21960437297821045, 'train/grad_norm_pre_clip': 66.03736114501953, 'timing/data': 0.0003079769667237997, 'timing/model': 1.0645997149404138, 'learning_rate/qwen_vl_interface': 6.326315275046315e-06, 'learning_rate/action_model': 6.326315275046315e-05, 'epoch': 0.13}) +[2026-06-24 10:26:36,232][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.14875303208827972, 'train/grad_norm_pre_clip': 32.35349655151367, 'timing/data': 0.0003060630988329649, 'timing/model': 1.0629650768823922, 'learning_rate/qwen_vl_interface': 6.321302774361624e-06, 'learning_rate/action_model': 6.321302774361623e-05, 'epoch': 0.13}) +[2026-06-24 10:26:37,295][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.20700137317180634, 'train/grad_norm_pre_clip': 57.55180358886719, 'timing/data': 0.00030089402571320534, 'timing/model': 1.0613811342045665, 'learning_rate/qwen_vl_interface': 6.316289016443453e-06, 'learning_rate/action_model': 6.316289016443453e-05, 'epoch': 0.13}) +[2026-06-24 10:26:38,359][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.10620792210102081, 'train/grad_norm_pre_clip': 25.319440841674805, 'timing/data': 0.00030875811353325844, 'timing/model': 1.0624403641559184, 'learning_rate/qwen_vl_interface': 6.311274007175729e-06, 'learning_rate/action_model': 6.31127400717573e-05, 'epoch': 0.13}) +[2026-06-24 10:26:39,424][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.2158554196357727, 'train/grad_norm_pre_clip': 63.544281005859375, 'timing/data': 0.00033529894426465034, 'timing/model': 1.0636751030106097, 'learning_rate/qwen_vl_interface': 6.306257752443847e-06, 'learning_rate/action_model': 6.306257752443847e-05, 'epoch': 0.13}) +[2026-06-24 10:26:40,490][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.17968885600566864, 'train/grad_norm_pre_clip': 53.853179931640625, 'timing/data': 0.0003275240305811167, 'timing/model': 1.0643716538324952, 'learning_rate/qwen_vl_interface': 6.301240258134658e-06, 'learning_rate/action_model': 6.301240258134658e-05, 'epoch': 0.13}) +[2026-06-24 10:26:41,799][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.13655327260494232, 'train/grad_norm_pre_clip': 37.8199348449707, 'timing/data': 0.0003620490897446871, 'timing/model': 1.3071727540809661, 'learning_rate/qwen_vl_interface': 6.296221530136471e-06, 'learning_rate/action_model': 6.29622153013647e-05, 'epoch': 0.13}) +[2026-06-24 10:26:42,862][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.07257590442895889, 'train/grad_norm_pre_clip': 24.897109985351562, 'timing/data': 0.0003650540020316839, 'timing/model': 1.0615399880334735, 'learning_rate/qwen_vl_interface': 6.291201574339046e-06, 'learning_rate/action_model': 6.291201574339046e-05, 'epoch': 0.13}) +[2026-06-24 10:26:43,928][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.08271972090005875, 'train/grad_norm_pre_clip': 47.53895568847656, 'timing/data': 0.0005370161961764097, 'timing/model': 1.0639167879708111, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.13}) +[2026-06-24 10:26:44,993][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.24055340886116028, 'train/grad_norm_pre_clip': 45.07333755493164, 'timing/data': 0.00030030286870896816, 'timing/model': 1.0635329538490623, 'learning_rate/qwen_vl_interface': 6.281158002912698e-06, 'learning_rate/action_model': 6.281158002912699e-05, 'epoch': 0.13}) +[2026-06-24 10:26:46,060][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.11529737710952759, 'train/grad_norm_pre_clip': 58.27709197998047, 'timing/data': 0.00035230000503361225, 'timing/model': 1.06550314091146, 'learning_rate/qwen_vl_interface': 6.276134399070468e-06, 'learning_rate/action_model': 6.276134399070469e-05, 'epoch': 0.13}) +[2026-06-24 10:26:47,120][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.08186569809913635, 'train/grad_norm_pre_clip': 28.262277603149414, 'timing/data': 0.00034529692493379116, 'timing/model': 1.0584515798836946, 'learning_rate/qwen_vl_interface': 6.2711095910023655e-06, 'learning_rate/action_model': 6.271109591002366e-05, 'epoch': 0.13}) +[2026-06-24 10:26:48,185][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.223467618227005, 'train/grad_norm_pre_clip': 62.379669189453125, 'timing/data': 0.0003003929741680622, 'timing/model': 1.0628315028734505, 'learning_rate/qwen_vl_interface': 6.266083584605281e-06, 'learning_rate/action_model': 6.266083584605281e-05, 'epoch': 0.13}) +[2026-06-24 10:26:49,250][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.10444521903991699, 'train/grad_norm_pre_clip': 17.25751304626465, 'timing/data': 0.0002987098414450884, 'timing/model': 1.063738709082827, 'learning_rate/qwen_vl_interface': 6.261056385777515e-06, 'learning_rate/action_model': 6.261056385777515e-05, 'epoch': 0.13}) +[2026-06-24 10:26:50,316][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.2551107108592987, 'train/grad_norm_pre_clip': 44.54848861694336, 'timing/data': 0.0003605459351092577, 'timing/model': 1.0640668589621782, 'learning_rate/qwen_vl_interface': 6.256028000418764e-06, 'learning_rate/action_model': 6.256028000418763e-05, 'epoch': 0.13}) +[2026-06-24 10:26:51,379][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.02354278229176998, 'train/grad_norm_pre_clip': 7.8436384201049805, 'timing/data': 0.0004131349269300699, 'timing/model': 1.0617592490743846, 'learning_rate/qwen_vl_interface': 6.250998434430119e-06, 'learning_rate/action_model': 6.250998434430119e-05, 'epoch': 0.13}) +[2026-06-24 10:26:52,443][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.18128585815429688, 'train/grad_norm_pre_clip': 109.83236694335938, 'timing/data': 0.0002897030208259821, 'timing/model': 1.0622733680065721, 'learning_rate/qwen_vl_interface': 6.245967693714056e-06, 'learning_rate/action_model': 6.245967693714056e-05, 'epoch': 0.13}) +[2026-06-24 10:26:53,508][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.018243610858917236, 'train/grad_norm_pre_clip': 17.32920265197754, 'timing/data': 0.0003363611176609993, 'timing/model': 1.0637424259912223, 'learning_rate/qwen_vl_interface': 6.240935784174429e-06, 'learning_rate/action_model': 6.240935784174429e-05, 'epoch': 0.13}) +[2026-06-24 10:26:54,571][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.036489225924015045, 'train/grad_norm_pre_clip': 19.6095027923584, 'timing/data': 0.0005388499703258276, 'timing/model': 1.0611195960082114, 'learning_rate/qwen_vl_interface': 6.235902711716464e-06, 'learning_rate/action_model': 6.235902711716464e-05, 'epoch': 0.13}) +[2026-06-24 10:26:55,638][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.14501817524433136, 'train/grad_norm_pre_clip': 76.80162811279297, 'timing/data': 0.00030571408569812775, 'timing/model': 1.0644087949767709, 'learning_rate/qwen_vl_interface': 6.230868482246752e-06, 'learning_rate/action_model': 6.230868482246752e-05, 'epoch': 0.13}) +[2026-06-24 10:26:56,703][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.03663461655378342, 'train/grad_norm_pre_clip': 18.055667877197266, 'timing/data': 0.0003250089939683676, 'timing/model': 1.06383989402093, 'learning_rate/qwen_vl_interface': 6.225833101673242e-06, 'learning_rate/action_model': 6.225833101673242e-05, 'epoch': 0.13}) +[2026-06-24 10:26:57,768][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.2114841789007187, 'train/grad_norm_pre_clip': 60.10437774658203, 'timing/data': 0.0003428230993449688, 'timing/model': 1.06264963792637, 'learning_rate/qwen_vl_interface': 6.220796575905235e-06, 'learning_rate/action_model': 6.220796575905234e-05, 'epoch': 0.13}) +[2026-06-24 10:26:58,832][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.19975145161151886, 'train/grad_norm_pre_clip': 46.06931686401367, 'timing/data': 0.0003107530064880848, 'timing/model': 1.0631459630094469, 'learning_rate/qwen_vl_interface': 6.215758910853371e-06, 'learning_rate/action_model': 6.21575891085337e-05, 'epoch': 0.13}) +[2026-06-24 10:27:00,101][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.23216120898723602, 'train/grad_norm_pre_clip': 52.0225715637207, 'timing/data': 0.00031682406552135944, 'timing/model': 1.266915651038289, 'learning_rate/qwen_vl_interface': 6.210720112429633e-06, 'learning_rate/action_model': 6.210720112429632e-05, 'epoch': 0.13}) +[2026-06-24 10:27:01,178][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.1553158164024353, 'train/grad_norm_pre_clip': 137.0535125732422, 'timing/data': 0.00032301503233611584, 'timing/model': 1.0616153550799936, 'learning_rate/qwen_vl_interface': 6.20568018654733e-06, 'learning_rate/action_model': 6.20568018654733e-05, 'epoch': 0.13}) +[2026-06-24 10:27:02,254][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.28176185488700867, 'train/grad_norm_pre_clip': 121.2857894897461, 'timing/data': 0.0004862411879003048, 'timing/model': 1.0735601799096912, 'learning_rate/qwen_vl_interface': 6.200639139121099e-06, 'learning_rate/action_model': 6.200639139121099e-05, 'epoch': 0.13}) +[2026-06-24 10:27:03,319][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.040321558713912964, 'train/grad_norm_pre_clip': 24.821313858032227, 'timing/data': 0.0003579910844564438, 'timing/model': 1.0633468402083963, 'learning_rate/qwen_vl_interface': 6.1955969760668864e-06, 'learning_rate/action_model': 6.195596976066886e-05, 'epoch': 0.13}) +[2026-06-24 10:27:04,382][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.05619000270962715, 'train/grad_norm_pre_clip': 19.752086639404297, 'timing/data': 0.00033136201091110706, 'timing/model': 1.061850009020418, 'learning_rate/qwen_vl_interface': 6.190553703301953e-06, 'learning_rate/action_model': 6.190553703301952e-05, 'epoch': 0.13}) +[2026-06-24 10:27:05,446][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.060184210538864136, 'train/grad_norm_pre_clip': 51.5345573425293, 'timing/data': 0.0003217428456991911, 'timing/model': 1.062789635034278, 'learning_rate/qwen_vl_interface': 6.185509326744861e-06, 'learning_rate/action_model': 6.185509326744861e-05, 'epoch': 0.13}) +[2026-06-24 10:27:06,513][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.0989689826965332, 'train/grad_norm_pre_clip': 36.555442810058594, 'timing/data': 0.0003061641473323107, 'timing/model': 1.0646125471685082, 'learning_rate/qwen_vl_interface': 6.180463852315464e-06, 'learning_rate/action_model': 6.180463852315463e-05, 'epoch': 0.13}) +[2026-06-24 10:27:07,577][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.2209702432155609, 'train/grad_norm_pre_clip': 93.87327575683594, 'timing/data': 0.00039327703416347504, 'timing/model': 1.0631626551039517, 'learning_rate/qwen_vl_interface': 6.175417285934913e-06, 'learning_rate/action_model': 6.175417285934913e-05, 'epoch': 0.13}) +[2026-06-24 10:27:08,648][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.11339711397886276, 'train/grad_norm_pre_clip': 40.74688720703125, 'timing/data': 0.0003077271394431591, 'timing/model': 1.0685133889783174, 'learning_rate/qwen_vl_interface': 6.170369633525628e-06, 'learning_rate/action_model': 6.170369633525628e-05, 'epoch': 0.13}) +[2026-06-24 10:27:09,708][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.12895718216896057, 'train/grad_norm_pre_clip': 49.67860412597656, 'timing/data': 0.00033388612791895866, 'timing/model': 1.058764539891854, 'learning_rate/qwen_vl_interface': 6.165320901011316e-06, 'learning_rate/action_model': 6.165320901011316e-05, 'epoch': 0.13}) +[2026-06-24 10:27:10,771][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.19691325724124908, 'train/grad_norm_pre_clip': 49.701568603515625, 'timing/data': 0.00041219196282327175, 'timing/model': 1.0609704698435962, 'learning_rate/qwen_vl_interface': 6.160271094316941e-06, 'learning_rate/action_model': 6.16027109431694e-05, 'epoch': 0.13}) +[2026-06-24 10:27:11,837][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.004605364054441452, 'train/grad_norm_pre_clip': 2.6594409942626953, 'timing/data': 0.0003717271611094475, 'timing/model': 1.0642311258707196, 'learning_rate/qwen_vl_interface': 6.155220219368736e-06, 'learning_rate/action_model': 6.155220219368735e-05, 'epoch': 0.13}) +[2026-06-24 10:27:12,901][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.20954182744026184, 'train/grad_norm_pre_clip': 52.174442291259766, 'timing/data': 0.0003109730314463377, 'timing/model': 1.06284174695611, 'learning_rate/qwen_vl_interface': 6.150168282094184e-06, 'learning_rate/action_model': 6.150168282094184e-05, 'epoch': 0.13}) +[2026-06-24 10:27:13,967][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.23684389889240265, 'train/grad_norm_pre_clip': 89.81896209716797, 'timing/data': 0.0003134768921881914, 'timing/model': 1.0641720050480217, 'learning_rate/qwen_vl_interface': 6.145115288422013e-06, 'learning_rate/action_model': 6.145115288422013e-05, 'epoch': 0.13}) +[2026-06-24 10:27:15,032][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.04317028447985649, 'train/grad_norm_pre_clip': 20.980379104614258, 'timing/data': 0.0003267228603363037, 'timing/model': 1.0634942601900548, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.13}) +[2026-06-24 10:27:16,098][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.27388429641723633, 'train/grad_norm_pre_clip': 86.34086608886719, 'timing/data': 0.0003098610322922468, 'timing/model': 1.0648961118422449, 'learning_rate/qwen_vl_interface': 6.135006155605927e-06, 'learning_rate/action_model': 6.135006155605927e-05, 'epoch': 0.13}) +[2026-06-24 10:27:17,370][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.23605820536613464, 'train/grad_norm_pre_clip': 57.55768966674805, 'timing/data': 0.00031006080098450184, 'timing/model': 1.270443978952244, 'learning_rate/qwen_vl_interface': 6.129950028325642e-06, 'learning_rate/action_model': 6.129950028325641e-05, 'epoch': 0.13}) +[2026-06-24 10:27:18,434][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.07626572996377945, 'train/grad_norm_pre_clip': 31.487529754638672, 'timing/data': 0.0004964710678905249, 'timing/model': 1.0618905890733004, 'learning_rate/qwen_vl_interface': 6.124892868374986e-06, 'learning_rate/action_model': 6.124892868374986e-05, 'epoch': 0.13}) +[2026-06-24 10:27:19,499][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.039951082319021225, 'train/grad_norm_pre_clip': 30.52117919921875, 'timing/data': 0.0004980429075658321, 'timing/model': 1.0625726720318198, 'learning_rate/qwen_vl_interface': 6.1198346816888175e-06, 'learning_rate/action_model': 6.119834681688818e-05, 'epoch': 0.13}) +[2026-06-24 10:27:20,562][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.18522541224956512, 'train/grad_norm_pre_clip': 60.80949020385742, 'timing/data': 0.00034918496385216713, 'timing/model': 1.0619658690411597, 'learning_rate/qwen_vl_interface': 6.1147754742032e-06, 'learning_rate/action_model': 6.114775474203199e-05, 'epoch': 0.13}) +[2026-06-24 10:27:21,626][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.15948180854320526, 'train/grad_norm_pre_clip': 50.30220413208008, 'timing/data': 0.0003496450372040272, 'timing/model': 1.0617965359706432, 'learning_rate/qwen_vl_interface': 6.109715251855397e-06, 'learning_rate/action_model': 6.109715251855397e-05, 'epoch': 0.13}) +[2026-06-24 10:27:22,691][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.03352748602628708, 'train/grad_norm_pre_clip': 11.759469032287598, 'timing/data': 0.00032420805655419827, 'timing/model': 1.0640687909908593, 'learning_rate/qwen_vl_interface': 6.104654020583859e-06, 'learning_rate/action_model': 6.104654020583858e-05, 'epoch': 0.13}) +[2026-06-24 10:27:23,756][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.02526465803384781, 'train/grad_norm_pre_clip': 10.637491226196289, 'timing/data': 0.00034849299117922783, 'timing/model': 1.0630605451297015, 'learning_rate/qwen_vl_interface': 6.099591786328222e-06, 'learning_rate/action_model': 6.099591786328223e-05, 'epoch': 0.13}) +[2026-06-24 10:27:24,821][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.10696382075548172, 'train/grad_norm_pre_clip': 40.607948303222656, 'timing/data': 0.00032168300822377205, 'timing/model': 1.0638654481153935, 'learning_rate/qwen_vl_interface': 6.094528555029304e-06, 'learning_rate/action_model': 6.094528555029304e-05, 'epoch': 0.13}) +[2026-06-24 10:27:25,883][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.3421328067779541, 'train/grad_norm_pre_clip': 82.61450958251953, 'timing/data': 0.0003187870606780052, 'timing/model': 1.0600465550087392, 'learning_rate/qwen_vl_interface': 6.089464332629086e-06, 'learning_rate/action_model': 6.0894643326290856e-05, 'epoch': 0.13}) +[2026-06-24 10:27:26,944][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.23264476656913757, 'train/grad_norm_pre_clip': 84.18001556396484, 'timing/data': 0.00032307603396475315, 'timing/model': 1.0590646190103143, 'learning_rate/qwen_vl_interface': 6.084399125070714e-06, 'learning_rate/action_model': 6.084399125070715e-05, 'epoch': 0.13}) +[2026-06-24 10:27:28,011][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.15093013644218445, 'train/grad_norm_pre_clip': 69.36053466796875, 'timing/data': 0.000364342937245965, 'timing/model': 1.065679132938385, 'learning_rate/qwen_vl_interface': 6.0793329382984955e-06, 'learning_rate/action_model': 6.0793329382984957e-05, 'epoch': 0.13}) +[2026-06-24 10:27:29,074][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.13989822566509247, 'train/grad_norm_pre_clip': 46.17103958129883, 'timing/data': 0.00030779698863625526, 'timing/model': 1.0610801868606359, 'learning_rate/qwen_vl_interface': 6.07426577825788e-06, 'learning_rate/action_model': 6.0742657782578794e-05, 'epoch': 0.13}) +[2026-06-24 10:27:30,139][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.08887044340372086, 'train/grad_norm_pre_clip': 44.37898635864258, 'timing/data': 0.00032404810190200806, 'timing/model': 1.0635337210260332, 'learning_rate/qwen_vl_interface': 6.069197650895461e-06, 'learning_rate/action_model': 6.069197650895461e-05, 'epoch': 0.13}) +[2026-06-24 10:27:31,211][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.4027656018733978, 'train/grad_norm_pre_clip': 82.26016235351562, 'timing/data': 0.0003758450038731098, 'timing/model': 1.070310679031536, 'learning_rate/qwen_vl_interface': 6.06412856215897e-06, 'learning_rate/action_model': 6.064128562158969e-05, 'epoch': 0.13}) +[2026-06-24 10:27:32,276][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.11448108404874802, 'train/grad_norm_pre_clip': 42.76043701171875, 'timing/data': 0.00032599084079265594, 'timing/model': 1.0638328201603144, 'learning_rate/qwen_vl_interface': 6.059058517997263e-06, 'learning_rate/action_model': 6.059058517997263e-05, 'epoch': 0.13}) +[2026-06-24 10:27:33,341][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.02466593123972416, 'train/grad_norm_pre_clip': 8.440665245056152, 'timing/data': 0.0002964860759675503, 'timing/model': 1.0638486850075424, 'learning_rate/qwen_vl_interface': 6.053987524360322e-06, 'learning_rate/action_model': 6.053987524360322e-05, 'epoch': 0.13}) +[2026-06-24 10:27:34,610][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.17321009933948517, 'train/grad_norm_pre_clip': 28.454586029052734, 'timing/data': 0.0003597140312194824, 'timing/model': 1.2675412779208273, 'learning_rate/qwen_vl_interface': 6.048915587199239e-06, 'learning_rate/action_model': 6.0489155871992386e-05, 'epoch': 0.13}) +[2026-06-24 10:27:35,675][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.10361208766698837, 'train/grad_norm_pre_clip': 62.58515167236328, 'timing/data': 0.00031185406260192394, 'timing/model': 1.0631917531136423, 'learning_rate/qwen_vl_interface': 6.043842712466215e-06, 'learning_rate/action_model': 6.043842712466214e-05, 'epoch': 0.13}) +[2026-06-24 10:27:36,739][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.03079480119049549, 'train/grad_norm_pre_clip': 12.870912551879883, 'timing/data': 0.0003233170136809349, 'timing/model': 1.0626554721966386, 'learning_rate/qwen_vl_interface': 6.03876890611455e-06, 'learning_rate/action_model': 6.03876890611455e-05, 'epoch': 0.14}) +[2026-06-24 10:27:37,804][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.04367271810770035, 'train/grad_norm_pre_clip': 39.567970275878906, 'timing/data': 0.0005705789662897587, 'timing/model': 1.0626776590943336, 'learning_rate/qwen_vl_interface': 6.033694174098641e-06, 'learning_rate/action_model': 6.033694174098641e-05, 'epoch': 0.14}) +[2026-06-24 10:27:38,870][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.18387345969676971, 'train/grad_norm_pre_clip': 53.50700378417969, 'timing/data': 0.00029464205726981163, 'timing/model': 1.0640590030234307, 'learning_rate/qwen_vl_interface': 6.028618522373966e-06, 'learning_rate/action_model': 6.028618522373966e-05, 'epoch': 0.14}) +[2026-06-24 10:27:39,934][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.33755287528038025, 'train/grad_norm_pre_clip': 84.65943908691406, 'timing/data': 0.0003177959006279707, 'timing/model': 1.0623234841041267, 'learning_rate/qwen_vl_interface': 6.023541956897087e-06, 'learning_rate/action_model': 6.023541956897086e-05, 'epoch': 0.14}) +[2026-06-24 10:27:40,999][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.1680464893579483, 'train/grad_norm_pre_clip': 50.29856491088867, 'timing/data': 0.0003107418306171894, 'timing/model': 1.0643077280838042, 'learning_rate/qwen_vl_interface': 6.018464483625635e-06, 'learning_rate/action_model': 6.0184644836256344e-05, 'epoch': 0.14}) +[2026-06-24 10:27:42,067][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.26568320393562317, 'train/grad_norm_pre_clip': 81.27628326416016, 'timing/data': 0.00033025979064404964, 'timing/model': 1.0655909278430045, 'learning_rate/qwen_vl_interface': 6.013386108518306e-06, 'learning_rate/action_model': 6.0133861085183065e-05, 'epoch': 0.14}) +[2026-06-24 10:27:43,131][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.12849365174770355, 'train/grad_norm_pre_clip': 78.01403045654297, 'timing/data': 0.00034127896651625633, 'timing/model': 1.063117615878582, 'learning_rate/qwen_vl_interface': 6.00830683753486e-06, 'learning_rate/action_model': 6.0083068375348596e-05, 'epoch': 0.14}) +[2026-06-24 10:27:44,195][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.18173815310001373, 'train/grad_norm_pre_clip': 51.95509719848633, 'timing/data': 0.0003118750173598528, 'timing/model': 1.0619137838948518, 'learning_rate/qwen_vl_interface': 6.0032266766361024e-06, 'learning_rate/action_model': 6.0032266766361026e-05, 'epoch': 0.14}) +[2026-06-24 10:27:45,258][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.0677984356880188, 'train/grad_norm_pre_clip': 26.39706039428711, 'timing/data': 0.0004621869884431362, 'timing/model': 1.061430122004822, 'learning_rate/qwen_vl_interface': 5.998145631783885e-06, 'learning_rate/action_model': 5.998145631783885e-05, 'epoch': 0.14}) +[2026-06-24 10:27:46,322][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.1559676229953766, 'train/grad_norm_pre_clip': 30.000059127807617, 'timing/data': 0.0003243980463594198, 'timing/model': 1.062366411788389, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.14}) +[2026-06-24 10:27:47,391][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.1786872148513794, 'train/grad_norm_pre_clip': 39.75666809082031, 'timing/data': 0.0004002898931503296, 'timing/model': 1.0668223989196122, 'learning_rate/qwen_vl_interface': 5.987980914071658e-06, 'learning_rate/action_model': 5.987980914071658e-05, 'epoch': 0.14}) +[2026-06-24 10:27:48,456][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.10866022109985352, 'train/grad_norm_pre_clip': 18.557828903198242, 'timing/data': 0.00037336978130042553, 'timing/model': 1.0637870540376753, 'learning_rate/qwen_vl_interface': 5.982897253140513e-06, 'learning_rate/action_model': 5.982897253140512e-05, 'epoch': 0.14}) +[2026-06-24 10:27:49,524][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.11089873313903809, 'train/grad_norm_pre_clip': 53.834373474121094, 'timing/data': 0.00032498990185558796, 'timing/model': 1.0664018800016493, 'learning_rate/qwen_vl_interface': 5.977812732113616e-06, 'learning_rate/action_model': 5.977812732113616e-05, 'epoch': 0.14}) +[2026-06-24 10:27:50,589][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.05649600550532341, 'train/grad_norm_pre_clip': 20.732728958129883, 'timing/data': 0.00033324514515697956, 'timing/model': 1.0638076420873404, 'learning_rate/qwen_vl_interface': 5.97272735695794e-06, 'learning_rate/action_model': 5.9727273569579404e-05, 'epoch': 0.14}) +[2026-06-24 10:27:51,923][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.11708592623472214, 'train/grad_norm_pre_clip': 39.93630599975586, 'timing/data': 0.0003595540765672922, 'timing/model': 1.3324093569535762, 'learning_rate/qwen_vl_interface': 5.967641133641455e-06, 'learning_rate/action_model': 5.967641133641455e-05, 'epoch': 0.14}) +[2026-06-24 10:27:52,989][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.29010188579559326, 'train/grad_norm_pre_clip': 72.65271759033203, 'timing/data': 0.00032300595194101334, 'timing/model': 1.0645060299430043, 'learning_rate/qwen_vl_interface': 5.9625540681331275e-06, 'learning_rate/action_model': 5.962554068133127e-05, 'epoch': 0.14}) +[2026-06-24 10:27:54,055][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.2257130742073059, 'train/grad_norm_pre_clip': 26.73759651184082, 'timing/data': 0.000296484911814332, 'timing/model': 1.0639872918836772, 'learning_rate/qwen_vl_interface': 5.957466166402912e-06, 'learning_rate/action_model': 5.957466166402912e-05, 'epoch': 0.14}) +[2026-06-24 10:27:55,119][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.09361455589532852, 'train/grad_norm_pre_clip': 18.742965698242188, 'timing/data': 0.00030990084633231163, 'timing/model': 1.0626534768380225, 'learning_rate/qwen_vl_interface': 5.952377434421744e-06, 'learning_rate/action_model': 5.952377434421743e-05, 'epoch': 0.14}) +[2026-06-24 10:27:56,184][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.20562289655208588, 'train/grad_norm_pre_clip': 34.44847106933594, 'timing/data': 0.00032918713986873627, 'timing/model': 1.0636407209094614, 'learning_rate/qwen_vl_interface': 5.947287878161534e-06, 'learning_rate/action_model': 5.947287878161534e-05, 'epoch': 0.14}) +[2026-06-24 10:27:57,249][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.0794072300195694, 'train/grad_norm_pre_clip': 32.27013397216797, 'timing/data': 0.0003472808748483658, 'timing/model': 1.0625547308009118, 'learning_rate/qwen_vl_interface': 5.94219750359516e-06, 'learning_rate/action_model': 5.9421975035951595e-05, 'epoch': 0.14}) +[2026-06-24 10:27:58,312][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.08578241616487503, 'train/grad_norm_pre_clip': 32.57999801635742, 'timing/data': 0.00032538012601435184, 'timing/model': 1.061679313890636, 'learning_rate/qwen_vl_interface': 5.937106316696462e-06, 'learning_rate/action_model': 5.937106316696461e-05, 'epoch': 0.14}) +[2026-06-24 10:27:59,377][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.1474076360464096, 'train/grad_norm_pre_clip': 69.48331451416016, 'timing/data': 0.00030367891304194927, 'timing/model': 1.063178728800267, 'learning_rate/qwen_vl_interface': 5.932014323440227e-06, 'learning_rate/action_model': 5.932014323440227e-05, 'epoch': 0.14}) +[2026-06-24 10:28:00,440][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.34200990200042725, 'train/grad_norm_pre_clip': 97.36450958251953, 'timing/data': 0.00029858993366360664, 'timing/model': 1.062337334966287, 'learning_rate/qwen_vl_interface': 5.9269215298021965e-06, 'learning_rate/action_model': 5.926921529802196e-05, 'epoch': 0.14}) +[2026-06-24 10:28:01,506][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.23590777814388275, 'train/grad_norm_pre_clip': 41.194488525390625, 'timing/data': 0.00032707303762435913, 'timing/model': 1.0641052750870585, 'learning_rate/qwen_vl_interface': 5.921827941759046e-06, 'learning_rate/action_model': 5.921827941759046e-05, 'epoch': 0.14}) +[2026-06-24 10:28:02,573][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.06644419580698013, 'train/grad_norm_pre_clip': 26.17919158935547, 'timing/data': 0.00034444499760866165, 'timing/model': 1.0652114120312035, 'learning_rate/qwen_vl_interface': 5.916733565288386e-06, 'learning_rate/action_model': 5.916733565288386e-05, 'epoch': 0.14}) +[2026-06-24 10:28:03,639][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.2303909957408905, 'train/grad_norm_pre_clip': 49.326881408691406, 'timing/data': 0.00029880995862185955, 'timing/model': 1.0641797429416329, 'learning_rate/qwen_vl_interface': 5.911638406368751e-06, 'learning_rate/action_model': 5.9116384063687504e-05, 'epoch': 0.14}) +[2026-06-24 10:28:04,705][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.2212275117635727, 'train/grad_norm_pre_clip': 47.70149612426758, 'timing/data': 0.0003861640579998493, 'timing/model': 1.0640720350202173, 'learning_rate/qwen_vl_interface': 5.9065424709795904e-06, 'learning_rate/action_model': 5.90654247097959e-05, 'epoch': 0.14}) +[2026-06-24 10:28:05,769][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.31394192576408386, 'train/grad_norm_pre_clip': 56.63252258300781, 'timing/data': 0.000350666930899024, 'timing/model': 1.0629961129743606, 'learning_rate/qwen_vl_interface': 5.901445765101273e-06, 'learning_rate/action_model': 5.901445765101272e-05, 'epoch': 0.14}) +[2026-06-24 10:28:06,832][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.10403093695640564, 'train/grad_norm_pre_clip': 23.647462844848633, 'timing/data': 0.0003034491091966629, 'timing/model': 1.0612731559667736, 'learning_rate/qwen_vl_interface': 5.896348294715063e-06, 'learning_rate/action_model': 5.8963482947150625e-05, 'epoch': 0.14}) +[2026-06-24 10:28:08,118][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.18889009952545166, 'train/grad_norm_pre_clip': 45.172733306884766, 'timing/data': 0.0005284210201352835, 'timing/model': 1.2836523579899222, 'learning_rate/qwen_vl_interface': 5.891250065803129e-06, 'learning_rate/action_model': 5.891250065803129e-05, 'epoch': 0.14}) +[2026-06-24 10:28:09,183][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.12105879932641983, 'train/grad_norm_pre_clip': 52.44350814819336, 'timing/data': 0.0003548350650817156, 'timing/model': 1.064144890056923, 'learning_rate/qwen_vl_interface': 5.886151084348525e-06, 'learning_rate/action_model': 5.886151084348525e-05, 'epoch': 0.14}) +[2026-06-24 10:28:10,248][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.0628204420208931, 'train/grad_norm_pre_clip': 14.74925708770752, 'timing/data': 0.0002958939876407385, 'timing/model': 1.0625864421017468, 'learning_rate/qwen_vl_interface': 5.88105135633519e-06, 'learning_rate/action_model': 5.88105135633519e-05, 'epoch': 0.14}) +[2026-06-24 10:28:11,313][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.10728663951158524, 'train/grad_norm_pre_clip': 30.573375701904297, 'timing/data': 0.00029951101168990135, 'timing/model': 1.0636454520281404, 'learning_rate/qwen_vl_interface': 5.875950887747939e-06, 'learning_rate/action_model': 5.8759508877479385e-05, 'epoch': 0.14}) +[2026-06-24 10:28:12,377][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.04013759270310402, 'train/grad_norm_pre_clip': 15.357941627502441, 'timing/data': 0.00030158599838614464, 'timing/model': 1.0630762490909547, 'learning_rate/qwen_vl_interface': 5.870849684572455e-06, 'learning_rate/action_model': 5.870849684572455e-05, 'epoch': 0.14}) +[2026-06-24 10:28:13,444][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.04713162034749985, 'train/grad_norm_pre_clip': 30.151994705200195, 'timing/data': 0.00030770711600780487, 'timing/model': 1.0648904850240797, 'learning_rate/qwen_vl_interface': 5.865747752795285e-06, 'learning_rate/action_model': 5.8657477527952853e-05, 'epoch': 0.14}) +[2026-06-24 10:28:14,513][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.04390643909573555, 'train/grad_norm_pre_clip': 21.53707504272461, 'timing/data': 0.0003404179587960243, 'timing/model': 1.0673628440126777, 'learning_rate/qwen_vl_interface': 5.86064509840383e-06, 'learning_rate/action_model': 5.8606450984038296e-05, 'epoch': 0.14}) +[2026-06-24 10:28:15,578][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.025295203551650047, 'train/grad_norm_pre_clip': 11.375005722045898, 'timing/data': 0.0003287261351943016, 'timing/model': 1.0637061630841345, 'learning_rate/qwen_vl_interface': 5.8555417273863406e-06, 'learning_rate/action_model': 5.85554172738634e-05, 'epoch': 0.14}) +[2026-06-24 10:28:16,644][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.04203616827726364, 'train/grad_norm_pre_clip': 15.408135414123535, 'timing/data': 0.0003421418368816376, 'timing/model': 1.0639004418626428, 'learning_rate/qwen_vl_interface': 5.8504376457319025e-06, 'learning_rate/action_model': 5.850437645731902e-05, 'epoch': 0.14}) +[2026-06-24 10:28:17,709][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.36072099208831787, 'train/grad_norm_pre_clip': 70.13702392578125, 'timing/data': 0.0003396261017769575, 'timing/model': 1.063829803140834, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.14}) +[2026-06-24 10:28:18,774][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.3037293255329132, 'train/grad_norm_pre_clip': 69.88677215576172, 'timing/data': 0.0003015149850398302, 'timing/model': 1.063795984024182, 'learning_rate/qwen_vl_interface': 5.840227374472719e-06, 'learning_rate/action_model': 5.8402273744727184e-05, 'epoch': 0.14}) +[2026-06-24 10:28:19,839][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.2570684254169464, 'train/grad_norm_pre_clip': 52.3231086730957, 'timing/data': 0.000366356922313571, 'timing/model': 1.063235678942874, 'learning_rate/qwen_vl_interface': 5.835121196850291e-06, 'learning_rate/action_model': 5.835121196850291e-05, 'epoch': 0.14}) +[2026-06-24 10:28:20,903][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.14148472249507904, 'train/grad_norm_pre_clip': 58.446144104003906, 'timing/data': 0.00045606493949890137, 'timing/model': 1.0619597299955785, 'learning_rate/qwen_vl_interface': 5.830014332555546e-06, 'learning_rate/action_model': 5.830014332555546e-05, 'epoch': 0.14}) +[2026-06-24 10:28:21,968][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.07349041104316711, 'train/grad_norm_pre_clip': 51.00397491455078, 'timing/data': 0.0003042309544980526, 'timing/model': 1.0628528378438205, 'learning_rate/qwen_vl_interface': 5.824906787581678e-06, 'learning_rate/action_model': 5.8249067875816774e-05, 'epoch': 0.14}) +[2026-06-24 10:28:23,034][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.034009870141744614, 'train/grad_norm_pre_clip': 19.368101119995117, 'timing/data': 0.0003505069762468338, 'timing/model': 1.0647331378422678, 'learning_rate/qwen_vl_interface': 5.81979856792267e-06, 'learning_rate/action_model': 5.8197985679226694e-05, 'epoch': 0.14}) +[2026-06-24 10:28:24,101][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.11306499689817429, 'train/grad_norm_pre_clip': 33.52964782714844, 'timing/data': 0.00031641311943531036, 'timing/model': 1.0655547420028597, 'learning_rate/qwen_vl_interface': 5.8146896795733066e-06, 'learning_rate/action_model': 5.814689679573306e-05, 'epoch': 0.14}) +[2026-06-24 10:28:25,462][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.16820791363716125, 'train/grad_norm_pre_clip': 89.57552337646484, 'timing/data': 0.0003053429536521435, 'timing/model': 1.3591835820116103, 'learning_rate/qwen_vl_interface': 5.809580128529151e-06, 'learning_rate/action_model': 5.809580128529151e-05, 'epoch': 0.14}) +[2026-06-24 10:28:26,527][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.4480930268764496, 'train/grad_norm_pre_clip': 71.6812515258789, 'timing/data': 0.00033433688804507256, 'timing/model': 1.0634062739554793, 'learning_rate/qwen_vl_interface': 5.804469920786544e-06, 'learning_rate/action_model': 5.804469920786544e-05, 'epoch': 0.14}) +[2026-06-24 10:28:27,591][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.16015103459358215, 'train/grad_norm_pre_clip': 52.44073486328125, 'timing/data': 0.0003090791869908571, 'timing/model': 1.0625080440659076, 'learning_rate/qwen_vl_interface': 5.7993590623426035e-06, 'learning_rate/action_model': 5.799359062342603e-05, 'epoch': 0.14}) +[2026-06-24 10:28:28,658][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.18986280262470245, 'train/grad_norm_pre_clip': 77.31198120117188, 'timing/data': 0.00030261697247624397, 'timing/model': 1.065694842953235, 'learning_rate/qwen_vl_interface': 5.794247559195202e-06, 'learning_rate/action_model': 5.794247559195202e-05, 'epoch': 0.14}) +[2026-06-24 10:28:29,724][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.16916655004024506, 'train/grad_norm_pre_clip': 67.67595672607422, 'timing/data': 0.00030383910052478313, 'timing/model': 1.0638906341046095, 'learning_rate/qwen_vl_interface': 5.789135417342978e-06, 'learning_rate/action_model': 5.789135417342978e-05, 'epoch': 0.14}) +[2026-06-24 10:28:30,790][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.14798420667648315, 'train/grad_norm_pre_clip': 45.88551712036133, 'timing/data': 0.00032377708703279495, 'timing/model': 1.064756111940369, 'learning_rate/qwen_vl_interface': 5.7840226427853094e-06, 'learning_rate/action_model': 5.784022642785309e-05, 'epoch': 0.14}) +[2026-06-24 10:28:31,854][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.1265423744916916, 'train/grad_norm_pre_clip': 65.53587341308594, 'timing/data': 0.00032714405097067356, 'timing/model': 1.0628289519809186, 'learning_rate/qwen_vl_interface': 5.778909241522327e-06, 'learning_rate/action_model': 5.7789092415223266e-05, 'epoch': 0.14}) +[2026-06-24 10:28:32,919][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.020259568467736244, 'train/grad_norm_pre_clip': 17.409908294677734, 'timing/data': 0.0003236269112676382, 'timing/model': 1.063591516111046, 'learning_rate/qwen_vl_interface': 5.773795219554887e-06, 'learning_rate/action_model': 5.773795219554887e-05, 'epoch': 0.14}) +[2026-06-24 10:28:33,984][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.049387622624635696, 'train/grad_norm_pre_clip': 26.325517654418945, 'timing/data': 0.0003068349324166775, 'timing/model': 1.0632208958268166, 'learning_rate/qwen_vl_interface': 5.768680582884582e-06, 'learning_rate/action_model': 5.768680582884583e-05, 'epoch': 0.14}) +[2026-06-24 10:28:35,050][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.07320161908864975, 'train/grad_norm_pre_clip': 30.098278045654297, 'timing/data': 0.00030906894244253635, 'timing/model': 1.0645373580045998, 'learning_rate/qwen_vl_interface': 5.763565337513725e-06, 'learning_rate/action_model': 5.7635653375137244e-05, 'epoch': 0.14}) +[2026-06-24 10:28:36,116][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.06415144354104996, 'train/grad_norm_pre_clip': 34.241416931152344, 'timing/data': 0.0003357690293341875, 'timing/model': 1.0643643881194293, 'learning_rate/qwen_vl_interface': 5.758449489445336e-06, 'learning_rate/action_model': 5.7584494894453356e-05, 'epoch': 0.14}) +[2026-06-24 10:28:37,182][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.05101536586880684, 'train/grad_norm_pre_clip': 37.04384994506836, 'timing/data': 0.00034167105332016945, 'timing/model': 1.0647432010155171, 'learning_rate/qwen_vl_interface': 5.753333044683153e-06, 'learning_rate/action_model': 5.7533330446831525e-05, 'epoch': 0.14}) +[2026-06-24 10:28:38,249][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.16336631774902344, 'train/grad_norm_pre_clip': 77.3707046508789, 'timing/data': 0.0003339869435876608, 'timing/model': 1.0646197199821472, 'learning_rate/qwen_vl_interface': 5.748216009231603e-06, 'learning_rate/action_model': 5.748216009231603e-05, 'epoch': 0.14}) +[2026-06-24 10:28:39,315][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.20853382349014282, 'train/grad_norm_pre_clip': 64.17106628417969, 'timing/data': 0.0003567389212548733, 'timing/model': 1.0647386128548533, 'learning_rate/qwen_vl_interface': 5.743098389095819e-06, 'learning_rate/action_model': 5.743098389095819e-05, 'epoch': 0.14}) +[2026-06-24 10:28:40,380][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.07820159941911697, 'train/grad_norm_pre_clip': 49.140010833740234, 'timing/data': 0.00034348410554230213, 'timing/model': 1.063225258141756, 'learning_rate/qwen_vl_interface': 5.737980190281608e-06, 'learning_rate/action_model': 5.7379801902816076e-05, 'epoch': 0.14}) +[2026-06-24 10:28:41,650][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.01918015629053116, 'train/grad_norm_pre_clip': 9.263120651245117, 'timing/data': 0.00031198514625430107, 'timing/model': 1.2688395250588655, 'learning_rate/qwen_vl_interface': 5.732861418795465e-06, 'learning_rate/action_model': 5.732861418795464e-05, 'epoch': 0.14}) +[2026-06-24 10:28:42,712][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.06869260966777802, 'train/grad_norm_pre_clip': 35.296852111816406, 'timing/data': 0.00029555405490100384, 'timing/model': 1.060440996894613, 'learning_rate/qwen_vl_interface': 5.7277420806445484e-06, 'learning_rate/action_model': 5.727742080644548e-05, 'epoch': 0.14}) +[2026-06-24 10:28:43,778][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.01901204325258732, 'train/grad_norm_pre_clip': 22.4691219329834, 'timing/data': 0.0003057329449802637, 'timing/model': 1.0638306860346347, 'learning_rate/qwen_vl_interface': 5.722622181836693e-06, 'learning_rate/action_model': 5.7226221818366935e-05, 'epoch': 0.14}) +[2026-06-24 10:28:44,840][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.26446717977523804, 'train/grad_norm_pre_clip': 104.38697814941406, 'timing/data': 0.0003423818852752447, 'timing/model': 1.0609422440174967, 'learning_rate/qwen_vl_interface': 5.717501728380382e-06, 'learning_rate/action_model': 5.7175017283803824e-05, 'epoch': 0.14}) +[2026-06-24 10:28:45,904][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.003418409265577793, 'train/grad_norm_pre_clip': 1.7247196435928345, 'timing/data': 0.0003749728202819824, 'timing/model': 1.061622172826901, 'learning_rate/qwen_vl_interface': 5.712380726284753e-06, 'learning_rate/action_model': 5.7123807262847526e-05, 'epoch': 0.14}) +[2026-06-24 10:28:46,970][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.14599430561065674, 'train/grad_norm_pre_clip': 55.507694244384766, 'timing/data': 0.000304951798170805, 'timing/model': 1.064813741017133, 'learning_rate/qwen_vl_interface': 5.707259181559586e-06, 'learning_rate/action_model': 5.7072591815595857e-05, 'epoch': 0.14}) +[2026-06-24 10:28:48,036][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.16952480375766754, 'train/grad_norm_pre_clip': 98.37769317626953, 'timing/data': 0.00035427394323050976, 'timing/model': 1.064321642043069, 'learning_rate/qwen_vl_interface': 5.702137100215299e-06, 'learning_rate/action_model': 5.702137100215299e-05, 'epoch': 0.14}) +[2026-06-24 10:28:49,100][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.05906100943684578, 'train/grad_norm_pre_clip': 62.47113037109375, 'timing/data': 0.0003098610322922468, 'timing/model': 1.0628629620186985, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.14}) +[2026-06-24 10:28:50,164][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.12120560556650162, 'train/grad_norm_pre_clip': 75.73487854003906, 'timing/data': 0.000321070896461606, 'timing/model': 1.0627471979241818, 'learning_rate/qwen_vl_interface': 5.691891351714187e-06, 'learning_rate/action_model': 5.691891351714187e-05, 'epoch': 0.14}) +[2026-06-24 10:28:51,229][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.20570750534534454, 'train/grad_norm_pre_clip': 81.01219177246094, 'timing/data': 0.00029148603789508343, 'timing/model': 1.063165325904265, 'learning_rate/qwen_vl_interface': 5.686767696581315e-06, 'learning_rate/action_model': 5.6867676965813155e-05, 'epoch': 0.14}) +[2026-06-24 10:28:52,296][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.28554967045783997, 'train/grad_norm_pre_clip': 111.43036651611328, 'timing/data': 0.00031298701651394367, 'timing/model': 1.065050384029746, 'learning_rate/qwen_vl_interface': 5.681643528877223e-06, 'learning_rate/action_model': 5.681643528877223e-05, 'epoch': 0.14}) +[2026-06-24 10:28:53,359][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.6520882248878479, 'train/grad_norm_pre_clip': 130.32777404785156, 'timing/data': 0.0003184771630913019, 'timing/model': 1.0621084228623658, 'learning_rate/qwen_vl_interface': 5.676518854615409e-06, 'learning_rate/action_model': 5.676518854615409e-05, 'epoch': 0.14}) +[2026-06-24 10:28:54,425][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.15687833726406097, 'train/grad_norm_pre_clip': 31.086227416992188, 'timing/data': 0.0003291571047157049, 'timing/model': 1.0642363161314279, 'learning_rate/qwen_vl_interface': 5.671393679809961e-06, 'learning_rate/action_model': 5.671393679809961e-05, 'epoch': 0.14}) +[2026-06-24 10:28:55,489][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.2356509566307068, 'train/grad_norm_pre_clip': 26.305171966552734, 'timing/data': 0.00034674094058573246, 'timing/model': 1.062107247998938, 'learning_rate/qwen_vl_interface': 5.666268010475558e-06, 'learning_rate/action_model': 5.666268010475558e-05, 'epoch': 0.14}) +[2026-06-24 10:28:56,551][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.16122682392597198, 'train/grad_norm_pre_clip': 54.54029846191406, 'timing/data': 0.00033053895458579063, 'timing/model': 1.0611454688478261, 'learning_rate/qwen_vl_interface': 5.66114185262746e-06, 'learning_rate/action_model': 5.6611418526274595e-05, 'epoch': 0.14}) +[2026-06-24 10:28:57,791][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.015653453767299652, 'train/grad_norm_pre_clip': 10.147782325744629, 'timing/data': 0.000309159979224205, 'timing/model': 1.2385321860201657, 'learning_rate/qwen_vl_interface': 5.6560152122814956e-06, 'learning_rate/action_model': 5.6560152122814956e-05, 'epoch': 0.14}) +[2026-06-24 10:28:58,856][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.05843742936849594, 'train/grad_norm_pre_clip': 32.93536376953125, 'timing/data': 0.0003394868690520525, 'timing/model': 1.062920835800469, 'learning_rate/qwen_vl_interface': 5.650888095454065e-06, 'learning_rate/action_model': 5.650888095454064e-05, 'epoch': 0.14}) +[2026-06-24 10:28:59,921][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.22624798119068146, 'train/grad_norm_pre_clip': 67.08363342285156, 'timing/data': 0.000306073809042573, 'timing/model': 1.0630830340087414, 'learning_rate/qwen_vl_interface': 5.645760508162124e-06, 'learning_rate/action_model': 5.645760508162125e-05, 'epoch': 0.14}) +[2026-06-24 10:29:00,982][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.18807148933410645, 'train/grad_norm_pre_clip': 43.152469635009766, 'timing/data': 0.00031397887505590916, 'timing/model': 1.0598647438455373, 'learning_rate/qwen_vl_interface': 5.640632456423186e-06, 'learning_rate/action_model': 5.640632456423186e-05, 'epoch': 0.14}) +[2026-06-24 10:29:02,047][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.20006677508354187, 'train/grad_norm_pre_clip': 65.69962310791016, 'timing/data': 0.00029589515179395676, 'timing/model': 1.0630567229818553, 'learning_rate/qwen_vl_interface': 5.635503946255299e-06, 'learning_rate/action_model': 5.6355039462552995e-05, 'epoch': 0.14}) +[2026-06-24 10:29:03,114][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.20919804275035858, 'train/grad_norm_pre_clip': 82.74439239501953, 'timing/data': 0.00030764704570174217, 'timing/model': 1.0657156019005924, 'learning_rate/qwen_vl_interface': 5.6303749836770604e-06, 'learning_rate/action_model': 5.63037498367706e-05, 'epoch': 0.14}) +[2026-06-24 10:29:04,180][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.4236537516117096, 'train/grad_norm_pre_clip': 80.18569946289062, 'timing/data': 0.000310421921312809, 'timing/model': 1.0645879919175059, 'learning_rate/qwen_vl_interface': 5.6252455747075905e-06, 'learning_rate/action_model': 5.6252455747075905e-05, 'epoch': 0.14}) +[2026-06-24 10:29:05,243][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.1595688760280609, 'train/grad_norm_pre_clip': 47.85378646850586, 'timing/data': 0.00030030286870896816, 'timing/model': 1.061500595882535, 'learning_rate/qwen_vl_interface': 5.620115725366539e-06, 'learning_rate/action_model': 5.620115725366539e-05, 'epoch': 0.14}) +[2026-06-24 10:29:06,308][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.24389080703258514, 'train/grad_norm_pre_clip': 106.19776153564453, 'timing/data': 0.0003129758406430483, 'timing/model': 1.0635331310331821, 'learning_rate/qwen_vl_interface': 5.614985441674067e-06, 'learning_rate/action_model': 5.614985441674067e-05, 'epoch': 0.14}) +[2026-06-24 10:29:07,376][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.10775698721408844, 'train/grad_norm_pre_clip': 24.16933250427246, 'timing/data': 0.0003120249602943659, 'timing/model': 1.0657741071190685, 'learning_rate/qwen_vl_interface': 5.609854729650851e-06, 'learning_rate/action_model': 5.6098547296508505e-05, 'epoch': 0.14}) +[2026-06-24 10:29:08,443][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.10693345218896866, 'train/grad_norm_pre_clip': 25.084009170532227, 'timing/data': 0.00031193485483527184, 'timing/model': 1.0658429199829698, 'learning_rate/qwen_vl_interface': 5.6047235953180634e-06, 'learning_rate/action_model': 5.604723595318063e-05, 'epoch': 0.14}) +[2026-06-24 10:29:09,510][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.1754855215549469, 'train/grad_norm_pre_clip': 24.573474884033203, 'timing/data': 0.0003063639160245657, 'timing/model': 1.0650299899280071, 'learning_rate/qwen_vl_interface': 5.599592044697382e-06, 'learning_rate/action_model': 5.5995920446973816e-05, 'epoch': 0.14}) +[2026-06-24 10:29:10,578][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.09834384173154831, 'train/grad_norm_pre_clip': 22.7924747467041, 'timing/data': 0.00030272803269326687, 'timing/model': 1.067147403024137, 'learning_rate/qwen_vl_interface': 5.594460083810962e-06, 'learning_rate/action_model': 5.594460083810962e-05, 'epoch': 0.14}) +[2026-06-24 10:29:11,644][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.06090598925948143, 'train/grad_norm_pre_clip': 15.98642635345459, 'timing/data': 0.00031629297882318497, 'timing/model': 1.0644145330879837, 'learning_rate/qwen_vl_interface': 5.589327718681446e-06, 'learning_rate/action_model': 5.589327718681445e-05, 'epoch': 0.14}) +[2026-06-24 10:29:12,706][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.14255958795547485, 'train/grad_norm_pre_clip': 34.05827713012695, 'timing/data': 0.0003140391781926155, 'timing/model': 1.060045203892514, 'learning_rate/qwen_vl_interface': 5.584194955331953e-06, 'learning_rate/action_model': 5.584194955331953e-05, 'epoch': 0.14}) +[2026-06-24 10:29:13,975][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.19942313432693481, 'train/grad_norm_pre_clip': 41.00046920776367, 'timing/data': 0.00031263590790331364, 'timing/model': 1.267729806015268, 'learning_rate/qwen_vl_interface': 5.579061799786064e-06, 'learning_rate/action_model': 5.579061799786064e-05, 'epoch': 0.14}) +[2026-06-24 10:29:15,040][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.07940831035375595, 'train/grad_norm_pre_clip': 14.48664665222168, 'timing/data': 0.00035438500344753265, 'timing/model': 1.0630901020485908, 'learning_rate/qwen_vl_interface': 5.573928258067828e-06, 'learning_rate/action_model': 5.5739282580678274e-05, 'epoch': 0.14}) +[2026-06-24 10:29:16,102][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.030470967292785645, 'train/grad_norm_pre_clip': 9.887200355529785, 'timing/data': 0.00031760591082274914, 'timing/model': 1.0607732420321554, 'learning_rate/qwen_vl_interface': 5.568794336201735e-06, 'learning_rate/action_model': 5.5687943362017345e-05, 'epoch': 0.14}) +[2026-06-24 10:29:17,168][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.17363384366035461, 'train/grad_norm_pre_clip': 30.360271453857422, 'timing/data': 0.0003603859804570675, 'timing/model': 1.0644579641520977, 'learning_rate/qwen_vl_interface': 5.563660040212733e-06, 'learning_rate/action_model': 5.563660040212733e-05, 'epoch': 0.14}) +[2026-06-24 10:29:18,233][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.18856102228164673, 'train/grad_norm_pre_clip': 28.58486557006836, 'timing/data': 0.00036492408253252506, 'timing/model': 1.0637268668506294, 'learning_rate/qwen_vl_interface': 5.558525376126201e-06, 'learning_rate/action_model': 5.5585253761262014e-05, 'epoch': 0.14}) +[2026-06-24 10:29:19,297][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.26807940006256104, 'train/grad_norm_pre_clip': 53.83574676513672, 'timing/data': 0.00034489715471863747, 'timing/model': 1.0621384929399937, 'learning_rate/qwen_vl_interface': 5.553390349967958e-06, 'learning_rate/action_model': 5.553390349967957e-05, 'epoch': 0.14}) +[2026-06-24 10:29:20,361][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.16121987998485565, 'train/grad_norm_pre_clip': 19.815589904785156, 'timing/data': 0.0003116140142083168, 'timing/model': 1.0629358210135251, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.14}) +[2026-06-24 10:29:21,428][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.07677120715379715, 'train/grad_norm_pre_clip': 19.31340217590332, 'timing/data': 0.0003372619394212961, 'timing/model': 1.0647905129007995, 'learning_rate/qwen_vl_interface': 5.543119235541703e-06, 'learning_rate/action_model': 5.543119235541703e-05, 'epoch': 0.14}) +[2026-06-24 10:29:22,493][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.09581451117992401, 'train/grad_norm_pre_clip': 28.611608505249023, 'timing/data': 0.00030677509494125843, 'timing/model': 1.0635492489673197, 'learning_rate/qwen_vl_interface': 5.537983159327418e-06, 'learning_rate/action_model': 5.537983159327418e-05, 'epoch': 0.14}) +[2026-06-24 10:29:23,558][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.17925918102264404, 'train/grad_norm_pre_clip': 38.89513397216797, 'timing/data': 0.0003339860122650862, 'timing/model': 1.0631805008742958, 'learning_rate/qwen_vl_interface': 5.532846745148854e-06, 'learning_rate/action_model': 5.532846745148853e-05, 'epoch': 0.14}) +[2026-06-24 10:29:24,625][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.13888582587242126, 'train/grad_norm_pre_clip': 32.381534576416016, 'timing/data': 0.0003400370478630066, 'timing/model': 1.0656141168437898, 'learning_rate/qwen_vl_interface': 5.527709999033881e-06, 'learning_rate/action_model': 5.527709999033881e-05, 'epoch': 0.14}) +[2026-06-24 10:29:25,691][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.07717060297727585, 'train/grad_norm_pre_clip': 16.899709701538086, 'timing/data': 0.0003057930152863264, 'timing/model': 1.0641847620718181, 'learning_rate/qwen_vl_interface': 5.522572927010754e-06, 'learning_rate/action_model': 5.522572927010754e-05, 'epoch': 0.14}) +[2026-06-24 10:29:26,756][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.10128822177648544, 'train/grad_norm_pre_clip': 24.1390438079834, 'timing/data': 0.000300463056191802, 'timing/model': 1.0633233899716288, 'learning_rate/qwen_vl_interface': 5.517435535108117e-06, 'learning_rate/action_model': 5.5174355351081165e-05, 'epoch': 0.14}) +[2026-06-24 10:29:27,819][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.2626263499259949, 'train/grad_norm_pre_clip': 41.38032150268555, 'timing/data': 0.00031810696236789227, 'timing/model': 1.0615886710584164, 'learning_rate/qwen_vl_interface': 5.5122978293549816e-06, 'learning_rate/action_model': 5.512297829354981e-05, 'epoch': 0.15}) +[2026-06-24 10:29:28,884][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1500 +[2026-06-24 10:29:53,937][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1500 +[2026-06-24 10:29:53,938][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1500 +[2026-06-24 10:30:25,075][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1500 +[2026-06-24 10:33:40,352][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-24 10:33:40,352][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.1312200278043747, 'train/grad_norm_pre_clip': 20.538658142089844, 'eval/action_loss/seconds': 25.05148616596125, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.07521944493055344, 'eval/latency_2/loss': 0.07521944493055344, 'eval/flappy/loss': 0.07521944493055344, 'eval/flappy/latency_2/loss': 0.07521944493055344, 'eval/flap_precision': 0.9062901155327343, 'eval/flap_recall': 0.9579375848032564, 'eval/flap_f1': 0.9313984168865437, 'eval/flap_rate_model': 0.21638888888888888, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9313984168865438, 'eval/latency_2/flap_precision': 0.9062901155327343, 'eval/latency_2/flap_recall': 0.9579375848032564, 'eval/latency_2/flap_f1': 0.9313984168865437, 'eval/latency_2/flap_rate_model': 0.21638888888888888, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9313984168865438, 'eval/action_classification/data_wait_seconds': 0.3303922414779663, 'eval/action_classification/predict_seconds': 30.722049713134766, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 153.74000000000365, 'rl_games_eval/mid_train/mean_length': 1258.2, 'rl_games_eval/mid_train/std_reward': 160.43479173796118, 'rl_games_eval/mid_train/std_length': 1292.4083565189449, 'rl_games_eval/mid_train/macro_mean_reward': 153.74000000000365, 'rl_games_eval/mid_train/macro_mean_length': 1258.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 153.74000000000365, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1258.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 160.43479173796118, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1292.4083565189449, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 153.74000000000365, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00031590298749506474, 'timing/model': 1.0639523370191455, 'learning_rate/qwen_vl_interface': 5.507159815780734e-06, 'learning_rate/action_model': 5.507159815780734e-05, 'epoch': 0.15}) +[2026-06-24 10:33:40,353][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_1500_state +[2026-06-24 10:33:40,355][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 10:33:43,269][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_1500_state/model.safetensors +[2026-06-24 10:33:48,910][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_1500_state/optimizer.bin +[2026-06-24 10:33:48,911][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_1500_state/sampler.bin +[2026-06-24 10:33:48,911][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-24 10:33:48,912][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-24 10:33:48,957][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-24 10:33:48,957][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 10:33:48,973][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 10:33:50,023][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.10672828555107117, 'train/grad_norm_pre_clip': 22.29920196533203, 'timing/data': 0.00041728303767740726, 'timing/model': 1.04879629611969, 'learning_rate/qwen_vl_interface': 5.50202150041512e-06, 'learning_rate/action_model': 5.50202150041512e-05, 'epoch': 0.15}) +[2026-06-24 10:33:51,288][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.05392734706401825, 'train/grad_norm_pre_clip': 19.38726806640625, 'timing/data': 0.00035540712997317314, 'timing/model': 1.2630016570910811, 'learning_rate/qwen_vl_interface': 5.496882889288235e-06, 'learning_rate/action_model': 5.496882889288235e-05, 'epoch': 0.15}) +[2026-06-24 10:33:52,337][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.183829203248024, 'train/grad_norm_pre_clip': 34.551971435546875, 'timing/data': 0.000316874124109745, 'timing/model': 1.0473377560265362, 'learning_rate/qwen_vl_interface': 5.491743988430531e-06, 'learning_rate/action_model': 5.491743988430531e-05, 'epoch': 0.15}) +[2026-06-24 10:33:53,388][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.11017931252717972, 'train/grad_norm_pre_clip': 30.434446334838867, 'timing/data': 0.00030983099713921547, 'timing/model': 1.0495218650903553, 'learning_rate/qwen_vl_interface': 5.486604803872789e-06, 'learning_rate/action_model': 5.486604803872789e-05, 'epoch': 0.15}) +[2026-06-24 10:33:54,437][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.10996586084365845, 'train/grad_norm_pre_clip': 26.12530517578125, 'timing/data': 0.0003185172099620104, 'timing/model': 1.0466591368895024, 'learning_rate/qwen_vl_interface': 5.481465341646133e-06, 'learning_rate/action_model': 5.4814653416461326e-05, 'epoch': 0.15}) +[2026-06-24 10:33:55,489][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.13019654154777527, 'train/grad_norm_pre_clip': 48.021915435791016, 'timing/data': 0.0003135381266474724, 'timing/model': 1.050357926869765, 'learning_rate/qwen_vl_interface': 5.476325607782006e-06, 'learning_rate/action_model': 5.476325607782006e-05, 'epoch': 0.15}) +[2026-06-24 10:33:56,542][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.03707008436322212, 'train/grad_norm_pre_clip': 14.639704704284668, 'timing/data': 0.0004921630024909973, 'timing/model': 1.0513453311286867, 'learning_rate/qwen_vl_interface': 5.47118560831217e-06, 'learning_rate/action_model': 5.47118560831217e-05, 'epoch': 0.15}) +[2026-06-24 10:33:57,597][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.2755543291568756, 'train/grad_norm_pre_clip': 63.8046989440918, 'timing/data': 0.0004990361630916595, 'timing/model': 1.0532212320249528, 'learning_rate/qwen_vl_interface': 5.466045349268704e-06, 'learning_rate/action_model': 5.4660453492687035e-05, 'epoch': 0.15}) +[2026-06-24 10:33:58,651][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.23432451486587524, 'train/grad_norm_pre_clip': 76.70867156982422, 'timing/data': 0.0003044409677386284, 'timing/model': 1.0524457090068609, 'learning_rate/qwen_vl_interface': 5.460904836683987e-06, 'learning_rate/action_model': 5.460904836683986e-05, 'epoch': 0.15}) +[2026-06-24 10:33:59,704][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.17684832215309143, 'train/grad_norm_pre_clip': 48.213436126708984, 'timing/data': 0.00033974694088101387, 'timing/model': 1.051521087065339, 'learning_rate/qwen_vl_interface': 5.455764076590696e-06, 'learning_rate/action_model': 5.455764076590696e-05, 'epoch': 0.15}) +[2026-06-24 10:34:00,758][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.1511293351650238, 'train/grad_norm_pre_clip': 42.1198844909668, 'timing/data': 0.00031488994136452675, 'timing/model': 1.0525859862100333, 'learning_rate/qwen_vl_interface': 5.450623075021799e-06, 'learning_rate/action_model': 5.450623075021799e-05, 'epoch': 0.15}) +[2026-06-24 10:34:01,811][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.021294767037034035, 'train/grad_norm_pre_clip': 22.925884246826172, 'timing/data': 0.00035436381585896015, 'timing/model': 1.0511186770163476, 'learning_rate/qwen_vl_interface': 5.445481838010551e-06, 'learning_rate/action_model': 5.445481838010551e-05, 'epoch': 0.15}) +[2026-06-24 10:34:02,865][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.045808400958776474, 'train/grad_norm_pre_clip': 31.19710350036621, 'timing/data': 0.0003233260940760374, 'timing/model': 1.0524599340278655, 'learning_rate/qwen_vl_interface': 5.440340371590477e-06, 'learning_rate/action_model': 5.440340371590477e-05, 'epoch': 0.15}) +[2026-06-24 10:34:03,920][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.003934301435947418, 'train/grad_norm_pre_clip': 2.234189748764038, 'timing/data': 0.0003290369641035795, 'timing/model': 1.0532940700650215, 'learning_rate/qwen_vl_interface': 5.435198681795375e-06, 'learning_rate/action_model': 5.435198681795375e-05, 'epoch': 0.15}) +[2026-06-24 10:34:04,975][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.31365153193473816, 'train/grad_norm_pre_clip': 65.81763458251953, 'timing/data': 0.0003032791428267956, 'timing/model': 1.0536417360417545, 'learning_rate/qwen_vl_interface': 5.430056774659305e-06, 'learning_rate/action_model': 5.4300567746593044e-05, 'epoch': 0.15}) +[2026-06-24 10:34:06,036][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.20949365198612213, 'train/grad_norm_pre_clip': 123.69852447509766, 'timing/data': 0.00032694311812520027, 'timing/model': 1.0595722910948098, 'learning_rate/qwen_vl_interface': 5.42491465621658e-06, 'learning_rate/action_model': 5.42491465621658e-05, 'epoch': 0.15}) +[2026-06-24 10:34:07,090][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.1789909154176712, 'train/grad_norm_pre_clip': 54.88502502441406, 'timing/data': 0.0003380139824002981, 'timing/model': 1.0523725089151412, 'learning_rate/qwen_vl_interface': 5.419772332501762e-06, 'learning_rate/action_model': 5.419772332501762e-05, 'epoch': 0.15}) +[2026-06-24 10:34:08,145][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.25981369614601135, 'train/grad_norm_pre_clip': 69.1108169555664, 'timing/data': 0.00038169603794813156, 'timing/model': 1.0528666470199823, 'learning_rate/qwen_vl_interface': 5.414629809549658e-06, 'learning_rate/action_model': 5.414629809549658e-05, 'epoch': 0.15}) +[2026-06-24 10:34:09,410][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.18098387122154236, 'train/grad_norm_pre_clip': 63.41416549682617, 'timing/data': 0.00032403809018433094, 'timing/model': 1.2635681619867682, 'learning_rate/qwen_vl_interface': 5.409487093395301e-06, 'learning_rate/action_model': 5.409487093395301e-05, 'epoch': 0.15}) +[2026-06-24 10:34:10,468][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.12197327613830566, 'train/grad_norm_pre_clip': 34.86338424682617, 'timing/data': 0.00031680380925536156, 'timing/model': 1.0562778308521956, 'learning_rate/qwen_vl_interface': 5.404344190073958e-06, 'learning_rate/action_model': 5.4043441900739577e-05, 'epoch': 0.15}) +[2026-06-24 10:34:11,520][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.12359913438558578, 'train/grad_norm_pre_clip': 39.87087631225586, 'timing/data': 0.00044914218597114086, 'timing/model': 1.0509924688376486, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 0.15}) +[2026-06-24 10:34:12,577][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.05195518210530281, 'train/grad_norm_pre_clip': 51.006797790527344, 'timing/data': 0.00030751596204936504, 'timing/model': 1.0549744090531021, 'learning_rate/qwen_vl_interface': 5.394057846072454e-06, 'learning_rate/action_model': 5.3940578460724533e-05, 'epoch': 0.15}) +[2026-06-24 10:34:13,634][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.13085338473320007, 'train/grad_norm_pre_clip': 22.216415405273438, 'timing/data': 0.00031636282801628113, 'timing/model': 1.055270376848057, 'learning_rate/qwen_vl_interface': 5.388914417463892e-06, 'learning_rate/action_model': 5.3889144174638914e-05, 'epoch': 0.15}) +[2026-06-24 10:34:14,690][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.140622079372406, 'train/grad_norm_pre_clip': 41.91924285888672, 'timing/data': 0.0003993490245193243, 'timing/model': 1.055007945979014, 'learning_rate/qwen_vl_interface': 5.3837708258315225e-06, 'learning_rate/action_model': 5.383770825831522e-05, 'epoch': 0.15}) +[2026-06-24 10:34:15,747][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.1660146862268448, 'train/grad_norm_pre_clip': 36.660316467285156, 'timing/data': 0.00030750688165426254, 'timing/model': 1.0551240569911897, 'learning_rate/qwen_vl_interface': 5.37862707721164e-06, 'learning_rate/action_model': 5.37862707721164e-05, 'epoch': 0.15}) +[2026-06-24 10:34:16,806][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.1859242022037506, 'train/grad_norm_pre_clip': 58.81001663208008, 'timing/data': 0.00031926902011036873, 'timing/model': 1.0572752540465444, 'learning_rate/qwen_vl_interface': 5.373483177640716e-06, 'learning_rate/action_model': 5.3734831776407154e-05, 'epoch': 0.15}) +[2026-06-24 10:34:17,863][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.0274476557970047, 'train/grad_norm_pre_clip': 13.158690452575684, 'timing/data': 0.00033333501778542995, 'timing/model': 1.0557020250707865, 'learning_rate/qwen_vl_interface': 5.3683391331554054e-06, 'learning_rate/action_model': 5.3683391331554056e-05, 'epoch': 0.15}) +[2026-06-24 10:34:18,921][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.16702593863010406, 'train/grad_norm_pre_clip': 52.81706619262695, 'timing/data': 0.000335038173943758, 'timing/model': 1.0564636948984116, 'learning_rate/qwen_vl_interface': 5.3631949497925306e-06, 'learning_rate/action_model': 5.363194949792531e-05, 'epoch': 0.15}) +[2026-06-24 10:34:19,977][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.07169436663389206, 'train/grad_norm_pre_clip': 45.7427864074707, 'timing/data': 0.0003214229363948107, 'timing/model': 1.0540056771133095, 'learning_rate/qwen_vl_interface': 5.358050633589078e-06, 'learning_rate/action_model': 5.3580506335890776e-05, 'epoch': 0.15}) +[2026-06-24 10:34:21,035][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.008202211931347847, 'train/grad_norm_pre_clip': 7.348402500152588, 'timing/data': 0.0003411101642996073, 'timing/model': 1.0564461960457265, 'learning_rate/qwen_vl_interface': 5.3529061905821884e-06, 'learning_rate/action_model': 5.352906190582189e-05, 'epoch': 0.15}) +[2026-06-24 10:34:22,093][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.12526240944862366, 'train/grad_norm_pre_clip': 32.284908294677734, 'timing/data': 0.0003228760324418545, 'timing/model': 1.056477967882529, 'learning_rate/qwen_vl_interface': 5.3477616268091505e-06, 'learning_rate/action_model': 5.34776162680915e-05, 'epoch': 0.15}) +[2026-06-24 10:34:23,151][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.19233061373233795, 'train/grad_norm_pre_clip': 36.893104553222656, 'timing/data': 0.00032979901880025864, 'timing/model': 1.0563854470383376, 'learning_rate/qwen_vl_interface': 5.3426169483074005e-06, 'learning_rate/action_model': 5.3426169483074005e-05, 'epoch': 0.15}) +[2026-06-24 10:34:24,211][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.19924117624759674, 'train/grad_norm_pre_clip': 58.27040481567383, 'timing/data': 0.00031367805786430836, 'timing/model': 1.0585806970484555, 'learning_rate/qwen_vl_interface': 5.3374721611145005e-06, 'learning_rate/action_model': 5.3374721611145005e-05, 'epoch': 0.15}) +[2026-06-24 10:34:25,272][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.09649535268545151, 'train/grad_norm_pre_clip': 54.752872467041016, 'timing/data': 0.00029931217432022095, 'timing/model': 1.059057954000309, 'learning_rate/qwen_vl_interface': 5.33232727126815e-06, 'learning_rate/action_model': 5.33232727126815e-05, 'epoch': 0.15}) +[2026-06-24 10:34:26,541][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.31206265091896057, 'train/grad_norm_pre_clip': 76.5231704711914, 'timing/data': 0.0003115038853138685, 'timing/model': 1.2674575769342482, 'learning_rate/qwen_vl_interface': 5.32718228480616e-06, 'learning_rate/action_model': 5.3271822848061605e-05, 'epoch': 0.15}) +[2026-06-24 10:34:27,599][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.09717662632465363, 'train/grad_norm_pre_clip': 24.7058048248291, 'timing/data': 0.0003062239848077297, 'timing/model': 1.056371449958533, 'learning_rate/qwen_vl_interface': 5.32203720776646e-06, 'learning_rate/action_model': 5.322037207766459e-05, 'epoch': 0.15}) +[2026-06-24 10:34:28,654][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.16129615902900696, 'train/grad_norm_pre_clip': 44.84486389160156, 'timing/data': 0.00031096418388187885, 'timing/model': 1.05351050500758, 'learning_rate/qwen_vl_interface': 5.3168920461870825e-06, 'learning_rate/action_model': 5.3168920461870816e-05, 'epoch': 0.15}) +[2026-06-24 10:34:29,713][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.2275257408618927, 'train/grad_norm_pre_clip': 51.484130859375, 'timing/data': 0.00031035207211971283, 'timing/model': 1.0576735350769013, 'learning_rate/qwen_vl_interface': 5.311746806106165e-06, 'learning_rate/action_model': 5.311746806106165e-05, 'epoch': 0.15}) +[2026-06-24 10:34:30,774][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.15418770909309387, 'train/grad_norm_pre_clip': 42.93293380737305, 'timing/data': 0.0003189889248460531, 'timing/model': 1.059558334061876, 'learning_rate/qwen_vl_interface': 5.306601493561931e-06, 'learning_rate/action_model': 5.3066014935619314e-05, 'epoch': 0.15}) +[2026-06-24 10:34:31,835][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.09484008699655533, 'train/grad_norm_pre_clip': 33.04161834716797, 'timing/data': 0.0002987401094287634, 'timing/model': 1.058633343083784, 'learning_rate/qwen_vl_interface': 5.30145611459269e-06, 'learning_rate/action_model': 5.30145611459269e-05, 'epoch': 0.15}) +[2026-06-24 10:34:32,894][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.1492442935705185, 'train/grad_norm_pre_clip': 32.296714782714844, 'timing/data': 0.00030566309578716755, 'timing/model': 1.0577581429388374, 'learning_rate/qwen_vl_interface': 5.296310675236833e-06, 'learning_rate/action_model': 5.296310675236833e-05, 'epoch': 0.15}) +[2026-06-24 10:34:33,953][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.20718783140182495, 'train/grad_norm_pre_clip': 46.78689193725586, 'timing/data': 0.000352049944922328, 'timing/model': 1.0580205649603158, 'learning_rate/qwen_vl_interface': 5.29116518153282e-06, 'learning_rate/action_model': 5.29116518153282e-05, 'epoch': 0.15}) +[2026-06-24 10:34:35,014][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.04928276687860489, 'train/grad_norm_pre_clip': 32.29134750366211, 'timing/data': 0.0003122449852526188, 'timing/model': 1.0586136288475245, 'learning_rate/qwen_vl_interface': 5.286019639519174e-06, 'learning_rate/action_model': 5.286019639519174e-05, 'epoch': 0.15}) +[2026-06-24 10:34:36,075][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.17390602827072144, 'train/grad_norm_pre_clip': 43.583580017089844, 'timing/data': 0.0003248387947678566, 'timing/model': 1.0596406778786331, 'learning_rate/qwen_vl_interface': 5.280874055234474e-06, 'learning_rate/action_model': 5.2808740552344734e-05, 'epoch': 0.15}) +[2026-06-24 10:34:37,134][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.09410622715950012, 'train/grad_norm_pre_clip': 15.235011100769043, 'timing/data': 0.00034149992279708385, 'timing/model': 1.0575124470051378, 'learning_rate/qwen_vl_interface': 5.27572843471735e-06, 'learning_rate/action_model': 5.2757284347173496e-05, 'epoch': 0.15}) +[2026-06-24 10:34:38,194][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.10039258748292923, 'train/grad_norm_pre_clip': 23.69473648071289, 'timing/data': 0.0003212329465895891, 'timing/model': 1.0582796190865338, 'learning_rate/qwen_vl_interface': 5.270582784006475e-06, 'learning_rate/action_model': 5.2705827840064746e-05, 'epoch': 0.15}) +[2026-06-24 10:34:39,256][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.12138406187295914, 'train/grad_norm_pre_clip': 45.32392120361328, 'timing/data': 0.0002924390137195587, 'timing/model': 1.0611282240133733, 'learning_rate/qwen_vl_interface': 5.265437109140557e-06, 'learning_rate/action_model': 5.2654371091405564e-05, 'epoch': 0.15}) +[2026-06-24 10:34:40,317][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.2183750420808792, 'train/grad_norm_pre_clip': 51.429019927978516, 'timing/data': 0.0003047108184546232, 'timing/model': 1.0592939048074186, 'learning_rate/qwen_vl_interface': 5.26029141615833e-06, 'learning_rate/action_model': 5.260291416158331e-05, 'epoch': 0.15}) +[2026-06-24 10:34:41,376][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.18947935104370117, 'train/grad_norm_pre_clip': 42.875579833984375, 'timing/data': 0.00033444701693952084, 'timing/model': 1.057470696978271, 'learning_rate/qwen_vl_interface': 5.255145711098556e-06, 'learning_rate/action_model': 5.255145711098556e-05, 'epoch': 0.15}) +[2026-06-24 10:34:42,435][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.043223049491643906, 'train/grad_norm_pre_clip': 14.559318542480469, 'timing/data': 0.0003004029858857393, 'timing/model': 1.057067773072049, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.15}) +[2026-06-24 10:34:43,704][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.12285212427377701, 'train/grad_norm_pre_clip': 30.064956665039062, 'timing/data': 0.00030874903313815594, 'timing/model': 1.2674615441355854, 'learning_rate/qwen_vl_interface': 5.244854288901445e-06, 'learning_rate/action_model': 5.2448542889014454e-05, 'epoch': 0.15}) +[2026-06-24 10:34:44,764][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.11102738976478577, 'train/grad_norm_pre_clip': 21.320903778076172, 'timing/data': 0.00032776501029729843, 'timing/model': 1.0590021458920091, 'learning_rate/qwen_vl_interface': 5.239708583841671e-06, 'learning_rate/action_model': 5.2397085838416704e-05, 'epoch': 0.15}) +[2026-06-24 10:34:45,829][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.16152916848659515, 'train/grad_norm_pre_clip': 64.14273834228516, 'timing/data': 0.0003389250487089157, 'timing/model': 1.0630314250010997, 'learning_rate/qwen_vl_interface': 5.234562890859443e-06, 'learning_rate/action_model': 5.2345628908594433e-05, 'epoch': 0.15}) +[2026-06-24 10:34:46,891][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.059749484062194824, 'train/grad_norm_pre_clip': 20.530481338500977, 'timing/data': 0.0003036998677998781, 'timing/model': 1.0607622819952667, 'learning_rate/qwen_vl_interface': 5.229417215993525e-06, 'learning_rate/action_model': 5.229417215993525e-05, 'epoch': 0.15}) +[2026-06-24 10:34:47,954][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.09814335405826569, 'train/grad_norm_pre_clip': 19.40464210510254, 'timing/data': 0.0003060440067201853, 'timing/model': 1.0607527850661427, 'learning_rate/qwen_vl_interface': 5.2242715652826525e-06, 'learning_rate/action_model': 5.224271565282652e-05, 'epoch': 0.15}) +[2026-06-24 10:34:49,015][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.0312882736325264, 'train/grad_norm_pre_clip': 17.179746627807617, 'timing/data': 0.0003318721428513527, 'timing/model': 1.0596097439993173, 'learning_rate/qwen_vl_interface': 5.219125944765529e-06, 'learning_rate/action_model': 5.2191259447655284e-05, 'epoch': 0.15}) +[2026-06-24 10:34:50,076][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.4960012137889862, 'train/grad_norm_pre_clip': 104.11445617675781, 'timing/data': 0.0003101120237261057, 'timing/model': 1.059629192110151, 'learning_rate/qwen_vl_interface': 5.213980360480828e-06, 'learning_rate/action_model': 5.213980360480828e-05, 'epoch': 0.15}) +[2026-06-24 10:34:51,138][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.17531229555606842, 'train/grad_norm_pre_clip': 62.52194595336914, 'timing/data': 0.0003125160001218319, 'timing/model': 1.0605679850559682, 'learning_rate/qwen_vl_interface': 5.208834818467183e-06, 'learning_rate/action_model': 5.208834818467182e-05, 'epoch': 0.15}) +[2026-06-24 10:34:52,199][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.07867234945297241, 'train/grad_norm_pre_clip': 20.881614685058594, 'timing/data': 0.00029965187422931194, 'timing/model': 1.0595765069592744, 'learning_rate/qwen_vl_interface': 5.203689324763168e-06, 'learning_rate/action_model': 5.203689324763168e-05, 'epoch': 0.15}) +[2026-06-24 10:34:53,262][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.11163435131311417, 'train/grad_norm_pre_clip': 77.50397491455078, 'timing/data': 0.00032889703288674355, 'timing/model': 1.0610471579711884, 'learning_rate/qwen_vl_interface': 5.198543885407311e-06, 'learning_rate/action_model': 5.198543885407311e-05, 'epoch': 0.15}) +[2026-06-24 10:34:54,322][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.15751859545707703, 'train/grad_norm_pre_clip': 42.465232849121094, 'timing/data': 0.000326832989230752, 'timing/model': 1.0586379901506007, 'learning_rate/qwen_vl_interface': 5.193398506438071e-06, 'learning_rate/action_model': 5.193398506438071e-05, 'epoch': 0.15}) +[2026-06-24 10:34:55,385][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.17476287484169006, 'train/grad_norm_pre_clip': 57.03962326049805, 'timing/data': 0.0003252399619668722, 'timing/model': 1.0608545099385083, 'learning_rate/qwen_vl_interface': 5.1882531938938365e-06, 'learning_rate/action_model': 5.188253193893836e-05, 'epoch': 0.15}) +[2026-06-24 10:34:56,447][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.19172988831996918, 'train/grad_norm_pre_clip': 82.09674835205078, 'timing/data': 0.00030441093258559704, 'timing/model': 1.0605537181254476, 'learning_rate/qwen_vl_interface': 5.183107953812917e-06, 'learning_rate/action_model': 5.183107953812917e-05, 'epoch': 0.15}) +[2026-06-24 10:34:57,508][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.12220616638660431, 'train/grad_norm_pre_clip': 56.620601654052734, 'timing/data': 0.0003325939178466797, 'timing/model': 1.0592539811041206, 'learning_rate/qwen_vl_interface': 5.177962792233542e-06, 'learning_rate/action_model': 5.177962792233542e-05, 'epoch': 0.15}) +[2026-06-24 10:34:58,569][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.20261666178703308, 'train/grad_norm_pre_clip': 31.72896385192871, 'timing/data': 0.0003258718643337488, 'timing/model': 1.0601669359020889, 'learning_rate/qwen_vl_interface': 5.172817715193843e-06, 'learning_rate/action_model': 5.1728177151938426e-05, 'epoch': 0.15}) +[2026-06-24 10:34:59,631][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.07105302810668945, 'train/grad_norm_pre_clip': 20.428781509399414, 'timing/data': 0.0003277040086686611, 'timing/model': 1.0606083360034972, 'learning_rate/qwen_vl_interface': 5.1676727287318516e-06, 'learning_rate/action_model': 5.1676727287318516e-05, 'epoch': 0.15}) +[2026-06-24 10:35:00,957][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.2087240070104599, 'train/grad_norm_pre_clip': 79.54871368408203, 'timing/data': 0.00032455893233418465, 'timing/model': 1.3239873589482158, 'learning_rate/qwen_vl_interface': 5.162527838885501e-06, 'learning_rate/action_model': 5.1625278388855006e-05, 'epoch': 0.15}) +[2026-06-24 10:35:02,022][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.07423318922519684, 'train/grad_norm_pre_clip': 20.33726692199707, 'timing/data': 0.00040339608676731586, 'timing/model': 1.0631686521228403, 'learning_rate/qwen_vl_interface': 5.1573830516926015e-06, 'learning_rate/action_model': 5.157383051692601e-05, 'epoch': 0.15}) +[2026-06-24 10:35:03,083][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.12560412287712097, 'train/grad_norm_pre_clip': 32.21126174926758, 'timing/data': 0.0004133349284529686, 'timing/model': 1.05897701298818, 'learning_rate/qwen_vl_interface': 5.15223837319085e-06, 'learning_rate/action_model': 5.15223837319085e-05, 'epoch': 0.15}) +[2026-06-24 10:35:04,146][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.09625595062971115, 'train/grad_norm_pre_clip': 27.341407775878906, 'timing/data': 0.00031557190231978893, 'timing/model': 1.062181934947148, 'learning_rate/qwen_vl_interface': 5.147093809417813e-06, 'learning_rate/action_model': 5.147093809417812e-05, 'epoch': 0.15}) +[2026-06-24 10:35:05,208][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.13790486752986908, 'train/grad_norm_pre_clip': 48.32282638549805, 'timing/data': 0.00035080802626907825, 'timing/model': 1.059848440112546, 'learning_rate/qwen_vl_interface': 5.141949366410923e-06, 'learning_rate/action_model': 5.141949366410923e-05, 'epoch': 0.15}) +[2026-06-24 10:35:06,268][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.06672243028879166, 'train/grad_norm_pre_clip': 23.835290908813477, 'timing/data': 0.0003553559072315693, 'timing/model': 1.0585589441470802, 'learning_rate/qwen_vl_interface': 5.13680505020747e-06, 'learning_rate/action_model': 5.13680505020747e-05, 'epoch': 0.15}) +[2026-06-24 10:35:07,332][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.04195595905184746, 'train/grad_norm_pre_clip': 17.451040267944336, 'timing/data': 0.000332533847540617, 'timing/model': 1.0619622787926346, 'learning_rate/qwen_vl_interface': 5.131660866844596e-06, 'learning_rate/action_model': 5.1316608668445955e-05, 'epoch': 0.15}) +[2026-06-24 10:35:08,396][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.03206661343574524, 'train/grad_norm_pre_clip': 15.495241165161133, 'timing/data': 0.00037221889942884445, 'timing/model': 1.0623912459705025, 'learning_rate/qwen_vl_interface': 5.126516822359286e-06, 'learning_rate/action_model': 5.1265168223592863e-05, 'epoch': 0.15}) +[2026-06-24 10:35:09,460][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.05860486254096031, 'train/grad_norm_pre_clip': 21.385936737060547, 'timing/data': 0.00033462815918028355, 'timing/model': 1.0620774689596146, 'learning_rate/qwen_vl_interface': 5.121372922788363e-06, 'learning_rate/action_model': 5.121372922788362e-05, 'epoch': 0.15}) +[2026-06-24 10:35:10,523][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.22921009361743927, 'train/grad_norm_pre_clip': 60.671485900878906, 'timing/data': 0.0003914439585059881, 'timing/model': 1.0611730341333896, 'learning_rate/qwen_vl_interface': 5.116229174168479e-06, 'learning_rate/action_model': 5.116229174168479e-05, 'epoch': 0.15}) +[2026-06-24 10:35:11,586][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.2280336320400238, 'train/grad_norm_pre_clip': 53.36979293823242, 'timing/data': 0.0003279149532318115, 'timing/model': 1.0616857081186026, 'learning_rate/qwen_vl_interface': 5.11108558253611e-06, 'learning_rate/action_model': 5.111085582536109e-05, 'epoch': 0.15}) +[2026-06-24 10:35:12,648][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.010144234634935856, 'train/grad_norm_pre_clip': 9.02664852142334, 'timing/data': 0.0003100119065493345, 'timing/model': 1.0610129779670388, 'learning_rate/qwen_vl_interface': 5.105942153927548e-06, 'learning_rate/action_model': 5.105942153927548e-05, 'epoch': 0.15}) +[2026-06-24 10:35:13,711][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.06010701507329941, 'train/grad_norm_pre_clip': 46.001426696777344, 'timing/data': 0.00031613302417099476, 'timing/model': 1.0614172781351954, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 0.15}) +[2026-06-24 10:35:14,771][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.2012401819229126, 'train/grad_norm_pre_clip': 61.80240249633789, 'timing/data': 0.00031965901143848896, 'timing/model': 1.0578024592250586, 'learning_rate/qwen_vl_interface': 5.095655809926044e-06, 'learning_rate/action_model': 5.0956558099260435e-05, 'epoch': 0.15}) +[2026-06-24 10:35:15,835][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.05377292260527611, 'train/grad_norm_pre_clip': 35.7261962890625, 'timing/data': 0.0003061641473323107, 'timing/model': 1.0627403538674116, 'learning_rate/qwen_vl_interface': 5.090512906604699e-06, 'learning_rate/action_model': 5.090512906604699e-05, 'epoch': 0.15}) +[2026-06-24 10:35:17,119][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.11823883652687073, 'train/grad_norm_pre_clip': 57.927734375, 'timing/data': 0.00029507302679121494, 'timing/model': 1.2827554140239954, 'learning_rate/qwen_vl_interface': 5.085370190450343e-06, 'learning_rate/action_model': 5.085370190450343e-05, 'epoch': 0.15}) +[2026-06-24 10:35:18,181][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.20005697011947632, 'train/grad_norm_pre_clip': 57.35734939575195, 'timing/data': 0.00033767311833798885, 'timing/model': 1.060409776167944, 'learning_rate/qwen_vl_interface': 5.080227667498239e-06, 'learning_rate/action_model': 5.080227667498238e-05, 'epoch': 0.15}) +[2026-06-24 10:35:19,247][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.04116452857851982, 'train/grad_norm_pre_clip': 21.552013397216797, 'timing/data': 0.000366586958989501, 'timing/model': 1.0641423689667135, 'learning_rate/qwen_vl_interface': 5.075085343783422e-06, 'learning_rate/action_model': 5.0750853437834224e-05, 'epoch': 0.15}) +[2026-06-24 10:35:20,308][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.07862363755702972, 'train/grad_norm_pre_clip': 44.212188720703125, 'timing/data': 0.00032996898517012596, 'timing/model': 1.0597257071640342, 'learning_rate/qwen_vl_interface': 5.069943225340698e-06, 'learning_rate/action_model': 5.0699432253406974e-05, 'epoch': 0.15}) +[2026-06-24 10:35:21,370][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.06865031272172928, 'train/grad_norm_pre_clip': 18.18571662902832, 'timing/data': 0.00030341907404363155, 'timing/model': 1.0602527169976383, 'learning_rate/qwen_vl_interface': 5.064801318204626e-06, 'learning_rate/action_model': 5.064801318204626e-05, 'epoch': 0.15}) +[2026-06-24 10:35:22,496][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.19661597907543182, 'train/grad_norm_pre_clip': 48.989646911621094, 'timing/data': 0.0006074490956962109, 'timing/model': 1.0609574208501726, 'learning_rate/qwen_vl_interface': 5.059659628409525e-06, 'learning_rate/action_model': 5.059659628409524e-05, 'epoch': 0.15}) +[2026-06-24 10:35:23,565][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.10007794946432114, 'train/grad_norm_pre_clip': 48.30007553100586, 'timing/data': 0.0003919149748980999, 'timing/model': 1.0667127878405154, 'learning_rate/qwen_vl_interface': 5.05451816198945e-06, 'learning_rate/action_model': 5.0545181619894496e-05, 'epoch': 0.15}) +[2026-06-24 10:35:24,628][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.04137318953871727, 'train/grad_norm_pre_clip': 48.41193771362305, 'timing/data': 0.0003080980386584997, 'timing/model': 1.0621710009872913, 'learning_rate/qwen_vl_interface': 5.049376924978201e-06, 'learning_rate/action_model': 5.049376924978201e-05, 'epoch': 0.15}) +[2026-06-24 10:35:25,690][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.08791661262512207, 'train/grad_norm_pre_clip': 29.76276206970215, 'timing/data': 0.0004975029733031988, 'timing/model': 1.0592833461705595, 'learning_rate/qwen_vl_interface': 5.044235923409305e-06, 'learning_rate/action_model': 5.044235923409305e-05, 'epoch': 0.15}) +[2026-06-24 10:35:26,753][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.10913649201393127, 'train/grad_norm_pre_clip': 43.696022033691406, 'timing/data': 0.0003671890590339899, 'timing/model': 1.0614769931416959, 'learning_rate/qwen_vl_interface': 5.0390951633160145e-06, 'learning_rate/action_model': 5.039095163316014e-05, 'epoch': 0.15}) +[2026-06-24 10:35:27,815][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.13452228903770447, 'train/grad_norm_pre_clip': 29.048311233520508, 'timing/data': 0.00032689282670617104, 'timing/model': 1.0602037669159472, 'learning_rate/qwen_vl_interface': 5.033954650731296e-06, 'learning_rate/action_model': 5.033954650731296e-05, 'epoch': 0.15}) +[2026-06-24 10:35:28,878][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.1787012368440628, 'train/grad_norm_pre_clip': 37.891693115234375, 'timing/data': 0.0005424378905445337, 'timing/model': 1.061525798169896, 'learning_rate/qwen_vl_interface': 5.028814391687832e-06, 'learning_rate/action_model': 5.0288143916878317e-05, 'epoch': 0.15}) +[2026-06-24 10:35:29,942][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.157640278339386, 'train/grad_norm_pre_clip': 36.32432174682617, 'timing/data': 0.0002995021641254425, 'timing/model': 1.062910827808082, 'learning_rate/qwen_vl_interface': 5.023674392217998e-06, 'learning_rate/action_model': 5.023674392217997e-05, 'epoch': 0.15}) +[2026-06-24 10:35:31,007][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.20963703095912933, 'train/grad_norm_pre_clip': 49.861995697021484, 'timing/data': 0.0003858341369777918, 'timing/model': 1.0627250610850751, 'learning_rate/qwen_vl_interface': 5.018534658353869e-06, 'learning_rate/action_model': 5.0185346583538685e-05, 'epoch': 0.15}) +[2026-06-24 10:35:32,068][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.07608037441968918, 'train/grad_norm_pre_clip': 50.988372802734375, 'timing/data': 0.00032271607778966427, 'timing/model': 1.0600482420995831, 'learning_rate/qwen_vl_interface': 5.013395196127212e-06, 'learning_rate/action_model': 5.0133951961272116e-05, 'epoch': 0.15}) +[2026-06-24 10:35:33,131][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.049256015568971634, 'train/grad_norm_pre_clip': 24.86492919921875, 'timing/data': 0.000298880971968174, 'timing/model': 1.0608066159766167, 'learning_rate/qwen_vl_interface': 5.00825601156947e-06, 'learning_rate/action_model': 5.0082560115694697e-05, 'epoch': 0.15}) +[2026-06-24 10:35:34,480][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.0941791981458664, 'train/grad_norm_pre_clip': 41.656497955322266, 'timing/data': 0.00035506580024957657, 'timing/model': 1.3480529838707298, 'learning_rate/qwen_vl_interface': 5.003117110711766e-06, 'learning_rate/action_model': 5.003117110711766e-05, 'epoch': 0.15}) +[2026-06-24 10:35:35,541][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.0547744482755661, 'train/grad_norm_pre_clip': 17.248825073242188, 'timing/data': 0.0003079078160226345, 'timing/model': 1.0592800870072097, 'learning_rate/qwen_vl_interface': 4.997978499584882e-06, 'learning_rate/action_model': 4.997978499584882e-05, 'epoch': 0.15}) +[2026-06-24 10:35:36,603][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.1019401103258133, 'train/grad_norm_pre_clip': 27.866506576538086, 'timing/data': 0.0005194940604269505, 'timing/model': 1.0593107258901, 'learning_rate/qwen_vl_interface': 4.992840184219267e-06, 'learning_rate/action_model': 4.9928401842192666e-05, 'epoch': 0.15}) +[2026-06-24 10:35:37,665][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.07595571875572205, 'train/grad_norm_pre_clip': 31.5455322265625, 'timing/data': 0.0003053429536521435, 'timing/model': 1.0609653620049357, 'learning_rate/qwen_vl_interface': 4.987702170645019e-06, 'learning_rate/action_model': 4.9877021706450185e-05, 'epoch': 0.15}) +[2026-06-24 10:35:38,729][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.07396179437637329, 'train/grad_norm_pre_clip': 33.41524887084961, 'timing/data': 0.00033405586145818233, 'timing/model': 1.0617386440280825, 'learning_rate/qwen_vl_interface': 4.982564464891886e-06, 'learning_rate/action_model': 4.982564464891886e-05, 'epoch': 0.16}) +[2026-06-24 10:35:39,790][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.24661503732204437, 'train/grad_norm_pre_clip': 128.020751953125, 'timing/data': 0.0003063250333070755, 'timing/model': 1.0602104892022908, 'learning_rate/qwen_vl_interface': 4.977427072989248e-06, 'learning_rate/action_model': 4.977427072989247e-05, 'epoch': 0.16}) +[2026-06-24 10:35:40,849][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.35012465715408325, 'train/grad_norm_pre_clip': 72.8405532836914, 'timing/data': 0.0002980688586831093, 'timing/model': 1.0567219369113445, 'learning_rate/qwen_vl_interface': 4.9722900009661215e-06, 'learning_rate/action_model': 4.972290000966121e-05, 'epoch': 0.16}) +[2026-06-24 10:35:41,908][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.11232595890760422, 'train/grad_norm_pre_clip': 43.375450134277344, 'timing/data': 0.0003043611068278551, 'timing/model': 1.0573034200351685, 'learning_rate/qwen_vl_interface': 4.9671532548511475e-06, 'learning_rate/action_model': 4.9671532548511474e-05, 'epoch': 0.16}) +[2026-06-24 10:35:42,971][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.22417500615119934, 'train/grad_norm_pre_clip': 81.9570083618164, 'timing/data': 0.0003045110497623682, 'timing/model': 1.0616703329142183, 'learning_rate/qwen_vl_interface': 4.962016840672583e-06, 'learning_rate/action_model': 4.962016840672583e-05, 'epoch': 0.16}) +[2026-06-24 10:35:44,032][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.10547865927219391, 'train/grad_norm_pre_clip': 24.150897979736328, 'timing/data': 0.00031051295809447765, 'timing/model': 1.0600577199365944, 'learning_rate/qwen_vl_interface': 4.956880764458298e-06, 'learning_rate/action_model': 4.9568807644582984e-05, 'epoch': 0.16}) +[2026-06-24 10:35:45,096][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.022292079403996468, 'train/grad_norm_pre_clip': 8.492384910583496, 'timing/data': 0.0005566738545894623, 'timing/model': 1.0616343191359192, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 0.16}) +[2026-06-24 10:35:46,160][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.25637704133987427, 'train/grad_norm_pre_clip': 50.595611572265625, 'timing/data': 0.00031017186120152473, 'timing/model': 1.0625363430008292, 'learning_rate/qwen_vl_interface': 4.9466096500320435e-06, 'learning_rate/action_model': 4.946609650032043e-05, 'epoch': 0.16}) +[2026-06-24 10:35:47,225][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.0704924538731575, 'train/grad_norm_pre_clip': 19.90154266357422, 'timing/data': 0.0003061241004616022, 'timing/model': 1.0628318791277707, 'learning_rate/qwen_vl_interface': 4.941474623873798e-06, 'learning_rate/action_model': 4.941474623873798e-05, 'epoch': 0.16}) +[2026-06-24 10:35:48,289][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.03574632853269577, 'train/grad_norm_pre_clip': 17.63034439086914, 'timing/data': 0.000332763884216547, 'timing/model': 1.0622717500664294, 'learning_rate/qwen_vl_interface': 4.936339959787268e-06, 'learning_rate/action_model': 4.9363399597872676e-05, 'epoch': 0.16}) +[2026-06-24 10:35:49,350][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.1854931116104126, 'train/grad_norm_pre_clip': 107.83788299560547, 'timing/data': 0.0003397171385586262, 'timing/model': 1.0596346848178655, 'learning_rate/qwen_vl_interface': 4.931205663798268e-06, 'learning_rate/action_model': 4.931205663798267e-05, 'epoch': 0.16}) +[2026-06-24 10:35:50,616][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.23319675028324127, 'train/grad_norm_pre_clip': 56.82099533081055, 'timing/data': 0.00034034810960292816, 'timing/model': 1.2643539439886808, 'learning_rate/qwen_vl_interface': 4.926071741932174e-06, 'learning_rate/action_model': 4.926071741932174e-05, 'epoch': 0.16}) +[2026-06-24 10:35:51,685][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.16571453213691711, 'train/grad_norm_pre_clip': 61.11922836303711, 'timing/data': 0.0003244890831410885, 'timing/model': 1.0680649029091, 'learning_rate/qwen_vl_interface': 4.920938200213936e-06, 'learning_rate/action_model': 4.920938200213936e-05, 'epoch': 0.16}) +[2026-06-24 10:35:52,749][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.2094525545835495, 'train/grad_norm_pre_clip': 48.384117126464844, 'timing/data': 0.0003037089481949806, 'timing/model': 1.0621718999464065, 'learning_rate/qwen_vl_interface': 4.915805044668048e-06, 'learning_rate/action_model': 4.9158050446680476e-05, 'epoch': 0.16}) +[2026-06-24 10:35:53,813][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.2311743199825287, 'train/grad_norm_pre_clip': 44.41843032836914, 'timing/data': 0.00039225583896040916, 'timing/model': 1.062464159913361, 'learning_rate/qwen_vl_interface': 4.9106722813185555e-06, 'learning_rate/action_model': 4.9106722813185554e-05, 'epoch': 0.16}) +[2026-06-24 10:35:54,877][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.06479541957378387, 'train/grad_norm_pre_clip': 23.011045455932617, 'timing/data': 0.00030749617144465446, 'timing/model': 1.062277961988002, 'learning_rate/qwen_vl_interface': 4.90553991618904e-06, 'learning_rate/action_model': 4.90553991618904e-05, 'epoch': 0.16}) +[2026-06-24 10:35:55,942][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.24713078141212463, 'train/grad_norm_pre_clip': 80.35466766357422, 'timing/data': 0.0003224948886781931, 'timing/model': 1.0627446291036904, 'learning_rate/qwen_vl_interface': 4.900407955302621e-06, 'learning_rate/action_model': 4.900407955302621e-05, 'epoch': 0.16}) +[2026-06-24 10:35:57,006][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.058876845985651016, 'train/grad_norm_pre_clip': 15.533491134643555, 'timing/data': 0.0003235870972275734, 'timing/model': 1.0626070939470083, 'learning_rate/qwen_vl_interface': 4.895276404681936e-06, 'learning_rate/action_model': 4.895276404681936e-05, 'epoch': 0.16}) +[2026-06-24 10:35:58,067][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.05010107532143593, 'train/grad_norm_pre_clip': 30.45368003845215, 'timing/data': 0.00032434891909360886, 'timing/model': 1.0596455498598516, 'learning_rate/qwen_vl_interface': 4.8901452703491505e-06, 'learning_rate/action_model': 4.8901452703491506e-05, 'epoch': 0.16}) +[2026-06-24 10:35:59,132][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.08815464377403259, 'train/grad_norm_pre_clip': 24.43408966064453, 'timing/data': 0.00029785814695060253, 'timing/model': 1.0630440760869533, 'learning_rate/qwen_vl_interface': 4.885014558325935e-06, 'learning_rate/action_model': 4.885014558325935e-05, 'epoch': 0.16}) +[2026-06-24 10:36:00,195][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.055368322879076004, 'train/grad_norm_pre_clip': 23.94919204711914, 'timing/data': 0.0003505079075694084, 'timing/model': 1.0615545939654112, 'learning_rate/qwen_vl_interface': 4.879884274633463e-06, 'learning_rate/action_model': 4.879884274633463e-05, 'epoch': 0.16}) +[2026-06-24 10:36:01,258][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.07605709135532379, 'train/grad_norm_pre_clip': 23.649978637695312, 'timing/data': 0.00038094515912234783, 'timing/model': 1.061945629073307, 'learning_rate/qwen_vl_interface': 4.8747544252924115e-06, 'learning_rate/action_model': 4.8747544252924107e-05, 'epoch': 0.16}) +[2026-06-24 10:36:02,323][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.11225803941488266, 'train/grad_norm_pre_clip': 35.86481857299805, 'timing/data': 0.0003452377859503031, 'timing/model': 1.0622557830065489, 'learning_rate/qwen_vl_interface': 4.869625016322941e-06, 'learning_rate/action_model': 4.869625016322941e-05, 'epoch': 0.16}) +[2026-06-24 10:36:03,386][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.0902673527598381, 'train/grad_norm_pre_clip': 31.291912078857422, 'timing/data': 0.00033287517726421356, 'timing/model': 1.0611263359896839, 'learning_rate/qwen_vl_interface': 4.864496053744703e-06, 'learning_rate/action_model': 4.864496053744702e-05, 'epoch': 0.16}) +[2026-06-24 10:36:04,451][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.20848023891448975, 'train/grad_norm_pre_clip': 72.26051330566406, 'timing/data': 0.0003404489252716303, 'timing/model': 1.0636610719375312, 'learning_rate/qwen_vl_interface': 4.859367543576816e-06, 'learning_rate/action_model': 4.859367543576816e-05, 'epoch': 0.16}) +[2026-06-24 10:36:05,513][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.15298987925052643, 'train/grad_norm_pre_clip': 42.64519500732422, 'timing/data': 0.00036406307481229305, 'timing/model': 1.0607147270347923, 'learning_rate/qwen_vl_interface': 4.854239491837876e-06, 'learning_rate/action_model': 4.854239491837876e-05, 'epoch': 0.16}) +[2026-06-24 10:36:06,777][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.19404210150241852, 'train/grad_norm_pre_clip': 32.44095230102539, 'timing/data': 0.0003137888852506876, 'timing/model': 1.261998777044937, 'learning_rate/qwen_vl_interface': 4.8491119045459355e-06, 'learning_rate/action_model': 4.8491119045459356e-05, 'epoch': 0.16}) +[2026-06-24 10:36:07,836][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.09701400250196457, 'train/grad_norm_pre_clip': 31.23697280883789, 'timing/data': 0.00030497205443680286, 'timing/model': 1.0579111210536212, 'learning_rate/qwen_vl_interface': 4.8439847877185055e-06, 'learning_rate/action_model': 4.843984787718505e-05, 'epoch': 0.16}) +[2026-06-24 10:36:08,897][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.040587231516838074, 'train/grad_norm_pre_clip': 22.204559326171875, 'timing/data': 0.0003546548541635275, 'timing/model': 1.0588168818503618, 'learning_rate/qwen_vl_interface': 4.8388581473725415e-06, 'learning_rate/action_model': 4.8388581473725416e-05, 'epoch': 0.16}) +[2026-06-24 10:36:09,957][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.06842342019081116, 'train/grad_norm_pre_clip': 31.208974838256836, 'timing/data': 0.00031617283821105957, 'timing/model': 1.0586923300288618, 'learning_rate/qwen_vl_interface': 4.833731989524444e-06, 'learning_rate/action_model': 4.833731989524443e-05, 'epoch': 0.16}) +[2026-06-24 10:36:11,022][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.020288264378905296, 'train/grad_norm_pre_clip': 15.96072006225586, 'timing/data': 0.00030679511837661266, 'timing/model': 1.0630634508561343, 'learning_rate/qwen_vl_interface': 4.828606320190041e-06, 'learning_rate/action_model': 4.828606320190041e-05, 'epoch': 0.16}) +[2026-06-24 10:36:12,085][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.09519746154546738, 'train/grad_norm_pre_clip': 37.72681427001953, 'timing/data': 0.0003128668759018183, 'timing/model': 1.0615278219338506, 'learning_rate/qwen_vl_interface': 4.823481145384593e-06, 'learning_rate/action_model': 4.823481145384592e-05, 'epoch': 0.16}) +[2026-06-24 10:36:13,149][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.03997870534658432, 'train/grad_norm_pre_clip': 27.5162296295166, 'timing/data': 0.0003397171385586262, 'timing/model': 1.0624263698700815, 'learning_rate/qwen_vl_interface': 4.818356471122778e-06, 'learning_rate/action_model': 4.8183564711227774e-05, 'epoch': 0.16}) +[2026-06-24 10:36:14,213][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.0198200773447752, 'train/grad_norm_pre_clip': 6.633027076721191, 'timing/data': 0.00039589195512235165, 'timing/model': 1.0623674611561, 'learning_rate/qwen_vl_interface': 4.813232303418686e-06, 'learning_rate/action_model': 4.8132323034186856e-05, 'epoch': 0.16}) +[2026-06-24 10:36:15,278][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.20003104209899902, 'train/grad_norm_pre_clip': 49.73630142211914, 'timing/data': 0.00030111405067145824, 'timing/model': 1.0629642319399863, 'learning_rate/qwen_vl_interface': 4.808108648285815e-06, 'learning_rate/action_model': 4.808108648285814e-05, 'epoch': 0.16}) +[2026-06-24 10:36:16,342][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.04146140068769455, 'train/grad_norm_pre_clip': 28.124113082885742, 'timing/data': 0.00033439695835113525, 'timing/model': 1.0629875068552792, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 0.16}) +[2026-06-24 10:36:17,404][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.08511047065258026, 'train/grad_norm_pre_clip': 37.36710739135742, 'timing/data': 0.00033916602842509747, 'timing/model': 1.0599781039636582, 'learning_rate/qwen_vl_interface': 4.7978628997847e-06, 'learning_rate/action_model': 4.7978628997847006e-05, 'epoch': 0.16}) +[2026-06-24 10:36:18,468][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.1520709991455078, 'train/grad_norm_pre_clip': 50.55079650878906, 'timing/data': 0.00032608094625175, 'timing/model': 1.0631061140447855, 'learning_rate/qwen_vl_interface': 4.792740818440415e-06, 'learning_rate/action_model': 4.792740818440415e-05, 'epoch': 0.16}) +[2026-06-24 10:36:19,533][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.21376939117908478, 'train/grad_norm_pre_clip': 56.7022590637207, 'timing/data': 0.00032000988721847534, 'timing/model': 1.0633223508484662, 'learning_rate/qwen_vl_interface': 4.7876192737152495e-06, 'learning_rate/action_model': 4.787619273715249e-05, 'epoch': 0.16}) +[2026-06-24 10:36:20,597][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.1029893234372139, 'train/grad_norm_pre_clip': 69.81307983398438, 'timing/data': 0.0003420610446482897, 'timing/model': 1.0622345109004527, 'learning_rate/qwen_vl_interface': 4.78249827161962e-06, 'learning_rate/action_model': 4.78249827161962e-05, 'epoch': 0.16}) +[2026-06-24 10:36:21,664][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.28165966272354126, 'train/grad_norm_pre_clip': 54.052513122558594, 'timing/data': 0.0003974249120801687, 'timing/model': 1.064911694964394, 'learning_rate/qwen_vl_interface': 4.777377818163308e-06, 'learning_rate/action_model': 4.7773778181633076e-05, 'epoch': 0.16}) +[2026-06-24 10:36:22,939][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.1568831205368042, 'train/grad_norm_pre_clip': 59.69916534423828, 'timing/data': 0.0003189078997820616, 'timing/model': 1.2730231271125376, 'learning_rate/qwen_vl_interface': 4.772257919355452e-06, 'learning_rate/action_model': 4.7722579193554515e-05, 'epoch': 0.16}) +[2026-06-24 10:36:24,002][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.0030065292958170176, 'train/grad_norm_pre_clip': 1.527797818183899, 'timing/data': 0.0003278248477727175, 'timing/model': 1.0618552449159324, 'learning_rate/qwen_vl_interface': 4.767138581204537e-06, 'learning_rate/action_model': 4.767138581204537e-05, 'epoch': 0.16}) +[2026-06-24 10:36:25,067][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.08884533494710922, 'train/grad_norm_pre_clip': 27.524072647094727, 'timing/data': 0.00030406098812818527, 'timing/model': 1.063509199069813, 'learning_rate/qwen_vl_interface': 4.762019809718393e-06, 'learning_rate/action_model': 4.762019809718393e-05, 'epoch': 0.16}) +[2026-06-24 10:36:26,130][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.029401270672678947, 'train/grad_norm_pre_clip': 72.47943878173828, 'timing/data': 0.00032514985650777817, 'timing/model': 1.0610820890869945, 'learning_rate/qwen_vl_interface': 4.756901610904181e-06, 'learning_rate/action_model': 4.756901610904181e-05, 'epoch': 0.16}) +[2026-06-24 10:36:27,193][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.1617266684770584, 'train/grad_norm_pre_clip': 50.62264633178711, 'timing/data': 0.00031730509363114834, 'timing/model': 1.0613567850086838, 'learning_rate/qwen_vl_interface': 4.7517839907683975e-06, 'learning_rate/action_model': 4.751783990768397e-05, 'epoch': 0.16}) +[2026-06-24 10:36:28,260][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.09921365976333618, 'train/grad_norm_pre_clip': 48.68498229980469, 'timing/data': 0.00030865916050970554, 'timing/model': 1.065064881928265, 'learning_rate/qwen_vl_interface': 4.746666955316849e-06, 'learning_rate/action_model': 4.7466669553168486e-05, 'epoch': 0.16}) +[2026-06-24 10:36:29,324][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.35675048828125, 'train/grad_norm_pre_clip': 109.71318817138672, 'timing/data': 0.00034168106503784657, 'timing/model': 1.0631525670178235, 'learning_rate/qwen_vl_interface': 4.741550510554666e-06, 'learning_rate/action_model': 4.7415505105546655e-05, 'epoch': 0.16}) +[2026-06-24 10:36:30,389][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.0878591239452362, 'train/grad_norm_pre_clip': 51.628257751464844, 'timing/data': 0.0005005691200494766, 'timing/model': 1.0623491320293397, 'learning_rate/qwen_vl_interface': 4.736434662486278e-06, 'learning_rate/action_model': 4.7364346624862774e-05, 'epoch': 0.16}) +[2026-06-24 10:36:31,454][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.18446297943592072, 'train/grad_norm_pre_clip': 78.61845397949219, 'timing/data': 0.0003280548844486475, 'timing/model': 1.0632040370255709, 'learning_rate/qwen_vl_interface': 4.731319417115419e-06, 'learning_rate/action_model': 4.7313194171154184e-05, 'epoch': 0.16}) +[2026-06-24 10:36:32,515][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.16909463703632355, 'train/grad_norm_pre_clip': 58.224266052246094, 'timing/data': 0.0003285969141870737, 'timing/model': 1.0600827829912305, 'learning_rate/qwen_vl_interface': 4.726204780445115e-06, 'learning_rate/action_model': 4.726204780445114e-05, 'epoch': 0.16}) +[2026-06-24 10:36:33,576][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.1034340038895607, 'train/grad_norm_pre_clip': 69.74276733398438, 'timing/data': 0.0003018360584974289, 'timing/model': 1.0592236339580268, 'learning_rate/qwen_vl_interface': 4.721090758477674e-06, 'learning_rate/action_model': 4.7210907584776745e-05, 'epoch': 0.16}) +[2026-06-24 10:36:34,636][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.46447205543518066, 'train/grad_norm_pre_clip': 111.51576232910156, 'timing/data': 0.0003009738866239786, 'timing/model': 1.0584269808605313, 'learning_rate/qwen_vl_interface': 4.715977357214691e-06, 'learning_rate/action_model': 4.715977357214691e-05, 'epoch': 0.16}) +[2026-06-24 10:36:35,701][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.08025748282670975, 'train/grad_norm_pre_clip': 27.594881057739258, 'timing/data': 0.0002937908284366131, 'timing/model': 1.0631877328269184, 'learning_rate/qwen_vl_interface': 4.710864582657023e-06, 'learning_rate/action_model': 4.7108645826570225e-05, 'epoch': 0.16}) +[2026-06-24 10:36:36,765][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.11160562187433243, 'train/grad_norm_pre_clip': 103.63237762451172, 'timing/data': 0.0003320830874145031, 'timing/model': 1.0626413689460605, 'learning_rate/qwen_vl_interface': 4.7057524408047975e-06, 'learning_rate/action_model': 4.705752440804798e-05, 'epoch': 0.16}) +[2026-06-24 10:36:37,830][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.14409680664539337, 'train/grad_norm_pre_clip': 73.3707275390625, 'timing/data': 0.00035311304964125156, 'timing/model': 1.0632652309723198, 'learning_rate/qwen_vl_interface': 4.700640937657397e-06, 'learning_rate/action_model': 4.700640937657397e-05, 'epoch': 0.16}) +[2026-06-24 10:36:39,189][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.08970708400011063, 'train/grad_norm_pre_clip': 33.440555572509766, 'timing/data': 0.0003795519005507231, 'timing/model': 1.3574160691350698, 'learning_rate/qwen_vl_interface': 4.695530079213455e-06, 'learning_rate/action_model': 4.6955300792134554e-05, 'epoch': 0.16}) +[2026-06-24 10:36:40,251][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.05737045034766197, 'train/grad_norm_pre_clip': 19.42981719970703, 'timing/data': 0.00032179406844079494, 'timing/model': 1.0603726690169424, 'learning_rate/qwen_vl_interface': 4.690419871470852e-06, 'learning_rate/action_model': 4.690419871470852e-05, 'epoch': 0.16}) +[2026-06-24 10:36:41,311][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.055092472583055496, 'train/grad_norm_pre_clip': 34.08434295654297, 'timing/data': 0.0003241179510951042, 'timing/model': 1.0591477630659938, 'learning_rate/qwen_vl_interface': 4.6853103204266946e-06, 'learning_rate/action_model': 4.6853103204266944e-05, 'epoch': 0.16}) +[2026-06-24 10:36:42,373][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.34696078300476074, 'train/grad_norm_pre_clip': 53.43653869628906, 'timing/data': 0.00029101595282554626, 'timing/model': 1.0600735500920564, 'learning_rate/qwen_vl_interface': 4.6802014320773305e-06, 'learning_rate/action_model': 4.680201432077331e-05, 'epoch': 0.16}) +[2026-06-24 10:36:43,433][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.0169823095202446, 'train/grad_norm_pre_clip': 9.341833114624023, 'timing/data': 0.0002974369563162327, 'timing/model': 1.0582589381374419, 'learning_rate/qwen_vl_interface': 4.675093212418324e-06, 'learning_rate/action_model': 4.675093212418324e-05, 'epoch': 0.16}) +[2026-06-24 10:36:44,497][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.19146911799907684, 'train/grad_norm_pre_clip': 53.19157409667969, 'timing/data': 0.000310773029923439, 'timing/model': 1.0624860688112676, 'learning_rate/qwen_vl_interface': 4.669985667444454e-06, 'learning_rate/action_model': 4.669985667444454e-05, 'epoch': 0.16}) +[2026-06-24 10:36:45,562][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.1483510583639145, 'train/grad_norm_pre_clip': 24.50100326538086, 'timing/data': 0.00032575102522969246, 'timing/model': 1.06397883198224, 'learning_rate/qwen_vl_interface': 4.664878803149711e-06, 'learning_rate/action_model': 4.66487880314971e-05, 'epoch': 0.16}) +[2026-06-24 10:36:46,626][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.04541221633553505, 'train/grad_norm_pre_clip': 18.890851974487305, 'timing/data': 0.0003028081264346838, 'timing/model': 1.0619961530901492, 'learning_rate/qwen_vl_interface': 4.659772625527284e-06, 'learning_rate/action_model': 4.6597726255272834e-05, 'epoch': 0.16}) +[2026-06-24 10:36:47,690][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.15646429359912872, 'train/grad_norm_pre_clip': 40.78544616699219, 'timing/data': 0.00031307688914239407, 'timing/model': 1.0623718900606036, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 0.16}) +[2026-06-24 10:36:48,752][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.06324755400419235, 'train/grad_norm_pre_clip': 27.028440475463867, 'timing/data': 0.0003218138590455055, 'timing/model': 1.0607907557860017, 'learning_rate/qwen_vl_interface': 4.649562354268097e-06, 'learning_rate/action_model': 4.6495623542680964e-05, 'epoch': 0.16}) +[2026-06-24 10:36:49,817][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.014386928640305996, 'train/grad_norm_pre_clip': 7.527883052825928, 'timing/data': 0.00033862493000924587, 'timing/model': 1.0634286748245358, 'learning_rate/qwen_vl_interface': 4.644458272613662e-06, 'learning_rate/action_model': 4.644458272613662e-05, 'epoch': 0.16}) +[2026-06-24 10:36:50,880][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.15131939947605133, 'train/grad_norm_pre_clip': 43.116294860839844, 'timing/data': 0.0003155921585857868, 'timing/model': 1.0611580279655755, 'learning_rate/qwen_vl_interface': 4.639354901596171e-06, 'learning_rate/action_model': 4.639354901596171e-05, 'epoch': 0.16}) +[2026-06-24 10:36:51,942][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.05698874965310097, 'train/grad_norm_pre_clip': 35.308998107910156, 'timing/data': 0.00030854903161525726, 'timing/model': 1.060617855982855, 'learning_rate/qwen_vl_interface': 4.634252247204718e-06, 'learning_rate/action_model': 4.634252247204717e-05, 'epoch': 0.16}) +[2026-06-24 10:36:53,011][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.16112369298934937, 'train/grad_norm_pre_clip': 41.226783752441406, 'timing/data': 0.0003493949770927429, 'timing/model': 1.0675857190508395, 'learning_rate/qwen_vl_interface': 4.629150315427547e-06, 'learning_rate/action_model': 4.629150315427547e-05, 'epoch': 0.16}) +[2026-06-24 10:36:54,284][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.10069619864225388, 'train/grad_norm_pre_clip': 48.868282318115234, 'timing/data': 0.00032433890737593174, 'timing/model': 1.27171206404455, 'learning_rate/qwen_vl_interface': 4.624049112252064e-06, 'learning_rate/action_model': 4.624049112252063e-05, 'epoch': 0.16}) +[2026-06-24 10:36:55,350][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.005258854012936354, 'train/grad_norm_pre_clip': 4.5179033279418945, 'timing/data': 0.00033658090978860855, 'timing/model': 1.064220919040963, 'learning_rate/qwen_vl_interface': 4.618948643664812e-06, 'learning_rate/action_model': 4.618948643664812e-05, 'epoch': 0.16}) +[2026-06-24 10:36:56,414][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.05622033774852753, 'train/grad_norm_pre_clip': 43.141754150390625, 'timing/data': 0.00030343886464834213, 'timing/model': 1.0626031758729368, 'learning_rate/qwen_vl_interface': 4.613848915651476e-06, 'learning_rate/action_model': 4.6138489156514764e-05, 'epoch': 0.16}) +[2026-06-24 10:36:57,479][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.07873550057411194, 'train/grad_norm_pre_clip': 27.19111442565918, 'timing/data': 0.0003229358699172735, 'timing/model': 1.0626795107964426, 'learning_rate/qwen_vl_interface': 4.608749934196871e-06, 'learning_rate/action_model': 4.6087499341968715e-05, 'epoch': 0.16}) +[2026-06-24 10:36:58,540][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.21572910249233246, 'train/grad_norm_pre_clip': 102.24267578125, 'timing/data': 0.0003115539439022541, 'timing/model': 1.0594593409914523, 'learning_rate/qwen_vl_interface': 4.603651705284938e-06, 'learning_rate/action_model': 4.603651705284938e-05, 'epoch': 0.16}) +[2026-06-24 10:36:59,601][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.026929320767521858, 'train/grad_norm_pre_clip': 19.95768928527832, 'timing/data': 0.0002987708430737257, 'timing/model': 1.0601482461206615, 'learning_rate/qwen_vl_interface': 4.5985542348987275e-06, 'learning_rate/action_model': 4.5985542348987275e-05, 'epoch': 0.16}) +[2026-06-24 10:37:00,663][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.055538278073072433, 'train/grad_norm_pre_clip': 25.010072708129883, 'timing/data': 0.00030525284819304943, 'timing/model': 1.0603441551793367, 'learning_rate/qwen_vl_interface': 4.593457529020411e-06, 'learning_rate/action_model': 4.59345752902041e-05, 'epoch': 0.16}) +[2026-06-24 10:37:01,728][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.24387042224407196, 'train/grad_norm_pre_clip': 71.4298324584961, 'timing/data': 0.00033034011721611023, 'timing/model': 1.0624650749377906, 'learning_rate/qwen_vl_interface': 4.588361593631251e-06, 'learning_rate/action_model': 4.5883615936312514e-05, 'epoch': 0.16}) +[2026-06-24 10:37:02,791][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.15744838118553162, 'train/grad_norm_pre_clip': 65.59031677246094, 'timing/data': 0.0003026172053068876, 'timing/model': 1.0621859929524362, 'learning_rate/qwen_vl_interface': 4.583266434711614e-06, 'learning_rate/action_model': 4.583266434711614e-05, 'epoch': 0.16}) +[2026-06-24 10:37:03,856][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.2007906138896942, 'train/grad_norm_pre_clip': 75.99889373779297, 'timing/data': 0.0005377689376473427, 'timing/model': 1.0623056679032743, 'learning_rate/qwen_vl_interface': 4.578172058240955e-06, 'learning_rate/action_model': 4.578172058240955e-05, 'epoch': 0.16}) +[2026-06-24 10:37:04,921][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.18597447872161865, 'train/grad_norm_pre_clip': 56.13098907470703, 'timing/data': 0.00031414907425642014, 'timing/model': 1.0643175751902163, 'learning_rate/qwen_vl_interface': 4.5730784701978046e-06, 'learning_rate/action_model': 4.573078470197804e-05, 'epoch': 0.16}) +[2026-06-24 10:37:05,986][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.5256863236427307, 'train/grad_norm_pre_clip': 79.94547271728516, 'timing/data': 0.0003082181792706251, 'timing/model': 1.06306376401335, 'learning_rate/qwen_vl_interface': 4.567985676559774e-06, 'learning_rate/action_model': 4.567985676559774e-05, 'epoch': 0.16}) +[2026-06-24 10:37:07,049][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.006566075142472982, 'train/grad_norm_pre_clip': 8.596353530883789, 'timing/data': 0.0003226750995963812, 'timing/model': 1.061251626117155, 'learning_rate/qwen_vl_interface': 4.562893683303539e-06, 'learning_rate/action_model': 4.5628936833035394e-05, 'epoch': 0.16}) +[2026-06-24 10:37:08,113][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.309099018573761, 'train/grad_norm_pre_clip': 66.75084686279297, 'timing/data': 0.0003046609926968813, 'timing/model': 1.0623698749113828, 'learning_rate/qwen_vl_interface': 4.557802496404841e-06, 'learning_rate/action_model': 4.55780249640484e-05, 'epoch': 0.16}) +[2026-06-24 10:37:09,177][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.01971522904932499, 'train/grad_norm_pre_clip': 15.959811210632324, 'timing/data': 0.0002973470836877823, 'timing/model': 1.0629148790612817, 'learning_rate/qwen_vl_interface': 4.552712121838466e-06, 'learning_rate/action_model': 4.552712121838466e-05, 'epoch': 0.16}) +[2026-06-24 10:37:10,241][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.030516967177391052, 'train/grad_norm_pre_clip': 23.6398868560791, 'timing/data': 0.00038510304875671864, 'timing/model': 1.0622726362198591, 'learning_rate/qwen_vl_interface': 4.547622565578258e-06, 'learning_rate/action_model': 4.5476225655782576e-05, 'epoch': 0.16}) +[2026-06-24 10:37:11,305][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.042139094322919846, 'train/grad_norm_pre_clip': 34.66202163696289, 'timing/data': 0.000316031975671649, 'timing/model': 1.0627260280307382, 'learning_rate/qwen_vl_interface': 4.54253383359709e-06, 'learning_rate/action_model': 4.54253383359709e-05, 'epoch': 0.16}) +[2026-06-24 10:37:12,370][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.00044707892811857164, 'train/grad_norm_pre_clip': 0.8925931453704834, 'timing/data': 0.0003076770808547735, 'timing/model': 1.0631593628786504, 'learning_rate/qwen_vl_interface': 4.537445931866874e-06, 'learning_rate/action_model': 4.537445931866874e-05, 'epoch': 0.16}) +[2026-06-24 10:37:13,713][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.01965450868010521, 'train/grad_norm_pre_clip': 15.923871040344238, 'timing/data': 0.0003077569417655468, 'timing/model': 1.3409070228226483, 'learning_rate/qwen_vl_interface': 4.532358866358546e-06, 'learning_rate/action_model': 4.532358866358546e-05, 'epoch': 0.16}) +[2026-06-24 10:37:14,776][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.17172424495220184, 'train/grad_norm_pre_clip': 56.50773620605469, 'timing/data': 0.00031482987105846405, 'timing/model': 1.0616794298402965, 'learning_rate/qwen_vl_interface': 4.52727264304206e-06, 'learning_rate/action_model': 4.52727264304206e-05, 'epoch': 0.16}) +[2026-06-24 10:37:15,842][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.11447466909885406, 'train/grad_norm_pre_clip': 66.58724212646484, 'timing/data': 0.0003452468663454056, 'timing/model': 1.064738499931991, 'learning_rate/qwen_vl_interface': 4.5221872678863846e-06, 'learning_rate/action_model': 4.5221872678863844e-05, 'epoch': 0.16}) +[2026-06-24 10:37:16,908][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.2191024124622345, 'train/grad_norm_pre_clip': 77.59351348876953, 'timing/data': 0.00029993196949362755, 'timing/model': 1.0639413851313293, 'learning_rate/qwen_vl_interface': 4.5171027468594885e-06, 'learning_rate/action_model': 4.517102746859488e-05, 'epoch': 0.16}) +[2026-06-24 10:37:17,972][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.13104037940502167, 'train/grad_norm_pre_clip': 55.56077575683594, 'timing/data': 0.00029822904616594315, 'timing/model': 1.063008206896484, 'learning_rate/qwen_vl_interface': 4.512019085928342e-06, 'learning_rate/action_model': 4.512019085928342e-05, 'epoch': 0.16}) +[2026-06-24 10:37:19,037][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.1047578677535057, 'train/grad_norm_pre_clip': 46.60386276245117, 'timing/data': 0.00030172616243362427, 'timing/model': 1.0635270809289068, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 0.16}) +[2026-06-24 10:37:20,101][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.10071396827697754, 'train/grad_norm_pre_clip': 70.27609252929688, 'timing/data': 0.0005181520245969296, 'timing/model': 1.0617011371068656, 'learning_rate/qwen_vl_interface': 4.501854368216117e-06, 'learning_rate/action_model': 4.501854368216116e-05, 'epoch': 0.16}) +[2026-06-24 10:37:21,165][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.07261341065168381, 'train/grad_norm_pre_clip': 30.902908325195312, 'timing/data': 0.0004729169886559248, 'timing/model': 1.062311985064298, 'learning_rate/qwen_vl_interface': 4.4967733233638995e-06, 'learning_rate/action_model': 4.496773323363899e-05, 'epoch': 0.16}) +[2026-06-24 10:37:22,229][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.009419308044016361, 'train/grad_norm_pre_clip': 9.95322322845459, 'timing/data': 0.00030292803421616554, 'timing/model': 1.0624575591646135, 'learning_rate/qwen_vl_interface': 4.4916931624651415e-06, 'learning_rate/action_model': 4.491693162465141e-05, 'epoch': 0.16}) +[2026-06-24 10:37:23,294][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.4516342282295227, 'train/grad_norm_pre_clip': 180.5625457763672, 'timing/data': 0.0003069050144404173, 'timing/model': 1.0629723460879177, 'learning_rate/qwen_vl_interface': 4.486613891481696e-06, 'learning_rate/action_model': 4.486613891481696e-05, 'epoch': 0.16}) +[2026-06-24 10:37:24,358][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.10898865014314651, 'train/grad_norm_pre_clip': 57.93635940551758, 'timing/data': 0.00034845294430851936, 'timing/model': 1.0628534650895745, 'learning_rate/qwen_vl_interface': 4.481535516374368e-06, 'learning_rate/action_model': 4.4815355163743674e-05, 'epoch': 0.16}) +[2026-06-24 10:37:25,422][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.007651751395314932, 'train/grad_norm_pre_clip': 6.767481803894043, 'timing/data': 0.0003038800787180662, 'timing/model': 1.0624060269910842, 'learning_rate/qwen_vl_interface': 4.476458043102916e-06, 'learning_rate/action_model': 4.4764580431029155e-05, 'epoch': 0.16}) +[2026-06-24 10:37:26,489][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.5148422718048096, 'train/grad_norm_pre_clip': 135.2701873779297, 'timing/data': 0.00033129192888736725, 'timing/model': 1.0653794270474464, 'learning_rate/qwen_vl_interface': 4.471381477626035e-06, 'learning_rate/action_model': 4.4713814776260346e-05, 'epoch': 0.16}) +[2026-06-24 10:37:27,553][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.15567436814308167, 'train/grad_norm_pre_clip': 86.80228424072266, 'timing/data': 0.00033167190849781036, 'timing/model': 1.0624011710751802, 'learning_rate/qwen_vl_interface': 4.466305825901361e-06, 'learning_rate/action_model': 4.466305825901361e-05, 'epoch': 0.16}) +[2026-06-24 10:37:28,621][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.04554888978600502, 'train/grad_norm_pre_clip': 17.881696701049805, 'timing/data': 0.00031891907565295696, 'timing/model': 1.0665767348837107, 'learning_rate/qwen_vl_interface': 4.46123109388545e-06, 'learning_rate/action_model': 4.4612310938854504e-05, 'epoch': 0.16}) +[2026-06-24 10:37:29,688][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.25757649540901184, 'train/grad_norm_pre_clip': 66.82781219482422, 'timing/data': 0.0004874039441347122, 'timing/model': 1.0647028821986169, 'learning_rate/qwen_vl_interface': 4.456157287533787e-06, 'learning_rate/action_model': 4.456157287533786e-05, 'epoch': 0.16}) +[2026-06-24 10:37:30,753][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.003906203666701913, 'train/grad_norm_pre_clip': 1.8351521492004395, 'timing/data': 0.0003023569006472826, 'timing/model': 1.0634753198828548, 'learning_rate/qwen_vl_interface': 4.451084412800763e-06, 'learning_rate/action_model': 4.451084412800763e-05, 'epoch': 0.17}) +[2026-06-24 10:37:32,058][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.060409050434827805, 'train/grad_norm_pre_clip': 32.82252883911133, 'timing/data': 0.0003520010504871607, 'timing/model': 1.303065022919327, 'learning_rate/qwen_vl_interface': 4.44601247563968e-06, 'learning_rate/action_model': 4.44601247563968e-05, 'epoch': 0.17}) +[2026-06-24 10:37:33,119][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.0465560182929039, 'train/grad_norm_pre_clip': 45.41108703613281, 'timing/data': 0.00034049898386001587, 'timing/model': 1.0596286768559366, 'learning_rate/qwen_vl_interface': 4.440941482002739e-06, 'learning_rate/action_model': 4.440941482002739e-05, 'epoch': 0.17}) +[2026-06-24 10:37:34,183][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.025246096774935722, 'train/grad_norm_pre_clip': 39.800785064697266, 'timing/data': 0.0003034390974789858, 'timing/model': 1.0621746650431305, 'learning_rate/qwen_vl_interface': 4.435871437841032e-06, 'learning_rate/action_model': 4.435871437841032e-05, 'epoch': 0.17}) +[2026-06-24 10:37:35,247][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.1850883960723877, 'train/grad_norm_pre_clip': 58.08030700683594, 'timing/data': 0.0004929641727358103, 'timing/model': 1.0627534920349717, 'learning_rate/qwen_vl_interface': 4.430802349104542e-06, 'learning_rate/action_model': 4.4308023491045416e-05, 'epoch': 0.17}) +[2026-06-24 10:37:36,312][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.1037026047706604, 'train/grad_norm_pre_clip': 42.165706634521484, 'timing/data': 0.00033223279751837254, 'timing/model': 1.0630212659016252, 'learning_rate/qwen_vl_interface': 4.425734221742121e-06, 'learning_rate/action_model': 4.425734221742121e-05, 'epoch': 0.17}) +[2026-06-24 10:37:37,375][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.15108229219913483, 'train/grad_norm_pre_clip': 107.033935546875, 'timing/data': 0.0003023468889296055, 'timing/model': 1.062058812007308, 'learning_rate/qwen_vl_interface': 4.420667061701505e-06, 'learning_rate/action_model': 4.420667061701505e-05, 'epoch': 0.17}) +[2026-06-24 10:37:38,439][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.009964847937226295, 'train/grad_norm_pre_clip': 7.957296848297119, 'timing/data': 0.000376635929569602, 'timing/model': 1.061556079890579, 'learning_rate/qwen_vl_interface': 4.415600874929285e-06, 'learning_rate/action_model': 4.415600874929285e-05, 'epoch': 0.17}) +[2026-06-24 10:37:39,503][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.08935854583978653, 'train/grad_norm_pre_clip': 19.676578521728516, 'timing/data': 0.00033812387846410275, 'timing/model': 1.0626070429570973, 'learning_rate/qwen_vl_interface': 4.410535667370915e-06, 'learning_rate/action_model': 4.410535667370915e-05, 'epoch': 0.17}) +[2026-06-24 10:37:40,563][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.12027943879365921, 'train/grad_norm_pre_clip': 58.231414794921875, 'timing/data': 0.00029142689891159534, 'timing/model': 1.0588233259040862, 'learning_rate/qwen_vl_interface': 4.405471444970698e-06, 'learning_rate/action_model': 4.405471444970698e-05, 'epoch': 0.17}) +[2026-06-24 10:37:41,629][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.09145999699831009, 'train/grad_norm_pre_clip': 32.49480438232422, 'timing/data': 0.00031201611272990704, 'timing/model': 1.063654180150479, 'learning_rate/qwen_vl_interface': 4.400408213671779e-06, 'learning_rate/action_model': 4.4004082136717784e-05, 'epoch': 0.17}) +[2026-06-24 10:37:42,694][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.04006091505289078, 'train/grad_norm_pre_clip': 21.377662658691406, 'timing/data': 0.00034983688965439796, 'timing/model': 1.0641228510066867, 'learning_rate/qwen_vl_interface': 4.395345979416143e-06, 'learning_rate/action_model': 4.3953459794161435e-05, 'epoch': 0.17}) +[2026-06-24 10:37:43,759][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.005818509496748447, 'train/grad_norm_pre_clip': 4.904073715209961, 'timing/data': 0.00030614505521953106, 'timing/model': 1.0628476398997009, 'learning_rate/qwen_vl_interface': 4.390284748144605e-06, 'learning_rate/action_model': 4.3902847481446044e-05, 'epoch': 0.17}) +[2026-06-24 10:37:44,824][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.04288838803768158, 'train/grad_norm_pre_clip': 31.421146392822266, 'timing/data': 0.0003085690550506115, 'timing/model': 1.0642382001969963, 'learning_rate/qwen_vl_interface': 4.3852245257968e-06, 'learning_rate/action_model': 4.385224525796801e-05, 'epoch': 0.17}) +[2026-06-24 10:37:45,889][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.04503683000802994, 'train/grad_norm_pre_clip': 50.79710006713867, 'timing/data': 0.0003063739277422428, 'timing/model': 1.0631628141272813, 'learning_rate/qwen_vl_interface': 4.380165318311183e-06, 'learning_rate/action_model': 4.380165318311183e-05, 'epoch': 0.17}) +[2026-06-24 10:37:46,954][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.05055835843086243, 'train/grad_norm_pre_clip': 51.761817932128906, 'timing/data': 0.00030417111702263355, 'timing/model': 1.0633535031229258, 'learning_rate/qwen_vl_interface': 4.375107131625016e-06, 'learning_rate/action_model': 4.3751071316250155e-05, 'epoch': 0.17}) +[2026-06-24 10:37:48,018][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.2170872986316681, 'train/grad_norm_pre_clip': 78.25141143798828, 'timing/data': 0.00033508893102407455, 'timing/model': 1.062362628057599, 'learning_rate/qwen_vl_interface': 4.370049971674359e-06, 'learning_rate/action_model': 4.370049971674359e-05, 'epoch': 0.17}) +[2026-06-24 10:37:49,084][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.14872725307941437, 'train/grad_norm_pre_clip': 55.50531768798828, 'timing/data': 0.00031206500716507435, 'timing/model': 1.0640007101465017, 'learning_rate/qwen_vl_interface': 4.364993844394074e-06, 'learning_rate/action_model': 4.364993844394074e-05, 'epoch': 0.17}) +[2026-06-24 10:37:50,341][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.009795068763196468, 'train/grad_norm_pre_clip': 12.887064933776855, 'timing/data': 0.00031266710720956326, 'timing/model': 1.255415173014626, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 0.17}) +[2026-06-24 10:37:51,407][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.061612341552972794, 'train/grad_norm_pre_clip': 49.89986038208008, 'timing/data': 0.0003231561277061701, 'timing/model': 1.0653161010704935, 'learning_rate/qwen_vl_interface': 4.3548847115779904e-06, 'learning_rate/action_model': 4.35488471157799e-05, 'epoch': 0.17}) +[2026-06-24 10:37:52,472][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.17271043360233307, 'train/grad_norm_pre_clip': 116.3988265991211, 'timing/data': 0.000302968081086874, 'timing/model': 1.06274714297615, 'learning_rate/qwen_vl_interface': 4.349831717905818e-06, 'learning_rate/action_model': 4.349831717905818e-05, 'epoch': 0.17}) +[2026-06-24 10:37:53,536][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.2699192762374878, 'train/grad_norm_pre_clip': 59.50873565673828, 'timing/data': 0.0003296178765594959, 'timing/model': 1.061906476970762, 'learning_rate/qwen_vl_interface': 4.344779780631266e-06, 'learning_rate/action_model': 4.344779780631266e-05, 'epoch': 0.17}) +[2026-06-24 10:37:54,602][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.09547985345125198, 'train/grad_norm_pre_clip': 58.41043472290039, 'timing/data': 0.00030659488402307034, 'timing/model': 1.064421305898577, 'learning_rate/qwen_vl_interface': 4.33972890568306e-06, 'learning_rate/action_model': 4.33972890568306e-05, 'epoch': 0.17}) +[2026-06-24 10:37:55,668][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.020589549094438553, 'train/grad_norm_pre_clip': 16.21543312072754, 'timing/data': 0.0002988011110574007, 'timing/model': 1.0648098660167307, 'learning_rate/qwen_vl_interface': 4.334679098988686e-06, 'learning_rate/action_model': 4.334679098988686e-05, 'epoch': 0.17}) +[2026-06-24 10:37:56,738][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.10856017470359802, 'train/grad_norm_pre_clip': 42.46643829345703, 'timing/data': 0.0003072360996156931, 'timing/model': 1.068115629022941, 'learning_rate/qwen_vl_interface': 4.329630366474373e-06, 'learning_rate/action_model': 4.329630366474373e-05, 'epoch': 0.17}) +[2026-06-24 10:37:57,804][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.3520137071609497, 'train/grad_norm_pre_clip': 207.3993682861328, 'timing/data': 0.00036189891397953033, 'timing/model': 1.0641280598938465, 'learning_rate/qwen_vl_interface': 4.324582714065088e-06, 'learning_rate/action_model': 4.3245827140650884e-05, 'epoch': 0.17}) +[2026-06-24 10:37:58,869][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.06440706551074982, 'train/grad_norm_pre_clip': 83.09198760986328, 'timing/data': 0.0003363711293786764, 'timing/model': 1.0639719481114298, 'learning_rate/qwen_vl_interface': 4.319536147684536e-06, 'learning_rate/action_model': 4.319536147684536e-05, 'epoch': 0.17}) +[2026-06-24 10:37:59,934][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.11941801011562347, 'train/grad_norm_pre_clip': 54.60977554321289, 'timing/data': 0.00031499098986387253, 'timing/model': 1.0628299980890006, 'learning_rate/qwen_vl_interface': 4.314490673255139e-06, 'learning_rate/action_model': 4.314490673255139e-05, 'epoch': 0.17}) +[2026-06-24 10:38:00,999][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.027475230395793915, 'train/grad_norm_pre_clip': 19.81058120727539, 'timing/data': 0.0003059140872210264, 'timing/model': 1.0638288729824126, 'learning_rate/qwen_vl_interface': 4.309446296698049e-06, 'learning_rate/action_model': 4.309446296698049e-05, 'epoch': 0.17}) +[2026-06-24 10:38:02,067][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.06429547071456909, 'train/grad_norm_pre_clip': 71.13680267333984, 'timing/data': 0.000294883968308568, 'timing/model': 1.066216462990269, 'learning_rate/qwen_vl_interface': 4.3044030239331155e-06, 'learning_rate/action_model': 4.304403023933115e-05, 'epoch': 0.17}) +[2026-06-24 10:38:03,132][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.1284477263689041, 'train/grad_norm_pre_clip': 82.45264434814453, 'timing/data': 0.0003681101370602846, 'timing/model': 1.0632315350230783, 'learning_rate/qwen_vl_interface': 4.299360860878903e-06, 'learning_rate/action_model': 4.299360860878903e-05, 'epoch': 0.17}) +[2026-06-24 10:38:04,197][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.3242359459400177, 'train/grad_norm_pre_clip': 86.87725830078125, 'timing/data': 0.00030874903313815594, 'timing/model': 1.0632461949717253, 'learning_rate/qwen_vl_interface': 4.294319813452671e-06, 'learning_rate/action_model': 4.294319813452671e-05, 'epoch': 0.17}) +[2026-06-24 10:38:05,263][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.03279872238636017, 'train/grad_norm_pre_clip': 11.807297706604004, 'timing/data': 0.00030107516795396805, 'timing/model': 1.0644953888840973, 'learning_rate/qwen_vl_interface': 4.289279887570369e-06, 'learning_rate/action_model': 4.289279887570369e-05, 'epoch': 0.17}) +[2026-06-24 10:38:06,328][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.17227961122989655, 'train/grad_norm_pre_clip': 137.7938995361328, 'timing/data': 0.000353784067556262, 'timing/model': 1.0640998780727386, 'learning_rate/qwen_vl_interface': 4.284241089146631e-06, 'learning_rate/action_model': 4.2842410891466306e-05, 'epoch': 0.17}) +[2026-06-24 10:38:07,588][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.04610581323504448, 'train/grad_norm_pre_clip': 15.241558074951172, 'timing/data': 0.00030715600587427616, 'timing/model': 1.2576864990405738, 'learning_rate/qwen_vl_interface': 4.279203424094768e-06, 'learning_rate/action_model': 4.2792034240947676e-05, 'epoch': 0.17}) +[2026-06-24 10:38:08,652][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.048531509935855865, 'train/grad_norm_pre_clip': 23.233875274658203, 'timing/data': 0.0003330849576741457, 'timing/model': 1.0626479981001467, 'learning_rate/qwen_vl_interface': 4.274166898326758e-06, 'learning_rate/action_model': 4.274166898326758e-05, 'epoch': 0.17}) +[2026-06-24 10:38:09,717][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.014961072243750095, 'train/grad_norm_pre_clip': 11.74860668182373, 'timing/data': 0.00039553199894726276, 'timing/model': 1.0640456709079444, 'learning_rate/qwen_vl_interface': 4.269131517753248e-06, 'learning_rate/action_model': 4.2691315177532485e-05, 'epoch': 0.17}) +[2026-06-24 10:38:10,782][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.0765167623758316, 'train/grad_norm_pre_clip': 55.104068756103516, 'timing/data': 0.00035263109020888805, 'timing/model': 1.0630337039474398, 'learning_rate/qwen_vl_interface': 4.2640972882835365e-06, 'learning_rate/action_model': 4.264097288283536e-05, 'epoch': 0.17}) +[2026-06-24 10:38:11,845][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.11477257311344147, 'train/grad_norm_pre_clip': 36.702022552490234, 'timing/data': 0.00033087003976106644, 'timing/model': 1.0616488659288734, 'learning_rate/qwen_vl_interface': 4.259064215825573e-06, 'learning_rate/action_model': 4.259064215825573e-05, 'epoch': 0.17}) +[2026-06-24 10:38:12,912][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.17588664591312408, 'train/grad_norm_pre_clip': 72.45182037353516, 'timing/data': 0.0003022169694304466, 'timing/model': 1.0648842179216444, 'learning_rate/qwen_vl_interface': 4.254032306285947e-06, 'learning_rate/action_model': 4.2540323062859466e-05, 'epoch': 0.17}) +[2026-06-24 10:38:13,980][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.08397093415260315, 'train/grad_norm_pre_clip': 31.25060272216797, 'timing/data': 0.00036051589995622635, 'timing/model': 1.0661396340001374, 'learning_rate/qwen_vl_interface': 4.249001565569883e-06, 'learning_rate/action_model': 4.2490015655698825e-05, 'epoch': 0.17}) +[2026-06-24 10:38:15,046][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.07307277619838715, 'train/grad_norm_pre_clip': 33.54564666748047, 'timing/data': 0.00033472804352641106, 'timing/model': 1.0646150859538466, 'learning_rate/qwen_vl_interface': 4.2439719995812375e-06, 'learning_rate/action_model': 4.243971999581238e-05, 'epoch': 0.17}) +[2026-06-24 10:38:16,110][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.2406986951828003, 'train/grad_norm_pre_clip': 78.79862976074219, 'timing/data': 0.0003221940714865923, 'timing/model': 1.062736292136833, 'learning_rate/qwen_vl_interface': 4.238943614222487e-06, 'learning_rate/action_model': 4.2389436142224865e-05, 'epoch': 0.17}) +[2026-06-24 10:38:17,173][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.17490163445472717, 'train/grad_norm_pre_clip': 110.86585235595703, 'timing/data': 0.0003163730725646019, 'timing/model': 1.0617732170503587, 'learning_rate/qwen_vl_interface': 4.23391641539472e-06, 'learning_rate/action_model': 4.2339164153947206e-05, 'epoch': 0.17}) +[2026-06-24 10:38:18,236][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1750 +[2026-06-24 10:38:42,979][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1750 +[2026-06-24 10:38:42,980][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1750 +[2026-06-24 10:39:14,212][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1750 +[2026-06-24 10:43:51,088][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1750 +[2026-06-24 10:43:51,089][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.020372580736875534, 'train/grad_norm_pre_clip': 16.561569213867188, 'eval/action_loss/seconds': 24.741631756071, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.07648404687643051, 'eval/latency_2/loss': 0.07648404687643051, 'eval/flappy/loss': 0.07648404687643051, 'eval/flappy/latency_2/loss': 0.07648404687643051, 'eval/flap_precision': 0.9149746192893401, 'eval/flap_recall': 0.9782903663500678, 'eval/flap_f1': 0.9455737704918032, 'eval/flap_rate_model': 0.21888888888888888, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9455737704918034, 'eval/latency_2/flap_precision': 0.9149746192893401, 'eval/latency_2/flap_recall': 0.9782903663500678, 'eval/latency_2/flap_f1': 0.9455737704918032, 'eval/latency_2/flap_rate_model': 0.21888888888888888, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9455737704918034, 'eval/action_classification/data_wait_seconds': 0.3132031559944153, 'eval/action_classification/predict_seconds': 30.828229904174805, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 217.38000000000935, 'rl_games_eval/mid_train/mean_length': 1770.0, 'rl_games_eval/mid_train/std_reward': 195.1522113633477, 'rl_games_eval/mid_train/std_length': 1572.0778606672125, 'rl_games_eval/mid_train/macro_mean_reward': 217.38000000000935, 'rl_games_eval/mid_train/macro_mean_length': 1770.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 217.38000000000935, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1770.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 195.1522113633477, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1572.0778606672125, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 217.38000000000935, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00030663516372442245, 'timing/model': 1.0614815619774163, 'learning_rate/qwen_vl_interface': 4.228890408997636e-06, 'learning_rate/action_model': 4.228890408997636e-05, 'epoch': 0.17}) +[2026-06-24 10:43:52,141][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.1027003824710846, 'train/grad_norm_pre_clip': 64.79270935058594, 'timing/data': 0.00033878604881465435, 'timing/model': 1.0502321140374988, 'learning_rate/qwen_vl_interface': 4.223865600929532e-06, 'learning_rate/action_model': 4.2238656009295325e-05, 'epoch': 0.17}) +[2026-06-24 10:43:53,415][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.3324066698551178, 'train/grad_norm_pre_clip': 135.4431610107422, 'timing/data': 0.0005275597795844078, 'timing/model': 1.2719013730529696, 'learning_rate/qwen_vl_interface': 4.218841997087301e-06, 'learning_rate/action_model': 4.218841997087301e-05, 'epoch': 0.17}) +[2026-06-24 10:43:54,465][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.10396780073642731, 'train/grad_norm_pre_clip': 37.987789154052734, 'timing/data': 0.0005589991342276335, 'timing/model': 1.0485773598775268, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 0.17}) +[2026-06-24 10:43:55,518][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.028649169951677322, 'train/grad_norm_pre_clip': 24.045379638671875, 'timing/data': 0.0003125169314444065, 'timing/model': 1.0511131489183754, 'learning_rate/qwen_vl_interface': 4.208798425660956e-06, 'learning_rate/action_model': 4.208798425660956e-05, 'epoch': 0.17}) +[2026-06-24 10:43:56,569][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.07566063106060028, 'train/grad_norm_pre_clip': 31.369394302368164, 'timing/data': 0.00030623399652540684, 'timing/model': 1.049252507975325, 'learning_rate/qwen_vl_interface': 4.203778469863529e-06, 'learning_rate/action_model': 4.203778469863529e-05, 'epoch': 0.17}) +[2026-06-24 10:43:57,621][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.07775608450174332, 'train/grad_norm_pre_clip': 32.54910659790039, 'timing/data': 0.0003148710820823908, 'timing/model': 1.050770718138665, 'learning_rate/qwen_vl_interface': 4.198759741865344e-06, 'learning_rate/action_model': 4.198759741865344e-05, 'epoch': 0.17}) +[2026-06-24 10:43:58,673][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.16061192750930786, 'train/grad_norm_pre_clip': 99.76221466064453, 'timing/data': 0.0003974861465394497, 'timing/model': 1.0500459379982203, 'learning_rate/qwen_vl_interface': 4.193742247556154e-06, 'learning_rate/action_model': 4.193742247556154e-05, 'epoch': 0.17}) +[2026-06-24 10:43:59,727][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.12100250273942947, 'train/grad_norm_pre_clip': 39.02046203613281, 'timing/data': 0.0003307999577373266, 'timing/model': 1.0522609951440245, 'learning_rate/qwen_vl_interface': 4.188725992824271e-06, 'learning_rate/action_model': 4.1887259928242715e-05, 'epoch': 0.17}) +[2026-06-24 10:44:00,779][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.02096780389547348, 'train/grad_norm_pre_clip': 14.173660278320312, 'timing/data': 0.0004058710765093565, 'timing/model': 1.0506337829865515, 'learning_rate/qwen_vl_interface': 4.183710983556547e-06, 'learning_rate/action_model': 4.183710983556547e-05, 'epoch': 0.17}) +[2026-06-24 10:44:01,836][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.022424403578042984, 'train/grad_norm_pre_clip': 15.066811561584473, 'timing/data': 0.0003034390974789858, 'timing/model': 1.0552068140823394, 'learning_rate/qwen_vl_interface': 4.178697225638378e-06, 'learning_rate/action_model': 4.1786972256383776e-05, 'epoch': 0.17}) +[2026-06-24 10:44:02,892][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.10838331282138824, 'train/grad_norm_pre_clip': 56.418243408203125, 'timing/data': 0.0003047620411962271, 'timing/model': 1.0542129799723625, 'learning_rate/qwen_vl_interface': 4.173684724953684e-06, 'learning_rate/action_model': 4.173684724953684e-05, 'epoch': 0.17}) +[2026-06-24 10:44:03,948][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.00955566018819809, 'train/grad_norm_pre_clip': 7.620681285858154, 'timing/data': 0.0003136878367513418, 'timing/model': 1.0545097370631993, 'learning_rate/qwen_vl_interface': 4.1686734873849185e-06, 'learning_rate/action_model': 4.1686734873849185e-05, 'epoch': 0.17}) +[2026-06-24 10:44:05,004][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.053645290434360504, 'train/grad_norm_pre_clip': 61.25613021850586, 'timing/data': 0.00034188199788331985, 'timing/model': 1.054633589927107, 'learning_rate/qwen_vl_interface': 4.1636635188130445e-06, 'learning_rate/action_model': 4.163663518813044e-05, 'epoch': 0.17}) +[2026-06-24 10:44:06,060][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.05081667751073837, 'train/grad_norm_pre_clip': 45.445594787597656, 'timing/data': 0.00029783882200717926, 'timing/model': 1.0548238481860608, 'learning_rate/qwen_vl_interface': 4.158654825117539e-06, 'learning_rate/action_model': 4.1586548251175386e-05, 'epoch': 0.17}) +[2026-06-24 10:44:07,117][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.32131820917129517, 'train/grad_norm_pre_clip': 85.18431854248047, 'timing/data': 0.0003132680431008339, 'timing/model': 1.054474372882396, 'learning_rate/qwen_vl_interface': 4.153647412176384e-06, 'learning_rate/action_model': 4.1536474121763846e-05, 'epoch': 0.17}) +[2026-06-24 10:44:08,393][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.10900390893220901, 'train/grad_norm_pre_clip': 32.780670166015625, 'timing/data': 0.00031740590929985046, 'timing/model': 1.274985812138766, 'learning_rate/qwen_vl_interface': 4.1486412858660574e-06, 'learning_rate/action_model': 4.148641285866057e-05, 'epoch': 0.17}) +[2026-06-24 10:44:09,447][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.025172190740704536, 'train/grad_norm_pre_clip': 20.210554122924805, 'timing/data': 0.0005586880724877119, 'timing/model': 1.0521891969256103, 'learning_rate/qwen_vl_interface': 4.143636452061529e-06, 'learning_rate/action_model': 4.1436364520615286e-05, 'epoch': 0.17}) +[2026-06-24 10:44:10,506][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.13687264919281006, 'train/grad_norm_pre_clip': 48.58887481689453, 'timing/data': 0.00032648188062012196, 'timing/model': 1.0575252990238369, 'learning_rate/qwen_vl_interface': 4.1386329166362454e-06, 'learning_rate/action_model': 4.1386329166362454e-05, 'epoch': 0.17}) +[2026-06-24 10:44:11,563][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.10021684318780899, 'train/grad_norm_pre_clip': 43.1306266784668, 'timing/data': 0.00035390397533774376, 'timing/model': 1.0547500730026513, 'learning_rate/qwen_vl_interface': 4.133630685462139e-06, 'learning_rate/action_model': 4.1336306854621396e-05, 'epoch': 0.17}) +[2026-06-24 10:44:12,621][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.035733822733163834, 'train/grad_norm_pre_clip': 12.760353088378906, 'timing/data': 0.00032656313851475716, 'timing/model': 1.0570441980380565, 'learning_rate/qwen_vl_interface': 4.1286297644096045e-06, 'learning_rate/action_model': 4.128629764409605e-05, 'epoch': 0.17}) +[2026-06-24 10:44:13,679][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.15222524106502533, 'train/grad_norm_pre_clip': 72.17194366455078, 'timing/data': 0.0003079678863286972, 'timing/model': 1.056103543844074, 'learning_rate/qwen_vl_interface': 4.1236301593475025e-06, 'learning_rate/action_model': 4.123630159347502e-05, 'epoch': 0.17}) +[2026-06-24 10:44:14,736][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.10529504716396332, 'train/grad_norm_pre_clip': 52.16748809814453, 'timing/data': 0.00033846497535705566, 'timing/model': 1.0550254718400538, 'learning_rate/qwen_vl_interface': 4.1186318761431476e-06, 'learning_rate/action_model': 4.118631876143148e-05, 'epoch': 0.17}) +[2026-06-24 10:44:15,793][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.24822171032428741, 'train/grad_norm_pre_clip': 102.62837982177734, 'timing/data': 0.0003262918908149004, 'timing/model': 1.0551493039820343, 'learning_rate/qwen_vl_interface': 4.113634920662301e-06, 'learning_rate/action_model': 4.113634920662302e-05, 'epoch': 0.17}) +[2026-06-24 10:44:16,849][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.3112773299217224, 'train/grad_norm_pre_clip': 29.416526794433594, 'timing/data': 0.00033572991378605366, 'timing/model': 1.0554308239370584, 'learning_rate/qwen_vl_interface': 4.108639298769173e-06, 'learning_rate/action_model': 4.1086392987691725e-05, 'epoch': 0.17}) +[2026-06-24 10:44:17,911][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.19514136016368866, 'train/grad_norm_pre_clip': 64.24430847167969, 'timing/data': 0.00032777502201497555, 'timing/model': 1.0603177228476852, 'learning_rate/qwen_vl_interface': 4.103645016326399e-06, 'learning_rate/action_model': 4.103645016326399e-05, 'epoch': 0.17}) +[2026-06-24 10:44:18,971][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.0420834980905056, 'train/grad_norm_pre_clip': 23.351577758789062, 'timing/data': 0.0003298788797110319, 'timing/model': 1.058515323093161, 'learning_rate/qwen_vl_interface': 4.098652079195051e-06, 'learning_rate/action_model': 4.09865207919505e-05, 'epoch': 0.17}) +[2026-06-24 10:44:20,034][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.09128154069185257, 'train/grad_norm_pre_clip': 39.000308990478516, 'timing/data': 0.00034260284155607224, 'timing/model': 1.0605991431511939, 'learning_rate/qwen_vl_interface': 4.093660493234617e-06, 'learning_rate/action_model': 4.0936604932346175e-05, 'epoch': 0.17}) +[2026-06-24 10:44:21,093][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.0891030952334404, 'train/grad_norm_pre_clip': 28.27348518371582, 'timing/data': 0.00031148409470915794, 'timing/model': 1.0577896619215608, 'learning_rate/qwen_vl_interface': 4.088670264303004e-06, 'learning_rate/action_model': 4.088670264303004e-05, 'epoch': 0.17}) +[2026-06-24 10:44:22,151][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.14298813045024872, 'train/grad_norm_pre_clip': 59.82841110229492, 'timing/data': 0.0003129180986434221, 'timing/model': 1.0562792380806059, 'learning_rate/qwen_vl_interface': 4.083681398256521e-06, 'learning_rate/action_model': 4.08368139825652e-05, 'epoch': 0.17}) +[2026-06-24 10:44:23,432][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.04030216112732887, 'train/grad_norm_pre_clip': 28.144094467163086, 'timing/data': 0.00030179601162672043, 'timing/model': 1.279844179050997, 'learning_rate/qwen_vl_interface': 4.078693900949881e-06, 'learning_rate/action_model': 4.078693900949881e-05, 'epoch': 0.17}) +[2026-06-24 10:44:24,489][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.022892922163009644, 'train/grad_norm_pre_clip': 13.095035552978516, 'timing/data': 0.0003489251248538494, 'timing/model': 1.054317108122632, 'learning_rate/qwen_vl_interface': 4.0737077782361914e-06, 'learning_rate/action_model': 4.073707778236191e-05, 'epoch': 0.17}) +[2026-06-24 10:44:25,548][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.09023166447877884, 'train/grad_norm_pre_clip': 44.91581344604492, 'timing/data': 0.0003038800787180662, 'timing/model': 1.0575164719484746, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 0.17}) +[2026-06-24 10:44:26,605][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.015427508391439915, 'train/grad_norm_pre_clip': 9.221721649169922, 'timing/data': 0.0004770951345562935, 'timing/model': 1.0553082800470293, 'learning_rate/qwen_vl_interface': 4.063739679992005e-06, 'learning_rate/action_model': 4.0637396799920054e-05, 'epoch': 0.17}) +[2026-06-24 10:44:27,665][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.20702965557575226, 'train/grad_norm_pre_clip': 127.25581359863281, 'timing/data': 0.0005734860897064209, 'timing/model': 1.0578866279684007, 'learning_rate/qwen_vl_interface': 4.05875771615963e-06, 'learning_rate/action_model': 4.0587577161596296e-05, 'epoch': 0.17}) +[2026-06-24 10:44:28,727][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.0321415513753891, 'train/grad_norm_pre_clip': 17.724502563476562, 'timing/data': 0.0003088598605245352, 'timing/model': 1.0606845878064632, 'learning_rate/qwen_vl_interface': 4.0537771503164265e-06, 'learning_rate/action_model': 4.053777150316426e-05, 'epoch': 0.17}) +[2026-06-24 10:44:29,786][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.04084344208240509, 'train/grad_norm_pre_clip': 17.6455020904541, 'timing/data': 0.0004965921398252249, 'timing/model': 1.0571775981225073, 'learning_rate/qwen_vl_interface': 4.048797988307366e-06, 'learning_rate/action_model': 4.048797988307366e-05, 'epoch': 0.17}) +[2026-06-24 10:44:30,846][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.16043025255203247, 'train/grad_norm_pre_clip': 56.25712966918945, 'timing/data': 0.0004952088929712772, 'timing/model': 1.0579424239695072, 'learning_rate/qwen_vl_interface': 4.043820235975774e-06, 'learning_rate/action_model': 4.043820235975774e-05, 'epoch': 0.17}) +[2026-06-24 10:44:31,906][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.07819689810276031, 'train/grad_norm_pre_clip': 45.51469039916992, 'timing/data': 0.0003429029602557421, 'timing/model': 1.0586532889865339, 'learning_rate/qwen_vl_interface': 4.038843899163319e-06, 'learning_rate/action_model': 4.0388438991633186e-05, 'epoch': 0.17}) +[2026-06-24 10:44:32,965][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.012030684389173985, 'train/grad_norm_pre_clip': 20.922273635864258, 'timing/data': 0.00030299811623990536, 'timing/model': 1.0577661348506808, 'learning_rate/qwen_vl_interface': 4.0338689837100096e-06, 'learning_rate/action_model': 4.0338689837100094e-05, 'epoch': 0.17}) +[2026-06-24 10:44:34,026][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.10610310733318329, 'train/grad_norm_pre_clip': 44.904205322265625, 'timing/data': 0.0004714441020041704, 'timing/model': 1.0582313381601125, 'learning_rate/qwen_vl_interface': 4.028895495454185e-06, 'learning_rate/action_model': 4.028895495454185e-05, 'epoch': 0.17}) +[2026-06-24 10:44:35,086][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.19315205514431, 'train/grad_norm_pre_clip': 61.66024398803711, 'timing/data': 0.0003621489740908146, 'timing/model': 1.0588274979963899, 'learning_rate/qwen_vl_interface': 4.023923440232515e-06, 'learning_rate/action_model': 4.023923440232515e-05, 'epoch': 0.17}) +[2026-06-24 10:44:36,148][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.00996527448296547, 'train/grad_norm_pre_clip': 13.979503631591797, 'timing/data': 0.00034137116745114326, 'timing/model': 1.0603051099460572, 'learning_rate/qwen_vl_interface': 4.018952823879978e-06, 'learning_rate/action_model': 4.018952823879978e-05, 'epoch': 0.17}) +[2026-06-24 10:44:37,210][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.014572994783520699, 'train/grad_norm_pre_clip': 11.61945629119873, 'timing/data': 0.0003080880269408226, 'timing/model': 1.0601064078509808, 'learning_rate/qwen_vl_interface': 4.013983652229872e-06, 'learning_rate/action_model': 4.013983652229871e-05, 'epoch': 0.17}) +[2026-06-24 10:44:38,494][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.08575966209173203, 'train/grad_norm_pre_clip': 34.21345138549805, 'timing/data': 0.00036373198963701725, 'timing/model': 1.2823957880027592, 'learning_rate/qwen_vl_interface': 4.009015931113795e-06, 'learning_rate/action_model': 4.009015931113795e-05, 'epoch': 0.17}) +[2026-06-24 10:44:39,555][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.01973392255604267, 'train/grad_norm_pre_clip': 43.886497497558594, 'timing/data': 0.0003246490377932787, 'timing/model': 1.0595455979928374, 'learning_rate/qwen_vl_interface': 4.0040496663616455e-06, 'learning_rate/action_model': 4.004049666361645e-05, 'epoch': 0.17}) +[2026-06-24 10:44:40,617][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.27751871943473816, 'train/grad_norm_pre_clip': 125.95861053466797, 'timing/data': 0.0003333049826323986, 'timing/model': 1.0604883069172502, 'learning_rate/qwen_vl_interface': 3.999084863801613e-06, 'learning_rate/action_model': 3.9990848638016124e-05, 'epoch': 0.17}) +[2026-06-24 10:44:41,678][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.06595845520496368, 'train/grad_norm_pre_clip': 20.418563842773438, 'timing/data': 0.00034776213578879833, 'timing/model': 1.0592734280508012, 'learning_rate/qwen_vl_interface': 3.994121529260167e-06, 'learning_rate/action_model': 3.994121529260167e-05, 'epoch': 0.17}) +[2026-06-24 10:44:42,740][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.21949851512908936, 'train/grad_norm_pre_clip': 52.0519905090332, 'timing/data': 0.00033416692167520523, 'timing/model': 1.0602390109561384, 'learning_rate/qwen_vl_interface': 3.98915966856206e-06, 'learning_rate/action_model': 3.98915966856206e-05, 'epoch': 0.17}) +[2026-06-24 10:44:43,800][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.24368806183338165, 'train/grad_norm_pre_clip': 45.68073272705078, 'timing/data': 0.00044403295032680035, 'timing/model': 1.058845145860687, 'learning_rate/qwen_vl_interface': 3.984199287530309e-06, 'learning_rate/action_model': 3.984199287530309e-05, 'epoch': 0.17}) +[2026-06-24 10:44:44,862][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.10094169527292252, 'train/grad_norm_pre_clip': 35.62803268432617, 'timing/data': 0.00030805799178779125, 'timing/model': 1.0605384740047157, 'learning_rate/qwen_vl_interface': 3.979240391986202e-06, 'learning_rate/action_model': 3.979240391986202e-05, 'epoch': 0.17}) +[2026-06-24 10:44:45,923][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.03544822707772255, 'train/grad_norm_pre_clip': 22.87714195251465, 'timing/data': 0.0003212930168956518, 'timing/model': 1.05906281596981, 'learning_rate/qwen_vl_interface': 3.974282987749275e-06, 'learning_rate/action_model': 3.974282987749274e-05, 'epoch': 0.17}) +[2026-06-24 10:44:46,985][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.3654078543186188, 'train/grad_norm_pre_clip': 46.89168167114258, 'timing/data': 0.00032124295830726624, 'timing/model': 1.0599753889255226, 'learning_rate/qwen_vl_interface': 3.969327080637318e-06, 'learning_rate/action_model': 3.9693270806373177e-05, 'epoch': 0.17}) +[2026-06-24 10:44:48,047][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.045377276837825775, 'train/grad_norm_pre_clip': 13.742889404296875, 'timing/data': 0.0003593340516090393, 'timing/model': 1.0605383741203696, 'learning_rate/qwen_vl_interface': 3.964372676466366e-06, 'learning_rate/action_model': 3.964372676466366e-05, 'epoch': 0.17}) +[2026-06-24 10:44:49,108][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.3449226915836334, 'train/grad_norm_pre_clip': 85.00635528564453, 'timing/data': 0.00032807607203722, 'timing/model': 1.0598221579566598, 'learning_rate/qwen_vl_interface': 3.959419781050686e-06, 'learning_rate/action_model': 3.959419781050685e-05, 'epoch': 0.17}) +[2026-06-24 10:44:50,170][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.01174168474972248, 'train/grad_norm_pre_clip': 19.699819564819336, 'timing/data': 0.0003456389531493187, 'timing/model': 1.0597584031056613, 'learning_rate/qwen_vl_interface': 3.95446840020278e-06, 'learning_rate/action_model': 3.95446840020278e-05, 'epoch': 0.17}) +[2026-06-24 10:44:51,231][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.1303839087486267, 'train/grad_norm_pre_clip': 38.858245849609375, 'timing/data': 0.00033999793231487274, 'timing/model': 1.0594505658373237, 'learning_rate/qwen_vl_interface': 3.949518539733365e-06, 'learning_rate/action_model': 3.949518539733365e-05, 'epoch': 0.17}) +[2026-06-24 10:44:52,292][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.209016352891922, 'train/grad_norm_pre_clip': 34.10247802734375, 'timing/data': 0.0003394070081412792, 'timing/model': 1.0593388469424099, 'learning_rate/qwen_vl_interface': 3.944570205451383e-06, 'learning_rate/action_model': 3.944570205451382e-05, 'epoch': 0.17}) +[2026-06-24 10:44:53,575][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.012639927677810192, 'train/grad_norm_pre_clip': 9.763738632202148, 'timing/data': 0.00029922183603048325, 'timing/model': 1.2818849391769618, 'learning_rate/qwen_vl_interface': 3.939623403163974e-06, 'learning_rate/action_model': 3.9396234031639744e-05, 'epoch': 0.17}) +[2026-06-24 10:44:54,638][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.0749654546380043, 'train/grad_norm_pre_clip': 24.71636962890625, 'timing/data': 0.0003214532043784857, 'timing/model': 1.0615246000234038, 'learning_rate/qwen_vl_interface': 3.934678138676493e-06, 'learning_rate/action_model': 3.934678138676493e-05, 'epoch': 0.18}) +[2026-06-24 10:44:55,703][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.22570477426052094, 'train/grad_norm_pre_clip': 110.31635284423828, 'timing/data': 0.0005098669789731503, 'timing/model': 1.0623897439800203, 'learning_rate/qwen_vl_interface': 3.929734417792478e-06, 'learning_rate/action_model': 3.929734417792478e-05, 'epoch': 0.18}) +[2026-06-24 10:44:56,766][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.05426623672246933, 'train/grad_norm_pre_clip': 37.55491638183594, 'timing/data': 0.00031878799200057983, 'timing/model': 1.0611422329675406, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 0.18}) +[2026-06-24 10:44:57,844][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.08204679936170578, 'train/grad_norm_pre_clip': 36.65570068359375, 'timing/data': 0.00031922897323966026, 'timing/model': 1.0611298198346049, 'learning_rate/qwen_vl_interface': 3.919851630039959e-06, 'learning_rate/action_model': 3.919851630039959e-05, 'epoch': 0.18}) +[2026-06-24 10:44:58,912][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.206660196185112, 'train/grad_norm_pre_clip': 74.35578918457031, 'timing/data': 0.00038138497620821, 'timing/model': 1.0659032170660794, 'learning_rate/qwen_vl_interface': 3.914912574769456e-06, 'learning_rate/action_model': 3.914912574769456e-05, 'epoch': 0.18}) +[2026-06-24 10:44:59,972][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.19212420284748077, 'train/grad_norm_pre_clip': 101.22525787353516, 'timing/data': 0.00029947212897241116, 'timing/model': 1.0588413178920746, 'learning_rate/qwen_vl_interface': 3.9099750862984126e-06, 'learning_rate/action_model': 3.909975086298413e-05, 'epoch': 0.18}) +[2026-06-24 10:45:01,035][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.18625591695308685, 'train/grad_norm_pre_clip': 54.444488525390625, 'timing/data': 0.00030395016074180603, 'timing/model': 1.0610767230391502, 'learning_rate/qwen_vl_interface': 3.905039170421243e-06, 'learning_rate/action_model': 3.905039170421243e-05, 'epoch': 0.18}) +[2026-06-24 10:45:02,100][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.14033684134483337, 'train/grad_norm_pre_clip': 91.49253845214844, 'timing/data': 0.0005311570130288601, 'timing/model': 1.0635633391793817, 'learning_rate/qwen_vl_interface': 3.9001048329305215e-06, 'learning_rate/action_model': 3.900104832930522e-05, 'epoch': 0.18}) +[2026-06-24 10:45:03,164][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.17458096146583557, 'train/grad_norm_pre_clip': 44.03086853027344, 'timing/data': 0.0003535731229931116, 'timing/model': 1.0624503418803215, 'learning_rate/qwen_vl_interface': 3.895172079616968e-06, 'learning_rate/action_model': 3.895172079616968e-05, 'epoch': 0.18}) +[2026-06-24 10:45:04,229][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.0954229012131691, 'train/grad_norm_pre_clip': 93.09864044189453, 'timing/data': 0.00042191194370388985, 'timing/model': 1.0625983201898634, 'learning_rate/qwen_vl_interface': 3.890240916269442e-06, 'learning_rate/action_model': 3.890240916269442e-05, 'epoch': 0.18}) +[2026-06-24 10:45:05,294][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.29527485370635986, 'train/grad_norm_pre_clip': 73.22187042236328, 'timing/data': 0.0003347380552440882, 'timing/model': 1.063359227962792, 'learning_rate/qwen_vl_interface': 3.885311348674942e-06, 'learning_rate/action_model': 3.885311348674942e-05, 'epoch': 0.18}) +[2026-06-24 10:45:06,357][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.011383065022528172, 'train/grad_norm_pre_clip': 8.020092010498047, 'timing/data': 0.00040714419446885586, 'timing/model': 1.0619653929024935, 'learning_rate/qwen_vl_interface': 3.880383382618584e-06, 'learning_rate/action_model': 3.880383382618584e-05, 'epoch': 0.18}) +[2026-06-24 10:45:07,423][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.06480708718299866, 'train/grad_norm_pre_clip': 45.67051696777344, 'timing/data': 0.0002970870118588209, 'timing/model': 1.0640690010040998, 'learning_rate/qwen_vl_interface': 3.8754570238836155e-06, 'learning_rate/action_model': 3.875457023883616e-05, 'epoch': 0.18}) +[2026-06-24 10:45:08,709][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.0495188906788826, 'train/grad_norm_pre_clip': 23.887439727783203, 'timing/data': 0.00034576887264847755, 'timing/model': 1.2843081660103053, 'learning_rate/qwen_vl_interface': 3.870532278251391e-06, 'learning_rate/action_model': 3.870532278251391e-05, 'epoch': 0.18}) +[2026-06-24 10:45:09,772][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.04217761382460594, 'train/grad_norm_pre_clip': 21.47892951965332, 'timing/data': 0.0003464000765234232, 'timing/model': 1.0612679871264845, 'learning_rate/qwen_vl_interface': 3.865609151501375e-06, 'learning_rate/action_model': 3.8656091515013745e-05, 'epoch': 0.18}) +[2026-06-24 10:45:10,836][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.05202949047088623, 'train/grad_norm_pre_clip': 23.622098922729492, 'timing/data': 0.0003477528225630522, 'timing/model': 1.0628940579481423, 'learning_rate/qwen_vl_interface': 3.8606876494111275e-06, 'learning_rate/action_model': 3.8606876494111275e-05, 'epoch': 0.18}) +[2026-06-24 10:45:11,929][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.13657160103321075, 'train/grad_norm_pre_clip': 54.38323974609375, 'timing/data': 0.00031381798908114433, 'timing/model': 1.0909324779640883, 'learning_rate/qwen_vl_interface': 3.8557677777563105e-06, 'learning_rate/action_model': 3.85576777775631e-05, 'epoch': 0.18}) +[2026-06-24 10:45:12,996][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.1244361475110054, 'train/grad_norm_pre_clip': 31.8083553314209, 'timing/data': 0.0003444068133831024, 'timing/model': 1.065773932961747, 'learning_rate/qwen_vl_interface': 3.850849542310663e-06, 'learning_rate/action_model': 3.850849542310663e-05, 'epoch': 0.18}) +[2026-06-24 10:45:14,074][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.07200095802545547, 'train/grad_norm_pre_clip': 26.84478759765625, 'timing/data': 0.000335638877004385, 'timing/model': 1.0759361609816551, 'learning_rate/qwen_vl_interface': 3.845932948846012e-06, 'learning_rate/action_model': 3.8459329488460115e-05, 'epoch': 0.18}) +[2026-06-24 10:45:15,186][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.004338312894105911, 'train/grad_norm_pre_clip': 1.7547149658203125, 'timing/data': 0.00030654482543468475, 'timing/model': 1.110507139004767, 'learning_rate/qwen_vl_interface': 3.841018003132252e-06, 'learning_rate/action_model': 3.841018003132252e-05, 'epoch': 0.18}) +[2026-06-24 10:45:16,304][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.09168988466262817, 'train/grad_norm_pre_clip': 40.60872268676758, 'timing/data': 0.00031382893212139606, 'timing/model': 1.1160631550010294, 'learning_rate/qwen_vl_interface': 3.8361047109373435e-06, 'learning_rate/action_model': 3.836104710937344e-05, 'epoch': 0.18}) +[2026-06-24 10:45:17,374][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.3146628737449646, 'train/grad_norm_pre_clip': 73.15747833251953, 'timing/data': 0.00031323707662522793, 'timing/model': 1.0689092378597707, 'learning_rate/qwen_vl_interface': 3.831193078027311e-06, 'learning_rate/action_model': 3.8311930780273114e-05, 'epoch': 0.18}) +[2026-06-24 10:45:18,493][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.007765069603919983, 'train/grad_norm_pre_clip': 6.0805487632751465, 'timing/data': 0.0002916669473052025, 'timing/model': 1.1176973530091345, 'learning_rate/qwen_vl_interface': 3.826283110166229e-06, 'learning_rate/action_model': 3.826283110166229e-05, 'epoch': 0.18}) +[2026-06-24 10:45:19,648][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.07275785505771637, 'train/grad_norm_pre_clip': 60.21237564086914, 'timing/data': 0.0003018260467797518, 'timing/model': 1.1530939592048526, 'learning_rate/qwen_vl_interface': 3.821374813116218e-06, 'learning_rate/action_model': 3.8213748131162186e-05, 'epoch': 0.18}) +[2026-06-24 10:45:20,791][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.034814272075891495, 'train/grad_norm_pre_clip': 14.784416198730469, 'timing/data': 0.00032392796128988266, 'timing/model': 1.141153915785253, 'learning_rate/qwen_vl_interface': 3.8164681926374385e-06, 'learning_rate/action_model': 3.8164681926374387e-05, 'epoch': 0.18}) +[2026-06-24 10:45:21,856][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.012072782032191753, 'train/grad_norm_pre_clip': 8.626731872558594, 'timing/data': 0.00046192598529160023, 'timing/model': 1.0629398098681122, 'learning_rate/qwen_vl_interface': 3.811563254488083e-06, 'learning_rate/action_model': 3.8115632544880824e-05, 'epoch': 0.18}) +[2026-06-24 10:45:22,921][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.09070931375026703, 'train/grad_norm_pre_clip': 43.945247650146484, 'timing/data': 0.0004938070196658373, 'timing/model': 1.06295000994578, 'learning_rate/qwen_vl_interface': 3.8066600044243674e-06, 'learning_rate/action_model': 3.8066600044243674e-05, 'epoch': 0.18}) +[2026-06-24 10:45:24,208][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.07107436656951904, 'train/grad_norm_pre_clip': 38.436073303222656, 'timing/data': 0.0003471910022199154, 'timing/model': 1.2855404790025204, 'learning_rate/qwen_vl_interface': 3.8017584482005315e-06, 'learning_rate/action_model': 3.8017584482005315e-05, 'epoch': 0.18}) +[2026-06-24 10:45:25,328][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.2919176518917084, 'train/grad_norm_pre_clip': 89.440673828125, 'timing/data': 0.0003442659508436918, 'timing/model': 1.118931232020259, 'learning_rate/qwen_vl_interface': 3.7968585915688223e-06, 'learning_rate/action_model': 3.796858591568822e-05, 'epoch': 0.18}) +[2026-06-24 10:45:26,405][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.050670966506004333, 'train/grad_norm_pre_clip': 34.866817474365234, 'timing/data': 0.00031168502755463123, 'timing/model': 1.075564440805465, 'learning_rate/qwen_vl_interface': 3.7919604402794984e-06, 'learning_rate/action_model': 3.7919604402794984e-05, 'epoch': 0.18}) +[2026-06-24 10:45:27,465][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.13358481228351593, 'train/grad_norm_pre_clip': 33.76124954223633, 'timing/data': 0.000325201079249382, 'timing/model': 1.0584118268452585, 'learning_rate/qwen_vl_interface': 3.787064000080809e-06, 'learning_rate/action_model': 3.787064000080809e-05, 'epoch': 0.18}) +[2026-06-24 10:45:28,531][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.02312181331217289, 'train/grad_norm_pre_clip': 20.846464157104492, 'timing/data': 0.00029940204694867134, 'timing/model': 1.0635639040265232, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.18}) +[2026-06-24 10:45:29,595][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.12677913904190063, 'train/grad_norm_pre_clip': 69.00684356689453, 'timing/data': 0.0003055729903280735, 'timing/model': 1.063344944966957, 'learning_rate/qwen_vl_interface': 3.7772762759383035e-06, 'learning_rate/action_model': 3.7772762759383036e-05, 'epoch': 0.18}) +[2026-06-24 10:45:30,661][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.048676300793886185, 'train/grad_norm_pre_clip': 48.77827835083008, 'timing/data': 0.0005025831051170826, 'timing/model': 1.063973082927987, 'learning_rate/qwen_vl_interface': 3.772385003480928e-06, 'learning_rate/action_model': 3.7723850034809276e-05, 'epoch': 0.18}) +[2026-06-24 10:45:31,769][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.021226057782769203, 'train/grad_norm_pre_clip': 31.850675582885742, 'timing/data': 0.0003731099423021078, 'timing/model': 1.105916173895821, 'learning_rate/qwen_vl_interface': 3.7674954650870533e-06, 'learning_rate/action_model': 3.7674954650870534e-05, 'epoch': 0.18}) +[2026-06-24 10:45:32,834][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.08293759077787399, 'train/grad_norm_pre_clip': 42.691959381103516, 'timing/data': 0.00029934197664260864, 'timing/model': 1.063209111103788, 'learning_rate/qwen_vl_interface': 3.762607666494825e-06, 'learning_rate/action_model': 3.762607666494825e-05, 'epoch': 0.18}) +[2026-06-24 10:45:33,903][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.11315641552209854, 'train/grad_norm_pre_clip': 108.20394897460938, 'timing/data': 0.000430587911978364, 'timing/model': 1.0671739210374653, 'learning_rate/qwen_vl_interface': 3.7577216134403487e-06, 'learning_rate/action_model': 3.757721613440349e-05, 'epoch': 0.18}) +[2026-06-24 10:45:34,972][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.07325679063796997, 'train/grad_norm_pre_clip': 43.35847473144531, 'timing/data': 0.0003640528302639723, 'timing/model': 1.0675668488256633, 'learning_rate/qwen_vl_interface': 3.7528373116576777e-06, 'learning_rate/action_model': 3.7528373116576776e-05, 'epoch': 0.18}) +[2026-06-24 10:45:36,038][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.2537544369697571, 'train/grad_norm_pre_clip': 94.68029022216797, 'timing/data': 0.000318658072501421, 'timing/model': 1.0644370869267732, 'learning_rate/qwen_vl_interface': 3.7479547668788153e-06, 'learning_rate/action_model': 3.747954766878815e-05, 'epoch': 0.18}) +[2026-06-24 10:45:37,102][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.022450152784585953, 'train/grad_norm_pre_clip': 15.111360549926758, 'timing/data': 0.0003622497897595167, 'timing/model': 1.0625338959507644, 'learning_rate/qwen_vl_interface': 3.7430739848336967e-06, 'learning_rate/action_model': 3.7430739848336965e-05, 'epoch': 0.18}) +[2026-06-24 10:45:38,166][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.19937224686145782, 'train/grad_norm_pre_clip': 54.678287506103516, 'timing/data': 0.00032905698753893375, 'timing/model': 1.0622636100742966, 'learning_rate/qwen_vl_interface': 3.7381949712501912e-06, 'learning_rate/action_model': 3.738194971250191e-05, 'epoch': 0.18}) +[2026-06-24 10:45:39,450][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.02473405748605728, 'train/grad_norm_pre_clip': 22.967416763305664, 'timing/data': 0.00030462094582617283, 'timing/model': 1.2820927700959146, 'learning_rate/qwen_vl_interface': 3.733317731854096e-06, 'learning_rate/action_model': 3.733317731854096e-05, 'epoch': 0.18}) +[2026-06-24 10:45:40,513][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.02242845669388771, 'train/grad_norm_pre_clip': 18.984704971313477, 'timing/data': 0.0003302099648863077, 'timing/model': 1.0618220518808812, 'learning_rate/qwen_vl_interface': 3.72844227236912e-06, 'learning_rate/action_model': 3.72844227236912e-05, 'epoch': 0.18}) +[2026-06-24 10:45:41,577][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.06840105354785919, 'train/grad_norm_pre_clip': 32.501190185546875, 'timing/data': 0.0003078368026763201, 'timing/model': 1.061703962041065, 'learning_rate/qwen_vl_interface': 3.723568598516889e-06, 'learning_rate/action_model': 3.723568598516889e-05, 'epoch': 0.18}) +[2026-06-24 10:45:42,640][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.019229089841246605, 'train/grad_norm_pre_clip': 31.984905242919922, 'timing/data': 0.00031104288063943386, 'timing/model': 1.0622335148509592, 'learning_rate/qwen_vl_interface': 3.7186967160169284e-06, 'learning_rate/action_model': 3.7186967160169286e-05, 'epoch': 0.18}) +[2026-06-24 10:45:43,703][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.02925940603017807, 'train/grad_norm_pre_clip': 26.14024543762207, 'timing/data': 0.0003099818713963032, 'timing/model': 1.0610352880321443, 'learning_rate/qwen_vl_interface': 3.713826630586666e-06, 'learning_rate/action_model': 3.713826630586666e-05, 'epoch': 0.18}) +[2026-06-24 10:45:44,765][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.028192171826958656, 'train/grad_norm_pre_clip': 11.70407485961914, 'timing/data': 0.00046572298742830753, 'timing/model': 1.0596674920525402, 'learning_rate/qwen_vl_interface': 3.708958347941417e-06, 'learning_rate/action_model': 3.708958347941417e-05, 'epoch': 0.18}) +[2026-06-24 10:45:45,829][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.0757003054022789, 'train/grad_norm_pre_clip': 82.97831726074219, 'timing/data': 0.0003138589672744274, 'timing/model': 1.0631262820679694, 'learning_rate/qwen_vl_interface': 3.7040918737943825e-06, 'learning_rate/action_model': 3.704091873794383e-05, 'epoch': 0.18}) +[2026-06-24 10:45:46,892][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.029528683051466942, 'train/grad_norm_pre_clip': 32.826377868652344, 'timing/data': 0.00032707303762435913, 'timing/model': 1.0604758190456778, 'learning_rate/qwen_vl_interface': 3.699227213856643e-06, 'learning_rate/action_model': 3.699227213856643e-05, 'epoch': 0.18}) +[2026-06-24 10:45:47,956][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.1679527908563614, 'train/grad_norm_pre_clip': 75.3900375366211, 'timing/data': 0.0003118750173598528, 'timing/model': 1.062396764056757, 'learning_rate/qwen_vl_interface': 3.6943643738371434e-06, 'learning_rate/action_model': 3.694364373837143e-05, 'epoch': 0.18}) +[2026-06-24 10:45:49,018][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.09361042082309723, 'train/grad_norm_pre_clip': 87.84434509277344, 'timing/data': 0.00032003014348447323, 'timing/model': 1.0604851280804724, 'learning_rate/qwen_vl_interface': 3.6895033594427e-06, 'learning_rate/action_model': 3.6895033594426996e-05, 'epoch': 0.18}) +[2026-06-24 10:45:50,079][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.30413010716438293, 'train/grad_norm_pre_clip': 94.41271209716797, 'timing/data': 0.0003190480638295412, 'timing/model': 1.059435669099912, 'learning_rate/qwen_vl_interface': 3.6846441763779834e-06, 'learning_rate/action_model': 3.684644176377983e-05, 'epoch': 0.18}) +[2026-06-24 10:45:51,143][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.13576437532901764, 'train/grad_norm_pre_clip': 63.45476150512695, 'timing/data': 0.0003174648154526949, 'timing/model': 1.0625282928813249, 'learning_rate/qwen_vl_interface': 3.679786830345517e-06, 'learning_rate/action_model': 3.6797868303455165e-05, 'epoch': 0.18}) +[2026-06-24 10:45:52,205][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.145217165350914, 'train/grad_norm_pre_clip': 93.73680877685547, 'timing/data': 0.000394290080294013, 'timing/model': 1.0604455249849707, 'learning_rate/qwen_vl_interface': 3.674931327045665e-06, 'learning_rate/action_model': 3.674931327045665e-05, 'epoch': 0.18}) +[2026-06-24 10:45:53,266][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.02939608134329319, 'train/grad_norm_pre_clip': 47.574546813964844, 'timing/data': 0.0002949440386146307, 'timing/model': 1.059437886113301, 'learning_rate/qwen_vl_interface': 3.6700776721766323e-06, 'learning_rate/action_model': 3.670077672176632e-05, 'epoch': 0.18}) +[2026-06-24 10:45:54,551][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.057131413370370865, 'train/grad_norm_pre_clip': 35.12373733520508, 'timing/data': 0.0003205819521099329, 'timing/model': 1.2832377071026713, 'learning_rate/qwen_vl_interface': 3.6652258714344525e-06, 'learning_rate/action_model': 3.6652258714344524e-05, 'epoch': 0.18}) +[2026-06-24 10:45:55,618][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.034585610032081604, 'train/grad_norm_pre_clip': 47.46293640136719, 'timing/data': 0.00031800614669919014, 'timing/model': 1.0649459920823574, 'learning_rate/qwen_vl_interface': 3.660375930512984e-06, 'learning_rate/action_model': 3.660375930512984e-05, 'epoch': 0.18}) +[2026-06-24 10:45:56,680][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.15762978792190552, 'train/grad_norm_pre_clip': 64.66803741455078, 'timing/data': 0.00032077194191515446, 'timing/model': 1.0609111310914159, 'learning_rate/qwen_vl_interface': 3.655527855103906e-06, 'learning_rate/action_model': 3.655527855103906e-05, 'epoch': 0.18}) +[2026-06-24 10:45:57,745][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.01916455291211605, 'train/grad_norm_pre_clip': 26.261877059936523, 'timing/data': 0.0005997051484882832, 'timing/model': 1.0625655062031, 'learning_rate/qwen_vl_interface': 3.6506816508966985e-06, 'learning_rate/action_model': 3.6506816508966985e-05, 'epoch': 0.18}) +[2026-06-24 10:45:58,807][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.23543599247932434, 'train/grad_norm_pre_clip': 82.64363098144531, 'timing/data': 0.000319348881021142, 'timing/model': 1.0600050629582256, 'learning_rate/qwen_vl_interface': 3.6458373235786563e-06, 'learning_rate/action_model': 3.6458373235786565e-05, 'epoch': 0.18}) +[2026-06-24 10:45:59,870][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.019717518240213394, 'train/grad_norm_pre_clip': 7.4823784828186035, 'timing/data': 0.0003176960162818432, 'timing/model': 1.0614823850337416, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 0.18}) +[2026-06-24 10:46:00,934][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.027447406202554703, 'train/grad_norm_pre_clip': 31.561038970947266, 'timing/data': 0.00030533294193446636, 'timing/model': 1.0625812869984657, 'learning_rate/qwen_vl_interface': 3.6361543223482077e-06, 'learning_rate/action_model': 3.636154322348207e-05, 'epoch': 0.18}) +[2026-06-24 10:46:01,999][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.07947880029678345, 'train/grad_norm_pre_clip': 70.58374786376953, 'timing/data': 0.0003210629802197218, 'timing/model': 1.063655513105914, 'learning_rate/qwen_vl_interface': 3.631315659799342e-06, 'learning_rate/action_model': 3.631315659799342e-05, 'epoch': 0.18}) +[2026-06-24 10:46:03,062][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.19667665660381317, 'train/grad_norm_pre_clip': 69.71633911132812, 'timing/data': 0.0004265410825610161, 'timing/model': 1.0613498180173337, 'learning_rate/qwen_vl_interface': 3.6264788968667112e-06, 'learning_rate/action_model': 3.6264788968667115e-05, 'epoch': 0.18}) +[2026-06-24 10:46:04,128][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.030667703598737717, 'train/grad_norm_pre_clip': 37.8153076171875, 'timing/data': 0.0002963568549603224, 'timing/model': 1.0642428160645068, 'learning_rate/qwen_vl_interface': 3.6216440392265232e-06, 'learning_rate/action_model': 3.621644039226523e-05, 'epoch': 0.18}) +[2026-06-24 10:46:05,191][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.11853034794330597, 'train/grad_norm_pre_clip': 152.992919921875, 'timing/data': 0.0003396070096641779, 'timing/model': 1.0619361279532313, 'learning_rate/qwen_vl_interface': 3.6168110925527566e-06, 'learning_rate/action_model': 3.616811092552756e-05, 'epoch': 0.18}) +[2026-06-24 10:46:06,253][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.10834372043609619, 'train/grad_norm_pre_clip': 46.495059967041016, 'timing/data': 0.0003001128789037466, 'timing/model': 1.0596506509464234, 'learning_rate/qwen_vl_interface': 3.6119800625171413e-06, 'learning_rate/action_model': 3.6119800625171416e-05, 'epoch': 0.18}) +[2026-06-24 10:46:07,317][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.06538934260606766, 'train/grad_norm_pre_clip': 34.988624572753906, 'timing/data': 0.0002968569751828909, 'timing/model': 1.062239760067314, 'learning_rate/qwen_vl_interface': 3.607150954789159e-06, 'learning_rate/action_model': 3.607150954789159e-05, 'epoch': 0.18}) +[2026-06-24 10:46:08,381][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.0006727741565555334, 'train/grad_norm_pre_clip': 0.4733603298664093, 'timing/data': 0.0003200802020728588, 'timing/model': 1.0625282318796962, 'learning_rate/qwen_vl_interface': 3.602323775036038e-06, 'learning_rate/action_model': 3.602323775036038e-05, 'epoch': 0.18}) +[2026-06-24 10:46:09,665][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.14462803304195404, 'train/grad_norm_pre_clip': 46.60011672973633, 'timing/data': 0.00033035012893378735, 'timing/model': 1.2824555989354849, 'learning_rate/qwen_vl_interface': 3.5974985289227416e-06, 'learning_rate/action_model': 3.5974985289227415e-05, 'epoch': 0.18}) +[2026-06-24 10:46:10,728][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.12678681313991547, 'train/grad_norm_pre_clip': 61.181583404541016, 'timing/data': 0.0003299592062830925, 'timing/model': 1.062131789047271, 'learning_rate/qwen_vl_interface': 3.5926752221119664e-06, 'learning_rate/action_model': 3.5926752221119664e-05, 'epoch': 0.18}) +[2026-06-24 10:46:11,793][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.24299436807632446, 'train/grad_norm_pre_clip': 97.82023620605469, 'timing/data': 0.0002957850228995085, 'timing/model': 1.063052216079086, 'learning_rate/qwen_vl_interface': 3.5878538602641298e-06, 'learning_rate/action_model': 3.58785386026413e-05, 'epoch': 0.18}) +[2026-06-24 10:46:12,857][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.0023931649047881365, 'train/grad_norm_pre_clip': 1.8149843215942383, 'timing/data': 0.0003137090243399143, 'timing/model': 1.0618690070696175, 'learning_rate/qwen_vl_interface': 3.5830344490373713e-06, 'learning_rate/action_model': 3.5830344490373714e-05, 'epoch': 0.18}) +[2026-06-24 10:46:13,919][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.07810230553150177, 'train/grad_norm_pre_clip': 79.23959350585938, 'timing/data': 0.00029995315708220005, 'timing/model': 1.060366537887603, 'learning_rate/qwen_vl_interface': 3.578216994087535e-06, 'learning_rate/action_model': 3.578216994087535e-05, 'epoch': 0.18}) +[2026-06-24 10:46:14,977][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.05599447339773178, 'train/grad_norm_pre_clip': 49.435020446777344, 'timing/data': 0.00031720614060759544, 'timing/model': 1.0567305460572243, 'learning_rate/qwen_vl_interface': 3.5734015010681745e-06, 'learning_rate/action_model': 3.5734015010681744e-05, 'epoch': 0.18}) +[2026-06-24 10:46:16,039][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.09595216810703278, 'train/grad_norm_pre_clip': 53.60296630859375, 'timing/data': 0.00032804603688418865, 'timing/model': 1.0604317209217697, 'learning_rate/qwen_vl_interface': 3.568587975630539e-06, 'learning_rate/action_model': 3.568587975630539e-05, 'epoch': 0.18}) +[2026-06-24 10:46:17,102][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.08419986814260483, 'train/grad_norm_pre_clip': 46.035118103027344, 'timing/data': 0.0003226750995963812, 'timing/model': 1.061997049022466, 'learning_rate/qwen_vl_interface': 3.56377642342357e-06, 'learning_rate/action_model': 3.56377642342357e-05, 'epoch': 0.18}) +[2026-06-24 10:46:18,166][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.09925086796283722, 'train/grad_norm_pre_clip': 38.8192253112793, 'timing/data': 0.00032498990185558796, 'timing/model': 1.0624219470191747, 'learning_rate/qwen_vl_interface': 3.5589668500938893e-06, 'learning_rate/action_model': 3.558966850093889e-05, 'epoch': 0.18}) +[2026-06-24 10:46:19,235][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.05943626910448074, 'train/grad_norm_pre_clip': 19.168601989746094, 'timing/data': 0.0003464601468294859, 'timing/model': 1.0664212319534272, 'learning_rate/qwen_vl_interface': 3.5541592612857996e-06, 'learning_rate/action_model': 3.5541592612857994e-05, 'epoch': 0.18}) +[2026-06-24 10:46:20,298][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.025957368314266205, 'train/grad_norm_pre_clip': 27.499065399169922, 'timing/data': 0.0002998129930347204, 'timing/model': 1.0615239329636097, 'learning_rate/qwen_vl_interface': 3.549353662641276e-06, 'learning_rate/action_model': 3.549353662641276e-05, 'epoch': 0.18}) +[2026-06-24 10:46:21,359][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.01698831282556057, 'train/grad_norm_pre_clip': 11.661100387573242, 'timing/data': 0.00031938892789185047, 'timing/model': 1.059702015016228, 'learning_rate/qwen_vl_interface': 3.5445500597999537e-06, 'learning_rate/action_model': 3.544550059799954e-05, 'epoch': 0.18}) +[2026-06-24 10:46:22,424][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.03203270211815834, 'train/grad_norm_pre_clip': 59.46170425415039, 'timing/data': 0.0003509081434458494, 'timing/model': 1.0632833850104362, 'learning_rate/qwen_vl_interface': 3.5397484583991317e-06, 'learning_rate/action_model': 3.5397484583991315e-05, 'epoch': 0.18}) +[2026-06-24 10:46:23,485][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.20007343590259552, 'train/grad_norm_pre_clip': 215.50234985351562, 'timing/data': 0.0003058440051972866, 'timing/model': 1.059939781902358, 'learning_rate/qwen_vl_interface': 3.5349488640737516e-06, 'learning_rate/action_model': 3.5349488640737515e-05, 'epoch': 0.18}) +[2026-06-24 10:46:24,768][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.24517859518527985, 'train/grad_norm_pre_clip': 52.175045013427734, 'timing/data': 0.00031515094451606274, 'timing/model': 1.2809630071278661, 'learning_rate/qwen_vl_interface': 3.5301512824564103e-06, 'learning_rate/action_model': 3.5301512824564104e-05, 'epoch': 0.18}) +[2026-06-24 10:46:25,831][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.028574666008353233, 'train/grad_norm_pre_clip': 40.002620697021484, 'timing/data': 0.00031830789521336555, 'timing/model': 1.0614412599243224, 'learning_rate/qwen_vl_interface': 3.5253557191773324e-06, 'learning_rate/action_model': 3.525355719177332e-05, 'epoch': 0.18}) +[2026-06-24 10:46:26,894][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.07506957650184631, 'train/grad_norm_pre_clip': 57.246795654296875, 'timing/data': 0.0003261419478803873, 'timing/model': 1.061635215068236, 'learning_rate/qwen_vl_interface': 3.5205621798643835e-06, 'learning_rate/action_model': 3.520562179864384e-05, 'epoch': 0.18}) +[2026-06-24 10:46:27,956][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.016470571979880333, 'train/grad_norm_pre_clip': 24.210561752319336, 'timing/data': 0.0002976988907903433, 'timing/model': 1.0607154199387878, 'learning_rate/qwen_vl_interface': 3.515770670143044e-06, 'learning_rate/action_model': 3.515770670143044e-05, 'epoch': 0.18}) +[2026-06-24 10:46:29,020][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.15036289393901825, 'train/grad_norm_pre_clip': 79.74601745605469, 'timing/data': 0.00030733714811503887, 'timing/model': 1.0622831920627505, 'learning_rate/qwen_vl_interface': 3.5109811956364186e-06, 'learning_rate/action_model': 3.510981195636418e-05, 'epoch': 0.18}) +[2026-06-24 10:46:30,082][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.11629871279001236, 'train/grad_norm_pre_clip': 44.78109359741211, 'timing/data': 0.00037122610956430435, 'timing/model': 1.0598158829379827, 'learning_rate/qwen_vl_interface': 3.5061937619652247e-06, 'learning_rate/action_model': 3.506193761965224e-05, 'epoch': 0.18}) +[2026-06-24 10:46:31,145][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.2742158770561218, 'train/grad_norm_pre_clip': 42.12858963012695, 'timing/data': 0.00032195402309298515, 'timing/model': 1.061612332938239, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 0.18}) +[2026-06-24 10:46:32,208][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.048008423298597336, 'train/grad_norm_pre_clip': 28.11330223083496, 'timing/data': 0.00031477981247007847, 'timing/model': 1.0616192969027907, 'learning_rate/qwen_vl_interface': 3.4966250396000047e-06, 'learning_rate/action_model': 3.496625039600005e-05, 'epoch': 0.18}) +[2026-06-24 10:46:33,270][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.1894904226064682, 'train/grad_norm_pre_clip': 58.29835891723633, 'timing/data': 0.00031500100158154964, 'timing/model': 1.060391193954274, 'learning_rate/qwen_vl_interface': 3.491843762135406e-06, 'learning_rate/action_model': 3.491843762135406e-05, 'epoch': 0.18}) +[2026-06-24 10:46:34,330][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.32762593030929565, 'train/grad_norm_pre_clip': 113.40181732177734, 'timing/data': 0.00030634389258921146, 'timing/model': 1.058727892814204, 'learning_rate/qwen_vl_interface': 3.487064547965084e-06, 'learning_rate/action_model': 3.487064547965084e-05, 'epoch': 0.18}) +[2026-06-24 10:46:35,394][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.03548693656921387, 'train/grad_norm_pre_clip': 35.70650100708008, 'timing/data': 0.00029474287293851376, 'timing/model': 1.0619910368695855, 'learning_rate/qwen_vl_interface': 3.482287402697709e-06, 'learning_rate/action_model': 3.482287402697709e-05, 'epoch': 0.18}) +[2026-06-24 10:46:36,460][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.0669129341840744, 'train/grad_norm_pre_clip': 25.63046646118164, 'timing/data': 0.00030665588565170765, 'timing/model': 1.0645096530206501, 'learning_rate/qwen_vl_interface': 3.4775123319395286e-06, 'learning_rate/action_model': 3.477512331939529e-05, 'epoch': 0.18}) +[2026-06-24 10:46:37,524][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.2205510139465332, 'train/grad_norm_pre_clip': 55.77603530883789, 'timing/data': 0.00033618113957345486, 'timing/model': 1.0627162579912692, 'learning_rate/qwen_vl_interface': 3.4727393412943555e-06, 'learning_rate/action_model': 3.4727393412943554e-05, 'epoch': 0.18}) +[2026-06-24 10:46:38,584][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.16340237855911255, 'train/grad_norm_pre_clip': 77.4804916381836, 'timing/data': 0.00030844914726912975, 'timing/model': 1.057778334012255, 'learning_rate/qwen_vl_interface': 3.4679684363635575e-06, 'learning_rate/action_model': 3.467968436363557e-05, 'epoch': 0.18}) +[2026-06-24 10:46:39,868][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.105735644698143, 'train/grad_norm_pre_clip': 55.56309127807617, 'timing/data': 0.0003806541208177805, 'timing/model': 1.2830087530892342, 'learning_rate/qwen_vl_interface': 3.4631996227460567e-06, 'learning_rate/action_model': 3.4631996227460566e-05, 'epoch': 0.18}) +[2026-06-24 10:46:40,934][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.052939239889383316, 'train/grad_norm_pre_clip': 35.564979553222656, 'timing/data': 0.0003165840171277523, 'timing/model': 1.063809199957177, 'learning_rate/qwen_vl_interface': 3.4584329060383244e-06, 'learning_rate/action_model': 3.4584329060383244e-05, 'epoch': 0.18}) +[2026-06-24 10:46:41,998][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.145575612783432, 'train/grad_norm_pre_clip': 119.74818420410156, 'timing/data': 0.00029394100420176983, 'timing/model': 1.0624985110480338, 'learning_rate/qwen_vl_interface': 3.453668291834367e-06, 'learning_rate/action_model': 3.453668291834367e-05, 'epoch': 0.18}) +[2026-06-24 10:46:43,060][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.15493309497833252, 'train/grad_norm_pre_clip': 60.50719451904297, 'timing/data': 0.0003076670691370964, 'timing/model': 1.0610831768717617, 'learning_rate/qwen_vl_interface': 3.448905785725724e-06, 'learning_rate/action_model': 3.448905785725724e-05, 'epoch': 0.18}) +[2026-06-24 10:46:44,127][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.06048028543591499, 'train/grad_norm_pre_clip': 21.18073844909668, 'timing/data': 0.00043448503129184246, 'timing/model': 1.0649707310367376, 'learning_rate/qwen_vl_interface': 3.444145393301463e-06, 'learning_rate/action_model': 3.444145393301463e-05, 'epoch': 0.18}) +[2026-06-24 10:46:45,191][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.05737359821796417, 'train/grad_norm_pre_clip': 45.09304428100586, 'timing/data': 0.0005338320042937994, 'timing/model': 1.0622071558609605, 'learning_rate/qwen_vl_interface': 3.4393871201481684e-06, 'learning_rate/action_model': 3.4393871201481684e-05, 'epoch': 0.18}) +[2026-06-24 10:46:46,256][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.11900622397661209, 'train/grad_norm_pre_clip': 48.529640197753906, 'timing/data': 0.00032376800663769245, 'timing/model': 1.0633952461648732, 'learning_rate/qwen_vl_interface': 3.4346309718499395e-06, 'learning_rate/action_model': 3.4346309718499395e-05, 'epoch': 0.18}) +[2026-06-24 10:46:47,317][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.027483513578772545, 'train/grad_norm_pre_clip': 13.834993362426758, 'timing/data': 0.0005319081246852875, 'timing/model': 1.058354409178719, 'learning_rate/qwen_vl_interface': 3.429876953988382e-06, 'learning_rate/action_model': 3.429876953988382e-05, 'epoch': 0.19}) +[2026-06-24 10:46:48,382][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.015006911009550095, 'train/grad_norm_pre_clip': 10.994254112243652, 'timing/data': 0.00031694513745605946, 'timing/model': 1.0636875058989972, 'learning_rate/qwen_vl_interface': 3.4251250721425988e-06, 'learning_rate/action_model': 3.425125072142599e-05, 'epoch': 0.19}) +[2026-06-24 10:46:49,448][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.28768518567085266, 'train/grad_norm_pre_clip': 45.348052978515625, 'timing/data': 0.0003496960271149874, 'timing/model': 1.063736047130078, 'learning_rate/qwen_vl_interface': 3.4203753318891875e-06, 'learning_rate/action_model': 3.4203753318891876e-05, 'epoch': 0.19}) +[2026-06-24 10:46:50,511][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.1246497854590416, 'train/grad_norm_pre_clip': 168.74957275390625, 'timing/data': 0.00037794001400470734, 'timing/model': 1.0619580510538071, 'learning_rate/qwen_vl_interface': 3.415627738802235e-06, 'learning_rate/action_model': 3.415627738802235e-05, 'epoch': 0.19}) +[2026-06-24 10:46:51,574][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.03580761328339577, 'train/grad_norm_pre_clip': 13.586731910705566, 'timing/data': 0.00032126298174262047, 'timing/model': 1.061057701939717, 'learning_rate/qwen_vl_interface': 3.4108822984533047e-06, 'learning_rate/action_model': 3.410882298453305e-05, 'epoch': 0.19}) +[2026-06-24 10:46:52,638][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.02134622633457184, 'train/grad_norm_pre_clip': 8.934124946594238, 'timing/data': 0.00030163582414388657, 'timing/model': 1.0630246340297163, 'learning_rate/qwen_vl_interface': 3.406139016411436e-06, 'learning_rate/action_model': 3.406139016411436e-05, 'epoch': 0.19}) +[2026-06-24 10:46:53,704][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.003585393773391843, 'train/grad_norm_pre_clip': 1.0871078968048096, 'timing/data': 0.0004018140025436878, 'timing/model': 1.0640360310208052, 'learning_rate/qwen_vl_interface': 3.401397898243134e-06, 'learning_rate/action_model': 3.4013978982431336e-05, 'epoch': 0.19}) +[2026-06-24 10:46:54,989][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.08470761775970459, 'train/grad_norm_pre_clip': 37.0756950378418, 'timing/data': 0.00030708592385053635, 'timing/model': 1.2829079229850322, 'learning_rate/qwen_vl_interface': 3.3966589495123657e-06, 'learning_rate/action_model': 3.396658949512366e-05, 'epoch': 0.19}) +[2026-06-24 10:46:56,053][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.2528788149356842, 'train/grad_norm_pre_clip': 20.15557098388672, 'timing/data': 0.0003345380537211895, 'timing/model': 1.0623770710080862, 'learning_rate/qwen_vl_interface': 3.3919221757805508e-06, 'learning_rate/action_model': 3.391922175780551e-05, 'epoch': 0.19}) +[2026-06-24 10:46:57,117][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.025606248527765274, 'train/grad_norm_pre_clip': 12.46045207977295, 'timing/data': 0.0003244290128350258, 'timing/model': 1.0626111100427806, 'learning_rate/qwen_vl_interface': 3.387187582606559e-06, 'learning_rate/action_model': 3.387187582606559e-05, 'epoch': 0.19}) +[2026-06-24 10:46:58,183][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.11211076378822327, 'train/grad_norm_pre_clip': 54.8765983581543, 'timing/data': 0.0002972169313579798, 'timing/model': 1.0645728609524667, 'learning_rate/qwen_vl_interface': 3.3824551755466966e-06, 'learning_rate/action_model': 3.3824551755466965e-05, 'epoch': 0.19}) +[2026-06-24 10:46:59,246][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.0240591149777174, 'train/grad_norm_pre_clip': 29.285629272460938, 'timing/data': 0.0003052428364753723, 'timing/model': 1.061280463822186, 'learning_rate/qwen_vl_interface': 3.377724960154707e-06, 'learning_rate/action_model': 3.3777249601547064e-05, 'epoch': 0.19}) +[2026-06-24 10:47:00,307][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.08105972409248352, 'train/grad_norm_pre_clip': 48.134769439697266, 'timing/data': 0.00031656399369239807, 'timing/model': 1.059944377047941, 'learning_rate/qwen_vl_interface': 3.372996941981763e-06, 'learning_rate/action_model': 3.372996941981762e-05, 'epoch': 0.19}) +[2026-06-24 10:47:01,371][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.1838918924331665, 'train/grad_norm_pre_clip': 56.97698974609375, 'timing/data': 0.00031957984901964664, 'timing/model': 1.0618836667854339, 'learning_rate/qwen_vl_interface': 3.368271126576455e-06, 'learning_rate/action_model': 3.368271126576455e-05, 'epoch': 0.19}) +[2026-06-24 10:47:02,435][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.16682592034339905, 'train/grad_norm_pre_clip': 54.97297668457031, 'timing/data': 0.0003074370324611664, 'timing/model': 1.0626133559271693, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 0.19}) +[2026-06-24 10:47:03,500][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.07739831507205963, 'train/grad_norm_pre_clip': 26.625694274902344, 'timing/data': 0.0003415909595787525, 'timing/model': 1.062704156851396, 'learning_rate/qwen_vl_interface': 3.3588261262501907e-06, 'learning_rate/action_model': 3.3588261262501905e-05, 'epoch': 0.19}) +[2026-06-24 10:47:04,563][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.010457034222781658, 'train/grad_norm_pre_clip': 7.36187219619751, 'timing/data': 0.0003115551080554724, 'timing/model': 1.062032838119194, 'learning_rate/qwen_vl_interface': 3.354106952413469e-06, 'learning_rate/action_model': 3.354106952413469e-05, 'epoch': 0.19}) +[2026-06-24 10:47:05,627][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.12794452905654907, 'train/grad_norm_pre_clip': 65.5124740600586, 'timing/data': 0.00035240105353295803, 'timing/model': 1.0622462178580463, 'learning_rate/qwen_vl_interface': 3.3493900035128385e-06, 'learning_rate/action_model': 3.349390003512838e-05, 'epoch': 0.19}) +[2026-06-24 10:47:06,694][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.029021531343460083, 'train/grad_norm_pre_clip': 27.0684814453125, 'timing/data': 0.0003125269431620836, 'timing/model': 1.0649623859208077, 'learning_rate/qwen_vl_interface': 3.3446752850839036e-06, 'learning_rate/action_model': 3.3446752850839034e-05, 'epoch': 0.19}) +[2026-06-24 10:47:07,756][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.18901321291923523, 'train/grad_norm_pre_clip': 90.58367919921875, 'timing/data': 0.00032749399542808533, 'timing/model': 1.0601634220220149, 'learning_rate/qwen_vl_interface': 3.339962802659649e-06, 'learning_rate/action_model': 3.339962802659649e-05, 'epoch': 0.19}) +[2026-06-24 10:47:08,818][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.0706569105386734, 'train/grad_norm_pre_clip': 52.369102478027344, 'timing/data': 0.00031369877979159355, 'timing/model': 1.0605814969167113, 'learning_rate/qwen_vl_interface': 3.3352525617704318e-06, 'learning_rate/action_model': 3.335252561770432e-05, 'epoch': 0.19}) +[2026-06-24 10:47:10,105][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.02249419502913952, 'train/grad_norm_pre_clip': 15.972115516662598, 'timing/data': 0.00031645409762859344, 'timing/model': 1.2852679688949138, 'learning_rate/qwen_vl_interface': 3.330544567943986e-06, 'learning_rate/action_model': 3.330544567943986e-05, 'epoch': 0.19}) +[2026-06-24 10:47:11,167][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.18746455013751984, 'train/grad_norm_pre_clip': 129.7462158203125, 'timing/data': 0.00032898690551519394, 'timing/model': 1.0612212878186256, 'learning_rate/qwen_vl_interface': 3.325838826705402e-06, 'learning_rate/action_model': 3.325838826705402e-05, 'epoch': 0.19}) +[2026-06-24 10:47:12,230][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.13582658767700195, 'train/grad_norm_pre_clip': 38.87746810913086, 'timing/data': 0.0003108331002295017, 'timing/model': 1.0612173210829496, 'learning_rate/qwen_vl_interface': 3.3211353435771317e-06, 'learning_rate/action_model': 3.3211353435771316e-05, 'epoch': 0.19}) +[2026-06-24 10:47:13,296][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.3250006139278412, 'train/grad_norm_pre_clip': 78.68087005615234, 'timing/data': 0.000310020986944437, 'timing/model': 1.0639207840431482, 'learning_rate/qwen_vl_interface': 3.3164341240789717e-06, 'learning_rate/action_model': 3.316434124078972e-05, 'epoch': 0.19}) +[2026-06-24 10:47:14,360][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.2645707428455353, 'train/grad_norm_pre_clip': 96.91354370117188, 'timing/data': 0.0003187779802829027, 'timing/model': 1.0624104209709913, 'learning_rate/qwen_vl_interface': 3.3117351737280684e-06, 'learning_rate/action_model': 3.311735173728068e-05, 'epoch': 0.19}) +[2026-06-24 10:47:15,427][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.06568236649036407, 'train/grad_norm_pre_clip': 61.368492126464844, 'timing/data': 0.0003466301131993532, 'timing/model': 1.065952837932855, 'learning_rate/qwen_vl_interface': 3.3070384980388974e-06, 'learning_rate/action_model': 3.307038498038897e-05, 'epoch': 0.19}) +[2026-06-24 10:47:16,491][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.22825223207473755, 'train/grad_norm_pre_clip': 142.8050537109375, 'timing/data': 0.0006213558372110128, 'timing/model': 1.0618959460407495, 'learning_rate/qwen_vl_interface': 3.3023441025232735e-06, 'learning_rate/action_model': 3.302344102523273e-05, 'epoch': 0.19}) +[2026-06-24 10:47:17,555][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.20498864352703094, 'train/grad_norm_pre_clip': 118.53739929199219, 'timing/data': 0.0003177060279995203, 'timing/model': 1.0625083658378571, 'learning_rate/qwen_vl_interface': 3.2976519926903307e-06, 'learning_rate/action_model': 3.2976519926903306e-05, 'epoch': 0.19}) +[2026-06-24 10:47:18,624][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.07800738513469696, 'train/grad_norm_pre_clip': 57.8193473815918, 'timing/data': 0.0005192949902266264, 'timing/model': 1.0663669970817864, 'learning_rate/qwen_vl_interface': 3.292962174046518e-06, 'learning_rate/action_model': 3.292962174046518e-05, 'epoch': 0.19}) +[2026-06-24 10:47:19,688][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.24858513474464417, 'train/grad_norm_pre_clip': 65.11175537109375, 'timing/data': 0.0005131131038069725, 'timing/model': 1.0622168888803571, 'learning_rate/qwen_vl_interface': 3.288274652095601e-06, 'learning_rate/action_model': 3.288274652095601e-05, 'epoch': 0.19}) +[2026-06-24 10:47:20,751][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.02084418199956417, 'train/grad_norm_pre_clip': 13.182761192321777, 'timing/data': 0.0003287468571215868, 'timing/model': 1.0614967800211161, 'learning_rate/qwen_vl_interface': 3.2835894323386476e-06, 'learning_rate/action_model': 3.283589432338647e-05, 'epoch': 0.19}) +[2026-06-24 10:47:21,816][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.022234154865145683, 'train/grad_norm_pre_clip': 21.436052322387695, 'timing/data': 0.0004926249384880066, 'timing/model': 1.0629206108860672, 'learning_rate/qwen_vl_interface': 3.2789065202740245e-06, 'learning_rate/action_model': 3.278906520274024e-05, 'epoch': 0.19}) +[2026-06-24 10:47:22,879][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.1156415194272995, 'train/grad_norm_pre_clip': 70.82011413574219, 'timing/data': 0.00030143512412905693, 'timing/model': 1.0623448120895773, 'learning_rate/qwen_vl_interface': 3.2742259213973887e-06, 'learning_rate/action_model': 3.2742259213973886e-05, 'epoch': 0.19}) +[2026-06-24 10:47:23,946][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.03594393655657768, 'train/grad_norm_pre_clip': 45.7010498046875, 'timing/data': 0.00032418803311884403, 'timing/model': 1.0653884429484606, 'learning_rate/qwen_vl_interface': 3.2695476412016846e-06, 'learning_rate/action_model': 3.269547641201684e-05, 'epoch': 0.19}) +[2026-06-24 10:47:25,231][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.05926927179098129, 'train/grad_norm_pre_clip': 72.29300689697266, 'timing/data': 0.00029412214644253254, 'timing/model': 1.282408651895821, 'learning_rate/qwen_vl_interface': 3.264871685177136e-06, 'learning_rate/action_model': 3.264871685177136e-05, 'epoch': 0.19}) +[2026-06-24 10:47:26,295][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.11719180643558502, 'train/grad_norm_pre_clip': 48.76115798950195, 'timing/data': 0.0005235618446022272, 'timing/model': 1.062837767181918, 'learning_rate/qwen_vl_interface': 3.260198058811235e-06, 'learning_rate/action_model': 3.2601980588112345e-05, 'epoch': 0.19}) +[2026-06-24 10:47:27,361][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.01652170717716217, 'train/grad_norm_pre_clip': 13.477235794067383, 'timing/data': 0.00032900809310376644, 'timing/model': 1.0641592480242252, 'learning_rate/qwen_vl_interface': 3.255526767588743e-06, 'learning_rate/action_model': 3.255526767588743e-05, 'epoch': 0.19}) +[2026-06-24 10:47:28,427][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.032660793513059616, 'train/grad_norm_pre_clip': 34.13328170776367, 'timing/data': 0.00032846699468791485, 'timing/model': 1.0644383518956602, 'learning_rate/qwen_vl_interface': 3.2508578169916833e-06, 'learning_rate/action_model': 3.250857816991683e-05, 'epoch': 0.19}) +[2026-06-24 10:47:29,488][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.2225835770368576, 'train/grad_norm_pre_clip': 75.71435546875, 'timing/data': 0.000390433007851243, 'timing/model': 1.0596191289369017, 'learning_rate/qwen_vl_interface': 3.246191212499325e-06, 'learning_rate/action_model': 3.246191212499325e-05, 'epoch': 0.19}) +[2026-06-24 10:47:30,554][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.10656369477510452, 'train/grad_norm_pre_clip': 33.97688674926758, 'timing/data': 0.0005318268667906523, 'timing/model': 1.063254470936954, 'learning_rate/qwen_vl_interface': 3.241526959588193e-06, 'learning_rate/action_model': 3.241526959588193e-05, 'epoch': 0.19}) +[2026-06-24 10:47:31,617][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.18503141403198242, 'train/grad_norm_pre_clip': 37.47451400756836, 'timing/data': 0.00031271716579794884, 'timing/model': 1.0617510410957038, 'learning_rate/qwen_vl_interface': 3.2368650637320443e-06, 'learning_rate/action_model': 3.236865063732044e-05, 'epoch': 0.19}) +[2026-06-24 10:47:32,682][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.10265008360147476, 'train/grad_norm_pre_clip': 56.96902847290039, 'timing/data': 0.0003059839364141226, 'timing/model': 1.063636698992923, 'learning_rate/qwen_vl_interface': 3.232205530401874e-06, 'learning_rate/action_model': 3.2322055304018736e-05, 'epoch': 0.19}) +[2026-06-24 10:47:33,746][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.040163032710552216, 'train/grad_norm_pre_clip': 15.46460247039795, 'timing/data': 0.00036498415283858776, 'timing/model': 1.0619583199732006, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 0.19}) +[2026-06-24 10:47:34,810][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.06512323021888733, 'train/grad_norm_pre_clip': 45.74525451660156, 'timing/data': 0.00034451717510819435, 'timing/model': 1.0629330398514867, 'learning_rate/qwen_vl_interface': 3.222893573189581e-06, 'learning_rate/action_model': 3.222893573189581e-05, 'epoch': 0.19}) +[2026-06-24 10:47:35,876][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.08953697234392166, 'train/grad_norm_pre_clip': 59.07868576049805, 'timing/data': 0.0003478829748928547, 'timing/model': 1.063679781043902, 'learning_rate/qwen_vl_interface': 3.2182411602355575e-06, 'learning_rate/action_model': 3.218241160235557e-05, 'epoch': 0.19}) +[2026-06-24 10:47:36,939][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.004335986915975809, 'train/grad_norm_pre_clip': 4.7425217628479, 'timing/data': 0.0003021170850843191, 'timing/model': 1.0620852699503303, 'learning_rate/qwen_vl_interface': 3.2135911316637004e-06, 'learning_rate/action_model': 3.2135911316637004e-05, 'epoch': 0.19}) +[2026-06-24 10:47:38,004][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.27669113874435425, 'train/grad_norm_pre_clip': 105.27654266357422, 'timing/data': 0.0003075969871133566, 'timing/model': 1.0630894149653614, 'learning_rate/qwen_vl_interface': 3.20894349293108e-06, 'learning_rate/action_model': 3.20894349293108e-05, 'epoch': 0.19}) +[2026-06-24 10:47:39,069][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.005650027189403772, 'train/grad_norm_pre_clip': 6.215041160583496, 'timing/data': 0.000307807931676507, 'timing/model': 1.0634822219144553, 'learning_rate/qwen_vl_interface': 3.2042982494919558e-06, 'learning_rate/action_model': 3.204298249491956e-05, 'epoch': 0.19}) +[2026-06-24 10:47:40,356][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.051012810319662094, 'train/grad_norm_pre_clip': 29.010997772216797, 'timing/data': 0.0003179870545864105, 'timing/model': 1.2852736248169094, 'learning_rate/qwen_vl_interface': 3.1996554067977825e-06, 'learning_rate/action_model': 3.199655406797782e-05, 'epoch': 0.19}) +[2026-06-24 10:47:41,422][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.11089171469211578, 'train/grad_norm_pre_clip': 61.181549072265625, 'timing/data': 0.0003268229775130749, 'timing/model': 1.0642753611318767, 'learning_rate/qwen_vl_interface': 3.1950149702971915e-06, 'learning_rate/action_model': 3.195014970297191e-05, 'epoch': 0.19}) +[2026-06-24 10:47:42,486][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.024953607469797134, 'train/grad_norm_pre_clip': 36.82100296020508, 'timing/data': 0.0003379541449248791, 'timing/model': 1.0625423190649599, 'learning_rate/qwen_vl_interface': 3.1903769454359955e-06, 'learning_rate/action_model': 3.190376945435996e-05, 'epoch': 0.19}) +[2026-06-24 10:47:43,549][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.017100892961025238, 'train/grad_norm_pre_clip': 17.959861755371094, 'timing/data': 0.0002942020073533058, 'timing/model': 1.0617608739994466, 'learning_rate/qwen_vl_interface': 3.1857413376571745e-06, 'learning_rate/action_model': 3.1857413376571746e-05, 'epoch': 0.19}) +[2026-06-24 10:47:44,612][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.0818537026643753, 'train/grad_norm_pre_clip': 64.5748291015625, 'timing/data': 0.0004893778823316097, 'timing/model': 1.0608215811662376, 'learning_rate/qwen_vl_interface': 3.1811081524008747e-06, 'learning_rate/action_model': 3.1811081524008746e-05, 'epoch': 0.19}) +[2026-06-24 10:47:45,678][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.013952155597507954, 'train/grad_norm_pre_clip': 9.745234489440918, 'timing/data': 0.0003174759913235903, 'timing/model': 1.064405628014356, 'learning_rate/qwen_vl_interface': 3.1764773951043937e-06, 'learning_rate/action_model': 3.1764773951043935e-05, 'epoch': 0.19}) +[2026-06-24 10:47:46,744][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.014830820262432098, 'train/grad_norm_pre_clip': 11.789921760559082, 'timing/data': 0.0003306001890450716, 'timing/model': 1.0643354170024395, 'learning_rate/qwen_vl_interface': 3.171849071202184e-06, 'learning_rate/action_model': 3.1718490712021835e-05, 'epoch': 0.19}) +[2026-06-24 10:47:47,810][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.14167658984661102, 'train/grad_norm_pre_clip': 24.746400833129883, 'timing/data': 0.00030299811623990536, 'timing/model': 1.063948780996725, 'learning_rate/qwen_vl_interface': 3.1672231861258437e-06, 'learning_rate/action_model': 3.1672231861258435e-05, 'epoch': 0.19}) +[2026-06-24 10:47:48,876][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.03540631756186485, 'train/grad_norm_pre_clip': 31.593727111816406, 'timing/data': 0.00043796212412416935, 'timing/model': 1.064357538940385, 'learning_rate/qwen_vl_interface': 3.162599745304103e-06, 'learning_rate/action_model': 3.162599745304103e-05, 'epoch': 0.19}) +[2026-06-24 10:47:49,940][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.12590773403644562, 'train/grad_norm_pre_clip': 57.65380096435547, 'timing/data': 0.00031993119046092033, 'timing/model': 1.0624721709173173, 'learning_rate/qwen_vl_interface': 3.157978754162831e-06, 'learning_rate/action_model': 3.157978754162831e-05, 'epoch': 0.19}) +[2026-06-24 10:47:51,005][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.02683897688984871, 'train/grad_norm_pre_clip': 16.70058250427246, 'timing/data': 0.00034321402199566364, 'timing/model': 1.063375957077369, 'learning_rate/qwen_vl_interface': 3.1533602181250145e-06, 'learning_rate/action_model': 3.153360218125014e-05, 'epoch': 0.19}) +[2026-06-24 10:47:52,069][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.0026054861955344677, 'train/grad_norm_pre_clip': 1.244189739227295, 'timing/data': 0.0003088400699198246, 'timing/model': 1.06254977802746, 'learning_rate/qwen_vl_interface': 3.148744142610764e-06, 'learning_rate/action_model': 3.148744142610764e-05, 'epoch': 0.19}) +[2026-06-24 10:47:53,134][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.07445340603590012, 'train/grad_norm_pre_clip': 64.42491912841797, 'timing/data': 0.00031999009661376476, 'timing/model': 1.063706167973578, 'learning_rate/qwen_vl_interface': 3.1441305330373027e-06, 'learning_rate/action_model': 3.1441305330373024e-05, 'epoch': 0.19}) +[2026-06-24 10:47:54,199][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.13832370936870575, 'train/grad_norm_pre_clip': 36.52976608276367, 'timing/data': 0.00030962005257606506, 'timing/model': 1.063460297184065, 'learning_rate/qwen_vl_interface': 3.139519394818959e-06, 'learning_rate/action_model': 3.139519394818959e-05, 'epoch': 0.19}) +[2026-06-24 10:47:55,483][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.022707924246788025, 'train/grad_norm_pre_clip': 26.9810848236084, 'timing/data': 0.0003301999531686306, 'timing/model': 1.2820755320135504, 'learning_rate/qwen_vl_interface': 3.1349107333671576e-06, 'learning_rate/action_model': 3.1349107333671574e-05, 'epoch': 0.19}) +[2026-06-24 10:47:56,548][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.15299111604690552, 'train/grad_norm_pre_clip': 85.69091796875, 'timing/data': 0.00045040505938231945, 'timing/model': 1.0639549258630723, 'learning_rate/qwen_vl_interface': 3.1303045540904205e-06, 'learning_rate/action_model': 3.13030455409042e-05, 'epoch': 0.19}) +[2026-06-24 10:47:57,616][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.16927474737167358, 'train/grad_norm_pre_clip': 111.66765594482422, 'timing/data': 0.00048501999117434025, 'timing/model': 1.065906337928027, 'learning_rate/qwen_vl_interface': 3.125700862394355e-06, 'learning_rate/action_model': 3.125700862394355e-05, 'epoch': 0.19}) +[2026-06-24 10:47:58,680][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.009515651501715183, 'train/grad_norm_pre_clip': 10.768146514892578, 'timing/data': 0.0003163039218634367, 'timing/model': 1.0625309951137751, 'learning_rate/qwen_vl_interface': 3.121099663681654e-06, 'learning_rate/action_model': 3.121099663681654e-05, 'epoch': 0.19}) +[2026-06-24 10:47:59,743][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.004695314448326826, 'train/grad_norm_pre_clip': 4.552763938903809, 'timing/data': 0.00033007003366947174, 'timing/model': 1.0611024349927902, 'learning_rate/qwen_vl_interface': 3.116500963352076e-06, 'learning_rate/action_model': 3.116500963352076e-05, 'epoch': 0.19}) +[2026-06-24 10:48:00,806][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.08840279281139374, 'train/grad_norm_pre_clip': 51.02489471435547, 'timing/data': 0.000306685920804739, 'timing/model': 1.0617861978244036, 'learning_rate/qwen_vl_interface': 3.1119047668024523e-06, 'learning_rate/action_model': 3.111904766802452e-05, 'epoch': 0.19}) +[2026-06-24 10:48:01,870][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.11339184641838074, 'train/grad_norm_pre_clip': 70.51978302001953, 'timing/data': 0.0003267331048846245, 'timing/model': 1.0618485151790082, 'learning_rate/qwen_vl_interface': 3.1073110794266746e-06, 'learning_rate/action_model': 3.1073110794266745e-05, 'epoch': 0.19}) +[2026-06-24 10:48:02,936][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.03202798217535019, 'train/grad_norm_pre_clip': 20.823272705078125, 'timing/data': 0.0003197400365024805, 'timing/model': 1.0644272149074823, 'learning_rate/qwen_vl_interface': 3.1027199066156942e-06, 'learning_rate/action_model': 3.102719906615694e-05, 'epoch': 0.19}) +[2026-06-24 10:48:04,001][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.05268750712275505, 'train/grad_norm_pre_clip': 41.934017181396484, 'timing/data': 0.00035313190892338753, 'timing/model': 1.06367814890109, 'learning_rate/qwen_vl_interface': 3.0981312537575086e-06, 'learning_rate/action_model': 3.0981312537575085e-05, 'epoch': 0.19}) +[2026-06-24 10:48:05,064][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.020755140110850334, 'train/grad_norm_pre_clip': 25.045259475708008, 'timing/data': 0.00033004884608089924, 'timing/model': 1.060814046068117, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 0.19}) +[2026-06-24 10:48:06,126][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.06883824616670609, 'train/grad_norm_pre_clip': 51.95769119262695, 'timing/data': 0.0003102619666606188, 'timing/model': 1.061222973978147, 'learning_rate/qwen_vl_interface': 3.0889615294367066e-06, 'learning_rate/action_model': 3.088961529436706e-05, 'epoch': 0.19}) +[2026-06-24 10:48:07,189][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.009307594038546085, 'train/grad_norm_pre_clip': 10.299039840698242, 'timing/data': 0.00029978202655911446, 'timing/model': 1.0613602709490806, 'learning_rate/qwen_vl_interface': 3.084380468735274e-06, 'learning_rate/action_model': 3.084380468735274e-05, 'epoch': 0.19}) +[2026-06-24 10:48:08,251][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.03143007308244705, 'train/grad_norm_pre_clip': 57.430023193359375, 'timing/data': 0.00030158599838614464, 'timing/model': 1.0599404878448695, 'learning_rate/qwen_vl_interface': 3.0798019495089888e-06, 'learning_rate/action_model': 3.079801949508989e-05, 'epoch': 0.19}) +[2026-06-24 10:48:09,314][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.17191395163536072, 'train/grad_norm_pre_clip': 58.01200485229492, 'timing/data': 0.00033050007186830044, 'timing/model': 1.0619255530182272, 'learning_rate/qwen_vl_interface': 3.075225977130995e-06, 'learning_rate/action_model': 3.075225977130995e-05, 'epoch': 0.19}) +[2026-06-24 10:48:10,603][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.3941778242588043, 'train/grad_norm_pre_clip': 114.22845458984375, 'timing/data': 0.00031068315729498863, 'timing/model': 1.2872791120316833, 'learning_rate/qwen_vl_interface': 3.07065255697145e-06, 'learning_rate/action_model': 3.07065255697145e-05, 'epoch': 0.19}) +[2026-06-24 10:48:11,668][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.32277238368988037, 'train/grad_norm_pre_clip': 72.3720474243164, 'timing/data': 0.0003261319361627102, 'timing/model': 1.062908989842981, 'learning_rate/qwen_vl_interface': 3.066081694397519e-06, 'learning_rate/action_model': 3.066081694397519e-05, 'epoch': 0.19}) +[2026-06-24 10:48:12,730][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.11066386103630066, 'train/grad_norm_pre_clip': 104.97669219970703, 'timing/data': 0.00033842516131699085, 'timing/model': 1.0602711299434304, 'learning_rate/qwen_vl_interface': 3.0615133947733634e-06, 'learning_rate/action_model': 3.061513394773363e-05, 'epoch': 0.19}) +[2026-06-24 10:48:13,796][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.25821971893310547, 'train/grad_norm_pre_clip': 76.3021011352539, 'timing/data': 0.000421751057729125, 'timing/model': 1.0647680789697915, 'learning_rate/qwen_vl_interface': 3.0569476634601393e-06, 'learning_rate/action_model': 3.056947663460139e-05, 'epoch': 0.19}) +[2026-06-24 10:48:14,863][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.09354409575462341, 'train/grad_norm_pre_clip': 106.7020492553711, 'timing/data': 0.0003098808228969574, 'timing/model': 1.064949369058013, 'learning_rate/qwen_vl_interface': 3.0523845058159835e-06, 'learning_rate/action_model': 3.0523845058159834e-05, 'epoch': 0.19}) +[2026-06-24 10:48:15,925][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.10713811963796616, 'train/grad_norm_pre_clip': 90.23873901367188, 'timing/data': 0.0003963231574743986, 'timing/model': 1.0602609219495207, 'learning_rate/qwen_vl_interface': 3.0478239271960165e-06, 'learning_rate/action_model': 3.0478239271960164e-05, 'epoch': 0.19}) +[2026-06-24 10:48:16,988][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.018915042281150818, 'train/grad_norm_pre_clip': 39.26881790161133, 'timing/data': 0.00038264808245003223, 'timing/model': 1.0617918048519641, 'learning_rate/qwen_vl_interface': 3.0432659329523317e-06, 'learning_rate/action_model': 3.0432659329523317e-05, 'epoch': 0.19}) +[2026-06-24 10:48:18,053][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.14576657116413116, 'train/grad_norm_pre_clip': 99.01697540283203, 'timing/data': 0.00031438982114195824, 'timing/model': 1.062720478978008, 'learning_rate/qwen_vl_interface': 3.038710528433991e-06, 'learning_rate/action_model': 3.038710528433991e-05, 'epoch': 0.19}) +[2026-06-24 10:48:19,119][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.11989405751228333, 'train/grad_norm_pre_clip': 63.0881462097168, 'timing/data': 0.0003029080107808113, 'timing/model': 1.0644328310154378, 'learning_rate/qwen_vl_interface': 3.0341577189870143e-06, 'learning_rate/action_model': 3.034157718987014e-05, 'epoch': 0.19}) +[2026-06-24 10:48:20,181][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.024544309824705124, 'train/grad_norm_pre_clip': 22.983190536499023, 'timing/data': 0.0003931981045752764, 'timing/model': 1.0606492629740387, 'learning_rate/qwen_vl_interface': 3.0296075099543743e-06, 'learning_rate/action_model': 3.029607509954374e-05, 'epoch': 0.19}) +[2026-06-24 10:48:21,241][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2000 +[2026-06-24 10:48:46,174][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2000 +[2026-06-24 10:48:46,174][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2000 +[2026-06-24 10:49:17,317][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2000 +[2026-06-24 10:51:46,370][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-06-24 10:51:46,371][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.10854806005954742, 'train/grad_norm_pre_clip': 56.5028076171875, 'eval/action_loss/seconds': 24.930826738942415, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.08313813805580139, 'eval/latency_2/loss': 0.08313813805580139, 'eval/flappy/loss': 0.08313813805580139, 'eval/flappy/latency_2/loss': 0.08313813805580139, 'eval/flap_precision': 0.956343792633015, 'eval/flap_recall': 0.9511533242876526, 'eval/flap_f1': 0.9537414965986396, 'eval/flap_rate_model': 0.2036111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9537414965986396, 'eval/latency_2/flap_precision': 0.956343792633015, 'eval/latency_2/flap_recall': 0.9511533242876526, 'eval/latency_2/flap_f1': 0.9537414965986396, 'eval/latency_2/flap_rate_model': 0.2036111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9537414965986396, 'eval/action_classification/data_wait_seconds': 0.33030927181243896, 'eval/action_classification/predict_seconds': 30.721773147583008, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 116.27999999999801, 'rl_games_eval/mid_train/mean_length': 956.0, 'rl_games_eval/mid_train/std_reward': 102.48608490912143, 'rl_games_eval/mid_train/std_length': 827.2419235024298, 'rl_games_eval/mid_train/macro_mean_reward': 116.27999999999801, 'rl_games_eval/mid_train/macro_mean_length': 956.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 116.27999999999801, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 956.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 102.48608490912143, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 827.2419235024298, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 116.27999999999801, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003213728778064251, 'timing/model': 1.0591002060100436, 'learning_rate/qwen_vl_interface': 3.025059906675997e-06, 'learning_rate/action_model': 3.025059906675997e-05, 'epoch': 0.19}) +[2026-06-24 10:51:46,371][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_2000_state +[2026-06-24 10:51:46,374][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 10:51:48,905][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_2000_state/model.safetensors +[2026-06-24 10:51:54,470][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_2000_state/optimizer.bin +[2026-06-24 10:51:54,471][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_2000_state/sampler.bin +[2026-06-24 10:51:54,471][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-24 10:51:54,472][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-24 10:51:54,514][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-24 10:51:54,514][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 10:51:54,524][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 10:51:55,572][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.22873352468013763, 'train/grad_norm_pre_clip': 153.60385131835938, 'timing/data': 0.0004029760602861643, 'timing/model': 1.046557636000216, 'learning_rate/qwen_vl_interface': 3.020514914488748e-06, 'learning_rate/action_model': 3.020514914488748e-05, 'epoch': 0.19}) +[2026-06-24 10:51:56,623][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.06194397062063217, 'train/grad_norm_pre_clip': 38.32224655151367, 'timing/data': 0.00033451803028583527, 'timing/model': 1.0490469371434301, 'learning_rate/qwen_vl_interface': 3.0159725387264317e-06, 'learning_rate/action_model': 3.0159725387264314e-05, 'epoch': 0.19}) +[2026-06-24 10:51:57,675][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.034745149314403534, 'train/grad_norm_pre_clip': 48.13714599609375, 'timing/data': 0.00032236496917903423, 'timing/model': 1.0507248560898006, 'learning_rate/qwen_vl_interface': 3.0114327847197734e-06, 'learning_rate/action_model': 3.011432784719773e-05, 'epoch': 0.19}) +[2026-06-24 10:51:58,723][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.0812709629535675, 'train/grad_norm_pre_clip': 94.38179016113281, 'timing/data': 0.0003382840659469366, 'timing/model': 1.0458947629667819, 'learning_rate/qwen_vl_interface': 3.0068956577964297e-06, 'learning_rate/action_model': 3.0068956577964298e-05, 'epoch': 0.19}) +[2026-06-24 10:51:59,780][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.00952662993222475, 'train/grad_norm_pre_clip': 13.332822799682617, 'timing/data': 0.0003433551173657179, 'timing/model': 1.0555259259417653, 'learning_rate/qwen_vl_interface': 3.0023611632809736e-06, 'learning_rate/action_model': 3.0023611632809735e-05, 'epoch': 0.19}) +[2026-06-24 10:52:00,840][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.006844537798315287, 'train/grad_norm_pre_clip': 6.4760918617248535, 'timing/data': 0.0003394470550119877, 'timing/model': 1.0581681139301509, 'learning_rate/qwen_vl_interface': 2.997829306494886e-06, 'learning_rate/action_model': 2.9978293064948858e-05, 'epoch': 0.19}) +[2026-06-24 10:52:01,889][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.16779683530330658, 'train/grad_norm_pre_clip': 66.84698486328125, 'timing/data': 0.00031093298457562923, 'timing/model': 1.0478346331510693, 'learning_rate/qwen_vl_interface': 2.993300092756554e-06, 'learning_rate/action_model': 2.9933000927565536e-05, 'epoch': 0.19}) +[2026-06-24 10:52:03,204][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.17234089970588684, 'train/grad_norm_pre_clip': 79.79715728759766, 'timing/data': 0.0003125660587102175, 'timing/model': 1.3134955239947885, 'learning_rate/qwen_vl_interface': 2.9887735273812583e-06, 'learning_rate/action_model': 2.9887735273812583e-05, 'epoch': 0.19}) +[2026-06-24 10:52:04,256][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.055200617760419846, 'train/grad_norm_pre_clip': 47.975006103515625, 'timing/data': 0.00035839201882481575, 'timing/model': 1.0501196319237351, 'learning_rate/qwen_vl_interface': 2.9842496156811786e-06, 'learning_rate/action_model': 2.9842496156811782e-05, 'epoch': 0.19}) +[2026-06-24 10:52:05,310][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.33994606137275696, 'train/grad_norm_pre_clip': 48.282554626464844, 'timing/data': 0.0003142599016427994, 'timing/model': 1.0531564010307193, 'learning_rate/qwen_vl_interface': 2.979728362965377e-06, 'learning_rate/action_model': 2.9797283629653766e-05, 'epoch': 0.19}) +[2026-06-24 10:52:06,378][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.16260196268558502, 'train/grad_norm_pre_clip': 70.88439178466797, 'timing/data': 0.0003810350317507982, 'timing/model': 1.0657935850322247, 'learning_rate/qwen_vl_interface': 2.9752097745397967e-06, 'learning_rate/action_model': 2.9752097745397966e-05, 'epoch': 0.19}) +[2026-06-24 10:52:07,572][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.337503582239151, 'train/grad_norm_pre_clip': 133.9420166015625, 'timing/data': 0.0005180221050977707, 'timing/model': 1.1924486821517348, 'learning_rate/qwen_vl_interface': 2.97069385570725e-06, 'learning_rate/action_model': 2.97069385570725e-05, 'epoch': 0.19}) +[2026-06-24 10:52:08,648][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.16103003919124603, 'train/grad_norm_pre_clip': 92.45529174804688, 'timing/data': 0.00034762313589453697, 'timing/model': 1.0743555829394609, 'learning_rate/qwen_vl_interface': 2.96618061176742e-06, 'learning_rate/action_model': 2.9661806117674202e-05, 'epoch': 0.19}) +[2026-06-24 10:52:09,704][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.022154036909341812, 'train/grad_norm_pre_clip': 25.20026397705078, 'timing/data': 0.0005282810889184475, 'timing/model': 1.0532326761167496, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 0.19}) +[2026-06-24 10:52:10,757][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.04566800221800804, 'train/grad_norm_pre_clip': 38.46805953979492, 'timing/data': 0.0003243889659643173, 'timing/model': 1.0522856879979372, 'learning_rate/qwen_vl_interface': 2.9571621697489433e-06, 'learning_rate/action_model': 2.9571621697489434e-05, 'epoch': 0.19}) +[2026-06-24 10:52:11,814][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.02971172146499157, 'train/grad_norm_pre_clip': 41.383663177490234, 'timing/data': 0.0003037999849766493, 'timing/model': 1.0546606548596174, 'learning_rate/qwen_vl_interface': 2.952656982253937e-06, 'learning_rate/action_model': 2.9526569822539368e-05, 'epoch': 0.2}) +[2026-06-24 10:52:12,871][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.11959154158830643, 'train/grad_norm_pre_clip': 44.87373733520508, 'timing/data': 0.00030061393044888973, 'timing/model': 1.0555600700899959, 'learning_rate/qwen_vl_interface': 2.948154490818926e-06, 'learning_rate/action_model': 2.9481544908189264e-05, 'epoch': 0.2}) +[2026-06-24 10:52:13,930][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.05782328546047211, 'train/grad_norm_pre_clip': 41.19374465942383, 'timing/data': 0.00034016789868474007, 'timing/model': 1.0573709218297154, 'learning_rate/qwen_vl_interface': 2.9436547007278303e-06, 'learning_rate/action_model': 2.9436547007278304e-05, 'epoch': 0.2}) +[2026-06-24 10:52:14,986][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.1754215508699417, 'train/grad_norm_pre_clip': 72.05891418457031, 'timing/data': 0.00032563088461756706, 'timing/model': 1.0552650201134384, 'learning_rate/qwen_vl_interface': 2.939157617261407e-06, 'learning_rate/action_model': 2.9391576172614072e-05, 'epoch': 0.2}) +[2026-06-24 10:52:16,044][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.21234498918056488, 'train/grad_norm_pre_clip': 94.31808471679688, 'timing/data': 0.00034903502091765404, 'timing/model': 1.0559508770238608, 'learning_rate/qwen_vl_interface': 2.9346632456972348e-06, 'learning_rate/action_model': 2.9346632456972346e-05, 'epoch': 0.2}) +[2026-06-24 10:52:17,099][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.036234501749277115, 'train/grad_norm_pre_clip': 36.16365432739258, 'timing/data': 0.0003252702299505472, 'timing/model': 1.0540576949715614, 'learning_rate/qwen_vl_interface': 2.930171591309707e-06, 'learning_rate/action_model': 2.9301715913097067e-05, 'epoch': 0.2}) +[2026-06-24 10:52:18,155][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.09973108768463135, 'train/grad_norm_pre_clip': 45.78879928588867, 'timing/data': 0.00030306819826364517, 'timing/model': 1.0544420059304684, 'learning_rate/qwen_vl_interface': 2.925682659370029e-06, 'learning_rate/action_model': 2.9256826593700286e-05, 'epoch': 0.2}) +[2026-06-24 10:52:19,425][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.12536843121051788, 'train/grad_norm_pre_clip': 67.7870864868164, 'timing/data': 0.0003346179146319628, 'timing/model': 1.267807936994359, 'learning_rate/qwen_vl_interface': 2.9211964551462157e-06, 'learning_rate/action_model': 2.9211964551462158e-05, 'epoch': 0.2}) +[2026-06-24 10:52:20,484][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.05061028152704239, 'train/grad_norm_pre_clip': 36.725276947021484, 'timing/data': 0.0004993169568479061, 'timing/model': 1.0569927729666233, 'learning_rate/qwen_vl_interface': 2.916712983903078e-06, 'learning_rate/action_model': 2.9167129839030776e-05, 'epoch': 0.2}) +[2026-06-24 10:52:21,543][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.006914781406521797, 'train/grad_norm_pre_clip': 5.9396514892578125, 'timing/data': 0.0003574211150407791, 'timing/model': 1.0574963190592825, 'learning_rate/qwen_vl_interface': 2.9122322509022147e-06, 'learning_rate/action_model': 2.912232250902215e-05, 'epoch': 0.2}) +[2026-06-24 10:52:22,601][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.0938565582036972, 'train/grad_norm_pre_clip': 59.78877639770508, 'timing/data': 0.00033620093017816544, 'timing/model': 1.056721735978499, 'learning_rate/qwen_vl_interface': 2.9077542614020217e-06, 'learning_rate/action_model': 2.9077542614020213e-05, 'epoch': 0.2}) +[2026-06-24 10:52:23,660][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.10511445254087448, 'train/grad_norm_pre_clip': 66.2682876586914, 'timing/data': 0.0002998120617121458, 'timing/model': 1.0568300390150398, 'learning_rate/qwen_vl_interface': 2.9032790206576626e-06, 'learning_rate/action_model': 2.9032790206576628e-05, 'epoch': 0.2}) +[2026-06-24 10:52:24,719][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.11187107115983963, 'train/grad_norm_pre_clip': 75.39164733886719, 'timing/data': 0.0004324407782405615, 'timing/model': 1.0573155800811946, 'learning_rate/qwen_vl_interface': 2.898806533921085e-06, 'learning_rate/action_model': 2.8988065339210853e-05, 'epoch': 0.2}) +[2026-06-24 10:52:25,778][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.07083785533905029, 'train/grad_norm_pre_clip': 30.775358200073242, 'timing/data': 0.00031052203848958015, 'timing/model': 1.057844082126394, 'learning_rate/qwen_vl_interface': 2.8943368064410038e-06, 'learning_rate/action_model': 2.8943368064410036e-05, 'epoch': 0.2}) +[2026-06-24 10:52:26,837][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.030129095539450645, 'train/grad_norm_pre_clip': 28.77466583251953, 'timing/data': 0.0004938968922942877, 'timing/model': 1.0572507090400904, 'learning_rate/qwen_vl_interface': 2.889869843462888e-06, 'learning_rate/action_model': 2.8898698434628878e-05, 'epoch': 0.2}) +[2026-06-24 10:52:27,894][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.023894835263490677, 'train/grad_norm_pre_clip': 39.27661895751953, 'timing/data': 0.00031940010376274586, 'timing/model': 1.0553568750619888, 'learning_rate/qwen_vl_interface': 2.8854056502289694e-06, 'learning_rate/action_model': 2.8854056502289696e-05, 'epoch': 0.2}) +[2026-06-24 10:52:28,956][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.005055077839642763, 'train/grad_norm_pre_clip': 4.894235610961914, 'timing/data': 0.0005220691673457623, 'timing/model': 1.060090968851, 'learning_rate/qwen_vl_interface': 2.8809442319782285e-06, 'learning_rate/action_model': 2.880944231978228e-05, 'epoch': 0.2}) +[2026-06-24 10:52:30,016][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.06868996471166611, 'train/grad_norm_pre_clip': 43.26507568359375, 'timing/data': 0.0005181219894438982, 'timing/model': 1.0580181900877506, 'learning_rate/qwen_vl_interface': 2.8764855939463887e-06, 'learning_rate/action_model': 2.8764855939463886e-05, 'epoch': 0.2}) +[2026-06-24 10:52:31,075][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.28577110171318054, 'train/grad_norm_pre_clip': 93.2667007446289, 'timing/data': 0.000526858028024435, 'timing/model': 1.0561114919837564, 'learning_rate/qwen_vl_interface': 2.8720297413659076e-06, 'learning_rate/action_model': 2.8720297413659075e-05, 'epoch': 0.2}) +[2026-06-24 10:52:32,134][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.1763283759355545, 'train/grad_norm_pre_clip': 45.19715118408203, 'timing/data': 0.00029769795946776867, 'timing/model': 1.0576660891529173, 'learning_rate/qwen_vl_interface': 2.8675766794659787e-06, 'learning_rate/action_model': 2.8675766794659787e-05, 'epoch': 0.2}) +[2026-06-24 10:52:33,193][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.020009534433484077, 'train/grad_norm_pre_clip': 9.4960355758667, 'timing/data': 0.0003067960496991873, 'timing/model': 1.05744037614204, 'learning_rate/qwen_vl_interface': 2.8631264134725162e-06, 'learning_rate/action_model': 2.863126413472516e-05, 'epoch': 0.2}) +[2026-06-24 10:52:34,315][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.0020883199758827686, 'train/grad_norm_pre_clip': 1.5104135274887085, 'timing/data': 0.0003355999942868948, 'timing/model': 1.058111635968089, 'learning_rate/qwen_vl_interface': 2.858678948608155e-06, 'learning_rate/action_model': 2.858678948608155e-05, 'epoch': 0.2}) +[2026-06-24 10:52:35,383][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.02051805704832077, 'train/grad_norm_pre_clip': 17.38502311706543, 'timing/data': 0.00042982702143490314, 'timing/model': 1.066178214037791, 'learning_rate/qwen_vl_interface': 2.8542342900922463e-06, 'learning_rate/action_model': 2.8542342900922465e-05, 'epoch': 0.2}) +[2026-06-24 10:52:36,738][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.07292657345533371, 'train/grad_norm_pre_clip': 70.24190521240234, 'timing/data': 0.00042916578240692616, 'timing/model': 1.3534206640906632, 'learning_rate/qwen_vl_interface': 2.8497924431408404e-06, 'learning_rate/action_model': 2.84979244314084e-05, 'epoch': 0.2}) +[2026-06-24 10:52:37,797][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.08583744615316391, 'train/grad_norm_pre_clip': 39.35856246948242, 'timing/data': 0.0003432838711887598, 'timing/model': 1.0578653651755303, 'learning_rate/qwen_vl_interface': 2.8453534129666936e-06, 'learning_rate/action_model': 2.8453534129666935e-05, 'epoch': 0.2}) +[2026-06-24 10:52:38,857][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.05697833001613617, 'train/grad_norm_pre_clip': 59.199832916259766, 'timing/data': 0.0003351690247654915, 'timing/model': 1.0581774509046227, 'learning_rate/qwen_vl_interface': 2.840917204779255e-06, 'learning_rate/action_model': 2.840917204779255e-05, 'epoch': 0.2}) +[2026-06-24 10:52:39,918][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.00946588534861803, 'train/grad_norm_pre_clip': 19.663223266601562, 'timing/data': 0.0003219849895685911, 'timing/model': 1.0593741561751813, 'learning_rate/qwen_vl_interface': 2.8364838237846644e-06, 'learning_rate/action_model': 2.8364838237846642e-05, 'epoch': 0.2}) +[2026-06-24 10:52:40,980][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.029775764793157578, 'train/grad_norm_pre_clip': 20.203289031982422, 'timing/data': 0.00031313812360167503, 'timing/model': 1.0603086699265987, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 0.2}) +[2026-06-24 10:52:42,043][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.03287743777036667, 'train/grad_norm_pre_clip': 35.477294921875, 'timing/data': 0.00032554101198911667, 'timing/model': 1.0610014100093395, 'learning_rate/qwen_vl_interface': 2.82762556418197e-06, 'learning_rate/action_model': 2.82762556418197e-05, 'epoch': 0.2}) +[2026-06-24 10:52:43,103][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.029931694269180298, 'train/grad_norm_pre_clip': 29.40074920654297, 'timing/data': 0.00032975897192955017, 'timing/model': 1.0589693170040846, 'learning_rate/qwen_vl_interface': 2.8232006959695313e-06, 'learning_rate/action_model': 2.8232006959695312e-05, 'epoch': 0.2}) +[2026-06-24 10:52:44,163][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.02010931260883808, 'train/grad_norm_pre_clip': 50.476070404052734, 'timing/data': 0.00029991287738084793, 'timing/model': 1.058071322971955, 'learning_rate/qwen_vl_interface': 2.818778675741247e-06, 'learning_rate/action_model': 2.8187786757412467e-05, 'epoch': 0.2}) +[2026-06-24 10:52:45,221][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.12428401410579681, 'train/grad_norm_pre_clip': 69.41027069091797, 'timing/data': 0.00036265095695853233, 'timing/model': 1.0564044450875372, 'learning_rate/qwen_vl_interface': 2.814359508686607e-06, 'learning_rate/action_model': 2.814359508686607e-05, 'epoch': 0.2}) +[2026-06-24 10:52:46,282][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.14164282381534576, 'train/grad_norm_pre_clip': 83.32300567626953, 'timing/data': 0.0003165442030876875, 'timing/model': 1.0591215339954942, 'learning_rate/qwen_vl_interface': 2.8099431999917465e-06, 'learning_rate/action_model': 2.8099431999917463e-05, 'epoch': 0.2}) +[2026-06-24 10:52:47,341][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.04972349852323532, 'train/grad_norm_pre_clip': 95.63056945800781, 'timing/data': 0.0003297789953649044, 'timing/model': 1.0583821469917893, 'learning_rate/qwen_vl_interface': 2.805529754839452e-06, 'learning_rate/action_model': 2.8055297548394515e-05, 'epoch': 0.2}) +[2026-06-24 10:52:48,402][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.14903181791305542, 'train/grad_norm_pre_clip': 74.49504852294922, 'timing/data': 0.00032581109553575516, 'timing/model': 1.0591418819967657, 'learning_rate/qwen_vl_interface': 2.801119178409146e-06, 'learning_rate/action_model': 2.8011191784091463e-05, 'epoch': 0.2}) +[2026-06-24 10:52:49,462][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.004710840526968241, 'train/grad_norm_pre_clip': 4.120985984802246, 'timing/data': 0.0003061250317841768, 'timing/model': 1.0584251079708338, 'learning_rate/qwen_vl_interface': 2.7967114758768892e-06, 'learning_rate/action_model': 2.7967114758768892e-05, 'epoch': 0.2}) +[2026-06-24 10:52:50,522][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.08400367200374603, 'train/grad_norm_pre_clip': 30.616567611694336, 'timing/data': 0.00033578998409211636, 'timing/model': 1.0588011441286653, 'learning_rate/qwen_vl_interface': 2.7923066524153607e-06, 'learning_rate/action_model': 2.7923066524153607e-05, 'epoch': 0.2}) +[2026-06-24 10:52:51,582][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.08521413058042526, 'train/grad_norm_pre_clip': 61.87028884887695, 'timing/data': 0.00032020104117691517, 'timing/model': 1.0582663509994745, 'learning_rate/qwen_vl_interface': 2.78790471319387e-06, 'learning_rate/action_model': 2.7879047131938697e-05, 'epoch': 0.2}) +[2026-06-24 10:52:52,847][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.026013506576418877, 'train/grad_norm_pre_clip': 38.57114028930664, 'timing/data': 0.0003180059138685465, 'timing/model': 1.2630505179986358, 'learning_rate/qwen_vl_interface': 2.7835056633783374e-06, 'learning_rate/action_model': 2.7835056633783374e-05, 'epoch': 0.2}) +[2026-06-24 10:52:53,909][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.09865809231996536, 'train/grad_norm_pre_clip': 61.95895767211914, 'timing/data': 0.0003998798783868551, 'timing/model': 1.0608541970141232, 'learning_rate/qwen_vl_interface': 2.779109508131291e-06, 'learning_rate/action_model': 2.779109508131291e-05, 'epoch': 0.2}) +[2026-06-24 10:52:54,971][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.044758204370737076, 'train/grad_norm_pre_clip': 69.35333251953125, 'timing/data': 0.00032932800240814686, 'timing/model': 1.059813714819029, 'learning_rate/qwen_vl_interface': 2.7747162526118665e-06, 'learning_rate/action_model': 2.7747162526118662e-05, 'epoch': 0.2}) +[2026-06-24 10:52:56,031][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.08566603064537048, 'train/grad_norm_pre_clip': 40.43755340576172, 'timing/data': 0.00031294720247387886, 'timing/model': 1.0586372178513557, 'learning_rate/qwen_vl_interface': 2.770325901975791e-06, 'learning_rate/action_model': 2.770325901975791e-05, 'epoch': 0.2}) +[2026-06-24 10:52:57,092][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.047546062618494034, 'train/grad_norm_pre_clip': 41.6965217590332, 'timing/data': 0.000325201079249382, 'timing/model': 1.0595985418185592, 'learning_rate/qwen_vl_interface': 2.765938461375387e-06, 'learning_rate/action_model': 2.7659384613753868e-05, 'epoch': 0.2}) +[2026-06-24 10:52:58,154][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.002210642211139202, 'train/grad_norm_pre_clip': 3.7438161373138428, 'timing/data': 0.00030105491168797016, 'timing/model': 1.0598096780013293, 'learning_rate/qwen_vl_interface': 2.7615539359595607e-06, 'learning_rate/action_model': 2.7615539359595603e-05, 'epoch': 0.2}) +[2026-06-24 10:52:59,216][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.013472017832100391, 'train/grad_norm_pre_clip': 32.487483978271484, 'timing/data': 0.0003120249602943659, 'timing/model': 1.0609281258657575, 'learning_rate/qwen_vl_interface': 2.757172330873798e-06, 'learning_rate/action_model': 2.757172330873798e-05, 'epoch': 0.2}) +[2026-06-24 10:53:00,278][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.2979170083999634, 'train/grad_norm_pre_clip': 231.14979553222656, 'timing/data': 0.0003241680096834898, 'timing/model': 1.0599230909720063, 'learning_rate/qwen_vl_interface': 2.752793651260153e-06, 'learning_rate/action_model': 2.752793651260153e-05, 'epoch': 0.2}) +[2026-06-24 10:53:01,340][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.008446835912764072, 'train/grad_norm_pre_clip': 6.302648067474365, 'timing/data': 0.00031056301668286324, 'timing/model': 1.06035584397614, 'learning_rate/qwen_vl_interface': 2.7484179022572534e-06, 'learning_rate/action_model': 2.7484179022572533e-05, 'epoch': 0.2}) +[2026-06-24 10:53:02,401][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.017848102375864983, 'train/grad_norm_pre_clip': 18.876007080078125, 'timing/data': 0.0003649250138550997, 'timing/model': 1.0597823769785464, 'learning_rate/qwen_vl_interface': 2.7440450890002828e-06, 'learning_rate/action_model': 2.744045089000283e-05, 'epoch': 0.2}) +[2026-06-24 10:53:03,464][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.01872883178293705, 'train/grad_norm_pre_clip': 50.14714431762695, 'timing/data': 0.000441898824647069, 'timing/model': 1.0611793999560177, 'learning_rate/qwen_vl_interface': 2.739675216620984e-06, 'learning_rate/action_model': 2.7396752166209837e-05, 'epoch': 0.2}) +[2026-06-24 10:53:04,526][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.17646507918834686, 'train/grad_norm_pre_clip': 68.60600280761719, 'timing/data': 0.0003031690139323473, 'timing/model': 1.0601366830524057, 'learning_rate/qwen_vl_interface': 2.7353082902476425e-06, 'learning_rate/action_model': 2.7353082902476422e-05, 'epoch': 0.2}) +[2026-06-24 10:53:05,593][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.05880124121904373, 'train/grad_norm_pre_clip': 45.749114990234375, 'timing/data': 0.00040702405385673046, 'timing/model': 1.0656984399538487, 'learning_rate/qwen_vl_interface': 2.7309443150050884e-06, 'learning_rate/action_model': 2.7309443150050884e-05, 'epoch': 0.2}) +[2026-06-24 10:53:06,653][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.1267441064119339, 'train/grad_norm_pre_clip': 155.87745666503906, 'timing/data': 0.00030390010215342045, 'timing/model': 1.0577410981059074, 'learning_rate/qwen_vl_interface': 2.726583296014692e-06, 'learning_rate/action_model': 2.726583296014692e-05, 'epoch': 0.2}) +[2026-06-24 10:53:07,717][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.049490299075841904, 'train/grad_norm_pre_clip': 27.959196090698242, 'timing/data': 0.0003010949585586786, 'timing/model': 1.062747031915933, 'learning_rate/qwen_vl_interface': 2.722225238394352e-06, 'learning_rate/action_model': 2.722225238394352e-05, 'epoch': 0.2}) +[2026-06-24 10:53:08,986][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.003589992644265294, 'train/grad_norm_pre_clip': 3.453416109085083, 'timing/data': 0.00033727288246154785, 'timing/model': 1.2675623500254005, 'learning_rate/qwen_vl_interface': 2.7178701472584934e-06, 'learning_rate/action_model': 2.7178701472584934e-05, 'epoch': 0.2}) +[2026-06-24 10:53:10,048][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.004913410171866417, 'train/grad_norm_pre_clip': 5.330920219421387, 'timing/data': 0.00033916719257831573, 'timing/model': 1.060545089887455, 'learning_rate/qwen_vl_interface': 2.7135180277180533e-06, 'learning_rate/action_model': 2.7135180277180533e-05, 'epoch': 0.2}) +[2026-06-24 10:53:11,111][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.22598308324813843, 'train/grad_norm_pre_clip': 116.24697875976562, 'timing/data': 0.00030940095894038677, 'timing/model': 1.061588800046593, 'learning_rate/qwen_vl_interface': 2.7091688848804904e-06, 'learning_rate/action_model': 2.7091688848804904e-05, 'epoch': 0.2}) +[2026-06-24 10:53:12,176][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.09553992748260498, 'train/grad_norm_pre_clip': 59.68800354003906, 'timing/data': 0.0003081390168517828, 'timing/model': 1.063221924006939, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 0.2}) +[2026-06-24 10:53:13,242][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.0033810969907790422, 'train/grad_norm_pre_clip': 5.134670257568359, 'timing/data': 0.00029613589867949486, 'timing/model': 1.0638380399905145, 'learning_rate/qwen_vl_interface': 2.700479549726341e-06, 'learning_rate/action_model': 2.7004795497263406e-05, 'epoch': 0.2}) +[2026-06-24 10:53:14,305][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.0009275000775232911, 'train/grad_norm_pre_clip': 0.9384662508964539, 'timing/data': 0.0003087390214204788, 'timing/model': 1.0622898461297154, 'learning_rate/qwen_vl_interface': 2.6961393676071728e-06, 'learning_rate/action_model': 2.6961393676071727e-05, 'epoch': 0.2}) +[2026-06-24 10:53:15,368][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.013148868456482887, 'train/grad_norm_pre_clip': 26.96246910095215, 'timing/data': 0.00035006715916097164, 'timing/model': 1.0608226920012385, 'learning_rate/qwen_vl_interface': 2.6918021825857054e-06, 'learning_rate/action_model': 2.6918021825857055e-05, 'epoch': 0.2}) +[2026-06-24 10:53:16,431][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.03126535564661026, 'train/grad_norm_pre_clip': 71.48222351074219, 'timing/data': 0.0003412999212741852, 'timing/model': 1.0613169099669904, 'learning_rate/qwen_vl_interface': 2.6874679997518677e-06, 'learning_rate/action_model': 2.6874679997518675e-05, 'epoch': 0.2}) +[2026-06-24 10:53:17,491][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.16838331520557404, 'train/grad_norm_pre_clip': 46.71207809448242, 'timing/data': 0.0003220748621970415, 'timing/model': 1.0589329770300537, 'learning_rate/qwen_vl_interface': 2.683136824192065e-06, 'learning_rate/action_model': 2.683136824192065e-05, 'epoch': 0.2}) +[2026-06-24 10:53:18,554][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.11578697711229324, 'train/grad_norm_pre_clip': 50.90524673461914, 'timing/data': 0.0003016360569745302, 'timing/model': 1.0612868240568787, 'learning_rate/qwen_vl_interface': 2.6788086609891746e-06, 'learning_rate/action_model': 2.6788086609891742e-05, 'epoch': 0.2}) +[2026-06-24 10:53:19,618][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.03673757612705231, 'train/grad_norm_pre_clip': 77.46033477783203, 'timing/data': 0.0003195099998265505, 'timing/model': 1.0622645691037178, 'learning_rate/qwen_vl_interface': 2.674483515222534e-06, 'learning_rate/action_model': 2.674483515222534e-05, 'epoch': 0.2}) +[2026-06-24 10:53:20,681][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.32981908321380615, 'train/grad_norm_pre_clip': 120.73381042480469, 'timing/data': 0.00032130302861332893, 'timing/model': 1.0617409558035433, 'learning_rate/qwen_vl_interface': 2.670161391967944e-06, 'learning_rate/action_model': 2.670161391967944e-05, 'epoch': 0.2}) +[2026-06-24 10:53:21,744][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.03210284560918808, 'train/grad_norm_pre_clip': 53.53556442260742, 'timing/data': 0.0003045620396733284, 'timing/model': 1.0608057822100818, 'learning_rate/qwen_vl_interface': 2.665842296297657e-06, 'learning_rate/action_model': 2.665842296297657e-05, 'epoch': 0.2}) +[2026-06-24 10:53:22,809][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.36012592911720276, 'train/grad_norm_pre_clip': 91.80646514892578, 'timing/data': 0.00032932800240814686, 'timing/model': 1.0638091580476612, 'learning_rate/qwen_vl_interface': 2.6615262332803764e-06, 'learning_rate/action_model': 2.6615262332803763e-05, 'epoch': 0.2}) +[2026-06-24 10:53:23,873][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.08177224546670914, 'train/grad_norm_pre_clip': 25.776376724243164, 'timing/data': 0.0003102119080722332, 'timing/model': 1.062237488804385, 'learning_rate/qwen_vl_interface': 2.657213207981236e-06, 'learning_rate/action_model': 2.6572132079812357e-05, 'epoch': 0.2}) +[2026-06-24 10:53:25,137][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.2713741958141327, 'train/grad_norm_pre_clip': 83.0177001953125, 'timing/data': 0.0003280660603195429, 'timing/model': 1.262618106091395, 'learning_rate/qwen_vl_interface': 2.652903225461812e-06, 'learning_rate/action_model': 2.6529032254618118e-05, 'epoch': 0.2}) +[2026-06-24 10:53:26,202][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.052315037697553635, 'train/grad_norm_pre_clip': 72.40445709228516, 'timing/data': 0.0003719679079949856, 'timing/model': 1.062837185105309, 'learning_rate/qwen_vl_interface': 2.648596290780112e-06, 'learning_rate/action_model': 2.648596290780112e-05, 'epoch': 0.2}) +[2026-06-24 10:53:27,264][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.04150852933526039, 'train/grad_norm_pre_clip': 112.68101501464844, 'timing/data': 0.00030274782329797745, 'timing/model': 1.0609490619972348, 'learning_rate/qwen_vl_interface': 2.6442924089905625e-06, 'learning_rate/action_model': 2.6442924089905625e-05, 'epoch': 0.2}) +[2026-06-24 10:53:28,326][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.11663185060024261, 'train/grad_norm_pre_clip': 119.29838562011719, 'timing/data': 0.0003285459242761135, 'timing/model': 1.060677922796458, 'learning_rate/qwen_vl_interface': 2.6399915851440104e-06, 'learning_rate/action_model': 2.6399915851440106e-05, 'epoch': 0.2}) +[2026-06-24 10:53:29,394][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.017153706401586533, 'train/grad_norm_pre_clip': 66.69062042236328, 'timing/data': 0.00030410103499889374, 'timing/model': 1.0661381599493325, 'learning_rate/qwen_vl_interface': 2.6356938242877088e-06, 'learning_rate/action_model': 2.6356938242877088e-05, 'epoch': 0.2}) +[2026-06-24 10:53:30,459][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.02917403168976307, 'train/grad_norm_pre_clip': 48.72083282470703, 'timing/data': 0.00035867304541170597, 'timing/model': 1.0635221619158983, 'learning_rate/qwen_vl_interface': 2.6313991314653208e-06, 'learning_rate/action_model': 2.631399131465321e-05, 'epoch': 0.2}) +[2026-06-24 10:53:31,523][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.011270208284258842, 'train/grad_norm_pre_clip': 26.50165557861328, 'timing/data': 0.0003162131179124117, 'timing/model': 1.0624458708334714, 'learning_rate/qwen_vl_interface': 2.627107511716907e-06, 'learning_rate/action_model': 2.627107511716907e-05, 'epoch': 0.2}) +[2026-06-24 10:53:32,585][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.00389387970790267, 'train/grad_norm_pre_clip': 3.9532229900360107, 'timing/data': 0.0003056139685213566, 'timing/model': 1.0602030740119517, 'learning_rate/qwen_vl_interface': 2.6228189700789254e-06, 'learning_rate/action_model': 2.622818970078925e-05, 'epoch': 0.2}) +[2026-06-24 10:53:33,647][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.007388133089989424, 'train/grad_norm_pre_clip': 66.35191345214844, 'timing/data': 0.0003090000245720148, 'timing/model': 1.0601062618661672, 'learning_rate/qwen_vl_interface': 2.6185335115842157e-06, 'learning_rate/action_model': 2.6185335115842158e-05, 'epoch': 0.2}) +[2026-06-24 10:53:34,709][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.4324877858161926, 'train/grad_norm_pre_clip': 85.29081726074219, 'timing/data': 0.00036464398726820946, 'timing/model': 1.0604688520543277, 'learning_rate/qwen_vl_interface': 2.614251141262001e-06, 'learning_rate/action_model': 2.6142511412620007e-05, 'epoch': 0.2}) +[2026-06-24 10:53:35,771][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.02880582958459854, 'train/grad_norm_pre_clip': 33.52831268310547, 'timing/data': 0.000310312956571579, 'timing/model': 1.061227196129039, 'learning_rate/qwen_vl_interface': 2.609971864137883e-06, 'learning_rate/action_model': 2.609971864137883e-05, 'epoch': 0.2}) +[2026-06-24 10:53:36,834][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.08737530559301376, 'train/grad_norm_pre_clip': 65.8666763305664, 'timing/data': 0.00031824689358472824, 'timing/model': 1.0613776678219438, 'learning_rate/qwen_vl_interface': 2.6056956852338334e-06, 'learning_rate/action_model': 2.6056956852338332e-05, 'epoch': 0.2}) +[2026-06-24 10:53:37,895][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.04155564308166504, 'train/grad_norm_pre_clip': 280.69354248046875, 'timing/data': 0.0003042810130864382, 'timing/model': 1.0595164250116795, 'learning_rate/qwen_vl_interface': 2.601422609568182e-06, 'learning_rate/action_model': 2.601422609568182e-05, 'epoch': 0.2}) +[2026-06-24 10:53:38,959][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.07176609337329865, 'train/grad_norm_pre_clip': 58.712867736816406, 'timing/data': 0.00032845698297023773, 'timing/model': 1.061671971110627, 'learning_rate/qwen_vl_interface': 2.5971526421556248e-06, 'learning_rate/action_model': 2.5971526421556248e-05, 'epoch': 0.2}) +[2026-06-24 10:53:40,023][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.007280867546796799, 'train/grad_norm_pre_clip': 25.775890350341797, 'timing/data': 0.00030141486786305904, 'timing/model': 1.0623938760254532, 'learning_rate/qwen_vl_interface': 2.592885788007205e-06, 'learning_rate/action_model': 2.592885788007205e-05, 'epoch': 0.2}) +[2026-06-24 10:53:41,374][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.4636158347129822, 'train/grad_norm_pre_clip': 93.85206604003906, 'timing/data': 0.0003022169694304466, 'timing/model': 1.3495858658570796, 'learning_rate/qwen_vl_interface': 2.5886220521303135e-06, 'learning_rate/action_model': 2.5886220521303135e-05, 'epoch': 0.2}) +[2026-06-24 10:53:42,435][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.026949355378746986, 'train/grad_norm_pre_clip': 46.00364685058594, 'timing/data': 0.0003173351287841797, 'timing/model': 1.0601189779117703, 'learning_rate/qwen_vl_interface': 2.5843614395286843e-06, 'learning_rate/action_model': 2.584361439528684e-05, 'epoch': 0.2}) +[2026-06-24 10:53:43,495][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.22027826309204102, 'train/grad_norm_pre_clip': 119.06240844726562, 'timing/data': 0.00031433021649718285, 'timing/model': 1.0577072820160538, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 0.2}) +[2026-06-24 10:53:44,558][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.03129942715167999, 'train/grad_norm_pre_clip': 67.85282135009766, 'timing/data': 0.0003035899717360735, 'timing/model': 1.0615859809331596, 'learning_rate/qwen_vl_interface': 2.5758496041477983e-06, 'learning_rate/action_model': 2.575849604147798e-05, 'epoch': 0.2}) +[2026-06-24 10:53:45,620][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.2355353832244873, 'train/grad_norm_pre_clip': 211.486572265625, 'timing/data': 0.00030117505230009556, 'timing/model': 1.0610670379828662, 'learning_rate/qwen_vl_interface': 2.5715983913576536e-06, 'learning_rate/action_model': 2.5715983913576536e-05, 'epoch': 0.2}) +[2026-06-24 10:53:46,682][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.14246141910552979, 'train/grad_norm_pre_clip': 87.09537506103516, 'timing/data': 0.00031168502755463123, 'timing/model': 1.060442706104368, 'learning_rate/qwen_vl_interface': 2.567350321820986e-06, 'learning_rate/action_model': 2.567350321820986e-05, 'epoch': 0.2}) +[2026-06-24 10:53:47,748][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.012901843525469303, 'train/grad_norm_pre_clip': 24.817445755004883, 'timing/data': 0.00030875997617840767, 'timing/model': 1.063969013048336, 'learning_rate/qwen_vl_interface': 2.5631054005231375e-06, 'learning_rate/action_model': 2.5631054005231374e-05, 'epoch': 0.2}) +[2026-06-24 10:53:48,813][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.0682360902428627, 'train/grad_norm_pre_clip': 59.75586700439453, 'timing/data': 0.0003069760277867317, 'timing/model': 1.0636635900009423, 'learning_rate/qwen_vl_interface': 2.5588636324457604e-06, 'learning_rate/action_model': 2.5588636324457603e-05, 'epoch': 0.2}) +[2026-06-24 10:53:49,877][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.08221698552370071, 'train/grad_norm_pre_clip': 99.72962188720703, 'timing/data': 0.00032266508787870407, 'timing/model': 1.0618802241515368, 'learning_rate/qwen_vl_interface': 2.554625022566807e-06, 'learning_rate/action_model': 2.5546250225668066e-05, 'epoch': 0.2}) +[2026-06-24 10:53:50,938][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.07605389505624771, 'train/grad_norm_pre_clip': 73.91802215576172, 'timing/data': 0.0003069150261580944, 'timing/model': 1.0595442310441285, 'learning_rate/qwen_vl_interface': 2.550389575860523e-06, 'learning_rate/action_model': 2.550389575860523e-05, 'epoch': 0.2}) +[2026-06-24 10:53:51,999][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.25754085183143616, 'train/grad_norm_pre_clip': 70.78407287597656, 'timing/data': 0.0002973980735987425, 'timing/model': 1.0592815079726279, 'learning_rate/qwen_vl_interface': 2.5461572972974384e-06, 'learning_rate/action_model': 2.5461572972974384e-05, 'epoch': 0.2}) +[2026-06-24 10:53:53,061][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.1538788080215454, 'train/grad_norm_pre_clip': 52.24323272705078, 'timing/data': 0.0003487248905003071, 'timing/model': 1.06120835384354, 'learning_rate/qwen_vl_interface': 2.541928191844369e-06, 'learning_rate/action_model': 2.5419281918443687e-05, 'epoch': 0.2}) +[2026-06-24 10:53:54,124][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.002216165652498603, 'train/grad_norm_pre_clip': 2.0828826427459717, 'timing/data': 0.0005197159480303526, 'timing/model': 1.0606014858931303, 'learning_rate/qwen_vl_interface': 2.5377022644644074e-06, 'learning_rate/action_model': 2.5377022644644072e-05, 'epoch': 0.2}) +[2026-06-24 10:53:55,188][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.008927997201681137, 'train/grad_norm_pre_clip': 13.290336608886719, 'timing/data': 0.000393799040466547, 'timing/model': 1.0618139510042965, 'learning_rate/qwen_vl_interface': 2.533479520116911e-06, 'learning_rate/action_model': 2.533479520116911e-05, 'epoch': 0.2}) +[2026-06-24 10:53:56,464][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.0984424278140068, 'train/grad_norm_pre_clip': 47.8446159362793, 'timing/data': 0.00032665301114320755, 'timing/model': 1.2742653689347208, 'learning_rate/qwen_vl_interface': 2.5292599637575094e-06, 'learning_rate/action_model': 2.5292599637575092e-05, 'epoch': 0.2}) +[2026-06-24 10:53:57,528][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.08666379004716873, 'train/grad_norm_pre_clip': 82.66620635986328, 'timing/data': 0.00036700814962387085, 'timing/model': 1.0626938419882208, 'learning_rate/qwen_vl_interface': 2.5250436003380836e-06, 'learning_rate/action_model': 2.5250436003380834e-05, 'epoch': 0.2}) +[2026-06-24 10:53:58,594][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.06945807486772537, 'train/grad_norm_pre_clip': 51.434871673583984, 'timing/data': 0.000313017051666975, 'timing/model': 1.0643287310376763, 'learning_rate/qwen_vl_interface': 2.5208304348067723e-06, 'learning_rate/action_model': 2.5208304348067723e-05, 'epoch': 0.2}) +[2026-06-24 10:53:59,658][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.005025356542319059, 'train/grad_norm_pre_clip': 8.071681022644043, 'timing/data': 0.00031164498068392277, 'timing/model': 1.0630122001748532, 'learning_rate/qwen_vl_interface': 2.5166204721079606e-06, 'learning_rate/action_model': 2.5166204721079606e-05, 'epoch': 0.2}) +[2026-06-24 10:54:00,721][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.08798237890005112, 'train/grad_norm_pre_clip': 77.64667510986328, 'timing/data': 0.0003203311935067177, 'timing/model': 1.0614534560590982, 'learning_rate/qwen_vl_interface': 2.5124137171822755e-06, 'learning_rate/action_model': 2.5124137171822754e-05, 'epoch': 0.2}) +[2026-06-24 10:54:01,786][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.025355171412229538, 'train/grad_norm_pre_clip': 57.78585433959961, 'timing/data': 0.00030096410773694515, 'timing/model': 1.062491824151948, 'learning_rate/qwen_vl_interface': 2.5082101749665745e-06, 'learning_rate/action_model': 2.5082101749665742e-05, 'epoch': 0.2}) +[2026-06-24 10:54:02,852][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.046291086822748184, 'train/grad_norm_pre_clip': 31.621110916137695, 'timing/data': 0.00032238499261438847, 'timing/model': 1.065140845021233, 'learning_rate/qwen_vl_interface': 2.5040098503939504e-06, 'learning_rate/action_model': 2.5040098503939503e-05, 'epoch': 0.21}) +[2026-06-24 10:54:03,918][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.12000039219856262, 'train/grad_norm_pre_clip': 44.13426208496094, 'timing/data': 0.0005280610639601946, 'timing/model': 1.0641577509231865, 'learning_rate/qwen_vl_interface': 2.49981274839372e-06, 'learning_rate/action_model': 2.4998127483937198e-05, 'epoch': 0.21}) +[2026-06-24 10:54:04,984][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.1419978141784668, 'train/grad_norm_pre_clip': 92.25743865966797, 'timing/data': 0.00035375403240323067, 'timing/model': 1.0642511560581625, 'learning_rate/qwen_vl_interface': 2.4956188738914124e-06, 'learning_rate/action_model': 2.4956188738914125e-05, 'epoch': 0.21}) +[2026-06-24 10:54:06,051][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.27076807618141174, 'train/grad_norm_pre_clip': 184.4204559326172, 'timing/data': 0.0004942070227116346, 'timing/model': 1.0642938159871846, 'learning_rate/qwen_vl_interface': 2.4914282318087767e-06, 'learning_rate/action_model': 2.4914282318087767e-05, 'epoch': 0.21}) +[2026-06-24 10:54:07,113][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.1680266410112381, 'train/grad_norm_pre_clip': 77.7907485961914, 'timing/data': 0.000310051953420043, 'timing/model': 1.0613069920800626, 'learning_rate/qwen_vl_interface': 2.4872408270637617e-06, 'learning_rate/action_model': 2.487240827063762e-05, 'epoch': 0.21}) +[2026-06-24 10:54:08,178][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.375823974609375, 'train/grad_norm_pre_clip': 84.89064025878906, 'timing/data': 0.00030414084903895855, 'timing/model': 1.0634706229902804, 'learning_rate/qwen_vl_interface': 2.4830566645705215e-06, 'learning_rate/action_model': 2.4830566645705212e-05, 'epoch': 0.21}) +[2026-06-24 10:54:09,242][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.11157718300819397, 'train/grad_norm_pre_clip': 86.43354797363281, 'timing/data': 0.00033601000905036926, 'timing/model': 1.06221052701585, 'learning_rate/qwen_vl_interface': 2.4788757492394034e-06, 'learning_rate/action_model': 2.4788757492394033e-05, 'epoch': 0.21}) +[2026-06-24 10:54:10,307][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.27003759145736694, 'train/grad_norm_pre_clip': 85.48181915283203, 'timing/data': 0.00032859621569514275, 'timing/model': 1.063542447052896, 'learning_rate/qwen_vl_interface': 2.474698085976947e-06, 'learning_rate/action_model': 2.474698085976947e-05, 'epoch': 0.21}) +[2026-06-24 10:54:11,371][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.03647712618112564, 'train/grad_norm_pre_clip': 26.292755126953125, 'timing/data': 0.0003351790364831686, 'timing/model': 1.062220708001405, 'learning_rate/qwen_vl_interface': 2.4705236796858688e-06, 'learning_rate/action_model': 2.4705236796858687e-05, 'epoch': 0.21}) +[2026-06-24 10:54:12,436][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.12813909351825714, 'train/grad_norm_pre_clip': 93.66959381103516, 'timing/data': 0.00032597198151052, 'timing/model': 1.063619202002883, 'learning_rate/qwen_vl_interface': 2.466352535265069e-06, 'learning_rate/action_model': 2.466352535265069e-05, 'epoch': 0.21}) +[2026-06-24 10:54:13,499][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.012107317335903645, 'train/grad_norm_pre_clip': 13.083117485046387, 'timing/data': 0.00031919917091727257, 'timing/model': 1.0611376359593123, 'learning_rate/qwen_vl_interface': 2.4621846576096173e-06, 'learning_rate/action_model': 2.4621846576096172e-05, 'epoch': 0.21}) +[2026-06-24 10:54:14,563][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.01708802580833435, 'train/grad_norm_pre_clip': 22.399639129638672, 'timing/data': 0.00030905893072485924, 'timing/model': 1.062179529806599, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 0.21}) +[2026-06-24 10:54:15,913][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.18201468884944916, 'train/grad_norm_pre_clip': 63.229148864746094, 'timing/data': 0.00032459921203553677, 'timing/model': 1.3479124039877206, 'learning_rate/qwen_vl_interface': 2.453858722155871e-06, 'learning_rate/action_model': 2.4538587221558706e-05, 'epoch': 0.21}) +[2026-06-24 10:54:16,978][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.005078122951090336, 'train/grad_norm_pre_clip': 4.537159442901611, 'timing/data': 0.0003115041181445122, 'timing/model': 1.063417554134503, 'learning_rate/qwen_vl_interface': 2.4497006741285213e-06, 'learning_rate/action_model': 2.4497006741285213e-05, 'epoch': 0.21}) +[2026-06-24 10:54:18,041][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.1192493885755539, 'train/grad_norm_pre_clip': 38.21979522705078, 'timing/data': 0.0003276141360402107, 'timing/model': 1.062032625079155, 'learning_rate/qwen_vl_interface': 2.4455459124084073e-06, 'learning_rate/action_model': 2.445545912408407e-05, 'epoch': 0.21}) +[2026-06-24 10:54:19,110][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.13480526208877563, 'train/grad_norm_pre_clip': 93.63922882080078, 'timing/data': 0.00032202410511672497, 'timing/model': 1.066665389109403, 'learning_rate/qwen_vl_interface': 2.4413944418713725e-06, 'learning_rate/action_model': 2.4413944418713726e-05, 'epoch': 0.21}) +[2026-06-24 10:54:20,174][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.021785106509923935, 'train/grad_norm_pre_clip': 35.99834060668945, 'timing/data': 0.00035193003714084625, 'timing/model': 1.0623742258176208, 'learning_rate/qwen_vl_interface': 2.4372462673894005e-06, 'learning_rate/action_model': 2.4372462673894006e-05, 'epoch': 0.21}) +[2026-06-24 10:54:21,238][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.07550345361232758, 'train/grad_norm_pre_clip': 89.68663024902344, 'timing/data': 0.000306685920804739, 'timing/model': 1.0632448110263795, 'learning_rate/qwen_vl_interface': 2.433101393830603e-06, 'learning_rate/action_model': 2.4331013938306026e-05, 'epoch': 0.21}) +[2026-06-24 10:54:22,300][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.15880133211612701, 'train/grad_norm_pre_clip': 58.770668029785156, 'timing/data': 0.00029469397850334644, 'timing/model': 1.0599251110106707, 'learning_rate/qwen_vl_interface': 2.42895982605922e-06, 'learning_rate/action_model': 2.4289598260592197e-05, 'epoch': 0.21}) +[2026-06-24 10:54:23,363][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.12298349291086197, 'train/grad_norm_pre_clip': 48.15484619140625, 'timing/data': 0.00035205110907554626, 'timing/model': 1.0613631799351424, 'learning_rate/qwen_vl_interface': 2.4248215689356126e-06, 'learning_rate/action_model': 2.4248215689356128e-05, 'epoch': 0.21}) +[2026-06-24 10:54:24,426][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.2559719979763031, 'train/grad_norm_pre_clip': 47.42848205566406, 'timing/data': 0.0003482741303741932, 'timing/model': 1.0615522658918053, 'learning_rate/qwen_vl_interface': 2.420686627316258e-06, 'learning_rate/action_model': 2.420686627316258e-05, 'epoch': 0.21}) +[2026-06-24 10:54:25,487][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.02130732871592045, 'train/grad_norm_pre_clip': 13.78296947479248, 'timing/data': 0.0003349278122186661, 'timing/model': 1.0600329139269888, 'learning_rate/qwen_vl_interface': 2.416555006053739e-06, 'learning_rate/action_model': 2.4165550060537388e-05, 'epoch': 0.21}) +[2026-06-24 10:54:26,548][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.20266512036323547, 'train/grad_norm_pre_clip': 90.57559204101562, 'timing/data': 0.0003010949585586786, 'timing/model': 1.0591768780723214, 'learning_rate/qwen_vl_interface': 2.4124267099967407e-06, 'learning_rate/action_model': 2.4124267099967406e-05, 'epoch': 0.21}) +[2026-06-24 10:54:27,613][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.048073891550302505, 'train/grad_norm_pre_clip': 37.435546875, 'timing/data': 0.00032769516110420227, 'timing/model': 1.0631205288227648, 'learning_rate/qwen_vl_interface': 2.408301743990051e-06, 'learning_rate/action_model': 2.408301743990051e-05, 'epoch': 0.21}) +[2026-06-24 10:54:28,673][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.37638312578201294, 'train/grad_norm_pre_clip': 96.46984100341797, 'timing/data': 0.00034472718834877014, 'timing/model': 1.059159796917811, 'learning_rate/qwen_vl_interface': 2.4041801128745463e-06, 'learning_rate/action_model': 2.4041801128745464e-05, 'epoch': 0.21}) +[2026-06-24 10:54:29,737][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.051092490553855896, 'train/grad_norm_pre_clip': 19.26289176940918, 'timing/data': 0.0003720289096236229, 'timing/model': 1.0617287170607597, 'learning_rate/qwen_vl_interface': 2.4000618214871934e-06, 'learning_rate/action_model': 2.4000618214871934e-05, 'epoch': 0.21}) +[2026-06-24 10:54:30,802][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.04467803239822388, 'train/grad_norm_pre_clip': 56.48847961425781, 'timing/data': 0.00034419610165059566, 'timing/model': 1.0634149319957942, 'learning_rate/qwen_vl_interface': 2.395946874661032e-06, 'learning_rate/action_model': 2.395946874661032e-05, 'epoch': 0.21}) +[2026-06-24 10:54:31,866][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.08377791941165924, 'train/grad_norm_pre_clip': 64.34949493408203, 'timing/data': 0.0003480631858110428, 'timing/model': 1.062492921948433, 'learning_rate/qwen_vl_interface': 2.3918352772251834e-06, 'learning_rate/action_model': 2.3918352772251834e-05, 'epoch': 0.21}) +[2026-06-24 10:54:32,929][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.06166769191622734, 'train/grad_norm_pre_clip': 46.33247375488281, 'timing/data': 0.00030689616687595844, 'timing/model': 1.062027398031205, 'learning_rate/qwen_vl_interface': 2.387727034004837e-06, 'learning_rate/action_model': 2.387727034004837e-05, 'epoch': 0.21}) +[2026-06-24 10:54:34,235][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.05852304771542549, 'train/grad_norm_pre_clip': 45.94919967651367, 'timing/data': 0.00032163294963538647, 'timing/model': 1.3039340218529105, 'learning_rate/qwen_vl_interface': 2.383622149821246e-06, 'learning_rate/action_model': 2.383622149821246e-05, 'epoch': 0.21}) +[2026-06-24 10:54:35,295][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.030211886391043663, 'train/grad_norm_pre_clip': 20.563518524169922, 'timing/data': 0.0003817661199718714, 'timing/model': 1.0583675201050937, 'learning_rate/qwen_vl_interface': 2.3795206294917183e-06, 'learning_rate/action_model': 2.3795206294917182e-05, 'epoch': 0.21}) +[2026-06-24 10:54:36,358][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.013821836560964584, 'train/grad_norm_pre_clip': 15.991914749145508, 'timing/data': 0.00029898108914494514, 'timing/model': 1.0622277150396258, 'learning_rate/qwen_vl_interface': 2.3754224778296165e-06, 'learning_rate/action_model': 2.3754224778296165e-05, 'epoch': 0.21}) +[2026-06-24 10:54:37,425][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.04370048642158508, 'train/grad_norm_pre_clip': 49.83222961425781, 'timing/data': 0.00032567093148827553, 'timing/model': 1.065160758793354, 'learning_rate/qwen_vl_interface': 2.37132769964435e-06, 'learning_rate/action_model': 2.37132769964435e-05, 'epoch': 0.21}) +[2026-06-24 10:54:38,490][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.048181477934122086, 'train/grad_norm_pre_clip': 30.77178955078125, 'timing/data': 0.000321884173899889, 'timing/model': 1.0628227009437978, 'learning_rate/qwen_vl_interface': 2.36723629974137e-06, 'learning_rate/action_model': 2.36723629974137e-05, 'epoch': 0.21}) +[2026-06-24 10:54:39,556][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.1344980001449585, 'train/grad_norm_pre_clip': 80.97164916992188, 'timing/data': 0.00031557190231978893, 'timing/model': 1.0643926581833512, 'learning_rate/qwen_vl_interface': 2.363148282922163e-06, 'learning_rate/action_model': 2.363148282922163e-05, 'epoch': 0.21}) +[2026-06-24 10:54:40,620][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.14915134012699127, 'train/grad_norm_pre_clip': 36.380008697509766, 'timing/data': 0.00031948997639119625, 'timing/model': 1.0629768220242113, 'learning_rate/qwen_vl_interface': 2.359063653984242e-06, 'learning_rate/action_model': 2.3590636539842418e-05, 'epoch': 0.21}) +[2026-06-24 10:54:41,686][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.032379310578107834, 'train/grad_norm_pre_clip': 18.614931106567383, 'timing/data': 0.00031837704591453075, 'timing/model': 1.0637385610025376, 'learning_rate/qwen_vl_interface': 2.3549824177211473e-06, 'learning_rate/action_model': 2.3549824177211475e-05, 'epoch': 0.21}) +[2026-06-24 10:54:42,749][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.11105915158987045, 'train/grad_norm_pre_clip': 64.68949127197266, 'timing/data': 0.000544652109965682, 'timing/model': 1.0612856280058622, 'learning_rate/qwen_vl_interface': 2.3509045789224385e-06, 'learning_rate/action_model': 2.3509045789224383e-05, 'epoch': 0.21}) +[2026-06-24 10:54:43,815][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.15333284437656403, 'train/grad_norm_pre_clip': 43.25423049926758, 'timing/data': 0.0003219840582460165, 'timing/model': 1.0643509810324758, 'learning_rate/qwen_vl_interface': 2.346830142373686e-06, 'learning_rate/action_model': 2.346830142373686e-05, 'epoch': 0.21}) +[2026-06-24 10:54:44,879][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.06446082144975662, 'train/grad_norm_pre_clip': 34.161964416503906, 'timing/data': 0.0003176771569997072, 'timing/model': 1.0622008559294045, 'learning_rate/qwen_vl_interface': 2.342759112856468e-06, 'learning_rate/action_model': 2.342759112856468e-05, 'epoch': 0.21}) +[2026-06-24 10:54:45,945][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.17509496212005615, 'train/grad_norm_pre_clip': 77.05319213867188, 'timing/data': 0.00032471888698637486, 'timing/model': 1.064738588873297, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 0.21}) +[2026-06-24 10:54:47,009][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.08383717387914658, 'train/grad_norm_pre_clip': 28.781831741333008, 'timing/data': 0.000327815068885684, 'timing/model': 1.062112341169268, 'learning_rate/qwen_vl_interface': 2.334627294022947e-06, 'learning_rate/action_model': 2.3346272940229472e-05, 'epoch': 0.21}) +[2026-06-24 10:54:48,075][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.026049628853797913, 'train/grad_norm_pre_clip': 15.821059226989746, 'timing/data': 0.00032410817220807076, 'timing/model': 1.0645538119133562, 'learning_rate/qwen_vl_interface': 2.330566514249788e-06, 'learning_rate/action_model': 2.3305665142497878e-05, 'epoch': 0.21}) +[2026-06-24 10:54:49,140][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.10305053740739822, 'train/grad_norm_pre_clip': 83.05479431152344, 'timing/data': 0.0003046810161322355, 'timing/model': 1.0634834428783506, 'learning_rate/qwen_vl_interface': 2.326509160594438e-06, 'learning_rate/action_model': 2.326509160594438e-05, 'epoch': 0.21}) +[2026-06-24 10:54:50,207][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.08018700778484344, 'train/grad_norm_pre_clip': 54.239479064941406, 'timing/data': 0.0003391869831830263, 'timing/model': 1.0651222390588373, 'learning_rate/qwen_vl_interface': 2.322455237818424e-06, 'learning_rate/action_model': 2.322455237818424e-05, 'epoch': 0.21}) +[2026-06-24 10:54:51,271][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.08832703530788422, 'train/grad_norm_pre_clip': 85.60750579833984, 'timing/data': 0.00031903921626508236, 'timing/model': 1.062600466888398, 'learning_rate/qwen_vl_interface': 2.3184047506792533e-06, 'learning_rate/action_model': 2.318404750679253e-05, 'epoch': 0.21}) +[2026-06-24 10:54:52,546][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.030594296753406525, 'train/grad_norm_pre_clip': 26.960874557495117, 'timing/data': 0.00031628296710550785, 'timing/model': 1.273601015098393, 'learning_rate/qwen_vl_interface': 2.314357703930396e-06, 'learning_rate/action_model': 2.314357703930396e-05, 'epoch': 0.21}) +[2026-06-24 10:54:53,610][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.015562858432531357, 'train/grad_norm_pre_clip': 12.629042625427246, 'timing/data': 0.0003318029921501875, 'timing/model': 1.0624890190083534, 'learning_rate/qwen_vl_interface': 2.3103141023212904e-06, 'learning_rate/action_model': 2.3103141023212905e-05, 'epoch': 0.21}) +[2026-06-24 10:54:54,676][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.11165013909339905, 'train/grad_norm_pre_clip': 52.03543472290039, 'timing/data': 0.00035628885962069035, 'timing/model': 1.064128956058994, 'learning_rate/qwen_vl_interface': 2.306273950597326e-06, 'learning_rate/action_model': 2.3062739505973256e-05, 'epoch': 0.21}) +[2026-06-24 10:54:55,740][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.017070222645998, 'train/grad_norm_pre_clip': 10.219527244567871, 'timing/data': 0.0003551968839019537, 'timing/model': 1.0625505850184709, 'learning_rate/qwen_vl_interface': 2.302237253499843e-06, 'learning_rate/action_model': 2.3022372534998432e-05, 'epoch': 0.21}) +[2026-06-24 10:54:56,803][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.11525383591651917, 'train/grad_norm_pre_clip': 75.72526550292969, 'timing/data': 0.00030489196069538593, 'timing/model': 1.060835856013, 'learning_rate/qwen_vl_interface': 2.2982040157661342e-06, 'learning_rate/action_model': 2.298204015766134e-05, 'epoch': 0.21}) +[2026-06-24 10:54:57,868][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.0918337032198906, 'train/grad_norm_pre_clip': 51.44040298461914, 'timing/data': 0.0003006539773195982, 'timing/model': 1.0639608311466873, 'learning_rate/qwen_vl_interface': 2.294174242129427e-06, 'learning_rate/action_model': 2.2941742421294268e-05, 'epoch': 0.21}) +[2026-06-24 10:54:58,933][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.26250746846199036, 'train/grad_norm_pre_clip': 209.80039978027344, 'timing/data': 0.0003435539547353983, 'timing/model': 1.0634891157969832, 'learning_rate/qwen_vl_interface': 2.290147937318888e-06, 'learning_rate/action_model': 2.2901479373188876e-05, 'epoch': 0.21}) +[2026-06-24 10:54:59,998][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.02520226500928402, 'train/grad_norm_pre_clip': 16.795228958129883, 'timing/data': 0.00029789889231324196, 'timing/model': 1.0631465730257332, 'learning_rate/qwen_vl_interface': 2.286125106059605e-06, 'learning_rate/action_model': 2.286125106059605e-05, 'epoch': 0.21}) +[2026-06-24 10:55:01,063][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.07468323409557343, 'train/grad_norm_pre_clip': 45.4133186340332, 'timing/data': 0.0003194599412381649, 'timing/model': 1.0637299180962145, 'learning_rate/qwen_vl_interface': 2.2821057530725985e-06, 'learning_rate/action_model': 2.2821057530725982e-05, 'epoch': 0.21}) +[2026-06-24 10:55:02,128][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.10301125049591064, 'train/grad_norm_pre_clip': 75.6841049194336, 'timing/data': 0.0003229361027479172, 'timing/model': 1.0631689860019833, 'learning_rate/qwen_vl_interface': 2.278089883074802e-06, 'learning_rate/action_model': 2.2780898830748022e-05, 'epoch': 0.21}) +[2026-06-24 10:55:03,194][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.01857520453631878, 'train/grad_norm_pre_clip': 9.193987846374512, 'timing/data': 0.00030888919718563557, 'timing/model': 1.064473724924028, 'learning_rate/qwen_vl_interface': 2.2740775007790637e-06, 'learning_rate/action_model': 2.274077500779064e-05, 'epoch': 0.21}) +[2026-06-24 10:55:04,261][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.07035274058580399, 'train/grad_norm_pre_clip': 56.766693115234375, 'timing/data': 0.000362620921805501, 'timing/model': 1.0643715730402619, 'learning_rate/qwen_vl_interface': 2.270068610894136e-06, 'learning_rate/action_model': 2.270068610894136e-05, 'epoch': 0.21}) +[2026-06-24 10:55:05,325][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.03426092118024826, 'train/grad_norm_pre_clip': 27.849586486816406, 'timing/data': 0.00033799512311816216, 'timing/model': 1.0627857269719243, 'learning_rate/qwen_vl_interface': 2.2660632181246765e-06, 'learning_rate/action_model': 2.2660632181246764e-05, 'epoch': 0.21}) +[2026-06-24 10:55:06,389][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.11967925727367401, 'train/grad_norm_pre_clip': 42.081886291503906, 'timing/data': 0.0005803888197988272, 'timing/model': 1.0620353191625327, 'learning_rate/qwen_vl_interface': 2.262061327171233e-06, 'learning_rate/action_model': 2.262061327171233e-05, 'epoch': 0.21}) +[2026-06-24 10:55:07,456][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.05186158046126366, 'train/grad_norm_pre_clip': 25.46815299987793, 'timing/data': 0.0002944120205938816, 'timing/model': 1.0656447929795831, 'learning_rate/qwen_vl_interface': 2.2580629427302513e-06, 'learning_rate/action_model': 2.258062942730251e-05, 'epoch': 0.21}) +[2026-06-24 10:55:08,522][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.06772585213184357, 'train/grad_norm_pre_clip': 47.3155632019043, 'timing/data': 0.0003362309653311968, 'timing/model': 1.064363137818873, 'learning_rate/qwen_vl_interface': 2.254068069494057e-06, 'learning_rate/action_model': 2.254068069494057e-05, 'epoch': 0.21}) +[2026-06-24 10:55:09,793][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.009115544147789478, 'train/grad_norm_pre_clip': 7.337806701660156, 'timing/data': 0.00030107516795396805, 'timing/model': 1.2693309960886836, 'learning_rate/qwen_vl_interface': 2.250076712150856e-06, 'learning_rate/action_model': 2.2500767121508556e-05, 'epoch': 0.21}) +[2026-06-24 10:55:10,859][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.18781930208206177, 'train/grad_norm_pre_clip': 82.16259002685547, 'timing/data': 0.0003182869404554367, 'timing/model': 1.0637135989964008, 'learning_rate/qwen_vl_interface': 2.2460888753847274e-06, 'learning_rate/action_model': 2.2460888753847272e-05, 'epoch': 0.21}) +[2026-06-24 10:55:11,924][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.0796605721116066, 'train/grad_norm_pre_clip': 50.10668182373047, 'timing/data': 0.0003153919242322445, 'timing/model': 1.0636031820904464, 'learning_rate/qwen_vl_interface': 2.242104563875621e-06, 'learning_rate/action_model': 2.242104563875621e-05, 'epoch': 0.21}) +[2026-06-24 10:55:12,990][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.11991913616657257, 'train/grad_norm_pre_clip': 48.5865592956543, 'timing/data': 0.0003651948645710945, 'timing/model': 1.0650089599657804, 'learning_rate/qwen_vl_interface': 2.2381237822993502e-06, 'learning_rate/action_model': 2.2381237822993505e-05, 'epoch': 0.21}) +[2026-06-24 10:55:14,055][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.06506801396608353, 'train/grad_norm_pre_clip': 27.826536178588867, 'timing/data': 0.00033585005439817905, 'timing/model': 1.0630104311276227, 'learning_rate/qwen_vl_interface': 2.2341465353275805e-06, 'learning_rate/action_model': 2.2341465353275806e-05, 'epoch': 0.21}) +[2026-06-24 10:55:15,122][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.10979826003313065, 'train/grad_norm_pre_clip': 74.47349548339844, 'timing/data': 0.00032947794534265995, 'timing/model': 1.0657016311306506, 'learning_rate/qwen_vl_interface': 2.230172827627836e-06, 'learning_rate/action_model': 2.230172827627836e-05, 'epoch': 0.21}) +[2026-06-24 10:55:16,187][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.02453063800930977, 'train/grad_norm_pre_clip': 33.14082336425781, 'timing/data': 0.00041042990051209927, 'timing/model': 1.0630230049137026, 'learning_rate/qwen_vl_interface': 2.2262026638634806e-06, 'learning_rate/action_model': 2.2262026638634805e-05, 'epoch': 0.21}) +[2026-06-24 10:55:17,255][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.23654258251190186, 'train/grad_norm_pre_clip': 98.97452545166016, 'timing/data': 0.0003030488733202219, 'timing/model': 1.0666702489834279, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 0.21}) +[2026-06-24 10:55:18,323][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.07548456639051437, 'train/grad_norm_pre_clip': 51.81999969482422, 'timing/data': 0.0003318730741739273, 'timing/model': 1.066187733085826, 'learning_rate/qwen_vl_interface': 2.2182729867736118e-06, 'learning_rate/action_model': 2.218272986773612e-05, 'epoch': 0.21}) +[2026-06-24 10:55:19,387][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.0670076385140419, 'train/grad_norm_pre_clip': 36.497554779052734, 'timing/data': 0.00033066095784306526, 'timing/model': 1.061966281849891, 'learning_rate/qwen_vl_interface': 2.2143134827540153e-06, 'learning_rate/action_model': 2.2143134827540152e-05, 'epoch': 0.21}) +[2026-06-24 10:55:20,452][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.025094017386436462, 'train/grad_norm_pre_clip': 31.648860931396484, 'timing/data': 0.00031927903182804585, 'timing/model': 1.063891404075548, 'learning_rate/qwen_vl_interface': 2.2103575412816335e-06, 'learning_rate/action_model': 2.2103575412816335e-05, 'epoch': 0.21}) +[2026-06-24 10:55:21,519][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.10000874102115631, 'train/grad_norm_pre_clip': 50.32821273803711, 'timing/data': 0.00033813389018177986, 'timing/model': 1.064911118010059, 'learning_rate/qwen_vl_interface': 2.206405166998985e-06, 'learning_rate/action_model': 2.2064051669989848e-05, 'epoch': 0.21}) +[2026-06-24 10:55:22,583][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.26356273889541626, 'train/grad_norm_pre_clip': 94.79918670654297, 'timing/data': 0.0003170550335198641, 'timing/model': 1.0629608791787177, 'learning_rate/qwen_vl_interface': 2.2024563645444027e-06, 'learning_rate/action_model': 2.2024563645444023e-05, 'epoch': 0.21}) +[2026-06-24 10:55:23,647][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.034930985420942307, 'train/grad_norm_pre_clip': 32.201988220214844, 'timing/data': 0.0004220418632030487, 'timing/model': 1.0625606880057603, 'learning_rate/qwen_vl_interface': 2.198511138552022e-06, 'learning_rate/action_model': 2.198511138552022e-05, 'epoch': 0.21}) +[2026-06-24 10:55:24,713][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.05710954964160919, 'train/grad_norm_pre_clip': 42.4107666015625, 'timing/data': 0.0003047320060431957, 'timing/model': 1.0638870981056243, 'learning_rate/qwen_vl_interface': 2.19456949365179e-06, 'learning_rate/action_model': 2.1945694936517898e-05, 'epoch': 0.21}) +[2026-06-24 10:55:25,779][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.033410508185625076, 'train/grad_norm_pre_clip': 88.52154541015625, 'timing/data': 0.0003134889993816614, 'timing/model': 1.0644916121382266, 'learning_rate/qwen_vl_interface': 2.1906314344694427e-06, 'learning_rate/action_model': 2.1906314344694426e-05, 'epoch': 0.21}) +[2026-06-24 10:55:27,049][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.23085275292396545, 'train/grad_norm_pre_clip': 91.29159545898438, 'timing/data': 0.0004938060883432627, 'timing/model': 1.2685314188711345, 'learning_rate/qwen_vl_interface': 2.1866969656265135e-06, 'learning_rate/action_model': 2.1866969656265133e-05, 'epoch': 0.21}) +[2026-06-24 10:55:28,113][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.013389026746153831, 'train/grad_norm_pre_clip': 10.815516471862793, 'timing/data': 0.0002938411198556423, 'timing/model': 1.0624244729988277, 'learning_rate/qwen_vl_interface': 2.1827660917403225e-06, 'learning_rate/action_model': 2.1827660917403223e-05, 'epoch': 0.21}) +[2026-06-24 10:55:29,183][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.04423924908041954, 'train/grad_norm_pre_clip': 38.63386917114258, 'timing/data': 0.000307347159832716, 'timing/model': 1.0676046039443463, 'learning_rate/qwen_vl_interface': 2.178838817423965e-06, 'learning_rate/action_model': 2.1788388174239648e-05, 'epoch': 0.21}) +[2026-06-24 10:55:30,250][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.0028721799608319998, 'train/grad_norm_pre_clip': 2.9860241413116455, 'timing/data': 0.0003007939085364342, 'timing/model': 1.065813141874969, 'learning_rate/qwen_vl_interface': 2.174915147286319e-06, 'learning_rate/action_model': 2.1749151472863188e-05, 'epoch': 0.21}) +[2026-06-24 10:55:31,317][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.06615155190229416, 'train/grad_norm_pre_clip': 61.39215850830078, 'timing/data': 0.00034829392097890377, 'timing/model': 1.065507299033925, 'learning_rate/qwen_vl_interface': 2.170995085932029e-06, 'learning_rate/action_model': 2.170995085932029e-05, 'epoch': 0.21}) +[2026-06-24 10:55:32,382][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.030778611078858376, 'train/grad_norm_pre_clip': 60.65556716918945, 'timing/data': 0.00032118288800120354, 'timing/model': 1.0638406411744654, 'learning_rate/qwen_vl_interface': 2.167078637961508e-06, 'learning_rate/action_model': 2.167078637961508e-05, 'epoch': 0.21}) +[2026-06-24 10:55:33,449][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.2644702196121216, 'train/grad_norm_pre_clip': 60.177040100097656, 'timing/data': 0.0003293980844318867, 'timing/model': 1.0651298509910703, 'learning_rate/qwen_vl_interface': 2.163165807970922e-06, 'learning_rate/action_model': 2.163165807970922e-05, 'epoch': 0.21}) +[2026-06-24 10:55:34,512][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.04435991123318672, 'train/grad_norm_pre_clip': 42.4847297668457, 'timing/data': 0.0003015059046447277, 'timing/model': 1.0612425559666008, 'learning_rate/qwen_vl_interface': 2.159256600552196e-06, 'learning_rate/action_model': 2.159256600552196e-05, 'epoch': 0.21}) +[2026-06-24 10:55:35,579][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.12335816770792007, 'train/grad_norm_pre_clip': 57.786373138427734, 'timing/data': 0.0003079280722886324, 'timing/model': 1.0651136310771108, 'learning_rate/qwen_vl_interface': 2.155351020293005e-06, 'learning_rate/action_model': 2.1553510202930053e-05, 'epoch': 0.21}) +[2026-06-24 10:55:36,643][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.10788024961948395, 'train/grad_norm_pre_clip': 96.49032592773438, 'timing/data': 0.00035973498597741127, 'timing/model': 1.062769962940365, 'learning_rate/qwen_vl_interface': 2.1514490717767623e-06, 'learning_rate/action_model': 2.151449071776762e-05, 'epoch': 0.21}) +[2026-06-24 10:55:37,714][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.06802557408809662, 'train/grad_norm_pre_clip': 39.066707611083984, 'timing/data': 0.00031852792017161846, 'timing/model': 1.0698166268412024, 'learning_rate/qwen_vl_interface': 2.1475507595826238e-06, 'learning_rate/action_model': 2.147550759582624e-05, 'epoch': 0.21}) +[2026-06-24 10:55:38,782][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.008615655824542046, 'train/grad_norm_pre_clip': 11.484274864196777, 'timing/data': 0.0003063851036131382, 'timing/model': 1.0661821861285716, 'learning_rate/qwen_vl_interface': 2.143656088285474e-06, 'learning_rate/action_model': 2.1436560882854738e-05, 'epoch': 0.21}) +[2026-06-24 10:55:39,851][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.14953021705150604, 'train/grad_norm_pre_clip': 34.304996490478516, 'timing/data': 0.0003261920064687729, 'timing/model': 1.0675241050776094, 'learning_rate/qwen_vl_interface': 2.1397650624559273e-06, 'learning_rate/action_model': 2.139765062455927e-05, 'epoch': 0.21}) +[2026-06-24 10:55:40,918][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.1643543243408203, 'train/grad_norm_pre_clip': 72.16877746582031, 'timing/data': 0.00029897107742726803, 'timing/model': 1.0652197510935366, 'learning_rate/qwen_vl_interface': 2.135877686660319e-06, 'learning_rate/action_model': 2.135877686660319e-05, 'epoch': 0.21}) +[2026-06-24 10:55:41,985][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.08132120221853256, 'train/grad_norm_pre_clip': 40.29549026489258, 'timing/data': 0.0003083581104874611, 'timing/model': 1.0656461610924453, 'learning_rate/qwen_vl_interface': 2.1319939654607034e-06, 'learning_rate/action_model': 2.1319939654607034e-05, 'epoch': 0.21}) +[2026-06-24 10:55:43,051][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.1164478287100792, 'train/grad_norm_pre_clip': 73.53012084960938, 'timing/data': 0.00032417895272374153, 'timing/model': 1.0643298709765077, 'learning_rate/qwen_vl_interface': 2.12811390341484e-06, 'learning_rate/action_model': 2.12811390341484e-05, 'epoch': 0.21}) +[2026-06-24 10:55:44,388][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.347167432308197, 'train/grad_norm_pre_clip': 110.59524536132812, 'timing/data': 0.00032035098411142826, 'timing/model': 1.335456890054047, 'learning_rate/qwen_vl_interface': 2.1242375050762003e-06, 'learning_rate/action_model': 2.1242375050762e-05, 'epoch': 0.21}) +[2026-06-24 10:55:45,455][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.19002537429332733, 'train/grad_norm_pre_clip': 52.561134338378906, 'timing/data': 0.00030469102784991264, 'timing/model': 1.0651810481213033, 'learning_rate/qwen_vl_interface': 2.1203647749939537e-06, 'learning_rate/action_model': 2.1203647749939536e-05, 'epoch': 0.21}) +[2026-06-24 10:55:46,524][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.06509430706501007, 'train/grad_norm_pre_clip': 46.22176742553711, 'timing/data': 0.0003110740799456835, 'timing/model': 1.0670047411695123, 'learning_rate/qwen_vl_interface': 2.116495717712963e-06, 'learning_rate/action_model': 2.1164957177129626e-05, 'epoch': 0.21}) +[2026-06-24 10:55:47,589][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.3284516930580139, 'train/grad_norm_pre_clip': 100.78321838378906, 'timing/data': 0.00034162099473178387, 'timing/model': 1.0635652369819582, 'learning_rate/qwen_vl_interface': 2.1126303377737847e-06, 'learning_rate/action_model': 2.1126303377737842e-05, 'epoch': 0.21}) +[2026-06-24 10:55:48,655][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.25436607003211975, 'train/grad_norm_pre_clip': 111.23674774169922, 'timing/data': 0.0003028789069503546, 'timing/model': 1.0638950860593468, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 0.21}) +[2026-06-24 10:55:49,721][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.0467492900788784, 'train/grad_norm_pre_clip': 36.92070770263672, 'timing/data': 0.0003055038396269083, 'timing/model': 1.0644214041531086, 'learning_rate/qwen_vl_interface': 2.1049106280614916e-06, 'learning_rate/action_model': 2.1049106280614915e-05, 'epoch': 0.21}) +[2026-06-24 10:55:50,787][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.11158870905637741, 'train/grad_norm_pre_clip': 75.11927032470703, 'timing/data': 0.0003508881200104952, 'timing/model': 1.0650114407762885, 'learning_rate/qwen_vl_interface': 2.101056307347888e-06, 'learning_rate/action_model': 2.101056307347888e-05, 'epoch': 0.21}) +[2026-06-24 10:55:51,852][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.07448355853557587, 'train/grad_norm_pre_clip': 35.32140350341797, 'timing/data': 0.000509526114910841, 'timing/model': 1.062775646103546, 'learning_rate/qwen_vl_interface': 2.097205682095105e-06, 'learning_rate/action_model': 2.0972056820951046e-05, 'epoch': 0.21}) +[2026-06-24 10:55:52,918][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.09643341600894928, 'train/grad_norm_pre_clip': 60.42102813720703, 'timing/data': 0.000420399010181427, 'timing/model': 1.0645734190475196, 'learning_rate/qwen_vl_interface': 2.0933587568220625e-06, 'learning_rate/action_model': 2.093358756822062e-05, 'epoch': 0.21}) +[2026-06-24 10:55:53,983][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.04445173963904381, 'train/grad_norm_pre_clip': 37.38362121582031, 'timing/data': 0.0003680500667542219, 'timing/model': 1.0627812570892274, 'learning_rate/qwen_vl_interface': 2.089515536043343e-06, 'learning_rate/action_model': 2.0895155360433434e-05, 'epoch': 0.21}) +[2026-06-24 10:55:55,046][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.022559477016329765, 'train/grad_norm_pre_clip': 15.937232971191406, 'timing/data': 0.00032482901588082314, 'timing/model': 1.062199414940551, 'learning_rate/qwen_vl_interface': 2.085676024269181e-06, 'learning_rate/action_model': 2.0856760242691813e-05, 'epoch': 0.22}) +[2026-06-24 10:55:56,113][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.02105642668902874, 'train/grad_norm_pre_clip': 12.383784294128418, 'timing/data': 0.0003064749762415886, 'timing/model': 1.0650702230632305, 'learning_rate/qwen_vl_interface': 2.081840226005459e-06, 'learning_rate/action_model': 2.0818402260054588e-05, 'epoch': 0.22}) +[2026-06-24 10:55:57,180][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.03175561875104904, 'train/grad_norm_pre_clip': 18.882293701171875, 'timing/data': 0.00035374402068555355, 'timing/model': 1.065030146157369, 'learning_rate/qwen_vl_interface': 2.078008145753697e-06, 'learning_rate/action_model': 2.0780081457536972e-05, 'epoch': 0.22}) +[2026-06-24 10:55:58,245][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.13941901922225952, 'train/grad_norm_pre_clip': 43.5509033203125, 'timing/data': 0.0003083879128098488, 'timing/model': 1.0639064779970795, 'learning_rate/qwen_vl_interface': 2.074179788011054e-06, 'learning_rate/action_model': 2.0741797880110543e-05, 'epoch': 0.22}) +[2026-06-24 10:55:59,312][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.014938953332602978, 'train/grad_norm_pre_clip': 9.625066757202148, 'timing/data': 0.0002956350799649954, 'timing/model': 1.064870365895331, 'learning_rate/qwen_vl_interface': 2.0703551572703233e-06, 'learning_rate/action_model': 2.070355157270323e-05, 'epoch': 0.22}) +[2026-06-24 10:56:00,585][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.07180161774158478, 'train/grad_norm_pre_clip': 44.517539978027344, 'timing/data': 0.00030874903313815594, 'timing/model': 1.2719082119874656, 'learning_rate/qwen_vl_interface': 2.0665342580199205e-06, 'learning_rate/action_model': 2.0665342580199204e-05, 'epoch': 0.22}) +[2026-06-24 10:56:01,650][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.02092881128191948, 'train/grad_norm_pre_clip': 17.975826263427734, 'timing/data': 0.0003213128075003624, 'timing/model': 1.0629516579210758, 'learning_rate/qwen_vl_interface': 2.0627170947438865e-06, 'learning_rate/action_model': 2.0627170947438863e-05, 'epoch': 0.22}) +[2026-06-24 10:56:02,712][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.12167920172214508, 'train/grad_norm_pre_clip': 86.12113189697266, 'timing/data': 0.00032730516977608204, 'timing/model': 1.0605670029763132, 'learning_rate/qwen_vl_interface': 2.0589036719218715e-06, 'learning_rate/action_model': 2.0589036719218712e-05, 'epoch': 0.22}) +[2026-06-24 10:56:03,776][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.15801434218883514, 'train/grad_norm_pre_clip': 44.96467971801758, 'timing/data': 0.0003289270680397749, 'timing/model': 1.0625576190650463, 'learning_rate/qwen_vl_interface': 2.0550939940291414e-06, 'learning_rate/action_model': 2.055093994029141e-05, 'epoch': 0.22}) +[2026-06-24 10:56:04,842][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.11058561503887177, 'train/grad_norm_pre_clip': 70.87226867675781, 'timing/data': 0.0003237379714846611, 'timing/model': 1.0642467779107392, 'learning_rate/qwen_vl_interface': 2.0512880655365644e-06, 'learning_rate/action_model': 2.0512880655365644e-05, 'epoch': 0.22}) +[2026-06-24 10:56:05,907][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.1312984675168991, 'train/grad_norm_pre_clip': 56.38478469848633, 'timing/data': 0.00030127610079944134, 'timing/model': 1.0632826189976186, 'learning_rate/qwen_vl_interface': 2.0474858909106122e-06, 'learning_rate/action_model': 2.047485890910612e-05, 'epoch': 0.22}) +[2026-06-24 10:56:06,971][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.3640459477901459, 'train/grad_norm_pre_clip': 74.05341339111328, 'timing/data': 0.00032066088169813156, 'timing/model': 1.0629201289266348, 'learning_rate/qwen_vl_interface': 2.043687474613347e-06, 'learning_rate/action_model': 2.043687474613347e-05, 'epoch': 0.22}) +[2026-06-24 10:56:08,036][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.25608670711517334, 'train/grad_norm_pre_clip': 41.724891662597656, 'timing/data': 0.0003071869723498821, 'timing/model': 1.0637703349348158, 'learning_rate/qwen_vl_interface': 2.0398928211024193e-06, 'learning_rate/action_model': 2.0398928211024193e-05, 'epoch': 0.22}) +[2026-06-24 10:56:09,100][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.24740362167358398, 'train/grad_norm_pre_clip': 69.03453063964844, 'timing/data': 0.0005510940682142973, 'timing/model': 1.0617887971457094, 'learning_rate/qwen_vl_interface': 2.0361019348310692e-06, 'learning_rate/action_model': 2.036101934831069e-05, 'epoch': 0.22}) +[2026-06-24 10:56:10,165][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.016221683472394943, 'train/grad_norm_pre_clip': 15.081219673156738, 'timing/data': 0.0003406889736652374, 'timing/model': 1.0631249151192605, 'learning_rate/qwen_vl_interface': 2.032314820248112e-06, 'learning_rate/action_model': 2.0323148202481122e-05, 'epoch': 0.22}) +[2026-06-24 10:56:11,228][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.10057609528303146, 'train/grad_norm_pre_clip': 35.4289665222168, 'timing/data': 0.00034390506334602833, 'timing/model': 1.0616661058738828, 'learning_rate/qwen_vl_interface': 2.0285314817979392e-06, 'learning_rate/action_model': 2.028531481797939e-05, 'epoch': 0.22}) +[2026-06-24 10:56:12,294][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.10796370357275009, 'train/grad_norm_pre_clip': 44.75284957885742, 'timing/data': 0.000496801920235157, 'timing/model': 1.0636178299319, 'learning_rate/qwen_vl_interface': 2.024751923920507e-06, 'learning_rate/action_model': 2.024751923920507e-05, 'epoch': 0.22}) +[2026-06-24 10:56:13,362][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.03936373069882393, 'train/grad_norm_pre_clip': 12.642845153808594, 'timing/data': 0.0003087688237428665, 'timing/model': 1.0671466609928757, 'learning_rate/qwen_vl_interface': 2.0209761510513376e-06, 'learning_rate/action_model': 2.0209761510513375e-05, 'epoch': 0.22}) +[2026-06-24 10:56:14,430][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.08175135403871536, 'train/grad_norm_pre_clip': 84.87171173095703, 'timing/data': 0.00039694388397037983, 'timing/model': 1.066327523905784, 'learning_rate/qwen_vl_interface': 2.017204167621511e-06, 'learning_rate/action_model': 2.017204167621511e-05, 'epoch': 0.22}) +[2026-06-24 10:56:15,498][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.10278821736574173, 'train/grad_norm_pre_clip': 29.136991500854492, 'timing/data': 0.00029731797985732555, 'timing/model': 1.0654479539953172, 'learning_rate/qwen_vl_interface': 2.0134359780576617e-06, 'learning_rate/action_model': 2.0134359780576615e-05, 'epoch': 0.22}) +[2026-06-24 10:56:16,565][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.01355844084173441, 'train/grad_norm_pre_clip': 10.605752944946289, 'timing/data': 0.000305964145809412, 'timing/model': 1.0659103910438716, 'learning_rate/qwen_vl_interface': 2.009671586781968e-06, 'learning_rate/action_model': 2.009671586781968e-05, 'epoch': 0.22}) +[2026-06-24 10:56:17,914][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.3149145245552063, 'train/grad_norm_pre_clip': 129.1685791015625, 'timing/data': 0.00040934793651103973, 'timing/model': 1.347100174985826, 'learning_rate/qwen_vl_interface': 2.0059109982121517e-06, 'learning_rate/action_model': 2.005910998212152e-05, 'epoch': 0.22}) +[2026-06-24 10:56:18,978][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.11882691830396652, 'train/grad_norm_pre_clip': 57.18191146850586, 'timing/data': 0.0005096858367323875, 'timing/model': 1.0623452821746469, 'learning_rate/qwen_vl_interface': 2.002154216761474e-06, 'learning_rate/action_model': 2.002154216761474e-05, 'epoch': 0.22}) +[2026-06-24 10:56:20,039][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.20878224074840546, 'train/grad_norm_pre_clip': 68.61753845214844, 'timing/data': 0.00030815787613391876, 'timing/model': 1.0594436679966748, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 0.22}) +[2026-06-24 10:56:21,099][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.07063460350036621, 'train/grad_norm_pre_clip': 37.617332458496094, 'timing/data': 0.00030949083156883717, 'timing/model': 1.058743414003402, 'learning_rate/qwen_vl_interface': 1.994652092848234e-06, 'learning_rate/action_model': 1.994652092848234e-05, 'epoch': 0.22}) +[2026-06-24 10:56:22,160][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.04142019897699356, 'train/grad_norm_pre_clip': 22.616247177124023, 'timing/data': 0.00030564400367438793, 'timing/model': 1.0594542080070823, 'learning_rate/qwen_vl_interface': 1.9909067591898306e-06, 'learning_rate/action_model': 1.9909067591898304e-05, 'epoch': 0.22}) +[2026-06-24 10:56:23,224][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.015920229256153107, 'train/grad_norm_pre_clip': 13.112704277038574, 'timing/data': 0.00031564198434352875, 'timing/model': 1.0624429460149258, 'learning_rate/qwen_vl_interface': 1.987165250258875e-06, 'learning_rate/action_model': 1.987165250258875e-05, 'epoch': 0.22}) +[2026-06-24 10:56:24,291][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2250 +[2026-06-24 10:56:49,159][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2250 +[2026-06-24 10:56:49,160][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2250 +[2026-06-24 10:57:20,291][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2250 +[2026-06-24 11:02:29,661][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2250 +[2026-06-24 11:02:29,662][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.18671105802059174, 'train/grad_norm_pre_clip': 53.748443603515625, 'eval/action_loss/seconds': 24.86695897113532, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.061446644365787506, 'eval/latency_2/loss': 0.061446644365787506, 'eval/flappy/loss': 0.061446644365787506, 'eval/flappy/latency_2/loss': 0.061446644365787506, 'eval/flap_precision': 0.9518716577540107, 'eval/flap_recall': 0.966078697421981, 'eval/flap_f1': 0.958922558922559, 'eval/flap_rate_model': 0.20777777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.958922558922559, 'eval/latency_2/flap_precision': 0.9518716577540107, 'eval/latency_2/flap_recall': 0.966078697421981, 'eval/latency_2/flap_f1': 0.958922558922559, 'eval/latency_2/flap_rate_model': 0.20777777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.958922558922559, 'eval/action_classification/data_wait_seconds': 0.29184284806251526, 'eval/action_classification/predict_seconds': 30.747209548950195, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 243.26000000001076, 'rl_games_eval/mid_train/mean_length': 1978.4, 'rl_games_eval/mid_train/std_reward': 199.3218462688027, 'rl_games_eval/mid_train/std_length': 1605.7832481378052, 'rl_games_eval/mid_train/macro_mean_reward': 243.26000000001076, 'rl_games_eval/mid_train/macro_mean_length': 1978.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 243.26000000001076, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1978.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 199.3218462688027, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1605.7832481378052, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 243.26000000001076, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005860400851815939, 'timing/model': 1.0647418010048568, 'learning_rate/qwen_vl_interface': 1.9834275704462388e-06, 'learning_rate/action_model': 1.9834275704462383e-05, 'epoch': 0.22}) +[2026-06-24 11:02:30,711][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.09258035570383072, 'train/grad_norm_pre_clip': 28.44526481628418, 'timing/data': 0.0003806259483098984, 'timing/model': 1.046984262065962, 'learning_rate/qwen_vl_interface': 1.9796937241382967e-06, 'learning_rate/action_model': 1.9796937241382968e-05, 'epoch': 0.22}) +[2026-06-24 11:02:31,762][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.03537992388010025, 'train/grad_norm_pre_clip': 47.079044342041016, 'timing/data': 0.0003189498092979193, 'timing/model': 1.0496255920734257, 'learning_rate/qwen_vl_interface': 1.9759637157169257e-06, 'learning_rate/action_model': 1.9759637157169255e-05, 'epoch': 0.22}) +[2026-06-24 11:02:33,074][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.07158993184566498, 'train/grad_norm_pre_clip': 44.79706954956055, 'timing/data': 0.00031536282040178776, 'timing/model': 1.3104984508827329, 'learning_rate/qwen_vl_interface': 1.9722375495594963e-06, 'learning_rate/action_model': 1.9722375495594964e-05, 'epoch': 0.22}) +[2026-06-24 11:02:34,126][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.10673244297504425, 'train/grad_norm_pre_clip': 52.41619873046875, 'timing/data': 0.00034370715729892254, 'timing/model': 1.0501284289639443, 'learning_rate/qwen_vl_interface': 1.968515230038873e-06, 'learning_rate/action_model': 1.9685152300388733e-05, 'epoch': 0.22}) +[2026-06-24 11:02:35,178][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.054567981511354446, 'train/grad_norm_pre_clip': 27.28265953063965, 'timing/data': 0.00032639410346746445, 'timing/model': 1.0504134770017117, 'learning_rate/qwen_vl_interface': 1.964796761523406e-06, 'learning_rate/action_model': 1.964796761523406e-05, 'epoch': 0.22}) +[2026-06-24 11:02:36,234][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.10559942573308945, 'train/grad_norm_pre_clip': 39.546363830566406, 'timing/data': 0.00030134711414575577, 'timing/model': 1.0541890219319612, 'learning_rate/qwen_vl_interface': 1.961082148376926e-06, 'learning_rate/action_model': 1.961082148376926e-05, 'epoch': 0.22}) +[2026-06-24 11:02:37,290][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.043222878128290176, 'train/grad_norm_pre_clip': 28.13220977783203, 'timing/data': 0.00033280602656304836, 'timing/model': 1.054375423816964, 'learning_rate/qwen_vl_interface': 1.9573713949587353e-06, 'learning_rate/action_model': 1.9573713949587354e-05, 'epoch': 0.22}) +[2026-06-24 11:02:38,346][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.00931750051677227, 'train/grad_norm_pre_clip': 4.757513523101807, 'timing/data': 0.0003561598714441061, 'timing/model': 1.0542286802083254, 'learning_rate/qwen_vl_interface': 1.953664505623612e-06, 'learning_rate/action_model': 1.9536645056236118e-05, 'epoch': 0.22}) +[2026-06-24 11:02:39,402][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.18245381116867065, 'train/grad_norm_pre_clip': 54.10121536254883, 'timing/data': 0.00033224490471184254, 'timing/model': 1.0553047051653266, 'learning_rate/qwen_vl_interface': 1.949961484721797e-06, 'learning_rate/action_model': 1.949961484721797e-05, 'epoch': 0.22}) +[2026-06-24 11:02:40,458][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.04059609770774841, 'train/grad_norm_pre_clip': 25.927534103393555, 'timing/data': 0.0003393390215933323, 'timing/model': 1.0544345600064844, 'learning_rate/qwen_vl_interface': 1.946262336598992e-06, 'learning_rate/action_model': 1.946262336598992e-05, 'epoch': 0.22}) +[2026-06-24 11:02:41,516][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.16275842487812042, 'train/grad_norm_pre_clip': 55.67064666748047, 'timing/data': 0.0003024190664291382, 'timing/model': 1.0556459110230207, 'learning_rate/qwen_vl_interface': 1.9425670655963537e-06, 'learning_rate/action_model': 1.9425670655963538e-05, 'epoch': 0.22}) +[2026-06-24 11:02:42,592][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.04096755385398865, 'train/grad_norm_pre_clip': 12.874961853027344, 'timing/data': 0.0003793740179389715, 'timing/model': 1.0606992889661342, 'learning_rate/qwen_vl_interface': 1.9388756760504858e-06, 'learning_rate/action_model': 1.9388756760504855e-05, 'epoch': 0.22}) +[2026-06-24 11:02:43,650][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.043999411165714264, 'train/grad_norm_pre_clip': 26.425392150878906, 'timing/data': 0.000313368858769536, 'timing/model': 1.0561875130515546, 'learning_rate/qwen_vl_interface': 1.935188172293441e-06, 'learning_rate/action_model': 1.935188172293441e-05, 'epoch': 0.22}) +[2026-06-24 11:02:44,709][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.005659644026309252, 'train/grad_norm_pre_clip': 3.1065778732299805, 'timing/data': 0.0004906030371785164, 'timing/model': 1.0565167241729796, 'learning_rate/qwen_vl_interface': 1.9315045586527103e-06, 'learning_rate/action_model': 1.9315045586527104e-05, 'epoch': 0.22}) +[2026-06-24 11:02:45,767][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.03820926696062088, 'train/grad_norm_pre_clip': 23.668577194213867, 'timing/data': 0.00033034197986125946, 'timing/model': 1.0565978391095996, 'learning_rate/qwen_vl_interface': 1.9278248394512216e-06, 'learning_rate/action_model': 1.9278248394512215e-05, 'epoch': 0.22}) +[2026-06-24 11:02:46,825][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.05144675448536873, 'train/grad_norm_pre_clip': 37.92097854614258, 'timing/data': 0.00033995998091995716, 'timing/model': 1.0564382299780846, 'learning_rate/qwen_vl_interface': 1.924149019007327e-06, 'learning_rate/action_model': 1.924149019007327e-05, 'epoch': 0.22}) +[2026-06-24 11:02:47,882][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.032729897648096085, 'train/grad_norm_pre_clip': 19.858699798583984, 'timing/data': 0.00030357111245393753, 'timing/model': 1.0560618829913437, 'learning_rate/qwen_vl_interface': 1.920477101634808e-06, 'learning_rate/action_model': 1.920477101634808e-05, 'epoch': 0.22}) +[2026-06-24 11:02:49,236][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.07189615070819855, 'train/grad_norm_pre_clip': 86.40074157714844, 'timing/data': 0.00030886102467775345, 'timing/model': 1.352506791939959, 'learning_rate/qwen_vl_interface': 1.9168090916428646e-06, 'learning_rate/action_model': 1.9168090916428645e-05, 'epoch': 0.22}) +[2026-06-24 11:02:50,297][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.07298266142606735, 'train/grad_norm_pre_clip': 56.689151763916016, 'timing/data': 0.0003162750508636236, 'timing/model': 1.0591228429693729, 'learning_rate/qwen_vl_interface': 1.9131449933361116e-06, 'learning_rate/action_model': 1.9131449933361114e-05, 'epoch': 0.22}) +[2026-06-24 11:02:51,355][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.07891667634248734, 'train/grad_norm_pre_clip': 51.23986053466797, 'timing/data': 0.00034846621565520763, 'timing/model': 1.0562311939429492, 'learning_rate/qwen_vl_interface': 1.90948481101457e-06, 'learning_rate/action_model': 1.9094848110145698e-05, 'epoch': 0.22}) +[2026-06-24 11:02:52,415][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.0078125623986125, 'train/grad_norm_pre_clip': 4.531671524047852, 'timing/data': 0.0003249009605497122, 'timing/model': 1.0584679038729519, 'learning_rate/qwen_vl_interface': 1.9058285489736703e-06, 'learning_rate/action_model': 1.9058285489736702e-05, 'epoch': 0.22}) +[2026-06-24 11:02:53,474][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.011201389133930206, 'train/grad_norm_pre_clip': 4.612448692321777, 'timing/data': 0.00032191607169806957, 'timing/model': 1.0574818598106503, 'learning_rate/qwen_vl_interface': 1.902176211504237e-06, 'learning_rate/action_model': 1.902176211504237e-05, 'epoch': 0.22}) +[2026-06-24 11:02:54,534][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.02025255560874939, 'train/grad_norm_pre_clip': 15.16148567199707, 'timing/data': 0.00031590485014021397, 'timing/model': 1.0583922050427645, 'learning_rate/qwen_vl_interface': 1.8985278028924938e-06, 'learning_rate/action_model': 1.8985278028924938e-05, 'epoch': 0.22}) +[2026-06-24 11:02:55,595][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.0439453050494194, 'train/grad_norm_pre_clip': 28.89813804626465, 'timing/data': 0.0003355010412633419, 'timing/model': 1.0588392780628055, 'learning_rate/qwen_vl_interface': 1.8948833274200492e-06, 'learning_rate/action_model': 1.8948833274200493e-05, 'epoch': 0.22}) +[2026-06-24 11:02:56,654][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.15190337598323822, 'train/grad_norm_pre_clip': 86.92546081542969, 'timing/data': 0.0003080188762396574, 'timing/model': 1.0582500509917736, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 0.22}) +[2026-06-24 11:02:57,714][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.01086924597620964, 'train/grad_norm_pre_clip': 12.670504570007324, 'timing/data': 0.0003476941492408514, 'timing/model': 1.057687204098329, 'learning_rate/qwen_vl_interface': 1.8876061929964188e-06, 'learning_rate/action_model': 1.8876061929964185e-05, 'epoch': 0.22}) +[2026-06-24 11:02:58,775][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.0326656848192215, 'train/grad_norm_pre_clip': 22.879541397094727, 'timing/data': 0.0003381059505045414, 'timing/model': 1.059711764100939, 'learning_rate/qwen_vl_interface': 1.883973542585356e-06, 'learning_rate/action_model': 1.8839735425853557e-05, 'epoch': 0.22}) +[2026-06-24 11:02:59,832][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.21501308679580688, 'train/grad_norm_pre_clip': 77.61567687988281, 'timing/data': 0.00031495303846895695, 'timing/model': 1.05534158507362, 'learning_rate/qwen_vl_interface': 1.8803448423938289e-06, 'learning_rate/action_model': 1.8803448423938287e-05, 'epoch': 0.22}) +[2026-06-24 11:03:00,893][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.06123762205243111, 'train/grad_norm_pre_clip': 66.56272888183594, 'timing/data': 0.000309632858261466, 'timing/model': 1.0594136470463127, 'learning_rate/qwen_vl_interface': 1.876720096680317e-06, 'learning_rate/action_model': 1.876720096680317e-05, 'epoch': 0.22}) +[2026-06-24 11:03:01,954][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.03553310036659241, 'train/grad_norm_pre_clip': 18.319988250732422, 'timing/data': 0.00032888888381421566, 'timing/model': 1.0597880738787353, 'learning_rate/qwen_vl_interface': 1.8730993096986652e-06, 'learning_rate/action_model': 1.8730993096986653e-05, 'epoch': 0.22}) +[2026-06-24 11:03:03,017][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.04125402867794037, 'train/grad_norm_pre_clip': 47.4181022644043, 'timing/data': 0.0003332069609314203, 'timing/model': 1.0610601981170475, 'learning_rate/qwen_vl_interface': 1.8694824856980641e-06, 'learning_rate/action_model': 1.8694824856980643e-05, 'epoch': 0.22}) +[2026-06-24 11:03:04,079][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.034370582550764084, 'train/grad_norm_pre_clip': 40.5927848815918, 'timing/data': 0.00031831907108426094, 'timing/model': 1.0602192680817097, 'learning_rate/qwen_vl_interface': 1.8658696289230624e-06, 'learning_rate/action_model': 1.8658696289230622e-05, 'epoch': 0.22}) +[2026-06-24 11:03:05,435][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.01673555001616478, 'train/grad_norm_pre_clip': 13.942839622497559, 'timing/data': 0.0003436461556702852, 'timing/model': 1.3540229429490864, 'learning_rate/qwen_vl_interface': 1.8622607436135449e-06, 'learning_rate/action_model': 1.8622607436135448e-05, 'epoch': 0.22}) +[2026-06-24 11:03:06,495][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.039125941693782806, 'train/grad_norm_pre_clip': 24.143888473510742, 'timing/data': 0.00030056596733629704, 'timing/model': 1.0586851430125535, 'learning_rate/qwen_vl_interface': 1.8586558340047417e-06, 'learning_rate/action_model': 1.8586558340047415e-05, 'epoch': 0.22}) +[2026-06-24 11:03:07,557][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.00176185870077461, 'train/grad_norm_pre_clip': 1.5108295679092407, 'timing/data': 0.0003024700563400984, 'timing/model': 1.06035643001087, 'learning_rate/qwen_vl_interface': 1.8550549043272151e-06, 'learning_rate/action_model': 1.855054904327215e-05, 'epoch': 0.22}) +[2026-06-24 11:03:08,618][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.05515040084719658, 'train/grad_norm_pre_clip': 75.30986785888672, 'timing/data': 0.0003720999229699373, 'timing/model': 1.0593671509996057, 'learning_rate/qwen_vl_interface': 1.8514579588068573e-06, 'learning_rate/action_model': 1.851457958806857e-05, 'epoch': 0.22}) +[2026-06-24 11:03:09,681][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.3153999447822571, 'train/grad_norm_pre_clip': 94.45914459228516, 'timing/data': 0.0003406708128750324, 'timing/model': 1.0614634559024125, 'learning_rate/qwen_vl_interface': 1.8478650016648814e-06, 'learning_rate/action_model': 1.8478650016648812e-05, 'epoch': 0.22}) +[2026-06-24 11:03:10,744][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.08102509379386902, 'train/grad_norm_pre_clip': 118.97442626953125, 'timing/data': 0.000317517202347517, 'timing/model': 1.061337101040408, 'learning_rate/qwen_vl_interface': 1.8442760371178251e-06, 'learning_rate/action_model': 1.844276037117825e-05, 'epoch': 0.22}) +[2026-06-24 11:03:11,806][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.17125174403190613, 'train/grad_norm_pre_clip': 123.83627319335938, 'timing/data': 0.00036041787825524807, 'timing/model': 1.0604708928149194, 'learning_rate/qwen_vl_interface': 1.840691069377537e-06, 'learning_rate/action_model': 1.840691069377537e-05, 'epoch': 0.22}) +[2026-06-24 11:03:12,869][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.14718492329120636, 'train/grad_norm_pre_clip': 51.1987419128418, 'timing/data': 0.00031342008151113987, 'timing/model': 1.0612469450570643, 'learning_rate/qwen_vl_interface': 1.837110102651179e-06, 'learning_rate/action_model': 1.837110102651179e-05, 'epoch': 0.22}) +[2026-06-24 11:03:13,933][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.03204674273729324, 'train/grad_norm_pre_clip': 35.37321090698242, 'timing/data': 0.00030531478114426136, 'timing/model': 1.0626664469018579, 'learning_rate/qwen_vl_interface': 1.8335331411412124e-06, 'learning_rate/action_model': 1.8335331411412124e-05, 'epoch': 0.22}) +[2026-06-24 11:03:14,996][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.15310080349445343, 'train/grad_norm_pre_clip': 66.60478210449219, 'timing/data': 0.0003373941872268915, 'timing/model': 1.0613845880143344, 'learning_rate/qwen_vl_interface': 1.8299601890454011e-06, 'learning_rate/action_model': 1.829960189045401e-05, 'epoch': 0.22}) +[2026-06-24 11:03:16,060][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.2565976679325104, 'train/grad_norm_pre_clip': 175.03892517089844, 'timing/data': 0.00031602499075233936, 'timing/model': 1.0623947300482541, 'learning_rate/qwen_vl_interface': 1.8263912505568028e-06, 'learning_rate/action_model': 1.826391250556803e-05, 'epoch': 0.22}) +[2026-06-24 11:03:17,122][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.03913490101695061, 'train/grad_norm_pre_clip': 38.40846252441406, 'timing/data': 0.00031041400507092476, 'timing/model': 1.060886634979397, 'learning_rate/qwen_vl_interface': 1.822826329863767e-06, 'learning_rate/action_model': 1.822826329863767e-05, 'epoch': 0.22}) +[2026-06-24 11:03:18,187][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.014745335094630718, 'train/grad_norm_pre_clip': 25.743410110473633, 'timing/data': 0.0003945122007280588, 'timing/model': 1.062823629938066, 'learning_rate/qwen_vl_interface': 1.8192654311499275e-06, 'learning_rate/action_model': 1.8192654311499274e-05, 'epoch': 0.22}) +[2026-06-24 11:03:19,251][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.006297962740063667, 'train/grad_norm_pre_clip': 15.659821510314941, 'timing/data': 0.0003176671452820301, 'timing/model': 1.0629133020993322, 'learning_rate/qwen_vl_interface': 1.8157085585941942e-06, 'learning_rate/action_model': 1.815708558594194e-05, 'epoch': 0.22}) +[2026-06-24 11:03:20,524][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.172357976436615, 'train/grad_norm_pre_clip': 98.9981918334961, 'timing/data': 0.000303089851513505, 'timing/model': 1.2715383968316019, 'learning_rate/qwen_vl_interface': 1.812155716370757e-06, 'learning_rate/action_model': 1.812155716370757e-05, 'epoch': 0.22}) +[2026-06-24 11:03:21,586][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.006354084704071283, 'train/grad_norm_pre_clip': 40.133724212646484, 'timing/data': 0.0003096919972449541, 'timing/model': 1.060572971124202, 'learning_rate/qwen_vl_interface': 1.8086069086490748e-06, 'learning_rate/action_model': 1.808606908649075e-05, 'epoch': 0.22}) +[2026-06-24 11:03:22,651][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.2644648551940918, 'train/grad_norm_pre_clip': 45.34514236450195, 'timing/data': 0.0003299410454928875, 'timing/model': 1.0636024030391127, 'learning_rate/qwen_vl_interface': 1.8050621395938712e-06, 'learning_rate/action_model': 1.805062139593871e-05, 'epoch': 0.22}) +[2026-06-24 11:03:23,716][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.013683884404599667, 'train/grad_norm_pre_clip': 16.67574691772461, 'timing/data': 0.00032858806662261486, 'timing/model': 1.0627770419232547, 'learning_rate/qwen_vl_interface': 1.8015214133651295e-06, 'learning_rate/action_model': 1.8015214133651294e-05, 'epoch': 0.22}) +[2026-06-24 11:03:24,780][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.0023704131599515676, 'train/grad_norm_pre_clip': 3.1563491821289062, 'timing/data': 0.00032411003485322, 'timing/model': 1.0628751090262085, 'learning_rate/qwen_vl_interface': 1.7979847341180881e-06, 'learning_rate/action_model': 1.7979847341180882e-05, 'epoch': 0.22}) +[2026-06-24 11:03:25,846][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.06760433316230774, 'train/grad_norm_pre_clip': 48.520286560058594, 'timing/data': 0.00034845503978431225, 'timing/model': 1.063827872974798, 'learning_rate/qwen_vl_interface': 1.7944521060032374e-06, 'learning_rate/action_model': 1.7944521060032374e-05, 'epoch': 0.22}) +[2026-06-24 11:03:26,913][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.15361173450946808, 'train/grad_norm_pre_clip': 99.08472442626953, 'timing/data': 0.0003134799189865589, 'timing/model': 1.0654495561029762, 'learning_rate/qwen_vl_interface': 1.7909235331663145e-06, 'learning_rate/action_model': 1.7909235331663144e-05, 'epoch': 0.22}) +[2026-06-24 11:03:27,980][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.1911100149154663, 'train/grad_norm_pre_clip': 129.41586303710938, 'timing/data': 0.00033235596492886543, 'timing/model': 1.0660452588927, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 0.22}) +[2026-06-24 11:03:29,047][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.09954088926315308, 'train/grad_norm_pre_clip': 55.786170959472656, 'timing/data': 0.0003543470520526171, 'timing/model': 1.064778378000483, 'learning_rate/qwen_vl_interface': 1.7838785698853936e-06, 'learning_rate/action_model': 1.7838785698853937e-05, 'epoch': 0.22}) +[2026-06-24 11:03:30,115][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.3969847559928894, 'train/grad_norm_pre_clip': 105.9458999633789, 'timing/data': 0.0003314430359750986, 'timing/model': 1.0662977278698236, 'learning_rate/qwen_vl_interface': 1.7803621877090518e-06, 'learning_rate/action_model': 1.780362187709052e-05, 'epoch': 0.22}) +[2026-06-24 11:03:31,182][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.01645398698747158, 'train/grad_norm_pre_clip': 19.691423416137695, 'timing/data': 0.0004861049819737673, 'timing/model': 1.0654082968831062, 'learning_rate/qwen_vl_interface': 1.7768498773459406e-06, 'learning_rate/action_model': 1.7768498773459403e-05, 'epoch': 0.22}) +[2026-06-24 11:03:32,251][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.000909355585463345, 'train/grad_norm_pre_clip': 0.6805241703987122, 'timing/data': 0.00033883703872561455, 'timing/model': 1.0669932200107723, 'learning_rate/qwen_vl_interface': 1.773341642917954e-06, 'learning_rate/action_model': 1.773341642917954e-05, 'epoch': 0.22}) +[2026-06-24 11:03:33,320][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.11251691728830338, 'train/grad_norm_pre_clip': 85.67935180664062, 'timing/data': 0.0003008160274475813, 'timing/model': 1.0675973889883608, 'learning_rate/qwen_vl_interface': 1.7698374885421994e-06, 'learning_rate/action_model': 1.7698374885421994e-05, 'epoch': 0.22}) +[2026-06-24 11:03:34,388][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.03618774563074112, 'train/grad_norm_pre_clip': 45.56037521362305, 'timing/data': 0.0003185190726071596, 'timing/model': 1.0669044370297343, 'learning_rate/qwen_vl_interface': 1.7663374183309946e-06, 'learning_rate/action_model': 1.7663374183309944e-05, 'epoch': 0.22}) +[2026-06-24 11:03:35,457][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.078727588057518, 'train/grad_norm_pre_clip': 41.945308685302734, 'timing/data': 0.00036817300133407116, 'timing/model': 1.0666189631447196, 'learning_rate/qwen_vl_interface': 1.7628414363918693e-06, 'learning_rate/action_model': 1.762841436391869e-05, 'epoch': 0.22}) +[2026-06-24 11:03:36,817][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.12382896989583969, 'train/grad_norm_pre_clip': 53.07277297973633, 'timing/data': 0.0003291189204901457, 'timing/model': 1.3584488418418914, 'learning_rate/qwen_vl_interface': 1.7593495468275518e-06, 'learning_rate/action_model': 1.7593495468275517e-05, 'epoch': 0.22}) +[2026-06-24 11:03:37,884][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.22530612349510193, 'train/grad_norm_pre_clip': 87.51654052734375, 'timing/data': 0.0003092619590461254, 'timing/model': 1.0652451219502836, 'learning_rate/qwen_vl_interface': 1.7558617537359722e-06, 'learning_rate/action_model': 1.755861753735972e-05, 'epoch': 0.22}) +[2026-06-24 11:03:38,947][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.026790019124746323, 'train/grad_norm_pre_clip': 38.57505416870117, 'timing/data': 0.0003125770017504692, 'timing/model': 1.061586000956595, 'learning_rate/qwen_vl_interface': 1.7523780612102455e-06, 'learning_rate/action_model': 1.7523780612102456e-05, 'epoch': 0.22}) +[2026-06-24 11:03:40,015][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.11993646621704102, 'train/grad_norm_pre_clip': 93.15687561035156, 'timing/data': 0.000338507117703557, 'timing/model': 1.0663729880470783, 'learning_rate/qwen_vl_interface': 1.7488984733386816e-06, 'learning_rate/action_model': 1.7488984733386816e-05, 'epoch': 0.22}) +[2026-06-24 11:03:41,083][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.00753258028998971, 'train/grad_norm_pre_clip': 7.663671016693115, 'timing/data': 0.0003260031808167696, 'timing/model': 1.0669031569268554, 'learning_rate/qwen_vl_interface': 1.7454229942047705e-06, 'learning_rate/action_model': 1.7454229942047705e-05, 'epoch': 0.22}) +[2026-06-24 11:03:42,151][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.13085824251174927, 'train/grad_norm_pre_clip': 114.95085906982422, 'timing/data': 0.00032888888381421566, 'timing/model': 1.0653386670164764, 'learning_rate/qwen_vl_interface': 1.741951627887181e-06, 'learning_rate/action_model': 1.741951627887181e-05, 'epoch': 0.22}) +[2026-06-24 11:03:43,217][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.0045699914917349815, 'train/grad_norm_pre_clip': 4.259806156158447, 'timing/data': 0.00039149587973952293, 'timing/model': 1.0646509560756385, 'learning_rate/qwen_vl_interface': 1.7384843784597539e-06, 'learning_rate/action_model': 1.7384843784597536e-05, 'epoch': 0.22}) +[2026-06-24 11:03:44,283][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.007822959683835506, 'train/grad_norm_pre_clip': 8.302756309509277, 'timing/data': 0.00033954810351133347, 'timing/model': 1.065048093907535, 'learning_rate/qwen_vl_interface': 1.735021249991498e-06, 'learning_rate/action_model': 1.735021249991498e-05, 'epoch': 0.22}) +[2026-06-24 11:03:45,350][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.0727897360920906, 'train/grad_norm_pre_clip': 44.84687805175781, 'timing/data': 0.0003351811319589615, 'timing/model': 1.064364091027528, 'learning_rate/qwen_vl_interface': 1.7315622465465899e-06, 'learning_rate/action_model': 1.7315622465465896e-05, 'epoch': 0.22}) +[2026-06-24 11:03:46,416][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.1427370011806488, 'train/grad_norm_pre_clip': 107.69254302978516, 'timing/data': 0.0004240879788994789, 'timing/model': 1.0647526439279318, 'learning_rate/qwen_vl_interface': 1.7281073721843615e-06, 'learning_rate/action_model': 1.7281073721843615e-05, 'epoch': 0.22}) +[2026-06-24 11:03:47,484][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.0701294019818306, 'train/grad_norm_pre_clip': 46.99188995361328, 'timing/data': 0.00033434899523854256, 'timing/model': 1.0662349238991737, 'learning_rate/qwen_vl_interface': 1.724656630959302e-06, 'learning_rate/action_model': 1.724656630959302e-05, 'epoch': 0.22}) +[2026-06-24 11:03:48,552][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.020091604441404343, 'train/grad_norm_pre_clip': 34.69258117675781, 'timing/data': 0.0005388439167290926, 'timing/model': 1.0665694659110159, 'learning_rate/qwen_vl_interface': 1.7212100269210461e-06, 'learning_rate/action_model': 1.721210026921046e-05, 'epoch': 0.22}) +[2026-06-24 11:03:49,620][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.08895133435726166, 'train/grad_norm_pre_clip': 93.85446166992188, 'timing/data': 0.0003171060234308243, 'timing/model': 1.0661927291657776, 'learning_rate/qwen_vl_interface': 1.7177675641143765e-06, 'learning_rate/action_model': 1.7177675641143766e-05, 'epoch': 0.22}) +[2026-06-24 11:03:50,688][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.3589164912700653, 'train/grad_norm_pre_clip': 133.8658447265625, 'timing/data': 0.0003140210174024105, 'timing/model': 1.0663758269511163, 'learning_rate/qwen_vl_interface': 1.7143292465792158e-06, 'learning_rate/action_model': 1.7143292465792156e-05, 'epoch': 0.22}) +[2026-06-24 11:03:51,955][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.053154878318309784, 'train/grad_norm_pre_clip': 23.94586181640625, 'timing/data': 0.00032861786894500256, 'timing/model': 1.2654760729055852, 'learning_rate/qwen_vl_interface': 1.7108950783506218e-06, 'learning_rate/action_model': 1.7108950783506216e-05, 'epoch': 0.23}) +[2026-06-24 11:03:53,023][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.15150976181030273, 'train/grad_norm_pre_clip': 155.71656799316406, 'timing/data': 0.00030992296524345875, 'timing/model': 1.0663089139852673, 'learning_rate/qwen_vl_interface': 1.7074650634587812e-06, 'learning_rate/action_model': 1.707465063458781e-05, 'epoch': 0.23}) +[2026-06-24 11:03:54,087][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.03479848429560661, 'train/grad_norm_pre_clip': 65.18982696533203, 'timing/data': 0.0002950849011540413, 'timing/model': 1.0628614930901676, 'learning_rate/qwen_vl_interface': 1.7040392059290055e-06, 'learning_rate/action_model': 1.7040392059290055e-05, 'epoch': 0.23}) +[2026-06-24 11:03:55,159][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.04321818798780441, 'train/grad_norm_pre_clip': 37.42272186279297, 'timing/data': 0.00031929113902151585, 'timing/model': 1.0696710229385644, 'learning_rate/qwen_vl_interface': 1.7006175097817313e-06, 'learning_rate/action_model': 1.700617509781731e-05, 'epoch': 0.23}) +[2026-06-24 11:03:56,227][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.1433446705341339, 'train/grad_norm_pre_clip': 111.07392883300781, 'timing/data': 0.0003374139778316021, 'timing/model': 1.0667755000758916, 'learning_rate/qwen_vl_interface': 1.6971999790325098e-06, 'learning_rate/action_model': 1.6971999790325098e-05, 'epoch': 0.23}) +[2026-06-24 11:03:57,294][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.1164667159318924, 'train/grad_norm_pre_clip': 111.96502685546875, 'timing/data': 0.0002928010653704405, 'timing/model': 1.0655133991967887, 'learning_rate/qwen_vl_interface': 1.6937866176920046e-06, 'learning_rate/action_model': 1.6937866176920046e-05, 'epoch': 0.23}) +[2026-06-24 11:03:58,364][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.27778804302215576, 'train/grad_norm_pre_clip': 91.56791687011719, 'timing/data': 0.00029495498165488243, 'timing/model': 1.0681674929801375, 'learning_rate/qwen_vl_interface': 1.6903774297659826e-06, 'learning_rate/action_model': 1.6903774297659824e-05, 'epoch': 0.23}) +[2026-06-24 11:03:59,432][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.05520499125123024, 'train/grad_norm_pre_clip': 68.5499267578125, 'timing/data': 0.0003385860472917557, 'timing/model': 1.0667413529008627, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 0.23}) +[2026-06-24 11:04:00,499][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.02142566628754139, 'train/grad_norm_pre_clip': 24.580162048339844, 'timing/data': 0.00035731191746890545, 'timing/model': 1.0648965069558471, 'learning_rate/qwen_vl_interface': 1.6835715901559783e-06, 'learning_rate/action_model': 1.683571590155978e-05, 'epoch': 0.23}) +[2026-06-24 11:04:01,566][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.17415402829647064, 'train/grad_norm_pre_clip': 100.2332534790039, 'timing/data': 0.00030268984846770763, 'timing/model': 1.0656895500142127, 'learning_rate/qwen_vl_interface': 1.680174946459029e-06, 'learning_rate/action_model': 1.680174946459029e-05, 'epoch': 0.23}) +[2026-06-24 11:04:02,634][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.11311931908130646, 'train/grad_norm_pre_clip': 92.48466491699219, 'timing/data': 0.00033567100763320923, 'timing/model': 1.0660561721306294, 'learning_rate/qwen_vl_interface': 1.676782492150618e-06, 'learning_rate/action_model': 1.676782492150618e-05, 'epoch': 0.23}) +[2026-06-24 11:04:03,703][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.003465769113972783, 'train/grad_norm_pre_clip': 6.291845321655273, 'timing/data': 0.0005620168522000313, 'timing/model': 1.0675233250949532, 'learning_rate/qwen_vl_interface': 1.6733942312119809e-06, 'learning_rate/action_model': 1.6733942312119807e-05, 'epoch': 0.23}) +[2026-06-24 11:04:04,774][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.014161202125251293, 'train/grad_norm_pre_clip': 9.974613189697266, 'timing/data': 0.0003229479771107435, 'timing/model': 1.0690985200926661, 'learning_rate/qwen_vl_interface': 1.6700101676194289e-06, 'learning_rate/action_model': 1.670010167619429e-05, 'epoch': 0.23}) +[2026-06-24 11:04:05,843][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.2491615265607834, 'train/grad_norm_pre_clip': 77.05506896972656, 'timing/data': 0.00031186616979539394, 'timing/model': 1.0668733071070164, 'learning_rate/qwen_vl_interface': 1.6666303053443502e-06, 'learning_rate/action_model': 1.66663030534435e-05, 'epoch': 0.23}) +[2026-06-24 11:04:06,910][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.046634793281555176, 'train/grad_norm_pre_clip': 49.422733306884766, 'timing/data': 0.00042762397788465023, 'timing/model': 1.0662581529468298, 'learning_rate/qwen_vl_interface': 1.6632546483532024e-06, 'learning_rate/action_model': 1.6632546483532024e-05, 'epoch': 0.23}) +[2026-06-24 11:04:07,978][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.08892671018838882, 'train/grad_norm_pre_clip': 109.71453857421875, 'timing/data': 0.0003004050813615322, 'timing/model': 1.0654702340252697, 'learning_rate/qwen_vl_interface': 1.659883200607505e-06, 'learning_rate/action_model': 1.6598832006075048e-05, 'epoch': 0.23}) +[2026-06-24 11:04:09,043][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.14289017021656036, 'train/grad_norm_pre_clip': 89.43179321289062, 'timing/data': 0.00030027516186237335, 'timing/model': 1.0639149218332022, 'learning_rate/qwen_vl_interface': 1.6565159660638407e-06, 'learning_rate/action_model': 1.6565159660638407e-05, 'epoch': 0.23}) +[2026-06-24 11:04:10,113][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.08030377328395844, 'train/grad_norm_pre_clip': 68.24893188476562, 'timing/data': 0.0003304718993604183, 'timing/model': 1.0682294790167361, 'learning_rate/qwen_vl_interface': 1.6531529486738466e-06, 'learning_rate/action_model': 1.6531529486738466e-05, 'epoch': 0.23}) +[2026-06-24 11:04:11,469][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.018749695271253586, 'train/grad_norm_pre_clip': 15.349030494689941, 'timing/data': 0.0002988819032907486, 'timing/model': 1.354099151911214, 'learning_rate/qwen_vl_interface': 1.6497941523842127e-06, 'learning_rate/action_model': 1.6497941523842124e-05, 'epoch': 0.23}) +[2026-06-24 11:04:12,537][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.0011054897913709283, 'train/grad_norm_pre_clip': 1.2406405210494995, 'timing/data': 0.00030381116084754467, 'timing/model': 1.066788572119549, 'learning_rate/qwen_vl_interface': 1.6464395811366704e-06, 'learning_rate/action_model': 1.6464395811366705e-05, 'epoch': 0.23}) +[2026-06-24 11:04:13,606][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.10287251323461533, 'train/grad_norm_pre_clip': 89.58783721923828, 'timing/data': 0.00033118296414613724, 'timing/model': 1.0671263809781522, 'learning_rate/qwen_vl_interface': 1.6430892388679988e-06, 'learning_rate/action_model': 1.6430892388679987e-05, 'epoch': 0.23}) +[2026-06-24 11:04:14,676][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.021659424528479576, 'train/grad_norm_pre_clip': 26.342031478881836, 'timing/data': 0.00038229883648455143, 'timing/model': 1.0691586958710104, 'learning_rate/qwen_vl_interface': 1.6397431295100091e-06, 'learning_rate/action_model': 1.639743129510009e-05, 'epoch': 0.23}) +[2026-06-24 11:04:15,745][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.00512788025662303, 'train/grad_norm_pre_clip': 8.07625675201416, 'timing/data': 0.00033098296262323856, 'timing/model': 1.066696829162538, 'learning_rate/qwen_vl_interface': 1.6364012569895471e-06, 'learning_rate/action_model': 1.636401256989547e-05, 'epoch': 0.23}) +[2026-06-24 11:04:16,813][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.0004987267893739045, 'train/grad_norm_pre_clip': 0.288097083568573, 'timing/data': 0.0003156231250613928, 'timing/model': 1.0668512620031834, 'learning_rate/qwen_vl_interface': 1.6330636252284888e-06, 'learning_rate/action_model': 1.6330636252284885e-05, 'epoch': 0.23}) +[2026-06-24 11:04:17,879][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.07016950100660324, 'train/grad_norm_pre_clip': 43.909549713134766, 'timing/data': 0.0003613000735640526, 'timing/model': 1.06380096799694, 'learning_rate/qwen_vl_interface': 1.6297302381437273e-06, 'learning_rate/action_model': 1.6297302381437272e-05, 'epoch': 0.23}) +[2026-06-24 11:04:18,946][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.03243168070912361, 'train/grad_norm_pre_clip': 24.17852020263672, 'timing/data': 0.00031511299312114716, 'timing/model': 1.0652724688407034, 'learning_rate/qwen_vl_interface': 1.6264010996471803e-06, 'learning_rate/action_model': 1.6264010996471802e-05, 'epoch': 0.23}) +[2026-06-24 11:04:20,014][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.05194895341992378, 'train/grad_norm_pre_clip': 22.010330200195312, 'timing/data': 0.0003505891654640436, 'timing/model': 1.0671702399849892, 'learning_rate/qwen_vl_interface': 1.6230762136457764e-06, 'learning_rate/action_model': 1.6230762136457763e-05, 'epoch': 0.23}) +[2026-06-24 11:04:21,081][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.08664783090353012, 'train/grad_norm_pre_clip': 122.96095275878906, 'timing/data': 0.00032711494714021683, 'timing/model': 1.064646636834368, 'learning_rate/qwen_vl_interface': 1.6197555840414567e-06, 'learning_rate/action_model': 1.6197555840414566e-05, 'epoch': 0.23}) +[2026-06-24 11:04:22,150][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.0001957364147529006, 'train/grad_norm_pre_clip': 0.12368538230657578, 'timing/data': 0.00030406215228140354, 'timing/model': 1.0676604479085654, 'learning_rate/qwen_vl_interface': 1.6164392147311619e-06, 'learning_rate/action_model': 1.616439214731162e-05, 'epoch': 0.23}) +[2026-06-24 11:04:23,219][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.06018882244825363, 'train/grad_norm_pre_clip': 69.01700592041016, 'timing/data': 0.00031287805177271366, 'timing/model': 1.0676857680082321, 'learning_rate/qwen_vl_interface': 1.6131271096068373e-06, 'learning_rate/action_model': 1.6131271096068373e-05, 'epoch': 0.23}) +[2026-06-24 11:04:24,289][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.4297131299972534, 'train/grad_norm_pre_clip': 57.443695068359375, 'timing/data': 0.00034333602525293827, 'timing/model': 1.0683841051068157, 'learning_rate/qwen_vl_interface': 1.6098192725554243e-06, 'learning_rate/action_model': 1.6098192725554243e-05, 'epoch': 0.23}) +[2026-06-24 11:04:25,360][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.10391981899738312, 'train/grad_norm_pre_clip': 124.33611297607422, 'timing/data': 0.00030698697082698345, 'timing/model': 1.0692014349624515, 'learning_rate/qwen_vl_interface': 1.6065157074588512e-06, 'learning_rate/action_model': 1.606515707458851e-05, 'epoch': 0.23}) +[2026-06-24 11:04:26,427][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.08186228573322296, 'train/grad_norm_pre_clip': 93.50434112548828, 'timing/data': 0.00030346098355948925, 'timing/model': 1.064746669959277, 'learning_rate/qwen_vl_interface': 1.6032164181940385e-06, 'learning_rate/action_model': 1.6032164181940384e-05, 'epoch': 0.23}) +[2026-06-24 11:04:27,497][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.36752083897590637, 'train/grad_norm_pre_clip': 83.75357055664062, 'timing/data': 0.00033713411539793015, 'timing/model': 1.069099238840863, 'learning_rate/qwen_vl_interface': 1.599921408632883e-06, 'learning_rate/action_model': 1.5999214086328828e-05, 'epoch': 0.23}) +[2026-06-24 11:04:28,566][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.04654518514871597, 'train/grad_norm_pre_clip': 30.75177764892578, 'timing/data': 0.00031933095306158066, 'timing/model': 1.0675843029748648, 'learning_rate/qwen_vl_interface': 1.5966306826422625e-06, 'learning_rate/action_model': 1.5966306826422626e-05, 'epoch': 0.23}) +[2026-06-24 11:04:29,882][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.058604754507541656, 'train/grad_norm_pre_clip': 40.02699279785156, 'timing/data': 0.0003188289701938629, 'timing/model': 1.3132168620359153, 'learning_rate/qwen_vl_interface': 1.593344244084029e-06, 'learning_rate/action_model': 1.593344244084029e-05, 'epoch': 0.23}) +[2026-06-24 11:04:30,951][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.06856626272201538, 'train/grad_norm_pre_clip': 50.30504608154297, 'timing/data': 0.00030961213633418083, 'timing/model': 1.0672955340705812, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 0.23}) +[2026-06-24 11:04:32,020][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.05042121186852455, 'train/grad_norm_pre_clip': 106.30786895751953, 'timing/data': 0.00033959909342229366, 'timing/model': 1.0679634630214423, 'learning_rate/qwen_vl_interface': 1.5867842446869603e-06, 'learning_rate/action_model': 1.58678424468696e-05, 'epoch': 0.23}) +[2026-06-24 11:04:33,090][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.19632743299007416, 'train/grad_norm_pre_clip': 127.0092544555664, 'timing/data': 0.00029747909866273403, 'timing/model': 1.0686460200231522, 'learning_rate/qwen_vl_interface': 1.5835106915466487e-06, 'learning_rate/action_model': 1.5835106915466484e-05, 'epoch': 0.23}) +[2026-06-24 11:04:34,161][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.010415333323180676, 'train/grad_norm_pre_clip': 51.492488861083984, 'timing/data': 0.00029695802368223667, 'timing/model': 1.0692593969870359, 'learning_rate/qwen_vl_interface': 1.5802414412357644e-06, 'learning_rate/action_model': 1.5802414412357643e-05, 'epoch': 0.23}) +[2026-06-24 11:04:35,228][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.09987261146306992, 'train/grad_norm_pre_clip': 97.0755386352539, 'timing/data': 0.00030999304726719856, 'timing/model': 1.0653650159947574, 'learning_rate/qwen_vl_interface': 1.5769764975909527e-06, 'learning_rate/action_model': 1.5769764975909526e-05, 'epoch': 0.23}) +[2026-06-24 11:04:36,297][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.32961204648017883, 'train/grad_norm_pre_clip': 85.70851135253906, 'timing/data': 0.0003306420985609293, 'timing/model': 1.0673635380808264, 'learning_rate/qwen_vl_interface': 1.573715864443809e-06, 'learning_rate/action_model': 1.573715864443809e-05, 'epoch': 0.23}) +[2026-06-24 11:04:37,365][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.02921607345342636, 'train/grad_norm_pre_clip': 68.89259338378906, 'timing/data': 0.00030625611543655396, 'timing/model': 1.0663204491138458, 'learning_rate/qwen_vl_interface': 1.5704595456208662e-06, 'learning_rate/action_model': 1.570459545620866e-05, 'epoch': 0.23}) +[2026-06-24 11:04:38,435][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.034947704523801804, 'train/grad_norm_pre_clip': 43.70258712768555, 'timing/data': 0.0003328360617160797, 'timing/model': 1.0679968041367829, 'learning_rate/qwen_vl_interface': 1.567207544943596e-06, 'learning_rate/action_model': 1.567207544943596e-05, 'epoch': 0.23}) +[2026-06-24 11:04:39,504][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.07586045563220978, 'train/grad_norm_pre_clip': 43.38187026977539, 'timing/data': 0.00032738596200942993, 'timing/model': 1.0674329369794577, 'learning_rate/qwen_vl_interface': 1.5639598662284031e-06, 'learning_rate/action_model': 1.563959866228403e-05, 'epoch': 0.23}) +[2026-06-24 11:04:40,573][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.06971164047718048, 'train/grad_norm_pre_clip': 77.9682388305664, 'timing/data': 0.00032046204432845116, 'timing/model': 1.0677706950809807, 'learning_rate/qwen_vl_interface': 1.5607165132866205e-06, 'learning_rate/action_model': 1.5607165132866204e-05, 'epoch': 0.23}) +[2026-06-24 11:04:41,643][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.00011859920050483197, 'train/grad_norm_pre_clip': 0.13889914751052856, 'timing/data': 0.000318238977342844, 'timing/model': 1.0683400901034474, 'learning_rate/qwen_vl_interface': 1.557477489924501e-06, 'learning_rate/action_model': 1.5574774899245008e-05, 'epoch': 0.23}) +[2026-06-24 11:04:42,713][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.2733403444290161, 'train/grad_norm_pre_clip': 65.99095916748047, 'timing/data': 0.00031473184935748577, 'timing/model': 1.068131259875372, 'learning_rate/qwen_vl_interface': 1.5542427999432197e-06, 'learning_rate/action_model': 1.5542427999432195e-05, 'epoch': 0.23}) +[2026-06-24 11:04:43,782][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.24152018129825592, 'train/grad_norm_pre_clip': 126.03215026855469, 'timing/data': 0.0003022078890353441, 'timing/model': 1.0671863870229572, 'learning_rate/qwen_vl_interface': 1.5510124471388675e-06, 'learning_rate/action_model': 1.5510124471388673e-05, 'epoch': 0.23}) +[2026-06-24 11:04:44,850][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.0881151407957077, 'train/grad_norm_pre_clip': 50.33632278442383, 'timing/data': 0.00030672596767544746, 'timing/model': 1.0673322440125048, 'learning_rate/qwen_vl_interface': 1.54778643530244e-06, 'learning_rate/action_model': 1.5477864353024397e-05, 'epoch': 0.23}) +[2026-06-24 11:04:45,921][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.05845395848155022, 'train/grad_norm_pre_clip': 92.22066497802734, 'timing/data': 0.00029699807055294514, 'timing/model': 1.0686450051143765, 'learning_rate/qwen_vl_interface': 1.544564768219845e-06, 'learning_rate/action_model': 1.544564768219845e-05, 'epoch': 0.23}) +[2026-06-24 11:04:46,990][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.0670645609498024, 'train/grad_norm_pre_clip': 99.05979919433594, 'timing/data': 0.00036030798219144344, 'timing/model': 1.0682633011601865, 'learning_rate/qwen_vl_interface': 1.5413474496718852e-06, 'learning_rate/action_model': 1.5413474496718853e-05, 'epoch': 0.23}) +[2026-06-24 11:04:48,270][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.12292990833520889, 'train/grad_norm_pre_clip': 52.630218505859375, 'timing/data': 0.0003333569038659334, 'timing/model': 1.2778550270013511, 'learning_rate/qwen_vl_interface': 1.538134483434265e-06, 'learning_rate/action_model': 1.538134483434265e-05, 'epoch': 0.23}) +[2026-06-24 11:04:49,339][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.1220197007060051, 'train/grad_norm_pre_clip': 58.95267868041992, 'timing/data': 0.0003403499722480774, 'timing/model': 1.067490633111447, 'learning_rate/qwen_vl_interface': 1.5349258732775791e-06, 'learning_rate/action_model': 1.534925873277579e-05, 'epoch': 0.23}) +[2026-06-24 11:04:50,407][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.006763841956853867, 'train/grad_norm_pre_clip': 8.171218872070312, 'timing/data': 0.0003116670995950699, 'timing/model': 1.0666555759962648, 'learning_rate/qwen_vl_interface': 1.531721622967311e-06, 'learning_rate/action_model': 1.531721622967311e-05, 'epoch': 0.23}) +[2026-06-24 11:04:51,478][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.00830711331218481, 'train/grad_norm_pre_clip': 8.119002342224121, 'timing/data': 0.00029951310716569424, 'timing/model': 1.069569098064676, 'learning_rate/qwen_vl_interface': 1.5285217362638246e-06, 'learning_rate/action_model': 1.5285217362638248e-05, 'epoch': 0.23}) +[2026-06-24 11:04:52,548][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.08617238700389862, 'train/grad_norm_pre_clip': 85.34933471679688, 'timing/data': 0.00033926800824701786, 'timing/model': 1.0681495021563023, 'learning_rate/qwen_vl_interface': 1.5253262169223668e-06, 'learning_rate/action_model': 1.5253262169223667e-05, 'epoch': 0.23}) +[2026-06-24 11:04:53,622][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.014624702744185925, 'train/grad_norm_pre_clip': 20.61158561706543, 'timing/data': 0.00029464391991496086, 'timing/model': 1.0721314640250057, 'learning_rate/qwen_vl_interface': 1.5221350686930567e-06, 'learning_rate/action_model': 1.5221350686930565e-05, 'epoch': 0.23}) +[2026-06-24 11:04:54,690][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.0005486419540829957, 'train/grad_norm_pre_clip': 0.36286869645118713, 'timing/data': 0.00037036696448922157, 'timing/model': 1.0659031048417091, 'learning_rate/qwen_vl_interface': 1.5189482953208867e-06, 'learning_rate/action_model': 1.5189482953208866e-05, 'epoch': 0.23}) +[2026-06-24 11:04:55,760][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.12589040398597717, 'train/grad_norm_pre_clip': 57.129676818847656, 'timing/data': 0.0003070179373025894, 'timing/model': 1.0690298478584737, 'learning_rate/qwen_vl_interface': 1.5157659005457117e-06, 'learning_rate/action_model': 1.5157659005457117e-05, 'epoch': 0.23}) +[2026-06-24 11:04:56,833][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.09129123389720917, 'train/grad_norm_pre_clip': 83.91999053955078, 'timing/data': 0.0002997640985995531, 'timing/model': 1.0706109940074384, 'learning_rate/qwen_vl_interface': 1.512587888102247e-06, 'learning_rate/action_model': 1.512587888102247e-05, 'epoch': 0.23}) +[2026-06-24 11:04:57,902][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.25036147236824036, 'train/grad_norm_pre_clip': 103.51691436767578, 'timing/data': 0.00032044295221567154, 'timing/model': 1.0677300509996712, 'learning_rate/qwen_vl_interface': 1.5094142617200691e-06, 'learning_rate/action_model': 1.509414261720069e-05, 'epoch': 0.23}) +[2026-06-24 11:04:58,973][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.05379826948046684, 'train/grad_norm_pre_clip': 45.74748611450195, 'timing/data': 0.0003113949205726385, 'timing/model': 1.0699658880475909, 'learning_rate/qwen_vl_interface': 1.506245025123605e-06, 'learning_rate/action_model': 1.5062450251236049e-05, 'epoch': 0.23}) +[2026-06-24 11:05:00,044][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.08116650581359863, 'train/grad_norm_pre_clip': 45.00685119628906, 'timing/data': 0.000315273180603981, 'timing/model': 1.068981290096417, 'learning_rate/qwen_vl_interface': 1.5030801820321317e-06, 'learning_rate/action_model': 1.5030801820321316e-05, 'epoch': 0.23}) +[2026-06-24 11:05:01,117][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.01540663093328476, 'train/grad_norm_pre_clip': 16.477378845214844, 'timing/data': 0.0003554178401827812, 'timing/model': 1.0710802199319005, 'learning_rate/qwen_vl_interface': 1.4999197361597672e-06, 'learning_rate/action_model': 1.4999197361597673e-05, 'epoch': 0.23}) +[2026-06-24 11:05:02,188][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.12250632792711258, 'train/grad_norm_pre_clip': 67.28316497802734, 'timing/data': 0.00030912202782928944, 'timing/model': 1.0701192771084607, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 0.23}) +[2026-06-24 11:05:03,259][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.10240533202886581, 'train/grad_norm_pre_clip': 87.96048736572266, 'timing/data': 0.0005188859067857265, 'timing/model': 1.0687973138410598, 'learning_rate/qwen_vl_interface': 1.4936120509030374e-06, 'learning_rate/action_model': 1.4936120509030374e-05, 'epoch': 0.23}) +[2026-06-24 11:05:04,333][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.24659262597560883, 'train/grad_norm_pre_clip': 83.37848663330078, 'timing/data': 0.0003017981071025133, 'timing/model': 1.0725196180865169, 'learning_rate/qwen_vl_interface': 1.4904648189210937e-06, 'learning_rate/action_model': 1.4904648189210937e-05, 'epoch': 0.23}) +[2026-06-24 11:05:05,610][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.011292877607047558, 'train/grad_norm_pre_clip': 17.85165786743164, 'timing/data': 0.0003115460276603699, 'timing/model': 1.2750101189594716, 'learning_rate/qwen_vl_interface': 1.4873219989630908e-06, 'learning_rate/action_model': 1.4873219989630907e-05, 'epoch': 0.23}) +[2026-06-24 11:05:06,679][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.1421986073255539, 'train/grad_norm_pre_clip': 74.3248291015625, 'timing/data': 0.000332514988258481, 'timing/model': 1.0672019941266626, 'learning_rate/qwen_vl_interface': 1.4841835947173017e-06, 'learning_rate/action_model': 1.4841835947173016e-05, 'epoch': 0.23}) +[2026-06-24 11:05:07,751][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.08737210929393768, 'train/grad_norm_pre_clip': 68.35118103027344, 'timing/data': 0.0003229470457881689, 'timing/model': 1.070774455089122, 'learning_rate/qwen_vl_interface': 1.4810496098668198e-06, 'learning_rate/action_model': 1.4810496098668197e-05, 'epoch': 0.23}) +[2026-06-24 11:05:08,823][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.3687816560268402, 'train/grad_norm_pre_clip': 90.8638916015625, 'timing/data': 0.0003400088753551245, 'timing/model': 1.0704127978533506, 'learning_rate/qwen_vl_interface': 1.47792004808955e-06, 'learning_rate/action_model': 1.47792004808955e-05, 'epoch': 0.23}) +[2026-06-24 11:05:09,895][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.08432041108608246, 'train/grad_norm_pre_clip': 45.256134033203125, 'timing/data': 0.00032456102780997753, 'timing/model': 1.0699462848715484, 'learning_rate/qwen_vl_interface': 1.4747949130582101e-06, 'learning_rate/action_model': 1.47479491305821e-05, 'epoch': 0.23}) +[2026-06-24 11:05:10,967][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.07387606054544449, 'train/grad_norm_pre_clip': 71.67804718017578, 'timing/data': 0.0003113050479441881, 'timing/model': 1.0704090280923992, 'learning_rate/qwen_vl_interface': 1.4716742084403166e-06, 'learning_rate/action_model': 1.4716742084403165e-05, 'epoch': 0.23}) +[2026-06-24 11:05:12,041][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.5567218661308289, 'train/grad_norm_pre_clip': 52.44294357299805, 'timing/data': 0.0003104440402239561, 'timing/model': 1.0728351338766515, 'learning_rate/qwen_vl_interface': 1.468557937898192e-06, 'learning_rate/action_model': 1.468557937898192e-05, 'epoch': 0.23}) +[2026-06-24 11:05:13,114][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.08316464722156525, 'train/grad_norm_pre_clip': 59.56510543823242, 'timing/data': 0.0004954419564455748, 'timing/model': 1.0711230679880828, 'learning_rate/qwen_vl_interface': 1.4654461050889527e-06, 'learning_rate/action_model': 1.4654461050889526e-05, 'epoch': 0.23}) +[2026-06-24 11:05:14,186][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.18077844381332397, 'train/grad_norm_pre_clip': 101.51026153564453, 'timing/data': 0.0003855449613183737, 'timing/model': 1.070276479003951, 'learning_rate/qwen_vl_interface': 1.4623387136645102e-06, 'learning_rate/action_model': 1.46233871366451e-05, 'epoch': 0.23}) +[2026-06-24 11:05:15,258][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.20408383011817932, 'train/grad_norm_pre_clip': 32.70214080810547, 'timing/data': 0.0003246909473091364, 'timing/model': 1.0706256399862468, 'learning_rate/qwen_vl_interface': 1.4592357672715585e-06, 'learning_rate/action_model': 1.4592357672715584e-05, 'epoch': 0.23}) +[2026-06-24 11:05:16,327][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.001041211886331439, 'train/grad_norm_pre_clip': 0.6241632699966431, 'timing/data': 0.0003297301009297371, 'timing/model': 1.0668689999729395, 'learning_rate/qwen_vl_interface': 1.4561372695515782e-06, 'learning_rate/action_model': 1.4561372695515783e-05, 'epoch': 0.23}) +[2026-06-24 11:05:17,397][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.010087940841913223, 'train/grad_norm_pre_clip': 5.782539367675781, 'timing/data': 0.00030597508884966373, 'timing/model': 1.0688775291200727, 'learning_rate/qwen_vl_interface': 1.4530432241408298e-06, 'learning_rate/action_model': 1.4530432241408298e-05, 'epoch': 0.23}) +[2026-06-24 11:05:18,469][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.1067819893360138, 'train/grad_norm_pre_clip': 59.084388732910156, 'timing/data': 0.0002999140415340662, 'timing/model': 1.0699056971352547, 'learning_rate/qwen_vl_interface': 1.4499536346703476e-06, 'learning_rate/action_model': 1.4499536346703476e-05, 'epoch': 0.23}) +[2026-06-24 11:05:19,541][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.08931097388267517, 'train/grad_norm_pre_clip': 58.13798904418945, 'timing/data': 0.0003162848297506571, 'timing/model': 1.070947750005871, 'learning_rate/qwen_vl_interface': 1.446868504765939e-06, 'learning_rate/action_model': 1.446868504765939e-05, 'epoch': 0.23}) +[2026-06-24 11:05:20,612][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.12315874546766281, 'train/grad_norm_pre_clip': 50.79377746582031, 'timing/data': 0.0002984821330755949, 'timing/model': 1.0692945150658488, 'learning_rate/qwen_vl_interface': 1.4437878380481718e-06, 'learning_rate/action_model': 1.4437878380481717e-05, 'epoch': 0.23}) +[2026-06-24 11:05:21,686][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.008905095979571342, 'train/grad_norm_pre_clip': 8.49032974243164, 'timing/data': 0.00033498997800052166, 'timing/model': 1.0722294270526618, 'learning_rate/qwen_vl_interface': 1.4407116381323827e-06, 'learning_rate/action_model': 1.4407116381323826e-05, 'epoch': 0.23}) +[2026-06-24 11:05:22,965][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.10791189968585968, 'train/grad_norm_pre_clip': 55.8840217590332, 'timing/data': 0.0003352900967001915, 'timing/model': 1.2774703421164304, 'learning_rate/qwen_vl_interface': 1.437639908628663e-06, 'learning_rate/action_model': 1.4376399086286629e-05, 'epoch': 0.23}) +[2026-06-24 11:05:24,038][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.10103140771389008, 'train/grad_norm_pre_clip': 61.08918380737305, 'timing/data': 0.00031685596331954, 'timing/model': 1.070913749979809, 'learning_rate/qwen_vl_interface': 1.43457265314186e-06, 'learning_rate/action_model': 1.43457265314186e-05, 'epoch': 0.23}) +[2026-06-24 11:05:25,109][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.10390422493219376, 'train/grad_norm_pre_clip': 61.71638107299805, 'timing/data': 0.0003351799678057432, 'timing/model': 1.0693993959575891, 'learning_rate/qwen_vl_interface': 1.4315098752715673e-06, 'learning_rate/action_model': 1.4315098752715672e-05, 'epoch': 0.23}) +[2026-06-24 11:05:26,177][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.03420327976346016, 'train/grad_norm_pre_clip': 31.889795303344727, 'timing/data': 0.0003211940638720989, 'timing/model': 1.066600756952539, 'learning_rate/qwen_vl_interface': 1.4284515786121239e-06, 'learning_rate/action_model': 1.4284515786121239e-05, 'epoch': 0.23}) +[2026-06-24 11:05:27,248][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.008880791254341602, 'train/grad_norm_pre_clip': 9.861730575561523, 'timing/data': 0.0003018879797309637, 'timing/model': 1.0688625541515648, 'learning_rate/qwen_vl_interface': 1.4253977667526122e-06, 'learning_rate/action_model': 1.4253977667526122e-05, 'epoch': 0.23}) +[2026-06-24 11:05:28,317][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.21639026701450348, 'train/grad_norm_pre_clip': 65.6808090209961, 'timing/data': 0.00029799994081258774, 'timing/model': 1.068288346985355, 'learning_rate/qwen_vl_interface': 1.4223484432768523e-06, 'learning_rate/action_model': 1.4223484432768521e-05, 'epoch': 0.23}) +[2026-06-24 11:05:29,388][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.06307023763656616, 'train/grad_norm_pre_clip': 50.49580764770508, 'timing/data': 0.00033208495005965233, 'timing/model': 1.0691902169492096, 'learning_rate/qwen_vl_interface': 1.4193036117633941e-06, 'learning_rate/action_model': 1.419303611763394e-05, 'epoch': 0.23}) +[2026-06-24 11:05:30,460][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.052254170179367065, 'train/grad_norm_pre_clip': 49.06074523925781, 'timing/data': 0.00032154517248272896, 'timing/model': 1.0699287499301136, 'learning_rate/qwen_vl_interface': 1.4162632757855157e-06, 'learning_rate/action_model': 1.4162632757855157e-05, 'epoch': 0.23}) +[2026-06-24 11:05:31,532][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.18258577585220337, 'train/grad_norm_pre_clip': 51.39522171020508, 'timing/data': 0.0003063760232180357, 'timing/model': 1.0699807510245591, 'learning_rate/qwen_vl_interface': 1.4132274389112222e-06, 'learning_rate/action_model': 1.413227438911222e-05, 'epoch': 0.23}) +[2026-06-24 11:05:32,604][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.28646886348724365, 'train/grad_norm_pre_clip': 66.84234619140625, 'timing/data': 0.00032754591666162014, 'timing/model': 1.071018256014213, 'learning_rate/qwen_vl_interface': 1.4101961047032367e-06, 'learning_rate/action_model': 1.4101961047032366e-05, 'epoch': 0.23}) +[2026-06-24 11:05:33,678][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.1642003059387207, 'train/grad_norm_pre_clip': 103.65593719482422, 'timing/data': 0.00030513410456478596, 'timing/model': 1.0719532479997724, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 0.23}) +[2026-06-24 11:05:34,753][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.1345657855272293, 'train/grad_norm_pre_clip': 111.36611938476562, 'timing/data': 0.0003268651198595762, 'timing/model': 1.0736639280803502, 'learning_rate/qwen_vl_interface': 1.4041469585106623e-06, 'learning_rate/action_model': 1.4041469585106622e-05, 'epoch': 0.23}) +[2026-06-24 11:05:35,829][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.08984119445085526, 'train/grad_norm_pre_clip': 54.67403793334961, 'timing/data': 0.00031805806793272495, 'timing/model': 1.074171356158331, 'learning_rate/qwen_vl_interface': 1.4011291536250809e-06, 'learning_rate/action_model': 1.4011291536250809e-05, 'epoch': 0.23}) +[2026-06-24 11:05:36,902][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.00233801850117743, 'train/grad_norm_pre_clip': 2.0642685890197754, 'timing/data': 0.0003432550001889467, 'timing/model': 1.0715719230938703, 'learning_rate/qwen_vl_interface': 1.39811586560382e-06, 'learning_rate/action_model': 1.3981158656038198e-05, 'epoch': 0.23}) +[2026-06-24 11:05:37,978][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.000983305973932147, 'train/grad_norm_pre_clip': 1.2370859384536743, 'timing/data': 0.00048615410923957825, 'timing/model': 1.0739884299691767, 'learning_rate/qwen_vl_interface': 1.3951070979831397e-06, 'learning_rate/action_model': 1.3951070979831397e-05, 'epoch': 0.23}) +[2026-06-24 11:05:39,050][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.0012785630533471704, 'train/grad_norm_pre_clip': 1.572458028793335, 'timing/data': 0.0004949811846017838, 'timing/model': 1.0706184059381485, 'learning_rate/qwen_vl_interface': 1.3921028542939977e-06, 'learning_rate/action_model': 1.3921028542939977e-05, 'epoch': 0.23}) +[2026-06-24 11:05:40,387][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.19240310788154602, 'train/grad_norm_pre_clip': 65.68465423583984, 'timing/data': 0.00030484399758279324, 'timing/model': 1.3352156959008425, 'learning_rate/qwen_vl_interface': 1.3891031380620398e-06, 'learning_rate/action_model': 1.3891031380620399e-05, 'epoch': 0.23}) +[2026-06-24 11:05:41,459][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.08258689194917679, 'train/grad_norm_pre_clip': 52.817012786865234, 'timing/data': 0.0003446780610829592, 'timing/model': 1.0705781169235706, 'learning_rate/qwen_vl_interface': 1.3861079528076008e-06, 'learning_rate/action_model': 1.3861079528076007e-05, 'epoch': 0.23}) +[2026-06-24 11:05:42,531][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.011261713691055775, 'train/grad_norm_pre_clip': 12.242796897888184, 'timing/data': 0.00030987197533249855, 'timing/model': 1.069614770123735, 'learning_rate/qwen_vl_interface': 1.3831173020456979e-06, 'learning_rate/action_model': 1.3831173020456977e-05, 'epoch': 0.23}) +[2026-06-24 11:05:43,603][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.0940750315785408, 'train/grad_norm_pre_clip': 50.21894836425781, 'timing/data': 0.00033798511140048504, 'timing/model': 1.0708508491516113, 'learning_rate/qwen_vl_interface': 1.3801311892860266e-06, 'learning_rate/action_model': 1.3801311892860267e-05, 'epoch': 0.24}) +[2026-06-24 11:05:44,675][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.13772985339164734, 'train/grad_norm_pre_clip': 36.98674392700195, 'timing/data': 0.00032655405811965466, 'timing/model': 1.0706574979703873, 'learning_rate/qwen_vl_interface': 1.3771496180329546e-06, 'learning_rate/action_model': 1.3771496180329546e-05, 'epoch': 0.24}) +[2026-06-24 11:05:45,747][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.010631154291331768, 'train/grad_norm_pre_clip': 13.679479598999023, 'timing/data': 0.0003382759168744087, 'timing/model': 1.070370523026213, 'learning_rate/qwen_vl_interface': 1.3741725917855245e-06, 'learning_rate/action_model': 1.3741725917855245e-05, 'epoch': 0.24}) +[2026-06-24 11:05:46,819][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.293222039937973, 'train/grad_norm_pre_clip': 64.45132446289062, 'timing/data': 0.0003044519107788801, 'timing/model': 1.0696077519096434, 'learning_rate/qwen_vl_interface': 1.3712001140374396e-06, 'learning_rate/action_model': 1.3712001140374394e-05, 'epoch': 0.24}) +[2026-06-24 11:05:47,892][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.14704114198684692, 'train/grad_norm_pre_clip': 61.39638900756836, 'timing/data': 0.0003057250287383795, 'timing/model': 1.071511666988954, 'learning_rate/qwen_vl_interface': 1.3682321882770697e-06, 'learning_rate/action_model': 1.3682321882770696e-05, 'epoch': 0.24}) +[2026-06-24 11:05:48,964][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.16623130440711975, 'train/grad_norm_pre_clip': 101.83642578125, 'timing/data': 0.00030305003747344017, 'timing/model': 1.070824058027938, 'learning_rate/qwen_vl_interface': 1.3652688179874415e-06, 'learning_rate/action_model': 1.3652688179874415e-05, 'epoch': 0.24}) +[2026-06-24 11:05:50,037][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.015945587307214737, 'train/grad_norm_pre_clip': 23.073774337768555, 'timing/data': 0.00032424903474748135, 'timing/model': 1.0712174808140844, 'learning_rate/qwen_vl_interface': 1.3623100066462342e-06, 'learning_rate/action_model': 1.362310006646234e-05, 'epoch': 0.24}) +[2026-06-24 11:05:51,110][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.16780702769756317, 'train/grad_norm_pre_clip': 55.9432487487793, 'timing/data': 0.00034701311960816383, 'timing/model': 1.0715114478953183, 'learning_rate/qwen_vl_interface': 1.3593557577257772e-06, 'learning_rate/action_model': 1.359355757725777e-05, 'epoch': 0.24}) +[2026-06-24 11:05:52,182][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.0639386922121048, 'train/grad_norm_pre_clip': 93.78328704833984, 'timing/data': 0.0003324151039123535, 'timing/model': 1.0707030619960278, 'learning_rate/qwen_vl_interface': 1.3564060746930469e-06, 'learning_rate/action_model': 1.3564060746930468e-05, 'epoch': 0.24}) +[2026-06-24 11:05:53,254][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.15317074954509735, 'train/grad_norm_pre_clip': 79.61555480957031, 'timing/data': 0.0003067268989980221, 'timing/model': 1.070643002865836, 'learning_rate/qwen_vl_interface': 1.353460961009662e-06, 'learning_rate/action_model': 1.3534609610096621e-05, 'epoch': 0.24}) +[2026-06-24 11:05:54,328][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.21878382563591003, 'train/grad_norm_pre_clip': 91.31600952148438, 'timing/data': 0.000336403027176857, 'timing/model': 1.0715104979462922, 'learning_rate/qwen_vl_interface': 1.3505204201318748e-06, 'learning_rate/action_model': 1.3505204201318747e-05, 'epoch': 0.24}) +[2026-06-24 11:05:55,398][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.15905945003032684, 'train/grad_norm_pre_clip': 58.91853713989258, 'timing/data': 0.0003597359172999859, 'timing/model': 1.0687146650161594, 'learning_rate/qwen_vl_interface': 1.3475844555105763e-06, 'learning_rate/action_model': 1.3475844555105762e-05, 'epoch': 0.24}) +[2026-06-24 11:05:56,692][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.0078000896610319614, 'train/grad_norm_pre_clip': 7.9599432945251465, 'timing/data': 0.00031236722134053707, 'timing/model': 1.2929205719847232, 'learning_rate/qwen_vl_interface': 1.3446530705912822e-06, 'learning_rate/action_model': 1.3446530705912821e-05, 'epoch': 0.24}) +[2026-06-24 11:05:57,768][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.11982511729001999, 'train/grad_norm_pre_clip': 85.6883773803711, 'timing/data': 0.0003352002240717411, 'timing/model': 1.0738211940042675, 'learning_rate/qwen_vl_interface': 1.341726268814137e-06, 'learning_rate/action_model': 1.341726268814137e-05, 'epoch': 0.24}) +[2026-06-24 11:05:58,841][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.008034349419176579, 'train/grad_norm_pre_clip': 8.251384735107422, 'timing/data': 0.0003988801036030054, 'timing/model': 1.071278987918049, 'learning_rate/qwen_vl_interface': 1.3388040536139066e-06, 'learning_rate/action_model': 1.3388040536139066e-05, 'epoch': 0.24}) +[2026-06-24 11:05:59,913][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.09710229933261871, 'train/grad_norm_pre_clip': 41.901634216308594, 'timing/data': 0.0003654169850051403, 'timing/model': 1.0702886609360576, 'learning_rate/qwen_vl_interface': 1.3358864284199715e-06, 'learning_rate/action_model': 1.3358864284199715e-05, 'epoch': 0.24}) +[2026-06-24 11:06:00,984][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.025694509968161583, 'train/grad_norm_pre_clip': 13.424088478088379, 'timing/data': 0.0003312118351459503, 'timing/model': 1.0696871830150485, 'learning_rate/qwen_vl_interface': 1.3329733966563282e-06, 'learning_rate/action_model': 1.3329733966563281e-05, 'epoch': 0.24}) +[2026-06-24 11:06:02,058][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.03083188273012638, 'train/grad_norm_pre_clip': 32.32469177246094, 'timing/data': 0.0005721659399569035, 'timing/model': 1.0717422009911388, 'learning_rate/qwen_vl_interface': 1.3300649617415814e-06, 'learning_rate/action_model': 1.3300649617415814e-05, 'epoch': 0.24}) +[2026-06-24 11:06:03,131][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.052671004086732864, 'train/grad_norm_pre_clip': 24.284229278564453, 'timing/data': 0.00032158405520021915, 'timing/model': 1.0712679629214108, 'learning_rate/qwen_vl_interface': 1.3271611270889434e-06, 'learning_rate/action_model': 1.3271611270889434e-05, 'epoch': 0.24}) +[2026-06-24 11:06:04,203][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.0031761913560330868, 'train/grad_norm_pre_clip': 5.965625286102295, 'timing/data': 0.0003118971362709999, 'timing/model': 1.0705502789933234, 'learning_rate/qwen_vl_interface': 1.324261896106223e-06, 'learning_rate/action_model': 1.324261896106223e-05, 'epoch': 0.24}) +[2026-06-24 11:06:05,275][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.08706997334957123, 'train/grad_norm_pre_clip': 53.842376708984375, 'timing/data': 0.00033534993417561054, 'timing/model': 1.0706579259131104, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 0.24}) +[2026-06-24 11:06:06,348][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.017668286338448524, 'train/grad_norm_pre_clip': 19.892017364501953, 'timing/data': 0.0003294600173830986, 'timing/model': 1.070928748929873, 'learning_rate/qwen_vl_interface': 1.3184772587547704e-06, 'learning_rate/action_model': 1.3184772587547704e-05, 'epoch': 0.24}) +[2026-06-24 11:06:07,420][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.1518872231245041, 'train/grad_norm_pre_clip': 80.0084228515625, 'timing/data': 0.00030352012254297733, 'timing/model': 1.0703282540198416, 'learning_rate/qwen_vl_interface': 1.3155918591746319e-06, 'learning_rate/action_model': 1.3155918591746319e-05, 'epoch': 0.24}) +[2026-06-24 11:06:08,493][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.30607739090919495, 'train/grad_norm_pre_clip': 48.912235260009766, 'timing/data': 0.0003239288926124573, 'timing/model': 1.071526451036334, 'learning_rate/qwen_vl_interface': 1.3127110768415915e-06, 'learning_rate/action_model': 1.3127110768415915e-05, 'epoch': 0.24}) +[2026-06-24 11:06:09,565][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.02005445957183838, 'train/grad_norm_pre_clip': 14.525781631469727, 'timing/data': 0.00043058907613158226, 'timing/model': 1.0706494059413671, 'learning_rate/qwen_vl_interface': 1.3098349151364088e-06, 'learning_rate/action_model': 1.3098349151364087e-05, 'epoch': 0.24}) +[2026-06-24 11:06:10,640][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.05770968645811081, 'train/grad_norm_pre_clip': 31.7569580078125, 'timing/data': 0.00031365896575152874, 'timing/model': 1.0734780579805374, 'learning_rate/qwen_vl_interface': 1.3069633774344199e-06, 'learning_rate/action_model': 1.3069633774344198e-05, 'epoch': 0.24}) +[2026-06-24 11:06:11,715][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.03628649562597275, 'train/grad_norm_pre_clip': 19.14128875732422, 'timing/data': 0.00031361007131636143, 'timing/model': 1.0732477700803429, 'learning_rate/qwen_vl_interface': 1.3040964671055343e-06, 'learning_rate/action_model': 1.3040964671055341e-05, 'epoch': 0.24}) +[2026-06-24 11:06:12,786][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.14741434156894684, 'train/grad_norm_pre_clip': 29.461763381958008, 'timing/data': 0.00032482086680829525, 'timing/model': 1.0691047748550773, 'learning_rate/qwen_vl_interface': 1.3012341875142318e-06, 'learning_rate/action_model': 1.3012341875142316e-05, 'epoch': 0.24}) +[2026-06-24 11:06:14,154][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.011974118649959564, 'train/grad_norm_pre_clip': 10.85517692565918, 'timing/data': 0.0005339139606803656, 'timing/model': 1.3660177951678634, 'learning_rate/qwen_vl_interface': 1.2983765420195546e-06, 'learning_rate/action_model': 1.2983765420195546e-05, 'epoch': 0.24}) +[2026-06-24 11:06:15,226][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.03926809877157211, 'train/grad_norm_pre_clip': 37.239620208740234, 'timing/data': 0.00030493317171931267, 'timing/model': 1.070244865026325, 'learning_rate/qwen_vl_interface': 1.295523533975111e-06, 'learning_rate/action_model': 1.295523533975111e-05, 'epoch': 0.24}) +[2026-06-24 11:06:16,297][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.19680696725845337, 'train/grad_norm_pre_clip': 118.65776824951172, 'timing/data': 0.00031290785409510136, 'timing/model': 1.0695240038912743, 'learning_rate/qwen_vl_interface': 1.2926751667290644e-06, 'learning_rate/action_model': 1.2926751667290644e-05, 'epoch': 0.24}) +[2026-06-24 11:06:17,373][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.010971674695611, 'train/grad_norm_pre_clip': 8.572708129882812, 'timing/data': 0.00047129602171480656, 'timing/model': 1.0737696001306176, 'learning_rate/qwen_vl_interface': 1.2898314436241328e-06, 'learning_rate/action_model': 1.2898314436241327e-05, 'epoch': 0.24}) +[2026-06-24 11:06:18,443][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.0038297083228826523, 'train/grad_norm_pre_clip': 2.5441973209381104, 'timing/data': 0.00032676407136023045, 'timing/model': 1.069147335132584, 'learning_rate/qwen_vl_interface': 1.2869923679975813e-06, 'learning_rate/action_model': 1.2869923679975812e-05, 'epoch': 0.24}) +[2026-06-24 11:06:19,516][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.003390502417460084, 'train/grad_norm_pre_clip': 3.3296260833740234, 'timing/data': 0.00031009293161332607, 'timing/model': 1.0708647770807147, 'learning_rate/qwen_vl_interface': 1.2841579431812254e-06, 'learning_rate/action_model': 1.2841579431812253e-05, 'epoch': 0.24}) +[2026-06-24 11:06:20,588][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.02230362594127655, 'train/grad_norm_pre_clip': 25.95773696899414, 'timing/data': 0.0004251000937074423, 'timing/model': 1.0697699841111898, 'learning_rate/qwen_vl_interface': 1.2813281725014196e-06, 'learning_rate/action_model': 1.2813281725014195e-05, 'epoch': 0.24}) +[2026-06-24 11:06:21,659][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.15382982790470123, 'train/grad_norm_pre_clip': 127.83042907714844, 'timing/data': 0.0002989419735968113, 'timing/model': 1.070333407027647, 'learning_rate/qwen_vl_interface': 1.2785030592790596e-06, 'learning_rate/action_model': 1.2785030592790595e-05, 'epoch': 0.24}) +[2026-06-24 11:06:22,734][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.4520517587661743, 'train/grad_norm_pre_clip': 126.28700256347656, 'timing/data': 0.00029506394639611244, 'timing/model': 1.0731735611334443, 'learning_rate/qwen_vl_interface': 1.2756826068295696e-06, 'learning_rate/action_model': 1.2756826068295696e-05, 'epoch': 0.24}) +[2026-06-24 11:06:23,808][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.21718013286590576, 'train/grad_norm_pre_clip': 116.92484283447266, 'timing/data': 0.0003622311633080244, 'timing/model': 1.0716390209272504, 'learning_rate/qwen_vl_interface': 1.2728668184629095e-06, 'learning_rate/action_model': 1.2728668184629094e-05, 'epoch': 0.24}) +[2026-06-24 11:06:24,881][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.14360837638378143, 'train/grad_norm_pre_clip': 31.309141159057617, 'timing/data': 0.00030142697505652905, 'timing/model': 1.0715468220878392, 'learning_rate/qwen_vl_interface': 1.2700556974835644e-06, 'learning_rate/action_model': 1.2700556974835643e-05, 'epoch': 0.24}) +[2026-06-24 11:06:25,955][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.006113090552389622, 'train/grad_norm_pre_clip': 7.36660099029541, 'timing/data': 0.00032825698144733906, 'timing/model': 1.0728830890730023, 'learning_rate/qwen_vl_interface': 1.26724924719054e-06, 'learning_rate/action_model': 1.26724924719054e-05, 'epoch': 0.24}) +[2026-06-24 11:06:27,028][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.020087722688913345, 'train/grad_norm_pre_clip': 14.561441421508789, 'timing/data': 0.0003516310825943947, 'timing/model': 1.0714701358228922, 'learning_rate/qwen_vl_interface': 1.2644474708773651e-06, 'learning_rate/action_model': 1.2644474708773651e-05, 'epoch': 0.24}) +[2026-06-24 11:06:28,101][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.009872587397694588, 'train/grad_norm_pre_clip': 12.708809852600098, 'timing/data': 0.00032418896444141865, 'timing/model': 1.0716996828559786, 'learning_rate/qwen_vl_interface': 1.261650371832076e-06, 'learning_rate/action_model': 1.261650371832076e-05, 'epoch': 0.24}) +[2026-06-24 11:06:29,172][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.21095728874206543, 'train/grad_norm_pre_clip': 36.2374153137207, 'timing/data': 0.0003047629725188017, 'timing/model': 1.0694705261848867, 'learning_rate/qwen_vl_interface': 1.2588579533372256e-06, 'learning_rate/action_model': 1.2588579533372255e-05, 'epoch': 0.24}) +[2026-06-24 11:06:30,450][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.1362786591053009, 'train/grad_norm_pre_clip': 57.54830551147461, 'timing/data': 0.00031900894828140736, 'timing/model': 1.2762196050025523, 'learning_rate/qwen_vl_interface': 1.2560702186698736e-06, 'learning_rate/action_model': 1.2560702186698737e-05, 'epoch': 0.24}) +[2026-06-24 11:06:31,527][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.03477993980050087, 'train/grad_norm_pre_clip': 27.020254135131836, 'timing/data': 0.0003110750112682581, 'timing/model': 1.07566850585863, 'learning_rate/qwen_vl_interface': 1.2532871711015818e-06, 'learning_rate/action_model': 1.2532871711015818e-05, 'epoch': 0.24}) +[2026-06-24 11:06:32,601][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.078343465924263, 'train/grad_norm_pre_clip': 41.3838996887207, 'timing/data': 0.0003022379241883755, 'timing/model': 1.0716532820370048, 'learning_rate/qwen_vl_interface': 1.2505088138984104e-06, 'learning_rate/action_model': 1.2505088138984102e-05, 'epoch': 0.24}) +[2026-06-24 11:06:33,673][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.1180223822593689, 'train/grad_norm_pre_clip': 44.11095428466797, 'timing/data': 0.00037075718864798546, 'timing/model': 1.0706579149700701, 'learning_rate/qwen_vl_interface': 1.247735150320917e-06, 'learning_rate/action_model': 1.247735150320917e-05, 'epoch': 0.24}) +[2026-06-24 11:06:34,749][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.05008985474705696, 'train/grad_norm_pre_clip': 58.19545364379883, 'timing/data': 0.00039823795668780804, 'timing/model': 1.0738150451797992, 'learning_rate/qwen_vl_interface': 1.2449661836241502e-06, 'learning_rate/action_model': 1.2449661836241502e-05, 'epoch': 0.24}) +[2026-06-24 11:06:35,821][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.057468462735414505, 'train/grad_norm_pre_clip': 48.69587707519531, 'timing/data': 0.00041171396151185036, 'timing/model': 1.0703870251309127, 'learning_rate/qwen_vl_interface': 1.2422019170576478e-06, 'learning_rate/action_model': 1.2422019170576477e-05, 'epoch': 0.24}) +[2026-06-24 11:06:36,893][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.0034422951284796, 'train/grad_norm_pre_clip': 3.6171839237213135, 'timing/data': 0.00033906707540154457, 'timing/model': 1.0706393939908594, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 0.24}) +[2026-06-24 11:06:37,966][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.1776289939880371, 'train/grad_norm_pre_clip': 79.4242935180664, 'timing/data': 0.0003574620932340622, 'timing/model': 1.071092741098255, 'learning_rate/qwen_vl_interface': 1.2366874972859958e-06, 'learning_rate/action_model': 1.2366874972859957e-05, 'epoch': 0.24}) +[2026-06-24 11:06:39,039][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.0009269971051253378, 'train/grad_norm_pre_clip': 0.5862944722175598, 'timing/data': 0.0003373140934854746, 'timing/model': 1.0715117729268968, 'learning_rate/qwen_vl_interface': 1.2339373505523245e-06, 'learning_rate/action_model': 1.2339373505523244e-05, 'epoch': 0.24}) +[2026-06-24 11:06:40,113][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.11496172845363617, 'train/grad_norm_pre_clip': 39.07537078857422, 'timing/data': 0.00036933389492332935, 'timing/model': 1.0721817470621318, 'learning_rate/qwen_vl_interface': 1.2311919168918665e-06, 'learning_rate/action_model': 1.2311919168918665e-05, 'epoch': 0.24}) +[2026-06-24 11:06:41,187][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.2163924276828766, 'train/grad_norm_pre_clip': 94.70386505126953, 'timing/data': 0.0003271449822932482, 'timing/model': 1.07186168897897, 'learning_rate/qwen_vl_interface': 1.2284511995265424e-06, 'learning_rate/action_model': 1.2284511995265425e-05, 'epoch': 0.24}) +[2026-06-24 11:06:42,261][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.11057089269161224, 'train/grad_norm_pre_clip': 65.17315673828125, 'timing/data': 0.00033468916080892086, 'timing/model': 1.072732771979645, 'learning_rate/qwen_vl_interface': 1.2257152016727343e-06, 'learning_rate/action_model': 1.2257152016727341e-05, 'epoch': 0.24}) +[2026-06-24 11:06:43,334][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.07208775728940964, 'train/grad_norm_pre_clip': 51.60679244995117, 'timing/data': 0.0003619601484388113, 'timing/model': 1.0710654219146818, 'learning_rate/qwen_vl_interface': 1.2229839265412893e-06, 'learning_rate/action_model': 1.2229839265412893e-05, 'epoch': 0.24}) +[2026-06-24 11:06:44,406][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.3729819059371948, 'train/grad_norm_pre_clip': 84.77677154541016, 'timing/data': 0.00033904705196619034, 'timing/model': 1.0703946151770651, 'learning_rate/qwen_vl_interface': 1.2202573773375107e-06, 'learning_rate/action_model': 1.2202573773375106e-05, 'epoch': 0.24}) +[2026-06-24 11:06:45,478][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.05361298844218254, 'train/grad_norm_pre_clip': 37.210426330566406, 'timing/data': 0.0003255519550293684, 'timing/model': 1.0710243238136172, 'learning_rate/qwen_vl_interface': 1.2175355572611569e-06, 'learning_rate/action_model': 1.217535557261157e-05, 'epoch': 0.24}) +[2026-06-24 11:06:46,763][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.12588396668434143, 'train/grad_norm_pre_clip': 77.29065704345703, 'timing/data': 0.00030768802389502525, 'timing/model': 1.2829912269953638, 'learning_rate/qwen_vl_interface': 1.2148184695064342e-06, 'learning_rate/action_model': 1.2148184695064342e-05, 'epoch': 0.24}) +[2026-06-24 11:06:47,838][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.09537491202354431, 'train/grad_norm_pre_clip': 57.35104751586914, 'timing/data': 0.00030859000980854034, 'timing/model': 1.0728504860308021, 'learning_rate/qwen_vl_interface': 1.212106117261996e-06, 'learning_rate/action_model': 1.212106117261996e-05, 'epoch': 0.24}) +[2026-06-24 11:06:48,910][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.03376394137740135, 'train/grad_norm_pre_clip': 43.32929992675781, 'timing/data': 0.000407105078920722, 'timing/model': 1.0705342961009592, 'learning_rate/qwen_vl_interface': 1.2093985037109385e-06, 'learning_rate/action_model': 1.2093985037109385e-05, 'epoch': 0.24}) +[2026-06-24 11:06:49,986][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.200973242521286, 'train/grad_norm_pre_clip': 48.251991271972656, 'timing/data': 0.00033876695670187473, 'timing/model': 1.0744810209143907, 'learning_rate/qwen_vl_interface': 1.2066956320307988e-06, 'learning_rate/action_model': 1.2066956320307987e-05, 'epoch': 0.24}) +[2026-06-24 11:06:51,060][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.13724417984485626, 'train/grad_norm_pre_clip': 114.09674072265625, 'timing/data': 0.00032199500128626823, 'timing/model': 1.0722372871823609, 'learning_rate/qwen_vl_interface': 1.2039975053935467e-06, 'learning_rate/action_model': 1.2039975053935468e-05, 'epoch': 0.24}) +[2026-06-24 11:06:52,136][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.13988493382930756, 'train/grad_norm_pre_clip': 77.72407531738281, 'timing/data': 0.0003271151799708605, 'timing/model': 1.074688188964501, 'learning_rate/qwen_vl_interface': 1.2013041269655837e-06, 'learning_rate/action_model': 1.2013041269655836e-05, 'epoch': 0.24}) +[2026-06-24 11:06:53,209][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.04171730577945709, 'train/grad_norm_pre_clip': 39.14115524291992, 'timing/data': 0.00038523413240909576, 'timing/model': 1.071340851020068, 'learning_rate/qwen_vl_interface': 1.198615499907739e-06, 'learning_rate/action_model': 1.198615499907739e-05, 'epoch': 0.24}) +[2026-06-24 11:06:54,284][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.010664071887731552, 'train/grad_norm_pre_clip': 11.888919830322266, 'timing/data': 0.00031820801086723804, 'timing/model': 1.073248271131888, 'learning_rate/qwen_vl_interface': 1.1959316273752664e-06, 'learning_rate/action_model': 1.1959316273752663e-05, 'epoch': 0.24}) +[2026-06-24 11:06:55,358][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.009923700243234634, 'train/grad_norm_pre_clip': 12.01603889465332, 'timing/data': 0.0003264241386204958, 'timing/model': 1.0728319231420755, 'learning_rate/qwen_vl_interface': 1.1932525125178416e-06, 'learning_rate/action_model': 1.1932525125178415e-05, 'epoch': 0.24}) +[2026-06-24 11:06:56,430][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.04773375764489174, 'train/grad_norm_pre_clip': 65.40483856201172, 'timing/data': 0.000329919857904315, 'timing/model': 1.0708447420038283, 'learning_rate/qwen_vl_interface': 1.1905781584795526e-06, 'learning_rate/action_model': 1.1905781584795525e-05, 'epoch': 0.24}) +[2026-06-24 11:06:57,503][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.037029512226581573, 'train/grad_norm_pre_clip': 27.022716522216797, 'timing/data': 0.00033583096228539944, 'timing/model': 1.0706175391096622, 'learning_rate/qwen_vl_interface': 1.1879085683989047e-06, 'learning_rate/action_model': 1.1879085683989048e-05, 'epoch': 0.24}) +[2026-06-24 11:06:58,575][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.028437718749046326, 'train/grad_norm_pre_clip': 31.54111099243164, 'timing/data': 0.0003226760309189558, 'timing/model': 1.0708462249021977, 'learning_rate/qwen_vl_interface': 1.1852437454088093e-06, 'learning_rate/action_model': 1.1852437454088091e-05, 'epoch': 0.24}) +[2026-06-24 11:06:59,649][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.06111859530210495, 'train/grad_norm_pre_clip': 27.403270721435547, 'timing/data': 0.00030459207482635975, 'timing/model': 1.0724741001613438, 'learning_rate/qwen_vl_interface': 1.1825836926365845e-06, 'learning_rate/action_model': 1.1825836926365844e-05, 'epoch': 0.24}) +[2026-06-24 11:07:00,724][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2500 +[2026-06-24 11:07:25,093][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2500 +[2026-06-24 11:07:25,125][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2500 +[2026-06-24 11:07:56,734][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2500 +[2026-06-24 11:13:57,639][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-06-24 11:13:57,640][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.03571353107690811, 'train/grad_norm_pre_clip': 28.718652725219727, 'eval/action_loss/seconds': 24.36698606889695, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.07106664776802063, 'eval/latency_2/loss': 0.07106664776802063, 'eval/flappy/loss': 0.07106664776802063, 'eval/flappy/latency_2/loss': 0.07106664776802063, 'eval/flap_precision': 0.9578804347826086, 'eval/flap_recall': 0.9565807327001357, 'eval/flap_f1': 0.9572301425661914, 'eval/flap_rate_model': 0.20444444444444446, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9572301425661913, 'eval/latency_2/flap_precision': 0.9578804347826086, 'eval/latency_2/flap_recall': 0.9565807327001357, 'eval/latency_2/flap_f1': 0.9572301425661914, 'eval/latency_2/flap_rate_model': 0.20444444444444446, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9572301425661913, 'eval/action_classification/data_wait_seconds': 0.30189821124076843, 'eval/action_classification/predict_seconds': 31.22376823425293, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 283.2200000000154, 'rl_games_eval/mid_train/mean_length': 2300.2, 'rl_games_eval/mid_train/std_reward': 198.73168242634333, 'rl_games_eval/mid_train/std_length': 1600.6886517995933, 'rl_games_eval/mid_train/macro_mean_reward': 283.2200000000154, 'rl_games_eval/mid_train/macro_mean_length': 2300.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 283.2200000000154, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2300.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 198.73168242634333, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1600.6886517995933, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 283.2200000000154, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003073981497436762, 'timing/model': 1.0736629290040582, 'learning_rate/qwen_vl_interface': 1.1799284132039522e-06, 'learning_rate/action_model': 1.1799284132039521e-05, 'epoch': 0.24}) +[2026-06-24 11:13:57,640][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_2500_state +[2026-06-24 11:13:57,643][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 11:14:00,169][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_2500_state/model.safetensors +[2026-06-24 11:14:05,740][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_2500_state/optimizer.bin +[2026-06-24 11:14:05,740][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_2500_state/sampler.bin +[2026-06-24 11:14:05,741][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-24 11:14:05,741][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-24 11:14:05,784][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-24 11:14:05,785][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 11:14:05,794][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 11:14:06,845][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.1742880940437317, 'train/grad_norm_pre_clip': 77.24994659423828, 'timing/data': 0.0004448958206921816, 'timing/model': 1.0491931098513305, 'learning_rate/qwen_vl_interface': 1.177277910227028e-06, 'learning_rate/action_model': 1.1772779102270278e-05, 'epoch': 0.24}) +[2026-06-24 11:14:07,903][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.21026180684566498, 'train/grad_norm_pre_clip': 103.47972869873047, 'timing/data': 0.00032568187452852726, 'timing/model': 1.0556880980730057, 'learning_rate/qwen_vl_interface': 1.1746321868163263e-06, 'learning_rate/action_model': 1.1746321868163262e-05, 'epoch': 0.24}) +[2026-06-24 11:14:08,958][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.344695508480072, 'train/grad_norm_pre_clip': 148.4129180908203, 'timing/data': 0.0003161639906466007, 'timing/model': 1.0537602449767292, 'learning_rate/qwen_vl_interface': 1.1719912460767511e-06, 'learning_rate/action_model': 1.171991246076751e-05, 'epoch': 0.24}) +[2026-06-24 11:14:10,013][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.28449931740760803, 'train/grad_norm_pre_clip': 66.43264770507812, 'timing/data': 0.00031108385883271694, 'timing/model': 1.0529939031694084, 'learning_rate/qwen_vl_interface': 1.169355091107594e-06, 'learning_rate/action_model': 1.169355091107594e-05, 'epoch': 0.24}) +[2026-06-24 11:14:11,074][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.011899372562766075, 'train/grad_norm_pre_clip': 21.731903076171875, 'timing/data': 0.00031196605414152145, 'timing/model': 1.0592160089872777, 'learning_rate/qwen_vl_interface': 1.1667237250025284e-06, 'learning_rate/action_model': 1.1667237250025285e-05, 'epoch': 0.24}) +[2026-06-24 11:14:12,141][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.10385365784168243, 'train/grad_norm_pre_clip': 50.21662139892578, 'timing/data': 0.0003273049369454384, 'timing/model': 1.065770298941061, 'learning_rate/qwen_vl_interface': 1.1640971508496099e-06, 'learning_rate/action_model': 1.1640971508496099e-05, 'epoch': 0.24}) +[2026-06-24 11:14:13,198][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.07200906425714493, 'train/grad_norm_pre_clip': 78.84089660644531, 'timing/data': 0.00032385787926614285, 'timing/model': 1.0551712971646339, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 0.24}) +[2026-06-24 11:14:14,255][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.11401081830263138, 'train/grad_norm_pre_clip': 111.5207748413086, 'timing/data': 0.0003255109768360853, 'timing/model': 1.0558489970862865, 'learning_rate/qwen_vl_interface': 1.1588583907243081e-06, 'learning_rate/action_model': 1.1588583907243081e-05, 'epoch': 0.24}) +[2026-06-24 11:14:15,610][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.014878876507282257, 'train/grad_norm_pre_clip': 18.753570556640625, 'timing/data': 0.00031527201645076275, 'timing/model': 1.3533685789443552, 'learning_rate/qwen_vl_interface': 1.1562462108999036e-06, 'learning_rate/action_model': 1.1562462108999036e-05, 'epoch': 0.24}) +[2026-06-24 11:14:16,668][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.2900867760181427, 'train/grad_norm_pre_clip': 93.92227935791016, 'timing/data': 0.0003337471280246973, 'timing/model': 1.0560762600507587, 'learning_rate/qwen_vl_interface': 1.1536388353235912e-06, 'learning_rate/action_model': 1.1536388353235911e-05, 'epoch': 0.24}) +[2026-06-24 11:14:17,726][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.032115865498781204, 'train/grad_norm_pre_clip': 35.30120849609375, 'timing/data': 0.00038813892751932144, 'timing/model': 1.0569808571599424, 'learning_rate/qwen_vl_interface': 1.1510362670552719e-06, 'learning_rate/action_model': 1.1510362670552718e-05, 'epoch': 0.24}) +[2026-06-24 11:14:18,784][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.11010158061981201, 'train/grad_norm_pre_clip': 84.22364807128906, 'timing/data': 0.000353554030880332, 'timing/model': 1.0563923090230674, 'learning_rate/qwen_vl_interface': 1.1484385091492054e-06, 'learning_rate/action_model': 1.1484385091492052e-05, 'epoch': 0.24}) +[2026-06-24 11:14:19,845][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.003647266188636422, 'train/grad_norm_pre_clip': 7.981949329376221, 'timing/data': 0.00033104210160672665, 'timing/model': 1.058784140041098, 'learning_rate/qwen_vl_interface': 1.145845564654006e-06, 'learning_rate/action_model': 1.145845564654006e-05, 'epoch': 0.24}) +[2026-06-24 11:14:20,904][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.253833144903183, 'train/grad_norm_pre_clip': 73.45711517333984, 'timing/data': 0.0003799141850322485, 'timing/model': 1.057898212922737, 'learning_rate/qwen_vl_interface': 1.1432574366126355e-06, 'learning_rate/action_model': 1.1432574366126356e-05, 'epoch': 0.24}) +[2026-06-24 11:14:21,965][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.15811626613140106, 'train/grad_norm_pre_clip': 70.37921142578125, 'timing/data': 0.0003711769822984934, 'timing/model': 1.0585279241204262, 'learning_rate/qwen_vl_interface': 1.1406741280624086e-06, 'learning_rate/action_model': 1.1406741280624086e-05, 'epoch': 0.24}) +[2026-06-24 11:14:23,027][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.041178636252880096, 'train/grad_norm_pre_clip': 19.697580337524414, 'timing/data': 0.00031505199149250984, 'timing/model': 1.0604447331279516, 'learning_rate/qwen_vl_interface': 1.138095642034983e-06, 'learning_rate/action_model': 1.1380956420349829e-05, 'epoch': 0.24}) +[2026-06-24 11:14:24,087][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.0691480040550232, 'train/grad_norm_pre_clip': 43.649169921875, 'timing/data': 0.00032428884878754616, 'timing/model': 1.0583440840709955, 'learning_rate/qwen_vl_interface': 1.1355219815563518e-06, 'learning_rate/action_model': 1.1355219815563516e-05, 'epoch': 0.24}) +[2026-06-24 11:14:25,150][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.09581094980239868, 'train/grad_norm_pre_clip': 106.17672729492188, 'timing/data': 0.000517501961439848, 'timing/model': 1.060978861991316, 'learning_rate/qwen_vl_interface': 1.132953149646853e-06, 'learning_rate/action_model': 1.1329531496468528e-05, 'epoch': 0.24}) +[2026-06-24 11:14:26,213][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.1765134334564209, 'train/grad_norm_pre_clip': 74.24398803710938, 'timing/data': 0.0003110840916633606, 'timing/model': 1.0616855560801923, 'learning_rate/qwen_vl_interface': 1.1303891493211508e-06, 'learning_rate/action_model': 1.1303891493211507e-05, 'epoch': 0.24}) +[2026-06-24 11:14:27,275][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.16538017988204956, 'train/grad_norm_pre_clip': 42.467647552490234, 'timing/data': 0.0005300862248986959, 'timing/model': 1.0601788349449635, 'learning_rate/qwen_vl_interface': 1.1278299835882433e-06, 'learning_rate/action_model': 1.1278299835882433e-05, 'epoch': 0.24}) +[2026-06-24 11:14:28,338][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.05439447611570358, 'train/grad_norm_pre_clip': 61.83782196044922, 'timing/data': 0.0005014219786971807, 'timing/model': 1.0612505648750812, 'learning_rate/qwen_vl_interface': 1.1252756554514545e-06, 'learning_rate/action_model': 1.1252756554514545e-05, 'epoch': 0.24}) +[2026-06-24 11:14:29,401][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.003437881590798497, 'train/grad_norm_pre_clip': 4.122446060180664, 'timing/data': 0.0003386661410331726, 'timing/model': 1.0609376560896635, 'learning_rate/qwen_vl_interface': 1.1227261679084316e-06, 'learning_rate/action_model': 1.1227261679084317e-05, 'epoch': 0.24}) +[2026-06-24 11:14:30,759][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.12718652188777924, 'train/grad_norm_pre_clip': 39.347415924072266, 'timing/data': 0.0003344891592860222, 'timing/model': 1.3565851119346917, 'learning_rate/qwen_vl_interface': 1.1201815239511389e-06, 'learning_rate/action_model': 1.120181523951139e-05, 'epoch': 0.24}) +[2026-06-24 11:14:31,822][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.24343180656433105, 'train/grad_norm_pre_clip': 90.62073516845703, 'timing/data': 0.0003795330412685871, 'timing/model': 1.0611385877709836, 'learning_rate/qwen_vl_interface': 1.1176417265658579e-06, 'learning_rate/action_model': 1.1176417265658579e-05, 'epoch': 0.24}) +[2026-06-24 11:14:32,886][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.005894665606319904, 'train/grad_norm_pre_clip': 3.9317548274993896, 'timing/data': 0.0003104028292000294, 'timing/model': 1.0626936168409884, 'learning_rate/qwen_vl_interface': 1.1151067787331832e-06, 'learning_rate/action_model': 1.1151067787331831e-05, 'epoch': 0.24}) +[2026-06-24 11:14:33,950][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.08747435361146927, 'train/grad_norm_pre_clip': 59.959800720214844, 'timing/data': 0.0003093210980296135, 'timing/model': 1.0624997611157596, 'learning_rate/qwen_vl_interface': 1.112576683428016e-06, 'learning_rate/action_model': 1.112576683428016e-05, 'epoch': 0.24}) +[2026-06-24 11:14:35,015][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.06194063276052475, 'train/grad_norm_pre_clip': 50.78825759887695, 'timing/data': 0.0003388868644833565, 'timing/model': 1.062968328827992, 'learning_rate/qwen_vl_interface': 1.1100514436195653e-06, 'learning_rate/action_model': 1.1100514436195652e-05, 'epoch': 0.24}) +[2026-06-24 11:14:36,079][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.182578906416893, 'train/grad_norm_pre_clip': 52.08714294433594, 'timing/data': 0.0003264229744672775, 'timing/model': 1.0625800078269094, 'learning_rate/qwen_vl_interface': 1.1075310622713398e-06, 'learning_rate/action_model': 1.1075310622713398e-05, 'epoch': 0.24}) +[2026-06-24 11:14:37,144][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.22788158059120178, 'train/grad_norm_pre_clip': 75.06562042236328, 'timing/data': 0.00033375690691173077, 'timing/model': 1.0639896630309522, 'learning_rate/qwen_vl_interface': 1.1050155423411474e-06, 'learning_rate/action_model': 1.1050155423411474e-05, 'epoch': 0.24}) +[2026-06-24 11:14:38,212][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.12873683869838715, 'train/grad_norm_pre_clip': 57.39227294921875, 'timing/data': 0.00034038908779621124, 'timing/model': 1.0664628469385207, 'learning_rate/qwen_vl_interface': 1.1025048867810908e-06, 'learning_rate/action_model': 1.1025048867810908e-05, 'epoch': 0.24}) +[2026-06-24 11:14:39,278][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.006182605400681496, 'train/grad_norm_pre_clip': 5.632788181304932, 'timing/data': 0.0003087799996137619, 'timing/model': 1.0643756550271064, 'learning_rate/qwen_vl_interface': 1.099999098537566e-06, 'learning_rate/action_model': 1.099999098537566e-05, 'epoch': 0.24}) +[2026-06-24 11:14:40,340][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.054261524230241776, 'train/grad_norm_pre_clip': 46.269649505615234, 'timing/data': 0.00032632285729050636, 'timing/model': 1.060493028955534, 'learning_rate/qwen_vl_interface': 1.0974981805512523e-06, 'learning_rate/action_model': 1.0974981805512521e-05, 'epoch': 0.24}) +[2026-06-24 11:14:41,408][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.2426060140132904, 'train/grad_norm_pre_clip': 106.81490325927734, 'timing/data': 0.00031982106156647205, 'timing/model': 1.0657766819931567, 'learning_rate/qwen_vl_interface': 1.0950021357571173e-06, 'learning_rate/action_model': 1.0950021357571173e-05, 'epoch': 0.25}) +[2026-06-24 11:14:42,474][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.05428682640194893, 'train/grad_norm_pre_clip': 93.01580047607422, 'timing/data': 0.0003440459258854389, 'timing/model': 1.0650872739497572, 'learning_rate/qwen_vl_interface': 1.0925109670844098e-06, 'learning_rate/action_model': 1.0925109670844097e-05, 'epoch': 0.25}) +[2026-06-24 11:14:43,542][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.015470723621547222, 'train/grad_norm_pre_clip': 8.09937572479248, 'timing/data': 0.00029806001111865044, 'timing/model': 1.0663125240243971, 'learning_rate/qwen_vl_interface': 1.0900246774566534e-06, 'learning_rate/action_model': 1.0900246774566533e-05, 'epoch': 0.25}) +[2026-06-24 11:14:44,610][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.045171771198511124, 'train/grad_norm_pre_clip': 20.555526733398438, 'timing/data': 0.00032339803874492645, 'timing/model': 1.0665651969611645, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 0.25}) +[2026-06-24 11:14:45,893][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.011323651298880577, 'train/grad_norm_pre_clip': 10.319563865661621, 'timing/data': 0.00032600201666355133, 'timing/model': 1.2807734850794077, 'learning_rate/qwen_vl_interface': 1.0850667470014643e-06, 'learning_rate/action_model': 1.0850667470014641e-05, 'epoch': 0.25}) +[2026-06-24 11:14:46,962][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.09031768888235092, 'train/grad_norm_pre_clip': 101.18305206298828, 'timing/data': 0.0003962349146604538, 'timing/model': 1.0675171599723399, 'learning_rate/qwen_vl_interface': 1.0825951119924388e-06, 'learning_rate/action_model': 1.0825951119924389e-05, 'epoch': 0.25}) +[2026-06-24 11:14:48,032][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.06880828738212585, 'train/grad_norm_pre_clip': 32.119720458984375, 'timing/data': 0.0003089511301368475, 'timing/model': 1.0685586740728468, 'learning_rate/qwen_vl_interface': 1.0801283676651732e-06, 'learning_rate/action_model': 1.080128367665173e-05, 'epoch': 0.25}) +[2026-06-24 11:14:49,099][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.013059657998383045, 'train/grad_norm_pre_clip': 8.210923194885254, 'timing/data': 0.0003689529839903116, 'timing/model': 1.0648545240983367, 'learning_rate/qwen_vl_interface': 1.0776665169145304e-06, 'learning_rate/action_model': 1.0776665169145302e-05, 'epoch': 0.25}) +[2026-06-24 11:14:50,167][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.12898971140384674, 'train/grad_norm_pre_clip': 67.061279296875, 'timing/data': 0.00030950200743973255, 'timing/model': 1.0665594548918307, 'learning_rate/qwen_vl_interface': 1.0752095626296272e-06, 'learning_rate/action_model': 1.0752095626296271e-05, 'epoch': 0.25}) +[2026-06-24 11:14:51,236][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.0003962054615840316, 'train/grad_norm_pre_clip': 0.22889196872711182, 'timing/data': 0.00032853707671165466, 'timing/model': 1.0669376128353179, 'learning_rate/qwen_vl_interface': 1.0727575076938376e-06, 'learning_rate/action_model': 1.0727575076938376e-05, 'epoch': 0.25}) +[2026-06-24 11:14:52,304][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.04996781423687935, 'train/grad_norm_pre_clip': 34.19718551635742, 'timing/data': 0.00047465204261243343, 'timing/model': 1.0662959760520607, 'learning_rate/qwen_vl_interface': 1.0703103549847847e-06, 'learning_rate/action_model': 1.0703103549847847e-05, 'epoch': 0.25}) +[2026-06-24 11:14:53,373][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.02167751081287861, 'train/grad_norm_pre_clip': 24.217845916748047, 'timing/data': 0.000312616815790534, 'timing/model': 1.067017318913713, 'learning_rate/qwen_vl_interface': 1.0678681073743394e-06, 'learning_rate/action_model': 1.0678681073743393e-05, 'epoch': 0.25}) +[2026-06-24 11:14:54,443][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.0017318601021543145, 'train/grad_norm_pre_clip': 0.7677063345909119, 'timing/data': 0.0004574200138449669, 'timing/model': 1.0681314601097256, 'learning_rate/qwen_vl_interface': 1.065430767728614e-06, 'learning_rate/action_model': 1.065430767728614e-05, 'epoch': 0.25}) +[2026-06-24 11:14:55,511][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.15150167047977448, 'train/grad_norm_pre_clip': 66.82524108886719, 'timing/data': 0.00031816703267395496, 'timing/model': 1.0666320910677314, 'learning_rate/qwen_vl_interface': 1.062998338907962e-06, 'learning_rate/action_model': 1.062998338907962e-05, 'epoch': 0.25}) +[2026-06-24 11:14:56,578][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.02825988084077835, 'train/grad_norm_pre_clip': 26.359342575073242, 'timing/data': 0.00030902004800736904, 'timing/model': 1.0660965640563518, 'learning_rate/qwen_vl_interface': 1.0605708237669746e-06, 'learning_rate/action_model': 1.0605708237669745e-05, 'epoch': 0.25}) +[2026-06-24 11:14:57,644][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.0031804691534489393, 'train/grad_norm_pre_clip': 3.847581386566162, 'timing/data': 0.00030634505674242973, 'timing/model': 1.0641101701185107, 'learning_rate/qwen_vl_interface': 1.0581482251544757e-06, 'learning_rate/action_model': 1.0581482251544756e-05, 'epoch': 0.25}) +[2026-06-24 11:14:58,713][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.11547508835792542, 'train/grad_norm_pre_clip': 82.72525024414062, 'timing/data': 0.0003210629802197218, 'timing/model': 1.0670636561699212, 'learning_rate/qwen_vl_interface': 1.0557305459135217e-06, 'learning_rate/action_model': 1.0557305459135217e-05, 'epoch': 0.25}) +[2026-06-24 11:14:59,781][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.018092477694153786, 'train/grad_norm_pre_clip': 17.36090087890625, 'timing/data': 0.000313048018142581, 'timing/model': 1.0672251367941499, 'learning_rate/qwen_vl_interface': 1.0533177888813917e-06, 'learning_rate/action_model': 1.0533177888813916e-05, 'epoch': 0.25}) +[2026-06-24 11:15:00,851][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.15485414862632751, 'train/grad_norm_pre_clip': 73.15119171142578, 'timing/data': 0.00032687396742403507, 'timing/model': 1.067818870069459, 'learning_rate/qwen_vl_interface': 1.05090995688959e-06, 'learning_rate/action_model': 1.0509099568895899e-05, 'epoch': 0.25}) +[2026-06-24 11:15:01,920][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.0053286924958229065, 'train/grad_norm_pre_clip': 4.850345134735107, 'timing/data': 0.00030945101752877235, 'timing/model': 1.0673084591981024, 'learning_rate/qwen_vl_interface': 1.0485070527638435e-06, 'learning_rate/action_model': 1.0485070527638436e-05, 'epoch': 0.25}) +[2026-06-24 11:15:02,994][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.044963281601667404, 'train/grad_norm_pre_clip': 19.17333221435547, 'timing/data': 0.0004027059767395258, 'timing/model': 1.0720184249803424, 'learning_rate/qwen_vl_interface': 1.046109079324094e-06, 'learning_rate/action_model': 1.0461090793240937e-05, 'epoch': 0.25}) +[2026-06-24 11:15:04,065][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.19302493333816528, 'train/grad_norm_pre_clip': 101.45974731445312, 'timing/data': 0.00030950107611715794, 'timing/model': 1.0696829750668257, 'learning_rate/qwen_vl_interface': 1.043716039384496e-06, 'learning_rate/action_model': 1.043716039384496e-05, 'epoch': 0.25}) +[2026-06-24 11:15:05,426][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.006318964064121246, 'train/grad_norm_pre_clip': 6.088985919952393, 'timing/data': 0.00032280711457133293, 'timing/model': 1.360156083945185, 'learning_rate/qwen_vl_interface': 1.0413279357534139e-06, 'learning_rate/action_model': 1.0413279357534138e-05, 'epoch': 0.25}) +[2026-06-24 11:15:06,498][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.030769523233175278, 'train/grad_norm_pre_clip': 18.40648078918457, 'timing/data': 0.0003289980813860893, 'timing/model': 1.0700776630546898, 'learning_rate/qwen_vl_interface': 1.038944771233422e-06, 'learning_rate/action_model': 1.0389447712334218e-05, 'epoch': 0.25}) +[2026-06-24 11:15:07,570][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.24285899102687836, 'train/grad_norm_pre_clip': 95.6136703491211, 'timing/data': 0.00037370179779827595, 'timing/model': 1.0703012098092586, 'learning_rate/qwen_vl_interface': 1.036566548621296e-06, 'learning_rate/action_model': 1.036566548621296e-05, 'epoch': 0.25}) +[2026-06-24 11:15:08,640][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.04139568656682968, 'train/grad_norm_pre_clip': 61.05683898925781, 'timing/data': 0.0002937421668320894, 'timing/model': 1.0680934318806976, 'learning_rate/qwen_vl_interface': 1.0341932707080146e-06, 'learning_rate/action_model': 1.0341932707080146e-05, 'epoch': 0.25}) +[2026-06-24 11:15:09,709][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.12219496071338654, 'train/grad_norm_pre_clip': 68.76366424560547, 'timing/data': 0.00042382581159472466, 'timing/model': 1.0674254149198532, 'learning_rate/qwen_vl_interface': 1.031824940278748e-06, 'learning_rate/action_model': 1.031824940278748e-05, 'epoch': 0.25}) +[2026-06-24 11:15:10,779][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.3763659596443176, 'train/grad_norm_pre_clip': 92.8177490234375, 'timing/data': 0.00031345896422863007, 'timing/model': 1.068460697075352, 'learning_rate/qwen_vl_interface': 1.0294615601128667e-06, 'learning_rate/action_model': 1.0294615601128665e-05, 'epoch': 0.25}) +[2026-06-24 11:15:11,849][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.016659630462527275, 'train/grad_norm_pre_clip': 13.063885688781738, 'timing/data': 0.0003153919242322445, 'timing/model': 1.0687657890375704, 'learning_rate/qwen_vl_interface': 1.0271031329839272e-06, 'learning_rate/action_model': 1.0271031329839273e-05, 'epoch': 0.25}) +[2026-06-24 11:15:12,919][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.06436412781476974, 'train/grad_norm_pre_clip': 29.162757873535156, 'timing/data': 0.0003126370720565319, 'timing/model': 1.0679126039613038, 'learning_rate/qwen_vl_interface': 1.0247496616596773e-06, 'learning_rate/action_model': 1.0247496616596772e-05, 'epoch': 0.25}) +[2026-06-24 11:15:13,987][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.09561292082071304, 'train/grad_norm_pre_clip': 75.60041046142578, 'timing/data': 0.00031955004669725895, 'timing/model': 1.0669853398576379, 'learning_rate/qwen_vl_interface': 1.0224011489020455e-06, 'learning_rate/action_model': 1.0224011489020455e-05, 'epoch': 0.25}) +[2026-06-24 11:15:15,055][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.02791598252952099, 'train/grad_norm_pre_clip': 38.31470489501953, 'timing/data': 0.0003978379536420107, 'timing/model': 1.066298948833719, 'learning_rate/qwen_vl_interface': 1.0200575974671412e-06, 'learning_rate/action_model': 1.020057597467141e-05, 'epoch': 0.25}) +[2026-06-24 11:15:16,128][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.1351320892572403, 'train/grad_norm_pre_clip': 32.30008316040039, 'timing/data': 0.00034348503686487675, 'timing/model': 1.0705117899924517, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 0.25}) +[2026-06-24 11:15:17,198][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.149129718542099, 'train/grad_norm_pre_clip': 62.390445709228516, 'timing/data': 0.0004921539220958948, 'timing/model': 1.0684845298528671, 'learning_rate/qwen_vl_interface': 1.015385389560844e-06, 'learning_rate/action_model': 1.015385389560844e-05, 'epoch': 0.25}) +[2026-06-24 11:15:18,268][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.0025785493198782206, 'train/grad_norm_pre_clip': 18.032821655273438, 'timing/data': 0.00031750602647662163, 'timing/model': 1.0688621080480516, 'learning_rate/qwen_vl_interface': 1.013056738572549e-06, 'learning_rate/action_model': 1.013056738572549e-05, 'epoch': 0.25}) +[2026-06-24 11:15:19,339][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.017165830358862877, 'train/grad_norm_pre_clip': 11.687007904052734, 'timing/data': 0.0003158431500196457, 'timing/model': 1.0687372309621423, 'learning_rate/qwen_vl_interface': 1.0107330598731674e-06, 'learning_rate/action_model': 1.0107330598731674e-05, 'epoch': 0.25}) +[2026-06-24 11:15:20,407][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.09316754341125488, 'train/grad_norm_pre_clip': 60.987178802490234, 'timing/data': 0.0003258730284869671, 'timing/model': 1.0668311880435795, 'learning_rate/qwen_vl_interface': 1.0084143561896667e-06, 'learning_rate/action_model': 1.0084143561896666e-05, 'epoch': 0.25}) +[2026-06-24 11:15:21,479][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.05557829886674881, 'train/grad_norm_pre_clip': 64.73969268798828, 'timing/data': 0.00030798884108662605, 'timing/model': 1.0703205408062786, 'learning_rate/qwen_vl_interface': 1.0061006302431746e-06, 'learning_rate/action_model': 1.0061006302431747e-05, 'epoch': 0.25}) +[2026-06-24 11:15:22,549][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.03076322004199028, 'train/grad_norm_pre_clip': 86.76034545898438, 'timing/data': 0.00034179212525486946, 'timing/model': 1.0686291719321162, 'learning_rate/qwen_vl_interface': 1.0037918847489792e-06, 'learning_rate/action_model': 1.0037918847489791e-05, 'epoch': 0.25}) +[2026-06-24 11:15:23,850][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.06117360666394234, 'train/grad_norm_pre_clip': 76.17478942871094, 'timing/data': 0.0003454990219324827, 'timing/model': 1.2990591959096491, 'learning_rate/qwen_vl_interface': 1.0014881224165193e-06, 'learning_rate/action_model': 1.0014881224165194e-05, 'epoch': 0.25}) +[2026-06-24 11:15:24,920][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.34364086389541626, 'train/grad_norm_pre_clip': 124.65376281738281, 'timing/data': 0.0003213940653949976, 'timing/model': 1.0689699479844421, 'learning_rate/qwen_vl_interface': 9.991893459493894e-07, 'learning_rate/action_model': 9.991893459493892e-06, 'epoch': 0.25}) +[2026-06-24 11:15:25,992][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.08312324434518814, 'train/grad_norm_pre_clip': 50.84934997558594, 'timing/data': 0.0003424929454922676, 'timing/model': 1.0701605619397014, 'learning_rate/qwen_vl_interface': 9.968955580453308e-07, 'learning_rate/action_model': 9.968955580453307e-06, 'epoch': 0.25}) +[2026-06-24 11:15:27,061][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.26574215292930603, 'train/grad_norm_pre_clip': 53.609954833984375, 'timing/data': 0.0003125870134681463, 'timing/model': 1.0671063330955803, 'learning_rate/qwen_vl_interface': 9.94606761396232e-07, 'learning_rate/action_model': 9.94606761396232e-06, 'epoch': 0.25}) +[2026-06-24 11:15:28,132][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.010973295196890831, 'train/grad_norm_pre_clip': 11.71845531463623, 'timing/data': 0.00032586208544671535, 'timing/model': 1.0691404619719833, 'learning_rate/qwen_vl_interface': 9.923229586881252e-07, 'learning_rate/action_model': 9.923229586881251e-06, 'epoch': 0.25}) +[2026-06-24 11:15:29,203][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.12164080888032913, 'train/grad_norm_pre_clip': 114.59093475341797, 'timing/data': 0.000497675035148859, 'timing/model': 1.0693152779713273, 'learning_rate/qwen_vl_interface': 9.900441526011767e-07, 'learning_rate/action_model': 9.900441526011766e-06, 'epoch': 0.25}) +[2026-06-24 11:15:30,275][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.07005472481250763, 'train/grad_norm_pre_clip': 79.70320892333984, 'timing/data': 0.0003497968427836895, 'timing/model': 1.0703464129474014, 'learning_rate/qwen_vl_interface': 9.877703458096949e-07, 'learning_rate/action_model': 9.877703458096948e-06, 'epoch': 0.25}) +[2026-06-24 11:15:31,345][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.21986065804958344, 'train/grad_norm_pre_clip': 93.36994171142578, 'timing/data': 0.00033295596949756145, 'timing/model': 1.068401669152081, 'learning_rate/qwen_vl_interface': 9.855015409821173e-07, 'learning_rate/action_model': 9.855015409821173e-06, 'epoch': 0.25}) +[2026-06-24 11:15:32,412][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.059567660093307495, 'train/grad_norm_pre_clip': 66.41226196289062, 'timing/data': 0.0003346391022205353, 'timing/model': 1.065349583979696, 'learning_rate/qwen_vl_interface': 9.832377407810148e-07, 'learning_rate/action_model': 9.832377407810148e-06, 'epoch': 0.25}) +[2026-06-24 11:15:33,482][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.02215777151286602, 'train/grad_norm_pre_clip': 31.472583770751953, 'timing/data': 0.00031979987397789955, 'timing/model': 1.0682003658730537, 'learning_rate/qwen_vl_interface': 9.809789478630804e-07, 'learning_rate/action_model': 9.809789478630804e-06, 'epoch': 0.25}) +[2026-06-24 11:15:34,553][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.4565979242324829, 'train/grad_norm_pre_clip': 70.60348510742188, 'timing/data': 0.0003325750585645437, 'timing/model': 1.0702923440840095, 'learning_rate/qwen_vl_interface': 9.787251648791351e-07, 'learning_rate/action_model': 9.787251648791352e-06, 'epoch': 0.25}) +[2026-06-24 11:15:35,625][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.12371883541345596, 'train/grad_norm_pre_clip': 65.83582305908203, 'timing/data': 0.0003247899003326893, 'timing/model': 1.0698553919792175, 'learning_rate/qwen_vl_interface': 9.764763944741182e-07, 'learning_rate/action_model': 9.764763944741182e-06, 'epoch': 0.25}) +[2026-06-24 11:15:36,698][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.002545418217778206, 'train/grad_norm_pre_clip': 5.0590691566467285, 'timing/data': 0.0003445269539952278, 'timing/model': 1.071107440860942, 'learning_rate/qwen_vl_interface': 9.742326392870866e-07, 'learning_rate/action_model': 9.742326392870865e-06, 'epoch': 0.25}) +[2026-06-24 11:15:37,771][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.005330297164618969, 'train/grad_norm_pre_clip': 8.751821517944336, 'timing/data': 0.0003364419098943472, 'timing/model': 1.0719426991418004, 'learning_rate/qwen_vl_interface': 9.719939019512133e-07, 'learning_rate/action_model': 9.719939019512134e-06, 'epoch': 0.25}) +[2026-06-24 11:15:38,842][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.04839819297194481, 'train/grad_norm_pre_clip': 52.50558090209961, 'timing/data': 0.00033443793654441833, 'timing/model': 1.0691104368306696, 'learning_rate/qwen_vl_interface': 9.697601850937794e-07, 'learning_rate/action_model': 9.697601850937792e-06, 'epoch': 0.25}) +[2026-06-24 11:15:39,912][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.0599311962723732, 'train/grad_norm_pre_clip': 56.66456985473633, 'timing/data': 0.0003417420666664839, 'timing/model': 1.068820449989289, 'learning_rate/qwen_vl_interface': 9.675314913361773e-07, 'learning_rate/action_model': 9.675314913361772e-06, 'epoch': 0.25}) +[2026-06-24 11:15:40,982][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.03114936500787735, 'train/grad_norm_pre_clip': 14.238993644714355, 'timing/data': 0.0003364619333297014, 'timing/model': 1.068490358069539, 'learning_rate/qwen_vl_interface': 9.653078232939033e-07, 'learning_rate/action_model': 9.653078232939033e-06, 'epoch': 0.25}) +[2026-06-24 11:15:42,264][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.012690797448158264, 'train/grad_norm_pre_clip': 14.988110542297363, 'timing/data': 0.00029886094853281975, 'timing/model': 1.2804820819292217, 'learning_rate/qwen_vl_interface': 9.630891835765562e-07, 'learning_rate/action_model': 9.63089183576556e-06, 'epoch': 0.25}) +[2026-06-24 11:15:43,335][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.29919368028640747, 'train/grad_norm_pre_clip': 89.19485473632812, 'timing/data': 0.0003239389043301344, 'timing/model': 1.069295516004786, 'learning_rate/qwen_vl_interface': 9.608755747878312e-07, 'learning_rate/action_model': 9.608755747878312e-06, 'epoch': 0.25}) +[2026-06-24 11:15:44,405][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.07284104824066162, 'train/grad_norm_pre_clip': 48.5428352355957, 'timing/data': 0.0003093609120696783, 'timing/model': 1.0677445218898356, 'learning_rate/qwen_vl_interface': 9.586669995255243e-07, 'learning_rate/action_model': 9.586669995255242e-06, 'epoch': 0.25}) +[2026-06-24 11:15:45,474][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.06949669122695923, 'train/grad_norm_pre_clip': 90.32862091064453, 'timing/data': 0.00029927282594144344, 'timing/model': 1.0683179358020425, 'learning_rate/qwen_vl_interface': 9.56463460381519e-07, 'learning_rate/action_model': 9.56463460381519e-06, 'epoch': 0.25}) +[2026-06-24 11:15:46,543][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.06301397830247879, 'train/grad_norm_pre_clip': 50.83841323852539, 'timing/data': 0.00035000708885490894, 'timing/model': 1.067586479941383, 'learning_rate/qwen_vl_interface': 9.542649599417924e-07, 'learning_rate/action_model': 9.542649599417923e-06, 'epoch': 0.25}) +[2026-06-24 11:15:47,617][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.0619710236787796, 'train/grad_norm_pre_clip': 113.27931213378906, 'timing/data': 0.00033525004982948303, 'timing/model': 1.0715382220223546, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 0.25}) +[2026-06-24 11:15:48,689][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.13759738206863403, 'train/grad_norm_pre_clip': 69.12074279785156, 'timing/data': 0.0002935519441962242, 'timing/model': 1.071085579926148, 'learning_rate/qwen_vl_interface': 9.498830854895102e-07, 'learning_rate/action_model': 9.498830854895102e-06, 'epoch': 0.25}) +[2026-06-24 11:15:49,761][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.3590143620967865, 'train/grad_norm_pre_clip': 79.43675994873047, 'timing/data': 0.00034315488301217556, 'timing/model': 1.0700251960661262, 'learning_rate/qwen_vl_interface': 9.476997166193283e-07, 'learning_rate/action_model': 9.476997166193283e-06, 'epoch': 0.25}) +[2026-06-24 11:15:50,832][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.0009417294058948755, 'train/grad_norm_pre_clip': 0.7117905020713806, 'timing/data': 0.0003165251109749079, 'timing/model': 1.069920456968248, 'learning_rate/qwen_vl_interface': 9.455213967381657e-07, 'learning_rate/action_model': 9.455213967381656e-06, 'epoch': 0.25}) +[2026-06-24 11:15:51,904][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.14302878081798553, 'train/grad_norm_pre_clip': 65.77306365966797, 'timing/data': 0.0005027651786804199, 'timing/model': 1.0691388780251145, 'learning_rate/qwen_vl_interface': 9.433481284024046e-07, 'learning_rate/action_model': 9.433481284024045e-06, 'epoch': 0.25}) +[2026-06-24 11:15:52,974][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.08114553987979889, 'train/grad_norm_pre_clip': 47.436161041259766, 'timing/data': 0.000308037968352437, 'timing/model': 1.0691680200397968, 'learning_rate/qwen_vl_interface': 9.411799141624933e-07, 'learning_rate/action_model': 9.411799141624933e-06, 'epoch': 0.25}) +[2026-06-24 11:15:54,044][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.24795731902122498, 'train/grad_norm_pre_clip': 56.24267578125, 'timing/data': 0.00031694513745605946, 'timing/model': 1.0683556720614433, 'learning_rate/qwen_vl_interface': 9.390167565629548e-07, 'learning_rate/action_model': 9.390167565629547e-06, 'epoch': 0.25}) +[2026-06-24 11:15:55,115][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.025328079238533974, 'train/grad_norm_pre_clip': 25.175880432128906, 'timing/data': 0.00033461814746260643, 'timing/model': 1.0687166079878807, 'learning_rate/qwen_vl_interface': 9.368586581423721e-07, 'learning_rate/action_model': 9.36858658142372e-06, 'epoch': 0.25}) +[2026-06-24 11:15:56,188][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.10691249370574951, 'train/grad_norm_pre_clip': 36.80098342895508, 'timing/data': 0.0003506590146571398, 'timing/model': 1.0715561811812222, 'learning_rate/qwen_vl_interface': 9.347056214333958e-07, 'learning_rate/action_model': 9.347056214333958e-06, 'epoch': 0.25}) +[2026-06-24 11:15:57,260][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.07386766374111176, 'train/grad_norm_pre_clip': 36.63253402709961, 'timing/data': 0.00031889905221760273, 'timing/model': 1.0702946889214218, 'learning_rate/qwen_vl_interface': 9.32557648962735e-07, 'learning_rate/action_model': 9.32557648962735e-06, 'epoch': 0.25}) +[2026-06-24 11:15:58,337][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.05768180266022682, 'train/grad_norm_pre_clip': 45.9590950012207, 'timing/data': 0.0003907838836312294, 'timing/model': 1.0755778511520475, 'learning_rate/qwen_vl_interface': 9.304147432511538e-07, 'learning_rate/action_model': 9.304147432511536e-06, 'epoch': 0.25}) +[2026-06-24 11:15:59,612][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.03361956402659416, 'train/grad_norm_pre_clip': 27.771751403808594, 'timing/data': 0.00029036495834589005, 'timing/model': 1.2736579079646617, 'learning_rate/qwen_vl_interface': 9.282769068134709e-07, 'learning_rate/action_model': 9.282769068134709e-06, 'epoch': 0.25}) +[2026-06-24 11:16:00,686][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.059188809245824814, 'train/grad_norm_pre_clip': 32.109764099121094, 'timing/data': 0.00033768382854759693, 'timing/model': 1.0719575169496238, 'learning_rate/qwen_vl_interface': 9.261441421585572e-07, 'learning_rate/action_model': 9.261441421585572e-06, 'epoch': 0.25}) +[2026-06-24 11:16:01,757][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.10809536278247833, 'train/grad_norm_pre_clip': 62.40671920776367, 'timing/data': 0.0003168450202792883, 'timing/model': 1.0694111720658839, 'learning_rate/qwen_vl_interface': 9.240164517893321e-07, 'learning_rate/action_model': 9.240164517893322e-06, 'epoch': 0.25}) +[2026-06-24 11:16:02,829][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.03458341956138611, 'train/grad_norm_pre_clip': 18.78370475769043, 'timing/data': 0.0003033902030438185, 'timing/model': 1.0710782120004296, 'learning_rate/qwen_vl_interface': 9.218938382027566e-07, 'learning_rate/action_model': 9.218938382027566e-06, 'epoch': 0.25}) +[2026-06-24 11:16:03,900][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.17418169975280762, 'train/grad_norm_pre_clip': 121.03514099121094, 'timing/data': 0.000535445986315608, 'timing/model': 1.0681833520065993, 'learning_rate/qwen_vl_interface': 9.197763038898372e-07, 'learning_rate/action_model': 9.197763038898371e-06, 'epoch': 0.25}) +[2026-06-24 11:16:04,971][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.012094438076019287, 'train/grad_norm_pre_clip': 10.8204345703125, 'timing/data': 0.0003242688253521919, 'timing/model': 1.0701189457904547, 'learning_rate/qwen_vl_interface': 9.176638513356187e-07, 'learning_rate/action_model': 9.176638513356186e-06, 'epoch': 0.25}) +[2026-06-24 11:16:06,043][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.13775506615638733, 'train/grad_norm_pre_clip': 62.23865509033203, 'timing/data': 0.000491051934659481, 'timing/model': 1.0696049979887903, 'learning_rate/qwen_vl_interface': 9.155564830191799e-07, 'learning_rate/action_model': 9.155564830191798e-06, 'epoch': 0.25}) +[2026-06-24 11:16:07,113][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.13962431252002716, 'train/grad_norm_pre_clip': 95.18595886230469, 'timing/data': 0.0003218448255211115, 'timing/model': 1.069138620980084, 'learning_rate/qwen_vl_interface': 9.134542014136374e-07, 'learning_rate/action_model': 9.134542014136374e-06, 'epoch': 0.25}) +[2026-06-24 11:16:08,187][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.016628973186016083, 'train/grad_norm_pre_clip': 9.457128524780273, 'timing/data': 0.0004972338210791349, 'timing/model': 1.0715933400206268, 'learning_rate/qwen_vl_interface': 9.113570089861343e-07, 'learning_rate/action_model': 9.113570089861342e-06, 'epoch': 0.25}) +[2026-06-24 11:16:09,259][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.14151357114315033, 'train/grad_norm_pre_clip': 46.69833755493164, 'timing/data': 0.0003291289322078228, 'timing/model': 1.0703338221646845, 'learning_rate/qwen_vl_interface': 9.09264908197842e-07, 'learning_rate/action_model': 9.092649081978421e-06, 'epoch': 0.25}) +[2026-06-24 11:16:10,335][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.06627508252859116, 'train/grad_norm_pre_clip': 32.06132125854492, 'timing/data': 0.0003568099346011877, 'timing/model': 1.0740645779296756, 'learning_rate/qwen_vl_interface': 9.071779015039591e-07, 'learning_rate/action_model': 9.071779015039591e-06, 'epoch': 0.25}) +[2026-06-24 11:16:11,408][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.010098845697939396, 'train/grad_norm_pre_clip': 11.827180862426758, 'timing/data': 0.0003193898592144251, 'timing/model': 1.0718529119621962, 'learning_rate/qwen_vl_interface': 9.050959913537037e-07, 'learning_rate/action_model': 9.050959913537038e-06, 'epoch': 0.25}) +[2026-06-24 11:16:12,481][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.009475510567426682, 'train/grad_norm_pre_clip': 9.411422729492188, 'timing/data': 0.00035566813312470913, 'timing/model': 1.0708775189705193, 'learning_rate/qwen_vl_interface': 9.030191801903122e-07, 'learning_rate/action_model': 9.030191801903122e-06, 'epoch': 0.25}) +[2026-06-24 11:16:13,553][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.06226214021444321, 'train/grad_norm_pre_clip': 67.38788604736328, 'timing/data': 0.0004533620085567236, 'timing/model': 1.070712526794523, 'learning_rate/qwen_vl_interface': 9.009474704510392e-07, 'learning_rate/action_model': 9.009474704510392e-06, 'epoch': 0.25}) +[2026-06-24 11:16:14,626][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.10538502037525177, 'train/grad_norm_pre_clip': 69.83596801757812, 'timing/data': 0.0003214641474187374, 'timing/model': 1.0714698969386518, 'learning_rate/qwen_vl_interface': 8.988808645671521e-07, 'learning_rate/action_model': 8.988808645671521e-06, 'epoch': 0.25}) +[2026-06-24 11:16:15,696][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.12720942497253418, 'train/grad_norm_pre_clip': 40.39960861206055, 'timing/data': 0.0003514599520713091, 'timing/model': 1.068422753131017, 'learning_rate/qwen_vl_interface': 8.968193649639264e-07, 'learning_rate/action_model': 8.968193649639263e-06, 'epoch': 0.25}) +[2026-06-24 11:16:16,970][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.01884087175130844, 'train/grad_norm_pre_clip': 23.321016311645508, 'timing/data': 0.00034000910818576813, 'timing/model': 1.2722254530526698, 'learning_rate/qwen_vl_interface': 8.947629740606487e-07, 'learning_rate/action_model': 8.947629740606486e-06, 'epoch': 0.25}) +[2026-06-24 11:16:18,042][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.13630738854408264, 'train/grad_norm_pre_clip': 45.99996566772461, 'timing/data': 0.0003008961211889982, 'timing/model': 1.0698538860306144, 'learning_rate/qwen_vl_interface': 8.927116942706054e-07, 'learning_rate/action_model': 8.927116942706054e-06, 'epoch': 0.25}) +[2026-06-24 11:16:19,114][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.04762069880962372, 'train/grad_norm_pre_clip': 44.39512252807617, 'timing/data': 0.00030893017537891865, 'timing/model': 1.0711709088645875, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 0.25}) +[2026-06-24 11:16:20,186][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.12264109402894974, 'train/grad_norm_pre_clip': 61.81930160522461, 'timing/data': 0.0003207828849554062, 'timing/model': 1.0705117180477828, 'learning_rate/qwen_vl_interface': 8.886244776533915e-07, 'learning_rate/action_model': 8.886244776533914e-06, 'epoch': 0.25}) +[2026-06-24 11:16:21,259][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.03335968032479286, 'train/grad_norm_pre_clip': 32.3643684387207, 'timing/data': 0.0003143199719488621, 'timing/model': 1.0708207071293145, 'learning_rate/qwen_vl_interface': 8.865885456227968e-07, 'learning_rate/action_model': 8.865885456227968e-06, 'epoch': 0.25}) +[2026-06-24 11:16:22,333][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.06453846395015717, 'train/grad_norm_pre_clip': 58.51458740234375, 'timing/data': 0.0003177959006279707, 'timing/model': 1.0724848909303546, 'learning_rate/qwen_vl_interface': 8.845577342985852e-07, 'learning_rate/action_model': 8.845577342985852e-06, 'epoch': 0.25}) +[2026-06-24 11:16:23,405][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.01313438918441534, 'train/grad_norm_pre_clip': 26.08552360534668, 'timing/data': 0.00032067298889160156, 'timing/model': 1.0702956169843674, 'learning_rate/qwen_vl_interface': 8.825320460640263e-07, 'learning_rate/action_model': 8.825320460640262e-06, 'epoch': 0.25}) +[2026-06-24 11:16:24,475][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.012909037061035633, 'train/grad_norm_pre_clip': 16.214553833007812, 'timing/data': 0.000327153829857707, 'timing/model': 1.0681479519698769, 'learning_rate/qwen_vl_interface': 8.805114832963792e-07, 'learning_rate/action_model': 8.80511483296379e-06, 'epoch': 0.25}) +[2026-06-24 11:16:25,545][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.1929049789905548, 'train/grad_norm_pre_clip': 60.26102828979492, 'timing/data': 0.00033456902019679546, 'timing/model': 1.068975605070591, 'learning_rate/qwen_vl_interface': 8.78496048366888e-07, 'learning_rate/action_model': 8.78496048366888e-06, 'epoch': 0.25}) +[2026-06-24 11:16:26,619][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.2295202612876892, 'train/grad_norm_pre_clip': 91.90457153320312, 'timing/data': 0.0003752550110220909, 'timing/model': 1.072496686130762, 'learning_rate/qwen_vl_interface': 8.764857436407744e-07, 'learning_rate/action_model': 8.764857436407745e-06, 'epoch': 0.25}) +[2026-06-24 11:16:27,694][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.017774548381567, 'train/grad_norm_pre_clip': 19.676368713378906, 'timing/data': 0.0004439740441739559, 'timing/model': 1.0722127419430763, 'learning_rate/qwen_vl_interface': 8.74480571477245e-07, 'learning_rate/action_model': 8.74480571477245e-06, 'epoch': 0.25}) +[2026-06-24 11:16:28,764][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.0984572246670723, 'train/grad_norm_pre_clip': 41.846309661865234, 'timing/data': 0.00030324910767376423, 'timing/model': 1.0688887960277498, 'learning_rate/qwen_vl_interface': 8.724805342294806e-07, 'learning_rate/action_model': 8.724805342294806e-06, 'epoch': 0.25}) +[2026-06-24 11:16:29,836][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.22637499868869781, 'train/grad_norm_pre_clip': 182.3977813720703, 'timing/data': 0.0003611489664763212, 'timing/model': 1.0698997231665999, 'learning_rate/qwen_vl_interface': 8.704856342446366e-07, 'learning_rate/action_model': 8.704856342446366e-06, 'epoch': 0.25}) +[2026-06-24 11:16:30,906][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.042550768703222275, 'train/grad_norm_pre_clip': 41.529022216796875, 'timing/data': 0.00032301596365869045, 'timing/model': 1.0690507870167494, 'learning_rate/qwen_vl_interface': 8.684958738638402e-07, 'learning_rate/action_model': 8.684958738638401e-06, 'epoch': 0.25}) +[2026-06-24 11:16:31,978][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.06492803990840912, 'train/grad_norm_pre_clip': 74.16503143310547, 'timing/data': 0.0005290140397846699, 'timing/model': 1.0698475018143654, 'learning_rate/qwen_vl_interface': 8.665112554221844e-07, 'learning_rate/action_model': 8.665112554221844e-06, 'epoch': 0.25}) +[2026-06-24 11:16:33,047][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.1455102562904358, 'train/grad_norm_pre_clip': 60.07834243774414, 'timing/data': 0.00032756500877439976, 'timing/model': 1.0670009541790932, 'learning_rate/qwen_vl_interface': 8.645317812487299e-07, 'learning_rate/action_model': 8.645317812487298e-06, 'epoch': 0.26}) +[2026-06-24 11:16:34,379][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.054606299847364426, 'train/grad_norm_pre_clip': 28.51706314086914, 'timing/data': 0.00031064311042428017, 'timing/model': 1.3309781239368021, 'learning_rate/qwen_vl_interface': 8.625574536664995e-07, 'learning_rate/action_model': 8.625574536664995e-06, 'epoch': 0.26}) +[2026-06-24 11:16:35,449][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.02383667603135109, 'train/grad_norm_pre_clip': 15.97605037689209, 'timing/data': 0.0003268439322710037, 'timing/model': 1.0681559089571238, 'learning_rate/qwen_vl_interface': 8.605882749924795e-07, 'learning_rate/action_model': 8.605882749924795e-06, 'epoch': 0.26}) +[2026-06-24 11:16:36,522][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.08404400199651718, 'train/grad_norm_pre_clip': 64.13282775878906, 'timing/data': 0.00032916804775595665, 'timing/model': 1.071580610005185, 'learning_rate/qwen_vl_interface': 8.586242475376064e-07, 'learning_rate/action_model': 8.586242475376063e-06, 'epoch': 0.26}) +[2026-06-24 11:16:37,593][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.12383165955543518, 'train/grad_norm_pre_clip': 67.85635375976562, 'timing/data': 0.0003035000991076231, 'timing/model': 1.0694541539996862, 'learning_rate/qwen_vl_interface': 8.566653736067781e-07, 'learning_rate/action_model': 8.56665373606778e-06, 'epoch': 0.26}) +[2026-06-24 11:16:38,664][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.06942939013242722, 'train/grad_norm_pre_clip': 40.2326545715332, 'timing/data': 0.000300354091450572, 'timing/model': 1.068913107039407, 'learning_rate/qwen_vl_interface': 8.547116554988421e-07, 'learning_rate/action_model': 8.547116554988422e-06, 'epoch': 0.26}) +[2026-06-24 11:16:39,735][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.02462751790881157, 'train/grad_norm_pre_clip': 21.037200927734375, 'timing/data': 0.0003368628676980734, 'timing/model': 1.0696719700936228, 'learning_rate/qwen_vl_interface': 8.527630955065959e-07, 'learning_rate/action_model': 8.527630955065959e-06, 'epoch': 0.26}) +[2026-06-24 11:16:40,807][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.3235372304916382, 'train/grad_norm_pre_clip': 243.10800170898438, 'timing/data': 0.00033759395591914654, 'timing/model': 1.0708835329860449, 'learning_rate/qwen_vl_interface': 8.508196959167822e-07, 'learning_rate/action_model': 8.508196959167822e-06, 'epoch': 0.26}) +[2026-06-24 11:16:41,878][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.15305863320827484, 'train/grad_norm_pre_clip': 33.135990142822266, 'timing/data': 0.0003117751330137253, 'timing/model': 1.0686720579396933, 'learning_rate/qwen_vl_interface': 8.488814590100887e-07, 'learning_rate/action_model': 8.488814590100887e-06, 'epoch': 0.26}) +[2026-06-24 11:16:42,949][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.006987982429563999, 'train/grad_norm_pre_clip': 7.9557695388793945, 'timing/data': 0.00031387992203235626, 'timing/model': 1.0699236681684852, 'learning_rate/qwen_vl_interface': 8.469483870611453e-07, 'learning_rate/action_model': 8.469483870611452e-06, 'epoch': 0.26}) +[2026-06-24 11:16:44,020][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.04187002405524254, 'train/grad_norm_pre_clip': 35.75083541870117, 'timing/data': 0.00037654698826372623, 'timing/model': 1.0689008079934865, 'learning_rate/qwen_vl_interface': 8.450204823385183e-07, 'learning_rate/action_model': 8.450204823385183e-06, 'epoch': 0.26}) +[2026-06-24 11:16:45,091][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.015233628451824188, 'train/grad_norm_pre_clip': 15.375152587890625, 'timing/data': 0.0003038509748876095, 'timing/model': 1.0698590849060565, 'learning_rate/qwen_vl_interface': 8.430977471047141e-07, 'learning_rate/action_model': 8.430977471047141e-06, 'epoch': 0.26}) +[2026-06-24 11:16:46,161][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.1205088198184967, 'train/grad_norm_pre_clip': 71.73934936523438, 'timing/data': 0.0005349649582058191, 'timing/model': 1.0674502479378134, 'learning_rate/qwen_vl_interface': 8.41180183616166e-07, 'learning_rate/action_model': 8.41180183616166e-06, 'epoch': 0.26}) +[2026-06-24 11:16:47,233][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.05876807123422623, 'train/grad_norm_pre_clip': 31.85441017150879, 'timing/data': 0.00035128905437886715, 'timing/model': 1.0706149500329047, 'learning_rate/qwen_vl_interface': 8.392677941232427e-07, 'learning_rate/action_model': 8.392677941232427e-06, 'epoch': 0.26}) +[2026-06-24 11:16:48,303][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.015485969372093678, 'train/grad_norm_pre_clip': 23.239896774291992, 'timing/data': 0.0002965969033539295, 'timing/model': 1.0686993200797588, 'learning_rate/qwen_vl_interface': 8.373605808702408e-07, 'learning_rate/action_model': 8.373605808702407e-06, 'epoch': 0.26}) +[2026-06-24 11:16:49,376][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.057783275842666626, 'train/grad_norm_pre_clip': 39.60029602050781, 'timing/data': 0.000489559955894947, 'timing/model': 1.0708868869114667, 'learning_rate/qwen_vl_interface': 8.354585460953814e-07, 'learning_rate/action_model': 8.354585460953814e-06, 'epoch': 0.26}) +[2026-06-24 11:16:50,652][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.19935302436351776, 'train/grad_norm_pre_clip': 91.59625244140625, 'timing/data': 0.0003260830417275429, 'timing/model': 1.2737304270267487, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 0.26}) +[2026-06-24 11:16:51,720][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.07831542938947678, 'train/grad_norm_pre_clip': 64.00198364257812, 'timing/data': 0.0003341580741107464, 'timing/model': 1.0669826599769294, 'learning_rate/qwen_vl_interface': 8.316700209025805e-07, 'learning_rate/action_model': 8.316700209025805e-06, 'epoch': 0.26}) +[2026-06-24 11:16:52,792][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.025743525475263596, 'train/grad_norm_pre_clip': 21.85960578918457, 'timing/data': 0.0003125069197267294, 'timing/model': 1.0703125428408384, 'learning_rate/qwen_vl_interface': 8.297835349306862e-07, 'learning_rate/action_model': 8.297835349306863e-06, 'epoch': 0.26}) +[2026-06-24 11:16:53,863][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.0629725307226181, 'train/grad_norm_pre_clip': 20.60505485534668, 'timing/data': 0.0003371338825672865, 'timing/model': 1.0693459461908787, 'learning_rate/qwen_vl_interface': 8.279022363290195e-07, 'learning_rate/action_model': 8.279022363290193e-06, 'epoch': 0.26}) +[2026-06-24 11:16:54,936][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.03379069268703461, 'train/grad_norm_pre_clip': 37.08432388305664, 'timing/data': 0.0005551830399781466, 'timing/model': 1.0712194629013538, 'learning_rate/qwen_vl_interface': 8.260261273053879e-07, 'learning_rate/action_model': 8.260261273053879e-06, 'epoch': 0.26}) +[2026-06-24 11:16:56,008][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.15487848222255707, 'train/grad_norm_pre_clip': 73.10900115966797, 'timing/data': 0.00047689606435596943, 'timing/model': 1.0697911391034722, 'learning_rate/qwen_vl_interface': 8.241552100615118e-07, 'learning_rate/action_model': 8.241552100615116e-06, 'epoch': 0.26}) +[2026-06-24 11:16:57,079][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.09658804535865784, 'train/grad_norm_pre_clip': 114.83223724365234, 'timing/data': 0.000351679977029562, 'timing/model': 1.0693875288125128, 'learning_rate/qwen_vl_interface': 8.222894867930146e-07, 'learning_rate/action_model': 8.222894867930147e-06, 'epoch': 0.26}) +[2026-06-24 11:16:58,150][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.0219236109405756, 'train/grad_norm_pre_clip': 35.23784637451172, 'timing/data': 0.000366969034075737, 'timing/model': 1.0694601740688086, 'learning_rate/qwen_vl_interface': 8.204289596894273e-07, 'learning_rate/action_model': 8.204289596894272e-06, 'epoch': 0.26}) +[2026-06-24 11:16:59,219][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.10559805482625961, 'train/grad_norm_pre_clip': 59.803863525390625, 'timing/data': 0.0002994420938193798, 'timing/model': 1.0677086431533098, 'learning_rate/qwen_vl_interface': 8.185736309341837e-07, 'learning_rate/action_model': 8.185736309341837e-06, 'epoch': 0.26}) +[2026-06-24 11:17:00,289][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.03909049555659294, 'train/grad_norm_pre_clip': 59.628963470458984, 'timing/data': 0.0003120768815279007, 'timing/model': 1.0685635157860816, 'learning_rate/qwen_vl_interface': 8.16723502704613e-07, 'learning_rate/action_model': 8.16723502704613e-06, 'epoch': 0.26}) +[2026-06-24 11:17:01,361][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.007236706558614969, 'train/grad_norm_pre_clip': 8.156266212463379, 'timing/data': 0.00033717299811542034, 'timing/model': 1.0700281348545104, 'learning_rate/qwen_vl_interface': 8.148785771719446e-07, 'learning_rate/action_model': 8.148785771719446e-06, 'epoch': 0.26}) +[2026-06-24 11:17:02,432][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.08600278943777084, 'train/grad_norm_pre_clip': 48.21144485473633, 'timing/data': 0.00033735414035618305, 'timing/model': 1.0696024948265404, 'learning_rate/qwen_vl_interface': 8.130388565013017e-07, 'learning_rate/action_model': 8.130388565013017e-06, 'epoch': 0.26}) +[2026-06-24 11:17:03,502][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.10597924143075943, 'train/grad_norm_pre_clip': 37.83147048950195, 'timing/data': 0.00029640598222613335, 'timing/model': 1.0686353058554232, 'learning_rate/qwen_vl_interface': 8.112043428516997e-07, 'learning_rate/action_model': 8.112043428516996e-06, 'epoch': 0.26}) +[2026-06-24 11:17:04,573][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.0511045828461647, 'train/grad_norm_pre_clip': 64.23030853271484, 'timing/data': 0.0003377748653292656, 'timing/model': 1.0686909588985145, 'learning_rate/qwen_vl_interface': 8.093750383760407e-07, 'learning_rate/action_model': 8.093750383760406e-06, 'epoch': 0.26}) +[2026-06-24 11:17:05,643][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.18443238735198975, 'train/grad_norm_pre_clip': 88.31835174560547, 'timing/data': 0.00030234805308282375, 'timing/model': 1.0687848739326, 'learning_rate/qwen_vl_interface': 8.075509452211175e-07, 'learning_rate/action_model': 8.075509452211174e-06, 'epoch': 0.26}) +[2026-06-24 11:17:06,714][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.18341264128684998, 'train/grad_norm_pre_clip': 48.708683013916016, 'timing/data': 0.0002989519853144884, 'timing/model': 1.0694924420677125, 'learning_rate/qwen_vl_interface': 8.057320655276027e-07, 'learning_rate/action_model': 8.057320655276026e-06, 'epoch': 0.26}) +[2026-06-24 11:17:08,074][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.03911096230149269, 'train/grad_norm_pre_clip': 36.23041534423828, 'timing/data': 0.0003133178688585758, 'timing/model': 1.3583042239770293, 'learning_rate/qwen_vl_interface': 8.039184014300547e-07, 'learning_rate/action_model': 8.039184014300546e-06, 'epoch': 0.26}) +[2026-06-24 11:17:09,146][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.008081425912678242, 'train/grad_norm_pre_clip': 11.818710327148438, 'timing/data': 0.00038074515759944916, 'timing/model': 1.0707430969923735, 'learning_rate/qwen_vl_interface': 8.021099550569088e-07, 'learning_rate/action_model': 8.021099550569087e-06, 'epoch': 0.26}) +[2026-06-24 11:17:10,217][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.05012241750955582, 'train/grad_norm_pre_clip': 64.0118408203125, 'timing/data': 0.000367951113730669, 'timing/model': 1.0688817298505455, 'learning_rate/qwen_vl_interface': 8.003067285304776e-07, 'learning_rate/action_model': 8.003067285304775e-06, 'epoch': 0.26}) +[2026-06-24 11:17:11,288][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.0996127501130104, 'train/grad_norm_pre_clip': 33.08186721801758, 'timing/data': 0.0003082188777625561, 'timing/model': 1.0697678120341152, 'learning_rate/qwen_vl_interface': 7.98508723966947e-07, 'learning_rate/action_model': 7.98508723966947e-06, 'epoch': 0.26}) +[2026-06-24 11:17:12,358][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.011599580757319927, 'train/grad_norm_pre_clip': 12.953349113464355, 'timing/data': 0.00036936416290700436, 'timing/model': 1.0681317280977964, 'learning_rate/qwen_vl_interface': 7.967159434763768e-07, 'learning_rate/action_model': 7.967159434763767e-06, 'epoch': 0.26}) +[2026-06-24 11:17:13,428][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.04459144547581673, 'train/grad_norm_pre_clip': 39.868526458740234, 'timing/data': 0.00038178684189915657, 'timing/model': 1.068371697096154, 'learning_rate/qwen_vl_interface': 7.94928389162695e-07, 'learning_rate/action_model': 7.949283891626949e-06, 'epoch': 0.26}) +[2026-06-24 11:17:14,500][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.11222422122955322, 'train/grad_norm_pre_clip': 142.8208770751953, 'timing/data': 0.0003325049765408039, 'timing/model': 1.0693811408709735, 'learning_rate/qwen_vl_interface': 7.931460631236942e-07, 'learning_rate/action_model': 7.931460631236942e-06, 'epoch': 0.26}) +[2026-06-24 11:17:15,571][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.053120359778404236, 'train/grad_norm_pre_clip': 57.11148452758789, 'timing/data': 0.0005236640572547913, 'timing/model': 1.06915064714849, 'learning_rate/qwen_vl_interface': 7.913689674510351e-07, 'learning_rate/action_model': 7.91368967451035e-06, 'epoch': 0.26}) +[2026-06-24 11:17:16,641][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.027045730501413345, 'train/grad_norm_pre_clip': 31.615951538085938, 'timing/data': 0.0003377748653292656, 'timing/model': 1.068635426927358, 'learning_rate/qwen_vl_interface': 7.895971042302381e-07, 'learning_rate/action_model': 7.89597104230238e-06, 'epoch': 0.26}) +[2026-06-24 11:17:17,711][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.037360984832048416, 'train/grad_norm_pre_clip': 33.68589401245117, 'timing/data': 0.00033945683389902115, 'timing/model': 1.0685992469079792, 'learning_rate/qwen_vl_interface': 7.878304755406827e-07, 'learning_rate/action_model': 7.878304755406828e-06, 'epoch': 0.26}) +[2026-06-24 11:17:18,780][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.011970406398177147, 'train/grad_norm_pre_clip': 12.497968673706055, 'timing/data': 0.0003275950439274311, 'timing/model': 1.067924730014056, 'learning_rate/qwen_vl_interface': 7.860690834556069e-07, 'learning_rate/action_model': 7.860690834556069e-06, 'epoch': 0.26}) +[2026-06-24 11:17:19,851][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.04667995497584343, 'train/grad_norm_pre_clip': 42.0103874206543, 'timing/data': 0.00047293794341385365, 'timing/model': 1.0690265779849142, 'learning_rate/qwen_vl_interface': 7.843129300421018e-07, 'learning_rate/action_model': 7.843129300421018e-06, 'epoch': 0.26}) +[2026-06-24 11:17:20,919][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.09314502030611038, 'train/grad_norm_pre_clip': 24.821853637695312, 'timing/data': 0.0003193190786987543, 'timing/model': 1.0660017339978367, 'learning_rate/qwen_vl_interface': 7.825620173611112e-07, 'learning_rate/action_model': 7.825620173611112e-06, 'epoch': 0.26}) +[2026-06-24 11:17:21,991][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.09490961581468582, 'train/grad_norm_pre_clip': 43.756221771240234, 'timing/data': 0.0003120959736406803, 'timing/model': 1.070273115998134, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 0.26}) +[2026-06-24 11:17:23,061][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.05613909661769867, 'train/grad_norm_pre_clip': 35.79418182373047, 'timing/data': 0.00031056394800543785, 'timing/model': 1.068363216938451, 'learning_rate/qwen_vl_interface': 7.790759224096968e-07, 'learning_rate/action_model': 7.790759224096967e-06, 'epoch': 0.26}) +[2026-06-24 11:17:24,341][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.4873047471046448, 'train/grad_norm_pre_clip': 122.73388671875, 'timing/data': 0.00031155580654740334, 'timing/model': 1.2782080757897347, 'learning_rate/qwen_vl_interface': 7.773407442303986e-07, 'learning_rate/action_model': 7.773407442303986e-06, 'epoch': 0.26}) +[2026-06-24 11:17:25,410][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.16806407272815704, 'train/grad_norm_pre_clip': 91.52648162841797, 'timing/data': 0.00035849306732416153, 'timing/model': 1.067367373034358, 'learning_rate/qwen_vl_interface': 7.756108149658635e-07, 'learning_rate/action_model': 7.756108149658635e-06, 'epoch': 0.26}) +[2026-06-24 11:17:26,480][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.13237912952899933, 'train/grad_norm_pre_clip': 70.73092651367188, 'timing/data': 0.000399499898776412, 'timing/model': 1.0681725637987256, 'learning_rate/qwen_vl_interface': 7.738861366462605e-07, 'learning_rate/action_model': 7.738861366462604e-06, 'epoch': 0.26}) +[2026-06-24 11:17:27,549][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.061252158135175705, 'train/grad_norm_pre_clip': 35.857879638671875, 'timing/data': 0.0003294989001005888, 'timing/model': 1.0681733540259302, 'learning_rate/qwen_vl_interface': 7.721667112955945e-07, 'learning_rate/action_model': 7.721667112955946e-06, 'epoch': 0.26}) +[2026-06-24 11:17:28,622][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.1006673276424408, 'train/grad_norm_pre_clip': 90.99158477783203, 'timing/data': 0.00031290692277252674, 'timing/model': 1.0714984070509672, 'learning_rate/qwen_vl_interface': 7.704525409317085e-07, 'learning_rate/action_model': 7.704525409317085e-06, 'epoch': 0.26}) +[2026-06-24 11:17:29,694][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.035205453634262085, 'train/grad_norm_pre_clip': 30.32850456237793, 'timing/data': 0.00037056603468954563, 'timing/model': 1.0698631438426673, 'learning_rate/qwen_vl_interface': 7.687436275662749e-07, 'learning_rate/action_model': 7.687436275662748e-06, 'epoch': 0.26}) +[2026-06-24 11:17:30,765][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.08732631802558899, 'train/grad_norm_pre_clip': 99.5692367553711, 'timing/data': 0.00033642188645899296, 'timing/model': 1.0690563879907131, 'learning_rate/qwen_vl_interface': 7.670399732048e-07, 'learning_rate/action_model': 7.670399732048e-06, 'epoch': 0.26}) +[2026-06-24 11:17:31,836][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.029045891016721725, 'train/grad_norm_pre_clip': 37.29978942871094, 'timing/data': 0.0002986609470099211, 'timing/model': 1.0696405940689147, 'learning_rate/qwen_vl_interface': 7.653415798466168e-07, 'learning_rate/action_model': 7.653415798466168e-06, 'epoch': 0.26}) +[2026-06-24 11:17:32,906][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.026013148948550224, 'train/grad_norm_pre_clip': 22.875133514404297, 'timing/data': 0.0003476629499346018, 'timing/model': 1.0688461791723967, 'learning_rate/qwen_vl_interface': 7.636484494848849e-07, 'learning_rate/action_model': 7.636484494848848e-06, 'epoch': 0.26}) +[2026-06-24 11:17:33,977][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.1779065728187561, 'train/grad_norm_pre_clip': 85.14128875732422, 'timing/data': 0.00032645417377352715, 'timing/model': 1.0688232448883355, 'learning_rate/qwen_vl_interface': 7.619605841065863e-07, 'learning_rate/action_model': 7.619605841065863e-06, 'epoch': 0.26}) +[2026-06-24 11:17:35,047][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.2352498173713684, 'train/grad_norm_pre_clip': 70.91914367675781, 'timing/data': 0.00034852512180805206, 'timing/model': 1.068432389991358, 'learning_rate/qwen_vl_interface': 7.602779856925258e-07, 'learning_rate/action_model': 7.602779856925257e-06, 'epoch': 0.26}) +[2026-06-24 11:17:36,118][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.15782764554023743, 'train/grad_norm_pre_clip': 81.48177337646484, 'timing/data': 0.0003151919227093458, 'timing/model': 1.0701021139975637, 'learning_rate/qwen_vl_interface': 7.586006562173262e-07, 'learning_rate/action_model': 7.586006562173262e-06, 'epoch': 0.26}) +[2026-06-24 11:17:37,190][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.09037448465824127, 'train/grad_norm_pre_clip': 52.30064010620117, 'timing/data': 0.0003179269842803478, 'timing/model': 1.0698254231829196, 'learning_rate/qwen_vl_interface': 7.569285976494262e-07, 'learning_rate/action_model': 7.569285976494261e-06, 'epoch': 0.26}) +[2026-06-24 11:17:38,260][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.026523981243371964, 'train/grad_norm_pre_clip': 46.85468673706055, 'timing/data': 0.0003062149044126272, 'timing/model': 1.0692229189444333, 'learning_rate/qwen_vl_interface': 7.55261811951081e-07, 'learning_rate/action_model': 7.55261811951081e-06, 'epoch': 0.26}) +[2026-06-24 11:17:39,332][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.005233106669038534, 'train/grad_norm_pre_clip': 10.569072723388672, 'timing/data': 0.0004215720109641552, 'timing/model': 1.0695009198971093, 'learning_rate/qwen_vl_interface': 7.536003010783552e-07, 'learning_rate/action_model': 7.536003010783551e-06, 'epoch': 0.26}) +[2026-06-24 11:17:40,615][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.06536070257425308, 'train/grad_norm_pre_clip': 33.7372932434082, 'timing/data': 0.00030678603798151016, 'timing/model': 1.2822281401604414, 'learning_rate/qwen_vl_interface': 7.519440669811251e-07, 'learning_rate/action_model': 7.5194406698112505e-06, 'epoch': 0.26}) +[2026-06-24 11:17:41,686][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.04931681603193283, 'train/grad_norm_pre_clip': 57.49409866333008, 'timing/data': 0.0003000840079039335, 'timing/model': 1.0692841298878193, 'learning_rate/qwen_vl_interface': 7.502931116030732e-07, 'learning_rate/action_model': 7.502931116030732e-06, 'epoch': 0.26}) +[2026-06-24 11:17:42,754][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.26733577251434326, 'train/grad_norm_pre_clip': 63.649757385253906, 'timing/data': 0.00032885815016925335, 'timing/model': 1.0657445709221065, 'learning_rate/qwen_vl_interface': 7.486474368816889e-07, 'learning_rate/action_model': 7.486474368816888e-06, 'epoch': 0.26}) +[2026-06-24 11:17:43,823][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.0036326898261904716, 'train/grad_norm_pre_clip': 3.2883026599884033, 'timing/data': 0.00038850982673466206, 'timing/model': 1.0678392329718918, 'learning_rate/qwen_vl_interface': 7.470070447482611e-07, 'learning_rate/action_model': 7.4700704474826105e-06, 'epoch': 0.26}) +[2026-06-24 11:17:44,895][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.10592883080244064, 'train/grad_norm_pre_clip': 128.17391967773438, 'timing/data': 0.0002901148982346058, 'timing/model': 1.070768543984741, 'learning_rate/qwen_vl_interface': 7.453719371278828e-07, 'learning_rate/action_model': 7.4537193712788275e-06, 'epoch': 0.26}) +[2026-06-24 11:17:45,967][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.014823266305029392, 'train/grad_norm_pre_clip': 15.313100814819336, 'timing/data': 0.00032592308707535267, 'timing/model': 1.069753214949742, 'learning_rate/qwen_vl_interface': 7.437421159394434e-07, 'learning_rate/action_model': 7.437421159394434e-06, 'epoch': 0.26}) +[2026-06-24 11:17:47,039][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.08855293691158295, 'train/grad_norm_pre_clip': 41.85829162597656, 'timing/data': 0.00034628016874194145, 'timing/model': 1.0710922179277986, 'learning_rate/qwen_vl_interface': 7.421175830956294e-07, 'learning_rate/action_model': 7.421175830956293e-06, 'epoch': 0.26}) +[2026-06-24 11:17:48,112][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.07680786401033401, 'train/grad_norm_pre_clip': 63.7248649597168, 'timing/data': 0.00031014299020171165, 'timing/model': 1.0707738688215613, 'learning_rate/qwen_vl_interface': 7.4049834050292e-07, 'learning_rate/action_model': 7.404983405029201e-06, 'epoch': 0.26}) +[2026-06-24 11:17:49,183][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.06387875229120255, 'train/grad_norm_pre_clip': 39.961463928222656, 'timing/data': 0.00043585896492004395, 'timing/model': 1.0693730970378965, 'learning_rate/qwen_vl_interface': 7.388843900615862e-07, 'learning_rate/action_model': 7.388843900615862e-06, 'epoch': 0.26}) +[2026-06-24 11:17:50,254][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.153046116232872, 'train/grad_norm_pre_clip': 53.067298889160156, 'timing/data': 0.00035114004276692867, 'timing/model': 1.069029298843816, 'learning_rate/qwen_vl_interface': 7.372757336656894e-07, 'learning_rate/action_model': 7.372757336656894e-06, 'epoch': 0.26}) +[2026-06-24 11:17:51,322][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.07594913244247437, 'train/grad_norm_pre_clip': 28.2159481048584, 'timing/data': 0.0003282970283180475, 'timing/model': 1.0665899370796978, 'learning_rate/qwen_vl_interface': 7.356723732030767e-07, 'learning_rate/action_model': 7.356723732030766e-06, 'epoch': 0.26}) +[2026-06-24 11:17:52,393][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.005106206052005291, 'train/grad_norm_pre_clip': 3.374882936477661, 'timing/data': 0.00029631704092025757, 'timing/model': 1.0695242460351437, 'learning_rate/qwen_vl_interface': 7.340743105553827e-07, 'learning_rate/action_model': 7.340743105553826e-06, 'epoch': 0.26}) +[2026-06-24 11:17:53,463][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.07407452911138535, 'train/grad_norm_pre_clip': 42.15293502807617, 'timing/data': 0.0003293489571660757, 'timing/model': 1.0689310810994357, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 0.26}) +[2026-06-24 11:17:54,535][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.02374400943517685, 'train/grad_norm_pre_clip': 16.315181732177734, 'timing/data': 0.0005302960053086281, 'timing/model': 1.0698085478506982, 'learning_rate/qwen_vl_interface': 7.30894086200188e-07, 'learning_rate/action_model': 7.30894086200188e-06, 'epoch': 0.26}) +[2026-06-24 11:17:55,606][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.26876384019851685, 'train/grad_norm_pre_clip': 87.75432586669922, 'timing/data': 0.0003278958611190319, 'timing/model': 1.0696324051823467, 'learning_rate/qwen_vl_interface': 7.293119282248585e-07, 'learning_rate/action_model': 7.293119282248584e-06, 'epoch': 0.26}) +[2026-06-24 11:17:56,864][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.23903727531433105, 'train/grad_norm_pre_clip': 105.29495239257812, 'timing/data': 0.00030781817622482777, 'timing/model': 1.2563196329865605, 'learning_rate/qwen_vl_interface': 7.277350755287837e-07, 'learning_rate/action_model': 7.277350755287837e-06, 'epoch': 0.26}) +[2026-06-24 11:17:57,935][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.010633363388478756, 'train/grad_norm_pre_clip': 21.280858993530273, 'timing/data': 0.0003522820770740509, 'timing/model': 1.0692869529593736, 'learning_rate/qwen_vl_interface': 7.261635299624877e-07, 'learning_rate/action_model': 7.261635299624878e-06, 'epoch': 0.26}) +[2026-06-24 11:17:59,007][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.0017393510788679123, 'train/grad_norm_pre_clip': 1.092875599861145, 'timing/data': 0.000387958949431777, 'timing/model': 1.0702924609649926, 'learning_rate/qwen_vl_interface': 7.245972933702652e-07, 'learning_rate/action_model': 7.2459729337026525e-06, 'epoch': 0.26}) +[2026-06-24 11:18:00,078][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.14829355478286743, 'train/grad_norm_pre_clip': 103.56906127929688, 'timing/data': 0.00029908190481364727, 'timing/model': 1.0689638832118362, 'learning_rate/qwen_vl_interface': 7.230363675901835e-07, 'learning_rate/action_model': 7.230363675901836e-06, 'epoch': 0.26}) +[2026-06-24 11:18:01,150][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.24795062839984894, 'train/grad_norm_pre_clip': 103.7338638305664, 'timing/data': 0.000391555018723011, 'timing/model': 1.0703366310335696, 'learning_rate/qwen_vl_interface': 7.214807544540757e-07, 'learning_rate/action_model': 7.214807544540757e-06, 'epoch': 0.26}) +[2026-06-24 11:18:02,223][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.09554940462112427, 'train/grad_norm_pre_clip': 75.68714904785156, 'timing/data': 0.0002979389391839504, 'timing/model': 1.0711512279231101, 'learning_rate/qwen_vl_interface': 7.199304557875411e-07, 'learning_rate/action_model': 7.199304557875411e-06, 'epoch': 0.26}) +[2026-06-24 11:18:03,296][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.18165940046310425, 'train/grad_norm_pre_clip': 86.10362243652344, 'timing/data': 0.0003707760479301214, 'timing/model': 1.0715540631208569, 'learning_rate/qwen_vl_interface': 7.183854734099403e-07, 'learning_rate/action_model': 7.183854734099403e-06, 'epoch': 0.26}) +[2026-06-24 11:18:04,371][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.0023031863383948803, 'train/grad_norm_pre_clip': 3.866024971008301, 'timing/data': 0.0003277561627328396, 'timing/model': 1.0732230769935995, 'learning_rate/qwen_vl_interface': 7.168458091343965e-07, 'learning_rate/action_model': 7.168458091343965e-06, 'epoch': 0.26}) +[2026-06-24 11:18:05,489][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.02069120481610298, 'train/grad_norm_pre_clip': 35.53728103637695, 'timing/data': 0.0003172049764543772, 'timing/model': 1.1165833119302988, 'learning_rate/qwen_vl_interface': 7.153114647677922e-07, 'learning_rate/action_model': 7.1531146476779224e-06, 'epoch': 0.26}) +[2026-06-24 11:18:06,693][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.06269432604312897, 'train/grad_norm_pre_clip': 65.52826690673828, 'timing/data': 0.00031643384136259556, 'timing/model': 1.2023322209715843, 'learning_rate/qwen_vl_interface': 7.137824421107658e-07, 'learning_rate/action_model': 7.137824421107657e-06, 'epoch': 0.26}) +[2026-06-24 11:18:07,771][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.16407109797000885, 'train/grad_norm_pre_clip': 92.357177734375, 'timing/data': 0.0003450480289757252, 'timing/model': 1.0767561881802976, 'learning_rate/qwen_vl_interface': 7.122587429577106e-07, 'learning_rate/action_model': 7.122587429577106e-06, 'epoch': 0.26}) +[2026-06-24 11:18:08,861][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.028800783678889275, 'train/grad_norm_pre_clip': 32.90122985839844, 'timing/data': 0.0003098919987678528, 'timing/model': 1.0876512399408966, 'learning_rate/qwen_vl_interface': 7.107403690967713e-07, 'learning_rate/action_model': 7.1074036909677135e-06, 'epoch': 0.26}) +[2026-06-24 11:18:09,947][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.024169834330677986, 'train/grad_norm_pre_clip': 14.514641761779785, 'timing/data': 0.0003157430328428745, 'timing/model': 1.0848505180329084, 'learning_rate/qwen_vl_interface': 7.092273223098452e-07, 'learning_rate/action_model': 7.092273223098451e-06, 'epoch': 0.26}) +[2026-06-24 11:18:11,045][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.01857377588748932, 'train/grad_norm_pre_clip': 33.69245910644531, 'timing/data': 0.0003136401064693928, 'timing/model': 1.09634452406317, 'learning_rate/qwen_vl_interface': 7.077196043725763e-07, 'learning_rate/action_model': 7.077196043725763e-06, 'epoch': 0.26}) +[2026-06-24 11:18:12,178][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.0571967214345932, 'train/grad_norm_pre_clip': 69.0341567993164, 'timing/data': 0.0003498971927911043, 'timing/model': 1.1310032960027456, 'learning_rate/qwen_vl_interface': 7.062172170543574e-07, 'learning_rate/action_model': 7.062172170543574e-06, 'epoch': 0.26}) +[2026-06-24 11:18:13,549][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.17036746442317963, 'train/grad_norm_pre_clip': 97.70979309082031, 'timing/data': 0.00035630003549158573, 'timing/model': 1.3691353101748973, 'learning_rate/qwen_vl_interface': 7.047201621183212e-07, 'learning_rate/action_model': 7.047201621183211e-06, 'epoch': 0.26}) +[2026-06-24 11:18:14,657][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.17893168330192566, 'train/grad_norm_pre_clip': 80.2569808959961, 'timing/data': 0.0003300798125565052, 'timing/model': 1.1069214949384332, 'learning_rate/qwen_vl_interface': 7.032284413213467e-07, 'learning_rate/action_model': 7.032284413213466e-06, 'epoch': 0.26}) +[2026-06-24 11:18:15,730][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.039813634008169174, 'train/grad_norm_pre_clip': 43.47024917602539, 'timing/data': 0.000329198082908988, 'timing/model': 1.0708142779767513, 'learning_rate/qwen_vl_interface': 7.017420564140503e-07, 'learning_rate/action_model': 7.017420564140504e-06, 'epoch': 0.26}) +[2026-06-24 11:18:16,864][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.07537665963172913, 'train/grad_norm_pre_clip': 112.7669677734375, 'timing/data': 0.0003098009619861841, 'timing/model': 1.132713021012023, 'learning_rate/qwen_vl_interface': 7.002610091407893e-07, 'learning_rate/action_model': 7.002610091407892e-06, 'epoch': 0.26}) +[2026-06-24 11:18:17,938][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.11830491572618484, 'train/grad_norm_pre_clip': 123.93402099609375, 'timing/data': 0.0003318830858916044, 'timing/model': 1.0724578420631588, 'learning_rate/qwen_vl_interface': 6.987853012396534e-07, 'learning_rate/action_model': 6.987853012396533e-06, 'epoch': 0.26}) +[2026-06-24 11:18:19,011][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.09396111965179443, 'train/grad_norm_pre_clip': 94.60660552978516, 'timing/data': 0.0003254711627960205, 'timing/model': 1.0718222470022738, 'learning_rate/qwen_vl_interface': 6.973149344424684e-07, 'learning_rate/action_model': 6.9731493444246825e-06, 'epoch': 0.26}) +[2026-06-24 11:18:20,095][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.04472348093986511, 'train/grad_norm_pre_clip': 38.06442642211914, 'timing/data': 0.00037410296499729156, 'timing/model': 1.0821657520718873, 'learning_rate/qwen_vl_interface': 6.958499104747915e-07, 'learning_rate/action_model': 6.9584991047479145e-06, 'epoch': 0.26}) +[2026-06-24 11:18:21,225][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.5148664712905884, 'train/grad_norm_pre_clip': 98.35184478759766, 'timing/data': 0.00030992296524345875, 'timing/model': 1.1286691899877042, 'learning_rate/qwen_vl_interface': 6.943902310559104e-07, 'learning_rate/action_model': 6.943902310559104e-06, 'epoch': 0.26}) +[2026-06-24 11:18:22,299][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.28681933879852295, 'train/grad_norm_pre_clip': 73.58716583251953, 'timing/data': 0.00033967779017984867, 'timing/model': 1.072578834136948, 'learning_rate/qwen_vl_interface': 6.929358978988407e-07, 'learning_rate/action_model': 6.929358978988406e-06, 'epoch': 0.26}) +[2026-06-24 11:18:23,373][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.1940479874610901, 'train/grad_norm_pre_clip': 63.42485427856445, 'timing/data': 0.0003234280738979578, 'timing/model': 1.0718889059498906, 'learning_rate/qwen_vl_interface': 6.914869127103219e-07, 'learning_rate/action_model': 6.914869127103219e-06, 'epoch': 0.26}) +[2026-06-24 11:18:24,444][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.020865051075816154, 'train/grad_norm_pre_clip': 15.745205879211426, 'timing/data': 0.0003425630275160074, 'timing/model': 1.0697302198968828, 'learning_rate/qwen_vl_interface': 6.900432771908193e-07, 'learning_rate/action_model': 6.9004327719081925e-06, 'epoch': 0.26}) +[2026-06-24 11:18:25,520][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.0320504754781723, 'train/grad_norm_pre_clip': 22.246007919311523, 'timing/data': 0.0003046328201889992, 'timing/model': 1.073889760999009, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 0.27}) +[2026-06-24 11:18:26,594][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.13319045305252075, 'train/grad_norm_pre_clip': 81.37391662597656, 'timing/data': 0.0003067359793931246, 'timing/model': 1.0724937079939991, 'learning_rate/qwen_vl_interface': 6.871720619293313e-07, 'learning_rate/action_model': 6.871720619293313e-06, 'epoch': 0.27}) +[2026-06-24 11:18:27,666][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.027553247287869453, 'train/grad_norm_pre_clip': 22.989137649536133, 'timing/data': 0.0003086789511144161, 'timing/model': 1.0703979309182614, 'learning_rate/qwen_vl_interface': 6.857444855568763e-07, 'learning_rate/action_model': 6.857444855568762e-06, 'epoch': 0.27}) +[2026-06-24 11:18:28,946][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.23498530685901642, 'train/grad_norm_pre_clip': 125.92484283447266, 'timing/data': 0.0003050931263715029, 'timing/model': 1.2783033410087228, 'learning_rate/qwen_vl_interface': 6.843222655924964e-07, 'learning_rate/action_model': 6.843222655924964e-06, 'epoch': 0.27}) +[2026-06-24 11:18:30,040][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.09604588896036148, 'train/grad_norm_pre_clip': 62.73888397216797, 'timing/data': 0.00032699410803616047, 'timing/model': 1.092604273930192, 'learning_rate/qwen_vl_interface': 6.829054037052448e-07, 'learning_rate/action_model': 6.829054037052447e-06, 'epoch': 0.27}) +[2026-06-24 11:18:31,133][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.1523551046848297, 'train/grad_norm_pre_clip': 57.50028991699219, 'timing/data': 0.0003973159473389387, 'timing/model': 1.0918995789252222, 'learning_rate/qwen_vl_interface': 6.814939015578889e-07, 'learning_rate/action_model': 6.814939015578889e-06, 'epoch': 0.27}) +[2026-06-24 11:18:32,207][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.07592974603176117, 'train/grad_norm_pre_clip': 62.266971588134766, 'timing/data': 0.00030404096469283104, 'timing/model': 1.0713954188395292, 'learning_rate/qwen_vl_interface': 6.800877608069048e-07, 'learning_rate/action_model': 6.800877608069048e-06, 'epoch': 0.27}) +[2026-06-24 11:18:33,292][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.25216200947761536, 'train/grad_norm_pre_clip': 74.14888763427734, 'timing/data': 0.00030298903584480286, 'timing/model': 1.0841072211042047, 'learning_rate/qwen_vl_interface': 6.786869831024765e-07, 'learning_rate/action_model': 6.786869831024765e-06, 'epoch': 0.27}) +[2026-06-24 11:18:34,439][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.03876551613211632, 'train/grad_norm_pre_clip': 23.938467025756836, 'timing/data': 0.00039409007877111435, 'timing/model': 1.1447909211274236, 'learning_rate/qwen_vl_interface': 6.772915700884952e-07, 'learning_rate/action_model': 6.772915700884952e-06, 'epoch': 0.27}) +[2026-06-24 11:18:35,512][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.02195003069937229, 'train/grad_norm_pre_clip': 25.860078811645508, 'timing/data': 0.0003155730664730072, 'timing/model': 1.071367733180523, 'learning_rate/qwen_vl_interface': 6.759015234025551e-07, 'learning_rate/action_model': 6.759015234025551e-06, 'epoch': 0.27}) +[2026-06-24 11:18:36,587][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.3479717969894409, 'train/grad_norm_pre_clip': 172.18942260742188, 'timing/data': 0.00029757898300886154, 'timing/model': 1.0739667320158333, 'learning_rate/qwen_vl_interface': 6.745168446759552e-07, 'learning_rate/action_model': 6.745168446759552e-06, 'epoch': 0.27}) +[2026-06-24 11:18:37,666][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2750 +[2026-06-24 11:19:02,222][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2750 +[2026-06-24 11:19:02,222][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2750 +[2026-06-24 11:19:33,684][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2750 +[2026-06-24 11:22:00,264][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2750 +[2026-06-24 11:22:00,265][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.2956385314464569, 'train/grad_norm_pre_clip': 136.151611328125, 'eval/action_loss/seconds': 24.553599643986672, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.06835377961397171, 'eval/latency_2/loss': 0.06835377961397171, 'eval/flappy/loss': 0.06835377961397171, 'eval/flappy/latency_2/loss': 0.06835377961397171, 'eval/flap_precision': 0.9617486338797814, 'eval/flap_recall': 0.9552238805970149, 'eval/flap_f1': 0.9584751531654188, 'eval/flap_rate_model': 0.20333333333333334, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9584751531654188, 'eval/latency_2/flap_precision': 0.9617486338797814, 'eval/latency_2/flap_recall': 0.9552238805970149, 'eval/latency_2/flap_f1': 0.9584751531654188, 'eval/latency_2/flap_rate_model': 0.20333333333333334, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9584751531654188, 'eval/action_classification/data_wait_seconds': 0.36240801215171814, 'eval/action_classification/predict_seconds': 31.00783920288086, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 112.99999999999913, 'rl_games_eval/mid_train/mean_length': 930.4, 'rl_games_eval/mid_train/std_reward': 113.37080752998062, 'rl_games_eval/mid_train/std_length': 914.9803495157696, 'rl_games_eval/mid_train/macro_mean_reward': 112.99999999999913, 'rl_games_eval/mid_train/macro_mean_length': 930.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 112.99999999999913, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 930.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 113.37080752998062, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 914.9803495157696, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 112.99999999999913, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003153020516037941, 'timing/model': 1.0779882629867643, 'learning_rate/qwen_vl_interface': 6.731375355336913e-07, 'learning_rate/action_model': 6.7313753553369125e-06, 'epoch': 0.27}) +[2026-06-24 11:22:01,319][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.16146551072597504, 'train/grad_norm_pre_clip': 44.66462707519531, 'timing/data': 0.00036370381712913513, 'timing/model': 1.052289682906121, 'learning_rate/qwen_vl_interface': 6.717635975944601e-07, 'learning_rate/action_model': 6.7176359759446e-06, 'epoch': 0.27}) +[2026-06-24 11:22:02,377][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.02627089060842991, 'train/grad_norm_pre_clip': 39.32228469848633, 'timing/data': 0.0003255018964409828, 'timing/model': 1.0562662999145687, 'learning_rate/qwen_vl_interface': 6.70395032470654e-07, 'learning_rate/action_model': 6.70395032470654e-06, 'epoch': 0.27}) +[2026-06-24 11:22:03,435][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.010247238911688328, 'train/grad_norm_pre_clip': 5.982447624206543, 'timing/data': 0.000315152108669281, 'timing/model': 1.0561085231602192, 'learning_rate/qwen_vl_interface': 6.690318417683606e-07, 'learning_rate/action_model': 6.690318417683605e-06, 'epoch': 0.27}) +[2026-06-24 11:22:04,492][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.17234408855438232, 'train/grad_norm_pre_clip': 138.3977813720703, 'timing/data': 0.0003468820359557867, 'timing/model': 1.0562388668768108, 'learning_rate/qwen_vl_interface': 6.676740270873606e-07, 'learning_rate/action_model': 6.676740270873607e-06, 'epoch': 0.27}) +[2026-06-24 11:22:05,553][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.058052413165569305, 'train/grad_norm_pre_clip': 51.6978759765625, 'timing/data': 0.00044548604637384415, 'timing/model': 1.0586416781879961, 'learning_rate/qwen_vl_interface': 6.663215900211247e-07, 'learning_rate/action_model': 6.663215900211248e-06, 'epoch': 0.27}) +[2026-06-24 11:22:06,611][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.007480937987565994, 'train/grad_norm_pre_clip': 9.42094612121582, 'timing/data': 0.00034744199365377426, 'timing/model': 1.0565085008274764, 'learning_rate/qwen_vl_interface': 6.649745321568127e-07, 'learning_rate/action_model': 6.649745321568127e-06, 'epoch': 0.27}) +[2026-06-24 11:22:07,670][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.31592392921447754, 'train/grad_norm_pre_clip': 59.486114501953125, 'timing/data': 0.0003151309210807085, 'timing/model': 1.0575532719958574, 'learning_rate/qwen_vl_interface': 6.636328550752731e-07, 'learning_rate/action_model': 6.636328550752731e-06, 'epoch': 0.27}) +[2026-06-24 11:22:08,731][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.005061783827841282, 'train/grad_norm_pre_clip': 6.414778709411621, 'timing/data': 0.000335930148139596, 'timing/model': 1.0592310421634465, 'learning_rate/qwen_vl_interface': 6.622965603510384e-07, 'learning_rate/action_model': 6.622965603510384e-06, 'epoch': 0.27}) +[2026-06-24 11:22:09,791][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.052413567900657654, 'train/grad_norm_pre_clip': 51.477333068847656, 'timing/data': 0.00032797595486044884, 'timing/model': 1.0586927810218185, 'learning_rate/qwen_vl_interface': 6.609656495523236e-07, 'learning_rate/action_model': 6.609656495523237e-06, 'epoch': 0.27}) +[2026-06-24 11:22:10,854][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.006759990006685257, 'train/grad_norm_pre_clip': 6.374939441680908, 'timing/data': 0.00034289387986063957, 'timing/model': 1.0604445990175009, 'learning_rate/qwen_vl_interface': 6.596401242410288e-07, 'learning_rate/action_model': 6.5964012424102885e-06, 'epoch': 0.27}) +[2026-06-24 11:22:11,918][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.014943815767765045, 'train/grad_norm_pre_clip': 19.80096435546875, 'timing/data': 0.00037378096021711826, 'timing/model': 1.0632985781412572, 'learning_rate/qwen_vl_interface': 6.583199859727298e-07, 'learning_rate/action_model': 6.583199859727298e-06, 'epoch': 0.27}) +[2026-06-24 11:22:12,981][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.07085511088371277, 'train/grad_norm_pre_clip': 28.810134887695312, 'timing/data': 0.00030913110822439194, 'timing/model': 1.0609366120770574, 'learning_rate/qwen_vl_interface': 6.570052362966832e-07, 'learning_rate/action_model': 6.570052362966832e-06, 'epoch': 0.27}) +[2026-06-24 11:22:14,040][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.2656967341899872, 'train/grad_norm_pre_clip': 76.67809295654297, 'timing/data': 0.00033952808007597923, 'timing/model': 1.0575998050626367, 'learning_rate/qwen_vl_interface': 6.556958767558215e-07, 'learning_rate/action_model': 6.556958767558215e-06, 'epoch': 0.27}) +[2026-06-24 11:22:15,386][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.17140817642211914, 'train/grad_norm_pre_clip': 82.8788833618164, 'timing/data': 0.0004025758244097233, 'timing/model': 1.3434038069099188, 'learning_rate/qwen_vl_interface': 6.543919088867499e-07, 'learning_rate/action_model': 6.543919088867499e-06, 'epoch': 0.27}) +[2026-06-24 11:22:16,448][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.017448341473937035, 'train/grad_norm_pre_clip': 30.457935333251953, 'timing/data': 0.0003034300170838833, 'timing/model': 1.0611173980869353, 'learning_rate/qwen_vl_interface': 6.530933342197487e-07, 'learning_rate/action_model': 6.530933342197487e-06, 'epoch': 0.27}) +[2026-06-24 11:22:17,511][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.24408383667469025, 'train/grad_norm_pre_clip': 54.91238021850586, 'timing/data': 0.00032455893233418465, 'timing/model': 1.061069105984643, 'learning_rate/qwen_vl_interface': 6.518001542787659e-07, 'learning_rate/action_model': 6.518001542787659e-06, 'epoch': 0.27}) +[2026-06-24 11:22:18,574][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.2692323625087738, 'train/grad_norm_pre_clip': 103.62664794921875, 'timing/data': 0.0003110240213572979, 'timing/model': 1.061683289008215, 'learning_rate/qwen_vl_interface': 6.505123705814226e-07, 'learning_rate/action_model': 6.505123705814225e-06, 'epoch': 0.27}) +[2026-06-24 11:22:19,638][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.025193486362695694, 'train/grad_norm_pre_clip': 39.17779541015625, 'timing/data': 0.00032310583628714085, 'timing/model': 1.0623856969177723, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 0.27}) +[2026-06-24 11:22:20,702][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.24922384321689606, 'train/grad_norm_pre_clip': 98.57118225097656, 'timing/data': 0.0003037909045815468, 'timing/model': 1.0627993231173605, 'learning_rate/qwen_vl_interface': 6.47952997956457e-07, 'learning_rate/action_model': 6.47952997956457e-06, 'epoch': 0.27}) +[2026-06-24 11:22:21,767][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.0937345027923584, 'train/grad_norm_pre_clip': 41.9774169921875, 'timing/data': 0.0003118349704891443, 'timing/model': 1.0627153851091862, 'learning_rate/qwen_vl_interface': 6.466814120324034e-07, 'learning_rate/action_model': 6.4668141203240325e-06, 'epoch': 0.27}) +[2026-06-24 11:22:22,831][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.17751045525074005, 'train/grad_norm_pre_clip': 80.48712158203125, 'timing/data': 0.00030916091054677963, 'timing/model': 1.063173825154081, 'learning_rate/qwen_vl_interface': 6.454152283591157e-07, 'learning_rate/action_model': 6.454152283591157e-06, 'epoch': 0.27}) +[2026-06-24 11:22:23,896][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.2694365084171295, 'train/grad_norm_pre_clip': 113.8707275390625, 'timing/data': 0.0002925088629126549, 'timing/model': 1.0632089201826602, 'learning_rate/qwen_vl_interface': 6.441544484225329e-07, 'learning_rate/action_model': 6.441544484225329e-06, 'epoch': 0.27}) +[2026-06-24 11:22:24,963][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.06051308661699295, 'train/grad_norm_pre_clip': 43.55254364013672, 'timing/data': 0.0005066220182925463, 'timing/model': 1.0649191308766603, 'learning_rate/qwen_vl_interface': 6.428990737022494e-07, 'learning_rate/action_model': 6.428990737022494e-06, 'epoch': 0.27}) +[2026-06-24 11:22:26,030][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.07716391980648041, 'train/grad_norm_pre_clip': 35.108375549316406, 'timing/data': 0.00030902097932994366, 'timing/model': 1.0654793020803481, 'learning_rate/qwen_vl_interface': 6.416491056715183e-07, 'learning_rate/action_model': 6.416491056715182e-06, 'epoch': 0.27}) +[2026-06-24 11:22:27,096][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.0053812069818377495, 'train/grad_norm_pre_clip': 5.297745227813721, 'timing/data': 0.0004889790434390306, 'timing/model': 1.0642322190105915, 'learning_rate/qwen_vl_interface': 6.404045457972455e-07, 'learning_rate/action_model': 6.404045457972455e-06, 'epoch': 0.27}) +[2026-06-24 11:22:28,168][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.005201819818466902, 'train/grad_norm_pre_clip': 4.508269786834717, 'timing/data': 0.0003456489648669958, 'timing/model': 1.0696401740424335, 'learning_rate/qwen_vl_interface': 6.391653955399931e-07, 'learning_rate/action_model': 6.391653955399932e-06, 'epoch': 0.27}) +[2026-06-24 11:22:29,236][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.13055065274238586, 'train/grad_norm_pre_clip': 54.69874572753906, 'timing/data': 0.00039055407978594303, 'timing/model': 1.0664679259061813, 'learning_rate/qwen_vl_interface': 6.37931656353971e-07, 'learning_rate/action_model': 6.37931656353971e-06, 'epoch': 0.27}) +[2026-06-24 11:22:30,306][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.0407571941614151, 'train/grad_norm_pre_clip': 24.775802612304688, 'timing/data': 0.00031027314253151417, 'timing/model': 1.0691565941087902, 'learning_rate/qwen_vl_interface': 6.367033296870422e-07, 'learning_rate/action_model': 6.367033296870421e-06, 'epoch': 0.27}) +[2026-06-24 11:22:31,374][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.060634031891822815, 'train/grad_norm_pre_clip': 36.78480529785156, 'timing/data': 0.0004066529218107462, 'timing/model': 1.066173099912703, 'learning_rate/qwen_vl_interface': 6.354804169807163e-07, 'learning_rate/action_model': 6.354804169807163e-06, 'epoch': 0.27}) +[2026-06-24 11:22:32,726][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.007108229678124189, 'train/grad_norm_pre_clip': 12.91054630279541, 'timing/data': 0.00035422504879534245, 'timing/model': 1.3503286608029157, 'learning_rate/qwen_vl_interface': 6.342629196701486e-07, 'learning_rate/action_model': 6.342629196701485e-06, 'epoch': 0.27}) +[2026-06-24 11:22:33,796][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.014347070828080177, 'train/grad_norm_pre_clip': 33.75248718261719, 'timing/data': 0.0003224448300898075, 'timing/model': 1.067741414066404, 'learning_rate/qwen_vl_interface': 6.330508391841409e-07, 'learning_rate/action_model': 6.33050839184141e-06, 'epoch': 0.27}) +[2026-06-24 11:22:34,861][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.1256515234708786, 'train/grad_norm_pre_clip': 68.58763122558594, 'timing/data': 0.0003093499690294266, 'timing/model': 1.064197177067399, 'learning_rate/qwen_vl_interface': 6.318441769451371e-07, 'learning_rate/action_model': 6.318441769451371e-06, 'epoch': 0.27}) +[2026-06-24 11:22:35,930][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.046448398381471634, 'train/grad_norm_pre_clip': 45.902530670166016, 'timing/data': 0.000311804935336113, 'timing/model': 1.066502084955573, 'learning_rate/qwen_vl_interface': 6.306429343692218e-07, 'learning_rate/action_model': 6.306429343692218e-06, 'epoch': 0.27}) +[2026-06-24 11:22:36,997][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.022825954481959343, 'train/grad_norm_pre_clip': 27.522186279296875, 'timing/data': 0.0003056940622627735, 'timing/model': 1.0663056049961597, 'learning_rate/qwen_vl_interface': 6.294471128661206e-07, 'learning_rate/action_model': 6.294471128661206e-06, 'epoch': 0.27}) +[2026-06-24 11:22:38,066][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.06962717324495316, 'train/grad_norm_pre_clip': 34.20718002319336, 'timing/data': 0.0002946129534393549, 'timing/model': 1.0669632598292083, 'learning_rate/qwen_vl_interface': 6.282567138391972e-07, 'learning_rate/action_model': 6.282567138391972e-06, 'epoch': 0.27}) +[2026-06-24 11:22:39,135][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.09490592777729034, 'train/grad_norm_pre_clip': 60.857948303222656, 'timing/data': 0.0003105930518358946, 'timing/model': 1.0670072410721332, 'learning_rate/qwen_vl_interface': 6.2707173868545e-07, 'learning_rate/action_model': 6.2707173868545005e-06, 'epoch': 0.27}) +[2026-06-24 11:22:40,204][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.004296126775443554, 'train/grad_norm_pre_clip': 4.27865743637085, 'timing/data': 0.0003295291680842638, 'timing/model': 1.0681019960902631, 'learning_rate/qwen_vl_interface': 6.258921887955142e-07, 'learning_rate/action_model': 6.258921887955142e-06, 'epoch': 0.27}) +[2026-06-24 11:22:41,273][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.08827634900808334, 'train/grad_norm_pre_clip': 33.625240325927734, 'timing/data': 0.0004065141547471285, 'timing/model': 1.0670598279684782, 'learning_rate/qwen_vl_interface': 6.24718065553657e-07, 'learning_rate/action_model': 6.247180655536569e-06, 'epoch': 0.27}) +[2026-06-24 11:22:42,343][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.071400485932827, 'train/grad_norm_pre_clip': 49.21574783325195, 'timing/data': 0.0003802650608122349, 'timing/model': 1.0678349831141531, 'learning_rate/qwen_vl_interface': 6.235493703377773e-07, 'learning_rate/action_model': 6.235493703377773e-06, 'epoch': 0.27}) +[2026-06-24 11:22:43,413][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.05264519900083542, 'train/grad_norm_pre_clip': 24.02800941467285, 'timing/data': 0.0003921361640095711, 'timing/model': 1.0685479608364403, 'learning_rate/qwen_vl_interface': 6.22386104519405e-07, 'learning_rate/action_model': 6.223861045194049e-06, 'epoch': 0.27}) +[2026-06-24 11:22:44,479][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.0708957388997078, 'train/grad_norm_pre_clip': 55.81253433227539, 'timing/data': 0.0003335659857839346, 'timing/model': 1.0647490669507533, 'learning_rate/qwen_vl_interface': 6.212282694636955e-07, 'learning_rate/action_model': 6.212282694636955e-06, 'epoch': 0.27}) +[2026-06-24 11:22:45,548][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.005677711684256792, 'train/grad_norm_pre_clip': 2.668938636779785, 'timing/data': 0.00030084489844739437, 'timing/model': 1.0669607790187001, 'learning_rate/qwen_vl_interface': 6.200758665294343e-07, 'learning_rate/action_model': 6.2007586652943425e-06, 'epoch': 0.27}) +[2026-06-24 11:22:46,619][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.3435726463794708, 'train/grad_norm_pre_clip': 44.52387619018555, 'timing/data': 0.00042246305383741856, 'timing/model': 1.06931127095595, 'learning_rate/qwen_vl_interface': 6.189288970690293e-07, 'learning_rate/action_model': 6.189288970690293e-06, 'epoch': 0.27}) +[2026-06-24 11:22:47,687][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.0028598241042345762, 'train/grad_norm_pre_clip': 2.528719663619995, 'timing/data': 0.0003869060892611742, 'timing/model': 1.066145887132734, 'learning_rate/qwen_vl_interface': 6.177873624285149e-07, 'learning_rate/action_model': 6.177873624285148e-06, 'epoch': 0.27}) +[2026-06-24 11:22:48,754][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.03437384217977524, 'train/grad_norm_pre_clip': 34.086387634277344, 'timing/data': 0.0003185281530022621, 'timing/model': 1.0654090018942952, 'learning_rate/qwen_vl_interface': 6.166512639475429e-07, 'learning_rate/action_model': 6.1665126394754285e-06, 'epoch': 0.27}) +[2026-06-24 11:22:50,098][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.1195632815361023, 'train/grad_norm_pre_clip': 77.25177001953125, 'timing/data': 0.0003685019910335541, 'timing/model': 1.342375841923058, 'learning_rate/qwen_vl_interface': 6.155206029593903e-07, 'learning_rate/action_model': 6.155206029593903e-06, 'epoch': 0.27}) +[2026-06-24 11:22:51,168][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.13507597148418427, 'train/grad_norm_pre_clip': 84.8416519165039, 'timing/data': 0.00031183590181171894, 'timing/model': 1.0683342940174043, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 0.27}) +[2026-06-24 11:22:52,236][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.06831676512956619, 'train/grad_norm_pre_clip': 39.4880485534668, 'timing/data': 0.00035629887133836746, 'timing/model': 1.0666140129324049, 'learning_rate/qwen_vl_interface': 6.132755987627329e-07, 'learning_rate/action_model': 6.132755987627328e-06, 'epoch': 0.27}) +[2026-06-24 11:22:53,306][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.09034112840890884, 'train/grad_norm_pre_clip': 95.32938385009766, 'timing/data': 0.00031340913847088814, 'timing/model': 1.0680922479368746, 'learning_rate/qwen_vl_interface': 6.121612581888654e-07, 'learning_rate/action_model': 6.121612581888654e-06, 'epoch': 0.27}) +[2026-06-24 11:22:54,375][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.015166450291872025, 'train/grad_norm_pre_clip': 18.49565315246582, 'timing/data': 0.00037142797373235226, 'timing/model': 1.0678329800721258, 'learning_rate/qwen_vl_interface': 6.110523603770883e-07, 'learning_rate/action_model': 6.110523603770883e-06, 'epoch': 0.27}) +[2026-06-24 11:22:55,450][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.16868820786476135, 'train/grad_norm_pre_clip': 92.9003677368164, 'timing/data': 0.0002967969048768282, 'timing/model': 1.0732090848032385, 'learning_rate/qwen_vl_interface': 6.099489066287553e-07, 'learning_rate/action_model': 6.099489066287553e-06, 'epoch': 0.27}) +[2026-06-24 11:22:56,519][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.08098133653402328, 'train/grad_norm_pre_clip': 66.25148010253906, 'timing/data': 0.00033167307265102863, 'timing/model': 1.0670897220261395, 'learning_rate/qwen_vl_interface': 6.088508982388304e-07, 'learning_rate/action_model': 6.0885089823883045e-06, 'epoch': 0.27}) +[2026-06-24 11:22:57,588][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.013689275830984116, 'train/grad_norm_pre_clip': 12.710395812988281, 'timing/data': 0.0005438709631562233, 'timing/model': 1.0676216811407357, 'learning_rate/qwen_vl_interface': 6.077583364958879e-07, 'learning_rate/action_model': 6.077583364958879e-06, 'epoch': 0.27}) +[2026-06-24 11:22:58,656][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.04426465928554535, 'train/grad_norm_pre_clip': 40.128761291503906, 'timing/data': 0.00031334790401160717, 'timing/model': 1.0669522429816425, 'learning_rate/qwen_vl_interface': 6.066712226821099e-07, 'learning_rate/action_model': 6.0667122268210985e-06, 'epoch': 0.27}) +[2026-06-24 11:22:59,729][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.02466501295566559, 'train/grad_norm_pre_clip': 29.1505184173584, 'timing/data': 0.000306194880977273, 'timing/model': 1.0707262689247727, 'learning_rate/qwen_vl_interface': 6.055895580732848e-07, 'learning_rate/action_model': 6.0558955807328465e-06, 'epoch': 0.27}) +[2026-06-24 11:23:00,800][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.05355474725365639, 'train/grad_norm_pre_clip': 54.28694534301758, 'timing/data': 0.0005155778490006924, 'timing/model': 1.0699630251619965, 'learning_rate/qwen_vl_interface': 6.045133439388061e-07, 'learning_rate/action_model': 6.045133439388061e-06, 'epoch': 0.27}) +[2026-06-24 11:23:01,871][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.05229853466153145, 'train/grad_norm_pre_clip': 44.0105094909668, 'timing/data': 0.00033019017428159714, 'timing/model': 1.068724799202755, 'learning_rate/qwen_vl_interface': 6.034425815416723e-07, 'learning_rate/action_model': 6.034425815416722e-06, 'epoch': 0.27}) +[2026-06-24 11:23:02,941][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.1184888482093811, 'train/grad_norm_pre_clip': 74.13191986083984, 'timing/data': 0.0003085201606154442, 'timing/model': 1.068524882895872, 'learning_rate/qwen_vl_interface': 6.023772721384808e-07, 'learning_rate/action_model': 6.023772721384808e-06, 'epoch': 0.27}) +[2026-06-24 11:23:04,013][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.04604411870241165, 'train/grad_norm_pre_clip': 36.408931732177734, 'timing/data': 0.0003308418672531843, 'timing/model': 1.0707734948955476, 'learning_rate/qwen_vl_interface': 6.013174169794321e-07, 'learning_rate/action_model': 6.0131741697943205e-06, 'epoch': 0.27}) +[2026-06-24 11:23:05,082][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.2644738256931305, 'train/grad_norm_pre_clip': 81.63927459716797, 'timing/data': 0.0003120158798992634, 'timing/model': 1.0667974271345884, 'learning_rate/qwen_vl_interface': 6.00263017308326e-07, 'learning_rate/action_model': 6.00263017308326e-06, 'epoch': 0.27}) +[2026-06-24 11:23:06,152][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.037546414881944656, 'train/grad_norm_pre_clip': 33.407127380371094, 'timing/data': 0.00029630702920258045, 'timing/model': 1.06888384395279, 'learning_rate/qwen_vl_interface': 5.992140743625586e-07, 'learning_rate/action_model': 5.992140743625586e-06, 'epoch': 0.27}) +[2026-06-24 11:23:07,498][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.09174422174692154, 'train/grad_norm_pre_clip': 94.82797241210938, 'timing/data': 0.00031334906816482544, 'timing/model': 1.3440468739718199, 'learning_rate/qwen_vl_interface': 5.981705893731236e-07, 'learning_rate/action_model': 5.981705893731236e-06, 'epoch': 0.27}) +[2026-06-24 11:23:08,569][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.1508176624774933, 'train/grad_norm_pre_clip': 100.2068099975586, 'timing/data': 0.00043120887130498886, 'timing/model': 1.0698009049519897, 'learning_rate/qwen_vl_interface': 5.971325635646074e-07, 'learning_rate/action_model': 5.971325635646074e-06, 'epoch': 0.27}) +[2026-06-24 11:23:09,638][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.006040866021066904, 'train/grad_norm_pre_clip': 5.940723896026611, 'timing/data': 0.00034642103128135204, 'timing/model': 1.067549095954746, 'learning_rate/qwen_vl_interface': 5.960999981551927e-07, 'learning_rate/action_model': 5.960999981551927e-06, 'epoch': 0.27}) +[2026-06-24 11:23:10,711][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.0167013481259346, 'train/grad_norm_pre_clip': 14.194502830505371, 'timing/data': 0.0003194401506334543, 'timing/model': 1.0710258441977203, 'learning_rate/qwen_vl_interface': 5.950728943566512e-07, 'learning_rate/action_model': 5.950728943566512e-06, 'epoch': 0.27}) +[2026-06-24 11:23:11,782][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.08007247000932693, 'train/grad_norm_pre_clip': 33.99463653564453, 'timing/data': 0.00040932907722890377, 'timing/model': 1.0692035211250186, 'learning_rate/qwen_vl_interface': 5.940512533743476e-07, 'learning_rate/action_model': 5.940512533743476e-06, 'epoch': 0.27}) +[2026-06-24 11:23:12,852][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.0761619433760643, 'train/grad_norm_pre_clip': 40.33319854736328, 'timing/data': 0.0003056339919567108, 'timing/model': 1.068304531974718, 'learning_rate/qwen_vl_interface': 5.930350764072341e-07, 'learning_rate/action_model': 5.930350764072341e-06, 'epoch': 0.27}) +[2026-06-24 11:23:13,923][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.052038855850696564, 'train/grad_norm_pre_clip': 43.2762565612793, 'timing/data': 0.00030078389681875706, 'timing/model': 1.0698713730089366, 'learning_rate/qwen_vl_interface': 5.9202436464785e-07, 'learning_rate/action_model': 5.920243646478499e-06, 'epoch': 0.27}) +[2026-06-24 11:23:14,992][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.20026248693466187, 'train/grad_norm_pre_clip': 75.44914245605469, 'timing/data': 0.0003154419828206301, 'timing/model': 1.067631786921993, 'learning_rate/qwen_vl_interface': 5.910191192823226e-07, 'learning_rate/action_model': 5.910191192823225e-06, 'epoch': 0.27}) +[2026-06-24 11:23:16,065][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.003551210043951869, 'train/grad_norm_pre_clip': 4.160798072814941, 'timing/data': 0.0003066258504986763, 'timing/model': 1.070884183049202, 'learning_rate/qwen_vl_interface': 5.90019341490364e-07, 'learning_rate/action_model': 5.900193414903639e-06, 'epoch': 0.27}) +[2026-06-24 11:23:17,135][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.008537337183952332, 'train/grad_norm_pre_clip': 4.588105201721191, 'timing/data': 0.0003100931644439697, 'timing/model': 1.0683201679494232, 'learning_rate/qwen_vl_interface': 5.890250324452681e-07, 'learning_rate/action_model': 5.8902503244526815e-06, 'epoch': 0.27}) +[2026-06-24 11:23:18,203][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.06195478141307831, 'train/grad_norm_pre_clip': 24.68260383605957, 'timing/data': 0.0004180849064141512, 'timing/model': 1.0659993791487068, 'learning_rate/qwen_vl_interface': 5.880361933139124e-07, 'learning_rate/action_model': 5.880361933139124e-06, 'epoch': 0.27}) +[2026-06-24 11:23:19,276][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.1333186775445938, 'train/grad_norm_pre_clip': 73.9001235961914, 'timing/data': 0.000302368076518178, 'timing/model': 1.0709739399608225, 'learning_rate/qwen_vl_interface': 5.87052825256754e-07, 'learning_rate/action_model': 5.87052825256754e-06, 'epoch': 0.27}) +[2026-06-24 11:23:20,345][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.027476077899336815, 'train/grad_norm_pre_clip': 58.19572448730469, 'timing/data': 0.00034732301719486713, 'timing/model': 1.0680770392064005, 'learning_rate/qwen_vl_interface': 5.860749294278315e-07, 'learning_rate/action_model': 5.860749294278315e-06, 'epoch': 0.27}) +[2026-06-24 11:23:21,415][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.021631505340337753, 'train/grad_norm_pre_clip': 20.6807861328125, 'timing/data': 0.0003309408202767372, 'timing/model': 1.0682686588261276, 'learning_rate/qwen_vl_interface': 5.851025069747592e-07, 'learning_rate/action_model': 5.851025069747591e-06, 'epoch': 0.27}) +[2026-06-24 11:23:22,485][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.035955093801021576, 'train/grad_norm_pre_clip': 70.25653839111328, 'timing/data': 0.0003287170547991991, 'timing/model': 1.068427064921707, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 0.27}) +[2026-06-24 11:23:23,555][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.003898174036294222, 'train/grad_norm_pre_clip': 4.9924516677856445, 'timing/data': 0.00033236504532396793, 'timing/model': 1.06826161313802, 'learning_rate/qwen_vl_interface': 5.831740867545084e-07, 'learning_rate/action_model': 5.831740867545084e-06, 'epoch': 0.27}) +[2026-06-24 11:23:24,912][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.01884138211607933, 'train/grad_norm_pre_clip': 25.178394317626953, 'timing/data': 0.0003092610277235508, 'timing/model': 1.355540961958468, 'learning_rate/qwen_vl_interface': 5.822180912504383e-07, 'learning_rate/action_model': 5.822180912504383e-06, 'epoch': 0.27}) +[2026-06-24 11:23:25,984][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.19782625138759613, 'train/grad_norm_pre_clip': 56.798500061035156, 'timing/data': 0.0003400479909032583, 'timing/model': 1.0700796251185238, 'learning_rate/qwen_vl_interface': 5.812675736484333e-07, 'learning_rate/action_model': 5.812675736484333e-06, 'epoch': 0.27}) +[2026-06-24 11:23:27,054][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.23142263293266296, 'train/grad_norm_pre_clip': 122.82577514648438, 'timing/data': 0.0004879869520664215, 'timing/model': 1.0682612308301032, 'learning_rate/qwen_vl_interface': 5.803225350639779e-07, 'learning_rate/action_model': 5.803225350639779e-06, 'epoch': 0.27}) +[2026-06-24 11:23:28,126][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.02643641084432602, 'train/grad_norm_pre_clip': 13.07126235961914, 'timing/data': 0.000496201915666461, 'timing/model': 1.0696764369495213, 'learning_rate/qwen_vl_interface': 5.793829766061283e-07, 'learning_rate/action_model': 5.793829766061282e-06, 'epoch': 0.27}) +[2026-06-24 11:23:29,197][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.10061562806367874, 'train/grad_norm_pre_clip': 48.02906799316406, 'timing/data': 0.00030816812068223953, 'timing/model': 1.0696768669877201, 'learning_rate/qwen_vl_interface': 5.784488993775083e-07, 'learning_rate/action_model': 5.784488993775083e-06, 'epoch': 0.27}) +[2026-06-24 11:23:30,270][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.003801167942583561, 'train/grad_norm_pre_clip': 3.2110025882720947, 'timing/data': 0.00034874514676630497, 'timing/model': 1.0712990129832178, 'learning_rate/qwen_vl_interface': 5.775203044743101e-07, 'learning_rate/action_model': 5.775203044743101e-06, 'epoch': 0.27}) +[2026-06-24 11:23:31,341][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.04905352368950844, 'train/grad_norm_pre_clip': 71.00807189941406, 'timing/data': 0.00030966196209192276, 'timing/model': 1.069366523064673, 'learning_rate/qwen_vl_interface': 5.765971929862905e-07, 'learning_rate/action_model': 5.765971929862904e-06, 'epoch': 0.27}) +[2026-06-24 11:23:32,412][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.1135479286313057, 'train/grad_norm_pre_clip': 63.98576354980469, 'timing/data': 0.0003190489951521158, 'timing/model': 1.0698164771310985, 'learning_rate/qwen_vl_interface': 5.756795659967728e-07, 'learning_rate/action_model': 5.756795659967728e-06, 'epoch': 0.27}) +[2026-06-24 11:23:33,483][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.10411582887172699, 'train/grad_norm_pre_clip': 85.08380126953125, 'timing/data': 0.00032749492675065994, 'timing/model': 1.0695917340926826, 'learning_rate/qwen_vl_interface': 5.74767424582644e-07, 'learning_rate/action_model': 5.7476742458264396e-06, 'epoch': 0.27}) +[2026-06-24 11:23:34,555][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.44819480180740356, 'train/grad_norm_pre_clip': 106.92741394042969, 'timing/data': 0.00031119491904973984, 'timing/model': 1.0704989470541477, 'learning_rate/qwen_vl_interface': 5.738607698143521e-07, 'learning_rate/action_model': 5.738607698143521e-06, 'epoch': 0.27}) +[2026-06-24 11:23:35,626][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.008351163007318974, 'train/grad_norm_pre_clip': 13.029541969299316, 'timing/data': 0.00035356380976736546, 'timing/model': 1.0694285659119487, 'learning_rate/qwen_vl_interface': 5.729596027559069e-07, 'learning_rate/action_model': 5.729596027559068e-06, 'epoch': 0.27}) +[2026-06-24 11:23:36,699][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.11357086151838303, 'train/grad_norm_pre_clip': 45.59172058105469, 'timing/data': 0.000310633797198534, 'timing/model': 1.0714920710306615, 'learning_rate/qwen_vl_interface': 5.720639244648784e-07, 'learning_rate/action_model': 5.720639244648784e-06, 'epoch': 0.27}) +[2026-06-24 11:23:37,771][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.07432914525270462, 'train/grad_norm_pre_clip': 87.24666595458984, 'timing/data': 0.00030292803421616554, 'timing/model': 1.0703984261490405, 'learning_rate/qwen_vl_interface': 5.711737359923952e-07, 'learning_rate/action_model': 5.7117373599239516e-06, 'epoch': 0.27}) +[2026-06-24 11:23:38,844][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.06335697323083878, 'train/grad_norm_pre_clip': 90.4247055053711, 'timing/data': 0.0003000539727509022, 'timing/model': 1.0714427568018436, 'learning_rate/qwen_vl_interface': 5.702890383831424e-07, 'learning_rate/action_model': 5.702890383831424e-06, 'epoch': 0.27}) +[2026-06-24 11:23:39,918][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.03546059876680374, 'train/grad_norm_pre_clip': 31.098438262939453, 'timing/data': 0.000298250000923872, 'timing/model': 1.0725133861415088, 'learning_rate/qwen_vl_interface': 5.694098326753633e-07, 'learning_rate/action_model': 5.694098326753633e-06, 'epoch': 0.27}) +[2026-06-24 11:23:40,990][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.18705925345420837, 'train/grad_norm_pre_clip': 108.5048599243164, 'timing/data': 0.0002915880177170038, 'timing/model': 1.0700450500007719, 'learning_rate/qwen_vl_interface': 5.685361199008528e-07, 'learning_rate/action_model': 5.685361199008528e-06, 'epoch': 0.28}) +[2026-06-24 11:23:42,346][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.14178965985774994, 'train/grad_norm_pre_clip': 61.87617492675781, 'timing/data': 0.000311665004119277, 'timing/model': 1.354463733965531, 'learning_rate/qwen_vl_interface': 5.676679010849629e-07, 'learning_rate/action_model': 5.676679010849629e-06, 'epoch': 0.28}) +[2026-06-24 11:23:43,417][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.03332768753170967, 'train/grad_norm_pre_clip': 59.79802322387695, 'timing/data': 0.00031662406399846077, 'timing/model': 1.0689468670170754, 'learning_rate/qwen_vl_interface': 5.668051772465963e-07, 'learning_rate/action_model': 5.668051772465963e-06, 'epoch': 0.28}) +[2026-06-24 11:23:44,488][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.07793894410133362, 'train/grad_norm_pre_clip': 74.57109832763672, 'timing/data': 0.0003355101216584444, 'timing/model': 1.0695030901115388, 'learning_rate/qwen_vl_interface': 5.659479493982075e-07, 'learning_rate/action_model': 5.659479493982075e-06, 'epoch': 0.28}) +[2026-06-24 11:23:45,558][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.044776126742362976, 'train/grad_norm_pre_clip': 75.83525085449219, 'timing/data': 0.0003038009162992239, 'timing/model': 1.069215128896758, 'learning_rate/qwen_vl_interface': 5.650962185458016e-07, 'learning_rate/action_model': 5.650962185458015e-06, 'epoch': 0.28}) +[2026-06-24 11:23:46,630][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.06947531551122665, 'train/grad_norm_pre_clip': 57.1906623840332, 'timing/data': 0.000318148173391819, 'timing/model': 1.069540458964184, 'learning_rate/qwen_vl_interface': 5.642499856889312e-07, 'learning_rate/action_model': 5.642499856889311e-06, 'epoch': 0.28}) +[2026-06-24 11:23:47,701][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.027840986847877502, 'train/grad_norm_pre_clip': 19.302167892456055, 'timing/data': 0.00029315007850527763, 'timing/model': 1.0696122830267996, 'learning_rate/qwen_vl_interface': 5.634092518206982e-07, 'learning_rate/action_model': 5.634092518206982e-06, 'epoch': 0.28}) +[2026-06-24 11:23:48,772][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.0036304921377450228, 'train/grad_norm_pre_clip': 4.977677345275879, 'timing/data': 0.0003419918939471245, 'timing/model': 1.070160862058401, 'learning_rate/qwen_vl_interface': 5.625740179277508e-07, 'learning_rate/action_model': 5.625740179277509e-06, 'epoch': 0.28}) +[2026-06-24 11:23:49,844][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.054744940251111984, 'train/grad_norm_pre_clip': 76.63080596923828, 'timing/data': 0.00033445795997977257, 'timing/model': 1.0695304279215634, 'learning_rate/qwen_vl_interface': 5.617442849902824e-07, 'learning_rate/action_model': 5.617442849902823e-06, 'epoch': 0.28}) +[2026-06-24 11:23:50,920][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.09365377575159073, 'train/grad_norm_pre_clip': 85.0560073852539, 'timing/data': 0.000301075866445899, 'timing/model': 1.0744641521014273, 'learning_rate/qwen_vl_interface': 5.609200539820307e-07, 'learning_rate/action_model': 5.609200539820307e-06, 'epoch': 0.28}) +[2026-06-24 11:23:51,992][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.044057175517082214, 'train/grad_norm_pre_clip': 25.861392974853516, 'timing/data': 0.00029525510035455227, 'timing/model': 1.0708582298830152, 'learning_rate/qwen_vl_interface': 5.60101325870277e-07, 'learning_rate/action_model': 5.6010132587027695e-06, 'epoch': 0.28}) +[2026-06-24 11:23:53,066][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.11075205355882645, 'train/grad_norm_pre_clip': 72.36307525634766, 'timing/data': 0.0003350090701133013, 'timing/model': 1.0720475630369037, 'learning_rate/qwen_vl_interface': 5.592881016158437e-07, 'learning_rate/action_model': 5.592881016158436e-06, 'epoch': 0.28}) +[2026-06-24 11:23:54,139][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.11446275562047958, 'train/grad_norm_pre_clip': 44.54712677001953, 'timing/data': 0.0003029590006917715, 'timing/model': 1.0718524849507958, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 0.28}) +[2026-06-24 11:23:55,212][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.02571629174053669, 'train/grad_norm_pre_clip': 30.609895706176758, 'timing/data': 0.00035880296491086483, 'timing/model': 1.0715587730519474, 'learning_rate/qwen_vl_interface': 5.576781684899362e-07, 'learning_rate/action_model': 5.5767816848993615e-06, 'epoch': 0.28}) +[2026-06-24 11:23:56,288][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.11929567903280258, 'train/grad_norm_pre_clip': 74.06378173828125, 'timing/data': 0.00034519797191023827, 'timing/model': 1.0745130700524896, 'learning_rate/qwen_vl_interface': 5.568814615078076e-07, 'learning_rate/action_model': 5.568814615078076e-06, 'epoch': 0.28}) +[2026-06-24 11:23:57,362][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.014320116490125656, 'train/grad_norm_pre_clip': 10.511786460876465, 'timing/data': 0.0005440809763967991, 'timing/model': 1.0710384440608323, 'learning_rate/qwen_vl_interface': 5.560902621616898e-07, 'learning_rate/action_model': 5.560902621616897e-06, 'epoch': 0.28}) +[2026-06-24 11:23:58,431][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.012412553653120995, 'train/grad_norm_pre_clip': 36.93571090698242, 'timing/data': 0.0003166340757161379, 'timing/model': 1.0682162449229509, 'learning_rate/qwen_vl_interface': 5.553045713800993e-07, 'learning_rate/action_model': 5.5530457138009935e-06, 'epoch': 0.28}) +[2026-06-24 11:23:59,790][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.13236889243125916, 'train/grad_norm_pre_clip': 87.54329681396484, 'timing/data': 0.00031297793611884117, 'timing/model': 1.3574949309695512, 'learning_rate/qwen_vl_interface': 5.54524390085089e-07, 'learning_rate/action_model': 5.545243900850891e-06, 'epoch': 0.28}) +[2026-06-24 11:24:00,862][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.005170564632862806, 'train/grad_norm_pre_clip': 6.8941168785095215, 'timing/data': 0.0003296090289950371, 'timing/model': 1.0698919789865613, 'learning_rate/qwen_vl_interface': 5.537497191922434e-07, 'learning_rate/action_model': 5.537497191922434e-06, 'epoch': 0.28}) +[2026-06-24 11:24:01,939][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.01940169185400009, 'train/grad_norm_pre_clip': 31.063688278198242, 'timing/data': 0.00030256807804107666, 'timing/model': 1.0750394449569285, 'learning_rate/qwen_vl_interface': 5.529805596106833e-07, 'learning_rate/action_model': 5.529805596106832e-06, 'epoch': 0.28}) +[2026-06-24 11:24:03,013][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.02944297529757023, 'train/grad_norm_pre_clip': 34.975807189941406, 'timing/data': 0.00029640691354870796, 'timing/model': 1.0724199649412185, 'learning_rate/qwen_vl_interface': 5.522169122430596e-07, 'learning_rate/action_model': 5.522169122430596e-06, 'epoch': 0.28}) +[2026-06-24 11:24:04,085][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.02863396890461445, 'train/grad_norm_pre_clip': 29.457181930541992, 'timing/data': 0.00033649196848273277, 'timing/model': 1.0704566980712116, 'learning_rate/qwen_vl_interface': 5.514587779855559e-07, 'learning_rate/action_model': 5.514587779855558e-06, 'epoch': 0.28}) +[2026-06-24 11:24:05,158][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.1133616641163826, 'train/grad_norm_pre_clip': 50.409019470214844, 'timing/data': 0.0003168049734085798, 'timing/model': 1.0716296599712223, 'learning_rate/qwen_vl_interface': 5.507061577278841e-07, 'learning_rate/action_model': 5.507061577278841e-06, 'epoch': 0.28}) +[2026-06-24 11:24:06,232][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.09010162204504013, 'train/grad_norm_pre_clip': 39.53709411621094, 'timing/data': 0.0003034300170838833, 'timing/model': 1.0724126000422984, 'learning_rate/qwen_vl_interface': 5.49959052353286e-07, 'learning_rate/action_model': 5.49959052353286e-06, 'epoch': 0.28}) +[2026-06-24 11:24:07,307][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.002522227121517062, 'train/grad_norm_pre_clip': 2.062321901321411, 'timing/data': 0.00033359602093696594, 'timing/model': 1.0735138759482652, 'learning_rate/qwen_vl_interface': 5.492174627385315e-07, 'learning_rate/action_model': 5.492174627385315e-06, 'epoch': 0.28}) +[2026-06-24 11:24:08,381][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.20670107007026672, 'train/grad_norm_pre_clip': 74.2441635131836, 'timing/data': 0.0003862949088215828, 'timing/model': 1.0728184098843485, 'learning_rate/qwen_vl_interface': 5.484813897539175e-07, 'learning_rate/action_model': 5.484813897539176e-06, 'epoch': 0.28}) +[2026-06-24 11:24:09,457][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.15769869089126587, 'train/grad_norm_pre_clip': 128.29563903808594, 'timing/data': 0.0003455991391092539, 'timing/model': 1.0741539970040321, 'learning_rate/qwen_vl_interface': 5.47750834263267e-07, 'learning_rate/action_model': 5.47750834263267e-06, 'epoch': 0.28}) +[2026-06-24 11:24:10,532][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.05179961770772934, 'train/grad_norm_pre_clip': 53.781829833984375, 'timing/data': 0.00033170287497341633, 'timing/model': 1.0731632560491562, 'learning_rate/qwen_vl_interface': 5.470257971239264e-07, 'learning_rate/action_model': 5.470257971239264e-06, 'epoch': 0.28}) +[2026-06-24 11:24:11,606][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.1234472393989563, 'train/grad_norm_pre_clip': 85.09434509277344, 'timing/data': 0.00039952993392944336, 'timing/model': 1.0723508291412145, 'learning_rate/qwen_vl_interface': 5.463062791867685e-07, 'learning_rate/action_model': 5.463062791867685e-06, 'epoch': 0.28}) +[2026-06-24 11:24:12,681][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.21805836260318756, 'train/grad_norm_pre_clip': 88.03390502929688, 'timing/data': 0.00030678603798151016, 'timing/model': 1.073305512079969, 'learning_rate/qwen_vl_interface': 5.455922812961871e-07, 'learning_rate/action_model': 5.455922812961871e-06, 'epoch': 0.28}) +[2026-06-24 11:24:13,756][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.007053237408399582, 'train/grad_norm_pre_clip': 10.22743034362793, 'timing/data': 0.0003084999043494463, 'timing/model': 1.073010016232729, 'learning_rate/qwen_vl_interface': 5.448838042900986e-07, 'learning_rate/action_model': 5.4488380429009855e-06, 'epoch': 0.28}) +[2026-06-24 11:24:14,829][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.02732064761221409, 'train/grad_norm_pre_clip': 28.096216201782227, 'timing/data': 0.00030929106287658215, 'timing/model': 1.0718372529372573, 'learning_rate/qwen_vl_interface': 5.441808489999395e-07, 'learning_rate/action_model': 5.4418084899993955e-06, 'epoch': 0.28}) +[2026-06-24 11:24:15,905][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.07940889894962311, 'train/grad_norm_pre_clip': 50.13665771484375, 'timing/data': 0.00031156488694250584, 'timing/model': 1.0742985741235316, 'learning_rate/qwen_vl_interface': 5.434834162506678e-07, 'learning_rate/action_model': 5.434834162506678e-06, 'epoch': 0.28}) +[2026-06-24 11:24:17,264][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.1598142385482788, 'train/grad_norm_pre_clip': 27.252845764160156, 'timing/data': 0.0005120118148624897, 'timing/model': 1.3572619389742613, 'learning_rate/qwen_vl_interface': 5.427915068607599e-07, 'learning_rate/action_model': 5.427915068607598e-06, 'epoch': 0.28}) +[2026-06-24 11:24:18,339][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.05793296545743942, 'train/grad_norm_pre_clip': 40.63514709472656, 'timing/data': 0.00030920119024813175, 'timing/model': 1.0730859658215195, 'learning_rate/qwen_vl_interface': 5.421051216422092e-07, 'learning_rate/action_model': 5.4210512164220915e-06, 'epoch': 0.28}) +[2026-06-24 11:24:19,412][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.15546222031116486, 'train/grad_norm_pre_clip': 102.867919921875, 'timing/data': 0.0003617689944803715, 'timing/model': 1.0718562558759004, 'learning_rate/qwen_vl_interface': 5.414242614005273e-07, 'learning_rate/action_model': 5.414242614005273e-06, 'epoch': 0.28}) +[2026-06-24 11:24:20,486][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.08203835040330887, 'train/grad_norm_pre_clip': 51.766456604003906, 'timing/data': 0.00030311010777950287, 'timing/model': 1.0720596469473094, 'learning_rate/qwen_vl_interface': 5.407489269347418e-07, 'learning_rate/action_model': 5.4074892693474184e-06, 'epoch': 0.28}) +[2026-06-24 11:24:21,558][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.13566875457763672, 'train/grad_norm_pre_clip': 79.83262634277344, 'timing/data': 0.00030460208654403687, 'timing/model': 1.0705064400099218, 'learning_rate/qwen_vl_interface': 5.400791190373954e-07, 'learning_rate/action_model': 5.400791190373954e-06, 'epoch': 0.28}) +[2026-06-24 11:24:22,634][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.08112837374210358, 'train/grad_norm_pre_clip': 95.63331604003906, 'timing/data': 0.00035806302912533283, 'timing/model': 1.0740394128952175, 'learning_rate/qwen_vl_interface': 5.394148384945446e-07, 'learning_rate/action_model': 5.394148384945446e-06, 'epoch': 0.28}) +[2026-06-24 11:24:23,708][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.005292090121656656, 'train/grad_norm_pre_clip': 6.579274654388428, 'timing/data': 0.0003394379746168852, 'timing/model': 1.072482128860429, 'learning_rate/qwen_vl_interface': 5.387560860857599e-07, 'learning_rate/action_model': 5.387560860857599e-06, 'epoch': 0.28}) +[2026-06-24 11:24:24,781][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.3063674569129944, 'train/grad_norm_pre_clip': 99.58788299560547, 'timing/data': 0.0003340367693454027, 'timing/model': 1.07147177006118, 'learning_rate/qwen_vl_interface': 5.381028625841239e-07, 'learning_rate/action_model': 5.381028625841238e-06, 'epoch': 0.28}) +[2026-06-24 11:24:25,855][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.057507384568452835, 'train/grad_norm_pre_clip': 136.6605987548828, 'timing/data': 0.00033942796289920807, 'timing/model': 1.0724915149621665, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 0.28}) +[2026-06-24 11:24:26,931][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.32439157366752625, 'train/grad_norm_pre_clip': 28.222497940063477, 'timing/data': 0.0003414209932088852, 'timing/model': 1.0741974778939039, 'learning_rate/qwen_vl_interface': 5.368130053621847e-07, 'learning_rate/action_model': 5.368130053621847e-06, 'epoch': 0.28}) +[2026-06-24 11:24:28,001][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.004030370619148016, 'train/grad_norm_pre_clip': 5.287684440612793, 'timing/data': 0.0003585228696465492, 'timing/model': 1.0683092400431633, 'learning_rate/qwen_vl_interface': 5.361763731556008e-07, 'learning_rate/action_model': 5.361763731556008e-06, 'epoch': 0.28}) +[2026-06-24 11:24:29,078][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.023096714168787003, 'train/grad_norm_pre_clip': 31.905363082885742, 'timing/data': 0.0003179069608449936, 'timing/model': 1.0753475460223854, 'learning_rate/qwen_vl_interface': 5.355452728836029e-07, 'learning_rate/action_model': 5.355452728836028e-06, 'epoch': 0.28}) +[2026-06-24 11:24:30,153][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.06908056139945984, 'train/grad_norm_pre_clip': 60.235843658447266, 'timing/data': 0.0003149008844047785, 'timing/model': 1.0738111911341548, 'learning_rate/qwen_vl_interface': 5.349197052868213e-07, 'learning_rate/action_model': 5.3491970528682124e-06, 'epoch': 0.28}) +[2026-06-24 11:24:31,227][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.08155246078968048, 'train/grad_norm_pre_clip': 82.14039611816406, 'timing/data': 0.0003176960162818432, 'timing/model': 1.0720333810895681, 'learning_rate/qwen_vl_interface': 5.342996710993949e-07, 'learning_rate/action_model': 5.342996710993949e-06, 'epoch': 0.28}) +[2026-06-24 11:24:32,301][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.009853453375399113, 'train/grad_norm_pre_clip': 16.55950927734375, 'timing/data': 0.00031994120217859745, 'timing/model': 1.0727702239528298, 'learning_rate/qwen_vl_interface': 5.336851710489682e-07, 'learning_rate/action_model': 5.336851710489682e-06, 'epoch': 0.28}) +[2026-06-24 11:24:33,374][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.011976001784205437, 'train/grad_norm_pre_clip': 14.145467758178711, 'timing/data': 0.00030733714811503887, 'timing/model': 1.0712281980086118, 'learning_rate/qwen_vl_interface': 5.330762058566921e-07, 'learning_rate/action_model': 5.33076205856692e-06, 'epoch': 0.28}) +[2026-06-24 11:24:34,732][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.0725923627614975, 'train/grad_norm_pre_clip': 118.82147979736328, 'timing/data': 0.0003067958168685436, 'timing/model': 1.3560433469247073, 'learning_rate/qwen_vl_interface': 5.324727762372203e-07, 'learning_rate/action_model': 5.324727762372202e-06, 'epoch': 0.28}) +[2026-06-24 11:24:35,804][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.023127440363168716, 'train/grad_norm_pre_clip': 25.160442352294922, 'timing/data': 0.00039535295218229294, 'timing/model': 1.0709255179390311, 'learning_rate/qwen_vl_interface': 5.318748828987106e-07, 'learning_rate/action_model': 5.318748828987106e-06, 'epoch': 0.28}) +[2026-06-24 11:24:36,879][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.21308307349681854, 'train/grad_norm_pre_clip': 35.55662536621094, 'timing/data': 0.0003134678117930889, 'timing/model': 1.073458543047309, 'learning_rate/qwen_vl_interface': 5.312825265428249e-07, 'learning_rate/action_model': 5.3128252654282495e-06, 'epoch': 0.28}) +[2026-06-24 11:24:37,955][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.004434092435985804, 'train/grad_norm_pre_clip': 5.1022562980651855, 'timing/data': 0.00031570298597216606, 'timing/model': 1.0737649688962847, 'learning_rate/qwen_vl_interface': 5.306957078647266e-07, 'learning_rate/action_model': 5.306957078647265e-06, 'epoch': 0.28}) +[2026-06-24 11:24:39,028][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.011951012536883354, 'train/grad_norm_pre_clip': 15.41783332824707, 'timing/data': 0.0003033899702131748, 'timing/model': 1.0714231999590993, 'learning_rate/qwen_vl_interface': 5.301144275530798e-07, 'learning_rate/action_model': 5.301144275530798e-06, 'epoch': 0.28}) +[2026-06-24 11:24:40,102][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.05297442525625229, 'train/grad_norm_pre_clip': 29.649803161621094, 'timing/data': 0.0003239680081605911, 'timing/model': 1.0723728029988706, 'learning_rate/qwen_vl_interface': 5.29538686290049e-07, 'learning_rate/action_model': 5.29538686290049e-06, 'epoch': 0.28}) +[2026-06-24 11:24:41,174][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.18867157399654388, 'train/grad_norm_pre_clip': 93.89225769042969, 'timing/data': 0.00031057302840054035, 'timing/model': 1.0709368649404496, 'learning_rate/qwen_vl_interface': 5.289684847512987e-07, 'learning_rate/action_model': 5.289684847512986e-06, 'epoch': 0.28}) +[2026-06-24 11:24:42,246][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.09505295008420944, 'train/grad_norm_pre_clip': 61.440528869628906, 'timing/data': 0.0003453679382801056, 'timing/model': 1.0706301787868142, 'learning_rate/qwen_vl_interface': 5.284038236059925e-07, 'learning_rate/action_model': 5.284038236059924e-06, 'epoch': 0.28}) +[2026-06-24 11:24:43,318][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.08280547708272934, 'train/grad_norm_pre_clip': 63.001502990722656, 'timing/data': 0.0003522119950503111, 'timing/model': 1.0699582470115274, 'learning_rate/qwen_vl_interface': 5.278447035167919e-07, 'learning_rate/action_model': 5.278447035167919e-06, 'epoch': 0.28}) +[2026-06-24 11:24:44,394][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.031475480645895004, 'train/grad_norm_pre_clip': 20.463457107543945, 'timing/data': 0.0003448969218879938, 'timing/model': 1.074004493886605, 'learning_rate/qwen_vl_interface': 5.272911251398551e-07, 'learning_rate/action_model': 5.272911251398551e-06, 'epoch': 0.28}) +[2026-06-24 11:24:45,469][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.21276669204235077, 'train/grad_norm_pre_clip': 20.929365158081055, 'timing/data': 0.0003113651182502508, 'timing/model': 1.0737533520441502, 'learning_rate/qwen_vl_interface': 5.267430891248371e-07, 'learning_rate/action_model': 5.2674308912483705e-06, 'epoch': 0.28}) +[2026-06-24 11:24:46,542][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.059885501861572266, 'train/grad_norm_pre_clip': 97.65633392333984, 'timing/data': 0.0003297301009297371, 'timing/model': 1.0718306410126388, 'learning_rate/qwen_vl_interface': 5.262005961148884e-07, 'learning_rate/action_model': 5.2620059611488835e-06, 'epoch': 0.28}) +[2026-06-24 11:24:47,617][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.19689476490020752, 'train/grad_norm_pre_clip': 69.01851654052734, 'timing/data': 0.00031115394085645676, 'timing/model': 1.072949631838128, 'learning_rate/qwen_vl_interface': 5.256636467466554e-07, 'learning_rate/action_model': 5.256636467466554e-06, 'epoch': 0.28}) +[2026-06-24 11:24:48,695][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.22627690434455872, 'train/grad_norm_pre_clip': 170.94505310058594, 'timing/data': 0.0003216641489416361, 'timing/model': 1.0769793780054897, 'learning_rate/qwen_vl_interface': 5.251322416502782e-07, 'learning_rate/action_model': 5.251322416502782e-06, 'epoch': 0.28}) +[2026-06-24 11:24:49,769][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.2777578830718994, 'train/grad_norm_pre_clip': 154.0155487060547, 'timing/data': 0.000309912022203207, 'timing/model': 1.0722174949478358, 'learning_rate/qwen_vl_interface': 5.246063814493896e-07, 'learning_rate/action_model': 5.2460638144938955e-06, 'epoch': 0.28}) +[2026-06-24 11:24:50,844][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.021647589281201363, 'train/grad_norm_pre_clip': 18.321802139282227, 'timing/data': 0.0003212138544768095, 'timing/model': 1.0730089710559696, 'learning_rate/qwen_vl_interface': 5.240860667611162e-07, 'learning_rate/action_model': 5.240860667611162e-06, 'epoch': 0.28}) +[2026-06-24 11:24:52,199][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.20371273159980774, 'train/grad_norm_pre_clip': 92.76211547851562, 'timing/data': 0.0003386959433555603, 'timing/model': 1.3531982807908207, 'learning_rate/qwen_vl_interface': 5.235712981960761e-07, 'learning_rate/action_model': 5.23571298196076e-06, 'epoch': 0.28}) +[2026-06-24 11:24:53,272][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.033199746161699295, 'train/grad_norm_pre_clip': 33.934600830078125, 'timing/data': 0.00031057302840054035, 'timing/model': 1.0722291159909219, 'learning_rate/qwen_vl_interface': 5.230620763583791e-07, 'learning_rate/action_model': 5.230620763583792e-06, 'epoch': 0.28}) +[2026-06-24 11:24:54,348][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.14712406694889069, 'train/grad_norm_pre_clip': 68.91755676269531, 'timing/data': 0.0003411110956221819, 'timing/model': 1.0737738851457834, 'learning_rate/qwen_vl_interface': 5.225584018456259e-07, 'learning_rate/action_model': 5.225584018456258e-06, 'epoch': 0.28}) +[2026-06-24 11:24:55,426][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.05496428906917572, 'train/grad_norm_pre_clip': 58.88740539550781, 'timing/data': 0.00030933087691664696, 'timing/model': 1.07674639695324, 'learning_rate/qwen_vl_interface': 5.220602752489054e-07, 'learning_rate/action_model': 5.220602752489054e-06, 'epoch': 0.28}) +[2026-06-24 11:24:56,498][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.19910277426242828, 'train/grad_norm_pre_clip': 105.12368774414062, 'timing/data': 0.00032666302286088467, 'timing/model': 1.0699234341736883, 'learning_rate/qwen_vl_interface': 5.215676971527976e-07, 'learning_rate/action_model': 5.2156769715279754e-06, 'epoch': 0.28}) +[2026-06-24 11:24:57,573][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.1538594365119934, 'train/grad_norm_pre_clip': 86.89324951171875, 'timing/data': 0.00047630397602915764, 'timing/model': 1.073248096043244, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 0.28}) +[2026-06-24 11:24:58,649][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.029369620606303215, 'train/grad_norm_pre_clip': 25.632652282714844, 'timing/data': 0.00029476406052708626, 'timing/model': 1.0751150909345597, 'learning_rate/qwen_vl_interface': 5.205991887681787e-07, 'learning_rate/action_model': 5.205991887681787e-06, 'epoch': 0.28}) +[2026-06-24 11:24:59,723][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.22173988819122314, 'train/grad_norm_pre_clip': 67.98011779785156, 'timing/data': 0.00047273794189095497, 'timing/model': 1.0713121199514717, 'learning_rate/qwen_vl_interface': 5.201232596162664e-07, 'learning_rate/action_model': 5.2012325961626646e-06, 'epoch': 0.28}) +[2026-06-24 11:25:00,796][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.03521708399057388, 'train/grad_norm_pre_clip': 29.548322677612305, 'timing/data': 0.0005133540835231543, 'timing/model': 1.0719982569571584, 'learning_rate/qwen_vl_interface': 5.196528812381624e-07, 'learning_rate/action_model': 5.196528812381624e-06, 'epoch': 0.28}) +[2026-06-24 11:25:01,872][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.05473097786307335, 'train/grad_norm_pre_clip': 50.38329315185547, 'timing/data': 0.00047749700024724007, 'timing/model': 1.0732138191815466, 'learning_rate/qwen_vl_interface': 5.191880541858819e-07, 'learning_rate/action_model': 5.191880541858819e-06, 'epoch': 0.28}) +[2026-06-24 11:25:02,948][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.01892136223614216, 'train/grad_norm_pre_clip': 22.074560165405273, 'timing/data': 0.0004210001789033413, 'timing/model': 1.0739835938438773, 'learning_rate/qwen_vl_interface': 5.187287790049259e-07, 'learning_rate/action_model': 5.1872877900492595e-06, 'epoch': 0.28}) +[2026-06-24 11:25:04,023][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.017362721264362335, 'train/grad_norm_pre_clip': 20.131175994873047, 'timing/data': 0.0003336870577186346, 'timing/model': 1.0733941970393062, 'learning_rate/qwen_vl_interface': 5.182750562342784e-07, 'learning_rate/action_model': 5.182750562342784e-06, 'epoch': 0.28}) +[2026-06-24 11:25:05,097][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.03704611212015152, 'train/grad_norm_pre_clip': 41.9030647277832, 'timing/data': 0.00036553596146404743, 'timing/model': 1.072402753867209, 'learning_rate/qwen_vl_interface': 5.178268864064092e-07, 'learning_rate/action_model': 5.178268864064092e-06, 'epoch': 0.28}) +[2026-06-24 11:25:06,170][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.09720946848392487, 'train/grad_norm_pre_clip': 139.1481170654297, 'timing/data': 0.0003675499465316534, 'timing/model': 1.071950954850763, 'learning_rate/qwen_vl_interface': 5.173842700472699e-07, 'learning_rate/action_model': 5.173842700472698e-06, 'epoch': 0.28}) +[2026-06-24 11:25:07,245][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.16101351380348206, 'train/grad_norm_pre_clip': 105.17452239990234, 'timing/data': 0.00038179708644747734, 'timing/model': 1.072482313029468, 'learning_rate/qwen_vl_interface': 5.169472076762958e-07, 'learning_rate/action_model': 5.169472076762958e-06, 'epoch': 0.28}) +[2026-06-24 11:25:08,319][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.12417199462652206, 'train/grad_norm_pre_clip': 49.63854217529297, 'timing/data': 0.00033015012741088867, 'timing/model': 1.0726518998853862, 'learning_rate/qwen_vl_interface': 5.165156998064041e-07, 'learning_rate/action_model': 5.165156998064041e-06, 'epoch': 0.28}) +[2026-06-24 11:25:09,676][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.15071581304073334, 'train/grad_norm_pre_clip': 79.74026489257812, 'timing/data': 0.0003648551646620035, 'timing/model': 1.355825741076842, 'learning_rate/qwen_vl_interface': 5.16089746943993e-07, 'learning_rate/action_model': 5.16089746943993e-06, 'epoch': 0.28}) +[2026-06-24 11:25:10,750][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.004914950113743544, 'train/grad_norm_pre_clip': 3.765047788619995, 'timing/data': 0.00031559192575514317, 'timing/model': 1.0715469340793788, 'learning_rate/qwen_vl_interface': 5.156693495889422e-07, 'learning_rate/action_model': 5.156693495889422e-06, 'epoch': 0.28}) +[2026-06-24 11:25:11,826][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.0033277980983257294, 'train/grad_norm_pre_clip': 1.6117209196090698, 'timing/data': 0.00030762702226638794, 'timing/model': 1.0749878950882703, 'learning_rate/qwen_vl_interface': 5.152545082346106e-07, 'learning_rate/action_model': 5.152545082346105e-06, 'epoch': 0.28}) +[2026-06-24 11:25:12,903][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.12713685631752014, 'train/grad_norm_pre_clip': 74.89848327636719, 'timing/data': 0.00041835615411400795, 'timing/model': 1.074849353870377, 'learning_rate/qwen_vl_interface': 5.148452233678393e-07, 'learning_rate/action_model': 5.148452233678393e-06, 'epoch': 0.28}) +[2026-06-24 11:25:13,977][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.36422041058540344, 'train/grad_norm_pre_clip': 96.4872055053711, 'timing/data': 0.0002957850228995085, 'timing/model': 1.072462640935555, 'learning_rate/qwen_vl_interface': 5.144414954689448e-07, 'learning_rate/action_model': 5.144414954689448e-06, 'epoch': 0.28}) +[2026-06-24 11:25:15,048][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.028131699189543724, 'train/grad_norm_pre_clip': 36.39677810668945, 'timing/data': 0.0003353699576109648, 'timing/model': 1.0699450229294598, 'learning_rate/qwen_vl_interface': 5.140433250117262e-07, 'learning_rate/action_model': 5.1404332501172625e-06, 'epoch': 0.28}) +[2026-06-24 11:25:16,123][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.2074117362499237, 'train/grad_norm_pre_clip': 129.1442413330078, 'timing/data': 0.00038110604509711266, 'timing/model': 1.0731813500169665, 'learning_rate/qwen_vl_interface': 5.136507124634576e-07, 'learning_rate/action_model': 5.136507124634575e-06, 'epoch': 0.28}) +[2026-06-24 11:25:17,196][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.007160624023526907, 'train/grad_norm_pre_clip': 7.134449481964111, 'timing/data': 0.00031619402579963207, 'timing/model': 1.0710857340600342, 'learning_rate/qwen_vl_interface': 5.132636582848919e-07, 'learning_rate/action_model': 5.13263658284892e-06, 'epoch': 0.28}) +[2026-06-24 11:25:18,273][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.04150604456663132, 'train/grad_norm_pre_clip': 53.0373420715332, 'timing/data': 0.0003385960590094328, 'timing/model': 1.0754920279141515, 'learning_rate/qwen_vl_interface': 5.128821629302588e-07, 'learning_rate/action_model': 5.128821629302588e-06, 'epoch': 0.28}) +[2026-06-24 11:25:19,347][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.16179177165031433, 'train/grad_norm_pre_clip': 105.24855041503906, 'timing/data': 0.00032054190523922443, 'timing/model': 1.0727073978632689, 'learning_rate/qwen_vl_interface': 5.125062268472642e-07, 'learning_rate/action_model': 5.125062268472641e-06, 'epoch': 0.28}) +[2026-06-24 11:25:20,423][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.024906370788812637, 'train/grad_norm_pre_clip': 49.58468246459961, 'timing/data': 0.0003354698419570923, 'timing/model': 1.074618746060878, 'learning_rate/qwen_vl_interface': 5.121358504770903e-07, 'learning_rate/action_model': 5.1213585047709025e-06, 'epoch': 0.28}) +[2026-06-24 11:25:21,497][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.012130949646234512, 'train/grad_norm_pre_clip': 14.40949535369873, 'timing/data': 0.0003567601088434458, 'timing/model': 1.072762280004099, 'learning_rate/qwen_vl_interface': 5.11771034254394e-07, 'learning_rate/action_model': 5.117710342543941e-06, 'epoch': 0.28}) +[2026-06-24 11:25:22,573][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.12344755232334137, 'train/grad_norm_pre_clip': 80.9107894897461, 'timing/data': 0.0003306400030851364, 'timing/model': 1.074005413101986, 'learning_rate/qwen_vl_interface': 5.114117786073083e-07, 'learning_rate/action_model': 5.114117786073083e-06, 'epoch': 0.28}) +[2026-06-24 11:25:23,646][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.17985254526138306, 'train/grad_norm_pre_clip': 45.04039764404297, 'timing/data': 0.0003168950788676739, 'timing/model': 1.071883148048073, 'learning_rate/qwen_vl_interface': 5.110580839574388e-07, 'learning_rate/action_model': 5.110580839574388e-06, 'epoch': 0.28}) +[2026-06-24 11:25:24,721][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.0009014817187562585, 'train/grad_norm_pre_clip': 0.860161304473877, 'timing/data': 0.00030535412952303886, 'timing/model': 1.0727657650131732, 'learning_rate/qwen_vl_interface': 5.107099507198657e-07, 'learning_rate/action_model': 5.107099507198657e-06, 'epoch': 0.28}) +[2026-06-24 11:25:25,793][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.03674529492855072, 'train/grad_norm_pre_clip': 48.316951751708984, 'timing/data': 0.0003292080946266651, 'timing/model': 1.0709774361457676, 'learning_rate/qwen_vl_interface': 5.103673793031439e-07, 'learning_rate/action_model': 5.103673793031439e-06, 'epoch': 0.28}) +[2026-06-24 11:25:27,147][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.21367156505584717, 'train/grad_norm_pre_clip': 65.91036987304688, 'timing/data': 0.0003290078602731228, 'timing/model': 1.3518439920153469, 'learning_rate/qwen_vl_interface': 5.100303701092992e-07, 'learning_rate/action_model': 5.100303701092991e-06, 'epoch': 0.28}) +[2026-06-24 11:25:28,219][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.1858665645122528, 'train/grad_norm_pre_clip': 82.08845520019531, 'timing/data': 0.00032973894849419594, 'timing/model': 1.070861245971173, 'learning_rate/qwen_vl_interface': 5.096989235338312e-07, 'learning_rate/action_model': 5.096989235338312e-06, 'epoch': 0.28}) +[2026-06-24 11:25:29,294][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.15452739596366882, 'train/grad_norm_pre_clip': 59.45081329345703, 'timing/data': 0.0003280059900134802, 'timing/model': 1.0730988171417266, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 0.28}) +[2026-06-24 11:25:30,368][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.048894140869379044, 'train/grad_norm_pre_clip': 270.3531494140625, 'timing/data': 0.0003164240624755621, 'timing/model': 1.0722219299059361, 'learning_rate/qwen_vl_interface': 5.09052719787379e-07, 'learning_rate/action_model': 5.09052719787379e-06, 'epoch': 0.28}) +[2026-06-24 11:25:31,440][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.11646980047225952, 'train/grad_norm_pre_clip': 76.7153549194336, 'timing/data': 0.0003375639207661152, 'timing/model': 1.0712424400262535, 'learning_rate/qwen_vl_interface': 5.087379633747516e-07, 'learning_rate/action_model': 5.087379633747515e-06, 'epoch': 0.28}) +[2026-06-24 11:25:32,513][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.03338392451405525, 'train/grad_norm_pre_clip': 43.474571228027344, 'timing/data': 0.00033264397643506527, 'timing/model': 1.0707244970835745, 'learning_rate/qwen_vl_interface': 5.084287710972115e-07, 'learning_rate/action_model': 5.084287710972115e-06, 'epoch': 0.28}) +[2026-06-24 11:25:33,587][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.13058620691299438, 'train/grad_norm_pre_clip': 90.24554443359375, 'timing/data': 0.00031163590028882027, 'timing/model': 1.07284796400927, 'learning_rate/qwen_vl_interface': 5.081251433176135e-07, 'learning_rate/action_model': 5.081251433176135e-06, 'epoch': 0.29}) +[2026-06-24 11:25:34,661][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.11922017484903336, 'train/grad_norm_pre_clip': 83.31062316894531, 'timing/data': 0.0003042819444090128, 'timing/model': 1.0726576559245586, 'learning_rate/qwen_vl_interface': 5.078270803922819e-07, 'learning_rate/action_model': 5.078270803922819e-06, 'epoch': 0.29}) +[2026-06-24 11:25:35,735][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.06794707477092743, 'train/grad_norm_pre_clip': 75.01617431640625, 'timing/data': 0.00030608405359089375, 'timing/model': 1.0717952859122306, 'learning_rate/qwen_vl_interface': 5.075345826710093e-07, 'learning_rate/action_model': 5.0753458267100935e-06, 'epoch': 0.29}) +[2026-06-24 11:25:36,825][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.041362207382917404, 'train/grad_norm_pre_clip': 19.172365188598633, 'timing/data': 0.00031303800642490387, 'timing/model': 1.0732753651682287, 'learning_rate/qwen_vl_interface': 5.072476504970592e-07, 'learning_rate/action_model': 5.072476504970592e-06, 'epoch': 0.29}) +[2026-06-24 11:25:37,906][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.011734319850802422, 'train/grad_norm_pre_clip': 10.873343467712402, 'timing/data': 0.0004694222006946802, 'timing/model': 1.079281702870503, 'learning_rate/qwen_vl_interface': 5.069662842071619e-07, 'learning_rate/action_model': 5.069662842071619e-06, 'epoch': 0.29}) +[2026-06-24 11:25:38,982][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.18967951834201813, 'train/grad_norm_pre_clip': 122.32347869873047, 'timing/data': 0.0003073869738727808, 'timing/model': 1.0740233878605068, 'learning_rate/qwen_vl_interface': 5.06690484131516e-07, 'learning_rate/action_model': 5.06690484131516e-06, 'epoch': 0.29}) +[2026-06-24 11:25:40,057][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.015379863791167736, 'train/grad_norm_pre_clip': 12.893685340881348, 'timing/data': 0.00033121206797659397, 'timing/model': 1.0736632919870317, 'learning_rate/qwen_vl_interface': 5.064202505937891e-07, 'learning_rate/action_model': 5.06420250593789e-06, 'epoch': 0.29}) +[2026-06-24 11:25:41,131][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.0694684386253357, 'train/grad_norm_pre_clip': 63.05523681640625, 'timing/data': 0.0003435548860579729, 'timing/model': 1.0729043860919774, 'learning_rate/qwen_vl_interface': 5.061555839111142e-07, 'learning_rate/action_model': 5.061555839111142e-06, 'epoch': 0.29}) +[2026-06-24 11:25:42,205][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.01513968501240015, 'train/grad_norm_pre_clip': 23.189420700073242, 'timing/data': 0.00030454201623797417, 'timing/model': 1.0719252771232277, 'learning_rate/qwen_vl_interface': 5.058964843940938e-07, 'learning_rate/action_model': 5.058964843940938e-06, 'epoch': 0.29}) +[2026-06-24 11:25:43,281][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.010939194820821285, 'train/grad_norm_pre_clip': 12.792536735534668, 'timing/data': 0.00032805604860186577, 'timing/model': 1.0747512020170689, 'learning_rate/qwen_vl_interface': 5.056429523467948e-07, 'learning_rate/action_model': 5.056429523467948e-06, 'epoch': 0.29}) +[2026-06-24 11:25:44,643][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.12503382563591003, 'train/grad_norm_pre_clip': 39.94029235839844, 'timing/data': 0.0003340071998536587, 'timing/model': 1.3597384740132838, 'learning_rate/qwen_vl_interface': 5.05394988066751e-07, 'learning_rate/action_model': 5.053949880667509e-06, 'epoch': 0.29}) +[2026-06-24 11:25:45,718][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.2829436957836151, 'train/grad_norm_pre_clip': 99.20362854003906, 'timing/data': 0.00037347106263041496, 'timing/model': 1.0737870610319078, 'learning_rate/qwen_vl_interface': 5.051525918449625e-07, 'learning_rate/action_model': 5.051525918449625e-06, 'epoch': 0.29}) +[2026-06-24 11:25:46,794][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.1040300652384758, 'train/grad_norm_pre_clip': 39.7147331237793, 'timing/data': 0.0003355611115694046, 'timing/model': 1.074058088939637, 'learning_rate/qwen_vl_interface': 5.049157639658952e-07, 'learning_rate/action_model': 5.049157639658953e-06, 'epoch': 0.29}) +[2026-06-24 11:25:47,870][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.07106771320104599, 'train/grad_norm_pre_clip': 39.05457305908203, 'timing/data': 0.00032073212787508965, 'timing/model': 1.0743481740355492, 'learning_rate/qwen_vl_interface': 5.046845047074795e-07, 'learning_rate/action_model': 5.046845047074795e-06, 'epoch': 0.29}) +[2026-06-24 11:25:48,945][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.07810931652784348, 'train/grad_norm_pre_clip': 29.791194915771484, 'timing/data': 0.00030953087843954563, 'timing/model': 1.0742614699993283, 'learning_rate/qwen_vl_interface': 5.044588143411106e-07, 'learning_rate/action_model': 5.044588143411106e-06, 'epoch': 0.29}) +[2026-06-24 11:25:50,019][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.019480900838971138, 'train/grad_norm_pre_clip': 71.49600982666016, 'timing/data': 0.0004984859842807055, 'timing/model': 1.0718698590062559, 'learning_rate/qwen_vl_interface': 5.042386931316493e-07, 'learning_rate/action_model': 5.042386931316493e-06, 'epoch': 0.29}) +[2026-06-24 11:25:51,096][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.05772782489657402, 'train/grad_norm_pre_clip': 38.22639083862305, 'timing/data': 0.00031285802833735943, 'timing/model': 1.074569377116859, 'learning_rate/qwen_vl_interface': 5.040241413374195e-07, 'learning_rate/action_model': 5.040241413374194e-06, 'epoch': 0.29}) +[2026-06-24 11:25:52,172][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.12489179521799088, 'train/grad_norm_pre_clip': 67.1128158569336, 'timing/data': 0.00033162301406264305, 'timing/model': 1.0751443558838218, 'learning_rate/qwen_vl_interface': 5.038151592102107e-07, 'learning_rate/action_model': 5.0381515921021065e-06, 'epoch': 0.29}) +[2026-06-24 11:25:53,250][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.15903759002685547, 'train/grad_norm_pre_clip': 64.37580108642578, 'timing/data': 0.00033423700369894505, 'timing/model': 1.0756634899880737, 'learning_rate/qwen_vl_interface': 5.036117469952742e-07, 'learning_rate/action_model': 5.0361174699527415e-06, 'epoch': 0.29}) +[2026-06-24 11:25:54,325][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.07176299393177032, 'train/grad_norm_pre_clip': 37.49787902832031, 'timing/data': 0.0003037611022591591, 'timing/model': 1.0742928660474718, 'learning_rate/qwen_vl_interface': 5.034139049313255e-07, 'learning_rate/action_model': 5.034139049313254e-06, 'epoch': 0.29}) +[2026-06-24 11:25:55,399][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.03610031306743622, 'train/grad_norm_pre_clip': 23.84474754333496, 'timing/data': 0.0003127770032733679, 'timing/model': 1.0720533900894225, 'learning_rate/qwen_vl_interface': 5.032216332505435e-07, 'learning_rate/action_model': 5.032216332505435e-06, 'epoch': 0.29}) +[2026-06-24 11:25:56,474][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.07100202143192291, 'train/grad_norm_pre_clip': 55.26177978515625, 'timing/data': 0.0003230359870940447, 'timing/model': 1.0732171151321381, 'learning_rate/qwen_vl_interface': 5.030349321785703e-07, 'learning_rate/action_model': 5.030349321785703e-06, 'epoch': 0.29}) +[2026-06-24 11:25:57,546][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.08404236286878586, 'train/grad_norm_pre_clip': 44.6920051574707, 'timing/data': 0.00035533681511878967, 'timing/model': 1.0706959501840174, 'learning_rate/qwen_vl_interface': 5.028538019345091e-07, 'learning_rate/action_model': 5.028538019345091e-06, 'epoch': 0.29}) +[2026-06-24 11:25:58,620][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.007730638142675161, 'train/grad_norm_pre_clip': 7.465104579925537, 'timing/data': 0.00047203595750033855, 'timing/model': 1.0723023358732462, 'learning_rate/qwen_vl_interface': 5.026782427309269e-07, 'learning_rate/action_model': 5.0267824273092684e-06, 'epoch': 0.29}) +[2026-06-24 11:25:59,696][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.0030194963328540325, 'train/grad_norm_pre_clip': 1.5665218830108643, 'timing/data': 0.000370916910469532, 'timing/model': 1.0736706049647182, 'learning_rate/qwen_vl_interface': 5.025082547738522e-07, 'learning_rate/action_model': 5.025082547738522e-06, 'epoch': 0.29}) +[2026-06-24 11:26:00,772][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.12344830483198166, 'train/grad_norm_pre_clip': 77.7368392944336, 'timing/data': 0.0003173758741468191, 'timing/model': 1.0750466778408736, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 0.29}) +[2026-06-24 11:26:02,122][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.1385132521390915, 'train/grad_norm_pre_clip': 58.41370391845703, 'timing/data': 0.00032494007609784603, 'timing/model': 1.3484051211271435, 'learning_rate/qwen_vl_interface': 5.021849933906476e-07, 'learning_rate/action_model': 5.021849933906476e-06, 'epoch': 0.29}) +[2026-06-24 11:26:03,196][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.021103501319885254, 'train/grad_norm_pre_clip': 20.127450942993164, 'timing/data': 0.00032348791137337685, 'timing/model': 1.0717750031035393, 'learning_rate/qwen_vl_interface': 5.020317203438827e-07, 'learning_rate/action_model': 5.020317203438827e-06, 'epoch': 0.29}) +[2026-06-24 11:26:04,268][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.2252974808216095, 'train/grad_norm_pre_clip': 118.29232025146484, 'timing/data': 0.00037283007986843586, 'timing/model': 1.0713115821126848, 'learning_rate/qwen_vl_interface': 5.018840193023556e-07, 'learning_rate/action_model': 5.018840193023556e-06, 'epoch': 0.29}) +[2026-06-24 11:26:05,344][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.10544481873512268, 'train/grad_norm_pre_clip': 72.6130142211914, 'timing/data': 0.00030215783044695854, 'timing/model': 1.0744200688786805, 'learning_rate/qwen_vl_interface': 5.017418904394012e-07, 'learning_rate/action_model': 5.017418904394012e-06, 'epoch': 0.29}) +[2026-06-24 11:26:06,420][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.22501686215400696, 'train/grad_norm_pre_clip': 76.06814575195312, 'timing/data': 0.0003273149486631155, 'timing/model': 1.0741794269997627, 'learning_rate/qwen_vl_interface': 5.016053339218154e-07, 'learning_rate/action_model': 5.016053339218154e-06, 'epoch': 0.29}) +[2026-06-24 11:26:07,493][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.028791610151529312, 'train/grad_norm_pre_clip': 29.445659637451172, 'timing/data': 0.00031063309870660305, 'timing/model': 1.0707496758550406, 'learning_rate/qwen_vl_interface': 5.014743499098559e-07, 'learning_rate/action_model': 5.014743499098559e-06, 'epoch': 0.29}) +[2026-06-24 11:26:08,571][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.003250067587941885, 'train/grad_norm_pre_clip': 3.540679693222046, 'timing/data': 0.0002969370689243078, 'timing/model': 1.0773734729737043, 'learning_rate/qwen_vl_interface': 5.013489385572386e-07, 'learning_rate/action_model': 5.013489385572385e-06, 'epoch': 0.29}) +[2026-06-24 11:26:09,646][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.0205768421292305, 'train/grad_norm_pre_clip': 34.37480926513672, 'timing/data': 0.00032007088884711266, 'timing/model': 1.0731635771226138, 'learning_rate/qwen_vl_interface': 5.012291000111409e-07, 'learning_rate/action_model': 5.012291000111409e-06, 'epoch': 0.29}) +[2026-06-24 11:26:10,717][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.12821877002716064, 'train/grad_norm_pre_clip': 121.00226593017578, 'timing/data': 0.00038532307371497154, 'timing/model': 1.069112349068746, 'learning_rate/qwen_vl_interface': 5.011148344122004e-07, 'learning_rate/action_model': 5.011148344122004e-06, 'epoch': 0.29}) +[2026-06-24 11:26:11,791][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.0986865907907486, 'train/grad_norm_pre_clip': 144.26803588867188, 'timing/data': 0.00033081183210015297, 'timing/model': 1.072914948919788, 'learning_rate/qwen_vl_interface': 5.010061418945141e-07, 'learning_rate/action_model': 5.010061418945141e-06, 'epoch': 0.29}) +[2026-06-24 11:26:12,866][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.062122661620378494, 'train/grad_norm_pre_clip': 73.12479400634766, 'timing/data': 0.0002968769986182451, 'timing/model': 1.0734398651402444, 'learning_rate/qwen_vl_interface': 5.009030225856385e-07, 'learning_rate/action_model': 5.009030225856385e-06, 'epoch': 0.29}) +[2026-06-24 11:26:13,942][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.013762025162577629, 'train/grad_norm_pre_clip': 11.90569019317627, 'timing/data': 0.00035769096575677395, 'timing/model': 1.0742187660653144, 'learning_rate/qwen_vl_interface': 5.0080547660659e-07, 'learning_rate/action_model': 5.0080547660659e-06, 'epoch': 0.29}) +[2026-06-24 11:26:15,019][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.07391300797462463, 'train/grad_norm_pre_clip': 42.16312789916992, 'timing/data': 0.0003229370340704918, 'timing/model': 1.0751387740019709, 'learning_rate/qwen_vl_interface': 5.00713504071844e-07, 'learning_rate/action_model': 5.00713504071844e-06, 'epoch': 0.29}) +[2026-06-24 11:26:16,094][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.036104716360569, 'train/grad_norm_pre_clip': 37.21039962768555, 'timing/data': 0.0003109739627689123, 'timing/model': 1.0729267499409616, 'learning_rate/qwen_vl_interface': 5.006271050893358e-07, 'learning_rate/action_model': 5.0062710508933585e-06, 'epoch': 0.29}) +[2026-06-24 11:26:17,167][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.0018870967905968428, 'train/grad_norm_pre_clip': 1.3067513704299927, 'timing/data': 0.0003160329069942236, 'timing/model': 1.0718717682175338, 'learning_rate/qwen_vl_interface': 5.005462797604594e-07, 'learning_rate/action_model': 5.005462797604594e-06, 'epoch': 0.29}) +[2026-06-24 11:26:18,242][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.0301828533411026, 'train/grad_norm_pre_clip': 17.81427764892578, 'timing/data': 0.00031330808997154236, 'timing/model': 1.0737165920436382, 'learning_rate/qwen_vl_interface': 5.004710281800675e-07, 'learning_rate/action_model': 5.004710281800674e-06, 'epoch': 0.29}) +[2026-06-24 11:26:19,602][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.09423264861106873, 'train/grad_norm_pre_clip': 37.580142974853516, 'timing/data': 0.0005214789416640997, 'timing/model': 1.3579427590593696, 'learning_rate/qwen_vl_interface': 5.004013504364721e-07, 'learning_rate/action_model': 5.00401350436472e-06, 'epoch': 0.29}) +[2026-06-24 11:26:20,678][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.015357679687440395, 'train/grad_norm_pre_clip': 12.39312744140625, 'timing/data': 0.0003101020120084286, 'timing/model': 1.0745307391043752, 'learning_rate/qwen_vl_interface': 5.00337246611444e-07, 'learning_rate/action_model': 5.00337246611444e-06, 'epoch': 0.29}) +[2026-06-24 11:26:21,754][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.16116860508918762, 'train/grad_norm_pre_clip': 74.71302032470703, 'timing/data': 0.0005537099204957485, 'timing/model': 1.0733086639083922, 'learning_rate/qwen_vl_interface': 5.002787167802127e-07, 'learning_rate/action_model': 5.002787167802128e-06, 'epoch': 0.29}) +[2026-06-24 11:26:22,827][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.4092990756034851, 'train/grad_norm_pre_clip': 144.19985961914062, 'timing/data': 0.0003468110226094723, 'timing/model': 1.0717393159866333, 'learning_rate/qwen_vl_interface': 5.002257610114659e-07, 'learning_rate/action_model': 5.002257610114658e-06, 'epoch': 0.29}) +[2026-06-24 11:26:23,903][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.02762308344244957, 'train/grad_norm_pre_clip': 73.03453063964844, 'timing/data': 0.00032693403773009777, 'timing/model': 1.0746042949613184, 'learning_rate/qwen_vl_interface': 5.001783793673502e-07, 'learning_rate/action_model': 5.001783793673501e-06, 'epoch': 0.29}) +[2026-06-24 11:26:24,979][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.15763217210769653, 'train/grad_norm_pre_clip': 63.79631423950195, 'timing/data': 0.0005782661028206348, 'timing/model': 1.0731934860814363, 'learning_rate/qwen_vl_interface': 5.001365719034709e-07, 'learning_rate/action_model': 5.00136571903471e-06, 'epoch': 0.29}) +[2026-06-24 11:26:26,054][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.04991802200675011, 'train/grad_norm_pre_clip': 42.22878646850586, 'timing/data': 0.0003010861109942198, 'timing/model': 1.0739901210181415, 'learning_rate/qwen_vl_interface': 5.001003386688916e-07, 'learning_rate/action_model': 5.001003386688916e-06, 'epoch': 0.29}) +[2026-06-24 11:26:27,129][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.053897153586149216, 'train/grad_norm_pre_clip': 55.65354537963867, 'timing/data': 0.00032643298618495464, 'timing/model': 1.0733193610794842, 'learning_rate/qwen_vl_interface': 5.00069679706133e-07, 'learning_rate/action_model': 5.000696797061331e-06, 'epoch': 0.29}) +[2026-06-24 11:26:28,204][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.16277317702770233, 'train/grad_norm_pre_clip': 67.30294799804688, 'timing/data': 0.0003071569371968508, 'timing/model': 1.0728806471452117, 'learning_rate/qwen_vl_interface': 5.000445950511764e-07, 'learning_rate/action_model': 5.000445950511764e-06, 'epoch': 0.29}) +[2026-06-24 11:26:29,280][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.03272462636232376, 'train/grad_norm_pre_clip': 56.40065383911133, 'timing/data': 0.00032165413722395897, 'timing/model': 1.0750015079975128, 'learning_rate/qwen_vl_interface': 5.000250847334593e-07, 'learning_rate/action_model': 5.000250847334593e-06, 'epoch': 0.29}) +[2026-06-24 11:26:30,355][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.08090435713529587, 'train/grad_norm_pre_clip': 55.568172454833984, 'timing/data': 0.0003176960162818432, 'timing/model': 1.0731711911503226, 'learning_rate/qwen_vl_interface': 5.000111487758778e-07, 'learning_rate/action_model': 5.0001114877587785e-06, 'epoch': 0.29}) +[2026-06-24 11:26:31,430][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.07302585244178772, 'train/grad_norm_pre_clip': 84.65025329589844, 'timing/data': 0.0003253910690546036, 'timing/model': 1.072764387121424, 'learning_rate/qwen_vl_interface': 5.000027871947874e-07, 'learning_rate/action_model': 5.0000278719478735e-06, 'epoch': 0.29}) +[2026-06-24 11:26:32,507][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3000 +[2026-06-24 11:26:57,123][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3000 +[2026-06-24 11:26:57,124][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3000 +[2026-06-24 11:27:28,851][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3000 +[2026-06-24 11:33:54,943][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3000 +[2026-06-24 11:33:54,943][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.033983588218688965, 'train/grad_norm_pre_clip': 22.509431838989258, 'eval/action_loss/seconds': 24.614662195090204, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.0662546381354332, 'eval/latency_2/loss': 0.0662546381354332, 'eval/flappy/loss': 0.0662546381354332, 'eval/flappy/latency_2/loss': 0.0662546381354332, 'eval/flap_precision': 0.951935914552737, 'eval/flap_recall': 0.9674355495251018, 'eval/flap_f1': 0.9596231493943472, 'eval/flap_rate_model': 0.20805555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9596231493943473, 'eval/latency_2/flap_precision': 0.951935914552737, 'eval/latency_2/flap_recall': 0.9674355495251018, 'eval/latency_2/flap_f1': 0.9596231493943472, 'eval/latency_2/flap_rate_model': 0.20805555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9596231493943473, 'eval/action_classification/data_wait_seconds': 0.2646082937717438, 'eval/action_classification/predict_seconds': 31.35713768005371, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 303.72000000001447, 'rl_games_eval/mid_train/mean_length': 2465.6, 'rl_games_eval/mid_train/std_reward': 180.53576266215117, 'rl_games_eval/mid_train/std_length': 1454.0181016754914, 'rl_games_eval/mid_train/macro_mean_reward': 303.72000000001447, 'rl_games_eval/mid_train/macro_mean_length': 2465.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 303.72000000001447, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2465.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 180.53576266215117, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1454.0181016754914, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 303.72000000001447, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003019969444721937, 'timing/model': 1.0761500939261168, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 0.29}) +[2026-06-24 11:33:54,943][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_3000_state +[2026-06-24 11:33:54,946][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 11:33:57,468][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_3000_state/model.safetensors +[2026-06-24 11:34:03,143][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_3000_state/optimizer.bin +[2026-06-24 11:34:03,143][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_3000_state/sampler.bin +[2026-06-24 11:34:03,143][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-24 11:34:03,144][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-24 11:34:03,195][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-24 11:34:03,195][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 11:34:03,207][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 11:34:03,207][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/checkpoints +[2026-06-24 11:34:05,888][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..a8b218ce58ba73e40d7daed9ce97cd431a92d3b6 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/summary.jsonl @@ -0,0 +1,6 @@ +{"steps": 500} +{"steps": 1000} +{"steps": 1500} +{"steps": 2000} +{"steps": 2500} +{"steps": 3000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_4000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_4000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..64fa5f73fd0184d39941e85f0ebedd8c279c2b84 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_4000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:074a513f33ae5b81499478f5b7cd258757d271ce29077b30e299b49f5ba81bdb +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..3e82ec4569883ec8e6960dd582968c7b017b6313 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/config.full.yaml @@ -0,0 +1,224 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 4000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 250 + eval_num_batches: 50 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 1 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: data/flappy_fix_latency_2_200ep + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: none + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..8a5707087236a16355a7a8fe92d3edda53bfac43 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/config.yaml @@ -0,0 +1,103 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + eval_num_workers: 8 + include_state: true + latency_curriculum: + enabled: false + obs_image_size: null + per_device_batch_size: 32 + persistent_workers: true + pin_memory: true + prefetch_factor: 4 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps +rl_games: + env_eval: + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 1 + image_size: 224 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: '' + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: flappy +run_id: flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_action_classification_interval: null + eval_interval: 250 + eval_num_batches: 50 + freeze_llm_layers: [] + freeze_modules: '' + freeze_tied_embedding: false + freeze_vit: false + gradient_accumulation_steps: 1 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 4000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + profile_timing: + enabled: false + reload_modules: null + save_interval: 500 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..aa03e6cbf8e375d660528e17fafddb4804b596b7 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.6028500199317932, + 0.3971499800682068, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4890792667865753, + 0.4890792667865753, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 330734, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..0caa254376d5805f972b0b0411e0b5188e3634e3 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7959861159324646, + 0.2040138840675354, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4030573070049286, + 0.4030573070049286, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..a63a2ab382b74b32e2b15f52706e860a68121fd8 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra/.hydra/config.yaml @@ -0,0 +1,222 @@ +framework: + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 32 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 4000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 250 + eval_num_batches: 50 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: false + freeze_tied_embedding: false + freeze_llm_layers: [] + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 1 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: data/flappy_fix_latency_2_200ep + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: none + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps +output_dir: null +config_yaml: null +is_debug: false +version_id: 0.21 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..75842abb249a04905ba4fc434d657a852f20ad06 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra/.hydra/hydra.yaml @@ -0,0 +1,341 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - ++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++framework.qwenvl.attn_implementation=flash_attention_2 + - ++framework.qwenvl.enable_gradient_checkpointing=true + - ++framework.action_model.state_dim=7 + - ++framework.action_model.loss_type=discrete_ce + - ++framework.action_model.action_horizon=1 + - ++framework.action_model.future_action_window_size=0 + - ++framework.action_model.past_action_window_size=0 + - ++framework.action_model.action_dim=7 + - ++framework.action_model.action_env_dim=2 + - ++framework.name=QwenOFT + - ++datasets.vla_data.dataset_py=lerobot_datasets + - ++datasets.vla_data.include_state=true + - ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games + - ++datasets.vla_data.data_mix=flappy_train + - ++datasets.vla_data.eval_data_mix=null + - ++datasets.vla_data.custom_mixtures_path=null + - ++datasets.vla_data.action_type=discrete + - ++datasets.vla_data.sequential_step_sampling=false + - ++datasets.vla_data.eval_sequential_step_sampling=null + - ++datasets.vla_data.num_workers=8 + - ++datasets.vla_data.eval_num_workers=8 + - ++datasets.vla_data.prefetch_factor=4 + - ++datasets.vla_data.persistent_workers=true + - ++datasets.vla_data.pin_memory=true + - ++datasets.vla_data.shuffle=true + - ++datasets.vla_data.action_balance.enabled=false + - ++datasets.vla_data.action_balance.strategy=balanced_epoch + - ++datasets.vla_data.action_balance.action_key=action_id + - ++datasets.vla_data.action_balance.target_flap_fraction=0.3 + - ++datasets.vla_data.action_balance.noop_id=0 + - ++datasets.vla_data.action_balance.flap_id=1 + - ++datasets.vla_data.latency_curriculum.enabled=false + - ++datasets.vla_data.latency_curriculum.strategy=exclusive + - ++datasets.vla_data.latency_curriculum.latencies=null + - ++datasets.vla_data.latency_curriculum.phase_steps=null + - ++datasets.vla_data.per_device_batch_size=32 + - ++datasets.vla_data.load_all_data_for_training=true + - ++datasets.vla_data.num_obs_frames=1 + - ++datasets.vla_data.image_mode=single + - ++datasets.vla_data.stitch_grid=[2,2] + - ++datasets.vla_data.obs_image_size=null + - ++datasets.vla_data.video_backend=torchvision_av + - ++dataset.source_hf= + - ++dataset.config_name=null + - ++dataset.source_subdir=null + - ++dataset.converted_name=flappy_train + - ++dataset.single_source_hf= + - ++dataset.mixed_source_hf= + - ++dataset.single_converted_name=flappy_train + - ++dataset.mixed_converted_name=flappy_mixed_latency_train + - ++dataset.single_latency_filter=null + - ++dataset.mixed_latency_filter=null + - ++dataset.force_download=false + - ++dataset.setup_force=false + - ++dataset.skip_verification=false + - ++dataset.verify_rows=200 + - ++dataset.max_episodes=null + - ++dataset.episodes_per_latency=null + - ++dataset.latency_filter=null + - ++dataset.debug_subset.enabled=false + - ++dataset.debug_subset.max_episodes=5 + - ++dataset.debug_subset.suffix=debug + - ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct + - ++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + - ++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1 + - ++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt + - ++trainer.max_train_steps=4000 + - ++trainer.num_warmup_steps=100 + - ++trainer.save_interval=500 + - ++trainer.eval_interval=250 + - ++trainer.eval_num_batches=50 + - ++trainer.per_latency_eval_num_batches=null + - ++trainer.eval_action_classification=true + - ++trainer.eval_action_classification_interval=null + - ++trainer.cc_f1_tolerance=1 + - ++trainer.learning_rate.base=2e-05 + - ++trainer.learning_rate.qwen_vl_interface=1e-05 + - ++trainer.learning_rate.action_model=0.0001 + - ++trainer.lr_scheduler_type=cosine_with_min_lr + - ++trainer.scheduler_specific_kwargs.min_lr=1e-06 + - ++trainer.freeze_modules= + - ++trainer.freeze_vit=false + - ++trainer.freeze_tied_embedding=false + - ++trainer.freeze_llm_layers=[] + - ++trainer.loss_scale.vla=1.0 + - ++trainer.loss_scale.vlm=0.1 + - ++trainer.max_grad_norm=1.0 + - ++trainer.weight_decay=0.0 + - ++trainer.logging_frequency=1 + - ++trainer.profile_timing.enabled=false + - ++trainer.profile_timing.log_interval=10 + - ++trainer.gradient_clipping=1.0 + - ++trainer.gradient_accumulation_steps=1 + - ++trainer.distributed_backend=none + - ++trainer.is_resume=false + - ++trainer.pretrained_checkpoint=null + - ++trainer.resume_step=0 + - ++trainer.reload_modules=null + - ++trainer.optimizer.name=AdamW + - ++trainer.optimizer.betas=[0.9,0.95] + - ++trainer.optimizer.eps=1e-08 + - ++trainer.optimizer.weight_decay=1e-08 + - ++trainer.optimizer.fused=true + - ++trainer.save_format=pt + - ++workspace_dir=WORKSPACE_DIR + - ++run_root_dir=results/Checkpoints + - ++seed=42 + - ++wandb_entity=saberrr-zju + - ++wandb_project=starVLA_rl_games + - ++auth.env_file=null + - ++auth.hf_token_env=HF_TOKEN + - ++auth.wandb_api_key_env=WANDB_API_KEY + - ++paths.run_root_dir=results/Checkpoints + - ++paths.dataset_local_dir=data/flappy_fix_latency_2_200ep + - ++paths.dataset_cache_dir=null + - ++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml + - ++rl_games.model_alias=openvla + - ++rl_games.env_eval.image_size=224 + - ++rl_games.env_eval.frameskip=1 + - ++rl_games.env_eval.seed=42 + - ++rl_games.env_eval.fixed_episode_seeds=true + - ++rl_games.env_eval.latency_seed_stride=0 + - ++rl_games.env_eval.task_seed_stride=0 + - ++rl_games.env_eval.task_description= + - ++rl_games.env_eval.enabled=true + - ++rl_games.env_eval.distributed_mode=none + - ++rl_games.env_eval.vectorized.enabled=false + - ++rl_games.env_eval.vectorized.batch_size=1 + - ++rl_games.env_eval.latency.prompt_map_path=null + - ++rl_games.env_eval.latency.mode=single + - ++rl_games.env_eval.latency.values=[0] + - ++rl_games.env_eval.mid_train.enabled=true + - ++rl_games.env_eval.mid_train.interval_steps=250 + - ++rl_games.env_eval.mid_train.latencies=[2] + - ++rl_games.env_eval.mid_train.num_episodes=5 + - ++rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - ++rl_games.env_eval.post_train.enabled=false + - ++rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + - ++rl_games.env_eval.post_train.num_episodes=5 + - ++rl_games.env_eval.post_train.max_steps_per_episode=3600 + - ++rl_games.task=flappy + - ++rl_games.initialization_mode=bridge + - ++rl_games.action_carrier=bridge + - ++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct + - ++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action + - ++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct + - ++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct + - ++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + - ++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++checkpoint.load=none + - ++checkpoint.hf_repo_id=null + - ++checkpoint.save_best_model=false + - ++checkpoint.save_pt_file=false + - ++checkpoint.local.keep_last_n=1 + - ++checkpoint.sync.enabled=false + - ++checkpoint.sync.repo_id=null + - ++checkpoint.sync.keep_last_n=0 + - ++checkpoint.sync.sync_every_n_checkpoints=1 + - ++checkpoint.sync.resume_policy=local_latest + - ++run_id=flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps + - ++output_dir=null + - ++config_yaml=null + - ++is_debug=false + - ++version_id=0.21 + - ++run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints + - ++trainer.is_resume=false + - ++trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - ++trainer.resume_step=0 + - ++datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + - ++datasets.vla_data.data_mix=flappy_train__bridge + - ++datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - ++framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + job: + name: train_starvla_hydra + chdir: false + override_dirname: ++auth.env_file=null,++auth.hf_token_env=HF_TOKEN,++auth.wandb_api_key_env=WANDB_API_KEY,++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action,++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action,++checkpoint.hf_repo_id=null,++checkpoint.load=none,++checkpoint.local.keep_last_n=1,++checkpoint.save_best_model=false,++checkpoint.save_pt_file=false,++checkpoint.sync.enabled=false,++checkpoint.sync.keep_last_n=0,++checkpoint.sync.repo_id=null,++checkpoint.sync.resume_policy=local_latest,++checkpoint.sync.sync_every_n_checkpoints=1,++config_yaml=null,++dataset.config_name=null,++dataset.converted_name=flappy_train,++dataset.debug_subset.enabled=false,++dataset.debug_subset.max_episodes=5,++dataset.debug_subset.suffix=debug,++dataset.episodes_per_latency=null,++dataset.force_download=false,++dataset.latency_filter=null,++dataset.max_episodes=null,++dataset.mixed_converted_name=flappy_mixed_latency_train,++dataset.mixed_latency_filter=null,++dataset.mixed_source_hf=,++dataset.setup_force=false,++dataset.single_converted_name=flappy_train,++dataset.single_latency_filter=null,++dataset.single_source_hf=,++dataset.skip_verification=false,++dataset.source_hf=,++dataset.source_subdir=null,++dataset.verify_rows=200,++datasets.vla_data.action_balance.action_key=action_id,++datasets.vla_data.action_balance.enabled=false,++datasets.vla_data.action_balance.flap_id=1,++datasets.vla_data.action_balance.noop_id=0,++datasets.vla_data.action_balance.strategy=balanced_epoch,++datasets.vla_data.action_balance.target_flap_fraction=0.3,++datasets.vla_data.action_type=discrete,++datasets.vla_data.custom_mixtures_path=null,++datasets.vla_data.data_mix=flappy_train,++datasets.vla_data.data_mix=flappy_train__bridge,++datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep,++datasets.vla_data.data_root_dir=playground/Datasets/rl_games,++datasets.vla_data.dataset_py=lerobot_datasets,++datasets.vla_data.eval_data_mix=flappy_train__bridge__val,++datasets.vla_data.eval_data_mix=null,++datasets.vla_data.eval_num_workers=8,++datasets.vla_data.eval_sequential_step_sampling=null,++datasets.vla_data.image_mode=single,++datasets.vla_data.include_state=true,++datasets.vla_data.latency_curriculum.enabled=false,++datasets.vla_data.latency_curriculum.latencies=null,++datasets.vla_data.latency_curriculum.phase_steps=null,++datasets.vla_data.latency_curriculum.strategy=exclusive,++datasets.vla_data.load_all_data_for_training=true,++datasets.vla_data.num_obs_frames=1,++datasets.vla_data.num_workers=8,++datasets.vla_data.obs_image_size=null,++datasets.vla_data.per_device_batch_size=32,++datasets.vla_data.persistent_workers=true,++datasets.vla_data.pin_memory=true,++datasets.vla_data.prefetch_factor=4,++datasets.vla_data.sequential_step_sampling=false,++datasets.vla_data.shuffle=true,++datasets.vla_data.stitch_grid=[2,2],++datasets.vla_data.video_backend=torchvision_av,++framework.action_model.action_dim=7,++framework.action_model.action_env_dim=2,++framework.action_model.action_horizon=1,++framework.action_model.future_action_window_size=0,++framework.action_model.loss_type=discrete_ce,++framework.action_model.past_action_window_size=0,++framework.action_model.state_dim=7,++framework.name=QwenOFT,++framework.qwenvl.attn_implementation=flash_attention_2,++framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.enable_gradient_checkpointing=true,++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt,++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1,++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1,++is_debug=false,++output_dir=null,++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml,++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++paths.dataset_cache_dir=null,++paths.dataset_local_dir=data/flappy_fix_latency_2_200ep,++paths.run_root_dir=results/Checkpoints,++rl_games.action_carrier=bridge,++rl_games.env_eval.distributed_mode=none,++rl_games.env_eval.enabled=true,++rl_games.env_eval.fixed_episode_seeds=true,++rl_games.env_eval.frameskip=1,++rl_games.env_eval.image_size=224,++rl_games.env_eval.latency.mode=single,++rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json,++rl_games.env_eval.latency.prompt_map_path=null,++rl_games.env_eval.latency.values=[0],++rl_games.env_eval.latency_seed_stride=0,++rl_games.env_eval.mid_train.enabled=true,++rl_games.env_eval.mid_train.interval_steps=250,++rl_games.env_eval.mid_train.latencies=[2],++rl_games.env_eval.mid_train.max_steps_per_episode=3600,++rl_games.env_eval.mid_train.num_episodes=5,++rl_games.env_eval.post_train.enabled=false,++rl_games.env_eval.post_train.latencies=[0,1,2,3,4],++rl_games.env_eval.post_train.max_steps_per_episode=3600,++rl_games.env_eval.post_train.num_episodes=5,++rl_games.env_eval.seed=42,++rl_games.env_eval.task_description=,++rl_games.env_eval.task_seed_stride=0,++rl_games.env_eval.vectorized.batch_size=1,++rl_games.env_eval.vectorized.enabled=false,++rl_games.initialization_mode=bridge,++rl_games.model_alias=openvla,++rl_games.task=flappy,++run_id=flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps,++run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints,++run_root_dir=results/Checkpoints,++seed=42,++trainer.cc_f1_tolerance=1,++trainer.distributed_backend=none,++trainer.eval_action_classification=true,++trainer.eval_action_classification_interval=null,++trainer.eval_interval=250,++trainer.eval_num_batches=50,++trainer.freeze_llm_layers=[],++trainer.freeze_modules=,++trainer.freeze_tied_embedding=false,++trainer.freeze_vit=false,++trainer.gradient_accumulation_steps=1,++trainer.gradient_clipping=1.0,++trainer.is_resume=false,++trainer.is_resume=false,++trainer.learning_rate.action_model=0.0001,++trainer.learning_rate.base=2e-05,++trainer.learning_rate.qwen_vl_interface=1e-05,++trainer.logging_frequency=1,++trainer.loss_scale.vla=1.0,++trainer.loss_scale.vlm=0.1,++trainer.lr_scheduler_type=cosine_with_min_lr,++trainer.max_grad_norm=1.0,++trainer.max_train_steps=4000,++trainer.num_warmup_steps=100,++trainer.optimizer.betas=[0.9,0.95],++trainer.optimizer.eps=1e-08,++trainer.optimizer.fused=true,++trainer.optimizer.name=AdamW,++trainer.optimizer.weight_decay=1e-08,++trainer.per_latency_eval_num_batches=null,++trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,++trainer.pretrained_checkpoint=null,++trainer.profile_timing.enabled=false,++trainer.profile_timing.log_interval=10,++trainer.reload_modules=null,++trainer.resume_step=0,++trainer.resume_step=0,++trainer.save_format=pt,++trainer.save_interval=500,++trainer.scheduler_specific_kwargs.min_lr=1e-06,++trainer.weight_decay=0.0,++version_id=0.21,++wandb_entity=saberrr-zju,++wandb_project=starVLA_rl_games,++workspace_dir=WORKSPACE_DIR,env=flappy,init=bridge,mode=single,model=openvla + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/latency-sensitive-bench/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/latency-sensitive-bench/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..6c9b307325daa64a6a4f879c5f53d4db78663752 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra/.hydra/overrides.yaml @@ -0,0 +1,181 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- ++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++framework.qwenvl.attn_implementation=flash_attention_2 +- ++framework.qwenvl.enable_gradient_checkpointing=true +- ++framework.action_model.state_dim=7 +- ++framework.action_model.loss_type=discrete_ce +- ++framework.action_model.action_horizon=1 +- ++framework.action_model.future_action_window_size=0 +- ++framework.action_model.past_action_window_size=0 +- ++framework.action_model.action_dim=7 +- ++framework.action_model.action_env_dim=2 +- ++framework.name=QwenOFT +- ++datasets.vla_data.dataset_py=lerobot_datasets +- ++datasets.vla_data.include_state=true +- ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games +- ++datasets.vla_data.data_mix=flappy_train +- ++datasets.vla_data.eval_data_mix=null +- ++datasets.vla_data.custom_mixtures_path=null +- ++datasets.vla_data.action_type=discrete +- ++datasets.vla_data.sequential_step_sampling=false +- ++datasets.vla_data.eval_sequential_step_sampling=null +- ++datasets.vla_data.num_workers=8 +- ++datasets.vla_data.eval_num_workers=8 +- ++datasets.vla_data.prefetch_factor=4 +- ++datasets.vla_data.persistent_workers=true +- ++datasets.vla_data.pin_memory=true +- ++datasets.vla_data.shuffle=true +- ++datasets.vla_data.action_balance.enabled=false +- ++datasets.vla_data.action_balance.strategy=balanced_epoch +- ++datasets.vla_data.action_balance.action_key=action_id +- ++datasets.vla_data.action_balance.target_flap_fraction=0.3 +- ++datasets.vla_data.action_balance.noop_id=0 +- ++datasets.vla_data.action_balance.flap_id=1 +- ++datasets.vla_data.latency_curriculum.enabled=false +- ++datasets.vla_data.latency_curriculum.strategy=exclusive +- ++datasets.vla_data.latency_curriculum.latencies=null +- ++datasets.vla_data.latency_curriculum.phase_steps=null +- ++datasets.vla_data.per_device_batch_size=32 +- ++datasets.vla_data.load_all_data_for_training=true +- ++datasets.vla_data.num_obs_frames=1 +- ++datasets.vla_data.image_mode=single +- ++datasets.vla_data.stitch_grid=[2,2] +- ++datasets.vla_data.obs_image_size=null +- ++datasets.vla_data.video_backend=torchvision_av +- ++dataset.source_hf= +- ++dataset.config_name=null +- ++dataset.source_subdir=null +- ++dataset.converted_name=flappy_train +- ++dataset.single_source_hf= +- ++dataset.mixed_source_hf= +- ++dataset.single_converted_name=flappy_train +- ++dataset.mixed_converted_name=flappy_mixed_latency_train +- ++dataset.single_latency_filter=null +- ++dataset.mixed_latency_filter=null +- ++dataset.force_download=false +- ++dataset.setup_force=false +- ++dataset.skip_verification=false +- ++dataset.verify_rows=200 +- ++dataset.max_episodes=null +- ++dataset.episodes_per_latency=null +- ++dataset.latency_filter=null +- ++dataset.debug_subset.enabled=false +- ++dataset.debug_subset.max_episodes=5 +- ++dataset.debug_subset.suffix=debug +- ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct +- ++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 +- ++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1 +- ++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt +- ++trainer.max_train_steps=4000 +- ++trainer.num_warmup_steps=100 +- ++trainer.save_interval=500 +- ++trainer.eval_interval=250 +- ++trainer.eval_num_batches=50 +- ++trainer.per_latency_eval_num_batches=null +- ++trainer.eval_action_classification=true +- ++trainer.eval_action_classification_interval=null +- ++trainer.cc_f1_tolerance=1 +- ++trainer.learning_rate.base=2e-05 +- ++trainer.learning_rate.qwen_vl_interface=1e-05 +- ++trainer.learning_rate.action_model=0.0001 +- ++trainer.lr_scheduler_type=cosine_with_min_lr +- ++trainer.scheduler_specific_kwargs.min_lr=1e-06 +- ++trainer.freeze_modules= +- ++trainer.freeze_vit=false +- ++trainer.freeze_tied_embedding=false +- ++trainer.freeze_llm_layers=[] +- ++trainer.loss_scale.vla=1.0 +- ++trainer.loss_scale.vlm=0.1 +- ++trainer.max_grad_norm=1.0 +- ++trainer.weight_decay=0.0 +- ++trainer.logging_frequency=1 +- ++trainer.profile_timing.enabled=false +- ++trainer.profile_timing.log_interval=10 +- ++trainer.gradient_clipping=1.0 +- ++trainer.gradient_accumulation_steps=1 +- ++trainer.distributed_backend=none +- ++trainer.is_resume=false +- ++trainer.pretrained_checkpoint=null +- ++trainer.resume_step=0 +- ++trainer.reload_modules=null +- ++trainer.optimizer.name=AdamW +- ++trainer.optimizer.betas=[0.9,0.95] +- ++trainer.optimizer.eps=1e-08 +- ++trainer.optimizer.weight_decay=1e-08 +- ++trainer.optimizer.fused=true +- ++trainer.save_format=pt +- ++workspace_dir=WORKSPACE_DIR +- ++run_root_dir=results/Checkpoints +- ++seed=42 +- ++wandb_entity=saberrr-zju +- ++wandb_project=starVLA_rl_games +- ++auth.env_file=null +- ++auth.hf_token_env=HF_TOKEN +- ++auth.wandb_api_key_env=WANDB_API_KEY +- ++paths.run_root_dir=results/Checkpoints +- ++paths.dataset_local_dir=data/flappy_fix_latency_2_200ep +- ++paths.dataset_cache_dir=null +- ++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml +- ++rl_games.model_alias=openvla +- ++rl_games.env_eval.image_size=224 +- ++rl_games.env_eval.frameskip=1 +- ++rl_games.env_eval.seed=42 +- ++rl_games.env_eval.fixed_episode_seeds=true +- ++rl_games.env_eval.latency_seed_stride=0 +- ++rl_games.env_eval.task_seed_stride=0 +- ++rl_games.env_eval.task_description= +- ++rl_games.env_eval.enabled=true +- ++rl_games.env_eval.distributed_mode=none +- ++rl_games.env_eval.vectorized.enabled=false +- ++rl_games.env_eval.vectorized.batch_size=1 +- ++rl_games.env_eval.latency.prompt_map_path=null +- ++rl_games.env_eval.latency.mode=single +- ++rl_games.env_eval.latency.values=[0] +- ++rl_games.env_eval.mid_train.enabled=true +- ++rl_games.env_eval.mid_train.interval_steps=250 +- ++rl_games.env_eval.mid_train.latencies=[2] +- ++rl_games.env_eval.mid_train.num_episodes=5 +- ++rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- ++rl_games.env_eval.post_train.enabled=false +- ++rl_games.env_eval.post_train.latencies=[0,1,2,3,4] +- ++rl_games.env_eval.post_train.num_episodes=5 +- ++rl_games.env_eval.post_train.max_steps_per_episode=3600 +- ++rl_games.task=flappy +- ++rl_games.initialization_mode=bridge +- ++rl_games.action_carrier=bridge +- ++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct +- ++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action +- ++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct +- ++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct +- ++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action +- ++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++checkpoint.load=none +- ++checkpoint.hf_repo_id=null +- ++checkpoint.save_best_model=false +- ++checkpoint.save_pt_file=false +- ++checkpoint.local.keep_last_n=1 +- ++checkpoint.sync.enabled=false +- ++checkpoint.sync.repo_id=null +- ++checkpoint.sync.keep_last_n=0 +- ++checkpoint.sync.sync_every_n_checkpoints=1 +- ++checkpoint.sync.resume_policy=local_latest +- ++run_id=flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps +- ++output_dir=null +- ++config_yaml=null +- ++is_debug=false +- ++version_id=0.21 +- ++run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints +- ++trainer.is_resume=false +- ++trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- ++trainer.resume_step=0 +- ++datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep +- ++datasets.vla_data.data_mix=flappy_train__bridge +- ++datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- ++framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..034967590f38a644b8cda6d2b2fa1e6616b8ba97 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/hydra/train_starvla_hydra.log @@ -0,0 +1,4216 @@ +[2026-06-24 11:59:00,668][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21' +[2026-06-24 11:59:00,669][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-24 11:59:00,916][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured. +[2026-06-24 11:59:00,916][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub +[2026-06-24 11:59:01,466][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases +[2026-06-24 11:59:01,473][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-24 11:59:04,765][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-24 11:59:04,789][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-24 11:59:05,196][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-24 11:59:05,196][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-24 11:59:05,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-24 11:59:05,198][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-24 11:59:05,198][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-24 11:59:05,198][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-24 11:59:05,198][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-24 11:59:05,198][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-24 11:59:05,198][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-24 11:59:05,217][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-24 11:59:05,244][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/config.full.yaml +[2026-06-24 11:59:05,248][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/config.yaml +[2026-06-24 11:59:05,249][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-24 11:59:07,717][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-24 11:59:07,717][starVLA.training.train_starvla][INFO] - Total optimization steps = 4000 +[2026-06-24 11:59:07,717][starVLA.training.train_starvla][INFO] - Per device batch size = 32 +[2026-06-24 11:59:07,717][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 1 +[2026-06-24 11:59:07,717][starVLA.training.train_starvla][INFO] - Total batch size = 32 +[2026-06-24 11:59:30,003][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6939594745635986, 'train/grad_norm_pre_clip': 13.036202430725098, 'timing/data': 20.402398857986555, 'timing/model': 1.880269082961604, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-24 11:59:30,973][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6941108703613281, 'train/grad_norm_pre_clip': 0.8904328942298889, 'timing/data': 0.0006950818933546543, 'timing/model': 0.9663203458767384, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-24 11:59:31,915][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.693867027759552, 'train/grad_norm_pre_clip': 0.8979056477546692, 'timing/data': 0.0009125990327447653, 'timing/model': 0.9376492430455983, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-24 11:59:32,896][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.6873155236244202, 'train/grad_norm_pre_clip': 21.7597599029541, 'timing/data': 0.0007458371110260487, 'timing/model': 0.9769839351065457, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-24 11:59:33,874][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6791630983352661, 'train/grad_norm_pre_clip': 17.332332611083984, 'timing/data': 0.0008690380491316319, 'timing/model': 0.9738281350582838, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-24 11:59:34,844][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6820808053016663, 'train/grad_norm_pre_clip': 5.818544864654541, 'timing/data': 0.0007409281097352505, 'timing/model': 0.9665715210139751, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-24 11:59:35,790][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.6692202687263489, 'train/grad_norm_pre_clip': 6.416031360626221, 'timing/data': 0.0008436001371592283, 'timing/model': 0.9430014560930431, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-24 11:59:36,721][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.7121397256851196, 'train/grad_norm_pre_clip': 9.84980583190918, 'timing/data': 0.000822931993752718, 'timing/model': 0.9289265680126846, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-24 11:59:37,596][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.6455320119857788, 'train/grad_norm_pre_clip': 4.174345016479492, 'timing/data': 0.00034860288724303246, 'timing/model': 0.8733741948381066, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-24 11:59:38,471][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.5686480402946472, 'train/grad_norm_pre_clip': 14.105907440185547, 'timing/data': 0.00030127400532364845, 'timing/model': 0.8730485360138118, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-24 11:59:39,347][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.5732211470603943, 'train/grad_norm_pre_clip': 7.415318489074707, 'timing/data': 0.0002787620760500431, 'timing/model': 0.8743903399445117, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-24 11:59:40,221][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.7101293206214905, 'train/grad_norm_pre_clip': 12.758611679077148, 'timing/data': 0.00031369784846901894, 'timing/model': 0.8724460990633816, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-24 11:59:41,095][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.6268185377120972, 'train/grad_norm_pre_clip': 5.880255222320557, 'timing/data': 0.0002783921081572771, 'timing/model': 0.872788371052593, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.0}) +[2026-06-24 11:59:41,968][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.7040006518363953, 'train/grad_norm_pre_clip': 13.685173034667969, 'timing/data': 0.00029847887344658375, 'timing/model': 0.8706340258941054, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.0}) +[2026-06-24 11:59:42,843][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.7052491307258606, 'train/grad_norm_pre_clip': 14.262772560119629, 'timing/data': 0.00030384911224246025, 'timing/model': 0.8731875808443874, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.0}) +[2026-06-24 11:59:43,719][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.8065332770347595, 'train/grad_norm_pre_clip': 23.368288040161133, 'timing/data': 0.00029717711731791496, 'timing/model': 0.8750816208776087, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.0}) +[2026-06-24 11:59:44,596][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.7586247324943542, 'train/grad_norm_pre_clip': 18.664255142211914, 'timing/data': 0.0003206611145287752, 'timing/model': 0.8747449310030788, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.0}) +[2026-06-24 11:59:45,472][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.7088506817817688, 'train/grad_norm_pre_clip': 12.91820240020752, 'timing/data': 0.00031888787634670734, 'timing/model': 0.8750977730378509, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.0}) +[2026-06-24 11:59:46,348][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.7385762333869934, 'train/grad_norm_pre_clip': 16.846920013427734, 'timing/data': 0.0003265819977968931, 'timing/model': 0.8743014622014016, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.0}) +[2026-06-24 11:59:47,226][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.6896636486053467, 'train/grad_norm_pre_clip': 7.381400108337402, 'timing/data': 0.0004776150453835726, 'timing/model': 0.8760792738758028, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.0}) +[2026-06-24 11:59:48,104][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.6818586587905884, 'train/grad_norm_pre_clip': 5.505635738372803, 'timing/data': 0.00029207789339125156, 'timing/model': 0.8762458369601518, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.0}) +[2026-06-24 11:59:48,985][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.6882389187812805, 'train/grad_norm_pre_clip': 5.862356662750244, 'timing/data': 0.0004249860066920519, 'timing/model': 0.8787734429351985, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.0}) +[2026-06-24 11:59:49,865][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.7134860754013062, 'train/grad_norm_pre_clip': 20.158994674682617, 'timing/data': 0.00027293083257973194, 'timing/model': 0.8779341308400035, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.0}) +[2026-06-24 11:59:50,745][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.749304473400116, 'train/grad_norm_pre_clip': 29.18122673034668, 'timing/data': 0.0002943410072475672, 'timing/model': 0.8786521980073303, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.0}) +[2026-06-24 11:59:51,625][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.7255931496620178, 'train/grad_norm_pre_clip': 16.86631202697754, 'timing/data': 0.0003190168645232916, 'timing/model': 0.8785379850305617, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.0}) +[2026-06-24 11:59:52,505][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.7364541888237, 'train/grad_norm_pre_clip': 28.03182029724121, 'timing/data': 0.0003014439716935158, 'timing/model': 0.8778014059644192, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.0}) +[2026-06-24 11:59:53,386][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.6758642792701721, 'train/grad_norm_pre_clip': 10.918733596801758, 'timing/data': 0.00030625308863818645, 'timing/model': 0.8799607420805842, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.0}) +[2026-06-24 11:59:54,270][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.6914467215538025, 'train/grad_norm_pre_clip': 13.983187675476074, 'timing/data': 0.00028127687983214855, 'timing/model': 0.8825412280857563, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.0}) +[2026-06-24 11:59:55,153][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.6711454391479492, 'train/grad_norm_pre_clip': 10.923629760742188, 'timing/data': 0.00028772884979844093, 'timing/model': 0.8811183839570731, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.0}) +[2026-06-24 11:59:56,036][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.6757531762123108, 'train/grad_norm_pre_clip': 1.8972469568252563, 'timing/data': 0.00030247587710618973, 'timing/model': 0.8810144290328026, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.0}) +[2026-06-24 11:59:56,924][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.6642181277275085, 'train/grad_norm_pre_clip': 2.125958204269409, 'timing/data': 0.0005273479036986828, 'timing/model': 0.8859819669742137, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.0}) +[2026-06-24 11:59:57,806][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.6997585892677307, 'train/grad_norm_pre_clip': 9.794967651367188, 'timing/data': 0.00030320812948048115, 'timing/model': 0.8801888250745833, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.0}) +[2026-06-24 11:59:58,708][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.7122003436088562, 'train/grad_norm_pre_clip': 16.474075317382812, 'timing/data': 0.00029273913241922855, 'timing/model': 0.9006369039416313, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.0}) +[2026-06-24 11:59:59,591][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.7218528389930725, 'train/grad_norm_pre_clip': 17.865602493286133, 'timing/data': 0.0002858659718185663, 'timing/model': 0.8813550919294357, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.0}) +[2026-06-24 12:00:00,473][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.6647288203239441, 'train/grad_norm_pre_clip': 3.5779225826263428, 'timing/data': 0.0002835518680512905, 'timing/model': 0.8800487651024014, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.0}) +[2026-06-24 12:00:01,357][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.6637051701545715, 'train/grad_norm_pre_clip': 19.528432846069336, 'timing/data': 0.000281978165730834, 'timing/model': 0.8821004400961101, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.0}) +[2026-06-24 12:00:02,246][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.6490524411201477, 'train/grad_norm_pre_clip': 9.873353958129883, 'timing/data': 0.00027994485571980476, 'timing/model': 0.8872526129707694, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.0}) +[2026-06-24 12:00:03,128][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.6694239377975464, 'train/grad_norm_pre_clip': 17.611974716186523, 'timing/data': 0.00028271880000829697, 'timing/model': 0.8808478040155023, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.0}) +[2026-06-24 12:00:04,011][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.6612569093704224, 'train/grad_norm_pre_clip': 7.195664405822754, 'timing/data': 0.00046713510528206825, 'timing/model': 0.8811149958055466, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.0}) +[2026-06-24 12:00:04,892][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.6163384318351746, 'train/grad_norm_pre_clip': 14.354401588439941, 'timing/data': 0.0003603848163038492, 'timing/model': 0.8790700968820602, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.0}) +[2026-06-24 12:00:05,774][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.6410778760910034, 'train/grad_norm_pre_clip': 9.657115936279297, 'timing/data': 0.00028849090449512005, 'timing/model': 0.8802721200045198, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.0}) +[2026-06-24 12:00:06,654][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.612479031085968, 'train/grad_norm_pre_clip': 18.681501388549805, 'timing/data': 0.00028786901384592056, 'timing/model': 0.8787893429398537, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.0}) +[2026-06-24 12:00:07,538][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.6057854294776917, 'train/grad_norm_pre_clip': 12.989033699035645, 'timing/data': 0.0003020560834556818, 'timing/model': 0.8818281770218164, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.0}) +[2026-06-24 12:00:08,423][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.5729181170463562, 'train/grad_norm_pre_clip': 17.611106872558594, 'timing/data': 0.00031618191860616207, 'timing/model': 0.8838066041935235, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.0}) +[2026-06-24 12:00:09,308][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.5465272665023804, 'train/grad_norm_pre_clip': 12.4789400100708, 'timing/data': 0.00030431104823946953, 'timing/model': 0.8834056840278208, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.0}) +[2026-06-24 12:00:10,188][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.45360755920410156, 'train/grad_norm_pre_clip': 21.945390701293945, 'timing/data': 0.00030500185675919056, 'timing/model': 0.8777457410469651, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.0}) +[2026-06-24 12:00:11,072][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.43228042125701904, 'train/grad_norm_pre_clip': 18.239402770996094, 'timing/data': 0.00030710501596331596, 'timing/model': 0.8824745509773493, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.0}) +[2026-06-24 12:00:11,956][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.565536379814148, 'train/grad_norm_pre_clip': 59.517887115478516, 'timing/data': 0.00029850914143025875, 'timing/model': 0.8825084560085088, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.0}) +[2026-06-24 12:00:12,839][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.48318928480148315, 'train/grad_norm_pre_clip': 36.66328048706055, 'timing/data': 0.0003228648565709591, 'timing/model': 0.8815385100897402, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.0}) +[2026-06-24 12:00:13,854][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.4818522334098816, 'train/grad_norm_pre_clip': 53.938682556152344, 'timing/data': 0.00032878597266972065, 'timing/model': 1.0127816528547555, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.0}) +[2026-06-24 12:00:14,737][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.3492366671562195, 'train/grad_norm_pre_clip': 35.60288619995117, 'timing/data': 0.0002965859603136778, 'timing/model': 0.8817177780438215, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.0}) +[2026-06-24 12:00:15,621][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.6207365989685059, 'train/grad_norm_pre_clip': 44.42538833618164, 'timing/data': 0.0002496871165931225, 'timing/model': 0.8825385761447251, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.01}) +[2026-06-24 12:00:16,506][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.5053760409355164, 'train/grad_norm_pre_clip': 23.976858139038086, 'timing/data': 0.0003144091460853815, 'timing/model': 0.8833881891332567, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:17,389][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.5123131275177002, 'train/grad_norm_pre_clip': 19.95326042175293, 'timing/data': 0.0003005031030625105, 'timing/model': 0.8807106239255518, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.01}) +[2026-06-24 12:00:18,273][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.5790843367576599, 'train/grad_norm_pre_clip': 49.92840576171875, 'timing/data': 0.0003024269826710224, 'timing/model': 0.8829627239611, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:19,159][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.42594972252845764, 'train/grad_norm_pre_clip': 17.86495018005371, 'timing/data': 0.00031087291426956654, 'timing/model': 0.8835409679450095, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.01}) +[2026-06-24 12:00:20,044][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.5355709195137024, 'train/grad_norm_pre_clip': 67.17859649658203, 'timing/data': 0.0003571389243006706, 'timing/model': 0.8837689659558237, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.01}) +[2026-06-24 12:00:20,930][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.6200018525123596, 'train/grad_norm_pre_clip': 69.8698501586914, 'timing/data': 0.00029812916181981564, 'timing/model': 0.8837657910771668, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.01}) +[2026-06-24 12:00:21,815][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.511902391910553, 'train/grad_norm_pre_clip': 32.396759033203125, 'timing/data': 0.00029192701913416386, 'timing/model': 0.8833934250287712, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.01}) +[2026-06-24 12:00:22,700][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.5662831664085388, 'train/grad_norm_pre_clip': 47.4777946472168, 'timing/data': 0.0003282749094069004, 'timing/model': 0.8834519640076905, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.01}) +[2026-06-24 12:00:23,589][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.6020128726959229, 'train/grad_norm_pre_clip': 44.28144073486328, 'timing/data': 0.00029641599394381046, 'timing/model': 0.8874921849928796, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.01}) +[2026-06-24 12:00:24,474][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.5257717370986938, 'train/grad_norm_pre_clip': 39.146976470947266, 'timing/data': 0.00030223606154322624, 'timing/model': 0.8826700118370354, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.01}) +[2026-06-24 12:00:25,361][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.5487815141677856, 'train/grad_norm_pre_clip': 23.302978515625, 'timing/data': 0.00029027415439486504, 'timing/model': 0.8859550971537828, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.01}) +[2026-06-24 12:00:26,248][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.4987589716911316, 'train/grad_norm_pre_clip': 30.491741180419922, 'timing/data': 0.0004443931393325329, 'timing/model': 0.8839931830298156, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:27,135][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.6547001600265503, 'train/grad_norm_pre_clip': 51.00725173950195, 'timing/data': 0.00031170411966741085, 'timing/model': 0.8861055208835751, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:28,022][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.6077253222465515, 'train/grad_norm_pre_clip': 57.85679626464844, 'timing/data': 0.00030147493816912174, 'timing/model': 0.8849261428695172, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.01}) +[2026-06-24 12:00:28,906][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.3919277489185333, 'train/grad_norm_pre_clip': 41.89815139770508, 'timing/data': 0.0002846128772944212, 'timing/model': 0.8818658899981529, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.01}) +[2026-06-24 12:00:29,792][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.5037282705307007, 'train/grad_norm_pre_clip': 52.848777770996094, 'timing/data': 0.00028407201170921326, 'timing/model': 0.8843206600286067, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:30,682][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.6962780952453613, 'train/grad_norm_pre_clip': 61.287506103515625, 'timing/data': 0.00030295783653855324, 'timing/model': 0.889010296901688, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.01}) +[2026-06-24 12:00:31,567][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.4861232042312622, 'train/grad_norm_pre_clip': 52.32508087158203, 'timing/data': 0.00031816703267395496, 'timing/model': 0.883229237049818, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.01}) +[2026-06-24 12:00:32,455][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.3644658923149109, 'train/grad_norm_pre_clip': 48.10275650024414, 'timing/data': 0.00028710789047181606, 'timing/model': 0.8865042740944773, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.01}) +[2026-06-24 12:00:33,342][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.47918951511383057, 'train/grad_norm_pre_clip': 54.15022659301758, 'timing/data': 0.00032642087899148464, 'timing/model': 0.8851769480388612, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.01}) +[2026-06-24 12:00:34,230][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.3930624723434448, 'train/grad_norm_pre_clip': 68.02978515625, 'timing/data': 0.0002913661301136017, 'timing/model': 0.885553183965385, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.01}) +[2026-06-24 12:00:35,118][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.3172841966152191, 'train/grad_norm_pre_clip': 22.78586769104004, 'timing/data': 0.00029453099705278873, 'timing/model': 0.8862837329506874, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.01}) +[2026-06-24 12:00:36,007][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.3671870231628418, 'train/grad_norm_pre_clip': 37.79704284667969, 'timing/data': 0.0003041790332645178, 'timing/model': 0.8873662420082837, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:36,894][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.5712491869926453, 'train/grad_norm_pre_clip': 111.68146514892578, 'timing/data': 0.0002953039947897196, 'timing/model': 0.8859312341082841, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.01}) +[2026-06-24 12:00:37,784][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.6585264801979065, 'train/grad_norm_pre_clip': 86.9084701538086, 'timing/data': 0.00028748903423547745, 'timing/model': 0.8884729268029332, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.01}) +[2026-06-24 12:00:38,672][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.5362561345100403, 'train/grad_norm_pre_clip': 38.07193374633789, 'timing/data': 0.000329446978867054, 'timing/model': 0.8858373709954321, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:39,560][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.3653225600719452, 'train/grad_norm_pre_clip': 19.351430892944336, 'timing/data': 0.00029570399783551693, 'timing/model': 0.88645467790775, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:40,450][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.4803769290447235, 'train/grad_norm_pre_clip': 43.515350341796875, 'timing/data': 0.0003076661378145218, 'timing/model': 0.8885259900707752, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.01}) +[2026-06-24 12:00:41,338][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.5749239325523376, 'train/grad_norm_pre_clip': 39.07463455200195, 'timing/data': 0.00030355900526046753, 'timing/model': 0.8859636199194938, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.01}) +[2026-06-24 12:00:42,226][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.565599799156189, 'train/grad_norm_pre_clip': 42.076107025146484, 'timing/data': 0.00029720691964030266, 'timing/model': 0.887063100002706, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.01}) +[2026-06-24 12:00:43,116][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.5472124814987183, 'train/grad_norm_pre_clip': 59.31515884399414, 'timing/data': 0.00029678596183657646, 'timing/model': 0.8874012348242104, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.01}) +[2026-06-24 12:00:44,002][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.5132270455360413, 'train/grad_norm_pre_clip': 14.180854797363281, 'timing/data': 0.00031639309599995613, 'timing/model': 0.8846878930926323, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.01}) +[2026-06-24 12:00:44,890][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.6424981951713562, 'train/grad_norm_pre_clip': 89.72752380371094, 'timing/data': 0.00035501504316926, 'timing/model': 0.8861236129887402, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.01}) +[2026-06-24 12:00:45,781][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.7647888660430908, 'train/grad_norm_pre_clip': 80.35328674316406, 'timing/data': 0.0003240569494664669, 'timing/model': 0.8895189661998302, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.01}) +[2026-06-24 12:00:46,669][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.5386471152305603, 'train/grad_norm_pre_clip': 46.05589294433594, 'timing/data': 0.0003132468555122614, 'timing/model': 0.8859439790248871, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.01}) +[2026-06-24 12:00:47,557][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.5668224096298218, 'train/grad_norm_pre_clip': 12.563312530517578, 'timing/data': 0.000501600094139576, 'timing/model': 0.8866029039490968, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:48,447][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.5200134515762329, 'train/grad_norm_pre_clip': 44.694583892822266, 'timing/data': 0.00029309908859431744, 'timing/model': 0.8874423978850245, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:49,334][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.6063616871833801, 'train/grad_norm_pre_clip': 45.50492858886719, 'timing/data': 0.0003414400853216648, 'timing/model': 0.8861496259924024, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.01}) +[2026-06-24 12:00:50,225][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.47503507137298584, 'train/grad_norm_pre_clip': 29.774717330932617, 'timing/data': 0.0003719169180840254, 'timing/model': 0.8884407400619239, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.01}) +[2026-06-24 12:00:51,113][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.382965624332428, 'train/grad_norm_pre_clip': 17.2403507232666, 'timing/data': 0.00030139414593577385, 'timing/model': 0.8865419928915799, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:52,005][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.5833406448364258, 'train/grad_norm_pre_clip': 120.90963745117188, 'timing/data': 0.00031544105149805546, 'timing/model': 0.8904444559011608, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:52,897][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.5172449350357056, 'train/grad_norm_pre_clip': 108.27613067626953, 'timing/data': 0.00029445206746459007, 'timing/model': 0.8902075330261141, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.01}) +[2026-06-24 12:00:53,786][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 1.2695976495742798, 'train/grad_norm_pre_clip': 188.70977783203125, 'timing/data': 0.00031544105149805546, 'timing/model': 0.8871136570814997, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.01}) +[2026-06-24 12:00:54,675][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.5236830115318298, 'train/grad_norm_pre_clip': 176.9324951171875, 'timing/data': 0.0003127758391201496, 'timing/model': 0.8870268249884248, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.01}) +[2026-06-24 12:00:55,568][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.4717029631137848, 'train/grad_norm_pre_clip': 56.20669174194336, 'timing/data': 0.00030207610689103603, 'timing/model': 0.8917582100257277, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.01}) +[2026-06-24 12:00:56,457][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.5298907160758972, 'train/grad_norm_pre_clip': 32.11457824707031, 'timing/data': 0.00029677688144147396, 'timing/model': 0.8874532161280513, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.01}) +[2026-06-24 12:00:57,347][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.5260981917381287, 'train/grad_norm_pre_clip': 35.325469970703125, 'timing/data': 0.00048182299360632896, 'timing/model': 0.8877117009833455, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.01}) +[2026-06-24 12:00:58,236][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.44705191254615784, 'train/grad_norm_pre_clip': 27.057861328125, 'timing/data': 0.0003235260955989361, 'timing/model': 0.8867409431841224, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.01}) +[2026-06-24 12:00:59,127][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.4968222677707672, 'train/grad_norm_pre_clip': 13.979130744934082, 'timing/data': 0.00029577407985925674, 'timing/model': 0.8892571181058884, 'learning_rate/qwen_vl_interface': 9.999998458888285e-06, 'learning_rate/action_model': 9.999998458888285e-05, 'epoch': 0.01}) +[2026-06-24 12:01:00,015][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.610012948513031, 'train/grad_norm_pre_clip': 37.05908966064453, 'timing/data': 0.0003137469757348299, 'timing/model': 0.8866202079225332, 'learning_rate/qwen_vl_interface': 9.999993835554141e-06, 'learning_rate/action_model': 9.999993835554141e-05, 'epoch': 0.01}) +[2026-06-24 12:01:00,905][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.4952313303947449, 'train/grad_norm_pre_clip': 14.181517601013184, 'timing/data': 0.00035205017775297165, 'timing/model': 0.888574109878391, 'learning_rate/qwen_vl_interface': 9.999986130000566e-06, 'learning_rate/action_model': 9.999986130000566e-05, 'epoch': 0.01}) +[2026-06-24 12:01:01,795][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.47682660818099976, 'train/grad_norm_pre_clip': 49.973175048828125, 'timing/data': 0.00032601109705865383, 'timing/model': 0.8876561019569635, 'learning_rate/qwen_vl_interface': 9.99997534223256e-06, 'learning_rate/action_model': 9.99997534223256e-05, 'epoch': 0.01}) +[2026-06-24 12:01:02,685][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.5385527610778809, 'train/grad_norm_pre_clip': 33.489234924316406, 'timing/data': 0.0002990700304508209, 'timing/model': 0.8884049849584699, 'learning_rate/qwen_vl_interface': 9.999961472257127e-06, 'learning_rate/action_model': 9.999961472257127e-05, 'epoch': 0.01}) +[2026-06-24 12:01:03,574][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.4227660000324249, 'train/grad_norm_pre_clip': 18.219236373901367, 'timing/data': 0.00030095409601926804, 'timing/model': 0.8875214911531657, 'learning_rate/qwen_vl_interface': 9.999944520083262e-06, 'learning_rate/action_model': 9.999944520083261e-05, 'epoch': 0.01}) +[2026-06-24 12:01:04,464][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.48458272218704224, 'train/grad_norm_pre_clip': 15.68236255645752, 'timing/data': 0.0003092989791184664, 'timing/model': 0.8885072891134769, 'learning_rate/qwen_vl_interface': 9.999924485721968e-06, 'learning_rate/action_model': 9.999924485721968e-05, 'epoch': 0.01}) +[2026-06-24 12:01:05,355][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.416036993265152, 'train/grad_norm_pre_clip': 37.32638931274414, 'timing/data': 0.0002927579917013645, 'timing/model': 0.8893146619666368, 'learning_rate/qwen_vl_interface': 9.999901369186244e-06, 'learning_rate/action_model': 9.999901369186244e-05, 'epoch': 0.01}) +[2026-06-24 12:01:06,244][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.3694683611392975, 'train/grad_norm_pre_clip': 37.51213836669922, 'timing/data': 0.00032194284722208977, 'timing/model': 0.8867305121384561, 'learning_rate/qwen_vl_interface': 9.999875170491092e-06, 'learning_rate/action_model': 9.999875170491091e-05, 'epoch': 0.01}) +[2026-06-24 12:01:07,136][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.3919714093208313, 'train/grad_norm_pre_clip': 31.580406188964844, 'timing/data': 0.00030830688774585724, 'timing/model': 0.8908817800693214, 'learning_rate/qwen_vl_interface': 9.999845889653509e-06, 'learning_rate/action_model': 9.999845889653508e-05, 'epoch': 0.01}) +[2026-06-24 12:01:08,029][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.4470522403717041, 'train/grad_norm_pre_clip': 56.977394104003906, 'timing/data': 0.00044178799726068974, 'timing/model': 0.8907780069857836, 'learning_rate/qwen_vl_interface': 9.999813526692496e-06, 'learning_rate/action_model': 9.999813526692495e-05, 'epoch': 0.01}) +[2026-06-24 12:01:08,920][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.34065166115760803, 'train/grad_norm_pre_clip': 47.02188491821289, 'timing/data': 0.0003064349293708801, 'timing/model': 0.8898759470321238, 'learning_rate/qwen_vl_interface': 9.999778081629055e-06, 'learning_rate/action_model': 9.999778081629054e-05, 'epoch': 0.01}) +[2026-06-24 12:01:09,812][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.5338248014450073, 'train/grad_norm_pre_clip': 40.06691360473633, 'timing/data': 0.0002950630150735378, 'timing/model': 0.8904427909292281, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.01}) +[2026-06-24 12:01:10,704][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.35209110379219055, 'train/grad_norm_pre_clip': 36.120521545410156, 'timing/data': 0.00029454194009304047, 'timing/model': 0.8896228550001979, 'learning_rate/qwen_vl_interface': 9.999697945288881e-06, 'learning_rate/action_model': 9.999697945288881e-05, 'epoch': 0.01}) +[2026-06-24 12:01:11,596][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.7173625230789185, 'train/grad_norm_pre_clip': 82.76729583740234, 'timing/data': 0.0003001319710165262, 'timing/model': 0.8910635050851852, 'learning_rate/qwen_vl_interface': 9.999653254064153e-06, 'learning_rate/action_model': 9.999653254064152e-05, 'epoch': 0.01}) +[2026-06-24 12:01:12,484][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.44086599349975586, 'train/grad_norm_pre_clip': 64.99128723144531, 'timing/data': 0.0002887509763240814, 'timing/model': 0.8862553869839758, 'learning_rate/qwen_vl_interface': 9.99960548084099e-06, 'learning_rate/action_model': 9.99960548084099e-05, 'epoch': 0.01}) +[2026-06-24 12:01:13,378][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.42793744802474976, 'train/grad_norm_pre_clip': 36.37701416015625, 'timing/data': 0.0002964348532259464, 'timing/model': 0.8918122500181198, 'learning_rate/qwen_vl_interface': 9.999554625650398e-06, 'learning_rate/action_model': 9.999554625650399e-05, 'epoch': 0.01}) +[2026-06-24 12:01:14,271][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.46633532643318176, 'train/grad_norm_pre_clip': 73.1980972290039, 'timing/data': 0.00031260494142770767, 'timing/model': 0.8910236619412899, 'learning_rate/qwen_vl_interface': 9.999500688525376e-06, 'learning_rate/action_model': 9.999500688525375e-05, 'epoch': 0.01}) +[2026-06-24 12:01:15,162][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.54857337474823, 'train/grad_norm_pre_clip': 118.91484069824219, 'timing/data': 0.0003045909106731415, 'timing/model': 0.8898232250940055, 'learning_rate/qwen_vl_interface': 9.99944366950092e-06, 'learning_rate/action_model': 9.999443669500919e-05, 'epoch': 0.01}) +[2026-06-24 12:01:16,051][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.47795751690864563, 'train/grad_norm_pre_clip': 62.3336296081543, 'timing/data': 0.00029453192837536335, 'timing/model': 0.8874891218729317, 'learning_rate/qwen_vl_interface': 9.99938356861403e-06, 'learning_rate/action_model': 9.99938356861403e-05, 'epoch': 0.01}) +[2026-06-24 12:01:16,944][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.267109215259552, 'train/grad_norm_pre_clip': 19.7296142578125, 'timing/data': 0.00030221580527722836, 'timing/model': 0.890967319952324, 'learning_rate/qwen_vl_interface': 9.999320385903709e-06, 'learning_rate/action_model': 9.999320385903708e-05, 'epoch': 0.01}) +[2026-06-24 12:01:17,835][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.4078116714954376, 'train/grad_norm_pre_clip': 26.01515007019043, 'timing/data': 0.0003214632160961628, 'timing/model': 0.8895521590020508, 'learning_rate/qwen_vl_interface': 9.999254121410948e-06, 'learning_rate/action_model': 9.999254121410949e-05, 'epoch': 0.01}) +[2026-06-24 12:01:18,727][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.29586392641067505, 'train/grad_norm_pre_clip': 20.574628829956055, 'timing/data': 0.00029016402550041676, 'timing/model': 0.8905766999814659, 'learning_rate/qwen_vl_interface': 9.999184775178751e-06, 'learning_rate/action_model': 9.999184775178751e-05, 'epoch': 0.01}) +[2026-06-24 12:01:19,618][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.48820826411247253, 'train/grad_norm_pre_clip': 79.67755889892578, 'timing/data': 0.0003622579388320446, 'timing/model': 0.8891096620354801, 'learning_rate/qwen_vl_interface': 9.999112347252117e-06, 'learning_rate/action_model': 9.999112347252116e-05, 'epoch': 0.01}) +[2026-06-24 12:01:20,511][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.4711487889289856, 'train/grad_norm_pre_clip': 89.13493347167969, 'timing/data': 0.00030217599123716354, 'timing/model': 0.8915492950472981, 'learning_rate/qwen_vl_interface': 9.99903683767804e-06, 'learning_rate/action_model': 9.99903683767804e-05, 'epoch': 0.01}) +[2026-06-24 12:01:21,403][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.764660120010376, 'train/grad_norm_pre_clip': 330.7972717285156, 'timing/data': 0.0004948368296027184, 'timing/model': 0.8900392148643732, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.01}) +[2026-06-24 12:01:22,293][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.9511524438858032, 'train/grad_norm_pre_clip': 206.91064453125, 'timing/data': 0.00029856921173632145, 'timing/model': 0.8883194839581847, 'learning_rate/qwen_vl_interface': 9.998876573785547e-06, 'learning_rate/action_model': 9.998876573785547e-05, 'epoch': 0.01}) +[2026-06-24 12:01:23,184][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.5906075835227966, 'train/grad_norm_pre_clip': 159.0599822998047, 'timing/data': 0.00041090999729931355, 'timing/model': 0.889213610207662, 'learning_rate/qwen_vl_interface': 9.998791819571126e-06, 'learning_rate/action_model': 9.998791819571127e-05, 'epoch': 0.01}) +[2026-06-24 12:01:24,076][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.4871075749397278, 'train/grad_norm_pre_clip': 99.26620483398438, 'timing/data': 0.0003078670706599951, 'timing/model': 0.8900633431039751, 'learning_rate/qwen_vl_interface': 9.998703983917252e-06, 'learning_rate/action_model': 9.998703983917252e-05, 'epoch': 0.01}) +[2026-06-24 12:01:24,965][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.5449178218841553, 'train/grad_norm_pre_clip': 92.38337707519531, 'timing/data': 0.0003179558552801609, 'timing/model': 0.8869772320613265, 'learning_rate/qwen_vl_interface': 9.998613066880919e-06, 'learning_rate/action_model': 9.998613066880918e-05, 'epoch': 0.01}) +[2026-06-24 12:01:25,851][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.4080210030078888, 'train/grad_norm_pre_clip': 65.92313385009766, 'timing/data': 0.00032776384614408016, 'timing/model': 0.8846158981323242, 'learning_rate/qwen_vl_interface': 9.99851906852112e-06, 'learning_rate/action_model': 9.998519068521119e-05, 'epoch': 0.01}) +[2026-06-24 12:01:26,741][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.4508291780948639, 'train/grad_norm_pre_clip': 43.4412727355957, 'timing/data': 0.00031725410372018814, 'timing/model': 0.8876908009406179, 'learning_rate/qwen_vl_interface': 9.998421988898851e-06, 'learning_rate/action_model': 9.998421988898851e-05, 'epoch': 0.01}) +[2026-06-24 12:01:27,633][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.40940672159194946, 'train/grad_norm_pre_clip': 22.542932510375977, 'timing/data': 0.00033851410262286663, 'timing/model': 0.8902137589175254, 'learning_rate/qwen_vl_interface': 9.998321828077107e-06, 'learning_rate/action_model': 9.998321828077106e-05, 'epoch': 0.01}) +[2026-06-24 12:01:28,524][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.5345260500907898, 'train/grad_norm_pre_clip': 42.98475646972656, 'timing/data': 0.00037085404619574547, 'timing/model': 0.8891563168726861, 'learning_rate/qwen_vl_interface': 9.99821858612088e-06, 'learning_rate/action_model': 9.99821858612088e-05, 'epoch': 0.01}) +[2026-06-24 12:01:29,413][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.3120919466018677, 'train/grad_norm_pre_clip': 15.826919555664062, 'timing/data': 0.0004796180874109268, 'timing/model': 0.8876582011580467, 'learning_rate/qwen_vl_interface': 9.998112263097164e-06, 'learning_rate/action_model': 9.998112263097165e-05, 'epoch': 0.01}) +[2026-06-24 12:01:30,305][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.43585824966430664, 'train/grad_norm_pre_clip': 24.5789737701416, 'timing/data': 0.0003684009425342083, 'timing/model': 0.8900842380244285, 'learning_rate/qwen_vl_interface': 9.99800285907495e-06, 'learning_rate/action_model': 9.99800285907495e-05, 'epoch': 0.01}) +[2026-06-24 12:01:31,198][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.37602633237838745, 'train/grad_norm_pre_clip': 31.788795471191406, 'timing/data': 0.0002960949204862118, 'timing/model': 0.8913363651372492, 'learning_rate/qwen_vl_interface': 9.997890374125228e-06, 'learning_rate/action_model': 9.997890374125227e-05, 'epoch': 0.01}) +[2026-06-24 12:01:32,091][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.2647179365158081, 'train/grad_norm_pre_clip': 20.37786865234375, 'timing/data': 0.00030121486634016037, 'timing/model': 0.8907528219278902, 'learning_rate/qwen_vl_interface': 9.997774808320992e-06, 'learning_rate/action_model': 9.997774808320992e-05, 'epoch': 0.01}) +[2026-06-24 12:01:32,983][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.5770015120506287, 'train/grad_norm_pre_clip': 56.39310073852539, 'timing/data': 0.00031826714985072613, 'timing/model': 0.8902869299054146, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.01}) +[2026-06-24 12:01:33,874][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.5683633089065552, 'train/grad_norm_pre_clip': 46.1095085144043, 'timing/data': 0.00030032312497496605, 'timing/model': 0.8896797229535878, 'learning_rate/qwen_vl_interface': 9.997534434450922e-06, 'learning_rate/action_model': 9.997534434450921e-05, 'epoch': 0.01}) +[2026-06-24 12:01:34,768][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.2528446912765503, 'train/grad_norm_pre_clip': 18.181039810180664, 'timing/data': 0.00029817898757755756, 'timing/model': 0.8921931830700487, 'learning_rate/qwen_vl_interface': 9.997409626541065e-06, 'learning_rate/action_model': 9.997409626541066e-05, 'epoch': 0.01}) +[2026-06-24 12:01:35,660][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.5102232098579407, 'train/grad_norm_pre_clip': 25.764156341552734, 'timing/data': 0.0002888110466301441, 'timing/model': 0.8898263298906386, 'learning_rate/qwen_vl_interface': 9.997281738088647e-06, 'learning_rate/action_model': 9.997281738088646e-05, 'epoch': 0.01}) +[2026-06-24 12:01:36,553][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.33326733112335205, 'train/grad_norm_pre_clip': 28.570960998535156, 'timing/data': 0.0003089490346610546, 'timing/model': 0.8915249230340123, 'learning_rate/qwen_vl_interface': 9.997150769176648e-06, 'learning_rate/action_model': 9.997150769176647e-05, 'epoch': 0.01}) +[2026-06-24 12:01:37,447][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.4267406165599823, 'train/grad_norm_pre_clip': 32.65029525756836, 'timing/data': 0.00033973692916333675, 'timing/model': 0.892064664978534, 'learning_rate/qwen_vl_interface': 9.997016719890052e-06, 'learning_rate/action_model': 9.997016719890052e-05, 'epoch': 0.01}) +[2026-06-24 12:01:38,341][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.35954922437667847, 'train/grad_norm_pre_clip': 32.80815505981445, 'timing/data': 0.0005096641834825277, 'timing/model': 0.8924835911020637, 'learning_rate/qwen_vl_interface': 9.996879590315847e-06, 'learning_rate/action_model': 9.996879590315846e-05, 'epoch': 0.01}) +[2026-06-24 12:01:39,232][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.46736273169517517, 'train/grad_norm_pre_clip': 43.14303970336914, 'timing/data': 0.00032649189233779907, 'timing/model': 0.888754865154624, 'learning_rate/qwen_vl_interface': 9.99673938054301e-06, 'learning_rate/action_model': 9.996739380543009e-05, 'epoch': 0.01}) +[2026-06-24 12:01:40,124][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.34281694889068604, 'train/grad_norm_pre_clip': 19.10613441467285, 'timing/data': 0.00029438198544085026, 'timing/model': 0.8905209950171411, 'learning_rate/qwen_vl_interface': 9.996596090662525e-06, 'learning_rate/action_model': 9.996596090662525e-05, 'epoch': 0.01}) +[2026-06-24 12:01:41,016][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.380034863948822, 'train/grad_norm_pre_clip': 46.746089935302734, 'timing/data': 0.0002999119460582733, 'timing/model': 0.8900554240681231, 'learning_rate/qwen_vl_interface': 9.996449720767369e-06, 'learning_rate/action_model': 9.996449720767368e-05, 'epoch': 0.01}) +[2026-06-24 12:01:41,910][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.6087859272956848, 'train/grad_norm_pre_clip': 99.91576385498047, 'timing/data': 0.0003032579552382231, 'timing/model': 0.8922392160166055, 'learning_rate/qwen_vl_interface': 9.99630027095252e-06, 'learning_rate/action_model': 9.99630027095252e-05, 'epoch': 0.01}) +[2026-06-24 12:01:42,803][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.3415355682373047, 'train/grad_norm_pre_clip': 19.098590850830078, 'timing/data': 0.00029143597930669785, 'timing/model': 0.891719734063372, 'learning_rate/qwen_vl_interface': 9.996147741314956e-06, 'learning_rate/action_model': 9.996147741314957e-05, 'epoch': 0.01}) +[2026-06-24 12:01:43,696][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.40355074405670166, 'train/grad_norm_pre_clip': 58.791847229003906, 'timing/data': 0.0002950930502265692, 'timing/model': 0.8909775940701365, 'learning_rate/qwen_vl_interface': 9.99599213195365e-06, 'learning_rate/action_model': 9.99599213195365e-05, 'epoch': 0.01}) +[2026-06-24 12:01:44,588][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.2985555827617645, 'train/grad_norm_pre_clip': 25.105493545532227, 'timing/data': 0.00029925117269158363, 'timing/model': 0.8902860898524523, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.01}) +[2026-06-24 12:01:45,480][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.5306692719459534, 'train/grad_norm_pre_clip': 96.19686889648438, 'timing/data': 0.00032098102383315563, 'timing/model': 0.8901338039431721, 'learning_rate/qwen_vl_interface': 9.995671674465709e-06, 'learning_rate/action_model': 9.995671674465709e-05, 'epoch': 0.01}) +[2026-06-24 12:01:46,370][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.2522432506084442, 'train/grad_norm_pre_clip': 30.520050048828125, 'timing/data': 0.00028914096765220165, 'timing/model': 0.8883147868327796, 'learning_rate/qwen_vl_interface': 9.99550682654701e-06, 'learning_rate/action_model': 9.995506826547011e-05, 'epoch': 0.01}) +[2026-06-24 12:01:47,266][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.4072759747505188, 'train/grad_norm_pre_clip': 36.84912109375, 'timing/data': 0.00031564105302095413, 'timing/model': 0.89430445083417, 'learning_rate/qwen_vl_interface': 9.995338899320457e-06, 'learning_rate/action_model': 9.995338899320456e-05, 'epoch': 0.01}) +[2026-06-24 12:01:48,157][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.25785329937934875, 'train/grad_norm_pre_clip': 27.32999038696289, 'timing/data': 0.0003279251977801323, 'timing/model': 0.8896784239914268, 'learning_rate/qwen_vl_interface': 9.995167892895011e-06, 'learning_rate/action_model': 9.99516789289501e-05, 'epoch': 0.02}) +[2026-06-24 12:01:49,051][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.351788192987442, 'train/grad_norm_pre_clip': 51.13921356201172, 'timing/data': 0.00029609515331685543, 'timing/model': 0.8923715809360147, 'learning_rate/qwen_vl_interface': 9.994993807381635e-06, 'learning_rate/action_model': 9.994993807381635e-05, 'epoch': 0.02}) +[2026-06-24 12:01:49,943][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.37391242384910583, 'train/grad_norm_pre_clip': 47.4573860168457, 'timing/data': 0.00030837790109217167, 'timing/model': 0.8899383330717683, 'learning_rate/qwen_vl_interface': 9.994816642893294e-06, 'learning_rate/action_model': 9.994816642893293e-05, 'epoch': 0.02}) +[2026-06-24 12:01:50,836][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.36193791031837463, 'train/grad_norm_pre_clip': 94.9853286743164, 'timing/data': 0.00034777214750647545, 'timing/model': 0.8911686788778752, 'learning_rate/qwen_vl_interface': 9.994636399544948e-06, 'learning_rate/action_model': 9.994636399544947e-05, 'epoch': 0.02}) +[2026-06-24 12:01:51,728][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.36421826481819153, 'train/grad_norm_pre_clip': 50.48857879638672, 'timing/data': 0.00029833894222974777, 'timing/model': 0.890341839985922, 'learning_rate/qwen_vl_interface': 9.994453077453552e-06, 'learning_rate/action_model': 9.994453077453552e-05, 'epoch': 0.02}) +[2026-06-24 12:01:52,622][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.31074875593185425, 'train/grad_norm_pre_clip': 48.046756744384766, 'timing/data': 0.00030633294954895973, 'timing/model': 0.8927518958225846, 'learning_rate/qwen_vl_interface': 9.994266676738065e-06, 'learning_rate/action_model': 9.994266676738065e-05, 'epoch': 0.02}) +[2026-06-24 12:01:53,513][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.1672266125679016, 'train/grad_norm_pre_clip': 51.92571258544922, 'timing/data': 0.0003885081969201565, 'timing/model': 0.8889335240237415, 'learning_rate/qwen_vl_interface': 9.99407719751944e-06, 'learning_rate/action_model': 9.994077197519439e-05, 'epoch': 0.02}) +[2026-06-24 12:01:54,404][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.39492982625961304, 'train/grad_norm_pre_clip': 54.57312774658203, 'timing/data': 0.0003062840551137924, 'timing/model': 0.8895698660053313, 'learning_rate/qwen_vl_interface': 9.993884639920626e-06, 'learning_rate/action_model': 9.993884639920626e-05, 'epoch': 0.02}) +[2026-06-24 12:01:55,297][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.3028842806816101, 'train/grad_norm_pre_clip': 31.13198471069336, 'timing/data': 0.0004849289543926716, 'timing/model': 0.8902020310051739, 'learning_rate/qwen_vl_interface': 9.993689004066573e-06, 'learning_rate/action_model': 9.993689004066573e-05, 'epoch': 0.02}) +[2026-06-24 12:01:56,190][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.36576178669929504, 'train/grad_norm_pre_clip': 76.87428283691406, 'timing/data': 0.0002993911039084196, 'timing/model': 0.8915686930995435, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.02}) +[2026-06-24 12:01:57,082][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.4859684109687805, 'train/grad_norm_pre_clip': 82.40151977539062, 'timing/data': 0.00029583415016531944, 'timing/model': 0.8906332200858742, 'learning_rate/qwen_vl_interface': 9.993288498102531e-06, 'learning_rate/action_model': 9.993288498102532e-05, 'epoch': 0.02}) +[2026-06-24 12:01:57,974][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.33564281463623047, 'train/grad_norm_pre_clip': 67.19178771972656, 'timing/data': 0.0006242000963538885, 'timing/model': 0.889378810999915, 'learning_rate/qwen_vl_interface': 9.993083628252424e-06, 'learning_rate/action_model': 9.993083628252424e-05, 'epoch': 0.02}) +[2026-06-24 12:01:59,025][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.5808464884757996, 'train/grad_norm_pre_clip': 125.01959991455078, 'timing/data': 0.00032190303318202496, 'timing/model': 1.0491437290329486, 'learning_rate/qwen_vl_interface': 9.992875680666849e-06, 'learning_rate/action_model': 9.992875680666849e-05, 'epoch': 0.02}) +[2026-06-24 12:01:59,917][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.4103471040725708, 'train/grad_norm_pre_clip': 80.04033660888672, 'timing/data': 0.00030497205443680286, 'timing/model': 0.8899350850842893, 'learning_rate/qwen_vl_interface': 9.992664655480736e-06, 'learning_rate/action_model': 9.992664655480735e-05, 'epoch': 0.02}) +[2026-06-24 12:02:00,807][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.2482578009366989, 'train/grad_norm_pre_clip': 58.14551544189453, 'timing/data': 0.0002916068769991398, 'timing/model': 0.8887066550087184, 'learning_rate/qwen_vl_interface': 9.992450552831018e-06, 'learning_rate/action_model': 9.992450552831017e-05, 'epoch': 0.02}) +[2026-06-24 12:02:01,697][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.4107004702091217, 'train/grad_norm_pre_clip': 28.320539474487305, 'timing/data': 0.00029901997186243534, 'timing/model': 0.8885431590024382, 'learning_rate/qwen_vl_interface': 9.992233372856623e-06, 'learning_rate/action_model': 9.992233372856624e-05, 'epoch': 0.02}) +[2026-06-24 12:02:02,592][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.3467194437980652, 'train/grad_norm_pre_clip': 71.108154296875, 'timing/data': 0.0003619480412453413, 'timing/model': 0.8925955821759999, 'learning_rate/qwen_vl_interface': 9.99201311569848e-06, 'learning_rate/action_model': 9.99201311569848e-05, 'epoch': 0.02}) +[2026-06-24 12:02:03,487][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.21314233541488647, 'train/grad_norm_pre_clip': 35.24311065673828, 'timing/data': 0.0002989100757986307, 'timing/model': 0.8940795541275293, 'learning_rate/qwen_vl_interface': 9.99178978149951e-06, 'learning_rate/action_model': 9.99178978149951e-05, 'epoch': 0.02}) +[2026-06-24 12:02:04,380][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.16656556725502014, 'train/grad_norm_pre_clip': 20.049604415893555, 'timing/data': 0.00030051288194954395, 'timing/model': 0.8906721801031381, 'learning_rate/qwen_vl_interface': 9.99156337040463e-06, 'learning_rate/action_model': 9.991563370404629e-05, 'epoch': 0.02}) +[2026-06-24 12:02:05,272][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.4559907615184784, 'train/grad_norm_pre_clip': 31.65842628479004, 'timing/data': 0.00028968206606805325, 'timing/model': 0.8899984480813146, 'learning_rate/qwen_vl_interface': 9.991333882560759e-06, 'learning_rate/action_model': 9.991333882560759e-05, 'epoch': 0.02}) +[2026-06-24 12:02:06,166][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.30464106798171997, 'train/grad_norm_pre_clip': 28.00669288635254, 'timing/data': 0.00029551307670772076, 'timing/model': 0.8929797550663352, 'learning_rate/qwen_vl_interface': 9.991101318116806e-06, 'learning_rate/action_model': 9.991101318116806e-05, 'epoch': 0.02}) +[2026-06-24 12:02:07,061][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.3549554944038391, 'train/grad_norm_pre_clip': 73.57906341552734, 'timing/data': 0.00033231195993721485, 'timing/model': 0.892248434945941, 'learning_rate/qwen_vl_interface': 9.990865677223684e-06, 'learning_rate/action_model': 9.990865677223684e-05, 'epoch': 0.02}) +[2026-06-24 12:02:07,955][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.4254610240459442, 'train/grad_norm_pre_clip': 52.60879898071289, 'timing/data': 0.00032644206658005714, 'timing/model': 0.8926120779942721, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.02}) +[2026-06-24 12:02:08,848][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.3756531774997711, 'train/grad_norm_pre_clip': 55.26115417480469, 'timing/data': 0.00032560015097260475, 'timing/model': 0.8918791059404612, 'learning_rate/qwen_vl_interface': 9.990385166703534e-06, 'learning_rate/action_model': 9.990385166703533e-05, 'epoch': 0.02}) +[2026-06-24 12:02:09,742][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.39055418968200684, 'train/grad_norm_pre_clip': 51.81329345703125, 'timing/data': 0.0002900029066950083, 'timing/model': 0.8912862660363317, 'learning_rate/qwen_vl_interface': 9.990140297388306e-06, 'learning_rate/action_model': 9.990140297388307e-05, 'epoch': 0.02}) +[2026-06-24 12:02:10,638][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.3783178925514221, 'train/grad_norm_pre_clip': 22.361282348632812, 'timing/data': 0.0004933038726449013, 'timing/model': 0.894003216875717, 'learning_rate/qwen_vl_interface': 9.989892352247505e-06, 'learning_rate/action_model': 9.989892352247504e-05, 'epoch': 0.02}) +[2026-06-24 12:02:11,532][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.2515397071838379, 'train/grad_norm_pre_clip': 23.59543228149414, 'timing/data': 0.00030129379592835903, 'timing/model': 0.8928954920265824, 'learning_rate/qwen_vl_interface': 9.989641331442014e-06, 'learning_rate/action_model': 9.989641331442013e-05, 'epoch': 0.02}) +[2026-06-24 12:02:12,426][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.4486597180366516, 'train/grad_norm_pre_clip': 29.455196380615234, 'timing/data': 0.00030592293478548527, 'timing/model': 0.8923061389941722, 'learning_rate/qwen_vl_interface': 9.989387235134722e-06, 'learning_rate/action_model': 9.989387235134722e-05, 'epoch': 0.02}) +[2026-06-24 12:02:13,320][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.38002949953079224, 'train/grad_norm_pre_clip': 77.62366485595703, 'timing/data': 0.000313909025862813, 'timing/model': 0.8919411150272936, 'learning_rate/qwen_vl_interface': 9.98913006349051e-06, 'learning_rate/action_model': 9.989130063490508e-05, 'epoch': 0.02}) +[2026-06-24 12:02:14,216][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.3420954942703247, 'train/grad_norm_pre_clip': 33.202392578125, 'timing/data': 0.0003208010457456112, 'timing/model': 0.8938957480713725, 'learning_rate/qwen_vl_interface': 9.988869816676247e-06, 'learning_rate/action_model': 9.988869816676246e-05, 'epoch': 0.02}) +[2026-06-24 12:02:15,109][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.32831424474716187, 'train/grad_norm_pre_clip': 30.0166015625, 'timing/data': 0.0003005529288202524, 'timing/model': 0.8920052170287818, 'learning_rate/qwen_vl_interface': 9.988606494860811e-06, 'learning_rate/action_model': 9.988606494860812e-05, 'epoch': 0.02}) +[2026-06-24 12:02:15,999][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.36946338415145874, 'train/grad_norm_pre_clip': 68.17326354980469, 'timing/data': 0.0003295179922133684, 'timing/model': 0.8881957500707358, 'learning_rate/qwen_vl_interface': 9.98834009821507e-06, 'learning_rate/action_model': 9.988340098215069e-05, 'epoch': 0.02}) +[2026-06-24 12:02:16,893][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.26285263895988464, 'train/grad_norm_pre_clip': 44.31599807739258, 'timing/data': 0.000300463056191802, 'timing/model': 0.8917287799995393, 'learning_rate/qwen_vl_interface': 9.98807062691188e-06, 'learning_rate/action_model': 9.988070626911879e-05, 'epoch': 0.02}) +[2026-06-24 12:02:17,787][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.3170672059059143, 'train/grad_norm_pre_clip': 45.130592346191406, 'timing/data': 0.00030406005680561066, 'timing/model': 0.8930204520002007, 'learning_rate/qwen_vl_interface': 9.987798081126099e-06, 'learning_rate/action_model': 9.987798081126099e-05, 'epoch': 0.02}) +[2026-06-24 12:02:18,680][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.290334552526474, 'train/grad_norm_pre_clip': 80.14299774169922, 'timing/data': 0.000344736035913229, 'timing/model': 0.8904985471162945, 'learning_rate/qwen_vl_interface': 9.987522461034584e-06, 'learning_rate/action_model': 9.987522461034584e-05, 'epoch': 0.02}) +[2026-06-24 12:02:19,574][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.6692122220993042, 'train/grad_norm_pre_clip': 98.59053802490234, 'timing/data': 0.00048438785597682, 'timing/model': 0.8922074700240046, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.02}) +[2026-06-24 12:02:20,468][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.3051232099533081, 'train/grad_norm_pre_clip': 43.64794158935547, 'timing/data': 0.0002976069226861, 'timing/model': 0.8924699418712407, 'learning_rate/qwen_vl_interface': 9.98696199865172e-06, 'learning_rate/action_model': 9.98696199865172e-05, 'epoch': 0.02}) +[2026-06-24 12:02:21,363][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.23753324151039124, 'train/grad_norm_pre_clip': 33.89912033081055, 'timing/data': 0.0002845840062946081, 'timing/model': 0.8932425810489804, 'learning_rate/qwen_vl_interface': 9.986677156724054e-06, 'learning_rate/action_model': 9.986677156724054e-05, 'epoch': 0.02}) +[2026-06-24 12:02:22,262][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.18167026340961456, 'train/grad_norm_pre_clip': 63.95382308959961, 'timing/data': 0.0002984190359711647, 'timing/model': 0.8970727580599487, 'learning_rate/qwen_vl_interface': 9.986389241218003e-06, 'learning_rate/action_model': 9.986389241218003e-05, 'epoch': 0.02}) +[2026-06-24 12:02:23,156][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.19929884374141693, 'train/grad_norm_pre_clip': 29.77646255493164, 'timing/data': 0.000313387019559741, 'timing/model': 0.8926144950091839, 'learning_rate/qwen_vl_interface': 9.986098252320397e-06, 'learning_rate/action_model': 9.986098252320397e-05, 'epoch': 0.02}) +[2026-06-24 12:02:24,053][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.3319108486175537, 'train/grad_norm_pre_clip': 58.326725006103516, 'timing/data': 0.00033914484083652496, 'timing/model': 0.894965501036495, 'learning_rate/qwen_vl_interface': 9.985804190220054e-06, 'learning_rate/action_model': 9.985804190220053e-05, 'epoch': 0.02}) +[2026-06-24 12:02:24,948][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 1.0945101976394653, 'train/grad_norm_pre_clip': 248.0982208251953, 'timing/data': 0.0003008029889315367, 'timing/model': 0.8936368529684842, 'learning_rate/qwen_vl_interface': 9.985507055107788e-06, 'learning_rate/action_model': 9.985507055107788e-05, 'epoch': 0.02}) +[2026-06-24 12:02:25,843][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.7645319104194641, 'train/grad_norm_pre_clip': 194.64366149902344, 'timing/data': 0.0002903041895478964, 'timing/model': 0.8933566180057824, 'learning_rate/qwen_vl_interface': 9.985206847176406e-06, 'learning_rate/action_model': 9.985206847176406e-05, 'epoch': 0.02}) +[2026-06-24 12:02:26,739][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.7446610927581787, 'train/grad_norm_pre_clip': 224.66004943847656, 'timing/data': 0.000472324900329113, 'timing/model': 0.8936807359568775, 'learning_rate/qwen_vl_interface': 9.984903566620711e-06, 'learning_rate/action_model': 9.984903566620712e-05, 'epoch': 0.02}) +[2026-06-24 12:02:27,636][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.33074313402175903, 'train/grad_norm_pre_clip': 83.005126953125, 'timing/data': 0.0003091490361839533, 'timing/model': 0.894801507005468, 'learning_rate/qwen_vl_interface': 9.984597213637497e-06, 'learning_rate/action_model': 9.984597213637497e-05, 'epoch': 0.02}) +[2026-06-24 12:02:28,532][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.43706801533699036, 'train/grad_norm_pre_clip': 65.62651062011719, 'timing/data': 0.0003090489190071821, 'timing/model': 0.8942000709939748, 'learning_rate/qwen_vl_interface': 9.984287788425555e-06, 'learning_rate/action_model': 9.984287788425555e-05, 'epoch': 0.02}) +[2026-06-24 12:02:29,426][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.27794790267944336, 'train/grad_norm_pre_clip': 31.628877639770508, 'timing/data': 0.0003250290174037218, 'timing/model': 0.8929816989693791, 'learning_rate/qwen_vl_interface': 9.983975291185666e-06, 'learning_rate/action_model': 9.983975291185667e-05, 'epoch': 0.02}) +[2026-06-24 12:02:30,323][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.4496888518333435, 'train/grad_norm_pre_clip': 66.25758361816406, 'timing/data': 0.0004830150865018368, 'timing/model': 0.8943876039702445, 'learning_rate/qwen_vl_interface': 9.983659722120607e-06, 'learning_rate/action_model': 9.983659722120608e-05, 'epoch': 0.02}) +[2026-06-24 12:02:31,217][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.4260959029197693, 'train/grad_norm_pre_clip': 52.509334564208984, 'timing/data': 0.00030089402571320534, 'timing/model': 0.892778669949621, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.02}) +[2026-06-24 12:02:32,113][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.37251588702201843, 'train/grad_norm_pre_clip': 46.26768112182617, 'timing/data': 0.0003165528178215027, 'timing/model': 0.8941056460607797, 'learning_rate/qwen_vl_interface': 9.983019369336048e-06, 'learning_rate/action_model': 9.983019369336048e-05, 'epoch': 0.02}) +[2026-06-24 12:02:33,005][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.25153714418411255, 'train/grad_norm_pre_clip': 35.340309143066406, 'timing/data': 0.0003050218801945448, 'timing/model': 0.8904477441683412, 'learning_rate/qwen_vl_interface': 9.982694586032066e-06, 'learning_rate/action_model': 9.982694586032065e-05, 'epoch': 0.02}) +[2026-06-24 12:02:33,899][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.2936348021030426, 'train/grad_norm_pre_clip': 30.572181701660156, 'timing/data': 0.00035417405888438225, 'timing/model': 0.8921148590743542, 'learning_rate/qwen_vl_interface': 9.982366731733948e-06, 'learning_rate/action_model': 9.982366731733947e-05, 'epoch': 0.02}) +[2026-06-24 12:02:34,794][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.4414898753166199, 'train/grad_norm_pre_clip': 57.049774169921875, 'timing/data': 0.00030485098250210285, 'timing/model': 0.89301947504282, 'learning_rate/qwen_vl_interface': 9.982035806654437e-06, 'learning_rate/action_model': 9.982035806654437e-05, 'epoch': 0.02}) +[2026-06-24 12:02:35,690][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.2686220407485962, 'train/grad_norm_pre_clip': 32.81814956665039, 'timing/data': 0.0002906240988522768, 'timing/model': 0.8940988171380013, 'learning_rate/qwen_vl_interface': 9.981701811008267e-06, 'learning_rate/action_model': 9.981701811008267e-05, 'epoch': 0.02}) +[2026-06-24 12:02:36,584][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.3422243893146515, 'train/grad_norm_pre_clip': 58.44093704223633, 'timing/data': 0.00030368915759027004, 'timing/model': 0.8923254851251841, 'learning_rate/qwen_vl_interface': 9.981364745012163e-06, 'learning_rate/action_model': 9.981364745012163e-05, 'epoch': 0.02}) +[2026-06-24 12:02:37,483][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.3397659361362457, 'train/grad_norm_pre_clip': 51.336517333984375, 'timing/data': 0.0004216700326651335, 'timing/model': 0.8975330230314285, 'learning_rate/qwen_vl_interface': 9.981024608884842e-06, 'learning_rate/action_model': 9.981024608884843e-05, 'epoch': 0.02}) +[2026-06-24 12:02:38,376][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.33426013588905334, 'train/grad_norm_pre_clip': 76.12479400634766, 'timing/data': 0.00028700800612568855, 'timing/model': 0.8914943090640008, 'learning_rate/qwen_vl_interface': 9.980681402847017e-06, 'learning_rate/action_model': 9.980681402847016e-05, 'epoch': 0.02}) +[2026-06-24 12:02:39,273][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.5333954691886902, 'train/grad_norm_pre_clip': 76.81196594238281, 'timing/data': 0.0003650940489023924, 'timing/model': 0.8952340870164335, 'learning_rate/qwen_vl_interface': 9.980335127121391e-06, 'learning_rate/action_model': 9.98033512712139e-05, 'epoch': 0.02}) +[2026-06-24 12:02:40,169][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.4171872138977051, 'train/grad_norm_pre_clip': 77.49604797363281, 'timing/data': 0.0002984090242534876, 'timing/model': 0.8938589708413929, 'learning_rate/qwen_vl_interface': 9.979985781932658e-06, 'learning_rate/action_model': 9.979985781932657e-05, 'epoch': 0.02}) +[2026-06-24 12:02:41,064][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.263481080532074, 'train/grad_norm_pre_clip': 22.62261390686035, 'timing/data': 0.00030629406683146954, 'timing/model': 0.8933923980221152, 'learning_rate/qwen_vl_interface': 9.979633367507501e-06, 'learning_rate/action_model': 9.979633367507502e-05, 'epoch': 0.02}) +[2026-06-24 12:02:41,960][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.4152153432369232, 'train/grad_norm_pre_clip': 40.8065071105957, 'timing/data': 0.0003016549162566662, 'timing/model': 0.8946832669898868, 'learning_rate/qwen_vl_interface': 9.979277884074604e-06, 'learning_rate/action_model': 9.979277884074604e-05, 'epoch': 0.02}) +[2026-06-24 12:02:42,857][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.4602353274822235, 'train/grad_norm_pre_clip': 54.251426696777344, 'timing/data': 0.0003251391462981701, 'timing/model': 0.8951302459463477, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.02}) +[2026-06-24 12:02:43,752][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.4182955324649811, 'train/grad_norm_pre_clip': 81.78726959228516, 'timing/data': 0.0002967161126434803, 'timing/model': 0.8936330410651863, 'learning_rate/qwen_vl_interface': 9.978557711110245e-06, 'learning_rate/action_model': 9.978557711110244e-05, 'epoch': 0.02}) +[2026-06-24 12:02:44,648][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.3469885587692261, 'train/grad_norm_pre_clip': 65.33102416992188, 'timing/data': 0.00029345997609198093, 'timing/model': 0.893569272942841, 'learning_rate/qwen_vl_interface': 9.978193022046097e-06, 'learning_rate/action_model': 9.978193022046097e-05, 'epoch': 0.02}) +[2026-06-24 12:02:45,543][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.48358678817749023, 'train/grad_norm_pre_clip': 60.10219192504883, 'timing/data': 0.00033251289278268814, 'timing/model': 0.8939296090975404, 'learning_rate/qwen_vl_interface': 9.977825264908832e-06, 'learning_rate/action_model': 9.977825264908832e-05, 'epoch': 0.02}) +[2026-06-24 12:02:46,439][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.35927265882492065, 'train/grad_norm_pre_clip': 48.09634017944336, 'timing/data': 0.0003007939085364342, 'timing/model': 0.8946262840181589, 'learning_rate/qwen_vl_interface': 9.97745443993708e-06, 'learning_rate/action_model': 9.97745443993708e-05, 'epoch': 0.02}) +[2026-06-24 12:02:47,334][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.21390776336193085, 'train/grad_norm_pre_clip': 43.781131744384766, 'timing/data': 0.0002948429901152849, 'timing/model': 0.8928378941491246, 'learning_rate/qwen_vl_interface': 9.977080547371468e-06, 'learning_rate/action_model': 9.977080547371468e-05, 'epoch': 0.02}) +[2026-06-24 12:02:48,230][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.1858048141002655, 'train/grad_norm_pre_clip': 39.17136001586914, 'timing/data': 0.0003065250348299742, 'timing/model': 0.8948446149006486, 'learning_rate/qwen_vl_interface': 9.97670358745461e-06, 'learning_rate/action_model': 9.97670358745461e-05, 'epoch': 0.02}) +[2026-06-24 12:02:49,126][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.2682470977306366, 'train/grad_norm_pre_clip': 31.755430221557617, 'timing/data': 0.0002962250728160143, 'timing/model': 0.8935676629189402, 'learning_rate/qwen_vl_interface': 9.976323560431112e-06, 'learning_rate/action_model': 9.976323560431111e-05, 'epoch': 0.02}) +[2026-06-24 12:02:50,018][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.2832711637020111, 'train/grad_norm_pre_clip': 47.9371452331543, 'timing/data': 0.00034063798375427723, 'timing/model': 0.8902511689811945, 'learning_rate/qwen_vl_interface': 9.975940466547567e-06, 'learning_rate/action_model': 9.975940466547567e-05, 'epoch': 0.02}) +[2026-06-24 12:02:50,917][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.3435204029083252, 'train/grad_norm_pre_clip': 35.201416015625, 'timing/data': 0.00033807381987571716, 'timing/model': 0.896381467115134, 'learning_rate/qwen_vl_interface': 9.975554306052563e-06, 'learning_rate/action_model': 9.975554306052562e-05, 'epoch': 0.02}) +[2026-06-24 12:02:51,816][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.302511990070343, 'train/grad_norm_pre_clip': 58.192935943603516, 'timing/data': 0.0003083879128098488, 'timing/model': 0.8977521848864853, 'learning_rate/qwen_vl_interface': 9.975165079196674e-06, 'learning_rate/action_model': 9.975165079196673e-05, 'epoch': 0.02}) +[2026-06-24 12:02:52,711][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.390292227268219, 'train/grad_norm_pre_clip': 51.7353630065918, 'timing/data': 0.0003299778327345848, 'timing/model': 0.893289454979822, 'learning_rate/qwen_vl_interface': 9.974772786232466e-06, 'learning_rate/action_model': 9.974772786232466e-05, 'epoch': 0.02}) +[2026-06-24 12:02:53,609][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.3546586334705353, 'train/grad_norm_pre_clip': 52.68258285522461, 'timing/data': 0.0002998721320182085, 'timing/model': 0.8962338420096785, 'learning_rate/qwen_vl_interface': 9.974377427414493e-06, 'learning_rate/action_model': 9.974377427414493e-05, 'epoch': 0.02}) +[2026-06-24 12:02:54,507][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.44287773966789246, 'train/grad_norm_pre_clip': 83.28970336914062, 'timing/data': 0.00032041105441749096, 'timing/model': 0.8962409759406, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.02}) +[2026-06-24 12:02:55,405][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.2745663821697235, 'train/grad_norm_pre_clip': 47.79782485961914, 'timing/data': 0.00029892008751630783, 'timing/model': 0.8966556340456009, 'learning_rate/qwen_vl_interface': 9.973577513245417e-06, 'learning_rate/action_model': 9.973577513245417e-05, 'epoch': 0.02}) +[2026-06-24 12:02:56,302][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.42365387082099915, 'train/grad_norm_pre_clip': 130.14083862304688, 'timing/data': 0.00030202604830265045, 'timing/model': 0.8948604899924248, 'learning_rate/qwen_vl_interface': 9.973172958413372e-06, 'learning_rate/action_model': 9.97317295841337e-05, 'epoch': 0.02}) +[2026-06-24 12:02:57,194][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.34464532136917114, 'train/grad_norm_pre_clip': 81.17446899414062, 'timing/data': 0.00029742694459855556, 'timing/model': 0.8910885050427169, 'learning_rate/qwen_vl_interface': 9.97276533876567e-06, 'learning_rate/action_model': 9.97276533876567e-05, 'epoch': 0.02}) +[2026-06-24 12:02:58,087][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.25192558765411377, 'train/grad_norm_pre_clip': 80.9249038696289, 'timing/data': 0.00029425090178847313, 'timing/model': 0.8909652340225875, 'learning_rate/qwen_vl_interface': 9.972354654566816e-06, 'learning_rate/action_model': 9.972354654566816e-05, 'epoch': 0.02}) +[2026-06-24 12:02:58,988][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.3178113102912903, 'train/grad_norm_pre_clip': 44.75986862182617, 'timing/data': 0.0003186371177434921, 'timing/model': 0.899538226891309, 'learning_rate/qwen_vl_interface': 9.971940906083295e-06, 'learning_rate/action_model': 9.971940906083295e-05, 'epoch': 0.02}) +[2026-06-24 12:02:59,889][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.351895272731781, 'train/grad_norm_pre_clip': 63.92270278930664, 'timing/data': 0.00045881979167461395, 'timing/model': 0.8981888100970536, 'learning_rate/qwen_vl_interface': 9.971524093583586e-06, 'learning_rate/action_model': 9.971524093583586e-05, 'epoch': 0.02}) +[2026-06-24 12:03:00,786][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.16232827305793762, 'train/grad_norm_pre_clip': 75.34506225585938, 'timing/data': 0.00034036790020763874, 'timing/model': 0.8952650618739426, 'learning_rate/qwen_vl_interface': 9.971104217338154e-06, 'learning_rate/action_model': 9.971104217338153e-05, 'epoch': 0.02}) +[2026-06-24 12:03:01,679][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.30967220664024353, 'train/grad_norm_pre_clip': 82.39599609375, 'timing/data': 0.0002913959324359894, 'timing/model': 0.892096065916121, 'learning_rate/qwen_vl_interface': 9.970681277619452e-06, 'learning_rate/action_model': 9.970681277619452e-05, 'epoch': 0.02}) +[2026-06-24 12:03:02,573][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.2155628502368927, 'train/grad_norm_pre_clip': 28.25773811340332, 'timing/data': 0.0003283850383013487, 'timing/model': 0.8923025329131633, 'learning_rate/qwen_vl_interface': 9.97025527470192e-06, 'learning_rate/action_model': 9.970255274701919e-05, 'epoch': 0.02}) +[2026-06-24 12:03:03,469][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.3830280303955078, 'train/grad_norm_pre_clip': 101.90276336669922, 'timing/data': 0.00031335698440670967, 'timing/model': 0.8941633100621402, 'learning_rate/qwen_vl_interface': 9.969826208861987e-06, 'learning_rate/action_model': 9.969826208861987e-05, 'epoch': 0.02}) +[2026-06-24 12:03:04,364][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.29935574531555176, 'train/grad_norm_pre_clip': 60.201438903808594, 'timing/data': 0.0002908350434154272, 'timing/model': 0.8936196120921522, 'learning_rate/qwen_vl_interface': 9.96939408037807e-06, 'learning_rate/action_model': 9.96939408037807e-05, 'epoch': 0.02}) +[2026-06-24 12:03:05,264][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.2854411005973816, 'train/grad_norm_pre_clip': 34.8510856628418, 'timing/data': 0.0003128969110548496, 'timing/model': 0.8980261390097439, 'learning_rate/qwen_vl_interface': 9.968958889530573e-06, 'learning_rate/action_model': 9.968958889530573e-05, 'epoch': 0.02}) +[2026-06-24 12:03:06,161][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.26940134167671204, 'train/grad_norm_pre_clip': 58.92642593383789, 'timing/data': 0.0005376979243010283, 'timing/model': 0.8951884831767529, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.02}) +[2026-06-24 12:03:07,056][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.27031996846199036, 'train/grad_norm_pre_clip': 60.92008972167969, 'timing/data': 0.00031795608811080456, 'timing/model': 0.8929236710537225, 'learning_rate/qwen_vl_interface': 9.968079321876382e-06, 'learning_rate/action_model': 9.968079321876382e-05, 'epoch': 0.02}) +[2026-06-24 12:03:07,953][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.21982404589653015, 'train/grad_norm_pre_clip': 41.45768737792969, 'timing/data': 0.00033752317540347576, 'timing/model': 0.8954326619859785, 'learning_rate/qwen_vl_interface': 9.967634945640432e-06, 'learning_rate/action_model': 9.967634945640431e-05, 'epoch': 0.02}) +[2026-06-24 12:03:08,848][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.459952175617218, 'train/grad_norm_pre_clip': 117.18415069580078, 'timing/data': 0.0002955941017717123, 'timing/model': 0.8936227620579302, 'learning_rate/qwen_vl_interface': 9.967187508182385e-06, 'learning_rate/action_model': 9.967187508182385e-05, 'epoch': 0.02}) +[2026-06-24 12:03:09,744][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.31743305921554565, 'train/grad_norm_pre_clip': 93.17973327636719, 'timing/data': 0.0002934299409389496, 'timing/model': 0.8939823559485376, 'learning_rate/qwen_vl_interface': 9.966737009792577e-06, 'learning_rate/action_model': 9.966737009792577e-05, 'epoch': 0.02}) +[2026-06-24 12:03:10,641][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.23505856096744537, 'train/grad_norm_pre_clip': 50.01567840576172, 'timing/data': 0.00033316388726234436, 'timing/model': 0.8952771530020982, 'learning_rate/qwen_vl_interface': 9.966283450763333e-06, 'learning_rate/action_model': 9.966283450763333e-05, 'epoch': 0.02}) +[2026-06-24 12:03:11,536][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.16096621751785278, 'train/grad_norm_pre_clip': 39.26427459716797, 'timing/data': 0.0003237668424844742, 'timing/model': 0.893217294011265, 'learning_rate/qwen_vl_interface': 9.965826831388962e-06, 'learning_rate/action_model': 9.965826831388961e-05, 'epoch': 0.02}) +[2026-06-24 12:03:12,430][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 250 +[2026-06-24 12:03:54,886][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 250 +[2026-06-24 12:03:54,886][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 250 +[2026-06-24 12:03:55,084][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmpui5k_wlf/test.c -o /tmp/tmpui5k_wlf/test.o +[2026-06-24 12:03:55,101][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmpui5k_wlf/test.o -laio -o /tmp/tmpui5k_wlf/a.out +[2026-06-24 12:03:55,819][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmp31f5vd80/test.c -o /tmp/tmp31f5vd80/test.o +[2026-06-24 12:03:55,835][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmp31f5vd80/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp31f5vd80/a.out +[2026-06-24 12:03:55,874][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmpe6md63mq/test.c -o /tmp/tmpe6md63mq/test.o +[2026-06-24 12:03:55,890][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmpe6md63mq/test.o -laio -o /tmp/tmpe6md63mq/a.out +[2026-06-24 12:04:49,901][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 250 +[2026-06-24 12:05:00,187][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 250 +[2026-06-24 12:05:00,188][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.2625819444656372, 'train/grad_norm_pre_clip': 35.650535583496094, 'eval/action_loss/seconds': 42.453660253901035, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.16421595215797424, 'eval/latency_2/loss': 0.16421595215797424, 'eval/flappy/loss': 0.16421595215797424, 'eval/flappy/latency_2/loss': 0.16421595215797424, 'eval/flap_precision': 0.7071984435797666, 'eval/flap_recall': 0.9864314789687924, 'eval/flap_f1': 0.8237960339943343, 'eval/flap_rate_model': 0.28555555555555556, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8237960339943343, 'eval/latency_2/flap_precision': 0.7071984435797666, 'eval/latency_2/flap_recall': 0.9864314789687924, 'eval/latency_2/flap_f1': 0.8237960339943343, 'eval/latency_2/flap_rate_model': 0.28555555555555556, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8237960339943343, 'eval/action_classification/data_wait_seconds': 23.460111618041992, 'eval/action_classification/predict_seconds': 29.32460594177246, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 6.119999999999993, 'rl_games_eval/mid_train/mean_length': 66.8, 'rl_games_eval/mid_train/std_reward': 1.8059900331950844, 'rl_games_eval/mid_train/std_length': 14.4, 'rl_games_eval/mid_train/macro_mean_reward': 6.119999999999993, 'rl_games_eval/mid_train/macro_mean_length': 66.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 6.119999999999993, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 66.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 1.8059900331950844, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 14.4, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 6.119999999999993, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0002992108929902315, 'timing/model': 0.89289053180255, 'learning_rate/qwen_vl_interface': 9.965367151965757e-06, 'learning_rate/action_model': 9.965367151965756e-05, 'epoch': 0.02}) +[2026-06-24 12:05:01,242][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.4776592254638672, 'train/grad_norm_pre_clip': 84.17909240722656, 'timing/data': 0.00042417505756020546, 'timing/model': 1.052795710042119, 'learning_rate/qwen_vl_interface': 9.964904412792002e-06, 'learning_rate/action_model': 9.964904412792002e-05, 'epoch': 0.02}) +[2026-06-24 12:05:02,299][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.2539065182209015, 'train/grad_norm_pre_clip': 55.294532775878906, 'timing/data': 0.0003077159635722637, 'timing/model': 1.0547580681741238, 'learning_rate/qwen_vl_interface': 9.964438614167964e-06, 'learning_rate/action_model': 9.964438614167964e-05, 'epoch': 0.02}) +[2026-06-24 12:05:03,361][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.24110162258148193, 'train/grad_norm_pre_clip': 31.83380889892578, 'timing/data': 0.0003242169041186571, 'timing/model': 1.060357681941241, 'learning_rate/qwen_vl_interface': 9.963969756395892e-06, 'learning_rate/action_model': 9.96396975639589e-05, 'epoch': 0.02}) +[2026-06-24 12:05:04,422][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.33314475417137146, 'train/grad_norm_pre_clip': 41.04023361206055, 'timing/data': 0.0003163230139762163, 'timing/model': 1.0597103999461979, 'learning_rate/qwen_vl_interface': 9.963497839780025e-06, 'learning_rate/action_model': 9.963497839780024e-05, 'epoch': 0.02}) +[2026-06-24 12:05:05,485][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.5285628437995911, 'train/grad_norm_pre_clip': 100.34422302246094, 'timing/data': 0.0004581189714372158, 'timing/model': 1.0612053209915757, 'learning_rate/qwen_vl_interface': 9.963022864626582e-06, 'learning_rate/action_model': 9.963022864626581e-05, 'epoch': 0.02}) +[2026-06-24 12:05:06,546][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.5042048692703247, 'train/grad_norm_pre_clip': 106.76134490966797, 'timing/data': 0.00033200206235051155, 'timing/model': 1.0590968609321862, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.02}) +[2026-06-24 12:05:07,607][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.49577048420906067, 'train/grad_norm_pre_clip': 135.26177978515625, 'timing/data': 0.00037384103052318096, 'timing/model': 1.059446055907756, 'learning_rate/qwen_vl_interface': 9.962063739941783e-06, 'learning_rate/action_model': 9.962063739941783e-05, 'epoch': 0.02}) +[2026-06-24 12:05:08,884][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.44637787342071533, 'train/grad_norm_pre_clip': 53.76901626586914, 'timing/data': 0.000308578135445714, 'timing/model': 1.2759096471127123, 'learning_rate/qwen_vl_interface': 9.961579591032791e-06, 'learning_rate/action_model': 9.96157959103279e-05, 'epoch': 0.02}) +[2026-06-24 12:05:09,948][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.3697589039802551, 'train/grad_norm_pre_clip': 62.817745208740234, 'timing/data': 0.00030000205151736736, 'timing/model': 1.0622550810221583, 'learning_rate/qwen_vl_interface': 9.961092384830956e-06, 'learning_rate/action_model': 9.961092384830956e-05, 'epoch': 0.03}) +[2026-06-24 12:05:11,008][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.359317809343338, 'train/grad_norm_pre_clip': 54.04401397705078, 'timing/data': 0.0003557868767529726, 'timing/model': 1.057921402156353, 'learning_rate/qwen_vl_interface': 9.96060212165242e-06, 'learning_rate/action_model': 9.96060212165242e-05, 'epoch': 0.03}) +[2026-06-24 12:05:12,068][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.23557382822036743, 'train/grad_norm_pre_clip': 44.251243591308594, 'timing/data': 0.00035769096575677395, 'timing/model': 1.058853279100731, 'learning_rate/qwen_vl_interface': 9.96010880181531e-06, 'learning_rate/action_model': 9.960108801815309e-05, 'epoch': 0.03}) +[2026-06-24 12:05:13,131][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.2505744695663452, 'train/grad_norm_pre_clip': 23.463319778442383, 'timing/data': 0.0003149700351059437, 'timing/model': 1.061495820991695, 'learning_rate/qwen_vl_interface': 9.959612425639736e-06, 'learning_rate/action_model': 9.959612425639735e-05, 'epoch': 0.03}) +[2026-06-24 12:05:14,190][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.2215927094221115, 'train/grad_norm_pre_clip': 35.52462387084961, 'timing/data': 0.0003035888075828552, 'timing/model': 1.057166479062289, 'learning_rate/qwen_vl_interface': 9.95911299344779e-06, 'learning_rate/action_model': 9.959112993447789e-05, 'epoch': 0.03}) +[2026-06-24 12:05:15,254][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.3238428831100464, 'train/grad_norm_pre_clip': 28.086471557617188, 'timing/data': 0.00038298708386719227, 'timing/model': 1.062201094115153, 'learning_rate/qwen_vl_interface': 9.958610505563546e-06, 'learning_rate/action_model': 9.958610505563546e-05, 'epoch': 0.03}) +[2026-06-24 12:05:16,322][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.17958030104637146, 'train/grad_norm_pre_clip': 58.67812728881836, 'timing/data': 0.0007421111222356558, 'timing/model': 1.0655383181292564, 'learning_rate/qwen_vl_interface': 9.95810496231307e-06, 'learning_rate/action_model': 9.95810496231307e-05, 'epoch': 0.03}) +[2026-06-24 12:05:17,388][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.30917543172836304, 'train/grad_norm_pre_clip': 89.26861572265625, 'timing/data': 0.0003117539454251528, 'timing/model': 1.0643247950356454, 'learning_rate/qwen_vl_interface': 9.957596364024396e-06, 'learning_rate/action_model': 9.957596364024396e-05, 'epoch': 0.03}) +[2026-06-24 12:05:18,458][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.3455248773097992, 'train/grad_norm_pre_clip': 86.58922576904297, 'timing/data': 0.00041539897210896015, 'timing/model': 1.068624212173745, 'learning_rate/qwen_vl_interface': 9.957084711027553e-06, 'learning_rate/action_model': 9.957084711027552e-05, 'epoch': 0.03}) +[2026-06-24 12:05:19,522][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.1846778243780136, 'train/grad_norm_pre_clip': 29.8946475982666, 'timing/data': 0.0004852700512856245, 'timing/model': 1.0616508559323847, 'learning_rate/qwen_vl_interface': 9.956570003654544e-06, 'learning_rate/action_model': 9.956570003654544e-05, 'epoch': 0.03}) +[2026-06-24 12:05:20,597][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.35489439964294434, 'train/grad_norm_pre_clip': 80.28050994873047, 'timing/data': 0.0002968159969896078, 'timing/model': 1.0738103110343218, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.03}) +[2026-06-24 12:05:21,661][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.2774985730648041, 'train/grad_norm_pre_clip': 54.66050720214844, 'timing/data': 0.000325449975207448, 'timing/model': 1.0626402308698744, 'learning_rate/qwen_vl_interface': 9.955531427117964e-06, 'learning_rate/action_model': 9.955531427117964e-05, 'epoch': 0.03}) +[2026-06-24 12:05:22,724][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.22995369136333466, 'train/grad_norm_pre_clip': 34.85809326171875, 'timing/data': 0.00029765814542770386, 'timing/model': 1.06085937586613, 'learning_rate/qwen_vl_interface': 9.955007558628316e-06, 'learning_rate/action_model': 9.955007558628316e-05, 'epoch': 0.03}) +[2026-06-24 12:05:23,788][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.38460707664489746, 'train/grad_norm_pre_clip': 56.112178802490234, 'timing/data': 0.000304289860650897, 'timing/model': 1.0630874908529222, 'learning_rate/qwen_vl_interface': 9.954480637110346e-06, 'learning_rate/action_model': 9.954480637110344e-05, 'epoch': 0.03}) +[2026-06-24 12:05:25,063][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.3095514178276062, 'train/grad_norm_pre_clip': 35.49485778808594, 'timing/data': 0.00030478090047836304, 'timing/model': 1.272960245842114, 'learning_rate/qwen_vl_interface': 9.953950662905963e-06, 'learning_rate/action_model': 9.953950662905964e-05, 'epoch': 0.03}) +[2026-06-24 12:05:26,126][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.4358022212982178, 'train/grad_norm_pre_clip': 66.86927795410156, 'timing/data': 0.0002930690534412861, 'timing/model': 1.0622120930347592, 'learning_rate/qwen_vl_interface': 9.953417636359067e-06, 'learning_rate/action_model': 9.953417636359067e-05, 'epoch': 0.03}) +[2026-06-24 12:05:27,193][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.16585470736026764, 'train/grad_norm_pre_clip': 23.373310089111328, 'timing/data': 0.00048796506598591805, 'timing/model': 1.06421990506351, 'learning_rate/qwen_vl_interface': 9.95288155781553e-06, 'learning_rate/action_model': 9.95288155781553e-05, 'epoch': 0.03}) +[2026-06-24 12:05:28,255][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.3002619743347168, 'train/grad_norm_pre_clip': 39.03791427612305, 'timing/data': 0.0002958450932055712, 'timing/model': 1.0605299919843674, 'learning_rate/qwen_vl_interface': 9.952342427623209e-06, 'learning_rate/action_model': 9.952342427623208e-05, 'epoch': 0.03}) +[2026-06-24 12:05:29,318][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.1471143513917923, 'train/grad_norm_pre_clip': 44.51679229736328, 'timing/data': 0.00029721716418862343, 'timing/model': 1.0615892680361867, 'learning_rate/qwen_vl_interface': 9.95180024613194e-06, 'learning_rate/action_model': 9.951800246131939e-05, 'epoch': 0.03}) +[2026-06-24 12:05:30,382][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.16351601481437683, 'train/grad_norm_pre_clip': 45.83628463745117, 'timing/data': 0.00029909098520874977, 'timing/model': 1.0623322001192719, 'learning_rate/qwen_vl_interface': 9.951255013693537e-06, 'learning_rate/action_model': 9.951255013693536e-05, 'epoch': 0.03}) +[2026-06-24 12:05:31,446][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.12961682677268982, 'train/grad_norm_pre_clip': 18.307497024536133, 'timing/data': 0.00031231483444571495, 'timing/model': 1.0623650029301643, 'learning_rate/qwen_vl_interface': 9.950706730661797e-06, 'learning_rate/action_model': 9.950706730661797e-05, 'epoch': 0.03}) +[2026-06-24 12:05:32,510][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.1541968435049057, 'train/grad_norm_pre_clip': 30.811689376831055, 'timing/data': 0.0003131669946014881, 'timing/model': 1.063015362014994, 'learning_rate/qwen_vl_interface': 9.950155397392494e-06, 'learning_rate/action_model': 9.950155397392493e-05, 'epoch': 0.03}) +[2026-06-24 12:05:33,577][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.14579761028289795, 'train/grad_norm_pre_clip': 30.006088256835938, 'timing/data': 0.00033881491981446743, 'timing/model': 1.064495787024498, 'learning_rate/qwen_vl_interface': 9.949601014243382e-06, 'learning_rate/action_model': 9.949601014243382e-05, 'epoch': 0.03}) +[2026-06-24 12:05:34,645][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.3318730890750885, 'train/grad_norm_pre_clip': 63.360595703125, 'timing/data': 0.0004960088990628719, 'timing/model': 1.0666555839125067, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.03}) +[2026-06-24 12:05:35,715][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.24732132256031036, 'train/grad_norm_pre_clip': 107.76222229003906, 'timing/data': 0.00042172009125351906, 'timing/model': 1.0679564331658185, 'learning_rate/qwen_vl_interface': 9.948483099746645e-06, 'learning_rate/action_model': 9.948483099746645e-05, 'epoch': 0.03}) +[2026-06-24 12:05:36,779][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.3396296501159668, 'train/grad_norm_pre_clip': 113.7723617553711, 'timing/data': 0.0005177001003175974, 'timing/model': 1.0618625849019736, 'learning_rate/qwen_vl_interface': 9.947919569124423e-06, 'learning_rate/action_model': 9.947919569124422e-05, 'epoch': 0.03}) +[2026-06-24 12:05:37,845][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.40066930651664734, 'train/grad_norm_pre_clip': 143.5882568359375, 'timing/data': 0.00028578611090779305, 'timing/model': 1.0640444438904524, 'learning_rate/qwen_vl_interface': 9.947352990073196e-06, 'learning_rate/action_model': 9.947352990073195e-05, 'epoch': 0.03}) +[2026-06-24 12:05:38,909][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.15610551834106445, 'train/grad_norm_pre_clip': 102.48873138427734, 'timing/data': 0.000347210094332695, 'timing/model': 1.062738848151639, 'learning_rate/qwen_vl_interface': 9.946783362960612e-06, 'learning_rate/action_model': 9.946783362960612e-05, 'epoch': 0.03}) +[2026-06-24 12:05:39,974][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.30099934339523315, 'train/grad_norm_pre_clip': 137.30690002441406, 'timing/data': 0.0003016861155629158, 'timing/model': 1.0633320910856128, 'learning_rate/qwen_vl_interface': 9.946210688156297e-06, 'learning_rate/action_model': 9.946210688156297e-05, 'epoch': 0.03}) +[2026-06-24 12:05:41,336][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.331588476896286, 'train/grad_norm_pre_clip': 81.11073303222656, 'timing/data': 0.0003215731121599674, 'timing/model': 1.3599991609808058, 'learning_rate/qwen_vl_interface': 9.945634966031852e-06, 'learning_rate/action_model': 9.945634966031851e-05, 'epoch': 0.03}) +[2026-06-24 12:05:42,400][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.24355730414390564, 'train/grad_norm_pre_clip': 49.00358200073242, 'timing/data': 0.00029859994538128376, 'timing/model': 1.0633519790135324, 'learning_rate/qwen_vl_interface': 9.945056196960858e-06, 'learning_rate/action_model': 9.945056196960858e-05, 'epoch': 0.03}) +[2026-06-24 12:05:43,466][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.28766104578971863, 'train/grad_norm_pre_clip': 85.0761489868164, 'timing/data': 0.0003017759881913662, 'timing/model': 1.063724358100444, 'learning_rate/qwen_vl_interface': 9.944474381318871e-06, 'learning_rate/action_model': 9.94447438131887e-05, 'epoch': 0.03}) +[2026-06-24 12:05:44,528][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.521825909614563, 'train/grad_norm_pre_clip': 149.1311492919922, 'timing/data': 0.00032098195515573025, 'timing/model': 1.0610690428875387, 'learning_rate/qwen_vl_interface': 9.943889519483424e-06, 'learning_rate/action_model': 9.943889519483424e-05, 'epoch': 0.03}) +[2026-06-24 12:05:45,594][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.30430901050567627, 'train/grad_norm_pre_clip': 60.76222229003906, 'timing/data': 0.00030714506283402443, 'timing/model': 1.0642419790383428, 'learning_rate/qwen_vl_interface': 9.94330161183403e-06, 'learning_rate/action_model': 9.94330161183403e-05, 'epoch': 0.03}) +[2026-06-24 12:05:46,658][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.09194678068161011, 'train/grad_norm_pre_clip': 18.986278533935547, 'timing/data': 0.0003182960208505392, 'timing/model': 1.063047030940652, 'learning_rate/qwen_vl_interface': 9.942710658752175e-06, 'learning_rate/action_model': 9.942710658752174e-05, 'epoch': 0.03}) +[2026-06-24 12:05:47,726][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.2376844882965088, 'train/grad_norm_pre_clip': 24.20185661315918, 'timing/data': 0.0002919470425695181, 'timing/model': 1.0658776061609387, 'learning_rate/qwen_vl_interface': 9.942116660621319e-06, 'learning_rate/action_model': 9.94211666062132e-05, 'epoch': 0.03}) +[2026-06-24 12:05:48,791][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.5040177702903748, 'train/grad_norm_pre_clip': 101.25259399414062, 'timing/data': 0.00030589313246309757, 'timing/model': 1.0637805671431124, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.03}) +[2026-06-24 12:05:49,858][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.4417106509208679, 'train/grad_norm_pre_clip': 71.65660858154297, 'timing/data': 0.00034838286228477955, 'timing/model': 1.064794085919857, 'learning_rate/qwen_vl_interface': 9.940919530756347e-06, 'learning_rate/action_model': 9.940919530756347e-05, 'epoch': 0.03}) +[2026-06-24 12:05:50,931][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.24461983144283295, 'train/grad_norm_pre_clip': 66.50591278076172, 'timing/data': 0.0003072461113333702, 'timing/model': 1.0713490301277488, 'learning_rate/qwen_vl_interface': 9.940316399799032e-06, 'learning_rate/action_model': 9.940316399799033e-05, 'epoch': 0.03}) +[2026-06-24 12:05:52,019][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.20686037838459015, 'train/grad_norm_pre_clip': 53.507408142089844, 'timing/data': 0.0004342030733823776, 'timing/model': 1.0864243779797107, 'learning_rate/qwen_vl_interface': 9.939710225346329e-06, 'learning_rate/action_model': 9.939710225346329e-05, 'epoch': 0.03}) +[2026-06-24 12:05:53,323][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.2536970376968384, 'train/grad_norm_pre_clip': 56.82725524902344, 'timing/data': 0.00043833116069436073, 'timing/model': 1.3022760201711208, 'learning_rate/qwen_vl_interface': 9.939101007791575e-06, 'learning_rate/action_model': 9.939101007791575e-05, 'epoch': 0.03}) +[2026-06-24 12:05:54,388][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.37825143337249756, 'train/grad_norm_pre_clip': 228.24893188476562, 'timing/data': 0.00034434511326253414, 'timing/model': 1.0632269440684468, 'learning_rate/qwen_vl_interface': 9.938488747530087e-06, 'learning_rate/action_model': 9.938488747530086e-05, 'epoch': 0.03}) +[2026-06-24 12:05:55,455][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.49338498711586, 'train/grad_norm_pre_clip': 184.1640167236328, 'timing/data': 0.0003397860564291477, 'timing/model': 1.0654005440883338, 'learning_rate/qwen_vl_interface': 9.937873444959151e-06, 'learning_rate/action_model': 9.937873444959151e-05, 'epoch': 0.03}) +[2026-06-24 12:05:56,807][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.2495892345905304, 'train/grad_norm_pre_clip': 65.05897521972656, 'timing/data': 0.0003151900600641966, 'timing/model': 1.3503092979080975, 'learning_rate/qwen_vl_interface': 9.937255100478036e-06, 'learning_rate/action_model': 9.937255100478036e-05, 'epoch': 0.03}) +[2026-06-24 12:05:57,873][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.2161318063735962, 'train/grad_norm_pre_clip': 51.317726135253906, 'timing/data': 0.00034730089828372, 'timing/model': 1.0643275380134583, 'learning_rate/qwen_vl_interface': 9.93663371448797e-06, 'learning_rate/action_model': 9.93663371448797e-05, 'epoch': 0.03}) +[2026-06-24 12:05:58,938][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.1983914077281952, 'train/grad_norm_pre_clip': 42.63901138305664, 'timing/data': 0.00029255799017846584, 'timing/model': 1.0641059230547398, 'learning_rate/qwen_vl_interface': 9.936009287392173e-06, 'learning_rate/action_model': 9.936009287392171e-05, 'epoch': 0.03}) +[2026-06-24 12:06:00,006][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.46228963136672974, 'train/grad_norm_pre_clip': 83.67425537109375, 'timing/data': 0.0003375830128788948, 'timing/model': 1.0665213000029325, 'learning_rate/qwen_vl_interface': 9.935381819595822e-06, 'learning_rate/action_model': 9.935381819595822e-05, 'epoch': 0.03}) +[2026-06-24 12:06:01,075][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.4783848226070404, 'train/grad_norm_pre_clip': 76.5169448852539, 'timing/data': 0.00030697486363351345, 'timing/model': 1.0672484929673374, 'learning_rate/qwen_vl_interface': 9.934751311506078e-06, 'learning_rate/action_model': 9.934751311506077e-05, 'epoch': 0.03}) +[2026-06-24 12:06:02,141][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.885028600692749, 'train/grad_norm_pre_clip': 124.40052795410156, 'timing/data': 0.0005187720526009798, 'timing/model': 1.0632390708196908, 'learning_rate/qwen_vl_interface': 9.93411776353207e-06, 'learning_rate/action_model': 9.934117763532069e-05, 'epoch': 0.03}) +[2026-06-24 12:06:03,209][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.2896237373352051, 'train/grad_norm_pre_clip': 34.385128021240234, 'timing/data': 0.0002918671816587448, 'timing/model': 1.066347295884043, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.03}) +[2026-06-24 12:06:04,276][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.25868338346481323, 'train/grad_norm_pre_clip': 32.58661651611328, 'timing/data': 0.00029742694459855556, 'timing/model': 1.0655203969217837, 'learning_rate/qwen_vl_interface': 9.932841549577642e-06, 'learning_rate/action_model': 9.932841549577641e-05, 'epoch': 0.03}) +[2026-06-24 12:06:05,340][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.33662018179893494, 'train/grad_norm_pre_clip': 23.183185577392578, 'timing/data': 0.0003054630942642689, 'timing/model': 1.0625766720622778, 'learning_rate/qwen_vl_interface': 9.932198884425343e-06, 'learning_rate/action_model': 9.932198884425342e-05, 'epoch': 0.03}) +[2026-06-24 12:06:06,411][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.344905823469162, 'train/grad_norm_pre_clip': 58.06234359741211, 'timing/data': 0.0003093488048762083, 'timing/model': 1.0694902779068798, 'learning_rate/qwen_vl_interface': 9.931553181045022e-06, 'learning_rate/action_model': 9.931553181045022e-05, 'epoch': 0.03}) +[2026-06-24 12:06:07,479][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.4564608335494995, 'train/grad_norm_pre_clip': 50.99711608886719, 'timing/data': 0.0003112030681222677, 'timing/model': 1.0666275839321315, 'learning_rate/qwen_vl_interface': 9.930904439855667e-06, 'learning_rate/action_model': 9.930904439855667e-05, 'epoch': 0.03}) +[2026-06-24 12:06:08,546][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.3609887361526489, 'train/grad_norm_pre_clip': 69.9277114868164, 'timing/data': 0.0003104819916188717, 'timing/model': 1.0654565109871328, 'learning_rate/qwen_vl_interface': 9.930252661278245e-06, 'learning_rate/action_model': 9.930252661278244e-05, 'epoch': 0.03}) +[2026-06-24 12:06:09,614][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.2000306248664856, 'train/grad_norm_pre_clip': 25.354671478271484, 'timing/data': 0.0003046109341084957, 'timing/model': 1.0659832882229239, 'learning_rate/qwen_vl_interface': 9.92959784573568e-06, 'learning_rate/action_model': 9.92959784573568e-05, 'epoch': 0.03}) +[2026-06-24 12:06:10,683][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.21915747225284576, 'train/grad_norm_pre_clip': 12.513042449951172, 'timing/data': 0.0003285959828644991, 'timing/model': 1.067817596020177, 'learning_rate/qwen_vl_interface': 9.92893999365288e-06, 'learning_rate/action_model': 9.92893999365288e-05, 'epoch': 0.03}) +[2026-06-24 12:06:12,045][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.3259212374687195, 'train/grad_norm_pre_clip': 27.174049377441406, 'timing/data': 0.00035720993764698505, 'timing/model': 1.3603363789152354, 'learning_rate/qwen_vl_interface': 9.928279105456717e-06, 'learning_rate/action_model': 9.928279105456718e-05, 'epoch': 0.03}) +[2026-06-24 12:06:13,111][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.5203698873519897, 'train/grad_norm_pre_clip': 52.674259185791016, 'timing/data': 0.0003067550715059042, 'timing/model': 1.0645066839642823, 'learning_rate/qwen_vl_interface': 9.927615181576034e-06, 'learning_rate/action_model': 9.927615181576034e-05, 'epoch': 0.03}) +[2026-06-24 12:06:14,177][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.480234295129776, 'train/grad_norm_pre_clip': 90.96675872802734, 'timing/data': 0.00030086305923759937, 'timing/model': 1.0644791321828961, 'learning_rate/qwen_vl_interface': 9.926948222441644e-06, 'learning_rate/action_model': 9.926948222441643e-05, 'epoch': 0.03}) +[2026-06-24 12:06:15,246][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.24003559350967407, 'train/grad_norm_pre_clip': 36.18852996826172, 'timing/data': 0.0002928080502897501, 'timing/model': 1.0678398010786623, 'learning_rate/qwen_vl_interface': 9.926278228486328e-06, 'learning_rate/action_model': 9.926278228486327e-05, 'epoch': 0.03}) +[2026-06-24 12:06:16,316][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.3119572699069977, 'train/grad_norm_pre_clip': 48.84543228149414, 'timing/data': 0.0005128509365022182, 'timing/model': 1.0679202219471335, 'learning_rate/qwen_vl_interface': 9.92560520014484e-06, 'learning_rate/action_model': 9.92560520014484e-05, 'epoch': 0.03}) +[2026-06-24 12:06:17,385][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.3464902937412262, 'train/grad_norm_pre_clip': 137.1424102783203, 'timing/data': 0.00030135503038764, 'timing/model': 1.0669149491004646, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.03}) +[2026-06-24 12:06:18,456][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.7206791639328003, 'train/grad_norm_pre_clip': 199.51101684570312, 'timing/data': 0.0003095800057053566, 'timing/model': 1.0700357370078564, 'learning_rate/qwen_vl_interface': 9.924250042052197e-06, 'learning_rate/action_model': 9.924250042052197e-05, 'epoch': 0.03}) +[2026-06-24 12:06:19,526][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.4173867404460907, 'train/grad_norm_pre_clip': 117.75408935546875, 'timing/data': 0.0003014649264514446, 'timing/model': 1.0684153919573873, 'learning_rate/qwen_vl_interface': 9.92356791318039e-06, 'learning_rate/action_model': 9.92356791318039e-05, 'epoch': 0.03}) +[2026-06-24 12:06:20,594][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.45288339257240295, 'train/grad_norm_pre_clip': 89.27588653564453, 'timing/data': 0.00029438198544085026, 'timing/model': 1.0661302809603512, 'learning_rate/qwen_vl_interface': 9.922882751681104e-06, 'learning_rate/action_model': 9.922882751681104e-05, 'epoch': 0.03}) +[2026-06-24 12:06:21,659][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.14454449713230133, 'train/grad_norm_pre_clip': 16.445703506469727, 'timing/data': 0.00030428008176386356, 'timing/model': 1.0637103870976716, 'learning_rate/qwen_vl_interface': 9.922194557998932e-06, 'learning_rate/action_model': 9.922194557998933e-05, 'epoch': 0.03}) +[2026-06-24 12:06:22,729][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.23816120624542236, 'train/grad_norm_pre_clip': 37.154911041259766, 'timing/data': 0.00030814693309366703, 'timing/model': 1.0679685960058123, 'learning_rate/qwen_vl_interface': 9.921503332580439e-06, 'learning_rate/action_model': 9.921503332580439e-05, 'epoch': 0.03}) +[2026-06-24 12:06:23,797][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.3924829959869385, 'train/grad_norm_pre_clip': 48.16096878051758, 'timing/data': 0.0002960141282528639, 'timing/model': 1.0662088789977133, 'learning_rate/qwen_vl_interface': 9.92080907587415e-06, 'learning_rate/action_model': 9.92080907587415e-05, 'epoch': 0.03}) +[2026-06-24 12:06:24,864][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.1797083169221878, 'train/grad_norm_pre_clip': 27.84994125366211, 'timing/data': 0.0003383541479706764, 'timing/model': 1.0651635818649083, 'learning_rate/qwen_vl_interface': 9.92011178833056e-06, 'learning_rate/action_model': 9.92011178833056e-05, 'epoch': 0.03}) +[2026-06-24 12:06:25,935][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.2025020271539688, 'train/grad_norm_pre_clip': 27.688871383666992, 'timing/data': 0.000304230023175478, 'timing/model': 1.0698325911071151, 'learning_rate/qwen_vl_interface': 9.919411470402136e-06, 'learning_rate/action_model': 9.919411470402136e-05, 'epoch': 0.03}) +[2026-06-24 12:06:27,305][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.28628185391426086, 'train/grad_norm_pre_clip': 31.558692932128906, 'timing/data': 0.00029570399783551693, 'timing/model': 1.3686063471250236, 'learning_rate/qwen_vl_interface': 9.918708122543305e-06, 'learning_rate/action_model': 9.918708122543305e-05, 'epoch': 0.03}) +[2026-06-24 12:06:28,374][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.2251937836408615, 'train/grad_norm_pre_clip': 32.818824768066406, 'timing/data': 0.0003296982031315565, 'timing/model': 1.0671551558189094, 'learning_rate/qwen_vl_interface': 9.918001745210458e-06, 'learning_rate/action_model': 9.918001745210456e-05, 'epoch': 0.03}) +[2026-06-24 12:06:29,441][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.3282933235168457, 'train/grad_norm_pre_clip': 49.33900833129883, 'timing/data': 0.00031543103978037834, 'timing/model': 1.0660671980585903, 'learning_rate/qwen_vl_interface': 9.917292338861958e-06, 'learning_rate/action_model': 9.917292338861957e-05, 'epoch': 0.03}) +[2026-06-24 12:06:30,509][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.14694853127002716, 'train/grad_norm_pre_clip': 37.60805130004883, 'timing/data': 0.0005315961316227913, 'timing/model': 1.065171039197594, 'learning_rate/qwen_vl_interface': 9.916579903958134e-06, 'learning_rate/action_model': 9.916579903958133e-05, 'epoch': 0.03}) +[2026-06-24 12:06:31,577][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.10601232945919037, 'train/grad_norm_pre_clip': 29.597591400146484, 'timing/data': 0.0003040099982172251, 'timing/model': 1.066689636092633, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.03}) +[2026-06-24 12:06:32,647][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.19923560321331024, 'train/grad_norm_pre_clip': 44.589725494384766, 'timing/data': 0.00030097411945462227, 'timing/model': 1.0682100150734186, 'learning_rate/qwen_vl_interface': 9.915145950335631e-06, 'learning_rate/action_model': 9.915145950335632e-05, 'epoch': 0.03}) +[2026-06-24 12:06:33,714][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.1728639453649521, 'train/grad_norm_pre_clip': 54.24375534057617, 'timing/data': 0.0003176550380885601, 'timing/model': 1.0653078581672162, 'learning_rate/qwen_vl_interface': 9.914424432547434e-06, 'learning_rate/action_model': 9.914424432547433e-05, 'epoch': 0.03}) +[2026-06-24 12:06:34,785][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.1458796113729477, 'train/grad_norm_pre_clip': 55.0387077331543, 'timing/data': 0.00030153500847518444, 'timing/model': 1.0700809119734913, 'learning_rate/qwen_vl_interface': 9.913699888064862e-06, 'learning_rate/action_model': 9.913699888064861e-05, 'epoch': 0.03}) +[2026-06-24 12:06:35,861][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.32395288348197937, 'train/grad_norm_pre_clip': 70.26911163330078, 'timing/data': 0.00030393898487091064, 'timing/model': 1.0738067540805787, 'learning_rate/qwen_vl_interface': 9.912972317358066e-06, 'learning_rate/action_model': 9.912972317358066e-05, 'epoch': 0.03}) +[2026-06-24 12:06:36,929][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.1360182762145996, 'train/grad_norm_pre_clip': 37.06922149658203, 'timing/data': 0.0003086179494857788, 'timing/model': 1.0665304409340024, 'learning_rate/qwen_vl_interface': 9.912241720899157e-06, 'learning_rate/action_model': 9.912241720899157e-05, 'epoch': 0.03}) +[2026-06-24 12:06:37,995][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.5251325368881226, 'train/grad_norm_pre_clip': 69.80563354492188, 'timing/data': 0.0003116040024906397, 'timing/model': 1.0646814368665218, 'learning_rate/qwen_vl_interface': 9.911508099162215e-06, 'learning_rate/action_model': 9.911508099162214e-05, 'epoch': 0.03}) +[2026-06-24 12:06:39,068][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.20425692200660706, 'train/grad_norm_pre_clip': 48.91870880126953, 'timing/data': 0.0003309999592602253, 'timing/model': 1.0715233169030398, 'learning_rate/qwen_vl_interface': 9.910771452623276e-06, 'learning_rate/action_model': 9.910771452623275e-05, 'epoch': 0.03}) +[2026-06-24 12:06:40,137][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.3949478268623352, 'train/grad_norm_pre_clip': 78.358642578125, 'timing/data': 0.00029096496291458607, 'timing/model': 1.0676379781216383, 'learning_rate/qwen_vl_interface': 9.910031781760342e-06, 'learning_rate/action_model': 9.910031781760342e-05, 'epoch': 0.03}) +[2026-06-24 12:06:41,206][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.25228747725486755, 'train/grad_norm_pre_clip': 40.290008544921875, 'timing/data': 0.0003098610322922468, 'timing/model': 1.067199997836724, 'learning_rate/qwen_vl_interface': 9.909289087053378e-06, 'learning_rate/action_model': 9.909289087053378e-05, 'epoch': 0.03}) +[2026-06-24 12:06:42,564][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.15845666825771332, 'train/grad_norm_pre_clip': 44.21921920776367, 'timing/data': 0.00033514807000756264, 'timing/model': 1.356070469133556, 'learning_rate/qwen_vl_interface': 9.90854336898431e-06, 'learning_rate/action_model': 9.90854336898431e-05, 'epoch': 0.03}) +[2026-06-24 12:06:43,633][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.1777273565530777, 'train/grad_norm_pre_clip': 29.23564910888672, 'timing/data': 0.0003059930168092251, 'timing/model': 1.0676052288617939, 'learning_rate/qwen_vl_interface': 9.907794628037028e-06, 'learning_rate/action_model': 9.907794628037028e-05, 'epoch': 0.03}) +[2026-06-24 12:06:44,704][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.11060696095228195, 'train/grad_norm_pre_clip': 20.101560592651367, 'timing/data': 0.0003165330272167921, 'timing/model': 1.0698874059598893, 'learning_rate/qwen_vl_interface': 9.90704286469738e-06, 'learning_rate/action_model': 9.90704286469738e-05, 'epoch': 0.03}) +[2026-06-24 12:06:45,777][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.4874247610569, 'train/grad_norm_pre_clip': 114.18522644042969, 'timing/data': 0.0003206010442227125, 'timing/model': 1.0711383200250566, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.03}) +[2026-06-24 12:06:46,845][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.4710976481437683, 'train/grad_norm_pre_clip': 88.21491241455078, 'timing/data': 0.0003193679731339216, 'timing/model': 1.066380085190758, 'learning_rate/qwen_vl_interface': 9.905530272794195e-06, 'learning_rate/action_model': 9.905530272794195e-05, 'epoch': 0.03}) +[2026-06-24 12:06:47,916][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.16538631916046143, 'train/grad_norm_pre_clip': 33.7460823059082, 'timing/data': 0.0003151900600641966, 'timing/model': 1.0692702918313444, 'learning_rate/qwen_vl_interface': 9.904769445212163e-06, 'learning_rate/action_model': 9.904769445212162e-05, 'epoch': 0.03}) +[2026-06-24 12:06:48,985][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.13562582433223724, 'train/grad_norm_pre_clip': 18.895214080810547, 'timing/data': 0.000295763835310936, 'timing/model': 1.067616533022374, 'learning_rate/qwen_vl_interface': 9.904005597200772e-06, 'learning_rate/action_model': 9.904005597200772e-05, 'epoch': 0.03}) +[2026-06-24 12:06:50,064][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.1425120234489441, 'train/grad_norm_pre_clip': 37.817527770996094, 'timing/data': 0.00032414705492556095, 'timing/model': 1.0769378498662263, 'learning_rate/qwen_vl_interface': 9.903238729255678e-06, 'learning_rate/action_model': 9.903238729255677e-05, 'epoch': 0.03}) +[2026-06-24 12:06:51,136][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.391735702753067, 'train/grad_norm_pre_clip': 40.24717712402344, 'timing/data': 0.00036951201036572456, 'timing/model': 1.0708039470482618, 'learning_rate/qwen_vl_interface': 9.902468841874492e-06, 'learning_rate/action_model': 9.902468841874492e-05, 'epoch': 0.03}) +[2026-06-24 12:06:52,211][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.2982442378997803, 'train/grad_norm_pre_clip': 75.65157318115234, 'timing/data': 0.00033695087768137455, 'timing/model': 1.073096552863717, 'learning_rate/qwen_vl_interface': 9.901695935556784e-06, 'learning_rate/action_model': 9.901695935556783e-05, 'epoch': 0.03}) +[2026-06-24 12:06:53,281][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.2191680669784546, 'train/grad_norm_pre_clip': 71.90642547607422, 'timing/data': 0.0003030281513929367, 'timing/model': 1.0682730961125344, 'learning_rate/qwen_vl_interface': 9.900920010804088e-06, 'learning_rate/action_model': 9.900920010804088e-05, 'epoch': 0.03}) +[2026-06-24 12:06:54,353][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.27534911036491394, 'train/grad_norm_pre_clip': 46.55987548828125, 'timing/data': 0.0003047108184546232, 'timing/model': 1.0704248689580709, 'learning_rate/qwen_vl_interface': 9.900141068119891e-06, 'learning_rate/action_model': 9.900141068119891e-05, 'epoch': 0.03}) +[2026-06-24 12:06:55,423][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.2720630466938019, 'train/grad_norm_pre_clip': 68.2015151977539, 'timing/data': 0.0003020560834556818, 'timing/model': 1.0684613501653075, 'learning_rate/qwen_vl_interface': 9.89935910800964e-06, 'learning_rate/action_model': 9.89935910800964e-05, 'epoch': 0.03}) +[2026-06-24 12:06:56,494][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.33406010270118713, 'train/grad_norm_pre_clip': 51.20254898071289, 'timing/data': 0.000306474044919014, 'timing/model': 1.0700097819790244, 'learning_rate/qwen_vl_interface': 9.898574130980743e-06, 'learning_rate/action_model': 9.898574130980742e-05, 'epoch': 0.03}) +[2026-06-24 12:06:57,850][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.3700498640537262, 'train/grad_norm_pre_clip': 74.20782470703125, 'timing/data': 0.0003733290359377861, 'timing/model': 1.3537420271895826, 'learning_rate/qwen_vl_interface': 9.89778613754256e-06, 'learning_rate/action_model': 9.89778613754256e-05, 'epoch': 0.03}) +[2026-06-24 12:06:58,923][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.15945583581924438, 'train/grad_norm_pre_clip': 44.115386962890625, 'timing/data': 0.0003424310125410557, 'timing/model': 1.0712342879269272, 'learning_rate/qwen_vl_interface': 9.896995128206412e-06, 'learning_rate/action_model': 9.896995128206411e-05, 'epoch': 0.03}) +[2026-06-24 12:06:59,995][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.32874974608421326, 'train/grad_norm_pre_clip': 43.49995040893555, 'timing/data': 0.0003179549239575863, 'timing/model': 1.0705432330723852, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.03}) +[2026-06-24 12:07:01,067][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.36203330755233765, 'train/grad_norm_pre_clip': 73.73656463623047, 'timing/data': 0.0003064139746129513, 'timing/model': 1.0710353571921587, 'learning_rate/qwen_vl_interface': 9.89540406389529e-06, 'learning_rate/action_model': 9.89540406389529e-05, 'epoch': 0.03}) +[2026-06-24 12:07:02,139][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.297309935092926, 'train/grad_norm_pre_clip': 67.2193832397461, 'timing/data': 0.00030788686126470566, 'timing/model': 1.070480358088389, 'learning_rate/qwen_vl_interface': 9.894604009952742e-06, 'learning_rate/action_model': 9.89460400995274e-05, 'epoch': 0.04}) +[2026-06-24 12:07:03,208][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.29520782828330994, 'train/grad_norm_pre_clip': 77.71215057373047, 'timing/data': 0.0003040698356926441, 'timing/model': 1.0674185808748007, 'learning_rate/qwen_vl_interface': 9.893800942177073e-06, 'learning_rate/action_model': 9.893800942177073e-05, 'epoch': 0.04}) +[2026-06-24 12:07:04,281][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.24045099318027496, 'train/grad_norm_pre_clip': 48.53274154663086, 'timing/data': 0.0004971211310476065, 'timing/model': 1.0702908330131322, 'learning_rate/qwen_vl_interface': 9.892994861089393e-06, 'learning_rate/action_model': 9.892994861089393e-05, 'epoch': 0.04}) +[2026-06-24 12:07:05,354][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.3492327332496643, 'train/grad_norm_pre_clip': 73.3169174194336, 'timing/data': 0.0002925880253314972, 'timing/model': 1.0718228549230844, 'learning_rate/qwen_vl_interface': 9.892185767212756e-06, 'learning_rate/action_model': 9.892185767212756e-05, 'epoch': 0.04}) +[2026-06-24 12:07:06,424][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.8349153995513916, 'train/grad_norm_pre_clip': 125.35993194580078, 'timing/data': 0.00029915012419223785, 'timing/model': 1.0688631490338594, 'learning_rate/qwen_vl_interface': 9.891373661072172e-06, 'learning_rate/action_model': 9.891373661072173e-05, 'epoch': 0.04}) +[2026-06-24 12:07:07,496][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.1646505445241928, 'train/grad_norm_pre_clip': 41.59346389770508, 'timing/data': 0.0006610390264540911, 'timing/model': 1.0699144289828837, 'learning_rate/qwen_vl_interface': 9.89055854319461e-06, 'learning_rate/action_model': 9.890558543194611e-05, 'epoch': 0.04}) +[2026-06-24 12:07:08,572][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.34724605083465576, 'train/grad_norm_pre_clip': 47.3205680847168, 'timing/data': 0.0005131619982421398, 'timing/model': 1.0733511208090931, 'learning_rate/qwen_vl_interface': 9.889740414108994e-06, 'learning_rate/action_model': 9.889740414108995e-05, 'epoch': 0.04}) +[2026-06-24 12:07:09,647][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.25454917550086975, 'train/grad_norm_pre_clip': 60.64826202392578, 'timing/data': 0.0003933270927518606, 'timing/model': 1.073270620079711, 'learning_rate/qwen_vl_interface': 9.888919274346196e-06, 'learning_rate/action_model': 9.888919274346195e-05, 'epoch': 0.04}) +[2026-06-24 12:07:10,718][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.23705729842185974, 'train/grad_norm_pre_clip': 44.884063720703125, 'timing/data': 0.000326341949403286, 'timing/model': 1.0696252300404012, 'learning_rate/qwen_vl_interface': 9.888095124439043e-06, 'learning_rate/action_model': 9.888095124439043e-05, 'epoch': 0.04}) +[2026-06-24 12:07:11,788][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.22643887996673584, 'train/grad_norm_pre_clip': 29.286148071289062, 'timing/data': 0.0003363110590726137, 'timing/model': 1.068803641013801, 'learning_rate/qwen_vl_interface': 9.88726796492232e-06, 'learning_rate/action_model': 9.88726796492232e-05, 'epoch': 0.04}) +[2026-06-24 12:07:13,142][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.2078869640827179, 'train/grad_norm_pre_clip': 35.15410614013672, 'timing/data': 0.0002910757903009653, 'timing/model': 1.3524933699518442, 'learning_rate/qwen_vl_interface': 9.886437796332759e-06, 'learning_rate/action_model': 9.88643779633276e-05, 'epoch': 0.04}) +[2026-06-24 12:07:14,216][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.1567108929157257, 'train/grad_norm_pre_clip': 33.27595520019531, 'timing/data': 0.0003113429993391037, 'timing/model': 1.0722134909592569, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.04}) +[2026-06-24 12:07:15,286][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.4209290146827698, 'train/grad_norm_pre_clip': 81.56935119628906, 'timing/data': 0.00030624307692050934, 'timing/model': 1.068693745881319, 'learning_rate/qwen_vl_interface': 9.884768434091832e-06, 'learning_rate/action_model': 9.884768434091832e-05, 'epoch': 0.04}) +[2026-06-24 12:07:16,359][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.42831945419311523, 'train/grad_norm_pre_clip': 48.18362808227539, 'timing/data': 0.00029856013134121895, 'timing/model': 1.0713457770179957, 'learning_rate/qwen_vl_interface': 9.883929241523698e-06, 'learning_rate/action_model': 9.883929241523697e-05, 'epoch': 0.04}) +[2026-06-24 12:07:17,430][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.16103489696979523, 'train/grad_norm_pre_clip': 38.7380256652832, 'timing/data': 0.0005003081168979406, 'timing/model': 1.0690589528530836, 'learning_rate/qwen_vl_interface': 9.883087042049185e-06, 'learning_rate/action_model': 9.883087042049186e-05, 'epoch': 0.04}) +[2026-06-24 12:07:18,503][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.18565653264522552, 'train/grad_norm_pre_clip': 44.819087982177734, 'timing/data': 0.0003191379364579916, 'timing/model': 1.071269074920565, 'learning_rate/qwen_vl_interface': 9.882241836214794e-06, 'learning_rate/action_model': 9.882241836214794e-05, 'epoch': 0.04}) +[2026-06-24 12:07:19,573][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.0657048299908638, 'train/grad_norm_pre_clip': 14.426166534423828, 'timing/data': 0.0003306199796497822, 'timing/model': 1.0682029698509723, 'learning_rate/qwen_vl_interface': 9.881393624568965e-06, 'learning_rate/action_model': 9.881393624568965e-05, 'epoch': 0.04}) +[2026-06-24 12:07:20,643][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.08576726168394089, 'train/grad_norm_pre_clip': 57.52260208129883, 'timing/data': 0.00029710610397160053, 'timing/model': 1.0684829649981111, 'learning_rate/qwen_vl_interface': 9.880542407662098e-06, 'learning_rate/action_model': 9.880542407662097e-05, 'epoch': 0.04}) +[2026-06-24 12:07:21,713][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.29582130908966064, 'train/grad_norm_pre_clip': 157.32862854003906, 'timing/data': 0.00029369001276791096, 'timing/model': 1.068550683092326, 'learning_rate/qwen_vl_interface': 9.879688186046535e-06, 'learning_rate/action_model': 9.879688186046534e-05, 'epoch': 0.04}) +[2026-06-24 12:07:22,781][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.43142789602279663, 'train/grad_norm_pre_clip': 163.7509002685547, 'timing/data': 0.00034877401776611805, 'timing/model': 1.0660547360312194, 'learning_rate/qwen_vl_interface': 9.87883096027657e-06, 'learning_rate/action_model': 9.87883096027657e-05, 'epoch': 0.04}) +[2026-06-24 12:07:23,850][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.8749850392341614, 'train/grad_norm_pre_clip': 215.2623291015625, 'timing/data': 0.0003076160792261362, 'timing/model': 1.0677262290846556, 'learning_rate/qwen_vl_interface': 9.87797073090845e-06, 'learning_rate/action_model': 9.87797073090845e-05, 'epoch': 0.04}) +[2026-06-24 12:07:24,921][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.17528486251831055, 'train/grad_norm_pre_clip': 87.71400451660156, 'timing/data': 0.0003407790791243315, 'timing/model': 1.0691442270763218, 'learning_rate/qwen_vl_interface': 9.877107498500368e-06, 'learning_rate/action_model': 9.877107498500367e-05, 'epoch': 0.04}) +[2026-06-24 12:07:25,990][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.4484628736972809, 'train/grad_norm_pre_clip': 125.48250579833984, 'timing/data': 0.0003514089621603489, 'timing/model': 1.0675227791070938, 'learning_rate/qwen_vl_interface': 9.876241263612467e-06, 'learning_rate/action_model': 9.876241263612465e-05, 'epoch': 0.04}) +[2026-06-24 12:07:27,062][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.6940901279449463, 'train/grad_norm_pre_clip': 88.55223083496094, 'timing/data': 0.00030319811776280403, 'timing/model': 1.0702128501143306, 'learning_rate/qwen_vl_interface': 9.875372026806835e-06, 'learning_rate/action_model': 9.875372026806835e-05, 'epoch': 0.04}) +[2026-06-24 12:07:28,418][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.28298965096473694, 'train/grad_norm_pre_clip': 74.81288146972656, 'timing/data': 0.00031099305488169193, 'timing/model': 1.354884881991893, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.04}) +[2026-06-24 12:07:29,489][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.31418630480766296, 'train/grad_norm_pre_clip': 50.514198303222656, 'timing/data': 0.0003416698891669512, 'timing/model': 1.069285583915189, 'learning_rate/qwen_vl_interface': 9.873624549700485e-06, 'learning_rate/action_model': 9.873624549700484e-05, 'epoch': 0.04}) +[2026-06-24 12:07:30,562][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.22730474174022675, 'train/grad_norm_pre_clip': 51.17238235473633, 'timing/data': 0.000306163914501667, 'timing/model': 1.071027077967301, 'learning_rate/qwen_vl_interface': 9.872746310533683e-06, 'learning_rate/action_model': 9.872746310533684e-05, 'epoch': 0.04}) +[2026-06-24 12:07:31,632][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.4064069986343384, 'train/grad_norm_pre_clip': 40.144657135009766, 'timing/data': 0.0003127960953861475, 'timing/model': 1.0686257588677108, 'learning_rate/qwen_vl_interface': 9.871865071716991e-06, 'learning_rate/action_model': 9.87186507171699e-05, 'epoch': 0.04}) +[2026-06-24 12:07:32,703][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.1661219447851181, 'train/grad_norm_pre_clip': 22.5382080078125, 'timing/data': 0.0003022560849785805, 'timing/model': 1.069032431114465, 'learning_rate/qwen_vl_interface': 9.870980833822231e-06, 'learning_rate/action_model': 9.87098083382223e-05, 'epoch': 0.04}) +[2026-06-24 12:07:33,775][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.2287711650133133, 'train/grad_norm_pre_clip': 25.37110710144043, 'timing/data': 0.00035293190740048885, 'timing/model': 1.070886867120862, 'learning_rate/qwen_vl_interface': 9.87009359742318e-06, 'learning_rate/action_model': 9.87009359742318e-05, 'epoch': 0.04}) +[2026-06-24 12:07:34,847][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.233594611287117, 'train/grad_norm_pre_clip': 30.816102981567383, 'timing/data': 0.0002965449821203947, 'timing/model': 1.0705973859876394, 'learning_rate/qwen_vl_interface': 9.86920336309555e-06, 'learning_rate/action_model': 9.86920336309555e-05, 'epoch': 0.04}) +[2026-06-24 12:07:35,919][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.1971251666545868, 'train/grad_norm_pre_clip': 16.27005386352539, 'timing/data': 0.00030708499252796173, 'timing/model': 1.0702503961510956, 'learning_rate/qwen_vl_interface': 9.86831013141701e-06, 'learning_rate/action_model': 9.86831013141701e-05, 'epoch': 0.04}) +[2026-06-24 12:07:36,990][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.23921628296375275, 'train/grad_norm_pre_clip': 23.504121780395508, 'timing/data': 0.00034968508407473564, 'timing/model': 1.0693741939030588, 'learning_rate/qwen_vl_interface': 9.867413902967166e-06, 'learning_rate/action_model': 9.867413902967166e-05, 'epoch': 0.04}) +[2026-06-24 12:07:38,061][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.13030275702476501, 'train/grad_norm_pre_clip': 16.148059844970703, 'timing/data': 0.0003266220446676016, 'timing/model': 1.0697887821588665, 'learning_rate/qwen_vl_interface': 9.86651467832757e-06, 'learning_rate/action_model': 9.86651467832757e-05, 'epoch': 0.04}) +[2026-06-24 12:07:39,134][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.38250795006752014, 'train/grad_norm_pre_clip': 39.01053237915039, 'timing/data': 0.0003307189326733351, 'timing/model': 1.070706671802327, 'learning_rate/qwen_vl_interface': 9.865612458081718e-06, 'learning_rate/action_model': 9.865612458081719e-05, 'epoch': 0.04}) +[2026-06-24 12:07:40,206][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.13402704894542694, 'train/grad_norm_pre_clip': 25.49736785888672, 'timing/data': 0.0003623887896537781, 'timing/model': 1.0711015618871897, 'learning_rate/qwen_vl_interface': 9.864707242815057e-06, 'learning_rate/action_model': 9.864707242815056e-05, 'epoch': 0.04}) +[2026-06-24 12:07:41,278][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.08431810885667801, 'train/grad_norm_pre_clip': 17.43992805480957, 'timing/data': 0.0003218129277229309, 'timing/model': 1.0700488409493119, 'learning_rate/qwen_vl_interface': 9.863799033114966e-06, 'learning_rate/action_model': 9.863799033114966e-05, 'epoch': 0.04}) +[2026-06-24 12:07:42,347][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.39536407589912415, 'train/grad_norm_pre_clip': 167.18765258789062, 'timing/data': 0.0003548648674041033, 'timing/model': 1.067433639895171, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.04}) +[2026-06-24 12:07:43,709][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.26323631405830383, 'train/grad_norm_pre_clip': 74.03966522216797, 'timing/data': 0.00030679511837661266, 'timing/model': 1.3609063760377467, 'learning_rate/qwen_vl_interface': 9.861973632773748e-06, 'learning_rate/action_model': 9.861973632773748e-05, 'epoch': 0.04}) +[2026-06-24 12:07:44,785][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.12277331948280334, 'train/grad_norm_pre_clip': 37.71754455566406, 'timing/data': 0.0003901110030710697, 'timing/model': 1.0742804519832134, 'learning_rate/qwen_vl_interface': 9.861056443317105e-06, 'learning_rate/action_model': 9.861056443317105e-05, 'epoch': 0.04}) +[2026-06-24 12:07:45,858][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.22392719984054565, 'train/grad_norm_pre_clip': 70.72577667236328, 'timing/data': 0.00028891093097627163, 'timing/model': 1.0708003009203821, 'learning_rate/qwen_vl_interface': 9.860136261796e-06, 'learning_rate/action_model': 9.860136261795998e-05, 'epoch': 0.04}) +[2026-06-24 12:07:46,931][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.20351317524909973, 'train/grad_norm_pre_clip': 42.14813995361328, 'timing/data': 0.00039284583181142807, 'timing/model': 1.0709894069004804, 'learning_rate/qwen_vl_interface': 9.859213088807521e-06, 'learning_rate/action_model': 9.85921308880752e-05, 'epoch': 0.04}) +[2026-06-24 12:07:48,002][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.45556291937828064, 'train/grad_norm_pre_clip': 165.29269409179688, 'timing/data': 0.000403036130592227, 'timing/model': 1.0701524580363184, 'learning_rate/qwen_vl_interface': 9.858286924950714e-06, 'learning_rate/action_model': 9.858286924950713e-05, 'epoch': 0.04}) +[2026-06-24 12:07:49,073][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.4113503396511078, 'train/grad_norm_pre_clip': 130.38653564453125, 'timing/data': 0.00029288907535374165, 'timing/model': 1.0686868319753557, 'learning_rate/qwen_vl_interface': 9.85735777082655e-06, 'learning_rate/action_model': 9.857357770826551e-05, 'epoch': 0.04}) +[2026-06-24 12:07:50,146][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.2997645139694214, 'train/grad_norm_pre_clip': 32.35500717163086, 'timing/data': 0.0003288760781288147, 'timing/model': 1.0714792551007122, 'learning_rate/qwen_vl_interface': 9.856425627037951e-06, 'learning_rate/action_model': 9.856425627037952e-05, 'epoch': 0.04}) +[2026-06-24 12:07:51,219][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.20525281131267548, 'train/grad_norm_pre_clip': 76.41691589355469, 'timing/data': 0.0003172841388732195, 'timing/model': 1.0712446270044893, 'learning_rate/qwen_vl_interface': 9.855490494189777e-06, 'learning_rate/action_model': 9.855490494189776e-05, 'epoch': 0.04}) +[2026-06-24 12:07:52,291][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.10922610014677048, 'train/grad_norm_pre_clip': 70.8356704711914, 'timing/data': 0.00033044908195734024, 'timing/model': 1.0707199140451849, 'learning_rate/qwen_vl_interface': 9.854552372888819e-06, 'learning_rate/action_model': 9.854552372888819e-05, 'epoch': 0.04}) +[2026-06-24 12:07:53,365][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.2547038197517395, 'train/grad_norm_pre_clip': 34.99562072753906, 'timing/data': 0.0003227749839425087, 'timing/model': 1.072694364003837, 'learning_rate/qwen_vl_interface': 9.853611263743817e-06, 'learning_rate/action_model': 9.853611263743817e-05, 'epoch': 0.04}) +[2026-06-24 12:07:54,436][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.2916775643825531, 'train/grad_norm_pre_clip': 104.83458709716797, 'timing/data': 0.0003317408263683319, 'timing/model': 1.0696811589878052, 'learning_rate/qwen_vl_interface': 9.85266716736545e-06, 'learning_rate/action_model': 9.852667167365449e-05, 'epoch': 0.04}) +[2026-06-24 12:07:55,506][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.30261558294296265, 'train/grad_norm_pre_clip': 45.250823974609375, 'timing/data': 0.0002938108518719673, 'timing/model': 1.0682440458331257, 'learning_rate/qwen_vl_interface': 9.851720084366324e-06, 'learning_rate/action_model': 9.851720084366324e-05, 'epoch': 0.04}) +[2026-06-24 12:07:56,579][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.19273824989795685, 'train/grad_norm_pre_clip': 42.48739242553711, 'timing/data': 0.00030742608942091465, 'timing/model': 1.070852275006473, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.04}) +[2026-06-24 12:07:57,647][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.22574178874492645, 'train/grad_norm_pre_clip': 49.60792541503906, 'timing/data': 0.0003247978165745735, 'timing/model': 1.06653687893413, 'learning_rate/qwen_vl_interface': 9.849816960965959e-06, 'learning_rate/action_model': 9.849816960965959e-05, 'epoch': 0.04}) +[2026-06-24 12:07:59,012][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.2767319679260254, 'train/grad_norm_pre_clip': 101.07547760009766, 'timing/data': 0.00029334984719753265, 'timing/model': 1.363960969960317, 'learning_rate/qwen_vl_interface': 9.848860921799633e-06, 'learning_rate/action_model': 9.848860921799632e-05, 'epoch': 0.04}) +[2026-06-24 12:08:00,085][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.22203373908996582, 'train/grad_norm_pre_clip': 40.261680603027344, 'timing/data': 0.00032842508517205715, 'timing/model': 1.0709833120927215, 'learning_rate/qwen_vl_interface': 9.847901898482385e-06, 'learning_rate/action_model': 9.847901898482385e-05, 'epoch': 0.04}) +[2026-06-24 12:08:01,157][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.3062342405319214, 'train/grad_norm_pre_clip': 111.11347198486328, 'timing/data': 0.00041587906889617443, 'timing/model': 1.0700138660613447, 'learning_rate/qwen_vl_interface': 9.846939891636512e-06, 'learning_rate/action_model': 9.846939891636512e-05, 'epoch': 0.04}) +[2026-06-24 12:08:02,229][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.44321784377098083, 'train/grad_norm_pre_clip': 61.7984504699707, 'timing/data': 0.00035150907933712006, 'timing/model': 1.0705639170482755, 'learning_rate/qwen_vl_interface': 9.845974901886253e-06, 'learning_rate/action_model': 9.845974901886252e-05, 'epoch': 0.04}) +[2026-06-24 12:08:03,300][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.17457100749015808, 'train/grad_norm_pre_clip': 33.380863189697266, 'timing/data': 0.0005467550363391638, 'timing/model': 1.0692261410877109, 'learning_rate/qwen_vl_interface': 9.845006929857777e-06, 'learning_rate/action_model': 9.845006929857776e-05, 'epoch': 0.04}) +[2026-06-24 12:08:04,370][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.14745338261127472, 'train/grad_norm_pre_clip': 56.002113342285156, 'timing/data': 0.0004829750396311283, 'timing/model': 1.0680797428358346, 'learning_rate/qwen_vl_interface': 9.844035976179193e-06, 'learning_rate/action_model': 9.844035976179192e-05, 'epoch': 0.04}) +[2026-06-24 12:08:05,441][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.285526841878891, 'train/grad_norm_pre_clip': 27.03746795654297, 'timing/data': 0.0003259910736232996, 'timing/model': 1.069317403016612, 'learning_rate/qwen_vl_interface': 9.843062041480538e-06, 'learning_rate/action_model': 9.843062041480539e-05, 'epoch': 0.04}) +[2026-06-24 12:08:06,513][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.12781894207000732, 'train/grad_norm_pre_clip': 46.78266906738281, 'timing/data': 0.00029851007275283337, 'timing/model': 1.070523563073948, 'learning_rate/qwen_vl_interface': 9.842085126393793e-06, 'learning_rate/action_model': 9.842085126393792e-05, 'epoch': 0.04}) +[2026-06-24 12:08:07,583][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.26128020882606506, 'train/grad_norm_pre_clip': 33.59156036376953, 'timing/data': 0.00035469396971166134, 'timing/model': 1.0685560768470168, 'learning_rate/qwen_vl_interface': 9.841105231552863e-06, 'learning_rate/action_model': 9.841105231552864e-05, 'epoch': 0.04}) +[2026-06-24 12:08:08,655][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.507422685623169, 'train/grad_norm_pre_clip': 122.74006652832031, 'timing/data': 0.0005712909623980522, 'timing/model': 1.0695085930638015, 'learning_rate/qwen_vl_interface': 9.840122357593594e-06, 'learning_rate/action_model': 9.840122357593594e-05, 'epoch': 0.04}) +[2026-06-24 12:08:09,725][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.21174657344818115, 'train/grad_norm_pre_clip': 52.561893463134766, 'timing/data': 0.00032278383150696754, 'timing/model': 1.0685933479107916, 'learning_rate/qwen_vl_interface': 9.839136505153762e-06, 'learning_rate/action_model': 9.839136505153762e-05, 'epoch': 0.04}) +[2026-06-24 12:08:10,797][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.276199609041214, 'train/grad_norm_pre_clip': 40.4365234375, 'timing/data': 0.00032019009813666344, 'timing/model': 1.0709754128474742, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.04}) +[2026-06-24 12:08:11,871][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.21990111470222473, 'train/grad_norm_pre_clip': 36.270511627197266, 'timing/data': 0.0003063241019845009, 'timing/model': 1.071686645038426, 'learning_rate/qwen_vl_interface': 9.837155867393176e-06, 'learning_rate/action_model': 9.837155867393175e-05, 'epoch': 0.04}) +[2026-06-24 12:08:12,942][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.2702489495277405, 'train/grad_norm_pre_clip': 60.03709030151367, 'timing/data': 0.000322133069857955, 'timing/model': 1.0697478230576962, 'learning_rate/qwen_vl_interface': 9.836161083357635e-06, 'learning_rate/action_model': 9.836161083357635e-05, 'epoch': 0.04}) +[2026-06-24 12:08:14,306][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.28922122716903687, 'train/grad_norm_pre_clip': 79.27890014648438, 'timing/data': 0.0003222839441150427, 'timing/model': 1.3625330640934408, 'learning_rate/qwen_vl_interface': 9.835163323411957e-06, 'learning_rate/action_model': 9.835163323411958e-05, 'epoch': 0.04}) +[2026-06-24 12:08:15,374][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.4998490810394287, 'train/grad_norm_pre_clip': 97.55886840820312, 'timing/data': 0.0003142689820379019, 'timing/model': 1.0670245010405779, 'learning_rate/qwen_vl_interface': 9.834162588203581e-06, 'learning_rate/action_model': 9.83416258820358e-05, 'epoch': 0.04}) +[2026-06-24 12:08:16,443][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.19562169909477234, 'train/grad_norm_pre_clip': 35.222564697265625, 'timing/data': 0.00032140198163688183, 'timing/model': 1.0674834221135825, 'learning_rate/qwen_vl_interface': 9.833158878381872e-06, 'learning_rate/action_model': 9.83315887838187e-05, 'epoch': 0.04}) +[2026-06-24 12:08:17,512][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.10683877021074295, 'train/grad_norm_pre_clip': 19.585084915161133, 'timing/data': 0.00029783789068460464, 'timing/model': 1.0672130959574133, 'learning_rate/qwen_vl_interface': 9.832152194598123e-06, 'learning_rate/action_model': 9.832152194598122e-05, 'epoch': 0.04}) +[2026-06-24 12:08:18,581][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.10895330458879471, 'train/grad_norm_pre_clip': 39.50790786743164, 'timing/data': 0.00032498897053301334, 'timing/model': 1.0673054289072752, 'learning_rate/qwen_vl_interface': 9.831142537505563e-06, 'learning_rate/action_model': 9.831142537505562e-05, 'epoch': 0.04}) +[2026-06-24 12:08:19,653][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.2535402774810791, 'train/grad_norm_pre_clip': 47.26449966430664, 'timing/data': 0.00031325710006058216, 'timing/model': 1.0700210989452899, 'learning_rate/qwen_vl_interface': 9.830129907759347e-06, 'learning_rate/action_model': 9.830129907759347e-05, 'epoch': 0.04}) +[2026-06-24 12:08:20,723][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.18194471299648285, 'train/grad_norm_pre_clip': 40.025047302246094, 'timing/data': 0.000335057033225894, 'timing/model': 1.0689173908904195, 'learning_rate/qwen_vl_interface': 9.82911430601656e-06, 'learning_rate/action_model': 9.82911430601656e-05, 'epoch': 0.04}) +[2026-06-24 12:08:21,793][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.2451578676700592, 'train/grad_norm_pre_clip': 44.19099426269531, 'timing/data': 0.00031003099866211414, 'timing/model': 1.0685258880257607, 'learning_rate/qwen_vl_interface': 9.828095732936216e-06, 'learning_rate/action_model': 9.828095732936216e-05, 'epoch': 0.04}) +[2026-06-24 12:08:22,863][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.37912803888320923, 'train/grad_norm_pre_clip': 36.03673553466797, 'timing/data': 0.0003041699528694153, 'timing/model': 1.0680545249488205, 'learning_rate/qwen_vl_interface': 9.827074189179254e-06, 'learning_rate/action_model': 9.827074189179253e-05, 'epoch': 0.04}) +[2026-06-24 12:08:23,933][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.2787376940250397, 'train/grad_norm_pre_clip': 63.61552047729492, 'timing/data': 0.0003083581104874611, 'timing/model': 1.0683081410825253, 'learning_rate/qwen_vl_interface': 9.82604967540854e-06, 'learning_rate/action_model': 9.82604967540854e-05, 'epoch': 0.04}) +[2026-06-24 12:08:25,004][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.18363581597805023, 'train/grad_norm_pre_clip': 13.797001838684082, 'timing/data': 0.0003184070810675621, 'timing/model': 1.0700905921403319, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.04}) +[2026-06-24 12:08:26,074][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.3347359299659729, 'train/grad_norm_pre_clip': 46.73369216918945, 'timing/data': 0.0003412398509681225, 'timing/model': 1.0683201549109071, 'learning_rate/qwen_vl_interface': 9.823991740486982e-06, 'learning_rate/action_model': 9.823991740486982e-05, 'epoch': 0.04}) +[2026-06-24 12:08:27,145][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.6335001587867737, 'train/grad_norm_pre_clip': 79.41078186035156, 'timing/data': 0.00029986188746988773, 'timing/model': 1.0691172790247947, 'learning_rate/qwen_vl_interface': 9.822958320671506e-06, 'learning_rate/action_model': 9.822958320671506e-05, 'epoch': 0.04}) +[2026-06-24 12:08:28,219][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.12400715798139572, 'train/grad_norm_pre_clip': 35.85186767578125, 'timing/data': 0.00034188106656074524, 'timing/model': 1.0728838781360537, 'learning_rate/qwen_vl_interface': 9.821921933513023e-06, 'learning_rate/action_model': 9.821921933513023e-05, 'epoch': 0.04}) +[2026-06-24 12:08:29,582][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.36335936188697815, 'train/grad_norm_pre_clip': 42.054420471191406, 'timing/data': 0.0003308500163257122, 'timing/model': 1.3606390447821468, 'learning_rate/qwen_vl_interface': 9.820882579684037e-06, 'learning_rate/action_model': 9.820882579684037e-05, 'epoch': 0.04}) +[2026-06-24 12:08:30,653][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.2582772672176361, 'train/grad_norm_pre_clip': 61.104000091552734, 'timing/data': 0.00030021206475794315, 'timing/model': 1.0693048520479351, 'learning_rate/qwen_vl_interface': 9.819840259858968e-06, 'learning_rate/action_model': 9.819840259858968e-05, 'epoch': 0.04}) +[2026-06-24 12:08:31,722][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.14102822542190552, 'train/grad_norm_pre_clip': 41.735450744628906, 'timing/data': 0.0003050409723073244, 'timing/model': 1.0675083461683244, 'learning_rate/qwen_vl_interface': 9.818794974714167e-06, 'learning_rate/action_model': 9.818794974714168e-05, 'epoch': 0.04}) +[2026-06-24 12:08:32,793][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.18736369907855988, 'train/grad_norm_pre_clip': 25.058574676513672, 'timing/data': 0.00031776586547493935, 'timing/model': 1.069506720872596, 'learning_rate/qwen_vl_interface': 9.817746724927912e-06, 'learning_rate/action_model': 9.817746724927912e-05, 'epoch': 0.04}) +[2026-06-24 12:08:33,867][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.2197604924440384, 'train/grad_norm_pre_clip': 22.199045181274414, 'timing/data': 0.0003314411733299494, 'timing/model': 1.0729125440120697, 'learning_rate/qwen_vl_interface': 9.8166955111804e-06, 'learning_rate/action_model': 9.816695511180399e-05, 'epoch': 0.04}) +[2026-06-24 12:08:34,938][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.2222583293914795, 'train/grad_norm_pre_clip': 23.86163902282715, 'timing/data': 0.0003126258961856365, 'timing/model': 1.0687831370159984, 'learning_rate/qwen_vl_interface': 9.815641334153746e-06, 'learning_rate/action_model': 9.815641334153746e-05, 'epoch': 0.04}) +[2026-06-24 12:08:36,008][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.25627434253692627, 'train/grad_norm_pre_clip': 39.888916015625, 'timing/data': 0.00032138312235474586, 'timing/model': 1.0684689080808312, 'learning_rate/qwen_vl_interface': 9.814584194532001e-06, 'learning_rate/action_model': 9.814584194532001e-05, 'epoch': 0.04}) +[2026-06-24 12:08:37,081][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.3267052471637726, 'train/grad_norm_pre_clip': 47.448997497558594, 'timing/data': 0.0003044998738914728, 'timing/model': 1.0719099259003997, 'learning_rate/qwen_vl_interface': 9.813524093001128e-06, 'learning_rate/action_model': 9.813524093001129e-05, 'epoch': 0.04}) +[2026-06-24 12:08:38,152][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.17907945811748505, 'train/grad_norm_pre_clip': 32.322235107421875, 'timing/data': 0.0006249509751796722, 'timing/model': 1.0691371618304402, 'learning_rate/qwen_vl_interface': 9.81246103024902e-06, 'learning_rate/action_model': 9.81246103024902e-05, 'epoch': 0.04}) +[2026-06-24 12:08:39,229][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.1790827363729477, 'train/grad_norm_pre_clip': 52.60997772216797, 'timing/data': 0.00030860817059874535, 'timing/model': 1.0753655729349703, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.04}) +[2026-06-24 12:08:40,301][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.30239853262901306, 'train/grad_norm_pre_clip': 85.56855773925781, 'timing/data': 0.0003289759624749422, 'timing/model': 1.0700874340254813, 'learning_rate/qwen_vl_interface': 9.810326023842245e-06, 'learning_rate/action_model': 9.810326023842245e-05, 'epoch': 0.04}) +[2026-06-24 12:08:41,379][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.4687662720680237, 'train/grad_norm_pre_clip': 133.58973693847656, 'timing/data': 0.0003409779164940119, 'timing/model': 1.0707087579648942, 'learning_rate/qwen_vl_interface': 9.809254081572963e-06, 'learning_rate/action_model': 9.809254081572961e-05, 'epoch': 0.04}) +[2026-06-24 12:08:42,449][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.41766688227653503, 'train/grad_norm_pre_clip': 137.023681640625, 'timing/data': 0.000307456124573946, 'timing/model': 1.0688468599691987, 'learning_rate/qwen_vl_interface': 9.808179180853204e-06, 'learning_rate/action_model': 9.808179180853204e-05, 'epoch': 0.04}) +[2026-06-24 12:08:43,521][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.4206550419330597, 'train/grad_norm_pre_clip': 74.02409362792969, 'timing/data': 0.00031094299629330635, 'timing/model': 1.0700906810816377, 'learning_rate/qwen_vl_interface': 9.807101322380464e-06, 'learning_rate/action_model': 9.807101322380464e-05, 'epoch': 0.04}) +[2026-06-24 12:08:44,882][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.24790993332862854, 'train/grad_norm_pre_clip': 45.4378547668457, 'timing/data': 0.00035595684312283993, 'timing/model': 1.3596965679898858, 'learning_rate/qwen_vl_interface': 9.806020506854152e-06, 'learning_rate/action_model': 9.806020506854153e-05, 'epoch': 0.04}) +[2026-06-24 12:08:45,954][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.20274107158184052, 'train/grad_norm_pre_clip': 32.928768157958984, 'timing/data': 0.0003135879524052143, 'timing/model': 1.0700515189673752, 'learning_rate/qwen_vl_interface': 9.804936734975596e-06, 'learning_rate/action_model': 9.804936734975596e-05, 'epoch': 0.04}) +[2026-06-24 12:08:47,024][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.3489164412021637, 'train/grad_norm_pre_clip': 47.30186462402344, 'timing/data': 0.0003207509871572256, 'timing/model': 1.0684816869907081, 'learning_rate/qwen_vl_interface': 9.803850007448046e-06, 'learning_rate/action_model': 9.803850007448045e-05, 'epoch': 0.04}) +[2026-06-24 12:08:48,096][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.1930956095457077, 'train/grad_norm_pre_clip': 45.89081573486328, 'timing/data': 0.00042705098167061806, 'timing/model': 1.069928077980876, 'learning_rate/qwen_vl_interface': 9.802760324976666e-06, 'learning_rate/action_model': 9.802760324976667e-05, 'epoch': 0.04}) +[2026-06-24 12:08:49,170][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.5466662049293518, 'train/grad_norm_pre_clip': 49.64224624633789, 'timing/data': 0.00032561994157731533, 'timing/model': 1.072577603859827, 'learning_rate/qwen_vl_interface': 9.80166768826854e-06, 'learning_rate/action_model': 9.80166768826854e-05, 'epoch': 0.04}) +[2026-06-24 12:08:50,243][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.33260759711265564, 'train/grad_norm_pre_clip': 61.33934020996094, 'timing/data': 0.00031457003206014633, 'timing/model': 1.0710665329825133, 'learning_rate/qwen_vl_interface': 9.800572098032668e-06, 'learning_rate/action_model': 9.800572098032668e-05, 'epoch': 0.04}) +[2026-06-24 12:08:51,314][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.2269291877746582, 'train/grad_norm_pre_clip': 36.812076568603516, 'timing/data': 0.00030788686126470566, 'timing/model': 1.0694146680179983, 'learning_rate/qwen_vl_interface': 9.799473554979966e-06, 'learning_rate/action_model': 9.799473554979965e-05, 'epoch': 0.04}) +[2026-06-24 12:08:52,393][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.19997137784957886, 'train/grad_norm_pre_clip': 32.64533996582031, 'timing/data': 0.0005325479432940483, 'timing/model': 1.0766487040091306, 'learning_rate/qwen_vl_interface': 9.798372059823268e-06, 'learning_rate/action_model': 9.798372059823268e-05, 'epoch': 0.04}) +[2026-06-24 12:08:53,465][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.3037290573120117, 'train/grad_norm_pre_clip': 18.88576889038086, 'timing/data': 0.00030093383975327015, 'timing/model': 1.0706606139428914, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.04}) +[2026-06-24 12:08:54,537][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.30418917536735535, 'train/grad_norm_pre_clip': 29.244686126708984, 'timing/data': 0.00031227595172822475, 'timing/model': 1.0707517049741, 'learning_rate/qwen_vl_interface': 9.79616021605879e-06, 'learning_rate/action_model': 9.796160216058788e-05, 'epoch': 0.04}) +[2026-06-24 12:08:55,609][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.4353930950164795, 'train/grad_norm_pre_clip': 79.27022552490234, 'timing/data': 0.00037481216713786125, 'timing/model': 1.0704123091418296, 'learning_rate/qwen_vl_interface': 9.795049868886247e-06, 'learning_rate/action_model': 9.795049868886247e-05, 'epoch': 0.05}) +[2026-06-24 12:08:56,684][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.30160537362098694, 'train/grad_norm_pre_clip': 40.618980407714844, 'timing/data': 0.00030555296689271927, 'timing/model': 1.0726554428692907, 'learning_rate/qwen_vl_interface': 9.793936572480192e-06, 'learning_rate/action_model': 9.793936572480192e-05, 'epoch': 0.05}) +[2026-06-24 12:08:57,757][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.22209583222866058, 'train/grad_norm_pre_clip': 26.72163963317871, 'timing/data': 0.00032737315632402897, 'timing/model': 1.0719289521221071, 'learning_rate/qwen_vl_interface': 9.792820327563028e-06, 'learning_rate/action_model': 9.792820327563028e-05, 'epoch': 0.05}) +[2026-06-24 12:08:58,830][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.1873517483472824, 'train/grad_norm_pre_clip': 16.236967086791992, 'timing/data': 0.00037970184348523617, 'timing/model': 1.0707703919615597, 'learning_rate/qwen_vl_interface': 9.791701134859076e-06, 'learning_rate/action_model': 9.791701134859075e-05, 'epoch': 0.05}) +[2026-06-24 12:09:00,167][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.4965144991874695, 'train/grad_norm_pre_clip': 79.09793090820312, 'timing/data': 0.0003435241524130106, 'timing/model': 1.3357351860031486, 'learning_rate/qwen_vl_interface': 9.790578995094565e-06, 'learning_rate/action_model': 9.790578995094563e-05, 'epoch': 0.05}) +[2026-06-24 12:09:01,239][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.13013416528701782, 'train/grad_norm_pre_clip': 36.01079559326172, 'timing/data': 0.00031649298034608364, 'timing/model': 1.0706127979792655, 'learning_rate/qwen_vl_interface': 9.789453908997639e-06, 'learning_rate/action_model': 9.789453908997638e-05, 'epoch': 0.05}) +[2026-06-24 12:09:02,308][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.12241734564304352, 'train/grad_norm_pre_clip': 18.590192794799805, 'timing/data': 0.0003417900297790766, 'timing/model': 1.0672938090283424, 'learning_rate/qwen_vl_interface': 9.788325877298355e-06, 'learning_rate/action_model': 9.788325877298355e-05, 'epoch': 0.05}) +[2026-06-24 12:09:03,383][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.26380443572998047, 'train/grad_norm_pre_clip': 83.05093383789062, 'timing/data': 0.0003621389623731375, 'timing/model': 1.073481534840539, 'learning_rate/qwen_vl_interface': 9.787194900728682e-06, 'learning_rate/action_model': 9.787194900728683e-05, 'epoch': 0.05}) +[2026-06-24 12:09:04,455][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.251409113407135, 'train/grad_norm_pre_clip': 51.58628463745117, 'timing/data': 0.00033207214437425137, 'timing/model': 1.0708224319387227, 'learning_rate/qwen_vl_interface': 9.786060980022499e-06, 'learning_rate/action_model': 9.786060980022498e-05, 'epoch': 0.05}) +[2026-06-24 12:09:05,528][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.22592337429523468, 'train/grad_norm_pre_clip': 61.57810592651367, 'timing/data': 0.00034534791484475136, 'timing/model': 1.071145277004689, 'learning_rate/qwen_vl_interface': 9.784924115915592e-06, 'learning_rate/action_model': 9.784924115915592e-05, 'epoch': 0.05}) +[2026-06-24 12:09:06,602][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.21226714551448822, 'train/grad_norm_pre_clip': 49.763648986816406, 'timing/data': 0.0003163630608469248, 'timing/model': 1.0716164400801063, 'learning_rate/qwen_vl_interface': 9.783784309145663e-06, 'learning_rate/action_model': 9.783784309145661e-05, 'epoch': 0.05}) +[2026-06-24 12:09:07,674][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.40188178420066833, 'train/grad_norm_pre_clip': 101.29510498046875, 'timing/data': 0.00031592207960784435, 'timing/model': 1.0710396009963006, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.05}) +[2026-06-24 12:09:08,746][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.3775308132171631, 'train/grad_norm_pre_clip': 65.19258117675781, 'timing/data': 0.00031802617013454437, 'timing/model': 1.0697447049897164, 'learning_rate/qwen_vl_interface': 9.78149587057707e-06, 'learning_rate/action_model': 9.78149587057707e-05, 'epoch': 0.05}) +[2026-06-24 12:09:09,819][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.15288536250591278, 'train/grad_norm_pre_clip': 33.98660659790039, 'timing/data': 0.0003680391237139702, 'timing/model': 1.0713928830809891, 'learning_rate/qwen_vl_interface': 9.780347240263354e-06, 'learning_rate/action_model': 9.780347240263353e-05, 'epoch': 0.05}) +[2026-06-24 12:09:10,887][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.19135086238384247, 'train/grad_norm_pre_clip': 53.26542663574219, 'timing/data': 0.0003635010216385126, 'timing/model': 1.0663373300340027, 'learning_rate/qwen_vl_interface': 9.779195670256498e-06, 'learning_rate/action_model': 9.779195670256497e-05, 'epoch': 0.05}) +[2026-06-24 12:09:11,957][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.12755025923252106, 'train/grad_norm_pre_clip': 17.402273178100586, 'timing/data': 0.000289783114567399, 'timing/model': 1.068212625104934, 'learning_rate/qwen_vl_interface': 9.778041161303746e-06, 'learning_rate/action_model': 9.778041161303745e-05, 'epoch': 0.05}) +[2026-06-24 12:09:13,029][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.17790573835372925, 'train/grad_norm_pre_clip': 50.350372314453125, 'timing/data': 0.00033759186044335365, 'timing/model': 1.070578298997134, 'learning_rate/qwen_vl_interface': 9.776883714154242e-06, 'learning_rate/action_model': 9.776883714154241e-05, 'epoch': 0.05}) +[2026-06-24 12:09:14,100][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.17437879741191864, 'train/grad_norm_pre_clip': 64.12614440917969, 'timing/data': 0.0003043999895453453, 'timing/model': 1.0695476671680808, 'learning_rate/qwen_vl_interface': 9.775723329559044e-06, 'learning_rate/action_model': 9.775723329559044e-05, 'epoch': 0.05}) +[2026-06-24 12:09:15,467][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.10249552130699158, 'train/grad_norm_pre_clip': 20.99813461303711, 'timing/data': 0.00032729306258261204, 'timing/model': 1.365967026911676, 'learning_rate/qwen_vl_interface': 9.774560008271114e-06, 'learning_rate/action_model': 9.774560008271114e-05, 'epoch': 0.05}) +[2026-06-24 12:09:16,537][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.16089339554309845, 'train/grad_norm_pre_clip': 42.43324279785156, 'timing/data': 0.0003106121439486742, 'timing/model': 1.0679486021399498, 'learning_rate/qwen_vl_interface': 9.773393751045315e-06, 'learning_rate/action_model': 9.773393751045315e-05, 'epoch': 0.05}) +[2026-06-24 12:09:17,608][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.12007751315832138, 'train/grad_norm_pre_clip': 45.40414047241211, 'timing/data': 0.00033629988320171833, 'timing/model': 1.0690558080095798, 'learning_rate/qwen_vl_interface': 9.772224558638422e-06, 'learning_rate/action_model': 9.772224558638422e-05, 'epoch': 0.05}) +[2026-06-24 12:09:18,678][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.41541588306427, 'train/grad_norm_pre_clip': 74.49148559570312, 'timing/data': 0.00031543103978037834, 'timing/model': 1.0684136948548257, 'learning_rate/qwen_vl_interface': 9.77105243180911e-06, 'learning_rate/action_model': 9.771052431809109e-05, 'epoch': 0.05}) +[2026-06-24 12:09:19,750][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.29669591784477234, 'train/grad_norm_pre_clip': 72.00967407226562, 'timing/data': 0.00031585199758410454, 'timing/model': 1.0703320698812604, 'learning_rate/qwen_vl_interface': 9.76987737131796e-06, 'learning_rate/action_model': 9.76987737131796e-05, 'epoch': 0.05}) +[2026-06-24 12:09:20,822][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.04921744391322136, 'train/grad_norm_pre_clip': 14.959393501281738, 'timing/data': 0.0004115309566259384, 'timing/model': 1.070498582907021, 'learning_rate/qwen_vl_interface': 9.768699377927454e-06, 'learning_rate/action_model': 9.768699377927453e-05, 'epoch': 0.05}) +[2026-06-24 12:09:21,893][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.2163943499326706, 'train/grad_norm_pre_clip': 19.488662719726562, 'timing/data': 0.00034127896651625633, 'timing/model': 1.0695504259783775, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.05}) +[2026-06-24 12:09:22,969][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.4005740284919739, 'train/grad_norm_pre_clip': 63.96249008178711, 'timing/data': 0.00029641599394381046, 'timing/model': 1.0741018059197813, 'learning_rate/qwen_vl_interface': 9.76633459550783e-06, 'learning_rate/action_model': 9.76633459550783e-05, 'epoch': 0.05}) +[2026-06-24 12:09:24,040][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.4521070122718811, 'train/grad_norm_pre_clip': 57.27333450317383, 'timing/data': 0.00032594078220427036, 'timing/model': 1.06988391187042, 'learning_rate/qwen_vl_interface': 9.765147808013193e-06, 'learning_rate/action_model': 9.765147808013192e-05, 'epoch': 0.05}) +[2026-06-24 12:09:25,113][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.4310319423675537, 'train/grad_norm_pre_clip': 74.40773010253906, 'timing/data': 0.0002947321627289057, 'timing/model': 1.07106670900248, 'learning_rate/qwen_vl_interface': 9.763958090688165e-06, 'learning_rate/action_model': 9.763958090688164e-05, 'epoch': 0.05}) +[2026-06-24 12:09:26,183][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.1847849190235138, 'train/grad_norm_pre_clip': 67.77521514892578, 'timing/data': 0.00030655390582978725, 'timing/model': 1.0686185320373625, 'learning_rate/qwen_vl_interface': 9.762765444304737e-06, 'learning_rate/action_model': 9.762765444304737e-05, 'epoch': 0.05}) +[2026-06-24 12:09:27,252][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.0788361206650734, 'train/grad_norm_pre_clip': 51.44801330566406, 'timing/data': 0.00031887786462903023, 'timing/model': 1.0672736919950694, 'learning_rate/qwen_vl_interface': 9.761569869636808e-06, 'learning_rate/action_model': 9.761569869636808e-05, 'epoch': 0.05}) +[2026-06-24 12:09:28,320][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.14350873231887817, 'train/grad_norm_pre_clip': 37.5305061340332, 'timing/data': 0.000384259968996048, 'timing/model': 1.065778149990365, 'learning_rate/qwen_vl_interface': 9.760371367460172e-06, 'learning_rate/action_model': 9.760371367460171e-05, 'epoch': 0.05}) +[2026-06-24 12:09:29,393][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.26671165227890015, 'train/grad_norm_pre_clip': 61.04414749145508, 'timing/data': 0.0003321319818496704, 'timing/model': 1.070886072004214, 'learning_rate/qwen_vl_interface': 9.759169938552524e-06, 'learning_rate/action_model': 9.759169938552524e-05, 'epoch': 0.05}) +[2026-06-24 12:09:30,756][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.3681429326534271, 'train/grad_norm_pre_clip': 38.89790344238281, 'timing/data': 0.00033861491829156876, 'timing/model': 1.3610571350436658, 'learning_rate/qwen_vl_interface': 9.757965583693458e-06, 'learning_rate/action_model': 9.757965583693457e-05, 'epoch': 0.05}) +[2026-06-24 12:09:31,826][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.3074093759059906, 'train/grad_norm_pre_clip': 42.11506652832031, 'timing/data': 0.00037970091216266155, 'timing/model': 1.0685348268598318, 'learning_rate/qwen_vl_interface': 9.756758303664467e-06, 'learning_rate/action_model': 9.756758303664468e-05, 'epoch': 0.05}) +[2026-06-24 12:09:32,895][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 500 +[2026-06-24 12:09:57,285][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 500 +[2026-06-24 12:09:57,286][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 500 +[2026-06-24 12:10:29,951][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 500 +[2026-06-24 12:11:12,332][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-24 12:11:12,333][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.6162391901016235, 'train/grad_norm_pre_clip': 79.41155242919922, 'eval/action_loss/seconds': 24.388410250889137, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.150342658162117, 'eval/latency_2/loss': 0.150342658162117, 'eval/flappy/loss': 0.150342658162117, 'eval/flappy/latency_2/loss': 0.150342658162117, 'eval/flap_precision': 0.7631027253668763, 'eval/flap_recall': 0.9877883310719131, 'eval/flap_f1': 0.8610289769367238, 'eval/flap_rate_model': 0.265, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8610289769367238, 'eval/latency_2/flap_precision': 0.7631027253668763, 'eval/latency_2/flap_recall': 0.9877883310719131, 'eval/latency_2/flap_f1': 0.8610289769367238, 'eval/latency_2/flap_rate_model': 0.265, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8610289769367238, 'eval/action_classification/data_wait_seconds': 0.3076152503490448, 'eval/action_classification/predict_seconds': 32.26344680786133, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 31.500000000000178, 'rl_games_eval/mid_train/mean_length': 273.8, 'rl_games_eval/mid_train/std_reward': 14.203098253550309, 'rl_games_eval/mid_train/std_length': 114.50659369660771, 'rl_games_eval/mid_train/macro_mean_reward': 31.500000000000178, 'rl_games_eval/mid_train/macro_mean_length': 273.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 31.500000000000178, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 273.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 14.203098253550309, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 114.50659369660771, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 31.500000000000178, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0002975671086460352, 'timing/model': 1.0679532380308956, 'learning_rate/qwen_vl_interface': 9.755548099248942e-06, 'learning_rate/action_model': 9.75554809924894e-05, 'epoch': 0.05}) +[2026-06-24 12:11:12,333][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_500_state +[2026-06-24 12:11:12,336][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 12:11:14,852][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_500_state/model.safetensors +[2026-06-24 12:11:22,931][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_500_state/optimizer.bin +[2026-06-24 12:11:22,931][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_500_state/sampler.bin +[2026-06-24 12:11:22,931][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_500_state/sampler_1.bin +[2026-06-24 12:11:22,932][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-24 12:11:22,933][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-24 12:11:22,933][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 12:11:22,942][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 12:11:24,012][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.1863311529159546, 'train/grad_norm_pre_clip': 33.834007263183594, 'timing/data': 0.0003956311848014593, 'timing/model': 1.0676576478872448, 'learning_rate/qwen_vl_interface': 9.75433497123217e-06, 'learning_rate/action_model': 9.754334971232171e-05, 'epoch': 0.05}) +[2026-06-24 12:11:25,083][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.31418293714523315, 'train/grad_norm_pre_clip': 44.60988998413086, 'timing/data': 0.0004976529162377119, 'timing/model': 1.069209175882861, 'learning_rate/qwen_vl_interface': 9.753118920401341e-06, 'learning_rate/action_model': 9.75311892040134e-05, 'epoch': 0.05}) +[2026-06-24 12:11:26,149][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.3236776888370514, 'train/grad_norm_pre_clip': 37.620357513427734, 'timing/data': 0.0004392031114548445, 'timing/model': 1.0646406239829957, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.05}) +[2026-06-24 12:11:27,212][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.20876117050647736, 'train/grad_norm_pre_clip': 22.58991050720215, 'timing/data': 0.00034812185913324356, 'timing/model': 1.0613162950612605, 'learning_rate/qwen_vl_interface': 9.750678053455725e-06, 'learning_rate/action_model': 9.750678053455725e-05, 'epoch': 0.05}) +[2026-06-24 12:11:28,277][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.3598209023475647, 'train/grad_norm_pre_clip': 40.10207748413086, 'timing/data': 0.00030512199737131596, 'timing/model': 1.063032923033461, 'learning_rate/qwen_vl_interface': 9.749453238924793e-06, 'learning_rate/action_model': 9.749453238924792e-05, 'epoch': 0.05}) +[2026-06-24 12:11:29,343][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.23501719534397125, 'train/grad_norm_pre_clip': 30.855819702148438, 'timing/data': 0.00035513495095074177, 'timing/model': 1.0643221491482109, 'learning_rate/qwen_vl_interface': 9.7482255047475e-06, 'learning_rate/action_model': 9.7482255047475e-05, 'epoch': 0.05}) +[2026-06-24 12:11:30,681][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.14712496101856232, 'train/grad_norm_pre_clip': 23.355161666870117, 'timing/data': 0.0004194860812276602, 'timing/model': 1.33625942398794, 'learning_rate/qwen_vl_interface': 9.746994851720518e-06, 'learning_rate/action_model': 9.746994851720517e-05, 'epoch': 0.05}) +[2026-06-24 12:11:31,750][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.17890140414237976, 'train/grad_norm_pre_clip': 27.669336318969727, 'timing/data': 0.0003242669627070427, 'timing/model': 1.0674708001315594, 'learning_rate/qwen_vl_interface': 9.745761280642397e-06, 'learning_rate/action_model': 9.745761280642397e-05, 'epoch': 0.05}) +[2026-06-24 12:11:32,816][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.15259744226932526, 'train/grad_norm_pre_clip': 48.288848876953125, 'timing/data': 0.0003267920110374689, 'timing/model': 1.0644496190361679, 'learning_rate/qwen_vl_interface': 9.744524792313591e-06, 'learning_rate/action_model': 9.74452479231359e-05, 'epoch': 0.05}) +[2026-06-24 12:11:33,883][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.22643518447875977, 'train/grad_norm_pre_clip': 56.047080993652344, 'timing/data': 0.00044965208508074284, 'timing/model': 1.0654726868961006, 'learning_rate/qwen_vl_interface': 9.743285387536445e-06, 'learning_rate/action_model': 9.743285387536444e-05, 'epoch': 0.05}) +[2026-06-24 12:11:34,950][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.12133204936981201, 'train/grad_norm_pre_clip': 19.56168556213379, 'timing/data': 0.0003139078617095947, 'timing/model': 1.065157996956259, 'learning_rate/qwen_vl_interface': 9.742043067115193e-06, 'learning_rate/action_model': 9.742043067115192e-05, 'epoch': 0.05}) +[2026-06-24 12:11:36,015][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.2670832574367523, 'train/grad_norm_pre_clip': 34.304386138916016, 'timing/data': 0.0003344269935041666, 'timing/model': 1.062826498877257, 'learning_rate/qwen_vl_interface': 9.740797831855965e-06, 'learning_rate/action_model': 9.740797831855965e-05, 'epoch': 0.05}) +[2026-06-24 12:11:37,084][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.05892682820558548, 'train/grad_norm_pre_clip': 38.23036193847656, 'timing/data': 0.0003039189614355564, 'timing/model': 1.067381955916062, 'learning_rate/qwen_vl_interface': 9.73954968256678e-06, 'learning_rate/action_model': 9.73954968256678e-05, 'epoch': 0.05}) +[2026-06-24 12:11:38,150][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.12608887255191803, 'train/grad_norm_pre_clip': 27.278715133666992, 'timing/data': 0.0003050919622182846, 'timing/model': 1.0648728650994599, 'learning_rate/qwen_vl_interface': 9.738298620057547e-06, 'learning_rate/action_model': 9.738298620057546e-05, 'epoch': 0.05}) +[2026-06-24 12:11:39,221][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.17936468124389648, 'train/grad_norm_pre_clip': 36.33595657348633, 'timing/data': 0.0004969011060893536, 'timing/model': 1.068936730036512, 'learning_rate/qwen_vl_interface': 9.73704464514007e-06, 'learning_rate/action_model': 9.737044645140068e-05, 'epoch': 0.05}) +[2026-06-24 12:11:40,287][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.11432801932096481, 'train/grad_norm_pre_clip': 22.07550621032715, 'timing/data': 0.0005340108182281256, 'timing/model': 1.0638412008993328, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.05}) +[2026-06-24 12:11:41,353][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.17119812965393066, 'train/grad_norm_pre_clip': 48.715057373046875, 'timing/data': 0.0003035191912204027, 'timing/model': 1.0653095729649067, 'learning_rate/qwen_vl_interface': 9.734527961337028e-06, 'learning_rate/action_model': 9.734527961337028e-05, 'epoch': 0.05}) +[2026-06-24 12:11:42,421][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.31822624802589417, 'train/grad_norm_pre_clip': 34.83538818359375, 'timing/data': 0.0003041899763047695, 'timing/model': 1.0657031622249633, 'learning_rate/qwen_vl_interface': 9.733265254084514e-06, 'learning_rate/action_model': 9.733265254084513e-05, 'epoch': 0.05}) +[2026-06-24 12:11:43,491][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.2369304746389389, 'train/grad_norm_pre_clip': 195.8514862060547, 'timing/data': 0.00036704796366393566, 'timing/model': 1.0681105030234903, 'learning_rate/qwen_vl_interface': 9.73199963768985e-06, 'learning_rate/action_model': 9.731999637689849e-05, 'epoch': 0.05}) +[2026-06-24 12:11:44,561][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.11579501628875732, 'train/grad_norm_pre_clip': 75.39807891845703, 'timing/data': 0.00035728001967072487, 'timing/model': 1.0684639159590006, 'learning_rate/qwen_vl_interface': 9.73073111297428e-06, 'learning_rate/action_model': 9.73073111297428e-05, 'epoch': 0.05}) +[2026-06-24 12:11:45,632][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.2696652412414551, 'train/grad_norm_pre_clip': 44.29147720336914, 'timing/data': 0.0003339559771120548, 'timing/model': 1.06950913509354, 'learning_rate/qwen_vl_interface': 9.72945968076094e-06, 'learning_rate/action_model': 9.729459680760939e-05, 'epoch': 0.05}) +[2026-06-24 12:11:46,703][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.31077393889427185, 'train/grad_norm_pre_clip': 109.203857421875, 'timing/data': 0.0003419008571654558, 'timing/model': 1.069135745987296, 'learning_rate/qwen_vl_interface': 9.728185341874844e-06, 'learning_rate/action_model': 9.728185341874844e-05, 'epoch': 0.05}) +[2026-06-24 12:11:48,049][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.14804893732070923, 'train/grad_norm_pre_clip': 19.740516662597656, 'timing/data': 0.00030005304142832756, 'timing/model': 1.3450311301276088, 'learning_rate/qwen_vl_interface': 9.7269080971429e-06, 'learning_rate/action_model': 9.726908097142899e-05, 'epoch': 0.05}) +[2026-06-24 12:11:49,121][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.4389501214027405, 'train/grad_norm_pre_clip': 79.76116180419922, 'timing/data': 0.0003420710563659668, 'timing/model': 1.070326728047803, 'learning_rate/qwen_vl_interface': 9.725627947393896e-06, 'learning_rate/action_model': 9.725627947393896e-05, 'epoch': 0.05}) +[2026-06-24 12:11:50,192][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.5825014114379883, 'train/grad_norm_pre_clip': 89.9586410522461, 'timing/data': 0.0003214818425476551, 'timing/model': 1.0695183130446821, 'learning_rate/qwen_vl_interface': 9.72434489345851e-06, 'learning_rate/action_model': 9.72434489345851e-05, 'epoch': 0.05}) +[2026-06-24 12:11:51,262][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.13059082627296448, 'train/grad_norm_pre_clip': 74.44562530517578, 'timing/data': 0.00030324794352054596, 'timing/model': 1.0677911371458322, 'learning_rate/qwen_vl_interface': 9.723058936169298e-06, 'learning_rate/action_model': 9.723058936169297e-05, 'epoch': 0.05}) +[2026-06-24 12:11:52,333][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.16071084141731262, 'train/grad_norm_pre_clip': 62.496665954589844, 'timing/data': 0.0003220129292458296, 'timing/model': 1.0696904668584466, 'learning_rate/qwen_vl_interface': 9.721770076360707e-06, 'learning_rate/action_model': 9.721770076360706e-05, 'epoch': 0.05}) +[2026-06-24 12:11:53,403][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.11832979321479797, 'train/grad_norm_pre_clip': 22.256319046020508, 'timing/data': 0.00033412594348192215, 'timing/model': 1.0687546429689974, 'learning_rate/qwen_vl_interface': 9.720478314869061e-06, 'learning_rate/action_model': 9.72047831486906e-05, 'epoch': 0.05}) +[2026-06-24 12:11:54,473][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.16155704855918884, 'train/grad_norm_pre_clip': 33.068016052246094, 'timing/data': 0.000375563045963645, 'timing/model': 1.0674810169730335, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.05}) +[2026-06-24 12:11:55,542][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.1511823683977127, 'train/grad_norm_pre_clip': 34.714534759521484, 'timing/data': 0.0003135369624942541, 'timing/model': 1.0678499590139836, 'learning_rate/qwen_vl_interface': 9.717886090191334e-06, 'learning_rate/action_model': 9.717886090191333e-05, 'epoch': 0.05}) +[2026-06-24 12:11:56,614][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.38222426176071167, 'train/grad_norm_pre_clip': 73.61164855957031, 'timing/data': 0.0003136370796710253, 'timing/model': 1.0708760700654238, 'learning_rate/qwen_vl_interface': 9.716585628687317e-06, 'learning_rate/action_model': 9.716585628687316e-05, 'epoch': 0.05}) +[2026-06-24 12:11:57,687][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.08345292508602142, 'train/grad_norm_pre_clip': 15.916759490966797, 'timing/data': 0.0003492739051580429, 'timing/model': 1.070939278928563, 'learning_rate/qwen_vl_interface': 9.715282268864379e-06, 'learning_rate/action_model': 9.715282268864378e-05, 'epoch': 0.05}) +[2026-06-24 12:11:58,762][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.2784965932369232, 'train/grad_norm_pre_clip': 42.77595901489258, 'timing/data': 0.00031355698592960835, 'timing/model': 1.0733293080702424, 'learning_rate/qwen_vl_interface': 9.713976011568256e-06, 'learning_rate/action_model': 9.713976011568256e-05, 'epoch': 0.05}) +[2026-06-24 12:11:59,827][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.09756803512573242, 'train/grad_norm_pre_clip': 13.139110565185547, 'timing/data': 0.0003182960208505392, 'timing/model': 1.0638380518648773, 'learning_rate/qwen_vl_interface': 9.712666857646565e-06, 'learning_rate/action_model': 9.712666857646564e-05, 'epoch': 0.05}) +[2026-06-24 12:12:00,900][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.13916988670825958, 'train/grad_norm_pre_clip': 49.69000244140625, 'timing/data': 0.0003658151254057884, 'timing/model': 1.0705749879125506, 'learning_rate/qwen_vl_interface': 9.711354807948798e-06, 'learning_rate/action_model': 9.711354807948798e-05, 'epoch': 0.05}) +[2026-06-24 12:12:01,971][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.12375470995903015, 'train/grad_norm_pre_clip': 20.25988006591797, 'timing/data': 0.0003130468539893627, 'timing/model': 1.0696757519617677, 'learning_rate/qwen_vl_interface': 9.710039863326334e-06, 'learning_rate/action_model': 9.710039863326334e-05, 'epoch': 0.05}) +[2026-06-24 12:12:03,043][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.08513353019952774, 'train/grad_norm_pre_clip': 24.53010368347168, 'timing/data': 0.0003323429264128208, 'timing/model': 1.0706270060036331, 'learning_rate/qwen_vl_interface': 9.708722024632425e-06, 'learning_rate/action_model': 9.708722024632424e-05, 'epoch': 0.05}) +[2026-06-24 12:12:04,113][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.4395088851451874, 'train/grad_norm_pre_clip': 43.354393005371094, 'timing/data': 0.0003040689043700695, 'timing/model': 1.0682993938680738, 'learning_rate/qwen_vl_interface': 9.7074012927222e-06, 'learning_rate/action_model': 9.7074012927222e-05, 'epoch': 0.05}) +[2026-06-24 12:12:05,466][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.10252778232097626, 'train/grad_norm_pre_clip': 24.548465728759766, 'timing/data': 0.0005027521401643753, 'timing/model': 1.350576268043369, 'learning_rate/qwen_vl_interface': 9.706077668452671e-06, 'learning_rate/action_model': 9.706077668452671e-05, 'epoch': 0.05}) +[2026-06-24 12:12:06,536][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.10724657028913498, 'train/grad_norm_pre_clip': 45.56962966918945, 'timing/data': 0.0003168429248034954, 'timing/model': 1.0681714040692896, 'learning_rate/qwen_vl_interface': 9.704751152682722e-06, 'learning_rate/action_model': 9.704751152682721e-05, 'epoch': 0.05}) +[2026-06-24 12:12:07,610][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.0582624115049839, 'train/grad_norm_pre_clip': 24.06525993347168, 'timing/data': 0.00030527194030582905, 'timing/model': 1.0729443589225411, 'learning_rate/qwen_vl_interface': 9.703421746273115e-06, 'learning_rate/action_model': 9.703421746273114e-05, 'epoch': 0.05}) +[2026-06-24 12:12:08,681][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.11160672456026077, 'train/grad_norm_pre_clip': 42.737518310546875, 'timing/data': 0.00035562715493142605, 'timing/model': 1.068900991929695, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.05}) +[2026-06-24 12:12:09,753][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.3732281029224396, 'train/grad_norm_pre_clip': 78.72908782958984, 'timing/data': 0.0003146899398416281, 'timing/model': 1.0705118621699512, 'learning_rate/qwen_vl_interface': 9.700754264987347e-06, 'learning_rate/action_model': 9.700754264987347e-05, 'epoch': 0.05}) +[2026-06-24 12:12:10,828][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.20249305665493011, 'train/grad_norm_pre_clip': 67.31343078613281, 'timing/data': 0.0004494620952755213, 'timing/model': 1.0738722099922597, 'learning_rate/qwen_vl_interface': 9.69941619184209e-06, 'learning_rate/action_model': 9.699416191842089e-05, 'epoch': 0.05}) +[2026-06-24 12:12:11,902][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.3312405049800873, 'train/grad_norm_pre_clip': 118.63922882080078, 'timing/data': 0.0003133281134068966, 'timing/model': 1.0717617550399154, 'learning_rate/qwen_vl_interface': 9.69807523151897e-06, 'learning_rate/action_model': 9.698075231518969e-05, 'epoch': 0.05}) +[2026-06-24 12:12:12,974][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.3969248831272125, 'train/grad_norm_pre_clip': 132.74078369140625, 'timing/data': 0.00035212980583310127, 'timing/model': 1.0707395798526704, 'learning_rate/qwen_vl_interface': 9.696731384888125e-06, 'learning_rate/action_model': 9.696731384888125e-05, 'epoch': 0.05}) +[2026-06-24 12:12:14,050][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.10426916927099228, 'train/grad_norm_pre_clip': 64.93147277832031, 'timing/data': 0.00030926894396543503, 'timing/model': 1.0742333270609379, 'learning_rate/qwen_vl_interface': 9.695384652821562e-06, 'learning_rate/action_model': 9.695384652821561e-05, 'epoch': 0.05}) +[2026-06-24 12:12:15,122][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.32373809814453125, 'train/grad_norm_pre_clip': 29.452707290649414, 'timing/data': 0.0003284050617367029, 'timing/model': 1.0710586979985237, 'learning_rate/qwen_vl_interface': 9.694035036193162e-06, 'learning_rate/action_model': 9.694035036193161e-05, 'epoch': 0.05}) +[2026-06-24 12:12:16,197][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.1703484207391739, 'train/grad_norm_pre_clip': 31.866165161132812, 'timing/data': 0.0003023361787199974, 'timing/model': 1.073114010039717, 'learning_rate/qwen_vl_interface': 9.692682535878674e-06, 'learning_rate/action_model': 9.692682535878673e-05, 'epoch': 0.05}) +[2026-06-24 12:12:17,272][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.18002642691135406, 'train/grad_norm_pre_clip': 31.12152862548828, 'timing/data': 0.0003236960619688034, 'timing/model': 1.0730084120295942, 'learning_rate/qwen_vl_interface': 9.691327152755721e-06, 'learning_rate/action_model': 9.691327152755722e-05, 'epoch': 0.05}) +[2026-06-24 12:12:18,343][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.08146172761917114, 'train/grad_norm_pre_clip': 44.16047286987305, 'timing/data': 0.00034617912024259567, 'timing/model': 1.0693407501094043, 'learning_rate/qwen_vl_interface': 9.6899688877038e-06, 'learning_rate/action_model': 9.689968887703798e-05, 'epoch': 0.05}) +[2026-06-24 12:12:19,413][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.3432031273841858, 'train/grad_norm_pre_clip': 67.2265396118164, 'timing/data': 0.0003213919699192047, 'timing/model': 1.068207236006856, 'learning_rate/qwen_vl_interface': 9.688607741604269e-06, 'learning_rate/action_model': 9.688607741604269e-05, 'epoch': 0.05}) +[2026-06-24 12:12:20,485][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.27531036734580994, 'train/grad_norm_pre_clip': 51.68308639526367, 'timing/data': 0.000298850005492568, 'timing/model': 1.0704132409300655, 'learning_rate/qwen_vl_interface': 9.687243715340365e-06, 'learning_rate/action_model': 9.687243715340364e-05, 'epoch': 0.05}) +[2026-06-24 12:12:21,555][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.22439466416835785, 'train/grad_norm_pre_clip': 55.19340896606445, 'timing/data': 0.0003685408737510443, 'timing/model': 1.068580857012421, 'learning_rate/qwen_vl_interface': 9.685876809797188e-06, 'learning_rate/action_model': 9.685876809797188e-05, 'epoch': 0.05}) +[2026-06-24 12:12:22,907][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.2093675583600998, 'train/grad_norm_pre_clip': 74.02689361572266, 'timing/data': 0.00035525509156286716, 'timing/model': 1.3507051768247038, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.05}) +[2026-06-24 12:12:23,980][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.31970423460006714, 'train/grad_norm_pre_clip': 43.63534927368164, 'timing/data': 0.0003574700094759464, 'timing/model': 1.0713379650842398, 'learning_rate/qwen_vl_interface': 9.683134364422766e-06, 'learning_rate/action_model': 9.683134364422765e-05, 'epoch': 0.05}) +[2026-06-24 12:12:25,051][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.182433620095253, 'train/grad_norm_pre_clip': 12.103589057922363, 'timing/data': 0.00030250707641243935, 'timing/model': 1.0695651229470968, 'learning_rate/qwen_vl_interface': 9.681758826371064e-06, 'learning_rate/action_model': 9.681758826371063e-05, 'epoch': 0.05}) +[2026-06-24 12:12:26,124][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.16523440182209015, 'train/grad_norm_pre_clip': 20.347293853759766, 'timing/data': 0.0005315360613167286, 'timing/model': 1.0705427848733962, 'learning_rate/qwen_vl_interface': 9.680380412599174e-06, 'learning_rate/action_model': 9.680380412599173e-05, 'epoch': 0.05}) +[2026-06-24 12:12:27,198][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.16794231534004211, 'train/grad_norm_pre_clip': 23.077838897705078, 'timing/data': 0.00029644486494362354, 'timing/model': 1.072503900853917, 'learning_rate/qwen_vl_interface': 9.678999124001533e-06, 'learning_rate/action_model': 9.678999124001533e-05, 'epoch': 0.05}) +[2026-06-24 12:12:28,269][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.2223810851573944, 'train/grad_norm_pre_clip': 39.458282470703125, 'timing/data': 0.0003136279992759228, 'timing/model': 1.07008289382793, 'learning_rate/qwen_vl_interface': 9.67761496147445e-06, 'learning_rate/action_model': 9.677614961474449e-05, 'epoch': 0.05}) +[2026-06-24 12:12:29,339][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.30481016635894775, 'train/grad_norm_pre_clip': 53.1671142578125, 'timing/data': 0.00032448815181851387, 'timing/model': 1.067986334906891, 'learning_rate/qwen_vl_interface': 9.676227925916082e-06, 'learning_rate/action_model': 9.676227925916082e-05, 'epoch': 0.05}) +[2026-06-24 12:12:30,413][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.2370358556509018, 'train/grad_norm_pre_clip': 34.597259521484375, 'timing/data': 0.00032471888698637486, 'timing/model': 1.073037230875343, 'learning_rate/qwen_vl_interface': 9.674838018226472e-06, 'learning_rate/action_model': 9.674838018226471e-05, 'epoch': 0.05}) +[2026-06-24 12:12:31,491][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.2795785963535309, 'train/grad_norm_pre_clip': 52.45225143432617, 'timing/data': 0.0003199090715497732, 'timing/model': 1.075819875113666, 'learning_rate/qwen_vl_interface': 9.67344523930751e-06, 'learning_rate/action_model': 9.673445239307511e-05, 'epoch': 0.05}) +[2026-06-24 12:12:32,566][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.5556841492652893, 'train/grad_norm_pre_clip': 62.28582763671875, 'timing/data': 0.0003270020242780447, 'timing/model': 1.0736783328466117, 'learning_rate/qwen_vl_interface': 9.672049590062959e-06, 'learning_rate/action_model': 9.672049590062958e-05, 'epoch': 0.05}) +[2026-06-24 12:12:33,638][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.13047164678573608, 'train/grad_norm_pre_clip': 26.446151733398438, 'timing/data': 0.00033486797474324703, 'timing/model': 1.0709870599675924, 'learning_rate/qwen_vl_interface': 9.670651071398436e-06, 'learning_rate/action_model': 9.670651071398436e-05, 'epoch': 0.05}) +[2026-06-24 12:12:34,711][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.11920447647571564, 'train/grad_norm_pre_clip': 40.335514068603516, 'timing/data': 0.0003065939527004957, 'timing/model': 1.0704975528642535, 'learning_rate/qwen_vl_interface': 9.669249684221427e-06, 'learning_rate/action_model': 9.669249684221428e-05, 'epoch': 0.05}) +[2026-06-24 12:12:35,783][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.24757802486419678, 'train/grad_norm_pre_clip': 51.04084777832031, 'timing/data': 0.0003264008555561304, 'timing/model': 1.0703542539849877, 'learning_rate/qwen_vl_interface': 9.66784542944128e-06, 'learning_rate/action_model': 9.66784542944128e-05, 'epoch': 0.05}) +[2026-06-24 12:12:36,857][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.245519757270813, 'train/grad_norm_pre_clip': 130.72779846191406, 'timing/data': 0.00031109293922781944, 'timing/model': 1.0729179489426315, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.05}) +[2026-06-24 12:12:37,932][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.3269132077693939, 'train/grad_norm_pre_clip': 47.21556854248047, 'timing/data': 0.00030100392177700996, 'timing/model': 1.0733427859377116, 'learning_rate/qwen_vl_interface': 9.66502832071824e-06, 'learning_rate/action_model': 9.66502832071824e-05, 'epoch': 0.06}) +[2026-06-24 12:12:39,008][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.2748973071575165, 'train/grad_norm_pre_clip': 51.69913101196289, 'timing/data': 0.0003078770823776722, 'timing/model': 1.0739996270276606, 'learning_rate/qwen_vl_interface': 9.663615468603344e-06, 'learning_rate/action_model': 9.663615468603343e-05, 'epoch': 0.06}) +[2026-06-24 12:12:40,361][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.31438302993774414, 'train/grad_norm_pre_clip': 28.675750732421875, 'timing/data': 0.0003291158936917782, 'timing/model': 1.3515069158747792, 'learning_rate/qwen_vl_interface': 9.662199752541287e-06, 'learning_rate/action_model': 9.662199752541287e-05, 'epoch': 0.06}) +[2026-06-24 12:12:41,437][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.13837867975234985, 'train/grad_norm_pre_clip': 35.706485748291016, 'timing/data': 0.00031247595325112343, 'timing/model': 1.074217877117917, 'learning_rate/qwen_vl_interface': 9.660781173450711e-06, 'learning_rate/action_model': 9.660781173450712e-05, 'epoch': 0.06}) +[2026-06-24 12:12:42,507][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.19192305207252502, 'train/grad_norm_pre_clip': 31.481008529663086, 'timing/data': 0.0002971470821648836, 'timing/model': 1.069271638058126, 'learning_rate/qwen_vl_interface': 9.659359732252121e-06, 'learning_rate/action_model': 9.659359732252121e-05, 'epoch': 0.06}) +[2026-06-24 12:12:43,582][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.20700761675834656, 'train/grad_norm_pre_clip': 29.198917388916016, 'timing/data': 0.0005007889121770859, 'timing/model': 1.0723677498754114, 'learning_rate/qwen_vl_interface': 9.657935429867872e-06, 'learning_rate/action_model': 9.657935429867873e-05, 'epoch': 0.06}) +[2026-06-24 12:12:44,656][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.3099474310874939, 'train/grad_norm_pre_clip': 57.239707946777344, 'timing/data': 0.0003307298757135868, 'timing/model': 1.0729083940386772, 'learning_rate/qwen_vl_interface': 9.65650826722218e-06, 'learning_rate/action_model': 9.656508267222179e-05, 'epoch': 0.06}) +[2026-06-24 12:12:45,728][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.19455118477344513, 'train/grad_norm_pre_clip': 35.892974853515625, 'timing/data': 0.0003018658608198166, 'timing/model': 1.0702912700362504, 'learning_rate/qwen_vl_interface': 9.655078245241114e-06, 'learning_rate/action_model': 9.655078245241112e-05, 'epoch': 0.06}) +[2026-06-24 12:12:46,803][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.18035003542900085, 'train/grad_norm_pre_clip': 52.73690414428711, 'timing/data': 0.00030751689337193966, 'timing/model': 1.0732546031940728, 'learning_rate/qwen_vl_interface': 9.653645364852598e-06, 'learning_rate/action_model': 9.653645364852597e-05, 'epoch': 0.06}) +[2026-06-24 12:12:47,877][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.3552674949169159, 'train/grad_norm_pre_clip': 31.646316528320312, 'timing/data': 0.00036160810850560665, 'timing/model': 1.0720246478449553, 'learning_rate/qwen_vl_interface': 9.652209626986414e-06, 'learning_rate/action_model': 9.652209626986413e-05, 'epoch': 0.06}) +[2026-06-24 12:12:48,954][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.29821670055389404, 'train/grad_norm_pre_clip': 31.46348762512207, 'timing/data': 0.00033926614560186863, 'timing/model': 1.0751763449516147, 'learning_rate/qwen_vl_interface': 9.6507710325742e-06, 'learning_rate/action_model': 9.6507710325742e-05, 'epoch': 0.06}) +[2026-06-24 12:12:50,032][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.12372635304927826, 'train/grad_norm_pre_clip': 18.422969818115234, 'timing/data': 0.0003876269329339266, 'timing/model': 1.0770250700879842, 'learning_rate/qwen_vl_interface': 9.649329582549439e-06, 'learning_rate/action_model': 9.649329582549439e-05, 'epoch': 0.06}) +[2026-06-24 12:12:51,106][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.15429961681365967, 'train/grad_norm_pre_clip': 19.33470344543457, 'timing/data': 0.00037106615491211414, 'timing/model': 1.0718260169960558, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.06}) +[2026-06-24 12:12:52,180][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.14593273401260376, 'train/grad_norm_pre_clip': 29.6937198638916, 'timing/data': 0.00034955493174493313, 'timing/model': 1.0720438258722425, 'learning_rate/qwen_vl_interface': 9.646438119405503e-06, 'learning_rate/action_model': 9.646438119405503e-05, 'epoch': 0.06}) +[2026-06-24 12:12:53,255][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.2268322855234146, 'train/grad_norm_pre_clip': 32.2714958190918, 'timing/data': 0.00031127408146858215, 'timing/model': 1.0739880800247192, 'learning_rate/qwen_vl_interface': 9.644988108162566e-06, 'learning_rate/action_model': 9.644988108162567e-05, 'epoch': 0.06}) +[2026-06-24 12:12:54,329][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.286367267370224, 'train/grad_norm_pre_clip': 70.67900085449219, 'timing/data': 0.0003611268475651741, 'timing/model': 1.0725262099877, 'learning_rate/qwen_vl_interface': 9.643535245059562e-06, 'learning_rate/action_model': 9.643535245059561e-05, 'epoch': 0.06}) +[2026-06-24 12:12:55,403][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.04732130467891693, 'train/grad_norm_pre_clip': 7.235672950744629, 'timing/data': 0.000344475032761693, 'timing/model': 1.0721396058797836, 'learning_rate/qwen_vl_interface': 9.642079531039235e-06, 'learning_rate/action_model': 9.642079531039235e-05, 'epoch': 0.06}) +[2026-06-24 12:12:56,479][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.10844234377145767, 'train/grad_norm_pre_clip': 54.47344207763672, 'timing/data': 0.00031116302125155926, 'timing/model': 1.0739548071287572, 'learning_rate/qwen_vl_interface': 9.640620967046186e-06, 'learning_rate/action_model': 9.640620967046186e-05, 'epoch': 0.06}) +[2026-06-24 12:12:57,829][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.14370600879192352, 'train/grad_norm_pre_clip': 39.11723709106445, 'timing/data': 0.00032622087746858597, 'timing/model': 1.3489324480760843, 'learning_rate/qwen_vl_interface': 9.63915955402686e-06, 'learning_rate/action_model': 9.63915955402686e-05, 'epoch': 0.06}) +[2026-06-24 12:12:58,903][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.16481895744800568, 'train/grad_norm_pre_clip': 38.28403091430664, 'timing/data': 0.00031062192283570766, 'timing/model': 1.0721967939753085, 'learning_rate/qwen_vl_interface': 9.637695292929552e-06, 'learning_rate/action_model': 9.637695292929552e-05, 'epoch': 0.06}) +[2026-06-24 12:12:59,975][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.15455198287963867, 'train/grad_norm_pre_clip': 72.07853698730469, 'timing/data': 0.0003063138574361801, 'timing/model': 1.0707418089732528, 'learning_rate/qwen_vl_interface': 9.636228184704405e-06, 'learning_rate/action_model': 9.636228184704405e-05, 'epoch': 0.06}) +[2026-06-24 12:13:01,050][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.05021616443991661, 'train/grad_norm_pre_clip': 16.689729690551758, 'timing/data': 0.0003108629025518894, 'timing/model': 1.0729693111497909, 'learning_rate/qwen_vl_interface': 9.634758230303408e-06, 'learning_rate/action_model': 9.634758230303408e-05, 'epoch': 0.06}) +[2026-06-24 12:13:02,127][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.2757318913936615, 'train/grad_norm_pre_clip': 102.59510803222656, 'timing/data': 0.00033729197457432747, 'timing/model': 1.0749999480322003, 'learning_rate/qwen_vl_interface': 9.633285430680401e-06, 'learning_rate/action_model': 9.633285430680401e-05, 'epoch': 0.06}) +[2026-06-24 12:13:03,204][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.6059293150901794, 'train/grad_norm_pre_clip': 56.53355026245117, 'timing/data': 0.00030020205304026604, 'timing/model': 1.0760184382088482, 'learning_rate/qwen_vl_interface': 9.631809786791066e-06, 'learning_rate/action_model': 9.631809786791066e-05, 'epoch': 0.06}) +[2026-06-24 12:13:04,280][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.07333401590585709, 'train/grad_norm_pre_clip': 28.17860221862793, 'timing/data': 0.0003196790348738432, 'timing/model': 1.0738223120570183, 'learning_rate/qwen_vl_interface': 9.630331299592932e-06, 'learning_rate/action_model': 9.630331299592933e-05, 'epoch': 0.06}) +[2026-06-24 12:13:05,358][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.10067968815565109, 'train/grad_norm_pre_clip': 30.50161361694336, 'timing/data': 0.0003730389289557934, 'timing/model': 1.0765241049230099, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.06}) +[2026-06-24 12:13:06,435][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.5240778923034668, 'train/grad_norm_pre_clip': 58.66852569580078, 'timing/data': 0.0003425418399274349, 'timing/model': 1.0756865660659969, 'learning_rate/qwen_vl_interface': 9.627365799109617e-06, 'learning_rate/action_model': 9.627365799109617e-05, 'epoch': 0.06}) +[2026-06-24 12:13:07,515][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.8086226582527161, 'train/grad_norm_pre_clip': 143.32984924316406, 'timing/data': 0.0003462389577180147, 'timing/model': 1.0780008430592716, 'learning_rate/qwen_vl_interface': 9.625878787748712e-06, 'learning_rate/action_model': 9.625878787748713e-05, 'epoch': 0.06}) +[2026-06-24 12:13:08,588][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.3401772677898407, 'train/grad_norm_pre_clip': 34.87588882446289, 'timing/data': 0.00032062106765806675, 'timing/model': 1.0711716869845986, 'learning_rate/qwen_vl_interface': 9.624388936927572e-06, 'learning_rate/action_model': 9.624388936927571e-05, 'epoch': 0.06}) +[2026-06-24 12:13:09,663][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.28285372257232666, 'train/grad_norm_pre_clip': 50.30430603027344, 'timing/data': 0.0003093599807471037, 'timing/model': 1.0741417319513857, 'learning_rate/qwen_vl_interface': 9.622896247612943e-06, 'learning_rate/action_model': 9.622896247612942e-05, 'epoch': 0.06}) +[2026-06-24 12:13:10,739][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.2932509481906891, 'train/grad_norm_pre_clip': 83.05767059326172, 'timing/data': 0.0005407629068940878, 'timing/model': 1.0737214148975909, 'learning_rate/qwen_vl_interface': 9.621400720773414e-06, 'learning_rate/action_model': 9.621400720773414e-05, 'epoch': 0.06}) +[2026-06-24 12:13:11,816][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.18997229635715485, 'train/grad_norm_pre_clip': 44.00418472290039, 'timing/data': 0.0003422410227358341, 'timing/model': 1.075080931885168, 'learning_rate/qwen_vl_interface': 9.619902357379416e-06, 'learning_rate/action_model': 9.619902357379415e-05, 'epoch': 0.06}) +[2026-06-24 12:13:12,891][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.2746972143650055, 'train/grad_norm_pre_clip': 49.08606719970703, 'timing/data': 0.00030073407106101513, 'timing/model': 1.0731224529445171, 'learning_rate/qwen_vl_interface': 9.618401158403222e-06, 'learning_rate/action_model': 9.618401158403222e-05, 'epoch': 0.06}) +[2026-06-24 12:13:13,965][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.15766914188861847, 'train/grad_norm_pre_clip': 39.88549041748047, 'timing/data': 0.0002951838541775942, 'timing/model': 1.0730478439945728, 'learning_rate/qwen_vl_interface': 9.616897124818944e-06, 'learning_rate/action_model': 9.616897124818945e-05, 'epoch': 0.06}) +[2026-06-24 12:13:15,319][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.30166611075401306, 'train/grad_norm_pre_clip': 65.00889587402344, 'timing/data': 0.00030413013882935047, 'timing/model': 1.351808734005317, 'learning_rate/qwen_vl_interface': 9.615390257602532e-06, 'learning_rate/action_model': 9.615390257602531e-05, 'epoch': 0.06}) +[2026-06-24 12:13:16,396][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.42623481154441833, 'train/grad_norm_pre_clip': 67.64724731445312, 'timing/data': 0.00037106522358953953, 'timing/model': 1.0755551520269364, 'learning_rate/qwen_vl_interface': 9.613880557731776e-06, 'learning_rate/action_model': 9.613880557731776e-05, 'epoch': 0.06}) +[2026-06-24 12:13:17,467][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.16757431626319885, 'train/grad_norm_pre_clip': 23.668968200683594, 'timing/data': 0.0003068151418119669, 'timing/model': 1.069177431985736, 'learning_rate/qwen_vl_interface': 9.612368026186304e-06, 'learning_rate/action_model': 9.612368026186304e-05, 'epoch': 0.06}) +[2026-06-24 12:13:18,538][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.233164981007576, 'train/grad_norm_pre_clip': 54.12746810913086, 'timing/data': 0.0003284050617367029, 'timing/model': 1.0696222048718482, 'learning_rate/qwen_vl_interface': 9.610852663947582e-06, 'learning_rate/action_model': 9.610852663947583e-05, 'epoch': 0.06}) +[2026-06-24 12:13:19,607][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.22808940708637238, 'train/grad_norm_pre_clip': 25.66537094116211, 'timing/data': 0.0003065138589590788, 'timing/model': 1.067563307005912, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.06}) +[2026-06-24 12:13:20,679][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.18249393999576569, 'train/grad_norm_pre_clip': 33.981021881103516, 'timing/data': 0.00029813917353749275, 'timing/model': 1.0699748070910573, 'learning_rate/qwen_vl_interface': 9.607813451325431e-06, 'learning_rate/action_model': 9.60781345132543e-05, 'epoch': 0.06}) +[2026-06-24 12:13:21,752][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.2478533238172531, 'train/grad_norm_pre_clip': 38.07745361328125, 'timing/data': 0.00030470103956758976, 'timing/model': 1.0713919419795275, 'learning_rate/qwen_vl_interface': 9.606289602914113e-06, 'learning_rate/action_model': 9.606289602914112e-05, 'epoch': 0.06}) +[2026-06-24 12:13:22,833][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.3276968002319336, 'train/grad_norm_pre_clip': 27.714250564575195, 'timing/data': 0.0003403278533369303, 'timing/model': 1.079732562880963, 'learning_rate/qwen_vl_interface': 9.604762927753768e-06, 'learning_rate/action_model': 9.604762927753768e-05, 'epoch': 0.06}) +[2026-06-24 12:13:23,907][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.21428202092647552, 'train/grad_norm_pre_clip': 22.343894958496094, 'timing/data': 0.00047315703704953194, 'timing/model': 1.0725065520964563, 'learning_rate/qwen_vl_interface': 9.603233426835037e-06, 'learning_rate/action_model': 9.603233426835036e-05, 'epoch': 0.06}) +[2026-06-24 12:13:24,978][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.19740630686283112, 'train/grad_norm_pre_clip': 28.13134765625, 'timing/data': 0.0003047508653253317, 'timing/model': 1.0693195001222193, 'learning_rate/qwen_vl_interface': 9.601701101150399e-06, 'learning_rate/action_model': 9.601701101150398e-05, 'epoch': 0.06}) +[2026-06-24 12:13:26,051][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.2089054137468338, 'train/grad_norm_pre_clip': 55.619911193847656, 'timing/data': 0.00032901717349886894, 'timing/model': 1.0709730789531022, 'learning_rate/qwen_vl_interface': 9.600165951694162e-06, 'learning_rate/action_model': 9.60016595169416e-05, 'epoch': 0.06}) +[2026-06-24 12:13:27,125][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.14111240208148956, 'train/grad_norm_pre_clip': 39.201412200927734, 'timing/data': 0.0003142689820379019, 'timing/model': 1.0725176739506423, 'learning_rate/qwen_vl_interface': 9.598627979462466e-06, 'learning_rate/action_model': 9.598627979462467e-05, 'epoch': 0.06}) +[2026-06-24 12:13:28,197][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.2767614424228668, 'train/grad_norm_pre_clip': 29.899160385131836, 'timing/data': 0.00030154502019286156, 'timing/model': 1.0707244239747524, 'learning_rate/qwen_vl_interface': 9.59708718545329e-06, 'learning_rate/action_model': 9.597087185453288e-05, 'epoch': 0.06}) +[2026-06-24 12:13:29,271][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.18822753429412842, 'train/grad_norm_pre_clip': 23.37548065185547, 'timing/data': 0.00031783501617610455, 'timing/model': 1.0716679620090872, 'learning_rate/qwen_vl_interface': 9.595543570666433e-06, 'learning_rate/action_model': 9.595543570666432e-05, 'epoch': 0.06}) +[2026-06-24 12:13:30,348][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.33713221549987793, 'train/grad_norm_pre_clip': 66.69047546386719, 'timing/data': 0.0003054630942642689, 'timing/model': 1.0760320201516151, 'learning_rate/qwen_vl_interface': 9.593997136103533e-06, 'learning_rate/action_model': 9.593997136103533e-05, 'epoch': 0.06}) +[2026-06-24 12:13:31,422][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.31925278902053833, 'train/grad_norm_pre_clip': 68.88204956054688, 'timing/data': 0.0003050819505006075, 'timing/model': 1.0723695571068674, 'learning_rate/qwen_vl_interface': 9.592447882768053e-06, 'learning_rate/action_model': 9.592447882768052e-05, 'epoch': 0.06}) +[2026-06-24 12:13:32,764][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.09936365485191345, 'train/grad_norm_pre_clip': 21.79412841796875, 'timing/data': 0.0004487610422074795, 'timing/model': 1.3399932379834354, 'learning_rate/qwen_vl_interface': 9.590895811665284e-06, 'learning_rate/action_model': 9.590895811665285e-05, 'epoch': 0.06}) +[2026-06-24 12:13:33,838][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.16427522897720337, 'train/grad_norm_pre_clip': 28.348709106445312, 'timing/data': 0.0004970920272171497, 'timing/model': 1.0719344830140471, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.06}) +[2026-06-24 12:13:34,912][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.11944963037967682, 'train/grad_norm_pre_clip': 28.106407165527344, 'timing/data': 0.0004866220988333225, 'timing/model': 1.0728288090322167, 'learning_rate/qwen_vl_interface': 9.587783220188212e-06, 'learning_rate/action_model': 9.587783220188212e-05, 'epoch': 0.06}) +[2026-06-24 12:13:35,987][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.24236437678337097, 'train/grad_norm_pre_clip': 53.37470245361328, 'timing/data': 0.0003207111731171608, 'timing/model': 1.0735185721423477, 'learning_rate/qwen_vl_interface': 9.586222701833631e-06, 'learning_rate/action_model': 9.58622270183363e-05, 'epoch': 0.06}) +[2026-06-24 12:13:37,060][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.15511910617351532, 'train/grad_norm_pre_clip': 28.804906845092773, 'timing/data': 0.00036234897561371326, 'timing/model': 1.0706556469667703, 'learning_rate/qwen_vl_interface': 9.584659369751216e-06, 'learning_rate/action_model': 9.584659369751215e-05, 'epoch': 0.06}) +[2026-06-24 12:13:38,133][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.2353202998638153, 'train/grad_norm_pre_clip': 39.759178161621094, 'timing/data': 0.0003244581166654825, 'timing/model': 1.071931529091671, 'learning_rate/qwen_vl_interface': 9.583093224955397e-06, 'learning_rate/action_model': 9.583093224955397e-05, 'epoch': 0.06}) +[2026-06-24 12:13:39,215][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.26215606927871704, 'train/grad_norm_pre_clip': 44.050540924072266, 'timing/data': 0.00036174687556922436, 'timing/model': 1.0804778339806944, 'learning_rate/qwen_vl_interface': 9.58152426846243e-06, 'learning_rate/action_model': 9.581524268462429e-05, 'epoch': 0.06}) +[2026-06-24 12:13:40,293][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.05533275753259659, 'train/grad_norm_pre_clip': 16.283172607421875, 'timing/data': 0.0003109830431640148, 'timing/model': 1.0763444690965116, 'learning_rate/qwen_vl_interface': 9.579952501290388e-06, 'learning_rate/action_model': 9.579952501290387e-05, 'epoch': 0.06}) +[2026-06-24 12:13:41,368][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.20113083720207214, 'train/grad_norm_pre_clip': 37.4473762512207, 'timing/data': 0.0002990509383380413, 'timing/model': 1.0736489270348102, 'learning_rate/qwen_vl_interface': 9.57837792445918e-06, 'learning_rate/action_model': 9.578377924459181e-05, 'epoch': 0.06}) +[2026-06-24 12:13:42,439][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.22938664257526398, 'train/grad_norm_pre_clip': 27.32301139831543, 'timing/data': 0.0005217269062995911, 'timing/model': 1.0689835350494832, 'learning_rate/qwen_vl_interface': 9.57680053899053e-06, 'learning_rate/action_model': 9.576800538990529e-05, 'epoch': 0.06}) +[2026-06-24 12:13:43,518][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.08248580247163773, 'train/grad_norm_pre_clip': 27.25547218322754, 'timing/data': 0.0005251751281321049, 'timing/model': 1.076492447173223, 'learning_rate/qwen_vl_interface': 9.575220345907984e-06, 'learning_rate/action_model': 9.575220345907984e-05, 'epoch': 0.06}) +[2026-06-24 12:13:44,593][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.11352182179689407, 'train/grad_norm_pre_clip': 31.741544723510742, 'timing/data': 0.00038597313687205315, 'timing/model': 1.072998909978196, 'learning_rate/qwen_vl_interface': 9.573637346236915e-06, 'learning_rate/action_model': 9.573637346236915e-05, 'epoch': 0.06}) +[2026-06-24 12:13:45,666][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.22799764573574066, 'train/grad_norm_pre_clip': 27.41816520690918, 'timing/data': 0.0003459290601313114, 'timing/model': 1.0720515649300069, 'learning_rate/qwen_vl_interface': 9.572051541004513e-06, 'learning_rate/action_model': 9.572051541004513e-05, 'epoch': 0.06}) +[2026-06-24 12:13:46,743][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.26883813738822937, 'train/grad_norm_pre_clip': 46.54863739013672, 'timing/data': 0.00033216201700270176, 'timing/model': 1.0750432719942182, 'learning_rate/qwen_vl_interface': 9.57046293123979e-06, 'learning_rate/action_model': 9.570462931239789e-05, 'epoch': 0.06}) +[2026-06-24 12:13:47,819][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.21043255925178528, 'train/grad_norm_pre_clip': 38.35605239868164, 'timing/data': 0.0003456580452620983, 'timing/model': 1.0740013590548187, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.06}) +[2026-06-24 12:13:48,894][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.09397503733634949, 'train/grad_norm_pre_clip': 25.825712203979492, 'timing/data': 0.0003010048530995846, 'timing/model': 1.0739621159154922, 'learning_rate/qwen_vl_interface': 9.567277302238524e-06, 'learning_rate/action_model': 9.567277302238524e-05, 'epoch': 0.06}) +[2026-06-24 12:13:50,245][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.09234323352575302, 'train/grad_norm_pre_clip': 21.73434066772461, 'timing/data': 0.0003383839502930641, 'timing/model': 1.349248148035258, 'learning_rate/qwen_vl_interface': 9.565680285069104e-06, 'learning_rate/action_model': 9.565680285069103e-05, 'epoch': 0.06}) +[2026-06-24 12:13:51,320][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.4130404591560364, 'train/grad_norm_pre_clip': 121.27433013916016, 'timing/data': 0.0003358100075274706, 'timing/model': 1.0739602430257946, 'learning_rate/qwen_vl_interface': 9.564080467501601e-06, 'learning_rate/action_model': 9.5640804675016e-05, 'epoch': 0.06}) +[2026-06-24 12:13:52,394][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.06423614919185638, 'train/grad_norm_pre_clip': 26.37563705444336, 'timing/data': 0.0005319369956851006, 'timing/model': 1.0719174849800766, 'learning_rate/qwen_vl_interface': 9.562477850574118e-06, 'learning_rate/action_model': 9.562477850574119e-05, 'epoch': 0.06}) +[2026-06-24 12:13:53,471][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.12942387163639069, 'train/grad_norm_pre_clip': 39.16091537475586, 'timing/data': 0.00032285507768392563, 'timing/model': 1.074644936947152, 'learning_rate/qwen_vl_interface': 9.56087243532658e-06, 'learning_rate/action_model': 9.56087243532658e-05, 'epoch': 0.06}) +[2026-06-24 12:13:54,543][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.2105223387479782, 'train/grad_norm_pre_clip': 71.79234313964844, 'timing/data': 0.0003143390640616417, 'timing/model': 1.070845798123628, 'learning_rate/qwen_vl_interface': 9.55926422280072e-06, 'learning_rate/action_model': 9.55926422280072e-05, 'epoch': 0.06}) +[2026-06-24 12:13:55,622][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.023296819999814034, 'train/grad_norm_pre_clip': 9.148865699768066, 'timing/data': 0.000367077998816967, 'timing/model': 1.0774391640443355, 'learning_rate/qwen_vl_interface': 9.55765321404009e-06, 'learning_rate/action_model': 9.557653214040091e-05, 'epoch': 0.06}) +[2026-06-24 12:13:56,697][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.22005940973758698, 'train/grad_norm_pre_clip': 29.148088455200195, 'timing/data': 0.0004016819875687361, 'timing/model': 1.0736335930414498, 'learning_rate/qwen_vl_interface': 9.556039410090061e-06, 'learning_rate/action_model': 9.556039410090061e-05, 'epoch': 0.06}) +[2026-06-24 12:13:57,771][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.12384442239999771, 'train/grad_norm_pre_clip': 14.365241050720215, 'timing/data': 0.0003068449441343546, 'timing/model': 1.0725103178992867, 'learning_rate/qwen_vl_interface': 9.554422811997804e-06, 'learning_rate/action_model': 9.554422811997805e-05, 'epoch': 0.06}) +[2026-06-24 12:13:58,849][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.1948152482509613, 'train/grad_norm_pre_clip': 30.32903289794922, 'timing/data': 0.0003369611222296953, 'timing/model': 1.0754439258016646, 'learning_rate/qwen_vl_interface': 9.55280342081232e-06, 'learning_rate/action_model': 9.55280342081232e-05, 'epoch': 0.06}) +[2026-06-24 12:13:59,925][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.20681524276733398, 'train/grad_norm_pre_clip': 67.86294555664062, 'timing/data': 0.0003724780399352312, 'timing/model': 1.0745764400344342, 'learning_rate/qwen_vl_interface': 9.55118123758441e-06, 'learning_rate/action_model': 9.551181237584409e-05, 'epoch': 0.06}) +[2026-06-24 12:14:00,996][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.21249131858348846, 'train/grad_norm_pre_clip': 59.69404602050781, 'timing/data': 0.00034781196154654026, 'timing/model': 1.0697948299348354, 'learning_rate/qwen_vl_interface': 9.549556263366691e-06, 'learning_rate/action_model': 9.549556263366692e-05, 'epoch': 0.06}) +[2026-06-24 12:14:02,072][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.44436177611351013, 'train/grad_norm_pre_clip': 86.61475372314453, 'timing/data': 0.0003646130207926035, 'timing/model': 1.0744231541175395, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.06}) +[2026-06-24 12:14:03,150][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.23032300174236298, 'train/grad_norm_pre_clip': 93.8480453491211, 'timing/data': 0.00030799699015915394, 'timing/model': 1.0759085270110518, 'learning_rate/qwen_vl_interface': 9.546297946181354e-06, 'learning_rate/action_model': 9.546297946181354e-05, 'epoch': 0.06}) +[2026-06-24 12:14:04,223][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.30567803978919983, 'train/grad_norm_pre_clip': 63.647884368896484, 'timing/data': 0.00034555699676275253, 'timing/model': 1.0718976920470595, 'learning_rate/qwen_vl_interface': 9.544664605328022e-06, 'learning_rate/action_model': 9.544664605328022e-05, 'epoch': 0.06}) +[2026-06-24 12:14:05,298][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.21708685159683228, 'train/grad_norm_pre_clip': 42.56528091430664, 'timing/data': 0.0003069050144404173, 'timing/model': 1.0726453431416303, 'learning_rate/qwen_vl_interface': 9.543028477713455e-06, 'learning_rate/action_model': 9.543028477713453e-05, 'epoch': 0.06}) +[2026-06-24 12:14:06,373][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.1969190537929535, 'train/grad_norm_pre_clip': 53.468116760253906, 'timing/data': 0.0002977680414915085, 'timing/model': 1.0738572040572762, 'learning_rate/qwen_vl_interface': 9.541389564399316e-06, 'learning_rate/action_model': 9.541389564399315e-05, 'epoch': 0.06}) +[2026-06-24 12:14:07,712][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.14619560539722443, 'train/grad_norm_pre_clip': 37.099063873291016, 'timing/data': 0.0003328740131109953, 'timing/model': 1.3375977010000497, 'learning_rate/qwen_vl_interface': 9.53974786644908e-06, 'learning_rate/action_model': 9.53974786644908e-05, 'epoch': 0.06}) +[2026-06-24 12:14:08,787][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.25237083435058594, 'train/grad_norm_pre_clip': 42.75041198730469, 'timing/data': 0.00033680093474686146, 'timing/model': 1.0730459538754076, 'learning_rate/qwen_vl_interface': 9.538103384928029e-06, 'learning_rate/action_model': 9.538103384928027e-05, 'epoch': 0.06}) +[2026-06-24 12:14:09,866][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.1766507476568222, 'train/grad_norm_pre_clip': 28.3713436126709, 'timing/data': 0.00037557422183454037, 'timing/model': 1.0773122301325202, 'learning_rate/qwen_vl_interface': 9.536456120903244e-06, 'learning_rate/action_model': 9.536456120903244e-05, 'epoch': 0.06}) +[2026-06-24 12:14:10,944][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.39570802450180054, 'train/grad_norm_pre_clip': 48.195281982421875, 'timing/data': 0.00031553092412650585, 'timing/model': 1.0765464848373085, 'learning_rate/qwen_vl_interface': 9.534806075443619e-06, 'learning_rate/action_model': 9.534806075443619e-05, 'epoch': 0.06}) +[2026-06-24 12:14:12,017][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.2573184072971344, 'train/grad_norm_pre_clip': 30.570425033569336, 'timing/data': 0.00030597392469644547, 'timing/model': 1.0718328019138426, 'learning_rate/qwen_vl_interface': 9.533153249619853e-06, 'learning_rate/action_model': 9.533153249619853e-05, 'epoch': 0.06}) +[2026-06-24 12:14:13,095][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.2151133120059967, 'train/grad_norm_pre_clip': 49.92946243286133, 'timing/data': 0.000312174903228879, 'timing/model': 1.0763837308622897, 'learning_rate/qwen_vl_interface': 9.531497644504444e-06, 'learning_rate/action_model': 9.531497644504445e-05, 'epoch': 0.06}) +[2026-06-24 12:14:14,171][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.16569960117340088, 'train/grad_norm_pre_clip': 27.54080581665039, 'timing/data': 0.00031046103686094284, 'timing/model': 1.0738638888578862, 'learning_rate/qwen_vl_interface': 9.529839261171698e-06, 'learning_rate/action_model': 9.529839261171696e-05, 'epoch': 0.06}) +[2026-06-24 12:14:15,248][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.29487738013267517, 'train/grad_norm_pre_clip': 60.68818283081055, 'timing/data': 0.00029465206898748875, 'timing/model': 1.0757237949874252, 'learning_rate/qwen_vl_interface': 9.52817810069772e-06, 'learning_rate/action_model': 9.52817810069772e-05, 'epoch': 0.06}) +[2026-06-24 12:14:16,324][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.21757034957408905, 'train/grad_norm_pre_clip': 34.7945556640625, 'timing/data': 0.00033608986996114254, 'timing/model': 1.0749287949874997, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.06}) +[2026-06-24 12:14:17,402][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.22992943227291107, 'train/grad_norm_pre_clip': 15.908456802368164, 'timing/data': 0.0003345371223986149, 'timing/model': 1.0758603310678154, 'learning_rate/qwen_vl_interface': 9.524847452639509e-06, 'learning_rate/action_model': 9.524847452639509e-05, 'epoch': 0.06}) +[2026-06-24 12:14:18,476][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.23215551674365997, 'train/grad_norm_pre_clip': 28.40279197692871, 'timing/data': 0.00038183596916496754, 'timing/model': 1.0726363800931722, 'learning_rate/qwen_vl_interface': 9.523177967216496e-06, 'learning_rate/action_model': 9.523177967216496e-05, 'epoch': 0.06}) +[2026-06-24 12:14:19,557][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.30784446001052856, 'train/grad_norm_pre_clip': 22.771394729614258, 'timing/data': 0.0005280191544443369, 'timing/model': 1.0793096648994833, 'learning_rate/qwen_vl_interface': 9.521505708974694e-06, 'learning_rate/action_model': 9.521505708974694e-05, 'epoch': 0.06}) +[2026-06-24 12:14:20,633][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.2312183529138565, 'train/grad_norm_pre_clip': 30.994184494018555, 'timing/data': 0.0004991148598492146, 'timing/model': 1.0734515180811286, 'learning_rate/qwen_vl_interface': 9.519830678999214e-06, 'learning_rate/action_model': 9.519830678999213e-05, 'epoch': 0.06}) +[2026-06-24 12:14:21,711][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.05972236022353172, 'train/grad_norm_pre_clip': 7.927695274353027, 'timing/data': 0.0003054419066756964, 'timing/model': 1.0760842019226402, 'learning_rate/qwen_vl_interface': 9.518152878376961e-06, 'learning_rate/action_model': 9.51815287837696e-05, 'epoch': 0.06}) +[2026-06-24 12:14:22,788][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.11588934063911438, 'train/grad_norm_pre_clip': 19.512170791625977, 'timing/data': 0.00036498415283858776, 'timing/model': 1.0761396361049265, 'learning_rate/qwen_vl_interface': 9.516472308196646e-06, 'learning_rate/action_model': 9.516472308196645e-05, 'epoch': 0.06}) +[2026-06-24 12:14:23,868][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.28055766224861145, 'train/grad_norm_pre_clip': 33.13100814819336, 'timing/data': 0.0003483330365270376, 'timing/model': 1.0775124789215624, 'learning_rate/qwen_vl_interface': 9.51478896954877e-06, 'learning_rate/action_model': 9.51478896954877e-05, 'epoch': 0.06}) +[2026-06-24 12:14:25,214][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.16769637167453766, 'train/grad_norm_pre_clip': 36.85170364379883, 'timing/data': 0.0003255200572311878, 'timing/model': 1.3451736741699278, 'learning_rate/qwen_vl_interface': 9.51310286352563e-06, 'learning_rate/action_model': 9.51310286352563e-05, 'epoch': 0.06}) +[2026-06-24 12:14:26,293][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.14957420527935028, 'train/grad_norm_pre_clip': 21.298189163208008, 'timing/data': 0.0003558469470590353, 'timing/model': 1.0772300900425762, 'learning_rate/qwen_vl_interface': 9.51141399122133e-06, 'learning_rate/action_model': 9.511413991221331e-05, 'epoch': 0.06}) +[2026-06-24 12:14:27,367][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.44585931301116943, 'train/grad_norm_pre_clip': 115.51687622070312, 'timing/data': 0.00034273206256330013, 'timing/model': 1.0715237569529563, 'learning_rate/qwen_vl_interface': 9.509722353731757e-06, 'learning_rate/action_model': 9.509722353731756e-05, 'epoch': 0.06}) +[2026-06-24 12:14:28,446][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.52727210521698, 'train/grad_norm_pre_clip': 151.3824462890625, 'timing/data': 0.00029747700318694115, 'timing/model': 1.078352864831686, 'learning_rate/qwen_vl_interface': 9.508027952154595e-06, 'learning_rate/action_model': 9.508027952154593e-05, 'epoch': 0.06}) +[2026-06-24 12:14:29,520][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.22379961609840393, 'train/grad_norm_pre_clip': 85.08562469482422, 'timing/data': 0.0002955538220703602, 'timing/model': 1.0715798018500209, 'learning_rate/qwen_vl_interface': 9.506330787589321e-06, 'learning_rate/action_model': 9.506330787589322e-05, 'epoch': 0.06}) +[2026-06-24 12:14:30,596][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.33864423632621765, 'train/grad_norm_pre_clip': 100.1983871459961, 'timing/data': 0.0003074470441788435, 'timing/model': 1.0743832560256124, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.07}) +[2026-06-24 12:14:31,671][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.311924546957016, 'train/grad_norm_pre_clip': 52.97413635253906, 'timing/data': 0.0003183560911566019, 'timing/model': 1.074337510857731, 'learning_rate/qwen_vl_interface': 9.50292817390133e-06, 'learning_rate/action_model': 9.502928173901329e-05, 'epoch': 0.07}) +[2026-06-24 12:14:32,745][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.11882327497005463, 'train/grad_norm_pre_clip': 30.51088523864746, 'timing/data': 0.0003030381631106138, 'timing/model': 1.0717884639743716, 'learning_rate/qwen_vl_interface': 9.501222726986525e-06, 'learning_rate/action_model': 9.501222726986524e-05, 'epoch': 0.07}) +[2026-06-24 12:14:33,820][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.47791391611099243, 'train/grad_norm_pre_clip': 141.668212890625, 'timing/data': 0.0003112030681222677, 'timing/model': 1.0734238289296627, 'learning_rate/qwen_vl_interface': 9.49951452149945e-06, 'learning_rate/action_model': 9.499514521499451e-05, 'epoch': 0.07}) +[2026-06-24 12:14:34,897][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.059012025594711304, 'train/grad_norm_pre_clip': 25.067951202392578, 'timing/data': 0.00032472796738147736, 'timing/model': 1.0753331971354783, 'learning_rate/qwen_vl_interface': 9.497803558548539e-06, 'learning_rate/action_model': 9.497803558548539e-05, 'epoch': 0.07}) +[2026-06-24 12:14:35,975][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.13823427259922028, 'train/grad_norm_pre_clip': 55.81718063354492, 'timing/data': 0.0003075359854847193, 'timing/model': 1.076476511079818, 'learning_rate/qwen_vl_interface': 9.496089839244015e-06, 'learning_rate/action_model': 9.496089839244015e-05, 'epoch': 0.07}) +[2026-06-24 12:14:37,048][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.14463593065738678, 'train/grad_norm_pre_clip': 33.789608001708984, 'timing/data': 0.0003379939589649439, 'timing/model': 1.0713001401163638, 'learning_rate/qwen_vl_interface': 9.494373364697895e-06, 'learning_rate/action_model': 9.494373364697894e-05, 'epoch': 0.07}) +[2026-06-24 12:14:38,128][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.19302114844322205, 'train/grad_norm_pre_clip': 45.72141647338867, 'timing/data': 0.00037944107316434383, 'timing/model': 1.0715739228762686, 'learning_rate/qwen_vl_interface': 9.492654136023978e-06, 'learning_rate/action_model': 9.492654136023979e-05, 'epoch': 0.07}) +[2026-06-24 12:14:39,207][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.04792075976729393, 'train/grad_norm_pre_clip': 40.42304992675781, 'timing/data': 0.0003282749094069004, 'timing/model': 1.0776262648869306, 'learning_rate/qwen_vl_interface': 9.490932154337855e-06, 'learning_rate/action_model': 9.490932154337855e-05, 'epoch': 0.07}) +[2026-06-24 12:14:40,279][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.20411454141139984, 'train/grad_norm_pre_clip': 43.84425735473633, 'timing/data': 0.0003867950290441513, 'timing/model': 1.0709554951172322, 'learning_rate/qwen_vl_interface': 9.4892074207569e-06, 'learning_rate/action_model': 9.489207420756899e-05, 'epoch': 0.07}) +[2026-06-24 12:14:41,360][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.2900894284248352, 'train/grad_norm_pre_clip': 51.48858642578125, 'timing/data': 0.00031938892789185047, 'timing/model': 1.0785286568570882, 'learning_rate/qwen_vl_interface': 9.487479936400274e-06, 'learning_rate/action_model': 9.487479936400272e-05, 'epoch': 0.07}) +[2026-06-24 12:14:42,711][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.19079968333244324, 'train/grad_norm_pre_clip': 66.15187072753906, 'timing/data': 0.0003000630531460047, 'timing/model': 1.3497765948995948, 'learning_rate/qwen_vl_interface': 9.485749702388923e-06, 'learning_rate/action_model': 9.485749702388923e-05, 'epoch': 0.07}) +[2026-06-24 12:14:43,787][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.13761046528816223, 'train/grad_norm_pre_clip': 35.31607437133789, 'timing/data': 0.0003175749443471432, 'timing/model': 1.0745982329826802, 'learning_rate/qwen_vl_interface': 9.484016719845576e-06, 'learning_rate/action_model': 9.484016719845576e-05, 'epoch': 0.07}) +[2026-06-24 12:14:44,867][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.4438784718513489, 'train/grad_norm_pre_clip': 89.3053207397461, 'timing/data': 0.00031719496473670006, 'timing/model': 1.078367656096816, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.07}) +[2026-06-24 12:14:45,944][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.10285744071006775, 'train/grad_norm_pre_clip': 39.9372444152832, 'timing/data': 0.00035095703788101673, 'timing/model': 1.0754471430554986, 'learning_rate/qwen_vl_interface': 9.480542513662734e-06, 'learning_rate/action_model': 9.480542513662734e-05, 'epoch': 0.07}) +[2026-06-24 12:14:47,022][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.06795154511928558, 'train/grad_norm_pre_clip': 26.16599464416504, 'timing/data': 0.0003340761177241802, 'timing/model': 1.0759395060595125, 'learning_rate/qwen_vl_interface': 9.478801292277617e-06, 'learning_rate/action_model': 9.478801292277616e-05, 'epoch': 0.07}) +[2026-06-24 12:14:48,101][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.10699977725744247, 'train/grad_norm_pre_clip': 25.28062629699707, 'timing/data': 0.0003357790410518646, 'timing/model': 1.0770731319207698, 'learning_rate/qwen_vl_interface': 9.477057326869249e-06, 'learning_rate/action_model': 9.477057326869249e-05, 'epoch': 0.07}) +[2026-06-24 12:14:49,176][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.17718440294265747, 'train/grad_norm_pre_clip': 56.66829299926758, 'timing/data': 0.00032273493707180023, 'timing/model': 1.0737394040916115, 'learning_rate/qwen_vl_interface': 9.475310618569274e-06, 'learning_rate/action_model': 9.475310618569274e-05, 'epoch': 0.07}) +[2026-06-24 12:14:50,254][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.08548370748758316, 'train/grad_norm_pre_clip': 28.607744216918945, 'timing/data': 0.00032152305357158184, 'timing/model': 1.0767156819347292, 'learning_rate/qwen_vl_interface': 9.473561168511115e-06, 'learning_rate/action_model': 9.473561168511115e-05, 'epoch': 0.07}) +[2026-06-24 12:14:51,333][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.125527024269104, 'train/grad_norm_pre_clip': 37.8646354675293, 'timing/data': 0.00029347999952733517, 'timing/model': 1.0772371999919415, 'learning_rate/qwen_vl_interface': 9.471808977829967e-06, 'learning_rate/action_model': 9.471808977829967e-05, 'epoch': 0.07}) +[2026-06-24 12:14:52,414][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.24990713596343994, 'train/grad_norm_pre_clip': 63.87681579589844, 'timing/data': 0.0003355490043759346, 'timing/model': 1.0794870459940284, 'learning_rate/qwen_vl_interface': 9.470054047662808e-06, 'learning_rate/action_model': 9.470054047662808e-05, 'epoch': 0.07}) +[2026-06-24 12:14:53,489][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.3677007257938385, 'train/grad_norm_pre_clip': 54.96213150024414, 'timing/data': 0.0005044050049036741, 'timing/model': 1.072798231150955, 'learning_rate/qwen_vl_interface': 9.468296379148394e-06, 'learning_rate/action_model': 9.468296379148393e-05, 'epoch': 0.07}) +[2026-06-24 12:14:54,572][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.10533921420574188, 'train/grad_norm_pre_clip': 37.678592681884766, 'timing/data': 0.0003564779181033373, 'timing/model': 1.0813888108823448, 'learning_rate/qwen_vl_interface': 9.466535973427259e-06, 'learning_rate/action_model': 9.466535973427258e-05, 'epoch': 0.07}) +[2026-06-24 12:14:55,651][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.271533340215683, 'train/grad_norm_pre_clip': 58.19450759887695, 'timing/data': 0.000492502935230732, 'timing/model': 1.0768174140248448, 'learning_rate/qwen_vl_interface': 9.464772831641706e-06, 'learning_rate/action_model': 9.464772831641707e-05, 'epoch': 0.07}) +[2026-06-24 12:14:56,730][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.14964450895786285, 'train/grad_norm_pre_clip': 42.483421325683594, 'timing/data': 0.00031404895707964897, 'timing/model': 1.0773092559538782, 'learning_rate/qwen_vl_interface': 9.463006954935823e-06, 'learning_rate/action_model': 9.463006954935822e-05, 'epoch': 0.07}) +[2026-06-24 12:14:57,808][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.5805729627609253, 'train/grad_norm_pre_clip': 72.70499420166016, 'timing/data': 0.00030758604407310486, 'timing/model': 1.0762106080073863, 'learning_rate/qwen_vl_interface': 9.461238344455468e-06, 'learning_rate/action_model': 9.461238344455468e-05, 'epoch': 0.07}) +[2026-06-24 12:14:58,887][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.4810316264629364, 'train/grad_norm_pre_clip': 79.58084869384766, 'timing/data': 0.0005052969790995121, 'timing/model': 1.0771082299761474, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.07}) +[2026-06-24 12:15:00,237][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.19205287098884583, 'train/grad_norm_pre_clip': 51.96263885498047, 'timing/data': 0.0004042969085276127, 'timing/model': 1.3492346238344908, 'learning_rate/qwen_vl_interface': 9.457692926763638e-06, 'learning_rate/action_model': 9.457692926763638e-05, 'epoch': 0.07}) +[2026-06-24 12:15:01,316][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.16107137501239777, 'train/grad_norm_pre_clip': 18.78888511657715, 'timing/data': 0.0005131720099598169, 'timing/model': 1.076750700129196, 'learning_rate/qwen_vl_interface': 9.455916121852748e-06, 'learning_rate/action_model': 9.455916121852747e-05, 'epoch': 0.07}) +[2026-06-24 12:15:02,394][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.12145125865936279, 'train/grad_norm_pre_clip': 19.724218368530273, 'timing/data': 0.00030969991348683834, 'timing/model': 1.076024608919397, 'learning_rate/qwen_vl_interface': 9.454136587768549e-06, 'learning_rate/action_model': 9.454136587768548e-05, 'epoch': 0.07}) +[2026-06-24 12:15:03,473][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.36266952753067017, 'train/grad_norm_pre_clip': 46.399600982666016, 'timing/data': 0.0003079769667237997, 'timing/model': 1.077304719015956, 'learning_rate/qwen_vl_interface': 9.452354325665763e-06, 'learning_rate/action_model': 9.452354325665763e-05, 'epoch': 0.07}) +[2026-06-24 12:15:04,564][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.12859973311424255, 'train/grad_norm_pre_clip': 27.17205238342285, 'timing/data': 0.00032280408777296543, 'timing/model': 1.0898727611638606, 'learning_rate/qwen_vl_interface': 9.450569336700878e-06, 'learning_rate/action_model': 9.450569336700878e-05, 'epoch': 0.07}) +[2026-06-24 12:15:05,780][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.14250977337360382, 'train/grad_norm_pre_clip': 37.68497085571289, 'timing/data': 0.00033659092150628567, 'timing/model': 1.2145254409406334, 'learning_rate/qwen_vl_interface': 9.448781622032155e-06, 'learning_rate/action_model': 9.448781622032155e-05, 'epoch': 0.07}) +[2026-06-24 12:15:06,885][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.25903967022895813, 'train/grad_norm_pre_clip': 74.93540954589844, 'timing/data': 0.0003562280908226967, 'timing/model': 1.1027049999684095, 'learning_rate/qwen_vl_interface': 9.446991182819624e-06, 'learning_rate/action_model': 9.446991182819623e-05, 'epoch': 0.07}) +[2026-06-24 12:15:07,960][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.219654843211174, 'train/grad_norm_pre_clip': 24.201778411865234, 'timing/data': 0.0003422608133405447, 'timing/model': 1.073288092855364, 'learning_rate/qwen_vl_interface': 9.44519802022508e-06, 'learning_rate/action_model': 9.44519802022508e-05, 'epoch': 0.07}) +[2026-06-24 12:15:09,045][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.13698580861091614, 'train/grad_norm_pre_clip': 22.50853157043457, 'timing/data': 0.00032634101808071136, 'timing/model': 1.0832893250044435, 'learning_rate/qwen_vl_interface': 9.443402135412088e-06, 'learning_rate/action_model': 9.443402135412088e-05, 'epoch': 0.07}) +[2026-06-24 12:15:10,121][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.3757166564464569, 'train/grad_norm_pre_clip': 59.49143981933594, 'timing/data': 0.0003307489678263664, 'timing/model': 1.074756694957614, 'learning_rate/qwen_vl_interface': 9.441603529545977e-06, 'learning_rate/action_model': 9.441603529545976e-05, 'epoch': 0.07}) +[2026-06-24 12:15:11,199][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.1315610110759735, 'train/grad_norm_pre_clip': 30.724164962768555, 'timing/data': 0.000416190130636096, 'timing/model': 1.0760508209932595, 'learning_rate/qwen_vl_interface': 9.439802203793843e-06, 'learning_rate/action_model': 9.439802203793842e-05, 'epoch': 0.07}) +[2026-06-24 12:15:12,276][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.26897090673446655, 'train/grad_norm_pre_clip': 47.391666412353516, 'timing/data': 0.00033312407322227955, 'timing/model': 1.0750855621881783, 'learning_rate/qwen_vl_interface': 9.437998159324547e-06, 'learning_rate/action_model': 9.437998159324547e-05, 'epoch': 0.07}) +[2026-06-24 12:15:13,352][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.37565669417381287, 'train/grad_norm_pre_clip': 47.01066207885742, 'timing/data': 0.000355767086148262, 'timing/model': 1.0742128770798445, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.07}) +[2026-06-24 12:15:14,426][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.2963787615299225, 'train/grad_norm_pre_clip': 56.89879608154297, 'timing/data': 0.0003155518788844347, 'timing/model': 1.0727000921033323, 'learning_rate/qwen_vl_interface': 9.434381918918731e-06, 'learning_rate/action_model': 9.434381918918731e-05, 'epoch': 0.07}) +[2026-06-24 12:15:15,503][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.20139652490615845, 'train/grad_norm_pre_clip': 40.74191665649414, 'timing/data': 0.0003712060861289501, 'timing/model': 1.075851698871702, 'learning_rate/qwen_vl_interface': 9.432569725328753e-06, 'learning_rate/action_model': 9.432569725328753e-05, 'epoch': 0.07}) +[2026-06-24 12:15:16,583][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.2224133163690567, 'train/grad_norm_pre_clip': 31.38112449645996, 'timing/data': 0.0003001431468874216, 'timing/model': 1.0779724230524153, 'learning_rate/qwen_vl_interface': 9.430754817714686e-06, 'learning_rate/action_model': 9.430754817714687e-05, 'epoch': 0.07}) +[2026-06-24 12:15:17,920][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.3074835240840912, 'train/grad_norm_pre_clip': 56.875850677490234, 'timing/data': 0.00030201603658497334, 'timing/model': 1.3358816059771925, 'learning_rate/qwen_vl_interface': 9.42893719725421e-06, 'learning_rate/action_model': 9.42893719725421e-05, 'epoch': 0.07}) +[2026-06-24 12:15:19,000][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.1804897040128708, 'train/grad_norm_pre_clip': 15.380189895629883, 'timing/data': 0.00034872302785515785, 'timing/model': 1.0776467230170965, 'learning_rate/qwen_vl_interface': 9.427116865126757e-06, 'learning_rate/action_model': 9.427116865126756e-05, 'epoch': 0.07}) +[2026-06-24 12:15:20,080][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.1673116385936737, 'train/grad_norm_pre_clip': 27.61070442199707, 'timing/data': 0.000321763101965189, 'timing/model': 1.0787686661351472, 'learning_rate/qwen_vl_interface': 9.42529382251352e-06, 'learning_rate/action_model': 9.42529382251352e-05, 'epoch': 0.07}) +[2026-06-24 12:15:21,159][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.23042407631874084, 'train/grad_norm_pre_clip': 39.45886993408203, 'timing/data': 0.0003061140887439251, 'timing/model': 1.0769611690193415, 'learning_rate/qwen_vl_interface': 9.423468070597453e-06, 'learning_rate/action_model': 9.423468070597452e-05, 'epoch': 0.07}) +[2026-06-24 12:15:22,235][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.0654740259051323, 'train/grad_norm_pre_clip': 12.79898738861084, 'timing/data': 0.0003438950516283512, 'timing/model': 1.0750766471028328, 'learning_rate/qwen_vl_interface': 9.421639610563264e-06, 'learning_rate/action_model': 9.421639610563264e-05, 'epoch': 0.07}) +[2026-06-24 12:15:23,314][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.31799376010894775, 'train/grad_norm_pre_clip': 43.31328201293945, 'timing/data': 0.00029738806188106537, 'timing/model': 1.0769370838534087, 'learning_rate/qwen_vl_interface': 9.419808443597422e-06, 'learning_rate/action_model': 9.419808443597423e-05, 'epoch': 0.07}) +[2026-06-24 12:15:24,394][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.15205596387386322, 'train/grad_norm_pre_clip': 56.19940948486328, 'timing/data': 0.0003327140584588051, 'timing/model': 1.0786070670001209, 'learning_rate/qwen_vl_interface': 9.417974570888152e-06, 'learning_rate/action_model': 9.417974570888153e-05, 'epoch': 0.07}) +[2026-06-24 12:15:25,474][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.13322453200817108, 'train/grad_norm_pre_clip': 23.656497955322266, 'timing/data': 0.000394538976252079, 'timing/model': 1.0777963241562247, 'learning_rate/qwen_vl_interface': 9.416137993625437e-06, 'learning_rate/action_model': 9.416137993625437e-05, 'epoch': 0.07}) +[2026-06-24 12:15:26,548][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.1434670388698578, 'train/grad_norm_pre_clip': 26.08812141418457, 'timing/data': 0.00030854903161525726, 'timing/model': 1.0723616909235716, 'learning_rate/qwen_vl_interface': 9.414298713001007e-06, 'learning_rate/action_model': 9.414298713001006e-05, 'epoch': 0.07}) +[2026-06-24 12:15:27,629][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.14589141309261322, 'train/grad_norm_pre_clip': 38.50334167480469, 'timing/data': 0.0003210210707038641, 'timing/model': 1.0798947880975902, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.07}) +[2026-06-24 12:15:28,711][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.12601308524608612, 'train/grad_norm_pre_clip': 20.595800399780273, 'timing/data': 0.0003296569921076298, 'timing/model': 1.0797695328947157, 'learning_rate/qwen_vl_interface': 9.410612046442719e-06, 'learning_rate/action_model': 9.410612046442718e-05, 'epoch': 0.07}) +[2026-06-24 12:15:29,790][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.10558559000492096, 'train/grad_norm_pre_clip': 25.368215560913086, 'timing/data': 0.0003353189677000046, 'timing/model': 1.0780125609599054, 'learning_rate/qwen_vl_interface': 9.408764662901097e-06, 'learning_rate/action_model': 9.408764662901098e-05, 'epoch': 0.07}) +[2026-06-24 12:15:30,867][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.09766221791505814, 'train/grad_norm_pre_clip': 39.09424591064453, 'timing/data': 0.00030153500847518444, 'timing/model': 1.0749609631020576, 'learning_rate/qwen_vl_interface': 9.406914580782238e-06, 'learning_rate/action_model': 9.406914580782238e-05, 'epoch': 0.07}) +[2026-06-24 12:15:31,944][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.28657448291778564, 'train/grad_norm_pre_clip': 50.99586486816406, 'timing/data': 0.0003103320486843586, 'timing/model': 1.0754067790694535, 'learning_rate/qwen_vl_interface': 9.405061801286638e-06, 'learning_rate/action_model': 9.405061801286638e-05, 'epoch': 0.07}) +[2026-06-24 12:15:33,026][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.021868115290999413, 'train/grad_norm_pre_clip': 8.596887588500977, 'timing/data': 0.0003021170850843191, 'timing/model': 1.0804086320567876, 'learning_rate/qwen_vl_interface': 9.403206325616546e-06, 'learning_rate/action_model': 9.403206325616545e-05, 'epoch': 0.07}) +[2026-06-24 12:15:34,105][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.1819201111793518, 'train/grad_norm_pre_clip': 41.37468719482422, 'timing/data': 0.0003232050221413374, 'timing/model': 1.0771540328860283, 'learning_rate/qwen_vl_interface': 9.40134815497596e-06, 'learning_rate/action_model': 9.40134815497596e-05, 'epoch': 0.07}) +[2026-06-24 12:15:35,444][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.08775591850280762, 'train/grad_norm_pre_clip': 38.53840637207031, 'timing/data': 0.0003166431561112404, 'timing/model': 1.3375959531404078, 'learning_rate/qwen_vl_interface': 9.399487290570629e-06, 'learning_rate/action_model': 9.399487290570629e-05, 'epoch': 0.07}) +[2026-06-24 12:15:36,527][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.216226264834404, 'train/grad_norm_pre_clip': 75.97468566894531, 'timing/data': 0.0003122049383819103, 'timing/model': 1.0816322260070592, 'learning_rate/qwen_vl_interface': 9.397623733608042e-06, 'learning_rate/action_model': 9.397623733608042e-05, 'epoch': 0.07}) +[2026-06-24 12:15:37,604][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.07627136260271072, 'train/grad_norm_pre_clip': 39.04301071166992, 'timing/data': 0.0002940308768302202, 'timing/model': 1.0756418709643185, 'learning_rate/qwen_vl_interface': 9.395757485297447e-06, 'learning_rate/action_model': 9.395757485297446e-05, 'epoch': 0.07}) +[2026-06-24 12:15:38,685][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.170271098613739, 'train/grad_norm_pre_clip': 78.77108764648438, 'timing/data': 0.00034032692201435566, 'timing/model': 1.0789295108988881, 'learning_rate/qwen_vl_interface': 9.393888546849828e-06, 'learning_rate/action_model': 9.393888546849828e-05, 'epoch': 0.07}) +[2026-06-24 12:15:39,765][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.1933857649564743, 'train/grad_norm_pre_clip': 84.86280822753906, 'timing/data': 0.00030323793180286884, 'timing/model': 1.0782684129662812, 'learning_rate/qwen_vl_interface': 9.392016919477922e-06, 'learning_rate/action_model': 9.392016919477921e-05, 'epoch': 0.07}) +[2026-06-24 12:15:40,845][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.22264085710048676, 'train/grad_norm_pre_clip': 103.13600158691406, 'timing/data': 0.00032435799948871136, 'timing/model': 1.0788434811402112, 'learning_rate/qwen_vl_interface': 9.390142604396206e-06, 'learning_rate/action_model': 9.390142604396205e-05, 'epoch': 0.07}) +[2026-06-24 12:15:41,924][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.28674525022506714, 'train/grad_norm_pre_clip': 77.11312103271484, 'timing/data': 0.00031865714117884636, 'timing/model': 1.0774153950624168, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.07}) +[2026-06-24 12:15:43,004][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.021622905507683754, 'train/grad_norm_pre_clip': 14.88137435913086, 'timing/data': 0.00031435885466635227, 'timing/model': 1.0781451619695872, 'learning_rate/qwen_vl_interface': 9.386385915969975e-06, 'learning_rate/action_model': 9.386385915969976e-05, 'epoch': 0.07}) +[2026-06-24 12:15:44,082][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.18771779537200928, 'train/grad_norm_pre_clip': 60.76164627075195, 'timing/data': 0.00030597392469644547, 'timing/model': 1.07716997503303, 'learning_rate/qwen_vl_interface': 9.384503545063138e-06, 'learning_rate/action_model': 9.384503545063137e-05, 'epoch': 0.07}) +[2026-06-24 12:15:45,162][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.3033309280872345, 'train/grad_norm_pre_clip': 59.213706970214844, 'timing/data': 0.00030554295517504215, 'timing/model': 1.078352551907301, 'learning_rate/qwen_vl_interface': 9.382618491321837e-06, 'learning_rate/action_model': 9.382618491321837e-05, 'epoch': 0.07}) +[2026-06-24 12:15:46,244][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.06574937701225281, 'train/grad_norm_pre_clip': 36.71468734741211, 'timing/data': 0.0005565132014453411, 'timing/model': 1.0793111079838127, 'learning_rate/qwen_vl_interface': 9.380730755969265e-06, 'learning_rate/action_model': 9.380730755969264e-05, 'epoch': 0.07}) +[2026-06-24 12:15:47,325][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.2021401822566986, 'train/grad_norm_pre_clip': 60.42796325683594, 'timing/data': 0.00037211691960692406, 'timing/model': 1.0796626380179077, 'learning_rate/qwen_vl_interface': 9.378840340230353e-06, 'learning_rate/action_model': 9.378840340230352e-05, 'epoch': 0.07}) +[2026-06-24 12:15:48,404][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.06655511260032654, 'train/grad_norm_pre_clip': 24.49106788635254, 'timing/data': 0.0003133069258183241, 'timing/model': 1.0768955720122904, 'learning_rate/qwen_vl_interface': 9.376947245331769e-06, 'learning_rate/action_model': 9.376947245331767e-05, 'epoch': 0.07}) +[2026-06-24 12:15:49,486][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.1715884506702423, 'train/grad_norm_pre_clip': 81.16084289550781, 'timing/data': 0.0003262721002101898, 'timing/model': 1.0812486598733813, 'learning_rate/qwen_vl_interface': 9.375051472501922e-06, 'learning_rate/action_model': 9.375051472501923e-05, 'epoch': 0.07}) +[2026-06-24 12:15:50,565][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.1329905092716217, 'train/grad_norm_pre_clip': 55.60049819946289, 'timing/data': 0.00031025195494294167, 'timing/model': 1.0767046748660505, 'learning_rate/qwen_vl_interface': 9.373153022970961e-06, 'learning_rate/action_model': 9.373153022970961e-05, 'epoch': 0.07}) +[2026-06-24 12:15:51,644][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.13050240278244019, 'train/grad_norm_pre_clip': 43.852569580078125, 'timing/data': 0.00029848888516426086, 'timing/model': 1.0779016080778092, 'learning_rate/qwen_vl_interface': 9.371251897970768e-06, 'learning_rate/action_model': 9.371251897970768e-05, 'epoch': 0.07}) +[2026-06-24 12:15:52,989][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.05142736807465553, 'train/grad_norm_pre_clip': 18.36248016357422, 'timing/data': 0.00032100104726850986, 'timing/model': 1.3437077209819108, 'learning_rate/qwen_vl_interface': 9.369348098734961e-06, 'learning_rate/action_model': 9.369348098734961e-05, 'epoch': 0.07}) +[2026-06-24 12:15:54,072][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.1274096965789795, 'train/grad_norm_pre_clip': 38.56299591064453, 'timing/data': 0.0003076060675084591, 'timing/model': 1.0805470871273428, 'learning_rate/qwen_vl_interface': 9.367441626498897e-06, 'learning_rate/action_model': 9.367441626498896e-05, 'epoch': 0.07}) +[2026-06-24 12:15:55,152][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.20673702657222748, 'train/grad_norm_pre_clip': 49.676090240478516, 'timing/data': 0.0003464592155069113, 'timing/model': 1.0786018709186465, 'learning_rate/qwen_vl_interface': 9.365532482499663e-06, 'learning_rate/action_model': 9.365532482499662e-05, 'epoch': 0.07}) +[2026-06-24 12:15:56,233][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 750 +[2026-06-24 12:16:21,417][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 750 +[2026-06-24 12:16:21,418][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 750 +[2026-06-24 12:16:54,685][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 750 +[2026-06-24 12:18:06,475][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 750 +[2026-06-24 12:18:06,475][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.22879092395305634, 'train/grad_norm_pre_clip': 30.295063018798828, 'eval/action_loss/seconds': 25.182354893069714, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.12303699553012848, 'eval/latency_2/loss': 0.12303699553012848, 'eval/flappy/loss': 0.12303699553012848, 'eval/flappy/latency_2/loss': 0.12303699553012848, 'eval/flap_precision': 0.8436754176610979, 'eval/flap_recall': 0.9592944369063772, 'eval/flap_f1': 0.8977777777777778, 'eval/flap_rate_model': 0.23277777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8977777777777777, 'eval/latency_2/flap_precision': 0.8436754176610979, 'eval/latency_2/flap_recall': 0.9592944369063772, 'eval/latency_2/flap_f1': 0.8977777777777778, 'eval/latency_2/flap_rate_model': 0.23277777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8977777777777777, 'eval/action_classification/data_wait_seconds': 0.3399023413658142, 'eval/action_classification/predict_seconds': 32.83481979370117, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 51.259999999999664, 'rl_games_eval/mid_train/mean_length': 431.8, 'rl_games_eval/mid_train/std_reward': 39.57012004025212, 'rl_games_eval/mid_train/std_length': 320.79239392479366, 'rl_games_eval/mid_train/macro_mean_reward': 51.259999999999664, 'rl_games_eval/mid_train/macro_mean_length': 431.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 51.259999999999664, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 431.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 39.57012004025212, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 320.79239392479366, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 51.259999999999664, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003211209550499916, 'timing/model': 1.0795224860776216, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.07}) +[2026-06-24 12:18:07,534][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.14408916234970093, 'train/grad_norm_pre_clip': 215.91024780273438, 'timing/data': 0.0003615568857640028, 'timing/model': 1.0563300370704383, 'learning_rate/qwen_vl_interface': 9.361706184168714e-06, 'learning_rate/action_model': 9.361706184168713e-05, 'epoch': 0.07}) +[2026-06-24 12:18:08,599][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.03559008240699768, 'train/grad_norm_pre_clip': 12.351115226745605, 'timing/data': 0.00034935493022203445, 'timing/model': 1.0637684781104326, 'learning_rate/qwen_vl_interface': 9.359789032319843e-06, 'learning_rate/action_model': 9.359789032319842e-05, 'epoch': 0.07}) +[2026-06-24 12:18:09,664][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.2950391471385956, 'train/grad_norm_pre_clip': 40.63882064819336, 'timing/data': 0.00034767109900712967, 'timing/model': 1.0639332861173898, 'learning_rate/qwen_vl_interface': 9.357869213673484e-06, 'learning_rate/action_model': 9.357869213673485e-05, 'epoch': 0.07}) +[2026-06-24 12:18:10,730][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.1757625937461853, 'train/grad_norm_pre_clip': 27.787351608276367, 'timing/data': 0.0003293878398835659, 'timing/model': 1.0641967908013612, 'learning_rate/qwen_vl_interface': 9.355946729475393e-06, 'learning_rate/action_model': 9.355946729475393e-05, 'epoch': 0.07}) +[2026-06-24 12:18:11,795][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.30060508847236633, 'train/grad_norm_pre_clip': 87.3160629272461, 'timing/data': 0.00031841592863202095, 'timing/model': 1.063115163007751, 'learning_rate/qwen_vl_interface': 9.354021580973049e-06, 'learning_rate/action_model': 9.354021580973048e-05, 'epoch': 0.07}) +[2026-06-24 12:18:13,096][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.20814259350299835, 'train/grad_norm_pre_clip': 38.05388259887695, 'timing/data': 0.00034962501376867294, 'timing/model': 1.2999081418383867, 'learning_rate/qwen_vl_interface': 9.352093769415655e-06, 'learning_rate/action_model': 9.352093769415655e-05, 'epoch': 0.07}) +[2026-06-24 12:18:14,164][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.240381121635437, 'train/grad_norm_pre_clip': 39.626895904541016, 'timing/data': 0.000334177166223526, 'timing/model': 1.0655649250838906, 'learning_rate/qwen_vl_interface': 9.35016329605415e-06, 'learning_rate/action_model': 9.350163296054151e-05, 'epoch': 0.07}) +[2026-06-24 12:18:15,231][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.3274807035923004, 'train/grad_norm_pre_clip': 65.27718353271484, 'timing/data': 0.00032231397926807404, 'timing/model': 1.066089658997953, 'learning_rate/qwen_vl_interface': 9.348230162141198e-06, 'learning_rate/action_model': 9.348230162141198e-05, 'epoch': 0.07}) +[2026-06-24 12:18:16,301][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.3938504159450531, 'train/grad_norm_pre_clip': 140.02088928222656, 'timing/data': 0.0006806659512221813, 'timing/model': 1.0675918441265821, 'learning_rate/qwen_vl_interface': 9.346294368931188e-06, 'learning_rate/action_model': 9.346294368931188e-05, 'epoch': 0.07}) +[2026-06-24 12:18:17,368][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.09546986222267151, 'train/grad_norm_pre_clip': 22.626754760742188, 'timing/data': 0.0003279650118201971, 'timing/model': 1.0654096249490976, 'learning_rate/qwen_vl_interface': 9.344355917680235e-06, 'learning_rate/action_model': 9.344355917680235e-05, 'epoch': 0.07}) +[2026-06-24 12:18:18,439][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.27319151163101196, 'train/grad_norm_pre_clip': 42.60564422607422, 'timing/data': 0.0003326029982417822, 'timing/model': 1.0689710080623627, 'learning_rate/qwen_vl_interface': 9.342414809646179e-06, 'learning_rate/action_model': 9.342414809646179e-05, 'epoch': 0.07}) +[2026-06-24 12:18:19,516][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.22656017541885376, 'train/grad_norm_pre_clip': 34.03994369506836, 'timing/data': 0.0003892299719154835, 'timing/model': 1.0746591091156006, 'learning_rate/qwen_vl_interface': 9.340471046088581e-06, 'learning_rate/action_model': 9.34047104608858e-05, 'epoch': 0.07}) +[2026-06-24 12:18:20,586][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.2135927975177765, 'train/grad_norm_pre_clip': 39.82102966308594, 'timing/data': 0.0003708149306476116, 'timing/model': 1.0690777590498328, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.07}) +[2026-06-24 12:18:21,657][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.31518301367759705, 'train/grad_norm_pre_clip': 56.482147216796875, 'timing/data': 0.00032592006027698517, 'timing/model': 1.0692167191300541, 'learning_rate/qwen_vl_interface': 9.336575557449641e-06, 'learning_rate/action_model': 9.336575557449641e-05, 'epoch': 0.07}) +[2026-06-24 12:18:22,727][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.2387225329875946, 'train/grad_norm_pre_clip': 43.089115142822266, 'timing/data': 0.0003201100043952465, 'timing/model': 1.0677846760954708, 'learning_rate/qwen_vl_interface': 9.334623834896039e-06, 'learning_rate/action_model': 9.334623834896038e-05, 'epoch': 0.07}) +[2026-06-24 12:18:23,798][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.18297483026981354, 'train/grad_norm_pre_clip': 24.430952072143555, 'timing/data': 0.0005239918828010559, 'timing/model': 1.069505721097812, 'learning_rate/qwen_vl_interface': 9.332669461874376e-06, 'learning_rate/action_model': 9.332669461874374e-05, 'epoch': 0.07}) +[2026-06-24 12:18:24,876][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.5365508794784546, 'train/grad_norm_pre_clip': 92.34713745117188, 'timing/data': 0.000395060982555151, 'timing/model': 1.0767517311032861, 'learning_rate/qwen_vl_interface': 9.330712439652824e-06, 'learning_rate/action_model': 9.330712439652823e-05, 'epoch': 0.07}) +[2026-06-24 12:18:25,952][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.3825129270553589, 'train/grad_norm_pre_clip': 79.95288848876953, 'timing/data': 0.0003262918908149004, 'timing/model': 1.0742406852077693, 'learning_rate/qwen_vl_interface': 9.328752769501272e-06, 'learning_rate/action_model': 9.328752769501272e-05, 'epoch': 0.07}) +[2026-06-24 12:18:27,023][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.2683265805244446, 'train/grad_norm_pre_clip': 59.67052459716797, 'timing/data': 0.00039422791451215744, 'timing/model': 1.068936264840886, 'learning_rate/qwen_vl_interface': 9.326790452691332e-06, 'learning_rate/action_model': 9.326790452691332e-05, 'epoch': 0.07}) +[2026-06-24 12:18:28,098][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.274519145488739, 'train/grad_norm_pre_clip': 41.78730010986328, 'timing/data': 0.0003333049826323986, 'timing/model': 1.0739116889890283, 'learning_rate/qwen_vl_interface': 9.324825490496327e-06, 'learning_rate/action_model': 9.324825490496327e-05, 'epoch': 0.07}) +[2026-06-24 12:18:29,173][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.2303604930639267, 'train/grad_norm_pre_clip': 30.955049514770508, 'timing/data': 0.00033642002381384373, 'timing/model': 1.0730491830036044, 'learning_rate/qwen_vl_interface': 9.322857884191303e-06, 'learning_rate/action_model': 9.322857884191302e-05, 'epoch': 0.07}) +[2026-06-24 12:18:30,482][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.12797512114048004, 'train/grad_norm_pre_clip': 19.465526580810547, 'timing/data': 0.00039247493259608746, 'timing/model': 1.3068300569429994, 'learning_rate/qwen_vl_interface': 9.320887635053013e-06, 'learning_rate/action_model': 9.320887635053014e-05, 'epoch': 0.07}) +[2026-06-24 12:18:31,553][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.31902071833610535, 'train/grad_norm_pre_clip': 46.388301849365234, 'timing/data': 0.0003147199749946594, 'timing/model': 1.0699569778516889, 'learning_rate/qwen_vl_interface': 9.318914744359936e-06, 'learning_rate/action_model': 9.318914744359937e-05, 'epoch': 0.07}) +[2026-06-24 12:18:32,634][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.2720620036125183, 'train/grad_norm_pre_clip': 48.306617736816406, 'timing/data': 0.0003907021600753069, 'timing/model': 1.0790939710568637, 'learning_rate/qwen_vl_interface': 9.316939213392257e-06, 'learning_rate/action_model': 9.316939213392257e-05, 'epoch': 0.07}) +[2026-06-24 12:18:33,706][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.34700584411621094, 'train/grad_norm_pre_clip': 67.24081420898438, 'timing/data': 0.0004121719393879175, 'timing/model': 1.0698404100257903, 'learning_rate/qwen_vl_interface': 9.314961043431877e-06, 'learning_rate/action_model': 9.314961043431876e-05, 'epoch': 0.07}) +[2026-06-24 12:18:34,786][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.1843070387840271, 'train/grad_norm_pre_clip': 31.11812400817871, 'timing/data': 0.00038756709545850754, 'timing/model': 1.0783052339684218, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.08}) +[2026-06-24 12:18:35,863][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.2340872436761856, 'train/grad_norm_pre_clip': 34.027835845947266, 'timing/data': 0.00037349993363022804, 'timing/model': 1.0756869481410831, 'learning_rate/qwen_vl_interface': 9.310996791669175e-06, 'learning_rate/action_model': 9.310996791669175e-05, 'epoch': 0.08}) +[2026-06-24 12:18:36,939][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.13736103475093842, 'train/grad_norm_pre_clip': 39.64374542236328, 'timing/data': 0.00037904991768300533, 'timing/model': 1.0741410609334707, 'learning_rate/qwen_vl_interface': 9.309010712439213e-06, 'learning_rate/action_model': 9.309010712439213e-05, 'epoch': 0.08}) +[2026-06-24 12:18:38,014][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.23483598232269287, 'train/grad_norm_pre_clip': 35.60029983520508, 'timing/data': 0.00037853908725082874, 'timing/model': 1.0736426969524473, 'learning_rate/qwen_vl_interface': 9.307021999361267e-06, 'learning_rate/action_model': 9.307021999361267e-05, 'epoch': 0.08}) +[2026-06-24 12:18:39,093][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.2264482080936432, 'train/grad_norm_pre_clip': 45.17033767700195, 'timing/data': 0.00034006801433861256, 'timing/model': 1.0767320359591395, 'learning_rate/qwen_vl_interface': 9.305030653725794e-06, 'learning_rate/action_model': 9.305030653725794e-05, 'epoch': 0.08}) +[2026-06-24 12:18:40,168][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.3417872190475464, 'train/grad_norm_pre_clip': 62.40877914428711, 'timing/data': 0.00032762414775788784, 'timing/model': 1.0735563149210066, 'learning_rate/qwen_vl_interface': 9.303036676824954e-06, 'learning_rate/action_model': 9.303036676824953e-05, 'epoch': 0.08}) +[2026-06-24 12:18:41,246][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.3089597821235657, 'train/grad_norm_pre_clip': 63.29170227050781, 'timing/data': 0.00037281797267496586, 'timing/model': 1.07646768190898, 'learning_rate/qwen_vl_interface': 9.301040069952616e-06, 'learning_rate/action_model': 9.301040069952616e-05, 'epoch': 0.08}) +[2026-06-24 12:18:42,325][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.20907247066497803, 'train/grad_norm_pre_clip': 27.758108139038086, 'timing/data': 0.000561312073841691, 'timing/model': 1.0763434581458569, 'learning_rate/qwen_vl_interface': 9.299040834404362e-06, 'learning_rate/action_model': 9.299040834404361e-05, 'epoch': 0.08}) +[2026-06-24 12:18:43,407][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.3223849833011627, 'train/grad_norm_pre_clip': 35.159393310546875, 'timing/data': 0.0003158219624310732, 'timing/model': 1.0807123160921037, 'learning_rate/qwen_vl_interface': 9.297038971477468e-06, 'learning_rate/action_model': 9.297038971477466e-05, 'epoch': 0.08}) +[2026-06-24 12:18:44,487][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.3050561845302582, 'train/grad_norm_pre_clip': 56.496036529541016, 'timing/data': 0.0003533021081238985, 'timing/model': 1.078542860923335, 'learning_rate/qwen_vl_interface': 9.295034482470922e-06, 'learning_rate/action_model': 9.295034482470923e-05, 'epoch': 0.08}) +[2026-06-24 12:18:45,566][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.19042432308197021, 'train/grad_norm_pre_clip': 35.33993911743164, 'timing/data': 0.0003404079470783472, 'timing/model': 1.0771924899891019, 'learning_rate/qwen_vl_interface': 9.293027368685418e-06, 'learning_rate/action_model': 9.293027368685419e-05, 'epoch': 0.08}) +[2026-06-24 12:18:46,645][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.21965134143829346, 'train/grad_norm_pre_clip': 31.572433471679688, 'timing/data': 0.00032884604297578335, 'timing/model': 1.0766400350257754, 'learning_rate/qwen_vl_interface': 9.29101763142335e-06, 'learning_rate/action_model': 9.29101763142335e-05, 'epoch': 0.08}) +[2026-06-24 12:18:47,957][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.2592321038246155, 'train/grad_norm_pre_clip': 45.71551513671875, 'timing/data': 0.00032173306681215763, 'timing/model': 1.3104265299625695, 'learning_rate/qwen_vl_interface': 9.28900527198881e-06, 'learning_rate/action_model': 9.289005271988811e-05, 'epoch': 0.08}) +[2026-06-24 12:18:49,034][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.11907516419887543, 'train/grad_norm_pre_clip': 14.513845443725586, 'timing/data': 0.0003426019102334976, 'timing/model': 1.0758549538441002, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.08}) +[2026-06-24 12:18:50,113][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.1446584165096283, 'train/grad_norm_pre_clip': 15.997905731201172, 'timing/data': 0.00029825884848833084, 'timing/model': 1.0770708229392767, 'learning_rate/qwen_vl_interface': 9.284972691827226e-06, 'learning_rate/action_model': 9.284972691827224e-05, 'epoch': 0.08}) +[2026-06-24 12:18:51,188][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.19011515378952026, 'train/grad_norm_pre_clip': 53.25749206542969, 'timing/data': 0.0003446259070187807, 'timing/model': 1.072850463911891, 'learning_rate/qwen_vl_interface': 9.282952473716873e-06, 'learning_rate/action_model': 9.282952473716874e-05, 'epoch': 0.08}) +[2026-06-24 12:18:52,273][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.19004496932029724, 'train/grad_norm_pre_clip': 31.275537490844727, 'timing/data': 0.000327554065734148, 'timing/model': 1.0836113300174475, 'learning_rate/qwen_vl_interface': 9.280929638667447e-06, 'learning_rate/action_model': 9.280929638667447e-05, 'epoch': 0.08}) +[2026-06-24 12:18:53,351][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.2786697745323181, 'train/grad_norm_pre_clip': 55.030433654785156, 'timing/data': 0.0004270109347999096, 'timing/model': 1.0769965439103544, 'learning_rate/qwen_vl_interface': 9.278904187991542e-06, 'learning_rate/action_model': 9.278904187991541e-05, 'epoch': 0.08}) +[2026-06-24 12:18:54,427][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.2885719835758209, 'train/grad_norm_pre_clip': 54.372344970703125, 'timing/data': 0.00032648188062012196, 'timing/model': 1.0738523020409048, 'learning_rate/qwen_vl_interface': 9.27687612300345e-06, 'learning_rate/action_model': 9.27687612300345e-05, 'epoch': 0.08}) +[2026-06-24 12:18:55,503][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.41449877619743347, 'train/grad_norm_pre_clip': 104.42144012451172, 'timing/data': 0.00030550197698175907, 'timing/model': 1.0745075698941946, 'learning_rate/qwen_vl_interface': 9.27484544501916e-06, 'learning_rate/action_model': 9.274845445019159e-05, 'epoch': 0.08}) +[2026-06-24 12:18:56,589][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.11322538554668427, 'train/grad_norm_pre_clip': 33.57847595214844, 'timing/data': 0.0003892600070685148, 'timing/model': 1.084064971189946, 'learning_rate/qwen_vl_interface': 9.27281215535636e-06, 'learning_rate/action_model': 9.272812155356358e-05, 'epoch': 0.08}) +[2026-06-24 12:18:57,669][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.08382410556077957, 'train/grad_norm_pre_clip': 13.785669326782227, 'timing/data': 0.00034850300289690495, 'timing/model': 1.0788030209951103, 'learning_rate/qwen_vl_interface': 9.270776255334425e-06, 'learning_rate/action_model': 9.270776255334425e-05, 'epoch': 0.08}) +[2026-06-24 12:18:58,752][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.08752172440290451, 'train/grad_norm_pre_clip': 17.478269577026367, 'timing/data': 0.0003590129781514406, 'timing/model': 1.0811505098827183, 'learning_rate/qwen_vl_interface': 9.268737746274432e-06, 'learning_rate/action_model': 9.268737746274432e-05, 'epoch': 0.08}) +[2026-06-24 12:18:59,832][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.24692502617835999, 'train/grad_norm_pre_clip': 53.32673645019531, 'timing/data': 0.000314278993755579, 'timing/model': 1.077645461075008, 'learning_rate/qwen_vl_interface': 9.266696629499146e-06, 'learning_rate/action_model': 9.266696629499145e-05, 'epoch': 0.08}) +[2026-06-24 12:19:00,914][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.4751993417739868, 'train/grad_norm_pre_clip': 99.01728820800781, 'timing/data': 0.00030426005832850933, 'timing/model': 1.0804158828686923, 'learning_rate/qwen_vl_interface': 9.264652906333026e-06, 'learning_rate/action_model': 9.264652906333026e-05, 'epoch': 0.08}) +[2026-06-24 12:19:01,994][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.17849689722061157, 'train/grad_norm_pre_clip': 29.00688362121582, 'timing/data': 0.00037997192703187466, 'timing/model': 1.078805745812133, 'learning_rate/qwen_vl_interface': 9.26260657810222e-06, 'learning_rate/action_model': 9.262606578102221e-05, 'epoch': 0.08}) +[2026-06-24 12:19:03,072][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.35775697231292725, 'train/grad_norm_pre_clip': 76.89574432373047, 'timing/data': 0.00033957604318857193, 'timing/model': 1.0757422449532896, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.08}) +[2026-06-24 12:19:04,150][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.08580267429351807, 'train/grad_norm_pre_clip': 20.43202018737793, 'timing/data': 0.0003312109038233757, 'timing/model': 1.0767178831156343, 'learning_rate/qwen_vl_interface': 9.258506111759608e-06, 'learning_rate/action_model': 9.258506111759609e-05, 'epoch': 0.08}) +[2026-06-24 12:19:05,462][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.15646865963935852, 'train/grad_norm_pre_clip': 42.494781494140625, 'timing/data': 0.000345227075740695, 'timing/model': 1.3107048850506544, 'learning_rate/qwen_vl_interface': 9.256451976308548e-06, 'learning_rate/action_model': 9.25645197630855e-05, 'epoch': 0.08}) +[2026-06-24 12:19:06,537][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.17473730444908142, 'train/grad_norm_pre_clip': 74.65545654296875, 'timing/data': 0.0003308500163257122, 'timing/model': 1.07309063593857, 'learning_rate/qwen_vl_interface': 9.2543952411143e-06, 'learning_rate/action_model': 9.254395241114299e-05, 'epoch': 0.08}) +[2026-06-24 12:19:07,616][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.15564066171646118, 'train/grad_norm_pre_clip': 25.139556884765625, 'timing/data': 0.0003391350619494915, 'timing/model': 1.0771681568585336, 'learning_rate/qwen_vl_interface': 9.252335907511452e-06, 'learning_rate/action_model': 9.252335907511453e-05, 'epoch': 0.08}) +[2026-06-24 12:19:08,694][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.16429701447486877, 'train/grad_norm_pre_clip': 35.79572677612305, 'timing/data': 0.0003133269492536783, 'timing/model': 1.0765332079026848, 'learning_rate/qwen_vl_interface': 9.250273976836289e-06, 'learning_rate/action_model': 9.250273976836289e-05, 'epoch': 0.08}) +[2026-06-24 12:19:09,772][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.15885142982006073, 'train/grad_norm_pre_clip': 35.025455474853516, 'timing/data': 0.0003110028337687254, 'timing/model': 1.0762510688509792, 'learning_rate/qwen_vl_interface': 9.248209450426772e-06, 'learning_rate/action_model': 9.248209450426772e-05, 'epoch': 0.08}) +[2026-06-24 12:19:10,853][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.26835593581199646, 'train/grad_norm_pre_clip': 34.92403793334961, 'timing/data': 0.0004977730568498373, 'timing/model': 1.0789632520172745, 'learning_rate/qwen_vl_interface': 9.24614232962255e-06, 'learning_rate/action_model': 9.24614232962255e-05, 'epoch': 0.08}) +[2026-06-24 12:19:11,934][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.23005881905555725, 'train/grad_norm_pre_clip': 31.930606842041016, 'timing/data': 0.00031381798908114433, 'timing/model': 1.0798552639316767, 'learning_rate/qwen_vl_interface': 9.244072615764957e-06, 'learning_rate/action_model': 9.244072615764956e-05, 'epoch': 0.08}) +[2026-06-24 12:19:13,017][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.16232386231422424, 'train/grad_norm_pre_clip': 80.4303970336914, 'timing/data': 0.0003856830298900604, 'timing/model': 1.0808155441191047, 'learning_rate/qwen_vl_interface': 9.242000310197005e-06, 'learning_rate/action_model': 9.242000310197004e-05, 'epoch': 0.08}) +[2026-06-24 12:19:14,097][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.21967051923274994, 'train/grad_norm_pre_clip': 35.661746978759766, 'timing/data': 0.0003457278944551945, 'timing/model': 1.0792305839713663, 'learning_rate/qwen_vl_interface': 9.239925414263392e-06, 'learning_rate/action_model': 9.239925414263392e-05, 'epoch': 0.08}) +[2026-06-24 12:19:15,177][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.2811373770236969, 'train/grad_norm_pre_clip': 29.0987548828125, 'timing/data': 0.0003141281194984913, 'timing/model': 1.0781283879186958, 'learning_rate/qwen_vl_interface': 9.237847929310497e-06, 'learning_rate/action_model': 9.237847929310496e-05, 'epoch': 0.08}) +[2026-06-24 12:19:16,262][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.22042205929756165, 'train/grad_norm_pre_clip': 29.409440994262695, 'timing/data': 0.00036342116072773933, 'timing/model': 1.0825049800332636, 'learning_rate/qwen_vl_interface': 9.235767856686375e-06, 'learning_rate/action_model': 9.235767856686375e-05, 'epoch': 0.08}) +[2026-06-24 12:19:17,341][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.20288227498531342, 'train/grad_norm_pre_clip': 26.791807174682617, 'timing/data': 0.00046826712787151337, 'timing/model': 1.0772327810991555, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.08}) +[2026-06-24 12:19:18,419][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.3053660988807678, 'train/grad_norm_pre_clip': 39.12030792236328, 'timing/data': 0.00030617392621934414, 'timing/model': 1.076991117093712, 'learning_rate/qwen_vl_interface': 9.23159995382508e-06, 'learning_rate/action_model': 9.231599953825079e-05, 'epoch': 0.08}) +[2026-06-24 12:19:19,501][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.16418065130710602, 'train/grad_norm_pre_clip': 17.609636306762695, 'timing/data': 0.0005403219256550074, 'timing/model': 1.0798056819476187, 'learning_rate/qwen_vl_interface': 9.22951212629241e-06, 'learning_rate/action_model': 9.229512126292411e-05, 'epoch': 0.08}) +[2026-06-24 12:19:20,578][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.17433185875415802, 'train/grad_norm_pre_clip': 26.86966896057129, 'timing/data': 0.0003351380582898855, 'timing/model': 1.0748777170665562, 'learning_rate/qwen_vl_interface': 9.227421716497531e-06, 'learning_rate/action_model': 9.227421716497531e-05, 'epoch': 0.08}) +[2026-06-24 12:19:21,657][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.10518988966941833, 'train/grad_norm_pre_clip': 19.1162166595459, 'timing/data': 0.000527628930285573, 'timing/model': 1.0771427920553833, 'learning_rate/qwen_vl_interface': 9.225328725796879e-06, 'learning_rate/action_model': 9.22532872579688e-05, 'epoch': 0.08}) +[2026-06-24 12:19:22,965][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.17898240685462952, 'train/grad_norm_pre_clip': 22.955211639404297, 'timing/data': 0.00033452711068093777, 'timing/model': 1.3068113918416202, 'learning_rate/qwen_vl_interface': 9.223233155548579e-06, 'learning_rate/action_model': 9.223233155548578e-05, 'epoch': 0.08}) +[2026-06-24 12:19:24,042][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.2210739254951477, 'train/grad_norm_pre_clip': 26.655569076538086, 'timing/data': 0.00037698703818023205, 'timing/model': 1.0751157840713859, 'learning_rate/qwen_vl_interface': 9.22113500711242e-06, 'learning_rate/action_model': 9.221135007112419e-05, 'epoch': 0.08}) +[2026-06-24 12:19:25,122][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.14249935746192932, 'train/grad_norm_pre_clip': 20.317150115966797, 'timing/data': 0.00033523794263601303, 'timing/model': 1.0783188669010997, 'learning_rate/qwen_vl_interface': 9.219034281849867e-06, 'learning_rate/action_model': 9.219034281849867e-05, 'epoch': 0.08}) +[2026-06-24 12:19:26,208][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.1269877552986145, 'train/grad_norm_pre_clip': 22.388534545898438, 'timing/data': 0.00034946505911648273, 'timing/model': 1.0844001239165664, 'learning_rate/qwen_vl_interface': 9.216930981124062e-06, 'learning_rate/action_model': 9.216930981124062e-05, 'epoch': 0.08}) +[2026-06-24 12:19:27,285][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.09064829349517822, 'train/grad_norm_pre_clip': 24.997629165649414, 'timing/data': 0.0004917220212519169, 'timing/model': 1.075059057911858, 'learning_rate/qwen_vl_interface': 9.214825106299808e-06, 'learning_rate/action_model': 9.214825106299808e-05, 'epoch': 0.08}) +[2026-06-24 12:19:28,369][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.13626757264137268, 'train/grad_norm_pre_clip': 26.625167846679688, 'timing/data': 0.0003845400642603636, 'timing/model': 1.0818899390287697, 'learning_rate/qwen_vl_interface': 9.212716658743592e-06, 'learning_rate/action_model': 9.212716658743592e-05, 'epoch': 0.08}) +[2026-06-24 12:19:29,454][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.3347318172454834, 'train/grad_norm_pre_clip': 87.8972396850586, 'timing/data': 0.00042367400601506233, 'timing/model': 1.083563034888357, 'learning_rate/qwen_vl_interface': 9.210605639823558e-06, 'learning_rate/action_model': 9.210605639823557e-05, 'epoch': 0.08}) +[2026-06-24 12:19:30,535][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.20419135689735413, 'train/grad_norm_pre_clip': 51.07905960083008, 'timing/data': 0.00039659393951296806, 'timing/model': 1.0787865940947086, 'learning_rate/qwen_vl_interface': 9.208492050909523e-06, 'learning_rate/action_model': 9.208492050909523e-05, 'epoch': 0.08}) +[2026-06-24 12:19:31,620][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.4919685125350952, 'train/grad_norm_pre_clip': 105.7047119140625, 'timing/data': 0.0004629280883818865, 'timing/model': 1.0834614650812, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.08}) +[2026-06-24 12:19:32,697][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.25851938128471375, 'train/grad_norm_pre_clip': 90.72187805175781, 'timing/data': 0.0003495849668979645, 'timing/model': 1.075094164116308, 'learning_rate/qwen_vl_interface': 9.204257168587061e-06, 'learning_rate/action_model': 9.204257168587061e-05, 'epoch': 0.08}) +[2026-06-24 12:19:33,779][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.229948490858078, 'train/grad_norm_pre_clip': 36.84457778930664, 'timing/data': 0.00046644406393170357, 'timing/model': 1.080074837896973, 'learning_rate/qwen_vl_interface': 9.2021358779266e-06, 'learning_rate/action_model': 9.202135877926601e-05, 'epoch': 0.08}) +[2026-06-24 12:19:34,857][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.030073381960392, 'train/grad_norm_pre_clip': 7.935194492340088, 'timing/data': 0.0004079448990523815, 'timing/model': 1.0763498381711543, 'learning_rate/qwen_vl_interface': 9.200012022768077e-06, 'learning_rate/action_model': 9.200012022768077e-05, 'epoch': 0.08}) +[2026-06-24 12:19:35,938][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.10470554977655411, 'train/grad_norm_pre_clip': 38.89351272583008, 'timing/data': 0.0003219840582460165, 'timing/model': 1.0793541979510337, 'learning_rate/qwen_vl_interface': 9.19788560448964e-06, 'learning_rate/action_model': 9.197885604489639e-05, 'epoch': 0.08}) +[2026-06-24 12:19:37,023][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.15252988040447235, 'train/grad_norm_pre_clip': 37.757572174072266, 'timing/data': 0.0003547042142599821, 'timing/model': 1.0836602088529617, 'learning_rate/qwen_vl_interface': 9.195756624471094e-06, 'learning_rate/action_model': 9.195756624471094e-05, 'epoch': 0.08}) +[2026-06-24 12:19:38,103][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.029004409909248352, 'train/grad_norm_pre_clip': 10.007473945617676, 'timing/data': 0.0004067718982696533, 'timing/model': 1.0776147390715778, 'learning_rate/qwen_vl_interface': 9.193625084093912e-06, 'learning_rate/action_model': 9.193625084093911e-05, 'epoch': 0.08}) +[2026-06-24 12:19:39,181][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.15054048597812653, 'train/grad_norm_pre_clip': 48.859317779541016, 'timing/data': 0.00032757408916950226, 'timing/model': 1.0766446820925921, 'learning_rate/qwen_vl_interface': 9.191490984741227e-06, 'learning_rate/action_model': 9.191490984741227e-05, 'epoch': 0.08}) +[2026-06-24 12:19:40,482][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.23205994069576263, 'train/grad_norm_pre_clip': 50.745025634765625, 'timing/data': 0.0003326029982417822, 'timing/model': 1.299729202175513, 'learning_rate/qwen_vl_interface': 9.189354327797835e-06, 'learning_rate/action_model': 9.189354327797836e-05, 'epoch': 0.08}) +[2026-06-24 12:19:41,558][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.12577968835830688, 'train/grad_norm_pre_clip': 55.036468505859375, 'timing/data': 0.00031874701380729675, 'timing/model': 1.0743753788992763, 'learning_rate/qwen_vl_interface': 9.18721511465019e-06, 'learning_rate/action_model': 9.187215114650191e-05, 'epoch': 0.08}) +[2026-06-24 12:19:42,642][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.06904453784227371, 'train/grad_norm_pre_clip': 34.29292678833008, 'timing/data': 0.0003134768921881914, 'timing/model': 1.0817522529978305, 'learning_rate/qwen_vl_interface': 9.185073346686402e-06, 'learning_rate/action_model': 9.185073346686401e-05, 'epoch': 0.08}) +[2026-06-24 12:19:43,720][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.1431398093700409, 'train/grad_norm_pre_clip': 30.15802764892578, 'timing/data': 0.00033188192173838615, 'timing/model': 1.0766128120012581, 'learning_rate/qwen_vl_interface': 9.182929025296242e-06, 'learning_rate/action_model': 9.182929025296242e-05, 'epoch': 0.08}) +[2026-06-24 12:19:44,797][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.036193329840898514, 'train/grad_norm_pre_clip': 32.39313507080078, 'timing/data': 0.0003370919730514288, 'timing/model': 1.0752975770737976, 'learning_rate/qwen_vl_interface': 9.180782151871138e-06, 'learning_rate/action_model': 9.180782151871137e-05, 'epoch': 0.08}) +[2026-06-24 12:19:45,877][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.07585499435663223, 'train/grad_norm_pre_clip': 39.334957122802734, 'timing/data': 0.00033210311084985733, 'timing/model': 1.0787949010264128, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.08}) +[2026-06-24 12:19:46,956][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.45826277136802673, 'train/grad_norm_pre_clip': 58.86851501464844, 'timing/data': 0.00032084109261631966, 'timing/model': 1.077597538009286, 'learning_rate/qwen_vl_interface': 9.176480754490078e-06, 'learning_rate/action_model': 9.176480754490078e-05, 'epoch': 0.08}) +[2026-06-24 12:19:48,036][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.13714201748371124, 'train/grad_norm_pre_clip': 47.70465087890625, 'timing/data': 0.00031887786462903023, 'timing/model': 1.0774555921088904, 'learning_rate/qwen_vl_interface': 9.174326233325253e-06, 'learning_rate/action_model': 9.174326233325253e-05, 'epoch': 0.08}) +[2026-06-24 12:19:49,114][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.11180330812931061, 'train/grad_norm_pre_clip': 31.939476013183594, 'timing/data': 0.0003114731516689062, 'timing/model': 1.0766928030643612, 'learning_rate/qwen_vl_interface': 9.17216916570774e-06, 'learning_rate/action_model': 9.172169165707741e-05, 'epoch': 0.08}) +[2026-06-24 12:19:50,191][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.22592857480049133, 'train/grad_norm_pre_clip': 60.194313049316406, 'timing/data': 0.0003260709345340729, 'timing/model': 1.0753764249384403, 'learning_rate/qwen_vl_interface': 9.170009553037239e-06, 'learning_rate/action_model': 9.170009553037238e-05, 'epoch': 0.08}) +[2026-06-24 12:19:51,269][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.06634171307086945, 'train/grad_norm_pre_clip': 28.454833984375, 'timing/data': 0.00033915601670742035, 'timing/model': 1.076555254869163, 'learning_rate/qwen_vl_interface': 9.167847396715098e-06, 'learning_rate/action_model': 9.167847396715098e-05, 'epoch': 0.08}) +[2026-06-24 12:19:52,349][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.6924809217453003, 'train/grad_norm_pre_clip': 139.78587341308594, 'timing/data': 0.00039718393236398697, 'timing/model': 1.078646593960002, 'learning_rate/qwen_vl_interface': 9.165682698144317e-06, 'learning_rate/action_model': 9.165682698144317e-05, 'epoch': 0.08}) +[2026-06-24 12:19:53,432][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.07402923703193665, 'train/grad_norm_pre_clip': 46.79011154174805, 'timing/data': 0.00033355504274368286, 'timing/model': 1.081226518843323, 'learning_rate/qwen_vl_interface': 9.163515458729545e-06, 'learning_rate/action_model': 9.163515458729545e-05, 'epoch': 0.08}) +[2026-06-24 12:19:54,517][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.09802903234958649, 'train/grad_norm_pre_clip': 34.69119644165039, 'timing/data': 0.0003844010643661022, 'timing/model': 1.0832715521100909, 'learning_rate/qwen_vl_interface': 9.16134567987708e-06, 'learning_rate/action_model': 9.16134567987708e-05, 'epoch': 0.08}) +[2026-06-24 12:19:55,597][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.18035483360290527, 'train/grad_norm_pre_clip': 42.29526901245117, 'timing/data': 0.00033118994906544685, 'timing/model': 1.0781047597993165, 'learning_rate/qwen_vl_interface': 9.159173362994868e-06, 'learning_rate/action_model': 9.159173362994869e-05, 'epoch': 0.08}) +[2026-06-24 12:19:56,678][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.18752384185791016, 'train/grad_norm_pre_clip': 21.058116912841797, 'timing/data': 0.0003181658685207367, 'timing/model': 1.0790592178236693, 'learning_rate/qwen_vl_interface': 9.156998509492503e-06, 'learning_rate/action_model': 9.156998509492503e-05, 'epoch': 0.08}) +[2026-06-24 12:19:57,976][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.08210185915231705, 'train/grad_norm_pre_clip': 22.490142822265625, 'timing/data': 0.00032738293521106243, 'timing/model': 1.2968205930665135, 'learning_rate/qwen_vl_interface': 9.154821120781221e-06, 'learning_rate/action_model': 9.154821120781222e-05, 'epoch': 0.08}) +[2026-06-24 12:19:59,053][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.08746594190597534, 'train/grad_norm_pre_clip': 13.568718910217285, 'timing/data': 0.00044987304136157036, 'timing/model': 1.0749053941108286, 'learning_rate/qwen_vl_interface': 9.152641198273911e-06, 'learning_rate/action_model': 9.15264119827391e-05, 'epoch': 0.08}) +[2026-06-24 12:20:00,136][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.3407648801803589, 'train/grad_norm_pre_clip': 56.531612396240234, 'timing/data': 0.00032527907751500607, 'timing/model': 1.08159922901541, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.08}) +[2026-06-24 12:20:01,220][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.17021778225898743, 'train/grad_norm_pre_clip': 31.480382919311523, 'timing/data': 0.00030955998227000237, 'timing/model': 1.0822596061043441, 'learning_rate/qwen_vl_interface': 9.148273757530948e-06, 'learning_rate/action_model': 9.148273757530948e-05, 'epoch': 0.08}) +[2026-06-24 12:20:02,308][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.43296757340431213, 'train/grad_norm_pre_clip': 64.20707702636719, 'timing/data': 0.00044558499939739704, 'timing/model': 1.0862340650055557, 'learning_rate/qwen_vl_interface': 9.146086242129283e-06, 'learning_rate/action_model': 9.146086242129282e-05, 'epoch': 0.08}) +[2026-06-24 12:20:03,392][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.05142855644226074, 'train/grad_norm_pre_clip': 9.70344066619873, 'timing/data': 0.0004670959897339344, 'timing/model': 1.0817036649677902, 'learning_rate/qwen_vl_interface': 9.143896198599553e-06, 'learning_rate/action_model': 9.143896198599553e-05, 'epoch': 0.08}) +[2026-06-24 12:20:04,473][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.2435714304447174, 'train/grad_norm_pre_clip': 50.836978912353516, 'timing/data': 0.0005105570890009403, 'timing/model': 1.0789465489797294, 'learning_rate/qwen_vl_interface': 9.141703628362858e-06, 'learning_rate/action_model': 9.141703628362858e-05, 'epoch': 0.08}) +[2026-06-24 12:20:05,558][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.2851485311985016, 'train/grad_norm_pre_clip': 74.59508514404297, 'timing/data': 0.00038793706335127354, 'timing/model': 1.0831389860250056, 'learning_rate/qwen_vl_interface': 9.139508532841928e-06, 'learning_rate/action_model': 9.139508532841927e-05, 'epoch': 0.08}) +[2026-06-24 12:20:06,643][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.2433789223432541, 'train/grad_norm_pre_clip': 35.7010612487793, 'timing/data': 0.0003907519858330488, 'timing/model': 1.0837244829162955, 'learning_rate/qwen_vl_interface': 9.137310913461139e-06, 'learning_rate/action_model': 9.137310913461139e-05, 'epoch': 0.08}) +[2026-06-24 12:20:07,729][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.2612738013267517, 'train/grad_norm_pre_clip': 42.37482833862305, 'timing/data': 0.00038534216582775116, 'timing/model': 1.0844470360316336, 'learning_rate/qwen_vl_interface': 9.135110771646502e-06, 'learning_rate/action_model': 9.135110771646502e-05, 'epoch': 0.08}) +[2026-06-24 12:20:08,810][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.16712188720703125, 'train/grad_norm_pre_clip': 27.670581817626953, 'timing/data': 0.000315522076562047, 'timing/model': 1.079185085138306, 'learning_rate/qwen_vl_interface': 9.132908108825667e-06, 'learning_rate/action_model': 9.132908108825667e-05, 'epoch': 0.08}) +[2026-06-24 12:20:09,887][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.33336037397384644, 'train/grad_norm_pre_clip': 63.711700439453125, 'timing/data': 0.0003412498626857996, 'timing/model': 1.0757235288619995, 'learning_rate/qwen_vl_interface': 9.130702926427915e-06, 'learning_rate/action_model': 9.130702926427914e-05, 'epoch': 0.08}) +[2026-06-24 12:20:10,971][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.13470877707004547, 'train/grad_norm_pre_clip': 15.539813995361328, 'timing/data': 0.00029728678055107594, 'timing/model': 1.0815259930677712, 'learning_rate/qwen_vl_interface': 9.128495225884167e-06, 'learning_rate/action_model': 9.128495225884166e-05, 'epoch': 0.08}) +[2026-06-24 12:20:12,055][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.07775592803955078, 'train/grad_norm_pre_clip': 9.016661643981934, 'timing/data': 0.0003225139807909727, 'timing/model': 1.0829906268045306, 'learning_rate/qwen_vl_interface': 9.126285008626975e-06, 'learning_rate/action_model': 9.126285008626974e-05, 'epoch': 0.08}) +[2026-06-24 12:20:13,140][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.1768225133419037, 'train/grad_norm_pre_clip': 45.32493591308594, 'timing/data': 0.00031905784271657467, 'timing/model': 1.083244773093611, 'learning_rate/qwen_vl_interface': 9.124072276090524e-06, 'learning_rate/action_model': 9.124072276090524e-05, 'epoch': 0.08}) +[2026-06-24 12:20:14,226][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.20880767703056335, 'train/grad_norm_pre_clip': 34.91316223144531, 'timing/data': 0.00034256186336278915, 'timing/model': 1.0839664458762854, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.08}) +[2026-06-24 12:20:15,609][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.13641586899757385, 'train/grad_norm_pre_clip': 20.58536148071289, 'timing/data': 0.00038183480501174927, 'timing/model': 1.3816961268894374, 'learning_rate/qwen_vl_interface': 9.119639270924754e-06, 'learning_rate/action_model': 9.119639270924754e-05, 'epoch': 0.08}) +[2026-06-24 12:20:16,696][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.12302158027887344, 'train/grad_norm_pre_clip': 30.94552993774414, 'timing/data': 0.0003581319469958544, 'timing/model': 1.0844470378942788, 'learning_rate/qwen_vl_interface': 9.11741900117196e-06, 'learning_rate/action_model': 9.11741900117196e-05, 'epoch': 0.08}) +[2026-06-24 12:20:17,810][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.19553475081920624, 'train/grad_norm_pre_clip': 42.71676254272461, 'timing/data': 0.000422121025621891, 'timing/model': 1.1123056870419532, 'learning_rate/qwen_vl_interface': 9.115196221892969e-06, 'learning_rate/action_model': 9.115196221892969e-05, 'epoch': 0.08}) +[2026-06-24 12:20:18,895][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.14697620272636414, 'train/grad_norm_pre_clip': 38.094810485839844, 'timing/data': 0.0005115990061312914, 'timing/model': 1.0829438909422606, 'learning_rate/qwen_vl_interface': 9.112970934530111e-06, 'learning_rate/action_model': 9.112970934530111e-05, 'epoch': 0.08}) +[2026-06-24 12:20:19,978][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.10027779638767242, 'train/grad_norm_pre_clip': 27.021575927734375, 'timing/data': 0.0004464569501578808, 'timing/model': 1.0811796251218766, 'learning_rate/qwen_vl_interface': 9.110743140527353e-06, 'learning_rate/action_model': 9.110743140527352e-05, 'epoch': 0.08}) +[2026-06-24 12:20:21,062][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.3236510455608368, 'train/grad_norm_pre_clip': 84.34983825683594, 'timing/data': 0.00041397614404559135, 'timing/model': 1.0823901640251279, 'learning_rate/qwen_vl_interface': 9.108512841330288e-06, 'learning_rate/action_model': 9.108512841330287e-05, 'epoch': 0.08}) +[2026-06-24 12:20:22,145][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.31312763690948486, 'train/grad_norm_pre_clip': 63.34700393676758, 'timing/data': 0.0004038570914417505, 'timing/model': 1.0812943510245532, 'learning_rate/qwen_vl_interface': 9.106280038386129e-06, 'learning_rate/action_model': 9.10628003838613e-05, 'epoch': 0.08}) +[2026-06-24 12:20:23,230][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.19511500000953674, 'train/grad_norm_pre_clip': 44.34770965576172, 'timing/data': 0.0003702740650624037, 'timing/model': 1.0831132689490914, 'learning_rate/qwen_vl_interface': 9.104044733143723e-06, 'learning_rate/action_model': 9.104044733143723e-05, 'epoch': 0.08}) +[2026-06-24 12:20:24,314][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.17357972264289856, 'train/grad_norm_pre_clip': 35.95601272583008, 'timing/data': 0.00040008011274039745, 'timing/model': 1.0825067029800266, 'learning_rate/qwen_vl_interface': 9.10180692705353e-06, 'learning_rate/action_model': 9.10180692705353e-05, 'epoch': 0.08}) +[2026-06-24 12:20:25,397][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.09340403974056244, 'train/grad_norm_pre_clip': 27.136503219604492, 'timing/data': 0.00043180910870432854, 'timing/model': 1.080660092877224, 'learning_rate/qwen_vl_interface': 9.099566621567643e-06, 'learning_rate/action_model': 9.099566621567642e-05, 'epoch': 0.08}) +[2026-06-24 12:20:26,480][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.2316490262746811, 'train/grad_norm_pre_clip': 34.16109848022461, 'timing/data': 0.0004331418313086033, 'timing/model': 1.0815342210698873, 'learning_rate/qwen_vl_interface': 9.097323818139767e-06, 'learning_rate/action_model': 9.097323818139767e-05, 'epoch': 0.08}) +[2026-06-24 12:20:27,565][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.31671008467674255, 'train/grad_norm_pre_clip': 46.557373046875, 'timing/data': 0.00036725797690451145, 'timing/model': 1.0829145670868456, 'learning_rate/qwen_vl_interface': 9.095078518225237e-06, 'learning_rate/action_model': 9.095078518225236e-05, 'epoch': 0.09}) +[2026-06-24 12:20:28,655][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.14744533598423004, 'train/grad_norm_pre_clip': 29.67753028869629, 'timing/data': 0.0004609029274433851, 'timing/model': 1.088727992027998, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.09}) +[2026-06-24 12:20:29,743][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.07517524808645248, 'train/grad_norm_pre_clip': 17.934885025024414, 'timing/data': 0.0004120820667594671, 'timing/model': 1.0861655110493302, 'learning_rate/qwen_vl_interface': 9.09058043476563e-06, 'learning_rate/action_model': 9.09058043476563e-05, 'epoch': 0.09}) +[2026-06-24 12:20:30,834][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.28263771533966064, 'train/grad_norm_pre_clip': 61.06999588012695, 'timing/data': 0.00038114399649202824, 'timing/model': 1.0890077571384609, 'learning_rate/qwen_vl_interface': 9.088327654139311e-06, 'learning_rate/action_model': 9.08832765413931e-05, 'epoch': 0.09}) +[2026-06-24 12:20:32,133][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.02557721920311451, 'train/grad_norm_pre_clip': 8.637120246887207, 'timing/data': 0.0003892399836331606, 'timing/model': 1.2973812001291662, 'learning_rate/qwen_vl_interface': 9.08607238286385e-06, 'learning_rate/action_model': 9.08607238286385e-05, 'epoch': 0.09}) +[2026-06-24 12:20:33,217][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.08790121972560883, 'train/grad_norm_pre_clip': 23.354352951049805, 'timing/data': 0.00039265607483685017, 'timing/model': 1.0816416731104255, 'learning_rate/qwen_vl_interface': 9.083814622402667e-06, 'learning_rate/action_model': 9.083814622402666e-05, 'epoch': 0.09}) +[2026-06-24 12:20:34,303][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.23262837529182434, 'train/grad_norm_pre_clip': 44.408145904541016, 'timing/data': 0.00035966397263109684, 'timing/model': 1.0848263709340245, 'learning_rate/qwen_vl_interface': 9.0815543742208e-06, 'learning_rate/action_model': 9.081554374220799e-05, 'epoch': 0.09}) +[2026-06-24 12:20:35,394][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.1605348140001297, 'train/grad_norm_pre_clip': 43.45159912109375, 'timing/data': 0.00044356100261211395, 'timing/model': 1.0890342670027167, 'learning_rate/qwen_vl_interface': 9.079291639784897e-06, 'learning_rate/action_model': 9.079291639784896e-05, 'epoch': 0.09}) +[2026-06-24 12:20:36,477][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.11013560742139816, 'train/grad_norm_pre_clip': 41.34109115600586, 'timing/data': 0.00045367004349827766, 'timing/model': 1.0807091558817774, 'learning_rate/qwen_vl_interface': 9.077026420563223e-06, 'learning_rate/action_model': 9.077026420563223e-05, 'epoch': 0.09}) +[2026-06-24 12:20:37,568][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.13113978505134583, 'train/grad_norm_pre_clip': 29.09649085998535, 'timing/data': 0.0004480991046875715, 'timing/model': 1.0890068460721523, 'learning_rate/qwen_vl_interface': 9.074758718025652e-06, 'learning_rate/action_model': 9.074758718025652e-05, 'epoch': 0.09}) +[2026-06-24 12:20:38,656][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.09053069353103638, 'train/grad_norm_pre_clip': 21.182186126708984, 'timing/data': 0.0003909529186785221, 'timing/model': 1.086645821109414, 'learning_rate/qwen_vl_interface': 9.072488533643678e-06, 'learning_rate/action_model': 9.072488533643677e-05, 'epoch': 0.09}) +[2026-06-24 12:20:39,743][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.028469020500779152, 'train/grad_norm_pre_clip': 9.727063179016113, 'timing/data': 0.0003966230433434248, 'timing/model': 1.0850810389965773, 'learning_rate/qwen_vl_interface': 9.070215868890391e-06, 'learning_rate/action_model': 9.070215868890391e-05, 'epoch': 0.09}) +[2026-06-24 12:20:40,826][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.36708319187164307, 'train/grad_norm_pre_clip': 115.4019775390625, 'timing/data': 0.00042396411299705505, 'timing/model': 1.0816323559265584, 'learning_rate/qwen_vl_interface': 9.067940725240503e-06, 'learning_rate/action_model': 9.067940725240503e-05, 'epoch': 0.09}) +[2026-06-24 12:20:41,910][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.034558847546577454, 'train/grad_norm_pre_clip': 28.344449996948242, 'timing/data': 0.00036021508276462555, 'timing/model': 1.0821549049578607, 'learning_rate/qwen_vl_interface': 9.06566310417033e-06, 'learning_rate/action_model': 9.06566310417033e-05, 'epoch': 0.09}) +[2026-06-24 12:20:42,993][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.06718452274799347, 'train/grad_norm_pre_clip': 23.78891944885254, 'timing/data': 0.0003597040195018053, 'timing/model': 1.0806883568875492, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.09}) +[2026-06-24 12:20:44,076][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.23361749947071075, 'train/grad_norm_pre_clip': 55.54594802856445, 'timing/data': 0.0003849209751933813, 'timing/model': 1.0818967318627983, 'learning_rate/qwen_vl_interface': 9.061100435682429e-06, 'learning_rate/action_model': 9.061100435682429e-05, 'epoch': 0.09}) +[2026-06-24 12:20:45,159][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.17462888360023499, 'train/grad_norm_pre_clip': 33.73627853393555, 'timing/data': 0.00045556388795375824, 'timing/model': 1.0808755380567163, 'learning_rate/qwen_vl_interface': 9.058815391225363e-06, 'learning_rate/action_model': 9.058815391225362e-05, 'epoch': 0.09}) +[2026-06-24 12:20:46,246][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.20870022475719452, 'train/grad_norm_pre_clip': 69.9085464477539, 'timing/data': 0.00038433982990682125, 'timing/model': 1.0848701240029186, 'learning_rate/qwen_vl_interface': 9.056527875269344e-06, 'learning_rate/action_model': 9.056527875269344e-05, 'epoch': 0.09}) +[2026-06-24 12:20:47,332][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.12349924445152283, 'train/grad_norm_pre_clip': 27.24382209777832, 'timing/data': 0.00039313617162406445, 'timing/model': 1.084474422968924, 'learning_rate/qwen_vl_interface': 9.054237889298714e-06, 'learning_rate/action_model': 9.054237889298714e-05, 'epoch': 0.09}) +[2026-06-24 12:20:48,628][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.11931806802749634, 'train/grad_norm_pre_clip': 31.004831314086914, 'timing/data': 0.0003962430637329817, 'timing/model': 1.294365950860083, 'learning_rate/qwen_vl_interface': 9.05194543479942e-06, 'learning_rate/action_model': 9.05194543479942e-05, 'epoch': 0.09}) +[2026-06-24 12:20:49,710][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.1249256357550621, 'train/grad_norm_pre_clip': 21.360885620117188, 'timing/data': 0.00038231699727475643, 'timing/model': 1.0805901840794832, 'learning_rate/qwen_vl_interface': 9.049650513259007e-06, 'learning_rate/action_model': 9.049650513259008e-05, 'epoch': 0.09}) +[2026-06-24 12:20:50,791][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.6616916060447693, 'train/grad_norm_pre_clip': 80.78841400146484, 'timing/data': 0.00032596010714769363, 'timing/model': 1.07950891694054, 'learning_rate/qwen_vl_interface': 9.047353126166632e-06, 'learning_rate/action_model': 9.047353126166632e-05, 'epoch': 0.09}) +[2026-06-24 12:20:51,872][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.16682128608226776, 'train/grad_norm_pre_clip': 44.75825119018555, 'timing/data': 0.00035008578561246395, 'timing/model': 1.078659936087206, 'learning_rate/qwen_vl_interface': 9.045053275013038e-06, 'learning_rate/action_model': 9.045053275013038e-05, 'epoch': 0.09}) +[2026-06-24 12:20:52,956][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.39326828718185425, 'train/grad_norm_pre_clip': 42.488651275634766, 'timing/data': 0.0003372221253812313, 'timing/model': 1.0823090558405966, 'learning_rate/qwen_vl_interface': 9.042750961290578e-06, 'learning_rate/action_model': 9.042750961290578e-05, 'epoch': 0.09}) +[2026-06-24 12:20:54,033][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.18925902247428894, 'train/grad_norm_pre_clip': 49.34543228149414, 'timing/data': 0.00034983595833182335, 'timing/model': 1.0756281360518187, 'learning_rate/qwen_vl_interface': 9.040446186493195e-06, 'learning_rate/action_model': 9.040446186493194e-05, 'epoch': 0.09}) +[2026-06-24 12:20:55,115][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.22535744309425354, 'train/grad_norm_pre_clip': 31.320621490478516, 'timing/data': 0.00033616996370255947, 'timing/model': 1.080286055104807, 'learning_rate/qwen_vl_interface': 9.038138952116435e-06, 'learning_rate/action_model': 9.038138952116435e-05, 'epoch': 0.09}) +[2026-06-24 12:20:56,195][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.10908952355384827, 'train/grad_norm_pre_clip': 25.805463790893555, 'timing/data': 0.00034010782837867737, 'timing/model': 1.0791215628851205, 'learning_rate/qwen_vl_interface': 9.035829259657437e-06, 'learning_rate/action_model': 9.035829259657436e-05, 'epoch': 0.09}) +[2026-06-24 12:20:57,277][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.17875626683235168, 'train/grad_norm_pre_clip': 30.780019760131836, 'timing/data': 0.00033872504718601704, 'timing/model': 1.0796139149460942, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.09}) +[2026-06-24 12:20:58,357][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.16628067195415497, 'train/grad_norm_pre_clip': 27.597562789916992, 'timing/data': 0.0003135481383651495, 'timing/model': 1.0791764659807086, 'learning_rate/qwen_vl_interface': 9.031202506489254e-06, 'learning_rate/action_model': 9.031202506489253e-05, 'epoch': 0.09}) +[2026-06-24 12:20:59,443][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.15880627930164337, 'train/grad_norm_pre_clip': 23.677419662475586, 'timing/data': 0.000366956926882267, 'timing/model': 1.076599686872214, 'learning_rate/qwen_vl_interface': 9.02888544878232e-06, 'learning_rate/action_model': 9.028885448782319e-05, 'epoch': 0.09}) +[2026-06-24 12:21:00,524][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.1571749746799469, 'train/grad_norm_pre_clip': 21.318124771118164, 'timing/data': 0.0003910730592906475, 'timing/model': 1.079259310849011, 'learning_rate/qwen_vl_interface': 9.026565938997644e-06, 'learning_rate/action_model': 9.026565938997645e-05, 'epoch': 0.09}) +[2026-06-24 12:21:01,602][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.16051049530506134, 'train/grad_norm_pre_clip': 22.112207412719727, 'timing/data': 0.0004252470098435879, 'timing/model': 1.0765735879540443, 'learning_rate/qwen_vl_interface': 9.024243978640334e-06, 'learning_rate/action_model': 9.024243978640333e-05, 'epoch': 0.09}) +[2026-06-24 12:21:02,688][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.19262829422950745, 'train/grad_norm_pre_clip': 24.759702682495117, 'timing/data': 0.000328446039929986, 'timing/model': 1.0836908360943198, 'learning_rate/qwen_vl_interface': 9.02191956921708e-06, 'learning_rate/action_model': 9.02191956921708e-05, 'epoch': 0.09}) +[2026-06-24 12:21:03,799][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.1168828010559082, 'train/grad_norm_pre_clip': 26.87366485595703, 'timing/data': 0.0006308229640126228, 'timing/model': 1.1091464220080525, 'learning_rate/qwen_vl_interface': 9.019592712236172e-06, 'learning_rate/action_model': 9.019592712236171e-05, 'epoch': 0.09}) +[2026-06-24 12:21:05,107][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.23448731005191803, 'train/grad_norm_pre_clip': 24.325212478637695, 'timing/data': 0.00046208593994379044, 'timing/model': 1.3060815988574177, 'learning_rate/qwen_vl_interface': 9.017263409207478e-06, 'learning_rate/action_model': 9.017263409207478e-05, 'epoch': 0.09}) +[2026-06-24 12:21:06,195][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.22244778275489807, 'train/grad_norm_pre_clip': 22.26087188720703, 'timing/data': 0.0004944659303873777, 'timing/model': 1.0857606548815966, 'learning_rate/qwen_vl_interface': 9.014931661642455e-06, 'learning_rate/action_model': 9.014931661642455e-05, 'epoch': 0.09}) +[2026-06-24 12:21:07,285][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.23006626963615417, 'train/grad_norm_pre_clip': 38.344444274902344, 'timing/data': 0.00043853186070919037, 'timing/model': 1.0884018160868436, 'learning_rate/qwen_vl_interface': 9.012597471054154e-06, 'learning_rate/action_model': 9.012597471054154e-05, 'epoch': 0.09}) +[2026-06-24 12:21:08,375][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.13917721807956696, 'train/grad_norm_pre_clip': 31.122699737548828, 'timing/data': 0.0004339730367064476, 'timing/model': 1.088005623081699, 'learning_rate/qwen_vl_interface': 9.010260838957206e-06, 'learning_rate/action_model': 9.010260838957205e-05, 'epoch': 0.09}) +[2026-06-24 12:21:09,461][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.2858906686306, 'train/grad_norm_pre_clip': 44.18442153930664, 'timing/data': 0.00047572096809744835, 'timing/model': 1.083809208124876, 'learning_rate/qwen_vl_interface': 9.007921766867823e-06, 'learning_rate/action_model': 9.007921766867822e-05, 'epoch': 0.09}) +[2026-06-24 12:21:10,552][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.20485402643680573, 'train/grad_norm_pre_clip': 25.535017013549805, 'timing/data': 0.0003658849745988846, 'timing/model': 1.0896338040474802, 'learning_rate/qwen_vl_interface': 9.005580256303804e-06, 'learning_rate/action_model': 9.005580256303803e-05, 'epoch': 0.09}) +[2026-06-24 12:21:11,640][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.49788951873779297, 'train/grad_norm_pre_clip': 57.125186920166016, 'timing/data': 0.00045659602619707584, 'timing/model': 1.0863747869152576, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.09}) +[2026-06-24 12:21:12,728][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.21289174258708954, 'train/grad_norm_pre_clip': 26.484371185302734, 'timing/data': 0.00043801101855933666, 'timing/model': 1.085705533158034, 'learning_rate/qwen_vl_interface': 9.000889925830964e-06, 'learning_rate/action_model': 9.000889925830965e-05, 'epoch': 0.09}) +[2026-06-24 12:21:13,810][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.1456967145204544, 'train/grad_norm_pre_clip': 24.257904052734375, 'timing/data': 0.0003627000842243433, 'timing/model': 1.080096502089873, 'learning_rate/qwen_vl_interface': 8.998541108965648e-06, 'learning_rate/action_model': 8.998541108965648e-05, 'epoch': 0.09}) +[2026-06-24 12:21:14,899][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.2350207269191742, 'train/grad_norm_pre_clip': 31.711591720581055, 'timing/data': 0.0004261189606040716, 'timing/model': 1.0872613589745015, 'learning_rate/qwen_vl_interface': 8.996189859712704e-06, 'learning_rate/action_model': 8.996189859712703e-05, 'epoch': 0.09}) +[2026-06-24 12:21:15,984][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.2208489179611206, 'train/grad_norm_pre_clip': 35.64421081542969, 'timing/data': 0.00043175905011594296, 'timing/model': 1.0831999781075865, 'learning_rate/qwen_vl_interface': 8.993836179597833e-06, 'learning_rate/action_model': 8.993836179597833e-05, 'epoch': 0.09}) +[2026-06-24 12:21:17,069][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.2128102034330368, 'train/grad_norm_pre_clip': 29.424413681030273, 'timing/data': 0.0004061309155076742, 'timing/model': 1.08301349892281, 'learning_rate/qwen_vl_interface': 8.99148007014831e-06, 'learning_rate/action_model': 8.99148007014831e-05, 'epoch': 0.09}) +[2026-06-24 12:21:18,154][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.2399396151304245, 'train/grad_norm_pre_clip': 60.09429168701172, 'timing/data': 0.0003635711036622524, 'timing/model': 1.083498327061534, 'learning_rate/qwen_vl_interface': 8.989121532892989e-06, 'learning_rate/action_model': 8.989121532892988e-05, 'epoch': 0.09}) +[2026-06-24 12:21:19,237][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.18395479023456573, 'train/grad_norm_pre_clip': 35.688697814941406, 'timing/data': 0.00035812193527817726, 'timing/model': 1.081651216140017, 'learning_rate/qwen_vl_interface': 8.986760569362302e-06, 'learning_rate/action_model': 8.986760569362301e-05, 'epoch': 0.09}) +[2026-06-24 12:21:20,321][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.09990754723548889, 'train/grad_norm_pre_clip': 20.771312713623047, 'timing/data': 0.0003501360770314932, 'timing/model': 1.082588512916118, 'learning_rate/qwen_vl_interface': 8.984397181088251e-06, 'learning_rate/action_model': 8.984397181088251e-05, 'epoch': 0.09}) +[2026-06-24 12:21:21,703][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.1408519297838211, 'train/grad_norm_pre_clip': 29.437911987304688, 'timing/data': 0.00036507495678961277, 'timing/model': 1.3797759991139174, 'learning_rate/qwen_vl_interface': 8.982031369604412e-06, 'learning_rate/action_model': 8.982031369604411e-05, 'epoch': 0.09}) +[2026-06-24 12:21:22,787][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.15343569219112396, 'train/grad_norm_pre_clip': 21.928133010864258, 'timing/data': 0.00039088213816285133, 'timing/model': 1.0825162371620536, 'learning_rate/qwen_vl_interface': 8.979663136445936e-06, 'learning_rate/action_model': 8.979663136445935e-05, 'epoch': 0.09}) +[2026-06-24 12:21:23,872][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.2259085476398468, 'train/grad_norm_pre_clip': 21.715770721435547, 'timing/data': 0.0004223422147333622, 'timing/model': 1.0832915799692273, 'learning_rate/qwen_vl_interface': 8.977292483149542e-06, 'learning_rate/action_model': 8.977292483149542e-05, 'epoch': 0.09}) +[2026-06-24 12:21:24,955][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.335176557302475, 'train/grad_norm_pre_clip': 41.32577133178711, 'timing/data': 0.00034584803506731987, 'timing/model': 1.0807855539023876, 'learning_rate/qwen_vl_interface': 8.974919411253523e-06, 'learning_rate/action_model': 8.974919411253523e-05, 'epoch': 0.09}) +[2026-06-24 12:21:26,045][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.2103162556886673, 'train/grad_norm_pre_clip': 45.1009521484375, 'timing/data': 0.0003682798705995083, 'timing/model': 1.0884675509296358, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.09}) +[2026-06-24 12:21:27,132][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.37426263093948364, 'train/grad_norm_pre_clip': 56.28540802001953, 'timing/data': 0.0004115109331905842, 'timing/model': 1.0853950129821897, 'learning_rate/qwen_vl_interface': 8.970166017823621e-06, 'learning_rate/action_model': 8.97016601782362e-05, 'epoch': 0.09}) +[2026-06-24 12:21:28,223][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.30350103974342346, 'train/grad_norm_pre_clip': 66.1778793334961, 'timing/data': 0.0004331821110099554, 'timing/model': 1.0889326930046082, 'learning_rate/qwen_vl_interface': 8.96778569937416e-06, 'learning_rate/action_model': 8.96778569937416e-05, 'epoch': 0.09}) +[2026-06-24 12:21:29,313][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.20511992275714874, 'train/grad_norm_pre_clip': 22.863664627075195, 'timing/data': 0.0003849710337817669, 'timing/model': 1.0881379439961165, 'learning_rate/qwen_vl_interface': 8.965402968493922e-06, 'learning_rate/action_model': 8.965402968493922e-05, 'epoch': 0.09}) +[2026-06-24 12:21:30,398][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.05521930009126663, 'train/grad_norm_pre_clip': 8.098668098449707, 'timing/data': 0.0003775071818381548, 'timing/model': 1.083860858110711, 'learning_rate/qwen_vl_interface': 8.96301782672904e-06, 'learning_rate/action_model': 8.963017826729038e-05, 'epoch': 0.09}) +[2026-06-24 12:21:31,490][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.05847993865609169, 'train/grad_norm_pre_clip': 13.524559020996094, 'timing/data': 0.00040595117025077343, 'timing/model': 1.0895517938770354, 'learning_rate/qwen_vl_interface': 8.960630275627197e-06, 'learning_rate/action_model': 8.960630275627197e-05, 'epoch': 0.09}) +[2026-06-24 12:21:32,570][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.29068315029144287, 'train/grad_norm_pre_clip': 60.514564514160156, 'timing/data': 0.0003932570107281208, 'timing/model': 1.078607012052089, 'learning_rate/qwen_vl_interface': 8.958240316737655e-06, 'learning_rate/action_model': 8.958240316737656e-05, 'epoch': 0.09}) +[2026-06-24 12:21:33,654][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.28079020977020264, 'train/grad_norm_pre_clip': 84.42160034179688, 'timing/data': 0.0003796608652919531, 'timing/model': 1.0823278150055557, 'learning_rate/qwen_vl_interface': 8.955847951611233e-06, 'learning_rate/action_model': 8.955847951611232e-05, 'epoch': 0.09}) +[2026-06-24 12:21:34,736][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.019535955041646957, 'train/grad_norm_pre_clip': 7.007569789886475, 'timing/data': 0.0004522779490798712, 'timing/model': 1.0797210100572556, 'learning_rate/qwen_vl_interface': 8.953453181800307e-06, 'learning_rate/action_model': 8.953453181800306e-05, 'epoch': 0.09}) +[2026-06-24 12:21:35,819][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.053751926869153976, 'train/grad_norm_pre_clip': 20.080490112304688, 'timing/data': 0.00040037999860942364, 'timing/model': 1.0813234040979296, 'learning_rate/qwen_vl_interface': 8.951056008858819e-06, 'learning_rate/action_model': 8.95105600885882e-05, 'epoch': 0.09}) +[2026-06-24 12:21:37,110][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.06478822976350784, 'train/grad_norm_pre_clip': 16.826576232910156, 'timing/data': 0.00037501309998333454, 'timing/model': 1.2892549990210682, 'learning_rate/qwen_vl_interface': 8.94865643434227e-06, 'learning_rate/action_model': 8.948656434342269e-05, 'epoch': 0.09}) +[2026-06-24 12:21:38,190][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.04984155297279358, 'train/grad_norm_pre_clip': 25.370235443115234, 'timing/data': 0.0003692519385367632, 'timing/model': 1.0788287890609354, 'learning_rate/qwen_vl_interface': 8.946254459807717e-06, 'learning_rate/action_model': 8.946254459807717e-05, 'epoch': 0.09}) +[2026-06-24 12:21:39,275][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.07817709445953369, 'train/grad_norm_pre_clip': 21.100574493408203, 'timing/data': 0.00035383296199142933, 'timing/model': 1.083030743058771, 'learning_rate/qwen_vl_interface': 8.943850086813773e-06, 'learning_rate/action_model': 8.943850086813773e-05, 'epoch': 0.09}) +[2026-06-24 12:21:40,361][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.37849023938179016, 'train/grad_norm_pre_clip': 138.8944854736328, 'timing/data': 0.0004365178756415844, 'timing/model': 1.0836902090813965, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.09}) +[2026-06-24 12:21:41,448][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.17671723663806915, 'train/grad_norm_pre_clip': 61.224853515625, 'timing/data': 0.00046903896145522594, 'timing/model': 1.0849806379992515, 'learning_rate/qwen_vl_interface': 8.93903415168996e-06, 'learning_rate/action_model': 8.93903415168996e-05, 'epoch': 0.09}) +[2026-06-24 12:21:42,535][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.17210006713867188, 'train/grad_norm_pre_clip': 65.71194458007812, 'timing/data': 0.0004385518841445446, 'timing/model': 1.085790695855394, 'learning_rate/qwen_vl_interface': 8.9366225926851e-06, 'learning_rate/action_model': 8.936622592685098e-05, 'epoch': 0.09}) +[2026-06-24 12:21:43,621][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.0313284769654274, 'train/grad_norm_pre_clip': 28.530515670776367, 'timing/data': 0.00044955196790397167, 'timing/model': 1.08406339911744, 'learning_rate/qwen_vl_interface': 8.934208641470863e-06, 'learning_rate/action_model': 8.934208641470863e-05, 'epoch': 0.09}) +[2026-06-24 12:21:44,710][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.1614757925271988, 'train/grad_norm_pre_clip': 37.506263732910156, 'timing/data': 0.00047364714555442333, 'timing/model': 1.087034888099879, 'learning_rate/qwen_vl_interface': 8.931792299613637e-06, 'learning_rate/action_model': 8.931792299613636e-05, 'epoch': 0.09}) +[2026-06-24 12:21:45,802][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.28945764899253845, 'train/grad_norm_pre_clip': 124.90167999267578, 'timing/data': 0.0004225021693855524, 'timing/model': 1.0903693670406938, 'learning_rate/qwen_vl_interface': 8.92937356868136e-06, 'learning_rate/action_model': 8.92937356868136e-05, 'epoch': 0.09}) +[2026-06-24 12:21:46,890][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.03958369418978691, 'train/grad_norm_pre_clip': 37.16053771972656, 'timing/data': 0.0004604028072208166, 'timing/model': 1.085529217030853, 'learning_rate/qwen_vl_interface': 8.926952450243522e-06, 'learning_rate/action_model': 8.926952450243522e-05, 'epoch': 0.09}) +[2026-06-24 12:21:47,976][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.16464485228061676, 'train/grad_norm_pre_clip': 47.78163528442383, 'timing/data': 0.0004973020404577255, 'timing/model': 1.084735579090193, 'learning_rate/qwen_vl_interface': 8.924528945871157e-06, 'learning_rate/action_model': 8.924528945871156e-05, 'epoch': 0.09}) +[2026-06-24 12:21:49,066][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.5458593368530273, 'train/grad_norm_pre_clip': 119.2593002319336, 'timing/data': 0.0003808841574937105, 'timing/model': 1.087608122965321, 'learning_rate/qwen_vl_interface': 8.922103057136857e-06, 'learning_rate/action_model': 8.922103057136856e-05, 'epoch': 0.09}) +[2026-06-24 12:21:50,153][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.06069788709282875, 'train/grad_norm_pre_clip': 66.74178314208984, 'timing/data': 0.0004242050927132368, 'timing/model': 1.0852487811353058, 'learning_rate/qwen_vl_interface': 8.919674785614747e-06, 'learning_rate/action_model': 8.919674785614747e-05, 'epoch': 0.09}) +[2026-06-24 12:21:51,236][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.09366582334041595, 'train/grad_norm_pre_clip': 30.408756256103516, 'timing/data': 0.0004553939215838909, 'timing/model': 1.0815140220802277, 'learning_rate/qwen_vl_interface': 8.91724413288051e-06, 'learning_rate/action_model': 8.91724413288051e-05, 'epoch': 0.09}) +[2026-06-24 12:21:52,322][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.3000689446926117, 'train/grad_norm_pre_clip': 90.58032989501953, 'timing/data': 0.0004689190536737442, 'timing/model': 1.0844501750543714, 'learning_rate/qwen_vl_interface': 8.91481110051137e-06, 'learning_rate/action_model': 8.91481110051137e-05, 'epoch': 0.09}) +[2026-06-24 12:21:53,408][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.1379672735929489, 'train/grad_norm_pre_clip': 38.45014953613281, 'timing/data': 0.0004674461670219898, 'timing/model': 1.0833122020121664, 'learning_rate/qwen_vl_interface': 8.912375690086097e-06, 'learning_rate/action_model': 8.912375690086096e-05, 'epoch': 0.09}) +[2026-06-24 12:21:54,496][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.159246563911438, 'train/grad_norm_pre_clip': 70.4927749633789, 'timing/data': 0.000405740924179554, 'timing/model': 1.0864444330800325, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.09}) +[2026-06-24 12:21:55,584][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.058397211134433746, 'train/grad_norm_pre_clip': 23.34349822998047, 'timing/data': 0.0004587089642882347, 'timing/model': 1.0860834559425712, 'learning_rate/qwen_vl_interface': 8.907497741389932e-06, 'learning_rate/action_model': 8.907497741389932e-05, 'epoch': 0.09}) +[2026-06-24 12:21:56,964][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.16296246647834778, 'train/grad_norm_pre_clip': 68.26451110839844, 'timing/data': 0.0004427200183272362, 'timing/model': 1.3780599178280681, 'learning_rate/qwen_vl_interface': 8.905055206284289e-06, 'learning_rate/action_model': 8.905055206284288e-05, 'epoch': 0.09}) +[2026-06-24 12:21:58,067][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.25519242882728577, 'train/grad_norm_pre_clip': 96.44805145263672, 'timing/data': 0.00039125303737819195, 'timing/model': 1.1011662730015814, 'learning_rate/qwen_vl_interface': 8.902610299453005e-06, 'learning_rate/action_model': 8.902610299453004e-05, 'epoch': 0.09}) +[2026-06-24 12:21:59,154][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.0324719175696373, 'train/grad_norm_pre_clip': 14.40721607208252, 'timing/data': 0.00043197907507419586, 'timing/model': 1.0855226861312985, 'learning_rate/qwen_vl_interface': 8.900163022482554e-06, 'learning_rate/action_model': 8.900163022482553e-05, 'epoch': 0.09}) +[2026-06-24 12:22:00,245][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.12693172693252563, 'train/grad_norm_pre_clip': 58.15073776245117, 'timing/data': 0.0003781490959227085, 'timing/model': 1.088950620032847, 'learning_rate/qwen_vl_interface': 8.897713376960946e-06, 'learning_rate/action_model': 8.897713376960946e-05, 'epoch': 0.09}) +[2026-06-24 12:22:01,340][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.1726381927728653, 'train/grad_norm_pre_clip': 30.16960906982422, 'timing/data': 0.0006900329608470201, 'timing/model': 1.0928109439555556, 'learning_rate/qwen_vl_interface': 8.895261364477731e-06, 'learning_rate/action_model': 8.895261364477731e-05, 'epoch': 0.09}) +[2026-06-24 12:22:02,431][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.3789121210575104, 'train/grad_norm_pre_clip': 39.47504425048828, 'timing/data': 0.0005018799565732479, 'timing/model': 1.088463497115299, 'learning_rate/qwen_vl_interface': 8.892806986623993e-06, 'learning_rate/action_model': 8.892806986623992e-05, 'epoch': 0.09}) +[2026-06-24 12:22:03,520][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.30685725808143616, 'train/grad_norm_pre_clip': 38.273868560791016, 'timing/data': 0.00041184201836586, 'timing/model': 1.0876756901852787, 'learning_rate/qwen_vl_interface': 8.890350244992349e-06, 'learning_rate/action_model': 8.890350244992348e-05, 'epoch': 0.09}) +[2026-06-24 12:22:04,610][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.1373765468597412, 'train/grad_norm_pre_clip': 31.057857513427734, 'timing/data': 0.00048004905693233013, 'timing/model': 1.087852121097967, 'learning_rate/qwen_vl_interface': 8.887891141176954e-06, 'learning_rate/action_model': 8.887891141176954e-05, 'epoch': 0.09}) +[2026-06-24 12:22:05,699][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.22298510372638702, 'train/grad_norm_pre_clip': 54.43523025512695, 'timing/data': 0.000416299793869257, 'timing/model': 1.0870024689938873, 'learning_rate/qwen_vl_interface': 8.885429676773495e-06, 'learning_rate/action_model': 8.885429676773495e-05, 'epoch': 0.09}) +[2026-06-24 12:22:06,790][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.17202726006507874, 'train/grad_norm_pre_clip': 48.39702606201172, 'timing/data': 0.0004023138899356127, 'timing/model': 1.0893638350535184, 'learning_rate/qwen_vl_interface': 8.882965853379188e-06, 'learning_rate/action_model': 8.882965853379187e-05, 'epoch': 0.09}) +[2026-06-24 12:22:07,879][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.10248684883117676, 'train/grad_norm_pre_clip': 33.197505950927734, 'timing/data': 0.0004043478984385729, 'timing/model': 1.086895708926022, 'learning_rate/qwen_vl_interface': 8.88049967259278e-06, 'learning_rate/action_model': 8.880499672592778e-05, 'epoch': 0.09}) +[2026-06-24 12:22:08,967][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.14420534670352936, 'train/grad_norm_pre_clip': 39.30585861206055, 'timing/data': 0.00046418001875281334, 'timing/model': 1.085948393912986, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.09}) +[2026-06-24 12:22:10,051][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.05202481895685196, 'train/grad_norm_pre_clip': 8.266685485839844, 'timing/data': 0.0003905219491571188, 'timing/model': 1.0826768840197474, 'learning_rate/qwen_vl_interface': 8.875560245246307e-06, 'learning_rate/action_model': 8.875560245246307e-05, 'epoch': 0.09}) +[2026-06-24 12:22:11,136][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.3517928719520569, 'train/grad_norm_pre_clip': 48.377784729003906, 'timing/data': 0.0003749129828065634, 'timing/model': 1.0832643639296293, 'learning_rate/qwen_vl_interface': 8.873087001891381e-06, 'learning_rate/action_model': 8.87308700189138e-05, 'epoch': 0.09}) +[2026-06-24 12:22:12,220][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.12367995083332062, 'train/grad_norm_pre_clip': 16.97444725036621, 'timing/data': 0.00035983510315418243, 'timing/model': 1.0820769190322608, 'learning_rate/qwen_vl_interface': 8.870611407554635e-06, 'learning_rate/action_model': 8.870611407554636e-05, 'epoch': 0.09}) +[2026-06-24 12:22:13,302][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.11922736465930939, 'train/grad_norm_pre_clip': 16.15257453918457, 'timing/data': 0.0004209689795970917, 'timing/model': 1.0803021441679448, 'learning_rate/qwen_vl_interface': 8.868133463842456e-06, 'learning_rate/action_model': 8.868133463842457e-05, 'epoch': 0.09}) +[2026-06-24 12:22:14,385][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.15003664791584015, 'train/grad_norm_pre_clip': 26.803133010864258, 'timing/data': 0.0004205978475511074, 'timing/model': 1.0810069439467043, 'learning_rate/qwen_vl_interface': 8.865653172362755e-06, 'learning_rate/action_model': 8.865653172362754e-05, 'epoch': 0.09}) +[2026-06-24 12:22:15,719][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.248219296336174, 'train/grad_norm_pre_clip': 33.04884719848633, 'timing/data': 0.0003933170810341835, 'timing/model': 1.3326847639400512, 'learning_rate/qwen_vl_interface': 8.863170534724962e-06, 'learning_rate/action_model': 8.863170534724961e-05, 'epoch': 0.09}) +[2026-06-24 12:22:16,803][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.20327337086200714, 'train/grad_norm_pre_clip': 23.733488082885742, 'timing/data': 0.0003508168738335371, 'timing/model': 1.0816308939829469, 'learning_rate/qwen_vl_interface': 8.860685552540038e-06, 'learning_rate/action_model': 8.860685552540037e-05, 'epoch': 0.09}) +[2026-06-24 12:22:17,883][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.059981491416692734, 'train/grad_norm_pre_clip': 8.99927806854248, 'timing/data': 0.00036892108619213104, 'timing/model': 1.078516016015783, 'learning_rate/qwen_vl_interface': 8.858198227420461e-06, 'learning_rate/action_model': 8.85819822742046e-05, 'epoch': 0.09}) +[2026-06-24 12:22:18,963][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.2590418756008148, 'train/grad_norm_pre_clip': 50.06097412109375, 'timing/data': 0.00039414805360138416, 'timing/model': 1.0784385609440506, 'learning_rate/qwen_vl_interface': 8.855708560980225e-06, 'learning_rate/action_model': 8.855708560980225e-05, 'epoch': 0.09}) +[2026-06-24 12:22:20,048][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.20041503012180328, 'train/grad_norm_pre_clip': 71.17584228515625, 'timing/data': 0.00038675498217344284, 'timing/model': 1.083220693981275, 'learning_rate/qwen_vl_interface': 8.853216554834852e-06, 'learning_rate/action_model': 8.85321655483485e-05, 'epoch': 0.09}) +[2026-06-24 12:22:21,131][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.18850001692771912, 'train/grad_norm_pre_clip': 35.59951400756836, 'timing/data': 0.0004147379659116268, 'timing/model': 1.0816371159162372, 'learning_rate/qwen_vl_interface': 8.850722210601376e-06, 'learning_rate/action_model': 8.850722210601377e-05, 'epoch': 0.1}) +[2026-06-24 12:22:22,214][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.17979590594768524, 'train/grad_norm_pre_clip': 29.047222137451172, 'timing/data': 0.0003570488188415766, 'timing/model': 1.0808732248842716, 'learning_rate/qwen_vl_interface': 8.848225529898351e-06, 'learning_rate/action_model': 8.84822552989835e-05, 'epoch': 0.1}) +[2026-06-24 12:22:23,297][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.10586240142583847, 'train/grad_norm_pre_clip': 30.656890869140625, 'timing/data': 0.0004082049708813429, 'timing/model': 1.0814932060893625, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.1}) +[2026-06-24 12:22:24,402][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.07074660807847977, 'train/grad_norm_pre_clip': 20.591920852661133, 'timing/data': 0.00044258893467485905, 'timing/model': 1.1032493989914656, 'learning_rate/qwen_vl_interface': 8.84322516556544e-06, 'learning_rate/action_model': 8.84322516556544e-05, 'epoch': 0.1}) +[2026-06-24 12:22:25,493][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.25082752108573914, 'train/grad_norm_pre_clip': 68.04010009765625, 'timing/data': 0.0004539908841252327, 'timing/model': 1.0893024220131338, 'learning_rate/qwen_vl_interface': 8.840721485180237e-06, 'learning_rate/action_model': 8.840721485180236e-05, 'epoch': 0.1}) +[2026-06-24 12:22:26,589][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.07688309997320175, 'train/grad_norm_pre_clip': 23.829591751098633, 'timing/data': 0.000464871060103178, 'timing/model': 1.0939885638654232, 'learning_rate/qwen_vl_interface': 8.838215474814848e-06, 'learning_rate/action_model': 8.838215474814847e-05, 'epoch': 0.1}) +[2026-06-24 12:22:27,677][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.09009797871112823, 'train/grad_norm_pre_clip': 22.678464889526367, 'timing/data': 0.0004233228974044323, 'timing/model': 1.0857690698467195, 'learning_rate/qwen_vl_interface': 8.835707136095392e-06, 'learning_rate/action_model': 8.835707136095391e-05, 'epoch': 0.1}) +[2026-06-24 12:22:28,764][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.067145437002182, 'train/grad_norm_pre_clip': 29.00067138671875, 'timing/data': 0.0005018208175897598, 'timing/model': 1.0852298301178962, 'learning_rate/qwen_vl_interface': 8.833196470649505e-06, 'learning_rate/action_model': 8.833196470649505e-05, 'epoch': 0.1}) +[2026-06-24 12:22:29,851][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.2776842713356018, 'train/grad_norm_pre_clip': 47.04991912841797, 'timing/data': 0.00044735916890203953, 'timing/model': 1.0844136299565434, 'learning_rate/qwen_vl_interface': 8.83068348010633e-06, 'learning_rate/action_model': 8.83068348010633e-05, 'epoch': 0.1}) +[2026-06-24 12:22:30,938][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.22764259576797485, 'train/grad_norm_pre_clip': 51.22591781616211, 'timing/data': 0.00040295510552823544, 'timing/model': 1.0858545410446823, 'learning_rate/qwen_vl_interface': 8.82816816609652e-06, 'learning_rate/action_model': 8.828168166096521e-05, 'epoch': 0.1}) +[2026-06-24 12:22:32,024][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.2522002160549164, 'train/grad_norm_pre_clip': 21.083515167236328, 'timing/data': 0.0004093870520591736, 'timing/model': 1.084022887982428, 'learning_rate/qwen_vl_interface': 8.825650530252238e-06, 'learning_rate/action_model': 8.825650530252238e-05, 'epoch': 0.1}) +[2026-06-24 12:22:33,111][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.38544154167175293, 'train/grad_norm_pre_clip': 50.02939224243164, 'timing/data': 0.00041299406439065933, 'timing/model': 1.0850948269944638, 'learning_rate/qwen_vl_interface': 8.823130574207144e-06, 'learning_rate/action_model': 8.823130574207143e-05, 'epoch': 0.1}) +[2026-06-24 12:22:34,408][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.20994676649570465, 'train/grad_norm_pre_clip': 58.50048065185547, 'timing/data': 0.0003989981487393379, 'timing/model': 1.295764806913212, 'learning_rate/qwen_vl_interface': 8.82060829959641e-06, 'learning_rate/action_model': 8.820608299596411e-05, 'epoch': 0.1}) +[2026-06-24 12:22:35,492][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.06617245823144913, 'train/grad_norm_pre_clip': 21.308650970458984, 'timing/data': 0.0004588800948113203, 'timing/model': 1.0821479849983007, 'learning_rate/qwen_vl_interface': 8.81808370805672e-06, 'learning_rate/action_model': 8.818083708056721e-05, 'epoch': 0.1}) +[2026-06-24 12:22:36,576][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.34330594539642334, 'train/grad_norm_pre_clip': 82.49114990234375, 'timing/data': 0.00045617506839334965, 'timing/model': 1.0815933449193835, 'learning_rate/qwen_vl_interface': 8.815556801226251e-06, 'learning_rate/action_model': 8.815556801226251e-05, 'epoch': 0.1}) +[2026-06-24 12:22:37,661][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.202811136841774, 'train/grad_norm_pre_clip': 31.906633377075195, 'timing/data': 0.00041801296174526215, 'timing/model': 1.0833519320003688, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.1}) +[2026-06-24 12:22:38,742][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.22225317358970642, 'train/grad_norm_pre_clip': 48.97510528564453, 'timing/data': 0.00036809989251196384, 'timing/model': 1.0791207519359887, 'learning_rate/qwen_vl_interface': 8.810496048253203e-06, 'learning_rate/action_model': 8.810496048253202e-05, 'epoch': 0.1}) +[2026-06-24 12:22:39,824][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.4018307626247406, 'train/grad_norm_pre_clip': 98.66730499267578, 'timing/data': 0.0003874960821121931, 'timing/model': 1.08048089989461, 'learning_rate/qwen_vl_interface': 8.80796220539449e-06, 'learning_rate/action_model': 8.80796220539449e-05, 'epoch': 0.1}) +[2026-06-24 12:22:40,908][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1000 +[2026-06-24 12:23:07,173][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1000 +[2026-06-24 12:23:07,174][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1000 +[2026-06-24 12:23:40,426][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1000 +[2026-06-24 12:24:29,938][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-24 12:24:29,939][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.47819000482559204, 'train/grad_norm_pre_clip': 108.21360778808594, 'eval/action_loss/seconds': 26.26345086400397, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.1355978399515152, 'eval/latency_2/loss': 0.1355978399515152, 'eval/flappy/loss': 0.1355978399515152, 'eval/flappy/latency_2/loss': 0.1355978399515152, 'eval/flap_precision': 0.9730185497470489, 'eval/flap_recall': 0.7829036635006784, 'eval/flap_f1': 0.8676691729323308, 'eval/flap_rate_model': 0.16472222222222221, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.8676691729323308, 'eval/latency_2/flap_precision': 0.9730185497470489, 'eval/latency_2/flap_recall': 0.7829036635006784, 'eval/latency_2/flap_f1': 0.8676691729323308, 'eval/latency_2/flap_rate_model': 0.16472222222222221, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.8676691729323308, 'eval/action_classification/data_wait_seconds': 0.3137364685535431, 'eval/action_classification/predict_seconds': 32.835628509521484, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 33.600000000000165, 'rl_games_eval/mid_train/mean_length': 289.4, 'rl_games_eval/mid_train/std_reward': 20.155892438689175, 'rl_games_eval/mid_train/std_length': 162.70291945751927, 'rl_games_eval/mid_train/macro_mean_reward': 33.600000000000165, 'rl_games_eval/mid_train/macro_mean_length': 289.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 33.600000000000165, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 289.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 20.155892438689175, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 162.70291945751927, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 33.600000000000165, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0004121519159525633, 'timing/model': 1.0824070610105991, 'learning_rate/qwen_vl_interface': 8.805426053812731e-06, 'learning_rate/action_model': 8.805426053812731e-05, 'epoch': 0.1}) +[2026-06-24 12:24:29,939][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_1000_state +[2026-06-24 12:24:29,942][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 12:24:32,635][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_1000_state/model.safetensors +[2026-06-24 12:24:38,408][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_1000_state/optimizer.bin +[2026-06-24 12:24:38,408][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_1000_state/sampler.bin +[2026-06-24 12:24:38,409][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-24 12:24:38,409][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-24 12:24:38,466][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-24 12:24:38,467][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 12:24:38,477][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 12:24:39,539][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.1143263503909111, 'train/grad_norm_pre_clip': 27.705480575561523, 'timing/data': 0.00042196991853415966, 'timing/model': 1.0598311109934002, 'learning_rate/qwen_vl_interface': 8.802887595153607e-06, 'learning_rate/action_model': 8.802887595153606e-05, 'epoch': 0.1}) +[2026-06-24 12:24:40,609][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.21317285299301147, 'train/grad_norm_pre_clip': 30.64771270751953, 'timing/data': 0.00037671602331101894, 'timing/model': 1.0688123239669949, 'learning_rate/qwen_vl_interface': 8.800346831064294e-06, 'learning_rate/action_model': 8.800346831064294e-05, 'epoch': 0.1}) +[2026-06-24 12:24:41,680][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.21492040157318115, 'train/grad_norm_pre_clip': 25.58682632446289, 'timing/data': 0.00044949213042855263, 'timing/model': 1.0686229979619384, 'learning_rate/qwen_vl_interface': 8.797803763193471e-06, 'learning_rate/action_model': 8.797803763193471e-05, 'epoch': 0.1}) +[2026-06-24 12:24:43,033][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.3324001729488373, 'train/grad_norm_pre_clip': 43.094451904296875, 'timing/data': 0.00036381185054779053, 'timing/model': 1.3515245171729475, 'learning_rate/qwen_vl_interface': 8.795258393191302e-06, 'learning_rate/action_model': 8.795258393191301e-05, 'epoch': 0.1}) +[2026-06-24 12:24:44,114][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.15852268040180206, 'train/grad_norm_pre_clip': 24.08017921447754, 'timing/data': 0.0003958619199693203, 'timing/model': 1.0795021771918982, 'learning_rate/qwen_vl_interface': 8.792710722709451e-06, 'learning_rate/action_model': 8.792710722709451e-05, 'epoch': 0.1}) +[2026-06-24 12:24:45,183][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.11851399391889572, 'train/grad_norm_pre_clip': 16.106319427490234, 'timing/data': 0.0004049490671604872, 'timing/model': 1.0671091610565782, 'learning_rate/qwen_vl_interface': 8.790160753401076e-06, 'learning_rate/action_model': 8.790160753401077e-05, 'epoch': 0.1}) +[2026-06-24 12:24:46,250][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.11601553857326508, 'train/grad_norm_pre_clip': 19.74519157409668, 'timing/data': 0.00038394005969166756, 'timing/model': 1.065068518044427, 'learning_rate/qwen_vl_interface': 8.787608486920824e-06, 'learning_rate/action_model': 8.787608486920824e-05, 'epoch': 0.1}) +[2026-06-24 12:24:47,315][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.09726208448410034, 'train/grad_norm_pre_clip': 28.786598205566406, 'timing/data': 0.00041209207847714424, 'timing/model': 1.0633197401184589, 'learning_rate/qwen_vl_interface': 8.78505392492483e-06, 'learning_rate/action_model': 8.78505392492483e-05, 'epoch': 0.1}) +[2026-06-24 12:24:48,381][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.15042749047279358, 'train/grad_norm_pre_clip': 22.71491241455078, 'timing/data': 0.00043026707135140896, 'timing/model': 1.06428361707367, 'learning_rate/qwen_vl_interface': 8.782497069070725e-06, 'learning_rate/action_model': 8.782497069070725e-05, 'epoch': 0.1}) +[2026-06-24 12:24:49,451][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.21388886868953705, 'train/grad_norm_pre_clip': 38.19995880126953, 'timing/data': 0.0003784389700740576, 'timing/model': 1.0678781641181558, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.1}) +[2026-06-24 12:24:50,521][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.23260290920734406, 'train/grad_norm_pre_clip': 64.23485565185547, 'timing/data': 0.0003410489298403263, 'timing/model': 1.06863396987319, 'learning_rate/qwen_vl_interface': 8.777376482426129e-06, 'learning_rate/action_model': 8.777376482426128e-05, 'epoch': 0.1}) +[2026-06-24 12:24:51,587][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.0948844850063324, 'train/grad_norm_pre_clip': 13.576864242553711, 'timing/data': 0.0004365278873592615, 'timing/model': 1.0641704448498785, 'learning_rate/qwen_vl_interface': 8.774812754958332e-06, 'learning_rate/action_model': 8.774812754958331e-05, 'epoch': 0.1}) +[2026-06-24 12:24:52,656][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.08396334946155548, 'train/grad_norm_pre_clip': 18.05655860900879, 'timing/data': 0.0003463088069111109, 'timing/model': 1.0669588099699467, 'learning_rate/qwen_vl_interface': 8.772246740277805e-06, 'learning_rate/action_model': 8.772246740277805e-05, 'epoch': 0.1}) +[2026-06-24 12:24:53,720][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.27523258328437805, 'train/grad_norm_pre_clip': 29.126304626464844, 'timing/data': 0.0003427518531680107, 'timing/model': 1.063041419023648, 'learning_rate/qwen_vl_interface': 8.769678440049608e-06, 'learning_rate/action_model': 8.769678440049608e-05, 'epoch': 0.1}) +[2026-06-24 12:24:54,790][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.11675506085157394, 'train/grad_norm_pre_clip': 29.43207550048828, 'timing/data': 0.0003800520207732916, 'timing/model': 1.0675626120064408, 'learning_rate/qwen_vl_interface': 8.767107855940287e-06, 'learning_rate/action_model': 8.767107855940286e-05, 'epoch': 0.1}) +[2026-06-24 12:24:55,856][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.1314137876033783, 'train/grad_norm_pre_clip': 25.678014755249023, 'timing/data': 0.00035011605359613895, 'timing/model': 1.064955998212099, 'learning_rate/qwen_vl_interface': 8.764534989617861e-06, 'learning_rate/action_model': 8.76453498961786e-05, 'epoch': 0.1}) +[2026-06-24 12:24:56,926][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.19554035365581512, 'train/grad_norm_pre_clip': 36.70880126953125, 'timing/data': 0.0003270828165113926, 'timing/model': 1.0676109320484102, 'learning_rate/qwen_vl_interface': 8.761959842751837e-06, 'learning_rate/action_model': 8.761959842751837e-05, 'epoch': 0.1}) +[2026-06-24 12:24:57,999][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.2859485149383545, 'train/grad_norm_pre_clip': 37.33724594116211, 'timing/data': 0.0003807640168815851, 'timing/model': 1.071492827963084, 'learning_rate/qwen_vl_interface': 8.7593824170132e-06, 'learning_rate/action_model': 8.7593824170132e-05, 'epoch': 0.1}) +[2026-06-24 12:24:59,067][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.17307397723197937, 'train/grad_norm_pre_clip': 49.6905632019043, 'timing/data': 0.00043591693975031376, 'timing/model': 1.0665526201482862, 'learning_rate/qwen_vl_interface': 8.756802714074412e-06, 'learning_rate/action_model': 8.756802714074412e-05, 'epoch': 0.1}) +[2026-06-24 12:25:00,426][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.14923438429832458, 'train/grad_norm_pre_clip': 37.6906623840332, 'timing/data': 0.0003608160186558962, 'timing/model': 1.35746192606166, 'learning_rate/qwen_vl_interface': 8.75422073560942e-06, 'learning_rate/action_model': 8.754220735609419e-05, 'epoch': 0.1}) +[2026-06-24 12:25:01,509][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.07139647006988525, 'train/grad_norm_pre_clip': 18.846799850463867, 'timing/data': 0.0004680871497839689, 'timing/model': 1.0810027900151908, 'learning_rate/qwen_vl_interface': 8.751636483293635e-06, 'learning_rate/action_model': 8.751636483293634e-05, 'epoch': 0.1}) +[2026-06-24 12:25:02,591][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.0539066344499588, 'train/grad_norm_pre_clip': 16.003347396850586, 'timing/data': 0.00037005404010415077, 'timing/model': 1.0795924582052976, 'learning_rate/qwen_vl_interface': 8.749049958803952e-06, 'learning_rate/action_model': 8.749049958803951e-05, 'epoch': 0.1}) +[2026-06-24 12:25:03,673][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.11561592668294907, 'train/grad_norm_pre_clip': 31.10751724243164, 'timing/data': 0.0004087460692971945, 'timing/model': 1.0809043070767075, 'learning_rate/qwen_vl_interface': 8.74646116381874e-06, 'learning_rate/action_model': 8.74646116381874e-05, 'epoch': 0.1}) +[2026-06-24 12:25:04,753][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.19568099081516266, 'train/grad_norm_pre_clip': 37.23085403442383, 'timing/data': 0.00039830594323575497, 'timing/model': 1.0774432900361717, 'learning_rate/qwen_vl_interface': 8.74387010001784e-06, 'learning_rate/action_model': 8.743870100017838e-05, 'epoch': 0.1}) +[2026-06-24 12:25:05,836][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.21849068999290466, 'train/grad_norm_pre_clip': 35.9441032409668, 'timing/data': 0.0004404259379953146, 'timing/model': 1.0814275071024895, 'learning_rate/qwen_vl_interface': 8.741276769082562e-06, 'learning_rate/action_model': 8.741276769082562e-05, 'epoch': 0.1}) +[2026-06-24 12:25:06,912][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.1786147505044937, 'train/grad_norm_pre_clip': 47.551597595214844, 'timing/data': 0.0004451440181583166, 'timing/model': 1.0741608089301735, 'learning_rate/qwen_vl_interface': 8.738681172695694e-06, 'learning_rate/action_model': 8.738681172695694e-05, 'epoch': 0.1}) +[2026-06-24 12:25:07,991][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.22613154351711273, 'train/grad_norm_pre_clip': 33.40286636352539, 'timing/data': 0.00046692509204149246, 'timing/model': 1.077517308993265, 'learning_rate/qwen_vl_interface': 8.73608331254149e-06, 'learning_rate/action_model': 8.736083312541489e-05, 'epoch': 0.1}) +[2026-06-24 12:25:09,073][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.2263999730348587, 'train/grad_norm_pre_clip': 49.19373321533203, 'timing/data': 0.0004637790843844414, 'timing/model': 1.0798775320872664, 'learning_rate/qwen_vl_interface': 8.733483190305671e-06, 'learning_rate/action_model': 8.733483190305671e-05, 'epoch': 0.1}) +[2026-06-24 12:25:10,147][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.3068307340145111, 'train/grad_norm_pre_clip': 116.64608764648438, 'timing/data': 0.0004374198615550995, 'timing/model': 1.0722259229514748, 'learning_rate/qwen_vl_interface': 8.730880807675431e-06, 'learning_rate/action_model': 8.73088080767543e-05, 'epoch': 0.1}) +[2026-06-24 12:25:11,227][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.18012085556983948, 'train/grad_norm_pre_clip': 39.976444244384766, 'timing/data': 0.00042743096128106117, 'timing/model': 1.077585387043655, 'learning_rate/qwen_vl_interface': 8.728276166339426e-06, 'learning_rate/action_model': 8.728276166339426e-05, 'epoch': 0.1}) +[2026-06-24 12:25:12,306][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.35855844616889954, 'train/grad_norm_pre_clip': 46.58348846435547, 'timing/data': 0.0004127328284084797, 'timing/model': 1.0768844541162252, 'learning_rate/qwen_vl_interface': 8.725669267987779e-06, 'learning_rate/action_model': 8.725669267987778e-05, 'epoch': 0.1}) +[2026-06-24 12:25:13,385][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.3130453824996948, 'train/grad_norm_pre_clip': 55.90946960449219, 'timing/data': 0.0004441719502210617, 'timing/model': 1.0770750499796122, 'learning_rate/qwen_vl_interface': 8.723060114312079e-06, 'learning_rate/action_model': 8.723060114312079e-05, 'epoch': 0.1}) +[2026-06-24 12:25:14,466][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.2412118762731552, 'train/grad_norm_pre_clip': 56.478477478027344, 'timing/data': 0.0004675760865211487, 'timing/model': 1.0797376409173012, 'learning_rate/qwen_vl_interface': 8.720448707005377e-06, 'learning_rate/action_model': 8.720448707005376e-05, 'epoch': 0.1}) +[2026-06-24 12:25:15,543][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.40990978479385376, 'train/grad_norm_pre_clip': 123.349365234375, 'timing/data': 0.0003881279844790697, 'timing/model': 1.0748347719199955, 'learning_rate/qwen_vl_interface': 8.717835047762188e-06, 'learning_rate/action_model': 8.717835047762187e-05, 'epoch': 0.1}) +[2026-06-24 12:25:16,621][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.12765876948833466, 'train/grad_norm_pre_clip': 16.2828369140625, 'timing/data': 0.0004145270213484764, 'timing/model': 1.07626987202093, 'learning_rate/qwen_vl_interface': 8.715219138278486e-06, 'learning_rate/action_model': 8.715219138278486e-05, 'epoch': 0.1}) +[2026-06-24 12:25:17,983][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.15288753807544708, 'train/grad_norm_pre_clip': 27.049741744995117, 'timing/data': 0.0004070429131388664, 'timing/model': 1.3602000691462308, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.1}) +[2026-06-24 12:25:19,060][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.14760196208953857, 'train/grad_norm_pre_clip': 19.11971092224121, 'timing/data': 0.00045761698856949806, 'timing/model': 1.0755579271353781, 'learning_rate/qwen_vl_interface': 8.709980575380741e-06, 'learning_rate/action_model': 8.709980575380741e-05, 'epoch': 0.1}) +[2026-06-24 12:25:20,133][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.0451088547706604, 'train/grad_norm_pre_clip': 9.799840927124023, 'timing/data': 0.0003722070250660181, 'timing/model': 1.0712761939503253, 'learning_rate/qwen_vl_interface': 8.707357925365949e-06, 'learning_rate/action_model': 8.707357925365948e-05, 'epoch': 0.1}) +[2026-06-24 12:25:21,216][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.20082232356071472, 'train/grad_norm_pre_clip': 31.753664016723633, 'timing/data': 0.0004157288931310177, 'timing/model': 1.0804617991670966, 'learning_rate/qwen_vl_interface': 8.704733031909134e-06, 'learning_rate/action_model': 8.704733031909133e-05, 'epoch': 0.1}) +[2026-06-24 12:25:22,302][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.06955088675022125, 'train/grad_norm_pre_clip': 24.894241333007812, 'timing/data': 0.00046128383837640285, 'timing/model': 1.0841228091157973, 'learning_rate/qwen_vl_interface': 8.70210589671356e-06, 'learning_rate/action_model': 8.702105896713561e-05, 'epoch': 0.1}) +[2026-06-24 12:25:23,387][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.3816830515861511, 'train/grad_norm_pre_clip': 60.303287506103516, 'timing/data': 0.00040220399387180805, 'timing/model': 1.083734941901639, 'learning_rate/qwen_vl_interface': 8.699476521483952e-06, 'learning_rate/action_model': 8.699476521483952e-05, 'epoch': 0.1}) +[2026-06-24 12:25:24,467][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.07358727604150772, 'train/grad_norm_pre_clip': 14.716547012329102, 'timing/data': 0.000422121025621891, 'timing/model': 1.077766367001459, 'learning_rate/qwen_vl_interface': 8.696844907926478e-06, 'learning_rate/action_model': 8.696844907926477e-05, 'epoch': 0.1}) +[2026-06-24 12:25:25,548][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.14620579779148102, 'train/grad_norm_pre_clip': 35.26523208618164, 'timing/data': 0.00037160608917474747, 'timing/model': 1.0795960249379277, 'learning_rate/qwen_vl_interface': 8.694211057748766e-06, 'learning_rate/action_model': 8.694211057748765e-05, 'epoch': 0.1}) +[2026-06-24 12:25:26,626][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.09567566215991974, 'train/grad_norm_pre_clip': 31.41518783569336, 'timing/data': 0.00038506207056343555, 'timing/model': 1.0762710140552372, 'learning_rate/qwen_vl_interface': 8.69157497265989e-06, 'learning_rate/action_model': 8.69157497265989e-05, 'epoch': 0.1}) +[2026-06-24 12:25:27,713][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.15907436609268188, 'train/grad_norm_pre_clip': 30.5102596282959, 'timing/data': 0.0005491599440574646, 'timing/model': 1.0853607789613307, 'learning_rate/qwen_vl_interface': 8.68893665437038e-06, 'learning_rate/action_model': 8.688936654370378e-05, 'epoch': 0.1}) +[2026-06-24 12:25:28,789][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.12331084907054901, 'train/grad_norm_pre_clip': 33.753623962402344, 'timing/data': 0.0003699241206049919, 'timing/model': 1.0739538120105863, 'learning_rate/qwen_vl_interface': 8.68629610459221e-06, 'learning_rate/action_model': 8.686296104592209e-05, 'epoch': 0.1}) +[2026-06-24 12:25:29,868][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.18062850832939148, 'train/grad_norm_pre_clip': 53.48747253417969, 'timing/data': 0.0004214700311422348, 'timing/model': 1.0770084669347852, 'learning_rate/qwen_vl_interface': 8.683653325038805e-06, 'learning_rate/action_model': 8.683653325038804e-05, 'epoch': 0.1}) +[2026-06-24 12:25:30,946][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.21358288824558258, 'train/grad_norm_pre_clip': 33.95384979248047, 'timing/data': 0.00038099405355751514, 'timing/model': 1.0763368878979236, 'learning_rate/qwen_vl_interface': 8.681008317425037e-06, 'learning_rate/action_model': 8.681008317425037e-05, 'epoch': 0.1}) +[2026-06-24 12:25:32,030][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.0362040251493454, 'train/grad_norm_pre_clip': 7.649197101593018, 'timing/data': 0.0003790601622313261, 'timing/model': 1.0816468989942223, 'learning_rate/qwen_vl_interface': 8.678361083467221e-06, 'learning_rate/action_model': 8.67836108346722e-05, 'epoch': 0.1}) +[2026-06-24 12:25:33,111][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.0384536013007164, 'train/grad_norm_pre_clip': 8.055689811706543, 'timing/data': 0.0004502029623836279, 'timing/model': 1.0793602438643575, 'learning_rate/qwen_vl_interface': 8.67571162488312e-06, 'learning_rate/action_model': 8.67571162488312e-05, 'epoch': 0.1}) +[2026-06-24 12:25:34,191][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.173735111951828, 'train/grad_norm_pre_clip': 47.150413513183594, 'timing/data': 0.0004594409838318825, 'timing/model': 1.0782821639440954, 'learning_rate/qwen_vl_interface': 8.673059943391937e-06, 'learning_rate/action_model': 8.673059943391936e-05, 'epoch': 0.1}) +[2026-06-24 12:25:35,552][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.2238996922969818, 'train/grad_norm_pre_clip': 52.969364166259766, 'timing/data': 0.0003801428247243166, 'timing/model': 1.3598591010086238, 'learning_rate/qwen_vl_interface': 8.670406040714321e-06, 'learning_rate/action_model': 8.670406040714321e-05, 'epoch': 0.1}) +[2026-06-24 12:25:36,628][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.06123647466301918, 'train/grad_norm_pre_clip': 22.419391632080078, 'timing/data': 0.0003162529319524765, 'timing/model': 1.0737313760910183, 'learning_rate/qwen_vl_interface': 8.667749918572361e-06, 'learning_rate/action_model': 8.667749918572361e-05, 'epoch': 0.1}) +[2026-06-24 12:25:37,703][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.09086772054433823, 'train/grad_norm_pre_clip': 43.635345458984375, 'timing/data': 0.0003461390733718872, 'timing/model': 1.0734298508614302, 'learning_rate/qwen_vl_interface': 8.665091578689584e-06, 'learning_rate/action_model': 8.665091578689584e-05, 'epoch': 0.1}) +[2026-06-24 12:25:38,775][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.11065695434808731, 'train/grad_norm_pre_clip': 46.146331787109375, 'timing/data': 0.0003806839231401682, 'timing/model': 1.0703512411564589, 'learning_rate/qwen_vl_interface': 8.662431022790958e-06, 'learning_rate/action_model': 8.662431022790958e-05, 'epoch': 0.1}) +[2026-06-24 12:25:39,850][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.09370803087949753, 'train/grad_norm_pre_clip': 54.72958755493164, 'timing/data': 0.0003339559771120548, 'timing/model': 1.0736822839826345, 'learning_rate/qwen_vl_interface': 8.65976825260289e-06, 'learning_rate/action_model': 8.659768252602888e-05, 'epoch': 0.1}) +[2026-06-24 12:25:40,931][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.1947900652885437, 'train/grad_norm_pre_clip': 83.8631820678711, 'timing/data': 0.0003297280054539442, 'timing/model': 1.0796276358887553, 'learning_rate/qwen_vl_interface': 8.657103269853222e-06, 'learning_rate/action_model': 8.657103269853222e-05, 'epoch': 0.1}) +[2026-06-24 12:25:42,018][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.04644226282835007, 'train/grad_norm_pre_clip': 32.27078628540039, 'timing/data': 0.00040085194632411003, 'timing/model': 1.0850954230409116, 'learning_rate/qwen_vl_interface': 8.654436076271232e-06, 'learning_rate/action_model': 8.654436076271232e-05, 'epoch': 0.1}) +[2026-06-24 12:25:43,106][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.03463686630129814, 'train/grad_norm_pre_clip': 21.7524471282959, 'timing/data': 0.0005000971723347902, 'timing/model': 1.085718089947477, 'learning_rate/qwen_vl_interface': 8.651766673587635e-06, 'learning_rate/action_model': 8.651766673587635e-05, 'epoch': 0.1}) +[2026-06-24 12:25:44,193][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.04168650507926941, 'train/grad_norm_pre_clip': 23.74761199951172, 'timing/data': 0.000444894190877676, 'timing/model': 1.0851721270009875, 'learning_rate/qwen_vl_interface': 8.649095063534576e-06, 'learning_rate/action_model': 8.649095063534576e-05, 'epoch': 0.1}) +[2026-06-24 12:25:45,278][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.011428967118263245, 'train/grad_norm_pre_clip': 8.446568489074707, 'timing/data': 0.0004664938896894455, 'timing/model': 1.0834026418160647, 'learning_rate/qwen_vl_interface': 8.646421247845633e-06, 'learning_rate/action_model': 8.646421247845633e-05, 'epoch': 0.1}) +[2026-06-24 12:25:46,364][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.021133974194526672, 'train/grad_norm_pre_clip': 148.0925750732422, 'timing/data': 0.00046876794658601284, 'timing/model': 1.0839119551237673, 'learning_rate/qwen_vl_interface': 8.643745228255816e-06, 'learning_rate/action_model': 8.643745228255817e-05, 'epoch': 0.1}) +[2026-06-24 12:25:47,450][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.0787121132016182, 'train/grad_norm_pre_clip': 88.77262878417969, 'timing/data': 0.0004647308960556984, 'timing/model': 1.0841297039296478, 'learning_rate/qwen_vl_interface': 8.641067006501568e-06, 'learning_rate/action_model': 8.641067006501567e-05, 'epoch': 0.1}) +[2026-06-24 12:25:48,535][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.34273770451545715, 'train/grad_norm_pre_clip': 73.06710052490234, 'timing/data': 0.00048140203580260277, 'timing/model': 1.0830961770843714, 'learning_rate/qwen_vl_interface': 8.638386584320753e-06, 'learning_rate/action_model': 8.638386584320752e-05, 'epoch': 0.1}) +[2026-06-24 12:25:49,623][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.12900611758232117, 'train/grad_norm_pre_clip': 53.55874252319336, 'timing/data': 0.00045393011532723904, 'timing/model': 1.085554523859173, 'learning_rate/qwen_vl_interface': 8.635703963452673e-06, 'learning_rate/action_model': 8.635703963452673e-05, 'epoch': 0.1}) +[2026-06-24 12:25:50,712][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.14656522870063782, 'train/grad_norm_pre_clip': 32.53401565551758, 'timing/data': 0.0004454650916159153, 'timing/model': 1.0877106748521328, 'learning_rate/qwen_vl_interface': 8.633019145638048e-06, 'learning_rate/action_model': 8.633019145638048e-05, 'epoch': 0.1}) +[2026-06-24 12:25:51,797][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.15054012835025787, 'train/grad_norm_pre_clip': 68.08145141601562, 'timing/data': 0.0003867349587380886, 'timing/model': 1.082499478943646, 'learning_rate/qwen_vl_interface': 8.630332132619029e-06, 'learning_rate/action_model': 8.630332132619028e-05, 'epoch': 0.1}) +[2026-06-24 12:25:53,169][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.040157824754714966, 'train/grad_norm_pre_clip': 20.941944122314453, 'timing/data': 0.00043906294740736485, 'timing/model': 1.3709037320222706, 'learning_rate/qwen_vl_interface': 8.627642926139187e-06, 'learning_rate/action_model': 8.627642926139187e-05, 'epoch': 0.1}) +[2026-06-24 12:25:54,253][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.4210972785949707, 'train/grad_norm_pre_clip': 77.11275482177734, 'timing/data': 0.0004158590454608202, 'timing/model': 1.0814842469990253, 'learning_rate/qwen_vl_interface': 8.624951527943522e-06, 'learning_rate/action_model': 8.624951527943522e-05, 'epoch': 0.1}) +[2026-06-24 12:25:55,335][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.2154700756072998, 'train/grad_norm_pre_clip': 35.98896408081055, 'timing/data': 0.00046243704855442047, 'timing/model': 1.0805316620972008, 'learning_rate/qwen_vl_interface': 8.622257939778453e-06, 'learning_rate/action_model': 8.622257939778453e-05, 'epoch': 0.1}) +[2026-06-24 12:25:56,419][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.05907968431711197, 'train/grad_norm_pre_clip': 19.163169860839844, 'timing/data': 0.00043588713742792606, 'timing/model': 1.0825843180064112, 'learning_rate/qwen_vl_interface': 8.619562163391818e-06, 'learning_rate/action_model': 8.619562163391817e-05, 'epoch': 0.1}) +[2026-06-24 12:25:57,503][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.03522812947630882, 'train/grad_norm_pre_clip': 16.078472137451172, 'timing/data': 0.0004914910532534122, 'timing/model': 1.0812669990118593, 'learning_rate/qwen_vl_interface': 8.616864200532877e-06, 'learning_rate/action_model': 8.616864200532877e-05, 'epoch': 0.1}) +[2026-06-24 12:25:58,582][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.11301817744970322, 'train/grad_norm_pre_clip': 45.32933044433594, 'timing/data': 0.0003636518958956003, 'timing/model': 1.0782036590389907, 'learning_rate/qwen_vl_interface': 8.61416405295231e-06, 'learning_rate/action_model': 8.61416405295231e-05, 'epoch': 0.1}) +[2026-06-24 12:25:59,664][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.1379118710756302, 'train/grad_norm_pre_clip': 34.75594711303711, 'timing/data': 0.0003278539516031742, 'timing/model': 1.079670697916299, 'learning_rate/qwen_vl_interface': 8.611461722402214e-06, 'learning_rate/action_model': 8.611461722402215e-05, 'epoch': 0.1}) +[2026-06-24 12:26:00,749][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.03216546028852463, 'train/grad_norm_pre_clip': 10.147849082946777, 'timing/data': 0.0003619480412453413, 'timing/model': 1.0833753109909594, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.1}) +[2026-06-24 12:26:01,832][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.15175925195217133, 'train/grad_norm_pre_clip': 42.661624908447266, 'timing/data': 0.000463258009403944, 'timing/model': 1.0813728591892868, 'learning_rate/qwen_vl_interface': 8.606050519408901e-06, 'learning_rate/action_model': 8.606050519408901e-05, 'epoch': 0.1}) +[2026-06-24 12:26:02,912][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.11476440727710724, 'train/grad_norm_pre_clip': 39.91354751586914, 'timing/data': 0.0003685108385980129, 'timing/model': 1.0779252969659865, 'learning_rate/qwen_vl_interface': 8.603341650476953e-06, 'learning_rate/action_model': 8.603341650476952e-05, 'epoch': 0.1}) +[2026-06-24 12:26:03,988][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.38412943482398987, 'train/grad_norm_pre_clip': 50.99371337890625, 'timing/data': 0.0003710559103637934, 'timing/model': 1.07439374900423, 'learning_rate/qwen_vl_interface': 8.600630605598016e-06, 'learning_rate/action_model': 8.600630605598015e-05, 'epoch': 0.1}) +[2026-06-24 12:26:05,066][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.19369979202747345, 'train/grad_norm_pre_clip': 32.705406188964844, 'timing/data': 0.0003468089271336794, 'timing/model': 1.0763910319656134, 'learning_rate/qwen_vl_interface': 8.597917386531255e-06, 'learning_rate/action_model': 8.597917386531254e-05, 'epoch': 0.1}) +[2026-06-24 12:26:06,144][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.2984941899776459, 'train/grad_norm_pre_clip': 66.11502075195312, 'timing/data': 0.0003446361515671015, 'timing/model': 1.0760634890757501, 'learning_rate/qwen_vl_interface': 8.595201995037252e-06, 'learning_rate/action_model': 8.595201995037251e-05, 'epoch': 0.1}) +[2026-06-24 12:26:07,248][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.1605856716632843, 'train/grad_norm_pre_clip': 42.25347137451172, 'timing/data': 0.0005153960082679987, 'timing/model': 1.1019084937870502, 'learning_rate/qwen_vl_interface': 8.59248443287799e-06, 'learning_rate/action_model': 8.592484432877989e-05, 'epoch': 0.1}) +[2026-06-24 12:26:08,337][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.22256793081760406, 'train/grad_norm_pre_clip': 46.48733901977539, 'timing/data': 0.0005203851033002138, 'timing/model': 1.0872543819714338, 'learning_rate/qwen_vl_interface': 8.589764701816868e-06, 'learning_rate/action_model': 8.589764701816868e-05, 'epoch': 0.1}) +[2026-06-24 12:26:09,419][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.16156105697155, 'train/grad_norm_pre_clip': 51.09549331665039, 'timing/data': 0.00047055212780833244, 'timing/model': 1.0804890820290893, 'learning_rate/qwen_vl_interface': 8.587042803618693e-06, 'learning_rate/action_model': 8.587042803618693e-05, 'epoch': 0.1}) +[2026-06-24 12:26:10,770][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.20466798543930054, 'train/grad_norm_pre_clip': 59.706565856933594, 'timing/data': 0.00033519789576530457, 'timing/model': 1.3495038580149412, 'learning_rate/qwen_vl_interface': 8.584318740049672e-06, 'learning_rate/action_model': 8.584318740049672e-05, 'epoch': 0.1}) +[2026-06-24 12:26:11,849][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.19473247230052948, 'train/grad_norm_pre_clip': 56.45269012451172, 'timing/data': 0.00034802197478711605, 'timing/model': 1.0774392769671977, 'learning_rate/qwen_vl_interface': 8.581592512877419e-06, 'learning_rate/action_model': 8.581592512877418e-05, 'epoch': 0.1}) +[2026-06-24 12:26:12,931][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.13785463571548462, 'train/grad_norm_pre_clip': 32.10131072998047, 'timing/data': 0.00035469490103423595, 'timing/model': 1.0805612078402191, 'learning_rate/qwen_vl_interface': 8.578864123870955e-06, 'learning_rate/action_model': 8.578864123870955e-05, 'epoch': 0.11}) +[2026-06-24 12:26:14,013][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.15488837659358978, 'train/grad_norm_pre_clip': 20.17588996887207, 'timing/data': 0.0003306798171252012, 'timing/model': 1.0799792970065027, 'learning_rate/qwen_vl_interface': 8.576133574800703e-06, 'learning_rate/action_model': 8.576133574800703e-05, 'epoch': 0.11}) +[2026-06-24 12:26:15,095][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.09197259694337845, 'train/grad_norm_pre_clip': 25.000572204589844, 'timing/data': 0.0003439041320234537, 'timing/model': 1.080125241074711, 'learning_rate/qwen_vl_interface': 8.573400867438486e-06, 'learning_rate/action_model': 8.573400867438486e-05, 'epoch': 0.11}) +[2026-06-24 12:26:16,174][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.08268584311008453, 'train/grad_norm_pre_clip': 17.211875915527344, 'timing/data': 0.0003480222076177597, 'timing/model': 1.0773878209292889, 'learning_rate/qwen_vl_interface': 8.570666003557528e-06, 'learning_rate/action_model': 8.570666003557528e-05, 'epoch': 0.11}) +[2026-06-24 12:26:17,249][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.12321389466524124, 'train/grad_norm_pre_clip': 49.63576889038086, 'timing/data': 0.0003509370144456625, 'timing/model': 1.0735124580096453, 'learning_rate/qwen_vl_interface': 8.567928984932452e-06, 'learning_rate/action_model': 8.567928984932452e-05, 'epoch': 0.11}) +[2026-06-24 12:26:18,330][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.15813028812408447, 'train/grad_norm_pre_clip': 31.65164566040039, 'timing/data': 0.00036351103335618973, 'timing/model': 1.0796137819997966, 'learning_rate/qwen_vl_interface': 8.565189813339278e-06, 'learning_rate/action_model': 8.565189813339278e-05, 'epoch': 0.11}) +[2026-06-24 12:26:19,407][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.17896437644958496, 'train/grad_norm_pre_clip': 83.76116943359375, 'timing/data': 0.00045221811160445213, 'timing/model': 1.0744513380341232, 'learning_rate/qwen_vl_interface': 8.562448490555429e-06, 'learning_rate/action_model': 8.562448490555429e-05, 'epoch': 0.11}) +[2026-06-24 12:26:20,489][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.11655120551586151, 'train/grad_norm_pre_clip': 23.533851623535156, 'timing/data': 0.0003863237798213959, 'timing/model': 1.080734762828797, 'learning_rate/qwen_vl_interface': 8.559705018359716e-06, 'learning_rate/action_model': 8.559705018359716e-05, 'epoch': 0.11}) +[2026-06-24 12:26:21,568][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.05376837030053139, 'train/grad_norm_pre_clip': 24.341413497924805, 'timing/data': 0.0003348968457430601, 'timing/model': 1.077230495866388, 'learning_rate/qwen_vl_interface': 8.55695939853235e-06, 'learning_rate/action_model': 8.55695939853235e-05, 'epoch': 0.11}) +[2026-06-24 12:26:22,647][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.3600621521472931, 'train/grad_norm_pre_clip': 67.09524536132812, 'timing/data': 0.0003378530964255333, 'timing/model': 1.0770145610440522, 'learning_rate/qwen_vl_interface': 8.554211632854932e-06, 'learning_rate/action_model': 8.554211632854932e-05, 'epoch': 0.11}) +[2026-06-24 12:26:23,724][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.22371776401996613, 'train/grad_norm_pre_clip': 52.04677963256836, 'timing/data': 0.00036261999048292637, 'timing/model': 1.0754431569948792, 'learning_rate/qwen_vl_interface': 8.551461723110461e-06, 'learning_rate/action_model': 8.551461723110461e-05, 'epoch': 0.11}) +[2026-06-24 12:26:24,803][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.37049001455307007, 'train/grad_norm_pre_clip': 104.83648681640625, 'timing/data': 0.0003359699621796608, 'timing/model': 1.077593226917088, 'learning_rate/qwen_vl_interface': 8.548709671083321e-06, 'learning_rate/action_model': 8.54870967108332e-05, 'epoch': 0.11}) +[2026-06-24 12:26:25,879][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.08704405277967453, 'train/grad_norm_pre_clip': 38.20134353637695, 'timing/data': 0.00031921803019940853, 'timing/model': 1.074418436968699, 'learning_rate/qwen_vl_interface': 8.545955478559288e-06, 'learning_rate/action_model': 8.545955478559288e-05, 'epoch': 0.11}) +[2026-06-24 12:26:26,958][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.112375408411026, 'train/grad_norm_pre_clip': 37.19390869140625, 'timing/data': 0.0003107120282948017, 'timing/model': 1.0773972088936716, 'learning_rate/qwen_vl_interface': 8.54319914732553e-06, 'learning_rate/action_model': 8.543199147325529e-05, 'epoch': 0.11}) +[2026-06-24 12:26:28,299][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.34339889883995056, 'train/grad_norm_pre_clip': 82.66600036621094, 'timing/data': 0.00037874002009630203, 'timing/model': 1.3396874628961086, 'learning_rate/qwen_vl_interface': 8.540440679170599e-06, 'learning_rate/action_model': 8.540440679170598e-05, 'epoch': 0.11}) +[2026-06-24 12:26:29,374][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.06859949231147766, 'train/grad_norm_pre_clip': 22.898731231689453, 'timing/data': 0.0003719970118254423, 'timing/model': 1.072806014912203, 'learning_rate/qwen_vl_interface': 8.537680075884433e-06, 'learning_rate/action_model': 8.537680075884433e-05, 'epoch': 0.11}) +[2026-06-24 12:26:30,453][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.21539610624313354, 'train/grad_norm_pre_clip': 246.8334197998047, 'timing/data': 0.0003738801460713148, 'timing/model': 1.0773455020971596, 'learning_rate/qwen_vl_interface': 8.53491733925836e-06, 'learning_rate/action_model': 8.53491733925836e-05, 'epoch': 0.11}) +[2026-06-24 12:26:31,528][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.07709280401468277, 'train/grad_norm_pre_clip': 29.481863021850586, 'timing/data': 0.00032292609103024006, 'timing/model': 1.0733247068710625, 'learning_rate/qwen_vl_interface': 8.532152471085092e-06, 'learning_rate/action_model': 8.53215247108509e-05, 'epoch': 0.11}) +[2026-06-24 12:26:32,606][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.12146840244531631, 'train/grad_norm_pre_clip': 31.24639129638672, 'timing/data': 0.00033346517011523247, 'timing/model': 1.0764799700118601, 'learning_rate/qwen_vl_interface': 8.529385473158717e-06, 'learning_rate/action_model': 8.529385473158717e-05, 'epoch': 0.11}) +[2026-06-24 12:26:33,695][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.03158930316567421, 'train/grad_norm_pre_clip': 19.036333084106445, 'timing/data': 0.0003252290189266205, 'timing/model': 1.0875419708900154, 'learning_rate/qwen_vl_interface': 8.526616347274712e-06, 'learning_rate/action_model': 8.526616347274712e-05, 'epoch': 0.11}) +[2026-06-24 12:26:34,779][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.11978546530008316, 'train/grad_norm_pre_clip': 42.52426528930664, 'timing/data': 0.0004067630507051945, 'timing/model': 1.081708800047636, 'learning_rate/qwen_vl_interface': 8.523845095229935e-06, 'learning_rate/action_model': 8.523845095229935e-05, 'epoch': 0.11}) +[2026-06-24 12:26:35,858][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.1846228539943695, 'train/grad_norm_pre_clip': 37.43301010131836, 'timing/data': 0.0003173849545419216, 'timing/model': 1.077764007030055, 'learning_rate/qwen_vl_interface': 8.521071718822616e-06, 'learning_rate/action_model': 8.521071718822616e-05, 'epoch': 0.11}) +[2026-06-24 12:26:36,934][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.319697767496109, 'train/grad_norm_pre_clip': 46.70087432861328, 'timing/data': 0.00034198001958429813, 'timing/model': 1.0746921601239592, 'learning_rate/qwen_vl_interface': 8.518296219852375e-06, 'learning_rate/action_model': 8.518296219852375e-05, 'epoch': 0.11}) +[2026-06-24 12:26:38,016][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.13774651288986206, 'train/grad_norm_pre_clip': 49.19289779663086, 'timing/data': 0.0003126261290162802, 'timing/model': 1.0800709510222077, 'learning_rate/qwen_vl_interface': 8.515518600120201e-06, 'learning_rate/action_model': 8.5155186001202e-05, 'epoch': 0.11}) +[2026-06-24 12:26:39,094][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.2684474289417267, 'train/grad_norm_pre_clip': 104.67365264892578, 'timing/data': 0.0003319820389151573, 'timing/model': 1.0764230948407203, 'learning_rate/qwen_vl_interface': 8.512738861428458e-06, 'learning_rate/action_model': 8.512738861428458e-05, 'epoch': 0.11}) +[2026-06-24 12:26:40,174][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.07656808197498322, 'train/grad_norm_pre_clip': 26.097461700439453, 'timing/data': 0.0003361701965332031, 'timing/model': 1.0781566728837788, 'learning_rate/qwen_vl_interface': 8.509957005580892e-06, 'learning_rate/action_model': 8.509957005580892e-05, 'epoch': 0.11}) +[2026-06-24 12:26:41,254][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.03547127917408943, 'train/grad_norm_pre_clip': 30.151275634765625, 'timing/data': 0.00032575102522969246, 'timing/model': 1.0784264889080077, 'learning_rate/qwen_vl_interface': 8.507173034382617e-06, 'learning_rate/action_model': 8.507173034382617e-05, 'epoch': 0.11}) +[2026-06-24 12:26:42,336][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.04570131003856659, 'train/grad_norm_pre_clip': 28.688940048217773, 'timing/data': 0.0003428119234740734, 'timing/model': 1.0802379741799086, 'learning_rate/qwen_vl_interface': 8.504386949640124e-06, 'learning_rate/action_model': 8.504386949640124e-05, 'epoch': 0.11}) +[2026-06-24 12:26:43,412][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.35444074869155884, 'train/grad_norm_pre_clip': 59.39303970336914, 'timing/data': 0.00037595396861433983, 'timing/model': 1.0742775332182646, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.11}) +[2026-06-24 12:26:44,490][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.1113874614238739, 'train/grad_norm_pre_clip': 43.74222946166992, 'timing/data': 0.0003335548099130392, 'timing/model': 1.0764755019918084, 'learning_rate/qwen_vl_interface': 8.498808446755291e-06, 'learning_rate/action_model': 8.498808446755291e-05, 'epoch': 0.11}) +[2026-06-24 12:26:45,834][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.20394539833068848, 'train/grad_norm_pre_clip': 47.860443115234375, 'timing/data': 0.0003449569921940565, 'timing/model': 1.343248587101698, 'learning_rate/qwen_vl_interface': 8.49601603223278e-06, 'learning_rate/action_model': 8.49601603223278e-05, 'epoch': 0.11}) +[2026-06-24 12:26:46,912][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.10738784819841385, 'train/grad_norm_pre_clip': 50.46300506591797, 'timing/data': 0.0003252788446843624, 'timing/model': 1.0764680888969451, 'learning_rate/qwen_vl_interface': 8.493221511405708e-06, 'learning_rate/action_model': 8.493221511405707e-05, 'epoch': 0.11}) +[2026-06-24 12:26:47,994][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.1464826762676239, 'train/grad_norm_pre_clip': 38.86594772338867, 'timing/data': 0.0003108130767941475, 'timing/model': 1.0797486470546573, 'learning_rate/qwen_vl_interface': 8.490424886087407e-06, 'learning_rate/action_model': 8.490424886087407e-05, 'epoch': 0.11}) +[2026-06-24 12:26:49,073][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.14351394772529602, 'train/grad_norm_pre_clip': 50.72163772583008, 'timing/data': 0.0005075619556009769, 'timing/model': 1.0772133260034025, 'learning_rate/qwen_vl_interface': 8.487626158092578e-06, 'learning_rate/action_model': 8.487626158092579e-05, 'epoch': 0.11}) +[2026-06-24 12:26:50,152][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.12690134346485138, 'train/grad_norm_pre_clip': 54.92026901245117, 'timing/data': 0.0005698380991816521, 'timing/model': 1.076783841010183, 'learning_rate/qwen_vl_interface': 8.484825329237287e-06, 'learning_rate/action_model': 8.484825329237287e-05, 'epoch': 0.11}) +[2026-06-24 12:26:51,231][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.028612898662686348, 'train/grad_norm_pre_clip': 13.418086051940918, 'timing/data': 0.000359133118763566, 'timing/model': 1.0782801050227135, 'learning_rate/qwen_vl_interface': 8.48202240133896e-06, 'learning_rate/action_model': 8.482022401338959e-05, 'epoch': 0.11}) +[2026-06-24 12:26:52,310][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.07124878466129303, 'train/grad_norm_pre_clip': 34.16194534301758, 'timing/data': 0.00043359282426536083, 'timing/model': 1.076330001000315, 'learning_rate/qwen_vl_interface': 8.479217376216385e-06, 'learning_rate/action_model': 8.479217376216385e-05, 'epoch': 0.11}) +[2026-06-24 12:26:53,388][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.39505651593208313, 'train/grad_norm_pre_clip': 42.11815643310547, 'timing/data': 0.0003339259419590235, 'timing/model': 1.077066922094673, 'learning_rate/qwen_vl_interface': 8.476410255689714e-06, 'learning_rate/action_model': 8.476410255689713e-05, 'epoch': 0.11}) +[2026-06-24 12:26:54,470][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.13898834586143494, 'train/grad_norm_pre_clip': 39.82365036010742, 'timing/data': 0.0003305799327790737, 'timing/model': 1.0798569810576737, 'learning_rate/qwen_vl_interface': 8.473601041580457e-06, 'learning_rate/action_model': 8.473601041580457e-05, 'epoch': 0.11}) +[2026-06-24 12:26:55,549][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.17167478799819946, 'train/grad_norm_pre_clip': 45.83113098144531, 'timing/data': 0.00036719790659844875, 'timing/model': 1.0776581100653857, 'learning_rate/qwen_vl_interface': 8.470789735711483e-06, 'learning_rate/action_model': 8.470789735711483e-05, 'epoch': 0.11}) +[2026-06-24 12:26:56,631][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.17105986177921295, 'train/grad_norm_pre_clip': 37.79191589355469, 'timing/data': 0.0003287659492343664, 'timing/model': 1.0807989961467683, 'learning_rate/qwen_vl_interface': 8.467976339907015e-06, 'learning_rate/action_model': 8.467976339907016e-05, 'epoch': 0.11}) +[2026-06-24 12:26:57,712][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.2604016661643982, 'train/grad_norm_pre_clip': 36.752891540527344, 'timing/data': 0.00030469102784991264, 'timing/model': 1.0787825668230653, 'learning_rate/qwen_vl_interface': 8.465160855992639e-06, 'learning_rate/action_model': 8.465160855992638e-05, 'epoch': 0.11}) +[2026-06-24 12:26:58,792][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.08974941074848175, 'train/grad_norm_pre_clip': 19.148061752319336, 'timing/data': 0.00035453401505947113, 'timing/model': 1.078272480983287, 'learning_rate/qwen_vl_interface': 8.462343285795289e-06, 'learning_rate/action_model': 8.462343285795288e-05, 'epoch': 0.11}) +[2026-06-24 12:26:59,875][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.4928242862224579, 'train/grad_norm_pre_clip': 40.88237762451172, 'timing/data': 0.0003249191213399172, 'timing/model': 1.0811213399283588, 'learning_rate/qwen_vl_interface': 8.459523631143259e-06, 'learning_rate/action_model': 8.459523631143258e-05, 'epoch': 0.11}) +[2026-06-24 12:27:00,952][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.04417141154408455, 'train/grad_norm_pre_clip': 13.602913856506348, 'timing/data': 0.0003409278579056263, 'timing/model': 1.0757347350008786, 'learning_rate/qwen_vl_interface': 8.456701893866188e-06, 'learning_rate/action_model': 8.456701893866187e-05, 'epoch': 0.11}) +[2026-06-24 12:27:02,030][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.2736837565898895, 'train/grad_norm_pre_clip': 46.99909210205078, 'timing/data': 0.00035815080627799034, 'timing/model': 1.076422344893217, 'learning_rate/qwen_vl_interface': 8.453878075795072e-06, 'learning_rate/action_model': 8.453878075795072e-05, 'epoch': 0.11}) +[2026-06-24 12:27:03,387][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.21913693845272064, 'train/grad_norm_pre_clip': 49.90067672729492, 'timing/data': 0.0003327140584588051, 'timing/model': 1.355457038152963, 'learning_rate/qwen_vl_interface': 8.451052178762258e-06, 'learning_rate/action_model': 8.451052178762258e-05, 'epoch': 0.11}) +[2026-06-24 12:27:04,466][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.20720422267913818, 'train/grad_norm_pre_clip': 36.293418884277344, 'timing/data': 0.0003378030378371477, 'timing/model': 1.0773548420984298, 'learning_rate/qwen_vl_interface': 8.448224204601439e-06, 'learning_rate/action_model': 8.448224204601438e-05, 'epoch': 0.11}) +[2026-06-24 12:27:05,559][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.21823976933956146, 'train/grad_norm_pre_clip': 27.46940040588379, 'timing/data': 0.0004030859563499689, 'timing/model': 1.0823983040172607, 'learning_rate/qwen_vl_interface': 8.445394155147655e-06, 'learning_rate/action_model': 8.445394155147653e-05, 'epoch': 0.11}) +[2026-06-24 12:27:06,648][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.15021400153636932, 'train/grad_norm_pre_clip': 27.790332794189453, 'timing/data': 0.00046644406393170357, 'timing/model': 1.0867966569494456, 'learning_rate/qwen_vl_interface': 8.442562032237295e-06, 'learning_rate/action_model': 8.442562032237296e-05, 'epoch': 0.11}) +[2026-06-24 12:27:07,733][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.1673777997493744, 'train/grad_norm_pre_clip': 42.418880462646484, 'timing/data': 0.00048693292774260044, 'timing/model': 1.083086604019627, 'learning_rate/qwen_vl_interface': 8.439727837708094e-06, 'learning_rate/action_model': 8.439727837708095e-05, 'epoch': 0.11}) +[2026-06-24 12:27:08,820][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.2960544526576996, 'train/grad_norm_pre_clip': 62.295745849609375, 'timing/data': 0.0004396340809762478, 'timing/model': 1.0848741340450943, 'learning_rate/qwen_vl_interface': 8.436891573399132e-06, 'learning_rate/action_model': 8.436891573399131e-05, 'epoch': 0.11}) +[2026-06-24 12:27:09,901][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.19992807507514954, 'train/grad_norm_pre_clip': 37.802459716796875, 'timing/data': 0.00048511894419789314, 'timing/model': 1.0791789509821683, 'learning_rate/qwen_vl_interface': 8.434053241150826e-06, 'learning_rate/action_model': 8.434053241150825e-05, 'epoch': 0.11}) +[2026-06-24 12:27:10,985][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.13812923431396484, 'train/grad_norm_pre_clip': 24.52005958557129, 'timing/data': 0.0003937678411602974, 'timing/model': 1.0820930819027126, 'learning_rate/qwen_vl_interface': 8.43121284280494e-06, 'learning_rate/action_model': 8.431212842804939e-05, 'epoch': 0.11}) +[2026-06-24 12:27:12,066][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.13359296321868896, 'train/grad_norm_pre_clip': 13.410045623779297, 'timing/data': 0.0004569368902593851, 'timing/model': 1.0791640530806035, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.11}) +[2026-06-24 12:27:13,149][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.17759352922439575, 'train/grad_norm_pre_clip': 22.892358779907227, 'timing/data': 0.0003838299307972193, 'timing/model': 1.0819279130082577, 'learning_rate/qwen_vl_interface': 8.425525855194184e-06, 'learning_rate/action_model': 8.425525855194183e-05, 'epoch': 0.11}) +[2026-06-24 12:27:14,243][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.08198009431362152, 'train/grad_norm_pre_clip': 19.686277389526367, 'timing/data': 0.00045585399493575096, 'timing/model': 1.091811093967408, 'learning_rate/qwen_vl_interface': 8.422679269619538e-06, 'learning_rate/action_model': 8.422679269619538e-05, 'epoch': 0.11}) +[2026-06-24 12:27:15,325][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.21585237979888916, 'train/grad_norm_pre_clip': 39.96544647216797, 'timing/data': 0.0004695397801697254, 'timing/model': 1.0805297328624874, 'learning_rate/qwen_vl_interface': 8.419830625327758e-06, 'learning_rate/action_model': 8.419830625327758e-05, 'epoch': 0.11}) +[2026-06-24 12:27:16,413][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.0925287976861, 'train/grad_norm_pre_clip': 152.9033660888672, 'timing/data': 0.0004533391911536455, 'timing/model': 1.085479258093983, 'learning_rate/qwen_vl_interface': 8.416979924167303e-06, 'learning_rate/action_model': 8.416979924167303e-05, 'epoch': 0.11}) +[2026-06-24 12:27:17,501][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.25081995129585266, 'train/grad_norm_pre_clip': 71.83643341064453, 'timing/data': 0.00040247407741844654, 'timing/model': 1.086002548923716, 'learning_rate/qwen_vl_interface': 8.414127167987957e-06, 'learning_rate/action_model': 8.414127167987957e-05, 'epoch': 0.11}) +[2026-06-24 12:27:18,587][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.04056846722960472, 'train/grad_norm_pre_clip': 21.157196044921875, 'timing/data': 0.00045637483708560467, 'timing/model': 1.085026510991156, 'learning_rate/qwen_vl_interface': 8.411272358640843e-06, 'learning_rate/action_model': 8.411272358640844e-05, 'epoch': 0.11}) +[2026-06-24 12:27:19,672][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.07144951820373535, 'train/grad_norm_pre_clip': 27.569480895996094, 'timing/data': 0.0004010018892586231, 'timing/model': 1.082217135000974, 'learning_rate/qwen_vl_interface': 8.408415497978418e-06, 'learning_rate/action_model': 8.408415497978418e-05, 'epoch': 0.11}) +[2026-06-24 12:27:21,040][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.15860576927661896, 'train/grad_norm_pre_clip': 46.41848373413086, 'timing/data': 0.00045610498636960983, 'timing/model': 1.3667459259741008, 'learning_rate/qwen_vl_interface': 8.405556587854468e-06, 'learning_rate/action_model': 8.405556587854467e-05, 'epoch': 0.11}) +[2026-06-24 12:27:22,123][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.1786556839942932, 'train/grad_norm_pre_clip': 36.55208969116211, 'timing/data': 0.0003775879740715027, 'timing/model': 1.0809707001317292, 'learning_rate/qwen_vl_interface': 8.402695630124106e-06, 'learning_rate/action_model': 8.402695630124106e-05, 'epoch': 0.11}) +[2026-06-24 12:27:23,207][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.09624972194433212, 'train/grad_norm_pre_clip': 37.72323226928711, 'timing/data': 0.0003828878980129957, 'timing/model': 1.081956627080217, 'learning_rate/qwen_vl_interface': 8.399832626643778e-06, 'learning_rate/action_model': 8.399832626643778e-05, 'epoch': 0.11}) +[2026-06-24 12:27:24,293][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.2882532477378845, 'train/grad_norm_pre_clip': 69.88848114013672, 'timing/data': 0.0003757940139621496, 'timing/model': 1.084143707063049, 'learning_rate/qwen_vl_interface': 8.396967579271256e-06, 'learning_rate/action_model': 8.396967579271257e-05, 'epoch': 0.11}) +[2026-06-24 12:27:25,374][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.36612141132354736, 'train/grad_norm_pre_clip': 98.53894805908203, 'timing/data': 0.00043366290628910065, 'timing/model': 1.0792079861275852, 'learning_rate/qwen_vl_interface': 8.394100489865638e-06, 'learning_rate/action_model': 8.394100489865638e-05, 'epoch': 0.11}) +[2026-06-24 12:27:26,461][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.035427555441856384, 'train/grad_norm_pre_clip': 20.1234073638916, 'timing/data': 0.0003997290041297674, 'timing/model': 1.0855622440576553, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.11}) +[2026-06-24 12:27:27,548][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.10257565230131149, 'train/grad_norm_pre_clip': 49.93857955932617, 'timing/data': 0.00043729995377361774, 'timing/model': 1.0847935141064227, 'learning_rate/qwen_vl_interface': 8.388360192398131e-06, 'learning_rate/action_model': 8.38836019239813e-05, 'epoch': 0.11}) +[2026-06-24 12:27:28,634][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.0643424317240715, 'train/grad_norm_pre_clip': 27.901260375976562, 'timing/data': 0.0003966139629483223, 'timing/model': 1.0841779098846018, 'learning_rate/qwen_vl_interface': 8.385486988061058e-06, 'learning_rate/action_model': 8.385486988061058e-05, 'epoch': 0.11}) +[2026-06-24 12:27:29,713][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.33832377195358276, 'train/grad_norm_pre_clip': 45.57349395751953, 'timing/data': 0.00037615513429045677, 'timing/model': 1.077979935100302, 'learning_rate/qwen_vl_interface': 8.38261174914052e-06, 'learning_rate/action_model': 8.38261174914052e-05, 'epoch': 0.11}) +[2026-06-24 12:27:30,795][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.3436770439147949, 'train/grad_norm_pre_clip': 57.66648864746094, 'timing/data': 0.00042639882303774357, 'timing/model': 1.079904019832611, 'learning_rate/qwen_vl_interface': 8.379734477502229e-06, 'learning_rate/action_model': 8.379734477502228e-05, 'epoch': 0.11}) +[2026-06-24 12:27:31,876][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.14393270015716553, 'train/grad_norm_pre_clip': 47.6729850769043, 'timing/data': 0.0004267999902367592, 'timing/model': 1.0787509100046009, 'learning_rate/qwen_vl_interface': 8.376855175013214e-06, 'learning_rate/action_model': 8.376855175013213e-05, 'epoch': 0.11}) +[2026-06-24 12:27:32,952][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.43680232763290405, 'train/grad_norm_pre_clip': 74.91203308105469, 'timing/data': 0.00037024402990937233, 'timing/model': 1.0741497671697289, 'learning_rate/qwen_vl_interface': 8.373973843541822e-06, 'learning_rate/action_model': 8.373973843541822e-05, 'epoch': 0.11}) +[2026-06-24 12:27:34,033][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.07047184556722641, 'train/grad_norm_pre_clip': 18.03893280029297, 'timing/data': 0.0003487539943307638, 'timing/model': 1.0798867889679968, 'learning_rate/qwen_vl_interface': 8.371090484957722e-06, 'learning_rate/action_model': 8.371090484957722e-05, 'epoch': 0.11}) +[2026-06-24 12:27:35,113][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.3260861933231354, 'train/grad_norm_pre_clip': 118.18527221679688, 'timing/data': 0.0003428931813687086, 'timing/model': 1.0785665540024638, 'learning_rate/qwen_vl_interface': 8.368205101131891e-06, 'learning_rate/action_model': 8.368205101131891e-05, 'epoch': 0.11}) +[2026-06-24 12:27:36,192][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.05762339010834694, 'train/grad_norm_pre_clip': 14.337688446044922, 'timing/data': 0.0003715560305863619, 'timing/model': 1.0767524249386042, 'learning_rate/qwen_vl_interface': 8.365317693936624e-06, 'learning_rate/action_model': 8.365317693936623e-05, 'epoch': 0.11}) +[2026-06-24 12:27:37,271][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.08283558487892151, 'train/grad_norm_pre_clip': 17.25912857055664, 'timing/data': 0.0003302181139588356, 'timing/model': 1.0780094300862402, 'learning_rate/qwen_vl_interface': 8.362428265245526e-06, 'learning_rate/action_model': 8.362428265245526e-05, 'epoch': 0.11}) +[2026-06-24 12:27:38,624][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.01999996230006218, 'train/grad_norm_pre_clip': 11.530213356018066, 'timing/data': 0.00032456801272928715, 'timing/model': 1.3508071559481323, 'learning_rate/qwen_vl_interface': 8.359536816933518e-06, 'learning_rate/action_model': 8.359536816933518e-05, 'epoch': 0.11}) +[2026-06-24 12:27:39,703][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.1341836005449295, 'train/grad_norm_pre_clip': 36.408634185791016, 'timing/data': 0.0004881650675088167, 'timing/model': 1.0778239930514246, 'learning_rate/qwen_vl_interface': 8.356643350876832e-06, 'learning_rate/action_model': 8.356643350876832e-05, 'epoch': 0.11}) +[2026-06-24 12:27:40,781][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.052086565643548965, 'train/grad_norm_pre_clip': 15.503695487976074, 'timing/data': 0.0003237570635974407, 'timing/model': 1.0764725299086422, 'learning_rate/qwen_vl_interface': 8.353747868953002e-06, 'learning_rate/action_model': 8.353747868953003e-05, 'epoch': 0.11}) +[2026-06-24 12:27:41,859][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.15942448377609253, 'train/grad_norm_pre_clip': 46.940555572509766, 'timing/data': 0.000331551069393754, 'timing/model': 1.0759656301233917, 'learning_rate/qwen_vl_interface': 8.350850373040876e-06, 'learning_rate/action_model': 8.350850373040876e-05, 'epoch': 0.11}) +[2026-06-24 12:27:42,937][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.28452953696250916, 'train/grad_norm_pre_clip': 43.66636657714844, 'timing/data': 0.00036254990845918655, 'timing/model': 1.0768176370766014, 'learning_rate/qwen_vl_interface': 8.34795086502061e-06, 'learning_rate/action_model': 8.347950865020609e-05, 'epoch': 0.11}) +[2026-06-24 12:27:44,020][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.14371949434280396, 'train/grad_norm_pre_clip': 29.750720977783203, 'timing/data': 0.00032194401137530804, 'timing/model': 1.080836809007451, 'learning_rate/qwen_vl_interface': 8.34504934677366e-06, 'learning_rate/action_model': 8.34504934677366e-05, 'epoch': 0.11}) +[2026-06-24 12:27:45,101][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.15709596872329712, 'train/grad_norm_pre_clip': 36.68425750732422, 'timing/data': 0.00033525796607136726, 'timing/model': 1.0789808810222894, 'learning_rate/qwen_vl_interface': 8.342145820182795e-06, 'learning_rate/action_model': 8.342145820182793e-05, 'epoch': 0.11}) +[2026-06-24 12:27:46,179][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.031049704179167747, 'train/grad_norm_pre_clip': 12.778602600097656, 'timing/data': 0.0003196189645677805, 'timing/model': 1.0768530638888478, 'learning_rate/qwen_vl_interface': 8.339240287132073e-06, 'learning_rate/action_model': 8.339240287132073e-05, 'epoch': 0.11}) +[2026-06-24 12:27:47,271][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.12758798897266388, 'train/grad_norm_pre_clip': 63.21588897705078, 'timing/data': 0.0003711550962179899, 'timing/model': 1.0895158348139375, 'learning_rate/qwen_vl_interface': 8.336332749506872e-06, 'learning_rate/action_model': 8.336332749506872e-05, 'epoch': 0.11}) +[2026-06-24 12:27:48,360][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.09312687814235687, 'train/grad_norm_pre_clip': 38.516632080078125, 'timing/data': 0.0003923350013792515, 'timing/model': 1.0880243289284408, 'learning_rate/qwen_vl_interface': 8.333423209193857e-06, 'learning_rate/action_model': 8.333423209193856e-05, 'epoch': 0.11}) +[2026-06-24 12:27:49,451][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.032221633940935135, 'train/grad_norm_pre_clip': 18.68784523010254, 'timing/data': 0.0003801619168370962, 'timing/model': 1.0891885720193386, 'learning_rate/qwen_vl_interface': 8.330511668080997e-06, 'learning_rate/action_model': 8.330511668080997e-05, 'epoch': 0.11}) +[2026-06-24 12:27:50,538][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.2184169441461563, 'train/grad_norm_pre_clip': 45.725067138671875, 'timing/data': 0.00043910183012485504, 'timing/model': 1.0850748419761658, 'learning_rate/qwen_vl_interface': 8.32759812805756e-06, 'learning_rate/action_model': 8.32759812805756e-05, 'epoch': 0.11}) +[2026-06-24 12:27:51,625][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.12778091430664062, 'train/grad_norm_pre_clip': 26.653141021728516, 'timing/data': 0.0003952810075134039, 'timing/model': 1.08509132405743, 'learning_rate/qwen_vl_interface': 8.324682591014111e-06, 'learning_rate/action_model': 8.32468259101411e-05, 'epoch': 0.11}) +[2026-06-24 12:27:52,716][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.15512385964393616, 'train/grad_norm_pre_clip': 60.02083969116211, 'timing/data': 0.0004134050104767084, 'timing/model': 1.0888059549033642, 'learning_rate/qwen_vl_interface': 8.32176505884251e-06, 'learning_rate/action_model': 8.32176505884251e-05, 'epoch': 0.11}) +[2026-06-24 12:27:53,803][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.07992054522037506, 'train/grad_norm_pre_clip': 48.84346389770508, 'timing/data': 0.00039173406548798084, 'timing/model': 1.0852442600298673, 'learning_rate/qwen_vl_interface': 8.318845533435913e-06, 'learning_rate/action_model': 8.318845533435913e-05, 'epoch': 0.11}) +[2026-06-24 12:27:54,892][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.232231006026268, 'train/grad_norm_pre_clip': 53.42744827270508, 'timing/data': 0.00041765207424759865, 'timing/model': 1.0877561471424997, 'learning_rate/qwen_vl_interface': 8.315924016688767e-06, 'learning_rate/action_model': 8.315924016688766e-05, 'epoch': 0.11}) +[2026-06-24 12:27:56,271][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.21568064391613007, 'train/grad_norm_pre_clip': 73.03749084472656, 'timing/data': 0.00043789995834231377, 'timing/model': 1.377061320003122, 'learning_rate/qwen_vl_interface': 8.313000510496811e-06, 'learning_rate/action_model': 8.313000510496811e-05, 'epoch': 0.11}) +[2026-06-24 12:27:57,358][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.21642246842384338, 'train/grad_norm_pre_clip': 76.18311309814453, 'timing/data': 0.0003697429783642292, 'timing/model': 1.0854811249300838, 'learning_rate/qwen_vl_interface': 8.31007501675708e-06, 'learning_rate/action_model': 8.310075016757079e-05, 'epoch': 0.11}) +[2026-06-24 12:27:58,445][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.23424547910690308, 'train/grad_norm_pre_clip': 66.62908935546875, 'timing/data': 0.00043715001083910465, 'timing/model': 1.0843927939422429, 'learning_rate/qwen_vl_interface': 8.307147537367894e-06, 'learning_rate/action_model': 8.307147537367893e-05, 'epoch': 0.11}) +[2026-06-24 12:27:59,530][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.24350978434085846, 'train/grad_norm_pre_clip': 98.86810302734375, 'timing/data': 0.00045958091504871845, 'timing/model': 1.0834092709701508, 'learning_rate/qwen_vl_interface': 8.30421807422886e-06, 'learning_rate/action_model': 8.30421807422886e-05, 'epoch': 0.11}) +[2026-06-24 12:28:00,616][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.12942829728126526, 'train/grad_norm_pre_clip': 46.81461715698242, 'timing/data': 0.00045057409442961216, 'timing/model': 1.0839126941282302, 'learning_rate/qwen_vl_interface': 8.301286629240876e-06, 'learning_rate/action_model': 8.301286629240876e-05, 'epoch': 0.11}) +[2026-06-24 12:28:01,702][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.1809476912021637, 'train/grad_norm_pre_clip': 80.0797119140625, 'timing/data': 0.0003830869682133198, 'timing/model': 1.083939224947244, 'learning_rate/qwen_vl_interface': 8.298353204306127e-06, 'learning_rate/action_model': 8.298353204306127e-05, 'epoch': 0.11}) +[2026-06-24 12:28:02,785][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.04040759801864624, 'train/grad_norm_pre_clip': 14.928873062133789, 'timing/data': 0.0004877941682934761, 'timing/model': 1.08203110890463, 'learning_rate/qwen_vl_interface': 8.295417801328081e-06, 'learning_rate/action_model': 8.29541780132808e-05, 'epoch': 0.11}) +[2026-06-24 12:28:03,871][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.2825239598751068, 'train/grad_norm_pre_clip': 51.16365051269531, 'timing/data': 0.0003989678807556629, 'timing/model': 1.0838770179543644, 'learning_rate/qwen_vl_interface': 8.292480422211486e-06, 'learning_rate/action_model': 8.292480422211485e-05, 'epoch': 0.11}) +[2026-06-24 12:28:04,958][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.12099627405405045, 'train/grad_norm_pre_clip': 47.84512710571289, 'timing/data': 0.0004467971157282591, 'timing/model': 1.0850494150072336, 'learning_rate/qwen_vl_interface': 8.289541068862377e-06, 'learning_rate/action_model': 8.289541068862377e-05, 'epoch': 0.11}) +[2026-06-24 12:28:06,038][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.07521236687898636, 'train/grad_norm_pre_clip': 47.27712631225586, 'timing/data': 0.00047508999705314636, 'timing/model': 1.0778559849131852, 'learning_rate/qwen_vl_interface': 8.286599743188068e-06, 'learning_rate/action_model': 8.286599743188069e-05, 'epoch': 0.12}) +[2026-06-24 12:28:07,122][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.2836795449256897, 'train/grad_norm_pre_clip': 37.3711051940918, 'timing/data': 0.0003754030913114548, 'timing/model': 1.0828490820713341, 'learning_rate/qwen_vl_interface': 8.283656447097154e-06, 'learning_rate/action_model': 8.283656447097153e-05, 'epoch': 0.12}) +[2026-06-24 12:28:08,208][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.16562846302986145, 'train/grad_norm_pre_clip': 25.838485717773438, 'timing/data': 0.00037057395093142986, 'timing/model': 1.0843438929878175, 'learning_rate/qwen_vl_interface': 8.280711182499507e-06, 'learning_rate/action_model': 8.280711182499506e-05, 'epoch': 0.12}) +[2026-06-24 12:28:09,295][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.18251702189445496, 'train/grad_norm_pre_clip': 35.281280517578125, 'timing/data': 0.0003582609351724386, 'timing/model': 1.0845716500189155, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.12}) +[2026-06-24 12:28:10,382][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.1012723445892334, 'train/grad_norm_pre_clip': 16.749252319335938, 'timing/data': 0.00038897921331226826, 'timing/model': 1.085315584903583, 'learning_rate/qwen_vl_interface': 8.274814755429892e-06, 'learning_rate/action_model': 8.274814755429891e-05, 'epoch': 0.12}) +[2026-06-24 12:28:11,471][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.0754161849617958, 'train/grad_norm_pre_clip': 21.189903259277344, 'timing/data': 0.00040434813126921654, 'timing/model': 1.0878018939401954, 'learning_rate/qwen_vl_interface': 8.271863596784053e-06, 'learning_rate/action_model': 8.271863596784051e-05, 'epoch': 0.12}) +[2026-06-24 12:28:12,559][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.10763105750083923, 'train/grad_norm_pre_clip': 22.059555053710938, 'timing/data': 0.00037509319372475147, 'timing/model': 1.0858572598081082, 'learning_rate/qwen_vl_interface': 8.268910477283732e-06, 'learning_rate/action_model': 8.26891047728373e-05, 'epoch': 0.12}) +[2026-06-24 12:28:13,947][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.24484412372112274, 'train/grad_norm_pre_clip': 18.912817001342773, 'timing/data': 0.0005388399586081505, 'timing/model': 1.385675294091925, 'learning_rate/qwen_vl_interface': 8.265955398845177e-06, 'learning_rate/action_model': 8.265955398845177e-05, 'epoch': 0.12}) +[2026-06-24 12:28:15,032][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.04889126867055893, 'train/grad_norm_pre_clip': 24.305072784423828, 'timing/data': 0.00039525097236037254, 'timing/model': 1.0835370318964124, 'learning_rate/qwen_vl_interface': 8.26299836338591e-06, 'learning_rate/action_model': 8.26299836338591e-05, 'epoch': 0.12}) +[2026-06-24 12:28:16,118][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.0627703070640564, 'train/grad_norm_pre_clip': 14.723929405212402, 'timing/data': 0.0003517288714647293, 'timing/model': 1.0844804900698364, 'learning_rate/qwen_vl_interface': 8.260039372824715e-06, 'learning_rate/action_model': 8.260039372824715e-05, 'epoch': 0.12}) +[2026-06-24 12:28:17,205][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.14494146406650543, 'train/grad_norm_pre_clip': 26.964130401611328, 'timing/data': 0.000369952991604805, 'timing/model': 1.0852389209903777, 'learning_rate/qwen_vl_interface': 8.257078429081652e-06, 'learning_rate/action_model': 8.257078429081651e-05, 'epoch': 0.12}) +[2026-06-24 12:28:18,287][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.342681348323822, 'train/grad_norm_pre_clip': 53.62789535522461, 'timing/data': 0.00042280182242393494, 'timing/model': 1.08061836194247, 'learning_rate/qwen_vl_interface': 8.254115534078041e-06, 'learning_rate/action_model': 8.254115534078042e-05, 'epoch': 0.12}) +[2026-06-24 12:28:19,373][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.23072446882724762, 'train/grad_norm_pre_clip': 58.21607971191406, 'timing/data': 0.0003311608452349901, 'timing/model': 1.0833913090173155, 'learning_rate/qwen_vl_interface': 8.25115068973648e-06, 'learning_rate/action_model': 8.25115068973648e-05, 'epoch': 0.12}) +[2026-06-24 12:28:20,457][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.17604459822177887, 'train/grad_norm_pre_clip': 61.16275405883789, 'timing/data': 0.00035153888165950775, 'timing/model': 1.0828917329199612, 'learning_rate/qwen_vl_interface': 8.248183897980815e-06, 'learning_rate/action_model': 8.248183897980815e-05, 'epoch': 0.12}) +[2026-06-24 12:28:21,540][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.10280939936637878, 'train/grad_norm_pre_clip': 25.462509155273438, 'timing/data': 0.00033341487869620323, 'timing/model': 1.0808086090255529, 'learning_rate/qwen_vl_interface': 8.24521516073617e-06, 'learning_rate/action_model': 8.24521516073617e-05, 'epoch': 0.12}) +[2026-06-24 12:28:22,625][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.09558898955583572, 'train/grad_norm_pre_clip': 20.530487060546875, 'timing/data': 0.00034845294430851936, 'timing/model': 1.0837227900046855, 'learning_rate/qwen_vl_interface': 8.24224447992893e-06, 'learning_rate/action_model': 8.242244479928928e-05, 'epoch': 0.12}) +[2026-06-24 12:28:23,710][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.026421749964356422, 'train/grad_norm_pre_clip': 13.75645923614502, 'timing/data': 0.0003870660439133644, 'timing/model': 1.0838056260254234, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 0.12}) +[2026-06-24 12:28:24,800][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.21260806918144226, 'train/grad_norm_pre_clip': 49.91646957397461, 'timing/data': 0.00030629406683146954, 'timing/model': 1.0885226360987872, 'learning_rate/qwen_vl_interface': 8.236297295338474e-06, 'learning_rate/action_model': 8.236297295338474e-05, 'epoch': 0.12}) +[2026-06-24 12:28:25,883][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.10652372241020203, 'train/grad_norm_pre_clip': 20.174335479736328, 'timing/data': 0.0003346570301800966, 'timing/model': 1.0813403048086911, 'learning_rate/qwen_vl_interface': 8.233320795414327e-06, 'learning_rate/action_model': 8.233320795414327e-05, 'epoch': 0.12}) +[2026-06-24 12:28:26,967][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.010916377417743206, 'train/grad_norm_pre_clip': 3.55448055267334, 'timing/data': 0.0003645939286798239, 'timing/model': 1.0815571418497711, 'learning_rate/qwen_vl_interface': 8.230342359645705e-06, 'learning_rate/action_model': 8.230342359645705e-05, 'epoch': 0.12}) +[2026-06-24 12:28:28,055][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.1328853815793991, 'train/grad_norm_pre_clip': 31.67698097229004, 'timing/data': 0.00033027888275682926, 'timing/model': 1.0869257429149002, 'learning_rate/qwen_vl_interface': 8.227361989965285e-06, 'learning_rate/action_model': 8.227361989965284e-05, 'epoch': 0.12}) +[2026-06-24 12:28:29,140][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.1255025416612625, 'train/grad_norm_pre_clip': 35.30634307861328, 'timing/data': 0.0003198990598320961, 'timing/model': 1.0838046029675752, 'learning_rate/qwen_vl_interface': 8.22437968830699e-06, 'learning_rate/action_model': 8.22437968830699e-05, 'epoch': 0.12}) +[2026-06-24 12:28:30,225][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.0577765628695488, 'train/grad_norm_pre_clip': 53.33761978149414, 'timing/data': 0.0005353628657758236, 'timing/model': 1.0825472769793123, 'learning_rate/qwen_vl_interface': 8.221395456606012e-06, 'learning_rate/action_model': 8.221395456606011e-05, 'epoch': 0.12}) +[2026-06-24 12:28:31,585][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.12367624044418335, 'train/grad_norm_pre_clip': 29.48724365234375, 'timing/data': 0.0003358190879225731, 'timing/model': 1.3585518940817565, 'learning_rate/qwen_vl_interface': 8.21840929679878e-06, 'learning_rate/action_model': 8.218409296798778e-05, 'epoch': 0.12}) +[2026-06-24 12:28:32,666][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.15586254000663757, 'train/grad_norm_pre_clip': 75.68490600585938, 'timing/data': 0.00030909013003110886, 'timing/model': 1.0791370361112058, 'learning_rate/qwen_vl_interface': 8.215421210822982e-06, 'learning_rate/action_model': 8.215421210822982e-05, 'epoch': 0.12}) +[2026-06-24 12:28:33,748][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.15580256283283234, 'train/grad_norm_pre_clip': 82.23590850830078, 'timing/data': 0.0003456980921328068, 'timing/model': 1.0804554880596697, 'learning_rate/qwen_vl_interface': 8.212431200617557e-06, 'learning_rate/action_model': 8.212431200617556e-05, 'epoch': 0.12}) +[2026-06-24 12:28:34,832][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.14216023683547974, 'train/grad_norm_pre_clip': 29.190900802612305, 'timing/data': 0.00037323893047869205, 'timing/model': 1.082684394903481, 'learning_rate/qwen_vl_interface': 8.209439268122689e-06, 'learning_rate/action_model': 8.209439268122688e-05, 'epoch': 0.12}) +[2026-06-24 12:28:35,915][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.09789370745420456, 'train/grad_norm_pre_clip': 21.855703353881836, 'timing/data': 0.0003354479558765888, 'timing/model': 1.080835429020226, 'learning_rate/qwen_vl_interface': 8.206445415279808e-06, 'learning_rate/action_model': 8.206445415279807e-05, 'epoch': 0.12}) +[2026-06-24 12:28:37,000][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.4744863212108612, 'train/grad_norm_pre_clip': 123.48358154296875, 'timing/data': 0.0003815349191427231, 'timing/model': 1.0839203409850597, 'learning_rate/qwen_vl_interface': 8.203449644031595e-06, 'learning_rate/action_model': 8.203449644031596e-05, 'epoch': 0.12}) +[2026-06-24 12:28:38,089][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.17335286736488342, 'train/grad_norm_pre_clip': 88.57971954345703, 'timing/data': 0.00039226608350872993, 'timing/model': 1.0866217941511422, 'learning_rate/qwen_vl_interface': 8.200451956321974e-06, 'learning_rate/action_model': 8.200451956321974e-05, 'epoch': 0.12}) +[2026-06-24 12:28:39,179][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.09530914574861526, 'train/grad_norm_pre_clip': 40.250125885009766, 'timing/data': 0.0003692221362143755, 'timing/model': 1.0886854010168463, 'learning_rate/qwen_vl_interface': 8.197452354096112e-06, 'learning_rate/action_model': 8.197452354096112e-05, 'epoch': 0.12}) +[2026-06-24 12:28:40,267][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.0133624617010355, 'train/grad_norm_pre_clip': 5.76638126373291, 'timing/data': 0.00036415201611816883, 'timing/model': 1.0865629729814827, 'learning_rate/qwen_vl_interface': 8.194450839300418e-06, 'learning_rate/action_model': 8.194450839300417e-05, 'epoch': 0.12}) +[2026-06-24 12:28:41,357][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.10847876965999603, 'train/grad_norm_pre_clip': 25.62786102294922, 'timing/data': 0.0003473409451544285, 'timing/model': 1.0884075809735805, 'learning_rate/qwen_vl_interface': 8.191447413882539e-06, 'learning_rate/action_model': 8.191447413882539e-05, 'epoch': 0.12}) +[2026-06-24 12:28:42,447][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.3539589047431946, 'train/grad_norm_pre_clip': 71.41985321044922, 'timing/data': 0.00036100693978369236, 'timing/model': 1.0880896348971874, 'learning_rate/qwen_vl_interface': 8.188442079791369e-06, 'learning_rate/action_model': 8.188442079791369e-05, 'epoch': 0.12}) +[2026-06-24 12:28:43,537][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.29086509346961975, 'train/grad_norm_pre_clip': 48.827884674072266, 'timing/data': 0.00036865100264549255, 'timing/model': 1.0881353300064802, 'learning_rate/qwen_vl_interface': 8.185434838977036e-06, 'learning_rate/action_model': 8.185434838977037e-05, 'epoch': 0.12}) +[2026-06-24 12:28:44,623][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.1810767948627472, 'train/grad_norm_pre_clip': 60.43284606933594, 'timing/data': 0.00037442194297909737, 'timing/model': 1.084251252003014, 'learning_rate/qwen_vl_interface': 8.182425693390905e-06, 'learning_rate/action_model': 8.182425693390905e-05, 'epoch': 0.12}) +[2026-06-24 12:28:45,710][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.09329245984554291, 'train/grad_norm_pre_clip': 49.42784881591797, 'timing/data': 0.00035989400930702686, 'timing/model': 1.085532272933051, 'learning_rate/qwen_vl_interface': 8.179414644985582e-06, 'learning_rate/action_model': 8.179414644985582e-05, 'epoch': 0.12}) +[2026-06-24 12:28:46,800][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.0904054045677185, 'train/grad_norm_pre_clip': 35.35719299316406, 'timing/data': 0.0003171749413013458, 'timing/model': 1.087758765090257, 'learning_rate/qwen_vl_interface': 8.176401695714897e-06, 'learning_rate/action_model': 8.176401695714895e-05, 'epoch': 0.12}) +[2026-06-24 12:28:47,887][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.16299524903297424, 'train/grad_norm_pre_clip': 36.9990119934082, 'timing/data': 0.00035038613714277744, 'timing/model': 1.0858489458914846, 'learning_rate/qwen_vl_interface': 8.173386847533922e-06, 'learning_rate/action_model': 8.173386847533921e-05, 'epoch': 0.12}) +[2026-06-24 12:28:49,250][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.0875423327088356, 'train/grad_norm_pre_clip': 25.32526969909668, 'timing/data': 0.00037728692404925823, 'timing/model': 1.3614132478833199, 'learning_rate/qwen_vl_interface': 8.170370102398962e-06, 'learning_rate/action_model': 8.17037010239896e-05, 'epoch': 0.12}) +[2026-06-24 12:28:50,337][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.05852161720395088, 'train/grad_norm_pre_clip': 36.97548294067383, 'timing/data': 0.00033868500031530857, 'timing/model': 1.0852775052189827, 'learning_rate/qwen_vl_interface': 8.167351462267548e-06, 'learning_rate/action_model': 8.167351462267548e-05, 'epoch': 0.12}) +[2026-06-24 12:28:51,420][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.0920984297990799, 'train/grad_norm_pre_clip': 15.994709014892578, 'timing/data': 0.00030540209263563156, 'timing/model': 1.0813352970872074, 'learning_rate/qwen_vl_interface': 8.164330929098444e-06, 'learning_rate/action_model': 8.164330929098443e-05, 'epoch': 0.12}) +[2026-06-24 12:28:52,503][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.23139004409313202, 'train/grad_norm_pre_clip': 45.75809860229492, 'timing/data': 0.000367447966709733, 'timing/model': 1.0808311619330198, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.12}) +[2026-06-24 12:28:53,589][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.17665046453475952, 'train/grad_norm_pre_clip': 41.030155181884766, 'timing/data': 0.00036780000664293766, 'timing/model': 1.0845801481045783, 'learning_rate/qwen_vl_interface': 8.158284191488355e-06, 'learning_rate/action_model': 8.158284191488355e-05, 'epoch': 0.12}) +[2026-06-24 12:28:54,678][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.0723247230052948, 'train/grad_norm_pre_clip': 21.69020652770996, 'timing/data': 0.00032815500162541866, 'timing/model': 1.0868144449777901, 'learning_rate/qwen_vl_interface': 8.155257990971032e-06, 'learning_rate/action_model': 8.155257990971031e-05, 'epoch': 0.12}) +[2026-06-24 12:28:55,765][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.12696735560894012, 'train/grad_norm_pre_clip': 58.38862609863281, 'timing/data': 0.0003558569587767124, 'timing/model': 1.0858329059556127, 'learning_rate/qwen_vl_interface': 8.152229905263341e-06, 'learning_rate/action_model': 8.152229905263342e-05, 'epoch': 0.12}) +[2026-06-24 12:28:56,852][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.17506533861160278, 'train/grad_norm_pre_clip': 43.12434005737305, 'timing/data': 0.0004327408969402313, 'timing/model': 1.0850265151821077, 'learning_rate/qwen_vl_interface': 8.149199936330175e-06, 'learning_rate/action_model': 8.149199936330174e-05, 'epoch': 0.12}) +[2026-06-24 12:28:57,938][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.19981330633163452, 'train/grad_norm_pre_clip': 28.807592391967773, 'timing/data': 0.0003935769200325012, 'timing/model': 1.084302949020639, 'learning_rate/qwen_vl_interface': 8.146168086137646e-06, 'learning_rate/action_model': 8.146168086137646e-05, 'epoch': 0.12}) +[2026-06-24 12:28:59,023][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.08717954903841019, 'train/grad_norm_pre_clip': 23.534465789794922, 'timing/data': 0.0003915138076990843, 'timing/model': 1.0838426859118044, 'learning_rate/qwen_vl_interface': 8.143134356653087e-06, 'learning_rate/action_model': 8.143134356653087e-05, 'epoch': 0.12}) +[2026-06-24 12:29:00,111][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.13264821469783783, 'train/grad_norm_pre_clip': 30.183000564575195, 'timing/data': 0.0004347150679677725, 'timing/model': 1.0860430710017681, 'learning_rate/qwen_vl_interface': 8.140098749845056e-06, 'learning_rate/action_model': 8.140098749845055e-05, 'epoch': 0.12}) +[2026-06-24 12:29:01,200][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.06648855656385422, 'train/grad_norm_pre_clip': 14.637290954589844, 'timing/data': 0.0004475880414247513, 'timing/model': 1.086747500114143, 'learning_rate/qwen_vl_interface': 8.137061267683323e-06, 'learning_rate/action_model': 8.137061267683323e-05, 'epoch': 0.12}) +[2026-06-24 12:29:02,286][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.16219276189804077, 'train/grad_norm_pre_clip': 36.7894401550293, 'timing/data': 0.00041008903644979, 'timing/model': 1.0844962319824845, 'learning_rate/qwen_vl_interface': 8.134021912138879e-06, 'learning_rate/action_model': 8.134021912138879e-05, 'epoch': 0.12}) +[2026-06-24 12:29:03,375][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.09180424362421036, 'train/grad_norm_pre_clip': 33.84085464477539, 'timing/data': 0.000330920098349452, 'timing/model': 1.0869448999874294, 'learning_rate/qwen_vl_interface': 8.130980685183927e-06, 'learning_rate/action_model': 8.130980685183926e-05, 'epoch': 0.12}) +[2026-06-24 12:29:04,460][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.09583146870136261, 'train/grad_norm_pre_clip': 34.123233795166016, 'timing/data': 0.00030883913859725, 'timing/model': 1.0840379931032658, 'learning_rate/qwen_vl_interface': 8.127937588791887e-06, 'learning_rate/action_model': 8.127937588791886e-05, 'epoch': 0.12}) +[2026-06-24 12:29:05,548][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.28083381056785583, 'train/grad_norm_pre_clip': 40.169212341308594, 'timing/data': 0.00030595320276916027, 'timing/model': 1.0862890209537, 'learning_rate/qwen_vl_interface': 8.12489262493739e-06, 'learning_rate/action_model': 8.12489262493739e-05, 'epoch': 0.12}) +[2026-06-24 12:29:06,905][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.08794182538986206, 'train/grad_norm_pre_clip': 28.824344635009766, 'timing/data': 0.00032935780473053455, 'timing/model': 1.3549204408191144, 'learning_rate/qwen_vl_interface': 8.121845795596282e-06, 'learning_rate/action_model': 8.121845795596282e-05, 'epoch': 0.12}) +[2026-06-24 12:29:07,986][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.1967957615852356, 'train/grad_norm_pre_clip': 73.73225402832031, 'timing/data': 0.000341109000146389, 'timing/model': 1.0795755898579955, 'learning_rate/qwen_vl_interface': 8.118797102745614e-06, 'learning_rate/action_model': 8.118797102745615e-05, 'epoch': 0.12}) +[2026-06-24 12:29:09,067][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.15900819003582, 'train/grad_norm_pre_clip': 41.30913162231445, 'timing/data': 0.0003148599062114954, 'timing/model': 1.0791745278984308, 'learning_rate/qwen_vl_interface': 8.115746548363655e-06, 'learning_rate/action_model': 8.115746548363655e-05, 'epoch': 0.12}) +[2026-06-24 12:29:10,156][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.1540413498878479, 'train/grad_norm_pre_clip': 29.695865631103516, 'timing/data': 0.0003603850491344929, 'timing/model': 1.0871064430102706, 'learning_rate/qwen_vl_interface': 8.112694134429873e-06, 'learning_rate/action_model': 8.112694134429873e-05, 'epoch': 0.12}) +[2026-06-24 12:29:11,243][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.18587253987789154, 'train/grad_norm_pre_clip': 23.806583404541016, 'timing/data': 0.00036992295645177364, 'timing/model': 1.0853895959444344, 'learning_rate/qwen_vl_interface': 8.109639862924947e-06, 'learning_rate/action_model': 8.109639862924945e-05, 'epoch': 0.12}) +[2026-06-24 12:29:12,330][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.08273667097091675, 'train/grad_norm_pre_clip': 13.58137321472168, 'timing/data': 0.00036332104355096817, 'timing/model': 1.0854165670461953, 'learning_rate/qwen_vl_interface': 8.10658373583076e-06, 'learning_rate/action_model': 8.106583735830761e-05, 'epoch': 0.12}) +[2026-06-24 12:29:13,414][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1250 +[2026-06-24 12:29:38,920][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1250 +[2026-06-24 12:29:38,920][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1250 +[2026-06-24 12:30:11,952][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1250 +[2026-06-24 12:34:55,462][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1250 +[2026-06-24 12:34:55,462][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.47082725167274475, 'train/grad_norm_pre_clip': 74.42877197265625, 'eval/action_loss/seconds': 25.50360062881373, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.07854944467544556, 'eval/latency_2/loss': 0.07854944467544556, 'eval/flappy/loss': 0.07854944467544556, 'eval/flappy/latency_2/loss': 0.07854944467544556, 'eval/flap_precision': 0.9354395604395604, 'eval/flap_recall': 0.9240162822252375, 'eval/flap_f1': 0.9296928327645051, 'eval/flap_rate_model': 0.20222222222222222, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9296928327645051, 'eval/latency_2/flap_precision': 0.9354395604395604, 'eval/latency_2/flap_recall': 0.9240162822252375, 'eval/latency_2/flap_f1': 0.9296928327645051, 'eval/latency_2/flap_rate_model': 0.20222222222222222, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9296928327645051, 'eval/action_classification/data_wait_seconds': 0.3765620291233063, 'eval/action_classification/predict_seconds': 32.55843734741211, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 221.74000000000711, 'rl_games_eval/mid_train/mean_length': 1806.8, 'rl_games_eval/mid_train/std_reward': 183.6014226524479, 'rl_games_eval/mid_train/std_length': 1480.1807186962003, 'rl_games_eval/mid_train/macro_mean_reward': 221.74000000000711, 'rl_games_eval/mid_train/macro_mean_length': 1806.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 221.74000000000711, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1806.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 183.6014226524479, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1480.1807186962003, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 221.74000000000711, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00042705005034804344, 'timing/model': 1.0831085410900414, 'learning_rate/qwen_vl_interface': 8.1035257551304e-06, 'learning_rate/action_model': 8.1035257551304e-05, 'epoch': 0.12}) +[2026-06-24 12:34:56,525][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.12324675917625427, 'train/grad_norm_pre_clip': 21.234878540039062, 'timing/data': 0.0003569689579308033, 'timing/model': 1.060879283118993, 'learning_rate/qwen_vl_interface': 8.100465922808158e-06, 'learning_rate/action_model': 8.100465922808159e-05, 'epoch': 0.12}) +[2026-06-24 12:34:57,589][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.2561115622520447, 'train/grad_norm_pre_clip': 26.26859474182129, 'timing/data': 0.00032681203447282314, 'timing/model': 1.0626819198951125, 'learning_rate/qwen_vl_interface': 8.097404240849527e-06, 'learning_rate/action_model': 8.097404240849527e-05, 'epoch': 0.12}) +[2026-06-24 12:34:58,659][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.12937799096107483, 'train/grad_norm_pre_clip': 35.1122932434082, 'timing/data': 0.0003328931052237749, 'timing/model': 1.0685960641130805, 'learning_rate/qwen_vl_interface': 8.094340711241198e-06, 'learning_rate/action_model': 8.094340711241198e-05, 'epoch': 0.12}) +[2026-06-24 12:34:59,724][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.07273619621992111, 'train/grad_norm_pre_clip': 33.80298614501953, 'timing/data': 0.00031971908174455166, 'timing/model': 1.0636010230518878, 'learning_rate/qwen_vl_interface': 8.09127533597106e-06, 'learning_rate/action_model': 8.091275335971061e-05, 'epoch': 0.12}) +[2026-06-24 12:35:00,791][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.192188560962677, 'train/grad_norm_pre_clip': 51.56591033935547, 'timing/data': 0.00033563910983502865, 'timing/model': 1.065391869051382, 'learning_rate/qwen_vl_interface': 8.088208117028205e-06, 'learning_rate/action_model': 8.088208117028206e-05, 'epoch': 0.12}) +[2026-06-24 12:35:02,109][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.09574315696954727, 'train/grad_norm_pre_clip': 18.344369888305664, 'timing/data': 0.00031004101037979126, 'timing/model': 1.3159166281111538, 'learning_rate/qwen_vl_interface': 8.085139056402918e-06, 'learning_rate/action_model': 8.085139056402917e-05, 'epoch': 0.12}) +[2026-06-24 12:35:03,175][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.2512362599372864, 'train/grad_norm_pre_clip': 41.20476531982422, 'timing/data': 0.0003425120376050472, 'timing/model': 1.0650190920569003, 'learning_rate/qwen_vl_interface': 8.082068156086675e-06, 'learning_rate/action_model': 8.082068156086675e-05, 'epoch': 0.12}) +[2026-06-24 12:35:04,244][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.07552217692136765, 'train/grad_norm_pre_clip': 12.724906921386719, 'timing/data': 0.00036288006231188774, 'timing/model': 1.0666202809661627, 'learning_rate/qwen_vl_interface': 8.078995418072156e-06, 'learning_rate/action_model': 8.078995418072157e-05, 'epoch': 0.12}) +[2026-06-24 12:35:05,312][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.247272327542305, 'train/grad_norm_pre_clip': 71.31663513183594, 'timing/data': 0.0003948500379920006, 'timing/model': 1.066252832999453, 'learning_rate/qwen_vl_interface': 8.075920844353223e-06, 'learning_rate/action_model': 8.075920844353223e-05, 'epoch': 0.12}) +[2026-06-24 12:35:06,381][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.0929635614156723, 'train/grad_norm_pre_clip': 33.28562545776367, 'timing/data': 0.0003204310778528452, 'timing/model': 1.0674393770750612, 'learning_rate/qwen_vl_interface': 8.072844436924934e-06, 'learning_rate/action_model': 8.072844436924934e-05, 'epoch': 0.12}) +[2026-06-24 12:35:07,450][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.19606061279773712, 'train/grad_norm_pre_clip': 47.7279167175293, 'timing/data': 0.00031658285297453403, 'timing/model': 1.0672971799504012, 'learning_rate/qwen_vl_interface': 8.069766197783537e-06, 'learning_rate/action_model': 8.069766197783537e-05, 'epoch': 0.12}) +[2026-06-24 12:35:08,517][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.32836267352104187, 'train/grad_norm_pre_clip': 54.548824310302734, 'timing/data': 0.0003400570712983608, 'timing/model': 1.0653525048401207, 'learning_rate/qwen_vl_interface': 8.066686128926468e-06, 'learning_rate/action_model': 8.066686128926468e-05, 'epoch': 0.12}) +[2026-06-24 12:35:09,588][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.31129947304725647, 'train/grad_norm_pre_clip': 53.79631805419922, 'timing/data': 0.00030853901989758015, 'timing/model': 1.0693241490516812, 'learning_rate/qwen_vl_interface': 8.063604232352352e-06, 'learning_rate/action_model': 8.06360423235235e-05, 'epoch': 0.12}) +[2026-06-24 12:35:10,660][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.12148960679769516, 'train/grad_norm_pre_clip': 25.553308486938477, 'timing/data': 0.00030614412389695644, 'timing/model': 1.071139340987429, 'learning_rate/qwen_vl_interface': 8.060520510060996e-06, 'learning_rate/action_model': 8.060520510060995e-05, 'epoch': 0.12}) +[2026-06-24 12:35:11,734][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.2454025000333786, 'train/grad_norm_pre_clip': 58.439884185791016, 'timing/data': 0.000411211047321558, 'timing/model': 1.0722039160318673, 'learning_rate/qwen_vl_interface': 8.057434964053391e-06, 'learning_rate/action_model': 8.057434964053392e-05, 'epoch': 0.12}) +[2026-06-24 12:35:12,807][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.13148468732833862, 'train/grad_norm_pre_clip': 24.642885208129883, 'timing/data': 0.0002997610718011856, 'timing/model': 1.0711550400592387, 'learning_rate/qwen_vl_interface': 8.05434759633172e-06, 'learning_rate/action_model': 8.054347596331721e-05, 'epoch': 0.12}) +[2026-06-24 12:35:13,879][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.2014859914779663, 'train/grad_norm_pre_clip': 34.19530487060547, 'timing/data': 0.0003002830781042576, 'timing/model': 1.0712124980054796, 'learning_rate/qwen_vl_interface': 8.05125840889934e-06, 'learning_rate/action_model': 8.05125840889934e-05, 'epoch': 0.12}) +[2026-06-24 12:35:14,953][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.10446096211671829, 'train/grad_norm_pre_clip': 18.6680850982666, 'timing/data': 0.0003147898241877556, 'timing/model': 1.071943708928302, 'learning_rate/qwen_vl_interface': 8.048167403760792e-06, 'learning_rate/action_model': 8.048167403760792e-05, 'epoch': 0.12}) +[2026-06-24 12:35:16,026][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.2968611717224121, 'train/grad_norm_pre_clip': 55.376285552978516, 'timing/data': 0.00032264506444334984, 'timing/model': 1.0716609789524227, 'learning_rate/qwen_vl_interface': 8.045074582921797e-06, 'learning_rate/action_model': 8.045074582921797e-05, 'epoch': 0.12}) +[2026-06-24 12:35:17,099][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.4307815432548523, 'train/grad_norm_pre_clip': 53.46110534667969, 'timing/data': 0.0006380348931998014, 'timing/model': 1.0703999649267644, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.12}) +[2026-06-24 12:35:18,462][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.16019028425216675, 'train/grad_norm_pre_clip': 37.35344314575195, 'timing/data': 0.0003990079276263714, 'timing/model': 1.3618405140005052, 'learning_rate/qwen_vl_interface': 8.038883502171223e-06, 'learning_rate/action_model': 8.038883502171223e-05, 'epoch': 0.12}) +[2026-06-24 12:35:19,538][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.08446376770734787, 'train/grad_norm_pre_clip': 25.257190704345703, 'timing/data': 0.000350788002833724, 'timing/model': 1.0738644180819392, 'learning_rate/qwen_vl_interface': 8.035785246276973e-06, 'learning_rate/action_model': 8.035785246276973e-05, 'epoch': 0.12}) +[2026-06-24 12:35:20,613][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.29687756299972534, 'train/grad_norm_pre_clip': 43.348121643066406, 'timing/data': 0.00031483988277614117, 'timing/model': 1.0737021339591593, 'learning_rate/qwen_vl_interface': 8.03268518271692e-06, 'learning_rate/action_model': 8.032685182716919e-05, 'epoch': 0.12}) +[2026-06-24 12:35:21,685][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.13221143186092377, 'train/grad_norm_pre_clip': 36.50507736206055, 'timing/data': 0.00030520185828208923, 'timing/model': 1.070924618979916, 'learning_rate/qwen_vl_interface': 8.029583313502664e-06, 'learning_rate/action_model': 8.029583313502664e-05, 'epoch': 0.12}) +[2026-06-24 12:35:22,760][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.11631935834884644, 'train/grad_norm_pre_clip': 18.75802993774414, 'timing/data': 0.0003214818425476551, 'timing/model': 1.072710865875706, 'learning_rate/qwen_vl_interface': 8.026479640646974e-06, 'learning_rate/action_model': 8.026479640646973e-05, 'epoch': 0.12}) +[2026-06-24 12:35:23,835][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.17119085788726807, 'train/grad_norm_pre_clip': 33.43177795410156, 'timing/data': 0.00030708592385053635, 'timing/model': 1.0739263040013611, 'learning_rate/qwen_vl_interface': 8.023374166163787e-06, 'learning_rate/action_model': 8.023374166163787e-05, 'epoch': 0.12}) +[2026-06-24 12:35:24,907][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.11571512371301651, 'train/grad_norm_pre_clip': 31.839282989501953, 'timing/data': 0.00033065001480281353, 'timing/model': 1.0706932959146798, 'learning_rate/qwen_vl_interface': 8.020266892068214e-06, 'learning_rate/action_model': 8.020266892068214e-05, 'epoch': 0.12}) +[2026-06-24 12:35:25,979][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.4433249831199646, 'train/grad_norm_pre_clip': 75.2536849975586, 'timing/data': 0.00030617392621934414, 'timing/model': 1.0695474378298968, 'learning_rate/qwen_vl_interface': 8.017157820376532e-06, 'learning_rate/action_model': 8.017157820376532e-05, 'epoch': 0.12}) +[2026-06-24 12:35:27,049][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.21286869049072266, 'train/grad_norm_pre_clip': 39.125389099121094, 'timing/data': 0.0003064649645239115, 'timing/model': 1.068375141127035, 'learning_rate/qwen_vl_interface': 8.014046953106183e-06, 'learning_rate/action_model': 8.014046953106182e-05, 'epoch': 0.12}) +[2026-06-24 12:35:28,122][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.04458096995949745, 'train/grad_norm_pre_clip': 26.788511276245117, 'timing/data': 0.00030625401996076107, 'timing/model': 1.071855063084513, 'learning_rate/qwen_vl_interface': 8.010934292275775e-06, 'learning_rate/action_model': 8.010934292275775e-05, 'epoch': 0.12}) +[2026-06-24 12:35:29,195][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.15093448758125305, 'train/grad_norm_pre_clip': 51.718849182128906, 'timing/data': 0.0003867850173264742, 'timing/model': 1.0715193441137671, 'learning_rate/qwen_vl_interface': 8.007819839905079e-06, 'learning_rate/action_model': 8.007819839905078e-05, 'epoch': 0.12}) +[2026-06-24 12:35:30,269][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.3319159746170044, 'train/grad_norm_pre_clip': 43.33566665649414, 'timing/data': 0.00033025885932147503, 'timing/model': 1.0721774471458048, 'learning_rate/qwen_vl_interface': 8.004703598015032e-06, 'learning_rate/action_model': 8.004703598015031e-05, 'epoch': 0.12}) +[2026-06-24 12:35:31,341][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.148169606924057, 'train/grad_norm_pre_clip': 37.6297492980957, 'timing/data': 0.0003519090823829174, 'timing/model': 1.0699728750623763, 'learning_rate/qwen_vl_interface': 8.001585568627724e-06, 'learning_rate/action_model': 8.001585568627725e-05, 'epoch': 0.12}) +[2026-06-24 12:35:32,416][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.028848715126514435, 'train/grad_norm_pre_clip': 10.450304985046387, 'timing/data': 0.0003210220020264387, 'timing/model': 1.0732674901373684, 'learning_rate/qwen_vl_interface': 7.998465753766418e-06, 'learning_rate/action_model': 7.998465753766418e-05, 'epoch': 0.12}) +[2026-06-24 12:35:33,490][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.19658613204956055, 'train/grad_norm_pre_clip': 37.355873107910156, 'timing/data': 0.0003114030696451664, 'timing/model': 1.0732629119884223, 'learning_rate/qwen_vl_interface': 7.995344155455523e-06, 'learning_rate/action_model': 7.995344155455522e-05, 'epoch': 0.12}) +[2026-06-24 12:35:34,861][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.0948827713727951, 'train/grad_norm_pre_clip': 29.72260284423828, 'timing/data': 0.0003287261351943016, 'timing/model': 1.3688496500253677, 'learning_rate/qwen_vl_interface': 7.992220775720609e-06, 'learning_rate/action_model': 7.992220775720609e-05, 'epoch': 0.12}) +[2026-06-24 12:35:35,937][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.21133382618427277, 'train/grad_norm_pre_clip': 62.520652770996094, 'timing/data': 0.0004273410886526108, 'timing/model': 1.0745653631165624, 'learning_rate/qwen_vl_interface': 7.989095616588409e-06, 'learning_rate/action_model': 7.989095616588408e-05, 'epoch': 0.12}) +[2026-06-24 12:35:37,013][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.2155298888683319, 'train/grad_norm_pre_clip': 26.513525009155273, 'timing/data': 0.0003064649645239115, 'timing/model': 1.0739948530681431, 'learning_rate/qwen_vl_interface': 7.985968680086799e-06, 'learning_rate/action_model': 7.985968680086798e-05, 'epoch': 0.12}) +[2026-06-24 12:35:38,090][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.052596040070056915, 'train/grad_norm_pre_clip': 21.593334197998047, 'timing/data': 0.0003154110163450241, 'timing/model': 1.0756652129348367, 'learning_rate/qwen_vl_interface': 7.982839968244816e-06, 'learning_rate/action_model': 7.982839968244816e-05, 'epoch': 0.12}) +[2026-06-24 12:35:39,163][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.20306089520454407, 'train/grad_norm_pre_clip': 21.93717384338379, 'timing/data': 0.000301494961604476, 'timing/model': 1.0714668561704457, 'learning_rate/qwen_vl_interface': 7.97970948309265e-06, 'learning_rate/action_model': 7.979709483092649e-05, 'epoch': 0.12}) +[2026-06-24 12:35:40,236][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.3171728551387787, 'train/grad_norm_pre_clip': 52.19078063964844, 'timing/data': 0.00030245701782405376, 'timing/model': 1.0707004198338836, 'learning_rate/qwen_vl_interface': 7.976577226661635e-06, 'learning_rate/action_model': 7.976577226661634e-05, 'epoch': 0.12}) +[2026-06-24 12:35:41,313][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.29155233502388, 'train/grad_norm_pre_clip': 33.423221588134766, 'timing/data': 0.00030597299337387085, 'timing/model': 1.0756276228930801, 'learning_rate/qwen_vl_interface': 7.973443200984259e-06, 'learning_rate/action_model': 7.973443200984259e-05, 'epoch': 0.13}) +[2026-06-24 12:35:42,389][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.09132988750934601, 'train/grad_norm_pre_clip': 41.14405059814453, 'timing/data': 0.0003324029967188835, 'timing/model': 1.075061511946842, 'learning_rate/qwen_vl_interface': 7.970307408094158e-06, 'learning_rate/action_model': 7.970307408094157e-05, 'epoch': 0.13}) +[2026-06-24 12:35:43,464][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.23200100660324097, 'train/grad_norm_pre_clip': 24.28952980041504, 'timing/data': 0.0003060039598494768, 'timing/model': 1.073467114008963, 'learning_rate/qwen_vl_interface': 7.967169850026113e-06, 'learning_rate/action_model': 7.967169850026113e-05, 'epoch': 0.13}) +[2026-06-24 12:35:44,543][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.08162935078144073, 'train/grad_norm_pre_clip': 16.49337387084961, 'timing/data': 0.00035085692070424557, 'timing/model': 1.0765402631368488, 'learning_rate/qwen_vl_interface': 7.964030528816055e-06, 'learning_rate/action_model': 7.964030528816055e-05, 'epoch': 0.13}) +[2026-06-24 12:35:45,617][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.13890044391155243, 'train/grad_norm_pre_clip': 20.04456329345703, 'timing/data': 0.0005047160666435957, 'timing/model': 1.0722990659996867, 'learning_rate/qwen_vl_interface': 7.960889446501052e-06, 'learning_rate/action_model': 7.960889446501052e-05, 'epoch': 0.13}) +[2026-06-24 12:35:46,693][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.08648238331079483, 'train/grad_norm_pre_clip': 25.348243713378906, 'timing/data': 0.00032222410663962364, 'timing/model': 1.0744631909765303, 'learning_rate/qwen_vl_interface': 7.957746605119317e-06, 'learning_rate/action_model': 7.957746605119318e-05, 'epoch': 0.13}) +[2026-06-24 12:35:47,769][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.1534709930419922, 'train/grad_norm_pre_clip': 33.20729446411133, 'timing/data': 0.00034409412182867527, 'timing/model': 1.0746518541127443, 'learning_rate/qwen_vl_interface': 7.954602006710209e-06, 'learning_rate/action_model': 7.954602006710209e-05, 'epoch': 0.13}) +[2026-06-24 12:35:48,844][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.05409181863069534, 'train/grad_norm_pre_clip': 14.198229789733887, 'timing/data': 0.00031321682035923004, 'timing/model': 1.0728680018801242, 'learning_rate/qwen_vl_interface': 7.951455653314222e-06, 'learning_rate/action_model': 7.951455653314221e-05, 'epoch': 0.13}) +[2026-06-24 12:35:50,127][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.059426162391901016, 'train/grad_norm_pre_clip': 9.375661849975586, 'timing/data': 0.0004927539266645908, 'timing/model': 1.2812156160362065, 'learning_rate/qwen_vl_interface': 7.94830754697299e-06, 'learning_rate/action_model': 7.94830754697299e-05, 'epoch': 0.13}) +[2026-06-24 12:35:51,201][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.05945154279470444, 'train/grad_norm_pre_clip': 10.686993598937988, 'timing/data': 0.00035742903128266335, 'timing/model': 1.0721562979742885, 'learning_rate/qwen_vl_interface': 7.945157689729287e-06, 'learning_rate/action_model': 7.945157689729287e-05, 'epoch': 0.13}) +[2026-06-24 12:35:52,277][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.1486465334892273, 'train/grad_norm_pre_clip': 39.94731903076172, 'timing/data': 0.00030983006581664085, 'timing/model': 1.0746499409433454, 'learning_rate/qwen_vl_interface': 7.94200608362702e-06, 'learning_rate/action_model': 7.94200608362702e-05, 'epoch': 0.13}) +[2026-06-24 12:35:53,355][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.07975687831640244, 'train/grad_norm_pre_clip': 24.249528884887695, 'timing/data': 0.00029930099844932556, 'timing/model': 1.0764027261175215, 'learning_rate/qwen_vl_interface': 7.938852730711232e-06, 'learning_rate/action_model': 7.938852730711231e-05, 'epoch': 0.13}) +[2026-06-24 12:35:54,431][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.2058834582567215, 'train/grad_norm_pre_clip': 41.491573333740234, 'timing/data': 0.0003231959417462349, 'timing/model': 1.074209596030414, 'learning_rate/qwen_vl_interface': 7.935697633028099e-06, 'learning_rate/action_model': 7.935697633028099e-05, 'epoch': 0.13}) +[2026-06-24 12:35:55,506][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.07317934930324554, 'train/grad_norm_pre_clip': 17.420351028442383, 'timing/data': 0.0003447560593485832, 'timing/model': 1.0737102499697357, 'learning_rate/qwen_vl_interface': 7.932540792624931e-06, 'learning_rate/action_model': 7.93254079262493e-05, 'epoch': 0.13}) +[2026-06-24 12:35:56,585][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.052096713334321976, 'train/grad_norm_pre_clip': 19.070154190063477, 'timing/data': 0.00032257381826639175, 'timing/model': 1.0772371201310307, 'learning_rate/qwen_vl_interface': 7.929382211550166e-06, 'learning_rate/action_model': 7.929382211550165e-05, 'epoch': 0.13}) +[2026-06-24 12:35:57,662][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.10290604829788208, 'train/grad_norm_pre_clip': 49.48223114013672, 'timing/data': 0.00031906808726489544, 'timing/model': 1.075451604090631, 'learning_rate/qwen_vl_interface': 7.926221891853373e-06, 'learning_rate/action_model': 7.926221891853372e-05, 'epoch': 0.13}) +[2026-06-24 12:35:58,739][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.10006208717823029, 'train/grad_norm_pre_clip': 21.147045135498047, 'timing/data': 0.0003084479831159115, 'timing/model': 1.0757787970360368, 'learning_rate/qwen_vl_interface': 7.923059835585251e-06, 'learning_rate/action_model': 7.923059835585251e-05, 'epoch': 0.13}) +[2026-06-24 12:35:59,816][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.22071011364459991, 'train/grad_norm_pre_clip': 107.58711242675781, 'timing/data': 0.0003136878367513418, 'timing/model': 1.0750113290268928, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.13}) +[2026-06-24 12:36:00,890][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.3748174011707306, 'train/grad_norm_pre_clip': 87.14232635498047, 'timing/data': 0.000378879951313138, 'timing/model': 1.0721637818496674, 'learning_rate/qwen_vl_interface': 7.916730521543436e-06, 'learning_rate/action_model': 7.916730521543436e-05, 'epoch': 0.13}) +[2026-06-24 12:36:01,969][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.022372204810380936, 'train/grad_norm_pre_clip': 19.002334594726562, 'timing/data': 0.0003853621892631054, 'timing/model': 1.07771061710082, 'learning_rate/qwen_vl_interface': 7.913563267876765e-06, 'learning_rate/action_model': 7.913563267876766e-05, 'epoch': 0.13}) +[2026-06-24 12:36:03,040][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.13130274415016174, 'train/grad_norm_pre_clip': 52.81216812133789, 'timing/data': 0.00035240105353295803, 'timing/model': 1.0692993339616805, 'learning_rate/qwen_vl_interface': 7.910394285852811e-06, 'learning_rate/action_model': 7.910394285852811e-05, 'epoch': 0.13}) +[2026-06-24 12:36:04,121][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.013119629584252834, 'train/grad_norm_pre_clip': 11.818771362304688, 'timing/data': 0.0003113329876214266, 'timing/model': 1.0787985969800502, 'learning_rate/qwen_vl_interface': 7.907223577527886e-06, 'learning_rate/action_model': 7.907223577527886e-05, 'epoch': 0.13}) +[2026-06-24 12:36:05,202][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.0061664879322052, 'train/grad_norm_pre_clip': 2.5402557849884033, 'timing/data': 0.0003208210691809654, 'timing/model': 1.079950193874538, 'learning_rate/qwen_vl_interface': 7.904051144959432e-06, 'learning_rate/action_model': 7.904051144959431e-05, 'epoch': 0.13}) +[2026-06-24 12:36:06,571][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.006426672451198101, 'train/grad_norm_pre_clip': 3.1737866401672363, 'timing/data': 0.0003372221253812313, 'timing/model': 1.3672465470153838, 'learning_rate/qwen_vl_interface': 7.900876990206002e-06, 'learning_rate/action_model': 7.900876990206002e-05, 'epoch': 0.13}) +[2026-06-24 12:36:07,651][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.12229893356561661, 'train/grad_norm_pre_clip': 99.84013366699219, 'timing/data': 0.00035497499629855156, 'timing/model': 1.078909964999184, 'learning_rate/qwen_vl_interface': 7.897701115327276e-06, 'learning_rate/action_model': 7.897701115327276e-05, 'epoch': 0.13}) +[2026-06-24 12:36:08,726][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.5557427406311035, 'train/grad_norm_pre_clip': 71.05966186523438, 'timing/data': 0.0003061241004616022, 'timing/model': 1.0728693748824298, 'learning_rate/qwen_vl_interface': 7.894523522384043e-06, 'learning_rate/action_model': 7.894523522384042e-05, 'epoch': 0.13}) +[2026-06-24 12:36:09,805][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.010376064106822014, 'train/grad_norm_pre_clip': 11.969364166259766, 'timing/data': 0.00038536195643246174, 'timing/model': 1.076974007068202, 'learning_rate/qwen_vl_interface': 7.891344213438205e-06, 'learning_rate/action_model': 7.891344213438204e-05, 'epoch': 0.13}) +[2026-06-24 12:36:10,880][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.09518174082040787, 'train/grad_norm_pre_clip': 31.763395309448242, 'timing/data': 0.0002997110132128, 'timing/model': 1.0743506611324847, 'learning_rate/qwen_vl_interface': 7.888163190552783e-06, 'learning_rate/action_model': 7.888163190552783e-05, 'epoch': 0.13}) +[2026-06-24 12:36:11,956][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.42187294363975525, 'train/grad_norm_pre_clip': 75.4381332397461, 'timing/data': 0.0004256169777363539, 'timing/model': 1.0737584510352463, 'learning_rate/qwen_vl_interface': 7.88498045579191e-06, 'learning_rate/action_model': 7.88498045579191e-05, 'epoch': 0.13}) +[2026-06-24 12:36:13,034][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.7250751256942749, 'train/grad_norm_pre_clip': 167.74012756347656, 'timing/data': 0.00034624896943569183, 'timing/model': 1.0756922548171133, 'learning_rate/qwen_vl_interface': 7.881796011220827e-06, 'learning_rate/action_model': 7.881796011220827e-05, 'epoch': 0.13}) +[2026-06-24 12:36:14,111][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.23649285733699799, 'train/grad_norm_pre_clip': 42.71967315673828, 'timing/data': 0.0003127360250800848, 'timing/model': 1.075464899186045, 'learning_rate/qwen_vl_interface': 7.878609858905885e-06, 'learning_rate/action_model': 7.878609858905886e-05, 'epoch': 0.13}) +[2026-06-24 12:36:15,186][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.04155531898140907, 'train/grad_norm_pre_clip': 40.77690887451172, 'timing/data': 0.0003037499263882637, 'timing/model': 1.073891039006412, 'learning_rate/qwen_vl_interface': 7.875422000914544e-06, 'learning_rate/action_model': 7.875422000914544e-05, 'epoch': 0.13}) +[2026-06-24 12:36:16,262][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.19561539590358734, 'train/grad_norm_pre_clip': 42.24790954589844, 'timing/data': 0.0003216429613530636, 'timing/model': 1.0747237901668996, 'learning_rate/qwen_vl_interface': 7.872232439315371e-06, 'learning_rate/action_model': 7.87223243931537e-05, 'epoch': 0.13}) +[2026-06-24 12:36:17,340][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.28541553020477295, 'train/grad_norm_pre_clip': 46.237937927246094, 'timing/data': 0.00032180314883589745, 'timing/model': 1.0764510671142489, 'learning_rate/qwen_vl_interface': 7.869041176178038e-06, 'learning_rate/action_model': 7.869041176178037e-05, 'epoch': 0.13}) +[2026-06-24 12:36:18,418][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.22112585604190826, 'train/grad_norm_pre_clip': 51.47185134887695, 'timing/data': 0.00034523685462772846, 'timing/model': 1.0759960538707674, 'learning_rate/qwen_vl_interface': 7.865848213573316e-06, 'learning_rate/action_model': 7.865848213573317e-05, 'epoch': 0.13}) +[2026-06-24 12:36:19,494][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.21755729615688324, 'train/grad_norm_pre_clip': 67.3680191040039, 'timing/data': 0.0003211209550499916, 'timing/model': 1.0749948881566525, 'learning_rate/qwen_vl_interface': 7.862653553573092e-06, 'learning_rate/action_model': 7.862653553573092e-05, 'epoch': 0.13}) +[2026-06-24 12:36:20,576][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.2048921138048172, 'train/grad_norm_pre_clip': 58.776832580566406, 'timing/data': 0.0003133881837129593, 'timing/model': 1.0795954000204802, 'learning_rate/qwen_vl_interface': 7.859457198250342e-06, 'learning_rate/action_model': 7.859457198250342e-05, 'epoch': 0.13}) +[2026-06-24 12:36:21,864][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.2330721765756607, 'train/grad_norm_pre_clip': 55.44633483886719, 'timing/data': 0.0004036668688058853, 'timing/model': 1.2869986849837005, 'learning_rate/qwen_vl_interface': 7.856259149679146e-06, 'learning_rate/action_model': 7.856259149679146e-05, 'epoch': 0.13}) +[2026-06-24 12:36:22,939][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.09606139361858368, 'train/grad_norm_pre_clip': 19.396560668945312, 'timing/data': 0.0002989398781210184, 'timing/model': 1.072563922032714, 'learning_rate/qwen_vl_interface': 7.853059409934685e-06, 'learning_rate/action_model': 7.853059409934685e-05, 'epoch': 0.13}) +[2026-06-24 12:36:24,017][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.07605776935815811, 'train/grad_norm_pre_clip': 42.35253143310547, 'timing/data': 0.00034647900611162186, 'timing/model': 1.0766763701103628, 'learning_rate/qwen_vl_interface': 7.849857981093231e-06, 'learning_rate/action_model': 7.84985798109323e-05, 'epoch': 0.13}) +[2026-06-24 12:36:25,092][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.14864127337932587, 'train/grad_norm_pre_clip': 46.0535774230957, 'timing/data': 0.00029603508301079273, 'timing/model': 1.0735264259856194, 'learning_rate/qwen_vl_interface': 7.84665486523216e-06, 'learning_rate/action_model': 7.846654865232159e-05, 'epoch': 0.13}) +[2026-06-24 12:36:26,168][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.21271160244941711, 'train/grad_norm_pre_clip': 44.286346435546875, 'timing/data': 0.00029892008751630783, 'timing/model': 1.0742009109817445, 'learning_rate/qwen_vl_interface': 7.84345006442994e-06, 'learning_rate/action_model': 7.843450064429939e-05, 'epoch': 0.13}) +[2026-06-24 12:36:27,247][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.1935514509677887, 'train/grad_norm_pre_clip': 38.97089385986328, 'timing/data': 0.00030104396864771843, 'timing/model': 1.077252799877897, 'learning_rate/qwen_vl_interface': 7.840243580766129e-06, 'learning_rate/action_model': 7.840243580766128e-05, 'epoch': 0.13}) +[2026-06-24 12:36:28,323][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.02902965620160103, 'train/grad_norm_pre_clip': 18.33753776550293, 'timing/data': 0.00032286508940160275, 'timing/model': 1.0747021788265556, 'learning_rate/qwen_vl_interface': 7.83703541632138e-06, 'learning_rate/action_model': 7.83703541632138e-05, 'epoch': 0.13}) +[2026-06-24 12:36:29,400][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.2662380337715149, 'train/grad_norm_pre_clip': 66.02526092529297, 'timing/data': 0.00033115106634795666, 'timing/model': 1.0752291970420629, 'learning_rate/qwen_vl_interface': 7.833825573177435e-06, 'learning_rate/action_model': 7.833825573177436e-05, 'epoch': 0.13}) +[2026-06-24 12:36:30,476][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.18517331779003143, 'train/grad_norm_pre_clip': 36.73341369628906, 'timing/data': 0.00034280307590961456, 'timing/model': 1.0744390171021223, 'learning_rate/qwen_vl_interface': 7.83061405341713e-06, 'learning_rate/action_model': 7.83061405341713e-05, 'epoch': 0.13}) +[2026-06-24 12:36:31,549][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.05403101444244385, 'train/grad_norm_pre_clip': 10.948473930358887, 'timing/data': 0.00031548202969133854, 'timing/model': 1.0716922690626234, 'learning_rate/qwen_vl_interface': 7.827400859124384e-06, 'learning_rate/action_model': 7.827400859124384e-05, 'epoch': 0.13}) +[2026-06-24 12:36:32,626][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.10110313445329666, 'train/grad_norm_pre_clip': 34.74604797363281, 'timing/data': 0.000311804935336113, 'timing/model': 1.0755684529431164, 'learning_rate/qwen_vl_interface': 7.8241859923842e-06, 'learning_rate/action_model': 7.8241859923842e-05, 'epoch': 0.13}) +[2026-06-24 12:36:33,702][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.14115095138549805, 'train/grad_norm_pre_clip': 26.23086166381836, 'timing/data': 0.00030969991348683834, 'timing/model': 1.0737417701166123, 'learning_rate/qwen_vl_interface': 7.820969455282676e-06, 'learning_rate/action_model': 7.820969455282675e-05, 'epoch': 0.13}) +[2026-06-24 12:36:34,775][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.177550807595253, 'train/grad_norm_pre_clip': 26.930828094482422, 'timing/data': 0.0003248888533562422, 'timing/model': 1.072086488129571, 'learning_rate/qwen_vl_interface': 7.817751249906982e-06, 'learning_rate/action_model': 7.817751249906983e-05, 'epoch': 0.13}) +[2026-06-24 12:36:35,850][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.12853221595287323, 'train/grad_norm_pre_clip': 24.181734085083008, 'timing/data': 0.0003272830508649349, 'timing/model': 1.0728610700462013, 'learning_rate/qwen_vl_interface': 7.814531378345381e-06, 'learning_rate/action_model': 7.814531378345381e-05, 'epoch': 0.13}) +[2026-06-24 12:36:36,923][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.20765189826488495, 'train/grad_norm_pre_clip': 38.007572174072266, 'timing/data': 0.00032219383865594864, 'timing/model': 1.0714808229822665, 'learning_rate/qwen_vl_interface': 7.811309842687213e-06, 'learning_rate/action_model': 7.811309842687212e-05, 'epoch': 0.13}) +[2026-06-24 12:36:37,997][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.18945476412773132, 'train/grad_norm_pre_clip': 29.963674545288086, 'timing/data': 0.0005220388993620872, 'timing/model': 1.0724707781337202, 'learning_rate/qwen_vl_interface': 7.808086645022895e-06, 'learning_rate/action_model': 7.808086645022895e-05, 'epoch': 0.13}) +[2026-06-24 12:36:39,075][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.10018584877252579, 'train/grad_norm_pre_clip': 20.641239166259766, 'timing/data': 0.00029986188746988773, 'timing/model': 1.0761904988903552, 'learning_rate/qwen_vl_interface': 7.804861787443924e-06, 'learning_rate/action_model': 7.804861787443923e-05, 'epoch': 0.13}) +[2026-06-24 12:36:40,148][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.1331414431333542, 'train/grad_norm_pre_clip': 40.27263641357422, 'timing/data': 0.00030784704722464085, 'timing/model': 1.070923478109762, 'learning_rate/qwen_vl_interface': 7.801635272042879e-06, 'learning_rate/action_model': 7.801635272042877e-05, 'epoch': 0.13}) +[2026-06-24 12:36:41,509][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.07188328355550766, 'train/grad_norm_pre_clip': 19.648086547851562, 'timing/data': 0.0003414400853216648, 'timing/model': 1.3597822079900652, 'learning_rate/qwen_vl_interface': 7.798407100913407e-06, 'learning_rate/action_model': 7.798407100913407e-05, 'epoch': 0.13}) +[2026-06-24 12:36:42,584][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.22026212513446808, 'train/grad_norm_pre_clip': 44.90692138671875, 'timing/data': 0.0004548721481114626, 'timing/model': 1.0731429180596024, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.13}) +[2026-06-24 12:36:43,657][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.1781393438577652, 'train/grad_norm_pre_clip': 95.7392349243164, 'timing/data': 0.00031115300953388214, 'timing/model': 1.072120571974665, 'learning_rate/qwen_vl_interface': 7.79194579984916e-06, 'learning_rate/action_model': 7.791945799849161e-05, 'epoch': 0.13}) +[2026-06-24 12:36:44,730][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.0534876249730587, 'train/grad_norm_pre_clip': 19.754112243652344, 'timing/data': 0.00036484398879110813, 'timing/model': 1.0713509100023657, 'learning_rate/qwen_vl_interface': 7.788712674107054e-06, 'learning_rate/action_model': 7.788712674107055e-05, 'epoch': 0.13}) +[2026-06-24 12:36:45,806][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.23325112462043762, 'train/grad_norm_pre_clip': 40.80961227416992, 'timing/data': 0.0003596940077841282, 'timing/model': 1.0737966429442167, 'learning_rate/qwen_vl_interface': 7.785477901021859e-06, 'learning_rate/action_model': 7.785477901021859e-05, 'epoch': 0.13}) +[2026-06-24 12:36:46,878][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.03822011500597, 'train/grad_norm_pre_clip': 8.060725212097168, 'timing/data': 0.0005136118270456791, 'timing/model': 1.0706070859450847, 'learning_rate/qwen_vl_interface': 7.78224148269258e-06, 'learning_rate/action_model': 7.78224148269258e-05, 'epoch': 0.13}) +[2026-06-24 12:36:47,955][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.24158689379692078, 'train/grad_norm_pre_clip': 58.16289138793945, 'timing/data': 0.0005258449818938971, 'timing/model': 1.0750537691637874, 'learning_rate/qwen_vl_interface': 7.779003421219293e-06, 'learning_rate/action_model': 7.779003421219293e-05, 'epoch': 0.13}) +[2026-06-24 12:36:49,031][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.10606562346220016, 'train/grad_norm_pre_clip': 51.04892349243164, 'timing/data': 0.00032793497666716576, 'timing/model': 1.0738843181170523, 'learning_rate/qwen_vl_interface': 7.775763718703144e-06, 'learning_rate/action_model': 7.775763718703144e-05, 'epoch': 0.13}) +[2026-06-24 12:36:50,106][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.14126189053058624, 'train/grad_norm_pre_clip': 87.90937805175781, 'timing/data': 0.00031386781483888626, 'timing/model': 1.073303708806634, 'learning_rate/qwen_vl_interface': 7.772522377246339e-06, 'learning_rate/action_model': 7.77252237724634e-05, 'epoch': 0.13}) +[2026-06-24 12:36:51,185][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.020494339987635612, 'train/grad_norm_pre_clip': 10.224861145019531, 'timing/data': 0.0003435029648244381, 'timing/model': 1.0780121230054647, 'learning_rate/qwen_vl_interface': 7.769279398952152e-06, 'learning_rate/action_model': 7.769279398952151e-05, 'epoch': 0.13}) +[2026-06-24 12:36:52,262][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.14174646139144897, 'train/grad_norm_pre_clip': 45.6036491394043, 'timing/data': 0.00036309100687503815, 'timing/model': 1.075056134024635, 'learning_rate/qwen_vl_interface': 7.766034785924914e-06, 'learning_rate/action_model': 7.766034785924912e-05, 'epoch': 0.13}) +[2026-06-24 12:36:53,334][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.24485136568546295, 'train/grad_norm_pre_clip': 54.06951904296875, 'timing/data': 0.0003385350573807955, 'timing/model': 1.0701136509887874, 'learning_rate/qwen_vl_interface': 7.762788540270018e-06, 'learning_rate/action_model': 7.762788540270018e-05, 'epoch': 0.13}) +[2026-06-24 12:36:54,417][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.027969343587756157, 'train/grad_norm_pre_clip': 18.315351486206055, 'timing/data': 0.0003538229502737522, 'timing/model': 1.073414778104052, 'learning_rate/qwen_vl_interface': 7.759540664093919e-06, 'learning_rate/action_model': 7.759540664093919e-05, 'epoch': 0.13}) +[2026-06-24 12:36:55,490][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.17742857336997986, 'train/grad_norm_pre_clip': 57.681949615478516, 'timing/data': 0.00040327501483261585, 'timing/model': 1.0704969489015639, 'learning_rate/qwen_vl_interface': 7.756291159504132e-06, 'learning_rate/action_model': 7.756291159504132e-05, 'epoch': 0.13}) +[2026-06-24 12:36:56,566][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.1671312302350998, 'train/grad_norm_pre_clip': 44.18415832519531, 'timing/data': 0.0003334151115268469, 'timing/model': 1.0743590369820595, 'learning_rate/qwen_vl_interface': 7.75304002860922e-06, 'learning_rate/action_model': 7.75304002860922e-05, 'epoch': 0.13}) +[2026-06-24 12:36:57,640][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.15650171041488647, 'train/grad_norm_pre_clip': 38.22422790527344, 'timing/data': 0.00030531291849911213, 'timing/model': 1.0726908009964973, 'learning_rate/qwen_vl_interface': 7.749787273518807e-06, 'learning_rate/action_model': 7.749787273518807e-05, 'epoch': 0.13}) +[2026-06-24 12:36:58,717][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.08676924556493759, 'train/grad_norm_pre_clip': 44.93073654174805, 'timing/data': 0.0003149299882352352, 'timing/model': 1.0754424980841577, 'learning_rate/qwen_vl_interface': 7.746532896343575e-06, 'learning_rate/action_model': 7.746532896343575e-05, 'epoch': 0.13}) +[2026-06-24 12:37:00,035][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.14844995737075806, 'train/grad_norm_pre_clip': 26.82495880126953, 'timing/data': 0.0003184869419783354, 'timing/model': 1.3166698119603097, 'learning_rate/qwen_vl_interface': 7.743276899195249e-06, 'learning_rate/action_model': 7.743276899195249e-05, 'epoch': 0.13}) +[2026-06-24 12:37:01,109][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.05903943255543709, 'train/grad_norm_pre_clip': 30.445011138916016, 'timing/data': 0.0003182061482220888, 'timing/model': 1.0719595099799335, 'learning_rate/qwen_vl_interface': 7.740019284186612e-06, 'learning_rate/action_model': 7.740019284186612e-05, 'epoch': 0.13}) +[2026-06-24 12:37:02,179][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.0654730573296547, 'train/grad_norm_pre_clip': 24.6412410736084, 'timing/data': 0.00032092095352709293, 'timing/model': 1.0686314629856497, 'learning_rate/qwen_vl_interface': 7.736760053431496e-06, 'learning_rate/action_model': 7.736760053431495e-05, 'epoch': 0.13}) +[2026-06-24 12:37:03,250][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.14523014426231384, 'train/grad_norm_pre_clip': 40.5990104675293, 'timing/data': 0.0004120129160583019, 'timing/model': 1.0699089479167014, 'learning_rate/qwen_vl_interface': 7.73349920904478e-06, 'learning_rate/action_model': 7.73349920904478e-05, 'epoch': 0.13}) +[2026-06-24 12:37:04,322][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.06655501574277878, 'train/grad_norm_pre_clip': 13.90096378326416, 'timing/data': 0.0005131519865244627, 'timing/model': 1.0699994668830186, 'learning_rate/qwen_vl_interface': 7.73023675314239e-06, 'learning_rate/action_model': 7.73023675314239e-05, 'epoch': 0.13}) +[2026-06-24 12:37:05,400][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.0919920951128006, 'train/grad_norm_pre_clip': 42.13817596435547, 'timing/data': 0.00031486013904213905, 'timing/model': 1.0754045569337904, 'learning_rate/qwen_vl_interface': 7.726972687841298e-06, 'learning_rate/action_model': 7.726972687841299e-05, 'epoch': 0.13}) +[2026-06-24 12:37:06,475][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.20230916142463684, 'train/grad_norm_pre_clip': 66.57156372070312, 'timing/data': 0.0003890390507876873, 'timing/model': 1.073600887088105, 'learning_rate/qwen_vl_interface': 7.723707015259522e-06, 'learning_rate/action_model': 7.723707015259522e-05, 'epoch': 0.13}) +[2026-06-24 12:37:07,551][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.1575072705745697, 'train/grad_norm_pre_clip': 23.85011100769043, 'timing/data': 0.0003025869373232126, 'timing/model': 1.074045370100066, 'learning_rate/qwen_vl_interface': 7.72043973751612e-06, 'learning_rate/action_model': 7.720439737516119e-05, 'epoch': 0.13}) +[2026-06-24 12:37:08,628][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.14844787120819092, 'train/grad_norm_pre_clip': 23.381776809692383, 'timing/data': 0.0002959838602691889, 'timing/model': 1.075466020964086, 'learning_rate/qwen_vl_interface': 7.717170856731192e-06, 'learning_rate/action_model': 7.717170856731193e-05, 'epoch': 0.13}) +[2026-06-24 12:37:09,701][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.013035329058766365, 'train/grad_norm_pre_clip': 9.487762451171875, 'timing/data': 0.0003434529062360525, 'timing/model': 1.0721032698638737, 'learning_rate/qwen_vl_interface': 7.713900375025883e-06, 'learning_rate/action_model': 7.713900375025882e-05, 'epoch': 0.13}) +[2026-06-24 12:37:10,774][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.10099201649427414, 'train/grad_norm_pre_clip': 29.92005729675293, 'timing/data': 0.0003017159178853035, 'timing/model': 1.071175622055307, 'learning_rate/qwen_vl_interface': 7.71062829452237e-06, 'learning_rate/action_model': 7.710628294522369e-05, 'epoch': 0.13}) +[2026-06-24 12:37:11,850][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.2030387818813324, 'train/grad_norm_pre_clip': 60.17255401611328, 'timing/data': 0.000302337110042572, 'timing/model': 1.0744665688835084, 'learning_rate/qwen_vl_interface': 7.707354617343871e-06, 'learning_rate/action_model': 7.707354617343871e-05, 'epoch': 0.13}) +[2026-06-24 12:37:12,923][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.05585211515426636, 'train/grad_norm_pre_clip': 19.76261329650879, 'timing/data': 0.0003024269826710224, 'timing/model': 1.071145093999803, 'learning_rate/qwen_vl_interface': 7.704079345614639e-06, 'learning_rate/action_model': 7.704079345614639e-05, 'epoch': 0.13}) +[2026-06-24 12:37:13,999][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.20167352259159088, 'train/grad_norm_pre_clip': 53.645240783691406, 'timing/data': 0.00031741498969495296, 'timing/model': 1.075108311837539, 'learning_rate/qwen_vl_interface': 7.700802481459964e-06, 'learning_rate/action_model': 7.700802481459964e-05, 'epoch': 0.13}) +[2026-06-24 12:37:15,073][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.1786249727010727, 'train/grad_norm_pre_clip': 69.51528930664062, 'timing/data': 0.0003194189630448818, 'timing/model': 1.0719206479843706, 'learning_rate/qwen_vl_interface': 7.697524027006168e-06, 'learning_rate/action_model': 7.697524027006167e-05, 'epoch': 0.13}) +[2026-06-24 12:37:16,149][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.06156652793288231, 'train/grad_norm_pre_clip': 22.254497528076172, 'timing/data': 0.0003139488399028778, 'timing/model': 1.074468582868576, 'learning_rate/qwen_vl_interface': 7.694243984380603e-06, 'learning_rate/action_model': 7.694243984380603e-05, 'epoch': 0.13}) +[2026-06-24 12:37:17,224][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.13108795881271362, 'train/grad_norm_pre_clip': 48.01521301269531, 'timing/data': 0.00030697486363351345, 'timing/model': 1.073087465018034, 'learning_rate/qwen_vl_interface': 7.690962355711651e-06, 'learning_rate/action_model': 7.690962355711652e-05, 'epoch': 0.13}) +[2026-06-24 12:37:18,506][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.12466739118099213, 'train/grad_norm_pre_clip': 60.5217170715332, 'timing/data': 0.0003158319741487503, 'timing/model': 1.2803562271874398, 'learning_rate/qwen_vl_interface': 7.68767914312873e-06, 'learning_rate/action_model': 7.68767914312873e-05, 'epoch': 0.13}) +[2026-06-24 12:37:19,576][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.11454766988754272, 'train/grad_norm_pre_clip': 38.94385528564453, 'timing/data': 0.0003193079028278589, 'timing/model': 1.0685095458757132, 'learning_rate/qwen_vl_interface': 7.684394348762278e-06, 'learning_rate/action_model': 7.684394348762277e-05, 'epoch': 0.13}) +[2026-06-24 12:37:20,651][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.06816038489341736, 'train/grad_norm_pre_clip': 31.181398391723633, 'timing/data': 0.0004212697967886925, 'timing/model': 1.0733301190193743, 'learning_rate/qwen_vl_interface': 7.681107974743763e-06, 'learning_rate/action_model': 7.681107974743764e-05, 'epoch': 0.13}) +[2026-06-24 12:37:21,727][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.038172654807567596, 'train/grad_norm_pre_clip': 16.29020881652832, 'timing/data': 0.00031046196818351746, 'timing/model': 1.0748440260067582, 'learning_rate/qwen_vl_interface': 7.677820023205676e-06, 'learning_rate/action_model': 7.677820023205677e-05, 'epoch': 0.13}) +[2026-06-24 12:37:22,803][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.15688571333885193, 'train/grad_norm_pre_clip': 24.68310546875, 'timing/data': 0.0003608260303735733, 'timing/model': 1.0736891420092434, 'learning_rate/qwen_vl_interface': 7.674530496281536e-06, 'learning_rate/action_model': 7.674530496281536e-05, 'epoch': 0.13}) +[2026-06-24 12:37:23,877][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.2803100347518921, 'train/grad_norm_pre_clip': 45.5656623840332, 'timing/data': 0.00033593899570405483, 'timing/model': 1.0727184528950602, 'learning_rate/qwen_vl_interface': 7.67123939610588e-06, 'learning_rate/action_model': 7.67123939610588e-05, 'epoch': 0.13}) +[2026-06-24 12:37:24,952][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.18865366280078888, 'train/grad_norm_pre_clip': 36.95003128051758, 'timing/data': 0.0003098798915743828, 'timing/model': 1.0729379339609295, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.13}) +[2026-06-24 12:37:26,033][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.07903841137886047, 'train/grad_norm_pre_clip': 30.08987808227539, 'timing/data': 0.00043879193253815174, 'timing/model': 1.0796140460297465, 'learning_rate/qwen_vl_interface': 7.664652484543271e-06, 'learning_rate/action_model': 7.664652484543271e-05, 'epoch': 0.13}) +[2026-06-24 12:37:27,109][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.2498602271080017, 'train/grad_norm_pre_clip': 120.47337341308594, 'timing/data': 0.0003155311569571495, 'timing/model': 1.0746541609987617, 'learning_rate/qwen_vl_interface': 7.661356677430499e-06, 'learning_rate/action_model': 7.661356677430499e-05, 'epoch': 0.13}) +[2026-06-24 12:37:28,179][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.30908945202827454, 'train/grad_norm_pre_clip': 95.17145538330078, 'timing/data': 0.0003000821452587843, 'timing/model': 1.0677973409183323, 'learning_rate/qwen_vl_interface': 7.658059305614556e-06, 'learning_rate/action_model': 7.658059305614555e-05, 'epoch': 0.13}) +[2026-06-24 12:37:29,251][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.5308135747909546, 'train/grad_norm_pre_clip': 59.979164123535156, 'timing/data': 0.00030875904485583305, 'timing/model': 1.0703856309410185, 'learning_rate/qwen_vl_interface': 7.654760371235073e-06, 'learning_rate/action_model': 7.654760371235074e-05, 'epoch': 0.13}) +[2026-06-24 12:37:30,322][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.17849597334861755, 'train/grad_norm_pre_clip': 57.01789474487305, 'timing/data': 0.0003027778584510088, 'timing/model': 1.0696921411436051, 'learning_rate/qwen_vl_interface': 7.651459876432694e-06, 'learning_rate/action_model': 7.651459876432694e-05, 'epoch': 0.13}) +[2026-06-24 12:37:31,398][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.05832541733980179, 'train/grad_norm_pre_clip': 12.306778907775879, 'timing/data': 0.0002982090227305889, 'timing/model': 1.0744368832092732, 'learning_rate/qwen_vl_interface': 7.648157823349074e-06, 'learning_rate/action_model': 7.648157823349074e-05, 'epoch': 0.13}) +[2026-06-24 12:37:32,471][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.16146820783615112, 'train/grad_norm_pre_clip': 25.00178337097168, 'timing/data': 0.0003007028717547655, 'timing/model': 1.071558389812708, 'learning_rate/qwen_vl_interface': 7.64485421412688e-06, 'learning_rate/action_model': 7.644854214126879e-05, 'epoch': 0.13}) +[2026-06-24 12:37:33,546][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.2417377531528473, 'train/grad_norm_pre_clip': 38.82170867919922, 'timing/data': 0.0003256499767303467, 'timing/model': 1.0732117570005357, 'learning_rate/qwen_vl_interface': 7.641549050909784e-06, 'learning_rate/action_model': 7.641549050909784e-05, 'epoch': 0.13}) +[2026-06-24 12:37:34,623][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.07949237525463104, 'train/grad_norm_pre_clip': 36.66758728027344, 'timing/data': 0.000307655893266201, 'timing/model': 1.0761446449905634, 'learning_rate/qwen_vl_interface': 7.638242335842476e-06, 'learning_rate/action_model': 7.638242335842476e-05, 'epoch': 0.14}) +[2026-06-24 12:37:35,905][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.06205875799059868, 'train/grad_norm_pre_clip': 31.336076736450195, 'timing/data': 0.0003016451373696327, 'timing/model': 1.2801253241486847, 'learning_rate/qwen_vl_interface': 7.634934071070643e-06, 'learning_rate/action_model': 7.634934071070643e-05, 'epoch': 0.14}) +[2026-06-24 12:37:36,978][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.22214995324611664, 'train/grad_norm_pre_clip': 80.7959976196289, 'timing/data': 0.00031459005549550056, 'timing/model': 1.0714714659843594, 'learning_rate/qwen_vl_interface': 7.631624258740985e-06, 'learning_rate/action_model': 7.631624258740984e-05, 'epoch': 0.14}) +[2026-06-24 12:37:38,050][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.485603928565979, 'train/grad_norm_pre_clip': 67.10395812988281, 'timing/data': 0.00035808118991553783, 'timing/model': 1.0700061209499836, 'learning_rate/qwen_vl_interface': 7.628312901001203e-06, 'learning_rate/action_model': 7.628312901001204e-05, 'epoch': 0.14}) +[2026-06-24 12:37:39,124][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.1910388320684433, 'train/grad_norm_pre_clip': 41.4686279296875, 'timing/data': 0.00030862889252603054, 'timing/model': 1.072472742991522, 'learning_rate/qwen_vl_interface': 7.625e-06, 'learning_rate/action_model': 7.625e-05, 'epoch': 0.14}) +[2026-06-24 12:37:40,197][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.16484196484088898, 'train/grad_norm_pre_clip': 18.990095138549805, 'timing/data': 0.00033899489790201187, 'timing/model': 1.0720687848515809, 'learning_rate/qwen_vl_interface': 7.621685557887083e-06, 'learning_rate/action_model': 7.621685557887083e-05, 'epoch': 0.14}) +[2026-06-24 12:37:41,272][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.19939598441123962, 'train/grad_norm_pre_clip': 45.1781005859375, 'timing/data': 0.0003071362152695656, 'timing/model': 1.073178936028853, 'learning_rate/qwen_vl_interface': 7.618369576813157e-06, 'learning_rate/action_model': 7.618369576813157e-05, 'epoch': 0.14}) +[2026-06-24 12:37:42,344][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.15485627949237823, 'train/grad_norm_pre_clip': 30.72810935974121, 'timing/data': 0.00032482901588082314, 'timing/model': 1.0706725989002734, 'learning_rate/qwen_vl_interface': 7.615052058929924e-06, 'learning_rate/action_model': 7.615052058929923e-05, 'epoch': 0.14}) +[2026-06-24 12:37:43,418][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.09004691988229752, 'train/grad_norm_pre_clip': 17.86953353881836, 'timing/data': 0.00030988105572760105, 'timing/model': 1.0716841649264097, 'learning_rate/qwen_vl_interface': 7.611733006390087e-06, 'learning_rate/action_model': 7.611733006390086e-05, 'epoch': 0.14}) +[2026-06-24 12:37:44,491][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.1883092075586319, 'train/grad_norm_pre_clip': 26.620895385742188, 'timing/data': 0.0003067550715059042, 'timing/model': 1.07143815420568, 'learning_rate/qwen_vl_interface': 7.608412421347344e-06, 'learning_rate/action_model': 7.608412421347344e-05, 'epoch': 0.14}) +[2026-06-24 12:37:45,562][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.30912867188453674, 'train/grad_norm_pre_clip': 37.06059646606445, 'timing/data': 0.0003048619255423546, 'timing/model': 1.0703530709724873, 'learning_rate/qwen_vl_interface': 7.605090305956384e-06, 'learning_rate/action_model': 7.605090305956383e-05, 'epoch': 0.14}) +[2026-06-24 12:37:46,635][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.17469719052314758, 'train/grad_norm_pre_clip': 23.314496994018555, 'timing/data': 0.0003155821468681097, 'timing/model': 1.0714261520188302, 'learning_rate/qwen_vl_interface': 7.6017666623728954e-06, 'learning_rate/action_model': 7.601766662372896e-05, 'epoch': 0.14}) +[2026-06-24 12:37:47,708][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.1550523191690445, 'train/grad_norm_pre_clip': 38.754859924316406, 'timing/data': 0.0003296269569545984, 'timing/model': 1.0712405941449106, 'learning_rate/qwen_vl_interface': 7.598441492753551e-06, 'learning_rate/action_model': 7.598441492753551e-05, 'epoch': 0.14}) +[2026-06-24 12:37:48,782][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.05990884453058243, 'train/grad_norm_pre_clip': 17.864625930786133, 'timing/data': 0.00029866001568734646, 'timing/model': 1.0722644128836691, 'learning_rate/qwen_vl_interface': 7.595114799256018e-06, 'learning_rate/action_model': 7.595114799256017e-05, 'epoch': 0.14}) +[2026-06-24 12:37:49,858][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.12018553167581558, 'train/grad_norm_pre_clip': 25.35336685180664, 'timing/data': 0.0003100619651377201, 'timing/model': 1.074112466070801, 'learning_rate/qwen_vl_interface': 7.591786584038954e-06, 'learning_rate/action_model': 7.591786584038954e-05, 'epoch': 0.14}) +[2026-06-24 12:37:50,929][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.32115262746810913, 'train/grad_norm_pre_clip': 55.612892150878906, 'timing/data': 0.00032606092281639576, 'timing/model': 1.0696615451015532, 'learning_rate/qwen_vl_interface': 7.588456849261999e-06, 'learning_rate/action_model': 7.588456849261998e-05, 'epoch': 0.14}) +[2026-06-24 12:37:52,002][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.20084959268569946, 'train/grad_norm_pre_clip': 31.423904418945312, 'timing/data': 0.00032990798354148865, 'timing/model': 1.0718177759554237, 'learning_rate/qwen_vl_interface': 7.585125597085784e-06, 'learning_rate/action_model': 7.585125597085783e-05, 'epoch': 0.14}) +[2026-06-24 12:37:53,274][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.10213446617126465, 'train/grad_norm_pre_clip': 12.354040145874023, 'timing/data': 0.0003016861155629158, 'timing/model': 1.2703997280914336, 'learning_rate/qwen_vl_interface': 7.58179282967192e-06, 'learning_rate/action_model': 7.58179282967192e-05, 'epoch': 0.14}) +[2026-06-24 12:37:54,347][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.2238687425851822, 'train/grad_norm_pre_clip': 25.64832878112793, 'timing/data': 0.00036059506237506866, 'timing/model': 1.071452891919762, 'learning_rate/qwen_vl_interface': 7.578458549183005e-06, 'learning_rate/action_model': 7.578458549183005e-05, 'epoch': 0.14}) +[2026-06-24 12:37:55,421][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.07621312141418457, 'train/grad_norm_pre_clip': 19.872901916503906, 'timing/data': 0.0003123150672763586, 'timing/model': 1.0716636159922928, 'learning_rate/qwen_vl_interface': 7.575122757782617e-06, 'learning_rate/action_model': 7.575122757782617e-05, 'epoch': 0.14}) +[2026-06-24 12:37:56,491][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.15723079442977905, 'train/grad_norm_pre_clip': 38.58829879760742, 'timing/data': 0.0003086088690906763, 'timing/model': 1.0688505349680781, 'learning_rate/qwen_vl_interface': 7.571785457635316e-06, 'learning_rate/action_model': 7.571785457635317e-05, 'epoch': 0.14}) +[2026-06-24 12:37:57,564][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.14385448396205902, 'train/grad_norm_pre_clip': 22.513063430786133, 'timing/data': 0.00030742608942091465, 'timing/model': 1.0712486989796162, 'learning_rate/qwen_vl_interface': 7.568446650906637e-06, 'learning_rate/action_model': 7.568446650906636e-05, 'epoch': 0.14}) +[2026-06-24 12:37:58,640][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.2257886379957199, 'train/grad_norm_pre_clip': 25.463363647460938, 'timing/data': 0.00033435714431107044, 'timing/model': 1.0748100839555264, 'learning_rate/qwen_vl_interface': 7.565106339763099e-06, 'learning_rate/action_model': 7.565106339763099e-05, 'epoch': 0.14}) +[2026-06-24 12:37:59,713][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.19143158197402954, 'train/grad_norm_pre_clip': 18.143108367919922, 'timing/data': 0.0003028570208698511, 'timing/model': 1.0708270699251443, 'learning_rate/qwen_vl_interface': 7.561764526372191e-06, 'learning_rate/action_model': 7.561764526372191e-05, 'epoch': 0.14}) +[2026-06-24 12:38:00,788][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.04065864905714989, 'train/grad_norm_pre_clip': 9.47650146484375, 'timing/data': 0.0003109320532530546, 'timing/model': 1.0735713220201433, 'learning_rate/qwen_vl_interface': 7.5584212129023796e-06, 'learning_rate/action_model': 7.558421212902379e-05, 'epoch': 0.14}) +[2026-06-24 12:38:01,864][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.34146520495414734, 'train/grad_norm_pre_clip': 42.87059020996094, 'timing/data': 0.0003639019560068846, 'timing/model': 1.0748828200157732, 'learning_rate/qwen_vl_interface': 7.555076401523106e-06, 'learning_rate/action_model': 7.555076401523105e-05, 'epoch': 0.14}) +[2026-06-24 12:38:02,936][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.20881037414073944, 'train/grad_norm_pre_clip': 39.28690719604492, 'timing/data': 0.00032372702844440937, 'timing/model': 1.0701806680299342, 'learning_rate/qwen_vl_interface': 7.551730094404782e-06, 'learning_rate/action_model': 7.551730094404781e-05, 'epoch': 0.14}) +[2026-06-24 12:38:04,011][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.5424945950508118, 'train/grad_norm_pre_clip': 66.64551544189453, 'timing/data': 0.00036614597775042057, 'timing/model': 1.0724763600155711, 'learning_rate/qwen_vl_interface': 7.5483822937187876e-06, 'learning_rate/action_model': 7.548382293718787e-05, 'epoch': 0.14}) +[2026-06-24 12:38:05,083][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.14901597797870636, 'train/grad_norm_pre_clip': 43.44825744628906, 'timing/data': 0.00030947010964155197, 'timing/model': 1.0709397399332374, 'learning_rate/qwen_vl_interface': 7.545033001637478e-06, 'learning_rate/action_model': 7.545033001637479e-05, 'epoch': 0.14}) +[2026-06-24 12:38:06,156][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.1450204849243164, 'train/grad_norm_pre_clip': 23.651195526123047, 'timing/data': 0.0003486829809844494, 'timing/model': 1.07127801515162, 'learning_rate/qwen_vl_interface': 7.54168222033417e-06, 'learning_rate/action_model': 7.54168222033417e-05, 'epoch': 0.14}) +[2026-06-24 12:38:07,231][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.16163957118988037, 'train/grad_norm_pre_clip': 53.22993469238281, 'timing/data': 0.0003175849560648203, 'timing/model': 1.0734743908978999, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.14}) +[2026-06-24 12:38:08,304][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.06444887816905975, 'train/grad_norm_pre_clip': 16.10363006591797, 'timing/data': 0.00035706907510757446, 'timing/model': 1.0710812550969422, 'learning_rate/qwen_vl_interface': 7.534976198759667e-06, 'learning_rate/action_model': 7.534976198759666e-05, 'epoch': 0.14}) +[2026-06-24 12:38:09,380][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.04115603491663933, 'train/grad_norm_pre_clip': 6.640848636627197, 'timing/data': 0.00047017098404467106, 'timing/model': 1.074257749132812, 'learning_rate/qwen_vl_interface': 7.531620962839935e-06, 'learning_rate/action_model': 7.531620962839935e-05, 'epoch': 0.14}) +[2026-06-24 12:38:10,718][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.09883882105350494, 'train/grad_norm_pre_clip': 28.6199951171875, 'timing/data': 0.0003070649690926075, 'timing/model': 1.336756348144263, 'learning_rate/qwen_vl_interface': 7.5282642464011325e-06, 'learning_rate/action_model': 7.528264246401133e-05, 'epoch': 0.14}) +[2026-06-24 12:38:11,788][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.030769795179367065, 'train/grad_norm_pre_clip': 12.50203800201416, 'timing/data': 0.0003022761084139347, 'timing/model': 1.0683449369389564, 'learning_rate/qwen_vl_interface': 7.524906051621397e-06, 'learning_rate/action_model': 7.524906051621396e-05, 'epoch': 0.14}) +[2026-06-24 12:38:12,864][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.0359993577003479, 'train/grad_norm_pre_clip': 16.694856643676758, 'timing/data': 0.0003023669123649597, 'timing/model': 1.0737251800019294, 'learning_rate/qwen_vl_interface': 7.521546380679822e-06, 'learning_rate/action_model': 7.521546380679822e-05, 'epoch': 0.14}) +[2026-06-24 12:38:13,937][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.05520612746477127, 'train/grad_norm_pre_clip': 65.9513931274414, 'timing/data': 0.0003028071951121092, 'timing/model': 1.0716398518998176, 'learning_rate/qwen_vl_interface': 7.518185235756461e-06, 'learning_rate/action_model': 7.518185235756461e-05, 'epoch': 0.14}) +[2026-06-24 12:38:15,012][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.03979669511318207, 'train/grad_norm_pre_clip': 17.56467628479004, 'timing/data': 0.0003215919714421034, 'timing/model': 1.0734834969043732, 'learning_rate/qwen_vl_interface': 7.514822619032328e-06, 'learning_rate/action_model': 7.514822619032327e-05, 'epoch': 0.14}) +[2026-06-24 12:38:16,086][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.3177603781223297, 'train/grad_norm_pre_clip': 51.462242126464844, 'timing/data': 0.000348854111507535, 'timing/model': 1.0728637359570712, 'learning_rate/qwen_vl_interface': 7.511458532689385e-06, 'learning_rate/action_model': 7.511458532689385e-05, 'epoch': 0.14}) +[2026-06-24 12:38:17,162][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.4494810700416565, 'train/grad_norm_pre_clip': 72.77616119384766, 'timing/data': 0.0004842879716306925, 'timing/model': 1.073431700002402, 'learning_rate/qwen_vl_interface': 7.508092978910553e-06, 'learning_rate/action_model': 7.508092978910552e-05, 'epoch': 0.14}) +[2026-06-24 12:38:18,236][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.25098273158073425, 'train/grad_norm_pre_clip': 66.07716369628906, 'timing/data': 0.00032101082615554333, 'timing/model': 1.0729348990134895, 'learning_rate/qwen_vl_interface': 7.5047259598797015e-06, 'learning_rate/action_model': 7.504725959879701e-05, 'epoch': 0.14}) +[2026-06-24 12:38:19,310][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.28613877296447754, 'train/grad_norm_pre_clip': 101.15343475341797, 'timing/data': 0.00034269201569259167, 'timing/model': 1.0719115019310266, 'learning_rate/qwen_vl_interface': 7.501357477781654e-06, 'learning_rate/action_model': 7.501357477781654e-05, 'epoch': 0.14}) +[2026-06-24 12:38:20,385][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.12867818772792816, 'train/grad_norm_pre_clip': 35.88459777832031, 'timing/data': 0.0003758340608328581, 'timing/model': 1.0733907839749008, 'learning_rate/qwen_vl_interface': 7.497987534802185e-06, 'learning_rate/action_model': 7.497987534802184e-05, 'epoch': 0.14}) +[2026-06-24 12:38:21,459][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.08856580406427383, 'train/grad_norm_pre_clip': 61.20509338378906, 'timing/data': 0.0005141831934452057, 'timing/model': 1.072374840034172, 'learning_rate/qwen_vl_interface': 7.494616133128008e-06, 'learning_rate/action_model': 7.494616133128008e-05, 'epoch': 0.14}) +[2026-06-24 12:38:22,535][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.0987989529967308, 'train/grad_norm_pre_clip': 37.88967514038086, 'timing/data': 0.00032750400714576244, 'timing/model': 1.0738908709026873, 'learning_rate/qwen_vl_interface': 7.49124327494679e-06, 'learning_rate/action_model': 7.49124327494679e-05, 'epoch': 0.14}) +[2026-06-24 12:38:23,611][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.15128573775291443, 'train/grad_norm_pre_clip': 53.56230545043945, 'timing/data': 0.0003674379549920559, 'timing/model': 1.075042539043352, 'learning_rate/qwen_vl_interface': 7.487868962447147e-06, 'learning_rate/action_model': 7.487868962447147e-05, 'epoch': 0.14}) +[2026-06-24 12:38:24,686][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.53382408618927, 'train/grad_norm_pre_clip': 78.40990447998047, 'timing/data': 0.0003408989869058132, 'timing/model': 1.0733853029087186, 'learning_rate/qwen_vl_interface': 7.48449319781863e-06, 'learning_rate/action_model': 7.484493197818629e-05, 'epoch': 0.14}) +[2026-06-24 12:38:25,761][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.2897472679615021, 'train/grad_norm_pre_clip': 54.18060302734375, 'timing/data': 0.0003210620488971472, 'timing/model': 1.072820045053959, 'learning_rate/qwen_vl_interface': 7.481115983251735e-06, 'learning_rate/action_model': 7.481115983251735e-05, 'epoch': 0.14}) +[2026-06-24 12:38:27,054][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.09167297929525375, 'train/grad_norm_pre_clip': 38.08124542236328, 'timing/data': 0.0003433940000832081, 'timing/model': 1.2916980949230492, 'learning_rate/qwen_vl_interface': 7.477737320937902e-06, 'learning_rate/action_model': 7.477737320937902e-05, 'epoch': 0.14}) +[2026-06-24 12:38:28,126][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.1401892453432083, 'train/grad_norm_pre_clip': 61.99230194091797, 'timing/data': 0.0003065839409828186, 'timing/model': 1.0708018019795418, 'learning_rate/qwen_vl_interface': 7.474357213069507e-06, 'learning_rate/action_model': 7.474357213069507e-05, 'epoch': 0.14}) +[2026-06-24 12:38:29,198][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.2227686196565628, 'train/grad_norm_pre_clip': 40.80902862548828, 'timing/data': 0.00030377903021872044, 'timing/model': 1.0706436559557915, 'learning_rate/qwen_vl_interface': 7.470975661839868e-06, 'learning_rate/action_model': 7.470975661839868e-05, 'epoch': 0.14}) +[2026-06-24 12:38:30,268][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.028124328702688217, 'train/grad_norm_pre_clip': 32.2584342956543, 'timing/data': 0.00042939488776028156, 'timing/model': 1.0680835579987615, 'learning_rate/qwen_vl_interface': 7.4675926694432336e-06, 'learning_rate/action_model': 7.467592669443233e-05, 'epoch': 0.14}) +[2026-06-24 12:38:31,342][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.0416211262345314, 'train/grad_norm_pre_clip': 23.512102127075195, 'timing/data': 0.00030783587135374546, 'timing/model': 1.0724709099158645, 'learning_rate/qwen_vl_interface': 7.4642082380747905e-06, 'learning_rate/action_model': 7.46420823807479e-05, 'epoch': 0.14}) +[2026-06-24 12:38:32,416][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.1276313066482544, 'train/grad_norm_pre_clip': 42.80181884765625, 'timing/data': 0.00030123512260615826, 'timing/model': 1.0720693580806255, 'learning_rate/qwen_vl_interface': 7.460822369930664e-06, 'learning_rate/action_model': 7.460822369930664e-05, 'epoch': 0.14}) +[2026-06-24 12:38:33,492][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.06798682361841202, 'train/grad_norm_pre_clip': 26.876405715942383, 'timing/data': 0.0003156219609081745, 'timing/model': 1.0740316251758486, 'learning_rate/qwen_vl_interface': 7.457435067207901e-06, 'learning_rate/action_model': 7.4574350672079e-05, 'epoch': 0.14}) +[2026-06-24 12:38:34,567][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.1462801992893219, 'train/grad_norm_pre_clip': 46.617427825927734, 'timing/data': 0.0003416799008846283, 'timing/model': 1.0735509749501944, 'learning_rate/qwen_vl_interface': 7.454046332104491e-06, 'learning_rate/action_model': 7.45404633210449e-05, 'epoch': 0.14}) +[2026-06-24 12:38:35,638][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.0963473841547966, 'train/grad_norm_pre_clip': 24.49782943725586, 'timing/data': 0.00029784790240228176, 'timing/model': 1.0699628898873925, 'learning_rate/qwen_vl_interface': 7.450656166819344e-06, 'learning_rate/action_model': 7.450656166819344e-05, 'epoch': 0.14}) +[2026-06-24 12:38:36,707][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.20264309644699097, 'train/grad_norm_pre_clip': 52.64900588989258, 'timing/data': 0.00032148207537829876, 'timing/model': 1.0669600230176002, 'learning_rate/qwen_vl_interface': 7.447264573552303e-06, 'learning_rate/action_model': 7.447264573552302e-05, 'epoch': 0.14}) +[2026-06-24 12:38:37,776][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.1504717618227005, 'train/grad_norm_pre_clip': 21.713298797607422, 'timing/data': 0.00036725797690451145, 'timing/model': 1.067195993848145, 'learning_rate/qwen_vl_interface': 7.443871554504135e-06, 'learning_rate/action_model': 7.443871554504134e-05, 'epoch': 0.14}) +[2026-06-24 12:38:38,847][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.05159265547990799, 'train/grad_norm_pre_clip': 20.77759552001953, 'timing/data': 0.0003122449852526188, 'timing/model': 1.0697525858413428, 'learning_rate/qwen_vl_interface': 7.440477111876535e-06, 'learning_rate/action_model': 7.440477111876535e-05, 'epoch': 0.14}) +[2026-06-24 12:38:39,920][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.050972990691661835, 'train/grad_norm_pre_clip': 12.088848114013672, 'timing/data': 0.0003324530553072691, 'timing/model': 1.0710537449922413, 'learning_rate/qwen_vl_interface': 7.437081247872118e-06, 'learning_rate/action_model': 7.437081247872118e-05, 'epoch': 0.14}) +[2026-06-24 12:38:40,993][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.13564226031303406, 'train/grad_norm_pre_clip': 51.15898132324219, 'timing/data': 0.00035849190317094326, 'timing/model': 1.0718874868471175, 'learning_rate/qwen_vl_interface': 7.433683964694425e-06, 'learning_rate/action_model': 7.433683964694425e-05, 'epoch': 0.14}) +[2026-06-24 12:38:42,069][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.018831167370080948, 'train/grad_norm_pre_clip': 7.944185733795166, 'timing/data': 0.0003004029858857393, 'timing/model': 1.0746166920289397, 'learning_rate/qwen_vl_interface': 7.430285264547913e-06, 'learning_rate/action_model': 7.430285264547913e-05, 'epoch': 0.14}) +[2026-06-24 12:38:43,143][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.16289663314819336, 'train/grad_norm_pre_clip': 57.863037109375, 'timing/data': 0.0003565379884094, 'timing/model': 1.0723122640047222, 'learning_rate/qwen_vl_interface': 7.426885149637964e-06, 'learning_rate/action_model': 7.426885149637965e-05, 'epoch': 0.14}) +[2026-06-24 12:38:44,500][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.012314693070948124, 'train/grad_norm_pre_clip': 6.332342624664307, 'timing/data': 0.0003264418337494135, 'timing/model': 1.3547884030267596, 'learning_rate/qwen_vl_interface': 7.42348362217088e-06, 'learning_rate/action_model': 7.423483622170879e-05, 'epoch': 0.14}) +[2026-06-24 12:38:45,574][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.12962055206298828, 'train/grad_norm_pre_clip': 35.068389892578125, 'timing/data': 0.00033972691744565964, 'timing/model': 1.0729393882211298, 'learning_rate/qwen_vl_interface': 7.420080684353865e-06, 'learning_rate/action_model': 7.420080684353865e-05, 'epoch': 0.14}) +[2026-06-24 12:38:46,646][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.19035601615905762, 'train/grad_norm_pre_clip': 67.29350280761719, 'timing/data': 0.00032869610004127026, 'timing/model': 1.0702394880354404, 'learning_rate/qwen_vl_interface': 7.416676338395057e-06, 'learning_rate/action_model': 7.416676338395057e-05, 'epoch': 0.14}) +[2026-06-24 12:38:47,723][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.09737511724233627, 'train/grad_norm_pre_clip': 43.2786979675293, 'timing/data': 0.0003153712023049593, 'timing/model': 1.0757456470746547, 'learning_rate/qwen_vl_interface': 7.413270586503496e-06, 'learning_rate/action_model': 7.413270586503496e-05, 'epoch': 0.14}) +[2026-06-24 12:38:48,796][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.026578467339277267, 'train/grad_norm_pre_clip': 20.532886505126953, 'timing/data': 0.0005550098139792681, 'timing/model': 1.0709831828717142, 'learning_rate/qwen_vl_interface': 7.409863430889137e-06, 'learning_rate/action_model': 7.409863430889136e-05, 'epoch': 0.14}) +[2026-06-24 12:38:49,868][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.18613006174564362, 'train/grad_norm_pre_clip': 51.39580535888672, 'timing/data': 0.0003099909517914057, 'timing/model': 1.0703871150035411, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.14}) +[2026-06-24 12:38:50,942][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.12928378582000732, 'train/grad_norm_pre_clip': 65.71479797363281, 'timing/data': 0.0002984690945595503, 'timing/model': 1.0720830541104078, 'learning_rate/qwen_vl_interface': 7.403044917336404e-06, 'learning_rate/action_model': 7.403044917336404e-05, 'epoch': 0.14}) +[2026-06-24 12:38:52,018][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.3764432668685913, 'train/grad_norm_pre_clip': 166.2711944580078, 'timing/data': 0.00032563996501266956, 'timing/model': 1.074570455122739, 'learning_rate/qwen_vl_interface': 7.399633563822487e-06, 'learning_rate/action_model': 7.399633563822487e-05, 'epoch': 0.14}) +[2026-06-24 12:38:53,092][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.10986831039190292, 'train/grad_norm_pre_clip': 33.39875411987305, 'timing/data': 0.0003480520099401474, 'timing/model': 1.0722928659524769, 'learning_rate/qwen_vl_interface': 7.396220815434691e-06, 'learning_rate/action_model': 7.39622081543469e-05, 'epoch': 0.14}) +[2026-06-24 12:38:54,166][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.20578159391880035, 'train/grad_norm_pre_clip': 29.830156326293945, 'timing/data': 0.0003163230139762163, 'timing/model': 1.0722427829168737, 'learning_rate/qwen_vl_interface': 7.392806674387506e-06, 'learning_rate/action_model': 7.392806674387506e-05, 'epoch': 0.14}) +[2026-06-24 12:38:55,237][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.1332034021615982, 'train/grad_norm_pre_clip': 54.50197219848633, 'timing/data': 0.0003458780702203512, 'timing/model': 1.0701361750252545, 'learning_rate/qwen_vl_interface': 7.389391142896336e-06, 'learning_rate/action_model': 7.389391142896335e-05, 'epoch': 0.14}) +[2026-06-24 12:38:56,307][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.11716771125793457, 'train/grad_norm_pre_clip': 57.65471267700195, 'timing/data': 0.0003071660175919533, 'timing/model': 1.068347112974152, 'learning_rate/qwen_vl_interface': 7.38597422317748e-06, 'learning_rate/action_model': 7.385974223177479e-05, 'epoch': 0.14}) +[2026-06-24 12:38:57,384][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.12294767051935196, 'train/grad_norm_pre_clip': 51.220184326171875, 'timing/data': 0.00039664306677877903, 'timing/model': 1.074955265969038, 'learning_rate/qwen_vl_interface': 7.382555917448141e-06, 'learning_rate/action_model': 7.382555917448141e-05, 'epoch': 0.14}) +[2026-06-24 12:38:58,459][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.158924862742424, 'train/grad_norm_pre_clip': 32.1963005065918, 'timing/data': 0.0003189679700881243, 'timing/model': 1.073961935006082, 'learning_rate/qwen_vl_interface': 7.379136227926418e-06, 'learning_rate/action_model': 7.379136227926418e-05, 'epoch': 0.14}) +[2026-06-24 12:38:59,534][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.21420715749263763, 'train/grad_norm_pre_clip': 47.802547454833984, 'timing/data': 0.0003106631338596344, 'timing/model': 1.07285126298666, 'learning_rate/qwen_vl_interface': 7.3757151568313144e-06, 'learning_rate/action_model': 7.375715156831315e-05, 'epoch': 0.14}) +[2026-06-24 12:39:00,808][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.20959705114364624, 'train/grad_norm_pre_clip': 103.63168334960938, 'timing/data': 0.0003113928250968456, 'timing/model': 1.2724355249665678, 'learning_rate/qwen_vl_interface': 7.3722927063827235e-06, 'learning_rate/action_model': 7.372292706382723e-05, 'epoch': 0.14}) +[2026-06-24 12:39:01,884][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.4255139231681824, 'train/grad_norm_pre_clip': 84.70240020751953, 'timing/data': 0.00031755492091178894, 'timing/model': 1.0741945919580758, 'learning_rate/qwen_vl_interface': 7.368868878801437e-06, 'learning_rate/action_model': 7.368868878801437e-05, 'epoch': 0.14}) +[2026-06-24 12:39:02,959][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.2326297014951706, 'train/grad_norm_pre_clip': 42.49528503417969, 'timing/data': 0.00036885100416839123, 'timing/model': 1.0740393800660968, 'learning_rate/qwen_vl_interface': 7.365443676309139e-06, 'learning_rate/action_model': 7.365443676309139e-05, 'epoch': 0.14}) +[2026-06-24 12:39:04,033][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.12286415696144104, 'train/grad_norm_pre_clip': 33.37749481201172, 'timing/data': 0.0005202549509704113, 'timing/model': 1.072079346049577, 'learning_rate/qwen_vl_interface': 7.362017101128406e-06, 'learning_rate/action_model': 7.362017101128406e-05, 'epoch': 0.14}) +[2026-06-24 12:39:05,104][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.0924263596534729, 'train/grad_norm_pre_clip': 35.71455764770508, 'timing/data': 0.00034717982634902, 'timing/model': 1.0693886750377715, 'learning_rate/qwen_vl_interface': 7.358589155482707e-06, 'learning_rate/action_model': 7.358589155482707e-05, 'epoch': 0.14}) +[2026-06-24 12:39:06,179][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.2281108796596527, 'train/grad_norm_pre_clip': 33.47684860229492, 'timing/data': 0.0003259910736232996, 'timing/model': 1.0728406340349466, 'learning_rate/qwen_vl_interface': 7.355159841596397e-06, 'learning_rate/action_model': 7.355159841596397e-05, 'epoch': 0.14}) +[2026-06-24 12:39:07,254][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.13205987215042114, 'train/grad_norm_pre_clip': 16.122041702270508, 'timing/data': 0.00031013088300824165, 'timing/model': 1.073455877136439, 'learning_rate/qwen_vl_interface': 7.351729161694722e-06, 'learning_rate/action_model': 7.351729161694722e-05, 'epoch': 0.14}) +[2026-06-24 12:39:08,332][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.10999151319265366, 'train/grad_norm_pre_clip': 19.33951187133789, 'timing/data': 0.00030593411065638065, 'timing/model': 1.0771259549073875, 'learning_rate/qwen_vl_interface': 7.348297118003811e-06, 'learning_rate/action_model': 7.348297118003811e-05, 'epoch': 0.14}) +[2026-06-24 12:39:09,409][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.08679228276014328, 'train/grad_norm_pre_clip': 13.234530448913574, 'timing/data': 0.00035251094959676266, 'timing/model': 1.0748466129880399, 'learning_rate/qwen_vl_interface': 7.3448637127506815e-06, 'learning_rate/action_model': 7.344863712750682e-05, 'epoch': 0.14}) +[2026-06-24 12:39:10,479][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.10301630198955536, 'train/grad_norm_pre_clip': 29.75343894958496, 'timing/data': 0.00033497693948447704, 'timing/model': 1.0686858969274908, 'learning_rate/qwen_vl_interface': 7.341428948163232e-06, 'learning_rate/action_model': 7.341428948163231e-05, 'epoch': 0.14}) +[2026-06-24 12:39:11,558][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.07973000407218933, 'train/grad_norm_pre_clip': 17.73164939880371, 'timing/data': 0.0003112440463155508, 'timing/model': 1.0771403920371085, 'learning_rate/qwen_vl_interface': 7.337992826470245e-06, 'learning_rate/action_model': 7.337992826470245e-05, 'epoch': 0.14}) +[2026-06-24 12:39:12,629][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.19401328265666962, 'train/grad_norm_pre_clip': 32.89023971557617, 'timing/data': 0.0003111730329692364, 'timing/model': 1.069567889906466, 'learning_rate/qwen_vl_interface': 7.334555349901381e-06, 'learning_rate/action_model': 7.33455534990138e-05, 'epoch': 0.14}) +[2026-06-24 12:39:13,702][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.055155713111162186, 'train/grad_norm_pre_clip': 8.527527809143066, 'timing/data': 0.00031052203848958015, 'timing/model': 1.071377081098035, 'learning_rate/qwen_vl_interface': 7.331116520687184e-06, 'learning_rate/action_model': 7.331116520687184e-05, 'epoch': 0.14}) +[2026-06-24 12:39:14,776][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.04466965049505234, 'train/grad_norm_pre_clip': 13.632258415222168, 'timing/data': 0.0003398170229047537, 'timing/model': 1.0729443980380893, 'learning_rate/qwen_vl_interface': 7.32767634105907e-06, 'learning_rate/action_model': 7.32767634105907e-05, 'epoch': 0.14}) +[2026-06-24 12:39:15,850][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.14280395209789276, 'train/grad_norm_pre_clip': 24.95174789428711, 'timing/data': 0.00031692394986748695, 'timing/model': 1.0722296191379428, 'learning_rate/qwen_vl_interface': 7.324234813249334e-06, 'learning_rate/action_model': 7.324234813249335e-05, 'epoch': 0.14}) +[2026-06-24 12:39:17,141][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.19637221097946167, 'train/grad_norm_pre_clip': 32.240936279296875, 'timing/data': 0.00031533115543425083, 'timing/model': 1.2890041859354824, 'learning_rate/qwen_vl_interface': 7.320791939491149e-06, 'learning_rate/action_model': 7.320791939491149e-05, 'epoch': 0.14}) +[2026-06-24 12:39:18,212][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.13213476538658142, 'train/grad_norm_pre_clip': 48.09425354003906, 'timing/data': 0.0003149700351059437, 'timing/model': 1.0695931068621576, 'learning_rate/qwen_vl_interface': 7.317347722018557e-06, 'learning_rate/action_model': 7.317347722018557e-05, 'epoch': 0.14}) +[2026-06-24 12:39:19,281][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.09738624840974808, 'train/grad_norm_pre_clip': 16.472152709960938, 'timing/data': 0.0003051718231290579, 'timing/model': 1.0678084229584783, 'learning_rate/qwen_vl_interface': 7.313902163066472e-06, 'learning_rate/action_model': 7.313902163066471e-05, 'epoch': 0.14}) +[2026-06-24 12:39:20,355][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.1021774560213089, 'train/grad_norm_pre_clip': 23.737205505371094, 'timing/data': 0.00034927413798868656, 'timing/model': 1.0719686700031161, 'learning_rate/qwen_vl_interface': 7.3104552648706805e-06, 'learning_rate/action_model': 7.310455264870681e-05, 'epoch': 0.14}) +[2026-06-24 12:39:21,430][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.20876167714595795, 'train/grad_norm_pre_clip': 40.427982330322266, 'timing/data': 0.0003290870226919651, 'timing/model': 1.0737304619979113, 'learning_rate/qwen_vl_interface': 7.307007029667839e-06, 'learning_rate/action_model': 7.30700702966784e-05, 'epoch': 0.14}) +[2026-06-24 12:39:22,506][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.2341732531785965, 'train/grad_norm_pre_clip': 19.223613739013672, 'timing/data': 0.0003028979990631342, 'timing/model': 1.074313254095614, 'learning_rate/qwen_vl_interface': 7.303557459695468e-06, 'learning_rate/action_model': 7.303557459695469e-05, 'epoch': 0.14}) +[2026-06-24 12:39:23,579][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.15674811601638794, 'train/grad_norm_pre_clip': 64.38742065429688, 'timing/data': 0.00031328690238296986, 'timing/model': 1.0713838541414589, 'learning_rate/qwen_vl_interface': 7.300106557191956e-06, 'learning_rate/action_model': 7.300106557191956e-05, 'epoch': 0.14}) +[2026-06-24 12:39:24,657][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.09110758453607559, 'train/grad_norm_pre_clip': 20.83734893798828, 'timing/data': 0.00035961298272013664, 'timing/model': 1.075982472160831, 'learning_rate/qwen_vl_interface': 7.296654324396557e-06, 'learning_rate/action_model': 7.296654324396556e-05, 'epoch': 0.14}) +[2026-06-24 12:39:25,734][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.19628337025642395, 'train/grad_norm_pre_clip': 37.562400817871094, 'timing/data': 0.00032041012309491634, 'timing/model': 1.0756044138688594, 'learning_rate/qwen_vl_interface': 7.2932007635493875e-06, 'learning_rate/action_model': 7.293200763549387e-05, 'epoch': 0.14}) +[2026-06-24 12:39:26,809][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.3364560008049011, 'train/grad_norm_pre_clip': 58.694862365722656, 'timing/data': 0.0003033890388906002, 'timing/model': 1.0740253140684217, 'learning_rate/qwen_vl_interface': 7.289745876891427e-06, 'learning_rate/action_model': 7.289745876891426e-05, 'epoch': 0.15}) +[2026-06-24 12:39:27,882][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1500 +[2026-06-24 12:39:53,394][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1500 +[2026-06-24 12:39:53,395][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1500 +[2026-06-24 12:40:25,964][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1500 +[2026-06-24 12:41:58,012][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-24 12:41:58,013][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.14515891671180725, 'train/grad_norm_pre_clip': 22.58863639831543, 'eval/action_loss/seconds': 25.510891232872382, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.07175314426422119, 'eval/latency_2/loss': 0.07175314426422119, 'eval/flappy/loss': 0.07175314426422119, 'eval/flappy/latency_2/loss': 0.07175314426422119, 'eval/flap_precision': 0.8958868894601543, 'eval/flap_recall': 0.9457259158751696, 'eval/flap_f1': 0.9201320132013203, 'eval/flap_rate_model': 0.21611111111111111, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9201320132013202, 'eval/latency_2/flap_precision': 0.8958868894601543, 'eval/latency_2/flap_recall': 0.9457259158751696, 'eval/latency_2/flap_f1': 0.9201320132013203, 'eval/latency_2/flap_rate_model': 0.21611111111111111, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9201320132013202, 'eval/action_classification/data_wait_seconds': 0.30857303738594055, 'eval/action_classification/predict_seconds': 32.167205810546875, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 71.09999999999886, 'rl_games_eval/mid_train/mean_length': 592.4, 'rl_games_eval/mid_train/std_reward': 71.60558637424673, 'rl_games_eval/mid_train/std_length': 577.9593757350078, 'rl_games_eval/mid_train/macro_mean_reward': 71.09999999999886, 'rl_games_eval/mid_train/macro_mean_length': 592.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 71.09999999999886, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 592.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 71.60558637424673, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 577.9593757350078, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 71.09999999999886, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003156918101012707, 'timing/model': 1.0715506859123707, 'learning_rate/qwen_vl_interface': 7.286289666664508e-06, 'learning_rate/action_model': 7.286289666664508e-05, 'epoch': 0.15}) +[2026-06-24 12:41:58,013][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_1500_state +[2026-06-24 12:41:58,016][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 12:42:00,631][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_1500_state/model.safetensors +[2026-06-24 12:42:06,292][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_1500_state/optimizer.bin +[2026-06-24 12:42:06,295][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_1500_state/sampler.bin +[2026-06-24 12:42:06,296][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-24 12:42:06,296][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-24 12:42:06,305][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-24 12:42:06,305][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 12:42:06,358][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 12:42:07,497][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.11180935800075531, 'train/grad_norm_pre_clip': 19.296632766723633, 'timing/data': 0.00048823398537933826, 'timing/model': 1.1370912108104676, 'learning_rate/qwen_vl_interface': 7.282832135111334e-06, 'learning_rate/action_model': 7.282832135111334e-05, 'epoch': 0.15}) +[2026-06-24 12:42:08,561][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.08590985089540482, 'train/grad_norm_pre_clip': 18.90386390686035, 'timing/data': 0.00043414998799562454, 'timing/model': 1.0621954400558025, 'learning_rate/qwen_vl_interface': 7.2793732844754584e-06, 'learning_rate/action_model': 7.279373284475459e-05, 'epoch': 0.15}) +[2026-06-24 12:42:09,621][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.17283658683300018, 'train/grad_norm_pre_clip': 25.213546752929688, 'timing/data': 0.0003875808324664831, 'timing/model': 1.058222308056429, 'learning_rate/qwen_vl_interface': 7.275913117001291e-06, 'learning_rate/action_model': 7.27591311700129e-05, 'epoch': 0.15}) +[2026-06-24 12:42:10,685][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.07532870024442673, 'train/grad_norm_pre_clip': 15.773018836975098, 'timing/data': 0.00036694202572107315, 'timing/model': 1.0623821411281824, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.15}) +[2026-06-24 12:42:11,750][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.11624141782522202, 'train/grad_norm_pre_clip': 32.488590240478516, 'timing/data': 0.00031800800934433937, 'timing/model': 1.063379269093275, 'learning_rate/qwen_vl_interface': 7.268988840519994e-06, 'learning_rate/action_model': 7.268988840519993e-05, 'epoch': 0.15}) +[2026-06-24 12:42:12,816][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.11390839517116547, 'train/grad_norm_pre_clip': 18.452991485595703, 'timing/data': 0.00038318196311593056, 'timing/model': 1.0638503308873624, 'learning_rate/qwen_vl_interface': 7.265524736005951e-06, 'learning_rate/action_model': 7.26552473600595e-05, 'epoch': 0.15}) +[2026-06-24 12:42:13,881][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.0378134660422802, 'train/grad_norm_pre_clip': 9.42154598236084, 'timing/data': 0.00033909804187715054, 'timing/model': 1.0640902421437204, 'learning_rate/qwen_vl_interface': 7.262059323639792e-06, 'learning_rate/action_model': 7.262059323639792e-05, 'epoch': 0.15}) +[2026-06-24 12:42:14,944][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.19272996485233307, 'train/grad_norm_pre_clip': 46.835357666015625, 'timing/data': 0.0003364528529345989, 'timing/model': 1.06047030701302, 'learning_rate/qwen_vl_interface': 7.25859260567018e-06, 'learning_rate/action_model': 7.258592605670179e-05, 'epoch': 0.15}) +[2026-06-24 12:42:16,007][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.24765583872795105, 'train/grad_norm_pre_clip': 64.82062530517578, 'timing/data': 0.00034278491511940956, 'timing/model': 1.062185107031837, 'learning_rate/qwen_vl_interface': 7.2551245843466334e-06, 'learning_rate/action_model': 7.255124584346633e-05, 'epoch': 0.15}) +[2026-06-24 12:42:17,070][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.06391263008117676, 'train/grad_norm_pre_clip': 21.749427795410156, 'timing/data': 0.0004930519498884678, 'timing/model': 1.0610418221913278, 'learning_rate/qwen_vl_interface': 7.251655261919515e-06, 'learning_rate/action_model': 7.251655261919515e-05, 'epoch': 0.15}) +[2026-06-24 12:42:18,136][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.22688226401805878, 'train/grad_norm_pre_clip': 42.37836837768555, 'timing/data': 0.0003143409267067909, 'timing/model': 1.0643593838904053, 'learning_rate/qwen_vl_interface': 7.248184640640027e-06, 'learning_rate/action_model': 7.248184640640026e-05, 'epoch': 0.15}) +[2026-06-24 12:42:19,501][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.026007095351815224, 'train/grad_norm_pre_clip': 15.673578262329102, 'timing/data': 0.00039222906343638897, 'timing/model': 1.3630150449462235, 'learning_rate/qwen_vl_interface': 7.24471272276022e-06, 'learning_rate/action_model': 7.24471272276022e-05, 'epoch': 0.15}) +[2026-06-24 12:42:20,568][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.038510482758283615, 'train/grad_norm_pre_clip': 13.2641019821167, 'timing/data': 0.00030543399043381214, 'timing/model': 1.0649895970709622, 'learning_rate/qwen_vl_interface': 7.241239510532985e-06, 'learning_rate/action_model': 7.241239510532984e-05, 'epoch': 0.15}) +[2026-06-24 12:42:21,635][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.011902195401489735, 'train/grad_norm_pre_clip': 12.940423011779785, 'timing/data': 0.0003058440051972866, 'timing/model': 1.0652652462013066, 'learning_rate/qwen_vl_interface': 7.237765006212049e-06, 'learning_rate/action_model': 7.237765006212048e-05, 'epoch': 0.15}) +[2026-06-24 12:42:22,706][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.43906474113464355, 'train/grad_norm_pre_clip': 76.14767456054688, 'timing/data': 0.00034199305810034275, 'timing/model': 1.069556549191475, 'learning_rate/qwen_vl_interface': 7.2342892120519805e-06, 'learning_rate/action_model': 7.23428921205198e-05, 'epoch': 0.15}) +[2026-06-24 12:42:23,771][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.16626861691474915, 'train/grad_norm_pre_clip': 74.24571228027344, 'timing/data': 0.00036239204928278923, 'timing/model': 1.0635671999771148, 'learning_rate/qwen_vl_interface': 7.230812130308187e-06, 'learning_rate/action_model': 7.230812130308186e-05, 'epoch': 0.15}) +[2026-06-24 12:42:24,837][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.2636054754257202, 'train/grad_norm_pre_clip': 38.086238861083984, 'timing/data': 0.00032483087852597237, 'timing/model': 1.064457193017006, 'learning_rate/qwen_vl_interface': 7.227333763236907e-06, 'learning_rate/action_model': 7.227333763236906e-05, 'epoch': 0.15}) +[2026-06-24 12:42:25,905][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.25989285111427307, 'train/grad_norm_pre_clip': 41.957305908203125, 'timing/data': 0.00036970595829188824, 'timing/model': 1.0663915029726923, 'learning_rate/qwen_vl_interface': 7.223854113095214e-06, 'learning_rate/action_model': 7.223854113095214e-05, 'epoch': 0.15}) +[2026-06-24 12:42:26,973][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.19287431240081787, 'train/grad_norm_pre_clip': 21.656373977661133, 'timing/data': 0.00043542194180190563, 'timing/model': 1.0663631320931017, 'learning_rate/qwen_vl_interface': 7.220373182141021e-06, 'learning_rate/action_model': 7.22037318214102e-05, 'epoch': 0.15}) +[2026-06-24 12:42:28,043][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.03586522862315178, 'train/grad_norm_pre_clip': 30.141359329223633, 'timing/data': 0.0003121460322290659, 'timing/model': 1.0684917499311268, 'learning_rate/qwen_vl_interface': 7.216890972633057e-06, 'learning_rate/action_model': 7.216890972633057e-05, 'epoch': 0.15}) +[2026-06-24 12:42:29,114][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.08493834733963013, 'train/grad_norm_pre_clip': 37.686607360839844, 'timing/data': 0.0003041510935872793, 'timing/model': 1.0690660090185702, 'learning_rate/qwen_vl_interface': 7.213407486830899e-06, 'learning_rate/action_model': 7.213407486830899e-05, 'epoch': 0.15}) +[2026-06-24 12:42:30,184][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.01928534358739853, 'train/grad_norm_pre_clip': 10.157475471496582, 'timing/data': 0.00038895290344953537, 'timing/model': 1.068494631908834, 'learning_rate/qwen_vl_interface': 7.209922726994937e-06, 'learning_rate/action_model': 7.209922726994936e-05, 'epoch': 0.15}) +[2026-06-24 12:42:31,250][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.1595991551876068, 'train/grad_norm_pre_clip': 45.10986328125, 'timing/data': 0.0003162538632750511, 'timing/model': 1.0647818639408797, 'learning_rate/qwen_vl_interface': 7.206436695386396e-06, 'learning_rate/action_model': 7.206436695386396e-05, 'epoch': 0.15}) +[2026-06-24 12:42:32,319][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.1356523036956787, 'train/grad_norm_pre_clip': 29.563688278198242, 'timing/data': 0.0003369329497218132, 'timing/model': 1.0666848339606076, 'learning_rate/qwen_vl_interface': 7.202949394267326e-06, 'learning_rate/action_model': 7.202949394267326e-05, 'epoch': 0.15}) +[2026-06-24 12:42:33,389][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.10720033943653107, 'train/grad_norm_pre_clip': 22.205078125, 'timing/data': 0.0003202019724994898, 'timing/model': 1.0682755471207201, 'learning_rate/qwen_vl_interface': 7.199460825900596e-06, 'learning_rate/action_model': 7.199460825900596e-05, 'epoch': 0.15}) +[2026-06-24 12:42:34,459][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.1737685650587082, 'train/grad_norm_pre_clip': 42.73355484008789, 'timing/data': 0.0003334670327603817, 'timing/model': 1.0685021651443094, 'learning_rate/qwen_vl_interface': 7.1959709925499e-06, 'learning_rate/action_model': 7.1959709925499e-05, 'epoch': 0.15}) +[2026-06-24 12:42:35,526][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.04212145879864693, 'train/grad_norm_pre_clip': 17.710176467895508, 'timing/data': 0.00034221401438117027, 'timing/model': 1.06540984287858, 'learning_rate/qwen_vl_interface': 7.1924798964797545e-06, 'learning_rate/action_model': 7.192479896479754e-05, 'epoch': 0.15}) +[2026-06-24 12:42:36,875][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.12042194604873657, 'train/grad_norm_pre_clip': 37.894683837890625, 'timing/data': 0.00031688506714999676, 'timing/model': 1.3474566829390824, 'learning_rate/qwen_vl_interface': 7.188987539955495e-06, 'learning_rate/action_model': 7.188987539955494e-05, 'epoch': 0.15}) +[2026-06-24 12:42:37,944][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.02889499068260193, 'train/grad_norm_pre_clip': 16.31260108947754, 'timing/data': 0.00035818410106003284, 'timing/model': 1.0677108520176262, 'learning_rate/qwen_vl_interface': 7.185493925243271e-06, 'learning_rate/action_model': 7.185493925243271e-05, 'epoch': 0.15}) +[2026-06-24 12:42:39,009][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.027244577184319496, 'train/grad_norm_pre_clip': 27.515281677246094, 'timing/data': 0.00031356909312307835, 'timing/model': 1.0630808940622956, 'learning_rate/qwen_vl_interface': 7.181999054610053e-06, 'learning_rate/action_model': 7.181999054610052e-05, 'epoch': 0.15}) +[2026-06-24 12:42:40,076][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.10032590478658676, 'train/grad_norm_pre_clip': 30.278076171875, 'timing/data': 0.00031152484007179737, 'timing/model': 1.0661210529506207, 'learning_rate/qwen_vl_interface': 7.178502930323622e-06, 'learning_rate/action_model': 7.178502930323622e-05, 'epoch': 0.15}) +[2026-06-24 12:42:41,142][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.15253104269504547, 'train/grad_norm_pre_clip': 40.592674255371094, 'timing/data': 0.00036191102117300034, 'timing/model': 1.0642865560948849, 'learning_rate/qwen_vl_interface': 7.175005554652577e-06, 'learning_rate/action_model': 7.175005554652578e-05, 'epoch': 0.15}) +[2026-06-24 12:42:42,212][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.32934868335723877, 'train/grad_norm_pre_clip': 68.06794738769531, 'timing/data': 0.0003295701462775469, 'timing/model': 1.0675773748662323, 'learning_rate/qwen_vl_interface': 7.171506929866328e-06, 'learning_rate/action_model': 7.171506929866328e-05, 'epoch': 0.15}) +[2026-06-24 12:42:43,283][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.14328043162822723, 'train/grad_norm_pre_clip': 90.46058654785156, 'timing/data': 0.0003293389454483986, 'timing/model': 1.0693239758256823, 'learning_rate/qwen_vl_interface': 7.168007058235095e-06, 'learning_rate/action_model': 7.168007058235095e-05, 'epoch': 0.15}) +[2026-06-24 12:42:44,368][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.3386068344116211, 'train/grad_norm_pre_clip': 76.85090637207031, 'timing/data': 0.0004360429011285305, 'timing/model': 1.0837201080285013, 'learning_rate/qwen_vl_interface': 7.164505942029905e-06, 'learning_rate/action_model': 7.164505942029904e-05, 'epoch': 0.15}) +[2026-06-24 12:42:45,849][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.11118090897798538, 'train/grad_norm_pre_clip': 25.824260711669922, 'timing/data': 0.000500817084684968, 'timing/model': 1.4787861919030547, 'learning_rate/qwen_vl_interface': 7.161003583522594e-06, 'learning_rate/action_model': 7.161003583522593e-05, 'epoch': 0.15}) +[2026-06-24 12:42:46,918][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.12276848405599594, 'train/grad_norm_pre_clip': 42.75893020629883, 'timing/data': 0.0003080880269408226, 'timing/model': 1.0672682041767985, 'learning_rate/qwen_vl_interface': 7.157499984985804e-06, 'learning_rate/action_model': 7.157499984985803e-05, 'epoch': 0.15}) +[2026-06-24 12:42:47,987][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.289244681596756, 'train/grad_norm_pre_clip': 47.66238784790039, 'timing/data': 0.00031312694773077965, 'timing/model': 1.067580637987703, 'learning_rate/qwen_vl_interface': 7.153995148692985e-06, 'learning_rate/action_model': 7.153995148692985e-05, 'epoch': 0.15}) +[2026-06-24 12:42:49,057][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.20914261043071747, 'train/grad_norm_pre_clip': 71.61001586914062, 'timing/data': 0.0004039809573441744, 'timing/model': 1.068514975020662, 'learning_rate/qwen_vl_interface': 7.150489076918383e-06, 'learning_rate/action_model': 7.150489076918384e-05, 'epoch': 0.15}) +[2026-06-24 12:42:50,125][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.10365093499422073, 'train/grad_norm_pre_clip': 27.300273895263672, 'timing/data': 0.00031370879150927067, 'timing/model': 1.0664890939369798, 'learning_rate/qwen_vl_interface': 7.1469817719370535e-06, 'learning_rate/action_model': 7.146981771937053e-05, 'epoch': 0.15}) +[2026-06-24 12:42:51,196][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.17164987325668335, 'train/grad_norm_pre_clip': 48.66532516479492, 'timing/data': 0.0003144398797303438, 'timing/model': 1.068748682970181, 'learning_rate/qwen_vl_interface': 7.1434732360248505e-06, 'learning_rate/action_model': 7.14347323602485e-05, 'epoch': 0.15}) +[2026-06-24 12:42:52,264][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.24218441545963287, 'train/grad_norm_pre_clip': 60.570106506347656, 'timing/data': 0.00034690205939114094, 'timing/model': 1.0662796669639647, 'learning_rate/qwen_vl_interface': 7.139963471458418e-06, 'learning_rate/action_model': 7.139963471458418e-05, 'epoch': 0.15}) +[2026-06-24 12:42:53,332][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.05106062814593315, 'train/grad_norm_pre_clip': 44.744415283203125, 'timing/data': 0.00040621496737003326, 'timing/model': 1.0668471371755004, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.15}) +[2026-06-24 12:42:54,694][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.20797845721244812, 'train/grad_norm_pre_clip': 57.67973709106445, 'timing/data': 0.0003525631036609411, 'timing/model': 1.3599838218651712, 'learning_rate/qwen_vl_interface': 7.132940265473466e-06, 'learning_rate/action_model': 7.132940265473466e-05, 'epoch': 0.15}) +[2026-06-24 12:42:55,766][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.10234655439853668, 'train/grad_norm_pre_clip': 15.540865898132324, 'timing/data': 0.0003161029890179634, 'timing/model': 1.0699996131006628, 'learning_rate/qwen_vl_interface': 7.129426828612227e-06, 'learning_rate/action_model': 7.129426828612227e-05, 'epoch': 0.15}) +[2026-06-24 12:42:56,837][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.09960061311721802, 'train/grad_norm_pre_clip': 17.8023738861084, 'timing/data': 0.00035439711064100266, 'timing/model': 1.0696969029959291, 'learning_rate/qwen_vl_interface': 7.125912172211326e-06, 'learning_rate/action_model': 7.125912172211325e-05, 'epoch': 0.15}) +[2026-06-24 12:42:57,904][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.18659406900405884, 'train/grad_norm_pre_clip': 43.58070373535156, 'timing/data': 0.0004968780558556318, 'timing/model': 1.065843784948811, 'learning_rate/qwen_vl_interface': 7.122396298551381e-06, 'learning_rate/action_model': 7.12239629855138e-05, 'epoch': 0.15}) +[2026-06-24 12:42:58,975][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.22491638362407684, 'train/grad_norm_pre_clip': 50.979129791259766, 'timing/data': 0.0002991908695548773, 'timing/model': 1.0687720230780542, 'learning_rate/qwen_vl_interface': 7.118879209913806e-06, 'learning_rate/action_model': 7.118879209913805e-05, 'epoch': 0.15}) +[2026-06-24 12:43:00,045][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.16355283558368683, 'train/grad_norm_pre_clip': 58.94573211669922, 'timing/data': 0.00035119010135531425, 'timing/model': 1.068140562158078, 'learning_rate/qwen_vl_interface': 7.115360908580803e-06, 'learning_rate/action_model': 7.115360908580802e-05, 'epoch': 0.15}) +[2026-06-24 12:43:01,117][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.060757506638765335, 'train/grad_norm_pre_clip': 21.360149383544922, 'timing/data': 0.0003033189568668604, 'timing/model': 1.071081932168454, 'learning_rate/qwen_vl_interface': 7.1118413968353575e-06, 'learning_rate/action_model': 7.111841396835357e-05, 'epoch': 0.15}) +[2026-06-24 12:43:02,191][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.10055291652679443, 'train/grad_norm_pre_clip': 20.391618728637695, 'timing/data': 0.0003758571110665798, 'timing/model': 1.0722255420405418, 'learning_rate/qwen_vl_interface': 7.108320676961242e-06, 'learning_rate/action_model': 7.108320676961242e-05, 'epoch': 0.15}) +[2026-06-24 12:43:03,261][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.15752114355564117, 'train/grad_norm_pre_clip': 46.930274963378906, 'timing/data': 0.0003802250139415264, 'timing/model': 1.0686600857879966, 'learning_rate/qwen_vl_interface': 7.104798751243018e-06, 'learning_rate/action_model': 7.104798751243019e-05, 'epoch': 0.15}) +[2026-06-24 12:43:04,338][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.25969162583351135, 'train/grad_norm_pre_clip': 30.878639221191406, 'timing/data': 0.0003169949632138014, 'timing/model': 1.0746942299883813, 'learning_rate/qwen_vl_interface': 7.1012756219660195e-06, 'learning_rate/action_model': 7.101275621966019e-05, 'epoch': 0.15}) +[2026-06-24 12:43:05,413][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.07594367116689682, 'train/grad_norm_pre_clip': 30.768478393554688, 'timing/data': 0.00041574309580028057, 'timing/model': 1.0737342850770801, 'learning_rate/qwen_vl_interface': 7.09775129141637e-06, 'learning_rate/action_model': 7.097751291416369e-05, 'epoch': 0.15}) +[2026-06-24 12:43:06,486][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.1672498881816864, 'train/grad_norm_pre_clip': 33.226318359375, 'timing/data': 0.00031296699307858944, 'timing/model': 1.0711412751115859, 'learning_rate/qwen_vl_interface': 7.094225761880966e-06, 'learning_rate/action_model': 7.094225761880967e-05, 'epoch': 0.15}) +[2026-06-24 12:43:07,557][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.16454756259918213, 'train/grad_norm_pre_clip': 44.32733917236328, 'timing/data': 0.0003154720179736614, 'timing/model': 1.0691220809239894, 'learning_rate/qwen_vl_interface': 7.0906990356474895e-06, 'learning_rate/action_model': 7.090699035647489e-05, 'epoch': 0.15}) +[2026-06-24 12:43:08,629][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.3541322648525238, 'train/grad_norm_pre_clip': 102.68971252441406, 'timing/data': 0.0003061741590499878, 'timing/model': 1.0708411489613354, 'learning_rate/qwen_vl_interface': 7.087171115004393e-06, 'learning_rate/action_model': 7.087171115004393e-05, 'epoch': 0.15}) +[2026-06-24 12:43:09,704][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.0709022507071495, 'train/grad_norm_pre_clip': 20.321706771850586, 'timing/data': 0.00033326586708426476, 'timing/model': 1.0733893201686442, 'learning_rate/qwen_vl_interface': 7.083642002240905e-06, 'learning_rate/action_model': 7.083642002240905e-05, 'epoch': 0.15}) +[2026-06-24 12:43:10,774][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.10151014477014542, 'train/grad_norm_pre_clip': 32.350494384765625, 'timing/data': 0.00029345997609198093, 'timing/model': 1.0688499480020255, 'learning_rate/qwen_vl_interface': 7.080111699647031e-06, 'learning_rate/action_model': 7.080111699647031e-05, 'epoch': 0.15}) +[2026-06-24 12:43:12,134][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.1460922211408615, 'train/grad_norm_pre_clip': 18.62201499938965, 'timing/data': 0.0003030078951269388, 'timing/model': 1.3582595270127058, 'learning_rate/qwen_vl_interface': 7.076580209513542e-06, 'learning_rate/action_model': 7.076580209513542e-05, 'epoch': 0.15}) +[2026-06-24 12:43:13,207][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.095486119389534, 'train/grad_norm_pre_clip': 28.911420822143555, 'timing/data': 0.00030442094430327415, 'timing/model': 1.0714150010608137, 'learning_rate/qwen_vl_interface': 7.073047534131987e-06, 'learning_rate/action_model': 7.073047534131986e-05, 'epoch': 0.15}) +[2026-06-24 12:43:14,282][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.23930205404758453, 'train/grad_norm_pre_clip': 53.26899719238281, 'timing/data': 0.0003248590510338545, 'timing/model': 1.0727161520626396, 'learning_rate/qwen_vl_interface': 7.069513675794679e-06, 'learning_rate/action_model': 7.069513675794679e-05, 'epoch': 0.15}) +[2026-06-24 12:43:15,353][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.26629382371902466, 'train/grad_norm_pre_clip': 74.00865936279297, 'timing/data': 0.00030250707641243935, 'timing/model': 1.0694544999860227, 'learning_rate/qwen_vl_interface': 7.065978636794704e-06, 'learning_rate/action_model': 7.065978636794704e-05, 'epoch': 0.15}) +[2026-06-24 12:43:16,426][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.24085365235805511, 'train/grad_norm_pre_clip': 58.983238220214844, 'timing/data': 0.0003348691388964653, 'timing/model': 1.071000206982717, 'learning_rate/qwen_vl_interface': 7.062442419425904e-06, 'learning_rate/action_model': 7.062442419425903e-05, 'epoch': 0.15}) +[2026-06-24 12:43:17,498][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.25985902547836304, 'train/grad_norm_pre_clip': 40.3845329284668, 'timing/data': 0.00032487884163856506, 'timing/model': 1.0704880100674927, 'learning_rate/qwen_vl_interface': 7.058905025982895e-06, 'learning_rate/action_model': 7.058905025982894e-05, 'epoch': 0.15}) +[2026-06-24 12:43:18,570][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.08246307075023651, 'train/grad_norm_pre_clip': 21.94056510925293, 'timing/data': 0.0003001228906214237, 'timing/model': 1.071202082093805, 'learning_rate/qwen_vl_interface': 7.055366458761056e-06, 'learning_rate/action_model': 7.055366458761056e-05, 'epoch': 0.15}) +[2026-06-24 12:43:19,643][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.21290546655654907, 'train/grad_norm_pre_clip': 45.87014389038086, 'timing/data': 0.00029992195777595043, 'timing/model': 1.071426078910008, 'learning_rate/qwen_vl_interface': 7.051826720056522e-06, 'learning_rate/action_model': 7.051826720056521e-05, 'epoch': 0.15}) +[2026-06-24 12:43:20,713][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.06720791012048721, 'train/grad_norm_pre_clip': 19.37453269958496, 'timing/data': 0.00031868787482380867, 'timing/model': 1.0684241640847176, 'learning_rate/qwen_vl_interface': 7.04828581216619e-06, 'learning_rate/action_model': 7.04828581216619e-05, 'epoch': 0.15}) +[2026-06-24 12:43:21,788][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.1540135145187378, 'train/grad_norm_pre_clip': 20.693710327148438, 'timing/data': 0.0003355101216584444, 'timing/model': 1.0726162889041007, 'learning_rate/qwen_vl_interface': 7.04474373738772e-06, 'learning_rate/action_model': 7.04474373738772e-05, 'epoch': 0.15}) +[2026-06-24 12:43:22,863][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.17467568814754486, 'train/grad_norm_pre_clip': 37.29909896850586, 'timing/data': 0.00030993111431598663, 'timing/model': 1.0737750090193003, 'learning_rate/qwen_vl_interface': 7.041200498019523e-06, 'learning_rate/action_model': 7.041200498019523e-05, 'epoch': 0.15}) +[2026-06-24 12:43:23,935][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.1803974211215973, 'train/grad_norm_pre_clip': 31.1239013671875, 'timing/data': 0.0003204811364412308, 'timing/model': 1.0707487571053207, 'learning_rate/qwen_vl_interface': 7.037656096360771e-06, 'learning_rate/action_model': 7.037656096360771e-05, 'epoch': 0.15}) +[2026-06-24 12:43:25,008][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.1575745791196823, 'train/grad_norm_pre_clip': 29.08572006225586, 'timing/data': 0.000296375947073102, 'timing/model': 1.0714358380064368, 'learning_rate/qwen_vl_interface': 7.0341105347113844e-06, 'learning_rate/action_model': 7.034110534711384e-05, 'epoch': 0.15}) +[2026-06-24 12:43:26,080][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.11938299983739853, 'train/grad_norm_pre_clip': 22.889617919921875, 'timing/data': 0.00032433890737593174, 'timing/model': 1.069740281905979, 'learning_rate/qwen_vl_interface': 7.030563815372042e-06, 'learning_rate/action_model': 7.030563815372042e-05, 'epoch': 0.15}) +[2026-06-24 12:43:27,151][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.033065054565668106, 'train/grad_norm_pre_clip': 7.552535533905029, 'timing/data': 0.00034159189090132713, 'timing/model': 1.0698330490849912, 'learning_rate/qwen_vl_interface': 7.027015940644173e-06, 'learning_rate/action_model': 7.027015940644173e-05, 'epoch': 0.15}) +[2026-06-24 12:43:28,224][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.07239898294210434, 'train/grad_norm_pre_clip': 15.5662202835083, 'timing/data': 0.0003108629025518894, 'timing/model': 1.0716794119216502, 'learning_rate/qwen_vl_interface': 7.02346691282995e-06, 'learning_rate/action_model': 7.02346691282995e-05, 'epoch': 0.15}) +[2026-06-24 12:43:29,587][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.20322056114673615, 'train/grad_norm_pre_clip': 23.471529006958008, 'timing/data': 0.0003269040025770664, 'timing/model': 1.3609844059683383, 'learning_rate/qwen_vl_interface': 7.019916734232304e-06, 'learning_rate/action_model': 7.019916734232303e-05, 'epoch': 0.15}) +[2026-06-24 12:43:30,657][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.1847655326128006, 'train/grad_norm_pre_clip': 30.717119216918945, 'timing/data': 0.0003228159621357918, 'timing/model': 1.0683690158184618, 'learning_rate/qwen_vl_interface': 7.0163654071549035e-06, 'learning_rate/action_model': 7.016365407154904e-05, 'epoch': 0.15}) +[2026-06-24 12:43:31,729][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.021149015054106712, 'train/grad_norm_pre_clip': 8.665494918823242, 'timing/data': 0.00031264591962099075, 'timing/model': 1.0705576911568642, 'learning_rate/qwen_vl_interface': 7.012812933902167e-06, 'learning_rate/action_model': 7.012812933902167e-05, 'epoch': 0.15}) +[2026-06-24 12:43:32,801][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.0911276638507843, 'train/grad_norm_pre_clip': 21.735036849975586, 'timing/data': 0.0003190981224179268, 'timing/model': 1.0705104020889848, 'learning_rate/qwen_vl_interface': 7.009259316779258e-06, 'learning_rate/action_model': 7.009259316779257e-05, 'epoch': 0.15}) +[2026-06-24 12:43:33,872][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.13789412379264832, 'train/grad_norm_pre_clip': 32.38533020019531, 'timing/data': 0.00029807910323143005, 'timing/model': 1.0694807399995625, 'learning_rate/qwen_vl_interface': 7.005704558092078e-06, 'learning_rate/action_model': 7.005704558092078e-05, 'epoch': 0.15}) +[2026-06-24 12:43:34,945][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.031593725085258484, 'train/grad_norm_pre_clip': 8.283716201782227, 'timing/data': 0.0004971078597009182, 'timing/model': 1.071003373945132, 'learning_rate/qwen_vl_interface': 7.002148660147271e-06, 'learning_rate/action_model': 7.002148660147271e-05, 'epoch': 0.15}) +[2026-06-24 12:43:36,027][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.06367301940917969, 'train/grad_norm_pre_clip': 20.110960006713867, 'timing/data': 0.0004391670227050781, 'timing/model': 1.0797439590096474, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.15}) +[2026-06-24 12:43:37,098][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.2779647707939148, 'train/grad_norm_pre_clip': 53.421234130859375, 'timing/data': 0.0003050209488719702, 'timing/model': 1.069610157981515, 'learning_rate/qwen_vl_interface': 6.99503345571505e-06, 'learning_rate/action_model': 6.99503345571505e-05, 'epoch': 0.15}) +[2026-06-24 12:43:38,171][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.034496452659368515, 'train/grad_norm_pre_clip': 13.305184364318848, 'timing/data': 0.0003064139746129513, 'timing/model': 1.0712127350270748, 'learning_rate/qwen_vl_interface': 6.991474153844614e-06, 'learning_rate/action_model': 6.991474153844613e-05, 'epoch': 0.15}) +[2026-06-24 12:43:39,245][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.23228929936885834, 'train/grad_norm_pre_clip': 71.3525161743164, 'timing/data': 0.0003427029587328434, 'timing/model': 1.0729548789095134, 'learning_rate/qwen_vl_interface': 6.98791372195051e-06, 'learning_rate/action_model': 6.98791372195051e-05, 'epoch': 0.15}) +[2026-06-24 12:43:40,320][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.07218465954065323, 'train/grad_norm_pre_clip': 43.55899429321289, 'timing/data': 0.0003088591620326042, 'timing/model': 1.073190350085497, 'learning_rate/qwen_vl_interface': 6.984352162343056e-06, 'learning_rate/action_model': 6.984352162343057e-05, 'epoch': 0.15}) +[2026-06-24 12:43:41,391][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.34741562604904175, 'train/grad_norm_pre_clip': 68.07103729248047, 'timing/data': 0.00034736283123493195, 'timing/model': 1.069719335064292, 'learning_rate/qwen_vl_interface': 6.980789477333316e-06, 'learning_rate/action_model': 6.980789477333316e-05, 'epoch': 0.15}) +[2026-06-24 12:43:42,463][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.2165936529636383, 'train/grad_norm_pre_clip': 45.169681549072266, 'timing/data': 0.0003246990963816643, 'timing/model': 1.0699961259961128, 'learning_rate/qwen_vl_interface': 6.977225669233074e-06, 'learning_rate/action_model': 6.977225669233074e-05, 'epoch': 0.15}) +[2026-06-24 12:43:43,533][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.03057817928493023, 'train/grad_norm_pre_clip': 15.225661277770996, 'timing/data': 0.0003283759579062462, 'timing/model': 1.0685655509587377, 'learning_rate/qwen_vl_interface': 6.973660740354848e-06, 'learning_rate/action_model': 6.973660740354848e-05, 'epoch': 0.15}) +[2026-06-24 12:43:44,606][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.2870785593986511, 'train/grad_norm_pre_clip': 55.518096923828125, 'timing/data': 0.0004940319340676069, 'timing/model': 1.0709441669750959, 'learning_rate/qwen_vl_interface': 6.970094693011881e-06, 'learning_rate/action_model': 6.97009469301188e-05, 'epoch': 0.15}) +[2026-06-24 12:43:45,677][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.16161490976810455, 'train/grad_norm_pre_clip': 54.809654235839844, 'timing/data': 0.00035701110027730465, 'timing/model': 1.0700009630527347, 'learning_rate/qwen_vl_interface': 6.966527529518142e-06, 'learning_rate/action_model': 6.966527529518142e-05, 'epoch': 0.15}) +[2026-06-24 12:43:47,033][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.08709409087896347, 'train/grad_norm_pre_clip': 27.60201072692871, 'timing/data': 0.00033007911406457424, 'timing/model': 1.354112931061536, 'learning_rate/qwen_vl_interface': 6.962959252188325e-06, 'learning_rate/action_model': 6.962959252188325e-05, 'epoch': 0.15}) +[2026-06-24 12:43:48,107][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.06543245166540146, 'train/grad_norm_pre_clip': 45.07911682128906, 'timing/data': 0.0006080989260226488, 'timing/model': 1.0722395181655884, 'learning_rate/qwen_vl_interface': 6.9593898633378476e-06, 'learning_rate/action_model': 6.959389863337847e-05, 'epoch': 0.15}) +[2026-06-24 12:43:49,183][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.14960147440433502, 'train/grad_norm_pre_clip': 19.477550506591797, 'timing/data': 0.0003218441270291805, 'timing/model': 1.0742759120184928, 'learning_rate/qwen_vl_interface': 6.955819365282846e-06, 'learning_rate/action_model': 6.955819365282845e-05, 'epoch': 0.15}) +[2026-06-24 12:43:50,253][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.26323026418685913, 'train/grad_norm_pre_clip': 55.365840911865234, 'timing/data': 0.0003121839836239815, 'timing/model': 1.0682974110823125, 'learning_rate/qwen_vl_interface': 6.952247760340176e-06, 'learning_rate/action_model': 6.952247760340176e-05, 'epoch': 0.15}) +[2026-06-24 12:43:51,328][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.020055092871189117, 'train/grad_norm_pre_clip': 16.371994018554688, 'timing/data': 0.000329338014125824, 'timing/model': 1.0723474028054625, 'learning_rate/qwen_vl_interface': 6.948675050827418e-06, 'learning_rate/action_model': 6.948675050827419e-05, 'epoch': 0.15}) +[2026-06-24 12:43:52,397][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.014237014576792717, 'train/grad_norm_pre_clip': 7.735176086425781, 'timing/data': 0.0003092291299253702, 'timing/model': 1.0681139188818634, 'learning_rate/qwen_vl_interface': 6.945101239062862e-06, 'learning_rate/action_model': 6.945101239062863e-05, 'epoch': 0.15}) +[2026-06-24 12:43:53,472][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.12371399998664856, 'train/grad_norm_pre_clip': 58.35496520996094, 'timing/data': 0.0003134568687528372, 'timing/model': 1.0727780289016664, 'learning_rate/qwen_vl_interface': 6.941526327365513e-06, 'learning_rate/action_model': 6.941526327365513e-05, 'epoch': 0.15}) +[2026-06-24 12:43:54,543][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.07434216886758804, 'train/grad_norm_pre_clip': 32.94573974609375, 'timing/data': 0.00031424895860254765, 'timing/model': 1.0694801600184292, 'learning_rate/qwen_vl_interface': 6.937950318055098e-06, 'learning_rate/action_model': 6.937950318055098e-05, 'epoch': 0.15}) +[2026-06-24 12:43:55,617][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.09217221289873123, 'train/grad_norm_pre_clip': 46.62228012084961, 'timing/data': 0.00031335814855992794, 'timing/model': 1.072165380930528, 'learning_rate/qwen_vl_interface': 6.934373213452045e-06, 'learning_rate/action_model': 6.934373213452044e-05, 'epoch': 0.15}) +[2026-06-24 12:43:56,692][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.25446614623069763, 'train/grad_norm_pre_clip': 84.84148406982422, 'timing/data': 0.0003530129324644804, 'timing/model': 1.0740828581620008, 'learning_rate/qwen_vl_interface': 6.930795015877502e-06, 'learning_rate/action_model': 6.930795015877502e-05, 'epoch': 0.15}) +[2026-06-24 12:43:57,767][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.1316552311182022, 'train/grad_norm_pre_clip': 41.52439498901367, 'timing/data': 0.0003016951959580183, 'timing/model': 1.0730311491061002, 'learning_rate/qwen_vl_interface': 6.927215727653321e-06, 'learning_rate/action_model': 6.927215727653321e-05, 'epoch': 0.16}) +[2026-06-24 12:43:58,837][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.6555709838867188, 'train/grad_norm_pre_clip': 190.3420867919922, 'timing/data': 0.0003143788781017065, 'timing/model': 1.068416679976508, 'learning_rate/qwen_vl_interface': 6.9236353511020625e-06, 'learning_rate/action_model': 6.923635351102063e-05, 'epoch': 0.16}) +[2026-06-24 12:43:59,911][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.41585594415664673, 'train/grad_norm_pre_clip': 55.81082534790039, 'timing/data': 0.0003316218499094248, 'timing/model': 1.072841424960643, 'learning_rate/qwen_vl_interface': 6.920053888546995e-06, 'learning_rate/action_model': 6.920053888546994e-05, 'epoch': 0.16}) +[2026-06-24 12:44:00,983][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.060030099004507065, 'train/grad_norm_pre_clip': 36.952754974365234, 'timing/data': 0.00031331717036664486, 'timing/model': 1.069893943844363, 'learning_rate/qwen_vl_interface': 6.916471342312093e-06, 'learning_rate/action_model': 6.916471342312093e-05, 'epoch': 0.16}) +[2026-06-24 12:44:02,057][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.26927801966667175, 'train/grad_norm_pre_clip': 47.39739990234375, 'timing/data': 0.00032418686896562576, 'timing/model': 1.0722234640270472, 'learning_rate/qwen_vl_interface': 6.9128877147220274e-06, 'learning_rate/action_model': 6.912887714722027e-05, 'epoch': 0.16}) +[2026-06-24 12:44:03,126][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.13150617480278015, 'train/grad_norm_pre_clip': 33.28643798828125, 'timing/data': 0.00033692200668156147, 'timing/model': 1.0678962240926921, 'learning_rate/qwen_vl_interface': 6.909303008102179e-06, 'learning_rate/action_model': 6.909303008102179e-05, 'epoch': 0.16}) +[2026-06-24 12:44:04,487][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.026523372158408165, 'train/grad_norm_pre_clip': 6.995090484619141, 'timing/data': 0.00029874988831579685, 'timing/model': 1.3594171509612352, 'learning_rate/qwen_vl_interface': 6.905717224778623e-06, 'learning_rate/action_model': 6.905717224778622e-05, 'epoch': 0.16}) +[2026-06-24 12:44:05,559][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.1401459127664566, 'train/grad_norm_pre_clip': 33.525978088378906, 'timing/data': 0.0003075660206377506, 'timing/model': 1.0702322809956968, 'learning_rate/qwen_vl_interface': 6.902130367078135e-06, 'learning_rate/action_model': 6.902130367078134e-05, 'epoch': 0.16}) +[2026-06-24 12:44:06,636][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.050257399678230286, 'train/grad_norm_pre_clip': 21.68288230895996, 'timing/data': 0.00031115300953388214, 'timing/model': 1.0754065429791808, 'learning_rate/qwen_vl_interface': 6.8985424373281895e-06, 'learning_rate/action_model': 6.89854243732819e-05, 'epoch': 0.16}) +[2026-06-24 12:44:07,708][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.12955757975578308, 'train/grad_norm_pre_clip': 29.964797973632812, 'timing/data': 0.0003645149990916252, 'timing/model': 1.070502245100215, 'learning_rate/qwen_vl_interface': 6.894953437856954e-06, 'learning_rate/action_model': 6.894953437856953e-05, 'epoch': 0.16}) +[2026-06-24 12:44:08,781][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.12536528706550598, 'train/grad_norm_pre_clip': 57.30230712890625, 'timing/data': 0.0006445478647947311, 'timing/model': 1.0698733117897063, 'learning_rate/qwen_vl_interface': 6.891363370993292e-06, 'learning_rate/action_model': 6.891363370993293e-05, 'epoch': 0.16}) +[2026-06-24 12:44:09,854][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.16944141685962677, 'train/grad_norm_pre_clip': 41.54514694213867, 'timing/data': 0.0003045499324798584, 'timing/model': 1.0718938070349395, 'learning_rate/qwen_vl_interface': 6.88777223906676e-06, 'learning_rate/action_model': 6.88777223906676e-05, 'epoch': 0.16}) +[2026-06-24 12:44:10,927][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.1400446891784668, 'train/grad_norm_pre_clip': 32.05331039428711, 'timing/data': 0.0003467509523034096, 'timing/model': 1.0710477109532803, 'learning_rate/qwen_vl_interface': 6.8841800444076e-06, 'learning_rate/action_model': 6.8841800444076e-05, 'epoch': 0.16}) +[2026-06-24 12:44:11,999][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.22001191973686218, 'train/grad_norm_pre_clip': 36.80650329589844, 'timing/data': 0.0003416410181671381, 'timing/model': 1.0712320879101753, 'learning_rate/qwen_vl_interface': 6.8805867893467525e-06, 'learning_rate/action_model': 6.880586789346752e-05, 'epoch': 0.16}) +[2026-06-24 12:44:13,073][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.2491873949766159, 'train/grad_norm_pre_clip': 31.697416305541992, 'timing/data': 0.0003354188520461321, 'timing/model': 1.071520108031109, 'learning_rate/qwen_vl_interface': 6.87699247621584e-06, 'learning_rate/action_model': 6.87699247621584e-05, 'epoch': 0.16}) +[2026-06-24 12:44:14,146][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.08547130227088928, 'train/grad_norm_pre_clip': 41.08622360229492, 'timing/data': 0.0003133069258183241, 'timing/model': 1.071331816026941, 'learning_rate/qwen_vl_interface': 6.873397107347176e-06, 'learning_rate/action_model': 6.873397107347176e-05, 'epoch': 0.16}) +[2026-06-24 12:44:15,219][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.1093805804848671, 'train/grad_norm_pre_clip': 31.2191104888916, 'timing/data': 0.0003034980036318302, 'timing/model': 1.07217085105367, 'learning_rate/qwen_vl_interface': 6.869800685073752e-06, 'learning_rate/action_model': 6.869800685073752e-05, 'epoch': 0.16}) +[2026-06-24 12:44:16,291][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.0746259093284607, 'train/grad_norm_pre_clip': 24.287181854248047, 'timing/data': 0.00031719403341412544, 'timing/model': 1.0700835401657969, 'learning_rate/qwen_vl_interface': 6.86620321172925e-06, 'learning_rate/action_model': 6.86620321172925e-05, 'epoch': 0.16}) +[2026-06-24 12:44:17,361][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.1047908142209053, 'train/grad_norm_pre_clip': 32.23411560058594, 'timing/data': 0.00033124093897640705, 'timing/model': 1.0683330229949206, 'learning_rate/qwen_vl_interface': 6.862604689648029e-06, 'learning_rate/action_model': 6.862604689648029e-05, 'epoch': 0.16}) +[2026-06-24 12:44:18,430][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.09411781281232834, 'train/grad_norm_pre_clip': 20.010589599609375, 'timing/data': 0.0003177151083946228, 'timing/model': 1.067465967964381, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.16}) +[2026-06-24 12:44:19,502][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.06647708266973495, 'train/grad_norm_pre_clip': 17.36174201965332, 'timing/data': 0.0003130261320620775, 'timing/model': 1.0706649471540004, 'learning_rate/qwen_vl_interface': 6.855404508616285e-06, 'learning_rate/action_model': 6.855404508616285e-05, 'epoch': 0.16}) +[2026-06-24 12:44:20,575][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.15451399981975555, 'train/grad_norm_pre_clip': 26.043115615844727, 'timing/data': 0.0003104209899902344, 'timing/model': 1.0712474829051644, 'learning_rate/qwen_vl_interface': 6.851802854337881e-06, 'learning_rate/action_model': 6.851802854337881e-05, 'epoch': 0.16}) +[2026-06-24 12:44:21,947][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.13126863539218903, 'train/grad_norm_pre_clip': 27.867124557495117, 'timing/data': 0.000332011841237545, 'timing/model': 1.3699247550684959, 'learning_rate/qwen_vl_interface': 6.848200160666993e-06, 'learning_rate/action_model': 6.848200160666993e-05, 'epoch': 0.16}) +[2026-06-24 12:44:23,020][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.16414277255535126, 'train/grad_norm_pre_clip': 68.6563720703125, 'timing/data': 0.0003075660206377506, 'timing/model': 1.071356839966029, 'learning_rate/qwen_vl_interface': 6.844596429941374e-06, 'learning_rate/action_model': 6.844596429941373e-05, 'epoch': 0.16}) +[2026-06-24 12:44:24,094][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.24964064359664917, 'train/grad_norm_pre_clip': 64.95162200927734, 'timing/data': 0.0003354090731590986, 'timing/model': 1.0725159919820726, 'learning_rate/qwen_vl_interface': 6.8409916644994425e-06, 'learning_rate/action_model': 6.840991664499442e-05, 'epoch': 0.16}) +[2026-06-24 12:44:25,166][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.06736817955970764, 'train/grad_norm_pre_clip': 15.833479881286621, 'timing/data': 0.0003137271851301193, 'timing/model': 1.070467849029228, 'learning_rate/qwen_vl_interface': 6.837385866680294e-06, 'learning_rate/action_model': 6.837385866680293e-05, 'epoch': 0.16}) +[2026-06-24 12:44:26,236][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.2418115735054016, 'train/grad_norm_pre_clip': 49.46349334716797, 'timing/data': 0.00032206298783421516, 'timing/model': 1.068664373131469, 'learning_rate/qwen_vl_interface': 6.833779038823691e-06, 'learning_rate/action_model': 6.833779038823691e-05, 'epoch': 0.16}) +[2026-06-24 12:44:27,309][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.034387022256851196, 'train/grad_norm_pre_clip': 10.501338958740234, 'timing/data': 0.0003392361104488373, 'timing/model': 1.0709009091369808, 'learning_rate/qwen_vl_interface': 6.830171183270063e-06, 'learning_rate/action_model': 6.830171183270064e-05, 'epoch': 0.16}) +[2026-06-24 12:44:28,380][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.03990379348397255, 'train/grad_norm_pre_clip': 14.354795455932617, 'timing/data': 0.00031314603984355927, 'timing/model': 1.0703103709965944, 'learning_rate/qwen_vl_interface': 6.826562302360511e-06, 'learning_rate/action_model': 6.82656230236051e-05, 'epoch': 0.16}) +[2026-06-24 12:44:29,452][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.042161282151937485, 'train/grad_norm_pre_clip': 11.564494132995605, 'timing/data': 0.00029844907112419605, 'timing/model': 1.0697083021514118, 'learning_rate/qwen_vl_interface': 6.822952398436797e-06, 'learning_rate/action_model': 6.822952398436797e-05, 'epoch': 0.16}) +[2026-06-24 12:44:30,527][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.043899357318878174, 'train/grad_norm_pre_clip': 25.940387725830078, 'timing/data': 0.0003250890877097845, 'timing/model': 1.0736066619865596, 'learning_rate/qwen_vl_interface': 6.81934147384135e-06, 'learning_rate/action_model': 6.81934147384135e-05, 'epoch': 0.16}) +[2026-06-24 12:44:31,600][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.052042122930288315, 'train/grad_norm_pre_clip': 19.77656364440918, 'timing/data': 0.00032296497374773026, 'timing/model': 1.071316699963063, 'learning_rate/qwen_vl_interface': 6.815729530917257e-06, 'learning_rate/action_model': 6.815729530917256e-05, 'epoch': 0.16}) +[2026-06-24 12:44:32,675][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.10216441750526428, 'train/grad_norm_pre_clip': 31.30968475341797, 'timing/data': 0.00031013088300824165, 'timing/model': 1.0733688247855753, 'learning_rate/qwen_vl_interface': 6.812116572008271e-06, 'learning_rate/action_model': 6.812116572008271e-05, 'epoch': 0.16}) +[2026-06-24 12:44:33,749][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.02778562344610691, 'train/grad_norm_pre_clip': 14.916078567504883, 'timing/data': 0.00031404802575707436, 'timing/model': 1.072727401042357, 'learning_rate/qwen_vl_interface': 6.808502599458803e-06, 'learning_rate/action_model': 6.808502599458803e-05, 'epoch': 0.16}) +[2026-06-24 12:44:34,824][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.2403527945280075, 'train/grad_norm_pre_clip': 42.1036262512207, 'timing/data': 0.00030780606903135777, 'timing/model': 1.072489745914936, 'learning_rate/qwen_vl_interface': 6.8048876156139154e-06, 'learning_rate/action_model': 6.804887615613915e-05, 'epoch': 0.16}) +[2026-06-24 12:44:35,893][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.02758195996284485, 'train/grad_norm_pre_clip': 62.398433685302734, 'timing/data': 0.0003007519990205765, 'timing/model': 1.0678425068035722, 'learning_rate/qwen_vl_interface': 6.801271622819337e-06, 'learning_rate/action_model': 6.801271622819336e-05, 'epoch': 0.16}) +[2026-06-24 12:44:36,962][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.03484756872057915, 'train/grad_norm_pre_clip': 24.20058250427246, 'timing/data': 0.00031143310479819775, 'timing/model': 1.0673419539816678, 'learning_rate/qwen_vl_interface': 6.797654623421443e-06, 'learning_rate/action_model': 6.797654623421444e-05, 'epoch': 0.16}) +[2026-06-24 12:44:38,034][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.10593562573194504, 'train/grad_norm_pre_clip': 51.33216857910156, 'timing/data': 0.0003183360677212477, 'timing/model': 1.0704928589984775, 'learning_rate/qwen_vl_interface': 6.794036619767269e-06, 'learning_rate/action_model': 6.79403661976727e-05, 'epoch': 0.16}) +[2026-06-24 12:44:39,397][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.2257765233516693, 'train/grad_norm_pre_clip': 96.60027313232422, 'timing/data': 0.00042065116576850414, 'timing/model': 1.3608146200422198, 'learning_rate/qwen_vl_interface': 6.790417614204494e-06, 'learning_rate/action_model': 6.790417614204494e-05, 'epoch': 0.16}) +[2026-06-24 12:44:40,469][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.07034216076135635, 'train/grad_norm_pre_clip': 43.578617095947266, 'timing/data': 0.00031882687471807003, 'timing/model': 1.0713871028274298, 'learning_rate/qwen_vl_interface': 6.786797609081453e-06, 'learning_rate/action_model': 6.786797609081453e-05, 'epoch': 0.16}) +[2026-06-24 12:44:41,544][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.17343930900096893, 'train/grad_norm_pre_clip': 79.58761596679688, 'timing/data': 0.0003064628690481186, 'timing/model': 1.0732944628689438, 'learning_rate/qwen_vl_interface': 6.783176606747129e-06, 'learning_rate/action_model': 6.78317660674713e-05, 'epoch': 0.16}) +[2026-06-24 12:44:42,619][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.12051212042570114, 'train/grad_norm_pre_clip': 44.39561080932617, 'timing/data': 0.0003128950484097004, 'timing/model': 1.0729054710827768, 'learning_rate/qwen_vl_interface': 6.77955460955115e-06, 'learning_rate/action_model': 6.77955460955115e-05, 'epoch': 0.16}) +[2026-06-24 12:44:43,690][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.010725053027272224, 'train/grad_norm_pre_clip': 11.982358932495117, 'timing/data': 0.00034178001806139946, 'timing/model': 1.0693317078985274, 'learning_rate/qwen_vl_interface': 6.775931619843791e-06, 'learning_rate/action_model': 6.77593161984379e-05, 'epoch': 0.16}) +[2026-06-24 12:44:44,763][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.21806740760803223, 'train/grad_norm_pre_clip': 46.507686614990234, 'timing/data': 0.00030317693017423153, 'timing/model': 1.0716533961240202, 'learning_rate/qwen_vl_interface': 6.7723076399759715e-06, 'learning_rate/action_model': 6.772307639975972e-05, 'epoch': 0.16}) +[2026-06-24 12:44:45,836][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.15450017154216766, 'train/grad_norm_pre_clip': 45.33943557739258, 'timing/data': 0.0003231549635529518, 'timing/model': 1.0715513511095196, 'learning_rate/qwen_vl_interface': 6.7686826722992516e-06, 'learning_rate/action_model': 6.768682672299252e-05, 'epoch': 0.16}) +[2026-06-24 12:44:46,908][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.11425703763961792, 'train/grad_norm_pre_clip': 66.3840560913086, 'timing/data': 0.00030204514041543007, 'timing/model': 1.070642810082063, 'learning_rate/qwen_vl_interface': 6.765056719165833e-06, 'learning_rate/action_model': 6.765056719165834e-05, 'epoch': 0.16}) +[2026-06-24 12:44:47,981][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.1727869212627411, 'train/grad_norm_pre_clip': 97.4566421508789, 'timing/data': 0.00030163396149873734, 'timing/model': 1.0709662961307913, 'learning_rate/qwen_vl_interface': 6.761429782928558e-06, 'learning_rate/action_model': 6.761429782928557e-05, 'epoch': 0.16}) +[2026-06-24 12:44:49,057][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.42612817883491516, 'train/grad_norm_pre_clip': 84.55294036865234, 'timing/data': 0.0003028460778295994, 'timing/model': 1.0740323779173195, 'learning_rate/qwen_vl_interface': 6.757801865940908e-06, 'learning_rate/action_model': 6.757801865940908e-05, 'epoch': 0.16}) +[2026-06-24 12:44:50,132][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.045728687196969986, 'train/grad_norm_pre_clip': 24.670398712158203, 'timing/data': 0.0003300989046692848, 'timing/model': 1.0741780530661345, 'learning_rate/qwen_vl_interface': 6.754172970556996e-06, 'learning_rate/action_model': 6.754172970556996e-05, 'epoch': 0.16}) +[2026-06-24 12:44:51,206][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.1305045187473297, 'train/grad_norm_pre_clip': 36.78901672363281, 'timing/data': 0.0005067349411547184, 'timing/model': 1.071207941044122, 'learning_rate/qwen_vl_interface': 6.750543099131573e-06, 'learning_rate/action_model': 6.750543099131573e-05, 'epoch': 0.16}) +[2026-06-24 12:44:52,281][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.16695871949195862, 'train/grad_norm_pre_clip': 56.94407272338867, 'timing/data': 0.00032226298935711384, 'timing/model': 1.0738000930286944, 'learning_rate/qwen_vl_interface': 6.746912254020025e-06, 'learning_rate/action_model': 6.746912254020025e-05, 'epoch': 0.16}) +[2026-06-24 12:44:53,353][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.05537768453359604, 'train/grad_norm_pre_clip': 29.650121688842773, 'timing/data': 0.0003864448517560959, 'timing/model': 1.0705028190277517, 'learning_rate/qwen_vl_interface': 6.743280437578366e-06, 'learning_rate/action_model': 6.743280437578366e-05, 'epoch': 0.16}) +[2026-06-24 12:44:54,423][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.4254101812839508, 'train/grad_norm_pre_clip': 55.64680862426758, 'timing/data': 0.000346318818628788, 'timing/model': 1.0682475678622723, 'learning_rate/qwen_vl_interface': 6.7396476521632456e-06, 'learning_rate/action_model': 6.739647652163246e-05, 'epoch': 0.16}) +[2026-06-24 12:44:55,499][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.08651547878980637, 'train/grad_norm_pre_clip': 13.9679536819458, 'timing/data': 0.0003113420680165291, 'timing/model': 1.0738684630487114, 'learning_rate/qwen_vl_interface': 6.736013900131934e-06, 'learning_rate/action_model': 6.736013900131934e-05, 'epoch': 0.16}) +[2026-06-24 12:44:56,856][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.060560423880815506, 'train/grad_norm_pre_clip': 116.5005111694336, 'timing/data': 0.0003139078617095947, 'timing/model': 1.3561858190223575, 'learning_rate/qwen_vl_interface': 6.732379183842339e-06, 'learning_rate/action_model': 6.73237918384234e-05, 'epoch': 0.16}) +[2026-06-24 12:44:57,934][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.16769437491893768, 'train/grad_norm_pre_clip': 50.79224395751953, 'timing/data': 0.0004968950524926186, 'timing/model': 1.0759461331181228, 'learning_rate/qwen_vl_interface': 6.728743505652984e-06, 'learning_rate/action_model': 6.728743505652983e-05, 'epoch': 0.16}) +[2026-06-24 12:44:59,028][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.15444184839725494, 'train/grad_norm_pre_clip': 37.33211135864258, 'timing/data': 0.0003364000003784895, 'timing/model': 1.0922473778482527, 'learning_rate/qwen_vl_interface': 6.725106867923023e-06, 'learning_rate/action_model': 6.725106867923023e-05, 'epoch': 0.16}) +[2026-06-24 12:45:00,102][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.01870078407227993, 'train/grad_norm_pre_clip': 10.589943885803223, 'timing/data': 0.00030805589631199837, 'timing/model': 1.0724491372238845, 'learning_rate/qwen_vl_interface': 6.721469273012229e-06, 'learning_rate/action_model': 6.721469273012229e-05, 'epoch': 0.16}) +[2026-06-24 12:45:01,180][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.01270626112818718, 'train/grad_norm_pre_clip': 3.811807870864868, 'timing/data': 0.00047605601139366627, 'timing/model': 1.0759552179370075, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.16}) +[2026-06-24 12:45:02,253][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.33712220191955566, 'train/grad_norm_pre_clip': 39.553550720214844, 'timing/data': 0.0005312599241733551, 'timing/model': 1.0707851520273834, 'learning_rate/qwen_vl_interface': 6.714191221090352e-06, 'learning_rate/action_model': 6.714191221090352e-05, 'epoch': 0.16}) +[2026-06-24 12:45:03,327][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.03280579298734665, 'train/grad_norm_pre_clip': 18.296131134033203, 'timing/data': 0.0005666979122906923, 'timing/model': 1.072626214940101, 'learning_rate/qwen_vl_interface': 6.710550768801916e-06, 'learning_rate/action_model': 6.710550768801915e-05, 'epoch': 0.16}) +[2026-06-24 12:45:04,400][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.2306131273508072, 'train/grad_norm_pre_clip': 28.84542465209961, 'timing/data': 0.0005881588440388441, 'timing/model': 1.0706460191868246, 'learning_rate/qwen_vl_interface': 6.706909368777944e-06, 'learning_rate/action_model': 6.706909368777944e-05, 'epoch': 0.16}) +[2026-06-24 12:45:05,473][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.17119379341602325, 'train/grad_norm_pre_clip': 101.62831115722656, 'timing/data': 0.0003173339646309614, 'timing/model': 1.071295005036518, 'learning_rate/qwen_vl_interface': 6.703267023381299e-06, 'learning_rate/action_model': 6.7032670233813e-05, 'epoch': 0.16}) +[2026-06-24 12:45:06,545][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.08277231454849243, 'train/grad_norm_pre_clip': 19.596614837646484, 'timing/data': 0.00033975602127611637, 'timing/model': 1.0702739439439029, 'learning_rate/qwen_vl_interface': 6.69962373497546e-06, 'learning_rate/action_model': 6.69962373497546e-05, 'epoch': 0.16}) +[2026-06-24 12:45:07,617][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.16680000722408295, 'train/grad_norm_pre_clip': 47.39717483520508, 'timing/data': 0.0003090379759669304, 'timing/model': 1.0703800071496516, 'learning_rate/qwen_vl_interface': 6.69597950592452e-06, 'learning_rate/action_model': 6.69597950592452e-05, 'epoch': 0.16}) +[2026-06-24 12:45:08,719][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.06271396577358246, 'train/grad_norm_pre_clip': 21.611652374267578, 'timing/data': 0.0003206399269402027, 'timing/model': 1.0997992400079966, 'learning_rate/qwen_vl_interface': 6.692334338593175e-06, 'learning_rate/action_model': 6.692334338593175e-05, 'epoch': 0.16}) +[2026-06-24 12:45:09,796][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.015968116000294685, 'train/grad_norm_pre_clip': 8.289604187011719, 'timing/data': 0.00032356614246964455, 'timing/model': 1.0757400540169328, 'learning_rate/qwen_vl_interface': 6.688688235346736e-06, 'learning_rate/action_model': 6.688688235346736e-05, 'epoch': 0.16}) +[2026-06-24 12:45:10,871][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.08455345034599304, 'train/grad_norm_pre_clip': 26.215299606323242, 'timing/data': 0.0003339459653943777, 'timing/model': 1.0733119768556207, 'learning_rate/qwen_vl_interface': 6.685041198551125e-06, 'learning_rate/action_model': 6.685041198551124e-05, 'epoch': 0.16}) +[2026-06-24 12:45:11,947][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.10084078460931778, 'train/grad_norm_pre_clip': 36.39495086669922, 'timing/data': 0.00031020003370940685, 'timing/model': 1.0743112589698285, 'learning_rate/qwen_vl_interface': 6.681393230572857e-06, 'learning_rate/action_model': 6.681393230572856e-05, 'epoch': 0.16}) +[2026-06-24 12:45:13,022][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.18786203861236572, 'train/grad_norm_pre_clip': 39.216121673583984, 'timing/data': 0.0003480419982224703, 'timing/model': 1.0738475781399757, 'learning_rate/qwen_vl_interface': 6.677744333779063e-06, 'learning_rate/action_model': 6.677744333779062e-05, 'epoch': 0.16}) +[2026-06-24 12:45:14,390][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.16111040115356445, 'train/grad_norm_pre_clip': 42.30928421020508, 'timing/data': 0.0003072537947446108, 'timing/model': 1.3658583329524845, 'learning_rate/qwen_vl_interface': 6.674094510537471e-06, 'learning_rate/action_model': 6.674094510537471e-05, 'epoch': 0.16}) +[2026-06-24 12:45:15,467][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.009231620468199253, 'train/grad_norm_pre_clip': 4.667423248291016, 'timing/data': 0.00036908197216689587, 'timing/model': 1.076351708965376, 'learning_rate/qwen_vl_interface': 6.670443763216412e-06, 'learning_rate/action_model': 6.670443763216411e-05, 'epoch': 0.16}) +[2026-06-24 12:45:16,542][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.06737174093723297, 'train/grad_norm_pre_clip': 30.447309494018555, 'timing/data': 0.0003078458830714226, 'timing/model': 1.0729473461396992, 'learning_rate/qwen_vl_interface': 6.666792094184817e-06, 'learning_rate/action_model': 6.666792094184818e-05, 'epoch': 0.16}) +[2026-06-24 12:45:17,618][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.0950896292924881, 'train/grad_norm_pre_clip': 31.477155685424805, 'timing/data': 0.0003151698037981987, 'timing/model': 1.0743544239085168, 'learning_rate/qwen_vl_interface': 6.663139505812215e-06, 'learning_rate/action_model': 6.663139505812215e-05, 'epoch': 0.16}) +[2026-06-24 12:45:18,690][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.1969849318265915, 'train/grad_norm_pre_clip': 62.88675308227539, 'timing/data': 0.0003298982046544552, 'timing/model': 1.0706217368133366, 'learning_rate/qwen_vl_interface': 6.659486000468728e-06, 'learning_rate/action_model': 6.659486000468728e-05, 'epoch': 0.16}) +[2026-06-24 12:45:19,764][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.05168604850769043, 'train/grad_norm_pre_clip': 39.650230407714844, 'timing/data': 0.0003010921645909548, 'timing/model': 1.072068371111527, 'learning_rate/qwen_vl_interface': 6.6558315805250805e-06, 'learning_rate/action_model': 6.65583158052508e-05, 'epoch': 0.16}) +[2026-06-24 12:45:20,839][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.21284323930740356, 'train/grad_norm_pre_clip': 79.34381103515625, 'timing/data': 0.00030991900712251663, 'timing/model': 1.073507561115548, 'learning_rate/qwen_vl_interface': 6.652176248352581e-06, 'learning_rate/action_model': 6.65217624835258e-05, 'epoch': 0.16}) +[2026-06-24 12:45:21,914][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.19798941910266876, 'train/grad_norm_pre_clip': 52.16321563720703, 'timing/data': 0.0003401970025151968, 'timing/model': 1.0734325100202113, 'learning_rate/qwen_vl_interface': 6.64852000632314e-06, 'learning_rate/action_model': 6.648520006323139e-05, 'epoch': 0.16}) +[2026-06-24 12:45:22,988][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.2512091398239136, 'train/grad_norm_pre_clip': 40.582191467285156, 'timing/data': 0.0003021541051566601, 'timing/model': 1.0725531198550016, 'learning_rate/qwen_vl_interface': 6.644862856809247e-06, 'learning_rate/action_model': 6.644862856809247e-05, 'epoch': 0.16}) +[2026-06-24 12:45:24,064][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.2667001485824585, 'train/grad_norm_pre_clip': 67.5831069946289, 'timing/data': 0.00030789501033723354, 'timing/model': 1.074416683986783, 'learning_rate/qwen_vl_interface': 6.64120480218399e-06, 'learning_rate/action_model': 6.64120480218399e-05, 'epoch': 0.16}) +[2026-06-24 12:45:25,141][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.14586712419986725, 'train/grad_norm_pre_clip': 44.99705505371094, 'timing/data': 0.00031793396919965744, 'timing/model': 1.0755380289629102, 'learning_rate/qwen_vl_interface': 6.637545844821043e-06, 'learning_rate/action_model': 6.637545844821043e-05, 'epoch': 0.16}) +[2026-06-24 12:45:26,216][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.24761231243610382, 'train/grad_norm_pre_clip': 79.00896453857422, 'timing/data': 0.0003054209519177675, 'timing/model': 1.072926877066493, 'learning_rate/qwen_vl_interface': 6.633885987094661e-06, 'learning_rate/action_model': 6.633885987094661e-05, 'epoch': 0.16}) +[2026-06-24 12:45:27,286][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.015318759717047215, 'train/grad_norm_pre_clip': 12.960411071777344, 'timing/data': 0.0004919259808957577, 'timing/model': 1.068899200996384, 'learning_rate/qwen_vl_interface': 6.630225231379687e-06, 'learning_rate/action_model': 6.630225231379687e-05, 'epoch': 0.16}) +[2026-06-24 12:45:28,359][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.3473978042602539, 'train/grad_norm_pre_clip': 64.0374755859375, 'timing/data': 0.0003360991831868887, 'timing/model': 1.0705709320027381, 'learning_rate/qwen_vl_interface': 6.626563580051546e-06, 'learning_rate/action_model': 6.626563580051546e-05, 'epoch': 0.16}) +[2026-06-24 12:45:29,443][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.0205803532153368, 'train/grad_norm_pre_clip': 9.534653663635254, 'timing/data': 0.0005000410601496696, 'timing/model': 1.0822559539228678, 'learning_rate/qwen_vl_interface': 6.622901035486241e-06, 'learning_rate/action_model': 6.622901035486241e-05, 'epoch': 0.16}) +[2026-06-24 12:45:30,518][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.02044287510216236, 'train/grad_norm_pre_clip': 7.015081405639648, 'timing/data': 0.0005564577877521515, 'timing/model': 1.0731266781222075, 'learning_rate/qwen_vl_interface': 6.61923760006036e-06, 'learning_rate/action_model': 6.61923760006036e-05, 'epoch': 0.16}) +[2026-06-24 12:45:31,882][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.035691920667886734, 'train/grad_norm_pre_clip': 25.282197952270508, 'timing/data': 0.00041823508217930794, 'timing/model': 1.3626813529990613, 'learning_rate/qwen_vl_interface': 6.615573276151068e-06, 'learning_rate/action_model': 6.615573276151068e-05, 'epoch': 0.16}) +[2026-06-24 12:45:32,957][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.001428202842362225, 'train/grad_norm_pre_clip': 1.814279317855835, 'timing/data': 0.0003197879996150732, 'timing/model': 1.073569354834035, 'learning_rate/qwen_vl_interface': 6.6119080661361e-06, 'learning_rate/action_model': 6.6119080661361e-05, 'epoch': 0.16}) +[2026-06-24 12:45:34,030][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.1086367592215538, 'train/grad_norm_pre_clip': 46.028934478759766, 'timing/data': 0.00036481302231550217, 'timing/model': 1.0712522929534316, 'learning_rate/qwen_vl_interface': 6.608241972393774e-06, 'learning_rate/action_model': 6.608241972393774e-05, 'epoch': 0.16}) +[2026-06-24 12:45:35,104][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.1657242476940155, 'train/grad_norm_pre_clip': 37.065162658691406, 'timing/data': 0.00030383793637156487, 'timing/model': 1.0722768420819193, 'learning_rate/qwen_vl_interface': 6.60457499730298e-06, 'learning_rate/action_model': 6.60457499730298e-05, 'epoch': 0.16}) +[2026-06-24 12:45:36,176][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.35731106996536255, 'train/grad_norm_pre_clip': 77.96915435791016, 'timing/data': 0.00036136689595878124, 'timing/model': 1.0700312771368772, 'learning_rate/qwen_vl_interface': 6.600907143243176e-06, 'learning_rate/action_model': 6.600907143243176e-05, 'epoch': 0.16}) +[2026-06-24 12:45:37,242][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.23572444915771484, 'train/grad_norm_pre_clip': 49.174713134765625, 'timing/data': 0.0003120729234069586, 'timing/model': 1.0644432080443949, 'learning_rate/qwen_vl_interface': 6.5972384125943945e-06, 'learning_rate/action_model': 6.597238412594394e-05, 'epoch': 0.16}) +[2026-06-24 12:45:38,317][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.14497070014476776, 'train/grad_norm_pre_clip': 41.21073913574219, 'timing/data': 0.00031355605460703373, 'timing/model': 1.0733168520964682, 'learning_rate/qwen_vl_interface': 6.593568807737233e-06, 'learning_rate/action_model': 6.593568807737233e-05, 'epoch': 0.16}) +[2026-06-24 12:45:39,391][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.10534118860960007, 'train/grad_norm_pre_clip': 31.32435417175293, 'timing/data': 0.00033869408071041107, 'timing/model': 1.0725799140054733, 'learning_rate/qwen_vl_interface': 6.5898983310528584e-06, 'learning_rate/action_model': 6.589898331052858e-05, 'epoch': 0.16}) +[2026-06-24 12:45:40,462][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.05011451616883278, 'train/grad_norm_pre_clip': 19.552520751953125, 'timing/data': 0.0003052100073546171, 'timing/model': 1.0692941218148917, 'learning_rate/qwen_vl_interface': 6.586226984923004e-06, 'learning_rate/action_model': 6.586226984923004e-05, 'epoch': 0.16}) +[2026-06-24 12:45:41,539][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.08959399163722992, 'train/grad_norm_pre_clip': 56.62076187133789, 'timing/data': 0.0003886180929839611, 'timing/model': 1.0758091111201793, 'learning_rate/qwen_vl_interface': 6.582554771729967e-06, 'learning_rate/action_model': 6.582554771729966e-05, 'epoch': 0.16}) +[2026-06-24 12:45:42,615][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.016445841640233994, 'train/grad_norm_pre_clip': 19.863950729370117, 'timing/data': 0.0003149688709527254, 'timing/model': 1.074363216990605, 'learning_rate/qwen_vl_interface': 6.578881693856605e-06, 'learning_rate/action_model': 6.578881693856604e-05, 'epoch': 0.16}) +[2026-06-24 12:45:43,689][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.33614540100097656, 'train/grad_norm_pre_clip': 83.48368072509766, 'timing/data': 0.000348531873896718, 'timing/model': 1.0717557549942285, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.16}) +[2026-06-24 12:45:44,760][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.10996412485837936, 'train/grad_norm_pre_clip': 33.25243377685547, 'timing/data': 0.0003140370827168226, 'timing/model': 1.0693261621054262, 'learning_rate/qwen_vl_interface': 6.571532953603151e-06, 'learning_rate/action_model': 6.571532953603151e-05, 'epoch': 0.16}) +[2026-06-24 12:45:45,834][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.11365398019552231, 'train/grad_norm_pre_clip': 42.65139389038086, 'timing/data': 0.00043452600948512554, 'timing/model': 1.0718551941681653, 'learning_rate/qwen_vl_interface': 6.567857295991577e-06, 'learning_rate/action_model': 6.567857295991576e-05, 'epoch': 0.16}) +[2026-06-24 12:45:46,908][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.587183952331543, 'train/grad_norm_pre_clip': 60.52302551269531, 'timing/data': 0.0003429120406508446, 'timing/model': 1.0725253820419312, 'learning_rate/qwen_vl_interface': 6.56418078323671e-06, 'learning_rate/action_model': 6.56418078323671e-05, 'epoch': 0.16}) +[2026-06-24 12:45:47,980][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.0979505404829979, 'train/grad_norm_pre_clip': 54.87471008300781, 'timing/data': 0.0003259789664298296, 'timing/model': 1.070751957129687, 'learning_rate/qwen_vl_interface': 6.5605034177242e-06, 'learning_rate/action_model': 6.5605034177242e-05, 'epoch': 0.16}) +[2026-06-24 12:45:49,340][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.040154747664928436, 'train/grad_norm_pre_clip': 19.882965087890625, 'timing/data': 0.0003521388862282038, 'timing/model': 1.3581636098679155, 'learning_rate/qwen_vl_interface': 6.556825201840252e-06, 'learning_rate/action_model': 6.556825201840252e-05, 'epoch': 0.16}) +[2026-06-24 12:45:50,415][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.18789389729499817, 'train/grad_norm_pre_clip': 37.70463180541992, 'timing/data': 0.00031773396767675877, 'timing/model': 1.0734706190414727, 'learning_rate/qwen_vl_interface': 6.553146137971618e-06, 'learning_rate/action_model': 6.553146137971618e-05, 'epoch': 0.16}) +[2026-06-24 12:45:51,490][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.036916136741638184, 'train/grad_norm_pre_clip': 18.156906127929688, 'timing/data': 0.00030292593874037266, 'timing/model': 1.0728520390111953, 'learning_rate/qwen_vl_interface': 6.549466228505603e-06, 'learning_rate/action_model': 6.549466228505603e-05, 'epoch': 0.17}) +[2026-06-24 12:45:52,563][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.07148810476064682, 'train/grad_norm_pre_clip': 30.34939193725586, 'timing/data': 0.00031863595359027386, 'timing/model': 1.0717299620155245, 'learning_rate/qwen_vl_interface': 6.545785475830061e-06, 'learning_rate/action_model': 6.54578547583006e-05, 'epoch': 0.17}) +[2026-06-24 12:45:53,633][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.15089558064937592, 'train/grad_norm_pre_clip': 61.85044860839844, 'timing/data': 0.00031120097264647484, 'timing/model': 1.067842568969354, 'learning_rate/qwen_vl_interface': 6.54210388233339e-06, 'learning_rate/action_model': 6.542103882333391e-05, 'epoch': 0.17}) +[2026-06-24 12:45:54,705][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.10355556011199951, 'train/grad_norm_pre_clip': 23.04277992248535, 'timing/data': 0.00039322697557508945, 'timing/model': 1.070666842861101, 'learning_rate/qwen_vl_interface': 6.538421450404538e-06, 'learning_rate/action_model': 6.538421450404538e-05, 'epoch': 0.17}) +[2026-06-24 12:45:55,779][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.19332878291606903, 'train/grad_norm_pre_clip': 40.18158721923828, 'timing/data': 0.000307213980704546, 'timing/model': 1.0725467109587044, 'learning_rate/qwen_vl_interface': 6.534738182432999e-06, 'learning_rate/action_model': 6.534738182432999e-05, 'epoch': 0.17}) +[2026-06-24 12:45:56,857][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.10391939431428909, 'train/grad_norm_pre_clip': 25.670557022094727, 'timing/data': 0.0002939088735729456, 'timing/model': 1.0759310079738498, 'learning_rate/qwen_vl_interface': 6.531054080808798e-06, 'learning_rate/action_model': 6.531054080808798e-05, 'epoch': 0.17}) +[2026-06-24 12:45:57,931][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.07777518779039383, 'train/grad_norm_pre_clip': 17.181711196899414, 'timing/data': 0.00043166009709239006, 'timing/model': 1.0723983100615442, 'learning_rate/qwen_vl_interface': 6.5273691479225124e-06, 'learning_rate/action_model': 6.527369147922511e-05, 'epoch': 0.17}) +[2026-06-24 12:45:59,005][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.02014901489019394, 'train/grad_norm_pre_clip': 7.839004039764404, 'timing/data': 0.0003240660298615694, 'timing/model': 1.0722149370703846, 'learning_rate/qwen_vl_interface': 6.523683386165257e-06, 'learning_rate/action_model': 6.523683386165256e-05, 'epoch': 0.17}) +[2026-06-24 12:46:00,082][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.11925766617059708, 'train/grad_norm_pre_clip': 25.958051681518555, 'timing/data': 0.00033187982626259327, 'timing/model': 1.0749860650394112, 'learning_rate/qwen_vl_interface': 6.519996797928677e-06, 'learning_rate/action_model': 6.519996797928677e-05, 'epoch': 0.17}) +[2026-06-24 12:46:01,152][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.08128810673952103, 'train/grad_norm_pre_clip': 33.849700927734375, 'timing/data': 0.00032733194530010223, 'timing/model': 1.0686432200018317, 'learning_rate/qwen_vl_interface': 6.5163093856049664e-06, 'learning_rate/action_model': 6.516309385604967e-05, 'epoch': 0.17}) +[2026-06-24 12:46:02,225][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.10208380967378616, 'train/grad_norm_pre_clip': 41.53694152832031, 'timing/data': 0.00031451694667339325, 'timing/model': 1.0718456411268562, 'learning_rate/qwen_vl_interface': 6.512621151586842e-06, 'learning_rate/action_model': 6.512621151586842e-05, 'epoch': 0.17}) +[2026-06-24 12:46:03,300][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.0702686533331871, 'train/grad_norm_pre_clip': 23.41339874267578, 'timing/data': 0.00036240811459720135, 'timing/model': 1.0732522390317172, 'learning_rate/qwen_vl_interface': 6.508932098267561e-06, 'learning_rate/action_model': 6.50893209826756e-05, 'epoch': 0.17}) +[2026-06-24 12:46:04,373][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.047607265412807465, 'train/grad_norm_pre_clip': 21.70444107055664, 'timing/data': 0.00030620209872722626, 'timing/model': 1.0712686169426888, 'learning_rate/qwen_vl_interface': 6.505242228040908e-06, 'learning_rate/action_model': 6.505242228040907e-05, 'epoch': 0.17}) +[2026-06-24 12:46:05,449][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.08173317462205887, 'train/grad_norm_pre_clip': 29.818994522094727, 'timing/data': 0.00031791417859494686, 'timing/model': 1.0741812270134687, 'learning_rate/qwen_vl_interface': 6.501551543301203e-06, 'learning_rate/action_model': 6.501551543301203e-05, 'epoch': 0.17}) +[2026-06-24 12:46:06,808][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.1096728965640068, 'train/grad_norm_pre_clip': 56.129112243652344, 'timing/data': 0.00030715297907590866, 'timing/model': 1.3573162108659744, 'learning_rate/qwen_vl_interface': 6.497860046443288e-06, 'learning_rate/action_model': 6.497860046443288e-05, 'epoch': 0.17}) +[2026-06-24 12:46:07,881][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.03903348743915558, 'train/grad_norm_pre_clip': 15.592596054077148, 'timing/data': 0.0003528599627315998, 'timing/model': 1.0719088420737535, 'learning_rate/qwen_vl_interface': 6.494167739862538e-06, 'learning_rate/action_model': 6.494167739862538e-05, 'epoch': 0.17}) +[2026-06-24 12:46:08,955][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.18232882022857666, 'train/grad_norm_pre_clip': 23.37723159790039, 'timing/data': 0.0003957720473408699, 'timing/model': 1.0717234869953245, 'learning_rate/qwen_vl_interface': 6.490474625954849e-06, 'learning_rate/action_model': 6.490474625954849e-05, 'epoch': 0.17}) +[2026-06-24 12:46:10,028][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.1439375877380371, 'train/grad_norm_pre_clip': 42.71235275268555, 'timing/data': 0.00033138995058834553, 'timing/model': 1.0714310980401933, 'learning_rate/qwen_vl_interface': 6.486780707116642e-06, 'learning_rate/action_model': 6.486780707116642e-05, 'epoch': 0.17}) +[2026-06-24 12:46:11,097][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.013837690465152264, 'train/grad_norm_pre_clip': 6.6717000007629395, 'timing/data': 0.0003245358821004629, 'timing/model': 1.0678098029457033, 'learning_rate/qwen_vl_interface': 6.483085985744864e-06, 'learning_rate/action_model': 6.483085985744863e-05, 'epoch': 0.17}) +[2026-06-24 12:46:12,168][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.11075019836425781, 'train/grad_norm_pre_clip': 39.834991455078125, 'timing/data': 0.0003932979889214039, 'timing/model': 1.0691394379828125, 'learning_rate/qwen_vl_interface': 6.479390464236974e-06, 'learning_rate/action_model': 6.479390464236974e-05, 'epoch': 0.17}) +[2026-06-24 12:46:13,239][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.21759474277496338, 'train/grad_norm_pre_clip': 66.38135528564453, 'timing/data': 0.0003281128592789173, 'timing/model': 1.0693215099163353, 'learning_rate/qwen_vl_interface': 6.475694144990961e-06, 'learning_rate/action_model': 6.475694144990961e-05, 'epoch': 0.17}) +[2026-06-24 12:46:14,314][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.24284642934799194, 'train/grad_norm_pre_clip': 59.050296783447266, 'timing/data': 0.0004973940085619688, 'timing/model': 1.073327997000888, 'learning_rate/qwen_vl_interface': 6.471997030405326e-06, 'learning_rate/action_model': 6.471997030405325e-05, 'epoch': 0.17}) +[2026-06-24 12:46:15,386][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.06541036814451218, 'train/grad_norm_pre_clip': 37.70558547973633, 'timing/data': 0.0003588218241930008, 'timing/model': 1.0706175060477108, 'learning_rate/qwen_vl_interface': 6.468299122879082e-06, 'learning_rate/action_model': 6.468299122879081e-05, 'epoch': 0.17}) +[2026-06-24 12:46:16,459][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.03829784318804741, 'train/grad_norm_pre_clip': 22.542539596557617, 'timing/data': 0.0003220618236809969, 'timing/model': 1.0710996189154685, 'learning_rate/qwen_vl_interface': 6.464600424811765e-06, 'learning_rate/action_model': 6.464600424811764e-05, 'epoch': 0.17}) +[2026-06-24 12:46:17,534][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.20620545744895935, 'train/grad_norm_pre_clip': 69.47411346435547, 'timing/data': 0.0004979858640581369, 'timing/model': 1.072724532801658, 'learning_rate/qwen_vl_interface': 6.460900938603418e-06, 'learning_rate/action_model': 6.460900938603418e-05, 'epoch': 0.17}) +[2026-06-24 12:46:18,609][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.32486608624458313, 'train/grad_norm_pre_clip': 46.9524040222168, 'timing/data': 0.00033148983493447304, 'timing/model': 1.073724569985643, 'learning_rate/qwen_vl_interface': 6.457200666654601e-06, 'learning_rate/action_model': 6.457200666654601e-05, 'epoch': 0.17}) +[2026-06-24 12:46:19,679][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.026217032223939896, 'train/grad_norm_pre_clip': 18.128475189208984, 'timing/data': 0.0003140759654343128, 'timing/model': 1.0681741740554571, 'learning_rate/qwen_vl_interface': 6.453499611366376e-06, 'learning_rate/action_model': 6.453499611366376e-05, 'epoch': 0.17}) +[2026-06-24 12:46:20,751][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.38844290375709534, 'train/grad_norm_pre_clip': 90.61334228515625, 'timing/data': 0.0003110799007117748, 'timing/model': 1.0704920450225472, 'learning_rate/qwen_vl_interface': 6.449797775140321e-06, 'learning_rate/action_model': 6.44979777514032e-05, 'epoch': 0.17}) +[2026-06-24 12:46:21,824][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.0388132743537426, 'train/grad_norm_pre_clip': 23.717172622680664, 'timing/data': 0.00033097900450229645, 'timing/model': 1.0714636391494423, 'learning_rate/qwen_vl_interface': 6.446095160378514e-06, 'learning_rate/action_model': 6.446095160378514e-05, 'epoch': 0.17}) +[2026-06-24 12:46:22,902][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.15703770518302917, 'train/grad_norm_pre_clip': 61.739990234375, 'timing/data': 0.0003164210356771946, 'timing/model': 1.0766050620004535, 'learning_rate/qwen_vl_interface': 6.442391769483544e-06, 'learning_rate/action_model': 6.442391769483544e-05, 'epoch': 0.17}) +[2026-06-24 12:46:24,264][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.09889699518680573, 'train/grad_norm_pre_clip': 150.46554565429688, 'timing/data': 0.0003164310473948717, 'timing/model': 1.360320228850469, 'learning_rate/qwen_vl_interface': 6.4386876048585e-06, 'learning_rate/action_model': 6.4386876048585e-05, 'epoch': 0.17}) +[2026-06-24 12:46:25,339][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.29941338300704956, 'train/grad_norm_pre_clip': 42.29646682739258, 'timing/data': 0.00032947584986686707, 'timing/model': 1.0731882948894054, 'learning_rate/qwen_vl_interface': 6.434982668906975e-06, 'learning_rate/action_model': 6.434982668906974e-05, 'epoch': 0.17}) +[2026-06-24 12:46:26,417][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.12474203109741211, 'train/grad_norm_pre_clip': 35.353172302246094, 'timing/data': 0.0003189349081367254, 'timing/model': 1.0763986790552735, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.17}) +[2026-06-24 12:46:27,488][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.1666957139968872, 'train/grad_norm_pre_clip': 59.694602966308594, 'timing/data': 0.0003349161706864834, 'timing/model': 1.0695140028838068, 'learning_rate/qwen_vl_interface': 6.427570492641351e-06, 'learning_rate/action_model': 6.427570492641351e-05, 'epoch': 0.17}) +[2026-06-24 12:46:28,562][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.08001364022493362, 'train/grad_norm_pre_clip': 25.5014591217041, 'timing/data': 0.0003317000810056925, 'timing/model': 1.072633835952729, 'learning_rate/qwen_vl_interface': 6.423863257136931e-06, 'learning_rate/action_model': 6.42386325713693e-05, 'epoch': 0.17}) +[2026-06-24 12:46:29,641][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.0957859680056572, 'train/grad_norm_pre_clip': 35.525516510009766, 'timing/data': 0.00034455512650310993, 'timing/model': 1.0769138329196721, 'learning_rate/qwen_vl_interface': 6.42015525992539e-06, 'learning_rate/action_model': 6.42015525992539e-05, 'epoch': 0.17}) +[2026-06-24 12:46:30,717][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.13668087124824524, 'train/grad_norm_pre_clip': 62.396514892578125, 'timing/data': 0.00039093196392059326, 'timing/model': 1.074788625119254, 'learning_rate/qwen_vl_interface': 6.416446503412803e-06, 'learning_rate/action_model': 6.416446503412803e-05, 'epoch': 0.17}) +[2026-06-24 12:46:31,792][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.0846562534570694, 'train/grad_norm_pre_clip': 34.90428924560547, 'timing/data': 0.00039263605140149593, 'timing/model': 1.0729659819044173, 'learning_rate/qwen_vl_interface': 6.412736990005746e-06, 'learning_rate/action_model': 6.412736990005746e-05, 'epoch': 0.17}) +[2026-06-24 12:46:32,865][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.07661528885364532, 'train/grad_norm_pre_clip': 24.434988021850586, 'timing/data': 0.00034318212419748306, 'timing/model': 1.0713223200291395, 'learning_rate/qwen_vl_interface': 6.409026722111279e-06, 'learning_rate/action_model': 6.409026722111278e-05, 'epoch': 0.17}) +[2026-06-24 12:46:33,941][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.21989034116268158, 'train/grad_norm_pre_clip': 49.96995544433594, 'timing/data': 0.0003287338186055422, 'timing/model': 1.0745781399309635, 'learning_rate/qwen_vl_interface': 6.405315702136955e-06, 'learning_rate/action_model': 6.405315702136954e-05, 'epoch': 0.17}) +[2026-06-24 12:46:35,012][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.1377076804637909, 'train/grad_norm_pre_clip': 34.36784744262695, 'timing/data': 0.0003189360722899437, 'timing/model': 1.069311989005655, 'learning_rate/qwen_vl_interface': 6.401603932490815e-06, 'learning_rate/action_model': 6.401603932490815e-05, 'epoch': 0.17}) +[2026-06-24 12:46:36,083][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.10427459329366684, 'train/grad_norm_pre_clip': 41.8961296081543, 'timing/data': 0.0003386030439287424, 'timing/model': 1.0697031589224935, 'learning_rate/qwen_vl_interface': 6.3978914155813865e-06, 'learning_rate/action_model': 6.397891415581386e-05, 'epoch': 0.17}) +[2026-06-24 12:46:37,167][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.11021173000335693, 'train/grad_norm_pre_clip': 36.23617172241211, 'timing/data': 0.00033393409103155136, 'timing/model': 1.0814679460600019, 'learning_rate/qwen_vl_interface': 6.394178153817682e-06, 'learning_rate/action_model': 6.394178153817682e-05, 'epoch': 0.17}) +[2026-06-24 12:46:38,238][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.2623594105243683, 'train/grad_norm_pre_clip': 64.67948913574219, 'timing/data': 0.0003032160457223654, 'timing/model': 1.0700992490164936, 'learning_rate/qwen_vl_interface': 6.390464149609193e-06, 'learning_rate/action_model': 6.390464149609193e-05, 'epoch': 0.17}) +[2026-06-24 12:46:39,314][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1750 +[2026-06-24 12:47:03,653][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1750 +[2026-06-24 12:47:03,653][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1750 +[2026-06-24 12:47:36,931][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1750 +[2026-06-24 12:56:34,080][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1750 +[2026-06-24 12:56:34,081][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.058058395981788635, 'train/grad_norm_pre_clip': 22.439912796020508, 'eval/action_loss/seconds': 24.33643164904788, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.09205260127782822, 'eval/latency_2/loss': 0.09205260127782822, 'eval/flappy/loss': 0.09205260127782822, 'eval/flappy/latency_2/loss': 0.09205260127782822, 'eval/flap_precision': 0.851508120649652, 'eval/flap_recall': 0.9959294436906377, 'eval/flap_f1': 0.9180737961225766, 'eval/flap_rate_model': 0.23944444444444443, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9180737961225767, 'eval/latency_2/flap_precision': 0.851508120649652, 'eval/latency_2/flap_recall': 0.9959294436906377, 'eval/latency_2/flap_f1': 0.9180737961225766, 'eval/latency_2/flap_rate_model': 0.23944444444444443, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9180737961225767, 'eval/action_classification/data_wait_seconds': 0.30238035321235657, 'eval/action_classification/predict_seconds': 32.874549865722656, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 374.1000000000155, 'rl_games_eval/mid_train/mean_length': 3034.4, 'rl_games_eval/mid_train/std_reward': 96.30771516344092, 'rl_games_eval/mid_train/std_length': 773.4529332803645, 'rl_games_eval/mid_train/macro_mean_reward': 374.1000000000155, 'rl_games_eval/mid_train/macro_mean_length': 3034.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 374.1000000000155, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3034.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 96.30771516344092, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 773.4529332803645, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 374.1000000000155, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00033859279938042164, 'timing/model': 1.0750225358642638, 'learning_rate/qwen_vl_interface': 6.386749405365901e-06, 'learning_rate/action_model': 6.3867494053659e-05, 'epoch': 0.17}) +[2026-06-24 12:56:35,188][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.21041736006736755, 'train/grad_norm_pre_clip': 69.02632141113281, 'timing/data': 0.0004958061035722494, 'timing/model': 1.1045117580797523, 'learning_rate/qwen_vl_interface': 6.3830339234982596e-06, 'learning_rate/action_model': 6.38303392349826e-05, 'epoch': 0.17}) +[2026-06-24 12:56:36,270][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.3165821135044098, 'train/grad_norm_pre_clip': 205.63571166992188, 'timing/data': 0.0006147199310362339, 'timing/model': 1.0789447908755392, 'learning_rate/qwen_vl_interface': 6.3793177064172075e-06, 'learning_rate/action_model': 6.379317706417207e-05, 'epoch': 0.17}) +[2026-06-24 12:56:37,730][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.1478491723537445, 'train/grad_norm_pre_clip': 22.139028549194336, 'timing/data': 0.0004769698716700077, 'timing/model': 1.4586889161728323, 'learning_rate/qwen_vl_interface': 6.375600756534157e-06, 'learning_rate/action_model': 6.375600756534157e-05, 'epoch': 0.17}) +[2026-06-24 12:56:38,816][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.061157144606113434, 'train/grad_norm_pre_clip': 28.470792770385742, 'timing/data': 0.00039136805571615696, 'timing/model': 1.0834052949212492, 'learning_rate/qwen_vl_interface': 6.371883076260996e-06, 'learning_rate/action_model': 6.371883076260997e-05, 'epoch': 0.17}) +[2026-06-24 12:56:39,904][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.08783552795648575, 'train/grad_norm_pre_clip': 30.064197540283203, 'timing/data': 0.0003933620173484087, 'timing/model': 1.0861101041082293, 'learning_rate/qwen_vl_interface': 6.368164668010087e-06, 'learning_rate/action_model': 6.368164668010087e-05, 'epoch': 0.17}) +[2026-06-24 12:56:41,046][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.1017676293849945, 'train/grad_norm_pre_clip': 21.823436737060547, 'timing/data': 0.0004014880396425724, 'timing/model': 1.1406625530216843, 'learning_rate/qwen_vl_interface': 6.364445534194263e-06, 'learning_rate/action_model': 6.364445534194263e-05, 'epoch': 0.17}) +[2026-06-24 12:56:42,193][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.14189505577087402, 'train/grad_norm_pre_clip': 31.14511489868164, 'timing/data': 0.000643253093585372, 'timing/model': 1.1440622999798506, 'learning_rate/qwen_vl_interface': 6.360725677226834e-06, 'learning_rate/action_model': 6.360725677226834e-05, 'epoch': 0.17}) +[2026-06-24 12:56:43,391][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.14337775111198425, 'train/grad_norm_pre_clip': 23.877321243286133, 'timing/data': 0.00048700813204050064, 'timing/model': 1.195866493973881, 'learning_rate/qwen_vl_interface': 6.357005099521569e-06, 'learning_rate/action_model': 6.357005099521569e-05, 'epoch': 0.17}) +[2026-06-24 12:56:44,494][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.03992888703942299, 'train/grad_norm_pre_clip': 19.934398651123047, 'timing/data': 0.000473623163998127, 'timing/model': 1.100283200852573, 'learning_rate/qwen_vl_interface': 6.353283803492717e-06, 'learning_rate/action_model': 6.353283803492717e-05, 'epoch': 0.17}) +[2026-06-24 12:56:45,683][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.07486770302057266, 'train/grad_norm_pre_clip': 26.64995002746582, 'timing/data': 0.0014355478342622519, 'timing/model': 1.1851313570514321, 'learning_rate/qwen_vl_interface': 6.349561791554981e-06, 'learning_rate/action_model': 6.349561791554981e-05, 'epoch': 0.17}) +[2026-06-24 12:56:46,788][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.1364850401878357, 'train/grad_norm_pre_clip': 23.406572341918945, 'timing/data': 0.0005416420754045248, 'timing/model': 1.1023358248639852, 'learning_rate/qwen_vl_interface': 6.345839066123536e-06, 'learning_rate/action_model': 6.345839066123535e-05, 'epoch': 0.17}) +[2026-06-24 12:56:47,889][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.14246779680252075, 'train/grad_norm_pre_clip': 71.87379455566406, 'timing/data': 0.0004955949261784554, 'timing/model': 1.0987064361106604, 'learning_rate/qwen_vl_interface': 6.342115629614019e-06, 'learning_rate/action_model': 6.342115629614019e-05, 'epoch': 0.17}) +[2026-06-24 12:56:48,981][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.1482851803302765, 'train/grad_norm_pre_clip': 58.35364532470703, 'timing/data': 0.0007451248820871115, 'timing/model': 1.0887188161723316, 'learning_rate/qwen_vl_interface': 6.338391484442527e-06, 'learning_rate/action_model': 6.338391484442526e-05, 'epoch': 0.17}) +[2026-06-24 12:56:50,076][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.07524804025888443, 'train/grad_norm_pre_clip': 22.04466438293457, 'timing/data': 0.00048002484254539013, 'timing/model': 1.089574937010184, 'learning_rate/qwen_vl_interface': 6.334666633025616e-06, 'learning_rate/action_model': 6.334666633025615e-05, 'epoch': 0.17}) +[2026-06-24 12:56:51,191][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.1731904000043869, 'train/grad_norm_pre_clip': 42.75546646118164, 'timing/data': 0.0005530728958547115, 'timing/model': 1.1130126679781824, 'learning_rate/qwen_vl_interface': 6.3309410777803025e-06, 'learning_rate/action_model': 6.330941077780302e-05, 'epoch': 0.17}) +[2026-06-24 12:56:52,272][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.07253025472164154, 'train/grad_norm_pre_clip': 22.902650833129883, 'timing/data': 0.0005399181973189116, 'timing/model': 1.0786573409568518, 'learning_rate/qwen_vl_interface': 6.327214821124063e-06, 'learning_rate/action_model': 6.327214821124062e-05, 'epoch': 0.17}) +[2026-06-24 12:56:53,412][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.07838942110538483, 'train/grad_norm_pre_clip': 40.85563278198242, 'timing/data': 0.0006159420590847731, 'timing/model': 1.137389216106385, 'learning_rate/qwen_vl_interface': 6.323487865474818e-06, 'learning_rate/action_model': 6.323487865474817e-05, 'epoch': 0.17}) +[2026-06-24 12:56:54,864][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.2320047914981842, 'train/grad_norm_pre_clip': 57.47652053833008, 'timing/data': 0.0004071879666298628, 'timing/model': 1.4494778029620647, 'learning_rate/qwen_vl_interface': 6.319760213250952e-06, 'learning_rate/action_model': 6.319760213250951e-05, 'epoch': 0.17}) +[2026-06-24 12:56:55,988][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.07621162384748459, 'train/grad_norm_pre_clip': 17.999061584472656, 'timing/data': 0.0005342171061784029, 'timing/model': 1.1222537900321186, 'learning_rate/qwen_vl_interface': 6.316031866871297e-06, 'learning_rate/action_model': 6.316031866871297e-05, 'epoch': 0.17}) +[2026-06-24 12:56:57,084][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.06913086771965027, 'train/grad_norm_pre_clip': 15.078896522521973, 'timing/data': 0.0005494658835232258, 'timing/model': 1.0930696569848806, 'learning_rate/qwen_vl_interface': 6.31230282875514e-06, 'learning_rate/action_model': 6.31230282875514e-05, 'epoch': 0.17}) +[2026-06-24 12:56:58,202][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.13979767262935638, 'train/grad_norm_pre_clip': 36.28832244873047, 'timing/data': 0.0004908859264105558, 'timing/model': 1.1156639051623642, 'learning_rate/qwen_vl_interface': 6.308573101322209e-06, 'learning_rate/action_model': 6.308573101322209e-05, 'epoch': 0.17}) +[2026-06-24 12:56:59,309][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.12966053187847137, 'train/grad_norm_pre_clip': 78.04568481445312, 'timing/data': 0.0006476009730249643, 'timing/model': 1.1045426880009472, 'learning_rate/qwen_vl_interface': 6.304842686992685e-06, 'learning_rate/action_model': 6.304842686992684e-05, 'epoch': 0.17}) +[2026-06-24 12:57:00,410][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.22473803162574768, 'train/grad_norm_pre_clip': 65.70830535888672, 'timing/data': 0.0004286780022084713, 'timing/model': 1.0994750480167568, 'learning_rate/qwen_vl_interface': 6.301111588187194e-06, 'learning_rate/action_model': 6.301111588187194e-05, 'epoch': 0.17}) +[2026-06-24 12:57:01,517][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.3130893111228943, 'train/grad_norm_pre_clip': 22.121829986572266, 'timing/data': 0.00042778695933520794, 'timing/model': 1.105003491975367, 'learning_rate/qwen_vl_interface': 6.297379807326806e-06, 'learning_rate/action_model': 6.297379807326806e-05, 'epoch': 0.17}) +[2026-06-24 12:57:02,611][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.2852403223514557, 'train/grad_norm_pre_clip': 75.47480010986328, 'timing/data': 0.00040844990871846676, 'timing/model': 1.092134780017659, 'learning_rate/qwen_vl_interface': 6.293647346833033e-06, 'learning_rate/action_model': 6.293647346833032e-05, 'epoch': 0.17}) +[2026-06-24 12:57:03,707][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.1543702632188797, 'train/grad_norm_pre_clip': 50.38168716430664, 'timing/data': 0.0003999941982328892, 'timing/model': 1.0942679799627513, 'learning_rate/qwen_vl_interface': 6.289914209127826e-06, 'learning_rate/action_model': 6.289914209127825e-05, 'epoch': 0.17}) +[2026-06-24 12:57:04,809][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.10651888698339462, 'train/grad_norm_pre_clip': 21.046823501586914, 'timing/data': 0.0004511810839176178, 'timing/model': 1.100108603015542, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.17}) +[2026-06-24 12:57:05,903][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.08107421547174454, 'train/grad_norm_pre_clip': 23.033058166503906, 'timing/data': 0.0006439450662583113, 'timing/model': 1.0919405471067876, 'learning_rate/qwen_vl_interface': 6.282445911773122e-06, 'learning_rate/action_model': 6.282445911773121e-05, 'epoch': 0.17}) +[2026-06-24 12:57:07,003][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.129316508769989, 'train/grad_norm_pre_clip': 48.534488677978516, 'timing/data': 0.00041882903315126896, 'timing/model': 1.0975838960148394, 'learning_rate/qwen_vl_interface': 6.27871075696972e-06, 'learning_rate/action_model': 6.27871075696972e-05, 'epoch': 0.17}) +[2026-06-24 12:57:08,101][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.04377336427569389, 'train/grad_norm_pre_clip': 14.811358451843262, 'timing/data': 0.0004470029380172491, 'timing/model': 1.0968429860658944, 'learning_rate/qwen_vl_interface': 6.27497493464708e-06, 'learning_rate/action_model': 6.27497493464708e-05, 'epoch': 0.17}) +[2026-06-24 12:57:09,203][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.06828267872333527, 'train/grad_norm_pre_clip': 18.71678352355957, 'timing/data': 0.0004566218703985214, 'timing/model': 1.0999476979486644, 'learning_rate/qwen_vl_interface': 6.271238447229332e-06, 'learning_rate/action_model': 6.271238447229332e-05, 'epoch': 0.17}) +[2026-06-24 12:57:10,316][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.10427507013082504, 'train/grad_norm_pre_clip': 28.43790054321289, 'timing/data': 0.0004740238655358553, 'timing/model': 1.1111621628515422, 'learning_rate/qwen_vl_interface': 6.267501297141042e-06, 'learning_rate/action_model': 6.267501297141042e-05, 'epoch': 0.17}) +[2026-06-24 12:57:11,687][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.033505506813526154, 'train/grad_norm_pre_clip': 10.530641555786133, 'timing/data': 0.00040214904583990574, 'timing/model': 1.3683755730744451, 'learning_rate/qwen_vl_interface': 6.263763486807208e-06, 'learning_rate/action_model': 6.263763486807208e-05, 'epoch': 0.17}) +[2026-06-24 12:57:12,794][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.2453470677137375, 'train/grad_norm_pre_clip': 63.15550994873047, 'timing/data': 0.0005102020222693682, 'timing/model': 1.1052016648463905, 'learning_rate/qwen_vl_interface': 6.260025018653253e-06, 'learning_rate/action_model': 6.260025018653253e-05, 'epoch': 0.17}) +[2026-06-24 12:57:13,896][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.07614179700613022, 'train/grad_norm_pre_clip': 17.1893253326416, 'timing/data': 0.0005507988389581442, 'timing/model': 1.0992449901532382, 'learning_rate/qwen_vl_interface': 6.2562858951050294e-06, 'learning_rate/action_model': 6.256285895105029e-05, 'epoch': 0.17}) +[2026-06-24 12:57:14,988][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.03532272204756737, 'train/grad_norm_pre_clip': 7.534902095794678, 'timing/data': 0.0004199820104986429, 'timing/model': 1.0895650940947235, 'learning_rate/qwen_vl_interface': 6.252546118588814e-06, 'learning_rate/action_model': 6.252546118588813e-05, 'epoch': 0.17}) +[2026-06-24 12:57:16,077][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.15701648592948914, 'train/grad_norm_pre_clip': 37.30718231201172, 'timing/data': 0.0005336771719157696, 'timing/model': 1.0855026098433882, 'learning_rate/qwen_vl_interface': 6.248805691531306e-06, 'learning_rate/action_model': 6.248805691531307e-05, 'epoch': 0.17}) +[2026-06-24 12:57:17,219][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.04530010744929314, 'train/grad_norm_pre_clip': 20.46352195739746, 'timing/data': 0.0006994290743023157, 'timing/model': 1.1397138158790767, 'learning_rate/qwen_vl_interface': 6.245064616359628e-06, 'learning_rate/action_model': 6.245064616359628e-05, 'epoch': 0.17}) +[2026-06-24 12:57:18,312][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.045989006757736206, 'train/grad_norm_pre_clip': 26.93412208557129, 'timing/data': 0.0004399099852889776, 'timing/model': 1.090907437959686, 'learning_rate/qwen_vl_interface': 6.241322895501325e-06, 'learning_rate/action_model': 6.241322895501325e-05, 'epoch': 0.17}) +[2026-06-24 12:57:19,405][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.05779580771923065, 'train/grad_norm_pre_clip': 27.186365127563477, 'timing/data': 0.0005874980706721544, 'timing/model': 1.0908354299608618, 'learning_rate/qwen_vl_interface': 6.2375805313843564e-06, 'learning_rate/action_model': 6.237580531384356e-05, 'epoch': 0.17}) +[2026-06-24 12:57:20,603][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.1728391796350479, 'train/grad_norm_pre_clip': 35.78398895263672, 'timing/data': 0.0004004549700766802, 'timing/model': 1.196346300188452, 'learning_rate/qwen_vl_interface': 6.233837526437106e-06, 'learning_rate/action_model': 6.233837526437105e-05, 'epoch': 0.17}) +[2026-06-24 12:57:21,694][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.043552570044994354, 'train/grad_norm_pre_clip': 22.607126235961914, 'timing/data': 0.00038301199674606323, 'timing/model': 1.0878512570634484, 'learning_rate/qwen_vl_interface': 6.230093883088364e-06, 'learning_rate/action_model': 6.230093883088364e-05, 'epoch': 0.17}) +[2026-06-24 12:57:22,798][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.05388697609305382, 'train/grad_norm_pre_clip': 20.80217742919922, 'timing/data': 0.0004954040050506592, 'timing/model': 1.1025334210135043, 'learning_rate/qwen_vl_interface': 6.226349603767341e-06, 'learning_rate/action_model': 6.22634960376734e-05, 'epoch': 0.17}) +[2026-06-24 12:57:23,899][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.02955484949052334, 'train/grad_norm_pre_clip': 10.897810935974121, 'timing/data': 0.0004836120642721653, 'timing/model': 1.0982000180520117, 'learning_rate/qwen_vl_interface': 6.2226046909036616e-06, 'learning_rate/action_model': 6.222604690903662e-05, 'epoch': 0.17}) +[2026-06-24 12:57:24,992][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.02596389502286911, 'train/grad_norm_pre_clip': 14.680519104003906, 'timing/data': 0.0004798548761755228, 'timing/model': 1.0907318040262908, 'learning_rate/qwen_vl_interface': 6.218859146927358e-06, 'learning_rate/action_model': 6.218859146927358e-05, 'epoch': 0.17}) +[2026-06-24 12:57:26,097][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.4096205532550812, 'train/grad_norm_pre_clip': 63.85626983642578, 'timing/data': 0.0007130540907382965, 'timing/model': 1.1021406028885394, 'learning_rate/qwen_vl_interface': 6.215112974268871e-06, 'learning_rate/action_model': 6.21511297426887e-05, 'epoch': 0.17}) +[2026-06-24 12:57:27,208][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.06486743688583374, 'train/grad_norm_pre_clip': 19.178619384765625, 'timing/data': 0.00048430310562253, 'timing/model': 1.109799362020567, 'learning_rate/qwen_vl_interface': 6.211366175359055e-06, 'learning_rate/action_model': 6.211366175359054e-05, 'epoch': 0.17}) +[2026-06-24 12:57:28,584][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.16621831059455872, 'train/grad_norm_pre_clip': 78.36734008789062, 'timing/data': 0.00048656691797077656, 'timing/model': 1.372586838901043, 'learning_rate/qwen_vl_interface': 6.207618752629164e-06, 'learning_rate/action_model': 6.207618752629163e-05, 'epoch': 0.17}) +[2026-06-24 12:57:29,696][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.12796694040298462, 'train/grad_norm_pre_clip': 25.592493057250977, 'timing/data': 0.000729857012629509, 'timing/model': 1.108026672853157, 'learning_rate/qwen_vl_interface': 6.203870708510862e-06, 'learning_rate/action_model': 6.203870708510862e-05, 'epoch': 0.17}) +[2026-06-24 12:57:30,796][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.15339574217796326, 'train/grad_norm_pre_clip': 43.54024887084961, 'timing/data': 0.0004985600244253874, 'timing/model': 1.0975550659932196, 'learning_rate/qwen_vl_interface': 6.200122045436212e-06, 'learning_rate/action_model': 6.200122045436211e-05, 'epoch': 0.17}) +[2026-06-24 12:57:31,887][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.05438145995140076, 'train/grad_norm_pre_clip': 15.802186012268066, 'timing/data': 0.0007853000424802303, 'timing/model': 1.0880867899395525, 'learning_rate/qwen_vl_interface': 6.196372765837682e-06, 'learning_rate/action_model': 6.196372765837682e-05, 'epoch': 0.17}) +[2026-06-24 12:57:32,996][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.23792846500873566, 'train/grad_norm_pre_clip': 47.08942794799805, 'timing/data': 0.0005467210430651903, 'timing/model': 1.106778834015131, 'learning_rate/qwen_vl_interface': 6.19262287214814e-06, 'learning_rate/action_model': 6.19262287214814e-05, 'epoch': 0.17}) +[2026-06-24 12:57:34,083][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.0648801252245903, 'train/grad_norm_pre_clip': 21.87309455871582, 'timing/data': 0.0004159151576459408, 'timing/model': 1.085145089076832, 'learning_rate/qwen_vl_interface': 6.18887236680085e-06, 'learning_rate/action_model': 6.18887236680085e-05, 'epoch': 0.17}) +[2026-06-24 12:57:35,210][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.22490671277046204, 'train/grad_norm_pre_clip': 79.65560913085938, 'timing/data': 0.000576837919652462, 'timing/model': 1.1243004770949483, 'learning_rate/qwen_vl_interface': 6.185121252229475e-06, 'learning_rate/action_model': 6.185121252229474e-05, 'epoch': 0.17}) +[2026-06-24 12:57:36,309][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.0029200746212154627, 'train/grad_norm_pre_clip': 1.168658971786499, 'timing/data': 0.000505432952195406, 'timing/model': 1.0897335270419717, 'learning_rate/qwen_vl_interface': 6.181369530868071e-06, 'learning_rate/action_model': 6.181369530868071e-05, 'epoch': 0.17}) +[2026-06-24 12:57:37,421][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.1134280115365982, 'train/grad_norm_pre_clip': 32.149715423583984, 'timing/data': 0.00037977611646056175, 'timing/model': 1.1102108170744032, 'learning_rate/qwen_vl_interface': 6.17761720515109e-06, 'learning_rate/action_model': 6.17761720515109e-05, 'epoch': 0.17}) +[2026-06-24 12:57:38,524][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.1679924577474594, 'train/grad_norm_pre_clip': 28.41687774658203, 'timing/data': 0.0005212319083511829, 'timing/model': 1.1004296520259231, 'learning_rate/qwen_vl_interface': 6.173864277513376e-06, 'learning_rate/action_model': 6.173864277513376e-05, 'epoch': 0.17}) +[2026-06-24 12:57:39,641][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.06593678891658783, 'train/grad_norm_pre_clip': 34.63340377807617, 'timing/data': 0.000356792937964201, 'timing/model': 1.1148109959904104, 'learning_rate/qwen_vl_interface': 6.170110750390163e-06, 'learning_rate/action_model': 6.170110750390163e-05, 'epoch': 0.17}) +[2026-06-24 12:57:40,754][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.11351250112056732, 'train/grad_norm_pre_clip': 27.57485580444336, 'timing/data': 0.00042366888374090195, 'timing/model': 1.1074402960948646, 'learning_rate/qwen_vl_interface': 6.1663566262170725e-06, 'learning_rate/action_model': 6.166356626217072e-05, 'epoch': 0.18}) +[2026-06-24 12:57:41,913][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.2526254951953888, 'train/grad_norm_pre_clip': 117.59638977050781, 'timing/data': 0.0005853939801454544, 'timing/model': 1.156460725935176, 'learning_rate/qwen_vl_interface': 6.162601907430116e-06, 'learning_rate/action_model': 6.162601907430116e-05, 'epoch': 0.18}) +[2026-06-24 12:57:43,013][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.04126175493001938, 'train/grad_norm_pre_clip': 20.869003295898438, 'timing/data': 0.0004132289905101061, 'timing/model': 1.0980394759681076, 'learning_rate/qwen_vl_interface': 6.15884659646569e-06, 'learning_rate/action_model': 6.15884659646569e-05, 'epoch': 0.18}) +[2026-06-24 12:57:44,119][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.08928649872541428, 'train/grad_norm_pre_clip': 44.10576629638672, 'timing/data': 0.0005819071084260941, 'timing/model': 1.1039916079025716, 'learning_rate/qwen_vl_interface': 6.155090695760574e-06, 'learning_rate/action_model': 6.155090695760573e-05, 'epoch': 0.18}) +[2026-06-24 12:57:45,485][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.1444704234600067, 'train/grad_norm_pre_clip': 45.15180587768555, 'timing/data': 0.00039642699994146824, 'timing/model': 1.3638520308304578, 'learning_rate/qwen_vl_interface': 6.151334207751933e-06, 'learning_rate/action_model': 6.151334207751932e-05, 'epoch': 0.18}) +[2026-06-24 12:57:46,607][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.25563177466392517, 'train/grad_norm_pre_clip': 77.28723907470703, 'timing/data': 0.0007089369464665651, 'timing/model': 1.1193666199687868, 'learning_rate/qwen_vl_interface': 6.1475771348773115e-06, 'learning_rate/action_model': 6.147577134877312e-05, 'epoch': 0.18}) +[2026-06-24 12:57:47,753][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.19586218893527985, 'train/grad_norm_pre_clip': 48.403419494628906, 'timing/data': 0.000474745174869895, 'timing/model': 1.144315447891131, 'learning_rate/qwen_vl_interface': 6.14381947957463e-06, 'learning_rate/action_model': 6.14381947957463e-05, 'epoch': 0.18}) +[2026-06-24 12:57:49,021][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.12134046852588654, 'train/grad_norm_pre_clip': 52.05824279785156, 'timing/data': 0.0004486660473048687, 'timing/model': 1.2656972010154277, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.18}) +[2026-06-24 12:57:50,128][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.18904592096805573, 'train/grad_norm_pre_clip': 45.903587341308594, 'timing/data': 0.00045172194950282574, 'timing/model': 1.105248103151098, 'learning_rate/qwen_vl_interface': 6.136302431438677e-06, 'learning_rate/action_model': 6.136302431438677e-05, 'epoch': 0.18}) +[2026-06-24 12:57:51,225][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.07535916566848755, 'train/grad_norm_pre_clip': 12.529083251953125, 'timing/data': 0.0003885319456458092, 'timing/model': 1.0948497347999364, 'learning_rate/qwen_vl_interface': 6.132543043483137e-06, 'learning_rate/action_model': 6.132543043483136e-05, 'epoch': 0.18}) +[2026-06-24 12:57:52,413][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.1850922405719757, 'train/grad_norm_pre_clip': 34.299957275390625, 'timing/data': 0.00040339119732379913, 'timing/model': 1.1863826799672097, 'learning_rate/qwen_vl_interface': 6.1287830828549975e-06, 'learning_rate/action_model': 6.128783082854997e-05, 'epoch': 0.18}) +[2026-06-24 12:57:53,514][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.005328524857759476, 'train/grad_norm_pre_clip': 3.6634960174560547, 'timing/data': 0.00038196099922060966, 'timing/model': 1.0987930570263416, 'learning_rate/qwen_vl_interface': 6.125022551994055e-06, 'learning_rate/action_model': 6.125022551994056e-05, 'epoch': 0.18}) +[2026-06-24 12:57:54,618][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.06448780000209808, 'train/grad_norm_pre_clip': 23.509090423583984, 'timing/data': 0.0008063409477472305, 'timing/model': 1.1010417151264846, 'learning_rate/qwen_vl_interface': 6.121261453340481e-06, 'learning_rate/action_model': 6.12126145334048e-05, 'epoch': 0.18}) +[2026-06-24 12:57:55,717][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.0800606980919838, 'train/grad_norm_pre_clip': 24.293224334716797, 'timing/data': 0.0004631530027836561, 'timing/model': 1.0967848969157785, 'learning_rate/qwen_vl_interface': 6.117499789334806e-06, 'learning_rate/action_model': 6.117499789334807e-05, 'epoch': 0.18}) +[2026-06-24 12:57:56,820][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.05285247042775154, 'train/grad_norm_pre_clip': 17.092384338378906, 'timing/data': 0.0004982789978384972, 'timing/model': 1.101008289027959, 'learning_rate/qwen_vl_interface': 6.113737562417939e-06, 'learning_rate/action_model': 6.113737562417938e-05, 'epoch': 0.18}) +[2026-06-24 12:57:57,965][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.07765281200408936, 'train/grad_norm_pre_clip': 40.531455993652344, 'timing/data': 0.0004577529616653919, 'timing/model': 1.1429144048597664, 'learning_rate/qwen_vl_interface': 6.109974775031144e-06, 'learning_rate/action_model': 6.109974775031144e-05, 'epoch': 0.18}) +[2026-06-24 12:57:59,107][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.13813193142414093, 'train/grad_norm_pre_clip': 47.8465576171875, 'timing/data': 0.0006110721733421087, 'timing/model': 1.1388466919306666, 'learning_rate/qwen_vl_interface': 6.106211429616054e-06, 'learning_rate/action_model': 6.106211429616053e-05, 'epoch': 0.18}) +[2026-06-24 12:58:00,224][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.2095271199941635, 'train/grad_norm_pre_clip': 66.79817962646484, 'timing/data': 0.00041999202221632004, 'timing/model': 1.1154125698376447, 'learning_rate/qwen_vl_interface': 6.102447528614662e-06, 'learning_rate/action_model': 6.102447528614662e-05, 'epoch': 0.18}) +[2026-06-24 12:58:01,337][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.15985514223575592, 'train/grad_norm_pre_clip': 54.570106506347656, 'timing/data': 0.0005437950603663921, 'timing/model': 1.1101860338822007, 'learning_rate/qwen_vl_interface': 6.098683074469325e-06, 'learning_rate/action_model': 6.0986830744693246e-05, 'epoch': 0.18}) +[2026-06-24 12:58:02,756][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.013089203275740147, 'train/grad_norm_pre_clip': 11.147435188293457, 'timing/data': 0.00044821505434811115, 'timing/model': 1.417289744131267, 'learning_rate/qwen_vl_interface': 6.094918069622754e-06, 'learning_rate/action_model': 6.0949180696227534e-05, 'epoch': 0.18}) +[2026-06-24 12:58:03,923][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.17724202573299408, 'train/grad_norm_pre_clip': 61.5859375, 'timing/data': 0.0006796810775995255, 'timing/model': 1.164373076055199, 'learning_rate/qwen_vl_interface': 6.091152516518022e-06, 'learning_rate/action_model': 6.0911525165180216e-05, 'epoch': 0.18}) +[2026-06-24 12:58:05,034][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.2933602035045624, 'train/grad_norm_pre_clip': 61.77492141723633, 'timing/data': 0.00039633712731301785, 'timing/model': 1.1086271179374307, 'learning_rate/qwen_vl_interface': 6.087386417598554e-06, 'learning_rate/action_model': 6.0873864175985537e-05, 'epoch': 0.18}) +[2026-06-24 12:58:06,204][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.04850221052765846, 'train/grad_norm_pre_clip': 16.27932357788086, 'timing/data': 0.0004271760117262602, 'timing/model': 1.168969236081466, 'learning_rate/qwen_vl_interface': 6.083619775308129e-06, 'learning_rate/action_model': 6.083619775308129e-05, 'epoch': 0.18}) +[2026-06-24 12:58:07,299][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.0681307464838028, 'train/grad_norm_pre_clip': 19.6695556640625, 'timing/data': 0.00046127988025546074, 'timing/model': 1.0903191689867526, 'learning_rate/qwen_vl_interface': 6.079852592090885e-06, 'learning_rate/action_model': 6.079852592090884e-05, 'epoch': 0.18}) +[2026-06-24 12:58:08,395][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.0479544959962368, 'train/grad_norm_pre_clip': 16.518524169921875, 'timing/data': 0.0004064161330461502, 'timing/model': 1.0924890891183168, 'learning_rate/qwen_vl_interface': 6.076084870391303e-06, 'learning_rate/action_model': 6.076084870391302e-05, 'epoch': 0.18}) +[2026-06-24 12:58:09,494][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.04772631451487541, 'train/grad_norm_pre_clip': 12.34245777130127, 'timing/data': 0.0007597729563713074, 'timing/model': 1.0961862979456782, 'learning_rate/qwen_vl_interface': 6.072316612654215e-06, 'learning_rate/action_model': 6.072316612654215e-05, 'epoch': 0.18}) +[2026-06-24 12:58:10,588][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.10898661613464355, 'train/grad_norm_pre_clip': 22.960256576538086, 'timing/data': 0.0006067440845072269, 'timing/model': 1.0910492818802595, 'learning_rate/qwen_vl_interface': 6.068547821324808e-06, 'learning_rate/action_model': 6.068547821324807e-05, 'epoch': 0.18}) +[2026-06-24 12:58:11,688][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.23582953214645386, 'train/grad_norm_pre_clip': 83.3913345336914, 'timing/data': 0.0005446069408208132, 'timing/model': 1.0971661929506809, 'learning_rate/qwen_vl_interface': 6.064778498848605e-06, 'learning_rate/action_model': 6.064778498848604e-05, 'epoch': 0.18}) +[2026-06-24 12:58:12,804][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.17077720165252686, 'train/grad_norm_pre_clip': 36.677406311035156, 'timing/data': 0.0006905810441821814, 'timing/model': 1.113920127041638, 'learning_rate/qwen_vl_interface': 6.061008647671479e-06, 'learning_rate/action_model': 6.061008647671479e-05, 'epoch': 0.18}) +[2026-06-24 12:58:13,907][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.053726013749837875, 'train/grad_norm_pre_clip': 16.05771827697754, 'timing/data': 0.0003972891718149185, 'timing/model': 1.1003400639165193, 'learning_rate/qwen_vl_interface': 6.05723827023965e-06, 'learning_rate/action_model': 6.057238270239649e-05, 'epoch': 0.18}) +[2026-06-24 12:58:15,021][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.11094775795936584, 'train/grad_norm_pre_clip': 28.401466369628906, 'timing/data': 0.0005951619241386652, 'timing/model': 1.1120367529802024, 'learning_rate/qwen_vl_interface': 6.053467368999667e-06, 'learning_rate/action_model': 6.0534673689996665e-05, 'epoch': 0.18}) +[2026-06-24 12:58:16,129][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.023782148957252502, 'train/grad_norm_pre_clip': 11.23100757598877, 'timing/data': 0.0003885419573634863, 'timing/model': 1.1061692608054727, 'learning_rate/qwen_vl_interface': 6.049695946398433e-06, 'learning_rate/action_model': 6.049695946398433e-05, 'epoch': 0.18}) +[2026-06-24 12:58:17,265][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.09104284644126892, 'train/grad_norm_pre_clip': 20.816341400146484, 'timing/data': 0.00039605703204870224, 'timing/model': 1.13353228289634, 'learning_rate/qwen_vl_interface': 6.045924004883184e-06, 'learning_rate/action_model': 6.045924004883183e-05, 'epoch': 0.18}) +[2026-06-24 12:58:18,637][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.060802556574344635, 'train/grad_norm_pre_clip': 49.20659637451172, 'timing/data': 0.0007750610820949078, 'timing/model': 1.3692185441032052, 'learning_rate/qwen_vl_interface': 6.042151546901487e-06, 'learning_rate/action_model': 6.042151546901487e-05, 'epoch': 0.18}) +[2026-06-24 12:58:19,804][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.08942105621099472, 'train/grad_norm_pre_clip': 54.16060256958008, 'timing/data': 0.0004319138824939728, 'timing/model': 1.1654624710790813, 'learning_rate/qwen_vl_interface': 6.038378574901252e-06, 'learning_rate/action_model': 6.0383785749012514e-05, 'epoch': 0.18}) +[2026-06-24 12:58:20,895][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.02948995679616928, 'train/grad_norm_pre_clip': 19.759082794189453, 'timing/data': 0.00037073902785778046, 'timing/model': 1.0875922029372305, 'learning_rate/qwen_vl_interface': 6.0346050913307195e-06, 'learning_rate/action_model': 6.0346050913307195e-05, 'epoch': 0.18}) +[2026-06-24 12:58:21,984][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.03381284698843956, 'train/grad_norm_pre_clip': 18.358863830566406, 'timing/data': 0.0005317931063473225, 'timing/model': 1.0871351610403508, 'learning_rate/qwen_vl_interface': 6.03083109863846e-06, 'learning_rate/action_model': 6.03083109863846e-05, 'epoch': 0.18}) +[2026-06-24 12:58:23,087][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.1793784648180008, 'train/grad_norm_pre_clip': 90.06912231445312, 'timing/data': 0.000598657876253128, 'timing/model': 1.0991572218481451, 'learning_rate/qwen_vl_interface': 6.02705659927338e-06, 'learning_rate/action_model': 6.02705659927338e-05, 'epoch': 0.18}) +[2026-06-24 12:58:24,181][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.3169816732406616, 'train/grad_norm_pre_clip': 67.12783813476562, 'timing/data': 0.0004919881466776133, 'timing/model': 1.0921451728790998, 'learning_rate/qwen_vl_interface': 6.023281595684709e-06, 'learning_rate/action_model': 6.023281595684709e-05, 'epoch': 0.18}) +[2026-06-24 12:58:25,304][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.11856010556221008, 'train/grad_norm_pre_clip': 75.80489349365234, 'timing/data': 0.0005069461185485125, 'timing/model': 1.120392661076039, 'learning_rate/qwen_vl_interface': 6.019506090322003e-06, 'learning_rate/action_model': 6.019506090322004e-05, 'epoch': 0.18}) +[2026-06-24 12:58:26,398][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.32947978377342224, 'train/grad_norm_pre_clip': 69.47215270996094, 'timing/data': 0.00048106699250638485, 'timing/model': 1.0916880790609866, 'learning_rate/qwen_vl_interface': 6.015730085635153e-06, 'learning_rate/action_model': 6.015730085635153e-05, 'epoch': 0.18}) +[2026-06-24 12:58:27,506][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.013412467204034328, 'train/grad_norm_pre_clip': 6.360056400299072, 'timing/data': 0.0004937311168760061, 'timing/model': 1.105959963053465, 'learning_rate/qwen_vl_interface': 6.011953584074361e-06, 'learning_rate/action_model': 6.0119535840743614e-05, 'epoch': 0.18}) +[2026-06-24 12:58:28,616][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.1646953970193863, 'train/grad_norm_pre_clip': 67.72721099853516, 'timing/data': 0.0005294880829751492, 'timing/model': 1.1080017511267215, 'learning_rate/qwen_vl_interface': 6.008176588090162e-06, 'learning_rate/action_model': 6.0081765880901616e-05, 'epoch': 0.18}) +[2026-06-24 12:58:29,709][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.019951503723859787, 'train/grad_norm_pre_clip': 18.351482391357422, 'timing/data': 0.00039220904000103474, 'timing/model': 1.0898551030550152, 'learning_rate/qwen_vl_interface': 6.004399100133409e-06, 'learning_rate/action_model': 6.004399100133409e-05, 'epoch': 0.18}) +[2026-06-24 12:58:30,803][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.02020145207643509, 'train/grad_norm_pre_clip': 23.237016677856445, 'timing/data': 0.00047551700845360756, 'timing/model': 1.0915854540653527, 'learning_rate/qwen_vl_interface': 6.0006211226552655e-06, 'learning_rate/action_model': 6.0006211226552655e-05, 'epoch': 0.18}) +[2026-06-24 12:58:32,004][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.06297038495540619, 'train/grad_norm_pre_clip': 36.70741653442383, 'timing/data': 0.0004034200683236122, 'timing/model': 1.1994113789405674, 'learning_rate/qwen_vl_interface': 5.996842658107228e-06, 'learning_rate/action_model': 5.996842658107228e-05, 'epoch': 0.18}) +[2026-06-24 12:58:33,101][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.05075901001691818, 'train/grad_norm_pre_clip': 46.86371612548828, 'timing/data': 0.0007040770724415779, 'timing/model': 1.0947164332028478, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.18}) +[2026-06-24 12:58:34,193][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.009130332618951797, 'train/grad_norm_pre_clip': 9.261545181274414, 'timing/data': 0.00046555791050195694, 'timing/model': 1.0892306219320744, 'learning_rate/qwen_vl_interface': 5.989284277608993e-06, 'learning_rate/action_model': 5.989284277608993e-05, 'epoch': 0.18}) +[2026-06-24 12:58:35,297][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.05686509236693382, 'train/grad_norm_pre_clip': 66.15634155273438, 'timing/data': 0.0006991680711507797, 'timing/model': 1.0986371128819883, 'learning_rate/qwen_vl_interface': 5.985504366563347e-06, 'learning_rate/action_model': 5.985504366563347e-05, 'epoch': 0.18}) +[2026-06-24 12:58:36,396][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.27027612924575806, 'train/grad_norm_pre_clip': 89.9167709350586, 'timing/data': 0.0015000479761511087, 'timing/model': 1.096674472792074, 'learning_rate/qwen_vl_interface': 5.9817239782569045e-06, 'learning_rate/action_model': 5.9817239782569045e-05, 'epoch': 0.18}) +[2026-06-24 12:58:37,491][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.06524118781089783, 'train/grad_norm_pre_clip': 39.184242248535156, 'timing/data': 0.0016381579916924238, 'timing/model': 1.0899799929466099, 'learning_rate/qwen_vl_interface': 5.977943115142715e-06, 'learning_rate/action_model': 5.977943115142715e-05, 'epoch': 0.18}) +[2026-06-24 12:58:38,914][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.6238582134246826, 'train/grad_norm_pre_clip': 150.8480987548828, 'timing/data': 0.0005633819382637739, 'timing/model': 1.4208417672198266, 'learning_rate/qwen_vl_interface': 5.9741617796741415e-06, 'learning_rate/action_model': 5.9741617796741414e-05, 'epoch': 0.18}) +[2026-06-24 12:58:40,005][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.05265231430530548, 'train/grad_norm_pre_clip': 32.936336517333984, 'timing/data': 0.00041215703822672367, 'timing/model': 1.08971571107395, 'learning_rate/qwen_vl_interface': 5.970379974304851e-06, 'learning_rate/action_model': 5.9703799743048506e-05, 'epoch': 0.18}) +[2026-06-24 12:58:41,107][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.07980429381132126, 'train/grad_norm_pre_clip': 63.5150032043457, 'timing/data': 0.00039178901351988316, 'timing/model': 1.1000866889953613, 'learning_rate/qwen_vl_interface': 5.966597701488816e-06, 'learning_rate/action_model': 5.9665977014888154e-05, 'epoch': 0.18}) +[2026-06-24 12:58:42,198][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.033905282616615295, 'train/grad_norm_pre_clip': 41.064048767089844, 'timing/data': 0.00041177705861628056, 'timing/model': 1.0890462980605662, 'learning_rate/qwen_vl_interface': 5.962814963680315e-06, 'learning_rate/action_model': 5.9628149636803144e-05, 'epoch': 0.18}) +[2026-06-24 12:58:43,299][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.0931260734796524, 'train/grad_norm_pre_clip': 38.596256256103516, 'timing/data': 0.00040716701187193394, 'timing/model': 1.0978419980965555, 'learning_rate/qwen_vl_interface': 5.95903176333392e-06, 'learning_rate/action_model': 5.95903176333392e-05, 'epoch': 0.18}) +[2026-06-24 12:58:44,407][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.028432512655854225, 'train/grad_norm_pre_clip': 48.88565444946289, 'timing/data': 0.0004929089918732643, 'timing/model': 1.106767490040511, 'learning_rate/qwen_vl_interface': 5.955248102904513e-06, 'learning_rate/action_model': 5.955248102904512e-05, 'epoch': 0.18}) +[2026-06-24 12:58:45,513][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.12973372638225555, 'train/grad_norm_pre_clip': 38.57368469238281, 'timing/data': 0.0005238780286163092, 'timing/model': 1.1030519369523972, 'learning_rate/qwen_vl_interface': 5.951463984847269e-06, 'learning_rate/action_model': 5.951463984847269e-05, 'epoch': 0.18}) +[2026-06-24 12:58:46,610][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.034764815121889114, 'train/grad_norm_pre_clip': 21.409778594970703, 'timing/data': 0.0004063560627400875, 'timing/model': 1.0957186040468514, 'learning_rate/qwen_vl_interface': 5.94767941161766e-06, 'learning_rate/action_model': 5.9476794116176596e-05, 'epoch': 0.18}) +[2026-06-24 12:58:47,814][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.11526653915643692, 'train/grad_norm_pre_clip': 79.56504821777344, 'timing/data': 0.0003970479592680931, 'timing/model': 1.2022633189335465, 'learning_rate/qwen_vl_interface': 5.943894385671456e-06, 'learning_rate/action_model': 5.9438943856714556e-05, 'epoch': 0.18}) +[2026-06-24 12:58:48,918][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.1608918458223343, 'train/grad_norm_pre_clip': 74.25003814697266, 'timing/data': 0.00040984293445944786, 'timing/model': 1.1019117359537631, 'learning_rate/qwen_vl_interface': 5.940108909464719e-06, 'learning_rate/action_model': 5.940108909464719e-05, 'epoch': 0.18}) +[2026-06-24 12:58:50,057][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.07471464574337006, 'train/grad_norm_pre_clip': 37.88642501831055, 'timing/data': 0.0005302091594785452, 'timing/model': 1.1366056490223855, 'learning_rate/qwen_vl_interface': 5.936322985453802e-06, 'learning_rate/action_model': 5.936322985453802e-05, 'epoch': 0.18}) +[2026-06-24 12:58:51,216][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.16679498553276062, 'train/grad_norm_pre_clip': 91.07434844970703, 'timing/data': 0.0004950130823999643, 'timing/model': 1.1573853141162544, 'learning_rate/qwen_vl_interface': 5.9325366160953516e-06, 'learning_rate/action_model': 5.932536616095351e-05, 'epoch': 0.18}) +[2026-06-24 12:58:52,315][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.11238753795623779, 'train/grad_norm_pre_clip': 85.23796081542969, 'timing/data': 0.0003985809162259102, 'timing/model': 1.0971905142068863, 'learning_rate/qwen_vl_interface': 5.928749803846301e-06, 'learning_rate/action_model': 5.9287498038463006e-05, 'epoch': 0.18}) +[2026-06-24 12:58:53,525][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.05049905925989151, 'train/grad_norm_pre_clip': 64.75312805175781, 'timing/data': 0.00039066700264811516, 'timing/model': 1.2083657619077712, 'learning_rate/qwen_vl_interface': 5.924962551163872e-06, 'learning_rate/action_model': 5.9249625511638716e-05, 'epoch': 0.18}) +[2026-06-24 12:58:54,643][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.25477898120880127, 'train/grad_norm_pre_clip': 87.13483428955078, 'timing/data': 0.0005691831465810537, 'timing/model': 1.1154784839600325, 'learning_rate/qwen_vl_interface': 5.921174860505573e-06, 'learning_rate/action_model': 5.9211748605055725e-05, 'epoch': 0.18}) +[2026-06-24 12:58:55,825][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.2872750461101532, 'train/grad_norm_pre_clip': 98.7579574584961, 'timing/data': 0.0003940630704164505, 'timing/model': 1.1798962890170515, 'learning_rate/qwen_vl_interface': 5.9173867343291916e-06, 'learning_rate/action_model': 5.917386734329191e-05, 'epoch': 0.18}) +[2026-06-24 12:58:56,974][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.1410643458366394, 'train/grad_norm_pre_clip': 66.24598693847656, 'timing/data': 0.0003902760799974203, 'timing/model': 1.1463146121241152, 'learning_rate/qwen_vl_interface': 5.913598175092804e-06, 'learning_rate/action_model': 5.913598175092804e-05, 'epoch': 0.18}) +[2026-06-24 12:58:58,395][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.0033603047486394644, 'train/grad_norm_pre_clip': 2.456272602081299, 'timing/data': 0.0005465599242597818, 'timing/model': 1.418274398893118, 'learning_rate/qwen_vl_interface': 5.909809185254767e-06, 'learning_rate/action_model': 5.9098091852547665e-05, 'epoch': 0.18}) +[2026-06-24 12:58:59,505][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.2651461958885193, 'train/grad_norm_pre_clip': 103.38957977294922, 'timing/data': 0.0004778101574629545, 'timing/model': 1.1078476950060576, 'learning_rate/qwen_vl_interface': 5.90601976727371e-06, 'learning_rate/action_model': 5.906019767273711e-05, 'epoch': 0.18}) +[2026-06-24 12:59:00,600][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.12115845084190369, 'train/grad_norm_pre_clip': 40.810707092285156, 'timing/data': 0.0003875000402331352, 'timing/model': 1.0926206950098276, 'learning_rate/qwen_vl_interface': 5.902229923608552e-06, 'learning_rate/action_model': 5.902229923608552e-05, 'epoch': 0.18}) +[2026-06-24 12:59:01,704][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.217490553855896, 'train/grad_norm_pre_clip': 64.12515258789062, 'timing/data': 0.0004484050441533327, 'timing/model': 1.1018463841173798, 'learning_rate/qwen_vl_interface': 5.898439656718476e-06, 'learning_rate/action_model': 5.898439656718476e-05, 'epoch': 0.18}) +[2026-06-24 12:59:02,801][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.05020785331726074, 'train/grad_norm_pre_clip': 95.22737884521484, 'timing/data': 0.0004107749555259943, 'timing/model': 1.095834021922201, 'learning_rate/qwen_vl_interface': 5.894648969062947e-06, 'learning_rate/action_model': 5.894648969062947e-05, 'epoch': 0.18}) +[2026-06-24 12:59:03,900][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.174753800034523, 'train/grad_norm_pre_clip': 45.445404052734375, 'timing/data': 0.00038485508412122726, 'timing/model': 1.0961222050245851, 'learning_rate/qwen_vl_interface': 5.890857863101701e-06, 'learning_rate/action_model': 5.890857863101701e-05, 'epoch': 0.18}) +[2026-06-24 12:59:04,995][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.1346009224653244, 'train/grad_norm_pre_clip': 52.41975021362305, 'timing/data': 0.0005022368859499693, 'timing/model': 1.0932197351939976, 'learning_rate/qwen_vl_interface': 5.887066341294745e-06, 'learning_rate/action_model': 5.8870663412947445e-05, 'epoch': 0.18}) +[2026-06-24 12:59:06,088][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.3066018521785736, 'train/grad_norm_pre_clip': 59.909400939941406, 'timing/data': 0.0005145601462572813, 'timing/model': 1.090984069975093, 'learning_rate/qwen_vl_interface': 5.883274406102356e-06, 'learning_rate/action_model': 5.883274406102356e-05, 'epoch': 0.18}) +[2026-06-24 12:59:07,197][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.12177310883998871, 'train/grad_norm_pre_clip': 57.712379455566406, 'timing/data': 0.0005234070122241974, 'timing/model': 1.1070166279096156, 'learning_rate/qwen_vl_interface': 5.879482059985082e-06, 'learning_rate/action_model': 5.879482059985082e-05, 'epoch': 0.18}) +[2026-06-24 12:59:08,292][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.20056694746017456, 'train/grad_norm_pre_clip': 38.06550216674805, 'timing/data': 0.000451341038569808, 'timing/model': 1.092246409971267, 'learning_rate/qwen_vl_interface': 5.875689305403732e-06, 'learning_rate/action_model': 5.875689305403732e-05, 'epoch': 0.18}) +[2026-06-24 12:59:09,422][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.14014647901058197, 'train/grad_norm_pre_clip': 22.66034507751465, 'timing/data': 0.0008606419432908297, 'timing/model': 1.1269456578884274, 'learning_rate/qwen_vl_interface': 5.871896144819384e-06, 'learning_rate/action_model': 5.871896144819384e-05, 'epoch': 0.18}) +[2026-06-24 12:59:10,521][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.07373590022325516, 'train/grad_norm_pre_clip': 25.156620025634766, 'timing/data': 0.00039521511644124985, 'timing/model': 1.0979204948525876, 'learning_rate/qwen_vl_interface': 5.868102580693381e-06, 'learning_rate/action_model': 5.868102580693381e-05, 'epoch': 0.18}) +[2026-06-24 12:59:11,633][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.0332048274576664, 'train/grad_norm_pre_clip': 22.229843139648438, 'timing/data': 0.0003973089624196291, 'timing/model': 1.109535425202921, 'learning_rate/qwen_vl_interface': 5.864308615487323e-06, 'learning_rate/action_model': 5.864308615487323e-05, 'epoch': 0.18}) +[2026-06-24 12:59:12,745][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.07943211495876312, 'train/grad_norm_pre_clip': 34.289268493652344, 'timing/data': 0.0004002440255135298, 'timing/model': 1.1102089839987457, 'learning_rate/qwen_vl_interface': 5.860514251663075e-06, 'learning_rate/action_model': 5.860514251663075e-05, 'epoch': 0.18}) +[2026-06-24 12:59:13,924][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.149271160364151, 'train/grad_norm_pre_clip': 25.864974975585938, 'timing/data': 0.0004129679873585701, 'timing/model': 1.1775312251411378, 'learning_rate/qwen_vl_interface': 5.8567194916827575e-06, 'learning_rate/action_model': 5.8567194916827575e-05, 'epoch': 0.18}) +[2026-06-24 12:59:15,031][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.1866540014743805, 'train/grad_norm_pre_clip': 35.59226989746094, 'timing/data': 0.00044014002196490765, 'timing/model': 1.1043407139368355, 'learning_rate/qwen_vl_interface': 5.852924338008748e-06, 'learning_rate/action_model': 5.8529243380087474e-05, 'epoch': 0.18}) +[2026-06-24 12:59:16,211][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.09067709743976593, 'train/grad_norm_pre_clip': 22.256324768066406, 'timing/data': 0.0004298700951039791, 'timing/model': 1.178495439933613, 'learning_rate/qwen_vl_interface': 5.849128793103682e-06, 'learning_rate/action_model': 5.849128793103682e-05, 'epoch': 0.18}) +[2026-06-24 12:59:17,680][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.18202094733715057, 'train/grad_norm_pre_clip': 31.810667037963867, 'timing/data': 0.00045960606075823307, 'timing/model': 1.4660261000972241, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.18}) +[2026-06-24 12:59:18,793][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.09968194365501404, 'train/grad_norm_pre_clip': 24.56985092163086, 'timing/data': 0.0008154769893735647, 'timing/model': 1.1095685630571097, 'learning_rate/qwen_vl_interface': 5.841536539452181e-06, 'learning_rate/action_model': 5.841536539452181e-05, 'epoch': 0.18}) +[2026-06-24 12:59:19,889][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.1373179852962494, 'train/grad_norm_pre_clip': 31.4414119720459, 'timing/data': 0.0005842309910804033, 'timing/model': 1.0924832138698548, 'learning_rate/qwen_vl_interface': 5.83773983563228e-06, 'learning_rate/action_model': 5.8377398356322796e-05, 'epoch': 0.18}) +[2026-06-24 12:59:21,008][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.06624479591846466, 'train/grad_norm_pre_clip': 14.717293739318848, 'timing/data': 0.0005179361905902624, 'timing/model': 1.1168155809864402, 'learning_rate/qwen_vl_interface': 5.833942750434377e-06, 'learning_rate/action_model': 5.8339427504343766e-05, 'epoch': 0.18}) +[2026-06-24 12:59:22,112][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.16929295659065247, 'train/grad_norm_pre_clip': 15.463321685791016, 'timing/data': 0.0006676081102341413, 'timing/model': 1.1019394448958337, 'learning_rate/qwen_vl_interface': 5.830145286322365e-06, 'learning_rate/action_model': 5.830145286322365e-05, 'epoch': 0.18}) +[2026-06-24 12:59:23,296][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.10641602426767349, 'train/grad_norm_pre_clip': 31.807077407836914, 'timing/data': 0.0006342248525470495, 'timing/model': 1.1812858411576599, 'learning_rate/qwen_vl_interface': 5.826347445760375e-06, 'learning_rate/action_model': 5.8263474457603754e-05, 'epoch': 0.18}) +[2026-06-24 12:59:24,388][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.2036963403224945, 'train/grad_norm_pre_clip': 38.79899978637695, 'timing/data': 0.0005745720118284225, 'timing/model': 1.0893518719822168, 'learning_rate/qwen_vl_interface': 5.822549231212785e-06, 'learning_rate/action_model': 5.822549231212785e-05, 'epoch': 0.18}) +[2026-06-24 12:59:25,482][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.18214848637580872, 'train/grad_norm_pre_clip': 36.589988708496094, 'timing/data': 0.0007432801648974419, 'timing/model': 1.09077242994681, 'learning_rate/qwen_vl_interface': 5.8187506451442155e-06, 'learning_rate/action_model': 5.8187506451442155e-05, 'epoch': 0.18}) +[2026-06-24 12:59:26,622][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.033190250396728516, 'train/grad_norm_pre_clip': 9.89718246459961, 'timing/data': 0.0004967460408806801, 'timing/model': 1.1382833090610802, 'learning_rate/qwen_vl_interface': 5.814951690019527e-06, 'learning_rate/action_model': 5.814951690019527e-05, 'epoch': 0.18}) +[2026-06-24 12:59:27,692][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.10695699602365494, 'train/grad_norm_pre_clip': 18.098243713378906, 'timing/data': 0.0004152418114244938, 'timing/model': 1.0681438259780407, 'learning_rate/qwen_vl_interface': 5.811152368303819e-06, 'learning_rate/action_model': 5.811152368303819e-05, 'epoch': 0.18}) +[2026-06-24 12:59:28,761][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.18601389229297638, 'train/grad_norm_pre_clip': 30.566448211669922, 'timing/data': 0.0003810080233961344, 'timing/model': 1.0672474480234087, 'learning_rate/qwen_vl_interface': 5.807352682462432e-06, 'learning_rate/action_model': 5.807352682462432e-05, 'epoch': 0.18}) +[2026-06-24 12:59:29,835][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.1500711739063263, 'train/grad_norm_pre_clip': 57.0349235534668, 'timing/data': 0.0004166848957538605, 'timing/model': 1.0718355160206556, 'learning_rate/qwen_vl_interface': 5.803552634960942e-06, 'learning_rate/action_model': 5.8035526349609415e-05, 'epoch': 0.18}) +[2026-06-24 12:59:30,904][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.18815971910953522, 'train/grad_norm_pre_clip': 40.855560302734375, 'timing/data': 0.00040296907536685467, 'timing/model': 1.0676856159698218, 'learning_rate/qwen_vl_interface': 5.799752228265157e-06, 'learning_rate/action_model': 5.799752228265156e-05, 'epoch': 0.18}) +[2026-06-24 12:59:31,997][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.06594283133745193, 'train/grad_norm_pre_clip': 13.508190155029297, 'timing/data': 0.0006022150628268719, 'timing/model': 1.0903860160615295, 'learning_rate/qwen_vl_interface': 5.7959514648411175e-06, 'learning_rate/action_model': 5.795951464841117e-05, 'epoch': 0.18}) +[2026-06-24 12:59:33,101][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.15668103098869324, 'train/grad_norm_pre_clip': 17.767473220825195, 'timing/data': 0.0005799529608339071, 'timing/model': 1.09718735399656, 'learning_rate/qwen_vl_interface': 5.792150347155101e-06, 'learning_rate/action_model': 5.7921503471551007e-05, 'epoch': 0.18}) +[2026-06-24 12:59:34,251][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.06882544606924057, 'train/grad_norm_pre_clip': 21.042268753051758, 'timing/data': 0.0009560510516166687, 'timing/model': 1.1462243048008531, 'learning_rate/qwen_vl_interface': 5.78834887767361e-06, 'learning_rate/action_model': 5.7883488776736093e-05, 'epoch': 0.18}) +[2026-06-24 12:59:35,481][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.08787966519594193, 'train/grad_norm_pre_clip': 14.162219047546387, 'timing/data': 0.0007321401499211788, 'timing/model': 1.2261331530753523, 'learning_rate/qwen_vl_interface': 5.784547058863377e-06, 'learning_rate/action_model': 5.784547058863377e-05, 'epoch': 0.18}) +[2026-06-24 12:59:36,729][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.10136266052722931, 'train/grad_norm_pre_clip': 30.945093154907227, 'timing/data': 0.001210241112858057, 'timing/model': 1.2449129521846771, 'learning_rate/qwen_vl_interface': 5.780744893191361e-06, 'learning_rate/action_model': 5.7807448931913614e-05, 'epoch': 0.18}) +[2026-06-24 12:59:38,464][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.06120556220412254, 'train/grad_norm_pre_clip': 20.146230697631836, 'timing/data': 0.0006112318951636553, 'timing/model': 1.7331622140482068, 'learning_rate/qwen_vl_interface': 5.7769423831247474e-06, 'learning_rate/action_model': 5.776942383124747e-05, 'epoch': 0.18}) +[2026-06-24 12:59:39,530][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.034614793956279755, 'train/grad_norm_pre_clip': 7.998177528381348, 'timing/data': 0.0003319159150123596, 'timing/model': 1.064314090181142, 'learning_rate/qwen_vl_interface': 5.7731395311309405e-06, 'learning_rate/action_model': 5.77313953113094e-05, 'epoch': 0.19}) +[2026-06-24 12:59:40,594][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.025167398154735565, 'train/grad_norm_pre_clip': 11.795869827270508, 'timing/data': 0.0003119481261819601, 'timing/model': 1.0622387209441513, 'learning_rate/qwen_vl_interface': 5.769336339677573e-06, 'learning_rate/action_model': 5.769336339677572e-05, 'epoch': 0.19}) +[2026-06-24 12:59:41,663][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.2135678082704544, 'train/grad_norm_pre_clip': 42.21830749511719, 'timing/data': 0.0005009141750633717, 'timing/model': 1.0669926600530744, 'learning_rate/qwen_vl_interface': 5.76553281123249e-06, 'learning_rate/action_model': 5.7655328112324904e-05, 'epoch': 0.19}) +[2026-06-24 12:59:42,730][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.26468560099601746, 'train/grad_norm_pre_clip': 78.77196502685547, 'timing/data': 0.0005166030023247004, 'timing/model': 1.0648355861194432, 'learning_rate/qwen_vl_interface': 5.761728948263766e-06, 'learning_rate/action_model': 5.761728948263766e-05, 'epoch': 0.19}) +[2026-06-24 12:59:43,795][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.12216942012310028, 'train/grad_norm_pre_clip': 48.4613037109375, 'timing/data': 0.00033451104536652565, 'timing/model': 1.0625785149168223, 'learning_rate/qwen_vl_interface': 5.757924753239683e-06, 'learning_rate/action_model': 5.757924753239683e-05, 'epoch': 0.19}) +[2026-06-24 12:59:44,861][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.02806432545185089, 'train/grad_norm_pre_clip': 7.461988925933838, 'timing/data': 0.0003164559602737427, 'timing/model': 1.0652296720072627, 'learning_rate/qwen_vl_interface': 5.7541202286287405e-06, 'learning_rate/action_model': 5.75412022862874e-05, 'epoch': 0.19}) +[2026-06-24 12:59:45,928][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.002170235849916935, 'train/grad_norm_pre_clip': 1.4390274286270142, 'timing/data': 0.00033964007161557674, 'timing/model': 1.0648457799106836, 'learning_rate/qwen_vl_interface': 5.750315376899655e-06, 'learning_rate/action_model': 5.7503153768996554e-05, 'epoch': 0.19}) +[2026-06-24 12:59:46,997][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.07653544843196869, 'train/grad_norm_pre_clip': 25.475801467895508, 'timing/data': 0.0003352921921759844, 'timing/model': 1.0674760779365897, 'learning_rate/qwen_vl_interface': 5.746510200521354e-06, 'learning_rate/action_model': 5.746510200521355e-05, 'epoch': 0.19}) +[2026-06-24 12:59:48,063][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.2978663742542267, 'train/grad_norm_pre_clip': 28.512678146362305, 'timing/data': 0.0003168969415128231, 'timing/model': 1.0649691249709576, 'learning_rate/qwen_vl_interface': 5.742704701962974e-06, 'learning_rate/action_model': 5.7427047019629744e-05, 'epoch': 0.19}) +[2026-06-24 12:59:49,126][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.09542521834373474, 'train/grad_norm_pre_clip': 23.806808471679688, 'timing/data': 0.00033439998514950275, 'timing/model': 1.0611377670429647, 'learning_rate/qwen_vl_interface': 5.7388988836938645e-06, 'learning_rate/action_model': 5.738898883693864e-05, 'epoch': 0.19}) +[2026-06-24 12:59:50,196][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.06845538318157196, 'train/grad_norm_pre_clip': 55.88854217529297, 'timing/data': 0.00030987407080829144, 'timing/model': 1.0678272529039532, 'learning_rate/qwen_vl_interface': 5.735092748183574e-06, 'learning_rate/action_model': 5.7350927481835736e-05, 'epoch': 0.19}) +[2026-06-24 12:59:51,266][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.012697622179985046, 'train/grad_norm_pre_clip': 7.211709976196289, 'timing/data': 0.0003164971712976694, 'timing/model': 1.0687552629970014, 'learning_rate/qwen_vl_interface': 5.7312862979018676e-06, 'learning_rate/action_model': 5.731286297901868e-05, 'epoch': 0.19}) +[2026-06-24 12:59:52,441][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.1177748441696167, 'train/grad_norm_pre_clip': 42.12868118286133, 'timing/data': 0.0003143318463116884, 'timing/model': 1.1728601770009845, 'learning_rate/qwen_vl_interface': 5.727479535318708e-06, 'learning_rate/action_model': 5.727479535318707e-05, 'epoch': 0.19}) +[2026-06-24 12:59:53,745][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.27719101309776306, 'train/grad_norm_pre_clip': 72.03369140625, 'timing/data': 0.000732440035790205, 'timing/model': 1.301176039967686, 'learning_rate/qwen_vl_interface': 5.7236724629042595e-06, 'learning_rate/action_model': 5.7236724629042594e-05, 'epoch': 0.19}) +[2026-06-24 12:59:54,809][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.09506017714738846, 'train/grad_norm_pre_clip': 32.91254425048828, 'timing/data': 0.0003949741367250681, 'timing/model': 1.0629665269516408, 'learning_rate/qwen_vl_interface': 5.719865083128895e-06, 'learning_rate/action_model': 5.7198650831288944e-05, 'epoch': 0.19}) +[2026-06-24 12:59:55,876][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.28765615820884705, 'train/grad_norm_pre_clip': 92.50823211669922, 'timing/data': 0.00034608179703354836, 'timing/model': 1.0647088210098445, 'learning_rate/qwen_vl_interface': 5.716057398463179e-06, 'learning_rate/action_model': 5.7160573984631785e-05, 'epoch': 0.19}) +[2026-06-24 12:59:57,186][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.06776262819766998, 'train/grad_norm_pre_clip': 44.159481048583984, 'timing/data': 0.00032930984161794186, 'timing/model': 1.3082165529485792, 'learning_rate/qwen_vl_interface': 5.712249411377875e-06, 'learning_rate/action_model': 5.712249411377875e-05, 'epoch': 0.19}) +[2026-06-24 12:59:58,251][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.09757248312234879, 'train/grad_norm_pre_clip': 56.09177017211914, 'timing/data': 0.0003343401476740837, 'timing/model': 1.063944982830435, 'learning_rate/qwen_vl_interface': 5.708441124343948e-06, 'learning_rate/action_model': 5.708441124343947e-05, 'epoch': 0.19}) +[2026-06-24 12:59:59,317][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.14615032076835632, 'train/grad_norm_pre_clip': 92.00863647460938, 'timing/data': 0.00034380797296762466, 'timing/model': 1.063858391949907, 'learning_rate/qwen_vl_interface': 5.704632539832551e-06, 'learning_rate/action_model': 5.704632539832551e-05, 'epoch': 0.19}) +[2026-06-24 13:00:00,384][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.18811875581741333, 'train/grad_norm_pre_clip': 43.19356918334961, 'timing/data': 0.0003559298347681761, 'timing/model': 1.0657074667979032, 'learning_rate/qwen_vl_interface': 5.700823660315038e-06, 'learning_rate/action_model': 5.7008236603150375e-05, 'epoch': 0.19}) +[2026-06-24 13:00:01,451][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.11461922526359558, 'train/grad_norm_pre_clip': 34.64878845214844, 'timing/data': 0.0003533849958330393, 'timing/model': 1.0650979401543736, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.19}) +[2026-06-24 13:00:02,524][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.07126642018556595, 'train/grad_norm_pre_clip': 37.07186508178711, 'timing/data': 0.00038076797500252724, 'timing/model': 1.0711605560500175, 'learning_rate/qwen_vl_interface': 5.693205026148002e-06, 'learning_rate/action_model': 5.693205026148002e-05, 'epoch': 0.19}) +[2026-06-24 13:00:03,590][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.1265193372964859, 'train/grad_norm_pre_clip': 55.613555908203125, 'timing/data': 0.00035816500894725323, 'timing/model': 1.0649422998540103, 'learning_rate/qwen_vl_interface': 5.689395276442131e-06, 'learning_rate/action_model': 5.689395276442131e-05, 'epoch': 0.19}) +[2026-06-24 13:00:04,660][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.27093517780303955, 'train/grad_norm_pre_clip': 45.28232192993164, 'timing/data': 0.0003844350576400757, 'timing/model': 1.0677441719453782, 'learning_rate/qwen_vl_interface': 5.685585241617435e-06, 'learning_rate/action_model': 5.6855852416174346e-05, 'epoch': 0.19}) +[2026-06-24 13:00:05,731][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.23036183416843414, 'train/grad_norm_pre_clip': 53.77732467651367, 'timing/data': 0.0003405220340937376, 'timing/model': 1.0696886270307004, 'learning_rate/qwen_vl_interface': 5.681774924146205e-06, 'learning_rate/action_model': 5.6817749241462046e-05, 'epoch': 0.19}) +[2026-06-24 13:00:06,800][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.2356192171573639, 'train/grad_norm_pre_clip': 32.50358581542969, 'timing/data': 0.00035329489037394524, 'timing/model': 1.0675923679955304, 'learning_rate/qwen_vl_interface': 5.677964326500914e-06, 'learning_rate/action_model': 5.6779643265009144e-05, 'epoch': 0.19}) +[2026-06-24 13:00:07,872][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.012510445900261402, 'train/grad_norm_pre_clip': 13.897039413452148, 'timing/data': 0.00039208889938890934, 'timing/model': 1.0699985339306295, 'learning_rate/qwen_vl_interface': 5.674153451154218e-06, 'learning_rate/action_model': 5.674153451154218e-05, 'epoch': 0.19}) +[2026-06-24 13:00:08,951][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.28224214911460876, 'train/grad_norm_pre_clip': 45.374210357666016, 'timing/data': 0.0003307939041405916, 'timing/model': 1.0775852610822767, 'learning_rate/qwen_vl_interface': 5.670342300578954e-06, 'learning_rate/action_model': 5.670342300578954e-05, 'epoch': 0.19}) +[2026-06-24 13:00:10,020][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.252472460269928, 'train/grad_norm_pre_clip': 86.29046630859375, 'timing/data': 0.000367332948371768, 'timing/model': 1.0675330320373178, 'learning_rate/qwen_vl_interface': 5.666530877248134e-06, 'learning_rate/action_model': 5.666530877248134e-05, 'epoch': 0.19}) +[2026-06-24 13:00:11,090][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.09604878723621368, 'train/grad_norm_pre_clip': 44.931602478027344, 'timing/data': 0.0003294509369879961, 'timing/model': 1.0677973011042923, 'learning_rate/qwen_vl_interface': 5.6627191836349484e-06, 'learning_rate/action_model': 5.662719183634948e-05, 'epoch': 0.19}) +[2026-06-24 13:00:12,156][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.23011170327663422, 'train/grad_norm_pre_clip': 54.05080795288086, 'timing/data': 0.00035018986091017723, 'timing/model': 1.0646589500829577, 'learning_rate/qwen_vl_interface': 5.6589072222127665e-06, 'learning_rate/action_model': 5.658907222212766e-05, 'epoch': 0.19}) +[2026-06-24 13:00:13,223][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.0952102392911911, 'train/grad_norm_pre_clip': 34.44856643676758, 'timing/data': 0.00031889998354017735, 'timing/model': 1.0656488658860326, 'learning_rate/qwen_vl_interface': 5.655094995455128e-06, 'learning_rate/action_model': 5.655094995455128e-05, 'epoch': 0.19}) +[2026-06-24 13:00:14,293][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.0522945411503315, 'train/grad_norm_pre_clip': 26.804790496826172, 'timing/data': 0.000323339132592082, 'timing/model': 1.068119843956083, 'learning_rate/qwen_vl_interface': 5.6512825058357445e-06, 'learning_rate/action_model': 5.651282505835745e-05, 'epoch': 0.19}) +[2026-06-24 13:00:15,573][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.18051089346408844, 'train/grad_norm_pre_clip': 46.50490951538086, 'timing/data': 0.0003899249713867903, 'timing/model': 1.2777589270845056, 'learning_rate/qwen_vl_interface': 5.647469755828501e-06, 'learning_rate/action_model': 5.647469755828501e-05, 'epoch': 0.19}) +[2026-06-24 13:00:16,640][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.07212628424167633, 'train/grad_norm_pre_clip': 39.13514709472656, 'timing/data': 0.00031432206742465496, 'timing/model': 1.065841794013977, 'learning_rate/qwen_vl_interface': 5.643656747907446e-06, 'learning_rate/action_model': 5.643656747907445e-05, 'epoch': 0.19}) +[2026-06-24 13:00:17,711][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.02400454878807068, 'train/grad_norm_pre_clip': 8.121865272521973, 'timing/data': 0.00031403196044266224, 'timing/model': 1.06948739499785, 'learning_rate/qwen_vl_interface': 5.639843484546801e-06, 'learning_rate/action_model': 5.639843484546801e-05, 'epoch': 0.19}) +[2026-06-24 13:00:18,779][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.1567363739013672, 'train/grad_norm_pre_clip': 47.364654541015625, 'timing/data': 0.0003331177867949009, 'timing/model': 1.0664814829360694, 'learning_rate/qwen_vl_interface': 5.636029968220953e-06, 'learning_rate/action_model': 5.6360299682209525e-05, 'epoch': 0.19}) +[2026-06-24 13:00:19,852][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.05296392738819122, 'train/grad_norm_pre_clip': 22.63023567199707, 'timing/data': 0.00032797781750559807, 'timing/model': 1.0710884348955005, 'learning_rate/qwen_vl_interface': 5.632216201404449e-06, 'learning_rate/action_model': 5.632216201404449e-05, 'epoch': 0.19}) +[2026-06-24 13:00:20,929][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.03333773463964462, 'train/grad_norm_pre_clip': 17.951740264892578, 'timing/data': 0.00038479501381516457, 'timing/model': 1.0754562967922539, 'learning_rate/qwen_vl_interface': 5.628402186572002e-06, 'learning_rate/action_model': 5.6284021865720015e-05, 'epoch': 0.19}) +[2026-06-24 13:00:21,996][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.23537859320640564, 'train/grad_norm_pre_clip': 61.06656265258789, 'timing/data': 0.00046055810526013374, 'timing/model': 1.0655620761681348, 'learning_rate/qwen_vl_interface': 5.6245879261984835e-06, 'learning_rate/action_model': 5.6245879261984836e-05, 'epoch': 0.19}) +[2026-06-24 13:00:23,068][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.08411490172147751, 'train/grad_norm_pre_clip': 26.67073631286621, 'timing/data': 0.0004942519590258598, 'timing/model': 1.0699070941191167, 'learning_rate/qwen_vl_interface': 5.620773422758927e-06, 'learning_rate/action_model': 5.620773422758927e-05, 'epoch': 0.19}) +[2026-06-24 13:00:24,138][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.15543296933174133, 'train/grad_norm_pre_clip': 50.0105094909668, 'timing/data': 0.00032937992364168167, 'timing/model': 1.0679587149061263, 'learning_rate/qwen_vl_interface': 5.616958678728522e-06, 'learning_rate/action_model': 5.616958678728522e-05, 'epoch': 0.19}) +[2026-06-24 13:00:25,207][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.06060414016246796, 'train/grad_norm_pre_clip': 25.166852951049805, 'timing/data': 0.000338227953761816, 'timing/model': 1.0671491420362145, 'learning_rate/qwen_vl_interface': 5.6131436965826145e-06, 'learning_rate/action_model': 5.613143696582615e-05, 'epoch': 0.19}) +[2026-06-24 13:00:26,276][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.06416336447000504, 'train/grad_norm_pre_clip': 10.177538871765137, 'timing/data': 0.000324200838804245, 'timing/model': 1.0676838408689946, 'learning_rate/qwen_vl_interface': 5.609328478796705e-06, 'learning_rate/action_model': 5.6093284787967046e-05, 'epoch': 0.19}) +[2026-06-24 13:00:27,349][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.09434331208467484, 'train/grad_norm_pre_clip': 46.875450134277344, 'timing/data': 0.0003748559392988682, 'timing/model': 1.071340121096, 'learning_rate/qwen_vl_interface': 5.605513027846447e-06, 'learning_rate/action_model': 5.605513027846446e-05, 'epoch': 0.19}) +[2026-06-24 13:00:28,419][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.15637736022472382, 'train/grad_norm_pre_clip': 25.995153427124023, 'timing/data': 0.0003973690327256918, 'timing/model': 1.067762803984806, 'learning_rate/qwen_vl_interface': 5.601697346207644e-06, 'learning_rate/action_model': 5.601697346207644e-05, 'epoch': 0.19}) +[2026-06-24 13:00:29,487][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.03531388193368912, 'train/grad_norm_pre_clip': 34.14238739013672, 'timing/data': 0.0003680430818349123, 'timing/model': 1.0665813700761646, 'learning_rate/qwen_vl_interface': 5.597881436356252e-06, 'learning_rate/action_model': 5.597881436356252e-05, 'epoch': 0.19}) +[2026-06-24 13:00:30,553][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.47428491711616516, 'train/grad_norm_pre_clip': 88.64106750488281, 'timing/data': 0.0003203330561518669, 'timing/model': 1.06408910616301, 'learning_rate/qwen_vl_interface': 5.594065300768372e-06, 'learning_rate/action_model': 5.5940653007683716e-05, 'epoch': 0.19}) +[2026-06-24 13:00:31,623][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.03339812532067299, 'train/grad_norm_pre_clip': 25.73213768005371, 'timing/data': 0.00033606402575969696, 'timing/model': 1.0687164361588657, 'learning_rate/qwen_vl_interface': 5.590248941920255e-06, 'learning_rate/action_model': 5.5902489419202554e-05, 'epoch': 0.19}) +[2026-06-24 13:00:32,902][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.19273081421852112, 'train/grad_norm_pre_clip': 45.296607971191406, 'timing/data': 0.0003153339494019747, 'timing/model': 1.2775423331186175, 'learning_rate/qwen_vl_interface': 5.586432362288292e-06, 'learning_rate/action_model': 5.586432362288292e-05, 'epoch': 0.19}) +[2026-06-24 13:00:33,971][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.03892945870757103, 'train/grad_norm_pre_clip': 21.42327117919922, 'timing/data': 0.0003506310749799013, 'timing/model': 1.067096035927534, 'learning_rate/qwen_vl_interface': 5.582615564349024e-06, 'learning_rate/action_model': 5.582615564349024e-05, 'epoch': 0.19}) +[2026-06-24 13:00:35,043][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.007948593236505985, 'train/grad_norm_pre_clip': 4.337869644165039, 'timing/data': 0.0003340090624988079, 'timing/model': 1.0704994909465313, 'learning_rate/qwen_vl_interface': 5.578798550579126e-06, 'learning_rate/action_model': 5.578798550579125e-05, 'epoch': 0.19}) +[2026-06-24 13:00:36,114][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.014654206112027168, 'train/grad_norm_pre_clip': 7.566678524017334, 'timing/data': 0.0003472140524536371, 'timing/model': 1.0690936029423028, 'learning_rate/qwen_vl_interface': 5.5749813234554175e-06, 'learning_rate/action_model': 5.574981323455417e-05, 'epoch': 0.19}) +[2026-06-24 13:00:37,185][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.17952698469161987, 'train/grad_norm_pre_clip': 83.14237976074219, 'timing/data': 0.0005188270006328821, 'timing/model': 1.0691677150316536, 'learning_rate/qwen_vl_interface': 5.571163885454859e-06, 'learning_rate/action_model': 5.5711638854548583e-05, 'epoch': 0.19}) +[2026-06-24 13:00:38,254][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.05507393926382065, 'train/grad_norm_pre_clip': 25.401851654052734, 'timing/data': 0.0003544681239873171, 'timing/model': 1.0667342820670456, 'learning_rate/qwen_vl_interface': 5.567346239054541e-06, 'learning_rate/action_model': 5.567346239054541e-05, 'epoch': 0.19}) +[2026-06-24 13:00:39,327][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.055740341544151306, 'train/grad_norm_pre_clip': 40.345191955566406, 'timing/data': 0.000675383023917675, 'timing/model': 1.071473984979093, 'learning_rate/qwen_vl_interface': 5.563528386731692e-06, 'learning_rate/action_model': 5.5635283867316926e-05, 'epoch': 0.19}) +[2026-06-24 13:00:40,402][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.1322619765996933, 'train/grad_norm_pre_clip': 14.09656810760498, 'timing/data': 0.0003535458818078041, 'timing/model': 1.0731532201170921, 'learning_rate/qwen_vl_interface': 5.559710330963678e-06, 'learning_rate/action_model': 5.5597103309636775e-05, 'epoch': 0.19}) +[2026-06-24 13:00:41,477][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.07469376921653748, 'train/grad_norm_pre_clip': 24.522172927856445, 'timing/data': 0.00031876005232334137, 'timing/model': 1.0730617810040712, 'learning_rate/qwen_vl_interface': 5.555892074227993e-06, 'learning_rate/action_model': 5.555892074227993e-05, 'epoch': 0.19}) +[2026-06-24 13:00:42,549][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.05699259787797928, 'train/grad_norm_pre_clip': 207.08119201660156, 'timing/data': 0.0004527829587459564, 'timing/model': 1.0701380339451134, 'learning_rate/qwen_vl_interface': 5.552073619002263e-06, 'learning_rate/action_model': 5.552073619002263e-05, 'epoch': 0.19}) +[2026-06-24 13:00:43,621][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.026151910424232483, 'train/grad_norm_pre_clip': 13.964129447937012, 'timing/data': 0.00031998311169445515, 'timing/model': 1.0703570190817118, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.19}) +[2026-06-24 13:00:44,692][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.006000205408781767, 'train/grad_norm_pre_clip': 2.44327712059021, 'timing/data': 0.00031078583560884, 'timing/model': 1.0693481098860502, 'learning_rate/qwen_vl_interface': 5.544436122991806e-06, 'learning_rate/action_model': 5.544436122991806e-05, 'epoch': 0.19}) +[2026-06-24 13:00:45,763][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.06398351490497589, 'train/grad_norm_pre_clip': 70.35076904296875, 'timing/data': 0.00035619083791971207, 'timing/model': 1.0700641318690032, 'learning_rate/qwen_vl_interface': 5.54061708716297e-06, 'learning_rate/action_model': 5.540617087162969e-05, 'epoch': 0.19}) +[2026-06-24 13:00:46,834][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.09501753002405167, 'train/grad_norm_pre_clip': 50.04271697998047, 'timing/data': 0.00032083503901958466, 'timing/model': 1.0687935000751168, 'learning_rate/qwen_vl_interface': 5.536797862755859e-06, 'learning_rate/action_model': 5.5367978627558595e-05, 'epoch': 0.19}) +[2026-06-24 13:00:47,903][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.02756558731198311, 'train/grad_norm_pre_clip': 16.109045028686523, 'timing/data': 0.000293482793495059, 'timing/model': 1.0681050550192595, 'learning_rate/qwen_vl_interface': 5.53297845224873e-06, 'learning_rate/action_model': 5.5329784522487296e-05, 'epoch': 0.19}) +[2026-06-24 13:00:48,976][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.22066688537597656, 'train/grad_norm_pre_clip': 82.90858459472656, 'timing/data': 0.0003443090245127678, 'timing/model': 1.0704712648876011, 'learning_rate/qwen_vl_interface': 5.529158858119958e-06, 'learning_rate/action_model': 5.529158858119957e-05, 'epoch': 0.19}) +[2026-06-24 13:00:50,255][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.15262793004512787, 'train/grad_norm_pre_clip': 26.800052642822266, 'timing/data': 0.00034146406687796116, 'timing/model': 1.2777963080443442, 'learning_rate/qwen_vl_interface': 5.52533908284803e-06, 'learning_rate/action_model': 5.52533908284803e-05, 'epoch': 0.19}) +[2026-06-24 13:00:51,329][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.006479343399405479, 'train/grad_norm_pre_clip': 7.0006422996521, 'timing/data': 0.00031167687848210335, 'timing/model': 1.071974691003561, 'learning_rate/qwen_vl_interface': 5.521519128911561e-06, 'learning_rate/action_model': 5.521519128911561e-05, 'epoch': 0.19}) +[2026-06-24 13:00:52,402][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.030520321801304817, 'train/grad_norm_pre_clip': 26.357845306396484, 'timing/data': 0.0003305531572550535, 'timing/model': 1.0715479089412838, 'learning_rate/qwen_vl_interface': 5.5176989987892785e-06, 'learning_rate/action_model': 5.517698998789278e-05, 'epoch': 0.19}) +[2026-06-24 13:00:53,475][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.18520689010620117, 'train/grad_norm_pre_clip': 96.58084106445312, 'timing/data': 0.00038578687235713005, 'timing/model': 1.070995771093294, 'learning_rate/qwen_vl_interface': 5.513878694960021e-06, 'learning_rate/action_model': 5.513878694960021e-05, 'epoch': 0.19}) +[2026-06-24 13:00:54,544][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.07759414613246918, 'train/grad_norm_pre_clip': 48.15386199951172, 'timing/data': 0.0003712789621204138, 'timing/model': 1.0677969970274717, 'learning_rate/qwen_vl_interface': 5.510058219902744e-06, 'learning_rate/action_model': 5.5100582199027436e-05, 'epoch': 0.19}) +[2026-06-24 13:00:55,616][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.04745519533753395, 'train/grad_norm_pre_clip': 18.887243270874023, 'timing/data': 0.000316255958750844, 'timing/model': 1.0698154717683792, 'learning_rate/qwen_vl_interface': 5.506237576096512e-06, 'learning_rate/action_model': 5.506237576096511e-05, 'epoch': 0.19}) +[2026-06-24 13:00:56,685][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.049921367317438126, 'train/grad_norm_pre_clip': 28.99062728881836, 'timing/data': 0.0003534960560500622, 'timing/model': 1.0676657748408616, 'learning_rate/qwen_vl_interface': 5.502416766020497e-06, 'learning_rate/action_model': 5.502416766020496e-05, 'epoch': 0.19}) +[2026-06-24 13:00:57,758][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.03956654667854309, 'train/grad_norm_pre_clip': 21.398548126220703, 'timing/data': 0.0003208550624549389, 'timing/model': 1.071206466993317, 'learning_rate/qwen_vl_interface': 5.498595792153985e-06, 'learning_rate/action_model': 5.498595792153984e-05, 'epoch': 0.19}) +[2026-06-24 13:00:58,829][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.1761428564786911, 'train/grad_norm_pre_clip': 79.3814926147461, 'timing/data': 0.0005246680229902267, 'timing/model': 1.0689022589940578, 'learning_rate/qwen_vl_interface': 5.4947746569763595e-06, 'learning_rate/action_model': 5.494774656976359e-05, 'epoch': 0.19}) +[2026-06-24 13:00:59,902][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.09046492725610733, 'train/grad_norm_pre_clip': 65.78746795654297, 'timing/data': 0.00034139305353164673, 'timing/model': 1.0717763050924987, 'learning_rate/qwen_vl_interface': 5.490953362967119e-06, 'learning_rate/action_model': 5.4909533629671185e-05, 'epoch': 0.19}) +[2026-06-24 13:01:00,974][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.07620514184236526, 'train/grad_norm_pre_clip': 76.28485870361328, 'timing/data': 0.0003235188778489828, 'timing/model': 1.0700888889841735, 'learning_rate/qwen_vl_interface': 5.487131912605859e-06, 'learning_rate/action_model': 5.4871319126058584e-05, 'epoch': 0.19}) +[2026-06-24 13:01:02,046][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.261741042137146, 'train/grad_norm_pre_clip': 74.96003723144531, 'timing/data': 0.00033200602047145367, 'timing/model': 1.070537966908887, 'learning_rate/qwen_vl_interface': 5.48331030837227e-06, 'learning_rate/action_model': 5.4833103083722705e-05, 'epoch': 0.19}) +[2026-06-24 13:01:03,119][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.3949262499809265, 'train/grad_norm_pre_clip': 89.32572174072266, 'timing/data': 0.0003173879813402891, 'timing/model': 1.0713478759862483, 'learning_rate/qwen_vl_interface': 5.4794885527461596e-06, 'learning_rate/action_model': 5.479488552746159e-05, 'epoch': 0.19}) +[2026-06-24 13:01:04,201][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.2823074460029602, 'train/grad_norm_pre_clip': 55.17661666870117, 'timing/data': 0.0005160621367394924, 'timing/model': 1.0797419310547411, 'learning_rate/qwen_vl_interface': 5.4756666482074164e-06, 'learning_rate/action_model': 5.475666648207416e-05, 'epoch': 0.19}) +[2026-06-24 13:01:05,274][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.1617526113986969, 'train/grad_norm_pre_clip': 43.35612869262695, 'timing/data': 0.00031945109367370605, 'timing/model': 1.0714064701460302, 'learning_rate/qwen_vl_interface': 5.471844597236036e-06, 'learning_rate/action_model': 5.471844597236035e-05, 'epoch': 0.19}) +[2026-06-24 13:01:06,346][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.3586351275444031, 'train/grad_norm_pre_clip': 131.62525939941406, 'timing/data': 0.00041033304296433926, 'timing/model': 1.0699997628107667, 'learning_rate/qwen_vl_interface': 5.468022402312105e-06, 'learning_rate/action_model': 5.468022402312105e-05, 'epoch': 0.19}) +[2026-06-24 13:01:07,683][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.21751782298088074, 'train/grad_norm_pre_clip': 30.66189956665039, 'timing/data': 0.0003104349598288536, 'timing/model': 1.3360365580301732, 'learning_rate/qwen_vl_interface': 5.464200065915806e-06, 'learning_rate/action_model': 5.4642000659158054e-05, 'epoch': 0.19}) +[2026-06-24 13:01:08,755][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.18931366503238678, 'train/grad_norm_pre_clip': 39.92728805541992, 'timing/data': 0.00035929703153669834, 'timing/model': 1.0706733260303736, 'learning_rate/qwen_vl_interface': 5.460377590527407e-06, 'learning_rate/action_model': 5.460377590527407e-05, 'epoch': 0.19}) +[2026-06-24 13:01:09,836][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.018399307504296303, 'train/grad_norm_pre_clip': 27.195186614990234, 'timing/data': 0.00033827684819698334, 'timing/model': 1.07884028297849, 'learning_rate/qwen_vl_interface': 5.456554978627276e-06, 'learning_rate/action_model': 5.456554978627275e-05, 'epoch': 0.19}) +[2026-06-24 13:01:10,941][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.11886882036924362, 'train/grad_norm_pre_clip': 50.54631042480469, 'timing/data': 0.00046683987602591515, 'timing/model': 1.1029531098902225, 'learning_rate/qwen_vl_interface': 5.452732232695859e-06, 'learning_rate/action_model': 5.452732232695858e-05, 'epoch': 0.19}) +[2026-06-24 13:01:12,243][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.16799116134643555, 'train/grad_norm_pre_clip': 53.93311309814453, 'timing/data': 0.000388602027669549, 'timing/model': 1.3005705859977752, 'learning_rate/qwen_vl_interface': 5.4489093552136986e-06, 'learning_rate/action_model': 5.4489093552136984e-05, 'epoch': 0.19}) +[2026-06-24 13:01:13,314][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.0964074581861496, 'train/grad_norm_pre_clip': 60.405120849609375, 'timing/data': 0.00033666379749774933, 'timing/model': 1.069365780800581, 'learning_rate/qwen_vl_interface': 5.445086348661418e-06, 'learning_rate/action_model': 5.4450863486614177e-05, 'epoch': 0.19}) +[2026-06-24 13:01:14,384][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2000 +[2026-06-24 13:01:39,469][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2000 +[2026-06-24 13:01:39,470][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2000 +[2026-06-24 13:02:12,152][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2000 +[2026-06-24 13:06:00,186][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-06-24 13:06:00,187][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.1360180377960205, 'train/grad_norm_pre_clip': 50.614017486572266, 'eval/action_loss/seconds': 25.0837806800846, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.060856617987155914, 'eval/latency_2/loss': 0.060856617987155914, 'eval/flappy/loss': 0.060856617987155914, 'eval/flappy/latency_2/loss': 0.060856617987155914, 'eval/flap_precision': 0.9389110225763613, 'eval/flap_recall': 0.9592944369063772, 'eval/flap_f1': 0.9489932885906041, 'eval/flap_rate_model': 0.20916666666666667, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9489932885906042, 'eval/latency_2/flap_precision': 0.9389110225763613, 'eval/latency_2/flap_recall': 0.9592944369063772, 'eval/latency_2/flap_f1': 0.9489932885906041, 'eval/latency_2/flap_rate_model': 0.20916666666666667, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9489932885906042, 'eval/action_classification/data_wait_seconds': 0.32197144627571106, 'eval/action_classification/predict_seconds': 32.26560974121094, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 174.0400000000039, 'rl_games_eval/mid_train/mean_length': 1421.6, 'rl_games_eval/mid_train/std_reward': 141.63890143602075, 'rl_games_eval/mid_train/std_length': 1140.6557061620304, 'rl_games_eval/mid_train/macro_mean_reward': 174.0400000000039, 'rl_games_eval/mid_train/macro_mean_length': 1421.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 174.0400000000039, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1421.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 141.63890143602075, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1140.6557061620304, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 174.0400000000039, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00032476079650223255, 'timing/model': 1.0684755560941994, 'learning_rate/qwen_vl_interface': 5.4412632155197224e-06, 'learning_rate/action_model': 5.441263215519722e-05, 'epoch': 0.19}) +[2026-06-24 13:06:00,187][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_2000_state +[2026-06-24 13:06:00,191][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 13:06:03,484][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_2000_state/model.safetensors +[2026-06-24 13:06:11,458][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_2000_state/optimizer.bin +[2026-06-24 13:06:11,458][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_2000_state/sampler.bin +[2026-06-24 13:06:11,458][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-24 13:06:11,459][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-24 13:06:11,466][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-24 13:06:11,466][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 13:06:11,476][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 13:06:12,535][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.2915947139263153, 'train/grad_norm_pre_clip': 54.36164855957031, 'timing/data': 0.0004391968250274658, 'timing/model': 1.05733509385027, 'learning_rate/qwen_vl_interface': 5.437439958269402e-06, 'learning_rate/action_model': 5.4374399582694025e-05, 'epoch': 0.19}) +[2026-06-24 13:06:13,610][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.1372193694114685, 'train/grad_norm_pre_clip': 23.954267501831055, 'timing/data': 0.000515329884365201, 'timing/model': 1.0726751550100744, 'learning_rate/qwen_vl_interface': 5.433616579391331e-06, 'learning_rate/action_model': 5.433616579391331e-05, 'epoch': 0.19}) +[2026-06-24 13:06:14,670][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.052466318011283875, 'train/grad_norm_pre_clip': 23.390209197998047, 'timing/data': 0.0003970868419855833, 'timing/model': 1.058234048075974, 'learning_rate/qwen_vl_interface': 5.429793081366453e-06, 'learning_rate/action_model': 5.429793081366452e-05, 'epoch': 0.19}) +[2026-06-24 13:06:15,731][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.09356476366519928, 'train/grad_norm_pre_clip': 49.607872009277344, 'timing/data': 0.00032908888533711433, 'timing/model': 1.0598920050542802, 'learning_rate/qwen_vl_interface': 5.425969466675797e-06, 'learning_rate/action_model': 5.4259694666757965e-05, 'epoch': 0.19}) +[2026-06-24 13:06:16,793][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.018435968086123466, 'train/grad_norm_pre_clip': 9.191215515136719, 'timing/data': 0.0003341489937156439, 'timing/model': 1.0596856619231403, 'learning_rate/qwen_vl_interface': 5.4221457378004634e-06, 'learning_rate/action_model': 5.422145737800464e-05, 'epoch': 0.19}) +[2026-06-24 13:06:17,856][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.037691812962293625, 'train/grad_norm_pre_clip': 8.540836334228516, 'timing/data': 0.00032885908149182796, 'timing/model': 1.0617939070798457, 'learning_rate/qwen_vl_interface': 5.41832189722163e-06, 'learning_rate/action_model': 5.41832189722163e-05, 'epoch': 0.19}) +[2026-06-24 13:06:18,920][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.12088123708963394, 'train/grad_norm_pre_clip': 35.04911804199219, 'timing/data': 0.0003268758300691843, 'timing/model': 1.0626595518551767, 'learning_rate/qwen_vl_interface': 5.414497947420546e-06, 'learning_rate/action_model': 5.414497947420546e-05, 'epoch': 0.19}) +[2026-06-24 13:06:20,196][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.14424702525138855, 'train/grad_norm_pre_clip': 29.70470428466797, 'timing/data': 0.0003103951457887888, 'timing/model': 1.2741513329092413, 'learning_rate/qwen_vl_interface': 5.4106738908785295e-06, 'learning_rate/action_model': 5.4106738908785294e-05, 'epoch': 0.19}) +[2026-06-24 13:06:21,273][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.03503827750682831, 'train/grad_norm_pre_clip': 10.269340515136719, 'timing/data': 0.0003421138972043991, 'timing/model': 1.0751576588954777, 'learning_rate/qwen_vl_interface': 5.4068497300769684e-06, 'learning_rate/action_model': 5.406849730076968e-05, 'epoch': 0.19}) +[2026-06-24 13:06:22,342][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.22194963693618774, 'train/grad_norm_pre_clip': 18.637401580810547, 'timing/data': 0.00033498089760541916, 'timing/model': 1.0669953210745007, 'learning_rate/qwen_vl_interface': 5.403025467497323e-06, 'learning_rate/action_model': 5.4030254674973225e-05, 'epoch': 0.19}) +[2026-06-24 13:06:23,417][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.13131673634052277, 'train/grad_norm_pre_clip': 32.12334442138672, 'timing/data': 0.00036754203028976917, 'timing/model': 1.073263285215944, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 0.19}) +[2026-06-24 13:06:24,486][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.27598702907562256, 'train/grad_norm_pre_clip': 41.84177017211914, 'timing/data': 0.00034220400266349316, 'timing/model': 1.0682178540155292, 'learning_rate/qwen_vl_interface': 5.3953766469299204e-06, 'learning_rate/action_model': 5.3953766469299206e-05, 'epoch': 0.19}) +[2026-06-24 13:06:25,560][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.0765552893280983, 'train/grad_norm_pre_clip': 21.533262252807617, 'timing/data': 0.0003432461526244879, 'timing/model': 1.0720884159673005, 'learning_rate/qwen_vl_interface': 5.391552093905404e-06, 'learning_rate/action_model': 5.3915520939054034e-05, 'epoch': 0.19}) +[2026-06-24 13:06:26,638][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.047843191772699356, 'train/grad_norm_pre_clip': 19.11318016052246, 'timing/data': 0.0003525540232658386, 'timing/model': 1.076115150935948, 'learning_rate/qwen_vl_interface': 5.38772744902927e-06, 'learning_rate/action_model': 5.38772744902927e-05, 'epoch': 0.19}) +[2026-06-24 13:06:27,708][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.05783839896321297, 'train/grad_norm_pre_clip': 18.702144622802734, 'timing/data': 0.0003602178767323494, 'timing/model': 1.0687592199537903, 'learning_rate/qwen_vl_interface': 5.383902714783289e-06, 'learning_rate/action_model': 5.383902714783289e-05, 'epoch': 0.19}) +[2026-06-24 13:06:28,780][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.08598454296588898, 'train/grad_norm_pre_clip': 16.180734634399414, 'timing/data': 0.00031424197368323803, 'timing/model': 1.0700799219775945, 'learning_rate/qwen_vl_interface': 5.38007789364929e-06, 'learning_rate/action_model': 5.38007789364929e-05, 'epoch': 0.2}) +[2026-06-24 13:06:29,855][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.048915449529886246, 'train/grad_norm_pre_clip': 10.385080337524414, 'timing/data': 0.00035651191137731075, 'timing/model': 1.0735313950572163, 'learning_rate/qwen_vl_interface': 5.3762529881091595e-06, 'learning_rate/action_model': 5.376252988109159e-05, 'epoch': 0.2}) +[2026-06-24 13:06:30,928][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.08163350820541382, 'train/grad_norm_pre_clip': 30.782846450805664, 'timing/data': 0.00031743687577545643, 'timing/model': 1.0710111679509282, 'learning_rate/qwen_vl_interface': 5.372428000644835e-06, 'learning_rate/action_model': 5.372428000644835e-05, 'epoch': 0.2}) +[2026-06-24 13:06:32,002][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.08773057907819748, 'train/grad_norm_pre_clip': 30.290679931640625, 'timing/data': 0.0004451279528439045, 'timing/model': 1.0719258461613208, 'learning_rate/qwen_vl_interface': 5.36860293373831e-06, 'learning_rate/action_model': 5.368602933738309e-05, 'epoch': 0.2}) +[2026-06-24 13:06:33,073][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.3040703535079956, 'train/grad_norm_pre_clip': 100.6305160522461, 'timing/data': 0.0003189109265804291, 'timing/model': 1.0695732631720603, 'learning_rate/qwen_vl_interface': 5.364777789871629e-06, 'learning_rate/action_model': 5.364777789871629e-05, 'epoch': 0.2}) +[2026-06-24 13:06:34,144][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.027361765503883362, 'train/grad_norm_pre_clip': 11.015376091003418, 'timing/data': 0.00033246586099267006, 'timing/model': 1.0696230828762054, 'learning_rate/qwen_vl_interface': 5.360952571526886e-06, 'learning_rate/action_model': 5.360952571526886e-05, 'epoch': 0.2}) +[2026-06-24 13:06:35,216][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.16899272799491882, 'train/grad_norm_pre_clip': 64.54297637939453, 'timing/data': 0.00031752814538776875, 'timing/model': 1.070071232970804, 'learning_rate/qwen_vl_interface': 5.357127281186225e-06, 'learning_rate/action_model': 5.357127281186225e-05, 'epoch': 0.2}) +[2026-06-24 13:06:36,501][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.04323112219572067, 'train/grad_norm_pre_clip': 19.89927864074707, 'timing/data': 0.00032757711596786976, 'timing/model': 1.2834482400212437, 'learning_rate/qwen_vl_interface': 5.3533019213318325e-06, 'learning_rate/action_model': 5.3533019213318325e-05, 'epoch': 0.2}) +[2026-06-24 13:06:37,573][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.08250510692596436, 'train/grad_norm_pre_clip': 26.838571548461914, 'timing/data': 0.0004444571677595377, 'timing/model': 1.069912765873596, 'learning_rate/qwen_vl_interface': 5.349476494445947e-06, 'learning_rate/action_model': 5.3494764944459466e-05, 'epoch': 0.2}) +[2026-06-24 13:06:38,641][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.005407953169196844, 'train/grad_norm_pre_clip': 1.9515541791915894, 'timing/data': 0.0003180380444973707, 'timing/model': 1.0665879230946302, 'learning_rate/qwen_vl_interface': 5.345651003010842e-06, 'learning_rate/action_model': 5.345651003010842e-05, 'epoch': 0.2}) +[2026-06-24 13:06:39,718][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.05669022724032402, 'train/grad_norm_pre_clip': 32.07469940185547, 'timing/data': 0.0003259240183979273, 'timing/model': 1.0748004568740726, 'learning_rate/qwen_vl_interface': 5.34182544950884e-06, 'learning_rate/action_model': 5.3418254495088404e-05, 'epoch': 0.2}) +[2026-06-24 13:06:40,793][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.19781120121479034, 'train/grad_norm_pre_clip': 113.54476165771484, 'timing/data': 0.00036118016578257084, 'timing/model': 1.074136204086244, 'learning_rate/qwen_vl_interface': 5.337999836422303e-06, 'learning_rate/action_model': 5.3379998364223024e-05, 'epoch': 0.2}) +[2026-06-24 13:06:41,875][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.2008252739906311, 'train/grad_norm_pre_clip': 127.22462463378906, 'timing/data': 0.0003375951200723648, 'timing/model': 1.071638627909124, 'learning_rate/qwen_vl_interface': 5.3341741662336254e-06, 'learning_rate/action_model': 5.334174166233625e-05, 'epoch': 0.2}) +[2026-06-24 13:06:42,952][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.10423239320516586, 'train/grad_norm_pre_clip': 26.330114364624023, 'timing/data': 0.0003507810179144144, 'timing/model': 1.0744779990054667, 'learning_rate/qwen_vl_interface': 5.330348441425245e-06, 'learning_rate/action_model': 5.3303484414252446e-05, 'epoch': 0.2}) +[2026-06-24 13:06:44,030][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.019490910694003105, 'train/grad_norm_pre_clip': 20.939640045166016, 'timing/data': 0.00034600100480020046, 'timing/model': 1.0764368120580912, 'learning_rate/qwen_vl_interface': 5.326522664479634e-06, 'learning_rate/action_model': 5.326522664479634e-05, 'epoch': 0.2}) +[2026-06-24 13:06:45,101][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.018065964803099632, 'train/grad_norm_pre_clip': 22.155092239379883, 'timing/data': 0.00032303808256983757, 'timing/model': 1.0699407211504877, 'learning_rate/qwen_vl_interface': 5.3226968378792974e-06, 'learning_rate/action_model': 5.3226968378792974e-05, 'epoch': 0.2}) +[2026-06-24 13:06:46,173][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.011195017024874687, 'train/grad_norm_pre_clip': 7.290954113006592, 'timing/data': 0.00034733396023511887, 'timing/model': 1.0705700418911874, 'learning_rate/qwen_vl_interface': 5.31887096410677e-06, 'learning_rate/action_model': 5.3188709641067703e-05, 'epoch': 0.2}) +[2026-06-24 13:06:47,243][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.0706801563501358, 'train/grad_norm_pre_clip': 15.176116943359375, 'timing/data': 0.00035593006759881973, 'timing/model': 1.0678952529560775, 'learning_rate/qwen_vl_interface': 5.315045045644625e-06, 'learning_rate/action_model': 5.3150450456446244e-05, 'epoch': 0.2}) +[2026-06-24 13:06:48,318][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.2378183901309967, 'train/grad_norm_pre_clip': 73.78485870361328, 'timing/data': 0.00037628808058798313, 'timing/model': 1.0737409419380128, 'learning_rate/qwen_vl_interface': 5.311219084975453e-06, 'learning_rate/action_model': 5.311219084975453e-05, 'epoch': 0.2}) +[2026-06-24 13:06:49,392][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.23455750942230225, 'train/grad_norm_pre_clip': 60.87269973754883, 'timing/data': 0.00033068214543163776, 'timing/model': 1.0724948830902576, 'learning_rate/qwen_vl_interface': 5.307393084581883e-06, 'learning_rate/action_model': 5.307393084581883e-05, 'epoch': 0.2}) +[2026-06-24 13:06:50,465][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.014055092819035053, 'train/grad_norm_pre_clip': 9.066590309143066, 'timing/data': 0.00035946606658399105, 'timing/model': 1.0713821849785745, 'learning_rate/qwen_vl_interface': 5.303567046946563e-06, 'learning_rate/action_model': 5.303567046946563e-05, 'epoch': 0.2}) +[2026-06-24 13:06:51,538][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.0132686672732234, 'train/grad_norm_pre_clip': 39.72386932373047, 'timing/data': 0.0003291789907962084, 'timing/model': 1.0715532628819346, 'learning_rate/qwen_vl_interface': 5.299740974552168e-06, 'learning_rate/action_model': 5.299740974552167e-05, 'epoch': 0.2}) +[2026-06-24 13:06:52,815][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.0959111824631691, 'train/grad_norm_pre_clip': 122.32127380371094, 'timing/data': 0.0003089010715484619, 'timing/model': 1.2752109149005264, 'learning_rate/qwen_vl_interface': 5.295914869881395e-06, 'learning_rate/action_model': 5.2959148698813956e-05, 'epoch': 0.2}) +[2026-06-24 13:06:53,888][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.14017392694950104, 'train/grad_norm_pre_clip': 192.51881408691406, 'timing/data': 0.0003453099634498358, 'timing/model': 1.0715510710142553, 'learning_rate/qwen_vl_interface': 5.2920887354169605e-06, 'learning_rate/action_model': 5.29208873541696e-05, 'epoch': 0.2}) +[2026-06-24 13:06:54,960][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.08857960253953934, 'train/grad_norm_pre_clip': 35.549705505371094, 'timing/data': 0.00039078504778444767, 'timing/model': 1.069988687057048, 'learning_rate/qwen_vl_interface': 5.288262573641602e-06, 'learning_rate/action_model': 5.2882625736416024e-05, 'epoch': 0.2}) +[2026-06-24 13:06:56,034][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.14151062071323395, 'train/grad_norm_pre_clip': 62.661190032958984, 'timing/data': 0.00031858007423579693, 'timing/model': 1.0719562259037048, 'learning_rate/qwen_vl_interface': 5.284436387038076e-06, 'learning_rate/action_model': 5.2844363870380764e-05, 'epoch': 0.2}) +[2026-06-24 13:06:57,107][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.04455503076314926, 'train/grad_norm_pre_clip': 71.47374725341797, 'timing/data': 0.00032907910645008087, 'timing/model': 1.071236778050661, 'learning_rate/qwen_vl_interface': 5.280610178089154e-06, 'learning_rate/action_model': 5.280610178089154e-05, 'epoch': 0.2}) +[2026-06-24 13:06:58,182][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.01945674978196621, 'train/grad_norm_pre_clip': 11.608587265014648, 'timing/data': 0.00032487185671925545, 'timing/model': 1.0741492048837245, 'learning_rate/qwen_vl_interface': 5.276783949277618e-06, 'learning_rate/action_model': 5.276783949277617e-05, 'epoch': 0.2}) +[2026-06-24 13:06:59,258][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.2063412368297577, 'train/grad_norm_pre_clip': 105.58214569091797, 'timing/data': 0.0003254420589655638, 'timing/model': 1.073939897119999, 'learning_rate/qwen_vl_interface': 5.2729577030862675e-06, 'learning_rate/action_model': 5.272957703086268e-05, 'epoch': 0.2}) +[2026-06-24 13:07:00,336][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.07227811962366104, 'train/grad_norm_pre_clip': 101.66299438476562, 'timing/data': 0.000332915922626853, 'timing/model': 1.076144602848217, 'learning_rate/qwen_vl_interface': 5.2691314419979125e-06, 'learning_rate/action_model': 5.269131441997912e-05, 'epoch': 0.2}) +[2026-06-24 13:07:01,411][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.00958376843482256, 'train/grad_norm_pre_clip': 25.537595748901367, 'timing/data': 0.00032768608070909977, 'timing/model': 1.0736898190807551, 'learning_rate/qwen_vl_interface': 5.265305168495374e-06, 'learning_rate/action_model': 5.265305168495373e-05, 'epoch': 0.2}) +[2026-06-24 13:07:02,486][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.048871736973524094, 'train/grad_norm_pre_clip': 26.450069427490234, 'timing/data': 0.0003218159545212984, 'timing/model': 1.0735889761708677, 'learning_rate/qwen_vl_interface': 5.261478885061476e-06, 'learning_rate/action_model': 5.261478885061476e-05, 'epoch': 0.2}) +[2026-06-24 13:07:03,559][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.13762645423412323, 'train/grad_norm_pre_clip': 53.6436767578125, 'timing/data': 0.00031763804145157337, 'timing/model': 1.0714240381494164, 'learning_rate/qwen_vl_interface': 5.257652594179055e-06, 'learning_rate/action_model': 5.257652594179055e-05, 'epoch': 0.2}) +[2026-06-24 13:07:04,633][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.09926256537437439, 'train/grad_norm_pre_clip': 34.10256576538086, 'timing/data': 0.00032674497924745083, 'timing/model': 1.0723039200529456, 'learning_rate/qwen_vl_interface': 5.253826298330951e-06, 'learning_rate/action_model': 5.2538262983309504e-05, 'epoch': 0.2}) +[2026-06-24 13:07:05,706][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.217336505651474, 'train/grad_norm_pre_clip': 110.7043685913086, 'timing/data': 0.00036557787097990513, 'timing/model': 1.0718170600011945, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.2}) +[2026-06-24 13:07:06,783][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.0022348465863615274, 'train/grad_norm_pre_clip': 5.6610870361328125, 'timing/data': 0.00031677610240876675, 'timing/model': 1.075181398075074, 'learning_rate/qwen_vl_interface': 5.246173701669052e-06, 'learning_rate/action_model': 5.2461737016690514e-05, 'epoch': 0.2}) +[2026-06-24 13:07:07,855][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.15489408373832703, 'train/grad_norm_pre_clip': 47.63246154785156, 'timing/data': 0.0003211339935660362, 'timing/model': 1.0700931078754365, 'learning_rate/qwen_vl_interface': 5.242347405820946e-06, 'learning_rate/action_model': 5.242347405820945e-05, 'epoch': 0.2}) +[2026-06-24 13:07:09,223][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.033592674881219864, 'train/grad_norm_pre_clip': 75.23162841796875, 'timing/data': 0.00034893699921667576, 'timing/model': 1.3665337609127164, 'learning_rate/qwen_vl_interface': 5.238521114938525e-06, 'learning_rate/action_model': 5.238521114938525e-05, 'epoch': 0.2}) +[2026-06-24 13:07:10,296][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.30801498889923096, 'train/grad_norm_pre_clip': 95.98286437988281, 'timing/data': 0.0003478950820863247, 'timing/model': 1.0718164630234241, 'learning_rate/qwen_vl_interface': 5.234694831504628e-06, 'learning_rate/action_model': 5.234694831504627e-05, 'epoch': 0.2}) +[2026-06-24 13:07:11,366][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.3197941482067108, 'train/grad_norm_pre_clip': 104.30638885498047, 'timing/data': 0.00032532308250665665, 'timing/model': 1.0681442869827151, 'learning_rate/qwen_vl_interface': 5.230868558002088e-06, 'learning_rate/action_model': 5.2308685580020876e-05, 'epoch': 0.2}) +[2026-06-24 13:07:12,443][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.07059137523174286, 'train/grad_norm_pre_clip': 31.932037353515625, 'timing/data': 0.00033958000130951405, 'timing/model': 1.075570431072265, 'learning_rate/qwen_vl_interface': 5.2270422969137345e-06, 'learning_rate/action_model': 5.227042296913735e-05, 'epoch': 0.2}) +[2026-06-24 13:07:13,515][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.12161145359277725, 'train/grad_norm_pre_clip': 52.17084503173828, 'timing/data': 0.0003199819475412369, 'timing/model': 1.070342387072742, 'learning_rate/qwen_vl_interface': 5.223216050722384e-06, 'learning_rate/action_model': 5.2232160507223835e-05, 'epoch': 0.2}) +[2026-06-24 13:07:14,588][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.22707879543304443, 'train/grad_norm_pre_clip': 115.47087097167969, 'timing/data': 0.0003090919926762581, 'timing/model': 1.0715577211230993, 'learning_rate/qwen_vl_interface': 5.219389821910848e-06, 'learning_rate/action_model': 5.219389821910849e-05, 'epoch': 0.2}) +[2026-06-24 13:07:15,668][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.0889480710029602, 'train/grad_norm_pre_clip': 33.09376907348633, 'timing/data': 0.0005662059411406517, 'timing/model': 1.0771756011527032, 'learning_rate/qwen_vl_interface': 5.215563612961924e-06, 'learning_rate/action_model': 5.2155636129619234e-05, 'epoch': 0.2}) +[2026-06-24 13:07:16,744][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.09945445507764816, 'train/grad_norm_pre_clip': 33.989566802978516, 'timing/data': 0.0004112939350306988, 'timing/model': 1.0742568438872695, 'learning_rate/qwen_vl_interface': 5.211737426358399e-06, 'learning_rate/action_model': 5.211737426358399e-05, 'epoch': 0.2}) +[2026-06-24 13:07:17,821][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.2258564680814743, 'train/grad_norm_pre_clip': 62.69244384765625, 'timing/data': 0.0003171670250594616, 'timing/model': 1.0751215380150825, 'learning_rate/qwen_vl_interface': 5.207911264583042e-06, 'learning_rate/action_model': 5.207911264583042e-05, 'epoch': 0.2}) +[2026-06-24 13:07:18,900][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.023973144590854645, 'train/grad_norm_pre_clip': 10.995674133300781, 'timing/data': 0.00031976099126040936, 'timing/model': 1.0772565419320017, 'learning_rate/qwen_vl_interface': 5.2040851301186066e-06, 'learning_rate/action_model': 5.204085130118607e-05, 'epoch': 0.2}) +[2026-06-24 13:07:19,978][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.11776641011238098, 'train/grad_norm_pre_clip': 64.94769287109375, 'timing/data': 0.00031832000240683556, 'timing/model': 1.0764891940634698, 'learning_rate/qwen_vl_interface': 5.200259025447835e-06, 'learning_rate/action_model': 5.2002590254478345e-05, 'epoch': 0.2}) +[2026-06-24 13:07:21,055][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.05476924404501915, 'train/grad_norm_pre_clip': 60.20384979248047, 'timing/data': 0.000300074927508831, 'timing/model': 1.0753671990241855, 'learning_rate/qwen_vl_interface': 5.196432953053437e-06, 'learning_rate/action_model': 5.196432953053437e-05, 'epoch': 0.2}) +[2026-06-24 13:07:22,134][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.23922064900398254, 'train/grad_norm_pre_clip': 69.36795043945312, 'timing/data': 0.0003193709999322891, 'timing/model': 1.0778343980200589, 'learning_rate/qwen_vl_interface': 5.192606915418118e-06, 'learning_rate/action_model': 5.192606915418118e-05, 'epoch': 0.2}) +[2026-06-24 13:07:23,212][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.09132469445466995, 'train/grad_norm_pre_clip': 70.06304931640625, 'timing/data': 0.0003220259677618742, 'timing/model': 1.0763149349950254, 'learning_rate/qwen_vl_interface': 5.1887809150245475e-06, 'learning_rate/action_model': 5.188780915024547e-05, 'epoch': 0.2}) +[2026-06-24 13:07:24,502][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.08811287581920624, 'train/grad_norm_pre_clip': 55.73222351074219, 'timing/data': 0.00032622390426695347, 'timing/model': 1.2882266521919519, 'learning_rate/qwen_vl_interface': 5.184954954355377e-06, 'learning_rate/action_model': 5.184954954355377e-05, 'epoch': 0.2}) +[2026-06-24 13:07:25,576][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.10089173913002014, 'train/grad_norm_pre_clip': 46.97401428222656, 'timing/data': 0.00031756795942783356, 'timing/model': 1.0725240630563349, 'learning_rate/qwen_vl_interface': 5.18112903589323e-06, 'learning_rate/action_model': 5.1811290358932294e-05, 'epoch': 0.2}) +[2026-06-24 13:07:26,651][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.016618242487311363, 'train/grad_norm_pre_clip': 5.888832092285156, 'timing/data': 0.000361880986019969, 'timing/model': 1.0737612671218812, 'learning_rate/qwen_vl_interface': 5.177303162120704e-06, 'learning_rate/action_model': 5.177303162120704e-05, 'epoch': 0.2}) +[2026-06-24 13:07:27,726][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.0144022386521101, 'train/grad_norm_pre_clip': 12.618510246276855, 'timing/data': 0.00035256403498351574, 'timing/model': 1.0731844890397042, 'learning_rate/qwen_vl_interface': 5.173477335520367e-06, 'learning_rate/action_model': 5.173477335520367e-05, 'epoch': 0.2}) +[2026-06-24 13:07:28,803][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.15523715317249298, 'train/grad_norm_pre_clip': 44.10783386230469, 'timing/data': 0.0003610900603234768, 'timing/model': 1.0757708731107414, 'learning_rate/qwen_vl_interface': 5.1696515585747555e-06, 'learning_rate/action_model': 5.169651558574755e-05, 'epoch': 0.2}) +[2026-06-24 13:07:29,880][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.07562205940485, 'train/grad_norm_pre_clip': 18.030609130859375, 'timing/data': 0.0003576730377972126, 'timing/model': 1.074990320019424, 'learning_rate/qwen_vl_interface': 5.1658258337663765e-06, 'learning_rate/action_model': 5.165825833766377e-05, 'epoch': 0.2}) +[2026-06-24 13:07:30,957][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.015727799385786057, 'train/grad_norm_pre_clip': 14.05667781829834, 'timing/data': 0.00034078117460012436, 'timing/model': 1.0758675979450345, 'learning_rate/qwen_vl_interface': 5.1620001635777e-06, 'learning_rate/action_model': 5.1620001635776994e-05, 'epoch': 0.2}) +[2026-06-24 13:07:32,034][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.006688386667519808, 'train/grad_norm_pre_clip': 2.891664505004883, 'timing/data': 0.0003102840855717659, 'timing/model': 1.075565816136077, 'learning_rate/qwen_vl_interface': 5.158174550491161e-06, 'learning_rate/action_model': 5.158174550491161e-05, 'epoch': 0.2}) +[2026-06-24 13:07:33,112][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.04765647277235985, 'train/grad_norm_pre_clip': 28.479320526123047, 'timing/data': 0.00033697416074573994, 'timing/model': 1.0756345218978822, 'learning_rate/qwen_vl_interface': 5.15434899698916e-06, 'learning_rate/action_model': 5.15434899698916e-05, 'epoch': 0.2}) +[2026-06-24 13:07:34,184][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.027300508692860603, 'train/grad_norm_pre_clip': 29.0267276763916, 'timing/data': 0.00030917185358703136, 'timing/model': 1.0706747528165579, 'learning_rate/qwen_vl_interface': 5.1505235055540545e-06, 'learning_rate/action_model': 5.1505235055540545e-05, 'epoch': 0.2}) +[2026-06-24 13:07:35,263][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.15287227928638458, 'train/grad_norm_pre_clip': 33.50110626220703, 'timing/data': 0.00031383102759718895, 'timing/model': 1.0772614870220423, 'learning_rate/qwen_vl_interface': 5.146698078668169e-06, 'learning_rate/action_model': 5.1466980786681686e-05, 'epoch': 0.2}) +[2026-06-24 13:07:36,341][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.07193947583436966, 'train/grad_norm_pre_clip': 27.717761993408203, 'timing/data': 0.00032236590050160885, 'timing/model': 1.076559942914173, 'learning_rate/qwen_vl_interface': 5.1428727188137774e-06, 'learning_rate/action_model': 5.142872718813777e-05, 'epoch': 0.2}) +[2026-06-24 13:07:37,420][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.015642918646335602, 'train/grad_norm_pre_clip': 8.387340545654297, 'timing/data': 0.00036112009547650814, 'timing/model': 1.0779228440951556, 'learning_rate/qwen_vl_interface': 5.139047428473115e-06, 'learning_rate/action_model': 5.1390474284731146e-05, 'epoch': 0.2}) +[2026-06-24 13:07:38,496][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.16390225291252136, 'train/grad_norm_pre_clip': 62.4650764465332, 'timing/data': 0.00032965000718832016, 'timing/model': 1.0737532440107316, 'learning_rate/qwen_vl_interface': 5.135222210128373e-06, 'learning_rate/action_model': 5.1352222101283734e-05, 'epoch': 0.2}) +[2026-06-24 13:07:39,572][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.12951213121414185, 'train/grad_norm_pre_clip': 60.907318115234375, 'timing/data': 0.0003773199860006571, 'timing/model': 1.0750251340214163, 'learning_rate/qwen_vl_interface': 5.131397066261691e-06, 'learning_rate/action_model': 5.131397066261691e-05, 'epoch': 0.2}) +[2026-06-24 13:07:40,648][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.13087573647499084, 'train/grad_norm_pre_clip': 62.26322937011719, 'timing/data': 0.00040606502443552017, 'timing/model': 1.074152547866106, 'learning_rate/qwen_vl_interface': 5.1275719993551664e-06, 'learning_rate/action_model': 5.127571999355166e-05, 'epoch': 0.2}) +[2026-06-24 13:07:41,727][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.03956697881221771, 'train/grad_norm_pre_clip': 11.511393547058105, 'timing/data': 0.0003419530112296343, 'timing/model': 1.077080394839868, 'learning_rate/qwen_vl_interface': 5.1237470118908425e-06, 'learning_rate/action_model': 5.1237470118908426e-05, 'epoch': 0.2}) +[2026-06-24 13:07:42,806][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.24561761319637299, 'train/grad_norm_pre_clip': 98.28350830078125, 'timing/data': 0.0003409620840102434, 'timing/model': 1.077725525945425, 'learning_rate/qwen_vl_interface': 5.11992210635071e-06, 'learning_rate/action_model': 5.1199221063507095e-05, 'epoch': 0.2}) +[2026-06-24 13:07:44,178][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.16905950009822845, 'train/grad_norm_pre_clip': 90.89604187011719, 'timing/data': 0.00035127089358866215, 'timing/model': 1.3696306208148599, 'learning_rate/qwen_vl_interface': 5.116097285216712e-06, 'learning_rate/action_model': 5.116097285216712e-05, 'epoch': 0.2}) +[2026-06-24 13:07:45,255][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.13510501384735107, 'train/grad_norm_pre_clip': 42.39836502075195, 'timing/data': 0.0003078191075474024, 'timing/model': 1.0763487359508872, 'learning_rate/qwen_vl_interface': 5.112272550970732e-06, 'learning_rate/action_model': 5.1122725509707313e-05, 'epoch': 0.2}) +[2026-06-24 13:07:46,331][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.19721496105194092, 'train/grad_norm_pre_clip': 60.76218032836914, 'timing/data': 0.0003196920733898878, 'timing/model': 1.07423455407843, 'learning_rate/qwen_vl_interface': 5.108447906094598e-06, 'learning_rate/action_model': 5.108447906094598e-05, 'epoch': 0.2}) +[2026-06-24 13:07:47,407][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.10027878731489182, 'train/grad_norm_pre_clip': 39.22470474243164, 'timing/data': 0.0003230778966099024, 'timing/model': 1.073877909919247, 'learning_rate/qwen_vl_interface': 5.10462335307008e-06, 'learning_rate/action_model': 5.104623353070079e-05, 'epoch': 0.2}) +[2026-06-24 13:07:48,489][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.05124002322554588, 'train/grad_norm_pre_clip': 58.76316833496094, 'timing/data': 0.00034141214564442635, 'timing/model': 1.0805045571178198, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 0.2}) +[2026-06-24 13:07:49,568][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.02749868854880333, 'train/grad_norm_pre_clip': 23.3171329498291, 'timing/data': 0.00031514279544353485, 'timing/model': 1.077445956878364, 'learning_rate/qwen_vl_interface': 5.09697453250268e-06, 'learning_rate/action_model': 5.0969745325026793e-05, 'epoch': 0.2}) +[2026-06-24 13:07:50,648][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.03653557226061821, 'train/grad_norm_pre_clip': 25.999187469482422, 'timing/data': 0.00033374782651662827, 'timing/model': 1.077821291051805, 'learning_rate/qwen_vl_interface': 5.093150269923032e-06, 'learning_rate/action_model': 5.093150269923032e-05, 'epoch': 0.2}) +[2026-06-24 13:07:51,727][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.03886602818965912, 'train/grad_norm_pre_clip': 21.808095932006836, 'timing/data': 0.00035181199200451374, 'timing/model': 1.0774387379642576, 'learning_rate/qwen_vl_interface': 5.089326109121472e-06, 'learning_rate/action_model': 5.089326109121472e-05, 'epoch': 0.2}) +[2026-06-24 13:07:52,805][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.3095073401927948, 'train/grad_norm_pre_clip': 61.433616638183594, 'timing/data': 0.00037116906605660915, 'timing/model': 1.076612239005044, 'learning_rate/qwen_vl_interface': 5.085502052579455e-06, 'learning_rate/action_model': 5.085502052579455e-05, 'epoch': 0.2}) +[2026-06-24 13:07:53,881][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.05800755321979523, 'train/grad_norm_pre_clip': 19.567581176757812, 'timing/data': 0.00034107198007404804, 'timing/model': 1.0740403241943568, 'learning_rate/qwen_vl_interface': 5.08167810277837e-06, 'learning_rate/action_model': 5.08167810277837e-05, 'epoch': 0.2}) +[2026-06-24 13:07:54,958][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.056795794516801834, 'train/grad_norm_pre_clip': 48.640724182128906, 'timing/data': 0.00032311887480318546, 'timing/model': 1.0754163910169154, 'learning_rate/qwen_vl_interface': 5.0778542621995385e-06, 'learning_rate/action_model': 5.0778542621995385e-05, 'epoch': 0.2}) +[2026-06-24 13:07:56,031][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.016365280374884605, 'train/grad_norm_pre_clip': 11.34349250793457, 'timing/data': 0.0003076489083468914, 'timing/model': 1.0720519148744643, 'learning_rate/qwen_vl_interface': 5.074030533324205e-06, 'learning_rate/action_model': 5.0740305333242046e-05, 'epoch': 0.2}) +[2026-06-24 13:07:57,107][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.10593415796756744, 'train/grad_norm_pre_clip': 20.09864616394043, 'timing/data': 0.00031552393920719624, 'timing/model': 1.0747829920146614, 'learning_rate/qwen_vl_interface': 5.070206918633549e-06, 'learning_rate/action_model': 5.070206918633549e-05, 'epoch': 0.2}) +[2026-06-24 13:07:58,186][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.025380291044712067, 'train/grad_norm_pre_clip': 22.218631744384766, 'timing/data': 0.000327827176079154, 'timing/model': 1.0770426609087735, 'learning_rate/qwen_vl_interface': 5.066383420608669e-06, 'learning_rate/action_model': 5.066383420608669e-05, 'epoch': 0.2}) +[2026-06-24 13:07:59,262][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.16098159551620483, 'train/grad_norm_pre_clip': 32.55805206298828, 'timing/data': 0.00032541295513510704, 'timing/model': 1.07429383485578, 'learning_rate/qwen_vl_interface': 5.062560041730598e-06, 'learning_rate/action_model': 5.062560041730597e-05, 'epoch': 0.2}) +[2026-06-24 13:08:00,342][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.05080593377351761, 'train/grad_norm_pre_clip': 47.223148345947266, 'timing/data': 0.0003151930868625641, 'timing/model': 1.0777885881252587, 'learning_rate/qwen_vl_interface': 5.0587367844802795e-06, 'learning_rate/action_model': 5.058736784480279e-05, 'epoch': 0.2}) +[2026-06-24 13:08:01,422][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.21380941569805145, 'train/grad_norm_pre_clip': 53.027896881103516, 'timing/data': 0.00035071116872131824, 'timing/model': 1.0789511329494417, 'learning_rate/qwen_vl_interface': 5.054913651338583e-06, 'learning_rate/action_model': 5.054913651338583e-05, 'epoch': 0.2}) +[2026-06-24 13:08:02,735][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.05790426582098007, 'train/grad_norm_pre_clip': 40.87947463989258, 'timing/data': 0.0003203931264579296, 'timing/model': 1.3116670390591025, 'learning_rate/qwen_vl_interface': 5.0510906447863025e-06, 'learning_rate/action_model': 5.051090644786303e-05, 'epoch': 0.2}) +[2026-06-24 13:08:03,813][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.1523183137178421, 'train/grad_norm_pre_clip': 60.30034637451172, 'timing/data': 0.0003226881381124258, 'timing/model': 1.0760231351014227, 'learning_rate/qwen_vl_interface': 5.047267767304141e-06, 'learning_rate/action_model': 5.0472677673041415e-05, 'epoch': 0.2}) +[2026-06-24 13:08:04,887][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.10453978180885315, 'train/grad_norm_pre_clip': 47.93710708618164, 'timing/data': 0.00034337607212364674, 'timing/model': 1.0726212090812624, 'learning_rate/qwen_vl_interface': 5.043445021372726e-06, 'learning_rate/action_model': 5.043445021372726e-05, 'epoch': 0.2}) +[2026-06-24 13:08:05,962][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.04237000644207001, 'train/grad_norm_pre_clip': 30.61048698425293, 'timing/data': 0.0003481549210846424, 'timing/model': 1.0728046111762524, 'learning_rate/qwen_vl_interface': 5.0396224094725955e-06, 'learning_rate/action_model': 5.0396224094725954e-05, 'epoch': 0.2}) +[2026-06-24 13:08:07,037][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.03792077675461769, 'train/grad_norm_pre_clip': 16.51043128967285, 'timing/data': 0.000320653198286891, 'timing/model': 1.0741266661789268, 'learning_rate/qwen_vl_interface': 5.035799934084196e-06, 'learning_rate/action_model': 5.035799934084196e-05, 'epoch': 0.2}) +[2026-06-24 13:08:08,112][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.06506049633026123, 'train/grad_norm_pre_clip': 37.566566467285156, 'timing/data': 0.00031588412821292877, 'timing/model': 1.0733520239591599, 'learning_rate/qwen_vl_interface': 5.031977597687896e-06, 'learning_rate/action_model': 5.0319775976878955e-05, 'epoch': 0.2}) +[2026-06-24 13:08:09,191][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.09190387278795242, 'train/grad_norm_pre_clip': 29.215391159057617, 'timing/data': 0.0003213349264115095, 'timing/model': 1.0772718589287251, 'learning_rate/qwen_vl_interface': 5.028155402763965e-06, 'learning_rate/action_model': 5.0281554027639645e-05, 'epoch': 0.2}) +[2026-06-24 13:08:10,269][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.1384035050868988, 'train/grad_norm_pre_clip': 27.751325607299805, 'timing/data': 0.00033517112024128437, 'timing/model': 1.0757878720760345, 'learning_rate/qwen_vl_interface': 5.024333351792584e-06, 'learning_rate/action_model': 5.024333351792584e-05, 'epoch': 0.2}) +[2026-06-24 13:08:11,342][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.10866691917181015, 'train/grad_norm_pre_clip': 32.140830993652344, 'timing/data': 0.000300966901704669, 'timing/model': 1.0720942881889641, 'learning_rate/qwen_vl_interface': 5.0205114472538415e-06, 'learning_rate/action_model': 5.0205114472538414e-05, 'epoch': 0.2}) +[2026-06-24 13:08:12,419][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.028228938579559326, 'train/grad_norm_pre_clip': 12.819554328918457, 'timing/data': 0.0003051140811294317, 'timing/model': 1.075207824120298, 'learning_rate/qwen_vl_interface': 5.01668969162773e-06, 'learning_rate/action_model': 5.016689691627729e-05, 'epoch': 0.2}) +[2026-06-24 13:08:13,494][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.047632813453674316, 'train/grad_norm_pre_clip': 17.871110916137695, 'timing/data': 0.00030639697797596455, 'timing/model': 1.0736299720592797, 'learning_rate/qwen_vl_interface': 5.012868087394145e-06, 'learning_rate/action_model': 5.012868087394145e-05, 'epoch': 0.2}) +[2026-06-24 13:08:14,574][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.07546227425336838, 'train/grad_norm_pre_clip': 26.895000457763672, 'timing/data': 0.00033060205169022083, 'timing/model': 1.0783222292084247, 'learning_rate/qwen_vl_interface': 5.009046637032881e-06, 'learning_rate/action_model': 5.009046637032881e-05, 'epoch': 0.2}) +[2026-06-24 13:08:15,649][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.033871691673994064, 'train/grad_norm_pre_clip': 22.977800369262695, 'timing/data': 0.00034065102227032185, 'timing/model': 1.073261781828478, 'learning_rate/qwen_vl_interface': 5.005225343023641e-06, 'learning_rate/action_model': 5.0052253430236406e-05, 'epoch': 0.2}) +[2026-06-24 13:08:16,725][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.1441747099161148, 'train/grad_norm_pre_clip': 74.22305297851562, 'timing/data': 0.0003393080551177263, 'timing/model': 1.0742201120592654, 'learning_rate/qwen_vl_interface': 5.001404207846017e-06, 'learning_rate/action_model': 5.001404207846017e-05, 'epoch': 0.2}) +[2026-06-24 13:08:17,804][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.032621126621961594, 'train/grad_norm_pre_clip': 19.993459701538086, 'timing/data': 0.0003180890344083309, 'timing/model': 1.0772841959260404, 'learning_rate/qwen_vl_interface': 4.9975832339795045e-06, 'learning_rate/action_model': 4.9975832339795044e-05, 'epoch': 0.2}) +[2026-06-24 13:08:18,884][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.1122785285115242, 'train/grad_norm_pre_clip': 39.58999252319336, 'timing/data': 0.0003237188793718815, 'timing/model': 1.078280617017299, 'learning_rate/qwen_vl_interface': 4.99376242390349e-06, 'learning_rate/action_model': 4.99376242390349e-05, 'epoch': 0.2}) +[2026-06-24 13:08:19,961][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.1436583399772644, 'train/grad_norm_pre_clip': 60.85132598876953, 'timing/data': 0.000325232045724988, 'timing/model': 1.0757492000702769, 'learning_rate/qwen_vl_interface': 4.989941780097258e-06, 'learning_rate/action_model': 4.9899417800972575e-05, 'epoch': 0.2}) +[2026-06-24 13:08:21,237][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.09785569459199905, 'train/grad_norm_pre_clip': 47.97318649291992, 'timing/data': 0.00030564493499696255, 'timing/model': 1.2741436620708555, 'learning_rate/qwen_vl_interface': 4.98612130503998e-06, 'learning_rate/action_model': 4.98612130503998e-05, 'epoch': 0.21}) +[2026-06-24 13:08:22,317][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.12902626395225525, 'train/grad_norm_pre_clip': 30.684085845947266, 'timing/data': 0.00032303715124726295, 'timing/model': 1.078001962043345, 'learning_rate/qwen_vl_interface': 4.982301001210722e-06, 'learning_rate/action_model': 4.982301001210721e-05, 'epoch': 0.21}) +[2026-06-24 13:08:23,397][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.12996849417686462, 'train/grad_norm_pre_clip': 45.474571228027344, 'timing/data': 0.0003108549863100052, 'timing/model': 1.0791649580933154, 'learning_rate/qwen_vl_interface': 4.97848087108844e-06, 'learning_rate/action_model': 4.97848087108844e-05, 'epoch': 0.21}) +[2026-06-24 13:08:24,475][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.1306261271238327, 'train/grad_norm_pre_clip': 49.800235748291016, 'timing/data': 0.0003256129566580057, 'timing/model': 1.07573074195534, 'learning_rate/qwen_vl_interface': 4.9746609171519725e-06, 'learning_rate/action_model': 4.974660917151973e-05, 'epoch': 0.21}) +[2026-06-24 13:08:25,554][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.15249311923980713, 'train/grad_norm_pre_clip': 34.898685455322266, 'timing/data': 0.0003129187971353531, 'timing/model': 1.0777682829648256, 'learning_rate/qwen_vl_interface': 4.970841141880044e-06, 'learning_rate/action_model': 4.970841141880044e-05, 'epoch': 0.21}) +[2026-06-24 13:08:26,631][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.17435647547245026, 'train/grad_norm_pre_clip': 47.56394577026367, 'timing/data': 0.0003443078603595495, 'timing/model': 1.0752271581441164, 'learning_rate/qwen_vl_interface': 4.967021547751271e-06, 'learning_rate/action_model': 4.967021547751271e-05, 'epoch': 0.21}) +[2026-06-24 13:08:27,710][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.09625169634819031, 'train/grad_norm_pre_clip': 35.43862533569336, 'timing/data': 0.0003279570955783129, 'timing/model': 1.0776212299242616, 'learning_rate/qwen_vl_interface': 4.963202137244141e-06, 'learning_rate/action_model': 4.963202137244141e-05, 'epoch': 0.21}) +[2026-06-24 13:08:28,790][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.1613142192363739, 'train/grad_norm_pre_clip': 49.98409652709961, 'timing/data': 0.0003156340681016445, 'timing/model': 1.0784088191576302, 'learning_rate/qwen_vl_interface': 4.959382912837032e-06, 'learning_rate/action_model': 4.9593829128370315e-05, 'epoch': 0.21}) +[2026-06-24 13:08:29,873][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.13031379878520966, 'train/grad_norm_pre_clip': 43.66888427734375, 'timing/data': 0.0003526641521602869, 'timing/model': 1.0814502900466323, 'learning_rate/qwen_vl_interface': 4.9555638770081965e-06, 'learning_rate/action_model': 4.955563877008196e-05, 'epoch': 0.21}) +[2026-06-24 13:08:30,956][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.09882338345050812, 'train/grad_norm_pre_clip': 32.705650329589844, 'timing/data': 0.0004128171131014824, 'timing/model': 1.0802899829577655, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 0.21}) +[2026-06-24 13:08:32,039][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.0677831768989563, 'train/grad_norm_pre_clip': 33.53369140625, 'timing/data': 0.00045397388748824596, 'timing/model': 1.081309838918969, 'learning_rate/qwen_vl_interface': 4.947926380997739e-06, 'learning_rate/action_model': 4.947926380997739e-05, 'epoch': 0.21}) +[2026-06-24 13:08:33,122][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.027331950142979622, 'train/grad_norm_pre_clip': 11.288768768310547, 'timing/data': 0.000321374973282218, 'timing/model': 1.0817356368061155, 'learning_rate/qwen_vl_interface': 4.9441079257720075e-06, 'learning_rate/action_model': 4.944107925772007e-05, 'epoch': 0.21}) +[2026-06-24 13:08:34,205][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.195075124502182, 'train/grad_norm_pre_clip': 38.48435974121094, 'timing/data': 0.0003889820072799921, 'timing/model': 1.0807645560707897, 'learning_rate/qwen_vl_interface': 4.940289669036323e-06, 'learning_rate/action_model': 4.940289669036323e-05, 'epoch': 0.21}) +[2026-06-24 13:08:35,285][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.008914285339415073, 'train/grad_norm_pre_clip': 4.189752101898193, 'timing/data': 0.00034455908462405205, 'timing/model': 1.0790026269387454, 'learning_rate/qwen_vl_interface': 4.936471613268309e-06, 'learning_rate/action_model': 4.936471613268309e-05, 'epoch': 0.21}) +[2026-06-24 13:08:36,366][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.11796697229146957, 'train/grad_norm_pre_clip': 24.718002319335938, 'timing/data': 0.00030469312332570553, 'timing/model': 1.07874600100331, 'learning_rate/qwen_vl_interface': 4.932653760945462e-06, 'learning_rate/action_model': 4.9326537609454616e-05, 'epoch': 0.21}) +[2026-06-24 13:08:37,446][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.11205117404460907, 'train/grad_norm_pre_clip': 71.07763671875, 'timing/data': 0.0003271459136158228, 'timing/model': 1.0783674560952932, 'learning_rate/qwen_vl_interface': 4.928836114545143e-06, 'learning_rate/action_model': 4.9288361145451434e-05, 'epoch': 0.21}) +[2026-06-24 13:08:38,733][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.022062938660383224, 'train/grad_norm_pre_clip': 6.114213466644287, 'timing/data': 0.0003297210205346346, 'timing/model': 1.2859117279294878, 'learning_rate/qwen_vl_interface': 4.925018676544583e-06, 'learning_rate/action_model': 4.925018676544582e-05, 'epoch': 0.21}) +[2026-06-24 13:08:39,816][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.09309236705303192, 'train/grad_norm_pre_clip': 33.499210357666016, 'timing/data': 0.0003208939451724291, 'timing/model': 1.081434000050649, 'learning_rate/qwen_vl_interface': 4.921201449420876e-06, 'learning_rate/action_model': 4.921201449420876e-05, 'epoch': 0.21}) +[2026-06-24 13:08:40,898][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.09911426901817322, 'train/grad_norm_pre_clip': 65.3062973022461, 'timing/data': 0.0003210240975022316, 'timing/model': 1.0803292379714549, 'learning_rate/qwen_vl_interface': 4.917384435650977e-06, 'learning_rate/action_model': 4.917384435650977e-05, 'epoch': 0.21}) +[2026-06-24 13:08:41,981][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.17851781845092773, 'train/grad_norm_pre_clip': 53.67753219604492, 'timing/data': 0.00032525183632969856, 'timing/model': 1.0809429988730699, 'learning_rate/qwen_vl_interface': 4.9135676377117085e-06, 'learning_rate/action_model': 4.9135676377117085e-05, 'epoch': 0.21}) +[2026-06-24 13:08:43,065][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.25295233726501465, 'train/grad_norm_pre_clip': 39.63936233520508, 'timing/data': 0.00036922493018209934, 'timing/model': 1.082566041033715, 'learning_rate/qwen_vl_interface': 4.909751058079747e-06, 'learning_rate/action_model': 4.909751058079747e-05, 'epoch': 0.21}) +[2026-06-24 13:08:44,147][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.013941377401351929, 'train/grad_norm_pre_clip': 9.11945629119873, 'timing/data': 0.0003509500529617071, 'timing/model': 1.080629392992705, 'learning_rate/qwen_vl_interface': 4.905934699231629e-06, 'learning_rate/action_model': 4.905934699231629e-05, 'epoch': 0.21}) +[2026-06-24 13:08:45,230][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.2452535182237625, 'train/grad_norm_pre_clip': 66.12852478027344, 'timing/data': 0.0003785230219364166, 'timing/model': 1.080747501924634, 'learning_rate/qwen_vl_interface': 4.902118563643751e-06, 'learning_rate/action_model': 4.902118563643751e-05, 'epoch': 0.21}) +[2026-06-24 13:08:46,311][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.11032982170581818, 'train/grad_norm_pre_clip': 51.60287857055664, 'timing/data': 0.0003166461829096079, 'timing/model': 1.079036959912628, 'learning_rate/qwen_vl_interface': 4.8983026537923565e-06, 'learning_rate/action_model': 4.898302653792357e-05, 'epoch': 0.21}) +[2026-06-24 13:08:47,392][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.247872456908226, 'train/grad_norm_pre_clip': 74.46571350097656, 'timing/data': 0.00031260796822607517, 'timing/model': 1.0801972500048578, 'learning_rate/qwen_vl_interface': 4.8944869721535544e-06, 'learning_rate/action_model': 4.8944869721535544e-05, 'epoch': 0.21}) +[2026-06-24 13:08:48,476][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.05288046598434448, 'train/grad_norm_pre_clip': 17.79936981201172, 'timing/data': 0.00035458686761558056, 'timing/model': 1.0819724171888083, 'learning_rate/qwen_vl_interface': 4.890671521203298e-06, 'learning_rate/action_model': 4.890671521203298e-05, 'epoch': 0.21}) +[2026-06-24 13:08:49,557][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.030506964772939682, 'train/grad_norm_pre_clip': 12.310571670532227, 'timing/data': 0.0003051040694117546, 'timing/model': 1.0792843960225582, 'learning_rate/qwen_vl_interface': 4.886856303417386e-06, 'learning_rate/action_model': 4.8868563034173856e-05, 'epoch': 0.21}) +[2026-06-24 13:08:50,639][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.12334119528532028, 'train/grad_norm_pre_clip': 64.42839050292969, 'timing/data': 0.00031548389233648777, 'timing/model': 1.0807900361251086, 'learning_rate/qwen_vl_interface': 4.883041321271479e-06, 'learning_rate/action_model': 4.8830413212714796e-05, 'epoch': 0.21}) +[2026-06-24 13:08:51,723][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.03736153244972229, 'train/grad_norm_pre_clip': 22.58741569519043, 'timing/data': 0.0003172771539539099, 'timing/model': 1.0817007271107286, 'learning_rate/qwen_vl_interface': 4.879226577241073e-06, 'learning_rate/action_model': 4.8792265772410735e-05, 'epoch': 0.21}) +[2026-06-24 13:08:52,801][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.06420058757066727, 'train/grad_norm_pre_clip': 19.199600219726562, 'timing/data': 0.00031002284958958626, 'timing/model': 1.0769271708559245, 'learning_rate/qwen_vl_interface': 4.875412073801517e-06, 'learning_rate/action_model': 4.875412073801517e-05, 'epoch': 0.21}) +[2026-06-24 13:08:53,881][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.020716045051813126, 'train/grad_norm_pre_clip': 11.992915153503418, 'timing/data': 0.0003184590023010969, 'timing/model': 1.0782467708922923, 'learning_rate/qwen_vl_interface': 4.871597813428e-06, 'learning_rate/action_model': 4.8715978134279997e-05, 'epoch': 0.21}) +[2026-06-24 13:08:54,966][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.008303392678499222, 'train/grad_norm_pre_clip': 3.8427374362945557, 'timing/data': 0.0003601768985390663, 'timing/model': 1.08336988883093, 'learning_rate/qwen_vl_interface': 4.867783798595552e-06, 'learning_rate/action_model': 4.8677837985955524e-05, 'epoch': 0.21}) +[2026-06-24 13:08:56,252][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.08487438410520554, 'train/grad_norm_pre_clip': 52.28982925415039, 'timing/data': 0.00032537197694182396, 'timing/model': 1.2846109201200306, 'learning_rate/qwen_vl_interface': 4.863970031779049e-06, 'learning_rate/action_model': 4.8639700317790486e-05, 'epoch': 0.21}) +[2026-06-24 13:08:57,339][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.02817039005458355, 'train/grad_norm_pre_clip': 22.756948471069336, 'timing/data': 0.0003694850020110607, 'timing/model': 1.0853162850253284, 'learning_rate/qwen_vl_interface': 4.8601565154531995e-06, 'learning_rate/action_model': 4.860156515453199e-05, 'epoch': 0.21}) +[2026-06-24 13:08:58,424][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.18696904182434082, 'train/grad_norm_pre_clip': 76.6679916381836, 'timing/data': 0.0003639049828052521, 'timing/model': 1.0834884531795979, 'learning_rate/qwen_vl_interface': 4.8563432520925555e-06, 'learning_rate/action_model': 4.856343252092556e-05, 'epoch': 0.21}) +[2026-06-24 13:08:59,510][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.15382152795791626, 'train/grad_norm_pre_clip': 82.01066589355469, 'timing/data': 0.0003989008255302906, 'timing/model': 1.0842472477816045, 'learning_rate/qwen_vl_interface': 4.852530244171502e-06, 'learning_rate/action_model': 4.852530244171502e-05, 'epoch': 0.21}) +[2026-06-24 13:09:00,595][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.10042640566825867, 'train/grad_norm_pre_clip': 94.78607940673828, 'timing/data': 0.0003395590465515852, 'timing/model': 1.083149969112128, 'learning_rate/qwen_vl_interface': 4.848717494164257e-06, 'learning_rate/action_model': 4.8487174941642564e-05, 'epoch': 0.21}) +[2026-06-24 13:09:01,679][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.2172069102525711, 'train/grad_norm_pre_clip': 71.09801483154297, 'timing/data': 0.000395825132727623, 'timing/model': 1.0820541149005294, 'learning_rate/qwen_vl_interface': 4.844905004544873e-06, 'learning_rate/action_model': 4.844905004544873e-05, 'epoch': 0.21}) +[2026-06-24 13:09:02,765][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.19553038477897644, 'train/grad_norm_pre_clip': 128.91603088378906, 'timing/data': 0.0003790329210460186, 'timing/model': 1.0844515629578382, 'learning_rate/qwen_vl_interface': 4.841092777787234e-06, 'learning_rate/action_model': 4.8410927777872336e-05, 'epoch': 0.21}) +[2026-06-24 13:09:03,849][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.18950751423835754, 'train/grad_norm_pre_clip': 42.310264587402344, 'timing/data': 0.00032549211755394936, 'timing/model': 1.08250295300968, 'learning_rate/qwen_vl_interface': 4.837280816365053e-06, 'learning_rate/action_model': 4.837280816365053e-05, 'epoch': 0.21}) +[2026-06-24 13:09:04,933][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.14894531667232513, 'train/grad_norm_pre_clip': 78.4415054321289, 'timing/data': 0.0003116270527243614, 'timing/model': 1.0825155738275498, 'learning_rate/qwen_vl_interface': 4.833469122751869e-06, 'learning_rate/action_model': 4.8334691227518686e-05, 'epoch': 0.21}) +[2026-06-24 13:09:06,015][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.09037736058235168, 'train/grad_norm_pre_clip': 31.831979751586914, 'timing/data': 0.0003345899749547243, 'timing/model': 1.0805047370959073, 'learning_rate/qwen_vl_interface': 4.829657699421048e-06, 'learning_rate/action_model': 4.829657699421048e-05, 'epoch': 0.21}) +[2026-06-24 13:09:07,099][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.11263799667358398, 'train/grad_norm_pre_clip': 61.054710388183594, 'timing/data': 0.0003218059428036213, 'timing/model': 1.0828411800321192, 'learning_rate/qwen_vl_interface': 4.825846548845783e-06, 'learning_rate/action_model': 4.825846548845783e-05, 'epoch': 0.21}) +[2026-06-24 13:09:08,185][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.05091565102338791, 'train/grad_norm_pre_clip': 60.00320816040039, 'timing/data': 0.0003292390611022711, 'timing/model': 1.084252413129434, 'learning_rate/qwen_vl_interface': 4.822035673499086e-06, 'learning_rate/action_model': 4.8220356734990854e-05, 'epoch': 0.21}) +[2026-06-24 13:09:09,269][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.14125677943229675, 'train/grad_norm_pre_clip': 58.17176055908203, 'timing/data': 0.00032100407406687737, 'timing/model': 1.0822153070475906, 'learning_rate/qwen_vl_interface': 4.818225075853797e-06, 'learning_rate/action_model': 4.8182250758537965e-05, 'epoch': 0.21}) +[2026-06-24 13:09:10,350][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.19242990016937256, 'train/grad_norm_pre_clip': 72.25778198242188, 'timing/data': 0.00034811487421393394, 'timing/model': 1.0796642829664052, 'learning_rate/qwen_vl_interface': 4.814414758382565e-06, 'learning_rate/action_model': 4.814414758382565e-05, 'epoch': 0.21}) +[2026-06-24 13:09:11,435][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.02321229688823223, 'train/grad_norm_pre_clip': 23.129283905029297, 'timing/data': 0.00032233702950179577, 'timing/model': 1.0832849240396172, 'learning_rate/qwen_vl_interface': 4.81060472355787e-06, 'learning_rate/action_model': 4.810604723557869e-05, 'epoch': 0.21}) +[2026-06-24 13:09:12,519][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.03339154273271561, 'train/grad_norm_pre_clip': 17.0430850982666, 'timing/data': 0.0003842730075120926, 'timing/model': 1.0823730251286179, 'learning_rate/qwen_vl_interface': 4.806794973852e-06, 'learning_rate/action_model': 4.806794973852e-05, 'epoch': 0.21}) +[2026-06-24 13:09:13,867][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.04554755613207817, 'train/grad_norm_pre_clip': 30.524948120117188, 'timing/data': 0.00036644889041781425, 'timing/model': 1.3463202668353915, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 0.21}) +[2026-06-24 13:09:14,948][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.008164525963366032, 'train/grad_norm_pre_clip': 3.8781988620758057, 'timing/data': 0.00033779582008719444, 'timing/model': 1.0791162899695337, 'learning_rate/qwen_vl_interface': 4.799176339684964e-06, 'learning_rate/action_model': 4.7991763396849636e-05, 'epoch': 0.21}) +[2026-06-24 13:09:16,030][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.1499175727367401, 'train/grad_norm_pre_clip': 65.72913360595703, 'timing/data': 0.00034626200795173645, 'timing/model': 1.0805821949616075, 'learning_rate/qwen_vl_interface': 4.795367460167448e-06, 'learning_rate/action_model': 4.795367460167448e-05, 'epoch': 0.21}) +[2026-06-24 13:09:17,110][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.11413773149251938, 'train/grad_norm_pre_clip': 41.48123550415039, 'timing/data': 0.00034907693043351173, 'timing/model': 1.078681284794584, 'learning_rate/qwen_vl_interface': 4.791558875656053e-06, 'learning_rate/action_model': 4.791558875656053e-05, 'epoch': 0.21}) +[2026-06-24 13:09:18,194][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.253624826669693, 'train/grad_norm_pre_clip': 46.1048698425293, 'timing/data': 0.0003177570179104805, 'timing/model': 1.0817151211667806, 'learning_rate/qwen_vl_interface': 4.787750588622126e-06, 'learning_rate/action_model': 4.787750588622126e-05, 'epoch': 0.21}) +[2026-06-24 13:09:19,277][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.11288489401340485, 'train/grad_norm_pre_clip': 73.79595184326172, 'timing/data': 0.00034754397347569466, 'timing/model': 1.0810126760043204, 'learning_rate/qwen_vl_interface': 4.783942601536823e-06, 'learning_rate/action_model': 4.783942601536823e-05, 'epoch': 0.21}) +[2026-06-24 13:09:20,362][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.04156595468521118, 'train/grad_norm_pre_clip': 23.49025535583496, 'timing/data': 0.00037121796049177647, 'timing/model': 1.0838383710943162, 'learning_rate/qwen_vl_interface': 4.780134916871107e-06, 'learning_rate/action_model': 4.780134916871107e-05, 'epoch': 0.21}) +[2026-06-24 13:09:21,451][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.18402938544750214, 'train/grad_norm_pre_clip': 94.94982147216797, 'timing/data': 0.000579240033403039, 'timing/model': 1.0862177358940244, 'learning_rate/qwen_vl_interface': 4.776327537095741e-06, 'learning_rate/action_model': 4.7763275370957404e-05, 'epoch': 0.21}) +[2026-06-24 13:09:22,536][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.1013183444738388, 'train/grad_norm_pre_clip': 37.78891372680664, 'timing/data': 0.00035474798642098904, 'timing/model': 1.0839181349147111, 'learning_rate/qwen_vl_interface': 4.772520464681294e-06, 'learning_rate/action_model': 4.7725204646812945e-05, 'epoch': 0.21}) +[2026-06-24 13:09:23,619][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.23967628180980682, 'train/grad_norm_pre_clip': 133.35617065429688, 'timing/data': 0.0004262018483132124, 'timing/model': 1.08130977419205, 'learning_rate/qwen_vl_interface': 4.768713702098134e-06, 'learning_rate/action_model': 4.7687137020981346e-05, 'epoch': 0.21}) +[2026-06-24 13:09:24,705][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.10224225372076035, 'train/grad_norm_pre_clip': 51.44112777709961, 'timing/data': 0.00036238180473446846, 'timing/model': 1.083837578073144, 'learning_rate/qwen_vl_interface': 4.764907251816427e-06, 'learning_rate/action_model': 4.764907251816427e-05, 'epoch': 0.21}) +[2026-06-24 13:09:25,789][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.10739519447088242, 'train/grad_norm_pre_clip': 28.082576751708984, 'timing/data': 0.00036760210059583187, 'timing/model': 1.0824054200202227, 'learning_rate/qwen_vl_interface': 4.7611011163061375e-06, 'learning_rate/action_model': 4.7611011163061376e-05, 'epoch': 0.21}) +[2026-06-24 13:09:26,877][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.04800659045577049, 'train/grad_norm_pre_clip': 19.509096145629883, 'timing/data': 0.0003252520691603422, 'timing/model': 1.0866718459874392, 'learning_rate/qwen_vl_interface': 4.757295298037025e-06, 'learning_rate/action_model': 4.7572952980370254e-05, 'epoch': 0.21}) +[2026-06-24 13:09:27,959][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.09709764271974564, 'train/grad_norm_pre_clip': 36.7803840637207, 'timing/data': 0.0003282679244875908, 'timing/model': 1.080743737053126, 'learning_rate/qwen_vl_interface': 4.753489799478647e-06, 'learning_rate/action_model': 4.7534897994786464e-05, 'epoch': 0.21}) +[2026-06-24 13:09:29,045][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.007524488028138876, 'train/grad_norm_pre_clip': 4.385206699371338, 'timing/data': 0.0003175071906298399, 'timing/model': 1.083721367875114, 'learning_rate/qwen_vl_interface': 4.749684623100346e-06, 'learning_rate/action_model': 4.749684623100346e-05, 'epoch': 0.21}) +[2026-06-24 13:09:30,335][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.1150103285908699, 'train/grad_norm_pre_clip': 43.22248840332031, 'timing/data': 0.00033398810774087906, 'timing/model': 1.2888388361316174, 'learning_rate/qwen_vl_interface': 4.745879771371261e-06, 'learning_rate/action_model': 4.74587977137126e-05, 'epoch': 0.21}) +[2026-06-24 13:09:31,420][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.13550081849098206, 'train/grad_norm_pre_clip': 60.66454315185547, 'timing/data': 0.0003332868218421936, 'timing/model': 1.0829106259625405, 'learning_rate/qwen_vl_interface': 4.742075246760319e-06, 'learning_rate/action_model': 4.742075246760319e-05, 'epoch': 0.21}) +[2026-06-24 13:09:32,500][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.12793336808681488, 'train/grad_norm_pre_clip': 26.51044273376465, 'timing/data': 0.0004136089701205492, 'timing/model': 1.0790662479121238, 'learning_rate/qwen_vl_interface': 4.738271051736234e-06, 'learning_rate/action_model': 4.738271051736234e-05, 'epoch': 0.21}) +[2026-06-24 13:09:33,589][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.02762400358915329, 'train/grad_norm_pre_clip': 13.310086250305176, 'timing/data': 0.0003633128944784403, 'timing/model': 1.0873365479055792, 'learning_rate/qwen_vl_interface': 4.734467188767511e-06, 'learning_rate/action_model': 4.734467188767511e-05, 'epoch': 0.21}) +[2026-06-24 13:09:34,675][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.0486859455704689, 'train/grad_norm_pre_clip': 22.200660705566406, 'timing/data': 0.00034354603849351406, 'timing/model': 1.0840362450107932, 'learning_rate/qwen_vl_interface': 4.730663660322428e-06, 'learning_rate/action_model': 4.730663660322429e-05, 'epoch': 0.21}) +[2026-06-24 13:09:35,754][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.09160076081752777, 'train/grad_norm_pre_clip': 39.937156677246094, 'timing/data': 0.0003380558919161558, 'timing/model': 1.077429440105334, 'learning_rate/qwen_vl_interface': 4.726860468869061e-06, 'learning_rate/action_model': 4.72686046886906e-05, 'epoch': 0.21}) +[2026-06-24 13:09:36,836][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.18933017551898956, 'train/grad_norm_pre_clip': 44.42600631713867, 'timing/data': 0.0003520331811159849, 'timing/model': 1.0802485819440335, 'learning_rate/qwen_vl_interface': 4.723057616875256e-06, 'learning_rate/action_model': 4.723057616875256e-05, 'epoch': 0.21}) +[2026-06-24 13:09:37,946][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.155983567237854, 'train/grad_norm_pre_clip': 39.34236145019531, 'timing/data': 0.00031355000101029873, 'timing/model': 1.1087090149521828, 'learning_rate/qwen_vl_interface': 4.71925510680864e-06, 'learning_rate/action_model': 4.719255106808639e-05, 'epoch': 0.21}) +[2026-06-24 13:09:39,031][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.039440520107746124, 'train/grad_norm_pre_clip': 23.410276412963867, 'timing/data': 0.00034894607961177826, 'timing/model': 1.0835523339919746, 'learning_rate/qwen_vl_interface': 4.715452941136624e-06, 'learning_rate/action_model': 4.7154529411366237e-05, 'epoch': 0.21}) +[2026-06-24 13:09:40,118][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.021485354751348495, 'train/grad_norm_pre_clip': 12.58523178100586, 'timing/data': 0.00031876005232334137, 'timing/model': 1.085457906126976, 'learning_rate/qwen_vl_interface': 4.71165112232639e-06, 'learning_rate/action_model': 4.7116511223263904e-05, 'epoch': 0.21}) +[2026-06-24 13:09:41,198][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.066022127866745, 'train/grad_norm_pre_clip': 24.2618350982666, 'timing/data': 0.00034835608676075935, 'timing/model': 1.077672102022916, 'learning_rate/qwen_vl_interface': 4.7078496528449e-06, 'learning_rate/action_model': 4.7078496528449e-05, 'epoch': 0.21}) +[2026-06-24 13:09:42,280][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.45359060168266296, 'train/grad_norm_pre_clip': 66.1448974609375, 'timing/data': 0.0003355010412633419, 'timing/model': 1.0807753391563892, 'learning_rate/qwen_vl_interface': 4.7040485351588845e-06, 'learning_rate/action_model': 4.7040485351588845e-05, 'epoch': 0.21}) +[2026-06-24 13:09:43,361][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.03653182461857796, 'train/grad_norm_pre_clip': 22.1414794921875, 'timing/data': 0.000366720138117671, 'timing/model': 1.079185624839738, 'learning_rate/qwen_vl_interface': 4.700247771734845e-06, 'learning_rate/action_model': 4.7002477717348456e-05, 'epoch': 0.21}) +[2026-06-24 13:09:44,440][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.03813168406486511, 'train/grad_norm_pre_clip': 28.648799896240234, 'timing/data': 0.0003502289764583111, 'timing/model': 1.077581935096532, 'learning_rate/qwen_vl_interface': 4.6964473650390595e-06, 'learning_rate/action_model': 4.6964473650390596e-05, 'epoch': 0.21}) +[2026-06-24 13:09:45,525][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.11393899470567703, 'train/grad_norm_pre_clip': 43.39338684082031, 'timing/data': 0.000334037933498621, 'timing/model': 1.0827614380978048, 'learning_rate/qwen_vl_interface': 4.692647317537569e-06, 'learning_rate/action_model': 4.692647317537568e-05, 'epoch': 0.21}) +[2026-06-24 13:09:46,611][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.16055737435817719, 'train/grad_norm_pre_clip': 48.760276794433594, 'timing/data': 0.0003519121091812849, 'timing/model': 1.0848873660434037, 'learning_rate/qwen_vl_interface': 4.688847631696182e-06, 'learning_rate/action_model': 4.688847631696182e-05, 'epoch': 0.21}) +[2026-06-24 13:09:47,989][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.029918573796749115, 'train/grad_norm_pre_clip': 20.787994384765625, 'timing/data': 0.000334019074216485, 'timing/model': 1.3764266478829086, 'learning_rate/qwen_vl_interface': 4.685048309980475e-06, 'learning_rate/action_model': 4.6850483099804755e-05, 'epoch': 0.21}) +[2026-06-24 13:09:49,072][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.08581110090017319, 'train/grad_norm_pre_clip': 53.416446685791016, 'timing/data': 0.00035616103559732437, 'timing/model': 1.0810826199594885, 'learning_rate/qwen_vl_interface': 4.6812493548557865e-06, 'learning_rate/action_model': 4.681249354855786e-05, 'epoch': 0.21}) +[2026-06-24 13:09:50,154][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.0015053552342578769, 'train/grad_norm_pre_clip': 0.9299976229667664, 'timing/data': 0.0003192010335624218, 'timing/model': 1.080683325883001, 'learning_rate/qwen_vl_interface': 4.677450768787216e-06, 'learning_rate/action_model': 4.677450768787216e-05, 'epoch': 0.21}) +[2026-06-24 13:09:51,235][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.0809042751789093, 'train/grad_norm_pre_clip': 35.53961944580078, 'timing/data': 0.00033968896605074406, 'timing/model': 1.0796984189655632, 'learning_rate/qwen_vl_interface': 4.673652554239626e-06, 'learning_rate/action_model': 4.673652554239625e-05, 'epoch': 0.21}) +[2026-06-24 13:09:52,319][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.14125509560108185, 'train/grad_norm_pre_clip': 95.29776000976562, 'timing/data': 0.0003044730983674526, 'timing/model': 1.082105467095971, 'learning_rate/qwen_vl_interface': 4.669854713677636e-06, 'learning_rate/action_model': 4.669854713677636e-05, 'epoch': 0.21}) +[2026-06-24 13:09:53,400][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.22717057168483734, 'train/grad_norm_pre_clip': 95.0744400024414, 'timing/data': 0.0003624518867582083, 'timing/model': 1.0788727509789169, 'learning_rate/qwen_vl_interface': 4.666057249565624e-06, 'learning_rate/action_model': 4.6660572495656246e-05, 'epoch': 0.21}) +[2026-06-24 13:09:54,483][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.07651318609714508, 'train/grad_norm_pre_clip': 55.36821365356445, 'timing/data': 0.00039670616388320923, 'timing/model': 1.0818125698715448, 'learning_rate/qwen_vl_interface': 4.662260164367722e-06, 'learning_rate/action_model': 4.6622601643677215e-05, 'epoch': 0.21}) +[2026-06-24 13:09:55,566][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.059020113199949265, 'train/grad_norm_pre_clip': 23.808549880981445, 'timing/data': 0.00033004116266965866, 'timing/model': 1.0815447429195046, 'learning_rate/qwen_vl_interface': 4.65846346054782e-06, 'learning_rate/action_model': 4.65846346054782e-05, 'epoch': 0.21}) +[2026-06-24 13:09:56,649][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.08528675138950348, 'train/grad_norm_pre_clip': 43.79838562011719, 'timing/data': 0.0003395089879631996, 'timing/model': 1.0812789399642497, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 0.21}) +[2026-06-24 13:09:57,730][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.03644663840532303, 'train/grad_norm_pre_clip': 13.533004760742188, 'timing/data': 0.00037359283305704594, 'timing/model': 1.0790562049951404, 'learning_rate/qwen_vl_interface': 4.650871206896319e-06, 'learning_rate/action_model': 4.650871206896319e-05, 'epoch': 0.21}) +[2026-06-24 13:09:58,810][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.0027407510206103325, 'train/grad_norm_pre_clip': 4.088042736053467, 'timing/data': 0.0002984008751809597, 'timing/model': 1.0785561418160796, 'learning_rate/qwen_vl_interface': 4.647075661991254e-06, 'learning_rate/action_model': 4.6470756619912543e-05, 'epoch': 0.21}) +[2026-06-24 13:09:59,889][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.10060111433267593, 'train/grad_norm_pre_clip': 27.66356658935547, 'timing/data': 0.0003275058697909117, 'timing/model': 1.0774556279648095, 'learning_rate/qwen_vl_interface': 4.643280508317244e-06, 'learning_rate/action_model': 4.6432805083172436e-05, 'epoch': 0.21}) +[2026-06-24 13:10:00,972][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.09572251886129379, 'train/grad_norm_pre_clip': 36.58103942871094, 'timing/data': 0.00035679200664162636, 'timing/model': 1.080655697034672, 'learning_rate/qwen_vl_interface': 4.639485748336926e-06, 'learning_rate/action_model': 4.639485748336926e-05, 'epoch': 0.21}) +[2026-06-24 13:10:02,053][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.09451698511838913, 'train/grad_norm_pre_clip': 40.128910064697266, 'timing/data': 0.00033441907726228237, 'timing/model': 1.079744549933821, 'learning_rate/qwen_vl_interface': 4.6356913845126765e-06, 'learning_rate/action_model': 4.635691384512676e-05, 'epoch': 0.21}) +[2026-06-24 13:10:03,133][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.38412609696388245, 'train/grad_norm_pre_clip': 111.27323150634766, 'timing/data': 0.0004098007921129465, 'timing/model': 1.078464839141816, 'learning_rate/qwen_vl_interface': 4.63189741930662e-06, 'learning_rate/action_model': 4.6318974193066195e-05, 'epoch': 0.21}) +[2026-06-24 13:10:04,421][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.8241491317749023, 'train/grad_norm_pre_clip': 124.59042358398438, 'timing/data': 0.000316765857860446, 'timing/model': 1.2860174130182713, 'learning_rate/qwen_vl_interface': 4.628103855180615e-06, 'learning_rate/action_model': 4.6281038551806156e-05, 'epoch': 0.21}) +[2026-06-24 13:10:05,502][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.37264499068260193, 'train/grad_norm_pre_clip': 86.96656799316406, 'timing/data': 0.00031764782033860683, 'timing/model': 1.07963703898713, 'learning_rate/qwen_vl_interface': 4.62431069459627e-06, 'learning_rate/action_model': 4.6243106945962697e-05, 'epoch': 0.21}) +[2026-06-24 13:10:06,583][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.10819535702466965, 'train/grad_norm_pre_clip': 63.40330123901367, 'timing/data': 0.0003246909473091364, 'timing/model': 1.0798300390597433, 'learning_rate/qwen_vl_interface': 4.62051794001492e-06, 'learning_rate/action_model': 4.62051794001492e-05, 'epoch': 0.21}) +[2026-06-24 13:10:07,668][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.5674721598625183, 'train/grad_norm_pre_clip': 99.89555358886719, 'timing/data': 0.0003251919988542795, 'timing/model': 1.0826130129862577, 'learning_rate/qwen_vl_interface': 4.616725593897645e-06, 'learning_rate/action_model': 4.616725593897645e-05, 'epoch': 0.21}) +[2026-06-24 13:10:08,748][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.38337841629981995, 'train/grad_norm_pre_clip': 80.75814819335938, 'timing/data': 0.0003925878554582596, 'timing/model': 1.0783694148994982, 'learning_rate/qwen_vl_interface': 4.612933658705257e-06, 'learning_rate/action_model': 4.6129336587052567e-05, 'epoch': 0.21}) +[2026-06-24 13:10:09,831][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.07747896760702133, 'train/grad_norm_pre_clip': 30.367372512817383, 'timing/data': 0.00031157699413597584, 'timing/model': 1.0815465010236949, 'learning_rate/qwen_vl_interface': 4.6091421368983005e-06, 'learning_rate/action_model': 4.6091421368983e-05, 'epoch': 0.21}) +[2026-06-24 13:10:10,912][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.05468976870179176, 'train/grad_norm_pre_clip': 81.77189636230469, 'timing/data': 0.0003119879402220249, 'timing/model': 1.0795066400896758, 'learning_rate/qwen_vl_interface': 4.605351030937054e-06, 'learning_rate/action_model': 4.605351030937055e-05, 'epoch': 0.21}) +[2026-06-24 13:10:11,996][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.10667604207992554, 'train/grad_norm_pre_clip': 35.51874923706055, 'timing/data': 0.00033524096943438053, 'timing/model': 1.0821024919860065, 'learning_rate/qwen_vl_interface': 4.601560343281526e-06, 'learning_rate/action_model': 4.601560343281526e-05, 'epoch': 0.21}) +[2026-06-24 13:10:13,080][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.04855896532535553, 'train/grad_norm_pre_clip': 28.109905242919922, 'timing/data': 0.00033385795541107655, 'timing/model': 1.0827591351699084, 'learning_rate/qwen_vl_interface': 4.59777007639145e-06, 'learning_rate/action_model': 4.59777007639145e-05, 'epoch': 0.21}) +[2026-06-24 13:10:14,161][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.0864168331027031, 'train/grad_norm_pre_clip': 27.454408645629883, 'timing/data': 0.0003324651625007391, 'timing/model': 1.0793008250184357, 'learning_rate/qwen_vl_interface': 4.59398023272629e-06, 'learning_rate/action_model': 4.59398023272629e-05, 'epoch': 0.21}) +[2026-06-24 13:10:15,242][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.03964701294898987, 'train/grad_norm_pre_clip': 18.35664176940918, 'timing/data': 0.00033082207664847374, 'timing/model': 1.0790724020916969, 'learning_rate/qwen_vl_interface': 4.590190814745234e-06, 'learning_rate/action_model': 4.590190814745234e-05, 'epoch': 0.22}) +[2026-06-24 13:10:16,327][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.1085515022277832, 'train/grad_norm_pre_clip': 38.61167526245117, 'timing/data': 0.0003471430391073227, 'timing/model': 1.084082396933809, 'learning_rate/qwen_vl_interface': 4.586401824907197e-06, 'learning_rate/action_model': 4.586401824907197e-05, 'epoch': 0.22}) +[2026-06-24 13:10:17,410][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.06614372879266739, 'train/grad_norm_pre_clip': 43.06718826293945, 'timing/data': 0.0003166359383612871, 'timing/model': 1.0811699738260359, 'learning_rate/qwen_vl_interface': 4.582613265670811e-06, 'learning_rate/action_model': 4.582613265670811e-05, 'epoch': 0.22}) +[2026-06-24 13:10:18,491][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.14833593368530273, 'train/grad_norm_pre_clip': 85.74420166015625, 'timing/data': 0.0003160152118653059, 'timing/model': 1.0790732859168202, 'learning_rate/qwen_vl_interface': 4.578825139494428e-06, 'learning_rate/action_model': 4.578825139494428e-05, 'epoch': 0.22}) +[2026-06-24 13:10:19,573][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.03865509480237961, 'train/grad_norm_pre_clip': 13.110627174377441, 'timing/data': 0.0003323550336062908, 'timing/model': 1.080029723001644, 'learning_rate/qwen_vl_interface': 4.57503744883613e-06, 'learning_rate/action_model': 4.5750374488361295e-05, 'epoch': 0.22}) +[2026-06-24 13:10:20,865][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.0777832493185997, 'train/grad_norm_pre_clip': 21.07489013671875, 'timing/data': 0.0003232189919799566, 'timing/model': 1.2907824518624693, 'learning_rate/qwen_vl_interface': 4.571250196153698e-06, 'learning_rate/action_model': 4.5712501961536985e-05, 'epoch': 0.22}) +[2026-06-24 13:10:21,948][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.04433014616370201, 'train/grad_norm_pre_clip': 24.1169376373291, 'timing/data': 0.0003449989017099142, 'timing/model': 1.0812570669222623, 'learning_rate/qwen_vl_interface': 4.567463383904649e-06, 'learning_rate/action_model': 4.5674633839046485e-05, 'epoch': 0.22}) +[2026-06-24 13:10:23,039][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.13422119617462158, 'train/grad_norm_pre_clip': 68.76725769042969, 'timing/data': 0.00032882788218557835, 'timing/model': 1.0890164349693805, 'learning_rate/qwen_vl_interface': 4.5636770145462e-06, 'learning_rate/action_model': 4.5636770145462e-05, 'epoch': 0.22}) +[2026-06-24 13:10:24,260][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.1970517933368683, 'train/grad_norm_pre_clip': 21.144989013671875, 'timing/data': 0.0006464170292019844, 'timing/model': 1.2197755419183522, 'learning_rate/qwen_vl_interface': 4.559891090535282e-06, 'learning_rate/action_model': 4.559891090535282e-05, 'epoch': 0.22}) +[2026-06-24 13:10:25,366][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.10531587898731232, 'train/grad_norm_pre_clip': 31.513280868530273, 'timing/data': 0.0003795039374381304, 'timing/model': 1.1034408109262586, 'learning_rate/qwen_vl_interface': 4.556105614328545e-06, 'learning_rate/action_model': 4.5561056143285455e-05, 'epoch': 0.22}) +[2026-06-24 13:10:26,447][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.16491474211215973, 'train/grad_norm_pre_clip': 40.4825553894043, 'timing/data': 0.0003469728399068117, 'timing/model': 1.0802257561590523, 'learning_rate/qwen_vl_interface': 4.55232058838234e-06, 'learning_rate/action_model': 4.55232058838234e-05, 'epoch': 0.22}) +[2026-06-24 13:10:27,530][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.3412765562534332, 'train/grad_norm_pre_clip': 47.15028381347656, 'timing/data': 0.00031361007131636143, 'timing/model': 1.0806413360405713, 'learning_rate/qwen_vl_interface': 4.548536015152732e-06, 'learning_rate/action_model': 4.5485360151527324e-05, 'epoch': 0.22}) +[2026-06-24 13:10:28,613][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.42963579297065735, 'train/grad_norm_pre_clip': 58.84342956542969, 'timing/data': 0.00032305787317454815, 'timing/model': 1.082329279044643, 'learning_rate/qwen_vl_interface': 4.544751897095488e-06, 'learning_rate/action_model': 4.5447518970954874e-05, 'epoch': 0.22}) +[2026-06-24 13:10:29,697][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.2007625848054886, 'train/grad_norm_pre_clip': 50.60905456542969, 'timing/data': 0.0003325559664517641, 'timing/model': 1.0816632730420679, 'learning_rate/qwen_vl_interface': 4.540968236666081e-06, 'learning_rate/action_model': 4.540968236666081e-05, 'epoch': 0.22}) +[2026-06-24 13:10:30,781][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.0761098861694336, 'train/grad_norm_pre_clip': 34.17551803588867, 'timing/data': 0.0005861229728907347, 'timing/model': 1.0820082197897136, 'learning_rate/qwen_vl_interface': 4.537185036319687e-06, 'learning_rate/action_model': 4.5371850363196874e-05, 'epoch': 0.22}) +[2026-06-24 13:10:31,861][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.1345125287771225, 'train/grad_norm_pre_clip': 24.166030883789062, 'timing/data': 0.000313169090077281, 'timing/model': 1.078339233994484, 'learning_rate/qwen_vl_interface': 4.533402298511184e-06, 'learning_rate/action_model': 4.533402298511184e-05, 'epoch': 0.22}) +[2026-06-24 13:10:32,943][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.0669795423746109, 'train/grad_norm_pre_clip': 18.514890670776367, 'timing/data': 0.000352643895894289, 'timing/model': 1.0803763950243592, 'learning_rate/qwen_vl_interface': 4.52962002569515e-06, 'learning_rate/action_model': 4.52962002569515e-05, 'epoch': 0.22}) +[2026-06-24 13:10:34,023][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.12284266948699951, 'train/grad_norm_pre_clip': 49.27381896972656, 'timing/data': 0.0003594658337533474, 'timing/model': 1.0786260080058128, 'learning_rate/qwen_vl_interface': 4.52583822032586e-06, 'learning_rate/action_model': 4.525838220325859e-05, 'epoch': 0.22}) +[2026-06-24 13:10:35,101][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.11233901232481003, 'train/grad_norm_pre_clip': 35.58074951171875, 'timing/data': 0.0003258830402046442, 'timing/model': 1.0768823239486665, 'learning_rate/qwen_vl_interface': 4.5220568848572864e-06, 'learning_rate/action_model': 4.5220568848572864e-05, 'epoch': 0.22}) +[2026-06-24 13:10:36,183][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.0768032819032669, 'train/grad_norm_pre_clip': 12.86828327178955, 'timing/data': 0.0003602078650146723, 'timing/model': 1.0801180608104914, 'learning_rate/qwen_vl_interface': 4.518276021743098e-06, 'learning_rate/action_model': 4.518276021743098e-05, 'epoch': 0.22}) +[2026-06-24 13:10:37,455][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.09582821279764175, 'train/grad_norm_pre_clip': 33.294334411621094, 'timing/data': 0.000442523043602705, 'timing/model': 1.2703005569055676, 'learning_rate/qwen_vl_interface': 4.514495633436653e-06, 'learning_rate/action_model': 4.5144956334366536e-05, 'epoch': 0.22}) +[2026-06-24 13:10:38,535][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.15741121768951416, 'train/grad_norm_pre_clip': 33.61091232299805, 'timing/data': 0.0003256530035287142, 'timing/model': 1.078704415122047, 'learning_rate/qwen_vl_interface': 4.510715722391009e-06, 'learning_rate/action_model': 4.5107157223910086e-05, 'epoch': 0.22}) +[2026-06-24 13:10:39,613][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.18151317536830902, 'train/grad_norm_pre_clip': 34.435646057128906, 'timing/data': 0.0003288090229034424, 'timing/model': 1.076634077122435, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 0.22}) +[2026-06-24 13:10:40,691][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.2109055072069168, 'train/grad_norm_pre_clip': 53.556941986083984, 'timing/data': 0.00031457189470529556, 'timing/model': 1.0762253359425813, 'learning_rate/qwen_vl_interface': 4.503157341892773e-06, 'learning_rate/action_model': 4.503157341892773e-05, 'epoch': 0.22}) +[2026-06-24 13:10:41,772][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.16122208535671234, 'train/grad_norm_pre_clip': 43.86003494262695, 'timing/data': 0.0005361291114240885, 'timing/model': 1.0785479731857777, 'learning_rate/qwen_vl_interface': 4.499378877344736e-06, 'learning_rate/action_model': 4.4993788773447356e-05, 'epoch': 0.22}) +[2026-06-24 13:10:42,855][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.042948510497808456, 'train/grad_norm_pre_clip': 11.803238868713379, 'timing/data': 0.00030871108174324036, 'timing/model': 1.0819779250305146, 'learning_rate/qwen_vl_interface': 4.495600899866593e-06, 'learning_rate/action_model': 4.495600899866593e-05, 'epoch': 0.22}) +[2026-06-24 13:10:43,939][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.014743626117706299, 'train/grad_norm_pre_clip': 5.421886444091797, 'timing/data': 0.0003655578475445509, 'timing/model': 1.0819018699694425, 'learning_rate/qwen_vl_interface': 4.491823411909839e-06, 'learning_rate/action_model': 4.491823411909838e-05, 'epoch': 0.22}) +[2026-06-24 13:10:45,021][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2250 +[2026-06-24 13:11:09,964][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2250 +[2026-06-24 13:11:09,965][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2250 +[2026-06-24 13:11:42,864][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2250 +[2026-06-24 13:16:15,598][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2250 +[2026-06-24 13:16:15,598][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.18131403625011444, 'train/grad_norm_pre_clip': 34.07521438598633, 'eval/action_loss/seconds': 24.940718235913664, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.061864208430051804, 'eval/latency_2/loss': 0.061864208430051804, 'eval/flappy/loss': 0.061864208430051804, 'eval/flappy/latency_2/loss': 0.061864208430051804, 'eval/flap_precision': 0.95, 'eval/flap_recall': 0.9538670284938942, 'eval/flap_f1': 0.951929587000677, 'eval/flap_rate_model': 0.20555555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.951929587000677, 'eval/latency_2/flap_precision': 0.95, 'eval/latency_2/flap_recall': 0.9538670284938942, 'eval/latency_2/flap_f1': 0.951929587000677, 'eval/latency_2/flap_rate_model': 0.20555555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.951929587000677, 'eval/action_classification/data_wait_seconds': 0.315312922000885, 'eval/action_classification/predict_seconds': 32.49160385131836, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 208.66000000000992, 'rl_games_eval/mid_train/mean_length': 1699.0, 'rl_games_eval/mid_train/std_reward': 196.09581943530796, 'rl_games_eval/mid_train/std_length': 1580.14505663246, 'rl_games_eval/mid_train/macro_mean_reward': 208.66000000000992, 'rl_games_eval/mid_train/macro_mean_length': 1699.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 208.66000000000992, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1699.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 196.09581943530796, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1580.14505663246, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 208.66000000000992, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003557091113179922, 'timing/model': 1.0812598790507764, 'learning_rate/qwen_vl_interface': 4.488046415925638e-06, 'learning_rate/action_model': 4.4880464159256383e-05, 'epoch': 0.22}) +[2026-06-24 13:16:16,665][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.12020302563905716, 'train/grad_norm_pre_clip': 32.64852523803711, 'timing/data': 0.00035060709342360497, 'timing/model': 1.064765237038955, 'learning_rate/qwen_vl_interface': 4.484269914364848e-06, 'learning_rate/action_model': 4.484269914364848e-05, 'epoch': 0.22}) +[2026-06-24 13:16:17,732][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.08531008660793304, 'train/grad_norm_pre_clip': 32.144775390625, 'timing/data': 0.0003718170337378979, 'timing/model': 1.0655101470183581, 'learning_rate/qwen_vl_interface': 4.480493909677998e-06, 'learning_rate/action_model': 4.480493909677998e-05, 'epoch': 0.22}) +[2026-06-24 13:16:18,800][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.12487395852804184, 'train/grad_norm_pre_clip': 36.120361328125, 'timing/data': 0.000332673080265522, 'timing/model': 1.0661678418982774, 'learning_rate/qwen_vl_interface': 4.476718404315294e-06, 'learning_rate/action_model': 4.476718404315293e-05, 'epoch': 0.22}) +[2026-06-24 13:16:19,881][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.11203537881374359, 'train/grad_norm_pre_clip': 40.804683685302734, 'timing/data': 0.00030512199737131596, 'timing/model': 1.0721810511313379, 'learning_rate/qwen_vl_interface': 4.472943400726621e-06, 'learning_rate/action_model': 4.472943400726621e-05, 'epoch': 0.22}) +[2026-06-24 13:16:20,949][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.041761256754398346, 'train/grad_norm_pre_clip': 12.614717483520508, 'timing/data': 0.0003984461072832346, 'timing/model': 1.0657161129638553, 'learning_rate/qwen_vl_interface': 4.469168901361541e-06, 'learning_rate/action_model': 4.4691689013615405e-05, 'epoch': 0.22}) +[2026-06-24 13:16:22,229][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.09409306198358536, 'train/grad_norm_pre_clip': 18.93619155883789, 'timing/data': 0.00033922609873116016, 'timing/model': 1.278568176086992, 'learning_rate/qwen_vl_interface': 4.4653949086692825e-06, 'learning_rate/action_model': 4.465394908669282e-05, 'epoch': 0.22}) +[2026-06-24 13:16:23,297][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.07541685551404953, 'train/grad_norm_pre_clip': 33.1827278137207, 'timing/data': 0.0003576700109988451, 'timing/model': 1.0662500318139791, 'learning_rate/qwen_vl_interface': 4.46162142509875e-06, 'learning_rate/action_model': 4.46162142509875e-05, 'epoch': 0.22}) +[2026-06-24 13:16:24,365][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.03225085884332657, 'train/grad_norm_pre_clip': 11.882996559143066, 'timing/data': 0.00033343490213155746, 'timing/model': 1.0665879489388317, 'learning_rate/qwen_vl_interface': 4.457848453098514e-06, 'learning_rate/action_model': 4.4578484530985144e-05, 'epoch': 0.22}) +[2026-06-24 13:16:25,436][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.12462877482175827, 'train/grad_norm_pre_clip': 12.64220905303955, 'timing/data': 0.00034179119393229485, 'timing/model': 1.069597610970959, 'learning_rate/qwen_vl_interface': 4.454075995116819e-06, 'learning_rate/action_model': 4.454075995116819e-05, 'epoch': 0.22}) +[2026-06-24 13:16:26,510][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.02677239663898945, 'train/grad_norm_pre_clip': 8.55119800567627, 'timing/data': 0.0003074959386140108, 'timing/model': 1.0726659838110209, 'learning_rate/qwen_vl_interface': 4.450304053601567e-06, 'learning_rate/action_model': 4.450304053601567e-05, 'epoch': 0.22}) +[2026-06-24 13:16:27,585][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.0898711085319519, 'train/grad_norm_pre_clip': 41.94050979614258, 'timing/data': 0.0003311100881546736, 'timing/model': 1.0730054820887744, 'learning_rate/qwen_vl_interface': 4.446532631000335e-06, 'learning_rate/action_model': 4.446532631000334e-05, 'epoch': 0.22}) +[2026-06-24 13:16:28,658][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.10140127688646317, 'train/grad_norm_pre_clip': 40.828094482421875, 'timing/data': 0.0004019630141556263, 'timing/model': 1.0706746389623731, 'learning_rate/qwen_vl_interface': 4.442761729760352e-06, 'learning_rate/action_model': 4.442761729760352e-05, 'epoch': 0.22}) +[2026-06-24 13:16:29,729][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.06283952295780182, 'train/grad_norm_pre_clip': 28.819713592529297, 'timing/data': 0.00035117799416184425, 'timing/model': 1.070198310073465, 'learning_rate/qwen_vl_interface': 4.438991352328522e-06, 'learning_rate/action_model': 4.4389913523285214e-05, 'epoch': 0.22}) +[2026-06-24 13:16:30,801][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.02392367087304592, 'train/grad_norm_pre_clip': 8.555222511291504, 'timing/data': 0.0003193081356585026, 'timing/model': 1.0699018770828843, 'learning_rate/qwen_vl_interface': 4.435221501151397e-06, 'learning_rate/action_model': 4.435221501151397e-05, 'epoch': 0.22}) +[2026-06-24 13:16:31,874][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.029152873903512955, 'train/grad_norm_pre_clip': 23.287925720214844, 'timing/data': 0.00036478321999311447, 'timing/model': 1.0713207290973514, 'learning_rate/qwen_vl_interface': 4.431452178675194e-06, 'learning_rate/action_model': 4.431452178675194e-05, 'epoch': 0.22}) +[2026-06-24 13:16:32,949][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.03393634408712387, 'train/grad_norm_pre_clip': 16.263643264770508, 'timing/data': 0.0003277440555393696, 'timing/model': 1.0734566538594663, 'learning_rate/qwen_vl_interface': 4.427683387345787e-06, 'learning_rate/action_model': 4.427683387345786e-05, 'epoch': 0.22}) +[2026-06-24 13:16:34,020][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.06515853852033615, 'train/grad_norm_pre_clip': 18.054773330688477, 'timing/data': 0.0003168040420860052, 'timing/model': 1.0696869338862598, 'learning_rate/qwen_vl_interface': 4.423915129608698e-06, 'learning_rate/action_model': 4.423915129608698e-05, 'epoch': 0.22}) +[2026-06-24 13:16:35,094][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.02995072863996029, 'train/grad_norm_pre_clip': 15.30093765258789, 'timing/data': 0.00034571788273751736, 'timing/model': 1.0718926507979631, 'learning_rate/qwen_vl_interface': 4.420147407909116e-06, 'learning_rate/action_model': 4.420147407909116e-05, 'epoch': 0.22}) +[2026-06-24 13:16:36,167][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.04609276354312897, 'train/grad_norm_pre_clip': 43.61829376220703, 'timing/data': 0.00035919295623898506, 'timing/model': 1.0715282319579273, 'learning_rate/qwen_vl_interface': 4.416380224691873e-06, 'learning_rate/action_model': 4.416380224691872e-05, 'epoch': 0.22}) +[2026-06-24 13:16:37,239][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.07850988954305649, 'train/grad_norm_pre_clip': 32.34684753417969, 'timing/data': 0.0003419099375605583, 'timing/model': 1.0708007020875812, 'learning_rate/qwen_vl_interface': 4.412613582401447e-06, 'learning_rate/action_model': 4.4126135824014475e-05, 'epoch': 0.22}) +[2026-06-24 13:16:38,314][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.004811769351363182, 'train/grad_norm_pre_clip': 3.477536201477051, 'timing/data': 0.0003704850096255541, 'timing/model': 1.0731190980877727, 'learning_rate/qwen_vl_interface': 4.40884748348198e-06, 'learning_rate/action_model': 4.40884748348198e-05, 'epoch': 0.22}) +[2026-06-24 13:16:39,390][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.006614496931433678, 'train/grad_norm_pre_clip': 4.529654502868652, 'timing/data': 0.00034112902358174324, 'timing/model': 1.0744026659522206, 'learning_rate/qwen_vl_interface': 4.405081930377247e-06, 'learning_rate/action_model': 4.405081930377247e-05, 'epoch': 0.22}) +[2026-06-24 13:16:40,464][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.035534076392650604, 'train/grad_norm_pre_clip': 26.537288665771484, 'timing/data': 0.00033986708149313927, 'timing/model': 1.0716967401094735, 'learning_rate/qwen_vl_interface': 4.401316925530676e-06, 'learning_rate/action_model': 4.4013169255306765e-05, 'epoch': 0.22}) +[2026-06-24 13:16:41,836][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.04742925614118576, 'train/grad_norm_pre_clip': 44.598724365234375, 'timing/data': 0.0003292171750217676, 'timing/model': 1.3704437529668212, 'learning_rate/qwen_vl_interface': 4.39755247138534e-06, 'learning_rate/action_model': 4.397552471385339e-05, 'epoch': 0.22}) +[2026-06-24 13:16:42,912][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.08319210261106491, 'train/grad_norm_pre_clip': 87.9931640625, 'timing/data': 0.0003773770295083523, 'timing/model': 1.0745626878924668, 'learning_rate/qwen_vl_interface': 4.393788570383948e-06, 'learning_rate/action_model': 4.393788570383948e-05, 'epoch': 0.22}) +[2026-06-24 13:16:43,987][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.0024440083652734756, 'train/grad_norm_pre_clip': 1.6749752759933472, 'timing/data': 0.0003727679140865803, 'timing/model': 1.0731118128169328, 'learning_rate/qwen_vl_interface': 4.390025224968859e-06, 'learning_rate/action_model': 4.390025224968858e-05, 'epoch': 0.22}) +[2026-06-24 13:16:45,060][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.01376818586140871, 'train/grad_norm_pre_clip': 13.944567680358887, 'timing/data': 0.0003459679428488016, 'timing/model': 1.0713818448130041, 'learning_rate/qwen_vl_interface': 4.386262437582062e-06, 'learning_rate/action_model': 4.386262437582062e-05, 'epoch': 0.22}) +[2026-06-24 13:16:46,129][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.22136786580085754, 'train/grad_norm_pre_clip': 141.87808227539062, 'timing/data': 0.00035383296199142933, 'timing/model': 1.0676062530837953, 'learning_rate/qwen_vl_interface': 4.382500210665195e-06, 'learning_rate/action_model': 4.3825002106651945e-05, 'epoch': 0.22}) +[2026-06-24 13:16:47,202][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.32694223523139954, 'train/grad_norm_pre_clip': 121.23694610595703, 'timing/data': 0.00034064799547195435, 'timing/model': 1.0712172829080373, 'learning_rate/qwen_vl_interface': 4.378738546659522e-06, 'learning_rate/action_model': 4.378738546659521e-05, 'epoch': 0.22}) +[2026-06-24 13:16:48,277][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.025006309151649475, 'train/grad_norm_pre_clip': 15.15488052368164, 'timing/data': 0.00031808484345674515, 'timing/model': 1.073876657988876, 'learning_rate/qwen_vl_interface': 4.374977448005945e-06, 'learning_rate/action_model': 4.3749774480059455e-05, 'epoch': 0.22}) +[2026-06-24 13:16:49,352][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.06447716057300568, 'train/grad_norm_pre_clip': 36.72214889526367, 'timing/data': 0.0003141791094094515, 'timing/model': 1.0733579790685326, 'learning_rate/qwen_vl_interface': 4.3712169171450044e-06, 'learning_rate/action_model': 4.3712169171450046e-05, 'epoch': 0.22}) +[2026-06-24 13:16:50,432][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.0009227867121808231, 'train/grad_norm_pre_clip': 1.146181583404541, 'timing/data': 0.0003164431545883417, 'timing/model': 1.0777350740972906, 'learning_rate/qwen_vl_interface': 4.367456956516864e-06, 'learning_rate/action_model': 4.367456956516864e-05, 'epoch': 0.22}) +[2026-06-24 13:16:51,507][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.023351354524493217, 'train/grad_norm_pre_clip': 33.077693939208984, 'timing/data': 0.00032440805807709694, 'timing/model': 1.074142404133454, 'learning_rate/qwen_vl_interface': 4.3636975685613236e-06, 'learning_rate/action_model': 4.3636975685613236e-05, 'epoch': 0.22}) +[2026-06-24 13:16:52,583][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.08450936526060104, 'train/grad_norm_pre_clip': 45.1851692199707, 'timing/data': 0.0003316011279821396, 'timing/model': 1.0736448550596833, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 0.22}) +[2026-06-24 13:16:53,661][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.000215574968024157, 'train/grad_norm_pre_clip': 0.37906399369239807, 'timing/data': 0.0003187078982591629, 'timing/model': 1.077135527972132, 'learning_rate/qwen_vl_interface': 4.356180520425371e-06, 'learning_rate/action_model': 4.356180520425371e-05, 'epoch': 0.22}) +[2026-06-24 13:16:54,735][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.0014428554568439722, 'train/grad_norm_pre_clip': 1.6606707572937012, 'timing/data': 0.0003544241189956665, 'timing/model': 1.072254505008459, 'learning_rate/qwen_vl_interface': 4.352422865122691e-06, 'learning_rate/action_model': 4.352422865122691e-05, 'epoch': 0.22}) +[2026-06-24 13:16:55,809][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.469425231218338, 'train/grad_norm_pre_clip': 140.4634552001953, 'timing/data': 0.00034205103293061256, 'timing/model': 1.071753207826987, 'learning_rate/qwen_vl_interface': 4.348665792248067e-06, 'learning_rate/action_model': 4.3486657922480666e-05, 'epoch': 0.22}) +[2026-06-24 13:16:56,885][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.09582161903381348, 'train/grad_norm_pre_clip': 280.99664306640625, 'timing/data': 0.00032050092704594135, 'timing/model': 1.0747305199038237, 'learning_rate/qwen_vl_interface': 4.344909304239427e-06, 'learning_rate/action_model': 4.344909304239427e-05, 'epoch': 0.22}) +[2026-06-24 13:16:57,962][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.5301606059074402, 'train/grad_norm_pre_clip': 205.72767639160156, 'timing/data': 0.0003758238162845373, 'timing/model': 1.0750350439921021, 'learning_rate/qwen_vl_interface': 4.341153403534311e-06, 'learning_rate/action_model': 4.34115340353431e-05, 'epoch': 0.22}) +[2026-06-24 13:16:59,043][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.3478677272796631, 'train/grad_norm_pre_clip': 116.26799774169922, 'timing/data': 0.00040764384903013706, 'timing/model': 1.079141120892018, 'learning_rate/qwen_vl_interface': 4.337398092569885e-06, 'learning_rate/action_model': 4.337398092569885e-05, 'epoch': 0.22}) +[2026-06-24 13:17:00,361][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.002777243498712778, 'train/grad_norm_pre_clip': 8.377551078796387, 'timing/data': 0.0003392361104488373, 'timing/model': 1.3159328668843955, 'learning_rate/qwen_vl_interface': 4.333643373782931e-06, 'learning_rate/action_model': 4.333643373782931e-05, 'epoch': 0.22}) +[2026-06-24 13:17:01,439][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.3039618134498596, 'train/grad_norm_pre_clip': 62.714378356933594, 'timing/data': 0.0003430729266256094, 'timing/model': 1.0765578779391944, 'learning_rate/qwen_vl_interface': 4.329889249609838e-06, 'learning_rate/action_model': 4.3298892496098386e-05, 'epoch': 0.22}) +[2026-06-24 13:17:02,518][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.2343965470790863, 'train/grad_norm_pre_clip': 238.26083374023438, 'timing/data': 0.0003226448316127062, 'timing/model': 1.078111673006788, 'learning_rate/qwen_vl_interface': 4.326135722486626e-06, 'learning_rate/action_model': 4.3261357224866256e-05, 'epoch': 0.22}) +[2026-06-24 13:17:03,595][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.006074648816138506, 'train/grad_norm_pre_clip': 11.572518348693848, 'timing/data': 0.0003563370555639267, 'timing/model': 1.0749706579372287, 'learning_rate/qwen_vl_interface': 4.322382794848911e-06, 'learning_rate/action_model': 4.322382794848911e-05, 'epoch': 0.22}) +[2026-06-24 13:17:04,671][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.0003232921881135553, 'train/grad_norm_pre_clip': 0.3308935761451721, 'timing/data': 0.00032985792495310307, 'timing/model': 1.0742118288762867, 'learning_rate/qwen_vl_interface': 4.318630469131931e-06, 'learning_rate/action_model': 4.3186304691319304e-05, 'epoch': 0.22}) +[2026-06-24 13:17:05,755][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 6.914164259796962e-05, 'train/grad_norm_pre_clip': 0.045894887298345566, 'timing/data': 0.00032223411835730076, 'timing/model': 1.0823970399796963, 'learning_rate/qwen_vl_interface': 4.314878747770527e-06, 'learning_rate/action_model': 4.314878747770527e-05, 'epoch': 0.22}) +[2026-06-24 13:17:06,831][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.08009958267211914, 'train/grad_norm_pre_clip': 188.998291015625, 'timing/data': 0.00039925798773765564, 'timing/model': 1.0741334171034396, 'learning_rate/qwen_vl_interface': 4.311127633199151e-06, 'learning_rate/action_model': 4.3111276331991504e-05, 'epoch': 0.22}) +[2026-06-24 13:17:07,909][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.006544441916048527, 'train/grad_norm_pre_clip': 14.132720947265625, 'timing/data': 0.00035552610643208027, 'timing/model': 1.0770344459451735, 'learning_rate/qwen_vl_interface': 4.307377127851861e-06, 'learning_rate/action_model': 4.307377127851861e-05, 'epoch': 0.22}) +[2026-06-24 13:17:08,984][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.31552183628082275, 'train/grad_norm_pre_clip': 97.86679077148438, 'timing/data': 0.00036318995989859104, 'timing/model': 1.0733402948826551, 'learning_rate/qwen_vl_interface': 4.303627234162319e-06, 'learning_rate/action_model': 4.3036272341623184e-05, 'epoch': 0.22}) +[2026-06-24 13:17:10,064][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.008565172553062439, 'train/grad_norm_pre_clip': 24.938758850097656, 'timing/data': 0.000366465887054801, 'timing/model': 1.078336311969906, 'learning_rate/qwen_vl_interface': 4.29987795456379e-06, 'learning_rate/action_model': 4.29987795456379e-05, 'epoch': 0.22}) +[2026-06-24 13:17:11,144][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.00981416180729866, 'train/grad_norm_pre_clip': 21.385093688964844, 'timing/data': 0.0003112431149929762, 'timing/model': 1.0779919479973614, 'learning_rate/qwen_vl_interface': 4.29612929148914e-06, 'learning_rate/action_model': 4.29612929148914e-05, 'epoch': 0.22}) +[2026-06-24 13:17:12,223][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.18459530174732208, 'train/grad_norm_pre_clip': 77.7514419555664, 'timing/data': 0.00034150993451476097, 'timing/model': 1.077633328968659, 'learning_rate/qwen_vl_interface': 4.292381247370838e-06, 'learning_rate/action_model': 4.2923812473708377e-05, 'epoch': 0.22}) +[2026-06-24 13:17:13,297][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.1302681565284729, 'train/grad_norm_pre_clip': 128.60386657714844, 'timing/data': 0.00035423412919044495, 'timing/model': 1.0717964030336589, 'learning_rate/qwen_vl_interface': 4.288633824640948e-06, 'learning_rate/action_model': 4.2886338246409474e-05, 'epoch': 0.22}) +[2026-06-24 13:17:14,390][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.2795078456401825, 'train/grad_norm_pre_clip': 63.17692565917969, 'timing/data': 0.0003953708801418543, 'timing/model': 1.088684604037553, 'learning_rate/qwen_vl_interface': 4.28488702573113e-06, 'learning_rate/action_model': 4.2848870257311296e-05, 'epoch': 0.22}) +[2026-06-24 13:17:15,641][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.26034942269325256, 'train/grad_norm_pre_clip': 147.57647705078125, 'timing/data': 0.0008352948352694511, 'timing/model': 1.2472010829951614, 'learning_rate/qwen_vl_interface': 4.281140853072644e-06, 'learning_rate/action_model': 4.281140853072645e-05, 'epoch': 0.22}) +[2026-06-24 13:17:16,755][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.5222309231758118, 'train/grad_norm_pre_clip': 100.61371612548828, 'timing/data': 0.0005456518847495317, 'timing/model': 1.1126071140170097, 'learning_rate/qwen_vl_interface': 4.277395309096338e-06, 'learning_rate/action_model': 4.277395309096338e-05, 'epoch': 0.22}) +[2026-06-24 13:17:17,851][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.0006517113652080297, 'train/grad_norm_pre_clip': 1.0248693227767944, 'timing/data': 0.0006038020364940166, 'timing/model': 1.0932177391368896, 'learning_rate/qwen_vl_interface': 4.27365039623266e-06, 'learning_rate/action_model': 4.27365039623266e-05, 'epoch': 0.22}) +[2026-06-24 13:17:19,166][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.00020230219524819404, 'train/grad_norm_pre_clip': 0.6184038519859314, 'timing/data': 0.0004196171648800373, 'timing/model': 1.3130529730115086, 'learning_rate/qwen_vl_interface': 4.269906116911639e-06, 'learning_rate/action_model': 4.269906116911638e-05, 'epoch': 0.22}) +[2026-06-24 13:17:20,264][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.23015965521335602, 'train/grad_norm_pre_clip': 102.29840850830078, 'timing/data': 0.0007409180980175734, 'timing/model': 1.0955194318667054, 'learning_rate/qwen_vl_interface': 4.2661624735628945e-06, 'learning_rate/action_model': 4.266162473562895e-05, 'epoch': 0.22}) +[2026-06-24 13:17:21,344][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.051448360085487366, 'train/grad_norm_pre_clip': 60.70317840576172, 'timing/data': 0.0004036370664834976, 'timing/model': 1.077637110138312, 'learning_rate/qwen_vl_interface': 4.262419468615644e-06, 'learning_rate/action_model': 4.2624194686156435e-05, 'epoch': 0.22}) +[2026-06-24 13:17:22,424][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.11322135478258133, 'train/grad_norm_pre_clip': 48.311283111572266, 'timing/data': 0.0003999900072813034, 'timing/model': 1.078520396957174, 'learning_rate/qwen_vl_interface': 4.258677104498675e-06, 'learning_rate/action_model': 4.2586771044986744e-05, 'epoch': 0.22}) +[2026-06-24 13:17:23,507][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.0657513365149498, 'train/grad_norm_pre_clip': 57.863826751708984, 'timing/data': 0.0007921650540083647, 'timing/model': 1.0803889208473265, 'learning_rate/qwen_vl_interface': 4.254935383640372e-06, 'learning_rate/action_model': 4.254935383640372e-05, 'epoch': 0.22}) +[2026-06-24 13:17:24,586][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.23980014026165009, 'train/grad_norm_pre_clip': 187.61940002441406, 'timing/data': 0.0005029530730098486, 'timing/model': 1.076690220972523, 'learning_rate/qwen_vl_interface': 4.2511943084686965e-06, 'learning_rate/action_model': 4.251194308468697e-05, 'epoch': 0.22}) +[2026-06-24 13:17:25,663][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.05745340511202812, 'train/grad_norm_pre_clip': 40.07009506225586, 'timing/data': 0.0004446841776371002, 'timing/model': 1.0757841048762202, 'learning_rate/qwen_vl_interface': 4.247453881411187e-06, 'learning_rate/action_model': 4.247453881411187e-05, 'epoch': 0.22}) +[2026-06-24 13:17:26,744][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.02085912600159645, 'train/grad_norm_pre_clip': 36.33649444580078, 'timing/data': 0.00039289682172238827, 'timing/model': 1.0789503909181803, 'learning_rate/qwen_vl_interface': 4.243714104894972e-06, 'learning_rate/action_model': 4.243714104894972e-05, 'epoch': 0.22}) +[2026-06-24 13:17:27,828][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.016660051420331, 'train/grad_norm_pre_clip': 72.35921478271484, 'timing/data': 0.0004375600256025791, 'timing/model': 1.0815999389160424, 'learning_rate/qwen_vl_interface': 4.239974981346748e-06, 'learning_rate/action_model': 4.239974981346748e-05, 'epoch': 0.22}) +[2026-06-24 13:17:28,909][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.34828487038612366, 'train/grad_norm_pre_clip': 86.92723846435547, 'timing/data': 0.0004394629504531622, 'timing/model': 1.0785200879909098, 'learning_rate/qwen_vl_interface': 4.236236513192793e-06, 'learning_rate/action_model': 4.236236513192793e-05, 'epoch': 0.22}) +[2026-06-24 13:17:29,992][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.2111080437898636, 'train/grad_norm_pre_clip': 38.29434585571289, 'timing/data': 0.0004179030656814575, 'timing/model': 1.0813120328821242, 'learning_rate/qwen_vl_interface': 4.2324987028589595e-06, 'learning_rate/action_model': 4.232498702858959e-05, 'epoch': 0.22}) +[2026-06-24 13:17:31,071][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.009069510735571384, 'train/grad_norm_pre_clip': 10.563640594482422, 'timing/data': 0.00040120212361216545, 'timing/model': 1.077671302948147, 'learning_rate/qwen_vl_interface': 4.228761552770669e-06, 'learning_rate/action_model': 4.228761552770669e-05, 'epoch': 0.22}) +[2026-06-24 13:17:32,150][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.036094702780246735, 'train/grad_norm_pre_clip': 37.197383880615234, 'timing/data': 0.0004371087998151779, 'timing/model': 1.0768526308238506, 'learning_rate/qwen_vl_interface': 4.225025065352921e-06, 'learning_rate/action_model': 4.225025065352921e-05, 'epoch': 0.22}) +[2026-06-24 13:17:33,225][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.19695793092250824, 'train/grad_norm_pre_clip': 65.87351989746094, 'timing/data': 0.000401392113417387, 'timing/model': 1.0726387521717697, 'learning_rate/qwen_vl_interface': 4.2212892430302795e-06, 'learning_rate/action_model': 4.2212892430302795e-05, 'epoch': 0.22}) +[2026-06-24 13:17:34,300][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.12344768643379211, 'train/grad_norm_pre_clip': 108.82768249511719, 'timing/data': 0.0003908530343323946, 'timing/model': 1.072873645927757, 'learning_rate/qwen_vl_interface': 4.21755408822688e-06, 'learning_rate/action_model': 4.21755408822688e-05, 'epoch': 0.22}) +[2026-06-24 13:17:35,378][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.1012929379940033, 'train/grad_norm_pre_clip': 89.28089904785156, 'timing/data': 0.000414437148720026, 'timing/model': 1.0766687609720975, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 0.22}) +[2026-06-24 13:17:36,455][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.14819659292697906, 'train/grad_norm_pre_clip': 45.17449188232422, 'timing/data': 0.00040680309757590294, 'timing/model': 1.0752492509782314, 'learning_rate/qwen_vl_interface': 4.210085790872176e-06, 'learning_rate/action_model': 4.2100857908721765e-05, 'epoch': 0.22}) +[2026-06-24 13:17:37,538][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.4195023477077484, 'train/grad_norm_pre_clip': 175.50669860839844, 'timing/data': 0.00045010284520685673, 'timing/model': 1.0806140180211514, 'learning_rate/qwen_vl_interface': 4.20635265316697e-06, 'learning_rate/action_model': 4.206352653166969e-05, 'epoch': 0.22}) +[2026-06-24 13:17:38,908][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.0893278643488884, 'train/grad_norm_pre_clip': 25.113807678222656, 'timing/data': 0.00045819906517863274, 'timing/model': 1.3688130229711533, 'learning_rate/qwen_vl_interface': 4.202620192673195e-06, 'learning_rate/action_model': 4.202620192673195e-05, 'epoch': 0.23}) +[2026-06-24 13:17:39,998][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.1681070178747177, 'train/grad_norm_pre_clip': 165.89498901367188, 'timing/data': 0.0003859829157590866, 'timing/model': 1.0876831361092627, 'learning_rate/qwen_vl_interface': 4.198888411812807e-06, 'learning_rate/action_model': 4.198888411812807e-05, 'epoch': 0.23}) +[2026-06-24 13:17:41,101][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.029265686869621277, 'train/grad_norm_pre_clip': 26.01230812072754, 'timing/data': 0.00043216999620199203, 'timing/model': 1.1007613609544933, 'learning_rate/qwen_vl_interface': 4.195157313007317e-06, 'learning_rate/action_model': 4.195157313007317e-05, 'epoch': 0.23}) +[2026-06-24 13:17:42,212][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.08848497271537781, 'train/grad_norm_pre_clip': 47.84970474243164, 'timing/data': 0.0006799041293561459, 'timing/model': 1.108299263054505, 'learning_rate/qwen_vl_interface': 4.1914268986777926e-06, 'learning_rate/action_model': 4.191426898677792e-05, 'epoch': 0.23}) +[2026-06-24 13:17:43,332][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.06008652225136757, 'train/grad_norm_pre_clip': 51.94406509399414, 'timing/data': 0.0008318191394209862, 'timing/model': 1.1169615308754146, 'learning_rate/qwen_vl_interface': 4.1876971712448626e-06, 'learning_rate/action_model': 4.187697171244863e-05, 'epoch': 0.23}) +[2026-06-24 13:17:44,432][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.38543885946273804, 'train/grad_norm_pre_clip': 83.68037414550781, 'timing/data': 0.000626114197075367, 'timing/model': 1.0981688180472702, 'learning_rate/qwen_vl_interface': 4.183968133128702e-06, 'learning_rate/action_model': 4.183968133128702e-05, 'epoch': 0.23}) +[2026-06-24 13:17:45,511][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.2733350694179535, 'train/grad_norm_pre_clip': 92.1302261352539, 'timing/data': 0.0003646139521151781, 'timing/model': 1.0773105828557163, 'learning_rate/qwen_vl_interface': 4.180239786749049e-06, 'learning_rate/action_model': 4.18023978674905e-05, 'epoch': 0.23}) +[2026-06-24 13:17:46,583][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.01874469220638275, 'train/grad_norm_pre_clip': 23.38416290283203, 'timing/data': 0.0003284849226474762, 'timing/model': 1.0709205069579184, 'learning_rate/qwen_vl_interface': 4.176512134525183e-06, 'learning_rate/action_model': 4.176512134525184e-05, 'epoch': 0.23}) +[2026-06-24 13:17:47,656][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.015819722786545753, 'train/grad_norm_pre_clip': 6.503567695617676, 'timing/data': 0.00037832907401025295, 'timing/model': 1.070545997004956, 'learning_rate/qwen_vl_interface': 4.172785178875939e-06, 'learning_rate/action_model': 4.1727851788759395e-05, 'epoch': 0.23}) +[2026-06-24 13:17:48,727][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.07255574315786362, 'train/grad_norm_pre_clip': 93.17952728271484, 'timing/data': 0.000344857107847929, 'timing/model': 1.0693503068760037, 'learning_rate/qwen_vl_interface': 4.169058922219698e-06, 'learning_rate/action_model': 4.1690589222196976e-05, 'epoch': 0.23}) +[2026-06-24 13:17:49,801][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.23249724507331848, 'train/grad_norm_pre_clip': 61.71852111816406, 'timing/data': 0.00031556212343275547, 'timing/model': 1.072294546989724, 'learning_rate/qwen_vl_interface': 4.165333366974385e-06, 'learning_rate/action_model': 4.165333366974385e-05, 'epoch': 0.23}) +[2026-06-24 13:17:50,875][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.010531795211136341, 'train/grad_norm_pre_clip': 13.903707504272461, 'timing/data': 0.00033164210617542267, 'timing/model': 1.0725156820844859, 'learning_rate/qwen_vl_interface': 4.161608515557475e-06, 'learning_rate/action_model': 4.161608515557475e-05, 'epoch': 0.23}) +[2026-06-24 13:17:51,944][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.021849455311894417, 'train/grad_norm_pre_clip': 17.73566246032715, 'timing/data': 0.00031632278114557266, 'timing/model': 1.067584760952741, 'learning_rate/qwen_vl_interface': 4.157884370385983e-06, 'learning_rate/action_model': 4.157884370385983e-05, 'epoch': 0.23}) +[2026-06-24 13:17:53,014][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.25865092873573303, 'train/grad_norm_pre_clip': 82.27815246582031, 'timing/data': 0.0003128359094262123, 'timing/model': 1.0689125719945878, 'learning_rate/qwen_vl_interface': 4.154160933876465e-06, 'learning_rate/action_model': 4.1541609338764654e-05, 'epoch': 0.23}) +[2026-06-24 13:17:54,089][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.1589839905500412, 'train/grad_norm_pre_clip': 41.370513916015625, 'timing/data': 0.00031656306236982346, 'timing/model': 1.0727919461205602, 'learning_rate/qwen_vl_interface': 4.150438208445022e-06, 'learning_rate/action_model': 4.150438208445021e-05, 'epoch': 0.23}) +[2026-06-24 13:17:55,162][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.1389344483613968, 'train/grad_norm_pre_clip': 77.33362579345703, 'timing/data': 0.00040146196261048317, 'timing/model': 1.0713900381233543, 'learning_rate/qwen_vl_interface': 4.146716196507285e-06, 'learning_rate/action_model': 4.146716196507285e-05, 'epoch': 0.23}) +[2026-06-24 13:17:56,239][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.04356458783149719, 'train/grad_norm_pre_clip': 51.452022552490234, 'timing/data': 0.0006290690507739782, 'timing/model': 1.074595828074962, 'learning_rate/qwen_vl_interface': 4.142994900478432e-06, 'learning_rate/action_model': 4.1429949004784315e-05, 'epoch': 0.23}) +[2026-06-24 13:17:57,551][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.10156943649053574, 'train/grad_norm_pre_clip': 66.109130859375, 'timing/data': 0.0003238269127905369, 'timing/model': 1.3102941028773785, 'learning_rate/qwen_vl_interface': 4.139274322773168e-06, 'learning_rate/action_model': 4.139274322773168e-05, 'epoch': 0.23}) +[2026-06-24 13:17:58,627][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.011474217288196087, 'train/grad_norm_pre_clip': 7.253912448883057, 'timing/data': 0.0003619890194386244, 'timing/model': 1.0747928430791944, 'learning_rate/qwen_vl_interface': 4.135554465805737e-06, 'learning_rate/action_model': 4.135554465805737e-05, 'epoch': 0.23}) +[2026-06-24 13:17:59,702][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.2004307359457016, 'train/grad_norm_pre_clip': 62.47782516479492, 'timing/data': 0.0003062740433961153, 'timing/model': 1.0734785599634051, 'learning_rate/qwen_vl_interface': 4.131835331989916e-06, 'learning_rate/action_model': 4.1318353319899156e-05, 'epoch': 0.23}) +[2026-06-24 13:18:00,775][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.07911691069602966, 'train/grad_norm_pre_clip': 63.52238464355469, 'timing/data': 0.0003158121835440397, 'timing/model': 1.071000105002895, 'learning_rate/qwen_vl_interface': 4.128116923739006e-06, 'learning_rate/action_model': 4.128116923739005e-05, 'epoch': 0.23}) +[2026-06-24 13:18:01,852][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.0987192913889885, 'train/grad_norm_pre_clip': 72.0845718383789, 'timing/data': 0.00035784090869128704, 'timing/model': 1.07572181802243, 'learning_rate/qwen_vl_interface': 4.124399243465845e-06, 'learning_rate/action_model': 4.1243992434658445e-05, 'epoch': 0.23}) +[2026-06-24 13:18:02,933][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.017268503084778786, 'train/grad_norm_pre_clip': 24.047191619873047, 'timing/data': 0.0005164281465113163, 'timing/model': 1.0786984888836741, 'learning_rate/qwen_vl_interface': 4.120682293582792e-06, 'learning_rate/action_model': 4.120682293582791e-05, 'epoch': 0.23}) +[2026-06-24 13:18:04,012][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.052884168922901154, 'train/grad_norm_pre_clip': 78.85200500488281, 'timing/data': 0.00037828879430890083, 'timing/model': 1.0775008550845087, 'learning_rate/qwen_vl_interface': 4.1169660765017416e-06, 'learning_rate/action_model': 4.1169660765017414e-05, 'epoch': 0.23}) +[2026-06-24 13:18:05,085][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.05842982232570648, 'train/grad_norm_pre_clip': 14.358421325683594, 'timing/data': 0.0003147199749946594, 'timing/model': 1.070815684972331, 'learning_rate/qwen_vl_interface': 4.113250594634103e-06, 'learning_rate/action_model': 4.113250594634102e-05, 'epoch': 0.23}) +[2026-06-24 13:18:06,158][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.05661211907863617, 'train/grad_norm_pre_clip': 44.780941009521484, 'timing/data': 0.0005171289667487144, 'timing/model': 1.0714733588974923, 'learning_rate/qwen_vl_interface': 4.109535850390808e-06, 'learning_rate/action_model': 4.109535850390808e-05, 'epoch': 0.23}) +[2026-06-24 13:18:07,233][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.05382620543241501, 'train/grad_norm_pre_clip': 12.033512115478516, 'timing/data': 0.0003253601025789976, 'timing/model': 1.0735819227993488, 'learning_rate/qwen_vl_interface': 4.105821846182322e-06, 'learning_rate/action_model': 4.1058218461823215e-05, 'epoch': 0.23}) +[2026-06-24 13:18:08,307][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.04031188413500786, 'train/grad_norm_pre_clip': 32.447418212890625, 'timing/data': 0.0005568340420722961, 'timing/model': 1.0717275871429592, 'learning_rate/qwen_vl_interface': 4.102108584418614e-06, 'learning_rate/action_model': 4.1021085844186136e-05, 'epoch': 0.23}) +[2026-06-24 13:18:09,382][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.002408119849860668, 'train/grad_norm_pre_clip': 1.7392215728759766, 'timing/data': 0.0003410989884287119, 'timing/model': 1.0737088229507208, 'learning_rate/qwen_vl_interface': 4.098396067509186e-06, 'learning_rate/action_model': 4.098396067509186e-05, 'epoch': 0.23}) +[2026-06-24 13:18:10,457][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.09769000113010406, 'train/grad_norm_pre_clip': 44.24384307861328, 'timing/data': 0.0003552059642970562, 'timing/model': 1.0725890940520912, 'learning_rate/qwen_vl_interface': 4.094684297863045e-06, 'learning_rate/action_model': 4.094684297863045e-05, 'epoch': 0.23}) +[2026-06-24 13:18:11,533][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.39339014887809753, 'train/grad_norm_pre_clip': 18.579273223876953, 'timing/data': 0.00032920786179602146, 'timing/model': 1.0749195739626884, 'learning_rate/qwen_vl_interface': 4.090973277888722e-06, 'learning_rate/action_model': 4.0909732778887215e-05, 'epoch': 0.23}) +[2026-06-24 13:18:12,606][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.17998425662517548, 'train/grad_norm_pre_clip': 104.12869262695312, 'timing/data': 0.0003313310444355011, 'timing/model': 1.0711778099648654, 'learning_rate/qwen_vl_interface': 4.0872630099942555e-06, 'learning_rate/action_model': 4.087263009994256e-05, 'epoch': 0.23}) +[2026-06-24 13:18:13,683][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.30325037240982056, 'train/grad_norm_pre_clip': 53.08278274536133, 'timing/data': 0.00034576794132590294, 'timing/model': 1.075205722823739, 'learning_rate/qwen_vl_interface': 4.083553496587197e-06, 'learning_rate/action_model': 4.0835534965871966e-05, 'epoch': 0.23}) +[2026-06-24 13:18:14,755][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.18925638496875763, 'train/grad_norm_pre_clip': 72.20802307128906, 'timing/data': 0.0003304791171103716, 'timing/model': 1.0701957810670137, 'learning_rate/qwen_vl_interface': 4.079844740074611e-06, 'learning_rate/action_model': 4.0798447400746115e-05, 'epoch': 0.23}) +[2026-06-24 13:18:16,048][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.18616653978824615, 'train/grad_norm_pre_clip': 71.87908935546875, 'timing/data': 0.0003083180636167526, 'timing/model': 1.2910412398632616, 'learning_rate/qwen_vl_interface': 4.0761367428630685e-06, 'learning_rate/action_model': 4.0761367428630685e-05, 'epoch': 0.23}) +[2026-06-24 13:18:17,207][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.10500973463058472, 'train/grad_norm_pre_clip': 37.4651985168457, 'timing/data': 0.0005613029934465885, 'timing/model': 1.156934225000441, 'learning_rate/qwen_vl_interface': 4.072429507358651e-06, 'learning_rate/action_model': 4.072429507358651e-05, 'epoch': 0.23}) +[2026-06-24 13:18:18,283][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.10949604213237762, 'train/grad_norm_pre_clip': 63.056724548339844, 'timing/data': 0.0005050969775766134, 'timing/model': 1.0742080688942224, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 0.23}) +[2026-06-24 13:18:19,355][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.12367197126150131, 'train/grad_norm_pre_clip': 42.808868408203125, 'timing/data': 0.0003797919489443302, 'timing/model': 1.069830052088946, 'learning_rate/qwen_vl_interface': 4.065017331093027e-06, 'learning_rate/action_model': 4.065017331093026e-05, 'epoch': 0.23}) +[2026-06-24 13:18:20,426][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.1410091668367386, 'train/grad_norm_pre_clip': 38.863494873046875, 'timing/data': 0.00035478500649333, 'timing/model': 1.069934719009325, 'learning_rate/qwen_vl_interface': 4.061312395141502e-06, 'learning_rate/action_model': 4.0613123951415015e-05, 'epoch': 0.23}) +[2026-06-24 13:18:21,501][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.014032901264727116, 'train/grad_norm_pre_clip': 15.998160362243652, 'timing/data': 0.00033418601378798485, 'timing/model': 1.07331403484568, 'learning_rate/qwen_vl_interface': 4.0576082305164575e-06, 'learning_rate/action_model': 4.057608230516457e-05, 'epoch': 0.23}) +[2026-06-24 13:18:22,574][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.017849067226052284, 'train/grad_norm_pre_clip': 15.984445571899414, 'timing/data': 0.00035010697320103645, 'timing/model': 1.0709716640412807, 'learning_rate/qwen_vl_interface': 4.053904839621487e-06, 'learning_rate/action_model': 4.053904839621487e-05, 'epoch': 0.23}) +[2026-06-24 13:18:23,646][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.1393301784992218, 'train/grad_norm_pre_clip': 48.28053665161133, 'timing/data': 0.0003435539547353983, 'timing/model': 1.0701722661033273, 'learning_rate/qwen_vl_interface': 4.050202224859682e-06, 'learning_rate/action_model': 4.050202224859682e-05, 'epoch': 0.23}) +[2026-06-24 13:18:24,718][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.08791663497686386, 'train/grad_norm_pre_clip': 34.63707733154297, 'timing/data': 0.00035769096575677395, 'timing/model': 1.0711086420342326, 'learning_rate/qwen_vl_interface': 4.046500388633625e-06, 'learning_rate/action_model': 4.046500388633625e-05, 'epoch': 0.23}) +[2026-06-24 13:18:25,787][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.054660532623529434, 'train/grad_norm_pre_clip': 32.37910842895508, 'timing/data': 0.0003306199796497822, 'timing/model': 1.0674101691693068, 'learning_rate/qwen_vl_interface': 4.042799333345401e-06, 'learning_rate/action_model': 4.042799333345401e-05, 'epoch': 0.23}) +[2026-06-24 13:18:26,859][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.06912482529878616, 'train/grad_norm_pre_clip': 21.65895652770996, 'timing/data': 0.0003574998117983341, 'timing/model': 1.0699395008850843, 'learning_rate/qwen_vl_interface': 4.0390990613965816e-06, 'learning_rate/action_model': 4.0390990613965814e-05, 'epoch': 0.23}) +[2026-06-24 13:18:27,932][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.19663386046886444, 'train/grad_norm_pre_clip': 64.42066192626953, 'timing/data': 0.00033734296448528767, 'timing/model': 1.07130823796615, 'learning_rate/qwen_vl_interface': 4.035399575188237e-06, 'learning_rate/action_model': 4.035399575188237e-05, 'epoch': 0.23}) +[2026-06-24 13:18:29,007][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.0015051794471219182, 'train/grad_norm_pre_clip': 0.5406044125556946, 'timing/data': 0.00032981904223561287, 'timing/model': 1.0731328590773046, 'learning_rate/qwen_vl_interface': 4.031700877120921e-06, 'learning_rate/action_model': 4.031700877120921e-05, 'epoch': 0.23}) +[2026-06-24 13:18:30,082][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.1727340966463089, 'train/grad_norm_pre_clip': 38.752620697021484, 'timing/data': 0.0003254399634897709, 'timing/model': 1.0736944728996605, 'learning_rate/qwen_vl_interface': 4.028002969594677e-06, 'learning_rate/action_model': 4.0280029695946766e-05, 'epoch': 0.23}) +[2026-06-24 13:18:31,171][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.16537924110889435, 'train/grad_norm_pre_clip': 24.561731338500977, 'timing/data': 0.0003097208682447672, 'timing/model': 1.0872559999115765, 'learning_rate/qwen_vl_interface': 4.0243058550090416e-06, 'learning_rate/action_model': 4.0243058550090416e-05, 'epoch': 0.23}) +[2026-06-24 13:18:32,365][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.03520536050200462, 'train/grad_norm_pre_clip': 13.554715156555176, 'timing/data': 0.0006424549501389265, 'timing/model': 1.1912405458278954, 'learning_rate/qwen_vl_interface': 4.0206095357630256e-06, 'learning_rate/action_model': 4.020609535763026e-05, 'epoch': 0.23}) +[2026-06-24 13:18:33,673][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.07155979424715042, 'train/grad_norm_pre_clip': 54.88273620605469, 'timing/data': 0.0005040550604462624, 'timing/model': 1.3062704240437597, 'learning_rate/qwen_vl_interface': 4.016914014255138e-06, 'learning_rate/action_model': 4.016914014255138e-05, 'epoch': 0.23}) +[2026-06-24 13:18:34,748][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.08085734397172928, 'train/grad_norm_pre_clip': 28.818416595458984, 'timing/data': 0.00032245391048491, 'timing/model': 1.072375797200948, 'learning_rate/qwen_vl_interface': 4.0132192928833606e-06, 'learning_rate/action_model': 4.01321929288336e-05, 'epoch': 0.23}) +[2026-06-24 13:18:35,823][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.046897657215595245, 'train/grad_norm_pre_clip': 7.0217695236206055, 'timing/data': 0.0003313319757580757, 'timing/model': 1.0733417270239443, 'learning_rate/qwen_vl_interface': 4.009525374045152e-06, 'learning_rate/action_model': 4.0095253740451524e-05, 'epoch': 0.23}) +[2026-06-24 13:18:36,896][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.06363996863365173, 'train/grad_norm_pre_clip': 31.028993606567383, 'timing/data': 0.000339356018230319, 'timing/model': 1.0714616130571812, 'learning_rate/qwen_vl_interface': 4.005832260137464e-06, 'learning_rate/action_model': 4.005832260137463e-05, 'epoch': 0.23}) +[2026-06-24 13:18:37,969][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.03604615107178688, 'train/grad_norm_pre_clip': 21.857118606567383, 'timing/data': 0.0003300691023468971, 'timing/model': 1.0718168790917844, 'learning_rate/qwen_vl_interface': 4.002139953556712e-06, 'learning_rate/action_model': 4.002139953556712e-05, 'epoch': 0.23}) +[2026-06-24 13:18:39,041][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.03858918696641922, 'train/grad_norm_pre_clip': 29.621416091918945, 'timing/data': 0.0003264120314270258, 'timing/model': 1.070280720014125, 'learning_rate/qwen_vl_interface': 3.998448456698799e-06, 'learning_rate/action_model': 3.9984484566987984e-05, 'epoch': 0.23}) +[2026-06-24 13:18:40,109][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.07699505239725113, 'train/grad_norm_pre_clip': 21.964635848999023, 'timing/data': 0.0003564590588212013, 'timing/model': 1.0661936569958925, 'learning_rate/qwen_vl_interface': 3.994757771959092e-06, 'learning_rate/action_model': 3.994757771959092e-05, 'epoch': 0.23}) +[2026-06-24 13:18:41,181][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.008171088993549347, 'train/grad_norm_pre_clip': 2.607671022415161, 'timing/data': 0.0003255200572311878, 'timing/model': 1.0710724650416523, 'learning_rate/qwen_vl_interface': 3.9910679017324405e-06, 'learning_rate/action_model': 3.9910679017324405e-05, 'epoch': 0.23}) +[2026-06-24 13:18:42,253][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.0041982210241258144, 'train/grad_norm_pre_clip': 2.707139730453491, 'timing/data': 0.0003301489632576704, 'timing/model': 1.0700044620316476, 'learning_rate/qwen_vl_interface': 3.9873788484131594e-06, 'learning_rate/action_model': 3.98737884841316e-05, 'epoch': 0.23}) +[2026-06-24 13:18:43,326][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.08610357344150543, 'train/grad_norm_pre_clip': 49.52748107910156, 'timing/data': 0.0003381140995770693, 'timing/model': 1.0714955490548164, 'learning_rate/qwen_vl_interface': 3.983690614395035e-06, 'learning_rate/action_model': 3.983690614395035e-05, 'epoch': 0.23}) +[2026-06-24 13:18:44,398][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.06341256201267242, 'train/grad_norm_pre_clip': 46.450687408447266, 'timing/data': 0.0003162031061947346, 'timing/model': 1.0704795129131526, 'learning_rate/qwen_vl_interface': 3.980003202071324e-06, 'learning_rate/action_model': 3.9800032020713235e-05, 'epoch': 0.23}) +[2026-06-24 13:18:45,470][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.28148531913757324, 'train/grad_norm_pre_clip': 51.537689208984375, 'timing/data': 0.0003273140173405409, 'timing/model': 1.0697943191044033, 'learning_rate/qwen_vl_interface': 3.9763166138347454e-06, 'learning_rate/action_model': 3.976316613834745e-05, 'epoch': 0.23}) +[2026-06-24 13:18:46,544][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.07534290105104446, 'train/grad_norm_pre_clip': 46.820674896240234, 'timing/data': 0.00037482287734746933, 'timing/model': 1.0725852518808097, 'learning_rate/qwen_vl_interface': 3.972630852077488e-06, 'learning_rate/action_model': 3.972630852077488e-05, 'epoch': 0.23}) +[2026-06-24 13:18:47,619][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.12025813013315201, 'train/grad_norm_pre_clip': 50.307437896728516, 'timing/data': 0.0003316819202154875, 'timing/model': 1.0732184699736536, 'learning_rate/qwen_vl_interface': 3.968945919191205e-06, 'learning_rate/action_model': 3.9689459191912046e-05, 'epoch': 0.23}) +[2026-06-24 13:18:48,691][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.04854210093617439, 'train/grad_norm_pre_clip': 28.265661239624023, 'timing/data': 0.0003214131575077772, 'timing/model': 1.0704536160919815, 'learning_rate/qwen_vl_interface': 3.965261817567003e-06, 'learning_rate/action_model': 3.965261817567003e-05, 'epoch': 0.23}) +[2026-06-24 13:18:49,765][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.029902437701821327, 'train/grad_norm_pre_clip': 24.60516357421875, 'timing/data': 0.0004964699037373066, 'timing/model': 1.072362223872915, 'learning_rate/qwen_vl_interface': 3.961578549595462e-06, 'learning_rate/action_model': 3.9615785495954616e-05, 'epoch': 0.23}) +[2026-06-24 13:18:51,047][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.18395084142684937, 'train/grad_norm_pre_clip': 39.970863342285156, 'timing/data': 0.00032794498838484287, 'timing/model': 1.2806051729712635, 'learning_rate/qwen_vl_interface': 3.95789611766661e-06, 'learning_rate/action_model': 3.95789611766661e-05, 'epoch': 0.23}) +[2026-06-24 13:18:52,123][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.18474537134170532, 'train/grad_norm_pre_clip': 64.75643157958984, 'timing/data': 0.0005111179780215025, 'timing/model': 1.0736551720183343, 'learning_rate/qwen_vl_interface': 3.954214524169941e-06, 'learning_rate/action_model': 3.954214524169941e-05, 'epoch': 0.23}) +[2026-06-24 13:18:53,198][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.004466901067644358, 'train/grad_norm_pre_clip': 2.8320510387420654, 'timing/data': 0.000319149112328887, 'timing/model': 1.0730745929758996, 'learning_rate/qwen_vl_interface': 3.950533771494399e-06, 'learning_rate/action_model': 3.950533771494399e-05, 'epoch': 0.23}) +[2026-06-24 13:18:54,271][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.07776357233524323, 'train/grad_norm_pre_clip': 47.344661712646484, 'timing/data': 0.00033766403794288635, 'timing/model': 1.0720834240783006, 'learning_rate/qwen_vl_interface': 3.946853862028382e-06, 'learning_rate/action_model': 3.946853862028382e-05, 'epoch': 0.23}) +[2026-06-24 13:18:55,349][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.01606757752597332, 'train/grad_norm_pre_clip': 30.268665313720703, 'timing/data': 0.0003348069731146097, 'timing/model': 1.0763912010006607, 'learning_rate/qwen_vl_interface': 3.94317479815975e-06, 'learning_rate/action_model': 3.943174798159749e-05, 'epoch': 0.23}) +[2026-06-24 13:18:56,425][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.28377774357795715, 'train/grad_norm_pre_clip': 68.18607330322266, 'timing/data': 0.0003488040529191494, 'timing/model': 1.0737477748189121, 'learning_rate/qwen_vl_interface': 3.9394965822758e-06, 'learning_rate/action_model': 3.9394965822757995e-05, 'epoch': 0.23}) +[2026-06-24 13:18:57,499][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.10186679661273956, 'train/grad_norm_pre_clip': 38.23693084716797, 'timing/data': 0.00033493805676698685, 'timing/model': 1.0723801399581134, 'learning_rate/qwen_vl_interface': 3.935819216763292e-06, 'learning_rate/action_model': 3.935819216763292e-05, 'epoch': 0.23}) +[2026-06-24 13:18:58,573][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.018670057877898216, 'train/grad_norm_pre_clip': 19.61894989013672, 'timing/data': 0.00033044908195734024, 'timing/model': 1.0722750930581242, 'learning_rate/qwen_vl_interface': 3.932142704008425e-06, 'learning_rate/action_model': 3.9321427040084255e-05, 'epoch': 0.23}) +[2026-06-24 13:18:59,649][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.44617730379104614, 'train/grad_norm_pre_clip': 52.45320510864258, 'timing/data': 0.00037306989543139935, 'timing/model': 1.0751107011456043, 'learning_rate/qwen_vl_interface': 3.92846704639685e-06, 'learning_rate/action_model': 3.9284670463968495e-05, 'epoch': 0.23}) +[2026-06-24 13:19:00,724][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.058483924716711044, 'train/grad_norm_pre_clip': 46.10493469238281, 'timing/data': 0.000413163797929883, 'timing/model': 1.0728520948905498, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 0.23}) +[2026-06-24 13:19:01,803][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.17699576914310455, 'train/grad_norm_pre_clip': 84.99208068847656, 'timing/data': 0.00035309186205267906, 'timing/model': 1.077334470115602, 'learning_rate/qwen_vl_interface': 3.921118306143396e-06, 'learning_rate/action_model': 3.921118306143396e-05, 'epoch': 0.23}) +[2026-06-24 13:19:02,878][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.15611234307289124, 'train/grad_norm_pre_clip': 23.656389236450195, 'timing/data': 0.00034295301884412766, 'timing/model': 1.0733483550138772, 'learning_rate/qwen_vl_interface': 3.917445228270035e-06, 'learning_rate/action_model': 3.9174452282700354e-05, 'epoch': 0.23}) +[2026-06-24 13:19:03,953][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.0016510923160240054, 'train/grad_norm_pre_clip': 0.6706105470657349, 'timing/data': 0.0003209109418094158, 'timing/model': 1.0739788771606982, 'learning_rate/qwen_vl_interface': 3.913773015076997e-06, 'learning_rate/action_model': 3.9137730150769966e-05, 'epoch': 0.23}) +[2026-06-24 13:19:05,027][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.07807918637990952, 'train/grad_norm_pre_clip': 75.0286636352539, 'timing/data': 0.0003159330226480961, 'timing/model': 1.0721326959319413, 'learning_rate/qwen_vl_interface': 3.9101016689471435e-06, 'learning_rate/action_model': 3.910101668947143e-05, 'epoch': 0.23}) +[2026-06-24 13:19:06,099][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.09295012056827545, 'train/grad_norm_pre_clip': 30.774930953979492, 'timing/data': 0.00031652417965233326, 'timing/model': 1.0706470599398017, 'learning_rate/qwen_vl_interface': 3.906431192262769e-06, 'learning_rate/action_model': 3.906431192262769e-05, 'epoch': 0.23}) +[2026-06-24 13:19:07,174][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.10716281086206436, 'train/grad_norm_pre_clip': 49.819068908691406, 'timing/data': 0.00031415908597409725, 'timing/model': 1.0732851459179074, 'learning_rate/qwen_vl_interface': 3.9027615874056075e-06, 'learning_rate/action_model': 3.902761587405607e-05, 'epoch': 0.23}) +[2026-06-24 13:19:08,456][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.09374327212572098, 'train/grad_norm_pre_clip': 38.17982482910156, 'timing/data': 0.0003380938433110714, 'timing/model': 1.2800774169154465, 'learning_rate/qwen_vl_interface': 3.899092856756825e-06, 'learning_rate/action_model': 3.8990928567568246e-05, 'epoch': 0.23}) +[2026-06-24 13:19:09,532][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.01776471734046936, 'train/grad_norm_pre_clip': 14.340742111206055, 'timing/data': 0.00034829298965632915, 'timing/model': 1.0746985881123692, 'learning_rate/qwen_vl_interface': 3.895425002697019e-06, 'learning_rate/action_model': 3.895425002697019e-05, 'epoch': 0.23}) +[2026-06-24 13:19:10,608][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.21520468592643738, 'train/grad_norm_pre_clip': 38.0532112121582, 'timing/data': 0.000333655159920454, 'timing/model': 1.0745925488881767, 'learning_rate/qwen_vl_interface': 3.891758027606227e-06, 'learning_rate/action_model': 3.891758027606227e-05, 'epoch': 0.23}) +[2026-06-24 13:19:11,684][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.19150330126285553, 'train/grad_norm_pre_clip': 97.46328735351562, 'timing/data': 0.00034063798375427723, 'timing/model': 1.0741502589080483, 'learning_rate/qwen_vl_interface': 3.888091933863902e-06, 'learning_rate/action_model': 3.888091933863902e-05, 'epoch': 0.23}) +[2026-06-24 13:19:12,761][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.11123267561197281, 'train/grad_norm_pre_clip': 61.13261413574219, 'timing/data': 0.00035396404564380646, 'timing/model': 1.0753841730765998, 'learning_rate/qwen_vl_interface': 3.884426723848934e-06, 'learning_rate/action_model': 3.884426723848934e-05, 'epoch': 0.23}) +[2026-06-24 13:19:13,836][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.016917310655117035, 'train/grad_norm_pre_clip': 18.036590576171875, 'timing/data': 0.00040172296576201916, 'timing/model': 1.073728848947212, 'learning_rate/qwen_vl_interface': 3.8807623999396414e-06, 'learning_rate/action_model': 3.8807623999396414e-05, 'epoch': 0.23}) +[2026-06-24 13:19:14,913][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.020104117691516876, 'train/grad_norm_pre_clip': 17.52760887145996, 'timing/data': 0.00031233602203428745, 'timing/model': 1.0750684500671923, 'learning_rate/qwen_vl_interface': 3.877098964513759e-06, 'learning_rate/action_model': 3.877098964513759e-05, 'epoch': 0.23}) +[2026-06-24 13:19:15,988][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.1426733136177063, 'train/grad_norm_pre_clip': 60.31209182739258, 'timing/data': 0.0003284560516476631, 'timing/model': 1.0731792580336332, 'learning_rate/qwen_vl_interface': 3.873436419948456e-06, 'learning_rate/action_model': 3.8734364199484554e-05, 'epoch': 0.23}) +[2026-06-24 13:19:17,067][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.055582210421562195, 'train/grad_norm_pre_clip': 38.731563568115234, 'timing/data': 0.0003192480653524399, 'timing/model': 1.0779753110837191, 'learning_rate/qwen_vl_interface': 3.869774768620315e-06, 'learning_rate/action_model': 3.869774768620315e-05, 'epoch': 0.23}) +[2026-06-24 13:19:18,147][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.05686100572347641, 'train/grad_norm_pre_clip': 26.412866592407227, 'timing/data': 0.000356798991560936, 'timing/model': 1.0784530770033598, 'learning_rate/qwen_vl_interface': 3.866114012905339e-06, 'learning_rate/action_model': 3.866114012905339e-05, 'epoch': 0.23}) +[2026-06-24 13:19:19,225][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.20212365686893463, 'train/grad_norm_pre_clip': 48.655792236328125, 'timing/data': 0.00036855111829936504, 'timing/model': 1.0764485578984022, 'learning_rate/qwen_vl_interface': 3.862454155178957e-06, 'learning_rate/action_model': 3.862454155178957e-05, 'epoch': 0.23}) +[2026-06-24 13:19:20,303][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.2626182436943054, 'train/grad_norm_pre_clip': 44.421051025390625, 'timing/data': 0.000339495949447155, 'timing/model': 1.0758257689885795, 'learning_rate/qwen_vl_interface': 3.858795197816009e-06, 'learning_rate/action_model': 3.8587951978160095e-05, 'epoch': 0.23}) +[2026-06-24 13:19:21,381][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.1408330649137497, 'train/grad_norm_pre_clip': 82.20510864257812, 'timing/data': 0.00031821709126234055, 'timing/model': 1.0763763808645308, 'learning_rate/qwen_vl_interface': 3.855137143190755e-06, 'learning_rate/action_model': 3.855137143190755e-05, 'epoch': 0.23}) +[2026-06-24 13:19:22,462][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.20276576280593872, 'train/grad_norm_pre_clip': 45.48979949951172, 'timing/data': 0.00042731105349957943, 'timing/model': 1.079069911967963, 'learning_rate/qwen_vl_interface': 3.851479993676863e-06, 'learning_rate/action_model': 3.851479993676863e-05, 'epoch': 0.23}) +[2026-06-24 13:19:23,537][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.021414808928966522, 'train/grad_norm_pre_clip': 20.928476333618164, 'timing/data': 0.00034192088060081005, 'timing/model': 1.0742120710201561, 'learning_rate/qwen_vl_interface': 3.84782375164742e-06, 'learning_rate/action_model': 3.8478237516474194e-05, 'epoch': 0.23}) +[2026-06-24 13:19:24,615][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.002770368941128254, 'train/grad_norm_pre_clip': 1.6915428638458252, 'timing/data': 0.0003526010550558567, 'timing/model': 1.0762285329401493, 'learning_rate/qwen_vl_interface': 3.8441684194749215e-06, 'learning_rate/action_model': 3.844168419474921e-05, 'epoch': 0.23}) +[2026-06-24 13:19:25,957][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.004316738806664944, 'train/grad_norm_pre_clip': 3.3937718868255615, 'timing/data': 0.000351910013705492, 'timing/model': 1.340551400091499, 'learning_rate/qwen_vl_interface': 3.840513999531271e-06, 'learning_rate/action_model': 3.840513999531271e-05, 'epoch': 0.23}) +[2026-06-24 13:19:27,036][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.0011384724639356136, 'train/grad_norm_pre_clip': 0.636965274810791, 'timing/data': 0.0003243179526180029, 'timing/model': 1.0773546060081571, 'learning_rate/qwen_vl_interface': 3.836860494187787e-06, 'learning_rate/action_model': 3.8368604941877864e-05, 'epoch': 0.23}) +[2026-06-24 13:19:28,113][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.23823398351669312, 'train/grad_norm_pre_clip': 43.76277160644531, 'timing/data': 0.00031524081714451313, 'timing/model': 1.0746654029935598, 'learning_rate/qwen_vl_interface': 3.833207905815184e-06, 'learning_rate/action_model': 3.8332079058151835e-05, 'epoch': 0.23}) +[2026-06-24 13:19:29,189][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.142847940325737, 'train/grad_norm_pre_clip': 61.625492095947266, 'timing/data': 0.0003366810269653797, 'timing/model': 1.07437312300317, 'learning_rate/qwen_vl_interface': 3.829556236783588e-06, 'learning_rate/action_model': 3.8295562367835884e-05, 'epoch': 0.23}) +[2026-06-24 13:19:30,268][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.005052529741078615, 'train/grad_norm_pre_clip': 2.6113431453704834, 'timing/data': 0.00036036595702171326, 'timing/model': 1.0773826870135963, 'learning_rate/qwen_vl_interface': 3.825905489462532e-06, 'learning_rate/action_model': 3.825905489462531e-05, 'epoch': 0.23}) +[2026-06-24 13:19:31,346][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.08237612247467041, 'train/grad_norm_pre_clip': 37.633148193359375, 'timing/data': 0.0003272630274295807, 'timing/model': 1.0767494190949947, 'learning_rate/qwen_vl_interface': 3.822255666220939e-06, 'learning_rate/action_model': 3.8222556662209384e-05, 'epoch': 0.24}) +[2026-06-24 13:19:32,422][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.14872877299785614, 'train/grad_norm_pre_clip': 18.3077449798584, 'timing/data': 0.0003511591348797083, 'timing/model': 1.0742492899298668, 'learning_rate/qwen_vl_interface': 3.818606769427146e-06, 'learning_rate/action_model': 3.8186067694271456e-05, 'epoch': 0.24}) +[2026-06-24 13:19:33,500][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.02106340043246746, 'train/grad_norm_pre_clip': 12.230908393859863, 'timing/data': 0.0003729090094566345, 'timing/model': 1.0756789401639253, 'learning_rate/qwen_vl_interface': 3.814958801448876e-06, 'learning_rate/action_model': 3.8149588014488755e-05, 'epoch': 0.24}) +[2026-06-24 13:19:34,576][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.1961512416601181, 'train/grad_norm_pre_clip': 48.64422607421875, 'timing/data': 0.00033016898669302464, 'timing/model': 1.0750848851166666, 'learning_rate/qwen_vl_interface': 3.8113117646532637e-06, 'learning_rate/action_model': 3.811311764653264e-05, 'epoch': 0.24}) +[2026-06-24 13:19:35,663][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.12897323071956635, 'train/grad_norm_pre_clip': 18.721179962158203, 'timing/data': 0.0003189980052411556, 'timing/model': 1.0853462680242956, 'learning_rate/qwen_vl_interface': 3.8076656614068268e-06, 'learning_rate/action_model': 3.8076656614068265e-05, 'epoch': 0.24}) +[2026-06-24 13:19:36,764][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.29503685235977173, 'train/grad_norm_pre_clip': 69.79074096679688, 'timing/data': 0.000438072020187974, 'timing/model': 1.0986054649110883, 'learning_rate/qwen_vl_interface': 3.8040204940754816e-06, 'learning_rate/action_model': 3.804020494075481e-05, 'epoch': 0.24}) +[2026-06-24 13:19:37,975][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.11880643665790558, 'train/grad_norm_pre_clip': 44.69192886352539, 'timing/data': 0.0004615860525518656, 'timing/model': 1.2089247680269182, 'learning_rate/qwen_vl_interface': 3.8003762650245418e-06, 'learning_rate/action_model': 3.8003762650245414e-05, 'epoch': 0.24}) +[2026-06-24 13:19:39,054][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.042488351464271545, 'train/grad_norm_pre_clip': 31.05708885192871, 'timing/data': 0.00040506990626454353, 'timing/model': 1.0769478981383145, 'learning_rate/qwen_vl_interface': 3.796732976618702e-06, 'learning_rate/action_model': 3.796732976618702e-05, 'epoch': 0.24}) +[2026-06-24 13:19:40,132][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.08286803215742111, 'train/grad_norm_pre_clip': 75.21240234375, 'timing/data': 0.0003229950089007616, 'timing/model': 1.076285595074296, 'learning_rate/qwen_vl_interface': 3.793090631222057e-06, 'learning_rate/action_model': 3.793090631222057e-05, 'epoch': 0.24}) +[2026-06-24 13:19:41,208][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.1160162091255188, 'train/grad_norm_pre_clip': 47.10381317138672, 'timing/data': 0.0003154110163450241, 'timing/model': 1.0753359929658473, 'learning_rate/qwen_vl_interface': 3.789449231198086e-06, 'learning_rate/action_model': 3.7894492311980855e-05, 'epoch': 0.24}) +[2026-06-24 13:19:42,510][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.14967215061187744, 'train/grad_norm_pre_clip': 80.08232879638672, 'timing/data': 0.0003329450264573097, 'timing/model': 1.2994624099228531, 'learning_rate/qwen_vl_interface': 3.7858087789096487e-06, 'learning_rate/action_model': 3.7858087789096484e-05, 'epoch': 0.24}) +[2026-06-24 13:19:43,588][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.11389770358800888, 'train/grad_norm_pre_clip': 53.447811126708984, 'timing/data': 0.0003438850399106741, 'timing/model': 1.0770124478731304, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 0.24}) +[2026-06-24 13:19:44,667][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.02814481407403946, 'train/grad_norm_pre_clip': 24.65764045715332, 'timing/data': 0.0003619289491325617, 'timing/model': 1.0771722870413214, 'learning_rate/qwen_vl_interface': 3.7785307269877706e-06, 'learning_rate/action_model': 3.7785307269877705e-05, 'epoch': 0.24}) +[2026-06-24 13:19:45,748][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.091395802795887, 'train/grad_norm_pre_clip': 92.10449981689453, 'timing/data': 0.0003503968473523855, 'timing/model': 1.0792633581440896, 'learning_rate/qwen_vl_interface': 3.774893132076978e-06, 'learning_rate/action_model': 3.774893132076978e-05, 'epoch': 0.24}) +[2026-06-24 13:19:46,831][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.012412160634994507, 'train/grad_norm_pre_clip': 8.630072593688965, 'timing/data': 0.000347101129591465, 'timing/model': 1.0820113900117576, 'learning_rate/qwen_vl_interface': 3.771256494347018e-06, 'learning_rate/action_model': 3.771256494347018e-05, 'epoch': 0.24}) +[2026-06-24 13:19:47,912][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.10683489590883255, 'train/grad_norm_pre_clip': 67.94514465332031, 'timing/data': 0.0003169539850205183, 'timing/model': 1.079027843894437, 'learning_rate/qwen_vl_interface': 3.7676208161576623e-06, 'learning_rate/action_model': 3.767620816157662e-05, 'epoch': 0.24}) +[2026-06-24 13:19:48,995][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.028789609670639038, 'train/grad_norm_pre_clip': 7.79813814163208, 'timing/data': 0.0003085988573729992, 'timing/model': 1.0809421529993415, 'learning_rate/qwen_vl_interface': 3.763986099868066e-06, 'learning_rate/action_model': 3.763986099868066e-05, 'epoch': 0.24}) +[2026-06-24 13:19:50,076][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.04662611335515976, 'train/grad_norm_pre_clip': 30.239389419555664, 'timing/data': 0.00031404895707964897, 'timing/model': 1.0795122811105102, 'learning_rate/qwen_vl_interface': 3.7603523478367555e-06, 'learning_rate/action_model': 3.760352347836755e-05, 'epoch': 0.24}) +[2026-06-24 13:19:51,157][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.03617088869214058, 'train/grad_norm_pre_clip': 12.959179878234863, 'timing/data': 0.0003197400365024805, 'timing/model': 1.0791324581950903, 'learning_rate/qwen_vl_interface': 3.756719562421634e-06, 'learning_rate/action_model': 3.756719562421634e-05, 'epoch': 0.24}) +[2026-06-24 13:19:52,239][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.0024851025082170963, 'train/grad_norm_pre_clip': 0.8316763639450073, 'timing/data': 0.0003573601134121418, 'timing/model': 1.0804495071060956, 'learning_rate/qwen_vl_interface': 3.753087745979977e-06, 'learning_rate/action_model': 3.753087745979977e-05, 'epoch': 0.24}) +[2026-06-24 13:19:53,322][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.12010671198368073, 'train/grad_norm_pre_clip': 47.4202880859375, 'timing/data': 0.0003275249619036913, 'timing/model': 1.0812150321435183, 'learning_rate/qwen_vl_interface': 3.7494569008684284e-06, 'learning_rate/action_model': 3.7494569008684284e-05, 'epoch': 0.24}) +[2026-06-24 13:19:54,403][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.02181084454059601, 'train/grad_norm_pre_clip': 17.811996459960938, 'timing/data': 0.0003628400154411793, 'timing/model': 1.0799097130075097, 'learning_rate/qwen_vl_interface': 3.7458270294430066e-06, 'learning_rate/action_model': 3.745827029443007e-05, 'epoch': 0.24}) +[2026-06-24 13:19:55,482][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.05866805836558342, 'train/grad_norm_pre_clip': 26.896032333374023, 'timing/data': 0.0003575508017092943, 'timing/model': 1.0777028531301767, 'learning_rate/qwen_vl_interface': 3.742198134059093e-06, 'learning_rate/action_model': 3.742198134059093e-05, 'epoch': 0.24}) +[2026-06-24 13:19:56,563][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.1962626576423645, 'train/grad_norm_pre_clip': 59.6910285949707, 'timing/data': 0.00032314611598849297, 'timing/model': 1.0788525179959834, 'learning_rate/qwen_vl_interface': 3.7385702170714434e-06, 'learning_rate/action_model': 3.738570217071443e-05, 'epoch': 0.24}) +[2026-06-24 13:19:57,646][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.01386538241058588, 'train/grad_norm_pre_clip': 13.827879905700684, 'timing/data': 0.00032815500162541866, 'timing/model': 1.0818931288085878, 'learning_rate/qwen_vl_interface': 3.7349432808341673e-06, 'learning_rate/action_model': 3.7349432808341674e-05, 'epoch': 0.24}) +[2026-06-24 13:19:58,727][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.04791341722011566, 'train/grad_norm_pre_clip': 40.315589904785156, 'timing/data': 0.00035868282429873943, 'timing/model': 1.0794312418438494, 'learning_rate/qwen_vl_interface': 3.731317327700749e-06, 'learning_rate/action_model': 3.731317327700749e-05, 'epoch': 0.24}) +[2026-06-24 13:20:00,109][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.04377667233347893, 'train/grad_norm_pre_clip': 17.817935943603516, 'timing/data': 0.00032482901588082314, 'timing/model': 1.3799652680754662, 'learning_rate/qwen_vl_interface': 3.72769236002403e-06, 'learning_rate/action_model': 3.72769236002403e-05, 'epoch': 0.24}) +[2026-06-24 13:20:01,192][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.09165018796920776, 'train/grad_norm_pre_clip': 24.657318115234375, 'timing/data': 0.0003503970801830292, 'timing/model': 1.0813771721441299, 'learning_rate/qwen_vl_interface': 3.7240683801562087e-06, 'learning_rate/action_model': 3.724068380156209e-05, 'epoch': 0.24}) +[2026-06-24 13:20:02,276][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.005337894428521395, 'train/grad_norm_pre_clip': 2.3963329792022705, 'timing/data': 0.00032392702996730804, 'timing/model': 1.0825358948204666, 'learning_rate/qwen_vl_interface': 3.720445390448851e-06, 'learning_rate/action_model': 3.7204453904488504e-05, 'epoch': 0.24}) +[2026-06-24 13:20:03,362][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.04545658826828003, 'train/grad_norm_pre_clip': 30.17447853088379, 'timing/data': 0.000388238113373518, 'timing/model': 1.0839713751338422, 'learning_rate/qwen_vl_interface': 3.7168233932528713e-06, 'learning_rate/action_model': 3.716823393252871e-05, 'epoch': 0.24}) +[2026-06-24 13:20:04,442][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.16628563404083252, 'train/grad_norm_pre_clip': 54.23690414428711, 'timing/data': 0.0003192389849573374, 'timing/model': 1.0784162050113082, 'learning_rate/qwen_vl_interface': 3.713202390918548e-06, 'learning_rate/action_model': 3.713202390918547e-05, 'epoch': 0.24}) +[2026-06-24 13:20:05,525][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.053473446518182755, 'train/grad_norm_pre_clip': 29.735340118408203, 'timing/data': 0.0003300388343632221, 'timing/model': 1.0820291799027473, 'learning_rate/qwen_vl_interface': 3.7095823857955075e-06, 'learning_rate/action_model': 3.709582385795507e-05, 'epoch': 0.24}) +[2026-06-24 13:20:06,607][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.017021004110574722, 'train/grad_norm_pre_clip': 15.253251075744629, 'timing/data': 0.0004261089488863945, 'timing/model': 1.0801265710033476, 'learning_rate/qwen_vl_interface': 3.705963380232733e-06, 'learning_rate/action_model': 3.7059633802327325e-05, 'epoch': 0.24}) +[2026-06-24 13:20:07,687][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.0030730331782251596, 'train/grad_norm_pre_clip': 1.37943434715271, 'timing/data': 0.0003288569860160351, 'timing/model': 1.0783191488590091, 'learning_rate/qwen_vl_interface': 3.7023453765785578e-06, 'learning_rate/action_model': 3.7023453765785574e-05, 'epoch': 0.24}) +[2026-06-24 13:20:08,771][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.019424276426434517, 'train/grad_norm_pre_clip': 23.969192504882812, 'timing/data': 0.00032594194635748863, 'timing/model': 1.0825629078317434, 'learning_rate/qwen_vl_interface': 3.6987283771806643e-06, 'learning_rate/action_model': 3.698728377180664e-05, 'epoch': 0.24}) +[2026-06-24 13:20:09,854][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.15347638726234436, 'train/grad_norm_pre_clip': 51.44526290893555, 'timing/data': 0.00032537011429667473, 'timing/model': 1.0815959831234068, 'learning_rate/qwen_vl_interface': 3.6951123843860857e-06, 'learning_rate/action_model': 3.695112384386086e-05, 'epoch': 0.24}) +[2026-06-24 13:20:10,936][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.2950145900249481, 'train/grad_norm_pre_clip': 104.67070007324219, 'timing/data': 0.0003336251247674227, 'timing/model': 1.0801268359646201, 'learning_rate/qwen_vl_interface': 3.6914974005412007e-06, 'learning_rate/action_model': 3.691497400541201e-05, 'epoch': 0.24}) +[2026-06-24 13:20:12,019][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.21168453991413116, 'train/grad_norm_pre_clip': 56.756038665771484, 'timing/data': 0.00031911791302263737, 'timing/model': 1.081144634168595, 'learning_rate/qwen_vl_interface': 3.68788342799173e-06, 'learning_rate/action_model': 3.6878834279917294e-05, 'epoch': 0.24}) +[2026-06-24 13:20:13,099][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.13673663139343262, 'train/grad_norm_pre_clip': 115.57676696777344, 'timing/data': 0.0004074340686202049, 'timing/model': 1.0782372299581766, 'learning_rate/qwen_vl_interface': 3.684270469082744e-06, 'learning_rate/action_model': 3.684270469082744e-05, 'epoch': 0.24}) +[2026-06-24 13:20:14,180][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.02446708455681801, 'train/grad_norm_pre_clip': 28.064777374267578, 'timing/data': 0.000318978913128376, 'timing/model': 1.0792889408767223, 'learning_rate/qwen_vl_interface': 3.6806585261586513e-06, 'learning_rate/action_model': 3.6806585261586515e-05, 'epoch': 0.24}) +[2026-06-24 13:20:15,260][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.0490228496491909, 'train/grad_norm_pre_clip': 37.86965560913086, 'timing/data': 0.0003307708539068699, 'timing/model': 1.07910372200422, 'learning_rate/qwen_vl_interface': 3.6770476015632044e-06, 'learning_rate/action_model': 3.6770476015632045e-05, 'epoch': 0.24}) +[2026-06-24 13:20:16,509][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.13636401295661926, 'train/grad_norm_pre_clip': 74.75267791748047, 'timing/data': 0.0003528119996190071, 'timing/model': 1.246614814037457, 'learning_rate/qwen_vl_interface': 3.673437697639491e-06, 'learning_rate/action_model': 3.673437697639491e-05, 'epoch': 0.24}) +[2026-06-24 13:20:17,588][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.1387917399406433, 'train/grad_norm_pre_clip': 69.115234375, 'timing/data': 0.0004491820000112057, 'timing/model': 1.0776501938235015, 'learning_rate/qwen_vl_interface': 3.6698288167299378e-06, 'learning_rate/action_model': 3.6698288167299376e-05, 'epoch': 0.24}) +[2026-06-24 13:20:18,667][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.13061603903770447, 'train/grad_norm_pre_clip': 33.43883514404297, 'timing/data': 0.00032859714701771736, 'timing/model': 1.077782240929082, 'learning_rate/qwen_vl_interface': 3.6662209611763115e-06, 'learning_rate/action_model': 3.6662209611763116e-05, 'epoch': 0.24}) +[2026-06-24 13:20:19,749][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.08270220458507538, 'train/grad_norm_pre_clip': 77.9585952758789, 'timing/data': 0.0003144899383187294, 'timing/model': 1.080012482125312, 'learning_rate/qwen_vl_interface': 3.6626141333197065e-06, 'learning_rate/action_model': 3.662614133319706e-05, 'epoch': 0.24}) +[2026-06-24 13:20:20,827][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.2029179334640503, 'train/grad_norm_pre_clip': 64.6353530883789, 'timing/data': 0.00034201215021312237, 'timing/model': 1.0764480349607766, 'learning_rate/qwen_vl_interface': 3.6590083355005587e-06, 'learning_rate/action_model': 3.659008335500558e-05, 'epoch': 0.24}) +[2026-06-24 13:20:21,905][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.16815459728240967, 'train/grad_norm_pre_clip': 40.46461868286133, 'timing/data': 0.0003148810938000679, 'timing/model': 1.0767414849251509, 'learning_rate/qwen_vl_interface': 3.655403570058629e-06, 'learning_rate/action_model': 3.655403570058629e-05, 'epoch': 0.24}) +[2026-06-24 13:20:22,985][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.0033013797365128994, 'train/grad_norm_pre_clip': 3.8545539379119873, 'timing/data': 0.00032896711491048336, 'timing/model': 1.0778809490147978, 'learning_rate/qwen_vl_interface': 3.6517998393330066e-06, 'learning_rate/action_model': 3.651799839333007e-05, 'epoch': 0.24}) +[2026-06-24 13:20:24,064][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.08007478713989258, 'train/grad_norm_pre_clip': 46.15117263793945, 'timing/data': 0.0003540839534252882, 'timing/model': 1.0771197008434683, 'learning_rate/qwen_vl_interface': 3.6481971456621216e-06, 'learning_rate/action_model': 3.648197145662122e-05, 'epoch': 0.24}) +[2026-06-24 13:20:25,148][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.003863009624183178, 'train/grad_norm_pre_clip': 5.502530097961426, 'timing/data': 0.0003014348912984133, 'timing/model': 1.0825926850084215, 'learning_rate/qwen_vl_interface': 3.644595491383715e-06, 'learning_rate/action_model': 3.644595491383715e-05, 'epoch': 0.24}) +[2026-06-24 13:20:26,229][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.21843937039375305, 'train/grad_norm_pre_clip': 93.84754943847656, 'timing/data': 0.0003384349402040243, 'timing/model': 1.0800399160943925, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 0.24}) +[2026-06-24 13:20:27,308][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.007414687424898148, 'train/grad_norm_pre_clip': 15.160449981689453, 'timing/data': 0.00032965908758342266, 'timing/model': 1.0777221058961004, 'learning_rate/qwen_vl_interface': 3.637395310351971e-06, 'learning_rate/action_model': 3.637395310351971e-05, 'epoch': 0.24}) +[2026-06-24 13:20:28,389][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.0942341759800911, 'train/grad_norm_pre_clip': 44.106048583984375, 'timing/data': 0.00031261611729860306, 'timing/model': 1.078425724990666, 'learning_rate/qwen_vl_interface': 3.6337967882707516e-06, 'learning_rate/action_model': 3.6337967882707515e-05, 'epoch': 0.24}) +[2026-06-24 13:20:29,468][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.25578004121780396, 'train/grad_norm_pre_clip': 55.994598388671875, 'timing/data': 0.00033551896922290325, 'timing/model': 1.0774505941662937, 'learning_rate/qwen_vl_interface': 3.63019931492625e-06, 'learning_rate/action_model': 3.63019931492625e-05, 'epoch': 0.24}) +[2026-06-24 13:20:30,545][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.2207699418067932, 'train/grad_norm_pre_clip': 53.1533088684082, 'timing/data': 0.0003300989046692848, 'timing/model': 1.0755331260152161, 'learning_rate/qwen_vl_interface': 3.6266028926528257e-06, 'learning_rate/action_model': 3.6266028926528254e-05, 'epoch': 0.24}) +[2026-06-24 13:20:31,624][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.048628244549036026, 'train/grad_norm_pre_clip': 51.96728515625, 'timing/data': 0.00031035207211971283, 'timing/model': 1.0773109791334718, 'learning_rate/qwen_vl_interface': 3.6230075237841604e-06, 'learning_rate/action_model': 3.62300752378416e-05, 'epoch': 0.24}) +[2026-06-24 13:20:32,912][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.3363804817199707, 'train/grad_norm_pre_clip': 70.76527404785156, 'timing/data': 0.0003390158526599407, 'timing/model': 1.2866088938899338, 'learning_rate/qwen_vl_interface': 3.619413210653248e-06, 'learning_rate/action_model': 3.619413210653248e-05, 'epoch': 0.24}) +[2026-06-24 13:20:33,990][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.11120559275150299, 'train/grad_norm_pre_clip': 106.47002410888672, 'timing/data': 0.0003468410577625036, 'timing/model': 1.0769173591397703, 'learning_rate/qwen_vl_interface': 3.6158199555924007e-06, 'learning_rate/action_model': 3.6158199555924006e-05, 'epoch': 0.24}) +[2026-06-24 13:20:35,066][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.04492519050836563, 'train/grad_norm_pre_clip': 36.675010681152344, 'timing/data': 0.00033966684713959694, 'timing/model': 1.0746845290996134, 'learning_rate/qwen_vl_interface': 3.6122277609332433e-06, 'learning_rate/action_model': 3.612227760933243e-05, 'epoch': 0.24}) +[2026-06-24 13:20:36,148][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.015027076005935669, 'train/grad_norm_pre_clip': 16.422170639038086, 'timing/data': 0.00031307688914239407, 'timing/model': 1.0804879639763385, 'learning_rate/qwen_vl_interface': 3.6086366290067088e-06, 'learning_rate/action_model': 3.6086366290067084e-05, 'epoch': 0.24}) +[2026-06-24 13:20:37,226][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.05310462787747383, 'train/grad_norm_pre_clip': 74.22309875488281, 'timing/data': 0.0003119248431175947, 'timing/model': 1.0761732880491763, 'learning_rate/qwen_vl_interface': 3.605046562143047e-06, 'learning_rate/action_model': 3.6050465621430466e-05, 'epoch': 0.24}) +[2026-06-24 13:20:38,307][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.2567813992500305, 'train/grad_norm_pre_clip': 60.70212936401367, 'timing/data': 0.000318978913128376, 'timing/model': 1.0789383728988469, 'learning_rate/qwen_vl_interface': 3.6014575626718112e-06, 'learning_rate/action_model': 3.601457562671811e-05, 'epoch': 0.24}) +[2026-06-24 13:20:39,392][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.03705913573503494, 'train/grad_norm_pre_clip': 62.0620231628418, 'timing/data': 0.0005618040449917316, 'timing/model': 1.0826135058887303, 'learning_rate/qwen_vl_interface': 3.597869632921866e-06, 'learning_rate/action_model': 3.597869632921866e-05, 'epoch': 0.24}) +[2026-06-24 13:20:40,472][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.10727795958518982, 'train/grad_norm_pre_clip': 54.30018615722656, 'timing/data': 0.00033588986843824387, 'timing/model': 1.0787258110940456, 'learning_rate/qwen_vl_interface': 3.5942827752213788e-06, 'learning_rate/action_model': 3.594282775221379e-05, 'epoch': 0.24}) +[2026-06-24 13:20:41,555][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.018201077356934547, 'train/grad_norm_pre_clip': 27.985767364501953, 'timing/data': 0.00035932380706071854, 'timing/model': 1.079823498846963, 'learning_rate/qwen_vl_interface': 3.590696991897822e-06, 'learning_rate/action_model': 3.590696991897822e-05, 'epoch': 0.24}) +[2026-06-24 13:20:42,640][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.014258584938943386, 'train/grad_norm_pre_clip': 19.952104568481445, 'timing/data': 0.0003732810728251934, 'timing/model': 1.0839446869213134, 'learning_rate/qwen_vl_interface': 3.5871122852779733e-06, 'learning_rate/action_model': 3.587112285277973e-05, 'epoch': 0.24}) +[2026-06-24 13:20:43,720][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.11176181584596634, 'train/grad_norm_pre_clip': 71.83562469482422, 'timing/data': 0.00032026111148297787, 'timing/model': 1.0781617180909961, 'learning_rate/qwen_vl_interface': 3.583528657687907e-06, 'learning_rate/action_model': 3.583528657687907e-05, 'epoch': 0.24}) +[2026-06-24 13:20:44,803][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.1895330548286438, 'train/grad_norm_pre_clip': 71.29349517822266, 'timing/data': 0.0003233770839869976, 'timing/model': 1.081011130940169, 'learning_rate/qwen_vl_interface': 3.5799461114530054e-06, 'learning_rate/action_model': 3.579946111453005e-05, 'epoch': 0.24}) +[2026-06-24 13:20:45,885][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.17190325260162354, 'train/grad_norm_pre_clip': 88.463134765625, 'timing/data': 0.0003246490377932787, 'timing/model': 1.0810724841430783, 'learning_rate/qwen_vl_interface': 3.576364648897939e-06, 'learning_rate/action_model': 3.576364648897939e-05, 'epoch': 0.24}) +[2026-06-24 13:20:46,968][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.10963041335344315, 'train/grad_norm_pre_clip': 82.66221618652344, 'timing/data': 0.000313909025862813, 'timing/model': 1.0806689960882068, 'learning_rate/qwen_vl_interface': 3.57278427234668e-06, 'learning_rate/action_model': 3.57278427234668e-05, 'epoch': 0.24}) +[2026-06-24 13:20:48,047][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.06315809488296509, 'train/grad_norm_pre_clip': 24.67235565185547, 'timing/data': 0.00038078497163951397, 'timing/model': 1.0781409610062838, 'learning_rate/qwen_vl_interface': 3.569204984122501e-06, 'learning_rate/action_model': 3.569204984122501e-05, 'epoch': 0.24}) +[2026-06-24 13:20:49,311][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2500 +[2026-06-24 13:21:13,656][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2500 +[2026-06-24 13:21:13,657][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2500 +[2026-06-24 13:21:46,568][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2500 +[2026-06-24 13:30:56,213][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-06-24 13:30:56,214][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.1130584254860878, 'train/grad_norm_pre_clip': 70.75817108154297, 'eval/action_loss/seconds': 24.343271889025345, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.06988238543272018, 'eval/latency_2/loss': 0.06988238543272018, 'eval/flappy/loss': 0.06988238543272018, 'eval/flappy/latency_2/loss': 0.06988238543272018, 'eval/flap_precision': 0.9141414141414141, 'eval/flap_recall': 0.9823609226594301, 'eval/flap_f1': 0.94702419882276, 'eval/flap_rate_model': 0.22, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.94702419882276, 'eval/latency_2/flap_precision': 0.9141414141414141, 'eval/latency_2/flap_recall': 0.9823609226594301, 'eval/latency_2/flap_f1': 0.94702419882276, 'eval/latency_2/flap_rate_model': 0.22, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.94702419882276, 'eval/action_classification/data_wait_seconds': 0.3342509865760803, 'eval/action_classification/predict_seconds': 32.4783821105957, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 419.8600000000212, 'rl_games_eval/mid_train/mean_length': 3401.6, 'rl_games_eval/mid_train/std_reward': 49.48000000000879, 'rl_games_eval/mid_train/std_length': 396.8, 'rl_games_eval/mid_train/macro_mean_reward': 419.8600000000212, 'rl_games_eval/mid_train/macro_mean_length': 3401.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 419.8600000000212, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3401.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 49.48000000000879, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 396.8, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 419.8600000000212, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005739270709455013, 'timing/model': 1.2619226449169219, 'learning_rate/qwen_vl_interface': 3.5656267865479555e-06, 'learning_rate/action_model': 3.565626786547955e-05, 'epoch': 0.24}) +[2026-06-24 13:30:56,214][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_2500_state +[2026-06-24 13:30:56,217][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 13:30:58,801][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_2500_state/model.safetensors +[2026-06-24 13:31:04,509][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_2500_state/optimizer.bin +[2026-06-24 13:31:04,510][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_2500_state/sampler.bin +[2026-06-24 13:31:04,510][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-24 13:31:04,511][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-24 13:31:04,516][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-24 13:31:04,516][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 13:31:04,526][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 13:31:05,653][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.052071601152420044, 'train/grad_norm_pre_clip': 46.50590133666992, 'timing/data': 0.00039600394666194916, 'timing/model': 1.1255369619466364, 'learning_rate/qwen_vl_interface': 3.5620496819449034e-06, 'learning_rate/action_model': 3.562049681944903e-05, 'epoch': 0.24}) +[2026-06-24 13:31:06,725][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.050660256296396255, 'train/grad_norm_pre_clip': 46.80464172363281, 'timing/data': 0.000404509948566556, 'timing/model': 1.070554964011535, 'learning_rate/qwen_vl_interface': 3.5584736726344867e-06, 'learning_rate/action_model': 3.5584736726344864e-05, 'epoch': 0.24}) +[2026-06-24 13:31:07,792][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.3268781900405884, 'train/grad_norm_pre_clip': 78.17792510986328, 'timing/data': 0.00037607597187161446, 'timing/model': 1.0644655181095004, 'learning_rate/qwen_vl_interface': 3.5548987609371393e-06, 'learning_rate/action_model': 3.554898760937139e-05, 'epoch': 0.24}) +[2026-06-24 13:31:08,859][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.2992936968803406, 'train/grad_norm_pre_clip': 69.45748138427734, 'timing/data': 0.0003218050114810467, 'timing/model': 1.065824502846226, 'learning_rate/qwen_vl_interface': 3.5513249491725825e-06, 'learning_rate/action_model': 3.5513249491725826e-05, 'epoch': 0.24}) +[2026-06-24 13:31:09,927][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.011745843105018139, 'train/grad_norm_pre_clip': 16.681020736694336, 'timing/data': 0.00038137612864375114, 'timing/model': 1.0661458389367908, 'learning_rate/qwen_vl_interface': 3.5477522396598236e-06, 'learning_rate/action_model': 3.5477522396598235e-05, 'epoch': 0.24}) +[2026-06-24 13:31:10,993][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.15314605832099915, 'train/grad_norm_pre_clip': 43.803375244140625, 'timing/data': 0.00034554884769022465, 'timing/model': 1.0647149730939418, 'learning_rate/qwen_vl_interface': 3.544180634717156e-06, 'learning_rate/action_model': 3.544180634717156e-05, 'epoch': 0.24}) +[2026-06-24 13:31:12,062][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.11864683777093887, 'train/grad_norm_pre_clip': 40.87312316894531, 'timing/data': 0.0003235279582440853, 'timing/model': 1.0675853379070759, 'learning_rate/qwen_vl_interface': 3.5406101366621536e-06, 'learning_rate/action_model': 3.540610136662154e-05, 'epoch': 0.24}) +[2026-06-24 13:31:13,126][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.08101722598075867, 'train/grad_norm_pre_clip': 55.388824462890625, 'timing/data': 0.0003511591348797083, 'timing/model': 1.0622243890538812, 'learning_rate/qwen_vl_interface': 3.5370407478116764e-06, 'learning_rate/action_model': 3.537040747811676e-05, 'epoch': 0.24}) +[2026-06-24 13:31:14,194][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.030226759612560272, 'train/grad_norm_pre_clip': 22.80232048034668, 'timing/data': 0.00034806295298039913, 'timing/model': 1.0655984599143267, 'learning_rate/qwen_vl_interface': 3.53347247048186e-06, 'learning_rate/action_model': 3.53347247048186e-05, 'epoch': 0.24}) +[2026-06-24 13:31:15,259][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.2539885640144348, 'train/grad_norm_pre_clip': 58.86280059814453, 'timing/data': 0.0003512399271130562, 'timing/model': 1.0639456580393016, 'learning_rate/qwen_vl_interface': 3.5299053069881214e-06, 'learning_rate/action_model': 3.529905306988121e-05, 'epoch': 0.24}) +[2026-06-24 13:31:16,324][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.018145611509680748, 'train/grad_norm_pre_clip': 8.201210975646973, 'timing/data': 0.0003426929470151663, 'timing/model': 1.0632432899437845, 'learning_rate/qwen_vl_interface': 3.526339259645154e-06, 'learning_rate/action_model': 3.526339259645154e-05, 'epoch': 0.24}) +[2026-06-24 13:31:17,694][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.03900924324989319, 'train/grad_norm_pre_clip': 25.492639541625977, 'timing/data': 0.00030912994407117367, 'timing/model': 1.367668794002384, 'learning_rate/qwen_vl_interface': 3.5227743307669267e-06, 'learning_rate/action_model': 3.522774330766927e-05, 'epoch': 0.24}) +[2026-06-24 13:31:18,764][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.0031254468485713005, 'train/grad_norm_pre_clip': 1.4703229665756226, 'timing/data': 0.00031381985172629356, 'timing/model': 1.0684699239209294, 'learning_rate/qwen_vl_interface': 3.5192105226666856e-06, 'learning_rate/action_model': 3.5192105226666855e-05, 'epoch': 0.24}) +[2026-06-24 13:31:19,836][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.1950283795595169, 'train/grad_norm_pre_clip': 50.23048400878906, 'timing/data': 0.00037211901508271694, 'timing/model': 1.0701318180654198, 'learning_rate/qwen_vl_interface': 3.5156478376569453e-06, 'learning_rate/action_model': 3.515647837656945e-05, 'epoch': 0.24}) +[2026-06-24 13:31:20,904][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.07609334588050842, 'train/grad_norm_pre_clip': 26.6645565032959, 'timing/data': 0.0003811861388385296, 'timing/model': 1.066233967896551, 'learning_rate/qwen_vl_interface': 3.512086278049492e-06, 'learning_rate/action_model': 3.512086278049492e-05, 'epoch': 0.24}) +[2026-06-24 13:31:21,981][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.047778282314538956, 'train/grad_norm_pre_clip': 26.344274520874023, 'timing/data': 0.00033984892070293427, 'timing/model': 1.0754512250423431, 'learning_rate/qwen_vl_interface': 3.508525846155387e-06, 'learning_rate/action_model': 3.508525846155386e-05, 'epoch': 0.24}) +[2026-06-24 13:31:23,053][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.15815643966197968, 'train/grad_norm_pre_clip': 80.61946868896484, 'timing/data': 0.00034488807432353497, 'timing/model': 1.070338121848181, 'learning_rate/qwen_vl_interface': 3.50496654428495e-06, 'learning_rate/action_model': 3.50496654428495e-05, 'epoch': 0.24}) +[2026-06-24 13:31:24,123][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.08297357708215714, 'train/grad_norm_pre_clip': 43.333343505859375, 'timing/data': 0.0003388458862900734, 'timing/model': 1.0691490108147264, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 0.24}) +[2026-06-24 13:31:25,192][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.21711322665214539, 'train/grad_norm_pre_clip': 61.707393646240234, 'timing/data': 0.0003186590038239956, 'timing/model': 1.0667758858762681, 'learning_rate/qwen_vl_interface': 3.4978513398527316e-06, 'learning_rate/action_model': 3.4978513398527314e-05, 'epoch': 0.24}) +[2026-06-24 13:31:26,262][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.09862100332975388, 'train/grad_norm_pre_clip': 26.846403121948242, 'timing/data': 0.00040980009362101555, 'timing/model': 1.068486169911921, 'learning_rate/qwen_vl_interface': 3.4942954419079228e-06, 'learning_rate/action_model': 3.494295441907923e-05, 'epoch': 0.24}) +[2026-06-24 13:31:27,332][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.01757744699716568, 'train/grad_norm_pre_clip': 17.636873245239258, 'timing/data': 0.0003282460384070873, 'timing/model': 1.0681667330209166, 'learning_rate/qwen_vl_interface': 3.490740683220743e-06, 'learning_rate/action_model': 3.490740683220743e-05, 'epoch': 0.24}) +[2026-06-24 13:31:28,405][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.008372155949473381, 'train/grad_norm_pre_clip': 4.4036641120910645, 'timing/data': 0.0003849428612738848, 'timing/model': 1.0718343439511955, 'learning_rate/qwen_vl_interface': 3.4871870660978324e-06, 'learning_rate/action_model': 3.4871870660978324e-05, 'epoch': 0.24}) +[2026-06-24 13:31:29,478][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.05576974153518677, 'train/grad_norm_pre_clip': 9.007993698120117, 'timing/data': 0.00034695188514888287, 'timing/model': 1.07112385192886, 'learning_rate/qwen_vl_interface': 3.4836345928450976e-06, 'learning_rate/action_model': 3.4836345928450974e-05, 'epoch': 0.24}) +[2026-06-24 13:31:30,551][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.20583821833133698, 'train/grad_norm_pre_clip': 46.750518798828125, 'timing/data': 0.000357871875166893, 'timing/model': 1.070886277826503, 'learning_rate/qwen_vl_interface': 3.4800832657676983e-06, 'learning_rate/action_model': 3.480083265767698e-05, 'epoch': 0.24}) +[2026-06-24 13:31:31,624][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.017825637012720108, 'train/grad_norm_pre_clip': 7.508296012878418, 'timing/data': 0.00031638401560485363, 'timing/model': 1.071641619084403, 'learning_rate/qwen_vl_interface': 3.4765330871700503e-06, 'learning_rate/action_model': 3.4765330871700504e-05, 'epoch': 0.24}) +[2026-06-24 13:31:32,694][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.09755223989486694, 'train/grad_norm_pre_clip': 41.20963668823242, 'timing/data': 0.000339368125423789, 'timing/model': 1.0686007291078568, 'learning_rate/qwen_vl_interface': 3.47298405935583e-06, 'learning_rate/action_model': 3.47298405935583e-05, 'epoch': 0.24}) +[2026-06-24 13:31:33,964][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.031191986054182053, 'train/grad_norm_pre_clip': 15.408759117126465, 'timing/data': 0.0003777791280299425, 'timing/model': 1.2685203100554645, 'learning_rate/qwen_vl_interface': 3.469436184627959e-06, 'learning_rate/action_model': 3.469436184627959e-05, 'epoch': 0.24}) +[2026-06-24 13:31:35,036][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.05540971830487251, 'train/grad_norm_pre_clip': 77.0875473022461, 'timing/data': 0.00037122704088687897, 'timing/model': 1.0699039718601853, 'learning_rate/qwen_vl_interface': 3.465889465288617e-06, 'learning_rate/action_model': 3.465889465288617e-05, 'epoch': 0.24}) +[2026-06-24 13:31:36,108][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.24560993909835815, 'train/grad_norm_pre_clip': 52.5476188659668, 'timing/data': 0.0003174061421304941, 'timing/model': 1.0707116620615125, 'learning_rate/qwen_vl_interface': 3.4623439036392308e-06, 'learning_rate/action_model': 3.462343903639231e-05, 'epoch': 0.24}) +[2026-06-24 13:31:37,182][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.11981376260519028, 'train/grad_norm_pre_clip': 45.16261672973633, 'timing/data': 0.0003212739247828722, 'timing/model': 1.0722338040359318, 'learning_rate/qwen_vl_interface': 3.4587995019804783e-06, 'learning_rate/action_model': 3.4587995019804783e-05, 'epoch': 0.24}) +[2026-06-24 13:31:38,260][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.02482953853905201, 'train/grad_norm_pre_clip': 33.211219787597656, 'timing/data': 0.0003544760402292013, 'timing/model': 1.0760798710398376, 'learning_rate/qwen_vl_interface': 3.4552562626122817e-06, 'learning_rate/action_model': 3.455256262612281e-05, 'epoch': 0.24}) +[2026-06-24 13:31:39,340][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.08517760038375854, 'train/grad_norm_pre_clip': 41.720760345458984, 'timing/data': 0.00040722498670220375, 'timing/model': 1.0783795530442148, 'learning_rate/qwen_vl_interface': 3.4517141878338106e-06, 'learning_rate/action_model': 3.451714187833811e-05, 'epoch': 0.24}) +[2026-06-24 13:31:40,419][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.22232839465141296, 'train/grad_norm_pre_clip': 86.21914672851562, 'timing/data': 0.0003202820662409067, 'timing/model': 1.0777957290410995, 'learning_rate/qwen_vl_interface': 3.4481732799434804e-06, 'learning_rate/action_model': 3.4481732799434805e-05, 'epoch': 0.25}) +[2026-06-24 13:31:41,494][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.03154399245977402, 'train/grad_norm_pre_clip': 12.293424606323242, 'timing/data': 0.00035918503999710083, 'timing/model': 1.0734317530877888, 'learning_rate/qwen_vl_interface': 3.4446335412389442e-06, 'learning_rate/action_model': 3.444633541238944e-05, 'epoch': 0.25}) +[2026-06-24 13:31:42,567][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.027003858238458633, 'train/grad_norm_pre_clip': 13.514483451843262, 'timing/data': 0.00045038596726953983, 'timing/model': 1.0710132119711488, 'learning_rate/qwen_vl_interface': 3.4410949740171056e-06, 'learning_rate/action_model': 3.4410949740171054e-05, 'epoch': 0.25}) +[2026-06-24 13:31:43,645][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.07191338390111923, 'train/grad_norm_pre_clip': 25.92286491394043, 'timing/data': 0.0003305310383439064, 'timing/model': 1.0756773960310966, 'learning_rate/qwen_vl_interface': 3.4375575805740984e-06, 'learning_rate/action_model': 3.437557580574099e-05, 'epoch': 0.25}) +[2026-06-24 13:31:44,718][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.008584327064454556, 'train/grad_norm_pre_clip': 5.33664083480835, 'timing/data': 0.0003382451832294464, 'timing/model': 1.07199302688241, 'learning_rate/qwen_vl_interface': 3.434021363205298e-06, 'learning_rate/action_model': 3.434021363205298e-05, 'epoch': 0.25}) +[2026-06-24 13:31:45,795][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.06494437158107758, 'train/grad_norm_pre_clip': 41.488807678222656, 'timing/data': 0.0003166450187563896, 'timing/model': 1.0748750229831785, 'learning_rate/qwen_vl_interface': 3.430486324205322e-06, 'learning_rate/action_model': 3.430486324205322e-05, 'epoch': 0.25}) +[2026-06-24 13:31:46,871][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.10397237539291382, 'train/grad_norm_pre_clip': 41.953819274902344, 'timing/data': 0.00033626193180680275, 'timing/model': 1.074511141050607, 'learning_rate/qwen_vl_interface': 3.4269524658680138e-06, 'learning_rate/action_model': 3.4269524658680135e-05, 'epoch': 0.25}) +[2026-06-24 13:31:47,947][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.018037017434835434, 'train/grad_norm_pre_clip': 8.146306037902832, 'timing/data': 0.00034374604001641273, 'timing/model': 1.0746472298633307, 'learning_rate/qwen_vl_interface': 3.423419790486459e-06, 'learning_rate/action_model': 3.423419790486459e-05, 'epoch': 0.25}) +[2026-06-24 13:31:49,023][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.18779665231704712, 'train/grad_norm_pre_clip': 50.6716423034668, 'timing/data': 0.0003170149866491556, 'timing/model': 1.0740586868487298, 'learning_rate/qwen_vl_interface': 3.4198883003529726e-06, 'learning_rate/action_model': 3.4198883003529724e-05, 'epoch': 0.25}) +[2026-06-24 13:31:50,101][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.00036253852886147797, 'train/grad_norm_pre_clip': 0.16993401944637299, 'timing/data': 0.0003124871291220188, 'timing/model': 1.076473694993183, 'learning_rate/qwen_vl_interface': 3.4163579977590957e-06, 'learning_rate/action_model': 3.4163579977590955e-05, 'epoch': 0.25}) +[2026-06-24 13:31:51,368][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.08216817677021027, 'train/grad_norm_pre_clip': 37.18647766113281, 'timing/data': 0.00031366990879178047, 'timing/model': 1.2653446570038795, 'learning_rate/qwen_vl_interface': 3.412828884995608e-06, 'learning_rate/action_model': 3.412828884995608e-05, 'epoch': 0.25}) +[2026-06-24 13:31:52,445][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.01674826070666313, 'train/grad_norm_pre_clip': 11.288999557495117, 'timing/data': 0.00032698409631848335, 'timing/model': 1.0756558130960912, 'learning_rate/qwen_vl_interface': 3.409300964352511e-06, 'learning_rate/action_model': 3.409300964352511e-05, 'epoch': 0.25}) +[2026-06-24 13:31:53,523][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.001972086727619171, 'train/grad_norm_pre_clip': 0.8681392073631287, 'timing/data': 0.00032544112764298916, 'timing/model': 1.0760906729847193, 'learning_rate/qwen_vl_interface': 3.405774238119034e-06, 'learning_rate/action_model': 3.405774238119034e-05, 'epoch': 0.25}) +[2026-06-24 13:31:54,602][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.028763841837644577, 'train/grad_norm_pre_clip': 37.33198547363281, 'timing/data': 0.00035223178565502167, 'timing/model': 1.0775560378096998, 'learning_rate/qwen_vl_interface': 3.4022487085836327e-06, 'learning_rate/action_model': 3.4022487085836326e-05, 'epoch': 0.25}) +[2026-06-24 13:31:55,678][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.05239592492580414, 'train/grad_norm_pre_clip': 64.56289672851562, 'timing/data': 0.00032632285729050636, 'timing/model': 1.0749940869864076, 'learning_rate/qwen_vl_interface': 3.398724378033981e-06, 'learning_rate/action_model': 3.398724378033981e-05, 'epoch': 0.25}) +[2026-06-24 13:31:56,756][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.0010622014524415135, 'train/grad_norm_pre_clip': 0.5301332473754883, 'timing/data': 0.0003106039948761463, 'timing/model': 1.0755691800732166, 'learning_rate/qwen_vl_interface': 3.3952012487569836e-06, 'learning_rate/action_model': 3.395201248756983e-05, 'epoch': 0.25}) +[2026-06-24 13:31:57,833][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.3078955411911011, 'train/grad_norm_pre_clip': 127.2433090209961, 'timing/data': 0.00036946404725313187, 'timing/model': 1.0759400080423802, 'learning_rate/qwen_vl_interface': 3.3916793230387574e-06, 'learning_rate/action_model': 3.3916793230387576e-05, 'epoch': 0.25}) +[2026-06-24 13:31:58,912][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.016313092783093452, 'train/grad_norm_pre_clip': 21.099559783935547, 'timing/data': 0.0003918060101568699, 'timing/model': 1.0766610950231552, 'learning_rate/qwen_vl_interface': 3.388158603164644e-06, 'learning_rate/action_model': 3.388158603164644e-05, 'epoch': 0.25}) +[2026-06-24 13:31:59,991][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.07399757951498032, 'train/grad_norm_pre_clip': 48.5131950378418, 'timing/data': 0.0003468319773674011, 'timing/model': 1.0773327490314841, 'learning_rate/qwen_vl_interface': 3.384639091419198e-06, 'learning_rate/action_model': 3.3846390914191975e-05, 'epoch': 0.25}) +[2026-06-24 13:32:01,072][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.002403972204774618, 'train/grad_norm_pre_clip': 2.060967445373535, 'timing/data': 0.000316153047606349, 'timing/model': 1.0800729191396385, 'learning_rate/qwen_vl_interface': 3.381120790086195e-06, 'learning_rate/action_model': 3.3811207900861943e-05, 'epoch': 0.25}) +[2026-06-24 13:32:02,148][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.06601424515247345, 'train/grad_norm_pre_clip': 18.655704498291016, 'timing/data': 0.0003095320425927639, 'timing/model': 1.0741195268929005, 'learning_rate/qwen_vl_interface': 3.3776037014486207e-06, 'learning_rate/action_model': 3.3776037014486206e-05, 'epoch': 0.25}) +[2026-06-24 13:32:03,227][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.2889091372489929, 'train/grad_norm_pre_clip': 92.83342742919922, 'timing/data': 0.000311525072902441, 'timing/model': 1.0769909829832613, 'learning_rate/qwen_vl_interface': 3.3740878277886758e-06, 'learning_rate/action_model': 3.374087827788676e-05, 'epoch': 0.25}) +[2026-06-24 13:32:04,307][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.022791961207985878, 'train/grad_norm_pre_clip': 30.928447723388672, 'timing/data': 0.00031723594292998314, 'timing/model': 1.078791217179969, 'learning_rate/qwen_vl_interface': 3.3705731713877735e-06, 'learning_rate/action_model': 3.370573171387774e-05, 'epoch': 0.25}) +[2026-06-24 13:32:05,383][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.11836300790309906, 'train/grad_norm_pre_clip': 45.016883850097656, 'timing/data': 0.0004084280226379633, 'timing/model': 1.074259506072849, 'learning_rate/qwen_vl_interface': 3.3670597345265336e-06, 'learning_rate/action_model': 3.3670597345265335e-05, 'epoch': 0.25}) +[2026-06-24 13:32:06,468][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.07284900546073914, 'train/grad_norm_pre_clip': 98.75714111328125, 'timing/data': 0.0003400391433387995, 'timing/model': 1.0833632908761501, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 0.25}) +[2026-06-24 13:32:07,545][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.07216498255729675, 'train/grad_norm_pre_clip': 44.35426712036133, 'timing/data': 0.0003107539378106594, 'timing/model': 1.07493175403215, 'learning_rate/qwen_vl_interface': 3.3600365285415844e-06, 'learning_rate/action_model': 3.3600365285415844e-05, 'epoch': 0.25}) +[2026-06-24 13:32:08,882][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.019085939973592758, 'train/grad_norm_pre_clip': 40.89731216430664, 'timing/data': 0.00033729407005012035, 'timing/model': 1.335432352963835, 'learning_rate/qwen_vl_interface': 3.356526763975152e-06, 'learning_rate/action_model': 3.356526763975152e-05, 'epoch': 0.25}) +[2026-06-24 13:32:09,961][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.3275134563446045, 'train/grad_norm_pre_clip': 69.918701171875, 'timing/data': 0.000418154988437891, 'timing/model': 1.0777527880854905, 'learning_rate/qwen_vl_interface': 3.353018228062947e-06, 'learning_rate/action_model': 3.3530182280629466e-05, 'epoch': 0.25}) +[2026-06-24 13:32:11,040][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.02385600283741951, 'train/grad_norm_pre_clip': 20.820556640625, 'timing/data': 0.00031210598535835743, 'timing/model': 1.0775975580327213, 'learning_rate/qwen_vl_interface': 3.3495109230816165e-06, 'learning_rate/action_model': 3.349510923081616e-05, 'epoch': 0.25}) +[2026-06-24 13:32:12,119][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.13182441890239716, 'train/grad_norm_pre_clip': 59.43503189086914, 'timing/data': 0.00031383917666971684, 'timing/model': 1.0772967559751123, 'learning_rate/qwen_vl_interface': 3.346004851307017e-06, 'learning_rate/action_model': 3.346004851307017e-05, 'epoch': 0.25}) +[2026-06-24 13:32:13,200][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.19332872331142426, 'train/grad_norm_pre_clip': 196.5980224609375, 'timing/data': 0.00037611695006489754, 'timing/model': 1.0789771738927811, 'learning_rate/qwen_vl_interface': 3.342500015014198e-06, 'learning_rate/action_model': 3.3425000150141976e-05, 'epoch': 0.25}) +[2026-06-24 13:32:14,278][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.004537344444543123, 'train/grad_norm_pre_clip': 7.922820568084717, 'timing/data': 0.0003268229775130749, 'timing/model': 1.0766456658020616, 'learning_rate/qwen_vl_interface': 3.3389964164774075e-06, 'learning_rate/action_model': 3.338996416477408e-05, 'epoch': 0.25}) +[2026-06-24 13:32:15,359][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.1760111004114151, 'train/grad_norm_pre_clip': 205.01193237304688, 'timing/data': 0.000338756013661623, 'timing/model': 1.0786631091032177, 'learning_rate/qwen_vl_interface': 3.335494057970097e-06, 'learning_rate/action_model': 3.335494057970097e-05, 'epoch': 0.25}) +[2026-06-24 13:32:16,436][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.11346273869276047, 'train/grad_norm_pre_clip': 63.52339172363281, 'timing/data': 0.0003317231312394142, 'timing/model': 1.075629360973835, 'learning_rate/qwen_vl_interface': 3.331992941764906e-06, 'learning_rate/action_model': 3.331992941764906e-05, 'epoch': 0.25}) +[2026-06-24 13:32:17,513][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.015536166727542877, 'train/grad_norm_pre_clip': 42.67384338378906, 'timing/data': 0.00038840994238853455, 'timing/model': 1.075596041046083, 'learning_rate/qwen_vl_interface': 3.328493070133672e-06, 'learning_rate/action_model': 3.328493070133672e-05, 'epoch': 0.25}) +[2026-06-24 13:32:18,593][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.007686661556363106, 'train/grad_norm_pre_clip': 5.67399787902832, 'timing/data': 0.00032303715124726295, 'timing/model': 1.0780767330434173, 'learning_rate/qwen_vl_interface': 3.324994445347424e-06, 'learning_rate/action_model': 3.324994445347424e-05, 'epoch': 0.25}) +[2026-06-24 13:32:19,674][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.03700840473175049, 'train/grad_norm_pre_clip': 55.83565139770508, 'timing/data': 0.00032590189948678017, 'timing/model': 1.0792939129751176, 'learning_rate/qwen_vl_interface': 3.3214970696763793e-06, 'learning_rate/action_model': 3.321497069676379e-05, 'epoch': 0.25}) +[2026-06-24 13:32:20,754][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.005192394368350506, 'train/grad_norm_pre_clip': 5.918140411376953, 'timing/data': 0.0003446480259299278, 'timing/model': 1.0787018819246441, 'learning_rate/qwen_vl_interface': 3.3180009453899496e-06, 'learning_rate/action_model': 3.318000945389949e-05, 'epoch': 0.25}) +[2026-06-24 13:32:21,832][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.11313130706548691, 'train/grad_norm_pre_clip': 122.86767578125, 'timing/data': 0.000315733952447772, 'timing/model': 1.0766359630506486, 'learning_rate/qwen_vl_interface': 3.3145060747567303e-06, 'learning_rate/action_model': 3.3145060747567304e-05, 'epoch': 0.25}) +[2026-06-24 13:32:22,914][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.21477831900119781, 'train/grad_norm_pre_clip': 72.17278289794922, 'timing/data': 0.0004389439709484577, 'timing/model': 1.0799026219174266, 'learning_rate/qwen_vl_interface': 3.311012460044507e-06, 'learning_rate/action_model': 3.311012460044507e-05, 'epoch': 0.25}) +[2026-06-24 13:32:23,996][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.5685111880302429, 'train/grad_norm_pre_clip': 91.37261962890625, 'timing/data': 0.0003263428807258606, 'timing/model': 1.0807587418239564, 'learning_rate/qwen_vl_interface': 3.3075201035202466e-06, 'learning_rate/action_model': 3.3075201035202466e-05, 'epoch': 0.25}) +[2026-06-24 13:32:25,298][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.04628570005297661, 'train/grad_norm_pre_clip': 33.716583251953125, 'timing/data': 0.00032095308415591717, 'timing/model': 1.2997159659862518, 'learning_rate/qwen_vl_interface': 3.3040290074501e-06, 'learning_rate/action_model': 3.3040290074501e-05, 'epoch': 0.25}) +[2026-06-24 13:32:26,378][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.20029973983764648, 'train/grad_norm_pre_clip': 17.191415786743164, 'timing/data': 0.0004490839783102274, 'timing/model': 1.0784392599016428, 'learning_rate/qwen_vl_interface': 3.3005391740994065e-06, 'learning_rate/action_model': 3.300539174099406e-05, 'epoch': 0.25}) +[2026-06-24 13:32:27,465][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.0003870393557008356, 'train/grad_norm_pre_clip': 0.3897651433944702, 'timing/data': 0.00039622397162020206, 'timing/model': 1.085518630919978, 'learning_rate/qwen_vl_interface': 3.2970506057326746e-06, 'learning_rate/action_model': 3.2970506057326745e-05, 'epoch': 0.25}) +[2026-06-24 13:32:28,546][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.03619833663105965, 'train/grad_norm_pre_clip': 33.19089889526367, 'timing/data': 0.0003293990157544613, 'timing/model': 1.0789898480288684, 'learning_rate/qwen_vl_interface': 3.293563304613605e-06, 'learning_rate/action_model': 3.293563304613605e-05, 'epoch': 0.25}) +[2026-06-24 13:32:29,631][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.00719407806172967, 'train/grad_norm_pre_clip': 12.131255149841309, 'timing/data': 0.0003076479770243168, 'timing/model': 1.0833693880122155, 'learning_rate/qwen_vl_interface': 3.2900772730050632e-06, 'learning_rate/action_model': 3.2900772730050635e-05, 'epoch': 0.25}) +[2026-06-24 13:32:30,710][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.20248284935951233, 'train/grad_norm_pre_clip': 125.25701904296875, 'timing/data': 0.0003169449046254158, 'timing/model': 1.0776877060998231, 'learning_rate/qwen_vl_interface': 3.2865925131691024e-06, 'learning_rate/action_model': 3.286592513169102e-05, 'epoch': 0.25}) +[2026-06-24 13:32:31,790][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.19682864844799042, 'train/grad_norm_pre_clip': 81.55792236328125, 'timing/data': 0.00033932807855308056, 'timing/model': 1.0781372729688883, 'learning_rate/qwen_vl_interface': 3.283109027366944e-06, 'learning_rate/action_model': 3.283109027366944e-05, 'epoch': 0.25}) +[2026-06-24 13:32:32,872][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.006351195741444826, 'train/grad_norm_pre_clip': 8.179210662841797, 'timing/data': 0.00036096805706620216, 'timing/model': 1.0803006489295512, 'learning_rate/qwen_vl_interface': 3.2796268178589813e-06, 'learning_rate/action_model': 3.279626817858981e-05, 'epoch': 0.25}) +[2026-06-24 13:32:33,953][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.47310948371887207, 'train/grad_norm_pre_clip': 54.00740432739258, 'timing/data': 0.0003168859984725714, 'timing/model': 1.079498474020511, 'learning_rate/qwen_vl_interface': 3.276145886904786e-06, 'learning_rate/action_model': 3.2761458869047855e-05, 'epoch': 0.25}) +[2026-06-24 13:32:35,033][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.14063699543476105, 'train/grad_norm_pre_clip': 161.2659149169922, 'timing/data': 0.00033696298487484455, 'timing/model': 1.0789373219013214, 'learning_rate/qwen_vl_interface': 3.2726662367630953e-06, 'learning_rate/action_model': 3.272666236763095e-05, 'epoch': 0.25}) +[2026-06-24 13:32:36,115][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.003051535226404667, 'train/grad_norm_pre_clip': 5.261579990386963, 'timing/data': 0.0003428040072321892, 'timing/model': 1.0801225111354142, 'learning_rate/qwen_vl_interface': 3.269187869691814e-06, 'learning_rate/action_model': 3.269187869691814e-05, 'epoch': 0.25}) +[2026-06-24 13:32:37,196][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.002387954853475094, 'train/grad_norm_pre_clip': 3.0540149211883545, 'timing/data': 0.0003151718992739916, 'timing/model': 1.0791009131353348, 'learning_rate/qwen_vl_interface': 3.2657107879480215e-06, 'learning_rate/action_model': 3.265710787948021e-05, 'epoch': 0.25}) +[2026-06-24 13:32:38,279][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.05898040533065796, 'train/grad_norm_pre_clip': 108.67291259765625, 'timing/data': 0.00035450607538223267, 'timing/model': 1.0810571799520403, 'learning_rate/qwen_vl_interface': 3.2622349937879537e-06, 'learning_rate/action_model': 3.2622349937879537e-05, 'epoch': 0.25}) +[2026-06-24 13:32:39,360][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.05906480550765991, 'train/grad_norm_pre_clip': 31.941749572753906, 'timing/data': 0.00032841693609952927, 'timing/model': 1.080152133014053, 'learning_rate/qwen_vl_interface': 3.2587604894670177e-06, 'learning_rate/action_model': 3.258760489467018e-05, 'epoch': 0.25}) +[2026-06-24 13:32:40,444][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.02212250791490078, 'train/grad_norm_pre_clip': 10.06289291381836, 'timing/data': 0.00036826194263994694, 'timing/model': 1.0814979530405253, 'learning_rate/qwen_vl_interface': 3.2552872772397805e-06, 'learning_rate/action_model': 3.25528727723978e-05, 'epoch': 0.25}) +[2026-06-24 13:32:41,523][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.013444514945149422, 'train/grad_norm_pre_clip': 33.67182159423828, 'timing/data': 0.0003103830385953188, 'timing/model': 1.077886810991913, 'learning_rate/qwen_vl_interface': 3.251815359359975e-06, 'learning_rate/action_model': 3.251815359359975e-05, 'epoch': 0.25}) +[2026-06-24 13:32:42,896][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.3239526152610779, 'train/grad_norm_pre_clip': 154.2202911376953, 'timing/data': 0.00037141796201467514, 'timing/model': 1.3713403011206537, 'learning_rate/qwen_vl_interface': 3.2483447380804878e-06, 'learning_rate/action_model': 3.2483447380804874e-05, 'epoch': 0.25}) +[2026-06-24 13:32:43,973][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.016144482418894768, 'train/grad_norm_pre_clip': 22.688392639160156, 'timing/data': 0.00038034492172300816, 'timing/model': 1.0754602260421962, 'learning_rate/qwen_vl_interface': 3.244875415653368e-06, 'learning_rate/action_model': 3.2448754156533675e-05, 'epoch': 0.25}) +[2026-06-24 13:32:45,054][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.19801317155361176, 'train/grad_norm_pre_clip': 157.7885284423828, 'timing/data': 0.00034410692751407623, 'timing/model': 1.0797425529453903, 'learning_rate/qwen_vl_interface': 3.2414073943298235e-06, 'learning_rate/action_model': 3.241407394329823e-05, 'epoch': 0.25}) +[2026-06-24 13:32:46,137][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.060620881617069244, 'train/grad_norm_pre_clip': 72.42160034179688, 'timing/data': 0.00043276301585137844, 'timing/model': 1.0811737950425595, 'learning_rate/qwen_vl_interface': 3.2379406763602093e-06, 'learning_rate/action_model': 3.237940676360209e-05, 'epoch': 0.25}) +[2026-06-24 13:32:47,215][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.06661321222782135, 'train/grad_norm_pre_clip': 102.5641098022461, 'timing/data': 0.0004405379295349121, 'timing/model': 1.0759850281756371, 'learning_rate/qwen_vl_interface': 3.23447526399405e-06, 'learning_rate/action_model': 3.23447526399405e-05, 'epoch': 0.25}) +[2026-06-24 13:32:48,298][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.03395800665020943, 'train/grad_norm_pre_clip': 39.5421028137207, 'timing/data': 0.0003270551096647978, 'timing/model': 1.0808153050020337, 'learning_rate/qwen_vl_interface': 3.231011159480008e-06, 'learning_rate/action_model': 3.231011159480008e-05, 'epoch': 0.25}) +[2026-06-24 13:32:49,381][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.4474927484989166, 'train/grad_norm_pre_clip': 169.6642608642578, 'timing/data': 0.0003331161569803953, 'timing/model': 1.0817251249682158, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 0.25}) +[2026-06-24 13:32:50,466][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.0034603928215801716, 'train/grad_norm_pre_clip': 5.0188374519348145, 'timing/data': 0.0003160939086228609, 'timing/model': 1.083378012990579, 'learning_rate/qwen_vl_interface': 3.2240868829987114e-06, 'learning_rate/action_model': 3.224086882998711e-05, 'epoch': 0.25}) +[2026-06-24 13:32:51,547][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.10537312924861908, 'train/grad_norm_pre_clip': 143.97972106933594, 'timing/data': 0.00034551904536783695, 'timing/model': 1.0795537098310888, 'learning_rate/qwen_vl_interface': 3.220626715524543e-06, 'learning_rate/action_model': 3.220626715524543e-05, 'epoch': 0.25}) +[2026-06-24 13:32:52,624][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.08966876566410065, 'train/grad_norm_pre_clip': 130.5314483642578, 'timing/data': 0.00034450716339051723, 'timing/model': 1.0748794309329242, 'learning_rate/qwen_vl_interface': 3.217167864888667e-06, 'learning_rate/action_model': 3.217167864888667e-05, 'epoch': 0.25}) +[2026-06-24 13:32:53,704][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.32857340574264526, 'train/grad_norm_pre_clip': 64.4841079711914, 'timing/data': 0.00031120399944484234, 'timing/model': 1.0789973000064492, 'learning_rate/qwen_vl_interface': 3.213710333335492e-06, 'learning_rate/action_model': 3.2137103333354915e-05, 'epoch': 0.25}) +[2026-06-24 13:32:54,782][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.08780952543020248, 'train/grad_norm_pre_clip': 79.46765899658203, 'timing/data': 0.0003462410531938076, 'timing/model': 1.0765218811575323, 'learning_rate/qwen_vl_interface': 3.210254123108577e-06, 'learning_rate/action_model': 3.210254123108577e-05, 'epoch': 0.25}) +[2026-06-24 13:32:55,863][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.3044598698616028, 'train/grad_norm_pre_clip': 90.21170806884766, 'timing/data': 0.0003258830402046442, 'timing/model': 1.079127207165584, 'learning_rate/qwen_vl_interface': 3.2067992364506136e-06, 'learning_rate/action_model': 3.206799236450613e-05, 'epoch': 0.25}) +[2026-06-24 13:32:56,945][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.07226240634918213, 'train/grad_norm_pre_clip': 49.29478073120117, 'timing/data': 0.00041425786912441254, 'timing/model': 1.080346830189228, 'learning_rate/qwen_vl_interface': 3.2033456756034437e-06, 'learning_rate/action_model': 3.203345675603443e-05, 'epoch': 0.25}) +[2026-06-24 13:32:58,030][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.1318635642528534, 'train/grad_norm_pre_clip': 95.8919677734375, 'timing/data': 0.0003615790046751499, 'timing/model': 1.083711853949353, 'learning_rate/qwen_vl_interface': 3.199893442808046e-06, 'learning_rate/action_model': 3.199893442808046e-05, 'epoch': 0.25}) +[2026-06-24 13:32:59,316][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.05213109776377678, 'train/grad_norm_pre_clip': 31.87128257751465, 'timing/data': 0.00035056890919804573, 'timing/model': 1.284164336975664, 'learning_rate/qwen_vl_interface': 3.196442540304532e-06, 'learning_rate/action_model': 3.196442540304532e-05, 'epoch': 0.25}) +[2026-06-24 13:33:00,396][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.05128306522965431, 'train/grad_norm_pre_clip': 24.232980728149414, 'timing/data': 0.0003416920080780983, 'timing/model': 1.0781109558884054, 'learning_rate/qwen_vl_interface': 3.192992970332162e-06, 'learning_rate/action_model': 3.192992970332162e-05, 'epoch': 0.25}) +[2026-06-24 13:33:01,475][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.23998230695724487, 'train/grad_norm_pre_clip': 72.55689239501953, 'timing/data': 0.0003369830083101988, 'timing/model': 1.0778839921113104, 'learning_rate/qwen_vl_interface': 3.1895447351293215e-06, 'learning_rate/action_model': 3.1895447351293215e-05, 'epoch': 0.25}) +[2026-06-24 13:33:02,560][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.11817024648189545, 'train/grad_norm_pre_clip': 38.46102523803711, 'timing/data': 0.0003131080884486437, 'timing/model': 1.0838142139837146, 'learning_rate/qwen_vl_interface': 3.186097836933528e-06, 'learning_rate/action_model': 3.1860978369335285e-05, 'epoch': 0.25}) +[2026-06-24 13:33:03,641][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.210245743393898, 'train/grad_norm_pre_clip': 90.3239974975586, 'timing/data': 0.000318027101457119, 'timing/model': 1.078604097943753, 'learning_rate/qwen_vl_interface': 3.1826522779814453e-06, 'learning_rate/action_model': 3.182652277981445e-05, 'epoch': 0.25}) +[2026-06-24 13:33:04,724][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.01631629839539528, 'train/grad_norm_pre_clip': 21.042036056518555, 'timing/data': 0.00030851992778480053, 'timing/model': 1.082088336115703, 'learning_rate/qwen_vl_interface': 3.1792080605088517e-06, 'learning_rate/action_model': 3.179208060508851e-05, 'epoch': 0.25}) +[2026-06-24 13:33:05,810][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.1506417840719223, 'train/grad_norm_pre_clip': 43.42346954345703, 'timing/data': 0.0003231070004403591, 'timing/model': 1.0844741500914097, 'learning_rate/qwen_vl_interface': 3.1757651867506654e-06, 'learning_rate/action_model': 3.175765186750666e-05, 'epoch': 0.25}) +[2026-06-24 13:33:06,897][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.05298400670289993, 'train/grad_norm_pre_clip': 29.942968368530273, 'timing/data': 0.00033898698166012764, 'timing/model': 1.0850175831001252, 'learning_rate/qwen_vl_interface': 3.1723236589409322e-06, 'learning_rate/action_model': 3.172323658940932e-05, 'epoch': 0.25}) +[2026-06-24 13:33:07,981][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.028427015990018845, 'train/grad_norm_pre_clip': 10.491639137268066, 'timing/data': 0.0003900530282407999, 'timing/model': 1.0822884279768914, 'learning_rate/qwen_vl_interface': 3.1688834793128175e-06, 'learning_rate/action_model': 3.168883479312818e-05, 'epoch': 0.25}) +[2026-06-24 13:33:09,067][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.08125010132789612, 'train/grad_norm_pre_clip': 105.24011993408203, 'timing/data': 0.0003102831542491913, 'timing/model': 1.0839998349547386, 'learning_rate/qwen_vl_interface': 3.165444650098619e-06, 'learning_rate/action_model': 3.165444650098619e-05, 'epoch': 0.25}) +[2026-06-24 13:33:10,150][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.07223723828792572, 'train/grad_norm_pre_clip': 29.116498947143555, 'timing/data': 0.000363111961632967, 'timing/model': 1.081617632182315, 'learning_rate/qwen_vl_interface': 3.1620071735297548e-06, 'learning_rate/action_model': 3.1620071735297544e-05, 'epoch': 0.25}) +[2026-06-24 13:33:11,231][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.0033089672215282917, 'train/grad_norm_pre_clip': 2.327800750732422, 'timing/data': 0.0003226660192012787, 'timing/model': 1.0796417610254139, 'learning_rate/qwen_vl_interface': 3.158571051836769e-06, 'learning_rate/action_model': 3.158571051836769e-05, 'epoch': 0.25}) +[2026-06-24 13:33:12,314][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.037395767867565155, 'train/grad_norm_pre_clip': 43.255516052246094, 'timing/data': 0.00032388907857239246, 'timing/model': 1.0813980230595917, 'learning_rate/qwen_vl_interface': 3.1551362872493196e-06, 'learning_rate/action_model': 3.155136287249319e-05, 'epoch': 0.25}) +[2026-06-24 13:33:13,395][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.05583725869655609, 'train/grad_norm_pre_clip': 72.29554748535156, 'timing/data': 0.00033793505281209946, 'timing/model': 1.0790911209769547, 'learning_rate/qwen_vl_interface': 3.151702881996189e-06, 'learning_rate/action_model': 3.151702881996189e-05, 'epoch': 0.25}) +[2026-06-24 13:33:14,478][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.03558911383152008, 'train/grad_norm_pre_clip': 17.280353546142578, 'timing/data': 0.00031559308990836143, 'timing/model': 1.081288781017065, 'learning_rate/qwen_vl_interface': 3.1482708383052803e-06, 'learning_rate/action_model': 3.14827083830528e-05, 'epoch': 0.25}) +[2026-06-24 13:33:15,771][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.034330159425735474, 'train/grad_norm_pre_clip': 31.957212448120117, 'timing/data': 0.000420781085267663, 'timing/model': 1.2922256861347705, 'learning_rate/qwen_vl_interface': 3.144840158403604e-06, 'learning_rate/action_model': 3.144840158403604e-05, 'epoch': 0.25}) +[2026-06-24 13:33:16,854][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.003221086924895644, 'train/grad_norm_pre_clip': 1.6730858087539673, 'timing/data': 0.00035598897375166416, 'timing/model': 1.0806438531726599, 'learning_rate/qwen_vl_interface': 3.141410844517294e-06, 'learning_rate/action_model': 3.141410844517294e-05, 'epoch': 0.25}) +[2026-06-24 13:33:17,937][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.18608176708221436, 'train/grad_norm_pre_clip': 31.088655471801758, 'timing/data': 0.0003436850383877754, 'timing/model': 1.0818277390208095, 'learning_rate/qwen_vl_interface': 3.1379828988715955e-06, 'learning_rate/action_model': 3.1379828988715956e-05, 'epoch': 0.25}) +[2026-06-24 13:33:19,020][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.15804098546504974, 'train/grad_norm_pre_clip': 74.61630249023438, 'timing/data': 0.0004034989979118109, 'timing/model': 1.0808162910398096, 'learning_rate/qwen_vl_interface': 3.1345563236908626e-06, 'learning_rate/action_model': 3.134556323690863e-05, 'epoch': 0.25}) +[2026-06-24 13:33:20,103][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.23873157799243927, 'train/grad_norm_pre_clip': 115.72886657714844, 'timing/data': 0.0003243789542466402, 'timing/model': 1.081684044096619, 'learning_rate/qwen_vl_interface': 3.1311311211985645e-06, 'learning_rate/action_model': 3.1311311211985644e-05, 'epoch': 0.25}) +[2026-06-24 13:33:21,185][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.01896706037223339, 'train/grad_norm_pre_clip': 26.49991798400879, 'timing/data': 0.00033461907878518105, 'timing/model': 1.0800344259478152, 'learning_rate/qwen_vl_interface': 3.127707293617277e-06, 'learning_rate/action_model': 3.1277072936172766e-05, 'epoch': 0.25}) +[2026-06-24 13:33:22,270][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.03970217704772949, 'train/grad_norm_pre_clip': 46.42405319213867, 'timing/data': 0.00039648497477173805, 'timing/model': 1.0839350861497223, 'learning_rate/qwen_vl_interface': 3.124284843168688e-06, 'learning_rate/action_model': 3.1242848431686875e-05, 'epoch': 0.25}) +[2026-06-24 13:33:23,355][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.017029430717229843, 'train/grad_norm_pre_clip': 40.129859924316406, 'timing/data': 0.000528302974998951, 'timing/model': 1.0828408929519355, 'learning_rate/qwen_vl_interface': 3.1208637720735813e-06, 'learning_rate/action_model': 3.1208637720735814e-05, 'epoch': 0.25}) +[2026-06-24 13:33:24,439][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.002043297979980707, 'train/grad_norm_pre_clip': 4.077902793884277, 'timing/data': 0.00035483716055750847, 'timing/model': 1.081476490944624, 'learning_rate/qwen_vl_interface': 3.117444082551861e-06, 'learning_rate/action_model': 3.117444082551861e-05, 'epoch': 0.25}) +[2026-06-24 13:33:25,525][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.015661999583244324, 'train/grad_norm_pre_clip': 17.108707427978516, 'timing/data': 0.00032405881211161613, 'timing/model': 1.0848627148661762, 'learning_rate/qwen_vl_interface': 3.114025776822523e-06, 'learning_rate/action_model': 3.1140257768225225e-05, 'epoch': 0.25}) +[2026-06-24 13:33:26,609][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.16117745637893677, 'train/grad_norm_pre_clip': 67.49650573730469, 'timing/data': 0.00030910992063581944, 'timing/model': 1.0826050848700106, 'learning_rate/qwen_vl_interface': 3.1106088571036647e-06, 'learning_rate/action_model': 3.1106088571036644e-05, 'epoch': 0.25}) +[2026-06-24 13:33:27,692][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.013056552037596703, 'train/grad_norm_pre_clip': 7.531402587890625, 'timing/data': 0.00031323800794780254, 'timing/model': 1.080967750865966, 'learning_rate/qwen_vl_interface': 3.1071933256124954e-06, 'learning_rate/action_model': 3.1071933256124955e-05, 'epoch': 0.25}) +[2026-06-24 13:33:28,778][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.10834267735481262, 'train/grad_norm_pre_clip': 45.185211181640625, 'timing/data': 0.000358393182978034, 'timing/model': 1.0847494178451598, 'learning_rate/qwen_vl_interface': 3.1037791845653116e-06, 'learning_rate/action_model': 3.1037791845653115e-05, 'epoch': 0.25}) +[2026-06-24 13:33:29,864][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.38889798521995544, 'train/grad_norm_pre_clip': 96.06643676757812, 'timing/data': 0.00030960096046328545, 'timing/model': 1.0841731859836727, 'learning_rate/qwen_vl_interface': 3.1003664361775138e-06, 'learning_rate/action_model': 3.1003664361775135e-05, 'epoch': 0.25}) +[2026-06-24 13:33:30,949][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.021757980808615685, 'train/grad_norm_pre_clip': 26.94510269165039, 'timing/data': 0.0003474028781056404, 'timing/model': 1.0837566060945392, 'learning_rate/qwen_vl_interface': 3.0969550826635987e-06, 'learning_rate/action_model': 3.0969550826635986e-05, 'epoch': 0.25}) +[2026-06-24 13:33:32,241][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.07868118584156036, 'train/grad_norm_pre_clip': 43.576534271240234, 'timing/data': 0.0004592030309140682, 'timing/model': 1.2902175968047231, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 0.25}) +[2026-06-24 13:33:33,322][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.20377419888973236, 'train/grad_norm_pre_clip': 67.34461975097656, 'timing/data': 0.00034734210930764675, 'timing/model': 1.0785659591201693, 'learning_rate/qwen_vl_interface': 3.0901365691108633e-06, 'learning_rate/action_model': 3.090136569110863e-05, 'epoch': 0.26}) +[2026-06-24 13:33:34,410][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.05465732887387276, 'train/grad_norm_pre_clip': 15.523117065429688, 'timing/data': 0.0003314830828458071, 'timing/model': 1.0865224979352206, 'learning_rate/qwen_vl_interface': 3.0867294134965043e-06, 'learning_rate/action_model': 3.086729413496504e-05, 'epoch': 0.26}) +[2026-06-24 13:33:35,493][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.009526271373033524, 'train/grad_norm_pre_clip': 6.457429885864258, 'timing/data': 0.0003163139335811138, 'timing/model': 1.0817905098665506, 'learning_rate/qwen_vl_interface': 3.0833236616049445e-06, 'learning_rate/action_model': 3.083323661604944e-05, 'epoch': 0.26}) +[2026-06-24 13:33:36,574][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.11143626272678375, 'train/grad_norm_pre_clip': 57.3583869934082, 'timing/data': 0.00032424996607005596, 'timing/model': 1.0796610610559583, 'learning_rate/qwen_vl_interface': 3.0799193156461347e-06, 'learning_rate/action_model': 3.079919315646135e-05, 'epoch': 0.26}) +[2026-06-24 13:33:37,656][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.2061740607023239, 'train/grad_norm_pre_clip': 127.0179443359375, 'timing/data': 0.0003412810619920492, 'timing/model': 1.0801225998438895, 'learning_rate/qwen_vl_interface': 3.0765163778291226e-06, 'learning_rate/action_model': 3.076516377829122e-05, 'epoch': 0.26}) +[2026-06-24 13:33:38,736][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.05728180333971977, 'train/grad_norm_pre_clip': 53.65299606323242, 'timing/data': 0.0003378340043127537, 'timing/model': 1.0785306610632688, 'learning_rate/qwen_vl_interface': 3.073114850362037e-06, 'learning_rate/action_model': 3.0731148503620365e-05, 'epoch': 0.26}) +[2026-06-24 13:33:39,822][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.04240167886018753, 'train/grad_norm_pre_clip': 37.64977264404297, 'timing/data': 0.00031208619475364685, 'timing/model': 1.0838982970453799, 'learning_rate/qwen_vl_interface': 3.069714735452088e-06, 'learning_rate/action_model': 3.069714735452088e-05, 'epoch': 0.26}) +[2026-06-24 13:33:40,905][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.23812855780124664, 'train/grad_norm_pre_clip': 59.789615631103516, 'timing/data': 0.0003064759075641632, 'timing/model': 1.0815261220559478, 'learning_rate/qwen_vl_interface': 3.066316035305577e-06, 'learning_rate/action_model': 3.066316035305577e-05, 'epoch': 0.26}) +[2026-06-24 13:33:41,993][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.16392771899700165, 'train/grad_norm_pre_clip': 62.43565368652344, 'timing/data': 0.0004330140072852373, 'timing/model': 1.0858140599448234, 'learning_rate/qwen_vl_interface': 3.062918752127885e-06, 'learning_rate/action_model': 3.062918752127885e-05, 'epoch': 0.26}) +[2026-06-24 13:33:43,076][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.007017331663519144, 'train/grad_norm_pre_clip': 8.022829055786133, 'timing/data': 0.0003418021369725466, 'timing/model': 1.0815952559933066, 'learning_rate/qwen_vl_interface': 3.0595228881234666e-06, 'learning_rate/action_model': 3.0595228881234665e-05, 'epoch': 0.26}) +[2026-06-24 13:33:44,159][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.0154024763032794, 'train/grad_norm_pre_clip': 4.606180667877197, 'timing/data': 0.00046310992911458015, 'timing/model': 1.081457257969305, 'learning_rate/qwen_vl_interface': 3.056128445495866e-06, 'learning_rate/action_model': 3.056128445495865e-05, 'epoch': 0.26}) +[2026-06-24 13:33:45,242][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.02089347131550312, 'train/grad_norm_pre_clip': 31.89679527282715, 'timing/data': 0.00032773613929748535, 'timing/model': 1.0812868899665773, 'learning_rate/qwen_vl_interface': 3.0527354264476976e-06, 'learning_rate/action_model': 3.0527354264476974e-05, 'epoch': 0.26}) +[2026-06-24 13:33:46,329][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.2510084807872772, 'train/grad_norm_pre_clip': 67.7594985961914, 'timing/data': 0.0005771249998360872, 'timing/model': 1.0846436789724976, 'learning_rate/qwen_vl_interface': 3.049343833180658e-06, 'learning_rate/action_model': 3.049343833180658e-05, 'epoch': 0.26}) +[2026-06-24 13:33:47,412][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.1791999340057373, 'train/grad_norm_pre_clip': 108.04911804199219, 'timing/data': 0.00038062501698732376, 'timing/model': 1.0811339060310274, 'learning_rate/qwen_vl_interface': 3.0459536678955094e-06, 'learning_rate/action_model': 3.0459536678955092e-05, 'epoch': 0.26}) +[2026-06-24 13:33:48,791][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.0042066434398293495, 'train/grad_norm_pre_clip': 5.0753021240234375, 'timing/data': 0.0003707071300595999, 'timing/model': 1.377765147946775, 'learning_rate/qwen_vl_interface': 3.042564932792099e-06, 'learning_rate/action_model': 3.0425649327920992e-05, 'epoch': 0.26}) +[2026-06-24 13:33:49,878][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.09952231496572495, 'train/grad_norm_pre_clip': 64.98885345458984, 'timing/data': 0.00031020306050777435, 'timing/model': 1.0855882288888097, 'learning_rate/qwen_vl_interface': 3.0391776300693394e-06, 'learning_rate/action_model': 3.0391776300693393e-05, 'epoch': 0.26}) +[2026-06-24 13:33:50,963][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.07830125093460083, 'train/grad_norm_pre_clip': 55.47799301147461, 'timing/data': 0.00031863804906606674, 'timing/model': 1.0833572391420603, 'learning_rate/qwen_vl_interface': 3.0357917619252093e-06, 'learning_rate/action_model': 3.0357917619252092e-05, 'epoch': 0.26}) +[2026-06-24 13:33:52,050][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.10760583728551865, 'train/grad_norm_pre_clip': 100.89620208740234, 'timing/data': 0.00036737113259732723, 'timing/model': 1.0846120479982346, 'learning_rate/qwen_vl_interface': 3.032407330556769e-06, 'learning_rate/action_model': 3.0324073305567684e-05, 'epoch': 0.26}) +[2026-06-24 13:33:53,134][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.0577126108109951, 'train/grad_norm_pre_clip': 46.25387954711914, 'timing/data': 0.00034507899545133114, 'timing/model': 1.0825834390707314, 'learning_rate/qwen_vl_interface': 3.029024338160134e-06, 'learning_rate/action_model': 3.0290243381601334e-05, 'epoch': 0.26}) +[2026-06-24 13:33:54,221][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.11123163253068924, 'train/grad_norm_pre_clip': 72.09282684326172, 'timing/data': 0.00032390886917710304, 'timing/model': 1.085684928111732, 'learning_rate/qwen_vl_interface': 3.025642786930493e-06, 'learning_rate/action_model': 3.025642786930493e-05, 'epoch': 0.26}) +[2026-06-24 13:33:55,305][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.18517763912677765, 'train/grad_norm_pre_clip': 115.67593383789062, 'timing/data': 0.0003475029952824116, 'timing/model': 1.0820759790949523, 'learning_rate/qwen_vl_interface': 3.0222626790620995e-06, 'learning_rate/action_model': 3.0222626790620996e-05, 'epoch': 0.26}) +[2026-06-24 13:33:56,390][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.22244159877300262, 'train/grad_norm_pre_clip': 126.21504974365234, 'timing/data': 0.0003279659431427717, 'timing/model': 1.083583652973175, 'learning_rate/qwen_vl_interface': 3.0188840167482664e-06, 'learning_rate/action_model': 3.0188840167482663e-05, 'epoch': 0.26}) +[2026-06-24 13:33:57,479][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.297995388507843, 'train/grad_norm_pre_clip': 53.14533233642578, 'timing/data': 0.00035241199657320976, 'timing/model': 1.0871507269330323, 'learning_rate/qwen_vl_interface': 3.0155068021813714e-06, 'learning_rate/action_model': 3.0155068021813715e-05, 'epoch': 0.26}) +[2026-06-24 13:33:58,566][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.004472777713090181, 'train/grad_norm_pre_clip': 4.086312294006348, 'timing/data': 0.0003356211818754673, 'timing/model': 1.0853709250222892, 'learning_rate/qwen_vl_interface': 3.0121310375528527e-06, 'learning_rate/action_model': 3.0121310375528527e-05, 'epoch': 0.26}) +[2026-06-24 13:33:59,654][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.054320115596055984, 'train/grad_norm_pre_clip': 58.4542236328125, 'timing/data': 0.0003557878080755472, 'timing/model': 1.086767098866403, 'learning_rate/qwen_vl_interface': 3.00875672505321e-06, 'learning_rate/action_model': 3.00875672505321e-05, 'epoch': 0.26}) +[2026-06-24 13:34:00,740][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.0403069444000721, 'train/grad_norm_pre_clip': 56.054283142089844, 'timing/data': 0.00038115610368549824, 'timing/model': 1.0843725628219545, 'learning_rate/qwen_vl_interface': 3.0053838668719944e-06, 'learning_rate/action_model': 3.0053838668719943e-05, 'epoch': 0.26}) +[2026-06-24 13:34:01,827][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.0014765947125852108, 'train/grad_norm_pre_clip': 0.5467697978019714, 'timing/data': 0.00032152398489415646, 'timing/model': 1.0854210539255291, 'learning_rate/qwen_vl_interface': 3.0020124651978165e-06, 'learning_rate/action_model': 3.0020124651978166e-05, 'epoch': 0.26}) +[2026-06-24 13:34:02,911][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.1218210980296135, 'train/grad_norm_pre_clip': 42.366085052490234, 'timing/data': 0.0003286271821707487, 'timing/model': 1.0827042709570378, 'learning_rate/qwen_vl_interface': 2.9986425222183463e-06, 'learning_rate/action_model': 2.998642522218346e-05, 'epoch': 0.26}) +[2026-06-24 13:34:04,198][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.16231299936771393, 'train/grad_norm_pre_clip': 35.71400451660156, 'timing/data': 0.00034380611032247543, 'timing/model': 1.2848136872053146, 'learning_rate/qwen_vl_interface': 2.9952740401202996e-06, 'learning_rate/action_model': 2.9952740401202995e-05, 'epoch': 0.26}) +[2026-06-24 13:34:05,279][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.05841464549303055, 'train/grad_norm_pre_clip': 77.148681640625, 'timing/data': 0.00032832706347107887, 'timing/model': 1.0797640210948884, 'learning_rate/qwen_vl_interface': 2.9919070210894484e-06, 'learning_rate/action_model': 2.9919070210894484e-05, 'epoch': 0.26}) +[2026-06-24 13:34:06,367][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.014577612280845642, 'train/grad_norm_pre_clip': 15.488420486450195, 'timing/data': 0.00034401589073240757, 'timing/model': 1.086354263825342, 'learning_rate/qwen_vl_interface': 2.9885414673106176e-06, 'learning_rate/action_model': 2.988541467310618e-05, 'epoch': 0.26}) +[2026-06-24 13:34:07,447][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.13903452455997467, 'train/grad_norm_pre_clip': 43.4708251953125, 'timing/data': 0.0003249610308557749, 'timing/model': 1.0789686411153525, 'learning_rate/qwen_vl_interface': 2.9851773809676745e-06, 'learning_rate/action_model': 2.985177380967674e-05, 'epoch': 0.26}) +[2026-06-24 13:34:08,532][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.05125483125448227, 'train/grad_norm_pre_clip': 31.175735473632812, 'timing/data': 0.00030516297556459904, 'timing/model': 1.0832653949037194, 'learning_rate/qwen_vl_interface': 2.9818147642435395e-06, 'learning_rate/action_model': 2.9818147642435395e-05, 'epoch': 0.26}) +[2026-06-24 13:34:09,619][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.005746710579842329, 'train/grad_norm_pre_clip': 9.162677764892578, 'timing/data': 0.00034943711943924427, 'timing/model': 1.0851205219514668, 'learning_rate/qwen_vl_interface': 2.9784536193201797e-06, 'learning_rate/action_model': 2.9784536193201796e-05, 'epoch': 0.26}) +[2026-06-24 13:34:10,704][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.07983790338039398, 'train/grad_norm_pre_clip': 82.73129272460938, 'timing/data': 0.00034481799229979515, 'timing/model': 1.0836841859854758, 'learning_rate/qwen_vl_interface': 2.9750939483786057e-06, 'learning_rate/action_model': 2.9750939483786056e-05, 'epoch': 0.26}) +[2026-06-24 13:34:11,788][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.13617007434368134, 'train/grad_norm_pre_clip': 30.4013671875, 'timing/data': 0.00033749407157301903, 'timing/model': 1.0824774799402803, 'learning_rate/qwen_vl_interface': 2.9717357535988682e-06, 'learning_rate/action_model': 2.9717357535988683e-05, 'epoch': 0.26}) +[2026-06-24 13:34:12,872][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.002038288861513138, 'train/grad_norm_pre_clip': 1.823535442352295, 'timing/data': 0.00031729601323604584, 'timing/model': 1.08261112193577, 'learning_rate/qwen_vl_interface': 2.968379037160066e-06, 'learning_rate/action_model': 2.9683790371600657e-05, 'epoch': 0.26}) +[2026-06-24 13:34:13,952][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.0437515489757061, 'train/grad_norm_pre_clip': 38.755367279052734, 'timing/data': 0.00030320906080305576, 'timing/model': 1.0783923577982932, 'learning_rate/qwen_vl_interface': 2.9650238012403365e-06, 'learning_rate/action_model': 2.9650238012403364e-05, 'epoch': 0.26}) +[2026-06-24 13:34:15,038][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.05328662693500519, 'train/grad_norm_pre_clip': 54.87220764160156, 'timing/data': 0.0003233968745917082, 'timing/model': 1.0845454998780042, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 0.26}) +[2026-06-24 13:34:16,122][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.024922756478190422, 'train/grad_norm_pre_clip': 50.78710174560547, 'timing/data': 0.00031581311486661434, 'timing/model': 1.0818082788027823, 'learning_rate/qwen_vl_interface': 2.9583177796658326e-06, 'learning_rate/action_model': 2.9583177796658324e-05, 'epoch': 0.26}) +[2026-06-24 13:34:17,209][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.020975805819034576, 'train/grad_norm_pre_clip': 25.28872299194336, 'timing/data': 0.00033595110289752483, 'timing/model': 1.085769894067198, 'learning_rate/qwen_vl_interface': 2.9549669983625233e-06, 'learning_rate/action_model': 2.9549669983625234e-05, 'epoch': 0.26}) +[2026-06-24 13:34:18,296][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.01641525886952877, 'train/grad_norm_pre_clip': 23.938323974609375, 'timing/data': 0.00031578284688293934, 'timing/model': 1.0851436289958656, 'learning_rate/qwen_vl_interface': 2.9516177062812123e-06, 'learning_rate/action_model': 2.9516177062812122e-05, 'epoch': 0.26}) +[2026-06-24 13:34:19,378][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.009803451597690582, 'train/grad_norm_pre_clip': 7.343501567840576, 'timing/data': 0.0003280059900134802, 'timing/model': 1.080348932882771, 'learning_rate/qwen_vl_interface': 2.94826990559522e-06, 'learning_rate/action_model': 2.94826990559522e-05, 'epoch': 0.26}) +[2026-06-24 13:34:20,459][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.018152816221117973, 'train/grad_norm_pre_clip': 28.23763084411621, 'timing/data': 0.00036742095835506916, 'timing/model': 1.079725171905011, 'learning_rate/qwen_vl_interface': 2.9449235984768946e-06, 'learning_rate/action_model': 2.9449235984768946e-05, 'epoch': 0.26}) +[2026-06-24 13:34:21,544][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.10532752424478531, 'train/grad_norm_pre_clip': 24.80719757080078, 'timing/data': 0.0003296900540590286, 'timing/model': 1.0833722359966487, 'learning_rate/qwen_vl_interface': 2.9415787870976216e-06, 'learning_rate/action_model': 2.9415787870976212e-05, 'epoch': 0.26}) +[2026-06-24 13:34:22,629][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.3210853934288025, 'train/grad_norm_pre_clip': 127.19931030273438, 'timing/data': 0.0005269800312817097, 'timing/model': 1.083161351038143, 'learning_rate/qwen_vl_interface': 2.9382354736278097e-06, 'learning_rate/action_model': 2.9382354736278096e-05, 'epoch': 0.26}) +[2026-06-24 13:34:24,011][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.053844090551137924, 'train/grad_norm_pre_clip': 29.819990158081055, 'timing/data': 0.0003172860015183687, 'timing/model': 1.3798308800905943, 'learning_rate/qwen_vl_interface': 2.9348936602369036e-06, 'learning_rate/action_model': 2.9348936602369037e-05, 'epoch': 0.26}) +[2026-06-24 13:34:25,093][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.5993798971176147, 'train/grad_norm_pre_clip': 139.46658325195312, 'timing/data': 0.00035828305408358574, 'timing/model': 1.0801410300191492, 'learning_rate/qwen_vl_interface': 2.931553349093364e-06, 'learning_rate/action_model': 2.931553349093364e-05, 'epoch': 0.26}) +[2026-06-24 13:34:26,175][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.031878046691417694, 'train/grad_norm_pre_clip': 33.0908203125, 'timing/data': 0.0003192590083926916, 'timing/model': 1.0809450328815728, 'learning_rate/qwen_vl_interface': 2.9282145423646856e-06, 'learning_rate/action_model': 2.9282145423646858e-05, 'epoch': 0.26}) +[2026-06-24 13:34:27,259][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.12809684872627258, 'train/grad_norm_pre_clip': 57.296607971191406, 'timing/data': 0.00032453006133437157, 'timing/model': 1.0824953569099307, 'learning_rate/qwen_vl_interface': 2.9248772422173855e-06, 'learning_rate/action_model': 2.9248772422173852e-05, 'epoch': 0.26}) +[2026-06-24 13:34:28,341][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.010781697928905487, 'train/grad_norm_pre_clip': 4.323387622833252, 'timing/data': 0.0004449461121112108, 'timing/model': 1.0805974639952183, 'learning_rate/qwen_vl_interface': 2.9215414508169964e-06, 'learning_rate/action_model': 2.9215414508169963e-05, 'epoch': 0.26}) +[2026-06-24 13:34:29,427][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.13159428536891937, 'train/grad_norm_pre_clip': 71.06391906738281, 'timing/data': 0.0003353608772158623, 'timing/model': 1.083545383065939, 'learning_rate/qwen_vl_interface': 2.918207170328081e-06, 'learning_rate/action_model': 2.918207170328081e-05, 'epoch': 0.26}) +[2026-06-24 13:34:30,509][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.145324245095253, 'train/grad_norm_pre_clip': 80.89166259765625, 'timing/data': 0.0003125770017504692, 'timing/model': 1.08103052014485, 'learning_rate/qwen_vl_interface': 2.914874402914219e-06, 'learning_rate/action_model': 2.914874402914219e-05, 'epoch': 0.26}) +[2026-06-24 13:34:31,591][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.08625081181526184, 'train/grad_norm_pre_clip': 29.905637741088867, 'timing/data': 0.0003552772104740143, 'timing/model': 1.0798326910007745, 'learning_rate/qwen_vl_interface': 2.9115431507380025e-06, 'learning_rate/action_model': 2.9115431507380024e-05, 'epoch': 0.26}) +[2026-06-24 13:34:32,674][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.027542714029550552, 'train/grad_norm_pre_clip': 70.0718765258789, 'timing/data': 0.0004020559135824442, 'timing/model': 1.082167279208079, 'learning_rate/qwen_vl_interface': 2.9082134159610474e-06, 'learning_rate/action_model': 2.9082134159610474e-05, 'epoch': 0.26}) +[2026-06-24 13:34:33,757][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.14680513739585876, 'train/grad_norm_pre_clip': 74.9535140991211, 'timing/data': 0.0003181879874318838, 'timing/model': 1.0809404759202152, 'learning_rate/qwen_vl_interface': 2.9048852007439825e-06, 'learning_rate/action_model': 2.9048852007439827e-05, 'epoch': 0.26}) +[2026-06-24 13:34:34,843][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.03663952276110649, 'train/grad_norm_pre_clip': 28.07067108154297, 'timing/data': 0.0003185691311955452, 'timing/model': 1.0844912191387266, 'learning_rate/qwen_vl_interface': 2.901558507246452e-06, 'learning_rate/action_model': 2.901558507246452e-05, 'epoch': 0.26}) +[2026-06-24 13:34:35,928][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.1600237339735031, 'train/grad_norm_pre_clip': 36.88161849975586, 'timing/data': 0.0003148110117763281, 'timing/model': 1.0829845890402794, 'learning_rate/qwen_vl_interface': 2.898233337627108e-06, 'learning_rate/action_model': 2.8982333376271075e-05, 'epoch': 0.26}) +[2026-06-24 13:34:37,015][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.1666678935289383, 'train/grad_norm_pre_clip': 76.2059555053711, 'timing/data': 0.00030915113165974617, 'timing/model': 1.0852951819542795, 'learning_rate/qwen_vl_interface': 2.8949096940436165e-06, 'learning_rate/action_model': 2.8949096940436164e-05, 'epoch': 0.26}) +[2026-06-24 13:34:38,098][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.07356031984090805, 'train/grad_norm_pre_clip': 40.54109573364258, 'timing/data': 0.0003554981667548418, 'timing/model': 1.0822052189614624, 'learning_rate/qwen_vl_interface': 2.8915875786526597e-06, 'learning_rate/action_model': 2.8915875786526596e-05, 'epoch': 0.26}) +[2026-06-24 13:34:39,184][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.05264967307448387, 'train/grad_norm_pre_clip': 53.93688201904297, 'timing/data': 0.00033638207241892815, 'timing/model': 1.0835514990612864, 'learning_rate/qwen_vl_interface': 2.888266993609914e-06, 'learning_rate/action_model': 2.8882669936099138e-05, 'epoch': 0.26}) +[2026-06-24 13:34:40,268][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.1084781214594841, 'train/grad_norm_pre_clip': 116.27239227294922, 'timing/data': 0.0003308509476482868, 'timing/model': 1.082451453898102, 'learning_rate/qwen_vl_interface': 2.8849479410700787e-06, 'learning_rate/action_model': 2.8849479410700785e-05, 'epoch': 0.26}) +[2026-06-24 13:34:41,352][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.06427137553691864, 'train/grad_norm_pre_clip': 70.043212890625, 'timing/data': 0.00033197388984262943, 'timing/model': 1.082842499949038, 'learning_rate/qwen_vl_interface': 2.8816304231868453e-06, 'learning_rate/action_model': 2.881630423186845e-05, 'epoch': 0.26}) +[2026-06-24 13:34:42,678][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.0464044027030468, 'train/grad_norm_pre_clip': 30.983516693115234, 'timing/data': 0.0004022051580250263, 'timing/model': 1.3236101779621094, 'learning_rate/qwen_vl_interface': 2.8783144421129177e-06, 'learning_rate/action_model': 2.8783144421129178e-05, 'epoch': 0.26}) +[2026-06-24 13:34:43,759][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.27389732003211975, 'train/grad_norm_pre_clip': 59.47484588623047, 'timing/data': 0.00030733784660696983, 'timing/model': 1.0798171099741012, 'learning_rate/qwen_vl_interface': 2.8750000000000013e-06, 'learning_rate/action_model': 2.875000000000001e-05, 'epoch': 0.26}) +[2026-06-24 13:34:44,844][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.009574594907462597, 'train/grad_norm_pre_clip': 6.986159801483154, 'timing/data': 0.0003346179146319628, 'timing/model': 1.083075893810019, 'learning_rate/qwen_vl_interface': 2.8716870989987983e-06, 'learning_rate/action_model': 2.8716870989987983e-05, 'epoch': 0.26}) +[2026-06-24 13:34:45,928][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.042603280395269394, 'train/grad_norm_pre_clip': 72.74617767333984, 'timing/data': 0.000327476067468524, 'timing/model': 1.0830138591118157, 'learning_rate/qwen_vl_interface': 2.8683757412590154e-06, 'learning_rate/action_model': 2.8683757412590152e-05, 'epoch': 0.26}) +[2026-06-24 13:34:47,011][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.05320719629526138, 'train/grad_norm_pre_clip': 40.0384407043457, 'timing/data': 0.0003136489540338516, 'timing/model': 1.0815553409047425, 'learning_rate/qwen_vl_interface': 2.865065928929357e-06, 'learning_rate/action_model': 2.8650659289293567e-05, 'epoch': 0.26}) +[2026-06-24 13:34:48,100][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.08059478551149368, 'train/grad_norm_pre_clip': 46.93573760986328, 'timing/data': 0.00031469203531742096, 'timing/model': 1.0870466690976173, 'learning_rate/qwen_vl_interface': 2.8617576641575265e-06, 'learning_rate/action_model': 2.8617576641575266e-05, 'epoch': 0.26}) +[2026-06-24 13:34:49,182][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.14277337491512299, 'train/grad_norm_pre_clip': 92.74687194824219, 'timing/data': 0.00032260618172585964, 'timing/model': 1.0809483940247446, 'learning_rate/qwen_vl_interface': 2.8584509490902167e-06, 'learning_rate/action_model': 2.8584509490902168e-05, 'epoch': 0.26}) +[2026-06-24 13:34:50,269][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.12439893931150436, 'train/grad_norm_pre_clip': 60.91547775268555, 'timing/data': 0.0003469719085842371, 'timing/model': 1.0849264189600945, 'learning_rate/qwen_vl_interface': 2.8551457858731224e-06, 'learning_rate/action_model': 2.8551457858731223e-05, 'epoch': 0.26}) +[2026-06-24 13:34:51,353][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.17935320734977722, 'train/grad_norm_pre_clip': 47.179039001464844, 'timing/data': 0.00035732099786400795, 'timing/model': 1.081963153090328, 'learning_rate/qwen_vl_interface': 2.851842176650928e-06, 'learning_rate/action_model': 2.8518421766509278e-05, 'epoch': 0.26}) +[2026-06-24 13:34:52,439][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.15215779840946198, 'train/grad_norm_pre_clip': 42.98225402832031, 'timing/data': 0.0003203118685632944, 'timing/model': 1.0848292289301753, 'learning_rate/qwen_vl_interface': 2.848540123567307e-06, 'learning_rate/action_model': 2.848540123567307e-05, 'epoch': 0.26}) +[2026-06-24 13:34:53,526][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.006709638517349958, 'train/grad_norm_pre_clip': 5.13014554977417, 'timing/data': 0.00037298095412552357, 'timing/model': 1.0855435440316796, 'learning_rate/qwen_vl_interface': 2.845239628764928e-06, 'learning_rate/action_model': 2.845239628764928e-05, 'epoch': 0.26}) +[2026-06-24 13:34:54,608][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.0578935332596302, 'train/grad_norm_pre_clip': 19.801807403564453, 'timing/data': 0.00036007585003972054, 'timing/model': 1.079786258051172, 'learning_rate/qwen_vl_interface': 2.8419406943854473e-06, 'learning_rate/action_model': 2.8419406943854472e-05, 'epoch': 0.26}) +[2026-06-24 13:34:55,707][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.05859439820051193, 'train/grad_norm_pre_clip': 23.39092445373535, 'timing/data': 0.0004142378456890583, 'timing/model': 1.0901000469457358, 'learning_rate/qwen_vl_interface': 2.8386433225695043e-06, 'learning_rate/action_model': 2.8386433225695043e-05, 'epoch': 0.26}) +[2026-06-24 13:34:56,795][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.18713189661502838, 'train/grad_norm_pre_clip': 86.43955993652344, 'timing/data': 0.00032370793633162975, 'timing/model': 1.0864492901600897, 'learning_rate/qwen_vl_interface': 2.8353475154567294e-06, 'learning_rate/action_model': 2.8353475154567294e-05, 'epoch': 0.26}) +[2026-06-24 13:34:57,883][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.10172848403453827, 'train/grad_norm_pre_clip': 53.07400131225586, 'timing/data': 0.00031850789673626423, 'timing/model': 1.0865133421029896, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 0.26}) +[2026-06-24 13:34:58,972][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.01185718085616827, 'train/grad_norm_pre_clip': 14.551647186279297, 'timing/data': 0.0003394379746168852, 'timing/model': 1.0873834178783, 'learning_rate/qwen_vl_interface': 2.8287606038941236e-06, 'learning_rate/action_model': 2.8287606038941235e-05, 'epoch': 0.26}) +[2026-06-24 13:35:00,054][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.0072939046658575535, 'train/grad_norm_pre_clip': 6.772488594055176, 'timing/data': 0.0003308509476482868, 'timing/model': 1.0813947969581932, 'learning_rate/qwen_vl_interface': 2.825469503718466e-06, 'learning_rate/action_model': 2.8254695037184657e-05, 'epoch': 0.26}) +[2026-06-24 13:35:01,329][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.1322525590658188, 'train/grad_norm_pre_clip': 266.8428955078125, 'timing/data': 0.00031982106156647205, 'timing/model': 1.272504101973027, 'learning_rate/qwen_vl_interface': 2.822179976794325e-06, 'learning_rate/action_model': 2.8221799767943254e-05, 'epoch': 0.26}) +[2026-06-24 13:35:02,412][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.09691169857978821, 'train/grad_norm_pre_clip': 59.55992126464844, 'timing/data': 0.00034439796581864357, 'timing/model': 1.0813010749407113, 'learning_rate/qwen_vl_interface': 2.8188920252562406e-06, 'learning_rate/action_model': 2.8188920252562408e-05, 'epoch': 0.26}) +[2026-06-24 13:35:03,499][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.23655381798744202, 'train/grad_norm_pre_clip': 84.40141296386719, 'timing/data': 0.00037839007563889027, 'timing/model': 1.0857237759046257, 'learning_rate/qwen_vl_interface': 2.815605651237723e-06, 'learning_rate/action_model': 2.8156056512377226e-05, 'epoch': 0.26}) +[2026-06-24 13:35:04,587][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.30765217542648315, 'train/grad_norm_pre_clip': 75.74032592773438, 'timing/data': 0.0003520110622048378, 'timing/model': 1.0858801209833473, 'learning_rate/qwen_vl_interface': 2.8123208568712716e-06, 'learning_rate/action_model': 2.8123208568712717e-05, 'epoch': 0.26}) +[2026-06-24 13:35:05,672][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.004813803359866142, 'train/grad_norm_pre_clip': 3.453711748123169, 'timing/data': 0.00035652006044983864, 'timing/model': 1.0834483939688653, 'learning_rate/qwen_vl_interface': 2.8090376442883494e-06, 'learning_rate/action_model': 2.809037644288349e-05, 'epoch': 0.26}) +[2026-06-24 13:35:06,756][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.018930071964859962, 'train/grad_norm_pre_clip': 15.450879096984863, 'timing/data': 0.0003188189584761858, 'timing/model': 1.082397200865671, 'learning_rate/qwen_vl_interface': 2.8057560156193985e-06, 'learning_rate/action_model': 2.8057560156193985e-05, 'epoch': 0.26}) +[2026-06-24 13:35:07,839][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.03001151792705059, 'train/grad_norm_pre_clip': 32.99424743652344, 'timing/data': 0.00031937984749674797, 'timing/model': 1.0814667439553887, 'learning_rate/qwen_vl_interface': 2.8024759729938343e-06, 'learning_rate/action_model': 2.8024759729938344e-05, 'epoch': 0.26}) +[2026-06-24 13:35:08,925][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.1729549914598465, 'train/grad_norm_pre_clip': 75.44190216064453, 'timing/data': 0.00047997222281992435, 'timing/model': 1.0833904000464827, 'learning_rate/qwen_vl_interface': 2.7991975185400368e-06, 'learning_rate/action_model': 2.7991975185400364e-05, 'epoch': 0.26}) +[2026-06-24 13:35:10,012][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.03857485204935074, 'train/grad_norm_pre_clip': 58.35535430908203, 'timing/data': 0.0003455390688031912, 'timing/model': 1.0853234031237662, 'learning_rate/qwen_vl_interface': 2.7959206543853614e-06, 'learning_rate/action_model': 2.7959206543853612e-05, 'epoch': 0.26}) +[2026-06-24 13:35:11,097][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.07270443439483643, 'train/grad_norm_pre_clip': 37.38809585571289, 'timing/data': 0.00032516103237867355, 'timing/model': 1.084106558933854, 'learning_rate/qwen_vl_interface': 2.79264538265613e-06, 'learning_rate/action_model': 2.79264538265613e-05, 'epoch': 0.26}) +[2026-06-24 13:35:12,182][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.0020816626492887735, 'train/grad_norm_pre_clip': 0.9803509712219238, 'timing/data': 0.00032879807986319065, 'timing/model': 1.0832982109859586, 'learning_rate/qwen_vl_interface': 2.7893717054776327e-06, 'learning_rate/action_model': 2.7893717054776327e-05, 'epoch': 0.26}) +[2026-06-24 13:35:13,268][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.1661674529314041, 'train/grad_norm_pre_clip': 68.6729965209961, 'timing/data': 0.00032081292010843754, 'timing/model': 1.0839816469233483, 'learning_rate/qwen_vl_interface': 2.786099624974119e-06, 'learning_rate/action_model': 2.786099624974119e-05, 'epoch': 0.26}) +[2026-06-24 13:35:14,353][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.08701255172491074, 'train/grad_norm_pre_clip': 48.83576202392578, 'timing/data': 0.00032380898483097553, 'timing/model': 1.0834814589470625, 'learning_rate/qwen_vl_interface': 2.7828291432688082e-06, 'learning_rate/action_model': 2.782829143268808e-05, 'epoch': 0.26}) +[2026-06-24 13:35:15,441][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.1156284511089325, 'train/grad_norm_pre_clip': 59.21151351928711, 'timing/data': 0.00033147307112812996, 'timing/model': 1.0865494939498603, 'learning_rate/qwen_vl_interface': 2.779560262483883e-06, 'learning_rate/action_model': 2.7795602624838828e-05, 'epoch': 0.26}) +[2026-06-24 13:35:16,533][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.02165962941944599, 'train/grad_norm_pre_clip': 13.739795684814453, 'timing/data': 0.00032057217322289944, 'timing/model': 1.0902970770839602, 'learning_rate/qwen_vl_interface': 2.7762929847404783e-06, 'learning_rate/action_model': 2.7762929847404785e-05, 'epoch': 0.26}) +[2026-06-24 13:35:17,620][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.06044360250234604, 'train/grad_norm_pre_clip': 55.16236877441406, 'timing/data': 0.00041002104990184307, 'timing/model': 1.0854111390653998, 'learning_rate/qwen_vl_interface': 2.773027312158703e-06, 'learning_rate/action_model': 2.773027312158703e-05, 'epoch': 0.26}) +[2026-06-24 13:35:18,904][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.17824366688728333, 'train/grad_norm_pre_clip': 64.96198272705078, 'timing/data': 0.0004576900973916054, 'timing/model': 1.2821256390307099, 'learning_rate/qwen_vl_interface': 2.7697632468576124e-06, 'learning_rate/action_model': 2.7697632468576124e-05, 'epoch': 0.26}) +[2026-06-24 13:35:19,989][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.02873421646654606, 'train/grad_norm_pre_clip': 65.38687133789062, 'timing/data': 0.0003484750632196665, 'timing/model': 1.0829192399978638, 'learning_rate/qwen_vl_interface': 2.76650079095522e-06, 'learning_rate/action_model': 2.7665007909552198e-05, 'epoch': 0.26}) +[2026-06-24 13:35:21,073][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.17476840317249298, 'train/grad_norm_pre_clip': 84.45716857910156, 'timing/data': 0.0004189671017229557, 'timing/model': 1.082597856875509, 'learning_rate/qwen_vl_interface': 2.763239946568505e-06, 'learning_rate/action_model': 2.763239946568505e-05, 'epoch': 0.26}) +[2026-06-24 13:35:22,159][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.4249044358730316, 'train/grad_norm_pre_clip': 66.3240737915039, 'timing/data': 0.0006037140265107155, 'timing/model': 1.0839152447879314, 'learning_rate/qwen_vl_interface': 2.7599807158133882e-06, 'learning_rate/action_model': 2.7599807158133884e-05, 'epoch': 0.26}) +[2026-06-24 13:35:23,245][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.32420632243156433, 'train/grad_norm_pre_clip': 39.25185012817383, 'timing/data': 0.00033256388269364834, 'timing/model': 1.0834846559446305, 'learning_rate/qwen_vl_interface': 2.7567231008047518e-06, 'learning_rate/action_model': 2.756723100804752e-05, 'epoch': 0.26}) +[2026-06-24 13:35:24,330][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.21123017370700836, 'train/grad_norm_pre_clip': 55.0355224609375, 'timing/data': 0.0003184590023010969, 'timing/model': 1.0834783259779215, 'learning_rate/qwen_vl_interface': 2.7534671036564283e-06, 'learning_rate/action_model': 2.753467103656428e-05, 'epoch': 0.26}) +[2026-06-24 13:35:25,418][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.11077579855918884, 'train/grad_norm_pre_clip': 63.1540641784668, 'timing/data': 0.00033739395439624786, 'timing/model': 1.0865839230827987, 'learning_rate/qwen_vl_interface': 2.7502127264811932e-06, 'learning_rate/action_model': 2.7502127264811935e-05, 'epoch': 0.26}) +[2026-06-24 13:35:26,509][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.08349316567182541, 'train/grad_norm_pre_clip': 66.597900390625, 'timing/data': 0.00033148215152323246, 'timing/model': 1.0898112880531698, 'learning_rate/qwen_vl_interface': 2.746959971390782e-06, 'learning_rate/action_model': 2.7469599713907817e-05, 'epoch': 0.27}) +[2026-06-24 13:35:27,595][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.26876765489578247, 'train/grad_norm_pre_clip': 90.76882934570312, 'timing/data': 0.0003442170564085245, 'timing/model': 1.0842860881239176, 'learning_rate/qwen_vl_interface': 2.743708840495869e-06, 'learning_rate/action_model': 2.743708840495869e-05, 'epoch': 0.27}) +[2026-06-24 13:35:28,681][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.009505609050393105, 'train/grad_norm_pre_clip': 10.968525886535645, 'timing/data': 0.0003497868310660124, 'timing/model': 1.0844779210165143, 'learning_rate/qwen_vl_interface': 2.740459335906081e-06, 'learning_rate/action_model': 2.740459335906081e-05, 'epoch': 0.27}) +[2026-06-24 13:35:29,769][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.344908744096756, 'train/grad_norm_pre_clip': 114.18224334716797, 'timing/data': 0.0004992480389773846, 'timing/model': 1.0861275519710034, 'learning_rate/qwen_vl_interface': 2.7372114597299827e-06, 'learning_rate/action_model': 2.7372114597299825e-05, 'epoch': 0.27}) +[2026-06-24 13:35:30,856][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.18252362310886383, 'train/grad_norm_pre_clip': 58.33161926269531, 'timing/data': 0.00032545183785259724, 'timing/model': 1.0850604579318315, 'learning_rate/qwen_vl_interface': 2.7339652140750876e-06, 'learning_rate/action_model': 2.7339652140750876e-05, 'epoch': 0.27}) +[2026-06-24 13:35:31,943][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.14629970490932465, 'train/grad_norm_pre_clip': 45.69179153442383, 'timing/data': 0.0003396589308977127, 'timing/model': 1.0851199810858816, 'learning_rate/qwen_vl_interface': 2.7307206010478498e-06, 'learning_rate/action_model': 2.7307206010478497e-05, 'epoch': 0.27}) +[2026-06-24 13:35:33,030][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.14328263700008392, 'train/grad_norm_pre_clip': 24.20073127746582, 'timing/data': 0.0003227870911359787, 'timing/model': 1.0851721209473908, 'learning_rate/qwen_vl_interface': 2.7274776227536616e-06, 'learning_rate/action_model': 2.7274776227536614e-05, 'epoch': 0.27}) +[2026-06-24 13:35:34,121][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.13652510941028595, 'train/grad_norm_pre_clip': 46.12154769897461, 'timing/data': 0.00031500100158154964, 'timing/model': 1.089970506960526, 'learning_rate/qwen_vl_interface': 2.7242362812968566e-06, 'learning_rate/action_model': 2.7242362812968563e-05, 'epoch': 0.27}) +[2026-06-24 13:35:35,206][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.0831548273563385, 'train/grad_norm_pre_clip': 28.36402702331543, 'timing/data': 0.00037264986895024776, 'timing/model': 1.0833208251278847, 'learning_rate/qwen_vl_interface': 2.7209965787807073e-06, 'learning_rate/action_model': 2.7209965787807075e-05, 'epoch': 0.27}) +[2026-06-24 13:35:36,492][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.05598772317171097, 'train/grad_norm_pre_clip': 25.442459106445312, 'timing/data': 0.0003393581137061119, 'timing/model': 1.2842068870086223, 'learning_rate/qwen_vl_interface': 2.717758517307423e-06, 'learning_rate/action_model': 2.7177585173074226e-05, 'epoch': 0.27}) +[2026-06-24 13:35:37,584][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.284122109413147, 'train/grad_norm_pre_clip': 91.59944915771484, 'timing/data': 0.00039460184052586555, 'timing/model': 1.0895589790306985, 'learning_rate/qwen_vl_interface': 2.714522098978142e-06, 'learning_rate/action_model': 2.714522098978142e-05, 'epoch': 0.27}) +[2026-06-24 13:35:38,673][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2750 +[2026-06-24 13:36:04,739][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2750 +[2026-06-24 13:36:04,740][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2750 +[2026-06-24 13:36:39,047][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2750 +[2026-06-24 13:43:02,418][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2750 +[2026-06-24 13:43:02,419][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.1226455345749855, 'train/grad_norm_pre_clip': 49.71760559082031, 'eval/action_loss/seconds': 26.064572115195915, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.06350108236074448, 'eval/latency_2/loss': 0.06350108236074448, 'eval/flappy/loss': 0.06350108236074448, 'eval/flappy/latency_2/loss': 0.06350108236074448, 'eval/flap_precision': 0.940026075619296, 'eval/flap_recall': 0.9782903663500678, 'eval/flap_f1': 0.958776595744681, 'eval/flap_rate_model': 0.21305555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9587765957446811, 'eval/latency_2/flap_precision': 0.940026075619296, 'eval/latency_2/flap_recall': 0.9782903663500678, 'eval/latency_2/flap_f1': 0.958776595744681, 'eval/latency_2/flap_rate_model': 0.21305555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9587765957446811, 'eval/action_classification/data_wait_seconds': 0.3964543342590332, 'eval/action_classification/predict_seconds': 33.802154541015625, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 290.720000000015, 'rl_games_eval/mid_train/mean_length': 2360.8, 'rl_games_eval/mid_train/std_reward': 191.76282642891158, 'rl_games_eval/mid_train/std_length': 1544.3919709711004, 'rl_games_eval/mid_train/macro_mean_reward': 290.720000000015, 'rl_games_eval/mid_train/macro_mean_length': 2360.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 290.720000000015, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2360.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 191.76282642891158, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1544.3919709711004, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 290.720000000015, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00035956595093011856, 'timing/model': 1.087816937128082, 'learning_rate/qwen_vl_interface': 2.7112873258929445e-06, 'learning_rate/action_model': 2.7112873258929443e-05, 'epoch': 0.27}) +[2026-06-24 13:43:03,489][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.10895199328660965, 'train/grad_norm_pre_clip': 43.846580505371094, 'timing/data': 0.000406614039093256, 'timing/model': 1.0689470020588487, 'learning_rate/qwen_vl_interface': 2.7080542001508407e-06, 'learning_rate/action_model': 2.7080542001508407e-05, 'epoch': 0.27}) +[2026-06-24 13:43:04,563][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.04085656255483627, 'train/grad_norm_pre_clip': 38.535972595214844, 'timing/data': 0.000361358979716897, 'timing/model': 1.0713822080288082, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 0.27}) +[2026-06-24 13:43:05,630][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.031231524422764778, 'train/grad_norm_pre_clip': 31.666196823120117, 'timing/data': 0.0003850331995636225, 'timing/model': 1.0655910959467292, 'learning_rate/qwen_vl_interface': 2.7015928990865954e-06, 'learning_rate/action_model': 2.7015928990865952e-05, 'epoch': 0.27}) +[2026-06-24 13:43:06,700][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.24578790366649628, 'train/grad_norm_pre_clip': 90.1358642578125, 'timing/data': 0.00034571997821331024, 'timing/model': 1.0684076100587845, 'learning_rate/qwen_vl_interface': 2.698364727957125e-06, 'learning_rate/action_model': 2.698364727957125e-05, 'epoch': 0.27}) +[2026-06-24 13:43:07,775][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.04086093604564667, 'train/grad_norm_pre_clip': 37.21894836425781, 'timing/data': 0.0003362919669598341, 'timing/model': 1.0733722699806094, 'learning_rate/qwen_vl_interface': 2.6951382125560767e-06, 'learning_rate/action_model': 2.6951382125560764e-05, 'epoch': 0.27}) +[2026-06-24 13:43:09,124][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.003838286269456148, 'train/grad_norm_pre_clip': 3.600362539291382, 'timing/data': 0.000386606901884079, 'timing/model': 1.3477883168961853, 'learning_rate/qwen_vl_interface': 2.6919133549771075e-06, 'learning_rate/action_model': 2.6919133549771076e-05, 'epoch': 0.27}) +[2026-06-24 13:43:10,198][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.18375088274478912, 'train/grad_norm_pre_clip': 53.31928634643555, 'timing/data': 0.0003663280513137579, 'timing/model': 1.0721573079936206, 'learning_rate/qwen_vl_interface': 2.688690157312789e-06, 'learning_rate/action_model': 2.6886901573127887e-05, 'epoch': 0.27}) +[2026-06-24 13:43:11,272][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.012016049586236477, 'train/grad_norm_pre_clip': 17.440549850463867, 'timing/data': 0.000323849031701684, 'timing/model': 1.0717919529415667, 'learning_rate/qwen_vl_interface': 2.6854686216546188e-06, 'learning_rate/action_model': 2.6854686216546186e-05, 'epoch': 0.27}) +[2026-06-24 13:43:12,345][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.055908240377902985, 'train/grad_norm_pre_clip': 24.598419189453125, 'timing/data': 0.0003485949710011482, 'timing/model': 1.0717516178265214, 'learning_rate/qwen_vl_interface': 2.6822487500930194e-06, 'learning_rate/action_model': 2.6822487500930193e-05, 'epoch': 0.27}) +[2026-06-24 13:43:13,422][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.029490021988749504, 'train/grad_norm_pre_clip': 20.494251251220703, 'timing/data': 0.0003834310919046402, 'timing/model': 1.0755295569542795, 'learning_rate/qwen_vl_interface': 2.6790305447173263e-06, 'learning_rate/action_model': 2.6790305447173264e-05, 'epoch': 0.27}) +[2026-06-24 13:43:14,498][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.0199888963252306, 'train/grad_norm_pre_clip': 17.3652400970459, 'timing/data': 0.00032621296122670174, 'timing/model': 1.074559744913131, 'learning_rate/qwen_vl_interface': 2.6758140076158005e-06, 'learning_rate/action_model': 2.6758140076158006e-05, 'epoch': 0.27}) +[2026-06-24 13:43:15,573][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.0825774073600769, 'train/grad_norm_pre_clip': 28.36623764038086, 'timing/data': 0.00031481217592954636, 'timing/model': 1.073407790856436, 'learning_rate/qwen_vl_interface': 2.6725991408756177e-06, 'learning_rate/action_model': 2.6725991408756172e-05, 'epoch': 0.27}) +[2026-06-24 13:43:16,648][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.17434924840927124, 'train/grad_norm_pre_clip': 61.64453887939453, 'timing/data': 0.00037564593367278576, 'timing/model': 1.0727662059944123, 'learning_rate/qwen_vl_interface': 2.6693859465828716e-06, 'learning_rate/action_model': 2.6693859465828713e-05, 'epoch': 0.27}) +[2026-06-24 13:43:17,721][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.12255743145942688, 'train/grad_norm_pre_clip': 64.33680725097656, 'timing/data': 0.00035047787241637707, 'timing/model': 1.0713123641908169, 'learning_rate/qwen_vl_interface': 2.666174426822565e-06, 'learning_rate/action_model': 2.6661744268225647e-05, 'epoch': 0.27}) +[2026-06-24 13:43:18,795][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.062459152191877365, 'train/grad_norm_pre_clip': 40.51240539550781, 'timing/data': 0.0003480040468275547, 'timing/model': 1.0722342380322516, 'learning_rate/qwen_vl_interface': 2.6629645836786217e-06, 'learning_rate/action_model': 2.6629645836786217e-05, 'epoch': 0.27}) +[2026-06-24 13:43:19,869][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.11266801506280899, 'train/grad_norm_pre_clip': 53.309913635253906, 'timing/data': 0.00036586704663932323, 'timing/model': 1.0722368329297751, 'learning_rate/qwen_vl_interface': 2.6597564192338738e-06, 'learning_rate/action_model': 2.6597564192338735e-05, 'epoch': 0.27}) +[2026-06-24 13:43:20,948][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.37726905941963196, 'train/grad_norm_pre_clip': 70.27263641357422, 'timing/data': 0.0003505588974803686, 'timing/model': 1.0771427031140774, 'learning_rate/qwen_vl_interface': 2.656549935570062e-06, 'learning_rate/action_model': 2.656549935570062e-05, 'epoch': 0.27}) +[2026-06-24 13:43:22,025][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.012363606132566929, 'train/grad_norm_pre_clip': 8.820948600769043, 'timing/data': 0.00032574194483458996, 'timing/model': 1.0757020669989288, 'learning_rate/qwen_vl_interface': 2.6533451347678406e-06, 'learning_rate/action_model': 2.6533451347678407e-05, 'epoch': 0.27}) +[2026-06-24 13:43:23,102][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.17265884578227997, 'train/grad_norm_pre_clip': 49.61845016479492, 'timing/data': 0.0003600670024752617, 'timing/model': 1.075305983191356, 'learning_rate/qwen_vl_interface': 2.6501420189067716e-06, 'learning_rate/action_model': 2.6501420189067716e-05, 'epoch': 0.27}) +[2026-06-24 13:43:24,176][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.06573375314474106, 'train/grad_norm_pre_clip': 31.96891212463379, 'timing/data': 0.00033125211484730244, 'timing/model': 1.07244742102921, 'learning_rate/qwen_vl_interface': 2.646940590065319e-06, 'learning_rate/action_model': 2.6469405900653187e-05, 'epoch': 0.27}) +[2026-06-24 13:43:25,256][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.0623060017824173, 'train/grad_norm_pre_clip': 57.684818267822266, 'timing/data': 0.0003219249192625284, 'timing/model': 1.0779109590221196, 'learning_rate/qwen_vl_interface': 2.6437408503208556e-06, 'learning_rate/action_model': 2.6437408503208556e-05, 'epoch': 0.27}) +[2026-06-24 13:43:26,611][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.159042090177536, 'train/grad_norm_pre_clip': 65.28240966796875, 'timing/data': 0.00032302713952958584, 'timing/model': 1.3531824848614633, 'learning_rate/qwen_vl_interface': 2.6405428017496594e-06, 'learning_rate/action_model': 2.6405428017496593e-05, 'epoch': 0.27}) +[2026-06-24 13:43:27,691][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.11019086837768555, 'train/grad_norm_pre_clip': 60.08560562133789, 'timing/data': 0.0005182139575481415, 'timing/model': 1.078374942066148, 'learning_rate/qwen_vl_interface': 2.6373464464269095e-06, 'learning_rate/action_model': 2.6373464464269095e-05, 'epoch': 0.27}) +[2026-06-24 13:43:28,770][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.13863590359687805, 'train/grad_norm_pre_clip': 25.331113815307617, 'timing/data': 0.00031627388671040535, 'timing/model': 1.0776997648645192, 'learning_rate/qwen_vl_interface': 2.6341517864266826e-06, 'learning_rate/action_model': 2.6341517864266824e-05, 'epoch': 0.27}) +[2026-06-24 13:43:29,848][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.02037023939192295, 'train/grad_norm_pre_clip': 12.982906341552734, 'timing/data': 0.0003442468587309122, 'timing/model': 1.0758430249989033, 'learning_rate/qwen_vl_interface': 2.6309588238219647e-06, 'learning_rate/action_model': 2.6309588238219644e-05, 'epoch': 0.27}) +[2026-06-24 13:43:30,926][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.00863930955529213, 'train/grad_norm_pre_clip': 4.0892014503479, 'timing/data': 0.00036655901931226254, 'timing/model': 1.0765573591925204, 'learning_rate/qwen_vl_interface': 2.627767560684632e-06, 'learning_rate/action_model': 2.627767560684632e-05, 'epoch': 0.27}) +[2026-06-24 13:43:32,004][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.15842266380786896, 'train/grad_norm_pre_clip': 30.59050941467285, 'timing/data': 0.00031259702518582344, 'timing/model': 1.0761647010222077, 'learning_rate/qwen_vl_interface': 2.624577999085456e-06, 'learning_rate/action_model': 2.624577999085456e-05, 'epoch': 0.27}) +[2026-06-24 13:43:33,083][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.03752609342336655, 'train/grad_norm_pre_clip': 12.929679870605469, 'timing/data': 0.00031301891431212425, 'timing/model': 1.0773694759700447, 'learning_rate/qwen_vl_interface': 2.621390141094116e-06, 'learning_rate/action_model': 2.621390141094116e-05, 'epoch': 0.27}) +[2026-06-24 13:43:34,163][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.06749438494443893, 'train/grad_norm_pre_clip': 29.40241241455078, 'timing/data': 0.00033058086410164833, 'timing/model': 1.0784422929864377, 'learning_rate/qwen_vl_interface': 2.618203988779174e-06, 'learning_rate/action_model': 2.6182039887791737e-05, 'epoch': 0.27}) +[2026-06-24 13:43:35,242][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.01611820049583912, 'train/grad_norm_pre_clip': 26.00726318359375, 'timing/data': 0.00031950115226209164, 'timing/model': 1.0772690379526466, 'learning_rate/qwen_vl_interface': 2.6150195442080907e-06, 'learning_rate/action_model': 2.6150195442080904e-05, 'epoch': 0.27}) +[2026-06-24 13:43:36,324][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.01021693553775549, 'train/grad_norm_pre_clip': 9.360584259033203, 'timing/data': 0.0003150920383632183, 'timing/model': 1.0811976320110261, 'learning_rate/qwen_vl_interface': 2.6118368094472187e-06, 'learning_rate/action_model': 2.6118368094472183e-05, 'epoch': 0.27}) +[2026-06-24 13:43:37,403][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.13505803048610687, 'train/grad_norm_pre_clip': 46.86835479736328, 'timing/data': 0.0003677110653370619, 'timing/model': 1.0771689109969884, 'learning_rate/qwen_vl_interface': 2.608655786561797e-06, 'learning_rate/action_model': 2.6086557865617973e-05, 'epoch': 0.27}) +[2026-06-24 13:43:38,485][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.09128045290708542, 'train/grad_norm_pre_clip': 43.5915641784668, 'timing/data': 0.0003859249409288168, 'timing/model': 1.0801917021162808, 'learning_rate/qwen_vl_interface': 2.605476477615959e-06, 'learning_rate/action_model': 2.605476477615959e-05, 'epoch': 0.27}) +[2026-06-24 13:43:39,566][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.017962221056222916, 'train/grad_norm_pre_clip': 12.52993392944336, 'timing/data': 0.00033514993265271187, 'timing/model': 1.0792985849548131, 'learning_rate/qwen_vl_interface': 2.6022988846727235e-06, 'learning_rate/action_model': 2.6022988846727238e-05, 'epoch': 0.27}) +[2026-06-24 13:43:40,646][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.10615470260381699, 'train/grad_norm_pre_clip': 55.58464431762695, 'timing/data': 0.00032381806522607803, 'timing/model': 1.0789515329524875, 'learning_rate/qwen_vl_interface': 2.5991230097939983e-06, 'learning_rate/action_model': 2.5991230097939985e-05, 'epoch': 0.27}) +[2026-06-24 13:43:41,725][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.08928927779197693, 'train/grad_norm_pre_clip': 41.19858169555664, 'timing/data': 0.00039606401696801186, 'timing/model': 1.0771647659130394, 'learning_rate/qwen_vl_interface': 2.5959488550405698e-06, 'learning_rate/action_model': 2.5959488550405697e-05, 'epoch': 0.27}) +[2026-06-24 13:43:42,807][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.013025755062699318, 'train/grad_norm_pre_clip': 8.56630802154541, 'timing/data': 0.0003199509810656309, 'timing/model': 1.0804333488922566, 'learning_rate/qwen_vl_interface': 2.5927764224721148e-06, 'learning_rate/action_model': 2.5927764224721145e-05, 'epoch': 0.27}) +[2026-06-24 13:43:44,162][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.13907071948051453, 'train/grad_norm_pre_clip': 23.402254104614258, 'timing/data': 0.0003486950881779194, 'timing/model': 1.353222772013396, 'learning_rate/qwen_vl_interface': 2.5896057141471914e-06, 'learning_rate/action_model': 2.5896057141471914e-05, 'epoch': 0.27}) +[2026-06-24 13:43:45,242][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.04740947484970093, 'train/grad_norm_pre_clip': 16.38545036315918, 'timing/data': 0.0003310220781713724, 'timing/model': 1.078143045073375, 'learning_rate/qwen_vl_interface': 2.5864367321232354e-06, 'learning_rate/action_model': 2.5864367321232352e-05, 'epoch': 0.27}) +[2026-06-24 13:43:46,321][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.07320915907621384, 'train/grad_norm_pre_clip': 21.993526458740234, 'timing/data': 0.0003555279690772295, 'timing/model': 1.0777561289723963, 'learning_rate/qwen_vl_interface': 2.583269478456566e-06, 'learning_rate/action_model': 2.5832694784565658e-05, 'epoch': 0.27}) +[2026-06-24 13:43:47,405][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.029307076707482338, 'train/grad_norm_pre_clip': 14.333590507507324, 'timing/data': 0.0003490559756755829, 'timing/model': 1.0821841310244054, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 0.27}) +[2026-06-24 13:43:48,487][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.01878947950899601, 'train/grad_norm_pre_clip': 7.7799859046936035, 'timing/data': 0.00033059087581932545, 'timing/model': 1.079883418045938, 'learning_rate/qwen_vl_interface': 2.5769401644147516e-06, 'learning_rate/action_model': 2.5769401644147516e-05, 'epoch': 0.27}) +[2026-06-24 13:43:49,568][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.2801496684551239, 'train/grad_norm_pre_clip': 61.31205368041992, 'timing/data': 0.0003286069259047508, 'timing/model': 1.0793675689492375, 'learning_rate/qwen_vl_interface': 2.5737781081466286e-06, 'learning_rate/action_model': 2.5737781081466285e-05, 'epoch': 0.27}) +[2026-06-24 13:43:50,648][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.00711976969614625, 'train/grad_norm_pre_clip': 5.259516716003418, 'timing/data': 0.00035053794272243977, 'timing/model': 1.079053338849917, 'learning_rate/qwen_vl_interface': 2.5706177884498352e-06, 'learning_rate/action_model': 2.570617788449835e-05, 'epoch': 0.27}) +[2026-06-24 13:43:51,728][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.021487202495336533, 'train/grad_norm_pre_clip': 17.484359741210938, 'timing/data': 0.0003245000261813402, 'timing/model': 1.0783400698564947, 'learning_rate/qwen_vl_interface': 2.5674592073750716e-06, 'learning_rate/action_model': 2.5674592073750713e-05, 'epoch': 0.27}) +[2026-06-24 13:43:52,809][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.09634232521057129, 'train/grad_norm_pre_clip': 36.90467071533203, 'timing/data': 0.00032459013164043427, 'timing/model': 1.0793501280713826, 'learning_rate/qwen_vl_interface': 2.564302366971901e-06, 'learning_rate/action_model': 2.564302366971901e-05, 'epoch': 0.27}) +[2026-06-24 13:43:53,892][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.1887694150209427, 'train/grad_norm_pre_clip': 85.95646667480469, 'timing/data': 0.00031976099126040936, 'timing/model': 1.080942772096023, 'learning_rate/qwen_vl_interface': 2.561147269288769e-06, 'learning_rate/action_model': 2.561147269288769e-05, 'epoch': 0.27}) +[2026-06-24 13:43:54,975][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.055442940443754196, 'train/grad_norm_pre_clip': 32.232357025146484, 'timing/data': 0.00037169898860156536, 'timing/model': 1.0810690498910844, 'learning_rate/qwen_vl_interface': 2.5579939163729823e-06, 'learning_rate/action_model': 2.557993916372982e-05, 'epoch': 0.27}) +[2026-06-24 13:43:56,056][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.10833318531513214, 'train/grad_norm_pre_clip': 58.27691650390625, 'timing/data': 0.0003864059690386057, 'timing/model': 1.0802935319952667, 'learning_rate/qwen_vl_interface': 2.5548423102707126e-06, 'learning_rate/action_model': 2.5548423102707126e-05, 'epoch': 0.27}) +[2026-06-24 13:43:57,138][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.07060157507658005, 'train/grad_norm_pre_clip': 91.82774353027344, 'timing/data': 0.000361097976565361, 'timing/model': 1.0798571130726486, 'learning_rate/qwen_vl_interface': 2.5516924530270103e-06, 'learning_rate/action_model': 2.5516924530270105e-05, 'epoch': 0.27}) +[2026-06-24 13:43:58,220][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.06005490571260452, 'train/grad_norm_pre_clip': 23.093490600585938, 'timing/data': 0.0003154631704092026, 'timing/model': 1.0805058220867068, 'learning_rate/qwen_vl_interface': 2.548544346685779e-06, 'learning_rate/action_model': 2.548544346685779e-05, 'epoch': 0.27}) +[2026-06-24 13:43:59,300][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.04068663716316223, 'train/grad_norm_pre_clip': 25.762392044067383, 'timing/data': 0.0003468310460448265, 'timing/model': 1.0786968930624425, 'learning_rate/qwen_vl_interface': 2.5453979932897916e-06, 'learning_rate/action_model': 2.5453979932897914e-05, 'epoch': 0.27}) +[2026-06-24 13:44:00,385][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.0202324315905571, 'train/grad_norm_pre_clip': 15.753372192382812, 'timing/data': 0.0003200620412826538, 'timing/model': 1.0825291259679943, 'learning_rate/qwen_vl_interface': 2.542253394880684e-06, 'learning_rate/action_model': 2.5422533948806842e-05, 'epoch': 0.27}) +[2026-06-24 13:44:01,741][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.00932338833808899, 'train/grad_norm_pre_clip': 7.31089448928833, 'timing/data': 0.0003296690993010998, 'timing/model': 1.3551106008235365, 'learning_rate/qwen_vl_interface': 2.5391105534989503e-06, 'learning_rate/action_model': 2.53911055349895e-05, 'epoch': 0.27}) +[2026-06-24 13:44:02,827][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.043520454317331314, 'train/grad_norm_pre_clip': 26.761152267456055, 'timing/data': 0.00041226414032280445, 'timing/model': 1.0836618659086525, 'learning_rate/qwen_vl_interface': 2.5359694711839453e-06, 'learning_rate/action_model': 2.5359694711839456e-05, 'epoch': 0.27}) +[2026-06-24 13:44:03,911][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.012787400744855404, 'train/grad_norm_pre_clip': 10.8937406539917, 'timing/data': 0.0005052001215517521, 'timing/model': 1.082404884044081, 'learning_rate/qwen_vl_interface': 2.532830149973886e-06, 'learning_rate/action_model': 2.532830149973886e-05, 'epoch': 0.27}) +[2026-06-24 13:44:04,991][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.01988961547613144, 'train/grad_norm_pre_clip': 10.862430572509766, 'timing/data': 0.00030999304726719856, 'timing/model': 1.07856031996198, 'learning_rate/qwen_vl_interface': 2.5296925919058433e-06, 'learning_rate/action_model': 2.5296925919058436e-05, 'epoch': 0.27}) +[2026-06-24 13:44:06,072][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.1309226155281067, 'train/grad_norm_pre_clip': 37.308143615722656, 'timing/data': 0.00032527116127312183, 'timing/model': 1.0791711679194123, 'learning_rate/qwen_vl_interface': 2.526556799015743e-06, 'learning_rate/action_model': 2.526556799015743e-05, 'epoch': 0.27}) +[2026-06-24 13:44:07,153][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.04273854196071625, 'train/grad_norm_pre_clip': 31.185258865356445, 'timing/data': 0.00034684198908507824, 'timing/model': 1.0796608191449195, 'learning_rate/qwen_vl_interface': 2.523422773338367e-06, 'learning_rate/action_model': 2.523422773338367e-05, 'epoch': 0.27}) +[2026-06-24 13:44:08,233][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.15963339805603027, 'train/grad_norm_pre_clip': 90.02172088623047, 'timing/data': 0.0003248110879212618, 'timing/model': 1.078392866998911, 'learning_rate/qwen_vl_interface': 2.5202905169073533e-06, 'learning_rate/action_model': 2.5202905169073532e-05, 'epoch': 0.27}) +[2026-06-24 13:44:09,321][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.06177355349063873, 'train/grad_norm_pre_clip': 32.883949279785156, 'timing/data': 0.00039248703978955746, 'timing/model': 1.0856877260375768, 'learning_rate/qwen_vl_interface': 2.517160031755185e-06, 'learning_rate/action_model': 2.5171600317551854e-05, 'epoch': 0.27}) +[2026-06-24 13:44:10,404][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.08180404454469681, 'train/grad_norm_pre_clip': 39.310733795166016, 'timing/data': 0.00047448184341192245, 'timing/model': 1.0815867099445313, 'learning_rate/qwen_vl_interface': 2.5140313199132025e-06, 'learning_rate/action_model': 2.5140313199132028e-05, 'epoch': 0.27}) +[2026-06-24 13:44:11,489][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.0780937597155571, 'train/grad_norm_pre_clip': 34.32902526855469, 'timing/data': 0.00033867591992020607, 'timing/model': 1.0833573781419545, 'learning_rate/qwen_vl_interface': 2.5109043834115948e-06, 'learning_rate/action_model': 2.5109043834115946e-05, 'epoch': 0.27}) +[2026-06-24 13:44:12,574][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.0030983551405370235, 'train/grad_norm_pre_clip': 3.561692714691162, 'timing/data': 0.00031960103660821915, 'timing/model': 1.0831236590165645, 'learning_rate/qwen_vl_interface': 2.507779224279392e-06, 'learning_rate/action_model': 2.5077792242793917e-05, 'epoch': 0.27}) +[2026-06-24 13:44:13,656][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.02957874909043312, 'train/grad_norm_pre_clip': 25.712472915649414, 'timing/data': 0.000326584093272686, 'timing/model': 1.0802304628305137, 'learning_rate/qwen_vl_interface': 2.504655844544479e-06, 'learning_rate/action_model': 2.504655844544479e-05, 'epoch': 0.27}) +[2026-06-24 13:44:14,741][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.13175329566001892, 'train/grad_norm_pre_clip': 46.95471954345703, 'timing/data': 0.00032937899231910706, 'timing/model': 1.0832448478322476, 'learning_rate/qwen_vl_interface': 2.501534246233583e-06, 'learning_rate/action_model': 2.5015342462335827e-05, 'epoch': 0.27}) +[2026-06-24 13:44:15,827][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.049503911286592484, 'train/grad_norm_pre_clip': 12.11352252960205, 'timing/data': 0.0003055538982152939, 'timing/model': 1.0847533938940614, 'learning_rate/qwen_vl_interface': 2.4984144313722765e-06, 'learning_rate/action_model': 2.4984144313722766e-05, 'epoch': 0.27}) +[2026-06-24 13:44:16,910][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.02132808417081833, 'train/grad_norm_pre_clip': 16.17890167236328, 'timing/data': 0.000353414099663496, 'timing/model': 1.081465075025335, 'learning_rate/qwen_vl_interface': 2.495296401984969e-06, 'learning_rate/action_model': 2.495296401984969e-05, 'epoch': 0.27}) +[2026-06-24 13:44:17,998][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.07040517032146454, 'train/grad_norm_pre_clip': 38.51933288574219, 'timing/data': 0.0003128279931843281, 'timing/model': 1.085966665064916, 'learning_rate/qwen_vl_interface': 2.4921801600949215e-06, 'learning_rate/action_model': 2.4921801600949217e-05, 'epoch': 0.27}) +[2026-06-24 13:44:19,354][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.002673415932804346, 'train/grad_norm_pre_clip': 4.126719951629639, 'timing/data': 0.0003389872144907713, 'timing/model': 1.3550223258789629, 'learning_rate/qwen_vl_interface': 2.4890657077242274e-06, 'learning_rate/action_model': 2.4890657077242274e-05, 'epoch': 0.27}) +[2026-06-24 13:44:20,445][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.013603931292891502, 'train/grad_norm_pre_clip': 10.93165111541748, 'timing/data': 0.0003889109939336777, 'timing/model': 1.089044569991529, 'learning_rate/qwen_vl_interface': 2.485953046893817e-06, 'learning_rate/action_model': 2.4859530468938165e-05, 'epoch': 0.27}) +[2026-06-24 13:44:21,530][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.08008871972560883, 'train/grad_norm_pre_clip': 25.016841888427734, 'timing/data': 0.00033929687924683094, 'timing/model': 1.083162385970354, 'learning_rate/qwen_vl_interface': 2.482842179623469e-06, 'learning_rate/action_model': 2.4828421796234693e-05, 'epoch': 0.27}) +[2026-06-24 13:44:22,610][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.11581077426671982, 'train/grad_norm_pre_clip': 50.08555603027344, 'timing/data': 0.00031575304456055164, 'timing/model': 1.0786557879764587, 'learning_rate/qwen_vl_interface': 2.4797331079317867e-06, 'learning_rate/action_model': 2.4797331079317865e-05, 'epoch': 0.27}) +[2026-06-24 13:44:23,696][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.010186663828790188, 'train/grad_norm_pre_clip': 25.548593521118164, 'timing/data': 0.0003553580027073622, 'timing/model': 1.0840284649748355, 'learning_rate/qwen_vl_interface': 2.4766258338362135e-06, 'learning_rate/action_model': 2.4766258338362136e-05, 'epoch': 0.27}) +[2026-06-24 13:44:24,781][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.08922641724348068, 'train/grad_norm_pre_clip': 53.11130905151367, 'timing/data': 0.00033205398358404636, 'timing/model': 1.083006695844233, 'learning_rate/qwen_vl_interface': 2.473520359353029e-06, 'learning_rate/action_model': 2.473520359353029e-05, 'epoch': 0.27}) +[2026-06-24 13:44:25,869][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.081952303647995, 'train/grad_norm_pre_clip': 61.76877212524414, 'timing/data': 0.00032425904646515846, 'timing/model': 1.0865298870485276, 'learning_rate/qwen_vl_interface': 2.4704166864973374e-06, 'learning_rate/action_model': 2.4704166864973373e-05, 'epoch': 0.27}) +[2026-06-24 13:44:26,955][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.0011920849792659283, 'train/grad_norm_pre_clip': 1.1324045658111572, 'timing/data': 0.00033729407005012035, 'timing/model': 1.0849923880305141, 'learning_rate/qwen_vl_interface': 2.467314817283081e-06, 'learning_rate/action_model': 2.467314817283081e-05, 'epoch': 0.27}) +[2026-06-24 13:44:28,040][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.01877930387854576, 'train/grad_norm_pre_clip': 14.896870613098145, 'timing/data': 0.0003457500133663416, 'timing/model': 1.0828233719803393, 'learning_rate/qwen_vl_interface': 2.464214753723028e-06, 'learning_rate/action_model': 2.464214753723028e-05, 'epoch': 0.27}) +[2026-06-24 13:44:29,124][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.14440329372882843, 'train/grad_norm_pre_clip': 84.66897583007812, 'timing/data': 0.00042827497236430645, 'timing/model': 1.0821148809045553, 'learning_rate/qwen_vl_interface': 2.461116497828778e-06, 'learning_rate/action_model': 2.461116497828778e-05, 'epoch': 0.27}) +[2026-06-24 13:44:30,210][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.09660730510950089, 'train/grad_norm_pre_clip': 71.07886505126953, 'timing/data': 0.00033110217191278934, 'timing/model': 1.084427336929366, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 0.27}) +[2026-06-24 13:44:31,296][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.031836964190006256, 'train/grad_norm_pre_clip': 11.556633949279785, 'timing/data': 0.0003346088342368603, 'timing/model': 1.0841856128536165, 'learning_rate/qwen_vl_interface': 2.4549254170782046e-06, 'learning_rate/action_model': 2.4549254170782042e-05, 'epoch': 0.27}) +[2026-06-24 13:44:32,381][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.06044350564479828, 'train/grad_norm_pre_clip': 38.40960693359375, 'timing/data': 0.00033366703428328037, 'timing/model': 1.0837469198741019, 'learning_rate/qwen_vl_interface': 2.4518325962392077e-06, 'learning_rate/action_model': 2.4518325962392076e-05, 'epoch': 0.27}) +[2026-06-24 13:44:33,468][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.001993217272683978, 'train/grad_norm_pre_clip': 3.3125433921813965, 'timing/data': 0.00031449110247194767, 'timing/model': 1.0848675260785967, 'learning_rate/qwen_vl_interface': 2.4487415911006594e-06, 'learning_rate/action_model': 2.4487415911006594e-05, 'epoch': 0.27}) +[2026-06-24 13:44:34,553][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.08510108292102814, 'train/grad_norm_pre_clip': 64.11772155761719, 'timing/data': 0.00030842889100313187, 'timing/model': 1.0839522369205952, 'learning_rate/qwen_vl_interface': 2.4456524036682793e-06, 'learning_rate/action_model': 2.445652403668279e-05, 'epoch': 0.27}) +[2026-06-24 13:44:35,642][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.06376770883798599, 'train/grad_norm_pre_clip': 64.85177612304688, 'timing/data': 0.0003203819505870342, 'timing/model': 1.0869602200109512, 'learning_rate/qwen_vl_interface': 2.44256503594661e-06, 'learning_rate/action_model': 2.44256503594661e-05, 'epoch': 0.27}) +[2026-06-24 13:44:36,988][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.1375654935836792, 'train/grad_norm_pre_clip': 63.57781219482422, 'timing/data': 0.0003155230078846216, 'timing/model': 1.3443803370464593, 'learning_rate/qwen_vl_interface': 2.4394794899390073e-06, 'learning_rate/action_model': 2.4394794899390072e-05, 'epoch': 0.27}) +[2026-06-24 13:44:38,070][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.4311846196651459, 'train/grad_norm_pre_clip': 117.61058807373047, 'timing/data': 0.00033631199039518833, 'timing/model': 1.0802074610255659, 'learning_rate/qwen_vl_interface': 2.4363957676476504e-06, 'learning_rate/action_model': 2.4363957676476502e-05, 'epoch': 0.27}) +[2026-06-24 13:44:39,155][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.005889580585062504, 'train/grad_norm_pre_clip': 12.785928726196289, 'timing/data': 0.00031464104540646076, 'timing/model': 1.0838957522064447, 'learning_rate/qwen_vl_interface': 2.4333138710735323e-06, 'learning_rate/action_model': 2.4333138710735324e-05, 'epoch': 0.27}) +[2026-06-24 13:44:40,238][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.12955541908740997, 'train/grad_norm_pre_clip': 71.2342300415039, 'timing/data': 0.0003444179892539978, 'timing/model': 1.081105720018968, 'learning_rate/qwen_vl_interface': 2.430233802216465e-06, 'learning_rate/action_model': 2.4302338022164647e-05, 'epoch': 0.27}) +[2026-06-24 13:44:41,326][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.13931366801261902, 'train/grad_norm_pre_clip': 45.7157096862793, 'timing/data': 0.00034666084684431553, 'timing/model': 1.0865663629956543, 'learning_rate/qwen_vl_interface': 2.427155563075068e-06, 'learning_rate/action_model': 2.4271555630750677e-05, 'epoch': 0.27}) +[2026-06-24 13:44:42,410][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.1574878692626953, 'train/grad_norm_pre_clip': 91.71842193603516, 'timing/data': 0.00034396699629724026, 'timing/model': 1.0822634068317711, 'learning_rate/qwen_vl_interface': 2.4240791556467788e-06, 'learning_rate/action_model': 2.4240791556467787e-05, 'epoch': 0.27}) +[2026-06-24 13:44:43,498][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.03869110718369484, 'train/grad_norm_pre_clip': 26.906789779663086, 'timing/data': 0.00031981104984879494, 'timing/model': 1.0862243010196835, 'learning_rate/qwen_vl_interface': 2.4210045819278456e-06, 'learning_rate/action_model': 2.4210045819278453e-05, 'epoch': 0.27}) +[2026-06-24 13:44:44,588][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.34333738684654236, 'train/grad_norm_pre_clip': 90.02864837646484, 'timing/data': 0.0003478638827800751, 'timing/model': 1.088008064078167, 'learning_rate/qwen_vl_interface': 2.4179318439133245e-06, 'learning_rate/action_model': 2.417931843913324e-05, 'epoch': 0.28}) +[2026-06-24 13:44:45,673][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.2487230747938156, 'train/grad_norm_pre_clip': 86.28902435302734, 'timing/data': 0.00035151001065969467, 'timing/model': 1.0834411610849202, 'learning_rate/qwen_vl_interface': 2.414860943597084e-06, 'learning_rate/action_model': 2.414860943597084e-05, 'epoch': 0.28}) +[2026-06-24 13:44:46,760][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.04194585233926773, 'train/grad_norm_pre_clip': 46.35554885864258, 'timing/data': 0.0003338269889354706, 'timing/model': 1.085424279095605, 'learning_rate/qwen_vl_interface': 2.4117918829717953e-06, 'learning_rate/action_model': 2.4117918829717955e-05, 'epoch': 0.28}) +[2026-06-24 13:44:47,846][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.05367809161543846, 'train/grad_norm_pre_clip': 55.8619270324707, 'timing/data': 0.0003686330746859312, 'timing/model': 1.0847873520106077, 'learning_rate/qwen_vl_interface': 2.4087246640289397e-06, 'learning_rate/action_model': 2.4087246640289397e-05, 'epoch': 0.28}) +[2026-06-24 13:44:48,932][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.024547435343265533, 'train/grad_norm_pre_clip': 47.14945602416992, 'timing/data': 0.0003280069213360548, 'timing/model': 1.0838782452046871, 'learning_rate/qwen_vl_interface': 2.405659288758804e-06, 'learning_rate/action_model': 2.4056592887588042e-05, 'epoch': 0.28}) +[2026-06-24 13:44:50,018][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.09061428159475327, 'train/grad_norm_pre_clip': 83.1586685180664, 'timing/data': 0.0003162948414683342, 'timing/model': 1.08458600519225, 'learning_rate/qwen_vl_interface': 2.402595759150474e-06, 'learning_rate/action_model': 2.4025957591504743e-05, 'epoch': 0.28}) +[2026-06-24 13:44:51,105][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.018951714038848877, 'train/grad_norm_pre_clip': 14.214094161987305, 'timing/data': 0.0003386570606380701, 'timing/model': 1.0849146728869528, 'learning_rate/qwen_vl_interface': 2.399534077191842e-06, 'learning_rate/action_model': 2.3995340771918418e-05, 'epoch': 0.28}) +[2026-06-24 13:44:52,192][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.0003568497486412525, 'train/grad_norm_pre_clip': 0.6812014579772949, 'timing/data': 0.00030844006687402725, 'timing/model': 1.0858858991414309, 'learning_rate/qwen_vl_interface': 2.3964742448695996e-06, 'learning_rate/action_model': 2.3964742448695994e-05, 'epoch': 0.28}) +[2026-06-24 13:44:53,283][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.10824200510978699, 'train/grad_norm_pre_clip': 89.87162780761719, 'timing/data': 0.00031671510078012943, 'timing/model': 1.0889751249924302, 'learning_rate/qwen_vl_interface': 2.393416264169242e-06, 'learning_rate/action_model': 2.3934162641692416e-05, 'epoch': 0.28}) +[2026-06-24 13:44:54,639][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.06975175440311432, 'train/grad_norm_pre_clip': 79.84479522705078, 'timing/data': 0.0014409369323402643, 'timing/model': 1.353704750072211, 'learning_rate/qwen_vl_interface': 2.390360137075055e-06, 'learning_rate/action_model': 2.3903601370750548e-05, 'epoch': 0.28}) +[2026-06-24 13:44:55,727][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.053724877536296844, 'train/grad_norm_pre_clip': 14.684223175048828, 'timing/data': 0.00032445997931063175, 'timing/model': 1.0857711748685688, 'learning_rate/qwen_vl_interface': 2.3873058655701277e-06, 'learning_rate/action_model': 2.3873058655701278e-05, 'epoch': 0.28}) +[2026-06-24 13:44:56,814][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.06254998594522476, 'train/grad_norm_pre_clip': 42.124454498291016, 'timing/data': 0.0003133390564471483, 'timing/model': 1.0857784480322152, 'learning_rate/qwen_vl_interface': 2.384253451636346e-06, 'learning_rate/action_model': 2.3842534516363462e-05, 'epoch': 0.28}) +[2026-06-24 13:44:57,901][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.12002244591712952, 'train/grad_norm_pre_clip': 39.346458435058594, 'timing/data': 0.00039695599116384983, 'timing/model': 1.084880701964721, 'learning_rate/qwen_vl_interface': 2.3812028972543847e-06, 'learning_rate/action_model': 2.3812028972543845e-05, 'epoch': 0.28}) +[2026-06-24 13:44:58,987][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.01765931025147438, 'train/grad_norm_pre_clip': 11.45543098449707, 'timing/data': 0.000355779193341732, 'timing/model': 1.084788919892162, 'learning_rate/qwen_vl_interface': 2.3781542044037195e-06, 'learning_rate/action_model': 2.3781542044037193e-05, 'epoch': 0.28}) +[2026-06-24 13:45:00,075][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.2300376445055008, 'train/grad_norm_pre_clip': 96.6204605102539, 'timing/data': 0.0003213339950889349, 'timing/model': 1.0867247791029513, 'learning_rate/qwen_vl_interface': 2.3751073750626124e-06, 'learning_rate/action_model': 2.3751073750626124e-05, 'epoch': 0.28}) +[2026-06-24 13:45:01,163][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.07296528667211533, 'train/grad_norm_pre_clip': 57.29496765136719, 'timing/data': 0.0003404589369893074, 'timing/model': 1.086477052886039, 'learning_rate/qwen_vl_interface': 2.372062411208114e-06, 'learning_rate/action_model': 2.3720624112081138e-05, 'epoch': 0.28}) +[2026-06-24 13:45:02,263][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.0019721188582479954, 'train/grad_norm_pre_clip': 2.836474895477295, 'timing/data': 0.0004252588842064142, 'timing/model': 1.0979718761518598, 'learning_rate/qwen_vl_interface': 2.3690193148160744e-06, 'learning_rate/action_model': 2.3690193148160743e-05, 'epoch': 0.28}) +[2026-06-24 13:45:03,351][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.060206521302461624, 'train/grad_norm_pre_clip': 52.2761344909668, 'timing/data': 0.00032699410803616047, 'timing/model': 1.0861997231841087, 'learning_rate/qwen_vl_interface': 2.3659780878611218e-06, 'learning_rate/action_model': 2.365978087861122e-05, 'epoch': 0.28}) +[2026-06-24 13:45:04,441][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.0053252228535711765, 'train/grad_norm_pre_clip': 7.616735935211182, 'timing/data': 0.00043146079406142235, 'timing/model': 1.088247282197699, 'learning_rate/qwen_vl_interface': 2.3629387323166766e-06, 'learning_rate/action_model': 2.3629387323166765e-05, 'epoch': 0.28}) +[2026-06-24 13:45:05,529][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.07500828057527542, 'train/grad_norm_pre_clip': 45.62421417236328, 'timing/data': 0.0003238888457417488, 'timing/model': 1.0859882170334458, 'learning_rate/qwen_vl_interface': 2.3599012501549453e-06, 'learning_rate/action_model': 2.3599012501549453e-05, 'epoch': 0.28}) +[2026-06-24 13:45:06,611][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.032305460423231125, 'train/grad_norm_pre_clip': 25.93202018737793, 'timing/data': 0.0003294600173830986, 'timing/model': 1.0811202190816402, 'learning_rate/qwen_vl_interface': 2.356865643346913e-06, 'learning_rate/action_model': 2.3568656433469134e-05, 'epoch': 0.28}) +[2026-06-24 13:45:07,699][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.014462011866271496, 'train/grad_norm_pre_clip': 13.274415969848633, 'timing/data': 0.0003266639541834593, 'timing/model': 1.0856786670628935, 'learning_rate/qwen_vl_interface': 2.3538319138623556e-06, 'learning_rate/action_model': 2.353831913862356e-05, 'epoch': 0.28}) +[2026-06-24 13:45:08,788][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.056410741060972214, 'train/grad_norm_pre_clip': 42.55592727661133, 'timing/data': 0.00034616002812981606, 'timing/model': 1.0879294571932405, 'learning_rate/qwen_vl_interface': 2.350800063669826e-06, 'learning_rate/action_model': 2.3508000636698257e-05, 'epoch': 0.28}) +[2026-06-24 13:45:09,877][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.1177997887134552, 'train/grad_norm_pre_clip': 59.288429260253906, 'timing/data': 0.00039590406231582165, 'timing/model': 1.087221086025238, 'learning_rate/qwen_vl_interface': 2.3477700947366602e-06, 'learning_rate/action_model': 2.3477700947366602e-05, 'epoch': 0.28}) +[2026-06-24 13:45:10,967][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.0018664601957425475, 'train/grad_norm_pre_clip': 1.2873538732528687, 'timing/data': 0.00032766605727374554, 'timing/model': 1.0885695009492338, 'learning_rate/qwen_vl_interface': 2.3447420090289694e-06, 'learning_rate/action_model': 2.3447420090289694e-05, 'epoch': 0.28}) +[2026-06-24 13:45:12,322][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.25753799080848694, 'train/grad_norm_pre_clip': 49.96247100830078, 'timing/data': 0.0003458301071077585, 'timing/model': 1.353054708102718, 'learning_rate/qwen_vl_interface': 2.3417158085116466e-06, 'learning_rate/action_model': 2.3417158085116466e-05, 'epoch': 0.28}) +[2026-06-24 13:45:13,410][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.10176969319581985, 'train/grad_norm_pre_clip': 76.0621109008789, 'timing/data': 0.00032733497209846973, 'timing/model': 1.0866555538959801, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 0.28}) +[2026-06-24 13:45:14,496][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.073032446205616, 'train/grad_norm_pre_clip': 54.79061508178711, 'timing/data': 0.00031708599999547005, 'timing/model': 1.0841386548709124, 'learning_rate/qwen_vl_interface': 2.335669070901558e-06, 'learning_rate/action_model': 2.335669070901558e-05, 'epoch': 0.28}) +[2026-06-24 13:45:15,584][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.01506910752505064, 'train/grad_norm_pre_clip': 27.254371643066406, 'timing/data': 0.0003399290144443512, 'timing/model': 1.086481929058209, 'learning_rate/qwen_vl_interface': 2.3326485377324532e-06, 'learning_rate/action_model': 2.3326485377324535e-05, 'epoch': 0.28}) +[2026-06-24 13:45:16,672][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.3354111909866333, 'train/grad_norm_pre_clip': 79.75900268554688, 'timing/data': 0.00033286609686911106, 'timing/model': 1.086087808944285, 'learning_rate/qwen_vl_interface': 2.329629897601039e-06, 'learning_rate/action_model': 2.329629897601039e-05, 'epoch': 0.28}) +[2026-06-24 13:45:17,762][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.04029414802789688, 'train/grad_norm_pre_clip': 25.616323471069336, 'timing/data': 0.0004976659547537565, 'timing/model': 1.0885625898372382, 'learning_rate/qwen_vl_interface': 2.32661315246608e-06, 'learning_rate/action_model': 2.32661315246608e-05, 'epoch': 0.28}) +[2026-06-24 13:45:18,854][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.011221326887607574, 'train/grad_norm_pre_clip': 14.393168449401855, 'timing/data': 0.00032643298618495464, 'timing/model': 1.089487207820639, 'learning_rate/qwen_vl_interface': 2.3235983042851052e-06, 'learning_rate/action_model': 2.3235983042851054e-05, 'epoch': 0.28}) +[2026-06-24 13:45:19,944][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.1513904482126236, 'train/grad_norm_pre_clip': 62.15818786621094, 'timing/data': 0.00031840894371271133, 'timing/model': 1.0890064449049532, 'learning_rate/qwen_vl_interface': 2.32058535501442e-06, 'learning_rate/action_model': 2.3205853550144202e-05, 'epoch': 0.28}) +[2026-06-24 13:45:21,030][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.12005530297756195, 'train/grad_norm_pre_clip': 28.945011138916016, 'timing/data': 0.00034295511431992054, 'timing/model': 1.0844785859808326, 'learning_rate/qwen_vl_interface': 2.3175743066090957e-06, 'learning_rate/action_model': 2.317574306609096e-05, 'epoch': 0.28}) +[2026-06-24 13:45:22,115][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.11233795434236526, 'train/grad_norm_pre_clip': 68.578125, 'timing/data': 0.0003426040057092905, 'timing/model': 1.0825457850005478, 'learning_rate/qwen_vl_interface': 2.3145651610229635e-06, 'learning_rate/action_model': 2.3145651610229634e-05, 'epoch': 0.28}) +[2026-06-24 13:45:23,196][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.034127797931432724, 'train/grad_norm_pre_clip': 41.57905578613281, 'timing/data': 0.0003265740815550089, 'timing/model': 1.0798178289551288, 'learning_rate/qwen_vl_interface': 2.3115579202086313e-06, 'learning_rate/action_model': 2.311557920208631e-05, 'epoch': 0.28}) +[2026-06-24 13:45:24,287][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.09704641252756119, 'train/grad_norm_pre_clip': 95.20893096923828, 'timing/data': 0.0003571209963411093, 'timing/model': 1.0894279200583696, 'learning_rate/qwen_vl_interface': 2.308552586117463e-06, 'learning_rate/action_model': 2.308552586117463e-05, 'epoch': 0.28}) +[2026-06-24 13:45:25,375][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.1284281313419342, 'train/grad_norm_pre_clip': 34.909912109375, 'timing/data': 0.00032434007152915, 'timing/model': 1.0857698551844805, 'learning_rate/qwen_vl_interface': 2.3055491606995834e-06, 'learning_rate/action_model': 2.3055491606995833e-05, 'epoch': 0.28}) +[2026-06-24 13:45:26,464][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.018808992579579353, 'train/grad_norm_pre_clip': 27.163578033447266, 'timing/data': 0.00031992094591259956, 'timing/model': 1.088152893818915, 'learning_rate/qwen_vl_interface': 2.302547645903889e-06, 'learning_rate/action_model': 2.302547645903889e-05, 'epoch': 0.28}) +[2026-06-24 13:45:27,550][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.003071750281378627, 'train/grad_norm_pre_clip': 2.78753924369812, 'timing/data': 0.0004044901579618454, 'timing/model': 1.084339878987521, 'learning_rate/qwen_vl_interface': 2.299548043678026e-06, 'learning_rate/action_model': 2.299548043678026e-05, 'epoch': 0.28}) +[2026-06-24 13:45:28,639][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.2186656892299652, 'train/grad_norm_pre_clip': 134.33480834960938, 'timing/data': 0.0003471330273896456, 'timing/model': 1.0866566230542958, 'learning_rate/qwen_vl_interface': 2.2965503559684045e-06, 'learning_rate/action_model': 2.2965503559684044e-05, 'epoch': 0.28}) +[2026-06-24 13:45:29,991][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.04552512615919113, 'train/grad_norm_pre_clip': 18.6149845123291, 'timing/data': 0.0003563300706446171, 'timing/model': 1.3506954340264201, 'learning_rate/qwen_vl_interface': 2.2935545847201935e-06, 'learning_rate/action_model': 2.2935545847201934e-05, 'epoch': 0.28}) +[2026-06-24 13:45:31,079][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.33145812153816223, 'train/grad_norm_pre_clip': 27.117631912231445, 'timing/data': 0.0003919468726962805, 'timing/model': 1.0857624739874154, 'learning_rate/qwen_vl_interface': 2.2905607318773136e-06, 'learning_rate/action_model': 2.2905607318773134e-05, 'epoch': 0.28}) +[2026-06-24 13:45:32,163][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.0036395147908478975, 'train/grad_norm_pre_clip': 5.730428695678711, 'timing/data': 0.000351310009136796, 'timing/model': 1.082440520869568, 'learning_rate/qwen_vl_interface': 2.2875687993824438e-06, 'learning_rate/action_model': 2.2875687993824434e-05, 'epoch': 0.28}) +[2026-06-24 13:45:33,252][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.09912727028131485, 'train/grad_norm_pre_clip': 62.13758087158203, 'timing/data': 0.00034253322519361973, 'timing/model': 1.0875094970688224, 'learning_rate/qwen_vl_interface': 2.284578789177018e-06, 'learning_rate/action_model': 2.284578789177018e-05, 'epoch': 0.28}) +[2026-06-24 13:45:34,342][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.04148617014288902, 'train/grad_norm_pre_clip': 45.29304504394531, 'timing/data': 0.000328657915815711, 'timing/model': 1.0882337789516896, 'learning_rate/qwen_vl_interface': 2.281590703201222e-06, 'learning_rate/action_model': 2.281590703201222e-05, 'epoch': 0.28}) +[2026-06-24 13:45:35,428][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.0777319148182869, 'train/grad_norm_pre_clip': 76.07435607910156, 'timing/data': 0.00038254912942647934, 'timing/model': 1.0845545348711312, 'learning_rate/qwen_vl_interface': 2.2786045433939903e-06, 'learning_rate/action_model': 2.27860454339399e-05, 'epoch': 0.28}) +[2026-06-24 13:45:36,516][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.01074153557419777, 'train/grad_norm_pre_clip': 18.498992919921875, 'timing/data': 0.00043277302756905556, 'timing/model': 1.0869846630375832, 'learning_rate/qwen_vl_interface': 2.27562031169301e-06, 'learning_rate/action_model': 2.27562031169301e-05, 'epoch': 0.28}) +[2026-06-24 13:45:37,604][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.03231772407889366, 'train/grad_norm_pre_clip': 24.023000717163086, 'timing/data': 0.00033243512734770775, 'timing/model': 1.0862591499462724, 'learning_rate/qwen_vl_interface': 2.2726380100347187e-06, 'learning_rate/action_model': 2.2726380100347185e-05, 'epoch': 0.28}) +[2026-06-24 13:45:38,690][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.1886928677558899, 'train/grad_norm_pre_clip': 159.25128173828125, 'timing/data': 0.0003255519550293684, 'timing/model': 1.084171456983313, 'learning_rate/qwen_vl_interface': 2.269657640354296e-06, 'learning_rate/action_model': 2.269657640354296e-05, 'epoch': 0.28}) +[2026-06-24 13:45:39,781][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.061514802277088165, 'train/grad_norm_pre_clip': 37.56147766113281, 'timing/data': 0.0003214238677173853, 'timing/model': 1.0894978269934654, 'learning_rate/qwen_vl_interface': 2.2666792045856735e-06, 'learning_rate/action_model': 2.2666792045856737e-05, 'epoch': 0.28}) +[2026-06-24 13:45:40,867][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.28056514263153076, 'train/grad_norm_pre_clip': 27.94733238220215, 'timing/data': 0.0003093509003520012, 'timing/model': 1.0840885420329869, 'learning_rate/qwen_vl_interface': 2.263702704661526e-06, 'learning_rate/action_model': 2.263702704661526e-05, 'epoch': 0.28}) +[2026-06-24 13:45:41,956][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.0009641361539252102, 'train/grad_norm_pre_clip': 0.9836628437042236, 'timing/data': 0.0004021760541945696, 'timing/model': 1.0873929420486093, 'learning_rate/qwen_vl_interface': 2.2607281425132728e-06, 'learning_rate/action_model': 2.260728142513273e-05, 'epoch': 0.28}) +[2026-06-24 13:45:43,045][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.015687277540564537, 'train/grad_norm_pre_clip': 12.881844520568848, 'timing/data': 0.0003399390261620283, 'timing/model': 1.0871792200487107, 'learning_rate/qwen_vl_interface': 2.257755520071073e-06, 'learning_rate/action_model': 2.2577555200710727e-05, 'epoch': 0.28}) +[2026-06-24 13:45:44,131][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.049260545521974564, 'train/grad_norm_pre_clip': 10.731804847717285, 'timing/data': 0.0003145819064229727, 'timing/model': 1.084622198017314, 'learning_rate/qwen_vl_interface': 2.2547848392638293e-06, 'learning_rate/action_model': 2.254784839263829e-05, 'epoch': 0.28}) +[2026-06-24 13:45:45,221][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.055162977427244186, 'train/grad_norm_pre_clip': 38.261695861816406, 'timing/data': 0.00033269496634602547, 'timing/model': 1.0883010090328753, 'learning_rate/qwen_vl_interface': 2.251816102019186e-06, 'learning_rate/action_model': 2.251816102019186e-05, 'epoch': 0.28}) +[2026-06-24 13:45:46,316][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.06866434961557388, 'train/grad_norm_pre_clip': 57.93191909790039, 'timing/data': 0.00035961600951850414, 'timing/model': 1.093503899173811, 'learning_rate/qwen_vl_interface': 2.248849310263521e-06, 'learning_rate/action_model': 2.248849310263521e-05, 'epoch': 0.28}) +[2026-06-24 13:45:47,671][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.10437659919261932, 'train/grad_norm_pre_clip': 49.70830535888672, 'timing/data': 0.0003358311951160431, 'timing/model': 1.3537933370098472, 'learning_rate/qwen_vl_interface': 2.2458844659219583e-06, 'learning_rate/action_model': 2.2458844659219583e-05, 'epoch': 0.28}) +[2026-06-24 13:45:48,763][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.015335469506680965, 'train/grad_norm_pre_clip': 16.234516143798828, 'timing/data': 0.0003335271030664444, 'timing/model': 1.0899582768324763, 'learning_rate/qwen_vl_interface': 2.242921570918351e-06, 'learning_rate/action_model': 2.242921570918351e-05, 'epoch': 0.28}) +[2026-06-24 13:45:49,848][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.1863034963607788, 'train/grad_norm_pre_clip': 29.442699432373047, 'timing/data': 0.000328797847032547, 'timing/model': 1.0841040648519993, 'learning_rate/qwen_vl_interface': 2.2399606271752847e-06, 'learning_rate/action_model': 2.2399606271752846e-05, 'epoch': 0.28}) +[2026-06-24 13:45:50,938][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.034342218190431595, 'train/grad_norm_pre_clip': 55.432613372802734, 'timing/data': 0.00031966017559170723, 'timing/model': 1.0878304180223495, 'learning_rate/qwen_vl_interface': 2.2370016366140916e-06, 'learning_rate/action_model': 2.2370016366140916e-05, 'epoch': 0.28}) +[2026-06-24 13:45:52,028][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.23164501786231995, 'train/grad_norm_pre_clip': 61.14013671875, 'timing/data': 0.0003250311128795147, 'timing/model': 1.0888992280233651, 'learning_rate/qwen_vl_interface': 2.234044601154824e-06, 'learning_rate/action_model': 2.2340446011548237e-05, 'epoch': 0.28}) +[2026-06-24 13:45:53,117][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.03776142746210098, 'train/grad_norm_pre_clip': 44.85572052001953, 'timing/data': 0.0003264739643782377, 'timing/model': 1.0868841409683228, 'learning_rate/qwen_vl_interface': 2.2310895227162697e-06, 'learning_rate/action_model': 2.23108952271627e-05, 'epoch': 0.28}) +[2026-06-24 13:45:54,202][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.1549353003501892, 'train/grad_norm_pre_clip': 78.9002685546875, 'timing/data': 0.00032639317214488983, 'timing/model': 1.0835685201454908, 'learning_rate/qwen_vl_interface': 2.228136403215951e-06, 'learning_rate/action_model': 2.2281364032159508e-05, 'epoch': 0.28}) +[2026-06-24 13:45:55,290][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.008307071402668953, 'train/grad_norm_pre_clip': 7.1356940269470215, 'timing/data': 0.00037115695886313915, 'timing/model': 1.0863175459671766, 'learning_rate/qwen_vl_interface': 2.2251852445701096e-06, 'learning_rate/action_model': 2.2251852445701095e-05, 'epoch': 0.28}) +[2026-06-24 13:45:56,378][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.21295061707496643, 'train/grad_norm_pre_clip': 104.53044128417969, 'timing/data': 0.0003965049982070923, 'timing/model': 1.0861434401012957, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 0.28}) +[2026-06-24 13:45:57,467][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.10148327052593231, 'train/grad_norm_pre_clip': 62.2217903137207, 'timing/data': 0.0003979478497058153, 'timing/model': 1.0874638019595295, 'learning_rate/qwen_vl_interface': 2.219288817500494e-06, 'learning_rate/action_model': 2.2192888175004938e-05, 'epoch': 0.28}) +[2026-06-24 13:45:58,552][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.110281802713871, 'train/grad_norm_pre_clip': 52.95486068725586, 'timing/data': 0.00032243598252534866, 'timing/model': 1.083617513999343, 'learning_rate/qwen_vl_interface': 2.216343552902848e-06, 'learning_rate/action_model': 2.2163435529028482e-05, 'epoch': 0.28}) +[2026-06-24 13:45:59,642][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.012365084141492844, 'train/grad_norm_pre_clip': 12.096765518188477, 'timing/data': 0.00033534993417561054, 'timing/model': 1.0882184610236436, 'learning_rate/qwen_vl_interface': 2.2134002568119335e-06, 'learning_rate/action_model': 2.2134002568119334e-05, 'epoch': 0.28}) +[2026-06-24 13:46:00,733][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.15801331400871277, 'train/grad_norm_pre_clip': 103.8453598022461, 'timing/data': 0.0003437860868871212, 'timing/model': 1.0901366260368377, 'learning_rate/qwen_vl_interface': 2.210458931137624e-06, 'learning_rate/action_model': 2.2104589311376235e-05, 'epoch': 0.28}) +[2026-06-24 13:46:01,825][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.11752267181873322, 'train/grad_norm_pre_clip': 48.747718811035156, 'timing/data': 0.00036945496685802937, 'timing/model': 1.0901722139678895, 'learning_rate/qwen_vl_interface': 2.2075195777885155e-06, 'learning_rate/action_model': 2.2075195777885156e-05, 'epoch': 0.28}) +[2026-06-24 13:46:02,919][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.027081603184342384, 'train/grad_norm_pre_clip': 21.15726089477539, 'timing/data': 0.00042764400132000446, 'timing/model': 1.0917411630507559, 'learning_rate/qwen_vl_interface': 2.20458219867192e-06, 'learning_rate/action_model': 2.20458219867192e-05, 'epoch': 0.28}) +[2026-06-24 13:46:04,010][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.2104884684085846, 'train/grad_norm_pre_clip': 42.93136215209961, 'timing/data': 0.00034608086571097374, 'timing/model': 1.0891979730222374, 'learning_rate/qwen_vl_interface': 2.201646795693872e-06, 'learning_rate/action_model': 2.201646795693872e-05, 'epoch': 0.28}) +[2026-06-24 13:46:05,375][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.02313738688826561, 'train/grad_norm_pre_clip': 16.304386138916016, 'timing/data': 0.0003419129643589258, 'timing/model': 1.363967424025759, 'learning_rate/qwen_vl_interface': 2.198713370759123e-06, 'learning_rate/action_model': 2.1987133707591228e-05, 'epoch': 0.28}) +[2026-06-24 13:46:06,462][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.045832425355911255, 'train/grad_norm_pre_clip': 45.604820251464844, 'timing/data': 0.0003075879067182541, 'timing/model': 1.0848476488608867, 'learning_rate/qwen_vl_interface': 2.195781925771142e-06, 'learning_rate/action_model': 2.1957819257711418e-05, 'epoch': 0.28}) +[2026-06-24 13:46:07,548][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.04461213946342468, 'train/grad_norm_pre_clip': 147.66879272460938, 'timing/data': 0.0003147509414702654, 'timing/model': 1.085284159053117, 'learning_rate/qwen_vl_interface': 2.192852462632108e-06, 'learning_rate/action_model': 2.1928524626321082e-05, 'epoch': 0.28}) +[2026-06-24 13:46:08,638][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.011429065838456154, 'train/grad_norm_pre_clip': 21.636688232421875, 'timing/data': 0.00032728491351008415, 'timing/model': 1.0876493249088526, 'learning_rate/qwen_vl_interface': 2.189924983242921e-06, 'learning_rate/action_model': 2.1899249832429208e-05, 'epoch': 0.28}) +[2026-06-24 13:46:09,727][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.04132433980703354, 'train/grad_norm_pre_clip': 31.86867332458496, 'timing/data': 0.00031903013586997986, 'timing/model': 1.0872839379590005, 'learning_rate/qwen_vl_interface': 2.1869994895031907e-06, 'learning_rate/action_model': 2.1869994895031905e-05, 'epoch': 0.28}) +[2026-06-24 13:46:10,817][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.060280941426754, 'train/grad_norm_pre_clip': 50.995967864990234, 'timing/data': 0.000343595864251256, 'timing/model': 1.0889097850304097, 'learning_rate/qwen_vl_interface': 2.184075983311234e-06, 'learning_rate/action_model': 2.184075983311234e-05, 'epoch': 0.28}) +[2026-06-24 13:46:11,908][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.17226561903953552, 'train/grad_norm_pre_clip': 88.60136413574219, 'timing/data': 0.0003345790319144726, 'timing/model': 1.088932137005031, 'learning_rate/qwen_vl_interface': 2.181154466564089e-06, 'learning_rate/action_model': 2.181154466564089e-05, 'epoch': 0.28}) +[2026-06-24 13:46:12,999][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.1283622533082962, 'train/grad_norm_pre_clip': 42.11757278442383, 'timing/data': 0.0003250709269195795, 'timing/model': 1.0895985588431358, 'learning_rate/qwen_vl_interface': 2.178234941157492e-06, 'learning_rate/action_model': 2.178234941157492e-05, 'epoch': 0.28}) +[2026-06-24 13:46:14,091][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.0464065782725811, 'train/grad_norm_pre_clip': 38.52423095703125, 'timing/data': 0.0003261228557676077, 'timing/model': 1.0907246558927, 'learning_rate/qwen_vl_interface': 2.175317408985889e-06, 'learning_rate/action_model': 2.1753174089858892e-05, 'epoch': 0.28}) +[2026-06-24 13:46:15,184][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.04216327145695686, 'train/grad_norm_pre_clip': 30.577972412109375, 'timing/data': 0.0003475330304354429, 'timing/model': 1.0906115039251745, 'learning_rate/qwen_vl_interface': 2.1724018719424413e-06, 'learning_rate/action_model': 2.1724018719424415e-05, 'epoch': 0.28}) +[2026-06-24 13:46:16,273][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.0032813020516186953, 'train/grad_norm_pre_clip': 2.337270498275757, 'timing/data': 0.00031863804906606674, 'timing/model': 1.0876519121229649, 'learning_rate/qwen_vl_interface': 2.1694883319190048e-06, 'learning_rate/action_model': 2.1694883319190046e-05, 'epoch': 0.28}) +[2026-06-24 13:46:17,366][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.06889994442462921, 'train/grad_norm_pre_clip': 44.82408142089844, 'timing/data': 0.00031243800185620785, 'timing/model': 1.091205531032756, 'learning_rate/qwen_vl_interface': 2.166576790806144e-06, 'learning_rate/action_model': 2.1665767908061442e-05, 'epoch': 0.28}) +[2026-06-24 13:46:18,455][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.20346884429454803, 'train/grad_norm_pre_clip': 87.52128601074219, 'timing/data': 0.00031981104984879494, 'timing/model': 1.0876604889053851, 'learning_rate/qwen_vl_interface': 2.16366725049313e-06, 'learning_rate/action_model': 2.1636672504931298e-05, 'epoch': 0.28}) +[2026-06-24 13:46:19,546][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.08224187791347504, 'train/grad_norm_pre_clip': 33.54741287231445, 'timing/data': 0.0003880099393427372, 'timing/model': 1.0887923771515489, 'learning_rate/qwen_vl_interface': 2.160759712867927e-06, 'learning_rate/action_model': 2.1607597128679268e-05, 'epoch': 0.28}) +[2026-06-24 13:46:20,636][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.1459173858165741, 'train/grad_norm_pre_clip': 59.507545471191406, 'timing/data': 0.0003231379669159651, 'timing/model': 1.0891611608676612, 'learning_rate/qwen_vl_interface': 2.157854179817207e-06, 'learning_rate/action_model': 2.157854179817207e-05, 'epoch': 0.28}) +[2026-06-24 13:46:21,728][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.0052741095423698425, 'train/grad_norm_pre_clip': 2.384584903717041, 'timing/data': 0.00031813792884349823, 'timing/model': 1.0895109879784286, 'learning_rate/qwen_vl_interface': 2.1549506532263392e-06, 'learning_rate/action_model': 2.1549506532263392e-05, 'epoch': 0.28}) +[2026-06-24 13:46:23,095][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.02447827160358429, 'train/grad_norm_pre_clip': 34.45005416870117, 'timing/data': 0.00032746512442827225, 'timing/model': 1.3653618739917874, 'learning_rate/qwen_vl_interface': 2.152049134979391e-06, 'learning_rate/action_model': 2.152049134979391e-05, 'epoch': 0.28}) +[2026-06-24 13:46:24,185][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.2641759216785431, 'train/grad_norm_pre_clip': 89.763427734375, 'timing/data': 0.0003163740038871765, 'timing/model': 1.0885616950690746, 'learning_rate/qwen_vl_interface': 2.1491496269591243e-06, 'learning_rate/action_model': 2.1491496269591242e-05, 'epoch': 0.28}) +[2026-06-24 13:46:25,279][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.10255563259124756, 'train/grad_norm_pre_clip': 57.837337493896484, 'timing/data': 0.00031738681718707085, 'timing/model': 1.0853030420839787, 'learning_rate/qwen_vl_interface': 2.1462521310469987e-06, 'learning_rate/action_model': 2.1462521310469987e-05, 'epoch': 0.28}) +[2026-06-24 13:46:26,373][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.009223660454154015, 'train/grad_norm_pre_clip': 9.721617698669434, 'timing/data': 0.0003970370162278414, 'timing/model': 1.0918961830902845, 'learning_rate/qwen_vl_interface': 2.1433566491231703e-06, 'learning_rate/action_model': 2.14335664912317e-05, 'epoch': 0.28}) +[2026-06-24 13:46:27,465][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.17187565565109253, 'train/grad_norm_pre_clip': 60.08055114746094, 'timing/data': 0.0003714680206030607, 'timing/model': 1.090233868919313, 'learning_rate/qwen_vl_interface': 2.140463183066482e-06, 'learning_rate/action_model': 2.140463183066482e-05, 'epoch': 0.28}) +[2026-06-24 13:46:28,556][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.1316538155078888, 'train/grad_norm_pre_clip': 51.00720977783203, 'timing/data': 0.0003077690489590168, 'timing/model': 1.090366859920323, 'learning_rate/qwen_vl_interface': 2.1375717347544745e-06, 'learning_rate/action_model': 2.1375717347544746e-05, 'epoch': 0.28}) +[2026-06-24 13:46:29,646][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.01019548811018467, 'train/grad_norm_pre_clip': 15.92760181427002, 'timing/data': 0.0003204420208930969, 'timing/model': 1.0884156830143183, 'learning_rate/qwen_vl_interface': 2.1346823060633793e-06, 'learning_rate/action_model': 2.1346823060633793e-05, 'epoch': 0.28}) +[2026-06-24 13:46:30,737][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.1862102597951889, 'train/grad_norm_pre_clip': 100.97187805175781, 'timing/data': 0.000327215064316988, 'timing/model': 1.0888184600044042, 'learning_rate/qwen_vl_interface': 2.1317948988681113e-06, 'learning_rate/action_model': 2.1317948988681114e-05, 'epoch': 0.28}) +[2026-06-24 13:46:31,825][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.22018322348594666, 'train/grad_norm_pre_clip': 78.68065643310547, 'timing/data': 0.0003139700274914503, 'timing/model': 1.0865660079289228, 'learning_rate/qwen_vl_interface': 2.1289095150422796e-06, 'learning_rate/action_model': 2.1289095150422798e-05, 'epoch': 0.28}) +[2026-06-24 13:46:32,914][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.09996648132801056, 'train/grad_norm_pre_clip': 54.507110595703125, 'timing/data': 0.00031055393628776073, 'timing/model': 1.086789048044011, 'learning_rate/qwen_vl_interface': 2.1260261564581784e-06, 'learning_rate/action_model': 2.126026156458178e-05, 'epoch': 0.28}) +[2026-06-24 13:46:34,001][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.13799694180488586, 'train/grad_norm_pre_clip': 60.322444915771484, 'timing/data': 0.00033568101935088634, 'timing/model': 1.0860359219368547, 'learning_rate/qwen_vl_interface': 2.1231448249867887e-06, 'learning_rate/action_model': 2.1231448249867888e-05, 'epoch': 0.28}) +[2026-06-24 13:46:35,092][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.0415288545191288, 'train/grad_norm_pre_clip': 59.18014144897461, 'timing/data': 0.00033920700661838055, 'timing/model': 1.0895308200269938, 'learning_rate/qwen_vl_interface': 2.120265522497772e-06, 'learning_rate/action_model': 2.120265522497772e-05, 'epoch': 0.28}) +[2026-06-24 13:46:36,183][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.13609018921852112, 'train/grad_norm_pre_clip': 59.87103271484375, 'timing/data': 0.00036181905306875706, 'timing/model': 1.0891554760746658, 'learning_rate/qwen_vl_interface': 2.1173882508594816e-06, 'learning_rate/action_model': 2.1173882508594812e-05, 'epoch': 0.28}) +[2026-06-24 13:46:37,275][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.009359985589981079, 'train/grad_norm_pre_clip': 7.418717861175537, 'timing/data': 0.0003239589277654886, 'timing/model': 1.0905743220355362, 'learning_rate/qwen_vl_interface': 2.1145130119389446e-06, 'learning_rate/action_model': 2.1145130119389444e-05, 'epoch': 0.28}) +[2026-06-24 13:46:38,369][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.22284583747386932, 'train/grad_norm_pre_clip': 53.7781982421875, 'timing/data': 0.0003402389120310545, 'timing/model': 1.0925050820223987, 'learning_rate/qwen_vl_interface': 2.1116398076018698e-06, 'learning_rate/action_model': 2.1116398076018694e-05, 'epoch': 0.29}) +[2026-06-24 13:46:39,461][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.08813890814781189, 'train/grad_norm_pre_clip': 44.8787956237793, 'timing/data': 0.00032775592990219593, 'timing/model': 1.090299196075648, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 0.29}) +[2026-06-24 13:46:40,823][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.012019116431474686, 'train/grad_norm_pre_clip': 11.57748031616211, 'timing/data': 0.00033994903787970543, 'timing/model': 1.3601794028654695, 'learning_rate/qwen_vl_interface': 2.1058995101343633e-06, 'learning_rate/action_model': 2.1058995101343632e-05, 'epoch': 0.29}) +[2026-06-24 13:46:41,912][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.04843172803521156, 'train/grad_norm_pre_clip': 20.401403427124023, 'timing/data': 0.00034364499151706696, 'timing/model': 1.0877161521930248, 'learning_rate/qwen_vl_interface': 2.1030324207287443e-06, 'learning_rate/action_model': 2.1030324207287444e-05, 'epoch': 0.29}) +[2026-06-24 13:46:43,001][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.02173193171620369, 'train/grad_norm_pre_clip': 17.337793350219727, 'timing/data': 0.0003176459576934576, 'timing/model': 1.0870992839336395, 'learning_rate/qwen_vl_interface': 2.100167373356224e-06, 'learning_rate/action_model': 2.1001673733562237e-05, 'epoch': 0.29}) +[2026-06-24 13:46:44,093][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.0984049141407013, 'train/grad_norm_pre_clip': 65.39530944824219, 'timing/data': 0.0003272539470344782, 'timing/model': 1.0902494648471475, 'learning_rate/qwen_vl_interface': 2.0973043698758955e-06, 'learning_rate/action_model': 2.0973043698758954e-05, 'epoch': 0.29}) +[2026-06-24 13:46:45,182][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.0652535930275917, 'train/grad_norm_pre_clip': 68.54222869873047, 'timing/data': 0.00032479106448590755, 'timing/model': 1.0879505849443376, 'learning_rate/qwen_vl_interface': 2.0944434121455335e-06, 'learning_rate/action_model': 2.0944434121455338e-05, 'epoch': 0.29}) +[2026-06-24 13:46:46,272][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.053453478962183, 'train/grad_norm_pre_clip': 76.42886352539062, 'timing/data': 0.0004207699093967676, 'timing/model': 1.087847501039505, 'learning_rate/qwen_vl_interface': 2.0915845020215815e-06, 'learning_rate/action_model': 2.0915845020215817e-05, 'epoch': 0.29}) +[2026-06-24 13:46:47,363][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.006544977426528931, 'train/grad_norm_pre_clip': 8.460955619812012, 'timing/data': 0.00031946017406880856, 'timing/model': 1.0892342070583254, 'learning_rate/qwen_vl_interface': 2.088727641359158e-06, 'learning_rate/action_model': 2.088727641359158e-05, 'epoch': 0.29}) +[2026-06-24 13:46:48,453][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.00669430335983634, 'train/grad_norm_pre_clip': 6.2527689933776855, 'timing/data': 0.00031632394529879093, 'timing/model': 1.088529041968286, 'learning_rate/qwen_vl_interface': 2.0858728320120443e-06, 'learning_rate/action_model': 2.0858728320120443e-05, 'epoch': 0.29}) +[2026-06-24 13:46:49,544][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.027431579306721687, 'train/grad_norm_pre_clip': 20.828468322753906, 'timing/data': 0.00034873606637120247, 'timing/model': 1.0891327180434018, 'learning_rate/qwen_vl_interface': 2.083020075832698e-06, 'learning_rate/action_model': 2.0830200758326977e-05, 'epoch': 0.29}) +[2026-06-24 13:46:50,635][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.37678757309913635, 'train/grad_norm_pre_clip': 131.80984497070312, 'timing/data': 0.0003496070858091116, 'timing/model': 1.0895450238604099, 'learning_rate/qwen_vl_interface': 2.0801693746722427e-06, 'learning_rate/action_model': 2.0801693746722425e-05, 'epoch': 0.29}) +[2026-06-24 13:46:51,724][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.11199670284986496, 'train/grad_norm_pre_clip': 50.55052947998047, 'timing/data': 0.0003278560470789671, 'timing/model': 1.0877421561162919, 'learning_rate/qwen_vl_interface': 2.077320730380464e-06, 'learning_rate/action_model': 2.077320730380464e-05, 'epoch': 0.29}) +[2026-06-24 13:46:52,817][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.09262370318174362, 'train/grad_norm_pre_clip': 40.5648078918457, 'timing/data': 0.00032699480652809143, 'timing/model': 1.0911203839350492, 'learning_rate/qwen_vl_interface': 2.074474144805818e-06, 'learning_rate/action_model': 2.0744741448058176e-05, 'epoch': 0.29}) +[2026-06-24 13:46:53,909][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.15459005534648895, 'train/grad_norm_pre_clip': 42.67754364013672, 'timing/data': 0.0003192299045622349, 'timing/model': 1.0906976808328182, 'learning_rate/qwen_vl_interface': 2.071629619795425e-06, 'learning_rate/action_model': 2.071629619795425e-05, 'epoch': 0.29}) +[2026-06-24 13:46:55,002][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.019898273050785065, 'train/grad_norm_pre_clip': 26.233980178833008, 'timing/data': 0.0003496268764138222, 'timing/model': 1.0909678388852626, 'learning_rate/qwen_vl_interface': 2.0687871571950627e-06, 'learning_rate/action_model': 2.0687871571950625e-05, 'epoch': 0.29}) +[2026-06-24 13:46:56,090][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.03696165233850479, 'train/grad_norm_pre_clip': 15.235800743103027, 'timing/data': 0.0005482309497892857, 'timing/model': 1.0859546870924532, 'learning_rate/qwen_vl_interface': 2.0659467588491768e-06, 'learning_rate/action_model': 2.0659467588491767e-05, 'epoch': 0.29}) +[2026-06-24 13:46:57,180][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.1074499785900116, 'train/grad_norm_pre_clip': 57.805843353271484, 'timing/data': 0.0003143011126667261, 'timing/model': 1.0888593171257526, 'learning_rate/qwen_vl_interface': 2.06310842660087e-06, 'learning_rate/action_model': 2.0631084266008697e-05, 'epoch': 0.29}) +[2026-06-24 13:46:58,544][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.30261659622192383, 'train/grad_norm_pre_clip': 109.97501373291016, 'timing/data': 0.0003394479863345623, 'timing/model': 1.3619799341540784, 'learning_rate/qwen_vl_interface': 2.0602721622919058e-06, 'learning_rate/action_model': 2.0602721622919054e-05, 'epoch': 0.29}) +[2026-06-24 13:46:59,631][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.010806083679199219, 'train/grad_norm_pre_clip': 7.00333833694458, 'timing/data': 0.0003462410531938076, 'timing/model': 1.0853414558805525, 'learning_rate/qwen_vl_interface': 2.0574379677627042e-06, 'learning_rate/action_model': 2.0574379677627042e-05, 'epoch': 0.29}) +[2026-06-24 13:47:00,720][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.018893472850322723, 'train/grad_norm_pre_clip': 6.791947841644287, 'timing/data': 0.00032754591666162014, 'timing/model': 1.0880750138312578, 'learning_rate/qwen_vl_interface': 2.0546058448523466e-06, 'learning_rate/action_model': 2.0546058448523466e-05, 'epoch': 0.29}) +[2026-06-24 13:47:01,810][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.05587274581193924, 'train/grad_norm_pre_clip': 40.72480773925781, 'timing/data': 0.0003147111274302006, 'timing/model': 1.0875567600596696, 'learning_rate/qwen_vl_interface': 2.051775795398563e-06, 'learning_rate/action_model': 2.0517757953985627e-05, 'epoch': 0.29}) +[2026-06-24 13:47:02,900][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.03904719278216362, 'train/grad_norm_pre_clip': 20.30460548400879, 'timing/data': 0.00031400006264448166, 'timing/model': 1.0888575250282884, 'learning_rate/qwen_vl_interface': 2.0489478212377424e-06, 'learning_rate/action_model': 2.0489478212377425e-05, 'epoch': 0.29}) +[2026-06-24 13:47:03,989][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.019659148529171944, 'train/grad_norm_pre_clip': 11.43252182006836, 'timing/data': 0.00031991093419492245, 'timing/model': 1.0875976881943643, 'learning_rate/qwen_vl_interface': 2.046121924204929e-06, 'learning_rate/action_model': 2.046121924204929e-05, 'epoch': 0.29}) +[2026-06-24 13:47:05,078][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.002348146867007017, 'train/grad_norm_pre_clip': 2.4386796951293945, 'timing/data': 0.00036875298246741295, 'timing/model': 1.0872600642032921, 'learning_rate/qwen_vl_interface': 2.0432981061338135e-06, 'learning_rate/action_model': 2.0432981061338135e-05, 'epoch': 0.29}) +[2026-06-24 13:47:06,164][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.05553494021296501, 'train/grad_norm_pre_clip': 42.36255645751953, 'timing/data': 0.0003515910357236862, 'timing/model': 1.0839129441883415, 'learning_rate/qwen_vl_interface': 2.040476368856742e-06, 'learning_rate/action_model': 2.0404763688567422e-05, 'epoch': 0.29}) +[2026-06-24 13:47:07,252][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.09869854152202606, 'train/grad_norm_pre_clip': 55.74165725708008, 'timing/data': 0.0003250811714679003, 'timing/model': 1.0866088210605085, 'learning_rate/qwen_vl_interface': 2.037656714204712e-06, 'learning_rate/action_model': 2.037656714204712e-05, 'epoch': 0.29}) +[2026-06-24 13:47:08,344][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.02627621218562126, 'train/grad_norm_pre_clip': 11.932343482971191, 'timing/data': 0.00034288386814296246, 'timing/model': 1.0899766699876636, 'learning_rate/qwen_vl_interface': 2.034839144007362e-06, 'learning_rate/action_model': 2.034839144007362e-05, 'epoch': 0.29}) +[2026-06-24 13:47:09,434][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.21548256278038025, 'train/grad_norm_pre_clip': 79.20429229736328, 'timing/data': 0.00036446494050323963, 'timing/model': 1.0888489121571183, 'learning_rate/qwen_vl_interface': 2.0320236600929848e-06, 'learning_rate/action_model': 2.032023660092985e-05, 'epoch': 0.29}) +[2026-06-24 13:47:10,526][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.13369473814964294, 'train/grad_norm_pre_clip': 39.470489501953125, 'timing/data': 0.0003280860837548971, 'timing/model': 1.0898326900787652, 'learning_rate/qwen_vl_interface': 2.029210264288518e-06, 'learning_rate/action_model': 2.029210264288518e-05, 'epoch': 0.29}) +[2026-06-24 13:47:11,615][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.391620397567749, 'train/grad_norm_pre_clip': 169.41795349121094, 'timing/data': 0.0003615089226514101, 'timing/model': 1.0879649298731238, 'learning_rate/qwen_vl_interface': 2.0263989584195443e-06, 'learning_rate/action_model': 2.0263989584195443e-05, 'epoch': 0.29}) +[2026-06-24 13:47:12,707][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.06887992471456528, 'train/grad_norm_pre_clip': 33.17380142211914, 'timing/data': 0.00031990092247724533, 'timing/model': 1.0901405101176351, 'learning_rate/qwen_vl_interface': 2.023589744310287e-06, 'learning_rate/action_model': 2.023589744310287e-05, 'epoch': 0.29}) +[2026-06-24 13:47:13,797][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.0008453989867120981, 'train/grad_norm_pre_clip': 0.7668255567550659, 'timing/data': 0.0003635429311543703, 'timing/model': 1.0884757719468325, 'learning_rate/qwen_vl_interface': 2.0207826237836155e-06, 'learning_rate/action_model': 2.0207826237836156e-05, 'epoch': 0.29}) +[2026-06-24 13:47:14,889][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.03146929293870926, 'train/grad_norm_pre_clip': 46.59025192260742, 'timing/data': 0.0003265140112489462, 'timing/model': 1.089943370083347, 'learning_rate/qwen_vl_interface': 2.017977598661042e-06, 'learning_rate/action_model': 2.0179775986610422e-05, 'epoch': 0.29}) +[2026-06-24 13:47:16,322][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.1866615265607834, 'train/grad_norm_pre_clip': 88.50404357910156, 'timing/data': 0.00034806388430297375, 'timing/model': 1.4311193081084639, 'learning_rate/qwen_vl_interface': 2.015174670762713e-06, 'learning_rate/action_model': 2.0151746707627128e-05, 'epoch': 0.29}) +[2026-06-24 13:47:17,682][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.07724004983901978, 'train/grad_norm_pre_clip': 90.02608489990234, 'timing/data': 0.0003958740271627903, 'timing/model': 1.3581396350637078, 'learning_rate/qwen_vl_interface': 2.012373841907422e-06, 'learning_rate/action_model': 2.012373841907422e-05, 'epoch': 0.29}) +[2026-06-24 13:47:18,767][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.11684418469667435, 'train/grad_norm_pre_clip': 124.26898956298828, 'timing/data': 0.00042346585541963577, 'timing/model': 1.0827624360099435, 'learning_rate/qwen_vl_interface': 2.009575113912595e-06, 'learning_rate/action_model': 2.0095751139125952e-05, 'epoch': 0.29}) +[2026-06-24 13:47:19,856][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.02250063046813011, 'train/grad_norm_pre_clip': 20.87445831298828, 'timing/data': 0.0003326251171529293, 'timing/model': 1.087741174036637, 'learning_rate/qwen_vl_interface': 2.006778488594293e-06, 'learning_rate/action_model': 2.0067784885942928e-05, 'epoch': 0.29}) +[2026-06-24 13:47:20,943][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.03262128680944443, 'train/grad_norm_pre_clip': 23.330007553100586, 'timing/data': 0.0003194008022546768, 'timing/model': 1.0851465470623225, 'learning_rate/qwen_vl_interface': 2.003983967767222e-06, 'learning_rate/action_model': 2.0039839677672217e-05, 'epoch': 0.29}) +[2026-06-24 13:47:22,031][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.04004056006669998, 'train/grad_norm_pre_clip': 59.840728759765625, 'timing/data': 0.0003285978455096483, 'timing/model': 1.0863932608626783, 'learning_rate/qwen_vl_interface': 2.0011915532447105e-06, 'learning_rate/action_model': 2.0011915532447104e-05, 'epoch': 0.29}) +[2026-06-24 13:47:23,122][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.004817905370146036, 'train/grad_norm_pre_clip': 5.026789665222168, 'timing/data': 0.00032312702387571335, 'timing/model': 1.089424948208034, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 0.29}) +[2026-06-24 13:47:24,209][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.05215713381767273, 'train/grad_norm_pre_clip': 17.729961395263672, 'timing/data': 0.0003624209202826023, 'timing/model': 1.0851401858963072, 'learning_rate/qwen_vl_interface': 1.995613050359877e-06, 'learning_rate/action_model': 1.995613050359877e-05, 'epoch': 0.29}) +[2026-06-24 13:47:25,296][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.01964166760444641, 'train/grad_norm_pre_clip': 12.30713176727295, 'timing/data': 0.000508615979924798, 'timing/model': 1.0851590419188142, 'learning_rate/qwen_vl_interface': 1.992826965617383e-06, 'learning_rate/action_model': 1.9928269656173828e-05, 'epoch': 0.29}) +[2026-06-24 13:47:26,379][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.09750289469957352, 'train/grad_norm_pre_clip': 58.56227111816406, 'timing/data': 0.0003192101139575243, 'timing/model': 1.0818548030219972, 'learning_rate/qwen_vl_interface': 1.9900429944191084e-06, 'learning_rate/action_model': 1.9900429944191085e-05, 'epoch': 0.29}) +[2026-06-24 13:47:27,467][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.1333209127187729, 'train/grad_norm_pre_clip': 77.91401672363281, 'timing/data': 0.0005171520169824362, 'timing/model': 1.085930393077433, 'learning_rate/qwen_vl_interface': 1.987261138571542e-06, 'learning_rate/action_model': 1.987261138571542e-05, 'epoch': 0.29}) +[2026-06-24 13:47:28,560][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.27800068259239197, 'train/grad_norm_pre_clip': 92.19903564453125, 'timing/data': 0.0005694499704986811, 'timing/model': 1.090892287902534, 'learning_rate/qwen_vl_interface': 1.984481399879802e-06, 'learning_rate/action_model': 1.9844813998798018e-05, 'epoch': 0.29}) +[2026-06-24 13:47:29,651][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.13160693645477295, 'train/grad_norm_pre_clip': 59.21830749511719, 'timing/data': 0.0005093479994684458, 'timing/model': 1.0885300489608198, 'learning_rate/qwen_vl_interface': 1.9817037801476257e-06, 'learning_rate/action_model': 1.9817037801476257e-05, 'epoch': 0.29}) +[2026-06-24 13:47:30,738][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.15340940654277802, 'train/grad_norm_pre_clip': 74.72537231445312, 'timing/data': 0.00032808701507747173, 'timing/model': 1.0859794758725911, 'learning_rate/qwen_vl_interface': 1.9789282811773837e-06, 'learning_rate/action_model': 1.9789282811773837e-05, 'epoch': 0.29}) +[2026-06-24 13:47:31,829][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.06682659685611725, 'train/grad_norm_pre_clip': 46.122467041015625, 'timing/data': 0.0003340670373290777, 'timing/model': 1.0892015621066093, 'learning_rate/qwen_vl_interface': 1.9761549047700674e-06, 'learning_rate/action_model': 1.9761549047700672e-05, 'epoch': 0.29}) +[2026-06-24 13:47:32,917][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.09920049458742142, 'train/grad_norm_pre_clip': 106.81116485595703, 'timing/data': 0.00032020104117691517, 'timing/model': 1.0860553588718176, 'learning_rate/qwen_vl_interface': 1.9733836527252885e-06, 'learning_rate/action_model': 1.9733836527252886e-05, 'epoch': 0.29}) +[2026-06-24 13:47:34,278][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.029607191681861877, 'train/grad_norm_pre_clip': 25.6622371673584, 'timing/data': 0.0003206131514161825, 'timing/model': 1.3597778999246657, 'learning_rate/qwen_vl_interface': 1.9706145268412843e-06, 'learning_rate/action_model': 1.9706145268412843e-05, 'epoch': 0.29}) +[2026-06-24 13:47:35,367][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.0013800561428070068, 'train/grad_norm_pre_clip': 0.6297529935836792, 'timing/data': 0.0003782310523092747, 'timing/model': 1.0866833301261067, 'learning_rate/qwen_vl_interface': 1.9678475289149096e-06, 'learning_rate/action_model': 1.9678475289149096e-05, 'epoch': 0.29}) +[2026-06-24 13:47:36,458][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.050993047654628754, 'train/grad_norm_pre_clip': 78.5237045288086, 'timing/data': 0.0003168350085616112, 'timing/model': 1.0895224560517818, 'learning_rate/qwen_vl_interface': 1.9650826607416403e-06, 'learning_rate/action_model': 1.9650826607416403e-05, 'epoch': 0.29}) +[2026-06-24 13:47:37,546][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.07825993746519089, 'train/grad_norm_pre_clip': 74.58338928222656, 'timing/data': 0.0003288381267338991, 'timing/model': 1.086276886984706, 'learning_rate/qwen_vl_interface': 1.9623199241155684e-06, 'learning_rate/action_model': 1.9623199241155683e-05, 'epoch': 0.29}) +[2026-06-24 13:47:38,633][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3000 +[2026-06-24 13:48:03,506][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3000 +[2026-06-24 13:48:03,507][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3000 +[2026-06-24 13:48:36,613][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3000 +[2026-06-24 13:54:42,248][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3000 +[2026-06-24 13:54:42,249][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.030007896944880486, 'train/grad_norm_pre_clip': 19.357791900634766, 'eval/action_loss/seconds': 24.871790723875165, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.060016099363565445, 'eval/latency_2/loss': 0.060016099363565445, 'eval/flappy/loss': 0.060016099363565445, 'eval/flappy/latency_2/loss': 0.060016099363565445, 'eval/flap_precision': 0.9643835616438357, 'eval/flap_recall': 0.9552238805970149, 'eval/flap_f1': 0.9597818677573279, 'eval/flap_rate_model': 0.20277777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9597818677573278, 'eval/latency_2/flap_precision': 0.9643835616438357, 'eval/latency_2/flap_recall': 0.9552238805970149, 'eval/latency_2/flap_f1': 0.9597818677573279, 'eval/latency_2/flap_rate_model': 0.20277777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9597818677573278, 'eval/action_classification/data_wait_seconds': 0.345988005399704, 'eval/action_classification/predict_seconds': 32.672306060791016, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 282.3200000000154, 'rl_games_eval/mid_train/mean_length': 2293.0, 'rl_games_eval/mid_train/std_reward': 199.98049304870963, 'rl_games_eval/mid_train/std_length': 1610.6818431956076, 'rl_games_eval/mid_train/macro_mean_reward': 282.3200000000154, 'rl_games_eval/mid_train/macro_mean_length': 2293.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 282.3200000000154, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2293.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 199.98049304870963, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1610.6818431956076, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 282.3200000000154, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00032776594161987305, 'timing/model': 1.0860338809434325, 'learning_rate/qwen_vl_interface': 1.959559320829403e-06, 'learning_rate/action_model': 1.9595593208294028e-05, 'epoch': 0.29}) +[2026-06-24 13:54:42,250][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_3000_state +[2026-06-24 13:54:42,252][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 13:54:44,887][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_3000_state/model.safetensors +[2026-06-24 13:54:50,579][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_3000_state/optimizer.bin +[2026-06-24 13:54:50,583][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_3000_state/sampler.bin +[2026-06-24 13:54:50,583][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-24 13:54:50,584][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-24 13:54:50,649][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-24 13:54:50,649][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 13:54:50,659][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 13:54:51,735][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.07001372426748276, 'train/grad_norm_pre_clip': 93.47711944580078, 'timing/data': 0.00041369698010385036, 'timing/model': 1.0704484099987894, 'learning_rate/qwen_vl_interface': 1.9568008526744718e-06, 'learning_rate/action_model': 1.956800852674472e-05, 'epoch': 0.29}) +[2026-06-24 13:54:53,725][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.25021040439605713, 'train/grad_norm_pre_clip': 102.76900482177734, 'timing/data': 0.0035082921385765076, 'timing/model': 1.9826455998700112, 'learning_rate/qwen_vl_interface': 1.9540445214407115e-06, 'learning_rate/action_model': 1.9540445214407112e-05, 'epoch': 0.29}) +[2026-06-24 13:54:54,808][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.10860250890254974, 'train/grad_norm_pre_clip': 47.819026947021484, 'timing/data': 0.0005023838020861149, 'timing/model': 1.080618903040886, 'learning_rate/qwen_vl_interface': 1.9512903289166797e-06, 'learning_rate/action_model': 1.9512903289166798e-05, 'epoch': 0.29}) +[2026-06-24 13:54:55,899][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.11412369459867477, 'train/grad_norm_pre_clip': 80.12425994873047, 'timing/data': 0.0006638069171458483, 'timing/model': 1.0884384061209857, 'learning_rate/qwen_vl_interface': 1.9485382768895404e-06, 'learning_rate/action_model': 1.9485382768895404e-05, 'epoch': 0.29}) +[2026-06-24 13:54:56,978][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.09650418162345886, 'train/grad_norm_pre_clip': 48.72385025024414, 'timing/data': 0.0004217319656163454, 'timing/model': 1.0771267421077937, 'learning_rate/qwen_vl_interface': 1.945788367145066e-06, 'learning_rate/action_model': 1.945788367145066e-05, 'epoch': 0.29}) +[2026-06-24 13:54:58,054][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.01826409064233303, 'train/grad_norm_pre_clip': 21.059673309326172, 'timing/data': 0.00044906395487487316, 'timing/model': 1.0743493949994445, 'learning_rate/qwen_vl_interface': 1.9430406014676508e-06, 'learning_rate/action_model': 1.943040601467651e-05, 'epoch': 0.29}) +[2026-06-24 13:54:59,132][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.09824761003255844, 'train/grad_norm_pre_clip': 35.98249053955078, 'timing/data': 0.00043088989332318306, 'timing/model': 1.0761742438189685, 'learning_rate/qwen_vl_interface': 1.9402949816402843e-06, 'learning_rate/action_model': 1.940294981640284e-05, 'epoch': 0.29}) +[2026-06-24 13:55:00,212][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.2571345865726471, 'train/grad_norm_pre_clip': 60.30637741088867, 'timing/data': 0.0003896220587193966, 'timing/model': 1.0782216920051724, 'learning_rate/qwen_vl_interface': 1.937551509444571e-06, 'learning_rate/action_model': 1.937551509444571e-05, 'epoch': 0.29}) +[2026-06-24 13:55:01,292][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.023498397320508957, 'train/grad_norm_pre_clip': 25.809524536132812, 'timing/data': 0.00041284598410129547, 'timing/model': 1.077597149880603, 'learning_rate/qwen_vl_interface': 1.9348101866607225e-06, 'learning_rate/action_model': 1.9348101866607223e-05, 'epoch': 0.29}) +[2026-06-24 13:55:02,371][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.041610751301050186, 'train/grad_norm_pre_clip': 40.439476013183594, 'timing/data': 0.0004339260049164295, 'timing/model': 1.0769555850420147, 'learning_rate/qwen_vl_interface': 1.9320710150675497e-06, 'learning_rate/action_model': 1.9320710150675496e-05, 'epoch': 0.29}) +[2026-06-24 13:55:03,449][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.06639180332422256, 'train/grad_norm_pre_clip': 19.48166275024414, 'timing/data': 0.0003950330428779125, 'timing/model': 1.0762006170116365, 'learning_rate/qwen_vl_interface': 1.929333996442473e-06, 'learning_rate/action_model': 1.9293339964424728e-05, 'epoch': 0.29}) +[2026-06-24 13:55:04,526][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.07955829799175262, 'train/grad_norm_pre_clip': 48.75249481201172, 'timing/data': 0.000414959155023098, 'timing/model': 1.0753849549219012, 'learning_rate/qwen_vl_interface': 1.9265991325615135e-06, 'learning_rate/action_model': 1.9265991325615135e-05, 'epoch': 0.29}) +[2026-06-24 13:55:05,607][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.05375539883971214, 'train/grad_norm_pre_clip': 31.26976776123047, 'timing/data': 0.0004060228820890188, 'timing/model': 1.079231824958697, 'learning_rate/qwen_vl_interface': 1.9238664251992975e-06, 'learning_rate/action_model': 1.9238664251992973e-05, 'epoch': 0.29}) +[2026-06-24 13:55:06,692][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.22334495186805725, 'train/grad_norm_pre_clip': 71.7672348022461, 'timing/data': 0.000504929106682539, 'timing/model': 1.0826696250587702, 'learning_rate/qwen_vl_interface': 1.921135876129045e-06, 'learning_rate/action_model': 1.921135876129045e-05, 'epoch': 0.29}) +[2026-06-24 13:55:07,779][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.1673164963722229, 'train/grad_norm_pre_clip': 620.8809814453125, 'timing/data': 0.0006103969644755125, 'timing/model': 1.0852967819664627, 'learning_rate/qwen_vl_interface': 1.9184074871225816e-06, 'learning_rate/action_model': 1.9184074871225815e-05, 'epoch': 0.29}) +[2026-06-24 13:55:08,863][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.024355020374059677, 'train/grad_norm_pre_clip': 40.49949645996094, 'timing/data': 0.0005122320726513863, 'timing/model': 1.0818755740765482, 'learning_rate/qwen_vl_interface': 1.9156812599503308e-06, 'learning_rate/action_model': 1.915681259950331e-05, 'epoch': 0.29}) +[2026-06-24 13:55:10,184][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.0880550667643547, 'train/grad_norm_pre_clip': 55.362396240234375, 'timing/data': 0.0004321020096540451, 'timing/model': 1.3187259279657155, 'learning_rate/qwen_vl_interface': 1.9129571963813087e-06, 'learning_rate/action_model': 1.9129571963813086e-05, 'epoch': 0.29}) +[2026-06-24 13:55:11,264][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.22216591238975525, 'train/grad_norm_pre_clip': 42.589603424072266, 'timing/data': 0.00037783989682793617, 'timing/model': 1.0784883319865912, 'learning_rate/qwen_vl_interface': 1.9102352981831318e-06, 'learning_rate/action_model': 1.9102352981831316e-05, 'epoch': 0.29}) +[2026-06-24 13:55:12,350][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.001628646394237876, 'train/grad_norm_pre_clip': 2.1236448287963867, 'timing/data': 0.0004258409608155489, 'timing/model': 1.0840672960039228, 'learning_rate/qwen_vl_interface': 1.9075155671220106e-06, 'learning_rate/action_model': 1.907515567122011e-05, 'epoch': 0.29}) +[2026-06-24 13:55:13,457][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.19276857376098633, 'train/grad_norm_pre_clip': 54.5482177734375, 'timing/data': 0.0003795728553086519, 'timing/model': 1.1052485960535705, 'learning_rate/qwen_vl_interface': 1.9047980049627513e-06, 'learning_rate/action_model': 1.904798004962751e-05, 'epoch': 0.29}) +[2026-06-24 13:55:14,545][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.2828776240348816, 'train/grad_norm_pre_clip': 86.74559020996094, 'timing/data': 0.0006971901748329401, 'timing/model': 1.0856346420478076, 'learning_rate/qwen_vl_interface': 1.9020826134687464e-06, 'learning_rate/action_model': 1.9020826134687465e-05, 'epoch': 0.29}) +[2026-06-24 13:55:15,625][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.08674398064613342, 'train/grad_norm_pre_clip': 109.19612121582031, 'timing/data': 0.0004741409793496132, 'timing/model': 1.0783831970766187, 'learning_rate/qwen_vl_interface': 1.8993693944019852e-06, 'learning_rate/action_model': 1.899369394401985e-05, 'epoch': 0.29}) +[2026-06-24 13:55:16,706][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.3117498755455017, 'train/grad_norm_pre_clip': 165.92776489257812, 'timing/data': 0.00044082803651690483, 'timing/model': 1.0793349558953196, 'learning_rate/qwen_vl_interface': 1.896658349523049e-06, 'learning_rate/action_model': 1.896658349523049e-05, 'epoch': 0.29}) +[2026-06-24 13:55:17,796][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.08588061481714249, 'train/grad_norm_pre_clip': 59.29643249511719, 'timing/data': 0.0005162500310689211, 'timing/model': 1.0878601861186326, 'learning_rate/qwen_vl_interface': 1.8939494805911002e-06, 'learning_rate/action_model': 1.8939494805911003e-05, 'epoch': 0.29}) +[2026-06-24 13:55:18,901][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.030122043564915657, 'train/grad_norm_pre_clip': 19.492048263549805, 'timing/data': 0.0008467810694128275, 'timing/model': 1.1020112368278205, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 0.29}) +[2026-06-24 13:55:20,009][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.0009931266540661454, 'train/grad_norm_pre_clip': 0.9419644474983215, 'timing/data': 0.0005394131876528263, 'timing/model': 1.1053938739933074, 'learning_rate/qwen_vl_interface': 1.8885382775977876e-06, 'learning_rate/action_model': 1.8885382775977875e-05, 'epoch': 0.29}) +[2026-06-24 13:55:21,117][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.009377951733767986, 'train/grad_norm_pre_clip': 7.9442009925842285, 'timing/data': 0.0004557170905172825, 'timing/model': 1.1059892228804529, 'learning_rate/qwen_vl_interface': 1.8858359470476897e-06, 'learning_rate/action_model': 1.8858359470476895e-05, 'epoch': 0.29}) +[2026-06-24 13:55:22,216][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.0847267434000969, 'train/grad_norm_pre_clip': 51.5768928527832, 'timing/data': 0.0006734458729624748, 'timing/model': 1.0965776899829507, 'learning_rate/qwen_vl_interface': 1.8831357994671245e-06, 'learning_rate/action_model': 1.8831357994671244e-05, 'epoch': 0.29}) +[2026-06-24 13:55:23,304][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.14857015013694763, 'train/grad_norm_pre_clip': 44.17833709716797, 'timing/data': 0.0004685209132730961, 'timing/model': 1.0862212011124939, 'learning_rate/qwen_vl_interface': 1.8804378366081841e-06, 'learning_rate/action_model': 1.880437836608184e-05, 'epoch': 0.29}) +[2026-06-24 13:55:24,390][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.0034718732349574566, 'train/grad_norm_pre_clip': 4.030407428741455, 'timing/data': 0.0003425239119678736, 'timing/model': 1.0842553381808102, 'learning_rate/qwen_vl_interface': 1.8777420602215484e-06, 'learning_rate/action_model': 1.8777420602215484e-05, 'epoch': 0.29}) +[2026-06-24 13:55:25,469][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.1323707550764084, 'train/grad_norm_pre_clip': 25.593860626220703, 'timing/data': 0.00033451803028583527, 'timing/model': 1.0771097810938954, 'learning_rate/qwen_vl_interface': 1.8750484720564794e-06, 'learning_rate/action_model': 1.875048472056479e-05, 'epoch': 0.29}) +[2026-06-24 13:55:26,548][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.0712645947933197, 'train/grad_norm_pre_clip': 34.144927978515625, 'timing/data': 0.00034801382571458817, 'timing/model': 1.0779773630201817, 'learning_rate/qwen_vl_interface': 1.872357073860814e-06, 'learning_rate/action_model': 1.872357073860814e-05, 'epoch': 0.29}) +[2026-06-24 13:55:27,919][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.23681367933750153, 'train/grad_norm_pre_clip': 55.99654006958008, 'timing/data': 0.00030998210422694683, 'timing/model': 1.3692321749404073, 'learning_rate/qwen_vl_interface': 1.8696678673809726e-06, 'learning_rate/action_model': 1.8696678673809725e-05, 'epoch': 0.29}) +[2026-06-24 13:55:29,000][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.10271962732076645, 'train/grad_norm_pre_clip': 78.50630950927734, 'timing/data': 0.0005122930742800236, 'timing/model': 1.0786418330390006, 'learning_rate/qwen_vl_interface': 1.8669808543619527e-06, 'learning_rate/action_model': 1.8669808543619526e-05, 'epoch': 0.29}) +[2026-06-24 13:55:30,077][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.03938785567879677, 'train/grad_norm_pre_clip': 35.94198989868164, 'timing/data': 0.00032185507006943226, 'timing/model': 1.0762851659674197, 'learning_rate/qwen_vl_interface': 1.8642960365473287e-06, 'learning_rate/action_model': 1.8642960365473285e-05, 'epoch': 0.29}) +[2026-06-24 13:55:31,155][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.028184054419398308, 'train/grad_norm_pre_clip': 11.262423515319824, 'timing/data': 0.0003196301404386759, 'timing/model': 1.0756487709004432, 'learning_rate/qwen_vl_interface': 1.8616134156792473e-06, 'learning_rate/action_model': 1.8616134156792472e-05, 'epoch': 0.29}) +[2026-06-24 13:55:32,237][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.045303527265787125, 'train/grad_norm_pre_clip': 48.452247619628906, 'timing/data': 0.0003081990871578455, 'timing/model': 1.0806731020566076, 'learning_rate/qwen_vl_interface': 1.8589329934984332e-06, 'learning_rate/action_model': 1.858932993498433e-05, 'epoch': 0.29}) +[2026-06-24 13:55:33,319][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.009926448576152325, 'train/grad_norm_pre_clip': 16.198049545288086, 'timing/data': 0.0003398088738322258, 'timing/model': 1.080526257166639, 'learning_rate/qwen_vl_interface': 1.856254771744185e-06, 'learning_rate/action_model': 1.856254771744185e-05, 'epoch': 0.29}) +[2026-06-24 13:55:34,400][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.012928676791489124, 'train/grad_norm_pre_clip': 17.197866439819336, 'timing/data': 0.0003265638370066881, 'timing/model': 1.0798523109406233, 'learning_rate/qwen_vl_interface': 1.8535787521543685e-06, 'learning_rate/action_model': 1.8535787521543685e-05, 'epoch': 0.29}) +[2026-06-24 13:55:35,482][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.02882768027484417, 'train/grad_norm_pre_clip': 18.842666625976562, 'timing/data': 0.0004786399658769369, 'timing/model': 1.0796232919674367, 'learning_rate/qwen_vl_interface': 1.8509049364654252e-06, 'learning_rate/action_model': 1.8509049364654253e-05, 'epoch': 0.29}) +[2026-06-24 13:55:36,565][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.07887925952672958, 'train/grad_norm_pre_clip': 26.36455726623535, 'timing/data': 0.0003328148741275072, 'timing/model': 1.081623530946672, 'learning_rate/qwen_vl_interface': 1.8482333264123648e-06, 'learning_rate/action_model': 1.8482333264123648e-05, 'epoch': 0.29}) +[2026-06-24 13:55:37,648][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.2183964103460312, 'train/grad_norm_pre_clip': 53.18422317504883, 'timing/data': 0.00039556296542286873, 'timing/model': 1.0811800381634384, 'learning_rate/qwen_vl_interface': 1.8455639237287688e-06, 'learning_rate/action_model': 1.8455639237287688e-05, 'epoch': 0.29}) +[2026-06-24 13:55:38,727][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.17741750180721283, 'train/grad_norm_pre_clip': 141.891845703125, 'timing/data': 0.0003461800515651703, 'timing/model': 1.0779384090565145, 'learning_rate/qwen_vl_interface': 1.8428967301467787e-06, 'learning_rate/action_model': 1.8428967301467787e-05, 'epoch': 0.29}) +[2026-06-24 13:55:39,810][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.03502080589532852, 'train/grad_norm_pre_clip': 22.41425132751465, 'timing/data': 0.0003700649831444025, 'timing/model': 1.0805848219897598, 'learning_rate/qwen_vl_interface': 1.840231747397111e-06, 'learning_rate/action_model': 1.840231747397111e-05, 'epoch': 0.29}) +[2026-06-24 13:55:40,891][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.14504610002040863, 'train/grad_norm_pre_clip': 83.86357879638672, 'timing/data': 0.000306346220895648, 'timing/model': 1.0802982938475907, 'learning_rate/qwen_vl_interface': 1.8375689772090443e-06, 'learning_rate/action_model': 1.8375689772090444e-05, 'epoch': 0.29}) +[2026-06-24 13:55:41,975][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.13565589487552643, 'train/grad_norm_pre_clip': 74.09562683105469, 'timing/data': 0.00031361915171146393, 'timing/model': 1.0819618301466107, 'learning_rate/qwen_vl_interface': 1.8349084213104168e-06, 'learning_rate/action_model': 1.834908421310417e-05, 'epoch': 0.29}) +[2026-06-24 13:55:43,060][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.03683144971728325, 'train/grad_norm_pre_clip': 15.302811622619629, 'timing/data': 0.0005208291113376617, 'timing/model': 1.0825408699456602, 'learning_rate/qwen_vl_interface': 1.8322500814276406e-06, 'learning_rate/action_model': 1.8322500814276404e-05, 'epoch': 0.29}) +[2026-06-24 13:55:44,352][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.15238341689109802, 'train/grad_norm_pre_clip': 63.01692199707031, 'timing/data': 0.0003822490107268095, 'timing/model': 1.2907332889735699, 'learning_rate/qwen_vl_interface': 1.8295939592856807e-06, 'learning_rate/action_model': 1.8295939592856807e-05, 'epoch': 0.29}) +[2026-06-24 13:55:45,437][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.07346691936254501, 'train/grad_norm_pre_clip': 42.21377182006836, 'timing/data': 0.00030752713792026043, 'timing/model': 1.0835589540656656, 'learning_rate/qwen_vl_interface': 1.8269400566080628e-06, 'learning_rate/action_model': 1.8269400566080628e-05, 'epoch': 0.3}) +[2026-06-24 13:55:46,524][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.009179691784083843, 'train/grad_norm_pre_clip': 11.046915054321289, 'timing/data': 0.0003352200146764517, 'timing/model': 1.0848636759910733, 'learning_rate/qwen_vl_interface': 1.8242883751168814e-06, 'learning_rate/action_model': 1.8242883751168812e-05, 'epoch': 0.3}) +[2026-06-24 13:55:47,609][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.0546940416097641, 'train/grad_norm_pre_clip': 25.517745971679688, 'timing/data': 0.0005134949460625648, 'timing/model': 1.0833841369021684, 'learning_rate/qwen_vl_interface': 1.8216389165327797e-06, 'learning_rate/action_model': 1.8216389165327798e-05, 'epoch': 0.3}) +[2026-06-24 13:55:48,698][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.03838358074426651, 'train/grad_norm_pre_clip': 23.77096176147461, 'timing/data': 0.00034301518462598324, 'timing/model': 1.087130376836285, 'learning_rate/qwen_vl_interface': 1.818991682574964e-06, 'learning_rate/action_model': 1.818991682574964e-05, 'epoch': 0.3}) +[2026-06-24 13:55:49,783][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.1330314427614212, 'train/grad_norm_pre_clip': 40.278263092041016, 'timing/data': 0.00033913692459464073, 'timing/model': 1.083313865819946, 'learning_rate/qwen_vl_interface': 1.8163466749611967e-06, 'learning_rate/action_model': 1.8163466749611965e-05, 'epoch': 0.3}) +[2026-06-24 13:55:50,869][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.024897700175642967, 'train/grad_norm_pre_clip': 6.081806182861328, 'timing/data': 0.0003116161096841097, 'timing/model': 1.0844220300205052, 'learning_rate/qwen_vl_interface': 1.8137038954077923e-06, 'learning_rate/action_model': 1.8137038954077922e-05, 'epoch': 0.3}) +[2026-06-24 13:55:51,961][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.1005759984254837, 'train/grad_norm_pre_clip': 83.44767761230469, 'timing/data': 0.00034413603134453297, 'timing/model': 1.0906155582051724, 'learning_rate/qwen_vl_interface': 1.8110633456296225e-06, 'learning_rate/action_model': 1.8110633456296224e-05, 'epoch': 0.3}) +[2026-06-24 13:55:53,042][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.005190160591155291, 'train/grad_norm_pre_clip': 4.108388423919678, 'timing/data': 0.0003144610673189163, 'timing/model': 1.0796997679863125, 'learning_rate/qwen_vl_interface': 1.8084250273401116e-06, 'learning_rate/action_model': 1.8084250273401116e-05, 'epoch': 0.3}) +[2026-06-24 13:55:54,128][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.030208151787519455, 'train/grad_norm_pre_clip': 16.848148345947266, 'timing/data': 0.0004720871802419424, 'timing/model': 1.083656664006412, 'learning_rate/qwen_vl_interface': 1.8057889422512379e-06, 'learning_rate/action_model': 1.805788942251238e-05, 'epoch': 0.3}) +[2026-06-24 13:55:55,212][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.06170095130801201, 'train/grad_norm_pre_clip': 39.15536117553711, 'timing/data': 0.00046702707186341286, 'timing/model': 1.082941550994292, 'learning_rate/qwen_vl_interface': 1.803155092073524e-06, 'learning_rate/action_model': 1.803155092073524e-05, 'epoch': 0.3}) +[2026-06-24 13:55:56,303][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.021159512922167778, 'train/grad_norm_pre_clip': 16.935348510742188, 'timing/data': 0.0003436650149524212, 'timing/model': 1.0889346331823617, 'learning_rate/qwen_vl_interface': 1.8005234785160493e-06, 'learning_rate/action_model': 1.8005234785160495e-05, 'epoch': 0.3}) +[2026-06-24 13:55:57,388][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.2682610750198364, 'train/grad_norm_pre_clip': 80.39575958251953, 'timing/data': 0.00032185507006943226, 'timing/model': 1.083291961112991, 'learning_rate/qwen_vl_interface': 1.7978941032864406e-06, 'learning_rate/action_model': 1.7978941032864404e-05, 'epoch': 0.3}) +[2026-06-24 13:55:58,477][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.10515885800123215, 'train/grad_norm_pre_clip': 41.41133499145508, 'timing/data': 0.0003673909232020378, 'timing/model': 1.088113090954721, 'learning_rate/qwen_vl_interface': 1.7952669680908676e-06, 'learning_rate/action_model': 1.7952669680908675e-05, 'epoch': 0.3}) +[2026-06-24 13:55:59,562][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.1596967577934265, 'train/grad_norm_pre_clip': 99.86528778076172, 'timing/data': 0.0003225060645490885, 'timing/model': 1.0834901239722967, 'learning_rate/qwen_vl_interface': 1.7926420746340522e-06, 'learning_rate/action_model': 1.7926420746340523e-05, 'epoch': 0.3}) +[2026-06-24 13:56:00,653][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.0032847661059349775, 'train/grad_norm_pre_clip': 2.4560341835021973, 'timing/data': 0.0003280660603195429, 'timing/model': 1.0886872499249876, 'learning_rate/qwen_vl_interface': 1.790019424619258e-06, 'learning_rate/action_model': 1.7900194246192578e-05, 'epoch': 0.3}) +[2026-06-24 13:56:02,010][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.008275684900581837, 'train/grad_norm_pre_clip': 9.035222053527832, 'timing/data': 0.0003408198244869709, 'timing/model': 1.3555128059815615, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 0.3}) +[2026-06-24 13:56:03,098][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.023546310141682625, 'train/grad_norm_pre_clip': 36.24882507324219, 'timing/data': 0.0003494969569146633, 'timing/model': 1.0861298069357872, 'learning_rate/qwen_vl_interface': 1.7847808617215162e-06, 'learning_rate/action_model': 1.7847808617215162e-05, 'epoch': 0.3}) +[2026-06-24 13:56:04,183][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.0808364674448967, 'train/grad_norm_pre_clip': 68.14269256591797, 'timing/data': 0.00032733497209846973, 'timing/model': 1.0834357261192054, 'learning_rate/qwen_vl_interface': 1.7821649522378134e-06, 'learning_rate/action_model': 1.7821649522378134e-05, 'epoch': 0.3}) +[2026-06-24 13:56:05,269][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.029136313125491142, 'train/grad_norm_pre_clip': 12.247424125671387, 'timing/data': 0.0003213339950889349, 'timing/model': 1.0853530489839613, 'learning_rate/qwen_vl_interface': 1.779551292994625e-06, 'learning_rate/action_model': 1.779551292994625e-05, 'epoch': 0.3}) +[2026-06-24 13:56:06,355][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.004475349094718695, 'train/grad_norm_pre_clip': 5.897462368011475, 'timing/data': 0.0003227971028536558, 'timing/model': 1.0844401840586215, 'learning_rate/qwen_vl_interface': 1.7769398856879217e-06, 'learning_rate/action_model': 1.7769398856879215e-05, 'epoch': 0.3}) +[2026-06-24 13:56:07,444][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.026084642857313156, 'train/grad_norm_pre_clip': 29.91629981994629, 'timing/data': 0.00033035106025636196, 'timing/model': 1.0872109308838844, 'learning_rate/qwen_vl_interface': 1.7743307320122227e-06, 'learning_rate/action_model': 1.7743307320122227e-05, 'epoch': 0.3}) +[2026-06-24 13:56:08,530][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.12213233858346939, 'train/grad_norm_pre_clip': 59.11317443847656, 'timing/data': 0.000337773934006691, 'timing/model': 1.0845546899363399, 'learning_rate/qwen_vl_interface': 1.7717238336605771e-06, 'learning_rate/action_model': 1.771723833660577e-05, 'epoch': 0.3}) +[2026-06-24 13:56:09,617][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.08823823928833008, 'train/grad_norm_pre_clip': 23.05042266845703, 'timing/data': 0.00032253609970211983, 'timing/model': 1.0850490310695022, 'learning_rate/qwen_vl_interface': 1.7691191923245702e-06, 'learning_rate/action_model': 1.7691191923245702e-05, 'epoch': 0.3}) +[2026-06-24 13:56:10,704][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.04973017796874046, 'train/grad_norm_pre_clip': 45.07016372680664, 'timing/data': 0.0003292490728199482, 'timing/model': 1.085001772036776, 'learning_rate/qwen_vl_interface': 1.7665168096943299e-06, 'learning_rate/action_model': 1.7665168096943298e-05, 'epoch': 0.3}) +[2026-06-24 13:56:11,793][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.09543929994106293, 'train/grad_norm_pre_clip': 86.58905029296875, 'timing/data': 0.0003146009985357523, 'timing/model': 1.088088012067601, 'learning_rate/qwen_vl_interface': 1.7639166874585117e-06, 'learning_rate/action_model': 1.7639166874585115e-05, 'epoch': 0.3}) +[2026-06-24 13:56:12,881][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.08864095062017441, 'train/grad_norm_pre_clip': 72.07654571533203, 'timing/data': 0.0003145909868180752, 'timing/model': 1.086247876053676, 'learning_rate/qwen_vl_interface': 1.7613188273043067e-06, 'learning_rate/action_model': 1.7613188273043066e-05, 'epoch': 0.3}) +[2026-06-24 13:56:13,972][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.1605086326599121, 'train/grad_norm_pre_clip': 98.26020812988281, 'timing/data': 0.00037681800313293934, 'timing/model': 1.0894907689653337, 'learning_rate/qwen_vl_interface': 1.7587232309174402e-06, 'learning_rate/action_model': 1.75872323091744e-05, 'epoch': 0.3}) +[2026-06-24 13:56:15,061][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.15552251040935516, 'train/grad_norm_pre_clip': 66.24107360839844, 'timing/data': 0.0003351401537656784, 'timing/model': 1.0873162848874927, 'learning_rate/qwen_vl_interface': 1.7561298999821632e-06, 'learning_rate/action_model': 1.7561298999821633e-05, 'epoch': 0.3}) +[2026-06-24 13:56:16,151][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.000652854738291353, 'train/grad_norm_pre_clip': 0.4388345777988434, 'timing/data': 0.00035559898242354393, 'timing/model': 1.0883411809336394, 'learning_rate/qwen_vl_interface': 1.753538836181262e-06, 'learning_rate/action_model': 1.753538836181262e-05, 'epoch': 0.3}) +[2026-06-24 13:56:17,239][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.0005108648329041898, 'train/grad_norm_pre_clip': 0.7122142314910889, 'timing/data': 0.00036858185194432735, 'timing/model': 1.086675371043384, 'learning_rate/qwen_vl_interface': 1.7509500411960491e-06, 'learning_rate/action_model': 1.750950041196049e-05, 'epoch': 0.3}) +[2026-06-24 13:56:18,545][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.10355085879564285, 'train/grad_norm_pre_clip': 95.02924346923828, 'timing/data': 0.0003162750508636236, 'timing/model': 1.3041811648290604, 'learning_rate/qwen_vl_interface': 1.7483635167063675e-06, 'learning_rate/action_model': 1.7483635167063675e-05, 'epoch': 0.3}) +[2026-06-24 13:56:19,634][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.16529712080955505, 'train/grad_norm_pre_clip': 51.55927658081055, 'timing/data': 0.0003490261733531952, 'timing/model': 1.0876086470670998, 'learning_rate/qwen_vl_interface': 1.7457792643905822e-06, 'learning_rate/action_model': 1.7457792643905823e-05, 'epoch': 0.3}) +[2026-06-24 13:56:20,724][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.006368579808622599, 'train/grad_norm_pre_clip': 8.547043800354004, 'timing/data': 0.0003089199308305979, 'timing/model': 1.088508928893134, 'learning_rate/qwen_vl_interface': 1.743197285925587e-06, 'learning_rate/action_model': 1.7431972859255872e-05, 'epoch': 0.3}) +[2026-06-24 13:56:21,820][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.377663791179657, 'train/grad_norm_pre_clip': 23.749431610107422, 'timing/data': 0.0003163449000567198, 'timing/model': 1.0940888051409274, 'learning_rate/qwen_vl_interface': 1.7406175829868018e-06, 'learning_rate/action_model': 1.740617582986802e-05, 'epoch': 0.3}) +[2026-06-24 13:56:22,910][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.04407215118408203, 'train/grad_norm_pre_clip': 21.659414291381836, 'timing/data': 0.00030528404749929905, 'timing/model': 1.0881311611738056, 'learning_rate/qwen_vl_interface': 1.7380401572481646e-06, 'learning_rate/action_model': 1.7380401572481644e-05, 'epoch': 0.3}) +[2026-06-24 13:56:23,994][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.26664870977401733, 'train/grad_norm_pre_clip': 81.49919128417969, 'timing/data': 0.00031676492653787136, 'timing/model': 1.082588115008548, 'learning_rate/qwen_vl_interface': 1.7354650103821403e-06, 'learning_rate/action_model': 1.7354650103821403e-05, 'epoch': 0.3}) +[2026-06-24 13:56:25,085][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.08351869881153107, 'train/grad_norm_pre_clip': 120.97588348388672, 'timing/data': 0.00039689592085778713, 'timing/model': 1.089360031997785, 'learning_rate/qwen_vl_interface': 1.7328921440597157e-06, 'learning_rate/action_model': 1.7328921440597157e-05, 'epoch': 0.3}) +[2026-06-24 13:56:26,172][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.17046736180782318, 'train/grad_norm_pre_clip': 229.84986877441406, 'timing/data': 0.0003114151768386364, 'timing/model': 1.0850885058753192, 'learning_rate/qwen_vl_interface': 1.7303215599503925e-06, 'learning_rate/action_model': 1.7303215599503925e-05, 'epoch': 0.3}) +[2026-06-24 13:56:27,260][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.06967493146657944, 'train/grad_norm_pre_clip': 59.40389633178711, 'timing/data': 0.00033341581001877785, 'timing/model': 1.086361767956987, 'learning_rate/qwen_vl_interface': 1.7277532597221966e-06, 'learning_rate/action_model': 1.7277532597221965e-05, 'epoch': 0.3}) +[2026-06-24 13:56:28,356][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.16806191205978394, 'train/grad_norm_pre_clip': 30.513330459594727, 'timing/data': 0.0003446179907768965, 'timing/model': 1.0939343888312578, 'learning_rate/qwen_vl_interface': 1.7251872450416704e-06, 'learning_rate/action_model': 1.7251872450416702e-05, 'epoch': 0.3}) +[2026-06-24 13:56:29,477][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.02942618355154991, 'train/grad_norm_pre_clip': 33.2340202331543, 'timing/data': 0.0004970240406692028, 'timing/model': 1.1186830231454223, 'learning_rate/qwen_vl_interface': 1.7226235175738717e-06, 'learning_rate/action_model': 1.7226235175738717e-05, 'epoch': 0.3}) +[2026-06-24 13:56:30,799][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.26288118958473206, 'train/grad_norm_pre_clip': 126.13792419433594, 'timing/data': 0.00034471810795366764, 'timing/model': 1.3208019360899925, 'learning_rate/qwen_vl_interface': 1.7200620789823769e-06, 'learning_rate/action_model': 1.7200620789823767e-05, 'epoch': 0.3}) +[2026-06-24 13:56:31,888][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.12434608489274979, 'train/grad_norm_pre_clip': 61.563541412353516, 'timing/data': 0.0003250108566135168, 'timing/model': 1.087103886064142, 'learning_rate/qwen_vl_interface': 1.717502930929276e-06, 'learning_rate/action_model': 1.717502930929276e-05, 'epoch': 0.3}) +[2026-06-24 13:56:32,977][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.046336714178323746, 'train/grad_norm_pre_clip': 35.923770904541016, 'timing/data': 0.00035437592305243015, 'timing/model': 1.0872323990333825, 'learning_rate/qwen_vl_interface': 1.7149460750751706e-06, 'learning_rate/action_model': 1.7149460750751704e-05, 'epoch': 0.3}) +[2026-06-24 13:56:34,066][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.16061894595623016, 'train/grad_norm_pre_clip': 85.18521881103516, 'timing/data': 0.00031693605706095695, 'timing/model': 1.0872283820062876, 'learning_rate/qwen_vl_interface': 1.7123915130791774e-06, 'learning_rate/action_model': 1.7123915130791772e-05, 'epoch': 0.3}) +[2026-06-24 13:56:35,153][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.07660895586013794, 'train/grad_norm_pre_clip': 65.54798889160156, 'timing/data': 0.0003163539804518223, 'timing/model': 1.0860292189754546, 'learning_rate/qwen_vl_interface': 1.709839246598925e-06, 'learning_rate/action_model': 1.709839246598925e-05, 'epoch': 0.3}) +[2026-06-24 13:56:36,536][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.0024159392341971397, 'train/grad_norm_pre_clip': 2.7332677841186523, 'timing/data': 0.0003270150627940893, 'timing/model': 1.3817300680093467, 'learning_rate/qwen_vl_interface': 1.7072892772905494e-06, 'learning_rate/action_model': 1.7072892772905494e-05, 'epoch': 0.3}) +[2026-06-24 13:56:37,625][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.0588298961520195, 'train/grad_norm_pre_clip': 67.6887435913086, 'timing/data': 0.00034025893546640873, 'timing/model': 1.0873901268932968, 'learning_rate/qwen_vl_interface': 1.7047416068086997e-06, 'learning_rate/action_model': 1.7047416068086998e-05, 'epoch': 0.3}) +[2026-06-24 13:56:38,715][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.001094745472073555, 'train/grad_norm_pre_clip': 1.0871667861938477, 'timing/data': 0.0003586940001696348, 'timing/model': 1.0881834488827735, 'learning_rate/qwen_vl_interface': 1.7021962368065317e-06, 'learning_rate/action_model': 1.7021962368065316e-05, 'epoch': 0.3}) +[2026-06-24 13:56:39,803][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.009072389453649521, 'train/grad_norm_pre_clip': 8.678695678710938, 'timing/data': 0.00031511206179857254, 'timing/model': 1.0859930941369385, 'learning_rate/qwen_vl_interface': 1.6996531689357062e-06, 'learning_rate/action_model': 1.6996531689357063e-05, 'epoch': 0.3}) +[2026-06-24 13:56:40,894][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.13424034416675568, 'train/grad_norm_pre_clip': 48.10997009277344, 'timing/data': 0.0003410610370337963, 'timing/model': 1.0895957539323717, 'learning_rate/qwen_vl_interface': 1.6971124048463942e-06, 'learning_rate/action_model': 1.6971124048463942e-05, 'epoch': 0.3}) +[2026-06-24 13:56:41,981][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.1549317091703415, 'train/grad_norm_pre_clip': 108.0633773803711, 'timing/data': 0.0006403529550880194, 'timing/model': 1.0847997330129147, 'learning_rate/qwen_vl_interface': 1.6945739461872692e-06, 'learning_rate/action_model': 1.6945739461872693e-05, 'epoch': 0.3}) +[2026-06-24 13:56:43,067][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.10965931415557861, 'train/grad_norm_pre_clip': 69.25953674316406, 'timing/data': 0.00031674490310251713, 'timing/model': 1.0842277889605612, 'learning_rate/qwen_vl_interface': 1.692037794605512e-06, 'learning_rate/action_model': 1.692037794605512e-05, 'epoch': 0.3}) +[2026-06-24 13:56:44,154][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.04712057113647461, 'train/grad_norm_pre_clip': 24.178003311157227, 'timing/data': 0.00033836509101092815, 'timing/model': 1.0857182901818305, 'learning_rate/qwen_vl_interface': 1.6895039517467999e-06, 'learning_rate/action_model': 1.6895039517467996e-05, 'epoch': 0.3}) +[2026-06-24 13:56:45,242][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.05929005146026611, 'train/grad_norm_pre_clip': 48.17316818237305, 'timing/data': 0.00034781312569975853, 'timing/model': 1.0869579000864178, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 0.3}) +[2026-06-24 13:56:46,333][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.013500496745109558, 'train/grad_norm_pre_clip': 8.746604919433594, 'timing/data': 0.00032153399661183357, 'timing/model': 1.0891096650157124, 'learning_rate/qwen_vl_interface': 1.6844431987737503e-06, 'learning_rate/action_model': 1.6844431987737502e-05, 'epoch': 0.3}) +[2026-06-24 13:56:47,422][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.22481364011764526, 'train/grad_norm_pre_clip': 175.95497131347656, 'timing/data': 0.000359745929017663, 'timing/model': 1.0874812849797308, 'learning_rate/qwen_vl_interface': 1.681916291943278e-06, 'learning_rate/action_model': 1.681916291943278e-05, 'epoch': 0.3}) +[2026-06-24 13:56:48,508][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.01968175545334816, 'train/grad_norm_pre_clip': 13.541749000549316, 'timing/data': 0.00032152398489415646, 'timing/model': 1.0845263130031526, 'learning_rate/qwen_vl_interface': 1.6793917004035885e-06, 'learning_rate/action_model': 1.6793917004035883e-05, 'epoch': 0.3}) +[2026-06-24 13:56:49,596][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.1908421516418457, 'train/grad_norm_pre_clip': 68.5038070678711, 'timing/data': 0.0003247500862926245, 'timing/model': 1.086011815117672, 'learning_rate/qwen_vl_interface': 1.6768694257928592e-06, 'learning_rate/action_model': 1.676869425792859e-05, 'epoch': 0.3}) +[2026-06-24 13:56:50,687][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.06120368093252182, 'train/grad_norm_pre_clip': 84.51387023925781, 'timing/data': 0.00035465601831674576, 'timing/model': 1.0892612808384001, 'learning_rate/qwen_vl_interface': 1.674349469747764e-06, 'learning_rate/action_model': 1.6743494697477642e-05, 'epoch': 0.3}) +[2026-06-24 13:56:51,780][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.37652456760406494, 'train/grad_norm_pre_clip': 71.38809967041016, 'timing/data': 0.00033243512734770775, 'timing/model': 1.0913330661132932, 'learning_rate/qwen_vl_interface': 1.6718318339034803e-06, 'learning_rate/action_model': 1.67183183390348e-05, 'epoch': 0.3}) +[2026-06-24 13:56:53,069][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.18519817292690277, 'train/grad_norm_pre_clip': 85.96904754638672, 'timing/data': 0.00031219678930938244, 'timing/model': 1.2881555878557265, 'learning_rate/qwen_vl_interface': 1.6693165198936705e-06, 'learning_rate/action_model': 1.6693165198936705e-05, 'epoch': 0.3}) +[2026-06-24 13:56:54,161][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.05903737246990204, 'train/grad_norm_pre_clip': 159.32032775878906, 'timing/data': 0.000340400030836463, 'timing/model': 1.089659352088347, 'learning_rate/qwen_vl_interface': 1.6668035293504957e-06, 'learning_rate/action_model': 1.666803529350496e-05, 'epoch': 0.3}) +[2026-06-24 13:56:55,249][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.030283816158771515, 'train/grad_norm_pre_clip': 24.946979522705078, 'timing/data': 0.00033476902171969414, 'timing/model': 1.086879258044064, 'learning_rate/qwen_vl_interface': 1.6642928639046089e-06, 'learning_rate/action_model': 1.6642928639046087e-05, 'epoch': 0.3}) +[2026-06-24 13:56:56,341][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.032483186572790146, 'train/grad_norm_pre_clip': 16.688980102539062, 'timing/data': 0.00031900894828140736, 'timing/model': 1.0906811829190701, 'learning_rate/qwen_vl_interface': 1.6617845251851542e-06, 'learning_rate/action_model': 1.661784525185154e-05, 'epoch': 0.3}) +[2026-06-24 13:56:57,429][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.2509493827819824, 'train/grad_norm_pre_clip': 41.30146408081055, 'timing/data': 0.00031819799914956093, 'timing/model': 1.085927892010659, 'learning_rate/qwen_vl_interface': 1.6592785148197635e-06, 'learning_rate/action_model': 1.6592785148197634e-05, 'epoch': 0.3}) +[2026-06-24 13:56:58,521][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.050433188676834106, 'train/grad_norm_pre_clip': 52.36721420288086, 'timing/data': 0.0003461299929767847, 'timing/model': 1.0909602190367877, 'learning_rate/qwen_vl_interface': 1.6567748344345602e-06, 'learning_rate/action_model': 1.65677483443456e-05, 'epoch': 0.3}) +[2026-06-24 13:56:59,611][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.03205464035272598, 'train/grad_norm_pre_clip': 19.169715881347656, 'timing/data': 0.0004614179488271475, 'timing/model': 1.0881259941961616, 'learning_rate/qwen_vl_interface': 1.6542734856541577e-06, 'learning_rate/action_model': 1.6542734856541576e-05, 'epoch': 0.3}) +[2026-06-24 13:57:00,704][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.14686043560504913, 'train/grad_norm_pre_clip': 98.54256439208984, 'timing/data': 0.00031493185088038445, 'timing/model': 1.0905422649811953, 'learning_rate/qwen_vl_interface': 1.6517744701016505e-06, 'learning_rate/action_model': 1.6517744701016505e-05, 'epoch': 0.3}) +[2026-06-24 13:57:01,797][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.027760295197367668, 'train/grad_norm_pre_clip': 62.2076416015625, 'timing/data': 0.0003130990080535412, 'timing/model': 1.0917290460783988, 'learning_rate/qwen_vl_interface': 1.649277789398624e-06, 'learning_rate/action_model': 1.649277789398624e-05, 'epoch': 0.3}) +[2026-06-24 13:57:02,897][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.09441354870796204, 'train/grad_norm_pre_clip': 65.85272979736328, 'timing/data': 0.0005107799079269171, 'timing/model': 1.0976125909946859, 'learning_rate/qwen_vl_interface': 1.6467834451651492e-06, 'learning_rate/action_model': 1.646783445165149e-05, 'epoch': 0.3}) +[2026-06-24 13:57:03,994][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.10574916005134583, 'train/grad_norm_pre_clip': 31.719301223754883, 'timing/data': 0.0005078339017927647, 'timing/model': 1.0959763380233198, 'learning_rate/qwen_vl_interface': 1.6442914390197755e-06, 'learning_rate/action_model': 1.6442914390197755e-05, 'epoch': 0.3}) +[2026-06-24 13:57:05,086][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.02285071089863777, 'train/grad_norm_pre_clip': 8.87972354888916, 'timing/data': 0.0003653171006590128, 'timing/model': 1.089436500100419, 'learning_rate/qwen_vl_interface': 1.6418017725795408e-06, 'learning_rate/action_model': 1.6418017725795408e-05, 'epoch': 0.3}) +[2026-06-24 13:57:06,179][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.09665023535490036, 'train/grad_norm_pre_clip': 96.27959442138672, 'timing/data': 0.0005102891009300947, 'timing/model': 1.0910759109538049, 'learning_rate/qwen_vl_interface': 1.6393144474599623e-06, 'learning_rate/action_model': 1.6393144474599622e-05, 'epoch': 0.3}) +[2026-06-24 13:57:07,272][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.02757364884018898, 'train/grad_norm_pre_clip': 30.6663761138916, 'timing/data': 0.0003226860426366329, 'timing/model': 1.0913400179706514, 'learning_rate/qwen_vl_interface': 1.6368294652750394e-06, 'learning_rate/action_model': 1.6368294652750393e-05, 'epoch': 0.3}) +[2026-06-24 13:57:08,364][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.003274860791862011, 'train/grad_norm_pre_clip': 2.166952133178711, 'timing/data': 0.000332023948431015, 'timing/model': 1.0910495319403708, 'learning_rate/qwen_vl_interface': 1.6343468276372471e-06, 'learning_rate/action_model': 1.634346827637247e-05, 'epoch': 0.3}) +[2026-06-24 13:57:09,665][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.17162133753299713, 'train/grad_norm_pre_clip': 89.45919036865234, 'timing/data': 0.00033622095361351967, 'timing/model': 1.2986755380406976, 'learning_rate/qwen_vl_interface': 1.6318665361575443e-06, 'learning_rate/action_model': 1.6318665361575442e-05, 'epoch': 0.3}) +[2026-06-24 13:57:10,753][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.10195816308259964, 'train/grad_norm_pre_clip': 65.49309539794922, 'timing/data': 0.0003553670831024647, 'timing/model': 1.0871712258085608, 'learning_rate/qwen_vl_interface': 1.6293885924453656e-06, 'learning_rate/action_model': 1.6293885924453655e-05, 'epoch': 0.3}) +[2026-06-24 13:57:11,843][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.008629241958260536, 'train/grad_norm_pre_clip': 16.355918884277344, 'timing/data': 0.00032126298174262047, 'timing/model': 1.0881933581549674, 'learning_rate/qwen_vl_interface': 1.6269129981086188e-06, 'learning_rate/action_model': 1.6269129981086186e-05, 'epoch': 0.3}) +[2026-06-24 13:57:12,928][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.0016655090730637312, 'train/grad_norm_pre_clip': 1.7046760320663452, 'timing/data': 0.00038729794323444366, 'timing/model': 1.0828800441231579, 'learning_rate/qwen_vl_interface': 1.6244397547536942e-06, 'learning_rate/action_model': 1.6244397547536942e-05, 'epoch': 0.3}) +[2026-06-24 13:57:14,017][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.015180874615907669, 'train/grad_norm_pre_clip': 31.29644203186035, 'timing/data': 0.0003304709680378437, 'timing/model': 1.0874379300512373, 'learning_rate/qwen_vl_interface': 1.6219688639854507e-06, 'learning_rate/action_model': 1.6219688639854506e-05, 'epoch': 0.3}) +[2026-06-24 13:57:15,106][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.25846433639526367, 'train/grad_norm_pre_clip': 110.19918823242188, 'timing/data': 0.0005191050004214048, 'timing/model': 1.0871078600175679, 'learning_rate/qwen_vl_interface': 1.6195003274072201e-06, 'learning_rate/action_model': 1.6195003274072203e-05, 'epoch': 0.3}) +[2026-06-24 13:57:16,195][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.022796621546149254, 'train/grad_norm_pre_clip': 31.937137603759766, 'timing/data': 0.0003435150720179081, 'timing/model': 1.0877030191477388, 'learning_rate/qwen_vl_interface': 1.6170341466208139e-06, 'learning_rate/action_model': 1.617034146620814e-05, 'epoch': 0.3}) +[2026-06-24 13:57:17,283][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.06815508753061295, 'train/grad_norm_pre_clip': 54.949302673339844, 'timing/data': 0.0003149320837110281, 'timing/model': 1.0860442728735507, 'learning_rate/qwen_vl_interface': 1.6145703232265055e-06, 'learning_rate/action_model': 1.6145703232265055e-05, 'epoch': 0.3}) +[2026-06-24 13:57:18,372][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.2528584599494934, 'train/grad_norm_pre_clip': 98.47431182861328, 'timing/data': 0.00031099398620426655, 'timing/model': 1.0880534080788493, 'learning_rate/qwen_vl_interface': 1.612108858823046e-06, 'learning_rate/action_model': 1.612108858823046e-05, 'epoch': 0.3}) +[2026-06-24 13:57:19,460][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.04206777364015579, 'train/grad_norm_pre_clip': 22.9395809173584, 'timing/data': 0.0005321300122886896, 'timing/model': 1.0858310931362212, 'learning_rate/qwen_vl_interface': 1.609649755007653e-06, 'learning_rate/action_model': 1.6096497550076527e-05, 'epoch': 0.3}) +[2026-06-24 13:57:20,550][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.008801077492535114, 'train/grad_norm_pre_clip': 22.513233184814453, 'timing/data': 0.0003223458770662546, 'timing/model': 1.0878126679453999, 'learning_rate/qwen_vl_interface': 1.6071930133760092e-06, 'learning_rate/action_model': 1.6071930133760092e-05, 'epoch': 0.3}) +[2026-06-24 13:57:21,641][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.013841643929481506, 'train/grad_norm_pre_clip': 18.030841827392578, 'timing/data': 0.000315273180603981, 'timing/model': 1.089516208972782, 'learning_rate/qwen_vl_interface': 1.6047386355222703e-06, 'learning_rate/action_model': 1.60473863552227e-05, 'epoch': 0.3}) +[2026-06-24 13:57:22,732][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.2948416471481323, 'train/grad_norm_pre_clip': 53.900428771972656, 'timing/data': 0.0003789321053773165, 'timing/model': 1.088953873841092, 'learning_rate/qwen_vl_interface': 1.602286623039055e-06, 'learning_rate/action_model': 1.6022866230390548e-05, 'epoch': 0.3}) +[2026-06-24 13:57:23,821][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.06591876596212387, 'train/grad_norm_pre_clip': 21.77597427368164, 'timing/data': 0.00034014880657196045, 'timing/model': 1.0879005349706858, 'learning_rate/qwen_vl_interface': 1.5998369775174483e-06, 'learning_rate/action_model': 1.599836977517448e-05, 'epoch': 0.3}) +[2026-06-24 13:57:24,911][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.01659700833261013, 'train/grad_norm_pre_clip': 31.00115966796875, 'timing/data': 0.0003590651322156191, 'timing/model': 1.0884202611632645, 'learning_rate/qwen_vl_interface': 1.5973897005469964e-06, 'learning_rate/action_model': 1.597389700546996e-05, 'epoch': 0.3}) +[2026-06-24 13:57:26,201][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.07903996855020523, 'train/grad_norm_pre_clip': 77.89900207519531, 'timing/data': 0.00033152312971651554, 'timing/model': 1.2887591461185366, 'learning_rate/qwen_vl_interface': 1.5949447937157116e-06, 'learning_rate/action_model': 1.5949447937157116e-05, 'epoch': 0.3}) +[2026-06-24 13:57:27,289][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.15019935369491577, 'train/grad_norm_pre_clip': 88.27213287353516, 'timing/data': 0.0004060729406774044, 'timing/model': 1.0857303091324866, 'learning_rate/qwen_vl_interface': 1.5925022586100697e-06, 'learning_rate/action_model': 1.5925022586100695e-05, 'epoch': 0.3}) +[2026-06-24 13:57:28,376][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.015321053564548492, 'train/grad_norm_pre_clip': 75.41632843017578, 'timing/data': 0.0003165740054100752, 'timing/model': 1.0858151579741389, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 0.3}) +[2026-06-24 13:57:29,466][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.018024278804659843, 'train/grad_norm_pre_clip': 62.66672134399414, 'timing/data': 0.0003043918404728174, 'timing/model': 1.088505914900452, 'learning_rate/qwen_vl_interface': 1.587624309913904e-06, 'learning_rate/action_model': 1.5876243099139037e-05, 'epoch': 0.3}) +[2026-06-24 13:57:30,556][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.018153581768274307, 'train/grad_norm_pre_clip': 7.660285949707031, 'timing/data': 0.0003012660890817642, 'timing/model': 1.088019261136651, 'learning_rate/qwen_vl_interface': 1.5851888994886291e-06, 'learning_rate/action_model': 1.5851888994886292e-05, 'epoch': 0.3}) +[2026-06-24 13:57:31,642][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.014922568574547768, 'train/grad_norm_pre_clip': 12.57199764251709, 'timing/data': 0.0003273750189691782, 'timing/model': 1.0845784968696535, 'learning_rate/qwen_vl_interface': 1.5827558671194912e-06, 'learning_rate/action_model': 1.582755867119491e-05, 'epoch': 0.3}) +[2026-06-24 13:57:32,732][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.040090929716825485, 'train/grad_norm_pre_clip': 14.865621566772461, 'timing/data': 0.00033387704752385616, 'timing/model': 1.0881275860592723, 'learning_rate/qwen_vl_interface': 1.5803252143852544e-06, 'learning_rate/action_model': 1.5803252143852544e-05, 'epoch': 0.3}) +[2026-06-24 13:57:33,818][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.05118449777364731, 'train/grad_norm_pre_clip': 26.726669311523438, 'timing/data': 0.00030794902704656124, 'timing/model': 1.0848982890602201, 'learning_rate/qwen_vl_interface': 1.5778969428631453e-06, 'learning_rate/action_model': 1.5778969428631453e-05, 'epoch': 0.3}) +[2026-06-24 13:57:34,906][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.0020618641283363104, 'train/grad_norm_pre_clip': 1.047644019126892, 'timing/data': 0.0003794231452047825, 'timing/model': 1.0855791990179569, 'learning_rate/qwen_vl_interface': 1.5754710541288437e-06, 'learning_rate/action_model': 1.5754710541288437e-05, 'epoch': 0.3}) +[2026-06-24 13:57:35,997][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.04732232168316841, 'train/grad_norm_pre_clip': 35.4376106262207, 'timing/data': 0.0003791418857872486, 'timing/model': 1.0891565908677876, 'learning_rate/qwen_vl_interface': 1.5730475497564784e-06, 'learning_rate/action_model': 1.5730475497564785e-05, 'epoch': 0.3}) +[2026-06-24 13:57:37,086][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.06737353652715683, 'train/grad_norm_pre_clip': 93.98834991455078, 'timing/data': 0.0003082400653511286, 'timing/model': 1.0872660800814629, 'learning_rate/qwen_vl_interface': 1.5706264313186412e-06, 'learning_rate/action_model': 1.570626431318641e-05, 'epoch': 0.3}) +[2026-06-24 13:57:38,175][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.21592977643013, 'train/grad_norm_pre_clip': 83.98833465576172, 'timing/data': 0.0003007850609719753, 'timing/model': 1.088290295097977, 'learning_rate/qwen_vl_interface': 1.5682077003863657e-06, 'learning_rate/action_model': 1.5682077003863657e-05, 'epoch': 0.3}) +[2026-06-24 13:57:39,266][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.03152918815612793, 'train/grad_norm_pre_clip': 26.164819717407227, 'timing/data': 0.00031393906101584435, 'timing/model': 1.0887605790048838, 'learning_rate/qwen_vl_interface': 1.565791358529139e-06, 'learning_rate/action_model': 1.5657913585291388e-05, 'epoch': 0.3}) +[2026-06-24 13:57:40,354][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.3916202187538147, 'train/grad_norm_pre_clip': 146.17213439941406, 'timing/data': 0.0003825090825557709, 'timing/model': 1.0861511370167136, 'learning_rate/qwen_vl_interface': 1.5633774073149025e-06, 'learning_rate/action_model': 1.5633774073149025e-05, 'epoch': 0.31}) +[2026-06-24 13:57:41,450][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.028641967102885246, 'train/grad_norm_pre_clip': 28.04336929321289, 'timing/data': 0.0003071569371968508, 'timing/model': 1.0877363260369748, 'learning_rate/qwen_vl_interface': 1.5609658483100407e-06, 'learning_rate/action_model': 1.5609658483100406e-05, 'epoch': 0.31}) +[2026-06-24 13:57:42,827][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.04796919226646423, 'train/grad_norm_pre_clip': 37.204612731933594, 'timing/data': 0.0004327429924160242, 'timing/model': 1.3746314190793782, 'learning_rate/qwen_vl_interface': 1.5585566830793883e-06, 'learning_rate/action_model': 1.558556683079388e-05, 'epoch': 0.31}) +[2026-06-24 13:57:43,920][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.27193355560302734, 'train/grad_norm_pre_clip': 114.53569030761719, 'timing/data': 0.0003297098446637392, 'timing/model': 1.0911980119999498, 'learning_rate/qwen_vl_interface': 1.556149913186228e-06, 'learning_rate/action_model': 1.5561499131862278e-05, 'epoch': 0.31}) +[2026-06-24 13:57:45,008][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.0650188997387886, 'train/grad_norm_pre_clip': 74.39361572265625, 'timing/data': 0.00034475792199373245, 'timing/model': 1.086907610995695, 'learning_rate/qwen_vl_interface': 1.5537455401922844e-06, 'learning_rate/action_model': 1.5537455401922844e-05, 'epoch': 0.31}) +[2026-06-24 13:57:46,099][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.0967865139245987, 'train/grad_norm_pre_clip': 57.89069747924805, 'timing/data': 0.000441879965364933, 'timing/model': 1.0889695561490953, 'learning_rate/qwen_vl_interface': 1.5513435656577302e-06, 'learning_rate/action_model': 1.5513435656577303e-05, 'epoch': 0.31}) +[2026-06-24 13:57:47,187][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.15276968479156494, 'train/grad_norm_pre_clip': 64.30810546875, 'timing/data': 0.0003735420759767294, 'timing/model': 1.0869332109577954, 'learning_rate/qwen_vl_interface': 1.5489439911411804e-06, 'learning_rate/action_model': 1.5489439911411803e-05, 'epoch': 0.31}) +[2026-06-24 13:57:48,277][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.1026042103767395, 'train/grad_norm_pre_clip': 58.83426284790039, 'timing/data': 0.0003163449000567198, 'timing/model': 1.0884131721686572, 'learning_rate/qwen_vl_interface': 1.5465468181996943e-06, 'learning_rate/action_model': 1.5465468181996943e-05, 'epoch': 0.31}) +[2026-06-24 13:57:49,366][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.0767759382724762, 'train/grad_norm_pre_clip': 75.61148834228516, 'timing/data': 0.00032818689942359924, 'timing/model': 1.0875618031714112, 'learning_rate/qwen_vl_interface': 1.5441520483887687e-06, 'learning_rate/action_model': 1.5441520483887686e-05, 'epoch': 0.31}) +[2026-06-24 13:57:50,454][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.16809135675430298, 'train/grad_norm_pre_clip': 159.01605224609375, 'timing/data': 0.00032233609817922115, 'timing/model': 1.086126446025446, 'learning_rate/qwen_vl_interface': 1.5417596832623451e-06, 'learning_rate/action_model': 1.5417596832623452e-05, 'epoch': 0.31}) +[2026-06-24 13:57:51,540][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.10441529750823975, 'train/grad_norm_pre_clip': 95.60365295410156, 'timing/data': 0.0003089399542659521, 'timing/model': 1.0847767011728138, 'learning_rate/qwen_vl_interface': 1.5393697243728047e-06, 'learning_rate/action_model': 1.539369724372805e-05, 'epoch': 0.31}) +[2026-06-24 13:57:52,631][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.2938031852245331, 'train/grad_norm_pre_clip': 115.05137634277344, 'timing/data': 0.0003140910994261503, 'timing/model': 1.0888821550179273, 'learning_rate/qwen_vl_interface': 1.536982173270963e-06, 'learning_rate/action_model': 1.536982173270963e-05, 'epoch': 0.31}) +[2026-06-24 13:57:53,721][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.04801320284605026, 'train/grad_norm_pre_clip': 51.53178405761719, 'timing/data': 0.0003412009682506323, 'timing/model': 1.0885239618364722, 'learning_rate/qwen_vl_interface': 1.5345970315060774e-06, 'learning_rate/action_model': 1.5345970315060772e-05, 'epoch': 0.31}) +[2026-06-24 13:57:54,808][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.007088012993335724, 'train/grad_norm_pre_clip': 10.50236701965332, 'timing/data': 0.00039481185376644135, 'timing/model': 1.08575258590281, 'learning_rate/qwen_vl_interface': 1.5322143006258406e-06, 'learning_rate/action_model': 1.5322143006258406e-05, 'epoch': 0.31}) +[2026-06-24 13:57:55,896][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.013619320467114449, 'train/grad_norm_pre_clip': 18.327245712280273, 'timing/data': 0.0003044919576495886, 'timing/model': 1.086059191962704, 'learning_rate/qwen_vl_interface': 1.5298339821763817e-06, 'learning_rate/action_model': 1.5298339821763815e-05, 'epoch': 0.31}) +[2026-06-24 13:57:56,984][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.11166302114725113, 'train/grad_norm_pre_clip': 51.60898971557617, 'timing/data': 0.00037446292117238045, 'timing/model': 1.0863895930815488, 'learning_rate/qwen_vl_interface': 1.5274560777022613e-06, 'learning_rate/action_model': 1.5274560777022613e-05, 'epoch': 0.31}) +[2026-06-24 13:57:58,277][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.050448380410671234, 'train/grad_norm_pre_clip': 25.665863037109375, 'timing/data': 0.0003215039614588022, 'timing/model': 1.2914375599939376, 'learning_rate/qwen_vl_interface': 1.5250805887464773e-06, 'learning_rate/action_model': 1.5250805887464773e-05, 'epoch': 0.31}) +[2026-06-24 13:57:59,358][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.06156422197818756, 'train/grad_norm_pre_clip': 27.708131790161133, 'timing/data': 0.0003159339539706707, 'timing/model': 1.0788946619722992, 'learning_rate/qwen_vl_interface': 1.5227075168504594e-06, 'learning_rate/action_model': 1.5227075168504593e-05, 'epoch': 0.31}) +[2026-06-24 13:58:00,441][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.1379113793373108, 'train/grad_norm_pre_clip': 47.372684478759766, 'timing/data': 0.0004913539160043001, 'timing/model': 1.0819374800194055, 'learning_rate/qwen_vl_interface': 1.5203368635540647e-06, 'learning_rate/action_model': 1.5203368635540647e-05, 'epoch': 0.31}) +[2026-06-24 13:58:01,526][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.12985338270664215, 'train/grad_norm_pre_clip': 77.47150421142578, 'timing/data': 0.00033236388117074966, 'timing/model': 1.0833774160128087, 'learning_rate/qwen_vl_interface': 1.5179686303955898e-06, 'learning_rate/action_model': 1.5179686303955898e-05, 'epoch': 0.31}) +[2026-06-24 13:58:02,613][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.09298858791589737, 'train/grad_norm_pre_clip': 32.123138427734375, 'timing/data': 0.00031377002596855164, 'timing/model': 1.085099503165111, 'learning_rate/qwen_vl_interface': 1.5156028189117522e-06, 'learning_rate/action_model': 1.5156028189117522e-05, 'epoch': 0.31}) +[2026-06-24 13:58:03,696][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.12534798681735992, 'train/grad_norm_pre_clip': 58.51124954223633, 'timing/data': 0.0003290779422968626, 'timing/model': 1.0818108138628304, 'learning_rate/qwen_vl_interface': 1.5132394306376988e-06, 'learning_rate/action_model': 1.5132394306376987e-05, 'epoch': 0.31}) +[2026-06-24 13:58:04,779][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.024781404063105583, 'train/grad_norm_pre_clip': 5.764720439910889, 'timing/data': 0.0003423329908400774, 'timing/model': 1.0807644079905003, 'learning_rate/qwen_vl_interface': 1.510878467107012e-06, 'learning_rate/action_model': 1.5108784671070119e-05, 'epoch': 0.31}) +[2026-06-24 13:58:05,863][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.010789114981889725, 'train/grad_norm_pre_clip': 17.117069244384766, 'timing/data': 0.0003878690768033266, 'timing/model': 1.0829036980867386, 'learning_rate/qwen_vl_interface': 1.5085199298516923e-06, 'learning_rate/action_model': 1.5085199298516923e-05, 'epoch': 0.31}) +[2026-06-24 13:58:06,947][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.12653742730617523, 'train/grad_norm_pre_clip': 68.85816955566406, 'timing/data': 0.0005093479994684458, 'timing/model': 1.0823124980088323, 'learning_rate/qwen_vl_interface': 1.5061638204021688e-06, 'learning_rate/action_model': 1.5061638204021689e-05, 'epoch': 0.31}) +[2026-06-24 13:58:08,031][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.02826453372836113, 'train/grad_norm_pre_clip': 14.525811195373535, 'timing/data': 0.000318288104608655, 'timing/model': 1.0817417160142213, 'learning_rate/qwen_vl_interface': 1.5038101402872971e-06, 'learning_rate/action_model': 1.503810140287297e-05, 'epoch': 0.31}) +[2026-06-24 13:58:09,114][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.013760355301201344, 'train/grad_norm_pre_clip': 13.120747566223145, 'timing/data': 0.00032818596810102463, 'timing/model': 1.0813147639855742, 'learning_rate/qwen_vl_interface': 1.5014588910343531e-06, 'learning_rate/action_model': 1.501458891034353e-05, 'epoch': 0.31}) +[2026-06-24 13:58:10,198][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.2931467592716217, 'train/grad_norm_pre_clip': 119.40377807617188, 'timing/data': 0.0003119960892945528, 'timing/model': 1.0832335301674902, 'learning_rate/qwen_vl_interface': 1.4991100741690365e-06, 'learning_rate/action_model': 1.4991100741690364e-05, 'epoch': 0.31}) +[2026-06-24 13:58:11,283][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.09853634238243103, 'train/grad_norm_pre_clip': 38.14662170410156, 'timing/data': 0.00032950914464890957, 'timing/model': 1.0829879390075803, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 0.31}) +[2026-06-24 13:58:12,365][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.01617550477385521, 'train/grad_norm_pre_clip': 8.497610092163086, 'timing/data': 0.00031628389842808247, 'timing/model': 1.08008186891675, 'learning_rate/qwen_vl_interface': 1.494419743696198e-06, 'learning_rate/action_model': 1.494419743696198e-05, 'epoch': 0.31}) +[2026-06-24 13:58:13,451][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.1221022680401802, 'train/grad_norm_pre_clip': 43.585243225097656, 'timing/data': 0.00043479702435433865, 'timing/model': 1.0849201509263366, 'learning_rate/qwen_vl_interface': 1.492078233132179e-06, 'learning_rate/action_model': 1.492078233132179e-05, 'epoch': 0.31}) +[2026-06-24 13:58:14,537][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.10659919679164886, 'train/grad_norm_pre_clip': 53.509674072265625, 'timing/data': 0.00032613310031592846, 'timing/model': 1.0839216059539467, 'learning_rate/qwen_vl_interface': 1.4897391610427953e-06, 'learning_rate/action_model': 1.4897391610427952e-05, 'epoch': 0.31}) +[2026-06-24 13:58:15,621][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.24143891036510468, 'train/grad_norm_pre_clip': 70.30552673339844, 'timing/data': 0.00036130892112851143, 'timing/model': 1.0826710870023817, 'learning_rate/qwen_vl_interface': 1.4874025289458465e-06, 'learning_rate/action_model': 1.4874025289458465e-05, 'epoch': 0.31}) +[2026-06-24 13:58:16,705][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.04377374425530434, 'train/grad_norm_pre_clip': 23.688148498535156, 'timing/data': 0.0003352200146764517, 'timing/model': 1.0823908799793571, 'learning_rate/qwen_vl_interface': 1.4850683383575456e-06, 'learning_rate/action_model': 1.4850683383575457e-05, 'epoch': 0.31}) +[2026-06-24 13:58:18,078][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.0012569708051159978, 'train/grad_norm_pre_clip': 2.2471587657928467, 'timing/data': 0.00037723896093666553, 'timing/model': 1.3709506560117006, 'learning_rate/qwen_vl_interface': 1.4827365907925249e-06, 'learning_rate/action_model': 1.4827365907925247e-05, 'epoch': 0.31}) +[2026-06-24 13:58:19,161][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.0652526319026947, 'train/grad_norm_pre_clip': 58.31807327270508, 'timing/data': 0.0003605268429964781, 'timing/model': 1.0812348588369787, 'learning_rate/qwen_vl_interface': 1.4804072877638292e-06, 'learning_rate/action_model': 1.4804072877638292e-05, 'epoch': 0.31}) +[2026-06-24 13:58:20,243][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.08764510601758957, 'train/grad_norm_pre_clip': 74.6252670288086, 'timing/data': 0.0003307410515844822, 'timing/model': 1.0802020290866494, 'learning_rate/qwen_vl_interface': 1.4780804307829204e-06, 'learning_rate/action_model': 1.4780804307829205e-05, 'epoch': 0.31}) +[2026-06-24 13:58:21,325][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.008800284937024117, 'train/grad_norm_pre_clip': 7.812977313995361, 'timing/data': 0.00032909889705479145, 'timing/model': 1.0811135428957641, 'learning_rate/qwen_vl_interface': 1.4757560213596677e-06, 'learning_rate/action_model': 1.4757560213596677e-05, 'epoch': 0.31}) +[2026-06-24 13:58:22,407][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.11460819095373154, 'train/grad_norm_pre_clip': 60.77030563354492, 'timing/data': 0.00034177303314208984, 'timing/model': 1.0803831301163882, 'learning_rate/qwen_vl_interface': 1.4734340610023564e-06, 'learning_rate/action_model': 1.4734340610023564e-05, 'epoch': 0.31}) +[2026-06-24 13:58:23,491][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.04333837702870369, 'train/grad_norm_pre_clip': 25.561359405517578, 'timing/data': 0.00031357002444565296, 'timing/model': 1.0821799570694566, 'learning_rate/qwen_vl_interface': 1.4711145512176825e-06, 'learning_rate/action_model': 1.4711145512176824e-05, 'epoch': 0.31}) +[2026-06-24 13:58:24,575][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.015292420983314514, 'train/grad_norm_pre_clip': 10.722420692443848, 'timing/data': 0.0003276858478784561, 'timing/model': 1.0822524449322373, 'learning_rate/qwen_vl_interface': 1.4687974935107464e-06, 'learning_rate/action_model': 1.4687974935107463e-05, 'epoch': 0.31}) +[2026-06-24 13:58:25,659][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.02449292503297329, 'train/grad_norm_pre_clip': 6.567361354827881, 'timing/data': 0.0003174969460815191, 'timing/model': 1.083017683122307, 'learning_rate/qwen_vl_interface': 1.4664828893850677e-06, 'learning_rate/action_model': 1.4664828893850676e-05, 'epoch': 0.31}) +[2026-06-24 13:58:26,744][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.07584904134273529, 'train/grad_norm_pre_clip': 67.06537628173828, 'timing/data': 0.0003419029526412487, 'timing/model': 1.0826678560115397, 'learning_rate/qwen_vl_interface': 1.4641707403425656e-06, 'learning_rate/action_model': 1.4641707403425655e-05, 'epoch': 0.31}) +[2026-06-24 13:58:27,829][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.09138510376214981, 'train/grad_norm_pre_clip': 78.39046478271484, 'timing/data': 0.00036730989813804626, 'timing/model': 1.0839182359632105, 'learning_rate/qwen_vl_interface': 1.4618610478835652e-06, 'learning_rate/action_model': 1.4618610478835652e-05, 'epoch': 0.31}) +[2026-06-24 13:58:28,915][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.04782909154891968, 'train/grad_norm_pre_clip': 65.4598388671875, 'timing/data': 0.00031809788197278976, 'timing/model': 1.0839382440317422, 'learning_rate/qwen_vl_interface': 1.4595538135068057e-06, 'learning_rate/action_model': 1.4595538135068057e-05, 'epoch': 0.31}) +[2026-06-24 13:58:30,000][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.0818134993314743, 'train/grad_norm_pre_clip': 68.19234466552734, 'timing/data': 0.00032958900555968285, 'timing/model': 1.0837269769981503, 'learning_rate/qwen_vl_interface': 1.4572490387094231e-06, 'learning_rate/action_model': 1.4572490387094232e-05, 'epoch': 0.31}) +[2026-06-24 13:58:31,090][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.02758810669183731, 'train/grad_norm_pre_clip': 27.963109970092773, 'timing/data': 0.0003407502081245184, 'timing/model': 1.0886750768404454, 'learning_rate/qwen_vl_interface': 1.454946724986962e-06, 'learning_rate/action_model': 1.454946724986962e-05, 'epoch': 0.31}) +[2026-06-24 13:58:32,178][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.13533590734004974, 'train/grad_norm_pre_clip': 97.92938232421875, 'timing/data': 0.00032515102066099644, 'timing/model': 1.0863735019229352, 'learning_rate/qwen_vl_interface': 1.4526468738333705e-06, 'learning_rate/action_model': 1.4526468738333704e-05, 'epoch': 0.31}) +[2026-06-24 13:58:33,262][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.16708217561244965, 'train/grad_norm_pre_clip': 64.20745849609375, 'timing/data': 0.0003287179861217737, 'timing/model': 1.0823728051036596, 'learning_rate/qwen_vl_interface': 1.4503494867409934e-06, 'learning_rate/action_model': 1.4503494867409934e-05, 'epoch': 0.31}) +[2026-06-24 13:58:34,345][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.23078866302967072, 'train/grad_norm_pre_clip': 64.76473236083984, 'timing/data': 0.00032851705327630043, 'timing/model': 1.081356615992263, 'learning_rate/qwen_vl_interface': 1.4480545652005825e-06, 'learning_rate/action_model': 1.4480545652005825e-05, 'epoch': 0.31}) +[2026-06-24 13:58:35,428][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.005126784089952707, 'train/grad_norm_pre_clip': 10.881611824035645, 'timing/data': 0.00030988198705017567, 'timing/model': 1.0816934588365257, 'learning_rate/qwen_vl_interface': 1.445762110701287e-06, 'learning_rate/action_model': 1.445762110701287e-05, 'epoch': 0.31}) +[2026-06-24 13:58:36,756][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.03857782855629921, 'train/grad_norm_pre_clip': 18.559398651123047, 'timing/data': 0.0003282870166003704, 'timing/model': 1.326485161902383, 'learning_rate/qwen_vl_interface': 1.4434721247306573e-06, 'learning_rate/action_model': 1.4434721247306573e-05, 'epoch': 0.31}) +[2026-06-24 13:58:37,838][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.12256588041782379, 'train/grad_norm_pre_clip': 28.466781616210938, 'timing/data': 0.0003369031473994255, 'timing/model': 1.0794205800630152, 'learning_rate/qwen_vl_interface': 1.4411846087746376e-06, 'learning_rate/action_model': 1.4411846087746375e-05, 'epoch': 0.31}) +[2026-06-24 13:58:38,918][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.1345452070236206, 'train/grad_norm_pre_clip': 148.1570281982422, 'timing/data': 0.00038337986916303635, 'timing/model': 1.0789374317973852, 'learning_rate/qwen_vl_interface': 1.4388995643175734e-06, 'learning_rate/action_model': 1.4388995643175734e-05, 'epoch': 0.31}) +[2026-06-24 13:58:39,997][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.10550887137651443, 'train/grad_norm_pre_clip': 67.8530502319336, 'timing/data': 0.00044900295324623585, 'timing/model': 1.0767995449714363, 'learning_rate/qwen_vl_interface': 1.436616992842207e-06, 'learning_rate/action_model': 1.4366169928422068e-05, 'epoch': 0.31}) +[2026-06-24 13:58:41,076][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.18126751482486725, 'train/grad_norm_pre_clip': 141.40576171875, 'timing/data': 0.0003391269128769636, 'timing/model': 1.0773306640330702, 'learning_rate/qwen_vl_interface': 1.434336895829671e-06, 'learning_rate/action_model': 1.434336895829671e-05, 'epoch': 0.31}) +[2026-06-24 13:58:42,156][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.1512327790260315, 'train/grad_norm_pre_clip': 134.8898468017578, 'timing/data': 0.0003310020547360182, 'timing/model': 1.0786302261985838, 'learning_rate/qwen_vl_interface': 1.4320592747594975e-06, 'learning_rate/action_model': 1.4320592747594974e-05, 'epoch': 0.31}) +[2026-06-24 13:58:43,236][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.01805979572236538, 'train/grad_norm_pre_clip': 12.66123104095459, 'timing/data': 0.00034209294244647026, 'timing/model': 1.0784826599992812, 'learning_rate/qwen_vl_interface': 1.42978413110961e-06, 'learning_rate/action_model': 1.42978413110961e-05, 'epoch': 0.31}) +[2026-06-24 13:58:44,316][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.17543229460716248, 'train/grad_norm_pre_clip': 29.0853214263916, 'timing/data': 0.00031164498068392277, 'timing/model': 1.0787276609335095, 'learning_rate/qwen_vl_interface': 1.4275114663563239e-06, 'learning_rate/action_model': 1.4275114663563239e-05, 'epoch': 0.31}) +[2026-06-24 13:58:45,396][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.005782662890851498, 'train/grad_norm_pre_clip': 5.432467937469482, 'timing/data': 0.00031385011970996857, 'timing/model': 1.0781093400437385, 'learning_rate/qwen_vl_interface': 1.4252412819743486e-06, 'learning_rate/action_model': 1.4252412819743486e-05, 'epoch': 0.31}) +[2026-06-24 13:58:46,475][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.17601236701011658, 'train/grad_norm_pre_clip': 72.24060821533203, 'timing/data': 0.000308099202811718, 'timing/model': 1.077358127804473, 'learning_rate/qwen_vl_interface': 1.422973579436779e-06, 'learning_rate/action_model': 1.422973579436779e-05, 'epoch': 0.31}) +[2026-06-24 13:58:47,554][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.01133317593485117, 'train/grad_norm_pre_clip': 8.81784725189209, 'timing/data': 0.0003221849910914898, 'timing/model': 1.0778665440157056, 'learning_rate/qwen_vl_interface': 1.4207083602151047e-06, 'learning_rate/action_model': 1.4207083602151047e-05, 'epoch': 0.31}) +[2026-06-24 13:58:48,631][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.14712320268154144, 'train/grad_norm_pre_clip': 28.79555320739746, 'timing/data': 0.0003475330304354429, 'timing/model': 1.0745513050351292, 'learning_rate/qwen_vl_interface': 1.4184456257792011e-06, 'learning_rate/action_model': 1.4184456257792012e-05, 'epoch': 0.31}) +[2026-06-24 13:58:49,722][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.1859249770641327, 'train/grad_norm_pre_clip': 81.86692810058594, 'timing/data': 0.0003271948080509901, 'timing/model': 1.0863273539580405, 'learning_rate/qwen_vl_interface': 1.416185377597334e-06, 'learning_rate/action_model': 1.4161853775973341e-05, 'epoch': 0.31}) +[2026-06-24 13:58:51,119][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.04967940226197243, 'train/grad_norm_pre_clip': 21.80455780029297, 'timing/data': 0.0008555680979043245, 'timing/model': 1.393073168117553, 'learning_rate/qwen_vl_interface': 1.4139276171361504e-06, 'learning_rate/action_model': 1.4139276171361504e-05, 'epoch': 0.31}) +[2026-06-24 13:58:52,213][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.018387315794825554, 'train/grad_norm_pre_clip': 19.026735305786133, 'timing/data': 0.0004602049011737108, 'timing/model': 1.0928380200639367, 'learning_rate/qwen_vl_interface': 1.411672345860689e-06, 'learning_rate/action_model': 1.411672345860689e-05, 'epoch': 0.31}) +[2026-06-24 13:58:53,290][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.10584988445043564, 'train/grad_norm_pre_clip': 74.31194305419922, 'timing/data': 0.00033162301406264305, 'timing/model': 1.0754674680065364, 'learning_rate/qwen_vl_interface': 1.4094195652343714e-06, 'learning_rate/action_model': 1.4094195652343715e-05, 'epoch': 0.31}) +[2026-06-24 13:58:54,370][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.07620900124311447, 'train/grad_norm_pre_clip': 34.60548782348633, 'timing/data': 0.0003411318175494671, 'timing/model': 1.0782108739949763, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 0.31}) +[2026-06-24 13:58:55,622][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.11770990490913391, 'train/grad_norm_pre_clip': 57.149871826171875, 'timing/data': 0.00032928911969065666, 'timing/model': 1.2503078470472246, 'learning_rate/qwen_vl_interface': 1.404921481774764e-06, 'learning_rate/action_model': 1.4049214817747641e-05, 'epoch': 0.31}) +[2026-06-24 13:58:56,701][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.01198149099946022, 'train/grad_norm_pre_clip': 9.937931060791016, 'timing/data': 0.00034791394136846066, 'timing/model': 1.077089638216421, 'learning_rate/qwen_vl_interface': 1.4026761818602344e-06, 'learning_rate/action_model': 1.4026761818602343e-05, 'epoch': 0.31}) +[2026-06-24 13:58:57,779][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.03997049853205681, 'train/grad_norm_pre_clip': 38.884273529052734, 'timing/data': 0.0003485949710011482, 'timing/model': 1.0768799250945449, 'learning_rate/qwen_vl_interface': 1.4004333784323588e-06, 'learning_rate/action_model': 1.4004333784323587e-05, 'epoch': 0.31}) +[2026-06-24 13:58:58,858][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.2084967941045761, 'train/grad_norm_pre_clip': 75.37789154052734, 'timing/data': 0.00031638494692742825, 'timing/model': 1.077537511009723, 'learning_rate/qwen_vl_interface': 1.3981930729464698e-06, 'learning_rate/action_model': 1.3981930729464699e-05, 'epoch': 0.31}) +[2026-06-24 13:58:59,941][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.023292874917387962, 'train/grad_norm_pre_clip': 18.956892013549805, 'timing/data': 0.0004231750499457121, 'timing/model': 1.0801730039529502, 'learning_rate/qwen_vl_interface': 1.3959552668562779e-06, 'learning_rate/action_model': 1.3959552668562778e-05, 'epoch': 0.31}) +[2026-06-24 13:59:01,018][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.18762557208538055, 'train/grad_norm_pre_clip': 64.38348388671875, 'timing/data': 0.00038927095010876656, 'timing/model': 1.075429572025314, 'learning_rate/qwen_vl_interface': 1.3937199616138716e-06, 'learning_rate/action_model': 1.3937199616138716e-05, 'epoch': 0.31}) +[2026-06-24 13:59:02,097][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.09622547775506973, 'train/grad_norm_pre_clip': 33.82304763793945, 'timing/data': 0.0003144009970128536, 'timing/model': 1.0779213840141892, 'learning_rate/qwen_vl_interface': 1.391487158669714e-06, 'learning_rate/action_model': 1.3914871586697139e-05, 'epoch': 0.31}) +[2026-06-24 13:59:03,179][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.013841846957802773, 'train/grad_norm_pre_clip': 19.731836318969727, 'timing/data': 0.00032298709265887737, 'timing/model': 1.0802357830107212, 'learning_rate/qwen_vl_interface': 1.3892568594726476e-06, 'learning_rate/action_model': 1.3892568594726476e-05, 'epoch': 0.31}) +[2026-06-24 13:59:04,259][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.10783398896455765, 'train/grad_norm_pre_clip': 51.04329299926758, 'timing/data': 0.00032500107772648335, 'timing/model': 1.0781568069942296, 'learning_rate/qwen_vl_interface': 1.387029065469891e-06, 'learning_rate/action_model': 1.3870290654698908e-05, 'epoch': 0.31}) +[2026-06-24 13:59:05,349][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.06181572005152702, 'train/grad_norm_pre_clip': 48.001399993896484, 'timing/data': 0.0003218650817871094, 'timing/model': 1.0879413839429617, 'learning_rate/qwen_vl_interface': 1.3848037781070317e-06, 'learning_rate/action_model': 1.3848037781070317e-05, 'epoch': 0.31}) +[2026-06-24 13:59:06,548][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.01066735852509737, 'train/grad_norm_pre_clip': 9.087386131286621, 'timing/data': 0.0006967890076339245, 'timing/model': 1.1967944439966232, 'learning_rate/qwen_vl_interface': 1.3825809988280393e-06, 'learning_rate/action_model': 1.382580998828039e-05, 'epoch': 0.31}) +[2026-06-24 13:59:07,631][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.018756069242954254, 'train/grad_norm_pre_clip': 21.953020095825195, 'timing/data': 0.00038053514435887337, 'timing/model': 1.0818531848490238, 'learning_rate/qwen_vl_interface': 1.3803607290752488e-06, 'learning_rate/action_model': 1.3803607290752487e-05, 'epoch': 0.31}) +[2026-06-24 13:59:08,709][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.02140776440501213, 'train/grad_norm_pre_clip': 20.892202377319336, 'timing/data': 0.00032300688326358795, 'timing/model': 1.0766172560397536, 'learning_rate/qwen_vl_interface': 1.378142970289367e-06, 'learning_rate/action_model': 1.378142970289367e-05, 'epoch': 0.31}) +[2026-06-24 13:59:09,788][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.14140728116035461, 'train/grad_norm_pre_clip': 72.45818328857422, 'timing/data': 0.00032580294646322727, 'timing/model': 1.0769589289557189, 'learning_rate/qwen_vl_interface': 1.3759277239094778e-06, 'learning_rate/action_model': 1.3759277239094778e-05, 'epoch': 0.31}) +[2026-06-24 13:59:10,863][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.010340246371924877, 'train/grad_norm_pre_clip': 17.548795700073242, 'timing/data': 0.0003461109008640051, 'timing/model': 1.0732035331893712, 'learning_rate/qwen_vl_interface': 1.3737149913730275e-06, 'learning_rate/action_model': 1.3737149913730273e-05, 'epoch': 0.31}) +[2026-06-24 13:59:11,938][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.054049666970968246, 'train/grad_norm_pre_clip': 54.04377746582031, 'timing/data': 0.00034549878910183907, 'timing/model': 1.0735951089300215, 'learning_rate/qwen_vl_interface': 1.3715047741158344e-06, 'learning_rate/action_model': 1.3715047741158344e-05, 'epoch': 0.31}) +[2026-06-24 13:59:13,219][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.0013728190679103136, 'train/grad_norm_pre_clip': 0.6752002835273743, 'timing/data': 0.00031741708517074585, 'timing/model': 1.2791373610962182, 'learning_rate/qwen_vl_interface': 1.3692970735720857e-06, 'learning_rate/action_model': 1.3692970735720857e-05, 'epoch': 0.31}) +[2026-06-24 13:59:14,296][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.1050659790635109, 'train/grad_norm_pre_clip': 71.80071258544922, 'timing/data': 0.0003280669916421175, 'timing/model': 1.0756992539390922, 'learning_rate/qwen_vl_interface': 1.3670918911743348e-06, 'learning_rate/action_model': 1.3670918911743346e-05, 'epoch': 0.31}) +[2026-06-24 13:59:15,373][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.01614227145910263, 'train/grad_norm_pre_clip': 6.451321601867676, 'timing/data': 0.0003188790287822485, 'timing/model': 1.0754892898257822, 'learning_rate/qwen_vl_interface': 1.3648892283534981e-06, 'learning_rate/action_model': 1.3648892283534981e-05, 'epoch': 0.31}) +[2026-06-24 13:59:16,451][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.2241114377975464, 'train/grad_norm_pre_clip': 99.17141723632812, 'timing/data': 0.000315873883664608, 'timing/model': 1.0758839219342917, 'learning_rate/qwen_vl_interface': 1.3626890865388615e-06, 'learning_rate/action_model': 1.3626890865388615e-05, 'epoch': 0.31}) +[2026-06-24 13:59:17,527][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.20907336473464966, 'train/grad_norm_pre_clip': 102.36255645751953, 'timing/data': 0.0003180780913680792, 'timing/model': 1.0747846369631588, 'learning_rate/qwen_vl_interface': 1.3604914671580741e-06, 'learning_rate/action_model': 1.360491467158074e-05, 'epoch': 0.31}) +[2026-06-24 13:59:18,606][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.10609466582536697, 'train/grad_norm_pre_clip': 48.556339263916016, 'timing/data': 0.000320622930303216, 'timing/model': 1.0767457629553974, 'learning_rate/qwen_vl_interface': 1.3582963716371444e-06, 'learning_rate/action_model': 1.3582963716371445e-05, 'epoch': 0.31}) +[2026-06-24 13:59:19,680][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.11264638602733612, 'train/grad_norm_pre_clip': 59.27577590942383, 'timing/data': 0.0003624008968472481, 'timing/model': 1.0727870240807533, 'learning_rate/qwen_vl_interface': 1.356103801400447e-06, 'learning_rate/action_model': 1.356103801400447e-05, 'epoch': 0.31}) +[2026-06-24 13:59:20,776][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.0582844503223896, 'train/grad_norm_pre_clip': 33.106380462646484, 'timing/data': 0.00032669398933649063, 'timing/model': 1.0943666559178382, 'learning_rate/qwen_vl_interface': 1.3539137578707195e-06, 'learning_rate/action_model': 1.3539137578707196e-05, 'epoch': 0.31}) +[2026-06-24 13:59:21,858][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.03071320801973343, 'train/grad_norm_pre_clip': 19.470874786376953, 'timing/data': 0.00047250697389245033, 'timing/model': 1.0799625390209258, 'learning_rate/qwen_vl_interface': 1.3517262424690534e-06, 'learning_rate/action_model': 1.3517262424690532e-05, 'epoch': 0.31}) +[2026-06-24 13:59:22,939][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.10439684242010117, 'train/grad_norm_pre_clip': 71.54354095458984, 'timing/data': 0.0005001400131732225, 'timing/model': 1.0792164269369096, 'learning_rate/qwen_vl_interface': 1.3495412566149058e-06, 'learning_rate/action_model': 1.3495412566149057e-05, 'epoch': 0.31}) +[2026-06-24 13:59:24,016][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.003938391339033842, 'train/grad_norm_pre_clip': 4.335025310516357, 'timing/data': 0.000519606051966548, 'timing/model': 1.0750946421176195, 'learning_rate/qwen_vl_interface': 1.3473588017260901e-06, 'learning_rate/action_model': 1.3473588017260901e-05, 'epoch': 0.31}) +[2026-06-24 13:59:25,098][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.01596219278872013, 'train/grad_norm_pre_clip': 11.38520336151123, 'timing/data': 0.0005163601599633694, 'timing/model': 1.080232216976583, 'learning_rate/qwen_vl_interface': 1.3451788792187795e-06, 'learning_rate/action_model': 1.3451788792187794e-05, 'epoch': 0.31}) +[2026-06-24 13:59:26,182][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.06446872651576996, 'train/grad_norm_pre_clip': 29.344684600830078, 'timing/data': 0.0005675570573657751, 'timing/model': 1.081445137038827, 'learning_rate/qwen_vl_interface': 1.3430014905074983e-06, 'learning_rate/action_model': 1.3430014905074984e-05, 'epoch': 0.31}) +[2026-06-24 13:59:27,260][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3250 +[2026-06-24 13:59:51,316][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3250 +[2026-06-24 13:59:51,316][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3250 +[2026-06-24 14:00:24,817][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3250 +[2026-06-24 14:09:59,463][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3250 +[2026-06-24 14:09:59,463][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.20356792211532593, 'train/grad_norm_pre_clip': 142.72140502929688, 'eval/action_loss/seconds': 24.05403699609451, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.06175939738750458, 'eval/latency_2/loss': 0.06175939738750458, 'eval/flappy/loss': 0.06175939738750458, 'eval/flappy/latency_2/loss': 0.06175939738750458, 'eval/flap_precision': 0.9505347593582888, 'eval/flap_recall': 0.9647218453188603, 'eval/flap_f1': 0.9575757575757576, 'eval/flap_rate_model': 0.20777777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9575757575757575, 'eval/latency_2/flap_precision': 0.9505347593582888, 'eval/latency_2/flap_recall': 0.9647218453188603, 'eval/latency_2/flap_f1': 0.9575757575757576, 'eval/latency_2/flap_rate_model': 0.20777777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9575757575757575, 'eval/action_classification/data_wait_seconds': 0.38868948817253113, 'eval/action_classification/predict_seconds': 33.02823257446289, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 442.26000000002523, 'rl_games_eval/mid_train/mean_length': 3582.4, 'rl_games_eval/mid_train/std_reward': 4.680000000000791, 'rl_games_eval/mid_train/std_length': 35.2, 'rl_games_eval/mid_train/macro_mean_reward': 442.26000000002523, 'rl_games_eval/mid_train/macro_mean_length': 3582.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 442.26000000002523, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3582.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 4.680000000000791, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 35.2, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 442.26000000002523, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00034594000317156315, 'timing/model': 1.076738293049857, 'learning_rate/qwen_vl_interface': 1.3408266370051323e-06, 'learning_rate/action_model': 1.3408266370051322e-05, 'epoch': 0.31}) +[2026-06-24 14:10:00,526][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.23748131096363068, 'train/grad_norm_pre_clip': 142.47486877441406, 'timing/data': 0.00041472003795206547, 'timing/model': 1.0615678359754384, 'learning_rate/qwen_vl_interface': 1.3386543201229213e-06, 'learning_rate/action_model': 1.3386543201229213e-05, 'epoch': 0.31}) +[2026-06-24 14:10:01,600][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.17037585377693176, 'train/grad_norm_pre_clip': 105.88467407226562, 'timing/data': 0.0003188790287822485, 'timing/model': 1.0717455518897623, 'learning_rate/qwen_vl_interface': 1.3364845412704553e-06, 'learning_rate/action_model': 1.3364845412704551e-05, 'epoch': 0.31}) +[2026-06-24 14:10:02,675][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.02886919304728508, 'train/grad_norm_pre_clip': 18.89332389831543, 'timing/data': 0.00040231598541140556, 'timing/model': 1.0733299299608916, 'learning_rate/qwen_vl_interface': 1.3343173018556835e-06, 'learning_rate/action_model': 1.3343173018556834e-05, 'epoch': 0.31}) +[2026-06-24 14:10:03,746][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.09874314814805984, 'train/grad_norm_pre_clip': 57.18198776245117, 'timing/data': 0.00035079894587397575, 'timing/model': 1.0694358020555228, 'learning_rate/qwen_vl_interface': 1.3321526032849034e-06, 'learning_rate/action_model': 1.3321526032849033e-05, 'epoch': 0.31}) +[2026-06-24 14:10:04,818][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.2793554663658142, 'train/grad_norm_pre_clip': 122.37494659423828, 'timing/data': 0.00035647000186145306, 'timing/model': 1.0704025190789253, 'learning_rate/qwen_vl_interface': 1.3299904469627608e-06, 'learning_rate/action_model': 1.3299904469627608e-05, 'epoch': 0.31}) +[2026-06-24 14:10:05,892][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.05180305242538452, 'train/grad_norm_pre_clip': 33.56046676635742, 'timing/data': 0.0003406801261007786, 'timing/model': 1.0719661470502615, 'learning_rate/qwen_vl_interface': 1.3278308342922607e-06, 'learning_rate/action_model': 1.3278308342922608e-05, 'epoch': 0.32}) +[2026-06-24 14:10:06,968][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.01997181959450245, 'train/grad_norm_pre_clip': 27.47727394104004, 'timing/data': 0.0003685920964926481, 'timing/model': 1.0744421109557152, 'learning_rate/qwen_vl_interface': 1.3256737666747488e-06, 'learning_rate/action_model': 1.3256737666747488e-05, 'epoch': 0.32}) +[2026-06-24 14:10:08,041][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.18661297857761383, 'train/grad_norm_pre_clip': 76.42791748046875, 'timing/data': 0.00031720614060759544, 'timing/model': 1.071580182062462, 'learning_rate/qwen_vl_interface': 1.3235192455099237e-06, 'learning_rate/action_model': 1.3235192455099239e-05, 'epoch': 0.32}) +[2026-06-24 14:10:09,115][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.11182799935340881, 'train/grad_norm_pre_clip': 83.13704681396484, 'timing/data': 0.0003246399573981762, 'timing/model': 1.072213283041492, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 0.32}) +[2026-06-24 14:10:10,190][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.18559208512306213, 'train/grad_norm_pre_clip': 92.82868957519531, 'timing/data': 0.00032093306072056293, 'timing/model': 1.0732003380544484, 'learning_rate/qwen_vl_interface': 1.3192178481288637e-06, 'learning_rate/action_model': 1.3192178481288636e-05, 'epoch': 0.32}) +[2026-06-24 14:10:11,264][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.016812017187476158, 'train/grad_norm_pre_clip': 20.607763290405273, 'timing/data': 0.00031924014911055565, 'timing/model': 1.0724359119776636, 'learning_rate/qwen_vl_interface': 1.3170709747037582e-06, 'learning_rate/action_model': 1.3170709747037583e-05, 'epoch': 0.32}) +[2026-06-24 14:10:12,342][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.004439265467226505, 'train/grad_norm_pre_clip': 8.35403823852539, 'timing/data': 0.00033136317506432533, 'timing/model': 1.0763327649328858, 'learning_rate/qwen_vl_interface': 1.314926653313598e-06, 'learning_rate/action_model': 1.314926653313598e-05, 'epoch': 0.32}) +[2026-06-24 14:10:13,418][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.02037494070827961, 'train/grad_norm_pre_clip': 17.48872184753418, 'timing/data': 0.00040391902439296246, 'timing/model': 1.0740897699724883, 'learning_rate/qwen_vl_interface': 1.312784885349811e-06, 'learning_rate/action_model': 1.3127848853498109e-05, 'epoch': 0.32}) +[2026-06-24 14:10:14,787][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.17404687404632568, 'train/grad_norm_pre_clip': 94.30614471435547, 'timing/data': 0.0003128370735794306, 'timing/model': 1.3676642018835992, 'learning_rate/qwen_vl_interface': 1.3106456722021648e-06, 'learning_rate/action_model': 1.3106456722021649e-05, 'epoch': 0.32}) +[2026-06-24 14:10:15,864][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.05776018276810646, 'train/grad_norm_pre_clip': 73.1060791015625, 'timing/data': 0.0003412209916859865, 'timing/model': 1.0753323549870402, 'learning_rate/qwen_vl_interface': 1.3085090152587735e-06, 'learning_rate/action_model': 1.3085090152587734e-05, 'epoch': 0.32}) +[2026-06-24 14:10:16,940][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.043056439608335495, 'train/grad_norm_pre_clip': 56.285621643066406, 'timing/data': 0.00035891495645046234, 'timing/model': 1.074112863978371, 'learning_rate/qwen_vl_interface': 1.3063749159060905e-06, 'learning_rate/action_model': 1.3063749159060904e-05, 'epoch': 0.32}) +[2026-06-24 14:10:18,014][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.07701180130243301, 'train/grad_norm_pre_clip': 56.74333953857422, 'timing/data': 0.0003517910372465849, 'timing/model': 1.0722252267878503, 'learning_rate/qwen_vl_interface': 1.3042433755289088e-06, 'learning_rate/action_model': 1.3042433755289089e-05, 'epoch': 0.32}) +[2026-06-24 14:10:19,091][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.016934452578425407, 'train/grad_norm_pre_clip': 24.60620880126953, 'timing/data': 0.00031484197825193405, 'timing/model': 1.075789714930579, 'learning_rate/qwen_vl_interface': 1.302114395510362e-06, 'learning_rate/action_model': 1.3021143955103618e-05, 'epoch': 0.32}) +[2026-06-24 14:10:20,169][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.10989458113908768, 'train/grad_norm_pre_clip': 58.8749885559082, 'timing/data': 0.00033289589919149876, 'timing/model': 1.076065604109317, 'learning_rate/qwen_vl_interface': 1.299987977231922e-06, 'learning_rate/action_model': 1.299987977231922e-05, 'epoch': 0.32}) +[2026-06-24 14:10:21,243][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.014521951787173748, 'train/grad_norm_pre_clip': 24.717525482177734, 'timing/data': 0.0003279969096183777, 'timing/model': 1.0730641230475157, 'learning_rate/qwen_vl_interface': 1.2978641220734e-06, 'learning_rate/action_model': 1.2978641220734e-05, 'epoch': 0.32}) +[2026-06-24 14:10:22,320][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.029403261840343475, 'train/grad_norm_pre_clip': 32.6436653137207, 'timing/data': 0.0003275859635323286, 'timing/model': 1.0744813969358802, 'learning_rate/qwen_vl_interface': 1.295742831412941e-06, 'learning_rate/action_model': 1.2957428314129411e-05, 'epoch': 0.32}) +[2026-06-24 14:10:23,398][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.06571444123983383, 'train/grad_norm_pre_clip': 56.124942779541016, 'timing/data': 0.00034370599314570427, 'timing/model': 1.076303691137582, 'learning_rate/qwen_vl_interface': 1.2936241066270278e-06, 'learning_rate/action_model': 1.2936241066270278e-05, 'epoch': 0.32}) +[2026-06-24 14:10:24,478][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.07458777725696564, 'train/grad_norm_pre_clip': 239.60012817382812, 'timing/data': 0.0003497770521789789, 'timing/model': 1.0786614429671317, 'learning_rate/qwen_vl_interface': 1.2915079490904786e-06, 'learning_rate/action_model': 1.2915079490904786e-05, 'epoch': 0.32}) +[2026-06-24 14:10:25,559][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.018306132405996323, 'train/grad_norm_pre_clip': 18.19734764099121, 'timing/data': 0.0003462198656052351, 'timing/model': 1.079430087003857, 'learning_rate/qwen_vl_interface': 1.289394360176443e-06, 'learning_rate/action_model': 1.289394360176443e-05, 'epoch': 0.32}) +[2026-06-24 14:10:26,640][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.036234479397535324, 'train/grad_norm_pre_clip': 87.8538589477539, 'timing/data': 0.00036607799120247364, 'timing/model': 1.0788854740094393, 'learning_rate/qwen_vl_interface': 1.287283341256409e-06, 'learning_rate/action_model': 1.2872833412564089e-05, 'epoch': 0.32}) +[2026-06-24 14:10:27,715][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.08466760814189911, 'train/grad_norm_pre_clip': 77.46603393554688, 'timing/data': 0.0004653849173337221, 'timing/model': 1.0733952671289444, 'learning_rate/qwen_vl_interface': 1.2851748937001927e-06, 'learning_rate/action_model': 1.2851748937001926e-05, 'epoch': 0.32}) +[2026-06-24 14:10:28,794][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.009703705087304115, 'train/grad_norm_pre_clip': 12.532272338867188, 'timing/data': 0.00032676407136023045, 'timing/model': 1.077749890042469, 'learning_rate/qwen_vl_interface': 1.2830690188759402e-06, 'learning_rate/action_model': 1.28306901887594e-05, 'epoch': 0.32}) +[2026-06-24 14:10:29,875][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.0921686515212059, 'train/grad_norm_pre_clip': 67.63080596923828, 'timing/data': 0.000534404069185257, 'timing/model': 1.078735482878983, 'learning_rate/qwen_vl_interface': 1.280965718150135e-06, 'learning_rate/action_model': 1.280965718150135e-05, 'epoch': 0.32}) +[2026-06-24 14:10:31,248][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.09657430648803711, 'train/grad_norm_pre_clip': 109.39940643310547, 'timing/data': 0.00035456707701087, 'timing/model': 1.3714232791680843, 'learning_rate/qwen_vl_interface': 1.278864992887583e-06, 'learning_rate/action_model': 1.278864992887583e-05, 'epoch': 0.32}) +[2026-06-24 14:10:32,330][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.023378221318125725, 'train/grad_norm_pre_clip': 26.79258155822754, 'timing/data': 0.00032284692861139774, 'timing/model': 1.0801228799391538, 'learning_rate/qwen_vl_interface': 1.276766844451423e-06, 'learning_rate/action_model': 1.276766844451423e-05, 'epoch': 0.32}) +[2026-06-24 14:10:33,415][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.15268409252166748, 'train/grad_norm_pre_clip': 77.55974578857422, 'timing/data': 0.0003540550824254751, 'timing/model': 1.083480620989576, 'learning_rate/qwen_vl_interface': 1.2746712742031207e-06, 'learning_rate/action_model': 1.2746712742031205e-05, 'epoch': 0.32}) +[2026-06-24 14:10:34,497][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.09130662679672241, 'train/grad_norm_pre_clip': 30.702974319458008, 'timing/data': 0.0003300600219517946, 'timing/model': 1.0802765691187233, 'learning_rate/qwen_vl_interface': 1.2725782835024705e-06, 'learning_rate/action_model': 1.2725782835024705e-05, 'epoch': 0.32}) +[2026-06-24 14:10:35,578][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.08120127767324448, 'train/grad_norm_pre_clip': 112.8630142211914, 'timing/data': 0.0003793030045926571, 'timing/model': 1.0793145110365003, 'learning_rate/qwen_vl_interface': 1.270487873707589e-06, 'learning_rate/action_model': 1.270487873707589e-05, 'epoch': 0.32}) +[2026-06-24 14:10:36,659][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.02056177332997322, 'train/grad_norm_pre_clip': 14.491622924804688, 'timing/data': 0.00033791479654610157, 'timing/model': 1.079693484120071, 'learning_rate/qwen_vl_interface': 1.2684000461749207e-06, 'learning_rate/action_model': 1.2684000461749207e-05, 'epoch': 0.32}) +[2026-06-24 14:10:37,744][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.016414109617471695, 'train/grad_norm_pre_clip': 11.805339813232422, 'timing/data': 0.0003516108263283968, 'timing/model': 1.082956274971366, 'learning_rate/qwen_vl_interface': 1.2663148022592365e-06, 'learning_rate/action_model': 1.2663148022592363e-05, 'epoch': 0.32}) +[2026-06-24 14:10:38,828][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.04708685353398323, 'train/grad_norm_pre_clip': 30.469650268554688, 'timing/data': 0.00032475986517965794, 'timing/model': 1.082118660910055, 'learning_rate/qwen_vl_interface': 1.2642321433136253e-06, 'learning_rate/action_model': 1.2642321433136254e-05, 'epoch': 0.32}) +[2026-06-24 14:10:39,909][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.004453970119357109, 'train/grad_norm_pre_clip': 3.466672897338867, 'timing/data': 0.00032242597080767155, 'timing/model': 1.079977166140452, 'learning_rate/qwen_vl_interface': 1.262152070689504e-06, 'learning_rate/action_model': 1.262152070689504e-05, 'epoch': 0.32}) +[2026-06-24 14:10:40,990][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.011363144032657146, 'train/grad_norm_pre_clip': 20.900129318237305, 'timing/data': 0.00053234095685184, 'timing/model': 1.078284596791491, 'learning_rate/qwen_vl_interface': 1.2600745857366093e-06, 'learning_rate/action_model': 1.2600745857366091e-05, 'epoch': 0.32}) +[2026-06-24 14:10:42,072][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.016744397580623627, 'train/grad_norm_pre_clip': 19.120704650878906, 'timing/data': 0.00032334798015654087, 'timing/model': 1.0808988299686462, 'learning_rate/qwen_vl_interface': 1.2579996898029967e-06, 'learning_rate/action_model': 1.2579996898029966e-05, 'epoch': 0.32}) +[2026-06-24 14:10:43,154][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.08367639034986496, 'train/grad_norm_pre_clip': 26.571151733398438, 'timing/data': 0.00030938186682760715, 'timing/model': 1.0799580111633986, 'learning_rate/qwen_vl_interface': 1.2559273842350452e-06, 'learning_rate/action_model': 1.2559273842350452e-05, 'epoch': 0.32}) +[2026-06-24 14:10:44,236][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.1382572203874588, 'train/grad_norm_pre_clip': 43.95468521118164, 'timing/data': 0.0003118261229246855, 'timing/model': 1.0807639360427856, 'learning_rate/qwen_vl_interface': 1.2538576703774507e-06, 'learning_rate/action_model': 1.2538576703774507e-05, 'epoch': 0.32}) +[2026-06-24 14:10:45,316][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.004276209510862827, 'train/grad_norm_pre_clip': 8.592949867248535, 'timing/data': 0.00033238413743674755, 'timing/model': 1.078160973964259, 'learning_rate/qwen_vl_interface': 1.25179054957323e-06, 'learning_rate/action_model': 1.2517905495732298e-05, 'epoch': 0.32}) +[2026-06-24 14:10:46,601][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.0013788904761895537, 'train/grad_norm_pre_clip': 1.0161575078964233, 'timing/data': 0.00032874802127480507, 'timing/model': 1.2831281090620905, 'learning_rate/qwen_vl_interface': 1.2497260231637121e-06, 'learning_rate/action_model': 1.249726023163712e-05, 'epoch': 0.32}) +[2026-06-24 14:10:47,678][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.027635939419269562, 'train/grad_norm_pre_clip': 19.725431442260742, 'timing/data': 0.0003532941918820143, 'timing/model': 1.0759542980231345, 'learning_rate/qwen_vl_interface': 1.2476640924885478e-06, 'learning_rate/action_model': 1.2476640924885476e-05, 'epoch': 0.32}) +[2026-06-24 14:10:48,755][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.19363218545913696, 'train/grad_norm_pre_clip': 79.32215881347656, 'timing/data': 0.0003404701128602028, 'timing/model': 1.0747588509693742, 'learning_rate/qwen_vl_interface': 1.2456047588857022e-06, 'learning_rate/action_model': 1.2456047588857023e-05, 'epoch': 0.32}) +[2026-06-24 14:10:49,831][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.09001768380403519, 'train/grad_norm_pre_clip': 53.38690948486328, 'timing/data': 0.00033656181767582893, 'timing/model': 1.0745998818892986, 'learning_rate/qwen_vl_interface': 1.243548023691451e-06, 'learning_rate/action_model': 1.2435480236914509e-05, 'epoch': 0.32}) +[2026-06-24 14:10:50,911][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.025291630998253822, 'train/grad_norm_pre_clip': 13.080512046813965, 'timing/data': 0.00035986583679914474, 'timing/model': 1.0788250628393143, 'learning_rate/qwen_vl_interface': 1.2414938882403923e-06, 'learning_rate/action_model': 1.2414938882403923e-05, 'epoch': 0.32}) +[2026-06-24 14:10:51,995][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.06754682213068008, 'train/grad_norm_pre_clip': 29.763322830200195, 'timing/data': 0.0005254170391708612, 'timing/model': 1.0818209128919989, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 0.32}) +[2026-06-24 14:10:53,078][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.013224819675087929, 'train/grad_norm_pre_clip': 7.873900890350342, 'timing/data': 0.0003261328674852848, 'timing/model': 1.0812767709139735, 'learning_rate/qwen_vl_interface': 1.2373934218977795e-06, 'learning_rate/action_model': 1.2373934218977795e-05, 'epoch': 0.32}) +[2026-06-24 14:10:54,160][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.02042991854250431, 'train/grad_norm_pre_clip': 13.26923656463623, 'timing/data': 0.0003592350985854864, 'timing/model': 1.0799655669834465, 'learning_rate/qwen_vl_interface': 1.2353470936669757e-06, 'learning_rate/action_model': 1.2353470936669756e-05, 'epoch': 0.32}) +[2026-06-24 14:10:55,239][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.2975757122039795, 'train/grad_norm_pre_clip': 60.576438903808594, 'timing/data': 0.0003343080170452595, 'timing/model': 1.077884596074, 'learning_rate/qwen_vl_interface': 1.2333033705008549e-06, 'learning_rate/action_model': 1.2333033705008548e-05, 'epoch': 0.32}) +[2026-06-24 14:10:56,316][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.023449121043086052, 'train/grad_norm_pre_clip': 10.818279266357422, 'timing/data': 0.0003518511075526476, 'timing/model': 1.0753379110246897, 'learning_rate/qwen_vl_interface': 1.231262253725569e-06, 'learning_rate/action_model': 1.2312622537255688e-05, 'epoch': 0.32}) +[2026-06-24 14:10:57,391][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.10363539308309555, 'train/grad_norm_pre_clip': 41.828521728515625, 'timing/data': 0.00037666806019842625, 'timing/model': 1.0736862069461495, 'learning_rate/qwen_vl_interface': 1.2292237446655764e-06, 'learning_rate/action_model': 1.2292237446655764e-05, 'epoch': 0.32}) +[2026-06-24 14:10:58,470][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.004320567008107901, 'train/grad_norm_pre_clip': 5.899996757507324, 'timing/data': 0.00035098008811473846, 'timing/model': 1.0767969528678805, 'learning_rate/qwen_vl_interface': 1.2271878446436426e-06, 'learning_rate/action_model': 1.2271878446436425e-05, 'epoch': 0.32}) +[2026-06-24 14:10:59,551][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.14469215273857117, 'train/grad_norm_pre_clip': 124.5828857421875, 'timing/data': 0.00032781600020825863, 'timing/model': 1.0792061598040164, 'learning_rate/qwen_vl_interface': 1.2251545549808405e-06, 'learning_rate/action_model': 1.2251545549808404e-05, 'epoch': 0.32}) +[2026-06-24 14:11:00,631][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.06918305903673172, 'train/grad_norm_pre_clip': 38.5312614440918, 'timing/data': 0.0003179670311510563, 'timing/model': 1.0784091020468622, 'learning_rate/qwen_vl_interface': 1.223123876996551e-06, 'learning_rate/action_model': 1.2231238769965509e-05, 'epoch': 0.32}) +[2026-06-24 14:11:01,711][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.09975861757993698, 'train/grad_norm_pre_clip': 38.27811050415039, 'timing/data': 0.0003570208791643381, 'timing/model': 1.0785312519874424, 'learning_rate/qwen_vl_interface': 1.2210958120084596e-06, 'learning_rate/action_model': 1.2210958120084595e-05, 'epoch': 0.32}) +[2026-06-24 14:11:03,081][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.04411611706018448, 'train/grad_norm_pre_clip': 33.17774200439453, 'timing/data': 0.0003833300434052944, 'timing/model': 1.3683660889510065, 'learning_rate/qwen_vl_interface': 1.2190703613325532e-06, 'learning_rate/action_model': 1.2190703613325532e-05, 'epoch': 0.32}) +[2026-06-24 14:11:04,156][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.0174727700650692, 'train/grad_norm_pre_clip': 32.252567291259766, 'timing/data': 0.00032580201514065266, 'timing/model': 1.0736321059521288, 'learning_rate/qwen_vl_interface': 1.2170475262831262e-06, 'learning_rate/action_model': 1.2170475262831264e-05, 'epoch': 0.32}) +[2026-06-24 14:11:05,232][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.006623104214668274, 'train/grad_norm_pre_clip': 6.434869289398193, 'timing/data': 0.00031362916342914104, 'timing/model': 1.0743595929816365, 'learning_rate/qwen_vl_interface': 1.2150273081727758e-06, 'learning_rate/action_model': 1.2150273081727758e-05, 'epoch': 0.32}) +[2026-06-24 14:11:06,309][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.10363903641700745, 'train/grad_norm_pre_clip': 67.067626953125, 'timing/data': 0.000326354056596756, 'timing/model': 1.07514069089666, 'learning_rate/qwen_vl_interface': 1.2130097083123971e-06, 'learning_rate/action_model': 1.213009708312397e-05, 'epoch': 0.32}) +[2026-06-24 14:11:07,384][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.04231119528412819, 'train/grad_norm_pre_clip': 24.986385345458984, 'timing/data': 0.0004208809696137905, 'timing/model': 1.0732152620330453, 'learning_rate/qwen_vl_interface': 1.2109947280111892e-06, 'learning_rate/action_model': 1.210994728011189e-05, 'epoch': 0.32}) +[2026-06-24 14:11:08,461][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.3661566972732544, 'train/grad_norm_pre_clip': 124.0101089477539, 'timing/data': 0.0003388170152902603, 'timing/model': 1.075632505118847, 'learning_rate/qwen_vl_interface': 1.2089823685766515e-06, 'learning_rate/action_model': 1.2089823685766516e-05, 'epoch': 0.32}) +[2026-06-24 14:11:09,538][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.0035856622271239758, 'train/grad_norm_pre_clip': 7.759840488433838, 'timing/data': 0.00032119289971888065, 'timing/model': 1.075192257994786, 'learning_rate/qwen_vl_interface': 1.2069726313145836e-06, 'learning_rate/action_model': 1.2069726313145834e-05, 'epoch': 0.32}) +[2026-06-24 14:11:10,614][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.18017874658107758, 'train/grad_norm_pre_clip': 84.15606689453125, 'timing/data': 0.0003210729919373989, 'timing/model': 1.0745974101591855, 'learning_rate/qwen_vl_interface': 1.2049655175290795e-06, 'learning_rate/action_model': 1.2049655175290795e-05, 'epoch': 0.32}) +[2026-06-24 14:11:11,691][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.10366689413785934, 'train/grad_norm_pre_clip': 52.01401901245117, 'timing/data': 0.00031982199288904667, 'timing/model': 1.0755425170063972, 'learning_rate/qwen_vl_interface': 1.202961028522534e-06, 'learning_rate/action_model': 1.202961028522534e-05, 'epoch': 0.32}) +[2026-06-24 14:11:12,767][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.09741280227899551, 'train/grad_norm_pre_clip': 104.50871276855469, 'timing/data': 0.0003178578335791826, 'timing/model': 1.0746697469148785, 'learning_rate/qwen_vl_interface': 1.200959165595641e-06, 'learning_rate/action_model': 1.2009591655956409e-05, 'epoch': 0.32}) +[2026-06-24 14:11:13,842][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.04089587181806564, 'train/grad_norm_pre_clip': 33.238834381103516, 'timing/data': 0.0005156591068953276, 'timing/model': 1.0729002908337861, 'learning_rate/qwen_vl_interface': 1.1989599300473828e-06, 'learning_rate/action_model': 1.1989599300473828e-05, 'epoch': 0.32}) +[2026-06-24 14:11:14,918][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.044743169099092484, 'train/grad_norm_pre_clip': 31.412914276123047, 'timing/data': 0.0003905440680682659, 'timing/model': 1.0743676291313022, 'learning_rate/qwen_vl_interface': 1.1969633231750474e-06, 'learning_rate/action_model': 1.1969633231750474e-05, 'epoch': 0.32}) +[2026-06-24 14:11:15,994][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.3567415177822113, 'train/grad_norm_pre_clip': 14.039624214172363, 'timing/data': 0.00033701397478580475, 'timing/model': 1.0738748030271381, 'learning_rate/qwen_vl_interface': 1.1949693462742078e-06, 'learning_rate/action_model': 1.1949693462742079e-05, 'epoch': 0.32}) +[2026-06-24 14:11:17,070][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.31447556614875793, 'train/grad_norm_pre_clip': 113.6678237915039, 'timing/data': 0.0003523719497025013, 'timing/model': 1.0742391680832952, 'learning_rate/qwen_vl_interface': 1.1929780006387326e-06, 'learning_rate/action_model': 1.1929780006387326e-05, 'epoch': 0.32}) +[2026-06-24 14:11:18,355][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.034745410084724426, 'train/grad_norm_pre_clip': 8.307512283325195, 'timing/data': 0.00037549599073827267, 'timing/model': 1.283629324985668, 'learning_rate/qwen_vl_interface': 1.1909892875607889e-06, 'learning_rate/action_model': 1.1909892875607889e-05, 'epoch': 0.32}) +[2026-06-24 14:11:19,430][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.0018681787187233567, 'train/grad_norm_pre_clip': 1.6635087728500366, 'timing/data': 0.0003265729174017906, 'timing/model': 1.0731131320353597, 'learning_rate/qwen_vl_interface': 1.1890032083308267e-06, 'learning_rate/action_model': 1.1890032083308268e-05, 'epoch': 0.32}) +[2026-06-24 14:11:20,503][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.14817026257514954, 'train/grad_norm_pre_clip': 82.5414047241211, 'timing/data': 0.0003198611084371805, 'timing/model': 1.0715107009746134, 'learning_rate/qwen_vl_interface': 1.1870197642375935e-06, 'learning_rate/action_model': 1.1870197642375935e-05, 'epoch': 0.32}) +[2026-06-24 14:11:21,579][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.16878579556941986, 'train/grad_norm_pre_clip': 114.68917083740234, 'timing/data': 0.000323727959766984, 'timing/model': 1.074280435917899, 'learning_rate/qwen_vl_interface': 1.1850389565681246e-06, 'learning_rate/action_model': 1.1850389565681248e-05, 'epoch': 0.32}) +[2026-06-24 14:11:22,655][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.10708124190568924, 'train/grad_norm_pre_clip': 51.447811126708984, 'timing/data': 0.000311295036226511, 'timing/model': 1.0745903081260622, 'learning_rate/qwen_vl_interface': 1.1830607866077438e-06, 'learning_rate/action_model': 1.1830607866077437e-05, 'epoch': 0.32}) +[2026-06-24 14:11:23,731][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.02008621208369732, 'train/grad_norm_pre_clip': 10.998392105102539, 'timing/data': 0.00034862617030739784, 'timing/model': 1.0749464179389179, 'learning_rate/qwen_vl_interface': 1.181085255640064e-06, 'learning_rate/action_model': 1.181085255640064e-05, 'epoch': 0.32}) +[2026-06-24 14:11:24,807][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.07871190458536148, 'train/grad_norm_pre_clip': 53.058631896972656, 'timing/data': 0.0003357010427862406, 'timing/model': 1.0737580540589988, 'learning_rate/qwen_vl_interface': 1.1791123649469865e-06, 'learning_rate/action_model': 1.1791123649469865e-05, 'epoch': 0.32}) +[2026-06-24 14:11:25,884][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.0008608490461483598, 'train/grad_norm_pre_clip': 0.5759685635566711, 'timing/data': 0.00032565207220613956, 'timing/model': 1.0752802861388773, 'learning_rate/qwen_vl_interface': 1.1771421158086993e-06, 'learning_rate/action_model': 1.1771421158086993e-05, 'epoch': 0.32}) +[2026-06-24 14:11:26,960][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.018636558204889297, 'train/grad_norm_pre_clip': 9.06303882598877, 'timing/data': 0.0003800338599830866, 'timing/model': 1.075223988853395, 'learning_rate/qwen_vl_interface': 1.1751745095036737e-06, 'learning_rate/action_model': 1.1751745095036737e-05, 'epoch': 0.32}) +[2026-06-24 14:11:28,034][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.11930200457572937, 'train/grad_norm_pre_clip': 64.09720611572266, 'timing/data': 0.0003168061375617981, 'timing/model': 1.0723192209843546, 'learning_rate/qwen_vl_interface': 1.1732095473086697e-06, 'learning_rate/action_model': 1.1732095473086695e-05, 'epoch': 0.32}) +[2026-06-24 14:11:29,110][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.19302016496658325, 'train/grad_norm_pre_clip': 81.6977767944336, 'timing/data': 0.000395443057641387, 'timing/model': 1.0744942091405392, 'learning_rate/qwen_vl_interface': 1.17124723049873e-06, 'learning_rate/action_model': 1.17124723049873e-05, 'epoch': 0.32}) +[2026-06-24 14:11:30,186][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.028002789244055748, 'train/grad_norm_pre_clip': 34.584922790527344, 'timing/data': 0.00033866590820252895, 'timing/model': 1.074225313961506, 'learning_rate/qwen_vl_interface': 1.1692875603471776e-06, 'learning_rate/action_model': 1.1692875603471775e-05, 'epoch': 0.32}) +[2026-06-24 14:11:31,263][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.2496362030506134, 'train/grad_norm_pre_clip': 119.67436981201172, 'timing/data': 0.00034531787969172, 'timing/model': 1.0751349017955363, 'learning_rate/qwen_vl_interface': 1.1673305381256256e-06, 'learning_rate/action_model': 1.1673305381256256e-05, 'epoch': 0.32}) +[2026-06-24 14:11:32,340][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.11124896258115768, 'train/grad_norm_pre_clip': 76.09756469726562, 'timing/data': 0.0004027069080621004, 'timing/model': 1.075838646152988, 'learning_rate/qwen_vl_interface': 1.1653761651039622e-06, 'learning_rate/action_model': 1.1653761651039622e-05, 'epoch': 0.32}) +[2026-06-24 14:11:33,417][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.020167594775557518, 'train/grad_norm_pre_clip': 40.17511749267578, 'timing/data': 0.00031153601594269276, 'timing/model': 1.0752177080139518, 'learning_rate/qwen_vl_interface': 1.1634244425503585e-06, 'learning_rate/action_model': 1.1634244425503586e-05, 'epoch': 0.32}) +[2026-06-24 14:11:34,495][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.07874312251806259, 'train/grad_norm_pre_clip': 55.7385139465332, 'timing/data': 0.0003162939101457596, 'timing/model': 1.0762098028790206, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 0.32}) +[2026-06-24 14:11:35,575][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.04070044681429863, 'train/grad_norm_pre_clip': 32.35900115966797, 'timing/data': 0.00034403707832098007, 'timing/model': 1.0778114129789174, 'learning_rate/qwen_vl_interface': 1.1595289539114193e-06, 'learning_rate/action_model': 1.1595289539114193e-05, 'epoch': 0.32}) +[2026-06-24 14:11:36,651][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.06410863995552063, 'train/grad_norm_pre_clip': 69.81700134277344, 'timing/data': 0.0003187290858477354, 'timing/model': 1.0753747739363462, 'learning_rate/qwen_vl_interface': 1.1575851903538226e-06, 'learning_rate/action_model': 1.1575851903538227e-05, 'epoch': 0.32}) +[2026-06-24 14:11:38,016][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.07262343168258667, 'train/grad_norm_pre_clip': 63.52583312988281, 'timing/data': 0.0003204219974577427, 'timing/model': 1.363336490932852, 'learning_rate/qwen_vl_interface': 1.155644082319766e-06, 'learning_rate/action_model': 1.155644082319766e-05, 'epoch': 0.32}) +[2026-06-24 14:11:39,097][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.24105285108089447, 'train/grad_norm_pre_clip': 59.60924530029297, 'timing/data': 0.00047173607163131237, 'timing/model': 1.07876974507235, 'learning_rate/qwen_vl_interface': 1.1537056310688128e-06, 'learning_rate/action_model': 1.1537056310688127e-05, 'epoch': 0.32}) +[2026-06-24 14:11:40,175][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.040117599070072174, 'train/grad_norm_pre_clip': 23.984914779663086, 'timing/data': 0.00036153895780444145, 'timing/model': 1.0763775389641523, 'learning_rate/qwen_vl_interface': 1.1517698378588021e-06, 'learning_rate/action_model': 1.1517698378588021e-05, 'epoch': 0.32}) +[2026-06-24 14:11:41,253][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.18823368847370148, 'train/grad_norm_pre_clip': 58.93000411987305, 'timing/data': 0.00031406013295054436, 'timing/model': 1.0766445309855044, 'learning_rate/qwen_vl_interface': 1.1498367039458493e-06, 'learning_rate/action_model': 1.1498367039458494e-05, 'epoch': 0.32}) +[2026-06-24 14:11:42,333][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.03593819960951805, 'train/grad_norm_pre_clip': 31.833467483520508, 'timing/data': 0.0003105239011347294, 'timing/model': 1.0779996779747307, 'learning_rate/qwen_vl_interface': 1.147906230584346e-06, 'learning_rate/action_model': 1.1479062305843461e-05, 'epoch': 0.32}) +[2026-06-24 14:11:43,409][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.24109399318695068, 'train/grad_norm_pre_clip': 88.06294250488281, 'timing/data': 0.0003682919777929783, 'timing/model': 1.0746963899582624, 'learning_rate/qwen_vl_interface': 1.1459784190269527e-06, 'learning_rate/action_model': 1.1459784190269527e-05, 'epoch': 0.32}) +[2026-06-24 14:11:44,488][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.09971322864294052, 'train/grad_norm_pre_clip': 87.10245513916016, 'timing/data': 0.0003241489175707102, 'timing/model': 1.0774577287957072, 'learning_rate/qwen_vl_interface': 1.1440532705246066e-06, 'learning_rate/action_model': 1.1440532705246065e-05, 'epoch': 0.32}) +[2026-06-24 14:11:45,572][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.003625560784712434, 'train/grad_norm_pre_clip': 4.392186164855957, 'timing/data': 0.00031617400236427784, 'timing/model': 1.0815114881843328, 'learning_rate/qwen_vl_interface': 1.1421307863265163e-06, 'learning_rate/action_model': 1.1421307863265162e-05, 'epoch': 0.32}) +[2026-06-24 14:11:46,654][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.19813500344753265, 'train/grad_norm_pre_clip': 115.16118621826172, 'timing/data': 0.00036702887155115604, 'timing/model': 1.080629691015929, 'learning_rate/qwen_vl_interface': 1.1402109676801598e-06, 'learning_rate/action_model': 1.1402109676801599e-05, 'epoch': 0.32}) +[2026-06-24 14:11:47,734][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.22511157393455505, 'train/grad_norm_pre_clip': 95.83198547363281, 'timing/data': 0.00032400921918451786, 'timing/model': 1.078359215054661, 'learning_rate/qwen_vl_interface': 1.1382938158312865e-06, 'learning_rate/action_model': 1.1382938158312865e-05, 'epoch': 0.32}) +[2026-06-24 14:11:48,816][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.2027355432510376, 'train/grad_norm_pre_clip': 83.1717529296875, 'timing/data': 0.00032172491773962975, 'timing/model': 1.080088975140825, 'learning_rate/qwen_vl_interface': 1.136379332023916e-06, 'learning_rate/action_model': 1.1363793320239161e-05, 'epoch': 0.32}) +[2026-06-24 14:11:49,897][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.017778199166059494, 'train/grad_norm_pre_clip': 15.204258918762207, 'timing/data': 0.00033737393096089363, 'timing/model': 1.0796578340232372, 'learning_rate/qwen_vl_interface': 1.1344675175003378e-06, 'learning_rate/action_model': 1.1344675175003378e-05, 'epoch': 0.32}) +[2026-06-24 14:11:50,979][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.061377979815006256, 'train/grad_norm_pre_clip': 91.44490814208984, 'timing/data': 0.00030988198705017567, 'timing/model': 1.0808774770703167, 'learning_rate/qwen_vl_interface': 1.132558373501104e-06, 'learning_rate/action_model': 1.1325583735011038e-05, 'epoch': 0.32}) +[2026-06-24 14:11:52,061][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.016329273581504822, 'train/grad_norm_pre_clip': 23.086132049560547, 'timing/data': 0.00033392803743481636, 'timing/model': 1.0796349421143532, 'learning_rate/qwen_vl_interface': 1.130651901265039e-06, 'learning_rate/action_model': 1.1306519012650391e-05, 'epoch': 0.32}) +[2026-06-24 14:11:53,143][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.04407922923564911, 'train/grad_norm_pre_clip': 21.93490982055664, 'timing/data': 0.0004037190228700638, 'timing/model': 1.0808681100606918, 'learning_rate/qwen_vl_interface': 1.128748102029234e-06, 'learning_rate/action_model': 1.1287481020292339e-05, 'epoch': 0.32}) +[2026-06-24 14:11:54,227][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.12183784693479538, 'train/grad_norm_pre_clip': 58.69675064086914, 'timing/data': 0.00031415000557899475, 'timing/model': 1.0823775469325483, 'learning_rate/qwen_vl_interface': 1.1268469770290397e-06, 'learning_rate/action_model': 1.1268469770290396e-05, 'epoch': 0.32}) +[2026-06-24 14:11:55,309][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.017569471150636673, 'train/grad_norm_pre_clip': 13.484682083129883, 'timing/data': 0.00033685402013361454, 'timing/model': 1.0799037371762097, 'learning_rate/qwen_vl_interface': 1.1249485274980793e-06, 'learning_rate/action_model': 1.1249485274980792e-05, 'epoch': 0.32}) +[2026-06-24 14:11:56,632][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.05873331427574158, 'train/grad_norm_pre_clip': 64.42011260986328, 'timing/data': 0.00033139297738671303, 'timing/model': 1.3210990570951253, 'learning_rate/qwen_vl_interface': 1.1230527546682332e-06, 'learning_rate/action_model': 1.1230527546682333e-05, 'epoch': 0.32}) +[2026-06-24 14:11:57,722][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.00910451915115118, 'train/grad_norm_pre_clip': 14.262900352478027, 'timing/data': 0.00034776306711137295, 'timing/model': 1.084688579197973, 'learning_rate/qwen_vl_interface': 1.121159659769649e-06, 'learning_rate/action_model': 1.1211596597696489e-05, 'epoch': 0.32}) +[2026-06-24 14:11:58,835][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.1845162808895111, 'train/grad_norm_pre_clip': 78.04779815673828, 'timing/data': 0.0047343079932034016, 'timing/model': 1.1038918590638787, 'learning_rate/qwen_vl_interface': 1.1192692440307365e-06, 'learning_rate/action_model': 1.1192692440307364e-05, 'epoch': 0.32}) +[2026-06-24 14:11:59,915][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.22695153951644897, 'train/grad_norm_pre_clip': 93.23309326171875, 'timing/data': 0.00031526293605566025, 'timing/model': 1.0784392459318042, 'learning_rate/qwen_vl_interface': 1.117381508678164e-06, 'learning_rate/action_model': 1.1173815086781638e-05, 'epoch': 0.33}) +[2026-06-24 14:12:00,995][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.026845812797546387, 'train/grad_norm_pre_clip': 33.10556411743164, 'timing/data': 0.0003226171247661114, 'timing/model': 1.0783678418956697, 'learning_rate/qwen_vl_interface': 1.115496454936863e-06, 'learning_rate/action_model': 1.115496454936863e-05, 'epoch': 0.33}) +[2026-06-24 14:12:02,080][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.016063693910837173, 'train/grad_norm_pre_clip': 12.78730583190918, 'timing/data': 0.0003113551065325737, 'timing/model': 1.0830252051819116, 'learning_rate/qwen_vl_interface': 1.1136140840300249e-06, 'learning_rate/action_model': 1.1136140840300248e-05, 'epoch': 0.33}) +[2026-06-24 14:12:03,162][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.0968044176697731, 'train/grad_norm_pre_clip': 64.92236328125, 'timing/data': 0.0003552180714905262, 'timing/model': 1.0802619110327214, 'learning_rate/qwen_vl_interface': 1.1117343971791005e-06, 'learning_rate/action_model': 1.1117343971791006e-05, 'epoch': 0.33}) +[2026-06-24 14:12:04,244][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.09818135201931, 'train/grad_norm_pre_clip': 71.76815795898438, 'timing/data': 0.00032909796573221684, 'timing/model': 1.0810692489612848, 'learning_rate/qwen_vl_interface': 1.1098573956037963e-06, 'learning_rate/action_model': 1.1098573956037963e-05, 'epoch': 0.33}) +[2026-06-24 14:12:05,331][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.025023331865668297, 'train/grad_norm_pre_clip': 18.199031829833984, 'timing/data': 0.00037352205254137516, 'timing/model': 1.0846801849547774, 'learning_rate/qwen_vl_interface': 1.1079830805220796e-06, 'learning_rate/action_model': 1.1079830805220794e-05, 'epoch': 0.33}) +[2026-06-24 14:12:06,414][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.16113735735416412, 'train/grad_norm_pre_clip': 56.593727111816406, 'timing/data': 0.00033217412419617176, 'timing/model': 1.0813120359089226, 'learning_rate/qwen_vl_interface': 1.1061114531501728e-06, 'learning_rate/action_model': 1.1061114531501729e-05, 'epoch': 0.33}) +[2026-06-24 14:12:07,497][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.019050678238272667, 'train/grad_norm_pre_clip': 7.56262731552124, 'timing/data': 0.000337755074724555, 'timing/model': 1.0811668029055, 'learning_rate/qwen_vl_interface': 1.1042425147025543e-06, 'learning_rate/action_model': 1.1042425147025542e-05, 'epoch': 0.33}) +[2026-06-24 14:12:08,580][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.03036465309560299, 'train/grad_norm_pre_clip': 48.1496467590332, 'timing/data': 0.0003717790823429823, 'timing/model': 1.081884091021493, 'learning_rate/qwen_vl_interface': 1.1023762663919578e-06, 'learning_rate/action_model': 1.102376266391958e-05, 'epoch': 0.33}) +[2026-06-24 14:12:09,664][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.0945577546954155, 'train/grad_norm_pre_clip': 78.41600036621094, 'timing/data': 0.0003239191137254238, 'timing/model': 1.0821790948975831, 'learning_rate/qwen_vl_interface': 1.1005127094293736e-06, 'learning_rate/action_model': 1.1005127094293735e-05, 'epoch': 0.33}) +[2026-06-24 14:12:10,749][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.13084110617637634, 'train/grad_norm_pre_clip': 80.12284088134766, 'timing/data': 0.0003404100425541401, 'timing/model': 1.082918114028871, 'learning_rate/qwen_vl_interface': 1.09865184502404e-06, 'learning_rate/action_model': 1.0986518450240399e-05, 'epoch': 0.33}) +[2026-06-24 14:12:11,836][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.06442443281412125, 'train/grad_norm_pre_clip': 43.9033203125, 'timing/data': 0.00035476707853376865, 'timing/model': 1.0854492511134595, 'learning_rate/qwen_vl_interface': 1.096793674383454e-06, 'learning_rate/action_model': 1.0967936743834541e-05, 'epoch': 0.33}) +[2026-06-24 14:12:12,922][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.15687260031700134, 'train/grad_norm_pre_clip': 47.3997802734375, 'timing/data': 0.00044270209036767483, 'timing/model': 1.0844823140650988, 'learning_rate/qwen_vl_interface': 1.094938198713362e-06, 'learning_rate/action_model': 1.094938198713362e-05, 'epoch': 0.33}) +[2026-06-24 14:12:14,007][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.11467520147562027, 'train/grad_norm_pre_clip': 87.28633880615234, 'timing/data': 0.00033947802148759365, 'timing/model': 1.0833271129522473, 'learning_rate/qwen_vl_interface': 1.0930854192177632e-06, 'learning_rate/action_model': 1.093085419217763e-05, 'epoch': 0.33}) +[2026-06-24 14:12:15,305][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.005919542163610458, 'train/grad_norm_pre_clip': 8.936999320983887, 'timing/data': 0.00036202999763190746, 'timing/model': 1.2884261161088943, 'learning_rate/qwen_vl_interface': 1.0912353370989034e-06, 'learning_rate/action_model': 1.0912353370989034e-05, 'epoch': 0.33}) +[2026-06-24 14:12:16,392][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.005591238848865032, 'train/grad_norm_pre_clip': 10.617178916931152, 'timing/data': 0.00038004503585398197, 'timing/model': 1.0848359290976077, 'learning_rate/qwen_vl_interface': 1.0893879535572826e-06, 'learning_rate/action_model': 1.0893879535572825e-05, 'epoch': 0.33}) +[2026-06-24 14:12:17,472][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.27799102663993835, 'train/grad_norm_pre_clip': 109.8084945678711, 'timing/data': 0.00031842803582549095, 'timing/model': 1.0788417940493673, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 0.33}) +[2026-06-24 14:12:18,559][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.07390693575143814, 'train/grad_norm_pre_clip': 45.033302307128906, 'timing/data': 0.0003164249937981367, 'timing/model': 1.08550467598252, 'learning_rate/qwen_vl_interface': 1.0857012869989937e-06, 'learning_rate/action_model': 1.0857012869989935e-05, 'epoch': 0.33}) +[2026-06-24 14:12:19,643][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.05077133700251579, 'train/grad_norm_pre_clip': 42.957027435302734, 'timing/data': 0.00033724401146173477, 'timing/model': 1.0826020510867238, 'learning_rate/qwen_vl_interface': 1.0838620063745645e-06, 'learning_rate/action_model': 1.0838620063745644e-05, 'epoch': 0.33}) +[2026-06-24 14:12:20,728][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.10194966197013855, 'train/grad_norm_pre_clip': 55.89011764526367, 'timing/data': 0.0003116452135145664, 'timing/model': 1.0834589409641922, 'learning_rate/qwen_vl_interface': 1.0820254291118477e-06, 'learning_rate/action_model': 1.0820254291118478e-05, 'epoch': 0.33}) +[2026-06-24 14:12:21,812][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.22977744042873383, 'train/grad_norm_pre_clip': 42.70994567871094, 'timing/data': 0.00030185701325535774, 'timing/model': 1.0821286810096353, 'learning_rate/qwen_vl_interface': 1.0801915564025782e-06, 'learning_rate/action_model': 1.080191556402578e-05, 'epoch': 0.33}) +[2026-06-24 14:12:22,901][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.010272768326103687, 'train/grad_norm_pre_clip': 9.56717300415039, 'timing/data': 0.00034024985507130623, 'timing/model': 1.0871929589193314, 'learning_rate/qwen_vl_interface': 1.0783603894367377e-06, 'learning_rate/action_model': 1.0783603894367377e-05, 'epoch': 0.33}) +[2026-06-24 14:12:23,988][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.3299604654312134, 'train/grad_norm_pre_clip': 49.75115203857422, 'timing/data': 0.00031668483279645443, 'timing/model': 1.085158727131784, 'learning_rate/qwen_vl_interface': 1.076531929402549e-06, 'learning_rate/action_model': 1.0765319294025492e-05, 'epoch': 0.33}) +[2026-06-24 14:12:25,076][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.0005304927472025156, 'train/grad_norm_pre_clip': 0.5145897269248962, 'timing/data': 0.0003200620412826538, 'timing/model': 1.0863135459367186, 'learning_rate/qwen_vl_interface': 1.0747061774864812e-06, 'learning_rate/action_model': 1.0747061774864811e-05, 'epoch': 0.33}) +[2026-06-24 14:12:26,165][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.0642697662115097, 'train/grad_norm_pre_clip': 51.852535247802734, 'timing/data': 0.00039220694452524185, 'timing/model': 1.0873534008860588, 'learning_rate/qwen_vl_interface': 1.0728831348732435e-06, 'learning_rate/action_model': 1.0728831348732436e-05, 'epoch': 0.33}) +[2026-06-24 14:12:27,252][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.018077408894896507, 'train/grad_norm_pre_clip': 24.031232833862305, 'timing/data': 0.0003420230932533741, 'timing/model': 1.0850438810884953, 'learning_rate/qwen_vl_interface': 1.0710628027457909e-06, 'learning_rate/action_model': 1.0710628027457908e-05, 'epoch': 0.33}) +[2026-06-24 14:12:28,339][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.009315443225204945, 'train/grad_norm_pre_clip': 7.7897443771362305, 'timing/data': 0.00034918682649731636, 'timing/model': 1.0858272528275847, 'learning_rate/qwen_vl_interface': 1.0692451822853136e-06, 'learning_rate/action_model': 1.0692451822853135e-05, 'epoch': 0.33}) +[2026-06-24 14:12:29,426][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.01502048410475254, 'train/grad_norm_pre_clip': 15.860576629638672, 'timing/data': 0.0003279771190136671, 'timing/model': 1.0853817160241306, 'learning_rate/qwen_vl_interface': 1.0674302746712485e-06, 'learning_rate/action_model': 1.0674302746712483e-05, 'epoch': 0.33}) +[2026-06-24 14:12:30,517][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.18875761330127716, 'train/grad_norm_pre_clip': 73.11302185058594, 'timing/data': 0.0003415520768612623, 'timing/model': 1.0885856670793146, 'learning_rate/qwen_vl_interface': 1.0656180810812695e-06, 'learning_rate/action_model': 1.0656180810812694e-05, 'epoch': 0.33}) +[2026-06-24 14:12:31,603][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.08379720896482468, 'train/grad_norm_pre_clip': 32.082420349121094, 'timing/data': 0.0003230071160942316, 'timing/model': 1.08498247503303, 'learning_rate/qwen_vl_interface': 1.0638086026912873e-06, 'learning_rate/action_model': 1.0638086026912872e-05, 'epoch': 0.33}) +[2026-06-24 14:12:32,884][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.03830575942993164, 'train/grad_norm_pre_clip': 31.601526260375977, 'timing/data': 0.00034691207110881805, 'timing/model': 1.2791896001435816, 'learning_rate/qwen_vl_interface': 1.062001840675454e-06, 'learning_rate/action_model': 1.062001840675454e-05, 'epoch': 0.33}) +[2026-06-24 14:12:33,964][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.3133007287979126, 'train/grad_norm_pre_clip': 115.7175064086914, 'timing/data': 0.00033865589648485184, 'timing/model': 1.0781898410059512, 'learning_rate/qwen_vl_interface': 1.060197796206159e-06, 'learning_rate/action_model': 1.060197796206159e-05, 'epoch': 0.33}) +[2026-06-24 14:12:35,052][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.036824677139520645, 'train/grad_norm_pre_clip': 56.169715881347656, 'timing/data': 0.00031596398912370205, 'timing/model': 1.0863718069158494, 'learning_rate/qwen_vl_interface': 1.0583964704540247e-06, 'learning_rate/action_model': 1.0583964704540247e-05, 'epoch': 0.33}) +[2026-06-24 14:12:36,135][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.11793157458305359, 'train/grad_norm_pre_clip': 68.21847534179688, 'timing/data': 0.0003137090243399143, 'timing/model': 1.0816465360112488, 'learning_rate/qwen_vl_interface': 1.0565978645879138e-06, 'learning_rate/action_model': 1.0565978645879138e-05, 'epoch': 0.33}) +[2026-06-24 14:12:37,223][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.0036780540831387043, 'train/grad_norm_pre_clip': 4.229676723480225, 'timing/data': 0.00031046289950609207, 'timing/model': 1.0860757008194923, 'learning_rate/qwen_vl_interface': 1.054801979774921e-06, 'learning_rate/action_model': 1.054801979774921e-05, 'epoch': 0.33}) +[2026-06-24 14:12:38,309][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.04290593042969704, 'train/grad_norm_pre_clip': 33.64273452758789, 'timing/data': 0.00032297708094120026, 'timing/model': 1.0841183220036328, 'learning_rate/qwen_vl_interface': 1.0530088171803778e-06, 'learning_rate/action_model': 1.0530088171803779e-05, 'epoch': 0.33}) +[2026-06-24 14:12:39,396][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.08840984106063843, 'train/grad_norm_pre_clip': 44.140262603759766, 'timing/data': 0.00035768188536167145, 'timing/model': 1.0848248498514295, 'learning_rate/qwen_vl_interface': 1.0512183779678464e-06, 'learning_rate/action_model': 1.0512183779678464e-05, 'epoch': 0.33}) +[2026-06-24 14:12:40,483][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.16624930500984192, 'train/grad_norm_pre_clip': 89.14131927490234, 'timing/data': 0.00047440105117857456, 'timing/model': 1.0855404150206596, 'learning_rate/qwen_vl_interface': 1.0494306632991238e-06, 'learning_rate/action_model': 1.0494306632991239e-05, 'epoch': 0.33}) +[2026-06-24 14:12:41,568][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.051897287368774414, 'train/grad_norm_pre_clip': 31.15804672241211, 'timing/data': 0.00033616204746067524, 'timing/model': 1.083429436897859, 'learning_rate/qwen_vl_interface': 1.0476456743342399e-06, 'learning_rate/action_model': 1.0476456743342398e-05, 'epoch': 0.33}) +[2026-06-24 14:12:42,658][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.010092468932271004, 'train/grad_norm_pre_clip': 18.88382911682129, 'timing/data': 0.0003165749367326498, 'timing/model': 1.088023642078042, 'learning_rate/qwen_vl_interface': 1.0458634122314514e-06, 'learning_rate/action_model': 1.0458634122314515e-05, 'epoch': 0.33}) +[2026-06-24 14:12:43,745][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.00505022332072258, 'train/grad_norm_pre_clip': 2.881049394607544, 'timing/data': 0.0003470929805189371, 'timing/model': 1.0851662321947515, 'learning_rate/qwen_vl_interface': 1.0440838781472533e-06, 'learning_rate/action_model': 1.0440838781472533e-05, 'epoch': 0.33}) +[2026-06-24 14:12:44,829][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.003376503475010395, 'train/grad_norm_pre_clip': 2.9144561290740967, 'timing/data': 0.00038585509173572063, 'timing/model': 1.0825165330898017, 'learning_rate/qwen_vl_interface': 1.0423070732363634e-06, 'learning_rate/action_model': 1.0423070732363634e-05, 'epoch': 0.33}) +[2026-06-24 14:12:45,913][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.016579151153564453, 'train/grad_norm_pre_clip': 25.00054931640625, 'timing/data': 0.00032766605727374554, 'timing/model': 1.0829215941485018, 'learning_rate/qwen_vl_interface': 1.0405329986517301e-06, 'learning_rate/action_model': 1.0405329986517301e-05, 'epoch': 0.33}) +[2026-06-24 14:12:47,002][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.0653606504201889, 'train/grad_norm_pre_clip': 26.679540634155273, 'timing/data': 0.0003221349325031042, 'timing/model': 1.0872659089509398, 'learning_rate/qwen_vl_interface': 1.038761655544534e-06, 'learning_rate/action_model': 1.0387616555445339e-05, 'epoch': 0.33}) +[2026-06-24 14:12:48,090][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.03762941434979439, 'train/grad_norm_pre_clip': 60.531158447265625, 'timing/data': 0.0003314230125397444, 'timing/model': 1.0858706859871745, 'learning_rate/qwen_vl_interface': 1.0369930450641774e-06, 'learning_rate/action_model': 1.0369930450641775e-05, 'epoch': 0.33}) +[2026-06-24 14:12:49,177][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.12965479493141174, 'train/grad_norm_pre_clip': 73.64530181884766, 'timing/data': 0.0003185479436069727, 'timing/model': 1.086114495061338, 'learning_rate/qwen_vl_interface': 1.0352271683582949e-06, 'learning_rate/action_model': 1.0352271683582947e-05, 'epoch': 0.33}) +[2026-06-24 14:12:50,456][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.1029701828956604, 'train/grad_norm_pre_clip': 71.23897552490234, 'timing/data': 0.00033162301406264305, 'timing/model': 1.2765354851726443, 'learning_rate/qwen_vl_interface': 1.0334640265727432e-06, 'learning_rate/action_model': 1.0334640265727431e-05, 'epoch': 0.33}) +[2026-06-24 14:12:51,539][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.015536326915025711, 'train/grad_norm_pre_clip': 19.37276268005371, 'timing/data': 0.0003196210600435734, 'timing/model': 1.081988241057843, 'learning_rate/qwen_vl_interface': 1.0317036208516071e-06, 'learning_rate/action_model': 1.0317036208516072e-05, 'epoch': 0.33}) +[2026-06-24 14:12:52,628][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.12704741954803467, 'train/grad_norm_pre_clip': 56.793846130371094, 'timing/data': 0.00047154701314866543, 'timing/model': 1.0865432110149413, 'learning_rate/qwen_vl_interface': 1.029945952337193e-06, 'learning_rate/action_model': 1.0299459523371929e-05, 'epoch': 0.33}) +[2026-06-24 14:12:53,717][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.36724919080734253, 'train/grad_norm_pre_clip': 155.7406463623047, 'timing/data': 0.00033655203878879547, 'timing/model': 1.0879101189784706, 'learning_rate/qwen_vl_interface': 1.0281910221700345e-06, 'learning_rate/action_model': 1.0281910221700345e-05, 'epoch': 0.33}) +[2026-06-24 14:12:54,806][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.056629132479429245, 'train/grad_norm_pre_clip': 24.82937240600586, 'timing/data': 0.0003203118685632944, 'timing/model': 1.08723832690157, 'learning_rate/qwen_vl_interface': 1.026438831488887e-06, 'learning_rate/action_model': 1.026438831488887e-05, 'epoch': 0.33}) +[2026-06-24 14:12:55,896][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.017660584300756454, 'train/grad_norm_pre_clip': 23.456741333007812, 'timing/data': 0.00035273307003080845, 'timing/model': 1.0881161170545965, 'learning_rate/qwen_vl_interface': 1.0246893814307262e-06, 'learning_rate/action_model': 1.0246893814307261e-05, 'epoch': 0.33}) +[2026-06-24 14:12:56,982][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.14603260159492493, 'train/grad_norm_pre_clip': 67.55448913574219, 'timing/data': 0.00031927996315062046, 'timing/model': 1.0850347459781915, 'learning_rate/qwen_vl_interface': 1.0229426731307521e-06, 'learning_rate/action_model': 1.022942673130752e-05, 'epoch': 0.33}) +[2026-06-24 14:12:58,069][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.07430823892354965, 'train/grad_norm_pre_clip': 42.91176986694336, 'timing/data': 0.00034031993709504604, 'timing/model': 1.0855332328937948, 'learning_rate/qwen_vl_interface': 1.0211987077223863e-06, 'learning_rate/action_model': 1.0211987077223861e-05, 'epoch': 0.33}) +[2026-06-24 14:12:59,156][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.029319733381271362, 'train/grad_norm_pre_clip': 56.34431076049805, 'timing/data': 0.0004275029059499502, 'timing/model': 1.0848206938244402, 'learning_rate/qwen_vl_interface': 1.0194574863372662e-06, 'learning_rate/action_model': 1.0194574863372661e-05, 'epoch': 0.33}) +[2026-06-24 14:13:00,242][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.21168148517608643, 'train/grad_norm_pre_clip': 80.65821838378906, 'timing/data': 0.00031771697103977203, 'timing/model': 1.0843275769148022, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 0.33}) +[2026-06-24 14:13:01,328][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.04351162537932396, 'train/grad_norm_pre_clip': 56.56409454345703, 'timing/data': 0.00035998690873384476, 'timing/model': 1.084925752133131, 'learning_rate/qwen_vl_interface': 1.015983280154425e-06, 'learning_rate/action_model': 1.0159832801544249e-05, 'epoch': 0.33}) +[2026-06-24 14:13:02,416][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.009297232143580914, 'train/grad_norm_pre_clip': 9.619993209838867, 'timing/data': 0.00033183395862579346, 'timing/model': 1.0861218199133873, 'learning_rate/qwen_vl_interface': 1.014250297611079e-06, 'learning_rate/action_model': 1.014250297611079e-05, 'epoch': 0.33}) +[2026-06-24 14:13:03,508][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.017665263265371323, 'train/grad_norm_pre_clip': 10.66455364227295, 'timing/data': 0.0003090298268944025, 'timing/model': 1.0906255228910595, 'learning_rate/qwen_vl_interface': 1.0125200635997274e-06, 'learning_rate/action_model': 1.0125200635997273e-05, 'epoch': 0.33}) +[2026-06-24 14:13:04,592][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.19627171754837036, 'train/grad_norm_pre_clip': 38.800376892089844, 'timing/data': 0.00030873017385602, 'timing/model': 1.0819564019329846, 'learning_rate/qwen_vl_interface': 1.0107925792431008e-06, 'learning_rate/action_model': 1.0107925792431008e-05, 'epoch': 0.33}) +[2026-06-24 14:13:05,682][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.164442777633667, 'train/grad_norm_pre_clip': 80.03899383544922, 'timing/data': 0.00033461907878518105, 'timing/model': 1.0886023819912225, 'learning_rate/qwen_vl_interface': 1.0090678456621461e-06, 'learning_rate/action_model': 1.009067845662146e-05, 'epoch': 0.33}) +[2026-06-24 14:13:06,771][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.08374611288309097, 'train/grad_norm_pre_clip': 35.04469680786133, 'timing/data': 0.00032932916656136513, 'timing/model': 1.087617431068793, 'learning_rate/qwen_vl_interface': 1.0073458639760218e-06, 'learning_rate/action_model': 1.0073458639760217e-05, 'epoch': 0.33}) +[2026-06-24 14:13:08,112][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.1909644454717636, 'train/grad_norm_pre_clip': 72.13705444335938, 'timing/data': 0.00029793987050652504, 'timing/model': 1.339142284123227, 'learning_rate/qwen_vl_interface': 1.0056266353021058e-06, 'learning_rate/action_model': 1.0056266353021056e-05, 'epoch': 0.33}) +[2026-06-24 14:13:09,199][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.20800770819187164, 'train/grad_norm_pre_clip': 54.92109298706055, 'timing/data': 0.0003264741972088814, 'timing/model': 1.0852921209298074, 'learning_rate/qwen_vl_interface': 1.0039101607559858e-06, 'learning_rate/action_model': 1.0039101607559858e-05, 'epoch': 0.33}) +[2026-06-24 14:13:10,285][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.22615037858486176, 'train/grad_norm_pre_clip': 104.47682189941406, 'timing/data': 0.00038553494960069656, 'timing/model': 1.0845724779646844, 'learning_rate/qwen_vl_interface': 1.0021964414514616e-06, 'learning_rate/action_model': 1.0021964414514616e-05, 'epoch': 0.33}) +[2026-06-24 14:13:11,378][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.02757577784359455, 'train/grad_norm_pre_clip': 28.12345314025879, 'timing/data': 0.0003407709300518036, 'timing/model': 1.091725962003693, 'learning_rate/qwen_vl_interface': 1.0004854785005506e-06, 'learning_rate/action_model': 1.0004854785005504e-05, 'epoch': 0.33}) +[2026-06-24 14:13:12,467][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.06015472114086151, 'train/grad_norm_pre_clip': 27.675506591796875, 'timing/data': 0.0003588248509913683, 'timing/model': 1.0866760618519038, 'learning_rate/qwen_vl_interface': 9.987772730134753e-07, 'learning_rate/action_model': 9.987772730134753e-06, 'epoch': 0.33}) +[2026-06-24 14:13:13,551][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.12715579569339752, 'train/grad_norm_pre_clip': 60.446075439453125, 'timing/data': 0.0003286069259047508, 'timing/model': 1.0821979569736868, 'learning_rate/qwen_vl_interface': 9.970718260986728e-07, 'learning_rate/action_model': 9.970718260986728e-06, 'epoch': 0.33}) +[2026-06-24 14:13:14,638][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.03875922039151192, 'train/grad_norm_pre_clip': 14.859822273254395, 'timing/data': 0.000385083956643939, 'timing/model': 1.08602839615196, 'learning_rate/qwen_vl_interface': 9.953691388627887e-07, 'learning_rate/action_model': 9.953691388627886e-06, 'epoch': 0.33}) +[2026-06-24 14:13:15,726][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.0014965030131861567, 'train/grad_norm_pre_clip': 1.1536531448364258, 'timing/data': 0.000559310894459486, 'timing/model': 1.0855168239213526, 'learning_rate/qwen_vl_interface': 9.936692124106795e-07, 'learning_rate/action_model': 9.936692124106794e-06, 'epoch': 0.33}) +[2026-06-24 14:13:16,813][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.01555450540035963, 'train/grad_norm_pre_clip': 13.741230010986328, 'timing/data': 0.000366969034075737, 'timing/model': 1.0856698208954185, 'learning_rate/qwen_vl_interface': 9.91972047845407e-07, 'learning_rate/action_model': 9.919720478454069e-06, 'epoch': 0.33}) +[2026-06-24 14:13:17,899][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.02376493625342846, 'train/grad_norm_pre_clip': 46.04617691040039, 'timing/data': 0.000379153061658144, 'timing/model': 1.0839079699944705, 'learning_rate/qwen_vl_interface': 9.902776462682446e-07, 'learning_rate/action_model': 9.902776462682444e-06, 'epoch': 0.33}) +[2026-06-24 14:13:18,988][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.00034952411078847945, 'train/grad_norm_pre_clip': 0.38276830315589905, 'timing/data': 0.0003351089544594288, 'timing/model': 1.0876026838086545, 'learning_rate/qwen_vl_interface': 9.885860087786706e-07, 'learning_rate/action_model': 9.885860087786705e-06, 'epoch': 0.33}) +[2026-06-24 14:13:20,075][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.0908721387386322, 'train/grad_norm_pre_clip': 45.239906311035156, 'timing/data': 0.00033155293203890324, 'timing/model': 1.0854869380127639, 'learning_rate/qwen_vl_interface': 9.868971364743694e-07, 'learning_rate/action_model': 9.868971364743692e-06, 'epoch': 0.33}) +[2026-06-24 14:13:21,162][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.05379188433289528, 'train/grad_norm_pre_clip': 36.01528549194336, 'timing/data': 0.00032360805198550224, 'timing/model': 1.085328099085018, 'learning_rate/qwen_vl_interface': 9.852110304512326e-07, 'learning_rate/action_model': 9.852110304512325e-06, 'epoch': 0.33}) +[2026-06-24 14:13:22,244][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.001983514754101634, 'train/grad_norm_pre_clip': 2.1494555473327637, 'timing/data': 0.0003092808183282614, 'timing/model': 1.079912091838196, 'learning_rate/qwen_vl_interface': 9.835276918033565e-07, 'learning_rate/action_model': 9.835276918033565e-06, 'epoch': 0.33}) +[2026-06-24 14:13:23,329][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.043630097061395645, 'train/grad_norm_pre_clip': 74.14739990234375, 'timing/data': 0.0003257018979638815, 'timing/model': 1.0833065211772919, 'learning_rate/qwen_vl_interface': 9.818471216230402e-07, 'learning_rate/action_model': 9.818471216230401e-06, 'epoch': 0.33}) +[2026-06-24 14:13:24,627][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.02401987835764885, 'train/grad_norm_pre_clip': 18.246387481689453, 'timing/data': 0.0003743530251085758, 'timing/model': 1.2967712772078812, 'learning_rate/qwen_vl_interface': 9.801693210007877e-07, 'learning_rate/action_model': 9.801693210007876e-06, 'epoch': 0.33}) +[2026-06-24 14:13:25,714][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.006481842137873173, 'train/grad_norm_pre_clip': 3.868114709854126, 'timing/data': 0.00034973700530827045, 'timing/model': 1.0848112788517028, 'learning_rate/qwen_vl_interface': 9.784942910253061e-07, 'learning_rate/action_model': 9.78494291025306e-06, 'epoch': 0.33}) +[2026-06-24 14:13:26,799][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.01959836296737194, 'train/grad_norm_pre_clip': 16.57976722717285, 'timing/data': 0.0003173870500177145, 'timing/model': 1.0832664060872048, 'learning_rate/qwen_vl_interface': 9.76822032783505e-07, 'learning_rate/action_model': 9.768220327835051e-06, 'epoch': 0.33}) +[2026-06-24 14:13:27,889][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.23887492716312408, 'train/grad_norm_pre_clip': 117.6280288696289, 'timing/data': 0.00032760598696768284, 'timing/model': 1.0881251669488847, 'learning_rate/qwen_vl_interface': 9.75152547360493e-07, 'learning_rate/action_model': 9.751525473604928e-06, 'epoch': 0.33}) +[2026-06-24 14:13:28,975][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.014373877085745335, 'train/grad_norm_pre_clip': 7.92965841293335, 'timing/data': 0.00033801584504544735, 'timing/model': 1.0844260049052536, 'learning_rate/qwen_vl_interface': 9.734858358395808e-07, 'learning_rate/action_model': 9.734858358395807e-06, 'epoch': 0.33}) +[2026-06-24 14:13:30,066][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.020087920129299164, 'train/grad_norm_pre_clip': 23.460702896118164, 'timing/data': 0.00033668195828795433, 'timing/model': 1.0893713489640504, 'learning_rate/qwen_vl_interface': 9.71821899302282e-07, 'learning_rate/action_model': 9.71821899302282e-06, 'epoch': 0.33}) +[2026-06-24 14:13:31,154][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.21937257051467896, 'train/grad_norm_pre_clip': 68.37405395507812, 'timing/data': 0.000337463803589344, 'timing/model': 1.0869806858245283, 'learning_rate/qwen_vl_interface': 9.701607388283026e-07, 'learning_rate/action_model': 9.701607388283026e-06, 'epoch': 0.33}) +[2026-06-24 14:13:32,242][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.19842073321342468, 'train/grad_norm_pre_clip': 92.07786560058594, 'timing/data': 0.0003332970663905144, 'timing/model': 1.0857074030209333, 'learning_rate/qwen_vl_interface': 9.68502355495557e-07, 'learning_rate/action_model': 9.68502355495557e-06, 'epoch': 0.33}) +[2026-06-24 14:13:33,331][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.1749657690525055, 'train/grad_norm_pre_clip': 118.13533020019531, 'timing/data': 0.0003539652097970247, 'timing/model': 1.0879855249077082, 'learning_rate/qwen_vl_interface': 9.668467503801486e-07, 'learning_rate/action_model': 9.668467503801484e-06, 'epoch': 0.33}) +[2026-06-24 14:13:34,419][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.1058151125907898, 'train/grad_norm_pre_clip': 36.410179138183594, 'timing/data': 0.0003245000261813402, 'timing/model': 1.0861731090117246, 'learning_rate/qwen_vl_interface': 9.65193924556381e-07, 'learning_rate/action_model': 9.65193924556381e-06, 'epoch': 0.33}) +[2026-06-24 14:13:35,510][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.010318969376385212, 'train/grad_norm_pre_clip': 13.508792877197266, 'timing/data': 0.00033904705196619034, 'timing/model': 1.0894524219911546, 'learning_rate/qwen_vl_interface': 9.635438790967576e-07, 'learning_rate/action_model': 9.635438790967575e-06, 'epoch': 0.33}) +[2026-06-24 14:13:36,599][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.22492381930351257, 'train/grad_norm_pre_clip': 81.26803588867188, 'timing/data': 0.00032622297294437885, 'timing/model': 1.087761664064601, 'learning_rate/qwen_vl_interface': 9.61896615071973e-07, 'learning_rate/action_model': 9.61896615071973e-06, 'epoch': 0.33}) +[2026-06-24 14:13:37,684][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.07110283523797989, 'train/grad_norm_pre_clip': 66.52498626708984, 'timing/data': 0.00032267719507217407, 'timing/model': 1.082730308873579, 'learning_rate/qwen_vl_interface': 9.6025213355092e-07, 'learning_rate/action_model': 9.6025213355092e-06, 'epoch': 0.33}) +[2026-06-24 14:13:38,775][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.014973831363022327, 'train/grad_norm_pre_clip': 7.989709854125977, 'timing/data': 0.00032391794957220554, 'timing/model': 1.0892444008495659, 'learning_rate/qwen_vl_interface': 9.586104356006851e-07, 'learning_rate/action_model': 9.58610435600685e-06, 'epoch': 0.33}) +[2026-06-24 14:13:39,861][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.27431735396385193, 'train/grad_norm_pre_clip': 94.7433090209961, 'timing/data': 0.00032683415338397026, 'timing/model': 1.084541382966563, 'learning_rate/qwen_vl_interface': 9.569715222865463e-07, 'learning_rate/action_model': 9.569715222865463e-06, 'epoch': 0.33}) +[2026-06-24 14:13:40,949][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.05962582677602768, 'train/grad_norm_pre_clip': 18.860490798950195, 'timing/data': 0.00034680101089179516, 'timing/model': 1.086567509919405, 'learning_rate/qwen_vl_interface': 9.553353946719784e-07, 'learning_rate/action_model': 9.553353946719783e-06, 'epoch': 0.33}) +[2026-06-24 14:13:42,329][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.003055915702134371, 'train/grad_norm_pre_clip': 4.257670879364014, 'timing/data': 0.00033995904959738255, 'timing/model': 1.3791169228497893, 'learning_rate/qwen_vl_interface': 9.537020538186461e-07, 'learning_rate/action_model': 9.537020538186459e-06, 'epoch': 0.33}) +[2026-06-24 14:13:43,416][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.08280491828918457, 'train/grad_norm_pre_clip': 44.06932830810547, 'timing/data': 0.00034494814462959766, 'timing/model': 1.0852166020777076, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 0.33}) +[2026-06-24 14:13:44,500][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.045701541006565094, 'train/grad_norm_pre_clip': 37.50703048706055, 'timing/data': 0.0003340281546115875, 'timing/model': 1.081657133065164, 'learning_rate/qwen_vl_interface': 9.504437366333097e-07, 'learning_rate/action_model': 9.504437366333096e-06, 'epoch': 0.33}) +[2026-06-24 14:13:45,587][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.0980016440153122, 'train/grad_norm_pre_clip': 39.48039627075195, 'timing/data': 0.0003207030240446329, 'timing/model': 1.085788066033274, 'learning_rate/qwen_vl_interface': 9.488187624155914e-07, 'learning_rate/action_model': 9.488187624155913e-06, 'epoch': 0.33}) +[2026-06-24 14:13:46,675][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.06000853329896927, 'train/grad_norm_pre_clip': 34.55906677246094, 'timing/data': 0.00035676080733537674, 'timing/model': 1.0858109800610691, 'learning_rate/qwen_vl_interface': 9.471965791876822e-07, 'learning_rate/action_model': 9.471965791876822e-06, 'epoch': 0.33}) +[2026-06-24 14:13:47,765][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.03013036958873272, 'train/grad_norm_pre_clip': 50.04811096191406, 'timing/data': 0.00037541589699685574, 'timing/model': 1.0882486009504646, 'learning_rate/qwen_vl_interface': 9.455771880021956e-07, 'learning_rate/action_model': 9.455771880021954e-06, 'epoch': 0.33}) +[2026-06-24 14:13:48,852][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.16863079369068146, 'train/grad_norm_pre_clip': 45.904815673828125, 'timing/data': 0.0003402489237487316, 'timing/model': 1.0856234568636864, 'learning_rate/qwen_vl_interface': 9.439605899099412e-07, 'learning_rate/action_model': 9.439605899099411e-06, 'epoch': 0.33}) +[2026-06-24 14:13:49,942][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.0378667488694191, 'train/grad_norm_pre_clip': 21.94860076904297, 'timing/data': 0.0003553880378603935, 'timing/model': 1.0880308519117534, 'learning_rate/qwen_vl_interface': 9.423467859599094e-07, 'learning_rate/action_model': 9.423467859599093e-06, 'epoch': 0.33}) +[2026-06-24 14:13:51,027][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.31313925981521606, 'train/grad_norm_pre_clip': 64.88211059570312, 'timing/data': 0.00034936703741550446, 'timing/model': 1.084011638071388, 'learning_rate/qwen_vl_interface': 9.407357771992802e-07, 'learning_rate/action_model': 9.407357771992802e-06, 'epoch': 0.33}) +[2026-06-24 14:13:52,114][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.035296324640512466, 'train/grad_norm_pre_clip': 26.02600860595703, 'timing/data': 0.0003227770794183016, 'timing/model': 1.0855889930389822, 'learning_rate/qwen_vl_interface': 9.391275646734213e-07, 'learning_rate/action_model': 9.391275646734213e-06, 'epoch': 0.33}) +[2026-06-24 14:13:53,204][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.17617014050483704, 'train/grad_norm_pre_clip': 179.4581298828125, 'timing/data': 0.00032142503187060356, 'timing/model': 1.0879380870610476, 'learning_rate/qwen_vl_interface': 9.375221494258821e-07, 'learning_rate/action_model': 9.375221494258821e-06, 'epoch': 0.34}) +[2026-06-24 14:13:54,292][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.06141664460301399, 'train/grad_norm_pre_clip': 62.51377487182617, 'timing/data': 0.00032365787774324417, 'timing/model': 1.0866521710995585, 'learning_rate/qwen_vl_interface': 9.359195324984002e-07, 'learning_rate/action_model': 9.359195324984002e-06, 'epoch': 0.34}) +[2026-06-24 14:13:55,381][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.034223198890686035, 'train/grad_norm_pre_clip': 81.12960052490234, 'timing/data': 0.0003356409724801779, 'timing/model': 1.08726446214132, 'learning_rate/qwen_vl_interface': 9.343197149308967e-07, 'learning_rate/action_model': 9.343197149308967e-06, 'epoch': 0.34}) +[2026-06-24 14:13:56,468][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.10311893373727798, 'train/grad_norm_pre_clip': 47.756935119628906, 'timing/data': 0.00032706395722925663, 'timing/model': 1.0854817209765315, 'learning_rate/qwen_vl_interface': 9.327226977614766e-07, 'learning_rate/action_model': 9.327226977614766e-06, 'epoch': 0.34}) +[2026-06-24 14:13:57,555][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.00919102318584919, 'train/grad_norm_pre_clip': 6.4879984855651855, 'timing/data': 0.00036277202889323235, 'timing/model': 1.0847406978718936, 'learning_rate/qwen_vl_interface': 9.311284820264245e-07, 'learning_rate/action_model': 9.311284820264244e-06, 'epoch': 0.34}) +[2026-06-24 14:13:58,809][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.020200546830892563, 'train/grad_norm_pre_clip': 24.173742294311523, 'timing/data': 0.0003429139032959938, 'timing/model': 1.2531341242138296, 'learning_rate/qwen_vl_interface': 9.295370687602114e-07, 'learning_rate/action_model': 9.295370687602113e-06, 'epoch': 0.34}) +[2026-06-24 14:13:59,899][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.09069061279296875, 'train/grad_norm_pre_clip': 64.8887939453125, 'timing/data': 0.0003151628188788891, 'timing/model': 1.0884237908758223, 'learning_rate/qwen_vl_interface': 9.279484589954886e-07, 'learning_rate/action_model': 9.279484589954886e-06, 'epoch': 0.34}) +[2026-06-24 14:14:00,988][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.08831018954515457, 'train/grad_norm_pre_clip': 71.8379135131836, 'timing/data': 0.0003164049703627825, 'timing/model': 1.086946734925732, 'learning_rate/qwen_vl_interface': 9.263626537630857e-07, 'learning_rate/action_model': 9.263626537630856e-06, 'epoch': 0.34}) +[2026-06-24 14:14:02,075][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.02206548862159252, 'train/grad_norm_pre_clip': 13.403254508972168, 'timing/data': 0.00035968609154224396, 'timing/model': 1.0850931820459664, 'learning_rate/qwen_vl_interface': 9.247796540920162e-07, 'learning_rate/action_model': 9.247796540920162e-06, 'epoch': 0.34}) +[2026-06-24 14:14:03,165][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.5488842725753784, 'train/grad_norm_pre_clip': 65.62638092041016, 'timing/data': 0.000319531187415123, 'timing/model': 1.0889012180268764, 'learning_rate/qwen_vl_interface': 9.231994610094716e-07, 'learning_rate/action_model': 9.231994610094716e-06, 'epoch': 0.34}) +[2026-06-24 14:14:04,253][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.02080608531832695, 'train/grad_norm_pre_clip': 22.746109008789062, 'timing/data': 0.0003143709618598223, 'timing/model': 1.086485569132492, 'learning_rate/qwen_vl_interface': 9.216220755408203e-07, 'learning_rate/action_model': 9.216220755408204e-06, 'epoch': 0.34}) +[2026-06-24 14:14:05,344][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.1066448763012886, 'train/grad_norm_pre_clip': 144.84156799316406, 'timing/data': 0.0003173770383000374, 'timing/model': 1.0895669700112194, 'learning_rate/qwen_vl_interface': 9.200474987096121e-07, 'learning_rate/action_model': 9.200474987096122e-06, 'epoch': 0.34}) +[2026-06-24 14:14:06,431][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.011339406482875347, 'train/grad_norm_pre_clip': 13.503485679626465, 'timing/data': 0.0003176468890160322, 'timing/model': 1.0850113579072058, 'learning_rate/qwen_vl_interface': 9.184757315375722e-07, 'learning_rate/action_model': 9.184757315375723e-06, 'epoch': 0.34}) +[2026-06-24 14:14:07,522][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.013918115757405758, 'train/grad_norm_pre_clip': 48.96033477783203, 'timing/data': 0.00034915609285235405, 'timing/model': 1.08997459593229, 'learning_rate/qwen_vl_interface': 9.16906775044604e-07, 'learning_rate/action_model': 9.16906775044604e-06, 'epoch': 0.34}) +[2026-06-24 14:14:08,615][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.024487188085913658, 'train/grad_norm_pre_clip': 25.93312644958496, 'timing/data': 0.00033988896757364273, 'timing/model': 1.0904761590063572, 'learning_rate/qwen_vl_interface': 9.153406302487841e-07, 'learning_rate/action_model': 9.153406302487842e-06, 'epoch': 0.34}) +[2026-06-24 14:14:09,703][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.07141953706741333, 'train/grad_norm_pre_clip': 72.50930786132812, 'timing/data': 0.0003332660999149084, 'timing/model': 1.0867760151159018, 'learning_rate/qwen_vl_interface': 9.137772981663706e-07, 'learning_rate/action_model': 9.137772981663706e-06, 'epoch': 0.34}) +[2026-06-24 14:14:10,792][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.3386955261230469, 'train/grad_norm_pre_clip': 79.62437438964844, 'timing/data': 0.00035806302912533283, 'timing/model': 1.0873078349977732, 'learning_rate/qwen_vl_interface': 9.122167798117904e-07, 'learning_rate/action_model': 9.122167798117903e-06, 'epoch': 0.34}) +[2026-06-24 14:14:11,878][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.0234268456697464, 'train/grad_norm_pre_clip': 23.738157272338867, 'timing/data': 0.00037892209365963936, 'timing/model': 1.08472734503448, 'learning_rate/qwen_vl_interface': 9.106590761976454e-07, 'learning_rate/action_model': 9.106590761976455e-06, 'epoch': 0.34}) +[2026-06-24 14:14:12,973][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.010549320839345455, 'train/grad_norm_pre_clip': 7.711989879608154, 'timing/data': 0.00035738199949264526, 'timing/model': 1.0936672461684793, 'learning_rate/qwen_vl_interface': 9.091041883347162e-07, 'learning_rate/action_model': 9.091041883347163e-06, 'epoch': 0.34}) +[2026-06-24 14:14:14,066][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.0028061741031706333, 'train/grad_norm_pre_clip': 2.473982334136963, 'timing/data': 0.0003503479529172182, 'timing/model': 1.0904713901691139, 'learning_rate/qwen_vl_interface': 9.075521172319495e-07, 'learning_rate/action_model': 9.075521172319496e-06, 'epoch': 0.34}) +[2026-06-24 14:14:15,359][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.02939547784626484, 'train/grad_norm_pre_clip': 35.399925231933594, 'timing/data': 0.0003373140934854746, 'timing/model': 1.291431677993387, 'learning_rate/qwen_vl_interface': 9.06002863896469e-07, 'learning_rate/action_model': 9.060028638964689e-06, 'epoch': 0.34}) +[2026-06-24 14:14:16,447][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.013776729814708233, 'train/grad_norm_pre_clip': 14.428750991821289, 'timing/data': 0.0003790420014411211, 'timing/model': 1.0863376820925623, 'learning_rate/qwen_vl_interface': 9.044564293335687e-07, 'learning_rate/action_model': 9.044564293335687e-06, 'epoch': 0.34}) +[2026-06-24 14:14:17,536][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.08002157509326935, 'train/grad_norm_pre_clip': 60.1281852722168, 'timing/data': 0.00032804603688418865, 'timing/model': 1.0877690920606256, 'learning_rate/qwen_vl_interface': 9.029128145467116e-07, 'learning_rate/action_model': 9.029128145467115e-06, 'epoch': 0.34}) +[2026-06-24 14:14:18,625][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.06006760522723198, 'train/grad_norm_pre_clip': 52.23832702636719, 'timing/data': 0.0003257519565522671, 'timing/model': 1.0881032100878656, 'learning_rate/qwen_vl_interface': 9.013720205375337e-07, 'learning_rate/action_model': 9.013720205375337e-06, 'epoch': 0.34}) +[2026-06-24 14:14:19,714][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.1414923518896103, 'train/grad_norm_pre_clip': 68.40139770507812, 'timing/data': 0.00034374604001641273, 'timing/model': 1.0873621669597924, 'learning_rate/qwen_vl_interface': 8.998340483058395e-07, 'learning_rate/action_model': 8.998340483058395e-06, 'epoch': 0.34}) +[2026-06-24 14:14:20,806][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.0029646551702171564, 'train/grad_norm_pre_clip': 2.486823320388794, 'timing/data': 0.0003314430359750986, 'timing/model': 1.0900414921343327, 'learning_rate/qwen_vl_interface': 8.982988988496027e-07, 'learning_rate/action_model': 8.982988988496027e-06, 'epoch': 0.34}) +[2026-06-24 14:14:21,897][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.020078523084521294, 'train/grad_norm_pre_clip': 16.2225341796875, 'timing/data': 0.00033183395862579346, 'timing/model': 1.089218445820734, 'learning_rate/qwen_vl_interface': 8.967665731649638e-07, 'learning_rate/action_model': 8.967665731649638e-06, 'epoch': 0.34}) +[2026-06-24 14:14:22,990][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.014976691454648972, 'train/grad_norm_pre_clip': 8.477849960327148, 'timing/data': 0.00033656181767582893, 'timing/model': 1.091946001863107, 'learning_rate/qwen_vl_interface': 8.952370722462337e-07, 'learning_rate/action_model': 8.952370722462336e-06, 'epoch': 0.34}) +[2026-06-24 14:14:24,079][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.2583860158920288, 'train/grad_norm_pre_clip': 73.61969757080078, 'timing/data': 0.0007095439359545708, 'timing/model': 1.08647741493769, 'learning_rate/qwen_vl_interface': 8.937103970858888e-07, 'learning_rate/action_model': 8.937103970858887e-06, 'epoch': 0.34}) +[2026-06-24 14:14:25,170][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.3310241997241974, 'train/grad_norm_pre_clip': 129.8876953125, 'timing/data': 0.00032018194906413555, 'timing/model': 1.089415627066046, 'learning_rate/qwen_vl_interface': 8.921865486745709e-07, 'learning_rate/action_model': 8.921865486745709e-06, 'epoch': 0.34}) +[2026-06-24 14:14:26,264][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.12254641950130463, 'train/grad_norm_pre_clip': 50.80685043334961, 'timing/data': 0.0003126871306449175, 'timing/model': 1.0920722191222012, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 0.34}) +[2026-06-24 14:14:27,352][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.024271667003631592, 'train/grad_norm_pre_clip': 40.45555114746094, 'timing/data': 0.0003314830828458071, 'timing/model': 1.0867426239419729, 'learning_rate/qwen_vl_interface': 8.891473360524191e-07, 'learning_rate/action_model': 8.89147336052419e-06, 'epoch': 0.34}) +[2026-06-24 14:14:28,441][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.03919117525219917, 'train/grad_norm_pre_clip': 33.14350891113281, 'timing/data': 0.00033441907726228237, 'timing/model': 1.087508121971041, 'learning_rate/qwen_vl_interface': 8.87631973813697e-07, 'learning_rate/action_model': 8.87631973813697e-06, 'epoch': 0.34}) +[2026-06-24 14:14:29,531][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.20762719213962555, 'train/grad_norm_pre_clip': 53.203800201416016, 'timing/data': 0.00031735701486468315, 'timing/model': 1.0883079760242254, 'learning_rate/qwen_vl_interface': 8.861194422682247e-07, 'learning_rate/action_model': 8.861194422682247e-06, 'epoch': 0.34}) +[2026-06-24 14:14:30,620][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.009595392271876335, 'train/grad_norm_pre_clip': 23.833168029785156, 'timing/data': 0.0003208429552614689, 'timing/model': 1.0876728619914502, 'learning_rate/qwen_vl_interface': 8.846097423974691e-07, 'learning_rate/action_model': 8.84609742397469e-06, 'epoch': 0.34}) +[2026-06-24 14:14:31,917][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.016423111781477928, 'train/grad_norm_pre_clip': 11.073841094970703, 'timing/data': 0.0003527030348777771, 'timing/model': 1.294898474123329, 'learning_rate/qwen_vl_interface': 8.831028751810574e-07, 'learning_rate/action_model': 8.831028751810575e-06, 'epoch': 0.34}) +[2026-06-24 14:14:33,005][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.18851831555366516, 'train/grad_norm_pre_clip': 78.56107330322266, 'timing/data': 0.0003513710107654333, 'timing/model': 1.0868579810485244, 'learning_rate/qwen_vl_interface': 8.815988415967793e-07, 'learning_rate/action_model': 8.815988415967793e-06, 'epoch': 0.34}) +[2026-06-24 14:14:34,099][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3500 +[2026-06-24 14:14:59,312][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3500 +[2026-06-24 14:14:59,312][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3500 +[2026-06-24 14:15:32,466][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3500 +[2026-06-24 14:23:34,012][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3500 +[2026-06-24 14:23:34,012][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.11549915373325348, 'train/grad_norm_pre_clip': 59.44438552856445, 'eval/action_loss/seconds': 25.210602941922843, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.06450928002595901, 'eval/latency_2/loss': 0.06450928002595901, 'eval/flappy/loss': 0.06450928002595901, 'eval/flappy/latency_2/loss': 0.06450928002595901, 'eval/flap_precision': 0.9341935483870968, 'eval/flap_recall': 0.9823609226594301, 'eval/flap_f1': 0.9576719576719577, 'eval/flap_rate_model': 0.2152777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9576719576719578, 'eval/latency_2/flap_precision': 0.9341935483870968, 'eval/latency_2/flap_recall': 0.9823609226594301, 'eval/latency_2/flap_f1': 0.9576719576719577, 'eval/latency_2/flap_rate_model': 0.2152777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9576719576719578, 'eval/action_classification/data_wait_seconds': 0.3245052993297577, 'eval/action_classification/predict_seconds': 32.728599548339844, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 365.76000000002057, 'rl_games_eval/mid_train/mean_length': 2965.0, 'rl_games_eval/mid_train/std_reward': 157.6800000000101, 'rl_games_eval/mid_train/std_length': 1270.0, 'rl_games_eval/mid_train/macro_mean_reward': 365.76000000002057, 'rl_games_eval/mid_train/macro_mean_length': 2965.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 365.76000000002057, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2965.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 157.6800000000101, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1270.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 365.76000000002057, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00034913583658635616, 'timing/model': 1.092394245089963, 'learning_rate/qwen_vl_interface': 8.80097642620585e-07, 'learning_rate/action_model': 8.80097642620585e-06, 'epoch': 0.34}) +[2026-06-24 14:23:34,013][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_3500_state +[2026-06-24 14:23:34,015][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 14:23:36,763][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_3500_state/model.safetensors +[2026-06-24 14:23:43,038][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_3500_state/optimizer.bin +[2026-06-24 14:23:43,038][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_3500_state/sampler.bin +[2026-06-24 14:23:43,038][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-24 14:23:43,039][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-24 14:23:43,094][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-24 14:23:43,094][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 14:23:43,104][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 14:23:44,179][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.013116227462887764, 'train/grad_norm_pre_clip': 20.237199783325195, 'timing/data': 0.000463742995634675, 'timing/model': 1.0733288121409714, 'learning_rate/qwen_vl_interface': 8.785992792265885e-07, 'learning_rate/action_model': 8.785992792265884e-06, 'epoch': 0.34}) +[2026-06-24 14:23:45,244][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.01190380658954382, 'train/grad_norm_pre_clip': 6.730147838592529, 'timing/data': 0.0003759979736059904, 'timing/model': 1.0633537089452147, 'learning_rate/qwen_vl_interface': 8.771037523870579e-07, 'learning_rate/action_model': 8.771037523870578e-06, 'epoch': 0.34}) +[2026-06-24 14:23:46,317][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.08907200396060944, 'train/grad_norm_pre_clip': 47.115840911865234, 'timing/data': 0.0003430750221014023, 'timing/model': 1.070722945034504, 'learning_rate/qwen_vl_interface': 8.756110630724289e-07, 'learning_rate/action_model': 8.756110630724288e-06, 'epoch': 0.34}) +[2026-06-24 14:23:47,398][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.03092968836426735, 'train/grad_norm_pre_clip': 34.929073333740234, 'timing/data': 0.00048117502592504025, 'timing/model': 1.0790369589813054, 'learning_rate/qwen_vl_interface': 8.741212122512874e-07, 'learning_rate/action_model': 8.741212122512873e-06, 'epoch': 0.34}) +[2026-06-24 14:23:48,476][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.08727477490901947, 'train/grad_norm_pre_clip': 71.41287994384766, 'timing/data': 0.0005454570055007935, 'timing/model': 1.076978859025985, 'learning_rate/qwen_vl_interface': 8.726342008903833e-07, 'learning_rate/action_model': 8.726342008903834e-06, 'epoch': 0.34}) +[2026-06-24 14:23:49,543][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.2142540067434311, 'train/grad_norm_pre_clip': 116.12492370605469, 'timing/data': 0.0003324560821056366, 'timing/model': 1.0647746010217816, 'learning_rate/qwen_vl_interface': 8.71150029954623e-07, 'learning_rate/action_model': 8.71150029954623e-06, 'epoch': 0.34}) +[2026-06-24 14:23:50,610][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.0012165347579866648, 'train/grad_norm_pre_clip': 0.8791835904121399, 'timing/data': 0.000331162940710783, 'timing/model': 1.065843357006088, 'learning_rate/qwen_vl_interface': 8.696687004070666e-07, 'learning_rate/action_model': 8.696687004070666e-06, 'epoch': 0.34}) +[2026-06-24 14:23:51,677][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.13973368704319, 'train/grad_norm_pre_clip': 99.39532470703125, 'timing/data': 0.00038078706711530685, 'timing/model': 1.0649124190676957, 'learning_rate/qwen_vl_interface': 8.68190213208934e-07, 'learning_rate/action_model': 8.68190213208934e-06, 'epoch': 0.34}) +[2026-06-24 14:23:52,746][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.24046482145786285, 'train/grad_norm_pre_clip': 109.9278564453125, 'timing/data': 0.0003075490240007639, 'timing/model': 1.0669206629972905, 'learning_rate/qwen_vl_interface': 8.667145693195997e-07, 'learning_rate/action_model': 8.667145693195996e-06, 'epoch': 0.34}) +[2026-06-24 14:23:53,814][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.04626510664820671, 'train/grad_norm_pre_clip': 16.138093948364258, 'timing/data': 0.0003146019298583269, 'timing/model': 1.066778850974515, 'learning_rate/qwen_vl_interface': 8.652417696965934e-07, 'learning_rate/action_model': 8.652417696965935e-06, 'epoch': 0.34}) +[2026-06-24 14:23:54,887][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.33223867416381836, 'train/grad_norm_pre_clip': 67.58567810058594, 'timing/data': 0.0003435059916228056, 'timing/model': 1.071859718998894, 'learning_rate/qwen_vl_interface': 8.637718152955967e-07, 'learning_rate/action_model': 8.637718152955967e-06, 'epoch': 0.34}) +[2026-06-24 14:23:55,954][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.0025680141989141703, 'train/grad_norm_pre_clip': 2.1625497341156006, 'timing/data': 0.0003430759534239769, 'timing/model': 1.0646464051678777, 'learning_rate/qwen_vl_interface': 8.62304707070449e-07, 'learning_rate/action_model': 8.62304707070449e-06, 'epoch': 0.34}) +[2026-06-24 14:23:57,025][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.03815317526459694, 'train/grad_norm_pre_clip': 9.542847633361816, 'timing/data': 0.00034739403054118156, 'timing/model': 1.0694871309679002, 'learning_rate/qwen_vl_interface': 8.608404459731412e-07, 'learning_rate/action_model': 8.608404459731411e-06, 'epoch': 0.34}) +[2026-06-24 14:23:58,097][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.011363280937075615, 'train/grad_norm_pre_clip': 24.355506896972656, 'timing/data': 0.00036757183261215687, 'timing/model': 1.070670171873644, 'learning_rate/qwen_vl_interface': 8.593790329538147e-07, 'learning_rate/action_model': 8.593790329538148e-06, 'epoch': 0.34}) +[2026-06-24 14:23:59,167][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.0008914315840229392, 'train/grad_norm_pre_clip': 0.822413444519043, 'timing/data': 0.0003212951123714447, 'timing/model': 1.0684655669610947, 'learning_rate/qwen_vl_interface': 8.579204689607655e-07, 'learning_rate/action_model': 8.579204689607654e-06, 'epoch': 0.34}) +[2026-06-24 14:24:00,469][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.0012657587649300694, 'train/grad_norm_pre_clip': 1.6275981664657593, 'timing/data': 0.00035575893707573414, 'timing/model': 1.300247565843165, 'learning_rate/qwen_vl_interface': 8.564647549404406e-07, 'learning_rate/action_model': 8.564647549404405e-06, 'epoch': 0.34}) +[2026-06-24 14:24:01,540][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.044418103992938995, 'train/grad_norm_pre_clip': 19.271800994873047, 'timing/data': 0.0003652069717645645, 'timing/model': 1.0698832031339407, 'learning_rate/qwen_vl_interface': 8.550118918374354e-07, 'learning_rate/action_model': 8.550118918374354e-06, 'epoch': 0.34}) +[2026-06-24 14:24:02,610][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.06815505772829056, 'train/grad_norm_pre_clip': 42.63875961303711, 'timing/data': 0.00038449419662356377, 'timing/model': 1.0683637079782784, 'learning_rate/qwen_vl_interface': 8.535618805944981e-07, 'learning_rate/action_model': 8.535618805944981e-06, 'epoch': 0.34}) +[2026-06-24 14:24:03,684][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.22088377177715302, 'train/grad_norm_pre_clip': 73.74624633789062, 'timing/data': 0.0003586048260331154, 'timing/model': 1.0716162440367043, 'learning_rate/qwen_vl_interface': 8.52114722152525e-07, 'learning_rate/action_model': 8.521147221525249e-06, 'epoch': 0.34}) +[2026-06-24 14:24:04,757][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.08129838109016418, 'train/grad_norm_pre_clip': 78.78120422363281, 'timing/data': 0.000375926960259676, 'timing/model': 1.0714807740878314, 'learning_rate/qwen_vl_interface': 8.506704174505622e-07, 'learning_rate/action_model': 8.506704174505623e-06, 'epoch': 0.34}) +[2026-06-24 14:24:05,829][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.013253581710159779, 'train/grad_norm_pre_clip': 22.587209701538086, 'timing/data': 0.0003999930340796709, 'timing/model': 1.0702036840375513, 'learning_rate/qwen_vl_interface': 8.492289674258019e-07, 'learning_rate/action_model': 8.492289674258018e-06, 'epoch': 0.34}) +[2026-06-24 14:24:06,903][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.00815749540925026, 'train/grad_norm_pre_clip': 90.4848403930664, 'timing/data': 0.0003766289446502924, 'timing/model': 1.073044846067205, 'learning_rate/qwen_vl_interface': 8.477903730135863e-07, 'learning_rate/action_model': 8.477903730135862e-06, 'epoch': 0.34}) +[2026-06-24 14:24:07,990][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.129009410738945, 'train/grad_norm_pre_clip': 37.58749771118164, 'timing/data': 0.0003520518075674772, 'timing/model': 1.0849307018797845, 'learning_rate/qwen_vl_interface': 8.463546351474042e-07, 'learning_rate/action_model': 8.463546351474042e-06, 'epoch': 0.34}) +[2026-06-24 14:24:09,199][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.0029498424846678972, 'train/grad_norm_pre_clip': 3.2655675411224365, 'timing/data': 0.00040044402703642845, 'timing/model': 1.2069053701125085, 'learning_rate/qwen_vl_interface': 8.449217547588882e-07, 'learning_rate/action_model': 8.449217547588882e-06, 'epoch': 0.34}) +[2026-06-24 14:24:10,294][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.05435341224074364, 'train/grad_norm_pre_clip': 52.98141098022461, 'timing/data': 0.0003716200590133667, 'timing/model': 1.0933125070296228, 'learning_rate/qwen_vl_interface': 8.434917327778214e-07, 'learning_rate/action_model': 8.434917327778214e-06, 'epoch': 0.34}) +[2026-06-24 14:24:11,366][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.049212466925382614, 'train/grad_norm_pre_clip': 20.18382453918457, 'timing/data': 0.0003459018189460039, 'timing/model': 1.0708175799809396, 'learning_rate/qwen_vl_interface': 8.420645701321281e-07, 'learning_rate/action_model': 8.420645701321281e-06, 'epoch': 0.34}) +[2026-06-24 14:24:12,439][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.07232237607240677, 'train/grad_norm_pre_clip': 44.64311218261719, 'timing/data': 0.0003398801200091839, 'timing/model': 1.0708237669896334, 'learning_rate/qwen_vl_interface': 8.406402677478792e-07, 'learning_rate/action_model': 8.406402677478791e-06, 'epoch': 0.34}) +[2026-06-24 14:24:13,513][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.00036912114592269063, 'train/grad_norm_pre_clip': 0.43476128578186035, 'timing/data': 0.0003228078130632639, 'timing/model': 1.0727581318933517, 'learning_rate/qwen_vl_interface': 8.39218826549289e-07, 'learning_rate/action_model': 8.39218826549289e-06, 'epoch': 0.34}) +[2026-06-24 14:24:14,587][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.05341731756925583, 'train/grad_norm_pre_clip': 24.530582427978516, 'timing/data': 0.00033886684104800224, 'timing/model': 1.0724646640010178, 'learning_rate/qwen_vl_interface': 8.378002474587148e-07, 'learning_rate/action_model': 8.378002474587147e-06, 'epoch': 0.34}) +[2026-06-24 14:24:15,663][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.04455984756350517, 'train/grad_norm_pre_clip': 24.79473876953125, 'timing/data': 0.0003374649677425623, 'timing/model': 1.0743046319112182, 'learning_rate/qwen_vl_interface': 8.363845313966573e-07, 'learning_rate/action_model': 8.363845313966573e-06, 'epoch': 0.34}) +[2026-06-24 14:24:16,739][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.01327450480312109, 'train/grad_norm_pre_clip': 40.07012939453125, 'timing/data': 0.00031347014009952545, 'timing/model': 1.0740579620469362, 'learning_rate/qwen_vl_interface': 8.34971679281759e-07, 'learning_rate/action_model': 8.34971679281759e-06, 'epoch': 0.34}) +[2026-06-24 14:24:18,102][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.061821747571229935, 'train/grad_norm_pre_clip': 47.98555374145508, 'timing/data': 0.00033421884290874004, 'timing/model': 1.3616072558797896, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 0.34}) +[2026-06-24 14:24:19,176][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.03970376402139664, 'train/grad_norm_pre_clip': 41.071205139160156, 'timing/data': 0.00040654512122273445, 'timing/model': 1.0720642269589007, 'learning_rate/qwen_vl_interface': 8.321545705587216e-07, 'learning_rate/action_model': 8.321545705587215e-06, 'epoch': 0.34}) +[2026-06-24 14:24:20,249][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.014538008719682693, 'train/grad_norm_pre_clip': 34.53480911254883, 'timing/data': 0.0003423551097512245, 'timing/model': 1.0716395920608193, 'learning_rate/qwen_vl_interface': 8.307503157785726e-07, 'learning_rate/action_model': 8.307503157785726e-06, 'epoch': 0.34}) +[2026-06-24 14:24:21,319][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.10821119695901871, 'train/grad_norm_pre_clip': 83.7059326171875, 'timing/data': 0.00032614381052553654, 'timing/model': 1.0684171319007874, 'learning_rate/qwen_vl_interface': 8.293489286015653e-07, 'learning_rate/action_model': 8.293489286015653e-06, 'epoch': 0.34}) +[2026-06-24 14:24:22,392][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.015206500887870789, 'train/grad_norm_pre_clip': 26.114662170410156, 'timing/data': 0.00032894895412027836, 'timing/model': 1.0713894900400192, 'learning_rate/qwen_vl_interface': 8.27950409937043e-07, 'learning_rate/action_model': 8.27950409937043e-06, 'epoch': 0.34}) +[2026-06-24 14:24:23,464][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.028893664479255676, 'train/grad_norm_pre_clip': 37.93119430541992, 'timing/data': 0.00031623500399291515, 'timing/model': 1.070906907087192, 'learning_rate/qwen_vl_interface': 8.265547606924905e-07, 'learning_rate/action_model': 8.265547606924905e-06, 'epoch': 0.34}) +[2026-06-24 14:24:24,536][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.02253193035721779, 'train/grad_norm_pre_clip': 16.39009666442871, 'timing/data': 0.0003288588486611843, 'timing/model': 1.069519178941846, 'learning_rate/qwen_vl_interface': 8.251619817735295e-07, 'learning_rate/action_model': 8.251619817735295e-06, 'epoch': 0.34}) +[2026-06-24 14:24:25,614][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.030058957636356354, 'train/grad_norm_pre_clip': 25.796756744384766, 'timing/data': 0.0003213249146938324, 'timing/model': 1.076596864964813, 'learning_rate/qwen_vl_interface': 8.237720740839185e-07, 'learning_rate/action_model': 8.237720740839185e-06, 'epoch': 0.34}) +[2026-06-24 14:24:26,690][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.13972698152065277, 'train/grad_norm_pre_clip': 70.41374969482422, 'timing/data': 0.00032718502916395664, 'timing/model': 1.075046552112326, 'learning_rate/qwen_vl_interface': 8.223850385255536e-07, 'learning_rate/action_model': 8.223850385255536e-06, 'epoch': 0.34}) +[2026-06-24 14:24:27,764][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.02937949076294899, 'train/grad_norm_pre_clip': 13.614452362060547, 'timing/data': 0.0003190001007169485, 'timing/model': 1.071787622058764, 'learning_rate/qwen_vl_interface': 8.210008759984667e-07, 'learning_rate/action_model': 8.210008759984667e-06, 'epoch': 0.34}) +[2026-06-24 14:24:28,836][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.06491056084632874, 'train/grad_norm_pre_clip': 51.68695831298828, 'timing/data': 0.0003420740831643343, 'timing/model': 1.0705586548428982, 'learning_rate/qwen_vl_interface': 8.196195874008278e-07, 'learning_rate/action_model': 8.196195874008278e-06, 'epoch': 0.34}) +[2026-06-24 14:24:29,909][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.0831715390086174, 'train/grad_norm_pre_clip': 72.65340423583984, 'timing/data': 0.0003312830813229084, 'timing/model': 1.0712110470049083, 'learning_rate/qwen_vl_interface': 8.182411736289377e-07, 'learning_rate/action_model': 8.182411736289376e-06, 'epoch': 0.34}) +[2026-06-24 14:24:30,984][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.005247446242719889, 'train/grad_norm_pre_clip': 9.726350784301758, 'timing/data': 0.00036243186332285404, 'timing/model': 1.0739950311835855, 'learning_rate/qwen_vl_interface': 8.168656355772349e-07, 'learning_rate/action_model': 8.168656355772349e-06, 'epoch': 0.34}) +[2026-06-24 14:24:32,057][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.036735717207193375, 'train/grad_norm_pre_clip': 41.3775520324707, 'timing/data': 0.00034262496046721935, 'timing/model': 1.0714446359779686, 'learning_rate/qwen_vl_interface': 8.154929741382922e-07, 'learning_rate/action_model': 8.154929741382922e-06, 'epoch': 0.34}) +[2026-06-24 14:24:33,132][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.05472221225500107, 'train/grad_norm_pre_clip': 42.951171875, 'timing/data': 0.0003210240975022316, 'timing/model': 1.0732790420297533, 'learning_rate/qwen_vl_interface': 8.141231902028128e-07, 'learning_rate/action_model': 8.141231902028127e-06, 'epoch': 0.34}) +[2026-06-24 14:24:34,419][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.03007645718753338, 'train/grad_norm_pre_clip': 63.34876251220703, 'timing/data': 0.00040434091351926327, 'timing/model': 1.2848824569955468, 'learning_rate/qwen_vl_interface': 8.127562846596365e-07, 'learning_rate/action_model': 8.127562846596365e-06, 'epoch': 0.34}) +[2026-06-24 14:24:35,493][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.13032880425453186, 'train/grad_norm_pre_clip': 106.83568572998047, 'timing/data': 0.00035085086710751057, 'timing/model': 1.0730293409433216, 'learning_rate/qwen_vl_interface': 8.113922583957329e-07, 'learning_rate/action_model': 8.113922583957328e-06, 'epoch': 0.34}) +[2026-06-24 14:24:36,569][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.03548859432339668, 'train/grad_norm_pre_clip': 19.782739639282227, 'timing/data': 0.0003146519884467125, 'timing/model': 1.0743903080001473, 'learning_rate/qwen_vl_interface': 8.100311122962023e-07, 'learning_rate/action_model': 8.100311122962022e-06, 'epoch': 0.34}) +[2026-06-24 14:24:37,645][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.06682334095239639, 'train/grad_norm_pre_clip': 74.19458770751953, 'timing/data': 0.0003662700764834881, 'timing/model': 1.0742067070677876, 'learning_rate/qwen_vl_interface': 8.086728472442801e-07, 'learning_rate/action_model': 8.0867284724428e-06, 'epoch': 0.34}) +[2026-06-24 14:24:38,720][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.2071060985326767, 'train/grad_norm_pre_clip': 117.95841217041016, 'timing/data': 0.00031768810003995895, 'timing/model': 1.0733304631430656, 'learning_rate/qwen_vl_interface': 8.073174641213277e-07, 'learning_rate/action_model': 8.073174641213277e-06, 'epoch': 0.34}) +[2026-06-24 14:24:39,795][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.13241860270500183, 'train/grad_norm_pre_clip': 97.7507095336914, 'timing/data': 0.00031138607300817966, 'timing/model': 1.0732246681582183, 'learning_rate/qwen_vl_interface': 8.059649638068395e-07, 'learning_rate/action_model': 8.059649638068395e-06, 'epoch': 0.34}) +[2026-06-24 14:24:40,873][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.030975770205259323, 'train/grad_norm_pre_clip': 14.363432884216309, 'timing/data': 0.0003936709836125374, 'timing/model': 1.0767399489413947, 'learning_rate/qwen_vl_interface': 8.046153471784379e-07, 'learning_rate/action_model': 8.04615347178438e-06, 'epoch': 0.34}) +[2026-06-24 14:24:41,949][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.07002907991409302, 'train/grad_norm_pre_clip': 71.45893096923828, 'timing/data': 0.0003487970679998398, 'timing/model': 1.0744156991131604, 'learning_rate/qwen_vl_interface': 8.032686151118754e-07, 'learning_rate/action_model': 8.032686151118754e-06, 'epoch': 0.34}) +[2026-06-24 14:24:43,027][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.05219598114490509, 'train/grad_norm_pre_clip': 23.246929168701172, 'timing/data': 0.00031723687425255775, 'timing/model': 1.0759400120005012, 'learning_rate/qwen_vl_interface': 8.019247684810305e-07, 'learning_rate/action_model': 8.019247684810305e-06, 'epoch': 0.34}) +[2026-06-24 14:24:44,101][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.20614518225193024, 'train/grad_norm_pre_clip': 70.4323501586914, 'timing/data': 0.0003252420574426651, 'timing/model': 1.0731825169641525, 'learning_rate/qwen_vl_interface': 8.005838081579118e-07, 'learning_rate/action_model': 8.005838081579116e-06, 'epoch': 0.34}) +[2026-06-24 14:24:45,181][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.12361346185207367, 'train/grad_norm_pre_clip': 73.99210357666016, 'timing/data': 0.0003778110258281231, 'timing/model': 1.0780479600653052, 'learning_rate/qwen_vl_interface': 7.992457350126534e-07, 'learning_rate/action_model': 7.992457350126533e-06, 'epoch': 0.34}) +[2026-06-24 14:24:46,259][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.0277527067810297, 'train/grad_norm_pre_clip': 10.541631698608398, 'timing/data': 0.0003133900463581085, 'timing/model': 1.076205366058275, 'learning_rate/qwen_vl_interface': 7.979105499135157e-07, 'learning_rate/action_model': 7.979105499135155e-06, 'epoch': 0.34}) +[2026-06-24 14:24:47,338][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.1947975754737854, 'train/grad_norm_pre_clip': 92.77698516845703, 'timing/data': 0.000318117905408144, 'timing/model': 1.0773291059304029, 'learning_rate/qwen_vl_interface': 7.965782537268866e-07, 'learning_rate/action_model': 7.965782537268867e-06, 'epoch': 0.34}) +[2026-06-24 14:24:48,418][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.08913803100585938, 'train/grad_norm_pre_clip': 57.40931701660156, 'timing/data': 0.0003831221256405115, 'timing/model': 1.0779940120410174, 'learning_rate/qwen_vl_interface': 7.952488473172796e-07, 'learning_rate/action_model': 7.952488473172795e-06, 'epoch': 0.34}) +[2026-06-24 14:24:49,500][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.02483387105166912, 'train/grad_norm_pre_clip': 24.417993545532227, 'timing/data': 0.00031702592968940735, 'timing/model': 1.080466871149838, 'learning_rate/qwen_vl_interface': 7.939223315473299e-07, 'learning_rate/action_model': 7.9392233154733e-06, 'epoch': 0.34}) +[2026-06-24 14:24:50,791][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.019043004140257835, 'train/grad_norm_pre_clip': 13.010329246520996, 'timing/data': 0.0003092419356107712, 'timing/model': 1.2899476550519466, 'learning_rate/qwen_vl_interface': 7.925987072778005e-07, 'learning_rate/action_model': 7.925987072778004e-06, 'epoch': 0.34}) +[2026-06-24 14:24:51,876][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.11372444033622742, 'train/grad_norm_pre_clip': 87.0269775390625, 'timing/data': 0.0003424349706619978, 'timing/model': 1.0836727428250015, 'learning_rate/qwen_vl_interface': 7.912779753675764e-07, 'learning_rate/action_model': 7.912779753675765e-06, 'epoch': 0.34}) +[2026-06-24 14:24:52,956][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.009467294439673424, 'train/grad_norm_pre_clip': 11.448929786682129, 'timing/data': 0.0003456010017544031, 'timing/model': 1.0774846819695085, 'learning_rate/qwen_vl_interface': 7.899601366736675e-07, 'learning_rate/action_model': 7.899601366736675e-06, 'epoch': 0.34}) +[2026-06-24 14:24:54,036][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.14085356891155243, 'train/grad_norm_pre_clip': 91.65148162841797, 'timing/data': 0.00033526099286973476, 'timing/model': 1.0790244340896606, 'learning_rate/qwen_vl_interface': 7.886451920512028e-07, 'learning_rate/action_model': 7.886451920512028e-06, 'epoch': 0.34}) +[2026-06-24 14:24:55,117][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.34249788522720337, 'train/grad_norm_pre_clip': 71.89537811279297, 'timing/data': 0.0003319240640848875, 'timing/model': 1.0793201250489801, 'learning_rate/qwen_vl_interface': 7.873331423534372e-07, 'learning_rate/action_model': 7.873331423534373e-06, 'epoch': 0.35}) +[2026-06-24 14:24:56,195][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.025426072999835014, 'train/grad_norm_pre_clip': 16.869226455688477, 'timing/data': 0.0003201230429112911, 'timing/model': 1.0763031621463597, 'learning_rate/qwen_vl_interface': 7.860239884317456e-07, 'learning_rate/action_model': 7.860239884317455e-06, 'epoch': 0.35}) +[2026-06-24 14:24:57,278][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.20050348341464996, 'train/grad_norm_pre_clip': 62.03531265258789, 'timing/data': 0.00036745122633874416, 'timing/model': 1.0811732020229101, 'learning_rate/qwen_vl_interface': 7.847177311356214e-07, 'learning_rate/action_model': 7.847177311356215e-06, 'epoch': 0.35}) +[2026-06-24 14:24:58,361][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.015288886614143848, 'train/grad_norm_pre_clip': 36.725955963134766, 'timing/data': 0.00034333602525293827, 'timing/model': 1.081593558890745, 'learning_rate/qwen_vl_interface': 7.834143713126849e-07, 'learning_rate/action_model': 7.834143713126848e-06, 'epoch': 0.35}) +[2026-06-24 14:24:59,441][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.21471591293811798, 'train/grad_norm_pre_clip': 87.45625305175781, 'timing/data': 0.0004054729361087084, 'timing/model': 1.0780874050687999, 'learning_rate/qwen_vl_interface': 7.821139098086685e-07, 'learning_rate/action_model': 7.821139098086685e-06, 'epoch': 0.35}) +[2026-06-24 14:25:00,525][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.3952605426311493, 'train/grad_norm_pre_clip': 55.37752914428711, 'timing/data': 0.00034675304777920246, 'timing/model': 1.082622073823586, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 0.35}) +[2026-06-24 14:25:01,609][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.1788272112607956, 'train/grad_norm_pre_clip': 51.6536750793457, 'timing/data': 0.000391968060284853, 'timing/model': 1.0824898600112647, 'learning_rate/qwen_vl_interface': 7.795216851309407e-07, 'learning_rate/action_model': 7.795216851309406e-06, 'epoch': 0.35}) +[2026-06-24 14:25:02,694][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.09370715171098709, 'train/grad_norm_pre_clip': 62.03593444824219, 'timing/data': 0.00035970681346952915, 'timing/model': 1.0832910418976098, 'learning_rate/qwen_vl_interface': 7.782299236392958e-07, 'learning_rate/action_model': 7.782299236392957e-06, 'epoch': 0.35}) +[2026-06-24 14:25:03,776][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.010053771547973156, 'train/grad_norm_pre_clip': 28.43549919128418, 'timing/data': 0.0003401900175958872, 'timing/model': 1.0807741209864616, 'learning_rate/qwen_vl_interface': 7.769410638307034e-07, 'learning_rate/action_model': 7.769410638307034e-06, 'epoch': 0.35}) +[2026-06-24 14:25:04,857][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.13592348992824554, 'train/grad_norm_pre_clip': 83.87519073486328, 'timing/data': 0.00032655405811965466, 'timing/model': 1.0789586987812072, 'learning_rate/qwen_vl_interface': 7.756551065414919e-07, 'learning_rate/action_model': 7.756551065414918e-06, 'epoch': 0.35}) +[2026-06-24 14:25:05,939][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.13569781184196472, 'train/grad_norm_pre_clip': 39.8198127746582, 'timing/data': 0.00032468116842210293, 'timing/model': 1.08050049492158, 'learning_rate/qwen_vl_interface': 7.743720526061047e-07, 'learning_rate/action_model': 7.743720526061047e-06, 'epoch': 0.35}) +[2026-06-24 14:25:07,219][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.029948782175779343, 'train/grad_norm_pre_clip': 7.393404960632324, 'timing/data': 0.00033578183501958847, 'timing/model': 1.2780280190054327, 'learning_rate/qwen_vl_interface': 7.73091902857101e-07, 'learning_rate/action_model': 7.73091902857101e-06, 'epoch': 0.35}) +[2026-06-24 14:25:08,302][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.15212410688400269, 'train/grad_norm_pre_clip': 65.60508728027344, 'timing/data': 0.000462990952655673, 'timing/model': 1.0808371279854327, 'learning_rate/qwen_vl_interface': 7.718146581251565e-07, 'learning_rate/action_model': 7.718146581251565e-06, 'epoch': 0.35}) +[2026-06-24 14:25:09,382][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.2288402020931244, 'train/grad_norm_pre_clip': 77.45240783691406, 'timing/data': 0.0003424638416618109, 'timing/model': 1.0784429179038852, 'learning_rate/qwen_vl_interface': 7.705403192390617e-07, 'learning_rate/action_model': 7.705403192390617e-06, 'epoch': 0.35}) +[2026-06-24 14:25:10,467][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.3230839967727661, 'train/grad_norm_pre_clip': 67.84207153320312, 'timing/data': 0.00036554806865751743, 'timing/model': 1.0836642279755324, 'learning_rate/qwen_vl_interface': 7.6926888702572e-07, 'learning_rate/action_model': 7.6926888702572e-06, 'epoch': 0.35}) +[2026-06-24 14:25:11,549][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.001327595440670848, 'train/grad_norm_pre_clip': 0.9845957159996033, 'timing/data': 0.00031617493368685246, 'timing/model': 1.0808573311660439, 'learning_rate/qwen_vl_interface': 7.680003623101517e-07, 'learning_rate/action_model': 7.680003623101516e-06, 'epoch': 0.35}) +[2026-06-24 14:25:12,630][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.20443229377269745, 'train/grad_norm_pre_clip': 247.9139862060547, 'timing/data': 0.00031882012262940407, 'timing/model': 1.0784884870518, 'learning_rate/qwen_vl_interface': 7.667347459154884e-07, 'learning_rate/action_model': 7.667347459154884e-06, 'epoch': 0.35}) +[2026-06-24 14:25:13,712][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.02804538607597351, 'train/grad_norm_pre_clip': 49.148189544677734, 'timing/data': 0.0003318139351904392, 'timing/model': 1.0809689508751035, 'learning_rate/qwen_vl_interface': 7.654720386629732e-07, 'learning_rate/action_model': 7.654720386629732e-06, 'epoch': 0.35}) +[2026-06-24 14:25:14,791][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.1679278314113617, 'train/grad_norm_pre_clip': 70.00276947021484, 'timing/data': 0.00032041314989328384, 'timing/model': 1.077173543162644, 'learning_rate/qwen_vl_interface': 7.642122413719652e-07, 'learning_rate/action_model': 7.642122413719651e-06, 'epoch': 0.35}) +[2026-06-24 14:25:15,877][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.3950643539428711, 'train/grad_norm_pre_clip': 25.19619369506836, 'timing/data': 0.0003082409966737032, 'timing/model': 1.084517349023372, 'learning_rate/qwen_vl_interface': 7.62955354859932e-07, 'learning_rate/action_model': 7.629553548599318e-06, 'epoch': 0.35}) +[2026-06-24 14:25:16,960][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.1082042008638382, 'train/grad_norm_pre_clip': 77.17412567138672, 'timing/data': 0.0003359820693731308, 'timing/model': 1.0808254659641534, 'learning_rate/qwen_vl_interface': 7.617013799424536e-07, 'learning_rate/action_model': 7.617013799424535e-06, 'epoch': 0.35}) +[2026-06-24 14:25:18,042][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.06927115470170975, 'train/grad_norm_pre_clip': 23.475021362304688, 'timing/data': 0.0003178978804498911, 'timing/model': 1.0805453229695559, 'learning_rate/qwen_vl_interface': 7.604503174332217e-07, 'learning_rate/action_model': 7.604503174332217e-06, 'epoch': 0.35}) +[2026-06-24 14:25:19,129][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.11393041908740997, 'train/grad_norm_pre_clip': 42.916717529296875, 'timing/data': 0.00035989820025861263, 'timing/model': 1.0857049189507961, 'learning_rate/qwen_vl_interface': 7.592021681440356e-07, 'learning_rate/action_model': 7.592021681440356e-06, 'epoch': 0.35}) +[2026-06-24 14:25:20,218][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.06634403020143509, 'train/grad_norm_pre_clip': 88.35311126708984, 'timing/data': 0.0003415131941437721, 'timing/model': 1.08687639888376, 'learning_rate/qwen_vl_interface': 7.579569328848072e-07, 'learning_rate/action_model': 7.579569328848071e-06, 'epoch': 0.35}) +[2026-06-24 14:25:21,301][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.1344742327928543, 'train/grad_norm_pre_clip': 103.91758728027344, 'timing/data': 0.0003457409329712391, 'timing/model': 1.0812420099973679, 'learning_rate/qwen_vl_interface': 7.567146124635554e-07, 'learning_rate/action_model': 7.567146124635553e-06, 'epoch': 0.35}) +[2026-06-24 14:25:22,388][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.010375581681728363, 'train/grad_norm_pre_clip': 12.610901832580566, 'timing/data': 0.0003158240579068661, 'timing/model': 1.086085030110553, 'learning_rate/qwen_vl_interface': 7.554752076864089e-07, 'learning_rate/action_model': 7.554752076864088e-06, 'epoch': 0.35}) +[2026-06-24 14:25:23,770][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.048885706812143326, 'train/grad_norm_pre_clip': 40.73413848876953, 'timing/data': 0.00034135207533836365, 'timing/model': 1.3797710889484733, 'learning_rate/qwen_vl_interface': 7.542387193576035e-07, 'learning_rate/action_model': 7.542387193576035e-06, 'epoch': 0.35}) +[2026-06-24 14:25:24,860][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.009557889774441719, 'train/grad_norm_pre_clip': 6.741866111755371, 'timing/data': 0.0003564308863133192, 'timing/model': 1.0796557390131056, 'learning_rate/qwen_vl_interface': 7.530051482794833e-07, 'learning_rate/action_model': 7.530051482794832e-06, 'epoch': 0.35}) +[2026-06-24 14:25:25,945][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.09859870374202728, 'train/grad_norm_pre_clip': 77.53570556640625, 'timing/data': 0.00037030712701380253, 'timing/model': 1.0835300348699093, 'learning_rate/qwen_vl_interface': 7.517744952525003e-07, 'learning_rate/action_model': 7.5177449525250025e-06, 'epoch': 0.35}) +[2026-06-24 14:25:27,030][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.09504248201847076, 'train/grad_norm_pre_clip': 50.320552825927734, 'timing/data': 0.00031221797689795494, 'timing/model': 1.0829292701091617, 'learning_rate/qwen_vl_interface': 7.505467610752093e-07, 'learning_rate/action_model': 7.505467610752093e-06, 'epoch': 0.35}) +[2026-06-24 14:25:28,114][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.1165657564997673, 'train/grad_norm_pre_clip': 70.12036895751953, 'timing/data': 0.000372591195628047, 'timing/model': 1.0827772880438715, 'learning_rate/qwen_vl_interface': 7.493219465442761e-07, 'learning_rate/action_model': 7.4932194654427605e-06, 'epoch': 0.35}) +[2026-06-24 14:25:29,199][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.007193494588136673, 'train/grad_norm_pre_clip': 7.720542907714844, 'timing/data': 0.00033151404932141304, 'timing/model': 1.0837322410661727, 'learning_rate/qwen_vl_interface': 7.481000524544686e-07, 'learning_rate/action_model': 7.481000524544686e-06, 'epoch': 0.35}) +[2026-06-24 14:25:30,286][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.06560654193162918, 'train/grad_norm_pre_clip': 48.400184631347656, 'timing/data': 0.00031216791830956936, 'timing/model': 1.0853219768032432, 'learning_rate/qwen_vl_interface': 7.468810795986598e-07, 'learning_rate/action_model': 7.4688107959865976e-06, 'epoch': 0.35}) +[2026-06-24 14:25:31,372][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.019458957016468048, 'train/grad_norm_pre_clip': 20.730308532714844, 'timing/data': 0.0003279969096183777, 'timing/model': 1.0846481840126216, 'learning_rate/qwen_vl_interface': 7.456650287678292e-07, 'learning_rate/action_model': 7.456650287678292e-06, 'epoch': 0.35}) +[2026-06-24 14:25:32,461][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.147378608584404, 'train/grad_norm_pre_clip': 70.8105239868164, 'timing/data': 0.0003416521940380335, 'timing/model': 1.0866510178893805, 'learning_rate/qwen_vl_interface': 7.44451900751059e-07, 'learning_rate/action_model': 7.444519007510589e-06, 'epoch': 0.35}) +[2026-06-24 14:25:33,549][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.11591708660125732, 'train/grad_norm_pre_clip': 68.39205932617188, 'timing/data': 0.00031453208066523075, 'timing/model': 1.086661412846297, 'learning_rate/qwen_vl_interface': 7.432416963355343e-07, 'learning_rate/action_model': 7.432416963355343e-06, 'epoch': 0.35}) +[2026-06-24 14:25:34,635][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.067597895860672, 'train/grad_norm_pre_clip': 50.224700927734375, 'timing/data': 0.0003547980450093746, 'timing/model': 1.0844068059232086, 'learning_rate/qwen_vl_interface': 7.420344163065427e-07, 'learning_rate/action_model': 7.420344163065427e-06, 'epoch': 0.35}) +[2026-06-24 14:25:35,718][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.004674597643315792, 'train/grad_norm_pre_clip': 4.91774845123291, 'timing/data': 0.00036196201108396053, 'timing/model': 1.0813795940484852, 'learning_rate/qwen_vl_interface': 7.40830061447477e-07, 'learning_rate/action_model': 7.408300614474769e-06, 'epoch': 0.35}) +[2026-06-24 14:25:36,806][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.03499205410480499, 'train/grad_norm_pre_clip': 18.60992431640625, 'timing/data': 0.00034375698305666447, 'timing/model': 1.086136271012947, 'learning_rate/qwen_vl_interface': 7.396286325398297e-07, 'learning_rate/action_model': 7.396286325398297e-06, 'epoch': 0.35}) +[2026-06-24 14:25:37,892][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.11435011029243469, 'train/grad_norm_pre_clip': 65.94136810302734, 'timing/data': 0.00031627481803297997, 'timing/model': 1.085136384004727, 'learning_rate/qwen_vl_interface': 7.384301303631927e-07, 'learning_rate/action_model': 7.384301303631927e-06, 'epoch': 0.35}) +[2026-06-24 14:25:39,180][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.007763315923511982, 'train/grad_norm_pre_clip': 14.767788887023926, 'timing/data': 0.0003350710030645132, 'timing/model': 1.2858459800481796, 'learning_rate/qwen_vl_interface': 7.372345556952642e-07, 'learning_rate/action_model': 7.372345556952642e-06, 'epoch': 0.35}) +[2026-06-24 14:25:40,266][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.028432738035917282, 'train/grad_norm_pre_clip': 37.777164459228516, 'timing/data': 0.00035706209018826485, 'timing/model': 1.0841291879769415, 'learning_rate/qwen_vl_interface': 7.360419093118368e-07, 'learning_rate/action_model': 7.360419093118367e-06, 'epoch': 0.35}) +[2026-06-24 14:25:41,353][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.009445972740650177, 'train/grad_norm_pre_clip': 8.14048957824707, 'timing/data': 0.00033083301968872547, 'timing/model': 1.086060767993331, 'learning_rate/qwen_vl_interface': 7.348521919868076e-07, 'learning_rate/action_model': 7.348521919868076e-06, 'epoch': 0.35}) +[2026-06-24 14:25:42,440][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.0386907123029232, 'train/grad_norm_pre_clip': 35.92528533935547, 'timing/data': 0.0003882809542119503, 'timing/model': 1.0849808920174837, 'learning_rate/qwen_vl_interface': 7.336654044921715e-07, 'learning_rate/action_model': 7.336654044921714e-06, 'epoch': 0.35}) +[2026-06-24 14:25:43,528][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.10877778381109238, 'train/grad_norm_pre_clip': 37.469783782958984, 'timing/data': 0.00038514495827257633, 'timing/model': 1.0866199319716543, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 0.35}) +[2026-06-24 14:25:44,613][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.11708881705999374, 'train/grad_norm_pre_clip': 77.74178314208984, 'timing/data': 0.00031819893047213554, 'timing/model': 1.083547294139862, 'learning_rate/qwen_vl_interface': 7.313006220725475e-07, 'learning_rate/action_model': 7.3130062207254755e-06, 'epoch': 0.35}) +[2026-06-24 14:25:45,703][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.057773344218730927, 'train/grad_norm_pre_clip': 33.68556213378906, 'timing/data': 0.00043408595956861973, 'timing/model': 1.0881327691022307, 'learning_rate/qwen_vl_interface': 7.301226286820418e-07, 'learning_rate/action_model': 7.301226286820417e-06, 'epoch': 0.35}) +[2026-06-24 14:25:46,790][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.11510686576366425, 'train/grad_norm_pre_clip': 71.35466003417969, 'timing/data': 0.0004093910101801157, 'timing/model': 1.0850334509741515, 'learning_rate/qwen_vl_interface': 7.289475681908916e-07, 'learning_rate/action_model': 7.289475681908915e-06, 'epoch': 0.35}) +[2026-06-24 14:25:47,874][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.03257914260029793, 'train/grad_norm_pre_clip': 76.10990142822266, 'timing/data': 0.00034044100902974606, 'timing/model': 1.0819453860167414, 'learning_rate/qwen_vl_interface': 7.277754413615792e-07, 'learning_rate/action_model': 7.277754413615792e-06, 'epoch': 0.35}) +[2026-06-24 14:25:48,956][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.16722960770130157, 'train/grad_norm_pre_clip': 26.092710494995117, 'timing/data': 0.00033056200481951237, 'timing/model': 1.0811547730118036, 'learning_rate/qwen_vl_interface': 7.266062489546855e-07, 'learning_rate/action_model': 7.266062489546854e-06, 'epoch': 0.35}) +[2026-06-24 14:25:50,045][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.17343637347221375, 'train/grad_norm_pre_clip': 37.750946044921875, 'timing/data': 0.0003549579996615648, 'timing/model': 1.086692017968744, 'learning_rate/qwen_vl_interface': 7.254399917288885e-07, 'learning_rate/action_model': 7.254399917288884e-06, 'epoch': 0.35}) +[2026-06-24 14:25:51,134][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.07245659083127975, 'train/grad_norm_pre_clip': 54.148887634277344, 'timing/data': 0.0003219149075448513, 'timing/model': 1.0873317569494247, 'learning_rate/qwen_vl_interface': 7.242766704409574e-07, 'learning_rate/action_model': 7.242766704409573e-06, 'epoch': 0.35}) +[2026-06-24 14:25:52,220][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.0015182668576017022, 'train/grad_norm_pre_clip': 2.664548397064209, 'timing/data': 0.0005943591240793467, 'timing/model': 1.0842554341070354, 'learning_rate/qwen_vl_interface': 7.231162858457595e-07, 'learning_rate/action_model': 7.231162858457595e-06, 'epoch': 0.35}) +[2026-06-24 14:25:53,306][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.006519509945064783, 'train/grad_norm_pre_clip': 5.592460632324219, 'timing/data': 0.00030501396395266056, 'timing/model': 1.0837917169556022, 'learning_rate/qwen_vl_interface': 7.21958838696257e-07, 'learning_rate/action_model': 7.219588386962569e-06, 'epoch': 0.35}) +[2026-06-24 14:25:54,395][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.054358191788196564, 'train/grad_norm_pre_clip': 42.448219299316406, 'timing/data': 0.0003238590434193611, 'timing/model': 1.088100709952414, 'learning_rate/qwen_vl_interface': 7.208043297435028e-07, 'learning_rate/action_model': 7.208043297435029e-06, 'epoch': 0.35}) +[2026-06-24 14:25:55,484][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.15078924596309662, 'train/grad_norm_pre_clip': 63.29733657836914, 'timing/data': 0.0006852601654827595, 'timing/model': 1.08671118109487, 'learning_rate/qwen_vl_interface': 7.196527597366467e-07, 'learning_rate/action_model': 7.196527597366466e-06, 'epoch': 0.35}) +[2026-06-24 14:25:56,573][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.003711962141096592, 'train/grad_norm_pre_clip': 2.813849449157715, 'timing/data': 0.0005501250270754099, 'timing/model': 1.0870522479526699, 'learning_rate/qwen_vl_interface': 7.185041294229292e-07, 'learning_rate/action_model': 7.185041294229292e-06, 'epoch': 0.35}) +[2026-06-24 14:25:57,662][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.10826069116592407, 'train/grad_norm_pre_clip': 183.58346557617188, 'timing/data': 0.0007123698014765978, 'timing/model': 1.086656047962606, 'learning_rate/qwen_vl_interface': 7.173584395476852e-07, 'learning_rate/action_model': 7.173584395476851e-06, 'epoch': 0.35}) +[2026-06-24 14:25:59,042][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.15876932442188263, 'train/grad_norm_pre_clip': 60.11936950683594, 'timing/data': 0.0003324849531054497, 'timing/model': 1.3785369188990444, 'learning_rate/qwen_vl_interface': 7.162156908543383e-07, 'learning_rate/action_model': 7.162156908543382e-06, 'epoch': 0.35}) +[2026-06-24 14:26:00,130][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.041760772466659546, 'train/grad_norm_pre_clip': 64.94010162353516, 'timing/data': 0.0003256530035287142, 'timing/model': 1.0857631771359593, 'learning_rate/qwen_vl_interface': 7.150758840844084e-07, 'learning_rate/action_model': 7.1507588408440836e-06, 'epoch': 0.35}) +[2026-06-24 14:26:01,219][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.17434214055538177, 'train/grad_norm_pre_clip': 48.16880798339844, 'timing/data': 0.00034126220270991325, 'timing/model': 1.0873696359340101, 'learning_rate/qwen_vl_interface': 7.139390199775024e-07, 'learning_rate/action_model': 7.1393901997750245e-06, 'epoch': 0.35}) +[2026-06-24 14:26:02,308][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.21529199182987213, 'train/grad_norm_pre_clip': 85.49824523925781, 'timing/data': 0.0003215949982404709, 'timing/model': 1.087907933164388, 'learning_rate/qwen_vl_interface': 7.128050992713173e-07, 'learning_rate/action_model': 7.128050992713173e-06, 'epoch': 0.35}) +[2026-06-24 14:26:03,395][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.06725220382213593, 'train/grad_norm_pre_clip': 62.09514617919922, 'timing/data': 0.00032183597795665264, 'timing/model': 1.0856459829956293, 'learning_rate/qwen_vl_interface': 7.116741227016455e-07, 'learning_rate/action_model': 7.116741227016455e-06, 'epoch': 0.35}) +[2026-06-24 14:26:04,484][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.20316562056541443, 'train/grad_norm_pre_clip': 86.5447769165039, 'timing/data': 0.0003487959038466215, 'timing/model': 1.08664037915878, 'learning_rate/qwen_vl_interface': 7.105460910023627e-07, 'learning_rate/action_model': 7.105460910023627e-06, 'epoch': 0.35}) +[2026-06-24 14:26:05,573][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.17728887498378754, 'train/grad_norm_pre_clip': 84.22710418701172, 'timing/data': 0.00035459804348647594, 'timing/model': 1.0878698190208524, 'learning_rate/qwen_vl_interface': 7.09421004905437e-07, 'learning_rate/action_model': 7.09421004905437e-06, 'epoch': 0.35}) +[2026-06-24 14:26:06,658][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.22087247669696808, 'train/grad_norm_pre_clip': 52.39204788208008, 'timing/data': 0.0003323459532111883, 'timing/model': 1.0833423451986164, 'learning_rate/qwen_vl_interface': 7.082988651409257e-07, 'learning_rate/action_model': 7.082988651409257e-06, 'epoch': 0.35}) +[2026-06-24 14:26:07,750][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.04380272328853607, 'train/grad_norm_pre_clip': 40.98830032348633, 'timing/data': 0.000335911987349391, 'timing/model': 1.090414117090404, 'learning_rate/qwen_vl_interface': 7.071796724369718e-07, 'learning_rate/action_model': 7.071796724369718e-06, 'epoch': 0.35}) +[2026-06-24 14:26:08,840][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.015069171786308289, 'train/grad_norm_pre_clip': 24.044015884399414, 'timing/data': 0.00037288200110197067, 'timing/model': 1.0875094670336694, 'learning_rate/qwen_vl_interface': 7.060634275198078e-07, 'learning_rate/action_model': 7.0606342751980775e-06, 'epoch': 0.35}) +[2026-06-24 14:26:09,922][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.15396931767463684, 'train/grad_norm_pre_clip': 58.505428314208984, 'timing/data': 0.00034935795702040195, 'timing/model': 1.0805798128712922, 'learning_rate/qwen_vl_interface': 7.049501311137526e-07, 'learning_rate/action_model': 7.049501311137525e-06, 'epoch': 0.35}) +[2026-06-24 14:26:11,006][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.27771857380867004, 'train/grad_norm_pre_clip': 56.36811447143555, 'timing/data': 0.0003219260834157467, 'timing/model': 1.0821410759817809, 'learning_rate/qwen_vl_interface': 7.038397839412134e-07, 'learning_rate/action_model': 7.038397839412134e-06, 'epoch': 0.35}) +[2026-06-24 14:26:12,094][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.09190073609352112, 'train/grad_norm_pre_clip': 110.38832092285156, 'timing/data': 0.0003249009605497122, 'timing/model': 1.086994705023244, 'learning_rate/qwen_vl_interface': 7.027323867226807e-07, 'learning_rate/action_model': 7.0273238672268065e-06, 'epoch': 0.35}) +[2026-06-24 14:26:13,183][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.15410271286964417, 'train/grad_norm_pre_clip': 81.20301055908203, 'timing/data': 0.000329089118167758, 'timing/model': 1.0868449581321329, 'learning_rate/qwen_vl_interface': 7.016279401767333e-07, 'learning_rate/action_model': 7.0162794017673324e-06, 'epoch': 0.35}) +[2026-06-24 14:26:14,272][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.1826658844947815, 'train/grad_norm_pre_clip': 68.46444702148438, 'timing/data': 0.0003202119842171669, 'timing/model': 1.0874255360104144, 'learning_rate/qwen_vl_interface': 7.005264450200349e-07, 'learning_rate/action_model': 7.005264450200348e-06, 'epoch': 0.35}) +[2026-06-24 14:26:15,356][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.028100689873099327, 'train/grad_norm_pre_clip': 30.24976921081543, 'timing/data': 0.0003149330150336027, 'timing/model': 1.0823709710966796, 'learning_rate/qwen_vl_interface': 6.994279019673329e-07, 'learning_rate/action_model': 6.994279019673329e-06, 'epoch': 0.35}) +[2026-06-24 14:26:16,447][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.06878610700368881, 'train/grad_norm_pre_clip': 23.005996704101562, 'timing/data': 0.00036832294426858425, 'timing/model': 1.0901585000101477, 'learning_rate/qwen_vl_interface': 6.983323117314604e-07, 'learning_rate/action_model': 6.983323117314604e-06, 'epoch': 0.35}) +[2026-06-24 14:26:17,780][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.006402971222996712, 'train/grad_norm_pre_clip': 12.18281364440918, 'timing/data': 0.0003300318494439125, 'timing/model': 1.3307520139496773, 'learning_rate/qwen_vl_interface': 6.972396750233348e-07, 'learning_rate/action_model': 6.972396750233348e-06, 'epoch': 0.35}) +[2026-06-24 14:26:18,866][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.013434925116598606, 'train/grad_norm_pre_clip': 10.464073181152344, 'timing/data': 0.00032961112447082996, 'timing/model': 1.084948896896094, 'learning_rate/qwen_vl_interface': 6.961499925519549e-07, 'learning_rate/action_model': 6.961499925519549e-06, 'epoch': 0.35}) +[2026-06-24 14:26:19,953][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.028088266029953957, 'train/grad_norm_pre_clip': 35.85381317138672, 'timing/data': 0.00036471709609031677, 'timing/model': 1.0846779928542674, 'learning_rate/qwen_vl_interface': 6.950632650244045e-07, 'learning_rate/action_model': 6.950632650244045e-06, 'epoch': 0.35}) +[2026-06-24 14:26:21,041][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.018860356882214546, 'train/grad_norm_pre_clip': 31.388851165771484, 'timing/data': 0.00038446392863988876, 'timing/model': 1.0867395480163395, 'learning_rate/qwen_vl_interface': 6.939794931458489e-07, 'learning_rate/action_model': 6.939794931458489e-06, 'epoch': 0.35}) +[2026-06-24 14:26:22,133][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.04231691360473633, 'train/grad_norm_pre_clip': 63.48789596557617, 'timing/data': 0.00031322892755270004, 'timing/model': 1.089878219179809, 'learning_rate/qwen_vl_interface': 6.928986776195371e-07, 'learning_rate/action_model': 6.92898677619537e-06, 'epoch': 0.35}) +[2026-06-24 14:26:23,221][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.08360245823860168, 'train/grad_norm_pre_clip': 60.46706008911133, 'timing/data': 0.0003508199006319046, 'timing/model': 1.0870021029841155, 'learning_rate/qwen_vl_interface': 6.918208191467968e-07, 'learning_rate/action_model': 6.918208191467968e-06, 'epoch': 0.35}) +[2026-06-24 14:26:24,311][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.15402595698833466, 'train/grad_norm_pre_clip': 73.7813720703125, 'timing/data': 0.00034344615414738655, 'timing/model': 1.0884248460642993, 'learning_rate/qwen_vl_interface': 6.907459184270398e-07, 'learning_rate/action_model': 6.907459184270398e-06, 'epoch': 0.35}) +[2026-06-24 14:26:25,400][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.12690310180187225, 'train/grad_norm_pre_clip': 40.72941207885742, 'timing/data': 0.0003646360710263252, 'timing/model': 1.087079619988799, 'learning_rate/qwen_vl_interface': 6.896739761577575e-07, 'learning_rate/action_model': 6.896739761577575e-06, 'epoch': 0.35}) +[2026-06-24 14:26:26,488][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.00508869206532836, 'train/grad_norm_pre_clip': 5.924936771392822, 'timing/data': 0.00032087392173707485, 'timing/model': 1.0861387448385358, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 0.35}) +[2026-06-24 14:26:27,576][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.20452570915222168, 'train/grad_norm_pre_clip': 29.881397247314453, 'timing/data': 0.0003175979945808649, 'timing/model': 1.0868814189452678, 'learning_rate/qwen_vl_interface': 6.875389697509818e-07, 'learning_rate/action_model': 6.875389697509818e-06, 'epoch': 0.35}) +[2026-06-24 14:26:28,662][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.20633737742900848, 'train/grad_norm_pre_clip': 87.0213623046875, 'timing/data': 0.0003478941507637501, 'timing/model': 1.0836052191443741, 'learning_rate/qwen_vl_interface': 6.864759069988715e-07, 'learning_rate/action_model': 6.864759069988715e-06, 'epoch': 0.35}) +[2026-06-24 14:26:29,752][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.08257859945297241, 'train/grad_norm_pre_clip': 37.4594841003418, 'timing/data': 0.00034530903212726116, 'timing/model': 1.0885117570869625, 'learning_rate/qwen_vl_interface': 6.854158054679998e-07, 'learning_rate/action_model': 6.854158054679997e-06, 'epoch': 0.35}) +[2026-06-24 14:26:30,844][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.14412415027618408, 'train/grad_norm_pre_clip': 54.9077262878418, 'timing/data': 0.0003776499070227146, 'timing/model': 1.0902815428562462, 'learning_rate/qwen_vl_interface': 6.84358665846255e-07, 'learning_rate/action_model': 6.84358665846255e-06, 'epoch': 0.35}) +[2026-06-24 14:26:31,935][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.04935745894908905, 'train/grad_norm_pre_clip': 32.959041595458984, 'timing/data': 0.000353905837982893, 'timing/model': 1.0890341319609433, 'learning_rate/qwen_vl_interface': 6.833044888196032e-07, 'learning_rate/action_model': 6.833044888196031e-06, 'epoch': 0.35}) +[2026-06-24 14:26:33,027][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.007272383663803339, 'train/grad_norm_pre_clip': 7.968077659606934, 'timing/data': 0.00031972117722034454, 'timing/model': 1.0905860180500895, 'learning_rate/qwen_vl_interface': 6.822532750720889e-07, 'learning_rate/action_model': 6.822532750720889e-06, 'epoch': 0.35}) +[2026-06-24 14:26:34,119][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.026829950511455536, 'train/grad_norm_pre_clip': 21.310882568359375, 'timing/data': 0.0003355918452143669, 'timing/model': 1.0902380279731005, 'learning_rate/qwen_vl_interface': 6.812050252858329e-07, 'learning_rate/action_model': 6.812050252858329e-06, 'epoch': 0.35}) +[2026-06-24 14:26:35,207][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.027565453201532364, 'train/grad_norm_pre_clip': 7.695407390594482, 'timing/data': 0.00035267393104732037, 'timing/model': 1.0870224831160158, 'learning_rate/qwen_vl_interface': 6.801597401410339e-07, 'learning_rate/action_model': 6.801597401410338e-06, 'epoch': 0.35}) +[2026-06-24 14:26:36,497][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.09995684772729874, 'train/grad_norm_pre_clip': 97.53208923339844, 'timing/data': 0.0003205330576747656, 'timing/model': 1.287732084048912, 'learning_rate/qwen_vl_interface': 6.791174203159645e-07, 'learning_rate/action_model': 6.7911742031596445e-06, 'epoch': 0.35}) +[2026-06-24 14:26:37,588][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.005074820481240749, 'train/grad_norm_pre_clip': 8.313789367675781, 'timing/data': 0.0003207738045603037, 'timing/model': 1.0902392868883908, 'learning_rate/qwen_vl_interface': 6.780780664869766e-07, 'learning_rate/action_model': 6.780780664869765e-06, 'epoch': 0.35}) +[2026-06-24 14:26:38,679][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.18254342675209045, 'train/grad_norm_pre_clip': 26.448326110839844, 'timing/data': 0.00035775313153862953, 'timing/model': 1.0890414989553392, 'learning_rate/qwen_vl_interface': 6.770416793284942e-07, 'learning_rate/action_model': 6.7704167932849426e-06, 'epoch': 0.35}) +[2026-06-24 14:26:39,768][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.0406675785779953, 'train/grad_norm_pre_clip': 52.76594543457031, 'timing/data': 0.00032643391750752926, 'timing/model': 1.0874944040551782, 'learning_rate/qwen_vl_interface': 6.760082595130184e-07, 'learning_rate/action_model': 6.760082595130183e-06, 'epoch': 0.35}) +[2026-06-24 14:26:40,855][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.010501121170818806, 'train/grad_norm_pre_clip': 16.329980850219727, 'timing/data': 0.0003217048943042755, 'timing/model': 1.0853735979180783, 'learning_rate/qwen_vl_interface': 6.749778077111241e-07, 'learning_rate/action_model': 6.749778077111241e-06, 'epoch': 0.35}) +[2026-06-24 14:26:41,945][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.0008690443355590105, 'train/grad_norm_pre_clip': 0.6461020708084106, 'timing/data': 0.00036304304376244545, 'timing/model': 1.0884825440589339, 'learning_rate/qwen_vl_interface': 6.739503245914602e-07, 'learning_rate/action_model': 6.739503245914601e-06, 'epoch': 0.35}) +[2026-06-24 14:26:43,036][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.003425456117838621, 'train/grad_norm_pre_clip': 3.6779987812042236, 'timing/data': 0.0003490569069981575, 'timing/model': 1.0891812040936202, 'learning_rate/qwen_vl_interface': 6.729258108207484e-07, 'learning_rate/action_model': 6.729258108207484e-06, 'epoch': 0.35}) +[2026-06-24 14:26:44,129][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.18853016197681427, 'train/grad_norm_pre_clip': 66.34050750732422, 'timing/data': 0.00033008097670972347, 'timing/model': 1.0909255619626492, 'learning_rate/qwen_vl_interface': 6.719042670637857e-07, 'learning_rate/action_model': 6.719042670637857e-06, 'epoch': 0.35}) +[2026-06-24 14:26:45,219][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.008590271696448326, 'train/grad_norm_pre_clip': 14.055998802185059, 'timing/data': 0.00032200594432651997, 'timing/model': 1.0889115980826318, 'learning_rate/qwen_vl_interface': 6.708856939834402e-07, 'learning_rate/action_model': 6.708856939834401e-06, 'epoch': 0.35}) +[2026-06-24 14:26:46,314][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.015104953199625015, 'train/grad_norm_pre_clip': 15.515380859375, 'timing/data': 0.0003348099999129772, 'timing/model': 1.092850225046277, 'learning_rate/qwen_vl_interface': 6.69870092240654e-07, 'learning_rate/action_model': 6.698700922406539e-06, 'epoch': 0.35}) +[2026-06-24 14:26:47,405][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.17775702476501465, 'train/grad_norm_pre_clip': 66.22431945800781, 'timing/data': 0.00032254704274237156, 'timing/model': 1.0892328941263258, 'learning_rate/qwen_vl_interface': 6.688574624944386e-07, 'learning_rate/action_model': 6.688574624944385e-06, 'epoch': 0.35}) +[2026-06-24 14:26:48,496][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.019541753455996513, 'train/grad_norm_pre_clip': 10.55790901184082, 'timing/data': 0.0004065651446580887, 'timing/model': 1.0894317738711834, 'learning_rate/qwen_vl_interface': 6.678478054018788e-07, 'learning_rate/action_model': 6.678478054018787e-06, 'epoch': 0.35}) +[2026-06-24 14:26:49,586][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.08308440446853638, 'train/grad_norm_pre_clip': 30.396747589111328, 'timing/data': 0.00033384794369339943, 'timing/model': 1.088736753212288, 'learning_rate/qwen_vl_interface': 6.668411216181312e-07, 'learning_rate/action_model': 6.668411216181311e-06, 'epoch': 0.36}) +[2026-06-24 14:26:50,677][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.1279902160167694, 'train/grad_norm_pre_clip': 30.865612030029297, 'timing/data': 0.0003253519535064697, 'timing/model': 1.0892920140177011, 'learning_rate/qwen_vl_interface': 6.658374117964198e-07, 'learning_rate/action_model': 6.658374117964198e-06, 'epoch': 0.36}) +[2026-06-24 14:26:51,768][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.007685288321226835, 'train/grad_norm_pre_clip': 5.443531036376953, 'timing/data': 0.00034192297607660294, 'timing/model': 1.0884635199327022, 'learning_rate/qwen_vl_interface': 6.648366765880437e-07, 'learning_rate/action_model': 6.648366765880436e-06, 'epoch': 0.36}) +[2026-06-24 14:26:52,860][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.034214869141578674, 'train/grad_norm_pre_clip': 34.9251594543457, 'timing/data': 0.00038868095725774765, 'timing/model': 1.0911555110942572, 'learning_rate/qwen_vl_interface': 6.638389166423672e-07, 'learning_rate/action_model': 6.638389166423671e-06, 'epoch': 0.36}) +[2026-06-24 14:26:54,154][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.11721788346767426, 'train/grad_norm_pre_clip': 49.666202545166016, 'timing/data': 0.00034873606637120247, 'timing/model': 1.2916819751262665, 'learning_rate/qwen_vl_interface': 6.628441326068262e-07, 'learning_rate/action_model': 6.628441326068262e-06, 'epoch': 0.36}) +[2026-06-24 14:26:55,242][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.012787677347660065, 'train/grad_norm_pre_clip': 13.4116792678833, 'timing/data': 0.0003237389028072357, 'timing/model': 1.0867334720678627, 'learning_rate/qwen_vl_interface': 6.61852325126926e-07, 'learning_rate/action_model': 6.61852325126926e-06, 'epoch': 0.36}) +[2026-06-24 14:26:56,330][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.10407527536153793, 'train/grad_norm_pre_clip': 72.3382797241211, 'timing/data': 0.00031597400084137917, 'timing/model': 1.0862695439718664, 'learning_rate/qwen_vl_interface': 6.608634948462388e-07, 'learning_rate/action_model': 6.6086349484623875e-06, 'epoch': 0.36}) +[2026-06-24 14:26:57,424][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.09335483610630035, 'train/grad_norm_pre_clip': 38.82255172729492, 'timing/data': 0.00033269613049924374, 'timing/model': 1.0928463449236006, 'learning_rate/qwen_vl_interface': 6.598776424064063e-07, 'learning_rate/action_model': 6.598776424064062e-06, 'epoch': 0.36}) +[2026-06-24 14:26:58,515][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.0009930592495948076, 'train/grad_norm_pre_clip': 0.47590503096580505, 'timing/data': 0.00033600209280848503, 'timing/model': 1.0893833448644727, 'learning_rate/qwen_vl_interface': 6.588947684471376e-07, 'learning_rate/action_model': 6.588947684471377e-06, 'epoch': 0.36}) +[2026-06-24 14:26:59,606][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.04207826033234596, 'train/grad_norm_pre_clip': 73.50680541992188, 'timing/data': 0.0003853249363601208, 'timing/model': 1.0887177779804915, 'learning_rate/qwen_vl_interface': 6.579148736062087e-07, 'learning_rate/action_model': 6.579148736062087e-06, 'epoch': 0.36}) +[2026-06-24 14:27:00,701][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.2206658571958542, 'train/grad_norm_pre_clip': 41.70348358154297, 'timing/data': 0.00036050891503691673, 'timing/model': 1.0939063869882375, 'learning_rate/qwen_vl_interface': 6.569379585194629e-07, 'learning_rate/action_model': 6.56937958519463e-06, 'epoch': 0.36}) +[2026-06-24 14:27:01,792][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.10814595967531204, 'train/grad_norm_pre_clip': 103.3483657836914, 'timing/data': 0.00034065102227032185, 'timing/model': 1.0894018670078367, 'learning_rate/qwen_vl_interface': 6.559640238208092e-07, 'learning_rate/action_model': 6.559640238208093e-06, 'epoch': 0.36}) +[2026-06-24 14:27:02,883][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.003108715871348977, 'train/grad_norm_pre_clip': 2.032191276550293, 'timing/data': 0.000316165154799819, 'timing/model': 1.0891851659398526, 'learning_rate/qwen_vl_interface': 6.549930701422247e-07, 'learning_rate/action_model': 6.549930701422247e-06, 'epoch': 0.36}) +[2026-06-24 14:27:03,976][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.08227197080850601, 'train/grad_norm_pre_clip': 25.353601455688477, 'timing/data': 0.0003639049828052521, 'timing/model': 1.0909799090586603, 'learning_rate/qwen_vl_interface': 6.540250981137486e-07, 'learning_rate/action_model': 6.540250981137486e-06, 'epoch': 0.36}) +[2026-06-24 14:27:05,067][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.0983385443687439, 'train/grad_norm_pre_clip': 54.79710006713867, 'timing/data': 0.00033867708407342434, 'timing/model': 1.0898355210665613, 'learning_rate/qwen_vl_interface': 6.530601083634889e-07, 'learning_rate/action_model': 6.53060108363489e-06, 'epoch': 0.36}) +[2026-06-24 14:27:06,163][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.0010188844753429294, 'train/grad_norm_pre_clip': 0.6053597331047058, 'timing/data': 0.0003555179573595524, 'timing/model': 1.0942936441861093, 'learning_rate/qwen_vl_interface': 6.52098101517617e-07, 'learning_rate/action_model': 6.5209810151761695e-06, 'epoch': 0.36}) +[2026-06-24 14:27:07,257][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.005979249253869057, 'train/grad_norm_pre_clip': 17.491024017333984, 'timing/data': 0.0003567610401660204, 'timing/model': 1.0924635280389339, 'learning_rate/qwen_vl_interface': 6.511390782003676e-07, 'learning_rate/action_model': 6.511390782003676e-06, 'epoch': 0.36}) +[2026-06-24 14:27:08,352][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.037385545670986176, 'train/grad_norm_pre_clip': 26.239883422851562, 'timing/data': 0.0003451600205153227, 'timing/model': 1.0931782280094922, 'learning_rate/qwen_vl_interface': 6.501830390340416e-07, 'learning_rate/action_model': 6.501830390340416e-06, 'epoch': 0.36}) +[2026-06-24 14:27:09,445][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.018277309834957123, 'train/grad_norm_pre_clip': 10.052837371826172, 'timing/data': 0.0003708580043166876, 'timing/model': 1.0913996370509267, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 0.36}) +[2026-06-24 14:27:10,539][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.038027986884117126, 'train/grad_norm_pre_clip': 14.561882972717285, 'timing/data': 0.0003422340378165245, 'timing/model': 1.0922183338552713, 'learning_rate/qwen_vl_interface': 6.482799156336761e-07, 'learning_rate/action_model': 6.482799156336761e-06, 'epoch': 0.36}) +[2026-06-24 14:27:11,838][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.055939026176929474, 'train/grad_norm_pre_clip': 116.86030578613281, 'timing/data': 0.00031585502438247204, 'timing/model': 1.2973848371766508, 'learning_rate/qwen_vl_interface': 6.473328326345524e-07, 'learning_rate/action_model': 6.473328326345523e-06, 'epoch': 0.36}) +[2026-06-24 14:27:12,927][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.08542975783348083, 'train/grad_norm_pre_clip': 57.72322463989258, 'timing/data': 0.00045621790923178196, 'timing/model': 1.0876952051185071, 'learning_rate/qwen_vl_interface': 6.463887362561827e-07, 'learning_rate/action_model': 6.4638873625618265e-06, 'epoch': 0.36}) +[2026-06-24 14:27:14,019][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.036946624517440796, 'train/grad_norm_pre_clip': 45.17980194091797, 'timing/data': 0.00032355915755033493, 'timing/model': 1.0901474740821868, 'learning_rate/qwen_vl_interface': 6.454476271111824e-07, 'learning_rate/action_model': 6.4544762711118225e-06, 'epoch': 0.36}) +[2026-06-24 14:27:15,111][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.044001419097185135, 'train/grad_norm_pre_clip': 148.00436401367188, 'timing/data': 0.00035496801137924194, 'timing/model': 1.0900185380596668, 'learning_rate/qwen_vl_interface': 6.445095058102248e-07, 'learning_rate/action_model': 6.445095058102248e-06, 'epoch': 0.36}) +[2026-06-24 14:27:16,205][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.004078173544257879, 'train/grad_norm_pre_clip': 5.312100887298584, 'timing/data': 0.00032648397609591484, 'timing/model': 1.092814896022901, 'learning_rate/qwen_vl_interface': 6.435743729620484e-07, 'learning_rate/action_model': 6.4357437296204844e-06, 'epoch': 0.36}) +[2026-06-24 14:27:17,299][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.002580849453806877, 'train/grad_norm_pre_clip': 2.5453574657440186, 'timing/data': 0.0004120559897273779, 'timing/model': 1.09152418281883, 'learning_rate/qwen_vl_interface': 6.426422291734498e-07, 'learning_rate/action_model': 6.4264222917344975e-06, 'epoch': 0.36}) +[2026-06-24 14:27:18,391][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.005106442142277956, 'train/grad_norm_pre_clip': 6.742685317993164, 'timing/data': 0.0003596870228648186, 'timing/model': 1.0910488639492542, 'learning_rate/qwen_vl_interface': 6.417130750492875e-07, 'learning_rate/action_model': 6.417130750492875e-06, 'epoch': 0.36}) +[2026-06-24 14:27:19,485][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.01171824149787426, 'train/grad_norm_pre_clip': 9.531929969787598, 'timing/data': 0.0003256720956414938, 'timing/model': 1.0920401450712234, 'learning_rate/qwen_vl_interface': 6.407869111924799e-07, 'learning_rate/action_model': 6.4078691119247986e-06, 'epoch': 0.36}) +[2026-06-24 14:27:20,578][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.06603021174669266, 'train/grad_norm_pre_clip': 53.2988395690918, 'timing/data': 0.0003404810559004545, 'timing/model': 1.0909605009946972, 'learning_rate/qwen_vl_interface': 6.398637382040035e-07, 'learning_rate/action_model': 6.3986373820400345e-06, 'epoch': 0.36}) +[2026-06-24 14:27:21,671][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.07175085693597794, 'train/grad_norm_pre_clip': 72.27853393554688, 'timing/data': 0.00033579207956790924, 'timing/model': 1.091378642944619, 'learning_rate/qwen_vl_interface': 6.389435566828961e-07, 'learning_rate/action_model': 6.3894355668289605e-06, 'epoch': 0.36}) +[2026-06-24 14:27:22,763][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.006561198737472296, 'train/grad_norm_pre_clip': 4.298384189605713, 'timing/data': 0.00037042703479528427, 'timing/model': 1.0902866441756487, 'learning_rate/qwen_vl_interface': 6.38026367226253e-07, 'learning_rate/action_model': 6.3802636722625295e-06, 'epoch': 0.36}) +[2026-06-24 14:27:23,860][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.01777438074350357, 'train/grad_norm_pre_clip': 10.805477142333984, 'timing/data': 0.00041384901851415634, 'timing/model': 1.0959853529930115, 'learning_rate/qwen_vl_interface': 6.371121704292292e-07, 'learning_rate/action_model': 6.371121704292292e-06, 'epoch': 0.36}) +[2026-06-24 14:27:24,950][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.017590513452887535, 'train/grad_norm_pre_clip': 10.254796028137207, 'timing/data': 0.00031092483550310135, 'timing/model': 1.0883060100022703, 'learning_rate/qwen_vl_interface': 6.362009668850358e-07, 'learning_rate/action_model': 6.362009668850358e-06, 'epoch': 0.36}) +[2026-06-24 14:27:26,041][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.12153670936822891, 'train/grad_norm_pre_clip': 109.65620422363281, 'timing/data': 0.0003628721460700035, 'timing/model': 1.0890512780752033, 'learning_rate/qwen_vl_interface': 6.352927571849443e-07, 'learning_rate/action_model': 6.352927571849443e-06, 'epoch': 0.36}) +[2026-06-24 14:27:27,140][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.014748279005289078, 'train/grad_norm_pre_clip': 9.659645080566406, 'timing/data': 0.00034497911110520363, 'timing/model': 1.097947641974315, 'learning_rate/qwen_vl_interface': 6.343875419182818e-07, 'learning_rate/action_model': 6.343875419182818e-06, 'epoch': 0.36}) +[2026-06-24 14:27:28,232][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.0009503148030489683, 'train/grad_norm_pre_clip': 1.614279866218567, 'timing/data': 0.00035039917565882206, 'timing/model': 1.090089282952249, 'learning_rate/qwen_vl_interface': 6.33485321672431e-07, 'learning_rate/action_model': 6.3348532167243095e-06, 'epoch': 0.36}) +[2026-06-24 14:27:29,593][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.03998783230781555, 'train/grad_norm_pre_clip': 36.88553237915039, 'timing/data': 0.0003274159971624613, 'timing/model': 1.3592844689264894, 'learning_rate/qwen_vl_interface': 6.325860970328357e-07, 'learning_rate/action_model': 6.325860970328357e-06, 'epoch': 0.36}) +[2026-06-24 14:27:30,686][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.010443118400871754, 'train/grad_norm_pre_clip': 20.18770980834961, 'timing/data': 0.0003922779578715563, 'timing/model': 1.0909970230422914, 'learning_rate/qwen_vl_interface': 6.316898685829913e-07, 'learning_rate/action_model': 6.3168986858299125e-06, 'epoch': 0.36}) +[2026-06-24 14:27:31,782][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.010152989067137241, 'train/grad_norm_pre_clip': 16.353551864624023, 'timing/data': 0.0003340989351272583, 'timing/model': 1.0943112839013338, 'learning_rate/qwen_vl_interface': 6.307966369044497e-07, 'learning_rate/action_model': 6.307966369044498e-06, 'epoch': 0.36}) +[2026-06-24 14:27:32,876][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.13785018026828766, 'train/grad_norm_pre_clip': 57.5894775390625, 'timing/data': 0.00032804813235998154, 'timing/model': 1.0926120430231094, 'learning_rate/qwen_vl_interface': 6.299064025768217e-07, 'learning_rate/action_model': 6.299064025768217e-06, 'epoch': 0.36}) +[2026-06-24 14:27:33,967][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.15978464484214783, 'train/grad_norm_pre_clip': 64.01628112792969, 'timing/data': 0.00031968182884156704, 'timing/model': 1.0895772299263626, 'learning_rate/qwen_vl_interface': 6.290191661777689e-07, 'learning_rate/action_model': 6.290191661777689e-06, 'epoch': 0.36}) +[2026-06-24 14:27:35,062][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.15530192852020264, 'train/grad_norm_pre_clip': 74.24800872802734, 'timing/data': 0.0003318639937788248, 'timing/model': 1.0926676199305803, 'learning_rate/qwen_vl_interface': 6.281349282830097e-07, 'learning_rate/action_model': 6.2813492828300975e-06, 'epoch': 0.36}) +[2026-06-24 14:27:36,153][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.12935128808021545, 'train/grad_norm_pre_clip': 44.01752471923828, 'timing/data': 0.0003800150007009506, 'timing/model': 1.0900201650802046, 'learning_rate/qwen_vl_interface': 6.272536894663174e-07, 'learning_rate/action_model': 6.272536894663174e-06, 'epoch': 0.36}) +[2026-06-24 14:27:37,248][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.028493639081716537, 'train/grad_norm_pre_clip': 28.736719131469727, 'timing/data': 0.0003379459958523512, 'timing/model': 1.093041054904461, 'learning_rate/qwen_vl_interface': 6.26375450299517e-07, 'learning_rate/action_model': 6.26375450299517e-06, 'epoch': 0.36}) +[2026-06-24 14:27:38,338][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.029242737218737602, 'train/grad_norm_pre_clip': 68.13395690917969, 'timing/data': 0.0003737630322575569, 'timing/model': 1.088781536789611, 'learning_rate/qwen_vl_interface': 6.255002113524885e-07, 'learning_rate/action_model': 6.255002113524886e-06, 'epoch': 0.36}) +[2026-06-24 14:27:39,429][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.010718610137701035, 'train/grad_norm_pre_clip': 14.766557693481445, 'timing/data': 0.0004062638618052006, 'timing/model': 1.0883454128634185, 'learning_rate/qwen_vl_interface': 6.246279731931659e-07, 'learning_rate/action_model': 6.246279731931659e-06, 'epoch': 0.36}) +[2026-06-24 14:27:40,522][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.013795856386423111, 'train/grad_norm_pre_clip': 14.748034477233887, 'timing/data': 0.00034372718073427677, 'timing/model': 1.0916995098814368, 'learning_rate/qwen_vl_interface': 6.237587363875348e-07, 'learning_rate/action_model': 6.237587363875348e-06, 'epoch': 0.36}) +[2026-06-24 14:27:41,612][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.2281641662120819, 'train/grad_norm_pre_clip': 82.33771514892578, 'timing/data': 0.00033355806954205036, 'timing/model': 1.0889153068419546, 'learning_rate/qwen_vl_interface': 6.22892501499633e-07, 'learning_rate/action_model': 6.228925014996329e-06, 'epoch': 0.36}) +[2026-06-24 14:27:42,705][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.06335397809743881, 'train/grad_norm_pre_clip': 48.68357467651367, 'timing/data': 0.0003445078618824482, 'timing/model': 1.0910041029565036, 'learning_rate/qwen_vl_interface': 6.220292690915513e-07, 'learning_rate/action_model': 6.220292690915512e-06, 'epoch': 0.36}) +[2026-06-24 14:27:43,798][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.0952102392911911, 'train/grad_norm_pre_clip': 57.52238464355469, 'timing/data': 0.0003471330273896456, 'timing/model': 1.0918546097818762, 'learning_rate/qwen_vl_interface': 6.211690397234322e-07, 'learning_rate/action_model': 6.211690397234322e-06, 'epoch': 0.36}) +[2026-06-24 14:27:44,891][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.35482150316238403, 'train/grad_norm_pre_clip': 61.533512115478516, 'timing/data': 0.00040114508010447025, 'timing/model': 1.0912258131429553, 'learning_rate/qwen_vl_interface': 6.20311813953468e-07, 'learning_rate/action_model': 6.203118139534681e-06, 'epoch': 0.36}) +[2026-06-24 14:27:46,208][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.07200223207473755, 'train/grad_norm_pre_clip': 55.78650665283203, 'timing/data': 0.00034511019475758076, 'timing/model': 1.3147378880530596, 'learning_rate/qwen_vl_interface': 6.194575923379037e-07, 'learning_rate/action_model': 6.194575923379037e-06, 'epoch': 0.36}) +[2026-06-24 14:27:47,299][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.009583861567080021, 'train/grad_norm_pre_clip': 11.440083503723145, 'timing/data': 0.00036266189999878407, 'timing/model': 1.089638948906213, 'learning_rate/qwen_vl_interface': 6.186063754310352e-07, 'learning_rate/action_model': 6.1860637543103516e-06, 'epoch': 0.36}) +[2026-06-24 14:27:48,393][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.3204512894153595, 'train/grad_norm_pre_clip': 63.71697998046875, 'timing/data': 0.0003596560563892126, 'timing/model': 1.092157453065738, 'learning_rate/qwen_vl_interface': 6.177581637852075e-07, 'learning_rate/action_model': 6.177581637852075e-06, 'epoch': 0.36}) +[2026-06-24 14:27:49,485][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.03472842648625374, 'train/grad_norm_pre_clip': 29.734588623046875, 'timing/data': 0.00032717594876885414, 'timing/model': 1.0907856489066035, 'learning_rate/qwen_vl_interface': 6.169129579508154e-07, 'learning_rate/action_model': 6.1691295795081535e-06, 'epoch': 0.36}) +[2026-06-24 14:27:50,576][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.03173745051026344, 'train/grad_norm_pre_clip': 103.63423156738281, 'timing/data': 0.0003005450125783682, 'timing/model': 1.0897580720484257, 'learning_rate/qwen_vl_interface': 6.160707584763042e-07, 'learning_rate/action_model': 6.160707584763042e-06, 'epoch': 0.36}) +[2026-06-24 14:27:51,664][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.2334223985671997, 'train/grad_norm_pre_clip': 41.551300048828125, 'timing/data': 0.0003177579492330551, 'timing/model': 1.0856599481776357, 'learning_rate/qwen_vl_interface': 6.152315659081681e-07, 'learning_rate/action_model': 6.152315659081681e-06, 'epoch': 0.36}) +[2026-06-24 14:27:52,757][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.005398621782660484, 'train/grad_norm_pre_clip': 9.808211326599121, 'timing/data': 0.00033247610554099083, 'timing/model': 1.092028986895457, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 0.36}) +[2026-06-24 14:27:53,852][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.008710507303476334, 'train/grad_norm_pre_clip': 9.348533630371094, 'timing/data': 0.00031816912814974785, 'timing/model': 1.0933126381132752, 'learning_rate/qwen_vl_interface': 6.135622036672411e-07, 'learning_rate/action_model': 6.135622036672411e-06, 'epoch': 0.36}) +[2026-06-24 14:27:54,944][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.061320628970861435, 'train/grad_norm_pre_clip': 34.1015625, 'timing/data': 0.0003253021277487278, 'timing/model': 1.0900133980903774, 'learning_rate/qwen_vl_interface': 6.127320350776815e-07, 'learning_rate/action_model': 6.127320350776815e-06, 'epoch': 0.36}) +[2026-06-24 14:27:56,036][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.00022193664335645735, 'train/grad_norm_pre_clip': 0.19670447707176208, 'timing/data': 0.0003288791049271822, 'timing/model': 1.0906107779592276, 'learning_rate/qwen_vl_interface': 6.119048755609574e-07, 'learning_rate/action_model': 6.119048755609574e-06, 'epoch': 0.36}) +[2026-06-24 14:27:57,123][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.02391425333917141, 'train/grad_norm_pre_clip': 42.7082633972168, 'timing/data': 0.0003186790272593498, 'timing/model': 1.0847305210772902, 'learning_rate/qwen_vl_interface': 6.110807256538055e-07, 'learning_rate/action_model': 6.110807256538055e-06, 'epoch': 0.36}) +[2026-06-24 14:27:58,214][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.03171014413237572, 'train/grad_norm_pre_clip': 21.596263885498047, 'timing/data': 0.0003193409647792578, 'timing/model': 1.0901849288493395, 'learning_rate/qwen_vl_interface': 6.102595858910059e-07, 'learning_rate/action_model': 6.102595858910058e-06, 'epoch': 0.36}) +[2026-06-24 14:27:59,315][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.20568382740020752, 'train/grad_norm_pre_clip': 109.26472473144531, 'timing/data': 0.0003277070354670286, 'timing/model': 1.0986678949557245, 'learning_rate/qwen_vl_interface': 6.094414568053888e-07, 'learning_rate/action_model': 6.094414568053887e-06, 'epoch': 0.36}) +[2026-06-24 14:28:00,402][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.053540173918008804, 'train/grad_norm_pre_clip': 21.302234649658203, 'timing/data': 0.00031489296816289425, 'timing/model': 1.08579392102547, 'learning_rate/qwen_vl_interface': 6.086263389278282e-07, 'learning_rate/action_model': 6.086263389278281e-06, 'epoch': 0.36}) +[2026-06-24 14:28:01,494][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.05739232897758484, 'train/grad_norm_pre_clip': 48.961090087890625, 'timing/data': 0.0003631729632616043, 'timing/model': 1.0897573030088097, 'learning_rate/qwen_vl_interface': 6.078142327872453e-07, 'learning_rate/action_model': 6.078142327872453e-06, 'epoch': 0.36}) +[2026-06-24 14:28:02,588][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.3463420271873474, 'train/grad_norm_pre_clip': 65.50530242919922, 'timing/data': 0.000315043143928051, 'timing/model': 1.0932734948582947, 'learning_rate/qwen_vl_interface': 6.070051389106073e-07, 'learning_rate/action_model': 6.070051389106073e-06, 'epoch': 0.36}) +[2026-06-24 14:28:03,982][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.0800275132060051, 'train/grad_norm_pre_clip': 36.45103073120117, 'timing/data': 0.0003153339494019747, 'timing/model': 1.3914870941080153, 'learning_rate/qwen_vl_interface': 6.061990578229268e-07, 'learning_rate/action_model': 6.061990578229267e-06, 'epoch': 0.36}) +[2026-06-24 14:28:05,073][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.008948728442192078, 'train/grad_norm_pre_clip': 11.151161193847656, 'timing/data': 0.00031216698698699474, 'timing/model': 1.0898943899665028, 'learning_rate/qwen_vl_interface': 6.053959900472609e-07, 'learning_rate/action_model': 6.053959900472609e-06, 'epoch': 0.36}) +[2026-06-24 14:28:06,163][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.03691012039780617, 'train/grad_norm_pre_clip': 24.516870498657227, 'timing/data': 0.00034697307273745537, 'timing/model': 1.0884947429876775, 'learning_rate/qwen_vl_interface': 6.045959361047107e-07, 'learning_rate/action_model': 6.045959361047107e-06, 'epoch': 0.36}) +[2026-06-24 14:28:07,256][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.007714346516877413, 'train/grad_norm_pre_clip': 7.506587505340576, 'timing/data': 0.0003184899687767029, 'timing/model': 1.0910584519151598, 'learning_rate/qwen_vl_interface': 6.03798896514423e-07, 'learning_rate/action_model': 6.0379889651442295e-06, 'epoch': 0.36}) +[2026-06-24 14:28:08,348][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.02212417498230934, 'train/grad_norm_pre_clip': 8.24414348602295, 'timing/data': 0.00036297296173870564, 'timing/model': 1.0901790740899742, 'learning_rate/qwen_vl_interface': 6.030048717935892e-07, 'learning_rate/action_model': 6.030048717935892e-06, 'epoch': 0.36}) +[2026-06-24 14:28:09,438][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.09214524924755096, 'train/grad_norm_pre_clip': 44.20424270629883, 'timing/data': 0.0003475239500403404, 'timing/model': 1.0881756520830095, 'learning_rate/qwen_vl_interface': 6.022138624574414e-07, 'learning_rate/action_model': 6.0221386245744135e-06, 'epoch': 0.36}) +[2026-06-24 14:28:10,527][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.004999982193112373, 'train/grad_norm_pre_clip': 9.750351905822754, 'timing/data': 0.000339738791808486, 'timing/model': 1.0873483419418335, 'learning_rate/qwen_vl_interface': 6.014258690192585e-07, 'learning_rate/action_model': 6.014258690192584e-06, 'epoch': 0.36}) +[2026-06-24 14:28:11,619][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.06284179538488388, 'train/grad_norm_pre_clip': 78.28145599365234, 'timing/data': 0.0003690039739012718, 'timing/model': 1.091235099826008, 'learning_rate/qwen_vl_interface': 6.006408919903601e-07, 'learning_rate/action_model': 6.006408919903602e-06, 'epoch': 0.36}) +[2026-06-24 14:28:12,710][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.06302877515554428, 'train/grad_norm_pre_clip': 44.24775695800781, 'timing/data': 0.00033588195219635963, 'timing/model': 1.0890787488315254, 'learning_rate/qwen_vl_interface': 5.998589318801099e-07, 'learning_rate/action_model': 5.998589318801098e-06, 'epoch': 0.36}) +[2026-06-24 14:28:13,803][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.07390574365854263, 'train/grad_norm_pre_clip': 54.340675354003906, 'timing/data': 0.00034696212969720364, 'timing/model': 1.090789885027334, 'learning_rate/qwen_vl_interface': 5.990799891959124e-07, 'learning_rate/action_model': 5.990799891959124e-06, 'epoch': 0.36}) +[2026-06-24 14:28:14,890][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.028734365478157997, 'train/grad_norm_pre_clip': 35.27373123168945, 'timing/data': 0.0003825700841844082, 'timing/model': 1.086234371876344, 'learning_rate/qwen_vl_interface': 5.98304064443216e-07, 'learning_rate/action_model': 5.98304064443216e-06, 'epoch': 0.36}) +[2026-06-24 14:28:15,984][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.09562904387712479, 'train/grad_norm_pre_clip': 77.27667236328125, 'timing/data': 0.0003411120269447565, 'timing/model': 1.092107166070491, 'learning_rate/qwen_vl_interface': 5.975311581255093e-07, 'learning_rate/action_model': 5.975311581255092e-06, 'epoch': 0.36}) +[2026-06-24 14:28:17,073][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.33431288599967957, 'train/grad_norm_pre_clip': 137.40248107910156, 'timing/data': 0.00039021391421556473, 'timing/model': 1.0866835000924766, 'learning_rate/qwen_vl_interface': 5.967612707443225e-07, 'learning_rate/action_model': 5.967612707443225e-06, 'epoch': 0.36}) +[2026-06-24 14:28:18,159][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3750 +[2026-06-24 14:28:42,911][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3750 +[2026-06-24 14:28:42,912][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3750 +[2026-06-24 14:29:16,588][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3750 +[2026-06-24 14:39:01,250][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3750 +[2026-06-24 14:39:01,251][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.01816880702972412, 'train/grad_norm_pre_clip': 63.04476547241211, 'eval/action_loss/seconds': 24.74997561587952, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.06265363097190857, 'eval/latency_2/loss': 0.06265363097190857, 'eval/flappy/loss': 0.06265363097190857, 'eval/flappy/latency_2/loss': 0.06265363097190857, 'eval/flap_precision': 0.9399477806788512, 'eval/flap_recall': 0.9769335142469471, 'eval/flap_f1': 0.9580838323353293, 'eval/flap_rate_model': 0.2127777777777778, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9580838323353293, 'eval/latency_2/flap_precision': 0.9399477806788512, 'eval/latency_2/flap_recall': 0.9769335142469471, 'eval/latency_2/flap_f1': 0.9580838323353293, 'eval/latency_2/flap_rate_model': 0.2127777777777778, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9580838323353293, 'eval/action_classification/data_wait_seconds': 0.3485524654388428, 'eval/action_classification/predict_seconds': 33.229164123535156, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003467120695859194, 'timing/model': 1.0857363031245768, 'learning_rate/qwen_vl_interface': 5.959944027992285e-07, 'learning_rate/action_model': 5.959944027992286e-06, 'epoch': 0.36}) +[2026-06-24 14:39:02,322][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.3987874984741211, 'train/grad_norm_pre_clip': 44.46015548706055, 'timing/data': 0.0003810559865087271, 'timing/model': 1.069245332153514, 'learning_rate/qwen_vl_interface': 5.952305547878388e-07, 'learning_rate/action_model': 5.952305547878387e-06, 'epoch': 0.36}) +[2026-06-24 14:39:03,392][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.016154035925865173, 'train/grad_norm_pre_clip': 15.218888282775879, 'timing/data': 0.00034898589365184307, 'timing/model': 1.0691738871391863, 'learning_rate/qwen_vl_interface': 5.944697272058045e-07, 'learning_rate/action_model': 5.944697272058045e-06, 'epoch': 0.36}) +[2026-06-24 14:39:04,466][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.07568711787462234, 'train/grad_norm_pre_clip': 36.99167251586914, 'timing/data': 0.0003481449093669653, 'timing/model': 1.0714429719373584, 'learning_rate/qwen_vl_interface': 5.937119205468217e-07, 'learning_rate/action_model': 5.9371192054682165e-06, 'epoch': 0.36}) +[2026-06-24 14:39:05,540][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.23387102782726288, 'train/grad_norm_pre_clip': 66.77410125732422, 'timing/data': 0.0003177470061928034, 'timing/model': 1.0726855469401926, 'learning_rate/qwen_vl_interface': 5.9295713530262e-07, 'learning_rate/action_model': 5.9295713530262e-06, 'epoch': 0.36}) +[2026-06-24 14:39:06,611][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.01316246297210455, 'train/grad_norm_pre_clip': 12.997222900390625, 'timing/data': 0.00031704618595540524, 'timing/model': 1.0699005599599332, 'learning_rate/qwen_vl_interface': 5.922053719629726e-07, 'learning_rate/action_model': 5.922053719629726e-06, 'epoch': 0.36}) +[2026-06-24 14:39:07,686][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.0013522026129066944, 'train/grad_norm_pre_clip': 1.9461256265640259, 'timing/data': 0.00033369800075888634, 'timing/model': 1.073138396954164, 'learning_rate/qwen_vl_interface': 5.914566310156903e-07, 'learning_rate/action_model': 5.914566310156903e-06, 'epoch': 0.36}) +[2026-06-24 14:39:08,763][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.20057693123817444, 'train/grad_norm_pre_clip': 60.08091735839844, 'timing/data': 0.00034655211493372917, 'timing/model': 1.0749325789511204, 'learning_rate/qwen_vl_interface': 5.907109129466231e-07, 'learning_rate/action_model': 5.90710912946623e-06, 'epoch': 0.36}) +[2026-06-24 14:39:09,839][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.15772366523742676, 'train/grad_norm_pre_clip': 81.86656951904297, 'timing/data': 0.0003398689441382885, 'timing/model': 1.075142052024603, 'learning_rate/qwen_vl_interface': 5.899682182396596e-07, 'learning_rate/action_model': 5.899682182396596e-06, 'epoch': 0.36}) +[2026-06-24 14:39:10,917][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.28405964374542236, 'train/grad_norm_pre_clip': 75.40983581542969, 'timing/data': 0.0003453199751675129, 'timing/model': 1.0753650600090623, 'learning_rate/qwen_vl_interface': 5.892285473767255e-07, 'learning_rate/action_model': 5.892285473767256e-06, 'epoch': 0.36}) +[2026-06-24 14:39:11,993][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.11182969063520432, 'train/grad_norm_pre_clip': 55.32981872558594, 'timing/data': 0.0003843919839709997, 'timing/model': 1.0745187278371304, 'learning_rate/qwen_vl_interface': 5.884919008377862e-07, 'learning_rate/action_model': 5.884919008377861e-06, 'epoch': 0.36}) +[2026-06-24 14:39:13,071][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.009054294787347317, 'train/grad_norm_pre_clip': 7.606435298919678, 'timing/data': 0.0003216150216758251, 'timing/model': 1.0766628189012408, 'learning_rate/qwen_vl_interface': 5.877582791008432e-07, 'learning_rate/action_model': 5.877582791008432e-06, 'epoch': 0.36}) +[2026-06-24 14:39:14,150][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.008315876126289368, 'train/grad_norm_pre_clip': 9.213907241821289, 'timing/data': 0.00031251716427505016, 'timing/model': 1.0772458540741354, 'learning_rate/qwen_vl_interface': 5.870276826419353e-07, 'learning_rate/action_model': 5.870276826419353e-06, 'epoch': 0.36}) +[2026-06-24 14:39:15,229][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.14337657392024994, 'train/grad_norm_pre_clip': 86.81226348876953, 'timing/data': 0.0003479141741991043, 'timing/model': 1.0779247689060867, 'learning_rate/qwen_vl_interface': 5.863001119351395e-07, 'learning_rate/action_model': 5.863001119351395e-06, 'epoch': 0.36}) +[2026-06-24 14:39:16,532][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.13700659573078156, 'train/grad_norm_pre_clip': 81.93241882324219, 'timing/data': 0.0003646661061793566, 'timing/model': 1.3010985618457198, 'learning_rate/qwen_vl_interface': 5.855755674525672e-07, 'learning_rate/action_model': 5.855755674525673e-06, 'epoch': 0.36}) +[2026-06-24 14:39:17,608][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.09315472096204758, 'train/grad_norm_pre_clip': 76.19050598144531, 'timing/data': 0.000339999096468091, 'timing/model': 1.0742055249866098, 'learning_rate/qwen_vl_interface': 5.848540496643687e-07, 'learning_rate/action_model': 5.848540496643686e-06, 'epoch': 0.36}) +[2026-06-24 14:39:18,688][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.2184348851442337, 'train/grad_norm_pre_clip': 43.214874267578125, 'timing/data': 0.00033853622153401375, 'timing/model': 1.078651832882315, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 0.36}) +[2026-06-24 14:39:19,771][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.11490441858768463, 'train/grad_norm_pre_clip': 64.5914306640625, 'timing/data': 0.00032563204877078533, 'timing/model': 1.0811048450414091, 'learning_rate/qwen_vl_interface': 5.834200960418683e-07, 'learning_rate/action_model': 5.8342009604186825e-06, 'epoch': 0.36}) +[2026-06-24 14:39:20,850][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.17481888830661774, 'train/grad_norm_pre_clip': 85.95101928710938, 'timing/data': 0.0003657869528979063, 'timing/model': 1.0782095300965011, 'learning_rate/qwen_vl_interface': 5.827076611380423e-07, 'learning_rate/action_model': 5.8270766113804224e-06, 'epoch': 0.36}) +[2026-06-24 14:39:21,933][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.15164609253406525, 'train/grad_norm_pre_clip': 63.24749755859375, 'timing/data': 0.0003384270239621401, 'timing/model': 1.0813108901493251, 'learning_rate/qwen_vl_interface': 5.819982547895437e-07, 'learning_rate/action_model': 5.819982547895436e-06, 'epoch': 0.36}) +[2026-06-24 14:39:23,019][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.1321040689945221, 'train/grad_norm_pre_clip': 65.40717315673828, 'timing/data': 0.0004016959574073553, 'timing/model': 1.0838164100423455, 'learning_rate/qwen_vl_interface': 5.812918774566974e-07, 'learning_rate/action_model': 5.812918774566975e-06, 'epoch': 0.36}) +[2026-06-24 14:39:24,100][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.0506771057844162, 'train/grad_norm_pre_clip': 33.94834899902344, 'timing/data': 0.0003262630198150873, 'timing/model': 1.0792703521437943, 'learning_rate/qwen_vl_interface': 5.805885295978642e-07, 'learning_rate/action_model': 5.805885295978642e-06, 'epoch': 0.36}) +[2026-06-24 14:39:25,176][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.22809891402721405, 'train/grad_norm_pre_clip': 64.52271270751953, 'timing/data': 0.00031885900534689426, 'timing/model': 1.0750167430378497, 'learning_rate/qwen_vl_interface': 5.798882116694394e-07, 'learning_rate/action_model': 5.798882116694394e-06, 'epoch': 0.36}) +[2026-06-24 14:39:26,263][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.0767865777015686, 'train/grad_norm_pre_clip': 72.47692108154297, 'timing/data': 0.0003400889690965414, 'timing/model': 1.0849193099420518, 'learning_rate/qwen_vl_interface': 5.791909241258518e-07, 'learning_rate/action_model': 5.791909241258518e-06, 'epoch': 0.37}) +[2026-06-24 14:39:27,343][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.0817158967256546, 'train/grad_norm_pre_clip': 52.215396881103516, 'timing/data': 0.00034796399995684624, 'timing/model': 1.0782412642147392, 'learning_rate/qwen_vl_interface': 5.784966674195619e-07, 'learning_rate/action_model': 5.784966674195619e-06, 'epoch': 0.37}) +[2026-06-24 14:39:28,424][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.08261950314044952, 'train/grad_norm_pre_clip': 23.70270347595215, 'timing/data': 0.0003149618860334158, 'timing/model': 1.0800837171263993, 'learning_rate/qwen_vl_interface': 5.778054420010683e-07, 'learning_rate/action_model': 5.778054420010683e-06, 'epoch': 0.37}) +[2026-06-24 14:39:29,508][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.05654490739107132, 'train/grad_norm_pre_clip': 52.64811706542969, 'timing/data': 0.00036122906021773815, 'timing/model': 1.0820597510319203, 'learning_rate/qwen_vl_interface': 5.77117248318897e-07, 'learning_rate/action_model': 5.77117248318897e-06, 'epoch': 0.37}) +[2026-06-24 14:39:30,591][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.5049661993980408, 'train/grad_norm_pre_clip': 64.25947570800781, 'timing/data': 0.00031381985172629356, 'timing/model': 1.0811172579415143, 'learning_rate/qwen_vl_interface': 5.76432086819611e-07, 'learning_rate/action_model': 5.76432086819611e-06, 'epoch': 0.37}) +[2026-06-24 14:39:31,675][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.06813196837902069, 'train/grad_norm_pre_clip': 44.33478546142578, 'timing/data': 0.0003176468890160322, 'timing/model': 1.0826423950493336, 'learning_rate/qwen_vl_interface': 5.757499579478037e-07, 'learning_rate/action_model': 5.757499579478036e-06, 'epoch': 0.37}) +[2026-06-24 14:39:32,759][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.05748767778277397, 'train/grad_norm_pre_clip': 61.75545120239258, 'timing/data': 0.0003416819963604212, 'timing/model': 1.0819342718459666, 'learning_rate/qwen_vl_interface': 5.750708621461011e-07, 'learning_rate/action_model': 5.75070862146101e-06, 'epoch': 0.37}) +[2026-06-24 14:39:34,137][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.36419737339019775, 'train/grad_norm_pre_clip': 42.82459259033203, 'timing/data': 0.0003311620093882084, 'timing/model': 1.3763645750004798, 'learning_rate/qwen_vl_interface': 5.743947998551607e-07, 'learning_rate/action_model': 5.7439479985516065e-06, 'epoch': 0.37}) +[2026-06-24 14:39:35,222][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.007321770768612623, 'train/grad_norm_pre_clip': 5.835790634155273, 'timing/data': 0.0003260229714214802, 'timing/model': 1.084060347871855, 'learning_rate/qwen_vl_interface': 5.737217715136726e-07, 'learning_rate/action_model': 5.737217715136726e-06, 'epoch': 0.37}) +[2026-06-24 14:39:36,307][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.11452832818031311, 'train/grad_norm_pre_clip': 110.04827880859375, 'timing/data': 0.00031991093419492245, 'timing/model': 1.083499192027375, 'learning_rate/qwen_vl_interface': 5.730517775583573e-07, 'learning_rate/action_model': 5.730517775583573e-06, 'epoch': 0.37}) +[2026-06-24 14:39:37,390][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.21626734733581543, 'train/grad_norm_pre_clip': 55.04108810424805, 'timing/data': 0.00033161393366754055, 'timing/model': 1.0808424658607692, 'learning_rate/qwen_vl_interface': 5.723848184239668e-07, 'learning_rate/action_model': 5.723848184239668e-06, 'epoch': 0.37}) +[2026-06-24 14:39:38,474][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.07462850213050842, 'train/grad_norm_pre_clip': 35.53973388671875, 'timing/data': 0.0003365729935467243, 'timing/model': 1.081998327979818, 'learning_rate/qwen_vl_interface': 5.717208945432832e-07, 'learning_rate/action_model': 5.717208945432832e-06, 'epoch': 0.37}) +[2026-06-24 14:39:39,557][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.11062569916248322, 'train/grad_norm_pre_clip': 49.181884765625, 'timing/data': 0.00031967111863195896, 'timing/model': 1.0816285719629377, 'learning_rate/qwen_vl_interface': 5.710600063471205e-07, 'learning_rate/action_model': 5.710600063471204e-06, 'epoch': 0.37}) +[2026-06-24 14:39:40,644][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.07297361642122269, 'train/grad_norm_pre_clip': 20.013500213623047, 'timing/data': 0.0003614199813455343, 'timing/model': 1.085663648089394, 'learning_rate/qwen_vl_interface': 5.704021542643201e-07, 'learning_rate/action_model': 5.704021542643201e-06, 'epoch': 0.37}) +[2026-06-24 14:39:41,730][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.03664546459913254, 'train/grad_norm_pre_clip': 17.175024032592773, 'timing/data': 0.00031919009052217007, 'timing/model': 1.083794361911714, 'learning_rate/qwen_vl_interface': 5.697473387217568e-07, 'learning_rate/action_model': 5.697473387217568e-06, 'epoch': 0.37}) +[2026-06-24 14:39:42,812][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.0689668133854866, 'train/grad_norm_pre_clip': 58.94119644165039, 'timing/data': 0.00035185087472200394, 'timing/model': 1.0807820667978376, 'learning_rate/qwen_vl_interface': 5.690955601443324e-07, 'learning_rate/action_model': 5.690955601443324e-06, 'epoch': 0.37}) +[2026-06-24 14:39:43,894][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.03451039642095566, 'train/grad_norm_pre_clip': 11.546060562133789, 'timing/data': 0.00032845791429281235, 'timing/model': 1.0806601978838444, 'learning_rate/qwen_vl_interface': 5.684468189549794e-07, 'learning_rate/action_model': 5.684468189549794e-06, 'epoch': 0.37}) +[2026-06-24 14:39:44,979][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.0009345798753201962, 'train/grad_norm_pre_clip': 0.4961947202682495, 'timing/data': 0.00031622499227523804, 'timing/model': 1.0835008979775012, 'learning_rate/qwen_vl_interface': 5.678011155746586e-07, 'learning_rate/action_model': 5.678011155746586e-06, 'epoch': 0.37}) +[2026-06-24 14:39:46,064][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.08583089709281921, 'train/grad_norm_pre_clip': 50.4454345703125, 'timing/data': 0.00032024201937019825, 'timing/model': 1.083420375129208, 'learning_rate/qwen_vl_interface': 5.671584504223596e-07, 'learning_rate/action_model': 5.671584504223596e-06, 'epoch': 0.37}) +[2026-06-24 14:39:47,151][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.06460805237293243, 'train/grad_norm_pre_clip': 44.32162857055664, 'timing/data': 0.00032161385752260685, 'timing/model': 1.085033566923812, 'learning_rate/qwen_vl_interface': 5.665188239151019e-07, 'learning_rate/action_model': 5.665188239151018e-06, 'epoch': 0.37}) +[2026-06-24 14:39:48,237][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.060260891914367676, 'train/grad_norm_pre_clip': 41.597286224365234, 'timing/data': 0.00039675598964095116, 'timing/model': 1.0837375198025256, 'learning_rate/qwen_vl_interface': 5.658822364679306e-07, 'learning_rate/action_model': 5.658822364679307e-06, 'epoch': 0.37}) +[2026-06-24 14:39:49,322][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.08657241612672806, 'train/grad_norm_pre_clip': 61.62590789794922, 'timing/data': 0.0003225458785891533, 'timing/model': 1.0840730189811438, 'learning_rate/qwen_vl_interface': 5.652486884939225e-07, 'learning_rate/action_model': 5.652486884939225e-06, 'epoch': 0.37}) +[2026-06-24 14:39:50,616][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.04164180904626846, 'train/grad_norm_pre_clip': 15.027262687683105, 'timing/data': 0.0003409208729863167, 'timing/model': 1.2916558941360563, 'learning_rate/qwen_vl_interface': 5.646181804041781e-07, 'learning_rate/action_model': 5.646181804041781e-06, 'epoch': 0.37}) +[2026-06-24 14:39:51,706][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.1828835904598236, 'train/grad_norm_pre_clip': 87.49894714355469, 'timing/data': 0.0003253121394664049, 'timing/model': 1.0888199131004512, 'learning_rate/qwen_vl_interface': 5.639907126078277e-07, 'learning_rate/action_model': 5.639907126078277e-06, 'epoch': 0.37}) +[2026-06-24 14:39:52,790][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.0013425112701952457, 'train/grad_norm_pre_clip': 1.8011313676834106, 'timing/data': 0.00032495101913809776, 'timing/model': 1.0819850210100412, 'learning_rate/qwen_vl_interface': 5.633662855120295e-07, 'learning_rate/action_model': 5.633662855120294e-06, 'epoch': 0.37}) +[2026-06-24 14:39:53,877][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.00036295014433562756, 'train/grad_norm_pre_clip': 0.30548763275146484, 'timing/data': 0.00035570887848734856, 'timing/model': 1.0858782099094242, 'learning_rate/qwen_vl_interface': 5.627448995219659e-07, 'learning_rate/action_model': 5.627448995219658e-06, 'epoch': 0.37}) +[2026-06-24 14:39:54,967][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.008078050799667835, 'train/grad_norm_pre_clip': 5.42402982711792, 'timing/data': 0.00034459796734154224, 'timing/model': 1.0878572000656277, 'learning_rate/qwen_vl_interface': 5.621265550408484e-07, 'learning_rate/action_model': 5.621265550408485e-06, 'epoch': 0.37}) +[2026-06-24 14:39:56,055][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.03567484766244888, 'train/grad_norm_pre_clip': 31.02450942993164, 'timing/data': 0.00033779488876461983, 'timing/model': 1.086895260028541, 'learning_rate/qwen_vl_interface': 5.61511252469914e-07, 'learning_rate/action_model': 5.61511252469914e-06, 'epoch': 0.37}) +[2026-06-24 14:39:57,143][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.07340221852064133, 'train/grad_norm_pre_clip': 75.5516128540039, 'timing/data': 0.00032727583311498165, 'timing/model': 1.0863172828685492, 'learning_rate/qwen_vl_interface': 5.608989922084255e-07, 'learning_rate/action_model': 5.608989922084256e-06, 'epoch': 0.37}) +[2026-06-24 14:39:58,231][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.05581603944301605, 'train/grad_norm_pre_clip': 38.164955139160156, 'timing/data': 0.00032051303423941135, 'timing/model': 1.0857251090928912, 'learning_rate/qwen_vl_interface': 5.602897746536722e-07, 'learning_rate/action_model': 5.602897746536722e-06, 'epoch': 0.37}) +[2026-06-24 14:39:59,320][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.027194490656256676, 'train/grad_norm_pre_clip': 38.61578369140625, 'timing/data': 0.00039451196789741516, 'timing/model': 1.087680002907291, 'learning_rate/qwen_vl_interface': 5.596836002009682e-07, 'learning_rate/action_model': 5.596836002009683e-06, 'epoch': 0.37}) +[2026-06-24 14:40:00,410][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.04484427347779274, 'train/grad_norm_pre_clip': 23.165529251098633, 'timing/data': 0.00034824502654373646, 'timing/model': 1.088127352995798, 'learning_rate/qwen_vl_interface': 5.590804692436549e-07, 'learning_rate/action_model': 5.590804692436549e-06, 'epoch': 0.37}) +[2026-06-24 14:40:01,498][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.0355449914932251, 'train/grad_norm_pre_clip': 16.088274002075195, 'timing/data': 0.0004154809284955263, 'timing/model': 1.0864086509682238, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 0.37}) +[2026-06-24 14:40:02,586][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.43520522117614746, 'train/grad_norm_pre_clip': 84.59176635742188, 'timing/data': 0.00032153399661183357, 'timing/model': 1.0864121168851852, 'learning_rate/qwen_vl_interface': 5.57883339378682e-07, 'learning_rate/action_model': 5.57883339378682e-06, 'epoch': 0.37}) +[2026-06-24 14:40:03,671][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.27240315079689026, 'train/grad_norm_pre_clip': 109.51887512207031, 'timing/data': 0.0004386049695312977, 'timing/model': 1.0830113119445741, 'learning_rate/qwen_vl_interface': 5.572893412478278e-07, 'learning_rate/action_model': 5.572893412478278e-06, 'epoch': 0.37}) +[2026-06-24 14:40:04,758][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.20156855881214142, 'train/grad_norm_pre_clip': 98.7382583618164, 'timing/data': 0.00033128284849226475, 'timing/model': 1.0852019998710603, 'learning_rate/qwen_vl_interface': 5.566983881659715e-07, 'learning_rate/action_model': 5.5669838816597155e-06, 'epoch': 0.37}) +[2026-06-24 14:40:05,846][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.10038778185844421, 'train/grad_norm_pre_clip': 33.98958969116211, 'timing/data': 0.00033886684104800224, 'timing/model': 1.0865981238894165, 'learning_rate/qwen_vl_interface': 5.561104805165771e-07, 'learning_rate/action_model': 5.561104805165771e-06, 'epoch': 0.37}) +[2026-06-24 14:40:07,143][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.2786135673522949, 'train/grad_norm_pre_clip': 40.64594268798828, 'timing/data': 0.0003223859239369631, 'timing/model': 1.2957220929674804, 'learning_rate/qwen_vl_interface': 5.555256186811309e-07, 'learning_rate/action_model': 5.555256186811309e-06, 'epoch': 0.37}) +[2026-06-24 14:40:08,232][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.0645424872636795, 'train/grad_norm_pre_clip': 56.47168731689453, 'timing/data': 0.0003509300295263529, 'timing/model': 1.0868895889725536, 'learning_rate/qwen_vl_interface': 5.549438030391438e-07, 'learning_rate/action_model': 5.5494380303914375e-06, 'epoch': 0.37}) +[2026-06-24 14:40:09,321][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.05974990129470825, 'train/grad_norm_pre_clip': 16.042211532592773, 'timing/data': 0.00033261487260460854, 'timing/model': 1.0871401389595121, 'learning_rate/qwen_vl_interface': 5.543650339681486e-07, 'learning_rate/action_model': 5.543650339681487e-06, 'epoch': 0.37}) +[2026-06-24 14:40:10,419][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.23799550533294678, 'train/grad_norm_pre_clip': 92.15725708007812, 'timing/data': 0.0003585140220820904, 'timing/model': 1.0885210949927568, 'learning_rate/qwen_vl_interface': 5.537893118437038e-07, 'learning_rate/action_model': 5.537893118437037e-06, 'epoch': 0.37}) +[2026-06-24 14:40:11,510][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.1923018842935562, 'train/grad_norm_pre_clip': 66.75579071044922, 'timing/data': 0.00044965604320168495, 'timing/model': 1.0894602399785072, 'learning_rate/qwen_vl_interface': 5.532166370393886e-07, 'learning_rate/action_model': 5.5321663703938855e-06, 'epoch': 0.37}) +[2026-06-24 14:40:12,600][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.043405186384916306, 'train/grad_norm_pre_clip': 33.744171142578125, 'timing/data': 0.0004580509848892689, 'timing/model': 1.0882366690784693, 'learning_rate/qwen_vl_interface': 5.526470099268049e-07, 'learning_rate/action_model': 5.526470099268048e-06, 'epoch': 0.37}) +[2026-06-24 14:40:13,690][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.19293181598186493, 'train/grad_norm_pre_clip': 89.31368255615234, 'timing/data': 0.000330441165715456, 'timing/model': 1.0883296730462462, 'learning_rate/qwen_vl_interface': 5.520804308755785e-07, 'learning_rate/action_model': 5.520804308755785e-06, 'epoch': 0.37}) +[2026-06-24 14:40:14,777][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.023358017206192017, 'train/grad_norm_pre_clip': 5.445810794830322, 'timing/data': 0.0003470729570835829, 'timing/model': 1.084805035032332, 'learning_rate/qwen_vl_interface': 5.515169002533553e-07, 'learning_rate/action_model': 5.515169002533553e-06, 'epoch': 0.37}) +[2026-06-24 14:40:15,865][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.11485506594181061, 'train/grad_norm_pre_clip': 67.27674865722656, 'timing/data': 0.0003404109738767147, 'timing/model': 1.0866602018941194, 'learning_rate/qwen_vl_interface': 5.509564184258055e-07, 'learning_rate/action_model': 5.509564184258055e-06, 'epoch': 0.37}) +[2026-06-24 14:40:16,953][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.06995002925395966, 'train/grad_norm_pre_clip': 38.27213668823242, 'timing/data': 0.0003258930519223213, 'timing/model': 1.0860418290831149, 'learning_rate/qwen_vl_interface': 5.503989857566187e-07, 'learning_rate/action_model': 5.503989857566187e-06, 'epoch': 0.37}) +[2026-06-24 14:40:18,042][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.10560615360736847, 'train/grad_norm_pre_clip': 44.022491455078125, 'timing/data': 0.0003123881760984659, 'timing/model': 1.0872006460558623, 'learning_rate/qwen_vl_interface': 5.498446026075072e-07, 'learning_rate/action_model': 5.498446026075071e-06, 'epoch': 0.37}) +[2026-06-24 14:40:19,130][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.00024013698566704988, 'train/grad_norm_pre_clip': 0.08167154341936111, 'timing/data': 0.00033911713398993015, 'timing/model': 1.0872233679983765, 'learning_rate/qwen_vl_interface': 5.492932693382044e-07, 'learning_rate/action_model': 5.492932693382044e-06, 'epoch': 0.37}) +[2026-06-24 14:40:20,219][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.006234192755073309, 'train/grad_norm_pre_clip': 8.337567329406738, 'timing/data': 0.00033380696550011635, 'timing/model': 1.087325237924233, 'learning_rate/qwen_vl_interface': 5.487449863064641e-07, 'learning_rate/action_model': 5.487449863064641e-06, 'epoch': 0.37}) +[2026-06-24 14:40:21,314][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.027606721967458725, 'train/grad_norm_pre_clip': 14.94054126739502, 'timing/data': 0.0005182649474591017, 'timing/model': 1.0929004200734198, 'learning_rate/qwen_vl_interface': 5.481997538680614e-07, 'learning_rate/action_model': 5.481997538680614e-06, 'epoch': 0.37}) +[2026-06-24 14:40:22,401][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.14445364475250244, 'train/grad_norm_pre_clip': 49.50740051269531, 'timing/data': 0.0003504590131342411, 'timing/model': 1.0854293729644269, 'learning_rate/qwen_vl_interface': 5.476575723767918e-07, 'learning_rate/action_model': 5.476575723767918e-06, 'epoch': 0.37}) +[2026-06-24 14:40:23,689][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.03734244406223297, 'train/grad_norm_pre_clip': 33.71961975097656, 'timing/data': 0.000329658854752779, 'timing/model': 1.286315970821306, 'learning_rate/qwen_vl_interface': 5.471184421844707e-07, 'learning_rate/action_model': 5.4711844218447065e-06, 'epoch': 0.37}) +[2026-06-24 14:40:24,778][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.2078021913766861, 'train/grad_norm_pre_clip': 68.33856964111328, 'timing/data': 0.0003153530415147543, 'timing/model': 1.0870542761404067, 'learning_rate/qwen_vl_interface': 5.465823636409342e-07, 'learning_rate/action_model': 5.465823636409342e-06, 'epoch': 0.37}) +[2026-06-24 14:40:25,866][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.1121024340391159, 'train/grad_norm_pre_clip': 66.83385467529297, 'timing/data': 0.00035966699942946434, 'timing/model': 1.0864883109461516, 'learning_rate/qwen_vl_interface': 5.460493370940377e-07, 'learning_rate/action_model': 5.460493370940376e-06, 'epoch': 0.37}) +[2026-06-24 14:40:26,956][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.02275923267006874, 'train/grad_norm_pre_clip': 19.904510498046875, 'timing/data': 0.00032131397165358067, 'timing/model': 1.0881500749383122, 'learning_rate/qwen_vl_interface': 5.455193628896563e-07, 'learning_rate/action_model': 5.455193628896562e-06, 'epoch': 0.37}) +[2026-06-24 14:40:28,046][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.08136647194623947, 'train/grad_norm_pre_clip': 21.718332290649414, 'timing/data': 0.0003324651625007391, 'timing/model': 1.0890337449964136, 'learning_rate/qwen_vl_interface': 5.449924413716847e-07, 'learning_rate/action_model': 5.449924413716847e-06, 'epoch': 0.37}) +[2026-06-24 14:40:29,141][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.10625927895307541, 'train/grad_norm_pre_clip': 45.3774528503418, 'timing/data': 0.0004984168335795403, 'timing/model': 1.0925049709621817, 'learning_rate/qwen_vl_interface': 5.444685728820357e-07, 'learning_rate/action_model': 5.444685728820357e-06, 'epoch': 0.37}) +[2026-06-24 14:40:30,227][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.051997993141412735, 'train/grad_norm_pre_clip': 37.90780258178711, 'timing/data': 0.00034818402491509914, 'timing/model': 1.084570461884141, 'learning_rate/qwen_vl_interface': 5.439477577606429e-07, 'learning_rate/action_model': 5.4394775776064295e-06, 'epoch': 0.37}) +[2026-06-24 14:40:31,320][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.10368848592042923, 'train/grad_norm_pre_clip': 38.7193717956543, 'timing/data': 0.00034192297607660294, 'timing/model': 1.0908130479510874, 'learning_rate/qwen_vl_interface': 5.434299963454578e-07, 'learning_rate/action_model': 5.434299963454577e-06, 'epoch': 0.37}) +[2026-06-24 14:40:32,407][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.008168071508407593, 'train/grad_norm_pre_clip': 13.80418872833252, 'timing/data': 0.0004808339290320873, 'timing/model': 1.0855827340856194, 'learning_rate/qwen_vl_interface': 5.429152889724485e-07, 'learning_rate/action_model': 5.429152889724484e-06, 'epoch': 0.37}) +[2026-06-24 14:40:33,498][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.09743459522724152, 'train/grad_norm_pre_clip': 72.12083435058594, 'timing/data': 0.00031623500399291515, 'timing/model': 1.0895157069899142, 'learning_rate/qwen_vl_interface': 5.424036359756045e-07, 'learning_rate/action_model': 5.424036359756045e-06, 'epoch': 0.37}) +[2026-06-24 14:40:34,589][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.07603755593299866, 'train/grad_norm_pre_clip': 58.01080322265625, 'timing/data': 0.00033018109388649464, 'timing/model': 1.089562705019489, 'learning_rate/qwen_vl_interface': 5.418950376869317e-07, 'learning_rate/action_model': 5.418950376869316e-06, 'epoch': 0.37}) +[2026-06-24 14:40:35,678][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.18555490672588348, 'train/grad_norm_pre_clip': 55.47041320800781, 'timing/data': 0.0003244001418352127, 'timing/model': 1.0865140620153397, 'learning_rate/qwen_vl_interface': 5.413894944364535e-07, 'learning_rate/action_model': 5.413894944364536e-06, 'epoch': 0.37}) +[2026-06-24 14:40:36,769][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.006390510592609644, 'train/grad_norm_pre_clip': 9.934121131896973, 'timing/data': 0.00035288394428789616, 'timing/model': 1.0896770779509097, 'learning_rate/qwen_vl_interface': 5.40887006552212e-07, 'learning_rate/action_model': 5.40887006552212e-06, 'epoch': 0.37}) +[2026-06-24 14:40:37,860][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.056488439440727234, 'train/grad_norm_pre_clip': 42.02907943725586, 'timing/data': 0.0003525919746607542, 'timing/model': 1.0895600270014256, 'learning_rate/qwen_vl_interface': 5.403875743602658e-07, 'learning_rate/action_model': 5.403875743602658e-06, 'epoch': 0.37}) +[2026-06-24 14:40:38,949][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.0037802665028721094, 'train/grad_norm_pre_clip': 2.302952289581299, 'timing/data': 0.00032675405964255333, 'timing/model': 1.0869278570171446, 'learning_rate/qwen_vl_interface': 5.398911981846905e-07, 'learning_rate/action_model': 5.3989119818469045e-06, 'epoch': 0.37}) +[2026-06-24 14:40:40,332][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.05119771882891655, 'train/grad_norm_pre_clip': 28.433439254760742, 'timing/data': 0.0003523027990013361, 'timing/model': 1.3822012308519334, 'learning_rate/qwen_vl_interface': 5.393978783475803e-07, 'learning_rate/action_model': 5.393978783475803e-06, 'epoch': 0.37}) +[2026-06-24 14:40:41,427][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.07776898145675659, 'train/grad_norm_pre_clip': 81.0259017944336, 'timing/data': 0.00033295596949756145, 'timing/model': 1.0924561799038202, 'learning_rate/qwen_vl_interface': 5.389076151690446e-07, 'learning_rate/action_model': 5.389076151690446e-06, 'epoch': 0.37}) +[2026-06-24 14:40:42,516][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.18910519778728485, 'train/grad_norm_pre_clip': 173.1661834716797, 'timing/data': 0.00035138102248311043, 'timing/model': 1.087893680203706, 'learning_rate/qwen_vl_interface': 5.384204089672097e-07, 'learning_rate/action_model': 5.384204089672097e-06, 'epoch': 0.37}) +[2026-06-24 14:40:43,608][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.1374720185995102, 'train/grad_norm_pre_clip': 80.8996810913086, 'timing/data': 0.0003315738867968321, 'timing/model': 1.0897733040619642, 'learning_rate/qwen_vl_interface': 5.379362600582187e-07, 'learning_rate/action_model': 5.3793626005821865e-06, 'epoch': 0.37}) +[2026-06-24 14:40:44,700][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.2358410507440567, 'train/grad_norm_pre_clip': 151.3283233642578, 'timing/data': 0.0003592150751501322, 'timing/model': 1.091040164930746, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 0.37}) +[2026-06-24 14:40:45,788][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.07718829810619354, 'train/grad_norm_pre_clip': 40.31912612915039, 'timing/data': 0.00034643104299902916, 'timing/model': 1.0861522241029888, 'learning_rate/qwen_vl_interface': 5.369771353734196e-07, 'learning_rate/action_model': 5.369771353734197e-06, 'epoch': 0.37}) +[2026-06-24 14:40:46,876][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.06053546071052551, 'train/grad_norm_pre_clip': 70.93197631835938, 'timing/data': 0.0003644151147454977, 'timing/model': 1.0865622039418668, 'learning_rate/qwen_vl_interface': 5.365021602199772e-07, 'learning_rate/action_model': 5.3650216021997716e-06, 'epoch': 0.37}) +[2026-06-24 14:40:47,966][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.018396563827991486, 'train/grad_norm_pre_clip': 40.799346923828125, 'timing/data': 0.0003482140600681305, 'timing/model': 1.0886509900446981, 'learning_rate/qwen_vl_interface': 5.360302436041088e-07, 'learning_rate/action_model': 5.360302436041088e-06, 'epoch': 0.37}) +[2026-06-24 14:40:49,057][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.009883582592010498, 'train/grad_norm_pre_clip': 11.063811302185059, 'timing/data': 0.0003172168508172035, 'timing/model': 1.0889108781702816, 'learning_rate/qwen_vl_interface': 5.355613858320372e-07, 'learning_rate/action_model': 5.355613858320372e-06, 'epoch': 0.37}) +[2026-06-24 14:40:50,147][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.10988318920135498, 'train/grad_norm_pre_clip': 66.39429473876953, 'timing/data': 0.0003280371893197298, 'timing/model': 1.0880763290915638, 'learning_rate/qwen_vl_interface': 5.350955872079984e-07, 'learning_rate/action_model': 5.350955872079985e-06, 'epoch': 0.37}) +[2026-06-24 14:40:51,236][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.09654886275529861, 'train/grad_norm_pre_clip': 99.63626861572266, 'timing/data': 0.00032009207643568516, 'timing/model': 1.087317100027576, 'learning_rate/qwen_vl_interface': 5.346328480342436e-07, 'learning_rate/action_model': 5.346328480342437e-06, 'epoch': 0.37}) +[2026-06-24 14:40:52,332][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.06656036525964737, 'train/grad_norm_pre_clip': 40.966739654541016, 'timing/data': 0.0005856610368937254, 'timing/model': 1.094312631059438, 'learning_rate/qwen_vl_interface': 5.341731686110402e-07, 'learning_rate/action_model': 5.341731686110402e-06, 'epoch': 0.37}) +[2026-06-24 14:40:53,421][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.03746478259563446, 'train/grad_norm_pre_clip': 31.111623764038086, 'timing/data': 0.0003376950044184923, 'timing/model': 1.0876514771953225, 'learning_rate/qwen_vl_interface': 5.337165492366681e-07, 'learning_rate/action_model': 5.337165492366681e-06, 'epoch': 0.37}) +[2026-06-24 14:40:54,511][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.07465631514787674, 'train/grad_norm_pre_clip': 32.207767486572266, 'timing/data': 0.0003904239274561405, 'timing/model': 1.0879956530407071, 'learning_rate/qwen_vl_interface': 5.332629902074238e-07, 'learning_rate/action_model': 5.332629902074238e-06, 'epoch': 0.37}) +[2026-06-24 14:40:55,802][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.1271049678325653, 'train/grad_norm_pre_clip': 42.49666976928711, 'timing/data': 0.0003393178340047598, 'timing/model': 1.289662545081228, 'learning_rate/qwen_vl_interface': 5.328124918176157e-07, 'learning_rate/action_model': 5.328124918176157e-06, 'epoch': 0.37}) +[2026-06-24 14:40:56,889][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.11875995248556137, 'train/grad_norm_pre_clip': 54.8448371887207, 'timing/data': 0.00034375698305666447, 'timing/model': 1.0853081890381873, 'learning_rate/qwen_vl_interface': 5.323650543595683e-07, 'learning_rate/action_model': 5.323650543595683e-06, 'epoch': 0.37}) +[2026-06-24 14:40:57,977][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.08915160596370697, 'train/grad_norm_pre_clip': 30.28989028930664, 'timing/data': 0.0003434661775827408, 'timing/model': 1.0857969669159502, 'learning_rate/qwen_vl_interface': 5.319206781236184e-07, 'learning_rate/action_model': 5.319206781236184e-06, 'epoch': 0.37}) +[2026-06-24 14:40:59,064][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.09957414865493774, 'train/grad_norm_pre_clip': 74.53612518310547, 'timing/data': 0.00034542917273938656, 'timing/model': 1.0860298539046198, 'learning_rate/qwen_vl_interface': 5.314793633981169e-07, 'learning_rate/action_model': 5.314793633981169e-06, 'epoch': 0.37}) +[2026-06-24 14:41:00,154][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.08799081295728683, 'train/grad_norm_pre_clip': 49.53935623168945, 'timing/data': 0.000368333188816905, 'timing/model': 1.0877657369710505, 'learning_rate/qwen_vl_interface': 5.310411104694285e-07, 'learning_rate/action_model': 5.3104111046942854e-06, 'epoch': 0.37}) +[2026-06-24 14:41:01,240][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.019217414781451225, 'train/grad_norm_pre_clip': 20.563907623291016, 'timing/data': 0.00035615009255707264, 'timing/model': 1.0842163630295545, 'learning_rate/qwen_vl_interface': 5.306059196219306e-07, 'learning_rate/action_model': 5.306059196219306e-06, 'epoch': 0.37}) +[2026-06-24 14:41:02,329][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.14586225152015686, 'train/grad_norm_pre_clip': 59.070159912109375, 'timing/data': 0.0003652069717645645, 'timing/model': 1.0874376990832388, 'learning_rate/qwen_vl_interface': 5.301737911380134e-07, 'learning_rate/action_model': 5.301737911380133e-06, 'epoch': 0.37}) +[2026-06-24 14:41:03,418][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.027429716661572456, 'train/grad_norm_pre_clip': 32.2530403137207, 'timing/data': 0.0003116361331194639, 'timing/model': 1.0876428540796041, 'learning_rate/qwen_vl_interface': 5.297447252980807e-07, 'learning_rate/action_model': 5.297447252980806e-06, 'epoch': 0.37}) +[2026-06-24 14:41:04,508][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.13614578545093536, 'train/grad_norm_pre_clip': 59.10930252075195, 'timing/data': 0.00039259810000658035, 'timing/model': 1.0884725630749017, 'learning_rate/qwen_vl_interface': 5.293187223805486e-07, 'learning_rate/action_model': 5.2931872238054855e-06, 'epoch': 0.37}) +[2026-06-24 14:41:05,596][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.04428277909755707, 'train/grad_norm_pre_clip': 34.559627532958984, 'timing/data': 0.00036644889041781425, 'timing/model': 1.0858659909572452, 'learning_rate/qwen_vl_interface': 5.288957826618466e-07, 'learning_rate/action_model': 5.288957826618466e-06, 'epoch': 0.37}) +[2026-06-24 14:41:06,686][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.004164223559200764, 'train/grad_norm_pre_clip': 3.247520685195923, 'timing/data': 0.00031600380316376686, 'timing/model': 1.0886233940254897, 'learning_rate/qwen_vl_interface': 5.284759064164139e-07, 'learning_rate/action_model': 5.284759064164139e-06, 'epoch': 0.37}) +[2026-06-24 14:41:07,777][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.1151961013674736, 'train/grad_norm_pre_clip': 29.684709548950195, 'timing/data': 0.0003398589324206114, 'timing/model': 1.0891597319860011, 'learning_rate/qwen_vl_interface': 5.280590939167053e-07, 'learning_rate/action_model': 5.280590939167054e-06, 'epoch': 0.37}) +[2026-06-24 14:41:08,866][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.011531569994986057, 'train/grad_norm_pre_clip': 6.725011348724365, 'timing/data': 0.00034343497827649117, 'timing/model': 1.0874076569452882, 'learning_rate/qwen_vl_interface': 5.276453454331852e-07, 'learning_rate/action_model': 5.276453454331852e-06, 'epoch': 0.37}) +[2026-06-24 14:41:09,957][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.001180078717879951, 'train/grad_norm_pre_clip': 0.5880260467529297, 'timing/data': 0.0003355711232870817, 'timing/model': 1.0892914577852935, 'learning_rate/qwen_vl_interface': 5.272346612343307e-07, 'learning_rate/action_model': 5.272346612343306e-06, 'epoch': 0.37}) +[2026-06-24 14:41:11,048][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.0731375589966774, 'train/grad_norm_pre_clip': 39.7602424621582, 'timing/data': 0.0003446789924055338, 'timing/model': 1.0898671499453485, 'learning_rate/qwen_vl_interface': 5.268270415866297e-07, 'learning_rate/action_model': 5.268270415866297e-06, 'epoch': 0.37}) +[2026-06-24 14:41:12,138][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.15022116899490356, 'train/grad_norm_pre_clip': 162.90049743652344, 'timing/data': 0.00039907009340822697, 'timing/model': 1.0878878480289131, 'learning_rate/qwen_vl_interface': 5.264224867545831e-07, 'learning_rate/action_model': 5.2642248675458315e-06, 'epoch': 0.37}) +[2026-06-24 14:41:13,231][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.16310007870197296, 'train/grad_norm_pre_clip': 82.00049591064453, 'timing/data': 0.00031952117569744587, 'timing/model': 1.091345988912508, 'learning_rate/qwen_vl_interface': 5.260209970007018e-07, 'learning_rate/action_model': 5.260209970007018e-06, 'epoch': 0.37}) +[2026-06-24 14:41:14,323][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.04521339386701584, 'train/grad_norm_pre_clip': 23.925064086914062, 'timing/data': 0.0003460810985416174, 'timing/model': 1.090526367072016, 'learning_rate/qwen_vl_interface': 5.256225725855076e-07, 'learning_rate/action_model': 5.256225725855076e-06, 'epoch': 0.37}) +[2026-06-24 14:41:15,703][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.029076721519231796, 'train/grad_norm_pre_clip': 22.55919647216797, 'timing/data': 0.00033714412711560726, 'timing/model': 1.3787604591343552, 'learning_rate/qwen_vl_interface': 5.252272137675345e-07, 'learning_rate/action_model': 5.252272137675344e-06, 'epoch': 0.37}) +[2026-06-24 14:41:16,793][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.023906877264380455, 'train/grad_norm_pre_clip': 35.806514739990234, 'timing/data': 0.000330410897731781, 'timing/model': 1.088122304994613, 'learning_rate/qwen_vl_interface': 5.248349208033264e-07, 'learning_rate/action_model': 5.248349208033264e-06, 'epoch': 0.37}) +[2026-06-24 14:41:17,884][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.023738404735922813, 'train/grad_norm_pre_clip': 21.957826614379883, 'timing/data': 0.00035972706973552704, 'timing/model': 1.0897384828422219, 'learning_rate/qwen_vl_interface': 5.244456939474375e-07, 'learning_rate/action_model': 5.244456939474376e-06, 'epoch': 0.37}) +[2026-06-24 14:41:18,975][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.23042239248752594, 'train/grad_norm_pre_clip': 52.149173736572266, 'timing/data': 0.0003244299441576004, 'timing/model': 1.088433698983863, 'learning_rate/qwen_vl_interface': 5.240595334524335e-07, 'learning_rate/action_model': 5.240595334524335e-06, 'epoch': 0.37}) +[2026-06-24 14:41:20,065][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.30541372299194336, 'train/grad_norm_pre_clip': 64.97371673583984, 'timing/data': 0.0003275859635323286, 'timing/model': 1.0890769178513438, 'learning_rate/qwen_vl_interface': 5.236764395688895e-07, 'learning_rate/action_model': 5.236764395688895e-06, 'epoch': 0.38}) +[2026-06-24 14:41:21,161][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.033784929662942886, 'train/grad_norm_pre_clip': 44.62925720214844, 'timing/data': 0.00032104295678436756, 'timing/model': 1.093970888061449, 'learning_rate/qwen_vl_interface': 5.2329641254539e-07, 'learning_rate/action_model': 5.2329641254539e-06, 'epoch': 0.38}) +[2026-06-24 14:41:22,252][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.014585493132472038, 'train/grad_norm_pre_clip': 7.144839763641357, 'timing/data': 0.00034094112925231457, 'timing/model': 1.089568429859355, 'learning_rate/qwen_vl_interface': 5.229194526285321e-07, 'learning_rate/action_model': 5.229194526285321e-06, 'epoch': 0.38}) +[2026-06-24 14:41:23,354][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.04308527335524559, 'train/grad_norm_pre_clip': 47.26002883911133, 'timing/data': 0.0005275020375847816, 'timing/model': 1.09899913915433, 'learning_rate/qwen_vl_interface': 5.225455600629201e-07, 'learning_rate/action_model': 5.2254556006292016e-06, 'epoch': 0.38}) +[2026-06-24 14:41:24,448][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.21125301718711853, 'train/grad_norm_pre_clip': 136.4119110107422, 'timing/data': 0.0004576209466904402, 'timing/model': 1.0927056269720197, 'learning_rate/qwen_vl_interface': 5.22174735091169e-07, 'learning_rate/action_model': 5.22174735091169e-06, 'epoch': 0.38}) +[2026-06-24 14:41:25,539][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.03527718037366867, 'train/grad_norm_pre_clip': 30.24396324157715, 'timing/data': 0.0003548969980329275, 'timing/model': 1.0892711291089654, 'learning_rate/qwen_vl_interface': 5.218069779539029e-07, 'learning_rate/action_model': 5.218069779539029e-06, 'epoch': 0.38}) +[2026-06-24 14:41:26,633][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.1277376413345337, 'train/grad_norm_pre_clip': 88.86328887939453, 'timing/data': 0.0003424941096454859, 'timing/model': 1.091854274040088, 'learning_rate/qwen_vl_interface': 5.214422888897555e-07, 'learning_rate/action_model': 5.214422888897555e-06, 'epoch': 0.38}) +[2026-06-24 14:41:27,724][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.012657800689339638, 'train/grad_norm_pre_clip': 6.6053643226623535, 'timing/data': 0.0003560488112270832, 'timing/model': 1.0899687090422958, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 0.38}) +[2026-06-24 14:41:28,815][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.018851855769753456, 'train/grad_norm_pre_clip': 13.851946830749512, 'timing/data': 0.00033060205169022083, 'timing/model': 1.089570048963651, 'learning_rate/qwen_vl_interface': 5.207221159253976e-07, 'learning_rate/action_model': 5.207221159253975e-06, 'epoch': 0.38}) +[2026-06-24 14:41:29,904][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.3913601338863373, 'train/grad_norm_pre_clip': 33.695274353027344, 'timing/data': 0.00037181912921369076, 'timing/model': 1.0866920270491391, 'learning_rate/qwen_vl_interface': 5.203666324924993e-07, 'learning_rate/action_model': 5.203666324924992e-06, 'epoch': 0.38}) +[2026-06-24 14:41:30,997][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.03497053682804108, 'train/grad_norm_pre_clip': 31.63188934326172, 'timing/data': 0.0005773250013589859, 'timing/model': 1.091223576106131, 'learning_rate/qwen_vl_interface': 5.200142180673437e-07, 'learning_rate/action_model': 5.200142180673437e-06, 'epoch': 0.38}) +[2026-06-24 14:41:32,089][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.018627023324370384, 'train/grad_norm_pre_clip': 11.611985206604004, 'timing/data': 0.0003280669916421175, 'timing/model': 1.090611555147916, 'learning_rate/qwen_vl_interface': 5.196648728786098e-07, 'learning_rate/action_model': 5.196648728786098e-06, 'epoch': 0.38}) +[2026-06-24 14:41:33,182][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.049016814678907394, 'train/grad_norm_pre_clip': 44.93229675292969, 'timing/data': 0.0003398079425096512, 'timing/model': 1.0910319238901138, 'learning_rate/qwen_vl_interface': 5.193185971529831e-07, 'learning_rate/action_model': 5.1931859715298315e-06, 'epoch': 0.38}) +[2026-06-24 14:41:34,514][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.2854776680469513, 'train/grad_norm_pre_clip': 82.15601348876953, 'timing/data': 0.0003506201319396496, 'timing/model': 1.3305809409357607, 'learning_rate/qwen_vl_interface': 5.189753911151584e-07, 'learning_rate/action_model': 5.189753911151584e-06, 'epoch': 0.38}) +[2026-06-24 14:41:35,605][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.1649583876132965, 'train/grad_norm_pre_clip': 116.64424896240234, 'timing/data': 0.0003247009590268135, 'timing/model': 1.0891643408685923, 'learning_rate/qwen_vl_interface': 5.186352549878387e-07, 'learning_rate/action_model': 5.1863525498783864e-06, 'epoch': 0.38}) +[2026-06-24 14:41:36,695][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.19441130757331848, 'train/grad_norm_pre_clip': 63.143001556396484, 'timing/data': 0.00030961213633418083, 'timing/model': 1.0885370909236372, 'learning_rate/qwen_vl_interface': 5.182981889917344e-07, 'learning_rate/action_model': 5.182981889917344e-06, 'epoch': 0.38}) +[2026-06-24 14:41:37,786][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.08027725666761398, 'train/grad_norm_pre_clip': 55.655635833740234, 'timing/data': 0.0003221258521080017, 'timing/model': 1.0897611621767282, 'learning_rate/qwen_vl_interface': 5.179641933455638e-07, 'learning_rate/action_model': 5.179641933455638e-06, 'epoch': 0.38}) +[2026-06-24 14:41:38,877][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.03281036764383316, 'train/grad_norm_pre_clip': 21.921972274780273, 'timing/data': 0.00034532998688519, 'timing/model': 1.089406784856692, 'learning_rate/qwen_vl_interface': 5.17633268266053e-07, 'learning_rate/action_model': 5.1763326826605295e-06, 'epoch': 0.38}) +[2026-06-24 14:41:39,968][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.16286715865135193, 'train/grad_norm_pre_clip': 48.61048889160156, 'timing/data': 0.00031987112015485764, 'timing/model': 1.0895678161177784, 'learning_rate/qwen_vl_interface': 5.173054139679357e-07, 'learning_rate/action_model': 5.173054139679357e-06, 'epoch': 0.38}) +[2026-06-24 14:41:41,064][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.08013410866260529, 'train/grad_norm_pre_clip': 46.88784408569336, 'timing/data': 0.0003121469635516405, 'timing/model': 1.0933911039028317, 'learning_rate/qwen_vl_interface': 5.169806306639539e-07, 'learning_rate/action_model': 5.1698063066395385e-06, 'epoch': 0.38}) +[2026-06-24 14:41:42,157][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.015963440760970116, 'train/grad_norm_pre_clip': 25.36416244506836, 'timing/data': 0.00032992009073495865, 'timing/model': 1.0913645930122584, 'learning_rate/qwen_vl_interface': 5.166589185648543e-07, 'learning_rate/action_model': 5.1665891856485425e-06, 'epoch': 0.38}) +[2026-06-24 14:41:43,245][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.17364582419395447, 'train/grad_norm_pre_clip': 35.79387664794922, 'timing/data': 0.0003245898988097906, 'timing/model': 1.087015506112948, 'learning_rate/qwen_vl_interface': 5.163402778793937e-07, 'learning_rate/action_model': 5.163402778793937e-06, 'epoch': 0.38}) +[2026-06-24 14:41:44,338][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.1638631522655487, 'train/grad_norm_pre_clip': 78.98381805419922, 'timing/data': 0.00031651509925723076, 'timing/model': 1.0906618109438568, 'learning_rate/qwen_vl_interface': 5.160247088143344e-07, 'learning_rate/action_model': 5.160247088143343e-06, 'epoch': 0.38}) +[2026-06-24 14:41:45,431][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.020956994965672493, 'train/grad_norm_pre_clip': 26.497652053833008, 'timing/data': 0.00032718502916395664, 'timing/model': 1.092129679163918, 'learning_rate/qwen_vl_interface': 5.157122115744453e-07, 'learning_rate/action_model': 5.157122115744452e-06, 'epoch': 0.38}) +[2026-06-24 14:41:46,524][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.056748174130916595, 'train/grad_norm_pre_clip': 48.1464958190918, 'timing/data': 0.0003841719590127468, 'timing/model': 1.0907678687945008, 'learning_rate/qwen_vl_interface': 5.154027863625028e-07, 'learning_rate/action_model': 5.154027863625028e-06, 'epoch': 0.38}) +[2026-06-24 14:41:47,618][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.15188764035701752, 'train/grad_norm_pre_clip': 38.94576644897461, 'timing/data': 0.0003353001084178686, 'timing/model': 1.093031874857843, 'learning_rate/qwen_vl_interface': 5.150964333792892e-07, 'learning_rate/action_model': 5.150964333792893e-06, 'epoch': 0.38}) +[2026-06-24 14:41:48,712][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.17984101176261902, 'train/grad_norm_pre_clip': 55.185054779052734, 'timing/data': 0.00043101003393530846, 'timing/model': 1.0911420909687877, 'learning_rate/qwen_vl_interface': 5.147931528235941e-07, 'learning_rate/action_model': 5.1479315282359405e-06, 'epoch': 0.38}) +[2026-06-24 14:41:49,801][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.19871436059474945, 'train/grad_norm_pre_clip': 145.14981079101562, 'timing/data': 0.0003260830417275429, 'timing/model': 1.0878303130157292, 'learning_rate/qwen_vl_interface': 5.144929448922123e-07, 'learning_rate/action_model': 5.144929448922123e-06, 'epoch': 0.38}) +[2026-06-24 14:41:50,891][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.3112926781177521, 'train/grad_norm_pre_clip': 83.13744354248047, 'timing/data': 0.0003293389454483986, 'timing/model': 1.088085503084585, 'learning_rate/qwen_vl_interface': 5.141958097799466e-07, 'learning_rate/action_model': 5.141958097799465e-06, 'epoch': 0.38}) +[2026-06-24 14:41:51,981][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.021579809486865997, 'train/grad_norm_pre_clip': 15.841137886047363, 'timing/data': 0.0003370540216565132, 'timing/model': 1.0888555902056396, 'learning_rate/qwen_vl_interface': 5.139017476796034e-07, 'learning_rate/action_model': 5.139017476796034e-06, 'epoch': 0.38}) +[2026-06-24 14:41:53,264][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.022443344816565514, 'train/grad_norm_pre_clip': 19.78578758239746, 'timing/data': 0.0003232278395444155, 'timing/model': 1.28078856295906, 'learning_rate/qwen_vl_interface': 5.136107587819971e-07, 'learning_rate/action_model': 5.136107587819971e-06, 'epoch': 0.38}) +[2026-06-24 14:41:54,348][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.012485763989388943, 'train/grad_norm_pre_clip': 9.505531311035156, 'timing/data': 0.00040839705616235733, 'timing/model': 1.0828006439842284, 'learning_rate/qwen_vl_interface': 5.13322843275948e-07, 'learning_rate/action_model': 5.13322843275948e-06, 'epoch': 0.38}) +[2026-06-24 14:41:55,436][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.20460298657417297, 'train/grad_norm_pre_clip': 87.3801040649414, 'timing/data': 0.00032413890585303307, 'timing/model': 1.08622674504295, 'learning_rate/qwen_vl_interface': 5.130380013482798e-07, 'learning_rate/action_model': 5.130380013482798e-06, 'epoch': 0.38}) +[2026-06-24 14:41:56,527][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.019130805507302284, 'train/grad_norm_pre_clip': 12.168112754821777, 'timing/data': 0.0003154731821268797, 'timing/model': 1.0893784500658512, 'learning_rate/qwen_vl_interface': 5.127562331838241e-07, 'learning_rate/action_model': 5.12756233183824e-06, 'epoch': 0.38}) +[2026-06-24 14:41:57,617][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.0022200702223926783, 'train/grad_norm_pre_clip': 2.2543270587921143, 'timing/data': 0.0003350100014358759, 'timing/model': 1.087531897937879, 'learning_rate/qwen_vl_interface': 5.124775389654176e-07, 'learning_rate/action_model': 5.1247753896541754e-06, 'epoch': 0.38}) +[2026-06-24 14:41:58,708][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.4318009316921234, 'train/grad_norm_pre_clip': 90.47340393066406, 'timing/data': 0.000358623918145895, 'timing/model': 1.0895436678547412, 'learning_rate/qwen_vl_interface': 5.122019188739006e-07, 'learning_rate/action_model': 5.122019188739007e-06, 'epoch': 0.38}) +[2026-06-24 14:41:59,797][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.06943615525960922, 'train/grad_norm_pre_clip': 61.71316909790039, 'timing/data': 0.0003447779454290867, 'timing/model': 1.0880731239449233, 'learning_rate/qwen_vl_interface': 5.119293730881215e-07, 'learning_rate/action_model': 5.119293730881214e-06, 'epoch': 0.38}) +[2026-06-24 14:42:00,888][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.030155325308442116, 'train/grad_norm_pre_clip': 26.339754104614258, 'timing/data': 0.0003309918101876974, 'timing/model': 1.0894654609728605, 'learning_rate/qwen_vl_interface': 5.11659901784932e-07, 'learning_rate/action_model': 5.1165990178493195e-06, 'epoch': 0.38}) +[2026-06-24 14:42:01,980][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.18060392141342163, 'train/grad_norm_pre_clip': 68.70943450927734, 'timing/data': 0.00034247408621013165, 'timing/model': 1.090035482076928, 'learning_rate/qwen_vl_interface': 5.113935051391885e-07, 'learning_rate/action_model': 5.113935051391885e-06, 'epoch': 0.38}) +[2026-06-24 14:42:03,065][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.012606741860508919, 'train/grad_norm_pre_clip': 7.62757682800293, 'timing/data': 0.0003143108915537596, 'timing/model': 1.0831213819328696, 'learning_rate/qwen_vl_interface': 5.111301833237533e-07, 'learning_rate/action_model': 5.1113018332375325e-06, 'epoch': 0.38}) +[2026-06-24 14:42:04,155][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.031684815883636475, 'train/grad_norm_pre_clip': 36.33833694458008, 'timing/data': 0.00036133010871708393, 'timing/model': 1.088385347975418, 'learning_rate/qwen_vl_interface': 5.108699365094926e-07, 'learning_rate/action_model': 5.108699365094926e-06, 'epoch': 0.38}) +[2026-06-24 14:42:05,242][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.0386025533080101, 'train/grad_norm_pre_clip': 41.3779411315918, 'timing/data': 0.0003334770444780588, 'timing/model': 1.0855756551027298, 'learning_rate/qwen_vl_interface': 5.106127648652786e-07, 'learning_rate/action_model': 5.106127648652786e-06, 'epoch': 0.38}) +[2026-06-24 14:42:06,335][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.003201322164386511, 'train/grad_norm_pre_clip': 4.660310745239258, 'timing/data': 0.0003243200480937958, 'timing/model': 1.091644945088774, 'learning_rate/qwen_vl_interface': 5.103586685579867e-07, 'learning_rate/action_model': 5.1035866855798665e-06, 'epoch': 0.38}) +[2026-06-24 14:42:07,425][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.01072603091597557, 'train/grad_norm_pre_clip': 15.713342666625977, 'timing/data': 0.00033213384449481964, 'timing/model': 1.0883588059805334, 'learning_rate/qwen_vl_interface': 5.101076477524967e-07, 'learning_rate/action_model': 5.101076477524967e-06, 'epoch': 0.38}) +[2026-06-24 14:42:08,514][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.14197158813476562, 'train/grad_norm_pre_clip': 69.39823150634766, 'timing/data': 0.00037297094240784645, 'timing/model': 1.0874717868864536, 'learning_rate/qwen_vl_interface': 5.098597026116943e-07, 'learning_rate/action_model': 5.098597026116943e-06, 'epoch': 0.38}) +[2026-06-24 14:42:09,607][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.013158513233065605, 'train/grad_norm_pre_clip': 9.658617973327637, 'timing/data': 0.0003225160762667656, 'timing/model': 1.0912482859566808, 'learning_rate/qwen_vl_interface': 5.096148332964679e-07, 'learning_rate/action_model': 5.096148332964678e-06, 'epoch': 0.38}) +[2026-06-24 14:42:10,903][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.10623951256275177, 'train/grad_norm_pre_clip': 47.72692108154297, 'timing/data': 0.0003716489300131798, 'timing/model': 1.2939487618859857, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 0.38}) +[2026-06-24 14:42:11,995][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.07320858538150787, 'train/grad_norm_pre_clip': 48.14635467529297, 'timing/data': 0.0003396382089704275, 'timing/model': 1.0903602759353817, 'learning_rate/qwen_vl_interface': 5.091343227763185e-07, 'learning_rate/action_model': 5.091343227763185e-06, 'epoch': 0.38}) +[2026-06-24 14:42:13,087][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.0558098703622818, 'train/grad_norm_pre_clip': 52.47395706176758, 'timing/data': 0.0003200010396540165, 'timing/model': 1.0908566280268133, 'learning_rate/qwen_vl_interface': 5.088986818831944e-07, 'learning_rate/action_model': 5.088986818831944e-06, 'epoch': 0.38}) +[2026-06-24 14:42:14,179][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.05955442786216736, 'train/grad_norm_pre_clip': 61.664493560791016, 'timing/data': 0.00031562289223074913, 'timing/model': 1.090180056169629, 'learning_rate/qwen_vl_interface': 5.086661174392419e-07, 'learning_rate/action_model': 5.086661174392418e-06, 'epoch': 0.38}) +[2026-06-24 14:42:15,268][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.05674237012863159, 'train/grad_norm_pre_clip': 41.178260803222656, 'timing/data': 0.0003196210600435734, 'timing/model': 1.0876996009610593, 'learning_rate/qwen_vl_interface': 5.084366295953703e-07, 'learning_rate/action_model': 5.0843662959537026e-06, 'epoch': 0.38}) +[2026-06-24 14:42:16,362][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.04647427424788475, 'train/grad_norm_pre_clip': 55.72058868408203, 'timing/data': 0.0003821479622274637, 'timing/model': 1.0916824790183455, 'learning_rate/qwen_vl_interface': 5.082102185004907e-07, 'learning_rate/action_model': 5.082102185004907e-06, 'epoch': 0.38}) +[2026-06-24 14:42:17,455][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.10178336501121521, 'train/grad_norm_pre_clip': 74.65792083740234, 'timing/data': 0.0003289380110800266, 'timing/model': 1.091108919121325, 'learning_rate/qwen_vl_interface': 5.079868843015197e-07, 'learning_rate/action_model': 5.079868843015197e-06, 'epoch': 0.38}) +[2026-06-24 14:42:18,548][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.3082120418548584, 'train/grad_norm_pre_clip': 79.86572265625, 'timing/data': 0.00034172297455370426, 'timing/model': 1.0921619390137494, 'learning_rate/qwen_vl_interface': 5.077666271433767e-07, 'learning_rate/action_model': 5.077666271433767e-06, 'epoch': 0.38}) +[2026-06-24 14:42:19,644][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.0760791078209877, 'train/grad_norm_pre_clip': 89.84819030761719, 'timing/data': 0.00033955788239836693, 'timing/model': 1.0938878231681883, 'learning_rate/qwen_vl_interface': 5.075494471689833e-07, 'learning_rate/action_model': 5.075494471689833e-06, 'epoch': 0.38}) +[2026-06-24 14:42:20,736][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.08962073177099228, 'train/grad_norm_pre_clip': 71.60775756835938, 'timing/data': 0.0003216040786355734, 'timing/model': 1.09057492390275, 'learning_rate/qwen_vl_interface': 5.073353445192652e-07, 'learning_rate/action_model': 5.073353445192652e-06, 'epoch': 0.38}) +[2026-06-24 14:42:21,830][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.1769794225692749, 'train/grad_norm_pre_clip': 30.753767013549805, 'timing/data': 0.00035063899122178555, 'timing/model': 1.0919915668200701, 'learning_rate/qwen_vl_interface': 5.071243193331522e-07, 'learning_rate/action_model': 5.071243193331522e-06, 'epoch': 0.38}) +[2026-06-24 14:42:22,921][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.382442831993103, 'train/grad_norm_pre_clip': 57.36783981323242, 'timing/data': 0.0003579431213438511, 'timing/model': 1.0894614888820797, 'learning_rate/qwen_vl_interface': 5.069163717475756e-07, 'learning_rate/action_model': 5.069163717475756e-06, 'epoch': 0.38}) +[2026-06-24 14:42:24,019][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.07552283257246017, 'train/grad_norm_pre_clip': 38.6506233215332, 'timing/data': 0.0003519919700920582, 'timing/model': 1.0960606548469514, 'learning_rate/qwen_vl_interface': 5.067115018974702e-07, 'learning_rate/action_model': 5.067115018974702e-06, 'epoch': 0.38}) +[2026-06-24 14:42:25,111][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.04364108666777611, 'train/grad_norm_pre_clip': 63.16655731201172, 'timing/data': 0.0003528639208525419, 'timing/model': 1.091191960964352, 'learning_rate/qwen_vl_interface': 5.065097099157743e-07, 'learning_rate/action_model': 5.065097099157743e-06, 'epoch': 0.38}) +[2026-06-24 14:42:26,200][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.1833966076374054, 'train/grad_norm_pre_clip': 78.85563659667969, 'timing/data': 0.0003279871307313442, 'timing/model': 1.0870987051166594, 'learning_rate/qwen_vl_interface': 5.063109959334281e-07, 'learning_rate/action_model': 5.063109959334281e-06, 'epoch': 0.38}) +[2026-06-24 14:42:27,294][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.08569063246250153, 'train/grad_norm_pre_clip': 29.3468017578125, 'timing/data': 0.0003185588866472244, 'timing/model': 1.0919336949009448, 'learning_rate/qwen_vl_interface': 5.061153600793754e-07, 'learning_rate/action_model': 5.061153600793753e-06, 'epoch': 0.38}) +[2026-06-24 14:42:28,589][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.007078856695443392, 'train/grad_norm_pre_clip': 6.952536106109619, 'timing/data': 0.0003313629422336817, 'timing/model': 1.2934674839489162, 'learning_rate/qwen_vl_interface': 5.059228024805618e-07, 'learning_rate/action_model': 5.059228024805619e-06, 'epoch': 0.38}) +[2026-06-24 14:42:29,683][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.05044690519571304, 'train/grad_norm_pre_clip': 42.96412658691406, 'timing/data': 0.0003278669901192188, 'timing/model': 1.0920986128039658, 'learning_rate/qwen_vl_interface': 5.057333232619356e-07, 'learning_rate/action_model': 5.057333232619355e-06, 'epoch': 0.38}) +[2026-06-24 14:42:30,774][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.03626016899943352, 'train/grad_norm_pre_clip': 28.71847152709961, 'timing/data': 0.0003381760325282812, 'timing/model': 1.0899620449636132, 'learning_rate/qwen_vl_interface': 5.055469225464484e-07, 'learning_rate/action_model': 5.055469225464484e-06, 'epoch': 0.38}) +[2026-06-24 14:42:31,871][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.12952348589897156, 'train/grad_norm_pre_clip': 72.47001647949219, 'timing/data': 0.0003528241068124771, 'timing/model': 1.0954205971211195, 'learning_rate/qwen_vl_interface': 5.053636004550538e-07, 'learning_rate/action_model': 5.053636004550538e-06, 'epoch': 0.38}) +[2026-06-24 14:42:32,961][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.1388523131608963, 'train/grad_norm_pre_clip': 91.20806884765625, 'timing/data': 0.00031194696202874184, 'timing/model': 1.0886639740783721, 'learning_rate/qwen_vl_interface': 5.051833571067068e-07, 'learning_rate/action_model': 5.0518335710670685e-06, 'epoch': 0.38}) +[2026-06-24 14:42:34,067][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.04159919545054436, 'train/grad_norm_pre_clip': 13.518603324890137, 'timing/data': 0.0004567289724946022, 'timing/model': 1.103878035908565, 'learning_rate/qwen_vl_interface': 5.050061926183655e-07, 'learning_rate/action_model': 5.050061926183655e-06, 'epoch': 0.38}) +[2026-06-24 14:42:35,293][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.1310628205537796, 'train/grad_norm_pre_clip': 102.77326202392578, 'timing/data': 0.000413327943533659, 'timing/model': 1.223714712075889, 'learning_rate/qwen_vl_interface': 5.048321071049904e-07, 'learning_rate/action_model': 5.048321071049904e-06, 'epoch': 0.38}) +[2026-06-24 14:42:36,401][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.07991668581962585, 'train/grad_norm_pre_clip': 14.400256156921387, 'timing/data': 0.0004084981046617031, 'timing/model': 1.1068399930372834, 'learning_rate/qwen_vl_interface': 5.046611006795437e-07, 'learning_rate/action_model': 5.046611006795436e-06, 'epoch': 0.38}) +[2026-06-24 14:42:37,491][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.22769129276275635, 'train/grad_norm_pre_clip': 67.69747924804688, 'timing/data': 0.0003369739279150963, 'timing/model': 1.0881264100316912, 'learning_rate/qwen_vl_interface': 5.044931734529889e-07, 'learning_rate/action_model': 5.0449317345298885e-06, 'epoch': 0.38}) +[2026-06-24 14:42:38,581][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.044717684388160706, 'train/grad_norm_pre_clip': 28.098899841308594, 'timing/data': 0.0003672409802675247, 'timing/model': 1.0886965510435402, 'learning_rate/qwen_vl_interface': 5.043283255342926e-07, 'learning_rate/action_model': 5.043283255342926e-06, 'epoch': 0.38}) +[2026-06-24 14:42:39,675][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.12816046178340912, 'train/grad_norm_pre_clip': 37.61091232299805, 'timing/data': 0.00031879893504083157, 'timing/model': 1.0917690270580351, 'learning_rate/qwen_vl_interface': 5.041665570304227e-07, 'learning_rate/action_model': 5.041665570304226e-06, 'epoch': 0.38}) +[2026-06-24 14:42:40,766][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.041983507573604584, 'train/grad_norm_pre_clip': 33.71677780151367, 'timing/data': 0.0003329259343445301, 'timing/model': 1.0900737191550434, 'learning_rate/qwen_vl_interface': 5.040078680463499e-07, 'learning_rate/action_model': 5.040078680463498e-06, 'epoch': 0.38}) +[2026-06-24 14:42:41,855][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.05220123380422592, 'train/grad_norm_pre_clip': 38.306880950927734, 'timing/data': 0.0003466908819973469, 'timing/model': 1.0872229989618063, 'learning_rate/qwen_vl_interface': 5.038522586850444e-07, 'learning_rate/action_model': 5.0385225868504445e-06, 'epoch': 0.38}) +[2026-06-24 14:42:42,949][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.13601034879684448, 'train/grad_norm_pre_clip': 60.23640441894531, 'timing/data': 0.00032061198726296425, 'timing/model': 1.092321694130078, 'learning_rate/qwen_vl_interface': 5.0369972904748e-07, 'learning_rate/action_model': 5.0369972904748004e-06, 'epoch': 0.38}) +[2026-06-24 14:42:44,043][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.09020474553108215, 'train/grad_norm_pre_clip': 82.3006362915039, 'timing/data': 0.0003275249619036913, 'timing/model': 1.0919634199235588, 'learning_rate/qwen_vl_interface': 5.035502792326317e-07, 'learning_rate/action_model': 5.035502792326316e-06, 'epoch': 0.38}) +[2026-06-24 14:42:45,136][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.01448066346347332, 'train/grad_norm_pre_clip': 26.650310516357422, 'timing/data': 0.0003619999624788761, 'timing/model': 1.0918402369134128, 'learning_rate/qwen_vl_interface': 5.034039093374759e-07, 'learning_rate/action_model': 5.034039093374759e-06, 'epoch': 0.38}) +[2026-06-24 14:42:46,494][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.0245069433003664, 'train/grad_norm_pre_clip': 17.715898513793945, 'timing/data': 0.00032980996184051037, 'timing/model': 1.3561558269429952, 'learning_rate/qwen_vl_interface': 5.032606194569903e-07, 'learning_rate/action_model': 5.032606194569902e-06, 'epoch': 0.38}) +[2026-06-24 14:42:47,585][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.07984365522861481, 'train/grad_norm_pre_clip': 60.975582122802734, 'timing/data': 0.00033514993265271187, 'timing/model': 1.089447817998007, 'learning_rate/qwen_vl_interface': 5.031204096841541e-07, 'learning_rate/action_model': 5.0312040968415415e-06, 'epoch': 0.38}) +[2026-06-24 14:42:48,682][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.14406418800354004, 'train/grad_norm_pre_clip': 31.616626739501953, 'timing/data': 0.0003490159288048744, 'timing/model': 1.0955606801435351, 'learning_rate/qwen_vl_interface': 5.029832801099477e-07, 'learning_rate/action_model': 5.029832801099478e-06, 'epoch': 0.38}) +[2026-06-24 14:42:49,778][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.14927074313163757, 'train/grad_norm_pre_clip': 59.72550582885742, 'timing/data': 0.0005949689075350761, 'timing/model': 1.0937261208891869, 'learning_rate/qwen_vl_interface': 5.028492308233534e-07, 'learning_rate/action_model': 5.028492308233534e-06, 'epoch': 0.38}) +[2026-06-24 14:42:50,869][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.1712312400341034, 'train/grad_norm_pre_clip': 67.75533294677734, 'timing/data': 0.0003631629515439272, 'timing/model': 1.0895202921237797, 'learning_rate/qwen_vl_interface': 5.027182619113541e-07, 'learning_rate/action_model': 5.027182619113541e-06, 'epoch': 0.38}) +[2026-06-24 14:42:51,967][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.030258728191256523, 'train/grad_norm_pre_clip': 48.04918670654297, 'timing/data': 0.00036395411007106304, 'timing/model': 1.0965368740726262, 'learning_rate/qwen_vl_interface': 5.02590373458934e-07, 'learning_rate/action_model': 5.02590373458934e-06, 'epoch': 0.38}) +[2026-06-24 14:42:53,057][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.10706993192434311, 'train/grad_norm_pre_clip': 55.9364128112793, 'timing/data': 0.0003888709470629692, 'timing/model': 1.0883199551608413, 'learning_rate/qwen_vl_interface': 5.024655655490792e-07, 'learning_rate/action_model': 5.024655655490792e-06, 'epoch': 0.38}) +[2026-06-24 14:42:54,147][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.11662822961807251, 'train/grad_norm_pre_clip': 71.11103820800781, 'timing/data': 0.0003329759929329157, 'timing/model': 1.0879740330856293, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 0.38}) +[2026-06-24 14:42:55,241][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.0917041003704071, 'train/grad_norm_pre_clip': 34.091583251953125, 'timing/data': 0.00031270808540284634, 'timing/model': 1.0924227768555284, 'learning_rate/qwen_vl_interface': 5.022251916790099e-07, 'learning_rate/action_model': 5.022251916790099e-06, 'epoch': 0.38}) +[2026-06-24 14:42:56,329][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.019491050392389297, 'train/grad_norm_pre_clip': 19.0010986328125, 'timing/data': 0.0003235680051147938, 'timing/model': 1.0867985419463366, 'learning_rate/qwen_vl_interface': 5.021096258747724e-07, 'learning_rate/action_model': 5.021096258747724e-06, 'epoch': 0.38}) +[2026-06-24 14:42:57,420][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.06612464785575867, 'train/grad_norm_pre_clip': 53.63120651245117, 'timing/data': 0.00040693487972021103, 'timing/model': 1.089598486199975, 'learning_rate/qwen_vl_interface': 5.019971409250508e-07, 'learning_rate/action_model': 5.0199714092505085e-06, 'epoch': 0.38}) +[2026-06-24 14:42:58,514][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.048043012619018555, 'train/grad_norm_pre_clip': 22.768150329589844, 'timing/data': 0.0003183879889547825, 'timing/model': 1.09162210021168, 'learning_rate/qwen_vl_interface': 5.018877369028366e-07, 'learning_rate/action_model': 5.018877369028366e-06, 'epoch': 0.38}) +[2026-06-24 14:42:59,603][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.19807934761047363, 'train/grad_norm_pre_clip': 58.6154670715332, 'timing/data': 0.0003497579600661993, 'timing/model': 1.0877125489059836, 'learning_rate/qwen_vl_interface': 5.017814138791207e-07, 'learning_rate/action_model': 5.017814138791207e-06, 'epoch': 0.38}) +[2026-06-24 14:43:00,694][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.020749632269144058, 'train/grad_norm_pre_clip': 12.486638069152832, 'timing/data': 0.00032963906414806843, 'timing/model': 1.0895924021024257, 'learning_rate/qwen_vl_interface': 5.01678171922894e-07, 'learning_rate/action_model': 5.016781719228939e-06, 'epoch': 0.38}) +[2026-06-24 14:43:01,791][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.050536204129457474, 'train/grad_norm_pre_clip': 28.005701065063477, 'timing/data': 0.00033577089197933674, 'timing/model': 1.0953459690790623, 'learning_rate/qwen_vl_interface': 5.015780111011501e-07, 'learning_rate/action_model': 5.015780111011501e-06, 'epoch': 0.38}) +[2026-06-24 14:43:03,094][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.06425982713699341, 'train/grad_norm_pre_clip': 26.017013549804688, 'timing/data': 0.0003871680237352848, 'timing/model': 1.3014935681130737, 'learning_rate/qwen_vl_interface': 5.01480931478882e-07, 'learning_rate/action_model': 5.01480931478882e-06, 'epoch': 0.38}) +[2026-06-24 14:43:04,181][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.0669354498386383, 'train/grad_norm_pre_clip': 83.37374877929688, 'timing/data': 0.00033171381801366806, 'timing/model': 1.0852013148833066, 'learning_rate/qwen_vl_interface': 5.013869331190835e-07, 'learning_rate/action_model': 5.013869331190835e-06, 'epoch': 0.38}) +[2026-06-24 14:43:05,269][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.019043050706386566, 'train/grad_norm_pre_clip': 25.132062911987305, 'timing/data': 0.00038082594983279705, 'timing/model': 1.0861599969211966, 'learning_rate/qwen_vl_interface': 5.012960160827491e-07, 'learning_rate/action_model': 5.012960160827491e-06, 'epoch': 0.38}) +[2026-06-24 14:43:06,362][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.049470722675323486, 'train/grad_norm_pre_clip': 17.642162322998047, 'timing/data': 0.0003395490348339081, 'timing/model': 1.0909611620008945, 'learning_rate/qwen_vl_interface': 5.012081804288739e-07, 'learning_rate/action_model': 5.0120818042887395e-06, 'epoch': 0.38}) +[2026-06-24 14:43:07,455][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.008631016127765179, 'train/grad_norm_pre_clip': 11.699764251708984, 'timing/data': 0.00034953816793859005, 'timing/model': 1.091765033081174, 'learning_rate/qwen_vl_interface': 5.011234262144533e-07, 'learning_rate/action_model': 5.011234262144532e-06, 'epoch': 0.38}) +[2026-06-24 14:43:08,544][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.0020264331251382828, 'train/grad_norm_pre_clip': 1.0799838304519653, 'timing/data': 0.00037239002995193005, 'timing/model': 1.0867279930971563, 'learning_rate/qwen_vl_interface': 5.010417534944839e-07, 'learning_rate/action_model': 5.010417534944838e-06, 'epoch': 0.38}) +[2026-06-24 14:43:09,634][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.03888600692152977, 'train/grad_norm_pre_clip': 19.014562606811523, 'timing/data': 0.00033188401721417904, 'timing/model': 1.089116623159498, 'learning_rate/qwen_vl_interface': 5.009631623219619e-07, 'learning_rate/action_model': 5.009631623219619e-06, 'epoch': 0.38}) +[2026-06-24 14:43:10,727][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.24286918342113495, 'train/grad_norm_pre_clip': 68.11013793945312, 'timing/data': 0.0003358209505677223, 'timing/model': 1.0914430459961295, 'learning_rate/qwen_vl_interface': 5.008876527478842e-07, 'learning_rate/action_model': 5.008876527478842e-06, 'epoch': 0.38}) +[2026-06-24 14:43:11,820][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.01711435616016388, 'train/grad_norm_pre_clip': 5.456600666046143, 'timing/data': 0.00035096006467938423, 'timing/model': 1.0909202119801193, 'learning_rate/qwen_vl_interface': 5.008152248212485e-07, 'learning_rate/action_model': 5.008152248212485e-06, 'epoch': 0.38}) +[2026-06-24 14:43:12,910][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.013530305586755276, 'train/grad_norm_pre_clip': 10.752537727355957, 'timing/data': 0.0003202918451279402, 'timing/model': 1.088181713828817, 'learning_rate/qwen_vl_interface': 5.007458785890523e-07, 'learning_rate/action_model': 5.007458785890523e-06, 'epoch': 0.38}) +[2026-06-24 14:43:14,002][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.11000916361808777, 'train/grad_norm_pre_clip': 41.25049591064453, 'timing/data': 0.0003473828546702862, 'timing/model': 1.0908840340562165, 'learning_rate/qwen_vl_interface': 5.006796140962932e-07, 'learning_rate/action_model': 5.006796140962932e-06, 'epoch': 0.38}) +[2026-06-24 14:43:15,094][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.025725873187184334, 'train/grad_norm_pre_clip': 38.89501953125, 'timing/data': 0.0003740140236914158, 'timing/model': 1.0901563039515167, 'learning_rate/qwen_vl_interface': 5.006164313859703e-07, 'learning_rate/action_model': 5.006164313859703e-06, 'epoch': 0.39}) +[2026-06-24 14:43:16,188][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.1897304505109787, 'train/grad_norm_pre_clip': 42.358375549316406, 'timing/data': 0.0003101630136370659, 'timing/model': 1.0920155688654631, 'learning_rate/qwen_vl_interface': 5.005563304990811e-07, 'learning_rate/action_model': 5.0055633049908116e-06, 'epoch': 0.39}) +[2026-06-24 14:43:17,279][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.007792823016643524, 'train/grad_norm_pre_clip': 10.117082595825195, 'timing/data': 0.0003403499722480774, 'timing/model': 1.090089158155024, 'learning_rate/qwen_vl_interface': 5.004993114746255e-07, 'learning_rate/action_model': 5.004993114746256e-06, 'epoch': 0.39}) +[2026-06-24 14:43:18,373][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.020611867308616638, 'train/grad_norm_pre_clip': 21.77577018737793, 'timing/data': 0.0003271049354225397, 'timing/model': 1.0914185959845781, 'learning_rate/qwen_vl_interface': 5.004453743496026e-07, 'learning_rate/action_model': 5.004453743496026e-06, 'epoch': 0.39}) +[2026-06-24 14:43:19,464][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.1404784917831421, 'train/grad_norm_pre_clip': 47.47454071044922, 'timing/data': 0.00040748598985373974, 'timing/model': 1.0902655979152769, 'learning_rate/qwen_vl_interface': 5.003945191590103e-07, 'learning_rate/action_model': 5.003945191590104e-06, 'epoch': 0.39}) +[2026-06-24 14:43:20,852][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.07685671001672745, 'train/grad_norm_pre_clip': 77.44747924804688, 'timing/data': 0.0003221859224140644, 'timing/model': 1.3858784709591419, 'learning_rate/qwen_vl_interface': 5.003467459358492e-07, 'learning_rate/action_model': 5.003467459358492e-06, 'epoch': 0.39}) +[2026-06-24 14:43:21,944][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.1628391444683075, 'train/grad_norm_pre_clip': 63.32747268676758, 'timing/data': 0.00032698502764105797, 'timing/model': 1.0905751490499824, 'learning_rate/qwen_vl_interface': 5.003020547111186e-07, 'learning_rate/action_model': 5.003020547111185e-06, 'epoch': 0.39}) +[2026-06-24 14:43:23,038][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.02479194849729538, 'train/grad_norm_pre_clip': 26.07042121887207, 'timing/data': 0.000325602013617754, 'timing/model': 1.0924663529731333, 'learning_rate/qwen_vl_interface': 5.002604455138174e-07, 'learning_rate/action_model': 5.002604455138173e-06, 'epoch': 0.39}) +[2026-06-24 14:43:24,133][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.1096271350979805, 'train/grad_norm_pre_clip': 26.100147247314453, 'timing/data': 0.0003870967775583267, 'timing/model': 1.0932072349824011, 'learning_rate/qwen_vl_interface': 5.002219183709461e-07, 'learning_rate/action_model': 5.00221918370946e-06, 'epoch': 0.39}) +[2026-06-24 14:43:25,230][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.013232405297458172, 'train/grad_norm_pre_clip': 19.709903717041016, 'timing/data': 0.0003192999865859747, 'timing/model': 1.095483863959089, 'learning_rate/qwen_vl_interface': 5.001864733075048e-07, 'learning_rate/action_model': 5.001864733075047e-06, 'epoch': 0.39}) +[2026-06-24 14:43:26,325][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.11091575771570206, 'train/grad_norm_pre_clip': 51.92265319824219, 'timing/data': 0.0005748211406171322, 'timing/model': 1.0925633541774005, 'learning_rate/qwen_vl_interface': 5.001541103464922e-07, 'learning_rate/action_model': 5.001541103464922e-06, 'epoch': 0.39}) +[2026-06-24 14:43:27,418][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.18276247382164001, 'train/grad_norm_pre_clip': 72.35015869140625, 'timing/data': 0.0003466119524091482, 'timing/model': 1.0910898330621421, 'learning_rate/qwen_vl_interface': 5.001248295089095e-07, 'learning_rate/action_model': 5.001248295089095e-06, 'epoch': 0.39}) +[2026-06-24 14:43:28,512][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.015551875345408916, 'train/grad_norm_pre_clip': 20.958436965942383, 'timing/data': 0.0003241291269659996, 'timing/model': 1.0929142120294273, 'learning_rate/qwen_vl_interface': 5.000986308137567e-07, 'learning_rate/action_model': 5.0009863081375666e-06, 'epoch': 0.39}) +[2026-06-24 14:43:29,607][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.06574303656816483, 'train/grad_norm_pre_clip': 39.9048957824707, 'timing/data': 0.000321713974699378, 'timing/model': 1.0928544390480965, 'learning_rate/qwen_vl_interface': 5.000755142780331e-07, 'learning_rate/action_model': 5.00075514278033e-06, 'epoch': 0.39}) +[2026-06-24 14:43:30,700][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.056992851197719574, 'train/grad_norm_pre_clip': 38.912940979003906, 'timing/data': 0.0003644959069788456, 'timing/model': 1.0915217930451035, 'learning_rate/qwen_vl_interface': 5.000554799167387e-07, 'learning_rate/action_model': 5.000554799167387e-06, 'epoch': 0.39}) +[2026-06-24 14:43:31,792][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.09347223490476608, 'train/grad_norm_pre_clip': 53.75733947753906, 'timing/data': 0.0003141800407320261, 'timing/model': 1.090737477876246, 'learning_rate/qwen_vl_interface': 5.000385277428741e-07, 'learning_rate/action_model': 5.000385277428741e-06, 'epoch': 0.39}) +[2026-06-24 14:43:32,887][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.01310825441032648, 'train/grad_norm_pre_clip': 10.938667297363281, 'timing/data': 0.00033300602808594704, 'timing/model': 1.0925280540250242, 'learning_rate/qwen_vl_interface': 5.000246577674399e-07, 'learning_rate/action_model': 5.000246577674399e-06, 'epoch': 0.39}) +[2026-06-24 14:43:33,980][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.1850699782371521, 'train/grad_norm_pre_clip': 84.8387451171875, 'timing/data': 0.00032822717912495136, 'timing/model': 1.0913971569389105, 'learning_rate/qwen_vl_interface': 5.000138699994347e-07, 'learning_rate/action_model': 5.000138699994347e-06, 'epoch': 0.39}) +[2026-06-24 14:43:35,072][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.10102631151676178, 'train/grad_norm_pre_clip': 52.332069396972656, 'timing/data': 0.00031014299020171165, 'timing/model': 1.0906894248910248, 'learning_rate/qwen_vl_interface': 5.000061644458599e-07, 'learning_rate/action_model': 5.0000616444586e-06, 'epoch': 0.39}) +[2026-06-24 14:43:36,167][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.06963615864515305, 'train/grad_norm_pre_clip': 113.95923614501953, 'timing/data': 0.00031443103216588497, 'timing/model': 1.0929362690076232, 'learning_rate/qwen_vl_interface': 5.000015411117148e-07, 'learning_rate/action_model': 5.0000154111171486e-06, 'epoch': 0.39}) +[2026-06-24 14:43:37,462][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4000 +[2026-06-24 14:44:02,049][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4000 +[2026-06-24 14:44:02,050][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4000 +[2026-06-24 14:44:35,331][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4000 +[2026-06-24 14:54:35,299][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000 +[2026-06-24 14:54:35,300][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.004688928369432688, 'train/grad_norm_pre_clip': 10.738411903381348, 'eval/action_loss/seconds': 24.585160420043394, 'eval/action_loss/samples': 1600.0, 'eval/loss': 0.057327624410390854, 'eval/latency_2/loss': 0.057327624410390854, 'eval/flappy/loss': 0.057327624410390854, 'eval/flappy/latency_2/loss': 0.057327624410390854, 'eval/flap_precision': 0.951935914552737, 'eval/flap_recall': 0.9674355495251018, 'eval/flap_f1': 0.9596231493943472, 'eval/flap_rate_model': 0.20805555555555555, 'eval/flap_rate_teacher': 0.20472222222222222, 'eval/flap_cc_f1': 0.9596231493943473, 'eval/latency_2/flap_precision': 0.951935914552737, 'eval/latency_2/flap_recall': 0.9674355495251018, 'eval/latency_2/flap_f1': 0.9596231493943472, 'eval/latency_2/flap_rate_model': 0.20805555555555555, 'eval/latency_2/flap_rate_teacher': 0.20472222222222222, 'eval/latency_2/flap_cc_f1': 0.9596231493943473, 'eval/action_classification/data_wait_seconds': 0.33202919363975525, 'eval/action_classification/predict_seconds': 32.83888626098633, 'eval/action_classification/frames': 3600.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/mean_length': 3600.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/macro_mean_length': 3600.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 444.60000000002566, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 3600.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 444.60000000002566, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00036158901639282703, 'timing/model': 1.2944383420981467, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 0.39}) +[2026-06-24 14:54:35,300][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_4000_state +[2026-06-24 14:54:35,303][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-24 14:54:38,062][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_4000_state/model.safetensors +[2026-06-24 14:54:46,398][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_4000_state/optimizer.bin +[2026-06-24 14:54:46,398][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_4000_state/sampler.bin +[2026-06-24 14:54:46,398][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-24 14:54:46,399][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-24 14:54:46,432][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-24 14:54:46,432][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-24 14:54:46,443][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-24 14:54:46,443][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/checkpoints +[2026-06-24 14:54:48,953][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..ce7dad56d7abf845dd57d6f02d76df86c6434c49 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/summary.jsonl @@ -0,0 +1,8 @@ +{"steps": 500} +{"steps": 1000} +{"steps": 1500} +{"steps": 2000} +{"steps": 2500} +{"steps": 3000} +{"steps": 3500} +{"steps": 4000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..744c43045cebd5202ce9d5622a85e6889bc43be9 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8afaa32fca8bd02a62fd251a0457e809370c64803c90166b0962241a2089519a +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..983fd4470563dc37f5b6816b1c92e15f01a14e45 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/config.full.yaml @@ -0,0 +1,217 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 64 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_llm_bottom_ratio: 0.7778 + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_last_8_layers +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..d5bf09ba1f47fcacf0e1a845b563dfcc43cfffda --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/config.yaml @@ -0,0 +1,97 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + eval_num_workers: 8 + include_state: true + latency_curriculum: + enabled: false + obs_image_size: null + per_device_batch_size: 64 + prefetch_factor: 4 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers +rl_games: + env_eval: + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 1 + image_size: 224 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: '' + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: flappy +run_id: flappy_fix_latency_2_200ep_last_8_layers +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_action_classification_interval: null + eval_interval: 100 + eval_num_batches: 100 + freeze_llm_bottom_ratio: 0.7778 + freeze_modules: '' + gradient_accumulation_steps: 2 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + reload_modules: null + save_interval: 500 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..aa03e6cbf8e375d660528e17fafddb4804b596b7 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.6028500199317932, + 0.3971499800682068, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4890792667865753, + 0.4890792667865753, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 330734, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..0caa254376d5805f972b0b0411e0b5188e3634e3 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7959861159324646, + 0.2040138840675354, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4030573070049286, + 0.4030573070049286, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..a83d2b2bb815d498b36e16887669c71db0290e68 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/hydra/.hydra/config.yaml @@ -0,0 +1,215 @@ +framework: + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 64 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_llm_bottom_ratio: 0.7778 + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_last_8_layers +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..0f6e939f2d65ebae2056f090b06b0a94b2fd2fdf --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/hydra/.hydra/hydra.yaml @@ -0,0 +1,259 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=flappy_fix_latency_2_200ep_last_8_layers + - run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=saberrr-zju + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=500 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.eval_action_classification=true + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=2 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.freeze_llm_bottom_ratio=0.7778 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9,0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=false + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.latency_curriculum.enabled=false + - datasets.vla_data.latency_curriculum.strategy=exclusive + - datasets.vla_data.per_device_batch_size=64 + - datasets.vla_data.num_workers=8 + - datasets.vla_data.eval_num_workers=8 + - datasets.vla_data.prefetch_factor=4 + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=250 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[2] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.eval_num_workers=8,datasets.vla_data.include_state=true,datasets.vla_data.latency_curriculum.enabled=false,datasets.vla_data.latency_curriculum.strategy=exclusive,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.num_workers=8,datasets.vla_data.per_device_batch_size=64,datasets.vla_data.prefetch_factor=4,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=250,rl_games.env_eval.mid_train.latencies=[2],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=flappy_fix_latency_2_200ep_last_8_layers,run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_action_classification=true,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.freeze_llm_bottom_ratio=0.7778,trainer.gradient_accumulation_steps=2,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=false,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=500,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=saberrr-zju,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/latency-sensitive-bench/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/latency-sensitive-bench/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..4e484f4b8792fdbb32e95dbab15461e87fa10679 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/hydra/.hydra/overrides.yaml @@ -0,0 +1,99 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=flappy_fix_latency_2_200ep_last_8_layers +- run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=saberrr-zju +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=500 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.eval_action_classification=true +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=2 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.freeze_llm_bottom_ratio=0.7778 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9,0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=false +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.latency_curriculum.enabled=false +- datasets.vla_data.latency_curriculum.strategy=exclusive +- datasets.vla_data.per_device_batch_size=64 +- datasets.vla_data.num_workers=8 +- datasets.vla_data.eval_num_workers=8 +- datasets.vla_data.prefetch_factor=4 +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=250 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[2] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..71561c4b225e9ae814721f85518ff640d6c9f4b4 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/hydra/train_starvla_hydra.log @@ -0,0 +1,5375 @@ +[2026-06-23 03:14:42,000][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-23 03:14:42,133][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured. +[2026-06-23 03:14:42,133][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub +[2026-06-23 03:14:42,467][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases +[2026-06-23 03:14:42,510][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-23 03:14:52,813][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-23 03:14:52,886][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-23 03:14:53,504][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-23 03:14:53,504][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-23 03:14:53,504][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-23 03:14:53,504][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-23 03:14:53,505][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-23 03:14:53,505][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-23 03:14:53,505][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-23 03:14:53,505][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-23 03:14:53,506][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-23 03:14:53,506][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-23 03:14:53,506][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-23 03:14:53,506][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-23 03:14:53,506][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-23 03:14:53,507][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-23 03:14:53,507][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-23 03:14:53,507][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-23 03:14:53,507][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-23 03:14:53,507][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-23 03:14:53,508][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-23 03:14:53,508][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-23 03:14:53,508][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-23 03:14:53,508][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-23 03:14:53,509][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-23 03:14:53,509][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-23 03:14:53,509][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-23 03:14:53,509][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-23 03:14:53,509][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-23 03:14:53,510][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-23 03:14:53,510][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-23 03:14:53,510][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-23 03:14:53,510][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-23 03:14:53,570][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-23 03:14:53,618][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/config.full.yaml +[2026-06-23 03:14:53,630][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/config.yaml +[2026-06-23 03:14:53,630][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-23 03:14:56,505][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-23 03:14:56,505][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-23 03:14:56,505][starVLA.training.train_starvla][INFO] - Per device batch size = 64 +[2026-06-23 03:14:56,505][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 2 +[2026-06-23 03:14:56,505][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-23 03:15:55,535][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6946540474891663, 'train/grad_norm_pre_clip': 5.690869331359863, 'timing/data': 0.0013340220029931515, 'timing/model': 1.692872888990678, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-23 03:15:58,814][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6940727829933167, 'train/grad_norm_pre_clip': 3.278801202774048, 'timing/data': 0.0013990829756949097, 'timing/model': 1.6555311649863143, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-23 03:16:02,075][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6897010207176208, 'train/grad_norm_pre_clip': 8.528905868530273, 'timing/data': 0.0014812249864917248, 'timing/model': 1.6538535070139915, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-23 03:16:05,503][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.688062846660614, 'train/grad_norm_pre_clip': 3.1251912117004395, 'timing/data': 0.14608958701137453, 'timing/model': 1.6572901330073364, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 03:16:08,748][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6895344257354736, 'train/grad_norm_pre_clip': 0.4231378436088562, 'timing/data': 0.0011744379880838096, 'timing/model': 1.6439727669931017, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-23 03:16:12,000][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6533840298652649, 'train/grad_norm_pre_clip': 6.288812160491943, 'timing/data': 0.0013675639929715544, 'timing/model': 1.6477262459811755, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-23 03:16:15,266][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.7122907638549805, 'train/grad_norm_pre_clip': 4.9285101890563965, 'timing/data': 0.0011934840003959835, 'timing/model': 1.6607242509780917, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 03:16:18,639][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.6914711594581604, 'train/grad_norm_pre_clip': 2.821364402770996, 'timing/data': 0.12100718700094149, 'timing/model': 1.6465601230156608, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 03:16:21,911][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.6613055467605591, 'train/grad_norm_pre_clip': 0.2619297206401825, 'timing/data': 0.001414161000866443, 'timing/model': 1.6581783010042273, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-23 03:16:25,187][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.7236741185188293, 'train/grad_norm_pre_clip': 6.303442001342773, 'timing/data': 0.0013259380066301674, 'timing/model': 1.6572928450186737, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-23 03:16:28,469][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.6697921752929688, 'train/grad_norm_pre_clip': 1.4490469694137573, 'timing/data': 0.00158922397531569, 'timing/model': 1.659924965002574, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-23 03:16:31,886][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.6765981316566467, 'train/grad_norm_pre_clip': 1.710127592086792, 'timing/data': 0.14373111899476498, 'timing/model': 1.6579156150110066, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-23 03:16:35,177][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.660215437412262, 'train/grad_norm_pre_clip': 0.9736402034759521, 'timing/data': 0.0011513549834489822, 'timing/model': 1.6745742879866157, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-23 03:16:38,458][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.6499855518341064, 'train/grad_norm_pre_clip': 3.0270307064056396, 'timing/data': 0.0011550030030775815, 'timing/model': 1.6604129640036263, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-23 03:16:41,744][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.6545525789260864, 'train/grad_norm_pre_clip': 2.0860559940338135, 'timing/data': 0.0011556130193639547, 'timing/model': 1.6629276260209735, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-23 03:16:45,160][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.6530448794364929, 'train/grad_norm_pre_clip': 1.3499623537063599, 'timing/data': 0.13086258599651046, 'timing/model': 1.661949339992134, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 03:16:48,470][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.6950189471244812, 'train/grad_norm_pre_clip': 4.005676746368408, 'timing/data': 0.001390065997838974, 'timing/model': 1.67612042100518, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 03:16:51,774][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.7763049006462097, 'train/grad_norm_pre_clip': 10.977213859558105, 'timing/data': 0.001170690986327827, 'timing/model': 1.6665750980027951, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-23 03:16:55,088][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.668789803981781, 'train/grad_norm_pre_clip': 2.0056843757629395, 'timing/data': 0.0013230330077931285, 'timing/model': 1.6812283979961649, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-23 03:16:58,532][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.6908458471298218, 'train/grad_norm_pre_clip': 3.588165044784546, 'timing/data': 0.13822918897494674, 'timing/model': 1.6632896910014097, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-23 03:17:01,839][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.6591330170631409, 'train/grad_norm_pre_clip': 1.2484712600708008, 'timing/data': 0.0013422670017462224, 'timing/model': 1.6738053529988974, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-23 03:17:05,139][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.6827541589736938, 'train/grad_norm_pre_clip': 1.292233943939209, 'timing/data': 0.0011191059893462807, 'timing/model': 1.6671052029996645, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 03:17:08,435][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.6603158712387085, 'train/grad_norm_pre_clip': 4.277716159820557, 'timing/data': 0.0012390679912641644, 'timing/model': 1.6681825109990314, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 03:17:11,848][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.6772434711456299, 'train/grad_norm_pre_clip': 1.5877989530563354, 'timing/data': 0.11746156198205426, 'timing/model': 1.6681202960025985, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-23 03:17:15,180][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.6629431843757629, 'train/grad_norm_pre_clip': 4.0169677734375, 'timing/data': 0.0011465870193205774, 'timing/model': 1.702589770022314, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-23 03:17:18,496][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.6776620149612427, 'train/grad_norm_pre_clip': 0.7931178212165833, 'timing/data': 0.0013863590138498694, 'timing/model': 1.6736571750079747, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-23 03:17:21,809][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.6920851469039917, 'train/grad_norm_pre_clip': 3.5200934410095215, 'timing/data': 0.0011625659826677293, 'timing/model': 1.6814848259964492, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-23 03:17:25,276][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.6687462329864502, 'train/grad_norm_pre_clip': 0.9737517237663269, 'timing/data': 0.12185643799602985, 'timing/model': 1.710029982990818, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 03:17:28,583][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.6633320450782776, 'train/grad_norm_pre_clip': 2.8616437911987305, 'timing/data': 0.0012763459817506373, 'timing/model': 1.6750399849843234, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-23 03:17:31,892][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.6530700325965881, 'train/grad_norm_pre_clip': 2.5112853050231934, 'timing/data': 0.0011224019981455058, 'timing/model': 1.6739493100030813, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-23 03:17:35,201][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.6698265075683594, 'train/grad_norm_pre_clip': 1.8806384801864624, 'timing/data': 0.0013001099869143218, 'timing/model': 1.6741603869886603, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-23 03:17:38,644][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.6093916296958923, 'train/grad_norm_pre_clip': 4.341050624847412, 'timing/data': 0.13102776202140376, 'timing/model': 1.677331881975988, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-23 03:17:41,955][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.7063251733779907, 'train/grad_norm_pre_clip': 6.331014156341553, 'timing/data': 0.001341196009889245, 'timing/model': 1.6771640609949827, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-23 03:17:45,272][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.5953876972198486, 'train/grad_norm_pre_clip': 4.0507893562316895, 'timing/data': 0.0014042929979041219, 'timing/model': 1.6804186970111914, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-23 03:17:48,583][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.6765328049659729, 'train/grad_norm_pre_clip': 5.060420036315918, 'timing/data': 0.001271387009182945, 'timing/model': 1.6751986889867112, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01}) +[2026-06-23 03:17:52,026][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.6736423373222351, 'train/grad_norm_pre_clip': 4.595015525817871, 'timing/data': 0.12452672299696133, 'timing/model': 1.6824732790119015, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01}) +[2026-06-23 03:17:55,339][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.6573137044906616, 'train/grad_norm_pre_clip': 3.1962883472442627, 'timing/data': 0.0011560250131879002, 'timing/model': 1.6781833870045375, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.01}) +[2026-06-23 03:17:58,651][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.6146239042282104, 'train/grad_norm_pre_clip': 7.057959079742432, 'timing/data': 0.001329413993516937, 'timing/model': 1.6750944810046349, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.01}) +[2026-06-23 03:18:01,965][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.6527765989303589, 'train/grad_norm_pre_clip': 2.6417551040649414, 'timing/data': 0.0011248269875068218, 'timing/model': 1.6780272939940915, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 03:18:05,415][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.644156813621521, 'train/grad_norm_pre_clip': 3.571077346801758, 'timing/data': 0.1336036929860711, 'timing/model': 1.6823624270036817, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-23 03:18:08,742][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.6307909488677979, 'train/grad_norm_pre_clip': 5.869595050811768, 'timing/data': 0.0011515459918882698, 'timing/model': 1.6821570850152057, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-23 03:18:12,061][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.656738817691803, 'train/grad_norm_pre_clip': 2.2149770259857178, 'timing/data': 0.0011320699995849282, 'timing/model': 1.680907075991854, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-23 03:18:15,386][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.6318204402923584, 'train/grad_norm_pre_clip': 1.9858813285827637, 'timing/data': 0.0012340889952611178, 'timing/model': 1.6838974849961232, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-23 03:18:18,849][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.5671084523200989, 'train/grad_norm_pre_clip': 3.0929293632507324, 'timing/data': 0.13407268200535327, 'timing/model': 1.684583279013168, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 03:18:22,181][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.6221467852592468, 'train/grad_norm_pre_clip': 7.070681571960449, 'timing/data': 0.0011431100137997419, 'timing/model': 1.6895995390077587, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-23 03:18:25,524][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.6800584197044373, 'train/grad_norm_pre_clip': 9.698946952819824, 'timing/data': 0.0013012220151722431, 'timing/model': 1.6993426239932887, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-23 03:18:28,877][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.5798835754394531, 'train/grad_norm_pre_clip': 4.540755748748779, 'timing/data': 0.0011453949846327305, 'timing/model': 1.706788650015369, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-23 03:18:32,352][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.6486585736274719, 'train/grad_norm_pre_clip': 11.162042617797852, 'timing/data': 0.13289989900658838, 'timing/model': 1.6929684720234945, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-23 03:18:35,711][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.6063719987869263, 'train/grad_norm_pre_clip': 11.686315536499023, 'timing/data': 0.0011898969823960215, 'timing/model': 1.7105268680024892, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-23 03:18:39,048][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.6079931259155273, 'train/grad_norm_pre_clip': 18.559528350830078, 'timing/data': 0.0012709370057564229, 'timing/model': 1.6895616280089598, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-23 03:18:42,385][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.6651025414466858, 'train/grad_norm_pre_clip': 22.704862594604492, 'timing/data': 0.0012425339955370873, 'timing/model': 1.6900573909806553, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 03:18:45,855][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.7160358428955078, 'train/grad_norm_pre_clip': 24.352245330810547, 'timing/data': 0.13656230401829816, 'timing/model': 1.6872441250015981, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-23 03:18:49,197][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.5579184293746948, 'train/grad_norm_pre_clip': 7.155640602111816, 'timing/data': 0.001473840995458886, 'timing/model': 1.6973117800080217, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-23 03:18:52,540][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.585448145866394, 'train/grad_norm_pre_clip': 8.81711196899414, 'timing/data': 0.0012992479896638542, 'timing/model': 1.6999488059955183, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-23 03:18:55,883][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.5534517765045166, 'train/grad_norm_pre_clip': 6.083870887756348, 'timing/data': 0.0012411120114848018, 'timing/model': 1.697420925978804, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-23 03:18:59,366][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.5521068572998047, 'train/grad_norm_pre_clip': 13.835371971130371, 'timing/data': 0.1404838569869753, 'timing/model': 1.6895310169784352, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 03:19:02,710][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.6596535444259644, 'train/grad_norm_pre_clip': 22.420127868652344, 'timing/data': 0.001355010987026617, 'timing/model': 1.6901692429964896, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-23 03:19:06,044][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.7154439687728882, 'train/grad_norm_pre_clip': 35.31039047241211, 'timing/data': 0.0011623059981502593, 'timing/model': 1.685672938998323, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-23 03:19:09,381][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.5658001899719238, 'train/grad_norm_pre_clip': 10.474954605102539, 'timing/data': 0.001155182981165126, 'timing/model': 1.685157473984873, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.02}) +[2026-06-23 03:19:12,849][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.4761858582496643, 'train/grad_norm_pre_clip': 17.468658447265625, 'timing/data': 0.1360284429974854, 'timing/model': 1.6875733519846108, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-23 03:19:16,180][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.526817798614502, 'train/grad_norm_pre_clip': 10.289872169494629, 'timing/data': 0.0011616960109677166, 'timing/model': 1.6878542469930835, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.02}) +[2026-06-23 03:19:19,513][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.5264439582824707, 'train/grad_norm_pre_clip': 19.87690544128418, 'timing/data': 0.0011403660173527896, 'timing/model': 1.684725728991907, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.02}) +[2026-06-23 03:19:22,844][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.6279608607292175, 'train/grad_norm_pre_clip': 8.896917343139648, 'timing/data': 0.001188674010336399, 'timing/model': 1.6857053239946254, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.02}) +[2026-06-23 03:19:26,302][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.5181231498718262, 'train/grad_norm_pre_clip': 6.913410186767578, 'timing/data': 0.12895879297866486, 'timing/model': 1.6875255919876508, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.02}) +[2026-06-23 03:19:29,640][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.5033601522445679, 'train/grad_norm_pre_clip': 6.991709232330322, 'timing/data': 0.0012731310271192342, 'timing/model': 1.6911415549984667, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-23 03:19:32,982][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.4236248731613159, 'train/grad_norm_pre_clip': 9.43954849243164, 'timing/data': 0.0012634930026251823, 'timing/model': 1.6928083089878783, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-23 03:19:36,340][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.5944234132766724, 'train/grad_norm_pre_clip': 16.547462463378906, 'timing/data': 0.001299528987146914, 'timing/model': 1.6992743449809495, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-23 03:19:39,821][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.5508242249488831, 'train/grad_norm_pre_clip': 11.141862869262695, 'timing/data': 0.1323330739978701, 'timing/model': 1.6988683630188461, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 03:19:43,164][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.45608583092689514, 'train/grad_norm_pre_clip': 9.540006637573242, 'timing/data': 0.0012744330160785466, 'timing/model': 1.6915446480270475, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-23 03:19:46,502][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.4429157078266144, 'train/grad_norm_pre_clip': 6.12164306640625, 'timing/data': 0.001187472022138536, 'timing/model': 1.6908765910193324, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-23 03:19:49,871][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.47691965103149414, 'train/grad_norm_pre_clip': 8.667813301086426, 'timing/data': 0.002365518012084067, 'timing/model': 1.7064778429921716, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-23 03:19:53,355][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.49300143122673035, 'train/grad_norm_pre_clip': 10.576728820800781, 'timing/data': 0.12979429902043194, 'timing/model': 1.6917398289951961, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-23 03:19:56,714][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.4819239675998688, 'train/grad_norm_pre_clip': 19.581480026245117, 'timing/data': 0.0014628009812440723, 'timing/model': 1.7077685600088444, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-23 03:20:00,055][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.4864063262939453, 'train/grad_norm_pre_clip': 8.928106307983398, 'timing/data': 0.0013290439965203404, 'timing/model': 1.691159620997496, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-23 03:20:03,416][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.4404370188713074, 'train/grad_norm_pre_clip': 7.724667549133301, 'timing/data': 0.0011762119829654694, 'timing/model': 1.7051077309879474, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-23 03:20:06,904][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.437355101108551, 'train/grad_norm_pre_clip': 7.391083240509033, 'timing/data': 0.13480490300571546, 'timing/model': 1.7022317980008665, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-23 03:20:10,262][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.5206476449966431, 'train/grad_norm_pre_clip': 21.950496673583984, 'timing/data': 0.0011840169900096953, 'timing/model': 1.7008996230142657, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-23 03:20:13,610][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.6329012513160706, 'train/grad_norm_pre_clip': 27.227455139160156, 'timing/data': 0.0011726949887815863, 'timing/model': 1.6977538330247626, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 03:20:16,967][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.42893895506858826, 'train/grad_norm_pre_clip': 20.980426788330078, 'timing/data': 0.0012170280097052455, 'timing/model': 1.7086641430214513, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.03}) +[2026-06-23 03:20:20,462][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.562582790851593, 'train/grad_norm_pre_clip': 17.198448181152344, 'timing/data': 0.15042103800806217, 'timing/model': 1.6962911339942366, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-23 03:20:23,809][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.609701931476593, 'train/grad_norm_pre_clip': 32.80324935913086, 'timing/data': 0.0013242150016594678, 'timing/model': 1.6950442380039021, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.03}) +[2026-06-23 03:20:27,165][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.47967079281806946, 'train/grad_norm_pre_clip': 41.47761535644531, 'timing/data': 0.0011834449833258986, 'timing/model': 1.702545314998133, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.03}) +[2026-06-23 03:20:30,518][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.4052700996398926, 'train/grad_norm_pre_clip': 11.01090145111084, 'timing/data': 0.0011526379967108369, 'timing/model': 1.6946717739920132, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.03}) +[2026-06-23 03:20:33,988][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.3446905016899109, 'train/grad_norm_pre_clip': 11.139193534851074, 'timing/data': 0.13487542601069435, 'timing/model': 1.686133465002058, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.03}) +[2026-06-23 03:20:37,339][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.44456547498703003, 'train/grad_norm_pre_clip': 11.746562957763672, 'timing/data': 0.0011810810246970505, 'timing/model': 1.7041940380004235, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.03}) +[2026-06-23 03:20:40,673][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.476875901222229, 'train/grad_norm_pre_clip': 35.92345428466797, 'timing/data': 0.0011505449947435409, 'timing/model': 1.6875066519714892, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.03}) +[2026-06-23 03:20:44,025][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.5105422735214233, 'train/grad_norm_pre_clip': 38.61029052734375, 'timing/data': 0.0012908429780509323, 'timing/model': 1.7048726930224802, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.03}) +[2026-06-23 03:20:47,484][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.4370846748352051, 'train/grad_norm_pre_clip': 8.5381498336792, 'timing/data': 0.12583681000978686, 'timing/model': 1.6877911199990194, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 03:20:50,827][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.3945085406303406, 'train/grad_norm_pre_clip': 23.465312957763672, 'timing/data': 0.0011430610029492527, 'timing/model': 1.6932693910202943, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.03}) +[2026-06-23 03:20:54,169][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.4545430839061737, 'train/grad_norm_pre_clip': 21.898609161376953, 'timing/data': 0.0012528130027931184, 'timing/model': 1.691044890001649, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-23 03:20:57,515][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.4789391756057739, 'train/grad_norm_pre_clip': 9.310128211975098, 'timing/data': 0.001216586009832099, 'timing/model': 1.6948862379940692, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-23 03:21:00,986][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.5341721177101135, 'train/grad_norm_pre_clip': 20.978382110595703, 'timing/data': 0.12641948700184003, 'timing/model': 1.6931020590127446, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-23 03:21:04,334][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.3718561828136444, 'train/grad_norm_pre_clip': 14.474834442138672, 'timing/data': 0.0013546519912779331, 'timing/model': 1.6966701039928012, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-23 03:21:07,690][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.3671553432941437, 'train/grad_norm_pre_clip': 12.61278247833252, 'timing/data': 0.0011500130058266222, 'timing/model': 1.706268983980408, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-23 03:21:11,032][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.4055436849594116, 'train/grad_norm_pre_clip': 9.454019546508789, 'timing/data': 0.0012053760001435876, 'timing/model': 1.6941132639767602, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-23 03:21:14,508][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.40769028663635254, 'train/grad_norm_pre_clip': 12.339629173278809, 'timing/data': 0.13555142100085504, 'timing/model': 1.6912132560100872, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-23 03:21:17,849][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.3312327265739441, 'train/grad_norm_pre_clip': 15.626895904541016, 'timing/data': 0.001326729980064556, 'timing/model': 1.690922073990805, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-23 03:21:21,193][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.49632585048675537, 'train/grad_norm_pre_clip': 45.565216064453125, 'timing/data': 0.0013570160081144422, 'timing/model': 1.6929725350055378, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-23 03:21:24,536][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.45221856236457825, 'train/grad_norm_pre_clip': 29.60688591003418, 'timing/data': 0.0012037530250381678, 'timing/model': 1.693122968019452, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-23 03:21:28,016][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 100 +[2026-06-23 03:24:23,393][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 100 +[2026-06-23 03:24:23,393][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 100 +[2026-06-23 03:24:23,909][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmpnmorv1h1/test.c -o /tmp/tmpnmorv1h1/test.o +[2026-06-23 03:24:23,934][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmpnmorv1h1/test.o -laio -o /tmp/tmpnmorv1h1/a.out +[2026-06-23 03:24:24,460][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmpjizslq4j/test.c -o /tmp/tmpjizslq4j/test.o +[2026-06-23 03:24:24,493][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmpjizslq4j/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpjizslq4j/a.out +[2026-06-23 03:24:24,557][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmprqv6pc_p/test.c -o /tmp/tmprqv6pc_p/test.o +[2026-06-23 03:24:24,580][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmprqv6pc_p/test.o -laio -o /tmp/tmprqv6pc_p/a.out +[2026-06-23 03:26:41,291][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 100 +[2026-06-23 03:26:41,293][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.62533038854599, 'train/grad_norm_pre_clip': 55.78997802734375, 'eval/action_loss/seconds': 175.37346896302188, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.2962167263031006, 'eval/latency_2/loss': 0.2962167263031006, 'eval/flappy/loss': 0.2962167263031006, 'eval/flappy/latency_2/loss': 0.2962167263031006, 'eval/flap_precision': 0.7635658914728682, 'eval/flap_recall': 0.39878542510121456, 'eval/flap_f1': 0.523936170212766, 'eval/flap_rate_model': 0.1075, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.523936170212766, 'eval/latency_2/flap_precision': 0.7635658914728682, 'eval/latency_2/flap_recall': 0.39878542510121456, 'eval/latency_2/flap_f1': 0.523936170212766, 'eval/latency_2/flap_rate_model': 0.1075, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.523936170212766, 'eval/action_classification/data_wait_seconds': 49.50204849243164, 'eval/action_classification/predict_seconds': 84.97298431396484, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.13121829100418836, 'timing/model': 1.698451912991004, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.04}) +[2026-06-23 03:26:45,427][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.5739618539810181, 'train/grad_norm_pre_clip': 71.00418853759766, 'timing/data': 0.0013161599927116185, 'timing/model': 2.078004549985053, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.04}) +[2026-06-23 03:26:49,544][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.4708271622657776, 'train/grad_norm_pre_clip': 38.48441696166992, 'timing/data': 0.0013604220002889633, 'timing/model': 2.0752460900112055, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.04}) +[2026-06-23 03:26:53,658][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.438229501247406, 'train/grad_norm_pre_clip': 23.457839965820312, 'timing/data': 0.001198612997541204, 'timing/model': 2.0786325680091977, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.04}) +[2026-06-23 03:26:57,920][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.5991865396499634, 'train/grad_norm_pre_clip': 23.051145553588867, 'timing/data': 0.14558452399796806, 'timing/model': 2.0791342800075654, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.04}) +[2026-06-23 03:27:02,045][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.6855494379997253, 'train/grad_norm_pre_clip': 50.748043060302734, 'timing/data': 0.0013652619963977486, 'timing/model': 2.0765455630025826, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.04}) +[2026-06-23 03:27:06,185][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.4306260943412781, 'train/grad_norm_pre_clip': 18.573293685913086, 'timing/data': 0.0014315939915832132, 'timing/model': 2.09647393898922, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.04}) +[2026-06-23 03:27:10,329][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.3680110573768616, 'train/grad_norm_pre_clip': 24.434589385986328, 'timing/data': 0.001294739980949089, 'timing/model': 2.0858671159949154, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.04}) +[2026-06-23 03:27:14,610][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.3377796411514282, 'train/grad_norm_pre_clip': 23.879995346069336, 'timing/data': 0.13872330397134647, 'timing/model': 2.091838888998609, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.04}) +[2026-06-23 03:27:18,835][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.5036349296569824, 'train/grad_norm_pre_clip': 35.343299865722656, 'timing/data': 0.0012791619810741395, 'timing/model': 2.107312729000114, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.04}) +[2026-06-23 03:27:22,979][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.5070408582687378, 'train/grad_norm_pre_clip': 20.184589385986328, 'timing/data': 0.001247623993549496, 'timing/model': 2.082652613025857, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.04}) +[2026-06-23 03:27:27,112][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.391929030418396, 'train/grad_norm_pre_clip': 7.668471336364746, 'timing/data': 0.0012855340028181672, 'timing/model': 2.0878146899922285, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.04}) +[2026-06-23 03:27:31,412][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.3722255825996399, 'train/grad_norm_pre_clip': 6.246912002563477, 'timing/data': 0.1447100609948393, 'timing/model': 2.0880708170006983, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.04}) +[2026-06-23 03:27:35,540][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.46256113052368164, 'train/grad_norm_pre_clip': 7.46287202835083, 'timing/data': 0.0013676559901796281, 'timing/model': 2.0853924360126257, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.04}) +[2026-06-23 03:27:39,668][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.38970112800598145, 'train/grad_norm_pre_clip': 8.06683349609375, 'timing/data': 0.0012617809989023954, 'timing/model': 2.0859325069759507, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.04}) +[2026-06-23 03:27:43,804][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.505305826663971, 'train/grad_norm_pre_clip': 61.63502883911133, 'timing/data': 0.0012357320229057223, 'timing/model': 2.0912694169965107, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.04}) +[2026-06-23 03:27:48,092][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.40974757075309753, 'train/grad_norm_pre_clip': 15.061563491821289, 'timing/data': 0.14136812000651844, 'timing/model': 2.0910420909931418, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.04}) +[2026-06-23 03:27:52,486][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.3867482841014862, 'train/grad_norm_pre_clip': 11.650391578674316, 'timing/data': 0.0011814119934570044, 'timing/model': 2.3508528280071914, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.05}) +[2026-06-23 03:27:56,625][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.43172022700309753, 'train/grad_norm_pre_clip': 49.44468307495117, 'timing/data': 0.001411286008078605, 'timing/model': 2.089742098993156, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.05}) +[2026-06-23 03:28:00,782][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.5702276229858398, 'train/grad_norm_pre_clip': 59.978668212890625, 'timing/data': 0.0014562990108970553, 'timing/model': 2.101524691999657, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.05}) +[2026-06-23 03:28:05,086][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.33818483352661133, 'train/grad_norm_pre_clip': 16.38658905029297, 'timing/data': 0.15242026199121028, 'timing/model': 2.0974918580031954, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.05}) +[2026-06-23 03:28:09,245][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.39502477645874023, 'train/grad_norm_pre_clip': 20.775924682617188, 'timing/data': 0.0012448089837562293, 'timing/model': 2.1035466239845846, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.05}) +[2026-06-23 03:28:13,409][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.38957035541534424, 'train/grad_norm_pre_clip': 15.950684547424316, 'timing/data': 0.001253263995749876, 'timing/model': 2.106126011000015, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.05}) +[2026-06-23 03:28:17,558][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.5161951184272766, 'train/grad_norm_pre_clip': 18.951221466064453, 'timing/data': 0.0016323640011250973, 'timing/model': 2.0917238690017257, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.05}) +[2026-06-23 03:28:21,857][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.5057390928268433, 'train/grad_norm_pre_clip': 52.721763610839844, 'timing/data': 0.15688468801090494, 'timing/model': 2.0909345239924733, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.05}) +[2026-06-23 03:28:26,108][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.5334916710853577, 'train/grad_norm_pre_clip': 28.36347198486328, 'timing/data': 0.0012086419737897813, 'timing/model': 2.1997993150143884, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.05}) +[2026-06-23 03:28:30,247][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.4091947376728058, 'train/grad_norm_pre_clip': 30.40273094177246, 'timing/data': 0.0014242810138966888, 'timing/model': 2.0902671590156388, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.05}) +[2026-06-23 03:28:34,391][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.34048953652381897, 'train/grad_norm_pre_clip': 29.254005432128906, 'timing/data': 0.0014706359943374991, 'timing/model': 2.092438513005618, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.05}) +[2026-06-23 03:28:38,691][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.45680370926856995, 'train/grad_norm_pre_clip': 45.60117721557617, 'timing/data': 0.15237963901017793, 'timing/model': 2.0947619680082425, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.05}) +[2026-06-23 03:28:42,838][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.458543598651886, 'train/grad_norm_pre_clip': 26.495176315307617, 'timing/data': 0.001331850013229996, 'timing/model': 2.0947153650049586, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.05}) +[2026-06-23 03:28:46,975][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.3160853087902069, 'train/grad_norm_pre_clip': 10.932718276977539, 'timing/data': 0.0012212349975015968, 'timing/model': 2.08866846197634, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.05}) +[2026-06-23 03:28:51,136][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.4179944396018982, 'train/grad_norm_pre_clip': 18.981054306030273, 'timing/data': 0.0014229679945856333, 'timing/model': 2.101023202005308, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.05}) +[2026-06-23 03:28:55,424][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.3195324242115021, 'train/grad_norm_pre_clip': 12.723320007324219, 'timing/data': 0.15115551897906698, 'timing/model': 2.0871608939778525, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.05}) +[2026-06-23 03:28:59,622][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.39109891653060913, 'train/grad_norm_pre_clip': 9.17373275756836, 'timing/data': 0.0012366540031507611, 'timing/model': 2.127906735986471, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.05}) +[2026-06-23 03:29:04,028][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.2573736310005188, 'train/grad_norm_pre_clip': 9.814302444458008, 'timing/data': 0.0012380970001686364, 'timing/model': 2.103519372001756, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.05}) +[2026-06-23 03:29:08,213][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.39045920968055725, 'train/grad_norm_pre_clip': 50.58644485473633, 'timing/data': 0.0012217669864185154, 'timing/model': 2.1250418330018874, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.05}) +[2026-06-23 03:29:12,517][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.38193395733833313, 'train/grad_norm_pre_clip': 17.795156478881836, 'timing/data': 0.14833104400895536, 'timing/model': 2.0973503660061397, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.05}) +[2026-06-23 03:29:16,671][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.4481634497642517, 'train/grad_norm_pre_clip': 64.21119689941406, 'timing/data': 0.0012721999955829233, 'timing/model': 2.099340074986685, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.05}) +[2026-06-23 03:29:20,835][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.4243800938129425, 'train/grad_norm_pre_clip': 51.0115852355957, 'timing/data': 0.0013385109778027982, 'timing/model': 2.102744233998237, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.05}) +[2026-06-23 03:29:24,994][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.5307707786560059, 'train/grad_norm_pre_clip': 44.641265869140625, 'timing/data': 0.0016446970112156123, 'timing/model': 2.1011697400244884, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.05}) +[2026-06-23 03:29:29,297][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.4145594835281372, 'train/grad_norm_pre_clip': 42.40560531616211, 'timing/data': 0.1442996010009665, 'timing/model': 2.101648147974629, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.05}) +[2026-06-23 03:29:33,461][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.36324554681777954, 'train/grad_norm_pre_clip': 30.751192092895508, 'timing/data': 0.0014577120018657297, 'timing/model': 2.103971692005871, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.05}) +[2026-06-23 03:29:37,661][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.5916679501533508, 'train/grad_norm_pre_clip': 42.18558883666992, 'timing/data': 0.0013296350080054253, 'timing/model': 2.0997141919797286, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.05}) +[2026-06-23 03:29:41,831][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.3213268220424652, 'train/grad_norm_pre_clip': 28.301488876342773, 'timing/data': 0.0013883640058338642, 'timing/model': 2.108336679986678, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.06}) +[2026-06-23 03:29:46,129][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.5034498572349548, 'train/grad_norm_pre_clip': 88.49525451660156, 'timing/data': 0.13835260798805393, 'timing/model': 2.0989007970201783, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.06}) +[2026-06-23 03:29:50,316][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.49285683035850525, 'train/grad_norm_pre_clip': 83.16067504882812, 'timing/data': 0.001248525979463011, 'timing/model': 2.1162693760124967, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.06}) +[2026-06-23 03:29:54,484][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.7551605701446533, 'train/grad_norm_pre_clip': 111.14933013916016, 'timing/data': 0.001494058989919722, 'timing/model': 2.0987586910196114, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.06}) +[2026-06-23 03:29:58,654][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.35596826672554016, 'train/grad_norm_pre_clip': 28.487594604492188, 'timing/data': 0.0012750850000884384, 'timing/model': 2.109874874004163, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.06}) +[2026-06-23 03:30:02,955][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.4294176399707794, 'train/grad_norm_pre_clip': 14.349746704101562, 'timing/data': 0.1430174930137582, 'timing/model': 2.0989273489976767, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.06}) +[2026-06-23 03:30:07,130][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.712222695350647, 'train/grad_norm_pre_clip': 78.93864440917969, 'timing/data': 0.001415653998265043, 'timing/model': 2.1150847710086964, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.06}) +[2026-06-23 03:30:11,476][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.5508127212524414, 'train/grad_norm_pre_clip': 45.12153625488281, 'timing/data': 0.0012596370070241392, 'timing/model': 2.2873974920075852, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.06}) +[2026-06-23 03:30:15,653][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.4529566764831543, 'train/grad_norm_pre_clip': 45.40914535522461, 'timing/data': 0.001468241011025384, 'timing/model': 2.1039551029971335, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.06}) +[2026-06-23 03:30:19,957][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.4202895760536194, 'train/grad_norm_pre_clip': 30.399688720703125, 'timing/data': 0.14513975399313495, 'timing/model': 2.0983429120096844, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.06}) +[2026-06-23 03:30:24,128][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.4080824553966522, 'train/grad_norm_pre_clip': 24.112459182739258, 'timing/data': 0.0012473429960664362, 'timing/model': 2.1165952229930554, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.06}) +[2026-06-23 03:30:28,300][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.37242093682289124, 'train/grad_norm_pre_clip': 16.202207565307617, 'timing/data': 0.0013997950009070337, 'timing/model': 2.1045580150093883, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.06}) +[2026-06-23 03:30:32,465][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.3412550091743469, 'train/grad_norm_pre_clip': 11.002849578857422, 'timing/data': 0.0011847380083054304, 'timing/model': 2.1043497739883605, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.06}) +[2026-06-23 03:30:36,765][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.48708587884902954, 'train/grad_norm_pre_clip': 41.3165168762207, 'timing/data': 0.137922706984682, 'timing/model': 2.1023215519962832, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.06}) +[2026-06-23 03:30:40,928][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.5169931054115295, 'train/grad_norm_pre_clip': 43.393497467041016, 'timing/data': 0.0011519880208652467, 'timing/model': 2.103100251988508, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.06}) +[2026-06-23 03:30:45,112][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.5133867859840393, 'train/grad_norm_pre_clip': 51.69957733154297, 'timing/data': 0.0015745980199426413, 'timing/model': 2.1218339029874187, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.06}) +[2026-06-23 03:30:49,315][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.41042304039001465, 'train/grad_norm_pre_clip': 19.177499771118164, 'timing/data': 0.0011826340050902218, 'timing/model': 2.1045060559990816, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.06}) +[2026-06-23 03:30:53,621][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.4199106693267822, 'train/grad_norm_pre_clip': 38.85249328613281, 'timing/data': 0.14617249299772084, 'timing/model': 2.101832991989795, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.06}) +[2026-06-23 03:30:57,776][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.4359363913536072, 'train/grad_norm_pre_clip': 45.302490234375, 'timing/data': 0.0011827150010503829, 'timing/model': 2.099863567011198, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.06}) +[2026-06-23 03:31:01,936][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.4838603138923645, 'train/grad_norm_pre_clip': 30.539520263671875, 'timing/data': 0.0013760809961240739, 'timing/model': 2.1001412330078892, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.06}) +[2026-06-23 03:31:06,099][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.38675248622894287, 'train/grad_norm_pre_clip': 21.442344665527344, 'timing/data': 0.0011669559753499925, 'timing/model': 2.1026855540112592, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.06}) +[2026-06-23 03:31:10,414][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.47404855489730835, 'train/grad_norm_pre_clip': 35.94624328613281, 'timing/data': 0.15135396298137493, 'timing/model': 2.099282152019441, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.06}) +[2026-06-23 03:31:14,576][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.352396160364151, 'train/grad_norm_pre_clip': 20.924699783325195, 'timing/data': 0.001243106002220884, 'timing/model': 2.1041912799992133, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.06}) +[2026-06-23 03:31:18,763][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.2909809648990631, 'train/grad_norm_pre_clip': 21.08481788635254, 'timing/data': 0.0012608190008904785, 'timing/model': 2.12101675698068, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.06}) +[2026-06-23 03:31:23,183][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.4547687768936157, 'train/grad_norm_pre_clip': 26.3210506439209, 'timing/data': 0.0011405269906390458, 'timing/model': 2.354516332008643, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.06}) +[2026-06-23 03:31:27,503][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.31149429082870483, 'train/grad_norm_pre_clip': 21.522363662719727, 'timing/data': 0.14675342699047178, 'timing/model': 2.1053272470016964, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.07}) +[2026-06-23 03:31:31,678][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.49224308133125305, 'train/grad_norm_pre_clip': 20.71112060546875, 'timing/data': 0.0020448360010050237, 'timing/model': 2.108882827014895, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.07}) +[2026-06-23 03:31:35,856][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.2669702172279358, 'train/grad_norm_pre_clip': 13.330284118652344, 'timing/data': 0.0016469719994347543, 'timing/model': 2.10673057398526, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.07}) +[2026-06-23 03:31:40,034][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.2861214876174927, 'train/grad_norm_pre_clip': 14.222702026367188, 'timing/data': 0.0013011330156587064, 'timing/model': 2.110102603008272, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.07}) +[2026-06-23 03:31:44,351][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.352924644947052, 'train/grad_norm_pre_clip': 16.576866149902344, 'timing/data': 0.14252847598982044, 'timing/model': 2.106905441993149, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.07}) +[2026-06-23 03:31:48,523][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.3326701521873474, 'train/grad_norm_pre_clip': 26.13546371459961, 'timing/data': 0.0013853679993189871, 'timing/model': 2.1047686880046967, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.07}) +[2026-06-23 03:31:52,698][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.3205994665622711, 'train/grad_norm_pre_clip': 14.171976089477539, 'timing/data': 0.001341718016192317, 'timing/model': 2.1032760239904746, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.07}) +[2026-06-23 03:31:57,072][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.31901711225509644, 'train/grad_norm_pre_clip': 32.27388000488281, 'timing/data': 0.0014171579969115555, 'timing/model': 2.2885761549987365, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.07}) +[2026-06-23 03:32:01,406][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.2824622392654419, 'train/grad_norm_pre_clip': 15.596809387207031, 'timing/data': 0.16888125499826856, 'timing/model': 2.103047778015025, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.07}) +[2026-06-23 03:32:05,591][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.2784101665019989, 'train/grad_norm_pre_clip': 14.665180206298828, 'timing/data': 0.001243967009941116, 'timing/model': 2.1116961230000015, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.07}) +[2026-06-23 03:32:09,787][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.37066492438316345, 'train/grad_norm_pre_clip': 109.34358215332031, 'timing/data': 0.0012442780134733766, 'timing/model': 2.134497514984105, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.07}) +[2026-06-23 03:32:14,009][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.3622109293937683, 'train/grad_norm_pre_clip': 38.39094924926758, 'timing/data': 0.001406276976922527, 'timing/model': 2.147528342989972, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.07}) +[2026-06-23 03:32:18,329][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.443899929523468, 'train/grad_norm_pre_clip': 31.084558486938477, 'timing/data': 0.1529680170060601, 'timing/model': 2.103733313997509, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.07}) +[2026-06-23 03:32:22,507][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.36217787861824036, 'train/grad_norm_pre_clip': 26.702224731445312, 'timing/data': 0.0013406760117504746, 'timing/model': 2.1108230089885183, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.07}) +[2026-06-23 03:32:26,688][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.3835277259349823, 'train/grad_norm_pre_clip': 41.767093658447266, 'timing/data': 0.0012213659938424826, 'timing/model': 2.1097972370043863, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.07}) +[2026-06-23 03:32:30,884][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.3068927824497223, 'train/grad_norm_pre_clip': 40.38475799560547, 'timing/data': 0.0014377649931702763, 'timing/model': 2.1272062610078137, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.07}) +[2026-06-23 03:32:35,426][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.42610737681388855, 'train/grad_norm_pre_clip': 23.738380432128906, 'timing/data': 0.1383070010051597, 'timing/model': 2.1023568410018925, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.07}) +[2026-06-23 03:32:39,591][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.39810946583747864, 'train/grad_norm_pre_clip': 49.066505432128906, 'timing/data': 0.0012875180109404027, 'timing/model': 2.104466800024966, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.07}) +[2026-06-23 03:32:43,773][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.28606775403022766, 'train/grad_norm_pre_clip': 36.37653350830078, 'timing/data': 0.0015988130180630833, 'timing/model': 2.1160397169878706, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.07}) +[2026-06-23 03:32:47,947][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.32223790884017944, 'train/grad_norm_pre_clip': 15.61406421661377, 'timing/data': 0.001136919017881155, 'timing/model': 2.1152722609986085, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.07}) +[2026-06-23 03:32:52,266][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.2418103814125061, 'train/grad_norm_pre_clip': 26.36263656616211, 'timing/data': 0.14005555101903155, 'timing/model': 2.120935060986085, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.07}) +[2026-06-23 03:32:56,431][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.45093369483947754, 'train/grad_norm_pre_clip': 35.41566467285156, 'timing/data': 0.0014801130164414644, 'timing/model': 2.104363388993079, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.07}) +[2026-06-23 03:33:00,620][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.3095104396343231, 'train/grad_norm_pre_clip': 36.31117630004883, 'timing/data': 0.0013026759843342006, 'timing/model': 2.113351230014814, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.07}) +[2026-06-23 03:33:04,820][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.2703937292098999, 'train/grad_norm_pre_clip': 19.59630584716797, 'timing/data': 0.0012999610044062138, 'timing/model': 2.131456178991357, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.07}) +[2026-06-23 03:33:09,242][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.44121962785720825, 'train/grad_norm_pre_clip': 43.203857421875, 'timing/data': 0.15094194700941443, 'timing/model': 2.112639159982791, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.07}) +[2026-06-23 03:33:13,407][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.29565975069999695, 'train/grad_norm_pre_clip': 26.59395408630371, 'timing/data': 0.0014165460015647113, 'timing/model': 2.1043973510095384, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.07}) +[2026-06-23 03:33:17,607][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.3821091949939728, 'train/grad_norm_pre_clip': 66.29933166503906, 'timing/data': 0.0013155299820937216, 'timing/model': 2.1387391969910823, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.08}) +[2026-06-23 03:33:21,786][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.3813056945800781, 'train/grad_norm_pre_clip': 38.78237533569336, 'timing/data': 0.0012984779896214604, 'timing/model': 2.106326187000377, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.08}) +[2026-06-23 03:33:26,119][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.38484352827072144, 'train/grad_norm_pre_clip': 30.2950382232666, 'timing/data': 0.15021635798620991, 'timing/model': 2.1164069660007954, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.08}) +[2026-06-23 03:33:30,296][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.2842094302177429, 'train/grad_norm_pre_clip': 19.2183895111084, 'timing/data': 0.0012603880022652447, 'timing/model': 2.105305667006178, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.08}) +[2026-06-23 03:33:34,459][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.30918100476264954, 'train/grad_norm_pre_clip': 14.04293441772461, 'timing/data': 0.0014028410078026354, 'timing/model': 2.1019807280099485, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.08}) +[2026-06-23 03:33:38,627][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.3220195770263672, 'train/grad_norm_pre_clip': 35.327545166015625, 'timing/data': 0.0012792930065188557, 'timing/model': 2.104578547994606, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.08}) +[2026-06-23 03:33:42,978][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 200 +[2026-06-23 03:35:40,247][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 200 +[2026-06-23 03:35:40,248][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 200 +[2026-06-23 03:37:58,024][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 200 +[2026-06-23 03:37:58,026][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.41189444065093994, 'train/grad_norm_pre_clip': 56.909515380859375, 'eval/action_loss/seconds': 117.26501721399836, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.17774564027786255, 'eval/latency_2/loss': 0.17774564027786255, 'eval/flappy/loss': 0.17774564027786255, 'eval/flappy/latency_2/loss': 0.17774564027786255, 'eval/flap_precision': 0.7333682556312205, 'eval/flap_recall': 0.9446693657219973, 'eval/flap_f1': 0.8257151282807431, 'eval/flap_rate_model': 0.26513888888888887, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8257151282807431, 'eval/latency_2/flap_precision': 0.7333682556312205, 'eval/latency_2/flap_recall': 0.9446693657219973, 'eval/latency_2/flap_f1': 0.8257151282807431, 'eval/latency_2/flap_rate_model': 0.26513888888888887, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8257151282807431, 'eval/action_classification/data_wait_seconds': 50.36199188232422, 'eval/action_classification/predict_seconds': 85.65563201904297, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.14437681398703717, 'timing/model': 2.1472985379805323, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.08}) +[2026-06-23 03:38:02,179][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.26913565397262573, 'train/grad_norm_pre_clip': 13.985020637512207, 'timing/data': 0.0011908790038432926, 'timing/model': 2.077998608001508, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.08}) +[2026-06-23 03:38:06,300][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.3548406660556793, 'train/grad_norm_pre_clip': 31.597719192504883, 'timing/data': 0.0012194010196253657, 'timing/model': 2.083970078994753, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.08}) +[2026-06-23 03:38:10,448][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.38455021381378174, 'train/grad_norm_pre_clip': 31.046918869018555, 'timing/data': 0.0013522670196834952, 'timing/model': 2.1035105389892124, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.08}) +[2026-06-23 03:38:14,731][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.33944666385650635, 'train/grad_norm_pre_clip': 28.136629104614258, 'timing/data': 0.14699049800401554, 'timing/model': 2.0834904449875467, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.08}) +[2026-06-23 03:38:18,865][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.205766960978508, 'train/grad_norm_pre_clip': 16.921594619750977, 'timing/data': 0.001187984977150336, 'timing/model': 2.0918717260065023, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.08}) +[2026-06-23 03:38:23,258][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.2594090402126312, 'train/grad_norm_pre_clip': 34.52322769165039, 'timing/data': 0.0012289389851503074, 'timing/model': 2.1045916599978227, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.08}) +[2026-06-23 03:38:27,397][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.17134453356266022, 'train/grad_norm_pre_clip': 10.770254135131836, 'timing/data': 0.001223941013449803, 'timing/model': 2.089277327002492, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.08}) +[2026-06-23 03:38:31,667][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.3984368145465851, 'train/grad_norm_pre_clip': 56.060787200927734, 'timing/data': 0.1418396350054536, 'timing/model': 2.084442447987385, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.08}) +[2026-06-23 03:38:35,797][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.21092581748962402, 'train/grad_norm_pre_clip': 13.689922332763672, 'timing/data': 0.0013811110111419111, 'timing/model': 2.0843443979974836, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.08}) +[2026-06-23 03:38:39,930][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.3775717318058014, 'train/grad_norm_pre_clip': 28.455669403076172, 'timing/data': 0.0014342380163725466, 'timing/model': 2.0882821019913536, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.08}) +[2026-06-23 03:38:44,081][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.2778902053833008, 'train/grad_norm_pre_clip': 43.22579574584961, 'timing/data': 0.0013974109897390008, 'timing/model': 2.1035881119896658, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.08}) +[2026-06-23 03:38:48,370][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.39331141114234924, 'train/grad_norm_pre_clip': 28.207901000976562, 'timing/data': 0.15040336200036108, 'timing/model': 2.0899709229997825, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.08}) +[2026-06-23 03:38:52,513][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.28356871008872986, 'train/grad_norm_pre_clip': 27.368928909301758, 'timing/data': 0.001215163996675983, 'timing/model': 2.095515083987266, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.08}) +[2026-06-23 03:38:56,911][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.3987213969230652, 'train/grad_norm_pre_clip': 41.793025970458984, 'timing/data': 0.0014561190037056804, 'timing/model': 2.0911894059972838, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.08}) +[2026-06-23 03:39:01,066][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.48193126916885376, 'train/grad_norm_pre_clip': 80.90669250488281, 'timing/data': 0.0012363129935692996, 'timing/model': 2.09295283301617, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.08}) +[2026-06-23 03:39:05,358][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.5806555151939392, 'train/grad_norm_pre_clip': 68.5583267211914, 'timing/data': 0.1554157700156793, 'timing/model': 2.0878609310020693, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.08}) +[2026-06-23 03:39:09,502][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.38752809166908264, 'train/grad_norm_pre_clip': 82.83682250976562, 'timing/data': 0.00124239397700876, 'timing/model': 2.0931454520032275, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.08}) +[2026-06-23 03:39:13,644][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.2660897672176361, 'train/grad_norm_pre_clip': 26.79680633544922, 'timing/data': 0.001695531012956053, 'timing/model': 2.0934967920184135, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.08}) +[2026-06-23 03:39:17,804][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.41994762420654297, 'train/grad_norm_pre_clip': 35.63503646850586, 'timing/data': 0.001214523013914004, 'timing/model': 2.1062162239977624, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.08}) +[2026-06-23 03:39:22,096][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.3864882290363312, 'train/grad_norm_pre_clip': 54.9657096862793, 'timing/data': 0.15169657999649644, 'timing/model': 2.0914627769961953, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.09}) +[2026-06-23 03:39:26,240][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.2544434070587158, 'train/grad_norm_pre_clip': 23.259967803955078, 'timing/data': 0.0012563509808387607, 'timing/model': 2.093590138014406, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.09}) +[2026-06-23 03:39:30,501][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.23019781708717346, 'train/grad_norm_pre_clip': 20.118404388427734, 'timing/data': 0.0012354410137049854, 'timing/model': 2.1117744230141398, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.09}) +[2026-06-23 03:39:34,658][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.31917259097099304, 'train/grad_norm_pre_clip': 47.64414978027344, 'timing/data': 0.0012115680146962404, 'timing/model': 2.1033161020022817, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.09}) +[2026-06-23 03:39:38,959][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.23423618078231812, 'train/grad_norm_pre_clip': 23.42843246459961, 'timing/data': 0.15357707400107756, 'timing/model': 2.0950386059994344, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.09}) +[2026-06-23 03:39:43,111][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.5035431981086731, 'train/grad_norm_pre_clip': 20.89727783203125, 'timing/data': 0.0012160059995949268, 'timing/model': 2.100338104995899, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.09}) +[2026-06-23 03:39:47,274][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.2582487165927887, 'train/grad_norm_pre_clip': 39.8514404296875, 'timing/data': 0.0013403850025497377, 'timing/model': 2.1073535929899663, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.09}) +[2026-06-23 03:39:51,453][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.27444368600845337, 'train/grad_norm_pre_clip': 27.792625427246094, 'timing/data': 0.0012310130114201456, 'timing/model': 2.111151888995664, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.09}) +[2026-06-23 03:39:55,763][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.2449173927307129, 'train/grad_norm_pre_clip': 13.125242233276367, 'timing/data': 0.15524061699397862, 'timing/model': 2.0958452589984518, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.09}) +[2026-06-23 03:39:59,932][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.4122571349143982, 'train/grad_norm_pre_clip': 60.80272674560547, 'timing/data': 0.0015550120151601732, 'timing/model': 2.1081953599932604, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.09}) +[2026-06-23 03:40:04,144][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.3087514042854309, 'train/grad_norm_pre_clip': 27.983196258544922, 'timing/data': 0.0012516309798229486, 'timing/model': 2.144438910007011, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.09}) +[2026-06-23 03:40:08,306][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.397016704082489, 'train/grad_norm_pre_clip': 67.80561828613281, 'timing/data': 0.0014723999775014818, 'timing/model': 2.1010816039924975, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.09}) +[2026-06-23 03:40:12,615][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.21183128654956818, 'train/grad_norm_pre_clip': 35.313724517822266, 'timing/data': 0.14872530399588868, 'timing/model': 2.1000286499911454, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.09}) +[2026-06-23 03:40:16,784][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.2990550398826599, 'train/grad_norm_pre_clip': 24.274171829223633, 'timing/data': 0.0014254619891289622, 'timing/model': 2.1018951080041006, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.09}) +[2026-06-23 03:40:20,956][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.3133784234523773, 'train/grad_norm_pre_clip': 41.9408073425293, 'timing/data': 0.001446481008315459, 'timing/model': 2.1082564380194526, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.09}) +[2026-06-23 03:40:25,122][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.3359528183937073, 'train/grad_norm_pre_clip': 58.729679107666016, 'timing/data': 0.0014130299969110638, 'timing/model': 2.105691041011596, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.09}) +[2026-06-23 03:40:29,442][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.2918701767921448, 'train/grad_norm_pre_clip': 53.739013671875, 'timing/data': 0.1545503529778216, 'timing/model': 2.10531637998065, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.09}) +[2026-06-23 03:40:33,622][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.2749384939670563, 'train/grad_norm_pre_clip': 25.57598114013672, 'timing/data': 0.0013387329818215221, 'timing/model': 2.1009275850083213, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.09}) +[2026-06-23 03:40:37,795][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.23665067553520203, 'train/grad_norm_pre_clip': 25.733524322509766, 'timing/data': 0.0013800689775962383, 'timing/model': 2.1142475270025898, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.09}) +[2026-06-23 03:40:42,218][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.19473542273044586, 'train/grad_norm_pre_clip': 13.470436096191406, 'timing/data': 0.0011532400094438344, 'timing/model': 2.099345964990789, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.09}) +[2026-06-23 03:40:46,513][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.2450447380542755, 'train/grad_norm_pre_clip': 43.74313735961914, 'timing/data': 0.14218368197907694, 'timing/model': 2.0968996079755016, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.09}) +[2026-06-23 03:40:50,688][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.308994859457016, 'train/grad_norm_pre_clip': 21.968387603759766, 'timing/data': 0.0011423500254750252, 'timing/model': 2.107161711988738, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.09}) +[2026-06-23 03:40:54,872][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.37721872329711914, 'train/grad_norm_pre_clip': 45.386444091796875, 'timing/data': 0.00124818499898538, 'timing/model': 2.107733639975777, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.09}) +[2026-06-23 03:40:59,055][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.20792430639266968, 'train/grad_norm_pre_clip': 23.156789779663086, 'timing/data': 0.0011819429928436875, 'timing/model': 2.1050904679868836, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.09}) +[2026-06-23 03:41:03,363][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.2188761681318283, 'train/grad_norm_pre_clip': 14.405016899108887, 'timing/data': 0.1491478109965101, 'timing/model': 2.1020824009901844, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.09}) +[2026-06-23 03:41:07,523][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.2254975587129593, 'train/grad_norm_pre_clip': 26.17409896850586, 'timing/data': 0.001129968004534021, 'timing/model': 2.0990590779983904, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.09}) +[2026-06-23 03:41:11,676][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.31946471333503723, 'train/grad_norm_pre_clip': 44.77028274536133, 'timing/data': 0.0011307990062050521, 'timing/model': 2.0969239470141474, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.1}) +[2026-06-23 03:41:16,024][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.25106576085090637, 'train/grad_norm_pre_clip': 42.49769973754883, 'timing/data': 0.0011768249969463795, 'timing/model': 2.288389761990402, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.1}) +[2026-06-23 03:41:20,328][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.3378939926624298, 'train/grad_norm_pre_clip': 26.825353622436523, 'timing/data': 0.14221725898096338, 'timing/model': 2.1026440970017575, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.1}) +[2026-06-23 03:41:24,513][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.3476904630661011, 'train/grad_norm_pre_clip': 58.04676818847656, 'timing/data': 0.0012302229879423976, 'timing/model': 2.105671066994546, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.1}) +[2026-06-23 03:41:43,430][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 250 +[2026-06-23 03:41:43,431][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.24706636369228363, 'train/grad_norm_pre_clip': 26.35892105102539, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/mean_length': 31.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/macro_mean_length': 31.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 31.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 2.0000000000000013, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0012176389864180237, 'timing/model': 2.102018723002402, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.1}) +[2026-06-23 03:41:47,566][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.3527129590511322, 'train/grad_norm_pre_clip': 34.32670593261719, 'timing/data': 0.001219482976011932, 'timing/model': 2.0853784090140834, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.1}) +[2026-06-23 03:41:51,838][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.4161660373210907, 'train/grad_norm_pre_clip': 99.61033630371094, 'timing/data': 0.14305641400278546, 'timing/model': 2.085093367990339, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.1}) +[2026-06-23 03:41:55,974][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.26599225401878357, 'train/grad_norm_pre_clip': 59.49302291870117, 'timing/data': 0.0012299030204303563, 'timing/model': 2.0905161059927195, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.1}) +[2026-06-23 03:42:00,128][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.2057529091835022, 'train/grad_norm_pre_clip': 21.522647857666016, 'timing/data': 0.0013707019970752299, 'timing/model': 2.0972211320186034, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.1}) +[2026-06-23 03:42:04,537][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.1987137794494629, 'train/grad_norm_pre_clip': 18.81916618347168, 'timing/data': 0.0015488920034840703, 'timing/model': 2.3502310379990377, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.1}) +[2026-06-23 03:42:08,797][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.2707209587097168, 'train/grad_norm_pre_clip': 25.628459930419922, 'timing/data': 0.13315243797842413, 'timing/model': 2.0835740659967996, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.1}) +[2026-06-23 03:42:12,966][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.35860422253608704, 'train/grad_norm_pre_clip': 33.10816955566406, 'timing/data': 0.00118530000327155, 'timing/model': 2.107304484001361, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.1}) +[2026-06-23 03:42:17,134][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.3004228174686432, 'train/grad_norm_pre_clip': 44.48270797729492, 'timing/data': 0.0015236760082188994, 'timing/model': 2.1026994099956937, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.1}) +[2026-06-23 03:42:21,283][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.2852192521095276, 'train/grad_norm_pre_clip': 24.91064453125, 'timing/data': 0.0014166180044412613, 'timing/model': 2.0966706080071162, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.1}) +[2026-06-23 03:42:25,566][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.28130608797073364, 'train/grad_norm_pre_clip': 43.23019790649414, 'timing/data': 0.1416356809786521, 'timing/model': 2.0876764440035913, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.1}) +[2026-06-23 03:42:29,709][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.22816376388072968, 'train/grad_norm_pre_clip': 43.38177490234375, 'timing/data': 0.0012548190134111792, 'timing/model': 2.09178180800518, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.1}) +[2026-06-23 03:42:33,852][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.2824894189834595, 'train/grad_norm_pre_clip': 37.137550354003906, 'timing/data': 0.0011262719926889986, 'timing/model': 2.0920731340011116, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.1}) +[2026-06-23 03:42:38,119][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.20837293565273285, 'train/grad_norm_pre_clip': 19.38404655456543, 'timing/data': 0.0013333240058273077, 'timing/model': 2.2012723420048133, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.1}) +[2026-06-23 03:42:42,408][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.2443011999130249, 'train/grad_norm_pre_clip': 17.447242736816406, 'timing/data': 0.11330096999881789, 'timing/model': 2.1090142219909467, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.1}) +[2026-06-23 03:42:46,568][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.18237905204296112, 'train/grad_norm_pre_clip': 12.697317123413086, 'timing/data': 0.0014063890266697854, 'timing/model': 2.1008961990009993, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.1}) +[2026-06-23 03:42:50,720][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.2374773472547531, 'train/grad_norm_pre_clip': 20.05859375, 'timing/data': 0.0014558009861502796, 'timing/model': 2.0963902809889987, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.1}) +[2026-06-23 03:42:54,892][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.1568809598684311, 'train/grad_norm_pre_clip': 11.48725414276123, 'timing/data': 0.0014087329909671098, 'timing/model': 2.115301282989094, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.1}) +[2026-06-23 03:42:59,189][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.22663435339927673, 'train/grad_norm_pre_clip': 55.0015983581543, 'timing/data': 0.14743389800423756, 'timing/model': 2.0949204500066116, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.1}) +[2026-06-23 03:43:03,343][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.2900189161300659, 'train/grad_norm_pre_clip': 60.956268310546875, 'timing/data': 0.0017051010217983276, 'timing/model': 2.099254667002242, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.1}) +[2026-06-23 03:43:07,512][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.4038279056549072, 'train/grad_norm_pre_clip': 70.7764892578125, 'timing/data': 0.0014860549999866635, 'timing/model': 2.108271265984513, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.1}) +[2026-06-23 03:43:11,678][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.29322946071624756, 'train/grad_norm_pre_clip': 26.39930534362793, 'timing/data': 0.001497335993917659, 'timing/model': 2.1088430100062396, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.1}) +[2026-06-23 03:43:16,197][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.4402887225151062, 'train/grad_norm_pre_clip': 68.80534362792969, 'timing/data': 0.14634634400135837, 'timing/model': 2.1000439089839347, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.11}) +[2026-06-23 03:43:20,368][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.3911926746368408, 'train/grad_norm_pre_clip': 85.62267303466797, 'timing/data': 0.0012395099911373109, 'timing/model': 2.111277454998344, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.11}) +[2026-06-23 03:43:24,524][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.3413640558719635, 'train/grad_norm_pre_clip': 34.19609832763672, 'timing/data': 0.0013201490219216794, 'timing/model': 2.100156083004549, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.11}) +[2026-06-23 03:43:28,687][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.36808788776397705, 'train/grad_norm_pre_clip': 36.27083969116211, 'timing/data': 0.001433519006241113, 'timing/model': 2.103321850998327, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.11}) +[2026-06-23 03:43:32,996][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.3475677967071533, 'train/grad_norm_pre_clip': 84.81912994384766, 'timing/data': 0.15191974901244976, 'timing/model': 2.096445183007745, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.11}) +[2026-06-23 03:43:37,156][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.2889622449874878, 'train/grad_norm_pre_clip': 21.48936653137207, 'timing/data': 0.0012685239780694246, 'timing/model': 2.096457976003876, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.11}) +[2026-06-23 03:43:41,305][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.1927066296339035, 'train/grad_norm_pre_clip': 23.617385864257812, 'timing/data': 0.0011586910113692284, 'timing/model': 2.094498744001612, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.11}) +[2026-06-23 03:43:45,489][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.2740759551525116, 'train/grad_norm_pre_clip': 20.31744956970215, 'timing/data': 0.0016709170013200492, 'timing/model': 2.113695824984461, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.11}) +[2026-06-23 03:43:49,952][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.28581517934799194, 'train/grad_norm_pre_clip': 29.297544479370117, 'timing/data': 0.13644226899486966, 'timing/model': 2.238919612020254, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.11}) +[2026-06-23 03:43:54,128][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.39784300327301025, 'train/grad_norm_pre_clip': 86.267822265625, 'timing/data': 0.0012303130060900003, 'timing/model': 2.112625684007071, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.11}) +[2026-06-23 03:43:58,289][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.43694183230400085, 'train/grad_norm_pre_clip': 76.04076385498047, 'timing/data': 0.0013571569870691746, 'timing/model': 2.1016877929796465, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.11}) +[2026-06-23 03:44:02,468][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.3391943573951721, 'train/grad_norm_pre_clip': 23.052383422851562, 'timing/data': 0.0012784530117642134, 'timing/model': 2.1059620199957862, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.11}) +[2026-06-23 03:44:06,763][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.25675567984580994, 'train/grad_norm_pre_clip': 32.534393310546875, 'timing/data': 0.1385019310109783, 'timing/model': 2.096803917986108, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.11}) +[2026-06-23 03:44:10,946][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.28439879417419434, 'train/grad_norm_pre_clip': 35.49199295043945, 'timing/data': 0.0013597119832411408, 'timing/model': 2.110840522014769, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.11}) +[2026-06-23 03:44:15,124][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.22675010561943054, 'train/grad_norm_pre_clip': 27.22488021850586, 'timing/data': 0.0012167680251877755, 'timing/model': 2.120063853013562, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.11}) +[2026-06-23 03:44:19,306][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.2268252819776535, 'train/grad_norm_pre_clip': 26.295469284057617, 'timing/data': 0.0013712439977098256, 'timing/model': 2.114140638994286, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.11}) +[2026-06-23 03:44:23,864][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.3673943281173706, 'train/grad_norm_pre_clip': 61.832183837890625, 'timing/data': 0.13855707997572608, 'timing/model': 2.3610356020217296, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.11}) +[2026-06-23 03:44:28,084][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.317181259393692, 'train/grad_norm_pre_clip': 14.354808807373047, 'timing/data': 0.001240721991052851, 'timing/model': 2.1549447199795395, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.11}) +[2026-06-23 03:44:32,275][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.25326648354530334, 'train/grad_norm_pre_clip': 22.495004653930664, 'timing/data': 0.0013559260114561766, 'timing/model': 2.1031144440057687, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.11}) +[2026-06-23 03:44:36,456][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.23187491297721863, 'train/grad_norm_pre_clip': 19.894893646240234, 'timing/data': 0.001249077991815284, 'timing/model': 2.110869081981946, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.11}) +[2026-06-23 03:44:40,780][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.38483497500419617, 'train/grad_norm_pre_clip': 43.66135025024414, 'timing/data': 0.14253625899436884, 'timing/model': 2.1085086109815165, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.11}) +[2026-06-23 03:44:44,945][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.2630419135093689, 'train/grad_norm_pre_clip': 30.593229293823242, 'timing/data': 0.0012495179835241288, 'timing/model': 2.103406673995778, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.11}) +[2026-06-23 03:44:49,114][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.3482593297958374, 'train/grad_norm_pre_clip': 21.137285232543945, 'timing/data': 0.001246172992978245, 'timing/model': 2.1069073419785127, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.11}) +[2026-06-23 03:44:53,285][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.3307218551635742, 'train/grad_norm_pre_clip': 55.86977005004883, 'timing/data': 0.001258396019693464, 'timing/model': 2.1052576849760953, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.11}) +[2026-06-23 03:44:57,583][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.2595227062702179, 'train/grad_norm_pre_clip': 29.615270614624023, 'timing/data': 0.13753118898603134, 'timing/model': 2.099969353992492, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.11}) +[2026-06-23 03:45:01,789][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.353575736284256, 'train/grad_norm_pre_clip': 57.249290466308594, 'timing/data': 0.0011257599981036037, 'timing/model': 2.1004479289986193, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.11}) +[2026-06-23 03:45:05,944][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.27614450454711914, 'train/grad_norm_pre_clip': 32.16024398803711, 'timing/data': 0.0011397459893487394, 'timing/model': 2.097071431024233, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.12}) +[2026-06-23 03:45:10,100][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.2418181151151657, 'train/grad_norm_pre_clip': 28.916934967041016, 'timing/data': 0.001130409014876932, 'timing/model': 2.100761258014245, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.12}) +[2026-06-23 03:45:14,400][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 300 +[2026-06-23 03:47:07,543][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 300 +[2026-06-23 03:47:07,544][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 300 +[2026-06-23 03:49:25,073][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 300 +[2026-06-23 03:49:25,075][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.39492547512054443, 'train/grad_norm_pre_clip': 54.996463775634766, 'eval/action_loss/seconds': 113.14026153701707, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.15760403871536255, 'eval/latency_2/loss': 0.15760403871536255, 'eval/flappy/loss': 0.15760403871536255, 'eval/flappy/latency_2/loss': 0.15760403871536255, 'eval/flap_precision': 0.7387983706720977, 'eval/flap_recall': 0.9790823211875843, 'eval/flap_f1': 0.8421358096343587, 'eval/flap_rate_model': 0.2727777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8421358096343587, 'eval/latency_2/flap_precision': 0.7387983706720977, 'eval/latency_2/flap_recall': 0.9790823211875843, 'eval/latency_2/flap_f1': 0.8421358096343587, 'eval/latency_2/flap_rate_model': 0.2727777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8421358096343587, 'eval/action_classification/data_wait_seconds': 50.21460723876953, 'eval/action_classification/predict_seconds': 85.45789337158203, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.14591304399073124, 'timing/model': 2.099322691006819, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.12}) +[2026-06-23 03:49:29,219][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.27770522236824036, 'train/grad_norm_pre_clip': 34.705020904541016, 'timing/data': 0.0013108020066283643, 'timing/model': 2.0955513280059677, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.12}) +[2026-06-23 03:49:33,490][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.16657134890556335, 'train/grad_norm_pre_clip': 24.215591430664062, 'timing/data': 0.0012873290106654167, 'timing/model': 2.2374100539891515, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.12}) +[2026-06-23 03:49:37,610][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.13233788311481476, 'train/grad_norm_pre_clip': 24.5881404876709, 'timing/data': 0.0011729569814633578, 'timing/model': 2.085723180003697, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.12}) +[2026-06-23 03:49:41,864][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.336222380399704, 'train/grad_norm_pre_clip': 46.40286636352539, 'timing/data': 0.1473161930043716, 'timing/model': 2.0742073500005063, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.12}) +[2026-06-23 03:49:45,989][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.2139941304922104, 'train/grad_norm_pre_clip': 20.944705963134766, 'timing/data': 0.0012804860016331077, 'timing/model': 2.090343385003507, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.12}) +[2026-06-23 03:49:50,120][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.30439379811286926, 'train/grad_norm_pre_clip': 43.66838836669922, 'timing/data': 0.0011475000064820051, 'timing/model': 2.086468650988536, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.12}) +[2026-06-23 03:49:54,267][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.17846384644508362, 'train/grad_norm_pre_clip': 20.092988967895508, 'timing/data': 0.0014768879918847233, 'timing/model': 2.098929006024264, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.12}) +[2026-06-23 03:49:58,549][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.30527549982070923, 'train/grad_norm_pre_clip': 50.77703857421875, 'timing/data': 0.1446226359985303, 'timing/model': 2.0834762529993895, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.12}) +[2026-06-23 03:50:02,669][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.26734301447868347, 'train/grad_norm_pre_clip': 87.2909927368164, 'timing/data': 0.0013115729962009937, 'timing/model': 2.0816283939930145, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.12}) +[2026-06-23 03:50:06,793][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.23017394542694092, 'train/grad_norm_pre_clip': 76.24623107910156, 'timing/data': 0.0011652529938146472, 'timing/model': 2.082087585993577, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.12}) +[2026-06-23 03:50:11,175][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.3237288296222687, 'train/grad_norm_pre_clip': 73.2035903930664, 'timing/data': 0.0011333939910400659, 'timing/model': 2.0795419579953887, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.12}) +[2026-06-23 03:50:15,461][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.44243428111076355, 'train/grad_norm_pre_clip': 78.89091491699219, 'timing/data': 0.1546080719854217, 'timing/model': 2.089430089981761, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.12}) +[2026-06-23 03:50:19,589][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.3958563506603241, 'train/grad_norm_pre_clip': 47.29584884643555, 'timing/data': 0.0014592960069421679, 'timing/model': 2.0810058529896196, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.12}) +[2026-06-23 03:50:23,736][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.16899390518665314, 'train/grad_norm_pre_clip': 34.220123291015625, 'timing/data': 0.0012153850111644715, 'timing/model': 2.0882883709855378, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.12}) +[2026-06-23 03:50:27,910][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.35738319158554077, 'train/grad_norm_pre_clip': 75.56159973144531, 'timing/data': 0.0012539960152935237, 'timing/model': 2.1127847600146197, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.12}) +[2026-06-23 03:50:32,193][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.5142020583152771, 'train/grad_norm_pre_clip': 139.87295532226562, 'timing/data': 0.14521670000976883, 'timing/model': 2.090269521024311, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.12}) +[2026-06-23 03:50:36,335][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.5320470333099365, 'train/grad_norm_pre_clip': 145.71051025390625, 'timing/data': 0.0012054670078214258, 'timing/model': 2.091659333993448, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.12}) +[2026-06-23 03:50:40,491][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.19590003788471222, 'train/grad_norm_pre_clip': 50.57525634765625, 'timing/data': 0.0012255639885552227, 'timing/model': 2.1022136360115837, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.12}) +[2026-06-23 03:50:44,737][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.29969707131385803, 'train/grad_norm_pre_clip': 70.64059448242188, 'timing/data': 0.001191501010907814, 'timing/model': 2.1150190379994456, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.12}) +[2026-06-23 03:50:49,050][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.3579755127429962, 'train/grad_norm_pre_clip': 92.34070587158203, 'timing/data': 0.1472220239811577, 'timing/model': 2.1149460980086587, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.12}) +[2026-06-23 03:50:53,210][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.528963565826416, 'train/grad_norm_pre_clip': 100.1475601196289, 'timing/data': 0.0011982840078417212, 'timing/model': 2.0954792510019615, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.12}) +[2026-06-23 03:50:57,350][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.34698542952537537, 'train/grad_norm_pre_clip': 61.02589797973633, 'timing/data': 0.0011505559959914535, 'timing/model': 2.090914653002983, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.12}) +[2026-06-23 03:51:01,498][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.37041592597961426, 'train/grad_norm_pre_clip': 29.201007843017578, 'timing/data': 0.0011292860144749284, 'timing/model': 2.0930808109987993, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.13}) +[2026-06-23 03:51:05,791][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.359504759311676, 'train/grad_norm_pre_clip': 73.77932739257812, 'timing/data': 0.14258353298646398, 'timing/model': 2.100503418012522, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.13}) +[2026-06-23 03:51:09,936][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.22484874725341797, 'train/grad_norm_pre_clip': 59.115840911865234, 'timing/data': 0.0012209349952172488, 'timing/model': 2.094278412987478, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.13}) +[2026-06-23 03:51:14,094][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.36556708812713623, 'train/grad_norm_pre_clip': 48.58946990966797, 'timing/data': 0.0012680530198849738, 'timing/model': 2.103797513002064, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.13}) +[2026-06-23 03:51:18,262][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.15549610555171967, 'train/grad_norm_pre_clip': 18.946895599365234, 'timing/data': 0.0012525439960882068, 'timing/model': 2.108846275979886, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.13}) +[2026-06-23 03:51:22,841][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.17691318690776825, 'train/grad_norm_pre_clip': 35.93019104003906, 'timing/data': 0.15943489299388602, 'timing/model': 2.101809943007538, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.13}) +[2026-06-23 03:51:27,003][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.20588164031505585, 'train/grad_norm_pre_clip': 37.945594787597656, 'timing/data': 0.0011582199949771166, 'timing/model': 2.102098159986781, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.13}) +[2026-06-23 03:51:31,160][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.37693360447883606, 'train/grad_norm_pre_clip': 49.32057571411133, 'timing/data': 0.0012288799916859716, 'timing/model': 2.099830696999561, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.13}) +[2026-06-23 03:51:35,333][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.21406814455986023, 'train/grad_norm_pre_clip': 14.943157196044922, 'timing/data': 0.0012513709953054786, 'timing/model': 2.1040862450026907, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.13}) +[2026-06-23 03:51:39,658][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.26677966117858887, 'train/grad_norm_pre_clip': 24.959266662597656, 'timing/data': 0.16405279200989753, 'timing/model': 2.10307518200716, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.13}) +[2026-06-23 03:51:43,822][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.242560014128685, 'train/grad_norm_pre_clip': 29.83144760131836, 'timing/data': 0.0012074909755028784, 'timing/model': 2.106166228011716, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.13}) +[2026-06-23 03:51:47,993][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.12907999753952026, 'train/grad_norm_pre_clip': 19.35154151916504, 'timing/data': 0.001293770008487627, 'timing/model': 2.1124354650091846, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.13}) +[2026-06-23 03:51:52,167][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.33940425515174866, 'train/grad_norm_pre_clip': 56.78635787963867, 'timing/data': 0.0013619959936477244, 'timing/model': 2.1151511679927353, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.13}) +[2026-06-23 03:51:56,538][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.3311193883419037, 'train/grad_norm_pre_clip': 45.931175231933594, 'timing/data': 0.14699153299443424, 'timing/model': 2.102235644008033, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.13}) +[2026-06-23 03:52:00,696][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.3413843810558319, 'train/grad_norm_pre_clip': 44.568023681640625, 'timing/data': 0.0013904489751439542, 'timing/model': 2.1006658200058155, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.13}) +[2026-06-23 03:52:04,853][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.23720581829547882, 'train/grad_norm_pre_clip': 25.361408233642578, 'timing/data': 0.0012128200032748282, 'timing/model': 2.0973944530123845, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.13}) +[2026-06-23 03:52:09,011][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.5613356828689575, 'train/grad_norm_pre_clip': 88.40182495117188, 'timing/data': 0.0013969609863124788, 'timing/model': 2.099737868004013, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.13}) +[2026-06-23 03:52:13,321][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.4893702268600464, 'train/grad_norm_pre_clip': 83.07958221435547, 'timing/data': 0.14953488000901416, 'timing/model': 2.1017172790016048, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.13}) +[2026-06-23 03:52:17,480][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.7064704298973083, 'train/grad_norm_pre_clip': 137.31849670410156, 'timing/data': 0.0012032630038447678, 'timing/model': 2.102075727976626, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.13}) +[2026-06-23 03:52:21,647][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.29900556802749634, 'train/grad_norm_pre_clip': 67.4117660522461, 'timing/data': 0.001208502013469115, 'timing/model': 2.104834388999734, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.13}) +[2026-06-23 03:52:25,810][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.18520471453666687, 'train/grad_norm_pre_clip': 18.48381996154785, 'timing/data': 0.001418870990164578, 'timing/model': 2.102171806996921, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.13}) +[2026-06-23 03:52:30,401][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.280762255191803, 'train/grad_norm_pre_clip': 36.40577697753906, 'timing/data': 0.16023975401185453, 'timing/model': 2.3687504779954907, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.13}) +[2026-06-23 03:52:34,560][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.27166223526000977, 'train/grad_norm_pre_clip': 52.90169143676758, 'timing/data': 0.0012295710039325058, 'timing/model': 2.1010976339748595, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.13}) +[2026-06-23 03:52:38,721][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.2959689497947693, 'train/grad_norm_pre_clip': 54.7066650390625, 'timing/data': 0.0012203849910292774, 'timing/model': 2.103122221014928, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.13}) +[2026-06-23 03:52:42,881][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.17027468979358673, 'train/grad_norm_pre_clip': 37.383209228515625, 'timing/data': 0.0013369399821385741, 'timing/model': 2.102073038025992, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.13}) +[2026-06-23 03:52:47,194][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.36329326033592224, 'train/grad_norm_pre_clip': 16.854639053344727, 'timing/data': 0.15241348498966545, 'timing/model': 2.1000264240137767, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.13}) +[2026-06-23 03:52:51,356][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.3260349929332733, 'train/grad_norm_pre_clip': 55.67972946166992, 'timing/data': 0.0012928280048072338, 'timing/model': 2.1037145029986277, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.14}) +[2026-06-23 03:52:55,521][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.48171183466911316, 'train/grad_norm_pre_clip': 66.93905639648438, 'timing/data': 0.001248115993803367, 'timing/model': 2.105346244003158, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.14}) +[2026-06-23 03:52:59,699][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.5177386999130249, 'train/grad_norm_pre_clip': 87.8480224609375, 'timing/data': 0.0012514519912656397, 'timing/model': 2.1168273389921524, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.14}) +[2026-06-23 03:53:04,063][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.41876500844955444, 'train/grad_norm_pre_clip': 66.82398986816406, 'timing/data': 0.14714393898611888, 'timing/model': 2.1400297270156443, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.14}) +[2026-06-23 03:53:08,244][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.3392176032066345, 'train/grad_norm_pre_clip': 39.07529067993164, 'timing/data': 0.0011759419867303222, 'timing/model': 2.1026676839974243, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.14}) +[2026-06-23 03:53:12,410][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.2728070020675659, 'train/grad_norm_pre_clip': 46.676090240478516, 'timing/data': 0.001225665007950738, 'timing/model': 2.103938278014539, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.14}) +[2026-06-23 03:53:16,574][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.4006428122520447, 'train/grad_norm_pre_clip': 58.74363708496094, 'timing/data': 0.0011854499753098935, 'timing/model': 2.101205074985046, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.14}) +[2026-06-23 03:53:20,904][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.6233628392219543, 'train/grad_norm_pre_clip': 65.76506805419922, 'timing/data': 0.1626487809990067, 'timing/model': 2.101715423981659, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.14}) +[2026-06-23 03:53:25,079][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.2523755431175232, 'train/grad_norm_pre_clip': 39.106746673583984, 'timing/data': 0.0013994849869050086, 'timing/model': 2.10719158698339, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.14}) +[2026-06-23 03:53:29,270][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.23792590200901031, 'train/grad_norm_pre_clip': 34.02680969238281, 'timing/data': 0.0017857089987955987, 'timing/model': 2.1168446250085253, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.14}) +[2026-06-23 03:53:33,439][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.31108710169792175, 'train/grad_norm_pre_clip': 18.125160217285156, 'timing/data': 0.0011923619895242155, 'timing/model': 2.105739958002232, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.14}) +[2026-06-23 03:53:37,770][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.33877068758010864, 'train/grad_norm_pre_clip': 36.84098815917969, 'timing/data': 0.1580532609950751, 'timing/model': 2.1056474529905245, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.14}) +[2026-06-23 03:53:42,115][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.3192615509033203, 'train/grad_norm_pre_clip': 34.545875549316406, 'timing/data': 0.001629470003535971, 'timing/model': 2.1285012970038224, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.14}) +[2026-06-23 03:53:46,296][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.1504869908094406, 'train/grad_norm_pre_clip': 8.486937522888184, 'timing/data': 0.001355173997581005, 'timing/model': 2.118084030022146, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.14}) +[2026-06-23 03:53:50,473][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.26234689354896545, 'train/grad_norm_pre_clip': 19.533727645874023, 'timing/data': 0.0011427610006649047, 'timing/model': 2.1008695370110217, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.14}) +[2026-06-23 03:53:54,803][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.18162883818149567, 'train/grad_norm_pre_clip': 19.675931930541992, 'timing/data': 0.15096036702743731, 'timing/model': 2.1169833299936727, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.14}) +[2026-06-23 03:53:58,978][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.2154979705810547, 'train/grad_norm_pre_clip': 20.99677848815918, 'timing/data': 0.0012969060044270009, 'timing/model': 2.113108279008884, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.14}) +[2026-06-23 03:54:03,149][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.1773536503314972, 'train/grad_norm_pre_clip': 18.8335018157959, 'timing/data': 0.001112665981054306, 'timing/model': 2.104030877002515, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.14}) +[2026-06-23 03:54:07,324][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.2973560392856598, 'train/grad_norm_pre_clip': 35.275264739990234, 'timing/data': 0.0014573020162060857, 'timing/model': 2.113077094021719, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.14}) +[2026-06-23 03:54:11,641][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.22521661221981049, 'train/grad_norm_pre_clip': 32.04900360107422, 'timing/data': 0.14505334399291314, 'timing/model': 2.104247879993636, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.14}) +[2026-06-23 03:54:16,071][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.1887180358171463, 'train/grad_norm_pre_clip': 37.526649475097656, 'timing/data': 0.0013643399870488793, 'timing/model': 2.366991546994541, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.14}) +[2026-06-23 03:54:20,247][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.29418957233428955, 'train/grad_norm_pre_clip': 61.36444854736328, 'timing/data': 0.0013381720054894686, 'timing/model': 2.1161465749901254, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.14}) +[2026-06-23 03:54:24,448][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.20429591834545135, 'train/grad_norm_pre_clip': 27.412330627441406, 'timing/data': 0.0014441079983953387, 'timing/model': 2.110019662999548, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.14}) +[2026-06-23 03:54:28,768][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.14502905309200287, 'train/grad_norm_pre_clip': 21.162729263305664, 'timing/data': 0.14995856501627713, 'timing/model': 2.103334140003426, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.14}) +[2026-06-23 03:54:32,954][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.3534598648548126, 'train/grad_norm_pre_clip': 76.8938217163086, 'timing/data': 0.0011796889884863049, 'timing/model': 2.1195231419987977, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.14}) +[2026-06-23 03:54:37,122][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.3749195337295532, 'train/grad_norm_pre_clip': 76.60395812988281, 'timing/data': 0.0013163609837647527, 'timing/model': 2.106795166997472, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.14}) +[2026-06-23 03:54:41,300][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.24557611346244812, 'train/grad_norm_pre_clip': 40.50841522216797, 'timing/data': 0.0012302229879423976, 'timing/model': 2.1096348289866, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.15}) +[2026-06-23 03:54:45,616][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.333501935005188, 'train/grad_norm_pre_clip': 29.762310028076172, 'timing/data': 0.14793708798242733, 'timing/model': 2.1037850299908314, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.15}) +[2026-06-23 03:54:49,859][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.4933747351169586, 'train/grad_norm_pre_clip': 91.3364028930664, 'timing/data': 0.0013750500220339745, 'timing/model': 2.127394089999143, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.15}) +[2026-06-23 03:54:54,024][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.37074756622314453, 'train/grad_norm_pre_clip': 51.66512680053711, 'timing/data': 0.0012120689789298922, 'timing/model': 2.1020794629876036, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.15}) +[2026-06-23 03:54:58,212][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.4530399441719055, 'train/grad_norm_pre_clip': 79.5718765258789, 'timing/data': 0.0013986339909024537, 'timing/model': 2.105188233021181, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.15}) +[2026-06-23 03:55:02,528][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.20998165011405945, 'train/grad_norm_pre_clip': 40.03194046020508, 'timing/data': 0.15152533698710613, 'timing/model': 2.101102421991527, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.15}) +[2026-06-23 03:55:06,703][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.32083529233932495, 'train/grad_norm_pre_clip': 48.796871185302734, 'timing/data': 0.0014650759985670447, 'timing/model': 2.1126112499914598, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.15}) +[2026-06-23 03:55:10,885][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.194553941488266, 'train/grad_norm_pre_clip': 31.258304595947266, 'timing/data': 0.0014761770144104958, 'timing/model': 2.1115207220136654, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.15}) +[2026-06-23 03:55:15,063][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.25153687596321106, 'train/grad_norm_pre_clip': 60.97816848754883, 'timing/data': 0.0012747650034725666, 'timing/model': 2.114912133984035, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.15}) +[2026-06-23 03:55:19,377][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.28817853331565857, 'train/grad_norm_pre_clip': 37.82240676879883, 'timing/data': 0.1464123610057868, 'timing/model': 2.1036985669925343, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.15}) +[2026-06-23 03:55:23,601][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.23819392919540405, 'train/grad_norm_pre_clip': 39.578067779541016, 'timing/data': 0.0012212459987495095, 'timing/model': 2.157532106997678, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.15}) +[2026-06-23 03:55:27,790][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.24486491084098816, 'train/grad_norm_pre_clip': 40.32735824584961, 'timing/data': 0.001417278021108359, 'timing/model': 2.123677556985058, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.15}) +[2026-06-23 03:55:31,963][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.2906058430671692, 'train/grad_norm_pre_clip': 61.75617218017578, 'timing/data': 0.0012196230236440897, 'timing/model': 2.108728199003963, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.15}) +[2026-06-23 03:55:36,287][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.19828541576862335, 'train/grad_norm_pre_clip': 13.60985279083252, 'timing/data': 0.14905650398577563, 'timing/model': 2.1065320890047587, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.15}) +[2026-06-23 03:55:40,484][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.17738376557826996, 'train/grad_norm_pre_clip': 21.400169372558594, 'timing/data': 0.00120484599028714, 'timing/model': 2.1085927239910234, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.15}) +[2026-06-23 03:55:44,647][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.3043254017829895, 'train/grad_norm_pre_clip': 50.58858108520508, 'timing/data': 0.001297887007240206, 'timing/model': 2.10197786099161, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.15}) +[2026-06-23 03:55:48,810][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.15101908147335052, 'train/grad_norm_pre_clip': 27.361740112304688, 'timing/data': 0.001245802006451413, 'timing/model': 2.101945378992241, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.15}) +[2026-06-23 03:55:53,122][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.19890940189361572, 'train/grad_norm_pre_clip': 19.80866241455078, 'timing/data': 0.14449451700784266, 'timing/model': 2.1035213170107454, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.15}) +[2026-06-23 03:55:57,308][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.20651516318321228, 'train/grad_norm_pre_clip': 27.275211334228516, 'timing/data': 0.0014728709938935935, 'timing/model': 2.113917827984551, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.15}) +[2026-06-23 03:56:01,670][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.2705991268157959, 'train/grad_norm_pre_clip': 45.585575103759766, 'timing/data': 0.0012679620122071356, 'timing/model': 2.108544129994698, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.15}) +[2026-06-23 03:56:05,829][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.26483243703842163, 'train/grad_norm_pre_clip': 25.268451690673828, 'timing/data': 0.0011176450061611831, 'timing/model': 2.1000579109822866, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.15}) +[2026-06-23 03:56:10,128][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.30134424567222595, 'train/grad_norm_pre_clip': 35.49616241455078, 'timing/data': 0.14530975199886598, 'timing/model': 2.0983026959875133, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.15}) +[2026-06-23 03:56:14,322][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.2618660032749176, 'train/grad_norm_pre_clip': 17.995296478271484, 'timing/data': 0.0012710380251519382, 'timing/model': 2.1243608590157237, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.15}) +[2026-06-23 03:56:18,494][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.150369331240654, 'train/grad_norm_pre_clip': 14.520588874816895, 'timing/data': 0.0013396739959716797, 'timing/model': 2.103193063987419, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.15}) +[2026-06-23 03:56:22,660][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.18859344720840454, 'train/grad_norm_pre_clip': 28.18293571472168, 'timing/data': 0.0012575029977597296, 'timing/model': 2.1026464820024557, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.15}) +[2026-06-23 03:56:26,958][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 400 +[2026-06-23 03:58:16,350][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 400 +[2026-06-23 03:58:16,350][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 400 +[2026-06-23 04:00:33,822][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 400 +[2026-06-23 04:00:33,823][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.2592415511608124, 'train/grad_norm_pre_clip': 16.8085880279541, 'eval/action_loss/seconds': 109.38810474800994, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.1115216463804245, 'eval/latency_2/loss': 0.1115216463804245, 'eval/flappy/loss': 0.1115216463804245, 'eval/flappy/latency_2/loss': 0.1115216463804245, 'eval/flap_precision': 0.8830995323981295, 'eval/flap_recall': 0.8920377867746289, 'eval/flap_f1': 0.8875461564283316, 'eval/flap_rate_model': 0.20791666666666667, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8875461564283316, 'eval/latency_2/flap_precision': 0.8830995323981295, 'eval/latency_2/flap_recall': 0.8920377867746289, 'eval/latency_2/flap_f1': 0.8875461564283316, 'eval/latency_2/flap_rate_model': 0.20791666666666667, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8875461564283316, 'eval/action_classification/data_wait_seconds': 50.196712493896484, 'eval/action_classification/predict_seconds': 85.70860290527344, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.1463369780103676, 'timing/model': 2.0956950889958534, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.15}) +[2026-06-23 04:00:37,945][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.3193427324295044, 'train/grad_norm_pre_clip': 31.388587951660156, 'timing/data': 0.0014953619975131005, 'timing/model': 2.077304060017923, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.16}) +[2026-06-23 04:00:42,052][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.15510599315166473, 'train/grad_norm_pre_clip': 31.93114471435547, 'timing/data': 0.001122713991208002, 'timing/model': 2.073903507000068, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.16}) +[2026-06-23 04:00:46,168][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.17106416821479797, 'train/grad_norm_pre_clip': 39.065208435058594, 'timing/data': 0.0013179240049794316, 'timing/model': 2.075838850985747, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.16}) +[2026-06-23 04:00:50,430][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.21262134611606598, 'train/grad_norm_pre_clip': 19.884267807006836, 'timing/data': 0.14837575601995923, 'timing/model': 2.0775838010013103, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.16}) +[2026-06-23 04:00:54,549][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.14173999428749084, 'train/grad_norm_pre_clip': 27.435758590698242, 'timing/data': 0.0011049809982068837, 'timing/model': 2.0793201469932683, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.16}) +[2026-06-23 04:00:58,674][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.32888174057006836, 'train/grad_norm_pre_clip': 37.84393310546875, 'timing/data': 0.0013174139894545078, 'timing/model': 2.0841683620237745, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.16}) +[2026-06-23 04:01:02,798][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.25279417634010315, 'train/grad_norm_pre_clip': 26.375831604003906, 'timing/data': 0.001140695996582508, 'timing/model': 2.0820843709807377, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.16}) +[2026-06-23 04:01:07,172][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.17417114973068237, 'train/grad_norm_pre_clip': 28.96921730041504, 'timing/data': 0.14271569999982603, 'timing/model': 2.0825109269935638, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.16}) +[2026-06-23 04:01:11,305][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.2665143609046936, 'train/grad_norm_pre_clip': 42.729408264160156, 'timing/data': 0.0011453459737822413, 'timing/model': 2.0903190489916597, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.16}) +[2026-06-23 04:01:15,433][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.2736845314502716, 'train/grad_norm_pre_clip': 41.98808670043945, 'timing/data': 0.0011289450048934668, 'timing/model': 2.083310301997699, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.16}) +[2026-06-23 04:01:19,563][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.14180614054203033, 'train/grad_norm_pre_clip': 25.84051513671875, 'timing/data': 0.0012707679998129606, 'timing/model': 2.0852529980184045, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.16}) +[2026-06-23 04:01:23,836][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.21487775444984436, 'train/grad_norm_pre_clip': 14.393939971923828, 'timing/data': 0.14304796399665065, 'timing/model': 2.085608213004889, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.16}) +[2026-06-23 04:01:27,968][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.449757844209671, 'train/grad_norm_pre_clip': 67.9190444946289, 'timing/data': 0.0015738370129838586, 'timing/model': 2.0858539849868976, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.16}) +[2026-06-23 04:01:32,098][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.5012720823287964, 'train/grad_norm_pre_clip': 82.55580139160156, 'timing/data': 0.0011213719844818115, 'timing/model': 2.085272801021347, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.16}) +[2026-06-23 04:01:36,229][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.21341797709465027, 'train/grad_norm_pre_clip': 44.49995803833008, 'timing/data': 0.0012574720021802932, 'timing/model': 2.088107926014345, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.16}) +[2026-06-23 04:01:40,755][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.35463687777519226, 'train/grad_norm_pre_clip': 25.563175201416016, 'timing/data': 0.139422687003389, 'timing/model': 2.3345213510037865, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.16}) +[2026-06-23 04:01:44,898][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.33278846740722656, 'train/grad_norm_pre_clip': 51.232330322265625, 'timing/data': 0.0013643900165334344, 'timing/model': 2.0919067799986806, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.16}) +[2026-06-23 04:01:49,043][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.4416363537311554, 'train/grad_norm_pre_clip': 59.46248245239258, 'timing/data': 0.0013180749956518412, 'timing/model': 2.0910872159875, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.16}) +[2026-06-23 04:01:53,189][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.3427586853504181, 'train/grad_norm_pre_clip': 34.96331024169922, 'timing/data': 0.0011560650018509477, 'timing/model': 2.09716772599495, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.16}) +[2026-06-23 04:01:57,478][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.2651714086532593, 'train/grad_norm_pre_clip': 13.809247016906738, 'timing/data': 0.1428766270109918, 'timing/model': 2.0915179179864936, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.16}) +[2026-06-23 04:02:01,643][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.21793368458747864, 'train/grad_norm_pre_clip': 24.65850257873535, 'timing/data': 0.0013807810028083622, 'timing/model': 2.114559088018723, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.16}) +[2026-06-23 04:02:05,800][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.17996177077293396, 'train/grad_norm_pre_clip': 12.058815002441406, 'timing/data': 0.0012450999929569662, 'timing/model': 2.102982271986548, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.16}) +[2026-06-23 04:02:09,951][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.2683930993080139, 'train/grad_norm_pre_clip': 25.06516456604004, 'timing/data': 0.0012783019919879735, 'timing/model': 2.096819337020861, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.16}) +[2026-06-23 04:02:14,330][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.18767397105693817, 'train/grad_norm_pre_clip': 13.186849594116211, 'timing/data': 0.1562133519910276, 'timing/model': 2.1717792190029286, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.16}) +[2026-06-23 04:02:18,496][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.26641538739204407, 'train/grad_norm_pre_clip': 19.47114372253418, 'timing/data': 0.0012673409946728498, 'timing/model': 2.099879315996077, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.16}) +[2026-06-23 04:02:22,655][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.17114689946174622, 'train/grad_norm_pre_clip': 13.121574401855469, 'timing/data': 0.0013661639823112637, 'timing/model': 2.0986734080070164, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.16}) +[2026-06-23 04:02:26,810][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.18517954647541046, 'train/grad_norm_pre_clip': 16.179487228393555, 'timing/data': 0.0012523530167527497, 'timing/model': 2.097875544015551, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.17}) +[2026-06-23 04:02:31,112][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.12683835625648499, 'train/grad_norm_pre_clip': 19.483627319335938, 'timing/data': 0.1537826510029845, 'timing/model': 2.094963782990817, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.17}) +[2026-06-23 04:02:35,261][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.2064642310142517, 'train/grad_norm_pre_clip': 21.477426528930664, 'timing/data': 0.0011362989898771048, 'timing/model': 2.0932090709975455, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.17}) +[2026-06-23 04:02:39,419][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.21278062462806702, 'train/grad_norm_pre_clip': 48.41388702392578, 'timing/data': 0.0011190070072188973, 'timing/model': 2.1015707750048023, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.17}) +[2026-06-23 04:02:43,567][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.1730339378118515, 'train/grad_norm_pre_clip': 22.015878677368164, 'timing/data': 0.0014633229875471443, 'timing/model': 2.0949167030048557, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.17}) +[2026-06-23 04:02:47,867][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.2993849813938141, 'train/grad_norm_pre_clip': 90.43112182617188, 'timing/data': 0.15526207399670966, 'timing/model': 2.0914737620041706, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.17}) +[2026-06-23 04:02:52,212][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.3568492829799652, 'train/grad_norm_pre_clip': 28.249298095703125, 'timing/data': 0.0011140379938296974, 'timing/model': 2.0911714719841257, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.17}) +[2026-06-23 04:02:56,370][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.23924748599529266, 'train/grad_norm_pre_clip': 25.95477294921875, 'timing/data': 0.0012159559992142022, 'timing/model': 2.1013142360025086, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.17}) +[2026-06-23 04:03:00,528][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.2611655294895172, 'train/grad_norm_pre_clip': 43.569007873535156, 'timing/data': 0.0011979219852946699, 'timing/model': 2.1005122939823195, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.17}) +[2026-06-23 04:03:04,826][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.19005903601646423, 'train/grad_norm_pre_clip': 37.40148162841797, 'timing/data': 0.13812419699388556, 'timing/model': 2.0952634679852054, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.17}) +[2026-06-23 04:03:08,985][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.17860044538974762, 'train/grad_norm_pre_clip': 21.464601516723633, 'timing/data': 0.0012123900232836604, 'timing/model': 2.1015766380005516, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.17}) +[2026-06-23 04:03:13,155][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.29781460762023926, 'train/grad_norm_pre_clip': 45.356868743896484, 'timing/data': 0.0012100049934815615, 'timing/model': 2.115926013007993, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.17}) +[2026-06-23 04:03:17,312][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.195118710398674, 'train/grad_norm_pre_clip': 56.89439392089844, 'timing/data': 0.0012537350121419877, 'timing/model': 2.0959704800043255, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.17}) +[2026-06-23 04:03:21,627][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.09727301448583603, 'train/grad_norm_pre_clip': 29.847557067871094, 'timing/data': 0.1542270709760487, 'timing/model': 2.1041417230153456, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.17}) +[2026-06-23 04:03:25,847][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.15889084339141846, 'train/grad_norm_pre_clip': 21.413408279418945, 'timing/data': 0.0013962090015411377, 'timing/model': 2.106613315991126, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.17}) +[2026-06-23 04:03:30,030][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.29836568236351013, 'train/grad_norm_pre_clip': 61.457000732421875, 'timing/data': 0.0013448350073304027, 'timing/model': 2.123373344016727, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.17}) +[2026-06-23 04:03:34,200][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.2756577730178833, 'train/grad_norm_pre_clip': 48.15758514404297, 'timing/data': 0.0011654130066744983, 'timing/model': 2.101675123005407, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.17}) +[2026-06-23 04:03:38,508][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.30691105127334595, 'train/grad_norm_pre_clip': 33.35221862792969, 'timing/data': 0.1430662159982603, 'timing/model': 2.0989663579966873, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.17}) +[2026-06-23 04:03:42,681][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.1332062929868698, 'train/grad_norm_pre_clip': 18.503442764282227, 'timing/data': 0.0013089280109852552, 'timing/model': 2.11000739200972, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.17}) +[2026-06-23 04:03:46,860][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.2254772186279297, 'train/grad_norm_pre_clip': 18.51951026916504, 'timing/data': 0.0011976120003964752, 'timing/model': 2.1189811270160135, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.17}) +[2026-06-23 04:03:51,035][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.17790545523166656, 'train/grad_norm_pre_clip': 21.34040641784668, 'timing/data': 0.0013829039817210287, 'timing/model': 2.1081799440144096, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.17}) +[2026-06-23 04:03:55,340][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.1908814013004303, 'train/grad_norm_pre_clip': 40.660484313964844, 'timing/data': 0.14094795199343935, 'timing/model': 2.103116760001285, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.17}) +[2026-06-23 04:03:59,602][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.26286402344703674, 'train/grad_norm_pre_clip': 19.054452896118164, 'timing/data': 0.0013163209951017052, 'timing/model': 2.201060224993853, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.17}) +[2026-06-23 04:04:03,768][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.20489685237407684, 'train/grad_norm_pre_clip': 20.381103515625, 'timing/data': 0.0013566359994001687, 'timing/model': 2.1053292249853257, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.17}) +[2026-06-23 04:04:07,930][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.23025178909301758, 'train/grad_norm_pre_clip': 21.3907527923584, 'timing/data': 0.0011665050114970654, 'timing/model': 2.105831852997653, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.17}) +[2026-06-23 04:04:12,251][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.24074789881706238, 'train/grad_norm_pre_clip': 55.8630256652832, 'timing/data': 0.14887063502101228, 'timing/model': 2.1068171819788404, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.17}) +[2026-06-23 04:04:16,441][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.19363664090633392, 'train/grad_norm_pre_clip': 18.110708236694336, 'timing/data': 0.0012601780181284994, 'timing/model': 2.107962847017916, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.18}) +[2026-06-23 04:04:20,610][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.3148278295993805, 'train/grad_norm_pre_clip': 121.31798553466797, 'timing/data': 0.0011216120037715882, 'timing/model': 2.0993229059968144, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.18}) +[2026-06-23 04:04:24,772][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.31622055172920227, 'train/grad_norm_pre_clip': 65.22862243652344, 'timing/data': 0.0011149889905937016, 'timing/model': 2.1077447019924875, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.18}) +[2026-06-23 04:04:29,080][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.34225302934646606, 'train/grad_norm_pre_clip': 59.19952392578125, 'timing/data': 0.14544529898557812, 'timing/model': 2.100453663006192, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.18}) +[2026-06-23 04:04:33,244][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.12107237428426743, 'train/grad_norm_pre_clip': 21.09965705871582, 'timing/data': 0.001202601008117199, 'timing/model': 2.1069158079917543, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.18}) +[2026-06-23 04:04:37,444][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.1481170952320099, 'train/grad_norm_pre_clip': 18.153226852416992, 'timing/data': 0.0012055269908159971, 'timing/model': 2.1089486929995473, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.18}) +[2026-06-23 04:04:41,639][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.377098947763443, 'train/grad_norm_pre_clip': 59.520668029785156, 'timing/data': 0.0014847019920125604, 'timing/model': 2.1300207349995617, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.18}) +[2026-06-23 04:04:45,948][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.26290765404701233, 'train/grad_norm_pre_clip': 63.1501579284668, 'timing/data': 0.13597817099071108, 'timing/model': 2.0991067519935314, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.18}) +[2026-06-23 04:04:50,121][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.21180391311645508, 'train/grad_norm_pre_clip': 38.346961975097656, 'timing/data': 0.0014302310009952635, 'timing/model': 2.1098667659971397, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.18}) +[2026-06-23 04:04:54,285][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.22781796753406525, 'train/grad_norm_pre_clip': 31.643993377685547, 'timing/data': 0.0013448940007947385, 'timing/model': 2.1005784070002846, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.18}) +[2026-06-23 04:04:58,473][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.18421810865402222, 'train/grad_norm_pre_clip': 34.35776901245117, 'timing/data': 0.0014312939892988652, 'timing/model': 2.116916677012341, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.18}) +[2026-06-23 04:05:02,785][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.14207573235034943, 'train/grad_norm_pre_clip': 29.55980110168457, 'timing/data': 0.14408680898486637, 'timing/model': 2.105396616010694, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.18}) +[2026-06-23 04:05:06,973][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.3427700698375702, 'train/grad_norm_pre_clip': 42.61980438232422, 'timing/data': 0.0014730409893672913, 'timing/model': 2.1166515040094964, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.18}) +[2026-06-23 04:05:11,394][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.09647537022829056, 'train/grad_norm_pre_clip': 17.041351318359375, 'timing/data': 0.001347919984254986, 'timing/model': 2.3592154030047823, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.18}) +[2026-06-23 04:05:15,576][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.21358637511730194, 'train/grad_norm_pre_clip': 23.252534866333008, 'timing/data': 0.001173527998616919, 'timing/model': 2.1161152519925963, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.18}) +[2026-06-23 04:05:19,879][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.2592438757419586, 'train/grad_norm_pre_clip': 44.27426528930664, 'timing/data': 0.1548117590136826, 'timing/model': 2.09533163101878, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.18}) +[2026-06-23 04:05:24,047][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.11741675436496735, 'train/grad_norm_pre_clip': 16.803396224975586, 'timing/data': 0.0011790870048571378, 'timing/model': 2.1033474059950095, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.18}) +[2026-06-23 04:05:28,211][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.1338895708322525, 'train/grad_norm_pre_clip': 38.25316619873047, 'timing/data': 0.001143300993135199, 'timing/model': 2.1052093140024226, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.18}) +[2026-06-23 04:05:32,389][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.25105375051498413, 'train/grad_norm_pre_clip': 41.14521408081055, 'timing/data': 0.0012149540125392377, 'timing/model': 2.106028147012694, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.18}) +[2026-06-23 04:05:36,700][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.23464101552963257, 'train/grad_norm_pre_clip': 45.284122467041016, 'timing/data': 0.1464561749889981, 'timing/model': 2.1035418729879893, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.18}) +[2026-06-23 04:05:40,872][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.20516301691532135, 'train/grad_norm_pre_clip': 25.725833892822266, 'timing/data': 0.0011918309901375324, 'timing/model': 2.1059803559910506, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.18}) +[2026-06-23 04:05:45,151][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.21204982697963715, 'train/grad_norm_pre_clip': 31.990644454956055, 'timing/data': 0.0012189520057290792, 'timing/model': 2.217595142981736, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.18}) +[2026-06-23 04:05:49,322][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.1113496869802475, 'train/grad_norm_pre_clip': 19.423860549926758, 'timing/data': 0.0012609589903149754, 'timing/model': 2.1094218009966426, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.18}) +[2026-06-23 04:05:53,639][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.12722326815128326, 'train/grad_norm_pre_clip': 14.78674602508545, 'timing/data': 0.15172084898222238, 'timing/model': 2.1026920400036033, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.18}) +[2026-06-23 04:05:57,804][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.28198713064193726, 'train/grad_norm_pre_clip': 45.8757209777832, 'timing/data': 0.001390699006151408, 'timing/model': 2.105077265005093, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.18}) +[2026-06-23 04:06:01,967][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.2919842302799225, 'train/grad_norm_pre_clip': 47.1696662902832, 'timing/data': 0.0013340239820536226, 'timing/model': 2.101161130995024, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.18}) +[2026-06-23 04:06:06,139][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.29609355330467224, 'train/grad_norm_pre_clip': 38.28792190551758, 'timing/data': 0.0012371650082059205, 'timing/model': 2.1059727370156907, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.19}) +[2026-06-23 04:06:10,442][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.09139347076416016, 'train/grad_norm_pre_clip': 28.4323787689209, 'timing/data': 0.14120207898668014, 'timing/model': 2.1032134990091436, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.19}) +[2026-06-23 04:06:14,611][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.4269559681415558, 'train/grad_norm_pre_clip': 126.64258575439453, 'timing/data': 0.0014022600080352277, 'timing/model': 2.1032398769748397, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.19}) +[2026-06-23 04:06:19,029][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.30626705288887024, 'train/grad_norm_pre_clip': 73.77566528320312, 'timing/data': 0.0011903079866897315, 'timing/model': 2.3579043159843422, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.19}) +[2026-06-23 04:06:23,200][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.46512311697006226, 'train/grad_norm_pre_clip': 112.6328353881836, 'timing/data': 0.001215975993545726, 'timing/model': 2.100888101995224, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.19}) +[2026-06-23 04:06:27,509][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.3539974093437195, 'train/grad_norm_pre_clip': 74.23574829101562, 'timing/data': 0.14423706498928368, 'timing/model': 2.103145631001098, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.19}) +[2026-06-23 04:06:31,697][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.18875914812088013, 'train/grad_norm_pre_clip': 29.990293502807617, 'timing/data': 0.0012173580180387944, 'timing/model': 2.1258115870004985, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.19}) +[2026-06-23 04:06:35,872][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.24476705491542816, 'train/grad_norm_pre_clip': 27.25580406188965, 'timing/data': 0.0012487169879022986, 'timing/model': 2.100403284013737, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.19}) +[2026-06-23 04:06:40,049][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.19934025406837463, 'train/grad_norm_pre_clip': 26.84381103515625, 'timing/data': 0.00135345000308007, 'timing/model': 2.1172810590069275, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.19}) +[2026-06-23 04:06:44,374][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.16339111328125, 'train/grad_norm_pre_clip': 29.461278915405273, 'timing/data': 0.1512936590006575, 'timing/model': 2.109154616977321, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.19}) +[2026-06-23 04:06:48,576][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.20240047574043274, 'train/grad_norm_pre_clip': 15.375136375427246, 'timing/data': 0.0014434460026677698, 'timing/model': 2.1252139400166925, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.19}) +[2026-06-23 04:06:52,751][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.29161393642425537, 'train/grad_norm_pre_clip': 69.59069061279297, 'timing/data': 0.0012629929988179356, 'timing/model': 2.1069001109863166, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.19}) +[2026-06-23 04:06:57,171][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.13331764936447144, 'train/grad_norm_pre_clip': 17.85120391845703, 'timing/data': 0.0012946009810548276, 'timing/model': 2.1088793459930457, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.19}) +[2026-06-23 04:07:01,490][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.14370635151863098, 'train/grad_norm_pre_clip': 14.983851432800293, 'timing/data': 0.14144713198766112, 'timing/model': 2.109268975997111, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.19}) +[2026-06-23 04:07:05,665][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.14698898792266846, 'train/grad_norm_pre_clip': 32.33475875854492, 'timing/data': 0.0013450350088533014, 'timing/model': 2.1069749880116433, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.19}) +[2026-06-23 04:07:09,837][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.12713538110256195, 'train/grad_norm_pre_clip': 22.04584312438965, 'timing/data': 0.00122299900976941, 'timing/model': 2.107835084985709, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.19}) +[2026-06-23 04:07:14,013][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.12739065289497375, 'train/grad_norm_pre_clip': 22.599836349487305, 'timing/data': 0.0013240159896668047, 'timing/model': 2.113918881979771, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.19}) +[2026-06-23 04:07:18,326][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.22378040850162506, 'train/grad_norm_pre_clip': 34.986454010009766, 'timing/data': 0.15159649899578653, 'timing/model': 2.1026005349995103, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.19}) +[2026-06-23 04:07:22,499][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.3077892065048218, 'train/grad_norm_pre_clip': 55.575103759765625, 'timing/data': 0.0011687990045174956, 'timing/model': 2.110250805999385, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.19}) +[2026-06-23 04:07:26,656][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.23017586767673492, 'train/grad_norm_pre_clip': 55.015010833740234, 'timing/data': 0.0011325019877403975, 'timing/model': 2.0997402600187343, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.19}) +[2026-06-23 04:07:30,932][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.14732009172439575, 'train/grad_norm_pre_clip': 19.101789474487305, 'timing/data': 0.001162025990197435, 'timing/model': 2.1384669129911344, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.19}) +[2026-06-23 04:07:35,243][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 500 +[2026-06-23 04:09:21,792][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 500 +[2026-06-23 04:09:21,793][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 500 +[2026-06-23 04:11:39,043][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 500 +[2026-06-23 04:13:03,748][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-23 04:13:03,750][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.25135886669158936, 'train/grad_norm_pre_clip': 68.50296020507812, 'eval/action_loss/seconds': 106.54492208999, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.14792080223560333, 'eval/latency_2/loss': 0.14792080223560333, 'eval/flappy/loss': 0.14792080223560333, 'eval/flappy/latency_2/loss': 0.14792080223560333, 'eval/flap_precision': 0.7811158798283262, 'eval/flap_recall': 0.9824561403508771, 'eval/flap_f1': 0.8702928870292888, 'eval/flap_rate_model': 0.2588888888888889, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8702928870292888, 'eval/latency_2/flap_precision': 0.7811158798283262, 'eval/latency_2/flap_recall': 0.9824561403508771, 'eval/latency_2/flap_f1': 0.8702928870292888, 'eval/latency_2/flap_rate_model': 0.2588888888888889, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8702928870292888, 'eval/action_classification/data_wait_seconds': 50.28413772583008, 'eval/action_classification/predict_seconds': 85.24114990234375, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 19.920000000000094, 'rl_games_eval/mid_train/mean_length': 179.6, 'rl_games_eval/mid_train/std_reward': 21.023834093713926, 'rl_games_eval/mid_train/std_length': 170.6113712505705, 'rl_games_eval/mid_train/macro_mean_reward': 19.920000000000094, 'rl_games_eval/mid_train/macro_mean_length': 179.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 19.920000000000094, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 179.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 21.023834093713926, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 170.6113712505705, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 19.920000000000094, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.15210109000327066, 'timing/model': 2.1008229790022597, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.19}) +[2026-06-23 04:13:03,751][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_500_state +[2026-06-23 04:13:03,758][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 04:13:14,298][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_500_state/model.safetensors +[2026-06-23 04:13:21,366][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_500_state/optimizer.bin +[2026-06-23 04:13:21,367][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_500_state/sampler.bin +[2026-06-23 04:13:21,367][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_500_state/sampler_1.bin +[2026-06-23 04:13:21,368][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-23 04:13:21,371][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-23 04:13:21,371][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 04:13:21,397][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 04:13:25,255][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.339787095785141, 'train/grad_norm_pre_clip': 69.58438110351562, 'timing/data': 0.0011996249959338456, 'timing/model': 1.9507123879739083, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.19}) +[2026-06-23 04:13:29,129][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.21153542399406433, 'train/grad_norm_pre_clip': 47.533058166503906, 'timing/data': 0.001120781002100557, 'timing/model': 1.961844199016923, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.19}) +[2026-06-23 04:13:33,079][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.21641112864017487, 'train/grad_norm_pre_clip': 46.14798355102539, 'timing/data': 0.0012289800215512514, 'timing/model': 2.0317500320088584, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.19}) +[2026-06-23 04:13:37,127][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.203692227602005, 'train/grad_norm_pre_clip': 61.33972930908203, 'timing/data': 0.1596123719937168, 'timing/model': 1.9677362109941896, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.2}) +[2026-06-23 04:13:41,014][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.30463171005249023, 'train/grad_norm_pre_clip': 47.89025115966797, 'timing/data': 0.0012139119789935648, 'timing/model': 1.9651657629874535, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.2}) +[2026-06-23 04:13:44,928][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.21859723329544067, 'train/grad_norm_pre_clip': 66.25186157226562, 'timing/data': 0.001681666006334126, 'timing/model': 1.9854332100076135, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.2}) +[2026-06-23 04:13:48,828][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.27901405096054077, 'train/grad_norm_pre_clip': 51.28418731689453, 'timing/data': 0.0012289099977351725, 'timing/model': 1.9724524929770268, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.2}) +[2026-06-23 04:13:52,889][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.15983429551124573, 'train/grad_norm_pre_clip': 42.69247817993164, 'timing/data': 0.15305632198578678, 'timing/model': 1.9757211300020572, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.2}) +[2026-06-23 04:13:56,824][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.147447407245636, 'train/grad_norm_pre_clip': 35.59381866455078, 'timing/data': 0.0013349450018722564, 'timing/model': 1.9969673630257603, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.2}) +[2026-06-23 04:14:00,757][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.1405973583459854, 'train/grad_norm_pre_clip': 19.362268447875977, 'timing/data': 0.001643434981815517, 'timing/model': 1.9938953590171877, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.2}) +[2026-06-23 04:14:04,685][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.12764321267604828, 'train/grad_norm_pre_clip': 33.62973403930664, 'timing/data': 0.0012866769975516945, 'timing/model': 1.9852993900130969, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.2}) +[2026-06-23 04:14:08,888][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.16224762797355652, 'train/grad_norm_pre_clip': 28.05668830871582, 'timing/data': 0.16117483700509183, 'timing/model': 1.9853354869992472, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.2}) +[2026-06-23 04:14:12,838][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.0986785814166069, 'train/grad_norm_pre_clip': 12.172832489013672, 'timing/data': 0.0012784720165655017, 'timing/model': 2.0008652919786982, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.2}) +[2026-06-23 04:14:16,788][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.28658804297447205, 'train/grad_norm_pre_clip': 50.827728271484375, 'timing/data': 0.0012213459995109588, 'timing/model': 2.0049323340062983, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.2}) +[2026-06-23 04:14:20,723][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.2004704773426056, 'train/grad_norm_pre_clip': 21.524757385253906, 'timing/data': 0.001232976996107027, 'timing/model': 1.9883154499984812, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.2}) +[2026-06-23 04:14:24,829][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.1352858543395996, 'train/grad_norm_pre_clip': 30.40208625793457, 'timing/data': 0.16120255898567848, 'timing/model': 1.9904434900090564, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.2}) +[2026-06-23 04:14:28,772][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.16742803156375885, 'train/grad_norm_pre_clip': 22.475183486938477, 'timing/data': 0.0014545459998771548, 'timing/model': 1.9915216120134573, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.2}) +[2026-06-23 04:14:32,736][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.2771252691745758, 'train/grad_norm_pre_clip': 62.693729400634766, 'timing/data': 0.0013339339930098504, 'timing/model': 2.0096698440029286, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.2}) +[2026-06-23 04:14:36,700][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.33768194913864136, 'train/grad_norm_pre_clip': 69.94649505615234, 'timing/data': 0.001382513000862673, 'timing/model': 1.9971537140081637, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.2}) +[2026-06-23 04:14:40,837][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.1546916514635086, 'train/grad_norm_pre_clip': 17.40566635131836, 'timing/data': 0.1505634470086079, 'timing/model': 2.0287278150208294, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.2}) +[2026-06-23 04:14:44,792][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.14818908274173737, 'train/grad_norm_pre_clip': 26.792394638061523, 'timing/data': 0.001209294016007334, 'timing/model': 1.998966212006053, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.2}) +[2026-06-23 04:14:48,750][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.19745376706123352, 'train/grad_norm_pre_clip': 33.870399475097656, 'timing/data': 0.0012460920261219144, 'timing/model': 1.9980056280037388, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.2}) +[2026-06-23 04:14:52,710][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.23024506866931915, 'train/grad_norm_pre_clip': 32.20158004760742, 'timing/data': 0.0012806649901904166, 'timing/model': 2.0025775209942367, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.2}) +[2026-06-23 04:14:56,822][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.12774641811847687, 'train/grad_norm_pre_clip': 24.898632049560547, 'timing/data': 0.15080930100521073, 'timing/model': 2.003370235004695, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.2}) +[2026-06-23 04:15:00,782][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.16344669461250305, 'train/grad_norm_pre_clip': 32.034637451171875, 'timing/data': 0.0012147430097684264, 'timing/model': 2.0005191790114623, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.2}) +[2026-06-23 04:15:04,875][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.19772927463054657, 'train/grad_norm_pre_clip': 19.702688217163086, 'timing/data': 0.0014202430029399693, 'timing/model': 2.015242473979015, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.2}) +[2026-06-23 04:15:08,899][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.17766857147216797, 'train/grad_norm_pre_clip': 19.815034866333008, 'timing/data': 0.001378526008920744, 'timing/model': 2.0029549459868576, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.2}) +[2026-06-23 04:15:13,058][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.12575551867485046, 'train/grad_norm_pre_clip': 20.036937713623047, 'timing/data': 0.19899275098578073, 'timing/model': 2.0019665650033858, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.2}) +[2026-06-23 04:15:17,199][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.17398260533809662, 'train/grad_norm_pre_clip': 34.87105941772461, 'timing/data': 0.0013923140068072826, 'timing/model': 2.0146632039977703, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.2}) +[2026-06-23 04:15:21,173][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.1739560067653656, 'train/grad_norm_pre_clip': 26.49465560913086, 'timing/data': 0.0015203020011540502, 'timing/model': 2.00658852502238, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.21}) +[2026-06-23 04:15:25,159][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.2461862713098526, 'train/grad_norm_pre_clip': 24.027151107788086, 'timing/data': 0.0016040059854276478, 'timing/model': 2.020395136991283, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.21}) +[2026-06-23 04:15:29,293][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.2682018280029297, 'train/grad_norm_pre_clip': 27.874061584472656, 'timing/data': 0.16564797298633493, 'timing/model': 2.0065869959944393, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.21}) +[2026-06-23 04:15:33,269][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.24707497656345367, 'train/grad_norm_pre_clip': 29.894304275512695, 'timing/data': 0.0013264819863252342, 'timing/model': 2.0089745639998, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.21}) +[2026-06-23 04:15:37,245][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.25847703218460083, 'train/grad_norm_pre_clip': 44.60707092285156, 'timing/data': 0.0015986259968485683, 'timing/model': 2.010295029002009, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.21}) +[2026-06-23 04:15:41,222][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.20805436372756958, 'train/grad_norm_pre_clip': 38.68949890136719, 'timing/data': 0.0012022430018987507, 'timing/model': 2.0127148759784177, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.21}) +[2026-06-23 04:15:45,355][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.20424163341522217, 'train/grad_norm_pre_clip': 24.76104736328125, 'timing/data': 0.15585364401340485, 'timing/model': 2.0057067809975706, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.21}) +[2026-06-23 04:15:49,575][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.12812283635139465, 'train/grad_norm_pre_clip': 13.384540557861328, 'timing/data': 0.0014137339894659817, 'timing/model': 2.249526022991631, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.21}) +[2026-06-23 04:15:53,568][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.22496479749679565, 'train/grad_norm_pre_clip': 21.724489212036133, 'timing/data': 0.0013713149819523096, 'timing/model': 2.0136855989985634, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.21}) +[2026-06-23 04:15:57,551][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.20821456611156464, 'train/grad_norm_pre_clip': 18.643640518188477, 'timing/data': 0.0012296529894229025, 'timing/model': 2.0145458539773244, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.21}) +[2026-06-23 04:16:01,686][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.20703324675559998, 'train/grad_norm_pre_clip': 26.06545066833496, 'timing/data': 0.15680056699784473, 'timing/model': 2.00911613900098, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.21}) +[2026-06-23 04:16:05,665][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.23612479865550995, 'train/grad_norm_pre_clip': 34.987613677978516, 'timing/data': 0.0012161790218669921, 'timing/model': 2.0130909020081162, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.21}) +[2026-06-23 04:16:09,675][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.12017624080181122, 'train/grad_norm_pre_clip': 30.169254302978516, 'timing/data': 0.0012598699831869453, 'timing/model': 2.033185049018357, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.21}) +[2026-06-23 04:16:13,654][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.28838202357292175, 'train/grad_norm_pre_clip': 45.96595001220703, 'timing/data': 0.0014104779984336346, 'timing/model': 2.01038349600276, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.21}) +[2026-06-23 04:16:17,785][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.1584692895412445, 'train/grad_norm_pre_clip': 21.832246780395508, 'timing/data': 0.15673098000115715, 'timing/model': 2.008126696018735, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.21}) +[2026-06-23 04:16:21,761][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.2677273750305176, 'train/grad_norm_pre_clip': 40.44804763793945, 'timing/data': 0.0012418449914548546, 'timing/model': 2.010942278022412, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.21}) +[2026-06-23 04:16:25,980][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.2646142244338989, 'train/grad_norm_pre_clip': 28.61139488220215, 'timing/data': 0.0014115899975877255, 'timing/model': 2.0091202330077067, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.21}) +[2026-06-23 04:16:29,958][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.23410716652870178, 'train/grad_norm_pre_clip': 27.692523956298828, 'timing/data': 0.0011477819934953004, 'timing/model': 2.0121250840020366, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.21}) +[2026-06-23 04:16:34,334][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.1597791314125061, 'train/grad_norm_pre_clip': 17.659738540649414, 'timing/data': 0.4017201219976414, 'timing/model': 2.0098788140167017, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.21}) +[2026-06-23 04:16:38,311][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.18010126054286957, 'train/grad_norm_pre_clip': 40.745574951171875, 'timing/data': 0.0011480830144137144, 'timing/model': 2.00947204901604, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.21}) +[2026-06-23 04:16:42,314][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.07011277228593826, 'train/grad_norm_pre_clip': 14.804631233215332, 'timing/data': 0.0012578559981193393, 'timing/model': 2.024658662005095, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.21}) +[2026-06-23 04:16:46,304][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.19678519666194916, 'train/grad_norm_pre_clip': 19.882465362548828, 'timing/data': 0.0011959710100200027, 'timing/model': 2.0168538999860175, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.21}) +[2026-06-23 04:16:50,444][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.1643168330192566, 'train/grad_norm_pre_clip': 18.139020919799805, 'timing/data': 0.16004262698697858, 'timing/model': 2.010448622982949, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.21}) +[2026-06-23 04:16:54,426][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.12870614230632782, 'train/grad_norm_pre_clip': 15.157340049743652, 'timing/data': 0.0011517390084918588, 'timing/model': 2.0119380219839513, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.21}) +[2026-06-23 04:16:58,606][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.21990883350372314, 'train/grad_norm_pre_clip': 38.619903564453125, 'timing/data': 0.0013076780014671385, 'timing/model': 2.026925507991109, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.21}) +[2026-06-23 04:17:02,598][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.1755581796169281, 'train/grad_norm_pre_clip': 21.484716415405273, 'timing/data': 0.0012179109908174723, 'timing/model': 2.0163406950014178, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.21}) +[2026-06-23 04:17:06,737][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.12162140011787415, 'train/grad_norm_pre_clip': 17.702157974243164, 'timing/data': 0.14976845600176603, 'timing/model': 2.014609862002544, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.22}) +[2026-06-23 04:17:10,736][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.12235704064369202, 'train/grad_norm_pre_clip': 20.562305450439453, 'timing/data': 0.0014053889899514616, 'timing/model': 2.0263591810071375, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.22}) +[2026-06-23 04:17:14,748][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.3285300135612488, 'train/grad_norm_pre_clip': 32.784759521484375, 'timing/data': 0.001231255999300629, 'timing/model': 2.0337154979933985, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.22}) +[2026-06-23 04:17:18,726][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.30057936906814575, 'train/grad_norm_pre_clip': 53.03285598754883, 'timing/data': 0.0011376329930499196, 'timing/model': 2.010893525002757, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.22}) +[2026-06-23 04:17:22,859][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.19182471930980682, 'train/grad_norm_pre_clip': 39.99893569946289, 'timing/data': 0.15194756799610332, 'timing/model': 2.0102265989989974, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.22}) +[2026-06-23 04:17:26,839][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.14531075954437256, 'train/grad_norm_pre_clip': 33.760318756103516, 'timing/data': 0.0011369420099072158, 'timing/model': 2.012997179001104, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.22}) +[2026-06-23 04:17:30,881][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.3176656663417816, 'train/grad_norm_pre_clip': 52.88766098022461, 'timing/data': 0.0014010699815116823, 'timing/model': 2.0675102949899156, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.22}) +[2026-06-23 04:17:34,877][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.18755823373794556, 'train/grad_norm_pre_clip': 21.795991897583008, 'timing/data': 0.0012317569926381111, 'timing/model': 2.021248298988212, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.22}) +[2026-06-23 04:17:39,020][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.18357083201408386, 'train/grad_norm_pre_clip': 28.783578872680664, 'timing/data': 0.15595796500565484, 'timing/model': 2.0150428950146306, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.22}) +[2026-06-23 04:17:43,029][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.09777966141700745, 'train/grad_norm_pre_clip': 29.56623077392578, 'timing/data': 0.0014101660053711385, 'timing/model': 2.027261520997854, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.22}) +[2026-06-23 04:17:47,037][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.11068528145551682, 'train/grad_norm_pre_clip': 29.787511825561523, 'timing/data': 0.001211339986184612, 'timing/model': 2.0271782769996207, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.22}) +[2026-06-23 04:17:51,052][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.10106831043958664, 'train/grad_norm_pre_clip': 17.732872009277344, 'timing/data': 0.0012438199773896486, 'timing/model': 2.037056336994283, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.22}) +[2026-06-23 04:17:55,206][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.14641627669334412, 'train/grad_norm_pre_clip': 33.82595443725586, 'timing/data': 0.15329277198179625, 'timing/model': 2.022726233000867, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.22}) +[2026-06-23 04:17:59,208][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.25781136751174927, 'train/grad_norm_pre_clip': 55.28705596923828, 'timing/data': 0.0012290919839870185, 'timing/model': 2.021292632998666, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.22}) +[2026-06-23 04:18:03,209][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.15529648959636688, 'train/grad_norm_pre_clip': 57.94041061401367, 'timing/data': 0.0012263169919606298, 'timing/model': 2.02559731999645, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.22}) +[2026-06-23 04:18:07,377][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.1611163169145584, 'train/grad_norm_pre_clip': 52.4890251159668, 'timing/data': 0.0012243340024724603, 'timing/model': 2.0176643939921632, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.22}) +[2026-06-23 04:18:11,540][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.27791276574134827, 'train/grad_norm_pre_clip': 28.657211303710938, 'timing/data': 0.1602621589845512, 'timing/model': 2.027232089982135, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.22}) +[2026-06-23 04:18:15,544][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.14489635825157166, 'train/grad_norm_pre_clip': 45.18129348754883, 'timing/data': 0.0012302749964874238, 'timing/model': 2.017955119983526, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.22}) +[2026-06-23 04:18:19,554][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.06847581267356873, 'train/grad_norm_pre_clip': 21.830596923828125, 'timing/data': 0.00125137300346978, 'timing/model': 2.021042648004368, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.22}) +[2026-06-23 04:18:23,561][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.1845693737268448, 'train/grad_norm_pre_clip': 29.250051498413086, 'timing/data': 0.0014290420222096145, 'timing/model': 2.0288905460038222, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.22}) +[2026-06-23 04:18:27,753][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.2208566963672638, 'train/grad_norm_pre_clip': 28.673738479614258, 'timing/data': 0.16347622399916872, 'timing/model': 2.0416396119981073, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.22}) +[2026-06-23 04:18:31,760][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.1018303707242012, 'train/grad_norm_pre_clip': 22.38895606994629, 'timing/data': 0.0012692859745584428, 'timing/model': 2.0298533349996433, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.22}) +[2026-06-23 04:18:35,746][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.11365051567554474, 'train/grad_norm_pre_clip': 18.496726989746094, 'timing/data': 0.0011688309896271676, 'timing/model': 2.014257306989748, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.22}) +[2026-06-23 04:18:39,761][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.09066846966743469, 'train/grad_norm_pre_clip': 18.75623321533203, 'timing/data': 0.0015049419889692217, 'timing/model': 2.015210740006296, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.22}) +[2026-06-23 04:18:43,905][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.20750336349010468, 'train/grad_norm_pre_clip': 51.054229736328125, 'timing/data': 0.15121218701824546, 'timing/model': 2.023227380006574, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.22}) +[2026-06-23 04:18:47,938][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.2857521176338196, 'train/grad_norm_pre_clip': 57.06515121459961, 'timing/data': 0.00142518401844427, 'timing/model': 2.058698248991277, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.22}) +[2026-06-23 04:18:51,956][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.1895621418952942, 'train/grad_norm_pre_clip': 23.35546112060547, 'timing/data': 0.0014821700169704854, 'timing/model': 2.033719986997312, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.23}) +[2026-06-23 04:18:55,946][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.18539461493492126, 'train/grad_norm_pre_clip': 44.4937629699707, 'timing/data': 0.001250280998647213, 'timing/model': 2.0153126530058216, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.23}) +[2026-06-23 04:19:00,093][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.18507233262062073, 'train/grad_norm_pre_clip': 22.175975799560547, 'timing/data': 0.16077450700686313, 'timing/model': 2.0144687639840413, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.23}) +[2026-06-23 04:19:04,089][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.24297301471233368, 'train/grad_norm_pre_clip': 27.333419799804688, 'timing/data': 0.0012043360038660467, 'timing/model': 2.0245623440132476, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.23}) +[2026-06-23 04:19:08,096][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.28259986639022827, 'train/grad_norm_pre_clip': 41.35620880126953, 'timing/data': 0.0013824159977957606, 'timing/model': 2.033790924004279, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.23}) +[2026-06-23 04:19:12,329][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.13002027571201324, 'train/grad_norm_pre_clip': 18.724809646606445, 'timing/data': 0.0012163990177214146, 'timing/model': 2.2585146549972706, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.23}) +[2026-06-23 04:19:16,471][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.12508220970630646, 'train/grad_norm_pre_clip': 25.0380916595459, 'timing/data': 0.15066430298611522, 'timing/model': 2.0170717230066657, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.23}) +[2026-06-23 04:19:20,479][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.39849650859832764, 'train/grad_norm_pre_clip': 27.45224380493164, 'timing/data': 0.0013881759950891137, 'timing/model': 2.033233107009437, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.23}) +[2026-06-23 04:19:24,491][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.244734525680542, 'train/grad_norm_pre_clip': 29.839080810546875, 'timing/data': 0.0012196939787827432, 'timing/model': 2.0339374920004047, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.23}) +[2026-06-23 04:19:28,481][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.32548004388809204, 'train/grad_norm_pre_clip': 56.50105667114258, 'timing/data': 0.0014263369957916439, 'timing/model': 2.015228322998155, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.23}) +[2026-06-23 04:19:32,621][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.280608594417572, 'train/grad_norm_pre_clip': 53.14968490600586, 'timing/data': 0.15566673499415629, 'timing/model': 2.015071864996571, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.23}) +[2026-06-23 04:19:36,612][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.2008907049894333, 'train/grad_norm_pre_clip': 43.42024612426758, 'timing/data': 0.001138895022450015, 'timing/model': 2.024893362977309, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.23}) +[2026-06-23 04:19:40,613][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.250802218914032, 'train/grad_norm_pre_clip': 21.038379669189453, 'timing/data': 0.0011126870231237262, 'timing/model': 2.0136015939933714, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.23}) +[2026-06-23 04:19:44,613][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.12520940601825714, 'train/grad_norm_pre_clip': 16.060083389282227, 'timing/data': 0.00161280098836869, 'timing/model': 2.023270189005416, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.23}) +[2026-06-23 04:19:49,013][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.11929704248905182, 'train/grad_norm_pre_clip': 23.04131317138672, 'timing/data': 0.16312018598546274, 'timing/model': 2.0095952750125434, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.23}) +[2026-06-23 04:19:53,004][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.16250841319561005, 'train/grad_norm_pre_clip': 13.590133666992188, 'timing/data': 0.0011091000051237643, 'timing/model': 2.022799444006523, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.23}) +[2026-06-23 04:19:57,001][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.19742348790168762, 'train/grad_norm_pre_clip': 25.963571548461914, 'timing/data': 0.001367046992527321, 'timing/model': 2.024785116984276, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.23}) +[2026-06-23 04:20:00,997][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.11033499240875244, 'train/grad_norm_pre_clip': 12.31117057800293, 'timing/data': 0.0013045610103290528, 'timing/model': 2.0229009189934004, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.23}) +[2026-06-23 04:20:05,146][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 600 +[2026-06-23 04:22:01,520][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 600 +[2026-06-23 04:22:01,521][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 600 +[2026-06-23 04:24:08,091][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 600 +[2026-06-23 04:24:08,092][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.1978667974472046, 'train/grad_norm_pre_clip': 18.19789695739746, 'eval/action_loss/seconds': 116.37102154898457, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.1009010374546051, 'eval/latency_2/loss': 0.1009010374546051, 'eval/flappy/loss': 0.1009010374546051, 'eval/flappy/latency_2/loss': 0.1009010374546051, 'eval/flap_precision': 0.914001421464108, 'eval/flap_recall': 0.8677462887989204, 'eval/flap_f1': 0.8902734510211147, 'eval/flap_rate_model': 0.19541666666666666, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8902734510211147, 'eval/latency_2/flap_precision': 0.914001421464108, 'eval/latency_2/flap_recall': 0.8677462887989204, 'eval/latency_2/flap_f1': 0.8902734510211147, 'eval/latency_2/flap_rate_model': 0.19541666666666666, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8902734510211147, 'eval/action_classification/data_wait_seconds': 50.2283935546875, 'eval/action_classification/predict_seconds': 74.64827728271484, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.15388556200196035, 'timing/model': 2.014443358988501, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.23}) +[2026-06-23 04:24:12,227][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.1807241290807724, 'train/grad_norm_pre_clip': 25.471765518188477, 'timing/data': 0.0012950630043633282, 'timing/model': 1.9916493839991745, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.23}) +[2026-06-23 04:24:16,166][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.2074311375617981, 'train/grad_norm_pre_clip': 21.95115852355957, 'timing/data': 0.0011208619980607182, 'timing/model': 1.985905409004772, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.23}) +[2026-06-23 04:24:20,115][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.27694377303123474, 'train/grad_norm_pre_clip': 28.0709228515625, 'timing/data': 0.0011607150081545115, 'timing/model': 2.007226409012219, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.23}) +[2026-06-23 04:24:24,216][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.21579097211360931, 'train/grad_norm_pre_clip': 32.75543212890625, 'timing/data': 0.1527163489954546, 'timing/model': 1.9924707320169546, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.23}) +[2026-06-23 04:24:28,161][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.20370152592658997, 'train/grad_norm_pre_clip': 30.589035034179688, 'timing/data': 0.0012114690034650266, 'timing/model': 1.9931648150086403, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.23}) +[2026-06-23 04:24:32,106][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.14088895916938782, 'train/grad_norm_pre_clip': 24.490190505981445, 'timing/data': 0.001213591982377693, 'timing/model': 1.9928109720058274, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.23}) +[2026-06-23 04:24:36,060][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.16159673035144806, 'train/grad_norm_pre_clip': 39.30843734741211, 'timing/data': 0.001215156022226438, 'timing/model': 1.9979045549989678, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.23}) +[2026-06-23 04:24:40,156][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.1897900402545929, 'train/grad_norm_pre_clip': 15.15095329284668, 'timing/data': 0.14258326101116836, 'timing/model': 1.9951992660062388, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.24}) +[2026-06-23 04:24:44,357][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.09975864738225937, 'train/grad_norm_pre_clip': 16.27023696899414, 'timing/data': 0.0014491479960270226, 'timing/model': 2.247309853002662, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.24}) +[2026-06-23 04:24:48,329][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.1517481803894043, 'train/grad_norm_pre_clip': 19.67121696472168, 'timing/data': 0.0016806760104373097, 'timing/model': 2.0022626699937973, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.24}) +[2026-06-23 04:24:52,291][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.250409334897995, 'train/grad_norm_pre_clip': 37.94615173339844, 'timing/data': 0.0014612000086344779, 'timing/model': 2.0005700629844796, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.24}) +[2026-06-23 04:24:56,402][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.07396719604730606, 'train/grad_norm_pre_clip': 13.863374710083008, 'timing/data': 0.15034959898912348, 'timing/model': 2.0033847659942694, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.24}) +[2026-06-23 04:25:00,360][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.16191597282886505, 'train/grad_norm_pre_clip': 26.089956283569336, 'timing/data': 0.0013922819925937802, 'timing/model': 2.0006230830040295, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.24}) +[2026-06-23 04:25:04,326][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.13847042620182037, 'train/grad_norm_pre_clip': 32.2515754699707, 'timing/data': 0.0012673019955400378, 'timing/model': 2.0034786309988704, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.24}) +[2026-06-23 04:25:08,309][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.1711721420288086, 'train/grad_norm_pre_clip': 21.914968490600586, 'timing/data': 0.001394997991155833, 'timing/model': 2.0196585310040973, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.24}) +[2026-06-23 04:25:12,422][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.19657909870147705, 'train/grad_norm_pre_clip': 49.686275482177734, 'timing/data': 0.14964206900913268, 'timing/model': 2.0010211629851256, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.24}) +[2026-06-23 04:25:16,476][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.16652241349220276, 'train/grad_norm_pre_clip': 24.776506423950195, 'timing/data': 0.0012025019968859851, 'timing/model': 2.0836105170019437, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.24}) +[2026-06-23 04:25:20,444][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.2196984440088272, 'train/grad_norm_pre_clip': 38.24138641357422, 'timing/data': 0.0012361250119283795, 'timing/model': 2.006887594994623, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.24}) +[2026-06-23 04:25:24,427][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.22749896347522736, 'train/grad_norm_pre_clip': 49.383914947509766, 'timing/data': 0.0012717809877358377, 'timing/model': 2.018525732011767, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.24}) +[2026-06-23 04:25:28,555][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.166860431432724, 'train/grad_norm_pre_clip': 27.751317977905273, 'timing/data': 0.15503763701417483, 'timing/model': 2.0110768289887346, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.24}) +[2026-06-23 04:25:32,527][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.07901360839605331, 'train/grad_norm_pre_clip': 16.476383209228516, 'timing/data': 0.0012355830112937838, 'timing/model': 2.0073135269922204, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.24}) +[2026-06-23 04:25:36,501][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.27825137972831726, 'train/grad_norm_pre_clip': 56.57760238647461, 'timing/data': 0.0012676329934038222, 'timing/model': 2.006869046977954, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.24}) +[2026-06-23 04:25:40,497][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.21648922562599182, 'train/grad_norm_pre_clip': 43.7589111328125, 'timing/data': 0.0014420750085264444, 'timing/model': 2.0136233090015594, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.24}) +[2026-06-23 04:25:44,617][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.1148107498884201, 'train/grad_norm_pre_clip': 19.10810089111328, 'timing/data': 0.14866489899577573, 'timing/model': 2.007238221995067, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.24}) +[2026-06-23 04:25:48,601][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.09947343170642853, 'train/grad_norm_pre_clip': 29.289958953857422, 'timing/data': 0.001210787013405934, 'timing/model': 2.006259746005526, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.24}) +[2026-06-23 04:25:52,629][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.2204030603170395, 'train/grad_norm_pre_clip': 38.65757751464844, 'timing/data': 0.00139189200126566, 'timing/model': 2.0276899070013314, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.24}) +[2026-06-23 04:25:56,634][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.11232610046863556, 'train/grad_norm_pre_clip': 24.055898666381836, 'timing/data': 0.0012583750067278743, 'timing/model': 2.0212195540079847, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.24}) +[2026-06-23 04:26:00,767][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.1461668163537979, 'train/grad_norm_pre_clip': 25.01487159729004, 'timing/data': 0.15214877401012927, 'timing/model': 2.01637801900506, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.24}) +[2026-06-23 04:26:04,739][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.0714329183101654, 'train/grad_norm_pre_clip': 13.21707820892334, 'timing/data': 0.0011111230123788118, 'timing/model': 2.0080678449885454, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.24}) +[2026-06-23 04:26:08,735][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.13259421288967133, 'train/grad_norm_pre_clip': 29.23052406311035, 'timing/data': 0.0013824240013491362, 'timing/model': 2.03132057399489, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.24}) +[2026-06-23 04:26:12,715][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.24725787341594696, 'train/grad_norm_pre_clip': 28.362329483032227, 'timing/data': 0.0012586959928739816, 'timing/model': 2.0103949289768934, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.24}) +[2026-06-23 04:26:16,851][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.17873282730579376, 'train/grad_norm_pre_clip': 23.931947708129883, 'timing/data': 0.1562574239796959, 'timing/model': 2.0115321530029178, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.24}) +[2026-06-23 04:26:20,853][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.052298784255981445, 'train/grad_norm_pre_clip': 9.738773345947266, 'timing/data': 0.0014730919792782515, 'timing/model': 2.0275525680044666, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.24}) +[2026-06-23 04:26:24,885][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.15670450031757355, 'train/grad_norm_pre_clip': 45.945762634277344, 'timing/data': 0.001215906988363713, 'timing/model': 2.060332751017995, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.25}) +[2026-06-23 04:26:28,872][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.11393928527832031, 'train/grad_norm_pre_clip': 29.3399600982666, 'timing/data': 0.0012069899821653962, 'timing/model': 2.0118081900000107, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.25}) +[2026-06-23 04:26:33,012][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.07311709970235825, 'train/grad_norm_pre_clip': 16.401294708251953, 'timing/data': 0.15951896799379028, 'timing/model': 2.012625160976313, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.25}) +[2026-06-23 04:26:36,998][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.11004648357629776, 'train/grad_norm_pre_clip': 36.291141510009766, 'timing/data': 0.0013413380074780434, 'timing/model': 2.0169174860056955, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.25}) +[2026-06-23 04:26:41,075][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.21273599565029144, 'train/grad_norm_pre_clip': 49.30117416381836, 'timing/data': 0.001296766015002504, 'timing/model': 2.0158519289980177, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.25}) +[2026-06-23 04:26:45,058][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.13166357576847076, 'train/grad_norm_pre_clip': 46.89224624633789, 'timing/data': 0.0012160369951743633, 'timing/model': 2.0112412110029254, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.25}) +[2026-06-23 04:26:49,111][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.21133115887641907, 'train/grad_norm_pre_clip': 23.14324188232422, 'timing/data': 0.060767724993638694, 'timing/model': 2.0163082469953224, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.25}) +[2026-06-23 04:26:53,103][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.5963506698608398, 'train/grad_norm_pre_clip': 130.1705322265625, 'timing/data': 0.0013542920059990138, 'timing/model': 2.0197048069967423, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.25}) +[2026-06-23 04:26:57,175][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.17258894443511963, 'train/grad_norm_pre_clip': 38.224769592285156, 'timing/data': 0.0013739489950239658, 'timing/model': 2.0151058610063046, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.25}) +[2026-06-23 04:27:01,400][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.39412930607795715, 'train/grad_norm_pre_clip': 93.76544952392578, 'timing/data': 0.0012696659832727164, 'timing/model': 2.0191930060100276, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.25}) +[2026-06-23 04:27:05,446][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.2174103856086731, 'train/grad_norm_pre_clip': 60.66361618041992, 'timing/data': 0.05748103198129684, 'timing/model': 2.016448826994747, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.25}) +[2026-06-23 04:27:09,438][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.15512831509113312, 'train/grad_norm_pre_clip': 30.11049461364746, 'timing/data': 0.0011924130085390061, 'timing/model': 2.0189706000091974, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.25}) +[2026-06-23 04:27:13,524][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.19200864434242249, 'train/grad_norm_pre_clip': 66.9797592163086, 'timing/data': 0.0014628919889219105, 'timing/model': 2.019376303011086, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.25}) +[2026-06-23 04:27:17,520][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.26489654183387756, 'train/grad_norm_pre_clip': 80.3593521118164, 'timing/data': 0.0014118790277279913, 'timing/model': 2.0193039829900954, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.25}) +[2026-06-23 04:27:21,569][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.3697918653488159, 'train/grad_norm_pre_clip': 100.23778533935547, 'timing/data': 0.053251706005539745, 'timing/model': 2.01955601098598, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.25}) +[2026-06-23 04:27:25,572][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.5293471813201904, 'train/grad_norm_pre_clip': 140.26243591308594, 'timing/data': 0.001384769013384357, 'timing/model': 2.022199669998372, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.25}) +[2026-06-23 04:27:29,654][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.298053503036499, 'train/grad_norm_pre_clip': 53.480525970458984, 'timing/data': 0.0014899929810781032, 'timing/model': 2.0193540250184014, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.25}) +[2026-06-23 04:27:33,898][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.22910061478614807, 'train/grad_norm_pre_clip': 36.973243713378906, 'timing/data': 0.001212280010804534, 'timing/model': 2.01582518700161, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.25}) +[2026-06-23 04:27:37,951][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.3194694221019745, 'train/grad_norm_pre_clip': 52.00340270996094, 'timing/data': 0.05251482399762608, 'timing/model': 2.026207913993858, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.25}) +[2026-06-23 04:27:41,953][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.22837947309017181, 'train/grad_norm_pre_clip': 38.11288070678711, 'timing/data': 0.0014260850148275495, 'timing/model': 2.0149365609977394, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.25}) +[2026-06-23 04:27:46,034][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.2272419035434723, 'train/grad_norm_pre_clip': 47.990760803222656, 'timing/data': 0.0012417449906934053, 'timing/model': 2.02117319800891, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.25}) +[2026-06-23 04:27:50,053][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.11130893975496292, 'train/grad_norm_pre_clip': 15.656085014343262, 'timing/data': 0.0012748359877150506, 'timing/model': 2.0326849129924085, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.25}) +[2026-06-23 04:27:54,116][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.26491427421569824, 'train/grad_norm_pre_clip': 31.759292602539062, 'timing/data': 0.07131369799026288, 'timing/model': 2.0170516390062403, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.25}) +[2026-06-23 04:27:58,111][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.2683335542678833, 'train/grad_norm_pre_clip': 54.5499267578125, 'timing/data': 0.0012739940138999373, 'timing/model': 2.0175840570009314, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.25}) +[2026-06-23 04:28:02,191][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.20455536246299744, 'train/grad_norm_pre_clip': 30.98110008239746, 'timing/data': 0.0012200039927847683, 'timing/model': 2.018660829984583, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.25}) +[2026-06-23 04:28:06,240][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.27384278178215027, 'train/grad_norm_pre_clip': 34.285945892333984, 'timing/data': 0.0012340400135144591, 'timing/model': 2.0291333539935295, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.26}) +[2026-06-23 04:28:10,286][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.20317669212818146, 'train/grad_norm_pre_clip': 30.4943904876709, 'timing/data': 0.054266958992229775, 'timing/model': 2.01763641499565, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.26}) +[2026-06-23 04:28:14,282][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.28915128111839294, 'train/grad_norm_pre_clip': 28.480775833129883, 'timing/data': 0.001208293018862605, 'timing/model': 2.0180432200140785, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.26}) +[2026-06-23 04:28:18,372][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.16261033713817596, 'train/grad_norm_pre_clip': 26.560131072998047, 'timing/data': 0.0015125350037124008, 'timing/model': 2.0275114589894656, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.26}) +[2026-06-23 04:28:22,365][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.19809584319591522, 'train/grad_norm_pre_clip': 26.414514541625977, 'timing/data': 0.0012544370256364346, 'timing/model': 2.019788232020801, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.26}) +[2026-06-23 04:28:26,430][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.2717866599559784, 'train/grad_norm_pre_clip': 19.987119674682617, 'timing/data': 0.06573913598549552, 'timing/model': 2.022698824002873, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.26}) +[2026-06-23 04:28:30,459][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.1740339696407318, 'train/grad_norm_pre_clip': 17.31157112121582, 'timing/data': 0.0013658640091307461, 'timing/model': 2.0359933830040973, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.26}) +[2026-06-23 04:28:34,552][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.18118473887443542, 'train/grad_norm_pre_clip': 24.85948371887207, 'timing/data': 0.001199655991513282, 'timing/model': 2.0333712910069153, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.26}) +[2026-06-23 04:28:38,545][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.25418633222579956, 'train/grad_norm_pre_clip': 46.1861457824707, 'timing/data': 0.0017259989981539547, 'timing/model': 2.0199412500078324, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.26}) +[2026-06-23 04:28:42,821][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.11419571936130524, 'train/grad_norm_pre_clip': 21.99526596069336, 'timing/data': 0.0556925919954665, 'timing/model': 2.020140774024185, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.26}) +[2026-06-23 04:28:46,824][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.15607260167598724, 'train/grad_norm_pre_clip': 19.482040405273438, 'timing/data': 0.0012291610182728618, 'timing/model': 2.026554727985058, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.26}) +[2026-06-23 04:28:50,921][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.19412177801132202, 'train/grad_norm_pre_clip': 17.335250854492188, 'timing/data': 0.001244529994437471, 'timing/model': 2.034106776001863, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.26}) +[2026-06-23 04:28:54,924][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.2587171196937561, 'train/grad_norm_pre_clip': 17.42464828491211, 'timing/data': 0.0014254739799071103, 'timing/model': 2.024540665006498, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.26}) +[2026-06-23 04:28:58,990][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.2854987382888794, 'train/grad_norm_pre_clip': 59.80514907836914, 'timing/data': 0.06559882502187975, 'timing/model': 2.0225842850049958, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.26}) +[2026-06-23 04:29:02,998][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.2260373830795288, 'train/grad_norm_pre_clip': 26.295534133911133, 'timing/data': 0.0012870280188508332, 'timing/model': 2.0279672760225367, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.26}) +[2026-06-23 04:29:07,106][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.18834474682807922, 'train/grad_norm_pre_clip': 33.563880920410156, 'timing/data': 0.0016716789978090674, 'timing/model': 2.0345141860016156, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.26}) +[2026-06-23 04:29:11,125][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.26978713274002075, 'train/grad_norm_pre_clip': 45.945919036865234, 'timing/data': 0.0012837919930461794, 'timing/model': 2.0427816860028543, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.26}) +[2026-06-23 04:29:15,208][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.20026588439941406, 'train/grad_norm_pre_clip': 41.79519271850586, 'timing/data': 0.052467502013314515, 'timing/model': 2.016885823017219, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.26}) +[2026-06-23 04:29:19,206][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.23089873790740967, 'train/grad_norm_pre_clip': 37.110130310058594, 'timing/data': 0.0013559449871536344, 'timing/model': 2.020254011003999, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.26}) +[2026-06-23 04:29:23,288][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.19917519390583038, 'train/grad_norm_pre_clip': 40.66017150878906, 'timing/data': 0.0013743900053668767, 'timing/model': 2.018542453995906, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.26}) +[2026-06-23 04:29:27,286][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.11634253710508347, 'train/grad_norm_pre_clip': 17.88998794555664, 'timing/data': 0.001280967000639066, 'timing/model': 2.019712368986802, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.26}) +[2026-06-23 04:29:31,344][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.17075686156749725, 'train/grad_norm_pre_clip': 33.91815948486328, 'timing/data': 0.0649051119980868, 'timing/model': 2.0175401009910274, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.26}) +[2026-06-23 04:29:35,361][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.2827780842781067, 'train/grad_norm_pre_clip': 51.39931106567383, 'timing/data': 0.0012006880133412778, 'timing/model': 2.0387098899809644, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.26}) +[2026-06-23 04:29:39,461][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.2506938874721527, 'train/grad_norm_pre_clip': 52.7212028503418, 'timing/data': 0.0017217809800058603, 'timing/model': 2.035618333000457, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.26}) +[2026-06-23 04:29:43,485][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.17911769449710846, 'train/grad_norm_pre_clip': 23.906673431396484, 'timing/data': 0.001465576991904527, 'timing/model': 2.046313567989273, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.26}) +[2026-06-23 04:29:47,792][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.3357304632663727, 'train/grad_norm_pre_clip': 31.752464294433594, 'timing/data': 0.06545944599201903, 'timing/model': 2.265693116991315, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.26}) +[2026-06-23 04:29:51,805][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.40500879287719727, 'train/grad_norm_pre_clip': 70.61661529541016, 'timing/data': 0.00136993097839877, 'timing/model': 2.036512449005386, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.27}) +[2026-06-23 04:29:55,891][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.28225061297416687, 'train/grad_norm_pre_clip': 46.312374114990234, 'timing/data': 0.0013685289886780083, 'timing/model': 2.0208052509988192, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.27}) +[2026-06-23 04:29:59,902][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.268958181142807, 'train/grad_norm_pre_clip': 66.11067962646484, 'timing/data': 0.0012669300194829702, 'timing/model': 2.0222672180098016, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.27}) +[2026-06-23 04:30:03,957][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.20659716427326202, 'train/grad_norm_pre_clip': 44.53964614868164, 'timing/data': 0.054630176979117095, 'timing/model': 2.0201217780122533, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.27}) +[2026-06-23 04:30:07,958][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.1589697003364563, 'train/grad_norm_pre_clip': 160.47174072265625, 'timing/data': 0.0014440380036830902, 'timing/model': 2.0208577039884403, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.27}) +[2026-06-23 04:30:12,057][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.12805378437042236, 'train/grad_norm_pre_clip': 30.221776962280273, 'timing/data': 0.0013471790007315576, 'timing/model': 2.0371447509969585, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.27}) +[2026-06-23 04:30:16,067][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.22139745950698853, 'train/grad_norm_pre_clip': 44.810333251953125, 'timing/data': 0.0012654580059461296, 'timing/model': 2.028123156982474, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.27}) +[2026-06-23 04:30:20,208][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.17249120771884918, 'train/grad_norm_pre_clip': 26.46547508239746, 'timing/data': 0.05978066200623289, 'timing/model': 2.1031214599788655, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.27}) +[2026-06-23 04:30:24,207][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.21622809767723083, 'train/grad_norm_pre_clip': 30.154691696166992, 'timing/data': 0.0012239409843459725, 'timing/model': 2.0209425310022198, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.27}) +[2026-06-23 04:30:28,293][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.0884011909365654, 'train/grad_norm_pre_clip': 13.505990028381348, 'timing/data': 0.0013421500043477863, 'timing/model': 2.017953563015908, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.27}) +[2026-06-23 04:30:32,281][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.18878766894340515, 'train/grad_norm_pre_clip': 22.952688217163086, 'timing/data': 0.0012692740128841251, 'timing/model': 2.014153708005324, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.27}) +[2026-06-23 04:30:36,335][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.22117210924625397, 'train/grad_norm_pre_clip': 22.82396697998047, 'timing/data': 0.06304898101370782, 'timing/model': 2.017047289002221, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.27}) +[2026-06-23 04:30:40,327][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.14243236184120178, 'train/grad_norm_pre_clip': 14.522491455078125, 'timing/data': 0.0014043150003999472, 'timing/model': 2.015340111014666, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.27}) +[2026-06-23 04:30:44,409][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.091719850897789, 'train/grad_norm_pre_clip': 16.23768424987793, 'timing/data': 0.0012882599839940667, 'timing/model': 2.0207541179843247, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.27}) +[2026-06-23 04:30:48,402][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.19293320178985596, 'train/grad_norm_pre_clip': 41.49261474609375, 'timing/data': 0.0012447090120986104, 'timing/model': 2.018187912006397, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.27}) +[2026-06-23 04:30:52,453][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 700 +[2026-06-23 04:32:48,990][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 700 +[2026-06-23 04:32:48,991][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 700 +[2026-06-23 04:34:55,624][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 700 +[2026-06-23 04:34:55,626][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.1253616362810135, 'train/grad_norm_pre_clip': 19.47079086303711, 'eval/action_loss/seconds': 116.53366997398552, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.1034395694732666, 'eval/latency_2/loss': 0.1034395694732666, 'eval/flappy/loss': 0.1034395694732666, 'eval/flappy/latency_2/loss': 0.1034395694732666, 'eval/flap_precision': 0.8474178403755869, 'eval/flap_recall': 0.9743589743589743, 'eval/flap_f1': 0.90646578782172, 'eval/flap_rate_model': 0.23666666666666666, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9064657878217199, 'eval/latency_2/flap_precision': 0.8474178403755869, 'eval/latency_2/flap_recall': 0.9743589743589743, 'eval/latency_2/flap_f1': 0.90646578782172, 'eval/latency_2/flap_rate_model': 0.23666666666666666, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9064657878217199, 'eval/action_classification/data_wait_seconds': 50.407562255859375, 'eval/action_classification/predict_seconds': 74.5501480102539, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.05423436299315654, 'timing/model': 2.0204019180091564, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.27}) +[2026-06-23 04:34:59,580][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.11332672089338303, 'train/grad_norm_pre_clip': 16.091901779174805, 'timing/data': 0.0012005280004814267, 'timing/model': 1.9942338420078158, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.27}) +[2026-06-23 04:35:03,603][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.10357707738876343, 'train/grad_norm_pre_clip': 21.04374122619629, 'timing/data': 0.0012413229851517826, 'timing/model': 1.990244699001778, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.27}) +[2026-06-23 04:35:07,578][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.1889190822839737, 'train/grad_norm_pre_clip': 35.40047073364258, 'timing/data': 0.0014750950213056058, 'timing/model': 2.0012314839987084, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.27}) +[2026-06-23 04:35:11,598][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.20024839043617249, 'train/grad_norm_pre_clip': 35.78229904174805, 'timing/data': 0.07959548200597055, 'timing/model': 1.993477896990953, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.27}) +[2026-06-23 04:35:15,542][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.1673499196767807, 'train/grad_norm_pre_clip': 34.28978729248047, 'timing/data': 0.0012408620095811784, 'timing/model': 1.9936600500077475, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.27}) +[2026-06-23 04:35:19,589][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.15660995244979858, 'train/grad_norm_pre_clip': 39.33573913574219, 'timing/data': 0.0012212270230520517, 'timing/model': 1.9931823649967555, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.27}) +[2026-06-23 04:35:23,589][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.0711626261472702, 'train/grad_norm_pre_clip': 14.178841590881348, 'timing/data': 0.0014246119826566428, 'timing/model': 2.013749400008237, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.27}) +[2026-06-23 04:35:27,611][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.10483911633491516, 'train/grad_norm_pre_clip': 29.81882667541504, 'timing/data': 0.06582632299978286, 'timing/model': 1.9983542450063396, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.27}) +[2026-06-23 04:35:31,571][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.19426168501377106, 'train/grad_norm_pre_clip': 23.797197341918945, 'timing/data': 0.0013388929946813732, 'timing/model': 2.0003945210191887, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.27}) +[2026-06-23 04:35:35,726][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.1303642988204956, 'train/grad_norm_pre_clip': 28.782155990600586, 'timing/data': 0.0014205850020516664, 'timing/model': 2.0265603699954227, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.27}) +[2026-06-23 04:35:39,766][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.23461118340492249, 'train/grad_norm_pre_clip': 39.95747375488281, 'timing/data': 0.0014188110071700066, 'timing/model': 2.015469937992748, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.28}) +[2026-06-23 04:35:43,821][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.2717359960079193, 'train/grad_norm_pre_clip': 43.88968276977539, 'timing/data': 0.07740291999652982, 'timing/model': 2.017196745000547, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.28}) +[2026-06-23 04:35:47,790][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.22372829914093018, 'train/grad_norm_pre_clip': 74.0218276977539, 'timing/data': 0.0012265760160516948, 'timing/model': 2.0056514340103604, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.28}) +[2026-06-23 04:35:51,853][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.060240261256694794, 'train/grad_norm_pre_clip': 17.28771209716797, 'timing/data': 0.0012843430158682168, 'timing/model': 2.006856467021862, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.28}) +[2026-06-23 04:35:55,939][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.10042744874954224, 'train/grad_norm_pre_clip': 18.944442749023438, 'timing/data': 0.0012920269800815731, 'timing/model': 2.1209401509840973, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.28}) +[2026-06-23 04:35:59,969][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.1651945263147354, 'train/grad_norm_pre_clip': 23.568613052368164, 'timing/data': 0.05433829899993725, 'timing/model': 2.005880848999368, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.28}) +[2026-06-23 04:36:03,942][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.1729832887649536, 'train/grad_norm_pre_clip': 30.548641204833984, 'timing/data': 0.001251853012945503, 'timing/model': 2.007350622996455, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.28}) +[2026-06-23 04:36:08,022][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.15121394395828247, 'train/grad_norm_pre_clip': 31.809968948364258, 'timing/data': 0.0013045899977441877, 'timing/model': 2.0231193419895135, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.28}) +[2026-06-23 04:36:12,014][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.16588304936885834, 'train/grad_norm_pre_clip': 35.09505844116211, 'timing/data': 0.0012628429976757616, 'timing/model': 2.025295039988123, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.28}) +[2026-06-23 04:36:16,045][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.1625858098268509, 'train/grad_norm_pre_clip': 25.98828887939453, 'timing/data': 0.05318272399017587, 'timing/model': 2.010988026013365, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.28}) +[2026-06-23 04:36:20,025][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.24970643222332, 'train/grad_norm_pre_clip': 34.322086334228516, 'timing/data': 0.0012576529989019036, 'timing/model': 2.012210419983603, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.28}) +[2026-06-23 04:36:24,111][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.16346313059329987, 'train/grad_norm_pre_clip': 30.84905242919922, 'timing/data': 0.0013977120106574148, 'timing/model': 2.0267266670125537, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.28}) +[2026-06-23 04:36:28,228][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.19008956849575043, 'train/grad_norm_pre_clip': 24.830615997314453, 'timing/data': 0.0012412130017764866, 'timing/model': 2.149772636999842, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.28}) +[2026-06-23 04:36:32,268][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.19386877119541168, 'train/grad_norm_pre_clip': 58.704776763916016, 'timing/data': 0.06663443401339464, 'timing/model': 2.0099500579817686, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.28}) +[2026-06-23 04:36:36,247][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.12916287779808044, 'train/grad_norm_pre_clip': 16.943307876586914, 'timing/data': 0.0011356679897289723, 'timing/model': 2.011814981000498, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.28}) +[2026-06-23 04:36:40,323][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.19000649452209473, 'train/grad_norm_pre_clip': 25.712305068969727, 'timing/data': 0.0011419799993745983, 'timing/model': 2.0227916080038995, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.28}) +[2026-06-23 04:36:44,305][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.15880443155765533, 'train/grad_norm_pre_clip': 27.632892608642578, 'timing/data': 0.0012883099843747914, 'timing/model': 2.0143441089894623, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.28}) +[2026-06-23 04:36:48,354][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.22104686498641968, 'train/grad_norm_pre_clip': 36.216190338134766, 'timing/data': 0.06105367501731962, 'timing/model': 2.016870202991413, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.28}) +[2026-06-23 04:36:52,342][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.13353252410888672, 'train/grad_norm_pre_clip': 27.43303108215332, 'timing/data': 0.001267322018975392, 'timing/model': 2.0164454269979615, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.28}) +[2026-06-23 04:36:56,435][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.2467869520187378, 'train/grad_norm_pre_clip': 41.14390182495117, 'timing/data': 0.001267100975383073, 'timing/model': 2.031387739989441, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.28}) +[2026-06-23 04:37:00,424][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.19268247485160828, 'train/grad_norm_pre_clip': 36.13346481323242, 'timing/data': 0.0013863309868611395, 'timing/model': 2.0160070619895123, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.28}) +[2026-06-23 04:37:04,704][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.19455519318580627, 'train/grad_norm_pre_clip': 24.75006866455078, 'timing/data': 0.05420798697741702, 'timing/model': 2.0132408190111164, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.28}) +[2026-06-23 04:37:08,700][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.18416202068328857, 'train/grad_norm_pre_clip': 48.72626495361328, 'timing/data': 0.0012214069720357656, 'timing/model': 2.019274384016171, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.28}) +[2026-06-23 04:37:12,780][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.23277989029884338, 'train/grad_norm_pre_clip': 35.486209869384766, 'timing/data': 0.001338993984973058, 'timing/model': 2.0173695019911975, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.28}) +[2026-06-23 04:37:16,771][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.20821760594844818, 'train/grad_norm_pre_clip': 26.2548828125, 'timing/data': 0.0013184059935156256, 'timing/model': 2.0158726240042597, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.28}) +[2026-06-23 04:37:20,815][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.205128014087677, 'train/grad_norm_pre_clip': 24.381628036499023, 'timing/data': 0.05742473900318146, 'timing/model': 2.0116218100010883, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.28}) +[2026-06-23 04:37:24,819][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.12087999284267426, 'train/grad_norm_pre_clip': 16.393156051635742, 'timing/data': 0.0011463479895610362, 'timing/model': 2.0263577409787104, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.29}) +[2026-06-23 04:37:28,894][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.31244733929634094, 'train/grad_norm_pre_clip': 64.29843139648438, 'timing/data': 0.0012338199885562062, 'timing/model': 2.013628212007461, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.29}) +[2026-06-23 04:37:32,878][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.16537629067897797, 'train/grad_norm_pre_clip': 41.7490234375, 'timing/data': 0.0011677169823087752, 'timing/model': 2.0132413649989758, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.29}) +[2026-06-23 04:37:37,003][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.3804982006549835, 'train/grad_norm_pre_clip': 52.99892044067383, 'timing/data': 0.05235373100731522, 'timing/model': 2.1043350340041798, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.29}) +[2026-06-23 04:37:40,981][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.25417810678482056, 'train/grad_norm_pre_clip': 37.03369140625, 'timing/data': 0.0011133170046377927, 'timing/model': 2.0118052780162543, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.29}) +[2026-06-23 04:37:45,072][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.16117613017559052, 'train/grad_norm_pre_clip': 23.62313461303711, 'timing/data': 0.0012161069898866117, 'timing/model': 2.0183589329826646, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.29}) +[2026-06-23 04:37:49,083][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.1968664973974228, 'train/grad_norm_pre_clip': 25.1580753326416, 'timing/data': 0.0014580030110664666, 'timing/model': 2.038063204003265, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.29}) +[2026-06-23 04:37:53,119][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.3372829258441925, 'train/grad_norm_pre_clip': 50.31193161010742, 'timing/data': 0.053630218986654654, 'timing/model': 2.0146520209964365, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.29}) +[2026-06-23 04:37:57,123][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.36201635003089905, 'train/grad_norm_pre_clip': 54.84695053100586, 'timing/data': 0.0014279879978857934, 'timing/model': 2.0298559970106, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.29}) +[2026-06-23 04:38:01,234][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.4588484466075897, 'train/grad_norm_pre_clip': 87.55426788330078, 'timing/data': 0.0013615349889732897, 'timing/model': 2.0482845210062806, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.29}) +[2026-06-23 04:38:05,221][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.20082513988018036, 'train/grad_norm_pre_clip': 31.683168411254883, 'timing/data': 0.0012236209877301008, 'timing/model': 2.0145576409995556, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.29}) +[2026-06-23 04:38:09,271][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.17865371704101562, 'train/grad_norm_pre_clip': 23.850950241088867, 'timing/data': 0.05528861199854873, 'timing/model': 2.0198442959808744, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.29}) +[2026-06-23 04:38:13,307][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.23116371035575867, 'train/grad_norm_pre_clip': 41.19451904296875, 'timing/data': 0.001291195017984137, 'timing/model': 2.0240258590201847, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.29}) +[2026-06-23 04:43:46,348][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 750 +[2026-06-23 04:43:46,350][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.18949222564697266, 'train/grad_norm_pre_clip': 35.91844940185547, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 83.5799999999984, 'rl_games_eval/mid_train/mean_length': 692.0, 'rl_games_eval/mid_train/std_reward': 71.19664037017279, 'rl_games_eval/mid_train/std_length': 575.6933211354809, 'rl_games_eval/mid_train/macro_mean_reward': 83.5799999999984, 'rl_games_eval/mid_train/macro_mean_length': 692.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 83.5799999999984, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 692.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 71.19664037017279, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 575.6933211354809, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 83.5799999999984, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0014866959827486426, 'timing/model': 2.023297692008782, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.29}) +[2026-06-23 04:43:50,235][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.1818854808807373, 'train/grad_norm_pre_clip': 23.55222511291504, 'timing/data': 0.00125013999058865, 'timing/model': 1.961700210988056, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.29}) +[2026-06-23 04:43:54,181][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.18212032318115234, 'train/grad_norm_pre_clip': 16.148035049438477, 'timing/data': 0.07242790100281127, 'timing/model': 1.9582625899929553, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.29}) +[2026-06-23 04:43:58,091][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.14066165685653687, 'train/grad_norm_pre_clip': 15.031220436096191, 'timing/data': 0.0012885710166301578, 'timing/model': 1.9856726260040887, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.29}) +[2026-06-23 04:44:02,300][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.23498672246932983, 'train/grad_norm_pre_clip': 48.717445373535156, 'timing/data': 0.001192842988530174, 'timing/model': 2.203695345990127, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.29}) +[2026-06-23 04:44:06,179][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.20945414900779724, 'train/grad_norm_pre_clip': 27.236209869384766, 'timing/data': 0.0012338790111243725, 'timing/model': 1.9609828820102848, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.29}) +[2026-06-23 04:44:10,140][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.34811410307884216, 'train/grad_norm_pre_clip': 41.73625946044922, 'timing/data': 0.05674443798488937, 'timing/model': 1.9759093669999857, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.29}) +[2026-06-23 04:44:14,073][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.20700550079345703, 'train/grad_norm_pre_clip': 25.569114685058594, 'timing/data': 0.0013404259807430208, 'timing/model': 1.9919412529852707, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.29}) +[2026-06-23 04:44:18,056][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.2235012948513031, 'train/grad_norm_pre_clip': 35.98380661010742, 'timing/data': 0.001130337972426787, 'timing/model': 1.9693568200164009, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.29}) +[2026-06-23 04:44:21,960][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.10157088190317154, 'train/grad_norm_pre_clip': 17.297142028808594, 'timing/data': 0.0011647010105662048, 'timing/model': 1.9745798469812144, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.29}) +[2026-06-23 04:44:25,945][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.2293555736541748, 'train/grad_norm_pre_clip': 34.67169952392578, 'timing/data': 0.06498557300074026, 'timing/model': 1.9897740059823263, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.29}) +[2026-06-23 04:44:29,872][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.10616269707679749, 'train/grad_norm_pre_clip': 19.509897232055664, 'timing/data': 0.0012028820056002587, 'timing/model': 1.9916039459931199, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.29}) +[2026-06-23 04:44:33,915][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.1848462074995041, 'train/grad_norm_pre_clip': 45.68692398071289, 'timing/data': 0.0011325619998387992, 'timing/model': 2.021559543994954, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.29}) +[2026-06-23 04:44:37,828][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.2750774621963501, 'train/grad_norm_pre_clip': 44.34790802001953, 'timing/data': 0.0011709340033121407, 'timing/model': 1.9805476789770182, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.3}) +[2026-06-23 04:44:41,816][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.1688096672296524, 'train/grad_norm_pre_clip': 32.96115493774414, 'timing/data': 0.05987321797874756, 'timing/model': 1.9847918990126345, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.3}) +[2026-06-23 04:44:45,749][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.1867174506187439, 'train/grad_norm_pre_clip': 24.01498031616211, 'timing/data': 0.0012280289956834167, 'timing/model': 1.9904499000112992, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.3}) +[2026-06-23 04:44:49,769][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.08701705932617188, 'train/grad_norm_pre_clip': 11.492177963256836, 'timing/data': 0.0013786170165985823, 'timing/model': 1.9860260870191269, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.3}) +[2026-06-23 04:44:53,702][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.15060244500637054, 'train/grad_norm_pre_clip': 27.175390243530273, 'timing/data': 0.0011940859840251505, 'timing/model': 1.9882581499987282, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.3}) +[2026-06-23 04:44:57,702][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.1286081075668335, 'train/grad_norm_pre_clip': 25.444026947021484, 'timing/data': 0.05618103698361665, 'timing/model': 1.9945368579938076, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.3}) +[2026-06-23 04:45:01,639][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.1319190114736557, 'train/grad_norm_pre_clip': 17.197521209716797, 'timing/data': 0.0011476899962872267, 'timing/model': 1.989491877000546, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.3}) +[2026-06-23 04:45:05,682][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.19553330540657043, 'train/grad_norm_pre_clip': 30.406713485717773, 'timing/data': 0.0012071389937773347, 'timing/model': 1.994600793987047, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.3}) +[2026-06-23 04:45:09,795][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.21967250108718872, 'train/grad_norm_pre_clip': 38.60871887207031, 'timing/data': 0.0011583399900700897, 'timing/model': 1.9950291849963833, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.3}) +[2026-06-23 04:45:13,793][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.21415366232395172, 'train/grad_norm_pre_clip': 23.112892150878906, 'timing/data': 0.05045240398612805, 'timing/model': 1.994277494988637, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.3}) +[2026-06-23 04:45:17,750][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.25266140699386597, 'train/grad_norm_pre_clip': 52.051918029785156, 'timing/data': 0.0011338639887981117, 'timing/model': 2.009167541982606, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.3}) +[2026-06-23 04:45:21,780][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.09311395138502121, 'train/grad_norm_pre_clip': 19.017349243164062, 'timing/data': 0.0012448400084394962, 'timing/model': 1.9926163339987397, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.3}) +[2026-06-23 04:45:25,749][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.19295001029968262, 'train/grad_norm_pre_clip': 31.46733283996582, 'timing/data': 0.0012705480039585382, 'timing/model': 2.013402282987954, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.3}) +[2026-06-23 04:45:29,766][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.13065002858638763, 'train/grad_norm_pre_clip': 25.50033187866211, 'timing/data': 0.06396498600952327, 'timing/model': 1.998900471982779, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.3}) +[2026-06-23 04:45:33,732][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.20247499644756317, 'train/grad_norm_pre_clip': 38.24860382080078, 'timing/data': 0.001199274993268773, 'timing/model': 2.006515689980006, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.3}) +[2026-06-23 04:45:37,785][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.15515659749507904, 'train/grad_norm_pre_clip': 19.544906616210938, 'timing/data': 0.0013826339854858816, 'timing/model': 2.0011797719926108, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.3}) +[2026-06-23 04:45:42,012][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.2030211240053177, 'train/grad_norm_pre_clip': 28.585838317871094, 'timing/data': 0.001221696991706267, 'timing/model': 2.2687650590087287, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.3}) +[2026-06-23 04:45:46,024][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.28443869948387146, 'train/grad_norm_pre_clip': 42.246097564697266, 'timing/data': 0.05115620500873774, 'timing/model': 1.9994222829991486, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.3}) +[2026-06-23 04:45:49,995][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.15409217774868011, 'train/grad_norm_pre_clip': 32.85436248779297, 'timing/data': 0.0011877939978148788, 'timing/model': 2.004412851994857, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.3}) +[2026-06-23 04:45:54,053][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.1607055962085724, 'train/grad_norm_pre_clip': 24.00379180908203, 'timing/data': 0.001448135997634381, 'timing/model': 2.0032023749954533, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.3}) +[2026-06-23 04:45:58,036][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.06624634563922882, 'train/grad_norm_pre_clip': 8.05830192565918, 'timing/data': 0.0013549840077757835, 'timing/model': 2.0212387340143323, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.3}) +[2026-06-23 04:46:02,057][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.1586828976869583, 'train/grad_norm_pre_clip': 27.7047176361084, 'timing/data': 0.051731857005506754, 'timing/model': 2.0066073430061806, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.3}) +[2026-06-23 04:46:06,048][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.11849711835384369, 'train/grad_norm_pre_clip': 15.252753257751465, 'timing/data': 0.0012303629773668945, 'timing/model': 2.026050482003484, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.3}) +[2026-06-23 04:46:10,108][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.1814929097890854, 'train/grad_norm_pre_clip': 19.334196090698242, 'timing/data': 0.0012286999844945967, 'timing/model': 2.007590498018544, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.3}) +[2026-06-23 04:46:14,164][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.1381375640630722, 'train/grad_norm_pre_clip': 24.366777420043945, 'timing/data': 0.0013764930190518498, 'timing/model': 2.0845157860021573, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.3}) +[2026-06-23 04:46:18,215][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.15148195624351501, 'train/grad_norm_pre_clip': 15.515427589416504, 'timing/data': 0.06675336800981313, 'timing/model': 2.013768959994195, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.3}) +[2026-06-23 04:46:22,212][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.13784299790859222, 'train/grad_norm_pre_clip': 16.76418113708496, 'timing/data': 0.0011620769801083952, 'timing/model': 2.0277989479945973, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.31}) +[2026-06-23 04:46:26,277][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.28973647952079773, 'train/grad_norm_pre_clip': 40.66120529174805, 'timing/data': 0.0013244769943412393, 'timing/model': 2.0055014460231178, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.31}) +[2026-06-23 04:46:30,271][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.12436296045780182, 'train/grad_norm_pre_clip': 20.997634887695312, 'timing/data': 0.0016230279870796949, 'timing/model': 2.0110809720063116, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.31}) +[2026-06-23 04:46:34,301][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.11192415654659271, 'train/grad_norm_pre_clip': 14.09564208984375, 'timing/data': 0.052269601001171395, 'timing/model': 2.01028654401307, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.31}) +[2026-06-23 04:46:38,288][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.1303565353155136, 'train/grad_norm_pre_clip': 22.21790885925293, 'timing/data': 0.0013749199861194938, 'timing/model': 2.0149400000227615, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.31}) +[2026-06-23 04:46:42,350][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.18850883841514587, 'train/grad_norm_pre_clip': 28.245895385742188, 'timing/data': 0.0013807310024276376, 'timing/model': 2.0082462139835116, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.31}) +[2026-06-23 04:46:46,331][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.35782408714294434, 'train/grad_norm_pre_clip': 70.4474868774414, 'timing/data': 0.001502606988651678, 'timing/model': 2.012286507000681, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.31}) +[2026-06-23 04:46:50,473][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.2552962005138397, 'train/grad_norm_pre_clip': 49.814517974853516, 'timing/data': 0.06023450498469174, 'timing/model': 2.0106168720230926, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.31}) +[2026-06-23 04:46:54,448][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.0922137200832367, 'train/grad_norm_pre_clip': 23.123315811157227, 'timing/data': 0.0012486070045270026, 'timing/model': 2.0081366689992137, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.31}) +[2026-06-23 04:46:58,521][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.1259160041809082, 'train/grad_norm_pre_clip': 24.04800033569336, 'timing/data': 0.0011088580067735165, 'timing/model': 2.0264248640160076, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.31}) +[2026-06-23 04:47:02,502][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.19196312129497528, 'train/grad_norm_pre_clip': 27.76033592224121, 'timing/data': 0.0013413289852906018, 'timing/model': 2.012281264003832, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.31}) +[2026-06-23 04:47:06,557][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 800 +[2026-06-23 04:49:02,535][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 800 +[2026-06-23 04:49:02,535][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 800 +[2026-06-23 04:51:09,758][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 800 +[2026-06-23 04:51:09,760][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.16816556453704834, 'train/grad_norm_pre_clip': 32.1400146484375, 'eval/action_loss/seconds': 115.97367460097303, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.09424236416816711, 'eval/latency_2/loss': 0.09424236416816711, 'eval/flappy/loss': 0.09424236416816711, 'eval/flappy/latency_2/loss': 0.09424236416816711, 'eval/flap_precision': 0.8634441087613293, 'eval/flap_recall': 0.9642375168690959, 'eval/flap_f1': 0.9110615237488046, 'eval/flap_rate_model': 0.2298611111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9110615237488046, 'eval/latency_2/flap_precision': 0.8634441087613293, 'eval/latency_2/flap_recall': 0.9642375168690959, 'eval/latency_2/flap_f1': 0.9110615237488046, 'eval/latency_2/flap_rate_model': 0.2298611111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9110615237488046, 'eval/action_classification/data_wait_seconds': 50.88166809082031, 'eval/action_classification/predict_seconds': 74.6269302368164, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06381989800138399, 'timing/model': 2.0155795569880866, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.31}) +[2026-06-23 04:51:13,879][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.2269689440727234, 'train/grad_norm_pre_clip': 51.58674240112305, 'timing/data': 0.0012427880137693137, 'timing/model': 2.1510729999863543, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.31}) +[2026-06-23 04:51:17,902][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.11791247874498367, 'train/grad_norm_pre_clip': 25.4392147064209, 'timing/data': 0.001210677990457043, 'timing/model': 1.9902819929993711, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.31}) +[2026-06-23 04:51:21,843][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.1250840723514557, 'train/grad_norm_pre_clip': 15.481490135192871, 'timing/data': 0.0012183920189272612, 'timing/model': 1.9924011629773304, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.31}) +[2026-06-23 04:51:25,869][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.31890201568603516, 'train/grad_norm_pre_clip': 46.26319122314453, 'timing/data': 0.05425419600214809, 'timing/model': 2.005650269013131, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.31}) +[2026-06-23 04:51:29,833][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.22982095181941986, 'train/grad_norm_pre_clip': 42.97177505493164, 'timing/data': 0.001210257993079722, 'timing/model': 2.0134897840034682, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.31}) +[2026-06-23 04:51:33,879][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.31429269909858704, 'train/grad_norm_pre_clip': 46.87403106689453, 'timing/data': 0.0014425870031118393, 'timing/model': 1.993694256991148, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.31}) +[2026-06-23 04:51:37,834][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.08443902432918549, 'train/grad_norm_pre_clip': 15.281578063964844, 'timing/data': 0.001257233991054818, 'timing/model': 1.996612635004567, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.31}) +[2026-06-23 04:51:41,849][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.10788624733686447, 'train/grad_norm_pre_clip': 35.06352996826172, 'timing/data': 0.05809981399215758, 'timing/model': 1.9987408009765204, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.31}) +[2026-06-23 04:51:45,803][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.2180778831243515, 'train/grad_norm_pre_clip': 43.38401794433594, 'timing/data': 0.0016549189749639481, 'timing/model': 1.9956974959932268, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.31}) +[2026-06-23 04:51:50,096][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.4158558249473572, 'train/grad_norm_pre_clip': 100.29586029052734, 'timing/data': 0.0012121410109102726, 'timing/model': 1.9994954180147033, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.31}) +[2026-06-23 04:51:54,056][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.2148468792438507, 'train/grad_norm_pre_clip': 36.03173828125, 'timing/data': 0.0014774010051041842, 'timing/model': 1.9992697940033395, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.31}) +[2026-06-23 04:51:58,073][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.1588902473449707, 'train/grad_norm_pre_clip': 25.24141502380371, 'timing/data': 0.05553774500731379, 'timing/model': 2.0042515349923633, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.31}) +[2026-06-23 04:52:02,037][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.23117539286613464, 'train/grad_norm_pre_clip': 26.24899673461914, 'timing/data': 0.0012684149842243642, 'timing/model': 2.003497354016872, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.31}) +[2026-06-23 04:52:06,092][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.24486973881721497, 'train/grad_norm_pre_clip': 38.72397232055664, 'timing/data': 0.0012219990021549165, 'timing/model': 2.004816934990231, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.32}) +[2026-06-23 04:52:10,067][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.1925683468580246, 'train/grad_norm_pre_clip': 39.91530990600586, 'timing/data': 0.0015281250234693289, 'timing/model': 2.008454332011752, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.32}) +[2026-06-23 04:52:14,099][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.17463712394237518, 'train/grad_norm_pre_clip': 34.13105773925781, 'timing/data': 0.06660264800302684, 'timing/model': 2.004563713009702, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.32}) +[2026-06-23 04:52:18,062][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.07828257232904434, 'train/grad_norm_pre_clip': 15.17220401763916, 'timing/data': 0.0012917879794258624, 'timing/model': 2.0015319300000556, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.32}) +[2026-06-23 04:52:22,247][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.14817793667316437, 'train/grad_norm_pre_clip': 32.13615417480469, 'timing/data': 0.0012639070046134293, 'timing/model': 2.0078771430125926, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.32}) +[2026-06-23 04:52:26,244][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.3214099705219269, 'train/grad_norm_pre_clip': 58.680049896240234, 'timing/data': 0.0014463539991993457, 'timing/model': 2.0310686000157148, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.32}) +[2026-06-23 04:52:30,293][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.22562234103679657, 'train/grad_norm_pre_clip': 32.13544845581055, 'timing/data': 0.06539029499981552, 'timing/model': 2.018793770985212, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.32}) +[2026-06-23 04:52:34,283][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.20644541084766388, 'train/grad_norm_pre_clip': 31.724597930908203, 'timing/data': 0.0013997369969729334, 'timing/model': 2.021485840989044, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.32}) +[2026-06-23 04:52:38,352][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.10274868458509445, 'train/grad_norm_pre_clip': 21.622289657592773, 'timing/data': 0.0012309449957683682, 'timing/model': 2.004388506989926, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.32}) +[2026-06-23 04:52:42,335][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.12773288786411285, 'train/grad_norm_pre_clip': 22.036712646484375, 'timing/data': 0.0015048719942569733, 'timing/model': 2.01276088700979, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.32}) +[2026-06-23 04:52:46,384][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.21045592427253723, 'train/grad_norm_pre_clip': 36.12511444091797, 'timing/data': 0.06933796100202017, 'timing/model': 2.012161352991825, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.32}) +[2026-06-23 04:52:50,404][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.22050955891609192, 'train/grad_norm_pre_clip': 31.779312133789062, 'timing/data': 0.001322405005339533, 'timing/model': 2.040198878006777, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.32}) +[2026-06-23 04:52:54,746][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.20337484776973724, 'train/grad_norm_pre_clip': 20.409242630004883, 'timing/data': 0.0012568639940582216, 'timing/model': 2.281449199974304, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.32}) +[2026-06-23 04:52:58,739][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.06939895451068878, 'train/grad_norm_pre_clip': 8.627326011657715, 'timing/data': 0.0012580149923451245, 'timing/model': 2.0257145979849156, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.32}) +[2026-06-23 04:53:02,794][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.10812462121248245, 'train/grad_norm_pre_clip': 10.890446662902832, 'timing/data': 0.0669536909845192, 'timing/model': 2.0148244700103533, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.32}) +[2026-06-23 04:53:06,768][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.12572726607322693, 'train/grad_norm_pre_clip': 24.710662841796875, 'timing/data': 0.0012461139995139092, 'timing/model': 2.004580968990922, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.32}) +[2026-06-23 04:53:10,825][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.29044628143310547, 'train/grad_norm_pre_clip': 27.072816848754883, 'timing/data': 0.0011362400255165994, 'timing/model': 2.006800736009609, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.32}) +[2026-06-23 04:53:14,811][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.17561262845993042, 'train/grad_norm_pre_clip': 38.09329605102539, 'timing/data': 0.0011615580006036907, 'timing/model': 2.0245395329839084, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.32}) +[2026-06-23 04:53:18,852][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.1756225973367691, 'train/grad_norm_pre_clip': 18.020078659057617, 'timing/data': 0.05954601598205045, 'timing/model': 2.0128975760017056, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.32}) +[2026-06-23 04:53:22,835][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.2461404651403427, 'train/grad_norm_pre_clip': 32.69572830200195, 'timing/data': 0.001119459979236126, 'timing/model': 2.0159276720078196, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.32}) +[2026-06-23 04:53:26,903][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.1694158911705017, 'train/grad_norm_pre_clip': 24.64744758605957, 'timing/data': 0.0011281860060989857, 'timing/model': 2.007329343003221, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.32}) +[2026-06-23 04:53:30,965][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.12991680204868317, 'train/grad_norm_pre_clip': 12.753327369689941, 'timing/data': 0.0012761090183630586, 'timing/model': 2.0943221490015276, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.32}) +[2026-06-23 04:53:35,011][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.12960410118103027, 'train/grad_norm_pre_clip': 17.367401123046875, 'timing/data': 0.06677391400444321, 'timing/model': 2.012167095002951, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.32}) +[2026-06-23 04:53:38,994][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.14964668452739716, 'train/grad_norm_pre_clip': 20.86701011657715, 'timing/data': 0.001144545996794477, 'timing/model': 2.011782900983235, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.32}) +[2026-06-23 04:53:43,079][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.14648506045341492, 'train/grad_norm_pre_clip': 27.740243911743164, 'timing/data': 0.0011064050195273012, 'timing/model': 2.0277236929978244, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.32}) +[2026-06-23 04:53:47,076][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.1346515268087387, 'train/grad_norm_pre_clip': 22.93976402282715, 'timing/data': 0.0011605249892454594, 'timing/model': 2.0244435850181617, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.32}) +[2026-06-23 04:53:51,107][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.1442328840494156, 'train/grad_norm_pre_clip': 22.398767471313477, 'timing/data': 0.05580523700336926, 'timing/model': 2.006915278994711, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.33}) +[2026-06-23 04:53:55,081][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.16632016003131866, 'train/grad_norm_pre_clip': 24.47942543029785, 'timing/data': 0.0011169550125487149, 'timing/model': 2.0087773769919295, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.33}) +[2026-06-23 04:53:59,146][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.16439031064510345, 'train/grad_norm_pre_clip': 30.364410400390625, 'timing/data': 0.0013248590112198144, 'timing/model': 2.0090342670155223, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.33}) +[2026-06-23 04:54:03,124][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.09116439521312714, 'train/grad_norm_pre_clip': 14.796090126037598, 'timing/data': 0.0011543540167622268, 'timing/model': 2.0087288189970423, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.33}) +[2026-06-23 04:54:07,371][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.13713203370571136, 'train/grad_norm_pre_clip': 25.98412322998047, 'timing/data': 0.05628022400196642, 'timing/model': 2.0087136960064527, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.33}) +[2026-06-23 04:54:11,347][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.18634694814682007, 'train/grad_norm_pre_clip': 18.242006301879883, 'timing/data': 0.0011148909979965538, 'timing/model': 2.0088384529808536, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.33}) +[2026-06-23 04:54:15,413][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.127462238073349, 'train/grad_norm_pre_clip': 15.58064079284668, 'timing/data': 0.0011397360067348927, 'timing/model': 2.012698234000709, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.33}) +[2026-06-23 04:54:19,415][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.14358492195606232, 'train/grad_norm_pre_clip': 18.095197677612305, 'timing/data': 0.0013080890057608485, 'timing/model': 2.0252577560022473, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.33}) +[2026-06-23 04:54:23,458][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.12647821009159088, 'train/grad_norm_pre_clip': 15.867948532104492, 'timing/data': 0.055629769020015374, 'timing/model': 2.0150758319941815, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.33}) +[2026-06-23 04:54:27,447][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.08712335675954819, 'train/grad_norm_pre_clip': 10.919939994812012, 'timing/data': 0.0011824159882962704, 'timing/model': 2.0178066669905093, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.33}) +[2026-06-23 04:54:31,529][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.08403270691633224, 'train/grad_norm_pre_clip': 37.90724563598633, 'timing/data': 0.001141579996328801, 'timing/model': 2.0197529300057795, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.33}) +[2026-06-23 04:54:35,518][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.11525322496891022, 'train/grad_norm_pre_clip': 18.389781951904297, 'timing/data': 0.0011571490031201392, 'timing/model': 2.0165286470146384, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.33}) +[2026-06-23 04:54:39,604][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.09495554864406586, 'train/grad_norm_pre_clip': 18.70188331604004, 'timing/data': 0.055143639008747414, 'timing/model': 2.0167130050249398, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.33}) +[2026-06-23 04:54:43,597][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.07444009184837341, 'train/grad_norm_pre_clip': 19.513656616210938, 'timing/data': 0.001109640987124294, 'timing/model': 2.019309130992042, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.33}) +[2026-06-23 04:54:47,679][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.21840938925743103, 'train/grad_norm_pre_clip': 48.56961441040039, 'timing/data': 0.0011072960041929036, 'timing/model': 2.016310033010086, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.33}) +[2026-06-23 04:54:51,677][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.16086456179618835, 'train/grad_norm_pre_clip': 26.061309814453125, 'timing/data': 0.001150025986135006, 'timing/model': 2.019534641003702, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.33}) +[2026-06-23 04:54:55,730][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.23970499634742737, 'train/grad_norm_pre_clip': 37.08504104614258, 'timing/data': 0.0605003590171691, 'timing/model': 2.019299214007333, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.33}) +[2026-06-23 04:54:59,719][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.2078164964914322, 'train/grad_norm_pre_clip': 24.809078216552734, 'timing/data': 0.0011626489867921919, 'timing/model': 2.0167538459936623, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.33}) +[2026-06-23 04:55:03,801][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.05903884395956993, 'train/grad_norm_pre_clip': 28.74901008605957, 'timing/data': 0.0012610109988600016, 'timing/model': 2.0156297750072554, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.33}) +[2026-06-23 04:55:07,808][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.14867372810840607, 'train/grad_norm_pre_clip': 23.994102478027344, 'timing/data': 0.0011512080091051757, 'timing/model': 2.030919504002668, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.33}) +[2026-06-23 04:55:12,100][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.24618525803089142, 'train/grad_norm_pre_clip': 41.006404876708984, 'timing/data': 0.05462270398857072, 'timing/model': 2.2632833779789507, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.33}) +[2026-06-23 04:55:16,087][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.2938707172870636, 'train/grad_norm_pre_clip': 34.29279327392578, 'timing/data': 0.0011496049992274493, 'timing/model': 2.0163001969922334, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.33}) +[2026-06-23 04:55:20,164][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.2496664822101593, 'train/grad_norm_pre_clip': 45.609493255615234, 'timing/data': 0.0011385440011508763, 'timing/model': 2.016922254988458, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.33}) +[2026-06-23 04:55:24,154][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.2724182605743408, 'train/grad_norm_pre_clip': 24.208345413208008, 'timing/data': 0.001150927011622116, 'timing/model': 2.0147439560096245, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.33}) +[2026-06-23 04:55:28,198][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.14281636476516724, 'train/grad_norm_pre_clip': 19.91178321838379, 'timing/data': 0.05538319799234159, 'timing/model': 2.0170897000061814, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.33}) +[2026-06-23 04:55:32,202][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.1827998012304306, 'train/grad_norm_pre_clip': 97.42654418945312, 'timing/data': 0.0011130869970656931, 'timing/model': 2.0315537819988094, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.33}) +[2026-06-23 04:55:36,284][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.1480976641178131, 'train/grad_norm_pre_clip': 39.74180603027344, 'timing/data': 0.001124698988860473, 'timing/model': 2.0170369079860393, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.34}) +[2026-06-23 04:55:40,280][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.09914174675941467, 'train/grad_norm_pre_clip': 22.509607315063477, 'timing/data': 0.0013328839850146323, 'timing/model': 2.019139763986459, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.34}) +[2026-06-23 04:55:44,339][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.17337359488010406, 'train/grad_norm_pre_clip': 37.02877426147461, 'timing/data': 0.06771897798171267, 'timing/model': 2.014507871994283, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.34}) +[2026-06-23 04:55:48,535][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.09371214359998703, 'train/grad_norm_pre_clip': 21.21891975402832, 'timing/data': 0.0012549899984151125, 'timing/model': 2.021889742987696, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.34}) +[2026-06-23 04:55:52,615][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.1607646495103836, 'train/grad_norm_pre_clip': 23.74798583984375, 'timing/data': 0.0012547490186989307, 'timing/model': 2.0171811589971185, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.34}) +[2026-06-23 04:55:56,609][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.1217406615614891, 'train/grad_norm_pre_clip': 17.175613403320312, 'timing/data': 0.001250141009222716, 'timing/model': 2.016749519010773, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.34}) +[2026-06-23 04:56:00,655][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.10844509303569794, 'train/grad_norm_pre_clip': 16.48845672607422, 'timing/data': 0.05679511500056833, 'timing/model': 2.0164050910098013, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.34}) +[2026-06-23 04:56:04,646][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.3260049819946289, 'train/grad_norm_pre_clip': 19.576248168945312, 'timing/data': 0.0012245239922776818, 'timing/model': 2.0148955199983902, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.34}) +[2026-06-23 04:56:08,746][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.1093653216958046, 'train/grad_norm_pre_clip': 12.053537368774414, 'timing/data': 0.0011711549886967987, 'timing/model': 2.03618178202305, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.34}) +[2026-06-23 04:56:12,737][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.2445923537015915, 'train/grad_norm_pre_clip': 21.229454040527344, 'timing/data': 0.0013486529933288693, 'timing/model': 2.009357823000755, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.34}) +[2026-06-23 04:56:16,784][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.1037856861948967, 'train/grad_norm_pre_clip': 9.640036582946777, 'timing/data': 0.05821933600236662, 'timing/model': 2.020064290991286, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.34}) +[2026-06-23 04:56:21,030][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.11079400777816772, 'train/grad_norm_pre_clip': 13.571577072143555, 'timing/data': 0.001225885993335396, 'timing/model': 2.268707101990003, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.34}) +[2026-06-23 04:56:25,122][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.16370472311973572, 'train/grad_norm_pre_clip': 16.970067977905273, 'timing/data': 0.0011767650139518082, 'timing/model': 2.0256786349928007, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.34}) +[2026-06-23 04:56:29,124][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.1048278659582138, 'train/grad_norm_pre_clip': 27.314558029174805, 'timing/data': 0.0013975129986647516, 'timing/model': 2.0192652420082595, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.34}) +[2026-06-23 04:56:33,170][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.24043098092079163, 'train/grad_norm_pre_clip': 20.901535034179688, 'timing/data': 0.05379164099576883, 'timing/model': 2.0176179360132664, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.34}) +[2026-06-23 04:56:37,185][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.06901205331087112, 'train/grad_norm_pre_clip': 16.52779197692871, 'timing/data': 0.0012258259812369943, 'timing/model': 2.0353193820046727, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.34}) +[2026-06-23 04:56:41,275][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.12159106880426407, 'train/grad_norm_pre_clip': 27.32568359375, 'timing/data': 0.0011935560032725334, 'timing/model': 2.0286657940014265, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.34}) +[2026-06-23 04:56:45,288][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.08989742398262024, 'train/grad_norm_pre_clip': 14.45238208770752, 'timing/data': 0.001552039000671357, 'timing/model': 2.031016577995615, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.34}) +[2026-06-23 04:56:49,351][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.14368082582950592, 'train/grad_norm_pre_clip': 19.531713485717773, 'timing/data': 0.05838413801393472, 'timing/model': 2.024736680992646, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.34}) +[2026-06-23 04:56:53,424][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.09850746393203735, 'train/grad_norm_pre_clip': 29.902507781982422, 'timing/data': 0.001249950990313664, 'timing/model': 2.0958087389881257, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.34}) +[2026-06-23 04:56:57,525][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.1358557790517807, 'train/grad_norm_pre_clip': 19.523012161254883, 'timing/data': 0.0014399309875443578, 'timing/model': 2.0309570170065854, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.34}) +[2026-06-23 04:57:01,519][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.10874173045158386, 'train/grad_norm_pre_clip': 20.732120513916016, 'timing/data': 0.0012999220052734017, 'timing/model': 2.0153136949811596, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.34}) +[2026-06-23 04:57:05,669][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.16378113627433777, 'train/grad_norm_pre_clip': 48.78146743774414, 'timing/data': 0.05466218799119815, 'timing/model': 2.0202847530017607, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.34}) +[2026-06-23 04:57:09,753][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.11855313926935196, 'train/grad_norm_pre_clip': 32.76554489135742, 'timing/data': 0.0012194240116514266, 'timing/model': 2.0277956939826254, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.34}) +[2026-06-23 04:57:13,876][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.18884460628032684, 'train/grad_norm_pre_clip': 44.232269287109375, 'timing/data': 0.001273523987038061, 'timing/model': 2.0335269560164306, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.34}) +[2026-06-23 04:57:17,866][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.19882892072200775, 'train/grad_norm_pre_clip': 25.637571334838867, 'timing/data': 0.0014723709900863469, 'timing/model': 2.0161307860107627, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.34}) +[2026-06-23 04:57:21,928][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.14005377888679504, 'train/grad_norm_pre_clip': 24.967838287353516, 'timing/data': 0.06175517599331215, 'timing/model': 2.0233722800039686, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.35}) +[2026-06-23 04:57:25,936][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.20262393355369568, 'train/grad_norm_pre_clip': 20.05556869506836, 'timing/data': 0.0012859170092269778, 'timing/model': 2.033092886005761, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.35}) +[2026-06-23 04:57:30,044][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.16101764142513275, 'train/grad_norm_pre_clip': 20.26543426513672, 'timing/data': 0.0011998079717159271, 'timing/model': 2.0112056939979084, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.35}) +[2026-06-23 04:57:34,041][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.29331958293914795, 'train/grad_norm_pre_clip': 30.11471176147461, 'timing/data': 0.001225434010848403, 'timing/model': 2.022644836019026, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.35}) +[2026-06-23 04:57:38,088][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.18050944805145264, 'train/grad_norm_pre_clip': 42.46396255493164, 'timing/data': 0.0615115649998188, 'timing/model': 2.016348952020053, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.35}) +[2026-06-23 04:57:42,077][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.22422052919864655, 'train/grad_norm_pre_clip': 37.524017333984375, 'timing/data': 0.0013663960271514952, 'timing/model': 2.0143930269987322, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.35}) +[2026-06-23 04:57:46,173][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.09069143980741501, 'train/grad_norm_pre_clip': 21.620542526245117, 'timing/data': 0.0012007990153506398, 'timing/model': 2.012355963990558, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.35}) +[2026-06-23 04:57:50,194][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.14816832542419434, 'train/grad_norm_pre_clip': 244.62054443359375, 'timing/data': 0.0015342450060416013, 'timing/model': 2.0346068470098544, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.35}) +[2026-06-23 04:57:54,231][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 900 +[2026-06-23 04:59:49,163][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 900 +[2026-06-23 04:59:49,163][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 900 +[2026-06-23 05:01:56,190][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 900 +[2026-06-23 05:01:56,191][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.13797789812088013, 'train/grad_norm_pre_clip': 23.61318588256836, 'eval/action_loss/seconds': 114.92810397100402, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.08848872780799866, 'eval/latency_2/loss': 0.08848872780799866, 'eval/flappy/loss': 0.08848872780799866, 'eval/flappy/latency_2/loss': 0.08848872780799866, 'eval/flap_precision': 0.8828996282527881, 'eval/flap_recall': 0.9615384615384616, 'eval/flap_f1': 0.9205426356589147, 'eval/flap_rate_model': 0.22416666666666665, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9205426356589146, 'eval/latency_2/flap_precision': 0.8828996282527881, 'eval/latency_2/flap_recall': 0.9615384615384616, 'eval/latency_2/flap_f1': 0.9205426356589147, 'eval/latency_2/flap_rate_model': 0.22416666666666665, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9205426356589146, 'eval/action_classification/data_wait_seconds': 50.36698913574219, 'eval/action_classification/predict_seconds': 74.94490051269531, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.05546935700112954, 'timing/model': 2.012929084012285, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.35}) +[2026-06-23 05:02:00,144][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.14635832607746124, 'train/grad_norm_pre_clip': 43.645294189453125, 'timing/data': 0.001169841008959338, 'timing/model': 2.006136037001852, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.35}) +[2026-06-23 05:02:04,183][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.06717108190059662, 'train/grad_norm_pre_clip': 17.169904708862305, 'timing/data': 0.0015373310016002506, 'timing/model': 1.993165954016149, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.35}) +[2026-06-23 05:02:08,144][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.2904222309589386, 'train/grad_norm_pre_clip': 40.650413513183594, 'timing/data': 0.001658193999901414, 'timing/model': 2.0041886279941536, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.35}) +[2026-06-23 05:02:12,140][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.2290646880865097, 'train/grad_norm_pre_clip': 23.4670467376709, 'timing/data': 0.05681046299287118, 'timing/model': 1.9919636310078204, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.35}) +[2026-06-23 05:02:16,089][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.1918525993824005, 'train/grad_norm_pre_clip': 42.13015365600586, 'timing/data': 0.0012085339985787868, 'timing/model': 1.9991069870011415, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.35}) +[2026-06-23 05:02:20,169][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.09685201942920685, 'train/grad_norm_pre_clip': 13.774490356445312, 'timing/data': 0.001384870003676042, 'timing/model': 2.0186231060069986, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.35}) +[2026-06-23 05:02:24,123][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.10553537309169769, 'train/grad_norm_pre_clip': 15.802299499511719, 'timing/data': 0.0015327019791584462, 'timing/model': 1.9958253820077516, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.35}) +[2026-06-23 05:02:28,259][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.29054516553878784, 'train/grad_norm_pre_clip': 46.40238952636719, 'timing/data': 0.06029275699984282, 'timing/model': 1.9909923349914607, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.35}) +[2026-06-23 05:02:32,206][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.15749938786029816, 'train/grad_norm_pre_clip': 25.611366271972656, 'timing/data': 0.0011363600206095725, 'timing/model': 1.994778327003587, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.35}) +[2026-06-23 05:02:36,251][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.14383907616138458, 'train/grad_norm_pre_clip': 30.202285766601562, 'timing/data': 0.0011163429880980402, 'timing/model': 1.9969516470155213, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.35}) +[2026-06-23 05:02:40,196][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.1390177309513092, 'train/grad_norm_pre_clip': 18.08781623840332, 'timing/data': 0.0011386440019123256, 'timing/model': 1.9931748939852696, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.35}) +[2026-06-23 05:02:44,206][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.2620493173599243, 'train/grad_norm_pre_clip': 34.456214904785156, 'timing/data': 0.058178228995529935, 'timing/model': 1.999085600982653, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.35}) +[2026-06-23 05:02:48,158][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.15046165883541107, 'train/grad_norm_pre_clip': 21.84469985961914, 'timing/data': 0.0011340150085743517, 'timing/model': 1.9996549280185718, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.35}) +[2026-06-23 05:02:52,202][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.061749815940856934, 'train/grad_norm_pre_clip': 12.942253112792969, 'timing/data': 0.0012960750027559698, 'timing/model': 1.998606095003197, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.35}) +[2026-06-23 05:02:56,158][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.17372845113277435, 'train/grad_norm_pre_clip': 27.891468048095703, 'timing/data': 0.0011619669967330992, 'timing/model': 1.9996368740103208, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.35}) +[2026-06-23 05:03:00,338][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.11097637563943863, 'train/grad_norm_pre_clip': 31.012924194335938, 'timing/data': 0.05664367298595607, 'timing/model': 2.0033262679935433, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.35}) +[2026-06-23 05:03:04,304][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.11755003035068512, 'train/grad_norm_pre_clip': 18.103565216064453, 'timing/data': 0.0013920120254624635, 'timing/model': 2.004898207000224, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.35}) +[2026-06-23 05:03:08,361][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.11757397651672363, 'train/grad_norm_pre_clip': 21.658565521240234, 'timing/data': 0.001155183999799192, 'timing/model': 2.007566432992462, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.36}) +[2026-06-23 05:03:12,328][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.14557521045207977, 'train/grad_norm_pre_clip': 32.5863151550293, 'timing/data': 0.0013222529960330576, 'timing/model': 2.008796281996183, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.36}) +[2026-06-23 05:03:16,352][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.2666063904762268, 'train/grad_norm_pre_clip': 58.54202651977539, 'timing/data': 0.06062173401005566, 'timing/model': 2.004777861002367, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.36}) +[2026-06-23 05:03:20,346][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.13873456418514252, 'train/grad_norm_pre_clip': 21.045663833618164, 'timing/data': 0.0013544730027206242, 'timing/model': 2.028289121983107, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.36}) +[2026-06-23 05:03:24,413][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.15409356355667114, 'train/grad_norm_pre_clip': 26.37915802001953, 'timing/data': 0.0011016659846063703, 'timing/model': 2.0066913279879373, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.36}) +[2026-06-23 05:03:28,380][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.18050749599933624, 'train/grad_norm_pre_clip': 19.605133056640625, 'timing/data': 0.0013169929734431207, 'timing/model': 2.0047743800096214, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.36}) +[2026-06-23 05:03:32,657][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.049671705812215805, 'train/grad_norm_pre_clip': 10.580028533935547, 'timing/data': 0.05614626198075712, 'timing/model': 2.255049166997196, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.36}) +[2026-06-23 05:03:36,634][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.13138902187347412, 'train/grad_norm_pre_clip': 25.209306716918945, 'timing/data': 0.0012928390060551465, 'timing/model': 2.0080283240240533, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.36}) +[2026-06-23 05:03:40,691][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.17273341119289398, 'train/grad_norm_pre_clip': 34.6699104309082, 'timing/data': 0.001149965013610199, 'timing/model': 2.006056445010472, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.36}) +[2026-06-23 05:03:44,664][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.04820920526981354, 'train/grad_norm_pre_clip': 16.508760452270508, 'timing/data': 0.0011343259830027819, 'timing/model': 2.0067861499846913, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.36}) +[2026-06-23 05:03:48,697][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.20134717226028442, 'train/grad_norm_pre_clip': 26.42805290222168, 'timing/data': 0.06285391701385379, 'timing/model': 2.006270422018133, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.36}) +[2026-06-23 05:03:52,667][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.09883664548397064, 'train/grad_norm_pre_clip': 24.4779109954834, 'timing/data': 0.0011274529970251024, 'timing/model': 2.008321025990881, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.36}) +[2026-06-23 05:03:56,754][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.11598841845989227, 'train/grad_norm_pre_clip': 16.580137252807617, 'timing/data': 0.001167446986073628, 'timing/model': 2.0267970739805605, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.36}) +[2026-06-23 05:04:00,731][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.2578715682029724, 'train/grad_norm_pre_clip': 47.5646858215332, 'timing/data': 0.0011476200015749782, 'timing/model': 2.009204008994857, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.36}) +[2026-06-23 05:04:04,764][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.11031437665224075, 'train/grad_norm_pre_clip': 20.40753173828125, 'timing/data': 0.057591962977312505, 'timing/model': 2.0111981210066006, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.36}) +[2026-06-23 05:04:08,848][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.05670175328850746, 'train/grad_norm_pre_clip': 11.710742950439453, 'timing/data': 0.001279093004995957, 'timing/model': 2.031331651000073, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.36}) +[2026-06-23 05:04:12,914][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.20087403059005737, 'train/grad_norm_pre_clip': 43.282100677490234, 'timing/data': 0.0011107019963674247, 'timing/model': 2.0094310060085263, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.36}) +[2026-06-23 05:04:16,896][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.11946702003479004, 'train/grad_norm_pre_clip': 16.724210739135742, 'timing/data': 0.0011526099988259375, 'timing/model': 2.0150116750155576, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.36}) +[2026-06-23 05:04:20,951][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.12118889391422272, 'train/grad_norm_pre_clip': 23.142576217651367, 'timing/data': 0.07042248200741597, 'timing/model': 2.012468049011659, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.36}) +[2026-06-23 05:04:24,958][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.0620938278734684, 'train/grad_norm_pre_clip': 15.579849243164062, 'timing/data': 0.0013028180110268295, 'timing/model': 2.0326620709965937, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.36}) +[2026-06-23 05:04:29,062][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.23436638712882996, 'train/grad_norm_pre_clip': 40.63518524169922, 'timing/data': 0.0012678030179813504, 'timing/model': 2.0388148360070772, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.36}) +[2026-06-23 05:04:33,069][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.06584727764129639, 'train/grad_norm_pre_clip': 16.596967697143555, 'timing/data': 0.001437225000699982, 'timing/model': 2.0161203060124535, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.36}) +[2026-06-23 05:04:37,130][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.3036455512046814, 'train/grad_norm_pre_clip': 36.96947479248047, 'timing/data': 0.0700851879955735, 'timing/model': 2.0168624139914755, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.36}) +[2026-06-23 05:04:41,401][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.16568011045455933, 'train/grad_norm_pre_clip': 44.71861267089844, 'timing/data': 0.0012916669948026538, 'timing/model': 2.296724224986974, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.36}) +[2026-06-23 05:04:45,505][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.19134163856506348, 'train/grad_norm_pre_clip': 53.692752838134766, 'timing/data': 0.0012164970103185624, 'timing/model': 2.0373649509856477, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.36}) +[2026-06-23 05:04:49,497][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.2788607180118561, 'train/grad_norm_pre_clip': 49.335933685302734, 'timing/data': 0.0014148140035104007, 'timing/model': 2.01931208698079, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.36}) +[2026-06-23 05:04:53,554][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.18898537755012512, 'train/grad_norm_pre_clip': 30.76636505126953, 'timing/data': 0.0619011539965868, 'timing/model': 2.019403902988415, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.37}) +[2026-06-23 05:04:57,555][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.2736462652683258, 'train/grad_norm_pre_clip': 23.519174575805664, 'timing/data': 0.0011117240064777434, 'timing/model': 2.0274145519942977, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.37}) +[2026-06-23 05:05:01,656][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.12104181200265884, 'train/grad_norm_pre_clip': 25.597820281982422, 'timing/data': 0.001258686010260135, 'timing/model': 2.014474846975645, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.37}) +[2026-06-23 05:05:05,652][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.10493436455726624, 'train/grad_norm_pre_clip': 12.073748588562012, 'timing/data': 0.001317393995122984, 'timing/model': 2.0147615839960054, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.37}) +[2026-06-23 05:05:09,714][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.18053720891475677, 'train/grad_norm_pre_clip': 49.509368896484375, 'timing/data': 0.06579071099986322, 'timing/model': 2.021591550001176, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.37}) +[2026-06-23 05:05:13,708][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.14335040748119354, 'train/grad_norm_pre_clip': 40.05617141723633, 'timing/data': 0.0012273880129214376, 'timing/model': 2.020217616984155, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.37}) +[2026-06-23 05:05:18,068][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.07446042448282242, 'train/grad_norm_pre_clip': 24.567873001098633, 'timing/data': 0.0013854699791409075, 'timing/model': 2.0115996889944654, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.37}) +[2026-06-23 05:05:22,065][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.10604534298181534, 'train/grad_norm_pre_clip': 31.504182815551758, 'timing/data': 0.0018256139883305877, 'timing/model': 2.0220245090022217, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.37}) +[2026-06-23 05:05:26,129][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.47048118710517883, 'train/grad_norm_pre_clip': 73.03041076660156, 'timing/data': 0.06933254501200281, 'timing/model': 2.0142258139967453, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.37}) +[2026-06-23 05:05:30,132][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.22608567774295807, 'train/grad_norm_pre_clip': 60.91503143310547, 'timing/data': 0.0013970310101285577, 'timing/model': 2.0192610640078783, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.37}) +[2026-06-23 05:05:34,217][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.3005097210407257, 'train/grad_norm_pre_clip': 36.317657470703125, 'timing/data': 0.00136140501126647, 'timing/model': 2.0212944579834584, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.37}) +[2026-06-23 05:05:38,230][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.20331209897994995, 'train/grad_norm_pre_clip': 21.427289962768555, 'timing/data': 0.0012562719930429012, 'timing/model': 2.0379803469986655, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.37}) +[2026-06-23 05:05:42,285][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.04372018203139305, 'train/grad_norm_pre_clip': 7.022222518920898, 'timing/data': 0.060883772996021435, 'timing/model': 2.018724461988313, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.37}) +[2026-06-23 05:05:46,287][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.15530581772327423, 'train/grad_norm_pre_clip': 23.706165313720703, 'timing/data': 0.0012026120093651116, 'timing/model': 2.025249275000533, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.37}) +[2026-06-23 05:05:50,585][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.22769322991371155, 'train/grad_norm_pre_clip': 50.27400207519531, 'timing/data': 0.001480456005083397, 'timing/model': 2.2126198309997562, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.37}) +[2026-06-23 05:05:54,585][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.15304729342460632, 'train/grad_norm_pre_clip': 29.23222541809082, 'timing/data': 0.0012460119905881584, 'timing/model': 2.0252954329771455, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.37}) +[2026-06-23 05:05:58,657][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.09426859021186829, 'train/grad_norm_pre_clip': 18.525997161865234, 'timing/data': 0.07028404300217517, 'timing/model': 2.020243754988769, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.37}) +[2026-06-23 05:06:02,655][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.08759048581123352, 'train/grad_norm_pre_clip': 18.421022415161133, 'timing/data': 0.0012502600147854537, 'timing/model': 2.0204711809928995, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.37}) +[2026-06-23 05:06:06,757][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.10775142908096313, 'train/grad_norm_pre_clip': 12.761887550354004, 'timing/data': 0.0012509709922596812, 'timing/model': 2.0179924689873587, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.37}) +[2026-06-23 05:06:10,748][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.2934260964393616, 'train/grad_norm_pre_clip': 62.68219757080078, 'timing/data': 0.0013187459844630212, 'timing/model': 2.0169476240116637, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.37}) +[2026-06-23 05:06:14,814][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.3314090669155121, 'train/grad_norm_pre_clip': 72.19187927246094, 'timing/data': 0.06626605801284313, 'timing/model': 2.0191493289894424, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.37}) +[2026-06-23 05:06:18,811][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.22480717301368713, 'train/grad_norm_pre_clip': 45.51388168334961, 'timing/data': 0.0014383780071511865, 'timing/model': 2.02037007099716, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.37}) +[2026-06-23 05:06:22,907][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.16368798911571503, 'train/grad_norm_pre_clip': 25.973491668701172, 'timing/data': 0.0012182710052002221, 'timing/model': 2.034040760016069, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.37}) +[2026-06-23 05:06:26,940][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.09997395426034927, 'train/grad_norm_pre_clip': 20.956884384155273, 'timing/data': 0.0011511180200614035, 'timing/model': 2.0185343319899403, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.37}) +[2026-06-23 05:06:30,991][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.19427475333213806, 'train/grad_norm_pre_clip': 31.23318862915039, 'timing/data': 0.06253717601066455, 'timing/model': 2.0167080749815796, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.37}) +[2026-06-23 05:06:34,987][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.39385026693344116, 'train/grad_norm_pre_clip': 65.44486236572266, 'timing/data': 0.0012072509853169322, 'timing/model': 2.023175401991466, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.38}) +[2026-06-23 05:06:39,074][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.18230389058589935, 'train/grad_norm_pre_clip': 32.96658706665039, 'timing/data': 0.001225363987032324, 'timing/model': 2.0212602200044785, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.38}) +[2026-06-23 05:06:43,074][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.166346937417984, 'train/grad_norm_pre_clip': 41.83935546875, 'timing/data': 0.0012480069999583066, 'timing/model': 2.0240483680099715, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.38}) +[2026-06-23 05:06:47,124][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.08498480916023254, 'train/grad_norm_pre_clip': 16.168546676635742, 'timing/data': 0.0573873000103049, 'timing/model': 2.0170047199935652, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.38}) +[2026-06-23 05:06:51,142][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.21608158946037292, 'train/grad_norm_pre_clip': 25.563159942626953, 'timing/data': 0.0017677370051387697, 'timing/model': 2.0417432689864654, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.38}) +[2026-06-23 05:06:55,246][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.0877595990896225, 'train/grad_norm_pre_clip': 21.80718231201172, 'timing/data': 0.001398352993419394, 'timing/model': 2.0337889380170964, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.38}) +[2026-06-23 05:06:59,361][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.09566842764616013, 'train/grad_norm_pre_clip': 21.136398315429688, 'timing/data': 0.0014185210166033357, 'timing/model': 2.134704978001537, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.38}) +[2026-06-23 05:07:03,422][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.26710817217826843, 'train/grad_norm_pre_clip': 19.677114486694336, 'timing/data': 0.06884988999809138, 'timing/model': 2.0184476240247022, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.38}) +[2026-06-23 05:07:07,419][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.09711746871471405, 'train/grad_norm_pre_clip': 13.387737274169922, 'timing/data': 0.0012536860012914985, 'timing/model': 2.021502376010176, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.38}) +[2026-06-23 05:07:11,515][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.17232128977775574, 'train/grad_norm_pre_clip': 29.5614013671875, 'timing/data': 0.0019215900101698935, 'timing/model': 2.016101652989164, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.38}) +[2026-06-23 05:07:15,527][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.13613000512123108, 'train/grad_norm_pre_clip': 30.49380111694336, 'timing/data': 0.0014086029841564596, 'timing/model': 2.033401516993763, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.38}) +[2026-06-23 05:07:19,580][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.14621709287166595, 'train/grad_norm_pre_clip': 20.35563850402832, 'timing/data': 0.0627934189978987, 'timing/model': 2.0155646940111183, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.38}) +[2026-06-23 05:07:23,568][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.12078538537025452, 'train/grad_norm_pre_clip': 22.127445220947266, 'timing/data': 0.001274234993616119, 'timing/model': 2.0150043159956113, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.38}) +[2026-06-23 05:07:27,657][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.11549945175647736, 'train/grad_norm_pre_clip': 38.730186462402344, 'timing/data': 0.0012341710098553449, 'timing/model': 2.015605273976689, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.38}) +[2026-06-23 05:07:31,781][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.16300854086875916, 'train/grad_norm_pre_clip': 16.599477767944336, 'timing/data': 0.0014036440115887672, 'timing/model': 2.151151691999985, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.38}) +[2026-06-23 05:07:35,840][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.22861678898334503, 'train/grad_norm_pre_clip': 23.339597702026367, 'timing/data': 0.0675157850200776, 'timing/model': 2.0189122979936656, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.38}) +[2026-06-23 05:07:39,849][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.13878758251667023, 'train/grad_norm_pre_clip': 31.530681610107422, 'timing/data': 0.0014365539827849716, 'timing/model': 2.0335401149932295, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.38}) +[2026-06-23 05:07:43,938][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.12379719316959381, 'train/grad_norm_pre_clip': 15.557876586914062, 'timing/data': 0.0012173790019005537, 'timing/model': 2.020963863003999, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.38}) +[2026-06-23 05:07:47,928][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.1527193784713745, 'train/grad_norm_pre_clip': 13.71086597442627, 'timing/data': 0.0013948369887657464, 'timing/model': 2.017018240992911, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.38}) +[2026-06-23 05:07:51,987][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.18242739140987396, 'train/grad_norm_pre_clip': 44.26139450073242, 'timing/data': 0.06575460999738425, 'timing/model': 2.018165877001593, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.38}) +[2026-06-23 05:07:55,980][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.13116997480392456, 'train/grad_norm_pre_clip': 15.771013259887695, 'timing/data': 0.0012170689878985286, 'timing/model': 2.017200711008627, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.38}) +[2026-06-23 05:08:00,083][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.09771791100502014, 'train/grad_norm_pre_clip': 12.684791564941406, 'timing/data': 0.0015215710154734552, 'timing/model': 2.0356263749999925, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.38}) +[2026-06-23 05:08:04,108][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.07831775397062302, 'train/grad_norm_pre_clip': 11.784543991088867, 'timing/data': 0.0012792540073860437, 'timing/model': 2.0486489349859767, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.38}) +[2026-06-23 05:08:08,400][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.07743968814611435, 'train/grad_norm_pre_clip': 14.449764251708984, 'timing/data': 0.05559508697479032, 'timing/model': 2.019121212011669, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.38}) +[2026-06-23 05:08:12,425][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.11659428477287292, 'train/grad_norm_pre_clip': 20.4554443359375, 'timing/data': 0.0014881000097375363, 'timing/model': 2.0339672720001545, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.38}) +[2026-06-23 05:08:16,532][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.15691019594669342, 'train/grad_norm_pre_clip': 18.121156692504883, 'timing/data': 0.001741898973705247, 'timing/model': 2.0305734679859597, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.38}) +[2026-06-23 05:08:20,549][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.15358515083789825, 'train/grad_norm_pre_clip': 15.146968841552734, 'timing/data': 0.0018213249859400094, 'timing/model': 2.037244321021717, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.39}) +[2026-06-23 05:08:24,603][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.2766086161136627, 'train/grad_norm_pre_clip': 26.851423263549805, 'timing/data': 0.054745456989621744, 'timing/model': 2.0187481619941536, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.39}) +[2026-06-23 05:08:28,602][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.16933095455169678, 'train/grad_norm_pre_clip': 32.35740661621094, 'timing/data': 0.0012230799766257405, 'timing/model': 2.021328266011551, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.39}) +[2026-06-23 05:08:32,688][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.15190862119197845, 'train/grad_norm_pre_clip': 20.783510208129883, 'timing/data': 0.0014106759917922318, 'timing/model': 2.0189066889870446, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.39}) +[2026-06-23 05:08:36,684][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.12141726166009903, 'train/grad_norm_pre_clip': 12.657203674316406, 'timing/data': 0.001429001014912501, 'timing/model': 2.018106108007487, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.39}) +[2026-06-23 05:08:40,833][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1000 +[2026-06-23 05:10:34,957][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1000 +[2026-06-23 05:10:34,958][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1000 +[2026-06-23 05:12:41,748][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1000 +[2026-06-23 05:22:41,181][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-23 05:22:41,183][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.07124984264373779, 'train/grad_norm_pre_clip': 17.075035095214844, 'eval/action_loss/seconds': 114.12088971197954, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07729317992925644, 'eval/latency_2/loss': 0.07729317992925644, 'eval/flappy/loss': 0.07729317992925644, 'eval/flappy/latency_2/loss': 0.07729317992925644, 'eval/flap_precision': 0.9456214689265536, 'eval/flap_recall': 0.9035087719298246, 'eval/flap_f1': 0.9240855762594893, 'eval/flap_rate_model': 0.19666666666666666, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9240855762594893, 'eval/latency_2/flap_precision': 0.9456214689265536, 'eval/latency_2/flap_recall': 0.9035087719298246, 'eval/latency_2/flap_f1': 0.9240855762594893, 'eval/latency_2/flap_rate_model': 0.19666666666666666, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9240855762594893, 'eval/action_classification/data_wait_seconds': 50.15794372558594, 'eval/action_classification/predict_seconds': 74.92803955078125, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 153.99999999999991, 'rl_games_eval/mid_train/mean_length': 1261.2, 'rl_games_eval/mid_train/std_reward': 120.2648743399357, 'rl_games_eval/mid_train/std_length': 970.7072473202206, 'rl_games_eval/mid_train/macro_mean_reward': 153.99999999999991, 'rl_games_eval/mid_train/macro_mean_length': 1261.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 153.99999999999991, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1261.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 120.2648743399357, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 970.7072473202206, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 153.99999999999991, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.06688274297630414, 'timing/model': 2.1071858039940707, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.39}) +[2026-06-23 05:22:41,184][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_1000_state +[2026-06-23 05:22:41,190][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 05:22:51,264][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_1000_state/model.safetensors +[2026-06-23 05:22:56,542][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_1000_state/optimizer.bin +[2026-06-23 05:22:56,543][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_1000_state/sampler.bin +[2026-06-23 05:22:56,543][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-23 05:22:56,544][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-23 05:22:56,659][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-23 05:22:56,660][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 05:22:56,684][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 05:23:00,546][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.16408903896808624, 'train/grad_norm_pre_clip': 28.062374114990234, 'timing/data': 0.0014630930090788752, 'timing/model': 1.9534800069814082, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.39}) +[2026-06-23 05:23:04,506][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.18250533938407898, 'train/grad_norm_pre_clip': 54.202186584472656, 'timing/data': 0.0013429410173557699, 'timing/model': 1.9540010199998505, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.39}) +[2026-06-23 05:23:08,404][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.09470073878765106, 'train/grad_norm_pre_clip': 16.701820373535156, 'timing/data': 0.0014158060075715184, 'timing/model': 1.9781184800085612, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.39}) +[2026-06-23 05:23:12,351][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.0902317613363266, 'train/grad_norm_pre_clip': 13.581988334655762, 'timing/data': 0.06782809199648909, 'timing/model': 1.9596677490044385, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.39}) +[2026-06-23 05:23:16,243][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.12149886786937714, 'train/grad_norm_pre_clip': 30.466869354248047, 'timing/data': 0.0014753549767192453, 'timing/model': 1.969736341998214, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.39}) +[2026-06-23 05:23:20,251][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.19034872949123383, 'train/grad_norm_pre_clip': 44.85337829589844, 'timing/data': 0.0013798899890389293, 'timing/model': 1.987374390999321, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.39}) +[2026-06-23 05:23:24,147][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.15302877128124237, 'train/grad_norm_pre_clip': 28.290342330932617, 'timing/data': 0.0014105149894021451, 'timing/model': 1.966788944992004, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.39}) +[2026-06-23 05:23:28,344][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.20564161241054535, 'train/grad_norm_pre_clip': 38.50178146362305, 'timing/data': 0.054469106980832294, 'timing/model': 2.2174589310016017, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.39}) +[2026-06-23 05:23:32,246][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.07630178332328796, 'train/grad_norm_pre_clip': 13.263142585754395, 'timing/data': 0.0012129300157539546, 'timing/model': 1.9727127389924135, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.39}) +[2026-06-23 05:23:36,253][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.26018261909484863, 'train/grad_norm_pre_clip': 64.07586669921875, 'timing/data': 0.001226103981025517, 'timing/model': 1.9886407259909902, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.39}) +[2026-06-23 05:23:40,172][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.16050049662590027, 'train/grad_norm_pre_clip': 42.32362365722656, 'timing/data': 0.0016817660070955753, 'timing/model': 1.9776047589839436, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.39}) +[2026-06-23 05:23:44,148][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.10367762297391891, 'train/grad_norm_pre_clip': 18.807064056396484, 'timing/data': 0.06125927701941691, 'timing/model': 1.9795774540107232, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.39}) +[2026-06-23 05:23:48,069][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.2453899383544922, 'train/grad_norm_pre_clip': 34.06049346923828, 'timing/data': 0.0012045850162394345, 'timing/model': 1.979932532995008, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.39}) +[2026-06-23 05:23:52,101][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.11189472675323486, 'train/grad_norm_pre_clip': 19.825206756591797, 'timing/data': 0.001375268999254331, 'timing/model': 2.005349863000447, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.39}) +[2026-06-23 05:23:56,049][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.1531793624162674, 'train/grad_norm_pre_clip': 28.789724349975586, 'timing/data': 0.0016752239898778498, 'timing/model': 2.006006117007928, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.39}) +[2026-06-23 05:24:00,043][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.16511747241020203, 'train/grad_norm_pre_clip': 33.842254638671875, 'timing/data': 0.05399620599928312, 'timing/model': 1.9895002479897812, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.39}) +[2026-06-23 05:24:04,236][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.25341543555259705, 'train/grad_norm_pre_clip': 30.511451721191406, 'timing/data': 0.0012296809873078018, 'timing/model': 1.983566854003584, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.39}) +[2026-06-23 05:24:08,261][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.14236752688884735, 'train/grad_norm_pre_clip': 19.929351806640625, 'timing/data': 0.001422055996954441, 'timing/model': 1.9885135820077267, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.39}) +[2026-06-23 05:24:12,228][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.10810451954603195, 'train/grad_norm_pre_clip': 17.723861694335938, 'timing/data': 0.0015185340016614646, 'timing/model': 2.01234798898804, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.39}) +[2026-06-23 05:24:16,248][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.0919748991727829, 'train/grad_norm_pre_clip': 20.721864700317383, 'timing/data': 0.06346872099675238, 'timing/model': 2.0047354720009025, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.39}) +[2026-06-23 05:24:20,198][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.1784447729587555, 'train/grad_norm_pre_clip': 23.202512741088867, 'timing/data': 0.0011817119957413524, 'timing/model': 1.995734372001607, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.4}) +[2026-06-23 05:24:24,237][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.1399955153465271, 'train/grad_norm_pre_clip': 19.924715042114258, 'timing/data': 0.0012359330139588565, 'timing/model': 1.9942893740080763, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.4}) +[2026-06-23 05:24:28,217][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.1891939491033554, 'train/grad_norm_pre_clip': 40.76639938354492, 'timing/data': 0.0012716779892798513, 'timing/model': 2.0154908600088675, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.4}) +[2026-06-23 05:24:32,237][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.11250130832195282, 'train/grad_norm_pre_clip': 23.217830657958984, 'timing/data': 0.057492696010740474, 'timing/model': 2.001229544985108, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.4}) +[2026-06-23 05:24:36,450][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.10549737513065338, 'train/grad_norm_pre_clip': 17.026044845581055, 'timing/data': 0.0012306729913689196, 'timing/model': 2.013789432006888, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.4}) +[2026-06-23 05:24:40,502][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.25380435585975647, 'train/grad_norm_pre_clip': 48.456478118896484, 'timing/data': 0.001401468995027244, 'timing/model': 2.0016577689966653, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.4}) +[2026-06-23 05:24:44,462][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.21133480966091156, 'train/grad_norm_pre_clip': 49.050804138183594, 'timing/data': 0.0014055559877306223, 'timing/model': 2.0021912759984843, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.4}) +[2026-06-23 05:24:48,482][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.32285600900650024, 'train/grad_norm_pre_clip': 47.406131744384766, 'timing/data': 0.0557305820111651, 'timing/model': 2.0049183929804713, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.4}) +[2026-06-23 05:24:52,449][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.23393207788467407, 'train/grad_norm_pre_clip': 23.715757369995117, 'timing/data': 0.0012801149860024452, 'timing/model': 2.0032977419905365, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.4}) +[2026-06-23 05:24:56,516][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.030774928629398346, 'train/grad_norm_pre_clip': 5.6873626708984375, 'timing/data': 0.0011796980106737465, 'timing/model': 2.0063161170110106, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.4}) +[2026-06-23 05:25:00,489][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.22451144456863403, 'train/grad_norm_pre_clip': 43.01454544067383, 'timing/data': 0.001493087998824194, 'timing/model': 2.0087690479995217, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.4}) +[2026-06-23 05:25:04,523][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.24215194582939148, 'train/grad_norm_pre_clip': 65.3724365234375, 'timing/data': 0.06657876798999496, 'timing/model': 2.0052680730004795, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.4}) +[2026-06-23 05:25:08,542][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.2589878737926483, 'train/grad_norm_pre_clip': 73.96696472167969, 'timing/data': 0.0012433960218913853, 'timing/model': 2.0101928990043234, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.4}) +[2026-06-23 05:25:12,610][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.1666598618030548, 'train/grad_norm_pre_clip': 41.4536247253418, 'timing/data': 0.0012671610165853053, 'timing/model': 2.0135006089985836, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.4}) +[2026-06-23 05:25:16,588][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.12693384289741516, 'train/grad_norm_pre_clip': 16.022335052490234, 'timing/data': 0.0014293499989435077, 'timing/model': 2.0121009350114036, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.4}) +[2026-06-23 05:25:20,623][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.08434943854808807, 'train/grad_norm_pre_clip': 17.34680938720703, 'timing/data': 0.056779491977067664, 'timing/model': 2.007911628985312, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.4}) +[2026-06-23 05:25:24,625][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.10794587433338165, 'train/grad_norm_pre_clip': 26.540985107421875, 'timing/data': 0.001505721011199057, 'timing/model': 2.0311956010118593, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.4}) +[2026-06-23 05:25:28,694][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.0895204246044159, 'train/grad_norm_pre_clip': 21.7138671875, 'timing/data': 0.001329505001194775, 'timing/model': 2.006487870006822, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.4}) +[2026-06-23 05:25:32,698][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.15734057128429413, 'train/grad_norm_pre_clip': 19.3503360748291, 'timing/data': 0.0014688829833175987, 'timing/model': 2.0319667459989432, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.4}) +[2026-06-23 05:25:36,758][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.0861218199133873, 'train/grad_norm_pre_clip': 20.36255645751953, 'timing/data': 0.06421832201885991, 'timing/model': 2.0253179279970936, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.4}) +[2026-06-23 05:25:40,820][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.13303422927856445, 'train/grad_norm_pre_clip': 26.046916961669922, 'timing/data': 0.0012668189883697778, 'timing/model': 2.0861977339955047, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.4}) +[2026-06-23 05:25:44,884][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.1373862475156784, 'train/grad_norm_pre_clip': 24.971057891845703, 'timing/data': 0.0011179450084455311, 'timing/model': 2.000532864010893, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.4}) +[2026-06-23 05:25:48,866][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.06007418781518936, 'train/grad_norm_pre_clip': 11.101271629333496, 'timing/data': 0.0012203340011183172, 'timing/model': 2.0204805250104982, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.4}) +[2026-06-23 05:25:52,893][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.11616859585046768, 'train/grad_norm_pre_clip': 31.047218322753906, 'timing/data': 0.056501160986954346, 'timing/model': 2.005977189022815, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.4}) +[2026-06-23 05:25:56,861][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.10840047895908356, 'train/grad_norm_pre_clip': 21.57954978942871, 'timing/data': 0.001240379991941154, 'timing/model': 2.006785942008719, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.4}) +[2026-06-23 05:26:01,164][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.1438211053609848, 'train/grad_norm_pre_clip': 22.753084182739258, 'timing/data': 0.0012988089874852449, 'timing/model': 2.0190991339914035, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.4}) +[2026-06-23 05:26:05,148][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.04288691282272339, 'train/grad_norm_pre_clip': 12.95855712890625, 'timing/data': 0.0015191760030575097, 'timing/model': 2.012235797010362, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.41}) +[2026-06-23 05:26:09,404][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.1219855323433876, 'train/grad_norm_pre_clip': 21.100738525390625, 'timing/data': 0.2532357279851567, 'timing/model': 2.0317028610152192, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.41}) +[2026-06-23 05:26:13,388][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.21556171774864197, 'train/grad_norm_pre_clip': 24.079301834106445, 'timing/data': 0.0014924569986760616, 'timing/model': 2.013714517990593, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.41}) +[2026-06-23 05:26:17,694][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.10842232406139374, 'train/grad_norm_pre_clip': 34.79683303833008, 'timing/data': 0.001384948001941666, 'timing/model': 2.0132984000083525, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.41}) +[2026-06-23 05:26:21,718][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.2308337241411209, 'train/grad_norm_pre_clip': 27.749130249023438, 'timing/data': 0.0014786009851377457, 'timing/model': 2.0396496780158486, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.41}) +[2026-06-23 05:26:25,781][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.1411338597536087, 'train/grad_norm_pre_clip': 14.833691596984863, 'timing/data': 0.06370715101365931, 'timing/model': 2.022300994984107, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.41}) +[2026-06-23 05:26:29,788][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.16104398667812347, 'train/grad_norm_pre_clip': 26.498802185058594, 'timing/data': 0.0012062379973940551, 'timing/model': 2.014973572018789, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.41}) +[2026-06-23 05:26:33,866][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.06661465764045715, 'train/grad_norm_pre_clip': 13.325202941894531, 'timing/data': 0.0013197269872762263, 'timing/model': 2.0148896560131107, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.41}) +[2026-06-23 05:26:37,860][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.22224824130535126, 'train/grad_norm_pre_clip': 45.40702819824219, 'timing/data': 0.001304829987930134, 'timing/model': 2.026295260991901, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.41}) +[2026-06-23 05:26:42,113][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.18108965456485748, 'train/grad_norm_pre_clip': 28.752073287963867, 'timing/data': 0.09526893400470726, 'timing/model': 2.061793202010449, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.41}) +[2026-06-23 05:26:46,111][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.1013563945889473, 'train/grad_norm_pre_clip': 24.168113708496094, 'timing/data': 0.0011276729928795248, 'timing/model': 2.0108553229947574, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.41}) +[2026-06-23 05:26:50,208][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.2917307913303375, 'train/grad_norm_pre_clip': 73.02863311767578, 'timing/data': 0.0012785719882231206, 'timing/model': 2.0090384159993846, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.41}) +[2026-06-23 05:26:54,207][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.27877387404441833, 'train/grad_norm_pre_clip': 75.91471099853516, 'timing/data': 0.0014597759873140603, 'timing/model': 2.0166077550093178, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.41}) +[2026-06-23 05:26:58,250][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.4487307369709015, 'train/grad_norm_pre_clip': 80.68099212646484, 'timing/data': 0.05355052501545288, 'timing/model': 2.0167354800214525, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.41}) +[2026-06-23 05:27:02,252][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.3281305432319641, 'train/grad_norm_pre_clip': 53.11722946166992, 'timing/data': 0.0012884799798484892, 'timing/model': 2.0265547040035017, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.41}) +[2026-06-23 05:27:06,347][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.18395164608955383, 'train/grad_norm_pre_clip': 24.241769790649414, 'timing/data': 0.0011916309886146337, 'timing/model': 2.018253961985465, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.41}) +[2026-06-23 05:27:10,336][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.157274067401886, 'train/grad_norm_pre_clip': 23.265832901000977, 'timing/data': 0.0015068629872985184, 'timing/model': 2.017921797989402, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.41}) +[2026-06-23 05:27:14,410][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.1887732893228531, 'train/grad_norm_pre_clip': 32.115028381347656, 'timing/data': 0.06512019399087876, 'timing/model': 2.0234762049803976, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.41}) +[2026-06-23 05:27:18,398][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.13980932533740997, 'train/grad_norm_pre_clip': 24.821929931640625, 'timing/data': 0.0011783060035668314, 'timing/model': 2.0148385599895846, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.41}) +[2026-06-23 05:27:22,570][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.25986501574516296, 'train/grad_norm_pre_clip': 37.03438186645508, 'timing/data': 0.001185339002404362, 'timing/model': 2.111227156012319, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.41}) +[2026-06-23 05:27:26,566][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.10489164292812347, 'train/grad_norm_pre_clip': 15.450090408325195, 'timing/data': 0.001396339008351788, 'timing/model': 2.020734607009217, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.41}) +[2026-06-23 05:27:30,616][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.107865110039711, 'train/grad_norm_pre_clip': 12.727551460266113, 'timing/data': 0.05322730899206363, 'timing/model': 2.017749922000803, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.41}) +[2026-06-23 05:27:34,609][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.1745702475309372, 'train/grad_norm_pre_clip': 16.100780487060547, 'timing/data': 0.0012126700021326542, 'timing/model': 2.019042598985834, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.41}) +[2026-06-23 05:27:38,687][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.1732991635799408, 'train/grad_norm_pre_clip': 20.310848236083984, 'timing/data': 0.001171443989733234, 'timing/model': 2.0165739670046605, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.41}) +[2026-06-23 05:27:42,683][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.28175100684165955, 'train/grad_norm_pre_clip': 35.12908172607422, 'timing/data': 0.0012558400048874319, 'timing/model': 2.020478427002672, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.41}) +[2026-06-23 05:27:46,734][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.15591512620449066, 'train/grad_norm_pre_clip': 16.058685302734375, 'timing/data': 0.05465373498736881, 'timing/model': 2.021928953996394, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.41}) +[2026-06-23 05:27:50,731][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.16156592965126038, 'train/grad_norm_pre_clip': 15.269271850585938, 'timing/data': 0.001208761997986585, 'timing/model': 2.0169367310008965, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.42}) +[2026-06-23 05:27:54,818][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.12815600633621216, 'train/grad_norm_pre_clip': 30.046415328979492, 'timing/data': 0.0015509439981542528, 'timing/model': 2.0233585899986792, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.42}) +[2026-06-23 05:27:59,042][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.18198083341121674, 'train/grad_norm_pre_clip': 29.89368438720703, 'timing/data': 0.0012416329991538078, 'timing/model': 2.0344013040012214, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.42}) +[2026-06-23 05:28:03,114][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.09863021969795227, 'train/grad_norm_pre_clip': 9.478002548217773, 'timing/data': 0.06671958099468611, 'timing/model': 2.029275395005243, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.42}) +[2026-06-23 05:28:07,105][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.18103691935539246, 'train/grad_norm_pre_clip': 14.908086776733398, 'timing/data': 0.0011218019935768098, 'timing/model': 2.0144225389813073, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.42}) +[2026-06-23 05:28:11,177][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.05758170410990715, 'train/grad_norm_pre_clip': 11.075272560119629, 'timing/data': 0.0011121039860881865, 'timing/model': 2.014778613985982, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.42}) +[2026-06-23 05:28:15,191][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.09527316689491272, 'train/grad_norm_pre_clip': 14.22757339477539, 'timing/data': 0.0012486259802244604, 'timing/model': 2.035638380009914, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.42}) +[2026-06-23 05:28:19,259][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.1289391666650772, 'train/grad_norm_pre_clip': 12.70150375366211, 'timing/data': 0.0647605849953834, 'timing/model': 2.015074715978699, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.42}) +[2026-06-23 05:28:23,250][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.14830005168914795, 'train/grad_norm_pre_clip': 41.0002555847168, 'timing/data': 0.0011203400208614767, 'timing/model': 2.015464002994122, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.42}) +[2026-06-23 05:28:27,332][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.13944315910339355, 'train/grad_norm_pre_clip': 13.508707046508789, 'timing/data': 0.001114569022320211, 'timing/model': 2.018405182985589, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.42}) +[2026-06-23 05:28:31,576][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.1579749435186386, 'train/grad_norm_pre_clip': 19.3751163482666, 'timing/data': 0.0012266959820408374, 'timing/model': 2.2588781540107448, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.42}) +[2026-06-23 05:28:35,625][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.1761307567358017, 'train/grad_norm_pre_clip': 35.93722152709961, 'timing/data': 0.05275907798204571, 'timing/model': 2.020366501004901, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.42}) +[2026-06-23 05:28:39,624][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.10660909861326218, 'train/grad_norm_pre_clip': 14.42566204071045, 'timing/data': 0.0016156639903783798, 'timing/model': 2.0211522390018217, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.42}) +[2026-06-23 05:28:43,708][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.10215537995100021, 'train/grad_norm_pre_clip': 17.28548240661621, 'timing/data': 0.0012267960119061172, 'timing/model': 2.0195536340179387, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.42}) +[2026-06-23 05:28:47,707][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.21087662875652313, 'train/grad_norm_pre_clip': 28.893417358398438, 'timing/data': 0.0012388780014589429, 'timing/model': 2.0221947789832484, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.42}) +[2026-06-23 05:28:51,771][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.08731906861066818, 'train/grad_norm_pre_clip': 18.687484741210938, 'timing/data': 0.06433358701178804, 'timing/model': 2.0180951640068088, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.42}) +[2026-06-23 05:28:55,764][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.1457197666168213, 'train/grad_norm_pre_clip': 24.531949996948242, 'timing/data': 0.0012323260016273707, 'timing/model': 2.018537447991548, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.42}) +[2026-06-23 05:28:59,846][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.03770904615521431, 'train/grad_norm_pre_clip': 7.582152366638184, 'timing/data': 0.0015382819983642548, 'timing/model': 2.0209119810024276, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.42}) +[2026-06-23 05:29:04,122][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.03507417067885399, 'train/grad_norm_pre_clip': 15.411311149597168, 'timing/data': 0.0014598759880755097, 'timing/model': 2.2990339289826807, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.42}) +[2026-06-23 05:29:08,172][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.21774813532829285, 'train/grad_norm_pre_clip': 23.231449127197266, 'timing/data': 0.05609666497912258, 'timing/model': 2.0155319289769977, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.42}) +[2026-06-23 05:29:12,154][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.16395889222621918, 'train/grad_norm_pre_clip': 43.038204193115234, 'timing/data': 0.0011340639903210104, 'timing/model': 2.0138589550042525, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.42}) +[2026-06-23 05:29:16,265][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.06199472397565842, 'train/grad_norm_pre_clip': 18.80340003967285, 'timing/data': 0.0011374320019967854, 'timing/model': 2.013889548019506, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.42}) +[2026-06-23 05:29:20,248][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.1970716118812561, 'train/grad_norm_pre_clip': 36.85060119628906, 'timing/data': 0.0011406060075387359, 'timing/model': 2.0112522130075376, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.42}) +[2026-06-23 05:29:24,288][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.18475334346294403, 'train/grad_norm_pre_clip': 23.27151870727539, 'timing/data': 0.055722183984471485, 'timing/model': 2.014396988990484, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.42}) +[2026-06-23 05:29:28,276][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.136685311794281, 'train/grad_norm_pre_clip': 20.607398986816406, 'timing/data': 0.00123664501006715, 'timing/model': 2.014279627997894, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.42}) +[2026-06-23 05:29:32,361][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.15888822078704834, 'train/grad_norm_pre_clip': 41.14186096191406, 'timing/data': 0.0012701969826593995, 'timing/model': 2.0195590869989246, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.42}) +[2026-06-23 05:29:36,472][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.14505712687969208, 'train/grad_norm_pre_clip': 88.34942626953125, 'timing/data': 0.0012632030120585114, 'timing/model': 2.124171342002228, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.43}) +[2026-06-23 05:29:40,536][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1100 +[2026-06-23 05:31:35,834][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1100 +[2026-06-23 05:31:35,834][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1100 +[2026-06-23 05:33:43,191][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1100 +[2026-06-23 05:33:43,193][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.23920820653438568, 'train/grad_norm_pre_clip': 28.88243865966797, 'eval/action_loss/seconds': 115.29426363899256, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07885781675577164, 'eval/latency_2/loss': 0.07885781675577164, 'eval/flappy/loss': 0.07885781675577164, 'eval/flappy/latency_2/loss': 0.07885781675577164, 'eval/flap_precision': 0.8831488314883149, 'eval/flap_recall': 0.9689608636977058, 'eval/flap_f1': 0.9240669240669241, 'eval/flap_rate_model': 0.22583333333333333, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9240669240669241, 'eval/latency_2/flap_precision': 0.8831488314883149, 'eval/latency_2/flap_recall': 0.9689608636977058, 'eval/latency_2/flap_f1': 0.9240669240669241, 'eval/latency_2/flap_rate_model': 0.22583333333333333, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9240669240669241, 'eval/action_classification/data_wait_seconds': 50.699092864990234, 'eval/action_classification/predict_seconds': 74.94705200195312, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06974495502072386, 'timing/model': 2.0201501620176714, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.43}) +[2026-06-23 05:33:47,157][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.14423643052577972, 'train/grad_norm_pre_clip': 19.758188247680664, 'timing/data': 0.0013571669987868518, 'timing/model': 1.987965941982111, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.43}) +[2026-06-23 05:33:51,188][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.1711866855621338, 'train/grad_norm_pre_clip': 25.334800720214844, 'timing/data': 0.0011280840262770653, 'timing/model': 1.9906900270143524, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.43}) +[2026-06-23 05:33:55,129][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.17151615023612976, 'train/grad_norm_pre_clip': 18.504161834716797, 'timing/data': 0.0011706119985319674, 'timing/model': 1.9949260569992475, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.43}) +[2026-06-23 05:33:59,136][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.050338469445705414, 'train/grad_norm_pre_clip': 11.413060188293457, 'timing/data': 0.06859248797991313, 'timing/model': 1.9893035040004179, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.43}) +[2026-06-23 05:34:03,325][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.10332921147346497, 'train/grad_norm_pre_clip': 26.092439651489258, 'timing/data': 0.0011021359823644161, 'timing/model': 2.2406188719905913, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.43}) +[2026-06-23 05:34:07,361][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.0667576864361763, 'train/grad_norm_pre_clip': 19.8231143951416, 'timing/data': 0.0015080159937497228, 'timing/model': 1.9954391450155526, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.43}) +[2026-06-23 05:34:11,318][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.08060014992952347, 'train/grad_norm_pre_clip': 11.97169303894043, 'timing/data': 0.0011612550006248057, 'timing/model': 2.005351931002224, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.43}) +[2026-06-23 05:34:15,340][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.07972866296768188, 'train/grad_norm_pre_clip': 21.872220993041992, 'timing/data': 0.05717298400122672, 'timing/model': 2.008877707994543, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.43}) +[2026-06-23 05:34:19,322][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.16502588987350464, 'train/grad_norm_pre_clip': 23.837385177612305, 'timing/data': 0.0012501800083555281, 'timing/model': 2.0205694799951743, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.43}) +[2026-06-23 05:34:23,377][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.07841988652944565, 'train/grad_norm_pre_clip': 12.88257122039795, 'timing/data': 0.0011997960100416094, 'timing/model': 2.0023484340053983, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.43}) +[2026-06-23 05:34:27,349][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.08816594630479813, 'train/grad_norm_pre_clip': 38.54759216308594, 'timing/data': 0.0013186460128054023, 'timing/model': 2.010095798992552, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.43}) +[2026-06-23 05:34:31,380][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.19112378358840942, 'train/grad_norm_pre_clip': 10.119766235351562, 'timing/data': 0.06465411401586607, 'timing/model': 2.0025564350071363, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.43}) +[2026-06-23 05:34:35,417][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.12113580107688904, 'train/grad_norm_pre_clip': 18.892356872558594, 'timing/data': 0.0012286099954508245, 'timing/model': 2.0773194060020614, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.43}) +[2026-06-23 05:34:39,521][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.1305336207151413, 'train/grad_norm_pre_clip': 47.10298156738281, 'timing/data': 0.0012495790142565966, 'timing/model': 2.0222158490214497, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.43}) +[2026-06-23 05:34:43,500][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.107141874730587, 'train/grad_norm_pre_clip': 16.956140518188477, 'timing/data': 0.0012586250086314976, 'timing/model': 2.010957866004901, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.43}) +[2026-06-23 05:34:47,538][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.16500361263751984, 'train/grad_norm_pre_clip': 47.079566955566406, 'timing/data': 0.056106353003997356, 'timing/model': 2.014453006006079, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.43}) +[2026-06-23 05:34:51,525][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.20443515479564667, 'train/grad_norm_pre_clip': 53.433753967285156, 'timing/data': 0.001276477996725589, 'timing/model': 2.015798904991243, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.43}) +[2026-06-23 05:34:55,616][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.0534810945391655, 'train/grad_norm_pre_clip': 14.301048278808594, 'timing/data': 0.0011439929949119687, 'timing/model': 2.0232768139976542, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.43}) +[2026-06-23 05:34:59,620][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.1437372863292694, 'train/grad_norm_pre_clip': 27.75018882751465, 'timing/data': 0.0011604739993344992, 'timing/model': 2.036761014984222, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.43}) +[2026-06-23 05:35:03,657][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.18956834077835083, 'train/grad_norm_pre_clip': 22.379121780395508, 'timing/data': 0.05882461598957889, 'timing/model': 2.009242724016076, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.43}) +[2026-06-23 05:35:07,630][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.17541471123695374, 'train/grad_norm_pre_clip': 39.67398452758789, 'timing/data': 0.0014156149991322309, 'timing/model': 2.0079232200223487, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.43}) +[2026-06-23 05:35:11,963][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.10049371421337128, 'train/grad_norm_pre_clip': 14.742199897766113, 'timing/data': 0.0012755859934259206, 'timing/model': 2.0112099050020333, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.43}) +[2026-06-23 05:35:15,941][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.18617700040340424, 'train/grad_norm_pre_clip': 27.478132247924805, 'timing/data': 0.0014719690079800785, 'timing/model': 2.0092502089974005, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.43}) +[2026-06-23 05:35:19,975][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.08731779456138611, 'train/grad_norm_pre_clip': 19.238407135009766, 'timing/data': 0.059030807024100795, 'timing/model': 2.0102111370069906, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.44}) +[2026-06-23 05:35:23,977][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.05507044121623039, 'train/grad_norm_pre_clip': 23.790822982788086, 'timing/data': 0.001209264009958133, 'timing/model': 2.015980808995664, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.44}) +[2026-06-23 05:35:28,054][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.15053386986255646, 'train/grad_norm_pre_clip': 30.535348892211914, 'timing/data': 0.0013914710143581033, 'timing/model': 2.0144511369871907, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.44}) +[2026-06-23 05:35:32,057][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.17080749571323395, 'train/grad_norm_pre_clip': 22.022119522094727, 'timing/data': 0.0013608939771074802, 'timing/model': 2.020792805007659, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.44}) +[2026-06-23 05:35:36,116][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.04551239684224129, 'train/grad_norm_pre_clip': 12.40095329284668, 'timing/data': 0.07008934597251937, 'timing/model': 2.016622194001684, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.44}) +[2026-06-23 05:35:40,112][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.08516416698694229, 'train/grad_norm_pre_clip': 26.415863037109375, 'timing/data': 0.0012310640013311058, 'timing/model': 2.0184115819865838, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.44}) +[2026-06-23 05:35:44,234][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.04890253767371178, 'train/grad_norm_pre_clip': 16.975353240966797, 'timing/data': 0.0012266149860806763, 'timing/model': 2.0305456510104705, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.44}) +[2026-06-23 05:35:48,231][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.1629316210746765, 'train/grad_norm_pre_clip': 20.876113891601562, 'timing/data': 0.0012865870085079223, 'timing/model': 2.0197281909931917, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.44}) +[2026-06-23 05:35:52,283][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.1843196451663971, 'train/grad_norm_pre_clip': 52.5232048034668, 'timing/data': 0.05929294100496918, 'timing/model': 2.019428099010838, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.44}) +[2026-06-23 05:35:56,276][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.0973271057009697, 'train/grad_norm_pre_clip': 36.46320724487305, 'timing/data': 0.00136032301816158, 'timing/model': 2.0168269480054732, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.44}) +[2026-06-23 05:36:00,363][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.19738517701625824, 'train/grad_norm_pre_clip': 45.49915313720703, 'timing/data': 0.0012106659996788949, 'timing/model': 2.0212604099942837, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.44}) +[2026-06-23 05:36:04,367][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.09004580974578857, 'train/grad_norm_pre_clip': 22.043224334716797, 'timing/data': 0.0014299009926617146, 'timing/model': 2.0299594230018556, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.44}) +[2026-06-23 05:36:08,423][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.13407954573631287, 'train/grad_norm_pre_clip': 28.528669357299805, 'timing/data': 0.05989756900817156, 'timing/model': 2.015903419989627, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.44}) +[2026-06-23 05:36:12,430][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.08403123915195465, 'train/grad_norm_pre_clip': 36.93824768066406, 'timing/data': 0.0016071490244939923, 'timing/model': 2.0234592519991565, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.44}) +[2026-06-23 05:36:16,610][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.29249733686447144, 'train/grad_norm_pre_clip': 76.04061126708984, 'timing/data': 0.001302325981669128, 'timing/model': 2.1139739240170456, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.44}) +[2026-06-23 05:36:20,613][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.3166825473308563, 'train/grad_norm_pre_clip': 60.686737060546875, 'timing/data': 0.0014998609840404242, 'timing/model': 2.025969193986384, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.44}) +[2026-06-23 05:36:24,667][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.09721370041370392, 'train/grad_norm_pre_clip': 22.021934509277344, 'timing/data': 0.05415758700110018, 'timing/model': 2.0215871109976433, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.44}) +[2026-06-23 05:36:28,662][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.09010743349790573, 'train/grad_norm_pre_clip': 20.953086853027344, 'timing/data': 0.0011981729767285287, 'timing/model': 2.0184924280038103, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.44}) +[2026-06-23 05:36:32,752][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.04080106317996979, 'train/grad_norm_pre_clip': 16.54181671142578, 'timing/data': 0.001174920005723834, 'timing/model': 2.025262408016715, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.44}) +[2026-06-23 05:36:36,751][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.20724895596504211, 'train/grad_norm_pre_clip': 37.60000228881836, 'timing/data': 0.0013554140168707818, 'timing/model': 2.022678075998556, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.44}) +[2026-06-23 05:36:40,844][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.09845104813575745, 'train/grad_norm_pre_clip': 19.85130500793457, 'timing/data': 0.06574776698835194, 'timing/model': 2.0189307459804695, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.44}) +[2026-06-23 05:36:44,857][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.22818653285503387, 'train/grad_norm_pre_clip': 34.456756591796875, 'timing/data': 0.0022092809958849102, 'timing/model': 2.031544671015581, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.44}) +[2026-06-23 05:36:48,965][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.16406410932540894, 'train/grad_norm_pre_clip': 26.301860809326172, 'timing/data': 0.001211347000207752, 'timing/model': 2.03920709900558, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.44}) +[2026-06-23 05:36:53,175][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.3756633400917053, 'train/grad_norm_pre_clip': 35.90774154663086, 'timing/data': 0.0012438679987099022, 'timing/model': 2.0231247120245825, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.44}) +[2026-06-23 05:36:57,245][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.12133315950632095, 'train/grad_norm_pre_clip': 36.39010238647461, 'timing/data': 0.05590061700786464, 'timing/model': 2.0318799689994194, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.44}) +[2026-06-23 05:37:01,256][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.06374411284923553, 'train/grad_norm_pre_clip': 19.769941329956055, 'timing/data': 0.0012129300157539546, 'timing/model': 2.0350220759864897, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.44}) +[2026-06-23 05:37:05,346][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.18105021119117737, 'train/grad_norm_pre_clip': 22.587331771850586, 'timing/data': 0.0012590159894898534, 'timing/model': 2.021489960985491, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.45}) +[2026-06-23 05:37:09,374][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.11052049696445465, 'train/grad_norm_pre_clip': 16.147476196289062, 'timing/data': 0.001512134011136368, 'timing/model': 2.0449544100265484, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.45}) +[2026-06-23 05:37:13,443][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.1668853908777237, 'train/grad_norm_pre_clip': 23.781524658203125, 'timing/data': 0.0645947179873474, 'timing/model': 2.025764212012291, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.45}) +[2026-06-23 05:37:17,471][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.13455267250537872, 'train/grad_norm_pre_clip': 29.902481079101562, 'timing/data': 0.0013855190190952271, 'timing/model': 2.0429659119981807, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.45}) +[2026-06-23 05:37:21,560][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.09677170217037201, 'train/grad_norm_pre_clip': 14.780838012695312, 'timing/data': 0.0012442790030036122, 'timing/model': 2.021884660993237, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.45}) +[2026-06-23 05:37:25,795][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.1536525934934616, 'train/grad_norm_pre_clip': 25.621278762817383, 'timing/data': 0.0013087179977446795, 'timing/model': 2.2580283309798688, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.45}) +[2026-06-23 05:37:29,859][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.20712991058826447, 'train/grad_norm_pre_clip': 21.2740421295166, 'timing/data': 0.06278116902103648, 'timing/model': 2.0226257649774197, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.45}) +[2026-06-23 05:37:33,873][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.18752118945121765, 'train/grad_norm_pre_clip': 20.699886322021484, 'timing/data': 0.001509328983956948, 'timing/model': 2.0354835299949627, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.45}) +[2026-06-23 05:37:37,959][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.11774718016386032, 'train/grad_norm_pre_clip': 15.148601531982422, 'timing/data': 0.0010339900036342442, 'timing/model': 2.021799865004141, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.45}) +[2026-06-23 05:37:41,966][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.11700335890054703, 'train/grad_norm_pre_clip': 6.408709526062012, 'timing/data': 0.0015027059998828918, 'timing/model': 2.0254399020050187, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.45}) +[2026-06-23 05:37:46,025][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.08065848052501678, 'train/grad_norm_pre_clip': 21.56161117553711, 'timing/data': 0.05979963002027944, 'timing/model': 2.017881320003653, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.45}) +[2026-06-23 05:37:50,040][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.16030333936214447, 'train/grad_norm_pre_clip': 45.147010803222656, 'timing/data': 0.0014468530134763569, 'timing/model': 2.039826033025747, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.45}) +[2026-06-23 05:37:54,146][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.1748427152633667, 'train/grad_norm_pre_clip': 34.290283203125, 'timing/data': 0.0014480350073426962, 'timing/model': 2.03956195700448, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.45}) +[2026-06-23 05:37:58,398][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.16339576244354248, 'train/grad_norm_pre_clip': 24.684370040893555, 'timing/data': 0.0012223180092405528, 'timing/model': 2.2775238979957066, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.45}) +[2026-06-23 05:38:02,466][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.17916379868984222, 'train/grad_norm_pre_clip': 16.150920867919922, 'timing/data': 0.06271452599321492, 'timing/model': 2.019614282005932, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.45}) +[2026-06-23 05:38:06,469][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.1445322185754776, 'train/grad_norm_pre_clip': 31.710308074951172, 'timing/data': 0.0012646059913095087, 'timing/model': 2.0253177910053637, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.45}) +[2026-06-23 05:38:10,560][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.25176575779914856, 'train/grad_norm_pre_clip': 61.59886932373047, 'timing/data': 0.0013306580076459795, 'timing/model': 2.018592520005768, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.45}) +[2026-06-23 05:38:14,552][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.17294622957706451, 'train/grad_norm_pre_clip': 27.189416885375977, 'timing/data': 0.0012780610122717917, 'timing/model': 2.0170691880048253, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.45}) +[2026-06-23 05:38:18,605][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.13925927877426147, 'train/grad_norm_pre_clip': 21.90593719482422, 'timing/data': 0.058580138022080064, 'timing/model': 2.0193512059922796, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.45}) +[2026-06-23 05:38:22,605][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.19417838752269745, 'train/grad_norm_pre_clip': 30.84395980834961, 'timing/data': 0.0014379670028574765, 'timing/model': 2.0224408760259394, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.45}) +[2026-06-23 05:38:26,694][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.10531981289386749, 'train/grad_norm_pre_clip': 22.130525588989258, 'timing/data': 0.0012294609914533794, 'timing/model': 2.0204117390094325, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.45}) +[2026-06-23 05:38:30,761][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.17375025153160095, 'train/grad_norm_pre_clip': 29.035696029663086, 'timing/data': 0.0014123590080998838, 'timing/model': 2.0891967630013824, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.45}) +[2026-06-23 05:38:34,819][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.2592753767967224, 'train/grad_norm_pre_clip': 37.017578125, 'timing/data': 0.0576133519934956, 'timing/model': 2.0221781870059203, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.45}) +[2026-06-23 05:38:38,823][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.32265982031822205, 'train/grad_norm_pre_clip': 35.40272521972656, 'timing/data': 0.0012126299843657762, 'timing/model': 2.022835573006887, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.45}) +[2026-06-23 05:38:42,923][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.18030677735805511, 'train/grad_norm_pre_clip': 39.44863510131836, 'timing/data': 0.001249388005817309, 'timing/model': 2.0213604110176675, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.45}) +[2026-06-23 05:38:46,918][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.12769795954227448, 'train/grad_norm_pre_clip': 15.169371604919434, 'timing/data': 0.0012878889974672347, 'timing/model': 2.0187080030154902, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.45}) +[2026-06-23 05:38:50,981][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.08869179338216782, 'train/grad_norm_pre_clip': 11.059027671813965, 'timing/data': 0.05710200799512677, 'timing/model': 2.0258893519931007, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.46}) +[2026-06-23 05:38:54,982][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.18688490986824036, 'train/grad_norm_pre_clip': 43.937034606933594, 'timing/data': 0.0013750809885095805, 'timing/model': 2.022946957993554, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.46}) +[2026-06-23 05:38:59,079][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.21665462851524353, 'train/grad_norm_pre_clip': 51.53319549560547, 'timing/data': 0.0011319509940221906, 'timing/model': 2.018340551003348, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.46}) +[2026-06-23 05:39:03,079][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.2086435854434967, 'train/grad_norm_pre_clip': 38.777442932128906, 'timing/data': 0.0011438530054874718, 'timing/model': 2.0223990000085905, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.46}) +[2026-06-23 05:39:07,375][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.09068191796541214, 'train/grad_norm_pre_clip': 22.195276260375977, 'timing/data': 0.054310350999003276, 'timing/model': 2.2589828700001817, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.46}) +[2026-06-23 05:39:11,375][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.24664291739463806, 'train/grad_norm_pre_clip': 18.616741180419922, 'timing/data': 0.0012089230003766716, 'timing/model': 2.0208477189880796, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.46}) +[2026-06-23 05:39:15,478][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.16633272171020508, 'train/grad_norm_pre_clip': 19.546079635620117, 'timing/data': 0.001220203994307667, 'timing/model': 2.0384859559999313, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.46}) +[2026-06-23 05:39:19,489][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.09488730877637863, 'train/grad_norm_pre_clip': 12.710670471191406, 'timing/data': 0.001420663989847526, 'timing/model': 2.033803286991315, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.46}) +[2026-06-23 05:39:23,561][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.1887989193201065, 'train/grad_norm_pre_clip': 36.41659927368164, 'timing/data': 0.06842824199702591, 'timing/model': 2.024172487988835, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.46}) +[2026-06-23 05:39:27,563][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.1783038079738617, 'train/grad_norm_pre_clip': 27.445791244506836, 'timing/data': 0.0011925230210181326, 'timing/model': 2.021694972005207, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.46}) +[2026-06-23 05:39:31,661][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.22579678893089294, 'train/grad_norm_pre_clip': 40.20295333862305, 'timing/data': 0.0012144839856773615, 'timing/model': 2.025629993993789, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.46}) +[2026-06-23 05:39:35,672][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.11475452035665512, 'train/grad_norm_pre_clip': 25.24793243408203, 'timing/data': 0.0019815399718936533, 'timing/model': 2.0284189480007626, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.46}) +[2026-06-23 05:39:39,768][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.16608130931854248, 'train/grad_norm_pre_clip': 21.314151763916016, 'timing/data': 0.05563678199541755, 'timing/model': 2.0626011780113913, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.46}) +[2026-06-23 05:39:43,774][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.14202909171581268, 'train/grad_norm_pre_clip': 17.70627784729004, 'timing/data': 0.001206760003697127, 'timing/model': 2.0244399220100604, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.46}) +[2026-06-23 05:39:47,857][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.1449776440858841, 'train/grad_norm_pre_clip': 13.802592277526855, 'timing/data': 0.0012022510054521263, 'timing/model': 2.019291752978461, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.46}) +[2026-06-23 05:39:51,858][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.12493748962879181, 'train/grad_norm_pre_clip': 26.626054763793945, 'timing/data': 0.001405165996402502, 'timing/model': 2.0207082419947255, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.46}) +[2026-06-23 05:39:55,919][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.11324124038219452, 'train/grad_norm_pre_clip': 28.1142520904541, 'timing/data': 0.058645088982302696, 'timing/model': 2.0215018799935933, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.46}) +[2026-06-23 05:39:59,920][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.12481345981359482, 'train/grad_norm_pre_clip': 15.224563598632812, 'timing/data': 0.0013319410209078342, 'timing/model': 2.0219510319875553, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.46}) +[2026-06-23 05:40:04,017][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.09697172045707703, 'train/grad_norm_pre_clip': 15.995610237121582, 'timing/data': 0.001248827000381425, 'timing/model': 2.026380457013147, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.46}) +[2026-06-23 05:40:08,024][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.08743107318878174, 'train/grad_norm_pre_clip': 18.144678115844727, 'timing/data': 0.0015184649964794517, 'timing/model': 2.0236442830064334, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.46}) +[2026-06-23 05:40:12,085][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.2378368079662323, 'train/grad_norm_pre_clip': 51.036415100097656, 'timing/data': 0.055751645006239414, 'timing/model': 2.024336272996152, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.46}) +[2026-06-23 05:40:16,264][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.17145954072475433, 'train/grad_norm_pre_clip': 29.942834854125977, 'timing/data': 0.0016228679742198437, 'timing/model': 2.0505272400041576, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.46}) +[2026-06-23 05:40:20,356][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.128134623169899, 'train/grad_norm_pre_clip': 26.000839233398438, 'timing/data': 0.0012522039760369807, 'timing/model': 2.02260345700779, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.46}) +[2026-06-23 05:40:24,371][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.16158579289913177, 'train/grad_norm_pre_clip': 20.358999252319336, 'timing/data': 0.0012535550049506128, 'timing/model': 2.0324400499812327, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.46}) +[2026-06-23 05:40:28,429][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1200 +[2026-06-23 05:42:22,840][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1200 +[2026-06-23 05:42:22,841][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1200 +[2026-06-23 05:44:29,941][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1200 +[2026-06-23 05:44:29,942][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.13267727196216583, 'train/grad_norm_pre_clip': 12.940722465515137, 'eval/action_loss/seconds': 114.4077372669999, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07388824224472046, 'eval/latency_2/loss': 0.07388824224472046, 'eval/flappy/loss': 0.07388824224472046, 'eval/flappy/latency_2/loss': 0.07388824224472046, 'eval/flap_precision': 0.9128137384412153, 'eval/flap_recall': 0.9325236167341431, 'eval/flap_f1': 0.9225634178905208, 'eval/flap_rate_model': 0.21027777777777779, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9225634178905208, 'eval/latency_2/flap_precision': 0.9128137384412153, 'eval/latency_2/flap_recall': 0.9325236167341431, 'eval/latency_2/flap_f1': 0.9225634178905208, 'eval/latency_2/flap_rate_model': 0.21027777777777779, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9225634178905208, 'eval/action_classification/data_wait_seconds': 50.51486587524414, 'eval/action_classification/predict_seconds': 74.8690185546875, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.05739042200730182, 'timing/model': 2.021196460002102, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.46}) +[2026-06-23 05:44:33,890][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.15040770173072815, 'train/grad_norm_pre_clip': 34.9253044128418, 'timing/data': 0.0013827040093019605, 'timing/model': 1.9935092690284364, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.46}) +[2026-06-23 05:44:37,908][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.2173655927181244, 'train/grad_norm_pre_clip': 51.18627166748047, 'timing/data': 0.0011107119789812714, 'timing/model': 1.9841690140019637, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.47}) +[2026-06-23 05:44:42,098][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.16429758071899414, 'train/grad_norm_pre_clip': 25.44074821472168, 'timing/data': 0.0011676980066113174, 'timing/model': 2.005858705000719, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.47}) +[2026-06-23 05:44:46,117][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.12996527552604675, 'train/grad_norm_pre_clip': 33.52518844604492, 'timing/data': 0.07291814099880867, 'timing/model': 1.996020160993794, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.47}) +[2026-06-23 05:44:50,069][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.19204656779766083, 'train/grad_norm_pre_clip': 49.12283706665039, 'timing/data': 0.0013229940086603165, 'timing/model': 1.9976105489768088, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.47}) +[2026-06-23 05:44:54,142][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.16959483921527863, 'train/grad_norm_pre_clip': 18.305320739746094, 'timing/data': 0.0014015290071256459, 'timing/model': 2.028562902996782, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.47}) +[2026-06-23 05:44:58,101][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.13330717384815216, 'train/grad_norm_pre_clip': 20.474903106689453, 'timing/data': 0.001453324977774173, 'timing/model': 1.9994331389898434, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.47}) +[2026-06-23 05:45:02,128][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.09804603457450867, 'train/grad_norm_pre_clip': 11.87187671661377, 'timing/data': 0.0730990849842783, 'timing/model': 1.9985385769978166, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.47}) +[2026-06-23 05:45:06,100][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.21892568469047546, 'train/grad_norm_pre_clip': 31.952951431274414, 'timing/data': 0.0014753059949725866, 'timing/model': 2.0169448730011936, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.47}) +[2026-06-23 05:45:10,147][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.21647673845291138, 'train/grad_norm_pre_clip': 33.78786849975586, 'timing/data': 0.001243336999323219, 'timing/model': 1.999525232007727, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.47}) +[2026-06-23 05:45:14,337][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.19653813540935516, 'train/grad_norm_pre_clip': 29.518850326538086, 'timing/data': 0.0011956790112890303, 'timing/model': 1.9991223190154415, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.47}) +[2026-06-23 05:45:18,374][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.15427477657794952, 'train/grad_norm_pre_clip': 23.66679573059082, 'timing/data': 0.07321468900772743, 'timing/model': 2.0057444729900453, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.47}) +[2026-06-23 05:45:22,358][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.12812335789203644, 'train/grad_norm_pre_clip': 30.691909790039062, 'timing/data': 0.0012543580087367445, 'timing/model': 2.0210167289769743, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.47}) +[2026-06-23 05:45:26,423][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.3046334981918335, 'train/grad_norm_pre_clip': 24.566667556762695, 'timing/data': 0.0012375969963613898, 'timing/model': 2.0095147810061462, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.47}) +[2026-06-23 05:45:30,401][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.12058623880147934, 'train/grad_norm_pre_clip': 24.12900161743164, 'timing/data': 0.0012539870222099125, 'timing/model': 2.009993176005082, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.47}) +[2026-06-23 05:45:34,430][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.08955024182796478, 'train/grad_norm_pre_clip': 13.68699836730957, 'timing/data': 0.05516833998262882, 'timing/model': 2.009543827007292, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.47}) +[2026-06-23 05:45:38,410][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.13042865693569183, 'train/grad_norm_pre_clip': 14.239523887634277, 'timing/data': 0.0012114389974158257, 'timing/model': 2.011564690998057, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.47}) +[2026-06-23 05:45:42,476][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.11569184809923172, 'train/grad_norm_pre_clip': 16.05744743347168, 'timing/data': 0.0012238509953022003, 'timing/model': 2.0107709860021714, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.47}) +[2026-06-23 05:45:46,502][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.08451710641384125, 'train/grad_norm_pre_clip': 17.468029022216797, 'timing/data': 0.0013689579791389406, 'timing/model': 2.021690118999686, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.47}) +[2026-06-23 05:45:50,529][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.10424209386110306, 'train/grad_norm_pre_clip': 17.888626098632812, 'timing/data': 0.055930959002580494, 'timing/model': 2.0062522990047, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.47}) +[2026-06-23 05:45:54,507][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.16205182671546936, 'train/grad_norm_pre_clip': 25.144609451293945, 'timing/data': 0.0013001220067963004, 'timing/model': 2.011808937008027, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.47}) +[2026-06-23 05:45:58,588][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.17554360628128052, 'train/grad_norm_pre_clip': 20.88311767578125, 'timing/data': 0.0013828350056428462, 'timing/model': 2.02017568401061, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.47}) +[2026-06-23 05:46:02,577][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.0362585112452507, 'train/grad_norm_pre_clip': 10.231135368347168, 'timing/data': 0.0013700209965463728, 'timing/model': 2.0155027720029466, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.47}) +[2026-06-23 05:46:06,628][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.09321087598800659, 'train/grad_norm_pre_clip': 21.55429458618164, 'timing/data': 0.059757417999207973, 'timing/model': 2.0188982260006014, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.47}) +[2026-06-23 05:46:10,615][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.37652379274368286, 'train/grad_norm_pre_clip': 41.479129791259766, 'timing/data': 0.0012196839961688966, 'timing/model': 2.014536358998157, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.47}) +[2026-06-23 05:46:14,690][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.08031106740236282, 'train/grad_norm_pre_clip': 22.542993545532227, 'timing/data': 0.0012763979902956635, 'timing/model': 2.0125711000000592, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.47}) +[2026-06-23 05:46:18,673][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.05967375636100769, 'train/grad_norm_pre_clip': 9.717400550842285, 'timing/data': 0.001245200983248651, 'timing/model': 2.013387006998528, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.47}) +[2026-06-23 05:46:22,899][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.12393610924482346, 'train/grad_norm_pre_clip': 23.870861053466797, 'timing/data': 0.05181486500077881, 'timing/model': 2.0137960739957634, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.48}) +[2026-06-23 05:46:26,886][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.0491509884595871, 'train/grad_norm_pre_clip': 14.982467651367188, 'timing/data': 0.0012013299856334925, 'timing/model': 2.017525824980112, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.48}) +[2026-06-23 05:46:30,956][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.15464118123054504, 'train/grad_norm_pre_clip': 20.348743438720703, 'timing/data': 0.0013072250003460795, 'timing/model': 2.013698155002203, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.48}) +[2026-06-23 05:46:34,944][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.10228326171636581, 'train/grad_norm_pre_clip': 17.979875564575195, 'timing/data': 0.0012423350126482546, 'timing/model': 2.016864595003426, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.48}) +[2026-06-23 05:46:39,000][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.1353461742401123, 'train/grad_norm_pre_clip': 23.017332077026367, 'timing/data': 0.06459131900919601, 'timing/model': 2.0181391109945253, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.48}) +[2026-06-23 05:46:43,013][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.07603022456169128, 'train/grad_norm_pre_clip': 13.678092002868652, 'timing/data': 0.0012876690016128123, 'timing/model': 2.037278450996382, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.48}) +[2026-06-23 05:46:47,119][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.1546511948108673, 'train/grad_norm_pre_clip': 30.97972297668457, 'timing/data': 0.0013925429957453161, 'timing/model': 2.0219100100221112, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.48}) +[2026-06-23 05:46:51,138][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.2644355893135071, 'train/grad_norm_pre_clip': 55.57307434082031, 'timing/data': 0.001247886015335098, 'timing/model': 2.0416951430088375, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.48}) +[2026-06-23 05:46:55,238][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.09840893000364304, 'train/grad_norm_pre_clip': 29.152647018432617, 'timing/data': 0.06902565201744437, 'timing/model': 2.0202062430034857, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.48}) +[2026-06-23 05:46:59,240][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.08709754049777985, 'train/grad_norm_pre_clip': 10.811872482299805, 'timing/data': 0.001160273008281365, 'timing/model': 2.021820377005497, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.48}) +[2026-06-23 05:47:03,344][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.10427916049957275, 'train/grad_norm_pre_clip': 19.602622985839844, 'timing/data': 0.0011325730010867119, 'timing/model': 2.039238927012775, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.48}) +[2026-06-23 05:47:07,340][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.09837619215250015, 'train/grad_norm_pre_clip': 19.910982131958008, 'timing/data': 0.0011792289733421057, 'timing/model': 2.0175094039877877, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.48}) +[2026-06-23 05:47:11,458][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.11824674159288406, 'train/grad_norm_pre_clip': 21.03996467590332, 'timing/data': 0.052251462009735405, 'timing/model': 2.022175424004672, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.48}) +[2026-06-23 05:47:15,511][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.19000652432441711, 'train/grad_norm_pre_clip': 46.3375129699707, 'timing/data': 0.0011202089954167604, 'timing/model': 2.017685940983938, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.48}) +[2026-06-23 05:47:19,615][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.32600682973861694, 'train/grad_norm_pre_clip': 36.968753814697266, 'timing/data': 0.0011308390239719301, 'timing/model': 2.014063490991248, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.48}) +[2026-06-23 05:47:23,603][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.023946816101670265, 'train/grad_norm_pre_clip': 6.512008190155029, 'timing/data': 0.001469494978664443, 'timing/model': 2.0162997730076313, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.48}) +[2026-06-23 05:47:27,908][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.136818528175354, 'train/grad_norm_pre_clip': 18.268117904663086, 'timing/data': 0.06625882000662386, 'timing/model': 2.265999961993657, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.48}) +[2026-06-23 05:47:31,899][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.07100430130958557, 'train/grad_norm_pre_clip': 13.971746444702148, 'timing/data': 0.0011534309887792915, 'timing/model': 2.017380309989676, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.48}) +[2026-06-23 05:47:35,986][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.35237741470336914, 'train/grad_norm_pre_clip': 53.457881927490234, 'timing/data': 0.0011383540113456547, 'timing/model': 2.019688063999638, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.48}) +[2026-06-23 05:47:39,982][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.17187675833702087, 'train/grad_norm_pre_clip': 56.15410614013672, 'timing/data': 0.0012045149924233556, 'timing/model': 2.0205423140141647, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.48}) +[2026-06-23 05:47:44,057][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.24685576558113098, 'train/grad_norm_pre_clip': 69.76716613769531, 'timing/data': 0.06972132599912584, 'timing/model': 2.025792464002734, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.48}) +[2026-06-23 05:47:48,060][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.13463367521762848, 'train/grad_norm_pre_clip': 37.34135055541992, 'timing/data': 0.0012169289984740317, 'timing/model': 2.0212651309848297, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.48}) +[2026-06-23 06:04:17,255][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1250 +[2026-06-23 06:04:17,257][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.12088827043771744, 'train/grad_norm_pre_clip': 38.05501937866211, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 255.94000000000815, 'rl_games_eval/mid_train/mean_length': 2080.0, 'rl_games_eval/mid_train/std_reward': 168.94732433514338, 'rl_games_eval/mid_train/std_length': 1361.7158293858524, 'rl_games_eval/mid_train/macro_mean_reward': 255.94000000000815, 'rl_games_eval/mid_train/macro_mean_length': 2080.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 255.94000000000815, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2080.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 168.94732433514338, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1361.7158293858524, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 255.94000000000815, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.001354762993287295, 'timing/model': 2.0220505019824486, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.48}) +[2026-06-23 06:04:21,324][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.16883134841918945, 'train/grad_norm_pre_clip': 14.036345481872559, 'timing/data': 0.0014859459770377725, 'timing/model': 2.156570582999848, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.48}) +[2026-06-23 06:04:25,248][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.18182821571826935, 'train/grad_norm_pre_clip': 34.038509368896484, 'timing/data': 0.05464394501177594, 'timing/model': 1.9548434130265377, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.48}) +[2026-06-23 06:04:29,123][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.16047760844230652, 'train/grad_norm_pre_clip': 35.57059860229492, 'timing/data': 0.0012272280000615865, 'timing/model': 1.960081817000173, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.48}) +[2026-06-23 06:04:33,101][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.12946467101573944, 'train/grad_norm_pre_clip': 30.46852684020996, 'timing/data': 0.0011904200073331594, 'timing/model': 1.9661653639923315, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.49}) +[2026-06-23 06:04:37,005][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.13062259554862976, 'train/grad_norm_pre_clip': 17.90920066833496, 'timing/data': 0.0014305940130725503, 'timing/model': 1.9814095899928361, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.49}) +[2026-06-23 06:04:40,959][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.11189372837543488, 'train/grad_norm_pre_clip': 13.434793472290039, 'timing/data': 0.05579171600402333, 'timing/model': 1.9704083020042162, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.49}) +[2026-06-23 06:04:44,883][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.06159209460020065, 'train/grad_norm_pre_clip': 9.133098602294922, 'timing/data': 0.0012313750048633665, 'timing/model': 1.9907694940047804, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.49}) +[2026-06-23 06:04:48,910][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.1498817354440689, 'train/grad_norm_pre_clip': 19.843429565429688, 'timing/data': 0.0013533119927160442, 'timing/model': 1.993997652985854, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.49}) +[2026-06-23 06:04:52,869][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.10851163417100906, 'train/grad_norm_pre_clip': 23.953506469726562, 'timing/data': 0.0012897140113636851, 'timing/model': 2.021694461989682, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.49}) +[2026-06-23 06:04:56,843][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.07290205359458923, 'train/grad_norm_pre_clip': 13.225351333618164, 'timing/data': 0.054171871015569195, 'timing/model': 1.982720520987641, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.49}) +[2026-06-23 06:05:00,770][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.10167749226093292, 'train/grad_norm_pre_clip': 17.023479461669922, 'timing/data': 0.001201099978061393, 'timing/model': 1.9859105600044131, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.49}) +[2026-06-23 06:05:04,769][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.1968507021665573, 'train/grad_norm_pre_clip': 27.784029006958008, 'timing/data': 0.0011236659775022417, 'timing/model': 1.9781415459874552, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.49}) +[2026-06-23 06:05:08,721][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.19344739615917206, 'train/grad_norm_pre_clip': 17.050928115844727, 'timing/data': 0.0017604230088181794, 'timing/model': 2.0113777209771797, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.49}) +[2026-06-23 06:05:12,718][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.14321905374526978, 'train/grad_norm_pre_clip': 23.477405548095703, 'timing/data': 0.05527411698130891, 'timing/model': 1.9918107150006108, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.49}) +[2026-06-23 06:05:16,648][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.11263161897659302, 'train/grad_norm_pre_clip': 22.29404640197754, 'timing/data': 0.0013488020049408078, 'timing/model': 1.985674098978052, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.49}) +[2026-06-23 06:05:20,676][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.2246393859386444, 'train/grad_norm_pre_clip': 30.449861526489258, 'timing/data': 0.0014259550080168992, 'timing/model': 1.9936840340087656, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.49}) +[2026-06-23 06:05:24,614][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.11340897530317307, 'train/grad_norm_pre_clip': 29.06682777404785, 'timing/data': 0.0012456020049285144, 'timing/model': 1.9859556469891686, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.49}) +[2026-06-23 06:05:28,701][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.0852019265294075, 'train/grad_norm_pre_clip': 14.708020210266113, 'timing/data': 0.06467578999581747, 'timing/model': 1.9885310199751984, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.49}) +[2026-06-23 06:05:32,640][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.23740707337856293, 'train/grad_norm_pre_clip': 29.071191787719727, 'timing/data': 0.0011054520145989954, 'timing/model': 1.989200140000321, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.49}) +[2026-06-23 06:05:36,685][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.15975457429885864, 'train/grad_norm_pre_clip': 27.615196228027344, 'timing/data': 0.0012020319991279393, 'timing/model': 1.9999623929907102, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.49}) +[2026-06-23 06:05:40,645][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.23339271545410156, 'train/grad_norm_pre_clip': 42.112152099609375, 'timing/data': 0.0013827260117977858, 'timing/model': 2.0065293929947075, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.49}) +[2026-06-23 06:05:44,654][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.1215343177318573, 'train/grad_norm_pre_clip': 20.557600021362305, 'timing/data': 0.055157380003947765, 'timing/model': 1.9964204359857831, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.49}) +[2026-06-23 06:05:48,623][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.05582055449485779, 'train/grad_norm_pre_clip': 20.98699951171875, 'timing/data': 0.0011793189914897084, 'timing/model': 2.0081678289861884, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.49}) +[2026-06-23 06:05:52,665][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.12729132175445557, 'train/grad_norm_pre_clip': 20.498430252075195, 'timing/data': 0.001357900007860735, 'timing/model': 2.0027416859811638, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.49}) +[2026-06-23 06:05:56,640][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.14954376220703125, 'train/grad_norm_pre_clip': 21.005104064941406, 'timing/data': 0.0012516529823187739, 'timing/model': 2.00283401098568, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.49}) +[2026-06-23 06:06:00,846][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.04600825905799866, 'train/grad_norm_pre_clip': 7.264446258544922, 'timing/data': 0.061420473997714, 'timing/model': 2.18089165200945, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.49}) +[2026-06-23 06:06:04,804][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.17465746402740479, 'train/grad_norm_pre_clip': 33.82174301147461, 'timing/data': 0.0012373770005069673, 'timing/model': 2.0025324930029456, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.49}) +[2026-06-23 06:06:08,850][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.24111220240592957, 'train/grad_norm_pre_clip': 40.49384307861328, 'timing/data': 0.001187765010399744, 'timing/model': 1.9998037939949427, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.49}) +[2026-06-23 06:06:12,813][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.09736353904008865, 'train/grad_norm_pre_clip': 21.062519073486328, 'timing/data': 0.0011828260030597448, 'timing/model': 2.0037885119963903, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.49}) +[2026-06-23 06:06:16,856][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.05732307210564613, 'train/grad_norm_pre_clip': 9.824066162109375, 'timing/data': 0.06416061701020226, 'timing/model': 2.011439675989095, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.5}) +[2026-06-23 06:06:20,833][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.2128535956144333, 'train/grad_norm_pre_clip': 29.656089782714844, 'timing/data': 0.0012291510065551847, 'timing/model': 2.012897945009172, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.5}) +[2026-06-23 06:06:24,901][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.2791211009025574, 'train/grad_norm_pre_clip': 50.882476806640625, 'timing/data': 0.0014082920097280294, 'timing/model': 2.008787853992544, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.5}) +[2026-06-23 06:06:28,898][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.21592164039611816, 'train/grad_norm_pre_clip': 40.665157318115234, 'timing/data': 0.0012464530009310693, 'timing/model': 2.0277139540121425, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.5}) +[2026-06-23 06:06:32,934][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.2658628225326538, 'train/grad_norm_pre_clip': 54.28463363647461, 'timing/data': 0.0542218099872116, 'timing/model': 2.0120586580014788, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.5}) +[2026-06-23 06:06:37,172][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.10852475464344025, 'train/grad_norm_pre_clip': 17.28032112121582, 'timing/data': 0.0014210560184437782, 'timing/model': 2.024825515021803, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.5}) +[2026-06-23 06:06:41,234][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.0974467471241951, 'train/grad_norm_pre_clip': 13.735902786254883, 'timing/data': 0.0011438039946369827, 'timing/model': 2.007966695004143, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.5}) +[2026-06-23 06:06:45,206][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.1681865006685257, 'train/grad_norm_pre_clip': 33.06874465942383, 'timing/data': 0.001159812993137166, 'timing/model': 2.0069857980124652, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.5}) +[2026-06-23 06:06:49,291][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.1554601788520813, 'train/grad_norm_pre_clip': 31.13707160949707, 'timing/data': 0.10434113699011505, 'timing/model': 2.0153356520168018, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.5}) +[2026-06-23 06:06:53,271][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.27155181765556335, 'train/grad_norm_pre_clip': 45.92542266845703, 'timing/data': 0.0011292669805698097, 'timing/model': 2.013838521001162, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.5}) +[2026-06-23 06:06:57,354][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.14049489796161652, 'train/grad_norm_pre_clip': 21.48599624633789, 'timing/data': 0.0012222479854244739, 'timing/model': 2.0262392819859087, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.5}) +[2026-06-23 06:07:01,334][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.152409166097641, 'train/grad_norm_pre_clip': 19.148916244506836, 'timing/data': 0.001153962017269805, 'timing/model': 2.010308797005564, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.5}) +[2026-06-23 06:07:05,371][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.24781343340873718, 'train/grad_norm_pre_clip': 48.54220199584961, 'timing/data': 0.053776604996528476, 'timing/model': 2.0128883400175255, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.5}) +[2026-06-23 06:07:09,395][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.1567617654800415, 'train/grad_norm_pre_clip': 19.960067749023438, 'timing/data': 0.0011822849919553846, 'timing/model': 2.0117796859994996, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.5}) +[2026-06-23 06:07:13,464][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.1622638702392578, 'train/grad_norm_pre_clip': 23.637529373168945, 'timing/data': 0.001291786989895627, 'timing/model': 2.011507718998473, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.5}) +[2026-06-23 06:07:17,445][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.08492948859930038, 'train/grad_norm_pre_clip': 10.870466232299805, 'timing/data': 0.0011727770033758134, 'timing/model': 2.0119088149804156, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.5}) +[2026-06-23 06:07:21,490][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.23726379871368408, 'train/grad_norm_pre_clip': 20.25116729736328, 'timing/data': 0.060742317989934236, 'timing/model': 2.0125219030014705, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.5}) +[2026-06-23 06:07:25,473][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.07181747257709503, 'train/grad_norm_pre_clip': 13.995970726013184, 'timing/data': 0.0011325230007059872, 'timing/model': 2.013223855989054, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.5}) +[2026-06-23 06:07:29,566][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.12125784903764725, 'train/grad_norm_pre_clip': 29.670452117919922, 'timing/data': 0.0014824300014879555, 'timing/model': 2.018309147999389, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.5}) +[2026-06-23 06:07:33,599][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.11979778856039047, 'train/grad_norm_pre_clip': 13.23186206817627, 'timing/data': 0.001400528009980917, 'timing/model': 2.0373153449909296, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.5}) +[2026-06-23 06:07:37,657][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1300 +[2026-06-23 06:09:32,863][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1300 +[2026-06-23 06:09:32,864][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1300 +[2026-06-23 06:11:39,462][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1300 +[2026-06-23 06:11:39,464][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.08913998305797577, 'train/grad_norm_pre_clip': 12.172773361206055, 'eval/action_loss/seconds': 115.20125089300564, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.0635925903916359, 'eval/latency_2/loss': 0.0635925903916359, 'eval/flappy/loss': 0.0635925903916359, 'eval/flappy/latency_2/loss': 0.0635925903916359, 'eval/flap_precision': 0.9208306294613887, 'eval/flap_recall': 0.9574898785425101, 'eval/flap_f1': 0.9388025140588818, 'eval/flap_rate_model': 0.2140277777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9388025140588818, 'eval/latency_2/flap_precision': 0.9208306294613887, 'eval/latency_2/flap_recall': 0.9574898785425101, 'eval/latency_2/flap_f1': 0.9388025140588818, 'eval/latency_2/flap_rate_model': 0.2140277777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9388025140588818, 'eval/action_classification/data_wait_seconds': 50.28218460083008, 'eval/action_classification/predict_seconds': 74.68290710449219, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06415753701003268, 'timing/model': 2.015619287005393, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.5}) +[2026-06-23 06:11:43,413][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.13478299975395203, 'train/grad_norm_pre_clip': 13.354814529418945, 'timing/data': 0.0012207149993628263, 'timing/model': 1.993172501999652, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.5}) +[2026-06-23 06:11:47,449][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.09794631600379944, 'train/grad_norm_pre_clip': 19.084980010986328, 'timing/data': 0.00142204700387083, 'timing/model': 1.9929616599984001, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.5}) +[2026-06-23 06:11:51,408][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.08641961216926575, 'train/grad_norm_pre_clip': 15.051039695739746, 'timing/data': 0.001516251009888947, 'timing/model': 1.997002494026674, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.5}) +[2026-06-23 06:11:55,424][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.1414734423160553, 'train/grad_norm_pre_clip': 22.907697677612305, 'timing/data': 0.06626435401267372, 'timing/model': 1.996332912996877, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.5}) +[2026-06-23 06:11:59,379][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.12365178763866425, 'train/grad_norm_pre_clip': 13.708062171936035, 'timing/data': 0.0013627279840875417, 'timing/model': 1.9963221910002176, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.51}) +[2026-06-23 06:12:03,576][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.177657812833786, 'train/grad_norm_pre_clip': 18.082353591918945, 'timing/data': 0.0011462480179034173, 'timing/model': 2.153589274996193, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.51}) +[2026-06-23 06:12:07,538][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.19926157593727112, 'train/grad_norm_pre_clip': 21.016332626342773, 'timing/data': 0.001556487026391551, 'timing/model': 2.0098265889973845, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.51}) +[2026-06-23 06:12:11,544][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.17513209581375122, 'train/grad_norm_pre_clip': 25.457090377807617, 'timing/data': 0.06187349901301786, 'timing/model': 1.9920243810047396, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.51}) +[2026-06-23 06:12:15,514][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.07026146352291107, 'train/grad_norm_pre_clip': 13.947407722473145, 'timing/data': 0.0011182469897903502, 'timing/model': 2.015228735981509, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.51}) +[2026-06-23 06:12:19,563][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.06131743639707565, 'train/grad_norm_pre_clip': 13.146080017089844, 'timing/data': 0.0012418139958754182, 'timing/model': 2.0054112220241223, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.51}) +[2026-06-23 06:12:23,528][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.12527725100517273, 'train/grad_norm_pre_clip': 19.010610580444336, 'timing/data': 0.0013420300092548132, 'timing/model': 2.0049612789880484, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.51}) +[2026-06-23 06:12:27,559][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.10072499513626099, 'train/grad_norm_pre_clip': 15.483771324157715, 'timing/data': 0.05645076901419088, 'timing/model': 2.009892577014398, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.51}) +[2026-06-23 06:12:31,521][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.22750937938690186, 'train/grad_norm_pre_clip': 33.757381439208984, 'timing/data': 0.0011889470042660832, 'timing/model': 2.003051884006709, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.51}) +[2026-06-23 06:12:35,808][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.1736871302127838, 'train/grad_norm_pre_clip': 25.969451904296875, 'timing/data': 0.0012495090195443481, 'timing/model': 2.237084253982175, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.51}) +[2026-06-23 06:12:39,768][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.1964949369430542, 'train/grad_norm_pre_clip': 25.63446807861328, 'timing/data': 0.0011351379798725247, 'timing/model': 2.003575967013603, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.51}) +[2026-06-23 06:12:43,800][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.08851500600576401, 'train/grad_norm_pre_clip': 15.118860244750977, 'timing/data': 0.055188828002428636, 'timing/model': 2.0156266120029613, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.51}) +[2026-06-23 06:12:47,795][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.07814273238182068, 'train/grad_norm_pre_clip': 19.950550079345703, 'timing/data': 0.0012347919982858002, 'timing/model': 2.0251887880149297, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.51}) +[2026-06-23 06:12:51,876][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.10243310034275055, 'train/grad_norm_pre_clip': 27.66852378845215, 'timing/data': 0.001405015995260328, 'timing/model': 2.0118965649744496, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.51}) +[2026-06-23 06:12:55,858][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.13153627514839172, 'train/grad_norm_pre_clip': 12.279210090637207, 'timing/data': 0.0012799350079149008, 'timing/model': 2.010116012999788, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.51}) +[2026-06-23 06:12:59,898][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.1584736406803131, 'train/grad_norm_pre_clip': 29.9031925201416, 'timing/data': 0.06773712299764156, 'timing/model': 2.008487980987411, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.51}) +[2026-06-23 06:13:03,875][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.10986658930778503, 'train/grad_norm_pre_clip': 22.217552185058594, 'timing/data': 0.0012409129994921386, 'timing/model': 2.010854191001272, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.51}) +[2026-06-23 06:13:07,953][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.0852862000465393, 'train/grad_norm_pre_clip': 19.301681518554688, 'timing/data': 0.00122760800877586, 'timing/model': 2.0146335419849493, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.51}) +[2026-06-23 06:13:12,175][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.15832246840000153, 'train/grad_norm_pre_clip': 30.458911895751953, 'timing/data': 0.0018009880150202662, 'timing/model': 2.0154587640136015, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.51}) +[2026-06-23 06:13:16,230][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.1892094612121582, 'train/grad_norm_pre_clip': 30.997386932373047, 'timing/data': 0.07141538898576982, 'timing/model': 2.014223778998712, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.51}) +[2026-06-23 06:13:20,217][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.04485500603914261, 'train/grad_norm_pre_clip': 8.071245193481445, 'timing/data': 0.0014173589879646897, 'timing/model': 2.013266388006741, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.51}) +[2026-06-23 06:13:24,288][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.20206323266029358, 'train/grad_norm_pre_clip': 46.60503005981445, 'timing/data': 0.0012118199956603348, 'timing/model': 2.0122448980109766, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.51}) +[2026-06-23 06:13:28,266][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.03600117191672325, 'train/grad_norm_pre_clip': 9.406425476074219, 'timing/data': 0.0011972719803452492, 'timing/model': 2.0145147999865003, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.51}) +[2026-06-23 06:13:32,299][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.05054091289639473, 'train/grad_norm_pre_clip': 13.999473571777344, 'timing/data': 0.05619521500193514, 'timing/model': 2.009750734985573, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.51}) +[2026-06-23 06:13:36,283][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.07760169357061386, 'train/grad_norm_pre_clip': 42.56043243408203, 'timing/data': 0.001142370980232954, 'timing/model': 2.0159246300172526, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.51}) +[2026-06-23 06:13:40,372][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.08097243309020996, 'train/grad_norm_pre_clip': 27.43400764465332, 'timing/data': 0.0011345360253471881, 'timing/model': 2.027466444997117, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.51}) +[2026-06-23 06:13:44,441][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.3740542531013489, 'train/grad_norm_pre_clip': 29.442556381225586, 'timing/data': 0.001320750016020611, 'timing/model': 2.099585712014232, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.52}) +[2026-06-23 06:13:48,495][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.231643408536911, 'train/grad_norm_pre_clip': 29.928800582885742, 'timing/data': 0.06633011301164515, 'timing/model': 2.016229757020483, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.52}) +[2026-06-23 06:13:52,480][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.1533031165599823, 'train/grad_norm_pre_clip': 19.447843551635742, 'timing/data': 0.001142550987424329, 'timing/model': 2.0127057140052784, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.52}) +[2026-06-23 06:13:56,555][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.12138789147138596, 'train/grad_norm_pre_clip': 24.148801803588867, 'timing/data': 0.0011460680107120425, 'timing/model': 2.0122605079959612, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.52}) +[2026-06-23 06:14:00,543][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.16261309385299683, 'train/grad_norm_pre_clip': 18.032032012939453, 'timing/data': 0.0011602540034800768, 'timing/model': 2.0173845140088815, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.52}) +[2026-06-23 06:14:04,592][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.11598902940750122, 'train/grad_norm_pre_clip': 26.266447067260742, 'timing/data': 0.06429004200617783, 'timing/model': 2.013006432011025, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.52}) +[2026-06-23 06:14:08,582][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.08552398532629013, 'train/grad_norm_pre_clip': 24.742223739624023, 'timing/data': 0.0011868039728142321, 'timing/model': 2.0159512320242357, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 0.52}) +[2026-06-23 06:14:12,658][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.07429233938455582, 'train/grad_norm_pre_clip': 11.568876266479492, 'timing/data': 0.0012064790062140673, 'timing/model': 2.014030089980224, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 0.52}) +[2026-06-23 06:14:16,647][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.0858859270811081, 'train/grad_norm_pre_clip': 16.241844177246094, 'timing/data': 0.0012521529861260206, 'timing/model': 2.0164050860039424, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 0.52}) +[2026-06-23 06:14:20,892][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.02956985868513584, 'train/grad_norm_pre_clip': 7.8783063888549805, 'timing/data': 0.05825562198879197, 'timing/model': 2.013329651003005, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.52}) +[2026-06-23 06:14:24,878][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.07880839705467224, 'train/grad_norm_pre_clip': 14.291254043579102, 'timing/data': 0.00120592798339203, 'timing/model': 2.014743912004633, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 0.52}) +[2026-06-23 06:14:28,949][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.06976892054080963, 'train/grad_norm_pre_clip': 15.115912437438965, 'timing/data': 0.0012096450082026422, 'timing/model': 2.0116240750066936, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 0.52}) +[2026-06-23 06:14:32,934][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.07637836784124374, 'train/grad_norm_pre_clip': 19.635883331298828, 'timing/data': 0.0012172400020062923, 'timing/model': 2.0157982660166454, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 0.52}) +[2026-06-23 06:14:36,974][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.07618775963783264, 'train/grad_norm_pre_clip': 14.449316024780273, 'timing/data': 0.05559905301197432, 'timing/model': 2.01347700500628, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 0.52}) +[2026-06-23 06:14:40,963][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.1766781508922577, 'train/grad_norm_pre_clip': 25.67165184020996, 'timing/data': 0.001227378990733996, 'timing/model': 2.0191866630048025, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 0.52}) +[2026-06-23 06:14:45,055][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.11085387319326401, 'train/grad_norm_pre_clip': 19.659503936767578, 'timing/data': 0.0014188109780661762, 'timing/model': 2.031406340014655, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 0.52}) +[2026-06-23 06:14:49,046][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.19101522862911224, 'train/grad_norm_pre_clip': 26.234569549560547, 'timing/data': 0.0012675320031121373, 'timing/model': 2.0154141480161343, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 0.52}) +[2026-06-23 06:14:53,307][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.14890140295028687, 'train/grad_norm_pre_clip': 24.295921325683594, 'timing/data': 0.05538052899646573, 'timing/model': 2.013569536997238, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 0.52}) +[2026-06-23 06:14:57,317][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.13072967529296875, 'train/grad_norm_pre_clip': 27.16375160217285, 'timing/data': 0.001262903999304399, 'timing/model': 2.0360793079889845, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 0.52}) +[2026-06-23 06:15:01,396][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.054659005254507065, 'train/grad_norm_pre_clip': 15.343932151794434, 'timing/data': 0.0012349820171948522, 'timing/model': 2.014793665992329, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.52}) +[2026-06-23 06:15:05,401][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.13617312908172607, 'train/grad_norm_pre_clip': 23.89322853088379, 'timing/data': 0.0013571479939855635, 'timing/model': 2.0307980549987406, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 0.52}) +[2026-06-23 06:15:09,464][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.11366777867078781, 'train/grad_norm_pre_clip': 19.803874969482422, 'timing/data': 0.058323868026491255, 'timing/model': 2.022163177985931, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 0.52}) +[2026-06-23 06:15:13,464][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.03719882667064667, 'train/grad_norm_pre_clip': 11.496928215026855, 'timing/data': 0.00124676400446333, 'timing/model': 2.0198466060101055, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 0.52}) +[2026-06-23 06:15:17,555][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.21191169321537018, 'train/grad_norm_pre_clip': 35.19455337524414, 'timing/data': 0.001442795997718349, 'timing/model': 2.0204627529892605, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 0.52}) +[2026-06-23 06:15:21,562][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.056627050042152405, 'train/grad_norm_pre_clip': 11.225756645202637, 'timing/data': 0.0011844190012197942, 'timing/model': 2.0351816070033237, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 0.52}) +[2026-06-23 06:15:25,642][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.14096398651599884, 'train/grad_norm_pre_clip': 21.641517639160156, 'timing/data': 0.05351815797621384, 'timing/model': 2.0149660329916514, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 0.52}) +[2026-06-23 06:15:29,632][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.10099563747644424, 'train/grad_norm_pre_clip': 13.830719947814941, 'timing/data': 0.0011090589978266507, 'timing/model': 2.01723725697957, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 0.53}) +[2026-06-23 06:15:33,699][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.04670209810137749, 'train/grad_norm_pre_clip': 15.206799507141113, 'timing/data': 0.0012826700112782419, 'timing/model': 2.0123889970127493, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 0.53}) +[2026-06-23 06:15:37,703][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.12894020974636078, 'train/grad_norm_pre_clip': 37.02450942993164, 'timing/data': 0.0011471289908513427, 'timing/model': 2.0319023100018967, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 0.53}) +[2026-06-23 06:15:41,752][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.1643892079591751, 'train/grad_norm_pre_clip': 36.09914779663086, 'timing/data': 0.05726417902042158, 'timing/model': 2.017509153985884, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.53}) +[2026-06-23 06:15:45,735][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.22409000992774963, 'train/grad_norm_pre_clip': 37.30261993408203, 'timing/data': 0.0011820239888038486, 'timing/model': 2.011867147986777, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 0.53}) +[2026-06-23 06:15:49,829][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.05363873019814491, 'train/grad_norm_pre_clip': 23.687223434448242, 'timing/data': 0.0011087890015915036, 'timing/model': 2.0425928790064063, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 0.53}) +[2026-06-23 06:15:53,818][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.08965302258729935, 'train/grad_norm_pre_clip': 28.34699249267578, 'timing/data': 0.0012161469785496593, 'timing/model': 2.0174916769901756, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 0.53}) +[2026-06-23 06:15:57,864][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.10846127569675446, 'train/grad_norm_pre_clip': 38.18821334838867, 'timing/data': 0.06259097199654207, 'timing/model': 2.0140581500018016, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 0.53}) +[2026-06-23 06:16:02,031][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.11751667410135269, 'train/grad_norm_pre_clip': 29.914684295654297, 'timing/data': 0.0011770440032705665, 'timing/model': 2.0153875950200018, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 0.53}) +[2026-06-23 06:16:06,099][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.24038511514663696, 'train/grad_norm_pre_clip': 60.33454895019531, 'timing/data': 0.001129968004534021, 'timing/model': 2.015368558000773, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 0.53}) +[2026-06-23 06:16:10,095][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.1594284623861313, 'train/grad_norm_pre_clip': 33.569576263427734, 'timing/data': 0.0012336389918345958, 'timing/model': 2.0204848799912725, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 0.53}) +[2026-06-23 06:16:14,150][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.19890505075454712, 'train/grad_norm_pre_clip': 18.98322296142578, 'timing/data': 0.061104363994672894, 'timing/model': 2.0180111580120865, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 0.53}) +[2026-06-23 06:16:18,139][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.3065715730190277, 'train/grad_norm_pre_clip': 39.757530212402344, 'timing/data': 0.0011328630207572132, 'timing/model': 2.0182789199752733, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 0.53}) +[2026-06-23 06:16:22,213][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.08314169943332672, 'train/grad_norm_pre_clip': 24.54180908203125, 'timing/data': 0.0011358989868313074, 'timing/model': 2.0150488000072073, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.53}) +[2026-06-23 06:16:26,203][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.1411600261926651, 'train/grad_norm_pre_clip': 15.766429901123047, 'timing/data': 0.0011721259797923267, 'timing/model': 2.017407968000043, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 0.53}) +[2026-06-23 06:16:30,280][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.13958631455898285, 'train/grad_norm_pre_clip': 20.515201568603516, 'timing/data': 0.06903128098929301, 'timing/model': 2.034951771987835, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 0.53}) +[2026-06-23 06:16:34,535][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.13537509739398956, 'train/grad_norm_pre_clip': 26.030141830444336, 'timing/data': 0.001257412979612127, 'timing/model': 2.277367367001716, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 0.53}) +[2026-06-23 06:16:38,616][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.05586914345622063, 'train/grad_norm_pre_clip': 18.088571548461914, 'timing/data': 0.0012200540222693235, 'timing/model': 2.022266109008342, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.53}) +[2026-06-23 06:16:42,610][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.1072269156575203, 'train/grad_norm_pre_clip': 23.778446197509766, 'timing/data': 0.0014261049800552428, 'timing/model': 2.015616040007444, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 0.53}) +[2026-06-23 06:16:46,660][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.1153310090303421, 'train/grad_norm_pre_clip': 17.62052345275879, 'timing/data': 0.05375250999350101, 'timing/model': 2.0212339889840223, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 0.53}) +[2026-06-23 06:16:50,681][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.11955513805150986, 'train/grad_norm_pre_clip': 19.828264236450195, 'timing/data': 0.0013981329975649714, 'timing/model': 2.038796287000878, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 0.53}) +[2026-06-23 06:16:54,778][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.07424364984035492, 'train/grad_norm_pre_clip': 13.172441482543945, 'timing/data': 0.0012397909886203706, 'timing/model': 2.028489021002315, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 0.53}) +[2026-06-23 06:16:58,788][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.17495739459991455, 'train/grad_norm_pre_clip': 17.330846786499023, 'timing/data': 0.0012950930104125291, 'timing/model': 2.030603886989411, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 0.53}) +[2026-06-23 06:17:02,843][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.25877830386161804, 'train/grad_norm_pre_clip': 21.035053253173828, 'timing/data': 0.05188340600579977, 'timing/model': 2.022571473993594, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.53}) +[2026-06-23 06:17:06,896][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.1298084259033203, 'train/grad_norm_pre_clip': 15.769957542419434, 'timing/data': 0.0013155500055290759, 'timing/model': 2.0685331459972076, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 0.53}) +[2026-06-23 06:17:10,983][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.16869033873081207, 'train/grad_norm_pre_clip': 17.793127059936523, 'timing/data': 0.0013781969901174307, 'timing/model': 2.021883419016376, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 0.53}) +[2026-06-23 06:17:14,997][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.1077871173620224, 'train/grad_norm_pre_clip': 10.233756065368652, 'timing/data': 0.0015226239920593798, 'timing/model': 2.0321407400188036, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 0.54}) +[2026-06-23 06:17:19,053][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.12821301817893982, 'train/grad_norm_pre_clip': 19.51468276977539, 'timing/data': 0.05518690898315981, 'timing/model': 2.0223872869974002, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 0.54}) +[2026-06-23 06:17:23,059][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.09308719635009766, 'train/grad_norm_pre_clip': 17.524269104003906, 'timing/data': 0.0012240009964443743, 'timing/model': 2.0211976769787725, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 0.54}) +[2026-06-23 06:17:27,161][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.16145071387290955, 'train/grad_norm_pre_clip': 28.63544464111328, 'timing/data': 0.001238288008607924, 'timing/model': 2.0344589080195874, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 0.54}) +[2026-06-23 06:17:31,175][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.07230629771947861, 'train/grad_norm_pre_clip': 19.459869384765625, 'timing/data': 0.0015295959892682731, 'timing/model': 2.029333927988773, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 0.54}) +[2026-06-23 06:17:35,226][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.09258905798196793, 'train/grad_norm_pre_clip': 10.549139976501465, 'timing/data': 0.05508705499232747, 'timing/model': 2.0175039780151565, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 0.54}) +[2026-06-23 06:17:39,237][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.16898629069328308, 'train/grad_norm_pre_clip': 23.172189712524414, 'timing/data': 0.0012009890051558614, 'timing/model': 2.035433526005363, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 0.54}) +[2026-06-23 06:17:43,361][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.1852552890777588, 'train/grad_norm_pre_clip': 27.2829532623291, 'timing/data': 0.0012220980133861303, 'timing/model': 2.0194137239886913, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.54}) +[2026-06-23 06:17:47,362][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.11345291882753372, 'train/grad_norm_pre_clip': 18.801483154296875, 'timing/data': 0.0013772939855698496, 'timing/model': 2.0195803300011903, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 0.54}) +[2026-06-23 06:17:51,434][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.07713419944047928, 'train/grad_norm_pre_clip': 16.302379608154297, 'timing/data': 0.06606911699054763, 'timing/model': 2.0249198649835307, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 0.54}) +[2026-06-23 06:17:55,437][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.27367427945137024, 'train/grad_norm_pre_clip': 41.586273193359375, 'timing/data': 0.0014041649992577732, 'timing/model': 2.0237228719925042, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 0.54}) +[2026-06-23 06:17:59,574][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.24926891922950745, 'train/grad_norm_pre_clip': 27.794523239135742, 'timing/data': 0.0012602389906533062, 'timing/model': 2.0739527200057637, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 0.54}) +[2026-06-23 06:18:03,575][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.07065519690513611, 'train/grad_norm_pre_clip': 18.808313369750977, 'timing/data': 0.0014199840079527348, 'timing/model': 2.0237453910231125, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.54}) +[2026-06-23 06:18:07,638][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.1834615021944046, 'train/grad_norm_pre_clip': 32.55765914916992, 'timing/data': 0.055381489015417174, 'timing/model': 2.0243796209979337, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 0.54}) +[2026-06-23 06:18:11,642][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.16754023730754852, 'train/grad_norm_pre_clip': 18.8536434173584, 'timing/data': 0.0013954380119685084, 'timing/model': 2.023580963985296, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 0.54}) +[2026-06-23 06:18:15,782][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.1157824695110321, 'train/grad_norm_pre_clip': 26.926712036132812, 'timing/data': 0.0012436879915185273, 'timing/model': 2.0726156649761833, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 0.54}) +[2026-06-23 06:18:19,792][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.22802092134952545, 'train/grad_norm_pre_clip': 41.76776885986328, 'timing/data': 0.0013352969835978001, 'timing/model': 2.0234397609892767, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 0.54}) +[2026-06-23 06:18:23,846][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1400 +[2026-06-23 06:20:18,756][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1400 +[2026-06-23 06:20:18,757][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1400 +[2026-06-23 06:22:25,479][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1400 +[2026-06-23 06:22:25,481][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.13033176958560944, 'train/grad_norm_pre_clip': 18.33019256591797, 'eval/action_loss/seconds': 114.90571264200844, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06611471623182297, 'eval/latency_2/loss': 0.06611471623182297, 'eval/flappy/loss': 0.06611471623182297, 'eval/flappy/latency_2/loss': 0.06611471623182297, 'eval/flap_precision': 0.8991282689912827, 'eval/flap_recall': 0.9743589743589743, 'eval/flap_f1': 0.9352331606217616, 'eval/flap_rate_model': 0.22305555555555556, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9352331606217616, 'eval/latency_2/flap_precision': 0.8991282689912827, 'eval/latency_2/flap_recall': 0.9743589743589743, 'eval/latency_2/flap_f1': 0.9352331606217616, 'eval/latency_2/flap_rate_model': 0.22305555555555556, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9352331606217616, 'eval/action_classification/data_wait_seconds': 50.28559875488281, 'eval/action_classification/predict_seconds': 74.66342163085938, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.05543939600465819, 'timing/model': 2.0217299799842294, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.54}) +[2026-06-23 06:22:29,421][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.0441780611872673, 'train/grad_norm_pre_clip': 7.22280216217041, 'timing/data': 0.0011968910112045705, 'timing/model': 1.9878972230071668, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 0.54}) +[2026-06-23 06:22:33,456][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.09404925256967545, 'train/grad_norm_pre_clip': 19.71216583251953, 'timing/data': 0.001259206997929141, 'timing/model': 1.990109840000514, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 0.54}) +[2026-06-23 06:22:37,415][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.21857260167598724, 'train/grad_norm_pre_clip': 29.206239700317383, 'timing/data': 0.0014021109964232892, 'timing/model': 2.0017063269915525, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 0.54}) +[2026-06-23 06:22:41,540][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.11213133484125137, 'train/grad_norm_pre_clip': 18.400545120239258, 'timing/data': 0.057934553013183177, 'timing/model': 1.9939303279970773, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 0.54}) +[2026-06-23 06:22:45,564][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.13084900379180908, 'train/grad_norm_pre_clip': 17.010164260864258, 'timing/data': 0.0013747000193689018, 'timing/model': 2.00560219300678, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 0.54}) +[2026-06-23 06:22:49,623][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.09023762494325638, 'train/grad_norm_pre_clip': 25.112815856933594, 'timing/data': 0.0013041399943176657, 'timing/model': 1.996971725980984, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 0.54}) +[2026-06-23 06:22:53,576][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.2334488034248352, 'train/grad_norm_pre_clip': 30.325225830078125, 'timing/data': 0.0012890920042991638, 'timing/model': 1.9990875550138298, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 0.54}) +[2026-06-23 06:22:57,588][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.10426318645477295, 'train/grad_norm_pre_clip': 18.787763595581055, 'timing/data': 0.05572152600507252, 'timing/model': 1.999746259971289, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 0.54}) +[2026-06-23 06:23:01,549][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.06868650019168854, 'train/grad_norm_pre_clip': 13.850646018981934, 'timing/data': 0.0014540069969370961, 'timing/model': 2.0019737650000025, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 0.55}) +[2026-06-23 06:23:05,611][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.15606456995010376, 'train/grad_norm_pre_clip': 31.933643341064453, 'timing/data': 0.001367466989904642, 'timing/model': 2.018133753008442, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.55}) +[2026-06-23 06:23:09,594][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.12044896185398102, 'train/grad_norm_pre_clip': 33.3170051574707, 'timing/data': 0.0012665599933825433, 'timing/model': 2.017277731996728, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 0.55}) +[2026-06-23 06:23:13,747][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.25527405738830566, 'train/grad_norm_pre_clip': 55.585357666015625, 'timing/data': 0.05259262301842682, 'timing/model': 2.143725945003098, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 0.55}) +[2026-06-23 06:23:17,729][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.14414581656455994, 'train/grad_norm_pre_clip': 49.384552001953125, 'timing/data': 0.0012004979944322258, 'timing/model': 2.0211079259752296, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 0.55}) +[2026-06-23 06:23:21,768][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.08007633686065674, 'train/grad_norm_pre_clip': 13.052605628967285, 'timing/data': 0.001175140991108492, 'timing/model': 1.9979991390137002, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 0.55}) +[2026-06-23 06:23:25,728][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.07745806127786636, 'train/grad_norm_pre_clip': 14.79158878326416, 'timing/data': 0.001179398997919634, 'timing/model': 2.0005382689996623, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 0.55}) +[2026-06-23 06:23:29,742][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.21648424863815308, 'train/grad_norm_pre_clip': 36.440345764160156, 'timing/data': 0.05601630901219323, 'timing/model': 2.0007678119873162, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 0.55}) +[2026-06-23 06:23:33,710][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.21011777222156525, 'train/grad_norm_pre_clip': 31.738250732421875, 'timing/data': 0.001127091993112117, 'timing/model': 2.0077072640124243, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 0.55}) +[2026-06-23 06:23:37,762][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.12099923193454742, 'train/grad_norm_pre_clip': 27.74506187438965, 'timing/data': 0.0012871389917563647, 'timing/model': 2.004883866000455, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 0.55}) +[2026-06-23 06:23:41,723][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.10664437711238861, 'train/grad_norm_pre_clip': 19.75302505493164, 'timing/data': 0.0011174740211572498, 'timing/model': 2.0026006160187535, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 0.55}) +[2026-06-23 06:23:45,983][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.07009845227003098, 'train/grad_norm_pre_clip': 12.615747451782227, 'timing/data': 0.0532202540198341, 'timing/model': 2.243343387992354, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.55}) +[2026-06-23 06:23:49,951][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.13887934386730194, 'train/grad_norm_pre_clip': 20.581180572509766, 'timing/data': 0.0012516030110418797, 'timing/model': 2.006351876014378, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 0.55}) +[2026-06-23 06:23:54,009][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.10856887698173523, 'train/grad_norm_pre_clip': 27.103044509887695, 'timing/data': 0.0011037090152967721, 'timing/model': 2.0068034710129723, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 0.55}) +[2026-06-23 06:23:57,994][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.18096747994422913, 'train/grad_norm_pre_clip': 13.083407402038574, 'timing/data': 0.0013489820121321827, 'timing/model': 2.0179052710009273, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.55}) +[2026-06-23 06:24:02,027][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.2779199481010437, 'train/grad_norm_pre_clip': 28.892488479614258, 'timing/data': 0.058762943983310834, 'timing/model': 2.007357293012319, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 0.55}) +[2026-06-23 06:24:06,009][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.28066486120224, 'train/grad_norm_pre_clip': 23.125001907348633, 'timing/data': 0.0013312799856066704, 'timing/model': 2.014514384994982, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 0.55}) +[2026-06-23 06:24:10,087][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.10024115443229675, 'train/grad_norm_pre_clip': 14.550751686096191, 'timing/data': 0.0011858810030389577, 'timing/model': 2.019669620000059, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 0.55}) +[2026-06-23 06:24:14,083][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.11992575228214264, 'train/grad_norm_pre_clip': 20.942237854003906, 'timing/data': 0.0013164220144972205, 'timing/model': 2.025292407022789, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 0.55}) +[2026-06-23 06:24:18,129][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.16685575246810913, 'train/grad_norm_pre_clip': 20.429264068603516, 'timing/data': 0.05313788098283112, 'timing/model': 2.017272730008699, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 0.55}) +[2026-06-23 06:24:22,366][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.07115238904953003, 'train/grad_norm_pre_clip': 13.866069793701172, 'timing/data': 0.0012161470076534897, 'timing/model': 2.0202370759798214, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 0.55}) +[2026-06-23 06:24:26,452][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.08924274146556854, 'train/grad_norm_pre_clip': 17.93897247314453, 'timing/data': 0.0014473650080617517, 'timing/model': 2.0302324529911857, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.55}) +[2026-06-23 06:24:30,456][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.11369200050830841, 'train/grad_norm_pre_clip': 22.918731689453125, 'timing/data': 0.0014873580075800419, 'timing/model': 2.0310157999920193, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 0.55}) +[2026-06-23 06:24:34,502][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.04678727686405182, 'train/grad_norm_pre_clip': 7.568082332611084, 'timing/data': 0.052958991000195965, 'timing/model': 2.018330460996367, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 0.55}) +[2026-06-23 06:24:38,493][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.09034983813762665, 'train/grad_norm_pre_clip': 12.349058151245117, 'timing/data': 0.0012269870203454047, 'timing/model': 2.015952397021465, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 0.55}) +[2026-06-23 06:24:42,587][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.1509290635585785, 'train/grad_norm_pre_clip': 18.610595703125, 'timing/data': 0.0013174139894545078, 'timing/model': 2.032174138992559, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 0.55}) +[2026-06-23 06:24:46,586][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.078220434486866, 'train/grad_norm_pre_clip': 10.51876163482666, 'timing/data': 0.0012458210112527013, 'timing/model': 2.024421884008916, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 0.56}) +[2026-06-23 06:24:50,633][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.12911441922187805, 'train/grad_norm_pre_clip': 20.56908416748047, 'timing/data': 0.05445471400162205, 'timing/model': 2.0201326189853717, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 0.56}) +[2026-06-23 06:24:54,729][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.1973520964384079, 'train/grad_norm_pre_clip': 18.007038116455078, 'timing/data': 0.0012086530041415244, 'timing/model': 2.12041521500214, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 0.56}) +[2026-06-23 06:24:58,811][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.06444327533245087, 'train/grad_norm_pre_clip': 11.070850372314453, 'timing/data': 0.0014914150233380497, 'timing/model': 2.0186921560089104, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 0.56}) +[2026-06-23 06:25:02,808][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.07632172107696533, 'train/grad_norm_pre_clip': 13.756253242492676, 'timing/data': 0.0012479260039981455, 'timing/model': 2.021531801990932, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 0.56}) +[2026-06-23 06:25:06,863][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.21290569007396698, 'train/grad_norm_pre_clip': 32.829185485839844, 'timing/data': 0.06340253600501455, 'timing/model': 2.0175141180225182, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.56}) +[2026-06-23 06:25:10,873][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.12174016237258911, 'train/grad_norm_pre_clip': 27.22569465637207, 'timing/data': 0.0013621269899886101, 'timing/model': 2.033820474985987, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 0.56}) +[2026-06-23 06:25:14,955][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.1306672990322113, 'train/grad_norm_pre_clip': 12.486547470092773, 'timing/data': 0.0013393140106927603, 'timing/model': 2.0210102970013395, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 0.56}) +[2026-06-23 06:25:18,955][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.13712438941001892, 'train/grad_norm_pre_clip': 15.491676330566406, 'timing/data': 0.0012420350103639066, 'timing/model': 2.0248272779863328, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 0.56}) +[2026-06-23 06:25:23,040][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.05995173752307892, 'train/grad_norm_pre_clip': 15.653117179870605, 'timing/data': 0.06919110199669376, 'timing/model': 2.0246844639768824, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 0.56}) +[2026-06-23 06:25:27,029][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.20098242163658142, 'train/grad_norm_pre_clip': 22.64337158203125, 'timing/data': 0.0011465080024208874, 'timing/model': 2.01705460602534, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 0.56}) +[2026-06-23 06:25:31,377][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.16195259988307953, 'train/grad_norm_pre_clip': 32.92228698730469, 'timing/data': 0.0014073399943299592, 'timing/model': 2.034937507007271, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 0.56}) +[2026-06-23 06:25:35,381][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.03838352859020233, 'train/grad_norm_pre_clip': 13.255786895751953, 'timing/data': 0.0014104259898886085, 'timing/model': 2.0224870310048573, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 0.56}) +[2026-06-23 06:25:39,441][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.09238352626562119, 'train/grad_norm_pre_clip': 10.45117473602295, 'timing/data': 0.05314317098236643, 'timing/model': 2.029811985004926, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 0.56}) +[2026-06-23 06:25:43,441][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.12647071480751038, 'train/grad_norm_pre_clip': 18.666902542114258, 'timing/data': 0.0011956189991906285, 'timing/model': 2.0204191399971023, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 0.56}) +[2026-06-23 06:25:47,525][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.18066708743572235, 'train/grad_norm_pre_clip': 30.833702087402344, 'timing/data': 0.0012264460092410445, 'timing/model': 2.0233128550171386, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.56}) +[2026-06-23 06:25:51,522][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.16252847015857697, 'train/grad_norm_pre_clip': 32.79422378540039, 'timing/data': 0.0012558400048874319, 'timing/model': 2.0186974779935554, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 0.56}) +[2026-06-23 06:25:55,572][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.1668689250946045, 'train/grad_norm_pre_clip': 24.815122604370117, 'timing/data': 0.052337151020765305, 'timing/model': 2.019662549981149, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 0.56}) +[2026-06-23 06:25:59,575][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.12597496807575226, 'train/grad_norm_pre_clip': 18.471763610839844, 'timing/data': 0.0012233100133016706, 'timing/model': 2.023453334986698, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 0.56}) +[2026-06-23 06:26:03,903][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.10042008757591248, 'train/grad_norm_pre_clip': 31.2971248626709, 'timing/data': 0.0014540769916493446, 'timing/model': 2.2648621940170415, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 0.56}) +[2026-06-23 06:26:07,906][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.16716539859771729, 'train/grad_norm_pre_clip': 49.69826126098633, 'timing/data': 0.001282859971979633, 'timing/model': 2.023644525994314, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 0.56}) +[2026-06-23 06:26:11,961][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.17120803892612457, 'train/grad_norm_pre_clip': 16.427709579467773, 'timing/data': 0.052724427019711584, 'timing/model': 2.017955745017389, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 0.56}) +[2026-06-23 06:26:15,979][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.09753292798995972, 'train/grad_norm_pre_clip': 11.084242820739746, 'timing/data': 0.0012051969824824482, 'timing/model': 2.0185810099937953, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 0.56}) +[2026-06-23 06:26:20,063][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.1979016214609146, 'train/grad_norm_pre_clip': 18.11821174621582, 'timing/data': 0.0013019760081078857, 'timing/model': 2.0207134890079033, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 0.56}) +[2026-06-23 06:26:24,061][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.08150853216648102, 'train/grad_norm_pre_clip': 10.152511596679688, 'timing/data': 0.0012218570045661181, 'timing/model': 2.0160451630072203, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 0.56}) +[2026-06-23 06:26:28,123][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.1257195770740509, 'train/grad_norm_pre_clip': 32.007568359375, 'timing/data': 0.06462824100162834, 'timing/model': 2.0179314379929565, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.57}) +[2026-06-23 06:26:32,120][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.13345080614089966, 'train/grad_norm_pre_clip': 17.30030059814453, 'timing/data': 0.0012089840020053089, 'timing/model': 2.018499826983316, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 0.57}) +[2026-06-23 06:26:36,325][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.0985550731420517, 'train/grad_norm_pre_clip': 14.382208824157715, 'timing/data': 0.0013044500083196908, 'timing/model': 2.144182052987162, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 0.57}) +[2026-06-23 06:26:40,318][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.10526115447282791, 'train/grad_norm_pre_clip': 15.685153007507324, 'timing/data': 0.0012432069925125688, 'timing/model': 2.0227646059938706, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 0.57}) +[2026-06-23 06:26:44,346][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.16786356270313263, 'train/grad_norm_pre_clip': 19.381732940673828, 'timing/data': 0.04921329900389537, 'timing/model': 2.010185413004365, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 0.57}) +[2026-06-23 06:26:48,338][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.0989602655172348, 'train/grad_norm_pre_clip': 19.068220138549805, 'timing/data': 0.0011201700253877789, 'timing/model': 2.020182154024951, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.57}) +[2026-06-23 06:26:52,424][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.11933421343564987, 'train/grad_norm_pre_clip': 24.266368865966797, 'timing/data': 0.0012372260098345578, 'timing/model': 2.017009180999594, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 0.57}) +[2026-06-23 06:26:56,413][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.10761644691228867, 'train/grad_norm_pre_clip': 13.194449424743652, 'timing/data': 0.0012078120198566467, 'timing/model': 2.0141674900078215, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 0.57}) +[2026-06-23 06:27:00,458][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.12912462651729584, 'train/grad_norm_pre_clip': 15.965591430664062, 'timing/data': 0.05101565897348337, 'timing/model': 2.0231638510013, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 0.57}) +[2026-06-23 06:27:04,458][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.10264035314321518, 'train/grad_norm_pre_clip': 14.500384330749512, 'timing/data': 0.0013861710031051189, 'timing/model': 2.023687148001045, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 0.57}) +[2026-06-23 06:27:08,538][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.14024314284324646, 'train/grad_norm_pre_clip': 33.14845657348633, 'timing/data': 0.0013394140114542097, 'timing/model': 2.014228871004889, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.57}) +[2026-06-23 06:27:12,744][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.11741377413272858, 'train/grad_norm_pre_clip': 25.47057342529297, 'timing/data': 0.0012383790162857622, 'timing/model': 2.0157087450206745, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 0.57}) +[2026-06-23 06:27:16,796][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.1506814956665039, 'train/grad_norm_pre_clip': 32.16855239868164, 'timing/data': 0.05288284900598228, 'timing/model': 2.0273389740032144, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.57}) +[2026-06-23 06:27:20,810][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.1363169550895691, 'train/grad_norm_pre_clip': 22.36649513244629, 'timing/data': 0.0011377829941920936, 'timing/model': 2.0354741140035912, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 0.57}) +[2026-06-23 06:27:24,892][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.13863420486450195, 'train/grad_norm_pre_clip': 23.243186950683594, 'timing/data': 0.0012062589812558144, 'timing/model': 2.022243006998906, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 0.57}) +[2026-06-23 06:27:28,888][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.10179811716079712, 'train/grad_norm_pre_clip': 13.52251148223877, 'timing/data': 0.001251472014700994, 'timing/model': 2.021586754010059, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 0.57}) +[2026-06-23 06:27:32,950][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.09047408401966095, 'train/grad_norm_pre_clip': 19.18987274169922, 'timing/data': 0.06470205599907786, 'timing/model': 2.0209239889809396, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 0.57}) +[2026-06-23 06:27:36,951][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.10499462485313416, 'train/grad_norm_pre_clip': 17.751298904418945, 'timing/data': 0.0012682040105573833, 'timing/model': 2.0203397989971563, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 0.57}) +[2026-06-23 06:27:41,045][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.09316005557775497, 'train/grad_norm_pre_clip': 19.115432739257812, 'timing/data': 0.0013724260206799954, 'timing/model': 2.0278349869768135, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 0.57}) +[2026-06-23 06:27:45,296][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.1410079002380371, 'train/grad_norm_pre_clip': 14.541641235351562, 'timing/data': 0.0013754009851254523, 'timing/model': 2.023990502988454, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 0.57}) +[2026-06-23 06:27:49,352][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.14497296512126923, 'train/grad_norm_pre_clip': 17.525283813476562, 'timing/data': 0.05232020901166834, 'timing/model': 2.0171468979970086, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.57}) +[2026-06-23 06:27:53,357][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.06528323888778687, 'train/grad_norm_pre_clip': 11.61738395690918, 'timing/data': 0.00118761500925757, 'timing/model': 2.029665408015717, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 0.57}) +[2026-06-23 06:27:57,435][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.16422827541828156, 'train/grad_norm_pre_clip': 16.494277954101562, 'timing/data': 0.001211458002217114, 'timing/model': 2.0173713520052843, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 0.57}) +[2026-06-23 06:28:01,437][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.11597725749015808, 'train/grad_norm_pre_clip': 31.95976448059082, 'timing/data': 0.0014492679911199957, 'timing/model': 2.0199935770069715, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 0.57}) +[2026-06-23 06:28:05,495][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.06434851884841919, 'train/grad_norm_pre_clip': 17.883861541748047, 'timing/data': 0.05670461998670362, 'timing/model': 2.0171108600043226, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 0.57}) +[2026-06-23 06:28:09,479][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.15001480281352997, 'train/grad_norm_pre_clip': 31.491878509521484, 'timing/data': 0.0013484410010278225, 'timing/model': 2.0114605389826465, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 0.57}) +[2026-06-23 06:28:13,578][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.09381825476884842, 'train/grad_norm_pre_clip': 18.734577178955078, 'timing/data': 0.0011433620238676667, 'timing/model': 2.043735070998082, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 0.58}) +[2026-06-23 06:28:17,617][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.07807193696498871, 'train/grad_norm_pre_clip': 14.484366416931152, 'timing/data': 0.0014303319912869483, 'timing/model': 2.0318784739938565, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 0.58}) +[2026-06-23 06:28:21,672][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.07394052296876907, 'train/grad_norm_pre_clip': 13.717813491821289, 'timing/data': 0.05435634101741016, 'timing/model': 2.019624303007731, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 0.58}) +[2026-06-23 06:28:25,673][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.19856759905815125, 'train/grad_norm_pre_clip': 36.624671936035156, 'timing/data': 0.0013608649896923453, 'timing/model': 2.01708169397898, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 0.58}) +[2026-06-23 06:28:29,756][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.01579217240214348, 'train/grad_norm_pre_clip': 6.335048198699951, 'timing/data': 0.0011971020139753819, 'timing/model': 2.0186782730161212, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.58}) +[2026-06-23 06:28:33,769][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.1586441844701767, 'train/grad_norm_pre_clip': 20.66085433959961, 'timing/data': 0.0013028070097789168, 'timing/model': 2.037127570976736, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 0.58}) +[2026-06-23 06:28:37,837][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.045947007834911346, 'train/grad_norm_pre_clip': 17.438011169433594, 'timing/data': 0.05471812901669182, 'timing/model': 2.031149665999692, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 0.58}) +[2026-06-23 06:28:41,830][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.11102481931447983, 'train/grad_norm_pre_clip': 16.779666900634766, 'timing/data': 0.00121818101615645, 'timing/model': 2.0145804819767363, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 0.58}) +[2026-06-23 06:28:45,919][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.17189374566078186, 'train/grad_norm_pre_clip': 28.566577911376953, 'timing/data': 0.0013819729792885482, 'timing/model': 2.0197286930051632, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 0.58}) +[2026-06-23 06:28:49,947][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.07239419966936111, 'train/grad_norm_pre_clip': 14.195082664489746, 'timing/data': 0.0016241709818132222, 'timing/model': 2.044842805014923, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 0.58}) +[2026-06-23 06:28:54,202][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.15701010823249817, 'train/grad_norm_pre_clip': 49.78771209716797, 'timing/data': 0.06022496399236843, 'timing/model': 2.0279423279862385, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 0.58}) +[2026-06-23 06:28:58,197][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.0725618377327919, 'train/grad_norm_pre_clip': 12.998387336730957, 'timing/data': 0.0011920719989575446, 'timing/model': 2.0220903750159778, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 0.58}) +[2026-06-23 06:29:02,278][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.10187914967536926, 'train/grad_norm_pre_clip': 17.957469940185547, 'timing/data': 0.001227186992764473, 'timing/model': 2.0199800460250117, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 0.58}) +[2026-06-23 06:29:06,260][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.3217146396636963, 'train/grad_norm_pre_clip': 38.6019172668457, 'timing/data': 0.0013870330003555864, 'timing/model': 2.0121128690079786, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 0.58}) +[2026-06-23 06:29:10,325][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1500 +[2026-06-23 06:31:04,953][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1500 +[2026-06-23 06:31:04,954][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1500 +[2026-06-23 06:33:11,699][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1500 +[2026-06-23 06:43:19,306][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-23 06:43:19,307][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.10832241177558899, 'train/grad_norm_pre_clip': 40.401832580566406, 'eval/action_loss/seconds': 114.62396363299922, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06995566189289093, 'eval/latency_2/loss': 0.06995566189289093, 'eval/flappy/loss': 0.06995566189289093, 'eval/flappy/latency_2/loss': 0.06995566189289093, 'eval/flap_precision': 0.885036496350365, 'eval/flap_recall': 0.9817813765182186, 'eval/flap_f1': 0.9309021113243763, 'eval/flap_rate_model': 0.22833333333333333, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9309021113243763, 'eval/latency_2/flap_precision': 0.885036496350365, 'eval/latency_2/flap_recall': 0.9817813765182186, 'eval/latency_2/flap_f1': 0.9309021113243763, 'eval/latency_2/flap_rate_model': 0.22833333333333333, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9309021113243763, 'eval/action_classification/data_wait_seconds': 50.254459381103516, 'eval/action_classification/predict_seconds': 74.81978607177734, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 156.02000000000382, 'rl_games_eval/mid_train/mean_length': 1275.6, 'rl_games_eval/mid_train/std_reward': 164.3303063953899, 'rl_games_eval/mid_train/std_length': 1324.9588069068411, 'rl_games_eval/mid_train/macro_mean_reward': 156.02000000000382, 'rl_games_eval/mid_train/macro_mean_length': 1275.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 156.02000000000382, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1275.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 164.3303063953899, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1324.9588069068411, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 156.02000000000382, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.06264878300135024, 'timing/model': 2.022892106004292, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.58}) +[2026-06-23 06:43:19,308][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_1500_state +[2026-06-23 06:43:19,316][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 06:43:28,050][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_1500_state/model.safetensors +[2026-06-23 06:43:31,997][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_1500_state/optimizer.bin +[2026-06-23 06:43:31,998][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_1500_state/sampler.bin +[2026-06-23 06:43:31,998][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-23 06:43:31,999][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-23 06:43:32,129][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-23 06:43:32,130][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 06:43:32,153][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 06:43:36,273][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.3148338794708252, 'train/grad_norm_pre_clip': 29.744943618774414, 'timing/data': 0.0012286199780646712, 'timing/model': 1.9674803100060672, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 0.58}) +[2026-06-23 06:43:40,255][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.08970693498849869, 'train/grad_norm_pre_clip': 28.953367233276367, 'timing/data': 0.0013996860070619732, 'timing/model': 1.9823030660045333, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 0.58}) +[2026-06-23 06:43:44,248][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.11458563804626465, 'train/grad_norm_pre_clip': 25.08677101135254, 'timing/data': 0.0012617910106200725, 'timing/model': 1.954347006016178, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 0.58}) +[2026-06-23 06:43:48,197][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.1271381974220276, 'train/grad_norm_pre_clip': 31.27503776550293, 'timing/data': 0.05412057199282572, 'timing/model': 1.9689963719865773, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 0.58}) +[2026-06-23 06:43:52,084][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.06809163093566895, 'train/grad_norm_pre_clip': 8.493814468383789, 'timing/data': 0.0012563119817059487, 'timing/model': 1.9650024930015206, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 0.58}) +[2026-06-23 06:43:56,057][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.09236650168895721, 'train/grad_norm_pre_clip': 14.572172164916992, 'timing/data': 0.0013915009913034737, 'timing/model': 1.9628070580074564, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 0.58}) +[2026-06-23 06:43:59,947][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.11404845863580704, 'train/grad_norm_pre_clip': 13.615734100341797, 'timing/data': 0.0013935549941379577, 'timing/model': 1.9663833860249724, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 0.58}) +[2026-06-23 06:44:03,902][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.23218685388565063, 'train/grad_norm_pre_clip': 19.006757736206055, 'timing/data': 0.05483247700612992, 'timing/model': 1.9732173520023935, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 0.58}) +[2026-06-23 06:44:07,813][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.07453326135873795, 'train/grad_norm_pre_clip': 13.648642539978027, 'timing/data': 0.0012084929912816733, 'timing/model': 1.9769291210104711, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 0.58}) +[2026-06-23 06:44:11,814][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.3281402289867401, 'train/grad_norm_pre_clip': 30.60085105895996, 'timing/data': 0.0012012689840048552, 'timing/model': 1.9760115449898876, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.58}) +[2026-06-23 06:44:15,940][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.14745618402957916, 'train/grad_norm_pre_clip': 24.02777099609375, 'timing/data': 0.0013954579771962017, 'timing/model': 2.1848489599942695, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 0.58}) +[2026-06-23 06:44:19,933][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.0776355192065239, 'train/grad_norm_pre_clip': 11.615030288696289, 'timing/data': 0.055839017004473135, 'timing/model': 2.003101546986727, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 0.59}) +[2026-06-23 06:44:23,857][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.229075089097023, 'train/grad_norm_pre_clip': 44.451385498046875, 'timing/data': 0.0014223880134522915, 'timing/model': 1.9780835050041787, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 0.59}) +[2026-06-23 06:44:27,866][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.1777891367673874, 'train/grad_norm_pre_clip': 36.12479782104492, 'timing/data': 0.001233359012985602, 'timing/model': 1.982131715019932, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 0.59}) +[2026-06-23 06:44:31,788][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.08302947878837585, 'train/grad_norm_pre_clip': 25.93578338623047, 'timing/data': 0.0013009039976168424, 'timing/model': 1.9827327959937975, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 0.59}) +[2026-06-23 06:44:35,788][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.0961361676454544, 'train/grad_norm_pre_clip': 17.4959774017334, 'timing/data': 0.06786680899676867, 'timing/model': 1.9875516390020493, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 0.59}) +[2026-06-23 06:44:39,723][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.02915840782225132, 'train/grad_norm_pre_clip': 5.557948112487793, 'timing/data': 0.0012556699803099036, 'timing/model': 1.9874806979787536, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 0.59}) +[2026-06-23 06:44:43,746][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.057266104966402054, 'train/grad_norm_pre_clip': 12.003564834594727, 'timing/data': 0.0012085630150977522, 'timing/model': 1.9882456630002707, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 0.59}) +[2026-06-23 06:44:47,758][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.14056803286075592, 'train/grad_norm_pre_clip': 19.39130210876465, 'timing/data': 0.0014021609968040138, 'timing/model': 2.063498649018584, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 0.59}) +[2026-06-23 06:44:51,754][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.08092425018548965, 'train/grad_norm_pre_clip': 17.702600479125977, 'timing/data': 0.05396913201548159, 'timing/model': 1.9930955739982892, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.59}) +[2026-06-23 06:44:55,705][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.17567948997020721, 'train/grad_norm_pre_clip': 24.82862091064453, 'timing/data': 0.001385440002195537, 'timing/model': 1.9962828129937407, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.59}) +[2026-06-23 06:44:59,740][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.08890971541404724, 'train/grad_norm_pre_clip': 16.035799026489258, 'timing/data': 0.0012124500062782317, 'timing/model': 1.9947143349854741, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 0.59}) +[2026-06-23 06:45:03,706][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.04108040779829025, 'train/grad_norm_pre_clip': 9.740760803222656, 'timing/data': 0.00141815998358652, 'timing/model': 2.0124675770057365, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 0.59}) +[2026-06-23 06:45:07,708][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.1113043799996376, 'train/grad_norm_pre_clip': 18.839632034301758, 'timing/data': 0.053991042979760095, 'timing/model': 1.9961107959970832, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 0.59}) +[2026-06-23 06:45:11,674][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.13589397072792053, 'train/grad_norm_pre_clip': 22.254728317260742, 'timing/data': 0.001192783995065838, 'timing/model': 2.0106152050138917, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 0.59}) +[2026-06-23 06:45:15,720][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.12386295944452286, 'train/grad_norm_pre_clip': 22.075061798095703, 'timing/data': 0.0013288049958646297, 'timing/model': 1.9980185519962106, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 0.59}) +[2026-06-23 06:45:19,671][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.04228881374001503, 'train/grad_norm_pre_clip': 18.896116256713867, 'timing/data': 0.0012319069937802851, 'timing/model': 1.9960792589990888, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 0.59}) +[2026-06-23 06:45:23,756][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.09573052078485489, 'train/grad_norm_pre_clip': 30.636770248413086, 'timing/data': 0.06409440000425093, 'timing/model': 1.9987532899831422, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 0.59}) +[2026-06-23 06:45:27,716][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.23850074410438538, 'train/grad_norm_pre_clip': 27.727909088134766, 'timing/data': 0.001195028016809374, 'timing/model': 2.001442018983653, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 0.59}) +[2026-06-23 06:45:31,755][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.3379925489425659, 'train/grad_norm_pre_clip': 38.1573600769043, 'timing/data': 0.0015639199991710484, 'timing/model': 1.9991587499971502, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.59}) +[2026-06-23 06:45:35,713][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.12722381949424744, 'train/grad_norm_pre_clip': 23.664798736572266, 'timing/data': 0.0014062180125620216, 'timing/model': 1.9997379500127863, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 0.59}) +[2026-06-23 06:45:39,721][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.06456643342971802, 'train/grad_norm_pre_clip': 13.690825462341309, 'timing/data': 0.04969506597262807, 'timing/model': 2.0000622189836577, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 0.59}) +[2026-06-23 06:45:43,710][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.1620308756828308, 'train/grad_norm_pre_clip': 19.63052749633789, 'timing/data': 0.0012109780218452215, 'timing/model': 2.0250966630119365, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 0.59}) +[2026-06-23 06:45:47,765][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.07983963191509247, 'train/grad_norm_pre_clip': 15.681981086730957, 'timing/data': 0.0011077870149165392, 'timing/model': 2.002901275001932, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 0.59}) +[2026-06-23 06:45:51,732][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.12567158043384552, 'train/grad_norm_pre_clip': 20.0408992767334, 'timing/data': 0.001176854013465345, 'timing/model': 2.0052561369957402, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.59}) +[2026-06-23 06:45:55,927][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.10552316159009933, 'train/grad_norm_pre_clip': 14.094792366027832, 'timing/data': 0.055068493995349854, 'timing/model': 2.177676708990475, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 0.59}) +[2026-06-23 06:45:59,900][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.0934092327952385, 'train/grad_norm_pre_clip': 32.48126220703125, 'timing/data': 0.0013922820216976106, 'timing/model': 2.0068324399762787, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 0.59}) +[2026-06-23 06:46:03,956][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.1880112737417221, 'train/grad_norm_pre_clip': 43.072303771972656, 'timing/data': 0.0011064349964726716, 'timing/model': 2.0072988719912246, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 0.6}) +[2026-06-23 06:46:07,930][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.21921126544475555, 'train/grad_norm_pre_clip': 52.536373138427734, 'timing/data': 0.0011574590171221644, 'timing/model': 2.0090214459924027, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 0.6}) +[2026-06-23 06:46:11,961][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.2219233214855194, 'train/grad_norm_pre_clip': 33.413909912109375, 'timing/data': 0.055781001021387056, 'timing/model': 2.0086312649946194, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.6}) +[2026-06-23 06:46:15,957][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.0542474091053009, 'train/grad_norm_pre_clip': 12.617470741271973, 'timing/data': 0.0012119390012230724, 'timing/model': 2.028122276009526, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 0.6}) +[2026-06-23 06:46:20,038][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.27226462960243225, 'train/grad_norm_pre_clip': 20.073575973510742, 'timing/data': 0.0014253839908633381, 'timing/model': 2.0166853060072754, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 0.6}) +[2026-06-23 06:46:24,044][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.09441786259412766, 'train/grad_norm_pre_clip': 13.74993896484375, 'timing/data': 0.0013134070031810552, 'timing/model': 2.0315385090070777, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 0.6}) +[2026-06-23 06:46:28,098][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.07951635867357254, 'train/grad_norm_pre_clip': 19.03539276123047, 'timing/data': 0.05654355901060626, 'timing/model': 2.0254495099943597, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 0.6}) +[2026-06-23 06:46:32,341][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.17844977974891663, 'train/grad_norm_pre_clip': 21.693580627441406, 'timing/data': 0.0014260939788073301, 'timing/model': 2.021189619001234, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 0.6}) +[2026-06-23 06:46:36,593][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.13413386046886444, 'train/grad_norm_pre_clip': 34.20808410644531, 'timing/data': 0.001226745982421562, 'timing/model': 2.032189772988204, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 0.6}) +[2026-06-23 06:46:40,623][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.12132839113473892, 'train/grad_norm_pre_clip': 15.490803718566895, 'timing/data': 0.0016756659897509962, 'timing/model': 2.056519223988289, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 0.6}) +[2026-06-23 06:46:44,939][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.12929794192314148, 'train/grad_norm_pre_clip': 24.185253143310547, 'timing/data': 0.31154182998579927, 'timing/model': 2.0320922840037383, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 0.6}) +[2026-06-23 06:46:48,963][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.05911211669445038, 'train/grad_norm_pre_clip': 7.116920471191406, 'timing/data': 0.001858473988249898, 'timing/model': 2.032210653007496, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 0.6}) +[2026-06-23 06:46:53,037][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.091701440513134, 'train/grad_norm_pre_clip': 17.05522918701172, 'timing/data': 0.0011999369889963418, 'timing/model': 2.0170855580072384, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.6}) +[2026-06-23 06:46:57,039][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.05977758392691612, 'train/grad_norm_pre_clip': 16.006486892700195, 'timing/data': 0.0012199539924040437, 'timing/model': 2.024645164987305, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 0.6}) +[2026-06-23 06:47:01,088][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.050567056983709335, 'train/grad_norm_pre_clip': 6.6541595458984375, 'timing/data': 0.057533929008059204, 'timing/model': 2.0181426419876516, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 0.6}) +[2026-06-23 06:47:05,116][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.14122749865055084, 'train/grad_norm_pre_clip': 15.800673484802246, 'timing/data': 0.0012539669987745583, 'timing/model': 2.0175134299788624, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 0.6}) +[2026-06-23 06:47:09,205][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.09029706567525864, 'train/grad_norm_pre_clip': 19.707378387451172, 'timing/data': 0.0013418089947663248, 'timing/model': 2.024594613001682, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 0.6}) +[2026-06-23 06:47:13,212][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.2425549179315567, 'train/grad_norm_pre_clip': 28.122663497924805, 'timing/data': 0.0012454119860194623, 'timing/model': 2.0311658539867494, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 0.6}) +[2026-06-23 06:47:17,257][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.11206088960170746, 'train/grad_norm_pre_clip': 15.490744590759277, 'timing/data': 0.056998871004907414, 'timing/model': 2.018344526004512, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 0.6}) +[2026-06-23 06:47:21,263][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.05536729469895363, 'train/grad_norm_pre_clip': 16.73457908630371, 'timing/data': 0.0012356030056253076, 'timing/model': 2.019530838006176, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 0.6}) +[2026-06-23 06:47:25,350][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.11930475383996964, 'train/grad_norm_pre_clip': 15.370553016662598, 'timing/data': 0.0012337189982645214, 'timing/model': 2.0234391589765437, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 0.6}) +[2026-06-23 06:47:29,379][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.15760935842990875, 'train/grad_norm_pre_clip': 17.26533317565918, 'timing/data': 0.0014097450184635818, 'timing/model': 2.041928380000172, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 0.6}) +[2026-06-23 06:47:33,458][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.07587994635105133, 'train/grad_norm_pre_clip': 14.787182807922363, 'timing/data': 0.06532480401801877, 'timing/model': 2.0287161800079048, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.6}) +[2026-06-23 06:47:37,588][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.19342085719108582, 'train/grad_norm_pre_clip': 22.395971298217773, 'timing/data': 0.001258123986190185, 'timing/model': 2.1547861360013485, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 0.6}) +[2026-06-23 06:47:41,668][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.056839246302843094, 'train/grad_norm_pre_clip': 14.322583198547363, 'timing/data': 0.0015090079978108406, 'timing/model': 2.018821117992047, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 0.6}) +[2026-06-23 06:47:45,679][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.20151054859161377, 'train/grad_norm_pre_clip': 41.19330978393555, 'timing/data': 0.0011493840138427913, 'timing/model': 2.037528521992499, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 0.6}) +[2026-06-23 06:47:49,734][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.08157035708427429, 'train/grad_norm_pre_clip': 18.03545379638672, 'timing/data': 0.06448086400632747, 'timing/model': 2.017265424976358, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 0.61}) +[2026-06-23 06:47:53,732][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.16346319019794464, 'train/grad_norm_pre_clip': 19.823034286499023, 'timing/data': 0.0012425860040821135, 'timing/model': 2.0210082499834243, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 0.61}) +[2026-06-23 06:47:57,815][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.14631478488445282, 'train/grad_norm_pre_clip': 20.04464340209961, 'timing/data': 0.0011993260122835636, 'timing/model': 2.0168551570095588, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 0.61}) +[2026-06-23 06:48:01,816][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.08002422749996185, 'train/grad_norm_pre_clip': 15.253928184509277, 'timing/data': 0.0011547629837878048, 'timing/model': 2.025530776998494, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 0.61}) +[2026-06-23 06:48:05,869][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.12987291812896729, 'train/grad_norm_pre_clip': 18.821048736572266, 'timing/data': 0.05581608501961455, 'timing/model': 2.020945625990862, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 0.61}) +[2026-06-23 06:48:09,869][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.06913452595472336, 'train/grad_norm_pre_clip': 14.520567893981934, 'timing/data': 0.0014220369921531528, 'timing/model': 2.019629754999187, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 0.61}) +[2026-06-23 06:48:14,234][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.08002567291259766, 'train/grad_norm_pre_clip': 25.59150505065918, 'timing/data': 0.0013497230247594416, 'timing/model': 2.038707571977284, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.61}) +[2026-06-23 06:48:18,258][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.16014403104782104, 'train/grad_norm_pre_clip': 111.92591094970703, 'timing/data': 0.0012679029896389693, 'timing/model': 2.0456774210033473, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 0.61}) +[2026-06-23 06:48:22,306][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.1265927255153656, 'train/grad_norm_pre_clip': 20.35584831237793, 'timing/data': 0.051299640006618574, 'timing/model': 2.022130286000902, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 0.61}) +[2026-06-23 06:48:26,306][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.04508199915289879, 'train/grad_norm_pre_clip': 9.193840980529785, 'timing/data': 0.0011436529748607427, 'timing/model': 2.0218862839974463, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 0.61}) +[2026-06-23 06:48:30,404][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.08704972267150879, 'train/grad_norm_pre_clip': 24.682817459106445, 'timing/data': 0.0012063090107403696, 'timing/model': 2.0343138559837826, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 0.61}) +[2026-06-23 06:48:34,425][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.17418961226940155, 'train/grad_norm_pre_clip': 42.44535446166992, 'timing/data': 0.0013042100181337446, 'timing/model': 2.0313062299974263, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 0.61}) +[2026-06-23 06:48:38,693][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.0649312436580658, 'train/grad_norm_pre_clip': 15.074814796447754, 'timing/data': 0.0918719410256017, 'timing/model': 2.0719817109929863, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 0.61}) +[2026-06-23 06:48:42,704][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.05891108885407448, 'train/grad_norm_pre_clip': 14.520811080932617, 'timing/data': 0.001345074997516349, 'timing/model': 2.03363050398184, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 0.61}) +[2026-06-23 06:48:46,848][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.17268502712249756, 'train/grad_norm_pre_clip': 40.13694763183594, 'timing/data': 0.0012311750033404678, 'timing/model': 2.017279283987591, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 0.61}) +[2026-06-23 06:48:50,841][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.141595259308815, 'train/grad_norm_pre_clip': 24.489477157592773, 'timing/data': 0.001277180010220036, 'timing/model': 2.0175550740095787, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 0.61}) +[2026-06-23 06:48:54,909][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.18766869604587555, 'train/grad_norm_pre_clip': 30.540348052978516, 'timing/data': 0.07074388497858308, 'timing/model': 2.0215396959974896, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.61}) +[2026-06-23 06:48:58,920][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.11870142072439194, 'train/grad_norm_pre_clip': 22.988178253173828, 'timing/data': 0.0012629239936359227, 'timing/model': 2.0338548200088553, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 0.61}) +[2026-06-23 06:49:03,000][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.24128463864326477, 'train/grad_norm_pre_clip': 46.47679901123047, 'timing/data': 0.0012182600039523095, 'timing/model': 2.0178646870190278, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 0.61}) +[2026-06-23 06:49:06,993][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.18581418693065643, 'train/grad_norm_pre_clip': 55.07926940917969, 'timing/data': 0.001262953010154888, 'timing/model': 2.0182690850051586, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 0.61}) +[2026-06-23 06:49:11,042][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.03720425069332123, 'train/grad_norm_pre_clip': 14.292001724243164, 'timing/data': 0.051549322000937536, 'timing/model': 2.0226622609770857, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 0.61}) +[2026-06-23 06:49:15,054][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.1834372878074646, 'train/grad_norm_pre_clip': 40.224708557128906, 'timing/data': 0.001259547017980367, 'timing/model': 2.0312321249803063, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 0.61}) +[2026-06-23 06:49:19,288][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.06538237631320953, 'train/grad_norm_pre_clip': 17.757579803466797, 'timing/data': 0.0013671760098077357, 'timing/model': 2.1719474800047465, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 0.61}) +[2026-06-23 06:49:23,288][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.32695236802101135, 'train/grad_norm_pre_clip': 53.9265251159668, 'timing/data': 0.001241013000253588, 'timing/model': 2.0229865809960756, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 0.61}) +[2026-06-23 06:49:27,358][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.13081151247024536, 'train/grad_norm_pre_clip': 28.59833526611328, 'timing/data': 0.06675569800427184, 'timing/model': 2.0220591780089308, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 0.61}) +[2026-06-23 06:49:31,353][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.1655322164297104, 'train/grad_norm_pre_clip': 39.54051971435547, 'timing/data': 0.0014575830136891454, 'timing/model': 2.018194988981122, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 0.61}) +[2026-06-23 06:49:35,436][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.2717256546020508, 'train/grad_norm_pre_clip': 39.16618728637695, 'timing/data': 0.0012177600001450628, 'timing/model': 2.0179847799881827, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.62}) +[2026-06-23 06:49:39,440][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.029419658705592155, 'train/grad_norm_pre_clip': 14.8331298828125, 'timing/data': 0.0014253240078687668, 'timing/model': 2.02656521400786, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 0.62}) +[2026-06-23 06:49:43,489][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.4695948362350464, 'train/grad_norm_pre_clip': 74.92855072021484, 'timing/data': 0.051890792994527146, 'timing/model': 2.019887075992301, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 0.62}) +[2026-06-23 06:49:47,491][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.17929305136203766, 'train/grad_norm_pre_clip': 40.872772216796875, 'timing/data': 0.0012077310238964856, 'timing/model': 2.0253219369915314, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 0.62}) +[2026-06-23 06:49:51,571][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.26454678177833557, 'train/grad_norm_pre_clip': 41.95501708984375, 'timing/data': 0.0012059079890605062, 'timing/model': 2.021026772010373, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 0.62}) +[2026-06-23 06:49:55,810][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.5468974113464355, 'train/grad_norm_pre_clip': 70.24632263183594, 'timing/data': 0.0011563269945327193, 'timing/model': 2.0178239149972796, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 0.62}) +[2026-06-23 06:49:59,865][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.2146838754415512, 'train/grad_norm_pre_clip': 45.44389343261719, 'timing/data': 0.062785182002699, 'timing/model': 2.0164604659948964, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 0.62}) +[2026-06-23 06:50:03,859][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.1590103656053543, 'train/grad_norm_pre_clip': 25.252117156982422, 'timing/data': 0.0012711289746221155, 'timing/model': 2.019064416992478, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 0.62}) +[2026-06-23 06:50:07,938][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.16281510889530182, 'train/grad_norm_pre_clip': 46.51639938354492, 'timing/data': 0.0011156620166730136, 'timing/model': 2.022033553017536, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 0.62}) +[2026-06-23 06:50:11,953][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.0825759544968605, 'train/grad_norm_pre_clip': 25.58698844909668, 'timing/data': 0.0011523699795361608, 'timing/model': 2.03963952101185, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 0.62}) +[2026-06-23 06:50:16,016][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1600 +[2026-06-23 06:52:09,904][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1600 +[2026-06-23 06:52:09,904][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1600 +[2026-06-23 06:54:17,509][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1600 +[2026-06-23 06:54:17,511][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.10700388252735138, 'train/grad_norm_pre_clip': 20.857006072998047, 'eval/action_loss/seconds': 113.88415672801784, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.1117619127035141, 'eval/latency_2/loss': 0.1117619127035141, 'eval/flappy/loss': 0.1117619127035141, 'eval/flappy/latency_2/loss': 0.1117619127035141, 'eval/flap_precision': 0.7947198275862069, 'eval/flap_recall': 0.99527665317139, 'eval/flap_f1': 0.883762732174955, 'eval/flap_rate_model': 0.2577777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8837627321749552, 'eval/latency_2/flap_precision': 0.7947198275862069, 'eval/latency_2/flap_recall': 0.99527665317139, 'eval/latency_2/flap_f1': 0.883762732174955, 'eval/latency_2/flap_rate_model': 0.2577777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8837627321749552, 'eval/action_classification/data_wait_seconds': 50.403839111328125, 'eval/action_classification/predict_seconds': 75.39453125, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06462210501194932, 'timing/model': 2.0201799280184787, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.62}) +[2026-06-23 06:54:21,472][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.08103293925523758, 'train/grad_norm_pre_clip': 13.783414840698242, 'timing/data': 0.0012231699947733432, 'timing/model': 1.9987189380044583, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 0.62}) +[2026-06-23 06:54:25,508][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.1624104380607605, 'train/grad_norm_pre_clip': 33.14549255371094, 'timing/data': 0.0011760430061258376, 'timing/model': 1.995685803995002, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 0.62}) +[2026-06-23 06:54:29,461][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.18533027172088623, 'train/grad_norm_pre_clip': 40.023521423339844, 'timing/data': 0.0013589109876193106, 'timing/model': 2.003095873980783, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 0.62}) +[2026-06-23 06:54:33,487][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.04646241292357445, 'train/grad_norm_pre_clip': 7.950466632843018, 'timing/data': 0.06792316498467699, 'timing/model': 2.003192142001353, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 0.62}) +[2026-06-23 06:54:37,460][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.06347513943910599, 'train/grad_norm_pre_clip': 9.004789352416992, 'timing/data': 0.001391932019032538, 'timing/model': 2.0168951800151262, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.62}) +[2026-06-23 06:54:41,506][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.06508861482143402, 'train/grad_norm_pre_clip': 8.333857536315918, 'timing/data': 0.0011683590128086507, 'timing/model': 1.9978980809974018, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 0.62}) +[2026-06-23 06:54:45,466][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.19009748101234436, 'train/grad_norm_pre_clip': 35.40725326538086, 'timing/data': 0.0012336400104686618, 'timing/model': 2.0051025230204687, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 0.62}) +[2026-06-23 06:54:49,487][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.27581191062927246, 'train/grad_norm_pre_clip': 57.96173095703125, 'timing/data': 0.06535844600875862, 'timing/model': 1.996673068992095, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 0.62}) +[2026-06-23 06:54:53,624][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.1340084820985794, 'train/grad_norm_pre_clip': 31.783615112304688, 'timing/data': 0.001200358005007729, 'timing/model': 2.020173158001853, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 0.62}) +[2026-06-23 06:54:57,688][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.3575853407382965, 'train/grad_norm_pre_clip': 69.81025695800781, 'timing/data': 0.001280024996958673, 'timing/model': 2.008149883011356, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.62}) +[2026-06-23 06:55:01,667][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.225009948015213, 'train/grad_norm_pre_clip': 35.89169692993164, 'timing/data': 0.0015645909879822284, 'timing/model': 2.012628950004, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 0.62}) +[2026-06-23 06:55:05,703][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.08282554894685745, 'train/grad_norm_pre_clip': 15.780324935913086, 'timing/data': 0.06784753501415253, 'timing/model': 2.0012685300025623, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 0.62}) +[2026-06-23 06:55:09,667][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.14085033535957336, 'train/grad_norm_pre_clip': 21.73891830444336, 'timing/data': 0.0011253400007262826, 'timing/model': 2.003368859994225, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 0.62}) +[2026-06-23 06:55:13,737][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.1319807767868042, 'train/grad_norm_pre_clip': 22.8795223236084, 'timing/data': 0.0015878530102781951, 'timing/model': 2.0204001870006323, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 0.62}) +[2026-06-23 06:55:17,732][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.027351705357432365, 'train/grad_norm_pre_clip': 6.044266223907471, 'timing/data': 0.0012522839824669063, 'timing/model': 2.0277326960058417, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 0.63}) +[2026-06-23 06:55:21,757][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.1948188990354538, 'train/grad_norm_pre_clip': 30.83535385131836, 'timing/data': 0.05705735899391584, 'timing/model': 2.002040527004283, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 0.63}) +[2026-06-23 06:55:25,765][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.0879630371928215, 'train/grad_norm_pre_clip': 17.940528869628906, 'timing/data': 0.0011938960233237594, 'timing/model': 2.0437734729785006, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 0.63}) +[2026-06-23 06:55:29,830][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.07406340539455414, 'train/grad_norm_pre_clip': 10.799385070800781, 'timing/data': 0.0012067700154148042, 'timing/model': 2.010537445981754, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 0.63}) +[2026-06-23 06:55:33,807][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.1754395067691803, 'train/grad_norm_pre_clip': 30.79599952697754, 'timing/data': 0.0012203349906485528, 'timing/model': 2.012994916003663, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.63}) +[2026-06-23 06:55:37,836][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.06441423296928406, 'train/grad_norm_pre_clip': 13.264373779296875, 'timing/data': 0.05525500900694169, 'timing/model': 2.0081145989825018, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.63}) +[2026-06-23 06:55:41,817][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.09788688272237778, 'train/grad_norm_pre_clip': 14.49154281616211, 'timing/data': 0.0012047560012433678, 'timing/model': 2.010034397011623, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 0.63}) +[2026-06-23 06:55:45,885][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.06604300439357758, 'train/grad_norm_pre_clip': 14.015124320983887, 'timing/data': 0.001350305014057085, 'timing/model': 2.007370033999905, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 0.63}) +[2026-06-23 06:55:49,873][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.06129191070795059, 'train/grad_norm_pre_clip': 15.618704795837402, 'timing/data': 0.0012520540039986372, 'timing/model': 2.0213809210108593, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 0.63}) +[2026-06-23 06:55:53,939][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.11356092989444733, 'train/grad_norm_pre_clip': 18.38144874572754, 'timing/data': 0.055532070982735604, 'timing/model': 2.0112146380124614, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 0.63}) +[2026-06-23 06:55:57,919][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.08511535078287125, 'train/grad_norm_pre_clip': 16.197710037231445, 'timing/data': 0.0013268309994600713, 'timing/model': 2.009365111007355, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 0.63}) +[2026-06-23 06:56:02,164][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.07719744741916656, 'train/grad_norm_pre_clip': 26.45639419555664, 'timing/data': 0.0015295360062737018, 'timing/model': 2.0176068879954983, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 0.63}) +[2026-06-23 06:56:06,147][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.1637466847896576, 'train/grad_norm_pre_clip': 19.101566314697266, 'timing/data': 0.001485604007029906, 'timing/model': 2.010492533998331, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 0.63}) +[2026-06-23 06:56:10,204][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.09501351416110992, 'train/grad_norm_pre_clip': 31.704891204833984, 'timing/data': 0.0662081360060256, 'timing/model': 2.0158880529925227, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 0.63}) +[2026-06-23 06:56:14,189][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.09343326091766357, 'train/grad_norm_pre_clip': 17.87030029296875, 'timing/data': 0.0012318759982008487, 'timing/model': 2.0158951060147956, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 0.63}) +[2026-06-23 06:56:18,286][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.021682675927877426, 'train/grad_norm_pre_clip': 5.8399977684021, 'timing/data': 0.0012547880178317428, 'timing/model': 2.032971345994156, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.63}) +[2026-06-23 06:56:22,282][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.17492058873176575, 'train/grad_norm_pre_clip': 29.290328979492188, 'timing/data': 0.0012691649899352342, 'timing/model': 2.024482240987709, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 0.63}) +[2026-06-23 06:56:26,325][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.1023702472448349, 'train/grad_norm_pre_clip': 18.917991638183594, 'timing/data': 0.05339328901027329, 'timing/model': 2.0206522960215807, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 0.63}) +[2026-06-23 06:56:30,327][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.1554378718137741, 'train/grad_norm_pre_clip': 32.22145080566406, 'timing/data': 0.0012210269924253225, 'timing/model': 2.0227554810117, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 0.63}) +[2026-06-23 06:56:34,676][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.11460647732019424, 'train/grad_norm_pre_clip': 19.286649703979492, 'timing/data': 0.001421346009010449, 'timing/model': 2.0192438939993735, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 0.63}) +[2026-06-23 06:56:38,667][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.15002532303333282, 'train/grad_norm_pre_clip': 24.986568450927734, 'timing/data': 0.001264927996089682, 'timing/model': 2.018498405988794, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 0.63}) +[2026-06-23 06:56:42,730][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.15105238556861877, 'train/grad_norm_pre_clip': 17.89168357849121, 'timing/data': 0.06540923999273218, 'timing/model': 2.021621938009048, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 0.63}) +[2026-06-23 06:56:46,722][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.1557958424091339, 'train/grad_norm_pre_clip': 24.97730255126953, 'timing/data': 0.001260999997612089, 'timing/model': 2.0174039950070437, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 0.63}) +[2026-06-23 06:56:50,822][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.09649740159511566, 'train/grad_norm_pre_clip': 15.166874885559082, 'timing/data': 0.0015568369999527931, 'timing/model': 2.031211795983836, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 0.63}) +[2026-06-23 06:56:54,815][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.07176893204450607, 'train/grad_norm_pre_clip': 26.054853439331055, 'timing/data': 0.0013262299762573093, 'timing/model': 2.0181117320025805, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 0.63}) +[2026-06-23 06:56:58,883][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.18145312368869781, 'train/grad_norm_pre_clip': 28.40496253967285, 'timing/data': 0.06843847699929029, 'timing/model': 2.020337975001894, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.63}) +[2026-06-23 06:57:02,883][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.08258265256881714, 'train/grad_norm_pre_clip': 24.98301124572754, 'timing/data': 0.001246873987838626, 'timing/model': 2.018882066011429, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 0.64}) +[2026-06-23 06:57:07,006][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.05773939564824104, 'train/grad_norm_pre_clip': 15.435896873474121, 'timing/data': 0.001212650997331366, 'timing/model': 2.0525561019894667, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 0.64}) +[2026-06-23 06:57:11,018][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.09494201093912125, 'train/grad_norm_pre_clip': 26.77263641357422, 'timing/data': 0.001221957994857803, 'timing/model': 2.036234806990251, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 0.64}) +[2026-06-23 06:57:15,079][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.10458812862634659, 'train/grad_norm_pre_clip': 27.296480178833008, 'timing/data': 0.05799662400386296, 'timing/model': 2.023472588014556, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 0.64}) +[2026-06-23 06:57:19,097][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.08896396309137344, 'train/grad_norm_pre_clip': 35.39753341674805, 'timing/data': 0.0013373210094869137, 'timing/model': 2.038551648001885, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 0.64}) +[2026-06-23 06:57:23,192][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.09662893414497375, 'train/grad_norm_pre_clip': 18.14167022705078, 'timing/data': 0.0014563809963874519, 'timing/model': 2.0222321260080207, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 0.64}) +[2026-06-23 06:57:27,192][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.19309107959270477, 'train/grad_norm_pre_clip': 36.260887145996094, 'timing/data': 0.0012388280010782182, 'timing/model': 2.0215966609830502, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 0.64}) +[2026-06-23 06:57:31,254][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.10927333682775497, 'train/grad_norm_pre_clip': 30.69637107849121, 'timing/data': 0.061917387996800244, 'timing/model': 2.0196947979857214, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 0.64}) +[2026-06-23 06:57:35,254][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.1777799278497696, 'train/grad_norm_pre_clip': 36.160579681396484, 'timing/data': 0.001449177012545988, 'timing/model': 2.0204479100066237, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 0.64}) +[2026-06-23 06:57:39,339][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.20676100254058838, 'train/grad_norm_pre_clip': 40.75261688232422, 'timing/data': 0.0012209459964651614, 'timing/model': 2.019760438997764, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.64}) +[2026-06-23 06:57:43,528][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.05279506742954254, 'train/grad_norm_pre_clip': 12.25882339477539, 'timing/data': 0.0014272969856392592, 'timing/model': 2.210285192995798, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 0.64}) +[2026-06-23 06:57:47,595][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.10662611573934555, 'train/grad_norm_pre_clip': 15.610416412353516, 'timing/data': 0.06760423598461784, 'timing/model': 2.0201321359782014, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 0.64}) +[2026-06-23 06:57:51,600][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.1795780062675476, 'train/grad_norm_pre_clip': 31.069293975830078, 'timing/data': 0.0012519130250439048, 'timing/model': 2.024932214990258, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 0.64}) +[2026-06-23 06:57:55,695][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.2840847373008728, 'train/grad_norm_pre_clip': 24.999755859375, 'timing/data': 0.001242504979018122, 'timing/model': 2.0238453570054844, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 0.64}) +[2026-06-23 06:57:59,717][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.11930152028799057, 'train/grad_norm_pre_clip': 29.14978790283203, 'timing/data': 0.001295903988648206, 'timing/model': 2.0394828189746477, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 0.64}) +[2026-06-23 06:58:03,777][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.05900217592716217, 'train/grad_norm_pre_clip': 14.35704517364502, 'timing/data': 0.05516675699618645, 'timing/model': 2.026054107991513, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 0.64}) +[2026-06-23 06:58:07,791][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.09168844670057297, 'train/grad_norm_pre_clip': 14.1827392578125, 'timing/data': 0.0012661800137721002, 'timing/model': 2.034948020998854, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 0.64}) +[2026-06-23 06:58:11,883][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.08881884813308716, 'train/grad_norm_pre_clip': 12.809303283691406, 'timing/data': 0.0012585649965330958, 'timing/model': 2.0188538650108967, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 0.64}) +[2026-06-23 06:58:15,923][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.0936756357550621, 'train/grad_norm_pre_clip': 23.925764083862305, 'timing/data': 0.0012690250005107373, 'timing/model': 2.066675310023129, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 0.64}) +[2026-06-23 06:58:19,989][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.11075634509325027, 'train/grad_norm_pre_clip': 60.529075622558594, 'timing/data': 0.07249206601409242, 'timing/model': 2.0162453560042195, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.64}) +[2026-06-23 06:58:23,992][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.09595544636249542, 'train/grad_norm_pre_clip': 20.000131607055664, 'timing/data': 0.0012199240154586732, 'timing/model': 2.0240077580092475, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 0.64}) +[2026-06-23 06:58:28,079][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.0902586355805397, 'train/grad_norm_pre_clip': 10.724218368530273, 'timing/data': 0.0011932050110772252, 'timing/model': 2.0166182850080077, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 0.64}) +[2026-06-23 06:58:32,089][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.05306054651737213, 'train/grad_norm_pre_clip': 18.877180099487305, 'timing/data': 0.0012657789920922369, 'timing/model': 2.034485657000914, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 0.64}) +[2026-06-23 06:58:36,154][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.16278615593910217, 'train/grad_norm_pre_clip': 25.242000579833984, 'timing/data': 0.07327934101340361, 'timing/model': 2.0181412490201183, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 0.64}) +[2026-06-23 06:58:40,155][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.1375388503074646, 'train/grad_norm_pre_clip': 37.71364212036133, 'timing/data': 0.0012668610143009573, 'timing/model': 2.0214733250031713, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 0.64}) +[2026-06-23 06:58:44,265][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.08248823136091232, 'train/grad_norm_pre_clip': 11.837658882141113, 'timing/data': 0.001391750993207097, 'timing/model': 2.043090706982184, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 0.64}) +[2026-06-23 06:58:48,269][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.0823499783873558, 'train/grad_norm_pre_clip': 19.333908081054688, 'timing/data': 0.0012961839966010302, 'timing/model': 2.022838678007247, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 0.65}) +[2026-06-23 06:58:52,575][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.10097426176071167, 'train/grad_norm_pre_clip': 14.911015510559082, 'timing/data': 0.056934011983685195, 'timing/model': 2.0266684039961547, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.65}) +[2026-06-23 06:58:56,590][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.08843529969453812, 'train/grad_norm_pre_clip': 15.346521377563477, 'timing/data': 0.0012171890120953321, 'timing/model': 2.0389358490065206, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 0.65}) +[2026-06-23 06:59:00,683][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.17062604427337646, 'train/grad_norm_pre_clip': 37.1054801940918, 'timing/data': 0.0012272369931451976, 'timing/model': 2.0221016769937705, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.65}) +[2026-06-23 06:59:04,689][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.1680225282907486, 'train/grad_norm_pre_clip': 17.56697654724121, 'timing/data': 0.0013056629977654666, 'timing/model': 2.021494304994121, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 0.65}) +[2026-06-23 06:59:08,757][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.17412693798542023, 'train/grad_norm_pre_clip': 37.54414367675781, 'timing/data': 0.06892819199129008, 'timing/model': 2.0193650809815153, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 0.65}) +[2026-06-23 06:59:12,759][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.01640690304338932, 'train/grad_norm_pre_clip': 6.29587459564209, 'timing/data': 0.0013794589904136956, 'timing/model': 2.024308463995112, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 0.65}) +[2026-06-23 06:59:16,849][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.038201704621315, 'train/grad_norm_pre_clip': 15.472860336303711, 'timing/data': 0.0012090340023860335, 'timing/model': 2.0208134170097765, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 0.65}) +[2026-06-23 06:59:20,847][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.18673409521579742, 'train/grad_norm_pre_clip': 40.741127014160156, 'timing/data': 0.0012381970009300858, 'timing/model': 2.0228187419997994, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.65}) +[2026-06-23 06:59:24,964][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.1024746522307396, 'train/grad_norm_pre_clip': 24.3080997467041, 'timing/data': 0.05765582498861477, 'timing/model': 2.018012432003161, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 0.65}) +[2026-06-23 06:59:28,980][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.08554655313491821, 'train/grad_norm_pre_clip': 17.89970588684082, 'timing/data': 0.0012693159806076437, 'timing/model': 2.024527540022973, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 0.65}) +[2026-06-23 06:59:33,082][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.08178718388080597, 'train/grad_norm_pre_clip': 13.426566123962402, 'timing/data': 0.0011956490052398294, 'timing/model': 2.038430546002928, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 0.65}) +[2026-06-23 06:59:37,075][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.10018213838338852, 'train/grad_norm_pre_clip': 14.309656143188477, 'timing/data': 0.001298668998060748, 'timing/model': 2.018812471011188, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 0.65}) +[2026-06-23 06:59:41,148][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.09400469064712524, 'train/grad_norm_pre_clip': 20.24909782409668, 'timing/data': 0.0699934299918823, 'timing/model': 2.024043265002547, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.65}) +[2026-06-23 06:59:45,158][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.16295798122882843, 'train/grad_norm_pre_clip': 33.842262268066406, 'timing/data': 0.0013144680124241859, 'timing/model': 2.0225917329953518, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 0.65}) +[2026-06-23 06:59:49,252][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.07322809845209122, 'train/grad_norm_pre_clip': 17.242305755615234, 'timing/data': 0.0013542719825636595, 'timing/model': 2.022799627011409, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 0.65}) +[2026-06-23 06:59:53,264][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.09176801890134811, 'train/grad_norm_pre_clip': 30.88672637939453, 'timing/data': 0.00125013999058865, 'timing/model': 2.030681299016578, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 0.65}) +[2026-06-23 06:59:57,540][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.18860748410224915, 'train/grad_norm_pre_clip': 42.24153137207031, 'timing/data': 0.06738214599317871, 'timing/model': 2.2289689669851214, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 0.65}) +[2026-06-23 07:00:01,565][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.026332473382353783, 'train/grad_norm_pre_clip': 11.418489456176758, 'timing/data': 0.0014862359967082739, 'timing/model': 2.041746503993636, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 0.65}) +[2026-06-23 07:00:05,668][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.1616320163011551, 'train/grad_norm_pre_clip': 30.653118133544922, 'timing/data': 0.001257422991329804, 'timing/model': 2.035323496995261, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 0.65}) +[2026-06-23 07:00:09,672][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.08446795493364334, 'train/grad_norm_pre_clip': 14.78968334197998, 'timing/data': 0.001417349005350843, 'timing/model': 2.024876804993255, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 0.65}) +[2026-06-23 07:00:13,755][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.13314217329025269, 'train/grad_norm_pre_clip': 29.744049072265625, 'timing/data': 0.07045021399972029, 'timing/model': 2.0308016130002216, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 0.65}) +[2026-06-23 07:00:17,774][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.18934883177280426, 'train/grad_norm_pre_clip': 41.83048629760742, 'timing/data': 0.0012149940012022853, 'timing/model': 2.038081932987552, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 0.65}) +[2026-06-23 07:00:21,854][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.06486175209283829, 'train/grad_norm_pre_clip': 19.729612350463867, 'timing/data': 0.001131951023126021, 'timing/model': 2.0161697790026665, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.65}) +[2026-06-23 07:00:25,847][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.2185325026512146, 'train/grad_norm_pre_clip': 35.813838958740234, 'timing/data': 0.001139836007496342, 'timing/model': 2.017456054018112, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 0.65}) +[2026-06-23 07:00:29,983][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.20627281069755554, 'train/grad_norm_pre_clip': 34.19889831542969, 'timing/data': 0.10684357400168665, 'timing/model': 2.0324153839901555, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 0.65}) +[2026-06-23 07:00:34,222][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.24244347214698792, 'train/grad_norm_pre_clip': 45.152523040771484, 'timing/data': 0.0012043349852319807, 'timing/model': 2.0188940720108803, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 0.66}) +[2026-06-23 07:00:38,305][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.10121508687734604, 'train/grad_norm_pre_clip': 25.569416046142578, 'timing/data': 0.0013940749922767282, 'timing/model': 2.017072757007554, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 0.66}) +[2026-06-23 07:00:42,333][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.21643203496932983, 'train/grad_norm_pre_clip': 34.897186279296875, 'timing/data': 0.0015428210026584566, 'timing/model': 2.0473585029831156, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 0.66}) +[2026-06-23 07:00:46,399][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.059022992849349976, 'train/grad_norm_pre_clip': 10.71127700805664, 'timing/data': 0.06690047599840909, 'timing/model': 2.0220636180019937, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 0.66}) +[2026-06-23 07:00:50,429][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.09735429286956787, 'train/grad_norm_pre_clip': 20.3806209564209, 'timing/data': 0.001412559999153018, 'timing/model': 2.043526021996513, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 0.66}) +[2026-06-23 07:00:54,534][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.0713813304901123, 'train/grad_norm_pre_clip': 13.835086822509766, 'timing/data': 0.0013817030121572316, 'timing/model': 2.0374282560078427, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 0.66}) +[2026-06-23 07:00:58,540][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.061240002512931824, 'train/grad_norm_pre_clip': 10.80873966217041, 'timing/data': 0.0013031679845880717, 'timing/model': 2.0242948009981774, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 0.66}) +[2026-06-23 07:01:02,619][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1700 +[2026-06-23 07:02:56,769][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1700 +[2026-06-23 07:02:56,769][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1700 +[2026-06-23 07:05:03,864][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1700 +[2026-06-23 07:05:03,865][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.19103021919727325, 'train/grad_norm_pre_clip': 45.60739517211914, 'eval/action_loss/seconds': 114.14612993499031, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07107002288103104, 'eval/latency_2/loss': 0.07107002288103104, 'eval/flappy/loss': 0.07107002288103104, 'eval/flappy/latency_2/loss': 0.07107002288103104, 'eval/flap_precision': 0.9002493765586035, 'eval/flap_recall': 0.9743589743589743, 'eval/flap_f1': 0.9358392741412831, 'eval/flap_rate_model': 0.22277777777777777, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9358392741412832, 'eval/latency_2/flap_precision': 0.9002493765586035, 'eval/latency_2/flap_recall': 0.9743589743589743, 'eval/latency_2/flap_f1': 0.9358392741412831, 'eval/latency_2/flap_rate_model': 0.22277777777777777, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9358392741412832, 'eval/action_classification/data_wait_seconds': 50.39736557006836, 'eval/action_classification/predict_seconds': 74.9246597290039, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.07345134901697747, 'timing/model': 2.02505531598581, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.66}) +[2026-06-23 07:05:07,814][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.050233010202646255, 'train/grad_norm_pre_clip': 11.87298583984375, 'timing/data': 0.0014588350022677332, 'timing/model': 2.004508846992394, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 0.66}) +[2026-06-23 07:05:11,873][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.11043250560760498, 'train/grad_norm_pre_clip': 15.885677337646484, 'timing/data': 0.0014059570094104856, 'timing/model': 2.0129705919825938, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 0.66}) +[2026-06-23 07:05:15,827][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.10342399030923843, 'train/grad_norm_pre_clip': 17.047515869140625, 'timing/data': 0.0012171289999969304, 'timing/model': 2.000896532990737, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 0.66}) +[2026-06-23 07:05:19,845][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.09626492112874985, 'train/grad_norm_pre_clip': 16.72844886779785, 'timing/data': 0.06244792902725749, 'timing/model': 1.9971304849896114, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 0.66}) +[2026-06-23 07:05:23,803][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.09202932566404343, 'train/grad_norm_pre_clip': 23.674461364746094, 'timing/data': 0.001237435993971303, 'timing/model': 2.0026837149925996, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 0.66}) +[2026-06-23 07:05:28,136][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.1053084209561348, 'train/grad_norm_pre_clip': 13.487325668334961, 'timing/data': 0.0014917860098648816, 'timing/model': 2.2778124549950007, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 0.66}) +[2026-06-23 07:05:32,091][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.0556429848074913, 'train/grad_norm_pre_clip': 12.92392349243164, 'timing/data': 0.0015113519912119955, 'timing/model': 1.9976957189792302, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 0.66}) +[2026-06-23 07:05:36,119][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.06798320263624191, 'train/grad_norm_pre_clip': 13.650535583496094, 'timing/data': 0.06423265699413605, 'timing/model': 2.0093545309791807, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 0.66}) +[2026-06-23 07:05:40,093][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.17828142642974854, 'train/grad_norm_pre_clip': 16.862485885620117, 'timing/data': 0.0012437679979484528, 'timing/model': 2.0102535810146946, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 0.66}) +[2026-06-23 07:05:44,159][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.129991814494133, 'train/grad_norm_pre_clip': 26.16079330444336, 'timing/data': 0.0016300720162689686, 'timing/model': 2.003261682984885, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.66}) +[2026-06-23 07:05:48,141][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.03505208343267441, 'train/grad_norm_pre_clip': 15.38225269317627, 'timing/data': 0.0012718390207737684, 'timing/model': 2.0133475180191454, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 0.66}) +[2026-06-23 07:05:52,183][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.08867361396551132, 'train/grad_norm_pre_clip': 18.897045135498047, 'timing/data': 0.06597358200815506, 'timing/model': 2.008353865996469, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 0.66}) +[2026-06-23 07:05:56,150][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.15714548528194427, 'train/grad_norm_pre_clip': 29.32733917236328, 'timing/data': 0.0012566089862957597, 'timing/model': 2.0058821399870794, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 0.66}) +[2026-06-23 07:06:00,322][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.12305644899606705, 'train/grad_norm_pre_clip': 20.294240951538086, 'timing/data': 0.001227525994181633, 'timing/model': 2.122937218984589, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 0.66}) +[2026-06-23 07:06:04,336][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.15428923070430756, 'train/grad_norm_pre_clip': 35.12800979614258, 'timing/data': 0.0012633720180019736, 'timing/model': 2.041408463002881, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 0.66}) +[2026-06-23 07:06:08,378][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.07868200540542603, 'train/grad_norm_pre_clip': 17.177417755126953, 'timing/data': 0.05405092900036834, 'timing/model': 2.0132629419968, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 0.66}) +[2026-06-23 07:06:12,346][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.08554122596979141, 'train/grad_norm_pre_clip': 16.37032699584961, 'timing/data': 0.001107313990360126, 'timing/model': 2.0060181360167917, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.66}) +[2026-06-23 07:06:16,419][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.1244024783372879, 'train/grad_norm_pre_clip': 18.3070011138916, 'timing/data': 0.0012279670045245439, 'timing/model': 2.0112521620176267, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 0.66}) +[2026-06-23 07:06:20,401][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.1096004918217659, 'train/grad_norm_pre_clip': 21.976404190063477, 'timing/data': 0.0012564189964905381, 'timing/model': 2.0121056420030072, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 0.67}) +[2026-06-23 07:06:24,455][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.12069865316152573, 'train/grad_norm_pre_clip': 15.441670417785645, 'timing/data': 0.06056139900465496, 'timing/model': 2.024427308992017, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.67}) +[2026-06-23 07:06:28,462][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.02085249498486519, 'train/grad_norm_pre_clip': 4.58601713180542, 'timing/data': 0.0012010469799861312, 'timing/model': 2.0209256109956186, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 0.67}) +[2026-06-23 07:06:32,546][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.07966959476470947, 'train/grad_norm_pre_clip': 12.211731910705566, 'timing/data': 0.0014279560127761215, 'timing/model': 2.0219480840023607, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 0.67}) +[2026-06-23 07:06:36,784][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.22445935010910034, 'train/grad_norm_pre_clip': 18.084951400756836, 'timing/data': 0.001251520006917417, 'timing/model': 2.0342106340103783, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 0.67}) +[2026-06-23 07:06:40,830][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.10728402435779572, 'train/grad_norm_pre_clip': 25.368181228637695, 'timing/data': 0.057145867991494015, 'timing/model': 2.016892512998311, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 0.67}) +[2026-06-23 07:06:44,817][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.10394242405891418, 'train/grad_norm_pre_clip': 28.634798049926758, 'timing/data': 0.0012210329878143966, 'timing/model': 2.015192067978205, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 0.67}) +[2026-06-23 07:06:48,894][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.11826439201831818, 'train/grad_norm_pre_clip': 25.74269676208496, 'timing/data': 0.0010304019961040467, 'timing/model': 2.015243747009663, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 0.67}) +[2026-06-23 07:06:52,885][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.04743114113807678, 'train/grad_norm_pre_clip': 19.34124755859375, 'timing/data': 0.001241242018295452, 'timing/model': 2.0176709209918045, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 0.67}) +[2026-06-23 07:06:56,928][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.10375314950942993, 'train/grad_norm_pre_clip': 17.328929901123047, 'timing/data': 0.053856759011978284, 'timing/model': 2.0164152969955467, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 0.67}) +[2026-06-23 07:07:00,919][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.1804591864347458, 'train/grad_norm_pre_clip': 16.345962524414062, 'timing/data': 0.0012214439921081066, 'timing/model': 2.018112313002348, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 0.67}) +[2026-06-23 07:07:05,000][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.08834248781204224, 'train/grad_norm_pre_clip': 14.59138011932373, 'timing/data': 0.001184666994959116, 'timing/model': 2.0202090450038668, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.67}) +[2026-06-23 07:07:09,253][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.0518873929977417, 'train/grad_norm_pre_clip': 9.793631553649902, 'timing/data': 0.0012821269920095801, 'timing/model': 2.029586360004032, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 0.67}) +[2026-06-23 07:07:13,315][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.10322616249322891, 'train/grad_norm_pre_clip': 19.880603790283203, 'timing/data': 0.06736013700719923, 'timing/model': 2.0181378710258286, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 0.67}) +[2026-06-23 07:07:17,310][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.05623902752995491, 'train/grad_norm_pre_clip': 14.828160285949707, 'timing/data': 0.0013324499886948615, 'timing/model': 2.0202775829820894, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 0.67}) +[2026-06-23 07:07:21,418][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.06370251625776291, 'train/grad_norm_pre_clip': 10.368672370910645, 'timing/data': 0.0015051980153657496, 'timing/model': 2.0328961679770146, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 0.67}) +[2026-06-23 07:07:25,429][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.09019770473241806, 'train/grad_norm_pre_clip': 27.51154899597168, 'timing/data': 0.0012973439879715443, 'timing/model': 2.0343134359864052, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 0.67}) +[2026-06-23 07:07:29,489][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.10602358728647232, 'train/grad_norm_pre_clip': 21.569679260253906, 'timing/data': 0.058024749014293775, 'timing/model': 2.0226364029804245, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 0.67}) +[2026-06-23 07:07:33,491][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.05218644067645073, 'train/grad_norm_pre_clip': 20.406845092773438, 'timing/data': 0.0013729039928875864, 'timing/model': 2.020183899992844, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 0.67}) +[2026-06-23 07:07:37,578][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.03630644083023071, 'train/grad_norm_pre_clip': 8.544790267944336, 'timing/data': 0.0011959679832216352, 'timing/model': 2.0207871650054585, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 0.67}) +[2026-06-23 07:07:41,610][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.09132132679224014, 'train/grad_norm_pre_clip': 13.46947193145752, 'timing/data': 0.001428546995157376, 'timing/model': 2.0196859530115034, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 0.67}) +[2026-06-23 07:07:45,671][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.22524423897266388, 'train/grad_norm_pre_clip': 17.632640838623047, 'timing/data': 0.06686578001244925, 'timing/model': 2.0188286420016084, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.67}) +[2026-06-23 07:07:49,669][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.20819148421287537, 'train/grad_norm_pre_clip': 51.21510696411133, 'timing/data': 0.001185528002679348, 'timing/model': 2.021575685008429, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 0.67}) +[2026-06-23 07:07:53,756][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.08759578317403793, 'train/grad_norm_pre_clip': 26.351917266845703, 'timing/data': 0.0014324539806693792, 'timing/model': 2.020017499977257, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 0.67}) +[2026-06-23 07:07:57,755][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.16864193975925446, 'train/grad_norm_pre_clip': 44.6241455078125, 'timing/data': 0.0012157450255472213, 'timing/model': 2.021391307003796, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 0.67}) +[2026-06-23 07:08:01,819][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.17668400704860687, 'train/grad_norm_pre_clip': 23.2130069732666, 'timing/data': 0.0667376870114822, 'timing/model': 2.018578546005301, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 0.67}) +[2026-06-23 07:08:05,817][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.13933199644088745, 'train/grad_norm_pre_clip': 22.084304809570312, 'timing/data': 0.0011929820175282657, 'timing/model': 2.0213423979876097, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.68}) +[2026-06-23 07:08:09,911][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.0929771214723587, 'train/grad_norm_pre_clip': 17.8045711517334, 'timing/data': 0.00122053298400715, 'timing/model': 2.0237979420053307, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 0.68}) +[2026-06-23 07:08:13,913][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.23245471715927124, 'train/grad_norm_pre_clip': 43.601097106933594, 'timing/data': 0.0017785030067898333, 'timing/model': 2.0226637280138675, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 0.68}) +[2026-06-23 07:08:18,141][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.10141441971063614, 'train/grad_norm_pre_clip': 27.797691345214844, 'timing/data': 0.0549227999872528, 'timing/model': 2.017352013004711, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 0.68}) +[2026-06-23 07:08:22,138][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.19152678549289703, 'train/grad_norm_pre_clip': 31.48920249938965, 'timing/data': 0.0012088610092177987, 'timing/model': 2.0212703970028087, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 0.68}) +[2026-06-23 07:09:13,377][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1750 +[2026-06-23 07:09:13,379][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.23153652250766754, 'train/grad_norm_pre_clip': 48.197757720947266, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 10.020000000000042, 'rl_games_eval/mid_train/mean_length': 98.6, 'rl_games_eval/mid_train/std_reward': 13.210510966650844, 'rl_games_eval/mid_train/std_length': 106.95344781726301, 'rl_games_eval/mid_train/macro_mean_reward': 10.020000000000042, 'rl_games_eval/mid_train/macro_mean_length': 98.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 10.020000000000042, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 98.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 13.210510966650844, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 106.95344781726301, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 10.020000000000042, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.001199694990646094, 'timing/model': 2.03648435298237, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.68}) +[2026-06-23 07:09:17,298][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.22882211208343506, 'train/grad_norm_pre_clip': 49.267112731933594, 'timing/data': 0.0012489260116126388, 'timing/model': 1.9766086020099465, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 0.68}) +[2026-06-23 07:09:21,309][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.09263424575328827, 'train/grad_norm_pre_clip': 18.28188705444336, 'timing/data': 0.06854869300150312, 'timing/model': 2.0026562380080577, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 0.68}) +[2026-06-23 07:09:25,307][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.18235202133655548, 'train/grad_norm_pre_clip': 28.729944229125977, 'timing/data': 0.0012410109920892864, 'timing/model': 1.9860330049996264, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 0.68}) +[2026-06-23 07:09:29,442][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.12799203395843506, 'train/grad_norm_pre_clip': 15.103914260864258, 'timing/data': 0.0013877120218239725, 'timing/model': 1.9978464530140627, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 0.68}) +[2026-06-23 07:09:33,522][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.16020435094833374, 'train/grad_norm_pre_clip': 21.947065353393555, 'timing/data': 0.0012639430060517043, 'timing/model': 2.1310335710004438, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 0.68}) +[2026-06-23 07:09:37,534][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.17458784580230713, 'train/grad_norm_pre_clip': 22.311481475830078, 'timing/data': 0.06279487197753042, 'timing/model': 1.9930349250207655, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 0.68}) +[2026-06-23 07:09:41,471][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.14772163331508636, 'train/grad_norm_pre_clip': 21.993486404418945, 'timing/data': 0.0011520370026119053, 'timing/model': 1.9903202799905557, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 0.68}) +[2026-06-23 07:09:45,496][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.10313737392425537, 'train/grad_norm_pre_clip': 13.295029640197754, 'timing/data': 0.001113115024054423, 'timing/model': 1.9866186220024247, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 0.68}) +[2026-06-23 07:09:49,437][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.1573382318019867, 'train/grad_norm_pre_clip': 23.392786026000977, 'timing/data': 0.0012342990085016936, 'timing/model': 1.9947930999915116, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 0.68}) +[2026-06-23 07:09:53,446][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.1348344087600708, 'train/grad_norm_pre_clip': 17.305994033813477, 'timing/data': 0.06568384898127988, 'timing/model': 1.9950693270075135, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.68}) +[2026-06-23 07:09:57,420][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.10354466736316681, 'train/grad_norm_pre_clip': 13.472817420959473, 'timing/data': 0.0012157750024925917, 'timing/model': 2.015280627994798, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 0.68}) +[2026-06-23 07:10:01,471][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.3305986225605011, 'train/grad_norm_pre_clip': 46.53972244262695, 'timing/data': 0.0013910290144849569, 'timing/model': 1.9993981010047719, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 0.68}) +[2026-06-23 07:10:05,450][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.24868172407150269, 'train/grad_norm_pre_clip': 42.26799011230469, 'timing/data': 0.001254645991139114, 'timing/model': 2.017864549998194, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 0.68}) +[2026-06-23 07:10:09,714][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.18958085775375366, 'train/grad_norm_pre_clip': 22.408447265625, 'timing/data': 0.055055157019523904, 'timing/model': 1.999032758991234, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 0.68}) +[2026-06-23 07:10:13,679][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.0852871686220169, 'train/grad_norm_pre_clip': 10.388384819030762, 'timing/data': 0.0013606429856736213, 'timing/model': 2.00582616799511, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 0.68}) +[2026-06-23 07:10:17,749][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.07309674471616745, 'train/grad_norm_pre_clip': 16.40745735168457, 'timing/data': 0.0011878829973284155, 'timing/model': 2.0203455110022333, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.68}) +[2026-06-23 07:10:21,723][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.07301728427410126, 'train/grad_norm_pre_clip': 11.87735652923584, 'timing/data': 0.0012839309929404408, 'timing/model': 2.009565162996296, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 0.68}) +[2026-06-23 07:10:25,747][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.09085698425769806, 'train/grad_norm_pre_clip': 9.479647636413574, 'timing/data': 0.056442970992065966, 'timing/model': 2.007498772989493, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 0.68}) +[2026-06-23 07:10:29,737][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.12223796546459198, 'train/grad_norm_pre_clip': 18.547821044921875, 'timing/data': 0.0011839359940495342, 'timing/model': 2.023486927995691, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 0.68}) +[2026-06-23 07:10:33,786][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.17668022215366364, 'train/grad_norm_pre_clip': 32.725067138671875, 'timing/data': 0.001145885995356366, 'timing/model': 2.0014521690027323, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.69}) +[2026-06-23 07:10:37,755][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.11759573221206665, 'train/grad_norm_pre_clip': 20.026243209838867, 'timing/data': 0.0012290790036786348, 'timing/model': 2.009395510016475, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 0.69}) +[2026-06-23 07:10:41,888][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.09831666201353073, 'train/grad_norm_pre_clip': 16.081806182861328, 'timing/data': 0.061738130025332794, 'timing/model': 2.1036646389984526, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 0.69}) +[2026-06-23 07:10:45,873][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.0976693257689476, 'train/grad_norm_pre_clip': 11.538871765136719, 'timing/data': 0.0012014280073344707, 'timing/model': 2.017228783981409, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 0.69}) +[2026-06-23 07:10:49,972][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.0358315110206604, 'train/grad_norm_pre_clip': 6.579208850860596, 'timing/data': 0.0017830619763117284, 'timing/model': 2.027592944999924, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 0.69}) +[2026-06-23 07:10:53,964][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.08020798116922379, 'train/grad_norm_pre_clip': 11.64597225189209, 'timing/data': 0.0012806039885617793, 'timing/model': 2.015387574007036, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 0.69}) +[2026-06-23 07:10:58,009][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.12733404338359833, 'train/grad_norm_pre_clip': 22.790454864501953, 'timing/data': 0.06456480998895131, 'timing/model': 2.0118897009815555, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 0.69}) +[2026-06-23 07:11:01,990][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.08458507061004639, 'train/grad_norm_pre_clip': 17.592266082763672, 'timing/data': 0.0012006360047962517, 'timing/model': 2.0111685230222065, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 0.69}) +[2026-06-23 07:11:06,058][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.16144756972789764, 'train/grad_norm_pre_clip': 25.66509437561035, 'timing/data': 0.0011330019915476441, 'timing/model': 2.011823748005554, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 0.69}) +[2026-06-23 07:11:10,044][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.19055317342281342, 'train/grad_norm_pre_clip': 23.84844398498535, 'timing/data': 0.001191349991131574, 'timing/model': 2.0176241009903606, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 0.69}) +[2026-06-23 07:11:14,078][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.12042845785617828, 'train/grad_norm_pre_clip': 22.966474533081055, 'timing/data': 0.053370623994851485, 'timing/model': 2.011803339002654, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.69}) +[2026-06-23 07:11:18,098][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.04718684032559395, 'train/grad_norm_pre_clip': 10.205986976623535, 'timing/data': 0.001117714011343196, 'timing/model': 2.0131919769919477, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 0.69}) +[2026-06-23 07:11:22,171][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.09589982032775879, 'train/grad_norm_pre_clip': 26.46869659423828, 'timing/data': 0.0011286939843557775, 'timing/model': 2.015111686021555, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 0.69}) +[2026-06-23 07:11:26,169][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.1013796404004097, 'train/grad_norm_pre_clip': 17.430391311645508, 'timing/data': 0.0012881889997515827, 'timing/model': 2.027429503010353, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 0.69}) +[2026-06-23 07:11:30,234][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.1559268683195114, 'train/grad_norm_pre_clip': 31.181772232055664, 'timing/data': 0.06591372497496195, 'timing/model': 2.0180173979897518, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 0.69}) +[2026-06-23 07:11:34,249][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.2299319952726364, 'train/grad_norm_pre_clip': 48.55681610107422, 'timing/data': 0.0012328669836279005, 'timing/model': 2.026873553986661, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 0.69}) +[2026-06-23 07:11:38,359][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.23516836762428284, 'train/grad_norm_pre_clip': 64.25646209716797, 'timing/data': 0.0014496170042548329, 'timing/model': 2.0203927510010544, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 0.69}) +[2026-06-23 07:11:42,363][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.08640070259571075, 'train/grad_norm_pre_clip': 21.42803192138672, 'timing/data': 0.0013138259819243103, 'timing/model': 2.0269548619980924, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 0.69}) +[2026-06-23 07:11:46,422][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.1350775957107544, 'train/grad_norm_pre_clip': 34.82536315917969, 'timing/data': 0.053990551008610055, 'timing/model': 2.0220770589949097, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 0.69}) +[2026-06-23 07:11:50,463][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.0512319840490818, 'train/grad_norm_pre_clip': 14.807527542114258, 'timing/data': 0.0016716370009817183, 'timing/model': 2.0574463889934123, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 0.69}) +[2026-06-23 07:11:54,548][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.06950142979621887, 'train/grad_norm_pre_clip': 29.325761795043945, 'timing/data': 0.0011956079979427159, 'timing/model': 2.0240407710080035, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.69}) +[2026-06-23 07:11:58,552][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.17447055876255035, 'train/grad_norm_pre_clip': 45.534141540527344, 'timing/data': 0.0014919039967935532, 'timing/model': 2.021835691994056, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 0.69}) +[2026-06-23 07:12:02,618][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.13120883703231812, 'train/grad_norm_pre_clip': 23.356491088867188, 'timing/data': 0.06117602999438532, 'timing/model': 2.024031009990722, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 0.69}) +[2026-06-23 07:12:06,613][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.17490078508853912, 'train/grad_norm_pre_clip': 32.55709457397461, 'timing/data': 0.001130296994233504, 'timing/model': 2.018934490974061, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 0.69}) +[2026-06-23 07:12:10,695][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.09247733652591705, 'train/grad_norm_pre_clip': 20.331083297729492, 'timing/data': 0.0011396949994377792, 'timing/model': 2.0155311200069264, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 0.69}) +[2026-06-23 07:12:14,687][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.236612007021904, 'train/grad_norm_pre_clip': 28.837507247924805, 'timing/data': 0.0011469380115158856, 'timing/model': 2.018043021002086, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 0.69}) +[2026-06-23 07:12:18,732][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.06027654930949211, 'train/grad_norm_pre_clip': 12.441266059875488, 'timing/data': 0.057028883980819955, 'timing/model': 2.014705239998875, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 0.7}) +[2026-06-23 07:12:22,724][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.1342499554157257, 'train/grad_norm_pre_clip': 39.969635009765625, 'timing/data': 0.0011268810194451362, 'timing/model': 2.016692522010999, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 0.7}) +[2026-06-23 07:12:26,994][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.10374178737401962, 'train/grad_norm_pre_clip': 28.01641082763672, 'timing/data': 0.0011975109810009599, 'timing/model': 2.015388426982099, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 0.7}) +[2026-06-23 07:12:31,005][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.21237722039222717, 'train/grad_norm_pre_clip': 35.19083786010742, 'timing/data': 0.0011539309925865382, 'timing/model': 2.036816356005147, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 0.7}) +[2026-06-23 07:12:35,071][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1800 +[2026-06-23 07:14:29,445][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1800 +[2026-06-23 07:14:29,445][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1800 +[2026-06-23 07:16:36,397][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1800 +[2026-06-23 07:16:36,399][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.035900015383958817, 'train/grad_norm_pre_clip': 12.260099411010742, 'eval/action_loss/seconds': 114.36965408199467, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.05845174938440323, 'eval/latency_2/loss': 0.05845174938440323, 'eval/flappy/loss': 0.05845174938440323, 'eval/flappy/latency_2/loss': 0.05845174938440323, 'eval/flap_precision': 0.9347684279191129, 'eval/flap_recall': 0.9669365721997301, 'eval/flap_f1': 0.9505804311774462, 'eval/flap_rate_model': 0.21291666666666667, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9505804311774461, 'eval/latency_2/flap_precision': 0.9347684279191129, 'eval/latency_2/flap_recall': 0.9669365721997301, 'eval/latency_2/flap_f1': 0.9505804311774462, 'eval/latency_2/flap_rate_model': 0.21291666666666667, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9505804311774461, 'eval/action_classification/data_wait_seconds': 50.36863327026367, 'eval/action_classification/predict_seconds': 74.81109619140625, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06453637799131684, 'timing/model': 2.0249519610079005, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.7}) +[2026-06-23 07:16:40,360][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.13278037309646606, 'train/grad_norm_pre_clip': 13.225290298461914, 'timing/data': 0.0013735069951508194, 'timing/model': 2.0048945270245895, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 0.7}) +[2026-06-23 07:16:44,385][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.1285843551158905, 'train/grad_norm_pre_clip': 18.112321853637695, 'timing/data': 0.0013924119994044304, 'timing/model': 1.9909525559924077, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 0.7}) +[2026-06-23 07:16:48,579][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.1678536832332611, 'train/grad_norm_pre_clip': 23.041704177856445, 'timing/data': 0.0012499889999162406, 'timing/model': 2.242450085992459, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 0.7}) +[2026-06-23 07:16:52,574][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.26771602034568787, 'train/grad_norm_pre_clip': 31.251012802124023, 'timing/data': 0.053790450998349115, 'timing/model': 1.9923090870142914, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 0.7}) +[2026-06-23 07:16:56,542][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.15765438973903656, 'train/grad_norm_pre_clip': 21.496126174926758, 'timing/data': 0.0011812210141215473, 'timing/model': 2.0129538509936538, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 0.7}) +[2026-06-23 07:17:00,582][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.18166066706180573, 'train/grad_norm_pre_clip': 30.36785888671875, 'timing/data': 0.0013046200037933886, 'timing/model': 1.9986955049971584, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 0.7}) +[2026-06-23 07:17:04,548][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.07494116574525833, 'train/grad_norm_pre_clip': 16.75448226928711, 'timing/data': 0.0012175989977549762, 'timing/model': 2.0088000039977487, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 0.7}) +[2026-06-23 07:17:08,587][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.0792236253619194, 'train/grad_norm_pre_clip': 21.254663467407227, 'timing/data': 0.0683857680123765, 'timing/model': 2.009085897996556, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 0.7}) +[2026-06-23 07:17:12,561][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.1040688008069992, 'train/grad_norm_pre_clip': 21.100271224975586, 'timing/data': 0.0012254439934622496, 'timing/model': 2.012718975020107, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 0.7}) +[2026-06-23 07:17:16,625][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.09141344577074051, 'train/grad_norm_pre_clip': 11.390958786010742, 'timing/data': 0.0012334779894445091, 'timing/model': 2.0210442479874473, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.7}) +[2026-06-23 07:17:20,622][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.10270103067159653, 'train/grad_norm_pre_clip': 11.058284759521484, 'timing/data': 0.0014056060172151774, 'timing/model': 2.0236979559995234, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 0.7}) +[2026-06-23 07:17:24,700][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.1004357635974884, 'train/grad_norm_pre_clip': 16.12214469909668, 'timing/data': 0.06401110897422768, 'timing/model': 2.003932167979656, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 0.7}) +[2026-06-23 07:17:28,682][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.2778497040271759, 'train/grad_norm_pre_clip': 45.17962646484375, 'timing/data': 0.001167556009022519, 'timing/model': 2.0211135770077817, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 0.7}) +[2026-06-23 07:17:32,734][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.12609602510929108, 'train/grad_norm_pre_clip': 28.388755798339844, 'timing/data': 0.0011779050109907985, 'timing/model': 2.001363377989037, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 0.7}) +[2026-06-23 07:17:36,697][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.09921073168516159, 'train/grad_norm_pre_clip': 14.843335151672363, 'timing/data': 0.0012189409753773361, 'timing/model': 2.0044270920043346, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.7}) +[2026-06-23 07:17:40,727][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.07462280243635178, 'train/grad_norm_pre_clip': 11.491851806640625, 'timing/data': 0.058888108003884554, 'timing/model': 2.009103035001317, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 0.7}) +[2026-06-23 07:17:44,712][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.051412489265203476, 'train/grad_norm_pre_clip': 14.618692398071289, 'timing/data': 0.001156124984845519, 'timing/model': 2.0135172340087593, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 0.7}) +[2026-06-23 07:17:48,792][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.18393166363239288, 'train/grad_norm_pre_clip': 27.415172576904297, 'timing/data': 0.0012331180041655898, 'timing/model': 2.0309032300137915, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 0.7}) +[2026-06-23 07:17:52,773][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.19226762652397156, 'train/grad_norm_pre_clip': 45.748451232910156, 'timing/data': 0.0013879629841540009, 'timing/model': 2.0151282729930244, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 0.7}) +[2026-06-23 07:17:56,944][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.2944943606853485, 'train/grad_norm_pre_clip': 43.0340576171875, 'timing/data': 0.05994916899362579, 'timing/model': 2.146471404004842, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.7}) +[2026-06-23 07:18:00,934][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.06484171748161316, 'train/grad_norm_pre_clip': 40.621700286865234, 'timing/data': 0.0013812299876008183, 'timing/model': 2.0181716659863014, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 0.7}) +[2026-06-23 07:18:05,024][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.10256417095661163, 'train/grad_norm_pre_clip': 20.542085647583008, 'timing/data': 0.0013023460051044822, 'timing/model': 2.025997365999501, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 0.71}) +[2026-06-23 07:18:09,019][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.0750633180141449, 'train/grad_norm_pre_clip': 8.641541481018066, 'timing/data': 0.0013711020001210272, 'timing/model': 2.0255012869893108, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 0.71}) +[2026-06-23 07:18:13,061][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.11371671408414841, 'train/grad_norm_pre_clip': 15.138102531433105, 'timing/data': 0.06279262198950164, 'timing/model': 2.012422156985849, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 0.71}) +[2026-06-23 07:18:17,046][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.06337231397628784, 'train/grad_norm_pre_clip': 12.879168510437012, 'timing/data': 0.0012202739890199155, 'timing/model': 2.01399552897783, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 0.71}) +[2026-06-23 07:18:21,122][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.017112892121076584, 'train/grad_norm_pre_clip': 4.407569408416748, 'timing/data': 0.0013310780050233006, 'timing/model': 2.018502309016185, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 0.71}) +[2026-06-23 07:18:25,100][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.11396703124046326, 'train/grad_norm_pre_clip': 14.247441291809082, 'timing/data': 0.0011789480049628764, 'timing/model': 2.008512954984326, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 0.71}) +[2026-06-23 07:18:29,398][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.03695806488394737, 'train/grad_norm_pre_clip': 7.710461139678955, 'timing/data': 0.0635970089933835, 'timing/model': 2.2612729360116646, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 0.71}) +[2026-06-23 07:18:33,392][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.19286338984966278, 'train/grad_norm_pre_clip': 24.29292106628418, 'timing/data': 0.0012895920081064105, 'timing/model': 2.024801600986393, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 0.71}) +[2026-06-23 07:18:37,466][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.08416572213172913, 'train/grad_norm_pre_clip': 15.035600662231445, 'timing/data': 0.0011315589945297688, 'timing/model': 2.0238924979930744, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.71}) +[2026-06-23 07:18:41,455][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.08175370842218399, 'train/grad_norm_pre_clip': 20.15723419189453, 'timing/data': 0.001266008010134101, 'timing/model': 2.0180545920156874, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 0.71}) +[2026-06-23 07:18:45,519][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.04309308901429176, 'train/grad_norm_pre_clip': 13.943504333496094, 'timing/data': 0.06987145700259134, 'timing/model': 2.017125901998952, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 0.71}) +[2026-06-23 07:18:49,499][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.19332265853881836, 'train/grad_norm_pre_clip': 31.287471771240234, 'timing/data': 0.0011377320042811334, 'timing/model': 2.009333355992567, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 0.71}) +[2026-06-23 07:18:53,573][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.1394258588552475, 'train/grad_norm_pre_clip': 34.079002380371094, 'timing/data': 0.001227747998200357, 'timing/model': 2.0168187670060433, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 0.71}) +[2026-06-23 07:18:57,558][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.20095837116241455, 'train/grad_norm_pre_clip': 47.5045166015625, 'timing/data': 0.0012095839774701744, 'timing/model': 2.0144499279849697, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 0.71}) +[2026-06-23 07:19:01,604][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.014857955276966095, 'train/grad_norm_pre_clip': 3.9494869709014893, 'timing/data': 0.06198957999004051, 'timing/model': 2.0132829080102965, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 0.71}) +[2026-06-23 07:19:05,840][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.0801483690738678, 'train/grad_norm_pre_clip': 26.342695236206055, 'timing/data': 0.001208251982461661, 'timing/model': 2.0179848480038345, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 0.71}) +[2026-06-23 07:19:09,926][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.1219949796795845, 'train/grad_norm_pre_clip': 34.54578399658203, 'timing/data': 0.001190428010886535, 'timing/model': 2.028188279015012, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 0.71}) +[2026-06-23 07:19:13,922][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.25387266278266907, 'train/grad_norm_pre_clip': 35.51223373413086, 'timing/data': 0.0014086809824220836, 'timing/model': 2.0127923589898273, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 0.71}) +[2026-06-23 07:19:17,979][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.03934205323457718, 'train/grad_norm_pre_clip': 18.5451602935791, 'timing/data': 0.06521172198699787, 'timing/model': 2.022470288997283, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.71}) +[2026-06-23 07:19:21,969][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.14570757746696472, 'train/grad_norm_pre_clip': 42.07254409790039, 'timing/data': 0.00121720798779279, 'timing/model': 2.0153088980005123, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 0.71}) +[2026-06-23 07:19:26,047][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.304659366607666, 'train/grad_norm_pre_clip': 38.204769134521484, 'timing/data': 0.001227486995048821, 'timing/model': 2.0199500960006844, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 0.71}) +[2026-06-23 07:19:30,047][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.28638145327568054, 'train/grad_norm_pre_clip': 94.53617095947266, 'timing/data': 0.001199024001834914, 'timing/model': 2.0173189020133577, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 0.71}) +[2026-06-23 07:19:34,094][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.22306068241596222, 'train/grad_norm_pre_clip': 24.13923454284668, 'timing/data': 0.053176193992840126, 'timing/model': 2.0215348839992657, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 0.71}) +[2026-06-23 07:19:38,175][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.093086376786232, 'train/grad_norm_pre_clip': 22.910385131835938, 'timing/data': 0.0011332530120853335, 'timing/model': 2.108659343997715, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 0.71}) +[2026-06-23 07:19:42,284][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.16322316229343414, 'train/grad_norm_pre_clip': 42.57381820678711, 'timing/data': 0.0012255539768375456, 'timing/model': 2.0383480679884087, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 0.71}) +[2026-06-23 07:19:46,284][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.0837104320526123, 'train/grad_norm_pre_clip': 14.549588203430176, 'timing/data': 0.001257191994227469, 'timing/model': 2.023307454015594, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 0.71}) +[2026-06-23 07:19:50,335][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.12899890542030334, 'train/grad_norm_pre_clip': 27.4381046295166, 'timing/data': 0.05341363401385024, 'timing/model': 2.0203097929770593, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 0.72}) +[2026-06-23 07:19:54,352][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.04909174144268036, 'train/grad_norm_pre_clip': 13.209056854248047, 'timing/data': 0.001282298006117344, 'timing/model': 2.0388319980120286, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 0.72}) +[2026-06-23 07:19:58,434][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.03938848525285721, 'train/grad_norm_pre_clip': 13.8985595703125, 'timing/data': 0.0012352910125628114, 'timing/model': 2.0157829220115673, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.72}) +[2026-06-23 07:20:02,429][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.1733919084072113, 'train/grad_norm_pre_clip': 32.76990509033203, 'timing/data': 0.001160052022896707, 'timing/model': 2.0227932650013827, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 0.72}) +[2026-06-23 07:20:06,483][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.08073864132165909, 'train/grad_norm_pre_clip': 19.541528701782227, 'timing/data': 0.05223632900742814, 'timing/model': 2.023520351998741, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 0.72}) +[2026-06-23 07:20:10,502][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.0693078488111496, 'train/grad_norm_pre_clip': 11.262713432312012, 'timing/data': 0.001237946009496227, 'timing/model': 2.0258498809998855, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 0.72}) +[2026-06-23 07:20:14,830][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.11670465022325516, 'train/grad_norm_pre_clip': 25.2507266998291, 'timing/data': 0.0012170880218036473, 'timing/model': 2.031682297005318, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 0.72}) +[2026-06-23 07:20:18,843][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.1799314022064209, 'train/grad_norm_pre_clip': 24.64137840270996, 'timing/data': 0.0015025440079625696, 'timing/model': 2.0324065900058486, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 0.72}) +[2026-06-23 07:20:22,895][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.15153680741786957, 'train/grad_norm_pre_clip': 18.782339096069336, 'timing/data': 0.06289014697540551, 'timing/model': 2.0169708959874697, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 0.72}) +[2026-06-23 07:20:26,885][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.1705639362335205, 'train/grad_norm_pre_clip': 24.615131378173828, 'timing/data': 0.0011076859955210239, 'timing/model': 2.019495543994708, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 0.72}) +[2026-06-23 07:20:30,962][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.1455516517162323, 'train/grad_norm_pre_clip': 34.23027420043945, 'timing/data': 0.0013028460089117289, 'timing/model': 2.019588469993323, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 0.72}) +[2026-06-23 07:20:34,958][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.041654717177152634, 'train/grad_norm_pre_clip': 12.779385566711426, 'timing/data': 0.0012345499999355525, 'timing/model': 2.02118445202359, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 0.72}) +[2026-06-23 07:20:39,008][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.17822477221488953, 'train/grad_norm_pre_clip': 24.32309341430664, 'timing/data': 0.051870885014068335, 'timing/model': 2.0213149890187196, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.72}) +[2026-06-23 07:20:43,006][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.05394143983721733, 'train/grad_norm_pre_clip': 13.909248352050781, 'timing/data': 0.0012080110027454793, 'timing/model': 2.0225101800169796, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 0.72}) +[2026-06-23 07:20:47,332][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.12223365902900696, 'train/grad_norm_pre_clip': 15.02579116821289, 'timing/data': 0.001197631994727999, 'timing/model': 2.2632009419903625, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 0.72}) +[2026-06-23 07:20:51,330][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.13500459492206573, 'train/grad_norm_pre_clip': 17.897951126098633, 'timing/data': 0.0012406220193952322, 'timing/model': 2.0203325589827728, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 0.72}) +[2026-06-23 07:20:55,386][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.12332849949598312, 'train/grad_norm_pre_clip': 20.696001052856445, 'timing/data': 0.06162683601723984, 'timing/model': 2.020418261003215, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.72}) +[2026-06-23 07:20:59,404][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.12312096357345581, 'train/grad_norm_pre_clip': 17.174238204956055, 'timing/data': 0.001376322004944086, 'timing/model': 2.043597066978691, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 0.72}) +[2026-06-23 07:21:03,473][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.03959088400006294, 'train/grad_norm_pre_clip': 10.62878131866455, 'timing/data': 0.001161835010861978, 'timing/model': 2.014123593980912, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 0.72}) +[2026-06-23 07:21:07,461][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.09604278206825256, 'train/grad_norm_pre_clip': 19.927227020263672, 'timing/data': 0.0014050760073587298, 'timing/model': 2.0154008669778705, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 0.72}) +[2026-06-23 07:21:11,498][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.09241869300603867, 'train/grad_norm_pre_clip': 13.207390785217285, 'timing/data': 0.05147102801129222, 'timing/model': 2.013711712002987, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 0.72}) +[2026-06-23 07:21:15,496][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.14837761223316193, 'train/grad_norm_pre_clip': 24.234127044677734, 'timing/data': 0.0011061429977416992, 'timing/model': 2.025659645005362, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 0.72}) +[2026-06-23 07:21:19,685][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.14241240918636322, 'train/grad_norm_pre_clip': 22.19757843017578, 'timing/data': 0.001158360013505444, 'timing/model': 2.131341786007397, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.72}) +[2026-06-23 07:21:23,694][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.080911785364151, 'train/grad_norm_pre_clip': 14.430173873901367, 'timing/data': 0.0013237950042821467, 'timing/model': 2.0338098469947, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 0.72}) +[2026-06-23 07:21:27,751][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.044572509825229645, 'train/grad_norm_pre_clip': 8.05297565460205, 'timing/data': 0.06605538498843089, 'timing/model': 2.019204454991268, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 0.72}) +[2026-06-23 07:21:31,744][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.10284766554832458, 'train/grad_norm_pre_clip': 20.405630111694336, 'timing/data': 0.0011317100143060088, 'timing/model': 2.0180091619840823, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 0.72}) +[2026-06-23 07:21:35,819][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.0927826389670372, 'train/grad_norm_pre_clip': 12.682353019714355, 'timing/data': 0.0011433220061007887, 'timing/model': 2.018089574004989, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 0.73}) +[2026-06-23 07:21:39,819][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.032734308391809464, 'train/grad_norm_pre_clip': 6.757214546203613, 'timing/data': 0.0014149040216580033, 'timing/model': 2.019521062000422, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 0.73}) +[2026-06-23 07:21:43,870][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.05007149651646614, 'train/grad_norm_pre_clip': 23.51333236694336, 'timing/data': 0.05067737301578745, 'timing/model': 2.0205028300115373, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 0.73}) +[2026-06-23 07:21:47,864][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.06223823502659798, 'train/grad_norm_pre_clip': 13.895928382873535, 'timing/data': 0.0012430660135578364, 'timing/model': 2.0179716180136893, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 0.73}) +[2026-06-23 07:21:51,942][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.15243205428123474, 'train/grad_norm_pre_clip': 30.284597396850586, 'timing/data': 0.001135867991251871, 'timing/model': 2.0199146009981632, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 0.73}) +[2026-06-23 07:21:56,165][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.15132823586463928, 'train/grad_norm_pre_clip': 18.68047523498535, 'timing/data': 0.0011520380212459713, 'timing/model': 2.017904242995428, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 0.73}) +[2026-06-23 07:22:00,214][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.10620103776454926, 'train/grad_norm_pre_clip': 20.018980026245117, 'timing/data': 0.05122938202111982, 'timing/model': 2.0211416159872897, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.73}) +[2026-06-23 07:22:04,216][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.11017580330371857, 'train/grad_norm_pre_clip': 26.718017578125, 'timing/data': 0.0011309089895803481, 'timing/model': 2.0234819629986305, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 0.73}) +[2026-06-23 07:22:08,297][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.12166382372379303, 'train/grad_norm_pre_clip': 18.097402572631836, 'timing/data': 0.001332561980234459, 'timing/model': 2.0182481500087306, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 0.73}) +[2026-06-23 07:22:12,296][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.10547547042369843, 'train/grad_norm_pre_clip': 16.722442626953125, 'timing/data': 0.0012059769942425191, 'timing/model': 2.0217657510074787, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 0.73}) +[2026-06-23 07:22:16,334][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.13175790011882782, 'train/grad_norm_pre_clip': 20.022476196289062, 'timing/data': 0.05095818202244118, 'timing/model': 2.0137768590066116, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 0.73}) +[2026-06-23 07:22:20,326][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.16269689798355103, 'train/grad_norm_pre_clip': 20.827180862426758, 'timing/data': 0.0011192870151717216, 'timing/model': 2.017672087997198, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.73}) +[2026-06-23 07:22:24,415][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.0712229460477829, 'train/grad_norm_pre_clip': 24.716493606567383, 'timing/data': 0.0011428400175645947, 'timing/model': 2.031730307993712, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 0.73}) +[2026-06-23 07:22:28,640][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.06323635578155518, 'train/grad_norm_pre_clip': 19.714855194091797, 'timing/data': 0.0011773350124713033, 'timing/model': 2.0159412079956383, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 0.73}) +[2026-06-23 07:22:32,683][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.09774928539991379, 'train/grad_norm_pre_clip': 12.3800048828125, 'timing/data': 0.051164424017770216, 'timing/model': 2.0199471510131843, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 0.73}) +[2026-06-23 07:22:36,681][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.06163303181529045, 'train/grad_norm_pre_clip': 21.164113998413086, 'timing/data': 0.0011201599845662713, 'timing/model': 2.019808669021586, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 0.73}) +[2026-06-23 07:22:40,761][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.11630120128393173, 'train/grad_norm_pre_clip': 27.521976470947266, 'timing/data': 0.001116182014811784, 'timing/model': 2.0184482600016054, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.73}) +[2026-06-23 07:22:44,772][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.15063586831092834, 'train/grad_norm_pre_clip': 24.394685745239258, 'timing/data': 0.0013636289804708213, 'timing/model': 2.026310175016988, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 0.73}) +[2026-06-23 07:22:48,832][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.09575995802879333, 'train/grad_norm_pre_clip': 27.124469757080078, 'timing/data': 0.058190231007756665, 'timing/model': 2.0240734980034176, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 0.73}) +[2026-06-23 07:22:52,837][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.07951238006353378, 'train/grad_norm_pre_clip': 17.080808639526367, 'timing/data': 0.0011989639897365123, 'timing/model': 2.0258251470222604, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 0.73}) +[2026-06-23 07:22:56,933][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.02885524369776249, 'train/grad_norm_pre_clip': 8.299148559570312, 'timing/data': 0.0012191009882371873, 'timing/model': 2.029050648008706, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 0.73}) +[2026-06-23 07:23:00,975][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.0956915095448494, 'train/grad_norm_pre_clip': 17.16107749938965, 'timing/data': 0.0015070440131239593, 'timing/model': 2.0312371299951337, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 0.73}) +[2026-06-23 07:23:05,036][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.17749740183353424, 'train/grad_norm_pre_clip': 31.93734359741211, 'timing/data': 0.05910202601808123, 'timing/model': 2.023777226015227, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 0.73}) +[2026-06-23 07:23:09,040][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.16973231732845306, 'train/grad_norm_pre_clip': 28.110980987548828, 'timing/data': 0.0012344300048425794, 'timing/model': 2.024671970983036, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 0.73}) +[2026-06-23 07:23:13,128][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.1244700625538826, 'train/grad_norm_pre_clip': 32.02740478515625, 'timing/data': 0.0012386789894662797, 'timing/model': 2.0240594730130397, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 0.73}) +[2026-06-23 07:23:17,130][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.08239526301622391, 'train/grad_norm_pre_clip': 12.183542251586914, 'timing/data': 0.0012995110009796917, 'timing/model': 2.0204867779975757, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 0.73}) +[2026-06-23 07:23:21,186][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1900 +[2026-06-23 07:25:15,815][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1900 +[2026-06-23 07:25:15,816][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1900 +[2026-06-23 07:27:22,686][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1900 +[2026-06-23 07:27:22,688][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.15061165392398834, 'train/grad_norm_pre_clip': 14.409992218017578, 'eval/action_loss/seconds': 114.62604641300277, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.05931830406188965, 'eval/latency_2/loss': 0.05931830406188965, 'eval/flappy/loss': 0.05931830406188965, 'eval/flappy/latency_2/loss': 0.05931830406188965, 'eval/flap_precision': 0.9025, 'eval/flap_recall': 0.9743589743589743, 'eval/flap_f1': 0.9370538611291369, 'eval/flap_rate_model': 0.2222222222222222, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9370538611291369, 'eval/latency_2/flap_precision': 0.9025, 'eval/latency_2/flap_recall': 0.9743589743589743, 'eval/latency_2/flap_f1': 0.9370538611291369, 'eval/latency_2/flap_rate_model': 0.2222222222222222, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9370538611291369, 'eval/action_classification/data_wait_seconds': 50.22459030151367, 'eval/action_classification/predict_seconds': 74.87393188476562, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.05505462098517455, 'timing/model': 2.0223561140010133, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.74}) +[2026-06-23 07:27:26,834][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.1397847831249237, 'train/grad_norm_pre_clip': 23.19667625427246, 'timing/data': 0.001185760018415749, 'timing/model': 2.202104544005124, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 0.74}) +[2026-06-23 07:27:30,858][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.09864818304777145, 'train/grad_norm_pre_clip': 63.2940673828125, 'timing/data': 0.0011164619936607778, 'timing/model': 1.9894253950042184, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 0.74}) +[2026-06-23 07:27:34,801][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.12468552589416504, 'train/grad_norm_pre_clip': 27.643081665039062, 'timing/data': 0.0011621259909588844, 'timing/model': 1.9950689219986089, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 0.74}) +[2026-06-23 07:27:38,809][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.17907722294330597, 'train/grad_norm_pre_clip': 21.864219665527344, 'timing/data': 0.06676427400088869, 'timing/model': 1.9932018640101887, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 0.74}) +[2026-06-23 07:27:42,760][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.06920960545539856, 'train/grad_norm_pre_clip': 23.380558013916016, 'timing/data': 0.001106723997509107, 'timing/model': 1.9991350249911193, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 0.74}) +[2026-06-23 07:27:46,813][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.16198506951332092, 'train/grad_norm_pre_clip': 15.531601905822754, 'timing/data': 0.0013091480068396777, 'timing/model': 2.007980327005498, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 0.74}) +[2026-06-23 07:27:50,763][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.098757803440094, 'train/grad_norm_pre_clip': 10.754746437072754, 'timing/data': 0.0011786370014306158, 'timing/model': 1.9952919309725985, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 0.74}) +[2026-06-23 07:27:54,774][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.17539529502391815, 'train/grad_norm_pre_clip': 41.38923263549805, 'timing/data': 0.059763983997981995, 'timing/model': 1.9970124300161842, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 0.74}) +[2026-06-23 07:27:58,739][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.1355351209640503, 'train/grad_norm_pre_clip': 29.349069595336914, 'timing/data': 0.001281937991734594, 'timing/model': 2.0076666029926855, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 0.74}) +[2026-06-23 07:28:03,051][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.08606279641389847, 'train/grad_norm_pre_clip': 16.5, 'timing/data': 0.0012503399921115488, 'timing/model': 2.0055053629912436, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.74}) +[2026-06-23 07:28:07,036][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.13361959159374237, 'train/grad_norm_pre_clip': 19.382253646850586, 'timing/data': 0.0012997309968341142, 'timing/model': 2.0116027280164417, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 0.74}) +[2026-06-23 07:28:11,098][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.1071922779083252, 'train/grad_norm_pre_clip': 14.805384635925293, 'timing/data': 0.08539385700714774, 'timing/model': 2.0132531780109275, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 0.74}) +[2026-06-23 07:28:15,090][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.2179008573293686, 'train/grad_norm_pre_clip': 38.22027587890625, 'timing/data': 0.001255889976164326, 'timing/model': 2.028249611001229, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.74}) +[2026-06-23 07:28:19,170][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.10185124725103378, 'train/grad_norm_pre_clip': 23.22432518005371, 'timing/data': 0.0013810509990435094, 'timing/model': 2.025482232013019, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 0.74}) +[2026-06-23 07:28:23,162][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.0668923631310463, 'train/grad_norm_pre_clip': 10.607819557189941, 'timing/data': 0.0012627730029635131, 'timing/model': 2.0275057310063858, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 0.74}) +[2026-06-23 07:28:27,205][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.11422111093997955, 'train/grad_norm_pre_clip': 19.782672882080078, 'timing/data': 0.05946600399329327, 'timing/model': 2.017390739027178, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 0.74}) +[2026-06-23 07:28:31,196][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.07656145840883255, 'train/grad_norm_pre_clip': 11.340703964233398, 'timing/data': 0.0015281219966709614, 'timing/model': 2.0177105409966316, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 0.74}) +[2026-06-23 07:28:35,474][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.07310493290424347, 'train/grad_norm_pre_clip': 10.630973815917969, 'timing/data': 0.001215345022501424, 'timing/model': 2.219255311996676, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 0.74}) +[2026-06-23 07:28:39,478][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.13596220314502716, 'train/grad_norm_pre_clip': 18.33656120300293, 'timing/data': 0.0014847629936411977, 'timing/model': 2.0297894770046696, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 0.74}) +[2026-06-23 07:28:43,542][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.059778716415166855, 'train/grad_norm_pre_clip': 13.503243446350098, 'timing/data': 0.07362451701192185, 'timing/model': 2.020108386001084, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.74}) +[2026-06-23 07:28:47,542][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.09302616119384766, 'train/grad_norm_pre_clip': 15.88062572479248, 'timing/data': 0.0014094639918766916, 'timing/model': 2.0264143960084766, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 0.74}) +[2026-06-23 07:28:51,601][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.13455039262771606, 'train/grad_norm_pre_clip': 19.83928680419922, 'timing/data': 0.0012120490137021989, 'timing/model': 2.01163659102167, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 0.74}) +[2026-06-23 07:28:55,590][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.1193365827202797, 'train/grad_norm_pre_clip': 23.786264419555664, 'timing/data': 0.0014363630034495145, 'timing/model': 2.0197481469949707, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 0.74}) +[2026-06-23 07:28:59,649][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.07448981702327728, 'train/grad_norm_pre_clip': 18.669286727905273, 'timing/data': 0.06563760398421437, 'timing/model': 2.024934487999417, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 0.74}) +[2026-06-23 07:29:03,660][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.11490444093942642, 'train/grad_norm_pre_clip': 22.45429039001465, 'timing/data': 0.0011824750108644366, 'timing/model': 2.030296114011435, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 0.75}) +[2026-06-23 07:29:07,777][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.15712951123714447, 'train/grad_norm_pre_clip': 24.379634857177734, 'timing/data': 0.0012329670134931803, 'timing/model': 2.0534684579761233, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 0.75}) +[2026-06-23 07:29:11,791][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.17294323444366455, 'train/grad_norm_pre_clip': 22.947561264038086, 'timing/data': 0.0013405659992713481, 'timing/model': 2.0388454890053254, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 0.75}) +[2026-06-23 07:29:15,868][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.17188872396945953, 'train/grad_norm_pre_clip': 30.944873809814453, 'timing/data': 0.06453781301388517, 'timing/model': 2.0341907929978333, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 0.75}) +[2026-06-23 07:29:19,866][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.1608314961194992, 'train/grad_norm_pre_clip': 37.867000579833984, 'timing/data': 0.001388604985550046, 'timing/model': 2.0159347520093434, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 0.75}) +[2026-06-23 07:29:23,957][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.20608772337436676, 'train/grad_norm_pre_clip': 22.62769317626953, 'timing/data': 0.0012112980184610933, 'timing/model': 2.021424284990644, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.75}) +[2026-06-23 07:29:27,952][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.13188140094280243, 'train/grad_norm_pre_clip': 27.3742618560791, 'timing/data': 0.0014343899965751916, 'timing/model': 2.01941734898719, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 0.75}) +[2026-06-23 07:29:32,023][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.09595669060945511, 'train/grad_norm_pre_clip': 19.923585891723633, 'timing/data': 0.0667197440052405, 'timing/model': 2.024385386990616, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 0.75}) +[2026-06-23 07:29:36,021][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.20221762359142303, 'train/grad_norm_pre_clip': 30.795381546020508, 'timing/data': 0.001282419019844383, 'timing/model': 2.0204329380067065, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 0.75}) +[2026-06-23 07:29:40,141][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.0763663724064827, 'train/grad_norm_pre_clip': 17.223880767822266, 'timing/data': 0.001114017009967938, 'timing/model': 2.0580272730148863, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 0.75}) +[2026-06-23 07:29:44,161][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.0939364805817604, 'train/grad_norm_pre_clip': 22.849285125732422, 'timing/data': 0.0013570270093623549, 'timing/model': 2.0403075130016077, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 0.75}) +[2026-06-23 07:29:48,218][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.05908825248479843, 'train/grad_norm_pre_clip': 15.541797637939453, 'timing/data': 0.05525071901502088, 'timing/model': 2.023668669979088, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 0.75}) +[2026-06-23 07:29:52,222][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.08676396310329437, 'train/grad_norm_pre_clip': 21.26871109008789, 'timing/data': 0.001208923989906907, 'timing/model': 2.02532715498819, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 0.75}) +[2026-06-23 07:29:56,477][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.08014320582151413, 'train/grad_norm_pre_clip': 16.618324279785156, 'timing/data': 0.0012482760066632181, 'timing/model': 2.016973948018858, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 0.75}) +[2026-06-23 07:30:00,485][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.20385754108428955, 'train/grad_norm_pre_clip': 30.61606788635254, 'timing/data': 0.0012618510227184743, 'timing/model': 2.026052458997583, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 0.75}) +[2026-06-23 07:30:04,562][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.0919027104973793, 'train/grad_norm_pre_clip': 23.642255783081055, 'timing/data': 0.05310107799596153, 'timing/model': 2.0268037479836494, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.75}) +[2026-06-23 07:30:08,555][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.07059293240308762, 'train/grad_norm_pre_clip': 11.138751983642578, 'timing/data': 0.001132211007643491, 'timing/model': 2.0177993169927504, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 0.75}) +[2026-06-23 07:30:12,636][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.183461993932724, 'train/grad_norm_pre_clip': 61.95921325683594, 'timing/data': 0.0012085420021321625, 'timing/model': 2.0174981110030785, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 0.75}) +[2026-06-23 07:30:16,718][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.10510401427745819, 'train/grad_norm_pre_clip': 15.756187438964844, 'timing/data': 0.001157758990302682, 'timing/model': 2.0194588049780577, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 0.75}) +[2026-06-23 07:30:20,776][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.13445636630058289, 'train/grad_norm_pre_clip': 37.83029556274414, 'timing/data': 0.06365075599751435, 'timing/model': 2.018843680009013, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 0.75}) +[2026-06-23 07:30:24,788][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.1514941304922104, 'train/grad_norm_pre_clip': 23.21932029724121, 'timing/data': 0.0011995960085187107, 'timing/model': 2.0357877100177575, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 0.75}) +[2026-06-23 07:30:28,871][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.07301872968673706, 'train/grad_norm_pre_clip': 11.508695602416992, 'timing/data': 0.0011239569867029786, 'timing/model': 2.018819238001015, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 0.75}) +[2026-06-23 07:30:32,864][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.07309582829475403, 'train/grad_norm_pre_clip': 18.709524154663086, 'timing/data': 0.0011952669883612543, 'timing/model': 2.0154469890112523, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 0.75}) +[2026-06-23 07:30:36,925][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.12316391617059708, 'train/grad_norm_pre_clip': 20.294946670532227, 'timing/data': 0.06705228899954818, 'timing/model': 2.0206438209861517, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 0.75}) +[2026-06-23 07:30:40,912][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.03164219483733177, 'train/grad_norm_pre_clip': 6.853984355926514, 'timing/data': 0.0016002669872250408, 'timing/model': 2.014663796988316, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 0.75}) +[2026-06-23 07:30:44,985][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.11236588656902313, 'train/grad_norm_pre_clip': 13.803800582885742, 'timing/data': 0.0012819280091207474, 'timing/model': 2.0160675510123838, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.75}) +[2026-06-23 07:30:48,977][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.12652276456356049, 'train/grad_norm_pre_clip': 19.64008140563965, 'timing/data': 0.001313135988311842, 'timing/model': 2.0168274659954477, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 0.76}) +[2026-06-23 07:30:53,268][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.07297694683074951, 'train/grad_norm_pre_clip': 12.84650993347168, 'timing/data': 0.05264424500637688, 'timing/model': 2.018987207004102, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 0.76}) +[2026-06-23 07:30:57,261][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.13830925524234772, 'train/grad_norm_pre_clip': 16.349124908447266, 'timing/data': 0.0011352569854352623, 'timing/model': 2.0165792609914206, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 0.76}) +[2026-06-23 07:31:01,337][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.0755610242486, 'train/grad_norm_pre_clip': 11.413494110107422, 'timing/data': 0.0011370709980838, 'timing/model': 2.016953863989329, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 0.76}) +[2026-06-23 07:31:05,331][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.07252825051546097, 'train/grad_norm_pre_clip': 10.341532707214355, 'timing/data': 0.001158579980256036, 'timing/model': 2.0180124509788584, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.76}) +[2026-06-23 07:31:09,374][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.056264184415340424, 'train/grad_norm_pre_clip': 18.493040084838867, 'timing/data': 0.054121404013130814, 'timing/model': 2.0145156029902864, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 0.76}) +[2026-06-23 07:31:13,369][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.1276489496231079, 'train/grad_norm_pre_clip': 21.231016159057617, 'timing/data': 0.0011525589798111469, 'timing/model': 2.0170021670055576, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 0.76}) +[2026-06-23 07:31:17,461][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.1841484159231186, 'train/grad_norm_pre_clip': 22.67080307006836, 'timing/data': 0.0012111169926356524, 'timing/model': 2.0330781940137967, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 0.76}) +[2026-06-23 07:31:21,477][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.1541348099708557, 'train/grad_norm_pre_clip': 40.899105072021484, 'timing/data': 0.001173796976218, 'timing/model': 2.0414200850063935, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 0.76}) +[2026-06-23 07:31:25,782][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.05342575162649155, 'train/grad_norm_pre_clip': 12.219738960266113, 'timing/data': 0.05551190199912526, 'timing/model': 2.266469178983243, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.76}) +[2026-06-23 07:31:29,787][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.0740712434053421, 'train/grad_norm_pre_clip': 15.040246963500977, 'timing/data': 0.0013294059899635613, 'timing/model': 2.0242842549923807, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 0.76}) +[2026-06-23 07:31:33,876][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.18174877762794495, 'train/grad_norm_pre_clip': 15.324129104614258, 'timing/data': 0.0016331360093317926, 'timing/model': 2.024323146993993, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.76}) +[2026-06-23 07:31:37,881][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.064354807138443, 'train/grad_norm_pre_clip': 49.02402877807617, 'timing/data': 0.0012637049949262291, 'timing/model': 2.025772847002372, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 0.76}) +[2026-06-23 07:31:41,946][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.0599064901471138, 'train/grad_norm_pre_clip': 20.491918563842773, 'timing/data': 0.05742565801483579, 'timing/model': 2.0238329129933845, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 0.76}) +[2026-06-23 07:31:45,952][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.12126482278108597, 'train/grad_norm_pre_clip': 34.2992057800293, 'timing/data': 0.0013794279948342592, 'timing/model': 2.0245496389979962, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 0.76}) +[2026-06-23 07:31:50,047][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.03837939351797104, 'train/grad_norm_pre_clip': 12.48835277557373, 'timing/data': 0.001246211992111057, 'timing/model': 2.026465789997019, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 0.76}) +[2026-06-23 07:31:54,056][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.1342976838350296, 'train/grad_norm_pre_clip': 40.25448226928711, 'timing/data': 0.0012877690023742616, 'timing/model': 2.026802033011336, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 0.76}) +[2026-06-23 07:31:58,257][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.15052899718284607, 'train/grad_norm_pre_clip': 30.625734329223633, 'timing/data': 0.05481217001215555, 'timing/model': 2.164991880999878, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 0.76}) +[2026-06-23 07:32:02,265][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.05841654911637306, 'train/grad_norm_pre_clip': 15.895757675170898, 'timing/data': 0.0012038339918944985, 'timing/model': 2.0272826830041595, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 0.76}) +[2026-06-23 07:32:06,372][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.23315614461898804, 'train/grad_norm_pre_clip': 45.600730895996094, 'timing/data': 0.001204244006657973, 'timing/model': 2.0369467370037455, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.76}) +[2026-06-23 07:32:10,375][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.2572251558303833, 'train/grad_norm_pre_clip': 35.991798400878906, 'timing/data': 0.0014367540134117007, 'timing/model': 2.0203630440228153, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 0.76}) +[2026-06-23 07:32:14,447][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.10881231725215912, 'train/grad_norm_pre_clip': 23.86686134338379, 'timing/data': 0.06760464000399224, 'timing/model': 2.022947790013859, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 0.76}) +[2026-06-23 07:32:18,455][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.08212170004844666, 'train/grad_norm_pre_clip': 16.574077606201172, 'timing/data': 0.0012798639945685863, 'timing/model': 2.026255559991114, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 0.76}) +[2026-06-23 07:32:22,545][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.05603359639644623, 'train/grad_norm_pre_clip': 14.886651992797852, 'timing/data': 0.0012536159774754196, 'timing/model': 2.0190737459925003, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 0.76}) +[2026-06-23 07:32:26,563][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.14503197371959686, 'train/grad_norm_pre_clip': 19.312789916992188, 'timing/data': 0.0014439969963859767, 'timing/model': 2.037046986981295, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 0.76}) +[2026-06-23 07:32:30,873][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.13165754079818726, 'train/grad_norm_pre_clip': 30.699167251586914, 'timing/data': 0.05597677099285647, 'timing/model': 2.277313974016579, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 0.76}) +[2026-06-23 07:32:34,892][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.06074442341923714, 'train/grad_norm_pre_clip': 14.882346153259277, 'timing/data': 0.0013906789827160537, 'timing/model': 2.039624080993235, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 0.77}) +[2026-06-23 07:32:38,986][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.1015259400010109, 'train/grad_norm_pre_clip': 65.33094787597656, 'timing/data': 0.001231864996952936, 'timing/model': 2.0236924120108597, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 0.77}) +[2026-06-23 07:32:42,986][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.2441401183605194, 'train/grad_norm_pre_clip': 35.44028854370117, 'timing/data': 0.0011509759933687747, 'timing/model': 2.025534185988363, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 0.77}) +[2026-06-23 07:32:47,043][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.11927363276481628, 'train/grad_norm_pre_clip': 28.1634464263916, 'timing/data': 0.06485738599440083, 'timing/model': 2.019980687997304, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.77}) +[2026-06-23 07:32:51,042][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.07036110758781433, 'train/grad_norm_pre_clip': 20.406373977661133, 'timing/data': 0.001222409016918391, 'timing/model': 2.02144741898519, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 0.77}) +[2026-06-23 07:32:55,134][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.12437495589256287, 'train/grad_norm_pre_clip': 17.622732162475586, 'timing/data': 0.0012195029994472861, 'timing/model': 2.025649891991634, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 0.77}) +[2026-06-23 07:32:59,140][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.05184190347790718, 'train/grad_norm_pre_clip': 14.072336196899414, 'timing/data': 0.0014075000071898103, 'timing/model': 2.025752203015145, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 0.77}) +[2026-06-23 07:33:03,211][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.1961764693260193, 'train/grad_norm_pre_clip': 28.229990005493164, 'timing/data': 0.0675794440030586, 'timing/model': 2.024130156991305, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 0.77}) +[2026-06-23 07:33:07,472][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.085858553647995, 'train/grad_norm_pre_clip': 17.807973861694336, 'timing/data': 0.001209934998769313, 'timing/model': 2.0269085809995886, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 0.77}) +[2026-06-23 07:33:11,553][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.05862504616379738, 'train/grad_norm_pre_clip': 16.057682037353516, 'timing/data': 0.0012061980087310076, 'timing/model': 2.0214293510071, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 0.77}) +[2026-06-23 07:33:15,565][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.16100621223449707, 'train/grad_norm_pre_clip': 21.376970291137695, 'timing/data': 0.00122413199278526, 'timing/model': 2.0344714840175584, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 0.77}) +[2026-06-23 07:33:19,652][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.047941166907548904, 'train/grad_norm_pre_clip': 12.45949649810791, 'timing/data': 0.0678886059904471, 'timing/model': 2.020911897008773, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 0.77}) +[2026-06-23 07:33:23,653][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.18938139081001282, 'train/grad_norm_pre_clip': 37.935394287109375, 'timing/data': 0.0012407420144882053, 'timing/model': 2.0224761069985107, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 0.77}) +[2026-06-23 07:33:27,770][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.10249918699264526, 'train/grad_norm_pre_clip': 25.972375869750977, 'timing/data': 0.0012509209918789566, 'timing/model': 2.041604362981161, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.77}) +[2026-06-23 07:33:31,794][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.054176297038793564, 'train/grad_norm_pre_clip': 12.49226188659668, 'timing/data': 0.0013120039948262274, 'timing/model': 2.042856025014771, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 0.77}) +[2026-06-23 07:33:35,857][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.0977112352848053, 'train/grad_norm_pre_clip': 27.598661422729492, 'timing/data': 0.053838802996324375, 'timing/model': 2.0233089000103064, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 0.77}) +[2026-06-23 07:33:39,984][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.051531463861465454, 'train/grad_norm_pre_clip': 12.560535430908203, 'timing/data': 0.0013754710089415312, 'timing/model': 2.1434523620118853, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 0.77}) +[2026-06-23 07:33:44,077][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.07070041447877884, 'train/grad_norm_pre_clip': 10.890467643737793, 'timing/data': 0.0017544309957884252, 'timing/model': 2.023812139988877, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 0.77}) +[2026-06-23 07:33:48,080][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.14215177297592163, 'train/grad_norm_pre_clip': 27.003355026245117, 'timing/data': 0.0012632829893846065, 'timing/model': 2.0244751069985796, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 0.77}) +[2026-06-23 07:33:52,149][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.09705695509910583, 'train/grad_norm_pre_clip': 25.440937042236328, 'timing/data': 0.06596678498317488, 'timing/model': 2.0221018840093166, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 0.77}) +[2026-06-23 07:33:56,148][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.08109527081251144, 'train/grad_norm_pre_clip': 29.588544845581055, 'timing/data': 0.0011746300151571631, 'timing/model': 2.020616471010726, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 0.77}) +[2026-06-23 07:34:00,231][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.09562699496746063, 'train/grad_norm_pre_clip': 24.305538177490234, 'timing/data': 0.0011291260016150773, 'timing/model': 2.019498826994095, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 0.77}) +[2026-06-23 07:34:04,228][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.07386325299739838, 'train/grad_norm_pre_clip': 15.349601745605469, 'timing/data': 0.0012031029909849167, 'timing/model': 2.019649677007692, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 0.77}) +[2026-06-23 07:34:08,274][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2000 +[2026-06-23 07:36:03,392][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2000 +[2026-06-23 07:36:03,393][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2000 +[2026-06-23 07:38:10,412][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2000 +[2026-06-23 07:51:05,836][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-06-23 07:51:05,838][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.03856541961431503, 'train/grad_norm_pre_clip': 11.653499603271484, 'eval/action_loss/seconds': 115.11478335398715, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04902122914791107, 'eval/latency_2/loss': 0.04902122914791107, 'eval/flappy/loss': 0.04902122914791107, 'eval/flappy/latency_2/loss': 0.04902122914791107, 'eval/flap_precision': 0.9395348837209302, 'eval/flap_recall': 0.9541160593792173, 'eval/flap_f1': 0.9467693337797122, 'eval/flap_rate_model': 0.20902777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9467693337797122, 'eval/latency_2/flap_precision': 0.9395348837209302, 'eval/latency_2/flap_recall': 0.9541160593792173, 'eval/latency_2/flap_f1': 0.9467693337797122, 'eval/latency_2/flap_rate_model': 0.20902777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9467693337797122, 'eval/action_classification/data_wait_seconds': 50.589622497558594, 'eval/action_classification/predict_seconds': 74.71989440917969, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 200.6800000000103, 'rl_games_eval/mid_train/mean_length': 1635.4, 'rl_games_eval/mid_train/std_reward': 200.38831702473323, 'rl_games_eval/mid_train/std_length': 1614.02622035703, 'rl_games_eval/mid_train/macro_mean_reward': 200.6800000000103, 'rl_games_eval/mid_train/macro_mean_length': 1635.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 200.6800000000103, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1635.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 200.38831702473323, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1614.02622035703, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 200.6800000000103, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.05176570298499428, 'timing/model': 2.0163454040011857, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.77}) +[2026-06-23 07:51:05,838][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_2000_state +[2026-06-23 07:51:05,846][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 07:51:14,790][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_2000_state/model.safetensors +[2026-06-23 07:51:18,710][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_2000_state/optimizer.bin +[2026-06-23 07:51:18,711][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_2000_state/sampler.bin +[2026-06-23 07:51:18,711][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-23 07:51:18,712][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-23 07:51:18,852][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-23 07:51:18,853][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 07:51:18,876][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 07:51:22,735][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.15113013982772827, 'train/grad_norm_pre_clip': 20.75532341003418, 'timing/data': 0.0011897289950866252, 'timing/model': 1.9504626330162864, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 0.77}) +[2026-06-23 07:51:26,698][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.07067011296749115, 'train/grad_norm_pre_clip': 18.33113670349121, 'timing/data': 0.0011878249933943152, 'timing/model': 1.9565369990014005, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 0.77}) +[2026-06-23 07:51:30,575][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.20694802701473236, 'train/grad_norm_pre_clip': 53.43598175048828, 'timing/data': 0.0014078120002523065, 'timing/model': 1.9620557480084244, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 0.78}) +[2026-06-23 07:51:34,552][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.13333889842033386, 'train/grad_norm_pre_clip': 39.83509063720703, 'timing/data': 0.0682934680080507, 'timing/model': 1.9735868609859608, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 0.78}) +[2026-06-23 07:51:38,447][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.3313956558704376, 'train/grad_norm_pre_clip': 48.8665885925293, 'timing/data': 0.0013858820020686835, 'timing/model': 1.967323260003468, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 0.78}) +[2026-06-23 07:51:42,433][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.175005704164505, 'train/grad_norm_pre_clip': 37.89889907836914, 'timing/data': 0.0012275290209800005, 'timing/model': 1.9682874610007275, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 0.78}) +[2026-06-23 07:51:46,618][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.07336797565221786, 'train/grad_norm_pre_clip': 22.33132553100586, 'timing/data': 0.0016616320062894374, 'timing/model': 2.2560476469807327, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 0.78}) +[2026-06-23 07:51:50,589][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.06709964573383331, 'train/grad_norm_pre_clip': 11.261025428771973, 'timing/data': 0.06437449197983369, 'timing/model': 1.975916077994043, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 0.78}) +[2026-06-23 07:51:54,505][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.10482210665941238, 'train/grad_norm_pre_clip': 14.16202449798584, 'timing/data': 0.001376383996102959, 'timing/model': 1.980806205974659, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 0.78}) +[2026-06-23 07:51:58,512][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.0684165433049202, 'train/grad_norm_pre_clip': 15.053064346313477, 'timing/data': 0.001633389008929953, 'timing/model': 1.9789411170058884, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.78}) +[2026-06-23 07:52:02,439][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.1403588503599167, 'train/grad_norm_pre_clip': 31.379310607910156, 'timing/data': 0.0014273279812186956, 'timing/model': 1.9845211490173824, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.78}) +[2026-06-23 07:52:06,433][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.12261167913675308, 'train/grad_norm_pre_clip': 35.5017204284668, 'timing/data': 0.06525569001678377, 'timing/model': 1.9826431070105173, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 0.78}) +[2026-06-23 07:52:10,368][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.15825851261615753, 'train/grad_norm_pre_clip': 30.128768920898438, 'timing/data': 0.0015943269827403128, 'timing/model': 1.9869983560056426, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 0.78}) +[2026-06-23 07:52:14,410][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.043555378913879395, 'train/grad_norm_pre_clip': 15.185510635375977, 'timing/data': 0.0014008189900778234, 'timing/model': 2.007459895015927, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 0.78}) +[2026-06-23 07:52:18,348][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.09017941355705261, 'train/grad_norm_pre_clip': 18.753131866455078, 'timing/data': 0.0012737340002786368, 'timing/model': 1.9896604029927403, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 0.78}) +[2026-06-23 07:52:22,394][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.025951674208045006, 'train/grad_norm_pre_clip': 6.390038967132568, 'timing/data': 0.0657704699842725, 'timing/model': 1.989729390013963, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 0.78}) +[2026-06-23 07:52:26,334][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.09405867010354996, 'train/grad_norm_pre_clip': 16.806848526000977, 'timing/data': 0.0012007799814455211, 'timing/model': 1.9921819930023048, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 0.78}) +[2026-06-23 07:52:30,385][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.21286122500896454, 'train/grad_norm_pre_clip': 51.814186096191406, 'timing/data': 0.0012844339944422245, 'timing/model': 2.010122201987542, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 0.78}) +[2026-06-23 07:52:34,334][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.059564389288425446, 'train/grad_norm_pre_clip': 19.998641967773438, 'timing/data': 0.0016033939900808036, 'timing/model': 1.9956290120026097, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 0.78}) +[2026-06-23 07:52:38,357][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.1725916564464569, 'train/grad_norm_pre_clip': 41.73991012573242, 'timing/data': 0.07299237698316574, 'timing/model': 1.9963741379906423, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.78}) +[2026-06-23 07:52:42,305][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.18619252741336823, 'train/grad_norm_pre_clip': 37.64531707763672, 'timing/data': 0.0012424059968907386, 'timing/model': 1.9962203729955945, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 0.78}) +[2026-06-23 07:52:46,354][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.09470568597316742, 'train/grad_norm_pre_clip': 22.907060623168945, 'timing/data': 0.0013496449973899871, 'timing/model': 2.0025459149910603, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 0.78}) +[2026-06-23 07:52:50,342][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.09261725842952728, 'train/grad_norm_pre_clip': 13.022844314575195, 'timing/data': 0.0015991849941201508, 'timing/model': 2.0197759210132062, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 0.78}) +[2026-06-23 07:52:54,494][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.12051477283239365, 'train/grad_norm_pre_clip': 36.061553955078125, 'timing/data': 0.06011160701746121, 'timing/model': 2.136927507002838, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 0.78}) +[2026-06-23 07:52:58,472][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.15626299381256104, 'train/grad_norm_pre_clip': 19.951221466064453, 'timing/data': 0.0013364399783313274, 'timing/model': 2.018414052989101, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.78}) +[2026-06-23 07:53:02,524][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.1561998873949051, 'train/grad_norm_pre_clip': 27.9022216796875, 'timing/data': 0.0012078620202373713, 'timing/model': 2.0025432569964323, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 0.78}) +[2026-06-23 07:53:06,493][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.08129072189331055, 'train/grad_norm_pre_clip': 26.820894241333008, 'timing/data': 0.0017922419938258827, 'timing/model': 2.0031852729734965, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 0.78}) +[2026-06-23 07:53:10,536][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.1111828163266182, 'train/grad_norm_pre_clip': 27.37808609008789, 'timing/data': 0.07160460800514556, 'timing/model': 2.0088691290002316, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 0.78}) +[2026-06-23 07:53:14,533][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.07880020141601562, 'train/grad_norm_pre_clip': 21.033702850341797, 'timing/data': 0.001501344988355413, 'timing/model': 2.0282263169938233, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 0.79}) +[2026-06-23 07:53:18,604][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.10399278253316879, 'train/grad_norm_pre_clip': 22.876291275024414, 'timing/data': 0.0013129159924574196, 'timing/model': 2.004295644990634, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.79}) +[2026-06-23 07:53:22,573][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.11210227012634277, 'train/grad_norm_pre_clip': 16.3231201171875, 'timing/data': 0.001295465015573427, 'timing/model': 2.0072949590103235, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 0.79}) +[2026-06-23 07:53:26,859][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.10303474217653275, 'train/grad_norm_pre_clip': 21.182418823242188, 'timing/data': 0.05655116299749352, 'timing/model': 2.257747580995783, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 0.79}) +[2026-06-23 07:53:30,825][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.055967673659324646, 'train/grad_norm_pre_clip': 22.522632598876953, 'timing/data': 0.001217210025060922, 'timing/model': 2.0045441040128935, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 0.79}) +[2026-06-23 07:53:34,886][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.23851490020751953, 'train/grad_norm_pre_clip': 41.62724685668945, 'timing/data': 0.0012215980095788836, 'timing/model': 2.007530303002568, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 0.79}) +[2026-06-23 07:53:38,880][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.1687352955341339, 'train/grad_norm_pre_clip': 41.421043395996094, 'timing/data': 0.0013906499953009188, 'timing/model': 2.0314540610124823, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 0.79}) +[2026-06-23 07:53:42,923][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.1245417669415474, 'train/grad_norm_pre_clip': 37.283138275146484, 'timing/data': 0.057986655010608956, 'timing/model': 2.018869582010666, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 0.79}) +[2026-06-23 07:53:46,904][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.19221261143684387, 'train/grad_norm_pre_clip': 42.887760162353516, 'timing/data': 0.0012238819908816367, 'timing/model': 2.013288227026351, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 0.79}) +[2026-06-23 07:53:50,997][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.18632271885871887, 'train/grad_norm_pre_clip': 49.706058502197266, 'timing/data': 0.0011932459892705083, 'timing/model': 2.0260484110040125, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 0.79}) +[2026-06-23 07:53:54,996][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.1308482140302658, 'train/grad_norm_pre_clip': 13.754097938537598, 'timing/data': 0.0014292520063463598, 'timing/model': 2.0263155860011466, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 0.79}) +[2026-06-23 07:53:59,056][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.09661202877759933, 'train/grad_norm_pre_clip': 15.837423324584961, 'timing/data': 0.06523872798425145, 'timing/model': 2.015225980023388, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.79}) +[2026-06-23 07:54:03,280][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.14310239255428314, 'train/grad_norm_pre_clip': 16.762107849121094, 'timing/data': 0.0012747359869536012, 'timing/model': 2.0173247790080495, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 0.79}) +[2026-06-23 07:54:07,376][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.07725954800844193, 'train/grad_norm_pre_clip': 11.122297286987305, 'timing/data': 0.001263505022507161, 'timing/model': 2.0390179840032943, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 0.79}) +[2026-06-23 07:54:11,376][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.09697424620389938, 'train/grad_norm_pre_clip': 30.458023071289062, 'timing/data': 0.001251923997187987, 'timing/model': 2.0281135810073465, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 0.79}) +[2026-06-23 07:54:15,416][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.19844530522823334, 'train/grad_norm_pre_clip': 43.57818603515625, 'timing/data': 0.06235089001711458, 'timing/model': 2.0098560269980226, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 0.79}) +[2026-06-23 07:54:19,420][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.12101907283067703, 'train/grad_norm_pre_clip': 29.333295822143555, 'timing/data': 0.0014375770115293562, 'timing/model': 2.02054015500471, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 0.79}) +[2026-06-23 07:54:23,710][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.047466374933719635, 'train/grad_norm_pre_clip': 11.702837944030762, 'timing/data': 0.0014446389977820218, 'timing/model': 2.012885447998997, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 0.79}) +[2026-06-23 07:54:27,712][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.1823398917913437, 'train/grad_norm_pre_clip': 20.115571975708008, 'timing/data': 0.0012740950041916221, 'timing/model': 2.0282656960189342, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 0.79}) +[2026-06-23 07:54:32,064][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.1086355522274971, 'train/grad_norm_pre_clip': 21.013261795043945, 'timing/data': 0.33198143797926605, 'timing/model': 2.050563886994496, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 0.79}) +[2026-06-23 07:54:36,147][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.04397089406847954, 'train/grad_norm_pre_clip': 7.320507526397705, 'timing/data': 0.0012479560100473464, 'timing/model': 2.108861707005417, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 0.79}) +[2026-06-23 07:54:40,244][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.06681843847036362, 'train/grad_norm_pre_clip': 9.070764541625977, 'timing/data': 0.0015859119885135442, 'timing/model': 2.030753856000956, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.79}) +[2026-06-23 07:54:44,246][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.16309553384780884, 'train/grad_norm_pre_clip': 23.4572811126709, 'timing/data': 0.001241093996213749, 'timing/model': 2.0285912300169002, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 0.79}) +[2026-06-23 07:54:48,282][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.11948313564062119, 'train/grad_norm_pre_clip': 20.331958770751953, 'timing/data': 0.05590154801029712, 'timing/model': 2.0102787769865245, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 0.79}) +[2026-06-23 07:54:52,272][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.10264279693365097, 'train/grad_norm_pre_clip': 22.32430648803711, 'timing/data': 0.0011405179975554347, 'timing/model': 2.016771547991084, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 0.79}) +[2026-06-23 07:54:56,360][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.08374606817960739, 'train/grad_norm_pre_clip': 16.581214904785156, 'timing/data': 0.0012070000229869038, 'timing/model': 2.017315103992587, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 0.79}) +[2026-06-23 07:55:00,362][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.13225308060646057, 'train/grad_norm_pre_clip': 18.24708366394043, 'timing/data': 0.0012653190060518682, 'timing/model': 2.0268568309838884, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 0.8}) +[2026-06-23 07:55:04,414][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.14365911483764648, 'train/grad_norm_pre_clip': 21.365751266479492, 'timing/data': 0.060902711004018784, 'timing/model': 2.015318448015023, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 0.8}) +[2026-06-23 07:55:08,413][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.08609626442193985, 'train/grad_norm_pre_clip': 12.000694274902344, 'timing/data': 0.0012728330038953573, 'timing/model': 2.020174344012048, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 0.8}) +[2026-06-23 07:55:12,752][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.15771077573299408, 'train/grad_norm_pre_clip': 19.790922164916992, 'timing/data': 0.0013822049950249493, 'timing/model': 2.0305875000194646, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 0.8}) +[2026-06-23 07:55:16,749][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.08766165375709534, 'train/grad_norm_pre_clip': 35.39714050292969, 'timing/data': 0.001269717002287507, 'timing/model': 2.0195414979825728, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 0.8}) +[2026-06-23 07:55:20,814][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.18835733830928802, 'train/grad_norm_pre_clip': 35.51871109008789, 'timing/data': 0.06462648100568913, 'timing/model': 2.021061338018626, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.8}) +[2026-06-23 07:55:24,813][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.036318339407444, 'train/grad_norm_pre_clip': 11.166196823120117, 'timing/data': 0.0012343609996605664, 'timing/model': 2.021692496986361, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 0.8}) +[2026-06-23 07:55:28,903][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.06681076437234879, 'train/grad_norm_pre_clip': 17.238372802734375, 'timing/data': 0.0012755870120599866, 'timing/model': 2.023864632996265, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 0.8}) +[2026-06-23 07:55:32,902][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.14563794434070587, 'train/grad_norm_pre_clip': 18.195621490478516, 'timing/data': 0.001402511988999322, 'timing/model': 2.019822024012683, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 0.8}) +[2026-06-23 07:55:36,954][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.07443959265947342, 'train/grad_norm_pre_clip': 16.99265480041504, 'timing/data': 0.055433289002394304, 'timing/model': 2.0194659080007114, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 0.8}) +[2026-06-23 07:55:40,964][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.1108424961566925, 'train/grad_norm_pre_clip': 14.995809555053711, 'timing/data': 0.0013254100049380213, 'timing/model': 2.0312048959895037, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 0.8}) +[2026-06-23 07:55:45,311][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.06279204785823822, 'train/grad_norm_pre_clip': 10.51665210723877, 'timing/data': 0.0013150999729987234, 'timing/model': 2.280812409997452, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 0.8}) +[2026-06-23 07:55:49,313][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.12463649362325668, 'train/grad_norm_pre_clip': 16.615869522094727, 'timing/data': 0.0013053830189164728, 'timing/model': 2.0229603429907, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 0.8}) +[2026-06-23 07:55:53,373][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.03333102539181709, 'train/grad_norm_pre_clip': 8.060365676879883, 'timing/data': 0.055594076024135575, 'timing/model': 2.0279312829952687, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 0.8}) +[2026-06-23 07:55:57,390][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.11918707937002182, 'train/grad_norm_pre_clip': 15.493976593017578, 'timing/data': 0.0012157669989392161, 'timing/model': 2.0364021490095183, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 0.8}) +[2026-06-23 07:56:01,499][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.05319713056087494, 'train/grad_norm_pre_clip': 12.184823989868164, 'timing/data': 0.0012529359955806285, 'timing/model': 2.037316051981179, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.8}) +[2026-06-23 07:56:05,539][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.17386798560619354, 'train/grad_norm_pre_clip': 44.496124267578125, 'timing/data': 0.0013768550124950707, 'timing/model': 2.0403658610011917, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 0.8}) +[2026-06-23 07:56:09,612][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.045160725712776184, 'train/grad_norm_pre_clip': 9.279707908630371, 'timing/data': 0.06993144200532697, 'timing/model': 2.0233099019969814, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 0.8}) +[2026-06-23 07:56:13,615][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.17382214963436127, 'train/grad_norm_pre_clip': 33.753875732421875, 'timing/data': 0.0012597080203704536, 'timing/model': 2.022686111013172, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 0.8}) +[2026-06-23 07:56:17,799][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.03052111715078354, 'train/grad_norm_pre_clip': 9.748082160949707, 'timing/data': 0.0011600040015764534, 'timing/model': 2.1206933229987044, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 0.8}) +[2026-06-23 07:56:21,823][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.04824819043278694, 'train/grad_norm_pre_clip': 10.74389362335205, 'timing/data': 0.0013696120004169643, 'timing/model': 2.046141684986651, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 0.8}) +[2026-06-23 07:56:25,871][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.1570519357919693, 'train/grad_norm_pre_clip': 17.119932174682617, 'timing/data': 0.056047032005153596, 'timing/model': 2.0192265459918417, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 0.8}) +[2026-06-23 07:56:29,884][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.13827788829803467, 'train/grad_norm_pre_clip': 27.295724868774414, 'timing/data': 0.001123385998653248, 'timing/model': 2.0211378810054157, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 0.8}) +[2026-06-23 07:56:33,992][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.07907942682504654, 'train/grad_norm_pre_clip': 26.826663970947266, 'timing/data': 0.0011316610034555197, 'timing/model': 2.0341625660075806, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 0.8}) +[2026-06-23 07:56:37,990][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.06198801472783089, 'train/grad_norm_pre_clip': 19.16175651550293, 'timing/data': 0.0014528750034514815, 'timing/model': 2.018554171983851, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 0.8}) +[2026-06-23 07:56:42,074][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.0652279332280159, 'train/grad_norm_pre_clip': 17.513633728027344, 'timing/data': 0.055543909984407946, 'timing/model': 2.0433923699893057, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.8}) +[2026-06-23 07:56:46,080][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.1570742279291153, 'train/grad_norm_pre_clip': 46.303653717041016, 'timing/data': 0.001245341991307214, 'timing/model': 2.0247671910037752, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 0.81}) +[2026-06-23 07:56:50,170][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.17474228143692017, 'train/grad_norm_pre_clip': 43.979373931884766, 'timing/data': 0.0012274380133021623, 'timing/model': 2.023827306984458, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 0.81}) +[2026-06-23 07:56:54,385][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.04002269729971886, 'train/grad_norm_pre_clip': 11.718382835388184, 'timing/data': 0.0012627140094991773, 'timing/model': 2.017226391006261, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 0.81}) +[2026-06-23 07:56:58,446][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.04888566955924034, 'train/grad_norm_pre_clip': 18.038972854614258, 'timing/data': 0.05443490098696202, 'timing/model': 2.0278171870158985, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 0.81}) +[2026-06-23 07:57:02,463][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.051498401910066605, 'train/grad_norm_pre_clip': 12.098221778869629, 'timing/data': 0.0012245129910297692, 'timing/model': 2.0371201829984784, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 0.81}) +[2026-06-23 07:57:06,567][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.09713888168334961, 'train/grad_norm_pre_clip': 19.689485549926758, 'timing/data': 0.0011914820061065257, 'timing/model': 2.041872968984535, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 0.81}) +[2026-06-23 07:57:10,594][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.09084941446781158, 'train/grad_norm_pre_clip': 24.747926712036133, 'timing/data': 0.0013803709880448878, 'timing/model': 2.0376965049945284, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 0.81}) +[2026-06-23 07:57:14,642][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.043969329446554184, 'train/grad_norm_pre_clip': 9.96042537689209, 'timing/data': 0.05717839999124408, 'timing/model': 2.0180930299975444, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 0.81}) +[2026-06-23 07:57:18,637][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.20337484776973724, 'train/grad_norm_pre_clip': 21.314470291137695, 'timing/data': 0.0011283660132903606, 'timing/model': 2.0209292950166855, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 0.81}) +[2026-06-23 07:57:22,728][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.10306183993816376, 'train/grad_norm_pre_clip': 24.76543617248535, 'timing/data': 0.0013592519972007722, 'timing/model': 2.022244238993153, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.81}) +[2026-06-23 07:57:26,977][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.032024506479501724, 'train/grad_norm_pre_clip': 8.03846263885498, 'timing/data': 0.0012936410203110427, 'timing/model': 2.0185370709805284, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 0.81}) +[2026-06-23 07:57:31,035][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.06913473457098007, 'train/grad_norm_pre_clip': 13.29747200012207, 'timing/data': 0.05655699098133482, 'timing/model': 2.0232356780034024, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 0.81}) +[2026-06-23 07:57:35,036][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.03275195509195328, 'train/grad_norm_pre_clip': 12.035168647766113, 'timing/data': 0.0012927190109621733, 'timing/model': 2.0210876720084343, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 0.81}) +[2026-06-23 07:57:39,122][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.09622668474912643, 'train/grad_norm_pre_clip': 18.813587188720703, 'timing/data': 0.001126932998886332, 'timing/model': 2.0194537960051093, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 0.81}) +[2026-06-23 07:57:43,124][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.029355695471167564, 'train/grad_norm_pre_clip': 11.375799179077148, 'timing/data': 0.0011527500173542649, 'timing/model': 2.026165374001721, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.81}) +[2026-06-23 07:57:47,191][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.045872341841459274, 'train/grad_norm_pre_clip': 12.715685844421387, 'timing/data': 0.05756365100387484, 'timing/model': 2.022026388993254, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 0.81}) +[2026-06-23 07:57:51,203][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.027581868693232536, 'train/grad_norm_pre_clip': 14.7161283493042, 'timing/data': 0.0014281189942266792, 'timing/model': 2.0314450279984158, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 0.81}) +[2026-06-23 07:57:55,297][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.12547898292541504, 'train/grad_norm_pre_clip': 22.899723052978516, 'timing/data': 0.0013415990106295794, 'timing/model': 2.024796536017675, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 0.81}) +[2026-06-23 07:57:59,366][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.1648682802915573, 'train/grad_norm_pre_clip': 40.28151321411133, 'timing/data': 0.001481156999943778, 'timing/model': 2.043565917003434, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 0.81}) +[2026-06-23 07:58:03,430][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2100 +[2026-06-23 07:59:57,508][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2100 +[2026-06-23 07:59:57,509][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2100 +[2026-06-23 08:02:04,414][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2100 +[2026-06-23 08:02:04,416][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.06273073703050613, 'train/grad_norm_pre_clip': 19.00798797607422, 'eval/action_loss/seconds': 114.07399601102225, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06072454899549484, 'eval/latency_2/loss': 0.06072454899549484, 'eval/flappy/loss': 0.06072454899549484, 'eval/flappy/latency_2/loss': 0.06072454899549484, 'eval/flap_precision': 0.9702127659574468, 'eval/flap_recall': 0.9230769230769231, 'eval/flap_f1': 0.9460580912863071, 'eval/flap_rate_model': 0.19583333333333333, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9460580912863072, 'eval/latency_2/flap_precision': 0.9702127659574468, 'eval/latency_2/flap_recall': 0.9230769230769231, 'eval/latency_2/flap_f1': 0.9460580912863071, 'eval/latency_2/flap_rate_model': 0.19583333333333333, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9460580912863072, 'eval/action_classification/data_wait_seconds': 50.40359115600586, 'eval/action_classification/predict_seconds': 74.90573120117188, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0570887720095925, 'timing/model': 2.0233475159911904, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.81}) +[2026-06-23 08:02:08,362][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.06586570292711258, 'train/grad_norm_pre_clip': 24.761178970336914, 'timing/data': 0.0011978339753113687, 'timing/model': 1.9929408250027336, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 0.81}) +[2026-06-23 08:02:12,385][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.45049965381622314, 'train/grad_norm_pre_clip': 85.95536041259766, 'timing/data': 0.001135016995249316, 'timing/model': 1.9910289779945742, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 0.81}) +[2026-06-23 08:02:16,579][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.08104439824819565, 'train/grad_norm_pre_clip': 32.017417907714844, 'timing/data': 0.0011390550062060356, 'timing/model': 1.9982525430095848, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 0.81}) +[2026-06-23 08:02:20,606][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.13647735118865967, 'train/grad_norm_pre_clip': 31.178747177124023, 'timing/data': 0.07591126600163989, 'timing/model': 1.9967584589903709, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 0.81}) +[2026-06-23 08:02:24,781][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.06938102096319199, 'train/grad_norm_pre_clip': 26.3741397857666, 'timing/data': 0.001190608978504315, 'timing/model': 1.9998385670187417, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 0.81}) +[2026-06-23 08:02:28,841][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.12746138870716095, 'train/grad_norm_pre_clip': 31.75835609436035, 'timing/data': 0.0016964849783107638, 'timing/model': 1.999697613005992, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 0.82}) +[2026-06-23 08:02:32,794][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.10340302437543869, 'train/grad_norm_pre_clip': 20.203594207763672, 'timing/data': 0.0012346309958957136, 'timing/model': 1.9997188390116207, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 0.82}) +[2026-06-23 08:02:36,808][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.061902161687612534, 'train/grad_norm_pre_clip': 12.584741592407227, 'timing/data': 0.055419869982870296, 'timing/model': 2.0004851419944316, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 0.82}) +[2026-06-23 08:02:40,774][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.11983087658882141, 'train/grad_norm_pre_clip': 23.930889129638672, 'timing/data': 0.0014097050006967038, 'timing/model': 2.0039314169844147, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.82}) +[2026-06-23 08:02:44,837][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.06942516565322876, 'train/grad_norm_pre_clip': 16.809354782104492, 'timing/data': 0.0012983990018256009, 'timing/model': 2.0042015229992103, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.82}) +[2026-06-23 08:02:48,807][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.0384550467133522, 'train/grad_norm_pre_clip': 15.876099586486816, 'timing/data': 0.0013070449931547046, 'timing/model': 2.0066864209948108, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 0.82}) +[2026-06-23 08:02:52,850][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.08806709200143814, 'train/grad_norm_pre_clip': 20.92943572998047, 'timing/data': 0.06466418199124746, 'timing/model': 2.0077600319928024, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 0.82}) +[2026-06-23 08:02:56,866][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.2445034235715866, 'train/grad_norm_pre_clip': 33.649009704589844, 'timing/data': 0.0014305539953056723, 'timing/model': 2.0079161470057443, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 0.82}) +[2026-06-23 08:03:00,946][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.23932130634784698, 'train/grad_norm_pre_clip': 23.785043716430664, 'timing/data': 0.0012245539983268827, 'timing/model': 2.0235838149965275, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 0.82}) +[2026-06-23 08:03:04,926][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.10791274905204773, 'train/grad_norm_pre_clip': 22.82279396057129, 'timing/data': 0.0012576240114867687, 'timing/model': 2.011994936998235, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 0.82}) +[2026-06-23 08:03:08,963][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.04652745649218559, 'train/grad_norm_pre_clip': 15.453095436096191, 'timing/data': 0.05683829999179579, 'timing/model': 2.0142727229977027, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 0.82}) +[2026-06-23 08:03:12,940][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.22538800537586212, 'train/grad_norm_pre_clip': 16.301712036132812, 'timing/data': 0.0011972019856330007, 'timing/model': 2.0106302390049677, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 0.82}) +[2026-06-23 08:03:17,017][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.2538856565952301, 'train/grad_norm_pre_clip': 88.61681365966797, 'timing/data': 0.0012072600075043738, 'timing/model': 2.0114304150047246, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 0.82}) +[2026-06-23 08:03:21,001][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.15166014432907104, 'train/grad_norm_pre_clip': 15.267705917358398, 'timing/data': 0.0012329290038906038, 'timing/model': 2.0148158379888628, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 0.82}) +[2026-06-23 08:03:25,044][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.08954700827598572, 'train/grad_norm_pre_clip': 15.322580337524414, 'timing/data': 0.05706336599541828, 'timing/model': 2.0144878280116245, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.82}) +[2026-06-23 08:03:29,105][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.1484018862247467, 'train/grad_norm_pre_clip': 23.400436401367188, 'timing/data': 0.0013899389887228608, 'timing/model': 2.088678409025306, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 0.82}) +[2026-06-23 08:03:33,183][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.05468774959445, 'train/grad_norm_pre_clip': 18.4830379486084, 'timing/data': 0.0011887069849763066, 'timing/model': 2.015230615012115, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 0.82}) +[2026-06-23 08:03:37,188][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.15909507870674133, 'train/grad_norm_pre_clip': 35.23212432861328, 'timing/data': 0.0012309550074860454, 'timing/model': 2.0345422670070548, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 0.82}) +[2026-06-23 08:03:41,235][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.1510649174451828, 'train/grad_norm_pre_clip': 23.472612380981445, 'timing/data': 0.06017221001093276, 'timing/model': 2.0137621850008145, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 0.82}) +[2026-06-23 08:03:45,243][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.09841393679380417, 'train/grad_norm_pre_clip': 18.41680908203125, 'timing/data': 0.001193375006550923, 'timing/model': 2.031939029984642, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 0.82}) +[2026-06-23 08:03:49,335][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.15765050053596497, 'train/grad_norm_pre_clip': 61.86075973510742, 'timing/data': 0.0012095650017727166, 'timing/model': 2.0212814929836895, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 0.82}) +[2026-06-23 08:03:53,332][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.2222234606742859, 'train/grad_norm_pre_clip': 26.696168899536133, 'timing/data': 0.0012426860048435628, 'timing/model': 2.020130242977757, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 0.82}) +[2026-06-23 08:03:57,387][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.14825960993766785, 'train/grad_norm_pre_clip': 22.135644912719727, 'timing/data': 0.056980321009177715, 'timing/model': 2.0216817470209207, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 0.82}) +[2026-06-23 08:04:01,387][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.20268064737319946, 'train/grad_norm_pre_clip': 37.74281311035156, 'timing/data': 0.0012429860071279109, 'timing/model': 2.0223306529805996, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 0.82}) +[2026-06-23 08:04:05,742][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.1287766844034195, 'train/grad_norm_pre_clip': 17.77292251586914, 'timing/data': 0.0013823840126860887, 'timing/model': 2.030511186021613, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.82}) +[2026-06-23 08:04:09,745][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.2289275974035263, 'train/grad_norm_pre_clip': 52.449623107910156, 'timing/data': 0.001239789999090135, 'timing/model': 2.0277372080017813, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 0.82}) +[2026-06-23 08:04:13,798][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.12289776653051376, 'train/grad_norm_pre_clip': 15.895465850830078, 'timing/data': 0.05557629599934444, 'timing/model': 2.0204012889880687, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 0.83}) +[2026-06-23 08:04:17,801][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.17097143828868866, 'train/grad_norm_pre_clip': 17.740623474121094, 'timing/data': 0.0014270969841163605, 'timing/model': 2.021593790006591, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 0.83}) +[2026-06-23 08:04:21,894][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.10832132399082184, 'train/grad_norm_pre_clip': 13.938474655151367, 'timing/data': 0.0013759430148638785, 'timing/model': 2.0200269319757354, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 0.83}) +[2026-06-23 08:04:25,890][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.15009805560112, 'train/grad_norm_pre_clip': 24.025774002075195, 'timing/data': 0.001256782008567825, 'timing/model': 2.020042026997544, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 0.83}) +[2026-06-23 08:04:29,954][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.05183402821421623, 'train/grad_norm_pre_clip': 8.35410213470459, 'timing/data': 0.0566999530128669, 'timing/model': 2.0284190230013337, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 0.83}) +[2026-06-23 08:04:33,957][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.1468394249677658, 'train/grad_norm_pre_clip': 18.058727264404297, 'timing/data': 0.0012359540269244462, 'timing/model': 2.021416011004476, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 0.83}) +[2026-06-23 08:04:38,109][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.08975207805633545, 'train/grad_norm_pre_clip': 17.755746841430664, 'timing/data': 0.001379680004902184, 'timing/model': 2.039097179978853, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 0.83}) +[2026-06-23 08:04:42,111][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.09517291188240051, 'train/grad_norm_pre_clip': 14.01616382598877, 'timing/data': 0.0012773009948432446, 'timing/model': 2.025129828019999, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 0.83}) +[2026-06-23 08:04:46,175][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.09874194115400314, 'train/grad_norm_pre_clip': 12.090337753295898, 'timing/data': 0.06203999000717886, 'timing/model': 2.0260492829838768, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.83}) +[2026-06-23 08:04:50,176][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.13359524309635162, 'train/grad_norm_pre_clip': 20.126619338989258, 'timing/data': 0.0014280790055636317, 'timing/model': 2.022072954016039, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 0.83}) +[2026-06-23 08:04:54,266][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.08797597885131836, 'train/grad_norm_pre_clip': 21.154373168945312, 'timing/data': 0.0012122599873691797, 'timing/model': 2.0195626829809044, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 0.83}) +[2026-06-23 08:04:58,276][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.2110644280910492, 'train/grad_norm_pre_clip': 44.6531867980957, 'timing/data': 0.0012569529935717583, 'timing/model': 2.027190986002097, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 0.83}) +[2026-06-23 08:05:02,381][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.1029965728521347, 'train/grad_norm_pre_clip': 13.487200736999512, 'timing/data': 0.07189183498849161, 'timing/model': 2.050602033006726, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 0.83}) +[2026-06-23 08:05:06,402][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.060516152530908585, 'train/grad_norm_pre_clip': 10.165186882019043, 'timing/data': 0.0012473140086513013, 'timing/model': 2.038369286019588, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 0.83}) +[2026-06-23 08:05:10,630][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.055100295692682266, 'train/grad_norm_pre_clip': 10.925004959106445, 'timing/data': 0.001312615000642836, 'timing/model': 2.1424025999731384, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 0.83}) +[2026-06-23 08:05:14,635][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.13980141282081604, 'train/grad_norm_pre_clip': 13.895515441894531, 'timing/data': 0.0012850950006395578, 'timing/model': 2.018409058015095, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 0.83}) +[2026-06-23 08:05:18,712][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.04318144917488098, 'train/grad_norm_pre_clip': 9.343563079833984, 'timing/data': 0.06969742898945697, 'timing/model': 2.0276786940230522, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 0.83}) +[2026-06-23 08:05:22,713][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.16956950724124908, 'train/grad_norm_pre_clip': 20.055343627929688, 'timing/data': 0.0012421860010363162, 'timing/model': 2.0224875190178864, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 0.83}) +[2026-06-23 08:05:26,808][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.08583860099315643, 'train/grad_norm_pre_clip': 11.884903907775879, 'timing/data': 0.001257122989045456, 'timing/model': 2.023422174010193, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.83}) +[2026-06-23 08:05:30,827][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.0516020841896534, 'train/grad_norm_pre_clip': 18.04500961303711, 'timing/data': 0.0014205649786163121, 'timing/model': 2.0367013990180567, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 0.83}) +[2026-06-23 08:05:34,913][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.15219946205615997, 'train/grad_norm_pre_clip': 38.0351676940918, 'timing/data': 0.06803701000171714, 'timing/model': 2.0227487339871004, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 0.83}) +[2026-06-23 08:05:38,905][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.03987951576709747, 'train/grad_norm_pre_clip': 10.662630081176758, 'timing/data': 0.00125361600657925, 'timing/model': 2.016977246996248, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 0.83}) +[2026-06-23 08:05:43,011][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.047147639095783234, 'train/grad_norm_pre_clip': 9.728456497192383, 'timing/data': 0.0014372059958986938, 'timing/model': 2.036367724998854, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 0.83}) +[2026-06-23 08:05:47,211][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.053836170583963394, 'train/grad_norm_pre_clip': 17.683246612548828, 'timing/data': 0.0012741950049530715, 'timing/model': 2.0199865429895, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 0.83}) +[2026-06-23 08:05:51,271][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.06584808230400085, 'train/grad_norm_pre_clip': 19.267913818359375, 'timing/data': 0.06646877201274037, 'timing/model': 2.018601732008392, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 0.83}) +[2026-06-23 08:05:55,266][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.0784703716635704, 'train/grad_norm_pre_clip': 27.84433364868164, 'timing/data': 0.0013953890011180192, 'timing/model': 2.01815536498907, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 0.83}) +[2026-06-23 08:05:59,352][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.04923464357852936, 'train/grad_norm_pre_clip': 11.156841278076172, 'timing/data': 0.0012192929862067103, 'timing/model': 2.0191735429980326, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.84}) +[2026-06-23 08:06:03,356][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.1072007343173027, 'train/grad_norm_pre_clip': 17.53789710998535, 'timing/data': 0.001266731007490307, 'timing/model': 2.0239140810153913, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 0.84}) +[2026-06-23 08:06:07,425][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.04788976535201073, 'train/grad_norm_pre_clip': 13.981605529785156, 'timing/data': 0.06810105801559985, 'timing/model': 2.0210261229949538, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.84}) +[2026-06-23 08:06:11,434][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.04190421104431152, 'train/grad_norm_pre_clip': 19.328645706176758, 'timing/data': 0.0014172579976730049, 'timing/model': 2.026195793005172, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 0.84}) +[2026-06-23 08:06:15,524][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.13313673436641693, 'train/grad_norm_pre_clip': 18.770854949951172, 'timing/data': 0.0012481669837143272, 'timing/model': 2.0198799399950076, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 0.84}) +[2026-06-23 08:06:19,762][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.08311687409877777, 'train/grad_norm_pre_clip': 13.521990776062012, 'timing/data': 0.0013621370017062873, 'timing/model': 2.258494391979184, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 0.84}) +[2026-06-23 08:06:23,827][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.09530441462993622, 'train/grad_norm_pre_clip': 19.234277725219727, 'timing/data': 0.06493319399305619, 'timing/model': 2.0217219699989073, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 0.84}) +[2026-06-23 08:06:27,852][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.11969728022813797, 'train/grad_norm_pre_clip': 20.82455825805664, 'timing/data': 0.001205888984259218, 'timing/model': 2.048863108997466, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.84}) +[2026-06-23 08:06:31,934][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.009379359893500805, 'train/grad_norm_pre_clip': 2.5164096355438232, 'timing/data': 0.0010378570004832, 'timing/model': 2.0204745199880563, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 0.84}) +[2026-06-23 08:06:35,932][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.12624983489513397, 'train/grad_norm_pre_clip': 21.793716430664062, 'timing/data': 0.00125510900397785, 'timing/model': 2.020348134014057, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 0.84}) +[2026-06-23 08:06:39,995][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.2133060246706009, 'train/grad_norm_pre_clip': 43.148433685302734, 'timing/data': 0.06673790898639709, 'timing/model': 2.0217264689854346, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 0.84}) +[2026-06-23 08:06:43,993][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.1180662140250206, 'train/grad_norm_pre_clip': 20.03858184814453, 'timing/data': 0.0012194029986858368, 'timing/model': 2.02138461501454, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 0.84}) +[2026-06-23 08:06:48,081][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.06425168365240097, 'train/grad_norm_pre_clip': 19.964996337890625, 'timing/data': 0.0012107869843021035, 'timing/model': 2.0199373890063725, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.84}) +[2026-06-23 08:06:52,323][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.06011878326535225, 'train/grad_norm_pre_clip': 19.79161262512207, 'timing/data': 0.0012224689999129623, 'timing/model': 2.2649727840034757, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 0.84}) +[2026-06-23 08:06:56,384][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.10444194823503494, 'train/grad_norm_pre_clip': 30.593990325927734, 'timing/data': 0.05384326199418865, 'timing/model': 2.0269894849916454, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 0.84}) +[2026-06-23 08:07:00,406][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.04467920586466789, 'train/grad_norm_pre_clip': 11.057202339172363, 'timing/data': 0.0012498990108724684, 'timing/model': 2.0410769139998592, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 0.84}) +[2026-06-23 08:07:04,512][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.1206260547041893, 'train/grad_norm_pre_clip': 24.37392807006836, 'timing/data': 0.001232828013598919, 'timing/model': 2.031255895009963, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 0.84}) +[2026-06-23 08:07:08,523][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.18415537476539612, 'train/grad_norm_pre_clip': 23.414562225341797, 'timing/data': 0.0013883960200473666, 'timing/model': 2.034245516988449, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 0.84}) +[2026-06-23 08:07:12,603][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.054237138479948044, 'train/grad_norm_pre_clip': 9.643856048583984, 'timing/data': 0.07096977700712159, 'timing/model': 2.022861198987812, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 0.84}) +[2026-06-23 08:07:16,621][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.09721912443637848, 'train/grad_norm_pre_clip': 15.061885833740234, 'timing/data': 0.001206719985930249, 'timing/model': 2.0402141920058057, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 0.84}) +[2026-06-23 08:07:20,711][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.17638354003429413, 'train/grad_norm_pre_clip': 30.505765914916992, 'timing/data': 0.0012176400050520897, 'timing/model': 2.019927554996684, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 0.84}) +[2026-06-23 08:07:24,783][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.1110575944185257, 'train/grad_norm_pre_clip': 23.695207595825195, 'timing/data': 0.001294382003834471, 'timing/model': 2.0961294079897925, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 0.84}) +[2026-06-23 08:07:28,855][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.10493044555187225, 'train/grad_norm_pre_clip': 27.780529022216797, 'timing/data': 0.07187912799417973, 'timing/model': 2.029613879014505, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.84}) +[2026-06-23 08:07:32,868][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.06593865156173706, 'train/grad_norm_pre_clip': 12.585620880126953, 'timing/data': 0.0013559249928221107, 'timing/model': 2.030689113016706, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 0.84}) +[2026-06-23 08:07:36,964][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.14213614165782928, 'train/grad_norm_pre_clip': 16.262378692626953, 'timing/data': 0.001403773989295587, 'timing/model': 2.0196240649966057, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 0.84}) +[2026-06-23 08:07:40,959][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.052852727472782135, 'train/grad_norm_pre_clip': 9.148880958557129, 'timing/data': 0.001155976002337411, 'timing/model': 2.020363901014207, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 0.84}) +[2026-06-23 08:07:45,025][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.08700808137655258, 'train/grad_norm_pre_clip': 20.494192123413086, 'timing/data': 0.06127771799219772, 'timing/model': 2.031430286995601, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 0.85}) +[2026-06-23 08:07:49,017][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.04986118525266647, 'train/grad_norm_pre_clip': 15.056387901306152, 'timing/data': 0.0011345369857735932, 'timing/model': 2.0134271590213757, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 0.85}) +[2026-06-23 08:07:53,108][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.08603977411985397, 'train/grad_norm_pre_clip': 18.973583221435547, 'timing/data': 0.001222780003445223, 'timing/model': 2.022720670007402, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 0.85}) +[2026-06-23 08:07:57,110][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.06904391944408417, 'train/grad_norm_pre_clip': 23.572959899902344, 'timing/data': 0.0013030169939156622, 'timing/model': 2.02218313497724, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 0.85}) +[2026-06-23 08:08:01,424][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.09649791568517685, 'train/grad_norm_pre_clip': 18.117595672607422, 'timing/data': 0.05791396301356144, 'timing/model': 2.275660084997071, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 0.85}) +[2026-06-23 08:08:05,423][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.08774825185537338, 'train/grad_norm_pre_clip': 100.10604858398438, 'timing/data': 0.001223722007125616, 'timing/model': 2.0189525420137215, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 0.85}) +[2026-06-23 08:08:09,506][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.06242190673947334, 'train/grad_norm_pre_clip': 15.868307113647461, 'timing/data': 0.0012112979893572628, 'timing/model': 2.0189272330026142, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.85}) +[2026-06-23 08:08:13,535][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.12978507578372955, 'train/grad_norm_pre_clip': 37.40424728393555, 'timing/data': 0.0012369549949653447, 'timing/model': 2.0514144510088954, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 0.85}) +[2026-06-23 08:08:17,610][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.10377383977174759, 'train/grad_norm_pre_clip': 20.65662384033203, 'timing/data': 0.06462801800807938, 'timing/model': 2.0209781489975285, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 0.85}) +[2026-06-23 08:08:21,631][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.08454936742782593, 'train/grad_norm_pre_clip': 16.16753387451172, 'timing/data': 0.001634811022086069, 'timing/model': 2.0363016109913588, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 0.85}) +[2026-06-23 08:08:25,719][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.08146478980779648, 'train/grad_norm_pre_clip': 24.28718376159668, 'timing/data': 0.0011894979979842901, 'timing/model': 2.0196578160102945, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 0.85}) +[2026-06-23 08:08:29,727][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.08715779334306717, 'train/grad_norm_pre_clip': 16.637344360351562, 'timing/data': 0.0012528749939519912, 'timing/model': 2.029429429996526, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 0.85}) +[2026-06-23 08:08:33,855][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.09957411885261536, 'train/grad_norm_pre_clip': 21.602571487426758, 'timing/data': 0.07068899701698683, 'timing/model': 2.075901729025645, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 0.85}) +[2026-06-23 08:08:37,856][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.0677391067147255, 'train/grad_norm_pre_clip': 20.539653778076172, 'timing/data': 0.0012364750145934522, 'timing/model': 2.0230326669989154, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 0.85}) +[2026-06-23 08:08:41,944][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.16447971761226654, 'train/grad_norm_pre_clip': 25.259794235229492, 'timing/data': 0.0011963710130658, 'timing/model': 2.0227416880079545, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 0.85}) +[2026-06-23 08:08:45,942][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.10338763892650604, 'train/grad_norm_pre_clip': 32.877525329589844, 'timing/data': 0.0012295719934627414, 'timing/model': 2.0221571649890393, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 0.85}) +[2026-06-23 08:08:50,016][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2200 +[2026-06-23 08:10:44,387][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2200 +[2026-06-23 08:10:44,388][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2200 +[2026-06-23 08:12:51,755][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2200 +[2026-06-23 08:12:51,757][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.15660029649734497, 'train/grad_norm_pre_clip': 32.86951446533203, 'eval/action_loss/seconds': 114.36879487999249, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04667578637599945, 'eval/latency_2/loss': 0.04667578637599945, 'eval/flappy/loss': 0.04667578637599945, 'eval/flappy/latency_2/loss': 0.04667578637599945, 'eval/flap_precision': 0.9571719918422842, 'eval/flap_recall': 0.9500674763832658, 'eval/flap_f1': 0.9536065018625127, 'eval/flap_rate_model': 0.20430555555555555, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9536065018625127, 'eval/latency_2/flap_precision': 0.9571719918422842, 'eval/latency_2/flap_recall': 0.9500674763832658, 'eval/latency_2/flap_f1': 0.9536065018625127, 'eval/latency_2/flap_rate_model': 0.20430555555555555, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9536065018625127, 'eval/action_classification/data_wait_seconds': 50.64501953125, 'eval/action_classification/predict_seconds': 74.9008560180664, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06590012900414877, 'timing/model': 2.0250310940027703, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.85}) +[2026-06-23 08:12:55,704][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.0448601208627224, 'train/grad_norm_pre_clip': 9.965888023376465, 'timing/data': 0.0015943059988785535, 'timing/model': 1.9930902259948198, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 0.85}) +[2026-06-23 08:12:59,778][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.11217457056045532, 'train/grad_norm_pre_clip': 23.328014373779297, 'timing/data': 0.001224842999363318, 'timing/model': 1.9943680159922224, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 0.85}) +[2026-06-23 08:13:03,739][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.05176942050457001, 'train/grad_norm_pre_clip': 8.037074089050293, 'timing/data': 0.0012097959988750517, 'timing/model': 2.00516286099446, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 0.85}) +[2026-06-23 08:13:07,757][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.01782972738146782, 'train/grad_norm_pre_clip': 5.749945163726807, 'timing/data': 0.06886582600418478, 'timing/model': 1.9952737470157444, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 0.85}) +[2026-06-23 08:13:11,729][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.10995210707187653, 'train/grad_norm_pre_clip': 25.788820266723633, 'timing/data': 0.0012017209955956787, 'timing/model': 2.0142820490000304, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 0.85}) +[2026-06-23 08:13:15,775][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.15342828631401062, 'train/grad_norm_pre_clip': 26.85435676574707, 'timing/data': 0.001127342984545976, 'timing/model': 2.000560663989745, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 0.85}) +[2026-06-23 08:13:19,736][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.05568297207355499, 'train/grad_norm_pre_clip': 21.47559356689453, 'timing/data': 0.001156978018116206, 'timing/model': 2.0072210609796457, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.85}) +[2026-06-23 08:13:23,759][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.09360139816999435, 'train/grad_norm_pre_clip': 18.51504135131836, 'timing/data': 0.0550797329924535, 'timing/model': 2.0051982610020787, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 0.85}) +[2026-06-23 08:13:27,724][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.05054007098078728, 'train/grad_norm_pre_clip': 17.961605072021484, 'timing/data': 0.0015326710126828402, 'timing/model': 2.0030236819875427, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 0.85}) +[2026-06-23 08:13:31,849][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.06877636164426804, 'train/grad_norm_pre_clip': 40.330997467041016, 'timing/data': 0.0011340550263412297, 'timing/model': 2.0807655280223116, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.86}) +[2026-06-23 08:13:35,818][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.02787036821246147, 'train/grad_norm_pre_clip': 11.102863311767578, 'timing/data': 0.001141920016380027, 'timing/model': 2.0069619370042346, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 0.86}) +[2026-06-23 08:13:39,862][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.11392845213413239, 'train/grad_norm_pre_clip': 25.677425384521484, 'timing/data': 0.06854974100133404, 'timing/model': 2.007777222985169, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 0.86}) +[2026-06-23 08:13:43,835][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.01860925741493702, 'train/grad_norm_pre_clip': 4.652127265930176, 'timing/data': 0.001224832987645641, 'timing/model': 2.007837544020731, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 0.86}) +[2026-06-23 08:13:47,899][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.07364129275083542, 'train/grad_norm_pre_clip': 14.169764518737793, 'timing/data': 0.001212851027958095, 'timing/model': 2.010818160022609, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 0.86}) +[2026-06-23 08:13:51,885][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.03171578049659729, 'train/grad_norm_pre_clip': 9.948258399963379, 'timing/data': 0.0012819890107493848, 'timing/model': 2.0189309300039895, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 0.86}) +[2026-06-23 08:13:55,928][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.07842247933149338, 'train/grad_norm_pre_clip': 21.43320655822754, 'timing/data': 0.05730135799967684, 'timing/model': 2.013995287998114, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 0.86}) +[2026-06-23 08:13:59,927][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.0548197440803051, 'train/grad_norm_pre_clip': 19.84381103515625, 'timing/data': 0.0012605189986061305, 'timing/model': 2.027775568974903, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 0.86}) +[2026-06-23 08:14:04,014][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.12930364906787872, 'train/grad_norm_pre_clip': 23.61526107788086, 'timing/data': 0.0013720649876631796, 'timing/model': 2.0278795220074244, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 0.86}) +[2026-06-23 08:14:08,286][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.18424102663993835, 'train/grad_norm_pre_clip': 43.46011734008789, 'timing/data': 0.0014247299986891448, 'timing/model': 2.028120230010245, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 0.86}) +[2026-06-23 08:14:12,335][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.09118442237377167, 'train/grad_norm_pre_clip': 16.19883918762207, 'timing/data': 0.055072693998226896, 'timing/model': 2.017176377004944, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.86}) +[2026-06-23 08:14:16,339][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.04991493374109268, 'train/grad_norm_pre_clip': 17.332788467407227, 'timing/data': 0.0011975009983871132, 'timing/model': 2.0280928119900636, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 0.86}) +[2026-06-23 08:14:20,440][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.11448407918214798, 'train/grad_norm_pre_clip': 46.55956268310547, 'timing/data': 0.0012354310019873083, 'timing/model': 2.0296696849982254, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 0.86}) +[2026-06-23 08:14:24,436][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.0484081469476223, 'train/grad_norm_pre_clip': 20.630813598632812, 'timing/data': 0.0015073530084919184, 'timing/model': 2.0170310309913475, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 0.86}) +[2026-06-23 08:14:28,498][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.09240853041410446, 'train/grad_norm_pre_clip': 22.985536575317383, 'timing/data': 0.0698918150155805, 'timing/model': 2.0175099069892894, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 0.86}) +[2026-06-23 08:14:32,496][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.013034332543611526, 'train/grad_norm_pre_clip': 4.417318344116211, 'timing/data': 0.001214973017340526, 'timing/model': 2.0223335219779983, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 0.86}) +[2026-06-23 08:14:36,585][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.24839134514331818, 'train/grad_norm_pre_clip': 64.29571533203125, 'timing/data': 0.0013888439862057567, 'timing/model': 2.024320109980181, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 0.86}) +[2026-06-23 08:14:40,644][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.025869088247418404, 'train/grad_norm_pre_clip': 9.133323669433594, 'timing/data': 0.0013096279872115701, 'timing/model': 2.0737513529893477, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 0.86}) +[2026-06-23 08:14:44,724][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.023331928998231888, 'train/grad_norm_pre_clip': 20.6376953125, 'timing/data': 0.06915502800256945, 'timing/model': 2.0263056659896392, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 0.86}) +[2026-06-23 08:14:48,725][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.08952601253986359, 'train/grad_norm_pre_clip': 32.19498825073242, 'timing/data': 0.0012105149799026549, 'timing/model': 2.02272057699156, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 0.86}) +[2026-06-23 08:14:52,839][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.11705823242664337, 'train/grad_norm_pre_clip': 35.73674774169922, 'timing/data': 0.0012174680014140904, 'timing/model': 2.035389868018683, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.86}) +[2026-06-23 08:14:56,874][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.035515736788511276, 'train/grad_norm_pre_clip': 11.943799018859863, 'timing/data': 0.0012737119977828115, 'timing/model': 2.0375388950051274, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 0.86}) +[2026-06-23 08:15:00,935][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.18821465969085693, 'train/grad_norm_pre_clip': 22.527738571166992, 'timing/data': 0.05540446299710311, 'timing/model': 2.025871123012621, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 0.86}) +[2026-06-23 08:15:04,977][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.1389773041009903, 'train/grad_norm_pre_clip': 33.935489654541016, 'timing/data': 0.0016924450173974037, 'timing/model': 2.040856398991309, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 0.86}) +[2026-06-23 08:15:09,078][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.1312684416770935, 'train/grad_norm_pre_clip': 29.233905792236328, 'timing/data': 0.001189897011499852, 'timing/model': 2.034314930002438, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 0.86}) +[2026-06-23 08:15:13,088][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.08976400643587112, 'train/grad_norm_pre_clip': 19.82883644104004, 'timing/data': 0.0011541309941094369, 'timing/model': 2.0327094449894503, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.86}) +[2026-06-23 08:15:17,256][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.2394135296344757, 'train/grad_norm_pre_clip': 45.31905746459961, 'timing/data': 0.053454706008778885, 'timing/model': 2.0173248189967126, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 0.87}) +[2026-06-23 08:15:21,251][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.22423632442951202, 'train/grad_norm_pre_clip': 35.730506896972656, 'timing/data': 0.0011302659986540675, 'timing/model': 2.019384771003388, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 0.87}) +[2026-06-23 08:15:25,356][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.08291059732437134, 'train/grad_norm_pre_clip': 22.724302291870117, 'timing/data': 0.0014398480125237256, 'timing/model': 2.0368357700062916, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 0.87}) +[2026-06-23 08:15:29,366][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.0802583396434784, 'train/grad_norm_pre_clip': 30.80780792236328, 'timing/data': 0.00115854901378043, 'timing/model': 2.035911622981075, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 0.87}) +[2026-06-23 08:15:33,424][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.09234669804573059, 'train/grad_norm_pre_clip': 28.41880989074707, 'timing/data': 0.053857080987654626, 'timing/model': 2.024128755001584, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.87}) +[2026-06-23 08:15:37,431][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.21072860062122345, 'train/grad_norm_pre_clip': 43.99188995361328, 'timing/data': 0.0011976109817624092, 'timing/model': 2.025536581990309, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 0.87}) +[2026-06-23 08:15:41,516][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.17789269983768463, 'train/grad_norm_pre_clip': 30.043842315673828, 'timing/data': 0.0011352049768902361, 'timing/model': 2.020394345978275, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 0.87}) +[2026-06-23 08:15:45,512][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.08203007280826569, 'train/grad_norm_pre_clip': 20.78375816345215, 'timing/data': 0.0011919410026166588, 'timing/model': 2.0205559610039927, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 0.87}) +[2026-06-23 08:15:49,705][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.09020517021417618, 'train/grad_norm_pre_clip': 15.277230262756348, 'timing/data': 0.051559619983891025, 'timing/model': 2.0211194200091995, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 0.87}) +[2026-06-23 08:15:53,709][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.21781077980995178, 'train/grad_norm_pre_clip': 13.905271530151367, 'timing/data': 0.0014149829803500324, 'timing/model': 2.0230593489832245, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 0.87}) +[2026-06-23 08:15:57,819][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.06791148334741592, 'train/grad_norm_pre_clip': 13.57956600189209, 'timing/data': 0.0014398889907170087, 'timing/model': 2.0395600750052836, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 0.87}) +[2026-06-23 08:16:01,816][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.08764272928237915, 'train/grad_norm_pre_clip': 11.716755867004395, 'timing/data': 0.0012721880048047751, 'timing/model': 2.019591941003455, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 0.87}) +[2026-06-23 08:16:05,867][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.10444755852222443, 'train/grad_norm_pre_clip': 19.25460433959961, 'timing/data': 0.05439560901140794, 'timing/model': 2.020040767005412, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 0.87}) +[2026-06-23 08:16:09,877][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.04325907304883003, 'train/grad_norm_pre_clip': 10.52988052368164, 'timing/data': 0.0012054149992763996, 'timing/model': 2.0310739260166883, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 0.87}) +[2026-06-23 08:28:30,356][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2250 +[2026-06-23 08:28:30,358][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.11318226158618927, 'train/grad_norm_pre_clip': 16.88060188293457, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 190.06000000001035, 'rl_games_eval/mid_train/mean_length': 1549.0, 'rl_games_eval/mid_train/std_reward': 208.48305063003372, 'rl_games_eval/mid_train/std_length': 1679.9416656539, 'rl_games_eval/mid_train/macro_mean_reward': 190.06000000001035, 'rl_games_eval/mid_train/macro_mean_length': 1549.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 190.06000000001035, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1549.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 208.48305063003372, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1679.9416656539, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 190.06000000001035, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0012031619844492525, 'timing/model': 2.0221072970016394, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.87}) +[2026-06-23 08:28:34,235][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.045179206877946854, 'train/grad_norm_pre_clip': 7.670698165893555, 'timing/data': 0.001446081994799897, 'timing/model': 1.9614528350066394, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 0.87}) +[2026-06-23 08:28:38,445][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.1516886204481125, 'train/grad_norm_pre_clip': 26.176055908203125, 'timing/data': 0.0675028350087814, 'timing/model': 2.2264365929877385, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 0.87}) +[2026-06-23 08:28:42,319][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.08479086309671402, 'train/grad_norm_pre_clip': 19.92096710205078, 'timing/data': 0.0012048560020048171, 'timing/model': 1.958928519015899, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 0.87}) +[2026-06-23 08:28:46,293][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.0533229298889637, 'train/grad_norm_pre_clip': 10.410591125488281, 'timing/data': 0.001105340983485803, 'timing/model': 1.9644788049918134, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 0.87}) +[2026-06-23 08:28:50,205][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.07798101752996445, 'train/grad_norm_pre_clip': 11.08358097076416, 'timing/data': 0.0013888959947507828, 'timing/model': 1.9874352249898948, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 0.87}) +[2026-06-23 08:28:54,157][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.08101034164428711, 'train/grad_norm_pre_clip': 17.40143585205078, 'timing/data': 0.05324518901761621, 'timing/model': 1.9701009579875972, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.87}) +[2026-06-23 08:28:58,060][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.11176013946533203, 'train/grad_norm_pre_clip': 33.38454818725586, 'timing/data': 0.0015470680082216859, 'timing/model': 1.9729273030243348, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 0.87}) +[2026-06-23 08:29:02,060][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.07381116598844528, 'train/grad_norm_pre_clip': 14.104283332824707, 'timing/data': 0.0013097199844196439, 'timing/model': 1.9756746609928086, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 0.87}) +[2026-06-23 08:29:05,989][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.04920482635498047, 'train/grad_norm_pre_clip': 9.560006141662598, 'timing/data': 0.0015445129829458892, 'timing/model': 1.9883386089932173, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 0.87}) +[2026-06-23 08:29:09,989][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.07104307413101196, 'train/grad_norm_pre_clip': 16.647113800048828, 'timing/data': 0.06873461199575104, 'timing/model': 1.9848351200053003, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.87}) +[2026-06-23 08:29:14,173][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.017089523375034332, 'train/grad_norm_pre_clip': 7.331851959228516, 'timing/data': 0.0012772490154020488, 'timing/model': 1.998120986012509, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 0.88}) +[2026-06-23 08:29:18,198][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.03490467742085457, 'train/grad_norm_pre_clip': 11.75779914855957, 'timing/data': 0.0011789480049628764, 'timing/model': 1.9922361400094815, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 0.88}) +[2026-06-23 08:29:22,134][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.13583990931510925, 'train/grad_norm_pre_clip': 27.546993255615234, 'timing/data': 0.0012558599992189556, 'timing/model': 1.9881342660228256, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 0.88}) +[2026-06-23 08:29:26,123][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.029240557923913002, 'train/grad_norm_pre_clip': 6.850588798522949, 'timing/data': 0.05014009299338795, 'timing/model': 1.9921779790020082, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 0.88}) +[2026-06-23 08:29:30,060][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.05603598803281784, 'train/grad_norm_pre_clip': 14.586627960205078, 'timing/data': 0.0013288950140122324, 'timing/model': 1.9907932740170509, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 0.88}) +[2026-06-23 08:29:34,109][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.12459751963615417, 'train/grad_norm_pre_clip': 28.95496940612793, 'timing/data': 0.001121081004384905, 'timing/model': 2.013240875006886, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 0.88}) +[2026-06-23 08:29:38,085][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.0787586197257042, 'train/grad_norm_pre_clip': 25.070783615112305, 'timing/data': 0.0015170219994615763, 'timing/model': 2.022199216997251, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 0.88}) +[2026-06-23 08:29:42,121][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.10725724697113037, 'train/grad_norm_pre_clip': 21.19978141784668, 'timing/data': 0.06899326099664904, 'timing/model': 2.0079642449854873, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 0.88}) +[2026-06-23 08:29:46,193][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.05968954414129257, 'train/grad_norm_pre_clip': 9.210789680480957, 'timing/data': 0.0014654969854746014, 'timing/model': 2.1108528320037294, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 0.88}) +[2026-06-23 08:29:50,254][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.11099217087030411, 'train/grad_norm_pre_clip': 19.81763458251953, 'timing/data': 0.0013731869985349476, 'timing/model': 2.012671307020355, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.88}) +[2026-06-23 08:29:54,220][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.11381034553050995, 'train/grad_norm_pre_clip': 37.381656646728516, 'timing/data': 0.0012754760100506246, 'timing/model': 2.0043097030138597, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 0.88}) +[2026-06-23 08:29:58,249][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.09277031570672989, 'train/grad_norm_pre_clip': 32.94138717651367, 'timing/data': 0.05466622399399057, 'timing/model': 2.0132565240201075, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 0.88}) +[2026-06-23 08:30:02,222][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.2027636021375656, 'train/grad_norm_pre_clip': 32.787105560302734, 'timing/data': 0.00125776301138103, 'timing/model': 2.0092934769927524, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 0.88}) +[2026-06-23 08:30:06,284][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.03036590665578842, 'train/grad_norm_pre_clip': 12.015637397766113, 'timing/data': 0.0012780709948856384, 'timing/model': 2.0094501590065192, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 0.88}) +[2026-06-23 08:30:10,460][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.0996030643582344, 'train/grad_norm_pre_clip': 29.769926071166992, 'timing/data': 0.0013178639928810298, 'timing/model': 2.210295804979978, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 0.88}) +[2026-06-23 08:30:14,523][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.08689390122890472, 'train/grad_norm_pre_clip': 19.7408447265625, 'timing/data': 0.08523369301110506, 'timing/model': 2.002941271988675, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 0.88}) +[2026-06-23 08:30:18,488][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.16607752442359924, 'train/grad_norm_pre_clip': 37.341270446777344, 'timing/data': 0.0011216830171179026, 'timing/model': 2.0039173380064312, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 0.88}) +[2026-06-23 08:30:22,773][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.060153309255838394, 'train/grad_norm_pre_clip': 21.367351531982422, 'timing/data': 0.0011117039830423892, 'timing/model': 2.0019678200187627, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 0.88}) +[2026-06-23 08:30:26,735][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.07049775123596191, 'train/grad_norm_pre_clip': 19.45660972595215, 'timing/data': 0.00131758299539797, 'timing/model': 2.002541893016314, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 0.88}) +[2026-06-23 08:30:30,758][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.07041184604167938, 'train/grad_norm_pre_clip': 19.615219116210938, 'timing/data': 0.05871891201240942, 'timing/model': 2.0035707269853447, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.88}) +[2026-06-23 08:30:34,720][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.019175950437784195, 'train/grad_norm_pre_clip': 13.515969276428223, 'timing/data': 0.0012198539916425943, 'timing/model': 2.003078000008827, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 0.88}) +[2026-06-23 08:30:38,784][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.07216300070285797, 'train/grad_norm_pre_clip': 27.70477867126465, 'timing/data': 0.0012765979918185622, 'timing/model': 2.021075140015455, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 0.88}) +[2026-06-23 08:30:42,756][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.030417688190937042, 'train/grad_norm_pre_clip': 12.261083602905273, 'timing/data': 0.0011685590143315494, 'timing/model': 2.008934705023421, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 0.88}) +[2026-06-23 08:30:46,787][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.07213486731052399, 'train/grad_norm_pre_clip': 24.548561096191406, 'timing/data': 0.05995710002025589, 'timing/model': 2.007316349016037, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 0.88}) +[2026-06-23 08:30:50,767][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.018376993015408516, 'train/grad_norm_pre_clip': 6.902878284454346, 'timing/data': 0.0011377519986126572, 'timing/model': 2.0144295519858133, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 0.88}) +[2026-06-23 08:30:55,069][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.025165457278490067, 'train/grad_norm_pre_clip': 14.545215606689453, 'timing/data': 0.001115349994506687, 'timing/model': 2.244109547988046, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 0.88}) +[2026-06-23 08:30:59,049][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.15293945372104645, 'train/grad_norm_pre_clip': 27.139650344848633, 'timing/data': 0.0012011989892926067, 'timing/model': 2.0114992760063615, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 0.89}) +[2026-06-23 08:31:03,089][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.10398048162460327, 'train/grad_norm_pre_clip': 20.868999481201172, 'timing/data': 0.06063759699463844, 'timing/model': 2.0119854060176294, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 0.89}) +[2026-06-23 08:31:07,084][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.07371620833873749, 'train/grad_norm_pre_clip': 25.238035202026367, 'timing/data': 0.0011659039882943034, 'timing/model': 2.0264195530035067, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 0.89}) +[2026-06-23 08:31:11,149][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.0517532080411911, 'train/grad_norm_pre_clip': 10.124717712402344, 'timing/data': 0.0011280439794063568, 'timing/model': 2.009569688991178, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.89}) +[2026-06-23 08:31:15,136][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.0763699859380722, 'train/grad_norm_pre_clip': 27.398319244384766, 'timing/data': 0.0011522290005814284, 'timing/model': 2.0141774140065536, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 0.89}) +[2026-06-23 08:31:19,168][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.06940487027168274, 'train/grad_norm_pre_clip': 18.48362922668457, 'timing/data': 0.050210093992063776, 'timing/model': 2.0120654760103207, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 0.89}) +[2026-06-23 08:31:23,157][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.04788897559046745, 'train/grad_norm_pre_clip': 10.044039726257324, 'timing/data': 0.0011075559887103736, 'timing/model': 2.0164451580203604, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 0.89}) +[2026-06-23 08:31:27,334][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.09896576404571533, 'train/grad_norm_pre_clip': 28.315509796142578, 'timing/data': 0.001109720004023984, 'timing/model': 2.115603940997971, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 0.89}) +[2026-06-23 08:31:31,323][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.08728892356157303, 'train/grad_norm_pre_clip': 65.18360900878906, 'timing/data': 0.0011635989940259606, 'timing/model': 2.015307441994082, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 0.89}) +[2026-06-23 08:31:35,384][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.0797610804438591, 'train/grad_norm_pre_clip': 19.49936866760254, 'timing/data': 0.06457480098470114, 'timing/model': 2.021629284019582, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 0.89}) +[2026-06-23 08:31:39,379][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.10823410004377365, 'train/grad_norm_pre_clip': 22.94220733642578, 'timing/data': 0.0011255790013819933, 'timing/model': 2.0191125980054494, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 0.89}) +[2026-06-23 08:31:43,457][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.055799663066864014, 'train/grad_norm_pre_clip': 13.875814437866211, 'timing/data': 0.001118646003305912, 'timing/model': 2.019461335003143, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 0.89}) +[2026-06-23 08:31:47,466][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.03815301135182381, 'train/grad_norm_pre_clip': 20.399612426757812, 'timing/data': 0.001186542009236291, 'timing/model': 2.0359392469981685, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 0.89}) +[2026-06-23 08:31:51,514][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2300 +[2026-06-23 08:33:46,071][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2300 +[2026-06-23 08:33:46,072][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2300 +[2026-06-23 08:35:53,491][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2300 +[2026-06-23 08:35:53,492][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.07310789823532104, 'train/grad_norm_pre_clip': 17.099149703979492, 'eval/action_loss/seconds': 114.55383774501388, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.050758637487888336, 'eval/latency_2/loss': 0.050758637487888336, 'eval/flappy/loss': 0.050758637487888336, 'eval/flappy/latency_2/loss': 0.050758637487888336, 'eval/flap_precision': 0.9448504983388705, 'eval/flap_recall': 0.9595141700404858, 'eval/flap_f1': 0.9521258788081687, 'eval/flap_rate_model': 0.20902777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9521258788081686, 'eval/latency_2/flap_precision': 0.9448504983388705, 'eval/latency_2/flap_recall': 0.9595141700404858, 'eval/latency_2/flap_f1': 0.9521258788081687, 'eval/latency_2/flap_rate_model': 0.20902777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9521258788081686, 'eval/action_classification/data_wait_seconds': 50.85429000854492, 'eval/action_classification/predict_seconds': 74.92073822021484, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.05919822902069427, 'timing/model': 2.0165488899801858, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.89}) +[2026-06-23 08:35:57,471][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.09362343698740005, 'train/grad_norm_pre_clip': 18.761180877685547, 'timing/data': 0.0013521280197892338, 'timing/model': 2.013302189006936, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 0.89}) +[2026-06-23 08:36:01,516][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.14107713103294373, 'train/grad_norm_pre_clip': 26.75632095336914, 'timing/data': 0.0012463020102586597, 'timing/model': 1.9976929950062186, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 0.89}) +[2026-06-23 08:36:05,470][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.14724917709827423, 'train/grad_norm_pre_clip': 36.28598403930664, 'timing/data': 0.0012337999942246825, 'timing/model': 1.9994918300071731, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 0.89}) +[2026-06-23 08:36:09,509][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.15549501776695251, 'train/grad_norm_pre_clip': 30.45282745361328, 'timing/data': 0.06806868000421673, 'timing/model': 2.0088655589788686, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 0.89}) +[2026-06-23 08:36:13,487][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.12763817608356476, 'train/grad_norm_pre_clip': 32.016483306884766, 'timing/data': 0.0012597469903994352, 'timing/model': 2.0112686400243547, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.89}) +[2026-06-23 08:36:17,564][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.08008500188589096, 'train/grad_norm_pre_clip': 15.009686470031738, 'timing/data': 0.0012417740072123706, 'timing/model': 2.0066833609889727, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 0.89}) +[2026-06-23 08:36:21,529][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.05810806527733803, 'train/grad_norm_pre_clip': 9.7492094039917, 'timing/data': 0.0015082760073710233, 'timing/model': 2.0055997999734245, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 0.89}) +[2026-06-23 08:36:25,828][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.23195795714855194, 'train/grad_norm_pre_clip': 34.56636047363281, 'timing/data': 0.06097326200688258, 'timing/model': 2.274359501025174, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 0.89}) +[2026-06-23 08:36:29,805][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.05554962530732155, 'train/grad_norm_pre_clip': 17.129018783569336, 'timing/data': 0.0012475140101742, 'timing/model': 2.0082103660097346, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 0.89}) +[2026-06-23 08:36:33,879][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.051614657044410706, 'train/grad_norm_pre_clip': 11.373902320861816, 'timing/data': 0.0014415830082725734, 'timing/model': 2.0127910200098995, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.89}) +[2026-06-23 08:36:37,865][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.10083215683698654, 'train/grad_norm_pre_clip': 27.30967140197754, 'timing/data': 0.0016635229985695332, 'timing/model': 2.0158066930016503, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 0.89}) +[2026-06-23 08:36:41,902][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.26673245429992676, 'train/grad_norm_pre_clip': 26.673139572143555, 'timing/data': 0.05604409598163329, 'timing/model': 2.0116373309865594, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 0.89}) +[2026-06-23 08:36:45,901][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.07394502311944962, 'train/grad_norm_pre_clip': 20.306781768798828, 'timing/data': 0.0012877390254288912, 'timing/model': 2.0230564690136816, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 0.9}) +[2026-06-23 08:36:49,986][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.07342399656772614, 'train/grad_norm_pre_clip': 16.160194396972656, 'timing/data': 0.0013652820198331028, 'timing/model': 2.027057461993536, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 0.9}) +[2026-06-23 08:36:54,006][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.08125222474336624, 'train/grad_norm_pre_clip': 15.632491111755371, 'timing/data': 0.0013049909903202206, 'timing/model': 2.0268311749969143, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 0.9}) +[2026-06-23 08:36:58,083][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.06567476689815521, 'train/grad_norm_pre_clip': 11.792169570922852, 'timing/data': 0.056421043002046645, 'timing/model': 2.0499438049737364, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 0.9}) +[2026-06-23 08:37:02,095][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.047807417809963226, 'train/grad_norm_pre_clip': 9.069598197937012, 'timing/data': 0.0013768629869446158, 'timing/model': 2.0324135109840427, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 0.9}) +[2026-06-23 08:37:06,181][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.13994164764881134, 'train/grad_norm_pre_clip': 20.40579605102539, 'timing/data': 0.001218862016685307, 'timing/model': 2.0284677629824728, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 0.9}) +[2026-06-23 08:37:10,180][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.05579396337270737, 'train/grad_norm_pre_clip': 14.204127311706543, 'timing/data': 0.0012639650085475296, 'timing/model': 2.020608655991964, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 0.9}) +[2026-06-23 08:37:14,277][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.03154020011425018, 'train/grad_norm_pre_clip': 4.8146820068359375, 'timing/data': 0.06518125699949451, 'timing/model': 2.0445501870126463, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.9}) +[2026-06-23 08:37:18,294][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.11222810298204422, 'train/grad_norm_pre_clip': 23.568801879882812, 'timing/data': 0.001462481974158436, 'timing/model': 2.040569595003035, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 0.9}) +[2026-06-23 08:37:22,386][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.16021601855754852, 'train/grad_norm_pre_clip': 33.456214904785156, 'timing/data': 0.0013939760101493448, 'timing/model': 2.0229524899914395, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 0.9}) +[2026-06-23 08:37:26,397][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.1150054931640625, 'train/grad_norm_pre_clip': 22.259891510009766, 'timing/data': 0.001181833998998627, 'timing/model': 2.0234068599820603, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 0.9}) +[2026-06-23 08:37:30,469][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.045725103467702866, 'train/grad_norm_pre_clip': 12.98031997680664, 'timing/data': 0.07067348400596529, 'timing/model': 2.0204759370244574, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 0.9}) +[2026-06-23 08:37:34,605][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.01621345430612564, 'train/grad_norm_pre_clip': 6.102692604064941, 'timing/data': 0.0011248679948039353, 'timing/model': 2.0267458419839386, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 0.9}) +[2026-06-23 08:37:38,702][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.17085786163806915, 'train/grad_norm_pre_clip': 22.939624786376953, 'timing/data': 0.0013901880010962486, 'timing/model': 2.0237774200213607, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 0.9}) +[2026-06-23 08:37:42,710][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.13261297345161438, 'train/grad_norm_pre_clip': 32.34030532836914, 'timing/data': 0.0012158759927842766, 'timing/model': 2.03581199402106, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 0.9}) +[2026-06-23 08:37:46,767][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.09523952752351761, 'train/grad_norm_pre_clip': 21.25738525390625, 'timing/data': 0.060300651006400585, 'timing/model': 2.0206992749881465, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 0.9}) +[2026-06-23 08:37:50,767][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.10968434810638428, 'train/grad_norm_pre_clip': 37.02882385253906, 'timing/data': 0.0011195680126547813, 'timing/model': 2.0244441049871966, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 0.9}) +[2026-06-23 08:37:54,873][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.20691829919815063, 'train/grad_norm_pre_clip': 44.16313171386719, 'timing/data': 0.001341649011010304, 'timing/model': 2.0353307780169416, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.9}) +[2026-06-23 08:37:58,879][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.11016160249710083, 'train/grad_norm_pre_clip': 25.937740325927734, 'timing/data': 0.0012270869920030236, 'timing/model': 2.029476625000825, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 0.9}) +[2026-06-23 08:38:02,953][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.09964516758918762, 'train/grad_norm_pre_clip': 14.669877052307129, 'timing/data': 0.0636789120035246, 'timing/model': 2.0203266319877002, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 0.9}) +[2026-06-23 08:38:07,212][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.06024348735809326, 'train/grad_norm_pre_clip': 52.62491989135742, 'timing/data': 0.0012008880148641765, 'timing/model': 2.280516686005285, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 0.9}) +[2026-06-23 08:38:11,325][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.2436302900314331, 'train/grad_norm_pre_clip': 53.42744064331055, 'timing/data': 0.0013869119866285473, 'timing/model': 2.022108286008006, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 0.9}) +[2026-06-23 08:38:15,374][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.07828409224748611, 'train/grad_norm_pre_clip': 21.733232498168945, 'timing/data': 0.001386290998198092, 'timing/model': 2.0427244339953177, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 0.9}) +[2026-06-23 08:38:19,432][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.08663252741098404, 'train/grad_norm_pre_clip': 30.777591705322266, 'timing/data': 0.05404896198888309, 'timing/model': 2.028755639010342, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 0.9}) +[2026-06-23 08:38:23,450][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.2004867047071457, 'train/grad_norm_pre_clip': 33.583251953125, 'timing/data': 0.0013918910117354244, 'timing/model': 2.0342608009814285, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 0.9}) +[2026-06-23 08:38:27,528][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.027208548039197922, 'train/grad_norm_pre_clip': 12.015863418579102, 'timing/data': 0.001129747019149363, 'timing/model': 2.0192438189988025, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 0.9}) +[2026-06-23 08:38:31,520][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.09005886316299438, 'train/grad_norm_pre_clip': 18.36456298828125, 'timing/data': 0.0011706420045811683, 'timing/model': 2.018882663018303, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 0.91}) +[2026-06-23 08:38:35,574][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.08227184414863586, 'train/grad_norm_pre_clip': 18.645334243774414, 'timing/data': 0.0622069530072622, 'timing/model': 2.019424746016739, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.91}) +[2026-06-23 08:38:39,609][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.0827980488538742, 'train/grad_norm_pre_clip': 15.259993553161621, 'timing/data': 0.0012850339990109205, 'timing/model': 2.0583199440152384, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 0.91}) +[2026-06-23 08:38:43,695][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.1414385885000229, 'train/grad_norm_pre_clip': 20.918201446533203, 'timing/data': 0.0011086679878644645, 'timing/model': 2.0215024070057552, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 0.91}) +[2026-06-23 08:38:47,696][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.018047001212835312, 'train/grad_norm_pre_clip': 9.129475593566895, 'timing/data': 0.001152688986621797, 'timing/model': 2.0256017220090143, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 0.91}) +[2026-06-23 08:38:51,746][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.08094432950019836, 'train/grad_norm_pre_clip': 26.48236083984375, 'timing/data': 0.05285954399732873, 'timing/model': 2.021515143016586, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 0.91}) +[2026-06-23 08:38:55,742][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.07726451009511948, 'train/grad_norm_pre_clip': 15.147171974182129, 'timing/data': 0.0011083970020990819, 'timing/model': 2.0198367330012843, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 0.91}) +[2026-06-23 08:38:59,840][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.08003070950508118, 'train/grad_norm_pre_clip': 23.924612045288086, 'timing/data': 0.001278641982935369, 'timing/model': 2.036172883992549, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 0.91}) +[2026-06-23 08:39:03,840][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.1637452244758606, 'train/grad_norm_pre_clip': 19.759000778198242, 'timing/data': 0.0013117629860062152, 'timing/model': 2.0234779020247515, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 0.91}) +[2026-06-23 08:39:07,892][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.04725130647420883, 'train/grad_norm_pre_clip': 16.88932228088379, 'timing/data': 0.05531397901359014, 'timing/model': 2.018399922992103, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 0.91}) +[2026-06-23 08:39:11,888][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.08816265314817429, 'train/grad_norm_pre_clip': 21.32183837890625, 'timing/data': 0.0011063640122301877, 'timing/model': 2.017992581008002, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 0.91}) +[2026-06-23 08:39:16,197][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.08913968503475189, 'train/grad_norm_pre_clip': 15.936838150024414, 'timing/data': 0.0012743239931296557, 'timing/model': 2.014006337005412, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.91}) +[2026-06-23 08:39:20,210][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.14936089515686035, 'train/grad_norm_pre_clip': 23.98436737060547, 'timing/data': 0.001316913025220856, 'timing/model': 2.0349319580127485, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 0.91}) +[2026-06-23 08:39:24,255][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.05900963768362999, 'train/grad_norm_pre_clip': 12.104676246643066, 'timing/data': 0.052417327009607106, 'timing/model': 2.017444732016884, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 0.91}) +[2026-06-23 08:39:28,247][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.07697469741106033, 'train/grad_norm_pre_clip': 13.834558486938477, 'timing/data': 0.0011192980164196342, 'timing/model': 2.018341930990573, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 0.91}) +[2026-06-23 08:39:32,326][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.03757982328534126, 'train/grad_norm_pre_clip': 9.406757354736328, 'timing/data': 0.0012545369972940534, 'timing/model': 2.0172940559859853, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.91}) +[2026-06-23 08:39:36,319][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.09270890802145004, 'train/grad_norm_pre_clip': 19.012489318847656, 'timing/data': 0.0013333639944903553, 'timing/model': 2.015562537999358, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 0.91}) +[2026-06-23 08:39:40,366][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.044037703424692154, 'train/grad_norm_pre_clip': 6.313157558441162, 'timing/data': 0.057080962986219674, 'timing/model': 2.0174632310227025, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 0.91}) +[2026-06-23 08:39:44,364][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.022314995527267456, 'train/grad_norm_pre_clip': 5.588910102844238, 'timing/data': 0.001130778982769698, 'timing/model': 2.022419237997383, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 0.91}) +[2026-06-23 08:39:48,655][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.06684558838605881, 'train/grad_norm_pre_clip': 13.376132011413574, 'timing/data': 0.0012580039910972118, 'timing/model': 2.0230927020020317, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 0.91}) +[2026-06-23 08:39:52,655][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.06072314828634262, 'train/grad_norm_pre_clip': 23.244937896728516, 'timing/data': 0.0011628279753495008, 'timing/model': 2.024505373003194, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 0.91}) +[2026-06-23 08:39:56,719][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.022133350372314453, 'train/grad_norm_pre_clip': 7.871860027313232, 'timing/data': 0.06183938498725183, 'timing/model': 2.022850737994304, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.91}) +[2026-06-23 08:40:00,724][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.06489834189414978, 'train/grad_norm_pre_clip': 11.043538093566895, 'timing/data': 0.0011174750106874853, 'timing/model': 2.0250819709908683, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 0.91}) +[2026-06-23 08:40:04,815][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.10369034111499786, 'train/grad_norm_pre_clip': 32.72118377685547, 'timing/data': 0.0013426700024865568, 'timing/model': 2.0224700870167, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 0.91}) +[2026-06-23 08:40:08,828][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.059034623205661774, 'train/grad_norm_pre_clip': 26.572553634643555, 'timing/data': 0.001147629984188825, 'timing/model': 2.020197518984787, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 0.91}) +[2026-06-23 08:40:12,907][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.0983099713921547, 'train/grad_norm_pre_clip': 18.309484481811523, 'timing/data': 0.06394693799666129, 'timing/model': 2.035546015977161, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 0.91}) +[2026-06-23 08:40:16,938][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.14219626784324646, 'train/grad_norm_pre_clip': 23.03557014465332, 'timing/data': 0.001208883011713624, 'timing/model': 2.0346518650185317, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 0.92}) +[2026-06-23 08:40:21,075][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.11687646061182022, 'train/grad_norm_pre_clip': 24.295448303222656, 'timing/data': 0.0012345610011834651, 'timing/model': 2.0578998049895745, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 0.92}) +[2026-06-23 08:40:25,091][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.09847798943519592, 'train/grad_norm_pre_clip': 20.362550735473633, 'timing/data': 0.001348731981124729, 'timing/model': 2.0363057820068207, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 0.92}) +[2026-06-23 08:40:29,173][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.05045554041862488, 'train/grad_norm_pre_clip': 10.678497314453125, 'timing/data': 0.06205778900766745, 'timing/model': 2.031181947008008, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 0.92}) +[2026-06-23 08:40:33,191][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.07405312359333038, 'train/grad_norm_pre_clip': 21.54670524597168, 'timing/data': 0.0012148139940109104, 'timing/model': 2.0314976120134816, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 0.92}) +[2026-06-23 08:40:37,302][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.12939120829105377, 'train/grad_norm_pre_clip': 19.92438316345215, 'timing/data': 0.0014190619986038655, 'timing/model': 2.031020280002849, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.92}) +[2026-06-23 08:40:41,311][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.11875415593385696, 'train/grad_norm_pre_clip': 29.450626373291016, 'timing/data': 0.0013061730132903904, 'timing/model': 2.026805936009623, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 0.92}) +[2026-06-23 08:40:45,372][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.05750953406095505, 'train/grad_norm_pre_clip': 13.599393844604492, 'timing/data': 0.05510543400305323, 'timing/model': 2.026371924992418, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 0.92}) +[2026-06-23 08:40:49,401][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.0342111699283123, 'train/grad_norm_pre_clip': 8.86864185333252, 'timing/data': 0.0012213759764563292, 'timing/model': 2.043990133999614, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 0.92}) +[2026-06-23 08:40:53,523][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.178147092461586, 'train/grad_norm_pre_clip': 33.35043716430664, 'timing/data': 0.0012002669973298907, 'timing/model': 2.0327051269996446, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 0.92}) +[2026-06-23 08:40:57,681][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.18042956292629242, 'train/grad_norm_pre_clip': 25.17879867553711, 'timing/data': 0.0012283189862500876, 'timing/model': 2.0283364670176525, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.92}) +[2026-06-23 08:41:01,762][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.21557465195655823, 'train/grad_norm_pre_clip': 18.214488983154297, 'timing/data': 0.06499969298602082, 'timing/model': 2.0320824579976033, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 0.92}) +[2026-06-23 08:41:05,777][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.016349399462342262, 'train/grad_norm_pre_clip': 6.665718078613281, 'timing/data': 0.0012153549760114402, 'timing/model': 2.0296377220074646, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 0.92}) +[2026-06-23 08:41:09,889][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.07521919906139374, 'train/grad_norm_pre_clip': 12.841264724731445, 'timing/data': 0.0012690940056927502, 'timing/model': 2.0333889439934865, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 0.92}) +[2026-06-23 08:41:13,907][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.10565424710512161, 'train/grad_norm_pre_clip': 21.76962661743164, 'timing/data': 0.0012652880104724318, 'timing/model': 2.029149432986742, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 0.92}) +[2026-06-23 08:41:17,979][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.10810060799121857, 'train/grad_norm_pre_clip': 17.348331451416016, 'timing/data': 0.05238814401673153, 'timing/model': 2.031546923011774, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.92}) +[2026-06-23 08:41:22,009][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.043470270931720734, 'train/grad_norm_pre_clip': 14.128275871276855, 'timing/data': 0.0012584539945237339, 'timing/model': 2.040782655996736, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 0.92}) +[2026-06-23 08:41:26,102][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.18321412801742554, 'train/grad_norm_pre_clip': 37.962223052978516, 'timing/data': 0.0014286689984146506, 'timing/model': 2.0244909889879636, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 0.92}) +[2026-06-23 08:41:30,355][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.041482046246528625, 'train/grad_norm_pre_clip': 12.296793937683105, 'timing/data': 0.0012349509925115854, 'timing/model': 2.26888006698573, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 0.92}) +[2026-06-23 08:41:34,420][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.07100757211446762, 'train/grad_norm_pre_clip': 19.68497085571289, 'timing/data': 0.06405896600335836, 'timing/model': 2.022347700985847, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 0.92}) +[2026-06-23 08:41:38,436][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.0634026750922203, 'train/grad_norm_pre_clip': 17.650651931762695, 'timing/data': 0.0012001969735138118, 'timing/model': 2.0252665450097993, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 0.92}) +[2026-06-23 08:41:42,531][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.0529748760163784, 'train/grad_norm_pre_clip': 12.820466995239258, 'timing/data': 0.0012069600052200258, 'timing/model': 2.0262905670097098, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 0.92}) +[2026-06-23 08:41:46,546][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.04201263189315796, 'train/grad_norm_pre_clip': 15.614771842956543, 'timing/data': 0.0012837120157200843, 'timing/model': 2.0279748199973255, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 0.92}) +[2026-06-23 08:41:50,619][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.16043423116207123, 'train/grad_norm_pre_clip': 41.05810546875, 'timing/data': 0.05943608598317951, 'timing/model': 2.0282290400064085, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 0.92}) +[2026-06-23 08:41:54,625][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.15700896084308624, 'train/grad_norm_pre_clip': 29.089529037475586, 'timing/data': 0.0013215920189395547, 'timing/model': 2.0264598219946492, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 0.92}) +[2026-06-23 08:41:58,738][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.12160039693117142, 'train/grad_norm_pre_clip': 28.37103271484375, 'timing/data': 0.0014162060106173158, 'timing/model': 2.042235999018885, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.92}) +[2026-06-23 08:42:02,999][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.19071029126644135, 'train/grad_norm_pre_clip': 95.9552001953125, 'timing/data': 0.001303117984207347, 'timing/model': 2.275561527989339, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 0.93}) +[2026-06-23 08:42:07,078][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.029848279431462288, 'train/grad_norm_pre_clip': 7.0746169090271, 'timing/data': 0.06722222999087535, 'timing/model': 2.0253367200202774, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 0.93}) +[2026-06-23 08:42:11,088][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.03794535622000694, 'train/grad_norm_pre_clip': 9.268470764160156, 'timing/data': 0.001210497022839263, 'timing/model': 2.027989992988296, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 0.93}) +[2026-06-23 08:42:15,171][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.047524046152830124, 'train/grad_norm_pre_clip': 20.908409118652344, 'timing/data': 0.001135839003836736, 'timing/model': 2.023753216024488, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 0.93}) +[2026-06-23 08:42:19,176][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.05476631969213486, 'train/grad_norm_pre_clip': 15.806281089782715, 'timing/data': 0.0011619069846346974, 'timing/model': 2.025814344990067, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 0.93}) +[2026-06-23 08:42:23,253][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.10736797004938126, 'train/grad_norm_pre_clip': 24.618806838989258, 'timing/data': 0.06915786600438878, 'timing/model': 2.022768939990783, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 0.93}) +[2026-06-23 08:42:27,262][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.08940683305263519, 'train/grad_norm_pre_clip': 13.586891174316406, 'timing/data': 0.001106954994611442, 'timing/model': 2.0270558999909554, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 0.93}) +[2026-06-23 08:42:31,376][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.05460368096828461, 'train/grad_norm_pre_clip': 19.32469367980957, 'timing/data': 0.001465317007387057, 'timing/model': 2.0386306950240396, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 0.93}) +[2026-06-23 08:42:35,407][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.07101864367723465, 'train/grad_norm_pre_clip': 13.324464797973633, 'timing/data': 0.0012234300083946437, 'timing/model': 2.0474040639819577, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 0.93}) +[2026-06-23 08:42:39,475][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2400 +[2026-06-23 08:44:33,737][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2400 +[2026-06-23 08:44:33,738][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2400 +[2026-06-23 08:46:40,624][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2400 +[2026-06-23 08:46:40,625][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.09313158690929413, 'train/grad_norm_pre_clip': 18.1638240814209, 'eval/action_loss/seconds': 114.25924460199894, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06145790219306946, 'eval/latency_2/loss': 0.06145790219306946, 'eval/flappy/loss': 0.06145790219306946, 'eval/flappy/latency_2/loss': 0.06145790219306946, 'eval/flap_precision': 0.9817251461988304, 'eval/flap_recall': 0.9062078272604588, 'eval/flap_f1': 0.9424561403508771, 'eval/flap_rate_model': 0.19, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9424561403508771, 'eval/latency_2/flap_precision': 0.9817251461988304, 'eval/latency_2/flap_recall': 0.9062078272604588, 'eval/latency_2/flap_f1': 0.9424561403508771, 'eval/latency_2/flap_rate_model': 0.19, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9424561403508771, 'eval/action_classification/data_wait_seconds': 50.1475944519043, 'eval/action_classification/predict_seconds': 75.12416076660156, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.05841141301789321, 'timing/model': 2.026466083014384, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.93}) +[2026-06-23 08:46:44,584][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.05943087115883827, 'train/grad_norm_pre_clip': 22.566123962402344, 'timing/data': 0.0015606039960402995, 'timing/model': 2.0086084979993757, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 0.93}) +[2026-06-23 08:46:48,620][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.20281527936458588, 'train/grad_norm_pre_clip': 58.39384078979492, 'timing/data': 0.001201890001539141, 'timing/model': 1.9995509939908516, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 0.93}) +[2026-06-23 08:46:52,572][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.046322282403707504, 'train/grad_norm_pre_clip': 8.473048210144043, 'timing/data': 0.0014176480181049556, 'timing/model': 1.996677446004469, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.93}) +[2026-06-23 08:46:56,567][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.15030942857265472, 'train/grad_norm_pre_clip': 36.20404052734375, 'timing/data': 0.0506983749801293, 'timing/model': 1.9943012170260772, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 0.93}) +[2026-06-23 08:47:00,754][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.08852753788232803, 'train/grad_norm_pre_clip': 23.1724853515625, 'timing/data': 0.0011002120154444128, 'timing/model': 1.9936330119962804, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 0.93}) +[2026-06-23 08:47:04,823][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.1158512756228447, 'train/grad_norm_pre_clip': 19.283525466918945, 'timing/data': 0.001327292004134506, 'timing/model': 2.0086683699919377, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 0.93}) +[2026-06-23 08:47:08,788][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.14374381303787231, 'train/grad_norm_pre_clip': 22.48560905456543, 'timing/data': 0.0011623869941104203, 'timing/model': 1.9971193940145895, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 0.93}) +[2026-06-23 08:47:12,818][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.13882219791412354, 'train/grad_norm_pre_clip': 19.99906349182129, 'timing/data': 0.057106813008431345, 'timing/model': 2.017597696976736, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 0.93}) +[2026-06-23 08:47:16,788][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.15574653446674347, 'train/grad_norm_pre_clip': 20.21661376953125, 'timing/data': 0.0012404320004861802, 'timing/model': 2.006278246000875, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 0.93}) +[2026-06-23 08:47:20,883][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.13390018045902252, 'train/grad_norm_pre_clip': 28.72622299194336, 'timing/data': 0.001326811994658783, 'timing/model': 2.0297407929901965, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.93}) +[2026-06-23 08:47:24,856][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.033399730920791626, 'train/grad_norm_pre_clip': 11.496768951416016, 'timing/data': 0.0012861360155511647, 'timing/model': 2.008161134988768, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 0.93}) +[2026-06-23 08:47:28,897][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.050585389137268066, 'train/grad_norm_pre_clip': 10.334049224853516, 'timing/data': 0.056882448989199474, 'timing/model': 2.0206477440078743, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 0.93}) +[2026-06-23 08:47:33,040][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.082168348133564, 'train/grad_norm_pre_clip': 32.31612777709961, 'timing/data': 0.0013993460161145777, 'timing/model': 2.1691829710034654, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 0.93}) +[2026-06-23 08:47:37,126][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.05125505477190018, 'train/grad_norm_pre_clip': 10.47850513458252, 'timing/data': 0.0014004979748278856, 'timing/model': 2.024704390001716, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 0.93}) +[2026-06-23 08:47:41,102][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.06946546584367752, 'train/grad_norm_pre_clip': 21.634597778320312, 'timing/data': 0.0013661239936482161, 'timing/model': 2.0066477509972174, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 0.93}) +[2026-06-23 08:47:45,158][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.11566530913114548, 'train/grad_norm_pre_clip': 19.49346351623535, 'timing/data': 0.05765273200813681, 'timing/model': 2.0297969990060665, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 0.94}) +[2026-06-23 08:47:49,144][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.07642007619142532, 'train/grad_norm_pre_clip': 13.652180671691895, 'timing/data': 0.0012203540245536715, 'timing/model': 2.0143641569884494, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 0.94}) +[2026-06-23 08:47:53,235][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.06059660762548447, 'train/grad_norm_pre_clip': 16.8934268951416, 'timing/data': 0.0013152800092939287, 'timing/model': 2.032591212016996, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 0.94}) +[2026-06-23 08:47:57,219][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.154195636510849, 'train/grad_norm_pre_clip': 33.334991455078125, 'timing/data': 0.0014246120117604733, 'timing/model': 2.011674168985337, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 0.94}) +[2026-06-23 08:48:01,277][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.2090364545583725, 'train/grad_norm_pre_clip': 51.65266418457031, 'timing/data': 0.059751730994321406, 'timing/model': 2.026665726007195, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.94}) +[2026-06-23 08:48:05,270][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.10143391042947769, 'train/grad_norm_pre_clip': 26.414335250854492, 'timing/data': 0.001280355005292222, 'timing/model': 2.020349427999463, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 0.94}) +[2026-06-23 08:48:09,624][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.10788454860448837, 'train/grad_norm_pre_clip': 36.86644744873047, 'timing/data': 0.0013322010054253042, 'timing/model': 2.03606569301337, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 0.94}) +[2026-06-23 08:48:13,624][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.08248576521873474, 'train/grad_norm_pre_clip': 14.909200668334961, 'timing/data': 0.0011771449935622513, 'timing/model': 2.0271582269924693, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 0.94}) +[2026-06-23 08:48:17,686][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.07479184865951538, 'train/grad_norm_pre_clip': 14.442582130432129, 'timing/data': 0.05784123099874705, 'timing/model': 2.0199240039801225, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 0.94}) +[2026-06-23 08:48:21,680][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.10933104157447815, 'train/grad_norm_pre_clip': 18.38387680053711, 'timing/data': 0.0011962820135522634, 'timing/model': 2.015367863990832, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 0.94}) +[2026-06-23 08:48:25,781][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.2241404801607132, 'train/grad_norm_pre_clip': 36.14399719238281, 'timing/data': 0.0012519950105343014, 'timing/model': 2.0240388070233166, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 0.94}) +[2026-06-23 08:48:29,785][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.0979142114520073, 'train/grad_norm_pre_clip': 25.22960662841797, 'timing/data': 0.0012615319865290076, 'timing/model': 2.0236722039990127, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 0.94}) +[2026-06-23 08:48:33,881][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.12226390093564987, 'train/grad_norm_pre_clip': 25.315366744995117, 'timing/data': 0.08990434298175387, 'timing/model': 2.0264876809960697, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 0.94}) +[2026-06-23 08:48:37,884][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.10483301430940628, 'train/grad_norm_pre_clip': 14.931671142578125, 'timing/data': 0.0014255649875849485, 'timing/model': 2.027630038006464, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 0.94}) +[2026-06-23 08:48:42,210][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.14639288187026978, 'train/grad_norm_pre_clip': 35.321022033691406, 'timing/data': 0.0014165090105962008, 'timing/model': 2.017286908987444, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 0.94}) +[2026-06-23 08:48:46,208][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.07509810477495193, 'train/grad_norm_pre_clip': 13.146139144897461, 'timing/data': 0.0015284650144167244, 'timing/model': 2.0214417949900962, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 0.94}) +[2026-06-23 08:48:50,264][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.08323368430137634, 'train/grad_norm_pre_clip': 17.70745849609375, 'timing/data': 0.06015354098053649, 'timing/model': 2.019934572977945, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 0.94}) +[2026-06-23 08:48:54,257][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.013959906063973904, 'train/grad_norm_pre_clip': 4.289026260375977, 'timing/data': 0.0013496649917215109, 'timing/model': 2.0179231689835433, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 0.94}) +[2026-06-23 08:48:58,346][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.08298595249652863, 'train/grad_norm_pre_clip': 17.107011795043945, 'timing/data': 0.0013572589959949255, 'timing/model': 2.023088601999916, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 0.94}) +[2026-06-23 08:49:02,340][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.10086015611886978, 'train/grad_norm_pre_clip': 14.301429748535156, 'timing/data': 0.0011585419997572899, 'timing/model': 2.0170354779984336, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 0.94}) +[2026-06-23 08:49:06,389][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.08513496816158295, 'train/grad_norm_pre_clip': 15.661316871643066, 'timing/data': 0.05752022098749876, 'timing/model': 2.0160075520107057, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 0.94}) +[2026-06-23 08:49:10,387][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.057821013033390045, 'train/grad_norm_pre_clip': 11.647872924804688, 'timing/data': 0.0011237880098633468, 'timing/model': 2.0193158620095346, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 0.94}) +[2026-06-23 08:49:14,661][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.05565108358860016, 'train/grad_norm_pre_clip': 13.78621768951416, 'timing/data': 0.0011260610190220177, 'timing/model': 2.207325185998343, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 0.94}) +[2026-06-23 08:49:18,660][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.11908844113349915, 'train/grad_norm_pre_clip': 21.678401947021484, 'timing/data': 0.0012050079822074622, 'timing/model': 2.0177752540039364, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 0.94}) +[2026-06-23 08:49:22,729][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.11922153830528259, 'train/grad_norm_pre_clip': 14.668176651000977, 'timing/data': 0.07216169501771219, 'timing/model': 2.02278780497727, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 0.94}) +[2026-06-23 08:49:26,740][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.08865252137184143, 'train/grad_norm_pre_clip': 54.85954284667969, 'timing/data': 0.0014485279971268028, 'timing/model': 2.0324713209993206, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 0.94}) +[2026-06-23 08:49:30,828][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.05471709743142128, 'train/grad_norm_pre_clip': 10.374468803405762, 'timing/data': 0.001126041985116899, 'timing/model': 2.021077282988699, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 0.95}) +[2026-06-23 08:49:34,847][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.04212179407477379, 'train/grad_norm_pre_clip': 14.528203010559082, 'timing/data': 0.0011601639853324741, 'timing/model': 2.042145164014073, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 0.95}) +[2026-06-23 08:49:38,902][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.046553004533052444, 'train/grad_norm_pre_clip': 15.63683032989502, 'timing/data': 0.05568815101287328, 'timing/model': 2.021590844000457, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 0.95}) +[2026-06-23 08:49:42,909][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.0783485397696495, 'train/grad_norm_pre_clip': 15.835713386535645, 'timing/data': 0.0011297090095467865, 'timing/model': 2.0258299760171212, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.95}) +[2026-06-23 08:49:47,011][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.0551617406308651, 'train/grad_norm_pre_clip': 11.398663520812988, 'timing/data': 0.0011189380020368844, 'timing/model': 2.024321984004928, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 0.95}) +[2026-06-23 08:49:51,064][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.12365638464689255, 'train/grad_norm_pre_clip': 18.982885360717773, 'timing/data': 0.001174732024082914, 'timing/model': 2.041365635988768, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 0.95}) +[2026-06-23 08:49:55,141][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.0847807303071022, 'train/grad_norm_pre_clip': 18.05744743347168, 'timing/data': 0.06632297800388187, 'timing/model': 2.029460924997693, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 0.95}) +[2026-06-23 08:49:59,159][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.07721377909183502, 'train/grad_norm_pre_clip': 15.746456146240234, 'timing/data': 0.0011137590045109391, 'timing/model': 2.035286019992782, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 0.95}) +[2026-06-23 08:50:03,271][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.09663824737071991, 'train/grad_norm_pre_clip': 28.13612174987793, 'timing/data': 0.0016092440055217594, 'timing/model': 2.040361306979321, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 0.95}) +[2026-06-23 08:50:07,274][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.17775586247444153, 'train/grad_norm_pre_clip': 25.978654861450195, 'timing/data': 0.0012048970093019307, 'timing/model': 2.0221321569988504, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 0.95}) +[2026-06-23 08:50:11,345][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.19187772274017334, 'train/grad_norm_pre_clip': 19.785400390625, 'timing/data': 0.05990758800180629, 'timing/model': 2.0282745379954576, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.95}) +[2026-06-23 08:50:15,349][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.04177389666438103, 'train/grad_norm_pre_clip': 11.652755737304688, 'timing/data': 0.001112255995394662, 'timing/model': 2.0239240329829045, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 0.95}) +[2026-06-23 08:50:19,478][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.10564950853586197, 'train/grad_norm_pre_clip': 27.08232307434082, 'timing/data': 0.0014363949885591865, 'timing/model': 2.03711118499632, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 0.95}) +[2026-06-23 08:50:23,585][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.15677137672901154, 'train/grad_norm_pre_clip': 19.340383529663086, 'timing/data': 0.001409314980264753, 'timing/model': 2.1276341500051785, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 0.95}) +[2026-06-23 08:50:27,652][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.1306096613407135, 'train/grad_norm_pre_clip': 24.298009872436523, 'timing/data': 0.058037369017256424, 'timing/model': 2.02628640900366, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 0.95}) +[2026-06-23 08:50:31,664][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.07796316593885422, 'train/grad_norm_pre_clip': 13.894466400146484, 'timing/data': 0.0011227249924559146, 'timing/model': 2.0282572300056927, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 0.95}) +[2026-06-23 08:50:35,761][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.07592520117759705, 'train/grad_norm_pre_clip': 10.741764068603516, 'timing/data': 0.0011339060147292912, 'timing/model': 2.0228412750002462, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 0.95}) +[2026-06-23 08:50:39,850][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.19943702220916748, 'train/grad_norm_pre_clip': 25.33885383605957, 'timing/data': 0.0011457980144768953, 'timing/model': 2.1078179460018873, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 0.95}) +[2026-06-23 08:50:43,998][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.14818638563156128, 'train/grad_norm_pre_clip': 30.15134048461914, 'timing/data': 0.0831308210035786, 'timing/model': 2.021701592981117, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 0.95}) +[2026-06-23 08:50:48,012][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.07633208483457565, 'train/grad_norm_pre_clip': 12.541130065917969, 'timing/data': 0.0012705980043392628, 'timing/model': 2.0306642770010512, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 0.95}) +[2026-06-23 08:50:52,111][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.21930715441703796, 'train/grad_norm_pre_clip': 30.197649002075195, 'timing/data': 0.0011114040098618716, 'timing/model': 2.024972573010018, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 0.95}) +[2026-06-23 08:50:56,248][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.1380276083946228, 'train/grad_norm_pre_clip': 28.658544540405273, 'timing/data': 0.001146598980994895, 'timing/model': 2.157786024006782, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 0.95}) +[2026-06-23 08:51:00,309][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.08311793953180313, 'train/grad_norm_pre_clip': 11.416952133178711, 'timing/data': 0.05703366099623963, 'timing/model': 2.0243575620115735, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 0.95}) +[2026-06-23 08:51:04,318][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.06903534382581711, 'train/grad_norm_pre_clip': 11.770865440368652, 'timing/data': 0.0012830320047214627, 'timing/model': 2.0274993869825266, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 0.95}) +[2026-06-23 08:51:08,416][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.09343163669109344, 'train/grad_norm_pre_clip': 11.3606538772583, 'timing/data': 0.0011123360018245876, 'timing/model': 2.0219323849887587, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 0.95}) +[2026-06-23 08:51:12,437][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.029217474162578583, 'train/grad_norm_pre_clip': 5.582225322723389, 'timing/data': 0.001308499020524323, 'timing/model': 2.0390177630179096, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 0.95}) +[2026-06-23 08:51:16,504][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.06107832491397858, 'train/grad_norm_pre_clip': 13.00235366821289, 'timing/data': 0.06475056300405413, 'timing/model': 2.021798162983032, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 0.96}) +[2026-06-23 08:51:20,514][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.15378770232200623, 'train/grad_norm_pre_clip': 19.059846878051758, 'timing/data': 0.0011253500124439597, 'timing/model': 2.0298254289955366, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 0.96}) +[2026-06-23 08:51:24,619][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.16144417226314545, 'train/grad_norm_pre_clip': 26.53533172607422, 'timing/data': 0.0012975890131201595, 'timing/model': 2.0287869879975915, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 0.96}) +[2026-06-23 08:51:28,632][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.14857976138591766, 'train/grad_norm_pre_clip': 29.95907211303711, 'timing/data': 0.0012740950041916221, 'timing/model': 2.0272287459811196, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 0.96}) +[2026-06-23 08:51:32,962][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.08084074407815933, 'train/grad_norm_pre_clip': 11.486895561218262, 'timing/data': 0.05291752997436561, 'timing/model': 2.041283681988716, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 0.96}) +[2026-06-23 08:51:36,982][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.17160961031913757, 'train/grad_norm_pre_clip': 18.433975219726562, 'timing/data': 0.0012619330082088709, 'timing/model': 2.025725302984938, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 0.96}) +[2026-06-23 08:51:41,086][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.0456673689186573, 'train/grad_norm_pre_clip': 8.511770248413086, 'timing/data': 0.0012418260157573968, 'timing/model': 2.0241516929818317, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 0.96}) +[2026-06-23 08:51:45,097][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.1232636421918869, 'train/grad_norm_pre_clip': 23.692157745361328, 'timing/data': 0.0012703679967671633, 'timing/model': 2.027628518990241, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 0.96}) +[2026-06-23 08:51:49,174][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.0630546361207962, 'train/grad_norm_pre_clip': 9.04117488861084, 'timing/data': 0.06726257299305871, 'timing/model': 2.025621767999837, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 0.96}) +[2026-06-23 08:51:53,169][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.07797294855117798, 'train/grad_norm_pre_clip': 13.692374229431152, 'timing/data': 0.001271390006877482, 'timing/model': 2.019987763982499, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 0.96}) +[2026-06-23 08:51:57,263][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.10733426362276077, 'train/grad_norm_pre_clip': 30.323122024536133, 'timing/data': 0.0012666409893427044, 'timing/model': 2.022711257013725, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 0.96}) +[2026-06-23 08:52:01,269][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.10946452617645264, 'train/grad_norm_pre_clip': 38.24211120605469, 'timing/data': 0.0011799909989349544, 'timing/model': 2.0261966369871516, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 0.96}) +[2026-06-23 08:52:05,419][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.17005079984664917, 'train/grad_norm_pre_clip': 19.005163192749023, 'timing/data': 0.05467314302222803, 'timing/model': 2.1167851430072915, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.96}) +[2026-06-23 08:52:09,443][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.08640772849321365, 'train/grad_norm_pre_clip': 16.37714958190918, 'timing/data': 0.0011296779848635197, 'timing/model': 2.0430993280024268, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 0.96}) +[2026-06-23 08:52:13,551][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.11426899582147598, 'train/grad_norm_pre_clip': 23.171558380126953, 'timing/data': 0.0015440029965247959, 'timing/model': 2.0274104480049573, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 0.96}) +[2026-06-23 08:52:17,569][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.09587536007165909, 'train/grad_norm_pre_clip': 20.777238845825195, 'timing/data': 0.0012583060015458614, 'timing/model': 2.030420110997511, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 0.96}) +[2026-06-23 08:52:21,649][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.05533454567193985, 'train/grad_norm_pre_clip': 11.744037628173828, 'timing/data': 0.05681935898610391, 'timing/model': 2.0408213549817447, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 0.96}) +[2026-06-23 08:52:25,667][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.09150100499391556, 'train/grad_norm_pre_clip': 15.223730087280273, 'timing/data': 0.001228501001605764, 'timing/model': 2.0298297090048436, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 0.96}) +[2026-06-23 08:52:29,768][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.12029697000980377, 'train/grad_norm_pre_clip': 19.139171600341797, 'timing/data': 0.001124056987464428, 'timing/model': 2.0274638060072903, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 0.96}) +[2026-06-23 08:52:33,795][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.09932825714349747, 'train/grad_norm_pre_clip': 22.590612411499023, 'timing/data': 0.001313537999521941, 'timing/model': 2.0373313130112365, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 0.96}) +[2026-06-23 08:52:37,880][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.09298479557037354, 'train/grad_norm_pre_clip': 27.75847053527832, 'timing/data': 0.0673014709900599, 'timing/model': 2.0297235419857316, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 0.96}) +[2026-06-23 08:52:41,941][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.10910564661026001, 'train/grad_norm_pre_clip': 105.60787963867188, 'timing/data': 0.0014120389823801816, 'timing/model': 2.0439438939793035, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 0.96}) +[2026-06-23 08:52:46,047][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.0367087796330452, 'train/grad_norm_pre_clip': 7.713464736938477, 'timing/data': 0.0011261419858783484, 'timing/model': 2.0256075939978473, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 0.96}) +[2026-06-23 08:52:50,055][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.07618218660354614, 'train/grad_norm_pre_clip': 22.09488868713379, 'timing/data': 0.0011608159984461963, 'timing/model': 2.025704252999276, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 0.96}) +[2026-06-23 08:52:54,127][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.021011747419834137, 'train/grad_norm_pre_clip': 6.450446605682373, 'timing/data': 0.060455747006926686, 'timing/model': 2.030694806977408, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 0.96}) +[2026-06-23 08:52:58,161][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.06730374693870544, 'train/grad_norm_pre_clip': 9.360925674438477, 'timing/data': 0.0012212470173835754, 'timing/model': 2.0468192690168507, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 0.96}) +[2026-06-23 08:53:02,296][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.242044135928154, 'train/grad_norm_pre_clip': 27.030109405517578, 'timing/data': 0.0014236019924283028, 'timing/model': 2.0464717670110986, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 0.97}) +[2026-06-23 08:53:06,313][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.11304544657468796, 'train/grad_norm_pre_clip': 18.965290069580078, 'timing/data': 0.0012738750083371997, 'timing/model': 2.029361063003307, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 0.97}) +[2026-06-23 08:53:10,385][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.04267038404941559, 'train/grad_norm_pre_clip': 9.151994705200195, 'timing/data': 0.05548487900523469, 'timing/model': 2.0308393050218, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 0.97}) +[2026-06-23 08:53:14,458][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.08905666321516037, 'train/grad_norm_pre_clip': 63.88164138793945, 'timing/data': 0.0013942770019639283, 'timing/model': 2.0816507160197943, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 0.97}) +[2026-06-23 08:53:18,574][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.04285937920212746, 'train/grad_norm_pre_clip': 11.801309585571289, 'timing/data': 0.0014487169974017888, 'timing/model': 2.0415879730135202, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 0.97}) +[2026-06-23 08:53:22,591][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.20323459804058075, 'train/grad_norm_pre_clip': 39.90744400024414, 'timing/data': 0.0012325179995968938, 'timing/model': 2.0294930189847946, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 0.97}) +[2026-06-23 08:53:26,661][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2500 +[2026-06-23 08:55:21,430][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2500 +[2026-06-23 08:55:21,431][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2500 +[2026-06-23 08:57:28,996][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2500 +[2026-06-23 09:08:41,632][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-06-23 09:08:41,634][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.07857496291399002, 'train/grad_norm_pre_clip': 21.88428497314453, 'eval/action_loss/seconds': 114.7653449359932, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04963912442326546, 'eval/latency_2/loss': 0.04963912442326546, 'eval/flappy/loss': 0.04963912442326546, 'eval/flappy/latency_2/loss': 0.04963912442326546, 'eval/flap_precision': 0.9590723055934516, 'eval/flap_recall': 0.9487179487179487, 'eval/flap_f1': 0.9538670284938942, 'eval/flap_rate_model': 0.2036111111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.953867028493894, 'eval/latency_2/flap_precision': 0.9590723055934516, 'eval/latency_2/flap_recall': 0.9487179487179487, 'eval/latency_2/flap_f1': 0.9538670284938942, 'eval/latency_2/flap_rate_model': 0.2036111111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.953867028493894, 'eval/action_classification/data_wait_seconds': 50.935218811035156, 'eval/action_classification/predict_seconds': 74.91557312011719, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 173.2800000000008, 'rl_games_eval/mid_train/mean_length': 1416.2, 'rl_games_eval/mid_train/std_reward': 151.64563165485944, 'rl_games_eval/mid_train/std_length': 1224.5701939864452, 'rl_games_eval/mid_train/macro_mean_reward': 173.2800000000008, 'rl_games_eval/mid_train/macro_mean_length': 1416.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 173.2800000000008, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1416.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 151.64563165485944, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1224.5701939864452, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 173.2800000000008, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.054451777978101745, 'timing/model': 2.0271220880094916, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 0.97}) +[2026-06-23 09:08:41,634][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_2500_state +[2026-06-23 09:08:41,642][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 09:08:50,546][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_2500_state/model.safetensors +[2026-06-23 09:08:54,453][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_2500_state/optimizer.bin +[2026-06-23 09:08:54,453][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_2500_state/sampler.bin +[2026-06-23 09:08:54,453][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-23 09:08:54,455][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-23 09:08:54,589][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-23 09:08:54,590][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 09:08:54,613][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 09:08:58,485][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.07264477759599686, 'train/grad_norm_pre_clip': 15.600651741027832, 'timing/data': 0.0014512320049107075, 'timing/model': 1.9556451549869962, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 0.97}) +[2026-06-23 09:09:02,443][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.08011719584465027, 'train/grad_norm_pre_clip': 13.028203010559082, 'timing/data': 0.001318886992521584, 'timing/model': 1.9525987840024754, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 0.97}) +[2026-06-23 09:09:06,312][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.06167474761605263, 'train/grad_norm_pre_clip': 15.312724113464355, 'timing/data': 0.0011428510188125074, 'timing/model': 1.9549567200010642, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 0.97}) +[2026-06-23 09:09:10,495][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.09166785329580307, 'train/grad_norm_pre_clip': 25.234651565551758, 'timing/data': 0.31222588900709525, 'timing/model': 1.9559244559786748, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 0.97}) +[2026-06-23 09:09:14,376][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.03929125517606735, 'train/grad_norm_pre_clip': 22.338193893432617, 'timing/data': 0.0011153209779877216, 'timing/model': 1.9626356349908747, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 0.97}) +[2026-06-23 09:09:18,349][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.020232532173395157, 'train/grad_norm_pre_clip': 9.365621566772461, 'timing/data': 0.0011285149957984686, 'timing/model': 1.964206777018262, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 0.97}) +[2026-06-23 09:09:22,236][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.006837019696831703, 'train/grad_norm_pre_clip': 1.7208389043807983, 'timing/data': 0.0012931289966218174, 'timing/model': 1.9660786940075923, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 0.97}) +[2026-06-23 09:09:26,182][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.055092405527830124, 'train/grad_norm_pre_clip': 20.25687599182129, 'timing/data': 0.0558164739923086, 'timing/model': 1.9680471390020102, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 0.97}) +[2026-06-23 09:09:30,092][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.09857656806707382, 'train/grad_norm_pre_clip': 27.784231185913086, 'timing/data': 0.0012646570103242993, 'timing/model': 1.9738904260157142, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 0.97}) +[2026-06-23 09:09:34,113][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.10830460488796234, 'train/grad_norm_pre_clip': 30.804397583007812, 'timing/data': 0.001265598984900862, 'timing/model': 1.9941769280121662, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 0.97}) +[2026-06-23 09:09:38,038][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.13421161472797394, 'train/grad_norm_pre_clip': 30.883346557617188, 'timing/data': 0.0011566169850993901, 'timing/model': 1.993030818994157, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 0.97}) +[2026-06-23 09:09:42,274][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.06592460721731186, 'train/grad_norm_pre_clip': 17.328861236572266, 'timing/data': 0.0638293789816089, 'timing/model': 2.2384794280224014, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 0.97}) +[2026-06-23 09:09:46,212][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.010109635069966316, 'train/grad_norm_pre_clip': 5.826082706451416, 'timing/data': 0.0012348610034678131, 'timing/model': 1.9893594829773065, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 0.97}) +[2026-06-23 09:09:50,242][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.0513814240694046, 'train/grad_norm_pre_clip': 28.07921600341797, 'timing/data': 0.00135550502454862, 'timing/model': 2.0003133910067845, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 0.97}) +[2026-06-23 09:09:54,178][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.07497462630271912, 'train/grad_norm_pre_clip': 27.109224319458008, 'timing/data': 0.0012557599984575063, 'timing/model': 1.9896840400178917, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.97}) +[2026-06-23 09:09:58,174][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.059795450419187546, 'train/grad_norm_pre_clip': 22.36442756652832, 'timing/data': 0.05370404999121092, 'timing/model': 1.9948118430038448, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 0.97}) +[2026-06-23 09:10:02,128][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.1553301066160202, 'train/grad_norm_pre_clip': 33.69011688232422, 'timing/data': 0.001360554015263915, 'timing/model': 1.9966082310129423, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 0.97}) +[2026-06-23 09:10:06,175][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.14740018546581268, 'train/grad_norm_pre_clip': 31.67462158203125, 'timing/data': 0.0013679179828613997, 'timing/model': 2.005791578005301, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 0.97}) +[2026-06-23 09:10:10,121][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.1981237828731537, 'train/grad_norm_pre_clip': 27.166772842407227, 'timing/data': 0.0012160670012235641, 'timing/model': 1.9925753370043822, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 0.97}) +[2026-06-23 09:10:14,195][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.11893132328987122, 'train/grad_norm_pre_clip': 41.59383773803711, 'timing/data': 0.058402520982781425, 'timing/model': 2.063106246001553, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 0.98}) +[2026-06-23 09:10:18,151][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.09777926653623581, 'train/grad_norm_pre_clip': 45.99921798706055, 'timing/data': 0.0013705930032301694, 'timing/model': 1.9985737079987302, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 0.98}) +[2026-06-23 09:10:22,200][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.14033186435699463, 'train/grad_norm_pre_clip': 17.872882843017578, 'timing/data': 0.001422308007022366, 'timing/model': 2.002743491990259, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 0.98}) +[2026-06-23 09:10:26,175][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.043605923652648926, 'train/grad_norm_pre_clip': 16.439327239990234, 'timing/data': 0.0012053069949615747, 'timing/model': 2.0179823170183226, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 0.98}) +[2026-06-23 09:10:30,218][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.002675382886081934, 'train/grad_norm_pre_clip': 1.2616149187088013, 'timing/data': 0.06681735898018815, 'timing/model': 2.0129666680004448, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 0.98}) +[2026-06-23 09:10:34,207][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.014750734902918339, 'train/grad_norm_pre_clip': 7.459670066833496, 'timing/data': 0.0012636450119316578, 'timing/model': 2.019121899997117, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 0.98}) +[2026-06-23 09:10:38,297][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.16874513030052185, 'train/grad_norm_pre_clip': 41.77876281738281, 'timing/data': 0.0012878399866167456, 'timing/model': 2.0151991310121957, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 0.98}) +[2026-06-23 09:10:42,298][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.06501086056232452, 'train/grad_norm_pre_clip': 13.83062744140625, 'timing/data': 0.0012666000111494213, 'timing/model': 2.017065346997697, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 0.98}) +[2026-06-23 09:10:46,342][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.1959782838821411, 'train/grad_norm_pre_clip': 40.766990661621094, 'timing/data': 0.060379785019904375, 'timing/model': 2.019743404991459, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 0.98}) +[2026-06-23 09:10:50,495][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.07594853639602661, 'train/grad_norm_pre_clip': 28.56317138671875, 'timing/data': 0.0012181710044387728, 'timing/model': 2.0219992120109964, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 0.98}) +[2026-06-23 09:10:54,579][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.04206506535410881, 'train/grad_norm_pre_clip': 23.501365661621094, 'timing/data': 0.0012432869989424944, 'timing/model': 2.023990198998945, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 0.98}) +[2026-06-23 09:10:58,564][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.07566474378108978, 'train/grad_norm_pre_clip': 27.527454376220703, 'timing/data': 0.0012943519977852702, 'timing/model': 2.013485158007825, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 0.98}) +[2026-06-23 09:11:02,619][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.08018581569194794, 'train/grad_norm_pre_clip': 27.201946258544922, 'timing/data': 0.06654998500016518, 'timing/model': 2.0189604129991494, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 0.98}) +[2026-06-23 09:11:06,613][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.1531544178724289, 'train/grad_norm_pre_clip': 38.818912506103516, 'timing/data': 0.0014005980046931654, 'timing/model': 2.0202126060030423, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 0.98}) +[2026-06-23 09:11:10,702][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.022502712905406952, 'train/grad_norm_pre_clip': 9.099220275878906, 'timing/data': 0.0013105420221108943, 'timing/model': 2.032429648010293, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 0.98}) +[2026-06-23 09:11:14,691][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.057011377066373825, 'train/grad_norm_pre_clip': 36.36943817138672, 'timing/data': 0.0014722199994139373, 'timing/model': 2.0177305880060885, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 0.98}) +[2026-06-23 09:11:18,744][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.028210289776325226, 'train/grad_norm_pre_clip': 18.273374557495117, 'timing/data': 0.05948498100042343, 'timing/model': 2.0197658269898966, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 0.98}) +[2026-06-23 09:11:22,788][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.0979534387588501, 'train/grad_norm_pre_clip': 29.86724281311035, 'timing/data': 0.0014112569915596396, 'timing/model': 2.066930031985976, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 0.98}) +[2026-06-23 09:11:26,879][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.02600349485874176, 'train/grad_norm_pre_clip': 9.869688034057617, 'timing/data': 0.001528584019979462, 'timing/model': 2.0313133109884802, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 0.98}) +[2026-06-23 09:11:30,868][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.11301184445619583, 'train/grad_norm_pre_clip': 72.5166244506836, 'timing/data': 0.0011528799950610846, 'timing/model': 2.014135138975689, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 0.98}) +[2026-06-23 09:11:34,930][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.11778996139764786, 'train/grad_norm_pre_clip': 28.267955780029297, 'timing/data': 0.0657008349953685, 'timing/model': 2.026234575023409, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 0.98}) +[2026-06-23 09:11:38,944][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.09215240925550461, 'train/grad_norm_pre_clip': 15.116765975952148, 'timing/data': 0.001457211998058483, 'timing/model': 2.0336103830195498, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 0.98}) +[2026-06-23 09:11:43,025][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.07494346052408218, 'train/grad_norm_pre_clip': 17.662776947021484, 'timing/data': 0.0012402420106809586, 'timing/model': 2.016633382998407, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 0.98}) +[2026-06-23 09:11:47,016][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.07935736328363419, 'train/grad_norm_pre_clip': 15.570268630981445, 'timing/data': 0.0012861760042142123, 'timing/model': 2.0178839919972233, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 0.98}) +[2026-06-23 09:11:51,081][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.21857643127441406, 'train/grad_norm_pre_clip': 46.28550720214844, 'timing/data': 0.06612851598765701, 'timing/model': 2.0211446760222316, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 0.98}) +[2026-06-23 09:11:55,081][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.06036493554711342, 'train/grad_norm_pre_clip': 15.148977279663086, 'timing/data': 0.001186231995234266, 'timing/model': 2.0233579440100584, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 0.98}) +[2026-06-23 09:11:59,559][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.06633967161178589, 'train/grad_norm_pre_clip': 30.699527740478516, 'timing/data': 0.0017732960113789886, 'timing/model': 2.017232635989785, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 0.99}) +[2026-06-23 09:12:03,549][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.0217597633600235, 'train/grad_norm_pre_clip': 6.377793788909912, 'timing/data': 0.001416447019437328, 'timing/model': 2.0147445269976743, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 0.99}) +[2026-06-23 09:12:07,610][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.12872914969921112, 'train/grad_norm_pre_clip': 32.399288177490234, 'timing/data': 0.06827031300053932, 'timing/model': 2.019714458991075, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 0.99}) +[2026-06-23 09:12:11,624][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.07706021517515182, 'train/grad_norm_pre_clip': 23.398792266845703, 'timing/data': 0.0014302130148280412, 'timing/model': 2.0361159440071788, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 0.99}) +[2026-06-23 09:12:15,720][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.07943680137395859, 'train/grad_norm_pre_clip': 28.610586166381836, 'timing/data': 0.0013102109951432794, 'timing/model': 2.0248550649848767, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.99}) +[2026-06-23 09:12:19,723][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.021947292611002922, 'train/grad_norm_pre_clip': 14.353734016418457, 'timing/data': 0.0012537669972516596, 'timing/model': 2.0217683920054697, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 0.99}) +[2026-06-23 09:12:23,808][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.010015224106609821, 'train/grad_norm_pre_clip': 7.009725093841553, 'timing/data': 0.07082865101983771, 'timing/model': 2.0325022019969765, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 0.99}) +[2026-06-23 09:12:27,826][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.06923288851976395, 'train/grad_norm_pre_clip': 20.96956443786621, 'timing/data': 0.0012413029908202589, 'timing/model': 2.0298783039906994, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 0.99}) +[2026-06-23 09:12:32,183][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.13266979157924652, 'train/grad_norm_pre_clip': 29.02184295654297, 'timing/data': 0.0012672620068769902, 'timing/model': 2.037673785991501, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 0.99}) +[2026-06-23 09:12:36,187][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.17492812871932983, 'train/grad_norm_pre_clip': 38.612327575683594, 'timing/data': 0.0012416349782142788, 'timing/model': 2.021141389006516, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 0.99}) +[2026-06-23 09:12:40,259][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.06067555770277977, 'train/grad_norm_pre_clip': 20.48773956298828, 'timing/data': 0.06433811798342504, 'timing/model': 2.0252703669830225, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 0.99}) +[2026-06-23 09:12:44,271][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.1156138926744461, 'train/grad_norm_pre_clip': 23.467782974243164, 'timing/data': 0.0012229999992996454, 'timing/model': 2.0269849830074236, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 0.99}) +[2026-06-23 09:12:48,358][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.055112164467573166, 'train/grad_norm_pre_clip': 15.931129455566406, 'timing/data': 0.0012222079967614263, 'timing/model': 2.0232235729927197, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 0.99}) +[2026-06-23 09:12:52,363][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.053884025663137436, 'train/grad_norm_pre_clip': 19.950794219970703, 'timing/data': 0.0012761869875248522, 'timing/model': 2.0238494479854126, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 0.99}) +[2026-06-23 09:12:56,419][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.07242631912231445, 'train/grad_norm_pre_clip': 63.434879302978516, 'timing/data': 0.055080130987334996, 'timing/model': 2.020951732993126, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 0.99}) +[2026-06-23 09:13:00,424][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.06118525564670563, 'train/grad_norm_pre_clip': 27.817705154418945, 'timing/data': 0.001180641003884375, 'timing/model': 2.025934639008483, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 0.99}) +[2026-06-23 09:13:04,547][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.10755389928817749, 'train/grad_norm_pre_clip': 18.039518356323242, 'timing/data': 0.001193073985632509, 'timing/model': 2.055015910998918, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 0.99}) +[2026-06-23 09:13:08,557][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.11414313316345215, 'train/grad_norm_pre_clip': 39.98349380493164, 'timing/data': 0.0012392100179567933, 'timing/model': 2.0253971139900386, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 0.99}) +[2026-06-23 09:13:12,630][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.05461592599749565, 'train/grad_norm_pre_clip': 13.129531860351562, 'timing/data': 0.061758269992424175, 'timing/model': 2.0284101409779396, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 0.99}) +[2026-06-23 09:13:16,654][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.09748415648937225, 'train/grad_norm_pre_clip': 26.300874710083008, 'timing/data': 0.0011921920231543481, 'timing/model': 2.0410547430219594, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 0.99}) +[2026-06-23 09:13:20,752][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.03845570608973503, 'train/grad_norm_pre_clip': 9.824067115783691, 'timing/data': 0.0012068799987901002, 'timing/model': 2.0276594320021104, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 0.99}) +[2026-06-23 09:13:24,766][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.1342620551586151, 'train/grad_norm_pre_clip': 19.055248260498047, 'timing/data': 0.0013894769945181906, 'timing/model': 2.0268871340085752, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 0.99}) +[2026-06-23 09:13:28,834][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.09542012214660645, 'train/grad_norm_pre_clip': 25.672155380249023, 'timing/data': 0.053190901002381, 'timing/model': 2.0309641160129104, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 0.99}) +[2026-06-23 09:13:32,861][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.09768868237733841, 'train/grad_norm_pre_clip': 20.99032974243164, 'timing/data': 0.0013066239771433175, 'timing/model': 2.0358218010223936, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 0.99}) +[2026-06-23 09:13:36,962][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.180062398314476, 'train/grad_norm_pre_clip': 29.599992752075195, 'timing/data': 0.0012876100081484765, 'timing/model': 2.029842991993064, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 0.99}) +[2026-06-23 09:13:41,389][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.10795391350984573, 'train/grad_norm_pre_clip': 21.527681350708008, 'timing/data': 0.0013716850080527365, 'timing/model': 2.4413209260092117, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 1.0}) +[2026-06-23 09:13:45,472][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.10677967220544815, 'train/grad_norm_pre_clip': 22.03362464904785, 'timing/data': 0.07031044198083691, 'timing/model': 2.0272872080095112, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 1.0}) +[2026-06-23 09:13:49,492][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.08572004735469818, 'train/grad_norm_pre_clip': 20.752695083618164, 'timing/data': 0.0012623619986698031, 'timing/model': 2.0328657850041054, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 1.0}) +[2026-06-23 09:13:53,595][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.04622625187039375, 'train/grad_norm_pre_clip': 10.222771644592285, 'timing/data': 0.0012716200144495815, 'timing/model': 2.031164910993539, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 1.0}) +[2026-06-23 09:13:57,620][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.07509227842092514, 'train/grad_norm_pre_clip': 19.993850708007812, 'timing/data': 0.001307144993916154, 'timing/model': 2.035868891980499, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 1.0}) +[2026-06-23 09:14:01,690][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.12592794001102448, 'train/grad_norm_pre_clip': 21.607437133789062, 'timing/data': 0.05147481901803985, 'timing/model': 2.031460436002817, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 1.0}) +[2026-06-23 09:14:05,714][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.05691816657781601, 'train/grad_norm_pre_clip': 12.64614486694336, 'timing/data': 0.0012766690051648766, 'timing/model': 2.033762148988899, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 1.0}) +[2026-06-23 09:14:09,816][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.11671900749206543, 'train/grad_norm_pre_clip': 18.36150360107422, 'timing/data': 0.001288209983613342, 'timing/model': 2.0308494660130236, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 1.0}) +[2026-06-23 09:14:13,837][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.058707769960165024, 'train/grad_norm_pre_clip': 11.016746520996094, 'timing/data': 0.0013682180142495781, 'timing/model': 2.033664846996544, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 1.0}) +[2026-06-23 09:14:18,383][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.046673353761434555, 'train/grad_norm_pre_clip': 10.655557632446289, 'timing/data': 0.04507071600528434, 'timing/model': 2.0322607920097653, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 1.0}) +[2026-06-23 09:14:22,410][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.07758849859237671, 'train/grad_norm_pre_clip': 12.04625129699707, 'timing/data': 0.0012950820091646165, 'timing/model': 2.0362037860031705, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 1.0}) +[2026-06-23 09:14:26,468][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.035852402448654175, 'train/grad_norm_pre_clip': 8.786982536315918, 'timing/data': 0.0013051620044279844, 'timing/model': 2.049347691994626, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 1.0}) +[2026-06-23 09:14:30,499][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.06825470179319382, 'train/grad_norm_pre_clip': 15.179901123046875, 'timing/data': 0.0013590010057669133, 'timing/model': 2.037548408989096, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 1.0}) +[2026-06-23 09:14:33,920][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.08286423981189728, 'train/grad_norm_pre_clip': 19.908903121948242, 'timing/data': 0.0010573530162218958, 'timing/model': 1.4281361760222353, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 1.0}) +[2026-06-23 09:14:41,622][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.04858536645770073, 'train/grad_norm_pre_clip': 17.272750854492188, 'timing/data': 0.02073920998373069, 'timing/model': 2.0548760830133688, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 1.0}) +[2026-06-23 09:14:45,745][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.09803912043571472, 'train/grad_norm_pre_clip': 23.647565841674805, 'timing/data': 0.0735617589962203, 'timing/model': 2.0388793460151646, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 1.0}) +[2026-06-23 09:14:49,792][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.045118290930986404, 'train/grad_norm_pre_clip': 10.869170188903809, 'timing/data': 0.0014496280055027455, 'timing/model': 2.018085784977302, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 1.0}) +[2026-06-23 09:14:54,133][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.10171163082122803, 'train/grad_norm_pre_clip': 31.866580963134766, 'timing/data': 0.08520181701169349, 'timing/model': 2.2692986050096806, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 1.0}) +[2026-06-23 09:14:58,137][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.0692388117313385, 'train/grad_norm_pre_clip': 21.58162498474121, 'timing/data': 0.0011445349955465645, 'timing/model': 2.0222721680183895, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 1.0}) +[2026-06-23 09:15:02,206][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.027023009955883026, 'train/grad_norm_pre_clip': 14.337262153625488, 'timing/data': 0.05248646799009293, 'timing/model': 2.0224523410142865, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 1.0}) +[2026-06-23 09:15:06,227][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.009824579581618309, 'train/grad_norm_pre_clip': 3.3611316680908203, 'timing/data': 0.0015520180168095976, 'timing/model': 2.0331584319937974, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 1.0}) +[2026-06-23 09:15:10,334][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.07877500355243683, 'train/grad_norm_pre_clip': 14.508171081542969, 'timing/data': 0.09027823500218801, 'timing/model': 2.0316438720037695, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 1.0}) +[2026-06-23 09:15:14,356][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.17419202625751495, 'train/grad_norm_pre_clip': 22.358304977416992, 'timing/data': 0.0012374460056889802, 'timing/model': 2.0345388510031626, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 1.0}) +[2026-06-23 09:15:18,442][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.05063740536570549, 'train/grad_norm_pre_clip': 20.9244441986084, 'timing/data': 0.0552596130000893, 'timing/model': 2.0265661889861804, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 1.0}) +[2026-06-23 09:15:22,478][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.022601911798119545, 'train/grad_norm_pre_clip': 7.392857074737549, 'timing/data': 0.0012467440101318061, 'timing/model': 2.048617872991599, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 1.0}) +[2026-06-23 09:15:26,600][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.03498610854148865, 'train/grad_norm_pre_clip': 8.967107772827148, 'timing/data': 0.08143922398448922, 'timing/model': 2.0563002869894262, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 1.0}) +[2026-06-23 09:15:30,627][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.114280566573143, 'train/grad_norm_pre_clip': 32.25322723388672, 'timing/data': 0.0012641159992199391, 'timing/model': 2.0395633619918954, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 1.01}) +[2026-06-23 09:15:34,738][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.05701427906751633, 'train/grad_norm_pre_clip': 15.178857803344727, 'timing/data': 0.07335680001415312, 'timing/model': 2.0358963450125884, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 1.01}) +[2026-06-23 09:15:38,770][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.07641629129648209, 'train/grad_norm_pre_clip': 13.311840057373047, 'timing/data': 0.0012309539888519794, 'timing/model': 2.0392502509930637, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.01}) +[2026-06-23 09:15:42,891][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2600 +[2026-06-23 09:17:37,227][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2600 +[2026-06-23 09:17:37,228][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2600 +[2026-06-23 09:19:44,207][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2600 +[2026-06-23 09:19:44,208][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.07894543558359146, 'train/grad_norm_pre_clip': 21.66971206665039, 'eval/action_loss/seconds': 114.33195865098969, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.055140405893325806, 'eval/latency_2/loss': 0.055140405893325806, 'eval/flappy/loss': 0.055140405893325806, 'eval/flappy/latency_2/loss': 0.055140405893325806, 'eval/flap_precision': 0.9351070733290071, 'eval/flap_recall': 0.9723346828609987, 'eval/flap_f1': 0.9533575917962289, 'eval/flap_rate_model': 0.2140277777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9533575917962289, 'eval/latency_2/flap_precision': 0.9351070733290071, 'eval/latency_2/flap_recall': 0.9723346828609987, 'eval/latency_2/flap_f1': 0.9533575917962289, 'eval/latency_2/flap_rate_model': 0.2140277777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9533575917962289, 'eval/action_classification/data_wait_seconds': 50.3559684753418, 'eval/action_classification/predict_seconds': 74.87662506103516, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.07784799800720066, 'timing/model': 2.034492520004278, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 1.01}) +[2026-06-23 09:19:48,159][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.11810797452926636, 'train/grad_norm_pre_clip': 22.070165634155273, 'timing/data': 0.0014245820057112724, 'timing/model': 1.9953048810129985, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 1.01}) +[2026-06-23 09:19:52,257][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.030665520578622818, 'train/grad_norm_pre_clip': 10.187849044799805, 'timing/data': 0.056661872018594295, 'timing/model': 1.9984482690051664, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 1.01}) +[2026-06-23 09:19:56,211][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.12474456429481506, 'train/grad_norm_pre_clip': 33.05245590209961, 'timing/data': 0.0013783460017293692, 'timing/model': 1.9934704609913751, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 1.01}) +[2026-06-23 09:20:00,251][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.1640881448984146, 'train/grad_norm_pre_clip': 42.70343017578125, 'timing/data': 0.08461966901086271, 'timing/model': 1.9982028449885547, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 1.01}) +[2026-06-23 09:20:04,224][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.12479186058044434, 'train/grad_norm_pre_clip': 48.27356719970703, 'timing/data': 0.00124676400446333, 'timing/model': 2.0113583419879433, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 1.01}) +[2026-06-23 09:20:08,258][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.1985330581665039, 'train/grad_norm_pre_clip': 44.479454040527344, 'timing/data': 0.055553002981469035, 'timing/model': 2.0078586140007246, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 1.01}) +[2026-06-23 09:20:12,240][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.08578971773386002, 'train/grad_norm_pre_clip': 37.96433639526367, 'timing/data': 0.0011012050090357661, 'timing/model': 2.015695284004323, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 1.01}) +[2026-06-23 09:20:16,295][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.07536822557449341, 'train/grad_norm_pre_clip': 29.960927963256836, 'timing/data': 0.07490059098927304, 'timing/model': 2.005077906011138, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 1.01}) +[2026-06-23 09:20:20,262][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.06091118976473808, 'train/grad_norm_pre_clip': 14.982542037963867, 'timing/data': 0.001192623982205987, 'timing/model': 2.0050750609952956, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 1.01}) +[2026-06-23 09:20:24,450][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.08324070274829865, 'train/grad_norm_pre_clip': 20.75343894958496, 'timing/data': 0.06706362101249397, 'timing/model': 2.140361263009254, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 1.01}) +[2026-06-23 09:20:28,465][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.0820779874920845, 'train/grad_norm_pre_clip': 37.0092887878418, 'timing/data': 0.0012844229931943119, 'timing/model': 2.0248309720191173, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 1.01}) +[2026-06-23 09:20:32,549][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.07471366226673126, 'train/grad_norm_pre_clip': 28.783117294311523, 'timing/data': 0.11331541801337153, 'timing/model': 2.0109432520112023, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 1.01}) +[2026-06-23 09:20:36,532][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.11221197247505188, 'train/grad_norm_pre_clip': 36.86716842651367, 'timing/data': 0.0012390890042297542, 'timing/model': 2.013052308000624, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 1.01}) +[2026-06-23 09:20:40,587][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.16437022387981415, 'train/grad_norm_pre_clip': 27.619722366333008, 'timing/data': 0.057766963000176474, 'timing/model': 2.011052372981794, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 1.01}) +[2026-06-23 09:20:44,580][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.07057322561740875, 'train/grad_norm_pre_clip': 18.52402687072754, 'timing/data': 0.0014193029783200473, 'timing/model': 2.0179592439963017, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 1.01}) +[2026-06-23 09:20:48,649][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.1055949404835701, 'train/grad_norm_pre_clip': 23.171722412109375, 'timing/data': 0.08657655399292707, 'timing/model': 2.015428938000696, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 1.01}) +[2026-06-23 09:20:52,647][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.020209738984704018, 'train/grad_norm_pre_clip': 9.93281364440918, 'timing/data': 0.0016739429847802967, 'timing/model': 2.0165868980111554, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 1.01}) +[2026-06-23 09:20:56,713][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.08085792511701584, 'train/grad_norm_pre_clip': 20.93882942199707, 'timing/data': 0.05678098200587556, 'timing/model': 2.021110765985213, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 1.01}) +[2026-06-23 09:21:00,959][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.1254817247390747, 'train/grad_norm_pre_clip': 27.611562728881836, 'timing/data': 0.001187624002341181, 'timing/model': 2.026342110999394, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 1.01}) +[2026-06-23 09:21:05,039][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.04636983200907707, 'train/grad_norm_pre_clip': 52.505428314208984, 'timing/data': 0.0829875529743731, 'timing/model': 2.0218350230134092, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 1.01}) +[2026-06-23 09:21:09,051][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.1601342409849167, 'train/grad_norm_pre_clip': 26.363073348999023, 'timing/data': 0.001387935015372932, 'timing/model': 2.032831610995345, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 1.01}) +[2026-06-23 09:21:13,120][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.055913206189870834, 'train/grad_norm_pre_clip': 16.382356643676758, 'timing/data': 0.05707365300622769, 'timing/model': 2.021087080007419, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 1.01}) +[2026-06-23 09:21:17,122][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.0932779535651207, 'train/grad_norm_pre_clip': 22.267375946044922, 'timing/data': 0.0012839520059060305, 'timing/model': 2.0223166220239364, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 1.02}) +[2026-06-23 09:21:21,199][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.12092567980289459, 'train/grad_norm_pre_clip': 31.80706214904785, 'timing/data': 0.08455603101174347, 'timing/model': 2.0162848769978154, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 1.02}) +[2026-06-23 09:21:25,336][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.07823130488395691, 'train/grad_norm_pre_clip': 17.264278411865234, 'timing/data': 0.0011746299860533327, 'timing/model': 2.0363816280150786, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 1.02}) +[2026-06-23 09:21:29,521][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.09012682735919952, 'train/grad_norm_pre_clip': 16.256839752197266, 'timing/data': 0.06881499601877294, 'timing/model': 2.019507421005983, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 1.02}) +[2026-06-23 09:21:33,561][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.02801457978785038, 'train/grad_norm_pre_clip': 9.267251014709473, 'timing/data': 0.0013995160115882754, 'timing/model': 2.0283454599848483, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 1.02}) +[2026-06-23 09:21:37,628][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.1396891176700592, 'train/grad_norm_pre_clip': 27.22346305847168, 'timing/data': 0.06956262799212709, 'timing/model': 2.020884464000119, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 1.02}) +[2026-06-23 09:21:41,650][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.06272507458925247, 'train/grad_norm_pre_clip': 17.6837215423584, 'timing/data': 0.0012116689758840948, 'timing/model': 2.0407033619994763, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 1.02}) +[2026-06-23 09:21:45,743][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.1021520122885704, 'train/grad_norm_pre_clip': 19.060516357421875, 'timing/data': 0.07207788500818424, 'timing/model': 2.019743445009226, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 1.02}) +[2026-06-23 09:21:49,749][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.030066970735788345, 'train/grad_norm_pre_clip': 11.508416175842285, 'timing/data': 0.0012094539997633547, 'timing/model': 2.025578075001249, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 1.02}) +[2026-06-23 09:21:53,821][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.05962269380688667, 'train/grad_norm_pre_clip': 10.426708221435547, 'timing/data': 0.07469712500460446, 'timing/model': 2.020156668004347, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 1.02}) +[2026-06-23 09:21:57,825][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.062342073768377304, 'train/grad_norm_pre_clip': 17.869089126586914, 'timing/data': 0.0012784019927494228, 'timing/model': 2.02582113299286, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 1.02}) +[2026-06-23 09:22:01,909][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.03473292663693428, 'train/grad_norm_pre_clip': 15.576963424682617, 'timing/data': 0.05672011899878271, 'timing/model': 2.024412480997853, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 1.02}) +[2026-06-23 09:22:06,168][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.14847348630428314, 'train/grad_norm_pre_clip': 33.06107711791992, 'timing/data': 0.0012158059980720282, 'timing/model': 2.2767394139955286, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 1.02}) +[2026-06-23 09:22:10,252][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.10506054013967514, 'train/grad_norm_pre_clip': 18.948516845703125, 'timing/data': 0.07560992098296992, 'timing/model': 2.0245011940132827, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 1.02}) +[2026-06-23 09:22:14,263][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.08040883392095566, 'train/grad_norm_pre_clip': 18.233797073364258, 'timing/data': 0.0013453249994199723, 'timing/model': 2.030264561995864, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 1.02}) +[2026-06-23 09:22:18,355][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.04389459639787674, 'train/grad_norm_pre_clip': 22.742324829101562, 'timing/data': 0.05772575701121241, 'timing/model': 2.030780405009864, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 1.02}) +[2026-06-23 09:22:22,395][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.09804914891719818, 'train/grad_norm_pre_clip': 20.35122299194336, 'timing/data': 0.0013489229895640165, 'timing/model': 2.045983360003447, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 1.02}) +[2026-06-23 09:22:26,492][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.08063992112874985, 'train/grad_norm_pre_clip': 15.02924633026123, 'timing/data': 0.08568426399142481, 'timing/model': 2.0283438420156017, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 1.02}) +[2026-06-23 09:22:30,511][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.0777701884508133, 'train/grad_norm_pre_clip': 15.808637619018555, 'timing/data': 0.001621128001715988, 'timing/model': 2.031514030008111, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 1.02}) +[2026-06-23 09:22:34,625][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.06219189986586571, 'train/grad_norm_pre_clip': 10.993173599243164, 'timing/data': 0.06834724801592529, 'timing/model': 2.0338986849819776, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 1.02}) +[2026-06-23 09:22:38,668][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.02398800104856491, 'train/grad_norm_pre_clip': 10.456123352050781, 'timing/data': 0.001450412004487589, 'timing/model': 2.0607296639936976, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 1.02}) +[2026-06-23 09:22:42,754][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.05309456214308739, 'train/grad_norm_pre_clip': 17.21870994567871, 'timing/data': 0.07647267400170676, 'timing/model': 2.027373067016015, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 1.02}) +[2026-06-23 09:22:46,766][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.09886821359395981, 'train/grad_norm_pre_clip': 17.46738624572754, 'timing/data': 0.001196283003082499, 'timing/model': 2.027872457023477, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 1.02}) +[2026-06-23 09:22:50,869][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.08365416526794434, 'train/grad_norm_pre_clip': 20.366491317749023, 'timing/data': 0.05804466002155095, 'timing/model': 2.0329764609923586, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 1.02}) +[2026-06-23 09:22:54,885][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.11139080673456192, 'train/grad_norm_pre_clip': 15.504507064819336, 'timing/data': 0.0013977049966342747, 'timing/model': 2.030728730984265, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 1.02}) +[2026-06-23 09:22:58,976][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.06242698058485985, 'train/grad_norm_pre_clip': 10.416117668151855, 'timing/data': 0.07479622898972593, 'timing/model': 2.029680178005947, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.02}) +[2026-06-23 09:23:03,006][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.10961873084306717, 'train/grad_norm_pre_clip': 20.163984298706055, 'timing/data': 0.0012239730276633054, 'timing/model': 2.0439515370235313, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 1.03}) +[2026-06-23 09:23:07,117][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.05654912441968918, 'train/grad_norm_pre_clip': 17.604915618896484, 'timing/data': 0.06891873700078577, 'timing/model': 2.0288378020050004, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 1.03}) +[2026-06-23 09:23:11,135][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.03790678828954697, 'train/grad_norm_pre_clip': 10.66342830657959, 'timing/data': 0.001222781022079289, 'timing/model': 2.0296904759889003, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 1.03}) +[2026-06-23 09:23:15,340][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.08263105899095535, 'train/grad_norm_pre_clip': 16.885053634643555, 'timing/data': 0.07412964300601743, 'timing/model': 2.025914077006746, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 1.03}) +[2026-06-23 09:23:19,355][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.1263227015733719, 'train/grad_norm_pre_clip': 25.037429809570312, 'timing/data': 0.0012281509989406914, 'timing/model': 2.028032512986101, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 1.03}) +[2026-06-23 09:23:23,442][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.02482466772198677, 'train/grad_norm_pre_clip': 10.647331237792969, 'timing/data': 0.05758959299419075, 'timing/model': 2.028775894024875, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 1.03}) +[2026-06-23 09:23:27,458][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.17659714818000793, 'train/grad_norm_pre_clip': 28.1844425201416, 'timing/data': 0.0012441610160749406, 'timing/model': 2.028594887990039, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 1.03}) +[2026-06-23 09:23:31,544][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.028786631301045418, 'train/grad_norm_pre_clip': 13.225571632385254, 'timing/data': 0.07534353499067947, 'timing/model': 2.0259871520102024, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 1.03}) +[2026-06-23 09:23:35,563][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.1632833182811737, 'train/grad_norm_pre_clip': 31.54172134399414, 'timing/data': 0.0011840900115203112, 'timing/model': 2.0346631849824917, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 1.03}) +[2026-06-23 09:23:39,663][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.13584935665130615, 'train/grad_norm_pre_clip': 30.174402236938477, 'timing/data': 0.06693900001118891, 'timing/model': 2.0273006560164504, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 1.03}) +[2026-06-23 09:23:43,689][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.054864220321178436, 'train/grad_norm_pre_clip': 16.662057876586914, 'timing/data': 0.0013529919961001724, 'timing/model': 2.04070042600506, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 1.03}) +[2026-06-23 09:23:47,987][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.1443033516407013, 'train/grad_norm_pre_clip': 24.32898712158203, 'timing/data': 0.08664182500797324, 'timing/model': 2.2291043139994144, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 1.03}) +[2026-06-23 09:23:52,038][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.08725502341985703, 'train/grad_norm_pre_clip': 14.693592071533203, 'timing/data': 0.0015956700081005692, 'timing/model': 2.06590384899755, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 1.03}) +[2026-06-23 09:23:56,156][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.11614182591438293, 'train/grad_norm_pre_clip': 24.111902236938477, 'timing/data': 0.07131380701321177, 'timing/model': 2.0342729719995987, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 1.03}) +[2026-06-23 09:24:00,177][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.10594987869262695, 'train/grad_norm_pre_clip': 21.080286026000977, 'timing/data': 0.0013164439878892154, 'timing/model': 2.034219172986923, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 1.03}) +[2026-06-23 09:24:04,284][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.07726427912712097, 'train/grad_norm_pre_clip': 18.926307678222656, 'timing/data': 0.08616685299784876, 'timing/model': 2.0315029979974497, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 1.03}) +[2026-06-23 09:24:08,319][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.03453690558671951, 'train/grad_norm_pre_clip': 8.766944885253906, 'timing/data': 0.001333575986791402, 'timing/model': 2.047277828998631, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 1.03}) +[2026-06-23 09:24:12,430][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.08016864955425262, 'train/grad_norm_pre_clip': 15.845010757446289, 'timing/data': 0.06623800899251364, 'timing/model': 2.0264654599886853, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 1.03}) +[2026-06-23 09:24:16,439][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.10966566950082779, 'train/grad_norm_pre_clip': 14.623929023742676, 'timing/data': 0.0012109699891880155, 'timing/model': 2.026800680992892, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 1.03}) +[2026-06-23 09:24:20,541][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.10232763737440109, 'train/grad_norm_pre_clip': 23.558605194091797, 'timing/data': 0.08870851999381557, 'timing/model': 2.031465184991248, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 1.03}) +[2026-06-23 09:24:24,810][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.11358319967985153, 'train/grad_norm_pre_clip': 21.345788955688477, 'timing/data': 0.0013798519794363528, 'timing/model': 2.040064144995995, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 1.03}) +[2026-06-23 09:24:28,922][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.13883818686008453, 'train/grad_norm_pre_clip': 22.30038070678711, 'timing/data': 0.06528527400223538, 'timing/model': 2.0386557599995285, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 1.03}) +[2026-06-23 09:24:32,942][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.09822281450033188, 'train/grad_norm_pre_clip': 20.908716201782227, 'timing/data': 0.001207573019200936, 'timing/model': 2.0296351500146557, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 1.03}) +[2026-06-23 09:24:37,046][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.0686568096280098, 'train/grad_norm_pre_clip': 19.294076919555664, 'timing/data': 0.08554634900065139, 'timing/model': 2.0322943170031067, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 1.03}) +[2026-06-23 09:24:41,082][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.15546846389770508, 'train/grad_norm_pre_clip': 14.668135643005371, 'timing/data': 0.0012829920160584152, 'timing/model': 2.048879426001804, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 1.03}) +[2026-06-23 09:24:45,185][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.0593089759349823, 'train/grad_norm_pre_clip': 17.321361541748047, 'timing/data': 0.0553296820144169, 'timing/model': 2.03513273599674, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 1.03}) +[2026-06-23 09:24:49,238][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.03504784405231476, 'train/grad_norm_pre_clip': 12.10933780670166, 'timing/data': 0.0012423770094756037, 'timing/model': 2.059593568992568, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 1.04}) +[2026-06-23 09:24:53,347][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.10818558186292648, 'train/grad_norm_pre_clip': 20.034423828125, 'timing/data': 0.09144313001888804, 'timing/model': 2.0301236460218206, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 1.04}) +[2026-06-23 09:24:57,407][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.04728362709283829, 'train/grad_norm_pre_clip': 12.143088340759277, 'timing/data': 0.0011917939991690218, 'timing/model': 2.042132270988077, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 1.04}) +[2026-06-23 09:25:01,503][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.09457613527774811, 'train/grad_norm_pre_clip': 14.194918632507324, 'timing/data': 0.05437235900899395, 'timing/model': 2.0335200580011588, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 1.04}) +[2026-06-23 09:25:05,530][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.02084873989224434, 'train/grad_norm_pre_clip': 8.44702434539795, 'timing/data': 0.0013404590135905892, 'timing/model': 2.0362946999957785, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 1.04}) +[2026-06-23 09:25:09,638][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.16461096704006195, 'train/grad_norm_pre_clip': 32.67943572998047, 'timing/data': 0.08660866599529982, 'timing/model': 2.0326030850119423, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 1.04}) +[2026-06-23 09:25:13,667][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.0851558968424797, 'train/grad_norm_pre_clip': 12.569146156311035, 'timing/data': 0.0011729089892469347, 'timing/model': 2.0376200140162837, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 1.04}) +[2026-06-23 09:25:17,753][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.058426789939403534, 'train/grad_norm_pre_clip': 9.605033874511719, 'timing/data': 0.05738258399651386, 'timing/model': 2.025802803982515, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 1.04}) +[2026-06-23 09:25:21,784][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.17124643921852112, 'train/grad_norm_pre_clip': 23.610637664794922, 'timing/data': 0.0011733599822036922, 'timing/model': 2.034187048993772, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 1.04}) +[2026-06-23 09:25:25,900][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.09456048905849457, 'train/grad_norm_pre_clip': 25.746652603149414, 'timing/data': 0.08599287399556488, 'timing/model': 2.038849060976645, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 1.04}) +[2026-06-23 09:25:30,192][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.058122601360082626, 'train/grad_norm_pre_clip': 9.953944206237793, 'timing/data': 0.0012430190108716488, 'timing/model': 2.2763552800170146, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 1.04}) +[2026-06-23 09:25:34,306][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.04551059007644653, 'train/grad_norm_pre_clip': 17.506847381591797, 'timing/data': 0.07129181100754067, 'timing/model': 2.0284890679758973, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 1.04}) +[2026-06-23 09:25:38,338][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.11766918748617172, 'train/grad_norm_pre_clip': 25.84261131286621, 'timing/data': 0.0016375069972127676, 'timing/model': 2.043584501981968, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 1.04}) +[2026-06-23 09:25:42,450][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.05491884797811508, 'train/grad_norm_pre_clip': 11.742242813110352, 'timing/data': 0.08439553898642771, 'timing/model': 2.0396181610121857, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 1.04}) +[2026-06-23 09:25:46,473][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.030896244570612907, 'train/grad_norm_pre_clip': 9.339812278747559, 'timing/data': 0.0012099269952159375, 'timing/model': 2.0349373490025755, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 1.04}) +[2026-06-23 09:25:50,573][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.04743979871273041, 'train/grad_norm_pre_clip': 17.36951446533203, 'timing/data': 0.05703043800895102, 'timing/model': 2.0332038369961083, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 1.04}) +[2026-06-23 09:25:54,603][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.12466734647750854, 'train/grad_norm_pre_clip': 26.064990997314453, 'timing/data': 0.0012071819801349193, 'timing/model': 2.037848015985219, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 1.04}) +[2026-06-23 09:25:58,704][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.07665533572435379, 'train/grad_norm_pre_clip': 20.047073364257812, 'timing/data': 0.07612818002235144, 'timing/model': 2.0339663529885, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 1.04}) +[2026-06-23 09:26:02,801][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.1758812963962555, 'train/grad_norm_pre_clip': 14.347565650939941, 'timing/data': 0.001330358994891867, 'timing/model': 2.1022304340149276, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 1.04}) +[2026-06-23 09:26:06,897][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.11970624327659607, 'train/grad_norm_pre_clip': 18.577743530273438, 'timing/data': 0.055474480002885684, 'timing/model': 2.034731243009446, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 1.04}) +[2026-06-23 09:26:10,924][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.020070038735866547, 'train/grad_norm_pre_clip': 6.147097587585449, 'timing/data': 0.0013400169846136123, 'timing/model': 2.0338349700032268, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 1.04}) +[2026-06-23 09:26:15,030][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.03977372124791145, 'train/grad_norm_pre_clip': 18.659645080566406, 'timing/data': 0.07628338399808854, 'timing/model': 2.0351279040041845, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 1.04}) +[2026-06-23 09:26:19,077][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.05850762501358986, 'train/grad_norm_pre_clip': 24.11094856262207, 'timing/data': 0.0013727379846386611, 'timing/model': 2.0530236049962696, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.04}) +[2026-06-23 09:26:23,187][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.1529352068901062, 'train/grad_norm_pre_clip': 19.899505615234375, 'timing/data': 0.058444950002012774, 'timing/model': 2.0346730099990964, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 1.04}) +[2026-06-23 09:26:27,211][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.05469495803117752, 'train/grad_norm_pre_clip': 12.831717491149902, 'timing/data': 0.0012042469752486795, 'timing/model': 2.0323393500002567, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 1.04}) +[2026-06-23 09:26:31,331][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2700 +[2026-06-23 09:28:26,907][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2700 +[2026-06-23 09:28:26,907][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2700 +[2026-06-23 09:30:34,403][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2700 +[2026-06-23 09:30:34,405][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.09776785224676132, 'train/grad_norm_pre_clip': 26.809799194335938, 'eval/action_loss/seconds': 115.57120567400125, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04696695879101753, 'eval/latency_2/loss': 0.04696695879101753, 'eval/flappy/loss': 0.04696695879101753, 'eval/flappy/latency_2/loss': 0.04696695879101753, 'eval/flap_precision': 0.9372549019607843, 'eval/flap_recall': 0.9676113360323887, 'eval/flap_f1': 0.952191235059761, 'eval/flap_rate_model': 0.2125, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.952191235059761, 'eval/latency_2/flap_precision': 0.9372549019607843, 'eval/latency_2/flap_recall': 0.9676113360323887, 'eval/latency_2/flap_f1': 0.952191235059761, 'eval/latency_2/flap_rate_model': 0.2125, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.952191235059761, 'eval/action_classification/data_wait_seconds': 50.281070709228516, 'eval/action_classification/predict_seconds': 75.52550506591797, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.08900847999029793, 'timing/model': 2.035534571012249, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 1.04}) +[2026-06-23 09:30:38,366][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.031120508909225464, 'train/grad_norm_pre_clip': 10.5437650680542, 'timing/data': 0.0019535210158210248, 'timing/model': 1.998937479016604, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 1.05}) +[2026-06-23 09:30:42,398][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.04290865734219551, 'train/grad_norm_pre_clip': 14.85769271850586, 'timing/data': 0.06497431700699963, 'timing/model': 1.9919363270164467, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 1.05}) +[2026-06-23 09:30:46,369][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.07447933405637741, 'train/grad_norm_pre_clip': 11.488194465637207, 'timing/data': 0.001495475007686764, 'timing/model': 2.007781188003719, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 1.05}) +[2026-06-23 09:30:50,434][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.08690690249204636, 'train/grad_norm_pre_clip': 23.4958553314209, 'timing/data': 0.09311248999438249, 'timing/model': 2.0045400069793686, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 1.05}) +[2026-06-23 09:30:54,401][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.062058355659246445, 'train/grad_norm_pre_clip': 26.30561637878418, 'timing/data': 0.0012092060060240328, 'timing/model': 2.005374910979299, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 1.05}) +[2026-06-23 09:30:58,444][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.03946205973625183, 'train/grad_norm_pre_clip': 9.997552871704102, 'timing/data': 0.05851929399068467, 'timing/model': 2.0029516229988076, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 1.05}) +[2026-06-23 09:31:02,426][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.14446571469306946, 'train/grad_norm_pre_clip': 27.262649536132812, 'timing/data': 0.0015172440034803003, 'timing/model': 2.017610801005503, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 1.05}) +[2026-06-23 09:31:06,710][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.03188077360391617, 'train/grad_norm_pre_clip': 9.811408996582031, 'timing/data': 0.07688941102242097, 'timing/model': 2.012335659994278, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 1.05}) +[2026-06-23 09:31:10,683][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.0801335871219635, 'train/grad_norm_pre_clip': 20.303972244262695, 'timing/data': 0.0012792339839506894, 'timing/model': 2.009191989985993, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 1.05}) +[2026-06-23 09:31:14,746][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.14777758717536926, 'train/grad_norm_pre_clip': 43.53890609741211, 'timing/data': 0.0688119689875748, 'timing/model': 2.0105326990014873, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 1.05}) +[2026-06-23 09:31:18,742][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.016467418521642685, 'train/grad_norm_pre_clip': 6.884251594543457, 'timing/data': 0.0012833630025852472, 'timing/model': 2.018202128005214, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 1.05}) +[2026-06-23 09:31:22,801][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.0452396497130394, 'train/grad_norm_pre_clip': 13.22810173034668, 'timing/data': 0.07867592098773457, 'timing/model': 2.0129407229833305, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 1.05}) +[2026-06-23 09:31:26,780][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.17732812464237213, 'train/grad_norm_pre_clip': 48.55228042602539, 'timing/data': 0.0014065900177229196, 'timing/model': 2.011939824995352, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 1.05}) +[2026-06-23 09:31:30,854][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.02416258119046688, 'train/grad_norm_pre_clip': 13.934928894042969, 'timing/data': 0.05746754800202325, 'timing/model': 2.0242906259954907, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 1.05}) +[2026-06-23 09:31:34,848][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.059799764305353165, 'train/grad_norm_pre_clip': 27.45962142944336, 'timing/data': 0.0013462489878293127, 'timing/model': 2.019365112995729, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 1.05}) +[2026-06-23 09:31:38,945][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.0376538522541523, 'train/grad_norm_pre_clip': 11.690607070922852, 'timing/data': 0.07775434601353481, 'timing/model': 2.0127730990061536, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 1.05}) +[2026-06-23 09:31:42,956][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.0650867223739624, 'train/grad_norm_pre_clip': 31.08904457092285, 'timing/data': 0.0012545889767352492, 'timing/model': 2.035726939997403, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 1.05}) +[2026-06-23 09:31:47,027][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.08369103819131851, 'train/grad_norm_pre_clip': 20.86087417602539, 'timing/data': 0.06808332999935374, 'timing/model': 2.013442600000417, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 1.05}) +[2026-06-23 09:31:51,017][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.06108175218105316, 'train/grad_norm_pre_clip': 21.850780487060547, 'timing/data': 0.0013423910131677985, 'timing/model': 2.017240060988115, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 1.05}) +[2026-06-23 09:31:55,091][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.00717430654913187, 'train/grad_norm_pre_clip': 3.38427734375, 'timing/data': 0.07929653598694131, 'timing/model': 2.0212972319859546, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 1.05}) +[2026-06-23 09:31:59,085][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.019880980253219604, 'train/grad_norm_pre_clip': 10.167292594909668, 'timing/data': 0.0012585759977810085, 'timing/model': 2.019580341002438, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 1.05}) +[2026-06-23 09:32:03,156][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.019540755078196526, 'train/grad_norm_pre_clip': 14.820651054382324, 'timing/data': 0.05800474897841923, 'timing/model': 2.0189895599905867, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 1.05}) +[2026-06-23 09:32:07,156][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.03564881160855293, 'train/grad_norm_pre_clip': 8.96522331237793, 'timing/data': 0.0012014310050290078, 'timing/model': 2.020990300021367, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 1.05}) +[2026-06-23 09:32:11,298][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.06967125087976456, 'train/grad_norm_pre_clip': 24.57623291015625, 'timing/data': 0.09073056001216173, 'timing/model': 2.0763111089763697, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 1.05}) +[2026-06-23 09:32:15,303][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.2004663050174713, 'train/grad_norm_pre_clip': 51.530540466308594, 'timing/data': 0.001263284997548908, 'timing/model': 2.023475377995055, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 1.05}) +[2026-06-23 09:32:19,393][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.10432745516300201, 'train/grad_norm_pre_clip': 39.9781379699707, 'timing/data': 0.05866392600000836, 'timing/model': 2.031194487994071, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 1.06}) +[2026-06-23 09:32:23,400][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.18233925104141235, 'train/grad_norm_pre_clip': 42.476253509521484, 'timing/data': 0.0015878250123932958, 'timing/model': 2.0190756089868955, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 1.06}) +[2026-06-23 09:32:27,486][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.040965620428323746, 'train/grad_norm_pre_clip': 14.09652328491211, 'timing/data': 0.07789293001405895, 'timing/model': 2.022564279992366, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 1.06}) +[2026-06-23 09:32:31,493][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.2047562301158905, 'train/grad_norm_pre_clip': 50.92330551147461, 'timing/data': 0.0013676479866262525, 'timing/model': 2.025907191011356, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 1.06}) +[2026-06-23 09:32:35,580][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.07421296089887619, 'train/grad_norm_pre_clip': 19.099971771240234, 'timing/data': 0.0693970340071246, 'timing/model': 2.019417698000325, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 1.06}) +[2026-06-23 09:32:39,577][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.026206407696008682, 'train/grad_norm_pre_clip': 14.584975242614746, 'timing/data': 0.0012293320032767951, 'timing/model': 2.020802149985684, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 1.06}) +[2026-06-23 09:32:43,669][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.1498388648033142, 'train/grad_norm_pre_clip': 29.325908660888672, 'timing/data': 0.08887639999738894, 'timing/model': 2.0239932300173678, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 1.06}) +[2026-06-23 09:32:47,774][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.0582839772105217, 'train/grad_norm_pre_clip': 31.530317306518555, 'timing/data': 0.001207171007990837, 'timing/model': 2.024015634000534, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 1.06}) +[2026-06-23 09:32:51,863][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.0760836973786354, 'train/grad_norm_pre_clip': 20.859481811523438, 'timing/data': 0.06673038500593975, 'timing/model': 2.028334498987533, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 1.06}) +[2026-06-23 09:32:55,880][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.04154162108898163, 'train/grad_norm_pre_clip': 15.525261878967285, 'timing/data': 0.0012775820214301348, 'timing/model': 2.040174908004701, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 1.06}) +[2026-06-23 09:32:59,983][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.12456629425287247, 'train/grad_norm_pre_clip': 34.540897369384766, 'timing/data': 0.09446055997977965, 'timing/model': 2.028320760000497, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 1.06}) +[2026-06-23 09:33:04,006][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.15752103924751282, 'train/grad_norm_pre_clip': 31.031984329223633, 'timing/data': 0.0012242330121807754, 'timing/model': 2.0301637889933772, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 1.06}) +[2026-06-23 09:33:08,098][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.22176942229270935, 'train/grad_norm_pre_clip': 45.767494201660156, 'timing/data': 0.06542780599556863, 'timing/model': 2.0297998179739807, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 1.06}) +[2026-06-23 09:33:12,117][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.09519898891448975, 'train/grad_norm_pre_clip': 36.077423095703125, 'timing/data': 0.0014270279789343476, 'timing/model': 2.0381559469969943, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 1.06}) +[2026-06-23 09:33:16,214][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.1151629164814949, 'train/grad_norm_pre_clip': 26.70293617248535, 'timing/data': 0.08597633597673848, 'timing/model': 2.026214017008897, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 1.06}) +[2026-06-23 09:33:20,377][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.06304536014795303, 'train/grad_norm_pre_clip': 16.57115936279297, 'timing/data': 0.0011849009897559881, 'timing/model': 2.179510157002369, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 1.06}) +[2026-06-23 09:33:24,475][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.07615459710359573, 'train/grad_norm_pre_clip': 18.72857093811035, 'timing/data': 0.057091675989795476, 'timing/model': 2.0281786730047315, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 1.06}) +[2026-06-23 09:33:28,497][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.06451989710330963, 'train/grad_norm_pre_clip': 17.815542221069336, 'timing/data': 0.0012142449850216508, 'timing/model': 2.0350742720183916, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 1.06}) +[2026-06-23 09:33:32,599][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.023384781554341316, 'train/grad_norm_pre_clip': 9.449118614196777, 'timing/data': 0.07860004098620266, 'timing/model': 2.031342787988251, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 1.06}) +[2026-06-23 09:33:36,620][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.0970679223537445, 'train/grad_norm_pre_clip': 18.049055099487305, 'timing/data': 0.0014043859846424311, 'timing/model': 2.034741062001558, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 1.06}) +[2026-06-23 09:33:40,712][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.037642285227775574, 'train/grad_norm_pre_clip': 15.694748878479004, 'timing/data': 0.05792161900899373, 'timing/model': 2.0288870350050274, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.06}) +[2026-06-23 09:33:44,736][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.08393007516860962, 'train/grad_norm_pre_clip': 23.15518569946289, 'timing/data': 0.001475066994316876, 'timing/model': 2.0330730260175187, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 1.06}) +[2026-06-23 09:33:48,830][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.13965635001659393, 'train/grad_norm_pre_clip': 36.2358283996582, 'timing/data': 0.07617636999930255, 'timing/model': 2.0297743219998665, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 1.06}) +[2026-06-23 09:33:52,850][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.05900855362415314, 'train/grad_norm_pre_clip': 19.175029754638672, 'timing/data': 0.001373398001305759, 'timing/model': 2.0340970149845816, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 1.06}) +[2026-06-23 09:47:34,219][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2750 +[2026-06-23 09:47:34,220][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.03852425143122673, 'train/grad_norm_pre_clip': 10.096199035644531, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 211.5600000000095, 'rl_games_eval/mid_train/mean_length': 1722.6, 'rl_games_eval/mid_train/std_reward': 198.2644758901727, 'rl_games_eval/mid_train/std_length': 1597.6542304265963, 'rl_games_eval/mid_train/macro_mean_reward': 211.5600000000095, 'rl_games_eval/mid_train/macro_mean_length': 1722.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 211.5600000000095, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1722.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 198.2644758901727, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1597.6542304265963, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 211.5600000000095, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.057742828998016194, 'timing/model': 2.0351302480266895, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 1.06}) +[2026-06-23 09:47:38,083][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.13142962753772736, 'train/grad_norm_pre_clip': 28.924789428710938, 'timing/data': 0.0012384389992803335, 'timing/model': 1.954573483002605, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 1.06}) +[2026-06-23 09:47:42,033][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.1569496989250183, 'train/grad_norm_pre_clip': 32.4625358581543, 'timing/data': 0.07638493098784238, 'timing/model': 1.9583045069884975, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 1.07}) +[2026-06-23 09:47:45,933][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.0924685150384903, 'train/grad_norm_pre_clip': 43.12920379638672, 'timing/data': 0.0011872340110130608, 'timing/model': 1.980093112011673, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 1.07}) +[2026-06-23 09:47:49,911][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.015251444652676582, 'train/grad_norm_pre_clip': 4.511620044708252, 'timing/data': 0.05804445699322969, 'timing/model': 1.9691317539836746, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 1.07}) +[2026-06-23 09:47:53,825][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.029182452708482742, 'train/grad_norm_pre_clip': 7.255953788757324, 'timing/data': 0.0013708529877476394, 'timing/model': 1.9809544699965045, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 1.07}) +[2026-06-23 09:47:57,808][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.02900543622672558, 'train/grad_norm_pre_clip': 13.86252498626709, 'timing/data': 0.08208011501119472, 'timing/model': 1.9723503089917358, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 1.07}) +[2026-06-23 09:48:01,707][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.24280759692192078, 'train/grad_norm_pre_clip': 37.77701950073242, 'timing/data': 0.0011117440008092672, 'timing/model': 1.9713214369839989, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 1.07}) +[2026-06-23 09:48:05,811][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.08142243325710297, 'train/grad_norm_pre_clip': 25.504323959350586, 'timing/data': 0.06725554901640862, 'timing/model': 1.9758921320026275, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 1.07}) +[2026-06-23 09:48:09,732][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.0378098264336586, 'train/grad_norm_pre_clip': 10.26076889038086, 'timing/data': 0.0012222490040585399, 'timing/model': 1.9842610959894955, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 1.07}) +[2026-06-23 09:48:13,746][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.02120886743068695, 'train/grad_norm_pre_clip': 12.631217002868652, 'timing/data': 0.08993515200563706, 'timing/model': 1.9843528539931867, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 1.07}) +[2026-06-23 09:48:17,680][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.13411284983158112, 'train/grad_norm_pre_clip': 25.659608840942383, 'timing/data': 0.0014288400125224143, 'timing/model': 1.9881351339863613, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 1.07}) +[2026-06-23 09:48:21,700][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.09617715328931808, 'train/grad_norm_pre_clip': 17.824554443359375, 'timing/data': 0.06795211602002382, 'timing/model': 1.9897958609799389, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 1.07}) +[2026-06-23 09:48:25,647][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.10408651828765869, 'train/grad_norm_pre_clip': 20.567157745361328, 'timing/data': 0.0014357229811139405, 'timing/model': 1.9944278199982364, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 1.07}) +[2026-06-23 09:48:29,674][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.10119680315256119, 'train/grad_norm_pre_clip': 22.897851943969727, 'timing/data': 0.08491842998773791, 'timing/model': 1.9916507260059007, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 1.07}) +[2026-06-23 09:48:33,645][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.13117261230945587, 'train/grad_norm_pre_clip': 23.411149978637695, 'timing/data': 0.0012161369959358126, 'timing/model': 2.020614701003069, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 1.07}) +[2026-06-23 09:48:37,873][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.05734018236398697, 'train/grad_norm_pre_clip': 27.801692962646484, 'timing/data': 0.05788134498288855, 'timing/model': 2.188137883989839, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 1.07}) +[2026-06-23 09:48:41,825][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.12900477647781372, 'train/grad_norm_pre_clip': 23.86391830444336, 'timing/data': 0.0011847190035041422, 'timing/model': 1.997742349980399, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 1.07}) +[2026-06-23 09:48:45,874][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.07058823108673096, 'train/grad_norm_pre_clip': 23.716094970703125, 'timing/data': 0.08773927300353535, 'timing/model': 1.9997793840011582, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 1.07}) +[2026-06-23 09:48:49,830][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.12693236768245697, 'train/grad_norm_pre_clip': 22.08175277709961, 'timing/data': 0.0011759329936467111, 'timing/model': 1.9997019389993511, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 1.07}) +[2026-06-23 09:48:53,876][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.02996855415403843, 'train/grad_norm_pre_clip': 5.9169816970825195, 'timing/data': 0.06874564298777841, 'timing/model': 2.0010081329965033, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 1.07}) +[2026-06-23 09:48:57,849][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.05187306925654411, 'train/grad_norm_pre_clip': 9.100479125976562, 'timing/data': 0.0013281749852467328, 'timing/model': 2.0078688909998164, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 1.07}) +[2026-06-23 09:49:01,890][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.058314234018325806, 'train/grad_norm_pre_clip': 20.786270141601562, 'timing/data': 0.07426057400880381, 'timing/model': 2.0017909670132212, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 1.07}) +[2026-06-23 09:49:05,955][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.034684669226408005, 'train/grad_norm_pre_clip': 13.172679901123047, 'timing/data': 0.0013220430118963122, 'timing/model': 2.0236976300075185, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 1.07}) +[2026-06-23 09:49:09,941][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.09444495290517807, 'train/grad_norm_pre_clip': 24.317001342773438, 'timing/data': 0.0012437689874786884, 'timing/model': 2.0235027779999655, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 1.07}) +[2026-06-23 09:49:13,958][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.0375247560441494, 'train/grad_norm_pre_clip': 18.369049072265625, 'timing/data': 0.0012351020122878253, 'timing/model': 2.0090884400124196, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 1.07}) +[2026-06-23 09:49:18,002][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.11216094344854355, 'train/grad_norm_pre_clip': 25.527559280395508, 'timing/data': 0.07065157499164343, 'timing/model': 2.010250266001094, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 1.07}) +[2026-06-23 09:49:22,073][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.03490554168820381, 'train/grad_norm_pre_clip': 14.440641403198242, 'timing/data': 0.0014239519950933754, 'timing/model': 2.021576125000138, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 1.07}) +[2026-06-23 09:49:26,061][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.1388772577047348, 'train/grad_norm_pre_clip': 27.88203239440918, 'timing/data': 0.0012597079912666231, 'timing/model': 2.022957296023378, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 1.08}) +[2026-06-23 09:49:30,046][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.11213034391403198, 'train/grad_norm_pre_clip': 19.408687591552734, 'timing/data': 0.001353591011138633, 'timing/model': 2.012285022006836, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 1.08}) +[2026-06-23 09:49:34,093][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.09520840644836426, 'train/grad_norm_pre_clip': 21.591150283813477, 'timing/data': 0.06711512900074013, 'timing/model': 2.0081828809925355, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 1.08}) +[2026-06-23 09:49:38,163][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.05454369634389877, 'train/grad_norm_pre_clip': 26.140674591064453, 'timing/data': 0.0013753310195170343, 'timing/model': 2.016124876012327, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 1.08}) +[2026-06-23 09:49:42,146][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.11826156824827194, 'train/grad_norm_pre_clip': 27.827377319335938, 'timing/data': 0.0014095549995545298, 'timing/model': 2.0159563240013085, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 1.08}) +[2026-06-23 09:49:46,390][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.10564691573381424, 'train/grad_norm_pre_clip': 35.00855255126953, 'timing/data': 0.0012040050060022622, 'timing/model': 2.27036844601389, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 1.08}) +[2026-06-23 09:49:50,454][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.07401484251022339, 'train/grad_norm_pre_clip': 27.10897445678711, 'timing/data': 0.06657805599388666, 'timing/model': 2.0223869259934872, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 1.08}) +[2026-06-23 09:49:54,556][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.11112546920776367, 'train/grad_norm_pre_clip': 23.40312385559082, 'timing/data': 0.0011870839807670563, 'timing/model': 2.0438657470222097, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 1.08}) +[2026-06-23 09:49:58,548][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.046904295682907104, 'train/grad_norm_pre_clip': 24.595067977905273, 'timing/data': 0.0012589060061145574, 'timing/model': 2.0163427640218288, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 1.08}) +[2026-06-23 09:50:02,545][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.15458545088768005, 'train/grad_norm_pre_clip': 47.35770034790039, 'timing/data': 0.0012115889985579997, 'timing/model': 2.017554542981088, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 1.08}) +[2026-06-23 09:50:06,606][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.10853702574968338, 'train/grad_norm_pre_clip': 20.510831832885742, 'timing/data': 0.07148841102025472, 'timing/model': 2.016413321980508, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 1.08}) +[2026-06-23 09:50:10,717][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.07576797902584076, 'train/grad_norm_pre_clip': 29.408437728881836, 'timing/data': 0.0013951889995951205, 'timing/model': 2.045829201990273, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 1.08}) +[2026-06-23 09:50:14,726][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.03184191882610321, 'train/grad_norm_pre_clip': 16.803743362426758, 'timing/data': 0.0012662489898502827, 'timing/model': 2.0284240859909914, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 1.08}) +[2026-06-23 09:50:18,885][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.14656738936901093, 'train/grad_norm_pre_clip': 47.3347053527832, 'timing/data': 0.0012563919881358743, 'timing/model': 2.1787749329814687, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 1.08}) +[2026-06-23 09:50:22,957][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.06096425652503967, 'train/grad_norm_pre_clip': 27.56584930419922, 'timing/data': 0.06611849699402228, 'timing/model': 2.0219327689846978, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 1.08}) +[2026-06-23 09:50:27,057][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.034154634922742844, 'train/grad_norm_pre_clip': 21.298606872558594, 'timing/data': 0.0011882160033565015, 'timing/model': 2.036763053998584, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 1.08}) +[2026-06-23 09:50:31,074][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.06214917451143265, 'train/grad_norm_pre_clip': 30.029150009155273, 'timing/data': 0.0012158869940321892, 'timing/model': 2.0378075129992794, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 1.08}) +[2026-06-23 09:50:35,093][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.025178490206599236, 'train/grad_norm_pre_clip': 7.565999507904053, 'timing/data': 0.0012262359960004687, 'timing/model': 2.0388747949909884, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.08}) +[2026-06-23 09:50:39,273][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.10388525575399399, 'train/grad_norm_pre_clip': 25.0180606842041, 'timing/data': 0.06776858598459512, 'timing/model': 2.0285708110022824, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 1.08}) +[2026-06-23 09:50:43,502][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.17930364608764648, 'train/grad_norm_pre_clip': 32.93044662475586, 'timing/data': 0.001368208002531901, 'timing/model': 2.037547950982116, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 1.08}) +[2026-06-23 09:50:47,513][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.10932762175798416, 'train/grad_norm_pre_clip': 37.95767593383789, 'timing/data': 0.0013059430057182908, 'timing/model': 2.0265724470082205, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 1.08}) +[2026-06-23 09:50:51,821][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.04291435703635216, 'train/grad_norm_pre_clip': 13.013648986816406, 'timing/data': 0.001311222993535921, 'timing/model': 2.3113412160018925, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 1.08}) +[2026-06-23 09:50:55,892][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2800 +[2026-06-23 09:52:50,345][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2800 +[2026-06-23 09:52:50,345][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2800 +[2026-06-23 09:54:58,044][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2800 +[2026-06-23 09:54:58,046][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.19428502023220062, 'train/grad_norm_pre_clip': 31.524274826049805, 'eval/action_loss/seconds': 114.44883900799323, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04646546393632889, 'eval/latency_2/loss': 0.04646546393632889, 'eval/flappy/loss': 0.04646546393632889, 'eval/flappy/latency_2/loss': 0.04646546393632889, 'eval/flap_precision': 0.9573748308525034, 'eval/flap_recall': 0.9547908232118758, 'eval/flap_f1': 0.956081081081081, 'eval/flap_rate_model': 0.20527777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.956081081081081, 'eval/latency_2/flap_precision': 0.9573748308525034, 'eval/latency_2/flap_recall': 0.9547908232118758, 'eval/latency_2/flap_f1': 0.956081081081081, 'eval/latency_2/flap_rate_model': 0.20527777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.956081081081081, 'eval/action_classification/data_wait_seconds': 50.54509735107422, 'eval/action_classification/predict_seconds': 75.49066925048828, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06124853697838262, 'timing/model': 2.0254189840052277, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 1.08}) +[2026-06-23 09:55:02,073][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.0662611871957779, 'train/grad_norm_pre_clip': 19.6810245513916, 'timing/data': 0.0011973530054092407, 'timing/model': 1.9967620329989586, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 1.08}) +[2026-06-23 09:55:06,040][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.037091366946697235, 'train/grad_norm_pre_clip': 13.026084899902344, 'timing/data': 0.001400708977598697, 'timing/model': 2.0142909459827933, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 1.08}) +[2026-06-23 09:55:10,010][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.04262934997677803, 'train/grad_norm_pre_clip': 13.852179527282715, 'timing/data': 0.001285364996874705, 'timing/model': 2.0125355109921657, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 1.08}) +[2026-06-23 09:55:14,033][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.16912266612052917, 'train/grad_norm_pre_clip': 55.79902648925781, 'timing/data': 0.06391586500103585, 'timing/model': 2.0006415309908334, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 1.09}) +[2026-06-23 09:55:18,221][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.0357348658144474, 'train/grad_norm_pre_clip': 16.31575584411621, 'timing/data': 0.0013129069993738085, 'timing/model': 2.0195243550115265, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 1.09}) +[2026-06-23 09:55:22,211][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.0400124192237854, 'train/grad_norm_pre_clip': 16.80929183959961, 'timing/data': 0.0013523189991246909, 'timing/model': 2.018862231023377, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 1.09}) +[2026-06-23 09:55:26,174][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.19561871886253357, 'train/grad_norm_pre_clip': 35.36521530151367, 'timing/data': 0.0012095659913029522, 'timing/model': 2.003279353986727, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 1.09}) +[2026-06-23 09:55:30,209][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.1105632334947586, 'train/grad_norm_pre_clip': 21.0647029876709, 'timing/data': 0.059423983009764925, 'timing/model': 2.0114739470009226, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 1.09}) +[2026-06-23 09:55:34,265][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.20899228751659393, 'train/grad_norm_pre_clip': 52.19303512573242, 'timing/data': 0.0012239720090292394, 'timing/model': 2.0058165120135527, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 1.09}) +[2026-06-23 09:55:38,245][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.110367052257061, 'train/grad_norm_pre_clip': 19.210283279418945, 'timing/data': 0.0012577750021591783, 'timing/model': 2.007782803993905, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 1.09}) +[2026-06-23 09:55:42,237][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.053661514073610306, 'train/grad_norm_pre_clip': 11.584649085998535, 'timing/data': 0.0016061890055425465, 'timing/model': 2.0200292149966117, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 1.09}) +[2026-06-23 09:55:46,286][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.06517627835273743, 'train/grad_norm_pre_clip': 14.17287540435791, 'timing/data': 0.0629919970233459, 'timing/model': 2.017710189014906, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 1.09}) +[2026-06-23 09:55:50,524][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.08060996979475021, 'train/grad_norm_pre_clip': 17.66481590270996, 'timing/data': 0.0012202950019855052, 'timing/model': 2.1788386780244764, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 1.09}) +[2026-06-23 09:55:54,519][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.15365742146968842, 'train/grad_norm_pre_clip': 18.513290405273438, 'timing/data': 0.0012825700105167925, 'timing/model': 2.0157971029984765, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 1.09}) +[2026-06-23 09:55:58,512][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.17785577476024628, 'train/grad_norm_pre_clip': 43.223995208740234, 'timing/data': 0.0012169489928055555, 'timing/model': 2.019929719011998, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 1.09}) +[2026-06-23 09:56:02,566][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.03294026106595993, 'train/grad_norm_pre_clip': 18.04206085205078, 'timing/data': 0.05952894702204503, 'timing/model': 2.0184026370116044, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 1.09}) +[2026-06-23 09:56:06,654][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.2041778564453125, 'train/grad_norm_pre_clip': 25.04658317565918, 'timing/data': 0.001172186020994559, 'timing/model': 2.0200779319857247, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 1.09}) +[2026-06-23 09:56:10,654][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.08496324717998505, 'train/grad_norm_pre_clip': 26.280269622802734, 'timing/data': 0.0012820889824070036, 'timing/model': 2.0235160650045145, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 1.09}) +[2026-06-23 09:56:14,664][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.09052697569131851, 'train/grad_norm_pre_clip': 15.853172302246094, 'timing/data': 0.0012552400003187358, 'timing/model': 2.0284956459945533, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 1.09}) +[2026-06-23 09:56:18,738][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.0634649470448494, 'train/grad_norm_pre_clip': 19.140060424804688, 'timing/data': 0.05649366800207645, 'timing/model': 2.0331679999944754, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 1.09}) +[2026-06-23 09:56:22,820][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.06311190873384476, 'train/grad_norm_pre_clip': 12.926302909851074, 'timing/data': 0.0012291609891690314, 'timing/model': 2.018864527984988, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 1.09}) +[2026-06-23 09:56:27,086][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.03631715476512909, 'train/grad_norm_pre_clip': 10.871421813964844, 'timing/data': 0.001627668010769412, 'timing/model': 2.0360515479987953, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 1.09}) +[2026-06-23 09:56:31,102][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.05687493830919266, 'train/grad_norm_pre_clip': 16.344806671142578, 'timing/data': 0.0012242929951753467, 'timing/model': 2.038641519000521, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 1.09}) +[2026-06-23 09:56:35,198][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.1004168689250946, 'train/grad_norm_pre_clip': 17.166484832763672, 'timing/data': 0.07029768300708383, 'timing/model': 2.041970971011324, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 1.09}) +[2026-06-23 09:56:39,306][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.08185449987649918, 'train/grad_norm_pre_clip': 26.96809959411621, 'timing/data': 0.0011204399925190955, 'timing/model': 2.046294689003844, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 1.09}) +[2026-06-23 09:56:43,305][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.10576079785823822, 'train/grad_norm_pre_clip': 18.230207443237305, 'timing/data': 0.0012230089923832566, 'timing/model': 2.0208739590016194, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 1.09}) +[2026-06-23 09:56:47,313][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.05545572191476822, 'train/grad_norm_pre_clip': 18.656944274902344, 'timing/data': 0.001199175021611154, 'timing/model': 2.027668862021528, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 1.09}) +[2026-06-23 09:56:51,382][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.0766688734292984, 'train/grad_norm_pre_clip': 14.14992618560791, 'timing/data': 0.07119873899500817, 'timing/model': 2.020308657985879, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 1.09}) +[2026-06-23 09:56:55,475][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.16722051799297333, 'train/grad_norm_pre_clip': 14.512584686279297, 'timing/data': 0.001188995986012742, 'timing/model': 2.0241035300132353, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 1.09}) +[2026-06-23 09:56:59,576][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.06090131774544716, 'train/grad_norm_pre_clip': 14.172589302062988, 'timing/data': 0.0015025549801066518, 'timing/model': 2.1109235860058106, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 1.1}) +[2026-06-23 09:57:03,583][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.040435101836919785, 'train/grad_norm_pre_clip': 16.76407241821289, 'timing/data': 0.0012039539869874716, 'timing/model': 2.0247770989954006, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 1.1}) +[2026-06-23 09:57:07,668][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.16189387440681458, 'train/grad_norm_pre_clip': 21.93437385559082, 'timing/data': 0.06023590100812726, 'timing/model': 2.038356333010597, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 1.1}) +[2026-06-23 09:57:11,795][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.09083961695432663, 'train/grad_norm_pre_clip': 20.418643951416016, 'timing/data': 0.001401918998453766, 'timing/model': 2.0403681909956504, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 1.1}) +[2026-06-23 09:57:15,828][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.05071144178509712, 'train/grad_norm_pre_clip': 17.84617042541504, 'timing/data': 0.0014480540121439844, 'timing/model': 2.047206103015924, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 1.1}) +[2026-06-23 09:57:19,856][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.08693896234035492, 'train/grad_norm_pre_clip': 14.94291877746582, 'timing/data': 0.0012439969868864864, 'timing/model': 2.032487388001755, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 1.1}) +[2026-06-23 09:57:23,932][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.1043170914053917, 'train/grad_norm_pre_clip': 29.437156677246094, 'timing/data': 0.05777369500719942, 'timing/model': 2.0367899470147677, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 1.1}) +[2026-06-23 09:57:28,035][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.04230432212352753, 'train/grad_norm_pre_clip': 11.387239456176758, 'timing/data': 0.0012178400065749884, 'timing/model': 2.030327181011671, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 1.1}) +[2026-06-23 09:57:32,051][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.03865300491452217, 'train/grad_norm_pre_clip': 8.050667762756348, 'timing/data': 0.0012632840080186725, 'timing/model': 2.0303277430066373, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 1.1}) +[2026-06-23 09:57:36,297][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.05363181233406067, 'train/grad_norm_pre_clip': 16.18303680419922, 'timing/data': 0.0013871020055375993, 'timing/model': 2.050546521990327, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 1.1}) +[2026-06-23 09:57:40,382][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.05623677745461464, 'train/grad_norm_pre_clip': 8.933185577392578, 'timing/data': 0.06980172399198636, 'timing/model': 2.0290399480145425, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 1.1}) +[2026-06-23 09:57:44,486][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.0960921198129654, 'train/grad_norm_pre_clip': 20.748748779296875, 'timing/data': 0.001234058989211917, 'timing/model': 2.032578264013864, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 1.1}) +[2026-06-23 09:57:48,506][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.06013641133904457, 'train/grad_norm_pre_clip': 9.17319107055664, 'timing/data': 0.0012346699950285256, 'timing/model': 2.0314415249740705, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 1.1}) +[2026-06-23 09:57:52,524][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.11050742119550705, 'train/grad_norm_pre_clip': 19.065814971923828, 'timing/data': 0.0012168280081823468, 'timing/model': 2.0299267880036496, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 1.1}) +[2026-06-23 09:57:56,608][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.11951874941587448, 'train/grad_norm_pre_clip': 33.8996467590332, 'timing/data': 0.05954831599956378, 'timing/model': 2.0326944820117205, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.1}) +[2026-06-23 09:58:00,713][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.047117941081523895, 'train/grad_norm_pre_clip': 12.309135437011719, 'timing/data': 0.0014259539893828332, 'timing/model': 2.028160707995994, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 1.1}) +[2026-06-23 09:58:04,733][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.08027220517396927, 'train/grad_norm_pre_clip': 19.152360916137695, 'timing/data': 0.0012495879782363772, 'timing/model': 2.0348930550098885, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 1.1}) +[2026-06-23 09:58:08,795][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.04437965899705887, 'train/grad_norm_pre_clip': 9.59753704071045, 'timing/data': 0.001232787995832041, 'timing/model': 2.035080581001239, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 1.1}) +[2026-06-23 09:58:12,886][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.03647590056061745, 'train/grad_norm_pre_clip': 9.74752426147461, 'timing/data': 0.06839298398699611, 'timing/model': 2.0319744140142575, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 1.1}) +[2026-06-23 09:58:16,996][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.030456678941845894, 'train/grad_norm_pre_clip': 8.803314208984375, 'timing/data': 0.0012453410017769784, 'timing/model': 2.030481146997772, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 1.1}) +[2026-06-23 09:58:21,010][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.052678391337394714, 'train/grad_norm_pre_clip': 20.91368865966797, 'timing/data': 0.0012906139891128987, 'timing/model': 2.0304478149919305, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 1.1}) +[2026-06-23 09:58:25,032][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.02630743570625782, 'train/grad_norm_pre_clip': 6.02325963973999, 'timing/data': 0.001424782007234171, 'timing/model': 2.035248212981969, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 1.1}) +[2026-06-23 09:58:29,128][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.040674738585948944, 'train/grad_norm_pre_clip': 14.53207015991211, 'timing/data': 0.06988403998548165, 'timing/model': 2.035132118006004, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 1.1}) +[2026-06-23 09:58:33,255][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.05130147561430931, 'train/grad_norm_pre_clip': 20.53342628479004, 'timing/data': 0.0012819480034522712, 'timing/model': 2.0487286349816713, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 1.1}) +[2026-06-23 09:58:37,292][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.07975449413061142, 'train/grad_norm_pre_clip': 16.731473922729492, 'timing/data': 0.0014149439812172204, 'timing/model': 2.047535831021378, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 1.1}) +[2026-06-23 09:58:41,566][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.07016746699810028, 'train/grad_norm_pre_clip': 18.960569381713867, 'timing/data': 0.0013815319980494678, 'timing/model': 2.2832240689895116, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 1.1}) +[2026-06-23 09:58:45,665][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.08626361191272736, 'train/grad_norm_pre_clip': 30.410808563232422, 'timing/data': 0.07185358900460415, 'timing/model': 2.040237409994006, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 1.11}) +[2026-06-23 09:58:49,782][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.04215187579393387, 'train/grad_norm_pre_clip': 13.148500442504883, 'timing/data': 0.0012572419946081936, 'timing/model': 2.0301986359991133, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 1.11}) +[2026-06-23 09:58:53,803][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.12181627750396729, 'train/grad_norm_pre_clip': 33.12337112426758, 'timing/data': 0.0012867860205005854, 'timing/model': 2.0313119220081717, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 1.11}) +[2026-06-23 09:58:57,856][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.11083372682332993, 'train/grad_norm_pre_clip': 25.256511688232422, 'timing/data': 0.0012252430024091154, 'timing/model': 2.0597522569878493, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 1.11}) +[2026-06-23 09:59:01,976][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.07382538914680481, 'train/grad_norm_pre_clip': 25.135089874267578, 'timing/data': 0.07134810698335059, 'timing/model': 2.0423540890042204, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 1.11}) +[2026-06-23 09:59:06,099][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.11399580538272858, 'train/grad_norm_pre_clip': 23.053409576416016, 'timing/data': 0.0020353099971543998, 'timing/model': 2.0365479229949415, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 1.11}) +[2026-06-23 09:59:10,105][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.10583235323429108, 'train/grad_norm_pre_clip': 41.91032028198242, 'timing/data': 0.001161385007435456, 'timing/model': 2.024227834015619, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 1.11}) +[2026-06-23 09:59:14,125][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.05061967670917511, 'train/grad_norm_pre_clip': 17.57504653930664, 'timing/data': 0.0011578580015338957, 'timing/model': 2.035603076015832, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 1.11}) +[2026-06-23 09:59:18,351][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.01902303472161293, 'train/grad_norm_pre_clip': 7.096930503845215, 'timing/data': 0.05992721099755727, 'timing/model': 2.0256427559943404, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 1.11}) +[2026-06-23 09:59:22,452][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.15846548974514008, 'train/grad_norm_pre_clip': 26.642868041992188, 'timing/data': 0.0011083769786637276, 'timing/model': 2.0265735059801955, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 1.11}) +[2026-06-23 09:59:26,475][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.13507886230945587, 'train/grad_norm_pre_clip': 28.30817985534668, 'timing/data': 0.001226104999659583, 'timing/model': 2.0415804559888784, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 1.11}) +[2026-06-23 09:59:30,501][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.09305021166801453, 'train/grad_norm_pre_clip': 40.210601806640625, 'timing/data': 0.001111924008000642, 'timing/model': 2.029093016986735, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 1.11}) +[2026-06-23 09:59:34,577][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.06080058217048645, 'train/grad_norm_pre_clip': 14.034299850463867, 'timing/data': 0.05942102801054716, 'timing/model': 2.0294493139954284, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 1.11}) +[2026-06-23 09:59:38,692][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.043486565351486206, 'train/grad_norm_pre_clip': 16.04156494140625, 'timing/data': 0.0011234459816478193, 'timing/model': 2.0436061049986165, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 1.11}) +[2026-06-23 09:59:42,712][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.05851282924413681, 'train/grad_norm_pre_clip': 12.611421585083008, 'timing/data': 0.00115066600847058, 'timing/model': 2.0307947680121288, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 1.11}) +[2026-06-23 09:59:46,729][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.03764279559254646, 'train/grad_norm_pre_clip': 11.332901000976562, 'timing/data': 0.0013047310058027506, 'timing/model': 2.0319493590213824, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 1.11}) +[2026-06-23 09:59:50,829][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.13624493777751923, 'train/grad_norm_pre_clip': 43.82988739013672, 'timing/data': 0.059039713989477605, 'timing/model': 2.0556664299801923, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 1.11}) +[2026-06-23 09:59:54,931][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.12318620830774307, 'train/grad_norm_pre_clip': 22.863624572753906, 'timing/data': 0.0011335439921822399, 'timing/model': 2.030289491027361, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 1.11}) +[2026-06-23 09:59:58,943][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.08842014521360397, 'train/grad_norm_pre_clip': 35.3525505065918, 'timing/data': 0.0011695209832396358, 'timing/model': 2.027849482023157, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 1.11}) +[2026-06-23 10:00:02,951][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.0076054600067436695, 'train/grad_norm_pre_clip': 3.8206071853637695, 'timing/data': 0.0012661180226132274, 'timing/model': 2.025083629996516, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 1.11}) +[2026-06-23 10:00:07,029][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.10695662349462509, 'train/grad_norm_pre_clip': 37.69730758666992, 'timing/data': 0.060326161998091266, 'timing/model': 2.0315667949907947, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 1.11}) +[2026-06-23 10:00:11,163][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.08786043524742126, 'train/grad_norm_pre_clip': 24.65492057800293, 'timing/data': 0.0012231500004418194, 'timing/model': 2.04909996301285, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 1.11}) +[2026-06-23 10:00:15,176][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.14297780394554138, 'train/grad_norm_pre_clip': 39.196189880371094, 'timing/data': 0.001259618002222851, 'timing/model': 2.0284557540144306, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 1.11}) +[2026-06-23 10:00:19,190][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.17770928144454956, 'train/grad_norm_pre_clip': 55.281314849853516, 'timing/data': 0.0012007280020043254, 'timing/model': 2.0289534619951155, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 1.11}) +[2026-06-23 10:00:23,281][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.06829002499580383, 'train/grad_norm_pre_clip': 16.25627326965332, 'timing/data': 0.06850790700991638, 'timing/model': 2.028593074006494, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 1.11}) +[2026-06-23 10:00:27,611][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.13520547747612, 'train/grad_norm_pre_clip': 25.715360641479492, 'timing/data': 0.0012230300053488463, 'timing/model': 2.075440365995746, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 1.11}) +[2026-06-23 10:00:31,623][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.09916336834430695, 'train/grad_norm_pre_clip': 24.389558792114258, 'timing/data': 0.0012236709881108254, 'timing/model': 2.025130864989478, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 1.12}) +[2026-06-23 10:00:35,644][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.09943391382694244, 'train/grad_norm_pre_clip': 22.387435913085938, 'timing/data': 0.001119397988077253, 'timing/model': 2.031418087979546, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 1.12}) +[2026-06-23 10:00:39,733][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.11917424947023392, 'train/grad_norm_pre_clip': 25.20111656188965, 'timing/data': 0.05935935801244341, 'timing/model': 2.0327765949768946, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 1.12}) +[2026-06-23 10:00:43,836][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.10956619679927826, 'train/grad_norm_pre_clip': 36.812564849853516, 'timing/data': 0.0012223180092405528, 'timing/model': 2.029831772990292, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 1.12}) +[2026-06-23 10:00:47,869][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.047054458409547806, 'train/grad_norm_pre_clip': 27.3354434967041, 'timing/data': 0.0012798949901480228, 'timing/model': 2.044297364016529, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 1.12}) +[2026-06-23 10:00:51,898][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.07424531131982803, 'train/grad_norm_pre_clip': 16.869550704956055, 'timing/data': 0.0014158150006551296, 'timing/model': 2.036082525999518, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 1.12}) +[2026-06-23 10:00:55,995][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.058526720851659775, 'train/grad_norm_pre_clip': 12.530527114868164, 'timing/data': 0.0699392129899934, 'timing/model': 2.0326701489975676, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 1.12}) +[2026-06-23 10:01:00,124][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.06203161180019379, 'train/grad_norm_pre_clip': 23.988903045654297, 'timing/data': 0.0011586910113692284, 'timing/model': 2.026246080000419, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 1.12}) +[2026-06-23 10:01:04,154][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.12095076590776443, 'train/grad_norm_pre_clip': 19.497825622558594, 'timing/data': 0.001246391999302432, 'timing/model': 2.0409452919848263, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 1.12}) +[2026-06-23 10:01:08,180][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.0919882208108902, 'train/grad_norm_pre_clip': 17.861055374145508, 'timing/data': 0.0012683129752986133, 'timing/model': 2.028154770989204, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 1.12}) +[2026-06-23 10:01:12,266][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.017750630155205727, 'train/grad_norm_pre_clip': 6.423496246337891, 'timing/data': 0.06521913499454968, 'timing/model': 2.026145836978685, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 1.12}) +[2026-06-23 10:01:16,366][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.1239255890250206, 'train/grad_norm_pre_clip': 28.056747436523438, 'timing/data': 0.0011194580001756549, 'timing/model': 2.0268136280064937, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.12}) +[2026-06-23 10:01:20,395][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.017251214012503624, 'train/grad_norm_pre_clip': 7.549111843109131, 'timing/data': 0.0014082720153965056, 'timing/model': 2.0330963019805495, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 1.12}) +[2026-06-23 10:01:24,438][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.091492660343647, 'train/grad_norm_pre_clip': 34.58055877685547, 'timing/data': 0.001245020015630871, 'timing/model': 2.0410895759996492, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 1.12}) +[2026-06-23 10:01:28,533][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.08333553373813629, 'train/grad_norm_pre_clip': 29.435434341430664, 'timing/data': 0.07200661502429284, 'timing/model': 2.0350460689805914, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 1.12}) +[2026-06-23 10:01:32,638][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.15260741114616394, 'train/grad_norm_pre_clip': 19.287477493286133, 'timing/data': 0.0012087840004824102, 'timing/model': 2.0274982310074847, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 1.12}) +[2026-06-23 10:01:36,898][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.10084990411996841, 'train/grad_norm_pre_clip': 19.92717170715332, 'timing/data': 0.0011570169881451875, 'timing/model': 2.026008216984337, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 1.12}) +[2026-06-23 10:01:40,952][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.09767553955316544, 'train/grad_norm_pre_clip': 23.76739501953125, 'timing/data': 0.0013489319826476276, 'timing/model': 2.052765916014323, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 1.12}) +[2026-06-23 10:01:45,031][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2900 +[2026-06-23 10:03:39,299][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2900 +[2026-06-23 10:03:39,300][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2900 +[2026-06-23 10:05:46,689][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2900 +[2026-06-23 10:05:46,691][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.1665489673614502, 'train/grad_norm_pre_clip': 22.49049186706543, 'eval/action_loss/seconds': 114.26427084300667, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04771409556269646, 'eval/latency_2/loss': 0.04771409556269646, 'eval/flappy/loss': 0.04771409556269646, 'eval/flappy/latency_2/loss': 0.04771409556269646, 'eval/flap_precision': 0.9692951849267272, 'eval/flap_recall': 0.937246963562753, 'eval/flap_f1': 0.9530017152658662, 'eval/flap_rate_model': 0.19902777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9530017152658661, 'eval/latency_2/flap_precision': 0.9692951849267272, 'eval/latency_2/flap_recall': 0.937246963562753, 'eval/latency_2/flap_f1': 0.9530017152658662, 'eval/latency_2/flap_rate_model': 0.19902777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9530017152658661, 'eval/action_classification/data_wait_seconds': 50.70091247558594, 'eval/action_classification/predict_seconds': 75.0041732788086, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06344739298219793, 'timing/model': 2.0299861889798194, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 1.12}) +[2026-06-23 10:05:50,734][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.2336103916168213, 'train/grad_norm_pre_clip': 60.438568115234375, 'timing/data': 0.0012561309849843383, 'timing/model': 1.999440274987137, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 1.12}) +[2026-06-23 10:05:54,682][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.035411279648542404, 'train/grad_norm_pre_clip': 16.360597610473633, 'timing/data': 0.0012581739865709096, 'timing/model': 1.99253628400038, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 1.12}) +[2026-06-23 10:05:58,672][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.12051023542881012, 'train/grad_norm_pre_clip': 36.0466423034668, 'timing/data': 0.0012079320149496198, 'timing/model': 2.0009674449975137, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 1.12}) +[2026-06-23 10:06:02,712][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.053360387682914734, 'train/grad_norm_pre_clip': 18.407760620117188, 'timing/data': 0.06665004100068472, 'timing/model': 2.0106093939975835, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 1.12}) +[2026-06-23 10:06:06,767][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.07962892204523087, 'train/grad_norm_pre_clip': 23.23597526550293, 'timing/data': 0.0013727069890592247, 'timing/model': 2.001874899986433, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 1.12}) +[2026-06-23 10:06:10,731][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.08758661895990372, 'train/grad_norm_pre_clip': 22.314132690429688, 'timing/data': 0.0012376469967421144, 'timing/model': 2.0037577119946945, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 1.12}) +[2026-06-23 10:06:14,699][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.2055654525756836, 'train/grad_norm_pre_clip': 31.686281204223633, 'timing/data': 0.0012506009952630848, 'timing/model': 2.005051217012806, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 1.13}) +[2026-06-23 10:06:18,736][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.0331149585545063, 'train/grad_norm_pre_clip': 7.0153045654296875, 'timing/data': 0.06144705897895619, 'timing/model': 2.0136415020097047, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 1.13}) +[2026-06-23 10:06:22,802][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.015116621740162373, 'train/grad_norm_pre_clip': 4.6408467292785645, 'timing/data': 0.0013294369855429977, 'timing/model': 2.0139453869778663, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 1.13}) +[2026-06-23 10:06:26,791][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.04465794190764427, 'train/grad_norm_pre_clip': 8.241147994995117, 'timing/data': 0.00137124300817959, 'timing/model': 2.0245391950011253, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 1.13}) +[2026-06-23 10:06:30,781][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.04326529800891876, 'train/grad_norm_pre_clip': 8.3180513381958, 'timing/data': 0.0016378760046791285, 'timing/model': 2.018546266015619, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 1.13}) +[2026-06-23 10:06:35,059][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.0527762845158577, 'train/grad_norm_pre_clip': 15.87626838684082, 'timing/data': 0.06043817498721182, 'timing/model': 2.010021619993495, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 1.13}) +[2026-06-23 10:06:39,145][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.0734425038099289, 'train/grad_norm_pre_clip': 23.08213233947754, 'timing/data': 0.0012893120001535863, 'timing/model': 2.027909674012335, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 1.13}) +[2026-06-23 10:06:43,151][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.11248718202114105, 'train/grad_norm_pre_clip': 19.621326446533203, 'timing/data': 0.0012936900020577013, 'timing/model': 2.031012750987429, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 1.13}) +[2026-06-23 10:06:47,139][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.03311363235116005, 'train/grad_norm_pre_clip': 9.189719200134277, 'timing/data': 0.0012459620193112642, 'timing/model': 2.014917446009349, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 1.13}) +[2026-06-23 10:06:51,181][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.12008145451545715, 'train/grad_norm_pre_clip': 19.23699188232422, 'timing/data': 0.05880638101371005, 'timing/model': 2.0148473309818655, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 1.13}) +[2026-06-23 10:06:55,283][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.06793506443500519, 'train/grad_norm_pre_clip': 19.840024948120117, 'timing/data': 0.0014436470228247344, 'timing/model': 2.0349353999772575, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 1.13}) +[2026-06-23 10:06:59,276][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.051602933555841446, 'train/grad_norm_pre_clip': 14.180140495300293, 'timing/data': 0.0014915450010448694, 'timing/model': 2.0176697830029298, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 1.13}) +[2026-06-23 10:07:03,276][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.07373465597629547, 'train/grad_norm_pre_clip': 14.962801933288574, 'timing/data': 0.0011871439928654581, 'timing/model': 2.0235809379955754, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 1.13}) +[2026-06-23 10:07:07,615][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.17227646708488464, 'train/grad_norm_pre_clip': 35.61738204956055, 'timing/data': 0.06027821998577565, 'timing/model': 2.0337515139835887, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 1.13}) +[2026-06-23 10:07:11,711][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.06762707233428955, 'train/grad_norm_pre_clip': 22.011577606201172, 'timing/data': 0.0012193529983051121, 'timing/model': 2.023884806985734, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 1.13}) +[2026-06-23 10:07:15,708][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.03649669885635376, 'train/grad_norm_pre_clip': 8.974556922912598, 'timing/data': 0.0017888550064526498, 'timing/model': 2.0193050490051974, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 1.13}) +[2026-06-23 10:07:19,719][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.045686352998018265, 'train/grad_norm_pre_clip': 13.461979866027832, 'timing/data': 0.0011352580040693283, 'timing/model': 2.032516683015274, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 1.13}) +[2026-06-23 10:07:23,807][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.08065165579319, 'train/grad_norm_pre_clip': 24.50514793395996, 'timing/data': 0.06850507098715752, 'timing/model': 2.0297273930045776, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 1.13}) +[2026-06-23 10:07:27,921][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.07334110885858536, 'train/grad_norm_pre_clip': 23.872068405151367, 'timing/data': 0.0012085929920431226, 'timing/model': 2.039501103980001, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 1.13}) +[2026-06-23 10:07:31,930][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.07303992658853531, 'train/grad_norm_pre_clip': 40.907806396484375, 'timing/data': 0.0012284489930607378, 'timing/model': 2.027632799989078, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 1.13}) +[2026-06-23 10:07:35,938][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.07200931012630463, 'train/grad_norm_pre_clip': 17.61947250366211, 'timing/data': 0.0011056929943151772, 'timing/model': 2.02791553499992, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 1.13}) +[2026-06-23 10:07:40,251][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.07362323254346848, 'train/grad_norm_pre_clip': 13.15201473236084, 'timing/data': 0.0693419179879129, 'timing/model': 2.264975575992139, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 1.13}) +[2026-06-23 10:07:44,365][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.0846250057220459, 'train/grad_norm_pre_clip': 15.324317932128906, 'timing/data': 0.0012393900251481682, 'timing/model': 2.0229177979927044, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 1.13}) +[2026-06-23 10:07:48,379][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.05994666740298271, 'train/grad_norm_pre_clip': 17.969440460205078, 'timing/data': 0.0013700209965463728, 'timing/model': 2.024663688003784, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 1.13}) +[2026-06-23 10:07:52,388][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.017858725041151047, 'train/grad_norm_pre_clip': 5.187312126159668, 'timing/data': 0.0014311640115920454, 'timing/model': 2.023202131007565, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 1.13}) +[2026-06-23 10:07:56,465][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.1652449369430542, 'train/grad_norm_pre_clip': 22.681833267211914, 'timing/data': 0.07616888202028349, 'timing/model': 2.022442029992817, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 1.13}) +[2026-06-23 10:08:00,567][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.12076270580291748, 'train/grad_norm_pre_clip': 26.59470558166504, 'timing/data': 0.0013374009868130088, 'timing/model': 2.0216810059791896, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 1.14}) +[2026-06-23 10:08:04,582][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.1396314948797226, 'train/grad_norm_pre_clip': 29.90130615234375, 'timing/data': 0.0012117589940316975, 'timing/model': 2.0303375229996163, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 1.14}) +[2026-06-23 10:08:08,607][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.07491512596607208, 'train/grad_norm_pre_clip': 28.2451171875, 'timing/data': 0.001241724006831646, 'timing/model': 2.0329214819939807, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 1.14}) +[2026-06-23 10:08:12,749][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.07140859216451645, 'train/grad_norm_pre_clip': 20.089357376098633, 'timing/data': 0.06139315199106932, 'timing/model': 2.096973949985113, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 1.14}) +[2026-06-23 10:08:16,853][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.22222508490085602, 'train/grad_norm_pre_clip': 25.99222755432129, 'timing/data': 0.0014175990072544664, 'timing/model': 2.0356528949923813, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 1.14}) +[2026-06-23 10:08:20,861][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.08608583360910416, 'train/grad_norm_pre_clip': 38.585792541503906, 'timing/data': 0.001492518000304699, 'timing/model': 2.021353268006351, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 1.14}) +[2026-06-23 10:08:24,886][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.0678996667265892, 'train/grad_norm_pre_clip': 10.606596946716309, 'timing/data': 0.001235221978276968, 'timing/model': 2.0295206510054413, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 1.14}) +[2026-06-23 10:08:28,964][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.1085723415017128, 'train/grad_norm_pre_clip': 24.18182945251465, 'timing/data': 0.05890956398798153, 'timing/model': 2.029546311998274, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 1.14}) +[2026-06-23 10:08:33,065][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.06123443320393562, 'train/grad_norm_pre_clip': 19.286611557006836, 'timing/data': 0.001188847003504634, 'timing/model': 2.032105322985444, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 1.14}) +[2026-06-23 10:08:37,078][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.02687869966030121, 'train/grad_norm_pre_clip': 10.77884578704834, 'timing/data': 0.0011532509815879166, 'timing/model': 2.0257582690101117, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.14}) +[2026-06-23 10:08:41,096][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.12288960069417953, 'train/grad_norm_pre_clip': 26.185937881469727, 'timing/data': 0.0013826850045006722, 'timing/model': 2.0277959869999904, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 1.14}) +[2026-06-23 10:08:45,194][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.05189766362309456, 'train/grad_norm_pre_clip': 9.578641891479492, 'timing/data': 0.07172977199661545, 'timing/model': 2.0301517929765396, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 1.14}) +[2026-06-23 10:08:49,559][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.12326427549123764, 'train/grad_norm_pre_clip': 28.63701820373535, 'timing/data': 0.0011346459796186537, 'timing/model': 2.0336438679951243, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 1.14}) +[2026-06-23 10:08:53,577][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.01616940088570118, 'train/grad_norm_pre_clip': 8.831663131713867, 'timing/data': 0.0011748899996746331, 'timing/model': 2.0336092469806317, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 1.14}) +[2026-06-23 10:08:57,605][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.02993047423660755, 'train/grad_norm_pre_clip': 10.181700706481934, 'timing/data': 0.0011859009973704815, 'timing/model': 2.0354875100019854, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 1.14}) +[2026-06-23 10:09:01,694][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.03288016840815544, 'train/grad_norm_pre_clip': 8.121577262878418, 'timing/data': 0.07774567700107582, 'timing/model': 2.0274549729947466, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 1.14}) +[2026-06-23 10:09:05,802][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.09021236002445221, 'train/grad_norm_pre_clip': 31.10692024230957, 'timing/data': 0.0012033040111418813, 'timing/model': 2.025662824016763, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 1.14}) +[2026-06-23 10:09:09,812][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.051455914974212646, 'train/grad_norm_pre_clip': 18.52572250366211, 'timing/data': 0.0012217669864185154, 'timing/model': 2.025194111018209, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 1.14}) +[2026-06-23 10:09:13,825][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.09578351676464081, 'train/grad_norm_pre_clip': 16.209440231323242, 'timing/data': 0.0013813820260111243, 'timing/model': 2.029195395996794, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 1.14}) +[2026-06-23 10:09:17,904][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.05919861048460007, 'train/grad_norm_pre_clip': 23.908660888671875, 'timing/data': 0.06347715397714637, 'timing/model': 2.0298390999960247, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 1.14}) +[2026-06-23 10:09:22,030][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.15278315544128418, 'train/grad_norm_pre_clip': 30.273813247680664, 'timing/data': 0.001261711004190147, 'timing/model': 2.0293894409842324, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 1.14}) +[2026-06-23 10:09:26,050][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.1398097574710846, 'train/grad_norm_pre_clip': 36.52149200439453, 'timing/data': 0.0013116830086801201, 'timing/model': 2.032892194984015, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 1.14}) +[2026-06-23 10:09:30,085][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.045415230095386505, 'train/grad_norm_pre_clip': 12.650525093078613, 'timing/data': 0.0011933139758184552, 'timing/model': 2.038511451013619, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 1.14}) +[2026-06-23 10:09:34,151][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.023570992052555084, 'train/grad_norm_pre_clip': 5.976811408996582, 'timing/data': 0.057231434009736404, 'timing/model': 2.023498535010731, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 1.14}) +[2026-06-23 10:09:38,272][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.04902191087603569, 'train/grad_norm_pre_clip': 30.737964630126953, 'timing/data': 0.0012832510110456496, 'timing/model': 2.0291209860006347, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 1.14}) +[2026-06-23 10:09:42,300][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.05371570959687233, 'train/grad_norm_pre_clip': 18.607446670532227, 'timing/data': 0.0013091989967506379, 'timing/model': 2.03854662401136, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 1.14}) +[2026-06-23 10:09:46,336][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.16734468936920166, 'train/grad_norm_pre_clip': 28.623998641967773, 'timing/data': 0.0013734270178247243, 'timing/model': 2.0478723870182876, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 1.15}) +[2026-06-23 10:09:50,429][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.03313853219151497, 'train/grad_norm_pre_clip': 6.980620861053467, 'timing/data': 0.0709706399939023, 'timing/model': 2.031646550982259, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 1.15}) +[2026-06-23 10:09:54,615][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.08322478085756302, 'train/grad_norm_pre_clip': 18.2901611328125, 'timing/data': 0.0011960399860981852, 'timing/model': 2.109222247992875, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 1.15}) +[2026-06-23 10:09:58,634][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.12113616615533829, 'train/grad_norm_pre_clip': 36.9684944152832, 'timing/data': 0.0012831909989472479, 'timing/model': 2.033753819996491, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 1.15}) +[2026-06-23 10:10:02,671][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.2173462063074112, 'train/grad_norm_pre_clip': 72.1838607788086, 'timing/data': 0.0015743589901831, 'timing/model': 2.043099019007059, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 1.15}) +[2026-06-23 10:10:06,772][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.0991361066699028, 'train/grad_norm_pre_clip': 22.568288803100586, 'timing/data': 0.06384020601399243, 'timing/model': 2.0416179049934726, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 1.15}) +[2026-06-23 10:10:10,875][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.1773877739906311, 'train/grad_norm_pre_clip': 21.978992462158203, 'timing/data': 0.0012244429963175207, 'timing/model': 2.028754781000316, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 1.15}) +[2026-06-23 10:10:14,894][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.12898647785186768, 'train/grad_norm_pre_clip': 24.173391342163086, 'timing/data': 0.0012657490151468664, 'timing/model': 2.029501593002351, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 1.15}) +[2026-06-23 10:10:18,909][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.06258627027273178, 'train/grad_norm_pre_clip': 13.685723304748535, 'timing/data': 0.0011810320138465613, 'timing/model': 2.0286424590158276, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 1.15}) +[2026-06-23 10:10:23,002][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.1653413325548172, 'train/grad_norm_pre_clip': 24.973257064819336, 'timing/data': 0.06886607300839387, 'timing/model': 2.039805716020055, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 1.15}) +[2026-06-23 10:10:27,098][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.030443988740444183, 'train/grad_norm_pre_clip': 9.37822437286377, 'timing/data': 0.0011230149830225855, 'timing/model': 2.02354889200069, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 1.15}) +[2026-06-23 10:10:31,318][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.11749203503131866, 'train/grad_norm_pre_clip': 12.281810760498047, 'timing/data': 0.0011739289911929518, 'timing/model': 2.023515713022789, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 1.15}) +[2026-06-23 10:10:35,325][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.08060993999242783, 'train/grad_norm_pre_clip': 14.352262496948242, 'timing/data': 0.0015682579833082855, 'timing/model': 2.024652516003698, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 1.15}) +[2026-06-23 10:10:39,401][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.05146557465195656, 'train/grad_norm_pre_clip': 9.619752883911133, 'timing/data': 0.07028639697819017, 'timing/model': 2.0244511669734493, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 1.15}) +[2026-06-23 10:10:43,496][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.044855996966362, 'train/grad_norm_pre_clip': 13.771068572998047, 'timing/data': 0.00111770597868599, 'timing/model': 2.0273311299970374, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 1.15}) +[2026-06-23 10:10:47,514][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.031115969642996788, 'train/grad_norm_pre_clip': 16.676311492919922, 'timing/data': 0.0011744099901989102, 'timing/model': 2.027590362005867, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 1.15}) +[2026-06-23 10:10:51,522][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.07020676881074905, 'train/grad_norm_pre_clip': 25.183780670166016, 'timing/data': 0.001131349999923259, 'timing/model': 2.025611006974941, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 1.15}) +[2026-06-23 10:10:55,586][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.07752441614866257, 'train/grad_norm_pre_clip': 15.925238609313965, 'timing/data': 0.05946327999117784, 'timing/model': 2.0237121699901763, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 1.15}) +[2026-06-23 10:10:59,677][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.04408740624785423, 'train/grad_norm_pre_clip': 11.835407257080078, 'timing/data': 0.001113667996833101, 'timing/model': 2.02441033301875, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 1.15}) +[2026-06-23 10:11:03,708][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.09336898475885391, 'train/grad_norm_pre_clip': 14.455034255981445, 'timing/data': 0.0011823949753306806, 'timing/model': 2.050856649992056, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 1.15}) +[2026-06-23 10:11:07,727][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.0934600979089737, 'train/grad_norm_pre_clip': 21.20421600341797, 'timing/data': 0.0012445499887689948, 'timing/model': 2.0373690109991003, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 1.15}) +[2026-06-23 10:11:11,892][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.08776883035898209, 'train/grad_norm_pre_clip': 18.135356903076172, 'timing/data': 0.06634566601132974, 'timing/model': 2.032199061010033, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 1.15}) +[2026-06-23 10:11:16,106][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.10262642055749893, 'train/grad_norm_pre_clip': 15.63533878326416, 'timing/data': 0.0012013899977318943, 'timing/model': 2.04112378397258, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 1.15}) +[2026-06-23 10:11:20,119][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.04513103887438774, 'train/grad_norm_pre_clip': 15.503959655761719, 'timing/data': 0.0011931740154977888, 'timing/model': 2.0269954709801823, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 1.15}) +[2026-06-23 10:11:24,139][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.0819031298160553, 'train/grad_norm_pre_clip': 26.222553253173828, 'timing/data': 0.0011256599973421544, 'timing/model': 2.036111608002102, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 1.15}) +[2026-06-23 10:11:28,231][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.04008742421865463, 'train/grad_norm_pre_clip': 13.436256408691406, 'timing/data': 0.08505066798534244, 'timing/model': 2.0243452489958145, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 1.15}) +[2026-06-23 10:11:32,323][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.22135736048221588, 'train/grad_norm_pre_clip': 36.70851135253906, 'timing/data': 0.0012532550026662648, 'timing/model': 2.022024883015547, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 1.16}) +[2026-06-23 10:11:36,351][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.05259186029434204, 'train/grad_norm_pre_clip': 11.520586013793945, 'timing/data': 0.0012328479788266122, 'timing/model': 2.0448801960155834, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 1.16}) +[2026-06-23 10:11:40,493][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.019336888566613197, 'train/grad_norm_pre_clip': 12.51020336151123, 'timing/data': 0.0011869330191984773, 'timing/model': 2.0250239390006755, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 1.16}) +[2026-06-23 10:11:44,567][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.12215405702590942, 'train/grad_norm_pre_clip': 22.617656707763672, 'timing/data': 0.056309512990992516, 'timing/model': 2.0256386670225766, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 1.16}) +[2026-06-23 10:11:48,679][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.04453705623745918, 'train/grad_norm_pre_clip': 12.896015167236328, 'timing/data': 0.0015102410106919706, 'timing/model': 2.0289435960003175, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 1.16}) +[2026-06-23 10:11:52,693][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.09974830597639084, 'train/grad_norm_pre_clip': 20.84624671936035, 'timing/data': 0.001378436980303377, 'timing/model': 2.0297768670134246, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 1.16}) +[2026-06-23 10:11:56,704][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.059008751064538956, 'train/grad_norm_pre_clip': 18.773597717285156, 'timing/data': 0.0012315849890001118, 'timing/model': 2.0257138419838157, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.16}) +[2026-06-23 10:12:00,775][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.04981514438986778, 'train/grad_norm_pre_clip': 11.601424217224121, 'timing/data': 0.06150502301170491, 'timing/model': 2.026947795005981, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 1.16}) +[2026-06-23 10:12:04,902][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.10176245868206024, 'train/grad_norm_pre_clip': 20.520526885986328, 'timing/data': 0.0012168889807071537, 'timing/model': 2.044982188992435, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 1.16}) +[2026-06-23 10:12:08,916][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.07104024291038513, 'train/grad_norm_pre_clip': 15.359241485595703, 'timing/data': 0.0013697910180781037, 'timing/model': 2.0269975200062618, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 1.16}) +[2026-06-23 10:12:13,172][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.0147476801648736, 'train/grad_norm_pre_clip': 6.733356952667236, 'timing/data': 0.001187994988868013, 'timing/model': 2.2712183620024007, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 1.16}) +[2026-06-23 10:12:17,265][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.04651707038283348, 'train/grad_norm_pre_clip': 18.192806243896484, 'timing/data': 0.07552185299573466, 'timing/model': 2.0313975349999964, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 1.16}) +[2026-06-23 10:12:21,368][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.052878495305776596, 'train/grad_norm_pre_clip': 131.5656280517578, 'timing/data': 0.0013481510104611516, 'timing/model': 2.0312678970221896, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 1.16}) +[2026-06-23 10:12:25,394][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.03798209875822067, 'train/grad_norm_pre_clip': 16.956933975219727, 'timing/data': 0.0012315760250203311, 'timing/model': 2.041708008997375, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 1.16}) +[2026-06-23 10:12:29,441][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.07706479728221893, 'train/grad_norm_pre_clip': 41.369659423828125, 'timing/data': 0.0014871180173940957, 'timing/model': 2.0581776829785667, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 1.16}) +[2026-06-23 10:12:33,529][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3000 +[2026-06-23 10:14:27,238][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3000 +[2026-06-23 10:14:27,239][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3000 +[2026-06-23 10:16:34,064][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3000 +[2026-06-23 10:36:48,544][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3000 +[2026-06-23 10:36:48,545][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.1284923106431961, 'train/grad_norm_pre_clip': 27.971406936645508, 'eval/action_loss/seconds': 113.70563947298797, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.047938037663698196, 'eval/latency_2/loss': 0.047938037663698196, 'eval/flappy/loss': 0.047938037663698196, 'eval/flappy/latency_2/loss': 0.047938037663698196, 'eval/flap_precision': 0.9620338983050848, 'eval/flap_recall': 0.9574898785425101, 'eval/flap_f1': 0.9597565099763272, 'eval/flap_rate_model': 0.2048611111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9597565099763271, 'eval/latency_2/flap_precision': 0.9620338983050848, 'eval/latency_2/flap_recall': 0.9574898785425101, 'eval/latency_2/flap_f1': 0.9597565099763272, 'eval/latency_2/flap_rate_model': 0.2048611111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9597565099763271, 'eval/action_classification/data_wait_seconds': 50.30683898925781, 'eval/action_classification/predict_seconds': 74.89663696289062, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 314.16000000001293, 'rl_games_eval/mid_train/mean_length': 2550.6, 'rl_games_eval/mid_train/std_reward': 168.18891283316998, 'rl_games_eval/mid_train/std_length': 1356.5594126318242, 'rl_games_eval/mid_train/macro_mean_reward': 314.16000000001293, 'rl_games_eval/mid_train/macro_mean_length': 2550.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 314.16000000001293, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2550.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 168.18891283316998, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1356.5594126318242, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 314.16000000001293, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.07120822198339738, 'timing/model': 2.027350745018339, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 1.16}) +[2026-06-23 10:36:48,546][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_3000_state +[2026-06-23 10:36:48,554][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 10:36:57,240][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_3000_state/model.safetensors +[2026-06-23 10:37:01,181][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_3000_state/optimizer.bin +[2026-06-23 10:37:01,181][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_3000_state/sampler.bin +[2026-06-23 10:37:01,181][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-23 10:37:01,182][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-23 10:37:01,314][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-23 10:37:01,314][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 10:37:01,338][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 10:37:05,306][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.06751469522714615, 'train/grad_norm_pre_clip': 23.106136322021484, 'timing/data': 0.0013426310033537447, 'timing/model': 1.9679298119735904, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 1.16}) +[2026-06-23 10:37:09,192][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.09838532656431198, 'train/grad_norm_pre_clip': 41.36412811279297, 'timing/data': 0.001286818995140493, 'timing/model': 1.9664083470124751, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 1.16}) +[2026-06-23 10:37:13,071][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.0551215223968029, 'train/grad_norm_pre_clip': 19.96552085876465, 'timing/data': 0.0011303889914415777, 'timing/model': 1.963847851002356, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 1.16}) +[2026-06-23 10:37:17,103][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.0616399347782135, 'train/grad_norm_pre_clip': 15.169055938720703, 'timing/data': 0.06166791799478233, 'timing/model': 2.053451223997399, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 1.16}) +[2026-06-23 10:37:21,090][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.08281159400939941, 'train/grad_norm_pre_clip': 22.07202911376953, 'timing/data': 0.0011228049988858402, 'timing/model': 1.9638695149915293, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 1.16}) +[2026-06-23 10:37:24,992][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.06355709582567215, 'train/grad_norm_pre_clip': 20.194826126098633, 'timing/data': 0.001233458984643221, 'timing/model': 1.9667619969986845, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 1.16}) +[2026-06-23 10:37:28,914][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.08980172872543335, 'train/grad_norm_pre_clip': 26.304851531982422, 'timing/data': 0.0011348169937264174, 'timing/model': 1.9743513900029939, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 1.16}) +[2026-06-23 10:37:32,900][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.03632368519902229, 'train/grad_norm_pre_clip': 15.599316596984863, 'timing/data': 0.06490899698110297, 'timing/model': 1.981076235009823, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 1.16}) +[2026-06-23 10:37:36,915][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.09566344320774078, 'train/grad_norm_pre_clip': 29.932815551757812, 'timing/data': 0.0013967520208097994, 'timing/model': 1.9839352359995246, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 1.16}) +[2026-06-23 10:37:40,847][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.04957751929759979, 'train/grad_norm_pre_clip': 18.798311233520508, 'timing/data': 0.001519838988315314, 'timing/model': 1.9946835070149973, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 1.16}) +[2026-06-23 10:37:44,782][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.09749103337526321, 'train/grad_norm_pre_clip': 23.909099578857422, 'timing/data': 0.0013493840233422816, 'timing/model': 1.9921837340225466, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 1.17}) +[2026-06-23 10:37:48,786][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.01651003584265709, 'train/grad_norm_pre_clip': 6.129909515380859, 'timing/data': 0.0659079129982274, 'timing/model': 1.9879519309906755, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 1.17}) +[2026-06-23 10:37:53,075][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.052307162433862686, 'train/grad_norm_pre_clip': 19.56303596496582, 'timing/data': 0.001133223995566368, 'timing/model': 2.0027522390009835, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 1.17}) +[2026-06-23 10:37:57,024][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.05487050861120224, 'train/grad_norm_pre_clip': 24.372360229492188, 'timing/data': 0.0012239619973115623, 'timing/model': 1.9936106520181056, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 1.17}) +[2026-06-23 10:38:01,008][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.008285848423838615, 'train/grad_norm_pre_clip': 2.977210521697998, 'timing/data': 0.0012313959887251258, 'timing/model': 2.0205509569786955, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 1.17}) +[2026-06-23 10:38:05,016][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.10144460946321487, 'train/grad_norm_pre_clip': 27.484525680541992, 'timing/data': 0.05834865299402736, 'timing/model': 1.9967221060069278, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 1.17}) +[2026-06-23 10:38:09,051][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.17480483651161194, 'train/grad_norm_pre_clip': 19.71964454650879, 'timing/data': 0.00139338499866426, 'timing/model': 1.9969278139760718, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 1.17}) +[2026-06-23 10:38:13,008][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.08584313839673996, 'train/grad_norm_pre_clip': 30.0502872467041, 'timing/data': 0.0012589469843078405, 'timing/model': 1.9962670700042509, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 1.17}) +[2026-06-23 10:38:16,969][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.04778296500444412, 'train/grad_norm_pre_clip': 17.990720748901367, 'timing/data': 0.0012317959917709231, 'timing/model': 2.0011577249970287, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 1.17}) +[2026-06-23 10:38:21,003][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.10014503449201584, 'train/grad_norm_pre_clip': 19.44464111328125, 'timing/data': 0.06634823599597439, 'timing/model': 2.0029971420008224, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 1.17}) +[2026-06-23 10:38:25,103][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.0434347428381443, 'train/grad_norm_pre_clip': 12.057831764221191, 'timing/data': 0.0013567869900725782, 'timing/model': 2.018852640001569, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 1.17}) +[2026-06-23 10:38:29,097][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.13455352187156677, 'train/grad_norm_pre_clip': 39.21246337890625, 'timing/data': 0.0014407819835469127, 'timing/model': 2.0221913480199873, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 1.17}) +[2026-06-23 10:38:33,069][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.04274863004684448, 'train/grad_norm_pre_clip': 16.970396041870117, 'timing/data': 0.0014712890260852873, 'timing/model': 2.0073588940140326, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 1.17}) +[2026-06-23 10:38:37,111][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.013358579948544502, 'train/grad_norm_pre_clip': 6.872419834136963, 'timing/data': 0.06800017901696265, 'timing/model': 2.0089941030018963, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 1.17}) +[2026-06-23 10:38:41,172][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.03146235644817352, 'train/grad_norm_pre_clip': 16.4925479888916, 'timing/data': 0.0012210370041429996, 'timing/model': 2.0109703720081598, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 1.17}) +[2026-06-23 10:38:45,154][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.106819286942482, 'train/grad_norm_pre_clip': 24.808483123779297, 'timing/data': 0.001265667990082875, 'timing/model': 2.0064316319767386, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 1.17}) +[2026-06-23 10:38:49,129][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.07593373209238052, 'train/grad_norm_pre_clip': 18.361698150634766, 'timing/data': 0.0014572729996871203, 'timing/model': 2.007967176992679, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 1.17}) +[2026-06-23 10:38:53,175][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.15685641765594482, 'train/grad_norm_pre_clip': 28.52499008178711, 'timing/data': 0.06861711799865589, 'timing/model': 2.0105730300128926, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 1.17}) +[2026-06-23 10:38:57,342][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.08728215843439102, 'train/grad_norm_pre_clip': 27.615968704223633, 'timing/data': 0.0014237510040402412, 'timing/model': 2.11241222298122, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 1.17}) +[2026-06-23 10:39:01,326][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.1474432647228241, 'train/grad_norm_pre_clip': 32.723934173583984, 'timing/data': 0.001444689987692982, 'timing/model': 2.0132112899736967, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 1.17}) +[2026-06-23 10:39:05,330][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.1476476490497589, 'train/grad_norm_pre_clip': 23.78723907470703, 'timing/data': 0.0012724410044029355, 'timing/model': 2.020610873005353, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 1.17}) +[2026-06-23 10:39:09,400][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.09304861724376678, 'train/grad_norm_pre_clip': 22.650842666625977, 'timing/data': 0.05994773600832559, 'timing/model': 2.028195366001455, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 1.17}) +[2026-06-23 10:39:13,474][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.13732868432998657, 'train/grad_norm_pre_clip': 27.047325134277344, 'timing/data': 0.001449930015951395, 'timing/model': 2.015708597988123, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 1.17}) +[2026-06-23 10:39:17,476][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.11132808029651642, 'train/grad_norm_pre_clip': 20.819555282592773, 'timing/data': 0.0012621519854292274, 'timing/model': 2.0312859030091204, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 1.17}) +[2026-06-23 10:39:21,487][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.06410406529903412, 'train/grad_norm_pre_clip': 14.859798431396484, 'timing/data': 0.001191361981909722, 'timing/model': 2.0339824039838277, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 1.17}) +[2026-06-23 10:39:25,548][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.27373993396759033, 'train/grad_norm_pre_clip': 59.14572525024414, 'timing/data': 0.05996861500898376, 'timing/model': 2.0182985560095403, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 1.17}) +[2026-06-23 10:39:29,633][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.04669172689318657, 'train/grad_norm_pre_clip': 13.324602127075195, 'timing/data': 0.0011215730046387762, 'timing/model': 2.0163598410144914, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 1.18}) +[2026-06-23 10:39:33,829][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.11476585268974304, 'train/grad_norm_pre_clip': 30.819435119628906, 'timing/data': 0.0012079020089004189, 'timing/model': 2.017811931989854, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 1.18}) +[2026-06-23 10:39:37,834][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.11870412528514862, 'train/grad_norm_pre_clip': 20.620811462402344, 'timing/data': 0.0011772949947044253, 'timing/model': 2.02643387299031, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 1.18}) +[2026-06-23 10:39:41,898][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.06822193413972855, 'train/grad_norm_pre_clip': 23.141647338867188, 'timing/data': 0.06796994101023301, 'timing/model': 2.0199018100101966, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.18}) +[2026-06-23 10:39:45,979][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.01728150062263012, 'train/grad_norm_pre_clip': 16.230161666870117, 'timing/data': 0.0013094999885652214, 'timing/model': 2.017519885004731, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 1.18}) +[2026-06-23 10:39:49,984][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.022509826347231865, 'train/grad_norm_pre_clip': 8.113853454589844, 'timing/data': 0.0013852999836672097, 'timing/model': 2.0275393679912668, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 1.18}) +[2026-06-23 10:39:54,012][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.06415629386901855, 'train/grad_norm_pre_clip': 15.019043922424316, 'timing/data': 0.0012468639761209488, 'timing/model': 2.0474215100111905, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 1.18}) +[2026-06-23 10:39:58,074][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.043009642511606216, 'train/grad_norm_pre_clip': 11.56921100616455, 'timing/data': 0.05878099999972619, 'timing/model': 2.0256230930099264, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 1.18}) +[2026-06-23 10:40:02,164][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.02720492333173752, 'train/grad_norm_pre_clip': 14.135478973388672, 'timing/data': 0.0012338200176600367, 'timing/model': 2.029418785008602, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 1.18}) +[2026-06-23 10:40:06,226][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.10428360104560852, 'train/grad_norm_pre_clip': 16.715715408325195, 'timing/data': 0.0012224389938637614, 'timing/model': 2.0738543909974396, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 1.18}) +[2026-06-23 10:40:10,268][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.0220822561532259, 'train/grad_norm_pre_clip': 9.311285018920898, 'timing/data': 0.0013691200001630932, 'timing/model': 2.065286645985907, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 1.18}) +[2026-06-23 10:40:14,347][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.11687722057104111, 'train/grad_norm_pre_clip': 19.82962417602539, 'timing/data': 0.07236562401521951, 'timing/model': 2.021900834981352, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 1.18}) +[2026-06-23 10:40:18,441][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.14738242328166962, 'train/grad_norm_pre_clip': 25.0533447265625, 'timing/data': 0.0011121459829155356, 'timing/model': 2.0342691889964044, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 1.18}) +[2026-06-23 10:40:22,468][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.018418077379465103, 'train/grad_norm_pre_clip': 9.255087852478027, 'timing/data': 0.0016188320005312562, 'timing/model': 2.0414418219879735, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 1.18}) +[2026-06-23 10:40:26,483][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.0490964911878109, 'train/grad_norm_pre_clip': 15.418476104736328, 'timing/data': 0.0011315309966448694, 'timing/model': 2.0289646619930863, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 1.18}) +[2026-06-23 10:40:30,553][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.09111320972442627, 'train/grad_norm_pre_clip': 25.004850387573242, 'timing/data': 0.061443269019946456, 'timing/model': 2.026078224997036, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 1.18}) +[2026-06-23 10:40:34,651][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.05510800704360008, 'train/grad_norm_pre_clip': 13.251416206359863, 'timing/data': 0.001198685000417754, 'timing/model': 2.0282975339796394, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 1.18}) +[2026-06-23 10:40:38,653][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.021524950861930847, 'train/grad_norm_pre_clip': 8.64815902709961, 'timing/data': 0.001412971003446728, 'timing/model': 2.0240587890148163, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 1.18}) +[2026-06-23 10:40:42,794][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.10338082164525986, 'train/grad_norm_pre_clip': 17.790939331054688, 'timing/data': 0.0012105570058338344, 'timing/model': 2.0257938030117657, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 1.18}) +[2026-06-23 10:40:46,864][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.061065685003995895, 'train/grad_norm_pre_clip': 20.023582458496094, 'timing/data': 0.05886993298190646, 'timing/model': 2.0280868179979734, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 1.18}) +[2026-06-23 10:40:50,964][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.09743516147136688, 'train/grad_norm_pre_clip': 16.938304901123047, 'timing/data': 0.0012667010014411062, 'timing/model': 2.0279090680123772, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 1.18}) +[2026-06-23 10:40:54,975][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.06520990282297134, 'train/grad_norm_pre_clip': 20.601593017578125, 'timing/data': 0.0012792439956683666, 'timing/model': 2.024955297005363, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 1.18}) +[2026-06-23 10:40:58,988][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.0770784243941307, 'train/grad_norm_pre_clip': 16.521020889282227, 'timing/data': 0.001285895996261388, 'timing/model': 2.03083758498542, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 1.18}) +[2026-06-23 10:41:03,055][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.08311355113983154, 'train/grad_norm_pre_clip': 28.55654525756836, 'timing/data': 0.05758270400110632, 'timing/model': 2.0245552730048075, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 1.18}) +[2026-06-23 10:41:07,151][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.11040513217449188, 'train/grad_norm_pre_clip': 25.322649002075195, 'timing/data': 0.0013988350110594183, 'timing/model': 2.027523436001502, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 1.18}) +[2026-06-23 10:41:11,167][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.09397252649068832, 'train/grad_norm_pre_clip': 22.0235652923584, 'timing/data': 0.0017201089940499514, 'timing/model': 2.029507060011383, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 1.19}) +[2026-06-23 10:41:15,446][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.14789873361587524, 'train/grad_norm_pre_clip': 23.656204223632812, 'timing/data': 0.0011886770080309361, 'timing/model': 2.2918843269872013, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 1.19}) +[2026-06-23 10:41:19,510][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.021301422268152237, 'train/grad_norm_pre_clip': 8.520008087158203, 'timing/data': 0.058274262002669275, 'timing/model': 2.0266221550118644, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 1.19}) +[2026-06-23 10:41:23,605][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.06208221986889839, 'train/grad_norm_pre_clip': 18.777374267578125, 'timing/data': 0.001189517992315814, 'timing/model': 2.031472937000217, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 1.19}) +[2026-06-23 10:41:27,651][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.05324624106287956, 'train/grad_norm_pre_clip': 15.614809036254883, 'timing/data': 0.0012178299948573112, 'timing/model': 2.035311219020514, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 1.19}) +[2026-06-23 10:41:31,676][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.09383709728717804, 'train/grad_norm_pre_clip': 18.67786979675293, 'timing/data': 0.0012465530016925186, 'timing/model': 2.0300679270003457, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 1.19}) +[2026-06-23 10:41:35,736][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.05509129911661148, 'train/grad_norm_pre_clip': 11.271815299987793, 'timing/data': 0.05326105700805783, 'timing/model': 2.0248846450122073, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 1.19}) +[2026-06-23 10:41:39,828][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.08689548075199127, 'train/grad_norm_pre_clip': 21.46036720275879, 'timing/data': 0.0012902040034532547, 'timing/model': 2.0258888990210835, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 1.19}) +[2026-06-23 10:41:43,833][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.13921837508678436, 'train/grad_norm_pre_clip': 22.9182186126709, 'timing/data': 0.0012450009817257524, 'timing/model': 2.0230364560266025, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 1.19}) +[2026-06-23 10:41:48,085][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.021014735102653503, 'train/grad_norm_pre_clip': 8.460000991821289, 'timing/data': 0.0014564110024366528, 'timing/model': 2.268628952006111, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 1.19}) +[2026-06-23 10:41:52,158][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.055777136236429214, 'train/grad_norm_pre_clip': 13.181758880615234, 'timing/data': 0.06780656799674034, 'timing/model': 2.0234642830037046, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 1.19}) +[2026-06-23 10:41:56,263][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.04307718947529793, 'train/grad_norm_pre_clip': 11.923667907714844, 'timing/data': 0.0011994370142929256, 'timing/model': 2.037973864993546, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 1.19}) +[2026-06-23 10:42:00,270][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.08310989290475845, 'train/grad_norm_pre_clip': 27.321685791015625, 'timing/data': 0.001609664992429316, 'timing/model': 2.025323073990876, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 1.19}) +[2026-06-23 10:42:04,282][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.006612068973481655, 'train/grad_norm_pre_clip': 4.360006809234619, 'timing/data': 0.0013532399898394942, 'timing/model': 2.029742188024102, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 1.19}) +[2026-06-23 10:42:08,367][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.03665967285633087, 'train/grad_norm_pre_clip': 16.456117630004883, 'timing/data': 0.06725661098607816, 'timing/model': 2.0270863290061243, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 1.19}) +[2026-06-23 10:42:12,481][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.06244516000151634, 'train/grad_norm_pre_clip': 19.412033081054688, 'timing/data': 0.0012374059879221022, 'timing/model': 2.0436850510013755, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 1.19}) +[2026-06-23 10:42:16,488][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.09044595062732697, 'train/grad_norm_pre_clip': 23.3460750579834, 'timing/data': 0.0013901790080126375, 'timing/model': 2.025863485003356, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 1.19}) +[2026-06-23 10:42:20,540][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.02339336834847927, 'train/grad_norm_pre_clip': 11.9635648727417, 'timing/data': 0.0012375559890642762, 'timing/model': 2.0669899949862156, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 1.19}) +[2026-06-23 10:42:24,624][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.03258276358246803, 'train/grad_norm_pre_clip': 10.4434175491333, 'timing/data': 0.06747102699591778, 'timing/model': 2.0301364779879805, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 1.19}) +[2026-06-23 10:42:28,749][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.01408402156084776, 'train/grad_norm_pre_clip': 6.285928249359131, 'timing/data': 0.0014044749841559678, 'timing/model': 2.0527370320050977, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 1.19}) +[2026-06-23 10:42:32,766][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.11788051575422287, 'train/grad_norm_pre_clip': 22.958877563476562, 'timing/data': 0.0014278680027928203, 'timing/model': 2.029297283006599, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 1.19}) +[2026-06-23 10:42:36,798][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.03998187184333801, 'train/grad_norm_pre_clip': 17.723962783813477, 'timing/data': 0.001330879982560873, 'timing/model': 2.044940290012164, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 1.19}) +[2026-06-23 10:42:40,897][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.06872760504484177, 'train/grad_norm_pre_clip': 23.962928771972656, 'timing/data': 0.07127395400311798, 'timing/model': 2.0294525070057716, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 1.19}) +[2026-06-23 10:42:45,001][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.06647948920726776, 'train/grad_norm_pre_clip': 26.368408203125, 'timing/data': 0.0012656690087169409, 'timing/model': 2.035675655992236, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 1.19}) +[2026-06-23 10:42:49,032][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.04903123527765274, 'train/grad_norm_pre_clip': 11.730566024780273, 'timing/data': 0.001447454997105524, 'timing/model': 2.040449545980664, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 1.19}) +[2026-06-23 10:42:53,050][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.12250418215990067, 'train/grad_norm_pre_clip': 27.37703514099121, 'timing/data': 0.0012800560216419399, 'timing/model': 2.0290692029811908, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 1.19}) +[2026-06-23 10:42:57,244][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.010966264642775059, 'train/grad_norm_pre_clip': 3.9494004249572754, 'timing/data': 0.05848861701088026, 'timing/model': 2.027593614999205, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 1.2}) +[2026-06-23 10:43:01,344][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.1241130381822586, 'train/grad_norm_pre_clip': 22.157691955566406, 'timing/data': 0.0012849839986301959, 'timing/model': 2.0304046280216426, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.2}) +[2026-06-23 10:43:05,361][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.06338082253932953, 'train/grad_norm_pre_clip': 17.84377098083496, 'timing/data': 0.001283430989133194, 'timing/model': 2.031396959995618, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 1.2}) +[2026-06-23 10:43:09,385][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.17561382055282593, 'train/grad_norm_pre_clip': 55.04426574707031, 'timing/data': 0.0012154169962741435, 'timing/model': 2.034470988990506, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 1.2}) +[2026-06-23 10:43:13,481][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.018408501520752907, 'train/grad_norm_pre_clip': 7.654933929443359, 'timing/data': 0.05854392901528627, 'timing/model': 2.0374319099937566, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 1.2}) +[2026-06-23 10:43:17,594][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.018368346616625786, 'train/grad_norm_pre_clip': 10.511273384094238, 'timing/data': 0.0013503049849532545, 'timing/model': 2.0306507560017053, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 1.2}) +[2026-06-23 10:43:21,617][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.036406584084033966, 'train/grad_norm_pre_clip': 15.469371795654297, 'timing/data': 0.0012418749975040555, 'timing/model': 2.034257549996255, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 1.2}) +[2026-06-23 10:43:25,640][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.04946392402052879, 'train/grad_norm_pre_clip': 15.353629112243652, 'timing/data': 0.0012315359781496227, 'timing/model': 2.036921221006196, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 1.2}) +[2026-06-23 10:43:29,949][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.11617352813482285, 'train/grad_norm_pre_clip': 29.752361297607422, 'timing/data': 0.06446997000603005, 'timing/model': 2.255214992997935, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 1.2}) +[2026-06-23 10:43:34,059][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.048126280307769775, 'train/grad_norm_pre_clip': 15.126200675964355, 'timing/data': 0.0011810930154751986, 'timing/model': 2.0261197469953913, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 1.2}) +[2026-06-23 10:43:38,064][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.15780574083328247, 'train/grad_norm_pre_clip': 37.79574966430664, 'timing/data': 0.00126952602295205, 'timing/model': 2.0246824479836505, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 1.2}) +[2026-06-23 10:43:42,074][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.06172642856836319, 'train/grad_norm_pre_clip': 22.772756576538086, 'timing/data': 0.0012456620170269161, 'timing/model': 2.0259320559853222, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 1.2}) +[2026-06-23 10:43:46,162][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3100 +[2026-06-23 10:45:40,667][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3100 +[2026-06-23 10:45:40,667][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3100 +[2026-06-23 10:47:47,842][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3100 +[2026-06-23 10:47:47,844][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.0167619027197361, 'train/grad_norm_pre_clip': 11.493402481079102, 'eval/action_loss/seconds': 114.50144204299431, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.05708307772874832, 'eval/latency_2/loss': 0.05708307772874832, 'eval/flappy/loss': 0.05708307772874832, 'eval/flappy/latency_2/loss': 0.05708307772874832, 'eval/flap_precision': 0.9284802043422733, 'eval/flap_recall': 0.9811066126855601, 'eval/flap_f1': 0.9540682414698163, 'eval/flap_rate_model': 0.2175, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9540682414698164, 'eval/latency_2/flap_precision': 0.9284802043422733, 'eval/latency_2/flap_recall': 0.9811066126855601, 'eval/latency_2/flap_f1': 0.9540682414698163, 'eval/latency_2/flap_rate_model': 0.2175, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9540682414698164, 'eval/action_classification/data_wait_seconds': 50.535648345947266, 'eval/action_classification/predict_seconds': 74.9544677734375, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.07100702100433409, 'timing/model': 2.021841736015631, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 1.2}) +[2026-06-23 10:47:51,902][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.014776485040783882, 'train/grad_norm_pre_clip': 5.526863098144531, 'timing/data': 0.0014011090097483248, 'timing/model': 2.0199611069983803, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 1.2}) +[2026-06-23 10:47:55,914][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.22117626667022705, 'train/grad_norm_pre_clip': 39.654930114746094, 'timing/data': 0.0015932240057736635, 'timing/model': 2.0167268800141755, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 1.2}) +[2026-06-23 10:47:59,870][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.13571599125862122, 'train/grad_norm_pre_clip': 45.68309783935547, 'timing/data': 0.0012262059899512678, 'timing/model': 1.9996781029913109, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 1.2}) +[2026-06-23 10:48:03,910][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.10384384542703629, 'train/grad_norm_pre_clip': 31.090810775756836, 'timing/data': 0.07272731699049473, 'timing/model': 2.0083694529894274, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 1.2}) +[2026-06-23 10:48:07,951][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.07115437835454941, 'train/grad_norm_pre_clip': 51.02821731567383, 'timing/data': 0.001178428006824106, 'timing/model': 1.9999006429861765, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 1.2}) +[2026-06-23 10:48:11,909][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.11700576543807983, 'train/grad_norm_pre_clip': 19.945005416870117, 'timing/data': 0.0012213570007588714, 'timing/model': 2.0006979450117797, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 1.2}) +[2026-06-23 10:48:15,871][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.026128189638257027, 'train/grad_norm_pre_clip': 12.233352661132812, 'timing/data': 0.0013624670100398362, 'timing/model': 2.002489314007107, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 1.2}) +[2026-06-23 10:48:19,915][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.00830609817057848, 'train/grad_norm_pre_clip': 6.112626552581787, 'timing/data': 0.06790484397788532, 'timing/model': 2.0123357599950396, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 1.2}) +[2026-06-23 10:48:23,984][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.14304307103157043, 'train/grad_norm_pre_clip': 27.060230255126953, 'timing/data': 0.0011987359903287143, 'timing/model': 2.0170693960099015, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 1.2}) +[2026-06-23 10:48:27,961][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.11962790042161942, 'train/grad_norm_pre_clip': 37.99650955200195, 'timing/data': 0.0014067390002310276, 'timing/model': 2.006922783999471, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 1.2}) +[2026-06-23 10:48:32,118][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.0777604728937149, 'train/grad_norm_pre_clip': 30.169452667236328, 'timing/data': 0.0011848000285681337, 'timing/model': 2.006841092981631, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 1.2}) +[2026-06-23 10:48:36,176][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.02795446291565895, 'train/grad_norm_pre_clip': 19.450027465820312, 'timing/data': 0.06302910501835868, 'timing/model': 2.019126966013573, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 1.2}) +[2026-06-23 10:48:40,247][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.05414378270506859, 'train/grad_norm_pre_clip': 15.343222618103027, 'timing/data': 0.001255590992514044, 'timing/model': 2.0156959139858373, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 1.2}) +[2026-06-23 10:48:44,248][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.11752506345510483, 'train/grad_norm_pre_clip': 22.722965240478516, 'timing/data': 0.001385630021104589, 'timing/model': 2.0144305949797854, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 1.21}) +[2026-06-23 10:48:48,246][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.2526191473007202, 'train/grad_norm_pre_clip': 46.63127899169922, 'timing/data': 0.0011274730204604566, 'timing/model': 2.0283263530000113, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 1.21}) +[2026-06-23 10:48:52,287][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.05046843737363815, 'train/grad_norm_pre_clip': 15.957135200500488, 'timing/data': 0.055276297993259504, 'timing/model': 2.013560986990342, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 1.21}) +[2026-06-23 10:48:56,353][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.07828257977962494, 'train/grad_norm_pre_clip': 24.53678321838379, 'timing/data': 0.0011169839999638498, 'timing/model': 2.0141935349965934, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 1.21}) +[2026-06-23 10:49:00,343][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.05413462966680527, 'train/grad_norm_pre_clip': 20.347820281982422, 'timing/data': 0.0014164580206852406, 'timing/model': 2.017479435977293, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 1.21}) +[2026-06-23 10:49:04,384][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.06693604588508606, 'train/grad_norm_pre_clip': 13.908488273620605, 'timing/data': 0.0013988960126880556, 'timing/model': 2.035049293015618, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 1.21}) +[2026-06-23 10:49:08,447][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.17518658936023712, 'train/grad_norm_pre_clip': 29.025541305541992, 'timing/data': 0.06592299201292917, 'timing/model': 2.0192751029971987, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 1.21}) +[2026-06-23 10:49:12,537][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.023276159539818764, 'train/grad_norm_pre_clip': 9.7733793258667, 'timing/data': 0.0012196839961688966, 'timing/model': 2.0232312330044806, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 1.21}) +[2026-06-23 10:49:16,549][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.08432894200086594, 'train/grad_norm_pre_clip': 31.158254623413086, 'timing/data': 0.001489082002080977, 'timing/model': 2.034642468992388, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 1.21}) +[2026-06-23 10:49:20,563][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.05835767835378647, 'train/grad_norm_pre_clip': 20.70856285095215, 'timing/data': 0.001289592997636646, 'timing/model': 2.028982838994125, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 1.21}) +[2026-06-23 10:49:24,613][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.052716270089149475, 'train/grad_norm_pre_clip': 26.350372314453125, 'timing/data': 0.053249121992848814, 'timing/model': 2.0198666829965077, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 1.21}) +[2026-06-23 10:49:28,702][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.1747957020998001, 'train/grad_norm_pre_clip': 46.556121826171875, 'timing/data': 0.0011987750185653567, 'timing/model': 2.020609714993043, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 1.21}) +[2026-06-23 10:49:32,710][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.018655596300959587, 'train/grad_norm_pre_clip': 8.007749557495117, 'timing/data': 0.0013600130041595548, 'timing/model': 2.034153723012423, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 1.21}) +[2026-06-23 10:49:36,740][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.13812464475631714, 'train/grad_norm_pre_clip': 21.18873405456543, 'timing/data': 0.001114089012844488, 'timing/model': 2.0515426860074513, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 1.21}) +[2026-06-23 10:49:40,798][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.07609955221414566, 'train/grad_norm_pre_clip': 25.41673469543457, 'timing/data': 0.06219676497858018, 'timing/model': 2.0206708460173104, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 1.21}) +[2026-06-23 10:49:44,880][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.06126198545098305, 'train/grad_norm_pre_clip': 53.7352409362793, 'timing/data': 0.0012623929942492396, 'timing/model': 2.017146792000858, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 1.21}) +[2026-06-23 10:49:48,876][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.0335460901260376, 'train/grad_norm_pre_clip': 16.988998413085938, 'timing/data': 0.001174089004052803, 'timing/model': 2.0197182430129033, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 1.21}) +[2026-06-23 10:49:52,885][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.13081218302249908, 'train/grad_norm_pre_clip': 22.550111770629883, 'timing/data': 0.0011208420037291944, 'timing/model': 2.03058010499808, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 1.21}) +[2026-06-23 10:49:56,947][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.011808054521679878, 'train/grad_norm_pre_clip': 9.830904006958008, 'timing/data': 0.06098464599926956, 'timing/model': 2.0237406649976037, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 1.21}) +[2026-06-23 10:50:01,056][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.03228488191962242, 'train/grad_norm_pre_clip': 16.891481399536133, 'timing/data': 0.00132370600476861, 'timing/model': 2.039621973002795, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 1.21}) +[2026-06-23 10:50:05,057][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.022645549848675728, 'train/grad_norm_pre_clip': 16.16067886352539, 'timing/data': 0.0011498749954625964, 'timing/model': 2.0221718949906062, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 1.21}) +[2026-06-23 10:50:09,060][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.15924005210399628, 'train/grad_norm_pre_clip': 56.29389953613281, 'timing/data': 0.0011650730157271028, 'timing/model': 2.023553497012472, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 1.21}) +[2026-06-23 10:50:13,257][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.02070162445306778, 'train/grad_norm_pre_clip': 7.686776638031006, 'timing/data': 0.05735310399904847, 'timing/model': 2.028336996998405, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 1.21}) +[2026-06-23 10:50:17,354][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.06090985983610153, 'train/grad_norm_pre_clip': 16.79192352294922, 'timing/data': 0.0014440779923461378, 'timing/model': 2.0275743460224476, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 1.21}) +[2026-06-23 10:50:21,385][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.10044825822114944, 'train/grad_norm_pre_clip': 22.721818923950195, 'timing/data': 0.0017674360133241862, 'timing/model': 2.04294661199674, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.21}) +[2026-06-23 10:50:25,726][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.06951861828565598, 'train/grad_norm_pre_clip': 16.456880569458008, 'timing/data': 0.001308538019657135, 'timing/model': 2.0373429559986107, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 1.21}) +[2026-06-23 10:50:29,911][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.0377609059214592, 'train/grad_norm_pre_clip': 14.731762886047363, 'timing/data': 0.11895284699858166, 'timing/model': 2.024319579009898, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 1.22}) +[2026-06-23 10:50:34,022][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.08111438155174255, 'train/grad_norm_pre_clip': 16.80011749267578, 'timing/data': 0.0013452460116241127, 'timing/model': 2.0370158609875944, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 1.22}) +[2026-06-23 10:50:38,061][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.07760414481163025, 'train/grad_norm_pre_clip': 24.647796630859375, 'timing/data': 0.001168860006146133, 'timing/model': 2.045382160024019, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 1.22}) +[2026-06-23 10:50:42,072][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.15583014488220215, 'train/grad_norm_pre_clip': 30.465984344482422, 'timing/data': 0.001242925995029509, 'timing/model': 2.0262348060205113, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 1.22}) +[2026-06-23 10:50:46,376][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.04042724147439003, 'train/grad_norm_pre_clip': 9.317584991455078, 'timing/data': 0.05678099702345207, 'timing/model': 2.2667469160223845, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 1.22}) +[2026-06-23 10:50:50,478][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.11057360470294952, 'train/grad_norm_pre_clip': 23.523237228393555, 'timing/data': 0.0013660939875990152, 'timing/model': 2.03530003700871, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 1.22}) +[2026-06-23 10:50:54,471][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.09076105058193207, 'train/grad_norm_pre_clip': 18.04655647277832, 'timing/data': 0.001147611008491367, 'timing/model': 2.0174245640228037, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 1.22}) +[2026-06-23 10:50:58,474][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.07238064706325531, 'train/grad_norm_pre_clip': 23.037460327148438, 'timing/data': 0.0012446100008673966, 'timing/model': 2.025450143002672, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 1.22}) +[2026-06-23 10:51:02,534][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.04581107199192047, 'train/grad_norm_pre_clip': 13.40585994720459, 'timing/data': 0.057449763000477105, 'timing/model': 2.023633837001398, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 1.22}) +[2026-06-23 10:51:06,609][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.06805720180273056, 'train/grad_norm_pre_clip': 14.269362449645996, 'timing/data': 0.0011062539997510612, 'timing/model': 2.0223559850128368, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 1.22}) +[2026-06-23 10:51:10,613][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.012188564985990524, 'train/grad_norm_pre_clip': 4.805490493774414, 'timing/data': 0.0011760130000766367, 'timing/model': 2.0236236170167103, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 1.22}) +[2026-06-23 10:51:14,624][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.02639913558959961, 'train/grad_norm_pre_clip': 10.624336242675781, 'timing/data': 0.00119580898899585, 'timing/model': 2.029627399984747, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 1.22}) +[2026-06-23 10:51:18,736][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.019161667674779892, 'train/grad_norm_pre_clip': 6.44583797454834, 'timing/data': 0.05840370600344613, 'timing/model': 2.0566087600018363, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 1.22}) +[2026-06-23 10:51:22,836][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.08801940083503723, 'train/grad_norm_pre_clip': 21.076696395874023, 'timing/data': 0.0013650530017912388, 'timing/model': 2.0275077820115257, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 1.22}) +[2026-06-23 10:51:26,853][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.04590284079313278, 'train/grad_norm_pre_clip': 32.23719024658203, 'timing/data': 0.001312645006692037, 'timing/model': 2.028587846987648, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 1.22}) +[2026-06-23 10:51:30,890][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.03326977416872978, 'train/grad_norm_pre_clip': 11.057096481323242, 'timing/data': 0.0014097050006967038, 'timing/model': 2.0448008380190004, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 1.22}) +[2026-06-23 10:51:34,974][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.1767289787530899, 'train/grad_norm_pre_clip': 53.869842529296875, 'timing/data': 0.06756619797670282, 'timing/model': 2.0313385690096766, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 1.22}) +[2026-06-23 10:51:39,078][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.06194022297859192, 'train/grad_norm_pre_clip': 16.48161506652832, 'timing/data': 0.0012444100284483284, 'timing/model': 2.033748732996173, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 1.22}) +[2026-06-23 10:51:43,111][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.011615152470767498, 'train/grad_norm_pre_clip': 7.098421096801758, 'timing/data': 0.0013897980097681284, 'timing/model': 2.04528925800696, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 1.22}) +[2026-06-23 10:51:47,121][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.09412485361099243, 'train/grad_norm_pre_clip': 25.991477966308594, 'timing/data': 0.0011935060028918087, 'timing/model': 2.027214316010941, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 1.22}) +[2026-06-23 10:51:51,185][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.1286674588918686, 'train/grad_norm_pre_clip': 23.00958251953125, 'timing/data': 0.05804660602007061, 'timing/model': 2.023884693015134, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 1.22}) +[2026-06-23 10:51:55,470][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.0033278174232691526, 'train/grad_norm_pre_clip': 2.935289144515991, 'timing/data': 0.0014858150098007172, 'timing/model': 2.0250921240076423, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 1.22}) +[2026-06-23 10:51:59,483][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.08704657852649689, 'train/grad_norm_pre_clip': 21.101627349853516, 'timing/data': 0.0011933959904126823, 'timing/model': 2.030092433007667, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 1.22}) +[2026-06-23 10:52:03,494][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.032652612775564194, 'train/grad_norm_pre_clip': 20.972368240356445, 'timing/data': 0.0011176049883943051, 'timing/model': 2.026673093001591, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 1.22}) +[2026-06-23 10:52:07,563][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.027048416435718536, 'train/grad_norm_pre_clip': 7.814891815185547, 'timing/data': 0.057719100994290784, 'timing/model': 2.0282614070165437, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 1.22}) +[2026-06-23 10:52:11,665][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.04086408391594887, 'train/grad_norm_pre_clip': 17.93042755126953, 'timing/data': 0.0011192279926035553, 'timing/model': 2.029104205983458, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 1.22}) +[2026-06-23 10:52:15,678][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.03542076051235199, 'train/grad_norm_pre_clip': 13.630643844604492, 'timing/data': 0.0011543830041773617, 'timing/model': 2.0286630890041124, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 1.23}) +[2026-06-23 10:52:19,692][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.06059560552239418, 'train/grad_norm_pre_clip': 18.501007080078125, 'timing/data': 0.0011270420218352228, 'timing/model': 2.028454633022193, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 1.23}) +[2026-06-23 10:52:23,760][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.15670092403888702, 'train/grad_norm_pre_clip': 110.13632202148438, 'timing/data': 0.05830798900569789, 'timing/model': 2.0270178580249194, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 1.23}) +[2026-06-23 10:52:27,900][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.16443784534931183, 'train/grad_norm_pre_clip': 43.60594177246094, 'timing/data': 0.0011370909924153239, 'timing/model': 2.0754276850202587, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 1.23}) +[2026-06-23 10:52:31,915][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.15541326999664307, 'train/grad_norm_pre_clip': 36.07661819458008, 'timing/data': 0.0011609149805735797, 'timing/model': 2.0296910160104744, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 1.23}) +[2026-06-23 10:52:35,929][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.054574742913246155, 'train/grad_norm_pre_clip': 30.69211196899414, 'timing/data': 0.0012778110103681684, 'timing/model': 2.028647628001636, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 1.23}) +[2026-06-23 10:52:40,004][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.07245045155286789, 'train/grad_norm_pre_clip': 39.40685272216797, 'timing/data': 0.06442620500456542, 'timing/model': 2.027370427007554, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 1.23}) +[2026-06-23 10:52:44,114][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.22311623394489288, 'train/grad_norm_pre_clip': 35.273406982421875, 'timing/data': 0.0011067460000049323, 'timing/model': 2.0429265330021735, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 1.23}) +[2026-06-23 10:52:48,127][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.15540610253810883, 'train/grad_norm_pre_clip': 36.53694534301758, 'timing/data': 0.001165533991297707, 'timing/model': 2.027604089002125, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 1.23}) +[2026-06-23 10:52:52,134][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.10726398229598999, 'train/grad_norm_pre_clip': 29.737659454345703, 'timing/data': 0.0011559959966689348, 'timing/model': 2.024042715987889, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 1.23}) +[2026-06-23 10:52:56,210][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.0665111169219017, 'train/grad_norm_pre_clip': 27.187816619873047, 'timing/data': 0.06631980399833992, 'timing/model': 2.027623682981357, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 1.23}) +[2026-06-23 10:53:00,300][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.09904520213603973, 'train/grad_norm_pre_clip': 13.452144622802734, 'timing/data': 0.0011366200051270425, 'timing/model': 2.02332205302082, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 1.23}) +[2026-06-23 10:53:04,501][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.040555790066719055, 'train/grad_norm_pre_clip': 18.088743209838867, 'timing/data': 0.0013411789841484278, 'timing/model': 2.216447834012797, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 1.23}) +[2026-06-23 10:53:08,508][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.053260624408721924, 'train/grad_norm_pre_clip': 29.31798553466797, 'timing/data': 0.0014623019960708916, 'timing/model': 2.0228532059991267, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 1.23}) +[2026-06-23 10:53:12,573][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.038255758583545685, 'train/grad_norm_pre_clip': 25.09701156616211, 'timing/data': 0.056885278987465426, 'timing/model': 2.027399266982684, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 1.23}) +[2026-06-23 10:53:16,668][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.15609152615070343, 'train/grad_norm_pre_clip': 40.601905822753906, 'timing/data': 0.0011315309966448694, 'timing/model': 2.028121261013439, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 1.23}) +[2026-06-23 10:53:20,681][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.06991542130708694, 'train/grad_norm_pre_clip': 25.138805389404297, 'timing/data': 0.0012420149869285524, 'timing/model': 2.0278293599840254, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 1.23}) +[2026-06-23 10:53:24,716][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.07970061898231506, 'train/grad_norm_pre_clip': 25.764514923095703, 'timing/data': 0.0012679730134550482, 'timing/model': 2.046298739995109, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 1.23}) +[2026-06-23 10:53:28,802][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.18802782893180847, 'train/grad_norm_pre_clip': 19.353057861328125, 'timing/data': 0.06913100901874714, 'timing/model': 2.0320487459830474, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 1.23}) +[2026-06-23 10:53:32,898][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.04093070328235626, 'train/grad_norm_pre_clip': 12.158722877502441, 'timing/data': 0.0012124800123274326, 'timing/model': 2.0262553799839225, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 1.23}) +[2026-06-23 10:53:37,133][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.15097945928573608, 'train/grad_norm_pre_clip': 32.89473342895508, 'timing/data': 0.001245541003299877, 'timing/model': 2.2509084699850064, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 1.23}) +[2026-06-23 10:53:41,152][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.09479863196611404, 'train/grad_norm_pre_clip': 27.782196044921875, 'timing/data': 0.001254888018593192, 'timing/model': 2.033120525011327, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.23}) +[2026-06-23 10:53:45,241][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.11294450610876083, 'train/grad_norm_pre_clip': 20.243370056152344, 'timing/data': 0.07046244901721366, 'timing/model': 2.030924477003282, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 1.23}) +[2026-06-23 10:53:49,339][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.16265951097011566, 'train/grad_norm_pre_clip': 28.462112426757812, 'timing/data': 0.0012081020104233176, 'timing/model': 2.0303402860008646, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 1.23}) +[2026-06-23 10:53:53,358][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.009913978166878223, 'train/grad_norm_pre_clip': 6.734472274780273, 'timing/data': 0.001478432008298114, 'timing/model': 2.0311498520022724, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 1.23}) +[2026-06-23 10:53:57,378][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.02631424553692341, 'train/grad_norm_pre_clip': 11.820517539978027, 'timing/data': 0.001446302980184555, 'timing/model': 2.0310308820044156, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 1.23}) +[2026-06-23 10:54:01,460][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.04216417297720909, 'train/grad_norm_pre_clip': 17.551647186279297, 'timing/data': 0.059622497006785125, 'timing/model': 2.033821908989921, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 1.24}) +[2026-06-23 10:54:05,563][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.020062649622559547, 'train/grad_norm_pre_clip': 12.596015930175781, 'timing/data': 0.0013450160040520132, 'timing/model': 2.031866134988377, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 1.24}) +[2026-06-23 10:54:09,618][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.16789697110652924, 'train/grad_norm_pre_clip': 32.9311637878418, 'timing/data': 0.0012461330043151975, 'timing/model': 2.068467300006887, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 1.24}) +[2026-06-23 10:54:13,638][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.11388755589723587, 'train/grad_norm_pre_clip': 25.89038848876953, 'timing/data': 0.0014513320056721568, 'timing/model': 2.030954418994952, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 1.24}) +[2026-06-23 10:54:17,718][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.03870898485183716, 'train/grad_norm_pre_clip': 11.926403999328613, 'timing/data': 0.058953100990038365, 'timing/model': 2.0333736590109766, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 1.24}) +[2026-06-23 10:54:21,825][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.0777415856719017, 'train/grad_norm_pre_clip': 22.465742111206055, 'timing/data': 0.0012082130124326795, 'timing/model': 2.033155996003188, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 1.24}) +[2026-06-23 10:54:25,851][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.07038877159357071, 'train/grad_norm_pre_clip': 18.186012268066406, 'timing/data': 0.0014028619916643947, 'timing/model': 2.03372380399378, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 1.24}) +[2026-06-23 10:54:29,877][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.06921190768480301, 'train/grad_norm_pre_clip': 14.716246604919434, 'timing/data': 0.0012108069786336273, 'timing/model': 2.0346672580053564, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 1.24}) +[2026-06-23 10:54:33,958][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3200 +[2026-06-23 10:56:28,277][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3200 +[2026-06-23 10:56:28,278][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3200 +[2026-06-23 10:58:35,768][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3200 +[2026-06-23 10:58:35,770][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.04847635328769684, 'train/grad_norm_pre_clip': 12.696881294250488, 'eval/action_loss/seconds': 114.3147960859933, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.040939811617136, 'eval/latency_2/loss': 0.040939811617136, 'eval/flappy/loss': 0.040939811617136, 'eval/flappy/latency_2/loss': 0.040939811617136, 'eval/flap_precision': 0.9504950495049505, 'eval/flap_recall': 0.97165991902834, 'eval/flap_f1': 0.960960960960961, 'eval/flap_rate_model': 0.21041666666666667, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9609609609609611, 'eval/latency_2/flap_precision': 0.9504950495049505, 'eval/latency_2/flap_recall': 0.97165991902834, 'eval/latency_2/flap_f1': 0.960960960960961, 'eval/latency_2/flap_rate_model': 0.21041666666666667, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9609609609609611, 'eval/action_classification/data_wait_seconds': 50.45212173461914, 'eval/action_classification/predict_seconds': 75.29338836669922, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06151061598211527, 'timing/model': 2.0311969549802598, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 1.24}) +[2026-06-23 10:58:39,802][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.06823771446943283, 'train/grad_norm_pre_clip': 9.530970573425293, 'timing/data': 0.00118504001875408, 'timing/model': 1.9885651540243998, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 1.24}) +[2026-06-23 10:58:43,758][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.1280316859483719, 'train/grad_norm_pre_clip': 27.505783081054688, 'timing/data': 0.0012761079997289926, 'timing/model': 2.004204724013107, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 1.24}) +[2026-06-23 10:58:47,720][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.06600296497344971, 'train/grad_norm_pre_clip': 22.272363662719727, 'timing/data': 0.0014882499817758799, 'timing/model': 2.001854911999544, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 1.24}) +[2026-06-23 10:58:51,752][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.03389088436961174, 'train/grad_norm_pre_clip': 7.816503047943115, 'timing/data': 0.06773475700174458, 'timing/model': 2.008257112989668, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 1.24}) +[2026-06-23 10:58:55,807][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.05321992188692093, 'train/grad_norm_pre_clip': 14.852385520935059, 'timing/data': 0.0013745809847023338, 'timing/model': 2.001568100997247, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 1.24}) +[2026-06-23 10:58:59,786][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.07445448637008667, 'train/grad_norm_pre_clip': 50.292381286621094, 'timing/data': 0.0012628340045921504, 'timing/model': 2.013820503023453, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 1.24}) +[2026-06-23 10:59:03,754][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.0693487897515297, 'train/grad_norm_pre_clip': 15.323098182678223, 'timing/data': 0.0014021199895069003, 'timing/model': 2.003226225002436, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 1.24}) +[2026-06-23 10:59:07,794][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.024243740364909172, 'train/grad_norm_pre_clip': 8.357251167297363, 'timing/data': 0.06689107898273505, 'timing/model': 2.011607702996116, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 1.24}) +[2026-06-23 10:59:12,134][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.030747998505830765, 'train/grad_norm_pre_clip': 13.59232234954834, 'timing/data': 0.0012492779933381826, 'timing/model': 2.0096757630235516, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 1.24}) +[2026-06-23 10:59:16,129][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.10460924357175827, 'train/grad_norm_pre_clip': 19.671724319458008, 'timing/data': 0.0014222880126908422, 'timing/model': 2.0256465649872553, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 1.24}) +[2026-06-23 10:59:20,127][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.03803255781531334, 'train/grad_norm_pre_clip': 16.012109756469727, 'timing/data': 0.001249088003532961, 'timing/model': 2.029049903008854, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 1.24}) +[2026-06-23 10:59:24,179][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.05882912501692772, 'train/grad_norm_pre_clip': 13.490793228149414, 'timing/data': 0.06230004399549216, 'timing/model': 2.015832248987863, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 1.24}) +[2026-06-23 10:59:28,270][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.03733391687273979, 'train/grad_norm_pre_clip': 9.532079696655273, 'timing/data': 0.0017004230176098645, 'timing/model': 2.0223269299895037, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 1.24}) +[2026-06-23 10:59:32,258][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.07048685103654861, 'train/grad_norm_pre_clip': 16.950733184814453, 'timing/data': 0.0012827300233766437, 'timing/model': 2.0156327889999375, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 1.24}) +[2026-06-23 10:59:36,250][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.13381558656692505, 'train/grad_norm_pre_clip': 23.140926361083984, 'timing/data': 0.0012394209916237742, 'timing/model': 2.017448945000069, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 1.24}) +[2026-06-23 10:59:40,290][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.04960670322179794, 'train/grad_norm_pre_clip': 26.43195343017578, 'timing/data': 0.055530355020891875, 'timing/model': 2.0161619359860197, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 1.24}) +[2026-06-23 10:59:44,421][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.04177981615066528, 'train/grad_norm_pre_clip': 20.445720672607422, 'timing/data': 0.001225104002514854, 'timing/model': 2.0175735850061756, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 1.25}) +[2026-06-23 10:59:48,436][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.027558866888284683, 'train/grad_norm_pre_clip': 12.125372886657715, 'timing/data': 0.0014108269824646413, 'timing/model': 2.0315684170054737, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 1.25}) +[2026-06-23 10:59:52,443][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.022101808339357376, 'train/grad_norm_pre_clip': 5.9222941398620605, 'timing/data': 0.0015113230037968606, 'timing/model': 2.0261395540146623, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 1.25}) +[2026-06-23 10:59:56,503][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.022209396585822105, 'train/grad_norm_pre_clip': 8.967691421508789, 'timing/data': 0.05738331901375204, 'timing/model': 2.0246073819871526, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 1.25}) +[2026-06-23 11:00:00,597][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.038256604224443436, 'train/grad_norm_pre_clip': 22.986364364624023, 'timing/data': 0.0012247830163687468, 'timing/model': 2.026212688011583, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 1.25}) +[2026-06-23 11:00:04,631][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.03492908179759979, 'train/grad_norm_pre_clip': 8.613879203796387, 'timing/data': 0.0013913110014982522, 'timing/model': 2.0402255929948296, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 1.25}) +[2026-06-23 11:00:08,634][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.01606847532093525, 'train/grad_norm_pre_clip': 7.433298587799072, 'timing/data': 0.0012094349949620664, 'timing/model': 2.023997784999665, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 1.25}) +[2026-06-23 11:00:12,725][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.17966721951961517, 'train/grad_norm_pre_clip': 28.354639053344727, 'timing/data': 0.0721253209921997, 'timing/model': 2.026664092001738, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 1.25}) +[2026-06-23 11:00:16,930][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.03458305448293686, 'train/grad_norm_pre_clip': 16.329662322998047, 'timing/data': 0.0012720810191240162, 'timing/model': 2.13493130900315, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 1.25}) +[2026-06-23 11:00:20,940][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.007815870456397533, 'train/grad_norm_pre_clip': 5.127011299133301, 'timing/data': 0.0015218420012388378, 'timing/model': 2.024942310003098, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 1.25}) +[2026-06-23 11:00:24,948][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.07040334492921829, 'train/grad_norm_pre_clip': 19.844568252563477, 'timing/data': 0.0015540019958280027, 'timing/model': 2.0235737900075037, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 1.25}) +[2026-06-23 11:00:29,020][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.11808783560991287, 'train/grad_norm_pre_clip': 14.749171257019043, 'timing/data': 0.0590745439985767, 'timing/model': 2.0290308840048965, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 1.25}) +[2026-06-23 11:00:33,135][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.08539355546236038, 'train/grad_norm_pre_clip': 40.3851318359375, 'timing/data': 0.0015313600015360862, 'timing/model': 2.04353999599698, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 1.25}) +[2026-06-23 11:00:37,147][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.052649740129709244, 'train/grad_norm_pre_clip': 17.63768768310547, 'timing/data': 0.0012810970074497163, 'timing/model': 2.0269238110049628, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 1.25}) +[2026-06-23 11:00:41,157][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.017937680706381798, 'train/grad_norm_pre_clip': 8.488200187683105, 'timing/data': 0.0011772050056606531, 'timing/model': 2.023333345015999, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 1.25}) +[2026-06-23 11:00:45,247][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.11546541005373001, 'train/grad_norm_pre_clip': 48.8826904296875, 'timing/data': 0.06686060200445354, 'timing/model': 2.0270020950119942, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 1.25}) +[2026-06-23 11:00:49,366][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.16889885067939758, 'train/grad_norm_pre_clip': 23.851179122924805, 'timing/data': 0.0013989450235385448, 'timing/model': 2.043206769012613, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 1.25}) +[2026-06-23 11:00:53,558][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.10958076268434525, 'train/grad_norm_pre_clip': 13.274391174316406, 'timing/data': 0.001316562993451953, 'timing/model': 2.0259159579873085, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 1.25}) +[2026-06-23 11:00:57,571][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.06260080635547638, 'train/grad_norm_pre_clip': 32.98932647705078, 'timing/data': 0.0012333290069364011, 'timing/model': 2.028024593979353, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 1.25}) +[2026-06-23 11:01:01,643][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.08160273730754852, 'train/grad_norm_pre_clip': 23.731435775756836, 'timing/data': 0.058748881972860545, 'timing/model': 2.0271172469947487, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.25}) +[2026-06-23 11:01:05,739][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.08307938277721405, 'train/grad_norm_pre_clip': 22.73761558532715, 'timing/data': 0.001370080979540944, 'timing/model': 2.0260738780198153, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 1.25}) +[2026-06-23 11:01:09,752][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.026596570387482643, 'train/grad_norm_pre_clip': 11.772530555725098, 'timing/data': 0.0013858010061085224, 'timing/model': 2.0263546479982324, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 1.25}) +[2026-06-23 11:01:13,779][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.041938696056604385, 'train/grad_norm_pre_clip': 13.860481262207031, 'timing/data': 0.0012204850208945572, 'timing/model': 2.0411976059840526, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 1.25}) +[2026-06-23 11:01:17,847][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.14326786994934082, 'train/grad_norm_pre_clip': 36.59845733642578, 'timing/data': 0.05839535899576731, 'timing/model': 2.0272292120207567, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 1.25}) +[2026-06-23 11:01:21,944][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.059999898076057434, 'train/grad_norm_pre_clip': 28.828563690185547, 'timing/data': 0.0012190029956400394, 'timing/model': 2.0245877419947647, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 1.25}) +[2026-06-23 11:01:25,984][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.023336851969361305, 'train/grad_norm_pre_clip': 10.051108360290527, 'timing/data': 0.0013940860226284713, 'timing/model': 2.0272052379732486, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 1.25}) +[2026-06-23 11:01:29,999][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.03512418642640114, 'train/grad_norm_pre_clip': 11.60878849029541, 'timing/data': 0.0014186020125634968, 'timing/model': 2.031242448982084, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 1.26}) +[2026-06-23 11:01:34,079][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.0485067255795002, 'train/grad_norm_pre_clip': 15.757471084594727, 'timing/data': 0.05826737202005461, 'timing/model': 2.037333032989409, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 1.26}) +[2026-06-23 11:01:38,198][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.27194803953170776, 'train/grad_norm_pre_clip': 32.241661071777344, 'timing/data': 0.001313165994361043, 'timing/model': 2.025758142990526, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 1.26}) +[2026-06-23 11:01:42,220][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.12073544412851334, 'train/grad_norm_pre_clip': 42.835205078125, 'timing/data': 0.0012401010026223958, 'timing/model': 2.0307291689969134, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 1.26}) +[2026-06-23 11:01:46,257][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.02153550088405609, 'train/grad_norm_pre_clip': 9.562999725341797, 'timing/data': 0.0012120600149501115, 'timing/model': 2.041090155020356, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 1.26}) +[2026-06-23 11:01:50,345][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.04844050109386444, 'train/grad_norm_pre_clip': 13.867206573486328, 'timing/data': 0.05729982501361519, 'timing/model': 2.0393074399908073, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 1.26}) +[2026-06-23 11:01:54,462][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.06093507632613182, 'train/grad_norm_pre_clip': 21.926836013793945, 'timing/data': 0.0014127210015431046, 'timing/model': 2.0290102220023982, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 1.26}) +[2026-06-23 11:19:17,786][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3250 +[2026-06-23 11:19:17,788][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.07369010895490646, 'train/grad_norm_pre_clip': 17.472370147705078, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 268.90000000001544, 'rl_games_eval/mid_train/mean_length': 2184.0, 'rl_games_eval/mid_train/std_reward': 215.18962800285104, 'rl_games_eval/mid_train/std_length': 1734.2502702897295, 'rl_games_eval/mid_train/macro_mean_reward': 268.90000000001544, 'rl_games_eval/mid_train/macro_mean_length': 2184.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 268.90000000001544, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2184.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 215.18962800285104, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1734.2502702897295, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 268.90000000001544, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0012666200054809451, 'timing/model': 2.033530755987158, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 1.26}) +[2026-06-23 11:19:21,662][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.09576642513275146, 'train/grad_norm_pre_clip': 21.445730209350586, 'timing/data': 0.0014337199972942472, 'timing/model': 1.959679933002917, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 1.26}) +[2026-06-23 11:19:25,608][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.05610888823866844, 'train/grad_norm_pre_clip': 23.19268035888672, 'timing/data': 0.0679945059819147, 'timing/model': 1.961408864997793, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 1.26}) +[2026-06-23 11:19:29,597][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.0410788394510746, 'train/grad_norm_pre_clip': 21.501117706298828, 'timing/data': 0.0015023070154711604, 'timing/model': 1.9858936149976216, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 1.26}) +[2026-06-23 11:19:33,478][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.07340598851442337, 'train/grad_norm_pre_clip': 22.732656478881836, 'timing/data': 0.0013609659799840301, 'timing/model': 1.9629236090113409, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 1.26}) +[2026-06-23 11:19:37,368][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.1441023349761963, 'train/grad_norm_pre_clip': 49.47720718383789, 'timing/data': 0.0015475699910894036, 'timing/model': 1.9652254219981842, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 1.26}) +[2026-06-23 11:19:41,342][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.04705890268087387, 'train/grad_norm_pre_clip': 22.074068069458008, 'timing/data': 0.06525187700754032, 'timing/model': 1.9820966199913528, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 1.26}) +[2026-06-23 11:19:45,585][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.19247306883335114, 'train/grad_norm_pre_clip': 49.17168426513672, 'timing/data': 0.0011709039972629398, 'timing/model': 1.9767982850025874, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 1.26}) +[2026-06-23 11:19:49,553][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.044111814349889755, 'train/grad_norm_pre_clip': 20.227317810058594, 'timing/data': 0.001351387007161975, 'timing/model': 2.024520050006686, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 1.26}) +[2026-06-23 11:19:53,482][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.13060998916625977, 'train/grad_norm_pre_clip': 24.010663986206055, 'timing/data': 0.0013669460022356361, 'timing/model': 1.9934807319950778, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 1.26}) +[2026-06-23 11:19:57,473][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.05073059722781181, 'train/grad_norm_pre_clip': 16.43798065185547, 'timing/data': 0.05906614498235285, 'timing/model': 1.99474059799104, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 1.26}) +[2026-06-23 11:20:01,505][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.1360815018415451, 'train/grad_norm_pre_clip': 16.39057731628418, 'timing/data': 0.001469725975766778, 'timing/model': 1.9932420660043135, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 1.26}) +[2026-06-23 11:20:05,459][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.14834123849868774, 'train/grad_norm_pre_clip': 34.5313720703125, 'timing/data': 0.001463744993088767, 'timing/model': 2.0088969460048247, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 1.26}) +[2026-06-23 11:20:09,394][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.022468365728855133, 'train/grad_norm_pre_clip': 12.712157249450684, 'timing/data': 0.0012109980161767453, 'timing/model': 1.9865130349935498, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 1.26}) +[2026-06-23 11:20:13,402][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.11908242106437683, 'train/grad_norm_pre_clip': 20.333454132080078, 'timing/data': 0.06858759198803455, 'timing/model': 1.990723833994707, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 1.26}) +[2026-06-23 11:20:17,502][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.10932634770870209, 'train/grad_norm_pre_clip': 24.8112850189209, 'timing/data': 0.001208512985613197, 'timing/model': 2.0102039339835756, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 1.26}) +[2026-06-23 11:20:21,459][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.14411665499210358, 'train/grad_norm_pre_clip': 24.05902671813965, 'timing/data': 0.0012504810001701117, 'timing/model': 1.9994615679897834, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 1.26}) +[2026-06-23 11:20:25,449][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.05932286009192467, 'train/grad_norm_pre_clip': 15.059904098510742, 'timing/data': 0.0012596790038514882, 'timing/model': 2.0236907250073273, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 1.26}) +[2026-06-23 11:20:29,468][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.29134494066238403, 'train/grad_norm_pre_clip': 32.58861541748047, 'timing/data': 0.05775428900960833, 'timing/model': 1.9926364189886954, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 1.26}) +[2026-06-23 11:20:33,498][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.09136907756328583, 'train/grad_norm_pre_clip': 19.40776824951172, 'timing/data': 0.001106954994611442, 'timing/model': 1.9938333190220874, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 1.27}) +[2026-06-23 11:20:37,467][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.06556879729032516, 'train/grad_norm_pre_clip': 14.673233985900879, 'timing/data': 0.0012369460018817335, 'timing/model': 2.005099417991005, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 1.27}) +[2026-06-23 11:20:41,456][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.08864269405603409, 'train/grad_norm_pre_clip': 22.264915466308594, 'timing/data': 0.001227708999067545, 'timing/model': 2.015703343990026, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 1.27}) +[2026-06-23 11:20:45,496][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.06238507106900215, 'train/grad_norm_pre_clip': 11.696962356567383, 'timing/data': 0.05878814001334831, 'timing/model': 2.015998876973754, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 1.27}) +[2026-06-23 11:20:49,581][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.026976361870765686, 'train/grad_norm_pre_clip': 9.990520477294922, 'timing/data': 0.0017809020064305514, 'timing/model': 2.0259508670133073, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 1.27}) +[2026-06-23 11:20:53,793][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.01030672062188387, 'train/grad_norm_pre_clip': 4.328226089477539, 'timing/data': 0.001241233985638246, 'timing/model': 2.0086818000127096, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 1.27}) +[2026-06-23 11:20:57,773][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.13703513145446777, 'train/grad_norm_pre_clip': 29.62671661376953, 'timing/data': 0.0012159170000813901, 'timing/model': 2.0111860760080162, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 1.27}) +[2026-06-23 11:21:01,818][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.11309269070625305, 'train/grad_norm_pre_clip': 17.91286277770996, 'timing/data': 0.06775750601082109, 'timing/model': 2.0118639760185033, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 1.27}) +[2026-06-23 11:21:05,883][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.14363716542720795, 'train/grad_norm_pre_clip': 26.44283103942871, 'timing/data': 0.0012346020084805787, 'timing/model': 2.0104747869772837, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 1.27}) +[2026-06-23 11:21:09,867][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.013280687853693962, 'train/grad_norm_pre_clip': 7.538357734680176, 'timing/data': 0.001226066000526771, 'timing/model': 2.0152801949880086, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 1.27}) +[2026-06-23 11:21:13,854][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.028382301330566406, 'train/grad_norm_pre_clip': 9.454071044921875, 'timing/data': 0.0014808870037086308, 'timing/model': 2.015703824989032, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 1.27}) +[2026-06-23 11:21:17,901][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.044606104493141174, 'train/grad_norm_pre_clip': 12.532499313354492, 'timing/data': 0.059125492989551276, 'timing/model': 2.0182411829882767, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 1.27}) +[2026-06-23 11:21:21,980][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.05013555660843849, 'train/grad_norm_pre_clip': 23.961517333984375, 'timing/data': 0.0013479609915520996, 'timing/model': 2.019135833979817, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 1.27}) +[2026-06-23 11:21:26,004][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.04491494223475456, 'train/grad_norm_pre_clip': 20.784822463989258, 'timing/data': 0.0012869380007032305, 'timing/model': 2.0201833079918288, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 1.27}) +[2026-06-23 11:21:30,004][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.09002070873975754, 'train/grad_norm_pre_clip': 17.151424407958984, 'timing/data': 0.0012538679875433445, 'timing/model': 2.0209365180053283, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 1.27}) +[2026-06-23 11:21:34,069][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.05882611870765686, 'train/grad_norm_pre_clip': 25.99095916748047, 'timing/data': 0.06867176599917002, 'timing/model': 2.019930675975047, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 1.27}) +[2026-06-23 11:21:38,154][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.012848232872784138, 'train/grad_norm_pre_clip': 3.2731709480285645, 'timing/data': 0.0012584560026880354, 'timing/model': 2.022038079012418, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.27}) +[2026-06-23 11:21:42,154][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.0756341889500618, 'train/grad_norm_pre_clip': 23.479434967041016, 'timing/data': 0.0012286709970794618, 'timing/model': 2.0220181399781723, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 1.27}) +[2026-06-23 11:21:46,174][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.07210584729909897, 'train/grad_norm_pre_clip': 19.075014114379883, 'timing/data': 0.0013924840022809803, 'timing/model': 2.042088440997759, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 1.27}) +[2026-06-23 11:21:50,231][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.07992714643478394, 'train/grad_norm_pre_clip': 12.275490760803223, 'timing/data': 0.0575172099925112, 'timing/model': 2.0219384389929473, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 1.27}) +[2026-06-23 11:21:54,338][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.06619220972061157, 'train/grad_norm_pre_clip': 13.239048957824707, 'timing/data': 0.0011888569861184806, 'timing/model': 2.03848874400137, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 1.27}) +[2026-06-23 11:21:58,407][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.12603236734867096, 'train/grad_norm_pre_clip': 27.0216121673584, 'timing/data': 0.0012840630079153925, 'timing/model': 2.0850655929825734, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 1.27}) +[2026-06-23 11:22:02,431][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.014649189077317715, 'train/grad_norm_pre_clip': 9.018134117126465, 'timing/data': 0.001386622985592112, 'timing/model': 2.0417926060035825, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 1.27}) +[2026-06-23 11:22:06,512][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.05581742152571678, 'train/grad_norm_pre_clip': 15.4036283493042, 'timing/data': 0.0726344290014822, 'timing/model': 2.0272924450109713, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 1.27}) +[2026-06-23 11:22:10,605][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.07161467522382736, 'train/grad_norm_pre_clip': 18.603557586669922, 'timing/data': 0.0011883259867317975, 'timing/model': 2.0255728559859563, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 1.27}) +[2026-06-23 11:22:14,615][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.038917046040296555, 'train/grad_norm_pre_clip': 9.734868049621582, 'timing/data': 0.001260299002751708, 'timing/model': 2.025450215995079, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 1.27}) +[2026-06-23 11:22:18,623][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.12932902574539185, 'train/grad_norm_pre_clip': 35.746463775634766, 'timing/data': 0.0012356439838185906, 'timing/model': 2.0267178069916554, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 1.28}) +[2026-06-23 11:22:22,692][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.04032480716705322, 'train/grad_norm_pre_clip': 15.253092765808105, 'timing/data': 0.058196275000227615, 'timing/model': 2.0264358530112077, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 1.28}) +[2026-06-23 11:22:26,787][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.05497424677014351, 'train/grad_norm_pre_clip': 21.262285232543945, 'timing/data': 0.0012040750007145107, 'timing/model': 2.0258967859845143, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 1.28}) +[2026-06-23 11:22:30,796][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.02804810181260109, 'train/grad_norm_pre_clip': 10.38152027130127, 'timing/data': 0.0012491790112107992, 'timing/model': 2.0275446699815802, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 1.28}) +[2026-06-23 11:22:35,065][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.023426230996847153, 'train/grad_norm_pre_clip': 7.751353740692139, 'timing/data': 0.0012603589857462794, 'timing/model': 2.0321950519864913, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 1.28}) +[2026-06-23 11:22:39,138][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3300 +[2026-06-23 11:24:33,663][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3300 +[2026-06-23 11:24:33,664][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3300 +[2026-06-23 11:26:41,298][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3300 +[2026-06-23 11:26:41,299][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.05390368029475212, 'train/grad_norm_pre_clip': 15.380778312683105, 'eval/action_loss/seconds': 114.52168137900298, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03993041440844536, 'eval/latency_2/loss': 0.03993041440844536, 'eval/flappy/loss': 0.03993041440844536, 'eval/flappy/latency_2/loss': 0.03993041440844536, 'eval/flap_precision': 0.9469200524246396, 'eval/flap_recall': 0.975033738191633, 'eval/flap_f1': 0.9607712765957447, 'eval/flap_rate_model': 0.21194444444444444, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9607712765957447, 'eval/latency_2/flap_precision': 0.9469200524246396, 'eval/latency_2/flap_recall': 0.975033738191633, 'eval/latency_2/flap_f1': 0.9607712765957447, 'eval/latency_2/flap_rate_model': 0.21194444444444444, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9607712765957447, 'eval/action_classification/data_wait_seconds': 50.8942985534668, 'eval/action_classification/predict_seconds': 75.0369873046875, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.05682546101161279, 'timing/model': 2.0303145649959333, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 1.28}) +[2026-06-23 11:26:45,355][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.08911589533090591, 'train/grad_norm_pre_clip': 16.492042541503906, 'timing/data': 0.0013575789926107973, 'timing/model': 2.0088866819860414, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 1.28}) +[2026-06-23 11:26:49,312][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.07637090981006622, 'train/grad_norm_pre_clip': 20.96384048461914, 'timing/data': 0.001266270992346108, 'timing/model': 1.9986971199978143, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 1.28}) +[2026-06-23 11:26:53,475][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.10522036254405975, 'train/grad_norm_pre_clip': 23.98430633544922, 'timing/data': 0.0014388190174940974, 'timing/model': 2.005422357993666, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 1.28}) +[2026-06-23 11:26:57,570][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.049320146441459656, 'train/grad_norm_pre_clip': 9.85706901550293, 'timing/data': 0.12857946602161974, 'timing/model': 2.006549689976964, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 1.28}) +[2026-06-23 11:27:01,914][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.07403567433357239, 'train/grad_norm_pre_clip': 24.22259521484375, 'timing/data': 0.0011991559877060354, 'timing/model': 2.0240711490041576, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 1.28}) +[2026-06-23 11:27:05,894][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.09081974625587463, 'train/grad_norm_pre_clip': 24.29484748840332, 'timing/data': 0.001194987999042496, 'timing/model': 2.018565002013929, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 1.28}) +[2026-06-23 11:27:09,899][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.034060779958963394, 'train/grad_norm_pre_clip': 12.89156723022461, 'timing/data': 0.0014962650020606816, 'timing/model': 2.019651709997561, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 1.28}) +[2026-06-23 11:27:13,949][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.009241584688425064, 'train/grad_norm_pre_clip': 5.895995616912842, 'timing/data': 0.07664883299730718, 'timing/model': 2.007563809020212, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 1.28}) +[2026-06-23 11:27:18,027][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.028481073677539825, 'train/grad_norm_pre_clip': 9.676139831542969, 'timing/data': 0.0010969570139423013, 'timing/model': 2.0141899030131754, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 1.28}) +[2026-06-23 11:27:22,011][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.08981665223836899, 'train/grad_norm_pre_clip': 21.10003662109375, 'timing/data': 0.0019105509854853153, 'timing/model': 2.017356694996124, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 1.28}) +[2026-06-23 11:27:26,001][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.0380171500146389, 'train/grad_norm_pre_clip': 23.55986976623535, 'timing/data': 0.001234892988577485, 'timing/model': 2.016709488001652, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 1.28}) +[2026-06-23 11:27:30,053][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.12067227810621262, 'train/grad_norm_pre_clip': 26.818159103393555, 'timing/data': 0.06339678197400644, 'timing/model': 2.0146206869976595, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 1.28}) +[2026-06-23 11:27:34,178][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.052443549036979675, 'train/grad_norm_pre_clip': 9.456130981445312, 'timing/data': 0.0012054679973516613, 'timing/model': 2.019260180008132, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 1.28}) +[2026-06-23 11:27:38,166][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.018126105889678, 'train/grad_norm_pre_clip': 7.737918376922607, 'timing/data': 0.00150537199806422, 'timing/model': 2.0155267599911895, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 1.28}) +[2026-06-23 11:27:42,165][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.07929942011833191, 'train/grad_norm_pre_clip': 15.49914264678955, 'timing/data': 0.0014204750186763704, 'timing/model': 2.021761170006357, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 1.28}) +[2026-06-23 11:27:46,219][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.09865449368953705, 'train/grad_norm_pre_clip': 30.438390731811523, 'timing/data': 0.05727172197657637, 'timing/model': 2.0210587400069926, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 1.28}) +[2026-06-23 11:27:50,305][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.008767257444560528, 'train/grad_norm_pre_clip': 5.146851062774658, 'timing/data': 0.0012637159961741418, 'timing/model': 2.0229773260070942, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 1.28}) +[2026-06-23 11:27:54,303][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.08508339524269104, 'train/grad_norm_pre_clip': 24.845685958862305, 'timing/data': 0.001275397022254765, 'timing/model': 2.022003735008184, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 1.28}) +[2026-06-23 11:27:58,315][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.09137158840894699, 'train/grad_norm_pre_clip': 30.950542449951172, 'timing/data': 0.0012716800265479833, 'timing/model': 2.0340429240022786, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 1.28}) +[2026-06-23 11:28:02,375][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.0720822885632515, 'train/grad_norm_pre_clip': 25.074087142944336, 'timing/data': 0.058327663980890065, 'timing/model': 2.0238831760070752, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 1.28}) +[2026-06-23 11:28:06,475][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.14629289507865906, 'train/grad_norm_pre_clip': 42.06391906738281, 'timing/data': 0.0012525039783213288, 'timing/model': 2.0350848369998857, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 1.29}) +[2026-06-23 11:28:10,697][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.02785695530474186, 'train/grad_norm_pre_clip': 25.58669090270996, 'timing/data': 0.001407260017003864, 'timing/model': 2.0239429339999333, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 1.29}) +[2026-06-23 11:28:14,705][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.01941666565835476, 'train/grad_norm_pre_clip': 13.20620059967041, 'timing/data': 0.0012061790039297193, 'timing/model': 2.0263790449826047, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 1.29}) +[2026-06-23 11:28:18,772][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.06645067036151886, 'train/grad_norm_pre_clip': 17.215049743652344, 'timing/data': 0.057634914002846926, 'timing/model': 2.025551605009241, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 1.29}) +[2026-06-23 11:28:22,868][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.06794391572475433, 'train/grad_norm_pre_clip': 22.836671829223633, 'timing/data': 0.0012117489823140204, 'timing/model': 2.0278383400000166, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 1.29}) +[2026-06-23 11:28:26,891][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.020946459844708443, 'train/grad_norm_pre_clip': 11.182235717773438, 'timing/data': 0.0014015600027050823, 'timing/model': 2.042071333999047, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 1.29}) +[2026-06-23 11:28:30,903][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.08573086559772491, 'train/grad_norm_pre_clip': 24.354389190673828, 'timing/data': 0.0012880110007245094, 'timing/model': 2.026689095015172, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 1.29}) +[2026-06-23 11:28:34,969][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.13276734948158264, 'train/grad_norm_pre_clip': 26.430334091186523, 'timing/data': 0.05747322499519214, 'timing/model': 2.025649061979493, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 1.29}) +[2026-06-23 11:28:39,074][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.15396547317504883, 'train/grad_norm_pre_clip': 35.778717041015625, 'timing/data': 0.0012146049994044006, 'timing/model': 2.0323305090132635, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 1.29}) +[2026-06-23 11:28:43,131][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.028210079297423363, 'train/grad_norm_pre_clip': 13.717259407043457, 'timing/data': 0.0011958900140598416, 'timing/model': 2.0284339360077865, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 1.29}) +[2026-06-23 11:28:47,142][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.029255997389554977, 'train/grad_norm_pre_clip': 9.862610816955566, 'timing/data': 0.001137021987233311, 'timing/model': 2.029198559001088, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 1.29}) +[2026-06-23 11:28:51,217][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.08828536421060562, 'train/grad_norm_pre_clip': 28.03927993774414, 'timing/data': 0.05682276299921796, 'timing/model': 2.0321836239891127, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 1.29}) +[2026-06-23 11:28:55,347][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.15185299515724182, 'train/grad_norm_pre_clip': 19.62740707397461, 'timing/data': 0.0012768100132234395, 'timing/model': 2.0609772260067984, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 1.29}) +[2026-06-23 11:28:59,378][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.05233527719974518, 'train/grad_norm_pre_clip': 29.351980209350586, 'timing/data': 0.0012487380008678883, 'timing/model': 2.0464615800010506, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.29}) +[2026-06-23 11:29:03,397][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.14661073684692383, 'train/grad_norm_pre_clip': 39.49689483642578, 'timing/data': 0.0013762939779553562, 'timing/model': 2.032095398986712, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 1.29}) +[2026-06-23 11:29:07,461][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.11188240349292755, 'train/grad_norm_pre_clip': 28.18019676208496, 'timing/data': 0.05838862599921413, 'timing/model': 2.023855259991251, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 1.29}) +[2026-06-23 11:29:11,577][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.06364694982767105, 'train/grad_norm_pre_clip': 22.569822311401367, 'timing/data': 0.0012645369861274958, 'timing/model': 2.047442509996472, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 1.29}) +[2026-06-23 11:29:15,650][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.13084588944911957, 'train/grad_norm_pre_clip': 29.841705322265625, 'timing/data': 0.0012482980091590434, 'timing/model': 2.085351387009723, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 1.29}) +[2026-06-23 11:29:19,672][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.06520161032676697, 'train/grad_norm_pre_clip': 31.21725082397461, 'timing/data': 0.0012455520045477897, 'timing/model': 2.028944732999662, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 1.29}) +[2026-06-23 11:29:23,770][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.11744072288274765, 'train/grad_norm_pre_clip': 22.563129425048828, 'timing/data': 0.06678299899795093, 'timing/model': 2.0292947989946697, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 1.29}) +[2026-06-23 11:29:27,875][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.026071352884173393, 'train/grad_norm_pre_clip': 21.15838623046875, 'timing/data': 0.0013960489886812866, 'timing/model': 2.030614605988376, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 1.29}) +[2026-06-23 11:29:31,885][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.07228211313486099, 'train/grad_norm_pre_clip': 16.518781661987305, 'timing/data': 0.0012369760079309344, 'timing/model': 2.027933960984228, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 1.29}) +[2026-06-23 11:29:35,899][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.07931078225374222, 'train/grad_norm_pre_clip': 27.1409969329834, 'timing/data': 0.0012272080057300627, 'timing/model': 2.03125963700586, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 1.29}) +[2026-06-23 11:29:39,972][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.042551811784505844, 'train/grad_norm_pre_clip': 15.314105987548828, 'timing/data': 0.05633177500567399, 'timing/model': 2.028595512994798, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 1.29}) +[2026-06-23 11:29:44,071][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.07400219142436981, 'train/grad_norm_pre_clip': 21.366615295410156, 'timing/data': 0.0012355830112937838, 'timing/model': 2.0321605799836107, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 1.29}) +[2026-06-23 11:29:48,089][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.06315299868583679, 'train/grad_norm_pre_clip': 16.825401306152344, 'timing/data': 0.0013737090048380196, 'timing/model': 2.029330208984902, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 1.29}) +[2026-06-23 11:29:52,362][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.050106968730688095, 'train/grad_norm_pre_clip': 14.187349319458008, 'timing/data': 0.0014651779783889651, 'timing/model': 2.0310419909947086, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 1.3}) +[2026-06-23 11:29:56,444][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.024089496582746506, 'train/grad_norm_pre_clip': 17.877172470092773, 'timing/data': 0.060639686009380966, 'timing/model': 2.0305057799851056, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 1.3}) +[2026-06-23 11:30:00,546][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.1208297610282898, 'train/grad_norm_pre_clip': 27.20400047302246, 'timing/data': 0.0012477560085244477, 'timing/model': 2.029492827015929, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 1.3}) +[2026-06-23 11:30:04,560][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.005602087359875441, 'train/grad_norm_pre_clip': 1.9084360599517822, 'timing/data': 0.0013300479913596064, 'timing/model': 2.028221846994711, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 1.3}) +[2026-06-23 11:30:08,581][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.015159960836172104, 'train/grad_norm_pre_clip': 5.395203590393066, 'timing/data': 0.001265007012989372, 'timing/model': 2.0292265030147973, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 1.3}) +[2026-06-23 11:30:12,656][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.054291341453790665, 'train/grad_norm_pre_clip': 14.787418365478516, 'timing/data': 0.05955839698435739, 'timing/model': 2.0287917089881375, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 1.3}) +[2026-06-23 11:30:16,762][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.04443147033452988, 'train/grad_norm_pre_clip': 8.660293579101562, 'timing/data': 0.00126116001047194, 'timing/model': 2.0300294249900617, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 1.3}) +[2026-06-23 11:30:20,783][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.09509194642305374, 'train/grad_norm_pre_clip': 21.902141571044922, 'timing/data': 0.0014889719896018505, 'timing/model': 2.028833715012297, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 1.3}) +[2026-06-23 11:30:25,032][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.19334596395492554, 'train/grad_norm_pre_clip': 48.62602615356445, 'timing/data': 0.0014285299985203892, 'timing/model': 2.2650071360112634, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 1.3}) +[2026-06-23 11:30:29,108][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.09645815938711166, 'train/grad_norm_pre_clip': 38.57582473754883, 'timing/data': 0.05757357898983173, 'timing/model': 2.0305303990026005, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 1.3}) +[2026-06-23 11:30:33,213][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.04022395610809326, 'train/grad_norm_pre_clip': 15.15433120727539, 'timing/data': 0.00132113101426512, 'timing/model': 2.0330143890168983, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 1.3}) +[2026-06-23 11:30:37,236][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.012657134793698788, 'train/grad_norm_pre_clip': 6.125850200653076, 'timing/data': 0.0012569430109579116, 'timing/model': 2.0345643530017696, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 1.3}) +[2026-06-23 11:30:41,258][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.050769392400979996, 'train/grad_norm_pre_clip': 19.976810455322266, 'timing/data': 0.00140620899037458, 'timing/model': 2.032239986991044, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 1.3}) +[2026-06-23 11:30:45,342][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.05296182259917259, 'train/grad_norm_pre_clip': 24.749601364135742, 'timing/data': 0.06031017599161714, 'timing/model': 2.0330123029998504, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 1.3}) +[2026-06-23 11:30:49,452][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.10814306139945984, 'train/grad_norm_pre_clip': 30.988990783691406, 'timing/data': 0.0012366849987301975, 'timing/model': 2.030164588999469, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 1.3}) +[2026-06-23 11:30:53,475][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.11177610605955124, 'train/grad_norm_pre_clip': 29.99695587158203, 'timing/data': 0.0014765690139029175, 'timing/model': 2.031027253018692, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 1.3}) +[2026-06-23 11:30:57,529][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.03546055406332016, 'train/grad_norm_pre_clip': 14.198295593261719, 'timing/data': 0.0012197140022180974, 'timing/model': 2.062880748999305, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 1.3}) +[2026-06-23 11:31:01,613][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.13980327546596527, 'train/grad_norm_pre_clip': 37.922481536865234, 'timing/data': 0.057301105989608914, 'timing/model': 2.0347519740171265, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 1.3}) +[2026-06-23 11:31:05,719][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.015115974470973015, 'train/grad_norm_pre_clip': 8.9345703125, 'timing/data': 0.0012250239960849285, 'timing/model': 2.0309771670144983, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 1.3}) +[2026-06-23 11:31:09,740][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.012012848630547523, 'train/grad_norm_pre_clip': 10.292854309082031, 'timing/data': 0.0014071199984755367, 'timing/model': 2.030616110016126, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 1.3}) +[2026-06-23 11:31:13,761][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.02046036534011364, 'train/grad_norm_pre_clip': 10.895256042480469, 'timing/data': 0.0011943370045628399, 'timing/model': 2.0329209079791326, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 1.3}) +[2026-06-23 11:31:17,850][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.05340386927127838, 'train/grad_norm_pre_clip': 27.689146041870117, 'timing/data': 0.059880232001887634, 'timing/model': 2.035047568002483, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 1.3}) +[2026-06-23 11:31:21,955][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.03399736434221268, 'train/grad_norm_pre_clip': 20.130352020263672, 'timing/data': 0.001239931007148698, 'timing/model': 2.0296461249818094, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 1.3}) +[2026-06-23 11:31:25,978][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.08067728579044342, 'train/grad_norm_pre_clip': 21.713754653930664, 'timing/data': 0.001235042989719659, 'timing/model': 2.0344992039899807, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 1.3}) +[2026-06-23 11:31:30,002][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.024854380637407303, 'train/grad_norm_pre_clip': 10.623785972595215, 'timing/data': 0.0012940710003022105, 'timing/model': 2.034461472998373, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 1.3}) +[2026-06-23 11:31:34,218][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.03779563307762146, 'train/grad_norm_pre_clip': 19.13169288635254, 'timing/data': 0.05830033199163154, 'timing/model': 2.032486682990566, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 1.31}) +[2026-06-23 11:31:38,336][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.06570475548505783, 'train/grad_norm_pre_clip': 38.61166000366211, 'timing/data': 0.0012286709970794618, 'timing/model': 2.0359394520055503, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 1.31}) +[2026-06-23 11:31:42,371][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.12659601867198944, 'train/grad_norm_pre_clip': 26.39191246032715, 'timing/data': 0.001267282001208514, 'timing/model': 2.0411086849926505, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 1.31}) +[2026-06-23 11:31:46,408][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.08521664887666702, 'train/grad_norm_pre_clip': 44.871273040771484, 'timing/data': 0.0011863829859066755, 'timing/model': 2.0404602759808768, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 1.31}) +[2026-06-23 11:31:50,491][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.012765453197062016, 'train/grad_norm_pre_clip': 7.952783107757568, 'timing/data': 0.057782866002526134, 'timing/model': 2.0338993610057514, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 1.31}) +[2026-06-23 11:31:54,596][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.06549156457185745, 'train/grad_norm_pre_clip': 25.520769119262695, 'timing/data': 0.0012091639800928533, 'timing/model': 2.0327174259873573, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 1.31}) +[2026-06-23 11:31:58,642][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.017806610092520714, 'train/grad_norm_pre_clip': 9.623327255249023, 'timing/data': 0.0014264159835875034, 'timing/model': 2.0522544010018464, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 1.31}) +[2026-06-23 11:32:02,670][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.14821304380893707, 'train/grad_norm_pre_clip': 28.88326644897461, 'timing/data': 0.0015821540146134794, 'timing/model': 2.0355510509980377, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 1.31}) +[2026-06-23 11:32:07,011][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.14096814393997192, 'train/grad_norm_pre_clip': 51.2297477722168, 'timing/data': 0.07182550698053092, 'timing/model': 2.281712159019662, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 1.31}) +[2026-06-23 11:32:11,118][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.025195004418492317, 'train/grad_norm_pre_clip': 10.5025634765625, 'timing/data': 0.0011991859937552363, 'timing/model': 2.0336197709839325, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 1.31}) +[2026-06-23 11:32:15,153][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.06563743948936462, 'train/grad_norm_pre_clip': 24.63909149169922, 'timing/data': 0.0012896829866804183, 'timing/model': 2.0461475320043974, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 1.31}) +[2026-06-23 11:32:19,180][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.0830225795507431, 'train/grad_norm_pre_clip': 29.089336395263672, 'timing/data': 0.0012012999795842916, 'timing/model': 2.0386299089877866, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.31}) +[2026-06-23 11:32:23,258][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.07023676484823227, 'train/grad_norm_pre_clip': 25.992599487304688, 'timing/data': 0.057648437010357156, 'timing/model': 2.034490387974074, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 1.31}) +[2026-06-23 11:32:27,380][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.05318313091993332, 'train/grad_norm_pre_clip': 32.21643829345703, 'timing/data': 0.0012326879950705916, 'timing/model': 2.0473231229989324, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 1.31}) +[2026-06-23 11:32:31,409][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.03882588446140289, 'train/grad_norm_pre_clip': 15.203523635864258, 'timing/data': 0.0014019410009495914, 'timing/model': 2.038160298980074, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 1.31}) +[2026-06-23 11:32:35,441][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.06376361101865768, 'train/grad_norm_pre_clip': 16.83795738220215, 'timing/data': 0.0016706169990357012, 'timing/model': 2.0365278010140173, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 1.31}) +[2026-06-23 11:32:39,565][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.06676105409860611, 'train/grad_norm_pre_clip': 25.165292739868164, 'timing/data': 0.06801015601377003, 'timing/model': 2.062787329981802, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 1.31}) +[2026-06-23 11:32:43,682][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.10376902669668198, 'train/grad_norm_pre_clip': 17.154138565063477, 'timing/data': 0.0013711639912799, 'timing/model': 2.0378787059744354, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 1.31}) +[2026-06-23 11:32:47,710][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.07364201545715332, 'train/grad_norm_pre_clip': 29.654983520507812, 'timing/data': 0.0014044660201761872, 'timing/model': 2.0382201869797427, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 1.31}) +[2026-06-23 11:32:51,750][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.04724365472793579, 'train/grad_norm_pre_clip': 11.637700080871582, 'timing/data': 0.0011932049819733948, 'timing/model': 2.048035795975011, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 1.31}) +[2026-06-23 11:32:55,833][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.005051781423389912, 'train/grad_norm_pre_clip': 2.1206846237182617, 'timing/data': 0.057977845979621634, 'timing/model': 2.0333562770101707, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 1.31}) +[2026-06-23 11:32:59,950][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.08004426211118698, 'train/grad_norm_pre_clip': 44.11249542236328, 'timing/data': 0.0012577839952427894, 'timing/model': 2.0321864949946757, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 1.31}) +[2026-06-23 11:33:03,978][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.017778825014829636, 'train/grad_norm_pre_clip': 9.787433624267578, 'timing/data': 0.0016795140109024942, 'timing/model': 2.032532033015741, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 1.31}) +[2026-06-23 11:33:07,996][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.08883500099182129, 'train/grad_norm_pre_clip': 23.850088119506836, 'timing/data': 0.001203123974846676, 'timing/model': 2.0334726810106076, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 1.31}) +[2026-06-23 11:33:12,076][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.0385994017124176, 'train/grad_norm_pre_clip': 15.551989555358887, 'timing/data': 0.05672527098795399, 'timing/model': 2.034780096000759, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 1.31}) +[2026-06-23 11:33:16,386][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.14436331391334534, 'train/grad_norm_pre_clip': 37.698768615722656, 'timing/data': 0.0011273839918430895, 'timing/model': 2.037686836993089, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 1.31}) +[2026-06-23 11:33:20,403][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.18733276426792145, 'train/grad_norm_pre_clip': 12.156647682189941, 'timing/data': 0.001168339018477127, 'timing/model': 2.030956689006416, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 1.32}) +[2026-06-23 11:33:24,419][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.09096105396747589, 'train/grad_norm_pre_clip': 24.72828483581543, 'timing/data': 0.0012125810026191175, 'timing/model': 2.0306705299881287, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 1.32}) +[2026-06-23 11:33:28,501][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3400 +[2026-06-23 11:35:22,622][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3400 +[2026-06-23 11:35:22,623][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3400 +[2026-06-23 11:37:29,850][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3400 +[2026-06-23 11:37:29,852][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.053437989205121994, 'train/grad_norm_pre_clip': 21.389347076416016, 'eval/action_loss/seconds': 114.1170470250072, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.040715910494327545, 'eval/latency_2/loss': 0.040715910494327545, 'eval/flappy/loss': 0.040715910494327545, 'eval/flappy/latency_2/loss': 0.040715910494327545, 'eval/flap_precision': 0.959541469993257, 'eval/flap_recall': 0.9601889338731444, 'eval/flap_f1': 0.9598650927487353, 'eval/flap_rate_model': 0.20597222222222222, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9598650927487353, 'eval/latency_2/flap_precision': 0.959541469993257, 'eval/latency_2/flap_recall': 0.9601889338731444, 'eval/latency_2/flap_f1': 0.9598650927487353, 'eval/latency_2/flap_rate_model': 0.20597222222222222, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9598650927487353, 'eval/action_classification/data_wait_seconds': 50.58713150024414, 'eval/action_classification/predict_seconds': 74.99097442626953, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06741974499891512, 'timing/model': 2.0297373740177136, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 1.32}) +[2026-06-23 11:37:33,895][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.03276849538087845, 'train/grad_norm_pre_clip': 23.30104637145996, 'timing/data': 0.0011907300213351846, 'timing/model': 1.99559666600544, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 1.32}) +[2026-06-23 11:37:38,096][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.0748615562915802, 'train/grad_norm_pre_clip': 23.606388092041016, 'timing/data': 0.0012382389977574348, 'timing/model': 2.24561610998353, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 1.32}) +[2026-06-23 11:37:42,059][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.09628049284219742, 'train/grad_norm_pre_clip': 36.442138671875, 'timing/data': 0.001227267988724634, 'timing/model': 2.0012872820079792, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 1.32}) +[2026-06-23 11:37:46,077][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.052483152598142624, 'train/grad_norm_pre_clip': 18.049734115600586, 'timing/data': 0.05913801898714155, 'timing/model': 1.9977681880118325, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 1.32}) +[2026-06-23 11:37:50,116][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.08642424643039703, 'train/grad_norm_pre_clip': 34.11799240112305, 'timing/data': 0.0011187279887963086, 'timing/model': 1.998272179014748, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 1.32}) +[2026-06-23 11:37:54,077][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.029043206945061684, 'train/grad_norm_pre_clip': 15.849272727966309, 'timing/data': 0.00114162900717929, 'timing/model': 2.0017403879901394, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 1.32}) +[2026-06-23 11:37:58,047][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.03572666645050049, 'train/grad_norm_pre_clip': 12.451371192932129, 'timing/data': 0.0011721359915100038, 'timing/model': 2.0093343119951896, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 1.32}) +[2026-06-23 11:38:02,082][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.08969376981258392, 'train/grad_norm_pre_clip': 14.347169876098633, 'timing/data': 0.059571269986918196, 'timing/model': 2.013420914008748, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 1.32}) +[2026-06-23 11:38:06,162][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.02015051059424877, 'train/grad_norm_pre_clip': 5.514662742614746, 'timing/data': 0.0011990650091320276, 'timing/model': 2.028252321993932, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 1.32}) +[2026-06-23 11:38:10,182][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.023858830332756042, 'train/grad_norm_pre_clip': 8.450900077819824, 'timing/data': 0.0013007439847569913, 'timing/model': 2.04940854900633, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 1.32}) +[2026-06-23 11:38:14,183][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.04684044420719147, 'train/grad_norm_pre_clip': 27.30927276611328, 'timing/data': 0.0013692100183106959, 'timing/model': 2.031118788989261, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 1.32}) +[2026-06-23 11:38:18,246][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.06021828204393387, 'train/grad_norm_pre_clip': 18.24180793762207, 'timing/data': 0.07150560498121195, 'timing/model': 2.0121530090109445, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 1.32}) +[2026-06-23 11:38:22,313][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.06797953695058823, 'train/grad_norm_pre_clip': 25.510862350463867, 'timing/data': 0.0011567070032469928, 'timing/model': 2.0113309190201107, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 1.32}) +[2026-06-23 11:38:26,297][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.05666666477918625, 'train/grad_norm_pre_clip': 16.63627815246582, 'timing/data': 0.001176603982457891, 'timing/model': 2.01510153801064, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 1.32}) +[2026-06-23 11:38:30,306][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.07667578011751175, 'train/grad_norm_pre_clip': 18.560678482055664, 'timing/data': 0.0015900479920674115, 'timing/model': 2.032001888990635, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 1.32}) +[2026-06-23 11:38:34,350][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.024017076939344406, 'train/grad_norm_pre_clip': 11.891372680664062, 'timing/data': 0.05705703399144113, 'timing/model': 2.0145705850154627, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 1.32}) +[2026-06-23 11:38:38,427][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.14933760464191437, 'train/grad_norm_pre_clip': 26.073280334472656, 'timing/data': 0.0011202699970453978, 'timing/model': 2.0161583900044207, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 1.32}) +[2026-06-23 11:38:42,420][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.07767916470766068, 'train/grad_norm_pre_clip': 25.16992950439453, 'timing/data': 0.0011570279893931001, 'timing/model': 2.01745577697875, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 1.32}) +[2026-06-23 11:38:46,569][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.02740694023668766, 'train/grad_norm_pre_clip': 10.681449890136719, 'timing/data': 0.0011619679862633348, 'timing/model': 2.017256809020182, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 1.32}) +[2026-06-23 11:38:50,619][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.08145638555288315, 'train/grad_norm_pre_clip': 9.663556098937988, 'timing/data': 0.05513210600474849, 'timing/model': 2.0199104929924943, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 1.32}) +[2026-06-23 11:38:54,693][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.05200662836432457, 'train/grad_norm_pre_clip': 15.32765007019043, 'timing/data': 0.0010963859967887402, 'timing/model': 2.017475403001299, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 1.32}) +[2026-06-23 11:38:58,683][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.004229512065649033, 'train/grad_norm_pre_clip': 2.2959139347076416, 'timing/data': 0.0011634080146905035, 'timing/model': 2.01840911002364, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 1.32}) +[2026-06-23 11:39:02,682][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.07596372812986374, 'train/grad_norm_pre_clip': 22.269351959228516, 'timing/data': 0.0011121339921373874, 'timing/model': 2.019018174993107, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 1.32}) +[2026-06-23 11:39:06,744][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.12025518715381622, 'train/grad_norm_pre_clip': 133.00306701660156, 'timing/data': 0.06010091898497194, 'timing/model': 2.025014847982675, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 1.33}) +[2026-06-23 11:39:10,837][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.07232887297868729, 'train/grad_norm_pre_clip': 53.26993942260742, 'timing/data': 0.001107775984564796, 'timing/model': 2.0233234539919067, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 1.33}) +[2026-06-23 11:39:14,838][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.04003129526972771, 'train/grad_norm_pre_clip': 22.98946762084961, 'timing/data': 0.001421905995812267, 'timing/model': 2.0223517989797983, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 1.33}) +[2026-06-23 11:39:18,874][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.04023586958646774, 'train/grad_norm_pre_clip': 15.203486442565918, 'timing/data': 0.001334172993665561, 'timing/model': 2.056447337992722, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 1.33}) +[2026-06-23 11:39:22,934][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.03499579057097435, 'train/grad_norm_pre_clip': 13.565339088439941, 'timing/data': 0.056438188010361046, 'timing/model': 2.0219867640116718, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 1.33}) +[2026-06-23 11:39:27,024][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.14947509765625, 'train/grad_norm_pre_clip': 28.877779006958008, 'timing/data': 0.0011245060013607144, 'timing/model': 2.0266311689920258, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 1.33}) +[2026-06-23 11:39:31,029][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.03150032460689545, 'train/grad_norm_pre_clip': 12.656983375549316, 'timing/data': 0.0012319649977143854, 'timing/model': 2.0226473710208666, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 1.33}) +[2026-06-23 11:39:35,037][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.09915435314178467, 'train/grad_norm_pre_clip': 53.82501220703125, 'timing/data': 0.0012339690001681447, 'timing/model': 2.027715659001842, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 1.33}) +[2026-06-23 11:39:39,137][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.24985484778881073, 'train/grad_norm_pre_clip': 72.27064514160156, 'timing/data': 0.08822622001753189, 'timing/model': 2.0300764980202075, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.33}) +[2026-06-23 11:39:43,243][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.04928576946258545, 'train/grad_norm_pre_clip': 20.389673233032227, 'timing/data': 0.0011885050043929368, 'timing/model': 2.0365660609968472, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 1.33}) +[2026-06-23 11:39:47,262][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.006853380240499973, 'train/grad_norm_pre_clip': 3.11134672164917, 'timing/data': 0.0014541860145982355, 'timing/model': 2.038422376004746, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 1.33}) +[2026-06-23 11:39:51,271][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.05113857239484787, 'train/grad_norm_pre_clip': 9.57263469696045, 'timing/data': 0.0011034480121452361, 'timing/model': 2.027658476989018, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 1.33}) +[2026-06-23 11:39:55,419][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.04543745145201683, 'train/grad_norm_pre_clip': 15.163094520568848, 'timing/data': 0.057755054003791884, 'timing/model': 2.0254630809940863, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 1.33}) +[2026-06-23 11:39:59,510][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.07771287858486176, 'train/grad_norm_pre_clip': 29.2822322845459, 'timing/data': 0.0012997199955862015, 'timing/model': 2.0236721630208194, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 1.33}) +[2026-06-23 11:40:03,520][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.0696895644068718, 'train/grad_norm_pre_clip': 20.660808563232422, 'timing/data': 0.0011769330012612045, 'timing/model': 2.025140190991806, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 1.33}) +[2026-06-23 11:40:07,532][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.12178868055343628, 'train/grad_norm_pre_clip': 30.604860305786133, 'timing/data': 0.0011225630005355924, 'timing/model': 2.025857894012006, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 1.33}) +[2026-06-23 11:40:11,609][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.2611754536628723, 'train/grad_norm_pre_clip': 54.6146354675293, 'timing/data': 0.06372134399134666, 'timing/model': 2.0293938310060184, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 1.33}) +[2026-06-23 11:40:15,700][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.1361648142337799, 'train/grad_norm_pre_clip': 43.33692169189453, 'timing/data': 0.0011157509870827198, 'timing/model': 2.0260410369955935, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 1.33}) +[2026-06-23 11:40:19,711][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.03774895519018173, 'train/grad_norm_pre_clip': 21.969789505004883, 'timing/data': 0.0011844670225400478, 'timing/model': 2.0280268210044596, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 1.33}) +[2026-06-23 11:40:23,733][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.0388912558555603, 'train/grad_norm_pre_clip': 17.456554412841797, 'timing/data': 0.0013781960005871952, 'timing/model': 2.03368862401112, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 1.33}) +[2026-06-23 11:40:28,014][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.15510089695453644, 'train/grad_norm_pre_clip': 33.95184326171875, 'timing/data': 0.06179198299651034, 'timing/model': 2.230045278003672, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 1.33}) +[2026-06-23 11:40:32,138][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.07865001261234283, 'train/grad_norm_pre_clip': 32.26609420776367, 'timing/data': 0.001300982985412702, 'timing/model': 2.047275801975047, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 1.33}) +[2026-06-23 11:40:36,167][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.10585787147283554, 'train/grad_norm_pre_clip': 31.806772232055664, 'timing/data': 0.0012286290002521127, 'timing/model': 2.033393721008906, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 1.33}) +[2026-06-23 11:40:40,203][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.148443341255188, 'train/grad_norm_pre_clip': 26.817962646484375, 'timing/data': 0.0015003809821791947, 'timing/model': 2.046997149998788, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 1.33}) +[2026-06-23 11:40:44,294][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.07461043447256088, 'train/grad_norm_pre_clip': 32.66267776489258, 'timing/data': 0.06889342900831252, 'timing/model': 2.031521357013844, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 1.33}) +[2026-06-23 11:40:48,425][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.1489664912223816, 'train/grad_norm_pre_clip': 50.448707580566406, 'timing/data': 0.0016919150075409561, 'timing/model': 2.050248297979124, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 1.33}) +[2026-06-23 11:40:52,450][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.11691666394472122, 'train/grad_norm_pre_clip': 35.12165069580078, 'timing/data': 0.0012241599906701595, 'timing/model': 2.0336803069803864, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 1.34}) +[2026-06-23 11:40:56,473][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.01854058913886547, 'train/grad_norm_pre_clip': 10.32084846496582, 'timing/data': 0.0013028760149609298, 'timing/model': 2.0338761339953635, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 1.34}) +[2026-06-23 11:41:00,582][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.07633823156356812, 'train/grad_norm_pre_clip': 29.58748435974121, 'timing/data': 0.05600489801145159, 'timing/model': 2.061463528982131, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 1.34}) +[2026-06-23 11:41:04,679][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.019920650869607925, 'train/grad_norm_pre_clip': 8.472248077392578, 'timing/data': 0.0012565299984999, 'timing/model': 2.0289619010000024, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 1.34}) +[2026-06-23 11:41:08,699][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.18861617147922516, 'train/grad_norm_pre_clip': 38.79635238647461, 'timing/data': 0.0012096340069547296, 'timing/model': 2.0313576399930753, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 1.34}) +[2026-06-23 11:41:12,738][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.06121175363659859, 'train/grad_norm_pre_clip': 25.274688720703125, 'timing/data': 0.0011573279916774482, 'timing/model': 2.052831143984804, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 1.34}) +[2026-06-23 11:41:16,810][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.07516281306743622, 'train/grad_norm_pre_clip': 28.54477310180664, 'timing/data': 0.05409855998004787, 'timing/model': 2.029621836001752, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 1.34}) +[2026-06-23 11:41:20,913][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.24088822305202484, 'train/grad_norm_pre_clip': 29.423898696899414, 'timing/data': 0.001135647005867213, 'timing/model': 2.0285568550170865, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 1.34}) +[2026-06-23 11:41:24,934][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.058624424040317535, 'train/grad_norm_pre_clip': 12.289846420288086, 'timing/data': 0.0011492929770611227, 'timing/model': 2.0335720920120366, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 1.34}) +[2026-06-23 11:41:28,952][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.1080794408917427, 'train/grad_norm_pre_clip': 26.683013916015625, 'timing/data': 0.0011263299966230989, 'timing/model': 2.0284985069883987, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 1.34}) +[2026-06-23 11:41:33,027][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.04440883919596672, 'train/grad_norm_pre_clip': 11.891637802124023, 'timing/data': 0.05886263001593761, 'timing/model': 2.029291317012394, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 1.34}) +[2026-06-23 11:41:37,286][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.0592038556933403, 'train/grad_norm_pre_clip': 18.45822525024414, 'timing/data': 0.0011187059863004833, 'timing/model': 2.027451675006887, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 1.34}) +[2026-06-23 11:41:41,304][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.050046760588884354, 'train/grad_norm_pre_clip': 16.054670333862305, 'timing/data': 0.0011464769777376205, 'timing/model': 2.0313785719918087, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 1.34}) +[2026-06-23 11:41:45,324][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.024578757584095, 'train/grad_norm_pre_clip': 15.50779914855957, 'timing/data': 0.001218139979755506, 'timing/model': 2.0297837000107393, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 1.34}) +[2026-06-23 11:41:49,410][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.058816440403461456, 'train/grad_norm_pre_clip': 11.992100715637207, 'timing/data': 0.06943616000353359, 'timing/model': 2.0287518290278967, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 1.34}) +[2026-06-23 11:41:53,514][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.0434369221329689, 'train/grad_norm_pre_clip': 19.235031127929688, 'timing/data': 0.001189717004308477, 'timing/model': 2.029475499992259, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 1.34}) +[2026-06-23 11:41:57,532][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.06233048439025879, 'train/grad_norm_pre_clip': 21.22968864440918, 'timing/data': 0.0012775999784935266, 'timing/model': 2.029747827007668, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 1.34}) +[2026-06-23 11:42:01,565][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.05213439464569092, 'train/grad_norm_pre_clip': 15.811859130859375, 'timing/data': 0.0012130899995099753, 'timing/model': 2.034097745985491, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 1.34}) +[2026-06-23 11:42:05,675][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.04692985862493515, 'train/grad_norm_pre_clip': 20.294801712036133, 'timing/data': 0.06594481400679797, 'timing/model': 2.0346307939908, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 1.34}) +[2026-06-23 11:42:10,027][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.024483738467097282, 'train/grad_norm_pre_clip': 10.38353157043457, 'timing/data': 0.0012234500027261674, 'timing/model': 2.034055577008985, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 1.34}) +[2026-06-23 11:42:14,046][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.060811951756477356, 'train/grad_norm_pre_clip': 18.67218780517578, 'timing/data': 0.0011716439912561327, 'timing/model': 2.0298903769871686, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 1.34}) +[2026-06-23 11:42:18,065][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.048328980803489685, 'train/grad_norm_pre_clip': 15.761380195617676, 'timing/data': 0.001109308999730274, 'timing/model': 2.0335063800157513, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 1.34}) +[2026-06-23 11:42:22,150][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.06945390999317169, 'train/grad_norm_pre_clip': 41.73281478881836, 'timing/data': 0.055839862005086616, 'timing/model': 2.0317278990114573, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 1.34}) +[2026-06-23 11:42:26,253][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.06951165199279785, 'train/grad_norm_pre_clip': 18.112173080444336, 'timing/data': 0.001317533984547481, 'timing/model': 2.0309202909993473, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 1.34}) +[2026-06-23 11:42:30,271][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.09191448241472244, 'train/grad_norm_pre_clip': 20.94038200378418, 'timing/data': 0.0012230889988131821, 'timing/model': 2.0286665610037744, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 1.34}) +[2026-06-23 11:42:34,295][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.024960460141301155, 'train/grad_norm_pre_clip': 7.858176231384277, 'timing/data': 0.0011104909935966134, 'timing/model': 2.0346847389882896, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 1.34}) +[2026-06-23 11:42:38,371][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.0773831233382225, 'train/grad_norm_pre_clip': 19.673809051513672, 'timing/data': 0.05343966098735109, 'timing/model': 2.032589103007922, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 1.35}) +[2026-06-23 11:42:42,736][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.20338018238544464, 'train/grad_norm_pre_clip': 22.289215087890625, 'timing/data': 0.001290553977014497, 'timing/model': 2.2892927300126757, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 1.35}) +[2026-06-23 11:42:46,766][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.1450381875038147, 'train/grad_norm_pre_clip': 21.505565643310547, 'timing/data': 0.0012823480064980686, 'timing/model': 2.033608883008128, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 1.35}) +[2026-06-23 11:42:50,798][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.1438170224428177, 'train/grad_norm_pre_clip': 37.038414001464844, 'timing/data': 0.0012228090199641883, 'timing/model': 2.03678583400324, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 1.35}) +[2026-06-23 11:42:54,895][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.06462603807449341, 'train/grad_norm_pre_clip': 18.65619468688965, 'timing/data': 0.06739632901735604, 'timing/model': 2.0371106959937606, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 1.35}) +[2026-06-23 11:42:59,008][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.07895077019929886, 'train/grad_norm_pre_clip': 16.899850845336914, 'timing/data': 0.0013473279832396656, 'timing/model': 2.0333467370073777, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.35}) +[2026-06-23 11:43:03,035][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.039777133613824844, 'train/grad_norm_pre_clip': 13.140070915222168, 'timing/data': 0.0012679609935730696, 'timing/model': 2.03426205998403, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 1.35}) +[2026-06-23 11:43:07,055][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.06052963063120842, 'train/grad_norm_pre_clip': 30.98180389404297, 'timing/data': 0.0012392889766488224, 'timing/model': 2.028067392006051, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 1.35}) +[2026-06-23 11:43:11,137][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.03326109051704407, 'train/grad_norm_pre_clip': 12.865347862243652, 'timing/data': 0.06308596098097041, 'timing/model': 2.030348196014529, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 1.35}) +[2026-06-23 11:43:15,241][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.034775663167238235, 'train/grad_norm_pre_clip': 16.665016174316406, 'timing/data': 0.0011071150074712932, 'timing/model': 2.0310605749837123, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 1.35}) +[2026-06-23 11:43:19,365][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.010488157160580158, 'train/grad_norm_pre_clip': 3.7795121669769287, 'timing/data': 0.0014615190157201141, 'timing/model': 2.03564278097474, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 1.35}) +[2026-06-23 11:43:23,392][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.10896609723567963, 'train/grad_norm_pre_clip': 31.049489974975586, 'timing/data': 0.0011961800046265125, 'timing/model': 2.034740827017231, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 1.35}) +[2026-06-23 11:43:27,481][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.17412839829921722, 'train/grad_norm_pre_clip': 20.176132202148438, 'timing/data': 0.057690765999723226, 'timing/model': 2.0322289379837457, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 1.35}) +[2026-06-23 11:43:31,588][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.05965154618024826, 'train/grad_norm_pre_clip': 25.087507247924805, 'timing/data': 0.0012521229800768197, 'timing/model': 2.0324897929967847, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 1.35}) +[2026-06-23 11:43:35,616][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.03776072710752487, 'train/grad_norm_pre_clip': 15.385570526123047, 'timing/data': 0.001471447991207242, 'timing/model': 2.0374836690025404, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 1.35}) +[2026-06-23 11:43:39,725][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.06476448476314545, 'train/grad_norm_pre_clip': 20.627365112304688, 'timing/data': 0.0013603830011561513, 'timing/model': 2.049544006003998, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 1.35}) +[2026-06-23 11:43:43,905][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.11863464117050171, 'train/grad_norm_pre_clip': 35.91229248046875, 'timing/data': 0.09725815101410262, 'timing/model': 2.0317311770049855, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 1.35}) +[2026-06-23 11:43:48,020][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.18224631249904633, 'train/grad_norm_pre_clip': 35.144954681396484, 'timing/data': 0.0012213060108479112, 'timing/model': 2.0328138080076315, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 1.35}) +[2026-06-23 11:43:52,086][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.043667227029800415, 'train/grad_norm_pre_clip': 11.518552780151367, 'timing/data': 0.0013844469795003533, 'timing/model': 2.074031943018781, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 1.35}) +[2026-06-23 11:43:56,113][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.03097645379602909, 'train/grad_norm_pre_clip': 16.198902130126953, 'timing/data': 0.0012073300022166222, 'timing/model': 2.0353868840029463, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 1.35}) +[2026-06-23 11:44:00,222][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.0675576850771904, 'train/grad_norm_pre_clip': 14.410679817199707, 'timing/data': 0.075010831002146, 'timing/model': 2.0329013040172867, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 1.35}) +[2026-06-23 11:44:04,333][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.0532282330095768, 'train/grad_norm_pre_clip': 15.736372947692871, 'timing/data': 0.0014036319917067885, 'timing/model': 2.0338722790183965, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 1.35}) +[2026-06-23 11:44:08,363][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.04470960423350334, 'train/grad_norm_pre_clip': 12.825848579406738, 'timing/data': 0.0015013339871075004, 'timing/model': 2.036356327997055, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 1.35}) +[2026-06-23 11:44:12,393][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.06917497515678406, 'train/grad_norm_pre_clip': 10.623230934143066, 'timing/data': 0.0012830590130761266, 'timing/model': 2.036012964992551, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 1.35}) +[2026-06-23 11:44:16,486][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3500 +[2026-06-23 11:46:10,869][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3500 +[2026-06-23 11:46:10,870][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3500 +[2026-06-23 11:48:18,860][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3500 +[2026-06-23 12:05:04,008][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3500 +[2026-06-23 12:05:04,009][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.013316726312041283, 'train/grad_norm_pre_clip': 5.777832984924316, 'eval/action_loss/seconds': 114.38003973200102, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03848486766219139, 'eval/latency_2/loss': 0.03848486766219139, 'eval/flappy/loss': 0.03848486766219139, 'eval/flappy/latency_2/loss': 0.03848486766219139, 'eval/flap_precision': 0.9564050972501676, 'eval/flap_recall': 0.9622132253711201, 'eval/flap_f1': 0.9593003699966364, 'eval/flap_rate_model': 0.20708333333333334, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9593003699966363, 'eval/latency_2/flap_precision': 0.9564050972501676, 'eval/latency_2/flap_recall': 0.9622132253711201, 'eval/latency_2/flap_f1': 0.9593003699966364, 'eval/latency_2/flap_rate_model': 0.20708333333333334, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9593003699966363, 'eval/action_classification/data_wait_seconds': 50.790435791015625, 'eval/action_classification/predict_seconds': 75.5156478881836, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 260.5200000000122, 'rl_games_eval/mid_train/mean_length': 2119.0, 'rl_games_eval/mid_train/std_reward': 189.18299500748955, 'rl_games_eval/mid_train/std_length': 1526.6031573398504, 'rl_games_eval/mid_train/macro_mean_reward': 260.5200000000122, 'rl_games_eval/mid_train/macro_mean_length': 2119.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 260.5200000000122, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2119.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 189.18299500748955, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1526.6031573398504, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 260.5200000000122, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.06569330199272372, 'timing/model': 2.034688338986598, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 1.35}) +[2026-06-23 12:05:04,010][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_3500_state +[2026-06-23 12:05:04,019][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 12:05:12,657][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_3500_state/model.safetensors +[2026-06-23 12:05:16,603][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_3500_state/optimizer.bin +[2026-06-23 12:05:16,604][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_3500_state/sampler.bin +[2026-06-23 12:05:16,604][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-23 12:05:16,605][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-23 12:05:16,743][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-23 12:05:16,744][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 12:05:16,768][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 12:05:20,758][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.07692328840494156, 'train/grad_norm_pre_clip': 14.092016220092773, 'timing/data': 0.001372286002151668, 'timing/model': 1.9628670969977975, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 1.35}) +[2026-06-23 12:05:24,638][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.06265836954116821, 'train/grad_norm_pre_clip': 26.29193115234375, 'timing/data': 0.0012974379933439195, 'timing/model': 1.9634100109979045, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 1.36}) +[2026-06-23 12:05:28,522][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.07687869668006897, 'train/grad_norm_pre_clip': 22.87034034729004, 'timing/data': 0.0012174190196674317, 'timing/model': 1.9644587599905208, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 1.36}) +[2026-06-23 12:05:32,481][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.1400943547487259, 'train/grad_norm_pre_clip': 26.851762771606445, 'timing/data': 0.06587912599206902, 'timing/model': 1.9700752689968795, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 1.36}) +[2026-06-23 12:05:36,473][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.04991157725453377, 'train/grad_norm_pre_clip': 22.973669052124023, 'timing/data': 0.0014583239972125739, 'timing/model': 1.9684472809894942, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 1.36}) +[2026-06-23 12:05:40,628][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.053046032786369324, 'train/grad_norm_pre_clip': 14.895781517028809, 'timing/data': 0.0012906850024592131, 'timing/model': 1.9714751260180492, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 1.36}) +[2026-06-23 12:05:44,536][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.05055246502161026, 'train/grad_norm_pre_clip': 10.746458053588867, 'timing/data': 0.0012611810234375298, 'timing/model': 1.9784215519903228, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 1.36}) +[2026-06-23 12:05:48,517][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.022685986012220383, 'train/grad_norm_pre_clip': 9.988261222839355, 'timing/data': 0.06464451499050483, 'timing/model': 1.9823389909870457, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 1.36}) +[2026-06-23 12:05:52,539][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.07780175656080246, 'train/grad_norm_pre_clip': 20.126842498779297, 'timing/data': 0.0011977030080743134, 'timing/model': 1.9943975830101408, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 1.36}) +[2026-06-23 12:05:56,462][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.05247953534126282, 'train/grad_norm_pre_clip': 25.237672805786133, 'timing/data': 0.001644078001845628, 'timing/model': 1.9847091429983266, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 1.36}) +[2026-06-23 12:06:00,389][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.054298996925354004, 'train/grad_norm_pre_clip': 13.20081901550293, 'timing/data': 0.0011486319999676198, 'timing/model': 1.9851743220060598, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 1.36}) +[2026-06-23 12:06:04,389][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.12202899158000946, 'train/grad_norm_pre_clip': 40.0155143737793, 'timing/data': 0.07025151498964988, 'timing/model': 1.984840336983325, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 1.36}) +[2026-06-23 12:06:08,411][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.03127901256084442, 'train/grad_norm_pre_clip': 15.677154541015625, 'timing/data': 0.0012443790037650615, 'timing/model': 1.9878416929859668, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 1.36}) +[2026-06-23 12:06:12,594][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.11304342746734619, 'train/grad_norm_pre_clip': 31.76703453063965, 'timing/data': 0.0011726559896487743, 'timing/model': 1.9895121800072957, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 1.36}) +[2026-06-23 12:06:16,534][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.12077884376049042, 'train/grad_norm_pre_clip': 23.343339920043945, 'timing/data': 0.001212811010191217, 'timing/model': 1.9913617690035608, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 1.36}) +[2026-06-23 12:06:20,531][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.015484927222132683, 'train/grad_norm_pre_clip': 7.88925838470459, 'timing/data': 0.05706331698456779, 'timing/model': 1.994355381000787, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 1.36}) +[2026-06-23 12:06:24,559][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.022709693759679794, 'train/grad_norm_pre_clip': 14.853835105895996, 'timing/data': 0.0011399259965401143, 'timing/model': 1.9921278859837912, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 1.36}) +[2026-06-23 12:06:28,504][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.1305290311574936, 'train/grad_norm_pre_clip': 33.618412017822266, 'timing/data': 0.0011736589949578047, 'timing/model': 1.9945228619908448, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 1.36}) +[2026-06-23 12:06:32,469][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.04826139286160469, 'train/grad_norm_pre_clip': 15.207293510437012, 'timing/data': 0.0011561959981918335, 'timing/model': 2.011416156979976, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 1.36}) +[2026-06-23 12:06:36,485][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.03629421442747116, 'train/grad_norm_pre_clip': 18.802352905273438, 'timing/data': 0.0576767009915784, 'timing/model': 1.9994369029882364, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 1.36}) +[2026-06-23 12:06:40,538][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.062285397201776505, 'train/grad_norm_pre_clip': 22.888925552368164, 'timing/data': 0.001313015993218869, 'timing/model': 2.0038573020137846, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 1.36}) +[2026-06-23 12:06:44,593][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.09071249514818192, 'train/grad_norm_pre_clip': 19.673694610595703, 'timing/data': 0.0011613859969656914, 'timing/model': 2.0181499120080844, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 1.36}) +[2026-06-23 12:06:48,561][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.007621231488883495, 'train/grad_norm_pre_clip': 3.059645652770996, 'timing/data': 0.001121451990911737, 'timing/model': 2.00433357298607, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 1.36}) +[2026-06-23 12:06:52,607][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.1574815809726715, 'train/grad_norm_pre_clip': 27.559354782104492, 'timing/data': 0.05905168200843036, 'timing/model': 2.013254066987429, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 1.36}) +[2026-06-23 12:06:56,684][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.06722511351108551, 'train/grad_norm_pre_clip': 29.009912490844727, 'timing/data': 0.0013024170184507966, 'timing/model': 2.0192110339994542, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 1.36}) +[2026-06-23 12:07:00,666][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.07917866110801697, 'train/grad_norm_pre_clip': 23.336957931518555, 'timing/data': 0.0013221029948908836, 'timing/model': 2.0134567829954904, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 1.36}) +[2026-06-23 12:07:04,667][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.10463398694992065, 'train/grad_norm_pre_clip': 18.34868049621582, 'timing/data': 0.0012152960116509348, 'timing/model': 2.0137052729842253, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 1.37}) +[2026-06-23 12:07:08,719][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.07647044211626053, 'train/grad_norm_pre_clip': 21.110536575317383, 'timing/data': 0.07021582999732345, 'timing/model': 2.0157831620017532, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 1.37}) +[2026-06-23 12:07:12,816][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.12213371694087982, 'train/grad_norm_pre_clip': 23.093658447265625, 'timing/data': 0.0015856800018809736, 'timing/model': 2.0167061089887284, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 1.37}) +[2026-06-23 12:07:16,991][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.059939898550510406, 'train/grad_norm_pre_clip': 18.599960327148438, 'timing/data': 0.0012908450153190643, 'timing/model': 2.1974123559775762, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.37}) +[2026-06-23 12:07:21,015][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.02266893908381462, 'train/grad_norm_pre_clip': 10.973236083984375, 'timing/data': 0.0012453309900593013, 'timing/model': 2.043804580986034, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 1.37}) +[2026-06-23 12:07:25,088][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.041407886892557144, 'train/grad_norm_pre_clip': 13.194962501525879, 'timing/data': 0.06663924100575969, 'timing/model': 2.023422981990734, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 1.37}) +[2026-06-23 12:07:29,200][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.08993444591760635, 'train/grad_norm_pre_clip': 25.063472747802734, 'timing/data': 0.001358650013571605, 'timing/model': 2.0256421029916964, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 1.37}) +[2026-06-23 12:07:33,220][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.025785479694604874, 'train/grad_norm_pre_clip': 11.646110534667969, 'timing/data': 0.0013200989924371243, 'timing/model': 2.0343450650107116, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 1.37}) +[2026-06-23 12:07:37,226][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.021284742280840874, 'train/grad_norm_pre_clip': 12.851999282836914, 'timing/data': 0.0012914869876112789, 'timing/model': 2.0244371379958466, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 1.37}) +[2026-06-23 12:07:41,289][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.17129744589328766, 'train/grad_norm_pre_clip': 47.75614929199219, 'timing/data': 0.05557365500135347, 'timing/model': 2.025492759014014, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 1.37}) +[2026-06-23 12:07:45,401][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.049796491861343384, 'train/grad_norm_pre_clip': 13.439488410949707, 'timing/data': 0.0012362739944364876, 'timing/model': 2.0244118419941515, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 1.37}) +[2026-06-23 12:07:49,417][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.09566917270421982, 'train/grad_norm_pre_clip': 20.985387802124023, 'timing/data': 0.0012732720060739666, 'timing/model': 2.0292838950117584, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 1.37}) +[2026-06-23 12:07:53,486][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.1136578842997551, 'train/grad_norm_pre_clip': 24.60535430908203, 'timing/data': 0.0012034229876007885, 'timing/model': 2.024622814002214, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 1.37}) +[2026-06-23 12:07:57,575][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.02432485669851303, 'train/grad_norm_pre_clip': 9.527112007141113, 'timing/data': 0.0637865909957327, 'timing/model': 2.0288054829870816, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 1.37}) +[2026-06-23 12:08:01,677][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.1085427775979042, 'train/grad_norm_pre_clip': 24.887374877929688, 'timing/data': 0.0013630790053866804, 'timing/model': 2.0300535899878014, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 1.37}) +[2026-06-23 12:08:05,721][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.05900388956069946, 'train/grad_norm_pre_clip': 32.573856353759766, 'timing/data': 0.00125709202256985, 'timing/model': 2.0563581340247765, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 1.37}) +[2026-06-23 12:08:09,761][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.04953620955348015, 'train/grad_norm_pre_clip': 17.998950958251953, 'timing/data': 0.00161396199837327, 'timing/model': 2.043420350004453, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 1.37}) +[2026-06-23 12:08:13,877][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.02362433262169361, 'train/grad_norm_pre_clip': 14.254525184631348, 'timing/data': 0.08734970900695771, 'timing/model': 2.036407261999557, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 1.37}) +[2026-06-23 12:08:17,981][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.06850048899650574, 'train/grad_norm_pre_clip': 33.838829040527344, 'timing/data': 0.0012322459951974452, 'timing/model': 2.0291297850199044, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 1.37}) +[2026-06-23 12:08:22,001][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.032132796943187714, 'train/grad_norm_pre_clip': 41.612911224365234, 'timing/data': 0.001404154987540096, 'timing/model': 2.0257769209856633, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 1.37}) +[2026-06-23 12:08:26,013][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.06083391606807709, 'train/grad_norm_pre_clip': 19.132837295532227, 'timing/data': 0.0011940659896936268, 'timing/model': 2.028292089002207, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 1.37}) +[2026-06-23 12:08:30,269][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.09774190187454224, 'train/grad_norm_pre_clip': 25.538890838623047, 'timing/data': 0.06499571501626633, 'timing/model': 2.030631181987701, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 1.37}) +[2026-06-23 12:08:34,366][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.10635971277952194, 'train/grad_norm_pre_clip': 21.926713943481445, 'timing/data': 0.0012578839960042387, 'timing/model': 2.024922464013798, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 1.37}) +[2026-06-23 12:08:38,389][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.013575350865721703, 'train/grad_norm_pre_clip': 5.528584957122803, 'timing/data': 0.0013694810040760785, 'timing/model': 2.0348290789988823, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 1.37}) +[2026-06-23 12:08:42,416][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.012962096370756626, 'train/grad_norm_pre_clip': 4.850608825683594, 'timing/data': 0.00121130800107494, 'timing/model': 2.041193570010364, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 1.37}) +[2026-06-23 12:08:46,508][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.04439004883170128, 'train/grad_norm_pre_clip': 15.559050559997559, 'timing/data': 0.06985977198928595, 'timing/model': 2.0284255589940585, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 1.37}) +[2026-06-23 12:08:50,635][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.20956118404865265, 'train/grad_norm_pre_clip': 26.537185668945312, 'timing/data': 0.001470106013584882, 'timing/model': 2.0476357449952047, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 1.38}) +[2026-06-23 12:08:54,657][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.04649879038333893, 'train/grad_norm_pre_clip': 12.215088844299316, 'timing/data': 0.0013868329988326877, 'timing/model': 2.0355772839975543, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 1.38}) +[2026-06-23 12:08:58,674][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.13392865657806396, 'train/grad_norm_pre_clip': 53.518985748291016, 'timing/data': 0.0012099759769625962, 'timing/model': 2.0295288790075574, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 1.38}) +[2026-06-23 12:09:02,792][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.03117401711642742, 'train/grad_norm_pre_clip': 17.904598236083984, 'timing/data': 0.05871504999231547, 'timing/model': 2.030150557984598, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 1.38}) +[2026-06-23 12:09:06,900][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.11918478459119797, 'train/grad_norm_pre_clip': 32.82353210449219, 'timing/data': 0.0015712829772382975, 'timing/model': 2.032532370008994, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 1.38}) +[2026-06-23 12:09:10,924][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.01677512377500534, 'train/grad_norm_pre_clip': 10.086813926696777, 'timing/data': 0.0012516030110418797, 'timing/model': 2.0338317310088314, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 1.38}) +[2026-06-23 12:09:14,963][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.03116828203201294, 'train/grad_norm_pre_clip': 14.738821029663086, 'timing/data': 0.0013514769962057471, 'timing/model': 2.0485177179798484, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 1.38}) +[2026-06-23 12:09:19,053][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.2906375825405121, 'train/grad_norm_pre_clip': 29.918989181518555, 'timing/data': 0.06658492097631097, 'timing/model': 2.0310428199882153, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 1.38}) +[2026-06-23 12:09:23,177][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.0642649233341217, 'train/grad_norm_pre_clip': 19.214754104614258, 'timing/data': 0.0012352230260148644, 'timing/model': 2.0471001099795103, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 1.38}) +[2026-06-23 12:09:27,205][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.18467974662780762, 'train/grad_norm_pre_clip': 40.113670349121094, 'timing/data': 0.0012475050170905888, 'timing/model': 2.030414809996728, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 1.38}) +[2026-06-23 12:09:31,243][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.026665600016713142, 'train/grad_norm_pre_clip': 8.878432273864746, 'timing/data': 0.0012448999914340675, 'timing/model': 2.0354722459742334, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 1.38}) +[2026-06-23 12:09:35,394][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.02432827465236187, 'train/grad_norm_pre_clip': 13.036067962646484, 'timing/data': 0.05856535301427357, 'timing/model': 2.103071291989181, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 1.38}) +[2026-06-23 12:09:39,547][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.06487230211496353, 'train/grad_norm_pre_clip': 17.065107345581055, 'timing/data': 0.0012852950021624565, 'timing/model': 2.0680064810148906, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 1.38}) +[2026-06-23 12:09:43,566][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.015221704728901386, 'train/grad_norm_pre_clip': 12.261626243591309, 'timing/data': 0.001168159011285752, 'timing/model': 2.031479109020438, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 1.38}) +[2026-06-23 12:09:47,588][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.10698233544826508, 'train/grad_norm_pre_clip': 26.37401008605957, 'timing/data': 0.0011451159953139722, 'timing/model': 2.0325737520179246, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 1.38}) +[2026-06-23 12:09:51,670][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.16679678857326508, 'train/grad_norm_pre_clip': 26.685474395751953, 'timing/data': 0.059212539985310286, 'timing/model': 2.0301598419900984, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 1.38}) +[2026-06-23 12:09:55,785][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.081133633852005, 'train/grad_norm_pre_clip': 19.731307983398438, 'timing/data': 0.0012423849839251488, 'timing/model': 2.0336579370195977, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 1.38}) +[2026-06-23 12:09:59,810][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.0933203250169754, 'train/grad_norm_pre_clip': 20.908292770385742, 'timing/data': 0.0014513720234390348, 'timing/model': 2.0329980879905634, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 1.38}) +[2026-06-23 12:10:03,833][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.033287640661001205, 'train/grad_norm_pre_clip': 18.647781372070312, 'timing/data': 0.0014789620181545615, 'timing/model': 2.0341731290100142, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 1.38}) +[2026-06-23 12:10:07,929][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.03945352882146835, 'train/grad_norm_pre_clip': 15.27280330657959, 'timing/data': 0.070946501014987, 'timing/model': 2.0333099239796866, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 1.38}) +[2026-06-23 12:10:12,074][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.03719310835003853, 'train/grad_norm_pre_clip': 17.886138916015625, 'timing/data': 0.0012403820001054555, 'timing/model': 2.034524999995483, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 1.38}) +[2026-06-23 12:10:16,116][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.05463702231645584, 'train/grad_norm_pre_clip': 27.231380462646484, 'timing/data': 0.0014462629915215075, 'timing/model': 2.052155067009153, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 1.38}) +[2026-06-23 12:10:20,142][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.017748938873410225, 'train/grad_norm_pre_clip': 6.8282599449157715, 'timing/data': 0.0013518180057872087, 'timing/model': 2.034049904003041, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 1.38}) +[2026-06-23 12:10:24,229][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.024928737431764603, 'train/grad_norm_pre_clip': 9.720876693725586, 'timing/data': 0.05825121400994249, 'timing/model': 2.0364018699910957, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 1.38}) +[2026-06-23 12:10:28,342][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.015080897137522697, 'train/grad_norm_pre_clip': 19.635805130004883, 'timing/data': 0.0012279490183573216, 'timing/model': 2.032629490015097, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 1.38}) +[2026-06-23 12:10:32,370][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.06362511217594147, 'train/grad_norm_pre_clip': 15.054265022277832, 'timing/data': 0.0014483259874396026, 'timing/model': 2.0361011149943806, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 1.38}) +[2026-06-23 12:10:36,404][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.038523297756910324, 'train/grad_norm_pre_clip': 12.695777893066406, 'timing/data': 0.001262914011022076, 'timing/model': 2.0393548179999925, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.39}) +[2026-06-23 12:10:40,500][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.09520012140274048, 'train/grad_norm_pre_clip': 35.87255096435547, 'timing/data': 0.06748478498775512, 'timing/model': 2.0356888649985194, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 1.39}) +[2026-06-23 12:10:44,671][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.0354592390358448, 'train/grad_norm_pre_clip': 13.005870819091797, 'timing/data': 0.001289021980483085, 'timing/model': 2.0879130529938266, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 1.39}) +[2026-06-23 12:10:48,704][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.1051245629787445, 'train/grad_norm_pre_clip': 30.251489639282227, 'timing/data': 0.0012794539798051119, 'timing/model': 2.0384210239863023, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 1.39}) +[2026-06-23 12:10:52,737][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.039070967584848404, 'train/grad_norm_pre_clip': 12.257163047790527, 'timing/data': 0.0012315050116740167, 'timing/model': 2.038566643022932, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 1.39}) +[2026-06-23 12:10:56,827][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.005909123457968235, 'train/grad_norm_pre_clip': 2.8284072875976562, 'timing/data': 0.0580549509904813, 'timing/model': 2.037136692000786, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 1.39}) +[2026-06-23 12:11:00,963][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.1697603017091751, 'train/grad_norm_pre_clip': 36.907596588134766, 'timing/data': 0.001329676975728944, 'timing/model': 2.0527356950042304, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 1.39}) +[2026-06-23 12:11:04,997][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.06333672255277634, 'train/grad_norm_pre_clip': 21.361446380615234, 'timing/data': 0.0013543720124289393, 'timing/model': 2.039007629995467, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 1.39}) +[2026-06-23 12:11:09,031][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.02923186868429184, 'train/grad_norm_pre_clip': 10.931015014648438, 'timing/data': 0.0012014899984933436, 'timing/model': 2.0400077880185563, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 1.39}) +[2026-06-23 12:11:13,125][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.02855607122182846, 'train/grad_norm_pre_clip': 13.407480239868164, 'timing/data': 0.059641531988745555, 'timing/model': 2.0387763930193614, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 1.39}) +[2026-06-23 12:11:17,251][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.015933888033032417, 'train/grad_norm_pre_clip': 11.322576522827148, 'timing/data': 0.0013004119973629713, 'timing/model': 2.0408736990066245, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 1.39}) +[2026-06-23 12:11:21,450][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.04818984493613243, 'train/grad_norm_pre_clip': 30.044004440307617, 'timing/data': 0.0012373469944577664, 'timing/model': 2.039201690990012, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 1.39}) +[2026-06-23 12:11:25,503][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.07025794684886932, 'train/grad_norm_pre_clip': 28.33420181274414, 'timing/data': 0.001203232997795567, 'timing/model': 2.0563461289857514, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 1.39}) +[2026-06-23 12:11:29,601][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.1127137839794159, 'train/grad_norm_pre_clip': 27.447893142700195, 'timing/data': 0.05940261101932265, 'timing/model': 2.041946191980969, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 1.39}) +[2026-06-23 12:11:33,729][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.08767644315958023, 'train/grad_norm_pre_clip': 23.098108291625977, 'timing/data': 0.0012514120026025921, 'timing/model': 2.0366688719950616, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 1.39}) +[2026-06-23 12:11:37,763][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.07103818655014038, 'train/grad_norm_pre_clip': 39.33203125, 'timing/data': 0.0013109419960528612, 'timing/model': 2.0369555020006374, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 1.39}) +[2026-06-23 12:11:41,797][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.10071706771850586, 'train/grad_norm_pre_clip': 24.332035064697266, 'timing/data': 0.0012333290069364011, 'timing/model': 2.0381967859866563, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 1.39}) +[2026-06-23 12:11:45,885][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.017779069021344185, 'train/grad_norm_pre_clip': 8.920379638671875, 'timing/data': 0.059102345985593274, 'timing/model': 2.035522435005987, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 1.39}) +[2026-06-23 12:11:50,027][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.1020844504237175, 'train/grad_norm_pre_clip': 34.487422943115234, 'timing/data': 0.001606267993338406, 'timing/model': 2.0511306959961075, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 1.39}) +[2026-06-23 12:11:54,313][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.07381521165370941, 'train/grad_norm_pre_clip': 16.599281311035156, 'timing/data': 0.0013006329827476293, 'timing/model': 2.0405496239836793, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 1.39}) +[2026-06-23 12:11:58,361][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.10586540400981903, 'train/grad_norm_pre_clip': 19.383159637451172, 'timing/data': 0.0017519670072942972, 'timing/model': 2.05268038000213, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 1.39}) +[2026-06-23 12:12:02,470][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3600 +[2026-06-23 12:13:56,609][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3600 +[2026-06-23 12:13:56,610][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3600 +[2026-06-23 12:16:04,358][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3600 +[2026-06-23 12:16:04,359][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.0608338937163353, 'train/grad_norm_pre_clip': 27.33292579650879, 'eval/action_loss/seconds': 114.13548601799994, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04284502565860748, 'eval/latency_2/loss': 0.04284502565860748, 'eval/flappy/loss': 0.04284502565860748, 'eval/flappy/latency_2/loss': 0.04284502565860748, 'eval/flap_precision': 0.9510905485789821, 'eval/flap_recall': 0.9709851551956815, 'eval/flap_f1': 0.9609348914858097, 'eval/flap_rate_model': 0.21013888888888888, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9609348914858096, 'eval/latency_2/flap_precision': 0.9510905485789821, 'eval/latency_2/flap_recall': 0.9709851551956815, 'eval/latency_2/flap_f1': 0.9609348914858097, 'eval/latency_2/flap_rate_model': 0.21013888888888888, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9609348914858096, 'eval/action_classification/data_wait_seconds': 50.559757232666016, 'eval/action_classification/predict_seconds': 75.20392608642578, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06804230500711128, 'timing/model': 2.045827036985429, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 1.39}) +[2026-06-23 12:16:08,403][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.11200663447380066, 'train/grad_norm_pre_clip': 45.9266471862793, 'timing/data': 0.001220687001477927, 'timing/model': 1.9975207669776864, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 1.39}) +[2026-06-23 12:16:12,372][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.061989977955818176, 'train/grad_norm_pre_clip': 18.114505767822266, 'timing/data': 0.0012316580105107278, 'timing/model': 2.006538443005411, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 1.39}) +[2026-06-23 12:16:16,348][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.06543387472629547, 'train/grad_norm_pre_clip': 23.651811599731445, 'timing/data': 0.0012623640068341047, 'timing/model': 2.0155343280057423, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 1.39}) +[2026-06-23 12:16:20,416][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.023577285930514336, 'train/grad_norm_pre_clip': 13.351198196411133, 'timing/data': 0.06869982599164359, 'timing/model': 2.0021239279885776, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 1.39}) +[2026-06-23 12:16:24,496][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.16472752392292023, 'train/grad_norm_pre_clip': 95.49154663085938, 'timing/data': 0.0014916579821147025, 'timing/model': 2.0229500709974673, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 1.4}) +[2026-06-23 12:16:28,491][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.06850125640630722, 'train/grad_norm_pre_clip': 25.936662673950195, 'timing/data': 0.0012981210020370781, 'timing/model': 2.0259569819900207, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 1.4}) +[2026-06-23 12:16:32,484][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.08775446563959122, 'train/grad_norm_pre_clip': 19.94382095336914, 'timing/data': 0.0017955609946511686, 'timing/model': 2.0209046590025537, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 1.4}) +[2026-06-23 12:16:36,529][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.0793788954615593, 'train/grad_norm_pre_clip': 40.165950775146484, 'timing/data': 0.05922694501350634, 'timing/model': 2.016850963002071, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 1.4}) +[2026-06-23 12:16:40,601][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.048844464123249054, 'train/grad_norm_pre_clip': 18.596214294433594, 'timing/data': 0.0011985670134890825, 'timing/model': 2.014951571007259, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 1.4}) +[2026-06-23 12:16:44,611][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.06278803199529648, 'train/grad_norm_pre_clip': 18.07128143310547, 'timing/data': 0.0016840639873407781, 'timing/model': 2.0351737969904207, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 1.4}) +[2026-06-23 12:16:48,604][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.12063741683959961, 'train/grad_norm_pre_clip': 27.992794036865234, 'timing/data': 0.0011985769961029291, 'timing/model': 2.0205957089783624, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 1.4}) +[2026-06-23 12:16:52,924][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.25390031933784485, 'train/grad_norm_pre_clip': 32.77009582519531, 'timing/data': 0.06940973299788311, 'timing/model': 2.268350600003032, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 1.4}) +[2026-06-23 12:16:57,000][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.09708963334560394, 'train/grad_norm_pre_clip': 19.48298454284668, 'timing/data': 0.001219986006617546, 'timing/model': 2.019861682987539, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 1.4}) +[2026-06-23 12:17:01,006][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.0035147094167768955, 'train/grad_norm_pre_clip': 1.3154910802841187, 'timing/data': 0.0012514539994299412, 'timing/model': 2.026089150022017, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 1.4}) +[2026-06-23 12:17:05,010][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.10630617290735245, 'train/grad_norm_pre_clip': 19.166353225708008, 'timing/data': 0.0012091159878764302, 'timing/model': 2.0206989430007525, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 1.4}) +[2026-06-23 12:17:09,073][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.04459605738520622, 'train/grad_norm_pre_clip': 8.451817512512207, 'timing/data': 0.0659538799955044, 'timing/model': 2.0203924470115453, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 1.4}) +[2026-06-23 12:17:13,159][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.06456537544727325, 'train/grad_norm_pre_clip': 19.121450424194336, 'timing/data': 0.0012296040076762438, 'timing/model': 2.0198026900179684, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 1.4}) +[2026-06-23 12:17:17,164][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.05990910157561302, 'train/grad_norm_pre_clip': 18.131080627441406, 'timing/data': 0.0012274500040803105, 'timing/model': 2.024091291008517, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 1.4}) +[2026-06-23 12:17:21,167][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.054684024304151535, 'train/grad_norm_pre_clip': 26.947675704956055, 'timing/data': 0.0012116210127715021, 'timing/model': 2.022539292986039, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 1.4}) +[2026-06-23 12:17:25,430][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.03751283884048462, 'train/grad_norm_pre_clip': 17.608495712280273, 'timing/data': 0.0701365220011212, 'timing/model': 2.2093868700030725, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 1.4}) +[2026-06-23 12:17:29,522][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.018927184864878654, 'train/grad_norm_pre_clip': 4.953854560852051, 'timing/data': 0.0013762449962086976, 'timing/model': 2.026464363996638, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 1.4}) +[2026-06-23 12:17:33,529][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.03951755911111832, 'train/grad_norm_pre_clip': 25.442092895507812, 'timing/data': 0.0012520460004452616, 'timing/model': 2.0254256049811374, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 1.4}) +[2026-06-23 12:17:37,536][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.11722209304571152, 'train/grad_norm_pre_clip': 20.797937393188477, 'timing/data': 0.001229693996720016, 'timing/model': 2.024064802011708, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 1.4}) +[2026-06-23 12:17:41,614][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.006273963022977114, 'train/grad_norm_pre_clip': 4.221597671508789, 'timing/data': 0.06969445198774338, 'timing/model': 2.0248263420071453, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 1.4}) +[2026-06-23 12:17:45,709][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.06621766835451126, 'train/grad_norm_pre_clip': 18.322521209716797, 'timing/data': 0.0012018819979857653, 'timing/model': 2.0264418220031075, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 1.4}) +[2026-06-23 12:17:49,724][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.03914288058876991, 'train/grad_norm_pre_clip': 19.444364547729492, 'timing/data': 0.0012398829858284444, 'timing/model': 2.0314382829819806, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 1.4}) +[2026-06-23 12:17:53,750][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.009032695554196835, 'train/grad_norm_pre_clip': 5.126850128173828, 'timing/data': 0.0013637020019814372, 'timing/model': 2.0414721700071823, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 1.4}) +[2026-06-23 12:17:57,859][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.18009130656719208, 'train/grad_norm_pre_clip': 37.7358283996582, 'timing/data': 0.05990779399871826, 'timing/model': 2.0661943049926776, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.4}) +[2026-06-23 12:18:01,978][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.039856959134340286, 'train/grad_norm_pre_clip': 25.57551383972168, 'timing/data': 0.001201372011564672, 'timing/model': 2.049854383978527, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 1.4}) +[2026-06-23 12:18:05,995][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.01927340403199196, 'train/grad_norm_pre_clip': 56.47749328613281, 'timing/data': 0.0017377629992552102, 'timing/model': 2.0315934919926804, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 1.4}) +[2026-06-23 12:18:10,011][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.010709591209888458, 'train/grad_norm_pre_clip': 7.446915626525879, 'timing/data': 0.0012402829888742417, 'timing/model': 2.0302542000135873, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 1.41}) +[2026-06-23 12:18:14,098][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.04246089234948158, 'train/grad_norm_pre_clip': 22.527420043945312, 'timing/data': 0.06726048601558432, 'timing/model': 2.0302738189930096, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 1.41}) +[2026-06-23 12:18:18,229][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.13024237751960754, 'train/grad_norm_pre_clip': 17.035531997680664, 'timing/data': 0.0012061010056640953, 'timing/model': 2.054685189999873, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 1.41}) +[2026-06-23 12:18:22,257][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.12072578072547913, 'train/grad_norm_pre_clip': 14.676224708557129, 'timing/data': 0.0012286920100450516, 'timing/model': 2.036170568986563, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 1.41}) +[2026-06-23 12:18:26,290][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.1200311928987503, 'train/grad_norm_pre_clip': 26.23670196533203, 'timing/data': 0.0016159479855559766, 'timing/model': 2.0393932120059617, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 1.41}) +[2026-06-23 12:18:30,376][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.06142803281545639, 'train/grad_norm_pre_clip': 11.347532272338867, 'timing/data': 0.05760525199002586, 'timing/model': 2.037893141998211, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 1.41}) +[2026-06-23 12:18:34,625][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.06594252586364746, 'train/grad_norm_pre_clip': 19.813562393188477, 'timing/data': 0.0013345470069907606, 'timing/model': 2.0419124070031103, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 1.41}) +[2026-06-23 12:18:38,653][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.02360955998301506, 'train/grad_norm_pre_clip': 15.339953422546387, 'timing/data': 0.0012389809999149293, 'timing/model': 2.0345631430100184, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 1.41}) +[2026-06-23 12:18:42,681][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.14047671854496002, 'train/grad_norm_pre_clip': 48.827735900878906, 'timing/data': 0.0012229910062160343, 'timing/model': 2.0360324140056036, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 1.41}) +[2026-06-23 12:18:46,773][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.14113831520080566, 'train/grad_norm_pre_clip': 29.79052734375, 'timing/data': 0.06956791898119263, 'timing/model': 2.0351326140225865, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 1.41}) +[2026-06-23 12:18:50,900][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.1285480111837387, 'train/grad_norm_pre_clip': 42.4577522277832, 'timing/data': 0.001184410008136183, 'timing/model': 2.0504092540068086, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 1.41}) +[2026-06-23 12:18:54,926][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.11888117343187332, 'train/grad_norm_pre_clip': 24.495309829711914, 'timing/data': 0.0013218639942351729, 'timing/model': 2.0354787439864594, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 1.41}) +[2026-06-23 12:18:58,977][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.04659591242671013, 'train/grad_norm_pre_clip': 21.547733306884766, 'timing/data': 0.001348692981991917, 'timing/model': 2.0565486249979585, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 1.41}) +[2026-06-23 12:19:03,080][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.07326874881982803, 'train/grad_norm_pre_clip': 47.9156608581543, 'timing/data': 0.05699060199549422, 'timing/model': 2.0476536940259393, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 1.41}) +[2026-06-23 12:19:07,457][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.06919123232364655, 'train/grad_norm_pre_clip': 22.785017013549805, 'timing/data': 0.0014130729832686484, 'timing/model': 2.2958394460147247, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 1.41}) +[2026-06-23 12:19:11,502][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.057864390313625336, 'train/grad_norm_pre_clip': 17.97208023071289, 'timing/data': 0.001242096011992544, 'timing/model': 2.051330149988644, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 1.41}) +[2026-06-23 12:19:15,544][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.07791166007518768, 'train/grad_norm_pre_clip': 21.229310989379883, 'timing/data': 0.001415667007677257, 'timing/model': 2.0432837779808324, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 1.41}) +[2026-06-23 12:19:19,653][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.02426224760711193, 'train/grad_norm_pre_clip': 7.922924518585205, 'timing/data': 0.06698449800023809, 'timing/model': 2.043600749981124, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 1.41}) +[2026-06-23 12:19:23,788][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.08989474922418594, 'train/grad_norm_pre_clip': 20.30829620361328, 'timing/data': 0.001412632002029568, 'timing/model': 2.0530476319836453, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 1.41}) +[2026-06-23 12:19:27,817][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.05753330886363983, 'train/grad_norm_pre_clip': 28.99271583557129, 'timing/data': 0.0011741910129785538, 'timing/model': 2.03002856700914, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 1.41}) +[2026-06-23 12:19:31,838][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.02365671843290329, 'train/grad_norm_pre_clip': 7.873926639556885, 'timing/data': 0.0011832579912152141, 'timing/model': 2.0303399289841764, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 1.41}) +[2026-06-23 12:19:35,924][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.06809104979038239, 'train/grad_norm_pre_clip': 31.653154373168945, 'timing/data': 0.058865044004051015, 'timing/model': 2.0379733179870527, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 1.41}) +[2026-06-23 12:19:40,519][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.16079288721084595, 'train/grad_norm_pre_clip': 43.91007995605469, 'timing/data': 0.0013772959937341511, 'timing/model': 2.048730179987615, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 1.41}) +[2026-06-23 12:19:44,580][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.10115721821784973, 'train/grad_norm_pre_clip': 30.324737548828125, 'timing/data': 0.0016870390099938959, 'timing/model': 2.056206568988273, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 1.41}) +[2026-06-23 12:19:48,639][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.18952208757400513, 'train/grad_norm_pre_clip': 42.78618621826172, 'timing/data': 0.0012430179922375828, 'timing/model': 2.0605942469846923, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 1.41}) +[2026-06-23 12:19:52,733][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.04289913550019264, 'train/grad_norm_pre_clip': 18.437435150146484, 'timing/data': 0.06146803600131534, 'timing/model': 2.0373206939839292, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 1.41}) +[2026-06-23 12:19:56,848][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.0984334796667099, 'train/grad_norm_pre_clip': 28.527780532836914, 'timing/data': 0.0012383489811327308, 'timing/model': 2.0372846940008458, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 1.42}) +[2026-06-23 12:20:00,880][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.03051869198679924, 'train/grad_norm_pre_clip': 11.705278396606445, 'timing/data': 0.0012534970010165125, 'timing/model': 2.0366371239942964, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 1.42}) +[2026-06-23 12:20:04,916][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.049529779702425, 'train/grad_norm_pre_clip': 14.893186569213867, 'timing/data': 0.0014198650023899972, 'timing/model': 2.0377849109936506, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 1.42}) +[2026-06-23 12:20:09,016][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.021197311580181122, 'train/grad_norm_pre_clip': 5.933566570281982, 'timing/data': 0.06285077999928035, 'timing/model': 2.0398551560065243, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 1.42}) +[2026-06-23 12:20:13,379][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.15015961229801178, 'train/grad_norm_pre_clip': 27.31256103515625, 'timing/data': 0.0011974740191362798, 'timing/model': 2.2795056920149364, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 1.42}) +[2026-06-23 12:20:17,416][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.017717162147164345, 'train/grad_norm_pre_clip': 8.887556076049805, 'timing/data': 0.001232949987752363, 'timing/model': 2.0407061520090792, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 1.42}) +[2026-06-23 12:20:21,469][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.08198653906583786, 'train/grad_norm_pre_clip': 23.473697662353516, 'timing/data': 0.0015874549862928689, 'timing/model': 2.052212833019439, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 1.42}) +[2026-06-23 12:20:25,587][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.02226804941892624, 'train/grad_norm_pre_clip': 12.316426277160645, 'timing/data': 0.06742588800261728, 'timing/model': 2.0531245160091203, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 1.42}) +[2026-06-23 12:20:29,707][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.025360411033034325, 'train/grad_norm_pre_clip': 9.578729629516602, 'timing/data': 0.001240673998836428, 'timing/model': 2.0382805789995473, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 1.42}) +[2026-06-23 12:20:33,752][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.09721845388412476, 'train/grad_norm_pre_clip': 24.59182357788086, 'timing/data': 0.0013044520164839923, 'timing/model': 2.0524552020069677, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 1.42}) +[2026-06-23 12:20:37,798][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.09895619750022888, 'train/grad_norm_pre_clip': 19.680301666259766, 'timing/data': 0.0013744609896093607, 'timing/model': 2.0511756040214095, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 1.42}) +[2026-06-23 12:20:41,899][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.06320986896753311, 'train/grad_norm_pre_clip': 16.448732376098633, 'timing/data': 0.05865064199315384, 'timing/model': 2.0435433209931944, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 1.42}) +[2026-06-23 12:20:46,020][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.018952270969748497, 'train/grad_norm_pre_clip': 5.614985466003418, 'timing/data': 0.0012365959992166609, 'timing/model': 2.0364577180007473, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 1.42}) +[2026-06-23 12:20:50,180][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.04187986999750137, 'train/grad_norm_pre_clip': 14.328540802001953, 'timing/data': 0.0012693669996224344, 'timing/model': 2.0406497069925535, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 1.42}) +[2026-06-23 12:20:54,230][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.1805616319179535, 'train/grad_norm_pre_clip': 40.93266296386719, 'timing/data': 0.001206931978231296, 'timing/model': 2.0538577629777137, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 1.42}) +[2026-06-23 12:20:58,332][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.03584488481283188, 'train/grad_norm_pre_clip': 15.355900764465332, 'timing/data': 0.06030874600401148, 'timing/model': 2.039588760002516, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 1.42}) +[2026-06-23 12:21:02,461][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.02750198170542717, 'train/grad_norm_pre_clip': 13.015201568603516, 'timing/data': 0.0013957199989818037, 'timing/model': 2.041862020996632, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 1.42}) +[2026-06-23 12:21:06,526][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.0162656307220459, 'train/grad_norm_pre_clip': 10.465167045593262, 'timing/data': 0.0013000229955650866, 'timing/model': 2.05858435400296, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 1.42}) +[2026-06-23 12:21:10,574][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.06399046629667282, 'train/grad_norm_pre_clip': 19.904909133911133, 'timing/data': 0.0012322379916440696, 'timing/model': 2.048918121989118, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 1.42}) +[2026-06-23 12:21:14,674][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.15188850462436676, 'train/grad_norm_pre_clip': 39.9784049987793, 'timing/data': 0.059046400972874835, 'timing/model': 2.04313303498202, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 1.42}) +[2026-06-23 12:21:18,808][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.09285058826208115, 'train/grad_norm_pre_clip': 28.25868797302246, 'timing/data': 0.0014719619939569384, 'timing/model': 2.0530800029810052, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.42}) +[2026-06-23 12:21:22,899][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.026375798508524895, 'train/grad_norm_pre_clip': 12.605467796325684, 'timing/data': 0.0012746969878207892, 'timing/model': 2.0911813999991864, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 1.42}) +[2026-06-23 12:21:26,941][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.054836858063936234, 'train/grad_norm_pre_clip': 17.049251556396484, 'timing/data': 0.0014180710131768137, 'timing/model': 2.041797812009463, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 1.42}) +[2026-06-23 12:21:31,039][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.10387729853391647, 'train/grad_norm_pre_clip': 37.16713333129883, 'timing/data': 0.058051529980730265, 'timing/model': 2.0414124990056735, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 1.42}) +[2026-06-23 12:21:35,163][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.026443762704730034, 'train/grad_norm_pre_clip': 10.689814567565918, 'timing/data': 0.0013665059814229608, 'timing/model': 2.0424836270103697, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 1.42}) +[2026-06-23 12:21:39,238][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.06307721883058548, 'train/grad_norm_pre_clip': 25.171077728271484, 'timing/data': 0.001263436017325148, 'timing/model': 2.0596460940141696, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 1.43}) +[2026-06-23 12:21:43,278][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.035604096949100494, 'train/grad_norm_pre_clip': 10.682262420654297, 'timing/data': 0.0013584319967776537, 'timing/model': 2.041816592012765, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 1.43}) +[2026-06-23 12:21:47,381][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.1024148166179657, 'train/grad_norm_pre_clip': 24.849353790283203, 'timing/data': 0.0582063960027881, 'timing/model': 2.041576226009056, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 1.43}) +[2026-06-23 12:21:51,510][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.05008678510785103, 'train/grad_norm_pre_clip': 15.012567520141602, 'timing/data': 0.0014415949990507215, 'timing/model': 2.0443231430253945, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 1.43}) +[2026-06-23 12:21:55,549][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.062108591198921204, 'train/grad_norm_pre_clip': 18.095775604248047, 'timing/data': 0.0013000530016142875, 'timing/model': 2.04180720099248, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 1.43}) +[2026-06-23 12:21:59,829][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.10720404237508774, 'train/grad_norm_pre_clip': 56.40925216674805, 'timing/data': 0.00124369899276644, 'timing/model': 2.0395763360138517, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 1.43}) +[2026-06-23 12:22:03,924][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.04241282865405083, 'train/grad_norm_pre_clip': 17.09819221496582, 'timing/data': 0.05846983098308556, 'timing/model': 2.0401534929987974, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 1.43}) +[2026-06-23 12:22:08,054][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.0736176148056984, 'train/grad_norm_pre_clip': 22.425649642944336, 'timing/data': 0.0012470450019463897, 'timing/model': 2.0507936110079754, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 1.43}) +[2026-06-23 12:22:12,087][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.013178542256355286, 'train/grad_norm_pre_clip': 4.971562385559082, 'timing/data': 0.001506725006038323, 'timing/model': 2.037356850982178, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 1.43}) +[2026-06-23 12:22:16,138][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.04961499571800232, 'train/grad_norm_pre_clip': 22.224802017211914, 'timing/data': 0.0012542680196929723, 'timing/model': 2.0551814020145684, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 1.43}) +[2026-06-23 12:22:20,243][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.09336251765489578, 'train/grad_norm_pre_clip': 26.14208984375, 'timing/data': 0.06739254400599748, 'timing/model': 2.0392103280173615, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 1.43}) +[2026-06-23 12:22:24,379][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.06757259368896484, 'train/grad_norm_pre_clip': 25.620140075683594, 'timing/data': 0.0014463640109170228, 'timing/model': 2.0458965270081535, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 1.43}) +[2026-06-23 12:22:28,426][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.05297013744711876, 'train/grad_norm_pre_clip': 28.57558250427246, 'timing/data': 0.0012920579756610096, 'timing/model': 2.045749484997941, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 1.43}) +[2026-06-23 12:22:32,530][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.02563636004924774, 'train/grad_norm_pre_clip': 8.721527099609375, 'timing/data': 0.0013441750197671354, 'timing/model': 2.0419123210012913, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 1.43}) +[2026-06-23 12:22:36,624][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.10479245334863663, 'train/grad_norm_pre_clip': 18.47165298461914, 'timing/data': 0.05869882201659493, 'timing/model': 2.0388993619999383, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 1.43}) +[2026-06-23 12:22:40,744][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.10369057208299637, 'train/grad_norm_pre_clip': 23.816102981567383, 'timing/data': 0.0012243629898875952, 'timing/model': 2.037900265015196, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 1.43}) +[2026-06-23 12:22:44,774][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.03527744486927986, 'train/grad_norm_pre_clip': 14.136336326599121, 'timing/data': 0.0012156670272815973, 'timing/model': 2.036793599021621, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 1.43}) +[2026-06-23 12:22:48,806][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.10927974432706833, 'train/grad_norm_pre_clip': 36.2120475769043, 'timing/data': 0.0011792599980253726, 'timing/model': 2.038273093989119, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 1.43}) +[2026-06-23 12:22:52,907][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3700 +[2026-06-23 12:24:47,801][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3700 +[2026-06-23 12:24:47,802][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3700 +[2026-06-23 12:26:55,470][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3700 +[2026-06-23 12:26:55,471][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.04639527201652527, 'train/grad_norm_pre_clip': 16.369173049926758, 'eval/action_loss/seconds': 114.89104226199561, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04220354184508324, 'eval/latency_2/loss': 0.04220354184508324, 'eval/flappy/loss': 0.04220354184508324, 'eval/flappy/latency_2/loss': 0.04220354184508324, 'eval/flap_precision': 0.9463350785340314, 'eval/flap_recall': 0.9757085020242915, 'eval/flap_f1': 0.9607973421926911, 'eval/flap_rate_model': 0.21222222222222223, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9607973421926912, 'eval/latency_2/flap_precision': 0.9463350785340314, 'eval/latency_2/flap_recall': 0.9757085020242915, 'eval/latency_2/flap_f1': 0.9607973421926911, 'eval/latency_2/flap_rate_model': 0.21222222222222223, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9607973421926912, 'eval/action_classification/data_wait_seconds': 50.430259704589844, 'eval/action_classification/predict_seconds': 75.58504486083984, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.07138187700184062, 'timing/model': 2.036993109009927, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 1.43}) +[2026-06-23 12:26:59,513][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.03821685537695885, 'train/grad_norm_pre_clip': 15.98947811126709, 'timing/data': 0.00137389000155963, 'timing/model': 2.003219070000341, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 1.43}) +[2026-06-23 12:27:03,510][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.05883214250206947, 'train/grad_norm_pre_clip': 18.924528121948242, 'timing/data': 0.001469876995543018, 'timing/model': 2.018976509978529, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 1.43}) +[2026-06-23 12:27:07,477][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.06291097402572632, 'train/grad_norm_pre_clip': 18.03298568725586, 'timing/data': 0.001347390003502369, 'timing/model': 2.0015829149924684, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 1.43}) +[2026-06-23 12:27:11,516][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.0492975227534771, 'train/grad_norm_pre_clip': 15.921192169189453, 'timing/data': 0.0676318779878784, 'timing/model': 2.0140738670015708, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 1.43}) +[2026-06-23 12:27:15,571][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.016183940693736076, 'train/grad_norm_pre_clip': 14.712244987487793, 'timing/data': 0.0012846649915445596, 'timing/model': 2.0131065619934816, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 1.43}) +[2026-06-23 12:27:19,554][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.021618351340293884, 'train/grad_norm_pre_clip': 12.928959846496582, 'timing/data': 0.001291306019993499, 'timing/model': 2.014860288996715, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 1.43}) +[2026-06-23 12:27:23,544][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.13050369918346405, 'train/grad_norm_pre_clip': 34.58378982543945, 'timing/data': 0.0013746110198553652, 'timing/model': 2.0241178149881307, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 1.43}) +[2026-06-23 12:27:27,722][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.011807412840425968, 'train/grad_norm_pre_clip': 5.095512866973877, 'timing/data': 0.0686241420044098, 'timing/model': 2.1424713379819877, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 1.44}) +[2026-06-23 12:27:31,795][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.005550510250031948, 'train/grad_norm_pre_clip': 2.702838182449341, 'timing/data': 0.0014181009901221842, 'timing/model': 2.0141543880163226, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 1.44}) +[2026-06-23 12:27:35,778][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.03850197046995163, 'train/grad_norm_pre_clip': 14.6591796875, 'timing/data': 0.0013002739869989455, 'timing/model': 2.015639827994164, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 1.44}) +[2026-06-23 12:27:39,772][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.0865616574883461, 'train/grad_norm_pre_clip': 24.531326293945312, 'timing/data': 0.0012810479965992272, 'timing/model': 2.0211693020246457, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 1.44}) +[2026-06-23 12:27:43,822][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.010847273282706738, 'train/grad_norm_pre_clip': 3.188042163848877, 'timing/data': 0.05697759400936775, 'timing/model': 2.020182242995361, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 1.44}) +[2026-06-23 12:27:47,897][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.034714628010988235, 'train/grad_norm_pre_clip': 13.757169723510742, 'timing/data': 0.0012023329909425229, 'timing/model': 2.0153752229816746, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 1.44}) +[2026-06-23 12:27:51,899][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.008580641821026802, 'train/grad_norm_pre_clip': 5.419933319091797, 'timing/data': 0.0012749560119118541, 'timing/model': 2.0258078910119366, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 1.44}) +[2026-06-23 12:27:55,931][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.06806589663028717, 'train/grad_norm_pre_clip': 29.869110107421875, 'timing/data': 0.0012846140016335994, 'timing/model': 2.029729744012002, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 1.44}) +[2026-06-23 12:28:00,247][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.08221931755542755, 'train/grad_norm_pre_clip': 37.566810607910156, 'timing/data': 0.0669765860075131, 'timing/model': 2.276624552003341, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 1.44}) +[2026-06-23 12:28:04,366][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.02861200086772442, 'train/grad_norm_pre_clip': 13.292031288146973, 'timing/data': 0.0012192540161777288, 'timing/model': 2.046439838974038, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 1.44}) +[2026-06-23 12:28:08,373][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.06989827752113342, 'train/grad_norm_pre_clip': 15.080633163452148, 'timing/data': 0.0014266970101743937, 'timing/model': 2.023917594982777, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 1.44}) +[2026-06-23 12:28:12,379][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.05813327059149742, 'train/grad_norm_pre_clip': 17.79357147216797, 'timing/data': 0.0012045059993397444, 'timing/model': 2.0277482210076414, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 1.44}) +[2026-06-23 12:28:16,453][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.12443270534276962, 'train/grad_norm_pre_clip': 24.893524169921875, 'timing/data': 0.055897537007695064, 'timing/model': 2.0263034149829764, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 1.44}) +[2026-06-23 12:28:20,558][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.010710545815527439, 'train/grad_norm_pre_clip': 5.406097412109375, 'timing/data': 0.0013051829882897437, 'timing/model': 2.0247220769815613, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 1.44}) +[2026-06-23 12:28:24,575][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.14146840572357178, 'train/grad_norm_pre_clip': 17.029577255249023, 'timing/data': 0.0012433579831849784, 'timing/model': 2.0305462279939093, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 1.44}) +[2026-06-23 12:28:28,593][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.06338801980018616, 'train/grad_norm_pre_clip': 24.923112869262695, 'timing/data': 0.001309429993852973, 'timing/model': 2.031596624001395, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 1.44}) +[2026-06-23 12:28:32,661][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.031412847340106964, 'train/grad_norm_pre_clip': 14.290264129638672, 'timing/data': 0.0652322189998813, 'timing/model': 2.023099303012714, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 1.44}) +[2026-06-23 12:28:36,843][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.01339974906295538, 'train/grad_norm_pre_clip': 5.176314830780029, 'timing/data': 0.0012400010018609464, 'timing/model': 2.0279811410000548, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 1.44}) +[2026-06-23 12:28:40,866][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.011120425537228584, 'train/grad_norm_pre_clip': 11.135581016540527, 'timing/data': 0.0013101219956297427, 'timing/model': 2.03844239198952, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 1.44}) +[2026-06-23 12:28:44,879][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.1139543354511261, 'train/grad_norm_pre_clip': 15.739484786987305, 'timing/data': 0.001334396016318351, 'timing/model': 2.0305121700221207, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 1.44}) +[2026-06-23 12:28:48,943][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.05439228191971779, 'train/grad_norm_pre_clip': 19.863727569580078, 'timing/data': 0.05657961600809358, 'timing/model': 2.023867623996921, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 1.44}) +[2026-06-23 12:28:53,040][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.015096347779035568, 'train/grad_norm_pre_clip': 5.69891357421875, 'timing/data': 0.0013750609941780567, 'timing/model': 2.0302118250110652, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 1.44}) +[2026-06-23 12:28:57,049][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.1085934191942215, 'train/grad_norm_pre_clip': 23.606962203979492, 'timing/data': 0.0012732140021398664, 'timing/model': 2.028956049005501, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 1.44}) +[2026-06-23 12:29:01,060][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.21783438324928284, 'train/grad_norm_pre_clip': 53.53734588623047, 'timing/data': 0.0012123009946662933, 'timing/model': 2.0279240969975945, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 1.44}) +[2026-06-23 12:29:05,139][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.05257779732346535, 'train/grad_norm_pre_clip': 16.471067428588867, 'timing/data': 0.06620801199460402, 'timing/model': 2.0267187349963933, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 1.44}) +[2026-06-23 12:29:09,506][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.003806755179539323, 'train/grad_norm_pre_clip': 1.3645124435424805, 'timing/data': 0.0011830260045826435, 'timing/model': 2.2910732140007894, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 1.44}) +[2026-06-23 12:29:13,531][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.0774211436510086, 'train/grad_norm_pre_clip': 17.928123474121094, 'timing/data': 0.001642765972064808, 'timing/model': 2.0375399260083213, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 1.45}) +[2026-06-23 12:29:17,553][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.0363057479262352, 'train/grad_norm_pre_clip': 18.66551971435547, 'timing/data': 0.00121645798208192, 'timing/model': 2.0311121819831897, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 1.45}) +[2026-06-23 12:29:21,645][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.09867367148399353, 'train/grad_norm_pre_clip': 28.821910858154297, 'timing/data': 0.060471528995549306, 'timing/model': 2.041587839019485, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 1.45}) +[2026-06-23 12:29:25,730][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.026252802461385727, 'train/grad_norm_pre_clip': 19.97893714904785, 'timing/data': 0.0010977190104313195, 'timing/model': 2.02732547398773, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 1.45}) +[2026-06-23 12:29:29,744][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.09555382281541824, 'train/grad_norm_pre_clip': 33.5059700012207, 'timing/data': 0.001284964004298672, 'timing/model': 2.03296997200232, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 1.45}) +[2026-06-23 12:29:33,771][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.1506711095571518, 'train/grad_norm_pre_clip': 37.70306396484375, 'timing/data': 0.001444659981643781, 'timing/model': 2.0320397579926066, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 1.45}) +[2026-06-23 12:29:37,856][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.09715287387371063, 'train/grad_norm_pre_clip': 26.5163631439209, 'timing/data': 0.0664251419948414, 'timing/model': 2.0286026510002557, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 1.45}) +[2026-06-23 12:29:41,953][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.05099223554134369, 'train/grad_norm_pre_clip': 24.9820556640625, 'timing/data': 0.001333374995738268, 'timing/model': 2.028709554026136, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 1.45}) +[2026-06-23 12:29:46,214][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.0750708058476448, 'train/grad_norm_pre_clip': 19.03250503540039, 'timing/data': 0.0012521040043793619, 'timing/model': 2.2732745629909914, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 1.45}) +[2026-06-23 12:29:50,240][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.11411824822425842, 'train/grad_norm_pre_clip': 23.419422149658203, 'timing/data': 0.0012109979870729148, 'timing/model': 2.034076774987625, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 1.45}) +[2026-06-23 12:29:54,330][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.029085922986268997, 'train/grad_norm_pre_clip': 24.398725509643555, 'timing/data': 0.06448508601170033, 'timing/model': 2.0375703680037986, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 1.45}) +[2026-06-23 12:29:58,445][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.007040947675704956, 'train/grad_norm_pre_clip': 3.729393243789673, 'timing/data': 0.0011514379875734448, 'timing/model': 2.0415586449962575, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 1.45}) +[2026-06-23 12:30:02,458][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.12726862728595734, 'train/grad_norm_pre_clip': 37.02131652832031, 'timing/data': 0.0011712450068444014, 'timing/model': 2.029770824010484, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 1.45}) +[2026-06-23 12:30:06,468][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.015820657834410667, 'train/grad_norm_pre_clip': 5.8349761962890625, 'timing/data': 0.0011215429985895753, 'timing/model': 2.027395240002079, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 1.45}) +[2026-06-23 12:30:10,552][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.10654804110527039, 'train/grad_norm_pre_clip': 23.462749481201172, 'timing/data': 0.06256235198816285, 'timing/model': 2.036780192021979, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 1.45}) +[2026-06-23 12:30:14,668][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.019268475472927094, 'train/grad_norm_pre_clip': 9.439592361450195, 'timing/data': 0.0012443690211512148, 'timing/model': 2.0497378960135393, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 1.45}) +[2026-06-23 12:47:36,869][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3750 +[2026-06-23 12:47:36,870][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.04767737165093422, 'train/grad_norm_pre_clip': 33.98775863647461, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 268.90000000001265, 'rl_games_eval/mid_train/mean_length': 2186.2, 'rl_games_eval/mid_train/std_reward': 184.73615780351173, 'rl_games_eval/mid_train/std_length': 1488.7764640804878, 'rl_games_eval/mid_train/macro_mean_reward': 268.90000000001265, 'rl_games_eval/mid_train/macro_mean_length': 2186.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 268.90000000001265, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2186.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 184.73615780351173, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1488.7764640804878, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 268.90000000001265, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0013002330088056624, 'timing/model': 2.135962478001602, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 1.45}) +[2026-06-23 12:47:40,744][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.018603088334202766, 'train/grad_norm_pre_clip': 15.262723922729492, 'timing/data': 0.0014045070274733007, 'timing/model': 1.9612933799799066, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 1.45}) +[2026-06-23 12:47:44,686][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.006236301269382238, 'train/grad_norm_pre_clip': 4.1517720222473145, 'timing/data': 0.06512156198732555, 'timing/model': 1.9597363339853473, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 1.45}) +[2026-06-23 12:47:48,670][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.2937748432159424, 'train/grad_norm_pre_clip': 47.47979736328125, 'timing/data': 0.0012881410075351596, 'timing/model': 1.9774303759913892, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 1.45}) +[2026-06-23 12:47:52,583][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.0040162717923521996, 'train/grad_norm_pre_clip': 3.491814613342285, 'timing/data': 0.0012024920142721385, 'timing/model': 1.9690143200277817, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 1.45}) +[2026-06-23 12:47:56,508][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.09793051332235336, 'train/grad_norm_pre_clip': 26.421653747558594, 'timing/data': 0.001220586011186242, 'timing/model': 1.982637626992073, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 1.45}) +[2026-06-23 12:48:00,574][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.1098446249961853, 'train/grad_norm_pre_clip': 16.83962059020996, 'timing/data': 0.059669812995707616, 'timing/model': 1.9806470090115909, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 1.45}) +[2026-06-23 12:48:04,577][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.0858108177781105, 'train/grad_norm_pre_clip': 33.35451126098633, 'timing/data': 0.0013678590185008943, 'timing/model': 1.989016403007554, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 1.45}) +[2026-06-23 12:48:08,493][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.16500450670719147, 'train/grad_norm_pre_clip': 43.7918586730957, 'timing/data': 0.001433340017683804, 'timing/model': 1.979776516003767, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 1.45}) +[2026-06-23 12:48:12,426][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.056519366800785065, 'train/grad_norm_pre_clip': 25.35845947265625, 'timing/data': 0.0013469700061250478, 'timing/model': 1.9947713259898592, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 1.45}) +[2026-06-23 12:48:16,398][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.021155577152967453, 'train/grad_norm_pre_clip': 16.877004623413086, 'timing/data': 0.05227001299499534, 'timing/model': 1.9807844550232403, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 1.46}) +[2026-06-23 12:48:20,409][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.05607933923602104, 'train/grad_norm_pre_clip': 22.272117614746094, 'timing/data': 0.0012863180018030107, 'timing/model': 1.9885552519990597, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 1.46}) +[2026-06-23 12:48:24,343][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.030316976830363274, 'train/grad_norm_pre_clip': 11.855169296264648, 'timing/data': 0.0011550449999049306, 'timing/model': 1.9907102079887409, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 1.46}) +[2026-06-23 12:48:28,295][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.0517251119017601, 'train/grad_norm_pre_clip': 34.77930450439453, 'timing/data': 0.001288421975914389, 'timing/model': 1.9990456500090659, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 1.46}) +[2026-06-23 12:48:32,504][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.01617279089987278, 'train/grad_norm_pre_clip': 8.808544158935547, 'timing/data': 0.06105868800659664, 'timing/model': 2.186364274995867, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 1.46}) +[2026-06-23 12:48:36,543][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.19083477556705475, 'train/grad_norm_pre_clip': 21.795129776000977, 'timing/data': 0.0011012160102836788, 'timing/model': 2.0073059950082097, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 1.46}) +[2026-06-23 12:48:40,491][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.10702618956565857, 'train/grad_norm_pre_clip': 33.953697204589844, 'timing/data': 0.001317875983659178, 'timing/model': 1.9933231199975125, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 1.46}) +[2026-06-23 12:48:44,440][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.021161455661058426, 'train/grad_norm_pre_clip': 14.975879669189453, 'timing/data': 0.0011488739983178675, 'timing/model': 1.9946211530186702, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 1.46}) +[2026-06-23 12:48:48,467][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.08616732805967331, 'train/grad_norm_pre_clip': 31.055076599121094, 'timing/data': 0.0683048719947692, 'timing/model': 2.0008724209910724, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 1.46}) +[2026-06-23 12:48:52,506][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.03621496260166168, 'train/grad_norm_pre_clip': 20.002634048461914, 'timing/data': 0.0011902400001417845, 'timing/model': 2.000078437005868, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 1.46}) +[2026-06-23 12:48:56,470][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.05630774423480034, 'train/grad_norm_pre_clip': 28.48187255859375, 'timing/data': 0.0012080729939043522, 'timing/model': 2.0025630829914007, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 1.46}) +[2026-06-23 12:49:00,436][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.044516514986753464, 'train/grad_norm_pre_clip': 15.002185821533203, 'timing/data': 0.0011207909847144037, 'timing/model': 2.0036095620016567, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 1.46}) +[2026-06-23 12:49:04,499][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.024246608838438988, 'train/grad_norm_pre_clip': 12.581720352172852, 'timing/data': 0.05794960999628529, 'timing/model': 2.040043332002824, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 1.46}) +[2026-06-23 12:49:08,551][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.010930913500487804, 'train/grad_norm_pre_clip': 6.723919868469238, 'timing/data': 0.0011076279915869236, 'timing/model': 2.0059216719819233, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 1.46}) +[2026-06-23 12:49:12,528][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.051804810762405396, 'train/grad_norm_pre_clip': 33.50710678100586, 'timing/data': 0.0012364959984552115, 'timing/model': 2.012357161002001, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 1.46}) +[2026-06-23 12:49:16,514][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.021466298028826714, 'train/grad_norm_pre_clip': 12.668173789978027, 'timing/data': 0.001516552991233766, 'timing/model': 2.016218912001932, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 1.46}) +[2026-06-23 12:49:20,567][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.008961557410657406, 'train/grad_norm_pre_clip': 6.138373374938965, 'timing/data': 0.05697426601545885, 'timing/model': 2.022682364011416, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 1.46}) +[2026-06-23 12:49:24,643][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.20512615144252777, 'train/grad_norm_pre_clip': 25.690567016601562, 'timing/data': 0.0012282300158403814, 'timing/model': 2.0194025140081067, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 1.46}) +[2026-06-23 12:49:28,656][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.0717572569847107, 'train/grad_norm_pre_clip': 16.78725814819336, 'timing/data': 0.0012633750156965107, 'timing/model': 2.037211546994513, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 1.46}) +[2026-06-23 12:49:32,661][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.04387471452355385, 'train/grad_norm_pre_clip': 23.789318084716797, 'timing/data': 0.0012315060012042522, 'timing/model': 2.024656223016791, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 1.46}) +[2026-06-23 12:49:36,708][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.06218566745519638, 'train/grad_norm_pre_clip': 38.94482421875, 'timing/data': 0.0558312330103945, 'timing/model': 2.0184365400054958, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 1.46}) +[2026-06-23 12:49:40,960][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.04678855836391449, 'train/grad_norm_pre_clip': 25.102203369140625, 'timing/data': 0.001118677988415584, 'timing/model': 2.0184845220064744, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 1.46}) +[2026-06-23 12:49:44,955][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.06904176622629166, 'train/grad_norm_pre_clip': 20.074501037597656, 'timing/data': 0.0012608109973371029, 'timing/model': 2.016751001996454, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 1.46}) +[2026-06-23 12:49:48,951][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.07104670256376266, 'train/grad_norm_pre_clip': 33.6682243347168, 'timing/data': 0.0011249489907640964, 'timing/model': 2.0213356300082523, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 1.46}) +[2026-06-23 12:49:53,008][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.07087168097496033, 'train/grad_norm_pre_clip': 25.230823516845703, 'timing/data': 0.05964545300230384, 'timing/model': 2.0206633110065013, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 1.46}) +[2026-06-23 12:49:57,089][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.04243295639753342, 'train/grad_norm_pre_clip': 20.390928268432617, 'timing/data': 0.0012438090052455664, 'timing/model': 2.0202060799929313, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 1.46}) +[2026-06-23 12:50:01,090][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.10011053830385208, 'train/grad_norm_pre_clip': 41.02408981323242, 'timing/data': 0.0011600949801504612, 'timing/model': 2.019858932006173, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 1.47}) +[2026-06-23 12:50:05,090][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.1076449304819107, 'train/grad_norm_pre_clip': 19.674543380737305, 'timing/data': 0.001111574994865805, 'timing/model': 2.0207189150096383, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 1.47}) +[2026-06-23 12:50:09,143][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.0024865285959094763, 'train/grad_norm_pre_clip': 1.3504674434661865, 'timing/data': 0.05232839600648731, 'timing/model': 2.024265955988085, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 1.47}) +[2026-06-23 12:50:13,483][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.07053421437740326, 'train/grad_norm_pre_clip': 29.76923179626465, 'timing/data': 0.0012290520244278014, 'timing/model': 2.0239215930050705, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 1.47}) +[2026-06-23 12:50:17,492][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.0830167755484581, 'train/grad_norm_pre_clip': 38.007442474365234, 'timing/data': 0.0014016510103829205, 'timing/model': 2.023592229001224, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 1.47}) +[2026-06-23 12:50:21,507][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.09146784991025925, 'train/grad_norm_pre_clip': 25.131799697875977, 'timing/data': 0.0013962610100861639, 'timing/model': 2.0290830729936715, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 1.47}) +[2026-06-23 12:50:25,574][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.023860525339841843, 'train/grad_norm_pre_clip': 12.844470024108887, 'timing/data': 0.05677244899561629, 'timing/model': 2.0277936029888224, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 1.47}) +[2026-06-23 12:50:29,667][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.08868396282196045, 'train/grad_norm_pre_clip': 26.229764938354492, 'timing/data': 0.0011822849919553846, 'timing/model': 2.034582936001243, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 1.47}) +[2026-06-23 12:50:33,684][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.07251583784818649, 'train/grad_norm_pre_clip': 23.097978591918945, 'timing/data': 0.0012143139902036637, 'timing/model': 2.025663643988082, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 1.47}) +[2026-06-23 12:50:37,716][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.10918626934289932, 'train/grad_norm_pre_clip': 25.885591506958008, 'timing/data': 0.0012698869977612048, 'timing/model': 2.030773925012909, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 1.47}) +[2026-06-23 12:50:41,793][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.10475000739097595, 'train/grad_norm_pre_clip': 27.72422981262207, 'timing/data': 0.059101995022501796, 'timing/model': 2.0317191570065916, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 1.47}) +[2026-06-23 12:50:46,136][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.038014672696590424, 'train/grad_norm_pre_clip': 22.311992645263672, 'timing/data': 0.0011539029947016388, 'timing/model': 2.2755377090070397, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 1.47}) +[2026-06-23 12:50:50,163][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.10581950843334198, 'train/grad_norm_pre_clip': 25.752593994140625, 'timing/data': 0.0012438090052455664, 'timing/model': 2.0414726739982143, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 1.47}) +[2026-06-23 12:50:54,180][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.019325707107782364, 'train/grad_norm_pre_clip': 6.199037075042725, 'timing/data': 0.0012226800026837736, 'timing/model': 2.0312540299782995, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 1.47}) +[2026-06-23 12:50:58,256][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3800 +[2026-06-23 12:52:52,206][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3800 +[2026-06-23 12:52:52,206][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3800 +[2026-06-23 12:54:59,785][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3800 +[2026-06-23 12:54:59,786][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.03204978257417679, 'train/grad_norm_pre_clip': 13.730850219726562, 'eval/action_loss/seconds': 113.94663719201344, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04336734861135483, 'eval/latency_2/loss': 0.04336734861135483, 'eval/flappy/loss': 0.04336734861135483, 'eval/flappy/latency_2/loss': 0.04336734861135483, 'eval/flap_precision': 0.9481286933683519, 'eval/flap_recall': 0.9743589743589743, 'eval/flap_f1': 0.9610648918469218, 'eval/flap_rate_model': 0.2115277777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9610648918469218, 'eval/latency_2/flap_precision': 0.9481286933683519, 'eval/latency_2/flap_recall': 0.9743589743589743, 'eval/latency_2/flap_f1': 0.9610648918469218, 'eval/latency_2/flap_rate_model': 0.2115277777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9610648918469218, 'eval/action_classification/data_wait_seconds': 50.598365783691406, 'eval/action_classification/predict_seconds': 75.14167785644531, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0624455549987033, 'timing/model': 2.0298185409919824, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 1.47}) +[2026-06-23 12:55:03,825][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.09557567536830902, 'train/grad_norm_pre_clip': 32.73818588256836, 'timing/data': 0.0012032240047119558, 'timing/model': 2.0004331370000727, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 1.47}) +[2026-06-23 12:55:07,787][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.03412141278386116, 'train/grad_norm_pre_clip': 11.008905410766602, 'timing/data': 0.0012359840038698167, 'timing/model': 2.0011908239976037, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 1.47}) +[2026-06-23 12:55:11,928][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.05058029294013977, 'train/grad_norm_pre_clip': 34.191463470458984, 'timing/data': 0.0014939209795556962, 'timing/model': 2.178968383988831, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 1.47}) +[2026-06-23 12:55:15,953][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.019774774089455605, 'train/grad_norm_pre_clip': 8.906010627746582, 'timing/data': 0.05629294499522075, 'timing/model': 2.012060939014191, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 1.47}) +[2026-06-23 12:55:19,996][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.07418468594551086, 'train/grad_norm_pre_clip': 73.3848876953125, 'timing/data': 0.0011009849840775132, 'timing/model': 2.001583481003763, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 1.47}) +[2026-06-23 12:55:23,970][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.10532140731811523, 'train/grad_norm_pre_clip': 39.972618103027344, 'timing/data': 0.0012201750068925321, 'timing/model': 2.0084953420155216, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 1.47}) +[2026-06-23 12:55:27,953][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.0024134735576808453, 'train/grad_norm_pre_clip': 1.32694673538208, 'timing/data': 0.0012135430006310344, 'timing/model': 2.015034715994261, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 1.47}) +[2026-06-23 12:55:31,985][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.013541318476200104, 'train/grad_norm_pre_clip': 7.520655155181885, 'timing/data': 0.05527505301870406, 'timing/model': 2.010585097974399, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 1.47}) +[2026-06-23 12:55:36,066][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.2418375313282013, 'train/grad_norm_pre_clip': 30.432472229003906, 'timing/data': 0.00121859202045016, 'timing/model': 2.027242612995906, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 1.47}) +[2026-06-23 12:55:40,057][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.0015972182154655457, 'train/grad_norm_pre_clip': 0.6237295269966125, 'timing/data': 0.0014260259922593832, 'timing/model': 2.015315214986913, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 1.47}) +[2026-06-23 12:55:44,047][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.090927854180336, 'train/grad_norm_pre_clip': 18.229503631591797, 'timing/data': 0.0012360440159682184, 'timing/model': 2.0177145059860777, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 1.47}) +[2026-06-23 12:55:48,138][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.04257356375455856, 'train/grad_norm_pre_clip': 19.344274520874023, 'timing/data': 0.06582045397954062, 'timing/model': 2.0204435070045292, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 1.48}) +[2026-06-23 12:55:52,222][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.008869890123605728, 'train/grad_norm_pre_clip': 11.398537635803223, 'timing/data': 0.001457855018088594, 'timing/model': 2.0239413890230935, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 1.48}) +[2026-06-23 12:55:56,221][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.0641714408993721, 'train/grad_norm_pre_clip': 31.89558219909668, 'timing/data': 0.001293812005314976, 'timing/model': 2.020896759000607, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 1.48}) +[2026-06-23 12:56:00,222][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.05039197951555252, 'train/grad_norm_pre_clip': 17.06195831298828, 'timing/data': 0.0012094659905415028, 'timing/model': 2.0231140239920933, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 1.48}) +[2026-06-23 12:56:04,357][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.06956371665000916, 'train/grad_norm_pre_clip': 25.402511596679688, 'timing/data': 0.0667221600015182, 'timing/model': 2.021488195983693, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 1.48}) +[2026-06-23 12:56:08,544][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.0912395492196083, 'train/grad_norm_pre_clip': 16.579219818115234, 'timing/data': 0.0013796600105706602, 'timing/model': 2.031107121001696, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 1.48}) +[2026-06-23 12:56:12,563][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.08427871763706207, 'train/grad_norm_pre_clip': 35.334716796875, 'timing/data': 0.0013376619899645448, 'timing/model': 2.0384365649952088, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 1.48}) +[2026-06-23 12:56:16,571][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.21159851551055908, 'train/grad_norm_pre_clip': 33.12818908691406, 'timing/data': 0.001192594994790852, 'timing/model': 2.0266680210188497, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 1.48}) +[2026-06-23 12:56:20,745][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.059045784175395966, 'train/grad_norm_pre_clip': 71.5992660522461, 'timing/data': 0.06427172099938616, 'timing/model': 2.127809953992255, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 1.48}) +[2026-06-23 12:56:24,844][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.047555383294820786, 'train/grad_norm_pre_clip': 11.785088539123535, 'timing/data': 0.0012304850097279996, 'timing/model': 2.0381288480130024, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 1.48}) +[2026-06-23 12:56:28,867][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.04363849386572838, 'train/grad_norm_pre_clip': 29.323579788208008, 'timing/data': 0.0012874089879915118, 'timing/model': 2.0412914689804893, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 1.48}) +[2026-06-23 12:56:32,880][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.03806453198194504, 'train/grad_norm_pre_clip': 11.476420402526855, 'timing/data': 0.0012354939826764166, 'timing/model': 2.02628230699338, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 1.48}) +[2026-06-23 12:56:36,950][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.021221229806542397, 'train/grad_norm_pre_clip': 9.406394958496094, 'timing/data': 0.059319827996660024, 'timing/model': 2.0279661459790077, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 1.48}) +[2026-06-23 12:56:41,053][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.05341630429029465, 'train/grad_norm_pre_clip': 24.513404846191406, 'timing/data': 0.0012013110099360347, 'timing/model': 2.0377913719858043, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 1.48}) +[2026-06-23 12:56:45,065][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.08731862902641296, 'train/grad_norm_pre_clip': 25.348966598510742, 'timing/data': 0.0013016960001550615, 'timing/model': 2.0263304469990544, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 1.48}) +[2026-06-23 12:56:49,080][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.005917115602642298, 'train/grad_norm_pre_clip': 3.885179042816162, 'timing/data': 0.001364322000881657, 'timing/model': 2.031123041990213, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 1.48}) +[2026-06-23 12:56:53,150][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.07372979074716568, 'train/grad_norm_pre_clip': 22.697357177734375, 'timing/data': 0.05923474000883289, 'timing/model': 2.0282487370132003, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 1.48}) +[2026-06-23 12:56:57,488][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.06217637285590172, 'train/grad_norm_pre_clip': 32.67264175415039, 'timing/data': 0.0011970620253123343, 'timing/model': 2.0249210730253253, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 1.48}) +[2026-06-23 12:57:01,518][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.045191291719675064, 'train/grad_norm_pre_clip': 27.65092658996582, 'timing/data': 0.001386371994158253, 'timing/model': 2.0443501410190947, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 1.48}) +[2026-06-23 12:57:05,533][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.0630464106798172, 'train/grad_norm_pre_clip': 17.181943893432617, 'timing/data': 0.001188666996313259, 'timing/model': 2.027972228010185, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 1.48}) +[2026-06-23 12:57:09,606][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.015376267954707146, 'train/grad_norm_pre_clip': 6.005741119384766, 'timing/data': 0.06443325901636854, 'timing/model': 2.02383358599036, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 1.48}) +[2026-06-23 12:57:13,712][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.011087700724601746, 'train/grad_norm_pre_clip': 3.9536030292510986, 'timing/data': 0.0013339160068426281, 'timing/model': 2.038798692985438, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 1.48}) +[2026-06-23 12:57:17,720][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.06363871693611145, 'train/grad_norm_pre_clip': 41.937129974365234, 'timing/data': 0.001446854992536828, 'timing/model': 2.026047502004076, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 1.48}) +[2026-06-23 12:57:21,731][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.05843796581029892, 'train/grad_norm_pre_clip': 19.573787689208984, 'timing/data': 0.0011937870003748685, 'timing/model': 2.0271326460060664, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 1.48}) +[2026-06-23 12:57:25,815][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.10961806774139404, 'train/grad_norm_pre_clip': 30.827957153320312, 'timing/data': 0.0553715490095783, 'timing/model': 2.025872823985992, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 1.48}) +[2026-06-23 12:57:29,964][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.04229436814785004, 'train/grad_norm_pre_clip': 26.82241439819336, 'timing/data': 0.0012544390046969056, 'timing/model': 2.0480012310144957, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 1.48}) +[2026-06-23 12:57:33,993][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.02024519443511963, 'train/grad_norm_pre_clip': 20.107425689697266, 'timing/data': 0.0013341459853108972, 'timing/model': 2.0347572000173386, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 1.49}) +[2026-06-23 12:57:38,013][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.09656170010566711, 'train/grad_norm_pre_clip': 51.43100357055664, 'timing/data': 0.0013749710051342845, 'timing/model': 2.031846227007918, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 1.49}) +[2026-06-23 12:57:42,086][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.12647095322608948, 'train/grad_norm_pre_clip': 18.672021865844727, 'timing/data': 0.05407635500887409, 'timing/model': 2.031758172000991, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 1.49}) +[2026-06-23 12:57:46,186][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.016960930079221725, 'train/grad_norm_pre_clip': 25.55548095703125, 'timing/data': 0.001224433013703674, 'timing/model': 2.030834582983516, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 1.49}) +[2026-06-23 12:57:50,213][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.03672760725021362, 'train/grad_norm_pre_clip': 16.632875442504883, 'timing/data': 0.001286156999412924, 'timing/model': 2.0359868930245284, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 1.49}) +[2026-06-23 12:57:54,237][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.004044901113957167, 'train/grad_norm_pre_clip': 2.2009570598602295, 'timing/data': 0.0013634499919135123, 'timing/model': 2.0343282749818172, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 1.49}) +[2026-06-23 12:57:58,321][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.06164173781871796, 'train/grad_norm_pre_clip': 46.4454231262207, 'timing/data': 0.06269424300990067, 'timing/model': 2.0316393230168615, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 1.49}) +[2026-06-23 12:58:02,427][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.031558021903038025, 'train/grad_norm_pre_clip': 16.742586135864258, 'timing/data': 0.0011930250038858503, 'timing/model': 2.033978013001615, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 1.49}) +[2026-06-23 12:58:06,681][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.08827229589223862, 'train/grad_norm_pre_clip': 39.47978210449219, 'timing/data': 0.0012548989907372743, 'timing/model': 2.038752383989049, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 1.49}) +[2026-06-23 12:58:10,712][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.14948835968971252, 'train/grad_norm_pre_clip': 19.042007446289062, 'timing/data': 0.0012093149998690933, 'timing/model': 2.0364185479993466, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 1.49}) +[2026-06-23 12:58:14,794][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.06032715365290642, 'train/grad_norm_pre_clip': 22.48642921447754, 'timing/data': 0.053804472991032526, 'timing/model': 2.0357783430081327, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 1.49}) +[2026-06-23 12:58:18,901][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.018419330939650536, 'train/grad_norm_pre_clip': 13.0582914352417, 'timing/data': 0.001195720979012549, 'timing/model': 2.0313755329989363, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 1.49}) +[2026-06-23 12:58:22,926][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.10001073032617569, 'train/grad_norm_pre_clip': 21.647661209106445, 'timing/data': 0.0012767899897880852, 'timing/model': 2.0348201979941223, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 1.49}) +[2026-06-23 12:58:26,949][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.02452686056494713, 'train/grad_norm_pre_clip': 14.800774574279785, 'timing/data': 0.0012209769920445979, 'timing/model': 2.0345873849873897, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 1.49}) +[2026-06-23 12:58:31,026][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.06475673615932465, 'train/grad_norm_pre_clip': 23.06987953186035, 'timing/data': 0.05678952898597345, 'timing/model': 2.032351960981032, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 1.49}) +[2026-06-23 12:58:35,137][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.04558548703789711, 'train/grad_norm_pre_clip': 18.928686141967773, 'timing/data': 0.001483140978962183, 'timing/model': 2.0346805619774386, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 1.49}) +[2026-06-23 12:58:39,185][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.1143469586968422, 'train/grad_norm_pre_clip': 34.25130081176758, 'timing/data': 0.0012261059891898185, 'timing/model': 2.0371541120111942, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 1.49}) +[2026-06-23 12:58:43,214][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.08107582479715347, 'train/grad_norm_pre_clip': 34.12152862548828, 'timing/data': 0.0012330379977356642, 'timing/model': 2.036705730977701, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 1.49}) +[2026-06-23 12:58:47,311][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.08439581096172333, 'train/grad_norm_pre_clip': 29.24399757385254, 'timing/data': 0.06497946599847637, 'timing/model': 2.040636783989612, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 1.49}) +[2026-06-23 12:58:51,423][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.018759096041321754, 'train/grad_norm_pre_clip': 13.687127113342285, 'timing/data': 0.0011727570090442896, 'timing/model': 2.035011534986552, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 1.49}) +[2026-06-23 12:58:55,453][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.028910178691148758, 'train/grad_norm_pre_clip': 11.798360824584961, 'timing/data': 0.0012365150032564998, 'timing/model': 2.037436675018398, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 1.49}) +[2026-06-23 12:58:59,479][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.17673465609550476, 'train/grad_norm_pre_clip': 31.84909439086914, 'timing/data': 0.0013829759845975786, 'timing/model': 2.0343914550030604, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 1.49}) +[2026-06-23 12:59:03,573][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.03656912222504616, 'train/grad_norm_pre_clip': 16.103363037109375, 'timing/data': 0.06421137601137161, 'timing/model': 2.03430514401407, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 1.49}) +[2026-06-23 12:59:07,688][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.028917605057358742, 'train/grad_norm_pre_clip': 17.973155975341797, 'timing/data': 0.0013639209792017937, 'timing/model': 2.034369561006315, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 1.49}) +[2026-06-23 12:59:11,774][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.04631870612502098, 'train/grad_norm_pre_clip': 25.572153091430664, 'timing/data': 0.0015038689889479429, 'timing/model': 2.09467604701058, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 1.49}) +[2026-06-23 12:59:15,822][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.0667959600687027, 'train/grad_norm_pre_clip': 24.300329208374023, 'timing/data': 0.0012209660199005157, 'timing/model': 2.0409306430083234, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 1.5}) +[2026-06-23 12:59:19,909][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.027413543313741684, 'train/grad_norm_pre_clip': 29.022369384765625, 'timing/data': 0.06558015500195324, 'timing/model': 2.0356122509983834, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 1.5}) +[2026-06-23 12:59:24,011][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.054022420197725296, 'train/grad_norm_pre_clip': 22.95395278930664, 'timing/data': 0.0013592919858638197, 'timing/model': 2.031015163025586, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 1.5}) +[2026-06-23 12:59:28,044][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.1656384915113449, 'train/grad_norm_pre_clip': 48.939674377441406, 'timing/data': 0.0016025519871618599, 'timing/model': 2.0419994019903243, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 1.5}) +[2026-06-23 12:59:32,106][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.13458916544914246, 'train/grad_norm_pre_clip': 18.72441864013672, 'timing/data': 0.0012429880152922124, 'timing/model': 2.0491251799976453, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 1.5}) +[2026-06-23 12:59:36,191][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.09975609183311462, 'train/grad_norm_pre_clip': 37.43613815307617, 'timing/data': 0.05413789799786173, 'timing/model': 2.035268587002065, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 1.5}) +[2026-06-23 12:59:40,299][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.024970771744847298, 'train/grad_norm_pre_clip': 10.26671314239502, 'timing/data': 0.0011829970171675086, 'timing/model': 2.0376845610153396, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 1.5}) +[2026-06-23 12:59:44,333][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.012119709514081478, 'train/grad_norm_pre_clip': 6.248103141784668, 'timing/data': 0.0014459820231422782, 'timing/model': 2.0384992939943913, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 1.5}) +[2026-06-23 12:59:48,627][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.11637195944786072, 'train/grad_norm_pre_clip': 25.1568660736084, 'timing/data': 0.0014301330083981156, 'timing/model': 2.0391700729960576, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 1.5}) +[2026-06-23 12:59:52,709][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.020149661228060722, 'train/grad_norm_pre_clip': 10.286863327026367, 'timing/data': 0.05359987300471403, 'timing/model': 2.03661032000673, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 1.5}) +[2026-06-23 12:59:56,821][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.12085108458995819, 'train/grad_norm_pre_clip': 42.189151763916016, 'timing/data': 0.0011867739958688617, 'timing/model': 2.037280827003997, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 1.5}) +[2026-06-23 13:00:00,865][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.02681810036301613, 'train/grad_norm_pre_clip': 23.374174118041992, 'timing/data': 0.0012982900079805404, 'timing/model': 2.0514292040024884, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 1.5}) +[2026-06-23 13:00:04,913][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.1066204085946083, 'train/grad_norm_pre_clip': 31.980209350585938, 'timing/data': 0.001370373007375747, 'timing/model': 2.0523724750091787, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 1.5}) +[2026-06-23 13:00:09,013][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.057896923273801804, 'train/grad_norm_pre_clip': 21.591337203979492, 'timing/data': 0.0713246609957423, 'timing/model': 2.0362922030035406, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 1.5}) +[2026-06-23 13:00:13,149][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.06584873795509338, 'train/grad_norm_pre_clip': 28.11650276184082, 'timing/data': 0.0012168480025138706, 'timing/model': 2.0614765270147473, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 1.5}) +[2026-06-23 13:00:17,185][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.05559263378381729, 'train/grad_norm_pre_clip': 31.367605209350586, 'timing/data': 0.0013325230102054775, 'timing/model': 2.036953313014237, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 1.5}) +[2026-06-23 13:00:21,473][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.09698568284511566, 'train/grad_norm_pre_clip': 24.519319534301758, 'timing/data': 0.0012197439791634679, 'timing/model': 2.293501979991561, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 1.5}) +[2026-06-23 13:00:25,565][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.09640967845916748, 'train/grad_norm_pre_clip': 22.775236129760742, 'timing/data': 0.05680260100052692, 'timing/model': 2.0381962949759327, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 1.5}) +[2026-06-23 13:00:29,689][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.027495412155985832, 'train/grad_norm_pre_clip': 16.723705291748047, 'timing/data': 0.001109670993173495, 'timing/model': 2.053944599989336, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 1.5}) +[2026-06-23 13:00:33,711][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.04123144969344139, 'train/grad_norm_pre_clip': 24.20262908935547, 'timing/data': 0.0011494240025058389, 'timing/model': 2.0318149979866575, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 1.5}) +[2026-06-23 13:00:37,740][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.05765604227781296, 'train/grad_norm_pre_clip': 20.90935516357422, 'timing/data': 0.001157730002887547, 'timing/model': 2.03793707099976, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 1.5}) +[2026-06-23 13:00:41,845][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.012192938476800919, 'train/grad_norm_pre_clip': 5.390373229980469, 'timing/data': 0.06663001500419341, 'timing/model': 2.0435314690112136, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 1.5}) +[2026-06-23 13:00:45,951][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.0934528261423111, 'train/grad_norm_pre_clip': 25.431447982788086, 'timing/data': 0.0012708790018223226, 'timing/model': 2.034512588987127, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 1.5}) +[2026-06-23 13:00:49,975][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.019374597817659378, 'train/grad_norm_pre_clip': 7.843750953674316, 'timing/data': 0.0011976340028923005, 'timing/model': 2.0336777739867102, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 1.5}) +[2026-06-23 13:00:54,058][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.03994070738554001, 'train/grad_norm_pre_clip': 19.370885848999023, 'timing/data': 0.0013370910019148141, 'timing/model': 2.0825786810019054, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 1.5}) +[2026-06-23 13:00:58,145][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.02470296062529087, 'train/grad_norm_pre_clip': 8.500697135925293, 'timing/data': 0.055294184014201164, 'timing/model': 2.0379550060024485, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 1.5}) +[2026-06-23 13:01:02,255][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.06779581308364868, 'train/grad_norm_pre_clip': 49.3661003112793, 'timing/data': 0.0012027229822706431, 'timing/model': 2.0362242849951144, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 1.51}) +[2026-06-23 13:01:06,296][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.07768909633159637, 'train/grad_norm_pre_clip': 29.89959144592285, 'timing/data': 0.001272470981348306, 'timing/model': 2.048542800010182, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 1.51}) +[2026-06-23 13:01:10,325][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.11526210606098175, 'train/grad_norm_pre_clip': 27.984668731689453, 'timing/data': 0.0018691139994189143, 'timing/model': 2.037169849994825, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 1.51}) +[2026-06-23 13:01:14,414][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.0734001025557518, 'train/grad_norm_pre_clip': 24.305749893188477, 'timing/data': 0.05652117100544274, 'timing/model': 2.036392762005562, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 1.51}) +[2026-06-23 13:01:18,525][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.02686757780611515, 'train/grad_norm_pre_clip': 11.391307830810547, 'timing/data': 0.0013438939931802452, 'timing/model': 2.0374354270170443, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 1.51}) +[2026-06-23 13:01:22,559][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.03785143047571182, 'train/grad_norm_pre_clip': 21.637454986572266, 'timing/data': 0.001280257012695074, 'timing/model': 2.0375507890130393, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 1.51}) +[2026-06-23 13:01:26,592][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.02916901186108589, 'train/grad_norm_pre_clip': 12.68630599975586, 'timing/data': 0.0017043900152202696, 'timing/model': 2.037854749010876, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 1.51}) +[2026-06-23 13:01:30,808][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.07068020105361938, 'train/grad_norm_pre_clip': 52.181663513183594, 'timing/data': 0.05383709000307135, 'timing/model': 2.039101472008042, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 1.51}) +[2026-06-23 13:01:34,930][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.05637471377849579, 'train/grad_norm_pre_clip': 36.65679931640625, 'timing/data': 0.0016185610147658736, 'timing/model': 2.0398114419949707, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 1.51}) +[2026-06-23 13:01:38,955][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.07232666015625, 'train/grad_norm_pre_clip': 36.13174819946289, 'timing/data': 0.0014246930077206343, 'timing/model': 2.0357467380235903, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 1.51}) +[2026-06-23 13:01:42,985][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.16041991114616394, 'train/grad_norm_pre_clip': 40.876773834228516, 'timing/data': 0.0012294820044189692, 'timing/model': 2.0355224320082925, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 1.51}) +[2026-06-23 13:01:47,080][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3900 +[2026-06-23 13:03:41,424][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3900 +[2026-06-23 13:03:41,424][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3900 +[2026-06-23 13:05:48,644][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3900 +[2026-06-23 13:05:48,645][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.06338005512952805, 'train/grad_norm_pre_clip': 15.113886833190918, 'eval/action_loss/seconds': 114.33959703601431, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04187161475419998, 'eval/latency_2/loss': 0.04187161475419998, 'eval/flappy/loss': 0.04187161475419998, 'eval/flappy/latency_2/loss': 0.04187161475419998, 'eval/flap_precision': 0.9585838343353373, 'eval/flap_recall': 0.9682860998650472, 'eval/flap_f1': 0.9634105404498153, 'eval/flap_rate_model': 0.20791666666666667, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9634105404498153, 'eval/latency_2/flap_precision': 0.9585838343353373, 'eval/latency_2/flap_recall': 0.9682860998650472, 'eval/latency_2/flap_f1': 0.9634105404498153, 'eval/latency_2/flap_rate_model': 0.20791666666666667, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9634105404498153, 'eval/action_classification/data_wait_seconds': 50.446327209472656, 'eval/action_classification/predict_seconds': 75.1322021484375, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06573832599679008, 'timing/model': 2.036101261008298, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 1.51}) +[2026-06-23 13:05:52,780][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.0939110815525055, 'train/grad_norm_pre_clip': 32.76312255859375, 'timing/data': 0.0011949390172958374, 'timing/model': 2.0953918939922005, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 1.51}) +[2026-06-23 13:05:56,765][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.0382625088095665, 'train/grad_norm_pre_clip': 16.769847869873047, 'timing/data': 0.0012244029785506427, 'timing/model': 2.0145318120194133, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 1.51}) +[2026-06-23 13:06:00,733][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.013212046585977077, 'train/grad_norm_pre_clip': 5.558804512023926, 'timing/data': 0.0012339900131337345, 'timing/model': 2.009846583008766, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 1.51}) +[2026-06-23 13:06:04,784][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.06832395493984222, 'train/grad_norm_pre_clip': 21.787137985229492, 'timing/data': 0.057151654007611796, 'timing/model': 2.0286692089866847, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 1.51}) +[2026-06-23 13:06:08,835][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.010070082731544971, 'train/grad_norm_pre_clip': 7.7493133544921875, 'timing/data': 0.0012002979929093271, 'timing/model': 2.007186273986008, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 1.51}) +[2026-06-23 13:06:12,808][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.008641651831567287, 'train/grad_norm_pre_clip': 5.980710506439209, 'timing/data': 0.0012209660199005157, 'timing/model': 2.0085063129954506, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 1.51}) +[2026-06-23 13:06:16,785][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.09033435583114624, 'train/grad_norm_pre_clip': 22.773515701293945, 'timing/data': 0.0012139839818701148, 'timing/model': 2.0092302989796735, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 1.51}) +[2026-06-23 13:06:20,823][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.13997039198875427, 'train/grad_norm_pre_clip': 61.47098159790039, 'timing/data': 0.05866125400643796, 'timing/model': 2.0162476890254766, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 1.51}) +[2026-06-23 13:06:24,896][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.05182116851210594, 'train/grad_norm_pre_clip': 20.50313949584961, 'timing/data': 0.0013601740065496415, 'timing/model': 2.0192673780256882, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 1.51}) +[2026-06-23 13:06:29,078][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.059221673756837845, 'train/grad_norm_pre_clip': 21.86794090270996, 'timing/data': 0.001256261981325224, 'timing/model': 2.03704590399866, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 1.51}) +[2026-06-23 13:06:33,090][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.03254155069589615, 'train/grad_norm_pre_clip': 22.86504364013672, 'timing/data': 0.001467152003897354, 'timing/model': 2.028830931027187, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 1.51}) +[2026-06-23 13:06:37,140][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.014478209428489208, 'train/grad_norm_pre_clip': 12.344209671020508, 'timing/data': 0.05808996700216085, 'timing/model': 2.019043712003622, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 1.51}) +[2026-06-23 13:06:41,224][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.018978718668222427, 'train/grad_norm_pre_clip': 11.66466999053955, 'timing/data': 0.0012452520022634417, 'timing/model': 2.0232999669970013, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 1.51}) +[2026-06-23 13:06:45,226][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.032795537263154984, 'train/grad_norm_pre_clip': 21.95909881591797, 'timing/data': 0.0012682439992204309, 'timing/model': 2.01938324799994, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 1.51}) +[2026-06-23 13:06:49,224][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.1416955441236496, 'train/grad_norm_pre_clip': 24.574283599853516, 'timing/data': 0.0012309449957683682, 'timing/model': 2.0195899789978284, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 1.52}) +[2026-06-23 13:06:53,284][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.13843174278736115, 'train/grad_norm_pre_clip': 24.745019912719727, 'timing/data': 0.05884323001373559, 'timing/model': 2.022711587982485, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 1.52}) +[2026-06-23 13:06:57,381][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.08908436447381973, 'train/grad_norm_pre_clip': 23.090564727783203, 'timing/data': 0.0012301329988986254, 'timing/model': 2.0309984709892888, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 1.52}) +[2026-06-23 13:07:01,407][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.03184976801276207, 'train/grad_norm_pre_clip': 14.211766242980957, 'timing/data': 0.0011549349874258041, 'timing/model': 2.0200769969960675, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 1.52}) +[2026-06-23 13:07:05,407][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.13455381989479065, 'train/grad_norm_pre_clip': 22.711977005004883, 'timing/data': 0.0011050330067519099, 'timing/model': 2.021342434018152, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 1.52}) +[2026-06-23 13:07:09,478][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.01374625414609909, 'train/grad_norm_pre_clip': 11.01146125793457, 'timing/data': 0.06460122202406637, 'timing/model': 2.026792165997904, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 1.52}) +[2026-06-23 13:07:13,593][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.06285415589809418, 'train/grad_norm_pre_clip': 21.90420150756836, 'timing/data': 0.0014640250010415912, 'timing/model': 2.0430407139938325, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 1.52}) +[2026-06-23 13:07:17,624][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.02247716300189495, 'train/grad_norm_pre_clip': 16.18404197692871, 'timing/data': 0.001426055998308584, 'timing/model': 2.0456896089890506, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 1.52}) +[2026-06-23 13:07:21,637][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.06643006205558777, 'train/grad_norm_pre_clip': 27.794384002685547, 'timing/data': 0.0012706480047199875, 'timing/model': 2.027274384017801, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 1.52}) +[2026-06-23 13:07:25,714][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.004676938988268375, 'train/grad_norm_pre_clip': 2.97003173828125, 'timing/data': 0.06746321101672947, 'timing/model': 2.027749742992455, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 1.52}) +[2026-06-23 13:07:29,814][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.06791214644908905, 'train/grad_norm_pre_clip': 35.12323760986328, 'timing/data': 0.0013965610123705119, 'timing/model': 2.02947147298255, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 1.52}) +[2026-06-23 13:07:33,822][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.06263571977615356, 'train/grad_norm_pre_clip': 29.3339786529541, 'timing/data': 0.0014067099837120622, 'timing/model': 2.022764917986933, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 1.52}) +[2026-06-23 13:07:38,078][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.06208100542426109, 'train/grad_norm_pre_clip': 35.633575439453125, 'timing/data': 0.001257043011719361, 'timing/model': 2.023299774999032, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 1.52}) +[2026-06-23 13:07:42,150][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.02103559859097004, 'train/grad_norm_pre_clip': 9.231284141540527, 'timing/data': 0.06717246302287094, 'timing/model': 2.025166654988425, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 1.52}) +[2026-06-23 13:07:46,276][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.20949116349220276, 'train/grad_norm_pre_clip': 36.407691955566406, 'timing/data': 0.0013364810147322714, 'timing/model': 2.048868087003939, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 1.52}) +[2026-06-23 13:07:50,287][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.08573121577501297, 'train/grad_norm_pre_clip': 31.281190872192383, 'timing/data': 0.0012466640037018806, 'timing/model': 2.027175756986253, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 1.52}) +[2026-06-23 13:07:54,309][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.04719496890902519, 'train/grad_norm_pre_clip': 21.353849411010742, 'timing/data': 0.0012217879993841052, 'timing/model': 2.0305403839738574, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 1.52}) +[2026-06-23 13:07:58,395][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.09841647744178772, 'train/grad_norm_pre_clip': 30.21384048461914, 'timing/data': 0.06758379298844375, 'timing/model': 2.03318687499268, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 1.52}) +[2026-06-23 13:08:02,520][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.038346078246831894, 'train/grad_norm_pre_clip': 20.677183151245117, 'timing/data': 0.0013902389910072088, 'timing/model': 2.0353983100212645, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 1.52}) +[2026-06-23 13:08:06,552][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.08106470853090286, 'train/grad_norm_pre_clip': 24.908702850341797, 'timing/data': 0.0013058440235909075, 'timing/model': 2.041585433995351, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 1.52}) +[2026-06-23 13:08:10,706][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.024399442598223686, 'train/grad_norm_pre_clip': 7.7059006690979, 'timing/data': 0.0014506510051432997, 'timing/model': 2.036118651012657, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 1.52}) +[2026-06-23 13:08:14,792][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.13096795976161957, 'train/grad_norm_pre_clip': 57.09385299682617, 'timing/data': 0.06774788600159809, 'timing/model': 2.030013165989658, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 1.52}) +[2026-06-23 13:08:18,896][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.010433807037770748, 'train/grad_norm_pre_clip': 5.1821980476379395, 'timing/data': 0.0014055380015634, 'timing/model': 2.029322668997338, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 1.52}) +[2026-06-23 13:08:22,929][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.01235644705593586, 'train/grad_norm_pre_clip': 5.4609503746032715, 'timing/data': 0.0014234400005079806, 'timing/model': 2.044172914000228, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 1.52}) +[2026-06-23 13:08:26,950][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.036530304700136185, 'train/grad_norm_pre_clip': 16.039440155029297, 'timing/data': 0.0012350720062386245, 'timing/model': 2.0319151889998466, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 1.52}) +[2026-06-23 13:08:31,042][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.16692718863487244, 'train/grad_norm_pre_clip': 44.79248046875, 'timing/data': 0.0695506070042029, 'timing/model': 2.0318688430124894, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 1.52}) +[2026-06-23 13:08:35,150][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.14472727477550507, 'train/grad_norm_pre_clip': 32.96824645996094, 'timing/data': 0.0013794889964628965, 'timing/model': 2.0306447900075, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 1.53}) +[2026-06-23 13:08:39,176][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.00704469159245491, 'train/grad_norm_pre_clip': 3.8359835147857666, 'timing/data': 0.0013263420260045677, 'timing/model': 2.035097518004477, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 1.53}) +[2026-06-23 13:08:43,457][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.0658612847328186, 'train/grad_norm_pre_clip': 27.623188018798828, 'timing/data': 0.0012121100153308362, 'timing/model': 2.292134232004173, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 1.53}) +[2026-06-23 13:08:47,549][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.11073032021522522, 'train/grad_norm_pre_clip': 42.79963302612305, 'timing/data': 0.06638112000655383, 'timing/model': 2.0343875650141854, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 1.53}) +[2026-06-23 13:08:51,659][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.09528285264968872, 'train/grad_norm_pre_clip': 30.471519470214844, 'timing/data': 0.0013831460091751069, 'timing/model': 2.0363933310145512, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 1.53}) +[2026-06-23 13:08:55,682][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.01830231584608555, 'train/grad_norm_pre_clip': 12.373589515686035, 'timing/data': 0.0012339499953668565, 'timing/model': 2.034765251009958, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 1.53}) +[2026-06-23 13:08:59,693][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.02643919363617897, 'train/grad_norm_pre_clip': 10.33716869354248, 'timing/data': 0.0013010239927098155, 'timing/model': 2.0261541670188308, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 1.53}) +[2026-06-23 13:09:03,760][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.03967977687716484, 'train/grad_norm_pre_clip': 14.956260681152344, 'timing/data': 0.05626926201512106, 'timing/model': 2.0263054560055025, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 1.53}) +[2026-06-23 13:09:07,862][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.08102389425039291, 'train/grad_norm_pre_clip': 19.385231018066406, 'timing/data': 0.0010972080053761601, 'timing/model': 2.030270852992544, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 1.53}) +[2026-06-23 13:09:11,872][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.08360939472913742, 'train/grad_norm_pre_clip': 42.013668060302734, 'timing/data': 0.0011546640016604215, 'timing/model': 2.027570982987527, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 1.53}) +[2026-06-23 13:09:16,122][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.04768018797039986, 'train/grad_norm_pre_clip': 37.22698974609375, 'timing/data': 0.0011298390163574368, 'timing/model': 2.2627287310024258, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 1.53}) +[2026-06-23 13:09:20,197][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.11972267180681229, 'train/grad_norm_pre_clip': 19.416866302490234, 'timing/data': 0.05730807199142873, 'timing/model': 2.0327419589739293, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 1.53}) +[2026-06-23 13:09:24,315][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.1813240796327591, 'train/grad_norm_pre_clip': 33.251224517822266, 'timing/data': 0.0017618359997868538, 'timing/model': 2.0355939590081107, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 1.53}) +[2026-06-23 13:09:28,340][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.012771910056471825, 'train/grad_norm_pre_clip': 5.6416425704956055, 'timing/data': 0.0011947879975195974, 'timing/model': 2.0344232239876874, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 1.53}) +[2026-06-23 13:09:32,364][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.05099214240908623, 'train/grad_norm_pre_clip': 37.99254608154297, 'timing/data': 0.0011046720028389245, 'timing/model': 2.035744415014051, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 1.53}) +[2026-06-23 13:09:36,457][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.024658245965838432, 'train/grad_norm_pre_clip': 19.86029624938965, 'timing/data': 0.06703823400312103, 'timing/model': 2.0313109310227446, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 1.53}) +[2026-06-23 13:09:40,570][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.030408663675189018, 'train/grad_norm_pre_clip': 14.981276512145996, 'timing/data': 0.0015292759926524013, 'timing/model': 2.0323324780038092, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 1.53}) +[2026-06-23 13:09:44,606][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.007295542396605015, 'train/grad_norm_pre_clip': 5.142452239990234, 'timing/data': 0.001384880015393719, 'timing/model': 2.048323461000109, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 1.53}) +[2026-06-23 13:09:48,666][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.03404425457119942, 'train/grad_norm_pre_clip': 17.91220474243164, 'timing/data': 0.0011956300004385412, 'timing/model': 2.072719396004686, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 1.53}) +[2026-06-23 13:09:52,750][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.05964414402842522, 'train/grad_norm_pre_clip': 34.77067565917969, 'timing/data': 0.06736974700470455, 'timing/model': 2.0296789330022875, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 1.53}) +[2026-06-23 13:09:56,857][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.02612181566655636, 'train/grad_norm_pre_clip': 9.833606719970703, 'timing/data': 0.0012107770016882569, 'timing/model': 2.0292704869934823, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 1.53}) +[2026-06-23 13:10:00,877][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.03247179463505745, 'train/grad_norm_pre_clip': 16.28977394104004, 'timing/data': 0.001228540000738576, 'timing/model': 2.0337661140074488, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 1.53}) +[2026-06-23 13:10:04,913][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.13263164460659027, 'train/grad_norm_pre_clip': 33.44865798950195, 'timing/data': 0.002018339990172535, 'timing/model': 2.040133672999218, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 1.53}) +[2026-06-23 13:10:09,003][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.11190919578075409, 'train/grad_norm_pre_clip': 29.044681549072266, 'timing/data': 0.06837659701704979, 'timing/model': 2.033312776009552, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 1.53}) +[2026-06-23 13:10:13,120][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.07818185538053513, 'train/grad_norm_pre_clip': 20.956026077270508, 'timing/data': 0.001669415010837838, 'timing/model': 2.0442697959952056, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 1.53}) +[2026-06-23 13:10:17,134][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.05748153477907181, 'train/grad_norm_pre_clip': 19.493846893310547, 'timing/data': 0.001310481980908662, 'timing/model': 2.0281451850023586, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 1.53}) +[2026-06-23 13:10:21,153][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.10289916396141052, 'train/grad_norm_pre_clip': 24.810104370117188, 'timing/data': 0.0012279190123081207, 'timing/model': 2.0301584849948995, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 1.54}) +[2026-06-23 13:10:25,346][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.11197438836097717, 'train/grad_norm_pre_clip': 34.502227783203125, 'timing/data': 0.06979362602578476, 'timing/model': 2.0325891060056165, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 1.54}) +[2026-06-23 13:10:29,451][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.015017816796898842, 'train/grad_norm_pre_clip': 7.407431602478027, 'timing/data': 0.0013624379935208708, 'timing/model': 2.033587712008739, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 1.54}) +[2026-06-23 13:10:33,471][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.032275713980197906, 'train/grad_norm_pre_clip': 19.47797203063965, 'timing/data': 0.00132344602025114, 'timing/model': 2.0298162210092414, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 1.54}) +[2026-06-23 13:10:37,499][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.012509824708104134, 'train/grad_norm_pre_clip': 11.173672676086426, 'timing/data': 0.0014380280044861138, 'timing/model': 2.0377027050126344, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 1.54}) +[2026-06-23 13:10:41,603][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.07319656014442444, 'train/grad_norm_pre_clip': 25.659456253051758, 'timing/data': 0.07091298498562537, 'timing/model': 2.0356201360118575, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 1.54}) +[2026-06-23 13:10:45,735][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.03418395295739174, 'train/grad_norm_pre_clip': 11.48514461517334, 'timing/data': 0.0012577750021591783, 'timing/model': 2.050659821019508, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 1.54}) +[2026-06-23 13:10:49,762][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.11792625486850739, 'train/grad_norm_pre_clip': 26.22614860534668, 'timing/data': 0.0016142830136232078, 'timing/model': 2.035181212995667, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 1.54}) +[2026-06-23 13:10:53,792][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.013824502006173134, 'train/grad_norm_pre_clip': 8.302105903625488, 'timing/data': 0.0013570279988925904, 'timing/model': 2.0370824559940957, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 1.54}) +[2026-06-23 13:10:58,107][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.0201954934746027, 'train/grad_norm_pre_clip': 12.78658676147461, 'timing/data': 0.058708208991447464, 'timing/model': 2.265664139005821, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 1.54}) +[2026-06-23 13:11:02,221][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.11051313579082489, 'train/grad_norm_pre_clip': 31.742855072021484, 'timing/data': 0.001255850016605109, 'timing/model': 2.035837805015035, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 1.54}) +[2026-06-23 13:11:06,261][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.1727617383003235, 'train/grad_norm_pre_clip': 31.713659286499023, 'timing/data': 0.0014228599902708083, 'timing/model': 2.0500636560027488, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 1.54}) +[2026-06-23 13:11:10,285][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.032698724418878555, 'train/grad_norm_pre_clip': 34.87190246582031, 'timing/data': 0.0013215630024205893, 'timing/model': 2.0333850809838623, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 1.54}) +[2026-06-23 13:11:14,370][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.09075657278299332, 'train/grad_norm_pre_clip': 25.862146377563477, 'timing/data': 0.056135253980755806, 'timing/model': 2.0366916620114353, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 1.54}) +[2026-06-23 13:11:18,484][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.11579608172178268, 'train/grad_norm_pre_clip': 23.767650604248047, 'timing/data': 0.0013980939984321594, 'timing/model': 2.0375403900106903, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 1.54}) +[2026-06-23 13:11:22,510][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.011037725023925304, 'train/grad_norm_pre_clip': 4.905412673950195, 'timing/data': 0.0013975530164316297, 'timing/model': 2.032791212986922, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 1.54}) +[2026-06-23 13:11:26,551][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.03625958412885666, 'train/grad_norm_pre_clip': 19.49634552001953, 'timing/data': 0.0012536880094558, 'timing/model': 2.0508335960039403, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 1.54}) +[2026-06-23 13:11:30,639][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.12831605970859528, 'train/grad_norm_pre_clip': 46.62577438354492, 'timing/data': 0.05944431902025826, 'timing/model': 2.0373502880102023, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 1.54}) +[2026-06-23 13:11:35,026][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.026816772297024727, 'train/grad_norm_pre_clip': 7.248237133026123, 'timing/data': 0.0013129560102242976, 'timing/model': 2.0429194789903704, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 1.54}) +[2026-06-23 13:11:39,052][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.030447082594037056, 'train/grad_norm_pre_clip': 19.855022430419922, 'timing/data': 0.0012682739761658013, 'timing/model': 2.0349323569971602, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 1.54}) +[2026-06-23 13:11:43,154][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.15865495800971985, 'train/grad_norm_pre_clip': 36.19561004638672, 'timing/data': 0.001234031020430848, 'timing/model': 2.0388971750217024, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 1.54}) +[2026-06-23 13:11:47,252][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.09765960276126862, 'train/grad_norm_pre_clip': 42.87242889404297, 'timing/data': 0.07114516399451531, 'timing/model': 2.0349022609880194, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 1.54}) +[2026-06-23 13:11:51,361][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.06272206455469131, 'train/grad_norm_pre_clip': 33.161197662353516, 'timing/data': 0.0013627390144392848, 'timing/model': 2.0362174410256557, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 1.54}) +[2026-06-23 13:11:55,396][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.03453999012708664, 'train/grad_norm_pre_clip': 17.011520385742188, 'timing/data': 0.001287118997424841, 'timing/model': 2.044853702012915, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 1.54}) +[2026-06-23 13:11:59,417][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.1085001677274704, 'train/grad_norm_pre_clip': 31.457263946533203, 'timing/data': 0.001389177021337673, 'timing/model': 2.031982564018108, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 1.54}) +[2026-06-23 13:12:03,496][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.0876651331782341, 'train/grad_norm_pre_clip': 24.8245906829834, 'timing/data': 0.05656702199485153, 'timing/model': 2.0346621650096495, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 1.54}) +[2026-06-23 13:12:07,762][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.02934020385146141, 'train/grad_norm_pre_clip': 21.194475173950195, 'timing/data': 0.001212241011671722, 'timing/model': 2.033320515998639, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 1.55}) +[2026-06-23 13:12:11,787][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.058562129735946655, 'train/grad_norm_pre_clip': 28.039752960205078, 'timing/data': 0.001407472009304911, 'timing/model': 2.0342286530067213, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 1.55}) +[2026-06-23 13:12:15,816][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.03577137738466263, 'train/grad_norm_pre_clip': 14.451882362365723, 'timing/data': 0.0012236610054969788, 'timing/model': 2.036041622981429, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 1.55}) +[2026-06-23 13:12:19,905][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.13046635687351227, 'train/grad_norm_pre_clip': 51.75276565551758, 'timing/data': 0.0567149949783925, 'timing/model': 2.035307705984451, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 1.55}) +[2026-06-23 13:12:24,017][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.020232301205396652, 'train/grad_norm_pre_clip': 9.618144989013672, 'timing/data': 0.0015758619992993772, 'timing/model': 2.03061217800132, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 1.55}) +[2026-06-23 13:12:28,060][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.009530295617878437, 'train/grad_norm_pre_clip': 7.987123012542725, 'timing/data': 0.0014292510168161243, 'timing/model': 2.05119468702469, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 1.55}) +[2026-06-23 13:12:32,111][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.013860952109098434, 'train/grad_norm_pre_clip': 8.36573600769043, 'timing/data': 0.0012083330075256526, 'timing/model': 2.0577705889882054, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 1.55}) +[2026-06-23 13:12:36,220][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4000 +[2026-06-23 13:14:31,038][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4000 +[2026-06-23 13:14:31,039][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4000 +[2026-06-23 13:16:38,465][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4000 +[2026-06-23 13:30:24,290][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000 +[2026-06-23 13:30:24,292][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.13327781856060028, 'train/grad_norm_pre_clip': 36.610595703125, 'eval/action_loss/seconds': 114.81447550500161, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04215487465262413, 'eval/latency_2/loss': 0.04215487465262413, 'eval/flappy/loss': 0.04215487465262413, 'eval/flappy/latency_2/loss': 0.04215487465262413, 'eval/flap_precision': 0.9653767820773931, 'eval/flap_recall': 0.9595141700404858, 'eval/flap_f1': 0.9624365482233502, 'eval/flap_rate_model': 0.20458333333333334, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9624365482233501, 'eval/latency_2/flap_precision': 0.9653767820773931, 'eval/latency_2/flap_recall': 0.9595141700404858, 'eval/latency_2/flap_f1': 0.9624365482233502, 'eval/latency_2/flap_rate_model': 0.20458333333333334, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9624365482233501, 'eval/action_classification/data_wait_seconds': 50.56044387817383, 'eval/action_classification/predict_seconds': 75.2252426147461, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 213.68000000001044, 'rl_games_eval/mid_train/mean_length': 1740.2, 'rl_games_eval/mid_train/std_reward': 188.710353717025, 'rl_games_eval/mid_train/std_length': 1519.861230507575, 'rl_games_eval/mid_train/macro_mean_reward': 213.68000000001044, 'rl_games_eval/mid_train/macro_mean_length': 1740.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 213.68000000001044, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1740.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 188.710353717025, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1519.861230507575, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 213.68000000001044, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.06773761598742567, 'timing/model': 2.044641294021858, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 1.55}) +[2026-06-23 13:30:24,293][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_4000_state +[2026-06-23 13:30:24,300][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 13:30:33,370][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_4000_state/model.safetensors +[2026-06-23 13:30:37,329][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_4000_state/optimizer.bin +[2026-06-23 13:30:37,330][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_4000_state/sampler.bin +[2026-06-23 13:30:37,330][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-23 13:30:37,331][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-23 13:30:37,469][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-23 13:30:37,469][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 13:30:37,496][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 13:30:41,485][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.10359163582324982, 'train/grad_norm_pre_clip': 34.623538970947266, 'timing/data': 0.0013915410090703517, 'timing/model': 1.9613195290148724, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 1.55}) +[2026-06-23 13:30:45,364][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.013918839395046234, 'train/grad_norm_pre_clip': 6.208139419555664, 'timing/data': 0.0012400319974403828, 'timing/model': 1.9616114259988535, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 1.55}) +[2026-06-23 13:30:49,284][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.06265673786401749, 'train/grad_norm_pre_clip': 20.790843963623047, 'timing/data': 0.0012341210094746202, 'timing/model': 1.9959549909981433, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 1.55}) +[2026-06-23 13:30:53,271][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.012231746688485146, 'train/grad_norm_pre_clip': 7.29651403427124, 'timing/data': 0.07140595800592564, 'timing/model': 1.9862820799753536, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 1.55}) +[2026-06-23 13:30:57,265][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.056722186505794525, 'train/grad_norm_pre_clip': 20.614456176757812, 'timing/data': 0.0012058990250807256, 'timing/model': 1.9741911799937952, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 1.55}) +[2026-06-23 13:31:01,346][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.03228403255343437, 'train/grad_norm_pre_clip': 15.53769302368164, 'timing/data': 0.0012294220214243978, 'timing/model': 2.149399932997767, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 1.55}) +[2026-06-23 13:31:05,283][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.17355377972126007, 'train/grad_norm_pre_clip': 42.126399993896484, 'timing/data': 0.0012467140040826052, 'timing/model': 1.9919682330219075, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 1.55}) +[2026-06-23 13:31:09,255][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.14367739856243134, 'train/grad_norm_pre_clip': 22.03508186340332, 'timing/data': 0.057672494993312284, 'timing/model': 1.979854241013527, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 1.55}) +[2026-06-23 13:31:13,262][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.11998885124921799, 'train/grad_norm_pre_clip': 27.855674743652344, 'timing/data': 0.0013399460003711283, 'timing/model': 1.981093428999884, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 1.55}) +[2026-06-23 13:31:17,195][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.06388632208108902, 'train/grad_norm_pre_clip': 23.237276077270508, 'timing/data': 0.001248228014446795, 'timing/model': 1.9928335150179919, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 1.55}) +[2026-06-23 13:31:21,130][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.10204622149467468, 'train/grad_norm_pre_clip': 27.03526496887207, 'timing/data': 0.001118726999266073, 'timing/model': 1.9860908800037578, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 1.55}) +[2026-06-23 13:31:25,139][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.07300177216529846, 'train/grad_norm_pre_clip': 20.640047073364258, 'timing/data': 0.07530417200177908, 'timing/model': 1.9914484109904151, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 1.55}) +[2026-06-23 13:31:29,163][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.05213667452335358, 'train/grad_norm_pre_clip': 16.95929718017578, 'timing/data': 0.0012875099782831967, 'timing/model': 1.989701029000571, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 1.55}) +[2026-06-23 13:31:33,360][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.00965492520481348, 'train/grad_norm_pre_clip': 6.7702178955078125, 'timing/data': 0.0011945280130021274, 'timing/model': 2.248105498001678, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 1.55}) +[2026-06-23 13:31:37,302][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.04763878136873245, 'train/grad_norm_pre_clip': 16.85512924194336, 'timing/data': 0.0011210020165890455, 'timing/model': 1.9934690480004065, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 1.55}) +[2026-06-23 13:31:41,309][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.01739867962896824, 'train/grad_norm_pre_clip': 5.513511657714844, 'timing/data': 0.05842154900892638, 'timing/model': 1.9932946839835495, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 1.55}) +[2026-06-23 13:31:45,346][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.02448548749089241, 'train/grad_norm_pre_clip': 10.157315254211426, 'timing/data': 0.0011251899995841086, 'timing/model': 1.99542363701039, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 1.55}) +[2026-06-23 13:31:49,310][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.062371786683797836, 'train/grad_norm_pre_clip': 21.658477783203125, 'timing/data': 0.001243036997038871, 'timing/model': 2.0047404050128534, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 1.56}) +[2026-06-23 13:31:53,288][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.01723824255168438, 'train/grad_norm_pre_clip': 13.167442321777344, 'timing/data': 0.0014187720080371946, 'timing/model': 2.0134748959972057, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 1.56}) +[2026-06-23 13:31:57,313][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.04733913019299507, 'train/grad_norm_pre_clip': 26.786996841430664, 'timing/data': 0.06657351600006223, 'timing/model': 2.0017140270210803, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 1.56}) +[2026-06-23 13:32:01,363][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.19452989101409912, 'train/grad_norm_pre_clip': 31.41085433959961, 'timing/data': 0.0011573689989745617, 'timing/model': 2.004975182004273, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 1.56}) +[2026-06-23 13:32:05,328][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.07840437442064285, 'train/grad_norm_pre_clip': 31.047855377197266, 'timing/data': 0.001147531991591677, 'timing/model': 2.0040845639887266, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 1.56}) +[2026-06-23 13:32:09,338][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.01874392293393612, 'train/grad_norm_pre_clip': 12.51222038269043, 'timing/data': 0.001140198000939563, 'timing/model': 2.007510955008911, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 1.56}) +[2026-06-23 13:32:13,381][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.05596960708498955, 'train/grad_norm_pre_clip': 20.878856658935547, 'timing/data': 0.07101378298830241, 'timing/model': 2.0088011480111163, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 1.56}) +[2026-06-23 13:32:17,442][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.12303567677736282, 'train/grad_norm_pre_clip': 39.99889373779297, 'timing/data': 0.0011184469913132489, 'timing/model': 2.011595190007938, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 1.56}) +[2026-06-23 13:32:21,422][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.106206014752388, 'train/grad_norm_pre_clip': 31.544992446899414, 'timing/data': 0.001185160013847053, 'timing/model': 2.0101697130012326, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 1.56}) +[2026-06-23 13:32:25,405][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.046529605984687805, 'train/grad_norm_pre_clip': 17.115570068359375, 'timing/data': 0.001132964011048898, 'timing/model': 2.0133663790184073, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 1.56}) +[2026-06-23 13:32:29,444][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.017636172473430634, 'train/grad_norm_pre_clip': 13.447258949279785, 'timing/data': 0.05559117699158378, 'timing/model': 2.0127810460107867, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 1.56}) +[2026-06-23 13:32:33,517][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.1259748786687851, 'train/grad_norm_pre_clip': 35.14246368408203, 'timing/data': 0.0011096909875050187, 'timing/model': 2.0144961240002885, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 1.56}) +[2026-06-23 13:32:37,506][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.0389404296875, 'train/grad_norm_pre_clip': 17.300046920776367, 'timing/data': 0.0011484820279292762, 'timing/model': 2.0169850819802377, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 1.56}) +[2026-06-23 13:32:41,497][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.01918034628033638, 'train/grad_norm_pre_clip': 15.021585464477539, 'timing/data': 0.0011215129925403744, 'timing/model': 2.017685324011836, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 1.56}) +[2026-06-23 13:32:45,718][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.022166147828102112, 'train/grad_norm_pre_clip': 11.876605033874512, 'timing/data': 0.057593598001403734, 'timing/model': 2.016362778027542, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 1.56}) +[2026-06-23 13:32:49,791][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.05179673433303833, 'train/grad_norm_pre_clip': 12.167343139648438, 'timing/data': 0.0011264420172665268, 'timing/model': 2.0159139560128096, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 1.56}) +[2026-06-23 13:32:53,784][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.08516997843980789, 'train/grad_norm_pre_clip': 36.0660285949707, 'timing/data': 0.0011861720122396946, 'timing/model': 2.0193600929924287, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 1.56}) +[2026-06-23 13:32:57,783][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.07329163700342178, 'train/grad_norm_pre_clip': 35.97575759887695, 'timing/data': 0.0011267329973634332, 'timing/model': 2.020923762000166, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 1.56}) +[2026-06-23 13:33:01,841][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.07660545408725739, 'train/grad_norm_pre_clip': 21.124248504638672, 'timing/data': 0.06529443099861965, 'timing/model': 2.018070716992952, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 1.56}) +[2026-06-23 13:33:05,924][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.04399280995130539, 'train/grad_norm_pre_clip': 27.027751922607422, 'timing/data': 0.0012820000119972974, 'timing/model': 2.022658984991722, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 1.56}) +[2026-06-23 13:33:09,929][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.04410490393638611, 'train/grad_norm_pre_clip': 18.828107833862305, 'timing/data': 0.001185070985229686, 'timing/model': 2.0231068519933615, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 1.56}) +[2026-06-23 13:33:13,939][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.12063624709844589, 'train/grad_norm_pre_clip': 39.35117721557617, 'timing/data': 0.0012831909989472479, 'timing/model': 2.0268900020164438, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 1.56}) +[2026-06-23 13:33:18,032][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.09492065012454987, 'train/grad_norm_pre_clip': 29.9645938873291, 'timing/data': 0.05345803400268778, 'timing/model': 2.0247979529958684, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 1.56}) +[2026-06-23 13:33:22,129][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.029555900022387505, 'train/grad_norm_pre_clip': 42.49169158935547, 'timing/data': 0.001228790992172435, 'timing/model': 2.0281518999836408, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 1.56}) +[2026-06-23 13:33:26,144][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.14578202366828918, 'train/grad_norm_pre_clip': 36.5299186706543, 'timing/data': 0.0012436289980541915, 'timing/model': 2.0293192469980568, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 1.56}) +[2026-06-23 13:33:30,174][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.06277080625295639, 'train/grad_norm_pre_clip': 22.030912399291992, 'timing/data': 0.0012303139956202358, 'timing/model': 2.045062972989399, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 1.56}) +[2026-06-23 13:33:34,242][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.13680294156074524, 'train/grad_norm_pre_clip': 18.84319305419922, 'timing/data': 0.05599135599914007, 'timing/model': 2.02972740700352, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 1.57}) +[2026-06-23 13:33:38,341][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.062180664390325546, 'train/grad_norm_pre_clip': 24.165658950805664, 'timing/data': 0.001422389002982527, 'timing/model': 2.0274184870067984, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 1.57}) +[2026-06-23 13:33:42,361][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.04046697914600372, 'train/grad_norm_pre_clip': 18.503969192504883, 'timing/data': 0.0012677930062636733, 'timing/model': 2.031524414982414, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 1.57}) +[2026-06-23 13:33:46,396][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.07736300677061081, 'train/grad_norm_pre_clip': 28.306955337524414, 'timing/data': 0.0013743699819315225, 'timing/model': 2.046815144014545, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 1.57}) +[2026-06-23 13:33:50,531][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.03785599768161774, 'train/grad_norm_pre_clip': 19.854387283325195, 'timing/data': 0.07005954600754194, 'timing/model': 2.0748442460026126, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 1.57}) +[2026-06-23 13:33:54,642][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.03975282981991768, 'train/grad_norm_pre_clip': 20.850942611694336, 'timing/data': 0.0014577240217477083, 'timing/model': 2.03122961500776, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 1.57}) +[2026-06-23 13:33:58,661][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.054046932607889175, 'train/grad_norm_pre_clip': 29.04100227355957, 'timing/data': 0.0013000819890294224, 'timing/model': 2.031396100996062, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 1.57}) +[2026-06-23 13:34:02,679][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.042815547436475754, 'train/grad_norm_pre_clip': 19.865554809570312, 'timing/data': 0.0012556510046124458, 'timing/model': 2.0311866420088336, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 1.57}) +[2026-06-23 13:34:06,761][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.06862510740756989, 'train/grad_norm_pre_clip': 31.07590675354004, 'timing/data': 0.06674909798312001, 'timing/model': 2.029585030017188, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 1.57}) +[2026-06-23 13:34:10,865][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.009433373808860779, 'train/grad_norm_pre_clip': 4.5935235023498535, 'timing/data': 0.0014087540039326996, 'timing/model': 2.029580429982161, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 1.57}) +[2026-06-23 13:34:14,898][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.0992092564702034, 'train/grad_norm_pre_clip': 29.32162857055664, 'timing/data': 0.001236055017216131, 'timing/model': 2.041918873001123, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 1.57}) +[2026-06-23 13:34:18,923][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.07482743263244629, 'train/grad_norm_pre_clip': 20.05916976928711, 'timing/data': 0.0014808370033279061, 'timing/model': 2.030958223011112, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 1.57}) +[2026-06-23 13:34:23,009][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.021300585940480232, 'train/grad_norm_pre_clip': 10.381155967712402, 'timing/data': 0.0595185240090359, 'timing/model': 2.0324793530162424, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 1.57}) +[2026-06-23 13:34:27,152][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.05719955265522003, 'train/grad_norm_pre_clip': 21.20318031311035, 'timing/data': 0.0012721710081677884, 'timing/model': 2.041110222984571, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 1.57}) +[2026-06-23 13:34:31,190][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.1517912745475769, 'train/grad_norm_pre_clip': 32.63593673706055, 'timing/data': 0.0012547789956443012, 'timing/model': 2.0294519340095576, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 1.57}) +[2026-06-23 13:34:35,234][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.055030614137649536, 'train/grad_norm_pre_clip': 21.427555084228516, 'timing/data': 0.001648216973990202, 'timing/model': 2.0394410449953284, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 1.57}) +[2026-06-23 13:34:39,325][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.04283592849969864, 'train/grad_norm_pre_clip': 16.274171829223633, 'timing/data': 0.06818252598168328, 'timing/model': 2.0305617330013774, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 1.57}) +[2026-06-23 13:34:43,427][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.061246566474437714, 'train/grad_norm_pre_clip': 19.216764450073242, 'timing/data': 0.0013512079895008355, 'timing/model': 2.030305838008644, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 1.57}) +[2026-06-23 13:34:47,448][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.06370008736848831, 'train/grad_norm_pre_clip': 29.95830726623535, 'timing/data': 0.0012520340096671134, 'timing/model': 2.03285056000459, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 1.57}) +[2026-06-23 13:34:51,483][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.11621692776679993, 'train/grad_norm_pre_clip': 37.577091217041016, 'timing/data': 0.001230855006724596, 'timing/model': 2.047365474019898, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 1.57}) +[2026-06-23 13:34:55,564][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.05529063940048218, 'train/grad_norm_pre_clip': 17.234947204589844, 'timing/data': 0.058211839001160115, 'timing/model': 2.0347315930121113, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 1.57}) +[2026-06-23 13:34:59,715][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.02735002711415291, 'train/grad_norm_pre_clip': 16.241003036499023, 'timing/data': 0.001406960014719516, 'timing/model': 2.0743190789944492, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 1.57}) +[2026-06-23 13:35:03,748][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.1564100682735443, 'train/grad_norm_pre_clip': 26.532869338989258, 'timing/data': 0.0013390439853537828, 'timing/model': 2.0404598870081827, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 1.57}) +[2026-06-23 13:35:07,791][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.03626761585474014, 'train/grad_norm_pre_clip': 20.428237915039062, 'timing/data': 0.0012672219891101122, 'timing/model': 2.052831551991403, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 1.57}) +[2026-06-23 13:35:11,894][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.047961246222257614, 'train/grad_norm_pre_clip': 24.439773559570312, 'timing/data': 0.06952777301194146, 'timing/model': 2.04062423799769, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 1.57}) +[2026-06-23 13:35:16,026][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.09423459321260452, 'train/grad_norm_pre_clip': 34.186073303222656, 'timing/data': 0.0014288309903349727, 'timing/model': 2.0551315199991222, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 1.57}) +[2026-06-23 13:35:20,054][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.09030331671237946, 'train/grad_norm_pre_clip': 10.266229629516602, 'timing/data': 0.001684843999100849, 'timing/model': 2.0368629330187105, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 1.58}) +[2026-06-23 13:35:24,107][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.11573564261198044, 'train/grad_norm_pre_clip': 48.23691940307617, 'timing/data': 0.0013004239881411195, 'timing/model': 2.0561913450073916, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 1.58}) +[2026-06-23 13:35:28,218][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.02934623323380947, 'train/grad_norm_pre_clip': 34.23250198364258, 'timing/data': 0.07945039198966697, 'timing/model': 2.041499359998852, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 1.58}) +[2026-06-23 13:35:32,344][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.03973791003227234, 'train/grad_norm_pre_clip': 11.503532409667969, 'timing/data': 0.001452385011361912, 'timing/model': 2.0457692999916617, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 1.58}) +[2026-06-23 13:35:36,537][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.010869311168789864, 'train/grad_norm_pre_clip': 11.930999755859375, 'timing/data': 0.0013942669902462512, 'timing/model': 2.0503700819972437, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 1.58}) +[2026-06-23 13:35:40,562][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.021415749564766884, 'train/grad_norm_pre_clip': 7.891144275665283, 'timing/data': 0.001489602989749983, 'timing/model': 2.0331516269943677, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 1.58}) +[2026-06-23 13:35:44,644][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.004585196264088154, 'train/grad_norm_pre_clip': 2.287684202194214, 'timing/data': 0.0585292249743361, 'timing/model': 2.0333941940043587, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 1.58}) +[2026-06-23 13:35:48,759][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.06982342153787613, 'train/grad_norm_pre_clip': 16.85169219970703, 'timing/data': 0.0012286610144656152, 'timing/model': 2.0405051150009967, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 1.58}) +[2026-06-23 13:35:52,796][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.06572532653808594, 'train/grad_norm_pre_clip': 23.91514015197754, 'timing/data': 0.001456913014408201, 'timing/model': 2.034958882984938, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 1.58}) +[2026-06-23 13:35:56,839][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.005733494181185961, 'train/grad_norm_pre_clip': 3.6916284561157227, 'timing/data': 0.00125111301895231, 'timing/model': 2.041663793992484, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 1.58}) +[2026-06-23 13:36:00,941][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.08998091518878937, 'train/grad_norm_pre_clip': 26.261171340942383, 'timing/data': 0.06604117900133133, 'timing/model': 2.0362565889954567, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 1.58}) +[2026-06-23 13:36:05,061][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.013578510843217373, 'train/grad_norm_pre_clip': 8.942397117614746, 'timing/data': 0.0012362650013528764, 'timing/model': 2.0406384249799885, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 1.58}) +[2026-06-23 13:36:09,350][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.015475509688258171, 'train/grad_norm_pre_clip': 11.157646179199219, 'timing/data': 0.0013020960032008588, 'timing/model': 2.0453296350024175, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 1.58}) +[2026-06-23 13:36:13,385][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.16688716411590576, 'train/grad_norm_pre_clip': 29.588336944580078, 'timing/data': 0.001369741978123784, 'timing/model': 2.0367935269896407, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 1.58}) +[2026-06-23 13:36:17,472][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.058654267340898514, 'train/grad_norm_pre_clip': 18.250524520874023, 'timing/data': 0.05875814802129753, 'timing/model': 2.0338374719722196, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 1.58}) +[2026-06-23 13:36:21,591][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.12968087196350098, 'train/grad_norm_pre_clip': 27.09471893310547, 'timing/data': 0.0012613410071935505, 'timing/model': 2.0399891000124626, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 1.58}) +[2026-06-23 13:36:25,633][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.027761057019233704, 'train/grad_norm_pre_clip': 23.749418258666992, 'timing/data': 0.0014121900021564215, 'timing/model': 2.050724109984003, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 1.58}) +[2026-06-23 13:36:29,666][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.07009577751159668, 'train/grad_norm_pre_clip': 26.7381591796875, 'timing/data': 0.0012148649839218706, 'timing/model': 2.0394146059989, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 1.58}) +[2026-06-23 13:36:33,760][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.06239086762070656, 'train/grad_norm_pre_clip': 22.93990707397461, 'timing/data': 0.06571574602276087, 'timing/model': 2.0323382569768, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 1.58}) +[2026-06-23 13:36:37,867][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.020807813853025436, 'train/grad_norm_pre_clip': 17.743228912353516, 'timing/data': 0.0011233759869355708, 'timing/model': 2.040242204006063, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 1.58}) +[2026-06-23 13:36:42,129][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.04494766145944595, 'train/grad_norm_pre_clip': 16.771438598632812, 'timing/data': 0.001162699016276747, 'timing/model': 2.272961586015299, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 1.58}) +[2026-06-23 13:36:46,162][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.09102433174848557, 'train/grad_norm_pre_clip': 23.52781105041504, 'timing/data': 0.0011455369822215289, 'timing/model': 2.04315238900017, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 1.58}) +[2026-06-23 13:36:50,244][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.16695120930671692, 'train/grad_norm_pre_clip': 30.13713264465332, 'timing/data': 0.056081319984514266, 'timing/model': 2.034371159010334, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 1.58}) +[2026-06-23 13:36:54,350][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.020682718604803085, 'train/grad_norm_pre_clip': 12.245438575744629, 'timing/data': 0.0012716399796772748, 'timing/model': 2.0335127209837083, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 1.58}) +[2026-06-23 13:36:58,398][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.245126873254776, 'train/grad_norm_pre_clip': 63.727901458740234, 'timing/data': 0.0012761580001097172, 'timing/model': 2.0470535359927453, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 1.58}) +[2026-06-23 13:37:02,429][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.16811974346637726, 'train/grad_norm_pre_clip': 36.326019287109375, 'timing/data': 0.0013029789843130857, 'timing/model': 2.0419291940052062, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 1.58}) +[2026-06-23 13:37:06,520][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.01703733205795288, 'train/grad_norm_pre_clip': 7.412222385406494, 'timing/data': 0.05283715401310474, 'timing/model': 2.0475502320041414, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 1.59}) +[2026-06-23 13:37:10,638][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.14241382479667664, 'train/grad_norm_pre_clip': 34.5190315246582, 'timing/data': 0.0013725970056839287, 'timing/model': 2.0368409880029503, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 1.59}) +[2026-06-23 13:37:14,670][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.08723516017198563, 'train/grad_norm_pre_clip': 27.568939208984375, 'timing/data': 0.0012742350227199495, 'timing/model': 2.0378196149831638, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 1.59}) +[2026-06-23 13:37:18,806][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.01555643044412136, 'train/grad_norm_pre_clip': 13.722574234008789, 'timing/data': 0.001248607994057238, 'timing/model': 2.036981504992582, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 1.59}) +[2026-06-23 13:37:22,882][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4100 +[2026-06-23 13:39:16,427][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4100 +[2026-06-23 13:39:16,427][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4100 +[2026-06-23 13:41:23,749][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4100 +[2026-06-23 13:41:23,750][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.03071451187133789, 'train/grad_norm_pre_clip': 29.341882705688477, 'eval/action_loss/seconds': 113.54053515399573, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04165314510464668, 'eval/latency_2/loss': 0.04165314510464668, 'eval/flappy/loss': 0.04165314510464668, 'eval/flappy/latency_2/loss': 0.04165314510464668, 'eval/flap_precision': 0.9505277044854882, 'eval/flap_recall': 0.9723346828609987, 'eval/flap_f1': 0.961307538358906, 'eval/flap_rate_model': 0.21055555555555555, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9613075383589059, 'eval/latency_2/flap_precision': 0.9505277044854882, 'eval/latency_2/flap_recall': 0.9723346828609987, 'eval/latency_2/flap_f1': 0.961307538358906, 'eval/latency_2/flap_rate_model': 0.21055555555555555, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9613075383589059, 'eval/action_classification/data_wait_seconds': 50.413631439208984, 'eval/action_classification/predict_seconds': 75.15959930419922, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.055636017001233995, 'timing/model': 2.0323774649878033, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 1.59}) +[2026-06-23 13:41:27,786][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.00617566704750061, 'train/grad_norm_pre_clip': 2.8850998878479004, 'timing/data': 0.0011468789889477193, 'timing/model': 2.001806399988709, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 1.59}) +[2026-06-23 13:41:31,753][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.021975498646497726, 'train/grad_norm_pre_clip': 21.138689041137695, 'timing/data': 0.0013783979811705649, 'timing/model': 1.996015846991213, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 1.59}) +[2026-06-23 13:41:35,715][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.04268546402454376, 'train/grad_norm_pre_clip': 19.603870391845703, 'timing/data': 0.0012150649854447693, 'timing/model': 2.0013436730077956, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 1.59}) +[2026-06-23 13:41:39,779][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.07657969743013382, 'train/grad_norm_pre_clip': 28.024385452270508, 'timing/data': 0.0531922280206345, 'timing/model': 2.054593587992713, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 1.59}) +[2026-06-23 13:41:43,862][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.0428343340754509, 'train/grad_norm_pre_clip': 21.217121124267578, 'timing/data': 0.001243298000190407, 'timing/model': 2.031340972025646, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 1.59}) +[2026-06-23 13:41:47,831][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.11658013612031937, 'train/grad_norm_pre_clip': 37.26628875732422, 'timing/data': 0.0014048159937374294, 'timing/model': 2.006538702000398, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 1.59}) +[2026-06-23 13:41:51,807][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.02273430861532688, 'train/grad_norm_pre_clip': 11.806158065795898, 'timing/data': 0.0012386390008032322, 'timing/model': 2.0062264440057334, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 1.59}) +[2026-06-23 13:41:55,855][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.05539022386074066, 'train/grad_norm_pre_clip': 15.912365913391113, 'timing/data': 0.07265609200112522, 'timing/model': 2.00813247699989, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 1.59}) +[2026-06-23 13:41:59,944][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.10825566202402115, 'train/grad_norm_pre_clip': 19.46058464050293, 'timing/data': 0.0012108380033168942, 'timing/model': 2.030898149998393, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 1.59}) +[2026-06-23 13:42:03,944][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.053836457431316376, 'train/grad_norm_pre_clip': 34.17045593261719, 'timing/data': 0.0014470340101979673, 'timing/model': 2.019819876004476, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 1.59}) +[2026-06-23 13:42:07,942][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.005039167124778032, 'train/grad_norm_pre_clip': 3.229260206222534, 'timing/data': 0.0012264270044397563, 'timing/model': 2.0146871189936064, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 1.59}) +[2026-06-23 13:42:12,074][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.04945583641529083, 'train/grad_norm_pre_clip': 35.261444091796875, 'timing/data': 0.12708245002431795, 'timing/model': 2.027190557011636, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 1.59}) +[2026-06-23 13:42:16,348][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.18059033155441284, 'train/grad_norm_pre_clip': 40.058677673339844, 'timing/data': 0.0012556909932754934, 'timing/model': 2.0330785179976374, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 1.59}) +[2026-06-23 13:42:20,344][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.10556447505950928, 'train/grad_norm_pre_clip': 29.786352157592773, 'timing/data': 0.0013053329894319177, 'timing/model': 2.0248260450025555, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 1.59}) +[2026-06-23 13:42:24,357][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.04387825354933739, 'train/grad_norm_pre_clip': 39.92160415649414, 'timing/data': 0.0013784980110358447, 'timing/model': 2.0324564870097674, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 1.59}) +[2026-06-23 13:42:28,421][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.046062201261520386, 'train/grad_norm_pre_clip': 19.89406394958496, 'timing/data': 0.06577994299004786, 'timing/model': 2.0225357720046304, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 1.59}) +[2026-06-23 13:42:32,521][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.04623541608452797, 'train/grad_norm_pre_clip': 21.235454559326172, 'timing/data': 0.0014071110053919256, 'timing/model': 2.029955003003124, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 1.59}) +[2026-06-23 13:42:36,528][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.039232391864061356, 'train/grad_norm_pre_clip': 17.639589309692383, 'timing/data': 0.0013864919892512262, 'timing/model': 2.0263100059819408, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 1.59}) +[2026-06-23 13:42:40,529][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.08475913852453232, 'train/grad_norm_pre_clip': 47.53612518310547, 'timing/data': 0.0015709539875388145, 'timing/model': 2.021396795025794, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 1.59}) +[2026-06-23 13:42:44,600][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.10192125290632248, 'train/grad_norm_pre_clip': 25.370933532714844, 'timing/data': 0.06722688701120205, 'timing/model': 2.0200242069840897, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 1.59}) +[2026-06-23 13:42:48,739][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.044926967471838, 'train/grad_norm_pre_clip': 28.852584838867188, 'timing/data': 0.0011701930197887123, 'timing/model': 2.025414769013878, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 1.59}) +[2026-06-23 13:42:52,750][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.11870861798524857, 'train/grad_norm_pre_clip': 22.525604248046875, 'timing/data': 0.0012544580094981939, 'timing/model': 2.027726600994356, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 1.6}) +[2026-06-23 13:42:56,760][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.060426484793424606, 'train/grad_norm_pre_clip': 26.769210815429688, 'timing/data': 0.0013909009867347777, 'timing/model': 2.02636341098696, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 1.6}) +[2026-06-23 13:43:00,842][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.20702975988388062, 'train/grad_norm_pre_clip': 33.25796127319336, 'timing/data': 0.06156822800403461, 'timing/model': 2.034333780989982, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 1.6}) +[2026-06-23 13:43:04,942][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.028472937643527985, 'train/grad_norm_pre_clip': 45.27942657470703, 'timing/data': 0.0013457669992931187, 'timing/model': 2.027790547988843, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 1.6}) +[2026-06-23 13:43:08,952][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.2847175598144531, 'train/grad_norm_pre_clip': 63.32208251953125, 'timing/data': 0.0013919829798396677, 'timing/model': 2.026237496989779, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 1.6}) +[2026-06-23 13:43:12,971][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.07004792988300323, 'train/grad_norm_pre_clip': 29.6754150390625, 'timing/data': 0.0012420760176610202, 'timing/model': 2.027595735999057, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 1.6}) +[2026-06-23 13:43:17,051][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.019542666152119637, 'train/grad_norm_pre_clip': 23.1801700592041, 'timing/data': 0.05954424801166169, 'timing/model': 2.029034525010502, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 1.6}) +[2026-06-23 13:43:21,176][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.16195179522037506, 'train/grad_norm_pre_clip': 30.751750946044922, 'timing/data': 0.001337052002782002, 'timing/model': 2.0472464229969773, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 1.6}) +[2026-06-23 13:43:25,441][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.026228221133351326, 'train/grad_norm_pre_clip': 16.289024353027344, 'timing/data': 0.001308759005041793, 'timing/model': 2.0300420759886038, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 1.6}) +[2026-06-23 13:43:29,459][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.07321368902921677, 'train/grad_norm_pre_clip': 31.261558532714844, 'timing/data': 0.0012420450220815837, 'timing/model': 2.0326992659829557, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 1.6}) +[2026-06-23 13:43:33,557][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.03249363601207733, 'train/grad_norm_pre_clip': 11.202486991882324, 'timing/data': 0.06997107798815705, 'timing/model': 2.0423141450155526, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 1.6}) +[2026-06-23 13:43:37,679][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.026082497090101242, 'train/grad_norm_pre_clip': 13.97801399230957, 'timing/data': 0.0013528599811252207, 'timing/model': 2.037509494984988, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 1.6}) +[2026-06-23 13:43:41,725][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.06738746166229248, 'train/grad_norm_pre_clip': 22.7584285736084, 'timing/data': 0.0014586460019927472, 'timing/model': 2.049349295994034, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 1.6}) +[2026-06-23 13:43:45,749][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.039416659623384476, 'train/grad_norm_pre_clip': 19.922975540161133, 'timing/data': 0.0017948269960470498, 'timing/model': 2.0310407079814468, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 1.6}) +[2026-06-23 13:43:49,826][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.07333038747310638, 'train/grad_norm_pre_clip': 22.18052101135254, 'timing/data': 0.05722175497794524, 'timing/model': 2.033874385000672, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 1.6}) +[2026-06-23 13:43:53,942][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.029400350525975227, 'train/grad_norm_pre_clip': 15.31789779663086, 'timing/data': 0.0013428519887384027, 'timing/model': 2.0348085509904195, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 1.6}) +[2026-06-23 13:43:58,162][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.05066655948758125, 'train/grad_norm_pre_clip': 9.428227424621582, 'timing/data': 0.001348021993180737, 'timing/model': 2.230708886985667, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 1.6}) +[2026-06-23 13:44:02,199][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.0299017783254385, 'train/grad_norm_pre_clip': 12.806666374206543, 'timing/data': 0.001150786003563553, 'timing/model': 2.0485753329994623, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 1.6}) +[2026-06-23 13:44:06,289][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.03412538394331932, 'train/grad_norm_pre_clip': 8.912900924682617, 'timing/data': 0.05804887300473638, 'timing/model': 2.037383408023743, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 1.6}) +[2026-06-23 13:44:10,414][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.033621616661548615, 'train/grad_norm_pre_clip': 22.476863861083984, 'timing/data': 0.0013592120085377246, 'timing/model': 2.03544431499904, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 1.6}) +[2026-06-23 13:44:14,442][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.0857129693031311, 'train/grad_norm_pre_clip': 18.763893127441406, 'timing/data': 0.0012869490019511431, 'timing/model': 2.0337284929992165, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 1.6}) +[2026-06-23 13:44:18,467][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.1562526524066925, 'train/grad_norm_pre_clip': 31.48868751525879, 'timing/data': 0.00111255599767901, 'timing/model': 2.0339950950001366, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 1.6}) +[2026-06-23 13:44:22,560][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.015374628826975822, 'train/grad_norm_pre_clip': 8.419044494628906, 'timing/data': 0.06897262300481088, 'timing/model': 2.0298077650077175, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 1.6}) +[2026-06-23 13:44:26,678][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.22813287377357483, 'train/grad_norm_pre_clip': 34.40934371948242, 'timing/data': 0.0012389000039547682, 'timing/model': 2.037822888989467, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 1.6}) +[2026-06-23 13:44:30,933][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.04244498163461685, 'train/grad_norm_pre_clip': 15.73184871673584, 'timing/data': 0.001425413996912539, 'timing/model': 2.2618861030205153, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 1.6}) +[2026-06-23 13:44:34,980][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.07489512860774994, 'train/grad_norm_pre_clip': 32.03904724121094, 'timing/data': 0.0013857510057277977, 'timing/model': 2.0528969659935683, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 1.6}) +[2026-06-23 13:44:39,070][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.025522148236632347, 'train/grad_norm_pre_clip': 20.864757537841797, 'timing/data': 0.05930068698944524, 'timing/model': 2.034250044001965, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 1.61}) +[2026-06-23 13:44:43,196][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.1746804565191269, 'train/grad_norm_pre_clip': 27.1954288482666, 'timing/data': 0.0012313560000620782, 'timing/model': 2.036707165010739, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 1.61}) +[2026-06-23 13:44:47,223][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.12696735560894012, 'train/grad_norm_pre_clip': 23.19664192199707, 'timing/data': 0.0012157369928900152, 'timing/model': 2.036485304008238, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 1.61}) +[2026-06-23 13:44:51,265][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.02348429337143898, 'train/grad_norm_pre_clip': 14.140555381774902, 'timing/data': 0.001370042999042198, 'timing/model': 2.0434689600078855, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 1.61}) +[2026-06-23 13:44:55,362][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.02536402828991413, 'train/grad_norm_pre_clip': 9.392383575439453, 'timing/data': 0.060125712014269084, 'timing/model': 2.0419039069965947, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 1.61}) +[2026-06-23 13:44:59,479][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.07960743457078934, 'train/grad_norm_pre_clip': 38.56170654296875, 'timing/data': 0.0013769050128757954, 'timing/model': 2.038554015016416, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 1.61}) +[2026-06-23 13:45:03,556][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.03629001975059509, 'train/grad_norm_pre_clip': 14.912393569946289, 'timing/data': 0.00124220599536784, 'timing/model': 2.063078058999963, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 1.61}) +[2026-06-23 13:45:07,595][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.12319430708885193, 'train/grad_norm_pre_clip': 20.39445686340332, 'timing/data': 0.0015356980147771537, 'timing/model': 2.0403387429832947, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 1.61}) +[2026-06-23 13:45:11,703][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.07561901211738586, 'train/grad_norm_pre_clip': 22.883390426635742, 'timing/data': 0.05801671399967745, 'timing/model': 2.0566345200059004, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 1.61}) +[2026-06-23 13:45:15,822][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.027061086148023605, 'train/grad_norm_pre_clip': 15.948735237121582, 'timing/data': 0.0012392000062391162, 'timing/model': 2.0369251529918984, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 1.61}) +[2026-06-23 13:45:19,869][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.03870678320527077, 'train/grad_norm_pre_clip': 18.660423278808594, 'timing/data': 0.0012267880083527416, 'timing/model': 2.052148799004499, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 1.61}) +[2026-06-23 13:45:23,933][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.05725586786866188, 'train/grad_norm_pre_clip': 23.552270889282227, 'timing/data': 0.0013602339895442128, 'timing/model': 2.0616458910226356, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 1.61}) +[2026-06-23 13:45:28,031][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.06648153811693192, 'train/grad_norm_pre_clip': 16.851579666137695, 'timing/data': 0.06566300601116382, 'timing/model': 2.036796074011363, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 1.61}) +[2026-06-23 13:45:32,170][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.04035170376300812, 'train/grad_norm_pre_clip': 26.10739517211914, 'timing/data': 0.0015270519943442196, 'timing/model': 2.0517898440011777, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 1.61}) +[2026-06-23 13:45:36,216][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.09344346076250076, 'train/grad_norm_pre_clip': 18.934621810913086, 'timing/data': 0.0012370559852570295, 'timing/model': 2.0503031060216017, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 1.61}) +[2026-06-23 13:45:40,349][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.037812601774930954, 'train/grad_norm_pre_clip': 11.309553146362305, 'timing/data': 0.001306404999922961, 'timing/model': 2.04681322601391, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 1.61}) +[2026-06-23 13:45:44,454][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.062060125172138214, 'train/grad_norm_pre_clip': 24.063861846923828, 'timing/data': 0.05938399099977687, 'timing/model': 2.0509613110043574, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 1.61}) +[2026-06-23 13:45:48,573][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.06954053044319153, 'train/grad_norm_pre_clip': 18.3436222076416, 'timing/data': 0.0011931550106965005, 'timing/model': 2.0364934020035435, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 1.61}) +[2026-06-23 13:45:52,641][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.16556236147880554, 'train/grad_norm_pre_clip': 30.286663055419922, 'timing/data': 0.0014483069826383144, 'timing/model': 2.069038380024722, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 1.61}) +[2026-06-23 13:45:56,694][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.04064300283789635, 'train/grad_norm_pre_clip': 26.61188507080078, 'timing/data': 0.0014221379824448377, 'timing/model': 2.0426934600109234, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 1.61}) +[2026-06-23 13:46:00,783][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.039678193628787994, 'train/grad_norm_pre_clip': 27.112857818603516, 'timing/data': 0.05891080599394627, 'timing/model': 2.0369995860091876, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 1.61}) +[2026-06-23 13:46:05,153][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.06619265675544739, 'train/grad_norm_pre_clip': 27.593721389770508, 'timing/data': 0.0013549840077757835, 'timing/model': 2.279660320986295, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 1.61}) +[2026-06-23 13:46:09,183][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.062455784529447556, 'train/grad_norm_pre_clip': 33.18973922729492, 'timing/data': 0.0012250739964656532, 'timing/model': 2.03698178200284, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 1.61}) +[2026-06-23 13:46:13,458][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.08246775716543198, 'train/grad_norm_pre_clip': 17.203989028930664, 'timing/data': 0.001252444984856993, 'timing/model': 2.2848239130107686, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 1.61}) +[2026-06-23 13:46:17,577][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.1102205142378807, 'train/grad_norm_pre_clip': 34.53104782104492, 'timing/data': 0.0888766180141829, 'timing/model': 2.038741344003938, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 1.61}) +[2026-06-23 13:46:21,694][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.04166609048843384, 'train/grad_norm_pre_clip': 27.602508544921875, 'timing/data': 0.0014694260025862604, 'timing/model': 2.035559972980991, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 1.62}) +[2026-06-23 13:46:25,720][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.13642075657844543, 'train/grad_norm_pre_clip': 17.496427536010742, 'timing/data': 0.0012182520003989339, 'timing/model': 2.0337761970004067, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 1.62}) +[2026-06-23 13:46:29,751][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.03810352459549904, 'train/grad_norm_pre_clip': 30.854398727416992, 'timing/data': 0.0015216419997159392, 'timing/model': 2.036842252011411, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 1.62}) +[2026-06-23 13:46:33,844][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.15255703032016754, 'train/grad_norm_pre_clip': 24.45125961303711, 'timing/data': 0.06612824599142186, 'timing/model': 2.0329184400034137, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 1.62}) +[2026-06-23 13:46:37,958][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.12846629321575165, 'train/grad_norm_pre_clip': 31.28630828857422, 'timing/data': 0.00135833999956958, 'timing/model': 2.034062755003106, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 1.62}) +[2026-06-23 13:46:41,986][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.10151733458042145, 'train/grad_norm_pre_clip': 28.22979736328125, 'timing/data': 0.0012211569992359728, 'timing/model': 2.0354373849986587, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 1.62}) +[2026-06-23 13:46:46,020][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.1837204098701477, 'train/grad_norm_pre_clip': 38.181732177734375, 'timing/data': 0.001260149001609534, 'timing/model': 2.0381937380007003, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 1.62}) +[2026-06-23 13:46:50,146][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.10054892301559448, 'train/grad_norm_pre_clip': 16.28567123413086, 'timing/data': 0.06034263200126588, 'timing/model': 2.0408357119886205, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 1.62}) +[2026-06-23 13:46:54,275][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.05866871029138565, 'train/grad_norm_pre_clip': 15.34434986114502, 'timing/data': 0.0013634600036311895, 'timing/model': 2.0352488450007513, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 1.62}) +[2026-06-23 13:46:58,329][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.059319090098142624, 'train/grad_norm_pre_clip': 49.188262939453125, 'timing/data': 0.0014228799846023321, 'timing/model': 2.0561154700117186, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 1.62}) +[2026-06-23 13:47:02,377][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.01580986939370632, 'train/grad_norm_pre_clip': 5.112617015838623, 'timing/data': 0.00117924899677746, 'timing/model': 2.051231262012152, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 1.62}) +[2026-06-23 13:47:06,484][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.07576030492782593, 'train/grad_norm_pre_clip': 21.25649070739746, 'timing/data': 0.07575992599595338, 'timing/model': 2.0375906430126633, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 1.62}) +[2026-06-23 13:47:10,600][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.10894294828176498, 'train/grad_norm_pre_clip': 19.864013671875, 'timing/data': 0.0013732379884459078, 'timing/model': 2.0377092809940223, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 1.62}) +[2026-06-23 13:47:14,654][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.022404814139008522, 'train/grad_norm_pre_clip': 16.296554565429688, 'timing/data': 0.0011575590178836137, 'timing/model': 2.0541415210173, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 1.62}) +[2026-06-23 13:47:18,688][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.049203235656023026, 'train/grad_norm_pre_clip': 14.335585594177246, 'timing/data': 0.0015432320069521666, 'timing/model': 2.041172981989803, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 1.62}) +[2026-06-23 13:47:23,010][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.015577493235468864, 'train/grad_norm_pre_clip': 7.011558532714844, 'timing/data': 0.05826879999949597, 'timing/model': 2.277438696997706, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 1.62}) +[2026-06-23 13:47:27,114][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.031979791820049286, 'train/grad_norm_pre_clip': 12.341425895690918, 'timing/data': 0.0012533060216810554, 'timing/model': 2.029860484006349, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 1.62}) +[2026-06-23 13:47:31,139][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.04248182475566864, 'train/grad_norm_pre_clip': 11.823209762573242, 'timing/data': 0.0011562769941519946, 'timing/model': 2.0384959029906895, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 1.62}) +[2026-06-23 13:47:35,169][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.050410568714141846, 'train/grad_norm_pre_clip': 23.183544158935547, 'timing/data': 0.0013877739838790148, 'timing/model': 2.034592619020259, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 1.62}) +[2026-06-23 13:47:39,251][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.04179558530449867, 'train/grad_norm_pre_clip': 10.068838119506836, 'timing/data': 0.05880478001199663, 'timing/model': 2.0320714190020226, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 1.62}) +[2026-06-23 13:47:43,386][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.017600147053599358, 'train/grad_norm_pre_clip': 7.086813449859619, 'timing/data': 0.0012521739990916103, 'timing/model': 2.0393260060227476, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 1.62}) +[2026-06-23 13:47:47,432][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.12439549714326859, 'train/grad_norm_pre_clip': 35.942596435546875, 'timing/data': 0.001380832021823153, 'timing/model': 2.050113142991904, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 1.62}) +[2026-06-23 13:47:51,477][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.13985952734947205, 'train/grad_norm_pre_clip': 16.992403030395508, 'timing/data': 0.0011924429854843765, 'timing/model': 2.0529080069973134, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 1.62}) +[2026-06-23 13:47:55,748][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.07187414169311523, 'train/grad_norm_pre_clip': 17.387800216674805, 'timing/data': 0.0582734790223185, 'timing/model': 2.2184363679843955, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 1.62}) +[2026-06-23 13:47:59,876][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.06741641461849213, 'train/grad_norm_pre_clip': 23.642580032348633, 'timing/data': 0.0014572240179404616, 'timing/model': 2.038302934000967, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 1.62}) +[2026-06-23 13:48:03,928][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.08407163619995117, 'train/grad_norm_pre_clip': 14.804577827453613, 'timing/data': 0.0012446400069165975, 'timing/model': 2.056919780006865, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 1.62}) +[2026-06-23 13:48:07,977][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.0056250328198075294, 'train/grad_norm_pre_clip': 4.383930683135986, 'timing/data': 0.0013540229992941022, 'timing/model': 2.05219541900442, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 1.63}) +[2026-06-23 13:48:12,074][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4200 +[2026-06-23 13:50:07,336][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4200 +[2026-06-23 13:50:07,336][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4200 +[2026-06-23 13:52:14,544][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4200 +[2026-06-23 13:52:14,545][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.11573471128940582, 'train/grad_norm_pre_clip': 24.138572692871094, 'eval/action_loss/seconds': 115.25834401298198, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03907468169927597, 'eval/latency_2/loss': 0.03907468169927597, 'eval/flappy/loss': 0.03907468169927597, 'eval/flappy/latency_2/loss': 0.03907468169927597, 'eval/flap_precision': 0.958, 'eval/flap_recall': 0.9696356275303644, 'eval/flap_f1': 0.9637826961770624, 'eval/flap_rate_model': 0.20833333333333334, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9637826961770625, 'eval/latency_2/flap_precision': 0.958, 'eval/latency_2/flap_recall': 0.9696356275303644, 'eval/latency_2/flap_f1': 0.9637826961770624, 'eval/latency_2/flap_rate_model': 0.20833333333333334, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9637826961770625, 'eval/action_classification/data_wait_seconds': 50.37306594848633, 'eval/action_classification/predict_seconds': 75.23931121826172, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.057846038980642334, 'timing/model': 2.0463739900151268, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 1.63}) +[2026-06-23 13:52:18,593][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.1129886656999588, 'train/grad_norm_pre_clip': 32.765342712402344, 'timing/data': 0.0012066999915987253, 'timing/model': 1.9984780160011724, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 1.63}) +[2026-06-23 13:52:22,655][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.028307368978857994, 'train/grad_norm_pre_clip': 10.468052864074707, 'timing/data': 0.0012827409955207258, 'timing/model': 2.0138109829858877, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 1.63}) +[2026-06-23 13:52:26,621][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.1321854442358017, 'train/grad_norm_pre_clip': 30.247623443603516, 'timing/data': 0.0013396650028880686, 'timing/model': 2.0038121129909996, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 1.63}) +[2026-06-23 13:52:30,658][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.05931362882256508, 'train/grad_norm_pre_clip': 17.608190536499023, 'timing/data': 0.07337483600713313, 'timing/model': 2.004308010014938, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 1.63}) +[2026-06-23 13:52:34,772][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.04124469310045242, 'train/grad_norm_pre_clip': 11.773979187011719, 'timing/data': 0.0013377220020629466, 'timing/model': 2.057779902999755, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 1.63}) +[2026-06-23 13:52:38,749][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.011908983811736107, 'train/grad_norm_pre_clip': 7.759162425994873, 'timing/data': 0.001393504993757233, 'timing/model': 2.0099442359933164, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 1.63}) +[2026-06-23 13:52:42,750][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.015525364316999912, 'train/grad_norm_pre_clip': 8.668111801147461, 'timing/data': 0.0014156360120978206, 'timing/model': 2.034238746011397, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 1.63}) +[2026-06-23 13:52:46,800][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.08887733519077301, 'train/grad_norm_pre_clip': 28.059907913208008, 'timing/data': 0.06149530201219022, 'timing/model': 2.010938202001853, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 1.63}) +[2026-06-23 13:52:50,871][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.04429032281041145, 'train/grad_norm_pre_clip': 22.334291458129883, 'timing/data': 0.0013795989798381925, 'timing/model': 2.0113449140044395, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 1.63}) +[2026-06-23 13:52:54,859][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.02971440739929676, 'train/grad_norm_pre_clip': 9.168437004089355, 'timing/data': 0.001250200002687052, 'timing/model': 2.0182891379809007, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 1.63}) +[2026-06-23 13:52:59,092][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.1031050905585289, 'train/grad_norm_pre_clip': 33.3190803527832, 'timing/data': 0.0012616519816219807, 'timing/model': 2.018360259011388, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 1.63}) +[2026-06-23 13:53:03,150][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.05452824756503105, 'train/grad_norm_pre_clip': 29.91090965270996, 'timing/data': 0.06961377599509433, 'timing/model': 2.01674591601477, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 1.63}) +[2026-06-23 13:53:07,237][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.0834098607301712, 'train/grad_norm_pre_clip': 24.64583396911621, 'timing/data': 0.0012116289872210473, 'timing/model': 2.020073585008504, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 1.63}) +[2026-06-23 13:53:11,238][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.028370913118124008, 'train/grad_norm_pre_clip': 17.049448013305664, 'timing/data': 0.001405086019076407, 'timing/model': 2.02092131099198, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 1.63}) +[2026-06-23 13:53:15,242][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.07902799546718597, 'train/grad_norm_pre_clip': 19.8373966217041, 'timing/data': 0.0012261760130058974, 'timing/model': 2.0241318049957044, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 1.63}) +[2026-06-23 13:53:19,307][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.06213901564478874, 'train/grad_norm_pre_clip': 27.757707595825195, 'timing/data': 0.05761472001904622, 'timing/model': 2.026269606023561, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 1.63}) +[2026-06-23 13:53:23,409][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.08844465762376785, 'train/grad_norm_pre_clip': 29.275835037231445, 'timing/data': 0.0012132020201534033, 'timing/model': 2.0305511990154628, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 1.63}) +[2026-06-23 13:53:27,421][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.11798764020204544, 'train/grad_norm_pre_clip': 30.06727409362793, 'timing/data': 0.00125660200137645, 'timing/model': 2.024194511002861, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 1.63}) +[2026-06-23 13:53:31,520][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.08295267820358276, 'train/grad_norm_pre_clip': 16.19219970703125, 'timing/data': 0.0013549840077757835, 'timing/model': 2.1170443599985447, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 1.63}) +[2026-06-23 13:53:35,580][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.176508828997612, 'train/grad_norm_pre_clip': 27.292728424072266, 'timing/data': 0.061427717009792104, 'timing/model': 2.0215625460259616, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 1.63}) +[2026-06-23 13:53:39,670][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.06922600418329239, 'train/grad_norm_pre_clip': 17.546680450439453, 'timing/data': 0.001229281013365835, 'timing/model': 2.0242318689997774, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 1.63}) +[2026-06-23 13:53:43,681][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.06769350171089172, 'train/grad_norm_pre_clip': 12.87824535369873, 'timing/data': 0.0012787230079993606, 'timing/model': 2.0302497499797028, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 1.63}) +[2026-06-23 13:53:47,700][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.04726661369204521, 'train/grad_norm_pre_clip': 19.422637939453125, 'timing/data': 0.001302185992244631, 'timing/model': 2.031624711002223, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 1.63}) +[2026-06-23 13:53:51,783][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.055820416659116745, 'train/grad_norm_pre_clip': 16.05183982849121, 'timing/data': 0.07083395100198686, 'timing/model': 2.0274534220225178, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 1.63}) +[2026-06-23 13:53:55,883][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.031275659799575806, 'train/grad_norm_pre_clip': 14.167469024658203, 'timing/data': 0.001610116014489904, 'timing/model': 2.029159986996092, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 1.64}) +[2026-06-23 13:53:59,896][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.01929647848010063, 'train/grad_norm_pre_clip': 8.430752754211426, 'timing/data': 0.0012461330043151975, 'timing/model': 2.027473248017486, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 1.64}) +[2026-06-23 13:54:03,930][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.08246532082557678, 'train/grad_norm_pre_clip': 28.00555992126465, 'timing/data': 0.0013459470064844936, 'timing/model': 2.041195329016773, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 1.64}) +[2026-06-23 13:54:08,199][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.012558664195239544, 'train/grad_norm_pre_clip': 5.362255096435547, 'timing/data': 0.06533266400219873, 'timing/model': 2.0281207040243316, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 1.64}) +[2026-06-23 13:54:12,299][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.007858981378376484, 'train/grad_norm_pre_clip': 2.97784423828125, 'timing/data': 0.0014082829875405878, 'timing/model': 2.028067486011423, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 1.64}) +[2026-06-23 13:54:16,313][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.1131862998008728, 'train/grad_norm_pre_clip': 27.61200523376465, 'timing/data': 0.0014367050025612116, 'timing/model': 2.025812968990067, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 1.64}) +[2026-06-23 13:54:20,329][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.11190574616193771, 'train/grad_norm_pre_clip': 34.048667907714844, 'timing/data': 0.0011191979865543544, 'timing/model': 2.032167664001463, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 1.64}) +[2026-06-23 13:54:24,406][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.04207868129014969, 'train/grad_norm_pre_clip': 17.441383361816406, 'timing/data': 0.05838807899272069, 'timing/model': 2.029241905984236, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 1.64}) +[2026-06-23 13:54:28,508][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.08604372292757034, 'train/grad_norm_pre_clip': 30.13361930847168, 'timing/data': 0.0015864519809838384, 'timing/model': 2.0260445659805555, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 1.64}) +[2026-06-23 13:54:32,532][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.11463278532028198, 'train/grad_norm_pre_clip': 24.692331314086914, 'timing/data': 0.001167188020190224, 'timing/model': 2.028875228977995, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 1.64}) +[2026-06-23 13:54:36,541][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.01320621743798256, 'train/grad_norm_pre_clip': 11.776372909545898, 'timing/data': 0.0013086590042803437, 'timing/model': 2.0267773609957658, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 1.64}) +[2026-06-23 13:54:40,644][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.08538247644901276, 'train/grad_norm_pre_clip': 31.234745025634766, 'timing/data': 0.06512491102330387, 'timing/model': 2.0289211529889144, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 1.64}) +[2026-06-23 13:54:44,743][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.07625936716794968, 'train/grad_norm_pre_clip': 12.319400787353516, 'timing/data': 0.0015480510191991925, 'timing/model': 2.0265369260159787, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 1.64}) +[2026-06-23 13:54:48,799][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.20528505742549896, 'train/grad_norm_pre_clip': 20.31022834777832, 'timing/data': 0.0016263459983747452, 'timing/model': 2.043957721994957, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 1.64}) +[2026-06-23 13:54:52,818][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.058893147855997086, 'train/grad_norm_pre_clip': 20.74641227722168, 'timing/data': 0.0012345110008027405, 'timing/model': 2.031089971016627, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 1.64}) +[2026-06-23 13:54:56,904][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.04395671933889389, 'train/grad_norm_pre_clip': 11.667315483093262, 'timing/data': 0.06847327901050448, 'timing/model': 2.030304438987514, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 1.64}) +[2026-06-23 13:55:01,007][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.05742976814508438, 'train/grad_norm_pre_clip': 21.735090255737305, 'timing/data': 0.0012195039889775217, 'timing/model': 2.028846084023826, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 1.64}) +[2026-06-23 13:55:05,025][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.027413900941610336, 'train/grad_norm_pre_clip': 18.34170150756836, 'timing/data': 0.0011441239912528545, 'timing/model': 2.031144081003731, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 1.64}) +[2026-06-23 13:55:09,045][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.026260776445269585, 'train/grad_norm_pre_clip': 11.80207633972168, 'timing/data': 0.0011320119956508279, 'timing/model': 2.031716878991574, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 1.64}) +[2026-06-23 13:55:13,380][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.030687900260090828, 'train/grad_norm_pre_clip': 20.630449295043945, 'timing/data': 0.06575870100641623, 'timing/model': 2.2810347840131726, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 1.64}) +[2026-06-23 13:55:17,493][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.02870429866015911, 'train/grad_norm_pre_clip': 13.879659652709961, 'timing/data': 0.0012267169950064272, 'timing/model': 2.0354056250071153, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 1.64}) +[2026-06-23 13:55:21,516][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.06650535762310028, 'train/grad_norm_pre_clip': 15.035558700561523, 'timing/data': 0.0011572490038815886, 'timing/model': 2.027690488001099, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 1.64}) +[2026-06-23 13:55:25,534][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.026693422347307205, 'train/grad_norm_pre_clip': 10.677705764770508, 'timing/data': 0.0011235860001761466, 'timing/model': 2.0294473330141045, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 1.64}) +[2026-06-23 13:55:29,605][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.09605731070041656, 'train/grad_norm_pre_clip': 18.941041946411133, 'timing/data': 0.05475349200423807, 'timing/model': 2.029549583006883, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 1.64}) +[2026-06-23 13:55:33,708][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.05332432687282562, 'train/grad_norm_pre_clip': 36.18513107299805, 'timing/data': 0.0011080169933848083, 'timing/model': 2.0293432010221295, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 1.64}) +[2026-06-23 14:13:00,190][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4250 +[2026-06-23 14:13:00,191][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.08763834834098816, 'train/grad_norm_pre_clip': 18.74553680419922, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 268.7800000000133, 'rl_games_eval/mid_train/mean_length': 2185.0, 'rl_games_eval/mid_train/std_reward': 191.31879573111533, 'rl_games_eval/mid_train/std_length': 1541.771837854097, 'rl_games_eval/mid_train/macro_mean_reward': 268.7800000000133, 'rl_games_eval/mid_train/macro_mean_length': 2185.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 268.7800000000133, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2185.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 191.31879573111533, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1541.771837854097, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 268.7800000000133, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0012396509991958737, 'timing/model': 2.0307110589928925, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 1.64}) +[2026-06-23 14:13:04,061][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.03802504763007164, 'train/grad_norm_pre_clip': 32.67832565307617, 'timing/data': 0.001270207023480907, 'timing/model': 1.9572959279757924, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 1.65}) +[2026-06-23 14:13:08,009][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.08004795014858246, 'train/grad_norm_pre_clip': 38.86851119995117, 'timing/data': 0.06979143200442195, 'timing/model': 1.9580269590078387, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 1.65}) +[2026-06-23 14:13:11,982][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.06370817124843597, 'train/grad_norm_pre_clip': 30.71222686767578, 'timing/data': 0.001225324987899512, 'timing/model': 1.9631258230074309, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 1.65}) +[2026-06-23 14:13:15,876][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.10860896855592728, 'train/grad_norm_pre_clip': 24.61026382446289, 'timing/data': 0.001257152995094657, 'timing/model': 1.9694957250030711, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 1.65}) +[2026-06-23 14:13:19,789][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.09976822882890701, 'train/grad_norm_pre_clip': 23.64102554321289, 'timing/data': 0.001214634015923366, 'timing/model': 1.9861943760188296, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 1.65}) +[2026-06-23 14:13:23,760][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.0508170910179615, 'train/grad_norm_pre_clip': 21.121793746948242, 'timing/data': 0.07060386400553398, 'timing/model': 1.971756352984812, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 1.65}) +[2026-06-23 14:13:27,784][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.005869513843208551, 'train/grad_norm_pre_clip': 2.7879762649536133, 'timing/data': 0.0012237619957886636, 'timing/model': 2.0022573930036742, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 1.65}) +[2026-06-23 14:13:31,805][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.010030210018157959, 'train/grad_norm_pre_clip': 6.625891208648682, 'timing/data': 0.0016414739948231727, 'timing/model': 1.9814041140198242, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 1.65}) +[2026-06-23 14:13:35,746][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.0595611073076725, 'train/grad_norm_pre_clip': 19.523242950439453, 'timing/data': 0.0014424860128201544, 'timing/model': 2.0023429410066456, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 1.65}) +[2026-06-23 14:13:39,737][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.06826309114694595, 'train/grad_norm_pre_clip': 18.8383731842041, 'timing/data': 0.06727912000496872, 'timing/model': 1.980174952012021, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 1.65}) +[2026-06-23 14:13:43,754][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.02724328637123108, 'train/grad_norm_pre_clip': 6.815170764923096, 'timing/data': 0.0015780359972268343, 'timing/model': 1.9864919469982851, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 1.65}) +[2026-06-23 14:13:47,682][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.02666235901415348, 'train/grad_norm_pre_clip': 11.713174819946289, 'timing/data': 0.0012666499824263155, 'timing/model': 1.985629742994206, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 1.65}) +[2026-06-23 14:13:51,635][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.0739959180355072, 'train/grad_norm_pre_clip': 17.98862648010254, 'timing/data': 0.0012941419845446944, 'timing/model': 2.0058518260193523, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 1.65}) +[2026-06-23 14:13:55,653][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.029517723247408867, 'train/grad_norm_pre_clip': 12.699963569641113, 'timing/data': 0.07690065097995102, 'timing/model': 1.99464329102193, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 1.65}) +[2026-06-23 14:13:59,685][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.13772305846214294, 'train/grad_norm_pre_clip': 54.1748046875, 'timing/data': 0.0012370760086923838, 'timing/model': 1.9957547749800142, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 1.65}) +[2026-06-23 14:14:03,647][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.06883709132671356, 'train/grad_norm_pre_clip': 39.7816047668457, 'timing/data': 0.0013499250053428113, 'timing/model': 2.01032215799205, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 1.65}) +[2026-06-23 14:14:07,849][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.08071140199899673, 'train/grad_norm_pre_clip': 44.9295768737793, 'timing/data': 0.0012436179968062788, 'timing/model': 1.9957821360148955, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 1.65}) +[2026-06-23 14:14:11,863][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.07512427866458893, 'train/grad_norm_pre_clip': 21.652841567993164, 'timing/data': 0.06167792799533345, 'timing/model': 1.9992016259930097, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 1.65}) +[2026-06-23 14:14:15,919][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.034725312143564224, 'train/grad_norm_pre_clip': 15.747332572937012, 'timing/data': 0.0012514530098997056, 'timing/model': 2.0110616449965164, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 1.65}) +[2026-06-23 14:14:19,874][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.09660510718822479, 'train/grad_norm_pre_clip': 33.03983688354492, 'timing/data': 0.0012454819807317108, 'timing/model': 1.9982787710032426, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 1.65}) +[2026-06-23 14:14:23,830][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.03921546787023544, 'train/grad_norm_pre_clip': 12.705934524536133, 'timing/data': 0.0012380180123727769, 'timing/model': 1.9997660100052599, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 1.65}) +[2026-06-23 14:14:27,848][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.07789165526628494, 'train/grad_norm_pre_clip': 28.716726303100586, 'timing/data': 0.058563681988744065, 'timing/model': 2.000710454012733, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 1.65}) +[2026-06-23 14:14:31,898][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.08026494085788727, 'train/grad_norm_pre_clip': 26.72707748413086, 'timing/data': 0.0012635049934033304, 'timing/model': 2.002142562007066, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 1.65}) +[2026-06-23 14:14:35,862][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.07537396252155304, 'train/grad_norm_pre_clip': 23.37786102294922, 'timing/data': 0.0013095100002828985, 'timing/model': 2.002014645986492, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 1.65}) +[2026-06-23 14:14:39,930][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.22290827333927155, 'train/grad_norm_pre_clip': 40.7924690246582, 'timing/data': 0.0013365010090637952, 'timing/model': 2.105052471975796, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 1.65}) +[2026-06-23 14:14:43,958][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.16710813343524933, 'train/grad_norm_pre_clip': 30.81605339050293, 'timing/data': 0.05971502000465989, 'timing/model': 2.0059875379956793, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 1.65}) +[2026-06-23 14:14:48,033][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.11546067893505096, 'train/grad_norm_pre_clip': 28.32146644592285, 'timing/data': 0.0015177449968177825, 'timing/model': 2.019666006002808, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 1.66}) +[2026-06-23 14:14:52,011][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.037076592445373535, 'train/grad_norm_pre_clip': 12.016271591186523, 'timing/data': 0.0012684440007433295, 'timing/model': 2.0095915980054997, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 1.66}) +[2026-06-23 14:14:55,988][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.10408508032560349, 'train/grad_norm_pre_clip': 24.773773193359375, 'timing/data': 0.0013637109950650483, 'timing/model': 2.010579626017716, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 1.66}) +[2026-06-23 14:15:00,026][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.045529939234256744, 'train/grad_norm_pre_clip': 21.134788513183594, 'timing/data': 0.06123551999917254, 'timing/model': 2.010589463985525, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 1.66}) +[2026-06-23 14:15:04,131][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.029542893171310425, 'train/grad_norm_pre_clip': 20.08625030517578, 'timing/data': 0.0012069910007994622, 'timing/model': 2.0414452190161683, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 1.66}) +[2026-06-23 14:15:08,140][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.011712062172591686, 'train/grad_norm_pre_clip': 7.112842082977295, 'timing/data': 0.0014652580139227211, 'timing/model': 2.0207272300031036, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 1.66}) +[2026-06-23 14:15:12,168][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.04446116089820862, 'train/grad_norm_pre_clip': 21.837142944335938, 'timing/data': 0.0012758979864884168, 'timing/model': 2.050231817003805, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 1.66}) +[2026-06-23 14:15:16,436][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.10935160517692566, 'train/grad_norm_pre_clip': 29.618865966796875, 'timing/data': 0.06402050799806602, 'timing/model': 2.0196942199836485, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 1.66}) +[2026-06-23 14:15:20,524][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.012248044833540916, 'train/grad_norm_pre_clip': 4.304467678070068, 'timing/data': 0.0012683940003626049, 'timing/model': 2.0208469400240574, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 1.66}) +[2026-06-23 14:15:24,525][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.110635906457901, 'train/grad_norm_pre_clip': 14.923566818237305, 'timing/data': 0.001211159018566832, 'timing/model': 2.0203025650116615, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 1.66}) +[2026-06-23 14:15:28,545][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.05253341794013977, 'train/grad_norm_pre_clip': 15.904509544372559, 'timing/data': 0.0011774860031437129, 'timing/model': 2.0258758249983657, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 1.66}) +[2026-06-23 14:15:32,619][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.035590074956417084, 'train/grad_norm_pre_clip': 11.230020523071289, 'timing/data': 0.06769249300123192, 'timing/model': 2.019467451988021, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 1.66}) +[2026-06-23 14:15:36,702][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.08520027995109558, 'train/grad_norm_pre_clip': 28.043729782104492, 'timing/data': 0.0011362999794073403, 'timing/model': 2.0233533039863687, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 1.66}) +[2026-06-23 14:15:40,701][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.06690513342618942, 'train/grad_norm_pre_clip': 19.539871215820312, 'timing/data': 0.0012963560002390295, 'timing/model': 2.0210620489961, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 1.66}) +[2026-06-23 14:15:44,713][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.07069568336009979, 'train/grad_norm_pre_clip': 15.944934844970703, 'timing/data': 0.0012890530051663518, 'timing/model': 2.023112418013625, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 1.66}) +[2026-06-23 14:15:48,826][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.0410236157476902, 'train/grad_norm_pre_clip': 14.449922561645508, 'timing/data': 0.05749778199242428, 'timing/model': 2.0228668650088366, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 1.66}) +[2026-06-23 14:15:52,913][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.11315242946147919, 'train/grad_norm_pre_clip': 20.685550689697266, 'timing/data': 0.0012194440059829503, 'timing/model': 2.021367723005824, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 1.66}) +[2026-06-23 14:15:56,928][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.03643518686294556, 'train/grad_norm_pre_clip': 14.377954483032227, 'timing/data': 0.0012807569873984903, 'timing/model': 2.024523906002287, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 1.66}) +[2026-06-23 14:16:00,932][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.005810974165797234, 'train/grad_norm_pre_clip': 5.868561267852783, 'timing/data': 0.0012259760114829987, 'timing/model': 2.022222683997825, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 1.66}) +[2026-06-23 14:16:04,995][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.059018272906541824, 'train/grad_norm_pre_clip': 18.204992294311523, 'timing/data': 0.05965786401065998, 'timing/model': 2.0206932479923125, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 1.66}) +[2026-06-23 14:16:09,086][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.06610889732837677, 'train/grad_norm_pre_clip': 21.637056350708008, 'timing/data': 0.0012227999977767467, 'timing/model': 2.027019148983527, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 1.66}) +[2026-06-23 14:16:13,100][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.046129997819662094, 'train/grad_norm_pre_clip': 14.156747817993164, 'timing/data': 0.0014305840013548732, 'timing/model': 2.0296153470117133, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 1.66}) +[2026-06-23 14:16:17,111][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.06896848976612091, 'train/grad_norm_pre_clip': 18.493684768676758, 'timing/data': 0.0012211870052851737, 'timing/model': 2.0301310700015165, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 1.66}) +[2026-06-23 14:16:21,415][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4300 +[2026-06-23 14:18:16,122][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4300 +[2026-06-23 14:18:16,123][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4300 +[2026-06-23 14:20:23,665][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4300 +[2026-06-23 14:20:23,666][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.04409907013177872, 'train/grad_norm_pre_clip': 29.39427947998047, 'eval/action_loss/seconds': 114.70384062902303, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03922010958194733, 'eval/latency_2/loss': 0.03922010958194733, 'eval/flappy/loss': 0.03922010958194733, 'eval/flappy/latency_2/loss': 0.03922010958194733, 'eval/flap_precision': 0.9566955363091273, 'eval/flap_recall': 0.9689608636977058, 'eval/flap_f1': 0.9627891384512237, 'eval/flap_rate_model': 0.20847222222222223, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9627891384512236, 'eval/latency_2/flap_precision': 0.9566955363091273, 'eval/latency_2/flap_recall': 0.9689608636977058, 'eval/latency_2/flap_f1': 0.9627891384512237, 'eval/latency_2/flap_rate_model': 0.20847222222222223, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9627891384512236, 'eval/action_classification/data_wait_seconds': 50.941959381103516, 'eval/action_classification/predict_seconds': 74.90888977050781, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.05780238501029089, 'timing/model': 2.26798633497674, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 1.66}) +[2026-06-23 14:20:27,696][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.04911138489842415, 'train/grad_norm_pre_clip': 15.789814949035645, 'timing/data': 0.0013072669971734285, 'timing/model': 1.986717049003346, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 1.66}) +[2026-06-23 14:20:31,643][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.0479254275560379, 'train/grad_norm_pre_clip': 25.754955291748047, 'timing/data': 0.001251071982551366, 'timing/model': 1.9927126200054772, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 1.66}) +[2026-06-23 14:20:35,588][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.023875538259744644, 'train/grad_norm_pre_clip': 14.338848114013672, 'timing/data': 0.0011980540002696216, 'timing/model': 1.9935117370041553, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 1.67}) +[2026-06-23 14:20:39,596][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.08016620576381683, 'train/grad_norm_pre_clip': 14.121499061584473, 'timing/data': 0.05684449497493915, 'timing/model': 1.9999424320121761, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 1.67}) +[2026-06-23 14:20:43,889][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.0044304803013801575, 'train/grad_norm_pre_clip': 1.5034189224243164, 'timing/data': 0.0013417899899650365, 'timing/model': 2.2437770569813438, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 1.67}) +[2026-06-23 14:20:47,891][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.05196505784988403, 'train/grad_norm_pre_clip': 20.21949577331543, 'timing/data': 0.0014029130106791854, 'timing/model': 2.0158619529975113, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 1.67}) +[2026-06-23 14:20:51,862][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.14311571419239044, 'train/grad_norm_pre_clip': 27.129220962524414, 'timing/data': 0.0012665510002989322, 'timing/model': 2.007844075997127, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 1.67}) +[2026-06-23 14:20:55,903][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.02023565210402012, 'train/grad_norm_pre_clip': 11.65899658203125, 'timing/data': 0.06499185197753832, 'timing/model': 2.0061447439948097, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 1.67}) +[2026-06-23 14:20:59,968][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.08129335939884186, 'train/grad_norm_pre_clip': 25.32163429260254, 'timing/data': 0.0012882610026281327, 'timing/model': 2.01392438201583, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 1.67}) +[2026-06-23 14:21:03,943][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.06359334290027618, 'train/grad_norm_pre_clip': 20.333065032958984, 'timing/data': 0.0012353730271570385, 'timing/model': 2.0094186549831647, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 1.67}) +[2026-06-23 14:21:07,937][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.03248037025332451, 'train/grad_norm_pre_clip': 19.306270599365234, 'timing/data': 0.0014432769967243075, 'timing/model': 2.016759240999818, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 1.67}) +[2026-06-23 14:21:11,989][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.08340562880039215, 'train/grad_norm_pre_clip': 33.79907989501953, 'timing/data': 0.0616793509980198, 'timing/model': 2.016825944010634, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 1.67}) +[2026-06-23 14:21:16,325][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.11167687922716141, 'train/grad_norm_pre_clip': 18.525123596191406, 'timing/data': 0.001311492989771068, 'timing/model': 2.021496756991837, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 1.67}) +[2026-06-23 14:21:20,327][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.0396660678088665, 'train/grad_norm_pre_clip': 24.19879722595215, 'timing/data': 0.0012350420001894236, 'timing/model': 2.013367723993724, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 1.67}) +[2026-06-23 14:21:24,314][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.01963687688112259, 'train/grad_norm_pre_clip': 6.449905872344971, 'timing/data': 0.001208433008287102, 'timing/model': 2.0131168510124553, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 1.67}) +[2026-06-23 14:21:28,374][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.09045151621103287, 'train/grad_norm_pre_clip': 29.07114028930664, 'timing/data': 0.07497560398769565, 'timing/model': 2.015364613995189, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 1.67}) +[2026-06-23 14:21:32,448][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.03009195812046528, 'train/grad_norm_pre_clip': 14.681419372558594, 'timing/data': 0.0011583610030356795, 'timing/model': 2.0167592520010658, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 1.67}) +[2026-06-23 14:21:36,447][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.05333546921610832, 'train/grad_norm_pre_clip': 16.367158889770508, 'timing/data': 0.0012932300160173327, 'timing/model': 2.0180486949975602, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 1.67}) +[2026-06-23 14:21:40,441][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.04557665437459946, 'train/grad_norm_pre_clip': 14.58244800567627, 'timing/data': 0.0013483510119840503, 'timing/model': 2.020254801987903, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 1.67}) +[2026-06-23 14:21:44,519][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.04124954342842102, 'train/grad_norm_pre_clip': 13.023094177246094, 'timing/data': 0.06911346002016217, 'timing/model': 2.0343546209915075, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 1.67}) +[2026-06-23 14:21:48,867][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.018279509618878365, 'train/grad_norm_pre_clip': 6.7836995124816895, 'timing/data': 0.0012184719962533563, 'timing/model': 2.2773179459909443, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 1.67}) +[2026-06-23 14:21:52,875][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.02639145590364933, 'train/grad_norm_pre_clip': 8.517692565917969, 'timing/data': 0.0012296830245759338, 'timing/model': 2.0226293209998403, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 1.67}) +[2026-06-23 14:21:56,879][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.07047368586063385, 'train/grad_norm_pre_clip': 27.778087615966797, 'timing/data': 0.001192925003124401, 'timing/model': 2.024198671017075, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 1.67}) +[2026-06-23 14:22:00,939][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.13118202984333038, 'train/grad_norm_pre_clip': 23.254501342773438, 'timing/data': 0.05636288502137177, 'timing/model': 2.0219481919775717, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 1.67}) +[2026-06-23 14:22:05,060][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.05047681927680969, 'train/grad_norm_pre_clip': 22.345260620117188, 'timing/data': 0.0013775959960184991, 'timing/model': 2.042849631019635, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 1.67}) +[2026-06-23 14:22:09,076][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.01981787197291851, 'train/grad_norm_pre_clip': 10.215267181396484, 'timing/data': 0.001264255988644436, 'timing/model': 2.0294082689797506, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 1.67}) +[2026-06-23 14:22:13,106][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.01859368197619915, 'train/grad_norm_pre_clip': 7.241501331329346, 'timing/data': 0.0013811120006721467, 'timing/model': 2.0350359409931116, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 1.67}) +[2026-06-23 14:22:17,181][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.06118987500667572, 'train/grad_norm_pre_clip': 18.050039291381836, 'timing/data': 0.06545368599472567, 'timing/model': 2.033567044971278, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 1.68}) +[2026-06-23 14:22:21,264][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.0845424234867096, 'train/grad_norm_pre_clip': 26.706985473632812, 'timing/data': 0.0012556900037452579, 'timing/model': 2.0208757299988065, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 1.68}) +[2026-06-23 14:22:25,379][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.04698684811592102, 'train/grad_norm_pre_clip': 24.170185089111328, 'timing/data': 0.0015053219976834953, 'timing/model': 2.024273229006212, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 1.68}) +[2026-06-23 14:22:29,392][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.13751764595508575, 'train/grad_norm_pre_clip': 26.2301025390625, 'timing/data': 0.00122693701996468, 'timing/model': 2.0308218710124493, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 1.68}) +[2026-06-23 14:22:33,547][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.05977287143468857, 'train/grad_norm_pre_clip': 21.38919448852539, 'timing/data': 0.06383500501397066, 'timing/model': 2.0219814339943696, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 1.68}) +[2026-06-23 14:22:37,737][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.10795387625694275, 'train/grad_norm_pre_clip': 30.870323181152344, 'timing/data': 0.001107888005208224, 'timing/model': 2.0370455940137617, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 1.68}) +[2026-06-23 14:22:41,756][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.08806885033845901, 'train/grad_norm_pre_clip': 22.9108943939209, 'timing/data': 0.0014216380077414215, 'timing/model': 2.037331011990318, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 1.68}) +[2026-06-23 14:22:45,765][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.045879993587732315, 'train/grad_norm_pre_clip': 21.090662002563477, 'timing/data': 0.0012777209922205657, 'timing/model': 2.0271505459968466, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 1.68}) +[2026-06-23 14:22:49,844][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.014951937831938267, 'train/grad_norm_pre_clip': 11.067106246948242, 'timing/data': 0.06722160399658605, 'timing/model': 2.028590758010978, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 1.68}) +[2026-06-23 14:22:53,949][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.05811073258519173, 'train/grad_norm_pre_clip': 26.023950576782227, 'timing/data': 0.001179921004222706, 'timing/model': 2.0403098570241127, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 1.68}) +[2026-06-23 14:22:57,983][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.19430193305015564, 'train/grad_norm_pre_clip': 38.653141021728516, 'timing/data': 0.0012772000045515597, 'timing/model': 2.0524687280121725, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 1.68}) +[2026-06-23 14:23:01,987][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.07435766607522964, 'train/grad_norm_pre_clip': 25.335702896118164, 'timing/data': 0.0012274390028323978, 'timing/model': 2.02530868301983, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 1.68}) +[2026-06-23 14:23:06,056][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.0988510474562645, 'train/grad_norm_pre_clip': 28.657825469970703, 'timing/data': 0.06869336398085579, 'timing/model': 2.0209462310012896, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 1.68}) +[2026-06-23 14:23:10,161][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.045785605907440186, 'train/grad_norm_pre_clip': 17.959505081176758, 'timing/data': 0.0012339009845163673, 'timing/model': 2.0370702700165566, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 1.68}) +[2026-06-23 14:23:14,164][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.056100647896528244, 'train/grad_norm_pre_clip': 21.525217056274414, 'timing/data': 0.001278754003578797, 'timing/model': 2.02336970999022, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 1.68}) +[2026-06-23 14:23:18,166][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.09669450670480728, 'train/grad_norm_pre_clip': 24.86544418334961, 'timing/data': 0.0012077720020897686, 'timing/model': 2.0243991520255804, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 1.68}) +[2026-06-23 14:23:22,247][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.09883034229278564, 'train/grad_norm_pre_clip': 23.285640716552734, 'timing/data': 0.07483438300550915, 'timing/model': 2.025549869023962, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 1.68}) +[2026-06-23 14:23:26,336][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.07149157673120499, 'train/grad_norm_pre_clip': 39.83185577392578, 'timing/data': 0.0014282999909482896, 'timing/model': 2.021316003985703, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 1.68}) +[2026-06-23 14:23:30,341][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.22414624691009521, 'train/grad_norm_pre_clip': 41.59605407714844, 'timing/data': 0.0012475650000851601, 'timing/model': 2.0234796529985033, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 1.68}) +[2026-06-23 14:23:34,585][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.11207977682352066, 'train/grad_norm_pre_clip': 20.02731704711914, 'timing/data': 0.0012410540075507015, 'timing/model': 2.028304981009569, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 1.68}) +[2026-06-23 14:23:38,649][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.10654284060001373, 'train/grad_norm_pre_clip': 27.217676162719727, 'timing/data': 0.05655871800263412, 'timing/model': 2.0243176609801594, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 1.68}) +[2026-06-23 14:23:42,743][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.10326990485191345, 'train/grad_norm_pre_clip': 29.12798500061035, 'timing/data': 0.0012143949861638248, 'timing/model': 2.0263496469997335, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 1.68}) +[2026-06-23 14:23:46,746][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.007984147407114506, 'train/grad_norm_pre_clip': 4.886331558227539, 'timing/data': 0.001227338012540713, 'timing/model': 2.023778797010891, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 1.68}) +[2026-06-23 14:23:50,749][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.013629348017275333, 'train/grad_norm_pre_clip': 6.675355911254883, 'timing/data': 0.0012390190095175058, 'timing/model': 2.0227981530188117, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 1.68}) +[2026-06-23 14:23:54,822][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.04809143394231796, 'train/grad_norm_pre_clip': 17.170366287231445, 'timing/data': 0.0576930430252105, 'timing/model': 2.0323317229922395, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 1.68}) +[2026-06-23 14:23:58,938][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.03720155358314514, 'train/grad_norm_pre_clip': 18.79754638671875, 'timing/data': 0.001213362003909424, 'timing/model': 2.0266277310147416, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 1.68}) +[2026-06-23 14:24:02,945][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.009915555827319622, 'train/grad_norm_pre_clip': 6.519328594207764, 'timing/data': 0.0013977430062368512, 'timing/model': 2.0260879829875194, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 1.69}) +[2026-06-23 14:24:07,032][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.07074591517448425, 'train/grad_norm_pre_clip': 11.727179527282715, 'timing/data': 0.0011686200159601867, 'timing/model': 2.04345082299551, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 1.69}) +[2026-06-23 14:24:11,109][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.03403264656662941, 'train/grad_norm_pre_clip': 11.197860717773438, 'timing/data': 0.07125271399854682, 'timing/model': 2.023025464004604, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 1.69}) +[2026-06-23 14:24:15,203][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.07268170267343521, 'train/grad_norm_pre_clip': 33.628875732421875, 'timing/data': 0.0013336639967747033, 'timing/model': 2.0262116529920604, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 1.69}) +[2026-06-23 14:24:19,213][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.011822612024843693, 'train/grad_norm_pre_clip': 4.3473920822143555, 'timing/data': 0.0012098159932065755, 'timing/model': 2.027174101996934, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 1.69}) +[2026-06-23 14:24:23,224][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.07046821713447571, 'train/grad_norm_pre_clip': 24.289377212524414, 'timing/data': 0.0011835870100185275, 'timing/model': 2.026463538000826, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 1.69}) +[2026-06-23 14:24:27,295][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.031453389674425125, 'train/grad_norm_pre_clip': 12.550868034362793, 'timing/data': 0.057233323983382434, 'timing/model': 2.029701893014135, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 1.69}) +[2026-06-23 14:24:31,393][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.12024031579494476, 'train/grad_norm_pre_clip': 42.711029052734375, 'timing/data': 0.001196071010781452, 'timing/model': 2.025436580996029, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 1.69}) +[2026-06-23 14:24:35,405][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.04828650504350662, 'train/grad_norm_pre_clip': 16.150096893310547, 'timing/data': 0.001219905010657385, 'timing/model': 2.028000068006804, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 1.69}) +[2026-06-23 14:24:39,563][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.1777932494878769, 'train/grad_norm_pre_clip': 28.75255584716797, 'timing/data': 0.001199055986944586, 'timing/model': 2.1748747280216776, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 1.69}) +[2026-06-23 14:24:43,652][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.03726634755730629, 'train/grad_norm_pre_clip': 9.002326011657715, 'timing/data': 0.0742629469896201, 'timing/model': 2.02857211502851, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 1.69}) +[2026-06-23 14:24:47,748][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.014490766450762749, 'train/grad_norm_pre_clip': 5.6049113273620605, 'timing/data': 0.0012149650137871504, 'timing/model': 2.0265915149939246, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 1.69}) +[2026-06-23 14:24:51,775][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.03401191905140877, 'train/grad_norm_pre_clip': 22.11037826538086, 'timing/data': 0.001244930987013504, 'timing/model': 2.042636347003281, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 1.69}) +[2026-06-23 14:24:55,789][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.08181580901145935, 'train/grad_norm_pre_clip': 16.585691452026367, 'timing/data': 0.0011842780222650617, 'timing/model': 2.028996219014516, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 1.69}) +[2026-06-23 14:24:59,872][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.047245994210243225, 'train/grad_norm_pre_clip': 10.883458137512207, 'timing/data': 0.06831726900418289, 'timing/model': 2.0291967800003476, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 1.69}) +[2026-06-23 14:25:04,001][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.10247534513473511, 'train/grad_norm_pre_clip': 24.61909294128418, 'timing/data': 0.0013507259718608111, 'timing/model': 2.0521560909983236, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 1.69}) +[2026-06-23 14:25:08,009][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.03077898733317852, 'train/grad_norm_pre_clip': 13.228906631469727, 'timing/data': 0.0011576790129765868, 'timing/model': 2.025237551977625, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 1.69}) +[2026-06-23 14:25:12,053][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.019244328141212463, 'train/grad_norm_pre_clip': 7.010122776031494, 'timing/data': 0.001240001991391182, 'timing/model': 2.0569140950101428, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 1.69}) +[2026-06-23 14:25:16,151][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.07646890729665756, 'train/grad_norm_pre_clip': 30.71795082092285, 'timing/data': 0.06580771098379046, 'timing/model': 2.0253998120024335, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 1.69}) +[2026-06-23 14:25:20,248][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.023748919367790222, 'train/grad_norm_pre_clip': 8.90214729309082, 'timing/data': 0.0013922929938416928, 'timing/model': 2.026604379003402, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 1.69}) +[2026-06-23 14:25:24,258][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.009112238883972168, 'train/grad_norm_pre_clip': 8.116201400756836, 'timing/data': 0.0012348120217211545, 'timing/model': 2.0278584269981366, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 1.69}) +[2026-06-23 14:25:28,276][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.05133763328194618, 'train/grad_norm_pre_clip': 33.52493667602539, 'timing/data': 0.0015408170002046973, 'timing/model': 2.0325295089860447, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 1.69}) +[2026-06-23 14:25:32,370][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.03764688968658447, 'train/grad_norm_pre_clip': 17.684961318969727, 'timing/data': 0.07820753700798377, 'timing/model': 2.031092302990146, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 1.69}) +[2026-06-23 14:25:36,478][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.030079472810029984, 'train/grad_norm_pre_clip': 15.001677513122559, 'timing/data': 0.0014204950130078942, 'timing/model': 2.0329250500071794, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 1.69}) +[2026-06-23 14:25:40,493][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.03340684622526169, 'train/grad_norm_pre_clip': 18.164813995361328, 'timing/data': 0.0012594469881150872, 'timing/model': 2.032302740000887, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 1.69}) +[2026-06-23 14:25:44,505][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.08524472266435623, 'train/grad_norm_pre_clip': 32.28569793701172, 'timing/data': 0.001370382000459358, 'timing/model': 2.0268905380216893, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 1.69}) +[2026-06-23 14:25:48,694][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.14115563035011292, 'train/grad_norm_pre_clip': 35.507049560546875, 'timing/data': 0.07159794098697603, 'timing/model': 2.1309616939979605, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 1.7}) +[2026-06-23 14:25:52,794][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.05036066472530365, 'train/grad_norm_pre_clip': 33.4126091003418, 'timing/data': 0.0011869730078615248, 'timing/model': 2.0308687399956398, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 1.7}) +[2026-06-23 14:25:56,813][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.05140090733766556, 'train/grad_norm_pre_clip': 20.36274528503418, 'timing/data': 0.001242216007085517, 'timing/model': 2.032704130979255, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 1.7}) +[2026-06-23 14:26:00,828][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.08534552156925201, 'train/grad_norm_pre_clip': 17.87814712524414, 'timing/data': 0.0012230300053488463, 'timing/model': 2.0290744839876425, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 1.7}) +[2026-06-23 14:26:04,910][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.03986738622188568, 'train/grad_norm_pre_clip': 14.051146507263184, 'timing/data': 0.058378410001751035, 'timing/model': 2.0334589880076237, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 1.7}) +[2026-06-23 14:26:09,022][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.05723264068365097, 'train/grad_norm_pre_clip': 10.32944107055664, 'timing/data': 0.001337822002824396, 'timing/model': 2.029903547023423, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 1.7}) +[2026-06-23 14:26:13,032][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.09075438231229782, 'train/grad_norm_pre_clip': 25.7445125579834, 'timing/data': 0.0014011189923621714, 'timing/model': 2.025080693012569, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 1.7}) +[2026-06-23 14:26:17,064][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.11247134953737259, 'train/grad_norm_pre_clip': 29.572166442871094, 'timing/data': 0.0017515970102977008, 'timing/model': 2.0412609460181557, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 1.7}) +[2026-06-23 14:26:21,132][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.11978094279766083, 'train/grad_norm_pre_clip': 32.348167419433594, 'timing/data': 0.056364137009950355, 'timing/model': 2.027231018000748, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 1.7}) +[2026-06-23 14:26:25,483][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.11470203846693039, 'train/grad_norm_pre_clip': 20.728633880615234, 'timing/data': 0.0012162670027464628, 'timing/model': 2.028703638992738, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 1.7}) +[2026-06-23 14:26:29,487][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.05658271536231041, 'train/grad_norm_pre_clip': 21.768489837646484, 'timing/data': 0.001228790992172435, 'timing/model': 2.0232258449832443, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 1.7}) +[2026-06-23 14:26:33,507][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.08321530371904373, 'train/grad_norm_pre_clip': 33.302974700927734, 'timing/data': 0.001412790996255353, 'timing/model': 2.035842902987497, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 1.7}) +[2026-06-23 14:26:37,579][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.018884629011154175, 'train/grad_norm_pre_clip': 8.169901847839355, 'timing/data': 0.06361925698001869, 'timing/model': 2.0270611029991414, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 1.7}) +[2026-06-23 14:26:41,689][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.05922595411539078, 'train/grad_norm_pre_clip': 20.213016510009766, 'timing/data': 0.0013424410135485232, 'timing/model': 2.0412780079932418, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 1.7}) +[2026-06-23 14:26:45,697][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.08284058421850204, 'train/grad_norm_pre_clip': 26.742908477783203, 'timing/data': 0.0012282899988349527, 'timing/model': 2.0261540170176886, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 1.7}) +[2026-06-23 14:26:49,709][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.03648034483194351, 'train/grad_norm_pre_clip': 20.168237686157227, 'timing/data': 0.0011871239985339344, 'timing/model': 2.024240111000836, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 1.7}) +[2026-06-23 14:26:53,790][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.07590450346469879, 'train/grad_norm_pre_clip': 18.689931869506836, 'timing/data': 0.06719049601815641, 'timing/model': 2.0295179140230175, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 1.7}) +[2026-06-23 14:26:57,933][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.15935459733009338, 'train/grad_norm_pre_clip': 22.79778480529785, 'timing/data': 0.0011608749919105321, 'timing/model': 2.022755845973734, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 1.7}) +[2026-06-23 14:27:01,944][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.04875167831778526, 'train/grad_norm_pre_clip': 21.006977081298828, 'timing/data': 0.001278863986954093, 'timing/model': 2.0271118270175066, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 1.7}) +[2026-06-23 14:27:05,961][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.046740684658288956, 'train/grad_norm_pre_clip': 15.194928169250488, 'timing/data': 0.0013097399787511677, 'timing/model': 2.0318898069963325, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 1.7}) +[2026-06-23 14:27:10,043][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4400 +[2026-06-23 14:29:03,954][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4400 +[2026-06-23 14:29:03,955][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4400 +[2026-06-23 14:31:11,212][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4400 +[2026-06-23 14:31:11,213][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.05003666505217552, 'train/grad_norm_pre_clip': 17.432369232177734, 'eval/action_loss/seconds': 113.90800998301711, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03972411900758743, 'eval/latency_2/loss': 0.03972411900758743, 'eval/flappy/loss': 0.03972411900758743, 'eval/flappy/latency_2/loss': 0.03972411900758743, 'eval/flap_precision': 0.9623149394347241, 'eval/flap_recall': 0.9649122807017544, 'eval/flap_f1': 0.9636118598382749, 'eval/flap_rate_model': 0.2063888888888889, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9636118598382749, 'eval/latency_2/flap_precision': 0.9623149394347241, 'eval/latency_2/flap_recall': 0.9649122807017544, 'eval/latency_2/flap_f1': 0.9636118598382749, 'eval/latency_2/flap_rate_model': 0.2063888888888889, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9636118598382749, 'eval/action_classification/data_wait_seconds': 50.520751953125, 'eval/action_classification/predict_seconds': 75.08695983886719, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06972556098480709, 'timing/model': 2.030598580982769, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 1.7}) +[2026-06-23 14:31:15,286][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.06110762804746628, 'train/grad_norm_pre_clip': 21.936872482299805, 'timing/data': 0.0012144950160291046, 'timing/model': 2.0175254389760084, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 1.7}) +[2026-06-23 14:31:19,227][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.17036154866218567, 'train/grad_norm_pre_clip': 38.23323440551758, 'timing/data': 0.00131350799347274, 'timing/model': 1.992394565982977, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 1.7}) +[2026-06-23 14:31:23,418][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.11325155198574066, 'train/grad_norm_pre_clip': 44.09053421020508, 'timing/data': 0.0011441149981692433, 'timing/model': 1.9967151260061655, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 1.7}) +[2026-06-23 14:31:27,449][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.19236181676387787, 'train/grad_norm_pre_clip': 25.804763793945312, 'timing/data': 0.07410777400946245, 'timing/model': 2.0048037320084404, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 1.7}) +[2026-06-23 14:31:31,491][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.02243737503886223, 'train/grad_norm_pre_clip': 8.676663398742676, 'timing/data': 0.0011296480079181492, 'timing/model': 1.9983510489983018, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 1.7}) +[2026-06-23 14:31:35,451][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.06503196060657501, 'train/grad_norm_pre_clip': 16.586816787719727, 'timing/data': 0.0012200159835629165, 'timing/model': 2.002917403995525, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 1.71}) +[2026-06-23 14:31:39,410][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.01637054793536663, 'train/grad_norm_pre_clip': 5.563248634338379, 'timing/data': 0.0011162329756189138, 'timing/model': 2.001532677997602, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 1.71}) +[2026-06-23 14:31:43,431][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.022391390055418015, 'train/grad_norm_pre_clip': 17.058805465698242, 'timing/data': 0.05677209701389074, 'timing/model': 2.0041645570017863, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 1.71}) +[2026-06-23 14:31:47,513][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.06031548231840134, 'train/grad_norm_pre_clip': 20.710111618041992, 'timing/data': 0.0014479160017799586, 'timing/model': 2.0225900009972975, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 1.71}) +[2026-06-23 14:31:51,492][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.0026507622096687555, 'train/grad_norm_pre_clip': 1.0595061779022217, 'timing/data': 0.0012528359948191792, 'timing/model': 2.010349378979299, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 1.71}) +[2026-06-23 14:31:55,613][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.11334602534770966, 'train/grad_norm_pre_clip': 27.511962890625, 'timing/data': 0.001214364980114624, 'timing/model': 2.0150943339976948, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 1.71}) +[2026-06-23 14:31:59,659][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.028535546734929085, 'train/grad_norm_pre_clip': 12.817240715026855, 'timing/data': 0.05967433098703623, 'timing/model': 2.015545402013231, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 1.71}) +[2026-06-23 14:32:03,734][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.11510267108678818, 'train/grad_norm_pre_clip': 32.586917877197266, 'timing/data': 0.0014596179826185107, 'timing/model': 2.0137224159843754, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 1.71}) +[2026-06-23 14:32:07,725][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.03224091976881027, 'train/grad_norm_pre_clip': 11.445737838745117, 'timing/data': 0.0012826000165659934, 'timing/model': 2.0154857199813705, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 1.71}) +[2026-06-23 14:32:11,725][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.029480628669261932, 'train/grad_norm_pre_clip': 8.431568145751953, 'timing/data': 0.0012873899831902236, 'timing/model': 2.0231303260079585, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 1.71}) +[2026-06-23 14:32:15,789][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.006133307237178087, 'train/grad_norm_pre_clip': 2.7192163467407227, 'timing/data': 0.06981693900888786, 'timing/model': 2.0199616720201448, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 1.71}) +[2026-06-23 14:32:19,879][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.132477805018425, 'train/grad_norm_pre_clip': 29.651762008666992, 'timing/data': 0.001240082026924938, 'timing/model': 2.0247385269904044, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 1.71}) +[2026-06-23 14:32:23,880][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.03808753192424774, 'train/grad_norm_pre_clip': 10.965449333190918, 'timing/data': 0.0012560219911392778, 'timing/model': 2.021245513984468, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 1.71}) +[2026-06-23 14:32:28,133][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.044053465127944946, 'train/grad_norm_pre_clip': 15.261261940002441, 'timing/data': 0.0012169689871370792, 'timing/model': 2.272619871015195, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 1.71}) +[2026-06-23 14:32:32,199][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.012827889062464237, 'train/grad_norm_pre_clip': 9.928794860839844, 'timing/data': 0.06202989397570491, 'timing/model': 2.0224112779833376, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 1.71}) +[2026-06-23 14:32:36,290][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.048074208199977875, 'train/grad_norm_pre_clip': 21.162105560302734, 'timing/data': 0.00121096899965778, 'timing/model': 2.0239116460143123, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 1.71}) +[2026-06-23 14:32:40,303][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.10018616914749146, 'train/grad_norm_pre_clip': 25.125795364379883, 'timing/data': 0.001255320996278897, 'timing/model': 2.0336094069934916, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 1.71}) +[2026-06-23 14:32:44,324][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.10326544940471649, 'train/grad_norm_pre_clip': 24.829734802246094, 'timing/data': 0.0012489389919210225, 'timing/model': 2.033788201981224, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 1.71}) +[2026-06-23 14:32:48,400][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.019164705649018288, 'train/grad_norm_pre_clip': 9.82005500793457, 'timing/data': 0.057208932004868984, 'timing/model': 2.0389240399817936, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 1.71}) +[2026-06-23 14:32:52,519][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.031801752746105194, 'train/grad_norm_pre_clip': 16.10387420654297, 'timing/data': 0.0014885410200804472, 'timing/model': 2.04145404099836, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 1.71}) +[2026-06-23 14:32:56,531][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.1666981428861618, 'train/grad_norm_pre_clip': 32.19437026977539, 'timing/data': 0.0012889530044049025, 'timing/model': 2.0277950070158113, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 1.71}) +[2026-06-23 14:33:00,780][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.05944401025772095, 'train/grad_norm_pre_clip': 40.864139556884766, 'timing/data': 0.0013765240146312863, 'timing/model': 2.2637049829936586, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 1.71}) +[2026-06-23 14:33:04,840][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.11810844391584396, 'train/grad_norm_pre_clip': 21.2073974609375, 'timing/data': 0.05332607397576794, 'timing/model': 2.0239574130100664, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 1.71}) +[2026-06-23 14:33:08,934][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.037064749747514725, 'train/grad_norm_pre_clip': 8.884058952331543, 'timing/data': 0.001454968994949013, 'timing/model': 2.025865605013678, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 1.71}) +[2026-06-23 14:33:12,957][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.12764473259449005, 'train/grad_norm_pre_clip': 46.84521484375, 'timing/data': 0.0012500410084612668, 'timing/model': 2.0424318330187816, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 1.71}) +[2026-06-23 14:33:16,973][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.0337599478662014, 'train/grad_norm_pre_clip': 10.407929420471191, 'timing/data': 0.001428671006578952, 'timing/model': 2.028620797995245, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 1.71}) +[2026-06-23 14:33:21,060][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.05430147796869278, 'train/grad_norm_pre_clip': 24.702260971069336, 'timing/data': 0.0748074010189157, 'timing/model': 2.025483642995823, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 1.72}) +[2026-06-23 14:33:25,156][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.02337646484375, 'train/grad_norm_pre_clip': 8.989981651306152, 'timing/data': 0.0013189179881010205, 'timing/model': 2.0246033309958875, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 1.72}) +[2026-06-23 14:33:29,168][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.006751410663127899, 'train/grad_norm_pre_clip': 2.3048269748687744, 'timing/data': 0.0012708900030702353, 'timing/model': 2.027848896017531, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 1.72}) +[2026-06-23 14:33:33,219][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.15894483029842377, 'train/grad_norm_pre_clip': 39.50009536743164, 'timing/data': 0.001445923000574112, 'timing/model': 2.0679876169888303, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 1.72}) +[2026-06-23 14:33:37,292][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.03134062886238098, 'train/grad_norm_pre_clip': 12.388494491577148, 'timing/data': 0.05880020098993555, 'timing/model': 2.0310208589944523, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 1.72}) +[2026-06-23 14:33:41,402][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.032122451812028885, 'train/grad_norm_pre_clip': 13.665262222290039, 'timing/data': 0.001258906995644793, 'timing/model': 2.0414841149759013, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 1.72}) +[2026-06-23 14:33:45,407][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.019322162494063377, 'train/grad_norm_pre_clip': 17.271793365478516, 'timing/data': 0.001230555004440248, 'timing/model': 2.0247379560023546, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 1.72}) +[2026-06-23 14:33:49,440][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.13346770405769348, 'train/grad_norm_pre_clip': 29.339900970458984, 'timing/data': 0.0014104569854680449, 'timing/model': 2.047518379986286, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 1.72}) +[2026-06-23 14:33:53,523][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.04852798581123352, 'train/grad_norm_pre_clip': 31.872920989990234, 'timing/data': 0.07212441001320258, 'timing/model': 2.0274512819887605, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 1.72}) +[2026-06-23 14:33:57,621][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.11744324117898941, 'train/grad_norm_pre_clip': 32.61201095581055, 'timing/data': 0.001332411979092285, 'timing/model': 2.0285026199999265, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 1.72}) +[2026-06-23 14:34:01,640][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.0751219391822815, 'train/grad_norm_pre_clip': 27.29743194580078, 'timing/data': 0.0012458229903131723, 'timing/model': 2.031405582005391, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 1.72}) +[2026-06-23 14:34:05,661][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.03217025473713875, 'train/grad_norm_pre_clip': 21.642772674560547, 'timing/data': 0.001247736014192924, 'timing/model': 2.0325687369913794, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 1.72}) +[2026-06-23 14:34:09,850][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.03298631310462952, 'train/grad_norm_pre_clip': 11.262608528137207, 'timing/data': 0.06657755599007942, 'timing/model': 2.0308577999821864, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 1.72}) +[2026-06-23 14:34:13,951][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.04687253758311272, 'train/grad_norm_pre_clip': 21.680099487304688, 'timing/data': 0.001373328996123746, 'timing/model': 2.028833784017479, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 1.72}) +[2026-06-23 14:34:17,969][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.05120731517672539, 'train/grad_norm_pre_clip': 13.543143272399902, 'timing/data': 0.0012522950128186494, 'timing/model': 2.030725506017916, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 1.72}) +[2026-06-23 14:34:21,986][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.051745086908340454, 'train/grad_norm_pre_clip': 24.71962547302246, 'timing/data': 0.0012543390039354563, 'timing/model': 2.0289179609972052, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 1.72}) +[2026-06-23 14:34:26,067][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.0568147674202919, 'train/grad_norm_pre_clip': 19.111963272094727, 'timing/data': 0.0577065079996828, 'timing/model': 2.0308434249891434, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 1.72}) +[2026-06-23 14:34:30,178][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.10422100871801376, 'train/grad_norm_pre_clip': 19.060441970825195, 'timing/data': 0.0012288219877518713, 'timing/model': 2.03513554399251, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 1.72}) +[2026-06-23 14:34:34,202][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.004206226672977209, 'train/grad_norm_pre_clip': 2.6725807189941406, 'timing/data': 0.0012734630145132542, 'timing/model': 2.0344629370083567, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 1.72}) +[2026-06-23 14:34:38,230][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.04906521737575531, 'train/grad_norm_pre_clip': 24.495357513427734, 'timing/data': 0.0013756419939454645, 'timing/model': 2.0358139910094906, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 1.72}) +[2026-06-23 14:34:42,538][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.07887569069862366, 'train/grad_norm_pre_clip': 39.331214904785156, 'timing/data': 0.06996837301994674, 'timing/model': 2.2456964220036753, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 1.72}) +[2026-06-23 14:34:46,651][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.07200990617275238, 'train/grad_norm_pre_clip': 21.943408966064453, 'timing/data': 0.001316532987402752, 'timing/model': 2.0351980819832534, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 1.72}) +[2026-06-23 14:34:50,680][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.05366286262869835, 'train/grad_norm_pre_clip': 19.612253189086914, 'timing/data': 0.0012544579803943634, 'timing/model': 2.036093263013754, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 1.72}) +[2026-06-23 14:34:54,701][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.04392459616065025, 'train/grad_norm_pre_clip': 20.210651397705078, 'timing/data': 0.0012165989901404828, 'timing/model': 2.031157474993961, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 1.72}) +[2026-06-23 14:34:58,784][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.02341495268046856, 'train/grad_norm_pre_clip': 19.418304443359375, 'timing/data': 0.05907511699479073, 'timing/model': 2.0338522119855043, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 1.72}) +[2026-06-23 14:35:02,882][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.04258684813976288, 'train/grad_norm_pre_clip': 18.83794403076172, 'timing/data': 0.001418132014805451, 'timing/model': 2.0268176250101533, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 1.72}) +[2026-06-23 14:35:06,898][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.08085241168737411, 'train/grad_norm_pre_clip': 22.553203582763672, 'timing/data': 0.0012528760125860572, 'timing/model': 2.0299495900108013, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 1.73}) +[2026-06-23 14:35:10,922][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.012829442508518696, 'train/grad_norm_pre_clip': 7.682097911834717, 'timing/data': 0.0013630290050059557, 'timing/model': 2.031982884014724, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 1.73}) +[2026-06-23 14:35:15,023][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.0065360525622963905, 'train/grad_norm_pre_clip': 5.151169776916504, 'timing/data': 0.07567930800723843, 'timing/model': 2.033911264996277, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 1.73}) +[2026-06-23 14:35:19,363][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.054866332560777664, 'train/grad_norm_pre_clip': 25.63351058959961, 'timing/data': 0.0013191989855840802, 'timing/model': 2.0302368080010638, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 1.73}) +[2026-06-23 14:35:23,410][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.06311560422182083, 'train/grad_norm_pre_clip': 14.955413818359375, 'timing/data': 0.0013977439957670867, 'timing/model': 2.0595293499936815, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 1.73}) +[2026-06-23 14:35:27,435][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.10519874095916748, 'train/grad_norm_pre_clip': 28.000200271606445, 'timing/data': 0.0011091190099250525, 'timing/model': 2.027262553019682, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 1.73}) +[2026-06-23 14:35:31,514][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.03598722442984581, 'train/grad_norm_pre_clip': 15.060240745544434, 'timing/data': 0.06775167398154736, 'timing/model': 2.026470826007426, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 1.73}) +[2026-06-23 14:35:35,618][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.055377811193466187, 'train/grad_norm_pre_clip': 33.44449234008789, 'timing/data': 0.0012105670175515115, 'timing/model': 2.033763662009733, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 1.73}) +[2026-06-23 14:35:39,629][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.10839759558439255, 'train/grad_norm_pre_clip': 21.37224006652832, 'timing/data': 0.0011569880007300526, 'timing/model': 2.027089711977169, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 1.73}) +[2026-06-23 14:35:43,645][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.05164077877998352, 'train/grad_norm_pre_clip': 13.685290336608887, 'timing/data': 0.0011182670132257044, 'timing/model': 2.029911924008047, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 1.73}) +[2026-06-23 14:35:47,743][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.059858642518520355, 'train/grad_norm_pre_clip': 15.363369941711426, 'timing/data': 0.07527580100577325, 'timing/model': 2.037159145984333, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 1.73}) +[2026-06-23 14:35:51,995][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.1146201565861702, 'train/grad_norm_pre_clip': 23.91598892211914, 'timing/data': 0.0013719660055357963, 'timing/model': 2.0314352929999586, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 1.73}) +[2026-06-23 14:35:56,019][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.01723800227046013, 'train/grad_norm_pre_clip': 7.143643856048584, 'timing/data': 0.001275988994166255, 'timing/model': 2.035777675017016, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 1.73}) +[2026-06-23 14:36:00,064][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.01293210405856371, 'train/grad_norm_pre_clip': 8.305564880371094, 'timing/data': 0.0012577850138768554, 'timing/model': 2.0520184919878375, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 1.73}) +[2026-06-23 14:36:04,162][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.03182491287589073, 'train/grad_norm_pre_clip': 9.763919830322266, 'timing/data': 0.0590700049942825, 'timing/model': 2.040455064008711, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 1.73}) +[2026-06-23 14:36:08,287][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.01059702504426241, 'train/grad_norm_pre_clip': 7.059393405914307, 'timing/data': 0.0012171699781902134, 'timing/model': 2.045015390001936, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 1.73}) +[2026-06-23 14:36:12,318][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.10740596801042557, 'train/grad_norm_pre_clip': 38.14193344116211, 'timing/data': 0.00125178400776349, 'timing/model': 2.034016274003079, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 1.73}) +[2026-06-23 14:36:16,343][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.1273774653673172, 'train/grad_norm_pre_clip': 29.124778747558594, 'timing/data': 0.0012883020099252462, 'timing/model': 2.035375784005737, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 1.73}) +[2026-06-23 14:36:20,438][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.06576676666736603, 'train/grad_norm_pre_clip': 35.20083999633789, 'timing/data': 0.05970705300569534, 'timing/model': 2.033062955015339, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 1.73}) +[2026-06-23 14:36:24,800][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.0483696311712265, 'train/grad_norm_pre_clip': 22.496051788330078, 'timing/data': 0.0012768000015057623, 'timing/model': 2.0303154840075877, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 1.73}) +[2026-06-23 14:36:28,818][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.04299576207995415, 'train/grad_norm_pre_clip': 23.65558624267578, 'timing/data': 0.0011653139954432845, 'timing/model': 2.0305692400143016, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 1.73}) +[2026-06-23 14:36:32,838][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.059099387377500534, 'train/grad_norm_pre_clip': 28.75214958190918, 'timing/data': 0.0013509769923985004, 'timing/model': 2.0332652179931756, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 1.73}) +[2026-06-23 14:36:36,919][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.05367891117930412, 'train/grad_norm_pre_clip': 13.038723945617676, 'timing/data': 0.06185249797999859, 'timing/model': 2.03227110800799, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 1.73}) +[2026-06-23 14:36:41,069][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.10808101296424866, 'train/grad_norm_pre_clip': 24.886709213256836, 'timing/data': 0.0014828099810983986, 'timing/model': 2.066718325979309, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 1.73}) +[2026-06-23 14:36:45,092][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.053866077214479446, 'train/grad_norm_pre_clip': 16.789085388183594, 'timing/data': 0.0011960410047322512, 'timing/model': 2.0343617660109885, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 1.73}) +[2026-06-23 14:36:49,127][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.05950749292969704, 'train/grad_norm_pre_clip': 15.54067611694336, 'timing/data': 0.0013006330118514597, 'timing/model': 2.038197902002139, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 1.74}) +[2026-06-23 14:36:53,212][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.04138439521193504, 'train/grad_norm_pre_clip': 10.375883102416992, 'timing/data': 0.06383264798205346, 'timing/model': 2.0337185690004844, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 1.74}) +[2026-06-23 14:36:57,556][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.09667179733514786, 'train/grad_norm_pre_clip': 18.957143783569336, 'timing/data': 0.0012238519848324358, 'timing/model': 2.2668206290109083, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 1.74}) +[2026-06-23 14:37:01,588][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.02626117318868637, 'train/grad_norm_pre_clip': 15.178824424743652, 'timing/data': 0.0012344709830358624, 'timing/model': 2.04898018197855, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 1.74}) +[2026-06-23 14:37:05,604][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.026420339941978455, 'train/grad_norm_pre_clip': 9.001785278320312, 'timing/data': 0.0011742110073100775, 'timing/model': 2.0307054599979892, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 1.74}) +[2026-06-23 14:37:09,711][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.10793055593967438, 'train/grad_norm_pre_clip': 26.66037940979004, 'timing/data': 0.07734543399419636, 'timing/model': 2.0425353629980236, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 1.74}) +[2026-06-23 14:37:13,827][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.004178717732429504, 'train/grad_norm_pre_clip': 3.361344337463379, 'timing/data': 0.001245793013367802, 'timing/model': 2.0334430639923085, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 1.74}) +[2026-06-23 14:37:17,869][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.08320455998182297, 'train/grad_norm_pre_clip': 18.13191032409668, 'timing/data': 0.0013043909857515246, 'timing/model': 2.0471212429984007, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 1.74}) +[2026-06-23 14:37:21,889][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.07005399465560913, 'train/grad_norm_pre_clip': 24.469924926757812, 'timing/data': 0.0011280050093773752, 'timing/model': 2.0319524930091575, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 1.74}) +[2026-06-23 14:37:25,970][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.06144766882061958, 'train/grad_norm_pre_clip': 33.981327056884766, 'timing/data': 0.060278588003711775, 'timing/model': 2.0328361919964664, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 1.74}) +[2026-06-23 14:37:30,090][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.09365026652812958, 'train/grad_norm_pre_clip': 39.494815826416016, 'timing/data': 0.001371304999338463, 'timing/model': 2.0323467050038744, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 1.74}) +[2026-06-23 14:37:34,240][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.020557718351483345, 'train/grad_norm_pre_clip': 11.99374771118164, 'timing/data': 0.0011931659828405827, 'timing/model': 2.0551857670070603, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 1.74}) +[2026-06-23 14:37:38,286][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.0404374860227108, 'train/grad_norm_pre_clip': 13.41649055480957, 'timing/data': 0.0012261960073374212, 'timing/model': 2.0502494090178516, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 1.74}) +[2026-06-23 14:37:42,382][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.026678644120693207, 'train/grad_norm_pre_clip': 13.33786678314209, 'timing/data': 0.0676220299792476, 'timing/model': 2.033521166013088, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 1.74}) +[2026-06-23 14:37:46,498][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.048509661108255386, 'train/grad_norm_pre_clip': 28.851638793945312, 'timing/data': 0.0011950579937547445, 'timing/model': 2.036586628993973, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 1.74}) +[2026-06-23 14:37:50,517][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.04054665192961693, 'train/grad_norm_pre_clip': 16.999067306518555, 'timing/data': 0.0013758130080532283, 'timing/model': 2.032026578002842, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 1.74}) +[2026-06-23 14:37:54,539][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.06391020864248276, 'train/grad_norm_pre_clip': 43.134422302246094, 'timing/data': 0.0012734240153804421, 'timing/model': 2.0322327360045165, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 1.74}) +[2026-06-23 14:37:58,625][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4500 +[2026-06-23 14:39:52,622][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4500 +[2026-06-23 14:39:52,623][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4500 +[2026-06-23 14:41:59,782][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4500 +[2026-06-23 15:05:35,119][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4500 +[2026-06-23 15:05:35,121][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.05016162618994713, 'train/grad_norm_pre_clip': 23.226224899291992, 'eval/action_loss/seconds': 113.99354420599411, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04019917547702789, 'eval/latency_2/loss': 0.04019917547702789, 'eval/flappy/loss': 0.04019917547702789, 'eval/flappy/latency_2/loss': 0.04019917547702789, 'eval/flap_precision': 0.956, 'eval/flap_recall': 0.9676113360323887, 'eval/flap_f1': 0.9617706237424547, 'eval/flap_rate_model': 0.20833333333333334, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9617706237424547, 'eval/latency_2/flap_precision': 0.956, 'eval/latency_2/flap_recall': 0.9676113360323887, 'eval/latency_2/flap_f1': 0.9617706237424547, 'eval/latency_2/flap_rate_model': 0.20833333333333334, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9617706237424547, 'eval/action_classification/data_wait_seconds': 50.00883483886719, 'eval/action_classification/predict_seconds': 75.39276123046875, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 366.6600000000206, 'rl_games_eval/mid_train/mean_length': 2972.2, 'rl_games_eval/mid_train/std_reward': 155.88000000001009, 'rl_games_eval/mid_train/std_length': 1255.6000000000001, 'rl_games_eval/mid_train/macro_mean_reward': 366.6600000000206, 'rl_games_eval/mid_train/macro_mean_length': 2972.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 366.6600000000206, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2972.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 155.88000000001009, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1255.6000000000001, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 366.6600000000206, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.05951193001237698, 'timing/model': 2.032716096000513, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 1.74}) +[2026-06-23 15:05:35,121][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_4500_state +[2026-06-23 15:05:35,128][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 15:05:44,149][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_4500_state/model.safetensors +[2026-06-23 15:05:48,101][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_4500_state/optimizer.bin +[2026-06-23 15:05:48,102][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_4500_state/sampler.bin +[2026-06-23 15:05:48,102][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-23 15:05:48,103][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-23 15:05:48,239][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-23 15:05:48,240][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 15:05:48,264][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 15:05:52,235][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.007550328504294157, 'train/grad_norm_pre_clip': 4.659342288970947, 'timing/data': 0.0012282200041227043, 'timing/model': 1.953001423011301, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 1.74}) +[2026-06-23 15:05:56,107][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.03462629392743111, 'train/grad_norm_pre_clip': 16.05077362060547, 'timing/data': 0.0012332689948379993, 'timing/model': 1.9578350170049816, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 1.74}) +[2026-06-23 15:05:59,996][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.059553518891334534, 'train/grad_norm_pre_clip': 15.633535385131836, 'timing/data': 0.0012031740043312311, 'timing/model': 1.9670222660060972, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 1.74}) +[2026-06-23 15:06:03,932][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.014701984822750092, 'train/grad_norm_pre_clip': 7.811847686767578, 'timing/data': 0.053759025991894305, 'timing/model': 1.961759088007966, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 1.74}) +[2026-06-23 15:06:07,893][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.043377023190259933, 'train/grad_norm_pre_clip': 25.889753341674805, 'timing/data': 0.0011638709984254092, 'timing/model': 1.9593909419781994, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 1.74}) +[2026-06-23 15:06:11,793][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.023164166137576103, 'train/grad_norm_pre_clip': 40.29607391357422, 'timing/data': 0.0011601339792832732, 'timing/model': 1.9827273419941776, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 1.74}) +[2026-06-23 15:06:15,702][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.04283386841416359, 'train/grad_norm_pre_clip': 20.831655502319336, 'timing/data': 0.0015214520099107176, 'timing/model': 1.9750136350048706, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 1.74}) +[2026-06-23 15:06:19,666][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.04513544589281082, 'train/grad_norm_pre_clip': 31.891860961914062, 'timing/data': 0.062112075014738366, 'timing/model': 1.9751009659958072, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 1.74}) +[2026-06-23 15:06:23,914][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.010225808247923851, 'train/grad_norm_pre_clip': 7.118027687072754, 'timing/data': 0.0013465889787767082, 'timing/model': 1.9843027330061886, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 1.75}) +[2026-06-23 15:06:27,822][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.12367338687181473, 'train/grad_norm_pre_clip': 34.28281784057617, 'timing/data': 0.0011572280200198293, 'timing/model': 1.9758809469931293, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 1.75}) +[2026-06-23 15:06:31,735][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.1606893241405487, 'train/grad_norm_pre_clip': 26.97264862060547, 'timing/data': 0.001152940996689722, 'timing/model': 1.976667251001345, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 1.75}) +[2026-06-23 15:06:35,726][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.14693637192249298, 'train/grad_norm_pre_clip': 35.25185012817383, 'timing/data': 0.06908966100309044, 'timing/model': 1.9800277019967325, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 1.75}) +[2026-06-23 15:06:39,741][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.05078812688589096, 'train/grad_norm_pre_clip': 26.635860443115234, 'timing/data': 0.0012405729794409126, 'timing/model': 1.9914605410012882, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 1.75}) +[2026-06-23 15:06:43,682][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.1494842767715454, 'train/grad_norm_pre_clip': 33.431251525878906, 'timing/data': 0.0012336799991317093, 'timing/model': 1.995933578000404, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 1.75}) +[2026-06-23 15:06:47,632][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.03818859905004501, 'train/grad_norm_pre_clip': 30.888607025146484, 'timing/data': 0.0013233750069048256, 'timing/model': 2.0057672129769344, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 1.75}) +[2026-06-23 15:06:51,623][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.05202966928482056, 'train/grad_norm_pre_clip': 21.828413009643555, 'timing/data': 0.054640455986373127, 'timing/model': 1.9920996330038179, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 1.75}) +[2026-06-23 15:06:55,710][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.04333895817399025, 'train/grad_norm_pre_clip': 29.8718318939209, 'timing/data': 0.0013443850039038807, 'timing/model': 2.0056056539760903, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 1.75}) +[2026-06-23 15:06:59,653][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.04261183738708496, 'train/grad_norm_pre_clip': 15.91600513458252, 'timing/data': 0.0011807619885075837, 'timing/model': 1.9911130889959168, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 1.75}) +[2026-06-23 15:07:03,594][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.09545601904392242, 'train/grad_norm_pre_clip': 23.616722106933594, 'timing/data': 0.0011730380065273494, 'timing/model': 1.9925947980082128, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 1.75}) +[2026-06-23 15:07:07,601][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.021271463483572006, 'train/grad_norm_pre_clip': 18.051490783691406, 'timing/data': 0.061127494991524145, 'timing/model': 1.995500096003525, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 1.75}) +[2026-06-23 15:07:11,635][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.0175771564245224, 'train/grad_norm_pre_clip': 20.099990844726562, 'timing/data': 0.001153090997831896, 'timing/model': 1.9984969460056163, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 1.75}) +[2026-06-23 15:07:15,586][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.013453321531414986, 'train/grad_norm_pre_clip': 8.01081371307373, 'timing/data': 0.00117103400407359, 'timing/model': 1.9972627139941324, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 1.75}) +[2026-06-23 15:07:19,538][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.11988149583339691, 'train/grad_norm_pre_clip': 39.85538101196289, 'timing/data': 0.0012986600049771369, 'timing/model': 1.997056523017818, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 1.75}) +[2026-06-23 15:07:23,546][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.025519654154777527, 'train/grad_norm_pre_clip': 20.319923400878906, 'timing/data': 0.055306446010945365, 'timing/model': 1.9980487760039978, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 1.75}) +[2026-06-23 15:07:27,688][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.01952028088271618, 'train/grad_norm_pre_clip': 13.302727699279785, 'timing/data': 0.0012079419975634664, 'timing/model': 2.1032600559992716, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 1.75}) +[2026-06-23 15:07:31,650][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.0944766253232956, 'train/grad_norm_pre_clip': 24.602144241333008, 'timing/data': 0.0011526299931574613, 'timing/model': 2.001159434992587, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 1.75}) +[2026-06-23 15:07:35,613][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.06775986403226852, 'train/grad_norm_pre_clip': 18.432567596435547, 'timing/data': 0.0011101120035164058, 'timing/model': 2.00612165700295, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 1.75}) +[2026-06-23 15:07:39,658][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.1729918122291565, 'train/grad_norm_pre_clip': 46.7399787902832, 'timing/data': 0.06477947501116432, 'timing/model': 2.008896495011868, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 1.75}) +[2026-06-23 15:07:43,722][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.06573911011219025, 'train/grad_norm_pre_clip': 22.839479446411133, 'timing/data': 0.0012904850009363145, 'timing/model': 2.011019077996025, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 1.75}) +[2026-06-23 15:07:47,700][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.028822554275393486, 'train/grad_norm_pre_clip': 15.820371627807617, 'timing/data': 0.0012772699992638081, 'timing/model': 2.011267146008322, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 1.75}) +[2026-06-23 15:07:51,690][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.0740308091044426, 'train/grad_norm_pre_clip': 32.3559684753418, 'timing/data': 0.0011223239998798817, 'timing/model': 2.0276598729833495, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 1.75}) +[2026-06-23 15:07:55,734][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.15019439160823822, 'train/grad_norm_pre_clip': 36.448143005371094, 'timing/data': 0.06682929300586693, 'timing/model': 2.010874261002755, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 1.75}) +[2026-06-23 15:07:59,801][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.014890627935528755, 'train/grad_norm_pre_clip': 6.689227104187012, 'timing/data': 0.0012722010142169893, 'timing/model': 2.014964599016821, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 1.75}) +[2026-06-23 15:08:03,999][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.05575394257903099, 'train/grad_norm_pre_clip': 31.783489227294922, 'timing/data': 0.001513736991910264, 'timing/model': 2.0139041490037926, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 1.75}) +[2026-06-23 15:08:07,978][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.05939900502562523, 'train/grad_norm_pre_clip': 26.963212966918945, 'timing/data': 0.0012308350123930722, 'timing/model': 2.0119164760108106, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 1.76}) +[2026-06-23 15:08:12,012][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.07544101774692535, 'train/grad_norm_pre_clip': 30.79633903503418, 'timing/data': 0.0576742030098103, 'timing/model': 2.0096085400145967, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 1.76}) +[2026-06-23 15:08:16,101][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.08994120359420776, 'train/grad_norm_pre_clip': 23.19332504272461, 'timing/data': 0.0013994759938213974, 'timing/model': 2.0384117939975113, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 1.76}) +[2026-06-23 15:08:20,122][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.07072854787111282, 'train/grad_norm_pre_clip': 26.676313400268555, 'timing/data': 0.0012579140020534396, 'timing/model': 2.0490806920279283, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 1.76}) +[2026-06-23 15:08:24,128][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.04343416914343834, 'train/grad_norm_pre_clip': 17.04450035095215, 'timing/data': 0.0018953429826069623, 'timing/model': 2.0295944079989567, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 1.76}) +[2026-06-23 15:08:28,189][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.06741161644458771, 'train/grad_norm_pre_clip': 46.0056266784668, 'timing/data': 0.06743460200959817, 'timing/model': 2.0206683709984645, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 1.76}) +[2026-06-23 15:08:32,267][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.09369327127933502, 'train/grad_norm_pre_clip': 19.061996459960938, 'timing/data': 0.0012368859897833318, 'timing/model': 2.020255687995814, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 1.76}) +[2026-06-23 15:08:36,300][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.10212051868438721, 'train/grad_norm_pre_clip': 20.89288902282715, 'timing/data': 0.0011676779831759632, 'timing/model': 2.0578877689840738, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 1.76}) +[2026-06-23 15:08:40,307][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.05756242573261261, 'train/grad_norm_pre_clip': 29.206600189208984, 'timing/data': 0.0011363999801687896, 'timing/model': 2.0331471340032294, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 1.76}) +[2026-06-23 15:08:44,368][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.01885993219912052, 'train/grad_norm_pre_clip': 11.27762508392334, 'timing/data': 0.05339879999519326, 'timing/model': 2.034631877002539, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 1.76}) +[2026-06-23 15:08:48,449][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.020485755056142807, 'train/grad_norm_pre_clip': 15.664441108703613, 'timing/data': 0.0011726479860953987, 'timing/model': 2.025223228003597, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 1.76}) +[2026-06-23 15:08:52,440][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.1531957983970642, 'train/grad_norm_pre_clip': 66.36088562011719, 'timing/data': 0.0012325980060268193, 'timing/model': 2.018068614997901, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 1.76}) +[2026-06-23 15:08:56,438][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.030391113832592964, 'train/grad_norm_pre_clip': 13.251181602478027, 'timing/data': 0.001225314976181835, 'timing/model': 2.0224228530132677, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 1.76}) +[2026-06-23 15:09:00,531][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.025651073083281517, 'train/grad_norm_pre_clip': 8.766855239868164, 'timing/data': 0.07085934100905433, 'timing/model': 2.023275568993995, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 1.76}) +[2026-06-23 15:09:04,619][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.04423796758055687, 'train/grad_norm_pre_clip': 16.971813201904297, 'timing/data': 0.0013614670024253428, 'timing/model': 2.028657175018452, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 1.76}) +[2026-06-23 15:09:08,634][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.02459874004125595, 'train/grad_norm_pre_clip': 13.569029808044434, 'timing/data': 0.001280666998354718, 'timing/model': 2.0273743629804812, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 1.76}) +[2026-06-23 15:09:12,791][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.04845479130744934, 'train/grad_norm_pre_clip': 19.485633850097656, 'timing/data': 0.0014298830064944923, 'timing/model': 2.025758998002857, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 1.76}) +[2026-06-23 15:09:16,855][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.11248001456260681, 'train/grad_norm_pre_clip': 29.440481185913086, 'timing/data': 0.06414880900410935, 'timing/model': 2.021876902988879, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 1.76}) +[2026-06-23 15:09:20,947][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.03912150859832764, 'train/grad_norm_pre_clip': 14.129364013671875, 'timing/data': 0.0014116289967205375, 'timing/model': 2.030994003987871, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 1.76}) +[2026-06-23 15:09:24,961][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.1201867163181305, 'train/grad_norm_pre_clip': 69.86054229736328, 'timing/data': 0.0012326189898885787, 'timing/model': 2.0331090120016597, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 1.76}) +[2026-06-23 15:09:28,986][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.1700972467660904, 'train/grad_norm_pre_clip': 19.91811180114746, 'timing/data': 0.0013261809945106506, 'timing/model': 2.040524326992454, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 1.76}) +[2026-06-23 15:09:33,036][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.04295229911804199, 'train/grad_norm_pre_clip': 16.537761688232422, 'timing/data': 0.0585825809976086, 'timing/model': 2.0158442240208387, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 1.76}) +[2026-06-23 15:09:37,142][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.030643707141280174, 'train/grad_norm_pre_clip': 19.309213638305664, 'timing/data': 0.0014011790044605732, 'timing/model': 2.0442327499913517, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 1.76}) +[2026-06-23 15:09:41,145][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.06392039358615875, 'train/grad_norm_pre_clip': 37.566375732421875, 'timing/data': 0.001235192990861833, 'timing/model': 2.02075557000353, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 1.76}) +[2026-06-23 15:09:45,416][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.044203273952007294, 'train/grad_norm_pre_clip': 21.1440486907959, 'timing/data': 0.0014616619737353176, 'timing/model': 2.2919246509845834, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 1.76}) +[2026-06-23 15:09:49,471][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.03322621062397957, 'train/grad_norm_pre_clip': 17.114032745361328, 'timing/data': 0.05589774899999611, 'timing/model': 2.0186313739977777, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 1.76}) +[2026-06-23 15:09:53,578][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.025811485946178436, 'train/grad_norm_pre_clip': 13.566084861755371, 'timing/data': 0.0013946079998277128, 'timing/model': 2.0454412440012675, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 1.77}) +[2026-06-23 15:09:57,587][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.06972076743841171, 'train/grad_norm_pre_clip': 25.513486862182617, 'timing/data': 0.001249319000635296, 'timing/model': 2.0253482190018985, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 1.77}) +[2026-06-23 15:10:01,610][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.05263575538992882, 'train/grad_norm_pre_clip': 27.71088409423828, 'timing/data': 0.001233620016137138, 'timing/model': 2.0410634629952256, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 1.77}) +[2026-06-23 15:10:05,669][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.04275444895029068, 'train/grad_norm_pre_clip': 20.7535400390625, 'timing/data': 0.0622161669889465, 'timing/model': 2.0214652709837537, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 1.77}) +[2026-06-23 15:10:09,757][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.06294189393520355, 'train/grad_norm_pre_clip': 26.225587844848633, 'timing/data': 0.0013261109997984022, 'timing/model': 2.0265915759955533, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 1.77}) +[2026-06-23 15:10:13,769][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.08943358808755875, 'train/grad_norm_pre_clip': 27.3850154876709, 'timing/data': 0.001284173020394519, 'timing/model': 2.026433322986122, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 1.77}) +[2026-06-23 15:10:18,032][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.012638632208108902, 'train/grad_norm_pre_clip': 6.7117390632629395, 'timing/data': 0.0013964110112283379, 'timing/model': 2.2814635969989467, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 1.77}) +[2026-06-23 15:10:22,095][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.04760532081127167, 'train/grad_norm_pre_clip': 18.720447540283203, 'timing/data': 0.0563169940141961, 'timing/model': 2.0237017050094437, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 1.77}) +[2026-06-23 15:10:26,209][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.017829254269599915, 'train/grad_norm_pre_clip': 15.736985206604004, 'timing/data': 0.001409535005223006, 'timing/model': 2.044953461998375, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 1.77}) +[2026-06-23 15:10:30,235][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.029266510158777237, 'train/grad_norm_pre_clip': 10.969198226928711, 'timing/data': 0.0014770499838050455, 'timing/model': 2.042398760007927, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 1.77}) +[2026-06-23 15:10:34,254][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.07120200246572495, 'train/grad_norm_pre_clip': 26.974123001098633, 'timing/data': 0.0014192629896569997, 'timing/model': 2.031069872988155, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 1.77}) +[2026-06-23 15:10:38,327][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.019744066521525383, 'train/grad_norm_pre_clip': 11.149290084838867, 'timing/data': 0.054399199987528846, 'timing/model': 2.0290985489846207, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 1.77}) +[2026-06-23 15:10:42,433][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.025459131225943565, 'train/grad_norm_pre_clip': 13.559429168701172, 'timing/data': 0.0014994109806139022, 'timing/model': 2.0408500169869512, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 1.77}) +[2026-06-23 15:10:46,448][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.0187987070530653, 'train/grad_norm_pre_clip': 9.227137565612793, 'timing/data': 0.0013156219793017954, 'timing/model': 2.0339633999974467, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 1.77}) +[2026-06-23 15:10:50,501][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.04224123805761337, 'train/grad_norm_pre_clip': 25.372159957885742, 'timing/data': 0.0014127409958746284, 'timing/model': 2.070403097983217, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 1.77}) +[2026-06-23 15:10:54,582][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.05389038100838661, 'train/grad_norm_pre_clip': 18.950145721435547, 'timing/data': 0.06569001797470264, 'timing/model': 2.0355558229784947, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 1.77}) +[2026-06-23 15:10:58,680][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.01114037074148655, 'train/grad_norm_pre_clip': 9.189520835876465, 'timing/data': 0.0014692849945276976, 'timing/model': 2.033381102985004, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 1.77}) +[2026-06-23 15:11:02,719][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.01973927952349186, 'train/grad_norm_pre_clip': 15.306024551391602, 'timing/data': 0.0012507520150393248, 'timing/model': 2.0494407450023573, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 1.77}) +[2026-06-23 15:11:06,725][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.10289786756038666, 'train/grad_norm_pre_clip': 41.349853515625, 'timing/data': 0.0012464739847928286, 'timing/model': 2.020890979008982, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 1.77}) +[2026-06-23 15:11:10,811][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.04277156665921211, 'train/grad_norm_pre_clip': 24.610370635986328, 'timing/data': 0.06203035399084911, 'timing/model': 2.0465343529940583, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 1.77}) +[2026-06-23 15:11:14,904][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.023121163249015808, 'train/grad_norm_pre_clip': 19.78818130493164, 'timing/data': 0.001198785990709439, 'timing/model': 2.0337731780018657, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 1.77}) +[2026-06-23 15:11:18,915][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.0392884686589241, 'train/grad_norm_pre_clip': 22.196231842041016, 'timing/data': 0.001292428991291672, 'timing/model': 2.027758381998865, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 1.77}) +[2026-06-23 15:11:22,942][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.027781566604971886, 'train/grad_norm_pre_clip': 13.076265335083008, 'timing/data': 0.0013719159760512412, 'timing/model': 2.0438659120118245, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 1.77}) +[2026-06-23 15:11:27,128][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.007316321600228548, 'train/grad_norm_pre_clip': 4.2693586349487305, 'timing/data': 0.05521968597895466, 'timing/model': 2.0297341629920993, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 1.77}) +[2026-06-23 15:11:31,207][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.12614072859287262, 'train/grad_norm_pre_clip': 50.032352447509766, 'timing/data': 0.0012356229999568313, 'timing/model': 2.020399791013915, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 1.77}) +[2026-06-23 15:11:35,233][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.051521170884370804, 'train/grad_norm_pre_clip': 18.956867218017578, 'timing/data': 0.0013898489996790886, 'timing/model': 2.021440373995574, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 1.77}) +[2026-06-23 15:11:39,239][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.06516984850168228, 'train/grad_norm_pre_clip': 262.24462890625, 'timing/data': 0.0012787230079993606, 'timing/model': 2.0236299310054164, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 1.78}) +[2026-06-23 15:11:43,305][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.0934443324804306, 'train/grad_norm_pre_clip': 21.62918472290039, 'timing/data': 0.060341111005982384, 'timing/model': 2.026094285014551, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 1.78}) +[2026-06-23 15:11:47,425][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.19923453032970428, 'train/grad_norm_pre_clip': 38.11923599243164, 'timing/data': 0.0012849439808633178, 'timing/model': 2.04169788400759, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 1.78}) +[2026-06-23 15:11:51,430][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.05493331700563431, 'train/grad_norm_pre_clip': 22.591318130493164, 'timing/data': 0.0013494050072040409, 'timing/model': 2.022926561010536, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 1.78}) +[2026-06-23 15:11:55,454][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.06263042986392975, 'train/grad_norm_pre_clip': 25.625274658203125, 'timing/data': 0.0013129059807397425, 'timing/model': 2.038190660998225, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 1.78}) +[2026-06-23 15:11:59,750][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.036101508885622025, 'train/grad_norm_pre_clip': 16.012508392333984, 'timing/data': 0.06785988798947074, 'timing/model': 2.2416865100094583, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 1.78}) +[2026-06-23 15:12:03,840][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.045822564512491226, 'train/grad_norm_pre_clip': 16.702301025390625, 'timing/data': 0.0014683140034321696, 'timing/model': 2.0216825309908018, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 1.78}) +[2026-06-23 15:12:07,836][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.025389764457941055, 'train/grad_norm_pre_clip': 12.68448543548584, 'timing/data': 0.0012795549992006272, 'timing/model': 2.0180868759925943, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 1.78}) +[2026-06-23 15:12:11,839][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.03688570111989975, 'train/grad_norm_pre_clip': 18.086400985717773, 'timing/data': 0.0011725370131898671, 'timing/model': 2.0226025320007466, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 1.78}) +[2026-06-23 15:12:15,914][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.02252057008445263, 'train/grad_norm_pre_clip': 13.000615119934082, 'timing/data': 0.0718644089938607, 'timing/model': 2.0254454860114492, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 1.78}) +[2026-06-23 15:12:19,989][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.015065985731780529, 'train/grad_norm_pre_clip': 10.232335090637207, 'timing/data': 0.0014101859997026622, 'timing/model': 2.0172556290053762, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 1.78}) +[2026-06-23 15:12:23,986][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.01354037132114172, 'train/grad_norm_pre_clip': 4.669424533843994, 'timing/data': 0.001313797984039411, 'timing/model': 2.020338937989436, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 1.78}) +[2026-06-23 15:12:27,985][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.05122900381684303, 'train/grad_norm_pre_clip': 22.46138572692871, 'timing/data': 0.0012054580147378147, 'timing/model': 2.021872342011193, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 1.78}) +[2026-06-23 15:12:32,056][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4600 +[2026-06-23 15:14:27,003][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4600 +[2026-06-23 15:14:27,004][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4600 +[2026-06-23 15:16:34,149][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4600 +[2026-06-23 15:16:34,151][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.019568827003240585, 'train/grad_norm_pre_clip': 11.442500114440918, 'eval/action_loss/seconds': 114.94359872498899, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04144303873181343, 'eval/latency_2/loss': 0.04144303873181343, 'eval/flappy/loss': 0.04144303873181343, 'eval/flappy/latency_2/loss': 0.04144303873181343, 'eval/flap_precision': 0.9597044996642041, 'eval/flap_recall': 0.9642375168690959, 'eval/flap_f1': 0.9619656681252104, 'eval/flap_rate_model': 0.20680555555555555, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9619656681252104, 'eval/latency_2/flap_precision': 0.9597044996642041, 'eval/latency_2/flap_recall': 0.9642375168690959, 'eval/latency_2/flap_f1': 0.9619656681252104, 'eval/latency_2/flap_rate_model': 0.20680555555555555, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9619656681252104, 'eval/action_classification/data_wait_seconds': 50.34064865112305, 'eval/action_classification/predict_seconds': 74.92625427246094, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0709537549992092, 'timing/model': 2.022706363000907, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 1.78}) +[2026-06-23 15:16:38,190][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.03155868127942085, 'train/grad_norm_pre_clip': 18.1466064453125, 'timing/data': 0.0012269669969100505, 'timing/model': 1.9993555439868942, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 1.78}) +[2026-06-23 15:16:42,140][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.02744399383664131, 'train/grad_norm_pre_clip': 21.67888641357422, 'timing/data': 0.001293972018174827, 'timing/model': 1.9965999010019004, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 1.78}) +[2026-06-23 15:16:46,093][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.01282668299973011, 'train/grad_norm_pre_clip': 5.7851715087890625, 'timing/data': 0.0013568269787356257, 'timing/model': 1.9986189829942305, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 1.78}) +[2026-06-23 15:16:50,112][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.022234424948692322, 'train/grad_norm_pre_clip': 8.073270797729492, 'timing/data': 0.06390047998866066, 'timing/model': 1.9992342409968842, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 1.78}) +[2026-06-23 15:16:54,157][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.06331197172403336, 'train/grad_norm_pre_clip': 26.99147605895996, 'timing/data': 0.0013476310123223811, 'timing/model': 1.9979394859983586, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 1.78}) +[2026-06-23 15:16:58,212][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.01083291694521904, 'train/grad_norm_pre_clip': 7.413050651550293, 'timing/data': 0.0013719550042878836, 'timing/model': 2.0165136950090528, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 1.78}) +[2026-06-23 15:17:02,177][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.08582932502031326, 'train/grad_norm_pre_clip': 18.98053550720215, 'timing/data': 0.0014227589999791235, 'timing/model': 2.0040781120187603, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 1.78}) +[2026-06-23 15:17:06,205][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.01230438519269228, 'train/grad_norm_pre_clip': 6.5994462966918945, 'timing/data': 0.060110735008493066, 'timing/model': 2.0057598509883974, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 1.78}) +[2026-06-23 15:17:10,263][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.08006905019283295, 'train/grad_norm_pre_clip': 33.28310775756836, 'timing/data': 0.0011838780192192644, 'timing/model': 2.0082906580064446, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 1.78}) +[2026-06-23 15:17:14,229][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.04914749786257744, 'train/grad_norm_pre_clip': 28.28900909423828, 'timing/data': 0.001254839007742703, 'timing/model': 2.003943773015635, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 1.78}) +[2026-06-23 15:17:18,198][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.007870472967624664, 'train/grad_norm_pre_clip': 4.149567127227783, 'timing/data': 0.0012149960093665868, 'timing/model': 2.0055954060226213, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 1.78}) +[2026-06-23 15:17:22,273][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.06477857381105423, 'train/grad_norm_pre_clip': 39.044429779052734, 'timing/data': 0.09724372401251458, 'timing/model': 2.012380245985696, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 1.78}) +[2026-06-23 15:17:26,334][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.03516177088022232, 'train/grad_norm_pre_clip': 23.238149642944336, 'timing/data': 0.0012155860022176057, 'timing/model': 2.0083080700132996, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 1.79}) +[2026-06-23 15:17:30,483][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.035647179931402206, 'train/grad_norm_pre_clip': 13.157866477966309, 'timing/data': 0.0014285910001490265, 'timing/model': 2.1794764600053895, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 1.79}) +[2026-06-23 15:17:34,487][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.043905943632125854, 'train/grad_norm_pre_clip': 25.404695510864258, 'timing/data': 0.0014048070006538182, 'timing/model': 2.0332389670074917, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 1.79}) +[2026-06-23 15:17:38,532][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.1317741721868515, 'train/grad_norm_pre_clip': 40.976173400878906, 'timing/data': 0.06047265499364585, 'timing/model': 2.0134217209997587, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 1.79}) +[2026-06-23 15:17:42,623][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.0012257590424269438, 'train/grad_norm_pre_clip': 0.9067116379737854, 'timing/data': 0.0012429769849404693, 'timing/model': 2.032898417004617, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 1.79}) +[2026-06-23 15:17:46,615][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.07161066681146622, 'train/grad_norm_pre_clip': 17.87103271484375, 'timing/data': 0.0012473160168156028, 'timing/model': 2.0168813450145535, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 1.79}) +[2026-06-23 15:17:50,604][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.0029982232954353094, 'train/grad_norm_pre_clip': 2.1874959468841553, 'timing/data': 0.0013459580077324063, 'timing/model': 2.0147393660154194, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 1.79}) +[2026-06-23 15:17:54,654][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.03100709617137909, 'train/grad_norm_pre_clip': 30.57425308227539, 'timing/data': 0.061441105004632846, 'timing/model': 2.0162845909944735, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 1.79}) +[2026-06-23 15:17:58,732][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.04539542272686958, 'train/grad_norm_pre_clip': 16.52766990661621, 'timing/data': 0.0014432170137297362, 'timing/model': 2.0189804539841134, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 1.79}) +[2026-06-23 15:18:02,722][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.08521649986505508, 'train/grad_norm_pre_clip': 23.042192459106445, 'timing/data': 0.0012592870043590665, 'timing/model': 2.016634678991977, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 1.79}) +[2026-06-23 15:18:06,768][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.07335993647575378, 'train/grad_norm_pre_clip': 27.308130264282227, 'timing/data': 0.0013581099919974804, 'timing/model': 2.0366423260129523, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 1.79}) +[2026-06-23 15:18:10,828][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.03998017683625221, 'train/grad_norm_pre_clip': 29.313928604125977, 'timing/data': 0.06311364599969238, 'timing/model': 2.0213064920098986, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 1.79}) +[2026-06-23 15:18:14,926][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.10337387770414352, 'train/grad_norm_pre_clip': 25.257917404174805, 'timing/data': 0.0011821050138678402, 'timing/model': 2.026707115001045, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 1.79}) +[2026-06-23 15:18:18,925][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.05412555858492851, 'train/grad_norm_pre_clip': 40.816566467285156, 'timing/data': 0.0012290710001252592, 'timing/model': 2.018122046982171, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 1.79}) +[2026-06-23 15:18:22,931][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.047436363995075226, 'train/grad_norm_pre_clip': 30.087892532348633, 'timing/data': 0.0012240820215083659, 'timing/model': 2.0208446189935785, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 1.79}) +[2026-06-23 15:18:26,991][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.036644455045461655, 'train/grad_norm_pre_clip': 20.22292709350586, 'timing/data': 0.0686031119839754, 'timing/model': 2.0176439940114506, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 1.79}) +[2026-06-23 15:18:31,064][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.07052024453878403, 'train/grad_norm_pre_clip': 27.40242576599121, 'timing/data': 0.0014343310031108558, 'timing/model': 2.0135026199859567, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 1.79}) +[2026-06-23 15:18:35,049][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.005811682436615229, 'train/grad_norm_pre_clip': 5.678991794586182, 'timing/data': 0.0011614260147325695, 'timing/model': 2.014490425004624, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 1.79}) +[2026-06-23 15:18:39,063][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.151133194565773, 'train/grad_norm_pre_clip': 41.755653381347656, 'timing/data': 0.0014069000026211143, 'timing/model': 2.023666754976148, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 1.79}) +[2026-06-23 15:18:43,328][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.06978525966405869, 'train/grad_norm_pre_clip': 41.719398498535156, 'timing/data': 0.06976363601279445, 'timing/model': 2.0318145439960063, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 1.79}) +[2026-06-23 15:18:47,415][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.045707907527685165, 'train/grad_norm_pre_clip': 14.144318580627441, 'timing/data': 0.0012554700078908354, 'timing/model': 2.0202165480004624, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 1.79}) +[2026-06-23 15:18:51,418][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.023426562547683716, 'train/grad_norm_pre_clip': 27.729459762573242, 'timing/data': 0.001280806987779215, 'timing/model': 2.022700478992192, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 1.79}) +[2026-06-23 15:18:55,429][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.19860081374645233, 'train/grad_norm_pre_clip': 35.803958892822266, 'timing/data': 0.001243557984707877, 'timing/model': 2.0237632729986217, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 1.79}) +[2026-06-23 15:18:59,485][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.05380240082740784, 'train/grad_norm_pre_clip': 18.260805130004883, 'timing/data': 0.056131831981474534, 'timing/model': 2.0207627480267547, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 1.79}) +[2026-06-23 15:19:03,574][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.07988189160823822, 'train/grad_norm_pre_clip': 28.524728775024414, 'timing/data': 0.0014723210188094527, 'timing/model': 2.0218568600248545, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 1.79}) +[2026-06-23 15:19:07,572][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.022446345537900925, 'train/grad_norm_pre_clip': 18.285884857177734, 'timing/data': 0.0011858420039061457, 'timing/model': 2.0219275600102264, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 1.79}) +[2026-06-23 15:19:11,581][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.15779663622379303, 'train/grad_norm_pre_clip': 29.3527889251709, 'timing/data': 0.001186353008961305, 'timing/model': 2.0272417819942348, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 1.8}) +[2026-06-23 15:19:15,676][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.09143167734146118, 'train/grad_norm_pre_clip': 45.50154113769531, 'timing/data': 0.057539443019777536, 'timing/model': 2.0248381900019012, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 1.8}) +[2026-06-23 15:19:19,758][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.07208956032991409, 'train/grad_norm_pre_clip': 27.557910919189453, 'timing/data': 0.0013827460061293095, 'timing/model': 2.0240600920224097, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 1.8}) +[2026-06-23 15:19:23,759][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.048774901777505875, 'train/grad_norm_pre_clip': 23.83759117126465, 'timing/data': 0.0012693569879047573, 'timing/model': 2.022247080982197, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 1.8}) +[2026-06-23 15:19:27,786][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.060603439807891846, 'train/grad_norm_pre_clip': 42.34989929199219, 'timing/data': 0.0013562270032707602, 'timing/model': 2.031233388988767, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 1.8}) +[2026-06-23 15:19:31,841][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.13479474186897278, 'train/grad_norm_pre_clip': 36.972450256347656, 'timing/data': 0.061759011005051434, 'timing/model': 2.0181305830192287, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 1.8}) +[2026-06-23 15:19:35,918][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.02435046061873436, 'train/grad_norm_pre_clip': 15.299361228942871, 'timing/data': 0.0012414249940775335, 'timing/model': 2.018508941982873, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 1.8}) +[2026-06-23 15:19:39,942][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.04868664965033531, 'train/grad_norm_pre_clip': 33.8917236328125, 'timing/data': 0.0012607009848579764, 'timing/model': 2.030538484978024, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 1.8}) +[2026-06-23 15:19:43,961][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.04143139347434044, 'train/grad_norm_pre_clip': 25.286483764648438, 'timing/data': 0.0013648620224557817, 'timing/model': 2.0273493909917306, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 1.8}) +[2026-06-23 15:19:48,071][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.07378437370061874, 'train/grad_norm_pre_clip': 27.57917022705078, 'timing/data': 0.0585129119863268, 'timing/model': 2.0637079480220564, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 1.8}) +[2026-06-23 15:19:52,186][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.046152953058481216, 'train/grad_norm_pre_clip': 25.8795223236084, 'timing/data': 0.0013532409793697298, 'timing/model': 2.0411669879977126, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 1.8}) +[2026-06-23 15:19:56,190][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.03608265519142151, 'train/grad_norm_pre_clip': 20.76137351989746, 'timing/data': 0.001244489016244188, 'timing/model': 2.019900012994185, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 1.8}) +[2026-06-23 15:20:00,209][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.05438191816210747, 'train/grad_norm_pre_clip': 22.944374084472656, 'timing/data': 0.0011638810101430863, 'timing/model': 2.0339228389784694, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 1.8}) +[2026-06-23 15:20:04,276][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.009510096162557602, 'train/grad_norm_pre_clip': 5.184913635253906, 'timing/data': 0.05715278902789578, 'timing/model': 2.024855932977516, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 1.8}) +[2026-06-23 15:20:08,373][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.06025521457195282, 'train/grad_norm_pre_clip': 19.646432876586914, 'timing/data': 0.0013632200134452432, 'timing/model': 2.0269555730046704, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 1.8}) +[2026-06-23 15:20:12,399][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.004929129965603352, 'train/grad_norm_pre_clip': 4.057835102081299, 'timing/data': 0.0012474950053729117, 'timing/model': 2.042966774024535, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 1.8}) +[2026-06-23 15:20:16,401][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.11427365243434906, 'train/grad_norm_pre_clip': 56.78385925292969, 'timing/data': 0.0013185770076233894, 'timing/model': 2.021453854016727, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 1.8}) +[2026-06-23 15:20:20,479][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.005511074792593718, 'train/grad_norm_pre_clip': 3.152432441711426, 'timing/data': 0.0548582180053927, 'timing/model': 2.038389936002204, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 1.8}) +[2026-06-23 15:20:24,617][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.10420766472816467, 'train/grad_norm_pre_clip': 37.458213806152344, 'timing/data': 0.0013973730092402548, 'timing/model': 2.024890645989217, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 1.8}) +[2026-06-23 15:20:28,640][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.006642749533057213, 'train/grad_norm_pre_clip': 4.736881732940674, 'timing/data': 0.0012747459986712784, 'timing/model': 2.0332405980152544, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 1.8}) +[2026-06-23 15:20:32,670][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.016644783318042755, 'train/grad_norm_pre_clip': 15.504441261291504, 'timing/data': 0.0014802760269958526, 'timing/model': 2.0329314169939607, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 1.8}) +[2026-06-23 15:20:36,751][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.12382391840219498, 'train/grad_norm_pre_clip': 15.529777526855469, 'timing/data': 0.06289484101580456, 'timing/model': 2.03920580400154, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 1.8}) +[2026-06-23 15:20:40,845][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.13887493312358856, 'train/grad_norm_pre_clip': 31.872804641723633, 'timing/data': 0.0015652719885110855, 'timing/model': 2.0218631099851336, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 1.8}) +[2026-06-23 15:20:44,864][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.03605961799621582, 'train/grad_norm_pre_clip': 30.74287223815918, 'timing/data': 0.0015014850068837404, 'timing/model': 2.0387688649934717, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 1.8}) +[2026-06-23 15:20:48,870][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.006105165928602219, 'train/grad_norm_pre_clip': 3.695297956466675, 'timing/data': 0.0013433829881250858, 'timing/model': 2.024734507984249, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 1.8}) +[2026-06-23 15:20:52,935][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.02548382431268692, 'train/grad_norm_pre_clip': 15.99805736541748, 'timing/data': 0.06388282700208947, 'timing/model': 2.019473124004435, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 1.81}) +[2026-06-23 15:20:57,090][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.05702124908566475, 'train/grad_norm_pre_clip': 37.272552490234375, 'timing/data': 0.001486436987761408, 'timing/model': 2.0834849090024363, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 1.81}) +[2026-06-23 15:21:01,103][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.06479571759700775, 'train/grad_norm_pre_clip': 41.809112548828125, 'timing/data': 0.0012730729940813035, 'timing/model': 2.0285027519857977, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 1.81}) +[2026-06-23 15:21:05,106][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.006279344670474529, 'train/grad_norm_pre_clip': 7.678070068359375, 'timing/data': 0.0013037489843554795, 'timing/model': 2.0250293420103844, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 1.81}) +[2026-06-23 15:21:09,191][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.03431668132543564, 'train/grad_norm_pre_clip': 30.452587127685547, 'timing/data': 0.06745297499583103, 'timing/model': 2.024933505017543, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 1.81}) +[2026-06-23 15:21:13,284][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.011692775413393974, 'train/grad_norm_pre_clip': 9.701619148254395, 'timing/data': 0.0011356089962646365, 'timing/model': 2.0203420800098684, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 1.81}) +[2026-06-23 15:21:17,283][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.07678813487291336, 'train/grad_norm_pre_clip': 23.13271713256836, 'timing/data': 0.001176785008283332, 'timing/model': 2.020612499996787, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 1.81}) +[2026-06-23 15:21:21,302][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.031700290739536285, 'train/grad_norm_pre_clip': 19.429540634155273, 'timing/data': 0.001116392988478765, 'timing/model': 2.0398163410136476, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 1.81}) +[2026-06-23 15:21:25,382][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.02309310808777809, 'train/grad_norm_pre_clip': 10.450929641723633, 'timing/data': 0.05643794799107127, 'timing/model': 2.0288801490096375, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 1.81}) +[2026-06-23 15:21:29,478][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.15418896079063416, 'train/grad_norm_pre_clip': 27.901338577270508, 'timing/data': 0.001186282024718821, 'timing/model': 2.0188529579900205, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 1.81}) +[2026-06-23 15:21:33,627][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.0778348445892334, 'train/grad_norm_pre_clip': 28.70183753967285, 'timing/data': 0.0012404629960656166, 'timing/model': 2.0193280350067653, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 1.81}) +[2026-06-23 15:21:37,624][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.04119059816002846, 'train/grad_norm_pre_clip': 26.304662704467773, 'timing/data': 0.0011108329927083105, 'timing/model': 2.021837303007487, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 1.81}) +[2026-06-23 15:21:41,805][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.03574172034859657, 'train/grad_norm_pre_clip': 13.864060401916504, 'timing/data': 0.06062361298245378, 'timing/model': 2.0231858949991874, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 1.81}) +[2026-06-23 15:21:46,000][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.004358474165201187, 'train/grad_norm_pre_clip': 2.9104714393615723, 'timing/data': 0.001229050976689905, 'timing/model': 2.03323551901849, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 1.81}) +[2026-06-23 15:21:50,023][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.03740542754530907, 'train/grad_norm_pre_clip': 18.440216064453125, 'timing/data': 0.0012449719943106174, 'timing/model': 2.040565907023847, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 1.81}) +[2026-06-23 15:21:54,038][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.01805446296930313, 'train/grad_norm_pre_clip': 13.724549293518066, 'timing/data': 0.0014631040103267878, 'timing/model': 2.0288601830252446, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 1.81}) +[2026-06-23 15:21:58,126][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.050909724086523056, 'train/grad_norm_pre_clip': 15.372652053833008, 'timing/data': 0.062464054994052276, 'timing/model': 2.030921801982913, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 1.81}) +[2026-06-23 15:22:02,227][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.01064174436032772, 'train/grad_norm_pre_clip': 24.148897171020508, 'timing/data': 0.0012665999820455909, 'timing/model': 2.02786899000057, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 1.81}) +[2026-06-23 15:22:06,496][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.11578724533319473, 'train/grad_norm_pre_clip': 32.64453125, 'timing/data': 0.001238028024090454, 'timing/model': 2.0222478119831067, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 1.81}) +[2026-06-23 15:22:10,508][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.07092873752117157, 'train/grad_norm_pre_clip': 49.18687057495117, 'timing/data': 0.001191130984807387, 'timing/model': 2.0286168830061797, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 1.81}) +[2026-06-23 15:22:14,585][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.1250157654285431, 'train/grad_norm_pre_clip': 68.65025329589844, 'timing/data': 0.07410043198615313, 'timing/model': 2.0243679489940405, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 1.81}) +[2026-06-23 15:22:18,678][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.06213519722223282, 'train/grad_norm_pre_clip': 65.21385192871094, 'timing/data': 0.0014409329742193222, 'timing/model': 2.0262618910055608, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 1.81}) +[2026-06-23 15:22:22,685][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.13673819601535797, 'train/grad_norm_pre_clip': 39.22002410888672, 'timing/data': 0.0012272089952602983, 'timing/model': 2.0268607570033055, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 1.81}) +[2026-06-23 15:22:26,688][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.034366801381111145, 'train/grad_norm_pre_clip': 27.45346450805664, 'timing/data': 0.0013243569992482662, 'timing/model': 2.0229262150241993, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 1.81}) +[2026-06-23 15:22:30,750][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.05196401849389076, 'train/grad_norm_pre_clip': 17.533845901489258, 'timing/data': 0.05571712501114234, 'timing/model': 2.0252781320014037, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 1.81}) +[2026-06-23 15:22:34,848][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.03511340916156769, 'train/grad_norm_pre_clip': 17.044340133666992, 'timing/data': 0.0013701319985557348, 'timing/model': 2.0366823299846146, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 1.81}) +[2026-06-23 15:22:39,117][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.04800797626376152, 'train/grad_norm_pre_clip': 34.0090217590332, 'timing/data': 0.0016647670126985759, 'timing/model': 2.285802198020974, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 1.82}) +[2026-06-23 15:22:43,125][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.10244773328304291, 'train/grad_norm_pre_clip': 19.95323371887207, 'timing/data': 0.0016252940113190562, 'timing/model': 2.023185103025753, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 1.82}) +[2026-06-23 15:22:47,203][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.02382161282002926, 'train/grad_norm_pre_clip': 10.910313606262207, 'timing/data': 0.06939846999011934, 'timing/model': 2.0269311770098284, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 1.82}) +[2026-06-23 15:22:51,290][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.14572006464004517, 'train/grad_norm_pre_clip': 45.5217170715332, 'timing/data': 0.0011850600130856037, 'timing/model': 2.0218860329769086, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 1.82}) +[2026-06-23 15:22:55,295][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.06963551789522171, 'train/grad_norm_pre_clip': 23.41335678100586, 'timing/data': 0.0012420850107446313, 'timing/model': 2.0230468579975422, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 1.82}) +[2026-06-23 15:22:59,301][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.049859244376420975, 'train/grad_norm_pre_clip': 30.10659408569336, 'timing/data': 0.0013104419922456145, 'timing/model': 2.0260548270016443, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 1.82}) +[2026-06-23 15:23:03,371][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.03296288102865219, 'train/grad_norm_pre_clip': 37.12812423706055, 'timing/data': 0.06330917699960992, 'timing/model': 2.026669362006942, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 1.82}) +[2026-06-23 15:23:07,484][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.11017806082963943, 'train/grad_norm_pre_clip': 28.72590446472168, 'timing/data': 0.0013621280086226761, 'timing/model': 2.042590688011842, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 1.82}) +[2026-06-23 15:23:11,496][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.1072407141327858, 'train/grad_norm_pre_clip': 16.249631881713867, 'timing/data': 0.0012443400046322495, 'timing/model': 2.0272519610007294, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 1.82}) +[2026-06-23 15:23:15,656][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.0999513790011406, 'train/grad_norm_pre_clip': 36.057167053222656, 'timing/data': 0.00136547299916856, 'timing/model': 2.0332431530114263, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 1.82}) +[2026-06-23 15:23:19,725][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4700 +[2026-06-23 15:25:14,167][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4700 +[2026-06-23 15:25:14,168][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4700 +[2026-06-23 15:27:21,312][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4700 +[2026-06-23 15:27:21,314][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.033417634665966034, 'train/grad_norm_pre_clip': 17.374248504638672, 'eval/action_loss/seconds': 114.43862794499728, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04249846935272217, 'eval/latency_2/loss': 0.04249846935272217, 'eval/flappy/loss': 0.04249846935272217, 'eval/flappy/latency_2/loss': 0.04249846935272217, 'eval/flap_precision': 0.9548472775564409, 'eval/flap_recall': 0.970310391363023, 'eval/flap_f1': 0.962516733601071, 'eval/flap_rate_model': 0.20916666666666667, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9625167336010709, 'eval/latency_2/flap_precision': 0.9548472775564409, 'eval/latency_2/flap_recall': 0.970310391363023, 'eval/latency_2/flap_f1': 0.962516733601071, 'eval/latency_2/flap_rate_model': 0.20916666666666667, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9625167336010709, 'eval/action_classification/data_wait_seconds': 50.4895133972168, 'eval/action_classification/predict_seconds': 74.92011260986328, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.06320089701330289, 'timing/model': 2.024455320992274, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 1.82}) +[2026-06-23 15:27:25,366][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.20048542320728302, 'train/grad_norm_pre_clip': 26.673452377319336, 'timing/data': 0.0015221229987218976, 'timing/model': 2.008368098991923, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 1.82}) +[2026-06-23 15:27:29,315][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.12314730882644653, 'train/grad_norm_pre_clip': 31.1544246673584, 'timing/data': 0.0012289909936953336, 'timing/model': 1.998958678013878, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 1.82}) +[2026-06-23 15:27:33,265][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.18103991448879242, 'train/grad_norm_pre_clip': 44.55768585205078, 'timing/data': 0.0013525999966077507, 'timing/model': 1.9968260359892156, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 1.82}) +[2026-06-23 15:27:37,439][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.02768288552761078, 'train/grad_norm_pre_clip': 22.820659637451172, 'timing/data': 0.06532693599001504, 'timing/model': 2.156020577007439, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 1.82}) +[2026-06-23 15:27:41,492][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.09155082702636719, 'train/grad_norm_pre_clip': 29.371294021606445, 'timing/data': 0.001372697006445378, 'timing/model': 2.0128537279961165, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 1.82}) +[2026-06-23 15:27:45,452][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.005827031563967466, 'train/grad_norm_pre_clip': 4.457337856292725, 'timing/data': 0.0012699270155280828, 'timing/model': 2.0003515620192047, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 1.82}) +[2026-06-23 15:27:49,418][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.07749084383249283, 'train/grad_norm_pre_clip': 28.008649826049805, 'timing/data': 0.0012673320015892386, 'timing/model': 2.0044909930147696, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 1.82}) +[2026-06-23 15:27:53,451][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.16286559402942657, 'train/grad_norm_pre_clip': 38.001644134521484, 'timing/data': 0.06245427700923756, 'timing/model': 2.0078538579982705, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 1.82}) +[2026-06-23 15:27:57,512][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.08733988553285599, 'train/grad_norm_pre_clip': 30.821765899658203, 'timing/data': 0.001336821005679667, 'timing/model': 2.008628920011688, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 1.82}) +[2026-06-23 15:28:01,504][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.0638197809457779, 'train/grad_norm_pre_clip': 33.195152282714844, 'timing/data': 0.0015074570255819708, 'timing/model': 2.0242748070158996, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 1.82}) +[2026-06-23 15:28:05,483][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.14202949404716492, 'train/grad_norm_pre_clip': 18.757658004760742, 'timing/data': 0.0014262860058806837, 'timing/model': 2.010300330002792, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 1.82}) +[2026-06-23 15:28:09,524][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.06496261060237885, 'train/grad_norm_pre_clip': 25.72478675842285, 'timing/data': 0.06273202001466416, 'timing/model': 2.010268320009345, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 1.82}) +[2026-06-23 15:28:13,662][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.01573397032916546, 'train/grad_norm_pre_clip': 10.42952823638916, 'timing/data': 0.0015418099937960505, 'timing/model': 2.0122074139944743, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 1.82}) +[2026-06-23 15:28:17,645][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.05995592474937439, 'train/grad_norm_pre_clip': 21.40863800048828, 'timing/data': 0.0012248640123289078, 'timing/model': 2.01251102599781, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 1.82}) +[2026-06-23 15:28:21,632][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.02129870094358921, 'train/grad_norm_pre_clip': 15.635022163391113, 'timing/data': 0.0012101869797334075, 'timing/model': 2.010806343983859, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 1.82}) +[2026-06-23 15:28:25,681][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.12488219887018204, 'train/grad_norm_pre_clip': 18.626293182373047, 'timing/data': 0.056794676987919956, 'timing/model': 2.018671830999665, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 1.83}) +[2026-06-23 15:28:29,764][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.09928218275308609, 'train/grad_norm_pre_clip': 30.696779251098633, 'timing/data': 0.0014398510102182627, 'timing/model': 2.015107484010514, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 1.83}) +[2026-06-23 15:28:33,769][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.019754888489842415, 'train/grad_norm_pre_clip': 7.010617733001709, 'timing/data': 0.0011714950087480247, 'timing/model': 2.0230542109929956, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 1.83}) +[2026-06-23 15:28:37,777][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.041519615799188614, 'train/grad_norm_pre_clip': 31.44683074951172, 'timing/data': 0.0012799550022464246, 'timing/model': 2.0300506509956904, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 1.83}) +[2026-06-23 15:28:41,835][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.06643363833427429, 'train/grad_norm_pre_clip': 38.29066848754883, 'timing/data': 0.06149075599387288, 'timing/model': 2.021576819999609, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 1.83}) +[2026-06-23 15:28:46,084][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.07151105999946594, 'train/grad_norm_pre_clip': 25.03866958618164, 'timing/data': 0.0014794540184084326, 'timing/model': 2.18227979898802, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 1.83}) +[2026-06-23 15:28:50,084][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.023362115025520325, 'train/grad_norm_pre_clip': 16.669771194458008, 'timing/data': 0.0013325019972398877, 'timing/model': 2.0223935100075323, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 1.83}) +[2026-06-23 15:28:54,084][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.0423196479678154, 'train/grad_norm_pre_clip': 23.8050537109375, 'timing/data': 0.0012039449939038604, 'timing/model': 2.018494574003853, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 1.83}) +[2026-06-23 15:28:58,138][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.02824348211288452, 'train/grad_norm_pre_clip': 23.547895431518555, 'timing/data': 0.05615629701060243, 'timing/model': 2.0254856039828155, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 1.83}) +[2026-06-23 15:29:02,245][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.11554916948080063, 'train/grad_norm_pre_clip': 64.30915832519531, 'timing/data': 0.0012342410045675933, 'timing/model': 2.0230524280050304, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 1.83}) +[2026-06-23 15:29:06,259][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.02852356992661953, 'train/grad_norm_pre_clip': 23.282482147216797, 'timing/data': 0.0016188719891943038, 'timing/model': 2.02430682600243, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 1.83}) +[2026-06-23 15:29:10,268][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.00802543107420206, 'train/grad_norm_pre_clip': 6.939042568206787, 'timing/data': 0.0013425009965430945, 'timing/model': 2.025977804005379, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 1.83}) +[2026-06-23 15:29:14,343][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.13254593312740326, 'train/grad_norm_pre_clip': 27.369009017944336, 'timing/data': 0.06850672300788574, 'timing/model': 2.0223033119982574, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 1.83}) +[2026-06-23 15:29:18,435][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.04040960222482681, 'train/grad_norm_pre_clip': 28.242229461669922, 'timing/data': 0.0013743799936491996, 'timing/model': 2.0245172640134115, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 1.83}) +[2026-06-23 15:29:22,679][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.014095297083258629, 'train/grad_norm_pre_clip': 12.943678855895996, 'timing/data': 0.0014158269914332777, 'timing/model': 2.026614631002303, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 1.83}) +[2026-06-23 15:29:26,690][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.14324907958507538, 'train/grad_norm_pre_clip': 28.29412841796875, 'timing/data': 0.0015207009855657816, 'timing/model': 2.0273229599988554, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 1.83}) +[2026-06-23 15:29:30,776][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.03332264721393585, 'train/grad_norm_pre_clip': 21.576221466064453, 'timing/data': 0.06831957699614577, 'timing/model': 2.0292430789850187, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 1.83}) +[2026-06-23 15:29:34,873][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.06778205931186676, 'train/grad_norm_pre_clip': 35.19483947753906, 'timing/data': 0.0012298630026634783, 'timing/model': 2.0277762280020397, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 1.83}) +[2026-06-23 15:29:38,887][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.07041162997484207, 'train/grad_norm_pre_clip': 13.030060768127441, 'timing/data': 0.0012309550074860454, 'timing/model': 2.0288246749842074, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 1.83}) +[2026-06-23 15:29:42,916][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.006795017048716545, 'train/grad_norm_pre_clip': 3.8825366497039795, 'timing/data': 0.0013403969933278859, 'timing/model': 2.0463851199892815, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 1.83}) +[2026-06-23 15:29:46,981][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.15838691592216492, 'train/grad_norm_pre_clip': 48.22767639160156, 'timing/data': 0.06471101701026782, 'timing/model': 2.0203966189874336, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 1.83}) +[2026-06-23 15:29:51,075][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.008764142170548439, 'train/grad_norm_pre_clip': 4.639638423919678, 'timing/data': 0.0012076419952791184, 'timing/model': 2.0273610110161826, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 1.83}) +[2026-06-23 15:29:55,181][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.010538856498897076, 'train/grad_norm_pre_clip': 9.053746223449707, 'timing/data': 0.0014286109944805503, 'timing/model': 2.1204030489898287, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 1.83}) +[2026-06-23 15:29:59,204][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.015631668269634247, 'train/grad_norm_pre_clip': 9.298198699951172, 'timing/data': 0.0012081230233889073, 'timing/model': 2.0341785000055097, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 1.83}) +[2026-06-23 15:30:03,268][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.004190311301499605, 'train/grad_norm_pre_clip': 1.7353078126907349, 'timing/data': 0.05595569600700401, 'timing/model': 2.023843389004469, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 1.83}) +[2026-06-23 15:30:07,370][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.056004662066698074, 'train/grad_norm_pre_clip': 24.01215362548828, 'timing/data': 0.0014935299986973405, 'timing/model': 2.0307647619920317, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 1.83}) +[2026-06-23 15:30:11,412][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.017409900203347206, 'train/grad_norm_pre_clip': 17.94571876525879, 'timing/data': 0.001391822996083647, 'timing/model': 2.053778143977979, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 1.84}) +[2026-06-23 15:30:15,439][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.06271380931138992, 'train/grad_norm_pre_clip': 19.4897403717041, 'timing/data': 0.0016495989984832704, 'timing/model': 2.0395623900112696, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 1.84}) +[2026-06-23 15:30:19,511][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.016025573015213013, 'train/grad_norm_pre_clip': 10.3508939743042, 'timing/data': 0.056717783998465165, 'timing/model': 2.0278367699938826, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 1.84}) +[2026-06-23 15:30:23,629][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.03639635816216469, 'train/grad_norm_pre_clip': 24.8907470703125, 'timing/data': 0.0014329490077216178, 'timing/model': 2.0429404540045653, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 1.84}) +[2026-06-23 15:30:27,657][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.11764310300350189, 'train/grad_norm_pre_clip': 25.08534049987793, 'timing/data': 0.001224242994794622, 'timing/model': 2.037666977994377, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 1.84}) +[2026-06-23 15:30:31,888][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.055393289774656296, 'train/grad_norm_pre_clip': 22.33681297302246, 'timing/data': 0.0013526100083254278, 'timing/model': 2.0378774169948883, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 1.84}) +[2026-06-23 15:30:35,965][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.0933174341917038, 'train/grad_norm_pre_clip': 18.704187393188477, 'timing/data': 0.06483717099763453, 'timing/model': 2.0280280130100437, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 1.84}) +[2026-06-23 15:30:40,064][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.0792122632265091, 'train/grad_norm_pre_clip': 25.797245025634766, 'timing/data': 0.0014698559825774282, 'timing/model': 2.026820471015526, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 1.84}) +[2026-06-23 15:48:38,420][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4750 +[2026-06-23 15:48:38,421][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.01130690984427929, 'train/grad_norm_pre_clip': 7.218193054199219, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 278.1800000000084, 'rl_games_eval/mid_train/mean_length': 2261.0, 'rl_games_eval/mid_train/std_reward': 164.18610659859004, 'rl_games_eval/mid_train/std_length': 1322.34791185981, 'rl_games_eval/mid_train/macro_mean_reward': 278.1800000000084, 'rl_games_eval/mid_train/macro_mean_length': 2261.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 278.1800000000084, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2261.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 164.18610659859004, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1322.34791185981, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 278.1800000000084, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.001274866983294487, 'timing/model': 2.0320181170245633, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 1.84}) +[2026-06-23 15:48:42,290][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.1464649885892868, 'train/grad_norm_pre_clip': 26.33342170715332, 'timing/data': 0.0012112669937778264, 'timing/model': 1.9580946670030244, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 1.84}) +[2026-06-23 15:48:46,475][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.09450561553239822, 'train/grad_norm_pre_clip': 26.977663040161133, 'timing/data': 0.30851036199601367, 'timing/model': 1.9607419640233275, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 1.84}) +[2026-06-23 15:48:50,466][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.08058550953865051, 'train/grad_norm_pre_clip': 18.757991790771484, 'timing/data': 0.001398612977936864, 'timing/model': 1.9771081999933813, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 1.84}) +[2026-06-23 15:48:54,383][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.04357055202126503, 'train/grad_norm_pre_clip': 13.2965087890625, 'timing/data': 0.0012489260116126388, 'timing/model': 1.9926045600150246, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 1.84}) +[2026-06-23 15:48:58,306][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.08713503181934357, 'train/grad_norm_pre_clip': 37.88677978515625, 'timing/data': 0.0011582690058276057, 'timing/model': 1.997739187994739, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 1.84}) +[2026-06-23 15:49:02,290][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.07199864089488983, 'train/grad_norm_pre_clip': 23.545886993408203, 'timing/data': 0.06650085697765462, 'timing/model': 1.9832673459895886, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 1.84}) +[2026-06-23 15:49:06,308][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.006447690539062023, 'train/grad_norm_pre_clip': 2.999972343444824, 'timing/data': 0.0014333059953060001, 'timing/model': 1.9999525400053244, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 1.84}) +[2026-06-23 15:49:10,220][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.07503117620944977, 'train/grad_norm_pre_clip': 36.14728546142578, 'timing/data': 0.0014829580031801015, 'timing/model': 1.9786751409992576, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 1.84}) +[2026-06-23 15:49:14,153][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.1005280613899231, 'train/grad_norm_pre_clip': 19.19477653503418, 'timing/data': 0.001383293973049149, 'timing/model': 1.9953479930118192, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 1.84}) +[2026-06-23 15:49:18,426][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.022920001298189163, 'train/grad_norm_pre_clip': 9.902268409729004, 'timing/data': 0.06526503700297326, 'timing/model': 2.248620282014599, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 1.84}) +[2026-06-23 15:49:22,466][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.08061408996582031, 'train/grad_norm_pre_clip': 21.580291748046875, 'timing/data': 0.0014546059828717262, 'timing/model': 2.002760245988611, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 1.84}) +[2026-06-23 15:49:26,405][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.1264033317565918, 'train/grad_norm_pre_clip': 32.12739562988281, 'timing/data': 0.001278420997550711, 'timing/model': 1.9947049290058203, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 1.84}) +[2026-06-23 15:49:30,356][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.09126322716474533, 'train/grad_norm_pre_clip': 47.14500045776367, 'timing/data': 0.0013137960049789399, 'timing/model': 2.00529903700226, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 1.84}) +[2026-06-23 15:49:34,366][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.04552139341831207, 'train/grad_norm_pre_clip': 13.490957260131836, 'timing/data': 0.05670318999909796, 'timing/model': 2.0068813349935226, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 1.84}) +[2026-06-23 15:49:38,408][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.053469929844141006, 'train/grad_norm_pre_clip': 21.427928924560547, 'timing/data': 0.0012408810143824667, 'timing/model': 1.9970043049834203, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 1.84}) +[2026-06-23 15:49:42,362][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.05901686102151871, 'train/grad_norm_pre_clip': 24.136014938354492, 'timing/data': 0.0012549670063890517, 'timing/model': 1.9963647410040721, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 1.84}) +[2026-06-23 15:49:46,332][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.03267139568924904, 'train/grad_norm_pre_clip': 14.20626163482666, 'timing/data': 0.0011566660250537097, 'timing/model': 2.0144408640044276, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 1.84}) +[2026-06-23 15:49:50,428][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.026065852493047714, 'train/grad_norm_pre_clip': 14.148686408996582, 'timing/data': 0.06228353700134903, 'timing/model': 2.078327788010938, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 1.85}) +[2026-06-23 15:49:54,487][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.031023794785141945, 'train/grad_norm_pre_clip': 18.318504333496094, 'timing/data': 0.0012924970069434494, 'timing/model': 2.0154383089975454, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 1.85}) +[2026-06-23 15:49:58,454][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.03252425789833069, 'train/grad_norm_pre_clip': 13.657732009887695, 'timing/data': 0.0011489219905342907, 'timing/model': 2.003308216982987, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 1.85}) +[2026-06-23 15:50:02,414][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.05016735568642616, 'train/grad_norm_pre_clip': 16.517160415649414, 'timing/data': 0.0011153890227433294, 'timing/model': 2.002401457983069, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 1.85}) +[2026-06-23 15:50:06,449][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.07637381553649902, 'train/grad_norm_pre_clip': 25.65217399597168, 'timing/data': 0.05630231401300989, 'timing/model': 2.0131876509985887, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 1.85}) +[2026-06-23 15:50:10,517][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.034150976687669754, 'train/grad_norm_pre_clip': 16.755210876464844, 'timing/data': 0.0014257929869927466, 'timing/model': 2.015994139976101, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 1.85}) +[2026-06-23 15:50:14,486][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.03286171704530716, 'train/grad_norm_pre_clip': 21.90949249267578, 'timing/data': 0.0011689189996104687, 'timing/model': 2.001195118005853, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 1.85}) +[2026-06-23 15:50:18,469][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.02856866642832756, 'train/grad_norm_pre_clip': 20.303098678588867, 'timing/data': 0.0012330370082054287, 'timing/model': 2.0141480620077346, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 1.85}) +[2026-06-23 15:50:22,497][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.09935668855905533, 'train/grad_norm_pre_clip': 67.3287124633789, 'timing/data': 0.055902486987179145, 'timing/model': 2.0057084030122496, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 1.85}) +[2026-06-23 15:50:26,577][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.028289536014199257, 'train/grad_norm_pre_clip': 20.50182342529297, 'timing/data': 0.0012675810139626265, 'timing/model': 2.0069973690260667, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 1.85}) +[2026-06-23 15:50:30,565][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.11999856680631638, 'train/grad_norm_pre_clip': 49.67654800415039, 'timing/data': 0.0012686429836321622, 'timing/model': 2.0168727800191846, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 1.85}) +[2026-06-23 15:50:34,559][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.042224593460559845, 'train/grad_norm_pre_clip': 28.670799255371094, 'timing/data': 0.0012551179970614612, 'timing/model': 2.0170902160170954, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 1.85}) +[2026-06-23 15:50:38,618][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.024818919599056244, 'train/grad_norm_pre_clip': 11.392441749572754, 'timing/data': 0.05644965899409726, 'timing/model': 2.028482989000622, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 1.85}) +[2026-06-23 15:50:42,699][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.05930815637111664, 'train/grad_norm_pre_clip': 28.297637939453125, 'timing/data': 0.0013666839804500341, 'timing/model': 2.025344480003696, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 1.85}) +[2026-06-23 15:50:46,703][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.11603205651044846, 'train/grad_norm_pre_clip': 27.812255859375, 'timing/data': 0.0013867210072930902, 'timing/model': 2.0295494780002628, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 1.85}) +[2026-06-23 15:50:50,700][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.033218879252672195, 'train/grad_norm_pre_clip': 15.29794979095459, 'timing/data': 0.0014863650139886886, 'timing/model': 2.021027566981502, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 1.85}) +[2026-06-23 15:50:54,765][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.15315783023834229, 'train/grad_norm_pre_clip': 38.74170684814453, 'timing/data': 0.059056445985333994, 'timing/model': 2.0284051049966365, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 1.85}) +[2026-06-23 15:50:58,947][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.01729516126215458, 'train/grad_norm_pre_clip': 9.315665245056152, 'timing/data': 0.001206748973345384, 'timing/model': 2.1232927330129314, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 1.85}) +[2026-06-23 15:51:02,956][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.12675142288208008, 'train/grad_norm_pre_clip': 33.352577209472656, 'timing/data': 0.0013750299985986203, 'timing/model': 2.0345379800128285, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 1.85}) +[2026-06-23 15:51:06,959][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.017895929515361786, 'train/grad_norm_pre_clip': 8.190851211547852, 'timing/data': 0.00136465099058114, 'timing/model': 2.0288856959959958, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 1.85}) +[2026-06-23 15:51:11,039][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.020209139212965965, 'train/grad_norm_pre_clip': 12.172588348388672, 'timing/data': 0.06422417997964658, 'timing/model': 2.0237795319990255, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 1.85}) +[2026-06-23 15:51:15,137][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.06021536514163017, 'train/grad_norm_pre_clip': 27.836978912353516, 'timing/data': 0.0012707659916486591, 'timing/model': 2.0210796190076508, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 1.85}) +[2026-06-23 15:51:19,154][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.1264897584915161, 'train/grad_norm_pre_clip': 37.913448333740234, 'timing/data': 0.0012383169960230589, 'timing/model': 2.0378282429883257, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 1.85}) +[2026-06-23 15:51:23,160][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.08401196449995041, 'train/grad_norm_pre_clip': 17.77056312561035, 'timing/data': 0.001210554997669533, 'timing/model': 2.0238124750030693, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 1.85}) +[2026-06-23 15:51:27,230][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.056588299572467804, 'train/grad_norm_pre_clip': 18.907014846801758, 'timing/data': 0.05547315100557171, 'timing/model': 2.0338124349946156, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 1.85}) +[2026-06-23 15:51:31,334][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.01129329763352871, 'train/grad_norm_pre_clip': 9.12524700164795, 'timing/data': 0.0014079200045671314, 'timing/model': 2.0278889710025396, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 1.85}) +[2026-06-23 15:51:35,532][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.04154516011476517, 'train/grad_norm_pre_clip': 17.684608459472656, 'timing/data': 0.0016780490113887936, 'timing/model': 2.0264543449738994, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 1.86}) +[2026-06-23 15:51:39,556][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.010138878598809242, 'train/grad_norm_pre_clip': 10.5149507522583, 'timing/data': 0.0012490859953686595, 'timing/model': 2.036814910999965, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 1.86}) +[2026-06-23 15:51:43,631][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.03845985606312752, 'train/grad_norm_pre_clip': 16.206804275512695, 'timing/data': 0.057484756980556995, 'timing/model': 2.0333005349966697, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 1.86}) +[2026-06-23 15:51:47,754][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.025044696405529976, 'train/grad_norm_pre_clip': 17.780961990356445, 'timing/data': 0.0014772879949305207, 'timing/model': 2.0445686479797587, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 1.86}) +[2026-06-23 15:51:51,784][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.04330617934465408, 'train/grad_norm_pre_clip': 16.405092239379883, 'timing/data': 0.0014030110032763332, 'timing/model': 2.0429690059972927, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 1.86}) +[2026-06-23 15:51:55,812][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.006867358926683664, 'train/grad_norm_pre_clip': 3.062772512435913, 'timing/data': 0.0012461809965316206, 'timing/model': 2.0450531180249527, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 1.86}) +[2026-06-23 15:51:59,902][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4800 +[2026-06-23 15:53:55,325][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4800 +[2026-06-23 15:53:55,325][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4800 +[2026-06-23 15:56:02,591][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4800 +[2026-06-23 15:56:02,592][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.04068402200937271, 'train/grad_norm_pre_clip': 18.630996704101562, 'eval/action_loss/seconds': 115.41915661099483, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04124416410923004, 'eval/latency_2/loss': 0.04124416410923004, 'eval/flappy/loss': 0.04124416410923004, 'eval/flappy/latency_2/loss': 0.04124416410923004, 'eval/flap_precision': 0.9548172757475083, 'eval/flap_recall': 0.9696356275303644, 'eval/flap_f1': 0.962169400736525, 'eval/flap_rate_model': 0.20902777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9621694007365249, 'eval/latency_2/flap_precision': 0.9548172757475083, 'eval/latency_2/flap_recall': 0.9696356275303644, 'eval/latency_2/flap_f1': 0.962169400736525, 'eval/latency_2/flap_rate_model': 0.20902777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9621694007365249, 'eval/action_classification/data_wait_seconds': 50.49650192260742, 'eval/action_classification/predict_seconds': 75.15345764160156, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.062260167993372306, 'timing/model': 2.0345888669835404, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 1.86}) +[2026-06-23 15:56:06,628][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.03359752520918846, 'train/grad_norm_pre_clip': 13.246562957763672, 'timing/data': 0.0013343040191102773, 'timing/model': 1.9910039420065004, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 1.86}) +[2026-06-23 15:56:10,582][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.10560919344425201, 'train/grad_norm_pre_clip': 34.30878829956055, 'timing/data': 0.0011668259976431727, 'timing/model': 2.0020639279973693, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 1.86}) +[2026-06-23 15:56:14,553][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.11350478231906891, 'train/grad_norm_pre_clip': 45.99108123779297, 'timing/data': 0.0011773540172725916, 'timing/model': 2.0178011739917565, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 1.86}) +[2026-06-23 15:56:18,574][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.027419669553637505, 'train/grad_norm_pre_clip': 9.140563011169434, 'timing/data': 0.06163623501197435, 'timing/model': 2.0010644769936334, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 1.86}) +[2026-06-23 15:56:22,637][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.02252405509352684, 'train/grad_norm_pre_clip': 16.33648681640625, 'timing/data': 0.0013932430010754615, 'timing/model': 2.007554181996966, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 1.86}) +[2026-06-23 15:56:26,599][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.0325680747628212, 'train/grad_norm_pre_clip': 22.0832576751709, 'timing/data': 0.0012572630075737834, 'timing/model': 2.0033999959996436, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 1.86}) +[2026-06-23 15:56:30,821][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.03167968615889549, 'train/grad_norm_pre_clip': 15.46490478515625, 'timing/data': 0.0012001770082861185, 'timing/model': 2.2585227329982445, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 1.86}) +[2026-06-23 15:56:34,861][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.050669021904468536, 'train/grad_norm_pre_clip': 28.55535888671875, 'timing/data': 0.0707554800028447, 'timing/model': 2.0080579849891365, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 1.86}) +[2026-06-23 15:56:38,931][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.04439545050263405, 'train/grad_norm_pre_clip': 17.161855697631836, 'timing/data': 0.0012708670110441744, 'timing/model': 2.0203693239891436, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 1.86}) +[2026-06-23 15:56:42,931][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.10560593008995056, 'train/grad_norm_pre_clip': 27.049270629882812, 'timing/data': 0.0012473840033635497, 'timing/model': 2.026280588004738, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 1.86}) +[2026-06-23 15:56:46,918][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.005970791447907686, 'train/grad_norm_pre_clip': 2.79552960395813, 'timing/data': 0.001294451009016484, 'timing/model': 2.015848085982725, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 1.86}) +[2026-06-23 15:56:50,970][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.008035542443394661, 'train/grad_norm_pre_clip': 9.694901466369629, 'timing/data': 0.0699093479779549, 'timing/model': 2.012960791005753, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 1.86}) +[2026-06-23 15:56:55,044][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.11111771315336227, 'train/grad_norm_pre_clip': 54.05654525756836, 'timing/data': 0.0013641100085806102, 'timing/model': 2.0156876149994787, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 1.86}) +[2026-06-23 15:56:59,042][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.0599883534014225, 'train/grad_norm_pre_clip': 37.60197448730469, 'timing/data': 0.001427987008355558, 'timing/model': 2.019680596975377, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 1.86}) +[2026-06-23 15:57:03,033][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.029766745865345, 'train/grad_norm_pre_clip': 19.53438949584961, 'timing/data': 0.0011528989998623729, 'timing/model': 2.0172910510154907, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 1.86}) +[2026-06-23 15:57:07,258][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.1273401528596878, 'train/grad_norm_pre_clip': 23.715192794799805, 'timing/data': 0.06103745198925026, 'timing/model': 2.0173229679930955, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 1.86}) +[2026-06-23 15:57:11,356][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.059516310691833496, 'train/grad_norm_pre_clip': 33.81559753417969, 'timing/data': 0.0013814319972880185, 'timing/model': 2.036344038002426, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 1.86}) +[2026-06-23 15:57:15,361][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.09132334589958191, 'train/grad_norm_pre_clip': 37.13246154785156, 'timing/data': 0.0013153399922885, 'timing/model': 2.0212336239928845, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 1.86}) +[2026-06-23 15:57:19,371][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.09630180895328522, 'train/grad_norm_pre_clip': 25.54464340209961, 'timing/data': 0.0014324160001706332, 'timing/model': 2.0292564000119455, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 1.87}) +[2026-06-23 15:57:23,438][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.13235123455524445, 'train/grad_norm_pre_clip': 51.30741500854492, 'timing/data': 0.057981438003480434, 'timing/model': 2.0243776160059497, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 1.87}) +[2026-06-23 15:57:27,518][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.10556460916996002, 'train/grad_norm_pre_clip': 42.73297119140625, 'timing/data': 0.001383596012601629, 'timing/model': 2.019075870019151, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 1.87}) +[2026-06-23 15:57:31,547][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.008717931807041168, 'train/grad_norm_pre_clip': 3.017712354660034, 'timing/data': 0.0013203989947214723, 'timing/model': 2.0445930790156126, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 1.87}) +[2026-06-23 15:57:35,554][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.04167670011520386, 'train/grad_norm_pre_clip': 16.243850708007812, 'timing/data': 0.0012131399998907, 'timing/model': 2.026244303007843, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 1.87}) +[2026-06-23 15:57:39,660][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.03622061014175415, 'train/grad_norm_pre_clip': 28.9327335357666, 'timing/data': 0.05713093900703825, 'timing/model': 2.0262729430105537, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 1.87}) +[2026-06-23 15:57:43,755][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.11290841549634933, 'train/grad_norm_pre_clip': 31.124595642089844, 'timing/data': 0.001198454003315419, 'timing/model': 2.0283797419979237, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 1.87}) +[2026-06-23 15:57:47,769][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.12774346768856049, 'train/grad_norm_pre_clip': 35.42010498046875, 'timing/data': 0.001286898012040183, 'timing/model': 2.029546493984526, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 1.87}) +[2026-06-23 15:57:51,781][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.05337146669626236, 'train/grad_norm_pre_clip': 28.628223419189453, 'timing/data': 0.0013348159845918417, 'timing/model': 2.027166022977326, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 1.87}) +[2026-06-23 15:57:55,860][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.04736711457371712, 'train/grad_norm_pre_clip': 20.379684448242188, 'timing/data': 0.06700101000024006, 'timing/model': 2.027935200021602, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 1.87}) +[2026-06-23 15:57:59,968][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.21036966145038605, 'train/grad_norm_pre_clip': 20.693010330200195, 'timing/data': 0.0012256840127520263, 'timing/model': 2.0374424679903314, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 1.87}) +[2026-06-23 15:58:03,977][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.016280608251690865, 'train/grad_norm_pre_clip': 11.906394004821777, 'timing/data': 0.001326319994404912, 'timing/model': 2.027997056982713, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 1.87}) +[2026-06-23 15:58:07,992][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.056443601846694946, 'train/grad_norm_pre_clip': 13.345952987670898, 'timing/data': 0.0014520020049531013, 'timing/model': 2.0281286370009184, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 1.87}) +[2026-06-23 15:58:12,111][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.041189003735780716, 'train/grad_norm_pre_clip': 9.252541542053223, 'timing/data': 0.06928136298665777, 'timing/model': 2.0633373690070584, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 1.87}) +[2026-06-23 15:58:16,209][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.04046407341957092, 'train/grad_norm_pre_clip': 21.244281768798828, 'timing/data': 0.0012131709954701364, 'timing/model': 2.03078896398074, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 1.87}) +[2026-06-23 15:58:20,219][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.1005854532122612, 'train/grad_norm_pre_clip': 43.84062194824219, 'timing/data': 0.001272551016882062, 'timing/model': 2.0249179049860686, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 1.87}) +[2026-06-23 15:58:24,232][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.007465375121682882, 'train/grad_norm_pre_clip': 6.547230243682861, 'timing/data': 0.0013587700086645782, 'timing/model': 2.0291876980045345, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 1.87}) +[2026-06-23 15:58:28,317][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.14173611998558044, 'train/grad_norm_pre_clip': 34.28850555419922, 'timing/data': 0.07685594499344006, 'timing/model': 2.0263185650110245, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 1.87}) +[2026-06-23 15:58:32,434][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.11503615230321884, 'train/grad_norm_pre_clip': 48.749507904052734, 'timing/data': 0.0013746789772994816, 'timing/model': 2.0434755830210634, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 1.87}) +[2026-06-23 15:58:36,449][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.03397371619939804, 'train/grad_norm_pre_clip': 20.276643753051758, 'timing/data': 0.0012769190070685, 'timing/model': 2.0296039450040553, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 1.87}) +[2026-06-23 15:58:40,462][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.033176615834236145, 'train/grad_norm_pre_clip': 12.6632661819458, 'timing/data': 0.0011659329757094383, 'timing/model': 2.0250161059957463, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 1.87}) +[2026-06-23 15:58:44,780][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.07193001359701157, 'train/grad_norm_pre_clip': 56.8133430480957, 'timing/data': 0.2606016270001419, 'timing/model': 2.0765760829963256, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 1.87}) +[2026-06-23 15:58:48,914][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.06827475130558014, 'train/grad_norm_pre_clip': 24.262088775634766, 'timing/data': 0.0011828650021925569, 'timing/model': 2.028478964988608, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 1.87}) +[2026-06-23 15:58:52,935][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.09009746462106705, 'train/grad_norm_pre_clip': 26.889545440673828, 'timing/data': 0.0012490270019043237, 'timing/model': 2.0304259969852865, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 1.87}) +[2026-06-23 15:58:56,965][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.007714603096246719, 'train/grad_norm_pre_clip': 3.7171273231506348, 'timing/data': 0.0012422250001691282, 'timing/model': 2.0426225929986686, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 1.87}) +[2026-06-23 15:59:01,053][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.028355969116091728, 'train/grad_norm_pre_clip': 21.735322952270508, 'timing/data': 0.06933366201701574, 'timing/model': 2.030754725972656, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 1.87}) +[2026-06-23 15:59:05,160][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.031019244343042374, 'train/grad_norm_pre_clip': 19.16837501525879, 'timing/data': 0.001354381995042786, 'timing/model': 2.03100456300308, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 1.88}) +[2026-06-23 15:59:09,181][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.005863895174115896, 'train/grad_norm_pre_clip': 2.3544371128082275, 'timing/data': 0.0012976170110050589, 'timing/model': 2.032562666980084, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 1.88}) +[2026-06-23 15:59:13,200][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.0842704027891159, 'train/grad_norm_pre_clip': 36.34381103515625, 'timing/data': 0.0013495729945134372, 'timing/model': 2.0313948150142096, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 1.88}) +[2026-06-23 15:59:17,292][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.013281255029141903, 'train/grad_norm_pre_clip': 10.017142295837402, 'timing/data': 0.07660201899125241, 'timing/model': 2.0298656160011888, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 1.88}) +[2026-06-23 15:59:21,434][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.02639092318713665, 'train/grad_norm_pre_clip': 16.74290657043457, 'timing/data': 0.0013457259919960052, 'timing/model': 2.07081778900465, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 1.88}) +[2026-06-23 15:59:25,452][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.07896824181079865, 'train/grad_norm_pre_clip': 19.18100929260254, 'timing/data': 0.0012563200143631548, 'timing/model': 2.032259749015793, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 1.88}) +[2026-06-23 15:59:29,468][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.04154334217309952, 'train/grad_norm_pre_clip': 20.655126571655273, 'timing/data': 0.0013456960150506347, 'timing/model': 2.027737952012103, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 1.88}) +[2026-06-23 15:59:33,541][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.04640306904911995, 'train/grad_norm_pre_clip': 17.168460845947266, 'timing/data': 0.05484940798487514, 'timing/model': 2.03052273101639, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 1.88}) +[2026-06-23 15:59:37,646][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.03724034130573273, 'train/grad_norm_pre_clip': 13.906929016113281, 'timing/data': 0.0013690989871975034, 'timing/model': 2.0318975470145233, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 1.88}) +[2026-06-23 15:59:41,663][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.040762852877378464, 'train/grad_norm_pre_clip': 15.357564926147461, 'timing/data': 0.0012240809737704694, 'timing/model': 2.0302069690078497, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 1.88}) +[2026-06-23 15:59:45,693][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.06315550208091736, 'train/grad_norm_pre_clip': 16.61522102355957, 'timing/data': 0.001231453992659226, 'timing/model': 2.0430710870132316, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 1.88}) +[2026-06-23 15:59:49,768][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.01222922746092081, 'train/grad_norm_pre_clip': 5.7093305587768555, 'timing/data': 0.05753862700657919, 'timing/model': 2.0287775990145747, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 1.88}) +[2026-06-23 15:59:53,873][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.07836059480905533, 'train/grad_norm_pre_clip': 21.81230926513672, 'timing/data': 0.001345486001810059, 'timing/model': 2.0317563770222478, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 1.88}) +[2026-06-23 15:59:58,047][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.018178492784500122, 'train/grad_norm_pre_clip': 9.141033172607422, 'timing/data': 0.00121653699898161, 'timing/model': 2.026804415974766, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 1.88}) +[2026-06-23 16:00:02,067][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.03532308712601662, 'train/grad_norm_pre_clip': 19.49680519104004, 'timing/data': 0.0012379670224618167, 'timing/model': 2.0302009450097103, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 1.88}) +[2026-06-23 16:00:06,149][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.13175716996192932, 'train/grad_norm_pre_clip': 32.539642333984375, 'timing/data': 0.05877155499183573, 'timing/model': 2.031800283002667, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 1.88}) +[2026-06-23 16:00:10,251][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.16813994944095612, 'train/grad_norm_pre_clip': 35.451080322265625, 'timing/data': 0.0012163670035079122, 'timing/model': 2.028902777004987, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 1.88}) +[2026-06-23 16:00:14,267][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.0822049006819725, 'train/grad_norm_pre_clip': 42.53434371948242, 'timing/data': 0.001260818011360243, 'timing/model': 2.02747590499348, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 1.88}) +[2026-06-23 16:00:18,283][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.015913521870970726, 'train/grad_norm_pre_clip': 13.104789733886719, 'timing/data': 0.001332261017523706, 'timing/model': 2.030597546021454, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 1.88}) +[2026-06-23 16:00:22,357][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.0924181267619133, 'train/grad_norm_pre_clip': 33.27045440673828, 'timing/data': 0.05463995298487134, 'timing/model': 2.031649885000661, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 1.88}) +[2026-06-23 16:00:26,460][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.002793385647237301, 'train/grad_norm_pre_clip': 3.0961813926696777, 'timing/data': 0.0013515570026356727, 'timing/model': 2.0293268390232697, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 1.88}) +[2026-06-23 16:00:30,741][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.033196330070495605, 'train/grad_norm_pre_clip': 16.28675079345703, 'timing/data': 0.0014410919975489378, 'timing/model': 2.031643554015318, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 1.88}) +[2026-06-23 16:00:34,770][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.013467371463775635, 'train/grad_norm_pre_clip': 7.991348743438721, 'timing/data': 0.0012120990140829235, 'timing/model': 2.0373766989796422, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 1.88}) +[2026-06-23 16:00:38,859][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.13149233162403107, 'train/grad_norm_pre_clip': 30.763391494750977, 'timing/data': 0.06661371199879795, 'timing/model': 2.032638450997183, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 1.88}) +[2026-06-23 16:00:42,965][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.026194997131824493, 'train/grad_norm_pre_clip': 16.10013771057129, 'timing/data': 0.0013771239900961518, 'timing/model': 2.0326022890221793, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 1.88}) +[2026-06-23 16:00:46,985][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.05040666088461876, 'train/grad_norm_pre_clip': 26.902244567871094, 'timing/data': 0.0012624019873328507, 'timing/model': 2.032708479993744, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 1.88}) +[2026-06-23 16:00:51,004][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.0883089229464531, 'train/grad_norm_pre_clip': 26.717544555664062, 'timing/data': 0.0011877040087711066, 'timing/model': 2.031534946989268, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 1.89}) +[2026-06-23 16:00:55,077][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.06960663944482803, 'train/grad_norm_pre_clip': 23.40462875366211, 'timing/data': 0.05773954000324011, 'timing/model': 2.0270728489849716, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 1.89}) +[2026-06-23 16:00:59,198][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.05824761837720871, 'train/grad_norm_pre_clip': 15.001605987548828, 'timing/data': 0.0012197530013509095, 'timing/model': 2.0496904839819763, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 1.89}) +[2026-06-23 16:01:03,464][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.005502123385667801, 'train/grad_norm_pre_clip': 2.5985569953918457, 'timing/data': 0.001683870010310784, 'timing/model': 2.2747385510010645, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 1.89}) +[2026-06-23 16:01:07,482][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.03273700922727585, 'train/grad_norm_pre_clip': 18.15283203125, 'timing/data': 0.001405346003593877, 'timing/model': 2.0312693449959625, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 1.89}) +[2026-06-23 16:01:11,570][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.026326673105359077, 'train/grad_norm_pre_clip': 19.35921859741211, 'timing/data': 0.06793531501898542, 'timing/model': 2.0330265350057743, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 1.89}) +[2026-06-23 16:01:15,674][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.02311677671968937, 'train/grad_norm_pre_clip': 9.211387634277344, 'timing/data': 0.0013475290033966303, 'timing/model': 2.0312787310103886, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 1.89}) +[2026-06-23 16:01:19,695][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.0603184811770916, 'train/grad_norm_pre_clip': 23.810239791870117, 'timing/data': 0.0012408220209181309, 'timing/model': 2.0303555699938443, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 1.89}) +[2026-06-23 16:01:23,714][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.040132664144039154, 'train/grad_norm_pre_clip': 20.28679656982422, 'timing/data': 0.0011692400148604065, 'timing/model': 2.03204589697998, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 1.89}) +[2026-06-23 16:01:27,807][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.04600952938199043, 'train/grad_norm_pre_clip': 22.54429817199707, 'timing/data': 0.0695703660021536, 'timing/model': 2.032779445988126, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 1.89}) +[2026-06-23 16:01:31,914][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.0382227785885334, 'train/grad_norm_pre_clip': 21.079708099365234, 'timing/data': 0.0012314649939071387, 'timing/model': 2.0303100179880857, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 1.89}) +[2026-06-23 16:01:35,933][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.02708578296005726, 'train/grad_norm_pre_clip': 11.514764785766602, 'timing/data': 0.0012479349970817566, 'timing/model': 2.0317772520065773, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 1.89}) +[2026-06-23 16:01:40,053][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.009998196735978127, 'train/grad_norm_pre_clip': 5.0585455894470215, 'timing/data': 0.0012516120041254908, 'timing/model': 2.029983813990839, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 1.89}) +[2026-06-23 16:01:44,126][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.04588555917143822, 'train/grad_norm_pre_clip': 17.031082153320312, 'timing/data': 0.05919643197557889, 'timing/model': 2.028539129998535, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 1.89}) +[2026-06-23 16:01:48,228][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.08872701972723007, 'train/grad_norm_pre_clip': 43.44728088378906, 'timing/data': 0.001167416019598022, 'timing/model': 2.03132127400022, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 1.89}) +[2026-06-23 16:01:52,250][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.049216821789741516, 'train/grad_norm_pre_clip': 17.968507766723633, 'timing/data': 0.0012712289753835648, 'timing/model': 2.0311351339914836, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 1.89}) +[2026-06-23 16:01:56,287][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.07905233651399612, 'train/grad_norm_pre_clip': 34.071380615234375, 'timing/data': 0.001196260011056438, 'timing/model': 2.0351462670078035, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 1.89}) +[2026-06-23 16:02:00,369][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.03571225330233574, 'train/grad_norm_pre_clip': 15.572729110717773, 'timing/data': 0.05696438901941292, 'timing/model': 2.034623706014827, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 1.89}) +[2026-06-23 16:02:04,474][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.08423307538032532, 'train/grad_norm_pre_clip': 18.279306411743164, 'timing/data': 0.0013862409978173673, 'timing/model': 2.0318828170129564, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 1.89}) +[2026-06-23 16:02:08,492][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.044181689620018005, 'train/grad_norm_pre_clip': 16.697799682617188, 'timing/data': 0.0012455210089683533, 'timing/model': 2.0299160769791342, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 1.89}) +[2026-06-23 16:02:12,710][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.01989368349313736, 'train/grad_norm_pre_clip': 15.045616149902344, 'timing/data': 0.0014641839952673763, 'timing/model': 2.227802787005203, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 1.89}) +[2026-06-23 16:02:16,786][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.18337482213974, 'train/grad_norm_pre_clip': 39.77754211425781, 'timing/data': 0.05820716600283049, 'timing/model': 2.0307571690063924, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 1.89}) +[2026-06-23 16:02:20,885][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.026545079424977303, 'train/grad_norm_pre_clip': 15.066866874694824, 'timing/data': 0.0014323659997899085, 'timing/model': 2.029681657004403, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 1.89}) +[2026-06-23 16:02:24,901][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.04282408207654953, 'train/grad_norm_pre_clip': 18.640413284301758, 'timing/data': 0.0014082510024309158, 'timing/model': 2.027525768004125, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 1.89}) +[2026-06-23 16:02:28,921][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.049612678587436676, 'train/grad_norm_pre_clip': 21.801008224487305, 'timing/data': 0.0013171029859222472, 'timing/model': 2.0323820630146656, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 1.89}) +[2026-06-23 16:02:33,015][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.33243444561958313, 'train/grad_norm_pre_clip': 42.956207275390625, 'timing/data': 0.06693971800268628, 'timing/model': 2.0365368610073347, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 1.89}) +[2026-06-23 16:02:37,123][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.023913487792015076, 'train/grad_norm_pre_clip': 27.150856018066406, 'timing/data': 0.0013644399878103286, 'timing/model': 2.032336275005946, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 1.9}) +[2026-06-23 16:02:41,175][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.044751815497875214, 'train/grad_norm_pre_clip': 28.140850067138672, 'timing/data': 0.0012374560174066573, 'timing/model': 2.0590440680098254, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 1.9}) +[2026-06-23 16:02:45,434][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.06175363436341286, 'train/grad_norm_pre_clip': 20.169544219970703, 'timing/data': 0.001344563002930954, 'timing/model': 2.264617348992033, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 1.9}) +[2026-06-23 16:02:49,522][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4900 +[2026-06-23 16:04:43,517][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4900 +[2026-06-23 16:04:43,518][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4900 +[2026-06-23 16:06:50,594][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4900 +[2026-06-23 16:06:50,596][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.01117561012506485, 'train/grad_norm_pre_clip': 4.897163391113281, 'eval/action_loss/seconds': 113.99073289599619, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.040438245981931686, 'eval/latency_2/loss': 0.040438245981931686, 'eval/flappy/loss': 0.040438245981931686, 'eval/flappy/latency_2/loss': 0.040438245981931686, 'eval/flap_precision': 0.9554817275747508, 'eval/flap_recall': 0.970310391363023, 'eval/flap_f1': 0.962838968865082, 'eval/flap_rate_model': 0.20902777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.962838968865082, 'eval/latency_2/flap_precision': 0.9554817275747508, 'eval/latency_2/flap_recall': 0.970310391363023, 'eval/latency_2/flap_f1': 0.962838968865082, 'eval/latency_2/flap_rate_model': 0.20902777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.962838968865082, 'eval/action_classification/data_wait_seconds': 50.41192626953125, 'eval/action_classification/predict_seconds': 75.0072250366211, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.05972840599133633, 'timing/model': 2.0354097439849284, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 1.9}) +[2026-06-23 16:06:54,632][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.018082428723573685, 'train/grad_norm_pre_clip': 12.554930686950684, 'timing/data': 0.001224841980729252, 'timing/model': 2.0066264450142626, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 1.9}) +[2026-06-23 16:06:58,598][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.07248927652835846, 'train/grad_norm_pre_clip': 20.04334259033203, 'timing/data': 0.0012733619951177388, 'timing/model': 2.0142995229980443, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 1.9}) +[2026-06-23 16:07:02,563][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.017315542325377464, 'train/grad_norm_pre_clip': 7.22578239440918, 'timing/data': 0.001344524003798142, 'timing/model': 2.0124678110005334, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 1.9}) +[2026-06-23 16:07:06,600][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.007101172581315041, 'train/grad_norm_pre_clip': 5.00610876083374, 'timing/data': 0.06582237099064514, 'timing/model': 2.0050142940017395, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 1.9}) +[2026-06-23 16:07:10,807][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.05929138883948326, 'train/grad_norm_pre_clip': 29.406843185424805, 'timing/data': 0.0012283789983484894, 'timing/model': 2.0013757739798166, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 1.9}) +[2026-06-23 16:07:14,787][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.1037176251411438, 'train/grad_norm_pre_clip': 31.39260482788086, 'timing/data': 0.0015410869964398444, 'timing/model': 2.0169206639984623, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 1.9}) +[2026-06-23 16:07:18,754][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.03784164413809776, 'train/grad_norm_pre_clip': 21.13266944885254, 'timing/data': 0.0013367799983825535, 'timing/model': 2.004873049998423, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 1.9}) +[2026-06-23 16:07:22,795][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.03903231769800186, 'train/grad_norm_pre_clip': 17.673763275146484, 'timing/data': 0.06162236101226881, 'timing/model': 2.0140498649852816, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 1.9}) +[2026-06-23 16:07:26,854][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.01041866559535265, 'train/grad_norm_pre_clip': 9.267834663391113, 'timing/data': 0.0013465670053847134, 'timing/model': 2.0081432959996164, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 1.9}) +[2026-06-23 16:07:30,830][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.18182505667209625, 'train/grad_norm_pre_clip': 30.607158660888672, 'timing/data': 0.0011422399838920683, 'timing/model': 2.0102926880063023, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 1.9}) +[2026-06-23 16:07:34,819][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.047708991914987564, 'train/grad_norm_pre_clip': 14.63781452178955, 'timing/data': 0.001462963002268225, 'timing/model': 2.01835033702082, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 1.9}) +[2026-06-23 16:07:38,867][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.046450190246105194, 'train/grad_norm_pre_clip': 12.766154289245605, 'timing/data': 0.059647162997862324, 'timing/model': 2.016977521998342, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 1.9}) +[2026-06-23 16:07:43,211][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.05532033368945122, 'train/grad_norm_pre_clip': 29.4006404876709, 'timing/data': 0.001213773008203134, 'timing/model': 2.0311598049884196, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 1.9}) +[2026-06-23 16:07:47,198][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.0837506353855133, 'train/grad_norm_pre_clip': 31.37957191467285, 'timing/data': 0.0012689539871644229, 'timing/model': 2.0143631669925526, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 1.9}) +[2026-06-23 16:07:51,195][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.06166143715381622, 'train/grad_norm_pre_clip': 20.278194427490234, 'timing/data': 0.0014394689933396876, 'timing/model': 2.022958015993936, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 1.9}) +[2026-06-23 16:07:55,238][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.09823545813560486, 'train/grad_norm_pre_clip': 22.417600631713867, 'timing/data': 0.05703358500613831, 'timing/model': 2.0155571560026146, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 1.9}) +[2026-06-23 16:07:59,311][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.015163187868893147, 'train/grad_norm_pre_clip': 17.83629608154297, 'timing/data': 0.0014198829885572195, 'timing/model': 2.016362124995794, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 1.9}) +[2026-06-23 16:08:03,290][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.07931675016880035, 'train/grad_norm_pre_clip': 44.29286193847656, 'timing/data': 0.0012172389833722264, 'timing/model': 2.010888466000324, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 1.9}) +[2026-06-23 16:08:07,280][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.052306462079286575, 'train/grad_norm_pre_clip': 17.61652946472168, 'timing/data': 0.0010934699967037886, 'timing/model': 2.014118874009, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 1.9}) +[2026-06-23 16:08:11,327][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.037563856691122055, 'train/grad_norm_pre_clip': 28.185781478881836, 'timing/data': 0.05966105900006369, 'timing/model': 2.0163280880078673, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 1.9}) +[2026-06-23 16:08:15,659][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.10336177051067352, 'train/grad_norm_pre_clip': 23.688983917236328, 'timing/data': 0.0013873630086891353, 'timing/model': 2.2716130999906454, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 1.9}) +[2026-06-23 16:08:19,657][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.013334359973669052, 'train/grad_norm_pre_clip': 10.180144309997559, 'timing/data': 0.0012507910141721368, 'timing/model': 2.0196303129778244, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 1.9}) +[2026-06-23 16:08:23,662][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.011981122195720673, 'train/grad_norm_pre_clip': 7.01266622543335, 'timing/data': 0.0011799499916378409, 'timing/model': 2.02283931299462, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 1.91}) +[2026-06-23 16:08:27,725][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.06020739674568176, 'train/grad_norm_pre_clip': 16.339082717895508, 'timing/data': 0.06307210100931115, 'timing/model': 2.0240203969879076, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 1.91}) +[2026-06-23 16:08:31,802][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.09265752136707306, 'train/grad_norm_pre_clip': 18.45860481262207, 'timing/data': 0.0012003769807051867, 'timing/model': 2.01886408199789, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 1.91}) +[2026-06-23 16:08:35,797][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.022689610719680786, 'train/grad_norm_pre_clip': 10.523027420043945, 'timing/data': 0.0011489920143503696, 'timing/model': 2.0178978189942427, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 1.91}) +[2026-06-23 16:08:39,798][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.0025070784613490105, 'train/grad_norm_pre_clip': 1.8535246849060059, 'timing/data': 0.0012825090088881552, 'timing/model': 2.0211957739957143, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 1.91}) +[2026-06-23 16:08:43,845][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.013767016120254993, 'train/grad_norm_pre_clip': 12.642602920532227, 'timing/data': 0.051425115001620725, 'timing/model': 2.020452402997762, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 1.91}) +[2026-06-23 16:08:47,931][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.06587279587984085, 'train/grad_norm_pre_clip': 19.61407470703125, 'timing/data': 0.0011703720083460212, 'timing/model': 2.0239265720010735, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 1.91}) +[2026-06-23 16:08:52,060][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.05744883045554161, 'train/grad_norm_pre_clip': 22.61558723449707, 'timing/data': 0.0011682580225169659, 'timing/model': 2.0339605619956274, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 1.91}) +[2026-06-23 16:08:56,057][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.10168415307998657, 'train/grad_norm_pre_clip': 33.484676361083984, 'timing/data': 0.0011117840185761452, 'timing/model': 2.0225236839905847, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 1.91}) +[2026-06-23 16:09:00,113][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.014459055848419666, 'train/grad_norm_pre_clip': 11.759261131286621, 'timing/data': 0.05922244899556972, 'timing/model': 2.0194854729925282, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 1.91}) +[2026-06-23 16:09:04,202][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.06065385788679123, 'train/grad_norm_pre_clip': 29.238174438476562, 'timing/data': 0.0013614049821626395, 'timing/model': 2.024513645010302, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 1.91}) +[2026-06-23 16:09:08,213][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.12160845845937729, 'train/grad_norm_pre_clip': 30.68466567993164, 'timing/data': 0.001371463993564248, 'timing/model': 2.0301120179938152, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 1.91}) +[2026-06-23 16:09:12,216][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.13918443024158478, 'train/grad_norm_pre_clip': 40.112579345703125, 'timing/data': 0.0011347460094839334, 'timing/model': 2.023233932995936, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 1.91}) +[2026-06-23 16:09:16,295][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.007780460175126791, 'train/grad_norm_pre_clip': 5.695115089416504, 'timing/data': 0.060920787014765665, 'timing/model': 2.028969621984288, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 1.91}) +[2026-06-23 16:09:20,422][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.023069145157933235, 'train/grad_norm_pre_clip': 12.783677101135254, 'timing/data': 0.0012800450203940272, 'timing/model': 2.050515226990683, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 1.91}) +[2026-06-23 16:09:24,493][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.016778187826275826, 'train/grad_norm_pre_clip': 8.232784271240234, 'timing/data': 0.001378997985739261, 'timing/model': 2.0665132850117516, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 1.91}) +[2026-06-23 16:09:28,507][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.07303063571453094, 'train/grad_norm_pre_clip': 25.187679290771484, 'timing/data': 0.001387783995596692, 'timing/model': 2.026714380015619, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 1.91}) +[2026-06-23 16:09:32,583][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.021213768050074577, 'train/grad_norm_pre_clip': 8.634793281555176, 'timing/data': 0.05544343899236992, 'timing/model': 2.033785153995268, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 1.91}) +[2026-06-23 16:09:36,707][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.004255456849932671, 'train/grad_norm_pre_clip': 3.213386058807373, 'timing/data': 0.0012600879999808967, 'timing/model': 2.0584032259939704, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 1.91}) +[2026-06-23 16:09:40,723][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.048221565783023834, 'train/grad_norm_pre_clip': 29.226604461669922, 'timing/data': 0.001268263004021719, 'timing/model': 2.030863841995597, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 1.91}) +[2026-06-23 16:09:44,763][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.04586680978536606, 'train/grad_norm_pre_clip': 33.20669937133789, 'timing/data': 0.0013396850263234228, 'timing/model': 2.0489072570053395, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 1.91}) +[2026-06-23 16:09:48,856][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.004465511068701744, 'train/grad_norm_pre_clip': 3.685756206512451, 'timing/data': 0.06634577998192981, 'timing/model': 2.036560798995197, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 1.91}) +[2026-06-23 16:09:52,975][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.04363587871193886, 'train/grad_norm_pre_clip': 14.521069526672363, 'timing/data': 0.001308386999880895, 'timing/model': 2.039496223005699, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 1.91}) +[2026-06-23 16:09:57,008][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.03134383261203766, 'train/grad_norm_pre_clip': 24.977924346923828, 'timing/data': 0.001452383992727846, 'timing/model': 2.043839719990501, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 1.91}) +[2026-06-23 16:10:01,270][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.015351034700870514, 'train/grad_norm_pre_clip': 14.172185897827148, 'timing/data': 0.001192513998830691, 'timing/model': 2.0294056150014512, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 1.91}) +[2026-06-23 16:10:05,351][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.06627283245325089, 'train/grad_norm_pre_clip': 19.84754180908203, 'timing/data': 0.06830135299242102, 'timing/model': 2.028854017989943, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 1.91}) +[2026-06-23 16:10:09,452][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.05127982795238495, 'train/grad_norm_pre_clip': 20.603137969970703, 'timing/data': 0.0012302830000407994, 'timing/model': 2.0292981500097085, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 1.92}) +[2026-06-23 16:10:13,473][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.06562244147062302, 'train/grad_norm_pre_clip': 25.393360137939453, 'timing/data': 0.0014323759824037552, 'timing/model': 2.0297555869910866, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 1.92}) +[2026-06-23 16:10:17,493][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.05520273745059967, 'train/grad_norm_pre_clip': 38.62760543823242, 'timing/data': 0.0013419099850580096, 'timing/model': 2.0335383570054546, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 1.92}) +[2026-06-23 16:10:21,576][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.029743004590272903, 'train/grad_norm_pre_clip': 11.67795467376709, 'timing/data': 0.06708559699472971, 'timing/model': 2.029533419001382, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 1.92}) +[2026-06-23 16:10:25,677][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.052344389259815216, 'train/grad_norm_pre_clip': 18.616405487060547, 'timing/data': 0.0012837419926654547, 'timing/model': 2.032149834994925, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 1.92}) +[2026-06-23 16:10:29,696][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.060646601021289825, 'train/grad_norm_pre_clip': 20.34547996520996, 'timing/data': 0.0012941610184498131, 'timing/model': 2.0300013569940347, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 1.92}) +[2026-06-23 16:10:33,778][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.12545502185821533, 'train/grad_norm_pre_clip': 26.834794998168945, 'timing/data': 0.001106434006942436, 'timing/model': 2.027954737015534, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 1.92}) +[2026-06-23 16:10:37,864][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.027324369177222252, 'train/grad_norm_pre_clip': 21.998310089111328, 'timing/data': 0.06983001797925681, 'timing/model': 2.0303723750112113, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 1.92}) +[2026-06-23 16:10:41,978][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.009384353645145893, 'train/grad_norm_pre_clip': 5.485566139221191, 'timing/data': 0.0012399500119499862, 'timing/model': 2.041766626003664, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 1.92}) +[2026-06-23 16:10:46,002][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.15385475754737854, 'train/grad_norm_pre_clip': 51.91819381713867, 'timing/data': 0.0014653169782832265, 'timing/model': 2.0328865030023735, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 1.92}) +[2026-06-23 16:10:50,025][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.08028660714626312, 'train/grad_norm_pre_clip': 27.861053466796875, 'timing/data': 0.0012074100086465478, 'timing/model': 2.032671004999429, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 1.92}) +[2026-06-23 16:10:54,096][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.20105546712875366, 'train/grad_norm_pre_clip': 26.50295639038086, 'timing/data': 0.053478339017601684, 'timing/model': 2.025404233019799, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 1.92}) +[2026-06-23 16:10:58,226][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.0027554105035960674, 'train/grad_norm_pre_clip': 2.2214622497558594, 'timing/data': 0.001408893003826961, 'timing/model': 2.053806034993613, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 1.92}) +[2026-06-23 16:11:02,249][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.07125537842512131, 'train/grad_norm_pre_clip': 28.94729232788086, 'timing/data': 0.0013214909995440394, 'timing/model': 2.035402315988904, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 1.92}) +[2026-06-23 16:11:06,455][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.07875761389732361, 'train/grad_norm_pre_clip': 33.182373046875, 'timing/data': 0.0014533750072587281, 'timing/model': 2.1926114249799866, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 1.92}) +[2026-06-23 16:11:10,537][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.07133470475673676, 'train/grad_norm_pre_clip': 31.66217803955078, 'timing/data': 0.06348651900771074, 'timing/model': 2.0282897890137974, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 1.92}) +[2026-06-23 16:11:14,645][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.009904865175485611, 'train/grad_norm_pre_clip': 5.830843925476074, 'timing/data': 0.0012065389892086387, 'timing/model': 2.0316518819890916, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 1.92}) +[2026-06-23 16:11:18,661][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.08618782460689545, 'train/grad_norm_pre_clip': 20.00748062133789, 'timing/data': 0.0012652370205614716, 'timing/model': 2.0305593440134544, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 1.92}) +[2026-06-23 16:11:22,675][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.015366376377642155, 'train/grad_norm_pre_clip': 13.091195106506348, 'timing/data': 0.0012455719988793135, 'timing/model': 2.030043252016185, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 1.92}) +[2026-06-23 16:11:26,768][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.02022465318441391, 'train/grad_norm_pre_clip': 9.114161491394043, 'timing/data': 0.06434326100861654, 'timing/model': 2.034098094998626, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 1.92}) +[2026-06-23 16:11:30,872][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.07440066337585449, 'train/grad_norm_pre_clip': 27.302431106567383, 'timing/data': 0.0012674010067712516, 'timing/model': 2.0331324620055966, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 1.92}) +[2026-06-23 16:11:34,908][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.02080025151371956, 'train/grad_norm_pre_clip': 16.755373001098633, 'timing/data': 0.0014160670107230544, 'timing/model': 2.0482438009930775, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 1.92}) +[2026-06-23 16:11:38,930][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.027430061250925064, 'train/grad_norm_pre_clip': 19.624549865722656, 'timing/data': 0.0014042639813851565, 'timing/model': 2.034037826990243, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 1.92}) +[2026-06-23 16:11:43,050][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.042308881878852844, 'train/grad_norm_pre_clip': 20.330785751342773, 'timing/data': 0.05947830399964005, 'timing/model': 2.0337871859956067, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 1.92}) +[2026-06-23 16:11:47,159][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.041630834341049194, 'train/grad_norm_pre_clip': 21.32181167602539, 'timing/data': 0.001416747021721676, 'timing/model': 2.033186777989613, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 1.92}) +[2026-06-23 16:11:51,204][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.036599528044462204, 'train/grad_norm_pre_clip': 15.773900032043457, 'timing/data': 0.0015270809817593545, 'timing/model': 2.053419128002133, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 1.93}) +[2026-06-23 16:11:55,246][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.07915277034044266, 'train/grad_norm_pre_clip': 22.805822372436523, 'timing/data': 0.0012085829803254455, 'timing/model': 2.0536372610076796, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 1.93}) +[2026-06-23 16:11:59,331][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.03762124851346016, 'train/grad_norm_pre_clip': 21.53556251525879, 'timing/data': 0.055133357003796846, 'timing/model': 2.0377114500151947, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 1.93}) +[2026-06-23 16:12:03,437][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.013911616988480091, 'train/grad_norm_pre_clip': 8.989684104919434, 'timing/data': 0.0013486510142683983, 'timing/model': 2.0339515160012525, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 1.93}) +[2026-06-23 16:12:07,465][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.014052963815629482, 'train/grad_norm_pre_clip': 15.073583602905273, 'timing/data': 0.0012473140086513013, 'timing/model': 2.03245526298997, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 1.93}) +[2026-06-23 16:12:11,485][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.0249251089990139, 'train/grad_norm_pre_clip': 13.689746856689453, 'timing/data': 0.0012146939989179373, 'timing/model': 2.034902063984191, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 1.93}) +[2026-06-23 16:12:15,726][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.018975578248500824, 'train/grad_norm_pre_clip': 14.645069122314453, 'timing/data': 0.05675170698668808, 'timing/model': 2.191534659010358, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 1.93}) +[2026-06-23 16:12:19,868][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.13248112797737122, 'train/grad_norm_pre_clip': 22.183637619018555, 'timing/data': 0.0016368479991797358, 'timing/model': 2.0699741230055224, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 1.93}) +[2026-06-23 16:12:23,901][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.02422056533396244, 'train/grad_norm_pre_clip': 14.325830459594727, 'timing/data': 0.0012353250058367848, 'timing/model': 2.0373930420028046, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 1.93}) +[2026-06-23 16:12:27,943][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.10278232395648956, 'train/grad_norm_pre_clip': 44.548912048339844, 'timing/data': 0.0012481889862101525, 'timing/model': 2.050816880975617, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 1.93}) +[2026-06-23 16:12:32,062][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.0602373443543911, 'train/grad_norm_pre_clip': 40.06714630126953, 'timing/data': 0.06669778097420931, 'timing/model': 2.055160435993457, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 1.93}) +[2026-06-23 16:12:36,174][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.024075184017419815, 'train/grad_norm_pre_clip': 19.257837295532227, 'timing/data': 0.0013670389889739454, 'timing/model': 2.0420661179814488, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 1.93}) +[2026-06-23 16:12:40,218][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.03533916175365448, 'train/grad_norm_pre_clip': 25.835037231445312, 'timing/data': 0.001433040015399456, 'timing/model': 2.045944582001539, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 1.93}) +[2026-06-23 16:12:44,240][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.018005887046456337, 'train/grad_norm_pre_clip': 11.89537239074707, 'timing/data': 0.00113109199446626, 'timing/model': 2.0329529570008162, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 1.93}) +[2026-06-23 16:12:48,322][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.0251786969602108, 'train/grad_norm_pre_clip': 21.744234085083008, 'timing/data': 0.05714311701012775, 'timing/model': 2.0292944829852786, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 1.93}) +[2026-06-23 16:12:52,683][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.03343126177787781, 'train/grad_norm_pre_clip': 13.740249633789062, 'timing/data': 0.0011919550015591085, 'timing/model': 2.034578054997837, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 1.93}) +[2026-06-23 16:12:56,712][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.09601236879825592, 'train/grad_norm_pre_clip': 25.603591918945312, 'timing/data': 0.0013382949982769787, 'timing/model': 2.0385832939937245, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 1.93}) +[2026-06-23 16:13:00,735][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.17179185152053833, 'train/grad_norm_pre_clip': 26.94196319580078, 'timing/data': 0.0011998990084975958, 'timing/model': 2.0317455519980285, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 1.93}) +[2026-06-23 16:13:04,820][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.09611858427524567, 'train/grad_norm_pre_clip': 29.03452491760254, 'timing/data': 0.0541912539920304, 'timing/model': 2.03787557099713, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 1.93}) +[2026-06-23 16:13:08,926][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.16883882880210876, 'train/grad_norm_pre_clip': 13.333142280578613, 'timing/data': 0.0011816949991043657, 'timing/model': 2.0319058340101037, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 1.93}) +[2026-06-23 16:13:12,949][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.07969648391008377, 'train/grad_norm_pre_clip': 22.326093673706055, 'timing/data': 0.0011888779990840703, 'timing/model': 2.0358575859863777, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 1.93}) +[2026-06-23 16:13:16,998][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.012552688829600811, 'train/grad_norm_pre_clip': 9.323102951049805, 'timing/data': 0.001303641009144485, 'timing/model': 2.054147975984961, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 1.93}) +[2026-06-23 16:13:21,080][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.022460108622908592, 'train/grad_norm_pre_clip': 13.80708122253418, 'timing/data': 0.05519589799223468, 'timing/model': 2.0326979199890047, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 1.93}) +[2026-06-23 16:13:25,252][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.012577175162732601, 'train/grad_norm_pre_clip': 10.090890884399414, 'timing/data': 0.0011664970079436898, 'timing/model': 2.0398365349974483, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 1.93}) +[2026-06-23 16:13:29,284][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.023455273360013962, 'train/grad_norm_pre_clip': 10.411578178405762, 'timing/data': 0.001230705005582422, 'timing/model': 2.036338510020869, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 1.93}) +[2026-06-23 16:13:33,322][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.03640856593847275, 'train/grad_norm_pre_clip': 32.38024139404297, 'timing/data': 0.0012263579992577434, 'timing/model': 2.0358009510091506, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 1.93}) +[2026-06-23 16:13:37,403][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 5000 +[2026-06-23 16:15:32,146][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 5000 +[2026-06-23 16:15:32,147][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 5000 +[2026-06-23 16:17:39,751][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 5000 +[2026-06-23 16:36:02,135][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 5000 +[2026-06-23 16:36:02,137][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.047863367944955826, 'train/grad_norm_pre_clip': 22.7413387298584, 'eval/action_loss/seconds': 114.73954293402494, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04150378704071045, 'eval/latency_2/loss': 0.04150378704071045, 'eval/flappy/loss': 0.04150378704071045, 'eval/flappy/latency_2/loss': 0.04150378704071045, 'eval/flap_precision': 0.958, 'eval/flap_recall': 0.9696356275303644, 'eval/flap_f1': 0.9637826961770624, 'eval/flap_rate_model': 0.20833333333333334, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9637826961770625, 'eval/latency_2/flap_precision': 0.958, 'eval/latency_2/flap_recall': 0.9696356275303644, 'eval/latency_2/flap_f1': 0.9637826961770624, 'eval/latency_2/flap_rate_model': 0.20833333333333334, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9637826961770625, 'eval/action_classification/data_wait_seconds': 50.753238677978516, 'eval/action_classification/predict_seconds': 75.1337661743164, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 285.0200000000155, 'rl_games_eval/mid_train/mean_length': 2314.6, 'rl_games_eval/mid_train/std_reward': 195.5323236705503, 'rl_games_eval/mid_train/std_length': 1574.9920126781594, 'rl_games_eval/mid_train/macro_mean_reward': 285.0200000000155, 'rl_games_eval/mid_train/macro_mean_length': 2314.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 285.0200000000155, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2314.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 195.5323236705503, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1574.9920126781594, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 285.0200000000155, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.05480774899479002, 'timing/model': 2.0357736849982757, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 1.94}) +[2026-06-23 16:36:02,137][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_5000_state +[2026-06-23 16:36:02,143][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 16:36:10,908][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_5000_state/model.safetensors +[2026-06-23 16:36:14,876][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_5000_state/optimizer.bin +[2026-06-23 16:36:14,876][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_5000_state/sampler.bin +[2026-06-23 16:36:14,876][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-23 16:36:14,878][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-23 16:36:15,328][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-23 16:36:15,328][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 16:36:15,352][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 16:36:15,352][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers/checkpoints +[2026-06-23 16:36:17,687][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..a0e21985e6b7ed5b88fddcc5f74f6ec45063b603 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers/summary.jsonl @@ -0,0 +1,10 @@ +{"steps": 500} +{"steps": 1000} +{"steps": 1500} +{"steps": 2000} +{"steps": 2500} +{"steps": 3000} +{"steps": 3500} +{"steps": 4000} +{"steps": 4500} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6f954fb25d53857377222246859f7a8449fc37fa --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:750786e2eac8b455ab6e8232416d692e8ad1d5b7a56d386b40bf53a9462ba436 +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..42ef7754481f8068af289c33bc0b716dffb428ed --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/config.full.yaml @@ -0,0 +1,219 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 64 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_llm_bottom_ratio: 0.7778 + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_last_8_layers_corrected +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..1f58ebcc65a35fba4bc588d6e94321e7695de0cc --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/config.yaml @@ -0,0 +1,99 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + eval_num_workers: 8 + include_state: true + latency_curriculum: + enabled: false + obs_image_size: null + per_device_batch_size: 64 + persistent_workers: true + pin_memory: true + prefetch_factor: 4 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected +rl_games: + env_eval: + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 1 + image_size: 224 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: '' + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: flappy +run_id: flappy_fix_latency_2_200ep_last_8_layers_corrected +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_action_classification_interval: null + eval_interval: 100 + eval_num_batches: 100 + freeze_llm_bottom_ratio: 0.7778 + freeze_modules: '' + gradient_accumulation_steps: 2 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + reload_modules: null + save_interval: 500 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..aa03e6cbf8e375d660528e17fafddb4804b596b7 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.6028500199317932, + 0.3971499800682068, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4890792667865753, + 0.4890792667865753, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 330734, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..0caa254376d5805f972b0b0411e0b5188e3634e3 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7959861159324646, + 0.2040138840675354, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4030573070049286, + 0.4030573070049286, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..cc6f0e5b8ed6a2cebedeaadf94fee0675a31841a --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra/.hydra/config.yaml @@ -0,0 +1,217 @@ +framework: + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 64 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: ${dataset.single_source_hf} + config_name: null + source_subdir: null + converted_name: ${dataset.single_converted_name} + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: ${dataset.single_latency_filter} + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: ${bridge_base_model.repo_id.${model}} +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_llm_bottom_ratio: 0.7778 + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: playground/Datasets/rl_games + dataset_cache_dir: null + base_model_dir: ${bridge_base_model.local_dir.${model}} + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: auto + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_last_8_layers_corrected +output_dir: null +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..fd2e042894fcd5b97209a8ac444e1b8aed993adf --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra/.hydra/hydra.yaml @@ -0,0 +1,266 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - run_id=flappy_fix_latency_2_200ep_last_8_layers_corrected + - run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints + - seed=42 + - wandb_entity=saberrr-zju + - wandb_project=starVLA_rl_games + - rl_games.env_eval.enabled=true + - checkpoint.sync.enabled=false + - checkpoint.sync.keep_last_n=0 + - checkpoint.local.keep_last_n=1 + - checkpoint.save_best_model=false + - checkpoint.save_pt_file=false + - trainer.is_resume=false + - trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - trainer.resume_step=0 + - trainer.max_train_steps=5000 + - trainer.num_warmup_steps=100 + - trainer.save_interval=500 + - trainer.eval_interval=100 + - trainer.eval_num_batches=100 + - trainer.eval_action_classification=true + - trainer.logging_frequency=1 + - trainer.gradient_accumulation_steps=2 + - trainer.distributed_backend=none + - trainer.learning_rate.base=2e-05 + - trainer.learning_rate.qwen_vl_interface=1e-05 + - trainer.learning_rate.action_model=0.0001 + - trainer.lr_scheduler_type=cosine_with_min_lr + - trainer.scheduler_specific_kwargs.min_lr=1e-06 + - trainer.freeze_llm_bottom_ratio=0.7778 + - trainer.loss_scale.vla=1.0 + - trainer.loss_scale.vlm=0.1 + - trainer.max_grad_norm=1.0 + - trainer.weight_decay=0.0 + - trainer.gradient_clipping=1.0 + - trainer.optimizer.name=AdamW + - trainer.optimizer.betas=[0.9,0.95] + - trainer.optimizer.eps=1e-08 + - trainer.optimizer.weight_decay=1e-08 + - trainer.optimizer.fused=true + - trainer.save_format=pt + - framework.name=QwenOFT + - framework.qwenvl.attn_implementation=flash_attention_2 + - framework.qwenvl.enable_gradient_checkpointing=true + - framework.action_model.action_dim=7 + - framework.action_model.action_env_dim=2 + - framework.action_model.state_dim=7 + - framework.action_model.loss_type=discrete_ce + - framework.action_model.action_horizon=1 + - framework.action_model.future_action_window_size=0 + - framework.action_model.past_action_window_size=0 + - datasets.vla_data.include_state=true + - datasets.vla_data.action_type=discrete + - datasets.vla_data.sequential_step_sampling=false + - datasets.vla_data.shuffle=true + - datasets.vla_data.num_workers=8 + - datasets.vla_data.eval_num_workers=8 + - datasets.vla_data.prefetch_factor=4 + - datasets.vla_data.persistent_workers=true + - datasets.vla_data.pin_memory=true + - datasets.vla_data.action_balance.enabled=false + - datasets.vla_data.action_balance.strategy=balanced_epoch + - datasets.vla_data.action_balance.action_key=action_id + - datasets.vla_data.action_balance.target_flap_fraction=0.3 + - datasets.vla_data.action_balance.noop_id=0 + - datasets.vla_data.action_balance.flap_id=1 + - datasets.vla_data.latency_curriculum.enabled=false + - datasets.vla_data.latency_curriculum.strategy=exclusive + - datasets.vla_data.per_device_batch_size=64 + - datasets.vla_data.num_workers=8 + - datasets.vla_data.eval_num_workers=8 + - datasets.vla_data.prefetch_factor=4 + - datasets.vla_data.persistent_workers=true + - datasets.vla_data.pin_memory=true + - datasets.vla_data.load_all_data_for_training=true + - datasets.vla_data.video_backend=torchvision_av + - datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + - datasets.vla_data.data_mix=flappy_train__bridge + - datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - rl_games.task=flappy + - rl_games.model_alias=openvla + - rl_games.initialization_mode=bridge + - rl_games.action_carrier=bridge + - rl_games.env_eval.distributed_mode=none + - rl_games.env_eval.latency.mode=single + - rl_games.env_eval.frameskip=1 + - rl_games.env_eval.image_size=224 + - rl_games.env_eval.seed=42 + - rl_games.env_eval.fixed_episode_seeds=true + - rl_games.env_eval.latency_seed_stride=0 + - rl_games.env_eval.task_seed_stride=0 + - rl_games.env_eval.latency.values=[0] + - rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + - rl_games.env_eval.mid_train.enabled=true + - rl_games.env_eval.mid_train.interval_steps=250 + - rl_games.env_eval.mid_train.num_episodes=5 + - rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - rl_games.env_eval.mid_train.latencies=[2] + - rl_games.env_eval.post_train.enabled=false + - rl_games.env_eval.post_train.num_episodes=5 + - rl_games.env_eval.post_train.max_steps_per_episode=3600 + - rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + job: + name: train_starvla_hydra + chdir: false + override_dirname: checkpoint.local.keep_last_n=1,checkpoint.save_best_model=false,checkpoint.save_pt_file=false,checkpoint.sync.enabled=false,checkpoint.sync.keep_last_n=0,datasets.vla_data.action_balance.action_key=action_id,datasets.vla_data.action_balance.enabled=false,datasets.vla_data.action_balance.flap_id=1,datasets.vla_data.action_balance.noop_id=0,datasets.vla_data.action_balance.strategy=balanced_epoch,datasets.vla_data.action_balance.target_flap_fraction=0.3,datasets.vla_data.action_type=discrete,datasets.vla_data.data_mix=flappy_train__bridge,datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep,datasets.vla_data.eval_data_mix=flappy_train__bridge__val,datasets.vla_data.eval_num_workers=8,datasets.vla_data.eval_num_workers=8,datasets.vla_data.include_state=true,datasets.vla_data.latency_curriculum.enabled=false,datasets.vla_data.latency_curriculum.strategy=exclusive,datasets.vla_data.load_all_data_for_training=true,datasets.vla_data.num_workers=8,datasets.vla_data.num_workers=8,datasets.vla_data.per_device_batch_size=64,datasets.vla_data.persistent_workers=true,datasets.vla_data.persistent_workers=true,datasets.vla_data.pin_memory=true,datasets.vla_data.pin_memory=true,datasets.vla_data.prefetch_factor=4,datasets.vla_data.prefetch_factor=4,datasets.vla_data.sequential_step_sampling=false,datasets.vla_data.shuffle=true,datasets.vla_data.video_backend=torchvision_av,env=flappy,framework.action_model.action_dim=7,framework.action_model.action_env_dim=2,framework.action_model.action_horizon=1,framework.action_model.future_action_window_size=0,framework.action_model.loss_type=discrete_ce,framework.action_model.past_action_window_size=0,framework.action_model.state_dim=7,framework.name=QwenOFT,framework.qwenvl.attn_implementation=flash_attention_2,framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,framework.qwenvl.enable_gradient_checkpointing=true,init=bridge,mode=single,model=openvla,rl_games.action_carrier=bridge,rl_games.env_eval.distributed_mode=none,rl_games.env_eval.enabled=true,rl_games.env_eval.fixed_episode_seeds=true,rl_games.env_eval.frameskip=1,rl_games.env_eval.image_size=224,rl_games.env_eval.latency.mode=single,rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json,rl_games.env_eval.latency.values=[0],rl_games.env_eval.latency_seed_stride=0,rl_games.env_eval.mid_train.enabled=true,rl_games.env_eval.mid_train.interval_steps=250,rl_games.env_eval.mid_train.latencies=[2],rl_games.env_eval.mid_train.max_steps_per_episode=3600,rl_games.env_eval.mid_train.num_episodes=5,rl_games.env_eval.post_train.enabled=false,rl_games.env_eval.post_train.latencies=[0,1,2,3,4],rl_games.env_eval.post_train.max_steps_per_episode=3600,rl_games.env_eval.post_train.num_episodes=5,rl_games.env_eval.seed=42,rl_games.env_eval.task_seed_stride=0,rl_games.initialization_mode=bridge,rl_games.model_alias=openvla,rl_games.task=flappy,run_id=flappy_fix_latency_2_200ep_last_8_layers_corrected,run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints,seed=42,trainer.distributed_backend=none,trainer.eval_action_classification=true,trainer.eval_interval=100,trainer.eval_num_batches=100,trainer.freeze_llm_bottom_ratio=0.7778,trainer.gradient_accumulation_steps=2,trainer.gradient_clipping=1.0,trainer.is_resume=false,trainer.learning_rate.action_model=0.0001,trainer.learning_rate.base=2e-05,trainer.learning_rate.qwen_vl_interface=1e-05,trainer.logging_frequency=1,trainer.loss_scale.vla=1.0,trainer.loss_scale.vlm=0.1,trainer.lr_scheduler_type=cosine_with_min_lr,trainer.max_grad_norm=1.0,trainer.max_train_steps=5000,trainer.num_warmup_steps=100,trainer.optimizer.betas=[0.9,0.95],trainer.optimizer.eps=1e-08,trainer.optimizer.fused=true,trainer.optimizer.name=AdamW,trainer.optimizer.weight_decay=1e-08,trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,trainer.resume_step=0,trainer.save_format=pt,trainer.save_interval=500,trainer.scheduler_specific_kwargs.min_lr=1e-06,trainer.weight_decay=0.0,wandb_entity=saberrr-zju,wandb_project=starVLA_rl_games + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/latency-sensitive-bench/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/latency-sensitive-bench/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..0ad9abf9609cd6299980f209bcac8c8a719e905e --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra/.hydra/overrides.yaml @@ -0,0 +1,106 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- run_id=flappy_fix_latency_2_200ep_last_8_layers_corrected +- run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints +- seed=42 +- wandb_entity=saberrr-zju +- wandb_project=starVLA_rl_games +- rl_games.env_eval.enabled=true +- checkpoint.sync.enabled=false +- checkpoint.sync.keep_last_n=0 +- checkpoint.local.keep_last_n=1 +- checkpoint.save_best_model=false +- checkpoint.save_pt_file=false +- trainer.is_resume=false +- trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- trainer.resume_step=0 +- trainer.max_train_steps=5000 +- trainer.num_warmup_steps=100 +- trainer.save_interval=500 +- trainer.eval_interval=100 +- trainer.eval_num_batches=100 +- trainer.eval_action_classification=true +- trainer.logging_frequency=1 +- trainer.gradient_accumulation_steps=2 +- trainer.distributed_backend=none +- trainer.learning_rate.base=2e-05 +- trainer.learning_rate.qwen_vl_interface=1e-05 +- trainer.learning_rate.action_model=0.0001 +- trainer.lr_scheduler_type=cosine_with_min_lr +- trainer.scheduler_specific_kwargs.min_lr=1e-06 +- trainer.freeze_llm_bottom_ratio=0.7778 +- trainer.loss_scale.vla=1.0 +- trainer.loss_scale.vlm=0.1 +- trainer.max_grad_norm=1.0 +- trainer.weight_decay=0.0 +- trainer.gradient_clipping=1.0 +- trainer.optimizer.name=AdamW +- trainer.optimizer.betas=[0.9,0.95] +- trainer.optimizer.eps=1e-08 +- trainer.optimizer.weight_decay=1e-08 +- trainer.optimizer.fused=true +- trainer.save_format=pt +- framework.name=QwenOFT +- framework.qwenvl.attn_implementation=flash_attention_2 +- framework.qwenvl.enable_gradient_checkpointing=true +- framework.action_model.action_dim=7 +- framework.action_model.action_env_dim=2 +- framework.action_model.state_dim=7 +- framework.action_model.loss_type=discrete_ce +- framework.action_model.action_horizon=1 +- framework.action_model.future_action_window_size=0 +- framework.action_model.past_action_window_size=0 +- datasets.vla_data.include_state=true +- datasets.vla_data.action_type=discrete +- datasets.vla_data.sequential_step_sampling=false +- datasets.vla_data.shuffle=true +- datasets.vla_data.num_workers=8 +- datasets.vla_data.eval_num_workers=8 +- datasets.vla_data.prefetch_factor=4 +- datasets.vla_data.persistent_workers=true +- datasets.vla_data.pin_memory=true +- datasets.vla_data.action_balance.enabled=false +- datasets.vla_data.action_balance.strategy=balanced_epoch +- datasets.vla_data.action_balance.action_key=action_id +- datasets.vla_data.action_balance.target_flap_fraction=0.3 +- datasets.vla_data.action_balance.noop_id=0 +- datasets.vla_data.action_balance.flap_id=1 +- datasets.vla_data.latency_curriculum.enabled=false +- datasets.vla_data.latency_curriculum.strategy=exclusive +- datasets.vla_data.per_device_batch_size=64 +- datasets.vla_data.num_workers=8 +- datasets.vla_data.eval_num_workers=8 +- datasets.vla_data.prefetch_factor=4 +- datasets.vla_data.persistent_workers=true +- datasets.vla_data.pin_memory=true +- datasets.vla_data.load_all_data_for_training=true +- datasets.vla_data.video_backend=torchvision_av +- datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep +- datasets.vla_data.data_mix=flappy_train__bridge +- datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- rl_games.task=flappy +- rl_games.model_alias=openvla +- rl_games.initialization_mode=bridge +- rl_games.action_carrier=bridge +- rl_games.env_eval.distributed_mode=none +- rl_games.env_eval.latency.mode=single +- rl_games.env_eval.frameskip=1 +- rl_games.env_eval.image_size=224 +- rl_games.env_eval.seed=42 +- rl_games.env_eval.fixed_episode_seeds=true +- rl_games.env_eval.latency_seed_stride=0 +- rl_games.env_eval.task_seed_stride=0 +- rl_games.env_eval.latency.values=[0] +- rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json +- rl_games.env_eval.mid_train.enabled=true +- rl_games.env_eval.mid_train.interval_steps=250 +- rl_games.env_eval.mid_train.num_episodes=5 +- rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- rl_games.env_eval.mid_train.latencies=[2] +- rl_games.env_eval.post_train.enabled=false +- rl_games.env_eval.post_train.num_episodes=5 +- rl_games.env_eval.post_train.max_steps_per_episode=3600 +- rl_games.env_eval.post_train.latencies=[0,1,2,3,4] diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..7a2e4e670503a2c54b2e3832b5fdb60f2a4ca60c --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/hydra/train_starvla_hydra.log @@ -0,0 +1,5375 @@ +[2026-06-23 10:13:11,247][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-23 10:13:11,945][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured. +[2026-06-23 10:13:11,946][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub +[2026-06-23 10:13:12,475][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases +[2026-06-23 10:13:12,504][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-23 10:13:17,826][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-23 10:13:17,851][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-23 10:13:18,670][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-23 10:13:18,671][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-23 10:13:18,671][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-23 10:13:18,671][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-23 10:13:18,671][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-23 10:13:18,671][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-23 10:13:18,671][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-23 10:13:18,671][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-23 10:13:18,671][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-23 10:13:18,671][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-23 10:13:18,672][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-23 10:13:18,672][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-23 10:13:18,672][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-23 10:13:18,672][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-23 10:13:18,672][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-23 10:13:18,672][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-23 10:13:18,672][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-23 10:13:18,672][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-23 10:13:18,672][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-23 10:13:18,700][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-23 10:13:18,722][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/config.full.yaml +[2026-06-23 10:13:18,728][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/config.yaml +[2026-06-23 10:13:18,728][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-23 10:13:21,613][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-23 10:13:21,613][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-23 10:13:21,613][starVLA.training.train_starvla][INFO] - Per device batch size = 64 +[2026-06-23 10:13:21,613][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 2 +[2026-06-23 10:13:21,613][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-23 10:13:58,204][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6946540474891663, 'train/grad_norm_pre_clip': 8.95699691772461, 'timing/data': 0.001659295056015253, 'timing/model': 2.0967132069636136, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-23 10:14:01,728][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6940727829933167, 'train/grad_norm_pre_clip': 10.540936470031738, 'timing/data': 0.001603872049599886, 'timing/model': 1.8258255789987743, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-23 10:14:05,590][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6895707249641418, 'train/grad_norm_pre_clip': 19.44685173034668, 'timing/data': 0.0031565360259264708, 'timing/model': 1.8645793619798496, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-23 10:14:09,790][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.6879895329475403, 'train/grad_norm_pre_clip': 10.032864570617676, 'timing/data': 0.0028859099838882685, 'timing/model': 1.9931411100551486, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 10:14:13,408][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6894322633743286, 'train/grad_norm_pre_clip': 2.311058521270752, 'timing/data': 0.002734459936618805, 'timing/model': 1.8397391840117052, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-23 10:14:16,714][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6523380875587463, 'train/grad_norm_pre_clip': 5.962258338928223, 'timing/data': 0.0023765950463712215, 'timing/model': 1.6800898920046166, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-23 10:14:19,734][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.7140339612960815, 'train/grad_norm_pre_clip': 2.2732553482055664, 'timing/data': 0.0008706169901415706, 'timing/model': 1.5162244559032843, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 10:14:22,686][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.6931897401809692, 'train/grad_norm_pre_clip': 4.707006931304932, 'timing/data': 0.0006526180077344179, 'timing/model': 1.5103050000034273, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 10:14:25,653][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.6616561412811279, 'train/grad_norm_pre_clip': 9.398393630981445, 'timing/data': 0.0005473659839481115, 'timing/model': 1.520067519042641, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-23 10:14:28,626][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.7269362211227417, 'train/grad_norm_pre_clip': 7.576957702636719, 'timing/data': 0.0007708399789407849, 'timing/model': 1.512185626081191, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-23 10:14:31,604][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.6703070998191833, 'train/grad_norm_pre_clip': 7.203150749206543, 'timing/data': 0.0005558440461754799, 'timing/model': 1.5170005980180576, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-23 10:14:34,581][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.6768601536750793, 'train/grad_norm_pre_clip': 2.1330463886260986, 'timing/data': 0.0005337420152500272, 'timing/model': 1.5205115970456973, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-23 10:14:37,567][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.660283088684082, 'train/grad_norm_pre_clip': 1.0477224588394165, 'timing/data': 0.0005532330833375454, 'timing/model': 1.5195531990611926, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-23 10:14:40,548][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.6498410105705261, 'train/grad_norm_pre_clip': 1.2303016185760498, 'timing/data': 0.000652817077934742, 'timing/model': 1.5216674769762903, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-23 10:14:43,526][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.659404993057251, 'train/grad_norm_pre_clip': 1.725275993347168, 'timing/data': 0.0005743280053138733, 'timing/model': 1.52380397496745, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-23 10:14:46,508][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.6620751619338989, 'train/grad_norm_pre_clip': 8.126851081848145, 'timing/data': 0.0005559960845857859, 'timing/model': 1.5187767000170425, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 10:14:49,520][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.6819552183151245, 'train/grad_norm_pre_clip': 6.830223083496094, 'timing/data': 0.0006645959801971912, 'timing/model': 1.5309354530181736, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 10:14:52,504][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.7147385478019714, 'train/grad_norm_pre_clip': 6.8961710929870605, 'timing/data': 0.0006874300306662917, 'timing/model': 1.5141494510462508, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-23 10:14:55,504][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.6737176775932312, 'train/grad_norm_pre_clip': 6.580173969268799, 'timing/data': 0.0006350089097395539, 'timing/model': 1.5286309000803158, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-23 10:14:58,512][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.681013286113739, 'train/grad_norm_pre_clip': 5.524428367614746, 'timing/data': 0.0005523600848391652, 'timing/model': 1.5318575879791752, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-23 10:15:01,501][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.6694365739822388, 'train/grad_norm_pre_clip': 1.5273363590240479, 'timing/data': 0.0005350220017135143, 'timing/model': 1.5230422179447487, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-23 10:15:04,507][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.6812246441841125, 'train/grad_norm_pre_clip': 5.195358753204346, 'timing/data': 0.0007966560078784823, 'timing/model': 1.5332821420161054, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 10:15:07,507][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.6537700891494751, 'train/grad_norm_pre_clip': 4.998237133026123, 'timing/data': 0.0005744470981881022, 'timing/model': 1.5253214250551537, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 10:15:10,552][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.6757407188415527, 'train/grad_norm_pre_clip': 2.1135923862457275, 'timing/data': 0.0005515619413927197, 'timing/model': 1.5461297859437764, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-23 10:15:13,564][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.6518971920013428, 'train/grad_norm_pre_clip': 3.2786788940429688, 'timing/data': 0.0006185959791764617, 'timing/model': 1.5277060630032793, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-23 10:15:16,575][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.6908829212188721, 'train/grad_norm_pre_clip': 10.207283020019531, 'timing/data': 0.0006382429273799062, 'timing/model': 1.5313466610386968, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-23 10:15:19,587][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.7147893309593201, 'train/grad_norm_pre_clip': 5.248040676116943, 'timing/data': 0.000567910959944129, 'timing/model': 1.5360729160020128, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-23 10:15:22,606][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.6717204451560974, 'train/grad_norm_pre_clip': 3.10201358795166, 'timing/data': 0.0005949169863015413, 'timing/model': 1.539540937053971, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 10:15:25,630][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.6599851846694946, 'train/grad_norm_pre_clip': 1.7777186632156372, 'timing/data': 0.0005587070481851697, 'timing/model': 1.5422145850025117, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-23 10:15:28,644][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.6563889980316162, 'train/grad_norm_pre_clip': 8.686017036437988, 'timing/data': 0.0005697259912267327, 'timing/model': 1.534327411907725, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-23 10:15:31,656][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.6671591997146606, 'train/grad_norm_pre_clip': 8.334672927856445, 'timing/data': 0.0006698879878968, 'timing/model': 1.5281012429622933, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-23 10:15:34,674][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.625405490398407, 'train/grad_norm_pre_clip': 12.471739768981934, 'timing/data': 0.0005767600378021598, 'timing/model': 1.5387306340271607, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-23 10:15:37,688][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.6909466981887817, 'train/grad_norm_pre_clip': 3.7073240280151367, 'timing/data': 0.000646978965960443, 'timing/model': 1.5422987029887736, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-23 10:15:40,700][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.5935577154159546, 'train/grad_norm_pre_clip': 3.03653621673584, 'timing/data': 0.000748916994780302, 'timing/model': 1.5380227459827438, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-23 10:15:43,857][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.6709893941879272, 'train/grad_norm_pre_clip': 3.916173219680786, 'timing/data': 0.0005663780029863119, 'timing/model': 1.5375967690488324, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01}) +[2026-06-23 10:15:46,877][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.6681737899780273, 'train/grad_norm_pre_clip': 6.575196743011475, 'timing/data': 0.000564487068913877, 'timing/model': 1.5422432459890842, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01}) +[2026-06-23 10:15:49,898][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.6512740850448608, 'train/grad_norm_pre_clip': 3.292530059814453, 'timing/data': 0.000662115984596312, 'timing/model': 1.536135277012363, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.01}) +[2026-06-23 10:15:52,925][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.6014984250068665, 'train/grad_norm_pre_clip': 4.268923759460449, 'timing/data': 0.0005588670028373599, 'timing/model': 1.537886338075623, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.01}) +[2026-06-23 10:15:55,944][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.6423981189727783, 'train/grad_norm_pre_clip': 4.807048320770264, 'timing/data': 0.0005584859754890203, 'timing/model': 1.5384471910074353, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 10:15:58,956][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.6415889263153076, 'train/grad_norm_pre_clip': 8.084813117980957, 'timing/data': 0.0007853100541979074, 'timing/model': 1.54114220791962, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-23 10:16:01,974][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.5975027084350586, 'train/grad_norm_pre_clip': 11.462723731994629, 'timing/data': 0.0005608409410342574, 'timing/model': 1.5335066210245714, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-23 10:16:05,006][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.6693543195724487, 'train/grad_norm_pre_clip': 11.854961395263672, 'timing/data': 0.0005856909556314349, 'timing/model': 1.5392410840140656, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-23 10:16:08,022][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.6289845108985901, 'train/grad_norm_pre_clip': 15.088276863098145, 'timing/data': 0.0005804069805890322, 'timing/model': 1.5467419310007244, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-23 10:16:11,058][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.5377486944198608, 'train/grad_norm_pre_clip': 7.8678765296936035, 'timing/data': 0.0006248470162972808, 'timing/model': 1.5485421469202265, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 10:16:14,083][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.5668773651123047, 'train/grad_norm_pre_clip': 7.110387325286865, 'timing/data': 0.0005649940576404333, 'timing/model': 1.5347346179187298, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-23 10:16:17,117][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.6021478176116943, 'train/grad_norm_pre_clip': 7.931105136871338, 'timing/data': 0.0005520139820873737, 'timing/model': 1.5424115699715912, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-23 10:16:20,152][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.5665763020515442, 'train/grad_norm_pre_clip': 12.397473335266113, 'timing/data': 0.0006430629873648286, 'timing/model': 1.5475155330495909, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-23 10:16:23,185][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.5816208720207214, 'train/grad_norm_pre_clip': 11.139395713806152, 'timing/data': 0.0005973330698907375, 'timing/model': 1.546455832081847, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-23 10:16:26,206][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.5425006151199341, 'train/grad_norm_pre_clip': 12.069296836853027, 'timing/data': 0.0006153460126370192, 'timing/model': 1.5429107929812744, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-23 10:16:29,225][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.48496687412261963, 'train/grad_norm_pre_clip': 15.768854141235352, 'timing/data': 0.0005612669046968222, 'timing/model': 1.5410837680101395, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-23 10:16:32,271][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.49796557426452637, 'train/grad_norm_pre_clip': 19.013479232788086, 'timing/data': 0.0005874729249626398, 'timing/model': 1.5647942700888962, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 10:16:35,316][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.5117044448852539, 'train/grad_norm_pre_clip': 14.812139511108398, 'timing/data': 0.0006033680401742458, 'timing/model': 1.5550254189874977, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-23 10:16:38,339][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.4989272952079773, 'train/grad_norm_pre_clip': 17.278514862060547, 'timing/data': 0.0006224549142643809, 'timing/model': 1.5352843690197915, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-23 10:16:41,365][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.550362765789032, 'train/grad_norm_pre_clip': 23.204771041870117, 'timing/data': 0.0008635309059172869, 'timing/model': 1.5447753409389406, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-23 10:16:44,383][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.5287950038909912, 'train/grad_norm_pre_clip': 14.818511962890625, 'timing/data': 0.0006512559484690428, 'timing/model': 1.5366180710261688, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-23 10:16:47,417][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.4713500440120697, 'train/grad_norm_pre_clip': 10.933185577392578, 'timing/data': 0.0005640690214931965, 'timing/model': 1.5438606289681047, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 10:16:50,456][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.5487114191055298, 'train/grad_norm_pre_clip': 24.884042739868164, 'timing/data': 0.0007540889782831073, 'timing/model': 1.5427810880355537, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-23 10:16:53,486][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.6187067627906799, 'train/grad_norm_pre_clip': 56.646305084228516, 'timing/data': 0.0005843819817528129, 'timing/model': 1.5427116889040917, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-23 10:16:56,556][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.510978102684021, 'train/grad_norm_pre_clip': 47.476036071777344, 'timing/data': 0.000817288993857801, 'timing/model': 1.53764923394192, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.02}) +[2026-06-23 10:16:59,569][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.44548842310905457, 'train/grad_norm_pre_clip': 15.848776817321777, 'timing/data': 0.0006002209847792983, 'timing/model': 1.5303239760687575, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-23 10:17:02,607][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.4617372453212738, 'train/grad_norm_pre_clip': 20.325763702392578, 'timing/data': 0.000867106020450592, 'timing/model': 1.5482085209805518, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.02}) +[2026-06-23 10:17:05,683][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.45248204469680786, 'train/grad_norm_pre_clip': 25.546062469482422, 'timing/data': 0.000572585966438055, 'timing/model': 1.5874354530824348, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.02}) +[2026-06-23 10:17:08,700][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.5957993865013123, 'train/grad_norm_pre_clip': 20.31700897216797, 'timing/data': 0.0005826029228046536, 'timing/model': 1.5316359039861709, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.02}) +[2026-06-23 10:17:11,714][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.4862250089645386, 'train/grad_norm_pre_clip': 25.59623908996582, 'timing/data': 0.0005742899375036359, 'timing/model': 1.5315225889207795, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.02}) +[2026-06-23 10:17:14,728][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.43787017464637756, 'train/grad_norm_pre_clip': 31.783660888671875, 'timing/data': 0.0006824469892308116, 'timing/model': 1.5359660409158096, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-23 10:17:17,769][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.34260815382003784, 'train/grad_norm_pre_clip': 23.51940155029297, 'timing/data': 0.0005743460496887565, 'timing/model': 1.5507640690775588, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-23 10:17:20,806][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.495596319437027, 'train/grad_norm_pre_clip': 20.65730857849121, 'timing/data': 0.000558825908228755, 'timing/model': 1.5509703629650176, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-23 10:17:23,827][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.5752513408660889, 'train/grad_norm_pre_clip': 67.66534423828125, 'timing/data': 0.0005744330119341612, 'timing/model': 1.5379378700163215, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 10:17:26,859][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.5267842411994934, 'train/grad_norm_pre_clip': 51.911319732666016, 'timing/data': 0.000577129889279604, 'timing/model': 1.542699869023636, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-23 10:17:29,884][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.43097877502441406, 'train/grad_norm_pre_clip': 48.91765594482422, 'timing/data': 0.0006192909786477685, 'timing/model': 1.5442120319930837, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-23 10:17:32,916][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.4603678286075592, 'train/grad_norm_pre_clip': 21.413541793823242, 'timing/data': 0.000678443000651896, 'timing/model': 1.544292801991105, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-23 10:17:35,954][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.4378911554813385, 'train/grad_norm_pre_clip': 27.89698028564453, 'timing/data': 0.0005673660198226571, 'timing/model': 1.5429879440926015, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-23 10:17:38,983][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.41888704895973206, 'train/grad_norm_pre_clip': 27.375049591064453, 'timing/data': 0.0007277210243046284, 'timing/model': 1.538147502928041, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-23 10:17:42,017][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.45633551478385925, 'train/grad_norm_pre_clip': 27.098661422729492, 'timing/data': 0.0005793790332973003, 'timing/model': 1.5416517879348248, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-23 10:17:45,050][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.41293179988861084, 'train/grad_norm_pre_clip': 16.949872970581055, 'timing/data': 0.0005834139883518219, 'timing/model': 1.5485582050168887, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-23 10:17:48,077][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.4142419099807739, 'train/grad_norm_pre_clip': 26.331178665161133, 'timing/data': 0.000567501992918551, 'timing/model': 1.539751980919391, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-23 10:17:51,112][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.43780529499053955, 'train/grad_norm_pre_clip': 32.44105529785156, 'timing/data': 0.0005872079636901617, 'timing/model': 1.5464194240048528, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-23 10:17:54,132][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.5656741857528687, 'train/grad_norm_pre_clip': 34.677791595458984, 'timing/data': 0.0006512129912152886, 'timing/model': 1.5379390550078824, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 10:17:57,159][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.40487486124038696, 'train/grad_norm_pre_clip': 26.735273361206055, 'timing/data': 0.0006283889524638653, 'timing/model': 1.5348222039174289, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.03}) +[2026-06-23 10:18:00,202][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.48031359910964966, 'train/grad_norm_pre_clip': 48.81460189819336, 'timing/data': 0.0007724020397290587, 'timing/model': 1.5502854670630768, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-23 10:18:03,245][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.5259612798690796, 'train/grad_norm_pre_clip': 25.573339462280273, 'timing/data': 0.000907503068447113, 'timing/model': 1.5428648919332772, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.03}) +[2026-06-23 10:18:06,287][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.39355969429016113, 'train/grad_norm_pre_clip': 16.573078155517578, 'timing/data': 0.0005852760514244437, 'timing/model': 1.549635597039014, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.03}) +[2026-06-23 10:18:09,319][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.4057491719722748, 'train/grad_norm_pre_clip': 31.659774780273438, 'timing/data': 0.0005999760469421744, 'timing/model': 1.5486065070144832, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.03}) +[2026-06-23 10:18:12,358][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.30000343918800354, 'train/grad_norm_pre_clip': 12.817178726196289, 'timing/data': 0.0006307479925453663, 'timing/model': 1.5398174279835075, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.03}) +[2026-06-23 10:18:15,405][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.40156635642051697, 'train/grad_norm_pre_clip': 21.34717559814453, 'timing/data': 0.0005832040915265679, 'timing/model': 1.546594810904935, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.03}) +[2026-06-23 10:18:18,440][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.45939144492149353, 'train/grad_norm_pre_clip': 85.07330322265625, 'timing/data': 0.0007707249606028199, 'timing/model': 1.5453826230950654, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.03}) +[2026-06-23 10:18:21,470][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.47838160395622253, 'train/grad_norm_pre_clip': 67.75701904296875, 'timing/data': 0.0008912079501897097, 'timing/model': 1.5396801619790494, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.03}) +[2026-06-23 10:18:24,513][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.42583927512168884, 'train/grad_norm_pre_clip': 52.23711395263672, 'timing/data': 0.0006187810795381665, 'timing/model': 1.5393905619857833, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 10:18:27,541][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.4608116149902344, 'train/grad_norm_pre_clip': 69.8941650390625, 'timing/data': 0.0005714900325983763, 'timing/model': 1.5368168910499662, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.03}) +[2026-06-23 10:18:30,578][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.38158687949180603, 'train/grad_norm_pre_clip': 40.83043670654297, 'timing/data': 0.0006068709772080183, 'timing/model': 1.5398717520292848, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-23 10:18:33,612][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.4506036639213562, 'train/grad_norm_pre_clip': 40.4702033996582, 'timing/data': 0.0005818490171805024, 'timing/model': 1.5414012630935758, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-23 10:18:36,642][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.4846823811531067, 'train/grad_norm_pre_clip': 16.492883682250977, 'timing/data': 0.0006287579890340567, 'timing/model': 1.546732553979382, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-23 10:18:39,682][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.34905755519866943, 'train/grad_norm_pre_clip': 35.75798034667969, 'timing/data': 0.0005722299683839083, 'timing/model': 1.5511675879824907, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-23 10:18:42,718][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.3262105882167816, 'train/grad_norm_pre_clip': 20.621580123901367, 'timing/data': 0.0007608089363202453, 'timing/model': 1.5470886409748346, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-23 10:18:45,746][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.36530745029449463, 'train/grad_norm_pre_clip': 23.259214401245117, 'timing/data': 0.0005770459538325667, 'timing/model': 1.5450408899923787, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-23 10:18:48,792][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.34907713532447815, 'train/grad_norm_pre_clip': 18.62803840637207, 'timing/data': 0.0005627679638564587, 'timing/model': 1.554358872002922, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-23 10:18:51,837][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.375771701335907, 'train/grad_norm_pre_clip': 45.00370788574219, 'timing/data': 0.0007102011004462838, 'timing/model': 1.5546186389401555, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-23 10:18:54,884][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.3219496011734009, 'train/grad_norm_pre_clip': 24.72798728942871, 'timing/data': 0.0005643139593303204, 'timing/model': 1.555785185075365, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-23 10:18:57,928][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.40000250935554504, 'train/grad_norm_pre_clip': 50.43971633911133, 'timing/data': 0.0006708550499752164, 'timing/model': 1.552095833932981, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-23 10:19:00,950][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 100 +[2026-06-23 10:21:33,577][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 100 +[2026-06-23 10:21:33,579][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 100 +[2026-06-23 10:21:34,486][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmpmnh430ui/test.c -o /tmp/tmpmnh430ui/test.o +[2026-06-23 10:21:34,513][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmpmnh430ui/test.o -laio -o /tmp/tmpmnh430ui/a.out +[2026-06-23 10:21:37,119][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmp4kc_ta7w/test.c -o /tmp/tmp4kc_ta7w/test.o +[2026-06-23 10:21:37,148][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmp4kc_ta7w/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmp4kc_ta7w/a.out +[2026-06-23 10:21:37,214][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmp4r0kf59l/test.c -o /tmp/tmp4r0kf59l/test.o +[2026-06-23 10:21:37,238][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmp4r0kf59l/test.o -laio -o /tmp/tmp4r0kf59l/a.out +[2026-06-23 10:23:26,923][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 100 +[2026-06-23 10:23:26,924][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.45130905508995056, 'train/grad_norm_pre_clip': 22.792850494384766, 'eval/action_loss/seconds': 152.6237718849443, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.22396358847618103, 'eval/latency_2/loss': 0.22396358847618103, 'eval/flappy/loss': 0.22396358847618103, 'eval/flappy/latency_2/loss': 0.22396358847618103, 'eval/flap_precision': 0.6930747922437673, 'eval/flap_recall': 0.8441295546558705, 'eval/flap_f1': 0.7611804076665651, 'eval/flap_rate_model': 0.25069444444444444, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.7611804076665651, 'eval/latency_2/flap_precision': 0.6930747922437673, 'eval/latency_2/flap_recall': 0.8441295546558705, 'eval/latency_2/flap_f1': 0.7611804076665651, 'eval/latency_2/flap_rate_model': 0.25069444444444444, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.7611804076665651, 'eval/action_classification/data_wait_seconds': 38.35670852661133, 'eval/action_classification/predict_seconds': 68.95975494384766, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006368429167196155, 'timing/model': 1.541296805953607, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.04}) +[2026-06-23 10:23:30,726][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.3514145612716675, 'train/grad_norm_pre_clip': 24.887638092041016, 'timing/data': 0.0006691199960187078, 'timing/model': 1.929250165936537, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.04}) +[2026-06-23 10:23:34,528][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.4231286346912384, 'train/grad_norm_pre_clip': 93.56633758544922, 'timing/data': 0.0006326530128717422, 'timing/model': 1.9287578159710392, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.04}) +[2026-06-23 10:23:38,329][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.4245409667491913, 'train/grad_norm_pre_clip': 55.43621826171875, 'timing/data': 0.0005900680553168058, 'timing/model': 1.9227433920605108, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.04}) +[2026-06-23 10:23:42,126][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.5222585201263428, 'train/grad_norm_pre_clip': 34.692352294921875, 'timing/data': 0.0007850829279050231, 'timing/model': 1.9222266279393807, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.04}) +[2026-06-23 10:23:45,942][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.29694056510925293, 'train/grad_norm_pre_clip': 59.62236404418945, 'timing/data': 0.0006140829063951969, 'timing/model': 1.9380623940378428, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.04}) +[2026-06-23 10:23:49,764][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.3449838161468506, 'train/grad_norm_pre_clip': 115.56965637207031, 'timing/data': 0.0007792279357090592, 'timing/model': 1.9451170169049874, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.04}) +[2026-06-23 10:23:53,581][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.27093416452407837, 'train/grad_norm_pre_clip': 27.679786682128906, 'timing/data': 0.0005757989129051566, 'timing/model': 1.9372518199961632, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.04}) +[2026-06-23 10:23:57,396][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.1796107441186905, 'train/grad_norm_pre_clip': 35.413856506347656, 'timing/data': 0.000560521031729877, 'timing/model': 1.9336079510394484, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.04}) +[2026-06-23 10:24:01,374][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.36574289202690125, 'train/grad_norm_pre_clip': 22.045757293701172, 'timing/data': 0.0006548459641635418, 'timing/model': 1.9293153480393812, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.04}) +[2026-06-23 10:24:05,190][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.45046502351760864, 'train/grad_norm_pre_clip': 119.72572326660156, 'timing/data': 0.0005956999957561493, 'timing/model': 1.933335970970802, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.04}) +[2026-06-23 10:24:08,988][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.38108259439468384, 'train/grad_norm_pre_clip': 67.4285888671875, 'timing/data': 0.0006117819575592875, 'timing/model': 1.9319351400481537, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.04}) +[2026-06-23 10:24:12,811][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.3736657500267029, 'train/grad_norm_pre_clip': 27.326099395751953, 'timing/data': 0.0005742940120398998, 'timing/model': 1.9431539590004832, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.04}) +[2026-06-23 10:24:16,634][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.4434342682361603, 'train/grad_norm_pre_clip': 104.79842376708984, 'timing/data': 0.0007976789493113756, 'timing/model': 1.9400338230188936, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.04}) +[2026-06-23 10:24:20,456][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.47264543175697327, 'train/grad_norm_pre_clip': 127.63542175292969, 'timing/data': 0.0007556111086159945, 'timing/model': 1.9458135720342398, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.04}) +[2026-06-23 10:24:24,287][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.41154223680496216, 'train/grad_norm_pre_clip': 33.920936584472656, 'timing/data': 0.000569593976251781, 'timing/model': 1.942495038965717, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.04}) +[2026-06-23 10:24:28,122][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.34372204542160034, 'train/grad_norm_pre_clip': 71.566162109375, 'timing/data': 0.0007131499005481601, 'timing/model': 1.9492964339442551, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.04}) +[2026-06-23 10:24:32,098][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.4100813567638397, 'train/grad_norm_pre_clip': 102.5654525756836, 'timing/data': 0.0005847499705851078, 'timing/model': 2.0545492579694837, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.05}) +[2026-06-23 10:24:35,917][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.3475145101547241, 'train/grad_norm_pre_clip': 68.07958984375, 'timing/data': 0.0005705939838662744, 'timing/model': 1.9357005989877507, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.05}) +[2026-06-23 10:24:39,739][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.3470557630062103, 'train/grad_norm_pre_clip': 27.102022171020508, 'timing/data': 0.0008002769900485873, 'timing/model': 1.9378787820460275, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.05}) +[2026-06-23 10:24:43,561][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.38914597034454346, 'train/grad_norm_pre_clip': 92.2689437866211, 'timing/data': 0.0005672280676662922, 'timing/model': 1.9343565149465576, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.05}) +[2026-06-23 10:24:47,385][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.374083012342453, 'train/grad_norm_pre_clip': 54.35182571411133, 'timing/data': 0.0005833449540659785, 'timing/model': 1.9417499200208113, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.05}) +[2026-06-23 10:24:51,225][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.38703444600105286, 'train/grad_norm_pre_clip': 58.3364143371582, 'timing/data': 0.0006108370143920183, 'timing/model': 1.9531777949305251, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.05}) +[2026-06-23 10:24:55,059][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.5059823989868164, 'train/grad_norm_pre_clip': 47.22731399536133, 'timing/data': 0.0007778219878673553, 'timing/model': 1.9444394790334627, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.05}) +[2026-06-23 10:24:58,881][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.3670967221260071, 'train/grad_norm_pre_clip': 40.24013900756836, 'timing/data': 0.0006051800446584821, 'timing/model': 1.9428778189467266, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.05}) +[2026-06-23 10:25:02,718][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.6363378167152405, 'train/grad_norm_pre_clip': 56.47262954711914, 'timing/data': 0.0005794770549982786, 'timing/model': 1.949118391959928, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.05}) +[2026-06-23 10:25:06,715][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.39266493916511536, 'train/grad_norm_pre_clip': 74.20761108398438, 'timing/data': 0.0005730760749429464, 'timing/model': 2.1146413029637188, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.05}) +[2026-06-23 10:25:10,535][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.34191665053367615, 'train/grad_norm_pre_clip': 67.1787338256836, 'timing/data': 0.0005995879182592034, 'timing/model': 1.9320125540252775, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.05}) +[2026-06-23 10:25:14,363][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.3442448675632477, 'train/grad_norm_pre_clip': 54.96002960205078, 'timing/data': 0.0007415119325742126, 'timing/model': 1.9460136650595814, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.05}) +[2026-06-23 10:25:18,194][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.44399383664131165, 'train/grad_norm_pre_clip': 96.94593048095703, 'timing/data': 0.0005902930861338973, 'timing/model': 1.9430615480523556, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.05}) +[2026-06-23 10:25:22,035][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.2782225012779236, 'train/grad_norm_pre_clip': 62.03471755981445, 'timing/data': 0.000667113927192986, 'timing/model': 1.9532091800356284, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.05}) +[2026-06-23 10:25:25,865][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.35855770111083984, 'train/grad_norm_pre_clip': 49.95551681518555, 'timing/data': 0.0005842639366164804, 'timing/model': 1.9440442729974166, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.05}) +[2026-06-23 10:25:29,697][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.31731119751930237, 'train/grad_norm_pre_clip': 76.95639038085938, 'timing/data': 0.0006041289307177067, 'timing/model': 1.9459288839716464, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.05}) +[2026-06-23 10:25:33,536][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.3550280034542084, 'train/grad_norm_pre_clip': 49.77029800415039, 'timing/data': 0.0007623280398547649, 'timing/model': 1.9550800350261852, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.05}) +[2026-06-23 10:25:37,415][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.2877732217311859, 'train/grad_norm_pre_clip': 49.816490173339844, 'timing/data': 0.0006259289802983403, 'timing/model': 1.9931313690030947, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.05}) +[2026-06-23 10:25:41,262][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.26094532012939453, 'train/grad_norm_pre_clip': 40.2732048034668, 'timing/data': 0.0005851740716025233, 'timing/model': 1.9448351960163563, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.05}) +[2026-06-23 10:25:45,103][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.34424975514411926, 'train/grad_norm_pre_clip': 65.01673126220703, 'timing/data': 0.0005690839607268572, 'timing/model': 1.9516572629800066, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.05}) +[2026-06-23 10:25:48,942][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.29951176047325134, 'train/grad_norm_pre_clip': 35.80802917480469, 'timing/data': 0.0005590820219367743, 'timing/model': 1.9451048660557717, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.05}) +[2026-06-23 10:25:52,792][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.33618396520614624, 'train/grad_norm_pre_clip': 49.93272018432617, 'timing/data': 0.0005812840536236763, 'timing/model': 1.950190277886577, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.05}) +[2026-06-23 10:25:56,621][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.41881558299064636, 'train/grad_norm_pre_clip': 27.09506607055664, 'timing/data': 0.0007625310681760311, 'timing/model': 1.9405862250132486, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.05}) +[2026-06-23 10:26:00,462][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.3593134880065918, 'train/grad_norm_pre_clip': 99.48121643066406, 'timing/data': 0.0005657220026478171, 'timing/model': 1.9521035460056737, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.05}) +[2026-06-23 10:26:04,305][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.41099753975868225, 'train/grad_norm_pre_clip': 90.85930633544922, 'timing/data': 0.0005666699726134539, 'timing/model': 1.9517608650494367, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.05}) +[2026-06-23 10:26:08,152][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.4979798197746277, 'train/grad_norm_pre_clip': 38.399070739746094, 'timing/data': 0.0005842640530318022, 'timing/model': 1.9522820060374215, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.05}) +[2026-06-23 10:26:12,163][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.35109832882881165, 'train/grad_norm_pre_clip': 99.94537353515625, 'timing/data': 0.0005767130060121417, 'timing/model': 2.128099243971519, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.06}) +[2026-06-23 10:26:15,991][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.37594524025917053, 'train/grad_norm_pre_clip': 78.6368179321289, 'timing/data': 0.0006180169293656945, 'timing/model': 1.9466121330624446, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.06}) +[2026-06-23 10:26:19,842][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.27732351422309875, 'train/grad_norm_pre_clip': 83.2802505493164, 'timing/data': 0.0005963980220258236, 'timing/model': 1.95273481705226, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.06}) +[2026-06-23 10:26:23,683][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.43533968925476074, 'train/grad_norm_pre_clip': 68.14498901367188, 'timing/data': 0.0005758380284532905, 'timing/model': 1.9523971740854904, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.06}) +[2026-06-23 10:26:27,534][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.2845989167690277, 'train/grad_norm_pre_clip': 22.97077751159668, 'timing/data': 0.0005766110261902213, 'timing/model': 1.9574424949241802, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.06}) +[2026-06-23 10:26:31,381][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.46469131112098694, 'train/grad_norm_pre_clip': 47.24345397949219, 'timing/data': 0.0006288510048761964, 'timing/model': 1.953702561906539, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.06}) +[2026-06-23 10:26:35,211][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.4056599736213684, 'train/grad_norm_pre_clip': 51.77322769165039, 'timing/data': 0.0013348909560590982, 'timing/model': 1.9398286920040846, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.06}) +[2026-06-23 10:26:39,056][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.3617127239704132, 'train/grad_norm_pre_clip': 40.41915512084961, 'timing/data': 0.0005743330111727118, 'timing/model': 1.942645504954271, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.06}) +[2026-06-23 10:26:42,942][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.27685678005218506, 'train/grad_norm_pre_clip': 20.96952247619629, 'timing/data': 0.0006274159532040358, 'timing/model': 1.983456325950101, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.06}) +[2026-06-23 10:26:46,773][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.36876559257507324, 'train/grad_norm_pre_clip': 128.36558532714844, 'timing/data': 0.0006191249703988433, 'timing/model': 1.9426197780994698, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.06}) +[2026-06-23 10:26:50,611][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.3575585186481476, 'train/grad_norm_pre_clip': 99.32102966308594, 'timing/data': 0.0006284191040322185, 'timing/model': 1.9469020529650152, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.06}) +[2026-06-23 10:26:54,436][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.2902679145336151, 'train/grad_norm_pre_clip': 76.46022033691406, 'timing/data': 0.0006644350942224264, 'timing/model': 1.9435132390353829, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.06}) +[2026-06-23 10:26:58,271][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.27979496121406555, 'train/grad_norm_pre_clip': 56.79916000366211, 'timing/data': 0.0005713510327041149, 'timing/model': 1.9508166190935299, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.06}) +[2026-06-23 10:27:02,109][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.3843437135219574, 'train/grad_norm_pre_clip': 140.3013458251953, 'timing/data': 0.0005874069174751639, 'timing/model': 1.9458823300665244, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.06}) +[2026-06-23 10:27:05,948][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.3824193477630615, 'train/grad_norm_pre_clip': 116.07938385009766, 'timing/data': 0.0007302999729290605, 'timing/model': 1.9488469489151612, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.06}) +[2026-06-23 10:27:09,779][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.26287513971328735, 'train/grad_norm_pre_clip': 26.845985412597656, 'timing/data': 0.000573151046410203, 'timing/model': 1.9498409100342542, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.06}) +[2026-06-23 10:27:13,623][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.740167498588562, 'train/grad_norm_pre_clip': 260.2415771484375, 'timing/data': 0.0005802569212391973, 'timing/model': 1.9463021729607135, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.06}) +[2026-06-23 10:27:17,468][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.7205519080162048, 'train/grad_norm_pre_clip': 274.0738525390625, 'timing/data': 0.0005747010000050068, 'timing/model': 1.949467959930189, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.06}) +[2026-06-23 10:27:21,297][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.5499262809753418, 'train/grad_norm_pre_clip': 215.85372924804688, 'timing/data': 0.0005776090547442436, 'timing/model': 1.9408832310000435, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.06}) +[2026-06-23 10:27:25,128][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.47961485385894775, 'train/grad_norm_pre_clip': 122.5236587524414, 'timing/data': 0.0008036179933696985, 'timing/model': 1.9441404179669917, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.06}) +[2026-06-23 10:27:28,963][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.33738425374031067, 'train/grad_norm_pre_clip': 47.608741760253906, 'timing/data': 0.0005988730117678642, 'timing/model': 1.9436727709835395, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.06}) +[2026-06-23 10:27:32,797][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.42224204540252686, 'train/grad_norm_pre_clip': 102.42955780029297, 'timing/data': 0.0005851290188729763, 'timing/model': 1.9384408079786226, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.06}) +[2026-06-23 10:27:36,626][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.23204576969146729, 'train/grad_norm_pre_clip': 39.93352127075195, 'timing/data': 0.0005684259813278913, 'timing/model': 1.9415530059486628, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.06}) +[2026-06-23 10:27:40,467][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.2667786777019501, 'train/grad_norm_pre_clip': 31.661701202392578, 'timing/data': 0.0006169750122353435, 'timing/model': 1.9461061139591038, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.06}) +[2026-06-23 10:27:44,320][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.34940117597579956, 'train/grad_norm_pre_clip': 30.508390426635742, 'timing/data': 0.0006308810552582145, 'timing/model': 1.9554056900087744, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.06}) +[2026-06-23 10:27:48,169][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.2700282633304596, 'train/grad_norm_pre_clip': 31.14351463317871, 'timing/data': 0.0005859939847141504, 'timing/model': 1.9536582250148058, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.07}) +[2026-06-23 10:27:52,012][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.5554205179214478, 'train/grad_norm_pre_clip': 112.9281997680664, 'timing/data': 0.0006277919746935368, 'timing/model': 1.9509898339165375, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.07}) +[2026-06-23 10:27:55,863][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.2698478400707245, 'train/grad_norm_pre_clip': 87.1503677368164, 'timing/data': 0.0005705519579350948, 'timing/model': 1.9564163360046223, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.07}) +[2026-06-23 10:27:59,693][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.34716376662254333, 'train/grad_norm_pre_clip': 95.7022476196289, 'timing/data': 0.0005940239643678069, 'timing/model': 1.946389351040125, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.07}) +[2026-06-23 10:28:03,520][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.27158379554748535, 'train/grad_norm_pre_clip': 39.500484466552734, 'timing/data': 0.0006969129899516702, 'timing/model': 1.9425707389600575, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.07}) +[2026-06-23 10:28:07,356][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.3361414670944214, 'train/grad_norm_pre_clip': 94.49496459960938, 'timing/data': 0.0006154810544103384, 'timing/model': 1.9429811380105093, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.07}) +[2026-06-23 10:28:11,194][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.2733154296875, 'train/grad_norm_pre_clip': 45.31679916381836, 'timing/data': 0.0005933750653639436, 'timing/model': 1.9517163869459182, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.07}) +[2026-06-23 10:28:15,023][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.26878389716148376, 'train/grad_norm_pre_clip': 49.6859245300293, 'timing/data': 0.0006075520068407059, 'timing/model': 1.9419259029673412, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.07}) +[2026-06-23 10:28:18,861][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.21142350137233734, 'train/grad_norm_pre_clip': 68.70110321044922, 'timing/data': 0.0005791009170934558, 'timing/model': 1.942936172010377, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.07}) +[2026-06-23 10:28:22,720][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.2726711332798004, 'train/grad_norm_pre_clip': 29.320261001586914, 'timing/data': 0.0005771310534328222, 'timing/model': 1.9483864590292796, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.07}) +[2026-06-23 10:28:26,573][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.34103140234947205, 'train/grad_norm_pre_clip': 37.092769622802734, 'timing/data': 0.0008745419327169657, 'timing/model': 1.9594648230122402, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.07}) +[2026-06-23 10:28:30,396][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.36608603596687317, 'train/grad_norm_pre_clip': 67.26451110839844, 'timing/data': 0.0008806380210444331, 'timing/model': 1.9400200449163094, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.07}) +[2026-06-23 10:28:34,225][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.29126033186912537, 'train/grad_norm_pre_clip': 34.361968994140625, 'timing/data': 0.0007261999417096376, 'timing/model': 1.942787024891004, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.07}) +[2026-06-23 10:28:38,067][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.30012962222099304, 'train/grad_norm_pre_clip': 48.794464111328125, 'timing/data': 0.0006166870007291436, 'timing/model': 1.9497169440146536, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.07}) +[2026-06-23 10:28:41,908][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.2631691098213196, 'train/grad_norm_pre_clip': 108.35772705078125, 'timing/data': 0.000585783040151, 'timing/model': 1.9461250179447234, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.07}) +[2026-06-23 10:28:45,750][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.33584582805633545, 'train/grad_norm_pre_clip': 103.41686248779297, 'timing/data': 0.0006328229792416096, 'timing/model': 1.9492532330332324, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.07}) +[2026-06-23 10:28:49,588][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.28365084528923035, 'train/grad_norm_pre_clip': 29.8538818359375, 'timing/data': 0.0006060340674594045, 'timing/model': 1.9482491589151323, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.07}) +[2026-06-23 10:28:53,419][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.308793842792511, 'train/grad_norm_pre_clip': 133.32154846191406, 'timing/data': 0.0006244730902835727, 'timing/model': 1.9382430400000885, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.07}) +[2026-06-23 10:28:57,247][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.41147908568382263, 'train/grad_norm_pre_clip': 149.13584899902344, 'timing/data': 0.0006422779988497496, 'timing/model': 1.9413605210138485, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.07}) +[2026-06-23 10:29:01,079][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.2529008984565735, 'train/grad_norm_pre_clip': 53.99518585205078, 'timing/data': 0.0007224859436973929, 'timing/model': 1.9442152519477531, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.07}) +[2026-06-23 10:29:04,934][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.206576868891716, 'train/grad_norm_pre_clip': 87.3258285522461, 'timing/data': 0.0006057610735297203, 'timing/model': 1.9545144640142098, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.07}) +[2026-06-23 10:29:08,763][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.4647883176803589, 'train/grad_norm_pre_clip': 166.44493103027344, 'timing/data': 0.0005959239788353443, 'timing/model': 1.943140204064548, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.07}) +[2026-06-23 10:29:12,603][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.20249253511428833, 'train/grad_norm_pre_clip': 68.66476440429688, 'timing/data': 0.000573636032640934, 'timing/model': 1.948632508981973, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.07}) +[2026-06-23 10:29:16,453][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.21188022196292877, 'train/grad_norm_pre_clip': 34.192012786865234, 'timing/data': 0.0005872260080650449, 'timing/model': 1.9545863450039178, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.07}) +[2026-06-23 10:29:20,285][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.7515369653701782, 'train/grad_norm_pre_clip': 246.24842834472656, 'timing/data': 0.0005969980265945196, 'timing/model': 1.9466020610416308, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.07}) +[2026-06-23 10:29:24,125][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.7982382774353027, 'train/grad_norm_pre_clip': 250.4536590576172, 'timing/data': 0.0005791729781776667, 'timing/model': 1.9442324070259929, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.07}) +[2026-06-23 10:29:27,946][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.7095430493354797, 'train/grad_norm_pre_clip': 253.64927673339844, 'timing/data': 0.000795771018601954, 'timing/model': 1.935219693928957, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.08}) +[2026-06-23 10:29:31,784][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.4008365273475647, 'train/grad_norm_pre_clip': 162.1265869140625, 'timing/data': 0.0007230399642139673, 'timing/model': 1.9454233170254156, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.08}) +[2026-06-23 10:29:35,619][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.3099076747894287, 'train/grad_norm_pre_clip': 40.343685150146484, 'timing/data': 0.0006294479826465249, 'timing/model': 1.9446050469996408, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.08}) +[2026-06-23 10:29:39,462][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.2757280468940735, 'train/grad_norm_pre_clip': 91.47309112548828, 'timing/data': 0.0005917440867051482, 'timing/model': 1.9505990769248456, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.08}) +[2026-06-23 10:29:43,295][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.326491117477417, 'train/grad_norm_pre_clip': 85.91427612304688, 'timing/data': 0.0007623709971085191, 'timing/model': 1.9481290839612484, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.08}) +[2026-06-23 10:29:47,143][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.22189687192440033, 'train/grad_norm_pre_clip': 22.3739013671875, 'timing/data': 0.0007654199143871665, 'timing/model': 1.9561153569957241, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.08}) +[2026-06-23 10:29:50,989][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 200 +[2026-06-23 10:31:58,303][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 200 +[2026-06-23 10:31:58,305][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 200 +[2026-06-23 10:33:23,267][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 200 +[2026-06-23 10:33:23,268][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.268943190574646, 'train/grad_norm_pre_clip': 42.849491119384766, 'eval/action_loss/seconds': 127.31104403792415, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.1726047843694687, 'eval/latency_2/loss': 0.1726047843694687, 'eval/flappy/loss': 0.1726047843694687, 'eval/flappy/latency_2/loss': 0.1726047843694687, 'eval/flap_precision': 0.9308565531475749, 'eval/flap_recall': 0.6086369770580297, 'eval/flap_f1': 0.7360261117911057, 'eval/flap_rate_model': 0.13458333333333333, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.7360261117911057, 'eval/latency_2/flap_precision': 0.9308565531475749, 'eval/latency_2/flap_recall': 0.6086369770580297, 'eval/latency_2/flap_f1': 0.7360261117911057, 'eval/latency_2/flap_rate_model': 0.13458333333333333, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.7360261117911057, 'eval/action_classification/data_wait_seconds': 1.215937614440918, 'eval/action_classification/predict_seconds': 83.58805084228516, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0007318159332498908, 'timing/model': 1.9482211549766362, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.08}) +[2026-06-23 10:33:27,087][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.44111818075180054, 'train/grad_norm_pre_clip': 129.61798095703125, 'timing/data': 0.0005808570422232151, 'timing/model': 1.9418852400267497, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.08}) +[2026-06-23 10:33:30,909][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.30411702394485474, 'train/grad_norm_pre_clip': 116.63895416259766, 'timing/data': 0.0005929049802944064, 'timing/model': 1.943325488944538, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.08}) +[2026-06-23 10:33:34,712][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.3804003894329071, 'train/grad_norm_pre_clip': 99.62747955322266, 'timing/data': 0.0006028120405972004, 'timing/model': 1.9194621309870854, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.08}) +[2026-06-23 10:33:38,544][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.25898876786231995, 'train/grad_norm_pre_clip': 36.73512268066406, 'timing/data': 0.0005811479641124606, 'timing/model': 1.947008894989267, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.08}) +[2026-06-23 10:33:42,391][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.16919609904289246, 'train/grad_norm_pre_clip': 29.57255744934082, 'timing/data': 0.000798060093075037, 'timing/model': 1.9496030599111691, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.08}) +[2026-06-23 10:33:46,212][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.22648239135742188, 'train/grad_norm_pre_clip': 89.49766540527344, 'timing/data': 0.0005963739240542054, 'timing/model': 1.9306183630833402, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.08}) +[2026-06-23 10:33:50,064][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.1920126974582672, 'train/grad_norm_pre_clip': 49.17192459106445, 'timing/data': 0.0005877260118722916, 'timing/model': 1.9434675379889086, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.08}) +[2026-06-23 10:33:53,902][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.2501823306083679, 'train/grad_norm_pre_clip': 22.08684539794922, 'timing/data': 0.0006031329976394773, 'timing/model': 1.947500568930991, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.08}) +[2026-06-23 10:33:57,735][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.19794806838035583, 'train/grad_norm_pre_clip': 60.73007583618164, 'timing/data': 0.0005923300050199032, 'timing/model': 1.944496872019954, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.08}) +[2026-06-23 10:34:01,573][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.3955748975276947, 'train/grad_norm_pre_clip': 49.15593719482422, 'timing/data': 0.0007462179055437446, 'timing/model': 1.9471089949365705, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.08}) +[2026-06-23 10:34:05,418][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.170720174908638, 'train/grad_norm_pre_clip': 50.33164978027344, 'timing/data': 0.0005940049886703491, 'timing/model': 1.9618194559589028, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.08}) +[2026-06-23 10:34:09,249][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.2932082712650299, 'train/grad_norm_pre_clip': 36.611572265625, 'timing/data': 0.0006065800553187728, 'timing/model': 1.9413329920498654, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.08}) +[2026-06-23 10:34:13,073][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.18933922052383423, 'train/grad_norm_pre_clip': 60.79327392578125, 'timing/data': 0.0006159100448712707, 'timing/model': 1.9414489029441029, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.08}) +[2026-06-23 10:34:16,918][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.3815315365791321, 'train/grad_norm_pre_clip': 84.59596252441406, 'timing/data': 0.0006686719134449959, 'timing/model': 1.9418730229372159, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.08}) +[2026-06-23 10:34:20,752][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.29697397351264954, 'train/grad_norm_pre_clip': 68.36058044433594, 'timing/data': 0.0005952609935775399, 'timing/model': 1.9441140140406787, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.08}) +[2026-06-23 10:34:24,588][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.3697599768638611, 'train/grad_norm_pre_clip': 49.30126190185547, 'timing/data': 0.0006075890269130468, 'timing/model': 1.9463073570514098, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.08}) +[2026-06-23 10:34:28,432][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.27892735600471497, 'train/grad_norm_pre_clip': 39.799652099609375, 'timing/data': 0.0008042490808293223, 'timing/model': 1.9500289540737867, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.08}) +[2026-06-23 10:34:32,275][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.3131130635738373, 'train/grad_norm_pre_clip': 136.56910705566406, 'timing/data': 0.0007127020508050919, 'timing/model': 1.9545268439687788, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.08}) +[2026-06-23 10:34:36,166][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.3413323163986206, 'train/grad_norm_pre_clip': 84.34420013427734, 'timing/data': 0.0005865059792995453, 'timing/model': 2.0004997400101274, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.08}) +[2026-06-23 10:34:40,010][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.2756408751010895, 'train/grad_norm_pre_clip': 65.60842895507812, 'timing/data': 0.0006104520289227366, 'timing/model': 1.9507581340149045, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.09}) +[2026-06-23 10:34:43,852][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.1925782412290573, 'train/grad_norm_pre_clip': 115.76226043701172, 'timing/data': 0.0009231260046362877, 'timing/model': 1.9558002670528367, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.09}) +[2026-06-23 10:34:47,685][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.32669469714164734, 'train/grad_norm_pre_clip': 127.84645080566406, 'timing/data': 0.0005818849895149469, 'timing/model': 1.9382795679848641, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.09}) +[2026-06-23 10:34:51,518][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.3463433086872101, 'train/grad_norm_pre_clip': 74.94673156738281, 'timing/data': 0.0007316389819607139, 'timing/model': 1.9473174000158906, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.09}) +[2026-06-23 10:34:55,368][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.2874782383441925, 'train/grad_norm_pre_clip': 41.925270080566406, 'timing/data': 0.0006967330118641257, 'timing/model': 1.9428229089826345, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.09}) +[2026-06-23 10:34:59,209][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.4093864858150482, 'train/grad_norm_pre_clip': 82.69607543945312, 'timing/data': 0.0007230320479720831, 'timing/model': 1.946752034011297, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.09}) +[2026-06-23 10:35:03,039][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.22315628826618195, 'train/grad_norm_pre_clip': 42.90036392211914, 'timing/data': 0.0007141359383240342, 'timing/model': 1.9442031499929726, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.09}) +[2026-06-23 10:35:06,867][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.2601456344127655, 'train/grad_norm_pre_clip': 49.42949676513672, 'timing/data': 0.0007388320518657565, 'timing/model': 1.9423292479477823, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.09}) +[2026-06-23 10:35:10,717][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.20327875018119812, 'train/grad_norm_pre_clip': 48.32700729370117, 'timing/data': 0.0005895630456507206, 'timing/model': 1.950897101080045, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.09}) +[2026-06-23 10:35:14,549][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.2544473707675934, 'train/grad_norm_pre_clip': 41.49581527709961, 'timing/data': 0.0005849390290677547, 'timing/model': 1.9383438379736617, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.09}) +[2026-06-23 10:35:18,383][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.2628588378429413, 'train/grad_norm_pre_clip': 60.6169319152832, 'timing/data': 0.000737446011044085, 'timing/model': 1.9467887709615752, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.09}) +[2026-06-23 10:35:22,221][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.44835522770881653, 'train/grad_norm_pre_clip': 108.40596008300781, 'timing/data': 0.0005824610125273466, 'timing/model': 1.9452953060390428, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.09}) +[2026-06-23 10:35:26,060][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.16693814098834991, 'train/grad_norm_pre_clip': 58.21738052368164, 'timing/data': 0.0005944330478087068, 'timing/model': 1.947813502047211, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.09}) +[2026-06-23 10:35:29,900][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.2865563929080963, 'train/grad_norm_pre_clip': 70.3778305053711, 'timing/data': 0.0006083190673962235, 'timing/model': 1.9490484431153163, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.09}) +[2026-06-23 10:35:33,732][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.28487440943717957, 'train/grad_norm_pre_clip': 64.85220336914062, 'timing/data': 0.0006721499375998974, 'timing/model': 1.9464621579973027, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.09}) +[2026-06-23 10:35:37,575][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.18111108243465424, 'train/grad_norm_pre_clip': 41.294700622558594, 'timing/data': 0.000581223051995039, 'timing/model': 1.9473317640367895, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.09}) +[2026-06-23 10:35:41,420][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.5120505094528198, 'train/grad_norm_pre_clip': 211.8285675048828, 'timing/data': 0.000659728073514998, 'timing/model': 1.9492956299800426, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.09}) +[2026-06-23 10:35:45,267][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.4056389033794403, 'train/grad_norm_pre_clip': 134.1416473388672, 'timing/data': 0.0006422089645639062, 'timing/model': 1.9598886789754033, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.09}) +[2026-06-23 10:35:49,116][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.2285917103290558, 'train/grad_norm_pre_clip': 109.81983947753906, 'timing/data': 0.0005901140393689275, 'timing/model': 1.9488041939912364, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.09}) +[2026-06-23 10:35:52,940][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.1794472187757492, 'train/grad_norm_pre_clip': 39.12528610229492, 'timing/data': 0.0005944179138168693, 'timing/model': 1.9362539659487084, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.09}) +[2026-06-23 10:35:56,793][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.2287958413362503, 'train/grad_norm_pre_clip': 91.09251403808594, 'timing/data': 0.0005881050601601601, 'timing/model': 1.9631950040347874, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.09}) +[2026-06-23 10:36:00,821][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.23750649392604828, 'train/grad_norm_pre_clip': 98.28975677490234, 'timing/data': 0.0006014020182192326, 'timing/model': 2.1333499449538067, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.09}) +[2026-06-23 10:36:04,659][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.23090051114559174, 'train/grad_norm_pre_clip': 27.883684158325195, 'timing/data': 0.0005888999439775944, 'timing/model': 1.944686808041297, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.09}) +[2026-06-23 10:36:08,490][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.23619361221790314, 'train/grad_norm_pre_clip': 81.7645034790039, 'timing/data': 0.000797517946921289, 'timing/model': 1.9446955160237849, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.09}) +[2026-06-23 10:36:12,334][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.2497725486755371, 'train/grad_norm_pre_clip': 109.92597198486328, 'timing/data': 0.0005861009703949094, 'timing/model': 1.9543006720487028, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.09}) +[2026-06-23 10:36:16,177][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.23692962527275085, 'train/grad_norm_pre_clip': 42.97312545776367, 'timing/data': 0.000890605035237968, 'timing/model': 1.9533842121018097, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.09}) +[2026-06-23 10:36:20,027][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.24216130375862122, 'train/grad_norm_pre_clip': 45.50084686279297, 'timing/data': 0.0005826260894536972, 'timing/model': 1.9513959169853479, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.1}) +[2026-06-23 10:36:23,891][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.1827579289674759, 'train/grad_norm_pre_clip': 33.102088928222656, 'timing/data': 0.0005916529335081577, 'timing/model': 1.9639826249331236, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.1}) +[2026-06-23 10:36:27,745][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.29534876346588135, 'train/grad_norm_pre_clip': 34.83491897583008, 'timing/data': 0.0005836900090798736, 'timing/model': 1.957905195071362, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.1}) +[2026-06-23 10:36:31,574][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.24314096570014954, 'train/grad_norm_pre_clip': 31.017475128173828, 'timing/data': 0.0007308580679818988, 'timing/model': 1.9419177379459143, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.1}) +[2026-06-23 10:36:44,016][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 250 +[2026-06-23 10:36:44,016][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.26244789361953735, 'train/grad_norm_pre_clip': 95.69417572021484, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 3.2799999999999985, 'rl_games_eval/mid_train/mean_length': 42.0, 'rl_games_eval/mid_train/std_reward': 2.559999999999994, 'rl_games_eval/mid_train/std_length': 22.0, 'rl_games_eval/mid_train/macro_mean_reward': 3.2799999999999985, 'rl_games_eval/mid_train/macro_mean_length': 42.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.2799999999999985, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 42.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 2.559999999999994, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 22.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 3.2799999999999985, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006384099833667278, 'timing/model': 1.951972502982244, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.1}) +[2026-06-23 10:36:47,850][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.3265865743160248, 'train/grad_norm_pre_clip': 62.42665100097656, 'timing/data': 0.0006133700953796506, 'timing/model': 1.9411423979327083, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.1}) +[2026-06-23 10:36:51,670][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.2515696585178375, 'train/grad_norm_pre_clip': 88.38453674316406, 'timing/data': 0.0005926369922235608, 'timing/model': 1.9386213700054213, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.1}) +[2026-06-23 10:36:55,570][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.27042505145072937, 'train/grad_norm_pre_clip': 135.30160522460938, 'timing/data': 0.0007545839762315154, 'timing/model': 2.0109747709939256, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.1}) +[2026-06-23 10:36:59,404][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.35954588651657104, 'train/grad_norm_pre_clip': 109.39570617675781, 'timing/data': 0.0006157970055937767, 'timing/model': 1.9472385320113972, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.1}) +[2026-06-23 10:37:03,245][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.2565568685531616, 'train/grad_norm_pre_clip': 122.5337142944336, 'timing/data': 0.0006148190004751086, 'timing/model': 1.94359014602378, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.1}) +[2026-06-23 10:37:07,107][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.21862418949604034, 'train/grad_norm_pre_clip': 24.41958999633789, 'timing/data': 0.0008125539170578122, 'timing/model': 1.9645181220257655, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.1}) +[2026-06-23 10:37:10,959][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.2997513711452484, 'train/grad_norm_pre_clip': 120.75206756591797, 'timing/data': 0.0008781420765444636, 'timing/model': 1.946738829021342, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.1}) +[2026-06-23 10:37:14,924][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.29032182693481445, 'train/grad_norm_pre_clip': 97.09446716308594, 'timing/data': 0.0005940740229561925, 'timing/model': 1.9504082649946213, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.1}) +[2026-06-23 10:37:18,762][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.21112120151519775, 'train/grad_norm_pre_clip': 80.00199127197266, 'timing/data': 0.0005877920193597674, 'timing/model': 1.9471656879177317, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.1}) +[2026-06-23 10:37:22,608][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.2087966501712799, 'train/grad_norm_pre_clip': 23.635480880737305, 'timing/data': 0.0005961969727650285, 'timing/model': 1.948907105019316, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.1}) +[2026-06-23 10:37:26,443][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.20878949761390686, 'train/grad_norm_pre_clip': 73.42095947265625, 'timing/data': 0.0005976969841867685, 'timing/model': 1.9547665639547631, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.1}) +[2026-06-23 10:37:30,281][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.2949526607990265, 'train/grad_norm_pre_clip': 67.36466217041016, 'timing/data': 0.0007056280737742782, 'timing/model': 1.9410594559740275, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.1}) +[2026-06-23 10:37:34,122][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.18804973363876343, 'train/grad_norm_pre_clip': 23.924606323242188, 'timing/data': 0.0006210430292412639, 'timing/model': 1.947566673043184, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.1}) +[2026-06-23 10:37:37,962][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.19047842919826508, 'train/grad_norm_pre_clip': 87.66192626953125, 'timing/data': 0.0005943130236119032, 'timing/model': 1.938154323026538, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.1}) +[2026-06-23 10:37:41,802][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.21146933734416962, 'train/grad_norm_pre_clip': 131.7475128173828, 'timing/data': 0.0006731579778715968, 'timing/model': 1.9467057531001046, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.1}) +[2026-06-23 10:37:45,718][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.22585390508174896, 'train/grad_norm_pre_clip': 53.39087677001953, 'timing/data': 0.0006501560565084219, 'timing/model': 2.0287166419439018, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.1}) +[2026-06-23 10:37:49,564][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.14107905328273773, 'train/grad_norm_pre_clip': 35.43241882324219, 'timing/data': 0.0006203539669513702, 'timing/model': 1.9583447839831933, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.1}) +[2026-06-23 10:37:53,389][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.2804321348667145, 'train/grad_norm_pre_clip': 120.29019165039062, 'timing/data': 0.0005937750684097409, 'timing/model': 1.933888502069749, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.1}) +[2026-06-23 10:37:57,228][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.17544890940189362, 'train/grad_norm_pre_clip': 87.53792572021484, 'timing/data': 0.000616264995187521, 'timing/model': 1.9496445999247953, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.1}) +[2026-06-23 10:38:01,075][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.29246532917022705, 'train/grad_norm_pre_clip': 35.306396484375, 'timing/data': 0.0005925000878050923, 'timing/model': 1.949419099953957, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.1}) +[2026-06-23 10:38:04,915][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.33682334423065186, 'train/grad_norm_pre_clip': 155.09388732910156, 'timing/data': 0.000585814006626606, 'timing/model': 1.9508345749927685, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.1}) +[2026-06-23 10:38:08,753][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.3880487382411957, 'train/grad_norm_pre_clip': 90.54088592529297, 'timing/data': 0.0005807341076433659, 'timing/model': 1.9430470400257036, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.11}) +[2026-06-23 10:38:12,609][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.2541467845439911, 'train/grad_norm_pre_clip': 40.96641159057617, 'timing/data': 0.0006854389794170856, 'timing/model': 1.9522076370194554, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.11}) +[2026-06-23 10:38:16,461][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.2468433380126953, 'train/grad_norm_pre_clip': 77.2857894897461, 'timing/data': 0.0005878369556739926, 'timing/model': 1.9613116689724848, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.11}) +[2026-06-23 10:38:20,415][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.3118935525417328, 'train/grad_norm_pre_clip': 81.22282409667969, 'timing/data': 0.0006800700211897492, 'timing/model': 2.067562108975835, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.11}) +[2026-06-23 10:38:24,263][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.19492106139659882, 'train/grad_norm_pre_clip': 36.17319869995117, 'timing/data': 0.000992087065242231, 'timing/model': 1.9492538390913978, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.11}) +[2026-06-23 10:38:28,121][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.26460129022598267, 'train/grad_norm_pre_clip': 40.96540069580078, 'timing/data': 0.000851286924444139, 'timing/model': 1.958748223958537, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.11}) +[2026-06-23 10:38:31,964][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.16339197754859924, 'train/grad_norm_pre_clip': 64.33909606933594, 'timing/data': 0.0009240040089935064, 'timing/model': 1.9480375400744379, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.11}) +[2026-06-23 10:38:35,805][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.18918123841285706, 'train/grad_norm_pre_clip': 58.76597595214844, 'timing/data': 0.0005934080109000206, 'timing/model': 1.9564105679746717, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.11}) +[2026-06-23 10:38:39,665][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.21188956499099731, 'train/grad_norm_pre_clip': 40.692604064941406, 'timing/data': 0.0006013530073687434, 'timing/model': 1.966616929974407, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.11}) +[2026-06-23 10:38:43,536][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.2682175636291504, 'train/grad_norm_pre_clip': 71.69818878173828, 'timing/data': 0.0005686590448021889, 'timing/model': 1.9572047520196065, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.11}) +[2026-06-23 10:38:47,384][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.3591359555721283, 'train/grad_norm_pre_clip': 59.11836624145508, 'timing/data': 0.0005830859299749136, 'timing/model': 1.9528564750216901, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.11}) +[2026-06-23 10:38:51,283][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.26825910806655884, 'train/grad_norm_pre_clip': 75.88985443115234, 'timing/data': 0.0007193449418991804, 'timing/model': 2.0074709199834615, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.11}) +[2026-06-23 10:38:55,126][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.19781532883644104, 'train/grad_norm_pre_clip': 54.75532531738281, 'timing/data': 0.0006029519718140364, 'timing/model': 1.9509349680738524, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.11}) +[2026-06-23 10:38:58,966][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.19805386662483215, 'train/grad_norm_pre_clip': 33.4753532409668, 'timing/data': 0.0006153170252218843, 'timing/model': 1.9570342479273677, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.11}) +[2026-06-23 10:39:02,816][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.26218658685684204, 'train/grad_norm_pre_clip': 126.83336639404297, 'timing/data': 0.0007407029625028372, 'timing/model': 1.9516833510715514, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.11}) +[2026-06-23 10:39:06,658][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.27138760685920715, 'train/grad_norm_pre_clip': 123.54924774169922, 'timing/data': 0.0005864789709448814, 'timing/model': 1.9433879900025204, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.11}) +[2026-06-23 10:39:10,502][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.2196618914604187, 'train/grad_norm_pre_clip': 29.304542541503906, 'timing/data': 0.00060281902551651, 'timing/model': 1.9577457000268623, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.11}) +[2026-06-23 10:39:14,345][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.2608806788921356, 'train/grad_norm_pre_clip': 63.466129302978516, 'timing/data': 0.0005981919821351767, 'timing/model': 1.9426945380400866, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.11}) +[2026-06-23 10:39:18,173][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.22410155832767487, 'train/grad_norm_pre_clip': 80.77735137939453, 'timing/data': 0.000603295979090035, 'timing/model': 1.9392033369513229, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.11}) +[2026-06-23 10:39:22,029][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.2016293704509735, 'train/grad_norm_pre_clip': 99.0254898071289, 'timing/data': 0.0006235559703782201, 'timing/model': 1.9651277009397745, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.11}) +[2026-06-23 10:39:25,901][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.3495978116989136, 'train/grad_norm_pre_clip': 85.42832946777344, 'timing/data': 0.0007747409399598837, 'timing/model': 1.950010740896687, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.11}) +[2026-06-23 10:39:29,735][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.24466723203659058, 'train/grad_norm_pre_clip': 38.9669303894043, 'timing/data': 0.0005831560119986534, 'timing/model': 1.947671030065976, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.11}) +[2026-06-23 10:39:33,589][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.40435221791267395, 'train/grad_norm_pre_clip': 178.44741821289062, 'timing/data': 0.000590489013120532, 'timing/model': 1.9547229879535735, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.11}) +[2026-06-23 10:39:37,435][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.26822948455810547, 'train/grad_norm_pre_clip': 110.66434478759766, 'timing/data': 0.0008504779543727636, 'timing/model': 1.9501249849563465, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.11}) +[2026-06-23 10:39:41,278][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.21963202953338623, 'train/grad_norm_pre_clip': 48.04201126098633, 'timing/data': 0.0005942920688539743, 'timing/model': 1.945391517947428, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.11}) +[2026-06-23 10:39:45,117][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.24699901044368744, 'train/grad_norm_pre_clip': 82.63317108154297, 'timing/data': 0.0007308189524337649, 'timing/model': 1.9480440219631419, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.11}) +[2026-06-23 10:39:48,963][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.3339639902114868, 'train/grad_norm_pre_clip': 119.85665893554688, 'timing/data': 0.0006507450016215444, 'timing/model': 1.9489707369357347, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.12}) +[2026-06-23 10:39:52,798][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.27398812770843506, 'train/grad_norm_pre_clip': 54.81208419799805, 'timing/data': 0.000674604088999331, 'timing/model': 1.948294679983519, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.12}) +[2026-06-23 10:39:56,632][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 300 +[2026-06-23 10:42:04,839][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 300 +[2026-06-23 10:42:04,841][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 300 +[2026-06-23 10:43:29,065][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 300 +[2026-06-23 10:43:29,066][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.29146307706832886, 'train/grad_norm_pre_clip': 30.702281951904297, 'eval/action_loss/seconds': 128.20479687000625, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.14433284103870392, 'eval/latency_2/loss': 0.14433284103870392, 'eval/flappy/loss': 0.14433284103870392, 'eval/flappy/latency_2/loss': 0.14433284103870392, 'eval/flap_precision': 0.9554545454545454, 'eval/flap_recall': 0.7091767881241565, 'eval/flap_f1': 0.8140975987606507, 'eval/flap_rate_model': 0.1527777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8140975987606506, 'eval/latency_2/flap_precision': 0.9554545454545454, 'eval/latency_2/flap_recall': 0.7091767881241565, 'eval/latency_2/flap_f1': 0.8140975987606507, 'eval/latency_2/flap_rate_model': 0.1527777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8140975987606506, 'eval/action_classification/data_wait_seconds': 1.1943049430847168, 'eval/action_classification/predict_seconds': 82.86691284179688, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005829429719597101, 'timing/model': 1.942155284108594, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.12}) +[2026-06-23 10:43:32,899][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.3799431025981903, 'train/grad_norm_pre_clip': 135.66433715820312, 'timing/data': 0.0006080829771235585, 'timing/model': 1.9412858149735257, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.12}) +[2026-06-23 10:43:36,726][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.2542155981063843, 'train/grad_norm_pre_clip': 135.6900634765625, 'timing/data': 0.0006144220242276788, 'timing/model': 1.9505924279801548, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.12}) +[2026-06-23 10:43:40,553][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.19766050577163696, 'train/grad_norm_pre_clip': 109.7295150756836, 'timing/data': 0.0008129049092531204, 'timing/model': 1.9387288800207898, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.12}) +[2026-06-23 10:43:44,380][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.2341095358133316, 'train/grad_norm_pre_clip': 57.61133575439453, 'timing/data': 0.0006329100579023361, 'timing/model': 1.9400953920558095, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.12}) +[2026-06-23 10:43:48,420][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.2002958506345749, 'train/grad_norm_pre_clip': 49.25826644897461, 'timing/data': 0.0005918319802731276, 'timing/model': 2.1503549580229446, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.12}) +[2026-06-23 10:43:52,257][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.32529088854789734, 'train/grad_norm_pre_clip': 80.46466064453125, 'timing/data': 0.0006369889015331864, 'timing/model': 1.944067791919224, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.12}) +[2026-06-23 10:43:56,091][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.15618282556533813, 'train/grad_norm_pre_clip': 21.889266967773438, 'timing/data': 0.0005709381075575948, 'timing/model': 1.9433439520653337, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.12}) +[2026-06-23 10:43:59,929][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.33628469705581665, 'train/grad_norm_pre_clip': 147.70352172851562, 'timing/data': 0.0006753529887646437, 'timing/model': 1.9347389420727268, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.12}) +[2026-06-23 10:44:03,774][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.3347844183444977, 'train/grad_norm_pre_clip': 152.59771728515625, 'timing/data': 0.0006048670038580894, 'timing/model': 1.959504113998264, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.12}) +[2026-06-23 10:44:07,610][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.2159682661294937, 'train/grad_norm_pre_clip': 96.85469818115234, 'timing/data': 0.0006670339498668909, 'timing/model': 1.9512725750682876, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.12}) +[2026-06-23 10:44:11,447][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.27563348412513733, 'train/grad_norm_pre_clip': 82.74129486083984, 'timing/data': 0.000688159023411572, 'timing/model': 1.946408404968679, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.12}) +[2026-06-23 10:44:15,294][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.4283573031425476, 'train/grad_norm_pre_clip': 136.21859741210938, 'timing/data': 0.0006317490478977561, 'timing/model': 1.9540325079578906, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.12}) +[2026-06-23 10:44:19,129][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.37096601724624634, 'train/grad_norm_pre_clip': 80.78731536865234, 'timing/data': 0.0006071309326216578, 'timing/model': 1.9416546649299562, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.12}) +[2026-06-23 10:44:23,158][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.14676262438297272, 'train/grad_norm_pre_clip': 33.11667251586914, 'timing/data': 0.000620257924310863, 'timing/model': 1.9595711810979992, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.12}) +[2026-06-23 10:44:27,007][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.2276770919561386, 'train/grad_norm_pre_clip': 129.9148406982422, 'timing/data': 0.000587266986258328, 'timing/model': 1.955462479032576, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.12}) +[2026-06-23 10:44:30,836][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.32320958375930786, 'train/grad_norm_pre_clip': 149.49342346191406, 'timing/data': 0.0005878550000488758, 'timing/model': 1.947424482088536, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.12}) +[2026-06-23 10:44:34,673][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.3679339587688446, 'train/grad_norm_pre_clip': 148.41897583007812, 'timing/data': 0.0005883750272914767, 'timing/model': 1.9490299379685894, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.12}) +[2026-06-23 10:44:38,515][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.1658770591020584, 'train/grad_norm_pre_clip': 28.12112808227539, 'timing/data': 0.0007573929615318775, 'timing/model': 1.957023827941157, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.12}) +[2026-06-23 10:44:42,352][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.2030903697013855, 'train/grad_norm_pre_clip': 45.416202545166016, 'timing/data': 0.0006110479589551687, 'timing/model': 1.9494521580636501, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.12}) +[2026-06-23 10:44:46,186][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.14574557542800903, 'train/grad_norm_pre_clip': 92.08438873291016, 'timing/data': 0.0007198540261015296, 'timing/model': 1.941908463020809, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.12}) +[2026-06-23 10:44:50,043][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.24395649135112762, 'train/grad_norm_pre_clip': 74.78856658935547, 'timing/data': 0.0008417570497840643, 'timing/model': 1.9544006399810314, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.12}) +[2026-06-23 10:44:54,047][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.20137403905391693, 'train/grad_norm_pre_clip': 37.075592041015625, 'timing/data': 0.0009851149516180158, 'timing/model': 2.1052928449353203, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.12}) +[2026-06-23 10:44:57,884][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.39581942558288574, 'train/grad_norm_pre_clip': 90.60780334472656, 'timing/data': 0.0010534710017964244, 'timing/model': 1.9415448190411553, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.13}) +[2026-06-23 10:45:01,745][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.22326114773750305, 'train/grad_norm_pre_clip': 41.7859992980957, 'timing/data': 0.0005885360296815634, 'timing/model': 1.965961784007959, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.13}) +[2026-06-23 10:45:05,596][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.06823572516441345, 'train/grad_norm_pre_clip': 34.79867935180664, 'timing/data': 0.00061474007088691, 'timing/model': 1.9563268250785768, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.13}) +[2026-06-23 10:45:09,425][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.31510987877845764, 'train/grad_norm_pre_clip': 98.45222473144531, 'timing/data': 0.0005867859581485391, 'timing/model': 1.9392335099400952, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.13}) +[2026-06-23 10:45:13,279][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.10294613242149353, 'train/grad_norm_pre_clip': 56.768531799316406, 'timing/data': 0.001299566007219255, 'timing/model': 1.9646205240860581, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.13}) +[2026-06-23 10:45:17,144][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.12339454889297485, 'train/grad_norm_pre_clip': 82.47049713134766, 'timing/data': 0.0006228310521692038, 'timing/model': 1.9556891890242696, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.13}) +[2026-06-23 10:45:21,017][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.22431981563568115, 'train/grad_norm_pre_clip': 91.18077087402344, 'timing/data': 0.0005928949685767293, 'timing/model': 1.9623584910295904, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.13}) +[2026-06-23 10:45:24,874][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.23498471081256866, 'train/grad_norm_pre_clip': 43.73991012573242, 'timing/data': 0.0005895320791751146, 'timing/model': 1.9695191700011492, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.13}) +[2026-06-23 10:45:28,847][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.22231078147888184, 'train/grad_norm_pre_clip': 35.10070037841797, 'timing/data': 0.000593438046053052, 'timing/model': 1.9426264930516481, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.13}) +[2026-06-23 10:45:32,699][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.2188470959663391, 'train/grad_norm_pre_clip': 62.29098129272461, 'timing/data': 0.0005786330439150333, 'timing/model': 1.9556572929723188, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.13}) +[2026-06-23 10:45:36,580][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.18543124198913574, 'train/grad_norm_pre_clip': 43.452274322509766, 'timing/data': 0.0005815939512103796, 'timing/model': 1.9592035410460085, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.13}) +[2026-06-23 10:45:40,422][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.10326502472162247, 'train/grad_norm_pre_clip': 43.98995590209961, 'timing/data': 0.0005767790134996176, 'timing/model': 1.9540242899674922, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.13}) +[2026-06-23 10:45:44,251][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.22485163807868958, 'train/grad_norm_pre_clip': 61.66805648803711, 'timing/data': 0.0006684110267087817, 'timing/model': 1.9396964780753478, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.13}) +[2026-06-23 10:45:48,093][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.3729557991027832, 'train/grad_norm_pre_clip': 126.98462677001953, 'timing/data': 0.0007975609041750431, 'timing/model': 1.9492375110276043, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.13}) +[2026-06-23 10:45:51,929][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.23197217285633087, 'train/grad_norm_pre_clip': 77.66085815429688, 'timing/data': 0.0005812660092487931, 'timing/model': 1.9458239229861647, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.13}) +[2026-06-23 10:45:55,779][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.2041853964328766, 'train/grad_norm_pre_clip': 43.66946029663086, 'timing/data': 0.000798239023424685, 'timing/model': 1.9544596869964153, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.13}) +[2026-06-23 10:45:59,810][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.19228710234165192, 'train/grad_norm_pre_clip': 56.642635345458984, 'timing/data': 0.000775469932705164, 'timing/model': 2.135054218932055, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.13}) +[2026-06-23 10:46:03,657][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.3307843804359436, 'train/grad_norm_pre_clip': 67.95901489257812, 'timing/data': 0.000841422937810421, 'timing/model': 1.9592330809682608, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.13}) +[2026-06-23 10:46:07,505][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.25997108221054077, 'train/grad_norm_pre_clip': 53.01380157470703, 'timing/data': 0.0006137650925666094, 'timing/model': 1.9520314150722697, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.13}) +[2026-06-23 10:46:11,340][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.16601845622062683, 'train/grad_norm_pre_clip': 51.0208740234375, 'timing/data': 0.0008109241025522351, 'timing/model': 1.9492167679127306, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.13}) +[2026-06-23 10:46:15,197][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.15695254504680634, 'train/grad_norm_pre_clip': 33.825862884521484, 'timing/data': 0.0008068940369412303, 'timing/model': 1.9506394088966772, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.13}) +[2026-06-23 10:46:19,036][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.18361227214336395, 'train/grad_norm_pre_clip': 54.62784957885742, 'timing/data': 0.0005904519930481911, 'timing/model': 1.9514413529541343, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.13}) +[2026-06-23 10:46:22,874][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.14452673494815826, 'train/grad_norm_pre_clip': 57.19245147705078, 'timing/data': 0.0005935870576649904, 'timing/model': 1.9496741930488497, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.13}) +[2026-06-23 10:46:26,730][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.12170469760894775, 'train/grad_norm_pre_clip': 59.85039138793945, 'timing/data': 0.0008254660060629249, 'timing/model': 1.9596054640132934, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.13}) +[2026-06-23 10:46:30,576][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.1032748892903328, 'train/grad_norm_pre_clip': 277.8690185546875, 'timing/data': 0.0005813080351799726, 'timing/model': 1.9593553548911586, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.13}) +[2026-06-23 10:46:34,612][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.31210780143737793, 'train/grad_norm_pre_clip': 65.57391357421875, 'timing/data': 0.000594218960031867, 'timing/model': 1.9472796770278364, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.13}) +[2026-06-23 10:46:38,449][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.2239549309015274, 'train/grad_norm_pre_clip': 100.91535949707031, 'timing/data': 0.0005929019534960389, 'timing/model': 1.9460643490310758, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.14}) +[2026-06-23 10:46:42,290][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.3807448446750641, 'train/grad_norm_pre_clip': 449.7778015136719, 'timing/data': 0.0007844420615583658, 'timing/model': 1.9495494751026854, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.14}) +[2026-06-23 10:46:46,144][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.16127148270606995, 'train/grad_norm_pre_clip': 51.98495101928711, 'timing/data': 0.0006398119730874896, 'timing/model': 1.9554086230928078, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.14}) +[2026-06-23 10:46:49,988][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.2523268759250641, 'train/grad_norm_pre_clip': 59.096858978271484, 'timing/data': 0.000983805046416819, 'timing/model': 1.9482081569731236, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.14}) +[2026-06-23 10:46:53,847][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.3840095102787018, 'train/grad_norm_pre_clip': 132.05120849609375, 'timing/data': 0.0006037940038368106, 'timing/model': 1.9650281890062615, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.14}) +[2026-06-23 10:46:57,706][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.19435186684131622, 'train/grad_norm_pre_clip': 103.85133361816406, 'timing/data': 0.0008191720116883516, 'timing/model': 1.9567442239494994, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.14}) +[2026-06-23 10:47:01,621][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.1588420867919922, 'train/grad_norm_pre_clip': 48.70124435424805, 'timing/data': 0.0006005710456520319, 'timing/model': 2.0175572510343045, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.14}) +[2026-06-23 10:47:05,601][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.2948331832885742, 'train/grad_norm_pre_clip': 49.698116302490234, 'timing/data': 0.0007522950181737542, 'timing/model': 2.0729496200801805, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.14}) +[2026-06-23 10:47:09,462][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.13415488600730896, 'train/grad_norm_pre_clip': 66.82203674316406, 'timing/data': 0.0005978050176054239, 'timing/model': 1.9519935989519581, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.14}) +[2026-06-23 10:47:13,323][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.14201222360134125, 'train/grad_norm_pre_clip': 50.40337371826172, 'timing/data': 0.0005872090114280581, 'timing/model': 1.9634784350637347, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.14}) +[2026-06-23 10:47:17,190][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.3170202374458313, 'train/grad_norm_pre_clip': 70.11161041259766, 'timing/data': 0.0006176279857754707, 'timing/model': 1.959321542060934, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.14}) +[2026-06-23 10:47:21,046][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.21095845103263855, 'train/grad_norm_pre_clip': 41.03672409057617, 'timing/data': 0.0006980670150369406, 'timing/model': 1.943282608059235, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.14}) +[2026-06-23 10:47:24,897][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.28126296401023865, 'train/grad_norm_pre_clip': 77.35103607177734, 'timing/data': 0.0006118660094216466, 'timing/model': 1.9618263019947335, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.14}) +[2026-06-23 10:47:28,778][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.1371251493692398, 'train/grad_norm_pre_clip': 28.445274353027344, 'timing/data': 0.000596315017901361, 'timing/model': 1.9861351200379431, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.14}) +[2026-06-23 10:47:32,635][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.1961083859205246, 'train/grad_norm_pre_clip': 48.50113296508789, 'timing/data': 0.000588066061027348, 'timing/model': 1.9534826129674911, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.14}) +[2026-06-23 10:47:36,494][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.16142748296260834, 'train/grad_norm_pre_clip': 50.64229202270508, 'timing/data': 0.0005943140713497996, 'timing/model': 1.9572102919919416, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.14}) +[2026-06-23 10:47:40,458][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.21643699705600739, 'train/grad_norm_pre_clip': 79.1548843383789, 'timing/data': 0.0006217859918251634, 'timing/model': 2.0665707289008424, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.14}) +[2026-06-23 10:47:44,358][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.185926154255867, 'train/grad_norm_pre_clip': 86.21529388427734, 'timing/data': 0.0006106169894337654, 'timing/model': 1.9589403219288215, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.14}) +[2026-06-23 10:47:48,216][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.274543821811676, 'train/grad_norm_pre_clip': 53.44527053833008, 'timing/data': 0.0008316370658576488, 'timing/model': 1.953272175975144, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.14}) +[2026-06-23 10:47:52,055][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.16214123368263245, 'train/grad_norm_pre_clip': 67.11933135986328, 'timing/data': 0.0005960679845884442, 'timing/model': 1.9473689229926094, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.14}) +[2026-06-23 10:47:55,905][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.16836494207382202, 'train/grad_norm_pre_clip': 82.93193817138672, 'timing/data': 0.000774538959376514, 'timing/model': 1.959375721984543, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.14}) +[2026-06-23 10:47:59,763][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.1923799216747284, 'train/grad_norm_pre_clip': 42.485843658447266, 'timing/data': 0.0008139029378071427, 'timing/model': 1.9550011730752885, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.14}) +[2026-06-23 10:48:03,615][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.27136752009391785, 'train/grad_norm_pre_clip': 108.4032211303711, 'timing/data': 0.0006666070548817515, 'timing/model': 1.9546758190263063, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.14}) +[2026-06-23 10:48:07,487][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.20340490341186523, 'train/grad_norm_pre_clip': 80.6460189819336, 'timing/data': 0.0006209739949554205, 'timing/model': 1.9649435490136966, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.14}) +[2026-06-23 10:48:11,362][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.12421124428510666, 'train/grad_norm_pre_clip': 82.60005950927734, 'timing/data': 0.0007284319726750255, 'timing/model': 1.9815925500588492, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.14}) +[2026-06-23 10:48:15,199][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.2055983692407608, 'train/grad_norm_pre_clip': 45.981571197509766, 'timing/data': 0.0005878560477867723, 'timing/model': 1.9485663690138608, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.14}) +[2026-06-23 10:48:19,053][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.3006054162979126, 'train/grad_norm_pre_clip': 89.27658081054688, 'timing/data': 0.0007264818996191025, 'timing/model': 1.9506747320992872, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.15}) +[2026-06-23 10:48:22,889][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.4172568917274475, 'train/grad_norm_pre_clip': 123.80372619628906, 'timing/data': 0.0006211630534380674, 'timing/model': 1.9445781710091978, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.15}) +[2026-06-23 10:48:26,751][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.406450092792511, 'train/grad_norm_pre_clip': 114.28553009033203, 'timing/data': 0.0006739059463143349, 'timing/model': 1.95067601907067, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.15}) +[2026-06-23 10:48:30,599][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.30021584033966064, 'train/grad_norm_pre_clip': 52.304893493652344, 'timing/data': 0.0006073049735277891, 'timing/model': 1.9494323699036613, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.15}) +[2026-06-23 10:48:34,453][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.4182392656803131, 'train/grad_norm_pre_clip': 119.77339172363281, 'timing/data': 0.000612420029938221, 'timing/model': 1.9538128649583086, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.15}) +[2026-06-23 10:48:38,297][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.3188965320587158, 'train/grad_norm_pre_clip': 107.51847839355469, 'timing/data': 0.0005993880331516266, 'timing/model': 1.9481311589479446, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.15}) +[2026-06-23 10:48:42,142][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.2061718851327896, 'train/grad_norm_pre_clip': 56.09994888305664, 'timing/data': 0.0006189360283315182, 'timing/model': 1.9458051580004394, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.15}) +[2026-06-23 10:48:46,145][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.1373860239982605, 'train/grad_norm_pre_clip': 53.37560272216797, 'timing/data': 0.0006319399690255523, 'timing/model': 2.0960245539899915, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.15}) +[2026-06-23 10:48:49,997][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.23060116171836853, 'train/grad_norm_pre_clip': 72.07115936279297, 'timing/data': 0.0005642389878630638, 'timing/model': 1.9452347139595076, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.15}) +[2026-06-23 10:48:53,847][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.284722238779068, 'train/grad_norm_pre_clip': 26.440670013427734, 'timing/data': 0.0005859020166099072, 'timing/model': 1.9536955719813704, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.15}) +[2026-06-23 10:48:57,686][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.16531307995319366, 'train/grad_norm_pre_clip': 19.934783935546875, 'timing/data': 0.0006137900054454803, 'timing/model': 1.9509450329933316, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.15}) +[2026-06-23 10:49:01,532][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.18210260570049286, 'train/grad_norm_pre_clip': 32.67534255981445, 'timing/data': 0.0007887110114097595, 'timing/model': 1.952002615085803, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.15}) +[2026-06-23 10:49:05,378][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.23955249786376953, 'train/grad_norm_pre_clip': 82.65428161621094, 'timing/data': 0.0006889919750392437, 'timing/model': 1.9462946419371292, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.15}) +[2026-06-23 10:49:09,218][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.17757025361061096, 'train/grad_norm_pre_clip': 39.88462829589844, 'timing/data': 0.0006298240041360259, 'timing/model': 1.9486185259884223, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.15}) +[2026-06-23 10:49:13,059][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.17386797070503235, 'train/grad_norm_pre_clip': 62.0829963684082, 'timing/data': 0.0005912899505347013, 'timing/model': 1.949094848940149, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.15}) +[2026-06-23 10:49:16,927][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.23608125746250153, 'train/grad_norm_pre_clip': 114.40916442871094, 'timing/data': 0.0006106150103732944, 'timing/model': 1.9560671489452943, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.15}) +[2026-06-23 10:49:20,846][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.12086685746908188, 'train/grad_norm_pre_clip': 51.507720947265625, 'timing/data': 0.0006161400815472007, 'timing/model': 1.946495944983326, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.15}) +[2026-06-23 10:49:24,726][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.2605045437812805, 'train/grad_norm_pre_clip': 96.05216217041016, 'timing/data': 0.0005894450005143881, 'timing/model': 1.9590411090757698, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.15}) +[2026-06-23 10:49:28,578][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.22354795038700104, 'train/grad_norm_pre_clip': 115.7680435180664, 'timing/data': 0.000852530007250607, 'timing/model': 1.9528331370092928, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.15}) +[2026-06-23 10:49:32,435][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.18103449046611786, 'train/grad_norm_pre_clip': 68.34444427490234, 'timing/data': 0.0006414599483832717, 'timing/model': 1.9492124320240691, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.15}) +[2026-06-23 10:49:36,285][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.27098000049591064, 'train/grad_norm_pre_clip': 67.73884582519531, 'timing/data': 0.0006405870663002133, 'timing/model': 1.953242757008411, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.15}) +[2026-06-23 10:49:40,122][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.2778180241584778, 'train/grad_norm_pre_clip': 53.41337585449219, 'timing/data': 0.0007237219251692295, 'timing/model': 1.947348474059254, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.15}) +[2026-06-23 10:49:43,981][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.2454015463590622, 'train/grad_norm_pre_clip': 75.93274688720703, 'timing/data': 0.0008019799133762717, 'timing/model': 1.9592350110178813, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.15}) +[2026-06-23 10:49:47,843][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.18383504450321198, 'train/grad_norm_pre_clip': 83.36412811279297, 'timing/data': 0.0006022329907864332, 'timing/model': 1.9515918410615996, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.15}) +[2026-06-23 10:49:51,723][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.20645195245742798, 'train/grad_norm_pre_clip': 27.145545959472656, 'timing/data': 0.0006695570191368461, 'timing/model': 1.9887704090215266, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.15}) +[2026-06-23 10:49:55,568][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 400 +[2026-06-23 10:52:04,160][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 400 +[2026-06-23 10:52:04,162][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 400 +[2026-06-23 10:53:28,888][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 400 +[2026-06-23 10:53:28,889][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.1774052530527115, 'train/grad_norm_pre_clip': 81.90648651123047, 'eval/action_loss/seconds': 128.58985276008025, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.10471580177545547, 'eval/latency_2/loss': 0.10471580177545547, 'eval/flappy/loss': 0.10471580177545547, 'eval/flappy/latency_2/loss': 0.10471580177545547, 'eval/flap_precision': 0.9321663019693655, 'eval/flap_recall': 0.8623481781376519, 'eval/flap_f1': 0.8958990536277603, 'eval/flap_rate_model': 0.19041666666666668, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8958990536277602, 'eval/latency_2/flap_precision': 0.9321663019693655, 'eval/latency_2/flap_recall': 0.8623481781376519, 'eval/latency_2/flap_f1': 0.8958990536277603, 'eval/latency_2/flap_rate_model': 0.19041666666666668, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8958990536277602, 'eval/action_classification/data_wait_seconds': 1.3214178085327148, 'eval/action_classification/predict_seconds': 83.236083984375, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005882709519937634, 'timing/model': 1.9475176519481465, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.15}) +[2026-06-23 10:53:32,721][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.23520787060260773, 'train/grad_norm_pre_clip': 44.177886962890625, 'timing/data': 0.0006317520746961236, 'timing/model': 1.9507319539552554, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.16}) +[2026-06-23 10:53:36,547][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.12113488465547562, 'train/grad_norm_pre_clip': 44.76082229614258, 'timing/data': 0.0005929990438744426, 'timing/model': 1.9347935960395262, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.16}) +[2026-06-23 10:53:40,360][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.09845758229494095, 'train/grad_norm_pre_clip': 40.17560577392578, 'timing/data': 0.0006579390028491616, 'timing/model': 1.9392191569786519, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.16}) +[2026-06-23 10:53:44,181][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.2613633871078491, 'train/grad_norm_pre_clip': 57.03903579711914, 'timing/data': 0.00074585503898561, 'timing/model': 1.9382355140987784, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.16}) +[2026-06-23 10:53:47,992][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.1544249802827835, 'train/grad_norm_pre_clip': 63.15050506591797, 'timing/data': 0.0005863569676876068, 'timing/model': 1.9348563209641725, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.16}) +[2026-06-23 10:53:51,833][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.30546796321868896, 'train/grad_norm_pre_clip': 78.42073059082031, 'timing/data': 0.0006139050237834454, 'timing/model': 1.9520646469900385, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.16}) +[2026-06-23 10:53:55,665][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.1926889568567276, 'train/grad_norm_pre_clip': 83.0772933959961, 'timing/data': 0.0006258089561015368, 'timing/model': 1.9363116309978068, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.16}) +[2026-06-23 10:53:59,517][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.14697371423244476, 'train/grad_norm_pre_clip': 31.473773956298828, 'timing/data': 0.0006828010082244873, 'timing/model': 1.9587307859910652, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.16}) +[2026-06-23 10:54:03,349][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.18228453397750854, 'train/grad_norm_pre_clip': 42.148250579833984, 'timing/data': 0.0006130519323050976, 'timing/model': 1.9401201780419797, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.16}) +[2026-06-23 10:54:07,193][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.18997187912464142, 'train/grad_norm_pre_clip': 32.27104568481445, 'timing/data': 0.0006116819567978382, 'timing/model': 1.9496152120409533, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.16}) +[2026-06-23 10:54:11,035][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.18120838701725006, 'train/grad_norm_pre_clip': 82.58467102050781, 'timing/data': 0.0006139999022707343, 'timing/model': 1.9530719969188794, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.16}) +[2026-06-23 10:54:14,867][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.22067853808403015, 'train/grad_norm_pre_clip': 49.002750396728516, 'timing/data': 0.0005937860114499927, 'timing/model': 1.9434030239935964, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.16}) +[2026-06-23 10:54:18,906][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.3286820650100708, 'train/grad_norm_pre_clip': 99.80277252197266, 'timing/data': 0.0006034740945324302, 'timing/model': 1.9611660030204803, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.16}) +[2026-06-23 10:54:22,743][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.24796625971794128, 'train/grad_norm_pre_clip': 97.53876495361328, 'timing/data': 0.000834694947116077, 'timing/model': 1.9455225139390677, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.16}) +[2026-06-23 10:54:26,585][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.12320312112569809, 'train/grad_norm_pre_clip': 33.520328521728516, 'timing/data': 0.0006018989952281117, 'timing/model': 1.9495017720619217, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.16}) +[2026-06-23 10:54:30,419][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.298279345035553, 'train/grad_norm_pre_clip': 81.91259765625, 'timing/data': 0.0008084659930318594, 'timing/model': 1.9494612520793453, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.16}) +[2026-06-23 10:54:34,271][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.32402896881103516, 'train/grad_norm_pre_clip': 85.37433624267578, 'timing/data': 0.0007332341047003865, 'timing/model': 1.9580289570149034, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.16}) +[2026-06-23 10:54:38,104][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.1938418745994568, 'train/grad_norm_pre_clip': 61.98910140991211, 'timing/data': 0.0007205390138551593, 'timing/model': 1.9438925340073183, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.16}) +[2026-06-23 10:54:41,950][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.31478458642959595, 'train/grad_norm_pre_clip': 100.94850158691406, 'timing/data': 0.0006045970367267728, 'timing/model': 1.9428199189715087, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.16}) +[2026-06-23 10:54:45,784][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.2880381941795349, 'train/grad_norm_pre_clip': 101.71759796142578, 'timing/data': 0.0006260479567572474, 'timing/model': 1.9434010270051658, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.16}) +[2026-06-23 10:54:49,840][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.21577602624893188, 'train/grad_norm_pre_clip': 84.71436309814453, 'timing/data': 0.0005754639860242605, 'timing/model': 2.1482241499470547, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.16}) +[2026-06-23 10:54:53,677][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.15530741214752197, 'train/grad_norm_pre_clip': 54.976322174072266, 'timing/data': 0.0005989730125293136, 'timing/model': 1.9481368570122868, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.16}) +[2026-06-23 10:54:57,540][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.25866934657096863, 'train/grad_norm_pre_clip': 79.54808807373047, 'timing/data': 0.0005909239407628775, 'timing/model': 1.9659451519837603, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.16}) +[2026-06-23 10:55:01,393][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.21654023230075836, 'train/grad_norm_pre_clip': 94.73963928222656, 'timing/data': 0.0005840520607307553, 'timing/model': 1.958489670068957, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.16}) +[2026-06-23 10:55:05,245][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.3062763810157776, 'train/grad_norm_pre_clip': 58.58444595336914, 'timing/data': 0.0005846520652994514, 'timing/model': 1.9569482130464166, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.16}) +[2026-06-23 10:55:09,086][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.1119716465473175, 'train/grad_norm_pre_clip': 50.51933288574219, 'timing/data': 0.0006171879358589649, 'timing/model': 1.945793300983496, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.16}) +[2026-06-23 10:55:12,932][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.22247932851314545, 'train/grad_norm_pre_clip': 65.8101577758789, 'timing/data': 0.0005940889241173863, 'timing/model': 1.9493933819467202, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.17}) +[2026-06-23 10:55:16,786][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.1720617413520813, 'train/grad_norm_pre_clip': 40.722145080566406, 'timing/data': 0.0007157339714467525, 'timing/model': 1.9489087199326605, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.17}) +[2026-06-23 10:55:20,643][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.14923357963562012, 'train/grad_norm_pre_clip': 38.437931060791016, 'timing/data': 0.0005968380719423294, 'timing/model': 1.9580109620001167, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.17}) +[2026-06-23 10:55:24,676][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.14802959561347961, 'train/grad_norm_pre_clip': 32.878684997558594, 'timing/data': 0.0006270839367061853, 'timing/model': 1.9498009209055454, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.17}) +[2026-06-23 10:55:28,514][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.14808617532253265, 'train/grad_norm_pre_clip': 34.339500427246094, 'timing/data': 0.0005871310131624341, 'timing/model': 1.956663863034919, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.17}) +[2026-06-23 10:55:32,369][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.1818520724773407, 'train/grad_norm_pre_clip': 54.99136734008789, 'timing/data': 0.0005978270201012492, 'timing/model': 1.9651555699529126, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.17}) +[2026-06-23 10:55:36,220][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.31874310970306396, 'train/grad_norm_pre_clip': 110.81782531738281, 'timing/data': 0.0006124670617282391, 'timing/model': 1.953598884982057, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.17}) +[2026-06-23 10:55:40,085][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.2206355631351471, 'train/grad_norm_pre_clip': 74.2306900024414, 'timing/data': 0.0005977660184726119, 'timing/model': 1.9577201739884913, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.17}) +[2026-06-23 10:55:43,929][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.12306559085845947, 'train/grad_norm_pre_clip': 25.459768295288086, 'timing/data': 0.0006318950327113271, 'timing/model': 1.9536824410315603, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.17}) +[2026-06-23 10:55:47,782][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.23012323677539825, 'train/grad_norm_pre_clip': 152.2833709716797, 'timing/data': 0.0006122570484876633, 'timing/model': 1.9519388970220461, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.17}) +[2026-06-23 10:55:51,629][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.32938551902770996, 'train/grad_norm_pre_clip': 166.1902313232422, 'timing/data': 0.0006100290920585394, 'timing/model': 1.9525545899523422, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.17}) +[2026-06-23 10:55:55,627][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.27023929357528687, 'train/grad_norm_pre_clip': 86.18321228027344, 'timing/data': 0.0006020469591021538, 'timing/model': 2.100003614090383, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.17}) +[2026-06-23 10:55:59,477][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.12337690591812134, 'train/grad_norm_pre_clip': 37.4647216796875, 'timing/data': 0.0008204010082408786, 'timing/model': 1.958381619071588, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.17}) +[2026-06-23 10:56:03,333][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.10867883265018463, 'train/grad_norm_pre_clip': 84.8367919921875, 'timing/data': 0.0005924060242250562, 'timing/model': 1.9624267710605636, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.17}) +[2026-06-23 10:56:07,193][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.1573883295059204, 'train/grad_norm_pre_clip': 54.04765701293945, 'timing/data': 0.0006400359561666846, 'timing/model': 1.9570292889839038, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.17}) +[2026-06-23 10:56:11,033][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.23400580883026123, 'train/grad_norm_pre_clip': 63.36557388305664, 'timing/data': 0.0006101710023358464, 'timing/model': 1.9468592870980501, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.17}) +[2026-06-23 10:56:14,895][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.18948078155517578, 'train/grad_norm_pre_clip': 47.23869705200195, 'timing/data': 0.0006014809478074312, 'timing/model': 1.9694746079621837, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.17}) +[2026-06-23 10:56:18,745][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.3264656066894531, 'train/grad_norm_pre_clip': 142.4749755859375, 'timing/data': 0.0006373990327119827, 'timing/model': 1.958208676893264, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.17}) +[2026-06-23 10:56:22,595][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.21157488226890564, 'train/grad_norm_pre_clip': 94.64502716064453, 'timing/data': 0.0005797599442303181, 'timing/model': 1.958718080073595, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.17}) +[2026-06-23 10:56:26,467][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.22757789492607117, 'train/grad_norm_pre_clip': 33.2238883972168, 'timing/data': 0.0009465659968554974, 'timing/model': 1.9677109119948, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.17}) +[2026-06-23 10:56:30,438][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.219279482960701, 'train/grad_norm_pre_clip': 92.18828582763672, 'timing/data': 0.000601058010943234, 'timing/model': 1.940267421072349, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.17}) +[2026-06-23 10:56:34,292][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.223719984292984, 'train/grad_norm_pre_clip': 108.53177642822266, 'timing/data': 0.0005842610262334347, 'timing/model': 1.9581267009489238, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.17}) +[2026-06-23 10:56:38,144][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.2270733118057251, 'train/grad_norm_pre_clip': 88.8427963256836, 'timing/data': 0.0006620270432904363, 'timing/model': 1.9500833180500194, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.17}) +[2026-06-23 10:56:42,005][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.17554092407226562, 'train/grad_norm_pre_clip': 20.786819458007812, 'timing/data': 0.0005874730413779616, 'timing/model': 1.9697342349682003, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.17}) +[2026-06-23 10:56:45,873][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.20775605738162994, 'train/grad_norm_pre_clip': 100.2004165649414, 'timing/data': 0.0005774319870397449, 'timing/model': 1.9725153120234609, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.17}) +[2026-06-23 10:56:49,728][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.4534551203250885, 'train/grad_norm_pre_clip': 179.50474548339844, 'timing/data': 0.0006616669707000256, 'timing/model': 1.949973972979933, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.17}) +[2026-06-23 10:56:53,573][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.2850075662136078, 'train/grad_norm_pre_clip': 130.64529418945312, 'timing/data': 0.0005938380490988493, 'timing/model': 1.9512079949490726, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.18}) +[2026-06-23 10:56:57,423][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.17140431702136993, 'train/grad_norm_pre_clip': 31.392061233520508, 'timing/data': 0.0006015379913151264, 'timing/model': 1.9508173880167305, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.18}) +[2026-06-23 10:57:01,357][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.19162249565124512, 'train/grad_norm_pre_clip': 91.51676177978516, 'timing/data': 0.0005920369876548648, 'timing/model': 2.044525410979986, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.18}) +[2026-06-23 10:57:05,199][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.2197011113166809, 'train/grad_norm_pre_clip': 89.01134490966797, 'timing/data': 0.0006037140265107155, 'timing/model': 1.9479264550609514, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.18}) +[2026-06-23 10:57:09,049][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.14038577675819397, 'train/grad_norm_pre_clip': 76.38553619384766, 'timing/data': 0.0005982639268040657, 'timing/model': 1.9555756540503353, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.18}) +[2026-06-23 10:57:12,897][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.11446728557348251, 'train/grad_norm_pre_clip': 62.56629180908203, 'timing/data': 0.0010468539549037814, 'timing/model': 1.9567854600027204, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.18}) +[2026-06-23 10:57:16,752][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.18914194405078888, 'train/grad_norm_pre_clip': 91.95978546142578, 'timing/data': 0.0009549950482323766, 'timing/model': 1.9534404820296913, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.18}) +[2026-06-23 10:57:20,608][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.2548355758190155, 'train/grad_norm_pre_clip': 160.69766235351562, 'timing/data': 0.0005904079880565405, 'timing/model': 1.9527458810480312, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.18}) +[2026-06-23 10:57:24,471][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.2616805136203766, 'train/grad_norm_pre_clip': 133.16058349609375, 'timing/data': 0.000630541006103158, 'timing/model': 1.9521489150356501, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.18}) +[2026-06-23 10:57:28,314][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.26349055767059326, 'train/grad_norm_pre_clip': 82.90286254882812, 'timing/data': 0.0006078940350562334, 'timing/model': 1.9438102709827945, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.18}) +[2026-06-23 10:57:32,160][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.10927727818489075, 'train/grad_norm_pre_clip': 20.690109252929688, 'timing/data': 0.0006033830577507615, 'timing/model': 1.952182154986076, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.18}) +[2026-06-23 10:57:36,199][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.14556993544101715, 'train/grad_norm_pre_clip': 60.894691467285156, 'timing/data': 0.0007118480280041695, 'timing/model': 2.132308026077226, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.18}) +[2026-06-23 10:57:40,054][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.39400234818458557, 'train/grad_norm_pre_clip': 114.36870574951172, 'timing/data': 0.0005896779475733638, 'timing/model': 1.953068993985653, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.18}) +[2026-06-23 10:57:43,889][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.19373543560504913, 'train/grad_norm_pre_clip': 106.6865463256836, 'timing/data': 0.0009121289476752281, 'timing/model': 1.9445044719614089, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.18}) +[2026-06-23 10:57:47,721][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.15201957523822784, 'train/grad_norm_pre_clip': 80.21684265136719, 'timing/data': 0.0006205900572240353, 'timing/model': 1.9442957499995828, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.18}) +[2026-06-23 10:57:51,574][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.23397213220596313, 'train/grad_norm_pre_clip': 81.68135070800781, 'timing/data': 0.0006114169955253601, 'timing/model': 1.950825144071132, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.18}) +[2026-06-23 10:57:55,435][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.16932149231433868, 'train/grad_norm_pre_clip': 78.49662780761719, 'timing/data': 0.000586440903134644, 'timing/model': 1.9601425690343603, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.18}) +[2026-06-23 10:57:59,288][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.3079979419708252, 'train/grad_norm_pre_clip': 101.22801208496094, 'timing/data': 0.0007757409475743771, 'timing/model': 1.9567479739198461, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.18}) +[2026-06-23 10:58:03,134][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.2575097978115082, 'train/grad_norm_pre_clip': 56.90617370605469, 'timing/data': 0.0009493449470028281, 'timing/model': 1.955843792995438, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.18}) +[2026-06-23 10:58:07,006][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.19039110839366913, 'train/grad_norm_pre_clip': 58.242835998535156, 'timing/data': 0.000592048978433013, 'timing/model': 1.9604726349934936, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.18}) +[2026-06-23 10:58:10,980][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.20190446078777313, 'train/grad_norm_pre_clip': 27.73919105529785, 'timing/data': 0.0008085999870672822, 'timing/model': 1.9532849240349606, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.18}) +[2026-06-23 10:58:14,827][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.20465748012065887, 'train/grad_norm_pre_clip': 23.23470115661621, 'timing/data': 0.000816997024230659, 'timing/model': 1.9545637370320037, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.18}) +[2026-06-23 10:58:18,670][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.13266852498054504, 'train/grad_norm_pre_clip': 39.7677116394043, 'timing/data': 0.0006252198945730925, 'timing/model': 1.9511819280451164, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.18}) +[2026-06-23 10:58:22,517][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.11606397479772568, 'train/grad_norm_pre_clip': 33.94862747192383, 'timing/data': 0.0005951699567958713, 'timing/model': 1.9548553199274465, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.18}) +[2026-06-23 10:58:26,365][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.15838485956192017, 'train/grad_norm_pre_clip': 39.49044418334961, 'timing/data': 0.0006522740004584193, 'timing/model': 1.956620411016047, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.18}) +[2026-06-23 10:58:30,206][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.2189631313085556, 'train/grad_norm_pre_clip': 44.071022033691406, 'timing/data': 0.000625939923338592, 'timing/model': 1.9517956549534574, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.18}) +[2026-06-23 10:58:34,045][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.20306077599525452, 'train/grad_norm_pre_clip': 27.30596160888672, 'timing/data': 0.0005920289549976587, 'timing/model': 1.948428994975984, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.19}) +[2026-06-23 10:58:37,883][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.12471814453601837, 'train/grad_norm_pre_clip': 58.502071380615234, 'timing/data': 0.0005953760119155049, 'timing/model': 1.9484802129445598, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.19}) +[2026-06-23 10:58:41,807][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.2176681011915207, 'train/grad_norm_pre_clip': 70.64096069335938, 'timing/data': 0.0005951839266344905, 'timing/model': 1.9575166678987443, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.19}) +[2026-06-23 10:58:45,718][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.16616082191467285, 'train/grad_norm_pre_clip': 63.64543533325195, 'timing/data': 0.0006024109898135066, 'timing/model': 1.9569168069865555, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.19}) +[2026-06-23 10:58:49,571][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.3075510561466217, 'train/grad_norm_pre_clip': 106.3999252319336, 'timing/data': 0.0006238759960979223, 'timing/model': 1.954435296007432, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.19}) +[2026-06-23 10:58:53,417][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.4295780658721924, 'train/grad_norm_pre_clip': 154.0966339111328, 'timing/data': 0.0006745990831404924, 'timing/model': 1.956790151889436, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.19}) +[2026-06-23 10:58:57,295][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.21793009340763092, 'train/grad_norm_pre_clip': 142.88668823242188, 'timing/data': 0.0006375309312716126, 'timing/model': 1.9697406460763887, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.19}) +[2026-06-23 10:59:01,147][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.21717935800552368, 'train/grad_norm_pre_clip': 95.49593353271484, 'timing/data': 0.000589504954405129, 'timing/model': 1.960692064021714, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.19}) +[2026-06-23 10:59:05,027][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.1996975541114807, 'train/grad_norm_pre_clip': 51.72386932373047, 'timing/data': 0.0005954069783911109, 'timing/model': 1.9749067099764943, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.19}) +[2026-06-23 10:59:08,898][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.31098416447639465, 'train/grad_norm_pre_clip': 155.7710418701172, 'timing/data': 0.0006167130777612329, 'timing/model': 1.9653367070714012, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.19}) +[2026-06-23 10:59:12,797][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.21130114793777466, 'train/grad_norm_pre_clip': 112.3508529663086, 'timing/data': 0.0005929709877818823, 'timing/model': 1.9959361840737984, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.19}) +[2026-06-23 10:59:16,651][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.14648227393627167, 'train/grad_norm_pre_clip': 36.03160858154297, 'timing/data': 0.0006034070393070579, 'timing/model': 1.9523807900259271, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.19}) +[2026-06-23 10:59:20,520][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.10743306577205658, 'train/grad_norm_pre_clip': 41.510189056396484, 'timing/data': 0.0006186800310388207, 'timing/model': 1.9689115789951757, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.19}) +[2026-06-23 10:59:24,373][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.10878127813339233, 'train/grad_norm_pre_clip': 59.611610412597656, 'timing/data': 0.0006282039685174823, 'timing/model': 1.9536531949415803, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.19}) +[2026-06-23 10:59:28,236][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.16913500428199768, 'train/grad_norm_pre_clip': 45.55949020385742, 'timing/data': 0.0005959749687463045, 'timing/model': 1.9526333298999816, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.19}) +[2026-06-23 10:59:32,093][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.11231417953968048, 'train/grad_norm_pre_clip': 31.103282928466797, 'timing/data': 0.0006079790182411671, 'timing/model': 1.9625912730116397, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.19}) +[2026-06-23 10:59:35,950][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.150692418217659, 'train/grad_norm_pre_clip': 77.0667953491211, 'timing/data': 0.0005978030385449529, 'timing/model': 1.9542820589849725, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.19}) +[2026-06-23 10:59:39,818][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.18908187747001648, 'train/grad_norm_pre_clip': 96.80157470703125, 'timing/data': 0.0006581599591299891, 'timing/model': 1.951320554013364, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.19}) +[2026-06-23 10:59:43,676][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.09795141965150833, 'train/grad_norm_pre_clip': 53.59537887573242, 'timing/data': 0.0005893229972571135, 'timing/model': 1.956842018989846, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.19}) +[2026-06-23 10:59:47,544][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.19954615831375122, 'train/grad_norm_pre_clip': 33.82973098754883, 'timing/data': 0.0006077879806980491, 'timing/model': 1.9552983220200986, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.19}) +[2026-06-23 10:59:51,390][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.13986968994140625, 'train/grad_norm_pre_clip': 30.036439895629883, 'timing/data': 0.0006086269859224558, 'timing/model': 1.9581934900488704, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.19}) +[2026-06-23 10:59:55,230][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 500 +[2026-06-23 11:02:04,913][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 500 +[2026-06-23 11:02:04,916][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 500 +[2026-06-23 11:03:29,227][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 500 +[2026-06-23 11:03:38,436][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-23 11:03:38,437][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.1124720498919487, 'train/grad_norm_pre_clip': 110.42305755615234, 'eval/action_loss/seconds': 129.67963352100924, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.12881311774253845, 'eval/latency_2/loss': 0.12881311774253845, 'eval/flappy/loss': 0.12881311774253845, 'eval/flappy/latency_2/loss': 0.12881311774253845, 'eval/flap_precision': 0.9783737024221453, 'eval/flap_recall': 0.7631578947368421, 'eval/flap_f1': 0.8574677786201669, 'eval/flap_rate_model': 0.16055555555555556, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8574677786201669, 'eval/latency_2/flap_precision': 0.9783737024221453, 'eval/latency_2/flap_recall': 0.7631578947368421, 'eval/latency_2/flap_f1': 0.8574677786201669, 'eval/latency_2/flap_rate_model': 0.16055555555555556, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8574677786201669, 'eval/action_classification/data_wait_seconds': 1.21974778175354, 'eval/action_classification/predict_seconds': 82.9171371459961, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 3.6599999999999975, 'rl_games_eval/mid_train/mean_length': 45.8, 'rl_games_eval/mid_train/std_reward': 2.48161237907937, 'rl_games_eval/mid_train/std_length': 21.40467238712146, 'rl_games_eval/mid_train/macro_mean_reward': 3.6599999999999975, 'rl_games_eval/mid_train/macro_mean_length': 45.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.6599999999999975, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 45.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 2.48161237907937, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 21.40467238712146, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 3.6599999999999975, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005915689980611205, 'timing/model': 1.9545120899565518, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.19}) +[2026-06-23 11:03:38,437][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_500_state +[2026-06-23 11:03:38,440][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 11:03:42,237][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_500_state/model.safetensors +[2026-06-23 11:03:45,136][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_500_state/optimizer.bin +[2026-06-23 11:03:45,136][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_500_state/sampler.bin +[2026-06-23 11:03:45,136][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_500_state/sampler_1.bin +[2026-06-23 11:03:45,137][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-23 11:03:45,138][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-23 11:03:45,138][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 11:03:45,159][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 11:03:48,910][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.2405761033296585, 'train/grad_norm_pre_clip': 130.22042846679688, 'timing/data': 0.0006531609687954187, 'timing/model': 1.9034407739527524, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.19}) +[2026-06-23 11:03:52,766][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.30643773078918457, 'train/grad_norm_pre_clip': 137.9949951171875, 'timing/data': 0.0006225509569048882, 'timing/model': 1.9115588680142537, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.19}) +[2026-06-23 11:03:56,525][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.15371836721897125, 'train/grad_norm_pre_clip': 42.21612548828125, 'timing/data': 0.0008829199941828847, 'timing/model': 1.9047512690303847, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.19}) +[2026-06-23 11:04:00,291][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.15802443027496338, 'train/grad_norm_pre_clip': 52.823856353759766, 'timing/data': 0.0006281830137595534, 'timing/model': 1.909047969034873, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.2}) +[2026-06-23 11:04:04,074][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.331418514251709, 'train/grad_norm_pre_clip': 80.15009307861328, 'timing/data': 0.0006656120531260967, 'timing/model': 1.92738269304391, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.2}) +[2026-06-23 11:04:07,846][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.16562505066394806, 'train/grad_norm_pre_clip': 26.425920486450195, 'timing/data': 0.0005804960383102298, 'timing/model': 1.9169009699253365, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.2}) +[2026-06-23 11:04:11,605][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.2831483483314514, 'train/grad_norm_pre_clip': 49.564170837402344, 'timing/data': 0.0007876870222389698, 'timing/model': 1.9108494400279596, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.2}) +[2026-06-23 11:04:15,391][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.15587739646434784, 'train/grad_norm_pre_clip': 44.05306625366211, 'timing/data': 0.0006126930238679051, 'timing/model': 1.921986442990601, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.2}) +[2026-06-23 11:04:19,177][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.09256234019994736, 'train/grad_norm_pre_clip': 45.618892669677734, 'timing/data': 0.000575978890992701, 'timing/model': 1.9214973410125822, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.2}) +[2026-06-23 11:04:22,971][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.10187868773937225, 'train/grad_norm_pre_clip': 43.699806213378906, 'timing/data': 0.000688052037730813, 'timing/model': 1.9220397740136832, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.2}) +[2026-06-23 11:04:26,915][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.09555981308221817, 'train/grad_norm_pre_clip': 26.087425231933594, 'timing/data': 0.0006191779393702745, 'timing/model': 1.9217712200479582, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.2}) +[2026-06-23 11:04:30,708][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.19726243615150452, 'train/grad_norm_pre_clip': 33.82552719116211, 'timing/data': 0.000611150055192411, 'timing/model': 1.923784130020067, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.2}) +[2026-06-23 11:04:34,502][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.05034944415092468, 'train/grad_norm_pre_clip': 27.689167022705078, 'timing/data': 0.0005903689889237285, 'timing/model': 1.924942706944421, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.2}) +[2026-06-23 11:04:38,306][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.241421639919281, 'train/grad_norm_pre_clip': 70.4270248413086, 'timing/data': 0.0007744150934740901, 'timing/model': 1.924963406054303, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.2}) +[2026-06-23 11:04:42,092][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.12229465693235397, 'train/grad_norm_pre_clip': 70.82694244384766, 'timing/data': 0.0006360829574987292, 'timing/model': 1.923514025984332, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.2}) +[2026-06-23 11:04:45,892][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.08256123960018158, 'train/grad_norm_pre_clip': 81.12710571289062, 'timing/data': 0.0006034580292180181, 'timing/model': 1.9259843269828707, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.2}) +[2026-06-23 11:04:49,745][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.25274211168289185, 'train/grad_norm_pre_clip': 88.63578796386719, 'timing/data': 0.0005937330424785614, 'timing/model': 1.9196107819443569, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.2}) +[2026-06-23 11:04:53,533][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.14190079271793365, 'train/grad_norm_pre_clip': 28.547605514526367, 'timing/data': 0.0005800150102004409, 'timing/model': 1.9241249540355057, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.2}) +[2026-06-23 11:04:57,495][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.15072816610336304, 'train/grad_norm_pre_clip': 23.76457405090332, 'timing/data': 0.0005969760240986943, 'timing/model': 2.09152672404889, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.2}) +[2026-06-23 11:05:01,291][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.24009732902050018, 'train/grad_norm_pre_clip': 93.84041595458984, 'timing/data': 0.0005790930008515716, 'timing/model': 1.9299547800328583, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.2}) +[2026-06-23 11:05:05,073][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.22030197083950043, 'train/grad_norm_pre_clip': 97.53495788574219, 'timing/data': 0.0005938509711995721, 'timing/model': 1.9176779399858788, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.2}) +[2026-06-23 11:05:08,889][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.20265504717826843, 'train/grad_norm_pre_clip': 52.2796745300293, 'timing/data': 0.0006020580185577273, 'timing/model': 1.9345155640039593, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.2}) +[2026-06-23 11:05:12,683][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.10922901332378387, 'train/grad_norm_pre_clip': 23.066986083984375, 'timing/data': 0.0006398149998858571, 'timing/model': 1.920470735989511, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.2}) +[2026-06-23 11:05:16,470][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.08018921315670013, 'train/grad_norm_pre_clip': 32.98386764526367, 'timing/data': 0.000590378069318831, 'timing/model': 1.9244261420099065, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.2}) +[2026-06-23 11:05:20,275][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.20554384589195251, 'train/grad_norm_pre_clip': 43.08419418334961, 'timing/data': 0.0005824370309710503, 'timing/model': 1.9213396979030222, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.2}) +[2026-06-23 11:05:24,083][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.2394496500492096, 'train/grad_norm_pre_clip': 95.60395812988281, 'timing/data': 0.0006785329896956682, 'timing/model': 1.9306996199302375, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.2}) +[2026-06-23 11:05:27,872][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.18956711888313293, 'train/grad_norm_pre_clip': 36.555992126464844, 'timing/data': 0.0006259649526327848, 'timing/model': 1.9272146750008687, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.2}) +[2026-06-23 11:05:31,847][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.07887335121631622, 'train/grad_norm_pre_clip': 36.596805572509766, 'timing/data': 0.0005906339501962066, 'timing/model': 1.9294055959908292, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.2}) +[2026-06-23 11:05:35,653][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.12680891156196594, 'train/grad_norm_pre_clip': 54.51327896118164, 'timing/data': 0.0005806899862363935, 'timing/model': 1.9258583799237385, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.2}) +[2026-06-23 11:05:39,460][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.21525521576404572, 'train/grad_norm_pre_clip': 36.25707244873047, 'timing/data': 0.0009366900194436312, 'timing/model': 1.9392902990803123, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.21}) +[2026-06-23 11:05:43,268][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.21534273028373718, 'train/grad_norm_pre_clip': 45.03035354614258, 'timing/data': 0.0005928800674155354, 'timing/model': 1.9320589849958196, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.21}) +[2026-06-23 11:05:47,066][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.202275812625885, 'train/grad_norm_pre_clip': 36.832584381103516, 'timing/data': 0.0005910010077059269, 'timing/model': 1.928046963061206, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.21}) +[2026-06-23 11:05:50,897][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.14509697258472443, 'train/grad_norm_pre_clip': 36.06534957885742, 'timing/data': 0.0007763339672237635, 'timing/model': 1.9222193559398875, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.21}) +[2026-06-23 11:05:54,710][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.14197494089603424, 'train/grad_norm_pre_clip': 29.773759841918945, 'timing/data': 0.0005854510236531496, 'timing/model': 1.9301703469827771, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.21}) +[2026-06-23 11:05:58,503][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.2274533212184906, 'train/grad_norm_pre_clip': 56.099517822265625, 'timing/data': 0.0008879369124770164, 'timing/model': 1.9272014430025592, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.21}) +[2026-06-23 11:06:02,297][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.18567895889282227, 'train/grad_norm_pre_clip': 59.64226150512695, 'timing/data': 0.0006629229756072164, 'timing/model': 1.928322701016441, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.21}) +[2026-06-23 11:06:06,250][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.14308790862560272, 'train/grad_norm_pre_clip': 66.0416259765625, 'timing/data': 0.0005982400616630912, 'timing/model': 1.9181547469925135, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.21}) +[2026-06-23 11:06:10,057][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.1836853176355362, 'train/grad_norm_pre_clip': 66.49466705322266, 'timing/data': 0.0005754940211772919, 'timing/model': 1.9268796499818563, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.21}) +[2026-06-23 11:06:13,846][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.2245873212814331, 'train/grad_norm_pre_clip': 28.483713150024414, 'timing/data': 0.0005740929627791047, 'timing/model': 1.9250802129972726, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.21}) +[2026-06-23 11:06:17,653][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.21279293298721313, 'train/grad_norm_pre_clip': 71.61514282226562, 'timing/data': 0.000620494014583528, 'timing/model': 1.9377229469828308, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.21}) +[2026-06-23 11:06:21,476][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.18291211128234863, 'train/grad_norm_pre_clip': 57.81814956665039, 'timing/data': 0.0005891530308872461, 'timing/model': 1.9587212259648368, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.21}) +[2026-06-23 11:06:25,273][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.07050321251153946, 'train/grad_norm_pre_clip': 18.885177612304688, 'timing/data': 0.0006612499710172415, 'timing/model': 1.9227283119689673, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.21}) +[2026-06-23 11:06:29,073][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.23207435011863708, 'train/grad_norm_pre_clip': 33.87709045410156, 'timing/data': 0.000575645943172276, 'timing/model': 1.931138099054806, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.21}) +[2026-06-23 11:06:32,892][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.11802009493112564, 'train/grad_norm_pre_clip': 39.75356674194336, 'timing/data': 0.0006003730231896043, 'timing/model': 1.9320979269687086, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.21}) +[2026-06-23 11:06:36,878][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.13414867222309113, 'train/grad_norm_pre_clip': 36.75122833251953, 'timing/data': 0.0006619800115004182, 'timing/model': 2.1081436849199235, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.21}) +[2026-06-23 11:06:40,686][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.19630269706249237, 'train/grad_norm_pre_clip': 46.39707946777344, 'timing/data': 0.0005865190178155899, 'timing/model': 1.9257910450687632, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.21}) +[2026-06-23 11:06:44,517][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.1381114274263382, 'train/grad_norm_pre_clip': 36.68465042114258, 'timing/data': 0.0005835969932377338, 'timing/model': 1.9555910170311108, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.21}) +[2026-06-23 11:06:48,323][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.11911898106336594, 'train/grad_norm_pre_clip': 57.522682189941406, 'timing/data': 0.0006043859757483006, 'timing/model': 1.930120492936112, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.21}) +[2026-06-23 11:06:52,132][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.080005943775177, 'train/grad_norm_pre_clip': 30.032686233520508, 'timing/data': 0.0006379769183695316, 'timing/model': 1.9331104640150443, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.21}) +[2026-06-23 11:06:55,934][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.06144319474697113, 'train/grad_norm_pre_clip': 35.35555648803711, 'timing/data': 0.0005917799426242709, 'timing/model': 1.9251814569579437, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.21}) +[2026-06-23 11:06:59,742][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.11965405195951462, 'train/grad_norm_pre_clip': 36.05046844482422, 'timing/data': 0.000585655914619565, 'timing/model': 1.9307830890174955, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.21}) +[2026-06-23 11:07:03,549][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.06454965472221375, 'train/grad_norm_pre_clip': 97.2584457397461, 'timing/data': 0.0005848429864272475, 'timing/model': 1.9328765380196273, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.21}) +[2026-06-23 11:07:07,363][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.15217357873916626, 'train/grad_norm_pre_clip': 43.356834411621094, 'timing/data': 0.0005851680180057883, 'timing/model': 1.941483118920587, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.21}) +[2026-06-23 11:07:11,351][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.33554789423942566, 'train/grad_norm_pre_clip': 143.5832977294922, 'timing/data': 0.0005937230307608843, 'timing/model': 1.930675123934634, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.21}) +[2026-06-23 11:07:15,158][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.24292393028736115, 'train/grad_norm_pre_clip': 147.94293212890625, 'timing/data': 0.0005950620397925377, 'timing/model': 1.9211961040273309, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.21}) +[2026-06-23 11:07:18,980][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.1077641099691391, 'train/grad_norm_pre_clip': 116.703369140625, 'timing/data': 0.0006176569731906056, 'timing/model': 1.943570151925087, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.22}) +[2026-06-23 11:07:22,793][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.05405309423804283, 'train/grad_norm_pre_clip': 75.66860961914062, 'timing/data': 0.0006761769764125347, 'timing/model': 1.9265282470732927, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.22}) +[2026-06-23 11:07:26,616][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.38592246174812317, 'train/grad_norm_pre_clip': 128.16452026367188, 'timing/data': 0.0006071609677746892, 'timing/model': 1.9340088019380346, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.22}) +[2026-06-23 11:07:30,418][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.35458168387413025, 'train/grad_norm_pre_clip': 78.97201538085938, 'timing/data': 0.0007702200673520565, 'timing/model': 1.920915122027509, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.22}) +[2026-06-23 11:07:34,234][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.05809780955314636, 'train/grad_norm_pre_clip': 39.173336029052734, 'timing/data': 0.0005876709474250674, 'timing/model': 1.940656555001624, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.22}) +[2026-06-23 11:07:38,037][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.2568836212158203, 'train/grad_norm_pre_clip': 62.28816604614258, 'timing/data': 0.0006012929370626807, 'timing/model': 1.927274925983511, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.22}) +[2026-06-23 11:07:41,851][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.19299212098121643, 'train/grad_norm_pre_clip': 77.4774398803711, 'timing/data': 0.0005916660884395242, 'timing/model': 1.9344554130220786, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.22}) +[2026-06-23 11:07:45,834][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.23412665724754333, 'train/grad_norm_pre_clip': 84.60079956054688, 'timing/data': 0.0005797959165647626, 'timing/model': 1.934041823958978, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.22}) +[2026-06-23 11:07:49,638][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.127748042345047, 'train/grad_norm_pre_clip': 44.978389739990234, 'timing/data': 0.0006381190614774823, 'timing/model': 1.9287797369761392, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.22}) +[2026-06-23 11:07:53,434][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.12762293219566345, 'train/grad_norm_pre_clip': 32.287471771240234, 'timing/data': 0.0005782679654657841, 'timing/model': 1.925057276035659, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.22}) +[2026-06-23 11:07:57,238][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.12489558756351471, 'train/grad_norm_pre_clip': 42.71925735473633, 'timing/data': 0.0005923178978264332, 'timing/model': 1.9224920379929245, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.22}) +[2026-06-23 11:08:01,056][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.08425557613372803, 'train/grad_norm_pre_clip': 39.86806869506836, 'timing/data': 0.0005930340848863125, 'timing/model': 1.9359761340310797, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.22}) +[2026-06-23 11:08:04,920][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.1284852921962738, 'train/grad_norm_pre_clip': 33.25332260131836, 'timing/data': 0.0005958660040050745, 'timing/model': 1.9334793620510027, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.22}) +[2026-06-23 11:08:08,730][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.10986637324094772, 'train/grad_norm_pre_clip': 20.73630142211914, 'timing/data': 0.0006101420149207115, 'timing/model': 1.9257992550265044, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.22}) +[2026-06-23 11:08:12,537][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.12712964415550232, 'train/grad_norm_pre_clip': 54.9610710144043, 'timing/data': 0.0006484959740191698, 'timing/model': 1.9203314000042155, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.22}) +[2026-06-23 11:08:16,504][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.14627975225448608, 'train/grad_norm_pre_clip': 28.537931442260742, 'timing/data': 0.0007771069649606943, 'timing/model': 2.094980144058354, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.22}) +[2026-06-23 11:08:20,311][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.24275188148021698, 'train/grad_norm_pre_clip': 43.96613311767578, 'timing/data': 0.0005935750668868423, 'timing/model': 1.9243957019643858, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.22}) +[2026-06-23 11:08:24,120][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.14357216656208038, 'train/grad_norm_pre_clip': 44.180301666259766, 'timing/data': 0.0005965050077065825, 'timing/model': 1.93502063897904, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.22}) +[2026-06-23 11:08:27,924][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.06831414997577667, 'train/grad_norm_pre_clip': 50.8111686706543, 'timing/data': 0.0009500980377197266, 'timing/model': 1.9257782149361446, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.22}) +[2026-06-23 11:08:31,738][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.18549744784832, 'train/grad_norm_pre_clip': 76.24419403076172, 'timing/data': 0.000594370998442173, 'timing/model': 1.9237855360843241, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.22}) +[2026-06-23 11:08:35,559][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.1694840043783188, 'train/grad_norm_pre_clip': 26.28952407836914, 'timing/data': 0.0006105010397732258, 'timing/model': 1.9436345780268312, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.22}) +[2026-06-23 11:08:39,368][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.08874771744012833, 'train/grad_norm_pre_clip': 82.88525390625, 'timing/data': 0.0005834329640492797, 'timing/model': 1.9356710839783773, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.22}) +[2026-06-23 11:08:43,182][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.07374254614114761, 'train/grad_norm_pre_clip': 26.725379943847656, 'timing/data': 0.000588919036090374, 'timing/model': 1.9332997419405729, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.22}) +[2026-06-23 11:08:46,996][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.08672033995389938, 'train/grad_norm_pre_clip': 62.38777542114258, 'timing/data': 0.0005942040588706732, 'timing/model': 1.9397044359939173, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.22}) +[2026-06-23 11:08:50,960][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.08022814244031906, 'train/grad_norm_pre_clip': 29.26263427734375, 'timing/data': 0.0009048579959198833, 'timing/model': 1.951128768036142, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.22}) +[2026-06-23 11:08:54,761][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.24325692653656006, 'train/grad_norm_pre_clip': 78.57644653320312, 'timing/data': 0.0006155379815027118, 'timing/model': 1.9266548759769648, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.22}) +[2026-06-23 11:08:58,573][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.32621055841445923, 'train/grad_norm_pre_clip': 136.68856811523438, 'timing/data': 0.000606511952355504, 'timing/model': 1.9299120209179819, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.23}) +[2026-06-23 11:09:02,401][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.34928253293037415, 'train/grad_norm_pre_clip': 92.55997467041016, 'timing/data': 0.0005876830546185374, 'timing/model': 1.9442686380352825, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.23}) +[2026-06-23 11:09:06,208][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.17137007415294647, 'train/grad_norm_pre_clip': 38.564186096191406, 'timing/data': 0.0005822499515488744, 'timing/model': 1.9316420000977814, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.23}) +[2026-06-23 11:09:10,001][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.39571240544319153, 'train/grad_norm_pre_clip': 107.74732971191406, 'timing/data': 0.0005864909617230296, 'timing/model': 1.9232983569381759, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.23}) +[2026-06-23 11:09:13,813][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.18574485182762146, 'train/grad_norm_pre_clip': 76.64779663085938, 'timing/data': 0.0005789739079773426, 'timing/model': 1.934374551055953, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.23}) +[2026-06-23 11:09:17,614][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.12490349262952805, 'train/grad_norm_pre_clip': 70.97129821777344, 'timing/data': 0.0005859519587829709, 'timing/model': 1.9285105608869344, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.23}) +[2026-06-23 11:09:21,601][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.10015576332807541, 'train/grad_norm_pre_clip': 49.29111099243164, 'timing/data': 0.0005724559305235744, 'timing/model': 2.1112052560783923, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.23}) +[2026-06-23 11:09:25,418][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.30440571904182434, 'train/grad_norm_pre_clip': 65.52279663085938, 'timing/data': 0.0006146109662950039, 'timing/model': 1.9365769799333066, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.23}) +[2026-06-23 11:09:29,214][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.17017336189746857, 'train/grad_norm_pre_clip': 37.555145263671875, 'timing/data': 0.0007640940602868795, 'timing/model': 1.9307034070370719, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.23}) +[2026-06-23 11:09:33,019][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.1768823266029358, 'train/grad_norm_pre_clip': 28.25692367553711, 'timing/data': 0.0007203540299087763, 'timing/model': 1.931463724002242, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.23}) +[2026-06-23 11:09:36,807][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.14430080354213715, 'train/grad_norm_pre_clip': 33.50870132446289, 'timing/data': 0.0005918939132243395, 'timing/model': 1.9211672689998522, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.23}) +[2026-06-23 11:09:40,614][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.13866524398326874, 'train/grad_norm_pre_clip': 58.48307800292969, 'timing/data': 0.0006500920280814171, 'timing/model': 1.9256981370272115, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.23}) +[2026-06-23 11:09:44,408][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.17963369190692902, 'train/grad_norm_pre_clip': 50.71399688720703, 'timing/data': 0.0005944129079580307, 'timing/model': 1.9222110930131748, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.23}) +[2026-06-23 11:09:48,219][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.09579918533563614, 'train/grad_norm_pre_clip': 36.37229537963867, 'timing/data': 0.0006357769016176462, 'timing/model': 1.937499918974936, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.23}) +[2026-06-23 11:09:52,016][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.08639732748270035, 'train/grad_norm_pre_clip': 50.945064544677734, 'timing/data': 0.0006154139991849661, 'timing/model': 1.9336229419568554, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.23}) +[2026-06-23 11:09:56,013][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.166387140750885, 'train/grad_norm_pre_clip': 36.56332015991211, 'timing/data': 0.0008258000016212463, 'timing/model': 1.933068384998478, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.23}) +[2026-06-23 11:09:59,820][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.11713100969791412, 'train/grad_norm_pre_clip': 49.42546463012695, 'timing/data': 0.0007340930169448256, 'timing/model': 1.9331081989221275, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.23}) +[2026-06-23 11:10:03,634][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.0678788274526596, 'train/grad_norm_pre_clip': 37.04269027709961, 'timing/data': 0.0008157480042427778, 'timing/model': 1.9339517180342227, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.23}) +[2026-06-23 11:10:07,450][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 600 +[2026-06-23 11:12:14,904][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 600 +[2026-06-23 11:12:14,906][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 600 +[2026-06-23 11:13:36,887][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 600 +[2026-06-23 11:13:36,887][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.16479246318340302, 'train/grad_norm_pre_clip': 29.077880859375, 'eval/action_loss/seconds': 127.45040879002772, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07668530941009521, 'eval/latency_2/loss': 0.07668530941009521, 'eval/flappy/loss': 0.07668530941009521, 'eval/flappy/latency_2/loss': 0.07668530941009521, 'eval/flap_precision': 0.9166666666666666, 'eval/flap_recall': 0.9352226720647774, 'eval/flap_f1': 0.9258517034068137, 'eval/flap_rate_model': 0.21, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9258517034068137, 'eval/latency_2/flap_precision': 0.9166666666666666, 'eval/latency_2/flap_recall': 0.9352226720647774, 'eval/latency_2/flap_f1': 0.9258517034068137, 'eval/latency_2/flap_rate_model': 0.21, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9258517034068137, 'eval/action_classification/data_wait_seconds': 1.4272382259368896, 'eval/action_classification/predict_seconds': 80.37405395507812, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.000597206992097199, 'timing/model': 1.93542776000686, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.23}) +[2026-06-23 11:13:40,677][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.13998191058635712, 'train/grad_norm_pre_clip': 30.743988037109375, 'timing/data': 0.0006068740040063858, 'timing/model': 1.9199397900374606, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.23}) +[2026-06-23 11:13:44,453][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.15739955008029938, 'train/grad_norm_pre_clip': 27.310256958007812, 'timing/data': 0.0005943399155512452, 'timing/model': 1.9235225439770147, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.23}) +[2026-06-23 11:13:48,246][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.1828767955303192, 'train/grad_norm_pre_clip': 32.15420913696289, 'timing/data': 0.0007154699414968491, 'timing/model': 1.9249287489801645, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.23}) +[2026-06-23 11:13:52,036][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.16969312727451324, 'train/grad_norm_pre_clip': 30.97175407409668, 'timing/data': 0.0006202469812706113, 'timing/model': 1.9256777520058677, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.23}) +[2026-06-23 11:13:55,821][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.21054041385650635, 'train/grad_norm_pre_clip': 36.124629974365234, 'timing/data': 0.0006006270414218307, 'timing/model': 1.918259844998829, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.23}) +[2026-06-23 11:13:59,605][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.10118164122104645, 'train/grad_norm_pre_clip': 48.99571990966797, 'timing/data': 0.0006103819468989968, 'timing/model': 1.9218691260321066, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.23}) +[2026-06-23 11:14:03,391][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.06561145186424255, 'train/grad_norm_pre_clip': 32.34789276123047, 'timing/data': 0.0006285340059548616, 'timing/model': 1.924952934961766, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.23}) +[2026-06-23 11:14:07,173][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.1727151870727539, 'train/grad_norm_pre_clip': 27.577375411987305, 'timing/data': 0.0005952599458396435, 'timing/model': 1.9247792609967291, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.24}) +[2026-06-23 11:14:10,955][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.07416544109582901, 'train/grad_norm_pre_clip': 42.864776611328125, 'timing/data': 0.0006744159618392587, 'timing/model': 1.914794722921215, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.24}) +[2026-06-23 11:14:14,756][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.1502353996038437, 'train/grad_norm_pre_clip': 60.494346618652344, 'timing/data': 0.0005850939778611064, 'timing/model': 1.932924625929445, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.24}) +[2026-06-23 11:14:18,729][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.1451859176158905, 'train/grad_norm_pre_clip': 80.0167465209961, 'timing/data': 0.0006923000328242779, 'timing/model': 1.92368673090823, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.24}) +[2026-06-23 11:14:22,525][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.028030013665556908, 'train/grad_norm_pre_clip': 21.7481746673584, 'timing/data': 0.0005971179343760014, 'timing/model': 1.9306258299620822, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.24}) +[2026-06-23 11:14:26,315][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.15661437809467316, 'train/grad_norm_pre_clip': 47.20913314819336, 'timing/data': 0.0006269830046221614, 'timing/model': 1.9155927489046007, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.24}) +[2026-06-23 11:14:30,112][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.09669899195432663, 'train/grad_norm_pre_clip': 20.85100746154785, 'timing/data': 0.0006932030664756894, 'timing/model': 1.937694341992028, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.24}) +[2026-06-23 11:14:33,898][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.15773265063762665, 'train/grad_norm_pre_clip': 27.561439514160156, 'timing/data': 0.0005849539302289486, 'timing/model': 1.921304352930747, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.24}) +[2026-06-23 11:14:37,686][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.13168752193450928, 'train/grad_norm_pre_clip': 63.37907409667969, 'timing/data': 0.000593886012211442, 'timing/model': 1.9213833409594372, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.24}) +[2026-06-23 11:14:41,487][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.11241672188043594, 'train/grad_norm_pre_clip': 53.96367263793945, 'timing/data': 0.0006430919747799635, 'timing/model': 1.924093251931481, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.24}) +[2026-06-23 11:14:45,284][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.10464294999837875, 'train/grad_norm_pre_clip': 39.909603118896484, 'timing/data': 0.000667835003696382, 'timing/model': 1.9289479850558564, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.24}) +[2026-06-23 11:14:49,078][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.11130641400814056, 'train/grad_norm_pre_clip': 32.38302993774414, 'timing/data': 0.0008243880001828074, 'timing/model': 1.9171785180224106, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.24}) +[2026-06-23 11:14:52,887][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.1267881542444229, 'train/grad_norm_pre_clip': 30.719776153564453, 'timing/data': 0.0007199079263955355, 'timing/model': 1.9401376850437373, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.24}) +[2026-06-23 11:14:56,682][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.10516071319580078, 'train/grad_norm_pre_clip': 38.78152847290039, 'timing/data': 0.000597139005549252, 'timing/model': 1.9296024850336835, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.24}) +[2026-06-23 11:15:00,481][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.1924998015165329, 'train/grad_norm_pre_clip': 46.54945755004883, 'timing/data': 0.0006771199405193329, 'timing/model': 1.9246629199478775, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.24}) +[2026-06-23 11:15:04,276][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.22655729949474335, 'train/grad_norm_pre_clip': 33.38649368286133, 'timing/data': 0.0006167900282889605, 'timing/model': 1.921405183034949, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.24}) +[2026-06-23 11:15:08,068][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.11525222659111023, 'train/grad_norm_pre_clip': 32.61029815673828, 'timing/data': 0.0006483970209956169, 'timing/model': 1.924141013994813, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.24}) +[2026-06-23 11:15:11,868][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.07163479179143906, 'train/grad_norm_pre_clip': 22.7724666595459, 'timing/data': 0.0006212850566953421, 'timing/model': 1.929577711969614, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.24}) +[2026-06-23 11:15:15,671][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.2773069739341736, 'train/grad_norm_pre_clip': 62.56552505493164, 'timing/data': 0.0006054670084267855, 'timing/model': 1.9222515620058402, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.24}) +[2026-06-23 11:15:19,471][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.12030339241027832, 'train/grad_norm_pre_clip': 26.900869369506836, 'timing/data': 0.0006734050111845136, 'timing/model': 1.9331085759913549, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.24}) +[2026-06-23 11:15:23,270][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.09764087200164795, 'train/grad_norm_pre_clip': 32.77324676513672, 'timing/data': 0.0006376099772751331, 'timing/model': 1.9275356139987707, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.24}) +[2026-06-23 11:15:27,072][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.0788813903927803, 'train/grad_norm_pre_clip': 37.583736419677734, 'timing/data': 0.000670108012855053, 'timing/model': 1.929601976997219, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.24}) +[2026-06-23 11:15:30,862][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.16588397324085236, 'train/grad_norm_pre_clip': 40.134220123291016, 'timing/data': 0.0006135760340839624, 'timing/model': 1.9234562099445611, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.24}) +[2026-06-23 11:15:34,672][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.2228061854839325, 'train/grad_norm_pre_clip': 36.49942398071289, 'timing/data': 0.0006652110023424029, 'timing/model': 1.9301280759973451, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.24}) +[2026-06-23 11:15:38,482][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.17464210093021393, 'train/grad_norm_pre_clip': 69.32160186767578, 'timing/data': 0.0006374310469254851, 'timing/model': 1.9355241630692035, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.24}) +[2026-06-23 11:15:42,290][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.08486474305391312, 'train/grad_norm_pre_clip': 30.13462257385254, 'timing/data': 0.000631702016107738, 'timing/model': 1.9333824620116502, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.24}) +[2026-06-23 11:15:46,099][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.14950469136238098, 'train/grad_norm_pre_clip': 53.18769836425781, 'timing/data': 0.0007041420321911573, 'timing/model': 1.9268161489162594, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.25}) +[2026-06-23 11:15:49,907][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.13266104459762573, 'train/grad_norm_pre_clip': 21.453855514526367, 'timing/data': 0.0005766360554844141, 'timing/model': 1.9395419380161911, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.25}) +[2026-06-23 11:15:53,724][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.04907474294304848, 'train/grad_norm_pre_clip': 23.41739845275879, 'timing/data': 0.0006387890316545963, 'timing/model': 1.9261350070592016, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.25}) +[2026-06-23 11:15:57,528][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.15479838848114014, 'train/grad_norm_pre_clip': 99.39563751220703, 'timing/data': 0.000622092979028821, 'timing/model': 1.9311706009320915, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.25}) +[2026-06-23 11:16:01,336][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.16171862185001373, 'train/grad_norm_pre_clip': 118.29978942871094, 'timing/data': 0.0005787040572613478, 'timing/model': 1.931711375946179, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.25}) +[2026-06-23 11:16:05,146][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.11433718353509903, 'train/grad_norm_pre_clip': 102.3525390625, 'timing/data': 0.0005815550684928894, 'timing/model': 1.931165596935898, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.25}) +[2026-06-23 11:16:08,957][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.25116249918937683, 'train/grad_norm_pre_clip': 65.81737518310547, 'timing/data': 0.0005911999614909291, 'timing/model': 1.93317910307087, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.25}) +[2026-06-23 11:16:12,773][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.3582414984703064, 'train/grad_norm_pre_clip': 151.9957733154297, 'timing/data': 0.000585394911468029, 'timing/model': 1.936810627928935, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.25}) +[2026-06-23 11:16:16,576][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.047027453780174255, 'train/grad_norm_pre_clip': 55.21925354003906, 'timing/data': 0.0005855270428583026, 'timing/model': 1.9308833149261773, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.25}) +[2026-06-23 11:16:20,380][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.13516202569007874, 'train/grad_norm_pre_clip': 28.40399742126465, 'timing/data': 0.0005892430199310184, 'timing/model': 1.9315535940695554, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.25}) +[2026-06-23 11:16:24,185][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.13981762528419495, 'train/grad_norm_pre_clip': 51.29227828979492, 'timing/data': 0.0006360670085996389, 'timing/model': 1.9315336639992893, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.25}) +[2026-06-23 11:16:28,213][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.19856250286102295, 'train/grad_norm_pre_clip': 81.59827423095703, 'timing/data': 0.0005898319650441408, 'timing/model': 2.14713917800691, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.25}) +[2026-06-23 11:16:32,014][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.08575456589460373, 'train/grad_norm_pre_clip': 43.29829788208008, 'timing/data': 0.000785500044003129, 'timing/model': 1.9315664200112224, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.25}) +[2026-06-23 11:16:35,844][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.0893043801188469, 'train/grad_norm_pre_clip': 41.64055633544922, 'timing/data': 0.0006396230310201645, 'timing/model': 1.9431486000539735, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.25}) +[2026-06-23 11:16:39,646][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.19220596551895142, 'train/grad_norm_pre_clip': 66.24632263183594, 'timing/data': 0.0006232579471543431, 'timing/model': 1.932900849962607, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.25}) +[2026-06-23 11:16:43,461][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.2950249910354614, 'train/grad_norm_pre_clip': 51.071048736572266, 'timing/data': 0.0005907939048483968, 'timing/model': 1.928439337061718, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.25}) +[2026-06-23 11:16:47,280][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.17634418606758118, 'train/grad_norm_pre_clip': 42.47968673706055, 'timing/data': 0.0005834749899804592, 'timing/model': 1.9369891409296542, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.25}) +[2026-06-23 11:16:51,082][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.16097810864448547, 'train/grad_norm_pre_clip': 72.44075012207031, 'timing/data': 0.0006075570126995444, 'timing/model': 1.931026228936389, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.25}) +[2026-06-23 11:16:54,889][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.16759133338928223, 'train/grad_norm_pre_clip': 30.458845138549805, 'timing/data': 0.00059243303257972, 'timing/model': 1.9303359009791166, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.25}) +[2026-06-23 11:16:58,757][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.1198187842965126, 'train/grad_norm_pre_clip': 26.374267578125, 'timing/data': 0.000711777014657855, 'timing/model': 1.9871631559217349, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.25}) +[2026-06-23 11:17:02,572][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.13594618439674377, 'train/grad_norm_pre_clip': 36.834598541259766, 'timing/data': 0.0006435359828174114, 'timing/model': 1.9371938069816679, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.25}) +[2026-06-23 11:17:06,381][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.1209486722946167, 'train/grad_norm_pre_clip': 39.781681060791016, 'timing/data': 0.0005813429597765207, 'timing/model': 1.9346694810083136, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.25}) +[2026-06-23 11:17:10,190][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.19340738654136658, 'train/grad_norm_pre_clip': 48.46406173706055, 'timing/data': 0.000597437028773129, 'timing/model': 1.932634275057353, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.25}) +[2026-06-23 11:17:13,989][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.07136326283216476, 'train/grad_norm_pre_clip': 23.43785858154297, 'timing/data': 0.0007704970194026828, 'timing/model': 1.9351679800311103, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.25}) +[2026-06-23 11:17:17,782][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.14355477690696716, 'train/grad_norm_pre_clip': 29.25806427001953, 'timing/data': 0.0005781420040875673, 'timing/model': 1.9211931000463665, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.25}) +[2026-06-23 11:17:21,599][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.22932161390781403, 'train/grad_norm_pre_clip': 36.812255859375, 'timing/data': 0.0005800860235467553, 'timing/model': 1.9390465279575437, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.26}) +[2026-06-23 11:17:25,404][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.1405634731054306, 'train/grad_norm_pre_clip': 33.37419128417969, 'timing/data': 0.0007568419678136706, 'timing/model': 1.9281173238996416, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.26}) +[2026-06-23 11:17:29,227][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.22245629131793976, 'train/grad_norm_pre_clip': 31.738962173461914, 'timing/data': 0.0006280239904299378, 'timing/model': 1.9388904920779169, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.26}) +[2026-06-23 11:17:33,234][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.10958361625671387, 'train/grad_norm_pre_clip': 20.489158630371094, 'timing/data': 0.0007204059511423111, 'timing/model': 2.126977405976504, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.26}) +[2026-06-23 11:17:37,029][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.1821170300245285, 'train/grad_norm_pre_clip': 59.87811279296875, 'timing/data': 0.0005693420534953475, 'timing/model': 1.9188480789307505, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.26}) +[2026-06-23 11:17:40,843][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.1752646118402481, 'train/grad_norm_pre_clip': 32.901458740234375, 'timing/data': 0.000677436008118093, 'timing/model': 1.9348898300668225, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.26}) +[2026-06-23 11:17:44,650][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.07740065455436707, 'train/grad_norm_pre_clip': 39.961002349853516, 'timing/data': 0.0006546389777213335, 'timing/model': 1.9297522159758955, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.26}) +[2026-06-23 11:17:48,444][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.17857515811920166, 'train/grad_norm_pre_clip': 112.62548828125, 'timing/data': 0.0005983990849927068, 'timing/model': 1.9288702580379322, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.26}) +[2026-06-23 11:17:52,255][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.40300339460372925, 'train/grad_norm_pre_clip': 180.65737915039062, 'timing/data': 0.0006192440632730722, 'timing/model': 1.9285047390731052, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.26}) +[2026-06-23 11:17:56,066][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.15610761940479279, 'train/grad_norm_pre_clip': 139.35032653808594, 'timing/data': 0.0007670930353924632, 'timing/model': 1.9464188540587202, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.26}) +[2026-06-23 11:17:59,866][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.18016086518764496, 'train/grad_norm_pre_clip': 83.74555206298828, 'timing/data': 0.0005856609204784036, 'timing/model': 1.9277907679788768, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.26}) +[2026-06-23 11:18:03,686][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.16774670779705048, 'train/grad_norm_pre_clip': 41.8253288269043, 'timing/data': 0.0005928199971094728, 'timing/model': 1.9405796730425209, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.26}) +[2026-06-23 11:18:07,679][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.1750662475824356, 'train/grad_norm_pre_clip': 91.03611755371094, 'timing/data': 0.0006112429546192288, 'timing/model': 1.935987617005594, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.26}) +[2026-06-23 11:18:11,493][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.36961907148361206, 'train/grad_norm_pre_clip': 131.91148376464844, 'timing/data': 0.000847143935970962, 'timing/model': 1.9309350220719352, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.26}) +[2026-06-23 11:18:15,306][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.20200155675411224, 'train/grad_norm_pre_clip': 56.699615478515625, 'timing/data': 0.0005815370241180062, 'timing/model': 1.9410936509957537, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.26}) +[2026-06-23 11:18:19,117][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.20712068676948547, 'train/grad_norm_pre_clip': 33.15908432006836, 'timing/data': 0.0005943340947851539, 'timing/model': 1.93452890496701, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.26}) +[2026-06-23 11:18:22,937][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.2118973284959793, 'train/grad_norm_pre_clip': 57.09730911254883, 'timing/data': 0.0006385419983416796, 'timing/model': 1.9185163559159264, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.26}) +[2026-06-23 11:18:26,754][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.13107377290725708, 'train/grad_norm_pre_clip': 57.26274108886719, 'timing/data': 0.0006855819374322891, 'timing/model': 1.9422445910749957, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.26}) +[2026-06-23 11:18:30,570][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.14057154953479767, 'train/grad_norm_pre_clip': 31.450908660888672, 'timing/data': 0.0008666629437357187, 'timing/model': 1.9423278720350936, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.26}) +[2026-06-23 11:18:34,381][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.10133843868970871, 'train/grad_norm_pre_clip': 22.620193481445312, 'timing/data': 0.0005773360608145595, 'timing/model': 1.9322790580336004, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.26}) +[2026-06-23 11:18:38,396][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.05173535272479057, 'train/grad_norm_pre_clip': 70.18115234375, 'timing/data': 0.0005828060675412416, 'timing/model': 2.1298207079526037, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.26}) +[2026-06-23 11:18:42,203][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.10546787083148956, 'train/grad_norm_pre_clip': 26.303895950317383, 'timing/data': 0.0007838059682399035, 'timing/model': 1.9368224909994751, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.26}) +[2026-06-23 11:18:46,004][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.1676739752292633, 'train/grad_norm_pre_clip': 33.38322448730469, 'timing/data': 0.0005795719334855676, 'timing/model': 1.933424917049706, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.26}) +[2026-06-23 11:18:49,816][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.11571558564901352, 'train/grad_norm_pre_clip': 41.95332717895508, 'timing/data': 0.0008991379290819168, 'timing/model': 1.9308002959005535, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.26}) +[2026-06-23 11:18:53,624][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.1275286078453064, 'train/grad_norm_pre_clip': 57.610225677490234, 'timing/data': 0.0006175190210342407, 'timing/model': 1.9309426499530673, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.26}) +[2026-06-23 11:18:57,441][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.360216349363327, 'train/grad_norm_pre_clip': 66.47672271728516, 'timing/data': 0.0005890040192753077, 'timing/model': 1.93272070994135, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.26}) +[2026-06-23 11:19:01,287][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.19707392156124115, 'train/grad_norm_pre_clip': 35.49927520751953, 'timing/data': 0.0005830090958625078, 'timing/model': 1.9590587329585105, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.27}) +[2026-06-23 11:19:05,097][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.15446843206882477, 'train/grad_norm_pre_clip': 41.05086135864258, 'timing/data': 0.000596686964854598, 'timing/model': 1.9400491459527984, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.27}) +[2026-06-23 11:19:08,912][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.16370877623558044, 'train/grad_norm_pre_clip': 42.81385040283203, 'timing/data': 0.0007857889868319035, 'timing/model': 1.9391700860578567, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.27}) +[2026-06-23 11:19:12,914][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.09788436442613602, 'train/grad_norm_pre_clip': 25.323551177978516, 'timing/data': 0.0008463079575449228, 'timing/model': 1.9337363559752703, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.27}) +[2026-06-23 11:19:16,732][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.09257710725069046, 'train/grad_norm_pre_clip': 32.56354522705078, 'timing/data': 0.0005767600378021598, 'timing/model': 1.931768600945361, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.27}) +[2026-06-23 11:19:20,534][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.04623261094093323, 'train/grad_norm_pre_clip': 49.477542877197266, 'timing/data': 0.0005865710554644465, 'timing/model': 1.932982413098216, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.27}) +[2026-06-23 11:19:24,413][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.11590488255023956, 'train/grad_norm_pre_clip': 30.11919593811035, 'timing/data': 0.0005881199613213539, 'timing/model': 1.9976036960724741, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.27}) +[2026-06-23 11:19:28,214][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.29721489548683167, 'train/grad_norm_pre_clip': 52.663455963134766, 'timing/data': 0.000613198964856565, 'timing/model': 1.9335253529716283, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.27}) +[2026-06-23 11:19:32,023][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.16538292169570923, 'train/grad_norm_pre_clip': 39.56028747558594, 'timing/data': 0.0006063879700377584, 'timing/model': 1.9290596869541332, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.27}) +[2026-06-23 11:19:35,838][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.05171273276209831, 'train/grad_norm_pre_clip': 20.9268798828125, 'timing/data': 0.00058846699539572, 'timing/model': 1.9393632699502632, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.27}) +[2026-06-23 11:19:39,637][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.2411211133003235, 'train/grad_norm_pre_clip': 133.60385131835938, 'timing/data': 0.0006008729105815291, 'timing/model': 1.931997537962161, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.27}) +[2026-06-23 11:19:43,608][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.13761723041534424, 'train/grad_norm_pre_clip': 86.67842864990234, 'timing/data': 0.0005973140941932797, 'timing/model': 2.1027299910783768, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.27}) +[2026-06-23 11:19:47,428][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.1645556390285492, 'train/grad_norm_pre_clip': 103.09575653076172, 'timing/data': 0.0007700610440224409, 'timing/model': 1.935360433999449, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.27}) +[2026-06-23 11:19:51,223][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.09068874269723892, 'train/grad_norm_pre_clip': 25.77231216430664, 'timing/data': 0.0005844250554218888, 'timing/model': 1.9183027879334986, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.27}) +[2026-06-23 11:19:55,019][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.22508442401885986, 'train/grad_norm_pre_clip': 57.12493133544922, 'timing/data': 0.0005997689440846443, 'timing/model': 1.926964296028018, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.27}) +[2026-06-23 11:19:58,828][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 700 +[2026-06-23 11:22:08,294][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 700 +[2026-06-23 11:22:08,296][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 700 +[2026-06-23 11:23:29,600][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 700 +[2026-06-23 11:23:29,601][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.09566500782966614, 'train/grad_norm_pre_clip': 56.525142669677734, 'eval/action_loss/seconds': 129.46408542303834, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07406891137361526, 'eval/latency_2/loss': 0.07406891137361526, 'eval/flappy/loss': 0.07406891137361526, 'eval/flappy/latency_2/loss': 0.07406891137361526, 'eval/flap_precision': 0.9022842639593909, 'eval/flap_recall': 0.9595141700404858, 'eval/flap_f1': 0.9300196206671025, 'eval/flap_rate_model': 0.21888888888888888, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9300196206671025, 'eval/latency_2/flap_precision': 0.9022842639593909, 'eval/latency_2/flap_recall': 0.9595141700404858, 'eval/latency_2/flap_f1': 0.9300196206671025, 'eval/latency_2/flap_rate_model': 0.21888888888888888, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9300196206671025, 'eval/action_classification/data_wait_seconds': 1.4611772298812866, 'eval/action_classification/predict_seconds': 79.65455627441406, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0007758020656183362, 'timing/model': 1.934728573076427, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.27}) +[2026-06-23 11:23:33,426][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.10588137805461884, 'train/grad_norm_pre_clip': 39.26112747192383, 'timing/data': 0.0005952740320935845, 'timing/model': 1.923787287902087, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.27}) +[2026-06-23 11:23:37,217][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.13082322478294373, 'train/grad_norm_pre_clip': 45.85783004760742, 'timing/data': 0.0006441900040954351, 'timing/model': 1.927936736959964, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.27}) +[2026-06-23 11:23:41,006][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.15144562721252441, 'train/grad_norm_pre_clip': 91.66861724853516, 'timing/data': 0.0007278049597516656, 'timing/model': 1.9232299680588767, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.27}) +[2026-06-23 11:23:44,784][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.12668655812740326, 'train/grad_norm_pre_clip': 21.237346649169922, 'timing/data': 0.0007546850247308612, 'timing/model': 1.9199244069168344, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.27}) +[2026-06-23 11:23:48,574][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.05090663582086563, 'train/grad_norm_pre_clip': 22.4554500579834, 'timing/data': 0.0007810399401932955, 'timing/model': 1.9222715010400862, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.27}) +[2026-06-23 11:23:52,371][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.1281941831111908, 'train/grad_norm_pre_clip': 37.75767135620117, 'timing/data': 0.0005924709839746356, 'timing/model': 1.9300350020639598, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.27}) +[2026-06-23 11:23:56,173][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.04880579933524132, 'train/grad_norm_pre_clip': 82.61817169189453, 'timing/data': 0.0005974179366603494, 'timing/model': 1.9353191269328818, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.27}) +[2026-06-23 11:23:59,966][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.1198803260922432, 'train/grad_norm_pre_clip': 40.79861831665039, 'timing/data': 0.0006542159244418144, 'timing/model': 1.920348175917752, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.27}) +[2026-06-23 11:24:03,764][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.12254825234413147, 'train/grad_norm_pre_clip': 28.78398895263672, 'timing/data': 0.0005872569745406508, 'timing/model': 1.938368092989549, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.27}) +[2026-06-23 11:24:07,553][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.17858462035655975, 'train/grad_norm_pre_clip': 58.788536071777344, 'timing/data': 0.0006588390097022057, 'timing/model': 1.9180709660286084, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.27}) +[2026-06-23 11:24:11,325][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.19944991171360016, 'train/grad_norm_pre_clip': 57.163963317871094, 'timing/data': 0.0006100530736148357, 'timing/model': 1.9140714419772848, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.28}) +[2026-06-23 11:24:15,112][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.17939278483390808, 'train/grad_norm_pre_clip': 53.14278793334961, 'timing/data': 0.0008144909515976906, 'timing/model': 1.926506862975657, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.28}) +[2026-06-23 11:24:18,912][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.20156149566173553, 'train/grad_norm_pre_clip': 52.32689666748047, 'timing/data': 0.0006707020802423358, 'timing/model': 1.9315608160104603, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.28}) +[2026-06-23 11:24:22,711][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.0851893424987793, 'train/grad_norm_pre_clip': 30.98761558532715, 'timing/data': 0.000861346023157239, 'timing/model': 1.9303299100138247, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.28}) +[2026-06-23 11:24:26,511][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.0808313712477684, 'train/grad_norm_pre_clip': 28.744365692138672, 'timing/data': 0.0007360100280493498, 'timing/model': 1.9294385750545189, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.28}) +[2026-06-23 11:24:30,311][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.10388776659965515, 'train/grad_norm_pre_clip': 22.6999568939209, 'timing/data': 0.0006118409801274538, 'timing/model': 1.9166966209886596, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.28}) +[2026-06-23 11:24:34,101][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.17418494820594788, 'train/grad_norm_pre_clip': 28.291627883911133, 'timing/data': 0.0006273160688579082, 'timing/model': 1.914888064027764, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.28}) +[2026-06-23 11:24:37,919][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.14176177978515625, 'train/grad_norm_pre_clip': 48.20193099975586, 'timing/data': 0.0006201919168233871, 'timing/model': 1.9338232240406796, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.28}) +[2026-06-23 11:24:41,754][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.07789359986782074, 'train/grad_norm_pre_clip': 35.957515716552734, 'timing/data': 0.0007239819969981909, 'timing/model': 1.9375594940502197, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.28}) +[2026-06-23 11:24:45,553][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.10544273257255554, 'train/grad_norm_pre_clip': 37.073570251464844, 'timing/data': 0.0006157959578558803, 'timing/model': 1.9262970520649105, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.28}) +[2026-06-23 11:24:49,345][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.16248054802417755, 'train/grad_norm_pre_clip': 49.91925048828125, 'timing/data': 0.0005934549262747169, 'timing/model': 1.918106768047437, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.28}) +[2026-06-23 11:24:53,136][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.1745551973581314, 'train/grad_norm_pre_clip': 35.38670349121094, 'timing/data': 0.0005963389994576573, 'timing/model': 1.920073573011905, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.28}) +[2026-06-23 11:24:56,936][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.13046672940254211, 'train/grad_norm_pre_clip': 76.03983306884766, 'timing/data': 0.0008590680081397295, 'timing/model': 1.9262436060234904, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.28}) +[2026-06-23 11:25:00,736][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.11536065489053726, 'train/grad_norm_pre_clip': 70.69633483886719, 'timing/data': 0.0006388210458680987, 'timing/model': 1.9201188910519704, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.28}) +[2026-06-23 11:25:04,529][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.0843866840004921, 'train/grad_norm_pre_clip': 42.51234817504883, 'timing/data': 0.0006037199636921287, 'timing/model': 1.926330326939933, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.28}) +[2026-06-23 11:25:08,335][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.14758996665477753, 'train/grad_norm_pre_clip': 33.245849609375, 'timing/data': 0.0005874469643458724, 'timing/model': 1.92808787105605, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.28}) +[2026-06-23 11:25:12,329][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.15523554384708405, 'train/grad_norm_pre_clip': 157.02012634277344, 'timing/data': 0.0005968629848212004, 'timing/model': 1.9321952529717237, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.28}) +[2026-06-23 11:25:16,136][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.08368714898824692, 'train/grad_norm_pre_clip': 28.111562728881836, 'timing/data': 0.0005893700290471315, 'timing/model': 1.9366266219876707, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.28}) +[2026-06-23 11:25:19,925][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.11463794112205505, 'train/grad_norm_pre_clip': 28.30173110961914, 'timing/data': 0.0006706790300086141, 'timing/model': 1.9178654489805922, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.28}) +[2026-06-23 11:25:23,726][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.15559199452400208, 'train/grad_norm_pre_clip': 30.09906005859375, 'timing/data': 0.0006382069550454617, 'timing/model': 1.9312598870601505, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.28}) +[2026-06-23 11:25:27,520][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.09052660316228867, 'train/grad_norm_pre_clip': 55.41730880737305, 'timing/data': 0.0007429880788549781, 'timing/model': 1.9294238151051104, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.28}) +[2026-06-23 11:25:31,319][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.1825675219297409, 'train/grad_norm_pre_clip': 60.577816009521484, 'timing/data': 0.0006087160436436534, 'timing/model': 1.923242133925669, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.28}) +[2026-06-23 11:25:35,128][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.07438007742166519, 'train/grad_norm_pre_clip': 35.642723083496094, 'timing/data': 0.0007963200332596898, 'timing/model': 1.9296222049742937, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.28}) +[2026-06-23 11:25:38,946][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.14489436149597168, 'train/grad_norm_pre_clip': 48.78727340698242, 'timing/data': 0.0006199210183694959, 'timing/model': 1.9419163899729028, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.28}) +[2026-06-23 11:25:42,832][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.2724404036998749, 'train/grad_norm_pre_clip': 87.82877349853516, 'timing/data': 0.0006166819948703051, 'timing/model': 2.0092091620899737, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.28}) +[2026-06-23 11:25:46,650][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.15151718258857727, 'train/grad_norm_pre_clip': 54.19339370727539, 'timing/data': 0.0005871220491826534, 'timing/model': 1.9460759300272912, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.28}) +[2026-06-23 11:25:50,450][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.19012711942195892, 'train/grad_norm_pre_clip': 50.5926628112793, 'timing/data': 0.0005851420573890209, 'timing/model': 1.9255656390450895, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.29}) +[2026-06-23 11:25:54,252][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.08339358121156693, 'train/grad_norm_pre_clip': 21.634830474853516, 'timing/data': 0.0006107119843363762, 'timing/model': 1.9246241219807416, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.29}) +[2026-06-23 11:25:58,064][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.058943621814250946, 'train/grad_norm_pre_clip': 66.42554473876953, 'timing/data': 0.0007780189625918865, 'timing/model': 1.9386390059953555, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.29}) +[2026-06-23 11:26:01,883][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.2210906594991684, 'train/grad_norm_pre_clip': 55.97871780395508, 'timing/data': 0.0006498949369415641, 'timing/model': 1.9436294459737837, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.29}) +[2026-06-23 11:26:05,676][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.1846979558467865, 'train/grad_norm_pre_clip': 45.995079040527344, 'timing/data': 0.0008673040429130197, 'timing/model': 1.9196270699612796, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.29}) +[2026-06-23 11:26:09,484][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.12324346601963043, 'train/grad_norm_pre_clip': 52.78456115722656, 'timing/data': 0.0006772879278287292, 'timing/model': 1.9353075000690296, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.29}) +[2026-06-23 11:26:13,289][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.17831841111183167, 'train/grad_norm_pre_clip': 67.85419464111328, 'timing/data': 0.0006568599492311478, 'timing/model': 1.927926383097656, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.29}) +[2026-06-23 11:26:17,151][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.2076224386692047, 'train/grad_norm_pre_clip': 32.20912551879883, 'timing/data': 0.0006435969844460487, 'timing/model': 1.9295416329987347, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.29}) +[2026-06-23 11:26:20,961][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.13018764555454254, 'train/grad_norm_pre_clip': 25.212486267089844, 'timing/data': 0.0006149809341877699, 'timing/model': 1.931901113013737, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.29}) +[2026-06-23 11:26:24,779][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.07629510760307312, 'train/grad_norm_pre_clip': 40.64555740356445, 'timing/data': 0.0005820479709655046, 'timing/model': 1.932328916969709, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.29}) +[2026-06-23 11:26:28,582][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.12103261798620224, 'train/grad_norm_pre_clip': 28.53596305847168, 'timing/data': 0.0006210770225152373, 'timing/model': 1.9284496679902077, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.29}) +[2026-06-23 11:26:32,383][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.18006587028503418, 'train/grad_norm_pre_clip': 59.6422119140625, 'timing/data': 0.0006050689844414592, 'timing/model': 1.9276720229536295, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.29}) +[2026-06-23 11:26:36,189][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.1434718519449234, 'train/grad_norm_pre_clip': 22.90155601501465, 'timing/data': 0.0006154630100354552, 'timing/model': 1.9385349790100008, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.29}) +[2026-06-23 11:27:25,959][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 750 +[2026-06-23 11:27:25,960][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.09174332767724991, 'train/grad_norm_pre_clip': 42.31746292114258, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 26.359999999999825, 'rl_games_eval/mid_train/mean_length': 229.6, 'rl_games_eval/mid_train/std_reward': 34.903329354088484, 'rl_games_eval/mid_train/std_length': 280.27172529529264, 'rl_games_eval/mid_train/macro_mean_reward': 26.359999999999825, 'rl_games_eval/mid_train/macro_mean_length': 229.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 26.359999999999825, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 229.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 34.903329354088484, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 280.27172529529264, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 26.359999999999825, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005830890731886029, 'timing/model': 1.9558237290475518, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.29}) +[2026-06-23 11:27:29,908][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.1698167324066162, 'train/grad_norm_pre_clip': 30.41748809814453, 'timing/data': 0.0006219310453161597, 'timing/model': 2.0923480469500646, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.29}) +[2026-06-23 11:27:33,659][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.09898639470338821, 'train/grad_norm_pre_clip': 22.140954971313477, 'timing/data': 0.0005852959584444761, 'timing/model': 1.9004226590041071, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.29}) +[2026-06-23 11:27:37,432][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.09822208434343338, 'train/grad_norm_pre_clip': 25.285165786743164, 'timing/data': 0.0006204459350556135, 'timing/model': 1.920600662007928, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.29}) +[2026-06-23 11:27:41,213][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.12781071662902832, 'train/grad_norm_pre_clip': 56.59786605834961, 'timing/data': 0.0006185220554471016, 'timing/model': 1.9177103850524873, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.29}) +[2026-06-23 11:27:44,981][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.1323583573102951, 'train/grad_norm_pre_clip': 31.24090576171875, 'timing/data': 0.0009271560702472925, 'timing/model': 1.9097836510045454, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.29}) +[2026-06-23 11:27:48,756][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.23876874148845673, 'train/grad_norm_pre_clip': 41.530460357666016, 'timing/data': 0.0008045739959925413, 'timing/model': 1.9161522820359096, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.29}) +[2026-06-23 11:27:52,537][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.10729148983955383, 'train/grad_norm_pre_clip': 46.27531433105469, 'timing/data': 0.0005985579919070005, 'timing/model': 1.9122678270796314, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.29}) +[2026-06-23 11:27:56,316][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.11044953018426895, 'train/grad_norm_pre_clip': 52.371856689453125, 'timing/data': 0.0008508460596203804, 'timing/model': 1.9206078168936074, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.29}) +[2026-06-23 11:28:00,112][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.10362639278173447, 'train/grad_norm_pre_clip': 56.0772590637207, 'timing/data': 0.0005814239848405123, 'timing/model': 1.9274596779141575, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.29}) +[2026-06-23 11:28:04,083][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.10173039138317108, 'train/grad_norm_pre_clip': 32.268516540527344, 'timing/data': 0.0005908090388402343, 'timing/model': 1.925062490045093, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.29}) +[2026-06-23 11:28:07,863][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.08828993141651154, 'train/grad_norm_pre_clip': 53.149009704589844, 'timing/data': 0.0005949890473857522, 'timing/model': 1.920496583916247, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.29}) +[2026-06-23 11:28:11,666][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.2135924994945526, 'train/grad_norm_pre_clip': 60.28966522216797, 'timing/data': 0.0005879080854356289, 'timing/model': 1.9331655889982358, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.29}) +[2026-06-23 11:28:15,454][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.10144779086112976, 'train/grad_norm_pre_clip': 57.31407928466797, 'timing/data': 0.0006094439886510372, 'timing/model': 1.921073195990175, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.3}) +[2026-06-23 11:28:19,239][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.12615841627120972, 'train/grad_norm_pre_clip': 46.99612045288086, 'timing/data': 0.0006469929357990623, 'timing/model': 1.9176437070127577, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.3}) +[2026-06-23 11:28:23,057][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.15961304306983948, 'train/grad_norm_pre_clip': 75.28718566894531, 'timing/data': 0.000674478942528367, 'timing/model': 1.9485801269765943, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.3}) +[2026-06-23 11:28:26,853][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.07160788029432297, 'train/grad_norm_pre_clip': 66.71688842773438, 'timing/data': 0.0007646589074283838, 'timing/model': 1.9247395159909502, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.3}) +[2026-06-23 11:28:30,633][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.06251045316457748, 'train/grad_norm_pre_clip': 29.605377197265625, 'timing/data': 0.0008097729878500104, 'timing/model': 1.9231412130175158, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.3}) +[2026-06-23 11:28:34,610][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.15219800174236298, 'train/grad_norm_pre_clip': 53.29015350341797, 'timing/data': 0.0006137460004538298, 'timing/model': 2.1047973870299757, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.3}) +[2026-06-23 11:28:38,416][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.1186499297618866, 'train/grad_norm_pre_clip': 65.3448257446289, 'timing/data': 0.0005850610323250294, 'timing/model': 1.9339361489983276, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.3}) +[2026-06-23 11:28:42,217][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.16365866363048553, 'train/grad_norm_pre_clip': 114.05503845214844, 'timing/data': 0.0005850209854543209, 'timing/model': 1.9334601170849055, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.3}) +[2026-06-23 11:28:46,016][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.24546903371810913, 'train/grad_norm_pre_clip': 79.22187042236328, 'timing/data': 0.0009615729795768857, 'timing/model': 1.9278100050287321, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.3}) +[2026-06-23 11:28:49,828][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.2683643698692322, 'train/grad_norm_pre_clip': 59.059871673583984, 'timing/data': 0.000868443981744349, 'timing/model': 1.9359016109956428, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.3}) +[2026-06-23 11:28:53,618][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.16816940903663635, 'train/grad_norm_pre_clip': 50.917259216308594, 'timing/data': 0.0008334389422088861, 'timing/model': 1.9261357359355316, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.3}) +[2026-06-23 11:28:57,420][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.0705762654542923, 'train/grad_norm_pre_clip': 25.812599182128906, 'timing/data': 0.0009444019524380565, 'timing/model': 1.926363738020882, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.3}) +[2026-06-23 11:29:01,218][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.07814563065767288, 'train/grad_norm_pre_clip': 32.75802230834961, 'timing/data': 0.0008744699880480766, 'timing/model': 1.934244908974506, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.3}) +[2026-06-23 11:29:05,018][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.09777826815843582, 'train/grad_norm_pre_clip': 32.7602424621582, 'timing/data': 0.0006667249836027622, 'timing/model': 1.9281759019941092, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.3}) +[2026-06-23 11:29:08,813][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.1726011484861374, 'train/grad_norm_pre_clip': 46.51925277709961, 'timing/data': 0.0005834890762344003, 'timing/model': 1.9300643779570237, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.3}) +[2026-06-23 11:29:12,619][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.13811463117599487, 'train/grad_norm_pre_clip': 91.17816162109375, 'timing/data': 0.0006035659462213516, 'timing/model': 1.9349107400048524, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.3}) +[2026-06-23 11:29:16,421][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.1270521581172943, 'train/grad_norm_pre_clip': 29.05409812927246, 'timing/data': 0.0006851539947092533, 'timing/model': 1.9319317180197686, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.3}) +[2026-06-23 11:29:20,264][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.18872056901454926, 'train/grad_norm_pre_clip': 937.61767578125, 'timing/data': 0.000588580034673214, 'timing/model': 1.9705271299462765, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.3}) +[2026-06-23 11:29:24,071][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.089325450360775, 'train/grad_norm_pre_clip': 20.856544494628906, 'timing/data': 0.0008664439665153623, 'timing/model': 1.935913326102309, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.3}) +[2026-06-23 11:29:27,875][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.07087179273366928, 'train/grad_norm_pre_clip': 100.28082275390625, 'timing/data': 0.0005857039941474795, 'timing/model': 1.931828440981917, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.3}) +[2026-06-23 11:29:31,668][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.05537153407931328, 'train/grad_norm_pre_clip': 44.491310119628906, 'timing/data': 0.000584061024710536, 'timing/model': 1.9237754710484296, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.3}) +[2026-06-23 11:29:35,465][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.15853765606880188, 'train/grad_norm_pre_clip': 67.55530548095703, 'timing/data': 0.0005981939611956477, 'timing/model': 1.9240840829443187, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.3}) +[2026-06-23 11:29:39,307][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.10771267861127853, 'train/grad_norm_pre_clip': 40.58848190307617, 'timing/data': 0.0006037199636921287, 'timing/model': 1.965023698983714, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.3}) +[2026-06-23 11:29:43,275][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.1014641672372818, 'train/grad_norm_pre_clip': 54.08246994018555, 'timing/data': 0.0005940550472587347, 'timing/model': 1.9195074869785458, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.3}) +[2026-06-23 11:29:47,087][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.07752961665391922, 'train/grad_norm_pre_clip': 26.007755279541016, 'timing/data': 0.0007501039654016495, 'timing/model': 1.9293810549424961, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.3}) +[2026-06-23 11:29:50,895][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.12905153632164001, 'train/grad_norm_pre_clip': 32.06553268432617, 'timing/data': 0.0007975490298122168, 'timing/model': 1.9350206339731812, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.3}) +[2026-06-23 11:29:54,697][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.08747085928916931, 'train/grad_norm_pre_clip': 49.19709777832031, 'timing/data': 0.0006142600905150175, 'timing/model': 1.933658775058575, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.31}) +[2026-06-23 11:29:58,487][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.23887981474399567, 'train/grad_norm_pre_clip': 51.34178924560547, 'timing/data': 0.0005881061078980565, 'timing/model': 1.921738795004785, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.31}) +[2026-06-23 11:30:02,293][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.15236632525920868, 'train/grad_norm_pre_clip': 77.12291717529297, 'timing/data': 0.0006024340400472283, 'timing/model': 1.9321072259917855, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.31}) +[2026-06-23 11:30:06,092][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.11594907194375992, 'train/grad_norm_pre_clip': 26.265043258666992, 'timing/data': 0.0006117390003055334, 'timing/model': 1.922351537970826, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.31}) +[2026-06-23 11:30:09,903][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.15313157439231873, 'train/grad_norm_pre_clip': 43.29813003540039, 'timing/data': 0.0007477989420294762, 'timing/model': 1.9377634129486978, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.31}) +[2026-06-23 11:30:13,710][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.056456513702869415, 'train/grad_norm_pre_clip': 41.20618438720703, 'timing/data': 0.0005744840018451214, 'timing/model': 1.9269656189717352, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.31}) +[2026-06-23 11:30:17,510][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.18372289836406708, 'train/grad_norm_pre_clip': 40.56013870239258, 'timing/data': 0.0008355779573321342, 'timing/model': 1.9314798400737345, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.31}) +[2026-06-23 11:30:21,316][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.12561312317848206, 'train/grad_norm_pre_clip': 40.87350845336914, 'timing/data': 0.0005908990278840065, 'timing/model': 1.9314782120054588, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.31}) +[2026-06-23 11:30:25,108][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.09375397115945816, 'train/grad_norm_pre_clip': 27.840618133544922, 'timing/data': 0.000593357952311635, 'timing/model': 1.9279078759718686, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.31}) +[2026-06-23 11:30:28,915][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.10481755435466766, 'train/grad_norm_pre_clip': 88.11165618896484, 'timing/data': 0.0005775929894298315, 'timing/model': 1.9328074359800667, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.31}) +[2026-06-23 11:30:32,716][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.158566415309906, 'train/grad_norm_pre_clip': 24.835023880004883, 'timing/data': 0.0006073139375075698, 'timing/model': 1.9208409480052069, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.31}) +[2026-06-23 11:30:36,519][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 800 +[2026-06-23 11:32:46,240][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 800 +[2026-06-23 11:32:46,242][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 800 +[2026-06-23 11:34:07,303][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 800 +[2026-06-23 11:34:07,304][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.10231868922710419, 'train/grad_norm_pre_clip': 27.71430015563965, 'eval/action_loss/seconds': 129.71830696996767, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07197540998458862, 'eval/latency_2/loss': 0.07197540998458862, 'eval/flappy/loss': 0.07197540998458862, 'eval/flappy/latency_2/loss': 0.07197540998458862, 'eval/flap_precision': 0.9588150289017341, 'eval/flap_recall': 0.8954116059379217, 'eval/flap_f1': 0.9260293091416609, 'eval/flap_rate_model': 0.1922222222222222, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9260293091416608, 'eval/latency_2/flap_precision': 0.9588150289017341, 'eval/latency_2/flap_recall': 0.8954116059379217, 'eval/latency_2/flap_f1': 0.9260293091416609, 'eval/latency_2/flap_rate_model': 0.1922222222222222, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9260293091416608, 'eval/action_classification/data_wait_seconds': 1.220320463180542, 'eval/action_classification/predict_seconds': 79.67793273925781, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.000601805979385972, 'timing/model': 1.9325881210388616, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.31}) +[2026-06-23 11:34:11,086][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.09434399008750916, 'train/grad_norm_pre_clip': 64.8160629272461, 'timing/data': 0.0006189659470692277, 'timing/model': 1.918479428975843, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.31}) +[2026-06-23 11:34:14,866][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.21801233291625977, 'train/grad_norm_pre_clip': 97.22103881835938, 'timing/data': 0.0006826519966125488, 'timing/model': 1.9201003590133041, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.31}) +[2026-06-23 11:34:18,685][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.08936735987663269, 'train/grad_norm_pre_clip': 51.84877395629883, 'timing/data': 0.0005924130091443658, 'timing/model': 1.9529653569916263, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.31}) +[2026-06-23 11:34:22,454][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.15150639414787292, 'train/grad_norm_pre_clip': 101.47296905517578, 'timing/data': 0.0007812440162524581, 'timing/model': 1.9169782230164856, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.31}) +[2026-06-23 11:34:26,250][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.26320475339889526, 'train/grad_norm_pre_clip': 102.24642944335938, 'timing/data': 0.0005914610810577869, 'timing/model': 1.9220263919560239, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.31}) +[2026-06-23 11:34:30,042][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.15817120671272278, 'train/grad_norm_pre_clip': 78.74600219726562, 'timing/data': 0.0008133840747177601, 'timing/model': 1.9313611080870032, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.31}) +[2026-06-23 11:34:33,830][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.060641106218099594, 'train/grad_norm_pre_clip': 30.539608001708984, 'timing/data': 0.0006908230716362596, 'timing/model': 1.9217833779985085, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.31}) +[2026-06-23 11:34:37,747][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.06572999060153961, 'train/grad_norm_pre_clip': 23.45412254333496, 'timing/data': 0.0006727359723299742, 'timing/model': 2.0493826150195673, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.31}) +[2026-06-23 11:34:41,528][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.19097423553466797, 'train/grad_norm_pre_clip': 104.48956298828125, 'timing/data': 0.0005962339928373694, 'timing/model': 1.9177958420477808, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.31}) +[2026-06-23 11:34:45,323][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.12158645689487457, 'train/grad_norm_pre_clip': 46.02433776855469, 'timing/data': 0.0005881610559299588, 'timing/model': 1.9383546740282327, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.31}) +[2026-06-23 11:34:49,149][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.1986236572265625, 'train/grad_norm_pre_clip': 47.279701232910156, 'timing/data': 0.0006299599772319198, 'timing/model': 1.9194282969692722, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.31}) +[2026-06-23 11:34:52,939][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.07181049883365631, 'train/grad_norm_pre_clip': 51.67914962768555, 'timing/data': 0.0006591998971998692, 'timing/model': 1.9162926930002868, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.31}) +[2026-06-23 11:34:56,729][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.2549470067024231, 'train/grad_norm_pre_clip': 77.20549774169922, 'timing/data': 0.0007829980459064245, 'timing/model': 1.921748491935432, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.31}) +[2026-06-23 11:35:00,531][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.2511933743953705, 'train/grad_norm_pre_clip': 41.04201126098633, 'timing/data': 0.0006124059436842799, 'timing/model': 1.9300901420647278, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.32}) +[2026-06-23 11:35:04,319][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.1357150673866272, 'train/grad_norm_pre_clip': 27.146150588989258, 'timing/data': 0.000582238077186048, 'timing/model': 1.9204979160567746, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.32}) +[2026-06-23 11:35:08,160][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.14324311912059784, 'train/grad_norm_pre_clip': 42.48041915893555, 'timing/data': 0.0006652739830315113, 'timing/model': 1.9710824510548264, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.32}) +[2026-06-23 11:35:11,954][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.05215523764491081, 'train/grad_norm_pre_clip': 28.820098876953125, 'timing/data': 0.0006127100205048919, 'timing/model': 1.9229440089548007, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.32}) +[2026-06-23 11:35:15,743][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.12021850049495697, 'train/grad_norm_pre_clip': 49.13811492919922, 'timing/data': 0.000591909047216177, 'timing/model': 1.9196506779408082, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.32}) +[2026-06-23 11:35:19,536][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.1458420604467392, 'train/grad_norm_pre_clip': 28.895418167114258, 'timing/data': 0.0005983259761705995, 'timing/model': 1.9265068910317495, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.32}) +[2026-06-23 11:35:23,318][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.12581181526184082, 'train/grad_norm_pre_clip': 31.510643005371094, 'timing/data': 0.000603104941546917, 'timing/model': 1.9137228360632434, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.32}) +[2026-06-23 11:35:27,122][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.1291419118642807, 'train/grad_norm_pre_clip': 65.841064453125, 'timing/data': 0.0006082260515540838, 'timing/model': 1.9366061959881335, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.32}) +[2026-06-23 11:35:30,926][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.07000572234392166, 'train/grad_norm_pre_clip': 23.04427719116211, 'timing/data': 0.0006019689608365297, 'timing/model': 1.9334086090093479, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.32}) +[2026-06-23 11:35:34,717][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.0865403562784195, 'train/grad_norm_pre_clip': 34.10552215576172, 'timing/data': 0.0006148930406197906, 'timing/model': 1.9214102450059727, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.32}) +[2026-06-23 11:35:38,529][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.26576581597328186, 'train/grad_norm_pre_clip': 69.00727081298828, 'timing/data': 0.0005721220513805747, 'timing/model': 1.9302429630188271, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.32}) +[2026-06-23 11:35:42,443][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.12272190302610397, 'train/grad_norm_pre_clip': 28.34724235534668, 'timing/data': 0.0006022639572620392, 'timing/model': 2.0458965419093147, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.32}) +[2026-06-23 11:35:46,257][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.06357736140489578, 'train/grad_norm_pre_clip': 42.52311325073242, 'timing/data': 0.0005893911002203822, 'timing/model': 1.9336338139837608, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.32}) +[2026-06-23 11:35:50,064][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.10018050670623779, 'train/grad_norm_pre_clip': 54.49010467529297, 'timing/data': 0.0005975109525024891, 'timing/model': 1.9403989949496463, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.32}) +[2026-06-23 11:35:53,879][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.08882419764995575, 'train/grad_norm_pre_clip': 30.38554573059082, 'timing/data': 0.0005894550122320652, 'timing/model': 1.9366411559749395, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.32}) +[2026-06-23 11:35:57,683][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.046258408576250076, 'train/grad_norm_pre_clip': 86.51473999023438, 'timing/data': 0.0007602840196341276, 'timing/model': 1.9291655459674075, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.32}) +[2026-06-23 11:36:01,504][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.18604373931884766, 'train/grad_norm_pre_clip': 57.66153335571289, 'timing/data': 0.0006437059491872787, 'timing/model': 1.9442960820160806, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.32}) +[2026-06-23 11:36:05,314][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.10330025106668472, 'train/grad_norm_pre_clip': 73.9937744140625, 'timing/data': 0.0007802509935572743, 'timing/model': 1.9341163010103628, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.32}) +[2026-06-23 11:36:09,126][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.08952637761831284, 'train/grad_norm_pre_clip': 41.15122604370117, 'timing/data': 0.0006030598888173699, 'timing/model': 1.931198533042334, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.32}) +[2026-06-23 11:36:12,923][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.14142341911792755, 'train/grad_norm_pre_clip': 41.711181640625, 'timing/data': 0.0005882330005988479, 'timing/model': 1.924175492953509, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.32}) +[2026-06-23 11:36:16,830][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.17906317114830017, 'train/grad_norm_pre_clip': 64.07208251953125, 'timing/data': 0.0007901119533926249, 'timing/model': 1.9271679209778085, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.32}) +[2026-06-23 11:36:20,625][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.11187227815389633, 'train/grad_norm_pre_clip': 36.26584243774414, 'timing/data': 0.0006003379821777344, 'timing/model': 1.9282063400605693, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.32}) +[2026-06-23 11:36:24,436][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.14422783255577087, 'train/grad_norm_pre_clip': 26.014633178710938, 'timing/data': 0.0008031979668885469, 'timing/model': 1.9323716369690374, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.32}) +[2026-06-23 11:36:28,228][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.14871346950531006, 'train/grad_norm_pre_clip': 39.72030258178711, 'timing/data': 0.0005923809949308634, 'timing/model': 1.9194663499947637, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.32}) +[2026-06-23 11:36:32,042][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.08625970035791397, 'train/grad_norm_pre_clip': 38.41048812866211, 'timing/data': 0.0005927329184487462, 'timing/model': 1.9432453180197626, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.32}) +[2026-06-23 11:36:35,855][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.08432070165872574, 'train/grad_norm_pre_clip': 35.76310729980469, 'timing/data': 0.0007474439917132258, 'timing/model': 1.93639665201772, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.32}) +[2026-06-23 11:36:39,664][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.10959168523550034, 'train/grad_norm_pre_clip': 19.065155029296875, 'timing/data': 0.0006030449876561761, 'timing/model': 1.933281878940761, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.33}) +[2026-06-23 11:36:43,471][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.1192946806550026, 'train/grad_norm_pre_clip': 31.946088790893555, 'timing/data': 0.0005846699932590127, 'timing/model': 1.926308787893504, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.33}) +[2026-06-23 11:36:47,362][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.16041821241378784, 'train/grad_norm_pre_clip': 48.41337203979492, 'timing/data': 0.0005824080435559154, 'timing/model': 2.018541133031249, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.33}) +[2026-06-23 11:36:51,170][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.09115537256002426, 'train/grad_norm_pre_clip': 31.168378829956055, 'timing/data': 0.0006019920110702515, 'timing/model': 1.9331699619069695, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.33}) +[2026-06-23 11:36:54,959][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.13134543597698212, 'train/grad_norm_pre_clip': 32.05799102783203, 'timing/data': 0.0006801079725846648, 'timing/model': 1.921210136031732, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.33}) +[2026-06-23 11:36:58,753][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.17890231311321259, 'train/grad_norm_pre_clip': 36.358055114746094, 'timing/data': 0.000603188993409276, 'timing/model': 1.9271793559892103, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.33}) +[2026-06-23 11:37:02,551][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.0958809107542038, 'train/grad_norm_pre_clip': 32.536224365234375, 'timing/data': 0.0006407279288396239, 'timing/model': 1.9270317159825936, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.33}) +[2026-06-23 11:37:06,357][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.17403683066368103, 'train/grad_norm_pre_clip': 69.90647888183594, 'timing/data': 0.0006549800746142864, 'timing/model': 1.933711899095215, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.33}) +[2026-06-23 11:37:10,157][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.224458247423172, 'train/grad_norm_pre_clip': 73.16387176513672, 'timing/data': 0.0005911329062655568, 'timing/model': 1.928697815979831, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.33}) +[2026-06-23 11:37:13,972][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.07137574255466461, 'train/grad_norm_pre_clip': 21.96929359436035, 'timing/data': 0.0006838690023869276, 'timing/model': 1.9302227579755709, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.33}) +[2026-06-23 11:37:17,789][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.09367109835147858, 'train/grad_norm_pre_clip': 54.20372772216797, 'timing/data': 0.0006024789763614535, 'timing/model': 1.9336578089278191, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.33}) +[2026-06-23 11:37:21,716][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.08519068360328674, 'train/grad_norm_pre_clip': 21.561180114746094, 'timing/data': 0.0006144109647721052, 'timing/model': 2.049794294987805, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.33}) +[2026-06-23 11:37:25,519][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.12907403707504272, 'train/grad_norm_pre_clip': 36.70597839355469, 'timing/data': 0.0005788960261270404, 'timing/model': 1.9306435600155964, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.33}) +[2026-06-23 11:37:29,323][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.09216529130935669, 'train/grad_norm_pre_clip': 31.693838119506836, 'timing/data': 0.0008592959493398666, 'timing/model': 1.9297243059845641, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.33}) +[2026-06-23 11:37:33,129][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.36271634697914124, 'train/grad_norm_pre_clip': 143.8696746826172, 'timing/data': 0.0006174800219014287, 'timing/model': 1.9316201130859554, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.33}) +[2026-06-23 11:37:36,931][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.2182888239622116, 'train/grad_norm_pre_clip': 103.27957153320312, 'timing/data': 0.0006405400345101953, 'timing/model': 1.9292781020049006, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.33}) +[2026-06-23 11:37:40,725][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.23388756811618805, 'train/grad_norm_pre_clip': 113.16316986083984, 'timing/data': 0.0006936050485819578, 'timing/model': 1.9264823159901425, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.33}) +[2026-06-23 11:37:44,533][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.15372702479362488, 'train/grad_norm_pre_clip': 39.0168571472168, 'timing/data': 0.0006059140432626009, 'timing/model': 1.9354494299041107, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.33}) +[2026-06-23 11:37:48,343][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.04195668175816536, 'train/grad_norm_pre_clip': 16.63526725769043, 'timing/data': 0.0005977930268272758, 'timing/model': 1.9350255000172183, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.33}) +[2026-06-23 11:37:52,184][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.1114601194858551, 'train/grad_norm_pre_clip': 113.74137878417969, 'timing/data': 0.0006162630161270499, 'timing/model': 1.9316306279506534, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.33}) +[2026-06-23 11:37:56,152][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.3212830424308777, 'train/grad_norm_pre_clip': 126.18816375732422, 'timing/data': 0.0006285450654104352, 'timing/model': 1.9446797019336373, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.33}) +[2026-06-23 11:37:59,950][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.2203698754310608, 'train/grad_norm_pre_clip': 64.70260620117188, 'timing/data': 0.0005985499592497945, 'timing/model': 1.9328641740139574, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.33}) +[2026-06-23 11:38:03,778][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.2544214427471161, 'train/grad_norm_pre_clip': 64.53258514404297, 'timing/data': 0.0005962610011920333, 'timing/model': 1.9484929259633645, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.33}) +[2026-06-23 11:38:07,581][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.2658044695854187, 'train/grad_norm_pre_clip': 80.13861846923828, 'timing/data': 0.00059171998873353, 'timing/model': 1.9306746900547296, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.33}) +[2026-06-23 11:38:11,377][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.10124199837446213, 'train/grad_norm_pre_clip': 44.58778762817383, 'timing/data': 0.0007147300057113171, 'timing/model': 1.928474664920941, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.33}) +[2026-06-23 11:38:15,187][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.21253177523612976, 'train/grad_norm_pre_clip': 57.184913635253906, 'timing/data': 0.0006069780793040991, 'timing/model': 1.9365853200433776, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.33}) +[2026-06-23 11:38:19,004][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.07772321999073029, 'train/grad_norm_pre_clip': 34.72298812866211, 'timing/data': 0.0010559629881754518, 'timing/model': 1.9493388779228553, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.34}) +[2026-06-23 11:38:22,803][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.07483513653278351, 'train/grad_norm_pre_clip': 67.51805114746094, 'timing/data': 0.000613537966273725, 'timing/model': 1.9331632549874485, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.34}) +[2026-06-23 11:38:26,816][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.14408335089683533, 'train/grad_norm_pre_clip': 59.76335906982422, 'timing/data': 0.0006435710238292813, 'timing/model': 2.1300518340431154, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.34}) +[2026-06-23 11:38:30,625][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.059159643948078156, 'train/grad_norm_pre_clip': 25.070636749267578, 'timing/data': 0.0006038669962435961, 'timing/model': 1.9326161299832165, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.34}) +[2026-06-23 11:38:34,455][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.1267627626657486, 'train/grad_norm_pre_clip': 71.86995697021484, 'timing/data': 0.0005772130098193884, 'timing/model': 1.9286745190620422, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.34}) +[2026-06-23 11:38:38,250][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.07090602815151215, 'train/grad_norm_pre_clip': 20.81058692932129, 'timing/data': 0.0005770439747720957, 'timing/model': 1.9289693309692666, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.34}) +[2026-06-23 11:38:42,070][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.0924975797533989, 'train/grad_norm_pre_clip': 71.98765563964844, 'timing/data': 0.0009658259805291891, 'timing/model': 1.9503556140698493, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.34}) +[2026-06-23 11:38:45,880][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.24385188519954681, 'train/grad_norm_pre_clip': 41.278236389160156, 'timing/data': 0.0005993189988657832, 'timing/model': 1.9430722589604557, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.34}) +[2026-06-23 11:38:49,693][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.12809588015079498, 'train/grad_norm_pre_clip': 34.1239128112793, 'timing/data': 0.0005889550084248185, 'timing/model': 1.9351409419905394, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.34}) +[2026-06-23 11:38:53,503][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.10834810882806778, 'train/grad_norm_pre_clip': 42.89267349243164, 'timing/data': 0.0006153290160000324, 'timing/model': 1.929291627020575, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.34}) +[2026-06-23 11:38:57,353][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.08982746303081512, 'train/grad_norm_pre_clip': 17.15086555480957, 'timing/data': 0.0006054911063984036, 'timing/model': 1.9247686680173501, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.34}) +[2026-06-23 11:39:01,346][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.14074137806892395, 'train/grad_norm_pre_clip': 43.52678680419922, 'timing/data': 0.0005914469948038459, 'timing/model': 1.9300733370473608, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.34}) +[2026-06-23 11:39:05,146][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.10351690649986267, 'train/grad_norm_pre_clip': 30.33945083618164, 'timing/data': 0.0005844919942319393, 'timing/model': 1.9202988240867853, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.34}) +[2026-06-23 11:39:08,945][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.2355070561170578, 'train/grad_norm_pre_clip': 106.2485580444336, 'timing/data': 0.000736398040316999, 'timing/model': 1.9276463349815458, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.34}) +[2026-06-23 11:39:12,748][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.37708181142807007, 'train/grad_norm_pre_clip': 140.7517547607422, 'timing/data': 0.0006091010291129351, 'timing/model': 1.9326267400756478, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.34}) +[2026-06-23 11:39:16,547][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.15114402770996094, 'train/grad_norm_pre_clip': 63.91285705566406, 'timing/data': 0.0005804080283269286, 'timing/model': 1.9277906810166314, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.34}) +[2026-06-23 11:39:20,370][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.09105980396270752, 'train/grad_norm_pre_clip': 37.70364761352539, 'timing/data': 0.0005822269013151526, 'timing/model': 1.9274886320345104, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.34}) +[2026-06-23 11:39:24,170][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.12756167352199554, 'train/grad_norm_pre_clip': 50.7424430847168, 'timing/data': 0.0006003489252179861, 'timing/model': 1.927130765048787, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.34}) +[2026-06-23 11:39:27,997][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.17115068435668945, 'train/grad_norm_pre_clip': 101.12517547607422, 'timing/data': 0.0005843190010637045, 'timing/model': 1.9450417869957164, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.34}) +[2026-06-23 11:39:31,816][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.10266681015491486, 'train/grad_norm_pre_clip': 67.03961181640625, 'timing/data': 0.0005981189897283912, 'timing/model': 1.9358814829029143, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.34}) +[2026-06-23 11:39:35,799][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.13243702054023743, 'train/grad_norm_pre_clip': 63.72175216674805, 'timing/data': 0.0006168920081108809, 'timing/model': 1.929647988989018, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.34}) +[2026-06-23 11:39:39,612][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.1921088993549347, 'train/grad_norm_pre_clip': 52.8745231628418, 'timing/data': 0.0007913539884611964, 'timing/model': 1.9330474360613152, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.34}) +[2026-06-23 11:39:43,426][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.35854372382164, 'train/grad_norm_pre_clip': 126.15657806396484, 'timing/data': 0.000612115953117609, 'timing/model': 1.934040330001153, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.34}) +[2026-06-23 11:39:47,228][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.2393111288547516, 'train/grad_norm_pre_clip': 118.76115417480469, 'timing/data': 0.0006065630586817861, 'timing/model': 1.9256906050723046, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.34}) +[2026-06-23 11:39:51,085][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.29327666759490967, 'train/grad_norm_pre_clip': 89.99327087402344, 'timing/data': 0.0005816359771415591, 'timing/model': 1.9791350299492478, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.34}) +[2026-06-23 11:39:54,889][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.09222407639026642, 'train/grad_norm_pre_clip': 43.08348083496094, 'timing/data': 0.0007815000135451555, 'timing/model': 1.9374715989688411, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.34}) +[2026-06-23 11:39:58,704][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.20184899866580963, 'train/grad_norm_pre_clip': 43.032737731933594, 'timing/data': 0.000599483959376812, 'timing/model': 1.933226719032973, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.35}) +[2026-06-23 11:40:02,528][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.3597210943698883, 'train/grad_norm_pre_clip': 115.75521850585938, 'timing/data': 0.0006131109548732638, 'timing/model': 1.9382182840490714, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.35}) +[2026-06-23 11:40:06,501][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.3684430718421936, 'train/grad_norm_pre_clip': 130.82786560058594, 'timing/data': 0.0005818849895149469, 'timing/model': 2.0917703439481556, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.35}) +[2026-06-23 11:40:10,301][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.45292529463768005, 'train/grad_norm_pre_clip': 123.71842193603516, 'timing/data': 0.0007628060411661863, 'timing/model': 1.929747385904193, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.35}) +[2026-06-23 11:40:14,113][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.11891817301511765, 'train/grad_norm_pre_clip': 64.02658081054688, 'timing/data': 0.0006666160188615322, 'timing/model': 1.932926559005864, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.35}) +[2026-06-23 11:40:17,923][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.13011343777179718, 'train/grad_norm_pre_clip': 112.24935913085938, 'timing/data': 0.000608753995038569, 'timing/model': 1.9329884649487212, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.35}) +[2026-06-23 11:40:21,808][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.25329092144966125, 'train/grad_norm_pre_clip': 133.23019409179688, 'timing/data': 0.0009153640130534768, 'timing/model': 1.9386467561125755, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.35}) +[2026-06-23 11:40:25,628][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.12296773493289948, 'train/grad_norm_pre_clip': 91.46157836914062, 'timing/data': 0.0005967170000076294, 'timing/model': 1.9317686759168282, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.35}) +[2026-06-23 11:40:29,435][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 900 +[2026-06-23 11:42:40,370][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 900 +[2026-06-23 11:42:40,371][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 900 +[2026-06-23 11:44:02,362][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 900 +[2026-06-23 11:44:02,363][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.2889958620071411, 'train/grad_norm_pre_clip': 137.1876220703125, 'eval/action_loss/seconds': 130.93317537405528, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.09024044871330261, 'eval/latency_2/loss': 0.09024044871330261, 'eval/flappy/loss': 0.09024044871330261, 'eval/flappy/latency_2/loss': 0.09024044871330261, 'eval/flap_precision': 0.964451313755796, 'eval/flap_recall': 0.8421052631578947, 'eval/flap_f1': 0.899135446685879, 'eval/flap_rate_model': 0.17972222222222223, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.899135446685879, 'eval/latency_2/flap_precision': 0.964451313755796, 'eval/latency_2/flap_recall': 0.8421052631578947, 'eval/latency_2/flap_f1': 0.899135446685879, 'eval/latency_2/flap_rate_model': 0.17972222222222223, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.899135446685879, 'eval/action_classification/data_wait_seconds': 1.607393741607666, 'eval/action_classification/predict_seconds': 80.18975067138672, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006119739264249802, 'timing/model': 1.9282905890140682, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.35}) +[2026-06-23 11:44:06,123][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.09391959011554718, 'train/grad_norm_pre_clip': 71.6171875, 'timing/data': 0.0007260780548676848, 'timing/model': 1.9099363799905404, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.35}) +[2026-06-23 11:44:09,916][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.06419344991445541, 'train/grad_norm_pre_clip': 26.80864715576172, 'timing/data': 0.0008043260313570499, 'timing/model': 1.924194875988178, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.35}) +[2026-06-23 11:44:13,698][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.18979014456272125, 'train/grad_norm_pre_clip': 61.59818649291992, 'timing/data': 0.0005886680446565151, 'timing/model': 1.916216942947358, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.35}) +[2026-06-23 11:44:17,473][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.23834703862667084, 'train/grad_norm_pre_clip': 47.641815185546875, 'timing/data': 0.0006805360317230225, 'timing/model': 1.9149513760348782, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.35}) +[2026-06-23 11:44:21,248][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.07637717574834824, 'train/grad_norm_pre_clip': 45.98100662231445, 'timing/data': 0.0006827249890193343, 'timing/model': 1.9137951990123838, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.35}) +[2026-06-23 11:44:25,046][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.11842625588178635, 'train/grad_norm_pre_clip': 22.220592498779297, 'timing/data': 0.0008246330544352531, 'timing/model': 1.9266097029903904, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.35}) +[2026-06-23 11:44:28,833][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.249928280711174, 'train/grad_norm_pre_clip': 74.96721649169922, 'timing/data': 0.0005931169725954533, 'timing/model': 1.918722445028834, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.35}) +[2026-06-23 11:44:32,711][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.18368662893772125, 'train/grad_norm_pre_clip': 73.25023651123047, 'timing/data': 0.0007914300076663494, 'timing/model': 1.9260513429762796, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.35}) +[2026-06-23 11:44:36,500][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.149249866604805, 'train/grad_norm_pre_clip': 70.3889389038086, 'timing/data': 0.0007595650386065245, 'timing/model': 1.9256432669935748, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.35}) +[2026-06-23 11:44:40,306][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.10883146524429321, 'train/grad_norm_pre_clip': 42.60788345336914, 'timing/data': 0.0006088149966672063, 'timing/model': 1.9329008030472323, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.35}) +[2026-06-23 11:44:44,103][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.12471025437116623, 'train/grad_norm_pre_clip': 38.60396957397461, 'timing/data': 0.0006060309242457151, 'timing/model': 1.919119497993961, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.35}) +[2026-06-23 11:44:47,896][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.37496256828308105, 'train/grad_norm_pre_clip': 103.76922607421875, 'timing/data': 0.0007543230894953012, 'timing/model': 1.9232113959733397, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.35}) +[2026-06-23 11:44:51,688][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.13011573255062103, 'train/grad_norm_pre_clip': 58.90843963623047, 'timing/data': 0.0006180809577926993, 'timing/model': 1.9242841949453577, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.35}) +[2026-06-23 11:44:55,493][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.065827876329422, 'train/grad_norm_pre_clip': 40.603023529052734, 'timing/data': 0.0006790970219299197, 'timing/model': 1.939815907040611, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.35}) +[2026-06-23 11:44:59,297][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.09704902023077011, 'train/grad_norm_pre_clip': 31.538908004760742, 'timing/data': 0.0007934810128062963, 'timing/model': 1.9345688089961186, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.35}) +[2026-06-23 11:45:03,088][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.08569122850894928, 'train/grad_norm_pre_clip': 64.11033630371094, 'timing/data': 0.0006171730346977711, 'timing/model': 1.9231236570049077, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.35}) +[2026-06-23 11:45:07,041][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.26043692231178284, 'train/grad_norm_pre_clip': 98.68708801269531, 'timing/data': 0.0005851610330864787, 'timing/model': 1.9241835259599611, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.35}) +[2026-06-23 11:45:10,841][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.1773858666419983, 'train/grad_norm_pre_clip': 93.39057159423828, 'timing/data': 0.0008615670958533883, 'timing/model': 1.934658158919774, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.36}) +[2026-06-23 11:45:14,637][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.22379492223262787, 'train/grad_norm_pre_clip': 88.67582702636719, 'timing/data': 0.0005892269546166062, 'timing/model': 1.9287152379984036, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.36}) +[2026-06-23 11:45:18,434][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.22437167167663574, 'train/grad_norm_pre_clip': 26.946199417114258, 'timing/data': 0.0005974810337647796, 'timing/model': 1.925769969006069, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.36}) +[2026-06-23 11:45:22,316][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.15288490056991577, 'train/grad_norm_pre_clip': 67.0521240234375, 'timing/data': 0.0006114660063758492, 'timing/model': 2.007561004953459, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.36}) +[2026-06-23 11:45:26,135][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.08814973384141922, 'train/grad_norm_pre_clip': 36.930389404296875, 'timing/data': 0.0005945839220657945, 'timing/model': 1.9378220300422981, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.36}) +[2026-06-23 11:45:29,945][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.2588413655757904, 'train/grad_norm_pre_clip': 71.718994140625, 'timing/data': 0.0005960420239716768, 'timing/model': 1.9420914980582893, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.36}) +[2026-06-23 11:45:33,742][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.07899726182222366, 'train/grad_norm_pre_clip': 21.10688018798828, 'timing/data': 0.000595684046857059, 'timing/model': 1.9222849000943825, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.36}) +[2026-06-23 11:45:37,697][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.10559316724538803, 'train/grad_norm_pre_clip': 50.88272476196289, 'timing/data': 0.0006740759126842022, 'timing/model': 2.0581041959812865, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.36}) +[2026-06-23 11:45:41,497][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.14637821912765503, 'train/grad_norm_pre_clip': 45.38800811767578, 'timing/data': 0.0005885960999876261, 'timing/model': 1.9291783230146393, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.36}) +[2026-06-23 11:45:45,301][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.06067477539181709, 'train/grad_norm_pre_clip': 27.770469665527344, 'timing/data': 0.0006778599927201867, 'timing/model': 1.9290467619430274, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.36}) +[2026-06-23 11:45:49,113][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.1875830888748169, 'train/grad_norm_pre_clip': 39.11063766479492, 'timing/data': 0.0008098609978333116, 'timing/model': 1.9307811610633507, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.36}) +[2026-06-23 11:45:52,918][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.09213468432426453, 'train/grad_norm_pre_clip': 25.102134704589844, 'timing/data': 0.0008126801112666726, 'timing/model': 1.9341243399539962, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.36}) +[2026-06-23 11:45:56,714][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.12976597249507904, 'train/grad_norm_pre_clip': 48.56309127807617, 'timing/data': 0.0006492849206551909, 'timing/model': 1.9244973049499094, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.36}) +[2026-06-23 11:46:00,535][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.23520241677761078, 'train/grad_norm_pre_clip': 38.00431442260742, 'timing/data': 0.0005956890527158976, 'timing/model': 1.9463256049202755, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.36}) +[2026-06-23 11:46:04,338][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.054235704243183136, 'train/grad_norm_pre_clip': 15.822418212890625, 'timing/data': 0.0006704599363729358, 'timing/model': 1.9287533439928666, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.36}) +[2026-06-23 11:46:08,145][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.04017239436507225, 'train/grad_norm_pre_clip': 22.965274810791016, 'timing/data': 0.0005941230338066816, 'timing/model': 1.9277695789933205, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.36}) +[2026-06-23 11:46:12,139][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.17116503417491913, 'train/grad_norm_pre_clip': 34.762264251708984, 'timing/data': 0.0007851650007069111, 'timing/model': 1.9281541859963909, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.36}) +[2026-06-23 11:46:15,948][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.16553780436515808, 'train/grad_norm_pre_clip': 35.8630256652832, 'timing/data': 0.0006907070055603981, 'timing/model': 1.926301678060554, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.36}) +[2026-06-23 11:46:19,753][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.15038812160491943, 'train/grad_norm_pre_clip': 35.29103469848633, 'timing/data': 0.0006091529503464699, 'timing/model': 1.9319544860627502, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.36}) +[2026-06-23 11:46:23,557][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.03150190785527229, 'train/grad_norm_pre_clip': 24.886083602905273, 'timing/data': 0.00083276501391083, 'timing/model': 1.941629431094043, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.36}) +[2026-06-23 11:46:27,364][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.22311264276504517, 'train/grad_norm_pre_clip': 34.787689208984375, 'timing/data': 0.0005852680187672377, 'timing/model': 1.930397466989234, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.36}) +[2026-06-23 11:46:31,151][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.04706147685647011, 'train/grad_norm_pre_clip': 53.98390579223633, 'timing/data': 0.0005829799920320511, 'timing/model': 1.91940760507714, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.36}) +[2026-06-23 11:46:34,974][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.1708756387233734, 'train/grad_norm_pre_clip': 41.694034576416016, 'timing/data': 0.0005926929879933596, 'timing/model': 1.9444102520355955, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.36}) +[2026-06-23 11:46:38,785][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.19983237981796265, 'train/grad_norm_pre_clip': 71.79119110107422, 'timing/data': 0.0006063641048967838, 'timing/model': 1.932032246957533, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.36}) +[2026-06-23 11:46:42,771][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.1420905441045761, 'train/grad_norm_pre_clip': 53.8675651550293, 'timing/data': 0.0005968640325590968, 'timing/model': 2.1229081149213016, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.36}) +[2026-06-23 11:46:46,580][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.1415768414735794, 'train/grad_norm_pre_clip': 38.93865966796875, 'timing/data': 0.0005939149996265769, 'timing/model': 1.9313226509839296, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.36}) +[2026-06-23 11:46:50,375][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.05986388772726059, 'train/grad_norm_pre_clip': 47.39548873901367, 'timing/data': 0.0009044329635798931, 'timing/model': 1.9210572040174156, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.37}) +[2026-06-23 11:46:54,178][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.33049818873405457, 'train/grad_norm_pre_clip': 33.601295471191406, 'timing/data': 0.0005872090114280581, 'timing/model': 1.927730633993633, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.37}) +[2026-06-23 11:46:57,993][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.19731870293617249, 'train/grad_norm_pre_clip': 60.95852279663086, 'timing/data': 0.0005976770771667361, 'timing/model': 1.9376232930226251, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.37}) +[2026-06-23 11:47:01,799][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.05191081389784813, 'train/grad_norm_pre_clip': 42.65528106689453, 'timing/data': 0.0006481390446424484, 'timing/model': 1.9297151119681075, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.37}) +[2026-06-23 11:47:05,614][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.12108036875724792, 'train/grad_norm_pre_clip': 32.096946716308594, 'timing/data': 0.0006725509883835912, 'timing/model': 1.9340255239512771, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.37}) +[2026-06-23 11:47:09,413][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.12901704013347626, 'train/grad_norm_pre_clip': 37.45490264892578, 'timing/data': 0.0005966470343992114, 'timing/model': 1.938185786944814, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.37}) +[2026-06-23 11:47:13,257][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.057182081043720245, 'train/grad_norm_pre_clip': 25.496767044067383, 'timing/data': 0.0005995170213282108, 'timing/model': 1.9751331900479272, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.37}) +[2026-06-23 11:47:17,288][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.12158079445362091, 'train/grad_norm_pre_clip': 52.11943435668945, 'timing/data': 0.0005929499166086316, 'timing/model': 1.9260927559807897, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.37}) +[2026-06-23 11:47:21,090][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.30752983689308167, 'train/grad_norm_pre_clip': 106.66450500488281, 'timing/data': 0.0006807600148022175, 'timing/model': 1.932880387059413, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.37}) +[2026-06-23 11:47:24,899][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.12967781722545624, 'train/grad_norm_pre_clip': 30.965023040771484, 'timing/data': 0.0006172609282657504, 'timing/model': 1.928090175963007, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.37}) +[2026-06-23 11:47:28,705][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.16846874356269836, 'train/grad_norm_pre_clip': 80.23585510253906, 'timing/data': 0.0007415169384330511, 'timing/model': 1.9301475989632308, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.37}) +[2026-06-23 11:47:32,515][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.20339637994766235, 'train/grad_norm_pre_clip': 84.81553649902344, 'timing/data': 0.000598476966843009, 'timing/model': 1.9413168170722201, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.37}) +[2026-06-23 11:47:36,319][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.09829530864953995, 'train/grad_norm_pre_clip': 68.90151977539062, 'timing/data': 0.000785763026215136, 'timing/model': 1.9273113620001823, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.37}) +[2026-06-23 11:47:40,118][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.10458861291408539, 'train/grad_norm_pre_clip': 23.690208435058594, 'timing/data': 0.0005795909091830254, 'timing/model': 1.9303577260579914, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.37}) +[2026-06-23 11:47:43,922][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.15609045326709747, 'train/grad_norm_pre_clip': 32.53824996948242, 'timing/data': 0.0006708860164508224, 'timing/model': 1.9244765709154308, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.37}) +[2026-06-23 11:47:47,722][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.0998564213514328, 'train/grad_norm_pre_clip': 51.995304107666016, 'timing/data': 0.0006588040851056576, 'timing/model': 1.9329394350061193, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.37}) +[2026-06-23 11:47:51,702][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.13856659829616547, 'train/grad_norm_pre_clip': 37.213462829589844, 'timing/data': 0.0006148220272734761, 'timing/model': 1.9319940749555826, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.37}) +[2026-06-23 11:47:55,497][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.09591272473335266, 'train/grad_norm_pre_clip': 30.146883010864258, 'timing/data': 0.0007864470826461911, 'timing/model': 1.9287669019540772, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.37}) +[2026-06-23 11:47:59,291][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.16114841401576996, 'train/grad_norm_pre_clip': 42.11095428466797, 'timing/data': 0.0005867730360478163, 'timing/model': 1.927928126999177, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.37}) +[2026-06-23 11:48:03,101][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.10720055550336838, 'train/grad_norm_pre_clip': 32.34867858886719, 'timing/data': 0.0005858560325577855, 'timing/model': 1.9374941700370982, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.37}) +[2026-06-23 11:48:06,903][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.10121668130159378, 'train/grad_norm_pre_clip': 32.35093307495117, 'timing/data': 0.0005898100789636374, 'timing/model': 1.9274284250568599, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.37}) +[2026-06-23 11:48:10,760][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.14889515936374664, 'train/grad_norm_pre_clip': 38.465110778808594, 'timing/data': 0.0006038419669494033, 'timing/model': 1.9885524499695748, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.37}) +[2026-06-23 11:48:14,577][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.07271572202444077, 'train/grad_norm_pre_clip': 14.971443176269531, 'timing/data': 0.0005901209078729153, 'timing/model': 1.9367719650035724, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.37}) +[2026-06-23 11:48:18,380][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.08664971590042114, 'train/grad_norm_pre_clip': 27.923669815063477, 'timing/data': 0.0006714960327371955, 'timing/model': 1.9235072410665452, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.37}) +[2026-06-23 11:48:22,322][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.1273401528596878, 'train/grad_norm_pre_clip': 30.8418025970459, 'timing/data': 0.0008021449903026223, 'timing/model': 2.0689891510410234, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.37}) +[2026-06-23 11:48:26,130][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.28066033124923706, 'train/grad_norm_pre_clip': 66.26747131347656, 'timing/data': 0.0008693520212545991, 'timing/model': 1.9399414139334112, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.38}) +[2026-06-23 11:48:29,942][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.24478130042552948, 'train/grad_norm_pre_clip': 68.82315063476562, 'timing/data': 0.0007771370001137257, 'timing/model': 1.9326687920838594, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.38}) +[2026-06-23 11:48:33,733][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.09711704403162003, 'train/grad_norm_pre_clip': 36.10883712768555, 'timing/data': 0.0005917799426242709, 'timing/model': 1.9229165419237688, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.38}) +[2026-06-23 11:48:37,548][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.12297219038009644, 'train/grad_norm_pre_clip': 35.46590805053711, 'timing/data': 0.0006024920148774981, 'timing/model': 1.9338487510103732, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.38}) +[2026-06-23 11:48:41,347][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.14854645729064941, 'train/grad_norm_pre_clip': 59.70035171508789, 'timing/data': 0.000675868010148406, 'timing/model': 1.9257892780005932, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.38}) +[2026-06-23 11:48:45,138][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.0945405438542366, 'train/grad_norm_pre_clip': 69.10279846191406, 'timing/data': 0.0006496571004390717, 'timing/model': 1.926637897035107, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.38}) +[2026-06-23 11:48:48,940][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.19228845834732056, 'train/grad_norm_pre_clip': 75.56976318359375, 'timing/data': 0.0006734770722687244, 'timing/model': 1.9398790730629116, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.38}) +[2026-06-23 11:48:52,737][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.3017183244228363, 'train/grad_norm_pre_clip': 58.262107849121094, 'timing/data': 0.0007495300378650427, 'timing/model': 1.9316411119652912, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.38}) +[2026-06-23 11:48:56,703][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.10739519447088242, 'train/grad_norm_pre_clip': 29.440393447875977, 'timing/data': 0.0005906539736315608, 'timing/model': 1.938813264016062, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.38}) +[2026-06-23 11:49:00,512][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.14523351192474365, 'train/grad_norm_pre_clip': 73.53553771972656, 'timing/data': 0.0006628719856962562, 'timing/model': 1.9340037970105186, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.38}) +[2026-06-23 11:49:04,313][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.08972860127687454, 'train/grad_norm_pre_clip': 37.862754821777344, 'timing/data': 0.0006302609108388424, 'timing/model': 1.9291198250139132, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.38}) +[2026-06-23 11:49:08,131][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.09247563034296036, 'train/grad_norm_pre_clip': 35.95900344848633, 'timing/data': 0.000586975016631186, 'timing/model': 1.93771090102382, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.38}) +[2026-06-23 11:49:11,923][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.07465918362140656, 'train/grad_norm_pre_clip': 24.314367294311523, 'timing/data': 0.0006657680496573448, 'timing/model': 1.9254147469764575, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.38}) +[2026-06-23 11:49:15,728][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.08773776143789291, 'train/grad_norm_pre_clip': 61.363868713378906, 'timing/data': 0.0007751910015940666, 'timing/model': 1.9310697939945385, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.38}) +[2026-06-23 11:49:19,544][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.14050744473934174, 'train/grad_norm_pre_clip': 34.95142364501953, 'timing/data': 0.0007687018951401114, 'timing/model': 1.933341309078969, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.38}) +[2026-06-23 11:49:23,375][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.14822988212108612, 'train/grad_norm_pre_clip': 63.638885498046875, 'timing/data': 0.000775866094045341, 'timing/model': 1.9534127549268305, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.38}) +[2026-06-23 11:49:27,251][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.05005267634987831, 'train/grad_norm_pre_clip': 19.188905715942383, 'timing/data': 0.0005945019656792283, 'timing/model': 2.004782511969097, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.38}) +[2026-06-23 11:49:31,065][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.1202981248497963, 'train/grad_norm_pre_clip': 29.217363357543945, 'timing/data': 0.0006677329074591398, 'timing/model': 1.939245339948684, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.38}) +[2026-06-23 11:49:34,879][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.13603748381137848, 'train/grad_norm_pre_clip': 39.49640655517578, 'timing/data': 0.000589251983910799, 'timing/model': 1.9334509080508724, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.38}) +[2026-06-23 11:49:38,688][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.14358171820640564, 'train/grad_norm_pre_clip': 60.586708068847656, 'timing/data': 0.0008245240896940231, 'timing/model': 1.9296202559489757, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.38}) +[2026-06-23 11:49:42,500][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.05573594942688942, 'train/grad_norm_pre_clip': 19.48935890197754, 'timing/data': 0.0006107689114287496, 'timing/model': 1.9307825649157166, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.38}) +[2026-06-23 11:49:46,311][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.07244126498699188, 'train/grad_norm_pre_clip': 30.267131805419922, 'timing/data': 0.0005834009498357773, 'timing/model': 1.9419012509752065, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.38}) +[2026-06-23 11:49:50,133][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.0666593462228775, 'train/grad_norm_pre_clip': 18.923236846923828, 'timing/data': 0.0006219270871952176, 'timing/model': 1.944595837034285, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.38}) +[2026-06-23 11:49:53,930][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.07868736982345581, 'train/grad_norm_pre_clip': 24.745006561279297, 'timing/data': 0.0007862600032240152, 'timing/model': 1.9272311220411211, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.38}) +[2026-06-23 11:49:57,741][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.041266486048698425, 'train/grad_norm_pre_clip': 20.20760154724121, 'timing/data': 0.0006358959944918752, 'timing/model': 1.9253807780332863, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.38}) +[2026-06-23 11:50:01,731][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.16995491087436676, 'train/grad_norm_pre_clip': 40.188560485839844, 'timing/data': 0.0008851630846038461, 'timing/model': 2.11041161196772, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.38}) +[2026-06-23 11:50:05,539][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.10360674560070038, 'train/grad_norm_pre_clip': 27.737213134765625, 'timing/data': 0.0005992300575599074, 'timing/model': 1.9310748520074412, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.39}) +[2026-06-23 11:50:09,345][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.29775431752204895, 'train/grad_norm_pre_clip': 89.51883697509766, 'timing/data': 0.0005937570240348577, 'timing/model': 1.9352689770748839, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.39}) +[2026-06-23 11:50:13,154][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.25042399764060974, 'train/grad_norm_pre_clip': 79.17936706542969, 'timing/data': 0.00088410300668329, 'timing/model': 1.934410692891106, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.39}) +[2026-06-23 11:50:16,964][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.1844843477010727, 'train/grad_norm_pre_clip': 55.48535919189453, 'timing/data': 0.000602366984821856, 'timing/model': 1.9369359100237489, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.39}) +[2026-06-23 11:50:20,770][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.10636366158723831, 'train/grad_norm_pre_clip': 40.573143005371094, 'timing/data': 0.0005938439862802625, 'timing/model': 1.9297952310880646, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.39}) +[2026-06-23 11:50:24,570][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1000 +[2026-06-23 11:52:34,287][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1000 +[2026-06-23 11:52:34,289][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1000 +[2026-06-23 11:53:55,860][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1000 +[2026-06-23 11:54:37,772][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-23 11:54:37,772][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.07738209515810013, 'train/grad_norm_pre_clip': 46.040557861328125, 'eval/action_loss/seconds': 129.71464459295385, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.08614929020404816, 'eval/latency_2/loss': 0.08614929020404816, 'eval/flappy/loss': 0.08614929020404816, 'eval/flappy/latency_2/loss': 0.08614929020404816, 'eval/flap_precision': 0.9797033567525371, 'eval/flap_recall': 0.8468286099865048, 'eval/flap_f1': 0.9084328628302569, 'eval/flap_rate_model': 0.17791666666666667, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9084328628302569, 'eval/latency_2/flap_precision': 0.9797033567525371, 'eval/latency_2/flap_recall': 0.8468286099865048, 'eval/latency_2/flap_f1': 0.9084328628302569, 'eval/latency_2/flap_rate_model': 0.17791666666666667, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9084328628302569, 'eval/action_classification/data_wait_seconds': 1.467551589012146, 'eval/action_classification/predict_seconds': 79.8968734741211, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 23.40000000000016, 'rl_games_eval/mid_train/mean_length': 207.2, 'rl_games_eval/mid_train/std_reward': 23.92488244485246, 'rl_games_eval/mid_train/std_length': 192.86617121724586, 'rl_games_eval/mid_train/macro_mean_reward': 23.40000000000016, 'rl_games_eval/mid_train/macro_mean_length': 207.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 23.40000000000016, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 207.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 23.92488244485246, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 192.86617121724586, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 23.40000000000016, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006218090420588851, 'timing/model': 1.928951709996909, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.39}) +[2026-06-23 11:54:37,772][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_1000_state +[2026-06-23 11:54:37,775][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 11:54:41,999][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_1000_state/model.safetensors +[2026-06-23 11:54:45,020][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_1000_state/optimizer.bin +[2026-06-23 11:54:45,021][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_1000_state/sampler.bin +[2026-06-23 11:54:45,021][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-23 11:54:45,021][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-23 11:54:45,056][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-23 11:54:45,056][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 11:54:45,068][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 11:54:48,811][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.20553933084011078, 'train/grad_norm_pre_clip': 81.43096160888672, 'timing/data': 0.000591946067288518, 'timing/model': 1.8887695169541985, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.39}) +[2026-06-23 11:54:52,557][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.15737496316432953, 'train/grad_norm_pre_clip': 112.15770721435547, 'timing/data': 0.0007207751041278243, 'timing/model': 1.9055986459134147, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.39}) +[2026-06-23 11:54:56,294][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.06544794142246246, 'train/grad_norm_pre_clip': 41.34951400756836, 'timing/data': 0.0006969500100240111, 'timing/model': 1.8917844570241868, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.39}) +[2026-06-23 11:55:00,050][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.06147081404924393, 'train/grad_norm_pre_clip': 64.76822662353516, 'timing/data': 0.0006092939293012023, 'timing/model': 1.9063080520136282, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.39}) +[2026-06-23 11:55:03,811][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.0994962528347969, 'train/grad_norm_pre_clip': 78.38186645507812, 'timing/data': 0.000848161056637764, 'timing/model': 1.90192874497734, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.39}) +[2026-06-23 11:55:07,574][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.07315663248300552, 'train/grad_norm_pre_clip': 25.219181060791016, 'timing/data': 0.0006099760066717863, 'timing/model': 1.9093410220229998, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.39}) +[2026-06-23 11:55:11,345][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.12429406493902206, 'train/grad_norm_pre_clip': 73.34610748291016, 'timing/data': 0.0006100289756432176, 'timing/model': 1.9115443730261177, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.39}) +[2026-06-23 11:55:15,110][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.16350862383842468, 'train/grad_norm_pre_clip': 42.76243591308594, 'timing/data': 0.0007934750756248832, 'timing/model': 1.9052198530407622, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.39}) +[2026-06-23 11:55:18,895][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.06250130385160446, 'train/grad_norm_pre_clip': 36.745018005371094, 'timing/data': 0.0006148549728095531, 'timing/model': 1.9207474190043285, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.39}) +[2026-06-23 11:55:22,663][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.18284767866134644, 'train/grad_norm_pre_clip': 94.8869400024414, 'timing/data': 0.0006398530676960945, 'timing/model': 1.913857122999616, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.39}) +[2026-06-23 11:55:26,433][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.22345444560050964, 'train/grad_norm_pre_clip': 91.30973052978516, 'timing/data': 0.0006842560833320022, 'timing/model': 1.913264647941105, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.39}) +[2026-06-23 11:55:30,211][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.10303915292024612, 'train/grad_norm_pre_clip': 60.98613739013672, 'timing/data': 0.0005943350261077285, 'timing/model': 1.910988025018014, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.39}) +[2026-06-23 11:55:33,985][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.12522131204605103, 'train/grad_norm_pre_clip': 68.70346069335938, 'timing/data': 0.0008022870169952512, 'timing/model': 1.9077587620122358, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.39}) +[2026-06-23 11:55:37,776][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.13216938078403473, 'train/grad_norm_pre_clip': 62.53409957885742, 'timing/data': 0.0008404059335589409, 'timing/model': 1.9213559870840982, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.39}) +[2026-06-23 11:55:41,752][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.13716188073158264, 'train/grad_norm_pre_clip': 44.87865447998047, 'timing/data': 0.0006013680249452591, 'timing/model': 1.918801676016301, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.39}) +[2026-06-23 11:55:45,535][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.13824844360351562, 'train/grad_norm_pre_clip': 34.610687255859375, 'timing/data': 0.0006099079037085176, 'timing/model': 1.9155428929952905, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.39}) +[2026-06-23 11:55:49,338][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.12260613590478897, 'train/grad_norm_pre_clip': 24.775436401367188, 'timing/data': 0.0005984079325571656, 'timing/model': 1.9354614259209484, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.39}) +[2026-06-23 11:55:53,119][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.12678420543670654, 'train/grad_norm_pre_clip': 35.8642463684082, 'timing/data': 0.00058253004681319, 'timing/model': 1.910535431932658, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.39}) +[2026-06-23 11:55:56,916][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.10568800568580627, 'train/grad_norm_pre_clip': 55.57268142700195, 'timing/data': 0.0006158079486340284, 'timing/model': 1.9282928090542555, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.39}) +[2026-06-23 11:56:00,718][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.07057932019233704, 'train/grad_norm_pre_clip': 36.72426986694336, 'timing/data': 0.0005859630182385445, 'timing/model': 1.926351243047975, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.39}) +[2026-06-23 11:56:04,519][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.2316308617591858, 'train/grad_norm_pre_clip': 55.79446792602539, 'timing/data': 0.0005864190170541406, 'timing/model': 1.930523735936731, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.4}) +[2026-06-23 11:56:08,297][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.14609922468662262, 'train/grad_norm_pre_clip': 57.602115631103516, 'timing/data': 0.000669401022605598, 'timing/model': 1.9142057579010725, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.4}) +[2026-06-23 11:56:12,089][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.07009339332580566, 'train/grad_norm_pre_clip': 26.814668655395508, 'timing/data': 0.0007970910519361496, 'timing/model': 1.9203061850275844, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.4}) +[2026-06-23 11:56:15,889][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.09949024021625519, 'train/grad_norm_pre_clip': 42.803226470947266, 'timing/data': 0.0006501820171251893, 'timing/model': 1.926860875915736, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.4}) +[2026-06-23 11:56:19,710][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.13028784096240997, 'train/grad_norm_pre_clip': 45.107608795166016, 'timing/data': 0.0005849420558661222, 'timing/model': 1.9450425939867273, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.4}) +[2026-06-23 11:56:23,497][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.1566312462091446, 'train/grad_norm_pre_clip': 34.693992614746094, 'timing/data': 0.0006006020121276379, 'timing/model': 1.9172809870215133, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.4}) +[2026-06-23 11:56:27,292][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.13439224660396576, 'train/grad_norm_pre_clip': 27.58592987060547, 'timing/data': 0.0007245680317282677, 'timing/model': 1.926524110022001, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.4}) +[2026-06-23 11:56:31,101][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.1675952970981598, 'train/grad_norm_pre_clip': 29.75592613220215, 'timing/data': 0.000712968991138041, 'timing/model': 1.928883565007709, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.4}) +[2026-06-23 11:56:34,908][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.1612468957901001, 'train/grad_norm_pre_clip': 45.742374420166016, 'timing/data': 0.0005976910470053554, 'timing/model': 1.9342447529779747, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.4}) +[2026-06-23 11:56:38,715][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.06924305856227875, 'train/grad_norm_pre_clip': 51.1268196105957, 'timing/data': 0.0006614699959754944, 'timing/model': 1.929219783982262, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.4}) +[2026-06-23 11:56:42,506][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.15183289349079132, 'train/grad_norm_pre_clip': 64.63970947265625, 'timing/data': 0.0006910159718245268, 'timing/model': 1.9270506019238383, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.4}) +[2026-06-23 11:56:46,307][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.0715092197060585, 'train/grad_norm_pre_clip': 20.56402587890625, 'timing/data': 0.0007508239941671491, 'timing/model': 1.9300243840552866, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.4}) +[2026-06-23 11:56:50,187][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.10672006011009216, 'train/grad_norm_pre_clip': 31.558942794799805, 'timing/data': 0.0006368979811668396, 'timing/model': 1.9364653589436784, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.4}) +[2026-06-23 11:56:53,991][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.15832732617855072, 'train/grad_norm_pre_clip': 53.515899658203125, 'timing/data': 0.0008346210233867168, 'timing/model': 1.9283519530436024, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.4}) +[2026-06-23 11:56:57,804][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.06302202492952347, 'train/grad_norm_pre_clip': 33.88751220703125, 'timing/data': 0.0007564709521830082, 'timing/model': 1.9345253020292148, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.4}) +[2026-06-23 11:57:01,604][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.05729769542813301, 'train/grad_norm_pre_clip': 36.87449264526367, 'timing/data': 0.0006024549948051572, 'timing/model': 1.9180004689842463, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.4}) +[2026-06-23 11:57:05,396][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.10915495455265045, 'train/grad_norm_pre_clip': 29.089479446411133, 'timing/data': 0.0006101489998400211, 'timing/model': 1.924393949098885, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.4}) +[2026-06-23 11:57:09,193][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.1247674748301506, 'train/grad_norm_pre_clip': 25.895139694213867, 'timing/data': 0.0005966470343992114, 'timing/model': 1.9206790860043839, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.4}) +[2026-06-23 11:57:12,992][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.15095120668411255, 'train/grad_norm_pre_clip': 45.064292907714844, 'timing/data': 0.0005983209703117609, 'timing/model': 1.9272316070273519, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.4}) +[2026-06-23 11:57:16,967][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.044541697949171066, 'train/grad_norm_pre_clip': 23.85903549194336, 'timing/data': 0.0006362629355862737, 'timing/model': 2.1051594460150227, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.4}) +[2026-06-23 11:57:20,764][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.12469879537820816, 'train/grad_norm_pre_clip': 24.341997146606445, 'timing/data': 0.000616719014942646, 'timing/model': 1.9259774700039998, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.4}) +[2026-06-23 11:57:24,569][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.0621652789413929, 'train/grad_norm_pre_clip': 20.231229782104492, 'timing/data': 0.000836877035908401, 'timing/model': 1.9377991470973939, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.4}) +[2026-06-23 11:57:28,366][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.04416430741548538, 'train/grad_norm_pre_clip': 28.36949920654297, 'timing/data': 0.0006358039099723101, 'timing/model': 1.9263154390500858, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.4}) +[2026-06-23 11:57:32,248][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.1341439038515091, 'train/grad_norm_pre_clip': 39.444881439208984, 'timing/data': 0.000636375043541193, 'timing/model': 1.9280102601042017, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.4}) +[2026-06-23 11:57:36,050][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.057341329753398895, 'train/grad_norm_pre_clip': 24.091096878051758, 'timing/data': 0.0006327390437945724, 'timing/model': 1.9304681749781594, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.4}) +[2026-06-23 11:57:39,848][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.059126127511262894, 'train/grad_norm_pre_clip': 32.35667419433594, 'timing/data': 0.0007194880163297057, 'timing/model': 1.9219631749438122, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.4}) +[2026-06-23 11:57:43,636][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.05933928117156029, 'train/grad_norm_pre_clip': 31.6435546875, 'timing/data': 0.0005821669474244118, 'timing/model': 1.9166956070112064, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.41}) +[2026-06-23 11:57:47,451][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.08899807184934616, 'train/grad_norm_pre_clip': 30.014545440673828, 'timing/data': 0.0005916330264881253, 'timing/model': 1.931083858013153, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.41}) +[2026-06-23 11:57:51,266][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.1435486078262329, 'train/grad_norm_pre_clip': 64.58316802978516, 'timing/data': 0.0006179980700835586, 'timing/model': 1.934006475028582, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.41}) +[2026-06-23 11:57:55,076][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.11732155829668045, 'train/grad_norm_pre_clip': 38.61509323120117, 'timing/data': 0.000613244017586112, 'timing/model': 1.9365515100071207, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.41}) +[2026-06-23 11:57:58,871][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.2047191858291626, 'train/grad_norm_pre_clip': 50.7261962890625, 'timing/data': 0.0005981569411233068, 'timing/model': 1.9312068719882518, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.41}) +[2026-06-23 11:58:02,684][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.06422800570726395, 'train/grad_norm_pre_clip': 30.572694778442383, 'timing/data': 0.0006337059894576669, 'timing/model': 1.9315181309357285, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.41}) +[2026-06-23 11:58:06,478][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.20328576862812042, 'train/grad_norm_pre_clip': 38.86787796020508, 'timing/data': 0.0006803990108892322, 'timing/model': 1.9282382800010964, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.41}) +[2026-06-23 11:58:10,285][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.07591862976551056, 'train/grad_norm_pre_clip': 41.130760192871094, 'timing/data': 0.0006328310118988156, 'timing/model': 1.9307726579718292, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.41}) +[2026-06-23 11:58:14,105][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.16674669086933136, 'train/grad_norm_pre_clip': 124.5656967163086, 'timing/data': 0.0005920999683439732, 'timing/model': 1.933260352932848, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.41}) +[2026-06-23 11:58:17,920][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.25505906343460083, 'train/grad_norm_pre_clip': 105.56828308105469, 'timing/data': 0.0006656890036538243, 'timing/model': 1.9342636769870296, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.41}) +[2026-06-23 11:58:21,737][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.04380245879292488, 'train/grad_norm_pre_clip': 47.91841125488281, 'timing/data': 0.0005916299996897578, 'timing/model': 1.9430622638901696, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.41}) +[2026-06-23 11:58:25,550][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.1820860505104065, 'train/grad_norm_pre_clip': 36.05397415161133, 'timing/data': 0.0007879119366407394, 'timing/model': 1.9330717369448394, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.41}) +[2026-06-23 11:58:29,351][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.12645582854747772, 'train/grad_norm_pre_clip': 55.737728118896484, 'timing/data': 0.0005721979541704059, 'timing/model': 1.9227480679983273, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.41}) +[2026-06-23 11:58:33,150][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.23005729913711548, 'train/grad_norm_pre_clip': 65.16947937011719, 'timing/data': 0.0007889390690252185, 'timing/model': 1.9299146620323882, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.41}) +[2026-06-23 11:58:36,955][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.12978778779506683, 'train/grad_norm_pre_clip': 21.904586791992188, 'timing/data': 0.0006619381019845605, 'timing/model': 1.9272444290108979, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.41}) +[2026-06-23 11:58:40,832][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.1395363062620163, 'train/grad_norm_pre_clip': 37.43870544433594, 'timing/data': 0.0007151570171117783, 'timing/model': 1.9380913120694458, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.41}) +[2026-06-23 11:58:44,633][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.1732717901468277, 'train/grad_norm_pre_clip': 32.7420654296875, 'timing/data': 0.0006779000395908952, 'timing/model': 1.9298676340840757, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.41}) +[2026-06-23 11:58:48,460][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.12923061847686768, 'train/grad_norm_pre_clip': 28.165843963623047, 'timing/data': 0.0006842080038040876, 'timing/model': 1.9476353080244735, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.41}) +[2026-06-23 11:58:52,279][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.08205031603574753, 'train/grad_norm_pre_clip': 30.215526580810547, 'timing/data': 0.0005854549817740917, 'timing/model': 1.9361076320055872, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.41}) +[2026-06-23 11:58:56,144][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.11436663568019867, 'train/grad_norm_pre_clip': 45.77129364013672, 'timing/data': 0.0005933099891990423, 'timing/model': 1.9912159199593589, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.41}) +[2026-06-23 11:58:59,950][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.1000853106379509, 'train/grad_norm_pre_clip': 27.253557205200195, 'timing/data': 0.0007354400586336851, 'timing/model': 1.9313369590090588, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.41}) +[2026-06-23 11:59:03,759][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.12291505187749863, 'train/grad_norm_pre_clip': 43.67751693725586, 'timing/data': 0.000610284972935915, 'timing/model': 1.9313845389988273, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.41}) +[2026-06-23 11:59:07,560][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.1852562427520752, 'train/grad_norm_pre_clip': 45.830535888671875, 'timing/data': 0.0006028809584677219, 'timing/model': 1.9287822900805622, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.41}) +[2026-06-23 11:59:11,369][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.12061848491430283, 'train/grad_norm_pre_clip': 17.338544845581055, 'timing/data': 0.0006035120459273458, 'timing/model': 1.9308544649975374, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.41}) +[2026-06-23 11:59:15,169][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.2964559495449066, 'train/grad_norm_pre_clip': 43.29607391357422, 'timing/data': 0.0005876390496268868, 'timing/model': 1.9335237120976672, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.41}) +[2026-06-23 11:59:18,963][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.11592243611812592, 'train/grad_norm_pre_clip': 46.75904083251953, 'timing/data': 0.0005937990499660373, 'timing/model': 1.9245272499974817, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.41}) +[2026-06-23 11:59:22,765][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.1650620698928833, 'train/grad_norm_pre_clip': 22.855464935302734, 'timing/data': 0.0005902920383960009, 'timing/model': 1.9308246340369806, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.42}) +[2026-06-23 11:59:26,774][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.07057295739650726, 'train/grad_norm_pre_clip': 35.42829132080078, 'timing/data': 0.0005857329815626144, 'timing/model': 1.9422757060965523, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.42}) +[2026-06-23 11:59:30,572][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.13293085992336273, 'train/grad_norm_pre_clip': 57.5914192199707, 'timing/data': 0.0006055820267647505, 'timing/model': 1.931054693995975, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.42}) +[2026-06-23 11:59:34,387][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.0937100350856781, 'train/grad_norm_pre_clip': 33.0223388671875, 'timing/data': 0.0005956330569460988, 'timing/model': 1.9424167190445587, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.42}) +[2026-06-23 11:59:38,210][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.1575600802898407, 'train/grad_norm_pre_clip': 47.68014907836914, 'timing/data': 0.0005770919378846884, 'timing/model': 1.9405653489520773, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.42}) +[2026-06-23 11:59:42,059][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.07999127358198166, 'train/grad_norm_pre_clip': 61.09638214111328, 'timing/data': 0.0008145120227709413, 'timing/model': 1.9335763619747013, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.42}) +[2026-06-23 11:59:45,884][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.13982224464416504, 'train/grad_norm_pre_clip': 46.49976348876953, 'timing/data': 0.0005688390228897333, 'timing/model': 1.9323465509805828, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.42}) +[2026-06-23 11:59:49,676][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.10606855154037476, 'train/grad_norm_pre_clip': 25.513864517211914, 'timing/data': 0.0006253969622775912, 'timing/model': 1.9203423309372738, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.42}) +[2026-06-23 11:59:53,490][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.0925740972161293, 'train/grad_norm_pre_clip': 39.41026306152344, 'timing/data': 0.0005887519801035523, 'timing/model': 1.9346546679735184, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.42}) +[2026-06-23 11:59:57,288][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.09240050613880157, 'train/grad_norm_pre_clip': 30.571435928344727, 'timing/data': 0.0007432610727846622, 'timing/model': 1.9263776810839772, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.42}) +[2026-06-23 12:00:01,313][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.2442125678062439, 'train/grad_norm_pre_clip': 33.57648849487305, 'timing/data': 0.0006048890063539147, 'timing/model': 1.9413099529920146, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.42}) +[2026-06-23 12:00:05,118][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.0888330489397049, 'train/grad_norm_pre_clip': 25.484846115112305, 'timing/data': 0.0005930910119786859, 'timing/model': 1.9316626689396799, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.42}) +[2026-06-23 12:00:08,915][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.09220249950885773, 'train/grad_norm_pre_clip': 39.15073013305664, 'timing/data': 0.0006044129841029644, 'timing/model': 1.9218410820467398, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.42}) +[2026-06-23 12:00:12,725][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.08439958840608597, 'train/grad_norm_pre_clip': 21.221036911010742, 'timing/data': 0.0006844060262665153, 'timing/model': 1.9377592708915472, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.42}) +[2026-06-23 12:00:16,524][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.23557592928409576, 'train/grad_norm_pre_clip': 52.67765808105469, 'timing/data': 0.0005918880924582481, 'timing/model': 1.9227316780015826, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.42}) +[2026-06-23 12:00:20,342][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.06625799089670181, 'train/grad_norm_pre_clip': 42.92732238769531, 'timing/data': 0.0007215139921754599, 'timing/model': 1.9400133970193565, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.42}) +[2026-06-23 12:00:24,168][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.10809522867202759, 'train/grad_norm_pre_clip': 43.356773376464844, 'timing/data': 0.0006185040110722184, 'timing/model': 1.9381225160323083, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.42}) +[2026-06-23 12:00:27,971][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.02723819762468338, 'train/grad_norm_pre_clip': 26.874826431274414, 'timing/data': 0.0008642700267955661, 'timing/model': 1.9371656030416489, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.42}) +[2026-06-23 12:00:31,945][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.028825918212532997, 'train/grad_norm_pre_clip': 24.962284088134766, 'timing/data': 0.0006047430215403438, 'timing/model': 2.10248378792312, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.42}) +[2026-06-23 12:00:35,748][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.12048912793397903, 'train/grad_norm_pre_clip': 23.88251304626465, 'timing/data': 0.0007957420311868191, 'timing/model': 1.926300661987625, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.42}) +[2026-06-23 12:00:39,565][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.1166929379105568, 'train/grad_norm_pre_clip': 54.07645797729492, 'timing/data': 0.000735271954908967, 'timing/model': 1.9354855819838122, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.42}) +[2026-06-23 12:00:43,429][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.07676801085472107, 'train/grad_norm_pre_clip': 72.8483657836914, 'timing/data': 0.000644543906673789, 'timing/model': 1.930385056999512, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.42}) +[2026-06-23 12:00:47,249][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.12756948173046112, 'train/grad_norm_pre_clip': 60.87461471557617, 'timing/data': 0.0006100169848650694, 'timing/model': 1.9392471259925514, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.42}) +[2026-06-23 12:00:51,061][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.13363699615001678, 'train/grad_norm_pre_clip': 33.36703109741211, 'timing/data': 0.0005908029852434993, 'timing/model': 1.9376149940071627, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.42}) +[2026-06-23 12:00:54,872][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.058368414640426636, 'train/grad_norm_pre_clip': 17.621360778808594, 'timing/data': 0.000737215974368155, 'timing/model': 1.9273324949899688, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.42}) +[2026-06-23 12:00:58,680][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.23254334926605225, 'train/grad_norm_pre_clip': 48.72926330566406, 'timing/data': 0.0006726770661771297, 'timing/model': 1.932306571980007, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.42}) +[2026-06-23 12:01:02,501][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.1916065514087677, 'train/grad_norm_pre_clip': 25.862106323242188, 'timing/data': 0.0005871399771422148, 'timing/model': 1.9384000389836729, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.43}) +[2026-06-23 12:01:06,312][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1100 +[2026-06-23 12:03:13,764][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1100 +[2026-06-23 12:03:13,766][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1100 +[2026-06-23 12:04:36,172][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1100 +[2026-06-23 12:04:36,173][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.12922263145446777, 'train/grad_norm_pre_clip': 28.898500442504883, 'eval/action_loss/seconds': 127.45009931200184, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.05752938240766525, 'eval/latency_2/loss': 0.05752938240766525, 'eval/flappy/loss': 0.05752938240766525, 'eval/flappy/latency_2/loss': 0.05752938240766525, 'eval/flap_precision': 0.9531795946890287, 'eval/flap_recall': 0.9203778677462888, 'eval/flap_f1': 0.9364915894267078, 'eval/flap_rate_model': 0.19875, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9364915894267078, 'eval/latency_2/flap_precision': 0.9531795946890287, 'eval/latency_2/flap_recall': 0.9203778677462888, 'eval/latency_2/flap_f1': 0.9364915894267078, 'eval/latency_2/flap_rate_model': 0.19875, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9364915894267078, 'eval/action_classification/data_wait_seconds': 1.5417460203170776, 'eval/action_classification/predict_seconds': 80.6575698852539, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005781910149380565, 'timing/model': 1.9393508600769565, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.43}) +[2026-06-23 12:04:39,963][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.09395463764667511, 'train/grad_norm_pre_clip': 20.16336441040039, 'timing/data': 0.0008037480292841792, 'timing/model': 1.9209885679883882, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.43}) +[2026-06-23 12:04:43,750][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.09530633687973022, 'train/grad_norm_pre_clip': 17.59397315979004, 'timing/data': 0.0006079300073906779, 'timing/model': 1.9257742849877104, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.43}) +[2026-06-23 12:04:47,535][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.1832486242055893, 'train/grad_norm_pre_clip': 43.78081512451172, 'timing/data': 0.0007280149729922414, 'timing/model': 1.920952565036714, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.43}) +[2026-06-23 12:04:51,323][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.05282626301050186, 'train/grad_norm_pre_clip': 13.186800956726074, 'timing/data': 0.0008204359328374267, 'timing/model': 1.9216056689620018, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.43}) +[2026-06-23 12:04:55,219][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.11445945501327515, 'train/grad_norm_pre_clip': 41.301429748535156, 'timing/data': 0.0005925450241193175, 'timing/model': 1.9205381820211187, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.43}) +[2026-06-23 12:04:58,992][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.0732974037528038, 'train/grad_norm_pre_clip': 44.86158752441406, 'timing/data': 0.0006098619196563959, 'timing/model': 1.910209926078096, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.43}) +[2026-06-23 12:05:02,789][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.11912507563829422, 'train/grad_norm_pre_clip': 52.063392639160156, 'timing/data': 0.0006007889751344919, 'timing/model': 1.9258663250366226, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.43}) +[2026-06-23 12:05:06,590][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.11098664999008179, 'train/grad_norm_pre_clip': 24.82380485534668, 'timing/data': 0.0005800490034744143, 'timing/model': 1.9363918310264125, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.43}) +[2026-06-23 12:05:10,389][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.13205736875534058, 'train/grad_norm_pre_clip': 42.65937423706055, 'timing/data': 0.0006082119653001428, 'timing/model': 1.9254465290578082, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.43}) +[2026-06-23 12:05:14,184][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.17032088339328766, 'train/grad_norm_pre_clip': 109.69911193847656, 'timing/data': 0.0006840609712526202, 'timing/model': 1.9314779019914567, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.43}) +[2026-06-23 12:05:17,964][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.08472897112369537, 'train/grad_norm_pre_clip': 38.344810485839844, 'timing/data': 0.0008427619468420744, 'timing/model': 1.9152415180578828, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.43}) +[2026-06-23 12:05:21,745][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.21253810822963715, 'train/grad_norm_pre_clip': 29.481332778930664, 'timing/data': 0.0006247450364753604, 'timing/model': 1.9143392170080915, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.43}) +[2026-06-23 12:05:25,604][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.08967780321836472, 'train/grad_norm_pre_clip': 85.1170654296875, 'timing/data': 0.0006125069921836257, 'timing/model': 1.985730893095024, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.43}) +[2026-06-23 12:05:29,411][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.2376127392053604, 'train/grad_norm_pre_clip': 91.0462875366211, 'timing/data': 0.000588921015150845, 'timing/model': 1.9307973220711574, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.43}) +[2026-06-23 12:05:33,202][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.17746414244174957, 'train/grad_norm_pre_clip': 87.02066040039062, 'timing/data': 0.0006067909998819232, 'timing/model': 1.9216793089872226, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.43}) +[2026-06-23 12:05:37,008][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.05520777404308319, 'train/grad_norm_pre_clip': 26.221960067749023, 'timing/data': 0.0007378639420494437, 'timing/model': 1.9221736500039697, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.43}) +[2026-06-23 12:05:40,820][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.1286032497882843, 'train/grad_norm_pre_clip': 45.527774810791016, 'timing/data': 0.0005863099358975887, 'timing/model': 1.9354340349091217, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.43}) +[2026-06-23 12:05:44,625][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.07779133319854736, 'train/grad_norm_pre_clip': 61.20564651489258, 'timing/data': 0.000601849053055048, 'timing/model': 1.9306314869318157, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.43}) +[2026-06-23 12:05:48,435][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.17090310156345367, 'train/grad_norm_pre_clip': 36.80736541748047, 'timing/data': 0.0006792290369048715, 'timing/model': 1.9383968560723588, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.43}) +[2026-06-23 12:05:52,237][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.0914364829659462, 'train/grad_norm_pre_clip': 46.28834533691406, 'timing/data': 0.0006040449952706695, 'timing/model': 1.9342268420150504, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.43}) +[2026-06-23 12:05:56,052][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.14286719262599945, 'train/grad_norm_pre_clip': 33.7608642578125, 'timing/data': 0.0006013400852680206, 'timing/model': 1.937879540026188, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.43}) +[2026-06-23 12:05:59,974][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.09209291636943817, 'train/grad_norm_pre_clip': 29.57309913635254, 'timing/data': 0.0006364760920405388, 'timing/model': 2.0395885029574856, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.43}) +[2026-06-23 12:06:03,780][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.3314395844936371, 'train/grad_norm_pre_clip': 104.4291000366211, 'timing/data': 0.0005913140485063195, 'timing/model': 1.9282126909820363, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.43}) +[2026-06-23 12:06:07,568][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.10337046533823013, 'train/grad_norm_pre_clip': 45.4085578918457, 'timing/data': 0.0005924740107730031, 'timing/model': 1.9282373980386183, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.44}) +[2026-06-23 12:06:11,372][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.0395917147397995, 'train/grad_norm_pre_clip': 14.527569770812988, 'timing/data': 0.0008426710264757276, 'timing/model': 1.9325048410100862, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.44}) +[2026-06-23 12:06:15,169][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.11292488127946854, 'train/grad_norm_pre_clip': 31.037216186523438, 'timing/data': 0.0006058219587430358, 'timing/model': 1.934539775014855, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.44}) +[2026-06-23 12:06:18,967][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.11482151597738266, 'train/grad_norm_pre_clip': 30.326200485229492, 'timing/data': 0.000604834989644587, 'timing/model': 1.9219546309905127, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.44}) +[2026-06-23 12:06:22,761][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.06666416674852371, 'train/grad_norm_pre_clip': 23.912790298461914, 'timing/data': 0.0005998199339956045, 'timing/model': 1.925773145048879, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.44}) +[2026-06-23 12:06:26,581][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.08644814044237137, 'train/grad_norm_pre_clip': 54.33076858520508, 'timing/data': 0.0006005619652569294, 'timing/model': 1.9334294219734147, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.44}) +[2026-06-23 12:06:30,394][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.0283329039812088, 'train/grad_norm_pre_clip': 15.716748237609863, 'timing/data': 0.0005881530232727528, 'timing/model': 1.9407953530317172, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.44}) +[2026-06-23 12:06:34,386][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.12059629708528519, 'train/grad_norm_pre_clip': 30.224571228027344, 'timing/data': 0.0006147990934550762, 'timing/model': 2.118924244074151, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.44}) +[2026-06-23 12:06:38,196][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.14349766075611115, 'train/grad_norm_pre_clip': 50.61719512939453, 'timing/data': 0.000605540000833571, 'timing/model': 1.9310276639880612, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.44}) +[2026-06-23 12:06:41,984][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.10739269852638245, 'train/grad_norm_pre_clip': 57.10012435913086, 'timing/data': 0.0005912730703130364, 'timing/model': 1.9213656350038946, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.44}) +[2026-06-23 12:06:45,799][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.17904096841812134, 'train/grad_norm_pre_clip': 47.24024200439453, 'timing/data': 0.0005906489677727222, 'timing/model': 1.9349484060658142, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.44}) +[2026-06-23 12:06:49,588][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.07883058488368988, 'train/grad_norm_pre_clip': 55.954105377197266, 'timing/data': 0.0006817129906266928, 'timing/model': 1.9251222419552505, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.44}) +[2026-06-23 12:06:53,393][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.14918144047260284, 'train/grad_norm_pre_clip': 39.646522521972656, 'timing/data': 0.000607079011388123, 'timing/model': 1.931166238966398, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.44}) +[2026-06-23 12:06:57,198][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.04847392812371254, 'train/grad_norm_pre_clip': 43.82569122314453, 'timing/data': 0.0007246299646794796, 'timing/model': 1.9344025129685178, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.44}) +[2026-06-23 12:07:01,011][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.14125414192676544, 'train/grad_norm_pre_clip': 40.67007827758789, 'timing/data': 0.000829984899610281, 'timing/model': 1.940454324008897, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.44}) +[2026-06-23 12:07:04,820][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.14119073748588562, 'train/grad_norm_pre_clip': 70.73309326171875, 'timing/data': 0.0006084650522097945, 'timing/model': 1.9317112399730831, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.44}) +[2026-06-23 12:07:08,772][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.08706574141979218, 'train/grad_norm_pre_clip': 27.21577262878418, 'timing/data': 0.0006209489656612277, 'timing/model': 1.9258838680107147, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.44}) +[2026-06-23 12:07:12,585][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.07736775279045105, 'train/grad_norm_pre_clip': 42.196441650390625, 'timing/data': 0.0005840270314365625, 'timing/model': 1.9399532629176974, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.44}) +[2026-06-23 12:07:16,389][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.08303561061620712, 'train/grad_norm_pre_clip': 25.310794830322266, 'timing/data': 0.0005972699727863073, 'timing/model': 1.9324561540270224, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.44}) +[2026-06-23 12:07:20,213][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.18247371912002563, 'train/grad_norm_pre_clip': 40.20771789550781, 'timing/data': 0.0007776119746267796, 'timing/model': 1.9457958890125155, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.44}) +[2026-06-23 12:07:24,012][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.09718111157417297, 'train/grad_norm_pre_clip': 51.51949691772461, 'timing/data': 0.0006614300655201077, 'timing/model': 1.9284505470423028, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.44}) +[2026-06-23 12:07:27,819][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.17723752558231354, 'train/grad_norm_pre_clip': 66.7413558959961, 'timing/data': 0.0007113589672371745, 'timing/model': 1.9247898940229788, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.44}) +[2026-06-23 12:07:31,634][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.12177116423845291, 'train/grad_norm_pre_clip': 39.65215301513672, 'timing/data': 0.0006035820115357637, 'timing/model': 1.9393909790087491, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.44}) +[2026-06-23 12:07:35,433][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.18050207197666168, 'train/grad_norm_pre_clip': 43.092777252197266, 'timing/data': 0.0005921699339523911, 'timing/model': 1.9236406560521573, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.44}) +[2026-06-23 12:07:39,371][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.07908355444669724, 'train/grad_norm_pre_clip': 43.09772872924805, 'timing/data': 0.0005966369062662125, 'timing/model': 2.0652835320215672, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.44}) +[2026-06-23 12:07:43,191][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.05028499290347099, 'train/grad_norm_pre_clip': 42.11414337158203, 'timing/data': 0.0006748060695827007, 'timing/model': 1.9389330139383674, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.44}) +[2026-06-23 12:07:46,996][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.11323651671409607, 'train/grad_norm_pre_clip': 56.3337287902832, 'timing/data': 0.0005764520028606057, 'timing/model': 1.9355015790788457, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.45}) +[2026-06-23 12:07:50,806][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.15171022713184357, 'train/grad_norm_pre_clip': 20.546600341796875, 'timing/data': 0.0007025389932096004, 'timing/model': 1.9291717009618878, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.45}) +[2026-06-23 12:07:54,619][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.1635599434375763, 'train/grad_norm_pre_clip': 47.03752899169922, 'timing/data': 0.0008444000268355012, 'timing/model': 1.9350676139583811, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.45}) +[2026-06-23 12:07:58,465][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.07901211082935333, 'train/grad_norm_pre_clip': 50.153648376464844, 'timing/data': 0.0005849440349265933, 'timing/model': 1.9715321420226246, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.45}) +[2026-06-23 12:08:02,278][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.1303611546754837, 'train/grad_norm_pre_clip': 77.61093139648438, 'timing/data': 0.0006431250367313623, 'timing/model': 1.9339560139924288, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.45}) +[2026-06-23 12:08:06,115][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.2039555460214615, 'train/grad_norm_pre_clip': 45.150142669677734, 'timing/data': 0.0011034091003239155, 'timing/model': 1.935267505934462, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.45}) +[2026-06-23 12:08:09,931][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.15500125288963318, 'train/grad_norm_pre_clip': 24.716060638427734, 'timing/data': 0.0007315210532397032, 'timing/model': 1.9305727840401232, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.45}) +[2026-06-23 12:08:13,870][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.14153583347797394, 'train/grad_norm_pre_clip': 36.88270568847656, 'timing/data': 0.0006242149975150824, 'timing/model': 1.928493418963626, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.45}) +[2026-06-23 12:08:17,683][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.1396181434392929, 'train/grad_norm_pre_clip': 76.11055755615234, 'timing/data': 0.0008051260374486446, 'timing/model': 1.9227675659349188, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.45}) +[2026-06-23 12:08:21,479][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.09831398725509644, 'train/grad_norm_pre_clip': 51.993717193603516, 'timing/data': 0.0006016130791977048, 'timing/model': 1.920840765000321, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.45}) +[2026-06-23 12:08:25,290][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.06902624666690826, 'train/grad_norm_pre_clip': 20.994590759277344, 'timing/data': 0.0006026640767231584, 'timing/model': 1.9394989920547232, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.45}) +[2026-06-23 12:08:29,095][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.10676389187574387, 'train/grad_norm_pre_clip': 53.32667541503906, 'timing/data': 0.0006035359110683203, 'timing/model': 1.9343697699951008, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.45}) +[2026-06-23 12:08:32,901][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.1556176245212555, 'train/grad_norm_pre_clip': 47.86121368408203, 'timing/data': 0.0007429409306496382, 'timing/model': 1.9365413460182026, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.45}) +[2026-06-23 12:08:36,714][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.11398716270923615, 'train/grad_norm_pre_clip': 29.303184509277344, 'timing/data': 0.0008282449562102556, 'timing/model': 1.9394517629407346, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.45}) +[2026-06-23 12:08:40,518][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.16856051981449127, 'train/grad_norm_pre_clip': 23.695148468017578, 'timing/data': 0.0005900370888411999, 'timing/model': 1.9273237250745296, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.45}) +[2026-06-23 12:08:44,539][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.04976436495780945, 'train/grad_norm_pre_clip': 35.251407623291016, 'timing/data': 0.0007290510693565011, 'timing/model': 2.137051739031449, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.45}) +[2026-06-23 12:08:48,348][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.16087475419044495, 'train/grad_norm_pre_clip': 57.2512092590332, 'timing/data': 0.0005871399771422148, 'timing/model': 1.93554826406762, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.45}) +[2026-06-23 12:08:52,171][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.035806138068437576, 'train/grad_norm_pre_clip': 35.75434494018555, 'timing/data': 0.0007995509076863527, 'timing/model': 1.9401624189922586, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.45}) +[2026-06-23 12:08:55,977][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.06900753825902939, 'train/grad_norm_pre_clip': 24.035511016845703, 'timing/data': 0.0007854490540921688, 'timing/model': 1.9331755249295384, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.45}) +[2026-06-23 12:08:59,784][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.17496810853481293, 'train/grad_norm_pre_clip': 25.6665096282959, 'timing/data': 0.0008879100205376744, 'timing/model': 1.935709871002473, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.45}) +[2026-06-23 12:09:03,581][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.09236056357622147, 'train/grad_norm_pre_clip': 37.99583435058594, 'timing/data': 0.0006174680311232805, 'timing/model': 1.926437965943478, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.45}) +[2026-06-23 12:09:07,393][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.13136684894561768, 'train/grad_norm_pre_clip': 41.19609832763672, 'timing/data': 0.0008700650651007891, 'timing/model': 1.9364144450519234, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.45}) +[2026-06-23 12:09:11,207][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.07588633894920349, 'train/grad_norm_pre_clip': 46.95652389526367, 'timing/data': 0.0005954880034551024, 'timing/model': 1.9250007029622793, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.45}) +[2026-06-23 12:09:15,014][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.22104808688163757, 'train/grad_norm_pre_clip': 31.26848602294922, 'timing/data': 0.0005943669239059091, 'timing/model': 1.931713734054938, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.45}) +[2026-06-23 12:09:19,012][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.05844675004482269, 'train/grad_norm_pre_clip': 11.803693771362305, 'timing/data': 0.0006898849969729781, 'timing/model': 1.9239772559376433, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.45}) +[2026-06-23 12:09:22,822][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.2544602155685425, 'train/grad_norm_pre_clip': 142.47084045410156, 'timing/data': 0.000577887985855341, 'timing/model': 1.9354299179976806, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.45}) +[2026-06-23 12:09:26,641][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.15928450226783752, 'train/grad_norm_pre_clip': 84.70365142822266, 'timing/data': 0.0006283440161496401, 'timing/model': 1.9355833659647033, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.46}) +[2026-06-23 12:09:30,452][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.23461250960826874, 'train/grad_norm_pre_clip': 159.49803161621094, 'timing/data': 0.0007424920331686735, 'timing/model': 1.9318318730220199, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.46}) +[2026-06-23 12:09:34,248][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.12176328152418137, 'train/grad_norm_pre_clip': 39.48349380493164, 'timing/data': 0.0005949130281805992, 'timing/model': 1.9331655219430104, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.46}) +[2026-06-23 12:09:38,057][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.08483283221721649, 'train/grad_norm_pre_clip': 26.83525276184082, 'timing/data': 0.0005856839707121253, 'timing/model': 1.9313721170183271, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.46}) +[2026-06-23 12:09:41,863][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.10228586941957474, 'train/grad_norm_pre_clip': 75.2670669555664, 'timing/data': 0.0005933570209890604, 'timing/model': 1.9346122558927163, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.46}) +[2026-06-23 12:09:45,663][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.30752766132354736, 'train/grad_norm_pre_clip': 82.19448852539062, 'timing/data': 0.0006019220454618335, 'timing/model': 1.9298571839462966, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.46}) +[2026-06-23 12:09:49,479][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.18994870781898499, 'train/grad_norm_pre_clip': 63.464454650878906, 'timing/data': 0.000801081070676446, 'timing/model': 1.9315161450067535, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.46}) +[2026-06-23 12:09:53,495][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.11893165856599808, 'train/grad_norm_pre_clip': 54.75250244140625, 'timing/data': 0.0006395040545612574, 'timing/model': 1.9623055680422112, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.46}) +[2026-06-23 12:09:57,297][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.12769751250743866, 'train/grad_norm_pre_clip': 40.54564666748047, 'timing/data': 0.0008850609883666039, 'timing/model': 1.9204564639367163, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.46}) +[2026-06-23 12:10:01,117][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.055158402770757675, 'train/grad_norm_pre_clip': 36.904075622558594, 'timing/data': 0.0007108219433575869, 'timing/model': 1.9368091269861907, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.46}) +[2026-06-23 12:10:04,922][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.06771577894687653, 'train/grad_norm_pre_clip': 41.34968948364258, 'timing/data': 0.0007941980147734284, 'timing/model': 1.9315500099910423, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.46}) +[2026-06-23 12:10:08,736][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.12210352718830109, 'train/grad_norm_pre_clip': 54.329933166503906, 'timing/data': 0.0005853659240528941, 'timing/model': 1.935196685953997, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.46}) +[2026-06-23 12:10:12,535][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.10951731353998184, 'train/grad_norm_pre_clip': 70.07886505126953, 'timing/data': 0.0007982159731909633, 'timing/model': 1.931176712969318, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.46}) +[2026-06-23 12:10:16,342][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.05297978222370148, 'train/grad_norm_pre_clip': 20.101886749267578, 'timing/data': 0.0005809599533677101, 'timing/model': 1.9324406259693205, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.46}) +[2026-06-23 12:10:20,152][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.1229267343878746, 'train/grad_norm_pre_clip': 72.20696258544922, 'timing/data': 0.0006148479878902435, 'timing/model': 1.940847376943566, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.46}) +[2026-06-23 12:10:24,127][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.08007897436618805, 'train/grad_norm_pre_clip': 46.0335578918457, 'timing/data': 0.0006097640143707395, 'timing/model': 2.100102229975164, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.46}) +[2026-06-23 12:10:27,931][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.042136430740356445, 'train/grad_norm_pre_clip': 42.742244720458984, 'timing/data': 0.0006313130725175142, 'timing/model': 1.9232566839782521, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.46}) +[2026-06-23 12:10:31,741][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.14581060409545898, 'train/grad_norm_pre_clip': 45.262123107910156, 'timing/data': 0.0008471189066767693, 'timing/model': 1.9326895449776202, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.46}) +[2026-06-23 12:10:35,556][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.14303481578826904, 'train/grad_norm_pre_clip': 49.414127349853516, 'timing/data': 0.000569078023545444, 'timing/model': 1.9297864419640973, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.46}) +[2026-06-23 12:10:39,368][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.09933048486709595, 'train/grad_norm_pre_clip': 81.83826446533203, 'timing/data': 0.0007282670121639967, 'timing/model': 1.9305184200638905, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.46}) +[2026-06-23 12:10:43,180][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.11484766751527786, 'train/grad_norm_pre_clip': 24.452646255493164, 'timing/data': 0.0006387759931385517, 'timing/model': 1.935321510070935, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.46}) +[2026-06-23 12:10:47,006][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.03614509478211403, 'train/grad_norm_pre_clip': 31.987367630004883, 'timing/data': 0.0005970830097794533, 'timing/model': 1.938933951081708, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.46}) +[2026-06-23 12:10:50,819][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.12387333065271378, 'train/grad_norm_pre_clip': 28.25375747680664, 'timing/data': 0.0006045629270374775, 'timing/model': 1.9225034769624472, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.46}) +[2026-06-23 12:10:54,611][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.10308697819709778, 'train/grad_norm_pre_clip': 35.57584762573242, 'timing/data': 0.0008326449897140265, 'timing/model': 1.9221840479876846, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.46}) +[2026-06-23 12:10:58,596][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1200 +[2026-06-23 12:13:08,115][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1200 +[2026-06-23 12:13:08,117][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1200 +[2026-06-23 12:14:28,702][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1200 +[2026-06-23 12:14:28,703][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.18292877078056335, 'train/grad_norm_pre_clip': 95.47624206542969, 'eval/action_loss/seconds': 129.51673246000428, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.0644681453704834, 'eval/latency_2/loss': 0.0644681453704834, 'eval/flappy/loss': 0.0644681453704834, 'eval/flappy/latency_2/loss': 0.0644681453704834, 'eval/flap_precision': 0.8927038626609443, 'eval/flap_recall': 0.9824561403508771, 'eval/flap_f1': 0.9354320591069707, 'eval/flap_rate_model': 0.22652777777777777, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9354320591069707, 'eval/latency_2/flap_precision': 0.8927038626609443, 'eval/latency_2/flap_recall': 0.9824561403508771, 'eval/latency_2/flap_f1': 0.9354320591069707, 'eval/latency_2/flap_rate_model': 0.22652777777777777, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9354320591069707, 'eval/action_classification/data_wait_seconds': 1.6939769983291626, 'eval/action_classification/predict_seconds': 78.65098571777344, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0007238880498334765, 'timing/model': 1.937595881987363, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.46}) +[2026-06-23 12:14:32,490][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.17609255015850067, 'train/grad_norm_pre_clip': 76.74288940429688, 'timing/data': 0.0007327690254896879, 'timing/model': 1.92639561498072, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.46}) +[2026-06-23 12:14:36,285][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.09658574312925339, 'train/grad_norm_pre_clip': 94.78779602050781, 'timing/data': 0.0006671160226687789, 'timing/model': 1.919679964077659, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.47}) +[2026-06-23 12:14:40,069][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.07311174273490906, 'train/grad_norm_pre_clip': 27.36358070373535, 'timing/data': 0.0006690130103379488, 'timing/model': 1.9240604660008103, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.47}) +[2026-06-23 12:14:43,847][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.04545485973358154, 'train/grad_norm_pre_clip': 32.32181167602539, 'timing/data': 0.000786833930760622, 'timing/model': 1.9113414239836857, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.47}) +[2026-06-23 12:14:47,757][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.10768259316682816, 'train/grad_norm_pre_clip': 40.214210510253906, 'timing/data': 0.0006094389827921987, 'timing/model': 1.935926582897082, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.47}) +[2026-06-23 12:14:51,547][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.12950795888900757, 'train/grad_norm_pre_clip': 66.44888305664062, 'timing/data': 0.0010555709013715386, 'timing/model': 1.9233098479453474, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.47}) +[2026-06-23 12:14:55,346][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.17054559290409088, 'train/grad_norm_pre_clip': 31.755321502685547, 'timing/data': 0.0006058679427951574, 'timing/model': 1.9233420169912279, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.47}) +[2026-06-23 12:14:59,140][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.06802631169557571, 'train/grad_norm_pre_clip': 37.69749069213867, 'timing/data': 0.0005961990682408214, 'timing/model': 1.9220454479800537, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.47}) +[2026-06-23 12:15:02,959][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.14175915718078613, 'train/grad_norm_pre_clip': 42.611053466796875, 'timing/data': 0.0005919560790061951, 'timing/model': 1.9188899070722982, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.47}) +[2026-06-23 12:15:06,756][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.12768621742725372, 'train/grad_norm_pre_clip': 47.48133850097656, 'timing/data': 0.0006066589849069715, 'timing/model': 1.9257574600633234, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.47}) +[2026-06-23 12:15:10,586][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.12011519074440002, 'train/grad_norm_pre_clip': 37.55841827392578, 'timing/data': 0.0005910849431529641, 'timing/model': 1.9215693429578096, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.47}) +[2026-06-23 12:15:14,377][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.13468538224697113, 'train/grad_norm_pre_clip': 78.493896484375, 'timing/data': 0.0006984949577599764, 'timing/model': 1.9217656180262566, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.47}) +[2026-06-23 12:15:18,268][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.07012971490621567, 'train/grad_norm_pre_clip': 35.65353012084961, 'timing/data': 0.0006307580042630434, 'timing/model': 2.0211102470057085, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.47}) +[2026-06-23 12:15:22,064][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.22594672441482544, 'train/grad_norm_pre_clip': 22.843034744262695, 'timing/data': 0.0007111619925126433, 'timing/model': 1.9145703659160063, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.47}) +[2026-06-23 12:15:25,852][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.09972996264696121, 'train/grad_norm_pre_clip': 27.530399322509766, 'timing/data': 0.0006062320899218321, 'timing/model': 1.9187093229265884, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.47}) +[2026-06-23 12:15:29,655][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.07709718495607376, 'train/grad_norm_pre_clip': 39.05496597290039, 'timing/data': 0.0006049940129742026, 'timing/model': 1.9269944700645283, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.47}) +[2026-06-23 12:15:33,439][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.12641513347625732, 'train/grad_norm_pre_clip': 25.976783752441406, 'timing/data': 0.0005954590160399675, 'timing/model': 1.9255300590302795, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.47}) +[2026-06-23 12:15:37,250][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.07001559436321259, 'train/grad_norm_pre_clip': 47.61838150024414, 'timing/data': 0.0005895650247111917, 'timing/model': 1.9297603700542822, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.47}) +[2026-06-23 12:15:41,049][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.02007298171520233, 'train/grad_norm_pre_clip': 15.601720809936523, 'timing/data': 0.0006046739872545004, 'timing/model': 1.9233323090011254, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.47}) +[2026-06-23 12:15:44,835][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.09399288147687912, 'train/grad_norm_pre_clip': 31.94925880432129, 'timing/data': 0.0006075630662962794, 'timing/model': 1.9248612329829484, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.47}) +[2026-06-23 12:15:48,702][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.1094975695014, 'train/grad_norm_pre_clip': 29.028226852416992, 'timing/data': 0.0007804760243743658, 'timing/model': 1.9388070169370621, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.47}) +[2026-06-23 12:15:52,581][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.16859079897403717, 'train/grad_norm_pre_clip': 57.71479034423828, 'timing/data': 0.0006018709391355515, 'timing/model': 1.9293393060797825, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.47}) +[2026-06-23 12:15:56,383][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.054821960628032684, 'train/grad_norm_pre_clip': 42.81358337402344, 'timing/data': 0.0007403589552268386, 'timing/model': 1.9290222669951618, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.47}) +[2026-06-23 12:16:00,203][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.10075581818819046, 'train/grad_norm_pre_clip': 53.83911895751953, 'timing/data': 0.0005996309919282794, 'timing/model': 1.9332700700033456, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.47}) +[2026-06-23 12:16:03,999][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.19716836512088776, 'train/grad_norm_pre_clip': 875.0846557617188, 'timing/data': 0.0006494090193882585, 'timing/model': 1.9174405980156735, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.47}) +[2026-06-23 12:16:07,791][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.01536730770021677, 'train/grad_norm_pre_clip': 26.74843978881836, 'timing/data': 0.0006901989690959454, 'timing/model': 1.9171315249986947, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.47}) +[2026-06-23 12:16:11,582][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.10444463789463043, 'train/grad_norm_pre_clip': 31.26621437072754, 'timing/data': 0.0006826990284025669, 'timing/model': 1.926352001959458, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.47}) +[2026-06-23 12:16:15,374][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.16413019597530365, 'train/grad_norm_pre_clip': 59.919578552246094, 'timing/data': 0.0005949609912931919, 'timing/model': 1.9305005570640787, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.48}) +[2026-06-23 12:16:19,179][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.06511886417865753, 'train/grad_norm_pre_clip': 40.41907501220703, 'timing/data': 0.0005846900166943669, 'timing/model': 1.9305146530969068, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.48}) +[2026-06-23 12:16:23,003][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.1416805237531662, 'train/grad_norm_pre_clip': 58.686702728271484, 'timing/data': 0.0006692539900541306, 'timing/model': 1.9474411639384925, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.48}) +[2026-06-23 12:16:26,814][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.12101338058710098, 'train/grad_norm_pre_clip': 34.995338439941406, 'timing/data': 0.0006641370709985495, 'timing/model': 1.9407074130140245, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.48}) +[2026-06-23 12:16:30,604][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.08080380409955978, 'train/grad_norm_pre_clip': 36.44534683227539, 'timing/data': 0.0005906270816922188, 'timing/model': 1.9241892859572545, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.48}) +[2026-06-23 12:16:34,388][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.08059363812208176, 'train/grad_norm_pre_clip': 96.35944366455078, 'timing/data': 0.0006437470437958837, 'timing/model': 1.9216220009839162, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.48}) +[2026-06-23 12:16:38,181][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.18862758576869965, 'train/grad_norm_pre_clip': 119.1486587524414, 'timing/data': 0.000658793025650084, 'timing/model': 1.9240379980765283, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.48}) +[2026-06-23 12:16:41,982][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.22581025958061218, 'train/grad_norm_pre_clip': 77.10517120361328, 'timing/data': 0.000819580047391355, 'timing/model': 1.9249146060319617, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.48}) +[2026-06-23 12:16:45,780][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.08091255277395248, 'train/grad_norm_pre_clip': 86.5367660522461, 'timing/data': 0.0005971150239929557, 'timing/model': 1.9305716709932312, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.48}) +[2026-06-23 12:16:49,616][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.08947058022022247, 'train/grad_norm_pre_clip': 58.498443603515625, 'timing/data': 0.0007342980243265629, 'timing/model': 1.926567148999311, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.48}) +[2026-06-23 12:16:53,425][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.10692224651575089, 'train/grad_norm_pre_clip': 53.52785110473633, 'timing/data': 0.0006248039426282048, 'timing/model': 1.9382064070086926, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.48}) +[2026-06-23 12:16:57,246][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.07839439809322357, 'train/grad_norm_pre_clip': 16.662460327148438, 'timing/data': 0.0006220909999683499, 'timing/model': 1.9329412849619985, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.48}) +[2026-06-23 12:17:01,052][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.10177664458751678, 'train/grad_norm_pre_clip': 32.659358978271484, 'timing/data': 0.0006080250022932887, 'timing/model': 1.925553097971715, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.48}) +[2026-06-23 12:17:04,850][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.0850757583975792, 'train/grad_norm_pre_clip': 41.292537689208984, 'timing/data': 0.0007841639453545213, 'timing/model': 1.9245598929701373, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.48}) +[2026-06-23 12:17:08,669][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.09608153998851776, 'train/grad_norm_pre_clip': 21.594776153564453, 'timing/data': 0.0007404509233310819, 'timing/model': 1.9337700720643625, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.48}) +[2026-06-23 12:17:12,466][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.1394723653793335, 'train/grad_norm_pre_clip': 77.52176666259766, 'timing/data': 0.0006006920011714101, 'timing/model': 1.9314870740054175, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.48}) +[2026-06-23 12:17:16,266][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.07720944285392761, 'train/grad_norm_pre_clip': 56.68780517578125, 'timing/data': 0.000619187019765377, 'timing/model': 1.9221454979851842, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.48}) +[2026-06-23 12:17:20,073][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.03683851659297943, 'train/grad_norm_pre_clip': 43.698001861572266, 'timing/data': 0.000616036937572062, 'timing/model': 1.9294609620701522, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.48}) +[2026-06-23 12:17:23,883][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.12899954617023468, 'train/grad_norm_pre_clip': 53.130455017089844, 'timing/data': 0.0006463450845330954, 'timing/model': 1.9304286470869556, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.48}) +[2026-06-23 12:17:27,697][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.039725419133901596, 'train/grad_norm_pre_clip': 11.50936508178711, 'timing/data': 0.0007702640723437071, 'timing/model': 1.9370769679080695, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.48}) +[2026-06-23 12:17:31,684][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.12982554733753204, 'train/grad_norm_pre_clip': 38.37739562988281, 'timing/data': 0.0006911660311743617, 'timing/model': 1.9310075170360506, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.48}) +[2026-06-23 12:17:35,487][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.1546659618616104, 'train/grad_norm_pre_clip': 47.16867446899414, 'timing/data': 0.0005922109121456742, 'timing/model': 1.9313556640408933, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.48}) +[2026-06-23 12:20:39,400][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1250 +[2026-06-23 12:20:39,401][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.1803414523601532, 'train/grad_norm_pre_clip': 42.64938735961914, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 108.73999999999823, 'rl_games_eval/mid_train/mean_length': 896.8, 'rl_games_eval/mid_train/std_reward': 97.07412837620372, 'rl_games_eval/mid_train/std_length': 783.4583332890141, 'rl_games_eval/mid_train/macro_mean_reward': 108.73999999999823, 'rl_games_eval/mid_train/macro_mean_length': 896.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 108.73999999999823, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 896.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 97.07412837620372, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 783.4583332890141, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 108.73999999999823, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006134889554232359, 'timing/model': 1.9416721710003912, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.48}) +[2026-06-23 12:20:43,137][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.09828726947307587, 'train/grad_norm_pre_clip': 27.83407211303711, 'timing/data': 0.0008279869798570871, 'timing/model': 1.8998041959712282, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.48}) +[2026-06-23 12:20:46,878][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.06779477745294571, 'train/grad_norm_pre_clip': 50.170379638671875, 'timing/data': 0.0006753290072083473, 'timing/model': 1.9034391320310533, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.48}) +[2026-06-23 12:20:50,631][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.07676380127668381, 'train/grad_norm_pre_clip': 30.32054328918457, 'timing/data': 0.0006131120026111603, 'timing/model': 1.9025543470634148, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.48}) +[2026-06-23 12:20:54,429][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.0641169622540474, 'train/grad_norm_pre_clip': 52.33149719238281, 'timing/data': 0.0005943089490756392, 'timing/model': 1.9121182140661404, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.49}) +[2026-06-23 12:20:58,186][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.21188649535179138, 'train/grad_norm_pre_clip': 49.7854118347168, 'timing/data': 0.0006784469587728381, 'timing/model': 1.905126768979244, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.49}) +[2026-06-23 12:21:01,993][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.14438749849796295, 'train/grad_norm_pre_clip': 34.570613861083984, 'timing/data': 0.0006554409628733993, 'timing/model': 1.9050839929841459, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.49}) +[2026-06-23 12:21:05,766][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.06674985587596893, 'train/grad_norm_pre_clip': 29.588029861450195, 'timing/data': 0.0006438699783757329, 'timing/model': 1.9203932490199804, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.49}) +[2026-06-23 12:21:09,529][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.16345523297786713, 'train/grad_norm_pre_clip': 52.88746643066406, 'timing/data': 0.0008147910702973604, 'timing/model': 1.9084803028963506, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.49}) +[2026-06-23 12:21:13,291][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.06719988584518433, 'train/grad_norm_pre_clip': 52.843875885009766, 'timing/data': 0.0005939709953963757, 'timing/model': 1.9091238839318976, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.49}) +[2026-06-23 12:21:17,044][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.06871747970581055, 'train/grad_norm_pre_clip': 51.32647705078125, 'timing/data': 0.0005911400076001883, 'timing/model': 1.9052238869480789, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.49}) +[2026-06-23 12:21:20,838][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.09399258345365524, 'train/grad_norm_pre_clip': 19.717628479003906, 'timing/data': 0.0005870319437235594, 'timing/model': 1.9098987380275503, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.49}) +[2026-06-23 12:21:24,601][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.215835303068161, 'train/grad_norm_pre_clip': 63.03939437866211, 'timing/data': 0.0005890850443392992, 'timing/model': 1.915760895004496, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.49}) +[2026-06-23 12:21:28,366][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.22028577327728271, 'train/grad_norm_pre_clip': 41.13768005371094, 'timing/data': 0.0005958708934485912, 'timing/model': 1.9177702500019222, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.49}) +[2026-06-23 12:21:32,133][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.10691199451684952, 'train/grad_norm_pre_clip': 29.05661964416504, 'timing/data': 0.0005943370051681995, 'timing/model': 1.9139503740007058, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.49}) +[2026-06-23 12:21:35,912][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.0774443969130516, 'train/grad_norm_pre_clip': 18.54486083984375, 'timing/data': 0.0005880399839952588, 'timing/model': 1.9239732719724998, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.49}) +[2026-06-23 12:21:39,688][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.13235299289226532, 'train/grad_norm_pre_clip': 50.99436950683594, 'timing/data': 0.0005834910552948713, 'timing/model': 1.9145892179803923, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.49}) +[2026-06-23 12:21:43,479][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.05450456216931343, 'train/grad_norm_pre_clip': 16.069923400878906, 'timing/data': 0.0006044590845704079, 'timing/model': 1.9270890590269119, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.49}) +[2026-06-23 12:21:47,269][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.09056629240512848, 'train/grad_norm_pre_clip': 23.66798210144043, 'timing/data': 0.0006022650049999356, 'timing/model': 1.9241058529587463, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.49}) +[2026-06-23 12:21:51,240][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.12740099430084229, 'train/grad_norm_pre_clip': 31.843080520629883, 'timing/data': 0.0005948509788140655, 'timing/model': 2.093097197008319, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.49}) +[2026-06-23 12:21:55,022][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.0837077870965004, 'train/grad_norm_pre_clip': 22.00705337524414, 'timing/data': 0.000609175069257617, 'timing/model': 1.920829465962015, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.49}) +[2026-06-23 12:21:58,818][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.24210494756698608, 'train/grad_norm_pre_clip': 95.33149719238281, 'timing/data': 0.0005950760096311569, 'timing/model': 1.9211773790884763, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.49}) +[2026-06-23 12:22:02,639][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.08004054427146912, 'train/grad_norm_pre_clip': 26.105703353881836, 'timing/data': 0.00080802699085325, 'timing/model': 1.9524725610390306, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.49}) +[2026-06-23 12:22:06,434][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.03172319009900093, 'train/grad_norm_pre_clip': 21.271394729614258, 'timing/data': 0.0006455900147557259, 'timing/model': 1.931178221013397, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.49}) +[2026-06-23 12:22:10,215][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.09992659837007523, 'train/grad_norm_pre_clip': 29.110376358032227, 'timing/data': 0.0006925689522176981, 'timing/model': 1.9167422610335052, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.49}) +[2026-06-23 12:22:14,007][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.14837925136089325, 'train/grad_norm_pre_clip': 199.01815795898438, 'timing/data': 0.0006481849122792482, 'timing/model': 1.921346358023584, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.49}) +[2026-06-23 12:22:17,803][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.07362661510705948, 'train/grad_norm_pre_clip': 28.639436721801758, 'timing/data': 0.0005893360357731581, 'timing/model': 1.9274781540734693, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.49}) +[2026-06-23 12:22:21,582][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.07152685523033142, 'train/grad_norm_pre_clip': 38.039581298828125, 'timing/data': 0.0006549329264089465, 'timing/model': 1.920732649974525, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.49}) +[2026-06-23 12:22:25,376][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.23293375968933105, 'train/grad_norm_pre_clip': 70.80166625976562, 'timing/data': 0.0005901249824091792, 'timing/model': 1.9300997250247747, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.49}) +[2026-06-23 12:22:29,172][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.10296375304460526, 'train/grad_norm_pre_clip': 33.38748550415039, 'timing/data': 0.0006904529873281717, 'timing/model': 1.938274796004407, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.49}) +[2026-06-23 12:22:32,976][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.06080564856529236, 'train/grad_norm_pre_clip': 63.153079986572266, 'timing/data': 0.0006057019345462322, 'timing/model': 1.9341578669846058, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.5}) +[2026-06-23 12:22:36,782][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.14664997160434723, 'train/grad_norm_pre_clip': 31.349315643310547, 'timing/data': 0.000594810931943357, 'timing/model': 1.926591576077044, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.5}) +[2026-06-23 12:22:40,572][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.1780230849981308, 'train/grad_norm_pre_clip': 58.515079498291016, 'timing/data': 0.0006033940007910132, 'timing/model': 1.9219562530051917, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.5}) +[2026-06-23 12:22:44,361][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.16365456581115723, 'train/grad_norm_pre_clip': 30.554515838623047, 'timing/data': 0.0006093130214139819, 'timing/model': 1.9174336629221216, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.5}) +[2026-06-23 12:22:48,161][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.1046372652053833, 'train/grad_norm_pre_clip': 21.366653442382812, 'timing/data': 0.0006446510087698698, 'timing/model': 1.9299684290308505, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.5}) +[2026-06-23 12:22:51,969][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.23639997839927673, 'train/grad_norm_pre_clip': 42.97164535522461, 'timing/data': 0.0007947100093588233, 'timing/model': 1.9282096980605274, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.5}) +[2026-06-23 12:22:55,768][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.15370601415634155, 'train/grad_norm_pre_clip': 62.19159698486328, 'timing/data': 0.0005884750280529261, 'timing/model': 1.927122401073575, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.5}) +[2026-06-23 12:22:59,573][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.1584586352109909, 'train/grad_norm_pre_clip': 102.01097106933594, 'timing/data': 0.0008372460724785924, 'timing/model': 1.931943706003949, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.5}) +[2026-06-23 12:23:03,364][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.04861459136009216, 'train/grad_norm_pre_clip': 19.52825927734375, 'timing/data': 0.0006617429899051785, 'timing/model': 1.9263331650290638, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.5}) +[2026-06-23 12:23:07,161][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.12840603291988373, 'train/grad_norm_pre_clip': 17.59187126159668, 'timing/data': 0.0006391189526766539, 'timing/model': 1.9313778060022742, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.5}) +[2026-06-23 12:23:10,961][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.1812569499015808, 'train/grad_norm_pre_clip': 57.525909423828125, 'timing/data': 0.0007434849394485354, 'timing/model': 1.9244009839603677, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.5}) +[2026-06-23 12:23:14,763][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.23889124393463135, 'train/grad_norm_pre_clip': 60.94859313964844, 'timing/data': 0.000618202961049974, 'timing/model': 1.9306872830493376, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.5}) +[2026-06-23 12:23:18,562][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.28677621483802795, 'train/grad_norm_pre_clip': 101.0472412109375, 'timing/data': 0.0005814379546791315, 'timing/model': 1.9294468110892922, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.5}) +[2026-06-23 12:23:22,358][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.15541967749595642, 'train/grad_norm_pre_clip': 65.24246978759766, 'timing/data': 0.0005966599564999342, 'timing/model': 1.931873254943639, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.5}) +[2026-06-23 12:23:26,151][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.1088910698890686, 'train/grad_norm_pre_clip': 17.05556869506836, 'timing/data': 0.0006799299735575914, 'timing/model': 1.9200149800162762, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.5}) +[2026-06-23 12:23:29,953][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.14737828075885773, 'train/grad_norm_pre_clip': 51.44356918334961, 'timing/data': 0.0005930199986323714, 'timing/model': 1.9327942640520632, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.5}) +[2026-06-23 12:23:33,748][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.23596972227096558, 'train/grad_norm_pre_clip': 53.01819610595703, 'timing/data': 0.0006015779217705131, 'timing/model': 1.9258041410939768, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.5}) +[2026-06-23 12:23:37,557][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.1456160992383957, 'train/grad_norm_pre_clip': 65.2093734741211, 'timing/data': 0.0006238430505618453, 'timing/model': 1.9250584989786148, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.5}) +[2026-06-23 12:23:41,391][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.14894546568393707, 'train/grad_norm_pre_clip': 56.36119079589844, 'timing/data': 0.0006109719397500157, 'timing/model': 1.951548260054551, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.5}) +[2026-06-23 12:23:45,198][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.17769365012645721, 'train/grad_norm_pre_clip': 23.733083724975586, 'timing/data': 0.0006205240497365594, 'timing/model': 1.9279277380555868, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.5}) +[2026-06-23 12:23:49,008][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1300 +[2026-06-23 12:25:56,574][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1300 +[2026-06-23 12:25:56,576][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1300 +[2026-06-23 12:27:20,773][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1300 +[2026-06-23 12:27:20,774][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.07678132504224777, 'train/grad_norm_pre_clip': 30.68048858642578, 'eval/action_loss/seconds': 127.56391045101918, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06606002151966095, 'eval/latency_2/loss': 0.06606002151966095, 'eval/flappy/loss': 0.06606002151966095, 'eval/flappy/latency_2/loss': 0.06606002151966095, 'eval/flap_precision': 0.9722222222222222, 'eval/flap_recall': 0.8974358974358975, 'eval/flap_f1': 0.9333333333333333, 'eval/flap_rate_model': 0.19, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9333333333333333, 'eval/latency_2/flap_precision': 0.9722222222222222, 'eval/latency_2/flap_recall': 0.8974358974358975, 'eval/latency_2/flap_f1': 0.9333333333333333, 'eval/latency_2/flap_rate_model': 0.19, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9333333333333333, 'eval/action_classification/data_wait_seconds': 1.4966423511505127, 'eval/action_classification/predict_seconds': 82.4989013671875, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0008328460389748216, 'timing/model': 1.9403721820563078, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.5}) +[2026-06-23 12:27:24,540][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.1637781858444214, 'train/grad_norm_pre_clip': 59.649681091308594, 'timing/data': 0.0008185330079868436, 'timing/model': 1.908259427989833, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.5}) +[2026-06-23 12:27:28,338][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.08156441152095795, 'train/grad_norm_pre_clip': 42.19090270996094, 'timing/data': 0.0005801660008728504, 'timing/model': 1.9361976420041174, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.5}) +[2026-06-23 12:27:32,120][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.12186773866415024, 'train/grad_norm_pre_clip': 44.46832275390625, 'timing/data': 0.0006024440517649055, 'timing/model': 1.9179804000305012, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.5}) +[2026-06-23 12:27:35,904][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.08206072449684143, 'train/grad_norm_pre_clip': 14.513768196105957, 'timing/data': 0.000773233943618834, 'timing/model': 1.9122679019346833, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.5}) +[2026-06-23 12:27:39,687][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.18156634271144867, 'train/grad_norm_pre_clip': 39.51838302612305, 'timing/data': 0.0006433459930121899, 'timing/model': 1.9198234429350123, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.51}) +[2026-06-23 12:27:43,481][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.2166159600019455, 'train/grad_norm_pre_clip': 49.034217834472656, 'timing/data': 0.0006260829977691174, 'timing/model': 1.9226113751064986, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.51}) +[2026-06-23 12:27:47,267][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.2229769229888916, 'train/grad_norm_pre_clip': 60.96383285522461, 'timing/data': 0.000594640034250915, 'timing/model': 1.9215815160423517, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.51}) +[2026-06-23 12:27:51,204][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.12184619903564453, 'train/grad_norm_pre_clip': 24.57037925720215, 'timing/data': 0.0006207589758560061, 'timing/model': 1.9216945389052853, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.51}) +[2026-06-23 12:27:54,986][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.07627270370721817, 'train/grad_norm_pre_clip': 37.81057357788086, 'timing/data': 0.000629195012152195, 'timing/model': 1.9261838470119983, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.51}) +[2026-06-23 12:27:58,777][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.04548671841621399, 'train/grad_norm_pre_clip': 16.3280086517334, 'timing/data': 0.0007955729961395264, 'timing/model': 1.9164528930559754, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.51}) +[2026-06-23 12:28:02,582][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.07064026594161987, 'train/grad_norm_pre_clip': 41.42863845825195, 'timing/data': 0.0006255940534174442, 'timing/model': 1.9347408310277387, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.51}) +[2026-06-23 12:28:06,380][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.09894541651010513, 'train/grad_norm_pre_clip': 49.88154602050781, 'timing/data': 0.0010930539574474096, 'timing/model': 1.9267684139776975, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.51}) +[2026-06-23 12:28:10,176][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.1267184466123581, 'train/grad_norm_pre_clip': 28.00035285949707, 'timing/data': 0.0006535440916195512, 'timing/model': 1.9268981930799782, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.51}) +[2026-06-23 12:28:13,989][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.16304093599319458, 'train/grad_norm_pre_clip': 45.68254470825195, 'timing/data': 0.0006603989750146866, 'timing/model': 1.937174506019801, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.51}) +[2026-06-23 12:28:17,793][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.20610283315181732, 'train/grad_norm_pre_clip': 59.59479522705078, 'timing/data': 0.0008238781010732055, 'timing/model': 1.9294409559806809, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.51}) +[2026-06-23 12:28:21,782][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.06387932598590851, 'train/grad_norm_pre_clip': 26.80577278137207, 'timing/data': 0.0006207709666341543, 'timing/model': 2.116603681934066, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.51}) +[2026-06-23 12:28:25,573][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.14084607362747192, 'train/grad_norm_pre_clip': 27.378520965576172, 'timing/data': 0.0006013419479131699, 'timing/model': 1.9271710959728807, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.51}) +[2026-06-23 12:28:29,374][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.09052852541208267, 'train/grad_norm_pre_clip': 35.95777893066406, 'timing/data': 0.0005835640477016568, 'timing/model': 1.9357790839858353, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.51}) +[2026-06-23 12:28:33,167][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.05843799188733101, 'train/grad_norm_pre_clip': 23.840803146362305, 'timing/data': 0.000615666969679296, 'timing/model': 1.917131299036555, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.51}) +[2026-06-23 12:28:36,956][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.11554767191410065, 'train/grad_norm_pre_clip': 39.449134826660156, 'timing/data': 0.0005829409928992391, 'timing/model': 1.9247784529579803, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.51}) +[2026-06-23 12:28:40,762][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.034046854823827744, 'train/grad_norm_pre_clip': 39.59107208251953, 'timing/data': 0.0007108589634299278, 'timing/model': 1.9370508670108393, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.51}) +[2026-06-23 12:28:44,572][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.04413660615682602, 'train/grad_norm_pre_clip': 21.85491943359375, 'timing/data': 0.0005833199247717857, 'timing/model': 1.9354341020807624, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.51}) +[2026-06-23 12:28:48,377][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.18533451855182648, 'train/grad_norm_pre_clip': 29.789260864257812, 'timing/data': 0.000623962958343327, 'timing/model': 1.929733217926696, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.51}) +[2026-06-23 12:28:52,174][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.27189528942108154, 'train/grad_norm_pre_clip': 52.331233978271484, 'timing/data': 0.0005887519801035523, 'timing/model': 1.9274474189151078, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.51}) +[2026-06-23 12:28:56,169][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.10836757719516754, 'train/grad_norm_pre_clip': 47.64736557006836, 'timing/data': 0.0006023410242050886, 'timing/model': 1.924726944998838, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.51}) +[2026-06-23 12:28:59,973][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.10916684567928314, 'train/grad_norm_pre_clip': 42.85435104370117, 'timing/data': 0.0005950520280748606, 'timing/model': 1.9327655180823058, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.51}) +[2026-06-23 12:29:03,782][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.08104459941387177, 'train/grad_norm_pre_clip': 25.546035766601562, 'timing/data': 0.0006458970019593835, 'timing/model': 1.9345588979776949, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.51}) +[2026-06-23 12:29:07,588][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.026180027052760124, 'train/grad_norm_pre_clip': 17.259885787963867, 'timing/data': 0.0006233870517462492, 'timing/model': 1.9236007500439882, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.51}) +[2026-06-23 12:29:11,394][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.08326615393161774, 'train/grad_norm_pre_clip': 24.04163932800293, 'timing/data': 0.0008765119127929211, 'timing/model': 1.9387353140627965, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.51}) +[2026-06-23 12:29:15,199][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.06961147487163544, 'train/grad_norm_pre_clip': 47.87482452392578, 'timing/data': 0.0006220899522304535, 'timing/model': 1.931644574040547, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.51}) +[2026-06-23 12:29:19,008][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.18157723546028137, 'train/grad_norm_pre_clip': 34.594425201416016, 'timing/data': 0.0006068629445508122, 'timing/model': 1.9364031789591536, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.52}) +[2026-06-23 12:29:22,805][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.13044674694538116, 'train/grad_norm_pre_clip': 33.989139556884766, 'timing/data': 0.0007744169561192393, 'timing/model': 1.921999281970784, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.52}) +[2026-06-23 12:29:26,608][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.08822629600763321, 'train/grad_norm_pre_clip': 48.67951965332031, 'timing/data': 0.0008363060187548399, 'timing/model': 1.9363166299881414, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.52}) +[2026-06-23 12:29:30,588][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.24253425002098083, 'train/grad_norm_pre_clip': 89.89643859863281, 'timing/data': 0.0005842039827257395, 'timing/model': 1.9365946629550308, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.52}) +[2026-06-23 12:29:34,407][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.2155495136976242, 'train/grad_norm_pre_clip': 62.20676803588867, 'timing/data': 0.0005805880064144731, 'timing/model': 1.9459309020312503, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.52}) +[2026-06-23 12:29:38,237][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.1234462633728981, 'train/grad_norm_pre_clip': 46.05950927734375, 'timing/data': 0.0005836320342496037, 'timing/model': 1.9276489099720493, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.52}) +[2026-06-23 12:29:42,035][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.1253810077905655, 'train/grad_norm_pre_clip': 25.66637420654297, 'timing/data': 0.0005949600599706173, 'timing/model': 1.933787735993974, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 0.52}) +[2026-06-23 12:29:45,837][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.08698619902133942, 'train/grad_norm_pre_clip': 33.96344757080078, 'timing/data': 0.0005964059382677078, 'timing/model': 1.9296316789695993, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 0.52}) +[2026-06-23 12:29:49,641][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.07045084238052368, 'train/grad_norm_pre_clip': 21.861738204956055, 'timing/data': 0.0005998850101605058, 'timing/model': 1.9310518349520862, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 0.52}) +[2026-06-23 12:29:53,430][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.038482509553432465, 'train/grad_norm_pre_clip': 18.246740341186523, 'timing/data': 0.0007360709132626653, 'timing/model': 1.9188210329739377, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.52}) +[2026-06-23 12:29:57,222][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.08453724533319473, 'train/grad_norm_pre_clip': 36.64216995239258, 'timing/data': 0.0009747400181367993, 'timing/model': 1.9264470039634034, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 0.52}) +[2026-06-23 12:30:01,177][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.14527447521686554, 'train/grad_norm_pre_clip': 35.60320281982422, 'timing/data': 0.0006159000331535935, 'timing/model': 2.0895144790410995, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 0.52}) +[2026-06-23 12:30:04,993][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.0860942080616951, 'train/grad_norm_pre_clip': 27.910385131835938, 'timing/data': 0.0005828379653394222, 'timing/model': 1.9346128349425271, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 0.52}) +[2026-06-23 12:30:08,808][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.07425248622894287, 'train/grad_norm_pre_clip': 23.798324584960938, 'timing/data': 0.0005934089422225952, 'timing/model': 1.9317164609674364, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 0.52}) +[2026-06-23 12:30:12,599][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.15806645154953003, 'train/grad_norm_pre_clip': 37.5584716796875, 'timing/data': 0.000596871948800981, 'timing/model': 1.9265460500027984, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 0.52}) +[2026-06-23 12:30:16,417][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.12642033398151398, 'train/grad_norm_pre_clip': 70.12966918945312, 'timing/data': 0.000636005075648427, 'timing/model': 1.9374533289810643, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 0.52}) +[2026-06-23 12:30:20,219][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.15037083625793457, 'train/grad_norm_pre_clip': 91.61869049072266, 'timing/data': 0.0006589679978787899, 'timing/model': 1.9287131009623408, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 0.52}) +[2026-06-23 12:30:24,020][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.15579834580421448, 'train/grad_norm_pre_clip': 55.64860534667969, 'timing/data': 0.0007137020584195852, 'timing/model': 1.9240678910864517, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 0.52}) +[2026-06-23 12:30:27,847][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.1423521488904953, 'train/grad_norm_pre_clip': 39.599422454833984, 'timing/data': 0.0006011970108374953, 'timing/model': 1.937650123029016, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 0.52}) +[2026-06-23 12:30:31,650][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.03613264858722687, 'train/grad_norm_pre_clip': 24.833223342895508, 'timing/data': 0.0006043700268492103, 'timing/model': 1.9349786980310455, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.52}) +[2026-06-23 12:30:35,466][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.1165245994925499, 'train/grad_norm_pre_clip': 36.99222946166992, 'timing/data': 0.0008810940198600292, 'timing/model': 1.932324492954649, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 0.52}) +[2026-06-23 12:30:39,278][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.10335609316825867, 'train/grad_norm_pre_clip': 26.854013442993164, 'timing/data': 0.0006595579907298088, 'timing/model': 1.929813945083879, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 0.52}) +[2026-06-23 12:30:43,090][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.06570743769407272, 'train/grad_norm_pre_clip': 42.70673751831055, 'timing/data': 0.0006704420084133744, 'timing/model': 1.9335205589886755, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 0.52}) +[2026-06-23 12:30:46,897][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.07887090742588043, 'train/grad_norm_pre_clip': 24.406557083129883, 'timing/data': 0.0007867569802328944, 'timing/model': 1.9322940829442814, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 0.52}) +[2026-06-23 12:30:50,694][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.08131664246320724, 'train/grad_norm_pre_clip': 31.072704315185547, 'timing/data': 0.0007008729735389352, 'timing/model': 1.9300641440786421, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 0.52}) +[2026-06-23 12:30:54,511][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.21976487338542938, 'train/grad_norm_pre_clip': 84.81686401367188, 'timing/data': 0.0006481470772996545, 'timing/model': 1.947028171038255, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 0.52}) +[2026-06-23 12:30:58,312][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.12846526503562927, 'train/grad_norm_pre_clip': 52.42515182495117, 'timing/data': 0.00058121292386204, 'timing/model': 1.9279137690318748, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 0.53}) +[2026-06-23 12:31:02,109][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.06046418100595474, 'train/grad_norm_pre_clip': 27.244630813598633, 'timing/data': 0.0006034059915691614, 'timing/model': 1.9284033649601042, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 0.53}) +[2026-06-23 12:31:05,928][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.10888443887233734, 'train/grad_norm_pre_clip': 38.005680084228516, 'timing/data': 0.0006479650037363172, 'timing/model': 1.9431677429238334, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 0.53}) +[2026-06-23 12:31:09,742][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.06821537762880325, 'train/grad_norm_pre_clip': 24.03164291381836, 'timing/data': 0.000604870030656457, 'timing/model': 1.9394997239578515, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.53}) +[2026-06-23 12:31:13,555][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.1729428917169571, 'train/grad_norm_pre_clip': 67.04256439208984, 'timing/data': 0.0005823170067742467, 'timing/model': 1.933034903020598, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 0.53}) +[2026-06-23 12:31:17,353][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.06250099092721939, 'train/grad_norm_pre_clip': 20.755435943603516, 'timing/data': 0.0006117590237408876, 'timing/model': 1.9314127530669793, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 0.53}) +[2026-06-23 12:31:21,164][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.05153867229819298, 'train/grad_norm_pre_clip': 33.98590850830078, 'timing/data': 0.0008415200281888247, 'timing/model': 1.9339449730468914, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 0.53}) +[2026-06-23 12:31:24,968][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.08532310277223587, 'train/grad_norm_pre_clip': 20.269432067871094, 'timing/data': 0.0006087450310587883, 'timing/model': 1.9331683469936252, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 0.53}) +[2026-06-23 12:31:28,763][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.09917842596769333, 'train/grad_norm_pre_clip': 41.84101486206055, 'timing/data': 0.0008074939250946045, 'timing/model': 1.9209114339901134, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 0.53}) +[2026-06-23 12:31:32,565][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.1039348766207695, 'train/grad_norm_pre_clip': 59.15110397338867, 'timing/data': 0.0007874820148572326, 'timing/model': 1.9237112870905548, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 0.53}) +[2026-06-23 12:31:36,356][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.12081851810216904, 'train/grad_norm_pre_clip': 20.06239128112793, 'timing/data': 0.0006042200839146972, 'timing/model': 1.920884639956057, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 0.53}) +[2026-06-23 12:31:40,362][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.25445958971977234, 'train/grad_norm_pre_clip': 42.757789611816406, 'timing/data': 0.0007719120476394892, 'timing/model': 2.1388430560473353, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 0.53}) +[2026-06-23 12:31:44,164][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.22907404601573944, 'train/grad_norm_pre_clip': 67.55975341796875, 'timing/data': 0.0005958200199529529, 'timing/model': 1.928059246041812, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 0.53}) +[2026-06-23 12:31:47,965][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.08104466646909714, 'train/grad_norm_pre_clip': 38.081634521484375, 'timing/data': 0.0005918310489505529, 'timing/model': 1.92759597604163, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.53}) +[2026-06-23 12:31:51,787][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.09923762828111649, 'train/grad_norm_pre_clip': 69.412841796875, 'timing/data': 0.0005947990575805306, 'timing/model': 1.9397887479281053, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 0.53}) +[2026-06-23 12:31:55,599][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.10371299833059311, 'train/grad_norm_pre_clip': 18.620264053344727, 'timing/data': 0.0005973180523142219, 'timing/model': 1.9434407430235296, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 0.53}) +[2026-06-23 12:31:59,420][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.07757154107093811, 'train/grad_norm_pre_clip': 15.764437675476074, 'timing/data': 0.000775407999753952, 'timing/model': 1.944458800018765, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 0.53}) +[2026-06-23 12:32:03,273][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.03205205127596855, 'train/grad_norm_pre_clip': 45.522193908691406, 'timing/data': 0.0007275049574673176, 'timing/model': 1.9772168600466102, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.53}) +[2026-06-23 12:32:07,089][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.07187378406524658, 'train/grad_norm_pre_clip': 25.584247589111328, 'timing/data': 0.0008529400220140815, 'timing/model': 1.9376365538919345, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 0.53}) +[2026-06-23 12:32:10,889][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.10897977650165558, 'train/grad_norm_pre_clip': 57.10844802856445, 'timing/data': 0.0007350640371441841, 'timing/model': 1.929365432006307, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 0.53}) +[2026-06-23 12:32:14,871][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.04411481320858002, 'train/grad_norm_pre_clip': 12.153976440429688, 'timing/data': 0.0005971349310129881, 'timing/model': 1.9348040679469705, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 0.53}) +[2026-06-23 12:32:18,676][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.06627680361270905, 'train/grad_norm_pre_clip': 20.739412307739258, 'timing/data': 0.0006030519725754857, 'timing/model': 1.929939356050454, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 0.53}) +[2026-06-23 12:32:22,487][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.11212797462940216, 'train/grad_norm_pre_clip': 22.6208553314209, 'timing/data': 0.0006877109408378601, 'timing/model': 1.9299615490017459, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 0.53}) +[2026-06-23 12:32:26,299][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.1446181684732437, 'train/grad_norm_pre_clip': 23.93482208251953, 'timing/data': 0.0006102439947426319, 'timing/model': 1.9327714160317555, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.53}) +[2026-06-23 12:32:30,106][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.14568911492824554, 'train/grad_norm_pre_clip': 30.798564910888672, 'timing/data': 0.0006071249954402447, 'timing/model': 1.9341020500287414, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 0.53}) +[2026-06-23 12:32:33,930][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.09870608896017075, 'train/grad_norm_pre_clip': 36.577293395996094, 'timing/data': 0.0006065670168027282, 'timing/model': 1.944598876987584, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 0.53}) +[2026-06-23 12:32:37,767][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.09141513705253601, 'train/grad_norm_pre_clip': 22.113407135009766, 'timing/data': 0.000856079044751823, 'timing/model': 1.9592070409562439, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 0.54}) +[2026-06-23 12:32:41,576][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.09585060179233551, 'train/grad_norm_pre_clip': 40.46101379394531, 'timing/data': 0.0006385769229382277, 'timing/model': 1.9329076289432123, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 0.54}) +[2026-06-23 12:32:45,550][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.06562644988298416, 'train/grad_norm_pre_clip': 77.18466186523438, 'timing/data': 0.0005885089049115777, 'timing/model': 2.0973908010637388, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 0.54}) +[2026-06-23 12:32:49,352][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.11769678443670273, 'train/grad_norm_pre_clip': 58.03546905517578, 'timing/data': 0.0006286690477281809, 'timing/model': 1.9288804880343378, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 0.54}) +[2026-06-23 12:32:53,171][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.03495648130774498, 'train/grad_norm_pre_clip': 20.547340393066406, 'timing/data': 0.0006334129720926285, 'timing/model': 1.9452957879984751, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 0.54}) +[2026-06-23 12:32:56,984][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.10136730223894119, 'train/grad_norm_pre_clip': 44.72869873046875, 'timing/data': 0.0005865750135853887, 'timing/model': 1.936465332051739, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 0.54}) +[2026-06-23 12:33:00,803][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.11658484488725662, 'train/grad_norm_pre_clip': 31.18236541748047, 'timing/data': 0.000596824917010963, 'timing/model': 1.9443870439426973, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 0.54}) +[2026-06-23 12:33:04,605][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.12910708785057068, 'train/grad_norm_pre_clip': 29.076534271240234, 'timing/data': 0.0006132809212431312, 'timing/model': 1.926600601989776, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.54}) +[2026-06-23 12:33:08,423][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.051404088735580444, 'train/grad_norm_pre_clip': 26.957748413085938, 'timing/data': 0.0007299230201169848, 'timing/model': 1.9419280380243436, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 0.54}) +[2026-06-23 12:33:12,239][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.05464712530374527, 'train/grad_norm_pre_clip': 22.839277267456055, 'timing/data': 0.0006245450349524617, 'timing/model': 1.9327980580274016, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 0.54}) +[2026-06-23 12:33:16,051][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.11748120933771133, 'train/grad_norm_pre_clip': 61.04976272583008, 'timing/data': 0.0006204840028658509, 'timing/model': 1.9324997559888288, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 0.54}) +[2026-06-23 12:33:19,872][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.0567173957824707, 'train/grad_norm_pre_clip': 33.919715881347656, 'timing/data': 0.000606970046646893, 'timing/model': 1.9465720480075106, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 0.54}) +[2026-06-23 12:33:23,681][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.19752776622772217, 'train/grad_norm_pre_clip': 80.45245361328125, 'timing/data': 0.0005914990324527025, 'timing/model': 1.9305711940396577, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.54}) +[2026-06-23 12:33:27,496][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.13579383492469788, 'train/grad_norm_pre_clip': 100.01235961914062, 'timing/data': 0.0007904209196567535, 'timing/model': 1.9325655290158466, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 0.54}) +[2026-06-23 12:33:31,305][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.12736254930496216, 'train/grad_norm_pre_clip': 53.766910552978516, 'timing/data': 0.0005836100317537785, 'timing/model': 1.93029974703677, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 0.54}) +[2026-06-23 12:33:35,107][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.06802795827388763, 'train/grad_norm_pre_clip': 41.244415283203125, 'timing/data': 0.0005906850565224886, 'timing/model': 1.9274748449679464, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 0.54}) +[2026-06-23 12:33:38,916][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.07935655862092972, 'train/grad_norm_pre_clip': 40.72431945800781, 'timing/data': 0.0005903580458834767, 'timing/model': 1.933728787000291, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 0.54}) +[2026-06-23 12:33:42,758][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1400 +[2026-06-23 12:35:49,562][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1400 +[2026-06-23 12:35:49,563][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1400 +[2026-06-23 12:37:13,015][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1400 +[2026-06-23 12:37:13,016][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.15047070384025574, 'train/grad_norm_pre_clip': 93.9392318725586, 'eval/action_loss/seconds': 126.80152592097875, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07775171101093292, 'eval/latency_2/loss': 0.07775171101093292, 'eval/flappy/loss': 0.07775171101093292, 'eval/flappy/latency_2/loss': 0.07775171101093292, 'eval/flap_precision': 0.9732739420935412, 'eval/flap_recall': 0.8846153846153846, 'eval/flap_f1': 0.926829268292683, 'eval/flap_rate_model': 0.18708333333333332, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9268292682926829, 'eval/latency_2/flap_precision': 0.9732739420935412, 'eval/latency_2/flap_recall': 0.8846153846153846, 'eval/latency_2/flap_f1': 0.926829268292683, 'eval/latency_2/flap_rate_model': 0.18708333333333332, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9268292682926829, 'eval/action_classification/data_wait_seconds': 1.516738772392273, 'eval/action_classification/predict_seconds': 81.73882293701172, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006203199736773968, 'timing/model': 1.9425839590840042, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.54}) +[2026-06-23 12:37:16,801][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.09756443649530411, 'train/grad_norm_pre_clip': 43.780555725097656, 'timing/data': 0.0006967460503801703, 'timing/model': 1.9186218129470944, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 0.54}) +[2026-06-23 12:37:20,620][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.09982698410749435, 'train/grad_norm_pre_clip': 67.4901123046875, 'timing/data': 0.0006027390481904149, 'timing/model': 1.9480159820523113, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 0.54}) +[2026-06-23 12:37:24,410][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.16054049134254456, 'train/grad_norm_pre_clip': 46.234806060791016, 'timing/data': 0.0005938980029895902, 'timing/model': 1.9251217079581693, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 0.54}) +[2026-06-23 12:37:28,192][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.19863206148147583, 'train/grad_norm_pre_clip': 45.540504455566406, 'timing/data': 0.0005928060272708535, 'timing/model': 1.910581364063546, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 0.54}) +[2026-06-23 12:37:31,985][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.1726466566324234, 'train/grad_norm_pre_clip': 81.65869903564453, 'timing/data': 0.00064900703728199, 'timing/model': 1.9149022219935432, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 0.54}) +[2026-06-23 12:37:35,770][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.10976964235305786, 'train/grad_norm_pre_clip': 24.085569381713867, 'timing/data': 0.0005908709717914462, 'timing/model': 1.9207877480657771, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 0.54}) +[2026-06-23 12:37:39,605][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.1584283709526062, 'train/grad_norm_pre_clip': 54.43094253540039, 'timing/data': 0.0006018480053171515, 'timing/model': 1.9677138249389827, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 0.54}) +[2026-06-23 12:37:43,378][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.15080104768276215, 'train/grad_norm_pre_clip': 28.856828689575195, 'timing/data': 0.0006061129970476031, 'timing/model': 1.9136330760084093, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 0.54}) +[2026-06-23 12:37:47,163][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.06117112562060356, 'train/grad_norm_pre_clip': 25.176719665527344, 'timing/data': 0.0007807740475982428, 'timing/model': 1.9188427480403334, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 0.55}) +[2026-06-23 12:37:50,961][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.09614110738039017, 'train/grad_norm_pre_clip': 35.48111343383789, 'timing/data': 0.0006060459418222308, 'timing/model': 1.931407060008496, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.55}) +[2026-06-23 12:37:54,759][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.11825386434793472, 'train/grad_norm_pre_clip': 47.11798858642578, 'timing/data': 0.0009031059453263879, 'timing/model': 1.9233446649741381, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 0.55}) +[2026-06-23 12:37:58,553][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.09382004290819168, 'train/grad_norm_pre_clip': 20.187286376953125, 'timing/data': 0.0006172489374876022, 'timing/model': 1.9255956639535725, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 0.55}) +[2026-06-23 12:38:02,354][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.0665750578045845, 'train/grad_norm_pre_clip': 19.377878189086914, 'timing/data': 0.0006042929599061608, 'timing/model': 1.925660380977206, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 0.55}) +[2026-06-23 12:38:06,153][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.06588941067457199, 'train/grad_norm_pre_clip': 26.333171844482422, 'timing/data': 0.0006063500186428428, 'timing/model': 1.925407467992045, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 0.55}) +[2026-06-23 12:38:09,956][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.06228545308113098, 'train/grad_norm_pre_clip': 50.532958984375, 'timing/data': 0.0006125009385868907, 'timing/model': 1.9321851140120998, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 0.55}) +[2026-06-23 12:38:13,863][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.1617535799741745, 'train/grad_norm_pre_clip': 21.805374145507812, 'timing/data': 0.0007249319460242987, 'timing/model': 2.034480069996789, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 0.55}) +[2026-06-23 12:38:17,666][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.23381531238555908, 'train/grad_norm_pre_clip': 25.214509963989258, 'timing/data': 0.0008324139053002, 'timing/model': 1.9328081379644573, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 0.55}) +[2026-06-23 12:38:21,467][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.10086050629615784, 'train/grad_norm_pre_clip': 27.464616775512695, 'timing/data': 0.0005872390465810895, 'timing/model': 1.9310351910535246, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 0.55}) +[2026-06-23 12:38:25,274][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.1101597398519516, 'train/grad_norm_pre_clip': 34.627017974853516, 'timing/data': 0.0006053789984434843, 'timing/model': 1.9301455640234053, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 0.55}) +[2026-06-23 12:38:29,066][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.1254526674747467, 'train/grad_norm_pre_clip': 47.20245361328125, 'timing/data': 0.0006001009605824947, 'timing/model': 1.9288916899822652, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.55}) +[2026-06-23 12:38:32,847][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.06947220861911774, 'train/grad_norm_pre_clip': 27.30794906616211, 'timing/data': 0.0006820999551564455, 'timing/model': 1.9196111720521003, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 0.55}) +[2026-06-23 12:38:36,661][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.05665801465511322, 'train/grad_norm_pre_clip': 27.17047691345215, 'timing/data': 0.0007013200083747506, 'timing/model': 1.9362108609639108, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 0.55}) +[2026-06-23 12:38:40,463][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.12661316990852356, 'train/grad_norm_pre_clip': 25.564613342285156, 'timing/data': 0.000774251064285636, 'timing/model': 1.9273095240350813, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.55}) +[2026-06-23 12:38:44,260][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.16298151016235352, 'train/grad_norm_pre_clip': 39.5893669128418, 'timing/data': 0.000586520996876061, 'timing/model': 1.9287663020659238, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 0.55}) +[2026-06-23 12:38:48,212][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.12742967903614044, 'train/grad_norm_pre_clip': 19.861637115478516, 'timing/data': 0.0005889319581910968, 'timing/model': 1.9320242090616375, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 0.55}) +[2026-06-23 12:38:52,027][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.06231852248311043, 'train/grad_norm_pre_clip': 18.222686767578125, 'timing/data': 0.0005945699522271752, 'timing/model': 1.9371487949974835, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 0.55}) +[2026-06-23 12:38:55,853][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.1364036649465561, 'train/grad_norm_pre_clip': 29.510650634765625, 'timing/data': 0.0005949740298092365, 'timing/model': 1.938431302900426, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 0.55}) +[2026-06-23 12:38:59,691][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.11619819700717926, 'train/grad_norm_pre_clip': 20.864261627197266, 'timing/data': 0.0006014100508764386, 'timing/model': 1.9743205159902573, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 0.55}) +[2026-06-23 12:39:03,503][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.07212873548269272, 'train/grad_norm_pre_clip': 33.891178131103516, 'timing/data': 0.0006066709756851196, 'timing/model': 1.928758793976158, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 0.55}) +[2026-06-23 12:39:07,293][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.04445253685116768, 'train/grad_norm_pre_clip': 14.99811840057373, 'timing/data': 0.0005974129308015108, 'timing/model': 1.9217318770242855, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.55}) +[2026-06-23 12:39:11,095][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.06194669380784035, 'train/grad_norm_pre_clip': 21.081647872924805, 'timing/data': 0.0010532510932534933, 'timing/model': 1.931781234103255, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 0.55}) +[2026-06-23 12:39:14,902][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.08306008577346802, 'train/grad_norm_pre_clip': 44.93587112426758, 'timing/data': 0.0006010429933667183, 'timing/model': 1.9306815169984475, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 0.55}) +[2026-06-23 12:39:18,797][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.093806192278862, 'train/grad_norm_pre_clip': 42.96489715576172, 'timing/data': 0.00063562102150172, 'timing/model': 2.0160382780013606, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 0.55}) +[2026-06-23 12:39:22,621][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.1371644288301468, 'train/grad_norm_pre_clip': 21.298782348632812, 'timing/data': 0.0006508700316771865, 'timing/model': 1.9429956190288067, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 0.55}) +[2026-06-23 12:39:26,429][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.04646393284201622, 'train/grad_norm_pre_clip': 16.367298126220703, 'timing/data': 0.0006020379951223731, 'timing/model': 1.9330750879598781, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 0.56}) +[2026-06-23 12:39:30,227][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.08004998415708542, 'train/grad_norm_pre_clip': 47.94175338745117, 'timing/data': 0.0006875899853184819, 'timing/model': 1.9352589109912515, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 0.56}) +[2026-06-23 12:39:34,021][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.1780514270067215, 'train/grad_norm_pre_clip': 34.70315170288086, 'timing/data': 0.0006063759792596102, 'timing/model': 1.933538265991956, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 0.56}) +[2026-06-23 12:39:37,809][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.07198789715766907, 'train/grad_norm_pre_clip': 46.8400764465332, 'timing/data': 0.0006316360086202621, 'timing/model': 1.923913941020146, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 0.56}) +[2026-06-23 12:39:41,596][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.08563140779733658, 'train/grad_norm_pre_clip': 39.174598693847656, 'timing/data': 0.0006529620150104165, 'timing/model': 1.9187016300857067, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 0.56}) +[2026-06-23 12:39:45,392][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.15300427377223969, 'train/grad_norm_pre_clip': 35.516456604003906, 'timing/data': 0.0006056269630789757, 'timing/model': 1.9323946730000898, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.56}) +[2026-06-23 12:39:49,232][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.06700082123279572, 'train/grad_norm_pre_clip': 36.226139068603516, 'timing/data': 0.000693426001816988, 'timing/model': 1.9348970549181104, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 0.56}) +[2026-06-23 12:39:53,227][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.13399849832057953, 'train/grad_norm_pre_clip': 29.479400634765625, 'timing/data': 0.0006497049471363425, 'timing/model': 2.1026096729328856, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 0.56}) +[2026-06-23 12:39:57,038][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.15601125359535217, 'train/grad_norm_pre_clip': 60.46598815917969, 'timing/data': 0.0005864560371264815, 'timing/model': 1.9358127360465005, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 0.56}) +[2026-06-23 12:40:00,835][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.0922410786151886, 'train/grad_norm_pre_clip': 43.984779357910156, 'timing/data': 0.0005979279521852732, 'timing/model': 1.9326893941033632, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 0.56}) +[2026-06-23 12:40:04,639][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.1953437179327011, 'train/grad_norm_pre_clip': 41.769657135009766, 'timing/data': 0.000609732000157237, 'timing/model': 1.9299588659778237, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 0.56}) +[2026-06-23 12:40:08,456][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.1657009720802307, 'train/grad_norm_pre_clip': 41.60732650756836, 'timing/data': 0.0008524390868842602, 'timing/model': 1.9461771269561723, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 0.56}) +[2026-06-23 12:40:12,255][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.030158942565321922, 'train/grad_norm_pre_clip': 43.09403991699219, 'timing/data': 0.0005917850648984313, 'timing/model': 1.9299314480740577, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 0.56}) +[2026-06-23 12:40:16,045][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.08554241061210632, 'train/grad_norm_pre_clip': 27.922964096069336, 'timing/data': 0.0006153139984235168, 'timing/model': 1.9240978279849514, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 0.56}) +[2026-06-23 12:40:19,851][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.10988222807645798, 'train/grad_norm_pre_clip': 39.874305725097656, 'timing/data': 0.0006004050374031067, 'timing/model': 1.9337983399163932, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 0.56}) +[2026-06-23 12:40:23,662][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.14073358476161957, 'train/grad_norm_pre_clip': 66.33149719238281, 'timing/data': 0.0005956160603091121, 'timing/model': 1.9332605520030484, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.56}) +[2026-06-23 12:40:27,618][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.14009425044059753, 'train/grad_norm_pre_clip': 54.252410888671875, 'timing/data': 0.0006907000206410885, 'timing/model': 1.9227949780179188, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 0.56}) +[2026-06-23 12:40:31,414][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.1094265952706337, 'train/grad_norm_pre_clip': 48.44080352783203, 'timing/data': 0.0006937789730727673, 'timing/model': 1.9206703470554203, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 0.56}) +[2026-06-23 12:40:35,212][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.05445479974150658, 'train/grad_norm_pre_clip': 36.27862548828125, 'timing/data': 0.0009880040306597948, 'timing/model': 1.9305835879640654, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 0.56}) +[2026-06-23 12:40:39,023][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.07863932102918625, 'train/grad_norm_pre_clip': 88.7486343383789, 'timing/data': 0.0006006190087646246, 'timing/model': 1.9366926540387794, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 0.56}) +[2026-06-23 12:40:42,842][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.11847642809152603, 'train/grad_norm_pre_clip': 39.54375076293945, 'timing/data': 0.0006386779714375734, 'timing/model': 1.93238051305525, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 0.56}) +[2026-06-23 12:40:46,657][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.1410936415195465, 'train/grad_norm_pre_clip': 29.200664520263672, 'timing/data': 0.0009180610068142414, 'timing/model': 1.9377780070062727, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 0.56}) +[2026-06-23 12:40:50,476][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.0702708289027214, 'train/grad_norm_pre_clip': 18.134593963623047, 'timing/data': 0.0005895780632272363, 'timing/model': 1.9350062120938674, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 0.56}) +[2026-06-23 12:40:54,286][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.13017718493938446, 'train/grad_norm_pre_clip': 26.310293197631836, 'timing/data': 0.0005899419775232673, 'timing/model': 1.9377635900164023, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 0.56}) +[2026-06-23 12:40:58,224][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.07416840642690659, 'train/grad_norm_pre_clip': 11.665374755859375, 'timing/data': 0.0005976069951429963, 'timing/model': 2.063695412944071, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 0.56}) +[2026-06-23 12:41:02,030][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.09016845375299454, 'train/grad_norm_pre_clip': 48.66819381713867, 'timing/data': 0.0005939899710938334, 'timing/model': 1.9334004839183763, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.57}) +[2026-06-23 12:41:05,839][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.07525409758090973, 'train/grad_norm_pre_clip': 12.087722778320312, 'timing/data': 0.0007078219205141068, 'timing/model': 1.9376382749760523, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 0.57}) +[2026-06-23 12:41:09,657][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.10124260187149048, 'train/grad_norm_pre_clip': 21.976484298706055, 'timing/data': 0.000700299977324903, 'timing/model': 1.93472064496018, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 0.57}) +[2026-06-23 12:41:13,479][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.1273467093706131, 'train/grad_norm_pre_clip': 44.790863037109375, 'timing/data': 0.0006337659433484077, 'timing/model': 1.935783245950006, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 0.57}) +[2026-06-23 12:41:17,299][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.12644904851913452, 'train/grad_norm_pre_clip': 27.111549377441406, 'timing/data': 0.0006204060046002269, 'timing/model': 1.9366053170524538, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 0.57}) +[2026-06-23 12:41:21,112][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.0891832485795021, 'train/grad_norm_pre_clip': 24.873971939086914, 'timing/data': 0.0006494979606941342, 'timing/model': 1.9312313889386132, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.57}) +[2026-06-23 12:41:24,941][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.07592593133449554, 'train/grad_norm_pre_clip': 23.135360717773438, 'timing/data': 0.0006749719614163041, 'timing/model': 1.9376406810479239, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 0.57}) +[2026-06-23 12:41:28,760][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.12763139605522156, 'train/grad_norm_pre_clip': 60.26329040527344, 'timing/data': 0.0005926260491833091, 'timing/model': 1.9391251449706033, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 0.57}) +[2026-06-23 12:41:32,703][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.08163508027791977, 'train/grad_norm_pre_clip': 41.228904724121094, 'timing/data': 0.000664462917484343, 'timing/model': 1.9244369741063565, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 0.57}) +[2026-06-23 12:41:36,523][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.023250754922628403, 'train/grad_norm_pre_clip': 25.407106399536133, 'timing/data': 0.000956367002800107, 'timing/model': 1.9423742140643299, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 0.57}) +[2026-06-23 12:41:40,319][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.15683358907699585, 'train/grad_norm_pre_clip': 51.083824157714844, 'timing/data': 0.0006197181064635515, 'timing/model': 1.929492020048201, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.57}) +[2026-06-23 12:41:44,122][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.15314088761806488, 'train/grad_norm_pre_clip': 85.27124786376953, 'timing/data': 0.0006153949070721865, 'timing/model': 1.9326507310615852, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 0.57}) +[2026-06-23 12:41:47,922][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.06631984561681747, 'train/grad_norm_pre_clip': 34.2591667175293, 'timing/data': 0.000608947011642158, 'timing/model': 1.9244116289773956, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.57}) +[2026-06-23 12:41:51,734][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.036695994436740875, 'train/grad_norm_pre_clip': 20.39628791809082, 'timing/data': 0.0005873379996046424, 'timing/model': 1.9416940259980038, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 0.57}) +[2026-06-23 12:41:55,563][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.17215760052204132, 'train/grad_norm_pre_clip': 65.40191650390625, 'timing/data': 0.0006136270239949226, 'timing/model': 1.9350071420194581, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 0.57}) +[2026-06-23 12:41:59,378][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.19067586958408356, 'train/grad_norm_pre_clip': 98.86128997802734, 'timing/data': 0.0006034440593793988, 'timing/model': 1.9397787969792262, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 0.57}) +[2026-06-23 12:42:03,271][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.14521904289722443, 'train/grad_norm_pre_clip': 80.37805938720703, 'timing/data': 0.0006319290259853005, 'timing/model': 2.013014190015383, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 0.57}) +[2026-06-23 12:42:07,097][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.04958798736333847, 'train/grad_norm_pre_clip': 38.39178466796875, 'timing/data': 0.0006044040201231837, 'timing/model': 1.9447726799407974, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 0.57}) +[2026-06-23 12:42:10,901][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.06799878925085068, 'train/grad_norm_pre_clip': 29.50527572631836, 'timing/data': 0.0006800350965932012, 'timing/model': 1.9253776039695367, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 0.57}) +[2026-06-23 12:42:14,712][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.1686316430568695, 'train/grad_norm_pre_clip': 43.3852424621582, 'timing/data': 0.0006778420647606254, 'timing/model': 1.9392781240167096, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 0.57}) +[2026-06-23 12:42:18,524][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.08992648124694824, 'train/grad_norm_pre_clip': 34.19722366333008, 'timing/data': 0.000595402903854847, 'timing/model': 1.9341162259224802, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.57}) +[2026-06-23 12:42:22,333][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.0658114030957222, 'train/grad_norm_pre_clip': 39.171485900878906, 'timing/data': 0.0006346810841932893, 'timing/model': 1.932202967000194, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 0.57}) +[2026-06-23 12:42:26,153][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.246307834982872, 'train/grad_norm_pre_clip': 40.367088317871094, 'timing/data': 0.0006046139169484377, 'timing/model': 1.9418631179723889, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 0.57}) +[2026-06-23 12:42:29,964][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.0657600536942482, 'train/grad_norm_pre_clip': 25.310867309570312, 'timing/data': 0.0007619099924340844, 'timing/model': 1.933307558996603, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 0.57}) +[2026-06-23 12:42:33,782][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.027751628309488297, 'train/grad_norm_pre_clip': 45.01176452636719, 'timing/data': 0.0006061389576643705, 'timing/model': 1.941982424003072, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 0.57}) +[2026-06-23 12:42:37,655][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.2611890435218811, 'train/grad_norm_pre_clip': 74.32353973388672, 'timing/data': 0.0005766400136053562, 'timing/model': 1.9383006209973246, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 0.57}) +[2026-06-23 12:42:41,465][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.08779677003622055, 'train/grad_norm_pre_clip': 35.107704162597656, 'timing/data': 0.0006395250093191862, 'timing/model': 1.9416960770031437, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 0.58}) +[2026-06-23 12:42:45,262][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.10302330553531647, 'train/grad_norm_pre_clip': 36.68837356567383, 'timing/data': 0.0005943739088252187, 'timing/model': 1.9292536560678855, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 0.58}) +[2026-06-23 12:42:49,064][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.05790914595127106, 'train/grad_norm_pre_clip': 34.1182975769043, 'timing/data': 0.0007861290359869599, 'timing/model': 1.9314528099494055, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 0.58}) +[2026-06-23 12:42:52,890][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.07007559388875961, 'train/grad_norm_pre_clip': 26.164644241333008, 'timing/data': 0.0005944340955466032, 'timing/model': 1.9487947069574147, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 0.58}) +[2026-06-23 12:42:56,699][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.052289992570877075, 'train/grad_norm_pre_clip': 39.744449615478516, 'timing/data': 0.0005895369686186314, 'timing/model': 1.93785820400808, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.58}) +[2026-06-23 12:43:00,506][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.10077278316020966, 'train/grad_norm_pre_clip': 43.086666107177734, 'timing/data': 0.000582739943638444, 'timing/model': 1.937419917085208, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 0.58}) +[2026-06-23 12:43:04,309][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.07486189901828766, 'train/grad_norm_pre_clip': 41.54199981689453, 'timing/data': 0.0006493729306384921, 'timing/model': 1.9288714609574527, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 0.58}) +[2026-06-23 12:43:08,155][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.13067607581615448, 'train/grad_norm_pre_clip': 20.600616455078125, 'timing/data': 0.0005782080115750432, 'timing/model': 1.969453016994521, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 0.58}) +[2026-06-23 12:43:11,983][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.13616234064102173, 'train/grad_norm_pre_clip': 35.90162658691406, 'timing/data': 0.0008335849270224571, 'timing/model': 1.9410140520194545, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 0.58}) +[2026-06-23 12:43:15,795][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.02920176275074482, 'train/grad_norm_pre_clip': 22.865421295166016, 'timing/data': 0.0005922049749642611, 'timing/model': 1.9359313769964501, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 0.58}) +[2026-06-23 12:43:19,607][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.10215477645397186, 'train/grad_norm_pre_clip': 26.305082321166992, 'timing/data': 0.0006201289361342788, 'timing/model': 1.9272004079539329, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 0.58}) +[2026-06-23 12:43:23,410][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.05208701267838478, 'train/grad_norm_pre_clip': 42.62158966064453, 'timing/data': 0.0006035380065441132, 'timing/model': 1.9361908789724112, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 0.58}) +[2026-06-23 12:43:27,235][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.06805519014596939, 'train/grad_norm_pre_clip': 30.254817962646484, 'timing/data': 0.0006010510260239244, 'timing/model': 1.939937026007101, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 0.58}) +[2026-06-23 12:43:31,053][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.2172573357820511, 'train/grad_norm_pre_clip': 38.004180908203125, 'timing/data': 0.0006038519786670804, 'timing/model': 1.935758821084164, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 0.58}) +[2026-06-23 12:43:34,868][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1500 +[2026-06-23 12:45:43,198][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1500 +[2026-06-23 12:45:43,199][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1500 +[2026-06-23 12:47:05,031][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1500 +[2026-06-23 12:52:47,500][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-23 12:52:47,501][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.11210650950670242, 'train/grad_norm_pre_clip': 25.527158737182617, 'eval/action_loss/seconds': 128.32660531403963, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04776176065206528, 'eval/latency_2/loss': 0.04776176065206528, 'eval/flappy/loss': 0.04776176065206528, 'eval/flappy/latency_2/loss': 0.04776176065206528, 'eval/flap_precision': 0.9359477124183007, 'eval/flap_recall': 0.9662618083670715, 'eval/flap_f1': 0.950863213811421, 'eval/flap_rate_model': 0.2125, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9508632138114209, 'eval/latency_2/flap_precision': 0.9359477124183007, 'eval/latency_2/flap_recall': 0.9662618083670715, 'eval/latency_2/flap_f1': 0.950863213811421, 'eval/latency_2/flap_rate_model': 0.2125, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9508632138114209, 'eval/action_classification/data_wait_seconds': 1.361587405204773, 'eval/action_classification/predict_seconds': 80.25891876220703, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 211.0200000000097, 'rl_games_eval/mid_train/mean_length': 1717.2, 'rl_games_eval/mid_train/std_reward': 195.658277616882, 'rl_games_eval/mid_train/std_length': 1577.5296383903537, 'rl_games_eval/mid_train/macro_mean_reward': 211.0200000000097, 'rl_games_eval/mid_train/macro_mean_length': 1717.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 211.0200000000097, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1717.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 195.658277616882, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1577.5296383903537, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 211.0200000000097, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005983239971101284, 'timing/model': 1.936922724940814, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.58}) +[2026-06-23 12:52:47,501][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_1500_state +[2026-06-23 12:52:47,503][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 12:52:51,262][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_1500_state/model.safetensors +[2026-06-23 12:52:54,243][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_1500_state/optimizer.bin +[2026-06-23 12:52:54,243][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_1500_state/sampler.bin +[2026-06-23 12:52:54,243][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-23 12:52:54,244][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-23 12:52:54,277][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-23 12:52:54,278][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 12:52:54,289][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 12:52:58,012][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.17618919909000397, 'train/grad_norm_pre_clip': 59.5999641418457, 'timing/data': 0.0007306140614673495, 'timing/model': 1.897691531921737, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 0.58}) +[2026-06-23 12:53:01,749][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.04465388134121895, 'train/grad_norm_pre_clip': 20.82267951965332, 'timing/data': 0.0006071329116821289, 'timing/model': 1.8929496880155057, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 0.58}) +[2026-06-23 12:53:05,476][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.19995254278182983, 'train/grad_norm_pre_clip': 75.87925720214844, 'timing/data': 0.0005903959972783923, 'timing/model': 1.8935221430147067, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 0.58}) +[2026-06-23 12:53:09,232][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.1346377432346344, 'train/grad_norm_pre_clip': 52.5468635559082, 'timing/data': 0.0006149410037323833, 'timing/model': 1.9162002849625424, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 0.58}) +[2026-06-23 12:53:12,987][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.05167283117771149, 'train/grad_norm_pre_clip': 17.929359436035156, 'timing/data': 0.0006238960195332766, 'timing/model': 1.9134677579859272, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 0.58}) +[2026-06-23 12:53:16,757][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.043010152876377106, 'train/grad_norm_pre_clip': 18.604387283325195, 'timing/data': 0.0005805420223623514, 'timing/model': 1.9186291770311072, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 0.58}) +[2026-06-23 12:53:20,704][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.14699164032936096, 'train/grad_norm_pre_clip': 22.692920684814453, 'timing/data': 0.0005754309240728617, 'timing/model': 1.9088052050210536, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 0.58}) +[2026-06-23 12:53:24,467][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.1987597793340683, 'train/grad_norm_pre_clip': 47.440826416015625, 'timing/data': 0.0008277620654553175, 'timing/model': 1.909625491942279, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 0.58}) +[2026-06-23 12:53:28,230][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.036395616829395294, 'train/grad_norm_pre_clip': 24.203954696655273, 'timing/data': 0.0008565880125388503, 'timing/model': 1.9138051921036094, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 0.58}) +[2026-06-23 12:53:32,003][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.2553020119667053, 'train/grad_norm_pre_clip': 46.310550689697266, 'timing/data': 0.0006535920547321439, 'timing/model': 1.9166308989515528, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.58}) +[2026-06-23 12:53:35,767][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.10020636767148972, 'train/grad_norm_pre_clip': 41.505226135253906, 'timing/data': 0.0007610640022903681, 'timing/model': 1.9088303809985518, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 0.58}) +[2026-06-23 12:53:39,538][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.05637870728969574, 'train/grad_norm_pre_clip': 43.397621154785156, 'timing/data': 0.0006036949343979359, 'timing/model': 1.9134149609599262, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 0.59}) +[2026-06-23 12:53:43,317][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.36326536536216736, 'train/grad_norm_pre_clip': 84.06288146972656, 'timing/data': 0.00060408900026232, 'timing/model': 1.9276176149724051, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 0.59}) +[2026-06-23 12:53:47,093][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.14692507684230804, 'train/grad_norm_pre_clip': 55.204586029052734, 'timing/data': 0.0006238389760255814, 'timing/model': 1.9176987109240144, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 0.59}) +[2026-06-23 12:53:51,051][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.0547664649784565, 'train/grad_norm_pre_clip': 30.942768096923828, 'timing/data': 0.0006030440563336015, 'timing/model': 2.0870517680887133, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 0.59}) +[2026-06-23 12:53:54,834][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.09248198568820953, 'train/grad_norm_pre_clip': 55.01992416381836, 'timing/data': 0.0006041050655767322, 'timing/model': 1.9273491080384701, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 0.59}) +[2026-06-23 12:53:58,618][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.023962965235114098, 'train/grad_norm_pre_clip': 33.31173324584961, 'timing/data': 0.0008184260223060846, 'timing/model': 1.9177947440184653, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 0.59}) +[2026-06-23 12:54:02,409][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.057005561888217926, 'train/grad_norm_pre_clip': 24.698034286499023, 'timing/data': 0.000599527033045888, 'timing/model': 1.925612869905308, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 0.59}) +[2026-06-23 12:54:06,193][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.044660188257694244, 'train/grad_norm_pre_clip': 27.881668090820312, 'timing/data': 0.0008989080088213086, 'timing/model': 1.9159820239292458, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 0.59}) +[2026-06-23 12:54:09,977][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.1056361049413681, 'train/grad_norm_pre_clip': 30.53412437438965, 'timing/data': 0.0006066259229555726, 'timing/model': 1.9165917040081695, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.59}) +[2026-06-23 12:54:13,773][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.17312054336071014, 'train/grad_norm_pre_clip': 42.89502716064453, 'timing/data': 0.0006176100578159094, 'timing/model': 1.928730295971036, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.59}) +[2026-06-23 12:54:17,574][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.13566572964191437, 'train/grad_norm_pre_clip': 43.70221710205078, 'timing/data': 0.0008253969717770815, 'timing/model': 1.9149766400223598, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 0.59}) +[2026-06-23 12:54:21,369][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.076093390583992, 'train/grad_norm_pre_clip': 28.47733497619629, 'timing/data': 0.0008011949248611927, 'timing/model': 1.9185261380625889, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 0.59}) +[2026-06-23 12:54:25,301][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.11329963058233261, 'train/grad_norm_pre_clip': 24.608356475830078, 'timing/data': 0.0005862410180270672, 'timing/model': 1.925740735954605, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 0.59}) +[2026-06-23 12:54:29,098][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.10619854927062988, 'train/grad_norm_pre_clip': 35.44049835205078, 'timing/data': 0.0005870100576430559, 'timing/model': 1.9309736520517617, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 0.59}) +[2026-06-23 12:54:32,883][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.05484309792518616, 'train/grad_norm_pre_clip': 47.147586822509766, 'timing/data': 0.0005837059579789639, 'timing/model': 1.921887725009583, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 0.59}) +[2026-06-23 12:54:36,679][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.05227452516555786, 'train/grad_norm_pre_clip': 18.12979507446289, 'timing/data': 0.0005899800453335047, 'timing/model': 1.926795858074911, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 0.59}) +[2026-06-23 12:54:40,460][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.03976890444755554, 'train/grad_norm_pre_clip': 17.120088577270508, 'timing/data': 0.0005916079971939325, 'timing/model': 1.922113404958509, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 0.59}) +[2026-06-23 12:54:44,267][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.1661178469657898, 'train/grad_norm_pre_clip': 41.08469009399414, 'timing/data': 0.0005791459698230028, 'timing/model': 1.9377087850589305, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 0.59}) +[2026-06-23 12:54:48,060][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.1894501894712448, 'train/grad_norm_pre_clip': 47.14851760864258, 'timing/data': 0.0006180029595270753, 'timing/model': 1.9271579319611192, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.59}) +[2026-06-23 12:54:51,863][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.03799634799361229, 'train/grad_norm_pre_clip': 33.440250396728516, 'timing/data': 0.0005953989457339048, 'timing/model': 1.9254022460663691, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 0.59}) +[2026-06-23 12:54:55,851][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.06409840285778046, 'train/grad_norm_pre_clip': 23.681026458740234, 'timing/data': 0.0005834989715367556, 'timing/model': 2.115561606013216, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 0.59}) +[2026-06-23 12:54:59,657][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.13339863717556, 'train/grad_norm_pre_clip': 28.042743682861328, 'timing/data': 0.0005844050319865346, 'timing/model': 1.9301602629711851, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 0.59}) +[2026-06-23 12:55:03,457][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.08053233474493027, 'train/grad_norm_pre_clip': 28.59768295288086, 'timing/data': 0.0009023519232869148, 'timing/model': 1.9350804090499878, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 0.59}) +[2026-06-23 12:55:07,256][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.10333669930696487, 'train/grad_norm_pre_clip': 33.50314712524414, 'timing/data': 0.0006935920100659132, 'timing/model': 1.9215209680842236, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.59}) +[2026-06-23 12:55:11,047][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.09567943215370178, 'train/grad_norm_pre_clip': 23.895082473754883, 'timing/data': 0.0006626979447901249, 'timing/model': 1.9255567520158365, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 0.59}) +[2026-06-23 12:55:14,850][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.11637634038925171, 'train/grad_norm_pre_clip': 49.005645751953125, 'timing/data': 0.0006720940582454205, 'timing/model': 1.9345402719918638, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 0.59}) +[2026-06-23 12:55:18,639][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.0982382670044899, 'train/grad_norm_pre_clip': 24.99113655090332, 'timing/data': 0.0006014059763401747, 'timing/model': 1.924625796964392, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 0.6}) +[2026-06-23 12:55:22,439][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.09351517260074615, 'train/grad_norm_pre_clip': 28.046977996826172, 'timing/data': 0.0009162410860881209, 'timing/model': 1.9278422640636563, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 0.6}) +[2026-06-23 12:55:26,229][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.18551133573055267, 'train/grad_norm_pre_clip': 37.187686920166016, 'timing/data': 0.000592566910199821, 'timing/model': 1.9269950560992584, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.6}) +[2026-06-23 12:55:30,227][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.057218171656131744, 'train/grad_norm_pre_clip': 48.372440338134766, 'timing/data': 0.0006158170290291309, 'timing/model': 1.9317418369464576, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 0.6}) +[2026-06-23 12:55:34,028][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.30969947576522827, 'train/grad_norm_pre_clip': 29.404306411743164, 'timing/data': 0.0006139399483799934, 'timing/model': 1.9358543650014326, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 0.6}) +[2026-06-23 12:55:37,822][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.10656702518463135, 'train/grad_norm_pre_clip': 26.00809669494629, 'timing/data': 0.0006482129683718085, 'timing/model': 1.9262471959227696, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 0.6}) +[2026-06-23 12:55:41,634][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.043779902160167694, 'train/grad_norm_pre_clip': 38.59077072143555, 'timing/data': 0.0005815719487145543, 'timing/model': 1.9343356350436807, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 0.6}) +[2026-06-23 12:55:45,455][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.24545244872570038, 'train/grad_norm_pre_clip': 31.612157821655273, 'timing/data': 0.000738644041121006, 'timing/model': 1.948178416932933, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 0.6}) +[2026-06-23 12:55:49,275][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.07234707474708557, 'train/grad_norm_pre_clip': 30.184890747070312, 'timing/data': 0.0005914510693401098, 'timing/model': 1.9410834449809045, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 0.6}) +[2026-06-23 12:55:53,153][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.10590648651123047, 'train/grad_norm_pre_clip': 20.754005432128906, 'timing/data': 0.0006355070509016514, 'timing/model': 1.9884830749360844, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 0.6}) +[2026-06-23 12:55:56,944][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.13301755487918854, 'train/grad_norm_pre_clip': 45.0483283996582, 'timing/data': 0.0005970200290903449, 'timing/model': 1.9279622649773955, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 0.6}) +[2026-06-23 12:56:00,755][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.043550051748752594, 'train/grad_norm_pre_clip': 12.264361381530762, 'timing/data': 0.0006661799270659685, 'timing/model': 1.9239482909906656, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 0.6}) +[2026-06-23 12:56:04,746][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.08404216170310974, 'train/grad_norm_pre_clip': 48.14045333862305, 'timing/data': 0.0006781769916415215, 'timing/model': 1.9293484470108524, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.6}) +[2026-06-23 12:56:08,559][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.04974203556776047, 'train/grad_norm_pre_clip': 53.818763732910156, 'timing/data': 0.0005998640554025769, 'timing/model': 1.9475028710439801, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 0.6}) +[2026-06-23 12:56:12,363][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.06191479414701462, 'train/grad_norm_pre_clip': 21.816362380981445, 'timing/data': 0.0006001460133120418, 'timing/model': 1.9308595689944923, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 0.6}) +[2026-06-23 12:56:16,158][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.11724618077278137, 'train/grad_norm_pre_clip': 23.670713424682617, 'timing/data': 0.0005903959972783923, 'timing/model': 1.9251775559969246, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 0.6}) +[2026-06-23 12:56:19,966][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.11543376743793488, 'train/grad_norm_pre_clip': 72.5661392211914, 'timing/data': 0.0006081439787521958, 'timing/model': 1.9307192949345335, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 0.6}) +[2026-06-23 12:56:23,770][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.17131581902503967, 'train/grad_norm_pre_clip': 53.44819259643555, 'timing/data': 0.000739614013582468, 'timing/model': 1.9296245800796896, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 0.6}) +[2026-06-23 12:56:27,600][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.07889744639396667, 'train/grad_norm_pre_clip': 29.412906646728516, 'timing/data': 0.0009063939796760678, 'timing/model': 1.9390678119380027, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 0.6}) +[2026-06-23 12:56:31,400][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.04684415087103844, 'train/grad_norm_pre_clip': 20.63370132446289, 'timing/data': 0.0006818310357630253, 'timing/model': 1.924131208914332, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 0.6}) +[2026-06-23 12:56:35,397][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.10945592820644379, 'train/grad_norm_pre_clip': 46.88521957397461, 'timing/data': 0.0006169449770823121, 'timing/model': 2.1149392910301685, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 0.6}) +[2026-06-23 12:56:39,220][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.14892812073230743, 'train/grad_norm_pre_clip': 43.34650802612305, 'timing/data': 0.0009972129482775927, 'timing/model': 1.9385201450204477, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 0.6}) +[2026-06-23 12:56:43,021][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.044738247990608215, 'train/grad_norm_pre_clip': 26.896682739257812, 'timing/data': 0.0007037789328023791, 'timing/model': 1.9197926339693367, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.6}) +[2026-06-23 12:56:46,816][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.23786599934101105, 'train/grad_norm_pre_clip': 51.41133499145508, 'timing/data': 0.0005992850055918097, 'timing/model': 1.928226210991852, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 0.6}) +[2026-06-23 12:56:50,617][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.09105076640844345, 'train/grad_norm_pre_clip': 48.49824523925781, 'timing/data': 0.0005796239711344242, 'timing/model': 1.9317549160914496, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 0.6}) +[2026-06-23 12:56:54,439][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.1634214073419571, 'train/grad_norm_pre_clip': 39.59505844116211, 'timing/data': 0.0008279919857159257, 'timing/model': 1.9407392790308222, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 0.6}) +[2026-06-23 12:56:58,244][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.08110110461711884, 'train/grad_norm_pre_clip': 37.906795501708984, 'timing/data': 0.0006070220842957497, 'timing/model': 1.921583256102167, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 0.61}) +[2026-06-23 12:57:02,042][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.1676488071680069, 'train/grad_norm_pre_clip': 37.00571823120117, 'timing/data': 0.0006055580452084541, 'timing/model': 1.931413834914565, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 0.61}) +[2026-06-23 12:57:05,886][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.17004050314426422, 'train/grad_norm_pre_clip': 24.437196731567383, 'timing/data': 0.0005922189448028803, 'timing/model': 1.9376143539557233, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 0.61}) +[2026-06-23 12:57:09,704][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.16353122889995575, 'train/grad_norm_pre_clip': 41.68339538574219, 'timing/data': 0.0005877240328118205, 'timing/model': 1.9375725809950382, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 0.61}) +[2026-06-23 12:57:13,509][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.11794289946556091, 'train/grad_norm_pre_clip': 45.77070617675781, 'timing/data': 0.0005758589832112193, 'timing/model': 1.9279226260259748, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 0.61}) +[2026-06-23 12:57:17,311][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.09077030420303345, 'train/grad_norm_pre_clip': 37.05370330810547, 'timing/data': 0.0005977930268272758, 'timing/model': 1.9328908909810707, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 0.61}) +[2026-06-23 12:57:21,122][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.11168687045574188, 'train/grad_norm_pre_clip': 32.11769104003906, 'timing/data': 0.000774347921833396, 'timing/model': 1.9319212780101225, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.61}) +[2026-06-23 12:57:24,930][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.24468252062797546, 'train/grad_norm_pre_clip': 35.93449783325195, 'timing/data': 0.0005830640438944101, 'timing/model': 1.9269586449954659, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 0.61}) +[2026-06-23 12:57:28,752][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.13408461213111877, 'train/grad_norm_pre_clip': 36.245018005371094, 'timing/data': 0.0005963100120425224, 'timing/model': 1.9426988379564136, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 0.61}) +[2026-06-23 12:57:32,566][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.08752674609422684, 'train/grad_norm_pre_clip': 51.32404708862305, 'timing/data': 0.0007474289741367102, 'timing/model': 1.9286253500031307, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 0.61}) +[2026-06-23 12:57:36,378][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.05955853685736656, 'train/grad_norm_pre_clip': 20.061357498168945, 'timing/data': 0.0006352940108627081, 'timing/model': 1.9352388490224257, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 0.61}) +[2026-06-23 12:57:40,209][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.05957111343741417, 'train/grad_norm_pre_clip': 25.492591857910156, 'timing/data': 0.0005881050601601601, 'timing/model': 1.9437734059756622, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 0.61}) +[2026-06-23 12:57:44,024][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.10610045492649078, 'train/grad_norm_pre_clip': 30.00619125366211, 'timing/data': 0.0008085370063781738, 'timing/model': 1.937562649953179, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 0.61}) +[2026-06-23 12:57:47,845][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.06029054895043373, 'train/grad_norm_pre_clip': 11.830363273620605, 'timing/data': 0.0006062360480427742, 'timing/model': 1.9323149789124727, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 0.61}) +[2026-06-23 12:57:51,671][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.18206283450126648, 'train/grad_norm_pre_clip': 57.23359680175781, 'timing/data': 0.0006025739712640643, 'timing/model': 1.9398713010596111, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 0.61}) +[2026-06-23 12:57:55,512][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.12958647310733795, 'train/grad_norm_pre_clip': 40.297664642333984, 'timing/data': 0.0005906590959057212, 'timing/model': 1.9583263639360666, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 0.61}) +[2026-06-23 12:57:59,310][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.1540958732366562, 'train/grad_norm_pre_clip': 38.21745300292969, 'timing/data': 0.0005933670327067375, 'timing/model': 1.924345211009495, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.61}) +[2026-06-23 12:58:03,118][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.11651818454265594, 'train/grad_norm_pre_clip': 19.753124237060547, 'timing/data': 0.0006117899902164936, 'timing/model': 1.933936545974575, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 0.61}) +[2026-06-23 12:58:06,939][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.13998042047023773, 'train/grad_norm_pre_clip': 53.21273422241211, 'timing/data': 0.0005914520006626844, 'timing/model': 1.9452987969852984, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 0.61}) +[2026-06-23 12:58:10,757][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.13758666813373566, 'train/grad_norm_pre_clip': 81.46435546875, 'timing/data': 0.000900162965990603, 'timing/model': 1.9346453220350668, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 0.61}) +[2026-06-23 12:58:14,748][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.14284050464630127, 'train/grad_norm_pre_clip': 105.28543090820312, 'timing/data': 0.0005954630905762315, 'timing/model': 2.1124956210842356, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 0.61}) +[2026-06-23 12:58:18,579][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.16353072226047516, 'train/grad_norm_pre_clip': 67.78572845458984, 'timing/data': 0.00060315802693367, 'timing/model': 1.9463067830074579, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 0.61}) +[2026-06-23 12:58:22,442][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.05355561152100563, 'train/grad_norm_pre_clip': 36.442413330078125, 'timing/data': 0.0006016590632498264, 'timing/model': 1.9919161500874907, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 0.61}) +[2026-06-23 12:58:26,254][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.11787256598472595, 'train/grad_norm_pre_clip': 62.41411590576172, 'timing/data': 0.0005863549886271358, 'timing/model': 1.9323128761025146, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 0.61}) +[2026-06-23 12:58:30,054][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.18246763944625854, 'train/grad_norm_pre_clip': 62.99065399169922, 'timing/data': 0.0006079640006646514, 'timing/model': 1.9264118960127234, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 0.61}) +[2026-06-23 12:58:33,878][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.15996518731117249, 'train/grad_norm_pre_clip': 86.00276947021484, 'timing/data': 0.000823729089461267, 'timing/model': 1.9431528369896114, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 0.61}) +[2026-06-23 12:58:37,688][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.43308213353157043, 'train/grad_norm_pre_clip': 96.00040435791016, 'timing/data': 0.0006110719405114651, 'timing/model': 1.93564035394229, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.62}) +[2026-06-23 12:58:41,509][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.12145133316516876, 'train/grad_norm_pre_clip': 54.485111236572266, 'timing/data': 0.000589850009419024, 'timing/model': 1.9409741460112855, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 0.62}) +[2026-06-23 12:58:45,329][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.17010749876499176, 'train/grad_norm_pre_clip': 33.49099349975586, 'timing/data': 0.000616266974247992, 'timing/model': 1.9356981730088592, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 0.62}) +[2026-06-23 12:58:49,136][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.06583350896835327, 'train/grad_norm_pre_clip': 26.319683074951172, 'timing/data': 0.0006018000422045588, 'timing/model': 1.9412384330062196, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 0.62}) +[2026-06-23 12:58:52,937][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.120016910135746, 'train/grad_norm_pre_clip': 52.56660461425781, 'timing/data': 0.0009325499413534999, 'timing/model': 1.9272193850483745, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 0.62}) +[2026-06-23 12:58:56,743][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.49249929189682007, 'train/grad_norm_pre_clip': 152.8016357421875, 'timing/data': 0.0005914309294894338, 'timing/model': 1.9258831309853122, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 0.62}) +[2026-06-23 12:59:00,553][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.28944161534309387, 'train/grad_norm_pre_clip': 94.14320373535156, 'timing/data': 0.0006702429382130504, 'timing/model': 1.9277868589852005, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 0.62}) +[2026-06-23 12:59:04,368][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.12665030360221863, 'train/grad_norm_pre_clip': 45.39144515991211, 'timing/data': 0.0005842479877173901, 'timing/model': 1.9332868120400235, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 0.62}) +[2026-06-23 12:59:08,171][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.24819767475128174, 'train/grad_norm_pre_clip': 98.47566986083984, 'timing/data': 0.0007513690507039428, 'timing/model': 1.926018501049839, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 0.62}) +[2026-06-23 12:59:11,994][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.0504862442612648, 'train/grad_norm_pre_clip': 16.54117774963379, 'timing/data': 0.0007344119949266315, 'timing/model': 1.9426691200351343, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 0.62}) +[2026-06-23 12:59:15,810][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1600 +[2026-06-23 13:01:24,226][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1600 +[2026-06-23 13:01:24,227][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1600 +[2026-06-23 13:02:45,872][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1600 +[2026-06-23 13:02:45,873][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.08473186194896698, 'train/grad_norm_pre_clip': 24.115432739257812, 'eval/action_loss/seconds': 128.41331375192385, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.08401693403720856, 'eval/latency_2/loss': 0.08401693403720856, 'eval/flappy/loss': 0.08401693403720856, 'eval/flappy/latency_2/loss': 0.08401693403720856, 'eval/flap_precision': 0.8549883990719258, 'eval/flap_recall': 0.9946018893387314, 'eval/flap_f1': 0.9195258889582034, 'eval/flap_rate_model': 0.23944444444444443, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9195258889582034, 'eval/latency_2/flap_precision': 0.8549883990719258, 'eval/latency_2/flap_recall': 0.9946018893387314, 'eval/latency_2/flap_f1': 0.9195258889582034, 'eval/latency_2/flap_rate_model': 0.23944444444444443, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9195258889582034, 'eval/action_classification/data_wait_seconds': 1.3879108428955078, 'eval/action_classification/predict_seconds': 80.04102325439453, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.00061495800036937, 'timing/model': 1.9449914109427482, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.62}) +[2026-06-23 13:02:49,652][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.0705423653125763, 'train/grad_norm_pre_clip': 38.95648956298828, 'timing/data': 0.0006090999813750386, 'timing/model': 1.9130270929308608, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 0.62}) +[2026-06-23 13:02:53,435][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.14272867143154144, 'train/grad_norm_pre_clip': 54.448341369628906, 'timing/data': 0.0007946069817990065, 'timing/model': 1.9150396869517863, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 0.62}) +[2026-06-23 13:02:57,226][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.156078040599823, 'train/grad_norm_pre_clip': 72.57659149169922, 'timing/data': 0.000661561032757163, 'timing/model': 1.9356511260848492, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 0.62}) +[2026-06-23 13:03:01,023][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.03539241477847099, 'train/grad_norm_pre_clip': 23.887033462524414, 'timing/data': 0.0005960729904472828, 'timing/model': 1.920289798057638, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 0.62}) +[2026-06-23 13:03:04,814][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.08176529407501221, 'train/grad_norm_pre_clip': 31.234357833862305, 'timing/data': 0.0006011800142005086, 'timing/model': 1.9236495409859344, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.62}) +[2026-06-23 13:03:08,785][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.07026605308055878, 'train/grad_norm_pre_clip': 17.375627517700195, 'timing/data': 0.0005945920711383224, 'timing/model': 1.9111943659372628, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 0.62}) +[2026-06-23 13:03:12,578][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.12872739136219025, 'train/grad_norm_pre_clip': 29.37800407409668, 'timing/data': 0.0008300709305331111, 'timing/model': 1.9167349729686975, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 0.62}) +[2026-06-23 13:03:16,383][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.10165493190288544, 'train/grad_norm_pre_clip': 39.72136306762695, 'timing/data': 0.0005833900067955256, 'timing/model': 1.928790649981238, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 0.62}) +[2026-06-23 13:03:20,176][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.04512547329068184, 'train/grad_norm_pre_clip': 20.288803100585938, 'timing/data': 0.0006018960848450661, 'timing/model': 1.9299413949484006, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 0.62}) +[2026-06-23 13:03:23,984][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.1616131067276001, 'train/grad_norm_pre_clip': 25.795955657958984, 'timing/data': 0.000605809036642313, 'timing/model': 1.9385387329384685, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.62}) +[2026-06-23 13:03:27,781][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.13384303450584412, 'train/grad_norm_pre_clip': 23.847171783447266, 'timing/data': 0.0006017650011926889, 'timing/model': 1.9351576389744878, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 0.62}) +[2026-06-23 13:03:31,591][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.08160771429538727, 'train/grad_norm_pre_clip': 19.69764518737793, 'timing/data': 0.0006231949664652348, 'timing/model': 1.9269400369375944, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 0.62}) +[2026-06-23 13:03:35,396][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.15201285481452942, 'train/grad_norm_pre_clip': 41.51205825805664, 'timing/data': 0.0006604549707844853, 'timing/model': 1.9343512069899589, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 0.62}) +[2026-06-23 13:03:39,255][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.15011058747768402, 'train/grad_norm_pre_clip': 44.94397735595703, 'timing/data': 0.0006160020129755139, 'timing/model': 1.9175585340708494, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 0.62}) +[2026-06-23 13:03:43,244][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.025047823786735535, 'train/grad_norm_pre_clip': 20.26580810546875, 'timing/data': 0.000607109977863729, 'timing/model': 1.9319018420064822, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 0.63}) +[2026-06-23 13:03:47,037][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.13020513951778412, 'train/grad_norm_pre_clip': 28.511760711669922, 'timing/data': 0.0006014599930495024, 'timing/model': 1.9264202399645, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 0.63}) +[2026-06-23 13:03:50,839][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.14239385724067688, 'train/grad_norm_pre_clip': 55.050716400146484, 'timing/data': 0.0005925459554418921, 'timing/model': 1.9369591180002317, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 0.63}) +[2026-06-23 13:03:54,626][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.05714373290538788, 'train/grad_norm_pre_clip': 27.309064865112305, 'timing/data': 0.0009236119221895933, 'timing/model': 1.9215295859612525, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 0.63}) +[2026-06-23 13:03:58,413][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.12090523540973663, 'train/grad_norm_pre_clip': 28.66310691833496, 'timing/data': 0.0006046009948477149, 'timing/model': 1.9257312680128962, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.63}) +[2026-06-23 13:04:02,232][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.05808011442422867, 'train/grad_norm_pre_clip': 24.30757713317871, 'timing/data': 0.0006182820070534945, 'timing/model': 1.9342979680513963, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.63}) +[2026-06-23 13:04:06,027][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.10681208968162537, 'train/grad_norm_pre_clip': 35.756561279296875, 'timing/data': 0.000576928025111556, 'timing/model': 1.9275624949950725, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 0.63}) +[2026-06-23 13:04:09,839][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.0330008827149868, 'train/grad_norm_pre_clip': 42.853843688964844, 'timing/data': 0.0006177281029522419, 'timing/model': 1.9302044350188226, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 0.63}) +[2026-06-23 13:04:13,820][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.06589656323194504, 'train/grad_norm_pre_clip': 34.62746810913086, 'timing/data': 0.000618139049038291, 'timing/model': 2.1088082829955965, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 0.63}) +[2026-06-23 13:04:17,621][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.08132018893957138, 'train/grad_norm_pre_clip': 20.32780647277832, 'timing/data': 0.0006012240191921592, 'timing/model': 1.9335589009569958, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 0.63}) +[2026-06-23 13:04:21,415][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.10485651344060898, 'train/grad_norm_pre_clip': 38.146568298339844, 'timing/data': 0.0008290390251204371, 'timing/model': 1.9222091890405864, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 0.63}) +[2026-06-23 13:04:25,227][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.13098682463169098, 'train/grad_norm_pre_clip': 65.23692321777344, 'timing/data': 0.0008144860621541739, 'timing/model': 1.9366877620341256, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 0.63}) +[2026-06-23 13:04:29,038][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.08712591230869293, 'train/grad_norm_pre_clip': 45.3283805847168, 'timing/data': 0.000666221952997148, 'timing/model': 1.9389996329555288, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 0.63}) +[2026-06-23 13:04:32,835][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.13422253727912903, 'train/grad_norm_pre_clip': 50.20981979370117, 'timing/data': 0.000653902068734169, 'timing/model': 1.9258884399896488, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 0.63}) +[2026-06-23 13:04:36,642][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.09501614421606064, 'train/grad_norm_pre_clip': 22.708072662353516, 'timing/data': 0.000597839942201972, 'timing/model': 1.93673817103263, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 0.63}) +[2026-06-23 13:04:40,473][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.011491455137729645, 'train/grad_norm_pre_clip': 45.60258865356445, 'timing/data': 0.0005870489403605461, 'timing/model': 1.947783425101079, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.63}) +[2026-06-23 13:04:44,286][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.11945383250713348, 'train/grad_norm_pre_clip': 51.47404861450195, 'timing/data': 0.0005991740617901087, 'timing/model': 1.9382401600014418, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 0.63}) +[2026-06-23 13:04:48,098][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.0942336767911911, 'train/grad_norm_pre_clip': 36.0737190246582, 'timing/data': 0.0006168410181999207, 'timing/model': 1.9288684129714966, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 0.63}) +[2026-06-23 13:04:51,906][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.1171591505408287, 'train/grad_norm_pre_clip': 37.09773635864258, 'timing/data': 0.0005974959349259734, 'timing/model': 1.9344764599809423, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 0.63}) +[2026-06-23 13:04:55,716][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.1036662608385086, 'train/grad_norm_pre_clip': 49.52410125732422, 'timing/data': 0.0007836659206077456, 'timing/model': 1.9348307939944789, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 0.63}) +[2026-06-23 13:04:59,519][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.09118854999542236, 'train/grad_norm_pre_clip': 32.5198974609375, 'timing/data': 0.0007778219878673553, 'timing/model': 1.9387904909672216, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 0.63}) +[2026-06-23 13:05:03,338][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.101416677236557, 'train/grad_norm_pre_clip': 39.40758514404297, 'timing/data': 0.0008143209852278233, 'timing/model': 1.9387228009290993, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 0.63}) +[2026-06-23 13:05:07,137][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.1287694126367569, 'train/grad_norm_pre_clip': 48.34630584716797, 'timing/data': 0.0005933979991823435, 'timing/model': 1.9209266430698335, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 0.63}) +[2026-06-23 13:05:10,965][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.07597935199737549, 'train/grad_norm_pre_clip': 28.723737716674805, 'timing/data': 0.0009177620522677898, 'timing/model': 1.9463078909320757, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 0.63}) +[2026-06-23 13:05:14,784][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.0663706362247467, 'train/grad_norm_pre_clip': 60.51726150512695, 'timing/data': 0.0006075940327718854, 'timing/model': 1.9320911020040512, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 0.63}) +[2026-06-23 13:05:18,594][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.14067161083221436, 'train/grad_norm_pre_clip': 54.874977111816406, 'timing/data': 0.0005757149774581194, 'timing/model': 1.9260146809974685, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.63}) +[2026-06-23 13:05:22,389][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.0460280105471611, 'train/grad_norm_pre_clip': 35.017940521240234, 'timing/data': 0.0005771199939772487, 'timing/model': 1.927265180973336, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 0.64}) +[2026-06-23 13:05:26,241][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.04273950308561325, 'train/grad_norm_pre_clip': 35.73536682128906, 'timing/data': 0.0006412829970940948, 'timing/model': 1.9388225850416347, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 0.64}) +[2026-06-23 13:05:30,058][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.09266316145658493, 'train/grad_norm_pre_clip': 22.22324562072754, 'timing/data': 0.0005948200123384595, 'timing/model': 1.9379944669781253, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 0.64}) +[2026-06-23 13:05:33,876][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.024208303540945053, 'train/grad_norm_pre_clip': 30.329111099243164, 'timing/data': 0.0006644929526373744, 'timing/model': 1.9341509079094976, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 0.64}) +[2026-06-23 13:05:37,675][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.0600431002676487, 'train/grad_norm_pre_clip': 29.09547996520996, 'timing/data': 0.0005893829511478543, 'timing/model': 1.9296441429760307, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 0.64}) +[2026-06-23 13:05:41,487][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.08121328055858612, 'train/grad_norm_pre_clip': 25.56414222717285, 'timing/data': 0.000825054943561554, 'timing/model': 1.9392110090702772, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 0.64}) +[2026-06-23 13:05:45,279][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.14644131064414978, 'train/grad_norm_pre_clip': 76.87047576904297, 'timing/data': 0.000652997987344861, 'timing/model': 1.9285418629879132, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 0.64}) +[2026-06-23 13:05:49,080][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.10368013381958008, 'train/grad_norm_pre_clip': 60.40062713623047, 'timing/data': 0.000686698011122644, 'timing/model': 1.9335568809183314, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 0.64}) +[2026-06-23 13:05:53,093][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.09764693677425385, 'train/grad_norm_pre_clip': 32.53977584838867, 'timing/data': 0.000853804056532681, 'timing/model': 1.9373322740430012, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 0.64}) +[2026-06-23 13:05:56,898][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.15284772217273712, 'train/grad_norm_pre_clip': 23.586240768432617, 'timing/data': 0.0005888930754736066, 'timing/model': 1.931252754991874, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.64}) +[2026-06-23 13:06:00,768][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.08248379826545715, 'train/grad_norm_pre_clip': 38.63808059692383, 'timing/data': 0.0007471860153600574, 'timing/model': 1.940421170904301, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 0.64}) +[2026-06-23 13:06:04,580][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.08636356145143509, 'train/grad_norm_pre_clip': 31.567014694213867, 'timing/data': 0.0005924629513174295, 'timing/model': 1.940535444067791, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 0.64}) +[2026-06-23 13:06:08,389][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.18806631863117218, 'train/grad_norm_pre_clip': 42.34440612792969, 'timing/data': 0.0006206760881468654, 'timing/model': 1.9340047900332138, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 0.64}) +[2026-06-23 13:06:12,203][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.15887576341629028, 'train/grad_norm_pre_clip': 23.735942840576172, 'timing/data': 0.0005981170106679201, 'timing/model': 1.9349152919603512, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 0.64}) +[2026-06-23 13:06:16,010][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.06989961117506027, 'train/grad_norm_pre_clip': 23.287538528442383, 'timing/data': 0.000586590962484479, 'timing/model': 1.9337190539808944, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 0.64}) +[2026-06-23 13:06:19,841][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.05919820815324783, 'train/grad_norm_pre_clip': 30.917396545410156, 'timing/data': 0.00068635493516922, 'timing/model': 1.9297981190029532, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 0.64}) +[2026-06-23 13:06:23,639][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.11546558886766434, 'train/grad_norm_pre_clip': 50.3442268371582, 'timing/data': 0.0006975990254431963, 'timing/model': 1.9292940760497004, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 0.64}) +[2026-06-23 13:06:27,628][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.08435191214084625, 'train/grad_norm_pre_clip': 42.40336990356445, 'timing/data': 0.0007141809910535812, 'timing/model': 1.9385475990129635, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 0.64}) +[2026-06-23 13:06:31,428][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.06634866446256638, 'train/grad_norm_pre_clip': 29.00140380859375, 'timing/data': 0.0008987130131572485, 'timing/model': 1.9306303460616618, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 0.64}) +[2026-06-23 13:06:35,247][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.10079000145196915, 'train/grad_norm_pre_clip': 20.596729278564453, 'timing/data': 0.0006002730224281549, 'timing/model': 1.9374380479566753, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.64}) +[2026-06-23 13:06:39,065][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.062095608562231064, 'train/grad_norm_pre_clip': 21.66257095336914, 'timing/data': 0.0006058579310774803, 'timing/model': 1.93585109396372, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 0.64}) +[2026-06-23 13:06:42,896][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.07086845487356186, 'train/grad_norm_pre_clip': 28.76157569885254, 'timing/data': 0.0007605000864714384, 'timing/model': 1.9356044699670747, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 0.64}) +[2026-06-23 13:06:46,699][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.08238688111305237, 'train/grad_norm_pre_clip': 30.930343627929688, 'timing/data': 0.0007820680039003491, 'timing/model': 1.92470933299046, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 0.64}) +[2026-06-23 13:06:50,505][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.156614288687706, 'train/grad_norm_pre_clip': 33.18244171142578, 'timing/data': 0.0005918930983170867, 'timing/model': 1.933335806010291, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 0.64}) +[2026-06-23 13:06:54,309][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.07154218852519989, 'train/grad_norm_pre_clip': 18.303035736083984, 'timing/data': 0.000664729974232614, 'timing/model': 1.9274065940408036, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 0.64}) +[2026-06-23 13:06:58,295][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.11586949974298477, 'train/grad_norm_pre_clip': 34.37083435058594, 'timing/data': 0.0006070379167795181, 'timing/model': 2.101140560931526, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 0.64}) +[2026-06-23 13:07:02,105][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.07008466869592667, 'train/grad_norm_pre_clip': 17.474315643310547, 'timing/data': 0.0007320709992200136, 'timing/model': 1.928871845942922, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 0.65}) +[2026-06-23 13:07:05,914][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.1092415302991867, 'train/grad_norm_pre_clip': 32.38766860961914, 'timing/data': 0.00060748599935323, 'timing/model': 1.9349903019610792, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.65}) +[2026-06-23 13:07:09,722][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.04139166325330734, 'train/grad_norm_pre_clip': 26.363542556762695, 'timing/data': 0.0009067479986697435, 'timing/model': 1.9261006889864802, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 0.65}) +[2026-06-23 13:07:13,560][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.12169016152620316, 'train/grad_norm_pre_clip': 39.08332443237305, 'timing/data': 0.0005923749413341284, 'timing/model': 1.9548382390057668, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.65}) +[2026-06-23 13:07:17,380][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.19199158251285553, 'train/grad_norm_pre_clip': 38.286563873291016, 'timing/data': 0.0008080049883574247, 'timing/model': 1.9426017380319536, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 0.65}) +[2026-06-23 13:07:21,214][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.09677479416131973, 'train/grad_norm_pre_clip': 30.855201721191406, 'timing/data': 0.0007917069597169757, 'timing/model': 1.9445952009409666, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 0.65}) +[2026-06-23 13:07:25,022][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.019482987001538277, 'train/grad_norm_pre_clip': 21.884014129638672, 'timing/data': 0.0008132769726216793, 'timing/model': 1.93505217297934, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 0.65}) +[2026-06-23 13:07:28,842][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.04353144392371178, 'train/grad_norm_pre_clip': 34.09966278076172, 'timing/data': 0.0006120260804891586, 'timing/model': 1.9389737700112164, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 0.65}) +[2026-06-23 13:07:32,662][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.13082994520664215, 'train/grad_norm_pre_clip': 31.93878746032715, 'timing/data': 0.000600005965679884, 'timing/model': 1.9436992279952392, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.65}) +[2026-06-23 13:07:36,489][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.07921308279037476, 'train/grad_norm_pre_clip': 39.851200103759766, 'timing/data': 0.000706432037986815, 'timing/model': 1.941299261059612, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 0.65}) +[2026-06-23 13:07:40,313][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.08349213004112244, 'train/grad_norm_pre_clip': 36.16596221923828, 'timing/data': 0.000595086021348834, 'timing/model': 1.9421193309826776, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 0.65}) +[2026-06-23 13:07:44,126][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.07166081666946411, 'train/grad_norm_pre_clip': 29.705060958862305, 'timing/data': 0.0005845240084454417, 'timing/model': 1.9375146459788084, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 0.65}) +[2026-06-23 13:07:47,935][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.11315321177244186, 'train/grad_norm_pre_clip': 33.153564453125, 'timing/data': 0.0008780889911577106, 'timing/model': 1.9271509109530598, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 0.65}) +[2026-06-23 13:07:51,743][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.08215750008821487, 'train/grad_norm_pre_clip': 48.70869064331055, 'timing/data': 0.000602921936661005, 'timing/model': 1.9361063810065389, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.65}) +[2026-06-23 13:07:55,553][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.12146152555942535, 'train/grad_norm_pre_clip': 38.247615814208984, 'timing/data': 0.0006161719793453813, 'timing/model': 1.937948351027444, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 0.65}) +[2026-06-23 13:07:59,359][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.051484450697898865, 'train/grad_norm_pre_clip': 23.66693687438965, 'timing/data': 0.0005996819818392396, 'timing/model': 1.9303389629349113, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 0.65}) +[2026-06-23 13:08:03,173][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.11299926787614822, 'train/grad_norm_pre_clip': 38.85588073730469, 'timing/data': 0.0007315690163522959, 'timing/model': 1.930461446987465, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 0.65}) +[2026-06-23 13:08:07,168][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.11378392577171326, 'train/grad_norm_pre_clip': 35.21094512939453, 'timing/data': 0.0006376398960128427, 'timing/model': 1.9320777399698272, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 0.65}) +[2026-06-23 13:08:10,983][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.03158089146018028, 'train/grad_norm_pre_clip': 15.530374526977539, 'timing/data': 0.000655350973829627, 'timing/model': 1.9343690749956295, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 0.65}) +[2026-06-23 13:08:14,790][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.08312004059553146, 'train/grad_norm_pre_clip': 33.085968017578125, 'timing/data': 0.0006286720745265484, 'timing/model': 1.9288632390089333, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 0.65}) +[2026-06-23 13:08:18,600][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.07253682613372803, 'train/grad_norm_pre_clip': 20.401926040649414, 'timing/data': 0.0006996870506554842, 'timing/model': 1.9300022679381073, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 0.65}) +[2026-06-23 13:08:22,396][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.11182042211294174, 'train/grad_norm_pre_clip': 29.17601203918457, 'timing/data': 0.0006413779919967055, 'timing/model': 1.9288682360202074, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 0.65}) +[2026-06-23 13:08:26,205][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.10729946941137314, 'train/grad_norm_pre_clip': 25.245866775512695, 'timing/data': 0.0006269959267228842, 'timing/model': 1.9306477779755369, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 0.65}) +[2026-06-23 13:08:30,019][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.06099438667297363, 'train/grad_norm_pre_clip': 36.15043640136719, 'timing/data': 0.0007337969727814198, 'timing/model': 1.9362163869664073, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.65}) +[2026-06-23 13:08:33,822][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.09289544820785522, 'train/grad_norm_pre_clip': 25.078197479248047, 'timing/data': 0.0005939330440014601, 'timing/model': 1.933983130962588, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 0.65}) +[2026-06-23 13:08:37,818][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.12101712822914124, 'train/grad_norm_pre_clip': 36.92949676513672, 'timing/data': 0.0005959869595244527, 'timing/model': 2.116815368994139, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 0.65}) +[2026-06-23 13:08:41,625][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.10109667479991913, 'train/grad_norm_pre_clip': 40.972991943359375, 'timing/data': 0.0006426629843190312, 'timing/model': 1.9277022699825466, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 0.66}) +[2026-06-23 13:08:45,442][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.11286415159702301, 'train/grad_norm_pre_clip': 43.82411575317383, 'timing/data': 0.0005894539644941688, 'timing/model': 1.9320335930678993, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 0.66}) +[2026-06-23 13:08:49,259][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.07488503307104111, 'train/grad_norm_pre_clip': 26.75343132019043, 'timing/data': 0.0005849209846928716, 'timing/model': 1.9351039190078154, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 0.66}) +[2026-06-23 13:08:53,080][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.053902942687273026, 'train/grad_norm_pre_clip': 20.88918685913086, 'timing/data': 0.0006016240222379565, 'timing/model': 1.9404226059559733, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 0.66}) +[2026-06-23 13:08:56,889][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.10401815176010132, 'train/grad_norm_pre_clip': 32.36107635498047, 'timing/data': 0.0005894049536436796, 'timing/model': 1.9329061349853873, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 0.66}) +[2026-06-23 13:09:00,701][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.014925091527402401, 'train/grad_norm_pre_clip': 21.5107364654541, 'timing/data': 0.0006908129435032606, 'timing/model': 1.940367801929824, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 0.66}) +[2026-06-23 13:09:04,516][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.05921587720513344, 'train/grad_norm_pre_clip': 25.385204315185547, 'timing/data': 0.0006003319285809994, 'timing/model': 1.939945358899422, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 0.66}) +[2026-06-23 13:09:08,323][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1700 +[2026-06-23 13:11:17,761][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1700 +[2026-06-23 13:11:17,763][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1700 +[2026-06-23 13:12:39,949][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1700 +[2026-06-23 13:12:39,950][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.08552789688110352, 'train/grad_norm_pre_clip': 43.92593002319336, 'eval/action_loss/seconds': 129.43495137803257, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.05031877011060715, 'eval/latency_2/loss': 0.05031877011060715, 'eval/flappy/loss': 0.05031877011060715, 'eval/flappy/latency_2/loss': 0.05031877011060715, 'eval/flap_precision': 0.944813829787234, 'eval/flap_recall': 0.9588394062078273, 'eval/flap_f1': 0.9517749497655726, 'eval/flap_rate_model': 0.2088888888888889, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9517749497655726, 'eval/latency_2/flap_precision': 0.944813829787234, 'eval/latency_2/flap_recall': 0.9588394062078273, 'eval/latency_2/flap_f1': 0.9517749497655726, 'eval/latency_2/flap_rate_model': 0.2088888888888889, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9517749497655726, 'eval/action_classification/data_wait_seconds': 1.143803596496582, 'eval/action_classification/predict_seconds': 80.81085205078125, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006222670199349523, 'timing/model': 1.9339888700051233, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.66}) +[2026-06-23 13:12:43,727][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.02890819124877453, 'train/grad_norm_pre_clip': 18.172697067260742, 'timing/data': 0.000591413932852447, 'timing/model': 1.919643153087236, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 0.66}) +[2026-06-23 13:12:47,507][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.10534955561161041, 'train/grad_norm_pre_clip': 36.09832000732422, 'timing/data': 0.0006351309129968286, 'timing/model': 1.9247550830477849, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 0.66}) +[2026-06-23 13:12:51,281][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.0848068967461586, 'train/grad_norm_pre_clip': 19.89167594909668, 'timing/data': 0.0005884339334443212, 'timing/model': 1.9181326929247007, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 0.66}) +[2026-06-23 13:12:55,061][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.07745208591222763, 'train/grad_norm_pre_clip': 19.745264053344727, 'timing/data': 0.0006046229973435402, 'timing/model': 1.9182853000238538, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 0.66}) +[2026-06-23 13:12:58,840][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.09127921611070633, 'train/grad_norm_pre_clip': 26.99903678894043, 'timing/data': 0.0006052410462871194, 'timing/model': 1.9124282409902662, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 0.66}) +[2026-06-23 13:13:02,623][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.08634142577648163, 'train/grad_norm_pre_clip': 25.54454231262207, 'timing/data': 0.0006389510817825794, 'timing/model': 1.9238628660095856, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 0.66}) +[2026-06-23 13:13:06,425][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.07744928449392319, 'train/grad_norm_pre_clip': 21.111892700195312, 'timing/data': 0.0005954449297860265, 'timing/model': 1.9305985879618675, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 0.66}) +[2026-06-23 13:13:10,216][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.03942205384373665, 'train/grad_norm_pre_clip': 12.853432655334473, 'timing/data': 0.0006859180284664035, 'timing/model': 1.9221318160416558, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 0.66}) +[2026-06-23 13:13:13,987][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.15014128386974335, 'train/grad_norm_pre_clip': 42.99113082885742, 'timing/data': 0.0005977670662105083, 'timing/model': 1.9135876829968765, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 0.66}) +[2026-06-23 13:13:17,790][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.07315082848072052, 'train/grad_norm_pre_clip': 68.17828369140625, 'timing/data': 0.000653519993647933, 'timing/model': 1.9254510349128395, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.66}) +[2026-06-23 13:13:21,572][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.02101192995905876, 'train/grad_norm_pre_clip': 32.64162063598633, 'timing/data': 0.0006910170195624232, 'timing/model': 1.9169695279560983, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 0.66}) +[2026-06-23 13:13:25,352][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.06729498505592346, 'train/grad_norm_pre_clip': 20.501602172851562, 'timing/data': 0.0006599780172109604, 'timing/model': 1.921196707058698, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 0.66}) +[2026-06-23 13:13:29,137][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.12442554533481598, 'train/grad_norm_pre_clip': 31.69907569885254, 'timing/data': 0.0006655820179730654, 'timing/model': 1.9258623529458418, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 0.66}) +[2026-06-23 13:13:32,926][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.12098930776119232, 'train/grad_norm_pre_clip': 22.57813262939453, 'timing/data': 0.0006052430253475904, 'timing/model': 1.9221908210311085, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 0.66}) +[2026-06-23 13:13:36,717][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.23447540402412415, 'train/grad_norm_pre_clip': 60.682350158691406, 'timing/data': 0.0007178430678322911, 'timing/model': 1.9180729859508574, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 0.66}) +[2026-06-23 13:13:40,501][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.021720729768276215, 'train/grad_norm_pre_clip': 26.512523651123047, 'timing/data': 0.0007054849993437529, 'timing/model': 1.9169478369876742, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 0.66}) +[2026-06-23 13:13:44,303][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.058136578649282455, 'train/grad_norm_pre_clip': 23.894819259643555, 'timing/data': 0.000649446970783174, 'timing/model': 1.9263831649441272, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.66}) +[2026-06-23 13:13:48,149][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.07719864696264267, 'train/grad_norm_pre_clip': 26.205642700195312, 'timing/data': 0.0006484990008175373, 'timing/model': 1.9760061840061098, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 0.66}) +[2026-06-23 13:13:51,961][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.07808304578065872, 'train/grad_norm_pre_clip': 24.73296356201172, 'timing/data': 0.0006673219613730907, 'timing/model': 1.9375015480909497, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 0.67}) +[2026-06-23 13:13:55,767][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.07683411240577698, 'train/grad_norm_pre_clip': 26.838653564453125, 'timing/data': 0.0005922960117459297, 'timing/model': 1.9262659640517086, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.67}) +[2026-06-23 13:13:59,561][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.01630854420363903, 'train/grad_norm_pre_clip': 9.509688377380371, 'timing/data': 0.0006055430276319385, 'timing/model': 1.9293915759772062, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 0.67}) +[2026-06-23 13:14:03,569][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.08992654085159302, 'train/grad_norm_pre_clip': 55.15109634399414, 'timing/data': 0.0007150289602577686, 'timing/model': 2.1262839749688283, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 0.67}) +[2026-06-23 13:14:07,385][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.20099374651908875, 'train/grad_norm_pre_clip': 25.186119079589844, 'timing/data': 0.0005906489677727222, 'timing/model': 1.9388841229956597, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 0.67}) +[2026-06-23 13:14:11,187][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.08824265003204346, 'train/grad_norm_pre_clip': 18.946090698242188, 'timing/data': 0.0005925670266151428, 'timing/model': 1.9281792619731277, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 0.67}) +[2026-06-23 13:14:14,988][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.11694598197937012, 'train/grad_norm_pre_clip': 31.502241134643555, 'timing/data': 0.0008890570607036352, 'timing/model': 1.9246338229859248, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 0.67}) +[2026-06-23 13:14:18,849][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.12690061330795288, 'train/grad_norm_pre_clip': 38.03972625732422, 'timing/data': 0.0007537769852206111, 'timing/model': 1.9786194149637595, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 0.67}) +[2026-06-23 13:14:22,641][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.05601479485630989, 'train/grad_norm_pre_clip': 31.091514587402344, 'timing/data': 0.00058985804207623, 'timing/model': 1.9294703929917887, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 0.67}) +[2026-06-23 13:14:26,436][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.09405608475208282, 'train/grad_norm_pre_clip': 56.940914154052734, 'timing/data': 0.0005935840308666229, 'timing/model': 1.9200875620590523, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 0.67}) +[2026-06-23 13:14:30,243][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.14693312346935272, 'train/grad_norm_pre_clip': 32.31298065185547, 'timing/data': 0.0007761999731883407, 'timing/model': 1.9303895119810477, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 0.67}) +[2026-06-23 13:14:34,047][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.05243900790810585, 'train/grad_norm_pre_clip': 15.847784996032715, 'timing/data': 0.0006029040087014437, 'timing/model': 1.9340827969135717, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.67}) +[2026-06-23 13:14:37,855][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.06040982902050018, 'train/grad_norm_pre_clip': 28.985883712768555, 'timing/data': 0.0006056500133126974, 'timing/model': 1.9365050139604136, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 0.67}) +[2026-06-23 13:14:41,683][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.14070427417755127, 'train/grad_norm_pre_clip': 51.011051177978516, 'timing/data': 0.000602441024966538, 'timing/model': 1.9461164250969887, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 0.67}) +[2026-06-23 13:14:45,498][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.07819264382123947, 'train/grad_norm_pre_clip': 21.717443466186523, 'timing/data': 0.000634909956716001, 'timing/model': 1.9368850510800257, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 0.67}) +[2026-06-23 13:14:49,311][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.07625321298837662, 'train/grad_norm_pre_clip': 19.29425048828125, 'timing/data': 0.00060387107077986, 'timing/model': 1.9365235719596967, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 0.67}) +[2026-06-23 13:14:53,132][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.07588932663202286, 'train/grad_norm_pre_clip': 42.37955856323242, 'timing/data': 0.0005950260674580932, 'timing/model': 1.9367359320167452, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 0.67}) +[2026-06-23 13:14:56,939][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.07620540261268616, 'train/grad_norm_pre_clip': 37.4805793762207, 'timing/data': 0.0006028350908309221, 'timing/model': 1.9325910730985925, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 0.67}) +[2026-06-23 13:15:00,747][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.04734406620264053, 'train/grad_norm_pre_clip': 14.037880897521973, 'timing/data': 0.0006080960156396031, 'timing/model': 1.9316975369583815, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 0.67}) +[2026-06-23 13:15:04,551][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.034630827605724335, 'train/grad_norm_pre_clip': 19.536211013793945, 'timing/data': 0.0006064031040295959, 'timing/model': 1.9318392920540646, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 0.67}) +[2026-06-23 13:15:08,371][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.060258179903030396, 'train/grad_norm_pre_clip': 43.57435989379883, 'timing/data': 0.0006020970176905394, 'timing/model': 1.9350530370138586, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 0.67}) +[2026-06-23 13:15:12,187][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.11782413721084595, 'train/grad_norm_pre_clip': 67.2783432006836, 'timing/data': 0.0005904550198465586, 'timing/model': 1.936704527004622, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.67}) +[2026-06-23 13:15:15,987][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.16500207781791687, 'train/grad_norm_pre_clip': 64.1735610961914, 'timing/data': 0.0006107130320742726, 'timing/model': 1.9259287239983678, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 0.67}) +[2026-06-23 13:15:19,809][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.02599283680319786, 'train/grad_norm_pre_clip': 21.52617073059082, 'timing/data': 0.0007981010712683201, 'timing/model': 1.937084986944683, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 0.67}) +[2026-06-23 13:15:23,610][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.07624977827072144, 'train/grad_norm_pre_clip': 38.06122589111328, 'timing/data': 0.0006204369710758328, 'timing/model': 1.921362585038878, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 0.67}) +[2026-06-23 13:15:27,419][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.13844461739063263, 'train/grad_norm_pre_clip': 33.63330841064453, 'timing/data': 0.0006079659797251225, 'timing/model': 1.9365540449507535, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 0.67}) +[2026-06-23 13:15:31,224][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.19352979958057404, 'train/grad_norm_pre_clip': 86.06048583984375, 'timing/data': 0.0006748070009052753, 'timing/model': 1.9300621959846467, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.68}) +[2026-06-23 13:15:35,061][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.030064068734645844, 'train/grad_norm_pre_clip': 22.57520866394043, 'timing/data': 0.0007248639594763517, 'timing/model': 1.9465574320638552, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 0.68}) +[2026-06-23 13:15:38,874][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.12525105476379395, 'train/grad_norm_pre_clip': 62.53623580932617, 'timing/data': 0.000720140989869833, 'timing/model': 1.9304212980205193, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 0.68}) +[2026-06-23 13:15:42,866][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.0897669866681099, 'train/grad_norm_pre_clip': 31.3381290435791, 'timing/data': 0.0006739581003785133, 'timing/model': 1.9327456189785153, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 0.68}) +[2026-06-23 13:15:46,686][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.04078628495335579, 'train/grad_norm_pre_clip': 34.765235900878906, 'timing/data': 0.0005828340072184801, 'timing/model': 1.9472394429612905, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 0.68}) +[2026-06-23 13:25:39,431][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1750 +[2026-06-23 13:25:39,432][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.05177344009280205, 'train/grad_norm_pre_clip': 26.163394927978516, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 361.70000000001505, 'rl_games_eval/mid_train/mean_length': 2934.2, 'rl_games_eval/mid_train/std_reward': 106.76950875602273, 'rl_games_eval/mid_train/std_length': 859.2597744570613, 'rl_games_eval/mid_train/macro_mean_reward': 361.70000000001505, 'rl_games_eval/mid_train/macro_mean_length': 2934.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 361.70000000001505, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2934.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 106.76950875602273, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 859.2597744570613, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 361.70000000001505, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006168090039864182, 'timing/model': 1.9232823189813644, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.68}) +[2026-06-23 13:25:43,165][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.0965302437543869, 'train/grad_norm_pre_clip': 25.230894088745117, 'timing/data': 0.0006924020126461983, 'timing/model': 1.8866347779985517, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 0.68}) +[2026-06-23 13:25:46,899][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.04544341191649437, 'train/grad_norm_pre_clip': 24.809677124023438, 'timing/data': 0.0007266260217875242, 'timing/model': 1.8967530120862648, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 0.68}) +[2026-06-23 13:25:50,635][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.048862237483263016, 'train/grad_norm_pre_clip': 27.335824966430664, 'timing/data': 0.0006329769967123866, 'timing/model': 1.9013259130297229, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 0.68}) +[2026-06-23 13:25:54,369][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.1499771773815155, 'train/grad_norm_pre_clip': 26.76237678527832, 'timing/data': 0.0005959970876574516, 'timing/model': 1.8997132930671796, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 0.68}) +[2026-06-23 13:25:58,112][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.1271408349275589, 'train/grad_norm_pre_clip': 24.191919326782227, 'timing/data': 0.0006115519208833575, 'timing/model': 1.8988135210238397, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 0.68}) +[2026-06-23 13:26:01,859][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.09680582582950592, 'train/grad_norm_pre_clip': 37.96350860595703, 'timing/data': 0.0006060819141566753, 'timing/model': 1.9031542990123853, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 0.68}) +[2026-06-23 13:26:05,616][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.09883275628089905, 'train/grad_norm_pre_clip': 20.974903106689453, 'timing/data': 0.0008041459368541837, 'timing/model': 1.9116862629307434, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 0.68}) +[2026-06-23 13:26:09,372][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.05591316148638725, 'train/grad_norm_pre_clip': 28.97649383544922, 'timing/data': 0.0005938620306551456, 'timing/model': 1.905554116005078, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 0.68}) +[2026-06-23 13:26:13,138][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.09958034008741379, 'train/grad_norm_pre_clip': 30.11625099182129, 'timing/data': 0.0006423840532079339, 'timing/model': 1.9133926739450544, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 0.68}) +[2026-06-23 13:26:17,078][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.1467224359512329, 'train/grad_norm_pre_clip': 31.758346557617188, 'timing/data': 0.0006025130860507488, 'timing/model': 1.9134055390022695, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.68}) +[2026-06-23 13:26:20,841][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.12160537391901016, 'train/grad_norm_pre_clip': 40.237857818603516, 'timing/data': 0.0006161450874060392, 'timing/model': 1.9051510959398001, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 0.68}) +[2026-06-23 13:26:24,610][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.16672161221504211, 'train/grad_norm_pre_clip': 37.7139892578125, 'timing/data': 0.0005971889477223158, 'timing/model': 1.9134465210372582, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 0.68}) +[2026-06-23 13:26:28,382][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.14502353966236115, 'train/grad_norm_pre_clip': 25.130107879638672, 'timing/data': 0.0005992020014673471, 'timing/model': 1.9171512050088495, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 0.68}) +[2026-06-23 13:26:32,157][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.1588383913040161, 'train/grad_norm_pre_clip': 47.823848724365234, 'timing/data': 0.0005949939368292689, 'timing/model': 1.915859620901756, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 0.68}) +[2026-06-23 13:26:35,926][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.11564938724040985, 'train/grad_norm_pre_clip': 25.63521957397461, 'timing/data': 0.0006078860023990273, 'timing/model': 1.911031326977536, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 0.68}) +[2026-06-23 13:26:39,699][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.046703748404979706, 'train/grad_norm_pre_clip': 20.0018253326416, 'timing/data': 0.0005956260720267892, 'timing/model': 1.920541015919298, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.68}) +[2026-06-23 13:26:43,469][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.06732357293367386, 'train/grad_norm_pre_clip': 31.970157623291016, 'timing/data': 0.000604828936047852, 'timing/model': 1.9108873800141737, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 0.68}) +[2026-06-23 13:26:47,308][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.06308191269636154, 'train/grad_norm_pre_clip': 19.731040954589844, 'timing/data': 0.0008177270647138357, 'timing/model': 1.9813873230013996, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 0.68}) +[2026-06-23 13:26:51,088][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.11913926899433136, 'train/grad_norm_pre_clip': 77.15352630615234, 'timing/data': 0.000697755953297019, 'timing/model': 1.9115239939419553, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 0.68}) +[2026-06-23 13:26:54,867][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.18361549079418182, 'train/grad_norm_pre_clip': 107.3974380493164, 'timing/data': 0.0007903219666332006, 'timing/model': 1.9156374329468235, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.69}) +[2026-06-23 13:26:58,661][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.1126944050192833, 'train/grad_norm_pre_clip': 44.7992057800293, 'timing/data': 0.0009378480026498437, 'timing/model': 1.9305028830422089, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 0.69}) +[2026-06-23 13:27:02,428][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.13221170008182526, 'train/grad_norm_pre_clip': 35.06929016113281, 'timing/data': 0.0006102160550653934, 'timing/model': 1.909121738979593, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 0.69}) +[2026-06-23 13:27:06,210][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.11658544093370438, 'train/grad_norm_pre_clip': 25.453887939453125, 'timing/data': 0.0006062029860913754, 'timing/model': 1.926378790056333, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 0.69}) +[2026-06-23 13:27:10,001][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.02178293466567993, 'train/grad_norm_pre_clip': 29.884891510009766, 'timing/data': 0.0006452919915318489, 'timing/model': 1.9317745260195807, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 0.69}) +[2026-06-23 13:27:13,790][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.12161778658628464, 'train/grad_norm_pre_clip': 35.9575309753418, 'timing/data': 0.0008436499629169703, 'timing/model': 1.919144223909825, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 0.69}) +[2026-06-23 13:27:17,589][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.07560485601425171, 'train/grad_norm_pre_clip': 32.897071838378906, 'timing/data': 0.0006022399757057428, 'timing/model': 1.925596215063706, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 0.69}) +[2026-06-23 13:27:21,491][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.0984850749373436, 'train/grad_norm_pre_clip': 26.012420654296875, 'timing/data': 0.0006268779980018735, 'timing/model': 2.0354609539499506, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 0.69}) +[2026-06-23 13:27:25,289][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.09722889959812164, 'train/grad_norm_pre_clip': 32.99497604370117, 'timing/data': 0.0009265240514650941, 'timing/model': 1.9263384600635618, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 0.69}) +[2026-06-23 13:27:29,092][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.09923451393842697, 'train/grad_norm_pre_clip': 34.88248825073242, 'timing/data': 0.0006992800626903772, 'timing/model': 1.927472846931778, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 0.69}) +[2026-06-23 13:27:32,884][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.06259187310934067, 'train/grad_norm_pre_clip': 27.355594635009766, 'timing/data': 0.000759118003770709, 'timing/model': 1.9264229879481718, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.69}) +[2026-06-23 13:27:36,688][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.032780397683382034, 'train/grad_norm_pre_clip': 16.730520248413086, 'timing/data': 0.0006226999685168266, 'timing/model': 1.9289890920044854, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 0.69}) +[2026-06-23 13:27:40,472][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.05268697813153267, 'train/grad_norm_pre_clip': 24.089229583740234, 'timing/data': 0.0006594540318474174, 'timing/model': 1.925399656058289, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 0.69}) +[2026-06-23 13:27:44,271][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.09273041784763336, 'train/grad_norm_pre_clip': 27.6127986907959, 'timing/data': 0.0006100189639255404, 'timing/model': 1.92856211704202, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 0.69}) +[2026-06-23 13:27:48,075][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.09937988966703415, 'train/grad_norm_pre_clip': 34.20350646972656, 'timing/data': 0.0006003440357744694, 'timing/model': 1.9349907749565318, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 0.69}) +[2026-06-23 13:27:51,863][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.26238781213760376, 'train/grad_norm_pre_clip': 59.8471565246582, 'timing/data': 0.0006175510352477431, 'timing/model': 1.9166836410295218, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 0.69}) +[2026-06-23 13:27:55,777][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.11584343761205673, 'train/grad_norm_pre_clip': 46.40647506713867, 'timing/data': 0.0006122790509834886, 'timing/model': 1.9318852700525895, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 0.69}) +[2026-06-23 13:27:59,589][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.07306890934705734, 'train/grad_norm_pre_clip': 19.77229118347168, 'timing/data': 0.0006096899742260575, 'timing/model': 1.9392458531074226, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 0.69}) +[2026-06-23 13:28:03,398][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.10089978575706482, 'train/grad_norm_pre_clip': 32.19120788574219, 'timing/data': 0.0006105429492890835, 'timing/model': 1.9261699620401487, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 0.69}) +[2026-06-23 13:28:07,218][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.04402777925133705, 'train/grad_norm_pre_clip': 46.18730926513672, 'timing/data': 0.000599929946474731, 'timing/model': 1.9245693139964715, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 0.69}) +[2026-06-23 13:28:11,030][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.03222036734223366, 'train/grad_norm_pre_clip': 26.982769012451172, 'timing/data': 0.0007741620065644383, 'timing/model': 1.9238816760480404, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.69}) +[2026-06-23 13:28:14,846][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.12438006699085236, 'train/grad_norm_pre_clip': 82.43663787841797, 'timing/data': 0.000740773044526577, 'timing/model': 1.9388379999436438, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 0.69}) +[2026-06-23 13:28:18,642][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.10835397988557816, 'train/grad_norm_pre_clip': 33.64672088623047, 'timing/data': 0.0006799729308113456, 'timing/model': 1.9240387219469994, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 0.69}) +[2026-06-23 13:28:22,445][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.05476024001836777, 'train/grad_norm_pre_clip': 22.589521408081055, 'timing/data': 0.0006169280968606472, 'timing/model': 1.9224779900396243, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 0.69}) +[2026-06-23 13:28:26,306][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.08584081381559372, 'train/grad_norm_pre_clip': 55.863746643066406, 'timing/data': 0.0006468979408964515, 'timing/model': 1.9926130530657247, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 0.69}) +[2026-06-23 13:28:30,110][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.2177429497241974, 'train/grad_norm_pre_clip': 69.93603515625, 'timing/data': 0.0006149780238047242, 'timing/model': 1.9285572440130636, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 0.69}) +[2026-06-23 13:28:33,925][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.05280197411775589, 'train/grad_norm_pre_clip': 36.07208251953125, 'timing/data': 0.0007229159818962216, 'timing/model': 1.9414486710447818, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 0.7}) +[2026-06-23 13:28:37,743][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.09420785307884216, 'train/grad_norm_pre_clip': 56.08668518066406, 'timing/data': 0.0006229809951037169, 'timing/model': 1.929638460977003, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 0.7}) +[2026-06-23 13:28:41,549][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.06918687373399734, 'train/grad_norm_pre_clip': 22.145273208618164, 'timing/data': 0.0006169140106067061, 'timing/model': 1.933176108985208, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 0.7}) +[2026-06-23 13:28:45,338][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.2392716109752655, 'train/grad_norm_pre_clip': 60.44368362426758, 'timing/data': 0.0006112359696999192, 'timing/model': 1.9163868769537657, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 0.7}) +[2026-06-23 13:28:49,136][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1800 +[2026-06-23 13:30:55,534][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1800 +[2026-06-23 13:30:55,535][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1800 +[2026-06-23 13:32:19,654][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1800 +[2026-06-23 13:32:19,655][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.06945774704217911, 'train/grad_norm_pre_clip': 49.607669830322266, 'eval/action_loss/seconds': 126.39495613402687, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06445325165987015, 'eval/latency_2/loss': 0.06445325165987015, 'eval/flappy/loss': 0.06445325165987015, 'eval/flappy/latency_2/loss': 0.06445325165987015, 'eval/flap_precision': 0.9865971705137752, 'eval/flap_recall': 0.8940620782726046, 'eval/flap_f1': 0.9380530973451328, 'eval/flap_rate_model': 0.18652777777777776, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9380530973451329, 'eval/latency_2/flap_precision': 0.9865971705137752, 'eval/latency_2/flap_recall': 0.8940620782726046, 'eval/latency_2/flap_f1': 0.9380530973451328, 'eval/latency_2/flap_rate_model': 0.18652777777777776, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9380530973451329, 'eval/action_classification/data_wait_seconds': 1.6119707822799683, 'eval/action_classification/predict_seconds': 82.31236267089844, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006258710054680705, 'timing/model': 1.9332029590150341, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.7}) +[2026-06-23 13:32:23,424][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.11696960777044296, 'train/grad_norm_pre_clip': 55.628902435302734, 'timing/data': 0.0006104919593781233, 'timing/model': 1.9123421229887754, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 0.7}) +[2026-06-23 13:32:27,202][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.08511646836996078, 'train/grad_norm_pre_clip': 22.338960647583008, 'timing/data': 0.0007963799871504307, 'timing/model': 1.9056801460683346, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 0.7}) +[2026-06-23 13:32:30,976][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.1568634957075119, 'train/grad_norm_pre_clip': 28.68909454345703, 'timing/data': 0.0008438440272584558, 'timing/model': 1.913881087093614, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 0.7}) +[2026-06-23 13:32:34,760][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.22440269589424133, 'train/grad_norm_pre_clip': 57.32663345336914, 'timing/data': 0.00064201804343611, 'timing/model': 1.9163780229864642, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 0.7}) +[2026-06-23 13:32:38,539][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.10256380587816238, 'train/grad_norm_pre_clip': 40.73078536987305, 'timing/data': 0.0006507670041173697, 'timing/model': 1.9183986389543861, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 0.7}) +[2026-06-23 13:32:42,344][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.0697551742196083, 'train/grad_norm_pre_clip': 18.7959041595459, 'timing/data': 0.0006103849736973643, 'timing/model': 1.919037436018698, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 0.7}) +[2026-06-23 13:32:46,165][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.12532596290111542, 'train/grad_norm_pre_clip': 21.25955581665039, 'timing/data': 0.0006087099900469184, 'timing/model': 1.950116960098967, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 0.7}) +[2026-06-23 13:32:49,959][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.03404763713479042, 'train/grad_norm_pre_clip': 32.643341064453125, 'timing/data': 0.0006416329415515065, 'timing/model': 1.9302947981050238, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 0.7}) +[2026-06-23 13:32:53,737][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.10073868185281754, 'train/grad_norm_pre_clip': 27.769874572753906, 'timing/data': 0.0008267930243164301, 'timing/model': 1.9184286970412359, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 0.7}) +[2026-06-23 13:32:57,538][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.10614421218633652, 'train/grad_norm_pre_clip': 20.471296310424805, 'timing/data': 0.0006459949072450399, 'timing/model': 1.923178724013269, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.7}) +[2026-06-23 13:33:01,328][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.10287854820489883, 'train/grad_norm_pre_clip': 26.49665069580078, 'timing/data': 0.0005951530765742064, 'timing/model': 1.9347486299229786, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 0.7}) +[2026-06-23 13:33:05,128][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.061037223786115646, 'train/grad_norm_pre_clip': 15.344841957092285, 'timing/data': 0.000587266986258328, 'timing/model': 1.9309860760113224, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 0.7}) +[2026-06-23 13:33:08,920][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.17907825112342834, 'train/grad_norm_pre_clip': 21.022899627685547, 'timing/data': 0.0008224979974329472, 'timing/model': 1.9212550000520423, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 0.7}) +[2026-06-23 13:33:12,703][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.08041540533304214, 'train/grad_norm_pre_clip': 17.808542251586914, 'timing/data': 0.00061765406280756, 'timing/model': 1.9152364250039682, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 0.7}) +[2026-06-23 13:33:16,501][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.06705674529075623, 'train/grad_norm_pre_clip': 23.60563087463379, 'timing/data': 0.0006131340051069856, 'timing/model': 1.9232603100826964, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.7}) +[2026-06-23 13:33:20,401][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.1107054203748703, 'train/grad_norm_pre_clip': 28.429651260375977, 'timing/data': 0.0008029980817809701, 'timing/model': 2.029428861918859, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 0.7}) +[2026-06-23 13:33:24,185][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.06929823011159897, 'train/grad_norm_pre_clip': 19.6729679107666, 'timing/data': 0.0006236800691112876, 'timing/model': 1.9193692320259288, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 0.7}) +[2026-06-23 13:33:27,987][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.15516217052936554, 'train/grad_norm_pre_clip': 39.98292922973633, 'timing/data': 0.0007448070682585239, 'timing/model': 1.9301390839973465, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 0.7}) +[2026-06-23 13:33:31,777][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.0785500779747963, 'train/grad_norm_pre_clip': 43.59590530395508, 'timing/data': 0.0005975170060992241, 'timing/model': 1.916284259990789, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 0.7}) +[2026-06-23 13:33:35,582][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.26522549986839294, 'train/grad_norm_pre_clip': 34.13017654418945, 'timing/data': 0.0006017920095473528, 'timing/model': 1.9317218740470707, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.7}) +[2026-06-23 13:33:39,376][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.04325852170586586, 'train/grad_norm_pre_clip': 15.048445701599121, 'timing/data': 0.0007687159813940525, 'timing/model': 1.9226223720470443, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 0.7}) +[2026-06-23 13:33:43,175][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.046914853155612946, 'train/grad_norm_pre_clip': 16.44775390625, 'timing/data': 0.0005863490514457226, 'timing/model': 1.9286166669335216, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 0.71}) +[2026-06-23 13:33:46,982][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.06549510359764099, 'train/grad_norm_pre_clip': 26.65143585205078, 'timing/data': 0.000597987906076014, 'timing/model': 1.9385554300388321, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 0.71}) +[2026-06-23 13:33:50,788][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.1123405173420906, 'train/grad_norm_pre_clip': 65.01285552978516, 'timing/data': 0.0006162310019135475, 'timing/model': 1.9339052609866485, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 0.71}) +[2026-06-23 13:33:54,707][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.10467976331710815, 'train/grad_norm_pre_clip': 48.18534851074219, 'timing/data': 0.0009099760791286826, 'timing/model': 1.9254841649672017, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 0.71}) +[2026-06-23 13:33:58,521][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.007944184355437756, 'train/grad_norm_pre_clip': 16.877195358276367, 'timing/data': 0.0006218999624252319, 'timing/model': 1.9295424990123138, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 0.71}) +[2026-06-23 13:34:02,313][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.10261279344558716, 'train/grad_norm_pre_clip': 51.26631164550781, 'timing/data': 0.0006371570052579045, 'timing/model': 1.9203505780315027, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 0.71}) +[2026-06-23 13:34:06,122][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.04375077411532402, 'train/grad_norm_pre_clip': 45.23064041137695, 'timing/data': 0.0006685819244012237, 'timing/model': 1.9331114799715579, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 0.71}) +[2026-06-23 13:34:09,922][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.22673605382442474, 'train/grad_norm_pre_clip': 34.655677795410156, 'timing/data': 0.0006232799496501684, 'timing/model': 1.9270551080117002, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 0.71}) +[2026-06-23 13:34:13,723][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.08353080600500107, 'train/grad_norm_pre_clip': 23.78316879272461, 'timing/data': 0.0006520709721371531, 'timing/model': 1.9269292700337246, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.71}) +[2026-06-23 13:34:17,533][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.08083008974790573, 'train/grad_norm_pre_clip': 31.290691375732422, 'timing/data': 0.0005833629984408617, 'timing/model': 1.9299379020230845, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 0.71}) +[2026-06-23 13:34:21,326][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.03141180798411369, 'train/grad_norm_pre_clip': 19.593421936035156, 'timing/data': 0.0006508249789476395, 'timing/model': 1.9299962359946221, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 0.71}) +[2026-06-23 13:34:25,206][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.11528883874416351, 'train/grad_norm_pre_clip': 23.231351852416992, 'timing/data': 0.0005936919478699565, 'timing/model': 2.014664249960333, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 0.71}) +[2026-06-23 13:34:29,061][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.07147451490163803, 'train/grad_norm_pre_clip': 42.187808990478516, 'timing/data': 0.0006294030463322997, 'timing/model': 1.9285090829944238, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 0.71}) +[2026-06-23 13:34:32,868][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.17185083031654358, 'train/grad_norm_pre_clip': 61.06673812866211, 'timing/data': 0.0005965559976175427, 'timing/model': 1.9259776920080185, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 0.71}) +[2026-06-23 13:34:36,690][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.020370207726955414, 'train/grad_norm_pre_clip': 38.20559310913086, 'timing/data': 0.0006011369405314326, 'timing/model': 1.9354067910462618, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 0.71}) +[2026-06-23 13:34:40,493][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.041451554745435715, 'train/grad_norm_pre_clip': 42.85989761352539, 'timing/data': 0.0006132490234449506, 'timing/model': 1.9313884270377457, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 0.71}) +[2026-06-23 13:34:44,296][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.05339541286230087, 'train/grad_norm_pre_clip': 54.50144958496094, 'timing/data': 0.0007952330633997917, 'timing/model': 1.9241761789890006, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 0.71}) +[2026-06-23 13:34:48,091][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.17005306482315063, 'train/grad_norm_pre_clip': 45.149566650390625, 'timing/data': 0.0006001009605824947, 'timing/model': 1.9264067880576476, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 0.71}) +[2026-06-23 13:34:51,895][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.029715457931160927, 'train/grad_norm_pre_clip': 29.06364631652832, 'timing/data': 0.0006705119740217924, 'timing/model': 1.9307393619092181, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.71}) +[2026-06-23 13:34:55,687][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.1601729691028595, 'train/grad_norm_pre_clip': 76.48220825195312, 'timing/data': 0.0006767869926989079, 'timing/model': 1.9237049110233784, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 0.71}) +[2026-06-23 13:34:59,613][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.22320371866226196, 'train/grad_norm_pre_clip': 45.68918228149414, 'timing/data': 0.0006091960240155458, 'timing/model': 2.049886188004166, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 0.71}) +[2026-06-23 13:35:03,407][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.13236141204833984, 'train/grad_norm_pre_clip': 59.74873352050781, 'timing/data': 0.0006062490865588188, 'timing/model': 1.9217124149436131, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 0.71}) +[2026-06-23 13:35:07,214][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.1612636297941208, 'train/grad_norm_pre_clip': 34.704856872558594, 'timing/data': 0.0006132220150902867, 'timing/model': 1.9213165499968454, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 0.71}) +[2026-06-23 13:35:11,021][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.07871484756469727, 'train/grad_norm_pre_clip': 26.88030242919922, 'timing/data': 0.0005941919516772032, 'timing/model': 1.9279321760404855, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 0.71}) +[2026-06-23 13:35:14,845][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.04322099685668945, 'train/grad_norm_pre_clip': 36.84241485595703, 'timing/data': 0.0005872390465810895, 'timing/model': 1.9437081230571494, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 0.71}) +[2026-06-23 13:35:18,665][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.06079065054655075, 'train/grad_norm_pre_clip': 27.879491806030273, 'timing/data': 0.0005889650201424956, 'timing/model': 1.9421406880719587, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 0.71}) +[2026-06-23 13:35:22,481][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.0657568946480751, 'train/grad_norm_pre_clip': 14.991829872131348, 'timing/data': 0.0006118940655142069, 'timing/model': 1.9302643589908257, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 0.72}) +[2026-06-23 13:35:26,283][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.06380118429660797, 'train/grad_norm_pre_clip': 17.329803466796875, 'timing/data': 0.0005924359429627657, 'timing/model': 1.931620639981702, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 0.72}) +[2026-06-23 13:35:30,131][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.03568989783525467, 'train/grad_norm_pre_clip': 12.917387962341309, 'timing/data': 0.000569180934689939, 'timing/model': 1.9648698390228674, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.72}) +[2026-06-23 13:35:33,927][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.13042117655277252, 'train/grad_norm_pre_clip': 61.09941482543945, 'timing/data': 0.0005962040740996599, 'timing/model': 1.9299746420001611, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 0.72}) +[2026-06-23 13:35:37,742][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.028833648189902306, 'train/grad_norm_pre_clip': 15.89085865020752, 'timing/data': 0.0005963539006188512, 'timing/model': 1.940144023974426, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 0.72}) +[2026-06-23 13:35:41,546][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.04038877412676811, 'train/grad_norm_pre_clip': 44.54121398925781, 'timing/data': 0.0008462970145046711, 'timing/model': 1.9271935890428722, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 0.72}) +[2026-06-23 13:35:45,345][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.08458512276411057, 'train/grad_norm_pre_clip': 38.611385345458984, 'timing/data': 0.0005877179792150855, 'timing/model': 1.9296223389683291, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 0.72}) +[2026-06-23 13:35:49,138][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.1732192188501358, 'train/grad_norm_pre_clip': 15.143120765686035, 'timing/data': 0.0005964839365333319, 'timing/model': 1.9291353420121595, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 0.72}) +[2026-06-23 13:35:52,957][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.027586162090301514, 'train/grad_norm_pre_clip': 33.29026794433594, 'timing/data': 0.0007180750835686922, 'timing/model': 1.9418877649586648, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 0.72}) +[2026-06-23 13:35:56,765][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.13453985750675201, 'train/grad_norm_pre_clip': 42.296871185302734, 'timing/data': 0.0006300830282270908, 'timing/model': 1.9286702250828966, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 0.72}) +[2026-06-23 13:36:00,646][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.0954514592885971, 'train/grad_norm_pre_clip': 51.45051193237305, 'timing/data': 0.0006121559999883175, 'timing/model': 1.9347041490254924, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 0.72}) +[2026-06-23 13:36:04,574][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.04273652657866478, 'train/grad_norm_pre_clip': 28.16994285583496, 'timing/data': 0.000740888062864542, 'timing/model': 2.0457938939798623, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 0.72}) +[2026-06-23 13:36:08,373][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.09348911792039871, 'train/grad_norm_pre_clip': 39.96194839477539, 'timing/data': 0.0006486970232799649, 'timing/model': 1.9206207699608058, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.72}) +[2026-06-23 13:36:12,178][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.1407286673784256, 'train/grad_norm_pre_clip': 36.14407730102539, 'timing/data': 0.0008682680781930685, 'timing/model': 1.9328442399855703, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 0.72}) +[2026-06-23 13:36:15,979][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.10785701870918274, 'train/grad_norm_pre_clip': 32.10971450805664, 'timing/data': 0.0005964890588074923, 'timing/model': 1.9290140949888155, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 0.72}) +[2026-06-23 13:36:19,785][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.08662126213312149, 'train/grad_norm_pre_clip': 27.350727081298828, 'timing/data': 0.0006034719990566373, 'timing/model': 1.9301369829336181, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 0.72}) +[2026-06-23 13:36:23,592][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.13615021109580994, 'train/grad_norm_pre_clip': 74.70338439941406, 'timing/data': 0.0006527340738102794, 'timing/model': 1.9260195629904047, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.72}) +[2026-06-23 13:36:27,383][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.1434747874736786, 'train/grad_norm_pre_clip': 54.55592727661133, 'timing/data': 0.0006107209483161569, 'timing/model': 1.9189008199609816, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 0.72}) +[2026-06-23 13:36:31,190][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.05759415403008461, 'train/grad_norm_pre_clip': 34.982147216796875, 'timing/data': 0.0008373769233003259, 'timing/model': 1.926189416903071, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 0.72}) +[2026-06-23 13:36:34,989][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.01789693906903267, 'train/grad_norm_pre_clip': 14.652923583984375, 'timing/data': 0.0006751999026164412, 'timing/model': 1.9226182119455189, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 0.72}) +[2026-06-23 13:36:38,900][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.06463166326284409, 'train/grad_norm_pre_clip': 49.489532470703125, 'timing/data': 0.0006134109571576118, 'timing/model': 1.9311601949157193, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 0.72}) +[2026-06-23 13:36:42,696][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.07717841863632202, 'train/grad_norm_pre_clip': 57.56742858886719, 'timing/data': 0.0005815459880977869, 'timing/model': 1.9209418710088357, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 0.72}) +[2026-06-23 13:36:46,490][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.06509450078010559, 'train/grad_norm_pre_clip': 25.5450439453125, 'timing/data': 0.0008128569461405277, 'timing/model': 1.928064723033458, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.72}) +[2026-06-23 13:36:50,279][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.10803607851266861, 'train/grad_norm_pre_clip': 31.313114166259766, 'timing/data': 0.000745701021514833, 'timing/model': 1.922816141974181, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 0.72}) +[2026-06-23 13:36:54,085][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.023859543725848198, 'train/grad_norm_pre_clip': 9.047789573669434, 'timing/data': 0.0006866310723125935, 'timing/model': 1.9300144099397585, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 0.72}) +[2026-06-23 13:36:57,893][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.11052151769399643, 'train/grad_norm_pre_clip': 35.360557556152344, 'timing/data': 0.0007457559695467353, 'timing/model': 1.93249320902396, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 0.72}) +[2026-06-23 13:37:01,723][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.07165880501270294, 'train/grad_norm_pre_clip': 32.5394401550293, 'timing/data': 0.0006082650506868958, 'timing/model': 1.9417385120177642, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 0.73}) +[2026-06-23 13:37:05,539][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.05891555920243263, 'train/grad_norm_pre_clip': 51.38125991821289, 'timing/data': 0.000592616037465632, 'timing/model': 1.9308510300470516, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 0.73}) +[2026-06-23 13:37:09,426][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.026705417782068253, 'train/grad_norm_pre_clip': 16.094860076904297, 'timing/data': 0.0006549910176545382, 'timing/model': 2.0119216960156336, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 0.73}) +[2026-06-23 13:37:13,229][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.06354111433029175, 'train/grad_norm_pre_clip': 23.364431381225586, 'timing/data': 0.0006862389855086803, 'timing/model': 1.932615148019977, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 0.73}) +[2026-06-23 13:37:17,043][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.06535156816244125, 'train/grad_norm_pre_clip': 42.00916290283203, 'timing/data': 0.0006513110129162669, 'timing/model': 1.933919640025124, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 0.73}) +[2026-06-23 13:37:20,847][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.09288159757852554, 'train/grad_norm_pre_clip': 34.95940017700195, 'timing/data': 0.0005902760894969106, 'timing/model': 1.930535031016916, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 0.73}) +[2026-06-23 13:37:24,642][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.09861377626657486, 'train/grad_norm_pre_clip': 46.72342300415039, 'timing/data': 0.0006349800387397408, 'timing/model': 1.9223722249735147, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.73}) +[2026-06-23 13:37:28,465][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.13594330847263336, 'train/grad_norm_pre_clip': 67.33487701416016, 'timing/data': 0.0006220419891178608, 'timing/model': 1.929061662987806, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 0.73}) +[2026-06-23 13:37:32,268][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.22971762716770172, 'train/grad_norm_pre_clip': 38.87432861328125, 'timing/data': 0.000676727038808167, 'timing/model': 1.930407639942132, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 0.73}) +[2026-06-23 13:37:36,085][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.11758238822221756, 'train/grad_norm_pre_clip': 28.35917091369629, 'timing/data': 0.0006201589712873101, 'timing/model': 1.930574515950866, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 0.73}) +[2026-06-23 13:37:39,911][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.09652651101350784, 'train/grad_norm_pre_clip': 25.86436653137207, 'timing/data': 0.0006071219686418772, 'timing/model': 1.9464261180255562, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 0.73}) +[2026-06-23 13:37:43,898][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.07427060604095459, 'train/grad_norm_pre_clip': 35.13458251953125, 'timing/data': 0.0006074090488255024, 'timing/model': 2.1160642609465867, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.73}) +[2026-06-23 13:37:47,714][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.033799830824136734, 'train/grad_norm_pre_clip': 35.087039947509766, 'timing/data': 0.0008160829311236739, 'timing/model': 1.9316737750777975, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 0.73}) +[2026-06-23 13:37:51,526][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.0883280336856842, 'train/grad_norm_pre_clip': 26.82842445373535, 'timing/data': 0.0006138620665296912, 'timing/model': 1.9300624600145966, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 0.73}) +[2026-06-23 13:37:55,324][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.06009587645530701, 'train/grad_norm_pre_clip': 24.31269645690918, 'timing/data': 0.0006254070904105902, 'timing/model': 1.9297318060416728, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 0.73}) +[2026-06-23 13:37:59,131][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.04241364076733589, 'train/grad_norm_pre_clip': 50.638092041015625, 'timing/data': 0.0005979390116408467, 'timing/model': 1.9373268049675971, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 0.73}) +[2026-06-23 13:38:02,959][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.0457136444747448, 'train/grad_norm_pre_clip': 35.954158782958984, 'timing/data': 0.0006249879952520132, 'timing/model': 1.9461552259745076, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.73}) +[2026-06-23 13:38:06,768][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.09519453346729279, 'train/grad_norm_pre_clip': 31.550552368164062, 'timing/data': 0.0006068390794098377, 'timing/model': 1.935063979937695, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 0.73}) +[2026-06-23 13:38:10,583][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.051721617579460144, 'train/grad_norm_pre_clip': 47.4521484375, 'timing/data': 0.0005971119971945882, 'timing/model': 1.9399774050107226, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 0.73}) +[2026-06-23 13:38:14,383][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.06140880286693573, 'train/grad_norm_pre_clip': 20.569536209106445, 'timing/data': 0.0005908419843763113, 'timing/model': 1.9292347519658506, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 0.73}) +[2026-06-23 13:38:18,342][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.03511759266257286, 'train/grad_norm_pre_clip': 22.97646141052246, 'timing/data': 0.0006048929644748569, 'timing/model': 1.920399968046695, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 0.73}) +[2026-06-23 13:38:22,153][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.08368942141532898, 'train/grad_norm_pre_clip': 25.85298728942871, 'timing/data': 0.000662120059132576, 'timing/model': 1.9370516730705276, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 0.73}) +[2026-06-23 13:38:25,954][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.10822594165802002, 'train/grad_norm_pre_clip': 38.28554916381836, 'timing/data': 0.0006510230014100671, 'timing/model': 1.9224939789855853, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 0.73}) +[2026-06-23 13:38:29,757][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.13641467690467834, 'train/grad_norm_pre_clip': 43.55733871459961, 'timing/data': 0.0006819620030000806, 'timing/model': 1.932162987999618, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 0.73}) +[2026-06-23 13:38:33,581][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.11349258571863174, 'train/grad_norm_pre_clip': 37.28192901611328, 'timing/data': 0.0005897700320929289, 'timing/model': 1.9384995809523389, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 0.73}) +[2026-06-23 13:38:37,399][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.08743542432785034, 'train/grad_norm_pre_clip': 45.104225158691406, 'timing/data': 0.0005922470008954406, 'timing/model': 1.9400106109678745, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 0.73}) +[2026-06-23 13:38:41,207][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1900 +[2026-06-23 13:40:49,349][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1900 +[2026-06-23 13:40:49,350][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1900 +[2026-06-23 13:42:11,629][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1900 +[2026-06-23 13:42:11,630][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.05670196935534477, 'train/grad_norm_pre_clip': 12.945600509643555, 'eval/action_loss/seconds': 128.13904184498824, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04340379685163498, 'eval/latency_2/loss': 0.04340379685163498, 'eval/flappy/loss': 0.04340379685163498, 'eval/flappy/latency_2/loss': 0.04340379685163498, 'eval/flap_precision': 0.9394136807817589, 'eval/flap_recall': 0.9730094466936572, 'eval/flap_f1': 0.9559164733178654, 'eval/flap_rate_model': 0.21319444444444444, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9559164733178653, 'eval/latency_2/flap_precision': 0.9394136807817589, 'eval/latency_2/flap_recall': 0.9730094466936572, 'eval/latency_2/flap_f1': 0.9559164733178654, 'eval/latency_2/flap_rate_model': 0.21319444444444444, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9559164733178653, 'eval/action_classification/data_wait_seconds': 1.3049869537353516, 'eval/action_classification/predict_seconds': 80.75554656982422, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0008081470150500536, 'timing/model': 1.9411404069978744, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.74}) +[2026-06-23 13:42:15,391][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.1334691345691681, 'train/grad_norm_pre_clip': 24.0069522857666, 'timing/data': 0.0006138649769127369, 'timing/model': 1.9085832010023296, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 0.74}) +[2026-06-23 13:42:19,171][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.04810639098286629, 'train/grad_norm_pre_clip': 17.302734375, 'timing/data': 0.0008069209288805723, 'timing/model': 1.917966163950041, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 0.74}) +[2026-06-23 13:42:22,956][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.1371302455663681, 'train/grad_norm_pre_clip': 23.216012954711914, 'timing/data': 0.0009236640762537718, 'timing/model': 1.9209535240661353, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 0.74}) +[2026-06-23 13:42:26,737][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.11065439879894257, 'train/grad_norm_pre_clip': 34.77321243286133, 'timing/data': 0.000646676984615624, 'timing/model': 1.9149444620124996, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 0.74}) +[2026-06-23 13:42:30,512][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.02798045612871647, 'train/grad_norm_pre_clip': 17.421472549438477, 'timing/data': 0.0006374720251187682, 'timing/model': 1.919664109009318, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 0.74}) +[2026-06-23 13:42:34,306][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.15310359001159668, 'train/grad_norm_pre_clip': 28.477746963500977, 'timing/data': 0.0005880709504708648, 'timing/model': 1.9349608100019395, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 0.74}) +[2026-06-23 13:42:38,102][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.08761639893054962, 'train/grad_norm_pre_clip': 40.885101318359375, 'timing/data': 0.0009259739890694618, 'timing/model': 1.9192054069135338, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 0.74}) +[2026-06-23 13:42:41,888][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.0992460548877716, 'train/grad_norm_pre_clip': 22.715404510498047, 'timing/data': 0.000680233002640307, 'timing/model': 1.9157399799441919, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 0.74}) +[2026-06-23 13:42:45,674][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.09865837544202805, 'train/grad_norm_pre_clip': 22.975391387939453, 'timing/data': 0.0005969010526314378, 'timing/model': 1.9260691810632125, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 0.74}) +[2026-06-23 13:42:49,466][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.05450628697872162, 'train/grad_norm_pre_clip': 20.948190689086914, 'timing/data': 0.0007950419094413519, 'timing/model': 1.924669414991513, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.74}) +[2026-06-23 13:42:53,269][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.1014837771654129, 'train/grad_norm_pre_clip': 27.239173889160156, 'timing/data': 0.000612802105024457, 'timing/model': 1.9270859499229118, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 0.74}) +[2026-06-23 13:42:57,065][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.10786285251379013, 'train/grad_norm_pre_clip': 42.19959259033203, 'timing/data': 0.0006242640083655715, 'timing/model': 1.928891666000709, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 0.74}) +[2026-06-23 13:43:00,852][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.10959568619728088, 'train/grad_norm_pre_clip': 44.88115310668945, 'timing/data': 0.0006940350867807865, 'timing/model': 1.926505925017409, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.74}) +[2026-06-23 13:43:04,691][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.06616925448179245, 'train/grad_norm_pre_clip': 21.636993408203125, 'timing/data': 0.0006023880559951067, 'timing/model': 1.9269540270324796, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 0.74}) +[2026-06-23 13:43:08,493][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.1581118106842041, 'train/grad_norm_pre_clip': 57.16769790649414, 'timing/data': 0.0006533539853990078, 'timing/model': 1.9361893660388887, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 0.74}) +[2026-06-23 13:43:12,481][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.131731316447258, 'train/grad_norm_pre_clip': 56.05290222167969, 'timing/data': 0.0008985709864646196, 'timing/model': 1.9468664770247415, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 0.74}) +[2026-06-23 13:43:16,270][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.09503766149282455, 'train/grad_norm_pre_clip': 167.2194061279297, 'timing/data': 0.0006352419732138515, 'timing/model': 1.929026477970183, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 0.74}) +[2026-06-23 13:43:20,068][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.09437422454357147, 'train/grad_norm_pre_clip': 21.14109992980957, 'timing/data': 0.0007872070418670774, 'timing/model': 1.9216502680210397, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 0.74}) +[2026-06-23 13:43:23,859][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.1434919387102127, 'train/grad_norm_pre_clip': 23.307828903198242, 'timing/data': 0.0006166960811242461, 'timing/model': 1.921162115992047, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 0.74}) +[2026-06-23 13:43:27,657][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.033562954515218735, 'train/grad_norm_pre_clip': 7.686402797698975, 'timing/data': 0.000595647026784718, 'timing/model': 1.9286126770311967, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.74}) +[2026-06-23 13:43:31,457][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.048922888934612274, 'train/grad_norm_pre_clip': 28.05979347229004, 'timing/data': 0.0006822540890425444, 'timing/model': 1.9260714179836214, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 0.74}) +[2026-06-23 13:43:35,239][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.07770693302154541, 'train/grad_norm_pre_clip': 19.723541259765625, 'timing/data': 0.0006896799895912409, 'timing/model': 1.9156164160231128, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 0.74}) +[2026-06-23 13:43:39,024][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.08385233581066132, 'train/grad_norm_pre_clip': 18.717039108276367, 'timing/data': 0.000682818004861474, 'timing/model': 1.9165077150100842, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 0.74}) +[2026-06-23 13:43:43,018][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.05193916708230972, 'train/grad_norm_pre_clip': 24.960411071777344, 'timing/data': 0.0006734700873494148, 'timing/model': 2.126044420991093, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 0.74}) +[2026-06-23 13:43:46,827][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.04655807465314865, 'train/grad_norm_pre_clip': 13.846439361572266, 'timing/data': 0.0006064169574528933, 'timing/model': 1.933116654981859, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 0.75}) +[2026-06-23 13:43:50,629][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.13383749127388, 'train/grad_norm_pre_clip': 23.283586502075195, 'timing/data': 0.0006049459334462881, 'timing/model': 1.92725268995855, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 0.75}) +[2026-06-23 13:43:54,435][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.16395775973796844, 'train/grad_norm_pre_clip': 19.02145767211914, 'timing/data': 0.0006888549542054534, 'timing/model': 1.9283418389968574, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 0.75}) +[2026-06-23 13:43:58,238][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.1742093712091446, 'train/grad_norm_pre_clip': 53.56413269042969, 'timing/data': 0.0005972629878669977, 'timing/model': 1.9309436009498313, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 0.75}) +[2026-06-23 13:44:02,062][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.1381814181804657, 'train/grad_norm_pre_clip': 50.098384857177734, 'timing/data': 0.0006200490752235055, 'timing/model': 1.9490767320385203, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 0.75}) +[2026-06-23 13:44:05,854][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.1165814995765686, 'train/grad_norm_pre_clip': 38.65538024902344, 'timing/data': 0.0007461149943992496, 'timing/model': 1.9225002439925447, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.75}) +[2026-06-23 13:44:09,656][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.03679358586668968, 'train/grad_norm_pre_clip': 28.567834854125977, 'timing/data': 0.0006540940375998616, 'timing/model': 1.9310492599615827, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 0.75}) +[2026-06-23 13:44:13,438][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.08234509080648422, 'train/grad_norm_pre_clip': 41.27999496459961, 'timing/data': 0.0008970130002126098, 'timing/model': 1.9182027800707147, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 0.75}) +[2026-06-23 13:44:17,273][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.07177021354436874, 'train/grad_norm_pre_clip': 58.50532913208008, 'timing/data': 0.0006192959845066071, 'timing/model': 1.928085780935362, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 0.75}) +[2026-06-23 13:44:21,095][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.1081671267747879, 'train/grad_norm_pre_clip': 58.81887435913086, 'timing/data': 0.0006215149769559503, 'timing/model': 1.9366532190470025, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 0.75}) +[2026-06-23 13:44:24,937][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.22322383522987366, 'train/grad_norm_pre_clip': 62.68942642211914, 'timing/data': 0.0005962230497971177, 'timing/model': 1.9310096009867266, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 0.75}) +[2026-06-23 13:44:28,756][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.01557285338640213, 'train/grad_norm_pre_clip': 43.94001388549805, 'timing/data': 0.0005840570665895939, 'timing/model': 1.947336651966907, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 0.75}) +[2026-06-23 13:44:32,570][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.05305828899145126, 'train/grad_norm_pre_clip': 66.28340148925781, 'timing/data': 0.0006670430302619934, 'timing/model': 1.9355808040127158, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 0.75}) +[2026-06-23 13:44:36,410][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.0276639461517334, 'train/grad_norm_pre_clip': 22.284955978393555, 'timing/data': 0.0006442220183089375, 'timing/model': 1.9313703960506245, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 0.75}) +[2026-06-23 13:44:40,232][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.07628405094146729, 'train/grad_norm_pre_clip': 38.57770919799805, 'timing/data': 0.0006033499957993627, 'timing/model': 1.944799138000235, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 0.75}) +[2026-06-23 13:44:44,076][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.02417982742190361, 'train/grad_norm_pre_clip': 16.26041603088379, 'timing/data': 0.0006304919952526689, 'timing/model': 1.9534500800073147, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.75}) +[2026-06-23 13:44:47,905][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.1092059463262558, 'train/grad_norm_pre_clip': 52.33175277709961, 'timing/data': 0.0006104530766606331, 'timing/model': 1.9519010999938473, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 0.75}) +[2026-06-23 13:44:51,731][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.15123093128204346, 'train/grad_norm_pre_clip': 36.117652893066406, 'timing/data': 0.0008654730627313256, 'timing/model': 1.9465259059797972, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 0.75}) +[2026-06-23 13:44:55,548][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.09831289201974869, 'train/grad_norm_pre_clip': 52.55351638793945, 'timing/data': 0.0006046360358595848, 'timing/model': 1.9467406260082498, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 0.75}) +[2026-06-23 13:44:59,362][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.0901980847120285, 'train/grad_norm_pre_clip': 34.2819709777832, 'timing/data': 0.00060489890165627, 'timing/model': 1.9469334980240092, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 0.75}) +[2026-06-23 13:45:03,179][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.052681513130664825, 'train/grad_norm_pre_clip': 21.592044830322266, 'timing/data': 0.0006736359791830182, 'timing/model': 1.9547208430012688, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 0.75}) +[2026-06-23 13:45:06,979][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.09478402137756348, 'train/grad_norm_pre_clip': 31.886951446533203, 'timing/data': 0.0006461790762841702, 'timing/model': 1.9237765580182895, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 0.75}) +[2026-06-23 13:45:10,778][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.08190010488033295, 'train/grad_norm_pre_clip': 48.40689468383789, 'timing/data': 0.000641272054053843, 'timing/model': 1.9312734130071476, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 0.75}) +[2026-06-23 13:45:14,647][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.1076049879193306, 'train/grad_norm_pre_clip': 29.439289093017578, 'timing/data': 0.0006134159630164504, 'timing/model': 1.988334437017329, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 0.75}) +[2026-06-23 13:45:18,459][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.01965297758579254, 'train/grad_norm_pre_clip': 27.398134231567383, 'timing/data': 0.0006107139633968472, 'timing/model': 1.9321900950744748, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 0.75}) +[2026-06-23 13:45:22,273][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.12460485845804214, 'train/grad_norm_pre_clip': 28.157350540161133, 'timing/data': 0.0006084690103307366, 'timing/model': 1.9286791040794924, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.75}) +[2026-06-23 13:45:26,084][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.06188574433326721, 'train/grad_norm_pre_clip': 31.88127899169922, 'timing/data': 0.0005934040527790785, 'timing/model': 1.935890355031006, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 0.76}) +[2026-06-23 13:45:29,889][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.11138425767421722, 'train/grad_norm_pre_clip': 57.666202545166016, 'timing/data': 0.0005931260529905558, 'timing/model': 1.9282756650354713, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 0.76}) +[2026-06-23 13:45:33,687][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.07757552713155746, 'train/grad_norm_pre_clip': 47.30316925048828, 'timing/data': 0.0005936339730396867, 'timing/model': 1.931636203895323, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 0.76}) +[2026-06-23 13:45:37,541][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.0861857533454895, 'train/grad_norm_pre_clip': 24.74428367614746, 'timing/data': 0.0005940379342064261, 'timing/model': 1.9716269930358976, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 0.76}) +[2026-06-23 13:45:41,340][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.037975624203681946, 'train/grad_norm_pre_clip': 18.657920837402344, 'timing/data': 0.0006133000133559108, 'timing/model': 1.930188077967614, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.76}) +[2026-06-23 13:45:45,151][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.0397462360560894, 'train/grad_norm_pre_clip': 36.392311096191406, 'timing/data': 0.0006028770003467798, 'timing/model': 1.933439576998353, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 0.76}) +[2026-06-23 13:45:48,969][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.07173604518175125, 'train/grad_norm_pre_clip': 23.487428665161133, 'timing/data': 0.0005950049962848425, 'timing/model': 1.9419770840322599, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 0.76}) +[2026-06-23 13:45:52,774][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.1995692104101181, 'train/grad_norm_pre_clip': 36.7861328125, 'timing/data': 0.0009140189504250884, 'timing/model': 1.9291195860132575, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 0.76}) +[2026-06-23 13:45:56,588][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.04784668609499931, 'train/grad_norm_pre_clip': 17.973825454711914, 'timing/data': 0.0006524119526147842, 'timing/model': 1.933021822012961, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 0.76}) +[2026-06-23 13:46:00,408][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.1242612972855568, 'train/grad_norm_pre_clip': 26.79178810119629, 'timing/data': 0.0006736699724569917, 'timing/model': 1.9387755719944835, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.76}) +[2026-06-23 13:46:04,243][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.06379088759422302, 'train/grad_norm_pre_clip': 29.75504493713379, 'timing/data': 0.0006111019756644964, 'timing/model': 1.9297174479579553, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 0.76}) +[2026-06-23 13:46:08,107][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.15828058123588562, 'train/grad_norm_pre_clip': 27.130847930908203, 'timing/data': 0.0005940919509157538, 'timing/model': 1.9946055409964174, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.76}) +[2026-06-23 13:46:11,917][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.03291051462292671, 'train/grad_norm_pre_clip': 29.929262161254883, 'timing/data': 0.0007244820008054376, 'timing/model': 1.9341144899372011, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 0.76}) +[2026-06-23 13:46:15,726][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.07071065902709961, 'train/grad_norm_pre_clip': 48.036251068115234, 'timing/data': 0.0006498469738289714, 'timing/model': 1.9311528149992228, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 0.76}) +[2026-06-23 13:46:19,533][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.08455675840377808, 'train/grad_norm_pre_clip': 24.307540893554688, 'timing/data': 0.0005948879988864064, 'timing/model': 1.9298254819586873, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 0.76}) +[2026-06-23 13:46:23,334][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.027125416323542595, 'train/grad_norm_pre_clip': 42.52671432495117, 'timing/data': 0.0005878469673916698, 'timing/model': 1.9364023690577596, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 0.76}) +[2026-06-23 13:46:27,136][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.09371499717235565, 'train/grad_norm_pre_clip': 57.945884704589844, 'timing/data': 0.0005808919668197632, 'timing/model': 1.921699836035259, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 0.76}) +[2026-06-23 13:46:30,929][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.11245805770158768, 'train/grad_norm_pre_clip': 32.53624725341797, 'timing/data': 0.0006081209285184741, 'timing/model': 1.9218710259301588, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 0.76}) +[2026-06-23 13:46:34,726][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.07090826332569122, 'train/grad_norm_pre_clip': 23.95909309387207, 'timing/data': 0.0006333909695968032, 'timing/model': 1.9255999649176374, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 0.76}) +[2026-06-23 13:46:38,533][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.06339520961046219, 'train/grad_norm_pre_clip': 35.68519592285156, 'timing/data': 0.000603565014898777, 'timing/model': 1.9304466129979119, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.76}) +[2026-06-23 13:46:42,345][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.21380789577960968, 'train/grad_norm_pre_clip': 48.10965347290039, 'timing/data': 0.0006058620056137443, 'timing/model': 1.933593557914719, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 0.76}) +[2026-06-23 13:46:46,155][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.07442140579223633, 'train/grad_norm_pre_clip': 16.58094596862793, 'timing/data': 0.0005941010313108563, 'timing/model': 1.9376571129541844, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 0.76}) +[2026-06-23 13:46:49,968][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.06281422078609467, 'train/grad_norm_pre_clip': 15.423324584960938, 'timing/data': 0.0005930230254307389, 'timing/model': 1.9306064279517159, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 0.76}) +[2026-06-23 13:46:53,782][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.06480341404676437, 'train/grad_norm_pre_clip': 25.03157615661621, 'timing/data': 0.000980651006102562, 'timing/model': 1.938012331025675, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 0.76}) +[2026-06-23 13:46:57,655][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.08346662670373917, 'train/grad_norm_pre_clip': 39.03630447387695, 'timing/data': 0.0006610630080103874, 'timing/model': 1.9954316560178995, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 0.76}) +[2026-06-23 13:47:01,462][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.09567105025053024, 'train/grad_norm_pre_clip': 24.98832893371582, 'timing/data': 0.0006026270566508174, 'timing/model': 1.929318929091096, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 0.76}) +[2026-06-23 13:47:05,262][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.039073195308446884, 'train/grad_norm_pre_clip': 20.41997528076172, 'timing/data': 0.0009142659837380052, 'timing/model': 1.927219785982743, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 0.77}) +[2026-06-23 13:47:09,064][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.06647973507642746, 'train/grad_norm_pre_clip': 28.150474548339844, 'timing/data': 0.000780488015152514, 'timing/model': 1.9320630059810355, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 0.77}) +[2026-06-23 13:47:12,884][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.14222930371761322, 'train/grad_norm_pre_clip': 33.47940444946289, 'timing/data': 0.0008124019950628281, 'timing/model': 1.9424236519262195, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 0.77}) +[2026-06-23 13:47:16,685][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.07140182703733444, 'train/grad_norm_pre_clip': 26.124868392944336, 'timing/data': 0.0006897730054333806, 'timing/model': 1.9283708180300891, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.77}) +[2026-06-23 13:47:20,501][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.05765258148312569, 'train/grad_norm_pre_clip': 38.93940734863281, 'timing/data': 0.0006047768983989954, 'timing/model': 1.9355271080275998, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 0.77}) +[2026-06-23 13:47:24,314][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.12463998794555664, 'train/grad_norm_pre_clip': 31.001798629760742, 'timing/data': 0.0005980939604341984, 'timing/model': 1.9355441150255501, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 0.77}) +[2026-06-23 13:47:28,124][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.04930231347680092, 'train/grad_norm_pre_clip': 18.302074432373047, 'timing/data': 0.0005984980380162597, 'timing/model': 1.9304615490837023, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 0.77}) +[2026-06-23 13:47:32,113][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.07601029425859451, 'train/grad_norm_pre_clip': 37.422508239746094, 'timing/data': 0.000599822960793972, 'timing/model': 2.1179479480488226, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 0.77}) +[2026-06-23 13:47:35,930][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.05407696217298508, 'train/grad_norm_pre_clip': 30.197431564331055, 'timing/data': 0.000790830934420228, 'timing/model': 1.9315372310811654, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 0.77}) +[2026-06-23 13:47:39,743][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.018084395676851273, 'train/grad_norm_pre_clip': 12.382512092590332, 'timing/data': 0.0006106910295784473, 'timing/model': 1.9319900799309835, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 0.77}) +[2026-06-23 13:47:43,552][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.058817360550165176, 'train/grad_norm_pre_clip': 22.12512969970703, 'timing/data': 0.0006143179489299655, 'timing/model': 1.9344806730514392, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 0.77}) +[2026-06-23 13:47:47,352][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.03409373760223389, 'train/grad_norm_pre_clip': 37.03084945678711, 'timing/data': 0.0006111060502007604, 'timing/model': 1.9253675199579448, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 0.77}) +[2026-06-23 13:47:51,203][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.07878243923187256, 'train/grad_norm_pre_clip': 52.113311767578125, 'timing/data': 0.0007388909580186009, 'timing/model': 1.9747812529094517, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 0.77}) +[2026-06-23 13:47:55,025][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.0955662876367569, 'train/grad_norm_pre_clip': 29.285566329956055, 'timing/data': 0.0006015680264681578, 'timing/model': 1.9503663330106065, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.77}) +[2026-06-23 13:47:58,842][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.0509938970208168, 'train/grad_norm_pre_clip': 25.022001266479492, 'timing/data': 0.0005936740199103951, 'timing/model': 1.9363239420345053, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 0.77}) +[2026-06-23 13:48:02,650][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.045742206275463104, 'train/grad_norm_pre_clip': 24.135347366333008, 'timing/data': 0.0005934269865974784, 'timing/model': 1.9277818680275232, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 0.77}) +[2026-06-23 13:48:06,462][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.03278777003288269, 'train/grad_norm_pre_clip': 45.578773498535156, 'timing/data': 0.0005968690384179354, 'timing/model': 1.934307181974873, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 0.77}) +[2026-06-23 13:48:10,271][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.12582114338874817, 'train/grad_norm_pre_clip': 58.84891891479492, 'timing/data': 0.0006360309198498726, 'timing/model': 1.9328108580084518, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 0.77}) +[2026-06-23 13:48:14,086][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.11299954354763031, 'train/grad_norm_pre_clip': 33.20707702636719, 'timing/data': 0.0005863209953531623, 'timing/model': 1.9374036899534985, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 0.77}) +[2026-06-23 13:48:17,896][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.07759296894073486, 'train/grad_norm_pre_clip': 52.41118621826172, 'timing/data': 0.0005907989107072353, 'timing/model': 1.9344630379928276, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 0.77}) +[2026-06-23 13:48:21,702][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.044639911502599716, 'train/grad_norm_pre_clip': 22.140594482421875, 'timing/data': 0.0006963859777897596, 'timing/model': 1.9327811499824747, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 0.77}) +[2026-06-23 13:48:25,521][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.08470160514116287, 'train/grad_norm_pre_clip': 34.491729736328125, 'timing/data': 0.0006033380050212145, 'timing/model': 1.9462481080554426, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 0.77}) +[2026-06-23 13:48:29,347][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.04922376945614815, 'train/grad_norm_pre_clip': 17.581022262573242, 'timing/data': 0.0007272729417309165, 'timing/model': 1.9385409249225631, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 0.77}) +[2026-06-23 13:48:33,149][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2000 +[2026-06-23 13:50:39,761][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2000 +[2026-06-23 13:50:39,763][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2000 +[2026-06-23 13:52:03,600][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2000 +[2026-06-23 13:55:14,498][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-06-23 13:55:14,499][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.03064841963350773, 'train/grad_norm_pre_clip': 15.124638557434082, 'eval/action_loss/seconds': 126.60993193101604, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.040764518082141876, 'eval/latency_2/loss': 0.040764518082141876, 'eval/flappy/loss': 0.040764518082141876, 'eval/flappy/latency_2/loss': 0.040764518082141876, 'eval/flap_precision': 0.9497021839841164, 'eval/flap_recall': 0.9682860998650472, 'eval/flap_f1': 0.9589041095890412, 'eval/flap_rate_model': 0.2098611111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9589041095890413, 'eval/latency_2/flap_precision': 0.9497021839841164, 'eval/latency_2/flap_recall': 0.9682860998650472, 'eval/latency_2/flap_f1': 0.9589041095890412, 'eval/latency_2/flap_rate_model': 0.2098611111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9589041095890413, 'eval/action_classification/data_wait_seconds': 1.733935832977295, 'eval/action_classification/predict_seconds': 81.93219757080078, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 115.8999999999974, 'rl_games_eval/mid_train/mean_length': 954.0, 'rl_games_eval/mid_train/std_reward': 81.98780397107578, 'rl_games_eval/mid_train/std_length': 661.8280743516401, 'rl_games_eval/mid_train/macro_mean_reward': 115.8999999999974, 'rl_games_eval/mid_train/macro_mean_length': 954.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 115.8999999999974, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 954.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 81.98780397107578, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 661.8280743516401, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 115.8999999999974, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006519269663840532, 'timing/model': 1.9307173170382157, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.77}) +[2026-06-23 13:55:14,499][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_2000_state +[2026-06-23 13:55:14,502][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 13:55:18,225][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_2000_state/model.safetensors +[2026-06-23 13:55:21,183][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_2000_state/optimizer.bin +[2026-06-23 13:55:21,183][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_2000_state/sampler.bin +[2026-06-23 13:55:21,183][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-23 13:55:21,184][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-23 13:55:21,214][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-23 13:55:21,214][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 13:55:21,228][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 13:55:24,979][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.08882071822881699, 'train/grad_norm_pre_clip': 40.408782958984375, 'timing/data': 0.0006093040574342012, 'timing/model': 1.8910357430577278, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 0.77}) +[2026-06-23 13:55:28,713][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.06928559392690659, 'train/grad_norm_pre_clip': 21.1903076171875, 'timing/data': 0.0008763460209593177, 'timing/model': 1.890522136935033, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 0.77}) +[2026-06-23 13:55:32,461][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.12342799454927444, 'train/grad_norm_pre_clip': 51.74626159667969, 'timing/data': 0.0006894319085404277, 'timing/model': 1.9047807230381295, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 0.78}) +[2026-06-23 13:55:36,215][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.0631239265203476, 'train/grad_norm_pre_clip': 28.90294075012207, 'timing/data': 0.0006134799914434552, 'timing/model': 1.9050410349154845, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 0.78}) +[2026-06-23 13:55:39,975][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.1865040510892868, 'train/grad_norm_pre_clip': 37.22856140136719, 'timing/data': 0.0006190889980643988, 'timing/model': 1.9115763049339876, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 0.78}) +[2026-06-23 13:55:43,738][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.04558612406253815, 'train/grad_norm_pre_clip': 20.24359703063965, 'timing/data': 0.0008488779421895742, 'timing/model': 1.9116156559903175, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 0.78}) +[2026-06-23 13:55:47,503][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.02319294959306717, 'train/grad_norm_pre_clip': 27.314579010009766, 'timing/data': 0.0006817140383645892, 'timing/model': 1.9057588790310547, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 0.78}) +[2026-06-23 13:55:51,260][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.06296674907207489, 'train/grad_norm_pre_clip': 23.335918426513672, 'timing/data': 0.0005892980843782425, 'timing/model': 1.9110268160002306, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 0.78}) +[2026-06-23 13:55:55,202][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.10582799464464188, 'train/grad_norm_pre_clip': 28.384178161621094, 'timing/data': 0.0005837399512529373, 'timing/model': 1.9054369369987398, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 0.78}) +[2026-06-23 13:55:58,977][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.03570741042494774, 'train/grad_norm_pre_clip': 28.23688507080078, 'timing/data': 0.0005957619287073612, 'timing/model': 1.918037711060606, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.78}) +[2026-06-23 13:56:02,745][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.04953497648239136, 'train/grad_norm_pre_clip': 25.807884216308594, 'timing/data': 0.0006698699435219169, 'timing/model': 1.9149182189721614, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.78}) +[2026-06-23 13:56:06,523][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.05278626084327698, 'train/grad_norm_pre_clip': 24.79336166381836, 'timing/data': 0.0006579060573130846, 'timing/model': 1.920636257971637, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 0.78}) +[2026-06-23 13:56:10,420][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.03009558841586113, 'train/grad_norm_pre_clip': 19.065109252929688, 'timing/data': 0.0006657750345766544, 'timing/model': 1.9130719649838284, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 0.78}) +[2026-06-23 13:56:14,220][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.03023098036646843, 'train/grad_norm_pre_clip': 15.821425437927246, 'timing/data': 0.0006657500052824616, 'timing/model': 1.909617303055711, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 0.78}) +[2026-06-23 13:56:18,004][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.06977038830518723, 'train/grad_norm_pre_clip': 28.97456932067871, 'timing/data': 0.0006129459943622351, 'timing/model': 1.9123322940431535, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 0.78}) +[2026-06-23 13:56:21,794][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.011817612685263157, 'train/grad_norm_pre_clip': 33.174041748046875, 'timing/data': 0.0006136288866400719, 'timing/model': 1.9212933359667659, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 0.78}) +[2026-06-23 13:56:25,761][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.025506503880023956, 'train/grad_norm_pre_clip': 26.555530548095703, 'timing/data': 0.0006143930368125439, 'timing/model': 2.0893911690218374, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 0.78}) +[2026-06-23 13:56:29,543][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.1455516666173935, 'train/grad_norm_pre_clip': 33.79460906982422, 'timing/data': 0.0005963250296190381, 'timing/model': 1.9184719190234318, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 0.78}) +[2026-06-23 13:56:33,333][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.05370371416211128, 'train/grad_norm_pre_clip': 33.51004409790039, 'timing/data': 0.0005999349523335695, 'timing/model': 1.9224134950200096, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 0.78}) +[2026-06-23 13:56:37,115][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.07770861685276031, 'train/grad_norm_pre_clip': 71.35269165039062, 'timing/data': 0.0006084690103307366, 'timing/model': 1.914735333994031, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.78}) +[2026-06-23 13:56:40,899][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.04418075829744339, 'train/grad_norm_pre_clip': 20.27053451538086, 'timing/data': 0.0006436220137402415, 'timing/model': 1.9165637369733304, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 0.78}) +[2026-06-23 13:56:44,684][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.06680569052696228, 'train/grad_norm_pre_clip': 53.080589294433594, 'timing/data': 0.0006299599772319198, 'timing/model': 1.9186541420640424, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 0.78}) +[2026-06-23 13:56:48,480][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.061343658715486526, 'train/grad_norm_pre_clip': 20.4753360748291, 'timing/data': 0.000601212028414011, 'timing/model': 1.9273288539843634, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 0.78}) +[2026-06-23 13:56:52,272][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.025118058547377586, 'train/grad_norm_pre_clip': 14.688697814941406, 'timing/data': 0.0006009929347783327, 'timing/model': 1.9263294449774548, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 0.78}) +[2026-06-23 13:56:56,062][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.18938805162906647, 'train/grad_norm_pre_clip': 64.34302520751953, 'timing/data': 0.0005988619523122907, 'timing/model': 1.92903284705244, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.78}) +[2026-06-23 13:57:00,042][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.23052352666854858, 'train/grad_norm_pre_clip': 54.29349899291992, 'timing/data': 0.0005884600104764104, 'timing/model': 1.9225678690709174, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 0.78}) +[2026-06-23 13:57:03,825][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.10739503055810928, 'train/grad_norm_pre_clip': 47.79802322387695, 'timing/data': 0.0006539609748870134, 'timing/model': 1.9172125218901783, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 0.78}) +[2026-06-23 13:57:07,619][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.0997430607676506, 'train/grad_norm_pre_clip': 84.55641174316406, 'timing/data': 0.0005935570225119591, 'timing/model': 1.9248649589717388, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 0.78}) +[2026-06-23 13:57:11,408][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.03540302813053131, 'train/grad_norm_pre_clip': 17.66881561279297, 'timing/data': 0.0006171800196170807, 'timing/model': 1.9267778149805963, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 0.79}) +[2026-06-23 13:57:15,197][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.05921820178627968, 'train/grad_norm_pre_clip': 27.33469009399414, 'timing/data': 0.0005760339554399252, 'timing/model': 1.9256868729135022, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.79}) +[2026-06-23 13:57:19,003][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.0733065977692604, 'train/grad_norm_pre_clip': 47.66811752319336, 'timing/data': 0.0006078699370846152, 'timing/model': 1.9333202739944682, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 0.79}) +[2026-06-23 13:57:22,790][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.17198996245861053, 'train/grad_norm_pre_clip': 50.606414794921875, 'timing/data': 0.0008232389809563756, 'timing/model': 1.924244966940023, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 0.79}) +[2026-06-23 13:57:26,586][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.004830247722566128, 'train/grad_norm_pre_clip': 20.1593074798584, 'timing/data': 0.0005822429666295648, 'timing/model': 1.9263065290870145, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 0.79}) +[2026-06-23 13:57:30,387][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.06512844562530518, 'train/grad_norm_pre_clip': 29.930906295776367, 'timing/data': 0.0005900630494579673, 'timing/model': 1.928087678970769, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 0.79}) +[2026-06-23 13:57:34,193][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.02538795955479145, 'train/grad_norm_pre_clip': 28.789548873901367, 'timing/data': 0.0006648669950664043, 'timing/model': 1.9326240820810199, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 0.79}) +[2026-06-23 13:57:38,012][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.03152278810739517, 'train/grad_norm_pre_clip': 28.069040298461914, 'timing/data': 0.0005828459979966283, 'timing/model': 1.924685250967741, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 0.79}) +[2026-06-23 13:57:41,805][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.08584174513816833, 'train/grad_norm_pre_clip': 33.28466033935547, 'timing/data': 0.0006035780534148216, 'timing/model': 1.926590527058579, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 0.79}) +[2026-06-23 13:57:45,614][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.10237225890159607, 'train/grad_norm_pre_clip': 40.93968200683594, 'timing/data': 0.00084633007645607, 'timing/model': 1.9259477849118412, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 0.79}) +[2026-06-23 13:57:49,411][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.0943567231297493, 'train/grad_norm_pre_clip': 58.316871643066406, 'timing/data': 0.0006117139710113406, 'timing/model': 1.9335332789923996, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 0.79}) +[2026-06-23 13:57:53,222][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.05529375746846199, 'train/grad_norm_pre_clip': 33.82387924194336, 'timing/data': 0.0006144100334495306, 'timing/model': 1.9307810529135168, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.79}) +[2026-06-23 13:57:57,014][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.12008645385503769, 'train/grad_norm_pre_clip': 28.981508255004883, 'timing/data': 0.0009530429961159825, 'timing/model': 1.9202888120198622, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 0.79}) +[2026-06-23 13:58:00,820][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.1403159201145172, 'train/grad_norm_pre_clip': 62.759796142578125, 'timing/data': 0.0008106570458039641, 'timing/model': 1.9293450380209833, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 0.79}) +[2026-06-23 13:58:04,626][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.08271260559558868, 'train/grad_norm_pre_clip': 44.39154815673828, 'timing/data': 0.000600390019826591, 'timing/model': 1.930419263895601, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 0.79}) +[2026-06-23 13:58:08,433][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.14555366337299347, 'train/grad_norm_pre_clip': 52.84505081176758, 'timing/data': 0.000896541983820498, 'timing/model': 1.9301154220011085, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 0.79}) +[2026-06-23 13:58:12,228][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.0589427687227726, 'train/grad_norm_pre_clip': 22.082454681396484, 'timing/data': 0.0005988950142636895, 'timing/model': 1.9244537190534174, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 0.79}) +[2026-06-23 13:58:16,048][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.03373816981911659, 'train/grad_norm_pre_clip': 18.884136199951172, 'timing/data': 0.0008014880586415529, 'timing/model': 1.9359065149910748, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 0.79}) +[2026-06-23 13:58:19,847][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.15447324514389038, 'train/grad_norm_pre_clip': 26.901884078979492, 'timing/data': 0.0006177669856697321, 'timing/model': 1.923222728073597, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 0.79}) +[2026-06-23 13:58:23,638][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.27655577659606934, 'train/grad_norm_pre_clip': 91.36463165283203, 'timing/data': 0.0006452410016208887, 'timing/model': 1.927463085972704, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 0.79}) +[2026-06-23 13:58:27,439][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.09224879741668701, 'train/grad_norm_pre_clip': 46.86577606201172, 'timing/data': 0.0008144039893522859, 'timing/model': 1.9246780059766024, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 0.79}) +[2026-06-23 13:58:31,238][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.10883592814207077, 'train/grad_norm_pre_clip': 49.986488342285156, 'timing/data': 0.0006757950177416205, 'timing/model': 1.9205409170826897, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.79}) +[2026-06-23 13:58:35,038][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.11506092548370361, 'train/grad_norm_pre_clip': 42.9957275390625, 'timing/data': 0.0008474080823361874, 'timing/model': 1.9280303079867736, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 0.79}) +[2026-06-23 13:58:38,852][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.09742888808250427, 'train/grad_norm_pre_clip': 38.09703063964844, 'timing/data': 0.0008405260741710663, 'timing/model': 1.932641533901915, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 0.79}) +[2026-06-23 13:58:42,655][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.052076902240514755, 'train/grad_norm_pre_clip': 16.893930435180664, 'timing/data': 0.000776752014644444, 'timing/model': 1.9274387249024585, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 0.79}) +[2026-06-23 13:58:46,468][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.05894901603460312, 'train/grad_norm_pre_clip': 23.12336540222168, 'timing/data': 0.0007527699926868081, 'timing/model': 1.9332398059777915, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 0.79}) +[2026-06-23 13:58:50,278][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.10049736499786377, 'train/grad_norm_pre_clip': 23.21988296508789, 'timing/data': 0.0007452919380739331, 'timing/model': 1.9300101370317861, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 0.8}) +[2026-06-23 13:58:54,085][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.06912674009799957, 'train/grad_norm_pre_clip': 25.400625228881836, 'timing/data': 0.0005985270254313946, 'timing/model': 1.9305178410140797, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 0.8}) +[2026-06-23 13:58:57,884][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.09980393946170807, 'train/grad_norm_pre_clip': 42.23150634765625, 'timing/data': 0.0006398479454219341, 'timing/model': 1.9277787669561803, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 0.8}) +[2026-06-23 13:59:01,763][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.13362285494804382, 'train/grad_norm_pre_clip': 37.84138488769531, 'timing/data': 0.0006633219309151173, 'timing/model': 1.9954000260913745, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 0.8}) +[2026-06-23 13:59:05,564][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.06962986290454865, 'train/grad_norm_pre_clip': 23.963790893554688, 'timing/data': 0.0006924020126461983, 'timing/model': 1.9285197190474719, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 0.8}) +[2026-06-23 13:59:09,372][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.1555604785680771, 'train/grad_norm_pre_clip': 30.320083618164062, 'timing/data': 0.0006117450539022684, 'timing/model': 1.9358867169357836, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.8}) +[2026-06-23 13:59:13,181][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.05633380636572838, 'train/grad_norm_pre_clip': 19.613374710083008, 'timing/data': 0.0005957909161224961, 'timing/model': 1.9388880740152672, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 0.8}) +[2026-06-23 13:59:16,998][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.08623913675546646, 'train/grad_norm_pre_clip': 36.44807052612305, 'timing/data': 0.0006019310094416142, 'timing/model': 1.9360976339085028, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 0.8}) +[2026-06-23 13:59:20,802][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.13686390221118927, 'train/grad_norm_pre_clip': 46.070762634277344, 'timing/data': 0.000584885012358427, 'timing/model': 1.937476916005835, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 0.8}) +[2026-06-23 13:59:24,622][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.08943705260753632, 'train/grad_norm_pre_clip': 30.889755249023438, 'timing/data': 0.000701078912243247, 'timing/model': 1.937484650989063, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 0.8}) +[2026-06-23 13:59:28,422][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.042253024876117706, 'train/grad_norm_pre_clip': 20.665613174438477, 'timing/data': 0.0005915670190006495, 'timing/model': 1.9310841859551147, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 0.8}) +[2026-06-23 13:59:32,223][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.07321853190660477, 'train/grad_norm_pre_clip': 12.852712631225586, 'timing/data': 0.0005900110118091106, 'timing/model': 1.9337056200020015, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 0.8}) +[2026-06-23 13:59:36,037][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.1341310292482376, 'train/grad_norm_pre_clip': 40.012508392333984, 'timing/data': 0.0006164780352264643, 'timing/model': 1.936829891987145, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 0.8}) +[2026-06-23 13:59:39,843][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.04740298166871071, 'train/grad_norm_pre_clip': 25.8675594329834, 'timing/data': 0.00059512909501791, 'timing/model': 1.9315728429937735, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 0.8}) +[2026-06-23 13:59:43,830][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.07066263258457184, 'train/grad_norm_pre_clip': 17.465999603271484, 'timing/data': 0.0005985230673104525, 'timing/model': 1.9331379900686443, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 0.8}) +[2026-06-23 13:59:47,635][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.04150236397981644, 'train/grad_norm_pre_clip': 11.29654598236084, 'timing/data': 0.0005871800240129232, 'timing/model': 1.9320789059856907, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.8}) +[2026-06-23 13:59:51,492][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.08932957053184509, 'train/grad_norm_pre_clip': 35.72686767578125, 'timing/data': 0.0005865109851583838, 'timing/model': 1.9287792169488966, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 0.8}) +[2026-06-23 13:59:55,305][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.02852533385157585, 'train/grad_norm_pre_clip': 11.898432731628418, 'timing/data': 0.000610093935392797, 'timing/model': 1.9349930540192872, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 0.8}) +[2026-06-23 13:59:59,117][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.10513539612293243, 'train/grad_norm_pre_clip': 39.71388244628906, 'timing/data': 0.0005956839304417372, 'timing/model': 1.929278994910419, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 0.8}) +[2026-06-23 14:00:02,932][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.01689760945737362, 'train/grad_norm_pre_clip': 11.244138717651367, 'timing/data': 0.0006221970543265343, 'timing/model': 1.9374563769670203, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 0.8}) +[2026-06-23 14:00:06,750][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.06638269871473312, 'train/grad_norm_pre_clip': 25.47138786315918, 'timing/data': 0.0007322509773075581, 'timing/model': 1.9462119669187814, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 0.8}) +[2026-06-23 14:00:10,582][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.11030824482440948, 'train/grad_norm_pre_clip': 42.1654167175293, 'timing/data': 0.0009353349450975657, 'timing/model': 1.9444674430415034, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 0.8}) +[2026-06-23 14:00:14,400][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.1661357879638672, 'train/grad_norm_pre_clip': 85.21351623535156, 'timing/data': 0.0007686719764024019, 'timing/model': 1.9341102899052203, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 0.8}) +[2026-06-23 14:00:18,211][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.09848438203334808, 'train/grad_norm_pre_clip': 37.10791778564453, 'timing/data': 0.0007767700590193272, 'timing/model': 1.9463150500087067, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 0.8}) +[2026-06-23 14:00:22,036][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.07032029330730438, 'train/grad_norm_pre_clip': 51.229026794433594, 'timing/data': 0.0006058349972590804, 'timing/model': 1.9455989379202947, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 0.8}) +[2026-06-23 14:00:25,882][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.05364164710044861, 'train/grad_norm_pre_clip': 33.900787353515625, 'timing/data': 0.0005990819772705436, 'timing/model': 1.9433410639176145, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.8}) +[2026-06-23 14:00:29,779][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.13331523537635803, 'train/grad_norm_pre_clip': 45.02079391479492, 'timing/data': 0.0006049060029909015, 'timing/model': 2.014262038981542, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 0.81}) +[2026-06-23 14:00:33,594][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.13610942661762238, 'train/grad_norm_pre_clip': 68.1447982788086, 'timing/data': 0.0006046859780326486, 'timing/model': 1.9466707300161943, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 0.81}) +[2026-06-23 14:00:37,405][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.01990988478064537, 'train/grad_norm_pre_clip': 23.879318237304688, 'timing/data': 0.0006155739538371563, 'timing/model': 1.9409441989846528, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 0.81}) +[2026-06-23 14:00:41,229][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.09033973515033722, 'train/grad_norm_pre_clip': 33.052974700927734, 'timing/data': 0.0006138229509815574, 'timing/model': 1.941343039041385, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 0.81}) +[2026-06-23 14:00:45,051][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.020411640405654907, 'train/grad_norm_pre_clip': 16.872346878051758, 'timing/data': 0.0006102100014686584, 'timing/model': 1.9356242520734668, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 0.81}) +[2026-06-23 14:00:48,878][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.1521376222372055, 'train/grad_norm_pre_clip': 33.536712646484375, 'timing/data': 0.0006103910272940993, 'timing/model': 1.9345753790112212, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 0.81}) +[2026-06-23 14:00:52,688][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.051214251667261124, 'train/grad_norm_pre_clip': 18.124065399169922, 'timing/data': 0.0006022279849275947, 'timing/model': 1.928772822022438, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 0.81}) +[2026-06-23 14:00:56,483][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.05988186597824097, 'train/grad_norm_pre_clip': 37.84868240356445, 'timing/data': 0.0006078209262341261, 'timing/model': 1.9178296200698242, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 0.81}) +[2026-06-23 14:01:00,291][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.1775393784046173, 'train/grad_norm_pre_clip': 35.8917350769043, 'timing/data': 0.0005915679503232241, 'timing/model': 1.924028862034902, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 0.81}) +[2026-06-23 14:01:04,106][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.08307211101055145, 'train/grad_norm_pre_clip': 28.32915496826172, 'timing/data': 0.0006798329995945096, 'timing/model': 1.9297042490215972, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.81}) +[2026-06-23 14:01:07,920][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.05034782364964485, 'train/grad_norm_pre_clip': 42.17577362060547, 'timing/data': 0.0005909200990572572, 'timing/model': 1.936937803053297, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 0.81}) +[2026-06-23 14:01:11,716][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.055828604847192764, 'train/grad_norm_pre_clip': 27.564802169799805, 'timing/data': 0.0005904519930481911, 'timing/model': 1.9242201859597117, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 0.81}) +[2026-06-23 14:01:15,530][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.015108378604054451, 'train/grad_norm_pre_clip': 7.377857685089111, 'timing/data': 0.0006077300058677793, 'timing/model': 1.9369885000633076, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 0.81}) +[2026-06-23 14:01:19,338][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.10236703604459763, 'train/grad_norm_pre_clip': 25.58610725402832, 'timing/data': 0.0007476239698007703, 'timing/model': 1.9298258230555803, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 0.81}) +[2026-06-23 14:01:23,137][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.06662154942750931, 'train/grad_norm_pre_clip': 32.07508087158203, 'timing/data': 0.0005975649692118168, 'timing/model': 1.9268528500106186, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.81}) +[2026-06-23 14:01:26,974][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.03220907971262932, 'train/grad_norm_pre_clip': 29.93951988220215, 'timing/data': 0.0006015599938109517, 'timing/model': 1.9412969229742885, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 0.81}) +[2026-06-23 14:01:30,778][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.047668423503637314, 'train/grad_norm_pre_clip': 28.566160202026367, 'timing/data': 0.0006394550437107682, 'timing/model': 1.9285454249475151, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 0.81}) +[2026-06-23 14:01:34,582][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.12748835980892181, 'train/grad_norm_pre_clip': 26.26271629333496, 'timing/data': 0.000645965919829905, 'timing/model': 1.9317538409959525, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 0.81}) +[2026-06-23 14:01:38,389][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.02133241668343544, 'train/grad_norm_pre_clip': 16.41350555419922, 'timing/data': 0.0007104889955371618, 'timing/model': 1.933987989090383, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 0.81}) +[2026-06-23 14:01:42,208][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2100 +[2026-06-23 14:03:47,377][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2100 +[2026-06-23 14:03:47,378][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2100 +[2026-06-23 14:05:10,858][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2100 +[2026-06-23 14:05:10,858][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.05003156140446663, 'train/grad_norm_pre_clip': 32.826255798339844, 'eval/action_loss/seconds': 125.16654925700277, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.043022118508815765, 'eval/latency_2/loss': 0.043022118508815765, 'eval/flappy/loss': 0.043022118508815765, 'eval/flappy/latency_2/loss': 0.043022118508815765, 'eval/flap_precision': 0.9649243466299863, 'eval/flap_recall': 0.946693657219973, 'eval/flap_f1': 0.9557220708446866, 'eval/flap_rate_model': 0.20194444444444445, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9557220708446866, 'eval/latency_2/flap_precision': 0.9649243466299863, 'eval/latency_2/flap_recall': 0.946693657219973, 'eval/latency_2/flap_f1': 0.9557220708446866, 'eval/latency_2/flap_rate_model': 0.20194444444444445, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9557220708446866, 'eval/action_classification/data_wait_seconds': 1.311061978340149, 'eval/action_classification/predict_seconds': 81.9886245727539, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.000757689937017858, 'timing/model': 1.9204463680507615, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.81}) +[2026-06-23 14:05:14,619][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.09594906121492386, 'train/grad_norm_pre_clip': 34.10555648803711, 'timing/data': 0.0006456889677792788, 'timing/model': 1.9128971169702709, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 0.81}) +[2026-06-23 14:05:18,399][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.23854759335517883, 'train/grad_norm_pre_clip': 39.40311050415039, 'timing/data': 0.0006516809808090329, 'timing/model': 1.9204563909443095, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 0.81}) +[2026-06-23 14:05:22,185][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.03805701807141304, 'train/grad_norm_pre_clip': 21.500268936157227, 'timing/data': 0.0006277529755607247, 'timing/model': 1.9216005509952083, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 0.81}) +[2026-06-23 14:05:25,977][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.0733136460185051, 'train/grad_norm_pre_clip': 32.68231964111328, 'timing/data': 0.0006641830550506711, 'timing/model': 1.9333464700030163, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 0.81}) +[2026-06-23 14:05:29,768][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.024518342688679695, 'train/grad_norm_pre_clip': 26.96854591369629, 'timing/data': 0.0006051310338079929, 'timing/model': 1.926649653003551, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 0.81}) +[2026-06-23 14:05:33,560][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.12327615916728973, 'train/grad_norm_pre_clip': 39.7224235534668, 'timing/data': 0.0006980589823797345, 'timing/model': 1.9189479190390557, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 0.82}) +[2026-06-23 14:05:37,351][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.054860424250364304, 'train/grad_norm_pre_clip': 36.121429443359375, 'timing/data': 0.0006249080179259181, 'timing/model': 1.9219983789371327, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 0.82}) +[2026-06-23 14:05:41,149][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.09156972914934158, 'train/grad_norm_pre_clip': 23.03124237060547, 'timing/data': 0.0006241940427571535, 'timing/model': 1.9145898660644889, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 0.82}) +[2026-06-23 14:05:44,952][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.1584121286869049, 'train/grad_norm_pre_clip': 34.48774337768555, 'timing/data': 0.0006124969804659486, 'timing/model': 1.9302372359670699, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.82}) +[2026-06-23 14:05:48,730][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.03178716078400612, 'train/grad_norm_pre_clip': 25.307998657226562, 'timing/data': 0.0005942659918218851, 'timing/model': 1.9130496779689565, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.82}) +[2026-06-23 14:05:52,527][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.09322474896907806, 'train/grad_norm_pre_clip': 27.817344665527344, 'timing/data': 0.0006019219290465117, 'timing/model': 1.9191799180116504, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 0.82}) +[2026-06-23 14:05:56,325][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.07960781455039978, 'train/grad_norm_pre_clip': 31.68246078491211, 'timing/data': 0.0006633989978581667, 'timing/model': 1.9243836849927902, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 0.82}) +[2026-06-23 14:06:00,122][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.1040574312210083, 'train/grad_norm_pre_clip': 37.46388244628906, 'timing/data': 0.0008000780362635851, 'timing/model': 1.9200781130930409, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 0.82}) +[2026-06-23 14:06:03,943][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.14996309578418732, 'train/grad_norm_pre_clip': 33.47567367553711, 'timing/data': 0.0008859110530465841, 'timing/model': 1.9323766470188275, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 0.82}) +[2026-06-23 14:06:07,740][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.03754560649394989, 'train/grad_norm_pre_clip': 22.805240631103516, 'timing/data': 0.0005891020409762859, 'timing/model': 1.9287725620670244, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 0.82}) +[2026-06-23 14:06:11,527][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.0847470760345459, 'train/grad_norm_pre_clip': 27.90007972717285, 'timing/data': 0.0006348589668050408, 'timing/model': 1.9273312370060012, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 0.82}) +[2026-06-23 14:06:15,503][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.20427103340625763, 'train/grad_norm_pre_clip': 20.96843719482422, 'timing/data': 0.000795696978457272, 'timing/model': 1.9286908180220053, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 0.82}) +[2026-06-23 14:06:19,295][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.16287773847579956, 'train/grad_norm_pre_clip': 63.235774993896484, 'timing/data': 0.0008731979178264737, 'timing/model': 1.9259424359770492, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 0.82}) +[2026-06-23 14:06:23,084][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.21455071866512299, 'train/grad_norm_pre_clip': 57.712318420410156, 'timing/data': 0.0007228809408843517, 'timing/model': 1.9261596859432757, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 0.82}) +[2026-06-23 14:06:26,871][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.14524583518505096, 'train/grad_norm_pre_clip': 58.88325119018555, 'timing/data': 0.0006793360225856304, 'timing/model': 1.9228813570225611, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.82}) +[2026-06-23 14:06:30,667][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.07709541916847229, 'train/grad_norm_pre_clip': 18.836334228515625, 'timing/data': 0.0006746079307049513, 'timing/model': 1.9246436670655385, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 0.82}) +[2026-06-23 14:06:34,453][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.03321422263979912, 'train/grad_norm_pre_clip': 22.092758178710938, 'timing/data': 0.000826739938929677, 'timing/model': 1.9181925259763375, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 0.82}) +[2026-06-23 14:06:38,254][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.08706878125667572, 'train/grad_norm_pre_clip': 28.542945861816406, 'timing/data': 0.0006163240177556872, 'timing/model': 1.9268874349072576, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 0.82}) +[2026-06-23 14:06:42,069][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.1489332616329193, 'train/grad_norm_pre_clip': 29.618797302246094, 'timing/data': 0.000598444021306932, 'timing/model': 1.9300294120330364, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 0.82}) +[2026-06-23 14:06:46,065][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.07146261632442474, 'train/grad_norm_pre_clip': 31.050121307373047, 'timing/data': 0.0005797339836135507, 'timing/model': 2.1239360270556062, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 0.82}) +[2026-06-23 14:06:49,867][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.11796147376298904, 'train/grad_norm_pre_clip': 33.59296798706055, 'timing/data': 0.0008401570376008749, 'timing/model': 1.9285521080018952, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 0.82}) +[2026-06-23 14:06:53,673][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.10554580390453339, 'train/grad_norm_pre_clip': 31.238435745239258, 'timing/data': 0.0006000229623168707, 'timing/model': 1.9309854679740965, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 0.82}) +[2026-06-23 14:06:57,467][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.12820987403392792, 'train/grad_norm_pre_clip': 34.6666259765625, 'timing/data': 0.0006274569313973188, 'timing/model': 1.9270282860379666, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 0.82}) +[2026-06-23 14:07:01,264][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.0778111070394516, 'train/grad_norm_pre_clip': 21.068906784057617, 'timing/data': 0.0005888650193810463, 'timing/model': 1.9305901379557326, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 0.82}) +[2026-06-23 14:07:05,056][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.06794700026512146, 'train/grad_norm_pre_clip': 11.556167602539062, 'timing/data': 0.0005986039759591222, 'timing/model': 1.9276452589547262, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.82}) +[2026-06-23 14:07:08,849][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.0553031787276268, 'train/grad_norm_pre_clip': 25.70210075378418, 'timing/data': 0.0005848589353263378, 'timing/model': 1.9285140470601618, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 0.82}) +[2026-06-23 14:07:12,675][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.11215938627719879, 'train/grad_norm_pre_clip': 15.095324516296387, 'timing/data': 0.0005995159735903144, 'timing/model': 1.9306745410431176, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 0.83}) +[2026-06-23 14:07:16,473][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.10351456701755524, 'train/grad_norm_pre_clip': 17.189355850219727, 'timing/data': 0.0006132069975137711, 'timing/model': 1.9336607520235702, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 0.83}) +[2026-06-23 14:07:20,329][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.07084465771913528, 'train/grad_norm_pre_clip': 30.402990341186523, 'timing/data': 0.0006090269889682531, 'timing/model': 1.9306978380773216, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 0.83}) +[2026-06-23 14:07:24,131][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.14835992455482483, 'train/grad_norm_pre_clip': 52.485660552978516, 'timing/data': 0.0006147889653220773, 'timing/model': 1.927457898040302, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 0.83}) +[2026-06-23 14:07:27,933][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.03417949005961418, 'train/grad_norm_pre_clip': 25.282148361206055, 'timing/data': 0.0006008010823279619, 'timing/model': 1.9249377690721303, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 0.83}) +[2026-06-23 14:07:31,734][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.08239318430423737, 'train/grad_norm_pre_clip': 18.60129165649414, 'timing/data': 0.0006303689442574978, 'timing/model': 1.9296536319889128, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 0.83}) +[2026-06-23 14:07:35,540][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.031056920066475868, 'train/grad_norm_pre_clip': 11.227944374084473, 'timing/data': 0.000606435933150351, 'timing/model': 1.9276849729940295, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 0.83}) +[2026-06-23 14:07:39,361][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.06250906735658646, 'train/grad_norm_pre_clip': 17.59160041809082, 'timing/data': 0.0006114790448918939, 'timing/model': 1.938357006991282, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 0.83}) +[2026-06-23 14:07:43,164][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.05569880083203316, 'train/grad_norm_pre_clip': 35.69281005859375, 'timing/data': 0.000611974042840302, 'timing/model': 1.9227522859582677, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.83}) +[2026-06-23 14:07:46,954][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.11105374246835709, 'train/grad_norm_pre_clip': 35.59063720703125, 'timing/data': 0.0006183049408718944, 'timing/model': 1.9199680549791083, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 0.83}) +[2026-06-23 14:07:50,775][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.02669965662062168, 'train/grad_norm_pre_clip': 13.955448150634766, 'timing/data': 0.0006064549088478088, 'timing/model': 1.9479370020562783, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 0.83}) +[2026-06-23 14:07:54,583][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.06665932387113571, 'train/grad_norm_pre_clip': 39.279178619384766, 'timing/data': 0.0006070949602872133, 'timing/model': 1.9265468979720026, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 0.83}) +[2026-06-23 14:07:58,408][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.12919728457927704, 'train/grad_norm_pre_clip': 35.19829559326172, 'timing/data': 0.0006460199365392327, 'timing/model': 1.9492586579872295, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 0.83}) +[2026-06-23 14:08:02,219][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.0589982345700264, 'train/grad_norm_pre_clip': 58.652740478515625, 'timing/data': 0.000616858946159482, 'timing/model': 1.9337523879949003, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 0.83}) +[2026-06-23 14:08:06,057][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.03587066009640694, 'train/grad_norm_pre_clip': 22.13082504272461, 'timing/data': 0.0006990740075707436, 'timing/model': 1.9596417079446837, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 0.83}) +[2026-06-23 14:08:09,861][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.1586524397134781, 'train/grad_norm_pre_clip': 21.78996467590332, 'timing/data': 0.0006122699705883861, 'timing/model': 1.930330109084025, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 0.83}) +[2026-06-23 14:08:13,679][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.014654709957540035, 'train/grad_norm_pre_clip': 18.285987854003906, 'timing/data': 0.0006051280070096254, 'timing/model': 1.9412156019825488, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 0.83}) +[2026-06-23 14:08:17,488][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.07510082423686981, 'train/grad_norm_pre_clip': 34.458988189697266, 'timing/data': 0.0006032399833202362, 'timing/model': 1.9377804310061038, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 0.83}) +[2026-06-23 14:08:21,354][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.08986185491085052, 'train/grad_norm_pre_clip': 26.478797912597656, 'timing/data': 0.0006082549225538969, 'timing/model': 1.9320203060051426, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.83}) +[2026-06-23 14:08:25,178][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.006401883903890848, 'train/grad_norm_pre_clip': 13.354680061340332, 'timing/data': 0.00062276900280267, 'timing/model': 1.9369208168936893, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 0.83}) +[2026-06-23 14:08:28,976][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.06432133913040161, 'train/grad_norm_pre_clip': 49.91351318359375, 'timing/data': 0.0006109230453148484, 'timing/model': 1.9282632289687172, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 0.83}) +[2026-06-23 14:08:32,782][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.05265863612294197, 'train/grad_norm_pre_clip': 38.463443756103516, 'timing/data': 0.0006018650019541383, 'timing/model': 1.9345498980255798, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 0.83}) +[2026-06-23 14:08:36,595][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.05486031994223595, 'train/grad_norm_pre_clip': 20.39515495300293, 'timing/data': 0.0006141390185803175, 'timing/model': 1.9401846249820665, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 0.83}) +[2026-06-23 14:08:40,399][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.04663849622011185, 'train/grad_norm_pre_clip': 27.94548225402832, 'timing/data': 0.0005944139556959271, 'timing/model': 1.929471709067002, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 0.83}) +[2026-06-23 14:08:44,210][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.012554335407912731, 'train/grad_norm_pre_clip': 11.959083557128906, 'timing/data': 0.0006873559905216098, 'timing/model': 1.934968258952722, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 0.83}) +[2026-06-23 14:08:48,015][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.054789938032627106, 'train/grad_norm_pre_clip': 42.11961364746094, 'timing/data': 0.0006971890106797218, 'timing/model': 1.9355037220520899, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 0.83}) +[2026-06-23 14:08:51,886][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.020717522129416466, 'train/grad_norm_pre_clip': 13.808653831481934, 'timing/data': 0.0006176449824124575, 'timing/model': 2.0019726819591597, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.84}) +[2026-06-23 14:08:55,699][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.06234756484627724, 'train/grad_norm_pre_clip': 22.194780349731445, 'timing/data': 0.0005993309896439314, 'timing/model': 1.9312922150129452, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 0.84}) +[2026-06-23 14:08:59,680][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.06173811852931976, 'train/grad_norm_pre_clip': 35.10894775390625, 'timing/data': 0.000629642978310585, 'timing/model': 1.9231034490512684, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.84}) +[2026-06-23 14:09:03,486][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.07135516405105591, 'train/grad_norm_pre_clip': 31.907543182373047, 'timing/data': 0.0005879730451852083, 'timing/model': 1.931367910001427, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 0.84}) +[2026-06-23 14:09:07,309][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.1505916267633438, 'train/grad_norm_pre_clip': 33.016658782958984, 'timing/data': 0.0005935869412496686, 'timing/model': 1.9415903249755502, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 0.84}) +[2026-06-23 14:09:11,115][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.1041792631149292, 'train/grad_norm_pre_clip': 26.975496292114258, 'timing/data': 0.0006273220060393214, 'timing/model': 1.926514336024411, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 0.84}) +[2026-06-23 14:09:14,924][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.11668339371681213, 'train/grad_norm_pre_clip': 28.89167022705078, 'timing/data': 0.0006682599196210504, 'timing/model': 1.9340987099567428, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 0.84}) +[2026-06-23 14:09:18,735][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.15116821229457855, 'train/grad_norm_pre_clip': 33.51526641845703, 'timing/data': 0.0006204450037330389, 'timing/model': 1.9340060229878873, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.84}) +[2026-06-23 14:09:22,530][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.03064919449388981, 'train/grad_norm_pre_clip': 31.611982345581055, 'timing/data': 0.0006010229699313641, 'timing/model': 1.929194289026782, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 0.84}) +[2026-06-23 14:09:26,335][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.1685265600681305, 'train/grad_norm_pre_clip': 34.687095642089844, 'timing/data': 0.0006261620437726378, 'timing/model': 1.931253905990161, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 0.84}) +[2026-06-23 14:09:30,238][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.10684197396039963, 'train/grad_norm_pre_clip': 33.40823745727539, 'timing/data': 0.0007996259955689311, 'timing/model': 1.998996949987486, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 0.84}) +[2026-06-23 14:09:34,041][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.06335457414388657, 'train/grad_norm_pre_clip': 29.488079071044922, 'timing/data': 0.0008362350054085255, 'timing/model': 1.9220864230301231, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 0.84}) +[2026-06-23 14:09:37,853][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.06363673508167267, 'train/grad_norm_pre_clip': 24.717782974243164, 'timing/data': 0.0005947400350123644, 'timing/model': 1.9305860209278762, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.84}) +[2026-06-23 14:09:41,686][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.13236728310585022, 'train/grad_norm_pre_clip': 45.714046478271484, 'timing/data': 0.0006077430443838239, 'timing/model': 1.9562781719723716, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 0.84}) +[2026-06-23 14:09:45,494][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.09796419739723206, 'train/grad_norm_pre_clip': 54.92094421386719, 'timing/data': 0.0006023349706083536, 'timing/model': 1.935254660085775, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 0.84}) +[2026-06-23 14:09:49,301][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.05468055605888367, 'train/grad_norm_pre_clip': 55.74044418334961, 'timing/data': 0.0008299830369651318, 'timing/model': 1.930527756921947, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 0.84}) +[2026-06-23 14:09:53,123][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.04602816328406334, 'train/grad_norm_pre_clip': 19.55446434020996, 'timing/data': 0.0006999600445851684, 'timing/model': 1.9365076560061425, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 0.84}) +[2026-06-23 14:09:56,932][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.13663174211978912, 'train/grad_norm_pre_clip': 46.05237579345703, 'timing/data': 0.0005924709839746356, 'timing/model': 1.932346921064891, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 0.84}) +[2026-06-23 14:10:00,853][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.058667488396167755, 'train/grad_norm_pre_clip': 33.88725662231445, 'timing/data': 0.0006007628981024027, 'timing/model': 1.987811911967583, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 0.84}) +[2026-06-23 14:10:04,717][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.0781986340880394, 'train/grad_norm_pre_clip': 35.284889221191406, 'timing/data': 0.0006120119942352176, 'timing/model': 1.9354073050199077, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 0.84}) +[2026-06-23 14:10:08,511][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.12040326744318008, 'train/grad_norm_pre_clip': 36.40401840209961, 'timing/data': 0.0006878250278532505, 'timing/model': 1.9304427369497716, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 0.84}) +[2026-06-23 14:10:12,316][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.13537949323654175, 'train/grad_norm_pre_clip': 56.92581558227539, 'timing/data': 0.0006120420293882489, 'timing/model': 1.9322120778961107, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 0.84}) +[2026-06-23 14:10:16,130][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.05652471259236336, 'train/grad_norm_pre_clip': 24.609134674072266, 'timing/data': 0.0005946590099483728, 'timing/model': 1.939140960923396, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.84}) +[2026-06-23 14:10:19,961][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.055981699377298355, 'train/grad_norm_pre_clip': 21.701494216918945, 'timing/data': 0.000657364958897233, 'timing/model': 1.9468491759616882, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 0.84}) +[2026-06-23 14:10:23,782][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.1163993626832962, 'train/grad_norm_pre_clip': 19.27124786376953, 'timing/data': 0.0006692130118608475, 'timing/model': 1.9407891780138016, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 0.84}) +[2026-06-23 14:10:27,576][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.060197051614522934, 'train/grad_norm_pre_clip': 33.9674072265625, 'timing/data': 0.0006147229578346014, 'timing/model': 1.9185670639853925, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 0.84}) +[2026-06-23 14:10:31,394][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.0661819651722908, 'train/grad_norm_pre_clip': 34.12714767456055, 'timing/data': 0.0007940530776977539, 'timing/model': 1.9398056030040607, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 0.85}) +[2026-06-23 14:10:35,255][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.04535241052508354, 'train/grad_norm_pre_clip': 14.158833503723145, 'timing/data': 0.000647687935270369, 'timing/model': 1.9868415950331837, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 0.85}) +[2026-06-23 14:10:39,072][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.046022891998291016, 'train/grad_norm_pre_clip': 22.165157318115234, 'timing/data': 0.0006336439400911331, 'timing/model': 1.9315816360758618, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 0.85}) +[2026-06-23 14:10:42,883][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.031209111213684082, 'train/grad_norm_pre_clip': 58.37516784667969, 'timing/data': 0.0009319439996033907, 'timing/model': 1.9279691959964111, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 0.85}) +[2026-06-23 14:10:46,702][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.06551823019981384, 'train/grad_norm_pre_clip': 31.664152145385742, 'timing/data': 0.0006730130407959223, 'timing/model': 1.9435069230385125, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 0.85}) +[2026-06-23 14:10:50,488][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.09418687224388123, 'train/grad_norm_pre_clip': 40.86993408203125, 'timing/data': 0.0006204789970070124, 'timing/model': 1.9212895369855687, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 0.85}) +[2026-06-23 14:10:54,296][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.05389091372489929, 'train/grad_norm_pre_clip': 39.6473503112793, 'timing/data': 0.0006159689510241151, 'timing/model': 1.9268681099638343, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.85}) +[2026-06-23 14:10:58,101][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.09284689277410507, 'train/grad_norm_pre_clip': 71.08617401123047, 'timing/data': 0.0006197199691087008, 'timing/model': 1.926235898048617, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 0.85}) +[2026-06-23 14:11:01,906][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.06111298128962517, 'train/grad_norm_pre_clip': 34.468563079833984, 'timing/data': 0.0006144880317151546, 'timing/model': 1.926268546958454, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 0.85}) +[2026-06-23 14:11:05,716][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.03094584494829178, 'train/grad_norm_pre_clip': 21.36770248413086, 'timing/data': 0.000776427099481225, 'timing/model': 1.932031331001781, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 0.85}) +[2026-06-23 14:11:09,536][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.19576552510261536, 'train/grad_norm_pre_clip': 53.973838806152344, 'timing/data': 0.0006255760090425611, 'timing/model': 1.9298324149567634, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 0.85}) +[2026-06-23 14:11:13,346][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.051480259746313095, 'train/grad_norm_pre_clip': 24.011171340942383, 'timing/data': 0.000600422965362668, 'timing/model': 1.9246186350937933, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 0.85}) +[2026-06-23 14:11:17,153][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.05194288492202759, 'train/grad_norm_pre_clip': 30.1311092376709, 'timing/data': 0.0006505839992314577, 'timing/model': 1.9285709770629182, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 0.85}) +[2026-06-23 14:11:20,958][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.1404857039451599, 'train/grad_norm_pre_clip': 44.22223663330078, 'timing/data': 0.0006048319628462195, 'timing/model': 1.9275203349534422, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 0.85}) +[2026-06-23 14:11:24,771][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.11593826860189438, 'train/grad_norm_pre_clip': 27.26685333251953, 'timing/data': 0.0006064740009605885, 'timing/model': 1.9383959249826148, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 0.85}) +[2026-06-23 14:11:28,574][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.10052728652954102, 'train/grad_norm_pre_clip': 22.92693328857422, 'timing/data': 0.0006077559664845467, 'timing/model': 1.9276468489551917, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 0.85}) +[2026-06-23 14:11:32,394][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2200 +[2026-06-23 14:13:39,853][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2200 +[2026-06-23 14:13:39,854][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2200 +[2026-06-23 14:15:03,617][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2200 +[2026-06-23 14:15:03,618][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.09945783764123917, 'train/grad_norm_pre_clip': 65.52645111083984, 'eval/action_loss/seconds': 127.45752425200772, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.052595850080251694, 'eval/latency_2/loss': 0.052595850080251694, 'eval/flappy/loss': 0.052595850080251694, 'eval/flappy/latency_2/loss': 0.052595850080251694, 'eval/flap_precision': 0.9826212889210717, 'eval/flap_recall': 0.9156545209176788, 'eval/flap_f1': 0.9479566887879847, 'eval/flap_rate_model': 0.19180555555555556, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9479566887879847, 'eval/latency_2/flap_precision': 0.9826212889210717, 'eval/latency_2/flap_recall': 0.9156545209176788, 'eval/latency_2/flap_f1': 0.9479566887879847, 'eval/latency_2/flap_rate_model': 0.19180555555555556, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9479566887879847, 'eval/action_classification/data_wait_seconds': 1.306728482246399, 'eval/action_classification/predict_seconds': 82.27059936523438, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.000621194951236248, 'timing/model': 1.9404792739078403, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.85}) +[2026-06-23 14:15:07,378][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.046455059200525284, 'train/grad_norm_pre_clip': 24.161638259887695, 'timing/data': 0.0006733299233019352, 'timing/model': 1.9094804619671777, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 0.85}) +[2026-06-23 14:15:11,172][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.0881899893283844, 'train/grad_norm_pre_clip': 50.079959869384766, 'timing/data': 0.0006104910280555487, 'timing/model': 1.9223097719950601, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 0.85}) +[2026-06-23 14:15:14,965][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.040753137320280075, 'train/grad_norm_pre_clip': 56.5989990234375, 'timing/data': 0.0005908990278840065, 'timing/model': 1.9247081551002339, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 0.85}) +[2026-06-23 14:15:18,742][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.04371695965528488, 'train/grad_norm_pre_clip': 51.94814682006836, 'timing/data': 0.0005960880080237985, 'timing/model': 1.919745650026016, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 0.85}) +[2026-06-23 14:15:22,527][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.09915924072265625, 'train/grad_norm_pre_clip': 26.330345153808594, 'timing/data': 0.0006195529131218791, 'timing/model': 1.925797580042854, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 0.85}) +[2026-06-23 14:15:26,327][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.13728077709674835, 'train/grad_norm_pre_clip': 38.57057571411133, 'timing/data': 0.0009079789742827415, 'timing/model': 1.9264401559485123, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 0.85}) +[2026-06-23 14:15:30,289][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.0241412203758955, 'train/grad_norm_pre_clip': 17.12694549560547, 'timing/data': 0.0006340179825201631, 'timing/model': 2.1073979310458526, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.85}) +[2026-06-23 14:15:34,082][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.04563554748892784, 'train/grad_norm_pre_clip': 14.212395668029785, 'timing/data': 0.0005847839638590813, 'timing/model': 1.9232683560112491, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 0.85}) +[2026-06-23 14:15:37,878][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.02177427150309086, 'train/grad_norm_pre_clip': 15.170854568481445, 'timing/data': 0.0006159689510241151, 'timing/model': 1.9242561699356884, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 0.85}) +[2026-06-23 14:15:41,676][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.03649620711803436, 'train/grad_norm_pre_clip': 23.401386260986328, 'timing/data': 0.0005980880232527852, 'timing/model': 1.9288208660436794, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.86}) +[2026-06-23 14:15:45,468][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.030175505205988884, 'train/grad_norm_pre_clip': 32.53022384643555, 'timing/data': 0.000615074997767806, 'timing/model': 1.9197435090318322, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 0.86}) +[2026-06-23 14:15:49,256][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.09094548225402832, 'train/grad_norm_pre_clip': 24.007627487182617, 'timing/data': 0.0005768389673903584, 'timing/model': 1.9189482959918678, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 0.86}) +[2026-06-23 14:15:53,057][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.08771704137325287, 'train/grad_norm_pre_clip': 49.27750015258789, 'timing/data': 0.000628227018751204, 'timing/model': 1.9306355469161645, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 0.86}) +[2026-06-23 14:15:56,859][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.0980810895562172, 'train/grad_norm_pre_clip': 64.44259643554688, 'timing/data': 0.0007544630207121372, 'timing/model': 1.9229725960176438, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 0.86}) +[2026-06-23 14:16:00,653][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.03740973025560379, 'train/grad_norm_pre_clip': 39.417720794677734, 'timing/data': 0.0006740179378539324, 'timing/model': 1.9334789370186627, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 0.86}) +[2026-06-23 14:16:04,618][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.08134225010871887, 'train/grad_norm_pre_clip': 46.07597351074219, 'timing/data': 0.0006036399863660336, 'timing/model': 1.9176638899371028, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 0.86}) +[2026-06-23 14:16:08,420][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.11179596930742264, 'train/grad_norm_pre_clip': 28.90437126159668, 'timing/data': 0.0005901569966226816, 'timing/model': 1.9257695419946685, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 0.86}) +[2026-06-23 14:16:12,236][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.17255960404872894, 'train/grad_norm_pre_clip': 40.28865432739258, 'timing/data': 0.0010159079683944583, 'timing/model': 1.9413294289261103, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 0.86}) +[2026-06-23 14:16:16,038][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.12172593921422958, 'train/grad_norm_pre_clip': 35.229278564453125, 'timing/data': 0.0009085920173674822, 'timing/model': 1.9333778710570186, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 0.86}) +[2026-06-23 14:16:19,837][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.11373031884431839, 'train/grad_norm_pre_clip': 104.5648422241211, 'timing/data': 0.0005861240206286311, 'timing/model': 1.9234409490600228, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.86}) +[2026-06-23 14:16:23,641][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.10631856322288513, 'train/grad_norm_pre_clip': 53.80110168457031, 'timing/data': 0.0006057480350136757, 'timing/model': 1.9241590099409223, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 0.86}) +[2026-06-23 14:16:27,436][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.17508965730667114, 'train/grad_norm_pre_clip': 100.2563705444336, 'timing/data': 0.0008421600796282291, 'timing/model': 1.927414545090869, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 0.86}) +[2026-06-23 14:16:31,223][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.0709543228149414, 'train/grad_norm_pre_clip': 48.1120719909668, 'timing/data': 0.0007322380552068353, 'timing/model': 1.9178560249274597, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 0.86}) +[2026-06-23 14:16:35,136][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.08730998635292053, 'train/grad_norm_pre_clip': 40.365257263183594, 'timing/data': 0.0006539589958265424, 'timing/model': 2.0463481749175116, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 0.86}) +[2026-06-23 14:16:38,923][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.023790666833519936, 'train/grad_norm_pre_clip': 26.844627380371094, 'timing/data': 0.0006095339776948094, 'timing/model': 1.9176589080598205, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 0.86}) +[2026-06-23 14:16:42,717][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.07890509814023972, 'train/grad_norm_pre_clip': 38.71504211425781, 'timing/data': 0.0006426370237022638, 'timing/model': 1.9293191789183766, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 0.86}) +[2026-06-23 14:16:46,516][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.030262604355812073, 'train/grad_norm_pre_clip': 32.24606704711914, 'timing/data': 0.0006194449961185455, 'timing/model': 1.9321221279678866, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 0.86}) +[2026-06-23 14:16:50,317][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.05964849144220352, 'train/grad_norm_pre_clip': 55.60551071166992, 'timing/data': 0.00057752407155931, 'timing/model': 1.9345780689036474, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 0.86}) +[2026-06-23 14:16:54,125][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.04167360067367554, 'train/grad_norm_pre_clip': 17.25861930847168, 'timing/data': 0.0006273370236158371, 'timing/model': 1.9289552010595798, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 0.86}) +[2026-06-23 14:16:57,926][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.09654716402292252, 'train/grad_norm_pre_clip': 73.7480239868164, 'timing/data': 0.0006015080725774169, 'timing/model': 1.9339580640662462, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.86}) +[2026-06-23 14:17:01,729][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.0272294320166111, 'train/grad_norm_pre_clip': 135.30703735351562, 'timing/data': 0.0006369560724124312, 'timing/model': 1.9331835129996762, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 0.86}) +[2026-06-23 14:17:05,529][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.17444048821926117, 'train/grad_norm_pre_clip': 39.69519805908203, 'timing/data': 0.0006236929912120104, 'timing/model': 1.9279269248945639, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 0.86}) +[2026-06-23 14:17:09,462][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.08453962206840515, 'train/grad_norm_pre_clip': 23.44622039794922, 'timing/data': 0.0006051979726180434, 'timing/model': 1.93496262608096, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 0.86}) +[2026-06-23 14:17:13,254][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.08031497150659561, 'train/grad_norm_pre_clip': 43.78193283081055, 'timing/data': 0.0006048850482329726, 'timing/model': 1.9214195189997554, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 0.86}) +[2026-06-23 14:17:17,060][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.09925054758787155, 'train/grad_norm_pre_clip': 34.4224967956543, 'timing/data': 0.0006605469388887286, 'timing/model': 1.9321393800200894, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.86}) +[2026-06-23 14:17:20,852][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.14330625534057617, 'train/grad_norm_pre_clip': 37.713829040527344, 'timing/data': 0.000593555043451488, 'timing/model': 1.9248400409705937, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 0.87}) +[2026-06-23 14:17:24,672][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.23975446820259094, 'train/grad_norm_pre_clip': 48.52446746826172, 'timing/data': 0.000592924072407186, 'timing/model': 1.9334498420357704, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 0.87}) +[2026-06-23 14:17:28,478][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.038060300052165985, 'train/grad_norm_pre_clip': 34.72551727294922, 'timing/data': 0.0006001170258969069, 'timing/model': 1.9270446939626709, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 0.87}) +[2026-06-23 14:17:32,290][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.043791212141513824, 'train/grad_norm_pre_clip': 28.1423282623291, 'timing/data': 0.0006793050561100245, 'timing/model': 1.937109079095535, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 0.87}) +[2026-06-23 14:17:36,093][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.10054130852222443, 'train/grad_norm_pre_clip': 31.412382125854492, 'timing/data': 0.0005875119240954518, 'timing/model': 1.9313634930877015, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.87}) +[2026-06-23 14:17:39,984][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.0851205512881279, 'train/grad_norm_pre_clip': 28.184980392456055, 'timing/data': 0.0008329490665346384, 'timing/model': 2.0193857919657603, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 0.87}) +[2026-06-23 14:17:43,777][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.12794682383537292, 'train/grad_norm_pre_clip': 30.059123992919922, 'timing/data': 0.0006431519286707044, 'timing/model': 1.9210533580044284, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 0.87}) +[2026-06-23 14:17:47,597][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.0866166427731514, 'train/grad_norm_pre_clip': 20.532726287841797, 'timing/data': 0.0006624459056183696, 'timing/model': 1.9283300200477242, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 0.87}) +[2026-06-23 14:17:51,466][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.04776241257786751, 'train/grad_norm_pre_clip': 16.44322967529297, 'timing/data': 0.0006404970772564411, 'timing/model': 1.9199476250214502, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 0.87}) +[2026-06-23 14:17:55,270][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.18140819668769836, 'train/grad_norm_pre_clip': 33.60028839111328, 'timing/data': 0.0008077939273789525, 'timing/model': 1.9385051920544356, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 0.87}) +[2026-06-23 14:17:59,082][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.028521429747343063, 'train/grad_norm_pre_clip': 22.58233642578125, 'timing/data': 0.0006378950783982873, 'timing/model': 1.930087812943384, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 0.87}) +[2026-06-23 14:18:02,893][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.08266670256853104, 'train/grad_norm_pre_clip': 33.989036560058594, 'timing/data': 0.0006492090178653598, 'timing/model': 1.9364616679959, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 0.87}) +[2026-06-23 14:18:06,690][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.08113864809274673, 'train/grad_norm_pre_clip': 32.61830139160156, 'timing/data': 0.0006413550581783056, 'timing/model': 1.9322103490121663, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 0.87}) +[2026-06-23 14:18:10,505][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.049063801765441895, 'train/grad_norm_pre_clip': 19.788427352905273, 'timing/data': 0.0006350880721583962, 'timing/model': 1.9290264300070703, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 0.87}) +[2026-06-23 14:24:12,260][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2250 +[2026-06-23 14:24:12,261][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.10367479175329208, 'train/grad_norm_pre_clip': 21.22691535949707, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 219.8400000000019, 'rl_games_eval/mid_train/mean_length': 1791.4, 'rl_games_eval/mid_train/std_reward': 122.89391522773798, 'rl_games_eval/mid_train/std_length': 989.3716389709177, 'rl_games_eval/mid_train/macro_mean_reward': 219.8400000000019, 'rl_games_eval/mid_train/macro_mean_length': 1791.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 219.8400000000019, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1791.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 122.89391522773798, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 989.3716389709177, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 219.8400000000019, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006287889555096626, 'timing/model': 1.9203141819452867, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.87}) +[2026-06-23 14:24:15,983][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.05281442776322365, 'train/grad_norm_pre_clip': 21.060405731201172, 'timing/data': 0.000605833949521184, 'timing/model': 1.891369041055441, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 0.87}) +[2026-06-23 14:24:19,711][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.19704702496528625, 'train/grad_norm_pre_clip': 33.64845657348633, 'timing/data': 0.000650492962449789, 'timing/model': 1.8939330329885706, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 0.87}) +[2026-06-23 14:24:23,447][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.0977935791015625, 'train/grad_norm_pre_clip': 52.758487701416016, 'timing/data': 0.0006108559900894761, 'timing/model': 1.8890408159932122, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 0.87}) +[2026-06-23 14:24:27,193][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.11977196484804153, 'train/grad_norm_pre_clip': 25.35492515563965, 'timing/data': 0.0005936619127169251, 'timing/model': 1.9001874139066786, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 0.87}) +[2026-06-23 14:24:30,953][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.06267897039651871, 'train/grad_norm_pre_clip': 29.0902042388916, 'timing/data': 0.000599829014390707, 'timing/model': 1.90943886898458, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 0.87}) +[2026-06-23 14:24:34,703][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.0849582776427269, 'train/grad_norm_pre_clip': 21.22188949584961, 'timing/data': 0.0006181540666148067, 'timing/model': 1.9041212430456653, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.87}) +[2026-06-23 14:24:38,460][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.09516406804323196, 'train/grad_norm_pre_clip': 23.872770309448242, 'timing/data': 0.0009119299938902259, 'timing/model': 1.89925381494686, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 0.87}) +[2026-06-23 14:24:42,284][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.0732264369726181, 'train/grad_norm_pre_clip': 24.3394718170166, 'timing/data': 0.0006431699730455875, 'timing/model': 1.9677851279266179, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 0.87}) +[2026-06-23 14:24:46,034][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.052159909158945084, 'train/grad_norm_pre_clip': 13.806835174560547, 'timing/data': 0.0005965130403637886, 'timing/model': 1.9021542889531702, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 0.87}) +[2026-06-23 14:24:49,814][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.05688878893852234, 'train/grad_norm_pre_clip': 34.077110290527344, 'timing/data': 0.0006134409923106432, 'timing/model': 1.9191476890118793, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.87}) +[2026-06-23 14:24:53,601][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.014697565697133541, 'train/grad_norm_pre_clip': 37.459205627441406, 'timing/data': 0.0006276710191741586, 'timing/model': 1.9244817349826917, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 0.88}) +[2026-06-23 14:24:57,403][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.033539656549692154, 'train/grad_norm_pre_clip': 12.749655723571777, 'timing/data': 0.000700408942066133, 'timing/model': 1.9138087020255625, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 0.88}) +[2026-06-23 14:25:01,157][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.11605003476142883, 'train/grad_norm_pre_clip': 46.05075454711914, 'timing/data': 0.0006000850116834044, 'timing/model': 1.9049016480566934, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 0.88}) +[2026-06-23 14:25:04,927][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.02734442614018917, 'train/grad_norm_pre_clip': 28.01451301574707, 'timing/data': 0.0006604730151593685, 'timing/model': 1.9149430410470814, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 0.88}) +[2026-06-23 14:25:08,715][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.06926834583282471, 'train/grad_norm_pre_clip': 33.87428283691406, 'timing/data': 0.000608155969530344, 'timing/model': 1.9182499339804053, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 0.88}) +[2026-06-23 14:25:12,498][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.06233128532767296, 'train/grad_norm_pre_clip': 20.290925979614258, 'timing/data': 0.000727347913198173, 'timing/model': 1.9205013399478048, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 0.88}) +[2026-06-23 14:25:16,392][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.07406982779502869, 'train/grad_norm_pre_clip': 54.76317596435547, 'timing/data': 0.0007886049570515752, 'timing/model': 2.0203976939665154, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 0.88}) +[2026-06-23 14:25:20,170][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.04812419041991234, 'train/grad_norm_pre_clip': 24.562145233154297, 'timing/data': 0.0005974629893898964, 'timing/model': 1.91738036100287, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 0.88}) +[2026-06-23 14:25:23,979][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.07317747920751572, 'train/grad_norm_pre_clip': 27.483905792236328, 'timing/data': 0.0006140340119600296, 'timing/model': 1.921118368045427, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 0.88}) +[2026-06-23 14:25:27,764][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.10001528263092041, 'train/grad_norm_pre_clip': 48.32036209106445, 'timing/data': 0.0006029660580679774, 'timing/model': 1.9229463380761445, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.88}) +[2026-06-23 14:25:31,546][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.08019928634166718, 'train/grad_norm_pre_clip': 38.060237884521484, 'timing/data': 0.0006088459631428123, 'timing/model': 1.925012714928016, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 0.88}) +[2026-06-23 14:25:35,329][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.03852774202823639, 'train/grad_norm_pre_clip': 27.123769760131836, 'timing/data': 0.0005941970739513636, 'timing/model': 1.921925044967793, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 0.88}) +[2026-06-23 14:25:39,121][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.13162285089492798, 'train/grad_norm_pre_clip': 26.821577072143555, 'timing/data': 0.0005925570148974657, 'timing/model': 1.9194690820295364, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 0.88}) +[2026-06-23 14:25:42,914][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.026087624952197075, 'train/grad_norm_pre_clip': 30.2442569732666, 'timing/data': 0.0005975859239697456, 'timing/model': 1.922695118933916, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 0.88}) +[2026-06-23 14:25:46,709][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.08667981624603271, 'train/grad_norm_pre_clip': 49.98146438598633, 'timing/data': 0.0006494859699159861, 'timing/model': 1.9226125729037449, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 0.88}) +[2026-06-23 14:25:50,688][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.072772316634655, 'train/grad_norm_pre_clip': 28.852554321289062, 'timing/data': 0.0005962259601801634, 'timing/model': 2.108429631916806, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 0.88}) +[2026-06-23 14:25:54,484][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.15958364307880402, 'train/grad_norm_pre_clip': 32.443458557128906, 'timing/data': 0.0006138830212876201, 'timing/model': 1.9263019879581407, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 0.88}) +[2026-06-23 14:25:58,277][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.08376286923885345, 'train/grad_norm_pre_clip': 31.10725975036621, 'timing/data': 0.000648070010356605, 'timing/model': 1.9252738120267168, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 0.88}) +[2026-06-23 14:26:02,070][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.07672189176082611, 'train/grad_norm_pre_clip': 25.183515548706055, 'timing/data': 0.0008613599929958582, 'timing/model': 1.9270160529995337, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 0.88}) +[2026-06-23 14:26:05,853][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.024306749925017357, 'train/grad_norm_pre_clip': 15.537156105041504, 'timing/data': 0.000778396031819284, 'timing/model': 1.915295536047779, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.88}) +[2026-06-23 14:26:09,647][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.01175619661808014, 'train/grad_norm_pre_clip': 18.946334838867188, 'timing/data': 0.0006146859377622604, 'timing/model': 1.9188844879390672, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 0.88}) +[2026-06-23 14:26:13,446][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.04101487994194031, 'train/grad_norm_pre_clip': 31.769189834594727, 'timing/data': 0.0006197370821610093, 'timing/model': 1.9200791710754856, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 0.88}) +[2026-06-23 14:26:17,247][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.05856148898601532, 'train/grad_norm_pre_clip': 55.26567077636719, 'timing/data': 0.0007214149227365851, 'timing/model': 1.9308227800065652, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 0.88}) +[2026-06-23 14:26:21,106][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.11354365944862366, 'train/grad_norm_pre_clip': 48.05852127075195, 'timing/data': 0.0007438199827447534, 'timing/model': 1.9879731700057164, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 0.88}) +[2026-06-23 14:26:24,903][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.033111199736595154, 'train/grad_norm_pre_clip': 45.433406829833984, 'timing/data': 0.000598252983763814, 'timing/model': 1.927725965040736, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 0.88}) +[2026-06-23 14:26:28,700][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.025308428332209587, 'train/grad_norm_pre_clip': 29.30399513244629, 'timing/data': 0.0006782930577173829, 'timing/model': 1.929956546984613, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 0.88}) +[2026-06-23 14:26:32,487][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.13364805281162262, 'train/grad_norm_pre_clip': 22.726696014404297, 'timing/data': 0.0008084509754553437, 'timing/model': 1.9165191990323365, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 0.89}) +[2026-06-23 14:26:36,269][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.0573665089905262, 'train/grad_norm_pre_clip': 26.117889404296875, 'timing/data': 0.0006245769327506423, 'timing/model': 1.9168737879954278, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 0.89}) +[2026-06-23 14:26:40,076][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.04988488927483559, 'train/grad_norm_pre_clip': 37.967498779296875, 'timing/data': 0.0007587349973618984, 'timing/model': 1.9329104200005531, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 0.89}) +[2026-06-23 14:26:43,879][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.020063038915395737, 'train/grad_norm_pre_clip': 19.362764358520508, 'timing/data': 0.000611822004429996, 'timing/model': 1.918355647008866, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.89}) +[2026-06-23 14:26:47,688][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.08387306332588196, 'train/grad_norm_pre_clip': 56.684696197509766, 'timing/data': 0.0006001190049573779, 'timing/model': 1.9392982419813052, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 0.89}) +[2026-06-23 14:26:51,495][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.05819233879446983, 'train/grad_norm_pre_clip': 50.505802154541016, 'timing/data': 0.0006330499891191721, 'timing/model': 1.9418890840606764, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 0.89}) +[2026-06-23 14:26:55,420][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.02528042532503605, 'train/grad_norm_pre_clip': 20.12495231628418, 'timing/data': 0.0005917759845033288, 'timing/model': 2.0460033719427884, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 0.89}) +[2026-06-23 14:26:59,221][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.0937354564666748, 'train/grad_norm_pre_clip': 40.89237594604492, 'timing/data': 0.0006770821055397391, 'timing/model': 1.9286882410524413, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 0.89}) +[2026-06-23 14:27:03,027][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.17960810661315918, 'train/grad_norm_pre_clip': 63.456871032714844, 'timing/data': 0.0007308860076591372, 'timing/model': 1.9341939929872751, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 0.89}) +[2026-06-23 14:27:06,830][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.09761668741703033, 'train/grad_norm_pre_clip': 69.76444244384766, 'timing/data': 0.0005932219792157412, 'timing/model': 1.9223000960191712, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 0.89}) +[2026-06-23 14:27:10,640][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.10038520395755768, 'train/grad_norm_pre_clip': 54.543365478515625, 'timing/data': 0.0005982089787721634, 'timing/model': 1.9325492220232263, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 0.89}) +[2026-06-23 14:27:14,438][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.060021448880434036, 'train/grad_norm_pre_clip': 36.868019104003906, 'timing/data': 0.0006328889867290854, 'timing/model': 1.92179162602406, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 0.89}) +[2026-06-23 14:27:18,249][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.037246763706207275, 'train/grad_norm_pre_clip': 32.220943450927734, 'timing/data': 0.0006149350665509701, 'timing/model': 1.9320841919397935, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 0.89}) +[2026-06-23 14:27:22,038][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2300 +[2026-06-23 14:29:29,617][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2300 +[2026-06-23 14:29:29,618][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2300 +[2026-06-23 14:30:53,047][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2300 +[2026-06-23 14:30:53,048][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.05602268874645233, 'train/grad_norm_pre_clip': 20.58481216430664, 'eval/action_loss/seconds': 127.57634264905937, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04832352325320244, 'eval/latency_2/loss': 0.04832352325320244, 'eval/flappy/loss': 0.04832352325320244, 'eval/flappy/latency_2/loss': 0.04832352325320244, 'eval/flap_precision': 0.9168765743073047, 'eval/flap_recall': 0.9824561403508771, 'eval/flap_f1': 0.9485342019543974, 'eval/flap_rate_model': 0.22055555555555556, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9485342019543974, 'eval/latency_2/flap_precision': 0.9168765743073047, 'eval/latency_2/flap_recall': 0.9824561403508771, 'eval/latency_2/flap_f1': 0.9485342019543974, 'eval/latency_2/flap_rate_model': 0.22055555555555556, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9485342019543974, 'eval/action_classification/data_wait_seconds': 1.405833125114441, 'eval/action_classification/predict_seconds': 81.84378814697266, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.000600716914050281, 'timing/model': 1.9218813319457695, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.89}) +[2026-06-23 14:30:56,806][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.21456824243068695, 'train/grad_norm_pre_clip': 65.14591979980469, 'timing/data': 0.0006074629491195083, 'timing/model': 1.90872113709338, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 0.89}) +[2026-06-23 14:31:00,584][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.16363196074962616, 'train/grad_norm_pre_clip': 60.22958755493164, 'timing/data': 0.0005908689927309752, 'timing/model': 1.9214323279447854, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 0.89}) +[2026-06-23 14:31:04,365][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.026991702616214752, 'train/grad_norm_pre_clip': 33.382694244384766, 'timing/data': 0.000584468012675643, 'timing/model': 1.9284320120932534, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 0.89}) +[2026-06-23 14:31:08,150][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.11023106426000595, 'train/grad_norm_pre_clip': 37.6622200012207, 'timing/data': 0.0005864320555701852, 'timing/model': 1.9227424920536578, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 0.89}) +[2026-06-23 14:31:11,931][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.08279155939817429, 'train/grad_norm_pre_clip': 35.63520431518555, 'timing/data': 0.0007560820085927844, 'timing/model': 1.9170826420886442, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.89}) +[2026-06-23 14:31:15,771][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.079237200319767, 'train/grad_norm_pre_clip': 22.757543563842773, 'timing/data': 0.0006129230605438352, 'timing/model': 1.9167921219486743, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 0.89}) +[2026-06-23 14:31:19,578][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.036033280193805695, 'train/grad_norm_pre_clip': 13.430113792419434, 'timing/data': 0.000585737987421453, 'timing/model': 1.9248061019461602, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 0.89}) +[2026-06-23 14:31:23,354][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.1945788711309433, 'train/grad_norm_pre_clip': 35.403778076171875, 'timing/data': 0.0006241929950192571, 'timing/model': 1.9148234279127792, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 0.89}) +[2026-06-23 14:31:27,135][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.10015217959880829, 'train/grad_norm_pre_clip': 41.17772674560547, 'timing/data': 0.0009224939858540893, 'timing/model': 1.921865899930708, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 0.89}) +[2026-06-23 14:31:30,923][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.12216866761445999, 'train/grad_norm_pre_clip': 80.81356048583984, 'timing/data': 0.00060468097217381, 'timing/model': 1.921018742956221, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.89}) +[2026-06-23 14:31:34,720][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.05334041267633438, 'train/grad_norm_pre_clip': 14.50640869140625, 'timing/data': 0.000615112017840147, 'timing/model': 1.9234167899703607, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 0.89}) +[2026-06-23 14:31:38,507][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.1962064653635025, 'train/grad_norm_pre_clip': 63.4058837890625, 'timing/data': 0.0008661759784445167, 'timing/model': 1.923039254033938, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 0.89}) +[2026-06-23 14:31:42,307][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.09796351194381714, 'train/grad_norm_pre_clip': 37.105323791503906, 'timing/data': 0.0006787519669160247, 'timing/model': 1.9298771420726553, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 0.9}) +[2026-06-23 14:31:46,153][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.05770013481378555, 'train/grad_norm_pre_clip': 17.130218505859375, 'timing/data': 0.0008047210285440087, 'timing/model': 1.9702176729915664, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 0.9}) +[2026-06-23 14:31:49,968][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.11876720935106277, 'train/grad_norm_pre_clip': 36.616455078125, 'timing/data': 0.0007527809357270598, 'timing/model': 1.93309827207122, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 0.9}) +[2026-06-23 14:31:53,757][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.0807160809636116, 'train/grad_norm_pre_clip': 51.25251007080078, 'timing/data': 0.000600236002355814, 'timing/model': 1.9280519150197506, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 0.9}) +[2026-06-23 14:31:57,571][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.09043296426534653, 'train/grad_norm_pre_clip': 34.16059112548828, 'timing/data': 0.0006024319445714355, 'timing/model': 1.937712004990317, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 0.9}) +[2026-06-23 14:32:01,369][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.13435864448547363, 'train/grad_norm_pre_clip': 28.895063400268555, 'timing/data': 0.0007493230514228344, 'timing/model': 1.9289737510262057, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 0.9}) +[2026-06-23 14:32:05,175][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.04841245338320732, 'train/grad_norm_pre_clip': 26.928266525268555, 'timing/data': 0.0006296059582382441, 'timing/model': 1.9348136309999973, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 0.9}) +[2026-06-23 14:32:08,976][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.033812955021858215, 'train/grad_norm_pre_clip': 36.83053970336914, 'timing/data': 0.0006055920384824276, 'timing/model': 1.925975868012756, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.9}) +[2026-06-23 14:32:12,782][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.07889185100793839, 'train/grad_norm_pre_clip': 27.569503784179688, 'timing/data': 0.0008460480021312833, 'timing/model': 1.9318296640412882, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 0.9}) +[2026-06-23 14:32:16,593][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.12342185527086258, 'train/grad_norm_pre_clip': 60.35973358154297, 'timing/data': 0.000719579984433949, 'timing/model': 1.9290403559571132, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 0.9}) +[2026-06-23 14:32:20,510][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.03936750069260597, 'train/grad_norm_pre_clip': 39.115169525146484, 'timing/data': 0.000816666055470705, 'timing/model': 2.0507634810637683, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 0.9}) +[2026-06-23 14:32:24,358][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.07017491012811661, 'train/grad_norm_pre_clip': 29.692354202270508, 'timing/data': 0.0005841179518029094, 'timing/model': 1.9446772249648347, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 0.9}) +[2026-06-23 14:32:28,182][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.06823158264160156, 'train/grad_norm_pre_clip': 27.632911682128906, 'timing/data': 0.0006409690249711275, 'timing/model': 1.9288834279868752, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 0.9}) +[2026-06-23 14:32:32,001][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.14001423120498657, 'train/grad_norm_pre_clip': 32.81210708618164, 'timing/data': 0.0006731280591338873, 'timing/model': 1.9461759310215712, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 0.9}) +[2026-06-23 14:32:35,804][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.11791762709617615, 'train/grad_norm_pre_clip': 23.16584587097168, 'timing/data': 0.0006159489275887609, 'timing/model': 1.9323564810911193, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 0.9}) +[2026-06-23 14:32:39,592][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.0773395448923111, 'train/grad_norm_pre_clip': 12.295273780822754, 'timing/data': 0.0006088929949328303, 'timing/model': 1.9239339899504557, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 0.9}) +[2026-06-23 14:32:43,407][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.05679652839899063, 'train/grad_norm_pre_clip': 16.07150650024414, 'timing/data': 0.0006199670024216175, 'timing/model': 1.9355242820456624, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 0.9}) +[2026-06-23 14:32:47,215][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.11204942315816879, 'train/grad_norm_pre_clip': 31.48712921142578, 'timing/data': 0.0006016340339556336, 'timing/model': 1.9351758059347048, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.9}) +[2026-06-23 14:32:51,017][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.05964730679988861, 'train/grad_norm_pre_clip': 22.917221069335938, 'timing/data': 0.0006169370608404279, 'timing/model': 1.9277950329706073, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 0.9}) +[2026-06-23 14:32:54,991][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.10877792537212372, 'train/grad_norm_pre_clip': 22.731292724609375, 'timing/data': 0.000598024926148355, 'timing/model': 2.0919625640381128, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 0.9}) +[2026-06-23 14:32:58,786][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.04388304799795151, 'train/grad_norm_pre_clip': 27.51979637145996, 'timing/data': 0.0006024659378454089, 'timing/model': 1.924032371956855, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 0.9}) +[2026-06-23 14:33:02,575][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.2180119901895523, 'train/grad_norm_pre_clip': 74.01445770263672, 'timing/data': 0.0006144919898360968, 'timing/model': 1.9271373110823333, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 0.9}) +[2026-06-23 14:33:06,369][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.06410621851682663, 'train/grad_norm_pre_clip': 35.96665573120117, 'timing/data': 0.0005942150019109249, 'timing/model': 1.9309796349843964, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 0.9}) +[2026-06-23 14:33:10,159][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.03798562288284302, 'train/grad_norm_pre_clip': 45.7352294921875, 'timing/data': 0.000601166975684464, 'timing/model': 1.9206312439637259, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 0.9}) +[2026-06-23 14:33:13,962][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.10661746561527252, 'train/grad_norm_pre_clip': 28.485368728637695, 'timing/data': 0.0007334959227591753, 'timing/model': 1.9256991629954427, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 0.9}) +[2026-06-23 14:33:17,772][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.00569158885627985, 'train/grad_norm_pre_clip': 19.30296516418457, 'timing/data': 0.000641072983853519, 'timing/model': 1.9214052699971944, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 0.9}) +[2026-06-23 14:33:21,584][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.056837618350982666, 'train/grad_norm_pre_clip': 26.596832275390625, 'timing/data': 0.0007723720045760274, 'timing/model': 1.932780978968367, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 0.91}) +[2026-06-23 14:33:25,408][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.07296808063983917, 'train/grad_norm_pre_clip': 39.387420654296875, 'timing/data': 0.0006031440570950508, 'timing/model': 1.954363107914105, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.91}) +[2026-06-23 14:33:29,211][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.066824771463871, 'train/grad_norm_pre_clip': 38.87619400024414, 'timing/data': 0.0006033359095454216, 'timing/model': 1.9228055030107498, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 0.91}) +[2026-06-23 14:33:33,018][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.05223516374826431, 'train/grad_norm_pre_clip': 22.477317810058594, 'timing/data': 0.0006609400734305382, 'timing/model': 1.9298193069407716, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 0.91}) +[2026-06-23 14:33:36,827][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.010898706503212452, 'train/grad_norm_pre_clip': 28.889362335205078, 'timing/data': 0.0006055410485714674, 'timing/model': 1.9229740389855579, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 0.91}) +[2026-06-23 14:33:40,639][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.05430055409669876, 'train/grad_norm_pre_clip': 23.820716857910156, 'timing/data': 0.0006142599740996957, 'timing/model': 1.9344402020797133, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 0.91}) +[2026-06-23 14:33:44,441][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.03801409900188446, 'train/grad_norm_pre_clip': 35.13083267211914, 'timing/data': 0.0005894519854336977, 'timing/model': 1.929694810998626, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 0.91}) +[2026-06-23 14:33:48,252][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.10723567008972168, 'train/grad_norm_pre_clip': 32.6434440612793, 'timing/data': 0.0005920700496062636, 'timing/model': 1.9375869589857757, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 0.91}) +[2026-06-23 14:33:52,058][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.04723073169589043, 'train/grad_norm_pre_clip': 31.69013786315918, 'timing/data': 0.0006120699690654874, 'timing/model': 1.9382451450219378, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 0.91}) +[2026-06-23 14:33:55,868][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.039577264338731766, 'train/grad_norm_pre_clip': 23.253297805786133, 'timing/data': 0.0007729619974270463, 'timing/model': 1.9345630209427327, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 0.91}) +[2026-06-23 14:33:59,689][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.052468191832304, 'train/grad_norm_pre_clip': 34.42107009887695, 'timing/data': 0.0005908999592065811, 'timing/model': 1.9385768959764391, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 0.91}) +[2026-06-23 14:34:03,505][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.12645527720451355, 'train/grad_norm_pre_clip': 67.73088073730469, 'timing/data': 0.0006691759917885065, 'timing/model': 1.9412432729732245, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.91}) +[2026-06-23 14:34:07,306][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.1307143121957779, 'train/grad_norm_pre_clip': 34.457889556884766, 'timing/data': 0.0006248001009225845, 'timing/model': 1.9289862720761448, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 0.91}) +[2026-06-23 14:34:11,101][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.03208203241229057, 'train/grad_norm_pre_clip': 36.72333908081055, 'timing/data': 0.0006008020136505365, 'timing/model': 1.9315027059055865, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 0.91}) +[2026-06-23 14:34:14,929][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.05026765912771225, 'train/grad_norm_pre_clip': 56.8044548034668, 'timing/data': 0.0005982109578326344, 'timing/model': 1.9561608899384737, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 0.91}) +[2026-06-23 14:34:18,740][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.06268089264631271, 'train/grad_norm_pre_clip': 22.26527214050293, 'timing/data': 0.0005943150026723742, 'timing/model': 1.9303487440338358, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.91}) +[2026-06-23 14:34:22,552][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.04687381535768509, 'train/grad_norm_pre_clip': 22.720632553100586, 'timing/data': 0.0008001290261745453, 'timing/model': 1.930514588020742, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 0.91}) +[2026-06-23 14:34:26,369][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.050801899284124374, 'train/grad_norm_pre_clip': 33.60096740722656, 'timing/data': 0.0006465740734711289, 'timing/model': 1.9390804810682312, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 0.91}) +[2026-06-23 14:34:30,169][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.016774170100688934, 'train/grad_norm_pre_clip': 21.27712631225586, 'timing/data': 0.0006101831095293164, 'timing/model': 1.9247499540215358, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 0.91}) +[2026-06-23 14:34:33,995][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.11365506052970886, 'train/grad_norm_pre_clip': 55.71860885620117, 'timing/data': 0.0008145129540935159, 'timing/model': 1.948830010020174, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 0.91}) +[2026-06-23 14:34:37,833][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.07589688897132874, 'train/grad_norm_pre_clip': 70.30225372314453, 'timing/data': 0.0005887700244784355, 'timing/model': 1.934316574013792, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 0.91}) +[2026-06-23 14:34:41,654][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.04334793612360954, 'train/grad_norm_pre_clip': 42.18928909301758, 'timing/data': 0.0006085220957174897, 'timing/model': 1.931406326009892, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.91}) +[2026-06-23 14:34:45,466][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.05286528542637825, 'train/grad_norm_pre_clip': 22.427213668823242, 'timing/data': 0.0006076729623600841, 'timing/model': 1.9396153460256755, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 0.91}) +[2026-06-23 14:34:49,269][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.05248516798019409, 'train/grad_norm_pre_clip': 21.224349975585938, 'timing/data': 0.0008123309817165136, 'timing/model': 1.928102502017282, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 0.91}) +[2026-06-23 14:34:53,083][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.03748767077922821, 'train/grad_norm_pre_clip': 29.189163208007812, 'timing/data': 0.0008362680673599243, 'timing/model': 1.9320294470526278, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 0.91}) +[2026-06-23 14:34:56,905][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.0746762678027153, 'train/grad_norm_pre_clip': 22.94699478149414, 'timing/data': 0.0006293749902397394, 'timing/model': 1.944214157992974, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 0.91}) +[2026-06-23 14:35:00,723][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.06730496883392334, 'train/grad_norm_pre_clip': 50.35855484008789, 'timing/data': 0.0006283690454438329, 'timing/model': 1.9316489199409261, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 0.92}) +[2026-06-23 14:35:04,719][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.15769809484481812, 'train/grad_norm_pre_clip': 59.254302978515625, 'timing/data': 0.0007558159995824099, 'timing/model': 2.1164683820679784, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 0.92}) +[2026-06-23 14:35:08,524][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.12181266397237778, 'train/grad_norm_pre_clip': 42.404029846191406, 'timing/data': 0.0007113470928743482, 'timing/model': 1.9324404690414667, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 0.92}) +[2026-06-23 14:35:12,331][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.048013415187597275, 'train/grad_norm_pre_clip': 11.441107749938965, 'timing/data': 0.0007015339797362685, 'timing/model': 1.9289392550708726, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 0.92}) +[2026-06-23 14:35:16,143][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.03454107418656349, 'train/grad_norm_pre_clip': 14.691475868225098, 'timing/data': 0.0007824399508535862, 'timing/model': 1.9333565480774269, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 0.92}) +[2026-06-23 14:35:19,936][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.04909319430589676, 'train/grad_norm_pre_clip': 25.444562911987305, 'timing/data': 0.0006853700615465641, 'timing/model': 1.9234688939759508, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.92}) +[2026-06-23 14:35:23,752][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.17914237082004547, 'train/grad_norm_pre_clip': 48.22148132324219, 'timing/data': 0.0005976150278002024, 'timing/model': 1.937961602001451, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 0.92}) +[2026-06-23 14:35:27,556][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.01572893187403679, 'train/grad_norm_pre_clip': 14.825081825256348, 'timing/data': 0.0005993240047246218, 'timing/model': 1.934583142050542, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 0.92}) +[2026-06-23 14:35:31,357][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.04262853041291237, 'train/grad_norm_pre_clip': 15.605807304382324, 'timing/data': 0.0006219620117917657, 'timing/model': 1.932834097999148, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 0.92}) +[2026-06-23 14:35:35,162][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.10600684583187103, 'train/grad_norm_pre_clip': 31.861358642578125, 'timing/data': 0.0006610989803448319, 'timing/model': 1.9334048760356382, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 0.92}) +[2026-06-23 14:35:39,144][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.10091198235750198, 'train/grad_norm_pre_clip': 39.053009033203125, 'timing/data': 0.000605165958404541, 'timing/model': 1.9293103270465508, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.92}) +[2026-06-23 14:35:42,944][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.15230335295200348, 'train/grad_norm_pre_clip': 30.12298011779785, 'timing/data': 0.000702467979863286, 'timing/model': 1.9238106360426173, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 0.92}) +[2026-06-23 14:35:46,759][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.006511430721729994, 'train/grad_norm_pre_clip': 25.42866325378418, 'timing/data': 0.0006442640442401171, 'timing/model': 1.9353155109565705, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 0.92}) +[2026-06-23 14:35:50,580][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.05325882509350777, 'train/grad_norm_pre_clip': 21.3115234375, 'timing/data': 0.00059500802308321, 'timing/model': 1.935271097929217, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 0.92}) +[2026-06-23 14:35:54,396][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.09221974015235901, 'train/grad_norm_pre_clip': 26.832698822021484, 'timing/data': 0.0006745300488546491, 'timing/model': 1.9385622369591147, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 0.92}) +[2026-06-23 14:35:58,204][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.048124682158231735, 'train/grad_norm_pre_clip': 12.427752494812012, 'timing/data': 0.0006412019720301032, 'timing/model': 1.9425620449474081, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.92}) +[2026-06-23 14:36:02,016][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.016093909740447998, 'train/grad_norm_pre_clip': 27.298667907714844, 'timing/data': 0.0006184200756251812, 'timing/model': 1.9326211449224502, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 0.92}) +[2026-06-23 14:36:05,822][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.08210372179746628, 'train/grad_norm_pre_clip': 36.33308029174805, 'timing/data': 0.0005901299882680178, 'timing/model': 1.9381380419945344, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 0.92}) +[2026-06-23 14:36:09,825][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.024908527731895447, 'train/grad_norm_pre_clip': 22.049959182739258, 'timing/data': 0.0006244779797270894, 'timing/model': 2.12573730701115, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 0.92}) +[2026-06-23 14:36:13,621][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.039239417761564255, 'train/grad_norm_pre_clip': 15.984177589416504, 'timing/data': 0.0005978390108793974, 'timing/model': 1.9262007840443403, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 0.92}) +[2026-06-23 14:36:17,421][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.0895855575799942, 'train/grad_norm_pre_clip': 37.876041412353516, 'timing/data': 0.0005997499683871865, 'timing/model': 1.9289401710266247, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 0.92}) +[2026-06-23 14:36:21,222][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.039329759776592255, 'train/grad_norm_pre_clip': 15.241168022155762, 'timing/data': 0.0006509080994874239, 'timing/model': 1.9364069469738752, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 0.92}) +[2026-06-23 14:36:25,018][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.0703277587890625, 'train/grad_norm_pre_clip': 38.06414031982422, 'timing/data': 0.000679123098962009, 'timing/model': 1.9287390119861811, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 0.92}) +[2026-06-23 14:36:28,830][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.15430805087089539, 'train/grad_norm_pre_clip': 73.25958251953125, 'timing/data': 0.0006214219611138105, 'timing/model': 1.9334534429945052, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 0.92}) +[2026-06-23 14:36:32,637][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.08492396771907806, 'train/grad_norm_pre_clip': 59.96165084838867, 'timing/data': 0.0006086879875510931, 'timing/model': 1.9351010789396241, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 0.92}) +[2026-06-23 14:36:36,437][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.08663010597229004, 'train/grad_norm_pre_clip': 38.132408142089844, 'timing/data': 0.0005980520509183407, 'timing/model': 1.9257585579762235, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.92}) +[2026-06-23 14:36:40,250][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.11018048971891403, 'train/grad_norm_pre_clip': 50.95180892944336, 'timing/data': 0.0006015920080244541, 'timing/model': 1.9366489470703527, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 0.93}) +[2026-06-23 14:36:44,072][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.036206986755132675, 'train/grad_norm_pre_clip': 20.773143768310547, 'timing/data': 0.0006129990797489882, 'timing/model': 1.9402927519986406, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 0.93}) +[2026-06-23 14:36:47,893][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.03832416236400604, 'train/grad_norm_pre_clip': 13.631685256958008, 'timing/data': 0.0008088339818641543, 'timing/model': 1.9388395830756053, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 0.93}) +[2026-06-23 14:36:51,690][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.08679641783237457, 'train/grad_norm_pre_clip': 49.08302688598633, 'timing/data': 0.0006927599897608161, 'timing/model': 1.923080577980727, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 0.93}) +[2026-06-23 14:36:55,499][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.06950023770332336, 'train/grad_norm_pre_clip': 38.035499572753906, 'timing/data': 0.0006004309980198741, 'timing/model': 1.9304907459300011, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 0.93}) +[2026-06-23 14:36:59,309][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.09252674877643585, 'train/grad_norm_pre_clip': 47.38448715209961, 'timing/data': 0.0006066400092095137, 'timing/model': 1.9352418479975313, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 0.93}) +[2026-06-23 14:37:03,115][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.1308141052722931, 'train/grad_norm_pre_clip': 48.45083999633789, 'timing/data': 0.000684048980474472, 'timing/model': 1.9299864190397784, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 0.93}) +[2026-06-23 14:37:06,920][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.03012477234005928, 'train/grad_norm_pre_clip': 20.638633728027344, 'timing/data': 0.0007545419503003359, 'timing/model': 1.921920660068281, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 0.93}) +[2026-06-23 14:37:10,750][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.04222675785422325, 'train/grad_norm_pre_clip': 24.322599411010742, 'timing/data': 0.000683707999996841, 'timing/model': 1.9652589380275458, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 0.93}) +[2026-06-23 14:37:14,734][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2400 +[2026-06-23 14:39:21,909][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2400 +[2026-06-23 14:39:21,911][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2400 +[2026-06-23 14:40:45,165][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2400 +[2026-06-23 14:40:45,165][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.07030967622995377, 'train/grad_norm_pre_clip': 24.071365356445312, 'eval/action_loss/seconds': 127.17219020193443, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.05799497291445732, 'eval/latency_2/loss': 0.05799497291445732, 'eval/flappy/loss': 0.05799497291445732, 'eval/flappy/latency_2/loss': 0.05799497291445732, 'eval/flap_precision': 0.9881305637982196, 'eval/flap_recall': 0.8987854251012146, 'eval/flap_f1': 0.9413427561837456, 'eval/flap_rate_model': 0.18722222222222223, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9413427561837457, 'eval/latency_2/flap_precision': 0.9881305637982196, 'eval/latency_2/flap_recall': 0.8987854251012146, 'eval/latency_2/flap_f1': 0.9413427561837456, 'eval/latency_2/flap_rate_model': 0.18722222222222223, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9413427561837457, 'eval/action_classification/data_wait_seconds': 1.2748043537139893, 'eval/action_classification/predict_seconds': 81.7945785522461, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0008112449431791902, 'timing/model': 2.107934106956236, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.93}) +[2026-06-23 14:40:48,949][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.1009880006313324, 'train/grad_norm_pre_clip': 107.56912231445312, 'timing/data': 0.0007232599891722202, 'timing/model': 1.9218207029625773, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 0.93}) +[2026-06-23 14:40:52,740][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.1995438039302826, 'train/grad_norm_pre_clip': 90.01790618896484, 'timing/data': 0.0006125149084255099, 'timing/model': 1.9209699500352144, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 0.93}) +[2026-06-23 14:40:56,526][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.06462308764457703, 'train/grad_norm_pre_clip': 47.35271072387695, 'timing/data': 0.0006434960523620248, 'timing/model': 1.9191885039908811, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.93}) +[2026-06-23 14:41:00,319][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.10545139759778976, 'train/grad_norm_pre_clip': 52.667808532714844, 'timing/data': 0.0006364090368151665, 'timing/model': 1.9306031669257209, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 0.93}) +[2026-06-23 14:41:04,304][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.059105753898620605, 'train/grad_norm_pre_clip': 28.753864288330078, 'timing/data': 0.0005994950188323855, 'timing/model': 1.918778700986877, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 0.93}) +[2026-06-23 14:41:08,104][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.06064536049962044, 'train/grad_norm_pre_clip': 13.394214630126953, 'timing/data': 0.000818532076664269, 'timing/model': 1.9241307779448107, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 0.93}) +[2026-06-23 14:41:11,892][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.11526240408420563, 'train/grad_norm_pre_clip': 24.32207679748535, 'timing/data': 0.0007983979303389788, 'timing/model': 1.9207063330104575, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 0.93}) +[2026-06-23 14:41:15,709][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.07598600536584854, 'train/grad_norm_pre_clip': 31.740915298461914, 'timing/data': 0.0006176580209285021, 'timing/model': 1.9321415829472244, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 0.93}) +[2026-06-23 14:41:19,524][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.19163084030151367, 'train/grad_norm_pre_clip': 109.48030090332031, 'timing/data': 0.0006919869920238853, 'timing/model': 1.938240387942642, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 0.93}) +[2026-06-23 14:41:23,329][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.07399994134902954, 'train/grad_norm_pre_clip': 48.0050163269043, 'timing/data': 0.0006145089864730835, 'timing/model': 1.9250834359554574, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.93}) +[2026-06-23 14:41:27,150][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.018564168363809586, 'train/grad_norm_pre_clip': 47.77718734741211, 'timing/data': 0.000618502963334322, 'timing/model': 1.942581244977191, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 0.93}) +[2026-06-23 14:41:30,960][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.05034307762980461, 'train/grad_norm_pre_clip': 33.12541580200195, 'timing/data': 0.0007681109709665179, 'timing/model': 1.9215010500047356, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 0.93}) +[2026-06-23 14:41:34,920][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.05042646825313568, 'train/grad_norm_pre_clip': 32.794342041015625, 'timing/data': 0.0006630809511989355, 'timing/model': 2.0951656360412017, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 0.93}) +[2026-06-23 14:41:38,718][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.11729570478200912, 'train/grad_norm_pre_clip': 37.86138153076172, 'timing/data': 0.000673800939694047, 'timing/model': 1.9320554489968345, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 0.93}) +[2026-06-23 14:41:42,536][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.03501468524336815, 'train/grad_norm_pre_clip': 23.787811279296875, 'timing/data': 0.0006249319994822145, 'timing/model': 1.9408961379667744, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 0.93}) +[2026-06-23 14:41:46,328][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.07315310835838318, 'train/grad_norm_pre_clip': 23.984460830688477, 'timing/data': 0.0006858799606561661, 'timing/model': 1.9270030680345371, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 0.94}) +[2026-06-23 14:41:50,128][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.0439029224216938, 'train/grad_norm_pre_clip': 24.424522399902344, 'timing/data': 0.000587497022934258, 'timing/model': 1.928042359999381, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 0.94}) +[2026-06-23 14:41:54,025][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.08432251960039139, 'train/grad_norm_pre_clip': 24.29022216796875, 'timing/data': 0.0008038349915295839, 'timing/model': 2.014912861981429, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 0.94}) +[2026-06-23 14:41:57,826][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.07495235651731491, 'train/grad_norm_pre_clip': 24.975393295288086, 'timing/data': 0.000658861012198031, 'timing/model': 1.9346044429112226, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 0.94}) +[2026-06-23 14:42:01,630][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.057546548545360565, 'train/grad_norm_pre_clip': 45.56034469604492, 'timing/data': 0.0007545379921793938, 'timing/model': 1.9349536339286715, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.94}) +[2026-06-23 14:42:05,430][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.023738985881209373, 'train/grad_norm_pre_clip': 27.725751876831055, 'timing/data': 0.0008118969853967428, 'timing/model': 1.933656172011979, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 0.94}) +[2026-06-23 14:42:09,411][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.1073012426495552, 'train/grad_norm_pre_clip': 32.35279083251953, 'timing/data': 0.0007329710060730577, 'timing/model': 1.926502319984138, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 0.94}) +[2026-06-23 14:42:13,200][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.11312469094991684, 'train/grad_norm_pre_clip': 41.203025817871094, 'timing/data': 0.0007748249918222427, 'timing/model': 1.9271949690300971, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 0.94}) +[2026-06-23 14:42:17,003][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.04927492141723633, 'train/grad_norm_pre_clip': 26.487850189208984, 'timing/data': 0.0005931489868089557, 'timing/model': 1.9303221789887175, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 0.94}) +[2026-06-23 14:42:20,801][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.06432843953371048, 'train/grad_norm_pre_clip': 18.647653579711914, 'timing/data': 0.0009205449605360627, 'timing/model': 1.9353105640038848, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 0.94}) +[2026-06-23 14:42:24,608][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.1270218938589096, 'train/grad_norm_pre_clip': 31.384733200073242, 'timing/data': 0.0006094499258324504, 'timing/model': 1.9308394800173119, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 0.94}) +[2026-06-23 14:42:28,423][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.0285978764295578, 'train/grad_norm_pre_clip': 41.62232208251953, 'timing/data': 0.000813826103694737, 'timing/model': 1.9374628990190104, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 0.94}) +[2026-06-23 14:42:32,218][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.05817897617816925, 'train/grad_norm_pre_clip': 14.102483749389648, 'timing/data': 0.0006643290398642421, 'timing/model': 1.9192764399340376, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 0.94}) +[2026-06-23 14:42:36,022][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.046691056340932846, 'train/grad_norm_pre_clip': 15.839608192443848, 'timing/data': 0.0006113379495218396, 'timing/model': 1.9364763019839302, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 0.94}) +[2026-06-23 14:42:39,966][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.0934395045042038, 'train/grad_norm_pre_clip': 47.566104888916016, 'timing/data': 0.0005978529807180166, 'timing/model': 2.0762532689841464, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 0.94}) +[2026-06-23 14:42:43,770][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.06797093152999878, 'train/grad_norm_pre_clip': 20.579010009765625, 'timing/data': 0.0006719109369441867, 'timing/model': 1.9337424929253757, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 0.94}) +[2026-06-23 14:42:47,584][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.06696345657110214, 'train/grad_norm_pre_clip': 39.43508529663086, 'timing/data': 0.0005914089269936085, 'timing/model': 1.933319309959188, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 0.94}) +[2026-06-23 14:42:51,393][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.018833883106708527, 'train/grad_norm_pre_clip': 20.893613815307617, 'timing/data': 0.0006525800563395023, 'timing/model': 1.935834203963168, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 0.94}) +[2026-06-23 14:42:55,191][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.037941236048936844, 'train/grad_norm_pre_clip': 27.802888870239258, 'timing/data': 0.0005911568878218532, 'timing/model': 1.9283009700011462, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 0.94}) +[2026-06-23 14:42:58,996][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.06653939932584763, 'train/grad_norm_pre_clip': 25.518413543701172, 'timing/data': 0.0006612170254811645, 'timing/model': 1.9304434469668195, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 0.94}) +[2026-06-23 14:43:02,798][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.06118904426693916, 'train/grad_norm_pre_clip': 39.643795013427734, 'timing/data': 0.0006377518875524402, 'timing/model': 1.9288214000407606, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 0.94}) +[2026-06-23 14:43:06,603][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.0664263367652893, 'train/grad_norm_pre_clip': 27.980648040771484, 'timing/data': 0.0007524579996243119, 'timing/model': 1.9329728169832379, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 0.94}) +[2026-06-23 14:43:10,408][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.03125132992863655, 'train/grad_norm_pre_clip': 33.179630279541016, 'timing/data': 0.0007259900448843837, 'timing/model': 1.9277407369809225, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 0.94}) +[2026-06-23 14:43:14,352][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.09772413969039917, 'train/grad_norm_pre_clip': 26.732200622558594, 'timing/data': 0.0006007369374856353, 'timing/model': 1.9260435090400279, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 0.94}) +[2026-06-23 14:43:18,168][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.13939426839351654, 'train/grad_norm_pre_clip': 38.83354949951172, 'timing/data': 0.000609269947744906, 'timing/model': 1.9334919700631872, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 0.94}) +[2026-06-23 14:43:21,976][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.06922617554664612, 'train/grad_norm_pre_clip': 47.32773971557617, 'timing/data': 0.0007891879649832845, 'timing/model': 1.9263703129254282, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 0.94}) +[2026-06-23 14:43:25,794][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.04094816371798515, 'train/grad_norm_pre_clip': 23.69507598876953, 'timing/data': 0.0005963889416307211, 'timing/model': 1.9440024599898607, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 0.95}) +[2026-06-23 14:43:29,592][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.029112016782164574, 'train/grad_norm_pre_clip': 34.44417190551758, 'timing/data': 0.0006022359011694789, 'timing/model': 1.9258517459966242, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 0.95}) +[2026-06-23 14:43:33,399][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.023856637999415398, 'train/grad_norm_pre_clip': 18.31477165222168, 'timing/data': 0.0006122899940237403, 'timing/model': 1.9261084001045674, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 0.95}) +[2026-06-23 14:43:37,206][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.03597540780901909, 'train/grad_norm_pre_clip': 19.98626136779785, 'timing/data': 0.0005991819780319929, 'timing/model': 1.93548732297495, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.95}) +[2026-06-23 14:43:41,017][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.04326143115758896, 'train/grad_norm_pre_clip': 26.86603546142578, 'timing/data': 0.000704668927937746, 'timing/model': 1.9284173869527876, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 0.95}) +[2026-06-23 14:43:44,890][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.09479004889726639, 'train/grad_norm_pre_clip': 27.79998779296875, 'timing/data': 0.0006652550073340535, 'timing/model': 2.006711142952554, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 0.95}) +[2026-06-23 14:43:48,706][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.09479687362909317, 'train/grad_norm_pre_clip': 33.50891876220703, 'timing/data': 0.0006041619926691055, 'timing/model': 1.9313035239465535, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 0.95}) +[2026-06-23 14:43:52,520][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.04455877095460892, 'train/grad_norm_pre_clip': 17.4581241607666, 'timing/data': 0.000708339037373662, 'timing/model': 1.9377841749228537, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 0.95}) +[2026-06-23 14:43:56,333][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.15253227949142456, 'train/grad_norm_pre_clip': 25.728090286254883, 'timing/data': 0.000584915978834033, 'timing/model': 1.9389636599225923, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 0.95}) +[2026-06-23 14:44:00,129][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.11409120261669159, 'train/grad_norm_pre_clip': 19.555644989013672, 'timing/data': 0.0006459630094468594, 'timing/model': 1.9289400440175086, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 0.95}) +[2026-06-23 14:44:03,959][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.30335357785224915, 'train/grad_norm_pre_clip': 43.57714080810547, 'timing/data': 0.0006527119548991323, 'timing/model': 1.9251724729547277, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.95}) +[2026-06-23 14:44:07,761][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.011270797811448574, 'train/grad_norm_pre_clip': 19.950586318969727, 'timing/data': 0.0005999930435791612, 'timing/model': 1.9316383430268615, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 0.95}) +[2026-06-23 14:44:11,557][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.09465385228395462, 'train/grad_norm_pre_clip': 25.699115753173828, 'timing/data': 0.0006580520421266556, 'timing/model': 1.9255482669686899, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 0.95}) +[2026-06-23 14:44:15,373][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.126645028591156, 'train/grad_norm_pre_clip': 50.59318923950195, 'timing/data': 0.0006855659885331988, 'timing/model': 1.9360975889721885, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 0.95}) +[2026-06-23 14:44:19,238][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.0740853101015091, 'train/grad_norm_pre_clip': 41.422786712646484, 'timing/data': 0.0006004510214552283, 'timing/model': 1.9289283209946007, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 0.95}) +[2026-06-23 14:44:23,060][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.052122630178928375, 'train/grad_norm_pre_clip': 22.22006607055664, 'timing/data': 0.0006682079983875155, 'timing/model': 1.9353148069931194, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 0.95}) +[2026-06-23 14:44:26,860][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.06779488176107407, 'train/grad_norm_pre_clip': 32.14083480834961, 'timing/data': 0.0007453319849446416, 'timing/model': 1.9298646530369297, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 0.95}) +[2026-06-23 14:44:30,669][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.15578114986419678, 'train/grad_norm_pre_clip': 26.926429748535156, 'timing/data': 0.0006679430371150374, 'timing/model': 1.9284578149672598, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 0.95}) +[2026-06-23 14:44:34,465][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.048939213156700134, 'train/grad_norm_pre_clip': 31.827512741088867, 'timing/data': 0.0007016460876911879, 'timing/model': 1.918957720976323, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 0.95}) +[2026-06-23 14:44:38,271][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.060543544590473175, 'train/grad_norm_pre_clip': 21.361827850341797, 'timing/data': 0.0006087040528655052, 'timing/model': 1.9236798449419439, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 0.95}) +[2026-06-23 14:44:42,079][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.10637678951025009, 'train/grad_norm_pre_clip': 28.98722267150879, 'timing/data': 0.0006241200026124716, 'timing/model': 1.9261059350101277, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 0.95}) +[2026-06-23 14:44:45,886][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.10880514234304428, 'train/grad_norm_pre_clip': 31.10612678527832, 'timing/data': 0.0005854490445926785, 'timing/model': 1.9375592319993302, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 0.95}) +[2026-06-23 14:44:49,714][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.048660922795534134, 'train/grad_norm_pre_clip': 24.52328109741211, 'timing/data': 0.000610062968917191, 'timing/model': 1.945560373016633, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 0.95}) +[2026-06-23 14:44:53,546][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.06610069423913956, 'train/grad_norm_pre_clip': 44.947242736816406, 'timing/data': 0.0006152769783511758, 'timing/model': 1.945077883079648, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 0.95}) +[2026-06-23 14:44:57,373][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.051648013293743134, 'train/grad_norm_pre_clip': 19.708831787109375, 'timing/data': 0.0006037970306351781, 'timing/model': 1.9446144590619951, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 0.95}) +[2026-06-23 14:45:01,178][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.03101605735719204, 'train/grad_norm_pre_clip': 13.449851036071777, 'timing/data': 0.0006152749992907047, 'timing/model': 1.9263801759807393, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 0.95}) +[2026-06-23 14:45:04,987][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.05391382798552513, 'train/grad_norm_pre_clip': 12.066178321838379, 'timing/data': 0.0006090359529480338, 'timing/model': 1.9402732569724321, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 0.96}) +[2026-06-23 14:45:08,805][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.15893647074699402, 'train/grad_norm_pre_clip': 27.29494857788086, 'timing/data': 0.0006661780644208193, 'timing/model': 1.9358560019172728, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 0.96}) +[2026-06-23 14:45:12,604][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.105603888630867, 'train/grad_norm_pre_clip': 29.778532028198242, 'timing/data': 0.0007593269692733884, 'timing/model': 1.9294591580983251, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 0.96}) +[2026-06-23 14:45:16,416][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.16317327320575714, 'train/grad_norm_pre_clip': 43.25897216796875, 'timing/data': 0.0006079149898141623, 'timing/model': 1.9244398419978097, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 0.96}) +[2026-06-23 14:45:20,214][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.10535267740488052, 'train/grad_norm_pre_clip': 42.936988830566406, 'timing/data': 0.0009662919910624623, 'timing/model': 1.9291581980651245, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 0.96}) +[2026-06-23 14:45:24,021][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.09997829049825668, 'train/grad_norm_pre_clip': 21.604969024658203, 'timing/data': 0.0006120160687714815, 'timing/model': 1.9276413939660415, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 0.96}) +[2026-06-23 14:45:27,816][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.04625633358955383, 'train/grad_norm_pre_clip': 14.726072311401367, 'timing/data': 0.0006199739873409271, 'timing/model': 1.928713301080279, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 0.96}) +[2026-06-23 14:45:31,636][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.0732540413737297, 'train/grad_norm_pre_clip': 18.579177856445312, 'timing/data': 0.0006687151035293937, 'timing/model': 1.933412536047399, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 0.96}) +[2026-06-23 14:45:35,436][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.036934129893779755, 'train/grad_norm_pre_clip': 21.93037986755371, 'timing/data': 0.0006070910021662712, 'timing/model': 1.9204368370119482, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 0.96}) +[2026-06-23 14:45:39,248][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.053985122591257095, 'train/grad_norm_pre_clip': 29.818071365356445, 'timing/data': 0.0006030689692124724, 'timing/model': 1.9310571909882128, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 0.96}) +[2026-06-23 14:45:43,044][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.10714787989854813, 'train/grad_norm_pre_clip': 27.670625686645508, 'timing/data': 0.0008235699497163296, 'timing/model': 1.9241367870708928, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 0.96}) +[2026-06-23 14:45:46,852][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.04557165876030922, 'train/grad_norm_pre_clip': 36.86396408081055, 'timing/data': 0.0009427809854969382, 'timing/model': 1.9308908749371767, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 0.96}) +[2026-06-23 14:45:50,684][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.21413227915763855, 'train/grad_norm_pre_clip': 39.633697509765625, 'timing/data': 0.0006161279743537307, 'timing/model': 1.950719007058069, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.96}) +[2026-06-23 14:45:54,490][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.06004440039396286, 'train/grad_norm_pre_clip': 19.68378257751465, 'timing/data': 0.0006000830326229334, 'timing/model': 1.9316659069154412, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 0.96}) +[2026-06-23 14:45:58,464][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.08652272075414658, 'train/grad_norm_pre_clip': 39.93010711669922, 'timing/data': 0.0007993060862645507, 'timing/model': 1.9300414479803294, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 0.96}) +[2026-06-23 14:46:02,264][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.07191778719425201, 'train/grad_norm_pre_clip': 32.80643844604492, 'timing/data': 0.0006132510025054216, 'timing/model': 1.9252170490799472, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 0.96}) +[2026-06-23 14:46:06,057][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.028781741857528687, 'train/grad_norm_pre_clip': 21.879638671875, 'timing/data': 0.0005988490302115679, 'timing/model': 1.9300634230021387, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 0.96}) +[2026-06-23 14:46:09,872][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.08224530518054962, 'train/grad_norm_pre_clip': 42.65152359008789, 'timing/data': 0.0008012939943000674, 'timing/model': 1.9391933680744842, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 0.96}) +[2026-06-23 14:46:13,686][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.10857005417346954, 'train/grad_norm_pre_clip': 61.58958053588867, 'timing/data': 0.0006140769692137837, 'timing/model': 1.932930313050747, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 0.96}) +[2026-06-23 14:46:17,492][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.04627009108662605, 'train/grad_norm_pre_clip': 22.74267578125, 'timing/data': 0.0007946339901536703, 'timing/model': 1.9402524839388207, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 0.96}) +[2026-06-23 14:46:21,286][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.038475535809993744, 'train/grad_norm_pre_clip': 17.371376037597656, 'timing/data': 0.0006234359461814165, 'timing/model': 1.9222582849906757, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 0.96}) +[2026-06-23 14:46:25,089][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.061545129865407944, 'train/grad_norm_pre_clip': 29.001609802246094, 'timing/data': 0.00066895701456815, 'timing/model': 1.9250698150135577, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 0.96}) +[2026-06-23 14:46:28,955][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.025943640619516373, 'train/grad_norm_pre_clip': 22.5153865814209, 'timing/data': 0.0005890779430046678, 'timing/model': 1.992461625020951, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 0.96}) +[2026-06-23 14:46:32,766][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.05244414508342743, 'train/grad_norm_pre_clip': 30.452970504760742, 'timing/data': 0.0007272220682352781, 'timing/model': 1.9323252009926364, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 0.96}) +[2026-06-23 14:46:36,575][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.013423782773315907, 'train/grad_norm_pre_clip': 12.040353775024414, 'timing/data': 0.000692634959705174, 'timing/model': 1.9292403140570968, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 0.96}) +[2026-06-23 14:46:40,394][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.07178709656000137, 'train/grad_norm_pre_clip': 18.285480499267578, 'timing/data': 0.0005950269987806678, 'timing/model': 1.9405569341033697, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 0.96}) +[2026-06-23 14:46:44,191][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.159408837556839, 'train/grad_norm_pre_clip': 36.36209487915039, 'timing/data': 0.0007707369513809681, 'timing/model': 1.9273214420536533, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 0.97}) +[2026-06-23 14:46:47,988][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.07525759190320969, 'train/grad_norm_pre_clip': 28.32642936706543, 'timing/data': 0.0007926109246909618, 'timing/model': 1.92885516397655, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 0.97}) +[2026-06-23 14:46:51,800][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.02342175878584385, 'train/grad_norm_pre_clip': 14.779340744018555, 'timing/data': 0.0009057309944182634, 'timing/model': 1.9366980040213093, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 0.97}) +[2026-06-23 14:46:55,637][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.03924470767378807, 'train/grad_norm_pre_clip': 51.493690490722656, 'timing/data': 0.0006048569921404123, 'timing/model': 1.9525388960028067, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 0.97}) +[2026-06-23 14:46:59,465][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.03977527469396591, 'train/grad_norm_pre_clip': 39.981849670410156, 'timing/data': 0.000798787921667099, 'timing/model': 1.9452300718985498, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 0.97}) +[2026-06-23 14:47:03,334][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.10603564977645874, 'train/grad_norm_pre_clip': 46.57044982910156, 'timing/data': 0.0006225650431588292, 'timing/model': 1.93363749794662, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 0.97}) +[2026-06-23 14:47:07,142][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2500 +[2026-06-23 14:49:13,738][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2500 +[2026-06-23 14:49:13,739][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2500 +[2026-06-23 14:50:36,388][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2500 +[2026-06-23 14:55:55,589][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-06-23 14:55:55,589][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.026385461911559105, 'train/grad_norm_pre_clip': 31.801713943481445, 'eval/action_loss/seconds': 126.59345282393042, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04444492980837822, 'eval/latency_2/loss': 0.04444492980837822, 'eval/flappy/loss': 0.04444492980837822, 'eval/flappy/latency_2/loss': 0.04444492980837822, 'eval/flap_precision': 0.9380245319561007, 'eval/flap_recall': 0.9804318488529015, 'eval/flap_f1': 0.9587594853183767, 'eval/flap_rate_model': 0.21513888888888888, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9587594853183766, 'eval/latency_2/flap_precision': 0.9380245319561007, 'eval/latency_2/flap_recall': 0.9804318488529015, 'eval/latency_2/flap_f1': 0.9587594853183767, 'eval/latency_2/flap_rate_model': 0.21513888888888888, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9587594853183766, 'eval/action_classification/data_wait_seconds': 1.4483649730682373, 'eval/action_classification/predict_seconds': 80.99742889404297, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 194.76000000001028, 'rl_games_eval/mid_train/mean_length': 1587.0, 'rl_games_eval/mid_train/std_reward': 205.5255176371176, 'rl_games_eval/mid_train/std_length': 1656.0664237886112, 'rl_games_eval/mid_train/macro_mean_reward': 194.76000000001028, 'rl_games_eval/mid_train/macro_mean_length': 1587.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 194.76000000001028, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1587.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 205.5255176371176, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1656.0664237886112, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 194.76000000001028, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006102300249040127, 'timing/model': 1.9356650200206786, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 0.97}) +[2026-06-23 14:55:55,590][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_2500_state +[2026-06-23 14:55:55,593][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 14:55:59,433][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_2500_state/model.safetensors +[2026-06-23 14:56:02,355][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_2500_state/optimizer.bin +[2026-06-23 14:56:02,355][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_2500_state/sampler.bin +[2026-06-23 14:56:02,355][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-23 14:56:02,356][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-23 14:56:02,382][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-23 14:56:02,382][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 14:56:02,411][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 14:56:06,136][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.06028037890791893, 'train/grad_norm_pre_clip': 23.51584815979004, 'timing/data': 0.0007718189153820276, 'timing/model': 1.8841595110716298, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 0.97}) +[2026-06-23 14:56:09,878][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.07018150389194489, 'train/grad_norm_pre_clip': 49.39393615722656, 'timing/data': 0.0006401389837265015, 'timing/model': 1.9004297410137951, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 0.97}) +[2026-06-23 14:56:13,697][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.08730368316173553, 'train/grad_norm_pre_clip': 31.600276947021484, 'timing/data': 0.0009012080263346434, 'timing/model': 1.9861240109894425, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 0.97}) +[2026-06-23 14:56:17,440][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.16920801997184753, 'train/grad_norm_pre_clip': 51.63686752319336, 'timing/data': 0.0006014870014041662, 'timing/model': 1.898911066004075, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 0.97}) +[2026-06-23 14:56:21,180][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.02809743396937847, 'train/grad_norm_pre_clip': 23.543107986450195, 'timing/data': 0.0008084949804469943, 'timing/model': 1.8979545349720865, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 0.97}) +[2026-06-23 14:56:24,932][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.030689267441630363, 'train/grad_norm_pre_clip': 16.5214900970459, 'timing/data': 0.0006007390329614282, 'timing/model': 1.9039451509015635, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 0.97}) +[2026-06-23 14:56:28,683][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.006902617402374744, 'train/grad_norm_pre_clip': 13.94409465789795, 'timing/data': 0.0006068069487810135, 'timing/model': 1.899977016961202, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 0.97}) +[2026-06-23 14:56:32,432][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.058030012995004654, 'train/grad_norm_pre_clip': 27.427448272705078, 'timing/data': 0.0005957259563729167, 'timing/model': 1.9016701149521396, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 0.97}) +[2026-06-23 14:56:36,189][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.04576984792947769, 'train/grad_norm_pre_clip': 39.522239685058594, 'timing/data': 0.0006021189037710428, 'timing/model': 1.904401381034404, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 0.97}) +[2026-06-23 14:56:39,955][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.04510882869362831, 'train/grad_norm_pre_clip': 21.818544387817383, 'timing/data': 0.0006066670175641775, 'timing/model': 1.9170247559668496, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 0.97}) +[2026-06-23 14:56:43,762][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.1314045637845993, 'train/grad_norm_pre_clip': 32.48738098144531, 'timing/data': 0.0006767519516870379, 'timing/model': 1.918579016928561, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 0.97}) +[2026-06-23 14:56:47,655][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.052430830895900726, 'train/grad_norm_pre_clip': 28.857440948486328, 'timing/data': 0.0006261060480028391, 'timing/model': 2.031878365087323, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 0.97}) +[2026-06-23 14:56:51,421][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.03924179822206497, 'train/grad_norm_pre_clip': 19.009977340698242, 'timing/data': 0.0006515909917652607, 'timing/model': 1.9135567039484158, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 0.97}) +[2026-06-23 14:56:55,190][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.022858818992972374, 'train/grad_norm_pre_clip': 19.069799423217773, 'timing/data': 0.0006227620178833604, 'timing/model': 1.916898027062416, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 0.97}) +[2026-06-23 14:56:58,961][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.09185772389173508, 'train/grad_norm_pre_clip': 23.715065002441406, 'timing/data': 0.0007788350339978933, 'timing/model': 1.91208858799655, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.97}) +[2026-06-23 14:57:02,736][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.08264515548944473, 'train/grad_norm_pre_clip': 41.93313217163086, 'timing/data': 0.0005971629871055484, 'timing/model': 1.918510297080502, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 0.97}) +[2026-06-23 14:57:06,578][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.14839139580726624, 'train/grad_norm_pre_clip': 40.39798355102539, 'timing/data': 0.00062289007473737, 'timing/model': 1.9848994850181043, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 0.97}) +[2026-06-23 14:57:10,348][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.07395616918802261, 'train/grad_norm_pre_clip': 47.78266906738281, 'timing/data': 0.0006975590949878097, 'timing/model': 1.9155933099100366, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 0.97}) +[2026-06-23 14:57:14,121][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.1307920664548874, 'train/grad_norm_pre_clip': 33.780765533447266, 'timing/data': 0.0006019680295139551, 'timing/model': 1.9126687359530479, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 0.97}) +[2026-06-23 14:57:17,938][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.087614044547081, 'train/grad_norm_pre_clip': 15.533934593200684, 'timing/data': 0.0006069749360904098, 'timing/model': 1.9552681799978018, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 0.98}) +[2026-06-23 14:57:21,720][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.029587289318442345, 'train/grad_norm_pre_clip': 15.994888305664062, 'timing/data': 0.0008091439958661795, 'timing/model': 1.9171888079727069, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 0.98}) +[2026-06-23 14:57:25,493][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.08483554422855377, 'train/grad_norm_pre_clip': 44.88823699951172, 'timing/data': 0.0006083260523155332, 'timing/model': 1.9124137960607186, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 0.98}) +[2026-06-23 14:57:29,278][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.05557646602392197, 'train/grad_norm_pre_clip': 26.474626541137695, 'timing/data': 0.0008003980619832873, 'timing/model': 1.925484150997363, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 0.98}) +[2026-06-23 14:57:33,072][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.00483296811580658, 'train/grad_norm_pre_clip': 21.572608947753906, 'timing/data': 0.0005950559861958027, 'timing/model': 1.9314939539181069, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 0.98}) +[2026-06-23 14:57:36,863][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.036059267818927765, 'train/grad_norm_pre_clip': 18.056316375732422, 'timing/data': 0.0007989270379766822, 'timing/model': 1.9251004280522466, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 0.98}) +[2026-06-23 14:57:40,647][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.052055951207876205, 'train/grad_norm_pre_clip': 48.101863861083984, 'timing/data': 0.0005928220925852656, 'timing/model': 1.9240553220734, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 0.98}) +[2026-06-23 14:57:44,448][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.07503233104944229, 'train/grad_norm_pre_clip': 41.112937927246094, 'timing/data': 0.0005992120131850243, 'timing/model': 1.931902616051957, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 0.98}) +[2026-06-23 14:57:48,235][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.11846625804901123, 'train/grad_norm_pre_clip': 29.466630935668945, 'timing/data': 0.000613359035924077, 'timing/model': 1.920713042956777, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 0.98}) +[2026-06-23 14:57:52,072][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.06410453468561172, 'train/grad_norm_pre_clip': 59.695228576660156, 'timing/data': 0.0006024580216035247, 'timing/model': 1.9249845820013434, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 0.98}) +[2026-06-23 14:57:55,889][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.014522871933877468, 'train/grad_norm_pre_clip': 23.320396423339844, 'timing/data': 0.0005868569714948535, 'timing/model': 1.935767735936679, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 0.98}) +[2026-06-23 14:57:59,675][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.1697193831205368, 'train/grad_norm_pre_clip': 77.70512390136719, 'timing/data': 0.000680833007209003, 'timing/model': 1.9283004309982061, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 0.98}) +[2026-06-23 14:58:03,476][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.05223677679896355, 'train/grad_norm_pre_clip': 40.03364562988281, 'timing/data': 0.0006017780397087336, 'timing/model': 1.9274181349901482, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 0.98}) +[2026-06-23 14:58:07,271][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.08105865120887756, 'train/grad_norm_pre_clip': 26.84906005859375, 'timing/data': 0.0008769590640440583, 'timing/model': 1.9297130280174315, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 0.98}) +[2026-06-23 14:58:11,075][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.01395051833242178, 'train/grad_norm_pre_clip': 10.128771781921387, 'timing/data': 0.0006962460465729237, 'timing/model': 1.9286018050042912, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 0.98}) +[2026-06-23 14:58:14,882][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.03580841049551964, 'train/grad_norm_pre_clip': 18.90097427368164, 'timing/data': 0.0008932249620556831, 'timing/model': 1.9276090380735695, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 0.98}) +[2026-06-23 14:58:18,676][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.03437881916761398, 'train/grad_norm_pre_clip': 17.176830291748047, 'timing/data': 0.0006071539828553796, 'timing/model': 1.918844826053828, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 0.98}) +[2026-06-23 14:58:22,466][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.0728689581155777, 'train/grad_norm_pre_clip': 21.598102569580078, 'timing/data': 0.0007854499854147434, 'timing/model': 1.9272631260100752, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 0.98}) +[2026-06-23 14:58:26,263][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.02639066055417061, 'train/grad_norm_pre_clip': 36.61259078979492, 'timing/data': 0.0007439689943566918, 'timing/model': 1.9253660860704258, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 0.98}) +[2026-06-23 14:58:30,071][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.0994795560836792, 'train/grad_norm_pre_clip': 26.024002075195312, 'timing/data': 0.0006037720013409853, 'timing/model': 1.9299104429082945, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 0.98}) +[2026-06-23 14:58:33,871][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.18571233749389648, 'train/grad_norm_pre_clip': 50.11216735839844, 'timing/data': 0.0006640290375798941, 'timing/model': 1.9244957520859316, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 0.98}) +[2026-06-23 14:58:37,660][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.10510233789682388, 'train/grad_norm_pre_clip': 23.52760887145996, 'timing/data': 0.0006126619409769773, 'timing/model': 1.914491966017522, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 0.98}) +[2026-06-23 14:58:41,468][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.08913488686084747, 'train/grad_norm_pre_clip': 36.29138946533203, 'timing/data': 0.0005992329679429531, 'timing/model': 1.9213698959210888, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 0.98}) +[2026-06-23 14:58:45,275][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.03248340263962746, 'train/grad_norm_pre_clip': 41.61675262451172, 'timing/data': 0.000623960979282856, 'timing/model': 1.9365689980331808, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 0.98}) +[2026-06-23 14:58:49,062][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.16802668571472168, 'train/grad_norm_pre_clip': 32.501243591308594, 'timing/data': 0.0006751639302819967, 'timing/model': 1.9245996159734204, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 0.98}) +[2026-06-23 14:58:52,860][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.04061674326658249, 'train/grad_norm_pre_clip': 8.866514205932617, 'timing/data': 0.0006579100154340267, 'timing/model': 1.9336088430136442, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 0.98}) +[2026-06-23 14:58:56,848][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.05974491313099861, 'train/grad_norm_pre_clip': 62.22732925415039, 'timing/data': 0.0005852410104125738, 'timing/model': 2.116971984039992, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 0.99}) +[2026-06-23 14:59:00,669][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.028098661452531815, 'train/grad_norm_pre_clip': 17.694889068603516, 'timing/data': 0.0007567669963464141, 'timing/model': 1.9498450639657676, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 0.99}) +[2026-06-23 14:59:04,469][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.11090760678052902, 'train/grad_norm_pre_clip': 40.95029067993164, 'timing/data': 0.0006032220553606749, 'timing/model': 1.9296140929218382, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 0.99}) +[2026-06-23 14:59:08,261][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.06383751332759857, 'train/grad_norm_pre_clip': 26.596118927001953, 'timing/data': 0.0008283979259431362, 'timing/model': 1.931391504011117, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 0.99}) +[2026-06-23 14:59:12,060][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.04016796126961708, 'train/grad_norm_pre_clip': 36.18440628051758, 'timing/data': 0.0006520450115203857, 'timing/model': 1.9287414480932057, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.99}) +[2026-06-23 14:59:15,868][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.009540446102619171, 'train/grad_norm_pre_clip': 18.74013900756836, 'timing/data': 0.0006651990115642548, 'timing/model': 1.9315529960440472, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 0.99}) +[2026-06-23 14:59:19,663][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.021881921216845512, 'train/grad_norm_pre_clip': 23.089372634887695, 'timing/data': 0.0006578390020877123, 'timing/model': 1.9307866599410772, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 0.99}) +[2026-06-23 14:59:23,462][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.04804655909538269, 'train/grad_norm_pre_clip': 33.425148010253906, 'timing/data': 0.0009213789599016309, 'timing/model': 1.9294467029394582, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 0.99}) +[2026-06-23 14:59:27,265][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.0713517889380455, 'train/grad_norm_pre_clip': 28.031978607177734, 'timing/data': 0.0006077649304643273, 'timing/model': 1.9308586829574779, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 0.99}) +[2026-06-23 14:59:31,107][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.12127760052680969, 'train/grad_norm_pre_clip': 28.862625122070312, 'timing/data': 0.0006203049561008811, 'timing/model': 1.9374085820745677, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 0.99}) +[2026-06-23 14:59:34,908][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.021951448172330856, 'train/grad_norm_pre_clip': 27.097614288330078, 'timing/data': 0.0006358419777825475, 'timing/model': 1.9290768479695544, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 0.99}) +[2026-06-23 14:59:38,704][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.04301406070590019, 'train/grad_norm_pre_clip': 16.21696662902832, 'timing/data': 0.0006011209916323423, 'timing/model': 1.9319223660277203, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 0.99}) +[2026-06-23 14:59:42,489][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.008494412526488304, 'train/grad_norm_pre_clip': 18.979368209838867, 'timing/data': 0.000660909921862185, 'timing/model': 1.9203847270691767, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 0.99}) +[2026-06-23 14:59:46,294][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.041105881333351135, 'train/grad_norm_pre_clip': 48.60781478881836, 'timing/data': 0.000662177917547524, 'timing/model': 1.9308662469265983, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 0.99}) +[2026-06-23 14:59:50,103][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.06699812412261963, 'train/grad_norm_pre_clip': 20.05767822265625, 'timing/data': 0.0007809040835127234, 'timing/model': 1.9266991600161418, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 0.99}) +[2026-06-23 14:59:53,901][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.05873342975974083, 'train/grad_norm_pre_clip': 34.89642333984375, 'timing/data': 0.0006136390147730708, 'timing/model': 1.9319014919456095, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 0.99}) +[2026-06-23 14:59:57,715][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.11633289605379105, 'train/grad_norm_pre_clip': 37.52760696411133, 'timing/data': 0.0006104859057813883, 'timing/model': 1.9366736380616203, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 0.99}) +[2026-06-23 15:00:01,521][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.0961218774318695, 'train/grad_norm_pre_clip': 20.03864097595215, 'timing/data': 0.0007018080214038491, 'timing/model': 1.9378411090001464, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 0.99}) +[2026-06-23 15:00:05,315][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.07473373413085938, 'train/grad_norm_pre_clip': 22.59193992614746, 'timing/data': 0.0005986469332128763, 'timing/model': 1.9223500801017508, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 0.99}) +[2026-06-23 15:00:09,126][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.07551456987857819, 'train/grad_norm_pre_clip': 36.25954055786133, 'timing/data': 0.000621977960690856, 'timing/model': 1.9429677050793543, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 0.99}) +[2026-06-23 15:00:12,928][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.0706891119480133, 'train/grad_norm_pre_clip': 33.28477096557617, 'timing/data': 0.0006003220332786441, 'timing/model': 1.926516390987672, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 0.99}) +[2026-06-23 15:00:16,731][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.11957839131355286, 'train/grad_norm_pre_clip': 23.329877853393555, 'timing/data': 0.0006569860270246863, 'timing/model': 1.9293099720962346, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 0.99}) +[2026-06-23 15:00:20,528][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.029000015929341316, 'train/grad_norm_pre_clip': 13.668548583984375, 'timing/data': 0.0007854050491005182, 'timing/model': 1.914133494021371, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 0.99}) +[2026-06-23 15:00:24,282][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.07506837695837021, 'train/grad_norm_pre_clip': 31.318466186523438, 'timing/data': 0.0007958200294524431, 'timing/model': 1.9050276890629902, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 0.99}) +[2026-06-23 15:00:28,036][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.1551034301519394, 'train/grad_norm_pre_clip': 41.63862228393555, 'timing/data': 0.0007401269394904375, 'timing/model': 1.9053737100912258, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 0.99}) +[2026-06-23 15:00:31,795][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.11707687377929688, 'train/grad_norm_pre_clip': 41.21491241455078, 'timing/data': 0.000607333960942924, 'timing/model': 1.9102620620978996, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 1.0}) +[2026-06-23 15:00:35,549][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.06572433561086655, 'train/grad_norm_pre_clip': 24.036672592163086, 'timing/data': 0.0006252860184758902, 'timing/model': 1.9051942229270935, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 1.0}) +[2026-06-23 15:00:39,304][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.02928844280540943, 'train/grad_norm_pre_clip': 35.05595779418945, 'timing/data': 0.0005853739567101002, 'timing/model': 1.9068392070475966, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 1.0}) +[2026-06-23 15:00:43,057][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.040640708059072495, 'train/grad_norm_pre_clip': 22.412687301635742, 'timing/data': 0.0006092529511079192, 'timing/model': 1.904668295988813, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 1.0}) +[2026-06-23 15:00:46,821][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.0718565508723259, 'train/grad_norm_pre_clip': 33.822052001953125, 'timing/data': 0.0007867040112614632, 'timing/model': 1.9145229910500348, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 1.0}) +[2026-06-23 15:00:50,597][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.14623981714248657, 'train/grad_norm_pre_clip': 49.02979278564453, 'timing/data': 0.0008515280205756426, 'timing/model': 1.9143710189964622, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 1.0}) +[2026-06-23 15:00:54,349][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.029977256432175636, 'train/grad_norm_pre_clip': 17.100419998168945, 'timing/data': 0.0007913259323686361, 'timing/model': 1.9036400349577889, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 1.0}) +[2026-06-23 15:00:58,103][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.16520154476165771, 'train/grad_norm_pre_clip': 38.79146194458008, 'timing/data': 0.000598698970861733, 'timing/model': 1.9048539289506152, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 1.0}) +[2026-06-23 15:01:01,859][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.060625653713941574, 'train/grad_norm_pre_clip': 29.217491149902344, 'timing/data': 0.0006066869245842099, 'timing/model': 1.9048714661039412, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 1.0}) +[2026-06-23 15:01:05,624][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.03238657861948013, 'train/grad_norm_pre_clip': 17.09914779663086, 'timing/data': 0.0006305680144578218, 'timing/model': 1.910723458044231, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 1.0}) +[2026-06-23 15:01:09,382][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.030695242807269096, 'train/grad_norm_pre_clip': 19.418169021606445, 'timing/data': 0.0006236110348254442, 'timing/model': 1.906881742994301, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 1.0}) +[2026-06-23 15:01:13,146][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.04423077404499054, 'train/grad_norm_pre_clip': 23.141616821289062, 'timing/data': 0.0006159750046208501, 'timing/model': 1.9054074989398941, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 1.0}) +[2026-06-23 15:01:16,910][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.06314408779144287, 'train/grad_norm_pre_clip': 22.249971389770508, 'timing/data': 0.0007987970020622015, 'timing/model': 1.9113035160116851, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 1.0}) +[2026-06-23 15:01:20,074][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.1225753128528595, 'train/grad_norm_pre_clip': 40.06114959716797, 'timing/data': 0.0005138569977134466, 'timing/model': 1.3118603309849277, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 1.0}) +[2026-06-23 15:01:30,097][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.042339276522397995, 'train/grad_norm_pre_clip': 23.670839309692383, 'timing/data': 0.0035529419546946883, 'timing/model': 2.308950514998287, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 1.0}) +[2026-06-23 15:01:34,714][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.05081567540764809, 'train/grad_norm_pre_clip': 16.839420318603516, 'timing/data': 0.0019367350032553077, 'timing/model': 2.3884142549941316, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 1.0}) +[2026-06-23 15:01:39,259][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.048056844621896744, 'train/grad_norm_pre_clip': 18.676525115966797, 'timing/data': 0.002710305037908256, 'timing/model': 2.204653925029561, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 1.0}) +[2026-06-23 15:01:43,873][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.033470455557107925, 'train/grad_norm_pre_clip': 20.530099868774414, 'timing/data': 0.0022662540432065725, 'timing/model': 2.4062603019410744, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 1.0}) +[2026-06-23 15:01:48,903][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.0681413784623146, 'train/grad_norm_pre_clip': 15.537830352783203, 'timing/data': 0.0023700540186837316, 'timing/model': 2.6759262969717383, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 1.0}) +[2026-06-23 15:01:52,756][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.03374423459172249, 'train/grad_norm_pre_clip': 30.53655242919922, 'timing/data': 0.000670114066451788, 'timing/model': 1.9247978508938104, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 1.0}) +[2026-06-23 15:01:56,546][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.019331805408000946, 'train/grad_norm_pre_clip': 45.17554473876953, 'timing/data': 0.0006213909946382046, 'timing/model': 1.9268413089448586, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 1.0}) +[2026-06-23 15:02:00,338][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.04092339798808098, 'train/grad_norm_pre_clip': 25.77610206604004, 'timing/data': 0.0006758479867130518, 'timing/model': 1.9290063480148092, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 1.0}) +[2026-06-23 15:02:04,127][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.12923403084278107, 'train/grad_norm_pre_clip': 57.09962844848633, 'timing/data': 0.0006508539663627744, 'timing/model': 1.922192609985359, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 1.0}) +[2026-06-23 15:02:07,920][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.03738407418131828, 'train/grad_norm_pre_clip': 25.978639602661133, 'timing/data': 0.0005714610451832414, 'timing/model': 1.9238000300247222, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 1.0}) +[2026-06-23 15:02:11,725][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.022280573844909668, 'train/grad_norm_pre_clip': 32.5632209777832, 'timing/data': 0.0006109370151534677, 'timing/model': 1.9346409509889781, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 1.0}) +[2026-06-23 15:02:15,517][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.015366862528026104, 'train/grad_norm_pre_clip': 10.778799057006836, 'timing/data': 0.0005928620230406523, 'timing/model': 1.9161846740171313, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 1.0}) +[2026-06-23 15:02:19,306][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.07391233742237091, 'train/grad_norm_pre_clip': 29.977279663085938, 'timing/data': 0.0008711270056664944, 'timing/model': 1.9132594839902595, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 1.01}) +[2026-06-23 15:02:23,126][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.02599288709461689, 'train/grad_norm_pre_clip': 38.3769645690918, 'timing/data': 0.0005533290095627308, 'timing/model': 1.9259886409854516, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 1.01}) +[2026-06-23 15:02:26,908][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.09180166572332382, 'train/grad_norm_pre_clip': 44.204830169677734, 'timing/data': 0.0005729250842705369, 'timing/model': 1.9214288200018927, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.01}) +[2026-06-23 15:02:30,724][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2600 +[2026-06-23 15:04:37,922][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2600 +[2026-06-23 15:04:37,923][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2600 +[2026-06-23 15:06:00,728][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2600 +[2026-06-23 15:06:00,729][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.07588141411542892, 'train/grad_norm_pre_clip': 32.417259216308594, 'eval/action_loss/seconds': 127.19560698908754, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03705192729830742, 'eval/latency_2/loss': 0.03705192729830742, 'eval/flappy/loss': 0.03705192729830742, 'eval/flappy/latency_2/loss': 0.03705192729830742, 'eval/flap_precision': 0.9630127774041695, 'eval/flap_recall': 0.9662618083670715, 'eval/flap_f1': 0.9646345570899293, 'eval/flap_rate_model': 0.20652777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9646345570899293, 'eval/latency_2/flap_precision': 0.9630127774041695, 'eval/latency_2/flap_recall': 0.9662618083670715, 'eval/latency_2/flap_f1': 0.9646345570899293, 'eval/latency_2/flap_rate_model': 0.20652777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9646345570899293, 'eval/action_classification/data_wait_seconds': 1.751070261001587, 'eval/action_classification/predict_seconds': 80.88145446777344, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005786769324913621, 'timing/model': 1.9460869770264253, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 1.01}) +[2026-06-23 15:06:04,501][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.13184621930122375, 'train/grad_norm_pre_clip': 35.62077331542969, 'timing/data': 0.0007855240255594254, 'timing/model': 1.9165447839768603, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 1.01}) +[2026-06-23 15:06:08,273][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.030267931520938873, 'train/grad_norm_pre_clip': 19.483938217163086, 'timing/data': 0.0005480030085891485, 'timing/model': 1.9098977949470282, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 1.01}) +[2026-06-23 15:06:12,042][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.09172643721103668, 'train/grad_norm_pre_clip': 41.374900817871094, 'timing/data': 0.0006596769671887159, 'timing/model': 1.9090358120156452, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 1.01}) +[2026-06-23 15:06:15,827][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.07599443197250366, 'train/grad_norm_pre_clip': 71.28163146972656, 'timing/data': 0.0005559140117838979, 'timing/model': 1.9201999839860946, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 1.01}) +[2026-06-23 15:06:19,608][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.10079115629196167, 'train/grad_norm_pre_clip': 79.74076080322266, 'timing/data': 0.0005713959690183401, 'timing/model': 1.9169901079731062, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 1.01}) +[2026-06-23 15:06:23,404][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.11360316723585129, 'train/grad_norm_pre_clip': 70.59736633300781, 'timing/data': 0.0007976919878274202, 'timing/model': 1.927510389010422, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 1.01}) +[2026-06-23 15:06:27,186][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.11643870174884796, 'train/grad_norm_pre_clip': 91.68634033203125, 'timing/data': 0.000627067987807095, 'timing/model': 1.9196795569732785, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 1.01}) +[2026-06-23 15:06:30,977][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.03409275412559509, 'train/grad_norm_pre_clip': 67.41144561767578, 'timing/data': 0.0005934940418228507, 'timing/model': 1.9292950549861416, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 1.01}) +[2026-06-23 15:06:34,759][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.045325133949518204, 'train/grad_norm_pre_clip': 17.858301162719727, 'timing/data': 0.0005797450430691242, 'timing/model': 1.9151056749979034, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 1.01}) +[2026-06-23 15:06:38,555][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.07612618058919907, 'train/grad_norm_pre_clip': 25.82048797607422, 'timing/data': 0.0005547610344365239, 'timing/model': 1.924418449983932, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 1.01}) +[2026-06-23 15:06:42,331][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.05429627373814583, 'train/grad_norm_pre_clip': 19.30911636352539, 'timing/data': 0.0005951999919489026, 'timing/model': 1.91194808203727, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 1.01}) +[2026-06-23 15:06:46,125][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.0769808366894722, 'train/grad_norm_pre_clip': 31.345081329345703, 'timing/data': 0.0008080289699137211, 'timing/model': 1.9278958620270714, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 1.01}) +[2026-06-23 15:06:49,928][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.14500799775123596, 'train/grad_norm_pre_clip': 39.77103042602539, 'timing/data': 0.0005771600408479571, 'timing/model': 1.9272219419945031, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 1.01}) +[2026-06-23 15:06:53,740][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.13709771633148193, 'train/grad_norm_pre_clip': 24.033212661743164, 'timing/data': 0.0005733720026910305, 'timing/model': 1.9417324910173193, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 1.01}) +[2026-06-23 15:06:57,567][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.029678665101528168, 'train/grad_norm_pre_clip': 21.362863540649414, 'timing/data': 0.0006003299495205283, 'timing/model': 1.917656707111746, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 1.01}) +[2026-06-23 15:07:01,378][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.1050046756863594, 'train/grad_norm_pre_clip': 29.777624130249023, 'timing/data': 0.0006252079037949443, 'timing/model': 1.9328862000256777, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 1.01}) +[2026-06-23 15:07:05,207][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.020412679761648178, 'train/grad_norm_pre_clip': 15.379547119140625, 'timing/data': 0.0005707331001758575, 'timing/model': 1.9265677459770814, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 1.01}) +[2026-06-23 15:07:09,006][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.057239942252635956, 'train/grad_norm_pre_clip': 19.738737106323242, 'timing/data': 0.0005845309933647513, 'timing/model': 1.9309253599494696, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 1.01}) +[2026-06-23 15:07:12,820][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.07673139870166779, 'train/grad_norm_pre_clip': 32.280059814453125, 'timing/data': 0.0008331240387633443, 'timing/model': 1.93895279802382, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 1.01}) +[2026-06-23 15:07:16,623][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.04757751524448395, 'train/grad_norm_pre_clip': 23.84785270690918, 'timing/data': 0.0006500609451904893, 'timing/model': 1.9246557879960164, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 1.01}) +[2026-06-23 15:07:20,428][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.07920857518911362, 'train/grad_norm_pre_clip': 25.51324462890625, 'timing/data': 0.0006236700573936105, 'timing/model': 1.93659519799985, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 1.01}) +[2026-06-23 15:07:24,229][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.06419312208890915, 'train/grad_norm_pre_clip': 24.82293128967285, 'timing/data': 0.0006490970263257623, 'timing/model': 1.9330122669925913, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 1.01}) +[2026-06-23 15:07:28,031][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.0730937123298645, 'train/grad_norm_pre_clip': 22.363847732543945, 'timing/data': 0.0005816880147904158, 'timing/model': 1.9279453899944201, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 1.02}) +[2026-06-23 15:07:31,840][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.04702620953321457, 'train/grad_norm_pre_clip': 27.960384368896484, 'timing/data': 0.0005840329686179757, 'timing/model': 1.9234092590631917, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 1.02}) +[2026-06-23 15:07:35,644][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.0532006211578846, 'train/grad_norm_pre_clip': 20.42892837524414, 'timing/data': 0.0006334519712254405, 'timing/model': 1.9232016790192574, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 1.02}) +[2026-06-23 15:07:39,447][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.09273599088191986, 'train/grad_norm_pre_clip': 25.913021087646484, 'timing/data': 0.0006241750670596957, 'timing/model': 1.9387522779870778, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 1.02}) +[2026-06-23 15:07:43,253][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.022006286308169365, 'train/grad_norm_pre_clip': 26.9503173828125, 'timing/data': 0.0006668280111625791, 'timing/model': 1.924731417093426, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 1.02}) +[2026-06-23 15:07:47,252][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.1137523204088211, 'train/grad_norm_pre_clip': 33.07510757446289, 'timing/data': 0.0005906090373173356, 'timing/model': 2.1095044310204685, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 1.02}) +[2026-06-23 15:07:51,050][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.045241475105285645, 'train/grad_norm_pre_clip': 16.855043411254883, 'timing/data': 0.0005947720492258668, 'timing/model': 1.9343706449726596, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 1.02}) +[2026-06-23 15:07:54,858][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.07639363408088684, 'train/grad_norm_pre_clip': 24.503376007080078, 'timing/data': 0.0005900880787521601, 'timing/model': 1.9290440250188112, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 1.02}) +[2026-06-23 15:07:58,649][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.025309506803750992, 'train/grad_norm_pre_clip': 24.160388946533203, 'timing/data': 0.000676720985211432, 'timing/model': 1.9234825359890237, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 1.02}) +[2026-06-23 15:08:02,456][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.05555399879813194, 'train/grad_norm_pre_clip': 15.680782318115234, 'timing/data': 0.0010681570274755359, 'timing/model': 1.9428743349853903, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 1.02}) +[2026-06-23 15:08:06,277][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.033364392817020416, 'train/grad_norm_pre_clip': 22.14177131652832, 'timing/data': 0.0005820520455017686, 'timing/model': 1.9428813860286027, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 1.02}) +[2026-06-23 15:08:10,070][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.01667010597884655, 'train/grad_norm_pre_clip': 25.522918701171875, 'timing/data': 0.0006280801026150584, 'timing/model': 1.920044341008179, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 1.02}) +[2026-06-23 15:08:13,859][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.03181608393788338, 'train/grad_norm_pre_clip': 50.12385177612305, 'timing/data': 0.0005749169504269958, 'timing/model': 1.9187116569373757, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 1.02}) +[2026-06-23 15:08:17,676][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.12434786558151245, 'train/grad_norm_pre_clip': 29.520671844482422, 'timing/data': 0.0007173490012064576, 'timing/model': 1.934802328934893, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 1.02}) +[2026-06-23 15:08:21,481][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.07590074837207794, 'train/grad_norm_pre_clip': 24.467195510864258, 'timing/data': 0.0006695740157738328, 'timing/model': 1.922725954093039, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 1.02}) +[2026-06-23 15:08:25,282][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.07988373935222626, 'train/grad_norm_pre_clip': 60.663909912109375, 'timing/data': 0.000583346001803875, 'timing/model': 1.9348120159702376, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 1.02}) +[2026-06-23 15:08:29,091][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.12820178270339966, 'train/grad_norm_pre_clip': 78.88346099853516, 'timing/data': 0.0005950239719823003, 'timing/model': 1.9274248289875686, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 1.02}) +[2026-06-23 15:08:32,895][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.09980738908052444, 'train/grad_norm_pre_clip': 53.405582427978516, 'timing/data': 0.0005845420528203249, 'timing/model': 1.9309111289912835, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 1.02}) +[2026-06-23 15:08:36,693][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.08206743746995926, 'train/grad_norm_pre_clip': 28.927412033081055, 'timing/data': 0.0005681479815393686, 'timing/model': 1.9291072930209339, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 1.02}) +[2026-06-23 15:08:40,496][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.04403930529952049, 'train/grad_norm_pre_clip': 20.730276107788086, 'timing/data': 0.0008341309148818254, 'timing/model': 1.9312914090696722, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 1.02}) +[2026-06-23 15:08:44,313][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.027036909013986588, 'train/grad_norm_pre_clip': 24.64487648010254, 'timing/data': 0.0005865490529686213, 'timing/model': 1.936452616006136, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 1.02}) +[2026-06-23 15:08:48,115][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.04498322308063507, 'train/grad_norm_pre_clip': 33.892120361328125, 'timing/data': 0.0006160999182611704, 'timing/model': 1.9355767259839922, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 1.02}) +[2026-06-23 15:08:51,939][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.06688249111175537, 'train/grad_norm_pre_clip': 18.1911678314209, 'timing/data': 0.0006265820702537894, 'timing/model': 1.9470587170217186, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 1.02}) +[2026-06-23 15:08:55,736][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.036209795624017715, 'train/grad_norm_pre_clip': 59.48508834838867, 'timing/data': 0.00069289596285671, 'timing/model': 1.9314481910550967, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 1.02}) +[2026-06-23 15:08:59,541][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.11390823870897293, 'train/grad_norm_pre_clip': 30.729747772216797, 'timing/data': 0.0006116270087659359, 'timing/model': 1.930670596077107, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 1.02}) +[2026-06-23 15:09:03,341][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.0871039628982544, 'train/grad_norm_pre_clip': 28.40517807006836, 'timing/data': 0.0006755120120942593, 'timing/model': 1.9301794270286337, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.02}) +[2026-06-23 15:09:07,148][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.11817742884159088, 'train/grad_norm_pre_clip': 36.056339263916016, 'timing/data': 0.0007559619843959808, 'timing/model': 1.930021160049364, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 1.03}) +[2026-06-23 15:09:10,954][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.009647143073379993, 'train/grad_norm_pre_clip': 36.354576110839844, 'timing/data': 0.0007350090891122818, 'timing/model': 1.9306959370151162, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 1.03}) +[2026-06-23 15:09:14,759][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.01725730486214161, 'train/grad_norm_pre_clip': 13.822152137756348, 'timing/data': 0.0005706839729100466, 'timing/model': 1.941702431999147, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 1.03}) +[2026-06-23 15:09:18,587][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.05960238724946976, 'train/grad_norm_pre_clip': 26.46410369873047, 'timing/data': 0.0005951259518042207, 'timing/model': 1.9551924329716712, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 1.03}) +[2026-06-23 15:09:22,394][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.10824541002511978, 'train/grad_norm_pre_clip': 50.21758270263672, 'timing/data': 0.0007797620492056012, 'timing/model': 1.9328431690810248, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 1.03}) +[2026-06-23 15:09:26,395][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.020927738398313522, 'train/grad_norm_pre_clip': 11.144006729125977, 'timing/data': 0.0005967740435153246, 'timing/model': 1.9210154190659523, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 1.03}) +[2026-06-23 15:09:30,202][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.2919011414051056, 'train/grad_norm_pre_clip': 76.25537109375, 'timing/data': 0.0005716810701414943, 'timing/model': 1.9248044659616426, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 1.03}) +[2026-06-23 15:09:33,993][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.020926354452967644, 'train/grad_norm_pre_clip': 17.573183059692383, 'timing/data': 0.0005867029540240765, 'timing/model': 1.9268164259847254, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 1.03}) +[2026-06-23 15:09:37,803][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.08801347762346268, 'train/grad_norm_pre_clip': 17.738584518432617, 'timing/data': 0.0005749730626121163, 'timing/model': 1.9306632219813764, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 1.03}) +[2026-06-23 15:09:41,611][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.1116846427321434, 'train/grad_norm_pre_clip': 28.869197845458984, 'timing/data': 0.00057307502720505, 'timing/model': 1.9329803580185398, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 1.03}) +[2026-06-23 15:09:45,422][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.04924364760518074, 'train/grad_norm_pre_clip': 19.477252960205078, 'timing/data': 0.0007949179271236062, 'timing/model': 1.933489890070632, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 1.03}) +[2026-06-23 15:09:49,231][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.09545263648033142, 'train/grad_norm_pre_clip': 35.33034133911133, 'timing/data': 0.0005806130357086658, 'timing/model': 1.9370427880203351, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 1.03}) +[2026-06-23 15:09:53,043][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.07580612599849701, 'train/grad_norm_pre_clip': 25.95608901977539, 'timing/data': 0.0006069389637559652, 'timing/model': 1.9383421699749306, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 1.03}) +[2026-06-23 15:09:57,047][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.1157090961933136, 'train/grad_norm_pre_clip': 45.313846588134766, 'timing/data': 0.0005784120876342058, 'timing/model': 2.121685422025621, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 1.03}) +[2026-06-23 15:10:00,861][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.09479058533906937, 'train/grad_norm_pre_clip': 28.28649139404297, 'timing/data': 0.0006441329605877399, 'timing/model': 1.9407629629131407, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 1.03}) +[2026-06-23 15:10:04,667][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.07436257600784302, 'train/grad_norm_pre_clip': 24.088621139526367, 'timing/data': 0.0005762859946116805, 'timing/model': 1.9425847160164267, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 1.03}) +[2026-06-23 15:10:08,459][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.04830103740096092, 'train/grad_norm_pre_clip': 26.605369567871094, 'timing/data': 0.0005833710310980678, 'timing/model': 1.9244816639693454, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 1.03}) +[2026-06-23 15:10:12,263][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.08597517758607864, 'train/grad_norm_pre_clip': 21.241453170776367, 'timing/data': 0.00076162104960531, 'timing/model': 1.9247942549409345, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 1.03}) +[2026-06-23 15:10:16,065][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.11087945848703384, 'train/grad_norm_pre_clip': 43.14741134643555, 'timing/data': 0.0005893299821764231, 'timing/model': 1.9309351110132411, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 1.03}) +[2026-06-23 15:10:19,859][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.06343657523393631, 'train/grad_norm_pre_clip': 16.687891006469727, 'timing/data': 0.0005954679800197482, 'timing/model': 1.9240028540370986, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 1.03}) +[2026-06-23 15:10:23,669][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.08478914946317673, 'train/grad_norm_pre_clip': 59.02294921875, 'timing/data': 0.0006238160422071815, 'timing/model': 1.9375802690628916, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 1.03}) +[2026-06-23 15:10:27,482][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.13147440552711487, 'train/grad_norm_pre_clip': 38.9395866394043, 'timing/data': 0.0005795019678771496, 'timing/model': 1.943732272950001, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 1.03}) +[2026-06-23 15:10:31,290][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.08555126935243607, 'train/grad_norm_pre_clip': 24.64490509033203, 'timing/data': 0.0006356709636747837, 'timing/model': 1.9402309840079397, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 1.03}) +[2026-06-23 15:10:35,165][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.04381595924496651, 'train/grad_norm_pre_clip': 29.1097412109375, 'timing/data': 0.000570531003177166, 'timing/model': 2.0018426480237395, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 1.03}) +[2026-06-23 15:10:38,976][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.13259167969226837, 'train/grad_norm_pre_clip': 33.73310089111328, 'timing/data': 0.0005896480288356543, 'timing/model': 1.9254075259668753, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 1.03}) +[2026-06-23 15:10:42,780][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.01830587163567543, 'train/grad_norm_pre_clip': 49.80726623535156, 'timing/data': 0.0005962960422039032, 'timing/model': 1.9267504889285192, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 1.03}) +[2026-06-23 15:10:46,608][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.03135824576020241, 'train/grad_norm_pre_clip': 12.910662651062012, 'timing/data': 0.0005880569806322455, 'timing/model': 1.9335100339958444, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 1.04}) +[2026-06-23 15:10:50,423][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.09515077620744705, 'train/grad_norm_pre_clip': 46.837745666503906, 'timing/data': 0.0005881499964743853, 'timing/model': 1.930521106114611, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 1.04}) +[2026-06-23 15:10:54,233][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.019119471311569214, 'train/grad_norm_pre_clip': 25.660322189331055, 'timing/data': 0.0005939099937677383, 'timing/model': 1.933179370011203, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 1.04}) +[2026-06-23 15:10:58,046][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.13635870814323425, 'train/grad_norm_pre_clip': 59.267574310302734, 'timing/data': 0.0005932459607720375, 'timing/model': 1.9381120309699327, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 1.04}) +[2026-06-23 15:11:01,851][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.03326967731118202, 'train/grad_norm_pre_clip': 44.945499420166016, 'timing/data': 0.0005807080306112766, 'timing/model': 1.9363803369924426, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 1.04}) +[2026-06-23 15:11:05,655][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.21969418227672577, 'train/grad_norm_pre_clip': 51.517845153808594, 'timing/data': 0.0007625010330229998, 'timing/model': 1.938949535950087, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 1.04}) +[2026-06-23 15:11:09,472][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.10496900975704193, 'train/grad_norm_pre_clip': 24.695486068725586, 'timing/data': 0.0005926440935581923, 'timing/model': 1.938083051936701, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 1.04}) +[2026-06-23 15:11:13,294][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.06063136085867882, 'train/grad_norm_pre_clip': 30.459760665893555, 'timing/data': 0.0005919799441471696, 'timing/model': 1.937388867023401, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 1.04}) +[2026-06-23 15:11:17,117][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.1000136137008667, 'train/grad_norm_pre_clip': 22.710329055786133, 'timing/data': 0.0005822359817102551, 'timing/model': 1.9209609850076959, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 1.04}) +[2026-06-23 15:11:20,922][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.08698122203350067, 'train/grad_norm_pre_clip': 19.42618751525879, 'timing/data': 0.000785224954597652, 'timing/model': 1.9298714300384745, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 1.04}) +[2026-06-23 15:11:24,731][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.05490323156118393, 'train/grad_norm_pre_clip': 19.66187286376953, 'timing/data': 0.0005833678878843784, 'timing/model': 1.9256948119727895, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 1.04}) +[2026-06-23 15:11:28,531][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.029611190780997276, 'train/grad_norm_pre_clip': 59.31147766113281, 'timing/data': 0.0005899389507248998, 'timing/model': 1.9214936119969934, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 1.04}) +[2026-06-23 15:11:32,415][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.058492064476013184, 'train/grad_norm_pre_clip': 18.280359268188477, 'timing/data': 0.0006376279052346945, 'timing/model': 1.9301532610552385, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 1.04}) +[2026-06-23 15:11:36,229][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.040190309286117554, 'train/grad_norm_pre_clip': 17.836017608642578, 'timing/data': 0.0005846340209245682, 'timing/model': 1.9367476491024718, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 1.04}) +[2026-06-23 15:11:40,069][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.04030395671725273, 'train/grad_norm_pre_clip': 32.21195983886719, 'timing/data': 0.0006107799708843231, 'timing/model': 1.928873839089647, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 1.04}) +[2026-06-23 15:11:43,878][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.06972825527191162, 'train/grad_norm_pre_clip': 31.74766731262207, 'timing/data': 0.0006841400172561407, 'timing/model': 1.9308798880083486, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 1.04}) +[2026-06-23 15:11:47,680][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.035420291125774384, 'train/grad_norm_pre_clip': 20.11491584777832, 'timing/data': 0.0005852399626746774, 'timing/model': 1.928176631918177, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 1.04}) +[2026-06-23 15:11:51,480][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.02522967755794525, 'train/grad_norm_pre_clip': 40.572418212890625, 'timing/data': 0.0005890559405088425, 'timing/model': 1.9281304209725931, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 1.04}) +[2026-06-23 15:11:55,280][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.10530740767717361, 'train/grad_norm_pre_clip': 33.55270767211914, 'timing/data': 0.0005921869305893779, 'timing/model': 1.9277305779978633, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 1.04}) +[2026-06-23 15:11:59,080][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.07900502532720566, 'train/grad_norm_pre_clip': 36.13875198364258, 'timing/data': 0.0005810640286654234, 'timing/model': 1.9291461739921942, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 1.04}) +[2026-06-23 15:12:02,889][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.005846654064953327, 'train/grad_norm_pre_clip': 33.432281494140625, 'timing/data': 0.0008245029021054506, 'timing/model': 1.9327497760532424, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 1.04}) +[2026-06-23 15:12:06,706][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.01004685927182436, 'train/grad_norm_pre_clip': 30.59597396850586, 'timing/data': 0.0005859399680048227, 'timing/model': 1.9378506960347295, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 1.04}) +[2026-06-23 15:12:10,693][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.026355601847171783, 'train/grad_norm_pre_clip': 23.084562301635742, 'timing/data': 0.000644739018753171, 'timing/model': 1.9346155510284007, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.04}) +[2026-06-23 15:12:14,500][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.15530045330524445, 'train/grad_norm_pre_clip': 48.25101852416992, 'timing/data': 0.0007710880599915981, 'timing/model': 1.9312258339487016, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 1.04}) +[2026-06-23 15:12:18,306][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.039279755204916, 'train/grad_norm_pre_clip': 15.06112289428711, 'timing/data': 0.0005991070065647364, 'timing/model': 1.9305180490482599, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 1.04}) +[2026-06-23 15:12:22,113][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2700 +[2026-06-23 15:14:29,702][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2700 +[2026-06-23 15:14:29,703][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2700 +[2026-06-23 15:15:52,514][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2700 +[2026-06-23 15:15:52,515][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.08747249841690063, 'train/grad_norm_pre_clip': 42.040767669677734, 'eval/action_loss/seconds': 127.58640000096057, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04143262282013893, 'eval/latency_2/loss': 0.04143262282013893, 'eval/flappy/loss': 0.04143262282013893, 'eval/flappy/latency_2/loss': 0.04143262282013893, 'eval/flap_precision': 0.9387886597938144, 'eval/flap_recall': 0.9831309041835358, 'eval/flap_f1': 0.96044825313118, 'eval/flap_rate_model': 0.21555555555555556, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.96044825313118, 'eval/latency_2/flap_precision': 0.9387886597938144, 'eval/latency_2/flap_recall': 0.9831309041835358, 'eval/latency_2/flap_f1': 0.96044825313118, 'eval/latency_2/flap_rate_model': 0.21555555555555556, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.96044825313118, 'eval/action_classification/data_wait_seconds': 1.6094822883605957, 'eval/action_classification/predict_seconds': 80.95661926269531, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005793130258098245, 'timing/model': 1.9332823210861534, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 1.04}) +[2026-06-23 15:15:56,467][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.048632748425006866, 'train/grad_norm_pre_clip': 75.86383819580078, 'timing/data': 0.000779353897087276, 'timing/model': 2.0991519211092964, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 1.05}) +[2026-06-23 15:16:00,253][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.013931822031736374, 'train/grad_norm_pre_clip': 7.139729976654053, 'timing/data': 0.0005820079240947962, 'timing/model': 1.9175101589644328, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 1.05}) +[2026-06-23 15:16:04,051][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.09206242114305496, 'train/grad_norm_pre_clip': 56.76734924316406, 'timing/data': 0.0005830059526488185, 'timing/model': 1.9304881090065464, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 1.05}) +[2026-06-23 15:16:07,833][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.03803115338087082, 'train/grad_norm_pre_clip': 27.87617301940918, 'timing/data': 0.0005936330417171121, 'timing/model': 1.9100023780483752, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 1.05}) +[2026-06-23 15:16:11,615][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.03247561305761337, 'train/grad_norm_pre_clip': 24.234960556030273, 'timing/data': 0.0005758010083809495, 'timing/model': 1.924670230015181, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 1.05}) +[2026-06-23 15:16:15,380][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.09820229560136795, 'train/grad_norm_pre_clip': 29.585832595825195, 'timing/data': 0.0005793230375275016, 'timing/model': 1.911642840015702, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 1.05}) +[2026-06-23 15:16:19,192][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.11047839373350143, 'train/grad_norm_pre_clip': 30.019948959350586, 'timing/data': 0.0006040739826858044, 'timing/model': 1.9376661310670897, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 1.05}) +[2026-06-23 15:16:22,974][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.026541467756032944, 'train/grad_norm_pre_clip': 34.31608581542969, 'timing/data': 0.0007754599209874868, 'timing/model': 1.9189077850896865, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 1.05}) +[2026-06-23 15:16:26,765][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.06340736895799637, 'train/grad_norm_pre_clip': 33.86671447753906, 'timing/data': 0.0006464340258389711, 'timing/model': 1.9346797720063478, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 1.05}) +[2026-06-23 15:16:30,566][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.1452207863330841, 'train/grad_norm_pre_clip': 100.30397033691406, 'timing/data': 0.0006000350695103407, 'timing/model': 1.9312255480326712, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 1.05}) +[2026-06-23 15:16:34,360][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.004515251610428095, 'train/grad_norm_pre_clip': 40.35708999633789, 'timing/data': 0.0005794259486719966, 'timing/model': 1.9229498719796538, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 1.05}) +[2026-06-23 15:16:38,129][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.04790117219090462, 'train/grad_norm_pre_clip': 21.033002853393555, 'timing/data': 0.0005888449959456921, 'timing/model': 1.9125997250666842, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 1.05}) +[2026-06-23 15:16:41,919][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.14726056158542633, 'train/grad_norm_pre_clip': 49.262332916259766, 'timing/data': 0.0005760829662904143, 'timing/model': 1.9229073679307476, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 1.05}) +[2026-06-23 15:16:45,707][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.01596686989068985, 'train/grad_norm_pre_clip': 11.119884490966797, 'timing/data': 0.0006293620681390166, 'timing/model': 1.9207524040248245, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 1.05}) +[2026-06-23 15:16:49,496][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.06785611063241959, 'train/grad_norm_pre_clip': 48.96501922607422, 'timing/data': 0.0006588558899238706, 'timing/model': 1.9203579580644146, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 1.05}) +[2026-06-23 15:16:53,283][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.020367702469229698, 'train/grad_norm_pre_clip': 49.38080596923828, 'timing/data': 0.0005961590213701129, 'timing/model': 1.922549964976497, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 1.05}) +[2026-06-23 15:16:57,068][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.08966609835624695, 'train/grad_norm_pre_clip': 71.0257339477539, 'timing/data': 0.0006010159850120544, 'timing/model': 1.9176203679526225, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 1.05}) +[2026-06-23 15:17:00,860][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.08951177448034286, 'train/grad_norm_pre_clip': 31.867128372192383, 'timing/data': 0.000827345997095108, 'timing/model': 1.9219213559990749, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 1.05}) +[2026-06-23 15:17:04,668][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.08267948776483536, 'train/grad_norm_pre_clip': 50.46392059326172, 'timing/data': 0.0006001419387757778, 'timing/model': 1.9306132880738005, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 1.05}) +[2026-06-23 15:17:08,472][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.006477927789092064, 'train/grad_norm_pre_clip': 20.185976028442383, 'timing/data': 0.0006685289554297924, 'timing/model': 1.941843911074102, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 1.05}) +[2026-06-23 15:17:12,274][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.018733885139226913, 'train/grad_norm_pre_clip': 21.19625473022461, 'timing/data': 0.0005811910377815366, 'timing/model': 1.9314409039216116, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 1.05}) +[2026-06-23 15:17:16,082][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.031241916120052338, 'train/grad_norm_pre_clip': 25.247488021850586, 'timing/data': 0.0005782309453934431, 'timing/model': 1.927335620042868, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 1.05}) +[2026-06-23 15:17:19,880][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.054707471281290054, 'train/grad_norm_pre_clip': 12.890284538269043, 'timing/data': 0.0006806739838793874, 'timing/model': 1.927224423037842, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 1.05}) +[2026-06-23 15:17:23,689][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.04435155913233757, 'train/grad_norm_pre_clip': 26.09214973449707, 'timing/data': 0.0005791620351374149, 'timing/model': 1.9354542839573696, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 1.05}) +[2026-06-23 15:17:27,492][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.13740818202495575, 'train/grad_norm_pre_clip': 55.706016540527344, 'timing/data': 0.0009202109649777412, 'timing/model': 1.930984259932302, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 1.05}) +[2026-06-23 15:17:31,290][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.05549095198512077, 'train/grad_norm_pre_clip': 25.510818481445312, 'timing/data': 0.0006495979614555836, 'timing/model': 1.9369962509954348, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 1.06}) +[2026-06-23 15:17:35,284][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.07526315003633499, 'train/grad_norm_pre_clip': 25.06916046142578, 'timing/data': 0.0005925219738855958, 'timing/model': 2.1225737320492044, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 1.06}) +[2026-06-23 15:17:39,090][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.03470119088888168, 'train/grad_norm_pre_clip': 37.71473693847656, 'timing/data': 0.0006640719948336482, 'timing/model': 1.941036578034982, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 1.06}) +[2026-06-23 15:17:42,873][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.12159672379493713, 'train/grad_norm_pre_clip': 46.3718376159668, 'timing/data': 0.00060396792832762, 'timing/model': 1.9232850620755926, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 1.06}) +[2026-06-23 15:17:46,668][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.05721627548336983, 'train/grad_norm_pre_clip': 20.29900360107422, 'timing/data': 0.0006008320488035679, 'timing/model': 1.9308203499531373, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 1.06}) +[2026-06-23 15:17:50,465][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.051069606095552444, 'train/grad_norm_pre_clip': 29.317241668701172, 'timing/data': 0.0006667299894616008, 'timing/model': 1.9382017420139164, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 1.06}) +[2026-06-23 15:17:54,260][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.09131142497062683, 'train/grad_norm_pre_clip': 30.657821655273438, 'timing/data': 0.0005939510883763433, 'timing/model': 1.9295104750199243, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 1.06}) +[2026-06-23 15:17:58,064][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.06039530411362648, 'train/grad_norm_pre_clip': 34.72232437133789, 'timing/data': 0.0006040689768269658, 'timing/model': 1.9324554189806804, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 1.06}) +[2026-06-23 15:18:01,851][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.055219244211912155, 'train/grad_norm_pre_clip': 17.94629669189453, 'timing/data': 0.0005806280532851815, 'timing/model': 1.9155254299985245, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 1.06}) +[2026-06-23 15:18:05,636][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.030094224959611893, 'train/grad_norm_pre_clip': 20.196853637695312, 'timing/data': 0.0005943920696154237, 'timing/model': 1.9213059920584783, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 1.06}) +[2026-06-23 15:18:09,434][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.015446803532540798, 'train/grad_norm_pre_clip': 12.245030403137207, 'timing/data': 0.0005880219396203756, 'timing/model': 1.9253948481054977, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 1.06}) +[2026-06-23 15:18:13,237][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.11302157491445541, 'train/grad_norm_pre_clip': 21.45018768310547, 'timing/data': 0.0005968180485069752, 'timing/model': 1.930488743004389, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 1.06}) +[2026-06-23 15:18:17,036][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.12039169669151306, 'train/grad_norm_pre_clip': 38.10576629638672, 'timing/data': 0.0005733160069212317, 'timing/model': 1.9274835829855874, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 1.06}) +[2026-06-23 15:18:20,841][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.0875530019402504, 'train/grad_norm_pre_clip': 36.900367736816406, 'timing/data': 0.0005866679130122066, 'timing/model': 1.9319087789626792, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 1.06}) +[2026-06-23 15:18:24,636][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.035112347453832626, 'train/grad_norm_pre_clip': 36.125701904296875, 'timing/data': 0.0007447788957506418, 'timing/model': 1.9249872559448704, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 1.06}) +[2026-06-23 15:18:28,429][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.07177668064832687, 'train/grad_norm_pre_clip': 30.49464988708496, 'timing/data': 0.0005863399710506201, 'timing/model': 1.9236613020766526, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 1.06}) +[2026-06-23 15:18:32,227][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.08711826801300049, 'train/grad_norm_pre_clip': 26.38302230834961, 'timing/data': 0.00059265096206218, 'timing/model': 1.9224739170167595, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 1.06}) +[2026-06-23 15:18:36,026][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.04866167902946472, 'train/grad_norm_pre_clip': 21.495738983154297, 'timing/data': 0.0006602099165320396, 'timing/model': 1.9226434129523113, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 1.06}) +[2026-06-23 15:18:39,849][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.027299392968416214, 'train/grad_norm_pre_clip': 18.67539405822754, 'timing/data': 0.0005951679777354002, 'timing/model': 1.9465274630347267, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 1.06}) +[2026-06-23 15:18:43,651][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.06861986219882965, 'train/grad_norm_pre_clip': 16.640899658203125, 'timing/data': 0.0006974689895287156, 'timing/model': 1.9273438119562343, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 1.06}) +[2026-06-23 15:18:47,463][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.05294784530997276, 'train/grad_norm_pre_clip': 49.86267852783203, 'timing/data': 0.0005773239536210895, 'timing/model': 1.9322444000281394, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.06}) +[2026-06-23 15:18:51,276][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.05394820123910904, 'train/grad_norm_pre_clip': 26.831127166748047, 'timing/data': 0.0007180829998105764, 'timing/model': 1.928217567037791, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 1.06}) +[2026-06-23 15:18:55,088][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.06799265742301941, 'train/grad_norm_pre_clip': 34.163700103759766, 'timing/data': 0.000588836963288486, 'timing/model': 1.9365735579049215, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 1.06}) +[2026-06-23 15:18:58,897][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.09485176205635071, 'train/grad_norm_pre_clip': 32.016143798828125, 'timing/data': 0.0006085879867896438, 'timing/model': 1.934122557984665, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 1.06}) +[2026-06-23 15:22:26,366][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2750 +[2026-06-23 15:22:26,367][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.05816146731376648, 'train/grad_norm_pre_clip': 32.88949966430664, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 123.74000000000481, 'rl_games_eval/mid_train/mean_length': 1015.8, 'rl_games_eval/mid_train/std_reward': 165.59867873870033, 'rl_games_eval/mid_train/std_length': 1333.9035047558725, 'rl_games_eval/mid_train/macro_mean_reward': 123.74000000000481, 'rl_games_eval/mid_train/macro_mean_length': 1015.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 123.74000000000481, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1015.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 165.59867873870033, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1333.9035047558725, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 123.74000000000481, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000581439002417028, 'timing/model': 1.9340685709612444, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 1.06}) +[2026-06-23 15:22:30,098][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.04955322667956352, 'train/grad_norm_pre_clip': 17.19144058227539, 'timing/data': 0.0006026879418641329, 'timing/model': 1.8848104790085927, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 1.06}) +[2026-06-23 15:22:33,826][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.04672488942742348, 'train/grad_norm_pre_clip': 19.360713958740234, 'timing/data': 0.0006008059717714787, 'timing/model': 1.8987005810486153, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 1.07}) +[2026-06-23 15:22:37,578][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.0946238785982132, 'train/grad_norm_pre_clip': 34.827476501464844, 'timing/data': 0.0007997580105438828, 'timing/model': 1.9080961359431967, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 1.07}) +[2026-06-23 15:22:41,332][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.017600074410438538, 'train/grad_norm_pre_clip': 12.089088439941406, 'timing/data': 0.0005694839637726545, 'timing/model': 1.9030970149906352, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 1.07}) +[2026-06-23 15:22:45,092][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.02047630399465561, 'train/grad_norm_pre_clip': 26.02159309387207, 'timing/data': 0.0006220500217750669, 'timing/model': 1.9126123300520703, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 1.07}) +[2026-06-23 15:22:49,029][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.04677623137831688, 'train/grad_norm_pre_clip': 21.77007484436035, 'timing/data': 0.0008086059242486954, 'timing/model': 2.086760482052341, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 1.07}) +[2026-06-23 15:22:52,806][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.14568212628364563, 'train/grad_norm_pre_clip': 28.593191146850586, 'timing/data': 0.0006018489366397262, 'timing/model': 1.9286416040267795, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 1.07}) +[2026-06-23 15:22:56,585][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.11037012189626694, 'train/grad_norm_pre_clip': 69.60826110839844, 'timing/data': 0.0007702319417148829, 'timing/model': 1.9166093250969425, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 1.07}) +[2026-06-23 15:23:00,353][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.038312606513500214, 'train/grad_norm_pre_clip': 23.176528930664062, 'timing/data': 0.0005805250257253647, 'timing/model': 1.9069376449333504, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 1.07}) +[2026-06-23 15:23:04,129][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.020653504878282547, 'train/grad_norm_pre_clip': 51.007041931152344, 'timing/data': 0.0007953770691528916, 'timing/model': 1.9119383280631155, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 1.07}) +[2026-06-23 15:23:07,907][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.07181644439697266, 'train/grad_norm_pre_clip': 55.9888801574707, 'timing/data': 0.0006215500179678202, 'timing/model': 1.919633990037255, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 1.07}) +[2026-06-23 15:23:11,675][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.05744866654276848, 'train/grad_norm_pre_clip': 19.07612419128418, 'timing/data': 0.0005785189568996429, 'timing/model': 1.9166800149250776, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 1.07}) +[2026-06-23 15:23:15,447][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.14384420216083527, 'train/grad_norm_pre_clip': 28.913286209106445, 'timing/data': 0.0005939590046182275, 'timing/model': 1.913904181914404, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 1.07}) +[2026-06-23 15:23:19,214][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.03435858339071274, 'train/grad_norm_pre_clip': 22.922388076782227, 'timing/data': 0.0005858000367879868, 'timing/model': 1.9075398610439152, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 1.07}) +[2026-06-23 15:23:23,177][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.050767555832862854, 'train/grad_norm_pre_clip': 16.79935073852539, 'timing/data': 0.0006148109678179026, 'timing/model': 1.9158386680064723, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 1.07}) +[2026-06-23 15:23:26,960][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.03125466778874397, 'train/grad_norm_pre_clip': 23.45533561706543, 'timing/data': 0.000613696058280766, 'timing/model': 1.9194509090157226, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 1.07}) +[2026-06-23 15:23:30,774][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.11488940566778183, 'train/grad_norm_pre_clip': 37.030548095703125, 'timing/data': 0.0009539489401504397, 'timing/model': 1.952989208046347, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 1.07}) +[2026-06-23 15:23:34,558][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.10980167239904404, 'train/grad_norm_pre_clip': 34.3860969543457, 'timing/data': 0.0005929230246692896, 'timing/model': 1.9214232260128483, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 1.07}) +[2026-06-23 15:23:38,330][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.08284904807806015, 'train/grad_norm_pre_clip': 29.184288024902344, 'timing/data': 0.0007254809606820345, 'timing/model': 1.9183107630815357, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 1.07}) +[2026-06-23 15:23:42,122][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.05021689832210541, 'train/grad_norm_pre_clip': 18.55399513244629, 'timing/data': 0.0005806219996884465, 'timing/model': 1.9169427399756387, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 1.07}) +[2026-06-23 15:23:45,910][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.03359813615679741, 'train/grad_norm_pre_clip': 10.197482109069824, 'timing/data': 0.000595786958001554, 'timing/model': 1.9260234570829198, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 1.07}) +[2026-06-23 15:23:49,706][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.022902658209204674, 'train/grad_norm_pre_clip': 18.667680740356445, 'timing/data': 0.0009801860433071852, 'timing/model': 1.9328847959404811, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 1.07}) +[2026-06-23 15:23:53,689][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.07479555904865265, 'train/grad_norm_pre_clip': 34.37297821044922, 'timing/data': 0.0006698010256513953, 'timing/model': 2.1061227830359712, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 1.07}) +[2026-06-23 15:23:57,471][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.051725834608078, 'train/grad_norm_pre_clip': 22.71351432800293, 'timing/data': 0.0005810399306938052, 'timing/model': 1.9194295200286433, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 1.07}) +[2026-06-23 15:24:01,260][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.02488473244011402, 'train/grad_norm_pre_clip': 30.403228759765625, 'timing/data': 0.0005900789983570576, 'timing/model': 1.922319503966719, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 1.07}) +[2026-06-23 15:24:05,060][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.19198393821716309, 'train/grad_norm_pre_clip': 27.77796173095703, 'timing/data': 0.0005793620366603136, 'timing/model': 1.9314364389283583, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 1.07}) +[2026-06-23 15:24:08,851][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.052780505269765854, 'train/grad_norm_pre_clip': 12.4558744430542, 'timing/data': 0.0005710460245609283, 'timing/model': 1.9309254600666463, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 1.07}) +[2026-06-23 15:24:12,638][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.16309025883674622, 'train/grad_norm_pre_clip': 57.208221435546875, 'timing/data': 0.0005936870584264398, 'timing/model': 1.9177767999935895, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 1.08}) +[2026-06-23 15:24:16,420][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.04987683519721031, 'train/grad_norm_pre_clip': 35.496212005615234, 'timing/data': 0.0005919190589338541, 'timing/model': 1.921348107047379, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 1.08}) +[2026-06-23 15:24:20,207][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.06096944212913513, 'train/grad_norm_pre_clip': 32.39643859863281, 'timing/data': 0.0005856320494785905, 'timing/model': 1.9289889680221677, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 1.08}) +[2026-06-23 15:24:24,006][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.061208438128232956, 'train/grad_norm_pre_clip': 31.698213577270508, 'timing/data': 0.0005861279787495732, 'timing/model': 1.9244532360462472, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 1.08}) +[2026-06-23 15:24:27,792][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.0366046279668808, 'train/grad_norm_pre_clip': 35.982688903808594, 'timing/data': 0.0005910589825361967, 'timing/model': 1.920957148890011, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 1.08}) +[2026-06-23 15:24:31,583][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.17599031329154968, 'train/grad_norm_pre_clip': 58.023521423339844, 'timing/data': 0.0005874029593542218, 'timing/model': 1.9266634720843285, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 1.08}) +[2026-06-23 15:24:35,381][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.0264343973249197, 'train/grad_norm_pre_clip': 24.030826568603516, 'timing/data': 0.000642133061774075, 'timing/model': 1.9231910390080884, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 1.08}) +[2026-06-23 15:24:39,165][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.1069423034787178, 'train/grad_norm_pre_clip': 36.397010803222656, 'timing/data': 0.0008710210677236319, 'timing/model': 1.917280868976377, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 1.08}) +[2026-06-23 15:24:42,980][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.017685541883111, 'train/grad_norm_pre_clip': 36.99414825439453, 'timing/data': 0.0005867849104106426, 'timing/model': 1.93432057404425, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 1.08}) +[2026-06-23 15:24:46,777][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.06331351399421692, 'train/grad_norm_pre_clip': 40.03684997558594, 'timing/data': 0.000664849067106843, 'timing/model': 1.9240698779467493, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 1.08}) +[2026-06-23 15:24:50,560][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.07561058551073074, 'train/grad_norm_pre_clip': 38.73039245605469, 'timing/data': 0.0006019959691911936, 'timing/model': 1.9185990900732577, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 1.08}) +[2026-06-23 15:24:54,367][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.08530636131763458, 'train/grad_norm_pre_clip': 45.027156829833984, 'timing/data': 0.0006006279727444053, 'timing/model': 1.9303002658998594, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 1.08}) +[2026-06-23 15:24:58,168][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.031119106337428093, 'train/grad_norm_pre_clip': 25.324731826782227, 'timing/data': 0.0006431880174204707, 'timing/model': 1.9258449530461803, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 1.08}) +[2026-06-23 15:25:01,951][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.09355179965496063, 'train/grad_norm_pre_clip': 44.771018981933594, 'timing/data': 0.0005918060196563601, 'timing/model': 1.918554339907132, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 1.08}) +[2026-06-23 15:25:05,763][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.0476670004427433, 'train/grad_norm_pre_clip': 31.17088508605957, 'timing/data': 0.0005973710212856531, 'timing/model': 1.9242969050537795, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 1.08}) +[2026-06-23 15:25:09,571][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.006517169997096062, 'train/grad_norm_pre_clip': 21.746131896972656, 'timing/data': 0.0005903110140934587, 'timing/model': 1.933564416016452, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 1.08}) +[2026-06-23 15:25:13,380][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.05223517119884491, 'train/grad_norm_pre_clip': 24.62248992919922, 'timing/data': 0.000595045043155551, 'timing/model': 1.9381238939240575, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 1.08}) +[2026-06-23 15:25:17,168][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.07408618927001953, 'train/grad_norm_pre_clip': 48.65321350097656, 'timing/data': 0.0006023639580234885, 'timing/model': 1.9185579750919715, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.08}) +[2026-06-23 15:25:20,951][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.11521877348423004, 'train/grad_norm_pre_clip': 35.09208297729492, 'timing/data': 0.0005693199345842004, 'timing/model': 1.9192561390809715, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 1.08}) +[2026-06-23 15:25:24,753][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.11315412074327469, 'train/grad_norm_pre_clip': 21.258434295654297, 'timing/data': 0.0005750240525230765, 'timing/model': 1.928076870040968, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 1.08}) +[2026-06-23 15:25:28,547][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.06266430020332336, 'train/grad_norm_pre_clip': 13.495992660522461, 'timing/data': 0.0006282099056988955, 'timing/model': 1.9258845950243995, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 1.08}) +[2026-06-23 15:25:32,485][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.035403575748205185, 'train/grad_norm_pre_clip': 16.262277603149414, 'timing/data': 0.0005988780176267028, 'timing/model': 2.074012165889144, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 1.08}) +[2026-06-23 15:25:36,273][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2800 +[2026-06-23 15:27:43,705][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2800 +[2026-06-23 15:27:43,706][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2800 +[2026-06-23 15:29:06,491][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2800 +[2026-06-23 15:29:06,492][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.08175923675298691, 'train/grad_norm_pre_clip': 28.52720832824707, 'eval/action_loss/seconds': 127.43008127599023, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03624390438199043, 'eval/latency_2/loss': 0.03624390438199043, 'eval/flappy/loss': 0.03624390438199043, 'eval/flappy/latency_2/loss': 0.03624390438199043, 'eval/flap_precision': 0.9501639344262295, 'eval/flap_recall': 0.9777327935222672, 'eval/flap_f1': 0.963751247090123, 'eval/flap_rate_model': 0.21180555555555555, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.963751247090123, 'eval/latency_2/flap_precision': 0.9501639344262295, 'eval/latency_2/flap_recall': 0.9777327935222672, 'eval/latency_2/flap_f1': 0.963751247090123, 'eval/latency_2/flap_rate_model': 0.21180555555555555, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.963751247090123, 'eval/action_classification/data_wait_seconds': 1.599132776260376, 'eval/action_classification/predict_seconds': 80.9957275390625, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005848479922860861, 'timing/model': 1.9252009650226682, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 1.08}) +[2026-06-23 15:29:10,257][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.03160740062594414, 'train/grad_norm_pre_clip': 15.2935152053833, 'timing/data': 0.0007666139863431454, 'timing/model': 1.9105166719527915, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 1.08}) +[2026-06-23 15:29:14,044][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.03904455900192261, 'train/grad_norm_pre_clip': 18.902191162109375, 'timing/data': 0.0007613750640302896, 'timing/model': 1.924228303018026, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 1.08}) +[2026-06-23 15:29:17,822][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.04776086285710335, 'train/grad_norm_pre_clip': 24.154767990112305, 'timing/data': 0.0006800319533795118, 'timing/model': 1.9159984409343451, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 1.08}) +[2026-06-23 15:29:21,736][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.1156369224190712, 'train/grad_norm_pre_clip': 40.504215240478516, 'timing/data': 0.0005979209672659636, 'timing/model': 1.9198937159962952, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 1.09}) +[2026-06-23 15:29:25,525][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.01783839799463749, 'train/grad_norm_pre_clip': 21.675174713134766, 'timing/data': 0.0007206120062619448, 'timing/model': 1.925761806895025, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 1.09}) +[2026-06-23 15:29:29,311][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.10483712702989578, 'train/grad_norm_pre_clip': 28.950170516967773, 'timing/data': 0.0006740959361195564, 'timing/model': 1.9176515960134566, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 1.09}) +[2026-06-23 15:29:33,109][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.08631009608507156, 'train/grad_norm_pre_clip': 31.464365005493164, 'timing/data': 0.0006033600075170398, 'timing/model': 1.929711845004931, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 1.09}) +[2026-06-23 15:29:36,908][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.20592960715293884, 'train/grad_norm_pre_clip': 50.15294647216797, 'timing/data': 0.0005795879987999797, 'timing/model': 1.9271348570473492, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 1.09}) +[2026-06-23 15:29:40,711][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.147476926445961, 'train/grad_norm_pre_clip': 26.755186080932617, 'timing/data': 0.0007757949642837048, 'timing/model': 1.9299566100817174, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 1.09}) +[2026-06-23 15:29:44,515][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.0603923499584198, 'train/grad_norm_pre_clip': 29.809234619140625, 'timing/data': 0.0006153429858386517, 'timing/model': 1.9296701790299267, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 1.09}) +[2026-06-23 15:29:48,315][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.028590071946382523, 'train/grad_norm_pre_clip': 16.398157119750977, 'timing/data': 0.0005840760422870517, 'timing/model': 1.9282652919646353, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 1.09}) +[2026-06-23 15:29:52,192][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.04963863641023636, 'train/grad_norm_pre_clip': 23.932859420776367, 'timing/data': 0.0006275459891185164, 'timing/model': 2.0013877419987693, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 1.09}) +[2026-06-23 15:29:55,978][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.0860109031200409, 'train/grad_norm_pre_clip': 27.67326545715332, 'timing/data': 0.0006661099614575505, 'timing/model': 1.9175918269902468, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 1.09}) +[2026-06-23 15:29:59,783][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.13420255482196808, 'train/grad_norm_pre_clip': 16.014362335205078, 'timing/data': 0.0007333920802921057, 'timing/model': 1.9372389960335568, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 1.09}) +[2026-06-23 15:30:03,596][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.07598447799682617, 'train/grad_norm_pre_clip': 31.654987335205078, 'timing/data': 0.0009116970468312502, 'timing/model': 1.9394894309807569, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 1.09}) +[2026-06-23 15:30:07,398][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.004274939186871052, 'train/grad_norm_pre_clip': 21.334566116333008, 'timing/data': 0.0006191940046846867, 'timing/model': 1.9234777950914577, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 1.09}) +[2026-06-23 15:30:11,186][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.12903381884098053, 'train/grad_norm_pre_clip': 29.019332885742188, 'timing/data': 0.0007823710329830647, 'timing/model': 1.9222213480388746, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 1.09}) +[2026-06-23 15:30:14,979][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.04786624759435654, 'train/grad_norm_pre_clip': 25.784992218017578, 'timing/data': 0.0005942520219832659, 'timing/model': 1.9320261999964714, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 1.09}) +[2026-06-23 15:30:18,791][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.07880651950836182, 'train/grad_norm_pre_clip': 33.1848030090332, 'timing/data': 0.0006002910668030381, 'timing/model': 1.9367376259760931, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 1.09}) +[2026-06-23 15:30:22,582][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.03063065931200981, 'train/grad_norm_pre_clip': 17.151281356811523, 'timing/data': 0.0006784630240872502, 'timing/model': 1.9199272929690778, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 1.09}) +[2026-06-23 15:30:26,582][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.06270269304513931, 'train/grad_norm_pre_clip': 20.55660629272461, 'timing/data': 0.000809194054454565, 'timing/model': 2.1247441599844024, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 1.09}) +[2026-06-23 15:30:30,374][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.056722089648246765, 'train/grad_norm_pre_clip': 41.94414520263672, 'timing/data': 0.0005853699985891581, 'timing/model': 1.9200319610536098, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 1.09}) +[2026-06-23 15:30:34,184][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.1712781935930252, 'train/grad_norm_pre_clip': 32.98676300048828, 'timing/data': 0.000581110012717545, 'timing/model': 1.941460717935115, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 1.09}) +[2026-06-23 15:30:37,972][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.11190854012966156, 'train/grad_norm_pre_clip': 25.224044799804688, 'timing/data': 0.0005830790614709258, 'timing/model': 1.9181292100111023, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 1.09}) +[2026-06-23 15:30:41,779][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.0687270238995552, 'train/grad_norm_pre_clip': 27.929725646972656, 'timing/data': 0.0005887410370633006, 'timing/model': 1.9264216390438378, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 1.09}) +[2026-06-23 15:30:45,575][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.1061413586139679, 'train/grad_norm_pre_clip': 50.03952407836914, 'timing/data': 0.0006358230020850897, 'timing/model': 1.9286244829418138, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 1.09}) +[2026-06-23 15:30:49,373][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.039573609828948975, 'train/grad_norm_pre_clip': 12.019526481628418, 'timing/data': 0.0005674600834026933, 'timing/model': 1.9325093540828675, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 1.09}) +[2026-06-23 15:30:53,178][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.04296168312430382, 'train/grad_norm_pre_clip': 25.287151336669922, 'timing/data': 0.0008486739825457335, 'timing/model': 1.925998140941374, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 1.09}) +[2026-06-23 15:30:56,967][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.07143139839172363, 'train/grad_norm_pre_clip': 22.467832565307617, 'timing/data': 0.0006784299621358514, 'timing/model': 1.9232994510093704, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 1.09}) +[2026-06-23 15:31:00,936][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.03981028497219086, 'train/grad_norm_pre_clip': 27.427337646484375, 'timing/data': 0.0007160590030252934, 'timing/model': 1.923532699001953, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 1.1}) +[2026-06-23 15:31:04,740][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.04261431097984314, 'train/grad_norm_pre_clip': 35.08084487915039, 'timing/data': 0.0006432129302993417, 'timing/model': 1.9234036148991436, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 1.1}) +[2026-06-23 15:31:08,563][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.12375351786613464, 'train/grad_norm_pre_clip': 35.89362716674805, 'timing/data': 0.0005971550708636642, 'timing/model': 1.9407199859851971, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 1.1}) +[2026-06-23 15:31:12,355][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.07379936426877975, 'train/grad_norm_pre_clip': 28.015422821044922, 'timing/data': 0.0006012240191921592, 'timing/model': 1.9231982950586826, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 1.1}) +[2026-06-23 15:31:16,165][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.023659251630306244, 'train/grad_norm_pre_clip': 16.726314544677734, 'timing/data': 0.0005852690665051341, 'timing/model': 1.9342795161064714, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 1.1}) +[2026-06-23 15:31:19,991][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.08793403953313828, 'train/grad_norm_pre_clip': 29.347976684570312, 'timing/data': 0.0006225400138646364, 'timing/model': 1.9356569460360333, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 1.1}) +[2026-06-23 15:31:23,804][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.05780896916985512, 'train/grad_norm_pre_clip': 31.63091278076172, 'timing/data': 0.0005790379364043474, 'timing/model': 1.937562732025981, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 1.1}) +[2026-06-23 15:31:27,603][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.03697466477751732, 'train/grad_norm_pre_clip': 26.7985782623291, 'timing/data': 0.0005799449281767011, 'timing/model': 1.9385472340509295, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 1.1}) +[2026-06-23 15:31:31,535][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.052391417324543, 'train/grad_norm_pre_clip': 15.500304222106934, 'timing/data': 0.000587497022934258, 'timing/model': 2.0597246940014884, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 1.1}) +[2026-06-23 15:31:35,333][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.0769478976726532, 'train/grad_norm_pre_clip': 32.72836685180664, 'timing/data': 0.0005818229401484132, 'timing/model': 1.9270566758932546, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 1.1}) +[2026-06-23 15:31:39,143][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.05664852634072304, 'train/grad_norm_pre_clip': 26.095609664916992, 'timing/data': 0.000839008018374443, 'timing/model': 1.934476719936356, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 1.1}) +[2026-06-23 15:31:42,953][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.06581812351942062, 'train/grad_norm_pre_clip': 25.00222396850586, 'timing/data': 0.0006111749680712819, 'timing/model': 1.930196093977429, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 1.1}) +[2026-06-23 15:31:46,748][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.06630654633045197, 'train/grad_norm_pre_clip': 25.013105392456055, 'timing/data': 0.0008766659302636981, 'timing/model': 1.9293300730641931, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 1.1}) +[2026-06-23 15:31:50,553][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.08550931513309479, 'train/grad_norm_pre_clip': 48.820533752441406, 'timing/data': 0.0005781929939985275, 'timing/model': 1.9249828319298103, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 1.1}) +[2026-06-23 15:31:54,361][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.07326587289571762, 'train/grad_norm_pre_clip': 42.01678466796875, 'timing/data': 0.0008111030329018831, 'timing/model': 1.9312763039488345, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.1}) +[2026-06-23 15:31:58,157][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.036892157047986984, 'train/grad_norm_pre_clip': 14.367707252502441, 'timing/data': 0.0005905620055273175, 'timing/model': 1.9241573140025139, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 1.1}) +[2026-06-23 15:32:01,962][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.05701388791203499, 'train/grad_norm_pre_clip': 17.631526947021484, 'timing/data': 0.0008909939788281918, 'timing/model': 1.9373714399989694, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 1.1}) +[2026-06-23 15:32:05,901][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.05426666513085365, 'train/grad_norm_pre_clip': 20.063634872436523, 'timing/data': 0.0005863480037078261, 'timing/model': 1.92974749696441, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 1.1}) +[2026-06-23 15:32:09,705][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.07390535622835159, 'train/grad_norm_pre_clip': 24.440589904785156, 'timing/data': 0.0005845359992235899, 'timing/model': 1.9264738790225238, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 1.1}) +[2026-06-23 15:32:13,507][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.02792184427380562, 'train/grad_norm_pre_clip': 15.342248916625977, 'timing/data': 0.0005733090220019221, 'timing/model': 1.9334724789950997, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 1.1}) +[2026-06-23 15:32:17,313][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.024687962606549263, 'train/grad_norm_pre_clip': 14.785384178161621, 'timing/data': 0.0007429949473589659, 'timing/model': 1.9206889970228076, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 1.1}) +[2026-06-23 15:32:21,100][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.01768980175256729, 'train/grad_norm_pre_clip': 24.1344051361084, 'timing/data': 0.0005936429370194674, 'timing/model': 1.9232948070857674, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 1.1}) +[2026-06-23 15:32:24,898][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.038453515619039536, 'train/grad_norm_pre_clip': 87.66558074951172, 'timing/data': 0.0005813209572806954, 'timing/model': 1.9288014579797164, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 1.1}) +[2026-06-23 15:32:28,705][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.05606868863105774, 'train/grad_norm_pre_clip': 31.825414657592773, 'timing/data': 0.0005909709725528955, 'timing/model': 1.9373385739745572, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 1.1}) +[2026-06-23 15:32:32,509][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.07425932586193085, 'train/grad_norm_pre_clip': 30.68821144104004, 'timing/data': 0.0006396890385076404, 'timing/model': 1.9243062279419973, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 1.1}) +[2026-06-23 15:32:36,480][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.049772609025239944, 'train/grad_norm_pre_clip': 16.461986541748047, 'timing/data': 0.0005985830212011933, 'timing/model': 2.105137465056032, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 1.1}) +[2026-06-23 15:32:40,269][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.04629352316260338, 'train/grad_norm_pre_clip': 45.66815948486328, 'timing/data': 0.0006085630739107728, 'timing/model': 1.9169922840083018, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 1.11}) +[2026-06-23 15:32:44,076][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.07587267458438873, 'train/grad_norm_pre_clip': 44.06161117553711, 'timing/data': 0.000599976978264749, 'timing/model': 1.9350830030161887, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 1.11}) +[2026-06-23 15:32:47,877][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.17169189453125, 'train/grad_norm_pre_clip': 23.17729949951172, 'timing/data': 0.0006204689852893353, 'timing/model': 1.9317584079690278, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 1.11}) +[2026-06-23 15:32:51,677][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.16510474681854248, 'train/grad_norm_pre_clip': 36.23378372192383, 'timing/data': 0.0005937260575592518, 'timing/model': 1.9291276329895481, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 1.11}) +[2026-06-23 15:32:55,462][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.07493900507688522, 'train/grad_norm_pre_clip': 21.638357162475586, 'timing/data': 0.0005844039842486382, 'timing/model': 1.9183358869049698, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 1.11}) +[2026-06-23 15:32:59,259][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.06565337628126144, 'train/grad_norm_pre_clip': 39.51753234863281, 'timing/data': 0.000619255006313324, 'timing/model': 1.9243806300219148, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 1.11}) +[2026-06-23 15:33:03,060][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.07476141303777695, 'train/grad_norm_pre_clip': 44.97641372680664, 'timing/data': 0.0005838050274178386, 'timing/model': 1.9238388849189505, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 1.11}) +[2026-06-23 15:33:06,900][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.039732351899147034, 'train/grad_norm_pre_clip': 45.87425231933594, 'timing/data': 0.0006050450028851628, 'timing/model': 1.9339922619983554, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 1.11}) +[2026-06-23 15:33:10,897][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.03550843521952629, 'train/grad_norm_pre_clip': 43.53190612792969, 'timing/data': 0.0008085329318419099, 'timing/model': 1.9317320990376174, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 1.11}) +[2026-06-23 15:33:14,697][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.16656872630119324, 'train/grad_norm_pre_clip': 54.276695251464844, 'timing/data': 0.0005916389636695385, 'timing/model': 1.9311259740497917, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 1.11}) +[2026-06-23 15:33:18,509][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.14031995832920074, 'train/grad_norm_pre_clip': 27.894126892089844, 'timing/data': 0.0005930769257247448, 'timing/model': 1.9366690349997953, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 1.11}) +[2026-06-23 15:33:22,304][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.09710139036178589, 'train/grad_norm_pre_clip': 33.183563232421875, 'timing/data': 0.000593325006775558, 'timing/model': 1.9312131199985743, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 1.11}) +[2026-06-23 15:33:26,108][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.04896031692624092, 'train/grad_norm_pre_clip': 58.55528259277344, 'timing/data': 0.0007387639489024878, 'timing/model': 1.9390842230059206, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 1.11}) +[2026-06-23 15:33:29,901][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.019846340641379356, 'train/grad_norm_pre_clip': 13.919313430786133, 'timing/data': 0.000581888947635889, 'timing/model': 1.9276818829821423, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 1.11}) +[2026-06-23 15:33:33,701][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.058509327471256256, 'train/grad_norm_pre_clip': 37.17395782470703, 'timing/data': 0.0008489559404551983, 'timing/model': 1.9223699929425493, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 1.11}) +[2026-06-23 15:33:37,512][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.08732348680496216, 'train/grad_norm_pre_clip': 25.294517517089844, 'timing/data': 0.0008122390136122704, 'timing/model': 1.9279953440418467, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 1.11}) +[2026-06-23 15:33:41,320][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.11439070105552673, 'train/grad_norm_pre_clip': 34.79276657104492, 'timing/data': 0.0007916410686448216, 'timing/model': 1.9297139670234174, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 1.11}) +[2026-06-23 15:33:45,313][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.08491423726081848, 'train/grad_norm_pre_clip': 15.103421211242676, 'timing/data': 0.0005860370583832264, 'timing/model': 1.9263732290128246, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 1.11}) +[2026-06-23 15:33:49,117][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.07278803735971451, 'train/grad_norm_pre_clip': 36.004608154296875, 'timing/data': 0.0005872900364920497, 'timing/model': 1.9262792230583727, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 1.11}) +[2026-06-23 15:33:52,920][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.030345648527145386, 'train/grad_norm_pre_clip': 26.13750648498535, 'timing/data': 0.0005901469849050045, 'timing/model': 1.924994730972685, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 1.11}) +[2026-06-23 15:33:56,733][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.08551614731550217, 'train/grad_norm_pre_clip': 35.77117919921875, 'timing/data': 0.0006185179809108377, 'timing/model': 1.9294421470258385, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 1.11}) +[2026-06-23 15:34:00,541][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.056313756853342056, 'train/grad_norm_pre_clip': 33.46706008911133, 'timing/data': 0.000630578026175499, 'timing/model': 1.9310632919659838, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 1.11}) +[2026-06-23 15:34:04,343][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.051011648029088974, 'train/grad_norm_pre_clip': 14.721020698547363, 'timing/data': 0.0006361729465425014, 'timing/model': 1.929778108955361, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 1.11}) +[2026-06-23 15:34:08,201][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.08133392035961151, 'train/grad_norm_pre_clip': 35.811248779296875, 'timing/data': 0.0008929439354687929, 'timing/model': 1.9400746440514922, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 1.11}) +[2026-06-23 15:34:12,018][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.10448883473873138, 'train/grad_norm_pre_clip': 35.19367218017578, 'timing/data': 0.0006061180029064417, 'timing/model': 1.9401484930422157, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 1.11}) +[2026-06-23 15:34:15,941][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.12107382714748383, 'train/grad_norm_pre_clip': 48.65448760986328, 'timing/data': 0.0005891670007258654, 'timing/model': 2.052083689952269, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 1.11}) +[2026-06-23 15:34:19,729][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.08306025713682175, 'train/grad_norm_pre_clip': 29.20698356628418, 'timing/data': 0.0006124479696154594, 'timing/model': 1.9223097889916971, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 1.12}) +[2026-06-23 15:34:23,565][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.0755191296339035, 'train/grad_norm_pre_clip': 38.172122955322266, 'timing/data': 0.0006081939209252596, 'timing/model': 1.9513639010256156, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 1.12}) +[2026-06-23 15:34:27,364][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.04904313385486603, 'train/grad_norm_pre_clip': 19.52664566040039, 'timing/data': 0.0005968159530311823, 'timing/model': 1.9284148690057918, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 1.12}) +[2026-06-23 15:34:31,167][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.08014002442359924, 'train/grad_norm_pre_clip': 69.90433502197266, 'timing/data': 0.0005845740670338273, 'timing/model': 1.9328709280816838, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 1.12}) +[2026-06-23 15:34:34,961][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.021141916513442993, 'train/grad_norm_pre_clip': 25.520238876342773, 'timing/data': 0.0006122119957581162, 'timing/model': 1.9295483500463888, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 1.12}) +[2026-06-23 15:34:38,776][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.03270542621612549, 'train/grad_norm_pre_clip': 17.14093589782715, 'timing/data': 0.0009677940979599953, 'timing/model': 1.9400577099295333, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 1.12}) +[2026-06-23 15:34:42,570][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.056078143417835236, 'train/grad_norm_pre_clip': 45.81066131591797, 'timing/data': 0.0005823590327054262, 'timing/model': 1.9259919590549543, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 1.12}) +[2026-06-23 15:34:46,352][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.03237864747643471, 'train/grad_norm_pre_clip': 24.034029006958008, 'timing/data': 0.0006339240353554487, 'timing/model': 1.9176428209757432, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 1.12}) +[2026-06-23 15:34:50,283][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.11092904210090637, 'train/grad_norm_pre_clip': 52.64643478393555, 'timing/data': 0.0006192619912326336, 'timing/model': 1.9276385250268504, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 1.12}) +[2026-06-23 15:34:54,072][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.07042460143566132, 'train/grad_norm_pre_clip': 24.86627960205078, 'timing/data': 0.0005892880726605654, 'timing/model': 1.9185338870156556, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 1.12}) +[2026-06-23 15:34:57,872][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.0342618003487587, 'train/grad_norm_pre_clip': 23.491064071655273, 'timing/data': 0.000574760022573173, 'timing/model': 1.9261402619304135, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 1.12}) +[2026-06-23 15:35:01,668][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.06316574662923813, 'train/grad_norm_pre_clip': 29.374610900878906, 'timing/data': 0.0005995659157633781, 'timing/model': 1.924866014975123, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.12}) +[2026-06-23 15:35:05,477][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.012348953634500504, 'train/grad_norm_pre_clip': 10.762042999267578, 'timing/data': 0.0006120990728959441, 'timing/model': 1.9261691540013999, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 1.12}) +[2026-06-23 15:35:09,285][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.07199163734912872, 'train/grad_norm_pre_clip': 47.9898681640625, 'timing/data': 0.0006236449116840959, 'timing/model': 1.935196242062375, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 1.12}) +[2026-06-23 15:35:13,091][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.06612658500671387, 'train/grad_norm_pre_clip': 41.08771514892578, 'timing/data': 0.0006665110122412443, 'timing/model': 1.9353910540230572, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 1.12}) +[2026-06-23 15:35:16,890][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.13278459012508392, 'train/grad_norm_pre_clip': 65.18827819824219, 'timing/data': 0.0008018639637157321, 'timing/model': 1.93555118504446, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 1.12}) +[2026-06-23 15:35:20,872][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.1079818457365036, 'train/grad_norm_pre_clip': 31.44692039489746, 'timing/data': 0.0006228119600564241, 'timing/model': 2.101903080008924, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 1.12}) +[2026-06-23 15:35:24,683][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.053080517798662186, 'train/grad_norm_pre_clip': 44.08931350708008, 'timing/data': 0.0008278139866888523, 'timing/model': 1.9359500920400023, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 1.12}) +[2026-06-23 15:35:28,573][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2900 +[2026-06-23 15:37:35,030][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2900 +[2026-06-23 15:37:35,033][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2900 +[2026-06-23 15:38:58,845][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2900 +[2026-06-23 15:38:58,846][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.2758212685585022, 'train/grad_norm_pre_clip': 57.4052848815918, 'eval/action_loss/seconds': 126.45411732106004, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.04822910949587822, 'eval/latency_2/loss': 0.04822910949587822, 'eval/flappy/loss': 0.04822910949587822, 'eval/flappy/latency_2/loss': 0.04822910949587822, 'eval/flap_precision': 0.985021398002853, 'eval/flap_recall': 0.9318488529014844, 'eval/flap_f1': 0.9576976421636615, 'eval/flap_rate_model': 0.1947222222222222, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9576976421636615, 'eval/latency_2/flap_precision': 0.985021398002853, 'eval/latency_2/flap_recall': 0.9318488529014844, 'eval/latency_2/flap_f1': 0.9576976421636615, 'eval/latency_2/flap_rate_model': 0.1947222222222222, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9576976421636615, 'eval/action_classification/data_wait_seconds': 1.7148354053497314, 'eval/action_classification/predict_seconds': 81.92414855957031, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005951999919489026, 'timing/model': 1.946307170088403, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 1.12}) +[2026-06-23 15:39:02,630][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.22379985451698303, 'train/grad_norm_pre_clip': 92.20793151855469, 'timing/data': 0.0005807980196550488, 'timing/model': 1.9234023219905794, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 1.12}) +[2026-06-23 15:39:06,409][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.06554300338029861, 'train/grad_norm_pre_clip': 39.96744155883789, 'timing/data': 0.0005998220294713974, 'timing/model': 1.9108799200039357, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 1.12}) +[2026-06-23 15:39:10,299][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.0854920893907547, 'train/grad_norm_pre_clip': 31.75261116027832, 'timing/data': 0.0007062219083309174, 'timing/model': 2.0271949290763587, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 1.12}) +[2026-06-23 15:39:14,077][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.032166242599487305, 'train/grad_norm_pre_clip': 29.93671226501465, 'timing/data': 0.0005924220895394683, 'timing/model': 1.9267526030307636, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 1.12}) +[2026-06-23 15:39:17,866][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.10890898108482361, 'train/grad_norm_pre_clip': 37.396610260009766, 'timing/data': 0.0005889879539608955, 'timing/model': 1.9266280579613522, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 1.12}) +[2026-06-23 15:39:21,652][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.12898746132850647, 'train/grad_norm_pre_clip': 59.65597915649414, 'timing/data': 0.0006421990692615509, 'timing/model': 1.9232143140397966, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 1.12}) +[2026-06-23 15:39:25,431][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.26191842555999756, 'train/grad_norm_pre_clip': 72.05162811279297, 'timing/data': 0.0009221530053764582, 'timing/model': 1.9106016890145838, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 1.13}) +[2026-06-23 15:39:29,219][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.06477387249469757, 'train/grad_norm_pre_clip': 27.690631866455078, 'timing/data': 0.000604191911406815, 'timing/model': 1.9200753029435873, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 1.13}) +[2026-06-23 15:39:32,996][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.020816514268517494, 'train/grad_norm_pre_clip': 11.939859390258789, 'timing/data': 0.0006021660519763827, 'timing/model': 1.9155921820783988, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 1.13}) +[2026-06-23 15:39:36,775][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.07751790434122086, 'train/grad_norm_pre_clip': 31.25489616394043, 'timing/data': 0.0008095690282061696, 'timing/model': 1.9154661149950698, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 1.13}) +[2026-06-23 15:39:40,574][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.0168326236307621, 'train/grad_norm_pre_clip': 20.997695922851562, 'timing/data': 0.0006201600190252066, 'timing/model': 1.9306525610154495, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 1.13}) +[2026-06-23 15:39:44,540][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.029055088758468628, 'train/grad_norm_pre_clip': 34.869239807128906, 'timing/data': 0.0005844749975949526, 'timing/model': 2.0984417219879106, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 1.13}) +[2026-06-23 15:39:48,320][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.04018276557326317, 'train/grad_norm_pre_clip': 41.70979309082031, 'timing/data': 0.0006213440792635083, 'timing/model': 1.9219695450738072, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 1.13}) +[2026-06-23 15:39:52,108][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.09189456701278687, 'train/grad_norm_pre_clip': 21.899682998657227, 'timing/data': 0.0006136440206319094, 'timing/model': 1.9264126259367913, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 1.13}) +[2026-06-23 15:39:55,891][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.013572999276220798, 'train/grad_norm_pre_clip': 46.34147262573242, 'timing/data': 0.0006495669949799776, 'timing/model': 1.9204761809669435, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 1.13}) +[2026-06-23 15:39:59,693][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.07570290565490723, 'train/grad_norm_pre_clip': 42.40991973876953, 'timing/data': 0.0007662950083613396, 'timing/model': 1.9263962500263005, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 1.13}) +[2026-06-23 15:40:03,500][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.08797155320644379, 'train/grad_norm_pre_clip': 38.29413986206055, 'timing/data': 0.0006187059916555882, 'timing/model': 1.935874519054778, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 1.13}) +[2026-06-23 15:40:07,289][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.02297298051416874, 'train/grad_norm_pre_clip': 28.180225372314453, 'timing/data': 0.0007933919550850987, 'timing/model': 1.9245831199223176, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 1.13}) +[2026-06-23 15:40:11,096][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.1077083870768547, 'train/grad_norm_pre_clip': 36.34132385253906, 'timing/data': 0.0006080589955672622, 'timing/model': 1.9414485159795731, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 1.13}) +[2026-06-23 15:40:14,904][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.1369665116071701, 'train/grad_norm_pre_clip': 49.96529769897461, 'timing/data': 0.0005895199719816446, 'timing/model': 1.9252022809814662, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 1.13}) +[2026-06-23 15:40:18,878][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.034771665930747986, 'train/grad_norm_pre_clip': 29.338083267211914, 'timing/data': 0.0006502369651570916, 'timing/model': 1.9289453610545024, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 1.13}) +[2026-06-23 15:40:22,698][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.04576472193002701, 'train/grad_norm_pre_clip': 25.121950149536133, 'timing/data': 0.0007629850879311562, 'timing/model': 1.9469744500238448, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 1.13}) +[2026-06-23 15:40:26,507][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.018229112029075623, 'train/grad_norm_pre_clip': 12.094820976257324, 'timing/data': 0.0005857560317963362, 'timing/model': 1.931652070954442, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 1.13}) +[2026-06-23 15:40:30,306][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.07791855186223984, 'train/grad_norm_pre_clip': 24.901596069335938, 'timing/data': 0.0005790609866380692, 'timing/model': 1.935462799970992, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 1.13}) +[2026-06-23 15:40:34,096][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.07147199660539627, 'train/grad_norm_pre_clip': 54.8358039855957, 'timing/data': 0.0006245580734685063, 'timing/model': 1.9196664459304884, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 1.13}) +[2026-06-23 15:40:37,896][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.05345472693443298, 'train/grad_norm_pre_clip': 16.690553665161133, 'timing/data': 0.0006094309501349926, 'timing/model': 1.9313180749304593, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 1.13}) +[2026-06-23 15:40:41,687][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.10358518362045288, 'train/grad_norm_pre_clip': 28.071060180664062, 'timing/data': 0.0005927740130573511, 'timing/model': 1.9230255630100146, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 1.13}) +[2026-06-23 15:40:45,487][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.07020982354879379, 'train/grad_norm_pre_clip': 22.74671173095703, 'timing/data': 0.0008177920244634151, 'timing/model': 1.9252350489841774, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 1.13}) +[2026-06-23 15:40:49,419][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.07446488738059998, 'train/grad_norm_pre_clip': 34.51908874511719, 'timing/data': 0.0005880000535398722, 'timing/model': 2.0615917349932715, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 1.13}) +[2026-06-23 15:40:53,223][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.09351867437362671, 'train/grad_norm_pre_clip': 37.18637466430664, 'timing/data': 0.0006378779653459787, 'timing/model': 1.9338924000039697, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 1.13}) +[2026-06-23 15:40:57,022][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.016200274229049683, 'train/grad_norm_pre_clip': 14.965978622436523, 'timing/data': 0.0006257150089368224, 'timing/model': 1.9257116399239749, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 1.13}) +[2026-06-23 15:41:00,831][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.13093356788158417, 'train/grad_norm_pre_clip': 43.849021911621094, 'timing/data': 0.000628557987511158, 'timing/model': 1.9293590700253844, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 1.13}) +[2026-06-23 15:41:04,632][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.041444286704063416, 'train/grad_norm_pre_clip': 15.879459381103516, 'timing/data': 0.0005803959211334586, 'timing/model': 1.929047314915806, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 1.14}) +[2026-06-23 15:41:08,435][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.16029144823551178, 'train/grad_norm_pre_clip': 31.28026008605957, 'timing/data': 0.0006803229916840792, 'timing/model': 1.9215416290098801, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 1.14}) +[2026-06-23 15:41:12,238][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.09902071952819824, 'train/grad_norm_pre_clip': 47.035804748535156, 'timing/data': 0.0006491689709946513, 'timing/model': 1.9324268019990996, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 1.14}) +[2026-06-23 15:41:16,045][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.046237338334321976, 'train/grad_norm_pre_clip': 32.9130744934082, 'timing/data': 0.000588505994528532, 'timing/model': 1.9296956929611042, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 1.14}) +[2026-06-23 15:41:19,834][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.1635144203901291, 'train/grad_norm_pre_clip': 34.512603759765625, 'timing/data': 0.0006198720075190067, 'timing/model': 1.9256955550517887, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 1.14}) +[2026-06-23 15:41:23,758][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.12834571301937103, 'train/grad_norm_pre_clip': 71.15740203857422, 'timing/data': 0.0008748499676585197, 'timing/model': 1.9379554219776765, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 1.14}) +[2026-06-23 15:41:27,572][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.04950643330812454, 'train/grad_norm_pre_clip': 18.03963851928711, 'timing/data': 0.0005894619971513748, 'timing/model': 1.9271734559442848, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 1.14}) +[2026-06-23 15:41:31,370][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.07838249951601028, 'train/grad_norm_pre_clip': 31.39559555053711, 'timing/data': 0.0006038200808688998, 'timing/model': 1.927950033918023, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 1.14}) +[2026-06-23 15:41:35,181][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.11729060113430023, 'train/grad_norm_pre_clip': 24.216901779174805, 'timing/data': 0.000588210066780448, 'timing/model': 1.93462587101385, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 1.14}) +[2026-06-23 15:41:38,985][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.01860129088163376, 'train/grad_norm_pre_clip': 22.730754852294922, 'timing/data': 0.0005856379866600037, 'timing/model': 1.9328279281035066, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.14}) +[2026-06-23 15:41:42,782][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.1456863135099411, 'train/grad_norm_pre_clip': 36.22323989868164, 'timing/data': 0.0006080820458009839, 'timing/model': 1.9197305750567466, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 1.14}) +[2026-06-23 15:41:46,580][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.0493868850171566, 'train/grad_norm_pre_clip': 22.975807189941406, 'timing/data': 0.0006308909505605698, 'timing/model': 1.9286059369333088, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 1.14}) +[2026-06-23 15:41:50,368][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.06554931402206421, 'train/grad_norm_pre_clip': 52.03966522216797, 'timing/data': 0.0006086110370233655, 'timing/model': 1.9159778149332851, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 1.14}) +[2026-06-23 15:41:54,243][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.01010122150182724, 'train/grad_norm_pre_clip': 13.381193161010742, 'timing/data': 0.0006130629917606711, 'timing/model': 2.0079018890392035, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 1.14}) +[2026-06-23 15:41:58,046][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.021900972351431847, 'train/grad_norm_pre_clip': 14.294477462768555, 'timing/data': 0.0006056210258975625, 'timing/model': 1.9303193980595097, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 1.14}) +[2026-06-23 15:42:01,844][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.07044723629951477, 'train/grad_norm_pre_clip': 53.24653244018555, 'timing/data': 0.0007734819082543254, 'timing/model': 1.921325761009939, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 1.14}) +[2026-06-23 15:42:05,653][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.025575250387191772, 'train/grad_norm_pre_clip': 18.55748748779297, 'timing/data': 0.0007884529186412692, 'timing/model': 1.9298849470214918, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 1.14}) +[2026-06-23 15:42:09,474][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.014958973973989487, 'train/grad_norm_pre_clip': 29.087186813354492, 'timing/data': 0.0005997910629957914, 'timing/model': 1.9388802280882373, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 1.14}) +[2026-06-23 15:42:13,292][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.13395360112190247, 'train/grad_norm_pre_clip': 49.2465934753418, 'timing/data': 0.0006293270271271467, 'timing/model': 1.9432545430026948, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 1.14}) +[2026-06-23 15:42:17,117][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.04112525284290314, 'train/grad_norm_pre_clip': 50.464134216308594, 'timing/data': 0.0006311510223895311, 'timing/model': 1.9375170210842043, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 1.14}) +[2026-06-23 15:42:20,929][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.16155225038528442, 'train/grad_norm_pre_clip': 54.54926681518555, 'timing/data': 0.0007004010258242488, 'timing/model': 1.937058944022283, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 1.14}) +[2026-06-23 15:42:24,770][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.09424847364425659, 'train/grad_norm_pre_clip': 55.93988800048828, 'timing/data': 0.0005937109235674143, 'timing/model': 1.9463217420270666, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 1.14}) +[2026-06-23 15:42:28,635][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.04879767447710037, 'train/grad_norm_pre_clip': 18.7222900390625, 'timing/data': 0.0005808989517390728, 'timing/model': 1.9288399070501328, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 1.14}) +[2026-06-23 15:42:32,443][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.010474830865859985, 'train/grad_norm_pre_clip': 28.298385620117188, 'timing/data': 0.0006403879960998893, 'timing/model': 1.931129968026653, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 1.14}) +[2026-06-23 15:42:36,250][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.027135508134961128, 'train/grad_norm_pre_clip': 37.738895416259766, 'timing/data': 0.0008533919462934136, 'timing/model': 1.9305171360028908, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 1.14}) +[2026-06-23 15:42:40,063][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.08235028386116028, 'train/grad_norm_pre_clip': 69.6674575805664, 'timing/data': 0.0005793430609628558, 'timing/model': 1.935511457035318, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 1.14}) +[2026-06-23 15:42:43,884][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.062243618071079254, 'train/grad_norm_pre_clip': 35.45738220214844, 'timing/data': 0.0006087740184739232, 'timing/model': 1.9392854480538517, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 1.15}) +[2026-06-23 15:42:47,677][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.0418415442109108, 'train/grad_norm_pre_clip': 21.860197067260742, 'timing/data': 0.0005841209786012769, 'timing/model': 1.9300000740913674, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 1.15}) +[2026-06-23 15:42:51,482][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.02664387784898281, 'train/grad_norm_pre_clip': 17.606143951416016, 'timing/data': 0.0006633329903706908, 'timing/model': 1.9309180669952184, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 1.15}) +[2026-06-23 15:42:55,284][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.08556092530488968, 'train/grad_norm_pre_clip': 27.195789337158203, 'timing/data': 0.0008836359484121203, 'timing/model': 1.9354137750342488, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 1.15}) +[2026-06-23 15:42:59,108][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.12412691116333008, 'train/grad_norm_pre_clip': 43.34378433227539, 'timing/data': 0.0005948090692982078, 'timing/model': 1.9517922940431163, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 1.15}) +[2026-06-23 15:43:02,894][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.041470423340797424, 'train/grad_norm_pre_clip': 21.6875057220459, 'timing/data': 0.0005974819650873542, 'timing/model': 1.9202062980039045, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 1.15}) +[2026-06-23 15:43:06,703][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.10736361145973206, 'train/grad_norm_pre_clip': 29.51287078857422, 'timing/data': 0.0006144959479570389, 'timing/model': 1.930290030897595, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 1.15}) +[2026-06-23 15:43:10,509][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.049027569591999054, 'train/grad_norm_pre_clip': 27.761693954467773, 'timing/data': 0.0005973719526082277, 'timing/model': 1.9308310720371082, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 1.15}) +[2026-06-23 15:43:14,304][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.053233761340379715, 'train/grad_norm_pre_clip': 31.924497604370117, 'timing/data': 0.0006204359233379364, 'timing/model': 1.9301429600454867, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 1.15}) +[2026-06-23 15:43:18,109][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.12949366867542267, 'train/grad_norm_pre_clip': 24.78937339782715, 'timing/data': 0.0007478949846699834, 'timing/model': 1.9260914389742538, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 1.15}) +[2026-06-23 15:43:21,907][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.018500497564673424, 'train/grad_norm_pre_clip': 27.32378387451172, 'timing/data': 0.0006577230524271727, 'timing/model': 1.932694896007888, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 1.15}) +[2026-06-23 15:43:25,719][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.09453783184289932, 'train/grad_norm_pre_clip': 48.07521438598633, 'timing/data': 0.0005978320259600878, 'timing/model': 1.93830386409536, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 1.15}) +[2026-06-23 15:43:29,519][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.05475682392716408, 'train/grad_norm_pre_clip': 23.724456787109375, 'timing/data': 0.0006016700062900782, 'timing/model': 1.9241902099456638, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 1.15}) +[2026-06-23 15:43:33,324][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.08683518320322037, 'train/grad_norm_pre_clip': 21.90671157836914, 'timing/data': 0.0005914770299568772, 'timing/model': 1.9245129349874333, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 1.15}) +[2026-06-23 15:43:37,129][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.056216612458229065, 'train/grad_norm_pre_clip': 24.49837875366211, 'timing/data': 0.000582645065151155, 'timing/model': 1.928395221941173, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 1.15}) +[2026-06-23 15:43:40,938][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.029134884476661682, 'train/grad_norm_pre_clip': 23.878299713134766, 'timing/data': 0.0005858370568603277, 'timing/model': 1.9410872880835086, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 1.15}) +[2026-06-23 15:43:44,756][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.04825522005558014, 'train/grad_norm_pre_clip': 23.725955963134766, 'timing/data': 0.0006085369968786836, 'timing/model': 1.9320334700169042, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 1.15}) +[2026-06-23 15:43:48,557][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.04961257055401802, 'train/grad_norm_pre_clip': 22.26589012145996, 'timing/data': 0.0006494830595329404, 'timing/model': 1.926755299908109, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 1.15}) +[2026-06-23 15:43:52,362][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.08185611665248871, 'train/grad_norm_pre_clip': 28.28793716430664, 'timing/data': 0.0006808590842410922, 'timing/model': 1.930794843006879, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 1.15}) +[2026-06-23 15:43:56,170][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.11308275908231735, 'train/grad_norm_pre_clip': 29.003387451171875, 'timing/data': 0.0005821450613439083, 'timing/model': 1.9357029420789331, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 1.15}) +[2026-06-23 15:43:59,970][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.05598732829093933, 'train/grad_norm_pre_clip': 37.264339447021484, 'timing/data': 0.0007762380409985781, 'timing/model': 1.9345638200175017, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 1.15}) +[2026-06-23 15:44:03,767][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.04578183963894844, 'train/grad_norm_pre_clip': 20.260862350463867, 'timing/data': 0.0005932989297434688, 'timing/model': 1.9324313029646873, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 1.15}) +[2026-06-23 15:44:07,574][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.08056976646184921, 'train/grad_norm_pre_clip': 26.82617950439453, 'timing/data': 0.0006035430124029517, 'timing/model': 1.9302249079337344, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 1.15}) +[2026-06-23 15:44:11,383][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.03284721449017525, 'train/grad_norm_pre_clip': 21.290456771850586, 'timing/data': 0.0006030560471117496, 'timing/model': 1.9313520810101181, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 1.15}) +[2026-06-23 15:44:15,193][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.053735487163066864, 'train/grad_norm_pre_clip': 28.029767990112305, 'timing/data': 0.0005744330119341612, 'timing/model': 1.9332451319787651, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 1.15}) +[2026-06-23 15:44:19,003][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.061352092772722244, 'train/grad_norm_pre_clip': 38.26756286621094, 'timing/data': 0.0008885529823601246, 'timing/model': 1.9337348350090906, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 1.15}) +[2026-06-23 15:44:22,802][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.0886869728565216, 'train/grad_norm_pre_clip': 45.165950775146484, 'timing/data': 0.0005984540330246091, 'timing/model': 1.9314213059842587, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 1.16}) +[2026-06-23 15:44:26,607][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.04687439650297165, 'train/grad_norm_pre_clip': 15.099649429321289, 'timing/data': 0.000604382948949933, 'timing/model': 1.930321078049019, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 1.16}) +[2026-06-23 15:44:30,399][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.0360790379345417, 'train/grad_norm_pre_clip': 33.219173431396484, 'timing/data': 0.0005882988916710019, 'timing/model': 1.9230288319522515, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 1.16}) +[2026-06-23 15:44:34,218][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.11955501139163971, 'train/grad_norm_pre_clip': 31.002607345581055, 'timing/data': 0.0006174689624458551, 'timing/model': 1.93320498499088, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 1.16}) +[2026-06-23 15:44:38,219][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.0415007509291172, 'train/grad_norm_pre_clip': 20.61965560913086, 'timing/data': 0.0008307509124279022, 'timing/model': 2.1252505800221115, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 1.16}) +[2026-06-23 15:44:42,013][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.04159468039870262, 'train/grad_norm_pre_clip': 19.999855041503906, 'timing/data': 0.0006011079531162977, 'timing/model': 1.9246776369400322, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 1.16}) +[2026-06-23 15:44:45,824][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.04077894613146782, 'train/grad_norm_pre_clip': 38.30101013183594, 'timing/data': 0.0005814640317112207, 'timing/model': 1.9324406660161912, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.16}) +[2026-06-23 15:44:49,636][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.03573015332221985, 'train/grad_norm_pre_clip': 13.106447219848633, 'timing/data': 0.0005748911062255502, 'timing/model': 1.9341567549854517, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 1.16}) +[2026-06-23 15:44:53,431][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.08964945375919342, 'train/grad_norm_pre_clip': 43.16248321533203, 'timing/data': 0.000584091991186142, 'timing/model': 1.926751377992332, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 1.16}) +[2026-06-23 15:44:57,262][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.08610295504331589, 'train/grad_norm_pre_clip': 29.335851669311523, 'timing/data': 0.0005926199955865741, 'timing/model': 1.9310940169962123, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 1.16}) +[2026-06-23 15:45:01,062][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.020677702501416206, 'train/grad_norm_pre_clip': 27.209142684936523, 'timing/data': 0.0005932890344411135, 'timing/model': 1.9197695900220424, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 1.16}) +[2026-06-23 15:45:04,883][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.03621210530400276, 'train/grad_norm_pre_clip': 20.33026695251465, 'timing/data': 0.0006040349835529923, 'timing/model': 1.9381706579588354, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 1.16}) +[2026-06-23 15:45:08,696][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.06059813126921654, 'train/grad_norm_pre_clip': 24.115333557128906, 'timing/data': 0.0005808740388602018, 'timing/model': 1.9327109040459618, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 1.16}) +[2026-06-23 15:45:12,692][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.09458022564649582, 'train/grad_norm_pre_clip': 49.948585510253906, 'timing/data': 0.0005892429035156965, 'timing/model': 1.9316463239956647, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 1.16}) +[2026-06-23 15:45:16,498][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.016037309542298317, 'train/grad_norm_pre_clip': 26.077268600463867, 'timing/data': 0.0005953209474682808, 'timing/model': 1.9348021369660273, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 1.16}) +[2026-06-23 15:45:20,296][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3000 +[2026-06-23 15:47:29,325][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3000 +[2026-06-23 15:47:29,327][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3000 +[2026-06-23 15:48:50,379][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3000 +[2026-06-23 15:57:01,402][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3000 +[2026-06-23 15:57:01,403][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.15023133158683777, 'train/grad_norm_pre_clip': 37.30182647705078, 'eval/action_loss/seconds': 129.0273146959953, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.035228878259658813, 'eval/latency_2/loss': 0.035228878259658813, 'eval/flappy/loss': 0.035228878259658813, 'eval/flappy/latency_2/loss': 0.035228878259658813, 'eval/flap_precision': 0.9670477471418965, 'eval/flap_recall': 0.970310391363023, 'eval/flap_f1': 0.9686763219939374, 'eval/flap_rate_model': 0.20652777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9686763219939374, 'eval/latency_2/flap_precision': 0.9670477471418965, 'eval/latency_2/flap_recall': 0.970310391363023, 'eval/latency_2/flap_f1': 0.9686763219939374, 'eval/latency_2/flap_rate_model': 0.20652777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9686763219939374, 'eval/action_classification/data_wait_seconds': 1.4048845767974854, 'eval/action_classification/predict_seconds': 79.41796112060547, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 300.4800000000097, 'rl_games_eval/mid_train/mean_length': 2440.8, 'rl_games_eval/mid_train/std_reward': 141.44521766395357, 'rl_games_eval/mid_train/std_length': 1138.7999648753068, 'rl_games_eval/mid_train/macro_mean_reward': 300.4800000000097, 'rl_games_eval/mid_train/macro_mean_length': 2440.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 300.4800000000097, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2440.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 141.44521766395357, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1138.7999648753068, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 300.4800000000097, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000574036967009306, 'timing/model': 1.9292271139565855, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 1.16}) +[2026-06-23 15:57:01,403][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_3000_state +[2026-06-23 15:57:01,406][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 15:57:05,144][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_3000_state/model.safetensors +[2026-06-23 15:57:08,014][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_3000_state/optimizer.bin +[2026-06-23 15:57:08,014][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_3000_state/sampler.bin +[2026-06-23 15:57:08,014][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-23 15:57:08,015][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-23 15:57:08,048][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-23 15:57:08,048][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 15:57:08,060][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 15:57:11,790][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.05029907822608948, 'train/grad_norm_pre_clip': 33.030147552490234, 'timing/data': 0.0008541489951312542, 'timing/model': 1.8771989579545334, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 1.16}) +[2026-06-23 15:57:15,522][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.07470957189798355, 'train/grad_norm_pre_clip': 62.51763916015625, 'timing/data': 0.000597959035076201, 'timing/model': 1.88709671003744, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 1.16}) +[2026-06-23 15:57:19,238][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.029258036985993385, 'train/grad_norm_pre_clip': 11.910768508911133, 'timing/data': 0.0006230700528249145, 'timing/model': 1.8826669129775837, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 1.16}) +[2026-06-23 15:57:22,982][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.025968190282583237, 'train/grad_norm_pre_clip': 14.083678245544434, 'timing/data': 0.0008095490047708154, 'timing/model': 1.9039016040042043, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 1.16}) +[2026-06-23 15:57:26,724][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.08569148182868958, 'train/grad_norm_pre_clip': 31.780561447143555, 'timing/data': 0.0007831279654055834, 'timing/model': 1.8976118930149823, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 1.16}) +[2026-06-23 15:57:30,464][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.08016716688871384, 'train/grad_norm_pre_clip': 38.79969024658203, 'timing/data': 0.0005780600477010012, 'timing/model': 1.9028247479582205, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 1.16}) +[2026-06-23 15:57:34,212][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.028108710423111916, 'train/grad_norm_pre_clip': 14.336265563964844, 'timing/data': 0.0005799020873382688, 'timing/model': 1.9026548070833087, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 1.16}) +[2026-06-23 15:57:37,968][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.04859780892729759, 'train/grad_norm_pre_clip': 50.87010955810547, 'timing/data': 0.0006893940735608339, 'timing/model': 1.904329746030271, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 1.16}) +[2026-06-23 15:57:41,718][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.05969604104757309, 'train/grad_norm_pre_clip': 24.90034294128418, 'timing/data': 0.0005871809553354979, 'timing/model': 1.900033910991624, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 1.16}) +[2026-06-23 15:57:45,496][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.059969544410705566, 'train/grad_norm_pre_clip': 36.18785858154297, 'timing/data': 0.000571846030652523, 'timing/model': 1.9216410520020872, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 1.16}) +[2026-06-23 15:57:49,279][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.09056664258241653, 'train/grad_norm_pre_clip': 49.8271598815918, 'timing/data': 0.0005837499629706144, 'timing/model': 1.9132605410413817, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 1.17}) +[2026-06-23 15:57:53,056][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.045075997710227966, 'train/grad_norm_pre_clip': 38.4088134765625, 'timing/data': 0.0006465089973062277, 'timing/model': 1.9060538640478626, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 1.17}) +[2026-06-23 15:57:56,832][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.031062224879860878, 'train/grad_norm_pre_clip': 20.330068588256836, 'timing/data': 0.000662298989482224, 'timing/model': 1.918190438998863, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 1.17}) +[2026-06-23 15:58:00,609][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.10167926549911499, 'train/grad_norm_pre_clip': 35.21490478515625, 'timing/data': 0.0006088860100135207, 'timing/model': 1.9169176679570228, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 1.17}) +[2026-06-23 15:58:04,562][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.028436830267310143, 'train/grad_norm_pre_clip': 31.72274398803711, 'timing/data': 0.0005970050115138292, 'timing/model': 2.097775962902233, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 1.17}) +[2026-06-23 15:58:08,350][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.04796427860856056, 'train/grad_norm_pre_clip': 32.20841979980469, 'timing/data': 0.0006374820368364453, 'timing/model': 1.914333546999842, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 1.17}) +[2026-06-23 15:58:12,122][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.16680797934532166, 'train/grad_norm_pre_clip': 20.71563720703125, 'timing/data': 0.0005774340825155377, 'timing/model': 1.9120208260137588, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 1.17}) +[2026-06-23 15:58:15,892][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.049928855150938034, 'train/grad_norm_pre_clip': 18.907493591308594, 'timing/data': 0.000603110995143652, 'timing/model': 1.9077476389938965, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 1.17}) +[2026-06-23 15:58:19,665][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.053462617099285126, 'train/grad_norm_pre_clip': 26.695117950439453, 'timing/data': 0.0007236399687826633, 'timing/model': 1.9185841080034152, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 1.17}) +[2026-06-23 15:58:23,433][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.08149009943008423, 'train/grad_norm_pre_clip': 28.40968894958496, 'timing/data': 0.0006109530804678798, 'timing/model': 1.9139658189378679, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 1.17}) +[2026-06-23 15:58:27,219][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.0464913547039032, 'train/grad_norm_pre_clip': 44.70663833618164, 'timing/data': 0.0005929809994995594, 'timing/model': 1.921971051953733, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 1.17}) +[2026-06-23 15:58:31,010][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.05369530990719795, 'train/grad_norm_pre_clip': 34.28174591064453, 'timing/data': 0.0005914400098845363, 'timing/model': 1.9200782530242577, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 1.17}) +[2026-06-23 15:58:34,793][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.07628334313631058, 'train/grad_norm_pre_clip': 42.608917236328125, 'timing/data': 0.0007321870652958751, 'timing/model': 1.9177205408923328, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 1.17}) +[2026-06-23 15:58:38,743][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.04410376027226448, 'train/grad_norm_pre_clip': 25.129056930541992, 'timing/data': 0.0005950049962848425, 'timing/model': 1.9130808640038595, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 1.17}) +[2026-06-23 15:58:42,514][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.007498550694435835, 'train/grad_norm_pre_clip': 14.869733810424805, 'timing/data': 0.0005656389985233545, 'timing/model': 1.9148436889518052, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 1.17}) +[2026-06-23 15:58:46,309][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.09709097445011139, 'train/grad_norm_pre_clip': 33.52963638305664, 'timing/data': 0.0006279830122366548, 'timing/model': 1.929279562085867, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 1.17}) +[2026-06-23 15:58:50,116][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.07072629034519196, 'train/grad_norm_pre_clip': 23.866628646850586, 'timing/data': 0.0006824380252510309, 'timing/model': 1.9365355699555948, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 1.17}) +[2026-06-23 15:58:53,908][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.0568992979824543, 'train/grad_norm_pre_clip': 18.899124145507812, 'timing/data': 0.000652504968456924, 'timing/model': 1.9272043909877539, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 1.17}) +[2026-06-23 15:58:57,707][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.051582999527454376, 'train/grad_norm_pre_clip': 35.367408752441406, 'timing/data': 0.0006257930072024465, 'timing/model': 1.9320557760074735, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 1.17}) +[2026-06-23 15:59:01,498][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.12661872804164886, 'train/grad_norm_pre_clip': 44.77567672729492, 'timing/data': 0.0005995420506224036, 'timing/model': 1.9250110799912363, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 1.17}) +[2026-06-23 15:59:05,287][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.0874970331788063, 'train/grad_norm_pre_clip': 42.881473541259766, 'timing/data': 0.0007972670719027519, 'timing/model': 1.9143672089558095, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 1.17}) +[2026-06-23 15:59:09,263][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.025411881506443024, 'train/grad_norm_pre_clip': 17.767765045166016, 'timing/data': 0.0005909290630370378, 'timing/model': 2.106927855988033, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 1.17}) +[2026-06-23 15:59:13,062][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.12082091718912125, 'train/grad_norm_pre_clip': 37.22279739379883, 'timing/data': 0.0005870499880984426, 'timing/model': 1.9226700470317155, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 1.17}) +[2026-06-23 15:59:16,914][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.09241797775030136, 'train/grad_norm_pre_clip': 23.64980125427246, 'timing/data': 0.0006720050005242229, 'timing/model': 1.931663200026378, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 1.17}) +[2026-06-23 15:59:20,719][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.049915190786123276, 'train/grad_norm_pre_clip': 25.2007999420166, 'timing/data': 0.000737329013645649, 'timing/model': 1.9213248960440978, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 1.17}) +[2026-06-23 15:59:24,517][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.21750973165035248, 'train/grad_norm_pre_clip': 53.88543701171875, 'timing/data': 0.0005961949937045574, 'timing/model': 1.9154387150192633, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 1.17}) +[2026-06-23 15:59:28,328][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.04047076776623726, 'train/grad_norm_pre_clip': 31.470653533935547, 'timing/data': 0.0007890879642218351, 'timing/model': 1.943397986004129, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 1.18}) +[2026-06-23 15:59:32,204][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.0733572393655777, 'train/grad_norm_pre_clip': 31.058046340942383, 'timing/data': 0.0005883920239284635, 'timing/model': 1.9657180780777708, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 1.18}) +[2026-06-23 15:59:36,000][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.10076455771923065, 'train/grad_norm_pre_clip': 34.23497009277344, 'timing/data': 0.0006373460637405515, 'timing/model': 1.9255143610062078, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 1.18}) +[2026-06-23 15:59:39,794][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.030745824798941612, 'train/grad_norm_pre_clip': 11.612444877624512, 'timing/data': 0.0006108570378273726, 'timing/model': 1.9302066139644012, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.18}) +[2026-06-23 15:59:43,758][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.019967718049883842, 'train/grad_norm_pre_clip': 21.684968948364258, 'timing/data': 0.0011568029876798391, 'timing/model': 1.9159061019308865, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 1.18}) +[2026-06-23 15:59:47,546][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.026239195838570595, 'train/grad_norm_pre_clip': 32.05469512939453, 'timing/data': 0.0005988630000501871, 'timing/model': 1.9233380829682574, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 1.18}) +[2026-06-23 15:59:51,344][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.038979850709438324, 'train/grad_norm_pre_clip': 18.336856842041016, 'timing/data': 0.0005861170357093215, 'timing/model': 1.9300590800121427, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 1.18}) +[2026-06-23 15:59:55,143][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.030917145311832428, 'train/grad_norm_pre_clip': 29.834749221801758, 'timing/data': 0.0006015810649842024, 'timing/model': 1.925223910016939, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 1.18}) +[2026-06-23 15:59:58,935][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.0038919912185519934, 'train/grad_norm_pre_clip': 22.941112518310547, 'timing/data': 0.0006022138986736536, 'timing/model': 1.9272246300242841, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 1.18}) +[2026-06-23 16:00:02,745][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.11917554587125778, 'train/grad_norm_pre_clip': 40.776737213134766, 'timing/data': 0.0005854950286448002, 'timing/model': 1.9352209600619972, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 1.18}) +[2026-06-23 16:00:06,528][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.007401659153401852, 'train/grad_norm_pre_clip': 20.99209976196289, 'timing/data': 0.0007818170124664903, 'timing/model': 1.916588393971324, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 1.18}) +[2026-06-23 16:00:10,331][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.09089808166027069, 'train/grad_norm_pre_clip': 37.176513671875, 'timing/data': 0.0005833740578964353, 'timing/model': 1.9220366380177438, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 1.18}) +[2026-06-23 16:00:14,256][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.07476874440908432, 'train/grad_norm_pre_clip': 70.76090240478516, 'timing/data': 0.0007441600318998098, 'timing/model': 2.0536415689857677, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 1.18}) +[2026-06-23 16:00:18,060][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.027964284643530846, 'train/grad_norm_pre_clip': 37.3665885925293, 'timing/data': 0.000705994083546102, 'timing/model': 1.9379365289350972, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 1.18}) +[2026-06-23 16:00:21,871][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.03433501720428467, 'train/grad_norm_pre_clip': 13.460131645202637, 'timing/data': 0.0005817109486088157, 'timing/model': 1.9393083970062435, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 1.18}) +[2026-06-23 16:00:25,666][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.056218232959508896, 'train/grad_norm_pre_clip': 23.402175903320312, 'timing/data': 0.0007112250896170735, 'timing/model': 1.9192103720270097, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 1.18}) +[2026-06-23 16:00:29,474][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.03382955864071846, 'train/grad_norm_pre_clip': 15.96294116973877, 'timing/data': 0.0006003470625728369, 'timing/model': 1.9261964660836384, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 1.18}) +[2026-06-23 16:00:33,288][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.04930286109447479, 'train/grad_norm_pre_clip': 25.391557693481445, 'timing/data': 0.0006631130818277597, 'timing/model': 1.9329690709710121, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 1.18}) +[2026-06-23 16:00:37,089][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.06696639209985733, 'train/grad_norm_pre_clip': 29.316055297851562, 'timing/data': 0.0005822099046781659, 'timing/model': 1.9363968319958076, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 1.18}) +[2026-06-23 16:00:40,917][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.045320577919483185, 'train/grad_norm_pre_clip': 25.096464157104492, 'timing/data': 0.0007963730022311211, 'timing/model': 1.9414438200183213, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 1.18}) +[2026-06-23 16:00:44,727][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.10791131854057312, 'train/grad_norm_pre_clip': 15.964776992797852, 'timing/data': 0.0005858769873157144, 'timing/model': 1.9374521669233218, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 1.18}) +[2026-06-23 16:00:48,692][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.04722514748573303, 'train/grad_norm_pre_clip': 38.35194778442383, 'timing/data': 0.0005982189904898405, 'timing/model': 1.9357372239464894, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 1.18}) +[2026-06-23 16:00:52,498][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.08495911210775375, 'train/grad_norm_pre_clip': 26.783342361450195, 'timing/data': 0.0005829479778185487, 'timing/model': 1.9342537220800295, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 1.18}) +[2026-06-23 16:00:56,284][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.09718472510576248, 'train/grad_norm_pre_clip': 49.840030670166016, 'timing/data': 0.0008087329333648086, 'timing/model': 1.9227083519799635, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 1.18}) +[2026-06-23 16:01:00,064][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.07731863111257553, 'train/grad_norm_pre_clip': 20.55390739440918, 'timing/data': 0.0006210750434547663, 'timing/model': 1.916226984001696, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 1.18}) +[2026-06-23 16:01:03,862][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.030311690643429756, 'train/grad_norm_pre_clip': 27.589645385742188, 'timing/data': 0.0009310200111940503, 'timing/model': 1.9229834589641541, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 1.19}) +[2026-06-23 16:01:07,661][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.053363412618637085, 'train/grad_norm_pre_clip': 14.461761474609375, 'timing/data': 0.0006587359821423888, 'timing/model': 1.9373214290244505, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 1.19}) +[2026-06-23 16:01:11,468][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.021467337384819984, 'train/grad_norm_pre_clip': 12.849150657653809, 'timing/data': 0.0006846069591119885, 'timing/model': 1.9298103960463777, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 1.19}) +[2026-06-23 16:01:15,264][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.05956822633743286, 'train/grad_norm_pre_clip': 50.1027717590332, 'timing/data': 0.0005890749162063003, 'timing/model': 1.9252639600308612, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 1.19}) +[2026-06-23 16:01:19,247][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.04629252851009369, 'train/grad_norm_pre_clip': 34.8124885559082, 'timing/data': 0.0006544929929077625, 'timing/model': 2.106828502030112, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 1.19}) +[2026-06-23 16:01:23,058][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.09008888155221939, 'train/grad_norm_pre_clip': 26.647077560424805, 'timing/data': 0.0008535549277439713, 'timing/model': 1.9368103459710255, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 1.19}) +[2026-06-23 16:01:26,870][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.058435048907995224, 'train/grad_norm_pre_clip': 24.304868698120117, 'timing/data': 0.0005870029563084245, 'timing/model': 1.931226241053082, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 1.19}) +[2026-06-23 16:01:30,690][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.04009987786412239, 'train/grad_norm_pre_clip': 30.191102981567383, 'timing/data': 0.0006887390045449138, 'timing/model': 1.9363411499653012, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 1.19}) +[2026-06-23 16:01:34,486][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.08295665681362152, 'train/grad_norm_pre_clip': 26.730587005615234, 'timing/data': 0.0005950919585302472, 'timing/model': 1.9234736689832062, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 1.19}) +[2026-06-23 16:01:38,272][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.03382759913802147, 'train/grad_norm_pre_clip': 41.775657653808594, 'timing/data': 0.00060048489831388, 'timing/model': 1.920439547044225, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 1.19}) +[2026-06-23 16:01:42,086][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.07861888408660889, 'train/grad_norm_pre_clip': 26.311630249023438, 'timing/data': 0.0006163920043036342, 'timing/model': 1.936494339024648, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 1.19}) +[2026-06-23 16:01:45,891][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.05623476579785347, 'train/grad_norm_pre_clip': 39.01193618774414, 'timing/data': 0.0005876069189980626, 'timing/model': 1.933223752072081, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 1.19}) +[2026-06-23 16:01:49,692][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.022791031748056412, 'train/grad_norm_pre_clip': 12.72244930267334, 'timing/data': 0.0006108660018071532, 'timing/model': 1.9290186299476773, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 1.19}) +[2026-06-23 16:01:53,670][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.004415029194205999, 'train/grad_norm_pre_clip': 13.602768898010254, 'timing/data': 0.0005823860410600901, 'timing/model': 1.9319397859508172, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 1.19}) +[2026-06-23 16:01:57,490][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.012044546194374561, 'train/grad_norm_pre_clip': 11.045027732849121, 'timing/data': 0.0005947320023551583, 'timing/model': 1.9383201010059565, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 1.19}) +[2026-06-23 16:02:01,325][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.05435553938150406, 'train/grad_norm_pre_clip': 32.15208053588867, 'timing/data': 0.0006146400701254606, 'timing/model': 1.926376119023189, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 1.19}) +[2026-06-23 16:02:05,137][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.05945946276187897, 'train/grad_norm_pre_clip': 19.380626678466797, 'timing/data': 0.0005847570719197392, 'timing/model': 1.9326664070831612, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 1.19}) +[2026-06-23 16:02:08,920][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.008161814883351326, 'train/grad_norm_pre_clip': 18.725067138671875, 'timing/data': 0.0009125939104706049, 'timing/model': 1.9189790700329468, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 1.19}) +[2026-06-23 16:02:12,710][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.03838803619146347, 'train/grad_norm_pre_clip': 41.26816177368164, 'timing/data': 0.0007285280153155327, 'timing/model': 1.9190680540632457, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 1.19}) +[2026-06-23 16:02:16,513][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.018005868420004845, 'train/grad_norm_pre_clip': 37.378074645996094, 'timing/data': 0.0006975519936531782, 'timing/model': 1.9299717139219865, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 1.19}) +[2026-06-23 16:02:20,311][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.16975568234920502, 'train/grad_norm_pre_clip': 47.0814208984375, 'timing/data': 0.0005851029418408871, 'timing/model': 1.9242101600393653, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 1.19}) +[2026-06-23 16:02:24,107][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.01708092726767063, 'train/grad_norm_pre_clip': 54.077850341796875, 'timing/data': 0.0005754079902544618, 'timing/model': 1.9315577080706134, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 1.19}) +[2026-06-23 16:02:28,058][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.0659780502319336, 'train/grad_norm_pre_clip': 59.3213005065918, 'timing/data': 0.00062568299472332, 'timing/model': 1.9281250020721927, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 1.19}) +[2026-06-23 16:02:31,851][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.05176051706075668, 'train/grad_norm_pre_clip': 34.560543060302734, 'timing/data': 0.0005846440326422453, 'timing/model': 1.9292948639485985, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 1.19}) +[2026-06-23 16:02:35,645][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.07844367623329163, 'train/grad_norm_pre_clip': 24.342269897460938, 'timing/data': 0.0005806029075756669, 'timing/model': 1.9284058519406244, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 1.19}) +[2026-06-23 16:02:39,447][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.08094814419746399, 'train/grad_norm_pre_clip': 24.25247573852539, 'timing/data': 0.0006732409819960594, 'timing/model': 1.932945283013396, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 1.19}) +[2026-06-23 16:02:43,243][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.06650154292583466, 'train/grad_norm_pre_clip': 53.135337829589844, 'timing/data': 0.0006120329489931464, 'timing/model': 1.9325824860716239, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 1.2}) +[2026-06-23 16:02:47,038][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.15578487515449524, 'train/grad_norm_pre_clip': 41.92502975463867, 'timing/data': 0.0007833499694243073, 'timing/model': 1.9277318109525368, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.2}) +[2026-06-23 16:02:50,847][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.0712244063615799, 'train/grad_norm_pre_clip': 27.0932674407959, 'timing/data': 0.0006473279790952802, 'timing/model': 1.9280555580044165, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 1.2}) +[2026-06-23 16:02:54,644][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.13126735389232635, 'train/grad_norm_pre_clip': 38.791866302490234, 'timing/data': 0.0006024069152772427, 'timing/model': 1.9305735930101946, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 1.2}) +[2026-06-23 16:02:58,608][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.009289348497986794, 'train/grad_norm_pre_clip': 16.558109283447266, 'timing/data': 0.0006642109947279096, 'timing/model': 2.0992284069070593, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 1.2}) +[2026-06-23 16:03:02,417][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.009646771475672722, 'train/grad_norm_pre_clip': 41.0355339050293, 'timing/data': 0.0005787099944427609, 'timing/model': 1.937669072067365, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 1.2}) +[2026-06-23 16:03:06,209][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.02721165306866169, 'train/grad_norm_pre_clip': 8.460630416870117, 'timing/data': 0.0005915160290896893, 'timing/model': 1.928032831987366, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 1.2}) +[2026-06-23 16:03:10,009][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.03182071074843407, 'train/grad_norm_pre_clip': 16.06015968322754, 'timing/data': 0.0005875800270587206, 'timing/model': 1.9290052549913526, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 1.2}) +[2026-06-23 16:03:13,833][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.1054920181632042, 'train/grad_norm_pre_clip': 17.047060012817383, 'timing/data': 0.0005937549285590649, 'timing/model': 1.9391991819720715, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 1.2}) +[2026-06-23 16:03:17,632][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.01877017319202423, 'train/grad_norm_pre_clip': 12.95102310180664, 'timing/data': 0.0005785929970443249, 'timing/model': 1.927241099998355, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 1.2}) +[2026-06-23 16:03:21,452][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.08640085905790329, 'train/grad_norm_pre_clip': 20.145389556884766, 'timing/data': 0.0006959140300750732, 'timing/model': 1.9436303459806368, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 1.2}) +[2026-06-23 16:03:25,235][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.0276210755109787, 'train/grad_norm_pre_clip': 35.87272644042969, 'timing/data': 0.000594708020798862, 'timing/model': 1.9206205049995333, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 1.2}) +[2026-06-23 16:03:29,045][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3100 +[2026-06-23 16:05:37,068][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3100 +[2026-06-23 16:05:37,070][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3100 +[2026-06-23 16:06:59,618][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3100 +[2026-06-23 16:06:59,619][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.027862681075930595, 'train/grad_norm_pre_clip': 53.4527587890625, 'eval/action_loss/seconds': 128.02176119596697, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.038271356374025345, 'eval/latency_2/loss': 0.038271356374025345, 'eval/flappy/loss': 0.038271356374025345, 'eval/flappy/latency_2/loss': 0.038271356374025345, 'eval/flap_precision': 0.95385629531971, 'eval/flap_recall': 0.9763832658569501, 'eval/flap_f1': 0.9649883294431477, 'eval/flap_rate_model': 0.21069444444444443, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9649883294431478, 'eval/latency_2/flap_precision': 0.95385629531971, 'eval/latency_2/flap_recall': 0.9763832658569501, 'eval/latency_2/flap_f1': 0.9649883294431477, 'eval/latency_2/flap_rate_model': 0.21069444444444443, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9649883294431478, 'eval/action_classification/data_wait_seconds': 1.5510441064834595, 'eval/action_classification/predict_seconds': 80.80986022949219, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006094640120863914, 'timing/model': 1.9319105290342122, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 1.2}) +[2026-06-23 16:07:03,405][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.03238551691174507, 'train/grad_norm_pre_clip': 21.312641143798828, 'timing/data': 0.0006464190082624555, 'timing/model': 1.9238466779934242, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 1.2}) +[2026-06-23 16:07:07,179][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.03486544266343117, 'train/grad_norm_pre_clip': 30.28064727783203, 'timing/data': 0.0006031399825587869, 'timing/model': 1.9173749609617516, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 1.2}) +[2026-06-23 16:07:10,966][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.08561467379331589, 'train/grad_norm_pre_clip': 26.935150146484375, 'timing/data': 0.0006098130252212286, 'timing/model': 1.928317275014706, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 1.2}) +[2026-06-23 16:07:14,748][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.08624403923749924, 'train/grad_norm_pre_clip': 49.87370681762695, 'timing/data': 0.0010844760108739138, 'timing/model': 1.9199303080094978, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 1.2}) +[2026-06-23 16:07:18,544][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.0376618355512619, 'train/grad_norm_pre_clip': 25.333879470825195, 'timing/data': 0.0006326950388029218, 'timing/model': 1.9267482860013843, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 1.2}) +[2026-06-23 16:07:22,508][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.12143412232398987, 'train/grad_norm_pre_clip': 66.35082244873047, 'timing/data': 0.000579686020500958, 'timing/model': 1.9179758480750024, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 1.2}) +[2026-06-23 16:07:26,297][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.006173060741275549, 'train/grad_norm_pre_clip': 4.881967067718506, 'timing/data': 0.0006540680769830942, 'timing/model': 1.9248862769454718, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 1.2}) +[2026-06-23 16:07:30,087][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.011146259494125843, 'train/grad_norm_pre_clip': 38.858455657958984, 'timing/data': 0.0005888169398531318, 'timing/model': 1.9227005189750344, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 1.2}) +[2026-06-23 16:07:33,881][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.10032631456851959, 'train/grad_norm_pre_clip': 32.43198776245117, 'timing/data': 0.0006018439307808876, 'timing/model': 1.9320239870576188, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 1.2}) +[2026-06-23 16:07:37,674][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.10579154640436172, 'train/grad_norm_pre_clip': 52.443260192871094, 'timing/data': 0.0008323340443894267, 'timing/model': 1.9294688329100609, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 1.2}) +[2026-06-23 16:07:41,464][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.02694697678089142, 'train/grad_norm_pre_clip': 20.029977798461914, 'timing/data': 0.0005817690398544073, 'timing/model': 1.9198227319866419, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 1.2}) +[2026-06-23 16:07:45,252][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.013779888860881329, 'train/grad_norm_pre_clip': 34.059181213378906, 'timing/data': 0.0006015470717102289, 'timing/model': 1.9259969720151275, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 1.2}) +[2026-06-23 16:07:49,044][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.03298277035355568, 'train/grad_norm_pre_clip': 41.64462661743164, 'timing/data': 0.0006052380194887519, 'timing/model': 1.9254301839973778, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 1.2}) +[2026-06-23 16:07:53,018][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.09233353286981583, 'train/grad_norm_pre_clip': 45.15445327758789, 'timing/data': 0.0005853479960933328, 'timing/model': 2.109998843050562, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 1.21}) +[2026-06-23 16:07:56,808][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.17718029022216797, 'train/grad_norm_pre_clip': 73.07476043701172, 'timing/data': 0.0007708959747105837, 'timing/model': 1.9317934790160507, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 1.21}) +[2026-06-23 16:08:00,586][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.06097111850976944, 'train/grad_norm_pre_clip': 67.30097961425781, 'timing/data': 0.0008852219907566905, 'timing/model': 1.9178411239990965, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 1.21}) +[2026-06-23 16:08:04,380][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.0762505754828453, 'train/grad_norm_pre_clip': 57.99884033203125, 'timing/data': 0.000613613985478878, 'timing/model': 1.9259405799675733, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 1.21}) +[2026-06-23 16:08:08,185][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.08669936656951904, 'train/grad_norm_pre_clip': 42.40129852294922, 'timing/data': 0.000598150072619319, 'timing/model': 1.9335607269313186, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 1.21}) +[2026-06-23 16:08:11,968][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.05428442731499672, 'train/grad_norm_pre_clip': 37.0953369140625, 'timing/data': 0.0005799270002171397, 'timing/model': 1.9226324410410598, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 1.21}) +[2026-06-23 16:08:15,759][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.1337485909461975, 'train/grad_norm_pre_clip': 58.593204498291016, 'timing/data': 0.0005950650665909052, 'timing/model': 1.919725724030286, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 1.21}) +[2026-06-23 16:08:19,555][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.018604211509227753, 'train/grad_norm_pre_clip': 23.958847045898438, 'timing/data': 0.0005981050198897719, 'timing/model': 1.9195260150590912, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 1.21}) +[2026-06-23 16:08:23,351][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.06535892933607101, 'train/grad_norm_pre_clip': 43.30874252319336, 'timing/data': 0.0006119840545579791, 'timing/model': 1.9221226189984009, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 1.21}) +[2026-06-23 16:08:27,326][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.09071669727563858, 'train/grad_norm_pre_clip': 44.48788833618164, 'timing/data': 0.0006031660595908761, 'timing/model': 1.9216205540578812, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 1.21}) +[2026-06-23 16:08:31,130][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.033422328531742096, 'train/grad_norm_pre_clip': 31.109148025512695, 'timing/data': 0.0005938649410381913, 'timing/model': 1.9328108539339155, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 1.21}) +[2026-06-23 16:08:34,928][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.14420099556446075, 'train/grad_norm_pre_clip': 55.17171859741211, 'timing/data': 0.0005798470228910446, 'timing/model': 1.9309168260078877, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 1.21}) +[2026-06-23 16:08:38,721][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.015056495554745197, 'train/grad_norm_pre_clip': 6.363494873046875, 'timing/data': 0.0006028229836374521, 'timing/model': 1.9229351689573377, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 1.21}) +[2026-06-23 16:08:42,511][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.0732806995511055, 'train/grad_norm_pre_clip': 26.058679580688477, 'timing/data': 0.0006101740291342139, 'timing/model': 1.9243081390159205, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 1.21}) +[2026-06-23 16:08:46,312][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.0430794283747673, 'train/grad_norm_pre_clip': 23.691993713378906, 'timing/data': 0.0007039520423859358, 'timing/model': 1.9291247449582443, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 1.21}) +[2026-06-23 16:08:50,105][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.01306326873600483, 'train/grad_norm_pre_clip': 34.516571044921875, 'timing/data': 0.0005845479900017381, 'timing/model': 1.9225580639904365, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 1.21}) +[2026-06-23 16:08:53,915][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.035184744745492935, 'train/grad_norm_pre_clip': 34.946250915527344, 'timing/data': 0.0006196289323270321, 'timing/model': 1.9289825840387493, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 1.21}) +[2026-06-23 16:08:57,939][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.038437049835920334, 'train/grad_norm_pre_clip': 38.77391052246094, 'timing/data': 0.0008165360195562243, 'timing/model': 2.1486551020061597, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 1.21}) +[2026-06-23 16:09:01,726][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.017774812877178192, 'train/grad_norm_pre_clip': 9.813568115234375, 'timing/data': 0.0006243269890546799, 'timing/model': 1.9250565289985389, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 1.21}) +[2026-06-23 16:09:05,534][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.01855713315308094, 'train/grad_norm_pre_clip': 27.55165672302246, 'timing/data': 0.0005949520273134112, 'timing/model': 1.9334094950463623, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 1.21}) +[2026-06-23 16:09:09,327][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.044077228754758835, 'train/grad_norm_pre_clip': 35.81882095336914, 'timing/data': 0.0005890349857509136, 'timing/model': 1.9251258539734408, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 1.21}) +[2026-06-23 16:09:13,125][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.062144290655851364, 'train/grad_norm_pre_clip': 23.47968292236328, 'timing/data': 0.0005990599747747183, 'timing/model': 1.9317264839773998, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 1.21}) +[2026-06-23 16:09:16,918][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.012187439948320389, 'train/grad_norm_pre_clip': 34.69866180419922, 'timing/data': 0.0006080500315874815, 'timing/model': 1.9292181509081274, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 1.21}) +[2026-06-23 16:09:20,714][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.08109906315803528, 'train/grad_norm_pre_clip': 88.46429443359375, 'timing/data': 0.0006381460698321462, 'timing/model': 1.9294194370741025, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 1.21}) +[2026-06-23 16:09:24,528][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.05044999346137047, 'train/grad_norm_pre_clip': 40.30609893798828, 'timing/data': 0.0006000419380143285, 'timing/model': 1.9534149010432884, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.21}) +[2026-06-23 16:09:28,321][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.07735686004161835, 'train/grad_norm_pre_clip': 24.33201789855957, 'timing/data': 0.0006950199604034424, 'timing/model': 1.9255752810277045, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 1.21}) +[2026-06-23 16:09:32,126][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.0455046072602272, 'train/grad_norm_pre_clip': 30.67327880859375, 'timing/data': 0.0005918530514463782, 'timing/model': 1.9296025111107156, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 1.22}) +[2026-06-23 16:09:35,934][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.08786657452583313, 'train/grad_norm_pre_clip': 23.229576110839844, 'timing/data': 0.000719055999070406, 'timing/model': 1.9302928980905563, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 1.22}) +[2026-06-23 16:09:39,732][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.050939206033945084, 'train/grad_norm_pre_clip': 20.40974998474121, 'timing/data': 0.0006687029963359237, 'timing/model': 1.9351720219710842, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 1.22}) +[2026-06-23 16:09:43,532][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.2117519974708557, 'train/grad_norm_pre_clip': 42.969635009765625, 'timing/data': 0.0005913280183449388, 'timing/model': 1.9225748060271144, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 1.22}) +[2026-06-23 16:09:47,322][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.02381785959005356, 'train/grad_norm_pre_clip': 34.65069580078125, 'timing/data': 0.0006242349045351148, 'timing/model': 1.919123366009444, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 1.22}) +[2026-06-23 16:09:51,133][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.09379800409078598, 'train/grad_norm_pre_clip': 46.76961135864258, 'timing/data': 0.0005933979991823435, 'timing/model': 1.942969391006045, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 1.22}) +[2026-06-23 16:09:54,930][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.0751069113612175, 'train/grad_norm_pre_clip': 47.74702072143555, 'timing/data': 0.0005906029837206006, 'timing/model': 1.925670136930421, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 1.22}) +[2026-06-23 16:09:58,723][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.03035399317741394, 'train/grad_norm_pre_clip': 19.451894760131836, 'timing/data': 0.0006579599576070905, 'timing/model': 1.9189069479471073, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 1.22}) +[2026-06-23 16:10:02,517][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.02346339449286461, 'train/grad_norm_pre_clip': 19.317781448364258, 'timing/data': 0.0009328880114480853, 'timing/model': 1.9275334780104458, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 1.22}) +[2026-06-23 16:10:06,302][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.0898885428905487, 'train/grad_norm_pre_clip': 23.262178421020508, 'timing/data': 0.0006319870008155704, 'timing/model': 1.9210636570351198, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 1.22}) +[2026-06-23 16:10:10,113][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.021298784762620926, 'train/grad_norm_pre_clip': 12.944076538085938, 'timing/data': 0.0005932850763201714, 'timing/model': 1.9347105369670317, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 1.22}) +[2026-06-23 16:10:13,920][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.03163345158100128, 'train/grad_norm_pre_clip': 36.54481506347656, 'timing/data': 0.0005975579842925072, 'timing/model': 1.931002754950896, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 1.22}) +[2026-06-23 16:10:17,716][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.010514367371797562, 'train/grad_norm_pre_clip': 24.864343643188477, 'timing/data': 0.0006509419763460755, 'timing/model': 1.931770452996716, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 1.22}) +[2026-06-23 16:10:21,520][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.018903328105807304, 'train/grad_norm_pre_clip': 7.919032096862793, 'timing/data': 0.0006001709261909127, 'timing/model': 1.9265794740058482, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 1.22}) +[2026-06-23 16:10:25,340][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.05196371674537659, 'train/grad_norm_pre_clip': 23.21516227722168, 'timing/data': 0.000591302989050746, 'timing/model': 1.9345241559203714, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 1.22}) +[2026-06-23 16:10:29,152][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.035107601433992386, 'train/grad_norm_pre_clip': 36.745948791503906, 'timing/data': 0.0005968979094177485, 'timing/model': 1.9373640440171584, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 1.22}) +[2026-06-23 16:10:32,943][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.10281472653150558, 'train/grad_norm_pre_clip': 27.713090896606445, 'timing/data': 0.0005964329466223717, 'timing/model': 1.9181568060303107, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 1.22}) +[2026-06-23 16:10:36,940][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.02596970461308956, 'train/grad_norm_pre_clip': 41.84203338623047, 'timing/data': 0.00072809390258044, 'timing/model': 2.12110718106851, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 1.22}) +[2026-06-23 16:10:40,746][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.004123488441109657, 'train/grad_norm_pre_clip': 16.889158248901367, 'timing/data': 0.0005956899840384722, 'timing/model': 1.9353890540078282, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 1.22}) +[2026-06-23 16:10:44,549][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.11612766236066818, 'train/grad_norm_pre_clip': 43.8291130065918, 'timing/data': 0.0006275529740378261, 'timing/model': 1.93082178698387, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 1.22}) +[2026-06-23 16:10:48,330][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.0691397488117218, 'train/grad_norm_pre_clip': 26.50022315979004, 'timing/data': 0.0005911809857934713, 'timing/model': 1.9188498069997877, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 1.22}) +[2026-06-23 16:10:52,143][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.02946227230131626, 'train/grad_norm_pre_clip': 39.7661247253418, 'timing/data': 0.0006133250426501036, 'timing/model': 1.938742593047209, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 1.22}) +[2026-06-23 16:10:55,940][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.07804998010396957, 'train/grad_norm_pre_clip': 28.20859146118164, 'timing/data': 0.0007274700328707695, 'timing/model': 1.9193449938902631, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 1.22}) +[2026-06-23 16:10:59,740][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.015074820257723331, 'train/grad_norm_pre_clip': 20.873897552490234, 'timing/data': 0.0005775629542768002, 'timing/model': 1.9302573889726773, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 1.22}) +[2026-06-23 16:11:03,544][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.012839599512517452, 'train/grad_norm_pre_clip': 22.217010498046875, 'timing/data': 0.0005917850648984313, 'timing/model': 1.9291825060499832, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 1.22}) +[2026-06-23 16:11:07,360][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.028683507815003395, 'train/grad_norm_pre_clip': 32.623905181884766, 'timing/data': 0.0007168829906731844, 'timing/model': 1.9306905990233645, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 1.22}) +[2026-06-23 16:11:11,345][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.02896934002637863, 'train/grad_norm_pre_clip': 34.860721588134766, 'timing/data': 0.0006328339222818613, 'timing/model': 1.9318870139541104, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 1.23}) +[2026-06-23 16:11:15,153][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.09011618047952652, 'train/grad_norm_pre_clip': 33.80317306518555, 'timing/data': 0.000682962010614574, 'timing/model': 1.9328612990211695, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 1.23}) +[2026-06-23 16:11:18,957][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.11873573064804077, 'train/grad_norm_pre_clip': 37.542354583740234, 'timing/data': 0.0005909259198233485, 'timing/model': 1.9329519579187036, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 1.23}) +[2026-06-23 16:11:22,763][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.13777130842208862, 'train/grad_norm_pre_clip': 26.8188533782959, 'timing/data': 0.0006988350069150329, 'timing/model': 1.9308164529502392, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 1.23}) +[2026-06-23 16:11:26,560][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.03651277348399162, 'train/grad_norm_pre_clip': 18.259258270263672, 'timing/data': 0.0007576530333608389, 'timing/model': 1.922732473933138, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 1.23}) +[2026-06-23 16:11:30,392][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.020063083618879318, 'train/grad_norm_pre_clip': 12.971665382385254, 'timing/data': 0.0005892449989914894, 'timing/model': 1.9483895699959248, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 1.23}) +[2026-06-23 16:11:34,213][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.0533536821603775, 'train/grad_norm_pre_clip': 24.95109748840332, 'timing/data': 0.0006188269471749663, 'timing/model': 1.949040396953933, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 1.23}) +[2026-06-23 16:11:38,041][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.14114214479923248, 'train/grad_norm_pre_clip': 33.76104736328125, 'timing/data': 0.0006052990211173892, 'timing/model': 1.9594493509503081, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 1.23}) +[2026-06-23 16:11:41,992][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.07187744230031967, 'train/grad_norm_pre_clip': 25.824037551879883, 'timing/data': 0.0006035800324752927, 'timing/model': 2.085852512042038, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 1.23}) +[2026-06-23 16:11:45,780][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.14169514179229736, 'train/grad_norm_pre_clip': 35.31849670410156, 'timing/data': 0.0006481260061264038, 'timing/model': 1.9177380069158971, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 1.23}) +[2026-06-23 16:11:49,574][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.01256613153964281, 'train/grad_norm_pre_clip': 19.99968147277832, 'timing/data': 0.0009573199786245823, 'timing/model': 1.9243230499560013, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 1.23}) +[2026-06-23 16:11:53,376][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.03004494123160839, 'train/grad_norm_pre_clip': 17.544160842895508, 'timing/data': 0.0005981649737805128, 'timing/model': 1.929930803948082, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 1.23}) +[2026-06-23 16:11:57,183][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.035797953605651855, 'train/grad_norm_pre_clip': 33.54167175292969, 'timing/data': 0.0006244060350582004, 'timing/model': 1.9363750179763883, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 1.23}) +[2026-06-23 16:12:00,986][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.010792553424835205, 'train/grad_norm_pre_clip': 34.82675552368164, 'timing/data': 0.0006891930243000388, 'timing/model': 1.9221419040113688, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 1.23}) +[2026-06-23 16:12:04,785][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.005298534408211708, 'train/grad_norm_pre_clip': 26.119247436523438, 'timing/data': 0.0006165410159155726, 'timing/model': 1.926648930995725, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 1.23}) +[2026-06-23 16:12:08,580][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.12468466907739639, 'train/grad_norm_pre_clip': 45.78861999511719, 'timing/data': 0.0007099549984559417, 'timing/model': 1.9219306040322408, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 1.23}) +[2026-06-23 16:12:12,379][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.032030314207077026, 'train/grad_norm_pre_clip': 28.988216400146484, 'timing/data': 0.0005960599519312382, 'timing/model': 1.9323276199866086, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 1.23}) +[2026-06-23 16:12:16,283][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.035662926733493805, 'train/grad_norm_pre_clip': 16.453800201416016, 'timing/data': 0.0006194290472194552, 'timing/model': 1.935446696006693, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 1.23}) +[2026-06-23 16:12:20,075][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.17664523422718048, 'train/grad_norm_pre_clip': 29.101594924926758, 'timing/data': 0.0005899500101804733, 'timing/model': 1.9191771920304745, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 1.23}) +[2026-06-23 16:12:23,863][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.020313618704676628, 'train/grad_norm_pre_clip': 17.70836639404297, 'timing/data': 0.0006129669491201639, 'timing/model': 1.9220862109214067, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 1.23}) +[2026-06-23 16:12:27,663][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.1303168088197708, 'train/grad_norm_pre_clip': 56.04781723022461, 'timing/data': 0.0006245230324566364, 'timing/model': 1.933067845995538, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 1.23}) +[2026-06-23 16:12:31,467][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.07725755870342255, 'train/grad_norm_pre_clip': 36.13950729370117, 'timing/data': 0.0006062410539016128, 'timing/model': 1.9371033660136163, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.23}) +[2026-06-23 16:12:35,273][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.049129534512758255, 'train/grad_norm_pre_clip': 24.843849182128906, 'timing/data': 0.000811578007414937, 'timing/model': 1.9329608549596742, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 1.23}) +[2026-06-23 16:12:39,111][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.10874658077955246, 'train/grad_norm_pre_clip': 26.348682403564453, 'timing/data': 0.0007058510091155767, 'timing/model': 1.956309522036463, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 1.23}) +[2026-06-23 16:12:42,902][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.003268939908593893, 'train/grad_norm_pre_clip': 23.3828067779541, 'timing/data': 0.000618249992839992, 'timing/model': 1.9229882740182802, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 1.23}) +[2026-06-23 16:12:46,792][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.022451797500252724, 'train/grad_norm_pre_clip': 23.48993682861328, 'timing/data': 0.0005965599557384849, 'timing/model': 2.021809774916619, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 1.23}) +[2026-06-23 16:12:50,595][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.04851403832435608, 'train/grad_norm_pre_clip': 16.508899688720703, 'timing/data': 0.00057411496527493, 'timing/model': 1.9266022499650717, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 1.24}) +[2026-06-23 16:12:54,384][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.037250544875860214, 'train/grad_norm_pre_clip': 32.91326904296875, 'timing/data': 0.0006032389355823398, 'timing/model': 1.9199254749109969, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 1.24}) +[2026-06-23 16:12:58,182][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.09430459141731262, 'train/grad_norm_pre_clip': 26.174964904785156, 'timing/data': 0.0005885349819436669, 'timing/model': 1.9300315650179982, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 1.24}) +[2026-06-23 16:13:01,975][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.03113553300499916, 'train/grad_norm_pre_clip': 18.83135223388672, 'timing/data': 0.0009980140021070838, 'timing/model': 1.9240350739564747, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 1.24}) +[2026-06-23 16:13:05,769][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.028812415897846222, 'train/grad_norm_pre_clip': 8.863595962524414, 'timing/data': 0.0005905689904466271, 'timing/model': 1.9284454609733075, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 1.24}) +[2026-06-23 16:13:09,578][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.035862281918525696, 'train/grad_norm_pre_clip': 26.92664909362793, 'timing/data': 0.0006239969516173005, 'timing/model': 1.9344345920253545, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 1.24}) +[2026-06-23 16:13:13,401][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.05620932951569557, 'train/grad_norm_pre_clip': 16.614025115966797, 'timing/data': 0.0005907490849494934, 'timing/model': 1.9375902089523152, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 1.24}) +[2026-06-23 16:13:17,226][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.06821370869874954, 'train/grad_norm_pre_clip': 16.6835994720459, 'timing/data': 0.0006077629514038563, 'timing/model': 1.9441546060843393, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 1.24}) +[2026-06-23 16:13:21,165][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3200 +[2026-06-23 16:15:28,754][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3200 +[2026-06-23 16:15:28,756][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3200 +[2026-06-23 16:16:51,785][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3200 +[2026-06-23 16:16:51,785][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.03923977538943291, 'train/grad_norm_pre_clip': 16.384815216064453, 'eval/action_loss/seconds': 127.58721447293647, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03356305882334709, 'eval/latency_2/loss': 0.03356305882334709, 'eval/flappy/loss': 0.03356305882334709, 'eval/flappy/latency_2/loss': 0.03356305882334709, 'eval/flap_precision': 0.9638312123241795, 'eval/flap_recall': 0.9709851551956815, 'eval/flap_f1': 0.9673949579831934, 'eval/flap_rate_model': 0.2073611111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9673949579831933, 'eval/latency_2/flap_precision': 0.9638312123241795, 'eval/latency_2/flap_recall': 0.9709851551956815, 'eval/latency_2/flap_f1': 0.9673949579831934, 'eval/latency_2/flap_rate_model': 0.2073611111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9673949579831933, 'eval/action_classification/data_wait_seconds': 1.189487099647522, 'eval/action_classification/predict_seconds': 81.6626968383789, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0008070759940892458, 'timing/model': 2.0594037709524855, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 1.24}) +[2026-06-23 16:16:55,560][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.04889455437660217, 'train/grad_norm_pre_clip': 33.01093673706055, 'timing/data': 0.0006214090390130877, 'timing/model': 1.9161104609956965, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 1.24}) +[2026-06-23 16:16:59,346][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.11011062562465668, 'train/grad_norm_pre_clip': 31.313966751098633, 'timing/data': 0.0006109679816290736, 'timing/model': 1.921374189085327, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 1.24}) +[2026-06-23 16:17:03,132][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.03970262035727501, 'train/grad_norm_pre_clip': 35.06874465942383, 'timing/data': 0.0005862360121682286, 'timing/model': 1.9274622119264677, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 1.24}) +[2026-06-23 16:17:06,933][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.062216758728027344, 'train/grad_norm_pre_clip': 26.4307804107666, 'timing/data': 0.0006699869409203529, 'timing/model': 1.9413897969061509, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 1.24}) +[2026-06-23 16:17:10,706][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.10415538400411606, 'train/grad_norm_pre_clip': 50.73224639892578, 'timing/data': 0.0005965320160612464, 'timing/model': 1.9155147019773722, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 1.24}) +[2026-06-23 16:17:14,486][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.11326050013303757, 'train/grad_norm_pre_clip': 46.00285339355469, 'timing/data': 0.0007566289277747273, 'timing/model': 1.9166065610479563, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 1.24}) +[2026-06-23 16:17:18,270][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.0682372972369194, 'train/grad_norm_pre_clip': 53.96561813354492, 'timing/data': 0.0007748809875920415, 'timing/model': 1.921223582001403, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 1.24}) +[2026-06-23 16:17:22,061][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.020446069538593292, 'train/grad_norm_pre_clip': 23.505691528320312, 'timing/data': 0.0006381940329447389, 'timing/model': 1.922814100049436, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 1.24}) +[2026-06-23 16:17:25,856][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.029224660247564316, 'train/grad_norm_pre_clip': 11.541155815124512, 'timing/data': 0.000624606036581099, 'timing/model': 1.92213219194673, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 1.24}) +[2026-06-23 16:17:29,641][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.11988166719675064, 'train/grad_norm_pre_clip': 38.25602340698242, 'timing/data': 0.0006017211126163602, 'timing/model': 1.9146965810796246, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 1.24}) +[2026-06-23 16:17:33,431][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.04650811478495598, 'train/grad_norm_pre_clip': 20.33008575439453, 'timing/data': 0.0006131361005827785, 'timing/model': 1.9219588210107759, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 1.24}) +[2026-06-23 16:17:37,229][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.022102104499936104, 'train/grad_norm_pre_clip': 26.70323944091797, 'timing/data': 0.0005994049133732915, 'timing/model': 1.935285450075753, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 1.24}) +[2026-06-23 16:17:41,032][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.05635951831936836, 'train/grad_norm_pre_clip': 40.20779800415039, 'timing/data': 0.0005948849720880389, 'timing/model': 1.9360677369404584, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 1.24}) +[2026-06-23 16:17:44,836][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.08149609714746475, 'train/grad_norm_pre_clip': 33.323326110839844, 'timing/data': 0.0006223960081115365, 'timing/model': 1.931531163980253, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 1.24}) +[2026-06-23 16:17:48,633][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.12708263099193573, 'train/grad_norm_pre_clip': 26.849966049194336, 'timing/data': 0.000810749945230782, 'timing/model': 1.925562446936965, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 1.24}) +[2026-06-23 16:17:52,437][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.058063410222530365, 'train/grad_norm_pre_clip': 25.020071029663086, 'timing/data': 0.0005981059512123466, 'timing/model': 1.9292527480283752, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 1.24}) +[2026-06-23 16:17:56,223][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.021009864285588264, 'train/grad_norm_pre_clip': 31.038257598876953, 'timing/data': 0.0006095130229368806, 'timing/model': 1.9204463980859146, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 1.25}) +[2026-06-23 16:18:00,003][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.027771804481744766, 'train/grad_norm_pre_clip': 22.759033203125, 'timing/data': 0.0008768800180405378, 'timing/model': 1.9164112500147894, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 1.25}) +[2026-06-23 16:18:03,790][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.03063678927719593, 'train/grad_norm_pre_clip': 14.269817352294922, 'timing/data': 0.0009027720661833882, 'timing/model': 1.918087178026326, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 1.25}) +[2026-06-23 16:18:07,638][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.021683236584067345, 'train/grad_norm_pre_clip': 10.512754440307617, 'timing/data': 0.0006740209646522999, 'timing/model': 1.9290234489599243, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 1.25}) +[2026-06-23 16:18:11,451][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.0253653172403574, 'train/grad_norm_pre_clip': 13.898268699645996, 'timing/data': 0.0006459910655394197, 'timing/model': 1.9303697999566793, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 1.25}) +[2026-06-23 16:18:15,258][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.0232662633061409, 'train/grad_norm_pre_clip': 31.703340530395508, 'timing/data': 0.0005998029373586178, 'timing/model': 1.9387243579840288, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 1.25}) +[2026-06-23 16:18:19,056][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.014011492021381855, 'train/grad_norm_pre_clip': 25.619382858276367, 'timing/data': 0.000584615976549685, 'timing/model': 1.9246678750496358, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 1.25}) +[2026-06-23 16:18:22,849][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.1534091979265213, 'train/grad_norm_pre_clip': 48.48185348510742, 'timing/data': 0.0005930929910391569, 'timing/model': 1.9270366959972307, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 1.25}) +[2026-06-23 16:18:26,648][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.019177595153450966, 'train/grad_norm_pre_clip': 16.836318969726562, 'timing/data': 0.000596099067479372, 'timing/model': 1.9285025290446356, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 1.25}) +[2026-06-23 16:18:30,453][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.009383093565702438, 'train/grad_norm_pre_clip': 18.833803176879883, 'timing/data': 0.0005769300041720271, 'timing/model': 1.9293581339297816, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 1.25}) +[2026-06-23 16:18:34,251][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.042324744164943695, 'train/grad_norm_pre_clip': 22.993770599365234, 'timing/data': 0.0006971340626478195, 'timing/model': 1.9182273119222373, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 1.25}) +[2026-06-23 16:18:38,053][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.10104839503765106, 'train/grad_norm_pre_clip': 35.72351837158203, 'timing/data': 0.0006086479406803846, 'timing/model': 1.930388386012055, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 1.25}) +[2026-06-23 16:18:41,872][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.048683732748031616, 'train/grad_norm_pre_clip': 29.51932144165039, 'timing/data': 0.000759148970246315, 'timing/model': 1.9213354649255052, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 1.25}) +[2026-06-23 16:18:45,671][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.059155598282814026, 'train/grad_norm_pre_clip': 28.20382308959961, 'timing/data': 0.000649923924356699, 'timing/model': 1.9309870690340176, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 1.25}) +[2026-06-23 16:18:49,473][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.0132970055565238, 'train/grad_norm_pre_clip': 15.303030967712402, 'timing/data': 0.0005885090213268995, 'timing/model': 1.929686106974259, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 1.25}) +[2026-06-23 16:18:53,268][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.09842199087142944, 'train/grad_norm_pre_clip': 53.45600128173828, 'timing/data': 0.0005940300179645419, 'timing/model': 1.9196118830004707, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 1.25}) +[2026-06-23 16:18:57,057][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.19520927965641022, 'train/grad_norm_pre_clip': 37.261085510253906, 'timing/data': 0.000604985048994422, 'timing/model': 1.9253867170773447, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 1.25}) +[2026-06-23 16:19:00,866][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.10053950548171997, 'train/grad_norm_pre_clip': 38.80099105834961, 'timing/data': 0.0006002410082146525, 'timing/model': 1.9311325860908255, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 1.25}) +[2026-06-23 16:19:04,672][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.06923303008079529, 'train/grad_norm_pre_clip': 45.60393524169922, 'timing/data': 0.0006085869390517473, 'timing/model': 1.935886927996762, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 1.25}) +[2026-06-23 16:19:08,466][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.09387189149856567, 'train/grad_norm_pre_clip': 42.39029312133789, 'timing/data': 0.0007819869788363576, 'timing/model': 1.9313142989994958, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.25}) +[2026-06-23 16:19:12,270][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.07048989832401276, 'train/grad_norm_pre_clip': 61.23057556152344, 'timing/data': 0.000595193007029593, 'timing/model': 1.9276282120263204, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 1.25}) +[2026-06-23 16:19:16,063][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.026279253885149956, 'train/grad_norm_pre_clip': 41.16062545776367, 'timing/data': 0.0005854459013789892, 'timing/model': 1.9291157249826938, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 1.25}) +[2026-06-23 16:19:19,875][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.0493880920112133, 'train/grad_norm_pre_clip': 21.056610107421875, 'timing/data': 0.0006315769860520959, 'timing/model': 1.9382071290165186, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 1.25}) +[2026-06-23 16:19:23,665][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.09651366621255875, 'train/grad_norm_pre_clip': 39.975467681884766, 'timing/data': 0.0006747549632564187, 'timing/model': 1.9199496469227597, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 1.25}) +[2026-06-23 16:19:27,444][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.045032743364572525, 'train/grad_norm_pre_clip': 42.89362716674805, 'timing/data': 0.0006043330067768693, 'timing/model': 1.9183871279237792, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 1.25}) +[2026-06-23 16:19:31,244][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.02088775299489498, 'train/grad_norm_pre_clip': 31.873912811279297, 'timing/data': 0.0007507470436394215, 'timing/model': 1.9258469810010865, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 1.25}) +[2026-06-23 16:19:35,041][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.03230399638414383, 'train/grad_norm_pre_clip': 116.72105407714844, 'timing/data': 0.0007293489761650562, 'timing/model': 1.919100511004217, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 1.26}) +[2026-06-23 16:19:38,849][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.08468181639909744, 'train/grad_norm_pre_clip': 35.62571334838867, 'timing/data': 0.0007927519036456943, 'timing/model': 1.9360052540432662, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 1.26}) +[2026-06-23 16:19:42,646][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.2024780660867691, 'train/grad_norm_pre_clip': 45.666114807128906, 'timing/data': 0.0005963239818811417, 'timing/model': 1.9303629759233445, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 1.26}) +[2026-06-23 16:19:46,459][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.016190731897950172, 'train/grad_norm_pre_clip': 7.015090465545654, 'timing/data': 0.0005884350975975394, 'timing/model': 1.9289426649920642, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 1.26}) +[2026-06-23 16:19:50,467][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.018718399107456207, 'train/grad_norm_pre_clip': 28.054859161376953, 'timing/data': 0.0005846179556101561, 'timing/model': 2.1335555789992213, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 1.26}) +[2026-06-23 16:19:54,268][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.05359449237585068, 'train/grad_norm_pre_clip': 29.391786575317383, 'timing/data': 0.0005763879744336009, 'timing/model': 1.939737058011815, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 1.26}) +[2026-06-23 16:19:58,072][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.010716482065618038, 'train/grad_norm_pre_clip': 60.7415885925293, 'timing/data': 0.0009355689398944378, 'timing/model': 1.9253761121071875, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 1.26}) +[2026-06-23 16:26:26,363][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3250 +[2026-06-23 16:26:26,363][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.03627549856901169, 'train/grad_norm_pre_clip': 18.428163528442383, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 236.180000000009, 'rl_games_eval/mid_train/mean_length': 1922.0, 'rl_games_eval/mid_train/std_reward': 183.11908038215006, 'rl_games_eval/mid_train/std_length': 1474.810496301135, 'rl_games_eval/mid_train/macro_mean_reward': 236.180000000009, 'rl_games_eval/mid_train/macro_mean_length': 1922.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 236.180000000009, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1922.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 183.11908038215006, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1474.810496301135, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 236.180000000009, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006031010998412967, 'timing/model': 1.9183813219424337, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 1.26}) +[2026-06-23 16:26:30,073][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.05955043435096741, 'train/grad_norm_pre_clip': 23.087039947509766, 'timing/data': 0.0005992400692775846, 'timing/model': 1.8892475799657404, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 1.26}) +[2026-06-23 16:26:33,799][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.016138235107064247, 'train/grad_norm_pre_clip': 8.111163139343262, 'timing/data': 0.0008273969870060682, 'timing/model': 1.8867219289531931, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 1.26}) +[2026-06-23 16:26:37,531][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.008266395889222622, 'train/grad_norm_pre_clip': 18.7429141998291, 'timing/data': 0.0005837590433657169, 'timing/model': 1.8922974430024624, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 1.26}) +[2026-06-23 16:26:41,277][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.06158207729458809, 'train/grad_norm_pre_clip': 14.898616790771484, 'timing/data': 0.0005944730946794152, 'timing/model': 1.9099857159890234, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 1.26}) +[2026-06-23 16:26:45,194][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.15727414190769196, 'train/grad_norm_pre_clip': 40.019779205322266, 'timing/data': 0.0005858190124854445, 'timing/model': 1.897934840992093, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 1.26}) +[2026-06-23 16:26:48,970][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.012535934336483479, 'train/grad_norm_pre_clip': 13.63438606262207, 'timing/data': 0.000613320036791265, 'timing/model': 1.9235521820373833, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 1.26}) +[2026-06-23 16:26:52,724][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.17636838555335999, 'train/grad_norm_pre_clip': 80.34217834472656, 'timing/data': 0.0005927459569647908, 'timing/model': 1.9087431030347943, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 1.26}) +[2026-06-23 16:26:56,478][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.048275429755449295, 'train/grad_norm_pre_clip': 48.72529983520508, 'timing/data': 0.0006001269211992621, 'timing/model': 1.894601930049248, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 1.26}) +[2026-06-23 16:27:00,242][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.07741034775972366, 'train/grad_norm_pre_clip': 18.698909759521484, 'timing/data': 0.0007191429613158107, 'timing/model': 1.90925647795666, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 1.26}) +[2026-06-23 16:27:04,008][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.04943779483437538, 'train/grad_norm_pre_clip': 29.468441009521484, 'timing/data': 0.0006085399072617292, 'timing/model': 1.9145755710778758, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 1.26}) +[2026-06-23 16:27:07,780][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.11219082027673721, 'train/grad_norm_pre_clip': 46.92625427246094, 'timing/data': 0.0005911770276725292, 'timing/model': 1.9140994519693777, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 1.26}) +[2026-06-23 16:27:11,555][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.07465120404958725, 'train/grad_norm_pre_clip': 39.959938049316406, 'timing/data': 0.0006195419700816274, 'timing/model': 1.9149322470184416, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 1.26}) +[2026-06-23 16:27:15,490][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.008021797984838486, 'train/grad_norm_pre_clip': 41.871543884277344, 'timing/data': 0.0005920479306951165, 'timing/model': 2.0789550240151584, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 1.26}) +[2026-06-23 16:27:19,262][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.04672030359506607, 'train/grad_norm_pre_clip': 31.998065948486328, 'timing/data': 0.000603226013481617, 'timing/model': 1.9237855039536953, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 1.26}) +[2026-06-23 16:27:23,027][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.08241938799619675, 'train/grad_norm_pre_clip': 29.054859161376953, 'timing/data': 0.0006151780253276229, 'timing/model': 1.90660237101838, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 1.26}) +[2026-06-23 16:27:26,797][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.07476375252008438, 'train/grad_norm_pre_clip': 30.1010799407959, 'timing/data': 0.000841505010612309, 'timing/model': 1.9139094260754064, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 1.26}) +[2026-06-23 16:27:30,575][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.09912538528442383, 'train/grad_norm_pre_clip': 30.237550735473633, 'timing/data': 0.0006007819902151823, 'timing/model': 1.9188966549700126, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 1.26}) +[2026-06-23 16:27:34,353][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.2753983438014984, 'train/grad_norm_pre_clip': 65.23371887207031, 'timing/data': 0.0006006970070302486, 'timing/model': 1.9169248349498957, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 1.26}) +[2026-06-23 16:27:38,161][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.07363089174032211, 'train/grad_norm_pre_clip': 32.465431213378906, 'timing/data': 0.0005962520372122526, 'timing/model': 1.9396813369821757, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 1.27}) +[2026-06-23 16:27:41,934][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.03002162277698517, 'train/grad_norm_pre_clip': 25.34599494934082, 'timing/data': 0.0006594089791178703, 'timing/model': 1.912959381006658, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 1.27}) +[2026-06-23 16:27:45,701][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.05746394768357277, 'train/grad_norm_pre_clip': 41.6969108581543, 'timing/data': 0.0006319729145616293, 'timing/model': 1.9113383420044556, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 1.27}) +[2026-06-23 16:27:49,492][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.11213364452123642, 'train/grad_norm_pre_clip': 37.971378326416016, 'timing/data': 0.0007295019458979368, 'timing/model': 1.9311765540624037, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 1.27}) +[2026-06-23 16:27:53,280][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.019150378182530403, 'train/grad_norm_pre_clip': 19.019643783569336, 'timing/data': 0.000642782892100513, 'timing/model': 1.9230132289230824, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 1.27}) +[2026-06-23 16:27:57,052][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.021216075867414474, 'train/grad_norm_pre_clip': 47.39353561401367, 'timing/data': 0.0006021489389240742, 'timing/model': 1.91368436999619, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 1.27}) +[2026-06-23 16:28:00,834][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.05990089103579521, 'train/grad_norm_pre_clip': 46.92988586425781, 'timing/data': 0.0006486549973487854, 'timing/model': 1.9215078840497881, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 1.27}) +[2026-06-23 16:28:04,615][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.08669595420360565, 'train/grad_norm_pre_clip': 27.70869255065918, 'timing/data': 0.0005736879538744688, 'timing/model': 1.9200991400284693, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 1.27}) +[2026-06-23 16:28:08,387][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.12851962447166443, 'train/grad_norm_pre_clip': 28.41979217529297, 'timing/data': 0.0005709169199690223, 'timing/model': 1.9127411139197648, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 1.27}) +[2026-06-23 16:28:12,176][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.005356078501790762, 'train/grad_norm_pre_clip': 19.337114334106445, 'timing/data': 0.0006083889165893197, 'timing/model': 1.9239314160076901, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 1.27}) +[2026-06-23 16:28:15,947][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.0273640975356102, 'train/grad_norm_pre_clip': 23.45391082763672, 'timing/data': 0.0008317740866914392, 'timing/model': 1.912664448027499, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 1.27}) +[2026-06-23 16:28:19,744][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.047852128744125366, 'train/grad_norm_pre_clip': 22.97628402709961, 'timing/data': 0.0005858250660821795, 'timing/model': 1.9221551450900733, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 1.27}) +[2026-06-23 16:28:23,542][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.05017156898975372, 'train/grad_norm_pre_clip': 20.76250648498535, 'timing/data': 0.0005786019610241055, 'timing/model': 1.9262501730117947, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 1.27}) +[2026-06-23 16:28:27,344][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.017346641048789024, 'train/grad_norm_pre_clip': 16.842924118041992, 'timing/data': 0.000587829970754683, 'timing/model': 1.927671167999506, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 1.27}) +[2026-06-23 16:28:31,148][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.04796218499541283, 'train/grad_norm_pre_clip': 18.64480972290039, 'timing/data': 0.0005930630723014474, 'timing/model': 1.9266112520126626, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 1.27}) +[2026-06-23 16:28:34,928][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.058130644261837006, 'train/grad_norm_pre_clip': 41.83747100830078, 'timing/data': 0.0005980260903015733, 'timing/model': 1.9188046799972653, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 1.27}) +[2026-06-23 16:28:38,731][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.01393839530646801, 'train/grad_norm_pre_clip': 17.830785751342773, 'timing/data': 0.0005865979474037886, 'timing/model': 1.9272537019569427, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.27}) +[2026-06-23 16:28:42,521][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.028131507337093353, 'train/grad_norm_pre_clip': 18.40672492980957, 'timing/data': 0.0005917239468544722, 'timing/model': 1.925377527019009, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 1.27}) +[2026-06-23 16:28:46,328][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.08585904538631439, 'train/grad_norm_pre_clip': 28.545345306396484, 'timing/data': 0.0007312579546123743, 'timing/model': 1.9289495699340478, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 1.27}) +[2026-06-23 16:28:50,116][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.07754569500684738, 'train/grad_norm_pre_clip': 27.781295776367188, 'timing/data': 0.0006200220668688416, 'timing/model': 1.9219170800643042, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 1.27}) +[2026-06-23 16:28:54,103][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.03972741216421127, 'train/grad_norm_pre_clip': 21.901100158691406, 'timing/data': 0.0006106450455263257, 'timing/model': 2.11804251302965, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 1.27}) +[2026-06-23 16:28:57,906][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.06332254409790039, 'train/grad_norm_pre_clip': 20.52316665649414, 'timing/data': 0.000928006018511951, 'timing/model': 1.9336873339489102, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 1.27}) +[2026-06-23 16:29:01,705][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.007248170208185911, 'train/grad_norm_pre_clip': 13.574165344238281, 'timing/data': 0.0007824120111763477, 'timing/model': 1.9188204419333488, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 1.27}) +[2026-06-23 16:29:05,505][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.029816782101988792, 'train/grad_norm_pre_clip': 12.712047576904297, 'timing/data': 0.0005862080724909902, 'timing/model': 1.9270121750887483, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 1.27}) +[2026-06-23 16:29:09,302][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.06795831769704819, 'train/grad_norm_pre_clip': 30.148468017578125, 'timing/data': 0.0006209020502865314, 'timing/model': 1.9275985939893872, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 1.27}) +[2026-06-23 16:29:13,097][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.036639776080846786, 'train/grad_norm_pre_clip': 42.12932586669922, 'timing/data': 0.000588053953833878, 'timing/model': 1.9239426690619439, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 1.27}) +[2026-06-23 16:29:16,968][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.09227646887302399, 'train/grad_norm_pre_clip': 58.8245964050293, 'timing/data': 0.0005848410073667765, 'timing/model': 1.9911387090105563, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 1.28}) +[2026-06-23 16:29:20,780][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.03544766455888748, 'train/grad_norm_pre_clip': 22.684375762939453, 'timing/data': 0.0005963969742879272, 'timing/model': 1.9376481720246375, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 1.28}) +[2026-06-23 16:29:24,559][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.05552177503705025, 'train/grad_norm_pre_clip': 24.68878936767578, 'timing/data': 0.0005935259396210313, 'timing/model': 1.917364046908915, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 1.28}) +[2026-06-23 16:29:28,343][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.06540891528129578, 'train/grad_norm_pre_clip': 39.99301528930664, 'timing/data': 0.0006533629493787885, 'timing/model': 1.9164127389667556, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 1.28}) +[2026-06-23 16:29:32,135][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.022624386474490166, 'train/grad_norm_pre_clip': 43.795902252197266, 'timing/data': 0.0006560009205713868, 'timing/model': 1.9238513869931921, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 1.28}) +[2026-06-23 16:29:35,939][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3300 +[2026-06-23 16:31:44,599][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3300 +[2026-06-23 16:31:44,601][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3300 +[2026-06-23 16:33:06,659][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3300 +[2026-06-23 16:33:06,660][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.05252159386873245, 'train/grad_norm_pre_clip': 35.21474075317383, 'eval/action_loss/seconds': 128.65755020093638, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.032092321664094925, 'eval/latency_2/loss': 0.032092321664094925, 'eval/flappy/loss': 0.032092321664094925, 'eval/flappy/latency_2/loss': 0.032092321664094925, 'eval/flap_precision': 0.9722033898305085, 'eval/flap_recall': 0.9676113360323887, 'eval/flap_f1': 0.9699019276293541, 'eval/flap_rate_model': 0.2048611111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.969901927629354, 'eval/latency_2/flap_precision': 0.9722033898305085, 'eval/latency_2/flap_recall': 0.9676113360323887, 'eval/latency_2/flap_f1': 0.9699019276293541, 'eval/latency_2/flap_rate_model': 0.2048611111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.969901927629354, 'eval/action_classification/data_wait_seconds': 1.4538631439208984, 'eval/action_classification/predict_seconds': 80.44019317626953, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.000998770003207028, 'timing/model': 1.9207382169552147, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 1.28}) +[2026-06-23 16:33:10,439][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.08522699773311615, 'train/grad_norm_pre_clip': 29.477144241333008, 'timing/data': 0.0006315719801932573, 'timing/model': 1.9166010590270162, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 1.28}) +[2026-06-23 16:33:14,225][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.08883678913116455, 'train/grad_norm_pre_clip': 53.310150146484375, 'timing/data': 0.0006515620043501258, 'timing/model': 1.9139490430243313, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 1.28}) +[2026-06-23 16:33:18,015][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.05746306851506233, 'train/grad_norm_pre_clip': 18.658754348754883, 'timing/data': 0.0008037009974941611, 'timing/model': 1.9244412209372967, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 1.28}) +[2026-06-23 16:33:21,801][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.07777340710163116, 'train/grad_norm_pre_clip': 27.941871643066406, 'timing/data': 0.0006602939683943987, 'timing/model': 1.923024753923528, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 1.28}) +[2026-06-23 16:33:25,576][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.02772112935781479, 'train/grad_norm_pre_clip': 8.29754638671875, 'timing/data': 0.0006514349952340126, 'timing/model': 1.918234828975983, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 1.28}) +[2026-06-23 16:33:29,366][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.04826189950108528, 'train/grad_norm_pre_clip': 39.23372268676758, 'timing/data': 0.0006534459535032511, 'timing/model': 1.9283946340437979, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 1.28}) +[2026-06-23 16:33:33,153][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.05565803870558739, 'train/grad_norm_pre_clip': 42.77341842651367, 'timing/data': 0.0006008399650454521, 'timing/model': 1.9208555329823866, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 1.28}) +[2026-06-23 16:33:36,924][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.009676399640738964, 'train/grad_norm_pre_clip': 13.204411506652832, 'timing/data': 0.0006363909924402833, 'timing/model': 1.9097485030069947, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 1.28}) +[2026-06-23 16:33:40,723][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.07588272541761398, 'train/grad_norm_pre_clip': 29.40238380432129, 'timing/data': 0.0006100679747760296, 'timing/model': 1.9195726310135797, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 1.28}) +[2026-06-23 16:33:44,517][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.1624734103679657, 'train/grad_norm_pre_clip': 29.936359405517578, 'timing/data': 0.0006268980214372277, 'timing/model': 1.927632977021858, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 1.28}) +[2026-06-23 16:33:48,487][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.04671519994735718, 'train/grad_norm_pre_clip': 62.60442352294922, 'timing/data': 0.000634283060207963, 'timing/model': 2.1009071780135855, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 1.28}) +[2026-06-23 16:33:52,303][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.09481295943260193, 'train/grad_norm_pre_clip': 27.05782699584961, 'timing/data': 0.0008859799709171057, 'timing/model': 1.9247287089237943, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 1.28}) +[2026-06-23 16:33:56,102][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.06626559793949127, 'train/grad_norm_pre_clip': 54.394622802734375, 'timing/data': 0.0006394690135493875, 'timing/model': 1.9208765759831294, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 1.28}) +[2026-06-23 16:33:59,892][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.0046963682398200035, 'train/grad_norm_pre_clip': 4.955916404724121, 'timing/data': 0.0006374949589371681, 'timing/model': 1.91813743498642, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 1.28}) +[2026-06-23 16:34:03,697][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.044947825372219086, 'train/grad_norm_pre_clip': 40.10026168823242, 'timing/data': 0.0005866680294275284, 'timing/model': 1.9364037160994485, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 1.28}) +[2026-06-23 16:34:07,492][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.09208746999502182, 'train/grad_norm_pre_clip': 33.107452392578125, 'timing/data': 0.000580127933062613, 'timing/model': 1.9185752840712667, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 1.28}) +[2026-06-23 16:34:11,287][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.019227439537644386, 'train/grad_norm_pre_clip': 37.39009475708008, 'timing/data': 0.0005983939627185464, 'timing/model': 1.9210303570143878, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 1.28}) +[2026-06-23 16:34:15,094][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.07962211966514587, 'train/grad_norm_pre_clip': 70.97639465332031, 'timing/data': 0.0005848050350323319, 'timing/model': 1.9362957720877603, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 1.28}) +[2026-06-23 16:34:18,892][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.06531558930873871, 'train/grad_norm_pre_clip': 31.839508056640625, 'timing/data': 0.0008738149190321565, 'timing/model': 1.9202847910346463, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 1.28}) +[2026-06-23 16:34:22,883][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.04214092716574669, 'train/grad_norm_pre_clip': 39.9799690246582, 'timing/data': 0.0006023349706083536, 'timing/model': 1.9269364950014278, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 1.28}) +[2026-06-23 16:34:26,676][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.08958865702152252, 'train/grad_norm_pre_clip': 51.87691879272461, 'timing/data': 0.000600020051933825, 'timing/model': 1.922767144977115, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 1.29}) +[2026-06-23 16:34:30,462][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.01353490725159645, 'train/grad_norm_pre_clip': 21.591201782226562, 'timing/data': 0.0006033569807186723, 'timing/model': 1.9211256389971823, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 1.29}) +[2026-06-23 16:34:34,265][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.018448425456881523, 'train/grad_norm_pre_clip': 13.071045875549316, 'timing/data': 0.0006065090419724584, 'timing/model': 1.9299568169517443, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 1.29}) +[2026-06-23 16:34:38,066][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.07729778438806534, 'train/grad_norm_pre_clip': 61.879005432128906, 'timing/data': 0.0005918670212849975, 'timing/model': 1.9264391990145668, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 1.29}) +[2026-06-23 16:34:41,866][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.056866664439439774, 'train/grad_norm_pre_clip': 22.179487228393555, 'timing/data': 0.0005866839783266187, 'timing/model': 1.9329244339605793, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 1.29}) +[2026-06-23 16:34:45,660][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.01837926171720028, 'train/grad_norm_pre_clip': 12.66701602935791, 'timing/data': 0.0005943169817328453, 'timing/model': 1.9253930610138923, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 1.29}) +[2026-06-23 16:34:49,465][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.07169369608163834, 'train/grad_norm_pre_clip': 43.963714599609375, 'timing/data': 0.0005853000329807401, 'timing/model': 1.9311792480293661, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 1.29}) +[2026-06-23 16:34:53,471][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.1354137361049652, 'train/grad_norm_pre_clip': 58.75358581542969, 'timing/data': 0.0008009730372577906, 'timing/model': 2.127405901090242, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 1.29}) +[2026-06-23 16:34:57,255][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.0910821482539177, 'train/grad_norm_pre_clip': 38.73643112182617, 'timing/data': 0.000862385961227119, 'timing/model': 1.916082554962486, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 1.29}) +[2026-06-23 16:35:01,057][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.03936045244336128, 'train/grad_norm_pre_clip': 33.71383285522461, 'timing/data': 0.0006009319331496954, 'timing/model': 1.9275698800338432, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 1.29}) +[2026-06-23 16:35:04,856][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.0234987810254097, 'train/grad_norm_pre_clip': 10.286184310913086, 'timing/data': 0.0006332410266622901, 'timing/model': 1.9188548070378602, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 1.29}) +[2026-06-23 16:35:08,646][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.07042859494686127, 'train/grad_norm_pre_clip': 67.16984558105469, 'timing/data': 0.0006435989635065198, 'timing/model': 1.925138734979555, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 1.29}) +[2026-06-23 16:35:12,434][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.12650874257087708, 'train/grad_norm_pre_clip': 56.15195083618164, 'timing/data': 0.0006341489497572184, 'timing/model': 1.920780391083099, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 1.29}) +[2026-06-23 16:35:16,237][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.03701665252447128, 'train/grad_norm_pre_clip': 35.80131530761719, 'timing/data': 0.0006014509126543999, 'timing/model': 1.9230827800929546, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.29}) +[2026-06-23 16:35:20,061][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.10463209450244904, 'train/grad_norm_pre_clip': 48.85948181152344, 'timing/data': 0.0005978019908070564, 'timing/model': 1.9406384859466925, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 1.29}) +[2026-06-23 16:35:23,881][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.05550511181354523, 'train/grad_norm_pre_clip': 26.307966232299805, 'timing/data': 0.0006950560491532087, 'timing/model': 1.9327486109687015, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 1.29}) +[2026-06-23 16:35:27,686][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.04463062062859535, 'train/grad_norm_pre_clip': 14.457015037536621, 'timing/data': 0.0006735939532518387, 'timing/model': 1.926411138032563, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 1.29}) +[2026-06-23 16:35:31,485][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.09281649440526962, 'train/grad_norm_pre_clip': 38.00605010986328, 'timing/data': 0.0005882079713046551, 'timing/model': 1.9264978490537032, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 1.29}) +[2026-06-23 16:35:35,280][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.08160379528999329, 'train/grad_norm_pre_clip': 37.571651458740234, 'timing/data': 0.00068344094324857, 'timing/model': 1.9196320010814816, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 1.29}) +[2026-06-23 16:35:39,071][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.0671161636710167, 'train/grad_norm_pre_clip': 50.09090805053711, 'timing/data': 0.0006161830388009548, 'timing/model': 1.9218244180083275, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 1.29}) +[2026-06-23 16:35:42,856][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.024062462151050568, 'train/grad_norm_pre_clip': 12.181143760681152, 'timing/data': 0.0006928980583325028, 'timing/model': 1.92117457697168, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 1.29}) +[2026-06-23 16:35:46,658][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.06794913113117218, 'train/grad_norm_pre_clip': 30.386310577392578, 'timing/data': 0.0005944069707766175, 'timing/model': 1.927592534921132, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 1.29}) +[2026-06-23 16:35:50,468][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.10974600911140442, 'train/grad_norm_pre_clip': 37.4600715637207, 'timing/data': 0.0006110789254307747, 'timing/model': 1.935517902020365, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 1.29}) +[2026-06-23 16:35:54,262][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.03380974382162094, 'train/grad_norm_pre_clip': 27.306640625, 'timing/data': 0.0006131819682195783, 'timing/model': 1.9204235801007599, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 1.29}) +[2026-06-23 16:35:58,061][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.07737012207508087, 'train/grad_norm_pre_clip': 57.06638717651367, 'timing/data': 0.0005912400083616376, 'timing/model': 1.9283939549932256, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 1.29}) +[2026-06-23 16:36:01,857][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.0752536728978157, 'train/grad_norm_pre_clip': 28.54350471496582, 'timing/data': 0.0005992900114506483, 'timing/model': 1.9316219659522176, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 1.29}) +[2026-06-23 16:36:05,664][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.03455037623643875, 'train/grad_norm_pre_clip': 18.103727340698242, 'timing/data': 0.0008624059846624732, 'timing/model': 1.9275653540389612, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 1.3}) +[2026-06-23 16:36:09,476][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.02607589401304722, 'train/grad_norm_pre_clip': 18.915679931640625, 'timing/data': 0.0006601709173992276, 'timing/model': 1.9320952149573714, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 1.3}) +[2026-06-23 16:36:13,281][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.11113250255584717, 'train/grad_norm_pre_clip': 49.229793548583984, 'timing/data': 0.0005866240244358778, 'timing/model': 1.934393881005235, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 1.3}) +[2026-06-23 16:36:17,092][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.0051038144156336784, 'train/grad_norm_pre_clip': 40.17842102050781, 'timing/data': 0.0005905450088903308, 'timing/model': 1.9281113240867853, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 1.3}) +[2026-06-23 16:36:20,896][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.01450154185295105, 'train/grad_norm_pre_clip': 13.776742935180664, 'timing/data': 0.0006041749147698283, 'timing/model': 1.9318850809941068, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 1.3}) +[2026-06-23 16:36:24,707][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.049626659601926804, 'train/grad_norm_pre_clip': 58.141197204589844, 'timing/data': 0.000607748981565237, 'timing/model': 1.9345653899945319, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 1.3}) +[2026-06-23 16:36:28,500][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.03353279456496239, 'train/grad_norm_pre_clip': 22.3501033782959, 'timing/data': 0.0006091860122978687, 'timing/model': 1.9301394369686022, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 1.3}) +[2026-06-23 16:36:32,284][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.05850421264767647, 'train/grad_norm_pre_clip': 23.88100242614746, 'timing/data': 0.0008387590060010552, 'timing/model': 1.917053794953972, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 1.3}) +[2026-06-23 16:36:36,097][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.10259249806404114, 'train/grad_norm_pre_clip': 34.17707061767578, 'timing/data': 0.0005984269082546234, 'timing/model': 1.9349642890738323, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 1.3}) +[2026-06-23 16:36:39,899][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.10989345610141754, 'train/grad_norm_pre_clip': 24.795835494995117, 'timing/data': 0.0006471079541370273, 'timing/model': 1.9396339560626075, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 1.3}) +[2026-06-23 16:36:43,698][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.07020880281925201, 'train/grad_norm_pre_clip': 65.12430572509766, 'timing/data': 0.0005980529822409153, 'timing/model': 1.9265740329865366, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 1.3}) +[2026-06-23 16:36:47,495][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.0179182980209589, 'train/grad_norm_pre_clip': 17.558576583862305, 'timing/data': 0.000607192050665617, 'timing/model': 1.9248210350051522, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 1.3}) +[2026-06-23 16:36:51,294][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.0863785594701767, 'train/grad_norm_pre_clip': 31.489086151123047, 'timing/data': 0.0005933709908276796, 'timing/model': 1.930343727930449, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 1.3}) +[2026-06-23 16:36:55,086][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.014328447170555592, 'train/grad_norm_pre_clip': 20.5007266998291, 'timing/data': 0.0006048290524631739, 'timing/model': 1.927152659976855, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 1.3}) +[2026-06-23 16:36:58,882][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.04604865983128548, 'train/grad_norm_pre_clip': 18.614215850830078, 'timing/data': 0.0005836860509589314, 'timing/model': 1.928795253043063, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 1.3}) +[2026-06-23 16:37:02,696][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.08155810832977295, 'train/grad_norm_pre_clip': 50.120880126953125, 'timing/data': 0.000815250095911324, 'timing/model': 1.9359443050343543, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 1.3}) +[2026-06-23 16:37:06,704][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.022280745208263397, 'train/grad_norm_pre_clip': 24.132347106933594, 'timing/data': 0.0006018489366397262, 'timing/model': 1.9391564410179853, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 1.3}) +[2026-06-23 16:37:10,495][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.049386367201805115, 'train/grad_norm_pre_clip': 35.42082595825195, 'timing/data': 0.000902898027561605, 'timing/model': 1.9236103930743411, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 1.3}) +[2026-06-23 16:37:14,303][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.024318335577845573, 'train/grad_norm_pre_clip': 50.66066360473633, 'timing/data': 0.0006549430545419455, 'timing/model': 1.9269508479628712, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 1.3}) +[2026-06-23 16:37:18,104][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.005631159991025925, 'train/grad_norm_pre_clip': 34.4530029296875, 'timing/data': 0.0008355140453204513, 'timing/model': 1.931309584993869, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 1.3}) +[2026-06-23 16:37:21,914][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.008591141551733017, 'train/grad_norm_pre_clip': 25.56587028503418, 'timing/data': 0.0005837910575792193, 'timing/model': 1.9277938299346715, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 1.3}) +[2026-06-23 16:37:25,711][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.016404928639531136, 'train/grad_norm_pre_clip': 23.922035217285156, 'timing/data': 0.0007825579959899187, 'timing/model': 1.9292522210162133, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 1.3}) +[2026-06-23 16:37:29,512][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.04603322967886925, 'train/grad_norm_pre_clip': 60.820064544677734, 'timing/data': 0.0006052539683878422, 'timing/model': 1.933332970016636, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 1.3}) +[2026-06-23 16:37:33,326][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.030357368290424347, 'train/grad_norm_pre_clip': 26.039104461669922, 'timing/data': 0.0006001000292599201, 'timing/model': 1.939115516957827, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 1.3}) +[2026-06-23 16:37:37,125][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.03842199966311455, 'train/grad_norm_pre_clip': 17.58797836303711, 'timing/data': 0.0005985180614516139, 'timing/model': 1.932728967978619, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 1.3}) +[2026-06-23 16:37:40,953][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.07364032417535782, 'train/grad_norm_pre_clip': 41.21684265136719, 'timing/data': 0.000596201978623867, 'timing/model': 1.94315052498132, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 1.31}) +[2026-06-23 16:37:44,764][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.04056698456406593, 'train/grad_norm_pre_clip': 47.53578186035156, 'timing/data': 0.000587044982239604, 'timing/model': 1.9340054349740967, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 1.31}) +[2026-06-23 16:37:48,550][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.08976195007562637, 'train/grad_norm_pre_clip': 68.41963958740234, 'timing/data': 0.0006192190339788795, 'timing/model': 1.9217356779845431, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 1.31}) +[2026-06-23 16:37:52,359][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.06403235346078873, 'train/grad_norm_pre_clip': 42.03425216674805, 'timing/data': 0.0005967780016362667, 'timing/model': 1.9289335140492767, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 1.31}) +[2026-06-23 16:37:56,163][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.009179776534438133, 'train/grad_norm_pre_clip': 5.580338954925537, 'timing/data': 0.0006308460142463446, 'timing/model': 1.9317643890390173, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 1.31}) +[2026-06-23 16:37:59,961][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.033289503306150436, 'train/grad_norm_pre_clip': 22.816362380981445, 'timing/data': 0.0006078979931771755, 'timing/model': 1.9235310680232942, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 1.31}) +[2026-06-23 16:38:03,769][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.025417819619178772, 'train/grad_norm_pre_clip': 13.601696968078613, 'timing/data': 0.0005866739666089416, 'timing/model': 1.929148938972503, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 1.31}) +[2026-06-23 16:38:07,579][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.1610114425420761, 'train/grad_norm_pre_clip': 35.3721923828125, 'timing/data': 0.0005811960436403751, 'timing/model': 1.9329129030229524, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 1.31}) +[2026-06-23 16:38:11,370][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.0415535643696785, 'train/grad_norm_pre_clip': 29.83089256286621, 'timing/data': 0.0008023639675229788, 'timing/model': 1.927595681976527, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 1.31}) +[2026-06-23 16:38:15,174][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.010845019482076168, 'train/grad_norm_pre_clip': 25.073476791381836, 'timing/data': 0.0007924720412120223, 'timing/model': 1.9371928399195895, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 1.31}) +[2026-06-23 16:38:18,981][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.061725519597530365, 'train/grad_norm_pre_clip': 24.797056198120117, 'timing/data': 0.0006182960933074355, 'timing/model': 1.9251714000711218, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 1.31}) +[2026-06-23 16:38:22,774][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.08108250796794891, 'train/grad_norm_pre_clip': 31.899944305419922, 'timing/data': 0.0005873390473425388, 'timing/model': 1.9232770300004631, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.31}) +[2026-06-23 16:38:26,597][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.03628719598054886, 'train/grad_norm_pre_clip': 30.864192962646484, 'timing/data': 0.0007062569493427873, 'timing/model': 1.9431128510041162, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 1.31}) +[2026-06-23 16:38:30,400][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.0346452072262764, 'train/grad_norm_pre_clip': 29.760833740234375, 'timing/data': 0.0005976689280942082, 'timing/model': 1.9340155320242047, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 1.31}) +[2026-06-23 16:38:34,222][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.058399323374032974, 'train/grad_norm_pre_clip': 52.548465728759766, 'timing/data': 0.0005999739514663815, 'timing/model': 1.9418550530681387, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 1.31}) +[2026-06-23 16:38:38,038][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.06068039685487747, 'train/grad_norm_pre_clip': 63.328433990478516, 'timing/data': 0.0005941749550402164, 'timing/model': 1.933233067044057, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 1.31}) +[2026-06-23 16:38:41,835][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.0478074811398983, 'train/grad_norm_pre_clip': 31.938621520996094, 'timing/data': 0.0006409140769392252, 'timing/model': 1.9229393049608916, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 1.31}) +[2026-06-23 16:38:45,652][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.032152920961380005, 'train/grad_norm_pre_clip': 14.976855278015137, 'timing/data': 0.0008084839209914207, 'timing/model': 1.932613654062152, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 1.31}) +[2026-06-23 16:38:49,463][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.07035637646913528, 'train/grad_norm_pre_clip': 24.367448806762695, 'timing/data': 0.0007145379204303026, 'timing/model': 1.9310891299974173, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 1.31}) +[2026-06-23 16:38:53,260][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.033780768513679504, 'train/grad_norm_pre_clip': 31.14291000366211, 'timing/data': 0.0007697230903431773, 'timing/model': 1.926469414960593, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 1.31}) +[2026-06-23 16:38:57,063][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.008219307288527489, 'train/grad_norm_pre_clip': 39.28338623046875, 'timing/data': 0.0005944269942119718, 'timing/model': 1.9294118511024863, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 1.31}) +[2026-06-23 16:39:00,870][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.03876369819045067, 'train/grad_norm_pre_clip': 36.678924560546875, 'timing/data': 0.000604286091402173, 'timing/model': 1.9372947741067037, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 1.31}) +[2026-06-23 16:39:04,675][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.11650495231151581, 'train/grad_norm_pre_clip': 63.50461959838867, 'timing/data': 0.0005961660062894225, 'timing/model': 1.9281958820065483, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 1.31}) +[2026-06-23 16:39:08,487][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.052032336592674255, 'train/grad_norm_pre_clip': 88.67546081542969, 'timing/data': 0.0008821369847282767, 'timing/model': 1.9361163940047845, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 1.31}) +[2026-06-23 16:39:12,291][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.03611290082335472, 'train/grad_norm_pre_clip': 58.1758918762207, 'timing/data': 0.0005829801084473729, 'timing/model': 1.923558181966655, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 1.31}) +[2026-06-23 16:39:16,294][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.09886246174573898, 'train/grad_norm_pre_clip': 34.063663482666016, 'timing/data': 0.0006183850346133113, 'timing/model': 2.124620346003212, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 1.31}) +[2026-06-23 16:39:20,162][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.10096292197704315, 'train/grad_norm_pre_clip': 25.726633071899414, 'timing/data': 0.0005940889241173863, 'timing/model': 1.953871478093788, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 1.32}) +[2026-06-23 16:39:23,978][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.07519780844449997, 'train/grad_norm_pre_clip': 26.930999755859375, 'timing/data': 0.0006094209384173155, 'timing/model': 1.9297241889871657, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 1.32}) +[2026-06-23 16:39:27,785][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3400 +[2026-06-23 16:41:36,653][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3400 +[2026-06-23 16:41:36,654][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3400 +[2026-06-23 16:42:57,959][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3400 +[2026-06-23 16:42:57,960][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.06420975923538208, 'train/grad_norm_pre_clip': 29.565351486206055, 'eval/action_loss/seconds': 128.86524753307458, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03246644511818886, 'eval/latency_2/loss': 0.03246644511818886, 'eval/flappy/loss': 0.03246644511818886, 'eval/flappy/latency_2/loss': 0.03246644511818886, 'eval/flap_precision': 0.9681139755766621, 'eval/flap_recall': 0.9628879892037787, 'eval/flap_f1': 0.9654939106901218, 'eval/flap_rate_model': 0.20472222222222222, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9654939106901219, 'eval/latency_2/flap_precision': 0.9681139755766621, 'eval/latency_2/flap_recall': 0.9628879892037787, 'eval/latency_2/flap_f1': 0.9654939106901218, 'eval/latency_2/flap_rate_model': 0.20472222222222222, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9654939106901219, 'eval/action_classification/data_wait_seconds': 1.5949106216430664, 'eval/action_classification/predict_seconds': 79.52468872070312, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005959359696134925, 'timing/model': 1.9422225820599124, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 1.32}) +[2026-06-23 16:43:01,737][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.018086936324834824, 'train/grad_norm_pre_clip': 25.13385581970215, 'timing/data': 0.0005981799913570285, 'timing/model': 1.9248292319243774, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 1.32}) +[2026-06-23 16:43:05,696][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.02254532463848591, 'train/grad_norm_pre_clip': 50.38178634643555, 'timing/data': 0.000591426040045917, 'timing/model': 1.9174924800172448, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 1.32}) +[2026-06-23 16:43:09,479][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.08280343562364578, 'train/grad_norm_pre_clip': 41.24116516113281, 'timing/data': 0.0006116789299994707, 'timing/model': 1.9223005709936842, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 1.32}) +[2026-06-23 16:43:13,268][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.03882450982928276, 'train/grad_norm_pre_clip': 24.358823776245117, 'timing/data': 0.0006383339641615748, 'timing/model': 1.9281602749833837, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 1.32}) +[2026-06-23 16:43:17,052][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.038737185299396515, 'train/grad_norm_pre_clip': 21.772232055664062, 'timing/data': 0.0005934500368312001, 'timing/model': 1.9228303560521454, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 1.32}) +[2026-06-23 16:43:20,836][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.010764161124825478, 'train/grad_norm_pre_clip': 37.88962173461914, 'timing/data': 0.0005934060318395495, 'timing/model': 1.920078472001478, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 1.32}) +[2026-06-23 16:43:24,642][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.04333759471774101, 'train/grad_norm_pre_clip': 18.968873977661133, 'timing/data': 0.0006088529480621219, 'timing/model': 1.935211066971533, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 1.32}) +[2026-06-23 16:43:28,423][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.06774068623781204, 'train/grad_norm_pre_clip': 20.164953231811523, 'timing/data': 0.0007989519508555532, 'timing/model': 1.9208958679810166, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 1.32}) +[2026-06-23 16:43:32,244][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.02464854158461094, 'train/grad_norm_pre_clip': 12.291847229003906, 'timing/data': 0.0006804800359532237, 'timing/model': 1.9168421239592135, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 1.32}) +[2026-06-23 16:43:36,200][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.03323884680867195, 'train/grad_norm_pre_clip': 22.747909545898438, 'timing/data': 0.0006242189556360245, 'timing/model': 2.096837230026722, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 1.32}) +[2026-06-23 16:43:39,976][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.03757288306951523, 'train/grad_norm_pre_clip': 21.51703453063965, 'timing/data': 0.0007383200572803617, 'timing/model': 1.9108396410010755, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 1.32}) +[2026-06-23 16:43:43,762][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.020423350855708122, 'train/grad_norm_pre_clip': 7.709963798522949, 'timing/data': 0.0006126939551904798, 'timing/model': 1.9151210190029815, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 1.32}) +[2026-06-23 16:43:47,548][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.04316055029630661, 'train/grad_norm_pre_clip': 46.10456466674805, 'timing/data': 0.0007356780115514994, 'timing/model': 1.9193894339259714, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 1.32}) +[2026-06-23 16:43:51,343][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.021791763603687286, 'train/grad_norm_pre_clip': 14.989710807800293, 'timing/data': 0.0006109359674155712, 'timing/model': 1.9223326640203595, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 1.32}) +[2026-06-23 16:43:55,199][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.02620602585375309, 'train/grad_norm_pre_clip': 20.96186065673828, 'timing/data': 0.0006423339946195483, 'timing/model': 1.9267294810852036, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 1.32}) +[2026-06-23 16:43:58,992][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.02496172860264778, 'train/grad_norm_pre_clip': 26.57670783996582, 'timing/data': 0.0008312850259244442, 'timing/model': 1.929300055024214, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 1.32}) +[2026-06-23 16:44:02,777][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.11870245635509491, 'train/grad_norm_pre_clip': 159.05328369140625, 'timing/data': 0.0007309000939130783, 'timing/model': 1.9154078719438985, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 1.32}) +[2026-06-23 16:44:06,562][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.09530028700828552, 'train/grad_norm_pre_clip': 29.229419708251953, 'timing/data': 0.000828977907076478, 'timing/model': 1.9268739050021395, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 1.32}) +[2026-06-23 16:44:10,337][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.021142473444342613, 'train/grad_norm_pre_clip': 13.43171215057373, 'timing/data': 0.0006047510541975498, 'timing/model': 1.9135277279419824, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 1.32}) +[2026-06-23 16:44:14,216][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.05674812197685242, 'train/grad_norm_pre_clip': 23.987525939941406, 'timing/data': 0.0008025220595300198, 'timing/model': 1.9261253760196269, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 1.32}) +[2026-06-23 16:44:18,002][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.03300214186310768, 'train/grad_norm_pre_clip': 54.25714111328125, 'timing/data': 0.0006147809326648712, 'timing/model': 1.919110232964158, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 1.32}) +[2026-06-23 16:44:21,789][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.01023316103965044, 'train/grad_norm_pre_clip': 37.70469665527344, 'timing/data': 0.0006260230438783765, 'timing/model': 1.9188640440115705, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 1.32}) +[2026-06-23 16:44:25,579][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.028323596343398094, 'train/grad_norm_pre_clip': 11.888497352600098, 'timing/data': 0.0006948639638721943, 'timing/model': 1.9178618430159986, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 1.32}) +[2026-06-23 16:44:29,391][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.0760405883193016, 'train/grad_norm_pre_clip': 38.48757553100586, 'timing/data': 0.0005776939215138555, 'timing/model': 1.938899454078637, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 1.33}) +[2026-06-23 16:44:33,191][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.026743164286017418, 'train/grad_norm_pre_clip': 20.39737892150879, 'timing/data': 0.0005928130121901631, 'timing/model': 1.9302719599800184, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 1.33}) +[2026-06-23 16:44:37,103][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.04272976890206337, 'train/grad_norm_pre_clip': 40.796974182128906, 'timing/data': 0.0006992389680817723, 'timing/model': 2.0400295350700617, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 1.33}) +[2026-06-23 16:44:40,917][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.054562509059906006, 'train/grad_norm_pre_clip': 30.032814025878906, 'timing/data': 0.0005910370964556932, 'timing/model': 1.9357405749615282, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 1.33}) +[2026-06-23 16:44:44,721][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.042118486016988754, 'train/grad_norm_pre_clip': 24.962909698486328, 'timing/data': 0.0006165329832583666, 'timing/model': 1.9213091799756512, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 1.33}) +[2026-06-23 16:44:48,525][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.1939030885696411, 'train/grad_norm_pre_clip': 53.748809814453125, 'timing/data': 0.0006451390217989683, 'timing/model': 1.9348593959584832, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 1.33}) +[2026-06-23 16:44:52,310][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.022893104702234268, 'train/grad_norm_pre_clip': 45.0360107421875, 'timing/data': 0.0006213329033926129, 'timing/model': 1.9144087369786575, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 1.33}) +[2026-06-23 16:44:56,139][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.031514722853899, 'train/grad_norm_pre_clip': 26.860666275024414, 'timing/data': 0.0006770309992134571, 'timing/model': 1.926435237051919, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 1.33}) +[2026-06-23 16:44:59,929][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.060971688479185104, 'train/grad_norm_pre_clip': 45.74178695678711, 'timing/data': 0.0006080759922042489, 'timing/model': 1.9258329959120601, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.33}) +[2026-06-23 16:45:03,748][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.040028903633356094, 'train/grad_norm_pre_clip': 26.070581436157227, 'timing/data': 0.0006012179655954242, 'timing/model': 1.9489820109447464, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 1.33}) +[2026-06-23 16:45:07,545][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.005975754465907812, 'train/grad_norm_pre_clip': 33.025306701660156, 'timing/data': 0.0006057519931346178, 'timing/model': 1.9258749620057642, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 1.33}) +[2026-06-23 16:45:11,345][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.028806710615754128, 'train/grad_norm_pre_clip': 7.1690449714660645, 'timing/data': 0.0005893770139664412, 'timing/model': 1.9319364900002256, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 1.33}) +[2026-06-23 16:45:15,332][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.04710690304636955, 'train/grad_norm_pre_clip': 36.866275787353516, 'timing/data': 0.0006620340282097459, 'timing/model': 2.104029997019097, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 1.33}) +[2026-06-23 16:45:19,134][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.033436186611652374, 'train/grad_norm_pre_clip': 55.20675277709961, 'timing/data': 0.0006302540423348546, 'timing/model': 1.929983232053928, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 1.33}) +[2026-06-23 16:45:22,929][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.0817575603723526, 'train/grad_norm_pre_clip': 28.1728515625, 'timing/data': 0.000914445030502975, 'timing/model': 1.919476076029241, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 1.33}) +[2026-06-23 16:45:26,742][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.09716834872961044, 'train/grad_norm_pre_clip': 38.82884979248047, 'timing/data': 0.00072134705260396, 'timing/model': 1.9339784139301628, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 1.33}) +[2026-06-23 16:45:30,526][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.15815481543540955, 'train/grad_norm_pre_clip': 54.31163787841797, 'timing/data': 0.000743082957342267, 'timing/model': 1.9193871449679136, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 1.33}) +[2026-06-23 16:45:34,321][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.04704548045992851, 'train/grad_norm_pre_clip': 24.315908432006836, 'timing/data': 0.0007996909553185105, 'timing/model': 1.9312898480566218, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 1.33}) +[2026-06-23 16:45:38,121][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.01062141265720129, 'train/grad_norm_pre_clip': 22.941490173339844, 'timing/data': 0.0006052289390936494, 'timing/model': 1.9276713330764323, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 1.33}) +[2026-06-23 16:45:41,923][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.041002314537763596, 'train/grad_norm_pre_clip': 34.730899810791016, 'timing/data': 0.0005891729379072785, 'timing/model': 1.9276683459756896, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 1.33}) +[2026-06-23 16:45:45,717][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.07478279620409012, 'train/grad_norm_pre_clip': 19.653242111206055, 'timing/data': 0.0005932800704613328, 'timing/model': 1.927511189947836, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 1.33}) +[2026-06-23 16:45:49,692][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.07223150879144669, 'train/grad_norm_pre_clip': 24.930376052856445, 'timing/data': 0.0005830629961565137, 'timing/model': 1.9311685239663348, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 1.33}) +[2026-06-23 16:45:53,495][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.08084908127784729, 'train/grad_norm_pre_clip': 56.96955871582031, 'timing/data': 0.0006134589202702045, 'timing/model': 1.9247435019351542, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 1.33}) +[2026-06-23 16:45:57,299][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.11772331595420837, 'train/grad_norm_pre_clip': 41.96290588378906, 'timing/data': 0.0006198739865794778, 'timing/model': 1.932854547048919, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 1.33}) +[2026-06-23 16:46:01,086][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.04873505234718323, 'train/grad_norm_pre_clip': 21.59824562072754, 'timing/data': 0.0006114300340414047, 'timing/model': 1.9246058959979564, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 1.33}) +[2026-06-23 16:46:04,898][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.16691310703754425, 'train/grad_norm_pre_clip': 81.59542846679688, 'timing/data': 0.0005779319908469915, 'timing/model': 1.9376963350223377, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 1.33}) +[2026-06-23 16:46:08,698][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.1354651153087616, 'train/grad_norm_pre_clip': 61.69083023071289, 'timing/data': 0.0005907999584451318, 'timing/model': 1.9281639809487388, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 1.34}) +[2026-06-23 16:46:12,510][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.014538291841745377, 'train/grad_norm_pre_clip': 29.491050720214844, 'timing/data': 0.0006006349576637149, 'timing/model': 1.9309623599983752, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 1.34}) +[2026-06-23 16:46:16,325][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.13707904517650604, 'train/grad_norm_pre_clip': 28.379690170288086, 'timing/data': 0.000602434971369803, 'timing/model': 1.9338920660084113, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 1.34}) +[2026-06-23 16:46:20,300][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.01073297020047903, 'train/grad_norm_pre_clip': 43.46031188964844, 'timing/data': 0.0006037360290065408, 'timing/model': 2.1053284099325538, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 1.34}) +[2026-06-23 16:46:24,123][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.11221873760223389, 'train/grad_norm_pre_clip': 34.83015441894531, 'timing/data': 0.0008077609818428755, 'timing/model': 1.9434071599971503, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 1.34}) +[2026-06-23 16:46:27,925][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.014497286640107632, 'train/grad_norm_pre_clip': 13.233527183532715, 'timing/data': 0.000623305095359683, 'timing/model': 1.92638584098313, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 1.34}) +[2026-06-23 16:46:31,737][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.04466252401471138, 'train/grad_norm_pre_clip': 39.50404357910156, 'timing/data': 0.0006032190285623074, 'timing/model': 1.9309473619796336, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 1.34}) +[2026-06-23 16:46:35,545][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.10940040647983551, 'train/grad_norm_pre_clip': 32.10011291503906, 'timing/data': 0.0005938360700383782, 'timing/model': 1.9385660890256986, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 1.34}) +[2026-06-23 16:46:39,337][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.04135889559984207, 'train/grad_norm_pre_clip': 48.5034065246582, 'timing/data': 0.0008123819716274738, 'timing/model': 1.9220806990051642, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 1.34}) +[2026-06-23 16:46:43,144][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.10143767297267914, 'train/grad_norm_pre_clip': 35.189430236816406, 'timing/data': 0.0005915400106459856, 'timing/model': 1.926936673000455, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 1.34}) +[2026-06-23 16:46:46,997][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.042937807738780975, 'train/grad_norm_pre_clip': 31.344762802124023, 'timing/data': 0.0005884820129722357, 'timing/model': 1.9815641529858112, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 1.34}) +[2026-06-23 16:46:50,788][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.08718652278184891, 'train/grad_norm_pre_clip': 70.81385040283203, 'timing/data': 0.0006032689707353711, 'timing/model': 1.924483231967315, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 1.34}) +[2026-06-23 16:46:54,814][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.06285153329372406, 'train/grad_norm_pre_clip': 48.86753845214844, 'timing/data': 0.0005840989761054516, 'timing/model': 1.9503645689692348, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 1.34}) +[2026-06-23 16:46:58,624][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.02147086150944233, 'train/grad_norm_pre_clip': 27.398923873901367, 'timing/data': 0.0006037279963493347, 'timing/model': 1.9369724600110203, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 1.34}) +[2026-06-23 16:47:02,407][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.08204562216997147, 'train/grad_norm_pre_clip': 32.81711959838867, 'timing/data': 0.0008133440278470516, 'timing/model': 1.9161601100349799, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 1.34}) +[2026-06-23 16:47:06,196][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.04721630737185478, 'train/grad_norm_pre_clip': 21.79998207092285, 'timing/data': 0.0006126700900495052, 'timing/model': 1.9218558879802004, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 1.34}) +[2026-06-23 16:47:10,008][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.06772327423095703, 'train/grad_norm_pre_clip': 23.810842514038086, 'timing/data': 0.000730465049855411, 'timing/model': 1.9349603280425072, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 1.34}) +[2026-06-23 16:47:13,816][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.028214234858751297, 'train/grad_norm_pre_clip': 17.55243682861328, 'timing/data': 0.0005916070658713579, 'timing/model': 1.9308416830608621, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 1.34}) +[2026-06-23 16:47:17,623][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.031434085220098495, 'train/grad_norm_pre_clip': 18.096757888793945, 'timing/data': 0.0006099959136918187, 'timing/model': 1.9354450550163165, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 1.34}) +[2026-06-23 16:47:21,437][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.04507051035761833, 'train/grad_norm_pre_clip': 28.543115615844727, 'timing/data': 0.0006151390261948109, 'timing/model': 1.9416292300447822, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 1.34}) +[2026-06-23 16:47:25,405][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.03959042206406593, 'train/grad_norm_pre_clip': 22.486255645751953, 'timing/data': 0.0006009220378473401, 'timing/model': 2.089852620032616, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 1.34}) +[2026-06-23 16:47:29,194][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.02932923473417759, 'train/grad_norm_pre_clip': 31.32126235961914, 'timing/data': 0.0006041009910404682, 'timing/model': 1.9213074919534847, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 1.34}) +[2026-06-23 16:47:33,029][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.08567457646131516, 'train/grad_norm_pre_clip': 58.25880432128906, 'timing/data': 0.0006437230622395873, 'timing/model': 1.9281789140077308, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 1.34}) +[2026-06-23 16:47:36,841][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.05475490540266037, 'train/grad_norm_pre_clip': 36.20892333984375, 'timing/data': 0.0007341790478676558, 'timing/model': 1.9292332470649853, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 1.34}) +[2026-06-23 16:47:40,648][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.10632188618183136, 'train/grad_norm_pre_clip': 32.97990036010742, 'timing/data': 0.0007040799828246236, 'timing/model': 1.9381905359914526, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 1.34}) +[2026-06-23 16:47:44,452][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.048966407775878906, 'train/grad_norm_pre_clip': 28.28652572631836, 'timing/data': 0.0005976000102236867, 'timing/model': 1.930579247069545, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 1.34}) +[2026-06-23 16:47:48,261][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.06604008376598358, 'train/grad_norm_pre_clip': 67.11231994628906, 'timing/data': 0.0005996329709887505, 'timing/model': 1.9276767950505018, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 1.35}) +[2026-06-23 16:47:52,050][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.23406684398651123, 'train/grad_norm_pre_clip': 48.9003791809082, 'timing/data': 0.0006110949907451868, 'timing/model': 1.9218578659929335, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 1.35}) +[2026-06-23 16:47:55,862][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.14551855623722076, 'train/grad_norm_pre_clip': 38.72383499145508, 'timing/data': 0.0006154559087008238, 'timing/model': 1.9355445560067892, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 1.35}) +[2026-06-23 16:47:59,668][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.13907206058502197, 'train/grad_norm_pre_clip': 62.2585563659668, 'timing/data': 0.0006072890246286988, 'timing/model': 1.9282102819997817, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 1.35}) +[2026-06-23 16:48:03,459][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.05028550326824188, 'train/grad_norm_pre_clip': 39.527740478515625, 'timing/data': 0.000610330025665462, 'timing/model': 1.9194808619795367, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 1.35}) +[2026-06-23 16:48:07,255][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.04831456393003464, 'train/grad_norm_pre_clip': 38.88349533081055, 'timing/data': 0.0007534440374001861, 'timing/model': 1.9243291550083086, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.35}) +[2026-06-23 16:48:11,055][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.01621631346642971, 'train/grad_norm_pre_clip': 30.097124099731445, 'timing/data': 0.0006333289202302694, 'timing/model': 1.9270397969521582, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 1.35}) +[2026-06-23 16:48:14,862][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.17959457635879517, 'train/grad_norm_pre_clip': 85.61585235595703, 'timing/data': 0.0008359870407730341, 'timing/model': 1.9377742200158536, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 1.35}) +[2026-06-23 16:48:18,672][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.05268353223800659, 'train/grad_norm_pre_clip': 32.61516571044922, 'timing/data': 0.0005937300156801939, 'timing/model': 1.9377211780520156, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 1.35}) +[2026-06-23 16:48:22,469][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.04766816645860672, 'train/grad_norm_pre_clip': 47.52635955810547, 'timing/data': 0.0006355829536914825, 'timing/model': 1.9228732900228351, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 1.35}) +[2026-06-23 16:48:26,276][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.008853472769260406, 'train/grad_norm_pre_clip': 23.321199417114258, 'timing/data': 0.0006471859524026513, 'timing/model': 1.9335311959730461, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 1.35}) +[2026-06-23 16:48:30,069][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.10193426162004471, 'train/grad_norm_pre_clip': 58.88296127319336, 'timing/data': 0.0005969490157440305, 'timing/model': 1.9208433679305017, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 1.35}) +[2026-06-23 16:48:33,885][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.11846482753753662, 'train/grad_norm_pre_clip': 27.92549705505371, 'timing/data': 0.0006000930443406105, 'timing/model': 1.935201738961041, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 1.35}) +[2026-06-23 16:48:37,715][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.020858826115727425, 'train/grad_norm_pre_clip': 33.227561950683594, 'timing/data': 0.000588569906540215, 'timing/model': 1.940432682982646, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 1.35}) +[2026-06-23 16:48:41,509][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.043369460850954056, 'train/grad_norm_pre_clip': 19.462635040283203, 'timing/data': 0.0006068319780752063, 'timing/model': 1.930104780010879, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 1.35}) +[2026-06-23 16:48:45,307][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.10363882035017014, 'train/grad_norm_pre_clip': 67.34981536865234, 'timing/data': 0.0005960830021649599, 'timing/model': 1.9296961450017989, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 1.35}) +[2026-06-23 16:48:49,093][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.039102669805288315, 'train/grad_norm_pre_clip': 24.642112731933594, 'timing/data': 0.000602749059908092, 'timing/model': 1.9217446710681543, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 1.35}) +[2026-06-23 16:48:52,890][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.2069701999425888, 'train/grad_norm_pre_clip': 90.06211853027344, 'timing/data': 0.000612420029938221, 'timing/model': 1.930392010952346, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 1.35}) +[2026-06-23 16:48:56,709][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.04334218427538872, 'train/grad_norm_pre_clip': 33.95232009887695, 'timing/data': 0.0005928239552304149, 'timing/model': 1.9400864450726658, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 1.35}) +[2026-06-23 16:49:00,523][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.040950555354356766, 'train/grad_norm_pre_clip': 22.105133056640625, 'timing/data': 0.0006541770417243242, 'timing/model': 1.944892646977678, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 1.35}) +[2026-06-23 16:49:04,527][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.16389478743076324, 'train/grad_norm_pre_clip': 61.9128532409668, 'timing/data': 0.0006128500681370497, 'timing/model': 2.129017331986688, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 1.35}) +[2026-06-23 16:49:08,333][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.05365169048309326, 'train/grad_norm_pre_clip': 42.23634338378906, 'timing/data': 0.0006083800690248609, 'timing/model': 1.9295179930049926, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 1.35}) +[2026-06-23 16:49:12,139][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.05247722566127777, 'train/grad_norm_pre_clip': 55.62342834472656, 'timing/data': 0.0006079970626160502, 'timing/model': 1.941355787916109, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 1.35}) +[2026-06-23 16:49:15,939][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.05337490513920784, 'train/grad_norm_pre_clip': 21.7679443359375, 'timing/data': 0.0006714670453220606, 'timing/model': 1.92825434403494, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 1.35}) +[2026-06-23 16:49:19,792][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3500 +[2026-06-23 16:51:27,415][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3500 +[2026-06-23 16:51:27,417][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3500 +[2026-06-23 16:52:49,960][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3500 +[2026-06-23 16:57:15,229][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3500 +[2026-06-23 16:57:15,229][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.007794021163135767, 'train/grad_norm_pre_clip': 29.75722312927246, 'eval/action_loss/seconds': 127.6201489130035, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.032517481595277786, 'eval/latency_2/loss': 0.032517481595277786, 'eval/flappy/loss': 0.032517481595277786, 'eval/flappy/latency_2/loss': 0.032517481595277786, 'eval/flap_precision': 0.9682217714672076, 'eval/flap_recall': 0.9662618083670715, 'eval/flap_f1': 0.9672407970280311, 'eval/flap_rate_model': 0.20541666666666666, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9672407970280311, 'eval/latency_2/flap_precision': 0.9682217714672076, 'eval/latency_2/flap_recall': 0.9662618083670715, 'eval/latency_2/flap_f1': 0.9672407970280311, 'eval/latency_2/flap_rate_model': 0.20541666666666666, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9672407970280311, 'eval/action_classification/data_wait_seconds': 1.465659260749817, 'eval/action_classification/predict_seconds': 80.88972473144531, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 161.04000000000397, 'rl_games_eval/mid_train/mean_length': 1316.8, 'rl_games_eval/mid_train/std_reward': 151.0126431793143, 'rl_games_eval/mid_train/std_length': 1216.3360390944601, 'rl_games_eval/mid_train/macro_mean_reward': 161.04000000000397, 'rl_games_eval/mid_train/macro_mean_length': 1316.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 161.04000000000397, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1316.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 151.0126431793143, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1216.3360390944601, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 161.04000000000397, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006046229973435402, 'timing/model': 1.9478610249934718, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 1.35}) +[2026-06-23 16:57:15,229][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_3500_state +[2026-06-23 16:57:15,232][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 16:57:18,944][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_3500_state/model.safetensors +[2026-06-23 16:57:21,895][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_3500_state/optimizer.bin +[2026-06-23 16:57:21,895][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_3500_state/sampler.bin +[2026-06-23 16:57:21,895][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-23 16:57:21,896][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-23 16:57:21,923][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-23 16:57:21,924][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 16:57:21,962][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 16:57:25,669][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.1108940914273262, 'train/grad_norm_pre_clip': 22.56689453125, 'timing/data': 0.0006103720515966415, 'timing/model': 1.88216340506915, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 1.35}) +[2026-06-23 16:57:29,409][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.05732675641775131, 'train/grad_norm_pre_clip': 23.22210121154785, 'timing/data': 0.0006375339580699801, 'timing/model': 1.9025297100888565, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 1.36}) +[2026-06-23 16:57:33,144][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.03529290854930878, 'train/grad_norm_pre_clip': 13.213385581970215, 'timing/data': 0.000615595025010407, 'timing/model': 1.8963218539720401, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 1.36}) +[2026-06-23 16:57:36,880][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.08984262496232986, 'train/grad_norm_pre_clip': 32.92696762084961, 'timing/data': 0.0005854939809069037, 'timing/model': 1.8940685730194673, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 1.36}) +[2026-06-23 16:57:40,627][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.04820120707154274, 'train/grad_norm_pre_clip': 43.3854866027832, 'timing/data': 0.0006746470462530851, 'timing/model': 1.9065982330357656, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 1.36}) +[2026-06-23 16:57:44,379][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.03298366814851761, 'train/grad_norm_pre_clip': 11.027338981628418, 'timing/data': 0.0005810919683426619, 'timing/model': 1.903330945991911, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 1.36}) +[2026-06-23 16:57:48,126][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.07058096677064896, 'train/grad_norm_pre_clip': 32.96094512939453, 'timing/data': 0.0006014429964125156, 'timing/model': 1.9043403980322182, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 1.36}) +[2026-06-23 16:57:51,894][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.026315269991755486, 'train/grad_norm_pre_clip': 15.07463264465332, 'timing/data': 0.0007176620420068502, 'timing/model': 1.9022944079479203, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 1.36}) +[2026-06-23 16:57:55,645][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.04407394677400589, 'train/grad_norm_pre_clip': 37.405311584472656, 'timing/data': 0.0006010829238221049, 'timing/model': 1.90316998003982, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 1.36}) +[2026-06-23 16:57:59,399][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.07143159955739975, 'train/grad_norm_pre_clip': 26.536279678344727, 'timing/data': 0.0005903539713472128, 'timing/model': 1.9076800890034065, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 1.36}) +[2026-06-23 16:58:03,361][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.018581459298729897, 'train/grad_norm_pre_clip': 15.42788028717041, 'timing/data': 0.0006029900396242738, 'timing/model': 2.0958079049596563, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 1.36}) +[2026-06-23 16:58:07,125][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.11320789158344269, 'train/grad_norm_pre_clip': 45.702247619628906, 'timing/data': 0.0005950520280748606, 'timing/model': 1.9179552309215069, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 1.36}) +[2026-06-23 16:58:10,898][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.06150585040450096, 'train/grad_norm_pre_clip': 38.631248474121094, 'timing/data': 0.0006043208995833993, 'timing/model': 1.909492919105105, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 1.36}) +[2026-06-23 16:58:14,678][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.10687393695116043, 'train/grad_norm_pre_clip': 34.38408660888672, 'timing/data': 0.0006989550311118364, 'timing/model': 1.914737054030411, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 1.36}) +[2026-06-23 16:58:18,449][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.14652177691459656, 'train/grad_norm_pre_clip': 46.57625961303711, 'timing/data': 0.00059874402359128, 'timing/model': 1.9158908349927515, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 1.36}) +[2026-06-23 16:58:22,230][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.01021566055715084, 'train/grad_norm_pre_clip': 12.014117240905762, 'timing/data': 0.0005853709299117327, 'timing/model': 1.9168984539574012, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 1.36}) +[2026-06-23 16:58:26,006][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.010470504872500896, 'train/grad_norm_pre_clip': 11.856325149536133, 'timing/data': 0.0007874550065025687, 'timing/model': 1.9181946149328724, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 1.36}) +[2026-06-23 16:58:29,787][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.13441535830497742, 'train/grad_norm_pre_clip': 34.00503921508789, 'timing/data': 0.0005875780479982495, 'timing/model': 1.9169987679924816, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 1.36}) +[2026-06-23 16:58:33,573][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.022165849804878235, 'train/grad_norm_pre_clip': 15.495747566223145, 'timing/data': 0.0007685070158913732, 'timing/model': 1.92223282600753, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 1.36}) +[2026-06-23 16:58:37,355][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.08867454528808594, 'train/grad_norm_pre_clip': 40.414344787597656, 'timing/data': 0.0006709439912810922, 'timing/model': 1.918193737976253, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 1.36}) +[2026-06-23 16:58:41,124][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.03014727495610714, 'train/grad_norm_pre_clip': 31.084630966186523, 'timing/data': 0.0006276119966059923, 'timing/model': 1.9089109500637278, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 1.36}) +[2026-06-23 16:58:44,913][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.07469723373651505, 'train/grad_norm_pre_clip': 22.311260223388672, 'timing/data': 0.0006894440157338977, 'timing/model': 1.919907644041814, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 1.36}) +[2026-06-23 16:58:48,703][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.009741310961544514, 'train/grad_norm_pre_clip': 17.86735725402832, 'timing/data': 0.000644140993244946, 'timing/model': 1.9146499950438738, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 1.36}) +[2026-06-23 16:58:52,480][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.1503295749425888, 'train/grad_norm_pre_clip': 45.95746994018555, 'timing/data': 0.0006065219640731812, 'timing/model': 1.9117989080259576, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 1.36}) +[2026-06-23 16:58:56,275][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.08901183307170868, 'train/grad_norm_pre_clip': 36.85747146606445, 'timing/data': 0.0005979649722576141, 'timing/model': 1.9232695989776403, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 1.36}) +[2026-06-23 16:59:00,060][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.05173796787858009, 'train/grad_norm_pre_clip': 38.73914337158203, 'timing/data': 0.0006075360579416156, 'timing/model': 1.9201447080122307, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 1.36}) +[2026-06-23 16:59:03,835][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.05758646875619888, 'train/grad_norm_pre_clip': 26.207365036010742, 'timing/data': 0.0006072450196370482, 'timing/model': 1.9096802970161662, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 1.37}) +[2026-06-23 16:59:07,631][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.058594826608896255, 'train/grad_norm_pre_clip': 46.09054183959961, 'timing/data': 0.0006189809646457434, 'timing/model': 1.9228472950635478, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 1.37}) +[2026-06-23 16:59:11,425][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.10580357164144516, 'train/grad_norm_pre_clip': 26.920778274536133, 'timing/data': 0.0005931789055466652, 'timing/model': 1.9269032439915463, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 1.37}) +[2026-06-23 16:59:15,209][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.047776829451322556, 'train/grad_norm_pre_clip': 20.45558738708496, 'timing/data': 0.0006590349366888404, 'timing/model': 1.9201762860175222, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.37}) +[2026-06-23 16:59:18,987][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.009519406594336033, 'train/grad_norm_pre_clip': 21.13152313232422, 'timing/data': 0.0006129919784143567, 'timing/model': 1.9144426700659096, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 1.37}) +[2026-06-23 16:59:22,777][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.05545379966497421, 'train/grad_norm_pre_clip': 31.130393981933594, 'timing/data': 0.0005836669588461518, 'timing/model': 1.9245335930027068, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 1.37}) +[2026-06-23 16:59:26,555][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.0483090803027153, 'train/grad_norm_pre_clip': 32.48859405517578, 'timing/data': 0.0005759820342063904, 'timing/model': 1.9197628360707313, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 1.37}) +[2026-06-23 16:59:30,366][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.025335311889648438, 'train/grad_norm_pre_clip': 19.085054397583008, 'timing/data': 0.000632600043900311, 'timing/model': 1.9354309629416093, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 1.37}) +[2026-06-23 16:59:34,156][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.01618524082005024, 'train/grad_norm_pre_clip': 31.972841262817383, 'timing/data': 0.0005941509734839201, 'timing/model': 1.9232534669572487, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 1.37}) +[2026-06-23 16:59:37,955][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.19584175944328308, 'train/grad_norm_pre_clip': 70.12156677246094, 'timing/data': 0.0005931539926677942, 'timing/model': 1.9318201399873942, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 1.37}) +[2026-06-23 16:59:41,746][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.07038004696369171, 'train/grad_norm_pre_clip': 17.78849983215332, 'timing/data': 0.0005847300635650754, 'timing/model': 1.9228790119523183, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 1.37}) +[2026-06-23 16:59:45,550][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.06694673746824265, 'train/grad_norm_pre_clip': 30.59232521057129, 'timing/data': 0.0006404640153050423, 'timing/model': 1.9445441500283778, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 1.37}) +[2026-06-23 16:59:49,358][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.11912493407726288, 'train/grad_norm_pre_clip': 48.89874267578125, 'timing/data': 0.0007251729257404804, 'timing/model': 1.9336609330493957, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 1.37}) +[2026-06-23 16:59:53,147][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.025463726371526718, 'train/grad_norm_pre_clip': 10.16899299621582, 'timing/data': 0.0006210940191522241, 'timing/model': 1.9158571810694411, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 1.37}) +[2026-06-23 16:59:56,932][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.01987026259303093, 'train/grad_norm_pre_clip': 45.68952941894531, 'timing/data': 0.0009831030620262027, 'timing/model': 1.9206191999837756, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 1.37}) +[2026-06-23 17:00:00,735][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.07108458131551743, 'train/grad_norm_pre_clip': 60.92276382446289, 'timing/data': 0.0006040430162101984, 'timing/model': 1.9351223499979824, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 1.37}) +[2026-06-23 17:00:04,541][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.03610682860016823, 'train/grad_norm_pre_clip': 18.605024337768555, 'timing/data': 0.0006116400472819805, 'timing/model': 1.9376574329799041, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 1.37}) +[2026-06-23 17:00:08,347][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.03250061720609665, 'train/grad_norm_pre_clip': 12.371334075927734, 'timing/data': 0.0005949549376964569, 'timing/model': 1.9337962550343946, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 1.37}) +[2026-06-23 17:00:12,157][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.06937813758850098, 'train/grad_norm_pre_clip': 28.473783493041992, 'timing/data': 0.0005973080405965447, 'timing/model': 1.938026115996763, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 1.37}) +[2026-06-23 17:00:16,135][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.008476356975734234, 'train/grad_norm_pre_clip': 15.694464683532715, 'timing/data': 0.0006043009925633669, 'timing/model': 1.9285869110608473, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 1.37}) +[2026-06-23 17:00:19,941][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.09989193826913834, 'train/grad_norm_pre_clip': 32.831016540527344, 'timing/data': 0.0006645190296694636, 'timing/model': 1.9290663409046829, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 1.37}) +[2026-06-23 17:00:23,743][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.05714435502886772, 'train/grad_norm_pre_clip': 33.537506103515625, 'timing/data': 0.0006165569648146629, 'timing/model': 1.92193278495688, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 1.37}) +[2026-06-23 17:00:27,530][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.11665444821119308, 'train/grad_norm_pre_clip': 42.48435974121094, 'timing/data': 0.000602340092882514, 'timing/model': 1.9188982580089942, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 1.37}) +[2026-06-23 17:00:31,330][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.009369505569338799, 'train/grad_norm_pre_clip': 18.507938385009766, 'timing/data': 0.0006001510191708803, 'timing/model': 1.9312444899696857, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 1.37}) +[2026-06-23 17:00:35,144][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.02900654636323452, 'train/grad_norm_pre_clip': 15.658980369567871, 'timing/data': 0.0007101009832695127, 'timing/model': 1.932406045962125, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 1.37}) +[2026-06-23 17:00:38,943][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.028201419860124588, 'train/grad_norm_pre_clip': 28.365671157836914, 'timing/data': 0.0006764839636161923, 'timing/model': 1.9300199470017105, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 1.37}) +[2026-06-23 17:00:42,746][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.23057620227336884, 'train/grad_norm_pre_clip': 37.239749908447266, 'timing/data': 0.0005903900600969791, 'timing/model': 1.925699239014648, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 1.38}) +[2026-06-23 17:00:46,735][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.0785730630159378, 'train/grad_norm_pre_clip': 16.213180541992188, 'timing/data': 0.0007887610699981451, 'timing/model': 2.110495980014093, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 1.38}) +[2026-06-23 17:00:50,540][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.09672893583774567, 'train/grad_norm_pre_clip': 43.18324661254883, 'timing/data': 0.0005952219944447279, 'timing/model': 1.932935282937251, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 1.38}) +[2026-06-23 17:00:54,359][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.03986033424735069, 'train/grad_norm_pre_clip': 35.66558074951172, 'timing/data': 0.0005953969666734338, 'timing/model': 1.934813686995767, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 1.38}) +[2026-06-23 17:00:58,149][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.1042785570025444, 'train/grad_norm_pre_clip': 39.60967254638672, 'timing/data': 0.0006466750055551529, 'timing/model': 1.917862404952757, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 1.38}) +[2026-06-23 17:01:01,951][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.007214643992483616, 'train/grad_norm_pre_clip': 22.897872924804688, 'timing/data': 0.0008871030295267701, 'timing/model': 1.929793471004814, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 1.38}) +[2026-06-23 17:01:05,755][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.03495771437883377, 'train/grad_norm_pre_clip': 44.794189453125, 'timing/data': 0.0008194530382752419, 'timing/model': 1.9287061599316075, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 1.38}) +[2026-06-23 17:01:09,575][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.21388177573680878, 'train/grad_norm_pre_clip': 49.83396911621094, 'timing/data': 0.0006093579577282071, 'timing/model': 1.9407484770053998, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 1.38}) +[2026-06-23 17:01:13,375][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.10905883461236954, 'train/grad_norm_pre_clip': 33.77085876464844, 'timing/data': 0.0008942220592871308, 'timing/model': 1.9353365829447284, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 1.38}) +[2026-06-23 17:01:17,188][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.08821016550064087, 'train/grad_norm_pre_clip': 59.3734016418457, 'timing/data': 0.0008119570557028055, 'timing/model': 1.9365225789370015, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 1.38}) +[2026-06-23 17:01:20,990][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.03795064613223076, 'train/grad_norm_pre_clip': 18.575462341308594, 'timing/data': 0.0006682099774479866, 'timing/model': 1.9248833250021562, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 1.38}) +[2026-06-23 17:01:24,783][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.02618074230849743, 'train/grad_norm_pre_clip': 14.580870628356934, 'timing/data': 0.0006018170388415456, 'timing/model': 1.92988458101172, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 1.38}) +[2026-06-23 17:01:28,593][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.05557292327284813, 'train/grad_norm_pre_clip': 18.675796508789062, 'timing/data': 0.0008061699336394668, 'timing/model': 1.9280802649445832, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 1.38}) +[2026-06-23 17:01:32,401][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.026571907103061676, 'train/grad_norm_pre_clip': 21.8824462890625, 'timing/data': 0.0005878220545127988, 'timing/model': 1.9263591589406133, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 1.38}) +[2026-06-23 17:01:36,197][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.09049323201179504, 'train/grad_norm_pre_clip': 39.198707580566406, 'timing/data': 0.000631108065135777, 'timing/model': 1.926218808046542, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 1.38}) +[2026-06-23 17:01:39,982][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.12504200637340546, 'train/grad_norm_pre_clip': 42.014320373535156, 'timing/data': 0.0006618789630010724, 'timing/model': 1.9242236749269068, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 1.38}) +[2026-06-23 17:01:43,782][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.04819400608539581, 'train/grad_norm_pre_clip': 20.401569366455078, 'timing/data': 0.0005840949015691876, 'timing/model': 1.9325827329885215, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 1.38}) +[2026-06-23 17:01:47,581][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.0892239660024643, 'train/grad_norm_pre_clip': 28.961698532104492, 'timing/data': 0.0005870870081707835, 'timing/model': 1.9368120559956878, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 1.38}) +[2026-06-23 17:01:51,413][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.051557816565036774, 'train/grad_norm_pre_clip': 42.956703186035156, 'timing/data': 0.0008185969199985266, 'timing/model': 1.919419453945011, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 1.38}) +[2026-06-23 17:01:55,227][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.03697855770587921, 'train/grad_norm_pre_clip': 35.07318878173828, 'timing/data': 0.0007028339896351099, 'timing/model': 1.9307564700720832, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 1.38}) +[2026-06-23 17:01:59,024][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.02768944390118122, 'train/grad_norm_pre_clip': 10.465835571289062, 'timing/data': 0.0006072350079193711, 'timing/model': 1.9298468199558556, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 1.38}) +[2026-06-23 17:02:02,833][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.035453375428915024, 'train/grad_norm_pre_clip': 22.451904296875, 'timing/data': 0.0007617620285600424, 'timing/model': 1.9393914869287983, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 1.38}) +[2026-06-23 17:02:06,657][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.019497742876410484, 'train/grad_norm_pre_clip': 13.64990234375, 'timing/data': 0.000705305952578783, 'timing/model': 1.9351655920036137, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 1.38}) +[2026-06-23 17:02:10,453][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.009526275098323822, 'train/grad_norm_pre_clip': 18.23130989074707, 'timing/data': 0.0005959629779681563, 'timing/model': 1.9296569779980928, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 1.38}) +[2026-06-23 17:02:14,259][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.004881225060671568, 'train/grad_norm_pre_clip': 16.535188674926758, 'timing/data': 0.0006047000642865896, 'timing/model': 1.93251600000076, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 1.38}) +[2026-06-23 17:02:18,074][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.11178711801767349, 'train/grad_norm_pre_clip': 26.474096298217773, 'timing/data': 0.0005933489883318543, 'timing/model': 1.9376047580735758, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 1.38}) +[2026-06-23 17:02:21,867][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.03551359102129936, 'train/grad_norm_pre_clip': 13.417627334594727, 'timing/data': 0.0005908560706302524, 'timing/model': 1.9249574790010229, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.39}) +[2026-06-23 17:02:25,669][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.10033785551786423, 'train/grad_norm_pre_clip': 44.87083053588867, 'timing/data': 0.0006154270377010107, 'timing/model': 1.9306338159367442, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 1.39}) +[2026-06-23 17:02:29,471][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.03732636570930481, 'train/grad_norm_pre_clip': 34.85770034790039, 'timing/data': 0.0006002250593155622, 'timing/model': 1.928533241036348, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 1.39}) +[2026-06-23 17:02:33,274][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.058854274451732635, 'train/grad_norm_pre_clip': 26.093326568603516, 'timing/data': 0.0006367419846355915, 'timing/model': 1.9214273799443617, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 1.39}) +[2026-06-23 17:02:37,083][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.04589112848043442, 'train/grad_norm_pre_clip': 34.804344177246094, 'timing/data': 0.0007551239104941487, 'timing/model': 1.9355253869434819, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 1.39}) +[2026-06-23 17:02:40,894][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.014493357390165329, 'train/grad_norm_pre_clip': 31.033273696899414, 'timing/data': 0.0005908739985898137, 'timing/model': 1.9361087110592052, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 1.39}) +[2026-06-23 17:02:44,692][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.10948587954044342, 'train/grad_norm_pre_clip': 40.38483428955078, 'timing/data': 0.0005772400181740522, 'timing/model': 1.926900637918152, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 1.39}) +[2026-06-23 17:02:48,499][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.06751803308725357, 'train/grad_norm_pre_clip': 23.44037628173828, 'timing/data': 0.0005937750684097409, 'timing/model': 1.9314479590393603, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 1.39}) +[2026-06-23 17:02:52,304][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.04124743491411209, 'train/grad_norm_pre_clip': 31.67919158935547, 'timing/data': 0.0009100820170715451, 'timing/model': 1.9279326139949262, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 1.39}) +[2026-06-23 17:02:56,099][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.03801010549068451, 'train/grad_norm_pre_clip': 27.793174743652344, 'timing/data': 0.0005904519930481911, 'timing/model': 1.9301548800431192, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 1.39}) +[2026-06-23 17:03:00,093][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.017089968547225, 'train/grad_norm_pre_clip': 27.54201889038086, 'timing/data': 0.0007695630192756653, 'timing/model': 1.9309451170265675, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 1.39}) +[2026-06-23 17:03:03,946][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.02410925179719925, 'train/grad_norm_pre_clip': 17.10283660888672, 'timing/data': 0.0006243590032681823, 'timing/model': 1.9822189200203866, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 1.39}) +[2026-06-23 17:03:07,750][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.05207471922039986, 'train/grad_norm_pre_clip': 28.651044845581055, 'timing/data': 0.0005975719541311264, 'timing/model': 1.9251170170027763, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 1.39}) +[2026-06-23 17:03:11,557][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.09495095908641815, 'train/grad_norm_pre_clip': 25.15300178527832, 'timing/data': 0.000585277914069593, 'timing/model': 1.9381729020969942, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 1.39}) +[2026-06-23 17:03:15,367][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.05263397842645645, 'train/grad_norm_pre_clip': 24.57689094543457, 'timing/data': 0.0006013100501149893, 'timing/model': 1.9291011659661308, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 1.39}) +[2026-06-23 17:03:19,183][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.03583609685301781, 'train/grad_norm_pre_clip': 29.703182220458984, 'timing/data': 0.0005940069677308202, 'timing/model': 1.9343995950184762, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 1.39}) +[2026-06-23 17:03:22,976][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.0974130854010582, 'train/grad_norm_pre_clip': 33.02253341674805, 'timing/data': 0.0006132959388196468, 'timing/model': 1.9232439519837499, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 1.39}) +[2026-06-23 17:03:26,764][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.026133045554161072, 'train/grad_norm_pre_clip': 34.82976531982422, 'timing/data': 0.0008384969551116228, 'timing/model': 1.9251079129753634, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 1.39}) +[2026-06-23 17:03:30,565][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.054964885115623474, 'train/grad_norm_pre_clip': 15.420580863952637, 'timing/data': 0.0006958739832043648, 'timing/model': 1.9201928289839998, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 1.39}) +[2026-06-23 17:03:34,370][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.04691863805055618, 'train/grad_norm_pre_clip': 20.3643798828125, 'timing/data': 0.0006199029739946127, 'timing/model': 1.9348759170388803, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 1.39}) +[2026-06-23 17:03:38,164][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.08038472384214401, 'train/grad_norm_pre_clip': 40.58315658569336, 'timing/data': 0.0005909759784117341, 'timing/model': 1.933582306955941, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 1.39}) +[2026-06-23 17:03:41,957][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3600 +[2026-06-23 17:05:49,959][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3600 +[2026-06-23 17:05:49,961][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3600 +[2026-06-23 17:07:13,022][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3600 +[2026-06-23 17:07:13,023][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.03273068740963936, 'train/grad_norm_pre_clip': 12.472230911254883, 'eval/action_loss/seconds': 127.99923597299494, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03351958841085434, 'eval/latency_2/loss': 0.03351958841085434, 'eval/flappy/loss': 0.03351958841085434, 'eval/flappy/latency_2/loss': 0.03351958841085434, 'eval/flap_precision': 0.9539776462853385, 'eval/flap_recall': 0.9790823211875843, 'eval/flap_f1': 0.9663669663669664, 'eval/flap_rate_model': 0.21125, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9663669663669664, 'eval/latency_2/flap_precision': 0.9539776462853385, 'eval/latency_2/flap_recall': 0.9790823211875843, 'eval/latency_2/flap_f1': 0.9663669663669664, 'eval/latency_2/flap_rate_model': 0.21125, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9663669663669664, 'eval/action_classification/data_wait_seconds': 1.200782299041748, 'eval/action_classification/predict_seconds': 81.64485931396484, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0007790119852870703, 'timing/model': 1.9228701640386134, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 1.39}) +[2026-06-23 17:07:16,787][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.06228964403271675, 'train/grad_norm_pre_clip': 32.64741134643555, 'timing/data': 0.0005965139716863632, 'timing/model': 1.9117637010058388, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 1.39}) +[2026-06-23 17:07:20,749][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.03737602010369301, 'train/grad_norm_pre_clip': 11.697419166564941, 'timing/data': 0.0007064479868859053, 'timing/model': 1.9139017189154401, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 1.39}) +[2026-06-23 17:07:24,514][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.035373371094465256, 'train/grad_norm_pre_clip': 19.09304428100586, 'timing/data': 0.0006195839960128069, 'timing/model': 1.9071841510012746, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 1.39}) +[2026-06-23 17:07:28,274][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.017710721120238304, 'train/grad_norm_pre_clip': 42.91310119628906, 'timing/data': 0.0006142660276964307, 'timing/model': 1.9101089119212702, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 1.39}) +[2026-06-23 17:07:32,048][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.08555836230516434, 'train/grad_norm_pre_clip': 57.59633255004883, 'timing/data': 0.0006049310322850943, 'timing/model': 1.9124718310777098, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 1.4}) +[2026-06-23 17:07:35,821][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.06018754094839096, 'train/grad_norm_pre_clip': 53.9289665222168, 'timing/data': 0.000665297033265233, 'timing/model': 1.915330200921744, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 1.4}) +[2026-06-23 17:07:39,618][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.05625439062714577, 'train/grad_norm_pre_clip': 23.90729331970215, 'timing/data': 0.0007094170432537794, 'timing/model': 1.9227013350464404, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 1.4}) +[2026-06-23 17:07:43,402][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.04330028221011162, 'train/grad_norm_pre_clip': 38.75325393676758, 'timing/data': 0.0006088160444051027, 'timing/model': 1.916459973086603, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 1.4}) +[2026-06-23 17:07:47,174][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.07658340781927109, 'train/grad_norm_pre_clip': 68.3531494140625, 'timing/data': 0.0005956239765509963, 'timing/model': 1.915707753971219, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 1.4}) +[2026-06-23 17:07:51,077][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.03523340821266174, 'train/grad_norm_pre_clip': 21.686315536499023, 'timing/data': 0.0007044500671327114, 'timing/model': 2.0456062799785286, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 1.4}) +[2026-06-23 17:07:54,865][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.06519278138875961, 'train/grad_norm_pre_clip': 20.370891571044922, 'timing/data': 0.0005880959797650576, 'timing/model': 1.9186774920672178, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 1.4}) +[2026-06-23 17:07:58,647][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.11778859049081802, 'train/grad_norm_pre_clip': 30.443754196166992, 'timing/data': 0.0005837560165673494, 'timing/model': 1.918456600047648, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 1.4}) +[2026-06-23 17:08:02,442][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.08328010886907578, 'train/grad_norm_pre_clip': 43.08747100830078, 'timing/data': 0.0006015250692144036, 'timing/model': 1.9261645909864455, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 1.4}) +[2026-06-23 17:08:06,230][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.011330201290547848, 'train/grad_norm_pre_clip': 10.282032012939453, 'timing/data': 0.0007913339650258422, 'timing/model': 1.9258865000447258, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 1.4}) +[2026-06-23 17:08:10,018][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.1118345558643341, 'train/grad_norm_pre_clip': 31.518869400024414, 'timing/data': 0.000607755035161972, 'timing/model': 1.9227814810583368, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 1.4}) +[2026-06-23 17:08:13,801][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.055523648858070374, 'train/grad_norm_pre_clip': 15.07003116607666, 'timing/data': 0.0006170619744807482, 'timing/model': 1.921001972979866, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 1.4}) +[2026-06-23 17:08:17,584][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.04173237830400467, 'train/grad_norm_pre_clip': 13.881457328796387, 'timing/data': 0.0006111729890108109, 'timing/model': 1.9173615969484672, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 1.4}) +[2026-06-23 17:08:21,382][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.057642653584480286, 'train/grad_norm_pre_clip': 21.424001693725586, 'timing/data': 0.0005960100097581744, 'timing/model': 1.925334868952632, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 1.4}) +[2026-06-23 17:08:25,313][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.05668475851416588, 'train/grad_norm_pre_clip': 30.645044326782227, 'timing/data': 0.0007991810562089086, 'timing/model': 2.05659388191998, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 1.4}) +[2026-06-23 17:08:29,104][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.038229454308748245, 'train/grad_norm_pre_clip': 29.28119659423828, 'timing/data': 0.0007750439690425992, 'timing/model': 1.9253928699763492, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 1.4}) +[2026-06-23 17:08:32,893][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.054225847125053406, 'train/grad_norm_pre_clip': 32.657623291015625, 'timing/data': 0.0006033820100128651, 'timing/model': 1.9206624160287902, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 1.4}) +[2026-06-23 17:08:36,689][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.020424123853445053, 'train/grad_norm_pre_clip': 10.016927719116211, 'timing/data': 0.000603034975938499, 'timing/model': 1.9264986669877544, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 1.4}) +[2026-06-23 17:08:40,504][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.11021102964878082, 'train/grad_norm_pre_clip': 23.93989372253418, 'timing/data': 0.0005876331124454737, 'timing/model': 1.9400398529833183, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 1.4}) +[2026-06-23 17:08:44,296][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.008211649022996426, 'train/grad_norm_pre_clip': 14.861686706542969, 'timing/data': 0.0008972580544650555, 'timing/model': 1.9206000569975004, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 1.4}) +[2026-06-23 17:08:48,096][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.048553016036748886, 'train/grad_norm_pre_clip': 22.481571197509766, 'timing/data': 0.0006292050238698721, 'timing/model': 1.9245734619908035, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 1.4}) +[2026-06-23 17:08:51,902][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.024317679926753044, 'train/grad_norm_pre_clip': 29.977739334106445, 'timing/data': 0.0006044598994776607, 'timing/model': 1.9310250960988924, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 1.4}) +[2026-06-23 17:08:55,712][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.0015328960726037621, 'train/grad_norm_pre_clip': 34.79890441894531, 'timing/data': 0.0005962079158052802, 'timing/model': 1.936274983920157, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 1.4}) +[2026-06-23 17:08:59,674][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.08115968108177185, 'train/grad_norm_pre_clip': 45.95213317871094, 'timing/data': 0.0006088860100135207, 'timing/model': 1.9305219009984285, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.4}) +[2026-06-23 17:09:03,490][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.031602829694747925, 'train/grad_norm_pre_clip': 26.083908081054688, 'timing/data': 0.0008228269871324301, 'timing/model': 1.9387255549663678, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 1.4}) +[2026-06-23 17:09:07,284][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.0054762521758675575, 'train/grad_norm_pre_clip': 12.85139274597168, 'timing/data': 0.0005964819574728608, 'timing/model': 1.9310541769955307, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 1.4}) +[2026-06-23 17:09:11,080][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.004967729561030865, 'train/grad_norm_pre_clip': 5.9833478927612305, 'timing/data': 0.0006107400404289365, 'timing/model': 1.928931770962663, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 1.41}) +[2026-06-23 17:09:14,873][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.08611167222261429, 'train/grad_norm_pre_clip': 39.768550872802734, 'timing/data': 0.0005979120032861829, 'timing/model': 1.9277693530311808, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 1.41}) +[2026-06-23 17:09:18,681][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.11297839879989624, 'train/grad_norm_pre_clip': 32.981327056884766, 'timing/data': 0.0006070189410820603, 'timing/model': 1.9285303810611367, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 1.41}) +[2026-06-23 17:09:22,471][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.03152322396636009, 'train/grad_norm_pre_clip': 27.8769588470459, 'timing/data': 0.0005982009461149573, 'timing/model': 1.924235848011449, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 1.41}) +[2026-06-23 17:09:26,266][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.08999631553888321, 'train/grad_norm_pre_clip': 54.5671272277832, 'timing/data': 0.0006273409817367792, 'timing/model': 1.915912869037129, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 1.41}) +[2026-06-23 17:09:30,189][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.048571884632110596, 'train/grad_norm_pre_clip': 37.62995529174805, 'timing/data': 0.0005855349591001868, 'timing/model': 2.0625868160277605, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 1.41}) +[2026-06-23 17:09:33,981][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.08346015959978104, 'train/grad_norm_pre_clip': 36.68800735473633, 'timing/data': 0.0007322559831663966, 'timing/model': 1.928246292984113, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 1.41}) +[2026-06-23 17:09:37,777][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.005115828942507505, 'train/grad_norm_pre_clip': 35.91271209716797, 'timing/data': 0.0006925039924681187, 'timing/model': 1.9330555889755487, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 1.41}) +[2026-06-23 17:09:41,584][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.053914740681648254, 'train/grad_norm_pre_clip': 33.62504577636719, 'timing/data': 0.0005827839486300945, 'timing/model': 1.941409622086212, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 1.41}) +[2026-06-23 17:09:45,376][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.09548866748809814, 'train/grad_norm_pre_clip': 40.66743850708008, 'timing/data': 0.0005929080070927739, 'timing/model': 1.9275120850652456, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 1.41}) +[2026-06-23 17:09:49,171][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.052549779415130615, 'train/grad_norm_pre_clip': 34.402793884277344, 'timing/data': 0.0006014289101585746, 'timing/model': 1.9283377369865775, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 1.41}) +[2026-06-23 17:09:52,979][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.1056050956249237, 'train/grad_norm_pre_clip': 36.349456787109375, 'timing/data': 0.0008064219728112221, 'timing/model': 1.935370382037945, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 1.41}) +[2026-06-23 17:09:56,787][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.022984052076935768, 'train/grad_norm_pre_clip': 16.686647415161133, 'timing/data': 0.0007709809578955173, 'timing/model': 1.93451772199478, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 1.41}) +[2026-06-23 17:10:00,596][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.026692986488342285, 'train/grad_norm_pre_clip': 45.72234344482422, 'timing/data': 0.0005983070004731417, 'timing/model': 1.9356391029432416, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 1.41}) +[2026-06-23 17:10:04,544][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.025472216308116913, 'train/grad_norm_pre_clip': 17.028106689453125, 'timing/data': 0.000620742910541594, 'timing/model': 1.918658239999786, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 1.41}) +[2026-06-23 17:10:08,335][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.050219181925058365, 'train/grad_norm_pre_clip': 32.16454315185547, 'timing/data': 0.0006026859628036618, 'timing/model': 1.9266399540938437, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 1.41}) +[2026-06-23 17:10:12,136][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.07750432193279266, 'train/grad_norm_pre_clip': 49.3630485534668, 'timing/data': 0.0008105880115181208, 'timing/model': 1.9259197429055348, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 1.41}) +[2026-06-23 17:10:15,942][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.03978823870420456, 'train/grad_norm_pre_clip': 24.169116973876953, 'timing/data': 0.0005956640234217048, 'timing/model': 1.9304292539600283, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 1.41}) +[2026-06-23 17:10:19,749][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.040290914475917816, 'train/grad_norm_pre_clip': 27.28917694091797, 'timing/data': 0.0005941130220890045, 'timing/model': 1.9301971930544823, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 1.41}) +[2026-06-23 17:10:23,558][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.039929233491420746, 'train/grad_norm_pre_clip': 42.81138229370117, 'timing/data': 0.0006445000180974603, 'timing/model': 1.9312101359246299, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 1.41}) +[2026-06-23 17:10:27,359][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.01788276433944702, 'train/grad_norm_pre_clip': 12.379358291625977, 'timing/data': 0.0006161479977890849, 'timing/model': 1.9312254670076072, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 1.41}) +[2026-06-23 17:10:31,165][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.045374177396297455, 'train/grad_norm_pre_clip': 40.18798828125, 'timing/data': 0.0007950009312480688, 'timing/model': 1.938210477004759, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 1.41}) +[2026-06-23 17:10:35,044][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.13541145622730255, 'train/grad_norm_pre_clip': 47.01593017578125, 'timing/data': 0.0007365669589489698, 'timing/model': 2.006799565977417, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 1.41}) +[2026-06-23 17:10:38,838][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.07543935626745224, 'train/grad_norm_pre_clip': 29.22209358215332, 'timing/data': 0.000594638055190444, 'timing/model': 1.9307028390467167, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 1.41}) +[2026-06-23 17:10:42,645][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.10533103346824646, 'train/grad_norm_pre_clip': 27.10350799560547, 'timing/data': 0.0006112889386713505, 'timing/model': 1.9327830279944465, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 1.41}) +[2026-06-23 17:10:46,437][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.033238623291254044, 'train/grad_norm_pre_clip': 23.691577911376953, 'timing/data': 0.0006362460553646088, 'timing/model': 1.9297012980096042, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 1.41}) +[2026-06-23 17:10:50,238][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.10170556604862213, 'train/grad_norm_pre_clip': 26.005149841308594, 'timing/data': 0.000614551012404263, 'timing/model': 1.935298308962956, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 1.42}) +[2026-06-23 17:10:54,036][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.0551024004817009, 'train/grad_norm_pre_clip': 24.520315170288086, 'timing/data': 0.0006142558995634317, 'timing/model': 1.922360735014081, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 1.42}) +[2026-06-23 17:10:57,849][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.056422535330057144, 'train/grad_norm_pre_clip': 25.20050048828125, 'timing/data': 0.0006710391025990248, 'timing/model': 1.9345280079869553, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 1.42}) +[2026-06-23 17:11:01,665][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.02476728893816471, 'train/grad_norm_pre_clip': 22.43670654296875, 'timing/data': 0.0007973100291565061, 'timing/model': 1.932674856041558, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 1.42}) +[2026-06-23 17:11:05,503][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.057803064584732056, 'train/grad_norm_pre_clip': 30.00535011291504, 'timing/data': 0.0007511080475524068, 'timing/model': 1.9328368969727308, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 1.42}) +[2026-06-23 17:11:09,422][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.020681418478488922, 'train/grad_norm_pre_clip': 27.219209671020508, 'timing/data': 0.0006352609489113092, 'timing/model': 2.046812467975542, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 1.42}) +[2026-06-23 17:11:13,221][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.060065608471632004, 'train/grad_norm_pre_clip': 31.37112045288086, 'timing/data': 0.0006157559109851718, 'timing/model': 1.9192110900767148, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 1.42}) +[2026-06-23 17:11:17,025][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.039058834314346313, 'train/grad_norm_pre_clip': 35.954254150390625, 'timing/data': 0.0005914389621466398, 'timing/model': 1.9342773000244051, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 1.42}) +[2026-06-23 17:11:20,826][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.018284443765878677, 'train/grad_norm_pre_clip': 14.488021850585938, 'timing/data': 0.0005970650818198919, 'timing/model': 1.9218388419831172, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 1.42}) +[2026-06-23 17:11:24,631][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.09791643172502518, 'train/grad_norm_pre_clip': 65.89812469482422, 'timing/data': 0.0006138930330052972, 'timing/model': 1.9220264900941402, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 1.42}) +[2026-06-23 17:11:28,427][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.10450533032417297, 'train/grad_norm_pre_clip': 38.98566818237305, 'timing/data': 0.0007866459200158715, 'timing/model': 1.9212001849664375, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 1.42}) +[2026-06-23 17:11:32,216][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.041242145001888275, 'train/grad_norm_pre_clip': 28.709028244018555, 'timing/data': 0.0005955570377409458, 'timing/model': 1.9198332990054041, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 1.42}) +[2026-06-23 17:11:36,022][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.035642944276332855, 'train/grad_norm_pre_clip': 40.703651428222656, 'timing/data': 0.0006029210053384304, 'timing/model': 1.9300870669540018, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 1.42}) +[2026-06-23 17:11:39,847][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.04769739881157875, 'train/grad_norm_pre_clip': 26.10735321044922, 'timing/data': 0.000583493965677917, 'timing/model': 1.953102090046741, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 1.42}) +[2026-06-23 17:11:43,658][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.24708691239356995, 'train/grad_norm_pre_clip': 80.75745391845703, 'timing/data': 0.0006003829184919596, 'timing/model': 1.9381308970041573, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 1.42}) +[2026-06-23 17:11:47,467][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.029692010954022408, 'train/grad_norm_pre_clip': 29.344188690185547, 'timing/data': 0.0005952779902145267, 'timing/model': 1.9344486150657758, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 1.42}) +[2026-06-23 17:11:51,252][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.03134044259786606, 'train/grad_norm_pre_clip': 58.81083679199219, 'timing/data': 0.0006028030766174197, 'timing/model': 1.9197352540213615, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 1.42}) +[2026-06-23 17:11:55,055][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.0020061021205037832, 'train/grad_norm_pre_clip': 22.12990379333496, 'timing/data': 0.0006103520281612873, 'timing/model': 1.9293330759974197, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 1.42}) +[2026-06-23 17:11:58,858][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.03454780951142311, 'train/grad_norm_pre_clip': 31.64055824279785, 'timing/data': 0.0006058349972590804, 'timing/model': 1.9338034130632877, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 1.42}) +[2026-06-23 17:12:02,681][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.12311103194952011, 'train/grad_norm_pre_clip': 52.67570495605469, 'timing/data': 0.0005944370059296489, 'timing/model': 1.9385700720595196, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 1.42}) +[2026-06-23 17:12:06,479][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.07873831689357758, 'train/grad_norm_pre_clip': 34.996185302734375, 'timing/data': 0.000601277919486165, 'timing/model': 1.9257703949697316, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.42}) +[2026-06-23 17:12:10,284][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.018030961975455284, 'train/grad_norm_pre_clip': 41.769283294677734, 'timing/data': 0.0005858120275661349, 'timing/model': 1.9284049960551783, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 1.42}) +[2026-06-23 17:12:14,209][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.07732345908880234, 'train/grad_norm_pre_clip': 46.38447952270508, 'timing/data': 0.000605622073635459, 'timing/model': 2.045693692052737, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 1.42}) +[2026-06-23 17:12:18,000][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.07235459983348846, 'train/grad_norm_pre_clip': 50.98766326904297, 'timing/data': 0.0008080150000751019, 'timing/model': 1.930898140068166, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 1.42}) +[2026-06-23 17:12:21,813][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.030777255073189735, 'train/grad_norm_pre_clip': 22.051746368408203, 'timing/data': 0.0006499039009213448, 'timing/model': 1.9334049350582063, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 1.42}) +[2026-06-23 17:12:25,623][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.1247749924659729, 'train/grad_norm_pre_clip': 59.717857360839844, 'timing/data': 0.0006935809506103396, 'timing/model': 1.9313342850655317, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 1.43}) +[2026-06-23 17:12:29,435][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.01604018360376358, 'train/grad_norm_pre_clip': 13.977299690246582, 'timing/data': 0.0006396110402420163, 'timing/model': 1.9336342599708587, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 1.43}) +[2026-06-23 17:12:33,239][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.0666942447423935, 'train/grad_norm_pre_clip': 50.74098587036133, 'timing/data': 0.0006176400929689407, 'timing/model': 1.927854512934573, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 1.43}) +[2026-06-23 17:12:37,052][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.04745617136359215, 'train/grad_norm_pre_clip': 29.10530662536621, 'timing/data': 0.0008001349633559585, 'timing/model': 1.9293456140439957, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 1.43}) +[2026-06-23 17:12:40,860][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.05724818632006645, 'train/grad_norm_pre_clip': 32.25096893310547, 'timing/data': 0.0005985730094835162, 'timing/model': 1.9302719460101798, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 1.43}) +[2026-06-23 17:12:44,677][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.06580985337495804, 'train/grad_norm_pre_clip': 34.492645263671875, 'timing/data': 0.0006167599931359291, 'timing/model': 1.939947257982567, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 1.43}) +[2026-06-23 17:12:48,601][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.049637459218502045, 'train/grad_norm_pre_clip': 25.805559158325195, 'timing/data': 0.0005976499523967505, 'timing/model': 1.9232601809781045, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 1.43}) +[2026-06-23 17:12:52,414][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.06270917505025864, 'train/grad_norm_pre_clip': 21.948219299316406, 'timing/data': 0.0006316660437732935, 'timing/model': 1.935434459010139, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 1.43}) +[2026-06-23 17:12:56,216][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.03016485832631588, 'train/grad_norm_pre_clip': 28.46018409729004, 'timing/data': 0.0006042050663381815, 'timing/model': 1.9308553209993988, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 1.43}) +[2026-06-23 17:13:00,073][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.03498614579439163, 'train/grad_norm_pre_clip': 38.08402633666992, 'timing/data': 0.0006034009857103229, 'timing/model': 1.988442427944392, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 1.43}) +[2026-06-23 17:13:03,888][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.07670363038778305, 'train/grad_norm_pre_clip': 30.645532608032227, 'timing/data': 0.0006000569555908442, 'timing/model': 1.9360161870718002, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 1.43}) +[2026-06-23 17:13:07,703][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.05685310438275337, 'train/grad_norm_pre_clip': 32.856475830078125, 'timing/data': 0.0005886360304430127, 'timing/model': 1.940919958986342, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 1.43}) +[2026-06-23 17:13:11,517][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.027541445568203926, 'train/grad_norm_pre_clip': 45.877628326416016, 'timing/data': 0.0007619319949299097, 'timing/model': 1.9337929150788113, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 1.43}) +[2026-06-23 17:13:15,309][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.023031001910567284, 'train/grad_norm_pre_clip': 14.42727279663086, 'timing/data': 0.0005914589855819941, 'timing/model': 1.9276637430302799, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 1.43}) +[2026-06-23 17:13:19,163][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.12503711879253387, 'train/grad_norm_pre_clip': 41.41202926635742, 'timing/data': 0.0005907649174332619, 'timing/model': 1.986700837034732, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 1.43}) +[2026-06-23 17:13:22,966][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.058751318603754044, 'train/grad_norm_pre_clip': 31.301267623901367, 'timing/data': 0.0006424459861591458, 'timing/model': 1.933586506987922, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 1.43}) +[2026-06-23 17:13:26,777][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.030054805800318718, 'train/grad_norm_pre_clip': 23.244590759277344, 'timing/data': 0.000624554930254817, 'timing/model': 1.9268471020041034, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 1.43}) +[2026-06-23 17:13:30,576][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.14503620564937592, 'train/grad_norm_pre_clip': 53.88492202758789, 'timing/data': 0.0006217950722202659, 'timing/model': 1.928192131104879, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 1.43}) +[2026-06-23 17:13:34,390][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3700 +[2026-06-23 17:15:41,273][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3700 +[2026-06-23 17:15:41,275][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3700 +[2026-06-23 17:17:04,091][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3700 +[2026-06-23 17:17:04,092][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.039628610014915466, 'train/grad_norm_pre_clip': 18.14203643798828, 'eval/action_loss/seconds': 126.88121383497491, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.033110834658145905, 'eval/latency_2/loss': 0.033110834658145905, 'eval/flappy/loss': 0.033110834658145905, 'eval/flappy/latency_2/loss': 0.033110834658145905, 'eval/flap_precision': 0.9576158940397351, 'eval/flap_recall': 0.9757085020242915, 'eval/flap_f1': 0.9665775401069518, 'eval/flap_rate_model': 0.20972222222222223, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9665775401069517, 'eval/latency_2/flap_precision': 0.9576158940397351, 'eval/latency_2/flap_recall': 0.9757085020242915, 'eval/latency_2/flap_f1': 0.9665775401069518, 'eval/latency_2/flap_rate_model': 0.20972222222222223, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9665775401069517, 'eval/action_classification/data_wait_seconds': 1.7186400890350342, 'eval/action_classification/predict_seconds': 80.91050720214844, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006060119485482574, 'timing/model': 1.9253914970904589, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 1.43}) +[2026-06-23 17:17:07,970][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.0449998639523983, 'train/grad_norm_pre_clip': 19.85576629638672, 'timing/data': 0.0006235640030354261, 'timing/model': 2.01515584893059, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 1.43}) +[2026-06-23 17:17:11,762][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.06125251203775406, 'train/grad_norm_pre_clip': 47.52253341674805, 'timing/data': 0.0006068809889256954, 'timing/model': 1.931292745983228, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 1.43}) +[2026-06-23 17:17:15,538][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.054123569279909134, 'train/grad_norm_pre_clip': 32.152435302734375, 'timing/data': 0.0005945439916104078, 'timing/model': 1.924019725061953, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 1.43}) +[2026-06-23 17:17:19,321][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.022860495373606682, 'train/grad_norm_pre_clip': 45.57369613647461, 'timing/data': 0.0006161719793453813, 'timing/model': 1.9236628769431263, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 1.43}) +[2026-06-23 17:17:23,117][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.026151463389396667, 'train/grad_norm_pre_clip': 26.894088745117188, 'timing/data': 0.0007724040187895298, 'timing/model': 1.92567665502429, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 1.43}) +[2026-06-23 17:17:26,913][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.020300379022955894, 'train/grad_norm_pre_clip': 17.275148391723633, 'timing/data': 0.0007289680652320385, 'timing/model': 1.9242617200361565, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 1.43}) +[2026-06-23 17:17:30,690][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.11741192638874054, 'train/grad_norm_pre_clip': 21.545854568481445, 'timing/data': 0.0006613009609282017, 'timing/model': 1.9180783230112866, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 1.43}) +[2026-06-23 17:17:34,474][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.028439372777938843, 'train/grad_norm_pre_clip': 37.094886779785156, 'timing/data': 0.0008112889481708407, 'timing/model': 1.9102908089989796, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 1.44}) +[2026-06-23 17:17:38,265][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.022549733519554138, 'train/grad_norm_pre_clip': 16.242450714111328, 'timing/data': 0.0006049929652363062, 'timing/model': 1.931112164980732, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 1.44}) +[2026-06-23 17:17:42,115][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.053625866770744324, 'train/grad_norm_pre_clip': 28.4127254486084, 'timing/data': 0.0005852089961990714, 'timing/model': 1.9210234499769285, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 1.44}) +[2026-06-23 17:17:45,904][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.04102472960948944, 'train/grad_norm_pre_clip': 28.29463005065918, 'timing/data': 0.0005879809614270926, 'timing/model': 1.920358816976659, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 1.44}) +[2026-06-23 17:17:49,686][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.01646832562983036, 'train/grad_norm_pre_clip': 12.020380020141602, 'timing/data': 0.0007730560610070825, 'timing/model': 1.9194948649965227, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 1.44}) +[2026-06-23 17:17:53,481][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.044674377888441086, 'train/grad_norm_pre_clip': 16.48710823059082, 'timing/data': 0.0006286109564825892, 'timing/model': 1.9203263099770993, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 1.44}) +[2026-06-23 17:17:57,282][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.010764727368950844, 'train/grad_norm_pre_clip': 31.465715408325195, 'timing/data': 0.0006049089133739471, 'timing/model': 1.9274016850395128, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 1.44}) +[2026-06-23 17:18:01,084][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.07952481508255005, 'train/grad_norm_pre_clip': 42.57952117919922, 'timing/data': 0.0007570059970021248, 'timing/model': 1.9356395340291783, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 1.44}) +[2026-06-23 17:18:04,915][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.031133132055401802, 'train/grad_norm_pre_clip': 24.4736385345459, 'timing/data': 0.0006211729487404227, 'timing/model': 1.924012367031537, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 1.44}) +[2026-06-23 17:18:08,710][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.08050152659416199, 'train/grad_norm_pre_clip': 25.38640022277832, 'timing/data': 0.0007616999791935086, 'timing/model': 1.925392627948895, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 1.44}) +[2026-06-23 17:18:12,537][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.04033409804105759, 'train/grad_norm_pre_clip': 17.63019371032715, 'timing/data': 0.0005843669641762972, 'timing/model': 1.948983003036119, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 1.44}) +[2026-06-23 17:18:16,311][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.0343463160097599, 'train/grad_norm_pre_clip': 12.52131462097168, 'timing/data': 0.0006678259233012795, 'timing/model': 1.9120541700394824, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 1.44}) +[2026-06-23 17:18:20,096][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.04728773608803749, 'train/grad_norm_pre_clip': 20.009721755981445, 'timing/data': 0.0006475929403677583, 'timing/model': 1.9247717559337616, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 1.44}) +[2026-06-23 17:18:23,890][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.003207338973879814, 'train/grad_norm_pre_clip': 18.982912063598633, 'timing/data': 0.0006137860473245382, 'timing/model': 1.9287273130612448, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 1.44}) +[2026-06-23 17:18:27,674][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.16356980800628662, 'train/grad_norm_pre_clip': 55.58786392211914, 'timing/data': 0.0007856589509174228, 'timing/model': 1.9192175469361246, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 1.44}) +[2026-06-23 17:18:31,469][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.051573943346738815, 'train/grad_norm_pre_clip': 36.63715362548828, 'timing/data': 0.0007731260266155005, 'timing/model': 1.9339713560184464, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 1.44}) +[2026-06-23 17:18:35,264][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.019943615421652794, 'train/grad_norm_pre_clip': 13.856423377990723, 'timing/data': 0.0005981129361316562, 'timing/model': 1.9265321439597756, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 1.44}) +[2026-06-23 17:18:39,060][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.016962990164756775, 'train/grad_norm_pre_clip': 14.581890106201172, 'timing/data': 0.0006664419779554009, 'timing/model': 1.9373642479768023, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 1.44}) +[2026-06-23 17:18:42,863][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.014753605239093304, 'train/grad_norm_pre_clip': 15.797966003417969, 'timing/data': 0.0006583608919754624, 'timing/model': 1.9252988359658048, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 1.44}) +[2026-06-23 17:18:46,789][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.15481622517108917, 'train/grad_norm_pre_clip': 32.783531188964844, 'timing/data': 0.0005977960536256433, 'timing/model': 2.0430654699448496, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 1.44}) +[2026-06-23 17:18:50,596][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.04089305177330971, 'train/grad_norm_pre_clip': 35.304595947265625, 'timing/data': 0.0006136489100754261, 'timing/model': 1.929032227024436, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 1.44}) +[2026-06-23 17:18:54,380][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.02216164954006672, 'train/grad_norm_pre_clip': 19.148855209350586, 'timing/data': 0.0005934969522058964, 'timing/model': 1.9215689420234412, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 1.44}) +[2026-06-23 17:18:58,172][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.06970400363206863, 'train/grad_norm_pre_clip': 38.48844909667969, 'timing/data': 0.0006096509750932455, 'timing/model': 1.924678066978231, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 1.44}) +[2026-06-23 17:19:01,975][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.1853644847869873, 'train/grad_norm_pre_clip': 34.853179931640625, 'timing/data': 0.000590119045227766, 'timing/model': 1.9314001150196418, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 1.44}) +[2026-06-23 17:19:05,776][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.03407187759876251, 'train/grad_norm_pre_clip': 26.42481803894043, 'timing/data': 0.000589251983910799, 'timing/model': 1.9290556060150266, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 1.44}) +[2026-06-23 17:19:09,568][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.00366537319496274, 'train/grad_norm_pre_clip': 19.54743003845215, 'timing/data': 0.0006170179694890976, 'timing/model': 1.925553405075334, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 1.44}) +[2026-06-23 17:19:13,373][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.02235650271177292, 'train/grad_norm_pre_clip': 31.230897903442383, 'timing/data': 0.0006300089880824089, 'timing/model': 1.9371135339606553, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 1.45}) +[2026-06-23 17:19:17,200][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.06318492442369461, 'train/grad_norm_pre_clip': 37.37348937988281, 'timing/data': 0.0005911390762776136, 'timing/model': 1.9427440220024437, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 1.45}) +[2026-06-23 17:19:21,098][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.16181635856628418, 'train/grad_norm_pre_clip': 41.937644958496094, 'timing/data': 0.0006358351092785597, 'timing/model': 1.9209004760487005, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 1.45}) +[2026-06-23 17:19:24,917][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.024003786966204643, 'train/grad_norm_pre_clip': 32.556949615478516, 'timing/data': 0.0005855830386281013, 'timing/model': 1.9364292409736663, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 1.45}) +[2026-06-23 17:19:28,718][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.055634185671806335, 'train/grad_norm_pre_clip': 28.98508644104004, 'timing/data': 0.0005854989867657423, 'timing/model': 1.9220432110596448, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 1.45}) +[2026-06-23 17:19:32,517][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.13131316006183624, 'train/grad_norm_pre_clip': 69.91226959228516, 'timing/data': 0.0005952389910817146, 'timing/model': 1.9293337620329112, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 1.45}) +[2026-06-23 17:19:36,318][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.05553172901272774, 'train/grad_norm_pre_clip': 51.17270278930664, 'timing/data': 0.0007587820291519165, 'timing/model': 1.9358026310801506, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 1.45}) +[2026-06-23 17:19:40,110][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.07085639238357544, 'train/grad_norm_pre_clip': 38.018272399902344, 'timing/data': 0.0005911800544708967, 'timing/model': 1.9233863010304049, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 1.45}) +[2026-06-23 17:19:43,908][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.021416455507278442, 'train/grad_norm_pre_clip': 33.08594512939453, 'timing/data': 0.0007761439774185419, 'timing/model': 1.9310098929563537, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 1.45}) +[2026-06-23 17:19:47,719][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.04230092838406563, 'train/grad_norm_pre_clip': 41.50952911376953, 'timing/data': 0.0005844340194016695, 'timing/model': 1.9376608970342204, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 1.45}) +[2026-06-23 17:19:51,582][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.022328438237309456, 'train/grad_norm_pre_clip': 18.977447509765625, 'timing/data': 0.0006588100222870708, 'timing/model': 1.9886245150119066, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 1.45}) +[2026-06-23 17:19:55,386][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.0053634424693882465, 'train/grad_norm_pre_clip': 22.728534698486328, 'timing/data': 0.0005930629558861256, 'timing/model': 1.9215638659661636, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 1.45}) +[2026-06-23 17:19:59,184][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.064033642411232, 'train/grad_norm_pre_clip': 42.952266693115234, 'timing/data': 0.0005923369899392128, 'timing/model': 1.9226565300486982, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 1.45}) +[2026-06-23 17:20:02,991][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.0313752219080925, 'train/grad_norm_pre_clip': 20.63749885559082, 'timing/data': 0.0006009920034557581, 'timing/model': 1.9269365470390767, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 1.45}) +[2026-06-23 17:20:06,803][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.11408829689025879, 'train/grad_norm_pre_clip': 21.032638549804688, 'timing/data': 0.000602042069658637, 'timing/model': 1.9327567289583385, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 1.45}) +[2026-06-23 17:20:10,606][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.006861664354801178, 'train/grad_norm_pre_clip': 21.362829208374023, 'timing/data': 0.0008134660311043262, 'timing/model': 1.936034303973429, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 1.45}) +[2026-06-23 17:23:08,966][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3750 +[2026-06-23 17:23:08,967][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.00762413302436471, 'train/grad_norm_pre_clip': 11.75856876373291, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 105.37999999999798, 'rl_games_eval/mid_train/mean_length': 868.6, 'rl_games_eval/mid_train/std_reward': 77.50794539916251, 'rl_games_eval/mid_train/std_length': 626.177802225534, 'rl_games_eval/mid_train/macro_mean_reward': 105.37999999999798, 'rl_games_eval/mid_train/macro_mean_length': 868.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 105.37999999999798, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 868.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 77.50794539916251, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 626.177802225534, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 105.37999999999798, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006288380827754736, 'timing/model': 1.9233786689583212, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 1.45}) +[2026-06-23 17:23:12,707][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.00478270323947072, 'train/grad_norm_pre_clip': 17.510454177856445, 'timing/data': 0.0006086969515308738, 'timing/model': 1.896388100925833, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 1.45}) +[2026-06-23 17:23:16,447][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.009336031042039394, 'train/grad_norm_pre_clip': 41.170509338378906, 'timing/data': 0.0007409550016745925, 'timing/model': 1.8971121039940044, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 1.45}) +[2026-06-23 17:23:20,183][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.17564910650253296, 'train/grad_norm_pre_clip': 55.632774353027344, 'timing/data': 0.0006417209515348077, 'timing/model': 1.8963473410112783, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 1.45}) +[2026-06-23 17:23:23,935][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.03960169106721878, 'train/grad_norm_pre_clip': 40.3191032409668, 'timing/data': 0.0005973170045763254, 'timing/model': 1.9021700579905882, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 1.45}) +[2026-06-23 17:23:27,703][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.053322289139032364, 'train/grad_norm_pre_clip': 15.50908088684082, 'timing/data': 0.0006056840065866709, 'timing/model': 1.9026647770078853, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 1.45}) +[2026-06-23 17:23:31,454][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.04303336888551712, 'train/grad_norm_pre_clip': 26.501188278198242, 'timing/data': 0.000616419012658298, 'timing/model': 1.9052908610319719, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 1.45}) +[2026-06-23 17:23:35,221][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.0737355649471283, 'train/grad_norm_pre_clip': 52.56471252441406, 'timing/data': 0.0008209589868783951, 'timing/model': 1.918613613001071, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 1.45}) +[2026-06-23 17:23:38,994][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.05785663425922394, 'train/grad_norm_pre_clip': 29.600383758544922, 'timing/data': 0.0006095650605857372, 'timing/model': 1.909567133989185, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 1.45}) +[2026-06-23 17:23:42,780][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.09056755155324936, 'train/grad_norm_pre_clip': 68.57640838623047, 'timing/data': 0.000601314939558506, 'timing/model': 1.9245757620083168, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 1.45}) +[2026-06-23 17:23:46,544][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.02973136492073536, 'train/grad_norm_pre_clip': 40.319271087646484, 'timing/data': 0.0005881290417164564, 'timing/model': 1.9098366290563717, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 1.46}) +[2026-06-23 17:23:50,321][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.01965606026351452, 'train/grad_norm_pre_clip': 33.84976577758789, 'timing/data': 0.0005884530255571008, 'timing/model': 1.9255048219347373, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 1.46}) +[2026-06-23 17:23:54,098][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.09924609959125519, 'train/grad_norm_pre_clip': 41.89276885986328, 'timing/data': 0.0005976320244371891, 'timing/model': 1.9179340130649507, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 1.46}) +[2026-06-23 17:23:57,879][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.026566926389932632, 'train/grad_norm_pre_clip': 14.027643203735352, 'timing/data': 0.0006127239903435111, 'timing/model': 1.9191560019971803, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 1.46}) +[2026-06-23 17:24:01,648][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.011635156348347664, 'train/grad_norm_pre_clip': 12.37760066986084, 'timing/data': 0.000588497961871326, 'timing/model': 1.9146179259987548, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 1.46}) +[2026-06-23 17:24:05,430][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.18036989867687225, 'train/grad_norm_pre_clip': 31.632984161376953, 'timing/data': 0.0006089020753279328, 'timing/model': 1.9193139770068228, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 1.46}) +[2026-06-23 17:24:09,220][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.14438609778881073, 'train/grad_norm_pre_clip': 63.53482437133789, 'timing/data': 0.0005833830218762159, 'timing/model': 1.9227508041076362, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 1.46}) +[2026-06-23 17:24:12,994][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.022098002955317497, 'train/grad_norm_pre_clip': 21.326107025146484, 'timing/data': 0.000844266964122653, 'timing/model': 1.9071206579683349, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 1.46}) +[2026-06-23 17:24:16,781][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.055940788239240646, 'train/grad_norm_pre_clip': 35.296321868896484, 'timing/data': 0.000675198039971292, 'timing/model': 1.9327761359745637, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 1.46}) +[2026-06-23 17:24:20,573][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.009111586958169937, 'train/grad_norm_pre_clip': 30.154687881469727, 'timing/data': 0.000613307929597795, 'timing/model': 1.9245714149437845, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 1.46}) +[2026-06-23 17:24:24,365][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.06863866001367569, 'train/grad_norm_pre_clip': 36.02020263671875, 'timing/data': 0.000772185972891748, 'timing/model': 1.9183842180063948, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 1.46}) +[2026-06-23 17:24:28,148][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.026995187625288963, 'train/grad_norm_pre_clip': 21.577306747436523, 'timing/data': 0.0006888100178912282, 'timing/model': 1.9223819669568911, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 1.46}) +[2026-06-23 17:24:31,936][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.025234457105398178, 'train/grad_norm_pre_clip': 12.526664733886719, 'timing/data': 0.0005994229577481747, 'timing/model': 1.9299797209678218, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 1.46}) +[2026-06-23 17:24:35,739][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.018547523766756058, 'train/grad_norm_pre_clip': 11.652374267578125, 'timing/data': 0.0006003790767863393, 'timing/model': 1.9375134949805215, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 1.46}) +[2026-06-23 17:24:39,704][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.027361920103430748, 'train/grad_norm_pre_clip': 19.508317947387695, 'timing/data': 0.0005916999652981758, 'timing/model': 2.0928441929863766, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 1.46}) +[2026-06-23 17:24:43,482][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.015534496866166592, 'train/grad_norm_pre_clip': 9.056746482849121, 'timing/data': 0.00080196897033602, 'timing/model': 1.9223263369640335, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 1.46}) +[2026-06-23 17:24:47,264][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.016942709684371948, 'train/grad_norm_pre_clip': 16.54428482055664, 'timing/data': 0.0006135549629107118, 'timing/model': 1.9184708499815315, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 1.46}) +[2026-06-23 17:24:51,064][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.21965207159519196, 'train/grad_norm_pre_clip': 42.01716232299805, 'timing/data': 0.0005831259768456221, 'timing/model': 1.928576659061946, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 1.46}) +[2026-06-23 17:24:54,855][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.05827364698052406, 'train/grad_norm_pre_clip': 20.56244659423828, 'timing/data': 0.0008160190191119909, 'timing/model': 1.917867159936577, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 1.46}) +[2026-06-23 17:24:58,645][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.06373704224824905, 'train/grad_norm_pre_clip': 39.59355545043945, 'timing/data': 0.000599999912083149, 'timing/model': 1.9256669050082564, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 1.46}) +[2026-06-23 17:25:02,428][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.02951127663254738, 'train/grad_norm_pre_clip': 54.50627517700195, 'timing/data': 0.0006602409994229674, 'timing/model': 1.924684987985529, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 1.46}) +[2026-06-23 17:25:06,221][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.03736522048711777, 'train/grad_norm_pre_clip': 23.327655792236328, 'timing/data': 0.000599607010371983, 'timing/model': 1.9212554100668058, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 1.46}) +[2026-06-23 17:25:10,021][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.08202257752418518, 'train/grad_norm_pre_clip': 45.633331298828125, 'timing/data': 0.000618050922639668, 'timing/model': 1.915209232014604, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 1.46}) +[2026-06-23 17:25:13,874][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.051170844584703445, 'train/grad_norm_pre_clip': 24.20387077331543, 'timing/data': 0.0006096300203353167, 'timing/model': 1.9356105460319668, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 1.46}) +[2026-06-23 17:25:17,672][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.11213735491037369, 'train/grad_norm_pre_clip': 58.71426010131836, 'timing/data': 0.0006045140326023102, 'timing/model': 1.9301083290483803, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 1.46}) +[2026-06-23 17:25:21,468][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.01902361959218979, 'train/grad_norm_pre_clip': 11.919892311096191, 'timing/data': 0.0005997670814394951, 'timing/model': 1.9176404650788754, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 1.46}) +[2026-06-23 17:25:25,257][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.04839257150888443, 'train/grad_norm_pre_clip': 41.381988525390625, 'timing/data': 0.000770026003010571, 'timing/model': 1.923956427956, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 1.47}) +[2026-06-23 17:25:29,046][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.039289165288209915, 'train/grad_norm_pre_clip': 24.659465789794922, 'timing/data': 0.0008087119786068797, 'timing/model': 1.923816432012245, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 1.47}) +[2026-06-23 17:25:32,827][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.004207493271678686, 'train/grad_norm_pre_clip': 12.624129295349121, 'timing/data': 0.0005967170000076294, 'timing/model': 1.9225556920282543, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 1.47}) +[2026-06-23 17:25:36,630][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.06382248550653458, 'train/grad_norm_pre_clip': 26.336143493652344, 'timing/data': 0.000586853944696486, 'timing/model': 1.933133395970799, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 1.47}) +[2026-06-23 17:25:40,438][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.050117023289203644, 'train/grad_norm_pre_clip': 57.574405670166016, 'timing/data': 0.0006541849579662085, 'timing/model': 1.9293602890102193, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 1.47}) +[2026-06-23 17:25:44,240][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.06086987257003784, 'train/grad_norm_pre_clip': 22.32830047607422, 'timing/data': 0.0009186669485643506, 'timing/model': 1.9349461289821193, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 1.47}) +[2026-06-23 17:25:48,219][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.03641991689801216, 'train/grad_norm_pre_clip': 56.66785430908203, 'timing/data': 0.0006648529088124633, 'timing/model': 1.9259327800245956, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 1.47}) +[2026-06-23 17:25:52,025][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.10100451856851578, 'train/grad_norm_pre_clip': 29.88801383972168, 'timing/data': 0.0006201909855008125, 'timing/model': 1.937791311996989, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 1.47}) +[2026-06-23 17:25:55,831][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.06209644302725792, 'train/grad_norm_pre_clip': 19.960433959960938, 'timing/data': 0.000806802068836987, 'timing/model': 1.9400187800638378, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 1.47}) +[2026-06-23 17:25:59,636][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.09016627073287964, 'train/grad_norm_pre_clip': 48.51923370361328, 'timing/data': 0.0007908160332590342, 'timing/model': 1.9202620820142329, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 1.47}) +[2026-06-23 17:26:03,441][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.12134765833616257, 'train/grad_norm_pre_clip': 41.44775390625, 'timing/data': 0.000588804017752409, 'timing/model': 1.934296483057551, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 1.47}) +[2026-06-23 17:26:07,250][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.06671500205993652, 'train/grad_norm_pre_clip': 51.49629211425781, 'timing/data': 0.0006034909747540951, 'timing/model': 1.9298241139622405, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 1.47}) +[2026-06-23 17:26:11,047][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.06177143752574921, 'train/grad_norm_pre_clip': 26.654193878173828, 'timing/data': 0.0006024329923093319, 'timing/model': 1.9283355289371684, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 1.47}) +[2026-06-23 17:26:14,844][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.029707541689276695, 'train/grad_norm_pre_clip': 19.309877395629883, 'timing/data': 0.0005909749306738377, 'timing/model': 1.919361719978042, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 1.47}) +[2026-06-23 17:26:18,690][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3800 +[2026-06-23 17:28:27,246][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3800 +[2026-06-23 17:28:27,248][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3800 +[2026-06-23 17:29:49,839][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3800 +[2026-06-23 17:29:49,840][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.03034239634871483, 'train/grad_norm_pre_clip': 40.23782730102539, 'eval/action_loss/seconds': 128.55384454398882, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03169264271855354, 'eval/latency_2/loss': 0.03169264271855354, 'eval/flappy/loss': 0.03169264271855354, 'eval/flappy/latency_2/loss': 0.03169264271855354, 'eval/flap_precision': 0.9682217714672076, 'eval/flap_recall': 0.9662618083670715, 'eval/flap_f1': 0.9672407970280311, 'eval/flap_rate_model': 0.20541666666666666, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9672407970280311, 'eval/latency_2/flap_precision': 0.9682217714672076, 'eval/latency_2/flap_recall': 0.9662618083670715, 'eval/latency_2/flap_f1': 0.9672407970280311, 'eval/latency_2/flap_rate_model': 0.20541666666666666, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9672407970280311, 'eval/action_classification/data_wait_seconds': 1.4971150159835815, 'eval/action_classification/predict_seconds': 80.90879821777344, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0008499089162796736, 'timing/model': 1.9775672849500552, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 1.47}) +[2026-06-23 17:29:53,606][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.04404719918966293, 'train/grad_norm_pre_clip': 20.42049789428711, 'timing/data': 0.000680247088894248, 'timing/model': 1.9094175660284236, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 1.47}) +[2026-06-23 17:29:57,380][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.02988019771873951, 'train/grad_norm_pre_clip': 27.5676212310791, 'timing/data': 0.0006169280968606472, 'timing/model': 1.9177152049960569, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 1.47}) +[2026-06-23 17:30:01,169][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.031185874715447426, 'train/grad_norm_pre_clip': 29.198978424072266, 'timing/data': 0.0007844370556995273, 'timing/model': 1.9251642830204219, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 1.47}) +[2026-06-23 17:30:04,955][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.049047235399484634, 'train/grad_norm_pre_clip': 26.79136085510254, 'timing/data': 0.0006086979992687702, 'timing/model': 1.9143804659834132, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 1.47}) +[2026-06-23 17:30:08,867][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.011808827519416809, 'train/grad_norm_pre_clip': 19.334144592285156, 'timing/data': 0.0006607209797948599, 'timing/model': 2.03525754599832, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 1.47}) +[2026-06-23 17:30:12,650][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.09234979748725891, 'train/grad_norm_pre_clip': 36.58538055419922, 'timing/data': 0.0005950569175183773, 'timing/model': 1.9201650869799778, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 1.47}) +[2026-06-23 17:30:16,428][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.010900676250457764, 'train/grad_norm_pre_clip': 10.910144805908203, 'timing/data': 0.0005922060227021575, 'timing/model': 1.9212448870530352, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 1.47}) +[2026-06-23 17:30:20,210][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.014915968291461468, 'train/grad_norm_pre_clip': 29.718881607055664, 'timing/data': 0.0008062609704211354, 'timing/model': 1.924090729909949, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 1.47}) +[2026-06-23 17:30:24,002][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.14068469405174255, 'train/grad_norm_pre_clip': 24.88335609436035, 'timing/data': 0.0005870889872312546, 'timing/model': 1.9225361649878323, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 1.47}) +[2026-06-23 17:30:27,781][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.0026735281571745872, 'train/grad_norm_pre_clip': 29.411104202270508, 'timing/data': 0.0006114370189607143, 'timing/model': 1.922715807100758, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 1.47}) +[2026-06-23 17:30:31,572][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.11318403482437134, 'train/grad_norm_pre_clip': 30.170743942260742, 'timing/data': 0.0005957060493528843, 'timing/model': 1.925171838956885, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 1.47}) +[2026-06-23 17:30:35,361][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.06097939610481262, 'train/grad_norm_pre_clip': 31.67558479309082, 'timing/data': 0.0005919970571994781, 'timing/model': 1.9140628339955583, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 1.48}) +[2026-06-23 17:30:39,149][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.024045540019869804, 'train/grad_norm_pre_clip': 22.721603393554688, 'timing/data': 0.0007863619830459356, 'timing/model': 1.9210375719703734, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 1.48}) +[2026-06-23 17:30:43,053][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.06075682491064072, 'train/grad_norm_pre_clip': 36.68400192260742, 'timing/data': 0.0005812320159748197, 'timing/model': 1.9186600770335644, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 1.48}) +[2026-06-23 17:30:46,855][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.05741054192185402, 'train/grad_norm_pre_clip': 32.33453369140625, 'timing/data': 0.0006154809379950166, 'timing/model': 1.9305544180097058, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 1.48}) +[2026-06-23 17:30:50,665][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.041809678077697754, 'train/grad_norm_pre_clip': 23.09183120727539, 'timing/data': 0.000593362026847899, 'timing/model': 1.9225179959321395, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 1.48}) +[2026-06-23 17:30:54,452][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.08666842430830002, 'train/grad_norm_pre_clip': 35.16922378540039, 'timing/data': 0.0005880369571968913, 'timing/model': 1.9271939140744507, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 1.48}) +[2026-06-23 17:30:58,249][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.03608227148652077, 'train/grad_norm_pre_clip': 40.74695587158203, 'timing/data': 0.0005897609516978264, 'timing/model': 1.9358053259784356, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 1.48}) +[2026-06-23 17:31:02,029][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.07172400504350662, 'train/grad_norm_pre_clip': 31.3389892578125, 'timing/data': 0.0006488290382549167, 'timing/model': 1.914027068996802, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 1.48}) +[2026-06-23 17:31:05,828][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.046434640884399414, 'train/grad_norm_pre_clip': 30.963701248168945, 'timing/data': 0.0005902870325371623, 'timing/model': 1.9253036639420316, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 1.48}) +[2026-06-23 17:31:09,611][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.028135253116488457, 'train/grad_norm_pre_clip': 31.665912628173828, 'timing/data': 0.0005861259996891022, 'timing/model': 1.918417381006293, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 1.48}) +[2026-06-23 17:31:13,454][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.040450721979141235, 'train/grad_norm_pre_clip': 30.610301971435547, 'timing/data': 0.0005945760058239102, 'timing/model': 1.983795019914396, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 1.48}) +[2026-06-23 17:31:17,244][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.030531972646713257, 'train/grad_norm_pre_clip': 23.24268341064453, 'timing/data': 0.000585590023547411, 'timing/model': 1.9277600429486483, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 1.48}) +[2026-06-23 17:31:21,036][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.0290119256824255, 'train/grad_norm_pre_clip': 19.485977172851562, 'timing/data': 0.0005868760636076331, 'timing/model': 1.9319318139459938, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 1.48}) +[2026-06-23 17:31:24,828][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.024300385266542435, 'train/grad_norm_pre_clip': 19.18243408203125, 'timing/data': 0.0005732550052925944, 'timing/model': 1.9290928900009021, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 1.48}) +[2026-06-23 17:31:28,622][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.07516631484031677, 'train/grad_norm_pre_clip': 25.0801944732666, 'timing/data': 0.0006806509336456656, 'timing/model': 1.9239462040131912, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 1.48}) +[2026-06-23 17:31:32,414][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.008402111008763313, 'train/grad_norm_pre_clip': 18.610782623291016, 'timing/data': 0.0006512219551950693, 'timing/model': 1.9175902650458738, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 1.48}) +[2026-06-23 17:31:36,209][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.062280312180519104, 'train/grad_norm_pre_clip': 37.39479064941406, 'timing/data': 0.0006667469860985875, 'timing/model': 1.9245292189298198, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 1.48}) +[2026-06-23 17:31:40,006][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.010171745903789997, 'train/grad_norm_pre_clip': 19.92093849182129, 'timing/data': 0.000590365962125361, 'timing/model': 1.9234426959883422, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 1.48}) +[2026-06-23 17:31:43,828][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.02311730571091175, 'train/grad_norm_pre_clip': 9.646302223205566, 'timing/data': 0.0006006530020385981, 'timing/model': 1.9330427129752934, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 1.48}) +[2026-06-23 17:31:47,722][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.04920098930597305, 'train/grad_norm_pre_clip': 16.925329208374023, 'timing/data': 0.0005990089848637581, 'timing/model': 2.030694643035531, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 1.48}) +[2026-06-23 17:31:51,518][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.03859357535839081, 'train/grad_norm_pre_clip': 18.171411514282227, 'timing/data': 0.0005975329549983144, 'timing/model': 1.9223730700323358, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 1.48}) +[2026-06-23 17:31:55,314][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.015481437556445599, 'train/grad_norm_pre_clip': 16.26264190673828, 'timing/data': 0.000614435994066298, 'timing/model': 1.9203715519979596, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 1.48}) +[2026-06-23 17:31:59,112][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.0285731703042984, 'train/grad_norm_pre_clip': 28.69179344177246, 'timing/data': 0.0008254580898210406, 'timing/model': 1.929244937025942, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 1.48}) +[2026-06-23 17:32:02,908][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.026229459792375565, 'train/grad_norm_pre_clip': 22.614803314208984, 'timing/data': 0.0008129320340231061, 'timing/model': 1.9304175130091608, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 1.48}) +[2026-06-23 17:32:06,705][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.07505467534065247, 'train/grad_norm_pre_clip': 44.754512786865234, 'timing/data': 0.0008507519960403442, 'timing/model': 1.9250097749754786, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 1.48}) +[2026-06-23 17:32:10,505][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.010675795376300812, 'train/grad_norm_pre_clip': 41.947662353515625, 'timing/data': 0.0006990520050749183, 'timing/model': 1.9262724400032312, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 1.48}) +[2026-06-23 17:32:14,312][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.012360936030745506, 'train/grad_norm_pre_clip': 41.742427825927734, 'timing/data': 0.0007930700667202473, 'timing/model': 1.9361159520922229, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 1.49}) +[2026-06-23 17:32:18,112][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.09817562997341156, 'train/grad_norm_pre_clip': 44.04193115234375, 'timing/data': 0.000597091973759234, 'timing/model': 1.9358411940047517, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 1.49}) +[2026-06-23 17:32:22,023][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.11777373403310776, 'train/grad_norm_pre_clip': 29.890810012817383, 'timing/data': 0.0006280549569055438, 'timing/model': 2.04330469109118, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 1.49}) +[2026-06-23 17:32:25,803][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.014649920165538788, 'train/grad_norm_pre_clip': 12.74565601348877, 'timing/data': 0.0006053559482097626, 'timing/model': 1.9173493289854378, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 1.49}) +[2026-06-23 17:32:29,600][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.06923840939998627, 'train/grad_norm_pre_clip': 26.050493240356445, 'timing/data': 0.0005952769424766302, 'timing/model': 1.9221043740399182, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 1.49}) +[2026-06-23 17:32:33,428][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.008897785097360611, 'train/grad_norm_pre_clip': 13.529284477233887, 'timing/data': 0.0006337560480460525, 'timing/model': 1.9373701509321108, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 1.49}) +[2026-06-23 17:32:37,237][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.03024781309068203, 'train/grad_norm_pre_clip': 37.537025451660156, 'timing/data': 0.0005949999904260039, 'timing/model': 1.9286615928867832, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 1.49}) +[2026-06-23 17:32:41,048][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.022869234904646873, 'train/grad_norm_pre_clip': 14.622465133666992, 'timing/data': 0.0005782230291515589, 'timing/model': 1.9399633790599182, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 1.49}) +[2026-06-23 17:32:44,844][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.06479789316654205, 'train/grad_norm_pre_clip': 32.91659164428711, 'timing/data': 0.0006161259952932596, 'timing/model': 1.9318656749092042, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 1.49}) +[2026-06-23 17:32:48,653][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.16121377050876617, 'train/grad_norm_pre_clip': 25.415964126586914, 'timing/data': 0.0007064810488373041, 'timing/model': 1.936339073930867, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 1.49}) +[2026-06-23 17:32:52,453][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.08089487999677658, 'train/grad_norm_pre_clip': 49.64524459838867, 'timing/data': 0.0005961140850558877, 'timing/model': 1.9326541849877685, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 1.49}) +[2026-06-23 17:32:56,416][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.03376425430178642, 'train/grad_norm_pre_clip': 30.34602165222168, 'timing/data': 0.000617834972217679, 'timing/model': 1.9248320859624073, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 1.49}) +[2026-06-23 17:33:00,209][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.08158412575721741, 'train/grad_norm_pre_clip': 41.54273986816406, 'timing/data': 0.0005847159773111343, 'timing/model': 1.924244262976572, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 1.49}) +[2026-06-23 17:33:04,018][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.03403472527861595, 'train/grad_norm_pre_clip': 24.663406372070312, 'timing/data': 0.0005916469963267446, 'timing/model': 1.9330012709833682, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 1.49}) +[2026-06-23 17:33:07,825][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.027096297591924667, 'train/grad_norm_pre_clip': 23.287153244018555, 'timing/data': 0.0008008800214156508, 'timing/model': 1.923288284917362, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 1.49}) +[2026-06-23 17:33:11,632][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.0474521704018116, 'train/grad_norm_pre_clip': 29.905994415283203, 'timing/data': 0.0005881949327886105, 'timing/model': 1.929945460986346, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 1.49}) +[2026-06-23 17:33:15,436][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.04869377613067627, 'train/grad_norm_pre_clip': 30.17816734313965, 'timing/data': 0.0006284519331529737, 'timing/model': 1.930459970026277, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 1.49}) +[2026-06-23 17:33:19,251][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.09444429725408554, 'train/grad_norm_pre_clip': 61.66386795043945, 'timing/data': 0.0006946770008653402, 'timing/model': 1.9467472200049087, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 1.49}) +[2026-06-23 17:33:23,055][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.028073135763406754, 'train/grad_norm_pre_clip': 40.77113723754883, 'timing/data': 0.0006131839472800493, 'timing/model': 1.9325797480996698, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 1.49}) +[2026-06-23 17:33:26,992][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.016404811292886734, 'train/grad_norm_pre_clip': 23.807605743408203, 'timing/data': 0.0006089990492910147, 'timing/model': 2.0688612739322707, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 1.49}) +[2026-06-23 17:33:30,776][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.04479750245809555, 'train/grad_norm_pre_clip': 32.393611907958984, 'timing/data': 0.0005916539812460542, 'timing/model': 1.9204944530501962, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 1.49}) +[2026-06-23 17:33:34,582][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.1071905866265297, 'train/grad_norm_pre_clip': 38.939430236816406, 'timing/data': 0.0005899760872125626, 'timing/model': 1.9248184710741043, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 1.49}) +[2026-06-23 17:33:38,369][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.04380800202488899, 'train/grad_norm_pre_clip': 29.588544845581055, 'timing/data': 0.0005903829587623477, 'timing/model': 1.9259307159809396, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 1.49}) +[2026-06-23 17:33:42,169][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.04963160306215286, 'train/grad_norm_pre_clip': 41.48082733154297, 'timing/data': 0.0005887039005756378, 'timing/model': 1.926388676976785, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 1.49}) +[2026-06-23 17:33:45,974][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.031576137989759445, 'train/grad_norm_pre_clip': 28.42202377319336, 'timing/data': 0.0007999080698937178, 'timing/model': 1.9290641769766808, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 1.49}) +[2026-06-23 17:33:49,772][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.010756168514490128, 'train/grad_norm_pre_clip': 20.744617462158203, 'timing/data': 0.00060235895216465, 'timing/model': 1.926600944949314, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 1.5}) +[2026-06-23 17:33:53,565][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.01035039871931076, 'train/grad_norm_pre_clip': 48.85469436645508, 'timing/data': 0.0005885979626327753, 'timing/model': 1.9292396360542625, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 1.5}) +[2026-06-23 17:33:57,364][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.04374344274401665, 'train/grad_norm_pre_clip': 19.287357330322266, 'timing/data': 0.0006115080323070288, 'timing/model': 1.9198576590279117, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 1.5}) +[2026-06-23 17:34:01,308][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.13177958130836487, 'train/grad_norm_pre_clip': 43.82836151123047, 'timing/data': 0.0005996740655973554, 'timing/model': 1.9330924450187013, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 1.5}) +[2026-06-23 17:34:05,110][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.06351571530103683, 'train/grad_norm_pre_clip': 38.538291931152344, 'timing/data': 0.0006017750129103661, 'timing/model': 1.9264646100346, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 1.5}) +[2026-06-23 17:34:08,912][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.09127169847488403, 'train/grad_norm_pre_clip': 42.653724670410156, 'timing/data': 0.0006140460027381778, 'timing/model': 1.9304779670201242, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 1.5}) +[2026-06-23 17:34:12,711][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.03577052801847458, 'train/grad_norm_pre_clip': 26.035123825073242, 'timing/data': 0.0006621350767090917, 'timing/model': 1.9236409589648247, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 1.5}) +[2026-06-23 17:34:16,504][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.02458510920405388, 'train/grad_norm_pre_clip': 30.08075714111328, 'timing/data': 0.0008138010744005442, 'timing/model': 1.9204122070223093, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 1.5}) +[2026-06-23 17:34:20,315][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.10657453536987305, 'train/grad_norm_pre_clip': 35.4587516784668, 'timing/data': 0.0005842200480401516, 'timing/model': 1.9290594490012154, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 1.5}) +[2026-06-23 17:34:24,173][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.009874124079942703, 'train/grad_norm_pre_clip': 25.647109985351562, 'timing/data': 0.0005902369739487767, 'timing/model': 1.9845856940373778, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 1.5}) +[2026-06-23 17:34:27,969][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.12352918833494186, 'train/grad_norm_pre_clip': 54.03316879272461, 'timing/data': 0.0006081740139052272, 'timing/model': 1.9273071490461007, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 1.5}) +[2026-06-23 17:34:31,848][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.02453913353383541, 'train/grad_norm_pre_clip': 30.533931732177734, 'timing/data': 0.0006196110043674707, 'timing/model': 2.0117694559739903, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 1.5}) +[2026-06-23 17:34:35,657][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.12192829698324203, 'train/grad_norm_pre_clip': 40.88142013549805, 'timing/data': 0.00061551493126899, 'timing/model': 1.9361866089748219, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 1.5}) +[2026-06-23 17:34:39,465][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.046483322978019714, 'train/grad_norm_pre_clip': 19.354433059692383, 'timing/data': 0.0005939259426668286, 'timing/model': 1.9368976670084521, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 1.5}) +[2026-06-23 17:34:43,264][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.06076285243034363, 'train/grad_norm_pre_clip': 31.606836318969727, 'timing/data': 0.0006787079619243741, 'timing/model': 1.925338158966042, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 1.5}) +[2026-06-23 17:34:47,075][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.0352771021425724, 'train/grad_norm_pre_clip': 20.52776527404785, 'timing/data': 0.00058742705732584, 'timing/model': 1.9424595140153542, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 1.5}) +[2026-06-23 17:34:50,860][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.047956764698028564, 'train/grad_norm_pre_clip': 52.28404235839844, 'timing/data': 0.0007288630586117506, 'timing/model': 1.9194149980321527, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 1.5}) +[2026-06-23 17:34:54,685][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.03919929265975952, 'train/grad_norm_pre_clip': 18.75272560119629, 'timing/data': 0.0007245179731398821, 'timing/model': 1.9382511740550399, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 1.5}) +[2026-06-23 17:34:58,496][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.020049721002578735, 'train/grad_norm_pre_clip': 36.94575500488281, 'timing/data': 0.0006058589788153768, 'timing/model': 1.93401463504415, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 1.5}) +[2026-06-23 17:35:02,298][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.1075909435749054, 'train/grad_norm_pre_clip': 51.96931838989258, 'timing/data': 0.0008235829882323742, 'timing/model': 1.9379489360144362, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 1.5}) +[2026-06-23 17:35:06,162][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.09395943582057953, 'train/grad_norm_pre_clip': 33.85894775390625, 'timing/data': 0.0006111129187047482, 'timing/model': 1.933272365015, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 1.5}) +[2026-06-23 17:35:09,971][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.009143627248704433, 'train/grad_norm_pre_clip': 7.845399856567383, 'timing/data': 0.0006279239896684885, 'timing/model': 1.934679076075554, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 1.5}) +[2026-06-23 17:35:13,782][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.09404633939266205, 'train/grad_norm_pre_clip': 42.575592041015625, 'timing/data': 0.0006167789688333869, 'timing/model': 1.9255701039219275, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 1.5}) +[2026-06-23 17:35:17,592][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.019300758838653564, 'train/grad_norm_pre_clip': 38.82585144042969, 'timing/data': 0.0006091010291129351, 'timing/model': 1.9336805500788614, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 1.5}) +[2026-06-23 17:35:21,390][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.040581900626420975, 'train/grad_norm_pre_clip': 32.1722412109375, 'timing/data': 0.0005964030278846622, 'timing/model': 1.92574459197931, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 1.5}) +[2026-06-23 17:35:25,181][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.026838062331080437, 'train/grad_norm_pre_clip': 21.849506378173828, 'timing/data': 0.0008158460259437561, 'timing/model': 1.9222488430095837, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 1.5}) +[2026-06-23 17:35:28,993][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.01935332641005516, 'train/grad_norm_pre_clip': 24.37053871154785, 'timing/data': 0.0006094219861552119, 'timing/model': 1.9316911869682372, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 1.51}) +[2026-06-23 17:35:32,801][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.08878318965435028, 'train/grad_norm_pre_clip': 35.371028900146484, 'timing/data': 0.0005962700815871358, 'timing/model': 1.9349043510155752, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 1.51}) +[2026-06-23 17:35:36,657][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.10219604521989822, 'train/grad_norm_pre_clip': 47.29378890991211, 'timing/data': 0.0005937239620834589, 'timing/model': 1.973802616004832, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 1.51}) +[2026-06-23 17:35:40,471][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.0562494620680809, 'train/grad_norm_pre_clip': 31.988086700439453, 'timing/data': 0.0005861719837412238, 'timing/model': 1.9357115849852562, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 1.51}) +[2026-06-23 17:35:44,292][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.049075137823820114, 'train/grad_norm_pre_clip': 31.4848690032959, 'timing/data': 0.0007644819561392069, 'timing/model': 1.9375919930171221, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 1.51}) +[2026-06-23 17:35:48,096][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.0443401001393795, 'train/grad_norm_pre_clip': 53.69877243041992, 'timing/data': 0.0005907909944653511, 'timing/model': 1.931339192087762, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 1.51}) +[2026-06-23 17:35:51,906][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.02108559012413025, 'train/grad_norm_pre_clip': 42.347801208496094, 'timing/data': 0.0008231370011344552, 'timing/model': 1.938557528075762, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 1.51}) +[2026-06-23 17:35:55,748][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.024257170036435127, 'train/grad_norm_pre_clip': 33.56120300292969, 'timing/data': 0.0005938659887760878, 'timing/model': 1.9638725059339777, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 1.51}) +[2026-06-23 17:35:59,547][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.05646718665957451, 'train/grad_norm_pre_clip': 48.643367767333984, 'timing/data': 0.000611363910138607, 'timing/model': 1.928124523954466, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 1.51}) +[2026-06-23 17:36:03,344][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.06172120198607445, 'train/grad_norm_pre_clip': 36.14037322998047, 'timing/data': 0.0006154910661280155, 'timing/model': 1.9274607819970697, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 1.51}) +[2026-06-23 17:36:07,154][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.12939834594726562, 'train/grad_norm_pre_clip': 33.12709426879883, 'timing/data': 0.0010472959838807583, 'timing/model': 1.9282338520279154, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 1.51}) +[2026-06-23 17:36:10,977][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3900 +[2026-06-23 17:38:19,776][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3900 +[2026-06-23 17:38:19,778][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3900 +[2026-06-23 17:39:41,658][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3900 +[2026-06-23 17:39:41,659][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.0799570307135582, 'train/grad_norm_pre_clip': 20.725988388061523, 'eval/action_loss/seconds': 128.79634342808276, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03278358653187752, 'eval/latency_2/loss': 0.03278358653187752, 'eval/flappy/loss': 0.03278358653187752, 'eval/flappy/latency_2/loss': 0.03278358653187752, 'eval/flap_precision': 0.9588859416445623, 'eval/flap_recall': 0.9757085020242915, 'eval/flap_f1': 0.9672240802675585, 'eval/flap_rate_model': 0.20944444444444443, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9672240802675585, 'eval/latency_2/flap_precision': 0.9588859416445623, 'eval/latency_2/flap_recall': 0.9757085020242915, 'eval/latency_2/flap_f1': 0.9672240802675585, 'eval/latency_2/flap_rate_model': 0.20944444444444443, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9672240802675585, 'eval/action_classification/data_wait_seconds': 1.5900189876556396, 'eval/action_classification/predict_seconds': 80.1014175415039, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006484780460596085, 'timing/model': 1.9331255939323455, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 1.51}) +[2026-06-23 17:39:45,445][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.12894074618816376, 'train/grad_norm_pre_clip': 49.15084457397461, 'timing/data': 0.0006091769319027662, 'timing/model': 1.9205701010068879, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 1.51}) +[2026-06-23 17:39:49,223][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.028339814394712448, 'train/grad_norm_pre_clip': 37.75836944580078, 'timing/data': 0.0007047279505059123, 'timing/model': 1.9155574770411476, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 1.51}) +[2026-06-23 17:39:52,996][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.008505814708769321, 'train/grad_norm_pre_clip': 20.855005264282227, 'timing/data': 0.0006077809957787395, 'timing/model': 1.9154478729469702, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 1.51}) +[2026-06-23 17:39:56,781][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.04148843511939049, 'train/grad_norm_pre_clip': 20.515552520751953, 'timing/data': 0.0006557999877259135, 'timing/model': 1.9169375789351761, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 1.51}) +[2026-06-23 17:40:00,742][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.010819773189723492, 'train/grad_norm_pre_clip': 9.692039489746094, 'timing/data': 0.0005918140523135662, 'timing/model': 1.9162526211002842, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 1.51}) +[2026-06-23 17:40:04,531][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.008937414735555649, 'train/grad_norm_pre_clip': 23.133861541748047, 'timing/data': 0.0006235679611563683, 'timing/model': 1.9182864610338584, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 1.51}) +[2026-06-23 17:40:08,313][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.06602640450000763, 'train/grad_norm_pre_clip': 27.095109939575195, 'timing/data': 0.0006193129811435938, 'timing/model': 1.9146092059090734, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 1.51}) +[2026-06-23 17:40:12,107][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.107827328145504, 'train/grad_norm_pre_clip': 54.68021774291992, 'timing/data': 0.0005965189775452018, 'timing/model': 1.9176124890800565, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 1.51}) +[2026-06-23 17:40:15,920][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.07230266183614731, 'train/grad_norm_pre_clip': 34.85072326660156, 'timing/data': 0.0005923520075157285, 'timing/model': 1.9420415950007737, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 1.51}) +[2026-06-23 17:40:19,721][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.05975141003727913, 'train/grad_norm_pre_clip': 34.86076736450195, 'timing/data': 0.0006097418954595923, 'timing/model': 1.9264705859823152, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 1.51}) +[2026-06-23 17:40:23,506][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.03434517979621887, 'train/grad_norm_pre_clip': 25.362985610961914, 'timing/data': 0.0006761889671906829, 'timing/model': 1.91921718895901, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 1.51}) +[2026-06-23 17:40:27,302][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.022367103025317192, 'train/grad_norm_pre_clip': 42.12480163574219, 'timing/data': 0.0007075580069795251, 'timing/model': 1.929478635895066, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 1.51}) +[2026-06-23 17:40:31,269][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.008445598185062408, 'train/grad_norm_pre_clip': 12.645662307739258, 'timing/data': 0.0007377489237114787, 'timing/model': 2.1118529699742794, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 1.51}) +[2026-06-23 17:40:35,070][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.05754527077078819, 'train/grad_norm_pre_clip': 41.9637565612793, 'timing/data': 0.0006745369173586369, 'timing/model': 1.9265921879559755, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 1.51}) +[2026-06-23 17:40:38,868][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.13405172526836395, 'train/grad_norm_pre_clip': 30.065013885498047, 'timing/data': 0.0006145799998193979, 'timing/model': 1.9277687690919265, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 1.52}) +[2026-06-23 17:40:42,672][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.037235938012599945, 'train/grad_norm_pre_clip': 37.26976776123047, 'timing/data': 0.0008059620158746839, 'timing/model': 1.9309789970284328, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 1.52}) +[2026-06-23 17:40:46,462][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.06574928760528564, 'train/grad_norm_pre_clip': 56.03175354003906, 'timing/data': 0.0006789950421079993, 'timing/model': 1.926870345021598, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 1.52}) +[2026-06-23 17:40:50,247][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.03197624906897545, 'train/grad_norm_pre_clip': 15.830269813537598, 'timing/data': 0.0005912249907851219, 'timing/model': 1.9165084100095555, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 1.52}) +[2026-06-23 17:40:54,050][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.09071775525808334, 'train/grad_norm_pre_clip': 32.751609802246094, 'timing/data': 0.0006118509918451309, 'timing/model': 1.9259844799526036, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 1.52}) +[2026-06-23 17:40:57,874][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.011260652914643288, 'train/grad_norm_pre_clip': 13.869119644165039, 'timing/data': 0.0006218180060386658, 'timing/model': 1.9284896400058642, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 1.52}) +[2026-06-23 17:41:01,661][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.04192321002483368, 'train/grad_norm_pre_clip': 28.321977615356445, 'timing/data': 0.0007657810347154737, 'timing/model': 1.9262185600819066, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 1.52}) +[2026-06-23 17:41:05,634][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.04956474155187607, 'train/grad_norm_pre_clip': 35.18571090698242, 'timing/data': 0.0006083609769120812, 'timing/model': 1.9153296690201387, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 1.52}) +[2026-06-23 17:41:09,425][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.028726210817694664, 'train/grad_norm_pre_clip': 23.524150848388672, 'timing/data': 0.0006321909604594111, 'timing/model': 1.9247774629620835, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 1.52}) +[2026-06-23 17:41:13,245][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.0026877988129854202, 'train/grad_norm_pre_clip': 31.030473709106445, 'timing/data': 0.0005969220073893666, 'timing/model': 1.946497175958939, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 1.52}) +[2026-06-23 17:41:17,059][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.06677541136741638, 'train/grad_norm_pre_clip': 24.052574157714844, 'timing/data': 0.0006021240260452032, 'timing/model': 1.9391929280245677, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 1.52}) +[2026-06-23 17:41:20,856][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.06123100221157074, 'train/grad_norm_pre_clip': 59.5511474609375, 'timing/data': 0.000609392998740077, 'timing/model': 1.9300049379235134, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 1.52}) +[2026-06-23 17:41:24,657][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.04052020609378815, 'train/grad_norm_pre_clip': 24.582948684692383, 'timing/data': 0.0006033660611137748, 'timing/model': 1.9292012739460915, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 1.52}) +[2026-06-23 17:41:28,450][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.03081141784787178, 'train/grad_norm_pre_clip': 16.61197280883789, 'timing/data': 0.0005922840209677815, 'timing/model': 1.9214971120236441, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 1.52}) +[2026-06-23 17:41:32,312][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.15388156473636627, 'train/grad_norm_pre_clip': 57.6429328918457, 'timing/data': 0.000594249926507473, 'timing/model': 1.9879195649409667, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 1.52}) +[2026-06-23 17:41:36,279][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.024576524272561073, 'train/grad_norm_pre_clip': 15.82413387298584, 'timing/data': 0.0008104019798338413, 'timing/model': 2.0849748300388455, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 1.52}) +[2026-06-23 17:41:40,076][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.031988535076379776, 'train/grad_norm_pre_clip': 64.69084930419922, 'timing/data': 0.0006251230370253325, 'timing/model': 1.9277005080366507, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 1.52}) +[2026-06-23 17:41:43,880][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.055360253900289536, 'train/grad_norm_pre_clip': 29.7420711517334, 'timing/data': 0.0007931120926514268, 'timing/model': 1.926634405972436, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 1.52}) +[2026-06-23 17:41:47,671][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.05292581394314766, 'train/grad_norm_pre_clip': 25.000211715698242, 'timing/data': 0.0005955399246886373, 'timing/model': 1.9236574260285124, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 1.52}) +[2026-06-23 17:41:51,495][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.02368689328432083, 'train/grad_norm_pre_clip': 35.854312896728516, 'timing/data': 0.0005953309591859579, 'timing/model': 1.921514410059899, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 1.52}) +[2026-06-23 17:41:55,330][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.0585065633058548, 'train/grad_norm_pre_clip': 18.995399475097656, 'timing/data': 0.0006034719990566373, 'timing/model': 1.9351353760575876, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 1.52}) +[2026-06-23 17:41:59,129][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.05687588453292847, 'train/grad_norm_pre_clip': 46.847537994384766, 'timing/data': 0.0005895450012758374, 'timing/model': 1.9271038479637355, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 1.52}) +[2026-06-23 17:42:02,917][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.005598492454737425, 'train/grad_norm_pre_clip': 34.30521011352539, 'timing/data': 0.0006556189619004726, 'timing/model': 1.9284188189776614, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 1.52}) +[2026-06-23 17:42:06,713][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.01279196236282587, 'train/grad_norm_pre_clip': 12.161985397338867, 'timing/data': 0.0006552130216732621, 'timing/model': 1.9235307209892198, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 1.52}) +[2026-06-23 17:42:10,494][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.09073691815137863, 'train/grad_norm_pre_clip': 36.052711486816406, 'timing/data': 0.0006122809136286378, 'timing/model': 1.9162433960009366, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 1.52}) +[2026-06-23 17:42:14,309][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.12703275680541992, 'train/grad_norm_pre_clip': 83.24681091308594, 'timing/data': 0.0005938829854130745, 'timing/model': 1.9420236110454425, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 1.52}) +[2026-06-23 17:42:18,123][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.11009343713521957, 'train/grad_norm_pre_clip': 48.285743713378906, 'timing/data': 0.0006101049948483706, 'timing/model': 1.9336005989462137, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 1.53}) +[2026-06-23 17:42:21,930][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.0059687793254852295, 'train/grad_norm_pre_clip': 13.599863052368164, 'timing/data': 0.000608316040597856, 'timing/model': 1.9237418889533728, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 1.53}) +[2026-06-23 17:42:25,726][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.03763844445347786, 'train/grad_norm_pre_clip': 33.22925567626953, 'timing/data': 0.0006266830023378134, 'timing/model': 1.931458231061697, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 1.53}) +[2026-06-23 17:42:29,520][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.043311432003974915, 'train/grad_norm_pre_clip': 28.850231170654297, 'timing/data': 0.0006071190582588315, 'timing/model': 1.9266098199877888, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 1.53}) +[2026-06-23 17:42:33,309][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.0850050076842308, 'train/grad_norm_pre_clip': 37.22698211669922, 'timing/data': 0.0005993120139464736, 'timing/model': 1.9264323291135952, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 1.53}) +[2026-06-23 17:42:37,118][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.011943692341446877, 'train/grad_norm_pre_clip': 15.28148365020752, 'timing/data': 0.0005905240541324019, 'timing/model': 1.9329805520828813, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 1.53}) +[2026-06-23 17:42:40,927][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.013133186846971512, 'train/grad_norm_pre_clip': 25.38153839111328, 'timing/data': 0.0006054219556972384, 'timing/model': 1.9293192309560254, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 1.53}) +[2026-06-23 17:42:44,736][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.023625070229172707, 'train/grad_norm_pre_clip': 27.00068473815918, 'timing/data': 0.0005962670547887683, 'timing/model': 1.9264194610295817, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 1.53}) +[2026-06-23 17:42:48,553][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.04196705296635628, 'train/grad_norm_pre_clip': 14.667932510375977, 'timing/data': 0.0005988420452922583, 'timing/model': 1.9384514739504084, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 1.53}) +[2026-06-23 17:42:52,360][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.06187407672405243, 'train/grad_norm_pre_clip': 43.123558044433594, 'timing/data': 0.0005940520204603672, 'timing/model': 1.9354113539448008, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 1.53}) +[2026-06-23 17:42:56,166][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.06031946837902069, 'train/grad_norm_pre_clip': 21.598953247070312, 'timing/data': 0.000602934043854475, 'timing/model': 1.9339143090182915, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 1.53}) +[2026-06-23 17:42:59,974][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.10675722360610962, 'train/grad_norm_pre_clip': 32.17327117919922, 'timing/data': 0.0007824529893696308, 'timing/model': 1.935909705935046, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 1.53}) +[2026-06-23 17:43:03,823][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.13415266573429108, 'train/grad_norm_pre_clip': 51.67379379272461, 'timing/data': 0.0007187570445239544, 'timing/model': 1.9415372010553256, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 1.53}) +[2026-06-23 17:43:07,631][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.06320469081401825, 'train/grad_norm_pre_clip': 55.1521110534668, 'timing/data': 0.0008212120737880468, 'timing/model': 1.9318446090910584, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 1.53}) +[2026-06-23 17:43:11,450][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.06857926398515701, 'train/grad_norm_pre_clip': 26.69371223449707, 'timing/data': 0.0006117270095273852, 'timing/model': 1.9309500199742615, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 1.53}) +[2026-06-23 17:43:15,436][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.07121345400810242, 'train/grad_norm_pre_clip': 39.455810546875, 'timing/data': 0.0008305000374093652, 'timing/model': 2.114485456957482, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 1.53}) +[2026-06-23 17:43:19,268][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.013722749426960945, 'train/grad_norm_pre_clip': 6.905643939971924, 'timing/data': 0.0006111649563536048, 'timing/model': 1.9522275800118223, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 1.53}) +[2026-06-23 17:43:23,071][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.018732838332653046, 'train/grad_norm_pre_clip': 32.8851203918457, 'timing/data': 0.0006159860640764236, 'timing/model': 1.9245448879664764, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 1.53}) +[2026-06-23 17:43:26,887][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.029147986322641373, 'train/grad_norm_pre_clip': 13.021092414855957, 'timing/data': 0.0006079550366848707, 'timing/model': 1.9431016150629148, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 1.53}) +[2026-06-23 17:43:30,681][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.07409178465604782, 'train/grad_norm_pre_clip': 66.13818359375, 'timing/data': 0.0006040689768269658, 'timing/model': 1.9305020129540935, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 1.53}) +[2026-06-23 17:43:34,491][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.02024751529097557, 'train/grad_norm_pre_clip': 10.894400596618652, 'timing/data': 0.0006196169415488839, 'timing/model': 1.929626132012345, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 1.53}) +[2026-06-23 17:43:38,292][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.04632965847849846, 'train/grad_norm_pre_clip': 25.184303283691406, 'timing/data': 0.0006165500963106751, 'timing/model': 1.9279383640969172, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 1.53}) +[2026-06-23 17:43:42,095][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.043199293315410614, 'train/grad_norm_pre_clip': 27.12518882751465, 'timing/data': 0.0005884750280529261, 'timing/model': 1.9287196929799393, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 1.53}) +[2026-06-23 17:43:45,912][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.11042485386133194, 'train/grad_norm_pre_clip': 29.23674964904785, 'timing/data': 0.0005955280503258109, 'timing/model': 1.9309754109708592, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 1.53}) +[2026-06-23 17:43:49,702][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.11387258768081665, 'train/grad_norm_pre_clip': 53.44047546386719, 'timing/data': 0.0005943999858573079, 'timing/model': 1.9234473860124126, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 1.53}) +[2026-06-23 17:43:53,511][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.046837568283081055, 'train/grad_norm_pre_clip': 53.13301467895508, 'timing/data': 0.0006100129103288054, 'timing/model': 1.9384527719812468, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 1.53}) +[2026-06-23 17:43:57,321][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.06687696278095245, 'train/grad_norm_pre_clip': 17.605539321899414, 'timing/data': 0.0005968679906800389, 'timing/model': 1.9256018330343068, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 1.54}) +[2026-06-23 17:44:01,117][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.054568950086832047, 'train/grad_norm_pre_clip': 26.82306671142578, 'timing/data': 0.000907114939764142, 'timing/model': 1.9248183110030368, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 1.54}) +[2026-06-23 17:44:04,924][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.017369328066706657, 'train/grad_norm_pre_clip': 24.32668685913086, 'timing/data': 0.0006033660611137748, 'timing/model': 1.92917422298342, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 1.54}) +[2026-06-23 17:44:08,727][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.10809354484081268, 'train/grad_norm_pre_clip': 68.91683197021484, 'timing/data': 0.0008320180932059884, 'timing/model': 1.9248966770246625, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 1.54}) +[2026-06-23 17:44:12,528][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.006153285037726164, 'train/grad_norm_pre_clip': 24.043079376220703, 'timing/data': 0.0005872970214113593, 'timing/model': 1.9306976499501616, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 1.54}) +[2026-06-23 17:44:16,324][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.058693982660770416, 'train/grad_norm_pre_clip': 25.81306266784668, 'timing/data': 0.000616756035014987, 'timing/model': 1.924356042058207, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 1.54}) +[2026-06-23 17:44:20,135][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.04844747856259346, 'train/grad_norm_pre_clip': 33.13264465332031, 'timing/data': 0.0005905600264668465, 'timing/model': 1.9329834970412776, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 1.54}) +[2026-06-23 17:44:23,946][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.09029955416917801, 'train/grad_norm_pre_clip': 35.306060791015625, 'timing/data': 0.0006564209470525384, 'timing/model': 1.9301814628997818, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 1.54}) +[2026-06-23 17:44:27,772][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.014313869178295135, 'train/grad_norm_pre_clip': 6.981924057006836, 'timing/data': 0.0005938870599493384, 'timing/model': 1.9390226029790938, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 1.54}) +[2026-06-23 17:44:31,582][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.02928353101015091, 'train/grad_norm_pre_clip': 18.088584899902344, 'timing/data': 0.0005962569266557693, 'timing/model': 1.9301684320671484, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 1.54}) +[2026-06-23 17:44:35,430][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.08239126205444336, 'train/grad_norm_pre_clip': 33.63834762573242, 'timing/data': 0.0007284639868885279, 'timing/model': 1.932002456043847, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 1.54}) +[2026-06-23 17:44:39,248][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.13189536333084106, 'train/grad_norm_pre_clip': 44.568565368652344, 'timing/data': 0.0006054830737411976, 'timing/model': 1.9388985579134896, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 1.54}) +[2026-06-23 17:44:43,052][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.05729685351252556, 'train/grad_norm_pre_clip': 40.59455108642578, 'timing/data': 0.0005895410431548953, 'timing/model': 1.9337005410343409, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 1.54}) +[2026-06-23 17:44:46,871][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.11336114257574081, 'train/grad_norm_pre_clip': 45.957515716552734, 'timing/data': 0.0006024090107530355, 'timing/model': 1.937989524914883, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 1.54}) +[2026-06-23 17:44:50,666][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.052428435534238815, 'train/grad_norm_pre_clip': 27.244060516357422, 'timing/data': 0.000893812975846231, 'timing/model': 1.9319547140039504, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 1.54}) +[2026-06-23 17:44:54,460][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.009279254823923111, 'train/grad_norm_pre_clip': 39.40339660644531, 'timing/data': 0.0005912639899179339, 'timing/model': 1.9229893860174343, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 1.54}) +[2026-06-23 17:44:58,297][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.015512874349951744, 'train/grad_norm_pre_clip': 63.997215270996094, 'timing/data': 0.0007247210014611483, 'timing/model': 1.9619154209503904, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 1.54}) +[2026-06-23 17:45:02,109][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.061571329832077026, 'train/grad_norm_pre_clip': 54.655941009521484, 'timing/data': 0.0006255819462239742, 'timing/model': 1.9381742170080543, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 1.54}) +[2026-06-23 17:45:05,934][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.048962872475385666, 'train/grad_norm_pre_clip': 24.563188552856445, 'timing/data': 0.0005998259875923395, 'timing/model': 1.9416941590607166, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 1.54}) +[2026-06-23 17:45:09,732][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.031801722943782806, 'train/grad_norm_pre_clip': 18.894954681396484, 'timing/data': 0.0006615601014345884, 'timing/model': 1.9280485160415992, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 1.54}) +[2026-06-23 17:45:13,532][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.17423643171787262, 'train/grad_norm_pre_clip': 73.40682220458984, 'timing/data': 0.0005988589255139232, 'timing/model': 1.927337885950692, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 1.54}) +[2026-06-23 17:45:17,339][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.03721192851662636, 'train/grad_norm_pre_clip': 21.02164649963379, 'timing/data': 0.0006069099763408303, 'timing/model': 1.9330067969858646, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 1.54}) +[2026-06-23 17:45:21,150][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.06879698485136032, 'train/grad_norm_pre_clip': 50.75897979736328, 'timing/data': 0.0006017079576849937, 'timing/model': 1.931363477022387, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 1.54}) +[2026-06-23 17:45:24,967][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.025194071233272552, 'train/grad_norm_pre_clip': 14.594679832458496, 'timing/data': 0.0005909251049160957, 'timing/model': 1.9396111849928275, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 1.54}) +[2026-06-23 17:45:28,959][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.08410090208053589, 'train/grad_norm_pre_clip': 44.8494987487793, 'timing/data': 0.0006070500239729881, 'timing/model': 1.9263257490238175, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 1.54}) +[2026-06-23 17:45:32,753][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.09523960202932358, 'train/grad_norm_pre_clip': 49.386016845703125, 'timing/data': 0.0006194949382916093, 'timing/model': 1.9197593359276652, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 1.54}) +[2026-06-23 17:45:36,562][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.014941032975912094, 'train/grad_norm_pre_clip': 52.68263244628906, 'timing/data': 0.0006137979216873646, 'timing/model': 1.9277693439507857, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 1.55}) +[2026-06-23 17:45:40,368][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.06901314854621887, 'train/grad_norm_pre_clip': 33.26681137084961, 'timing/data': 0.0006496109999716282, 'timing/model': 1.9325901370029896, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 1.55}) +[2026-06-23 17:45:44,173][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.04089409112930298, 'train/grad_norm_pre_clip': 18.153789520263672, 'timing/data': 0.0006492650136351585, 'timing/model': 1.9304887499893084, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 1.55}) +[2026-06-23 17:45:47,981][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.03946829214692116, 'train/grad_norm_pre_clip': 24.331424713134766, 'timing/data': 0.0006022220477461815, 'timing/model': 1.9361786210210994, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 1.55}) +[2026-06-23 17:45:51,782][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.0342797115445137, 'train/grad_norm_pre_clip': 27.752696990966797, 'timing/data': 0.0006659269565716386, 'timing/model': 1.9280767089221627, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 1.55}) +[2026-06-23 17:45:55,590][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.018543437123298645, 'train/grad_norm_pre_clip': 37.53971862792969, 'timing/data': 0.0006037909770384431, 'timing/model': 1.934314938960597, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 1.55}) +[2026-06-23 17:45:59,398][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.027513686567544937, 'train/grad_norm_pre_clip': 38.700035095214844, 'timing/data': 0.0008847629651427269, 'timing/model': 1.9367138739908114, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 1.55}) +[2026-06-23 17:46:03,205][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4000 +[2026-06-23 17:48:09,679][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4000 +[2026-06-23 17:48:09,680][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4000 +[2026-06-23 17:49:32,805][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4000 +[2026-06-23 17:56:38,894][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000 +[2026-06-23 17:56:38,894][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.11044824123382568, 'train/grad_norm_pre_clip': 37.700157165527344, 'eval/action_loss/seconds': 126.47170226601884, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03364407643675804, 'eval/latency_2/loss': 0.03364407643675804, 'eval/flappy/loss': 0.03364407643675804, 'eval/flappy/latency_2/loss': 0.03364407643675804, 'eval/flap_precision': 0.977961432506887, 'eval/flap_recall': 0.9581646423751687, 'eval/flap_f1': 0.9679618268575324, 'eval/flap_rate_model': 0.20166666666666666, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9679618268575323, 'eval/latency_2/flap_precision': 0.977961432506887, 'eval/latency_2/flap_recall': 0.9581646423751687, 'eval/latency_2/flap_f1': 0.9679618268575324, 'eval/latency_2/flap_rate_model': 0.20166666666666666, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9679618268575323, 'eval/action_classification/data_wait_seconds': 1.5996506214141846, 'eval/action_classification/predict_seconds': 81.3460693359375, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 261.3800000000085, 'rl_games_eval/mid_train/mean_length': 2125.4, 'rl_games_eval/mid_train/std_reward': 157.84373791824612, 'rl_games_eval/mid_train/std_length': 1270.7383050809478, 'rl_games_eval/mid_train/macro_mean_reward': 261.3800000000085, 'rl_games_eval/mid_train/macro_mean_length': 2125.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 261.3800000000085, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2125.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 157.84373791824612, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1270.7383050809478, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 261.3800000000085, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006216929759830236, 'timing/model': 1.938487765030004, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 1.55}) +[2026-06-23 17:56:38,894][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_4000_state +[2026-06-23 17:56:38,897][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 17:56:42,636][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_4000_state/model.safetensors +[2026-06-23 17:56:45,532][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_4000_state/optimizer.bin +[2026-06-23 17:56:45,532][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_4000_state/sampler.bin +[2026-06-23 17:56:45,532][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-23 17:56:45,533][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-23 17:56:45,558][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-23 17:56:45,558][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 17:56:45,569][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 17:56:49,367][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.08091588318347931, 'train/grad_norm_pre_clip': 41.641422271728516, 'timing/data': 0.0006228999700397253, 'timing/model': 1.9474063600646332, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 1.55}) +[2026-06-23 17:56:53,100][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.023997638374567032, 'train/grad_norm_pre_clip': 20.349241256713867, 'timing/data': 0.0006072929827496409, 'timing/model': 1.8971632350003347, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 1.55}) +[2026-06-23 17:56:56,835][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.05449344590306282, 'train/grad_norm_pre_clip': 42.550594329833984, 'timing/data': 0.0005950289778411388, 'timing/model': 1.8958080739248544, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 1.55}) +[2026-06-23 17:57:00,574][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.04034322127699852, 'train/grad_norm_pre_clip': 27.614255905151367, 'timing/data': 0.0006699149962514639, 'timing/model': 1.9045790589880198, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 1.55}) +[2026-06-23 17:57:04,319][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.0578475184738636, 'train/grad_norm_pre_clip': 27.4653377532959, 'timing/data': 0.0005950049962848425, 'timing/model': 1.9056732869939879, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 1.55}) +[2026-06-23 17:57:08,066][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.02211185172200203, 'train/grad_norm_pre_clip': 13.399059295654297, 'timing/data': 0.0006142889615148306, 'timing/model': 1.9046315299347043, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 1.55}) +[2026-06-23 17:57:11,885][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.15703164041042328, 'train/grad_norm_pre_clip': 29.67591667175293, 'timing/data': 0.0006075180135667324, 'timing/model': 1.971556906006299, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 1.55}) +[2026-06-23 17:57:15,636][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.07478699833154678, 'train/grad_norm_pre_clip': 37.919742584228516, 'timing/data': 0.000588549068197608, 'timing/model': 1.9037199170561507, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 1.55}) +[2026-06-23 17:57:19,391][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.05165313556790352, 'train/grad_norm_pre_clip': 35.13862228393555, 'timing/data': 0.0006240889197215438, 'timing/model': 1.9069121989887208, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 1.55}) +[2026-06-23 17:57:23,216][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.016640543937683105, 'train/grad_norm_pre_clip': 26.722232818603516, 'timing/data': 0.0005880540702491999, 'timing/model': 1.9087154309963807, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 1.55}) +[2026-06-23 17:57:26,970][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.07746445387601852, 'train/grad_norm_pre_clip': 22.309879302978516, 'timing/data': 0.000660609919577837, 'timing/model': 1.9020946249365807, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 1.55}) +[2026-06-23 17:57:30,744][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.09685716032981873, 'train/grad_norm_pre_clip': 82.93737030029297, 'timing/data': 0.0006550849648192525, 'timing/model': 1.918762914952822, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 1.55}) +[2026-06-23 17:57:34,524][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.04614455997943878, 'train/grad_norm_pre_clip': 40.412654876708984, 'timing/data': 0.0005955280503258109, 'timing/model': 1.9174968679435551, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 1.55}) +[2026-06-23 17:57:38,295][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.017968077212572098, 'train/grad_norm_pre_clip': 40.5093879699707, 'timing/data': 0.0005813419120386243, 'timing/model': 1.9201189040904865, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 1.55}) +[2026-06-23 17:57:42,094][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.03573291003704071, 'train/grad_norm_pre_clip': 27.328956604003906, 'timing/data': 0.0005856429925188422, 'timing/model': 1.945339962025173, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 1.55}) +[2026-06-23 17:57:45,857][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.01914188638329506, 'train/grad_norm_pre_clip': 17.498090744018555, 'timing/data': 0.0006184970261529088, 'timing/model': 1.9146577309584245, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 1.55}) +[2026-06-23 17:57:49,619][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.028716029599308968, 'train/grad_norm_pre_clip': 25.765544891357422, 'timing/data': 0.0006014590617269278, 'timing/model': 1.9082111159805208, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 1.55}) +[2026-06-23 17:57:53,451][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.05717844516038895, 'train/grad_norm_pre_clip': 18.680896759033203, 'timing/data': 0.000830297009088099, 'timing/model': 1.9652421349892393, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 1.56}) +[2026-06-23 17:57:57,225][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.014460659585893154, 'train/grad_norm_pre_clip': 9.85366439819336, 'timing/data': 0.0006079910090193152, 'timing/model': 1.9221088759368286, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 1.56}) +[2026-06-23 17:58:01,003][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.054780781269073486, 'train/grad_norm_pre_clip': 27.846948623657227, 'timing/data': 0.0009076340356841683, 'timing/model': 1.9225440860027447, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 1.56}) +[2026-06-23 17:58:04,780][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.1551578938961029, 'train/grad_norm_pre_clip': 36.15053939819336, 'timing/data': 0.0005987170152366161, 'timing/model': 1.9107068750308827, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 1.56}) +[2026-06-23 17:58:08,579][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.06591779738664627, 'train/grad_norm_pre_clip': 49.228233337402344, 'timing/data': 0.0006571609992533922, 'timing/model': 1.927965612965636, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 1.56}) +[2026-06-23 17:58:12,374][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.0330750048160553, 'train/grad_norm_pre_clip': 28.947107315063477, 'timing/data': 0.0006580649642273784, 'timing/model': 1.927490281057544, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 1.56}) +[2026-06-23 17:58:16,160][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.06986182928085327, 'train/grad_norm_pre_clip': 41.082481384277344, 'timing/data': 0.0006027310155332088, 'timing/model': 1.9225657050265, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 1.56}) +[2026-06-23 17:58:19,954][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.06460200995206833, 'train/grad_norm_pre_clip': 32.940101623535156, 'timing/data': 0.0007993279723450541, 'timing/model': 1.9297950640320778, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 1.56}) +[2026-06-23 17:58:23,744][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.08334434032440186, 'train/grad_norm_pre_clip': 42.79643630981445, 'timing/data': 0.0006038849242031574, 'timing/model': 1.923310825950466, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 1.56}) +[2026-06-23 17:58:27,545][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.03181516379117966, 'train/grad_norm_pre_clip': 34.283870697021484, 'timing/data': 0.0006043100729584694, 'timing/model': 1.9185745469294488, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 1.56}) +[2026-06-23 17:58:31,329][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.043945807963609695, 'train/grad_norm_pre_clip': 46.77405548095703, 'timing/data': 0.0005991130601614714, 'timing/model': 1.9209046360338107, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 1.56}) +[2026-06-23 17:58:35,112][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.06876067072153091, 'train/grad_norm_pre_clip': 27.941495895385742, 'timing/data': 0.0005870240274816751, 'timing/model': 1.912623033975251, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 1.56}) +[2026-06-23 17:58:38,905][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.03533351793885231, 'train/grad_norm_pre_clip': 29.092985153198242, 'timing/data': 0.0007536860648542643, 'timing/model': 1.926824340946041, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 1.56}) +[2026-06-23 17:58:42,700][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.009940855205059052, 'train/grad_norm_pre_clip': 21.683698654174805, 'timing/data': 0.0009497179416939616, 'timing/model': 1.929970259894617, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 1.56}) +[2026-06-23 17:58:46,496][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.03314312919974327, 'train/grad_norm_pre_clip': 20.84076499938965, 'timing/data': 0.0005971720675006509, 'timing/model': 1.924905023071915, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 1.56}) +[2026-06-23 17:58:50,298][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.03953548148274422, 'train/grad_norm_pre_clip': 18.76863670349121, 'timing/data': 0.0008174170507118106, 'timing/model': 1.9286420969292521, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 1.56}) +[2026-06-23 17:58:54,098][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.06039518117904663, 'train/grad_norm_pre_clip': 28.181968688964844, 'timing/data': 0.0005806430708616972, 'timing/model': 1.9272694380488247, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 1.56}) +[2026-06-23 17:58:57,894][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.044204432517290115, 'train/grad_norm_pre_clip': 41.67232131958008, 'timing/data': 0.0005965569289401174, 'timing/model': 1.9301773300394416, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 1.56}) +[2026-06-23 17:59:01,876][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.03836794197559357, 'train/grad_norm_pre_clip': 21.573535919189453, 'timing/data': 0.0007323800818994641, 'timing/model': 1.9180316920392215, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 1.56}) +[2026-06-23 17:59:05,673][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.018864864483475685, 'train/grad_norm_pre_clip': 19.67200469970703, 'timing/data': 0.0006747879087924957, 'timing/model': 1.9223495071055368, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 1.56}) +[2026-06-23 17:59:09,481][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.029831426218152046, 'train/grad_norm_pre_clip': 66.279541015625, 'timing/data': 0.0008857230423018336, 'timing/model': 1.9255919490242377, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 1.56}) +[2026-06-23 17:59:13,290][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.08704671263694763, 'train/grad_norm_pre_clip': 23.056991577148438, 'timing/data': 0.0005934260552749038, 'timing/model': 1.9331939480034634, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 1.56}) +[2026-06-23 17:59:17,079][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.09948795288801193, 'train/grad_norm_pre_clip': 44.53022384643555, 'timing/data': 0.0006058919243514538, 'timing/model': 1.9179932120023295, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 1.56}) +[2026-06-23 17:59:20,881][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.024321405217051506, 'train/grad_norm_pre_clip': 15.77010440826416, 'timing/data': 0.0005964199081063271, 'timing/model': 1.9320629839785397, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 1.56}) +[2026-06-23 17:59:24,673][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.12164957821369171, 'train/grad_norm_pre_clip': 29.36918830871582, 'timing/data': 0.0005991399521008134, 'timing/model': 1.9264134790282696, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 1.56}) +[2026-06-23 17:59:28,459][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.045232128351926804, 'train/grad_norm_pre_clip': 30.090330123901367, 'timing/data': 0.0006047800416126847, 'timing/model': 1.9226424829103053, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 1.56}) +[2026-06-23 17:59:32,418][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.07197720557451248, 'train/grad_norm_pre_clip': 48.84022903442383, 'timing/data': 0.0005891619948670268, 'timing/model': 2.1022358199115843, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 1.57}) +[2026-06-23 17:59:36,219][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.03658353537321091, 'train/grad_norm_pre_clip': 17.529850006103516, 'timing/data': 0.0007234018994495273, 'timing/model': 1.9262965829111636, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 1.57}) +[2026-06-23 17:59:40,020][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.04129016771912575, 'train/grad_norm_pre_clip': 29.19120979309082, 'timing/data': 0.0005973289953544736, 'timing/model': 1.918094099033624, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 1.57}) +[2026-06-23 17:59:43,805][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.029389534145593643, 'train/grad_norm_pre_clip': 22.49310874938965, 'timing/data': 0.0006278849905356765, 'timing/model': 1.9144324739463627, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 1.57}) +[2026-06-23 17:59:47,607][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.07082469016313553, 'train/grad_norm_pre_clip': 41.67355728149414, 'timing/data': 0.0006091780960559845, 'timing/model': 1.9290563699323684, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 1.57}) +[2026-06-23 17:59:51,433][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.07071886956691742, 'train/grad_norm_pre_clip': 28.91234016418457, 'timing/data': 0.0008447429863736033, 'timing/model': 1.9289649309357628, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 1.57}) +[2026-06-23 17:59:55,225][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.03770992532372475, 'train/grad_norm_pre_clip': 26.45937728881836, 'timing/data': 0.0005948130274191499, 'timing/model': 1.9292138990713283, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 1.57}) +[2026-06-23 17:59:59,035][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.038292694836854935, 'train/grad_norm_pre_clip': 14.234845161437988, 'timing/data': 0.0005966890603303909, 'timing/model': 1.9273895360529423, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 1.57}) +[2026-06-23 18:00:02,826][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.05865037441253662, 'train/grad_norm_pre_clip': 39.817359924316406, 'timing/data': 0.0007102950476109982, 'timing/model': 1.9226528720464557, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 1.57}) +[2026-06-23 18:00:06,805][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.009997447952628136, 'train/grad_norm_pre_clip': 29.06316566467285, 'timing/data': 0.0007067660335451365, 'timing/model': 1.9258494859095663, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 1.57}) +[2026-06-23 18:00:10,607][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.08986138552427292, 'train/grad_norm_pre_clip': 27.848594665527344, 'timing/data': 0.0006835869280621409, 'timing/model': 1.9262458310695365, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 1.57}) +[2026-06-23 18:00:14,410][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.07067108154296875, 'train/grad_norm_pre_clip': 23.394197463989258, 'timing/data': 0.0006424180464819074, 'timing/model': 1.926689863903448, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 1.57}) +[2026-06-23 18:00:18,212][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.011525842361152172, 'train/grad_norm_pre_clip': 9.255158424377441, 'timing/data': 0.00064574193675071, 'timing/model': 1.9332868510391563, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 1.57}) +[2026-06-23 18:00:22,013][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.04079093039035797, 'train/grad_norm_pre_clip': 20.581193923950195, 'timing/data': 0.0006594660226255655, 'timing/model': 1.927407516981475, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 1.57}) +[2026-06-23 18:00:25,815][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.1270119696855545, 'train/grad_norm_pre_clip': 46.1569938659668, 'timing/data': 0.0005924500292167068, 'timing/model': 1.9279628630029038, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 1.57}) +[2026-06-23 18:00:29,620][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.06081242859363556, 'train/grad_norm_pre_clip': 39.22486114501953, 'timing/data': 0.0006078489823266864, 'timing/model': 1.928712290013209, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 1.57}) +[2026-06-23 18:00:33,427][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.04103842005133629, 'train/grad_norm_pre_clip': 33.95783996582031, 'timing/data': 0.0006053050747141242, 'timing/model': 1.929712557932362, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 1.57}) +[2026-06-23 18:00:37,399][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.04459715262055397, 'train/grad_norm_pre_clip': 23.435516357421875, 'timing/data': 0.0006097969599068165, 'timing/model': 2.108380614896305, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 1.57}) +[2026-06-23 18:00:41,175][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.04081368073821068, 'train/grad_norm_pre_clip': 73.67070770263672, 'timing/data': 0.0005973110673949122, 'timing/model': 1.9154257219051942, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 1.57}) +[2026-06-23 18:00:44,968][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.05972990021109581, 'train/grad_norm_pre_clip': 33.10512161254883, 'timing/data': 0.0005975369131192565, 'timing/model': 1.9225381419528276, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 1.57}) +[2026-06-23 18:00:48,805][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.03247367963194847, 'train/grad_norm_pre_clip': 25.989757537841797, 'timing/data': 0.0006056849379092455, 'timing/model': 1.9373520349618047, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 1.57}) +[2026-06-23 18:00:52,598][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.04178887978196144, 'train/grad_norm_pre_clip': 19.929489135742188, 'timing/data': 0.0006052739918231964, 'timing/model': 1.920781932072714, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 1.57}) +[2026-06-23 18:00:56,389][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.0636645182967186, 'train/grad_norm_pre_clip': 32.529205322265625, 'timing/data': 0.0006384029984474182, 'timing/model': 1.9258796949870884, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 1.57}) +[2026-06-23 18:01:00,178][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.04919297248125076, 'train/grad_norm_pre_clip': 37.70781707763672, 'timing/data': 0.0005978539120405912, 'timing/model': 1.9240692210150883, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 1.57}) +[2026-06-23 18:01:03,975][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.03289693593978882, 'train/grad_norm_pre_clip': 25.013107299804688, 'timing/data': 0.0005876260111108422, 'timing/model': 1.927207553992048, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 1.57}) +[2026-06-23 18:01:07,779][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.04671531170606613, 'train/grad_norm_pre_clip': 23.50508689880371, 'timing/data': 0.0005848150467500091, 'timing/model': 1.9194765640422702, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 1.57}) +[2026-06-23 18:01:11,766][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.04508903995156288, 'train/grad_norm_pre_clip': 20.976951599121094, 'timing/data': 0.0006358149694278836, 'timing/model': 1.9413620110135525, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 1.58}) +[2026-06-23 18:01:15,562][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.1810590922832489, 'train/grad_norm_pre_clip': 81.58728790283203, 'timing/data': 0.000603717053309083, 'timing/model': 1.9289081760216504, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 1.58}) +[2026-06-23 18:01:19,363][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.029941072687506676, 'train/grad_norm_pre_clip': 17.346811294555664, 'timing/data': 0.0006179800257086754, 'timing/model': 1.9274867250351235, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 1.58}) +[2026-06-23 18:01:23,141][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.03554536774754524, 'train/grad_norm_pre_clip': 41.196537017822266, 'timing/data': 0.0005990489153191447, 'timing/model': 1.9174022369552404, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 1.58}) +[2026-06-23 18:01:26,938][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.03417159616947174, 'train/grad_norm_pre_clip': 27.704500198364258, 'timing/data': 0.0006126200314611197, 'timing/model': 1.9212241530185565, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 1.58}) +[2026-06-23 18:01:30,728][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.018530840054154396, 'train/grad_norm_pre_clip': 24.74314308166504, 'timing/data': 0.0006491870153695345, 'timing/model': 1.9179346499731764, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 1.58}) +[2026-06-23 18:01:34,533][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.007449822966009378, 'train/grad_norm_pre_clip': 16.355026245117188, 'timing/data': 0.0005997739499434829, 'timing/model': 1.927191826980561, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 1.58}) +[2026-06-23 18:01:38,323][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.05518988147377968, 'train/grad_norm_pre_clip': 15.235873222351074, 'timing/data': 0.0006111320108175278, 'timing/model': 1.9237297599902377, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 1.58}) +[2026-06-23 18:01:42,120][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.021517418324947357, 'train/grad_norm_pre_clip': 21.740203857421875, 'timing/data': 0.0006091530667617917, 'timing/model': 1.918183573987335, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 1.58}) +[2026-06-23 18:01:45,928][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.016886521130800247, 'train/grad_norm_pre_clip': 24.340118408203125, 'timing/data': 0.0005971299251541495, 'timing/model': 1.930537148960866, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 1.58}) +[2026-06-23 18:01:49,721][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.07558216899633408, 'train/grad_norm_pre_clip': 19.25330924987793, 'timing/data': 0.0009350409964099526, 'timing/model': 1.9307874210644513, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 1.58}) +[2026-06-23 18:01:53,508][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.008955879136919975, 'train/grad_norm_pre_clip': 12.313077926635742, 'timing/data': 0.0006091600516811013, 'timing/model': 1.921469078049995, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 1.58}) +[2026-06-23 18:01:57,311][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.018423020839691162, 'train/grad_norm_pre_clip': 36.52277374267578, 'timing/data': 0.0006043449975550175, 'timing/model': 1.9305230049649253, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 1.58}) +[2026-06-23 18:02:01,122][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.15451566874980927, 'train/grad_norm_pre_clip': 39.06120300292969, 'timing/data': 0.0005819620564579964, 'timing/model': 1.9405750299338251, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 1.58}) +[2026-06-23 18:02:04,930][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.040136829018592834, 'train/grad_norm_pre_clip': 19.80026626586914, 'timing/data': 0.0005905020516365767, 'timing/model': 1.9370168750174344, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 1.58}) +[2026-06-23 18:02:08,718][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.11227785795927048, 'train/grad_norm_pre_clip': 56.15669250488281, 'timing/data': 0.0007004499202594161, 'timing/model': 1.9189405760262161, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 1.58}) +[2026-06-23 18:02:12,513][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.042438920587301254, 'train/grad_norm_pre_clip': 37.192466735839844, 'timing/data': 0.0006816589739173651, 'timing/model': 1.9193284090142697, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 1.58}) +[2026-06-23 18:02:16,316][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.07331323623657227, 'train/grad_norm_pre_clip': 66.6741943359375, 'timing/data': 0.0006014350801706314, 'timing/model': 1.923814812907949, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 1.58}) +[2026-06-23 18:02:20,102][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.040670301765203476, 'train/grad_norm_pre_clip': 15.061477661132812, 'timing/data': 0.0006026829360052943, 'timing/model': 1.9181869060266763, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 1.58}) +[2026-06-23 18:02:23,918][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.0358455665409565, 'train/grad_norm_pre_clip': 22.15831756591797, 'timing/data': 0.0006457410054281354, 'timing/model': 1.9341235350584611, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 1.58}) +[2026-06-23 18:02:27,724][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.02631353959441185, 'train/grad_norm_pre_clip': 33.18181228637695, 'timing/data': 0.0006647859700024128, 'timing/model': 1.9252763480180874, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 1.58}) +[2026-06-23 18:02:31,534][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.11353155970573425, 'train/grad_norm_pre_clip': 40.10980224609375, 'timing/data': 0.0005945260636508465, 'timing/model': 1.9286215280881152, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 1.58}) +[2026-06-23 18:02:35,339][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.19688864052295685, 'train/grad_norm_pre_clip': 70.83879089355469, 'timing/data': 0.0009234800236299634, 'timing/model': 1.927549100946635, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 1.58}) +[2026-06-23 18:02:39,149][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.0363137386739254, 'train/grad_norm_pre_clip': 30.585660934448242, 'timing/data': 0.0006000939756631851, 'timing/model': 1.9353594479616731, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 1.58}) +[2026-06-23 18:02:42,945][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.13631640374660492, 'train/grad_norm_pre_clip': 61.79282760620117, 'timing/data': 0.000618733000010252, 'timing/model': 1.9248932730406523, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 1.58}) +[2026-06-23 18:02:46,749][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.0779448002576828, 'train/grad_norm_pre_clip': 70.37895202636719, 'timing/data': 0.0008021249668672681, 'timing/model': 1.9259329680353403, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 1.58}) +[2026-06-23 18:02:50,572][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.013978658244013786, 'train/grad_norm_pre_clip': 12.906850814819336, 'timing/data': 0.0005921749398112297, 'timing/model': 1.9338296098867431, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 1.59}) +[2026-06-23 18:02:54,366][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.09735913574695587, 'train/grad_norm_pre_clip': 37.56953811645508, 'timing/data': 0.0006660779472440481, 'timing/model': 1.92202871106565, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 1.59}) +[2026-06-23 18:02:58,169][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.06774070858955383, 'train/grad_norm_pre_clip': 36.10537338256836, 'timing/data': 0.0006159740732982755, 'timing/model': 1.9281496389303356, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 1.59}) +[2026-06-23 18:03:01,981][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.029533017426729202, 'train/grad_norm_pre_clip': 25.541465759277344, 'timing/data': 0.0006061779567971826, 'timing/model': 1.9276769959833473, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 1.59}) +[2026-06-23 18:03:05,791][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4100 +[2026-06-23 18:05:14,804][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4100 +[2026-06-23 18:05:14,805][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4100 +[2026-06-23 18:06:36,790][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4100 +[2026-06-23 18:06:36,791][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.014511295594274998, 'train/grad_norm_pre_clip': 21.019840240478516, 'eval/action_loss/seconds': 129.00963401410263, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.033243920654058456, 'eval/latency_2/loss': 0.033243920654058456, 'eval/flappy/loss': 0.033243920654058456, 'eval/flappy/latency_2/loss': 0.033243920654058456, 'eval/flap_precision': 0.977366255144033, 'eval/flap_recall': 0.9615384615384616, 'eval/flap_f1': 0.9693877551020408, 'eval/flap_rate_model': 0.2025, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9693877551020407, 'eval/latency_2/flap_precision': 0.977366255144033, 'eval/latency_2/flap_recall': 0.9615384615384616, 'eval/latency_2/flap_f1': 0.9693877551020408, 'eval/latency_2/flap_rate_model': 0.2025, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9693877551020407, 'eval/action_classification/data_wait_seconds': 1.3154186010360718, 'eval/action_classification/predict_seconds': 80.47588348388672, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006159339100122452, 'timing/model': 1.9245497590163723, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 1.59}) +[2026-06-23 18:06:40,575][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.00899251364171505, 'train/grad_norm_pre_clip': 9.133732795715332, 'timing/data': 0.0006010030629113317, 'timing/model': 1.9195851660333574, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 1.59}) +[2026-06-23 18:06:44,355][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.03109084814786911, 'train/grad_norm_pre_clip': 22.120567321777344, 'timing/data': 0.000798347988165915, 'timing/model': 1.9221891560591757, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 1.59}) +[2026-06-23 18:06:48,138][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.020687909796833992, 'train/grad_norm_pre_clip': 37.24490737915039, 'timing/data': 0.0006228559650480747, 'timing/model': 1.9160178329329938, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 1.59}) +[2026-06-23 18:06:51,928][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.056321993470191956, 'train/grad_norm_pre_clip': 27.83388900756836, 'timing/data': 0.0005961210699751973, 'timing/model': 1.9273479050025344, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 1.59}) +[2026-06-23 18:06:55,704][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.047675736248493195, 'train/grad_norm_pre_clip': 31.026464462280273, 'timing/data': 0.0005923230201005936, 'timing/model': 1.9175624520285055, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 1.59}) +[2026-06-23 18:06:59,535][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.09775247424840927, 'train/grad_norm_pre_clip': 44.99135971069336, 'timing/data': 0.0006077890284359455, 'timing/model': 1.9269182520220056, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 1.59}) +[2026-06-23 18:07:03,340][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.023540036752820015, 'train/grad_norm_pre_clip': 37.370506286621094, 'timing/data': 0.0005947210593149066, 'timing/model': 1.922266362933442, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 1.59}) +[2026-06-23 18:07:07,134][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.12893277406692505, 'train/grad_norm_pre_clip': 52.81946563720703, 'timing/data': 0.0006339780520647764, 'timing/model': 1.928306671907194, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 1.59}) +[2026-06-23 18:07:11,056][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.05584106221795082, 'train/grad_norm_pre_clip': 39.12662124633789, 'timing/data': 0.000610293005593121, 'timing/model': 1.9088279000716284, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 1.59}) +[2026-06-23 18:07:14,835][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.019547054544091225, 'train/grad_norm_pre_clip': 29.53141975402832, 'timing/data': 0.0006601359928026795, 'timing/model': 1.9129674000432715, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 1.59}) +[2026-06-23 18:07:18,621][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.006847315933555365, 'train/grad_norm_pre_clip': 22.217395782470703, 'timing/data': 0.0005969830090180039, 'timing/model': 1.9170321440324187, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 1.59}) +[2026-06-23 18:07:22,400][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.0236364733427763, 'train/grad_norm_pre_clip': 21.736543655395508, 'timing/data': 0.0008373371092602611, 'timing/model': 1.916964667965658, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 1.59}) +[2026-06-23 18:07:26,191][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.08350285142660141, 'train/grad_norm_pre_clip': 33.20130157470703, 'timing/data': 0.0005895220674574375, 'timing/model': 1.9245368789415807, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 1.59}) +[2026-06-23 18:07:29,984][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.07120895385742188, 'train/grad_norm_pre_clip': 44.90294647216797, 'timing/data': 0.0006099370075389743, 'timing/model': 1.9223041939549148, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 1.59}) +[2026-06-23 18:07:33,766][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.023707402870059013, 'train/grad_norm_pre_clip': 36.72092819213867, 'timing/data': 0.0006606109673157334, 'timing/model': 1.9243015690008178, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 1.59}) +[2026-06-23 18:07:37,571][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.06272561848163605, 'train/grad_norm_pre_clip': 35.08864212036133, 'timing/data': 0.0006038720021024346, 'timing/model': 1.9375392829533666, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 1.59}) +[2026-06-23 18:07:41,528][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.07130429893732071, 'train/grad_norm_pre_clip': 29.496978759765625, 'timing/data': 0.0006018689600750804, 'timing/model': 2.100538778002374, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 1.59}) +[2026-06-23 18:07:45,308][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.03952223062515259, 'train/grad_norm_pre_clip': 46.789306640625, 'timing/data': 0.0006913190009072423, 'timing/model': 1.912028933991678, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 1.59}) +[2026-06-23 18:07:49,102][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.07054319977760315, 'train/grad_norm_pre_clip': 26.251968383789062, 'timing/data': 0.0008531569037586451, 'timing/model': 1.9218378469813615, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 1.59}) +[2026-06-23 18:07:52,896][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.0659436360001564, 'train/grad_norm_pre_clip': 29.46767234802246, 'timing/data': 0.0006526620127260685, 'timing/model': 1.9303934349445626, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 1.59}) +[2026-06-23 18:07:56,681][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.016658954322338104, 'train/grad_norm_pre_clip': 50.05878448486328, 'timing/data': 0.0006086089415475726, 'timing/model': 1.9177599960239604, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 1.59}) +[2026-06-23 18:08:00,478][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.08510562777519226, 'train/grad_norm_pre_clip': 22.476322174072266, 'timing/data': 0.0005985310999676585, 'timing/model': 1.9233681900659576, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 1.6}) +[2026-06-23 18:08:04,270][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.03618388995528221, 'train/grad_norm_pre_clip': 29.858455657958984, 'timing/data': 0.0007393850246444345, 'timing/model': 1.9299186829011887, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 1.6}) +[2026-06-23 18:08:08,061][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.17111042141914368, 'train/grad_norm_pre_clip': 43.87325668334961, 'timing/data': 0.0006090609822422266, 'timing/model': 1.9286565450020134, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 1.6}) +[2026-06-23 18:08:11,853][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.025126954540610313, 'train/grad_norm_pre_clip': 28.94317626953125, 'timing/data': 0.0006033149547874928, 'timing/model': 1.9212813220219687, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 1.6}) +[2026-06-23 18:08:15,829][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.2176554799079895, 'train/grad_norm_pre_clip': 60.703460693359375, 'timing/data': 0.0006961750332266092, 'timing/model': 1.928396994015202, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 1.6}) +[2026-06-23 18:08:19,639][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.06724609434604645, 'train/grad_norm_pre_clip': 30.79776382446289, 'timing/data': 0.0006042219465598464, 'timing/model': 1.9318463909439743, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 1.6}) +[2026-06-23 18:08:23,432][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.009996341541409492, 'train/grad_norm_pre_clip': 9.018659591674805, 'timing/data': 0.000594518962316215, 'timing/model': 1.919194129994139, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 1.6}) +[2026-06-23 18:08:27,238][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.21752217411994934, 'train/grad_norm_pre_clip': 51.139488220214844, 'timing/data': 0.0008689829846844077, 'timing/model': 1.9330940999789163, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 1.6}) +[2026-06-23 18:08:31,091][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.017712337896227837, 'train/grad_norm_pre_clip': 23.247520446777344, 'timing/data': 0.0006048069335520267, 'timing/model': 1.933693516999483, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 1.6}) +[2026-06-23 18:08:34,900][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.05166452378034592, 'train/grad_norm_pre_clip': 23.647066116333008, 'timing/data': 0.0006933779222890735, 'timing/model': 1.9365771929733455, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 1.6}) +[2026-06-23 18:08:38,705][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.044184610247612, 'train/grad_norm_pre_clip': 17.611282348632812, 'timing/data': 0.0005843550898134708, 'timing/model': 1.9334890090394765, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 1.6}) +[2026-06-23 18:08:42,499][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.05053417757153511, 'train/grad_norm_pre_clip': 25.503416061401367, 'timing/data': 0.0006430140929296613, 'timing/model': 1.927143262932077, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 1.6}) +[2026-06-23 18:08:46,298][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.03334923833608627, 'train/grad_norm_pre_clip': 27.05404281616211, 'timing/data': 0.0007774629630148411, 'timing/model': 1.9241705388994887, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 1.6}) +[2026-06-23 18:08:50,260][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.02570231631398201, 'train/grad_norm_pre_clip': 16.3689022064209, 'timing/data': 0.0008590940851718187, 'timing/model': 1.9248437259811908, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 1.6}) +[2026-06-23 18:08:54,066][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.06534141302108765, 'train/grad_norm_pre_clip': 31.235605239868164, 'timing/data': 0.0006004150491207838, 'timing/model': 1.9353352470789105, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 1.6}) +[2026-06-23 18:08:57,866][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.03326115757226944, 'train/grad_norm_pre_clip': 24.081588745117188, 'timing/data': 0.0006283399416133761, 'timing/model': 1.9318040170473978, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 1.6}) +[2026-06-23 18:09:01,656][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.054995097219944, 'train/grad_norm_pre_clip': 36.60055923461914, 'timing/data': 0.0009189120028167963, 'timing/model': 1.9241420049220324, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 1.6}) +[2026-06-23 18:09:05,486][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.11579926311969757, 'train/grad_norm_pre_clip': 34.32904815673828, 'timing/data': 0.0006623530061915517, 'timing/model': 1.9511676570400596, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 1.6}) +[2026-06-23 18:09:09,307][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.0215739943087101, 'train/grad_norm_pre_clip': 9.544407844543457, 'timing/data': 0.0007004460785537958, 'timing/model': 1.9566062400117517, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 1.6}) +[2026-06-23 18:09:13,119][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.024919910356402397, 'train/grad_norm_pre_clip': 25.05171012878418, 'timing/data': 0.000697385985404253, 'timing/model': 1.9297216290142387, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 1.6}) +[2026-06-23 18:09:16,928][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.08870860189199448, 'train/grad_norm_pre_clip': 40.44801712036133, 'timing/data': 0.0005986019968986511, 'timing/model': 1.9392984919250011, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 1.6}) +[2026-06-23 18:09:20,916][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.10032372176647186, 'train/grad_norm_pre_clip': 29.102680206298828, 'timing/data': 0.0006442359881475568, 'timing/model': 2.1230238690041006, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 1.6}) +[2026-06-23 18:09:24,714][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.036511410027742386, 'train/grad_norm_pre_clip': 21.51289939880371, 'timing/data': 0.0006026360206305981, 'timing/model': 1.9343319550389424, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 1.6}) +[2026-06-23 18:09:28,553][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.10906442999839783, 'train/grad_norm_pre_clip': 43.73537826538086, 'timing/data': 0.0006062419852241874, 'timing/model': 1.9656116439728066, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 1.6}) +[2026-06-23 18:09:32,359][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.05074124038219452, 'train/grad_norm_pre_clip': 39.193031311035156, 'timing/data': 0.000680859899148345, 'timing/model': 1.9374018860980868, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 1.6}) +[2026-06-23 18:09:36,171][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.10552575439214706, 'train/grad_norm_pre_clip': 41.9100341796875, 'timing/data': 0.0006041129818186164, 'timing/model': 1.9328042110428214, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 1.6}) +[2026-06-23 18:09:39,974][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.018819913268089294, 'train/grad_norm_pre_clip': 24.09688949584961, 'timing/data': 0.000856642029248178, 'timing/model': 1.9312712930841371, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 1.61}) +[2026-06-23 18:09:43,783][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.13139529526233673, 'train/grad_norm_pre_clip': 24.520767211914062, 'timing/data': 0.000708002015016973, 'timing/model': 1.929735800018534, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 1.61}) +[2026-06-23 18:09:47,603][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.11200615763664246, 'train/grad_norm_pre_clip': 31.480979919433594, 'timing/data': 0.0005909790052101016, 'timing/model': 1.9373845850350335, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 1.61}) +[2026-06-23 18:09:51,404][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.039870522916316986, 'train/grad_norm_pre_clip': 21.10886001586914, 'timing/data': 0.0005948729813098907, 'timing/model': 1.924894168972969, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 1.61}) +[2026-06-23 18:09:55,193][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.048980146646499634, 'train/grad_norm_pre_clip': 27.929546356201172, 'timing/data': 0.000639398000203073, 'timing/model': 1.927395366015844, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 1.61}) +[2026-06-23 18:09:59,000][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.04908245801925659, 'train/grad_norm_pre_clip': 26.419103622436523, 'timing/data': 0.0006776349619030952, 'timing/model': 1.9354564999230206, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 1.61}) +[2026-06-23 18:10:02,803][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.057098567485809326, 'train/grad_norm_pre_clip': 24.182252883911133, 'timing/data': 0.0006026830524206161, 'timing/model': 1.925309563986957, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 1.61}) +[2026-06-23 18:10:06,604][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.10264600813388824, 'train/grad_norm_pre_clip': 30.2038516998291, 'timing/data': 0.00071797298733145, 'timing/model': 1.9230174970580265, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 1.61}) +[2026-06-23 18:10:10,412][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.0582747720181942, 'train/grad_norm_pre_clip': 16.980297088623047, 'timing/data': 0.000743580050766468, 'timing/model': 1.9273726230021566, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 1.61}) +[2026-06-23 18:10:14,214][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.07681348919868469, 'train/grad_norm_pre_clip': 30.163793563842773, 'timing/data': 0.0006062850588932633, 'timing/model': 1.931107635027729, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 1.61}) +[2026-06-23 18:10:18,020][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.017258064821362495, 'train/grad_norm_pre_clip': 15.258011817932129, 'timing/data': 0.0006102779880166054, 'timing/model': 1.9272140749963, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 1.61}) +[2026-06-23 18:10:21,837][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.03424987941980362, 'train/grad_norm_pre_clip': 33.15890121459961, 'timing/data': 0.0006418779958039522, 'timing/model': 1.938987719011493, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 1.61}) +[2026-06-23 18:10:25,637][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.0874742642045021, 'train/grad_norm_pre_clip': 32.516292572021484, 'timing/data': 0.0006718180375173688, 'timing/model': 1.9319979599677026, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 1.61}) +[2026-06-23 18:10:29,444][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.06612513214349747, 'train/grad_norm_pre_clip': 36.64091491699219, 'timing/data': 0.0006029990036040545, 'timing/model': 1.9293049549451098, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 1.61}) +[2026-06-23 18:10:33,245][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.05069594457745552, 'train/grad_norm_pre_clip': 32.69374465942383, 'timing/data': 0.0006635779282078147, 'timing/model': 1.9298920780420303, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 1.61}) +[2026-06-23 18:10:37,043][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.023803742602467537, 'train/grad_norm_pre_clip': 12.347826957702637, 'timing/data': 0.0005917759845033288, 'timing/model': 1.9290106269763783, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 1.61}) +[2026-06-23 18:10:40,844][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.06307034194469452, 'train/grad_norm_pre_clip': 32.875389099121094, 'timing/data': 0.0005849570734426379, 'timing/model': 1.9285314020235091, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 1.61}) +[2026-06-23 18:10:44,646][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.0481925792992115, 'train/grad_norm_pre_clip': 20.231857299804688, 'timing/data': 0.0005960710113868117, 'timing/model': 1.928163968026638, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 1.61}) +[2026-06-23 18:10:48,440][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.06280067563056946, 'train/grad_norm_pre_clip': 22.165361404418945, 'timing/data': 0.0005911190528422594, 'timing/model': 1.9252196389716119, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 1.61}) +[2026-06-23 18:10:52,253][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.01968526840209961, 'train/grad_norm_pre_clip': 35.17378616333008, 'timing/data': 0.0006025949260219932, 'timing/model': 1.9372388499323279, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 1.61}) +[2026-06-23 18:10:56,071][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.019843367859721184, 'train/grad_norm_pre_clip': 14.79384994506836, 'timing/data': 0.0005922910058870912, 'timing/model': 1.942685705027543, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 1.61}) +[2026-06-23 18:11:00,051][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.035371262580156326, 'train/grad_norm_pre_clip': 25.634401321411133, 'timing/data': 0.0005987499607726932, 'timing/model': 1.9269416199531406, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 1.61}) +[2026-06-23 18:11:03,852][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.03731090947985649, 'train/grad_norm_pre_clip': 35.27741622924805, 'timing/data': 0.000593801960349083, 'timing/model': 1.9266608140897006, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 1.61}) +[2026-06-23 18:11:07,683][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.018739424645900726, 'train/grad_norm_pre_clip': 20.18438720703125, 'timing/data': 0.0005930219776928425, 'timing/model': 1.9657899499870837, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 1.61}) +[2026-06-23 18:11:11,476][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.10869915783405304, 'train/grad_norm_pre_clip': 52.5302848815918, 'timing/data': 0.0006158009637147188, 'timing/model': 1.9311088190879673, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 1.61}) +[2026-06-23 18:11:15,285][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.09622782468795776, 'train/grad_norm_pre_clip': 43.97411346435547, 'timing/data': 0.0010020789923146367, 'timing/model': 1.9385396799771115, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 1.62}) +[2026-06-23 18:11:19,121][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.1558515876531601, 'train/grad_norm_pre_clip': 31.29532814025879, 'timing/data': 0.0006238359492272139, 'timing/model': 1.9639478250173852, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 1.62}) +[2026-06-23 18:11:22,928][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.015598180703818798, 'train/grad_norm_pre_clip': 17.80440902709961, 'timing/data': 0.0005943669239059091, 'timing/model': 1.9315371769480407, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 1.62}) +[2026-06-23 18:11:26,741][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.10145732760429382, 'train/grad_norm_pre_clip': 21.050920486450195, 'timing/data': 0.0007844110950827599, 'timing/model': 1.9405296290060505, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 1.62}) +[2026-06-23 18:11:30,534][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.053309131413698196, 'train/grad_norm_pre_clip': 47.39828872680664, 'timing/data': 0.0006112540140748024, 'timing/model': 1.923767733038403, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 1.62}) +[2026-06-23 18:11:34,502][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.08176495134830475, 'train/grad_norm_pre_clip': 50.16705322265625, 'timing/data': 0.0006240289658308029, 'timing/model': 1.9378182429354638, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 1.62}) +[2026-06-23 18:11:38,307][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.18454629182815552, 'train/grad_norm_pre_clip': 37.886714935302734, 'timing/data': 0.0005868779262527823, 'timing/model': 1.9261341640958562, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 1.62}) +[2026-06-23 18:11:42,120][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.1813780665397644, 'train/grad_norm_pre_clip': 62.28710174560547, 'timing/data': 0.0005881930701434612, 'timing/model': 1.9412249230081216, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 1.62}) +[2026-06-23 18:11:45,926][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.024708790704607964, 'train/grad_norm_pre_clip': 28.031429290771484, 'timing/data': 0.0005980810383334756, 'timing/model': 1.9413426740793511, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 1.62}) +[2026-06-23 18:11:49,731][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.02486622892320156, 'train/grad_norm_pre_clip': 33.237369537353516, 'timing/data': 0.0006274530896916986, 'timing/model': 1.9309547470184043, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 1.62}) +[2026-06-23 18:11:53,531][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.016335422173142433, 'train/grad_norm_pre_clip': 14.203858375549316, 'timing/data': 0.0006193530280143023, 'timing/model': 1.9173078479943797, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 1.62}) +[2026-06-23 18:11:57,325][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.04631587490439415, 'train/grad_norm_pre_clip': 25.577346801757812, 'timing/data': 0.0006083759944885969, 'timing/model': 1.9257002460071817, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 1.62}) +[2026-06-23 18:12:01,125][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.15276047587394714, 'train/grad_norm_pre_clip': 32.4330940246582, 'timing/data': 0.0006071709794923663, 'timing/model': 1.9349071249598637, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 1.62}) +[2026-06-23 18:12:05,066][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.027598347514867783, 'train/grad_norm_pre_clip': 34.32927322387695, 'timing/data': 0.0005887700244784355, 'timing/model': 2.067976407939568, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 1.62}) +[2026-06-23 18:12:08,876][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.05298051983118057, 'train/grad_norm_pre_clip': 24.74895668029785, 'timing/data': 0.0005930259358137846, 'timing/model': 1.9303836490726098, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 1.62}) +[2026-06-23 18:12:12,687][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.02988857589662075, 'train/grad_norm_pre_clip': 25.266937255859375, 'timing/data': 0.0005887179868295789, 'timing/model': 1.9358967159641907, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 1.62}) +[2026-06-23 18:12:16,490][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.04500285163521767, 'train/grad_norm_pre_clip': 32.96894073486328, 'timing/data': 0.0005991678917780519, 'timing/model': 1.9342764629982412, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 1.62}) +[2026-06-23 18:12:20,294][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.06751731038093567, 'train/grad_norm_pre_clip': 20.388391494750977, 'timing/data': 0.0005982340080663562, 'timing/model': 1.9266357090091333, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 1.62}) +[2026-06-23 18:12:24,116][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.047651346772909164, 'train/grad_norm_pre_clip': 27.600595474243164, 'timing/data': 0.0006084089400246739, 'timing/model': 1.9287901379866526, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 1.62}) +[2026-06-23 18:12:27,925][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.03129267692565918, 'train/grad_norm_pre_clip': 15.874306678771973, 'timing/data': 0.0010805740021169186, 'timing/model': 1.930560428998433, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 1.62}) +[2026-06-23 18:12:31,751][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.011587132699787617, 'train/grad_norm_pre_clip': 24.09800910949707, 'timing/data': 0.0007356860442087054, 'timing/model': 1.9466458219103515, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 1.62}) +[2026-06-23 18:12:35,549][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.10254751145839691, 'train/grad_norm_pre_clip': 45.23886489868164, 'timing/data': 0.0006180510390549898, 'timing/model': 1.9253943440271541, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 1.62}) +[2026-06-23 18:12:39,469][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.10685646533966064, 'train/grad_norm_pre_clip': 23.359619140625, 'timing/data': 0.0006330490577965975, 'timing/model': 1.9345148550346494, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 1.62}) +[2026-06-23 18:12:43,267][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.04204288870096207, 'train/grad_norm_pre_clip': 23.4793758392334, 'timing/data': 0.0006114080315455794, 'timing/model': 1.9258065900066867, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 1.62}) +[2026-06-23 18:12:47,070][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.09003346413373947, 'train/grad_norm_pre_clip': 45.290828704833984, 'timing/data': 0.0006226050900295377, 'timing/model': 1.93696863099467, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 1.62}) +[2026-06-23 18:12:50,875][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.09727494418621063, 'train/grad_norm_pre_clip': 53.55119323730469, 'timing/data': 0.0006566069787368178, 'timing/model': 1.9306931889150292, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 1.62}) +[2026-06-23 18:12:54,702][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.008274017833173275, 'train/grad_norm_pre_clip': 8.515427589416504, 'timing/data': 0.0005829980364069343, 'timing/model': 1.9332909469958395, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 1.63}) +[2026-06-23 18:12:58,513][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4200 +[2026-06-23 18:15:07,358][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4200 +[2026-06-23 18:15:07,359][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4200 +[2026-06-23 18:16:29,754][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4200 +[2026-06-23 18:16:29,755][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.07528264820575714, 'train/grad_norm_pre_clip': 58.37394714355469, 'eval/action_loss/seconds': 128.84212329308502, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03191108629107475, 'eval/latency_2/loss': 0.03191108629107475, 'eval/flappy/loss': 0.03191108629107475, 'eval/flappy/latency_2/loss': 0.03191108629107475, 'eval/flap_precision': 0.9638312123241795, 'eval/flap_recall': 0.9709851551956815, 'eval/flap_f1': 0.9673949579831934, 'eval/flap_rate_model': 0.2073611111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9673949579831933, 'eval/latency_2/flap_precision': 0.9638312123241795, 'eval/latency_2/flap_recall': 0.9709851551956815, 'eval/latency_2/flap_f1': 0.9673949579831934, 'eval/latency_2/flap_rate_model': 0.2073611111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9673949579831933, 'eval/action_classification/data_wait_seconds': 1.2461299896240234, 'eval/action_classification/predict_seconds': 80.95474243164062, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006107459776103497, 'timing/model': 1.937361766002141, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 1.63}) +[2026-06-23 18:16:33,540][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.06284671276807785, 'train/grad_norm_pre_clip': 43.19416427612305, 'timing/data': 0.0005909890169277787, 'timing/model': 1.9174188000615686, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 1.63}) +[2026-06-23 18:16:37,321][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.01247318647801876, 'train/grad_norm_pre_clip': 5.299070358276367, 'timing/data': 0.0007764389738440514, 'timing/model': 1.9151598459575325, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 1.63}) +[2026-06-23 18:16:41,104][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.05750355124473572, 'train/grad_norm_pre_clip': 22.38214111328125, 'timing/data': 0.000607294961810112, 'timing/model': 1.919375783065334, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 1.63}) +[2026-06-23 18:16:44,879][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.05205278471112251, 'train/grad_norm_pre_clip': 14.263833999633789, 'timing/data': 0.0006116339936852455, 'timing/model': 1.912432854063809, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 1.63}) +[2026-06-23 18:16:48,663][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.04630621522665024, 'train/grad_norm_pre_clip': 16.964021682739258, 'timing/data': 0.0005988810444250703, 'timing/model': 1.9178938580444083, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 1.63}) +[2026-06-23 18:16:52,464][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.006977559998631477, 'train/grad_norm_pre_clip': 16.607080459594727, 'timing/data': 0.0006112690316513181, 'timing/model': 1.9193326119566336, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 1.63}) +[2026-06-23 18:16:56,271][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.00521242106333375, 'train/grad_norm_pre_clip': 17.908260345458984, 'timing/data': 0.0006079879822209477, 'timing/model': 1.9372645559487864, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 1.63}) +[2026-06-23 18:17:00,247][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.08704466372728348, 'train/grad_norm_pre_clip': 37.99308395385742, 'timing/data': 0.0006156569579616189, 'timing/model': 2.109632539912127, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 1.63}) +[2026-06-23 18:17:04,044][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.0596269927918911, 'train/grad_norm_pre_clip': 37.27086639404297, 'timing/data': 0.0006085670320317149, 'timing/model': 1.9271602049702778, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 1.63}) +[2026-06-23 18:17:07,829][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.03914526477456093, 'train/grad_norm_pre_clip': 27.660871505737305, 'timing/data': 0.0006007390329614282, 'timing/model': 1.9202891119057313, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 1.63}) +[2026-06-23 18:17:11,614][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.034046802669763565, 'train/grad_norm_pre_clip': 30.57302474975586, 'timing/data': 0.0008863499388098717, 'timing/model': 1.9257839309284464, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 1.63}) +[2026-06-23 18:17:15,420][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.04469883814454079, 'train/grad_norm_pre_clip': 19.240108489990234, 'timing/data': 0.0006004789611324668, 'timing/model': 1.9437271460192278, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 1.63}) +[2026-06-23 18:17:19,219][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.07741615921258926, 'train/grad_norm_pre_clip': 28.482358932495117, 'timing/data': 0.0008411890594288707, 'timing/model': 1.9277936579892412, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 1.63}) +[2026-06-23 18:17:23,019][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.00963003933429718, 'train/grad_norm_pre_clip': 27.335346221923828, 'timing/data': 0.0006037229904904962, 'timing/model': 1.9322010630276054, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 1.63}) +[2026-06-23 18:17:26,804][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.045964717864990234, 'train/grad_norm_pre_clip': 43.0438117980957, 'timing/data': 0.0009222910739481449, 'timing/model': 1.9231914319097996, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 1.63}) +[2026-06-23 18:17:30,617][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.051198188215494156, 'train/grad_norm_pre_clip': 22.53057289123535, 'timing/data': 0.0007455000886693597, 'timing/model': 1.9356146689970046, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 1.63}) +[2026-06-23 18:17:34,421][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.04206133633852005, 'train/grad_norm_pre_clip': 26.367319107055664, 'timing/data': 0.000598215963691473, 'timing/model': 1.9228145690867677, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 1.63}) +[2026-06-23 18:17:38,209][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.06261392682790756, 'train/grad_norm_pre_clip': 38.297000885009766, 'timing/data': 0.0008499289397150278, 'timing/model': 1.9196669929660857, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 1.63}) +[2026-06-23 18:17:42,007][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.08444525301456451, 'train/grad_norm_pre_clip': 35.42120361328125, 'timing/data': 0.0007705949246883392, 'timing/model': 1.9320578830083832, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 1.63}) +[2026-06-23 18:17:45,809][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.13261385262012482, 'train/grad_norm_pre_clip': 35.978919982910156, 'timing/data': 0.0006619549822062254, 'timing/model': 1.919464375008829, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 1.63}) +[2026-06-23 18:17:49,612][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.08554623275995255, 'train/grad_norm_pre_clip': 41.22150802612305, 'timing/data': 0.0005908110179007053, 'timing/model': 1.9352650989312679, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 1.63}) +[2026-06-23 18:17:53,414][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.07351665198802948, 'train/grad_norm_pre_clip': 35.50274658203125, 'timing/data': 0.0005867989966645837, 'timing/model': 1.9276477430248633, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 1.63}) +[2026-06-23 18:17:57,201][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.025722283869981766, 'train/grad_norm_pre_clip': 50.938262939453125, 'timing/data': 0.0005851039895787835, 'timing/model': 1.9175351980375126, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 1.63}) +[2026-06-23 18:18:01,016][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.06645162403583527, 'train/grad_norm_pre_clip': 38.72098159790039, 'timing/data': 0.0005947049940004945, 'timing/model': 1.9384793960489333, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 1.63}) +[2026-06-23 18:18:04,823][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.029880233108997345, 'train/grad_norm_pre_clip': 21.68494415283203, 'timing/data': 0.0005899459356442094, 'timing/model': 1.9261604630155489, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 1.64}) +[2026-06-23 18:18:08,621][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.009202113375067711, 'train/grad_norm_pre_clip': 56.38971710205078, 'timing/data': 0.0006556590087711811, 'timing/model': 1.9293395829154179, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 1.64}) +[2026-06-23 18:18:12,412][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.06536295264959335, 'train/grad_norm_pre_clip': 39.47813034057617, 'timing/data': 0.0007495960453525186, 'timing/model': 1.9185854920651764, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 1.64}) +[2026-06-23 18:18:16,208][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.011515047401189804, 'train/grad_norm_pre_clip': 32.2322883605957, 'timing/data': 0.0007615069625899196, 'timing/model': 1.927380840992555, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 1.64}) +[2026-06-23 18:18:20,016][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.03243746981024742, 'train/grad_norm_pre_clip': 41.91875457763672, 'timing/data': 0.0008162549929693341, 'timing/model': 1.9255875569069758, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 1.64}) +[2026-06-23 18:18:23,812][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.10159716755151749, 'train/grad_norm_pre_clip': 55.487857818603516, 'timing/data': 0.0007070869905874133, 'timing/model': 1.9177933820756152, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 1.64}) +[2026-06-23 18:18:27,643][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.09103677421808243, 'train/grad_norm_pre_clip': 43.7663459777832, 'timing/data': 0.0008062220877036452, 'timing/model': 1.9519977909512818, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 1.64}) +[2026-06-23 18:18:31,436][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.038978755474090576, 'train/grad_norm_pre_clip': 91.4810562133789, 'timing/data': 0.0006087030051276088, 'timing/model': 1.9281832510605454, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 1.64}) +[2026-06-23 18:18:35,237][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.07116983085870743, 'train/grad_norm_pre_clip': 31.12044906616211, 'timing/data': 0.0005953639047220349, 'timing/model': 1.9297678059665486, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 1.64}) +[2026-06-23 18:18:39,030][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.0609421543776989, 'train/grad_norm_pre_clip': 27.056333541870117, 'timing/data': 0.0005921500269323587, 'timing/model': 1.9203380619874224, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 1.64}) +[2026-06-23 18:18:42,846][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.02974051609635353, 'train/grad_norm_pre_clip': 29.750619888305664, 'timing/data': 0.0006136189913377166, 'timing/model': 1.9317312618950382, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 1.64}) +[2026-06-23 18:18:46,639][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.010546749457716942, 'train/grad_norm_pre_clip': 16.825613021850586, 'timing/data': 0.0006617989856749773, 'timing/model': 1.918657566071488, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 1.64}) +[2026-06-23 18:18:50,428][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.0781424269080162, 'train/grad_norm_pre_clip': 13.483263969421387, 'timing/data': 0.0006074750563129783, 'timing/model': 1.920627843006514, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 1.64}) +[2026-06-23 18:18:54,231][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.12985093891620636, 'train/grad_norm_pre_clip': 33.578182220458984, 'timing/data': 0.0006430789362639189, 'timing/model': 1.9348583050305024, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 1.64}) +[2026-06-23 18:18:58,020][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.048203762620687485, 'train/grad_norm_pre_clip': 27.359874725341797, 'timing/data': 0.0006135640433058143, 'timing/model': 1.92667153198272, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 1.64}) +[2026-06-23 18:19:01,818][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.044692981988191605, 'train/grad_norm_pre_clip': 38.30820083618164, 'timing/data': 0.0006270098965615034, 'timing/model': 1.9305086210370064, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 1.64}) +[2026-06-23 18:19:05,610][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.03774339705705643, 'train/grad_norm_pre_clip': 16.175111770629883, 'timing/data': 0.0007505599642172456, 'timing/model': 1.9273858410306275, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 1.64}) +[2026-06-23 18:19:09,582][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.04951612278819084, 'train/grad_norm_pre_clip': 30.6693115234375, 'timing/data': 0.0005880859680473804, 'timing/model': 2.101139153004624, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 1.64}) +[2026-06-23 18:19:13,379][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.01741853915154934, 'train/grad_norm_pre_clip': 9.86081314086914, 'timing/data': 0.0006420029094442725, 'timing/model': 1.9209185240324587, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 1.64}) +[2026-06-23 18:19:17,180][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.015116902068257332, 'train/grad_norm_pre_clip': 26.1320743560791, 'timing/data': 0.0006073679542168975, 'timing/model': 1.9242248419905081, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 1.64}) +[2026-06-23 18:19:20,992][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.02552901953458786, 'train/grad_norm_pre_clip': 16.361238479614258, 'timing/data': 0.00086952093988657, 'timing/model': 1.9345254770014435, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 1.64}) +[2026-06-23 18:19:24,797][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.08818012475967407, 'train/grad_norm_pre_clip': 30.886587142944336, 'timing/data': 0.0006037980783730745, 'timing/model': 1.926590400049463, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 1.64}) +[2026-06-23 18:19:28,606][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.03521279990673065, 'train/grad_norm_pre_clip': 39.2756233215332, 'timing/data': 0.00060562405269593, 'timing/model': 1.9360829569632187, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 1.64}) +[2026-06-23 18:19:32,405][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.09756455570459366, 'train/grad_norm_pre_clip': 33.600826263427734, 'timing/data': 0.0005843989783897996, 'timing/model': 1.9290889180265367, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 1.64}) +[2026-06-23 18:19:36,223][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.057682670652866364, 'train/grad_norm_pre_clip': 29.100873947143555, 'timing/data': 0.0008560279384255409, 'timing/model': 1.9188777779927477, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 1.64}) +[2026-06-23 18:28:22,745][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4250 +[2026-06-23 18:28:22,746][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.07773426920175552, 'train/grad_norm_pre_clip': 23.224279403686523, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 321.42000000001417, 'rl_games_eval/mid_train/mean_length': 2608.4, 'rl_games_eval/mid_train/std_reward': 159.47787808973385, 'rl_games_eval/mid_train/std_length': 1284.1907334971702, 'rl_games_eval/mid_train/macro_mean_reward': 321.42000000001417, 'rl_games_eval/mid_train/macro_mean_length': 2608.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 321.42000000001417, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2608.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 159.47787808973385, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1284.1907334971702, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 321.42000000001417, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006146980449557304, 'timing/model': 1.9249761711107567, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 1.64}) +[2026-06-23 18:28:26,474][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.02253798581659794, 'train/grad_norm_pre_clip': 33.8865852355957, 'timing/data': 0.000827350071631372, 'timing/model': 1.8953199819661677, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 1.65}) +[2026-06-23 18:28:30,207][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.10154546052217484, 'train/grad_norm_pre_clip': 114.93656158447266, 'timing/data': 0.0009634799789637327, 'timing/model': 1.8962092890869826, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 1.65}) +[2026-06-23 18:28:33,976][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.0285724475979805, 'train/grad_norm_pre_clip': 19.405736923217773, 'timing/data': 0.0005954650696367025, 'timing/model': 1.8975171140627936, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 1.65}) +[2026-06-23 18:28:37,738][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.0899602621793747, 'train/grad_norm_pre_clip': 32.04130172729492, 'timing/data': 0.000597730977460742, 'timing/model': 1.9146320919971913, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 1.65}) +[2026-06-23 18:28:41,667][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.09586796164512634, 'train/grad_norm_pre_clip': 31.63654327392578, 'timing/data': 0.0005817790515720844, 'timing/model': 2.0837391290115193, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 1.65}) +[2026-06-23 18:28:45,410][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.0680125430226326, 'train/grad_norm_pre_clip': 25.255596160888672, 'timing/data': 0.000593140022829175, 'timing/model': 1.897853764006868, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 1.65}) +[2026-06-23 18:28:49,159][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.008924351073801517, 'train/grad_norm_pre_clip': 10.849451065063477, 'timing/data': 0.0006280710222199559, 'timing/model': 1.9064332679845393, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 1.65}) +[2026-06-23 18:28:52,920][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.014156538993120193, 'train/grad_norm_pre_clip': 6.942197799682617, 'timing/data': 0.0008168390486389399, 'timing/model': 1.9080297889886424, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 1.65}) +[2026-06-23 18:28:56,700][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.019452417269349098, 'train/grad_norm_pre_clip': 12.059853553771973, 'timing/data': 0.0005976200336590409, 'timing/model': 1.920091820997186, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 1.65}) +[2026-06-23 18:29:00,460][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.0360431931912899, 'train/grad_norm_pre_clip': 24.935302734375, 'timing/data': 0.0009095889981836081, 'timing/model': 1.905663117999211, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 1.65}) +[2026-06-23 18:29:04,227][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.024720678105950356, 'train/grad_norm_pre_clip': 15.702764511108398, 'timing/data': 0.0005990569479763508, 'timing/model': 1.9147586789913476, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 1.65}) +[2026-06-23 18:29:08,008][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.05649328604340553, 'train/grad_norm_pre_clip': 48.22124481201172, 'timing/data': 0.000588082941249013, 'timing/model': 1.9253263010177761, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 1.65}) +[2026-06-23 18:29:11,818][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.031737346202135086, 'train/grad_norm_pre_clip': 18.26348114013672, 'timing/data': 0.000591974938288331, 'timing/model': 1.94775454106275, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 1.65}) +[2026-06-23 18:29:15,596][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.029852593317627907, 'train/grad_norm_pre_clip': 34.306358337402344, 'timing/data': 0.0006464930484071374, 'timing/model': 1.923523616977036, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 1.65}) +[2026-06-23 18:29:19,382][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.08776643872261047, 'train/grad_norm_pre_clip': 36.02268600463867, 'timing/data': 0.0005851340247318149, 'timing/model': 1.9272807739907876, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 1.65}) +[2026-06-23 18:29:23,161][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.010607203468680382, 'train/grad_norm_pre_clip': 20.360965728759766, 'timing/data': 0.0006190850399434566, 'timing/model': 1.9248439579969272, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 1.65}) +[2026-06-23 18:29:26,943][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.07713768631219864, 'train/grad_norm_pre_clip': 50.696563720703125, 'timing/data': 0.0005984880262985826, 'timing/model': 1.9113702269969508, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 1.65}) +[2026-06-23 18:29:30,723][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.08662610501050949, 'train/grad_norm_pre_clip': 35.58966064453125, 'timing/data': 0.0007425900548696518, 'timing/model': 1.9217557270312682, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 1.65}) +[2026-06-23 18:29:34,513][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.028308581560850143, 'train/grad_norm_pre_clip': 23.853321075439453, 'timing/data': 0.0006080589955672622, 'timing/model': 1.9209876139648259, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 1.65}) +[2026-06-23 18:29:38,294][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.03849443420767784, 'train/grad_norm_pre_clip': 30.860322952270508, 'timing/data': 0.0008817219641059637, 'timing/model': 1.921776706003584, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 1.65}) +[2026-06-23 18:29:42,074][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.0638357475399971, 'train/grad_norm_pre_clip': 24.911174774169922, 'timing/data': 0.000595084042288363, 'timing/model': 1.9213297900278121, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 1.65}) +[2026-06-23 18:29:46,100][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.05141512677073479, 'train/grad_norm_pre_clip': 34.545066833496094, 'timing/data': 0.0006592120043933392, 'timing/model': 2.0353818769799545, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 1.65}) +[2026-06-23 18:29:49,887][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.02605714090168476, 'train/grad_norm_pre_clip': 19.60219955444336, 'timing/data': 0.0006201280048117042, 'timing/model': 1.9180169369792566, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 1.65}) +[2026-06-23 18:29:53,686][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.05716416612267494, 'train/grad_norm_pre_clip': 52.26288986206055, 'timing/data': 0.0007325049955397844, 'timing/model': 1.9300473319599405, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 1.65}) +[2026-06-23 18:29:57,481][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.20340704917907715, 'train/grad_norm_pre_clip': 49.65177536010742, 'timing/data': 0.0006035249680280685, 'timing/model': 1.9295381460106, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 1.65}) +[2026-06-23 18:30:01,273][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.13922543823719025, 'train/grad_norm_pre_clip': 31.907018661499023, 'timing/data': 0.000618630088865757, 'timing/model': 1.9284457550384104, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 1.65}) +[2026-06-23 18:30:05,075][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.12234361469745636, 'train/grad_norm_pre_clip': 37.380149841308594, 'timing/data': 0.0005859209923073649, 'timing/model': 1.923301124945283, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 1.66}) +[2026-06-23 18:30:08,865][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.0412026010453701, 'train/grad_norm_pre_clip': 27.996185302734375, 'timing/data': 0.0008560009300708771, 'timing/model': 1.9190191749949008, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 1.66}) +[2026-06-23 18:30:12,652][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.03116089478135109, 'train/grad_norm_pre_clip': 18.255905151367188, 'timing/data': 0.0005822440143674612, 'timing/model': 1.927458474994637, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 1.66}) +[2026-06-23 18:30:16,442][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.028312647715210915, 'train/grad_norm_pre_clip': 23.26121711730957, 'timing/data': 0.000748175079934299, 'timing/model': 1.9188705059932545, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 1.66}) +[2026-06-23 18:30:20,420][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.02185661345720291, 'train/grad_norm_pre_clip': 22.63324546813965, 'timing/data': 0.0006718359654769301, 'timing/model': 2.1130893190857023, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 1.66}) +[2026-06-23 18:30:24,238][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.023313019424676895, 'train/grad_norm_pre_clip': 20.39200782775879, 'timing/data': 0.000643312931060791, 'timing/model': 1.9184627429349348, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 1.66}) +[2026-06-23 18:30:28,040][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.03804086148738861, 'train/grad_norm_pre_clip': 37.1104621887207, 'timing/data': 0.0007873399881646037, 'timing/model': 1.9146148749860004, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 1.66}) +[2026-06-23 18:30:31,841][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.052850257605314255, 'train/grad_norm_pre_clip': 19.873802185058594, 'timing/data': 0.0006818940164521337, 'timing/model': 1.931789010995999, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 1.66}) +[2026-06-23 18:30:35,640][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.028403816744685173, 'train/grad_norm_pre_clip': 13.476227760314941, 'timing/data': 0.0005978220142424107, 'timing/model': 1.9274859729921445, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 1.66}) +[2026-06-23 18:30:39,444][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.09130071848630905, 'train/grad_norm_pre_clip': 31.364763259887695, 'timing/data': 0.0006401099963113666, 'timing/model': 1.9271741850534454, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 1.66}) +[2026-06-23 18:30:43,246][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.05779294669628143, 'train/grad_norm_pre_clip': 27.201515197753906, 'timing/data': 0.0007195150246843696, 'timing/model': 1.9230941479327157, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 1.66}) +[2026-06-23 18:30:47,043][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.03624437749385834, 'train/grad_norm_pre_clip': 22.05169677734375, 'timing/data': 0.0007491130381822586, 'timing/model': 1.929105072049424, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 1.66}) +[2026-06-23 18:30:50,839][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.08759929984807968, 'train/grad_norm_pre_clip': 53.231468200683594, 'timing/data': 0.0006263349205255508, 'timing/model': 1.9186124969273806, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 1.66}) +[2026-06-23 18:30:54,794][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.08580997586250305, 'train/grad_norm_pre_clip': 15.935833930969238, 'timing/data': 0.0008004480041563511, 'timing/model': 1.928041047998704, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 1.66}) +[2026-06-23 18:30:58,601][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.0277656652033329, 'train/grad_norm_pre_clip': 45.08312225341797, 'timing/data': 0.000587912043556571, 'timing/model': 1.9291078250389546, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 1.66}) +[2026-06-23 18:31:02,386][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.04795488715171814, 'train/grad_norm_pre_clip': 29.679975509643555, 'timing/data': 0.0005924950819462538, 'timing/model': 1.9225510549731553, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 1.66}) +[2026-06-23 18:31:06,191][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.10902898013591766, 'train/grad_norm_pre_clip': 28.782062530517578, 'timing/data': 0.0006236509652808309, 'timing/model': 1.928243848029524, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 1.66}) +[2026-06-23 18:31:09,994][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.022928036749362946, 'train/grad_norm_pre_clip': 13.547440528869629, 'timing/data': 0.0006352060008794069, 'timing/model': 1.9235599709209055, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 1.66}) +[2026-06-23 18:31:13,789][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.0016368841752409935, 'train/grad_norm_pre_clip': 3.856659173965454, 'timing/data': 0.0006788630271330476, 'timing/model': 1.9252328879665583, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 1.66}) +[2026-06-23 18:31:17,594][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.056264862418174744, 'train/grad_norm_pre_clip': 24.357738494873047, 'timing/data': 0.0005955910310149193, 'timing/model': 1.9366917489096522, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 1.66}) +[2026-06-23 18:31:21,395][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.04239905625581741, 'train/grad_norm_pre_clip': 21.01181411743164, 'timing/data': 0.0006609989795833826, 'timing/model': 1.9236460650572553, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 1.66}) +[2026-06-23 18:31:25,313][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.04547857120633125, 'train/grad_norm_pre_clip': 21.445924758911133, 'timing/data': 0.0005906710866838694, 'timing/model': 2.046549223945476, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 1.66}) +[2026-06-23 18:31:29,102][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.06106025353074074, 'train/grad_norm_pre_clip': 19.418216705322266, 'timing/data': 0.0005950459744781256, 'timing/model': 1.9252668441040441, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 1.66}) +[2026-06-23 18:31:32,899][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4300 +[2026-06-23 18:33:40,771][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4300 +[2026-06-23 18:33:40,772][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4300 +[2026-06-23 18:35:03,157][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4300 +[2026-06-23 18:35:03,158][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.08981084078550339, 'train/grad_norm_pre_clip': 61.30784606933594, 'eval/action_loss/seconds': 127.86859747802373, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.031873032450675964, 'eval/latency_2/loss': 0.031873032450675964, 'eval/flappy/loss': 0.031873032450675964, 'eval/flappy/latency_2/loss': 0.031873032450675964, 'eval/flap_precision': 0.9728445349626612, 'eval/flap_recall': 0.9669365721997301, 'eval/flap_f1': 0.969881556683587, 'eval/flap_rate_model': 0.20458333333333334, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9698815566835871, 'eval/latency_2/flap_precision': 0.9728445349626612, 'eval/latency_2/flap_recall': 0.9669365721997301, 'eval/latency_2/flap_f1': 0.969881556683587, 'eval/latency_2/flap_rate_model': 0.20458333333333334, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9698815566835871, 'eval/action_classification/data_wait_seconds': 1.2700490951538086, 'eval/action_classification/predict_seconds': 80.9358901977539, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.000682962010614574, 'timing/model': 1.9299750640057027, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 1.66}) +[2026-06-23 18:35:06,943][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.03492258861660957, 'train/grad_norm_pre_clip': 31.411638259887695, 'timing/data': 0.0007929089479148388, 'timing/model': 1.914120897068642, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 1.66}) +[2026-06-23 18:35:10,737][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.028922894969582558, 'train/grad_norm_pre_clip': 23.86461639404297, 'timing/data': 0.0006019009742885828, 'timing/model': 1.9228854310931638, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 1.66}) +[2026-06-23 18:35:14,540][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.011738363653421402, 'train/grad_norm_pre_clip': 31.38715171813965, 'timing/data': 0.0008043210254982114, 'timing/model': 1.910406123031862, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 1.67}) +[2026-06-23 18:35:18,315][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.054691530764102936, 'train/grad_norm_pre_clip': 34.33758544921875, 'timing/data': 0.0007981429807841778, 'timing/model': 1.9112193550681695, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 1.67}) +[2026-06-23 18:35:22,095][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.007492939941585064, 'train/grad_norm_pre_clip': 9.479172706604004, 'timing/data': 0.0006530169630423188, 'timing/model': 1.9236559240380302, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 1.67}) +[2026-06-23 18:35:25,873][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.030526936054229736, 'train/grad_norm_pre_clip': 22.776535034179688, 'timing/data': 0.0006151599809527397, 'timing/model': 1.9202932639745995, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 1.67}) +[2026-06-23 18:35:29,658][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.10486862063407898, 'train/grad_norm_pre_clip': 35.66606521606445, 'timing/data': 0.0006748620653524995, 'timing/model': 1.9208822409855202, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 1.67}) +[2026-06-23 18:35:33,449][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.024969004094600677, 'train/grad_norm_pre_clip': 25.53011131286621, 'timing/data': 0.0006198899354785681, 'timing/model': 1.9190604540053755, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 1.67}) +[2026-06-23 18:35:37,228][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.07162953168153763, 'train/grad_norm_pre_clip': 30.138460159301758, 'timing/data': 0.000810838071629405, 'timing/model': 1.9202062310650945, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 1.67}) +[2026-06-23 18:35:41,021][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.08707356452941895, 'train/grad_norm_pre_clip': 53.741783142089844, 'timing/data': 0.0006042749155312777, 'timing/model': 1.9296315349638462, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 1.67}) +[2026-06-23 18:35:44,827][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.052860964089632034, 'train/grad_norm_pre_clip': 25.904226303100586, 'timing/data': 0.0006153970025479794, 'timing/model': 1.9478404959663749, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 1.67}) +[2026-06-23 18:35:48,618][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.039400890469551086, 'train/grad_norm_pre_clip': 23.095722198486328, 'timing/data': 0.0008266289951279759, 'timing/model': 1.9233267910312861, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 1.67}) +[2026-06-23 18:35:52,429][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.12720143795013428, 'train/grad_norm_pre_clip': 34.06711959838867, 'timing/data': 0.0006957200821489096, 'timing/model': 1.9247122219530866, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 1.67}) +[2026-06-23 18:35:56,227][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.03809216618537903, 'train/grad_norm_pre_clip': 17.016952514648438, 'timing/data': 0.0006016050465404987, 'timing/model': 1.9292618359904736, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 1.67}) +[2026-06-23 18:36:00,020][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.013990605250000954, 'train/grad_norm_pre_clip': 7.934648513793945, 'timing/data': 0.0006034329999238253, 'timing/model': 1.9177355089923367, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 1.67}) +[2026-06-23 18:36:03,827][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.08232081681489944, 'train/grad_norm_pre_clip': 45.114139556884766, 'timing/data': 0.0007790760137140751, 'timing/model': 1.9332015480613336, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 1.67}) +[2026-06-23 18:36:07,637][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.0690799430012703, 'train/grad_norm_pre_clip': 59.70249557495117, 'timing/data': 0.0005842969985678792, 'timing/model': 1.9470521409530193, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 1.67}) +[2026-06-23 18:36:11,436][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.028898579999804497, 'train/grad_norm_pre_clip': 12.946112632751465, 'timing/data': 0.0005870439345017076, 'timing/model': 1.925469227018766, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 1.67}) +[2026-06-23 18:36:15,237][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.030321406200528145, 'train/grad_norm_pre_clip': 33.15336227416992, 'timing/data': 0.0007795250276103616, 'timing/model': 1.9202715420396999, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 1.67}) +[2026-06-23 18:36:19,046][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.03653480485081673, 'train/grad_norm_pre_clip': 36.7257194519043, 'timing/data': 0.000596456928178668, 'timing/model': 1.9335677290800959, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 1.67}) +[2026-06-23 18:36:22,837][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.0071422564797103405, 'train/grad_norm_pre_clip': 12.600420951843262, 'timing/data': 0.0005942840361967683, 'timing/model': 1.930907438043505, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 1.67}) +[2026-06-23 18:36:26,610][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.02243123948574066, 'train/grad_norm_pre_clip': 26.318126678466797, 'timing/data': 0.000599120045080781, 'timing/model': 1.9146172530017793, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 1.67}) +[2026-06-23 18:36:30,408][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.06322486698627472, 'train/grad_norm_pre_clip': 36.774696350097656, 'timing/data': 0.0006081720348447561, 'timing/model': 1.9257529890164733, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 1.67}) +[2026-06-23 18:36:34,215][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.10815632343292236, 'train/grad_norm_pre_clip': 38.68736267089844, 'timing/data': 0.0005949039477854967, 'timing/model': 1.9341267329873517, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 1.67}) +[2026-06-23 18:36:38,017][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.08074213564395905, 'train/grad_norm_pre_clip': 51.87458038330078, 'timing/data': 0.0006335079669952393, 'timing/model': 1.9359393230406567, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 1.67}) +[2026-06-23 18:36:41,831][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.008000376634299755, 'train/grad_norm_pre_clip': 16.76332664489746, 'timing/data': 0.0005886689759790897, 'timing/model': 1.9364120309473947, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 1.67}) +[2026-06-23 18:36:45,619][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.027350271120667458, 'train/grad_norm_pre_clip': 29.303979873657227, 'timing/data': 0.0006108309607952833, 'timing/model': 1.9186164850834757, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 1.67}) +[2026-06-23 18:36:49,421][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.08937224000692368, 'train/grad_norm_pre_clip': 35.85117721557617, 'timing/data': 0.0005961570423096418, 'timing/model': 1.9351329139899462, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 1.68}) +[2026-06-23 18:36:53,226][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.18415264785289764, 'train/grad_norm_pre_clip': 67.2494888305664, 'timing/data': 0.0010271610226482153, 'timing/model': 1.9332130859838799, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 1.68}) +[2026-06-23 18:36:57,038][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.02147325500845909, 'train/grad_norm_pre_clip': 19.305137634277344, 'timing/data': 0.000585838919505477, 'timing/model': 1.9297179409768432, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 1.68}) +[2026-06-23 18:37:00,852][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.1399734765291214, 'train/grad_norm_pre_clip': 58.77322769165039, 'timing/data': 0.0006121760234236717, 'timing/model': 1.9332636729814112, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 1.68}) +[2026-06-23 18:37:04,652][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.019995342940092087, 'train/grad_norm_pre_clip': 31.109418869018555, 'timing/data': 0.0007941350340843201, 'timing/model': 1.9273052649805322, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 1.68}) +[2026-06-23 18:37:08,448][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.06336566060781479, 'train/grad_norm_pre_clip': 35.1148567199707, 'timing/data': 0.0005926730809733272, 'timing/model': 1.9252830980112776, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 1.68}) +[2026-06-23 18:37:12,232][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.06486430764198303, 'train/grad_norm_pre_clip': 67.13310241699219, 'timing/data': 0.0006561520276591182, 'timing/model': 1.9201745269820094, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 1.68}) +[2026-06-23 18:37:16,038][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.028422940522432327, 'train/grad_norm_pre_clip': 19.296113967895508, 'timing/data': 0.0008094960357993841, 'timing/model': 1.9374271880369633, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 1.68}) +[2026-06-23 18:37:19,850][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.007365759927779436, 'train/grad_norm_pre_clip': 43.32572555541992, 'timing/data': 0.0007907010149210691, 'timing/model': 1.9397529050474986, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 1.68}) +[2026-06-23 18:37:23,839][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.02886352688074112, 'train/grad_norm_pre_clip': 30.530269622802734, 'timing/data': 0.0005940030096098781, 'timing/model': 2.110366013017483, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 1.68}) +[2026-06-23 18:37:27,646][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.10378263890743256, 'train/grad_norm_pre_clip': 53.140132904052734, 'timing/data': 0.0007750979857519269, 'timing/model': 1.9298975969431922, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 1.68}) +[2026-06-23 18:37:31,421][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.06034660339355469, 'train/grad_norm_pre_clip': 19.8508358001709, 'timing/data': 0.0006055820267647505, 'timing/model': 1.9164595670299605, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 1.68}) +[2026-06-23 18:37:35,227][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.03769180551171303, 'train/grad_norm_pre_clip': 24.801055908203125, 'timing/data': 0.0006265589036047459, 'timing/model': 1.9282237099250779, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 1.68}) +[2026-06-23 18:37:39,030][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.05353681370615959, 'train/grad_norm_pre_clip': 26.569150924682617, 'timing/data': 0.0006686320994049311, 'timing/model': 1.9287598199443892, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 1.68}) +[2026-06-23 18:37:42,829][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.015286359004676342, 'train/grad_norm_pre_clip': 16.92543601989746, 'timing/data': 0.0007790169911459088, 'timing/model': 1.9337166280020028, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 1.68}) +[2026-06-23 18:37:46,640][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.049215078353881836, 'train/grad_norm_pre_clip': 29.88567352294922, 'timing/data': 0.0005990930367261171, 'timing/model': 1.9391025319928303, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 1.68}) +[2026-06-23 18:37:50,453][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.049331292510032654, 'train/grad_norm_pre_clip': 30.338483810424805, 'timing/data': 0.0006031799130141735, 'timing/model': 1.9311259390087798, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 1.68}) +[2026-06-23 18:37:54,246][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.06585410982370377, 'train/grad_norm_pre_clip': 27.793859481811523, 'timing/data': 0.0006042649038136005, 'timing/model': 1.9191699330694973, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 1.68}) +[2026-06-23 18:37:58,233][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.20088905096054077, 'train/grad_norm_pre_clip': 43.35921859741211, 'timing/data': 0.0005906239384785295, 'timing/model': 1.9304622150957584, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 1.68}) +[2026-06-23 18:38:02,043][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.09504861384630203, 'train/grad_norm_pre_clip': 35.84242630004883, 'timing/data': 0.000600127037614584, 'timing/model': 1.9401828540721908, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 1.68}) +[2026-06-23 18:38:05,838][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.12263819575309753, 'train/grad_norm_pre_clip': 68.52750396728516, 'timing/data': 0.0006069929804652929, 'timing/model': 1.929651363985613, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 1.68}) +[2026-06-23 18:38:09,624][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.1872168481349945, 'train/grad_norm_pre_clip': 41.92127227783203, 'timing/data': 0.0006131730042397976, 'timing/model': 1.9178162150783464, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 1.68}) +[2026-06-23 18:38:13,437][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.006173210684210062, 'train/grad_norm_pre_clip': 26.610137939453125, 'timing/data': 0.000639689969830215, 'timing/model': 1.937439188011922, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 1.68}) +[2026-06-23 18:38:17,247][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.03536802530288696, 'train/grad_norm_pre_clip': 22.36884117126465, 'timing/data': 0.000655011972412467, 'timing/model': 1.93397027393803, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 1.68}) +[2026-06-23 18:38:21,037][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.04344690591096878, 'train/grad_norm_pre_clip': 19.253673553466797, 'timing/data': 0.0006312300683930516, 'timing/model': 1.9258014389779419, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 1.68}) +[2026-06-23 18:38:24,839][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.03407549113035202, 'train/grad_norm_pre_clip': 51.378543853759766, 'timing/data': 0.0008184739854186773, 'timing/model': 1.9240574990399182, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 1.68}) +[2026-06-23 18:38:28,837][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.015384066849946976, 'train/grad_norm_pre_clip': 38.01046371459961, 'timing/data': 0.0006034260150045156, 'timing/model': 2.120822925004177, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 1.69}) +[2026-06-23 18:38:32,650][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.04462480917572975, 'train/grad_norm_pre_clip': 20.661596298217773, 'timing/data': 0.0006067630602046847, 'timing/model': 1.9374898090027273, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 1.69}) +[2026-06-23 18:38:36,446][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.04591704159975052, 'train/grad_norm_pre_clip': 23.60396957397461, 'timing/data': 0.0008051670156419277, 'timing/model': 1.9212113780668005, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 1.69}) +[2026-06-23 18:38:40,246][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.09021259099245071, 'train/grad_norm_pre_clip': 43.35693359375, 'timing/data': 0.0006000340217724442, 'timing/model': 1.9291832580929622, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 1.69}) +[2026-06-23 18:38:44,050][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.013747514225542545, 'train/grad_norm_pre_clip': 14.852893829345703, 'timing/data': 0.0008319549961015582, 'timing/model': 1.925735897035338, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 1.69}) +[2026-06-23 18:38:47,862][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.04127081111073494, 'train/grad_norm_pre_clip': 27.32992935180664, 'timing/data': 0.0006115350406616926, 'timing/model': 1.9259631860768422, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 1.69}) +[2026-06-23 18:38:51,702][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.0419299490749836, 'train/grad_norm_pre_clip': 34.05088424682617, 'timing/data': 0.0006417250260710716, 'timing/model': 1.9602232880424708, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 1.69}) +[2026-06-23 18:38:55,514][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.04963304474949837, 'train/grad_norm_pre_clip': 42.55471420288086, 'timing/data': 0.0008399880025535822, 'timing/model': 1.9352028510766104, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 1.69}) +[2026-06-23 18:38:59,313][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.03430529683828354, 'train/grad_norm_pre_clip': 29.15134620666504, 'timing/data': 0.0007784590125083923, 'timing/model': 1.9251892720349133, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 1.69}) +[2026-06-23 18:39:03,302][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.06276693940162659, 'train/grad_norm_pre_clip': 21.7609920501709, 'timing/data': 0.0006122039631009102, 'timing/model': 1.936369423987344, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 1.69}) +[2026-06-23 18:39:07,094][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.05980836972594261, 'train/grad_norm_pre_clip': 20.898405075073242, 'timing/data': 0.0008324349764734507, 'timing/model': 1.9247270099585876, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 1.69}) +[2026-06-23 18:39:10,915][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.012535308487713337, 'train/grad_norm_pre_clip': 20.052783966064453, 'timing/data': 0.0006154719740152359, 'timing/model': 1.9455880189780146, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 1.69}) +[2026-06-23 18:39:14,718][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.022186025977134705, 'train/grad_norm_pre_clip': 23.446596145629883, 'timing/data': 0.0005958490073680878, 'timing/model': 1.9363680289825425, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 1.69}) +[2026-06-23 18:39:18,526][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.10860704630613327, 'train/grad_norm_pre_clip': 37.83043670654297, 'timing/data': 0.0006149460095912218, 'timing/model': 1.9315222919685766, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 1.69}) +[2026-06-23 18:39:22,331][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.017811447381973267, 'train/grad_norm_pre_clip': 19.962326049804688, 'timing/data': 0.000943336053751409, 'timing/model': 1.9304315529298037, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 1.69}) +[2026-06-23 18:39:26,141][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.08488396555185318, 'train/grad_norm_pre_clip': 32.05210876464844, 'timing/data': 0.000612196046859026, 'timing/model': 1.9340094599174336, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 1.69}) +[2026-06-23 18:39:29,979][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.028200805187225342, 'train/grad_norm_pre_clip': 20.864362716674805, 'timing/data': 0.0005974969826638699, 'timing/model': 1.9342512330040336, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 1.69}) +[2026-06-23 18:39:33,775][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.02366793341934681, 'train/grad_norm_pre_clip': 26.734590530395508, 'timing/data': 0.0006090679671615362, 'timing/model': 1.923800908960402, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 1.69}) +[2026-06-23 18:39:37,586][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.05373268947005272, 'train/grad_norm_pre_clip': 26.1699275970459, 'timing/data': 0.0006529800593852997, 'timing/model': 1.9341903839958832, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 1.69}) +[2026-06-23 18:39:41,446][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.015368578024208546, 'train/grad_norm_pre_clip': 24.49932289123535, 'timing/data': 0.0005978770786896348, 'timing/model': 1.9951367779867724, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 1.69}) +[2026-06-23 18:39:45,261][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.005471598356962204, 'train/grad_norm_pre_clip': 14.73687744140625, 'timing/data': 0.0005990999052301049, 'timing/model': 1.9330995560158044, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 1.69}) +[2026-06-23 18:39:49,070][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.05365331470966339, 'train/grad_norm_pre_clip': 32.66529846191406, 'timing/data': 0.0007073729066178203, 'timing/model': 1.928478137939237, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 1.69}) +[2026-06-23 18:39:52,875][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.06595204770565033, 'train/grad_norm_pre_clip': 17.974414825439453, 'timing/data': 0.0006837769178673625, 'timing/model': 1.9358289140509441, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 1.69}) +[2026-06-23 18:39:56,678][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.049326907843351364, 'train/grad_norm_pre_clip': 58.2352294921875, 'timing/data': 0.0007217649836093187, 'timing/model': 1.922532666940242, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 1.69}) +[2026-06-23 18:40:00,478][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.04118737578392029, 'train/grad_norm_pre_clip': 35.94681930541992, 'timing/data': 0.0009341239929199219, 'timing/model': 1.924587415996939, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 1.69}) +[2026-06-23 18:40:04,289][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.05332176014780998, 'train/grad_norm_pre_clip': 27.15951919555664, 'timing/data': 0.0009068490471690893, 'timing/model': 1.9360205649863929, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 1.69}) +[2026-06-23 18:40:08,106][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.08207971602678299, 'train/grad_norm_pre_clip': 88.86202239990234, 'timing/data': 0.0006182519719004631, 'timing/model': 1.9307598660234362, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 1.7}) +[2026-06-23 18:40:11,904][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.08294592052698135, 'train/grad_norm_pre_clip': 49.03627395629883, 'timing/data': 0.0006641240324825048, 'timing/model': 1.9234535299474373, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 1.7}) +[2026-06-23 18:40:15,702][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.074654720723629, 'train/grad_norm_pre_clip': 25.280757904052734, 'timing/data': 0.0005997230764478445, 'timing/model': 1.9262747429311275, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 1.7}) +[2026-06-23 18:40:19,504][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.08846868574619293, 'train/grad_norm_pre_clip': 27.391155242919922, 'timing/data': 0.0005847580032423139, 'timing/model': 1.9318950839806348, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 1.7}) +[2026-06-23 18:40:23,307][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.005030246917158365, 'train/grad_norm_pre_clip': 23.984939575195312, 'timing/data': 0.0006138840690255165, 'timing/model': 1.932129348977469, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 1.7}) +[2026-06-23 18:40:27,111][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.06359843909740448, 'train/grad_norm_pre_clip': 29.529813766479492, 'timing/data': 0.000659469049423933, 'timing/model': 1.9334756240714341, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 1.7}) +[2026-06-23 18:40:30,920][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.06616611033678055, 'train/grad_norm_pre_clip': 26.536645889282227, 'timing/data': 0.0006074750563129783, 'timing/model': 1.9255221330095083, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 1.7}) +[2026-06-23 18:40:34,732][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.08310338854789734, 'train/grad_norm_pre_clip': 24.468585968017578, 'timing/data': 0.000613568932749331, 'timing/model': 1.9370522730750963, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 1.7}) +[2026-06-23 18:40:38,532][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.05712762475013733, 'train/grad_norm_pre_clip': 22.90013885498047, 'timing/data': 0.0006010920042172074, 'timing/model': 1.9276488550240174, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 1.7}) +[2026-06-23 18:40:42,335][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.16850116848945618, 'train/grad_norm_pre_clip': 62.03251266479492, 'timing/data': 0.000760348979383707, 'timing/model': 1.9303030179580674, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 1.7}) +[2026-06-23 18:40:46,138][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.0683998391032219, 'train/grad_norm_pre_clip': 34.32252502441406, 'timing/data': 0.0006050799274817109, 'timing/model': 1.92726521496661, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 1.7}) +[2026-06-23 18:40:49,965][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.09037329256534576, 'train/grad_norm_pre_clip': 33.29487991333008, 'timing/data': 0.0011085569858551025, 'timing/model': 1.9552964320173487, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 1.7}) +[2026-06-23 18:40:53,777][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.00945034995675087, 'train/grad_norm_pre_clip': 16.6180477142334, 'timing/data': 0.0007809379603713751, 'timing/model': 1.9313572460087016, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 1.7}) +[2026-06-23 18:40:57,586][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.0396382212638855, 'train/grad_norm_pre_clip': 41.22177505493164, 'timing/data': 0.0006088670343160629, 'timing/model': 1.9252227190881968, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 1.7}) +[2026-06-23 18:41:01,386][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.11734791100025177, 'train/grad_norm_pre_clip': 29.274442672729492, 'timing/data': 0.0005952960345894098, 'timing/model': 1.9300666019553319, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 1.7}) +[2026-06-23 18:41:05,190][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.04346888139843941, 'train/grad_norm_pre_clip': 37.00620651245117, 'timing/data': 0.0006464309990406036, 'timing/model': 1.931851033004932, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 1.7}) +[2026-06-23 18:41:09,010][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.021154196932911873, 'train/grad_norm_pre_clip': 12.180853843688965, 'timing/data': 0.0006559590110555291, 'timing/model': 1.9419585820287466, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 1.7}) +[2026-06-23 18:41:12,813][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.12927751243114471, 'train/grad_norm_pre_clip': 53.59269332885742, 'timing/data': 0.0006488569779321551, 'timing/model': 1.9327334240078926, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 1.7}) +[2026-06-23 18:41:16,620][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.025328503921628, 'train/grad_norm_pre_clip': 20.772905349731445, 'timing/data': 0.0006039480213075876, 'timing/model': 1.9305700940312818, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 1.7}) +[2026-06-23 18:41:20,433][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.03404368460178375, 'train/grad_norm_pre_clip': 29.920852661132812, 'timing/data': 0.0006030240328982472, 'timing/model': 1.940060861990787, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 1.7}) +[2026-06-23 18:41:24,228][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4400 +[2026-06-23 18:43:29,960][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4400 +[2026-06-23 18:43:29,962][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4400 +[2026-06-23 18:44:54,074][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4400 +[2026-06-23 18:44:54,074][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.09898050129413605, 'train/grad_norm_pre_clip': 18.75771141052246, 'eval/action_loss/seconds': 125.72947845805902, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03176560997962952, 'eval/latency_2/loss': 0.03176560997962952, 'eval/flappy/loss': 0.03176560997962952, 'eval/flappy/latency_2/loss': 0.03176560997962952, 'eval/flap_precision': 0.9766001376462491, 'eval/flap_recall': 0.9574898785425101, 'eval/flap_f1': 0.9669505962521295, 'eval/flap_rate_model': 0.20180555555555554, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9669505962521295, 'eval/latency_2/flap_precision': 0.9766001376462491, 'eval/latency_2/flap_recall': 0.9574898785425101, 'eval/latency_2/flap_f1': 0.9669505962521295, 'eval/latency_2/flap_rate_model': 0.20180555555555554, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9669505962521295, 'eval/action_classification/data_wait_seconds': 1.6524560451507568, 'eval/action_classification/predict_seconds': 82.26738739013672, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006227780831977725, 'timing/model': 1.9189572129398584, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 1.7}) +[2026-06-23 18:44:57,843][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.0735013484954834, 'train/grad_norm_pre_clip': 32.80656051635742, 'timing/data': 0.0007968450663611293, 'timing/model': 1.9079578360542655, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 1.7}) +[2026-06-23 18:45:01,743][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.13840575516223907, 'train/grad_norm_pre_clip': 63.03230667114258, 'timing/data': 0.0006103739142417908, 'timing/model': 1.915171073982492, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 1.7}) +[2026-06-23 18:45:05,539][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.04000892862677574, 'train/grad_norm_pre_clip': 63.96052932739258, 'timing/data': 0.0008007739670574665, 'timing/model': 1.923431092989631, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 1.7}) +[2026-06-23 18:45:09,303][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.14650429785251617, 'train/grad_norm_pre_clip': 33.91187286376953, 'timing/data': 0.0006066289497539401, 'timing/model': 1.9134431179845706, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 1.7}) +[2026-06-23 18:45:13,082][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.018742017447948456, 'train/grad_norm_pre_clip': 11.573822975158691, 'timing/data': 0.0005934690125286579, 'timing/model': 1.9210818010615185, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 1.7}) +[2026-06-23 18:45:16,878][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.015243577770888805, 'train/grad_norm_pre_clip': 24.896764755249023, 'timing/data': 0.0009965330827981234, 'timing/model': 1.9300837500486523, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 1.71}) +[2026-06-23 18:45:20,657][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.027698880061507225, 'train/grad_norm_pre_clip': 36.26612854003906, 'timing/data': 0.000880512990988791, 'timing/model': 1.9188524120254442, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 1.71}) +[2026-06-23 18:45:24,444][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.013384604826569557, 'train/grad_norm_pre_clip': 20.87961769104004, 'timing/data': 0.0005960180424153805, 'timing/model': 1.923064881004393, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 1.71}) +[2026-06-23 18:45:28,228][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.05544900521636009, 'train/grad_norm_pre_clip': 49.74982452392578, 'timing/data': 0.0005843080580234528, 'timing/model': 1.9182630979921669, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 1.71}) +[2026-06-23 18:45:32,097][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.00218799221329391, 'train/grad_norm_pre_clip': 15.118915557861328, 'timing/data': 0.000807018019258976, 'timing/model': 2.007901046075858, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 1.71}) +[2026-06-23 18:45:35,879][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.04371549189090729, 'train/grad_norm_pre_clip': 40.26362228393555, 'timing/data': 0.0006101559847593307, 'timing/model': 1.9208155908854678, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 1.71}) +[2026-06-23 18:45:39,671][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.028815707191824913, 'train/grad_norm_pre_clip': 17.46055030822754, 'timing/data': 0.000989805907011032, 'timing/model': 1.9267897449899465, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 1.71}) +[2026-06-23 18:45:43,471][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.12527793645858765, 'train/grad_norm_pre_clip': 35.154762268066406, 'timing/data': 0.0006595690501853824, 'timing/model': 1.9298086959170178, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 1.71}) +[2026-06-23 18:45:47,274][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.036185123026371, 'train/grad_norm_pre_clip': 36.06965637207031, 'timing/data': 0.0007811069954186678, 'timing/model': 1.9226598109817132, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 1.71}) +[2026-06-23 18:45:51,064][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.04661296680569649, 'train/grad_norm_pre_clip': 25.35982894897461, 'timing/data': 0.0009312439942732453, 'timing/model': 1.929955048020929, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 1.71}) +[2026-06-23 18:45:54,861][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.01573454774916172, 'train/grad_norm_pre_clip': 13.512300491333008, 'timing/data': 0.000591604970395565, 'timing/model': 1.9244669080944732, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 1.71}) +[2026-06-23 18:45:58,644][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.04714685678482056, 'train/grad_norm_pre_clip': 29.873199462890625, 'timing/data': 0.0006137019954621792, 'timing/model': 1.9229031159775332, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 1.71}) +[2026-06-23 18:46:02,439][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.018357601016759872, 'train/grad_norm_pre_clip': 11.60512638092041, 'timing/data': 0.0006209929706528783, 'timing/model': 1.9244147640420124, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 1.71}) +[2026-06-23 18:46:06,296][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.08142312616109848, 'train/grad_norm_pre_clip': 30.492050170898438, 'timing/data': 0.0007282620063051581, 'timing/model': 1.9248138710390776, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 1.71}) +[2026-06-23 18:46:10,094][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.053511928766965866, 'train/grad_norm_pre_clip': 23.412233352661133, 'timing/data': 0.0008994509698823094, 'timing/model': 1.9281780769815668, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 1.71}) +[2026-06-23 18:46:13,876][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.02877384051680565, 'train/grad_norm_pre_clip': 28.92087745666504, 'timing/data': 0.0006253599422052503, 'timing/model': 1.9172717140754685, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 1.71}) +[2026-06-23 18:46:17,663][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.029406776651740074, 'train/grad_norm_pre_clip': 64.68872833251953, 'timing/data': 0.0006061519961804152, 'timing/model': 1.9155849640956149, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 1.71}) +[2026-06-23 18:46:21,470][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.1338537335395813, 'train/grad_norm_pre_clip': 31.86019515991211, 'timing/data': 0.0006433000089600682, 'timing/model': 1.9353238369803876, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 1.71}) +[2026-06-23 18:46:25,258][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.02233291044831276, 'train/grad_norm_pre_clip': 15.244940757751465, 'timing/data': 0.000616347067989409, 'timing/model': 1.9160603759810328, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 1.71}) +[2026-06-23 18:46:29,058][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.0519469678401947, 'train/grad_norm_pre_clip': 46.174686431884766, 'timing/data': 0.0007877199677750468, 'timing/model': 1.917330382973887, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 1.71}) +[2026-06-23 18:46:32,869][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.07389657199382782, 'train/grad_norm_pre_clip': 24.111995697021484, 'timing/data': 0.0005889820167794824, 'timing/model': 1.9400488429237157, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 1.71}) +[2026-06-23 18:46:36,693][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.046483300626277924, 'train/grad_norm_pre_clip': 39.32561111450195, 'timing/data': 0.0006111569236963987, 'timing/model': 1.961391295073554, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 1.71}) +[2026-06-23 18:46:40,487][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.09666237235069275, 'train/grad_norm_pre_clip': 38.10745620727539, 'timing/data': 0.0006148279644548893, 'timing/model': 1.9155394909903407, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 1.71}) +[2026-06-23 18:46:44,295][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.03264271467924118, 'train/grad_norm_pre_clip': 39.56300735473633, 'timing/data': 0.0007837309967726469, 'timing/model': 1.941229396034032, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 1.71}) +[2026-06-23 18:46:48,103][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.06356070190668106, 'train/grad_norm_pre_clip': 56.60226821899414, 'timing/data': 0.0007730539655312896, 'timing/model': 1.925799384014681, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 1.71}) +[2026-06-23 18:46:51,898][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.01585194654762745, 'train/grad_norm_pre_clip': 13.227344512939453, 'timing/data': 0.0007207000162452459, 'timing/model': 1.9249917670385912, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 1.71}) +[2026-06-23 18:46:55,696][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.0633871778845787, 'train/grad_norm_pre_clip': 51.114013671875, 'timing/data': 0.0006094770506024361, 'timing/model': 1.9264807470608503, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 1.72}) +[2026-06-23 18:46:59,504][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.03343087434768677, 'train/grad_norm_pre_clip': 12.671689987182617, 'timing/data': 0.000593848992139101, 'timing/model': 1.9345449939137325, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 1.72}) +[2026-06-23 18:47:03,302][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.008792647160589695, 'train/grad_norm_pre_clip': 11.048452377319336, 'timing/data': 0.0005952549399808049, 'timing/model': 1.9330063189845532, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 1.72}) +[2026-06-23 18:47:07,095][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.09281948953866959, 'train/grad_norm_pre_clip': 84.4105453491211, 'timing/data': 0.0006128100212663412, 'timing/model': 1.9245526869781315, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 1.72}) +[2026-06-23 18:47:10,917][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.011246309615671635, 'train/grad_norm_pre_clip': 16.51897430419922, 'timing/data': 0.000623369007371366, 'timing/model': 1.9375403589801863, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 1.72}) +[2026-06-23 18:47:14,725][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.02294163405895233, 'train/grad_norm_pre_clip': 42.193992614746094, 'timing/data': 0.0006044571055099368, 'timing/model': 1.9383434759220108, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 1.72}) +[2026-06-23 18:47:18,513][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.04574422538280487, 'train/grad_norm_pre_clip': 51.74598693847656, 'timing/data': 0.000605724984779954, 'timing/model': 1.9217684790492058, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 1.72}) +[2026-06-23 18:47:22,305][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.20049040019512177, 'train/grad_norm_pre_clip': 22.571308135986328, 'timing/data': 0.000790275982581079, 'timing/model': 1.9256235889624804, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 1.72}) +[2026-06-23 18:47:26,117][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.03074335679411888, 'train/grad_norm_pre_clip': 27.507204055786133, 'timing/data': 0.0006098300218582153, 'timing/model': 1.9297817900078371, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 1.72}) +[2026-06-23 18:47:29,922][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.11967851221561432, 'train/grad_norm_pre_clip': 36.192203521728516, 'timing/data': 0.001106353010982275, 'timing/model': 1.9295250909635797, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 1.72}) +[2026-06-23 18:47:33,720][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.03427466005086899, 'train/grad_norm_pre_clip': 51.45344543457031, 'timing/data': 0.0006940599996596575, 'timing/model': 1.927840293967165, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 1.72}) +[2026-06-23 18:47:37,521][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.029642242938280106, 'train/grad_norm_pre_clip': 15.196636199951172, 'timing/data': 0.000692047062329948, 'timing/model': 1.9355639539426193, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 1.72}) +[2026-06-23 18:47:41,322][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.028372609987854958, 'train/grad_norm_pre_clip': 27.490333557128906, 'timing/data': 0.0006239929934963584, 'timing/model': 1.9314324839506298, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 1.72}) +[2026-06-23 18:47:45,307][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.05854722484946251, 'train/grad_norm_pre_clip': 61.05263900756836, 'timing/data': 0.0007226259913295507, 'timing/model': 1.9327668140176684, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 1.72}) +[2026-06-23 18:47:49,119][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.06733671575784683, 'train/grad_norm_pre_clip': 33.44330978393555, 'timing/data': 0.0006011569639667869, 'timing/model': 1.927031734958291, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 1.72}) +[2026-06-23 18:47:52,910][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.04672892764210701, 'train/grad_norm_pre_clip': 60.22933578491211, 'timing/data': 0.0006111359689384699, 'timing/model': 1.9250944259110838, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 1.72}) +[2026-06-23 18:47:56,726][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.05685291439294815, 'train/grad_norm_pre_clip': 33.07371139526367, 'timing/data': 0.0006153699941933155, 'timing/model': 1.9312935309717432, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 1.72}) +[2026-06-23 18:48:00,535][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.09701181948184967, 'train/grad_norm_pre_clip': 23.930498123168945, 'timing/data': 0.0006143311038613319, 'timing/model': 1.930515358923003, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 1.72}) +[2026-06-23 18:48:04,332][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.0038998310919851065, 'train/grad_norm_pre_clip': 24.88681983947754, 'timing/data': 0.0006067759823054075, 'timing/model': 1.9226360509637743, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 1.72}) +[2026-06-23 18:48:08,146][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.041806019842624664, 'train/grad_norm_pre_clip': 39.17002868652344, 'timing/data': 0.0006202709628269076, 'timing/model': 1.9420557629782706, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 1.72}) +[2026-06-23 18:48:11,981][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.02607405185699463, 'train/grad_norm_pre_clip': 54.012813568115234, 'timing/data': 0.0008094980148598552, 'timing/model': 1.9376090340083465, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 1.72}) +[2026-06-23 18:48:15,854][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.05691841244697571, 'train/grad_norm_pre_clip': 42.65935516357422, 'timing/data': 0.0006169350817799568, 'timing/model': 1.9875570180593058, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 1.72}) +[2026-06-23 18:48:19,676][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.04061416909098625, 'train/grad_norm_pre_clip': 20.099151611328125, 'timing/data': 0.0006106909131631255, 'timing/model': 1.937374801025726, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 1.72}) +[2026-06-23 18:48:23,474][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.019048567861318588, 'train/grad_norm_pre_clip': 21.315418243408203, 'timing/data': 0.0007320409640669823, 'timing/model': 1.9278631309280172, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 1.72}) +[2026-06-23 18:48:27,283][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.015496643260121346, 'train/grad_norm_pre_clip': 9.196643829345703, 'timing/data': 0.0008210920495912433, 'timing/model': 1.9341163990320638, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 1.72}) +[2026-06-23 18:48:31,086][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.04320460185408592, 'train/grad_norm_pre_clip': 34.89738845825195, 'timing/data': 0.0005982080474495888, 'timing/model': 1.9303358510369435, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 1.72}) +[2026-06-23 18:48:34,897][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.0792127475142479, 'train/grad_norm_pre_clip': 17.638526916503906, 'timing/data': 0.0006019709398970008, 'timing/model': 1.9407493349863216, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 1.73}) +[2026-06-23 18:48:38,705][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.024351941421628, 'train/grad_norm_pre_clip': 22.83449363708496, 'timing/data': 0.000760673894546926, 'timing/model': 1.9354323390871286, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 1.73}) +[2026-06-23 18:48:42,516][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.009808183647692204, 'train/grad_norm_pre_clip': 13.848101615905762, 'timing/data': 0.0006319049280136824, 'timing/model': 1.9385705209570006, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 1.73}) +[2026-06-23 18:48:46,325][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.02697725035250187, 'train/grad_norm_pre_clip': 53.832298278808594, 'timing/data': 0.0007330359658226371, 'timing/model': 1.9335790990153328, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 1.73}) +[2026-06-23 18:48:50,199][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.044814009219408035, 'train/grad_norm_pre_clip': 29.785770416259766, 'timing/data': 0.000606477027758956, 'timing/model': 1.9340304200304672, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 1.73}) +[2026-06-23 18:48:54,000][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.02156398631632328, 'train/grad_norm_pre_clip': 48.41658401489258, 'timing/data': 0.0008910709293559194, 'timing/model': 1.9316733030136675, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 1.73}) +[2026-06-23 18:48:57,796][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.020588461309671402, 'train/grad_norm_pre_clip': 32.63980484008789, 'timing/data': 0.0006637399783357978, 'timing/model': 1.9253780259750783, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 1.73}) +[2026-06-23 18:49:01,595][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.013372768647968769, 'train/grad_norm_pre_clip': 30.22698402404785, 'timing/data': 0.0006038809660822153, 'timing/model': 1.9329205009853467, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 1.73}) +[2026-06-23 18:49:05,410][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.11021921038627625, 'train/grad_norm_pre_clip': 23.291873931884766, 'timing/data': 0.0006462420569732785, 'timing/model': 1.9296725679887459, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 1.73}) +[2026-06-23 18:49:09,203][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.059273526072502136, 'train/grad_norm_pre_clip': 32.81160354614258, 'timing/data': 0.0008054350037127733, 'timing/model': 1.9289012619992718, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 1.73}) +[2026-06-23 18:49:13,060][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.04183453321456909, 'train/grad_norm_pre_clip': 29.69612693786621, 'timing/data': 0.0005950490012764931, 'timing/model': 1.9331903610145673, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 1.73}) +[2026-06-23 18:49:16,859][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.0795121043920517, 'train/grad_norm_pre_clip': 18.635543823242188, 'timing/data': 0.0005820139776915312, 'timing/model': 1.9285824309336022, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 1.73}) +[2026-06-23 18:49:20,690][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.02432865835726261, 'train/grad_norm_pre_clip': 22.647371292114258, 'timing/data': 0.0006204190431162715, 'timing/model': 1.9481885000132024, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 1.73}) +[2026-06-23 18:49:24,497][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.04781292378902435, 'train/grad_norm_pre_clip': 30.933921813964844, 'timing/data': 0.0007144809933379292, 'timing/model': 1.9297447920544073, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 1.73}) +[2026-06-23 18:49:28,291][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.02105499617755413, 'train/grad_norm_pre_clip': 12.092781066894531, 'timing/data': 0.0005981150316074491, 'timing/model': 1.927263921010308, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 1.73}) +[2026-06-23 18:49:32,104][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.016259901225566864, 'train/grad_norm_pre_clip': 15.178960800170898, 'timing/data': 0.0006068770308047533, 'timing/model': 1.929566349950619, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 1.73}) +[2026-06-23 18:49:35,905][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.0388803593814373, 'train/grad_norm_pre_clip': 30.6167049407959, 'timing/data': 0.0006439299322664738, 'timing/model': 1.9191723270341754, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 1.73}) +[2026-06-23 18:49:39,708][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.04067598283290863, 'train/grad_norm_pre_clip': 29.976070404052734, 'timing/data': 0.0005932640051469207, 'timing/model': 1.930486152996309, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 1.73}) +[2026-06-23 18:49:43,517][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.06855548918247223, 'train/grad_norm_pre_clip': 41.8642463684082, 'timing/data': 0.0005960289854556322, 'timing/model': 1.9340393139282241, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 1.73}) +[2026-06-23 18:49:47,336][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.023168088868260384, 'train/grad_norm_pre_clip': 21.085519790649414, 'timing/data': 0.0007403610507026315, 'timing/model': 1.9359426849987358, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 1.73}) +[2026-06-23 18:49:51,147][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.025304701179265976, 'train/grad_norm_pre_clip': 28.630102157592773, 'timing/data': 0.0006130519323050976, 'timing/model': 1.931191192008555, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 1.73}) +[2026-06-23 18:49:54,957][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.036372341215610504, 'train/grad_norm_pre_clip': 21.38789176940918, 'timing/data': 0.0006248010322451591, 'timing/model': 1.9272656519897282, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 1.73}) +[2026-06-23 18:49:58,760][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.03138260543346405, 'train/grad_norm_pre_clip': 21.600582122802734, 'timing/data': 0.000720218988135457, 'timing/model': 1.9305361319566146, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 1.73}) +[2026-06-23 18:50:02,627][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.07552967220544815, 'train/grad_norm_pre_clip': 31.695507049560547, 'timing/data': 0.0006223530508577824, 'timing/model': 1.9909430240513757, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 1.73}) +[2026-06-23 18:50:06,436][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.08475368469953537, 'train/grad_norm_pre_clip': 48.163150787353516, 'timing/data': 0.0006274349289014935, 'timing/model': 1.9311501419870183, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 1.73}) +[2026-06-23 18:50:10,240][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.0876985713839531, 'train/grad_norm_pre_clip': 30.952709197998047, 'timing/data': 0.0006226289551705122, 'timing/model': 1.9285957909887657, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 1.74}) +[2026-06-23 18:50:14,049][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.02879704348742962, 'train/grad_norm_pre_clip': 16.625730514526367, 'timing/data': 0.0010611190227791667, 'timing/model': 1.9312002520309761, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 1.74}) +[2026-06-23 18:50:17,847][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.0474235899746418, 'train/grad_norm_pre_clip': 22.273632049560547, 'timing/data': 0.0006358679383993149, 'timing/model': 1.929629456019029, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 1.74}) +[2026-06-23 18:50:21,647][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.022169310599565506, 'train/grad_norm_pre_clip': 13.463135719299316, 'timing/data': 0.0006746199214830995, 'timing/model': 1.9379205190343782, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 1.74}) +[2026-06-23 18:50:25,441][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.0452820248901844, 'train/grad_norm_pre_clip': 25.636186599731445, 'timing/data': 0.0006515389541164041, 'timing/model': 1.9284567988943309, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 1.74}) +[2026-06-23 18:50:29,413][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.07323824614286423, 'train/grad_norm_pre_clip': 20.89775276184082, 'timing/data': 0.0008501650299876928, 'timing/model': 1.920288312016055, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 1.74}) +[2026-06-23 18:50:33,213][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.0018082506721839309, 'train/grad_norm_pre_clip': 24.794958114624023, 'timing/data': 0.0007199320243671536, 'timing/model': 1.9316609039669856, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 1.74}) +[2026-06-23 18:50:37,039][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.06223355978727341, 'train/grad_norm_pre_clip': 20.79530906677246, 'timing/data': 0.0006568270036950707, 'timing/model': 1.9495981649961323, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 1.74}) +[2026-06-23 18:50:40,855][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.054270029067993164, 'train/grad_norm_pre_clip': 29.63230323791504, 'timing/data': 0.0005992379738017917, 'timing/model': 1.9342338780406862, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 1.74}) +[2026-06-23 18:50:44,654][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.06921201199293137, 'train/grad_norm_pre_clip': 26.032323837280273, 'timing/data': 0.0005963579751551151, 'timing/model': 1.9260624010348693, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 1.74}) +[2026-06-23 18:50:48,463][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.055011410266160965, 'train/grad_norm_pre_clip': 27.69831657409668, 'timing/data': 0.0007890169508755207, 'timing/model': 1.934196308022365, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 1.74}) +[2026-06-23 18:50:52,261][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.009054153226315975, 'train/grad_norm_pre_clip': 22.063438415527344, 'timing/data': 0.0006728799780830741, 'timing/model': 1.9275193939683959, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 1.74}) +[2026-06-23 18:50:56,064][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.057672545313835144, 'train/grad_norm_pre_clip': 21.43702507019043, 'timing/data': 0.0006503410404548049, 'timing/model': 1.9334622300229967, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 1.74}) +[2026-06-23 18:50:59,874][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.017760543152689934, 'train/grad_norm_pre_clip': 27.825685501098633, 'timing/data': 0.0005922470008954406, 'timing/model': 1.9375542000634596, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 1.74}) +[2026-06-23 18:51:03,693][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.04410979896783829, 'train/grad_norm_pre_clip': 42.92349624633789, 'timing/data': 0.0006325370632112026, 'timing/model': 1.9413947520079091, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 1.74}) +[2026-06-23 18:51:07,503][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.0427669920027256, 'train/grad_norm_pre_clip': 39.527828216552734, 'timing/data': 0.0007370560197159648, 'timing/model': 1.936734214075841, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 1.74}) +[2026-06-23 18:51:11,301][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.0336732342839241, 'train/grad_norm_pre_clip': 24.19176483154297, 'timing/data': 0.000817954889498651, 'timing/model': 1.9214675960829481, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 1.74}) +[2026-06-23 18:51:15,117][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4500 +[2026-06-23 18:53:22,042][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4500 +[2026-06-23 18:53:22,044][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4500 +[2026-06-23 18:54:45,317][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4500 +[2026-06-23 18:59:21,413][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4500 +[2026-06-23 18:59:21,414][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.08912591636180878, 'train/grad_norm_pre_clip': 47.485355377197266, 'eval/action_loss/seconds': 126.92253986001015, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.030599934980273247, 'eval/latency_2/loss': 0.030599934980273247, 'eval/flappy/loss': 0.030599934980273247, 'eval/flappy/latency_2/loss': 0.030599934980273247, 'eval/flap_precision': 0.9701492537313433, 'eval/flap_recall': 0.9649122807017544, 'eval/flap_f1': 0.9675236806495263, 'eval/flap_rate_model': 0.20472222222222222, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9675236806495263, 'eval/latency_2/flap_precision': 0.9701492537313433, 'eval/latency_2/flap_recall': 0.9649122807017544, 'eval/latency_2/flap_f1': 0.9675236806495263, 'eval/latency_2/flap_rate_model': 0.20472222222222222, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9675236806495263, 'eval/action_classification/data_wait_seconds': 1.3046931028366089, 'eval/action_classification/predict_seconds': 81.80339050292969, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 169.44000000000344, 'rl_games_eval/mid_train/mean_length': 1384.6, 'rl_games_eval/mid_train/std_reward': 156.5375558771846, 'rl_games_eval/mid_train/std_length': 1261.0880381638706, 'rl_games_eval/mid_train/macro_mean_reward': 169.44000000000344, 'rl_games_eval/mid_train/macro_mean_length': 1384.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 169.44000000000344, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1384.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 156.5375558771846, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1261.0880381638706, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 169.44000000000344, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006429720669984818, 'timing/model': 1.9263255799887702, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 1.74}) +[2026-06-23 18:59:21,414][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_4500_state +[2026-06-23 18:59:21,417][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 18:59:25,162][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_4500_state/model.safetensors +[2026-06-23 18:59:28,468][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_4500_state/optimizer.bin +[2026-06-23 18:59:28,468][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_4500_state/sampler.bin +[2026-06-23 18:59:28,468][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-23 18:59:28,469][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-23 18:59:28,495][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-23 18:59:28,495][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 18:59:28,506][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 18:59:32,225][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.02568318322300911, 'train/grad_norm_pre_clip': 27.35881996154785, 'timing/data': 0.0009129520040005445, 'timing/model': 1.8786057230317965, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 1.74}) +[2026-06-23 18:59:35,957][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.03567037358880043, 'train/grad_norm_pre_clip': 20.481290817260742, 'timing/data': 0.0006449120119214058, 'timing/model': 1.894461756106466, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 1.74}) +[2026-06-23 18:59:39,681][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.07751108705997467, 'train/grad_norm_pre_clip': 27.60411834716797, 'timing/data': 0.0006074840202927589, 'timing/model': 1.8939771050354466, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 1.74}) +[2026-06-23 18:59:43,430][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.010895374231040478, 'train/grad_norm_pre_clip': 4.775639057159424, 'timing/data': 0.0006040470907464623, 'timing/model': 1.9060947069665417, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 1.74}) +[2026-06-23 18:59:47,168][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.02416970580816269, 'train/grad_norm_pre_clip': 27.12350845336914, 'timing/data': 0.0006082279141992331, 'timing/model': 1.8923159140395, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 1.74}) +[2026-06-23 18:59:50,928][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.022983524948358536, 'train/grad_norm_pre_clip': 14.411033630371094, 'timing/data': 0.0006453689420595765, 'timing/model': 1.8952394099906087, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 1.74}) +[2026-06-23 18:59:54,682][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.038572825491428375, 'train/grad_norm_pre_clip': 22.035120010375977, 'timing/data': 0.0006005690665915608, 'timing/model': 1.899000525008887, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 1.74}) +[2026-06-23 18:59:58,444][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.0380934476852417, 'train/grad_norm_pre_clip': 23.927953720092773, 'timing/data': 0.0007655690424144268, 'timing/model': 1.9094523569801822, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 1.74}) +[2026-06-23 19:00:02,202][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.013760192319750786, 'train/grad_norm_pre_clip': 10.664849281311035, 'timing/data': 0.0006929249502718449, 'timing/model': 1.9104798460612074, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 1.75}) +[2026-06-23 19:00:06,131][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.08296242356300354, 'train/grad_norm_pre_clip': 34.41767883300781, 'timing/data': 0.0008142549777403474, 'timing/model': 2.078707888023928, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 1.75}) +[2026-06-23 19:00:09,883][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.15129454433918, 'train/grad_norm_pre_clip': 43.95016860961914, 'timing/data': 0.0006520430324599147, 'timing/model': 1.9041015839902684, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 1.75}) +[2026-06-23 19:00:13,643][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.2149011492729187, 'train/grad_norm_pre_clip': 63.54046630859375, 'timing/data': 0.0006097229197621346, 'timing/model': 1.8991142149316147, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 1.75}) +[2026-06-23 19:00:17,415][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.059701841324567795, 'train/grad_norm_pre_clip': 41.26557540893555, 'timing/data': 0.0006003440357744694, 'timing/model': 1.9065518629504368, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 1.75}) +[2026-06-23 19:00:21,170][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.04829223081469536, 'train/grad_norm_pre_clip': 25.78702163696289, 'timing/data': 0.0007500560022890568, 'timing/model': 1.9008592349709943, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 1.75}) +[2026-06-23 19:00:24,935][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.07749900221824646, 'train/grad_norm_pre_clip': 50.221641540527344, 'timing/data': 0.0007377410074695945, 'timing/model': 1.9112708519678563, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 1.75}) +[2026-06-23 19:00:28,707][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.040683645755052567, 'train/grad_norm_pre_clip': 44.53426742553711, 'timing/data': 0.0008611140074208379, 'timing/model': 1.9087118869647384, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 1.75}) +[2026-06-23 19:00:32,480][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.0674908459186554, 'train/grad_norm_pre_clip': 33.30501174926758, 'timing/data': 0.0006051460513845086, 'timing/model': 1.9136138369794935, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 1.75}) +[2026-06-23 19:00:36,266][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.0333215706050396, 'train/grad_norm_pre_clip': 19.245792388916016, 'timing/data': 0.0006046830676496029, 'timing/model': 1.9149616070790216, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 1.75}) +[2026-06-23 19:00:40,232][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.05174030736088753, 'train/grad_norm_pre_clip': 17.301008224487305, 'timing/data': 0.000599483959376812, 'timing/model': 1.9166964649921283, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 1.75}) +[2026-06-23 19:00:44,007][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.024934686720371246, 'train/grad_norm_pre_clip': 19.987380981445312, 'timing/data': 0.000607670983299613, 'timing/model': 1.91319301398471, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 1.75}) +[2026-06-23 19:00:47,801][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.015576452016830444, 'train/grad_norm_pre_clip': 30.898696899414062, 'timing/data': 0.0006027900381013751, 'timing/model': 1.92336349200923, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 1.75}) +[2026-06-23 19:00:51,590][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.01865440607070923, 'train/grad_norm_pre_clip': 21.201242446899414, 'timing/data': 0.0006729490123689175, 'timing/model': 1.9338084779446945, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 1.75}) +[2026-06-23 19:00:55,368][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.038564808666706085, 'train/grad_norm_pre_clip': 48.54488754272461, 'timing/data': 0.0006371949566528201, 'timing/model': 1.9263859179336578, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 1.75}) +[2026-06-23 19:00:59,154][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.04743433743715286, 'train/grad_norm_pre_clip': 36.0313835144043, 'timing/data': 0.0005841320380568504, 'timing/model': 1.9202863520476967, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 1.75}) +[2026-06-23 19:01:02,937][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.012918573804199696, 'train/grad_norm_pre_clip': 30.562103271484375, 'timing/data': 0.0007058550836518407, 'timing/model': 1.9155115929897875, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 1.75}) +[2026-06-23 19:01:06,729][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.06937768310308456, 'train/grad_norm_pre_clip': 26.84475326538086, 'timing/data': 0.0007020391058176756, 'timing/model': 1.921752192080021, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 1.75}) +[2026-06-23 19:01:10,695][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.05858761817216873, 'train/grad_norm_pre_clip': 26.880338668823242, 'timing/data': 0.0008379750652238727, 'timing/model': 2.1076623770641163, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 1.75}) +[2026-06-23 19:01:14,491][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.07605739682912827, 'train/grad_norm_pre_clip': 50.5246696472168, 'timing/data': 0.00082809804007411, 'timing/model': 1.9318658669944853, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 1.75}) +[2026-06-23 19:01:18,295][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.04012307897210121, 'train/grad_norm_pre_clip': 17.08305549621582, 'timing/data': 0.0008429249282926321, 'timing/model': 1.9293254780350253, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 1.75}) +[2026-06-23 19:01:22,082][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.0404270738363266, 'train/grad_norm_pre_clip': 23.762828826904297, 'timing/data': 0.0009591180132701993, 'timing/model': 1.9145076459972188, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 1.75}) +[2026-06-23 19:01:25,885][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.022856377065181732, 'train/grad_norm_pre_clip': 21.316041946411133, 'timing/data': 0.0006005300674587488, 'timing/model': 1.9264318599598482, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 1.75}) +[2026-06-23 19:01:29,668][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.02568596601486206, 'train/grad_norm_pre_clip': 33.799617767333984, 'timing/data': 0.0006301930407062173, 'timing/model': 1.9204124530078843, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 1.75}) +[2026-06-23 19:01:33,461][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.02619730681180954, 'train/grad_norm_pre_clip': 33.32468795776367, 'timing/data': 0.000920227961614728, 'timing/model': 1.9260439110221341, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 1.75}) +[2026-06-23 19:01:37,254][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.04787534102797508, 'train/grad_norm_pre_clip': 29.155038833618164, 'timing/data': 0.0006698670331388712, 'timing/model': 1.9243989610113204, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 1.75}) +[2026-06-23 19:01:41,050][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.030307887122035027, 'train/grad_norm_pre_clip': 29.203439712524414, 'timing/data': 0.0006480589509010315, 'timing/model': 1.9223453040467575, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 1.76}) +[2026-06-23 19:01:45,040][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.07584688067436218, 'train/grad_norm_pre_clip': 32.95969009399414, 'timing/data': 0.0006131629925221205, 'timing/model': 1.9295537980506197, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 1.76}) +[2026-06-23 19:01:48,831][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.041801996529102325, 'train/grad_norm_pre_clip': 30.489742279052734, 'timing/data': 0.0006019949214532971, 'timing/model': 1.9250363080063835, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 1.76}) +[2026-06-23 19:01:52,626][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.04637640714645386, 'train/grad_norm_pre_clip': 35.94546127319336, 'timing/data': 0.0006086039356887341, 'timing/model': 1.925382937071845, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 1.76}) +[2026-06-23 19:01:56,415][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.053558144718408585, 'train/grad_norm_pre_clip': 35.78876495361328, 'timing/data': 0.0005868850275874138, 'timing/model': 1.9220503449905664, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 1.76}) +[2026-06-23 19:02:00,208][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.020672401413321495, 'train/grad_norm_pre_clip': 13.823212623596191, 'timing/data': 0.0006574139697477221, 'timing/model': 1.926924666040577, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 1.76}) +[2026-06-23 19:02:03,995][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.10770457237958908, 'train/grad_norm_pre_clip': 50.90669250488281, 'timing/data': 0.0006680949591100216, 'timing/model': 1.92467434797436, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 1.76}) +[2026-06-23 19:02:07,801][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.08981353044509888, 'train/grad_norm_pre_clip': 26.925222396850586, 'timing/data': 0.0005933190695941448, 'timing/model': 1.9384058290161192, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 1.76}) +[2026-06-23 19:02:11,606][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.016028504818677902, 'train/grad_norm_pre_clip': 8.404044151306152, 'timing/data': 0.0005930979968979955, 'timing/model': 1.9390074849361554, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 1.76}) +[2026-06-23 19:02:15,544][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.022355439141392708, 'train/grad_norm_pre_clip': 19.281938552856445, 'timing/data': 0.0006067109061405063, 'timing/model': 2.067982246982865, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 1.76}) +[2026-06-23 19:02:19,354][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.04172160476446152, 'train/grad_norm_pre_clip': 39.28697204589844, 'timing/data': 0.000828166026622057, 'timing/model': 1.9305506258970127, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 1.76}) +[2026-06-23 19:02:23,159][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.10672134906053543, 'train/grad_norm_pre_clip': 22.92978858947754, 'timing/data': 0.0006128849927335978, 'timing/model': 1.9269271219382063, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 1.76}) +[2026-06-23 19:02:26,956][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.022365320473909378, 'train/grad_norm_pre_clip': 20.04945182800293, 'timing/data': 0.0007923640077933669, 'timing/model': 1.930718574905768, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 1.76}) +[2026-06-23 19:02:30,755][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.029340043663978577, 'train/grad_norm_pre_clip': 17.756683349609375, 'timing/data': 0.0008678409503772855, 'timing/model': 1.933439312968403, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 1.76}) +[2026-06-23 19:02:34,554][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.061057012528181076, 'train/grad_norm_pre_clip': 35.91788101196289, 'timing/data': 0.0006467599887400866, 'timing/model': 1.929022430907935, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 1.76}) +[2026-06-23 19:02:38,356][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.016829190775752068, 'train/grad_norm_pre_clip': 25.436660766601562, 'timing/data': 0.0006565810181200504, 'timing/model': 1.9302253049099818, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 1.76}) +[2026-06-23 19:02:42,167][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.048452116549015045, 'train/grad_norm_pre_clip': 62.24184799194336, 'timing/data': 0.0006062089232727885, 'timing/model': 1.933846842031926, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 1.76}) +[2026-06-23 19:02:45,966][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.09930267184972763, 'train/grad_norm_pre_clip': 62.040260314941406, 'timing/data': 0.0006009469507262111, 'timing/model': 1.9282589009962976, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 1.76}) +[2026-06-23 19:02:49,877][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.032831862568855286, 'train/grad_norm_pre_clip': 16.937236785888672, 'timing/data': 0.000858612940646708, 'timing/model': 1.9367869179695845, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 1.76}) +[2026-06-23 19:02:53,697][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.02908310480415821, 'train/grad_norm_pre_clip': 28.41504669189453, 'timing/data': 0.0006124279461801052, 'timing/model': 1.9369572639698163, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 1.76}) +[2026-06-23 19:02:57,509][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.1409294158220291, 'train/grad_norm_pre_clip': 36.614864349365234, 'timing/data': 0.0006009710486978292, 'timing/model': 1.9279098639963195, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 1.76}) +[2026-06-23 19:03:01,308][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.020738819614052773, 'train/grad_norm_pre_clip': 15.686244010925293, 'timing/data': 0.0006010900251567364, 'timing/model': 1.9195268549956381, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 1.76}) +[2026-06-23 19:03:05,107][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.03033697046339512, 'train/grad_norm_pre_clip': 26.217269897460938, 'timing/data': 0.000814983039163053, 'timing/model': 1.9258517490234226, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 1.76}) +[2026-06-23 19:03:08,915][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.05053430050611496, 'train/grad_norm_pre_clip': 41.14143371582031, 'timing/data': 0.0006022279849275947, 'timing/model': 1.9286563820205629, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 1.76}) +[2026-06-23 19:03:12,720][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.06947284191846848, 'train/grad_norm_pre_clip': 50.51384353637695, 'timing/data': 0.0007750949589535594, 'timing/model': 1.9259873050032184, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 1.76}) +[2026-06-23 19:03:16,522][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.11077815294265747, 'train/grad_norm_pre_clip': 40.33353042602539, 'timing/data': 0.0006238579517230392, 'timing/model': 1.9284723900491372, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 1.76}) +[2026-06-23 19:03:20,417][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.018757637590169907, 'train/grad_norm_pre_clip': 18.588640213012695, 'timing/data': 0.0008515340741723776, 'timing/model': 2.021874504047446, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 1.77}) +[2026-06-23 19:03:24,218][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.022839181125164032, 'train/grad_norm_pre_clip': 37.28425216674805, 'timing/data': 0.0005940119735896587, 'timing/model': 1.9277793139917776, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 1.77}) +[2026-06-23 19:03:28,018][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.09142305701971054, 'train/grad_norm_pre_clip': 38.45973205566406, 'timing/data': 0.0006244269898161292, 'timing/model': 1.9307612259872258, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 1.77}) +[2026-06-23 19:03:31,826][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.015870586037635803, 'train/grad_norm_pre_clip': 39.91462326049805, 'timing/data': 0.0007174769416451454, 'timing/model': 1.9299685190198943, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 1.77}) +[2026-06-23 19:03:35,616][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.03058015927672386, 'train/grad_norm_pre_clip': 30.609373092651367, 'timing/data': 0.000846287002786994, 'timing/model': 1.926802882924676, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 1.77}) +[2026-06-23 19:03:39,440][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.0764680951833725, 'train/grad_norm_pre_clip': 26.099077224731445, 'timing/data': 0.0006132900016382337, 'timing/model': 1.9421258000656962, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 1.77}) +[2026-06-23 19:03:43,253][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.011567256413400173, 'train/grad_norm_pre_clip': 19.401866912841797, 'timing/data': 0.0006408230401575565, 'timing/model': 1.933772615971975, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 1.77}) +[2026-06-23 19:03:47,051][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.02401454746723175, 'train/grad_norm_pre_clip': 23.829496383666992, 'timing/data': 0.0006062419852241874, 'timing/model': 1.9289857540279627, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 1.77}) +[2026-06-23 19:03:50,854][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.021707957610487938, 'train/grad_norm_pre_clip': 27.30622100830078, 'timing/data': 0.0008731839479878545, 'timing/model': 1.9302026429213583, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 1.77}) +[2026-06-23 19:03:54,738][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.024163439869880676, 'train/grad_norm_pre_clip': 29.068361282348633, 'timing/data': 0.0005901800468564034, 'timing/model': 1.9306058880174533, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 1.77}) +[2026-06-23 19:03:58,548][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.05948173627257347, 'train/grad_norm_pre_clip': 34.1321907043457, 'timing/data': 0.0008935430087149143, 'timing/model': 1.9371201230678707, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 1.77}) +[2026-06-23 19:04:02,363][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.02683420479297638, 'train/grad_norm_pre_clip': 36.47441864013672, 'timing/data': 0.000584024004638195, 'timing/model': 1.9308371969964355, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 1.77}) +[2026-06-23 19:04:06,191][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.013770200312137604, 'train/grad_norm_pre_clip': 18.959312438964844, 'timing/data': 0.0008284570649266243, 'timing/model': 1.9277764200232923, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 1.77}) +[2026-06-23 19:04:09,997][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.017344243824481964, 'train/grad_norm_pre_clip': 15.137856483459473, 'timing/data': 0.0005990869831293821, 'timing/model': 1.9344011869980022, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 1.77}) +[2026-06-23 19:04:13,800][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.034593742340803146, 'train/grad_norm_pre_clip': 25.179611206054688, 'timing/data': 0.0007374320412054658, 'timing/model': 1.9301299289800227, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 1.77}) +[2026-06-23 19:04:17,606][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.02395172230899334, 'train/grad_norm_pre_clip': 34.42988967895508, 'timing/data': 0.0008269600803032517, 'timing/model': 1.9298835230292752, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 1.77}) +[2026-06-23 19:04:21,401][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.013525793328881264, 'train/grad_norm_pre_clip': 25.459932327270508, 'timing/data': 0.0006243790267035365, 'timing/model': 1.9228621199727058, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 1.77}) +[2026-06-23 19:04:25,223][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.016242504119873047, 'train/grad_norm_pre_clip': 16.89612579345703, 'timing/data': 0.000646354048512876, 'timing/model': 1.9526053189765662, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 1.77}) +[2026-06-23 19:04:29,018][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.11561711877584457, 'train/grad_norm_pre_clip': 48.59768295288086, 'timing/data': 0.000596456928178668, 'timing/model': 1.9309585510054603, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 1.77}) +[2026-06-23 19:04:32,826][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.031943581998348236, 'train/grad_norm_pre_clip': 21.69419288635254, 'timing/data': 0.0006069410592317581, 'timing/model': 1.9283898280700669, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 1.77}) +[2026-06-23 19:04:36,632][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.010108254849910736, 'train/grad_norm_pre_clip': 33.38418197631836, 'timing/data': 0.0006486150668933988, 'timing/model': 1.9310684750089422, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 1.77}) +[2026-06-23 19:04:40,425][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.03756660968065262, 'train/grad_norm_pre_clip': 21.295289993286133, 'timing/data': 0.0006314680213108659, 'timing/model': 1.928379593999125, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 1.77}) +[2026-06-23 19:04:44,243][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.0233580581843853, 'train/grad_norm_pre_clip': 10.743401527404785, 'timing/data': 0.0009026449406519532, 'timing/model': 1.9416361940093338, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 1.77}) +[2026-06-23 19:04:48,036][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.004748867359012365, 'train/grad_norm_pre_clip': 11.099807739257812, 'timing/data': 0.0005944949807599187, 'timing/model': 1.926570211071521, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 1.77}) +[2026-06-23 19:04:51,840][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.08899158239364624, 'train/grad_norm_pre_clip': 34.67596435546875, 'timing/data': 0.0008403100073337555, 'timing/model': 1.9180956820491701, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 1.77}) +[2026-06-23 19:04:55,648][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.04852134361863136, 'train/grad_norm_pre_clip': 25.266725540161133, 'timing/data': 0.0006235759938135743, 'timing/model': 1.9384492470417172, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 1.77}) +[2026-06-23 19:04:59,580][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.08673252165317535, 'train/grad_norm_pre_clip': 51.35186767578125, 'timing/data': 0.0008409220026805997, 'timing/model': 2.0569471509661525, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 1.78}) +[2026-06-23 19:05:03,378][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.04249334707856178, 'train/grad_norm_pre_clip': 54.62439727783203, 'timing/data': 0.0006289079319685698, 'timing/model': 1.9261481469729915, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 1.78}) +[2026-06-23 19:05:07,176][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.14864155650138855, 'train/grad_norm_pre_clip': 59.75651931762695, 'timing/data': 0.0005887129809707403, 'timing/model': 1.9293132269522175, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 1.78}) +[2026-06-23 19:05:10,963][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.012568406760692596, 'train/grad_norm_pre_clip': 10.572639465332031, 'timing/data': 0.0005919000832363963, 'timing/model': 1.9205157640390098, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 1.78}) +[2026-06-23 19:05:14,775][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.008799729868769646, 'train/grad_norm_pre_clip': 11.937417030334473, 'timing/data': 0.0006204199744388461, 'timing/model': 1.9307857099920511, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 1.78}) +[2026-06-23 19:05:18,574][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.03413951396942139, 'train/grad_norm_pre_clip': 25.672809600830078, 'timing/data': 0.0005958830006420612, 'timing/model': 1.9244636329822242, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 1.78}) +[2026-06-23 19:05:22,378][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.0744721069931984, 'train/grad_norm_pre_clip': 64.23676300048828, 'timing/data': 0.000621568993665278, 'timing/model': 1.9268154189921916, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 1.78}) +[2026-06-23 19:05:26,196][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.021146317943930626, 'train/grad_norm_pre_clip': 48.21848678588867, 'timing/data': 0.00059068005066365, 'timing/model': 1.9333542379317805, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 1.78}) +[2026-06-23 19:05:29,997][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.02074696309864521, 'train/grad_norm_pre_clip': 25.670358657836914, 'timing/data': 0.0006068269722163677, 'timing/model': 1.9307543160393834, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 1.78}) +[2026-06-23 19:05:33,982][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.03242901712656021, 'train/grad_norm_pre_clip': 41.52043914794922, 'timing/data': 0.000625019078142941, 'timing/model': 2.101389891002327, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 1.78}) +[2026-06-23 19:05:37,783][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.011502286419272423, 'train/grad_norm_pre_clip': 25.05202865600586, 'timing/data': 0.0007864580256864429, 'timing/model': 1.9213895670836791, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 1.78}) +[2026-06-23 19:05:41,606][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.024566536769270897, 'train/grad_norm_pre_clip': 19.070180892944336, 'timing/data': 0.0009165389928966761, 'timing/model': 1.9443347610067576, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 1.78}) +[2026-06-23 19:05:45,419][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.03028726764023304, 'train/grad_norm_pre_clip': 49.536197662353516, 'timing/data': 0.0006085350178182125, 'timing/model': 1.9326995690353215, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 1.78}) +[2026-06-23 19:05:49,223][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4600 +[2026-06-23 19:07:56,132][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4600 +[2026-06-23 19:07:56,133][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4600 +[2026-06-23 19:09:19,508][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4600 +[2026-06-23 19:09:19,509][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.01133037731051445, 'train/grad_norm_pre_clip': 27.938343048095703, 'eval/action_loss/seconds': 126.90525899396744, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.0323982909321785, 'eval/latency_2/loss': 0.0323982909321785, 'eval/flappy/loss': 0.0323982909321785, 'eval/flappy/latency_2/loss': 0.0323982909321785, 'eval/flap_precision': 0.9651006711409396, 'eval/flap_recall': 0.970310391363023, 'eval/flap_f1': 0.9676985195154777, 'eval/flap_rate_model': 0.20694444444444443, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9676985195154778, 'eval/latency_2/flap_precision': 0.9651006711409396, 'eval/latency_2/flap_recall': 0.970310391363023, 'eval/latency_2/flap_f1': 0.9676985195154777, 'eval/latency_2/flap_rate_model': 0.20694444444444443, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9676985195154778, 'eval/action_classification/data_wait_seconds': 1.2286231517791748, 'eval/action_classification/predict_seconds': 81.94302368164062, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006050039082765579, 'timing/model': 1.9301845260197297, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 1.78}) +[2026-06-23 19:09:23,314][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.024011453613638878, 'train/grad_norm_pre_clip': 43.077247619628906, 'timing/data': 0.0006511869141831994, 'timing/model': 1.9324736549751833, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 1.78}) +[2026-06-23 19:09:27,110][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.04180049151182175, 'train/grad_norm_pre_clip': 34.94727325439453, 'timing/data': 0.0006568559911102057, 'timing/model': 1.9333710329374298, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 1.78}) +[2026-06-23 19:09:30,919][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.017684003338217735, 'train/grad_norm_pre_clip': 22.812606811523438, 'timing/data': 0.0005964830052107573, 'timing/model': 1.9347759779775515, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 1.78}) +[2026-06-23 19:09:34,715][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.021457914263010025, 'train/grad_norm_pre_clip': 39.010780334472656, 'timing/data': 0.0008617480052635074, 'timing/model': 1.9296637719962746, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 1.78}) +[2026-06-23 19:09:38,512][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.05275098606944084, 'train/grad_norm_pre_clip': 32.46177291870117, 'timing/data': 0.0007699630223214626, 'timing/model': 1.9313399479724467, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 1.78}) +[2026-06-23 19:09:42,304][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.01950610801577568, 'train/grad_norm_pre_clip': 22.656763076782227, 'timing/data': 0.0006873490056023002, 'timing/model': 1.9283881050068885, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 1.78}) +[2026-06-23 19:09:46,094][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.02527095563709736, 'train/grad_norm_pre_clip': 29.585012435913086, 'timing/data': 0.0008189859800040722, 'timing/model': 1.9152121240040287, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 1.78}) +[2026-06-23 19:09:49,886][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.010380387306213379, 'train/grad_norm_pre_clip': 20.528226852416992, 'timing/data': 0.000916345976293087, 'timing/model': 1.926884796936065, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 1.78}) +[2026-06-23 19:09:53,857][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.053498610854148865, 'train/grad_norm_pre_clip': 42.46700668334961, 'timing/data': 0.0007639139657840133, 'timing/model': 2.101771099958569, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 1.78}) +[2026-06-23 19:09:57,650][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.07170012593269348, 'train/grad_norm_pre_clip': 41.54366683959961, 'timing/data': 0.0006216309266164899, 'timing/model': 1.9114183250349015, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 1.78}) +[2026-06-23 19:10:01,451][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.030645702034235, 'train/grad_norm_pre_clip': 27.198366165161133, 'timing/data': 0.0006295570638030767, 'timing/model': 1.9340711500262842, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 1.78}) +[2026-06-23 19:10:05,250][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.010797462426126003, 'train/grad_norm_pre_clip': 32.20178985595703, 'timing/data': 0.0006038809660822153, 'timing/model': 1.9252820740221068, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 1.78}) +[2026-06-23 19:10:09,029][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.030008915811777115, 'train/grad_norm_pre_clip': 56.294010162353516, 'timing/data': 0.0005947310710325837, 'timing/model': 1.9207768299384043, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 1.79}) +[2026-06-23 19:10:12,824][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.036961689591407776, 'train/grad_norm_pre_clip': 23.30678939819336, 'timing/data': 0.0006275320192798972, 'timing/model': 1.9274802419822663, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 1.79}) +[2026-06-23 19:10:16,619][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.020773043856024742, 'train/grad_norm_pre_clip': 18.160974502563477, 'timing/data': 0.0006334040081128478, 'timing/model': 1.929459081031382, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 1.79}) +[2026-06-23 19:10:20,401][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.09899082779884338, 'train/grad_norm_pre_clip': 49.3065299987793, 'timing/data': 0.0007137330248951912, 'timing/model': 1.918020120006986, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 1.79}) +[2026-06-23 19:10:24,214][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.0008375709294341505, 'train/grad_norm_pre_clip': 23.744674682617188, 'timing/data': 0.0008718090830370784, 'timing/model': 1.9351249999599531, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 1.79}) +[2026-06-23 19:10:28,196][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.05362584814429283, 'train/grad_norm_pre_clip': 22.921815872192383, 'timing/data': 0.0006022339221090078, 'timing/model': 1.929262330988422, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 1.79}) +[2026-06-23 19:10:31,986][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.0021415851078927517, 'train/grad_norm_pre_clip': 65.46469116210938, 'timing/data': 0.0006061679450795054, 'timing/model': 1.9248696350259706, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 1.79}) +[2026-06-23 19:10:35,781][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.022343629971146584, 'train/grad_norm_pre_clip': 55.45879364013672, 'timing/data': 0.0005801940569654107, 'timing/model': 1.922872163937427, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 1.79}) +[2026-06-23 19:10:39,566][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.0278024785220623, 'train/grad_norm_pre_clip': 13.804224014282227, 'timing/data': 0.0006120089674368501, 'timing/model': 1.922341061057523, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 1.79}) +[2026-06-23 19:10:43,364][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.06462420523166656, 'train/grad_norm_pre_clip': 45.93412399291992, 'timing/data': 0.0005827700952067971, 'timing/model': 1.9278787069488317, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 1.79}) +[2026-06-23 19:10:47,154][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.04201633110642433, 'train/grad_norm_pre_clip': 51.853145599365234, 'timing/data': 0.0007320840377360582, 'timing/model': 1.9234672040911391, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 1.79}) +[2026-06-23 19:10:50,938][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.02952827699482441, 'train/grad_norm_pre_clip': 88.3643569946289, 'timing/data': 0.0006094910204410553, 'timing/model': 1.9192737709963694, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 1.79}) +[2026-06-23 19:10:54,746][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.1603761464357376, 'train/grad_norm_pre_clip': 37.1994743347168, 'timing/data': 0.0006185539532452822, 'timing/model': 1.9311230549355969, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 1.79}) +[2026-06-23 19:10:58,743][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.04817810282111168, 'train/grad_norm_pre_clip': 53.82082748413086, 'timing/data': 0.0006298969965428114, 'timing/model': 2.1253682259703055, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 1.79}) +[2026-06-23 19:11:02,539][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.024318121373653412, 'train/grad_norm_pre_clip': 33.240909576416016, 'timing/data': 0.0005896029761061072, 'timing/model': 1.9275248190388083, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 1.79}) +[2026-06-23 19:11:06,345][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.032415151596069336, 'train/grad_norm_pre_clip': 44.090599060058594, 'timing/data': 0.0005993599770590663, 'timing/model': 1.9303571740165353, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 1.79}) +[2026-06-23 19:11:10,155][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.031248541548848152, 'train/grad_norm_pre_clip': 19.707836151123047, 'timing/data': 0.000767786055803299, 'timing/model': 1.9366138119949028, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 1.79}) +[2026-06-23 19:11:13,976][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.001986625837162137, 'train/grad_norm_pre_clip': 20.72231101989746, 'timing/data': 0.0007824499625712633, 'timing/model': 1.9372566629899666, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 1.79}) +[2026-06-23 19:11:17,784][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.10226994752883911, 'train/grad_norm_pre_clip': 48.24978256225586, 'timing/data': 0.0006077209254726768, 'timing/model': 1.9259369929786772, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 1.79}) +[2026-06-23 19:11:21,593][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.08571210503578186, 'train/grad_norm_pre_clip': 53.10081100463867, 'timing/data': 0.000775514985434711, 'timing/model': 1.9344364199787378, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 1.79}) +[2026-06-23 19:11:25,378][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.03961925208568573, 'train/grad_norm_pre_clip': 24.75551414489746, 'timing/data': 0.0007828359957784414, 'timing/model': 1.9212368170265108, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 1.79}) +[2026-06-23 19:11:29,195][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.010191511362791061, 'train/grad_norm_pre_clip': 40.8750114440918, 'timing/data': 0.0008979140548035502, 'timing/model': 1.9373355200514197, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 1.79}) +[2026-06-23 19:11:32,980][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.07249344885349274, 'train/grad_norm_pre_clip': 36.37530517578125, 'timing/data': 0.0006000030552968383, 'timing/model': 1.919563967967406, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 1.79}) +[2026-06-23 19:11:36,814][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.03348308429121971, 'train/grad_norm_pre_clip': 24.81583595275879, 'timing/data': 0.0006046909838914871, 'timing/model': 1.9551953870104626, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 1.79}) +[2026-06-23 19:11:40,605][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.046764716506004333, 'train/grad_norm_pre_clip': 35.64712905883789, 'timing/data': 0.0006033700192347169, 'timing/model': 1.9174077810021117, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 1.79}) +[2026-06-23 19:11:44,415][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.007214525248855352, 'train/grad_norm_pre_clip': 22.00242805480957, 'timing/data': 0.0006013229722157121, 'timing/model': 1.9331247629597783, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 1.79}) +[2026-06-23 19:11:48,219][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.10608138889074326, 'train/grad_norm_pre_clip': 45.172550201416016, 'timing/data': 0.0006239240756258368, 'timing/model': 1.9301524501061067, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 1.8}) +[2026-06-23 19:11:52,025][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.06390441209077835, 'train/grad_norm_pre_clip': 29.092723846435547, 'timing/data': 0.0006596510065719485, 'timing/model': 1.9301635320298374, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 1.8}) +[2026-06-23 19:11:55,876][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.02692730352282524, 'train/grad_norm_pre_clip': 26.164621353149414, 'timing/data': 0.0008377910126000643, 'timing/model': 1.9293127889977768, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 1.8}) +[2026-06-23 19:11:59,703][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.03210565820336342, 'train/grad_norm_pre_clip': 17.78978729248047, 'timing/data': 0.000633800053037703, 'timing/model': 1.9333535800687969, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 1.8}) +[2026-06-23 19:12:03,508][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.056348785758018494, 'train/grad_norm_pre_clip': 42.17227554321289, 'timing/data': 0.0007422150811180472, 'timing/model': 1.9322354330215603, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 1.8}) +[2026-06-23 19:12:07,304][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.16088013350963593, 'train/grad_norm_pre_clip': 50.576866149902344, 'timing/data': 0.0007155640050768852, 'timing/model': 1.9330510650761425, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 1.8}) +[2026-06-23 19:12:11,120][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.025147229433059692, 'train/grad_norm_pre_clip': 28.4583740234375, 'timing/data': 0.0007103250827640295, 'timing/model': 1.9364493750035763, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 1.8}) +[2026-06-23 19:12:14,916][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.05074573680758476, 'train/grad_norm_pre_clip': 86.26168060302734, 'timing/data': 0.000740888062864542, 'timing/model': 1.9317310199839994, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 1.8}) +[2026-06-23 19:12:18,723][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.06989697366952896, 'train/grad_norm_pre_clip': 74.27112579345703, 'timing/data': 0.0006012220401316881, 'timing/model': 1.9270825360435992, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 1.8}) +[2026-06-23 19:12:22,524][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.0380948930978775, 'train/grad_norm_pre_clip': 40.018638610839844, 'timing/data': 0.000601690961048007, 'timing/model': 1.9225079660536721, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 1.8}) +[2026-06-23 19:12:26,323][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.017362581565976143, 'train/grad_norm_pre_clip': 19.815547943115234, 'timing/data': 0.0005941250128671527, 'timing/model': 1.9278691819636151, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 1.8}) +[2026-06-23 19:12:30,128][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.040203891694545746, 'train/grad_norm_pre_clip': 21.19833755493164, 'timing/data': 0.0006212570006027818, 'timing/model': 1.9362859450047836, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 1.8}) +[2026-06-23 19:12:33,997][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.04843892902135849, 'train/grad_norm_pre_clip': 35.57986068725586, 'timing/data': 0.0006291649769991636, 'timing/model': 1.999467305955477, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 1.8}) +[2026-06-23 19:12:37,787][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.010603299364447594, 'train/grad_norm_pre_clip': 49.03565979003906, 'timing/data': 0.0006343210116028786, 'timing/model': 1.9274737959494814, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 1.8}) +[2026-06-23 19:12:41,628][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.03161565214395523, 'train/grad_norm_pre_clip': 34.909759521484375, 'timing/data': 0.0006258649518713355, 'timing/model': 1.9203182779019699, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 1.8}) +[2026-06-23 19:12:45,441][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.002111512701958418, 'train/grad_norm_pre_clip': 6.458787441253662, 'timing/data': 0.0006209599087014794, 'timing/model': 1.9379725330509245, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 1.8}) +[2026-06-23 19:12:49,259][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.06051184982061386, 'train/grad_norm_pre_clip': 51.50144958496094, 'timing/data': 0.0006104350322857499, 'timing/model': 1.936610742006451, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 1.8}) +[2026-06-23 19:12:53,057][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.0024430365301668644, 'train/grad_norm_pre_clip': 6.169067859649658, 'timing/data': 0.0007538000354543328, 'timing/model': 1.9240735820494592, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 1.8}) +[2026-06-23 19:12:56,854][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.09501989930868149, 'train/grad_norm_pre_clip': 46.88931655883789, 'timing/data': 0.0006238790228962898, 'timing/model': 1.9307463630102575, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 1.8}) +[2026-06-23 19:13:00,658][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.0038977544754743576, 'train/grad_norm_pre_clip': 53.72529220581055, 'timing/data': 0.0006047500064596534, 'timing/model': 1.9374003120465204, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 1.8}) +[2026-06-23 19:13:04,465][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.016902294009923935, 'train/grad_norm_pre_clip': 30.46111297607422, 'timing/data': 0.000617304933257401, 'timing/model': 1.933547527063638, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 1.8}) +[2026-06-23 19:13:08,441][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.08225464820861816, 'train/grad_norm_pre_clip': 51.0801887512207, 'timing/data': 0.0006021929439157248, 'timing/model': 2.10145414294675, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 1.8}) +[2026-06-23 19:13:12,306][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.11709441244602203, 'train/grad_norm_pre_clip': 53.38709259033203, 'timing/data': 0.0006788729224354029, 'timing/model': 1.9238509440328926, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 1.8}) +[2026-06-23 19:13:16,105][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.021022969856858253, 'train/grad_norm_pre_clip': 58.05482482910156, 'timing/data': 0.0006107819499447942, 'timing/model': 1.9256970649585128, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 1.8}) +[2026-06-23 19:13:19,913][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.031355760991573334, 'train/grad_norm_pre_clip': 41.352027893066406, 'timing/data': 0.0006049770163372159, 'timing/model': 1.9383105280576274, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 1.8}) +[2026-06-23 19:13:23,714][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.04147830605506897, 'train/grad_norm_pre_clip': 52.266727447509766, 'timing/data': 0.0008675899589434266, 'timing/model': 1.9317182230297476, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 1.81}) +[2026-06-23 19:13:27,504][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.040593165904283524, 'train/grad_norm_pre_clip': 35.77664566040039, 'timing/data': 0.0006157959578558803, 'timing/model': 1.918747336952947, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 1.81}) +[2026-06-23 19:13:31,308][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.08108401298522949, 'train/grad_norm_pre_clip': 62.42650604248047, 'timing/data': 0.0006131629925221205, 'timing/model': 1.9301828150637448, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 1.81}) +[2026-06-23 19:13:35,111][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.020668916404247284, 'train/grad_norm_pre_clip': 24.5994873046875, 'timing/data': 0.0007657340029254556, 'timing/model': 1.9253047660458833, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 1.81}) +[2026-06-23 19:13:38,902][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.029325230047106743, 'train/grad_norm_pre_clip': 23.8637638092041, 'timing/data': 0.0007354880217462778, 'timing/model': 1.9196866190759465, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 1.81}) +[2026-06-23 19:13:42,879][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.05711774900555611, 'train/grad_norm_pre_clip': 39.333980560302734, 'timing/data': 0.0006796990055590868, 'timing/model': 1.9224558040732518, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 1.81}) +[2026-06-23 19:13:46,688][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.09256584197282791, 'train/grad_norm_pre_clip': 32.16012191772461, 'timing/data': 0.0006693260511383414, 'timing/model': 1.935347810969688, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 1.81}) +[2026-06-23 19:13:50,493][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.024347849190235138, 'train/grad_norm_pre_clip': 17.863441467285156, 'timing/data': 0.0005977100227028131, 'timing/model': 1.93116683291737, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 1.81}) +[2026-06-23 19:13:54,293][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.04407527670264244, 'train/grad_norm_pre_clip': 45.31208419799805, 'timing/data': 0.0007992580067366362, 'timing/model': 1.9251877490896732, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 1.81}) +[2026-06-23 19:13:58,164][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.10612709075212479, 'train/grad_norm_pre_clip': 42.66394805908203, 'timing/data': 0.0006401559803634882, 'timing/model': 1.9291753770085052, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 1.81}) +[2026-06-23 19:14:01,950][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.07127411663532257, 'train/grad_norm_pre_clip': 45.0648193359375, 'timing/data': 0.0006281749811023474, 'timing/model': 1.92235940892715, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 1.81}) +[2026-06-23 19:14:05,756][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.03700471296906471, 'train/grad_norm_pre_clip': 36.25300598144531, 'timing/data': 0.0006604909431189299, 'timing/model': 1.92604767694138, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 1.81}) +[2026-06-23 19:14:09,562][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.040766190737485886, 'train/grad_norm_pre_clip': 29.370773315429688, 'timing/data': 0.0006369230104610324, 'timing/model': 1.9307029760675505, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 1.81}) +[2026-06-23 19:14:13,548][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.003765685949474573, 'train/grad_norm_pre_clip': 3.634884834289551, 'timing/data': 0.0007959710201248527, 'timing/model': 2.097733919043094, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 1.81}) +[2026-06-23 19:14:17,351][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.03915733844041824, 'train/grad_norm_pre_clip': 26.819110870361328, 'timing/data': 0.0006165470695123076, 'timing/model': 1.9354476019507274, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 1.81}) +[2026-06-23 19:14:21,153][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.02170468121767044, 'train/grad_norm_pre_clip': 22.276090621948242, 'timing/data': 0.0007729320786893368, 'timing/model': 1.9343835789477453, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 1.81}) +[2026-06-23 19:14:24,970][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.01977390982210636, 'train/grad_norm_pre_clip': 13.443087577819824, 'timing/data': 0.0007779729785397649, 'timing/model': 1.936089011025615, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 1.81}) +[2026-06-23 19:14:28,768][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.010958109050989151, 'train/grad_norm_pre_clip': 21.285268783569336, 'timing/data': 0.0005799069767817855, 'timing/model': 1.9278236499521881, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 1.81}) +[2026-06-23 19:14:32,566][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.08454669266939163, 'train/grad_norm_pre_clip': 49.31235122680664, 'timing/data': 0.0006001889705657959, 'timing/model': 1.9239989729830995, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 1.81}) +[2026-06-23 19:14:36,382][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.008533859625458717, 'train/grad_norm_pre_clip': 8.441910743713379, 'timing/data': 0.0006222659721970558, 'timing/model': 1.9442647519754246, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 1.81}) +[2026-06-23 19:14:40,179][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.013180019333958626, 'train/grad_norm_pre_clip': 15.416807174682617, 'timing/data': 0.0008875499479472637, 'timing/model': 1.9274185670074075, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 1.81}) +[2026-06-23 19:14:43,973][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.09341716021299362, 'train/grad_norm_pre_clip': 55.90813064575195, 'timing/data': 0.0006129669491201639, 'timing/model': 1.9216412659734488, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 1.81}) +[2026-06-23 19:14:47,782][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.12970644235610962, 'train/grad_norm_pre_clip': 47.87118911743164, 'timing/data': 0.0006466250633820891, 'timing/model': 1.93539909995161, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 1.81}) +[2026-06-23 19:14:51,584][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.03523336350917816, 'train/grad_norm_pre_clip': 35.06068420410156, 'timing/data': 0.0006259339861571789, 'timing/model': 1.9233267849776894, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 1.81}) +[2026-06-23 19:14:55,393][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.012986722402274609, 'train/grad_norm_pre_clip': 49.985206604003906, 'timing/data': 0.0005861569661647081, 'timing/model': 1.9309381149942055, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 1.81}) +[2026-06-23 19:14:59,202][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.021811898797750473, 'train/grad_norm_pre_clip': 13.72062873840332, 'timing/data': 0.00079542794264853, 'timing/model': 1.9366200639633462, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 1.81}) +[2026-06-23 19:15:03,028][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.03957011178135872, 'train/grad_norm_pre_clip': 30.93702507019043, 'timing/data': 0.0006202849326655269, 'timing/model': 1.944009842001833, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 1.82}) +[2026-06-23 19:15:06,830][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.013755647465586662, 'train/grad_norm_pre_clip': 47.23733901977539, 'timing/data': 0.0006013780366629362, 'timing/model': 1.9319964569294825, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 1.82}) +[2026-06-23 19:15:10,653][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.010670645162463188, 'train/grad_norm_pre_clip': 41.53165817260742, 'timing/data': 0.0006561810150742531, 'timing/model': 1.9337172079831362, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 1.82}) +[2026-06-23 19:15:14,468][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.05563909560441971, 'train/grad_norm_pre_clip': 63.34040832519531, 'timing/data': 0.0006134649738669395, 'timing/model': 1.9376925849355757, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 1.82}) +[2026-06-23 19:15:18,269][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.03198779746890068, 'train/grad_norm_pre_clip': 22.280193328857422, 'timing/data': 0.0005979370325803757, 'timing/model': 1.9328848390141502, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 1.82}) +[2026-06-23 19:15:22,246][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.08949872106313705, 'train/grad_norm_pre_clip': 94.11460876464844, 'timing/data': 0.0008216840215027332, 'timing/model': 1.9224363110261038, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 1.82}) +[2026-06-23 19:15:26,131][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.03223421052098274, 'train/grad_norm_pre_clip': 39.75995635986328, 'timing/data': 0.0006069230148568749, 'timing/model': 1.9560215899255127, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 1.82}) +[2026-06-23 19:15:29,922][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.04647326469421387, 'train/grad_norm_pre_clip': 42.526424407958984, 'timing/data': 0.0007524919928982854, 'timing/model': 1.920613085036166, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 1.82}) +[2026-06-23 19:15:33,736][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.092585988342762, 'train/grad_norm_pre_clip': 23.17913818359375, 'timing/data': 0.0006085559725761414, 'timing/model': 1.9364313459955156, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 1.82}) +[2026-06-23 19:15:37,533][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.0647822916507721, 'train/grad_norm_pre_clip': 157.4410400390625, 'timing/data': 0.0005947619210928679, 'timing/model': 1.9289492069510743, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 1.82}) +[2026-06-23 19:15:41,331][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4700 +[2026-06-23 19:17:50,215][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4700 +[2026-06-23 19:17:50,217][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4700 +[2026-06-23 19:19:12,528][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4700 +[2026-06-23 19:19:12,529][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.03633277118206024, 'train/grad_norm_pre_clip': 21.022830963134766, 'eval/action_loss/seconds': 128.8809460250195, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03262188285589218, 'eval/latency_2/loss': 0.03262188285589218, 'eval/flappy/loss': 0.03262188285589218, 'eval/flappy/latency_2/loss': 0.03262188285589218, 'eval/flap_precision': 0.967654986522911, 'eval/flap_recall': 0.9689608636977058, 'eval/flap_f1': 0.9683074848280513, 'eval/flap_rate_model': 0.2061111111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9683074848280513, 'eval/latency_2/flap_precision': 0.967654986522911, 'eval/latency_2/flap_recall': 0.9689608636977058, 'eval/latency_2/flap_f1': 0.9683074848280513, 'eval/latency_2/flap_rate_model': 0.2061111111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9683074848280513, 'eval/action_classification/data_wait_seconds': 1.5093674659729004, 'eval/action_classification/predict_seconds': 80.58406066894531, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0007855050498619676, 'timing/model': 1.9314895329298452, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 1.82}) +[2026-06-23 19:19:16,314][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.15688835084438324, 'train/grad_norm_pre_clip': 38.88733673095703, 'timing/data': 0.0005987549666315317, 'timing/model': 1.9177842179778963, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 1.82}) +[2026-06-23 19:19:20,086][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.13064691424369812, 'train/grad_norm_pre_clip': 22.41952133178711, 'timing/data': 0.0006063450127840042, 'timing/model': 1.9202514439821243, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 1.82}) +[2026-06-23 19:19:23,873][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.18612097203731537, 'train/grad_norm_pre_clip': 45.93244934082031, 'timing/data': 0.0009145719232037663, 'timing/model': 1.9283385179005563, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 1.82}) +[2026-06-23 19:19:27,653][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.06163153797388077, 'train/grad_norm_pre_clip': 51.46840286254883, 'timing/data': 0.0006341940024867654, 'timing/model': 1.9154801110271364, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 1.82}) +[2026-06-23 19:19:31,444][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.0295847337692976, 'train/grad_norm_pre_clip': 33.5174446105957, 'timing/data': 0.0006052319658920169, 'timing/model': 1.9194316769717261, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 1.82}) +[2026-06-23 19:19:35,209][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.0026907119899988174, 'train/grad_norm_pre_clip': 4.313976764678955, 'timing/data': 0.0006056289421394467, 'timing/model': 1.9128193089272827, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 1.82}) +[2026-06-23 19:19:38,999][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.06224721297621727, 'train/grad_norm_pre_clip': 19.705860137939453, 'timing/data': 0.0006058840081095695, 'timing/model': 1.9223016889300197, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 1.82}) +[2026-06-23 19:19:42,881][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.12058016657829285, 'train/grad_norm_pre_clip': 40.47643280029297, 'timing/data': 0.0006008690688759089, 'timing/model': 1.9312251990195364, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 1.82}) +[2026-06-23 19:19:46,653][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.05475696176290512, 'train/grad_norm_pre_clip': 42.793304443359375, 'timing/data': 0.0006521729519590735, 'timing/model': 1.9143761419691145, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 1.82}) +[2026-06-23 19:19:50,429][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.08810590952634811, 'train/grad_norm_pre_clip': 36.963016510009766, 'timing/data': 0.0006953190313652158, 'timing/model': 1.9158729369519278, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 1.82}) +[2026-06-23 19:19:54,224][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.1117226853966713, 'train/grad_norm_pre_clip': 35.84208679199219, 'timing/data': 0.0006135939620435238, 'timing/model': 1.918763748020865, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 1.82}) +[2026-06-23 19:19:58,013][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.058821529150009155, 'train/grad_norm_pre_clip': 33.891475677490234, 'timing/data': 0.000626124907284975, 'timing/model': 1.923143697087653, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 1.82}) +[2026-06-23 19:20:01,801][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.011033247224986553, 'train/grad_norm_pre_clip': 21.293895721435547, 'timing/data': 0.0006311869947239757, 'timing/model': 1.930553985061124, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 1.82}) +[2026-06-23 19:20:05,605][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.04740940034389496, 'train/grad_norm_pre_clip': 19.430906295776367, 'timing/data': 0.0006527109071612358, 'timing/model': 1.92070247605443, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 1.82}) +[2026-06-23 19:20:09,397][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.030370160937309265, 'train/grad_norm_pre_clip': 43.4954948425293, 'timing/data': 0.0005989730125293136, 'timing/model': 1.9279562690062448, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 1.82}) +[2026-06-23 19:20:13,238][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.12995344400405884, 'train/grad_norm_pre_clip': 40.47456741333008, 'timing/data': 0.0006096439901739359, 'timing/model': 1.9706508279778063, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 1.83}) +[2026-06-23 19:20:17,029][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.0382595956325531, 'train/grad_norm_pre_clip': 22.84915542602539, 'timing/data': 0.0006768950261175632, 'timing/model': 1.9238475820748135, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 1.83}) +[2026-06-23 19:20:20,819][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.03872595727443695, 'train/grad_norm_pre_clip': 29.799339294433594, 'timing/data': 0.0009123790077865124, 'timing/model': 1.9249119879677892, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 1.83}) +[2026-06-23 19:20:24,608][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.045602381229400635, 'train/grad_norm_pre_clip': 28.186960220336914, 'timing/data': 0.0008041999535635114, 'timing/model': 1.9257305490318686, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 1.83}) +[2026-06-23 19:20:28,408][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.048083528876304626, 'train/grad_norm_pre_clip': 24.752222061157227, 'timing/data': 0.0006475640693679452, 'timing/model': 1.9289364520227537, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 1.83}) +[2026-06-23 19:20:32,204][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.029453152790665627, 'train/grad_norm_pre_clip': 40.85418701171875, 'timing/data': 0.0006167680257931352, 'timing/model': 1.9251182320294902, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 1.83}) +[2026-06-23 19:20:36,007][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.06759525835514069, 'train/grad_norm_pre_clip': 25.88495445251465, 'timing/data': 0.0006851139478385448, 'timing/model': 1.9386265251087025, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 1.83}) +[2026-06-23 19:20:39,810][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.020004959776997566, 'train/grad_norm_pre_clip': 17.48934555053711, 'timing/data': 0.0006016740808263421, 'timing/model': 1.9286959969904274, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 1.83}) +[2026-06-23 19:20:43,616][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.020939959213137627, 'train/grad_norm_pre_clip': 17.293140411376953, 'timing/data': 0.0006108249071985483, 'timing/model': 1.9266771040856838, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 1.83}) +[2026-06-23 19:20:47,525][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.05073932558298111, 'train/grad_norm_pre_clip': 31.319421768188477, 'timing/data': 0.0005940699484199286, 'timing/model': 2.0456366590224206, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 1.83}) +[2026-06-23 19:20:51,342][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.009825706481933594, 'train/grad_norm_pre_clip': 21.312700271606445, 'timing/data': 0.0005915310466662049, 'timing/model': 1.9379865869414061, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 1.83}) +[2026-06-23 19:20:55,142][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.007207405287772417, 'train/grad_norm_pre_clip': 18.779399871826172, 'timing/data': 0.0006571159465238452, 'timing/model': 1.9254025070695207, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 1.83}) +[2026-06-23 19:20:58,944][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.06945116817951202, 'train/grad_norm_pre_clip': 33.970279693603516, 'timing/data': 0.000768003985285759, 'timing/model': 1.926326405024156, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 1.83}) +[2026-06-23 19:21:02,732][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.02610381878912449, 'train/grad_norm_pre_clip': 26.845447540283203, 'timing/data': 0.0005983499577268958, 'timing/model': 1.9263218130217865, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 1.83}) +[2026-06-23 19:21:06,532][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.02277272567152977, 'train/grad_norm_pre_clip': 14.690324783325195, 'timing/data': 0.000793528975918889, 'timing/model': 1.92568026506342, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 1.83}) +[2026-06-23 19:21:10,340][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.11013851314783096, 'train/grad_norm_pre_clip': 23.422819137573242, 'timing/data': 0.0006085439817979932, 'timing/model': 1.9264013960491866, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 1.83}) +[2026-06-23 19:21:14,149][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.03462602570652962, 'train/grad_norm_pre_clip': 25.349952697753906, 'timing/data': 0.0005920400144532323, 'timing/model': 1.9373738690046594, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 1.83}) +[2026-06-23 19:21:17,949][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.08884551376104355, 'train/grad_norm_pre_clip': 30.87906837463379, 'timing/data': 0.0006169249536469579, 'timing/model': 1.9219529619440436, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 1.83}) +[2026-06-23 19:21:21,878][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.06387922167778015, 'train/grad_norm_pre_clip': 19.237625122070312, 'timing/data': 0.0008138290140777826, 'timing/model': 2.0581056179944426, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 1.83}) +[2026-06-23 19:21:25,697][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.0024736078921705484, 'train/grad_norm_pre_clip': 6.252985000610352, 'timing/data': 0.0005915469955652952, 'timing/model': 1.9341994519345462, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 1.83}) +[2026-06-23 19:21:29,502][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.07904404401779175, 'train/grad_norm_pre_clip': 61.970985412597656, 'timing/data': 0.0008001849055290222, 'timing/model': 1.9344581889454275, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 1.83}) +[2026-06-23 19:21:33,301][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.007470325566828251, 'train/grad_norm_pre_clip': 20.80331802368164, 'timing/data': 0.0007857410237193108, 'timing/model': 1.9184034400386736, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 1.83}) +[2026-06-23 19:21:37,088][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.008976420387625694, 'train/grad_norm_pre_clip': 24.2882137298584, 'timing/data': 0.0008298569591715932, 'timing/model': 1.917189772007987, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 1.83}) +[2026-06-23 19:21:40,903][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.017209751531481743, 'train/grad_norm_pre_clip': 23.477083206176758, 'timing/data': 0.0006958560552448034, 'timing/model': 1.9348406699718907, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 1.83}) +[2026-06-23 19:21:44,708][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.010838278569281101, 'train/grad_norm_pre_clip': 9.476999282836914, 'timing/data': 0.0006607420509681106, 'timing/model': 1.9275729230139405, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 1.83}) +[2026-06-23 19:21:48,503][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.026374151930212975, 'train/grad_norm_pre_clip': 19.58164405822754, 'timing/data': 0.0007061510113999248, 'timing/model': 1.9213930789846927, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 1.83}) +[2026-06-23 19:21:52,316][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.008956512436270714, 'train/grad_norm_pre_clip': 13.840503692626953, 'timing/data': 0.000778881018050015, 'timing/model': 1.9320735910441726, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 1.84}) +[2026-06-23 19:21:56,316][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.05166199430823326, 'train/grad_norm_pre_clip': 24.325801849365234, 'timing/data': 0.0006242060335353017, 'timing/model': 1.9345648429589346, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 1.84}) +[2026-06-23 19:22:00,122][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.03346797451376915, 'train/grad_norm_pre_clip': 17.830659866333008, 'timing/data': 0.0006093160482123494, 'timing/model': 1.9372268280712888, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 1.84}) +[2026-06-23 19:22:03,924][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.02920643799006939, 'train/grad_norm_pre_clip': 18.843677520751953, 'timing/data': 0.0007013480644673109, 'timing/model': 1.9208751330152154, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 1.84}) +[2026-06-23 19:22:07,713][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.10677168518304825, 'train/grad_norm_pre_clip': 32.20466613769531, 'timing/data': 0.000623266096226871, 'timing/model': 1.9182034709956497, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 1.84}) +[2026-06-23 19:22:11,501][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.0886794850230217, 'train/grad_norm_pre_clip': 32.45753479003906, 'timing/data': 0.0007627509767189622, 'timing/model': 1.9200322339311242, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 1.84}) +[2026-06-23 19:22:15,283][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.04367313161492348, 'train/grad_norm_pre_clip': 27.798032760620117, 'timing/data': 0.0006876080296933651, 'timing/model': 1.9197466489858925, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 1.84}) +[2026-06-23 19:22:19,078][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.06142401322722435, 'train/grad_norm_pre_clip': 27.015348434448242, 'timing/data': 0.0007940349169075489, 'timing/model': 1.9312984980642796, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 1.84}) +[2026-06-23 19:29:37,656][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4750 +[2026-06-23 19:29:37,657][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.0183589905500412, 'train/grad_norm_pre_clip': 26.158952713012695, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 265.9800000000083, 'rl_games_eval/mid_train/mean_length': 2162.4, 'rl_games_eval/mid_train/std_reward': 151.80635559818955, 'rl_games_eval/mid_train/std_length': 1222.032830982867, 'rl_games_eval/mid_train/macro_mean_reward': 265.9800000000083, 'rl_games_eval/mid_train/macro_mean_length': 2162.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 265.9800000000083, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 2162.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 151.80635559818955, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1222.032830982867, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 265.9800000000083, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006055081030353904, 'timing/model': 1.9206162439659238, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 1.84}) +[2026-06-23 19:29:41,367][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.16742302477359772, 'train/grad_norm_pre_clip': 34.77958679199219, 'timing/data': 0.0005987299373373389, 'timing/model': 1.8770570328924805, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 1.84}) +[2026-06-23 19:29:45,100][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.11041299998760223, 'train/grad_norm_pre_clip': 40.341487884521484, 'timing/data': 0.0009646760299801826, 'timing/model': 1.9010377250378951, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 1.84}) +[2026-06-23 19:29:48,838][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.0523359514772892, 'train/grad_norm_pre_clip': 30.426759719848633, 'timing/data': 0.0006605829112231731, 'timing/model': 1.8988322570221499, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 1.84}) +[2026-06-23 19:29:52,561][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.016073787584900856, 'train/grad_norm_pre_clip': 19.459651947021484, 'timing/data': 0.00061461899895221, 'timing/model': 1.88777625199873, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 1.84}) +[2026-06-23 19:29:56,298][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.05976276472210884, 'train/grad_norm_pre_clip': 40.36797332763672, 'timing/data': 0.0007970510050654411, 'timing/model': 1.8946438489947468, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 1.84}) +[2026-06-23 19:30:00,055][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.06315328925848007, 'train/grad_norm_pre_clip': 37.91472625732422, 'timing/data': 0.0006638310151174664, 'timing/model': 1.9043232679832727, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 1.84}) +[2026-06-23 19:30:03,813][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.015216170810163021, 'train/grad_norm_pre_clip': 23.690692901611328, 'timing/data': 0.0006221570074558258, 'timing/model': 1.9037512199720368, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 1.84}) +[2026-06-23 19:30:07,558][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.057376645505428314, 'train/grad_norm_pre_clip': 32.38094711303711, 'timing/data': 0.0007392959669232368, 'timing/model': 1.8996916009346023, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 1.84}) +[2026-06-23 19:30:11,315][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.1084059625864029, 'train/grad_norm_pre_clip': 46.616966247558594, 'timing/data': 0.000581186031922698, 'timing/model': 1.9105841519776732, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 1.84}) +[2026-06-23 19:30:15,075][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.03020251914858818, 'train/grad_norm_pre_clip': 31.305831909179688, 'timing/data': 0.0007283800514414907, 'timing/model': 1.9087218660861254, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 1.84}) +[2026-06-23 19:30:18,845][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.041692767292261124, 'train/grad_norm_pre_clip': 23.537824630737305, 'timing/data': 0.0006294139893725514, 'timing/model': 1.9207186249550432, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 1.84}) +[2026-06-23 19:30:22,635][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.10614810138940811, 'train/grad_norm_pre_clip': 32.942352294921875, 'timing/data': 0.0006154450820758939, 'timing/model': 1.9221315729664639, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 1.84}) +[2026-06-23 19:30:26,407][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.050854574888944626, 'train/grad_norm_pre_clip': 35.02622604370117, 'timing/data': 0.0007681109709665179, 'timing/model': 1.919579559005797, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 1.84}) +[2026-06-23 19:30:30,184][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.037101421505212784, 'train/grad_norm_pre_clip': 35.92155838012695, 'timing/data': 0.000760308001190424, 'timing/model': 1.9222759120166302, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 1.84}) +[2026-06-23 19:30:33,958][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.054226528853178024, 'train/grad_norm_pre_clip': 27.688966751098633, 'timing/data': 0.00084270304068923, 'timing/model': 1.9246222929796204, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 1.84}) +[2026-06-23 19:30:37,933][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.057426538318395615, 'train/grad_norm_pre_clip': 19.550079345703125, 'timing/data': 0.0007779470179229975, 'timing/model': 1.9139092169934884, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 1.84}) +[2026-06-23 19:30:41,741][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.04638214036822319, 'train/grad_norm_pre_clip': 51.457706451416016, 'timing/data': 0.0008038550149649382, 'timing/model': 1.9212296219775453, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 1.84}) +[2026-06-23 19:30:45,528][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.04721975326538086, 'train/grad_norm_pre_clip': 41.481441497802734, 'timing/data': 0.0005920169642195106, 'timing/model': 1.9158702780259773, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 1.85}) +[2026-06-23 19:30:49,309][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.04659122973680496, 'train/grad_norm_pre_clip': 23.306446075439453, 'timing/data': 0.0006120760226622224, 'timing/model': 1.915330068091862, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 1.85}) +[2026-06-23 19:30:53,090][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.016027027741074562, 'train/grad_norm_pre_clip': 9.794029235839844, 'timing/data': 0.0006008159834891558, 'timing/model': 1.9233779920032248, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 1.85}) +[2026-06-23 19:30:56,868][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.006577656138688326, 'train/grad_norm_pre_clip': 34.611568450927734, 'timing/data': 0.0005928069585934281, 'timing/model': 1.9195827370276675, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 1.85}) +[2026-06-23 19:31:00,656][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.08855900168418884, 'train/grad_norm_pre_clip': 40.29241943359375, 'timing/data': 0.000598522019572556, 'timing/model': 1.922564159031026, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 1.85}) +[2026-06-23 19:31:04,442][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.02517838031053543, 'train/grad_norm_pre_clip': 29.71672248840332, 'timing/data': 0.0006231430452316999, 'timing/model': 1.9222627650015056, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 1.85}) +[2026-06-23 19:31:08,383][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.008117837831377983, 'train/grad_norm_pre_clip': 15.496784210205078, 'timing/data': 0.000628105946816504, 'timing/model': 2.0864016299601644, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 1.85}) +[2026-06-23 19:31:12,164][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.021871861070394516, 'train/grad_norm_pre_clip': 22.976764678955078, 'timing/data': 0.0006275050109252334, 'timing/model': 1.9106682959245518, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 1.85}) +[2026-06-23 19:31:15,943][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.03568895906209946, 'train/grad_norm_pre_clip': 21.527748107910156, 'timing/data': 0.000729234074242413, 'timing/model': 1.9193846519337967, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 1.85}) +[2026-06-23 19:31:19,727][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.006337579805403948, 'train/grad_norm_pre_clip': 15.620331764221191, 'timing/data': 0.0006062100874260068, 'timing/model': 1.9114369319286197, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 1.85}) +[2026-06-23 19:31:23,499][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.1246953159570694, 'train/grad_norm_pre_clip': 63.74112319946289, 'timing/data': 0.0008626899216324091, 'timing/model': 1.914918459020555, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 1.85}) +[2026-06-23 19:31:27,309][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.021197520196437836, 'train/grad_norm_pre_clip': 16.777568817138672, 'timing/data': 0.0008590179495513439, 'timing/model': 1.9357703829882666, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 1.85}) +[2026-06-23 19:31:31,101][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.024066630750894547, 'train/grad_norm_pre_clip': 37.59443283081055, 'timing/data': 0.0006033920217305422, 'timing/model': 1.9220809229882434, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 1.85}) +[2026-06-23 19:31:34,883][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.04486807808279991, 'train/grad_norm_pre_clip': 38.854026794433594, 'timing/data': 0.0005947819445282221, 'timing/model': 1.9216190190054476, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 1.85}) +[2026-06-23 19:31:38,678][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.09091684967279434, 'train/grad_norm_pre_clip': 37.965049743652344, 'timing/data': 0.0005856329808011651, 'timing/model': 1.9284803440095857, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 1.85}) +[2026-06-23 19:31:42,602][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.016371626406908035, 'train/grad_norm_pre_clip': 13.165203094482422, 'timing/data': 0.0005959890550002456, 'timing/model': 1.9263534429483116, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 1.85}) +[2026-06-23 19:31:46,405][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.04310179129242897, 'train/grad_norm_pre_clip': 35.822303771972656, 'timing/data': 0.0005947219906374812, 'timing/model': 1.936470699030906, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 1.85}) +[2026-06-23 19:31:50,197][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.048292260617017746, 'train/grad_norm_pre_clip': 51.96348190307617, 'timing/data': 0.0006535518914461136, 'timing/model': 1.919366622925736, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 1.85}) +[2026-06-23 19:31:53,992][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.0760505348443985, 'train/grad_norm_pre_clip': 32.99747848510742, 'timing/data': 0.0007210489129647613, 'timing/model': 1.9291076919762418, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 1.85}) +[2026-06-23 19:31:57,786][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.03521987795829773, 'train/grad_norm_pre_clip': 30.603971481323242, 'timing/data': 0.0006166569655761123, 'timing/model': 1.9218524659518152, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 1.85}) +[2026-06-23 19:32:01,569][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.012674140743911266, 'train/grad_norm_pre_clip': 20.48353385925293, 'timing/data': 0.0006675680633634329, 'timing/model': 1.9160241949139163, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 1.85}) +[2026-06-23 19:32:05,359][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.06016532704234123, 'train/grad_norm_pre_clip': 24.188337326049805, 'timing/data': 0.0008981159189715981, 'timing/model': 1.9156680960441008, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 1.85}) +[2026-06-23 19:32:09,148][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.07070335745811462, 'train/grad_norm_pre_clip': 34.94229507446289, 'timing/data': 0.0005980900023132563, 'timing/model': 1.925419259001501, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 1.85}) +[2026-06-23 19:32:13,106][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.004567768890410662, 'train/grad_norm_pre_clip': 8.084468841552734, 'timing/data': 0.000583347980864346, 'timing/model': 2.0938874890562147, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 1.85}) +[2026-06-23 19:32:16,916][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.03797877952456474, 'train/grad_norm_pre_clip': 14.131158828735352, 'timing/data': 0.0006072020623832941, 'timing/model': 1.933294471935369, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 1.85}) +[2026-06-23 19:32:20,716][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.005313954781740904, 'train/grad_norm_pre_clip': 7.617897987365723, 'timing/data': 0.0005890130996704102, 'timing/model': 1.9309118980308995, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 1.85}) +[2026-06-23 19:32:24,501][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.030945071950554848, 'train/grad_norm_pre_clip': 37.54344177246094, 'timing/data': 0.000739420996978879, 'timing/model': 1.9169643949717283, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 1.86}) +[2026-06-23 19:32:28,272][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.011269884184002876, 'train/grad_norm_pre_clip': 32.99010467529297, 'timing/data': 0.0006052269600331783, 'timing/model': 1.9111212890129536, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 1.86}) +[2026-06-23 19:32:32,055][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.025687068700790405, 'train/grad_norm_pre_clip': 11.405220985412598, 'timing/data': 0.0006078120786696672, 'timing/model': 1.9246389789041132, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 1.86}) +[2026-06-23 19:32:35,840][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.04488986358046532, 'train/grad_norm_pre_clip': 35.359134674072266, 'timing/data': 0.0006008779164403677, 'timing/model': 1.9219197010388598, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 1.86}) +[2026-06-23 19:32:39,635][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.04789208993315697, 'train/grad_norm_pre_clip': 23.800127029418945, 'timing/data': 0.0005929149920120835, 'timing/model': 1.9242988580372185, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 1.86}) +[2026-06-23 19:32:43,432][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.006833285558968782, 'train/grad_norm_pre_clip': 52.56283950805664, 'timing/data': 0.0005925230216234922, 'timing/model': 1.9240288980072364, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 1.86}) +[2026-06-23 19:32:47,387][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4800 +[2026-06-23 19:34:53,804][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4800 +[2026-06-23 19:34:53,805][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4800 +[2026-06-23 19:36:16,301][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4800 +[2026-06-23 19:36:16,302][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.051465436816215515, 'train/grad_norm_pre_clip': 31.902809143066406, 'eval/action_loss/seconds': 126.41464157600421, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03044542856514454, 'eval/latency_2/loss': 0.03044542856514454, 'eval/flappy/loss': 0.03044542856514454, 'eval/flappy/latency_2/loss': 0.03044542856514454, 'eval/flap_precision': 0.9681571815718157, 'eval/flap_recall': 0.9642375168690959, 'eval/flap_f1': 0.9661933739012846, 'eval/flap_rate_model': 0.205, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9661933739012845, 'eval/latency_2/flap_precision': 0.9681571815718157, 'eval/latency_2/flap_recall': 0.9642375168690959, 'eval/latency_2/flap_f1': 0.9661933739012846, 'eval/latency_2/flap_rate_model': 0.205, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9661933739012845, 'eval/action_classification/data_wait_seconds': 1.6195404529571533, 'eval/action_classification/predict_seconds': 80.69608306884766, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006137300515547395, 'timing/model': 1.9224341890076175, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 1.86}) +[2026-06-23 19:36:20,077][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.031157860532402992, 'train/grad_norm_pre_clip': 67.44522857666016, 'timing/data': 0.000716226059012115, 'timing/model': 1.9067250180523843, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 1.86}) +[2026-06-23 19:36:23,848][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.06286702305078506, 'train/grad_norm_pre_clip': 35.15935516357422, 'timing/data': 0.0006197720067575574, 'timing/model': 1.9083253989228979, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 1.86}) +[2026-06-23 19:36:27,624][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.14354754984378815, 'train/grad_norm_pre_clip': 54.76026153564453, 'timing/data': 0.0007080279756337404, 'timing/model': 1.9161682379199192, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 1.86}) +[2026-06-23 19:36:31,413][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.05469755455851555, 'train/grad_norm_pre_clip': 28.691816329956055, 'timing/data': 0.0006480059819296002, 'timing/model': 1.9271134920418262, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 1.86}) +[2026-06-23 19:36:35,236][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.03034939058125019, 'train/grad_norm_pre_clip': 25.938697814941406, 'timing/data': 0.0006105100037530065, 'timing/model': 1.9212035529781133, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 1.86}) +[2026-06-23 19:36:39,030][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.0770673081278801, 'train/grad_norm_pre_clip': 46.49649429321289, 'timing/data': 0.0008308669785037637, 'timing/model': 1.9309263419127092, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 1.86}) +[2026-06-23 19:36:42,850][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.024618016555905342, 'train/grad_norm_pre_clip': 14.125051498413086, 'timing/data': 0.0006127409869804978, 'timing/model': 1.925725303008221, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 1.86}) +[2026-06-23 19:36:46,641][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.014266488142311573, 'train/grad_norm_pre_clip': 17.265329360961914, 'timing/data': 0.00078048394061625, 'timing/model': 1.9176194429164752, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 1.86}) +[2026-06-23 19:36:50,423][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.03621598333120346, 'train/grad_norm_pre_clip': 32.01539993286133, 'timing/data': 0.0006051640957593918, 'timing/model': 1.9107687990181148, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 1.86}) +[2026-06-23 19:36:54,215][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.13448861241340637, 'train/grad_norm_pre_clip': 66.28947448730469, 'timing/data': 0.0006100580794736743, 'timing/model': 1.92270347091835, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 1.86}) +[2026-06-23 19:36:58,013][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.008355120196938515, 'train/grad_norm_pre_clip': 21.033334732055664, 'timing/data': 0.0006258100038394332, 'timing/model': 1.9273309839190915, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 1.86}) +[2026-06-23 19:37:01,794][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.011922648176550865, 'train/grad_norm_pre_clip': 25.767745971679688, 'timing/data': 0.0006421010475605726, 'timing/model': 1.9217952110338956, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 1.86}) +[2026-06-23 19:37:05,577][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.028159024193882942, 'train/grad_norm_pre_clip': 53.41862869262695, 'timing/data': 0.0008250169921666384, 'timing/model': 1.923364864080213, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 1.86}) +[2026-06-23 19:37:09,367][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.05904041603207588, 'train/grad_norm_pre_clip': 33.761783599853516, 'timing/data': 0.000589621951803565, 'timing/model': 1.9315094959456474, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 1.86}) +[2026-06-23 19:37:13,204][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.018420100212097168, 'train/grad_norm_pre_clip': 12.404163360595703, 'timing/data': 0.0008422819664701819, 'timing/model': 1.9339889860711992, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 1.86}) +[2026-06-23 19:37:17,005][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.11797305941581726, 'train/grad_norm_pre_clip': 26.476343154907227, 'timing/data': 0.000906718079932034, 'timing/model': 1.9244330510264263, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 1.86}) +[2026-06-23 19:37:20,783][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.08689778298139572, 'train/grad_norm_pre_clip': 34.45452880859375, 'timing/data': 0.000802651047706604, 'timing/model': 1.919029978918843, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 1.86}) +[2026-06-23 19:37:24,577][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.05859379842877388, 'train/grad_norm_pre_clip': 104.79096221923828, 'timing/data': 0.0006508940132334828, 'timing/model': 1.9296941299689934, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 1.86}) +[2026-06-23 19:37:28,368][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.05611253157258034, 'train/grad_norm_pre_clip': 23.400053024291992, 'timing/data': 0.0006020929431542754, 'timing/model': 1.924631552072242, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 1.87}) +[2026-06-23 19:37:32,161][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.09440510720014572, 'train/grad_norm_pre_clip': 47.41685104370117, 'timing/data': 0.0006033600075170398, 'timing/model': 1.922080855933018, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 1.87}) +[2026-06-23 19:37:35,960][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.14153529703617096, 'train/grad_norm_pre_clip': 52.99175262451172, 'timing/data': 0.0005994579987600446, 'timing/model': 1.9220155000220984, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 1.87}) +[2026-06-23 19:37:39,777][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.016615737229585648, 'train/grad_norm_pre_clip': 26.044981002807617, 'timing/data': 0.0005954579683020711, 'timing/model': 1.9321134750498459, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 1.87}) +[2026-06-23 19:37:43,576][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.03851654753088951, 'train/grad_norm_pre_clip': 26.25875473022461, 'timing/data': 0.0005876970244571567, 'timing/model': 1.9270880579715595, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 1.87}) +[2026-06-23 19:37:47,384][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.017278190702199936, 'train/grad_norm_pre_clip': 33.10529327392578, 'timing/data': 0.0005825909320265055, 'timing/model': 1.9373259990243241, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 1.87}) +[2026-06-23 19:37:51,165][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.10656781494617462, 'train/grad_norm_pre_clip': 36.027095794677734, 'timing/data': 0.0007735799299553037, 'timing/model': 1.9165612360229716, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 1.87}) +[2026-06-23 19:37:54,973][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.04626433178782463, 'train/grad_norm_pre_clip': 31.592628479003906, 'timing/data': 0.0005900620017200708, 'timing/model': 1.9305555180180818, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 1.87}) +[2026-06-23 19:37:58,773][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.03519147261977196, 'train/grad_norm_pre_clip': 27.177518844604492, 'timing/data': 0.0005914559587836266, 'timing/model': 1.9240038519492373, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 1.87}) +[2026-06-23 19:38:02,577][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.018984438851475716, 'train/grad_norm_pre_clip': 20.617916107177734, 'timing/data': 0.0007180209504440427, 'timing/model': 1.9351904640207067, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 1.87}) +[2026-06-23 19:38:06,381][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.15149666368961334, 'train/grad_norm_pre_clip': 31.049880981445312, 'timing/data': 0.0006165030645206571, 'timing/model': 1.9288633649703115, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 1.87}) +[2026-06-23 19:38:10,193][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.013088823296129704, 'train/grad_norm_pre_clip': 29.07101058959961, 'timing/data': 0.0008133109658956528, 'timing/model': 1.9302332369843498, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 1.87}) +[2026-06-23 19:38:14,172][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.053951021283864975, 'train/grad_norm_pre_clip': 21.817138671875, 'timing/data': 0.0006029140204191208, 'timing/model': 1.9277334050275385, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 1.87}) +[2026-06-23 19:38:17,989][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.04113331437110901, 'train/grad_norm_pre_clip': 47.22412109375, 'timing/data': 0.0006125379586592317, 'timing/model': 1.9419539219234139, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 1.87}) +[2026-06-23 19:38:21,787][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.03339797258377075, 'train/grad_norm_pre_clip': 35.812984466552734, 'timing/data': 0.0005896440707147121, 'timing/model': 1.9227645450737327, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 1.87}) +[2026-06-23 19:38:25,569][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.11559876054525375, 'train/grad_norm_pre_clip': 63.37004852294922, 'timing/data': 0.0006740790558978915, 'timing/model': 1.9182202549418435, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 1.87}) +[2026-06-23 19:38:29,365][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.006973977666348219, 'train/grad_norm_pre_clip': 8.071452140808105, 'timing/data': 0.0005992900114506483, 'timing/model': 1.927520632976666, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 1.87}) +[2026-06-23 19:38:33,186][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.14654825627803802, 'train/grad_norm_pre_clip': 60.266841888427734, 'timing/data': 0.0006547829834744334, 'timing/model': 1.9489154459442943, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 1.87}) +[2026-06-23 19:38:36,980][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.04545065015554428, 'train/grad_norm_pre_clip': 60.806983947753906, 'timing/data': 0.0006044520996510983, 'timing/model': 1.920717203989625, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 1.87}) +[2026-06-23 19:38:40,775][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.04294855520129204, 'train/grad_norm_pre_clip': 51.92924880981445, 'timing/data': 0.0006416819524019957, 'timing/model': 1.9201162389945239, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 1.87}) +[2026-06-23 19:38:44,635][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.03151242434978485, 'train/grad_norm_pre_clip': 18.638723373413086, 'timing/data': 0.0006866860203444958, 'timing/model': 1.9859106589574367, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 1.87}) +[2026-06-23 19:38:48,435][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.08660661429166794, 'train/grad_norm_pre_clip': 27.937664031982422, 'timing/data': 0.0006668269634246826, 'timing/model': 1.9383097250247374, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 1.87}) +[2026-06-23 19:38:52,223][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.03950507938861847, 'train/grad_norm_pre_clip': 31.083242416381836, 'timing/data': 0.0006220069481059909, 'timing/model': 1.919508249964565, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 1.87}) +[2026-06-23 19:38:56,037][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.07479728013277054, 'train/grad_norm_pre_clip': 51.80823516845703, 'timing/data': 0.0006314439233392477, 'timing/model': 1.9313529320061207, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 1.87}) +[2026-06-23 19:38:59,835][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.016020197421312332, 'train/grad_norm_pre_clip': 12.340036392211914, 'timing/data': 0.000731242005713284, 'timing/model': 1.935332567896694, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 1.87}) +[2026-06-23 19:39:03,622][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.011325555853545666, 'train/grad_norm_pre_clip': 35.60428237915039, 'timing/data': 0.0007360070012509823, 'timing/model': 1.9164534390438348, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 1.87}) +[2026-06-23 19:39:07,429][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.01698668673634529, 'train/grad_norm_pre_clip': 61.81950759887695, 'timing/data': 0.0006309900199994445, 'timing/model': 1.9255113920662552, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 1.88}) +[2026-06-23 19:39:11,213][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.019319674000144005, 'train/grad_norm_pre_clip': 22.537748336791992, 'timing/data': 0.0005839579971507192, 'timing/model': 1.9160368810407817, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 1.88}) +[2026-06-23 19:39:15,018][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.07610916346311569, 'train/grad_norm_pre_clip': 51.367183685302734, 'timing/data': 0.0006248889258131385, 'timing/model': 1.9309060239465907, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 1.88}) +[2026-06-23 19:39:18,880][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.020045476034283638, 'train/grad_norm_pre_clip': 31.28078842163086, 'timing/data': 0.0005931230261921883, 'timing/model': 1.9282433800399303, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 1.88}) +[2026-06-23 19:39:22,678][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.018453119322657585, 'train/grad_norm_pre_clip': 34.45241928100586, 'timing/data': 0.0006096420111134648, 'timing/model': 1.9220443080412224, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 1.88}) +[2026-06-23 19:39:26,496][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.06333937495946884, 'train/grad_norm_pre_clip': 37.862220764160156, 'timing/data': 0.0005970270140096545, 'timing/model': 1.925853792927228, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 1.88}) +[2026-06-23 19:39:30,293][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.06269382685422897, 'train/grad_norm_pre_clip': 61.28264617919922, 'timing/data': 0.0006005340255796909, 'timing/model': 1.929760577972047, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 1.88}) +[2026-06-23 19:39:34,100][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.017355836927890778, 'train/grad_norm_pre_clip': 24.980045318603516, 'timing/data': 0.0006110469112172723, 'timing/model': 1.9373221599962562, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 1.88}) +[2026-06-23 19:39:37,901][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.021925058215856552, 'train/grad_norm_pre_clip': 29.840473175048828, 'timing/data': 0.0006167279789224267, 'timing/model': 1.9264768989523873, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 1.88}) +[2026-06-23 19:39:41,708][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.04780205339193344, 'train/grad_norm_pre_clip': 44.955596923828125, 'timing/data': 0.0006901280721649528, 'timing/model': 1.9225983900250867, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 1.88}) +[2026-06-23 19:39:45,503][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.062408559024333954, 'train/grad_norm_pre_clip': 30.310529708862305, 'timing/data': 0.0007672540377825499, 'timing/model': 1.922610527020879, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 1.88}) +[2026-06-23 19:39:49,323][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.016195954754948616, 'train/grad_norm_pre_clip': 12.947531700134277, 'timing/data': 0.0007203600835055113, 'timing/model': 1.9454934580717236, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 1.88}) +[2026-06-23 19:39:53,131][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.03314247727394104, 'train/grad_norm_pre_clip': 33.315677642822266, 'timing/data': 0.0007127650314942002, 'timing/model': 1.9288767320103943, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 1.88}) +[2026-06-23 19:39:56,937][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.018163656815886497, 'train/grad_norm_pre_clip': 31.1710262298584, 'timing/data': 0.0006739149102941155, 'timing/model': 1.9225498309824616, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 1.88}) +[2026-06-23 19:40:00,738][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.025134216994047165, 'train/grad_norm_pre_clip': 29.29444122314453, 'timing/data': 0.0006698539946228266, 'timing/model': 1.9354205960407853, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 1.88}) +[2026-06-23 19:40:04,547][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.028202315792441368, 'train/grad_norm_pre_clip': 13.664331436157227, 'timing/data': 0.0005971849896013737, 'timing/model': 1.9367307830834761, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 1.88}) +[2026-06-23 19:40:08,361][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.10794349759817123, 'train/grad_norm_pre_clip': 44.300201416015625, 'timing/data': 0.0007304209284484386, 'timing/model': 1.9332353320205584, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 1.88}) +[2026-06-23 19:40:12,158][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.04836736619472504, 'train/grad_norm_pre_clip': 45.10577392578125, 'timing/data': 0.0006115720607340336, 'timing/model': 1.92583248403389, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 1.88}) +[2026-06-23 19:40:15,976][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.05371814966201782, 'train/grad_norm_pre_clip': 160.3138427734375, 'timing/data': 0.0005914829671382904, 'timing/model': 1.9344260749639943, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 1.88}) +[2026-06-23 19:40:19,773][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.039547983556985855, 'train/grad_norm_pre_clip': 25.569120407104492, 'timing/data': 0.0007082440424710512, 'timing/model': 1.923684854991734, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 1.88}) +[2026-06-23 19:40:23,594][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.0005002642283216119, 'train/grad_norm_pre_clip': 4.159768581390381, 'timing/data': 0.0006101160543039441, 'timing/model': 1.9328583430033177, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 1.88}) +[2026-06-23 19:40:27,413][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.01274973712861538, 'train/grad_norm_pre_clip': 16.918439865112305, 'timing/data': 0.0006572820711880922, 'timing/model': 1.9346483730478212, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 1.88}) +[2026-06-23 19:40:31,224][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.017440935596823692, 'train/grad_norm_pre_clip': 19.85842514038086, 'timing/data': 0.0006267090793699026, 'timing/model': 1.9320010900264606, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 1.88}) +[2026-06-23 19:40:35,017][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.06715066730976105, 'train/grad_norm_pre_clip': 36.401424407958984, 'timing/data': 0.0009856870165094733, 'timing/model': 1.921825592056848, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 1.88}) +[2026-06-23 19:40:38,813][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.012022691778838634, 'train/grad_norm_pre_clip': 27.16316795349121, 'timing/data': 0.0006816439563408494, 'timing/model': 1.9307628530077636, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 1.88}) +[2026-06-23 19:40:42,661][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.007990660145878792, 'train/grad_norm_pre_clip': 23.59035873413086, 'timing/data': 0.0006517969304695725, 'timing/model': 1.935478194965981, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 1.88}) +[2026-06-23 19:40:46,463][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.08418219536542892, 'train/grad_norm_pre_clip': 19.588581085205078, 'timing/data': 0.0007935630856081843, 'timing/model': 1.9308744769077748, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 1.89}) +[2026-06-23 19:40:50,269][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.04254600778222084, 'train/grad_norm_pre_clip': 20.845745086669922, 'timing/data': 0.0008701960323378444, 'timing/model': 1.925859358976595, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 1.89}) +[2026-06-23 19:40:54,065][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.015778953209519386, 'train/grad_norm_pre_clip': 33.77676010131836, 'timing/data': 0.0006038779392838478, 'timing/model': 1.9331592430826277, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 1.89}) +[2026-06-23 19:40:57,876][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.01073793787509203, 'train/grad_norm_pre_clip': 11.108243942260742, 'timing/data': 0.0006482959724962711, 'timing/model': 1.9304441619897261, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 1.89}) +[2026-06-23 19:41:01,669][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.06471285223960876, 'train/grad_norm_pre_clip': 54.98847579956055, 'timing/data': 0.0006336530204862356, 'timing/model': 1.9220453810412437, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 1.89}) +[2026-06-23 19:41:05,474][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.022793564945459366, 'train/grad_norm_pre_clip': 27.20822525024414, 'timing/data': 0.0008237180300056934, 'timing/model': 1.9405958520947024, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 1.89}) +[2026-06-23 19:41:09,293][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.016786711290478706, 'train/grad_norm_pre_clip': 9.370220184326172, 'timing/data': 0.0006269080331549048, 'timing/model': 1.9333533520111814, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 1.89}) +[2026-06-23 19:41:13,094][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.018022455275058746, 'train/grad_norm_pre_clip': 19.999420166015625, 'timing/data': 0.00059804308693856, 'timing/model': 1.9245422809617594, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 1.89}) +[2026-06-23 19:41:16,889][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.026149185374379158, 'train/grad_norm_pre_clip': 40.26769256591797, 'timing/data': 0.0007204280700534582, 'timing/model': 1.9202071010367945, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 1.89}) +[2026-06-23 19:41:20,691][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.04151947423815727, 'train/grad_norm_pre_clip': 33.798492431640625, 'timing/data': 0.0006948820082470775, 'timing/model': 1.9298620329936966, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 1.89}) +[2026-06-23 19:41:24,500][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.020763961598277092, 'train/grad_norm_pre_clip': 15.645423889160156, 'timing/data': 0.0006153719732537866, 'timing/model': 1.9374593120301142, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 1.89}) +[2026-06-23 19:41:28,285][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.013586272485554218, 'train/grad_norm_pre_clip': 23.732824325561523, 'timing/data': 0.0009069580119103193, 'timing/model': 1.9164970229612663, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 1.89}) +[2026-06-23 19:41:32,075][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.02399129793047905, 'train/grad_norm_pre_clip': 29.11502456665039, 'timing/data': 0.0007554240291938186, 'timing/model': 1.9190752080176026, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 1.89}) +[2026-06-23 19:41:35,893][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.042168933898210526, 'train/grad_norm_pre_clip': 35.35944747924805, 'timing/data': 0.0006087080109864473, 'timing/model': 1.9330727340420708, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 1.89}) +[2026-06-23 19:41:39,691][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.10601545870304108, 'train/grad_norm_pre_clip': 41.450523376464844, 'timing/data': 0.0007586119463667274, 'timing/model': 1.932632116950117, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 1.89}) +[2026-06-23 19:41:43,495][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.05559784546494484, 'train/grad_norm_pre_clip': 29.269123077392578, 'timing/data': 0.0006225690012797713, 'timing/model': 1.9257753300480545, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 1.89}) +[2026-06-23 19:41:47,321][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.03976748511195183, 'train/grad_norm_pre_clip': 24.18449592590332, 'timing/data': 0.0006036870181560516, 'timing/model': 1.948906845995225, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 1.89}) +[2026-06-23 19:41:51,125][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.05549623817205429, 'train/grad_norm_pre_clip': 107.12452697753906, 'timing/data': 0.0005942799616605043, 'timing/model': 1.9322142320452258, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 1.89}) +[2026-06-23 19:41:54,948][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.09027591347694397, 'train/grad_norm_pre_clip': 28.079660415649414, 'timing/data': 0.0006096179131418467, 'timing/model': 1.9379017789615318, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 1.89}) +[2026-06-23 19:41:58,766][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.0450318306684494, 'train/grad_norm_pre_clip': 64.0479736328125, 'timing/data': 0.0006357741076499224, 'timing/model': 1.938886313000694, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 1.89}) +[2026-06-23 19:42:02,577][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.010325176641345024, 'train/grad_norm_pre_clip': 11.224446296691895, 'timing/data': 0.0005884249694645405, 'timing/model': 1.9417800930095837, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 1.89}) +[2026-06-23 19:42:06,389][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.11727321147918701, 'train/grad_norm_pre_clip': 28.02410888671875, 'timing/data': 0.0006880280561745167, 'timing/model': 1.9283628900302574, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 1.89}) +[2026-06-23 19:42:10,192][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.02987469546496868, 'train/grad_norm_pre_clip': 39.647850036621094, 'timing/data': 0.0006809590850025415, 'timing/model': 1.9293612890178338, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 1.89}) +[2026-06-23 19:42:13,992][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.047096893191337585, 'train/grad_norm_pre_clip': 38.06482696533203, 'timing/data': 0.0006890619406476617, 'timing/model': 1.9263520700624213, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 1.89}) +[2026-06-23 19:42:17,779][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.06242121011018753, 'train/grad_norm_pre_clip': 41.48395538330078, 'timing/data': 0.0006593840662389994, 'timing/model': 1.9214766999939457, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 1.89}) +[2026-06-23 19:42:21,583][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.21169714629650116, 'train/grad_norm_pre_clip': 39.33698654174805, 'timing/data': 0.0005958460969850421, 'timing/model': 1.934007708914578, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 1.89}) +[2026-06-23 19:42:25,387][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.022921079769730568, 'train/grad_norm_pre_clip': 27.8882999420166, 'timing/data': 0.0006204360397532582, 'timing/model': 1.9368363620014861, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 1.9}) +[2026-06-23 19:42:29,189][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.03395244479179382, 'train/grad_norm_pre_clip': 13.36860179901123, 'timing/data': 0.0006093140691518784, 'timing/model': 1.924052131944336, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 1.9}) +[2026-06-23 19:42:33,007][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.0872395783662796, 'train/grad_norm_pre_clip': 20.437667846679688, 'timing/data': 0.0009014579700306058, 'timing/model': 1.9373303700704128, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 1.9}) +[2026-06-23 19:42:36,809][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4900 +[2026-06-23 19:44:45,051][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4900 +[2026-06-23 19:44:45,053][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4900 +[2026-06-23 19:46:08,047][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4900 +[2026-06-23 19:46:08,048][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.045244768261909485, 'train/grad_norm_pre_clip': 28.034305572509766, 'eval/action_loss/seconds': 128.23892732092645, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.0322595089673996, 'eval/latency_2/loss': 0.0322595089673996, 'eval/flappy/loss': 0.0322595089673996, 'eval/flappy/latency_2/loss': 0.0322595089673996, 'eval/flap_precision': 0.962, 'eval/flap_recall': 0.9736842105263158, 'eval/flap_f1': 0.9678068410462777, 'eval/flap_rate_model': 0.20833333333333334, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9678068410462777, 'eval/latency_2/flap_precision': 0.962, 'eval/latency_2/flap_recall': 0.9736842105263158, 'eval/latency_2/flap_f1': 0.9678068410462777, 'eval/latency_2/flap_rate_model': 0.20833333333333334, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9678068410462777, 'eval/action_classification/data_wait_seconds': 1.7135841846466064, 'eval/action_classification/predict_seconds': 81.0721435546875, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0007820670725777745, 'timing/model': 1.9230771420989186, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 1.9}) +[2026-06-23 19:46:11,834][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.02147454023361206, 'train/grad_norm_pre_clip': 23.787046432495117, 'timing/data': 0.0006082220934331417, 'timing/model': 1.9203908189665526, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 1.9}) +[2026-06-23 19:46:15,624][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.0732516497373581, 'train/grad_norm_pre_clip': 27.53812599182129, 'timing/data': 0.0006050869124010205, 'timing/model': 1.9221457639941946, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 1.9}) +[2026-06-23 19:46:19,411][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.02515631727874279, 'train/grad_norm_pre_clip': 24.19685173034668, 'timing/data': 0.0007535350741818547, 'timing/model': 1.9227659340249375, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 1.9}) +[2026-06-23 19:46:23,300][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.004181939177215099, 'train/grad_norm_pre_clip': 24.597476959228516, 'timing/data': 0.0005926770390942693, 'timing/model': 1.9170094049768522, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 1.9}) +[2026-06-23 19:46:27,075][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.05616813898086548, 'train/grad_norm_pre_clip': 41.678123474121094, 'timing/data': 0.000623362953774631, 'timing/model': 1.918667965917848, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 1.9}) +[2026-06-23 19:46:30,865][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.06381431967020035, 'train/grad_norm_pre_clip': 36.77290344238281, 'timing/data': 0.0005887099541723728, 'timing/model': 1.9201132439775392, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 1.9}) +[2026-06-23 19:46:34,658][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.023100726306438446, 'train/grad_norm_pre_clip': 24.81838607788086, 'timing/data': 0.0006033569807186723, 'timing/model': 1.9316238280152902, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 1.9}) +[2026-06-23 19:46:38,446][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.012863865122199059, 'train/grad_norm_pre_clip': 38.14262771606445, 'timing/data': 0.0006182929500937462, 'timing/model': 1.9186432430287823, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 1.9}) +[2026-06-23 19:46:42,238][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.021594326943159103, 'train/grad_norm_pre_clip': 30.87332534790039, 'timing/data': 0.0006835759850218892, 'timing/model': 1.928763019037433, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 1.9}) +[2026-06-23 19:46:46,020][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.10618282109498978, 'train/grad_norm_pre_clip': 47.381378173828125, 'timing/data': 0.0009024740429595113, 'timing/model': 1.9209998130099848, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 1.9}) +[2026-06-23 19:46:49,832][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.02692239172756672, 'train/grad_norm_pre_clip': 24.325319290161133, 'timing/data': 0.0006872930098325014, 'timing/model': 1.9302092819707468, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 1.9}) +[2026-06-23 19:46:53,694][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.05392435938119888, 'train/grad_norm_pre_clip': 28.357948303222656, 'timing/data': 0.0006186040118336678, 'timing/model': 2.0034010770032182, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 1.9}) +[2026-06-23 19:46:57,482][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.032124973833560944, 'train/grad_norm_pre_clip': 21.21410369873047, 'timing/data': 0.0006500609451904893, 'timing/model': 1.9246258880011737, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 1.9}) +[2026-06-23 19:47:01,295][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.05139189213514328, 'train/grad_norm_pre_clip': 23.73725700378418, 'timing/data': 0.0006134699797257781, 'timing/model': 1.9299237679224461, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 1.9}) +[2026-06-23 19:47:05,139][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.06216535344719887, 'train/grad_norm_pre_clip': 43.005523681640625, 'timing/data': 0.0005975519306957722, 'timing/model': 1.9749826119514182, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 1.9}) +[2026-06-23 19:47:08,923][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.08233728259801865, 'train/grad_norm_pre_clip': 25.667272567749023, 'timing/data': 0.0005891009932383895, 'timing/model': 1.9258808309677988, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 1.9}) +[2026-06-23 19:47:12,723][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.027451379224658012, 'train/grad_norm_pre_clip': 34.95452880859375, 'timing/data': 0.0008349569980055094, 'timing/model': 1.9317337770480663, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 1.9}) +[2026-06-23 19:47:16,515][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.033539846539497375, 'train/grad_norm_pre_clip': 38.17374801635742, 'timing/data': 0.0007198300445452332, 'timing/model': 1.9164418609580025, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 1.9}) +[2026-06-23 19:47:20,304][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.0967225581407547, 'train/grad_norm_pre_clip': 51.40336227416992, 'timing/data': 0.0006246799603104591, 'timing/model': 1.9161002750042826, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 1.9}) +[2026-06-23 19:47:24,090][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.012005489319562912, 'train/grad_norm_pre_clip': 24.37478256225586, 'timing/data': 0.0007409879472106695, 'timing/model': 1.9267515420215204, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 1.9}) +[2026-06-23 19:47:28,005][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.06841892004013062, 'train/grad_norm_pre_clip': 28.090742111206055, 'timing/data': 0.000603265012614429, 'timing/model': 2.044902357039973, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 1.9}) +[2026-06-23 19:47:31,803][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.006187081802636385, 'train/grad_norm_pre_clip': 22.38104248046875, 'timing/data': 0.0006099350284785032, 'timing/model': 1.926283412030898, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 1.9}) +[2026-06-23 19:47:35,641][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.014629274606704712, 'train/grad_norm_pre_clip': 19.656982421875, 'timing/data': 0.0006000240100547671, 'timing/model': 1.966435591923073, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 1.91}) +[2026-06-23 19:47:39,445][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.04613853991031647, 'train/grad_norm_pre_clip': 40.37933349609375, 'timing/data': 0.000633124029263854, 'timing/model': 1.927773401956074, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 1.91}) +[2026-06-23 19:47:43,246][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.0829683169722557, 'train/grad_norm_pre_clip': 59.0945930480957, 'timing/data': 0.0006238489877432585, 'timing/model': 1.93470053805504, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 1.91}) +[2026-06-23 19:47:47,044][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.039494775235652924, 'train/grad_norm_pre_clip': 40.047935485839844, 'timing/data': 0.0005922619020566344, 'timing/model': 1.928469633916393, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 1.91}) +[2026-06-23 19:47:50,837][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.007633579429239035, 'train/grad_norm_pre_clip': 16.236167907714844, 'timing/data': 0.0007452249992638826, 'timing/model': 1.9226648639887571, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 1.91}) +[2026-06-23 19:47:54,640][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.025169216096401215, 'train/grad_norm_pre_clip': 36.18535614013672, 'timing/data': 0.0006075759883970022, 'timing/model': 1.923271310981363, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 1.91}) +[2026-06-23 19:47:58,442][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.05522731691598892, 'train/grad_norm_pre_clip': 35.89443588256836, 'timing/data': 0.000610177987255156, 'timing/model': 1.9315442469669506, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 1.91}) +[2026-06-23 19:48:02,424][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.044075340032577515, 'train/grad_norm_pre_clip': 44.29707336425781, 'timing/data': 0.0006038090214133263, 'timing/model': 1.930854495964013, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 1.91}) +[2026-06-23 19:48:06,231][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.05249207466840744, 'train/grad_norm_pre_clip': 43.05278778076172, 'timing/data': 0.000653035007417202, 'timing/model': 1.9259497630409896, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 1.91}) +[2026-06-23 19:48:10,037][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.01210047211498022, 'train/grad_norm_pre_clip': 14.820693016052246, 'timing/data': 0.0007255080854520202, 'timing/model': 1.9350054030073807, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 1.91}) +[2026-06-23 19:48:13,836][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.08953944593667984, 'train/grad_norm_pre_clip': 46.785728454589844, 'timing/data': 0.0008215979905799031, 'timing/model': 1.9276508800685406, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 1.91}) +[2026-06-23 19:48:17,623][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.10505979508161545, 'train/grad_norm_pre_clip': 77.78089904785156, 'timing/data': 0.0006826519966125488, 'timing/model': 1.9231612840667367, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 1.91}) +[2026-06-23 19:48:21,426][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.09193102270364761, 'train/grad_norm_pre_clip': 29.834135055541992, 'timing/data': 0.000979752978309989, 'timing/model': 1.930013127042912, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 1.91}) +[2026-06-23 19:48:25,242][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.007539946585893631, 'train/grad_norm_pre_clip': 7.4732985496521, 'timing/data': 0.0006072790129110217, 'timing/model': 1.937761847046204, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 1.91}) +[2026-06-23 19:48:29,023][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.039077870547771454, 'train/grad_norm_pre_clip': 33.79145431518555, 'timing/data': 0.000858171028085053, 'timing/model': 1.9183023020159453, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 1.91}) +[2026-06-23 19:48:32,954][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.03442322090268135, 'train/grad_norm_pre_clip': 21.87778663635254, 'timing/data': 0.0006833299994468689, 'timing/model': 2.061127690016292, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 1.91}) +[2026-06-23 19:48:36,752][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.048871804028749466, 'train/grad_norm_pre_clip': 29.354251861572266, 'timing/data': 0.0005927769234403968, 'timing/model': 1.9279421319952235, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 1.91}) +[2026-06-23 19:48:40,556][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.01707891933619976, 'train/grad_norm_pre_clip': 24.45400619506836, 'timing/data': 0.000729974010027945, 'timing/model': 1.9316519839921966, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 1.91}) +[2026-06-23 19:48:44,337][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.015230190008878708, 'train/grad_norm_pre_clip': 51.03615188598633, 'timing/data': 0.000788407982327044, 'timing/model': 1.9168948999140412, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 1.91}) +[2026-06-23 19:48:48,147][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.018230197951197624, 'train/grad_norm_pre_clip': 21.50794792175293, 'timing/data': 0.0006717790383845568, 'timing/model': 1.9366095099831, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 1.91}) +[2026-06-23 19:48:51,961][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.02128119394183159, 'train/grad_norm_pre_clip': 31.241609573364258, 'timing/data': 0.000691442983224988, 'timing/model': 1.9310724439565092, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 1.91}) +[2026-06-23 19:48:55,748][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.008838105946779251, 'train/grad_norm_pre_clip': 13.800904273986816, 'timing/data': 0.0008227949729189277, 'timing/model': 1.9185239740181714, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 1.91}) +[2026-06-23 19:48:59,583][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.054357267916202545, 'train/grad_norm_pre_clip': 42.17353439331055, 'timing/data': 0.0006619259947910905, 'timing/model': 1.9592461350839585, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 1.91}) +[2026-06-23 19:49:03,390][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.01360254641622305, 'train/grad_norm_pre_clip': 15.657646179199219, 'timing/data': 0.0007720129797235131, 'timing/model': 1.9272372650448233, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 1.91}) +[2026-06-23 19:49:07,338][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.008052968420088291, 'train/grad_norm_pre_clip': 50.75717544555664, 'timing/data': 0.000732847023755312, 'timing/model': 1.9277247550198808, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 1.91}) +[2026-06-23 19:49:11,140][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.028723368421196938, 'train/grad_norm_pre_clip': 27.678617477416992, 'timing/data': 0.0006051339441910386, 'timing/model': 1.9309259300353006, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 1.91}) +[2026-06-23 19:49:14,944][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.03888009861111641, 'train/grad_norm_pre_clip': 49.7622184753418, 'timing/data': 0.0006067189387977123, 'timing/model': 1.9329453259706497, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 1.92}) +[2026-06-23 19:49:18,750][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.0489254854619503, 'train/grad_norm_pre_clip': 29.915172576904297, 'timing/data': 0.0005932049825787544, 'timing/model': 1.9344096760032699, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 1.92}) +[2026-06-23 19:49:22,551][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.048566024750471115, 'train/grad_norm_pre_clip': 25.652563095092773, 'timing/data': 0.0006142789497971535, 'timing/model': 1.924423680990003, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 1.92}) +[2026-06-23 19:49:26,357][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.023943452164530754, 'train/grad_norm_pre_clip': 11.613125801086426, 'timing/data': 0.0005963590228930116, 'timing/model': 1.9299446899676695, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 1.92}) +[2026-06-23 19:49:30,159][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.03260130435228348, 'train/grad_norm_pre_clip': 35.22821044921875, 'timing/data': 0.0006715400377288461, 'timing/model': 1.9250118689378724, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 1.92}) +[2026-06-23 19:49:33,963][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.02379494719207287, 'train/grad_norm_pre_clip': 19.134279251098633, 'timing/data': 0.0007268399931490421, 'timing/model': 1.9232209980254993, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 1.92}) +[2026-06-23 19:49:37,843][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.15112608671188354, 'train/grad_norm_pre_clip': 33.078125, 'timing/data': 0.0007405380019918084, 'timing/model': 2.005519847967662, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 1.92}) +[2026-06-23 19:49:41,653][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.009770470671355724, 'train/grad_norm_pre_clip': 12.891737937927246, 'timing/data': 0.000735615030862391, 'timing/model': 1.9335532210534438, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 1.92}) +[2026-06-23 19:49:45,455][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.06988689303398132, 'train/grad_norm_pre_clip': 53.04922866821289, 'timing/data': 0.0005996680120006204, 'timing/model': 1.9355433210730553, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 1.92}) +[2026-06-23 19:49:49,254][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.14743541181087494, 'train/grad_norm_pre_clip': 44.01416778564453, 'timing/data': 0.0005879619857296348, 'timing/model': 1.9305850930977613, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 1.92}) +[2026-06-23 19:49:53,058][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.059310056269168854, 'train/grad_norm_pre_clip': 37.88498306274414, 'timing/data': 0.0006012030644342303, 'timing/model': 1.9319457459496334, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 1.92}) +[2026-06-23 19:49:56,900][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.16553737223148346, 'train/grad_norm_pre_clip': 38.472564697265625, 'timing/data': 0.0006018300773575902, 'timing/model': 1.9628496289951727, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 1.92}) +[2026-06-23 19:50:00,706][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.006134992931038141, 'train/grad_norm_pre_clip': 23.787899017333984, 'timing/data': 0.0005922239506617188, 'timing/model': 1.9300599589478225, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 1.92}) +[2026-06-23 19:50:04,512][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.06926236301660538, 'train/grad_norm_pre_clip': 43.57198715209961, 'timing/data': 0.0006048150826245546, 'timing/model': 1.9326843430753797, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 1.92}) +[2026-06-23 19:50:08,309][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.07474297285079956, 'train/grad_norm_pre_clip': 29.547895431518555, 'timing/data': 0.0006164989899843931, 'timing/model': 1.9207523970399052, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 1.92}) +[2026-06-23 19:50:12,242][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.071070596575737, 'train/grad_norm_pre_clip': 34.830440521240234, 'timing/data': 0.0006780070252716541, 'timing/model': 2.0575736629543826, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 1.92}) +[2026-06-23 19:50:16,055][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.006446472369134426, 'train/grad_norm_pre_clip': 31.867538452148438, 'timing/data': 0.0007900919299572706, 'timing/model': 1.9331276600714773, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 1.92}) +[2026-06-23 19:50:19,866][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.015615819953382015, 'train/grad_norm_pre_clip': 19.262174606323242, 'timing/data': 0.0006037310231477022, 'timing/model': 1.9368088319897652, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 1.92}) +[2026-06-23 19:50:23,672][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.0058584390208125114, 'train/grad_norm_pre_clip': 26.06535530090332, 'timing/data': 0.0006800369592383504, 'timing/model': 1.9349450240842998, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 1.92}) +[2026-06-23 19:50:27,496][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.016682961955666542, 'train/grad_norm_pre_clip': 15.005365371704102, 'timing/data': 0.000581471947953105, 'timing/model': 1.9361364700598642, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 1.92}) +[2026-06-23 19:50:31,300][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.07866717129945755, 'train/grad_norm_pre_clip': 41.391197204589844, 'timing/data': 0.0005996719701215625, 'timing/model': 1.9319260399788618, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 1.92}) +[2026-06-23 19:50:35,125][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.008639766834676266, 'train/grad_norm_pre_clip': 28.964414596557617, 'timing/data': 0.0006356639787554741, 'timing/model': 1.9510545369703323, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 1.92}) +[2026-06-23 19:50:38,923][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.008094443008303642, 'train/grad_norm_pre_clip': 7.791703701019287, 'timing/data': 0.0006237269844859838, 'timing/model': 1.9307620460167527, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 1.92}) +[2026-06-23 19:50:42,722][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.03982705995440483, 'train/grad_norm_pre_clip': 41.93215560913086, 'timing/data': 0.0007245809538289905, 'timing/model': 1.9268740060506389, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 1.92}) +[2026-06-23 19:50:46,680][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.0148296058177948, 'train/grad_norm_pre_clip': 32.69044876098633, 'timing/data': 0.0006137869786471128, 'timing/model': 1.9332025150069967, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 1.92}) +[2026-06-23 19:50:50,487][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.03260355442762375, 'train/grad_norm_pre_clip': 13.708930969238281, 'timing/data': 0.0006687710992991924, 'timing/model': 1.9272865019738674, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 1.93}) +[2026-06-23 19:50:54,286][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.03413422778248787, 'train/grad_norm_pre_clip': 39.55284118652344, 'timing/data': 0.0005924600409343839, 'timing/model': 1.9346680539892986, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 1.93}) +[2026-06-23 19:50:58,094][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.014901939779520035, 'train/grad_norm_pre_clip': 20.609148025512695, 'timing/data': 0.0006060010055080056, 'timing/model': 1.9337509640026838, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 1.93}) +[2026-06-23 19:51:01,914][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.009568396024405956, 'train/grad_norm_pre_clip': 19.997779846191406, 'timing/data': 0.0006054609548300505, 'timing/model': 1.9462620429694653, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 1.93}) +[2026-06-23 19:51:05,723][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.016798337921500206, 'train/grad_norm_pre_clip': 30.4669246673584, 'timing/data': 0.0005946960300207138, 'timing/model': 1.9334043940762058, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 1.93}) +[2026-06-23 19:51:09,533][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.025843480601906776, 'train/grad_norm_pre_clip': 21.84569549560547, 'timing/data': 0.0008710360853001475, 'timing/model': 1.9343750340631232, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 1.93}) +[2026-06-23 19:51:13,345][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.024914385750889778, 'train/grad_norm_pre_clip': 25.527931213378906, 'timing/data': 0.0006498410366475582, 'timing/model': 1.9376460670027882, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 1.93}) +[2026-06-23 19:51:17,325][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.14066632091999054, 'train/grad_norm_pre_clip': 23.820947647094727, 'timing/data': 0.0005981980357319117, 'timing/model': 2.100663953111507, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 1.93}) +[2026-06-23 19:51:21,121][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.02981739677488804, 'train/grad_norm_pre_clip': 78.7283706665039, 'timing/data': 0.0006340979598462582, 'timing/model': 1.9289811489870772, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 1.93}) +[2026-06-23 19:51:24,920][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.05737879127264023, 'train/grad_norm_pre_clip': 27.670368194580078, 'timing/data': 0.0005770140560343862, 'timing/model': 1.929058738052845, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 1.93}) +[2026-06-23 19:51:28,714][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.04217683896422386, 'train/grad_norm_pre_clip': 44.96431350708008, 'timing/data': 0.0006984049687162042, 'timing/model': 1.9192453610012308, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 1.93}) +[2026-06-23 19:51:32,523][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.049097348004579544, 'train/grad_norm_pre_clip': 25.77703094482422, 'timing/data': 0.0006093719275668263, 'timing/model': 1.937886210042052, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 1.93}) +[2026-06-23 19:51:36,329][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.05117497965693474, 'train/grad_norm_pre_clip': 28.62836265563965, 'timing/data': 0.0006191639695316553, 'timing/model': 1.9407034810865298, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 1.93}) +[2026-06-23 19:51:40,133][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.040909912437200546, 'train/grad_norm_pre_clip': 40.293827056884766, 'timing/data': 0.000627067987807095, 'timing/model': 1.931310273008421, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 1.93}) +[2026-06-23 19:51:43,924][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.023283733054995537, 'train/grad_norm_pre_clip': 44.75204086303711, 'timing/data': 0.0007227490423247218, 'timing/model': 1.920204860973172, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 1.93}) +[2026-06-23 19:51:47,728][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.018038766458630562, 'train/grad_norm_pre_clip': 18.080425262451172, 'timing/data': 0.0006063840119168162, 'timing/model': 1.9337390980217606, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 1.93}) +[2026-06-23 19:51:51,725][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.06372081488370895, 'train/grad_norm_pre_clip': 11.961441040039062, 'timing/data': 0.0006060689920559525, 'timing/model': 1.9304821549449116, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 1.93}) +[2026-06-23 19:51:55,518][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.07585360109806061, 'train/grad_norm_pre_clip': 29.995689392089844, 'timing/data': 0.00059082405641675, 'timing/model': 1.918561325990595, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 1.93}) +[2026-06-23 19:51:59,360][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.03963572531938553, 'train/grad_norm_pre_clip': 36.31678009033203, 'timing/data': 0.0006158549804240465, 'timing/model': 1.929424270056188, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 1.93}) +[2026-06-23 19:52:03,163][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.12083710730075836, 'train/grad_norm_pre_clip': 30.85618782043457, 'timing/data': 0.0006072460673749447, 'timing/model': 1.9299877369776368, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 1.93}) +[2026-06-23 19:52:06,964][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.03509039431810379, 'train/grad_norm_pre_clip': 27.910676956176758, 'timing/data': 0.0006036839913576841, 'timing/model': 1.9300768879475072, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 1.93}) +[2026-06-23 19:52:10,758][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.008606205694377422, 'train/grad_norm_pre_clip': 10.677874565124512, 'timing/data': 0.0006192210130393505, 'timing/model': 1.9215772290481254, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 1.93}) +[2026-06-23 19:52:14,553][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.05473760515451431, 'train/grad_norm_pre_clip': 50.57927703857422, 'timing/data': 0.0006018949206918478, 'timing/model': 1.9237347870366648, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 1.93}) +[2026-06-23 19:52:18,359][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.02003164030611515, 'train/grad_norm_pre_clip': 23.854175567626953, 'timing/data': 0.0006091389805078506, 'timing/model': 1.9330880439374596, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 1.93}) +[2026-06-23 19:52:22,160][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.06641005724668503, 'train/grad_norm_pre_clip': 23.74993896484375, 'timing/data': 0.0008672400144860148, 'timing/model': 1.9270213029813021, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 1.93}) +[2026-06-23 19:52:26,160][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.08682411164045334, 'train/grad_norm_pre_clip': 42.203529357910156, 'timing/data': 0.0006087049841880798, 'timing/model': 1.9459905759431422, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 1.93}) +[2026-06-23 19:52:29,952][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 5000 +[2026-06-23 19:54:37,212][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 5000 +[2026-06-23 19:54:37,213][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 5000 +[2026-06-23 19:56:00,634][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 5000 +[2026-06-23 20:01:54,824][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 5000 +[2026-06-23 20:01:54,825][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.029060376808047295, 'train/grad_norm_pre_clip': 18.85932731628418, 'eval/action_loss/seconds': 127.25659483205527, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.03156288340687752, 'eval/latency_2/loss': 0.03156288340687752, 'eval/flappy/loss': 0.03156288340687752, 'eval/flappy/latency_2/loss': 0.03156288340687752, 'eval/flap_precision': 0.9708474576271187, 'eval/flap_recall': 0.9662618083670715, 'eval/flap_f1': 0.9685492052756172, 'eval/flap_rate_model': 0.2048611111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9685492052756173, 'eval/latency_2/flap_precision': 0.9708474576271187, 'eval/latency_2/flap_recall': 0.9662618083670715, 'eval/latency_2/flap_f1': 0.9685492052756172, 'eval/latency_2/flap_rate_model': 0.2048611111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9685492052756173, 'eval/action_classification/data_wait_seconds': 1.5402905941009521, 'eval/action_classification/predict_seconds': 81.69891357421875, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 217.48000000001034, 'rl_games_eval/mid_train/mean_length': 1771.0, 'rl_games_eval/mid_train/std_reward': 185.64113121828532, 'rl_games_eval/mid_train/std_length': 1494.9829430465086, 'rl_games_eval/mid_train/macro_mean_reward': 217.48000000001034, 'rl_games_eval/mid_train/macro_mean_length': 1771.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 217.48000000001034, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 1771.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 185.64113121828532, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 1494.9829430465086, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 217.48000000001034, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006135719595476985, 'timing/model': 1.9248917010845616, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 1.94}) +[2026-06-23 20:01:54,825][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_5000_state +[2026-06-23 20:01:54,828][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 20:01:58,638][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_5000_state/model.safetensors +[2026-06-23 20:02:01,551][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_5000_state/optimizer.bin +[2026-06-23 20:02:01,552][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_5000_state/sampler.bin +[2026-06-23 20:02:01,552][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-23 20:02:01,552][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-23 20:02:01,586][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-23 20:02:01,586][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 20:02:01,597][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 20:02:01,598][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected/checkpoints +[2026-06-23 20:02:04,877][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..a0e21985e6b7ed5b88fddcc5f74f6ec45063b603 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected/summary.jsonl @@ -0,0 +1,10 @@ +{"steps": 500} +{"steps": 1000} +{"steps": 1500} +{"steps": 2000} +{"steps": 2500} +{"steps": 3000} +{"steps": 3500} +{"steps": 4000} +{"steps": 4500} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_5000_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_5000_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..d36ee373b8187f63007ce6775e0692a9b6ff1473 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_5000_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b2375e35dca20389af04508aba967acffda126a63d3daa8344629d271bdcac3d +size 9138230516 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..2b2bf61be8218d9dadb7c0608c5be3cba5c80cb3 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/config.full.yaml @@ -0,0 +1,226 @@ +framework: + name: QwenOFT + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + action_model_type: MLP + action_dim: 7 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: discrete_ce + state_dim: 7 + action_horizon: 1 + action_env_dim: 2 +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 64 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: true + freeze_tied_embedding: true + freeze_llm_layers: + - 0 + - 27 + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: none + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: data/flappy_fix_latency_2_200ep + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: none + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_last_8_layers_corrected_v2 +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2 +config_yaml: null +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..f489e2bccdea01799ffd10cf5e2155c3e03b5e77 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/config.yaml @@ -0,0 +1,105 @@ +checkpoint: + local: + keep_last_n: 1 + save_best_model: false + save_pt_file: false + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_train__bridge + dataset_py: lerobot_datasets + eval_data_mix: flappy_train__bridge__val + eval_num_workers: 8 + include_state: true + latency_curriculum: + enabled: false + obs_image_size: null + per_device_batch_size: 64 + persistent_workers: true + pin_memory: true + prefetch_factor: 4 +framework: + action_model: + action_dim: 7 + action_env_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + loss_type: discrete_ce + state_dim: 7 + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + enable_gradient_checkpointing: true +output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2 +rl_games: + env_eval: + distributed_mode: none + enabled: true + fixed_episode_seeds: true + frameskip: 1 + image_size: 224 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + latency_seed_stride: 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + max_steps_per_episode: 3600 + num_episodes: 5 + seed: 42 + task_description: '' + task_seed_stride: 0 + vectorized: + enabled: false + model_alias: openvla + task: flappy +run_id: flappy_fix_latency_2_200ep_last_8_layers_corrected_v2 +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +trainer: + distributed_backend: none + eval_action_classification: true + eval_action_classification_interval: null + eval_interval: 100 + eval_num_batches: 100 + freeze_llm_layers: + - 0 + - 27 + freeze_modules: '' + freeze_tied_embedding: true + freeze_vit: true + gradient_accumulation_steps: 2 + gradient_clipping: 1.0 + is_resume: false + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + logging_frequency: 1 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 5000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: true + weight_decay: 1.0e-08 + per_latency_eval_num_batches: null + pretrained_checkpoint: null + profile_timing: + enabled: false + reload_modules: null + save_interval: 500 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..aa03e6cbf8e375d660528e17fafddb4804b596b7 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/dataset_statistics.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.6028500199317932, + 0.3971499800682068, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4890792667865753, + 0.4890792667865753, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 330734, + "num_trajectories": 180 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..0caa254376d5805f972b0b0411e0b5188e3634e3 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/dataset_statistics_eval.json @@ -0,0 +1,127 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7959861159324646, + 0.2040138840675354, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.4030573070049286, + 0.4030573070049286, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "mask": [ + true, + true, + true, + true, + true, + true, + true + ] + }, + "state": { + "mean": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "std": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "max": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "min": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ], + "q99": [ + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0, + 0.0 + ] + }, + "num_transitions": 72000, + "num_trajectories": 20 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra/.hydra/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra/.hydra/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..b235723c34a82aa1df75397225f286679debe452 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra/.hydra/config.yaml @@ -0,0 +1,224 @@ +framework: + qwenvl: + base_vlm: /workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + action_model: + state_dim: 7 + loss_type: discrete_ce + action_horizon: 1 + future_action_window_size: 0 + past_action_window_size: 0 + action_dim: 7 + action_env_dim: 2 + name: QwenOFT +datasets: + vla_data: + dataset_py: lerobot_datasets + include_state: true + data_root_dir: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + data_mix: flappy_train__bridge + eval_data_mix: flappy_train__bridge__val + custom_mixtures_path: null + action_type: discrete + sequential_step_sampling: false + eval_sequential_step_sampling: null + num_workers: 8 + eval_num_workers: 8 + prefetch_factor: 4 + persistent_workers: true + pin_memory: true + shuffle: true + action_balance: + enabled: false + strategy: balanced_epoch + action_key: action_id + target_flap_fraction: 0.3 + noop_id: 0 + flap_id: 1 + latency_curriculum: + enabled: false + strategy: exclusive + latencies: null + phase_steps: null + per_device_batch_size: 64 + load_all_data_for_training: true + num_obs_frames: 1 + image_mode: single + stitch_grid: + - 2 + - 2 + obs_image_size: null + video_backend: torchvision_av +dataset: + source_hf: '' + config_name: null + source_subdir: null + converted_name: flappy_train + single_source_hf: '' + mixed_source_hf: '' + single_converted_name: flappy_train + mixed_converted_name: flappy_mixed_latency_train + single_latency_filter: null + mixed_latency_filter: null + force_download: false + setup_force: false + skip_verification: false + verify_rows: 200 + max_episodes: null + episodes_per_latency: null + latency_filter: null + debug_subset: + enabled: false + max_episodes: 5 + suffix: debug +base_model: + repo_id: Qwen/Qwen3-VL-4B-Instruct +initialization: + checkpoint_local_dir: playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_hf_repo_id: StarVLA/Qwen3VL-OFT-Bridge-RT-1 + checkpoint_filename: checkpoints/steps_5000_pytorch_model.pt +trainer: + max_train_steps: 5000 + num_warmup_steps: 100 + save_interval: 500 + eval_interval: 100 + eval_num_batches: 100 + per_latency_eval_num_batches: null + eval_action_classification: true + eval_action_classification_interval: null + cc_f1_tolerance: 1 + learning_rate: + base: 2.0e-05 + qwen_vl_interface: 1.0e-05 + action_model: 0.0001 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + freeze_vit: true + freeze_tied_embedding: true + freeze_llm_layers: + - 0 + - 27 + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 1 + profile_timing: + enabled: false + log_interval: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: none + is_resume: false + pretrained_checkpoint: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + resume_step: 0 + reload_modules: null + optimizer: + name: AdamW + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: true + save_format: pt +workspace_dir: WORKSPACE_DIR +run_root_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +auth: + env_file: null + hf_token_env: HF_TOKEN + wandb_api_key_env: WANDB_API_KEY +paths: + run_root_dir: results/Checkpoints + dataset_local_dir: data/flappy_fix_latency_2_200ep + dataset_cache_dir: null + base_model_dir: playground/Pretrained_models/Qwen3-VL-4B-Instruct + accelerate_config: starVLA/config/deepseeds/deepspeed_zero2.yaml +launch: + use_accelerate: true + gpus: null + num_processes: 1 + dry_run: false +conda: + enabled: true + env_name: null +rl_games: + model_alias: openvla + env_eval: + image_size: 224 + frameskip: 1 + seed: 42 + fixed_episode_seeds: true + latency_seed_stride: 0 + task_seed_stride: 0 + task_description: '' + enabled: true + distributed_mode: none + vectorized: + enabled: false + batch_size: 1 + latency: + prompt_map_path: /workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + mode: single + values: + - 0 + mid_train: + enabled: true + interval_steps: 250 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 3600 + post_train: + enabled: false + latencies: + - 0 + - 1 + - 2 + - 3 + - 4 + num_episodes: 5 + max_steps_per_episode: 3600 + task: flappy + initialization_mode: bridge + action_carrier: bridge +model: openvla +env: flappy +init: bridge +bridge_base_model: + repo_id: + openvla: Qwen/Qwen3-VL-4B-Instruct + pi0: StarVLA/Qwen2.5-VL-3B-Instruct-Action + pi05: Qwen/Qwen3-VL-4B-Instruct + gr00t: Qwen/Qwen3-VL-4B-Instruct + local_dir: + openvla: playground/Pretrained_models/Qwen3-VL-4B-Instruct + pi0: playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + pi05: playground/Pretrained_models/Qwen3-VL-4B-Instruct + gr00t: playground/Pretrained_models/Qwen3-VL-4B-Instruct +mode: single +checkpoint: + load: none + hf_repo_id: null + save_best_model: false + save_pt_file: false + local: + keep_last_n: 1 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: flappy_fix_latency_2_200ep_last_8_layers_corrected_v2 +output_dir: null +config_yaml: null +is_debug: false +version_id: 0.21 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra/.hydra/hydra.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra/.hydra/hydra.yaml new file mode 100644 index 0000000000000000000000000000000000000000..51e77e675a4581f2d659202b530aca2aadfb0dc8 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra/.hydra/hydra.yaml @@ -0,0 +1,341 @@ +hydra: + run: + dir: ${run_root_dir}/${run_id}/hydra + sweep: + dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S} + subdir: ${hydra.job.num} + launcher: + _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher + sweeper: + _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper + max_batch_size: null + params: null + help: + app_name: ${hydra.job.name} + header: '${hydra.help.app_name} is powered by Hydra. + + ' + footer: 'Powered by Hydra (https://hydra.cc) + + Use --hydra-help to view Hydra specific help + + ' + template: '${hydra.help.header} + + == Configuration groups == + + Compose your configuration from those groups (group=option) + + + $APP_CONFIG_GROUPS + + + == Config == + + Override anything in the config (foo.bar=value) + + + $CONFIG + + + ${hydra.help.footer} + + ' + hydra_help: + template: 'Hydra (${hydra.runtime.version}) + + See https://hydra.cc for more info. + + + == Flags == + + $FLAGS_HELP + + + == Configuration groups == + + Compose your configuration from those groups (For example, append hydra/job_logging=disabled + to command line) + + + $HYDRA_CONFIG_GROUPS + + + Use ''--cfg hydra'' to Show the Hydra config. + + ' + hydra_help: ??? + hydra_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][HYDRA] %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + root: + level: INFO + handlers: + - console + loggers: + logging_example: + level: DEBUG + disable_existing_loggers: false + job_logging: + version: 1 + formatters: + simple: + format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s' + handlers: + console: + class: logging.StreamHandler + formatter: simple + stream: ext://sys.stdout + file: + class: logging.FileHandler + formatter: simple + filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log + root: + level: INFO + handlers: + - console + - file + disable_existing_loggers: false + env: {} + mode: RUN + searchpath: [] + callbacks: {} + output_subdir: .hydra + overrides: + hydra: + - hydra.mode=RUN + task: + - model=openvla + - env=flappy + - init=bridge + - mode=single + - ++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++framework.qwenvl.attn_implementation=flash_attention_2 + - ++framework.qwenvl.enable_gradient_checkpointing=true + - ++framework.action_model.state_dim=7 + - ++framework.action_model.loss_type=discrete_ce + - ++framework.action_model.action_horizon=1 + - ++framework.action_model.future_action_window_size=0 + - ++framework.action_model.past_action_window_size=0 + - ++framework.action_model.action_dim=7 + - ++framework.action_model.action_env_dim=2 + - ++framework.name=QwenOFT + - ++datasets.vla_data.dataset_py=lerobot_datasets + - ++datasets.vla_data.include_state=true + - ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games + - ++datasets.vla_data.data_mix=flappy_train + - ++datasets.vla_data.eval_data_mix=null + - ++datasets.vla_data.custom_mixtures_path=null + - ++datasets.vla_data.action_type=discrete + - ++datasets.vla_data.sequential_step_sampling=false + - ++datasets.vla_data.eval_sequential_step_sampling=null + - ++datasets.vla_data.num_workers=8 + - ++datasets.vla_data.eval_num_workers=8 + - ++datasets.vla_data.prefetch_factor=4 + - ++datasets.vla_data.persistent_workers=true + - ++datasets.vla_data.pin_memory=true + - ++datasets.vla_data.shuffle=true + - ++datasets.vla_data.action_balance.enabled=false + - ++datasets.vla_data.action_balance.strategy=balanced_epoch + - ++datasets.vla_data.action_balance.action_key=action_id + - ++datasets.vla_data.action_balance.target_flap_fraction=0.3 + - ++datasets.vla_data.action_balance.noop_id=0 + - ++datasets.vla_data.action_balance.flap_id=1 + - ++datasets.vla_data.latency_curriculum.enabled=false + - ++datasets.vla_data.latency_curriculum.strategy=exclusive + - ++datasets.vla_data.latency_curriculum.latencies=null + - ++datasets.vla_data.latency_curriculum.phase_steps=null + - ++datasets.vla_data.per_device_batch_size=64 + - ++datasets.vla_data.load_all_data_for_training=true + - ++datasets.vla_data.num_obs_frames=1 + - ++datasets.vla_data.image_mode=single + - ++datasets.vla_data.stitch_grid=[2,2] + - ++datasets.vla_data.obs_image_size=null + - ++datasets.vla_data.video_backend=torchvision_av + - ++dataset.source_hf= + - ++dataset.config_name=null + - ++dataset.source_subdir=null + - ++dataset.converted_name=flappy_train + - ++dataset.single_source_hf= + - ++dataset.mixed_source_hf= + - ++dataset.single_converted_name=flappy_train + - ++dataset.mixed_converted_name=flappy_mixed_latency_train + - ++dataset.single_latency_filter=null + - ++dataset.mixed_latency_filter=null + - ++dataset.force_download=false + - ++dataset.setup_force=false + - ++dataset.skip_verification=false + - ++dataset.verify_rows=200 + - ++dataset.max_episodes=null + - ++dataset.episodes_per_latency=null + - ++dataset.latency_filter=null + - ++dataset.debug_subset.enabled=false + - ++dataset.debug_subset.max_episodes=5 + - ++dataset.debug_subset.suffix=debug + - ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct + - ++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 + - ++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1 + - ++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt + - ++trainer.max_train_steps=5000 + - ++trainer.num_warmup_steps=100 + - ++trainer.save_interval=500 + - ++trainer.eval_interval=100 + - ++trainer.eval_num_batches=100 + - ++trainer.per_latency_eval_num_batches=null + - ++trainer.eval_action_classification=true + - ++trainer.eval_action_classification_interval=null + - ++trainer.cc_f1_tolerance=1 + - ++trainer.learning_rate.base=2e-05 + - ++trainer.learning_rate.qwen_vl_interface=1e-05 + - ++trainer.learning_rate.action_model=0.0001 + - ++trainer.lr_scheduler_type=cosine_with_min_lr + - ++trainer.scheduler_specific_kwargs.min_lr=1e-06 + - ++trainer.freeze_modules= + - ++trainer.freeze_vit=true + - ++trainer.freeze_tied_embedding=true + - ++trainer.freeze_llm_layers=[0,27] + - ++trainer.loss_scale.vla=1.0 + - ++trainer.loss_scale.vlm=0.1 + - ++trainer.max_grad_norm=1.0 + - ++trainer.weight_decay=0.0 + - ++trainer.logging_frequency=1 + - ++trainer.profile_timing.enabled=false + - ++trainer.profile_timing.log_interval=10 + - ++trainer.gradient_clipping=1.0 + - ++trainer.gradient_accumulation_steps=2 + - ++trainer.distributed_backend=none + - ++trainer.is_resume=false + - ++trainer.pretrained_checkpoint=null + - ++trainer.resume_step=0 + - ++trainer.reload_modules=null + - ++trainer.optimizer.name=AdamW + - ++trainer.optimizer.betas=[0.9,0.95] + - ++trainer.optimizer.eps=1e-08 + - ++trainer.optimizer.weight_decay=1e-08 + - ++trainer.optimizer.fused=true + - ++trainer.save_format=pt + - ++workspace_dir=WORKSPACE_DIR + - ++run_root_dir=results/Checkpoints + - ++seed=42 + - ++wandb_entity=saberrr-zju + - ++wandb_project=starVLA_rl_games + - ++auth.env_file=null + - ++auth.hf_token_env=HF_TOKEN + - ++auth.wandb_api_key_env=WANDB_API_KEY + - ++paths.run_root_dir=results/Checkpoints + - ++paths.dataset_local_dir=data/flappy_fix_latency_2_200ep + - ++paths.dataset_cache_dir=null + - ++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml + - ++rl_games.model_alias=openvla + - ++rl_games.env_eval.image_size=224 + - ++rl_games.env_eval.frameskip=1 + - ++rl_games.env_eval.seed=42 + - ++rl_games.env_eval.fixed_episode_seeds=true + - ++rl_games.env_eval.latency_seed_stride=0 + - ++rl_games.env_eval.task_seed_stride=0 + - ++rl_games.env_eval.task_description= + - ++rl_games.env_eval.enabled=true + - ++rl_games.env_eval.distributed_mode=none + - ++rl_games.env_eval.vectorized.enabled=false + - ++rl_games.env_eval.vectorized.batch_size=1 + - ++rl_games.env_eval.latency.prompt_map_path=null + - ++rl_games.env_eval.latency.mode=single + - ++rl_games.env_eval.latency.values=[0] + - ++rl_games.env_eval.mid_train.enabled=true + - ++rl_games.env_eval.mid_train.interval_steps=250 + - ++rl_games.env_eval.mid_train.latencies=[2] + - ++rl_games.env_eval.mid_train.num_episodes=5 + - ++rl_games.env_eval.mid_train.max_steps_per_episode=3600 + - ++rl_games.env_eval.post_train.enabled=false + - ++rl_games.env_eval.post_train.latencies=[0,1,2,3,4] + - ++rl_games.env_eval.post_train.num_episodes=5 + - ++rl_games.env_eval.post_train.max_steps_per_episode=3600 + - ++rl_games.task=flappy + - ++rl_games.initialization_mode=bridge + - ++rl_games.action_carrier=bridge + - ++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct + - ++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action + - ++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct + - ++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct + - ++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action + - ++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++checkpoint.load=none + - ++checkpoint.hf_repo_id=null + - ++checkpoint.save_best_model=false + - ++checkpoint.save_pt_file=false + - ++checkpoint.local.keep_last_n=1 + - ++checkpoint.sync.enabled=false + - ++checkpoint.sync.repo_id=null + - ++checkpoint.sync.keep_last_n=0 + - ++checkpoint.sync.sync_every_n_checkpoints=1 + - ++checkpoint.sync.resume_policy=local_latest + - ++run_id=flappy_fix_latency_2_200ep_last_8_layers_corrected_v2 + - ++output_dir=null + - ++config_yaml=null + - ++is_debug=false + - ++version_id=0.21 + - ++run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints + - ++trainer.is_resume=false + - ++trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt + - ++trainer.resume_step=0 + - ++datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep + - ++datasets.vla_data.data_mix=flappy_train__bridge + - ++datasets.vla_data.eval_data_mix=flappy_train__bridge__val + - ++framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct + - ++rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json + job: + name: train_starvla_hydra + chdir: false + override_dirname: ++auth.env_file=null,++auth.hf_token_env=HF_TOKEN,++auth.wandb_api_key_env=WANDB_API_KEY,++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action,++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action,++checkpoint.hf_repo_id=null,++checkpoint.load=none,++checkpoint.local.keep_last_n=1,++checkpoint.save_best_model=false,++checkpoint.save_pt_file=false,++checkpoint.sync.enabled=false,++checkpoint.sync.keep_last_n=0,++checkpoint.sync.repo_id=null,++checkpoint.sync.resume_policy=local_latest,++checkpoint.sync.sync_every_n_checkpoints=1,++config_yaml=null,++dataset.config_name=null,++dataset.converted_name=flappy_train,++dataset.debug_subset.enabled=false,++dataset.debug_subset.max_episodes=5,++dataset.debug_subset.suffix=debug,++dataset.episodes_per_latency=null,++dataset.force_download=false,++dataset.latency_filter=null,++dataset.max_episodes=null,++dataset.mixed_converted_name=flappy_mixed_latency_train,++dataset.mixed_latency_filter=null,++dataset.mixed_source_hf=,++dataset.setup_force=false,++dataset.single_converted_name=flappy_train,++dataset.single_latency_filter=null,++dataset.single_source_hf=,++dataset.skip_verification=false,++dataset.source_hf=,++dataset.source_subdir=null,++dataset.verify_rows=200,++datasets.vla_data.action_balance.action_key=action_id,++datasets.vla_data.action_balance.enabled=false,++datasets.vla_data.action_balance.flap_id=1,++datasets.vla_data.action_balance.noop_id=0,++datasets.vla_data.action_balance.strategy=balanced_epoch,++datasets.vla_data.action_balance.target_flap_fraction=0.3,++datasets.vla_data.action_type=discrete,++datasets.vla_data.custom_mixtures_path=null,++datasets.vla_data.data_mix=flappy_train,++datasets.vla_data.data_mix=flappy_train__bridge,++datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep,++datasets.vla_data.data_root_dir=playground/Datasets/rl_games,++datasets.vla_data.dataset_py=lerobot_datasets,++datasets.vla_data.eval_data_mix=flappy_train__bridge__val,++datasets.vla_data.eval_data_mix=null,++datasets.vla_data.eval_num_workers=8,++datasets.vla_data.eval_sequential_step_sampling=null,++datasets.vla_data.image_mode=single,++datasets.vla_data.include_state=true,++datasets.vla_data.latency_curriculum.enabled=false,++datasets.vla_data.latency_curriculum.latencies=null,++datasets.vla_data.latency_curriculum.phase_steps=null,++datasets.vla_data.latency_curriculum.strategy=exclusive,++datasets.vla_data.load_all_data_for_training=true,++datasets.vla_data.num_obs_frames=1,++datasets.vla_data.num_workers=8,++datasets.vla_data.obs_image_size=null,++datasets.vla_data.per_device_batch_size=64,++datasets.vla_data.persistent_workers=true,++datasets.vla_data.pin_memory=true,++datasets.vla_data.prefetch_factor=4,++datasets.vla_data.sequential_step_sampling=false,++datasets.vla_data.shuffle=true,++datasets.vla_data.stitch_grid=[2,2],++datasets.vla_data.video_backend=torchvision_av,++framework.action_model.action_dim=7,++framework.action_model.action_env_dim=2,++framework.action_model.action_horizon=1,++framework.action_model.future_action_window_size=0,++framework.action_model.loss_type=discrete_ce,++framework.action_model.past_action_window_size=0,++framework.action_model.state_dim=7,++framework.name=QwenOFT,++framework.qwenvl.attn_implementation=flash_attention_2,++framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.enable_gradient_checkpointing=true,++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt,++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1,++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1,++is_debug=false,++output_dir=null,++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml,++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++paths.dataset_cache_dir=null,++paths.dataset_local_dir=data/flappy_fix_latency_2_200ep,++paths.run_root_dir=results/Checkpoints,++rl_games.action_carrier=bridge,++rl_games.env_eval.distributed_mode=none,++rl_games.env_eval.enabled=true,++rl_games.env_eval.fixed_episode_seeds=true,++rl_games.env_eval.frameskip=1,++rl_games.env_eval.image_size=224,++rl_games.env_eval.latency.mode=single,++rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json,++rl_games.env_eval.latency.prompt_map_path=null,++rl_games.env_eval.latency.values=[0],++rl_games.env_eval.latency_seed_stride=0,++rl_games.env_eval.mid_train.enabled=true,++rl_games.env_eval.mid_train.interval_steps=250,++rl_games.env_eval.mid_train.latencies=[2],++rl_games.env_eval.mid_train.max_steps_per_episode=3600,++rl_games.env_eval.mid_train.num_episodes=5,++rl_games.env_eval.post_train.enabled=false,++rl_games.env_eval.post_train.latencies=[0,1,2,3,4],++rl_games.env_eval.post_train.max_steps_per_episode=3600,++rl_games.env_eval.post_train.num_episodes=5,++rl_games.env_eval.seed=42,++rl_games.env_eval.task_description=,++rl_games.env_eval.task_seed_stride=0,++rl_games.env_eval.vectorized.batch_size=1,++rl_games.env_eval.vectorized.enabled=false,++rl_games.initialization_mode=bridge,++rl_games.model_alias=openvla,++rl_games.task=flappy,++run_id=flappy_fix_latency_2_200ep_last_8_layers_corrected_v2,++run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints,++run_root_dir=results/Checkpoints,++seed=42,++trainer.cc_f1_tolerance=1,++trainer.distributed_backend=none,++trainer.eval_action_classification=true,++trainer.eval_action_classification_interval=null,++trainer.eval_interval=100,++trainer.eval_num_batches=100,++trainer.freeze_llm_layers=[0,27],++trainer.freeze_modules=,++trainer.freeze_tied_embedding=true,++trainer.freeze_vit=true,++trainer.gradient_accumulation_steps=2,++trainer.gradient_clipping=1.0,++trainer.is_resume=false,++trainer.is_resume=false,++trainer.learning_rate.action_model=0.0001,++trainer.learning_rate.base=2e-05,++trainer.learning_rate.qwen_vl_interface=1e-05,++trainer.logging_frequency=1,++trainer.loss_scale.vla=1.0,++trainer.loss_scale.vlm=0.1,++trainer.lr_scheduler_type=cosine_with_min_lr,++trainer.max_grad_norm=1.0,++trainer.max_train_steps=5000,++trainer.num_warmup_steps=100,++trainer.optimizer.betas=[0.9,0.95],++trainer.optimizer.eps=1e-08,++trainer.optimizer.fused=true,++trainer.optimizer.name=AdamW,++trainer.optimizer.weight_decay=1e-08,++trainer.per_latency_eval_num_batches=null,++trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt,++trainer.pretrained_checkpoint=null,++trainer.profile_timing.enabled=false,++trainer.profile_timing.log_interval=10,++trainer.reload_modules=null,++trainer.resume_step=0,++trainer.resume_step=0,++trainer.save_format=pt,++trainer.save_interval=500,++trainer.scheduler_specific_kwargs.min_lr=1e-06,++trainer.weight_decay=0.0,++version_id=0.21,++wandb_entity=saberrr-zju,++wandb_project=starVLA_rl_games,++workspace_dir=WORKSPACE_DIR,env=flappy,init=bridge,mode=single,model=openvla + id: ??? + num: ??? + config_name: train + env_set: {} + env_copy: [] + config: + override_dirname: + kv_sep: '=' + item_sep: ',' + exclude_keys: [] + runtime: + version: 1.3.3 + version_base: '1.1' + cwd: /workspace/latency-sensitive-bench/starVLA + config_sources: + - path: hydra.conf + schema: pkg + provider: hydra + - path: /workspace/latency-sensitive-bench/starVLA/examples/rl_games/config + schema: file + provider: main + - path: '' + schema: structured + provider: schema + output_dir: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra + choices: + cross_task_setup: null + checkpoint: default + mode: single + init: bridge + env: flappy + model: openvla + hydra/env: default + hydra/callbacks: null + hydra/job_logging: default + hydra/hydra_logging: default + hydra/hydra_help: default + hydra/help: default + hydra/sweeper: basic + hydra/launcher: basic + hydra/output: default + verbose: false diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra/.hydra/overrides.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra/.hydra/overrides.yaml new file mode 100644 index 0000000000000000000000000000000000000000..d5d22fbcd7f904d2b2ba3a5696809259b59f9249 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra/.hydra/overrides.yaml @@ -0,0 +1,181 @@ +- model=openvla +- env=flappy +- init=bridge +- mode=single +- ++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++framework.qwenvl.attn_implementation=flash_attention_2 +- ++framework.qwenvl.enable_gradient_checkpointing=true +- ++framework.action_model.state_dim=7 +- ++framework.action_model.loss_type=discrete_ce +- ++framework.action_model.action_horizon=1 +- ++framework.action_model.future_action_window_size=0 +- ++framework.action_model.past_action_window_size=0 +- ++framework.action_model.action_dim=7 +- ++framework.action_model.action_env_dim=2 +- ++framework.name=QwenOFT +- ++datasets.vla_data.dataset_py=lerobot_datasets +- ++datasets.vla_data.include_state=true +- ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games +- ++datasets.vla_data.data_mix=flappy_train +- ++datasets.vla_data.eval_data_mix=null +- ++datasets.vla_data.custom_mixtures_path=null +- ++datasets.vla_data.action_type=discrete +- ++datasets.vla_data.sequential_step_sampling=false +- ++datasets.vla_data.eval_sequential_step_sampling=null +- ++datasets.vla_data.num_workers=8 +- ++datasets.vla_data.eval_num_workers=8 +- ++datasets.vla_data.prefetch_factor=4 +- ++datasets.vla_data.persistent_workers=true +- ++datasets.vla_data.pin_memory=true +- ++datasets.vla_data.shuffle=true +- ++datasets.vla_data.action_balance.enabled=false +- ++datasets.vla_data.action_balance.strategy=balanced_epoch +- ++datasets.vla_data.action_balance.action_key=action_id +- ++datasets.vla_data.action_balance.target_flap_fraction=0.3 +- ++datasets.vla_data.action_balance.noop_id=0 +- ++datasets.vla_data.action_balance.flap_id=1 +- ++datasets.vla_data.latency_curriculum.enabled=false +- ++datasets.vla_data.latency_curriculum.strategy=exclusive +- ++datasets.vla_data.latency_curriculum.latencies=null +- ++datasets.vla_data.latency_curriculum.phase_steps=null +- ++datasets.vla_data.per_device_batch_size=64 +- ++datasets.vla_data.load_all_data_for_training=true +- ++datasets.vla_data.num_obs_frames=1 +- ++datasets.vla_data.image_mode=single +- ++datasets.vla_data.stitch_grid=[2,2] +- ++datasets.vla_data.obs_image_size=null +- ++datasets.vla_data.video_backend=torchvision_av +- ++dataset.source_hf= +- ++dataset.config_name=null +- ++dataset.source_subdir=null +- ++dataset.converted_name=flappy_train +- ++dataset.single_source_hf= +- ++dataset.mixed_source_hf= +- ++dataset.single_converted_name=flappy_train +- ++dataset.mixed_converted_name=flappy_mixed_latency_train +- ++dataset.single_latency_filter=null +- ++dataset.mixed_latency_filter=null +- ++dataset.force_download=false +- ++dataset.setup_force=false +- ++dataset.skip_verification=false +- ++dataset.verify_rows=200 +- ++dataset.max_episodes=null +- ++dataset.episodes_per_latency=null +- ++dataset.latency_filter=null +- ++dataset.debug_subset.enabled=false +- ++dataset.debug_subset.max_episodes=5 +- ++dataset.debug_subset.suffix=debug +- ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct +- ++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1 +- ++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1 +- ++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt +- ++trainer.max_train_steps=5000 +- ++trainer.num_warmup_steps=100 +- ++trainer.save_interval=500 +- ++trainer.eval_interval=100 +- ++trainer.eval_num_batches=100 +- ++trainer.per_latency_eval_num_batches=null +- ++trainer.eval_action_classification=true +- ++trainer.eval_action_classification_interval=null +- ++trainer.cc_f1_tolerance=1 +- ++trainer.learning_rate.base=2e-05 +- ++trainer.learning_rate.qwen_vl_interface=1e-05 +- ++trainer.learning_rate.action_model=0.0001 +- ++trainer.lr_scheduler_type=cosine_with_min_lr +- ++trainer.scheduler_specific_kwargs.min_lr=1e-06 +- ++trainer.freeze_modules= +- ++trainer.freeze_vit=true +- ++trainer.freeze_tied_embedding=true +- ++trainer.freeze_llm_layers=[0,27] +- ++trainer.loss_scale.vla=1.0 +- ++trainer.loss_scale.vlm=0.1 +- ++trainer.max_grad_norm=1.0 +- ++trainer.weight_decay=0.0 +- ++trainer.logging_frequency=1 +- ++trainer.profile_timing.enabled=false +- ++trainer.profile_timing.log_interval=10 +- ++trainer.gradient_clipping=1.0 +- ++trainer.gradient_accumulation_steps=2 +- ++trainer.distributed_backend=none +- ++trainer.is_resume=false +- ++trainer.pretrained_checkpoint=null +- ++trainer.resume_step=0 +- ++trainer.reload_modules=null +- ++trainer.optimizer.name=AdamW +- ++trainer.optimizer.betas=[0.9,0.95] +- ++trainer.optimizer.eps=1e-08 +- ++trainer.optimizer.weight_decay=1e-08 +- ++trainer.optimizer.fused=true +- ++trainer.save_format=pt +- ++workspace_dir=WORKSPACE_DIR +- ++run_root_dir=results/Checkpoints +- ++seed=42 +- ++wandb_entity=saberrr-zju +- ++wandb_project=starVLA_rl_games +- ++auth.env_file=null +- ++auth.hf_token_env=HF_TOKEN +- ++auth.wandb_api_key_env=WANDB_API_KEY +- ++paths.run_root_dir=results/Checkpoints +- ++paths.dataset_local_dir=data/flappy_fix_latency_2_200ep +- ++paths.dataset_cache_dir=null +- ++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml +- ++rl_games.model_alias=openvla +- ++rl_games.env_eval.image_size=224 +- ++rl_games.env_eval.frameskip=1 +- ++rl_games.env_eval.seed=42 +- ++rl_games.env_eval.fixed_episode_seeds=true +- ++rl_games.env_eval.latency_seed_stride=0 +- ++rl_games.env_eval.task_seed_stride=0 +- ++rl_games.env_eval.task_description= +- ++rl_games.env_eval.enabled=true +- ++rl_games.env_eval.distributed_mode=none +- ++rl_games.env_eval.vectorized.enabled=false +- ++rl_games.env_eval.vectorized.batch_size=1 +- ++rl_games.env_eval.latency.prompt_map_path=null +- ++rl_games.env_eval.latency.mode=single +- ++rl_games.env_eval.latency.values=[0] +- ++rl_games.env_eval.mid_train.enabled=true +- ++rl_games.env_eval.mid_train.interval_steps=250 +- ++rl_games.env_eval.mid_train.latencies=[2] +- ++rl_games.env_eval.mid_train.num_episodes=5 +- ++rl_games.env_eval.mid_train.max_steps_per_episode=3600 +- ++rl_games.env_eval.post_train.enabled=false +- ++rl_games.env_eval.post_train.latencies=[0,1,2,3,4] +- ++rl_games.env_eval.post_train.num_episodes=5 +- ++rl_games.env_eval.post_train.max_steps_per_episode=3600 +- ++rl_games.task=flappy +- ++rl_games.initialization_mode=bridge +- ++rl_games.action_carrier=bridge +- ++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct +- ++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action +- ++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct +- ++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct +- ++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action +- ++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++checkpoint.load=none +- ++checkpoint.hf_repo_id=null +- ++checkpoint.save_best_model=false +- ++checkpoint.save_pt_file=false +- ++checkpoint.local.keep_last_n=1 +- ++checkpoint.sync.enabled=false +- ++checkpoint.sync.repo_id=null +- ++checkpoint.sync.keep_last_n=0 +- ++checkpoint.sync.sync_every_n_checkpoints=1 +- ++checkpoint.sync.resume_policy=local_latest +- ++run_id=flappy_fix_latency_2_200ep_last_8_layers_corrected_v2 +- ++output_dir=null +- ++config_yaml=null +- ++is_debug=false +- ++version_id=0.21 +- ++run_root_dir=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints +- ++trainer.is_resume=false +- ++trainer.pretrained_checkpoint=/workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +- ++trainer.resume_step=0 +- ++datasets.vla_data.data_root_dir=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep +- ++datasets.vla_data.data_mix=flappy_train__bridge +- ++datasets.vla_data.eval_data_mix=flappy_train__bridge__val +- ++framework.qwenvl.base_vlm=/workspace/latency-sensitive-bench/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct +- ++rl_games.env_eval.latency.prompt_map_path=/workspace/latency-sensitive-bench/starVLA/data/flappy_fix_latency_2_200ep/flappy_train__bridge/latency_prompt_map.json diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra/train_starvla_hydra.log b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra/train_starvla_hydra.log new file mode 100644 index 0000000000000000000000000000000000000000..2a1f55be9d75168ccd85810cc1a60db72afb23fa --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/hydra/train_starvla_hydra.log @@ -0,0 +1,5376 @@ +[2026-06-23 16:43:43,090][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21' +[2026-06-23 16:43:43,090][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up +[2026-06-23 16:43:43,402][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured. +[2026-06-23 16:43:43,402][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub +[2026-06-23 16:43:43,689][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases +[2026-06-23 16:43:43,696][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking +[2026-06-23 16:43:46,959][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +[2026-06-23 16:43:46,983][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge` +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge'] +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin'] +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): [] +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain'] +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101'] +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): [] +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick'] +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka'] +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): [] +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot'] +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge'] +[2026-06-23 16:43:47,371][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): [] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): [] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): [] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): [] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): [] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor'] +[2026-06-23 16:43:47,372][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge'] +[2026-06-23 16:43:47,392][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val` +[2026-06-23 16:43:47,409][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/config.full.yaml +[2026-06-23 16:43:47,414][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/config.yaml +[2026-06-23 16:43:47,414][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch. +[2026-06-23 16:43:49,511][starVLA.training.train_starvla][INFO] - ***** Training Configuration ***** +[2026-06-23 16:43:49,512][starVLA.training.train_starvla][INFO] - Total optimization steps = 5000 +[2026-06-23 16:43:49,512][starVLA.training.train_starvla][INFO] - Per device batch size = 64 +[2026-06-23 16:43:49,512][starVLA.training.train_starvla][INFO] - Gradient accumulation steps = 2 +[2026-06-23 16:43:49,512][starVLA.training.train_starvla][INFO] - Total batch size = 128 +[2026-06-23 16:44:16,353][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6946540474891663, 'train/grad_norm_pre_clip': 8.944799423217773, 'timing/data': 0.0011549401096999645, 'timing/model': 1.4546807589940727, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0}) +[2026-06-23 16:44:19,282][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6940727829933167, 'train/grad_norm_pre_clip': 10.536144256591797, 'timing/data': 0.0012592158745974302, 'timing/model': 1.4585366111714393, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0}) +[2026-06-23 16:44:22,143][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6895287036895752, 'train/grad_norm_pre_clip': 19.43878746032715, 'timing/data': 0.0018488629721105099, 'timing/model': 1.4639013288542628, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0}) +[2026-06-23 16:44:25,074][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.6881852149963379, 'train/grad_norm_pre_clip': 10.03441333770752, 'timing/data': 0.0013697331305593252, 'timing/model': 1.498638565884903, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 16:44:27,927][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.689396858215332, 'train/grad_norm_pre_clip': 2.3071377277374268, 'timing/data': 0.0014562958385795355, 'timing/model': 1.4347805241122842, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0}) +[2026-06-23 16:44:30,806][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.652509331703186, 'train/grad_norm_pre_clip': 5.957568645477295, 'timing/data': 0.0014736989978700876, 'timing/model': 1.4573997349943966, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0}) +[2026-06-23 16:44:33,502][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.7139065861701965, 'train/grad_norm_pre_clip': 2.2412307262420654, 'timing/data': 0.0012994110584259033, 'timing/model': 1.3305918071419, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 16:44:36,093][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.6930562257766724, 'train/grad_norm_pre_clip': 4.663080215454102, 'timing/data': 0.000816793879494071, 'timing/model': 1.3193062858190387, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0}) +[2026-06-23 16:44:38,685][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.6616442203521729, 'train/grad_norm_pre_clip': 9.361183166503906, 'timing/data': 0.000630553811788559, 'timing/model': 1.3151041329838336, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0}) +[2026-06-23 16:44:41,275][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.7266539335250854, 'train/grad_norm_pre_clip': 7.538466930389404, 'timing/data': 0.0005347160622477531, 'timing/model': 1.319778606062755, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0}) +[2026-06-23 16:44:43,872][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.6703930497169495, 'train/grad_norm_pre_clip': 7.17041015625, 'timing/data': 0.0005341949872672558, 'timing/model': 1.322575211059302, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0}) +[2026-06-23 16:44:46,469][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.6768441796302795, 'train/grad_norm_pre_clip': 1.9383971691131592, 'timing/data': 0.0005817650817334652, 'timing/model': 1.324222391936928, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0}) +[2026-06-23 16:44:49,072][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.6604847311973572, 'train/grad_norm_pre_clip': 1.0827465057373047, 'timing/data': 0.0007078519556671381, 'timing/model': 1.3234827660489827, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.01}) +[2026-06-23 16:44:51,681][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.6503238677978516, 'train/grad_norm_pre_clip': 1.0727559328079224, 'timing/data': 0.0007055969908833504, 'timing/model': 1.326461432967335, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.01}) +[2026-06-23 16:44:54,296][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.659670889377594, 'train/grad_norm_pre_clip': 1.8442572355270386, 'timing/data': 0.00081488280557096, 'timing/model': 1.3280773849692196, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.01}) +[2026-06-23 16:44:56,929][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.6624654531478882, 'train/grad_norm_pre_clip': 8.280050277709961, 'timing/data': 0.0008823401294648647, 'timing/model': 1.3466648170724511, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 16:44:59,563][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.6822299361228943, 'train/grad_norm_pre_clip': 6.58579683303833, 'timing/data': 0.0006433508824557066, 'timing/model': 1.3364304970018566, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 16:45:02,184][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.713620662689209, 'train/grad_norm_pre_clip': 6.62531852722168, 'timing/data': 0.0005339339841157198, 'timing/model': 1.3332202220335603, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.01}) +[2026-06-23 16:45:04,810][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.674497127532959, 'train/grad_norm_pre_clip': 6.8437981605529785, 'timing/data': 0.00056058494374156, 'timing/model': 1.3415307798422873, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.01}) +[2026-06-23 16:45:07,441][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.6817030906677246, 'train/grad_norm_pre_clip': 5.8223490715026855, 'timing/data': 0.000582926906645298, 'timing/model': 1.3346241300459951, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.01}) +[2026-06-23 16:45:10,074][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.6706715226173401, 'train/grad_norm_pre_clip': 1.7392042875289917, 'timing/data': 0.000555835897102952, 'timing/model': 1.342532241018489, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.01}) +[2026-06-23 16:45:12,700][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.6814320087432861, 'train/grad_norm_pre_clip': 4.6155595779418945, 'timing/data': 0.0005931558553129435, 'timing/model': 1.3332469768356532, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 16:45:15,340][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.6548390984535217, 'train/grad_norm_pre_clip': 5.191258907318115, 'timing/data': 0.0006514850538223982, 'timing/model': 1.3393055270425975, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 16:45:17,976][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.6765066385269165, 'train/grad_norm_pre_clip': 2.1586833000183105, 'timing/data': 0.0008173380047082901, 'timing/model': 1.3385534880217165, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.01}) +[2026-06-23 16:45:20,612][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.6529338955879211, 'train/grad_norm_pre_clip': 3.038203001022339, 'timing/data': 0.0008581050205975771, 'timing/model': 1.3379873461090028, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.01}) +[2026-06-23 16:45:23,242][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.6914294362068176, 'train/grad_norm_pre_clip': 10.114646911621094, 'timing/data': 0.0005505559965968132, 'timing/model': 1.336519249016419, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.01}) +[2026-06-23 16:45:25,885][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.716418981552124, 'train/grad_norm_pre_clip': 5.166629314422607, 'timing/data': 0.0005603840108960867, 'timing/model': 1.3384538320824504, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.01}) +[2026-06-23 16:45:28,524][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.6741432547569275, 'train/grad_norm_pre_clip': 3.1486878395080566, 'timing/data': 0.0005686089862138033, 'timing/model': 1.3376360170077533, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01}) +[2026-06-23 16:45:31,170][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.6613386869430542, 'train/grad_norm_pre_clip': 0.8417428731918335, 'timing/data': 0.0006681869272142649, 'timing/model': 1.3489888680633157, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01}) +[2026-06-23 16:45:33,826][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.6563680171966553, 'train/grad_norm_pre_clip': 7.7472686767578125, 'timing/data': 0.0005729079712182283, 'timing/model': 1.3502338449470699, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01}) +[2026-06-23 16:45:36,470][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.6689667701721191, 'train/grad_norm_pre_clip': 6.767736434936523, 'timing/data': 0.0005749310366809368, 'timing/model': 1.3441684788558632, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01}) +[2026-06-23 16:45:39,121][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.6207846403121948, 'train/grad_norm_pre_clip': 10.72819995880127, 'timing/data': 0.0008141819853335619, 'timing/model': 1.349737510085106, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01}) +[2026-06-23 16:45:41,766][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.702960193157196, 'train/grad_norm_pre_clip': 2.8238024711608887, 'timing/data': 0.0005802910309284925, 'timing/model': 1.3505858399439603, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01}) +[2026-06-23 16:45:44,405][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.5985056757926941, 'train/grad_norm_pre_clip': 2.010753631591797, 'timing/data': 0.000574630917981267, 'timing/model': 1.3419133711140603, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01}) +[2026-06-23 16:45:47,053][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.6703206896781921, 'train/grad_norm_pre_clip': 1.418626070022583, 'timing/data': 0.0005607251077890396, 'timing/model': 1.3414890579879284, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01}) +[2026-06-23 16:45:49,701][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.6649979948997498, 'train/grad_norm_pre_clip': 2.1467607021331787, 'timing/data': 0.0005739799235016108, 'timing/model': 1.339847293915227, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01}) +[2026-06-23 16:45:52,350][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.6539419889450073, 'train/grad_norm_pre_clip': 6.719210624694824, 'timing/data': 0.0005582298617810011, 'timing/model': 1.3494116000365466, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.01}) +[2026-06-23 16:45:54,996][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.6382440328598022, 'train/grad_norm_pre_clip': 8.675786972045898, 'timing/data': 0.0005852410104125738, 'timing/model': 1.3446721299551427, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.01}) +[2026-06-23 16:45:57,645][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.6550688743591309, 'train/grad_norm_pre_clip': 8.207502365112305, 'timing/data': 0.0005668771918863058, 'timing/model': 1.3450536418240517, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 16:46:00,294][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.6552539467811584, 'train/grad_norm_pre_clip': 8.195706367492676, 'timing/data': 0.0005614960100501776, 'timing/model': 1.3449240839108825, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.02}) +[2026-06-23 16:46:02,947][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.6183385848999023, 'train/grad_norm_pre_clip': 3.3081440925598145, 'timing/data': 0.0006958788726478815, 'timing/model': 1.352912544971332, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.02}) +[2026-06-23 16:46:05,784][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.7020115256309509, 'train/grad_norm_pre_clip': 14.433232307434082, 'timing/data': 0.0005721359048038721, 'timing/model': 1.5254486969206482, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.02}) +[2026-06-23 16:46:08,424][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.6902675628662109, 'train/grad_norm_pre_clip': 21.47605323791504, 'timing/data': 0.0005605439655482769, 'timing/model': 1.3425584388896823, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.02}) +[2026-06-23 16:46:11,081][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.6167827844619751, 'train/grad_norm_pre_clip': 16.604612350463867, 'timing/data': 0.0008629739750176668, 'timing/model': 1.3484450969845057, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 16:46:13,727][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.6693941950798035, 'train/grad_norm_pre_clip': 15.571868896484375, 'timing/data': 0.0005709240213036537, 'timing/model': 1.3457149860914797, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.02}) +[2026-06-23 16:46:16,378][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.6908186674118042, 'train/grad_norm_pre_clip': 16.74713897705078, 'timing/data': 0.000591181917116046, 'timing/model': 1.3473922871053219, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.02}) +[2026-06-23 16:46:19,035][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.6041165590286255, 'train/grad_norm_pre_clip': 6.312260627746582, 'timing/data': 0.0006182121578603983, 'timing/model': 1.3506839440669864, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.02}) +[2026-06-23 16:46:21,689][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.6386321187019348, 'train/grad_norm_pre_clip': 3.8895490169525146, 'timing/data': 0.0005559148266911507, 'timing/model': 1.346560524078086, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.02}) +[2026-06-23 16:46:24,345][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.6933586001396179, 'train/grad_norm_pre_clip': 18.233028411865234, 'timing/data': 0.000732076819986105, 'timing/model': 1.3515127338469028, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.02}) +[2026-06-23 16:46:27,007][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.7027854919433594, 'train/grad_norm_pre_clip': 23.827285766601562, 'timing/data': 0.0007213270291686058, 'timing/model': 1.3528295068535954, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.02}) +[2026-06-23 16:46:29,654][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.6938318610191345, 'train/grad_norm_pre_clip': 21.9443302154541, 'timing/data': 0.0006361869163811207, 'timing/model': 1.3485754509456456, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 16:46:32,312][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.7063785195350647, 'train/grad_norm_pre_clip': 27.157258987426758, 'timing/data': 0.0005676171276718378, 'timing/model': 1.3569777000229806, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.02}) +[2026-06-23 16:46:34,979][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.6096354126930237, 'train/grad_norm_pre_clip': 5.422735214233398, 'timing/data': 0.0005692411214113235, 'timing/model': 1.35624573007226, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.02}) +[2026-06-23 16:46:37,647][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.6453347206115723, 'train/grad_norm_pre_clip': 9.387989044189453, 'timing/data': 0.0005870540626347065, 'timing/model': 1.3560210540890694, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.02}) +[2026-06-23 16:46:40,307][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.5724336504936218, 'train/grad_norm_pre_clip': 13.775856971740723, 'timing/data': 0.0005820239894092083, 'timing/model': 1.3489938008133322, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.02}) +[2026-06-23 16:46:42,960][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.5751225352287292, 'train/grad_norm_pre_clip': 10.181815147399902, 'timing/data': 0.0006312571931630373, 'timing/model': 1.3466729340143502, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.02}) +[2026-06-23 16:46:45,609][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.6130154728889465, 'train/grad_norm_pre_clip': 9.52898120880127, 'timing/data': 0.0005826160777360201, 'timing/model': 1.347591697005555, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.02}) +[2026-06-23 16:46:48,267][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.622344434261322, 'train/grad_norm_pre_clip': 14.535088539123535, 'timing/data': 0.0009293670300394297, 'timing/model': 1.3541044329758734, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.02}) +[2026-06-23 16:46:50,928][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.6116622686386108, 'train/grad_norm_pre_clip': 10.353826522827148, 'timing/data': 0.0005877451039850712, 'timing/model': 1.3514383661095053, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.02}) +[2026-06-23 16:46:53,586][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.5987502932548523, 'train/grad_norm_pre_clip': 20.907289505004883, 'timing/data': 0.0005835168994963169, 'timing/model': 1.3566193650476635, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.02}) +[2026-06-23 16:46:56,242][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.5866196155548096, 'train/grad_norm_pre_clip': 14.134398460388184, 'timing/data': 0.0005871239118278027, 'timing/model': 1.3492486800532788, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.02}) +[2026-06-23 16:46:58,907][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.5862785577774048, 'train/grad_norm_pre_clip': 7.2400336265563965, 'timing/data': 0.0006042369641363621, 'timing/model': 1.3489330229349434, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.02}) +[2026-06-23 16:47:01,559][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.6358269453048706, 'train/grad_norm_pre_clip': 5.422037124633789, 'timing/data': 0.000576073070988059, 'timing/model': 1.3498857200611383, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.02}) +[2026-06-23 16:47:04,212][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.5847740173339844, 'train/grad_norm_pre_clip': 6.435020923614502, 'timing/data': 0.0005826458800584078, 'timing/model': 1.3470027458388358, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.02}) +[2026-06-23 16:47:06,866][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.542911946773529, 'train/grad_norm_pre_clip': 9.946389198303223, 'timing/data': 0.0005700720939785242, 'timing/model': 1.346213968936354, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:47:09,523][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.4946267902851105, 'train/grad_norm_pre_clip': 4.792409420013428, 'timing/data': 0.0005916832014918327, 'timing/model': 1.3538933990057558, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.03}) +[2026-06-23 16:47:12,187][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.5845834016799927, 'train/grad_norm_pre_clip': 3.7576239109039307, 'timing/data': 0.0006803488358855247, 'timing/model': 1.3585225308779627, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.03}) +[2026-06-23 16:47:14,850][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.5939421057701111, 'train/grad_norm_pre_clip': 19.274826049804688, 'timing/data': 0.0005788980051875114, 'timing/model': 1.3507180311717093, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:47:17,517][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.4834255278110504, 'train/grad_norm_pre_clip': 9.12605094909668, 'timing/data': 0.0005863229744136333, 'timing/model': 1.3492149310186505, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.03}) +[2026-06-23 16:47:20,175][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.5391644239425659, 'train/grad_norm_pre_clip': 29.966720581054688, 'timing/data': 0.0007321869488805532, 'timing/model': 1.3493264578282833, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.03}) +[2026-06-23 16:47:22,841][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.5738831758499146, 'train/grad_norm_pre_clip': 31.82111930847168, 'timing/data': 0.0005952699575573206, 'timing/model': 1.3514064271003008, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.03}) +[2026-06-23 16:47:25,504][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.6540469527244568, 'train/grad_norm_pre_clip': 44.189788818359375, 'timing/data': 0.0005997479893267155, 'timing/model': 1.3520077548455447, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.03}) +[2026-06-23 16:47:28,168][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.6235831379890442, 'train/grad_norm_pre_clip': 36.0027961730957, 'timing/data': 0.0005796200130134821, 'timing/model': 1.3583394300658256, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.03}) +[2026-06-23 16:47:30,825][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.5563105344772339, 'train/grad_norm_pre_clip': 8.5347318649292, 'timing/data': 0.000582836102694273, 'timing/model': 1.3519606119953096, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.03}) +[2026-06-23 16:47:33,500][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.49428611993789673, 'train/grad_norm_pre_clip': 13.068419456481934, 'timing/data': 0.0005873339250683784, 'timing/model': 1.3560356721282005, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:47:36,163][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.5417532920837402, 'train/grad_norm_pre_clip': 48.841312408447266, 'timing/data': 0.0006022921297699213, 'timing/model': 1.3531501430552453, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.03}) +[2026-06-23 16:47:38,824][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.6497818827629089, 'train/grad_norm_pre_clip': 35.3759651184082, 'timing/data': 0.0006000879220664501, 'timing/model': 1.3547100538853556, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.03}) +[2026-06-23 16:47:41,488][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.6806208491325378, 'train/grad_norm_pre_clip': 29.01898193359375, 'timing/data': 0.000583417946472764, 'timing/model': 1.3480349578894675, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:47:44,155][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.5174652338027954, 'train/grad_norm_pre_clip': 4.511693477630615, 'timing/data': 0.0005753419827669859, 'timing/model': 1.359418093925342, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:47:46,823][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.5539789795875549, 'train/grad_norm_pre_clip': 3.957109212875366, 'timing/data': 0.0005909418687224388, 'timing/model': 1.3582010730169713, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.03}) +[2026-06-23 16:47:49,491][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.6633908152580261, 'train/grad_norm_pre_clip': 29.384519577026367, 'timing/data': 0.0005916131194680929, 'timing/model': 1.3585701638367027, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.03}) +[2026-06-23 16:47:52,161][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.6779143214225769, 'train/grad_norm_pre_clip': 39.812408447265625, 'timing/data': 0.0006781648844480515, 'timing/model': 1.3566292421892285, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.03}) +[2026-06-23 16:47:54,826][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.5073155164718628, 'train/grad_norm_pre_clip': 30.260696411132812, 'timing/data': 0.0005947891622781754, 'timing/model': 1.3533662729896605, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.03}) +[2026-06-23 16:47:57,481][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.4788665771484375, 'train/grad_norm_pre_clip': 3.4963717460632324, 'timing/data': 0.0005734178703278303, 'timing/model': 1.3508312841877341, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.03}) +[2026-06-23 16:48:00,148][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.6096518039703369, 'train/grad_norm_pre_clip': 36.26371383666992, 'timing/data': 0.0005804509855806828, 'timing/model': 1.3566818740218878, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.03}) +[2026-06-23 16:48:02,805][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.5422857999801636, 'train/grad_norm_pre_clip': 38.99480438232422, 'timing/data': 0.0005778470076620579, 'timing/model': 1.3453462661709636, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.03}) +[2026-06-23 16:48:05,469][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.6743289232254028, 'train/grad_norm_pre_clip': 36.89372253417969, 'timing/data': 0.0006354751531034708, 'timing/model': 1.3540226600598544, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.03}) +[2026-06-23 16:48:08,145][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.6124274730682373, 'train/grad_norm_pre_clip': 22.942546844482422, 'timing/data': 0.0005949190817773342, 'timing/model': 1.3625057409517467, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:48:10,817][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.42956554889678955, 'train/grad_norm_pre_clip': 12.303391456604004, 'timing/data': 0.0005772949662059546, 'timing/model': 1.357741178944707, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.03}) +[2026-06-23 16:48:13,495][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.5861825346946716, 'train/grad_norm_pre_clip': 37.729835510253906, 'timing/data': 0.0007153148762881756, 'timing/model': 1.3596768868155777, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.03}) +[2026-06-23 16:48:16,168][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.7312598824501038, 'train/grad_norm_pre_clip': 65.35990905761719, 'timing/data': 0.0005894589703530073, 'timing/model': 1.3628767838235945, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.04}) +[2026-06-23 16:48:18,849][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.7090097665786743, 'train/grad_norm_pre_clip': 41.14718246459961, 'timing/data': 0.0005873849149793386, 'timing/model': 1.3596063419245183, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.04}) +[2026-06-23 16:48:21,511][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.5249216556549072, 'train/grad_norm_pre_clip': 52.111534118652344, 'timing/data': 0.0005845790728926659, 'timing/model': 1.3499943551141769, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.04}) +[2026-06-23 16:48:24,181][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.48033127188682556, 'train/grad_norm_pre_clip': 32.81915283203125, 'timing/data': 0.0005752709694206715, 'timing/model': 1.3523003198206425, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.04}) +[2026-06-23 16:48:26,848][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.47512608766555786, 'train/grad_norm_pre_clip': 4.277822494506836, 'timing/data': 0.000706859864294529, 'timing/model': 1.3557606060057878, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.04}) +[2026-06-23 16:48:29,522][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.48802921175956726, 'train/grad_norm_pre_clip': 49.6537971496582, 'timing/data': 0.0005938569083809853, 'timing/model': 1.3584844158031046, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.04}) +[2026-06-23 16:48:32,188][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.5471540689468384, 'train/grad_norm_pre_clip': 65.34596252441406, 'timing/data': 0.0005931758787482977, 'timing/model': 1.3575387871824205, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.04}) +[2026-06-23 16:48:34,858][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.8750532865524292, 'train/grad_norm_pre_clip': 86.7702407836914, 'timing/data': 0.0008160751312971115, 'timing/model': 1.3586335750296712, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.04}) +[2026-06-23 16:48:37,526][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.9223215579986572, 'train/grad_norm_pre_clip': 85.44603729248047, 'timing/data': 0.0005904501304030418, 'timing/model': 1.358938517048955, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.04}) +[2026-06-23 16:48:40,197][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 100 +[2026-06-23 16:50:33,410][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 100 +[2026-06-23 16:50:33,411][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 100 +[2026-06-23 16:50:34,025][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmp476gf59r/test.c -o /tmp/tmp476gf59r/test.o +[2026-06-23 16:50:34,040][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmp476gf59r/test.o -laio -o /tmp/tmp476gf59r/a.out +[2026-06-23 16:50:35,533][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmpeuqn8enb/test.c -o /tmp/tmpeuqn8enb/test.o +[2026-06-23 16:50:35,552][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmpeuqn8enb/test.o -L/usr/local/cuda -L/usr/local/cuda/lib64 -lcufile -o /tmp/tmpeuqn8enb/a.out +[2026-06-23 16:50:35,597][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat -Wno-unused-result -Wsign-compare -DNDEBUG -fwrapv -O2 -Wall -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -fno-merge-constants -O2 -isystem /venv/starvla_rl_games_openvla/include -fPIC -c /tmp/tmp_3qfanrc/test.c -o /tmp/tmp_3qfanrc/test.o +[2026-06-23 16:50:35,613][root][INFO] - gcc -pthread -B /venv/starvla_rl_games_openvla/compiler_compat /tmp/tmp_3qfanrc/test.o -laio -o /tmp/tmp_3qfanrc/a.out +[2026-06-23 16:52:10,974][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 100 +[2026-06-23 16:52:10,975][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.5830697417259216, 'train/grad_norm_pre_clip': 54.80833435058594, 'eval/action_loss/seconds': 113.21036803093739, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.33306655287742615, 'eval/latency_2/loss': 0.33306655287742615, 'eval/flappy/loss': 0.33306655287742615, 'eval/flappy/latency_2/loss': 0.33306655287742615, 'eval/flap_precision': 0.5826161790017211, 'eval/flap_recall': 0.9136302294197031, 'eval/flap_f1': 0.7115081450341565, 'eval/flap_rate_model': 0.3227777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.7115081450341565, 'eval/latency_2/flap_precision': 0.5826161790017211, 'eval/latency_2/flap_recall': 0.9136302294197031, 'eval/latency_2/flap_f1': 0.7115081450341565, 'eval/latency_2/flap_rate_model': 0.3227777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.7115081450341565, 'eval/action_classification/data_wait_seconds': 28.344980239868164, 'eval/action_classification/predict_seconds': 65.62167358398438, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0007432270795106888, 'timing/model': 1.3575801611877978, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.04}) +[2026-06-23 16:52:14,303][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.45635536313056946, 'train/grad_norm_pre_clip': 18.54092788696289, 'timing/data': 0.0005782770458608866, 'timing/model': 1.6858711778186262, 'learning_rate/qwen_vl_interface': 9.999999023727212e-06, 'learning_rate/action_model': 9.99999902372721e-05, 'epoch': 0.04}) +[2026-06-23 16:52:17,629][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.47199034690856934, 'train/grad_norm_pre_clip': 21.553056716918945, 'timing/data': 0.0005775860045105219, 'timing/model': 1.6905123121105134, 'learning_rate/qwen_vl_interface': 9.999996094909247e-06, 'learning_rate/action_model': 9.999996094909248e-05, 'epoch': 0.04}) +[2026-06-23 16:52:20,959][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.6405388116836548, 'train/grad_norm_pre_clip': 57.811832427978516, 'timing/data': 0.0005837469361722469, 'timing/model': 1.6949589799623936, 'learning_rate/qwen_vl_interface': 9.999991213547314e-06, 'learning_rate/action_model': 9.999991213547313e-05, 'epoch': 0.04}) +[2026-06-23 16:52:24,585][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.7521196603775024, 'train/grad_norm_pre_clip': 49.30788040161133, 'timing/data': 0.0008023988921195269, 'timing/model': 1.6869054869748652, 'learning_rate/qwen_vl_interface': 9.999984379643413e-06, 'learning_rate/action_model': 9.999984379643412e-05, 'epoch': 0.04}) +[2026-06-23 16:52:27,912][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.4520431160926819, 'train/grad_norm_pre_clip': 34.657623291015625, 'timing/data': 0.0005803098902106285, 'timing/model': 1.6905347688589245, 'learning_rate/qwen_vl_interface': 9.999975593200357e-06, 'learning_rate/action_model': 9.999975593200357e-05, 'epoch': 0.04}) +[2026-06-23 16:52:31,244][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.5293540954589844, 'train/grad_norm_pre_clip': 36.62673568725586, 'timing/data': 0.0005826358683407307, 'timing/model': 1.6922907088883221, 'learning_rate/qwen_vl_interface': 9.999964854221756e-06, 'learning_rate/action_model': 9.999964854221754e-05, 'epoch': 0.04}) +[2026-06-23 16:52:34,570][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.4498567283153534, 'train/grad_norm_pre_clip': 3.7275068759918213, 'timing/data': 0.000623973086476326, 'timing/model': 1.69224590389058, 'learning_rate/qwen_vl_interface': 9.999952162712027e-06, 'learning_rate/action_model': 9.999952162712026e-05, 'epoch': 0.04}) +[2026-06-23 16:52:37,903][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.37782129645347595, 'train/grad_norm_pre_clip': 21.786725997924805, 'timing/data': 0.0006441909354180098, 'timing/model': 1.6934373041149229, 'learning_rate/qwen_vl_interface': 9.999937518676382e-06, 'learning_rate/action_model': 9.999937518676382e-05, 'epoch': 0.04}) +[2026-06-23 16:52:41,248][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.5951536297798157, 'train/grad_norm_pre_clip': 32.848411560058594, 'timing/data': 0.0006701699458062649, 'timing/model': 1.6956177209503949, 'learning_rate/qwen_vl_interface': 9.999920922120846e-06, 'learning_rate/action_model': 9.999920922120846e-05, 'epoch': 0.04}) +[2026-06-23 16:52:44,593][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.6307675242424011, 'train/grad_norm_pre_clip': 28.60017204284668, 'timing/data': 0.0005825748667120934, 'timing/model': 1.6974669061601162, 'learning_rate/qwen_vl_interface': 9.999902373052239e-06, 'learning_rate/action_model': 9.999902373052239e-05, 'epoch': 0.04}) +[2026-06-23 16:52:47,927][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.48019173741340637, 'train/grad_norm_pre_clip': 10.886971473693848, 'timing/data': 0.0006435401737689972, 'timing/model': 1.691924607846886, 'learning_rate/qwen_vl_interface': 9.999881871478188e-06, 'learning_rate/action_model': 9.999881871478187e-05, 'epoch': 0.04}) +[2026-06-23 16:52:51,262][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.4965249300003052, 'train/grad_norm_pre_clip': 18.067615509033203, 'timing/data': 0.0006473169196397066, 'timing/model': 1.69474172196351, 'learning_rate/qwen_vl_interface': 9.999859417407115e-06, 'learning_rate/action_model': 9.999859417407115e-05, 'epoch': 0.04}) +[2026-06-23 16:52:54,597][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.5789744257926941, 'train/grad_norm_pre_clip': 26.30364990234375, 'timing/data': 0.0006733760237693787, 'timing/model': 1.6927685730624944, 'learning_rate/qwen_vl_interface': 9.999835010848254e-06, 'learning_rate/action_model': 9.999835010848254e-05, 'epoch': 0.04}) +[2026-06-23 16:52:57,942][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.5090628862380981, 'train/grad_norm_pre_clip': 23.329626083374023, 'timing/data': 0.0005797301419079304, 'timing/model': 1.7015072589274496, 'learning_rate/qwen_vl_interface': 9.999808651811639e-06, 'learning_rate/action_model': 9.999808651811639e-05, 'epoch': 0.04}) +[2026-06-23 16:53:01,281][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.6013479232788086, 'train/grad_norm_pre_clip': 20.58997917175293, 'timing/data': 0.0005952990613877773, 'timing/model': 1.6953118219971657, 'learning_rate/qwen_vl_interface': 9.999780340308101e-06, 'learning_rate/action_model': 9.999780340308101e-05, 'epoch': 0.04}) +[2026-06-23 16:53:04,622][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.4867856800556183, 'train/grad_norm_pre_clip': 11.475362777709961, 'timing/data': 0.0005782169755548239, 'timing/model': 1.6989853181876242, 'learning_rate/qwen_vl_interface': 9.999750076349282e-06, 'learning_rate/action_model': 9.999750076349282e-05, 'epoch': 0.04}) +[2026-06-23 16:53:07,963][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.5139932632446289, 'train/grad_norm_pre_clip': 26.37379264831543, 'timing/data': 0.0006225300021469593, 'timing/model': 1.6955298918765038, 'learning_rate/qwen_vl_interface': 9.999717859947618e-06, 'learning_rate/action_model': 9.999717859947618e-05, 'epoch': 0.05}) +[2026-06-23 16:53:11,438][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.3948957026004791, 'train/grad_norm_pre_clip': 23.212238311767578, 'timing/data': 0.0006754791829735041, 'timing/model': 1.693997299997136, 'learning_rate/qwen_vl_interface': 9.999683691116356e-06, 'learning_rate/action_model': 9.999683691116355e-05, 'epoch': 0.05}) +[2026-06-23 16:53:14,779][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.47556060552597046, 'train/grad_norm_pre_clip': 3.6743409633636475, 'timing/data': 0.0005910210311412811, 'timing/model': 1.6969553898088634, 'learning_rate/qwen_vl_interface': 9.999647569869539e-06, 'learning_rate/action_model': 9.99964756986954e-05, 'epoch': 0.05}) +[2026-06-23 16:53:18,131][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.4994131326675415, 'train/grad_norm_pre_clip': 27.93500518798828, 'timing/data': 0.0005857308860868216, 'timing/model': 1.6971689730416983, 'learning_rate/qwen_vl_interface': 9.999609496222017e-06, 'learning_rate/action_model': 9.999609496222017e-05, 'epoch': 0.05}) +[2026-06-23 16:53:21,465][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.44800037145614624, 'train/grad_norm_pre_clip': 11.135580062866211, 'timing/data': 0.0005925740115344524, 'timing/model': 1.6929845861159265, 'learning_rate/qwen_vl_interface': 9.99956947018944e-06, 'learning_rate/action_model': 9.999569470189439e-05, 'epoch': 0.05}) +[2026-06-23 16:53:24,798][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.4422897398471832, 'train/grad_norm_pre_clip': 23.856616973876953, 'timing/data': 0.0005792179144918919, 'timing/model': 1.69263970782049, 'learning_rate/qwen_vl_interface': 9.99952749178826e-06, 'learning_rate/action_model': 9.999527491788259e-05, 'epoch': 0.05}) +[2026-06-23 16:53:28,133][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.5197353959083557, 'train/grad_norm_pre_clip': 31.29507827758789, 'timing/data': 0.0005776658654212952, 'timing/model': 1.694458135869354, 'learning_rate/qwen_vl_interface': 9.999483561035732e-06, 'learning_rate/action_model': 9.999483561035732e-05, 'epoch': 0.05}) +[2026-06-23 16:53:31,479][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.6023615002632141, 'train/grad_norm_pre_clip': 40.3929328918457, 'timing/data': 0.0006409150082617998, 'timing/model': 1.7051570878829807, 'learning_rate/qwen_vl_interface': 9.999437677949917e-06, 'learning_rate/action_model': 9.999437677949917e-05, 'epoch': 0.05}) +[2026-06-23 16:53:34,810][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.5639808773994446, 'train/grad_norm_pre_clip': 38.604339599609375, 'timing/data': 0.0005795599427074194, 'timing/model': 1.692191496025771, 'learning_rate/qwen_vl_interface': 9.999389842549674e-06, 'learning_rate/action_model': 9.999389842549673e-05, 'epoch': 0.05}) +[2026-06-23 16:53:38,141][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.5024905800819397, 'train/grad_norm_pre_clip': 11.188138961791992, 'timing/data': 0.000656444113701582, 'timing/model': 1.6936902338638902, 'learning_rate/qwen_vl_interface': 9.999340054854669e-06, 'learning_rate/action_model': 9.999340054854667e-05, 'epoch': 0.05}) +[2026-06-23 16:53:41,485][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.45026013255119324, 'train/grad_norm_pre_clip': 35.27657699584961, 'timing/data': 0.0006648199632763863, 'timing/model': 1.702539738966152, 'learning_rate/qwen_vl_interface': 9.999288314885362e-06, 'learning_rate/action_model': 9.999288314885362e-05, 'epoch': 0.05}) +[2026-06-23 16:53:44,835][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.45577821135520935, 'train/grad_norm_pre_clip': 24.544275283813477, 'timing/data': 0.0006003580056130886, 'timing/model': 1.697208276949823, 'learning_rate/qwen_vl_interface': 9.999234622663027e-06, 'learning_rate/action_model': 9.999234622663028e-05, 'epoch': 0.05}) +[2026-06-23 16:53:48,179][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.49423977732658386, 'train/grad_norm_pre_clip': 23.43486213684082, 'timing/data': 0.0005946280434727669, 'timing/model': 1.6981415620539337, 'learning_rate/qwen_vl_interface': 9.999178978209732e-06, 'learning_rate/action_model': 9.999178978209732e-05, 'epoch': 0.05}) +[2026-06-23 16:53:51,531][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.3925235867500305, 'train/grad_norm_pre_clip': 21.206249237060547, 'timing/data': 0.0005808519199490547, 'timing/model': 1.6998645409476012, 'learning_rate/qwen_vl_interface': 9.999121381548354e-06, 'learning_rate/action_model': 9.999121381548353e-05, 'epoch': 0.05}) +[2026-06-23 16:53:54,874][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.4642484784126282, 'train/grad_norm_pre_clip': 7.948774337768555, 'timing/data': 0.0008121959399431944, 'timing/model': 1.6987735009752214, 'learning_rate/qwen_vl_interface': 9.999061832702563e-06, 'learning_rate/action_model': 9.999061832702563e-05, 'epoch': 0.05}) +[2026-06-23 16:53:58,332][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.4165070652961731, 'train/grad_norm_pre_clip': 5.516689300537109, 'timing/data': 0.000600739847868681, 'timing/model': 1.8150670190807432, 'learning_rate/qwen_vl_interface': 9.999000331696842e-06, 'learning_rate/action_model': 9.999000331696841e-05, 'epoch': 0.05}) +[2026-06-23 16:54:01,670][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.4594925343990326, 'train/grad_norm_pre_clip': 6.171248435974121, 'timing/data': 0.0005781468935310841, 'timing/model': 1.6965885830577463, 'learning_rate/qwen_vl_interface': 9.99893687855647e-06, 'learning_rate/action_model': 9.99893687855647e-05, 'epoch': 0.05}) +[2026-06-23 16:54:05,015][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.36182788014411926, 'train/grad_norm_pre_clip': 18.04545021057129, 'timing/data': 0.0008995609823614359, 'timing/model': 1.7002946028951555, 'learning_rate/qwen_vl_interface': 9.998871473307529e-06, 'learning_rate/action_model': 9.99887147330753e-05, 'epoch': 0.05}) +[2026-06-23 16:54:08,362][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.39813685417175293, 'train/grad_norm_pre_clip': 9.405573844909668, 'timing/data': 0.0005720250774174929, 'timing/model': 1.6966257651802152, 'learning_rate/qwen_vl_interface': 9.998804115976907e-06, 'learning_rate/action_model': 9.998804115976908e-05, 'epoch': 0.05}) +[2026-06-23 16:54:11,709][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.5469050407409668, 'train/grad_norm_pre_clip': 30.210390090942383, 'timing/data': 0.0005824151448905468, 'timing/model': 1.6985881458967924, 'learning_rate/qwen_vl_interface': 9.99873480659229e-06, 'learning_rate/action_model': 9.99873480659229e-05, 'epoch': 0.05}) +[2026-06-23 16:54:15,052][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.4640955626964569, 'train/grad_norm_pre_clip': 18.234243392944336, 'timing/data': 0.0005713740829378366, 'timing/model': 1.696906405966729, 'learning_rate/qwen_vl_interface': 9.99866354518217e-06, 'learning_rate/action_model': 9.99866354518217e-05, 'epoch': 0.05}) +[2026-06-23 16:54:18,400][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.4276215732097626, 'train/grad_norm_pre_clip': 3.6583056449890137, 'timing/data': 0.0005814020987600088, 'timing/model': 1.6972464059945196, 'learning_rate/qwen_vl_interface': 9.99859033177584e-06, 'learning_rate/action_model': 9.99859033177584e-05, 'epoch': 0.05}) +[2026-06-23 16:54:21,749][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.5245540142059326, 'train/grad_norm_pre_clip': 30.14588165283203, 'timing/data': 0.0005945779848843813, 'timing/model': 1.704175831982866, 'learning_rate/qwen_vl_interface': 9.998515166403394e-06, 'learning_rate/action_model': 9.998515166403394e-05, 'epoch': 0.05}) +[2026-06-23 16:54:25,107][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.46754711866378784, 'train/grad_norm_pre_clip': 21.706621170043945, 'timing/data': 0.0006006390321999788, 'timing/model': 1.704370919847861, 'learning_rate/qwen_vl_interface': 9.99843804909573e-06, 'learning_rate/action_model': 9.99843804909573e-05, 'epoch': 0.05}) +[2026-06-23 16:54:28,452][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.4139397144317627, 'train/grad_norm_pre_clip': 26.218128204345703, 'timing/data': 0.0005878049414604902, 'timing/model': 1.7041566888801754, 'learning_rate/qwen_vl_interface': 9.998358979884548e-06, 'learning_rate/action_model': 9.998358979884547e-05, 'epoch': 0.05}) +[2026-06-23 16:54:31,809][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.5477834939956665, 'train/grad_norm_pre_clip': 15.449881553649902, 'timing/data': 0.0005714239086955786, 'timing/model': 1.7030286800581962, 'learning_rate/qwen_vl_interface': 9.998277958802352e-06, 'learning_rate/action_model': 9.998277958802351e-05, 'epoch': 0.05}) +[2026-06-23 16:54:35,152][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.35958990454673767, 'train/grad_norm_pre_clip': 10.507486343383789, 'timing/data': 0.0006192540749907494, 'timing/model': 1.6927396811079234, 'learning_rate/qwen_vl_interface': 9.998194985882442e-06, 'learning_rate/action_model': 9.998194985882442e-05, 'epoch': 0.06}) +[2026-06-23 16:54:38,497][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.38640296459198, 'train/grad_norm_pre_clip': 21.05106544494629, 'timing/data': 0.0005767240654677153, 'timing/model': 1.6992086020763963, 'learning_rate/qwen_vl_interface': 9.99811006115893e-06, 'learning_rate/action_model': 9.998110061158931e-05, 'epoch': 0.06}) +[2026-06-23 16:54:41,845][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.375843346118927, 'train/grad_norm_pre_clip': 29.93423080444336, 'timing/data': 0.0005779271014034748, 'timing/model': 1.702565643005073, 'learning_rate/qwen_vl_interface': 9.998023184666726e-06, 'learning_rate/action_model': 9.998023184666726e-05, 'epoch': 0.06}) +[2026-06-23 16:54:45,203][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.5264608860015869, 'train/grad_norm_pre_clip': 14.745915412902832, 'timing/data': 0.0007767900824546814, 'timing/model': 1.6962698609568179, 'learning_rate/qwen_vl_interface': 9.997934356441538e-06, 'learning_rate/action_model': 9.997934356441537e-05, 'epoch': 0.06}) +[2026-06-23 16:54:48,642][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.3750244677066803, 'train/grad_norm_pre_clip': 21.827220916748047, 'timing/data': 0.0005956690292805433, 'timing/model': 1.6947266510687768, 'learning_rate/qwen_vl_interface': 9.997843576519879e-06, 'learning_rate/action_model': 9.997843576519878e-05, 'epoch': 0.06}) +[2026-06-23 16:54:51,981][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.522350549697876, 'train/grad_norm_pre_clip': 27.70925521850586, 'timing/data': 0.0006542799528688192, 'timing/model': 1.6965370860416442, 'learning_rate/qwen_vl_interface': 9.99775084493907e-06, 'learning_rate/action_model': 9.997750844939069e-05, 'epoch': 0.06}) +[2026-06-23 16:54:55,333][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.5160297155380249, 'train/grad_norm_pre_clip': 43.78647994995117, 'timing/data': 0.0006827139295637608, 'timing/model': 1.6970973100978881, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.06}) +[2026-06-23 16:54:58,692][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.5367130041122437, 'train/grad_norm_pre_clip': 33.95270919799805, 'timing/data': 0.000652015907689929, 'timing/model': 1.7039960389956832, 'learning_rate/qwen_vl_interface': 9.99755952695327e-06, 'learning_rate/action_model': 9.997559526953269e-05, 'epoch': 0.06}) +[2026-06-23 16:55:02,045][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.4169313311576843, 'train/grad_norm_pre_clip': 9.554710388183594, 'timing/data': 0.0005605430342257023, 'timing/model': 1.7027838320937008, 'learning_rate/qwen_vl_interface': 9.99746094062692e-06, 'learning_rate/action_model': 9.99746094062692e-05, 'epoch': 0.06}) +[2026-06-23 16:55:05,406][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.41878268122673035, 'train/grad_norm_pre_clip': 6.644491672515869, 'timing/data': 0.0006195241585373878, 'timing/model': 1.7014261020813137, 'learning_rate/qwen_vl_interface': 9.997360402798706e-06, 'learning_rate/action_model': 9.997360402798705e-05, 'epoch': 0.06}) +[2026-06-23 16:55:08,749][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.4385838210582733, 'train/grad_norm_pre_clip': 20.864452362060547, 'timing/data': 0.0005925639998167753, 'timing/model': 1.698644574964419, 'learning_rate/qwen_vl_interface': 9.997257913509955e-06, 'learning_rate/action_model': 9.997257913509955e-05, 'epoch': 0.06}) +[2026-06-23 16:55:12,095][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.41399088501930237, 'train/grad_norm_pre_clip': 8.547830581665039, 'timing/data': 0.0006630758289247751, 'timing/model': 1.695810270961374, 'learning_rate/qwen_vl_interface': 9.997153472802794e-06, 'learning_rate/action_model': 9.997153472802795e-05, 'epoch': 0.06}) +[2026-06-23 16:55:15,443][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.3397423326969147, 'train/grad_norm_pre_clip': 17.709054946899414, 'timing/data': 0.0005801699589937925, 'timing/model': 1.700514875119552, 'learning_rate/qwen_vl_interface': 9.997047080720157e-06, 'learning_rate/action_model': 9.997047080720158e-05, 'epoch': 0.06}) +[2026-06-23 16:55:18,792][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.3443363606929779, 'train/grad_norm_pre_clip': 12.605485916137695, 'timing/data': 0.000627659959718585, 'timing/model': 1.6995323300361633, 'learning_rate/qwen_vl_interface': 9.996938737305777e-06, 'learning_rate/action_model': 9.996938737305776e-05, 'epoch': 0.06}) +[2026-06-23 16:55:22,139][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.3651037812232971, 'train/grad_norm_pre_clip': 4.028079032897949, 'timing/data': 0.0006886750925332308, 'timing/model': 1.7024876570794731, 'learning_rate/qwen_vl_interface': 9.996828442604189e-06, 'learning_rate/action_model': 9.99682844260419e-05, 'epoch': 0.06}) +[2026-06-23 16:55:25,495][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.35457152128219604, 'train/grad_norm_pre_clip': 23.998647689819336, 'timing/data': 0.0007380279712378979, 'timing/model': 1.7032635868526995, 'learning_rate/qwen_vl_interface': 9.996716196660733e-06, 'learning_rate/action_model': 9.996716196660731e-05, 'epoch': 0.06}) +[2026-06-23 16:55:28,842][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.6538069248199463, 'train/grad_norm_pre_clip': 82.99515533447266, 'timing/data': 0.0005813220050185919, 'timing/model': 1.700770374154672, 'learning_rate/qwen_vl_interface': 9.996601999521546e-06, 'learning_rate/action_model': 9.996601999521546e-05, 'epoch': 0.06}) +[2026-06-23 16:55:32,201][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.5162332653999329, 'train/grad_norm_pre_clip': 52.410003662109375, 'timing/data': 0.0005869630258530378, 'timing/model': 1.7031199608463794, 'learning_rate/qwen_vl_interface': 9.996485851233572e-06, 'learning_rate/action_model': 9.996485851233572e-05, 'epoch': 0.06}) +[2026-06-23 16:55:35,650][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.43281662464141846, 'train/grad_norm_pre_clip': 5.974752426147461, 'timing/data': 0.0005973831284791231, 'timing/model': 1.7908591150771827, 'learning_rate/qwen_vl_interface': 9.996367751844556e-06, 'learning_rate/action_model': 9.996367751844555e-05, 'epoch': 0.06}) +[2026-06-23 16:55:38,995][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.5973133444786072, 'train/grad_norm_pre_clip': 43.05056381225586, 'timing/data': 0.000584098044782877, 'timing/model': 1.6966218629386276, 'learning_rate/qwen_vl_interface': 9.996247701403043e-06, 'learning_rate/action_model': 9.996247701403042e-05, 'epoch': 0.06}) +[2026-06-23 16:55:42,348][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.5521486401557922, 'train/grad_norm_pre_clip': 48.40278625488281, 'timing/data': 0.0006545300129801035, 'timing/model': 1.7016901501920074, 'learning_rate/qwen_vl_interface': 9.99612569995838e-06, 'learning_rate/action_model': 9.99612569995838e-05, 'epoch': 0.06}) +[2026-06-23 16:55:45,705][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.5718597769737244, 'train/grad_norm_pre_clip': 69.02251434326172, 'timing/data': 0.0005742090288549662, 'timing/model': 1.708498280029744, 'learning_rate/qwen_vl_interface': 9.99600174756072e-06, 'learning_rate/action_model': 9.99600174756072e-05, 'epoch': 0.06}) +[2026-06-23 16:55:49,047][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.36879777908325195, 'train/grad_norm_pre_clip': 23.983848571777344, 'timing/data': 0.0006144549697637558, 'timing/model': 1.698559972224757, 'learning_rate/qwen_vl_interface': 9.995875844261012e-06, 'learning_rate/action_model': 9.995875844261013e-05, 'epoch': 0.06}) +[2026-06-23 16:55:52,409][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.41193830966949463, 'train/grad_norm_pre_clip': 30.40804100036621, 'timing/data': 0.0005969530902802944, 'timing/model': 1.7098363691475242, 'learning_rate/qwen_vl_interface': 9.995747990111015e-06, 'learning_rate/action_model': 9.995747990111013e-05, 'epoch': 0.06}) +[2026-06-23 16:55:55,761][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.45219460129737854, 'train/grad_norm_pre_clip': 40.73284149169922, 'timing/data': 0.0005967710167169571, 'timing/model': 1.7026944679673761, 'learning_rate/qwen_vl_interface': 9.99561818516328e-06, 'learning_rate/action_model': 9.99561818516328e-05, 'epoch': 0.06}) +[2026-06-23 16:55:59,119][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.39590615034103394, 'train/grad_norm_pre_clip': 67.04664611816406, 'timing/data': 0.0007092230953276157, 'timing/model': 1.707397439982742, 'learning_rate/qwen_vl_interface': 9.995486429471166e-06, 'learning_rate/action_model': 9.995486429471166e-05, 'epoch': 0.07}) +[2026-06-23 16:56:02,467][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.5472912788391113, 'train/grad_norm_pre_clip': 23.57006072998047, 'timing/data': 0.0006019619759172201, 'timing/model': 1.6967914190609008, 'learning_rate/qwen_vl_interface': 9.995352723088833e-06, 'learning_rate/action_model': 9.995352723088834e-05, 'epoch': 0.07}) +[2026-06-23 16:56:05,820][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.26272451877593994, 'train/grad_norm_pre_clip': 22.345012664794922, 'timing/data': 0.0005805010441690683, 'timing/model': 1.6990786700043827, 'learning_rate/qwen_vl_interface': 9.995217066071245e-06, 'learning_rate/action_model': 9.995217066071246e-05, 'epoch': 0.07}) +[2026-06-23 16:56:09,181][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.3995988368988037, 'train/grad_norm_pre_clip': 45.82781219482422, 'timing/data': 0.000590750016272068, 'timing/model': 1.7057747680228204, 'learning_rate/qwen_vl_interface': 9.995079458474163e-06, 'learning_rate/action_model': 9.995079458474164e-05, 'epoch': 0.07}) +[2026-06-23 16:56:12,537][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.3903021812438965, 'train/grad_norm_pre_clip': 27.14898109436035, 'timing/data': 0.0005702420603483915, 'timing/model': 1.7071314710192382, 'learning_rate/qwen_vl_interface': 9.994939900354154e-06, 'learning_rate/action_model': 9.994939900354154e-05, 'epoch': 0.07}) +[2026-06-23 16:56:15,898][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.3703787922859192, 'train/grad_norm_pre_clip': 4.429721355438232, 'timing/data': 0.0005893278867006302, 'timing/model': 1.7098068119958043, 'learning_rate/qwen_vl_interface': 9.994798391768582e-06, 'learning_rate/action_model': 9.994798391768582e-05, 'epoch': 0.07}) +[2026-06-23 16:56:19,251][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.39011910557746887, 'train/grad_norm_pre_clip': 6.293217182159424, 'timing/data': 0.0006295829080045223, 'timing/model': 1.7003278390038759, 'learning_rate/qwen_vl_interface': 9.99465493277562e-06, 'learning_rate/action_model': 9.994654932775618e-05, 'epoch': 0.07}) +[2026-06-23 16:56:22,610][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.34520572423934937, 'train/grad_norm_pre_clip': 30.79975128173828, 'timing/data': 0.0005844789557158947, 'timing/model': 1.7031209918204695, 'learning_rate/qwen_vl_interface': 9.994509523434235e-06, 'learning_rate/action_model': 9.994509523434235e-05, 'epoch': 0.07}) +[2026-06-23 16:56:26,180][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.40350374579429626, 'train/grad_norm_pre_clip': 20.03215217590332, 'timing/data': 0.0006656812038272619, 'timing/model': 1.9189938169438392, 'learning_rate/qwen_vl_interface': 9.994362163804202e-06, 'learning_rate/action_model': 9.9943621638042e-05, 'epoch': 0.07}) +[2026-06-23 16:56:29,524][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.4451880156993866, 'train/grad_norm_pre_clip': 12.426732063293457, 'timing/data': 0.000653448048979044, 'timing/model': 1.6992178910877556, 'learning_rate/qwen_vl_interface': 9.994212853946092e-06, 'learning_rate/action_model': 9.994212853946092e-05, 'epoch': 0.07}) +[2026-06-23 16:56:32,888][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.46936875581741333, 'train/grad_norm_pre_clip': 23.525850296020508, 'timing/data': 0.0006801390554755926, 'timing/model': 1.7038243131246418, 'learning_rate/qwen_vl_interface': 9.994061593921283e-06, 'learning_rate/action_model': 9.994061593921282e-05, 'epoch': 0.07}) +[2026-06-23 16:56:36,250][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.41057491302490234, 'train/grad_norm_pre_clip': 27.86371612548828, 'timing/data': 0.0006850280333310366, 'timing/model': 1.7019994461443275, 'learning_rate/qwen_vl_interface': 9.993908383791951e-06, 'learning_rate/action_model': 9.993908383791951e-05, 'epoch': 0.07}) +[2026-06-23 16:56:39,605][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.43741512298583984, 'train/grad_norm_pre_clip': 7.587493896484375, 'timing/data': 0.0008250109385699034, 'timing/model': 1.699597073951736, 'learning_rate/qwen_vl_interface': 9.993753223621076e-06, 'learning_rate/action_model': 9.993753223621077e-05, 'epoch': 0.07}) +[2026-06-23 16:56:42,957][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.44550490379333496, 'train/grad_norm_pre_clip': 56.10589599609375, 'timing/data': 0.0007846339140087366, 'timing/model': 1.7070931461639702, 'learning_rate/qwen_vl_interface': 9.99359611347244e-06, 'learning_rate/action_model': 9.993596113472439e-05, 'epoch': 0.07}) +[2026-06-23 16:56:46,318][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.40801841020584106, 'train/grad_norm_pre_clip': 41.47697830200195, 'timing/data': 0.0007673320360481739, 'timing/model': 1.7065000939182937, 'learning_rate/qwen_vl_interface': 9.99343705341062e-06, 'learning_rate/action_model': 9.99343705341062e-05, 'epoch': 0.07}) +[2026-06-23 16:56:49,665][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.40841910243034363, 'train/grad_norm_pre_clip': 28.365583419799805, 'timing/data': 0.0006651501171290874, 'timing/model': 1.7003522629383951, 'learning_rate/qwen_vl_interface': 9.993276043501004e-06, 'learning_rate/action_model': 9.993276043501004e-05, 'epoch': 0.07}) +[2026-06-23 16:56:53,019][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.40892618894577026, 'train/grad_norm_pre_clip': 5.158494472503662, 'timing/data': 0.0006095860153436661, 'timing/model': 1.7022992649581283, 'learning_rate/qwen_vl_interface': 9.993113083809776e-06, 'learning_rate/action_model': 9.993113083809775e-05, 'epoch': 0.07}) +[2026-06-23 16:56:56,364][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.40621069073677063, 'train/grad_norm_pre_clip': 34.12510299682617, 'timing/data': 0.0006966390646994114, 'timing/model': 1.69837364112027, 'learning_rate/qwen_vl_interface': 9.992948174403921e-06, 'learning_rate/action_model': 9.992948174403922e-05, 'epoch': 0.07}) +[2026-06-23 16:56:59,731][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.4580506384372711, 'train/grad_norm_pre_clip': 43.50822830200195, 'timing/data': 0.0006419469136744738, 'timing/model': 1.7120430651120842, 'learning_rate/qwen_vl_interface': 9.99278131535123e-06, 'learning_rate/action_model': 9.992781315351229e-05, 'epoch': 0.07}) +[2026-06-23 16:57:03,076][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.5021074414253235, 'train/grad_norm_pre_clip': 50.22848892211914, 'timing/data': 0.0005956999957561493, 'timing/model': 1.7003904290031642, 'learning_rate/qwen_vl_interface': 9.992612506720289e-06, 'learning_rate/action_model': 9.99261250672029e-05, 'epoch': 0.07}) +[2026-06-23 16:57:06,424][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.3496313989162445, 'train/grad_norm_pre_clip': 37.93864822387695, 'timing/data': 0.0006632870063185692, 'timing/model': 1.6978187391068786, 'learning_rate/qwen_vl_interface': 9.992441748580492e-06, 'learning_rate/action_model': 9.99244174858049e-05, 'epoch': 0.07}) +[2026-06-23 16:57:09,776][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.5077245235443115, 'train/grad_norm_pre_clip': 18.246374130249023, 'timing/data': 0.0008854640182107687, 'timing/model': 1.703487846069038, 'learning_rate/qwen_vl_interface': 9.992269041002028e-06, 'learning_rate/action_model': 9.992269041002027e-05, 'epoch': 0.07}) +[2026-06-23 16:57:13,229][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.5257584452629089, 'train/grad_norm_pre_clip': 71.38483428955078, 'timing/data': 0.0008444769773632288, 'timing/model': 1.804392232792452, 'learning_rate/qwen_vl_interface': 9.992094384055894e-06, 'learning_rate/action_model': 9.992094384055895e-05, 'epoch': 0.07}) +[2026-06-23 16:57:16,579][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.4525796175003052, 'train/grad_norm_pre_clip': 55.56303024291992, 'timing/data': 0.0006085941568017006, 'timing/model': 1.7054207760374993, 'learning_rate/qwen_vl_interface': 9.991917777813884e-06, 'learning_rate/action_model': 9.991917777813884e-05, 'epoch': 0.07}) +[2026-06-23 16:57:19,927][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.7880350351333618, 'train/grad_norm_pre_clip': 101.42698669433594, 'timing/data': 0.0005835772026330233, 'timing/model': 1.6937295070383698, 'learning_rate/qwen_vl_interface': 9.991739222348593e-06, 'learning_rate/action_model': 9.991739222348592e-05, 'epoch': 0.07}) +[2026-06-23 16:57:23,274][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.4932861626148224, 'train/grad_norm_pre_clip': 75.58744049072266, 'timing/data': 0.0005805310793220997, 'timing/model': 1.7014826021622866, 'learning_rate/qwen_vl_interface': 9.991558717733418e-06, 'learning_rate/action_model': 9.991558717733417e-05, 'epoch': 0.07}) +[2026-06-23 16:57:26,617][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.43487703800201416, 'train/grad_norm_pre_clip': 36.537960052490234, 'timing/data': 0.0005977130495011806, 'timing/model': 1.6968531650491059, 'learning_rate/qwen_vl_interface': 9.99137626404256e-06, 'learning_rate/action_model': 9.991376264042559e-05, 'epoch': 0.08}) +[2026-06-23 16:57:29,966][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.4198525846004486, 'train/grad_norm_pre_clip': 12.013372421264648, 'timing/data': 0.0008233380503952503, 'timing/model': 1.7021931649651378, 'learning_rate/qwen_vl_interface': 9.991191861351015e-06, 'learning_rate/action_model': 9.991191861351014e-05, 'epoch': 0.08}) +[2026-06-23 16:57:33,333][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.4941002428531647, 'train/grad_norm_pre_clip': 29.917068481445312, 'timing/data': 0.0005959300324320793, 'timing/model': 1.706763467984274, 'learning_rate/qwen_vl_interface': 9.991005509734587e-06, 'learning_rate/action_model': 9.991005509734587e-05, 'epoch': 0.08}) +[2026-06-23 16:57:36,691][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.4518526494503021, 'train/grad_norm_pre_clip': 19.12508773803711, 'timing/data': 0.0005860610399395227, 'timing/model': 1.7060702999588102, 'learning_rate/qwen_vl_interface': 9.990817209269878e-06, 'learning_rate/action_model': 9.990817209269878e-05, 'epoch': 0.08}) +[2026-06-23 16:57:40,047][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.4064328372478485, 'train/grad_norm_pre_clip': 10.724811553955078, 'timing/data': 0.000699163880199194, 'timing/model': 1.7027210290543735, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.08}) +[2026-06-23 16:57:43,408][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.44038674235343933, 'train/grad_norm_pre_clip': 34.509117126464844, 'timing/data': 0.0006026129703968763, 'timing/model': 1.704966904129833, 'learning_rate/qwen_vl_interface': 9.990434762106029e-06, 'learning_rate/action_model': 9.990434762106028e-05, 'epoch': 0.08}) +[2026-06-23 16:57:46,775][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 200 +[2026-06-23 16:59:20,405][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 200 +[2026-06-23 16:59:20,406][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 200 +[2026-06-23 17:00:34,115][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 200 +[2026-06-23 17:00:34,116][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.3995933532714844, 'train/grad_norm_pre_clip': 8.183698654174805, 'eval/action_loss/seconds': 93.62842905987054, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.22609199583530426, 'eval/latency_2/loss': 0.22609199583530426, 'eval/flappy/loss': 0.22609199583530426, 'eval/flappy/latency_2/loss': 0.22609199583530426, 'eval/flap_precision': 0.751060820367751, 'eval/flap_recall': 0.7165991902834008, 'eval/flap_f1': 0.7334254143646408, 'eval/flap_rate_model': 0.1963888888888889, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.7334254143646409, 'eval/latency_2/flap_precision': 0.751060820367751, 'eval/latency_2/flap_recall': 0.7165991902834008, 'eval/latency_2/flap_f1': 0.7334254143646408, 'eval/latency_2/flap_rate_model': 0.1963888888888889, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.7334254143646409, 'eval/action_classification/data_wait_seconds': 0.9995468258857727, 'eval/action_classification/predict_seconds': 72.57767486572266, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006800780538469553, 'timing/model': 1.7116951730567962, 'learning_rate/qwen_vl_interface': 9.990240615564099e-06, 'learning_rate/action_model': 9.990240615564099e-05, 'epoch': 0.08}) +[2026-06-23 17:00:37,468][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.3200586438179016, 'train/grad_norm_pre_clip': 8.090289115905762, 'timing/data': 0.0005848589353263378, 'timing/model': 1.6980013668071479, 'learning_rate/qwen_vl_interface': 9.990044520488305e-06, 'learning_rate/action_model': 9.990044520488305e-05, 'epoch': 0.08}) +[2026-06-23 17:00:40,828][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.35479292273521423, 'train/grad_norm_pre_clip': 17.742067337036133, 'timing/data': 0.0005820239894092083, 'timing/model': 1.7060791128315032, 'learning_rate/qwen_vl_interface': 9.989846476959258e-06, 'learning_rate/action_model': 9.989846476959257e-05, 'epoch': 0.08}) +[2026-06-23 17:00:44,182][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.43145209550857544, 'train/grad_norm_pre_clip': 22.044078826904297, 'timing/data': 0.00059840502217412, 'timing/model': 1.697078324854374, 'learning_rate/qwen_vl_interface': 9.989646485058361e-06, 'learning_rate/action_model': 9.989646485058361e-05, 'epoch': 0.08}) +[2026-06-23 17:00:47,546][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.44165274500846863, 'train/grad_norm_pre_clip': 13.62337589263916, 'timing/data': 0.000585711095482111, 'timing/model': 1.7095475131645799, 'learning_rate/qwen_vl_interface': 9.98944454486783e-06, 'learning_rate/action_model': 9.989444544867829e-05, 'epoch': 0.08}) +[2026-06-23 17:00:50,898][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.356073796749115, 'train/grad_norm_pre_clip': 31.64035415649414, 'timing/data': 0.0005866021383553743, 'timing/model': 1.7000734750181437, 'learning_rate/qwen_vl_interface': 9.989240656470667e-06, 'learning_rate/action_model': 9.989240656470666e-05, 'epoch': 0.08}) +[2026-06-23 17:00:54,270][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.3145641088485718, 'train/grad_norm_pre_clip': 6.982075214385986, 'timing/data': 0.0006450831424444914, 'timing/model': 1.7133340609725565, 'learning_rate/qwen_vl_interface': 9.98903481995069e-06, 'learning_rate/action_model': 9.989034819950689e-05, 'epoch': 0.08}) +[2026-06-23 17:00:57,631][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.3103937804698944, 'train/grad_norm_pre_clip': 55.244510650634766, 'timing/data': 0.0006683759856969118, 'timing/model': 1.7096248159650713, 'learning_rate/qwen_vl_interface': 9.988827035392506e-06, 'learning_rate/action_model': 9.988827035392507e-05, 'epoch': 0.08}) +[2026-06-23 17:01:00,992][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.3610662817955017, 'train/grad_norm_pre_clip': 15.809435844421387, 'timing/data': 0.0006003179587423801, 'timing/model': 1.7049989190418273, 'learning_rate/qwen_vl_interface': 9.988617302881529e-06, 'learning_rate/action_model': 9.988617302881529e-05, 'epoch': 0.08}) +[2026-06-23 17:01:04,355][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.30020976066589355, 'train/grad_norm_pre_clip': 9.305684089660645, 'timing/data': 0.0006021710578352213, 'timing/model': 1.712618162157014, 'learning_rate/qwen_vl_interface': 9.988405622503972e-06, 'learning_rate/action_model': 9.988405622503971e-05, 'epoch': 0.08}) +[2026-06-23 17:01:07,721][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.43793970346450806, 'train/grad_norm_pre_clip': 47.533058166503906, 'timing/data': 0.000598004087805748, 'timing/model': 1.7035815070848912, 'learning_rate/qwen_vl_interface': 9.988191994346848e-06, 'learning_rate/action_model': 9.988191994346848e-05, 'epoch': 0.08}) +[2026-06-23 17:01:11,077][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.41601327061653137, 'train/grad_norm_pre_clip': 63.0220832824707, 'timing/data': 0.000631256029009819, 'timing/model': 1.6994538460858166, 'learning_rate/qwen_vl_interface': 9.987976418497974e-06, 'learning_rate/action_model': 9.987976418497972e-05, 'epoch': 0.08}) +[2026-06-23 17:01:14,432][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.42567873001098633, 'train/grad_norm_pre_clip': 9.462371826171875, 'timing/data': 0.0005789068527519703, 'timing/model': 1.7077660418581218, 'learning_rate/qwen_vl_interface': 9.987758895045964e-06, 'learning_rate/action_model': 9.987758895045963e-05, 'epoch': 0.08}) +[2026-06-23 17:01:18,076][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.2893894612789154, 'train/grad_norm_pre_clip': 16.003070831298828, 'timing/data': 0.0006144240032881498, 'timing/model': 1.9906161960680038, 'learning_rate/qwen_vl_interface': 9.98753942408023e-06, 'learning_rate/action_model': 9.98753942408023e-05, 'epoch': 0.08}) +[2026-06-23 17:01:21,436][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.45990291237831116, 'train/grad_norm_pre_clip': 47.470237731933594, 'timing/data': 0.0005920419935137033, 'timing/model': 1.7088507730513811, 'learning_rate/qwen_vl_interface': 9.987318005690992e-06, 'learning_rate/action_model': 9.987318005690992e-05, 'epoch': 0.08}) +[2026-06-23 17:01:24,784][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.4527589976787567, 'train/grad_norm_pre_clip': 36.5611686706543, 'timing/data': 0.0005948380567133427, 'timing/model': 1.7032072129659355, 'learning_rate/qwen_vl_interface': 9.987094639969268e-06, 'learning_rate/action_model': 9.987094639969268e-05, 'epoch': 0.08}) +[2026-06-23 17:01:28,134][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.530813992023468, 'train/grad_norm_pre_clip': 56.144264221191406, 'timing/data': 0.00061645801179111, 'timing/model': 1.7022493791300803, 'learning_rate/qwen_vl_interface': 9.986869327006872e-06, 'learning_rate/action_model': 9.986869327006873e-05, 'epoch': 0.08}) +[2026-06-23 17:01:31,488][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.3802439570426941, 'train/grad_norm_pre_clip': 18.168840408325195, 'timing/data': 0.0006605810485780239, 'timing/model': 1.703386161942035, 'learning_rate/qwen_vl_interface': 9.986642066896424e-06, 'learning_rate/action_model': 9.986642066896423e-05, 'epoch': 0.08}) +[2026-06-23 17:01:34,850][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.3661303222179413, 'train/grad_norm_pre_clip': 9.230891227722168, 'timing/data': 0.0005982350558042526, 'timing/model': 1.703861721092835, 'learning_rate/qwen_vl_interface': 9.986412859731341e-06, 'learning_rate/action_model': 9.98641285973134e-05, 'epoch': 0.08}) +[2026-06-23 17:01:38,208][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.4239477515220642, 'train/grad_norm_pre_clip': 22.16522216796875, 'timing/data': 0.0005981340073049068, 'timing/model': 1.7020422380883247, 'learning_rate/qwen_vl_interface': 9.986181705605839e-06, 'learning_rate/action_model': 9.986181705605839e-05, 'epoch': 0.08}) +[2026-06-23 17:01:41,567][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.47714996337890625, 'train/grad_norm_pre_clip': 42.15768814086914, 'timing/data': 0.0005849788431078196, 'timing/model': 1.701934551820159, 'learning_rate/qwen_vl_interface': 9.98594860461494e-06, 'learning_rate/action_model': 9.98594860461494e-05, 'epoch': 0.09}) +[2026-06-23 17:01:44,926][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.3285447359085083, 'train/grad_norm_pre_clip': 3.568453550338745, 'timing/data': 0.0005967221222817898, 'timing/model': 1.70236780308187, 'learning_rate/qwen_vl_interface': 9.985713556854461e-06, 'learning_rate/action_model': 9.98571355685446e-05, 'epoch': 0.09}) +[2026-06-23 17:01:48,296][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.3861411213874817, 'train/grad_norm_pre_clip': 35.294410705566406, 'timing/data': 0.0006058691069483757, 'timing/model': 1.7055298411287367, 'learning_rate/qwen_vl_interface': 9.985476562421024e-06, 'learning_rate/action_model': 9.985476562421024e-05, 'epoch': 0.09}) +[2026-06-23 17:01:51,654][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.41883546113967896, 'train/grad_norm_pre_clip': 19.811725616455078, 'timing/data': 0.0006621340289711952, 'timing/model': 1.703747516963631, 'learning_rate/qwen_vl_interface': 9.985237621412047e-06, 'learning_rate/action_model': 9.985237621412046e-05, 'epoch': 0.09}) +[2026-06-23 17:01:55,020][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.32759276032447815, 'train/grad_norm_pre_clip': 19.98701286315918, 'timing/data': 0.0006440209690481424, 'timing/model': 1.70835694600828, 'learning_rate/qwen_vl_interface': 9.984996733925749e-06, 'learning_rate/action_model': 9.984996733925748e-05, 'epoch': 0.09}) +[2026-06-23 17:01:58,379][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.5185606479644775, 'train/grad_norm_pre_clip': 27.42436408996582, 'timing/data': 0.000602601794525981, 'timing/model': 1.7052796299103647, 'learning_rate/qwen_vl_interface': 9.98475390006115e-06, 'learning_rate/action_model': 9.98475390006115e-05, 'epoch': 0.09}) +[2026-06-23 17:02:01,744][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.3671223223209381, 'train/grad_norm_pre_clip': 41.071144104003906, 'timing/data': 0.0005995668470859528, 'timing/model': 1.704432524042204, 'learning_rate/qwen_vl_interface': 9.98450911991807e-06, 'learning_rate/action_model': 9.98450911991807e-05, 'epoch': 0.09}) +[2026-06-23 17:02:05,198][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.418883740901947, 'train/grad_norm_pre_clip': 9.126157760620117, 'timing/data': 0.0005898589733988047, 'timing/model': 1.802927321055904, 'learning_rate/qwen_vl_interface': 9.98426239359713e-06, 'learning_rate/action_model': 9.984262393597129e-05, 'epoch': 0.09}) +[2026-06-23 17:02:08,559][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.387419193983078, 'train/grad_norm_pre_clip': 14.330912590026855, 'timing/data': 0.0005977728869765997, 'timing/model': 1.702271023998037, 'learning_rate/qwen_vl_interface': 9.984013721199747e-06, 'learning_rate/action_model': 9.984013721199747e-05, 'epoch': 0.09}) +[2026-06-23 17:02:11,929][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.3962290585041046, 'train/grad_norm_pre_clip': 23.77394676208496, 'timing/data': 0.0006508829537779093, 'timing/model': 1.7089251251891255, 'learning_rate/qwen_vl_interface': 9.983763102828143e-06, 'learning_rate/action_model': 9.983763102828143e-05, 'epoch': 0.09}) +[2026-06-23 17:02:15,300][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.3438751697540283, 'train/grad_norm_pre_clip': 33.57857894897461, 'timing/data': 0.0006111178081482649, 'timing/model': 1.7046446478925645, 'learning_rate/qwen_vl_interface': 9.983510538585336e-06, 'learning_rate/action_model': 9.983510538585337e-05, 'epoch': 0.09}) +[2026-06-23 17:02:18,658][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.4572400748729706, 'train/grad_norm_pre_clip': 28.984130859375, 'timing/data': 0.0006865400355309248, 'timing/model': 1.704114078078419, 'learning_rate/qwen_vl_interface': 9.98325602857515e-06, 'learning_rate/action_model': 9.98325602857515e-05, 'epoch': 0.09}) +[2026-06-23 17:02:22,006][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.32210662961006165, 'train/grad_norm_pre_clip': 21.471399307250977, 'timing/data': 0.0006008590571582317, 'timing/model': 1.7036751890555024, 'learning_rate/qwen_vl_interface': 9.9829995729022e-06, 'learning_rate/action_model': 9.982999572902199e-05, 'epoch': 0.09}) +[2026-06-23 17:02:25,352][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.3247736096382141, 'train/grad_norm_pre_clip': 5.1678853034973145, 'timing/data': 0.000953983049839735, 'timing/model': 1.6988326059654355, 'learning_rate/qwen_vl_interface': 9.982741171671905e-06, 'learning_rate/action_model': 9.982741171671906e-05, 'epoch': 0.09}) +[2026-06-23 17:02:28,709][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.3264700472354889, 'train/grad_norm_pre_clip': 23.478029251098633, 'timing/data': 0.0006987929809838533, 'timing/model': 1.7074080971069634, 'learning_rate/qwen_vl_interface': 9.982480824990489e-06, 'learning_rate/action_model': 9.982480824990488e-05, 'epoch': 0.09}) +[2026-06-23 17:02:32,070][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.30968421697616577, 'train/grad_norm_pre_clip': 38.10609817504883, 'timing/data': 0.0006944949273020029, 'timing/model': 1.7052386971190572, 'learning_rate/qwen_vl_interface': 9.982218532964963e-06, 'learning_rate/action_model': 9.982218532964964e-05, 'epoch': 0.09}) +[2026-06-23 17:02:35,429][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.30241405963897705, 'train/grad_norm_pre_clip': 68.7917251586914, 'timing/data': 0.000639601843431592, 'timing/model': 1.7062157900072634, 'learning_rate/qwen_vl_interface': 9.981954295703152e-06, 'learning_rate/action_model': 9.981954295703152e-05, 'epoch': 0.09}) +[2026-06-23 17:02:38,786][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.3582153916358948, 'train/grad_norm_pre_clip': 22.3564510345459, 'timing/data': 0.0006639580242335796, 'timing/model': 1.7015747749246657, 'learning_rate/qwen_vl_interface': 9.981688113313672e-06, 'learning_rate/action_model': 9.981688113313672e-05, 'epoch': 0.09}) +[2026-06-23 17:02:42,160][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.3147951364517212, 'train/grad_norm_pre_clip': 11.05211353302002, 'timing/data': 0.0006901659071445465, 'timing/model': 1.711997717851773, 'learning_rate/qwen_vl_interface': 9.98141998590594e-06, 'learning_rate/action_model': 9.981419985905939e-05, 'epoch': 0.09}) +[2026-06-23 17:02:45,520][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.22568991780281067, 'train/grad_norm_pre_clip': 13.536911010742188, 'timing/data': 0.0006120400503277779, 'timing/model': 1.7101525079924613, 'learning_rate/qwen_vl_interface': 9.981149913590171e-06, 'learning_rate/action_model': 9.981149913590172e-05, 'epoch': 0.09}) +[2026-06-23 17:02:48,878][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.3091202676296234, 'train/grad_norm_pre_clip': 8.432181358337402, 'timing/data': 0.0005967211909592152, 'timing/model': 1.7046720751095563, 'learning_rate/qwen_vl_interface': 9.980877896477386e-06, 'learning_rate/action_model': 9.980877896477386e-05, 'epoch': 0.09}) +[2026-06-23 17:02:52,243][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.3881843090057373, 'train/grad_norm_pre_clip': 7.535196781158447, 'timing/data': 0.0006005291361361742, 'timing/model': 1.7040129229426384, 'learning_rate/qwen_vl_interface': 9.980603934679397e-06, 'learning_rate/action_model': 9.980603934679398e-05, 'epoch': 0.09}) +[2026-06-23 17:02:55,691][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.4292917549610138, 'train/grad_norm_pre_clip': 26.040254592895508, 'timing/data': 0.0005951381754130125, 'timing/model': 1.7008099309168756, 'learning_rate/qwen_vl_interface': 9.980328028308824e-06, 'learning_rate/action_model': 9.980328028308824e-05, 'epoch': 0.09}) +[2026-06-23 17:02:59,048][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.2831622362136841, 'train/grad_norm_pre_clip': 18.188419342041016, 'timing/data': 0.000684716971591115, 'timing/model': 1.7027017218060791, 'learning_rate/qwen_vl_interface': 9.980050177479075e-06, 'learning_rate/action_model': 9.980050177479076e-05, 'epoch': 0.09}) +[2026-06-23 17:03:02,408][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.2961927652359009, 'train/grad_norm_pre_clip': 12.67110538482666, 'timing/data': 0.0006006478797644377, 'timing/model': 1.7049780369270593, 'learning_rate/qwen_vl_interface': 9.97977038230437e-06, 'learning_rate/action_model': 9.97977038230437e-05, 'epoch': 0.09}) +[2026-06-23 17:03:05,771][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.44500699639320374, 'train/grad_norm_pre_clip': 78.69273376464844, 'timing/data': 0.0006989638786762953, 'timing/model': 1.7102187320124358, 'learning_rate/qwen_vl_interface': 9.97948864289972e-06, 'learning_rate/action_model': 9.97948864289972e-05, 'epoch': 0.09}) +[2026-06-23 17:03:09,142][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.30991673469543457, 'train/grad_norm_pre_clip': 33.13554000854492, 'timing/data': 0.0007461318746209145, 'timing/model': 1.713423180161044, 'learning_rate/qwen_vl_interface': 9.979204959380936e-06, 'learning_rate/action_model': 9.979204959380936e-05, 'epoch': 0.1}) +[2026-06-23 17:03:12,499][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.3256526291370392, 'train/grad_norm_pre_clip': 10.749935150146484, 'timing/data': 0.000587213784456253, 'timing/model': 1.7009550968650728, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.1}) +[2026-06-23 17:03:15,852][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.3718903064727783, 'train/grad_norm_pre_clip': 17.604490280151367, 'timing/data': 0.0007058267947286367, 'timing/model': 1.7064237808808684, 'learning_rate/qwen_vl_interface': 9.978631760468216e-06, 'learning_rate/action_model': 9.978631760468216e-05, 'epoch': 0.1}) +[2026-06-23 17:03:19,226][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.3564949929714203, 'train/grad_norm_pre_clip': 29.487028121948242, 'timing/data': 0.0005973819643259048, 'timing/model': 1.7180770051199943, 'learning_rate/qwen_vl_interface': 9.9783422453099e-06, 'learning_rate/action_model': 9.9783422453099e-05, 'epoch': 0.1}) +[2026-06-23 17:03:27,690][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 250 +[2026-06-23 17:03:27,691][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.36977699398994446, 'train/grad_norm_pre_clip': 4.884023189544678, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/mean_length': 31.0, 'rl_games_eval/mid_train/std_reward': 0.0, 'rl_games_eval/mid_train/std_length': 0.0, 'rl_games_eval/mid_train/macro_mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/macro_mean_length': 31.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 2.0000000000000013, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 31.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 0.0, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 2.0000000000000013, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005972220096737146, 'timing/model': 1.699948898050934, 'learning_rate/qwen_vl_interface': 9.978050786508693e-06, 'learning_rate/action_model': 9.978050786508692e-05, 'epoch': 0.1}) +[2026-06-23 17:03:31,033][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.46598008275032043, 'train/grad_norm_pre_clip': 27.96097755432129, 'timing/data': 0.0006648700218647718, 'timing/model': 1.6955052099656314, 'learning_rate/qwen_vl_interface': 9.9777573841844e-06, 'learning_rate/action_model': 9.977757384184401e-05, 'epoch': 0.1}) +[2026-06-23 17:03:34,388][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.4291519522666931, 'train/grad_norm_pre_clip': 93.92909240722656, 'timing/data': 0.0006934828124940395, 'timing/model': 1.7049610749818385, 'learning_rate/qwen_vl_interface': 9.97746203845763e-06, 'learning_rate/action_model': 9.97746203845763e-05, 'epoch': 0.1}) +[2026-06-23 17:03:37,736][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.3470454812049866, 'train/grad_norm_pre_clip': 42.380821228027344, 'timing/data': 0.000680048018693924, 'timing/model': 1.6981712630949914, 'learning_rate/qwen_vl_interface': 9.97716474944979e-06, 'learning_rate/action_model': 9.977164749449788e-05, 'epoch': 0.1}) +[2026-06-23 17:03:41,092][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.30323338508605957, 'train/grad_norm_pre_clip': 31.49677085876465, 'timing/data': 0.0006023319438099861, 'timing/model': 1.7099108409602195, 'learning_rate/qwen_vl_interface': 9.976865517283079e-06, 'learning_rate/action_model': 9.976865517283078e-05, 'epoch': 0.1}) +[2026-06-23 17:03:44,456][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.3015041947364807, 'train/grad_norm_pre_clip': 22.842498779296875, 'timing/data': 0.0006008590571582317, 'timing/model': 1.7038716690149158, 'learning_rate/qwen_vl_interface': 9.976564342080505e-06, 'learning_rate/action_model': 9.976564342080504e-05, 'epoch': 0.1}) +[2026-06-23 17:03:48,113][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.366362065076828, 'train/grad_norm_pre_clip': 40.84064483642578, 'timing/data': 0.0005977628752589226, 'timing/model': 2.005627625156194, 'learning_rate/qwen_vl_interface': 9.976261223965867e-06, 'learning_rate/action_model': 9.976261223965865e-05, 'epoch': 0.1}) +[2026-06-23 17:03:51,495][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.41336551308631897, 'train/grad_norm_pre_clip': 27.366905212402344, 'timing/data': 0.0005907798185944557, 'timing/model': 1.707791247870773, 'learning_rate/qwen_vl_interface': 9.975956163063767e-06, 'learning_rate/action_model': 9.975956163063767e-05, 'epoch': 0.1}) +[2026-06-23 17:03:54,847][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.3602399230003357, 'train/grad_norm_pre_clip': 22.266319274902344, 'timing/data': 0.0005975530948489904, 'timing/model': 1.697860793909058, 'learning_rate/qwen_vl_interface': 9.975649159499605e-06, 'learning_rate/action_model': 9.975649159499603e-05, 'epoch': 0.1}) +[2026-06-23 17:03:58,196][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.2988881468772888, 'train/grad_norm_pre_clip': 69.3356704711914, 'timing/data': 0.0005951279308646917, 'timing/model': 1.701407076092437, 'learning_rate/qwen_vl_interface': 9.975340213399573e-06, 'learning_rate/action_model': 9.975340213399573e-05, 'epoch': 0.1}) +[2026-06-23 17:04:01,550][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.5036240220069885, 'train/grad_norm_pre_clip': 109.19810485839844, 'timing/data': 0.0006013498641550541, 'timing/model': 1.7008297881111503, 'learning_rate/qwen_vl_interface': 9.975029324890674e-06, 'learning_rate/action_model': 9.975029324890674e-05, 'epoch': 0.1}) +[2026-06-23 17:04:04,909][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.29062408208847046, 'train/grad_norm_pre_clip': 73.30421447753906, 'timing/data': 0.0005991258658468723, 'timing/model': 1.7104760119691491, 'learning_rate/qwen_vl_interface': 9.974716494100698e-06, 'learning_rate/action_model': 9.974716494100697e-05, 'epoch': 0.1}) +[2026-06-23 17:04:08,277][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.4095328450202942, 'train/grad_norm_pre_clip': 66.98794555664062, 'timing/data': 0.0006051268428564072, 'timing/model': 1.7080081619787961, 'learning_rate/qwen_vl_interface': 9.97440172115824e-06, 'learning_rate/action_model': 9.974401721158239e-05, 'epoch': 0.1}) +[2026-06-23 17:04:11,639][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.26794689893722534, 'train/grad_norm_pre_clip': 7.002272129058838, 'timing/data': 0.00058798398822546, 'timing/model': 1.7051503569819033, 'learning_rate/qwen_vl_interface': 9.97408500619269e-06, 'learning_rate/action_model': 9.97408500619269e-05, 'epoch': 0.1}) +[2026-06-23 17:04:14,993][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.31592392921447754, 'train/grad_norm_pre_clip': 23.17779541015625, 'timing/data': 0.0005931539926677942, 'timing/model': 1.705602674977854, 'learning_rate/qwen_vl_interface': 9.973766349334239e-06, 'learning_rate/action_model': 9.973766349334239e-05, 'epoch': 0.1}) +[2026-06-23 17:04:18,350][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.3426734507083893, 'train/grad_norm_pre_clip': 33.46349334716797, 'timing/data': 0.0005887469742447138, 'timing/model': 1.7028881218284369, 'learning_rate/qwen_vl_interface': 9.973445750713875e-06, 'learning_rate/action_model': 9.973445750713875e-05, 'epoch': 0.1}) +[2026-06-23 17:04:21,706][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.3224298655986786, 'train/grad_norm_pre_clip': 7.3811354637146, 'timing/data': 0.0005943570286035538, 'timing/model': 1.7014679710846394, 'learning_rate/qwen_vl_interface': 9.97312321046338e-06, 'learning_rate/action_model': 9.97312321046338e-05, 'epoch': 0.1}) +[2026-06-23 17:04:25,060][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.35247278213500977, 'train/grad_norm_pre_clip': 45.27293395996094, 'timing/data': 0.0005884049460291862, 'timing/model': 1.7027958817780018, 'learning_rate/qwen_vl_interface': 9.972798728715343e-06, 'learning_rate/action_model': 9.972798728715344e-05, 'epoch': 0.1}) +[2026-06-23 17:04:28,410][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.2097626030445099, 'train/grad_norm_pre_clip': 20.882421493530273, 'timing/data': 0.000597432954236865, 'timing/model': 1.7022738889791071, 'learning_rate/qwen_vl_interface': 9.972472305603145e-06, 'learning_rate/action_model': 9.972472305603144e-05, 'epoch': 0.1}) +[2026-06-23 17:04:31,762][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.33536115288734436, 'train/grad_norm_pre_clip': 8.202571868896484, 'timing/data': 0.0005841779056936502, 'timing/model': 1.7040863980073482, 'learning_rate/qwen_vl_interface': 9.972143941260965e-06, 'learning_rate/action_model': 9.972143941260965e-05, 'epoch': 0.1}) +[2026-06-23 17:04:35,269][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.4238683879375458, 'train/grad_norm_pre_clip': 17.143274307250977, 'timing/data': 0.0005963800940662622, 'timing/model': 1.8526708639692515, 'learning_rate/qwen_vl_interface': 9.971813635823781e-06, 'learning_rate/action_model': 9.971813635823781e-05, 'epoch': 0.1}) +[2026-06-23 17:04:38,616][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.3425653576850891, 'train/grad_norm_pre_clip': 12.00615119934082, 'timing/data': 0.0005901891272515059, 'timing/model': 1.6971618849784136, 'learning_rate/qwen_vl_interface': 9.971481389427372e-06, 'learning_rate/action_model': 9.971481389427371e-05, 'epoch': 0.1}) +[2026-06-23 17:04:41,959][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.43769246339797974, 'train/grad_norm_pre_clip': 47.18698501586914, 'timing/data': 0.0007627441082149744, 'timing/model': 1.6981860070955008, 'learning_rate/qwen_vl_interface': 9.971147202208308e-06, 'learning_rate/action_model': 9.971147202208308e-05, 'epoch': 0.11}) +[2026-06-23 17:04:45,301][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.3159763813018799, 'train/grad_norm_pre_clip': 47.06418228149414, 'timing/data': 0.0005981740541756153, 'timing/model': 1.6960931411013007, 'learning_rate/qwen_vl_interface': 9.970811074303962e-06, 'learning_rate/action_model': 9.970811074303962e-05, 'epoch': 0.11}) +[2026-06-23 17:04:48,660][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.4264795482158661, 'train/grad_norm_pre_clip': 12.666080474853516, 'timing/data': 0.0006841958966106176, 'timing/model': 1.7025690178852528, 'learning_rate/qwen_vl_interface': 9.970473005852505e-06, 'learning_rate/action_model': 9.970473005852505e-05, 'epoch': 0.11}) +[2026-06-23 17:04:52,013][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.5701823234558105, 'train/grad_norm_pre_clip': 63.2889518737793, 'timing/data': 0.0005910198669880629, 'timing/model': 1.7011892120353878, 'learning_rate/qwen_vl_interface': 9.970132996992904e-06, 'learning_rate/action_model': 9.970132996992904e-05, 'epoch': 0.11}) +[2026-06-23 17:04:55,354][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.4282475709915161, 'train/grad_norm_pre_clip': 72.0802230834961, 'timing/data': 0.0006102160550653934, 'timing/model': 1.701217973139137, 'learning_rate/qwen_vl_interface': 9.969791047864921e-06, 'learning_rate/action_model': 9.969791047864921e-05, 'epoch': 0.11}) +[2026-06-23 17:04:58,698][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.3608797788619995, 'train/grad_norm_pre_clip': 47.89234924316406, 'timing/data': 0.0006599000189453363, 'timing/model': 1.6964298139791936, 'learning_rate/qwen_vl_interface': 9.969447158609122e-06, 'learning_rate/action_model': 9.96944715860912e-05, 'epoch': 0.11}) +[2026-06-23 17:05:02,047][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.2414751797914505, 'train/grad_norm_pre_clip': 56.74775695800781, 'timing/data': 0.0006727140862494707, 'timing/model': 1.6963217980228364, 'learning_rate/qwen_vl_interface': 9.969101329366864e-06, 'learning_rate/action_model': 9.969101329366863e-05, 'epoch': 0.11}) +[2026-06-23 17:05:05,404][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.455000638961792, 'train/grad_norm_pre_clip': 39.84619903564453, 'timing/data': 0.0006617030594497919, 'timing/model': 1.7049194709397852, 'learning_rate/qwen_vl_interface': 9.968753560280305e-06, 'learning_rate/action_model': 9.968753560280304e-05, 'epoch': 0.11}) +[2026-06-23 17:05:08,768][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.4010986089706421, 'train/grad_norm_pre_clip': 17.474143981933594, 'timing/data': 0.00076653016731143, 'timing/model': 1.706578667042777, 'learning_rate/qwen_vl_interface': 9.968403851492402e-06, 'learning_rate/action_model': 9.968403851492401e-05, 'epoch': 0.11}) +[2026-06-23 17:05:12,127][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.33125966787338257, 'train/grad_norm_pre_clip': 8.67719841003418, 'timing/data': 0.0006496799178421497, 'timing/model': 1.7010806321632117, 'learning_rate/qwen_vl_interface': 9.968052203146904e-06, 'learning_rate/action_model': 9.968052203146903e-05, 'epoch': 0.11}) +[2026-06-23 17:05:15,493][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.41592711210250854, 'train/grad_norm_pre_clip': 62.45009231567383, 'timing/data': 0.0005900389514863491, 'timing/model': 1.708020880818367, 'learning_rate/qwen_vl_interface': 9.967698615388363e-06, 'learning_rate/action_model': 9.967698615388362e-05, 'epoch': 0.11}) +[2026-06-23 17:05:18,854][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.5779409408569336, 'train/grad_norm_pre_clip': 139.23533630371094, 'timing/data': 0.0005929139442741871, 'timing/model': 1.7019024721812457, 'learning_rate/qwen_vl_interface': 9.967343088362123e-06, 'learning_rate/action_model': 9.967343088362122e-05, 'epoch': 0.11}) +[2026-06-23 17:05:22,208][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.37604716420173645, 'train/grad_norm_pre_clip': 75.71212005615234, 'timing/data': 0.000639019999653101, 'timing/model': 1.6997006419114769, 'learning_rate/qwen_vl_interface': 9.966985622214328e-06, 'learning_rate/action_model': 9.966985622214329e-05, 'epoch': 0.11}) +[2026-06-23 17:05:25,672][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.40208134055137634, 'train/grad_norm_pre_clip': 45.271480560302734, 'timing/data': 0.0005821241065859795, 'timing/model': 1.701454886002466, 'learning_rate/qwen_vl_interface': 9.966626217091922e-06, 'learning_rate/action_model': 9.966626217091922e-05, 'epoch': 0.11}) +[2026-06-23 17:05:29,034][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.29443806409835815, 'train/grad_norm_pre_clip': 60.01310348510742, 'timing/data': 0.0006092949770390987, 'timing/model': 1.704401439987123, 'learning_rate/qwen_vl_interface': 9.966264873142641e-06, 'learning_rate/action_model': 9.96626487314264e-05, 'epoch': 0.11}) +[2026-06-23 17:05:32,393][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.38677847385406494, 'train/grad_norm_pre_clip': 106.19271087646484, 'timing/data': 0.0008292780257761478, 'timing/model': 1.7110847160220146, 'learning_rate/qwen_vl_interface': 9.965901590515018e-06, 'learning_rate/action_model': 9.965901590515018e-05, 'epoch': 0.11}) +[2026-06-23 17:05:35,746][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.35394200682640076, 'train/grad_norm_pre_clip': 44.44528579711914, 'timing/data': 0.0006477870047092438, 'timing/model': 1.700935231987387, 'learning_rate/qwen_vl_interface': 9.965536369358386e-06, 'learning_rate/action_model': 9.965536369358386e-05, 'epoch': 0.11}) +[2026-06-23 17:05:39,114][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.487061470746994, 'train/grad_norm_pre_clip': 69.4893798828125, 'timing/data': 0.0006042150780558586, 'timing/model': 1.7107294041197747, 'learning_rate/qwen_vl_interface': 9.965169209822874e-06, 'learning_rate/action_model': 9.965169209822874e-05, 'epoch': 0.11}) +[2026-06-23 17:05:42,484][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.343657523393631, 'train/grad_norm_pre_clip': 44.857337951660156, 'timing/data': 0.0005889171734452248, 'timing/model': 1.7118569898884743, 'learning_rate/qwen_vl_interface': 9.96480011205941e-06, 'learning_rate/action_model': 9.964800112059409e-05, 'epoch': 0.11}) +[2026-06-23 17:05:45,854][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.3451688885688782, 'train/grad_norm_pre_clip': 65.29331970214844, 'timing/data': 0.0006003379821777344, 'timing/model': 1.7122016549110413, 'learning_rate/qwen_vl_interface': 9.964429076219711e-06, 'learning_rate/action_model': 9.964429076219711e-05, 'epoch': 0.11}) +[2026-06-23 17:05:49,218][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.43938660621643066, 'train/grad_norm_pre_clip': 55.79829788208008, 'timing/data': 0.0005832258611917496, 'timing/model': 1.706097684102133, 'learning_rate/qwen_vl_interface': 9.964056102456299e-06, 'learning_rate/action_model': 9.964056102456299e-05, 'epoch': 0.11}) +[2026-06-23 17:05:52,571][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.3104473054409027, 'train/grad_norm_pre_clip': 40.888031005859375, 'timing/data': 0.0005969719495624304, 'timing/model': 1.7047255530487746, 'learning_rate/qwen_vl_interface': 9.963681190922489e-06, 'learning_rate/action_model': 9.963681190922488e-05, 'epoch': 0.11}) +[2026-06-23 17:05:55,933][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.48369312286376953, 'train/grad_norm_pre_clip': 60.39799118041992, 'timing/data': 0.0006021810695528984, 'timing/model': 1.7093328179325908, 'learning_rate/qwen_vl_interface': 9.963304341772391e-06, 'learning_rate/action_model': 9.963304341772391e-05, 'epoch': 0.11}) +[2026-06-23 17:05:59,289][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.4810330271720886, 'train/grad_norm_pre_clip': 92.50286865234375, 'timing/data': 0.0006028430070728064, 'timing/model': 1.7073602590244263, 'learning_rate/qwen_vl_interface': 9.962925555160919e-06, 'learning_rate/action_model': 9.962925555160918e-05, 'epoch': 0.11}) +[2026-06-23 17:06:02,649][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.4808829128742218, 'train/grad_norm_pre_clip': 93.08790588378906, 'timing/data': 0.0006431289948523045, 'timing/model': 1.7061256268061697, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.11}) +[2026-06-23 17:06:06,010][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.368999719619751, 'train/grad_norm_pre_clip': 78.04293823242188, 'timing/data': 0.0006002380978316069, 'timing/model': 1.7126983450725675, 'learning_rate/qwen_vl_interface': 9.96216217017745e-06, 'learning_rate/action_model': 9.962162170177451e-05, 'epoch': 0.11}) +[2026-06-23 17:06:09,355][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.37018442153930664, 'train/grad_norm_pre_clip': 23.2054443359375, 'timing/data': 0.0005842980463057756, 'timing/model': 1.6993062279652804, 'learning_rate/qwen_vl_interface': 9.961777572119258e-06, 'learning_rate/action_model': 9.961777572119258e-05, 'epoch': 0.12}) +[2026-06-23 17:06:12,704][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.3120993673801422, 'train/grad_norm_pre_clip': 8.130866050720215, 'timing/data': 0.000599586870521307, 'timing/model': 1.7028661749791354, 'learning_rate/qwen_vl_interface': 9.961391037227286e-06, 'learning_rate/action_model': 9.961391037227284e-05, 'epoch': 0.12}) +[2026-06-23 17:06:16,320][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 300 +[2026-06-23 17:07:50,034][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 300 +[2026-06-23 17:07:50,035][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 300 +[2026-06-23 17:09:03,051][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 300 +[2026-06-23 17:09:03,052][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.43851032853126526, 'train/grad_norm_pre_clip': 25.127275466918945, 'eval/action_loss/seconds': 93.71254952298477, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.18159280717372894, 'eval/latency_2/loss': 0.18159280717372894, 'eval/flappy/loss': 0.18159280717372894, 'eval/flappy/latency_2/loss': 0.18159280717372894, 'eval/flap_precision': 0.775625, 'eval/flap_recall': 0.8373819163292847, 'eval/flap_f1': 0.8053212199870213, 'eval/flap_rate_model': 0.2222222222222222, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8053212199870213, 'eval/latency_2/flap_precision': 0.775625, 'eval/latency_2/flap_recall': 0.8373819163292847, 'eval/latency_2/flap_f1': 0.8053212199870213, 'eval/latency_2/flap_rate_model': 0.2222222222222222, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8053212199870213, 'eval/action_classification/data_wait_seconds': 0.9276175498962402, 'eval/action_classification/predict_seconds': 71.94412231445312, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0007548488210886717, 'timing/model': 1.7051787979435176, 'learning_rate/qwen_vl_interface': 9.961002565660419e-06, 'learning_rate/action_model': 9.961002565660419e-05, 'epoch': 0.12}) +[2026-06-23 17:09:06,413][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.4329957365989685, 'train/grad_norm_pre_clip': 17.04758644104004, 'timing/data': 0.0005876037757843733, 'timing/model': 1.70606261305511, 'learning_rate/qwen_vl_interface': 9.96061215757835e-06, 'learning_rate/action_model': 9.960612157578349e-05, 'epoch': 0.12}) +[2026-06-23 17:09:09,767][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.27651703357696533, 'train/grad_norm_pre_clip': 10.507758140563965, 'timing/data': 0.0005945269949734211, 'timing/model': 1.6989783409517258, 'learning_rate/qwen_vl_interface': 9.96021981314156e-06, 'learning_rate/action_model': 9.96021981314156e-05, 'epoch': 0.12}) +[2026-06-23 17:09:13,115][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.231602281332016, 'train/grad_norm_pre_clip': 10.867465019226074, 'timing/data': 0.0005946571473032236, 'timing/model': 1.7033211688976735, 'learning_rate/qwen_vl_interface': 9.959825532511322e-06, 'learning_rate/action_model': 9.959825532511322e-05, 'epoch': 0.12}) +[2026-06-23 17:09:16,464][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.4091424345970154, 'train/grad_norm_pre_clip': 40.038169860839844, 'timing/data': 0.0006127411033958197, 'timing/model': 1.6948673529550433, 'learning_rate/qwen_vl_interface': 9.959429315849715e-06, 'learning_rate/action_model': 9.959429315849715e-05, 'epoch': 0.12}) +[2026-06-23 17:09:19,832][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.3509543240070343, 'train/grad_norm_pre_clip': 17.109689712524414, 'timing/data': 0.0005876938812434673, 'timing/model': 1.7149628999177366, 'learning_rate/qwen_vl_interface': 9.959031163319606e-06, 'learning_rate/action_model': 9.959031163319607e-05, 'epoch': 0.12}) +[2026-06-23 17:09:23,196][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.4118976294994354, 'train/grad_norm_pre_clip': 34.25270462036133, 'timing/data': 0.0007447798270732164, 'timing/model': 1.7049454969819635, 'learning_rate/qwen_vl_interface': 9.958631075084664e-06, 'learning_rate/action_model': 9.958631075084664e-05, 'epoch': 0.12}) +[2026-06-23 17:09:26,856][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.25973737239837646, 'train/grad_norm_pre_clip': 6.855932712554932, 'timing/data': 0.0005960199050605297, 'timing/model': 2.0052733421325684, 'learning_rate/qwen_vl_interface': 9.958229051309346e-06, 'learning_rate/action_model': 9.958229051309346e-05, 'epoch': 0.12}) +[2026-06-23 17:09:30,217][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.30744117498397827, 'train/grad_norm_pre_clip': 69.0849380493164, 'timing/data': 0.0005813820753246546, 'timing/model': 1.7061036890372634, 'learning_rate/qwen_vl_interface': 9.95782509215891e-06, 'learning_rate/action_model': 9.95782509215891e-05, 'epoch': 0.12}) +[2026-06-23 17:09:33,583][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.4313088655471802, 'train/grad_norm_pre_clip': 63.13450241088867, 'timing/data': 0.0005882848054170609, 'timing/model': 1.7088660378940403, 'learning_rate/qwen_vl_interface': 9.95741919779941e-06, 'learning_rate/action_model': 9.95741919779941e-05, 'epoch': 0.12}) +[2026-06-23 17:09:36,947][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.20820851624011993, 'train/grad_norm_pre_clip': 20.146080017089844, 'timing/data': 0.0005954878870397806, 'timing/model': 1.7067302539944649, 'learning_rate/qwen_vl_interface': 9.957011368397692e-06, 'learning_rate/action_model': 9.95701136839769e-05, 'epoch': 0.12}) +[2026-06-23 17:09:40,304][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.5622063279151917, 'train/grad_norm_pre_clip': 138.7078094482422, 'timing/data': 0.0005890470929443836, 'timing/model': 1.7031886379700154, 'learning_rate/qwen_vl_interface': 9.956601604121398e-06, 'learning_rate/action_model': 9.956601604121397e-05, 'epoch': 0.12}) +[2026-06-23 17:09:43,656][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.6144891977310181, 'train/grad_norm_pre_clip': 147.2447509765625, 'timing/data': 0.0008148010820150375, 'timing/model': 1.7015908739995211, 'learning_rate/qwen_vl_interface': 9.95618990513897e-06, 'learning_rate/action_model': 9.95618990513897e-05, 'epoch': 0.12}) +[2026-06-23 17:09:47,016][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.49914857745170593, 'train/grad_norm_pre_clip': 97.75347900390625, 'timing/data': 0.00059477798640728, 'timing/model': 1.7044807630591094, 'learning_rate/qwen_vl_interface': 9.95577627161964e-06, 'learning_rate/action_model': 9.95577627161964e-05, 'epoch': 0.12}) +[2026-06-23 17:09:50,371][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.526696503162384, 'train/grad_norm_pre_clip': 113.2710189819336, 'timing/data': 0.0006351140327751637, 'timing/model': 1.7033936779480428, 'learning_rate/qwen_vl_interface': 9.955360703733438e-06, 'learning_rate/action_model': 9.955360703733437e-05, 'epoch': 0.12}) +[2026-06-23 17:09:53,734][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.3680781126022339, 'train/grad_norm_pre_clip': 19.147043228149414, 'timing/data': 0.0005983950104564428, 'timing/model': 1.7102851660456508, 'learning_rate/qwen_vl_interface': 9.954943201651187e-06, 'learning_rate/action_model': 9.954943201651187e-05, 'epoch': 0.12}) +[2026-06-23 17:09:57,087][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.2562641203403473, 'train/grad_norm_pre_clip': 60.845481872558594, 'timing/data': 0.0005933749489486217, 'timing/model': 1.701439102878794, 'learning_rate/qwen_vl_interface': 9.954523765544507e-06, 'learning_rate/action_model': 9.954523765544508e-05, 'epoch': 0.12}) +[2026-06-23 17:10:00,441][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.40936726331710815, 'train/grad_norm_pre_clip': 128.31626892089844, 'timing/data': 0.0005992061924189329, 'timing/model': 1.708126086043194, 'learning_rate/qwen_vl_interface': 9.954102395585814e-06, 'learning_rate/action_model': 9.954102395585813e-05, 'epoch': 0.12}) +[2026-06-23 17:10:03,803][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.31531479954719543, 'train/grad_norm_pre_clip': 45.81024169921875, 'timing/data': 0.0006258261855691671, 'timing/model': 1.7114369459450245, 'learning_rate/qwen_vl_interface': 9.953679091948313e-06, 'learning_rate/action_model': 9.953679091948314e-05, 'epoch': 0.12}) +[2026-06-23 17:10:07,167][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.450447142124176, 'train/grad_norm_pre_clip': 19.153837203979492, 'timing/data': 0.0006090949755162001, 'timing/model': 1.7069172519259155, 'learning_rate/qwen_vl_interface': 9.953253854806013e-06, 'learning_rate/action_model': 9.953253854806012e-05, 'epoch': 0.12}) +[2026-06-23 17:10:10,518][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.27293115854263306, 'train/grad_norm_pre_clip': 62.301143646240234, 'timing/data': 0.0006144840735942125, 'timing/model': 1.70007444312796, 'learning_rate/qwen_vl_interface': 9.95282668433371e-06, 'learning_rate/action_model': 9.952826684333709e-05, 'epoch': 0.12}) +[2026-06-23 17:10:13,875][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.3957611918449402, 'train/grad_norm_pre_clip': 69.55996704101562, 'timing/data': 0.0005968508776277304, 'timing/model': 1.7014267228078097, 'learning_rate/qwen_vl_interface': 9.952397580706998e-06, 'learning_rate/action_model': 9.952397580706998e-05, 'epoch': 0.12}) +[2026-06-23 17:10:17,476][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.3311787247657776, 'train/grad_norm_pre_clip': 9.742951393127441, 'timing/data': 0.0007878909818828106, 'timing/model': 1.9446361190639436, 'learning_rate/qwen_vl_interface': 9.951966544102267e-06, 'learning_rate/action_model': 9.951966544102266e-05, 'epoch': 0.12}) +[2026-06-23 17:10:20,838][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.32696062326431274, 'train/grad_norm_pre_clip': 9.43967056274414, 'timing/data': 0.0006033438257873058, 'timing/model': 1.7111493591219187, 'learning_rate/qwen_vl_interface': 9.951533574696696e-06, 'learning_rate/action_model': 9.951533574696697e-05, 'epoch': 0.13}) +[2026-06-23 17:10:24,199][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.251067578792572, 'train/grad_norm_pre_clip': 10.252644538879395, 'timing/data': 0.0006017808336764574, 'timing/model': 1.7022938870359212, 'learning_rate/qwen_vl_interface': 9.951098672668266e-06, 'learning_rate/action_model': 9.951098672668265e-05, 'epoch': 0.13}) +[2026-06-23 17:10:27,548][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.12383206933736801, 'train/grad_norm_pre_clip': 33.64324188232422, 'timing/data': 0.0006183420773595572, 'timing/model': 1.6994995360728353, 'learning_rate/qwen_vl_interface': 9.95066183819575e-06, 'learning_rate/action_model': 9.950661838195749e-05, 'epoch': 0.13}) +[2026-06-23 17:10:30,919][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.45323795080184937, 'train/grad_norm_pre_clip': 63.54175567626953, 'timing/data': 0.0005963100120425224, 'timing/model': 1.7160112501587719, 'learning_rate/qwen_vl_interface': 9.95022307145871e-06, 'learning_rate/action_model': 9.950223071458708e-05, 'epoch': 0.13}) +[2026-06-23 17:10:34,283][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.2845188081264496, 'train/grad_norm_pre_clip': 19.457103729248047, 'timing/data': 0.0006706600543111563, 'timing/model': 1.7083795780781657, 'learning_rate/qwen_vl_interface': 9.949782372637507e-06, 'learning_rate/action_model': 9.949782372637506e-05, 'epoch': 0.13}) +[2026-06-23 17:10:37,643][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.2856021821498871, 'train/grad_norm_pre_clip': 121.67176818847656, 'timing/data': 0.0005977030377835035, 'timing/model': 1.7059367168694735, 'learning_rate/qwen_vl_interface': 9.949339741913297e-06, 'learning_rate/action_model': 9.949339741913297e-05, 'epoch': 0.13}) +[2026-06-23 17:10:41,007][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.3777713179588318, 'train/grad_norm_pre_clip': 123.40208435058594, 'timing/data': 0.0006117690354585648, 'timing/model': 1.7044066300150007, 'learning_rate/qwen_vl_interface': 9.948895179468032e-06, 'learning_rate/action_model': 9.948895179468031e-05, 'epoch': 0.13}) +[2026-06-23 17:10:44,367][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.38921383023262024, 'train/grad_norm_pre_clip': 83.72684478759766, 'timing/data': 0.0005995968822389841, 'timing/model': 1.7056963630020618, 'learning_rate/qwen_vl_interface': 9.948448685484448e-06, 'learning_rate/action_model': 9.948448685484447e-05, 'epoch': 0.13}) +[2026-06-23 17:10:47,723][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.36861518025398254, 'train/grad_norm_pre_clip': 75.0794448852539, 'timing/data': 0.0006720928940922022, 'timing/model': 1.7047285430599004, 'learning_rate/qwen_vl_interface': 9.948000260146087e-06, 'learning_rate/action_model': 9.948000260146087e-05, 'epoch': 0.13}) +[2026-06-23 17:10:51,085][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.25796177983283997, 'train/grad_norm_pre_clip': 34.066734313964844, 'timing/data': 0.0007000849582254887, 'timing/model': 1.7098480809945613, 'learning_rate/qwen_vl_interface': 9.947549903637277e-06, 'learning_rate/action_model': 9.947549903637278e-05, 'epoch': 0.13}) +[2026-06-23 17:10:54,458][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.34207257628440857, 'train/grad_norm_pre_clip': 58.698089599609375, 'timing/data': 0.0006005889736115932, 'timing/model': 1.7125014860648662, 'learning_rate/qwen_vl_interface': 9.947097616143145e-06, 'learning_rate/action_model': 9.947097616143144e-05, 'epoch': 0.13}) +[2026-06-23 17:10:57,820][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.24523919820785522, 'train/grad_norm_pre_clip': 60.19343185424805, 'timing/data': 0.0006769320461899042, 'timing/model': 1.7039020960219204, 'learning_rate/qwen_vl_interface': 9.946643397849607e-06, 'learning_rate/action_model': 9.946643397849607e-05, 'epoch': 0.13}) +[2026-06-23 17:11:01,176][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.3569146692752838, 'train/grad_norm_pre_clip': 57.26567077636719, 'timing/data': 0.0007310330402106047, 'timing/model': 1.7059553631115705, 'learning_rate/qwen_vl_interface': 9.946187248943375e-06, 'learning_rate/action_model': 9.946187248943374e-05, 'epoch': 0.13}) +[2026-06-23 17:11:04,539][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.38171494007110596, 'train/grad_norm_pre_clip': 76.21602630615234, 'timing/data': 0.0005995158571749926, 'timing/model': 1.7082497240044177, 'learning_rate/qwen_vl_interface': 9.945729169611957e-06, 'learning_rate/action_model': 9.945729169611958e-05, 'epoch': 0.13}) +[2026-06-23 17:11:08,009][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.39115259051322937, 'train/grad_norm_pre_clip': 124.56246185302734, 'timing/data': 0.0006098458543419838, 'timing/model': 1.7120102578774095, 'learning_rate/qwen_vl_interface': 9.94526916004365e-06, 'learning_rate/action_model': 9.94526916004365e-05, 'epoch': 0.13}) +[2026-06-23 17:11:11,373][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.3784552812576294, 'train/grad_norm_pre_clip': 83.7375717163086, 'timing/data': 0.0006271181628108025, 'timing/model': 1.7105338019318879, 'learning_rate/qwen_vl_interface': 9.944807220427549e-06, 'learning_rate/action_model': 9.944807220427548e-05, 'epoch': 0.13}) +[2026-06-23 17:11:14,736][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.4092133343219757, 'train/grad_norm_pre_clip': 80.05195617675781, 'timing/data': 0.0005928939208388329, 'timing/model': 1.7090976019389927, 'learning_rate/qwen_vl_interface': 9.944343350953534e-06, 'learning_rate/action_model': 9.944343350953533e-05, 'epoch': 0.13}) +[2026-06-23 17:11:18,092][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.48398372530937195, 'train/grad_norm_pre_clip': 55.968650817871094, 'timing/data': 0.0006099969614297152, 'timing/model': 1.6997043211013079, 'learning_rate/qwen_vl_interface': 9.94387755181229e-06, 'learning_rate/action_model': 9.94387755181229e-05, 'epoch': 0.13}) +[2026-06-23 17:11:21,472][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.38428837060928345, 'train/grad_norm_pre_clip': 40.00684356689453, 'timing/data': 0.0005965710151940584, 'timing/model': 1.7138562109321356, 'learning_rate/qwen_vl_interface': 9.943409823195289e-06, 'learning_rate/action_model': 9.943409823195288e-05, 'epoch': 0.13}) +[2026-06-23 17:11:24,838][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.2595016360282898, 'train/grad_norm_pre_clip': 24.900510787963867, 'timing/data': 0.0007065169047564268, 'timing/model': 1.7105663190595806, 'learning_rate/qwen_vl_interface': 9.942940165294794e-06, 'learning_rate/action_model': 9.942940165294793e-05, 'epoch': 0.13}) +[2026-06-23 17:11:28,211][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.32096803188323975, 'train/grad_norm_pre_clip': 16.77671241760254, 'timing/data': 0.0006376081146299839, 'timing/model': 1.713731394149363, 'learning_rate/qwen_vl_interface': 9.942468578303865e-06, 'learning_rate/action_model': 9.942468578303865e-05, 'epoch': 0.13}) +[2026-06-23 17:11:31,574][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.39236918091773987, 'train/grad_norm_pre_clip': 59.29166793823242, 'timing/data': 0.0005876538343727589, 'timing/model': 1.7066274338867515, 'learning_rate/qwen_vl_interface': 9.941995062416354e-06, 'learning_rate/action_model': 9.941995062416354e-05, 'epoch': 0.13}) +[2026-06-23 17:11:34,948][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.26942676305770874, 'train/grad_norm_pre_clip': 65.95629119873047, 'timing/data': 0.0006064299959689379, 'timing/model': 1.7127554579637945, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.13}) +[2026-06-23 17:11:38,328][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.2742156684398651, 'train/grad_norm_pre_clip': 54.54932403564453, 'timing/data': 0.0005914818029850721, 'timing/model': 1.7195931649766862, 'learning_rate/qwen_vl_interface': 9.941042244730954e-06, 'learning_rate/action_model': 9.941042244730955e-05, 'epoch': 0.13}) +[2026-06-23 17:11:41,753][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.19829407334327698, 'train/grad_norm_pre_clip': 12.879670143127441, 'timing/data': 0.000782631104812026, 'timing/model': 1.7137399048078805, 'learning_rate/qwen_vl_interface': 9.940562943324736e-06, 'learning_rate/action_model': 9.940562943324735e-05, 'epoch': 0.13}) +[2026-06-23 17:11:45,134][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.470192551612854, 'train/grad_norm_pre_clip': 60.7548942565918, 'timing/data': 0.0006633868906646967, 'timing/model': 1.7166103080380708, 'learning_rate/qwen_vl_interface': 9.940081713805266e-06, 'learning_rate/action_model': 9.940081713805266e-05, 'epoch': 0.13}) +[2026-06-23 17:11:48,508][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.4544975459575653, 'train/grad_norm_pre_clip': 98.80607604980469, 'timing/data': 0.0007610500324517488, 'timing/model': 1.7089100279845297, 'learning_rate/qwen_vl_interface': 9.939598556370366e-06, 'learning_rate/action_model': 9.939598556370365e-05, 'epoch': 0.14}) +[2026-06-23 17:11:51,882][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.6415185332298279, 'train/grad_norm_pre_clip': 99.40989685058594, 'timing/data': 0.0006358751561492682, 'timing/model': 1.7124088509008288, 'learning_rate/qwen_vl_interface': 9.93911347121864e-06, 'learning_rate/action_model': 9.939113471218639e-05, 'epoch': 0.14}) +[2026-06-23 17:11:55,351][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.23721903562545776, 'train/grad_norm_pre_clip': 32.15178298950195, 'timing/data': 0.0006072709802538157, 'timing/model': 1.8056591339409351, 'learning_rate/qwen_vl_interface': 9.938626458549488e-06, 'learning_rate/action_model': 9.938626458549488e-05, 'epoch': 0.14}) +[2026-06-23 17:11:58,718][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.28581860661506653, 'train/grad_norm_pre_clip': 10.087644577026367, 'timing/data': 0.0005931740161031485, 'timing/model': 1.7138860940467566, 'learning_rate/qwen_vl_interface': 9.938137518563106e-06, 'learning_rate/action_model': 9.938137518563106e-05, 'epoch': 0.14}) +[2026-06-23 17:12:02,092][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.39889204502105713, 'train/grad_norm_pre_clip': 59.997581481933594, 'timing/data': 0.0007154939230531454, 'timing/model': 1.7112994519993663, 'learning_rate/qwen_vl_interface': 9.937646651460475e-06, 'learning_rate/action_model': 9.937646651460474e-05, 'epoch': 0.14}) +[2026-06-23 17:12:05,464][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.23844225704669952, 'train/grad_norm_pre_clip': 31.097900390625, 'timing/data': 0.0006485790945589542, 'timing/model': 1.7127750639338046, 'learning_rate/qwen_vl_interface': 9.937153857443375e-06, 'learning_rate/action_model': 9.937153857443374e-05, 'epoch': 0.14}) +[2026-06-23 17:12:08,837][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.30416789650917053, 'train/grad_norm_pre_clip': 39.36354446411133, 'timing/data': 0.0006840350106358528, 'timing/model': 1.712669595144689, 'learning_rate/qwen_vl_interface': 9.936659136714372e-06, 'learning_rate/action_model': 9.936659136714371e-05, 'epoch': 0.14}) +[2026-06-23 17:12:12,202][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.4689320921897888, 'train/grad_norm_pre_clip': 26.700063705444336, 'timing/data': 0.0006605410017073154, 'timing/model': 1.703534189146012, 'learning_rate/qwen_vl_interface': 9.936162489476827e-06, 'learning_rate/action_model': 9.936162489476828e-05, 'epoch': 0.14}) +[2026-06-23 17:12:15,557][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.2491922229528427, 'train/grad_norm_pre_clip': 27.53960418701172, 'timing/data': 0.0006715420167893171, 'timing/model': 1.7022120931651443, 'learning_rate/qwen_vl_interface': 9.935663915934899e-06, 'learning_rate/action_model': 9.935663915934898e-05, 'epoch': 0.14}) +[2026-06-23 17:12:18,919][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.3206256926059723, 'train/grad_norm_pre_clip': 21.12371063232422, 'timing/data': 0.0005982338916510344, 'timing/model': 1.702057623071596, 'learning_rate/qwen_vl_interface': 9.935163416293525e-06, 'learning_rate/action_model': 9.935163416293525e-05, 'epoch': 0.14}) +[2026-06-23 17:12:22,283][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.3386613726615906, 'train/grad_norm_pre_clip': 38.76872634887695, 'timing/data': 0.0006254250183701515, 'timing/model': 1.7091288180090487, 'learning_rate/qwen_vl_interface': 9.934660990758445e-06, 'learning_rate/action_model': 9.934660990758445e-05, 'epoch': 0.14}) +[2026-06-23 17:12:25,659][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.3788282871246338, 'train/grad_norm_pre_clip': 16.466033935546875, 'timing/data': 0.0006722428370267153, 'timing/model': 1.7071379648987204, 'learning_rate/qwen_vl_interface': 9.934156639536186e-06, 'learning_rate/action_model': 9.934156639536187e-05, 'epoch': 0.14}) +[2026-06-23 17:12:29,033][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.3820685148239136, 'train/grad_norm_pre_clip': 24.51034164428711, 'timing/data': 0.0007990319281816483, 'timing/model': 1.710709925973788, 'learning_rate/qwen_vl_interface': 9.93365036283407e-06, 'learning_rate/action_model': 9.93365036283407e-05, 'epoch': 0.14}) +[2026-06-23 17:12:32,401][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.2969837486743927, 'train/grad_norm_pre_clip': 52.64777755737305, 'timing/data': 0.0005954890511929989, 'timing/model': 1.7135044960305095, 'learning_rate/qwen_vl_interface': 9.933142160860208e-06, 'learning_rate/action_model': 9.933142160860208e-05, 'epoch': 0.14}) +[2026-06-23 17:12:35,768][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.276681512594223, 'train/grad_norm_pre_clip': 44.67169952392578, 'timing/data': 0.00059477798640728, 'timing/model': 1.7099353191442788, 'learning_rate/qwen_vl_interface': 9.932632033823499e-06, 'learning_rate/action_model': 9.932632033823499e-05, 'epoch': 0.14}) +[2026-06-23 17:12:39,155][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.2860128581523895, 'train/grad_norm_pre_clip': 57.43712615966797, 'timing/data': 0.0005898880772292614, 'timing/model': 1.724312821868807, 'learning_rate/qwen_vl_interface': 9.93211998193364e-06, 'learning_rate/action_model': 9.932119981933639e-05, 'epoch': 0.14}) +[2026-06-23 17:12:42,514][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.24566085636615753, 'train/grad_norm_pre_clip': 26.081329345703125, 'timing/data': 0.0005901788827031851, 'timing/model': 1.7057121419347823, 'learning_rate/qwen_vl_interface': 9.931606005401117e-06, 'learning_rate/action_model': 9.931606005401116e-05, 'epoch': 0.14}) +[2026-06-23 17:12:45,959][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.2971770167350769, 'train/grad_norm_pre_clip': 54.35209655761719, 'timing/data': 0.0008007038850337267, 'timing/model': 1.6999484510160983, 'learning_rate/qwen_vl_interface': 9.931090104437202e-06, 'learning_rate/action_model': 9.931090104437202e-05, 'epoch': 0.14}) +[2026-06-23 17:12:49,322][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.3708946108818054, 'train/grad_norm_pre_clip': 25.095792770385742, 'timing/data': 0.000582484994083643, 'timing/model': 1.7071319671813399, 'learning_rate/qwen_vl_interface': 9.930572279253967e-06, 'learning_rate/action_model': 9.930572279253967e-05, 'epoch': 0.14}) +[2026-06-23 17:12:52,698][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.35682979226112366, 'train/grad_norm_pre_clip': 57.134342193603516, 'timing/data': 0.0006633058656007051, 'timing/model': 1.7153299800120294, 'learning_rate/qwen_vl_interface': 9.930052530064269e-06, 'learning_rate/action_model': 9.93005253006427e-05, 'epoch': 0.14}) +[2026-06-23 17:12:56,067][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.2497212290763855, 'train/grad_norm_pre_clip': 54.73077392578125, 'timing/data': 0.0005966608878225088, 'timing/model': 1.7067600858863443, 'learning_rate/qwen_vl_interface': 9.929530857081757e-06, 'learning_rate/action_model': 9.929530857081757e-05, 'epoch': 0.14}) +[2026-06-23 17:12:59,433][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.21928398311138153, 'train/grad_norm_pre_clip': 32.49134826660156, 'timing/data': 0.0006033640820533037, 'timing/model': 1.7062491260003299, 'learning_rate/qwen_vl_interface': 9.92900726052087e-06, 'learning_rate/action_model': 9.929007260520869e-05, 'epoch': 0.14}) +[2026-06-23 17:13:02,803][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.269807904958725, 'train/grad_norm_pre_clip': 26.614086151123047, 'timing/data': 0.0006033140234649181, 'timing/model': 1.7130107588600367, 'learning_rate/qwen_vl_interface': 9.92848174059684e-06, 'learning_rate/action_model': 9.92848174059684e-05, 'epoch': 0.14}) +[2026-06-23 17:13:06,173][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.2252615988254547, 'train/grad_norm_pre_clip': 42.29799270629883, 'timing/data': 0.0005956590175628662, 'timing/model': 1.7094745840877295, 'learning_rate/qwen_vl_interface': 9.92795429752569e-06, 'learning_rate/action_model': 9.927954297525689e-05, 'epoch': 0.14}) +[2026-06-23 17:13:09,544][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.20670217275619507, 'train/grad_norm_pre_clip': 11.264856338500977, 'timing/data': 0.000598113052546978, 'timing/model': 1.7112640149425715, 'learning_rate/qwen_vl_interface': 9.927424931524228e-06, 'learning_rate/action_model': 9.927424931524228e-05, 'epoch': 0.14}) +[2026-06-23 17:13:12,909][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.2892948091030121, 'train/grad_norm_pre_clip': 26.17967414855957, 'timing/data': 0.0006015507970005274, 'timing/model': 1.7052526590414345, 'learning_rate/qwen_vl_interface': 9.926893642810058e-06, 'learning_rate/action_model': 9.926893642810059e-05, 'epoch': 0.14}) +[2026-06-23 17:13:16,280][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.2440105527639389, 'train/grad_norm_pre_clip': 45.858360290527344, 'timing/data': 0.0005896380171179771, 'timing/model': 1.7115540080703795, 'learning_rate/qwen_vl_interface': 9.926360431601575e-06, 'learning_rate/action_model': 9.926360431601574e-05, 'epoch': 0.15}) +[2026-06-23 17:13:19,654][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.39159390330314636, 'train/grad_norm_pre_clip': 28.820831298828125, 'timing/data': 0.0006592790596187115, 'timing/model': 1.7112778779119253, 'learning_rate/qwen_vl_interface': 9.92582529811796e-06, 'learning_rate/action_model': 9.925825298117959e-05, 'epoch': 0.15}) +[2026-06-23 17:13:23,026][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.5308107137680054, 'train/grad_norm_pre_clip': 140.35157775878906, 'timing/data': 0.0006528969388455153, 'timing/model': 1.708267912035808, 'learning_rate/qwen_vl_interface': 9.925288242579184e-06, 'learning_rate/action_model': 9.925288242579184e-05, 'epoch': 0.15}) +[2026-06-23 17:13:26,382][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.46417683362960815, 'train/grad_norm_pre_clip': 161.78146362304688, 'timing/data': 0.0005996271502226591, 'timing/model': 1.7052363948896527, 'learning_rate/qwen_vl_interface': 9.924749265206015e-06, 'learning_rate/action_model': 9.924749265206015e-05, 'epoch': 0.15}) +[2026-06-23 17:13:29,746][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.49145475029945374, 'train/grad_norm_pre_clip': 177.3033447265625, 'timing/data': 0.0005944070871919394, 'timing/model': 1.7083775450009853, 'learning_rate/qwen_vl_interface': 9.924208366220003e-06, 'learning_rate/action_model': 9.924208366220002e-05, 'epoch': 0.15}) +[2026-06-23 17:13:33,115][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.332453191280365, 'train/grad_norm_pre_clip': 70.36971282958984, 'timing/data': 0.0006084630731493235, 'timing/model': 1.7102566750254482, 'learning_rate/qwen_vl_interface': 9.923665545843492e-06, 'learning_rate/action_model': 9.923665545843491e-05, 'epoch': 0.15}) +[2026-06-23 17:13:36,622][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.3573497533798218, 'train/grad_norm_pre_clip': 40.7102165222168, 'timing/data': 0.0005868128500878811, 'timing/model': 1.7072641919367015, 'learning_rate/qwen_vl_interface': 9.923120804299615e-06, 'learning_rate/action_model': 9.923120804299613e-05, 'epoch': 0.15}) +[2026-06-23 17:13:39,984][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.3273771107196808, 'train/grad_norm_pre_clip': 110.76751708984375, 'timing/data': 0.0006644190289080143, 'timing/model': 1.7081051610875875, 'learning_rate/qwen_vl_interface': 9.922574141812295e-06, 'learning_rate/action_model': 9.922574141812294e-05, 'epoch': 0.15}) +[2026-06-23 17:13:43,347][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.4188014566898346, 'train/grad_norm_pre_clip': 117.05384826660156, 'timing/data': 0.000704634003341198, 'timing/model': 1.7076280128676444, 'learning_rate/qwen_vl_interface': 9.922025558606242e-06, 'learning_rate/action_model': 9.922025558606242e-05, 'epoch': 0.15}) +[2026-06-23 17:13:46,720][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.37446287274360657, 'train/grad_norm_pre_clip': 40.56474304199219, 'timing/data': 0.0007608989253640175, 'timing/model': 1.7208493140060455, 'learning_rate/qwen_vl_interface': 9.92147505490696e-06, 'learning_rate/action_model': 9.92147505490696e-05, 'epoch': 0.15}) +[2026-06-23 17:13:50,083][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.2965506613254547, 'train/grad_norm_pre_clip': 49.90761184692383, 'timing/data': 0.0006568338721990585, 'timing/model': 1.703827050048858, 'learning_rate/qwen_vl_interface': 9.920922630940741e-06, 'learning_rate/action_model': 9.920922630940742e-05, 'epoch': 0.15}) +[2026-06-23 17:13:53,456][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.3147311806678772, 'train/grad_norm_pre_clip': 69.08500671386719, 'timing/data': 0.000593534903600812, 'timing/model': 1.7111219761427492, 'learning_rate/qwen_vl_interface': 9.920368286934664e-06, 'learning_rate/action_model': 9.920368286934663e-05, 'epoch': 0.15}) +[2026-06-23 17:13:56,820][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.41128718852996826, 'train/grad_norm_pre_clip': 134.3166046142578, 'timing/data': 0.0006013899110257626, 'timing/model': 1.706848445115611, 'learning_rate/qwen_vl_interface': 9.9198120231166e-06, 'learning_rate/action_model': 9.9198120231166e-05, 'epoch': 0.15}) +[2026-06-23 17:14:00,183][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.33730000257492065, 'train/grad_norm_pre_clip': 63.72977828979492, 'timing/data': 0.0006690172012895346, 'timing/model': 1.7085815099999309, 'learning_rate/qwen_vl_interface': 9.919253839715208e-06, 'learning_rate/action_model': 9.919253839715208e-05, 'epoch': 0.15}) +[2026-06-23 17:14:03,555][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.3177650570869446, 'train/grad_norm_pre_clip': 32.50474548339844, 'timing/data': 0.0005876040086150169, 'timing/model': 1.7094847068656236, 'learning_rate/qwen_vl_interface': 9.918693736959935e-06, 'learning_rate/action_model': 9.918693736959934e-05, 'epoch': 0.15}) +[2026-06-23 17:14:06,939][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.47329816222190857, 'train/grad_norm_pre_clip': 124.24641418457031, 'timing/data': 0.0008780499920248985, 'timing/model': 1.7203499609604478, 'learning_rate/qwen_vl_interface': 9.918131715081019e-06, 'learning_rate/action_model': 9.918131715081018e-05, 'epoch': 0.15}) +[2026-06-23 17:14:10,319][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.2365788370370865, 'train/grad_norm_pre_clip': 58.066932678222656, 'timing/data': 0.0005978329572826624, 'timing/model': 1.7157011351082474, 'learning_rate/qwen_vl_interface': 9.917567774309486e-06, 'learning_rate/action_model': 9.917567774309485e-05, 'epoch': 0.15}) +[2026-06-23 17:14:13,680][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.23978936672210693, 'train/grad_norm_pre_clip': 25.082691192626953, 'timing/data': 0.0005838971119374037, 'timing/model': 1.7019393369555473, 'learning_rate/qwen_vl_interface': 9.91700191487715e-06, 'learning_rate/action_model': 9.91700191487715e-05, 'epoch': 0.15}) +[2026-06-23 17:14:17,051][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.26156315207481384, 'train/grad_norm_pre_clip': 66.1079330444336, 'timing/data': 0.0005940159317106009, 'timing/model': 1.7107410239987075, 'learning_rate/qwen_vl_interface': 9.916434137016614e-06, 'learning_rate/action_model': 9.916434137016613e-05, 'epoch': 0.15}) +[2026-06-23 17:14:20,427][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.2877464294433594, 'train/grad_norm_pre_clip': 65.66661071777344, 'timing/data': 0.0006035938858985901, 'timing/model': 1.719257263932377, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.15}) +[2026-06-23 17:14:23,908][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.43242472410202026, 'train/grad_norm_pre_clip': 92.24323272705078, 'timing/data': 0.0006031629163771868, 'timing/model': 1.8278793871868402, 'learning_rate/qwen_vl_interface': 9.915292826945303e-06, 'learning_rate/action_model': 9.915292826945303e-05, 'epoch': 0.15}) +[2026-06-23 17:14:27,272][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.37621140480041504, 'train/grad_norm_pre_clip': 28.540109634399414, 'timing/data': 0.0006071301177144051, 'timing/model': 1.7058093440718949, 'learning_rate/qwen_vl_interface': 9.914719295203676e-06, 'learning_rate/action_model': 9.914719295203676e-05, 'epoch': 0.15}) +[2026-06-23 17:14:30,642][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.3237105906009674, 'train/grad_norm_pre_clip': 23.912992477416992, 'timing/data': 0.0005918620154261589, 'timing/model': 1.7113844989798963, 'learning_rate/qwen_vl_interface': 9.91414384597215e-06, 'learning_rate/action_model': 9.91414384597215e-05, 'epoch': 0.15}) +[2026-06-23 17:14:34,014][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.275785893201828, 'train/grad_norm_pre_clip': 68.94530487060547, 'timing/data': 0.0006650590803474188, 'timing/model': 1.7127820840105414, 'learning_rate/qwen_vl_interface': 9.913566479487267e-06, 'learning_rate/action_model': 9.913566479487267e-05, 'epoch': 0.15}) +[2026-06-23 17:14:37,375][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.37353017926216125, 'train/grad_norm_pre_clip': 64.30636596679688, 'timing/data': 0.0006114989519119263, 'timing/model': 1.7042973299976438, 'learning_rate/qwen_vl_interface': 9.912987195986366e-06, 'learning_rate/action_model': 9.912987195986365e-05, 'epoch': 0.15}) +[2026-06-23 17:14:40,751][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 400 +[2026-06-23 17:16:14,627][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 400 +[2026-06-23 17:16:14,628][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 400 +[2026-06-23 17:17:28,046][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 400 +[2026-06-23 17:17:28,047][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.3037523031234741, 'train/grad_norm_pre_clip': 23.278108596801758, 'eval/action_loss/seconds': 93.87360102613457, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.1598178595304489, 'eval/latency_2/loss': 0.1598178595304489, 'eval/flappy/loss': 0.1598178595304489, 'eval/flappy/latency_2/loss': 0.1598178595304489, 'eval/flap_precision': 0.8847077662129704, 'eval/flap_recall': 0.7456140350877193, 'eval/flap_f1': 0.8092273892347126, 'eval/flap_rate_model': 0.17347222222222222, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8092273892347127, 'eval/latency_2/flap_precision': 0.8847077662129704, 'eval/latency_2/flap_recall': 0.7456140350877193, 'eval/latency_2/flap_f1': 0.8092273892347126, 'eval/latency_2/flap_rate_model': 0.17347222222222222, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8092273892347127, 'eval/action_classification/data_wait_seconds': 1.0172370672225952, 'eval/action_classification/predict_seconds': 72.28152465820312, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005998669657856226, 'timing/model': 1.7156654298305511, 'learning_rate/qwen_vl_interface': 9.912405995707562e-06, 'learning_rate/action_model': 9.912405995707561e-05, 'epoch': 0.15}) +[2026-06-23 17:17:31,411][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.3245590627193451, 'train/grad_norm_pre_clip': 24.635366439819336, 'timing/data': 0.0008018771186470985, 'timing/model': 1.705029138829559, 'learning_rate/qwen_vl_interface': 9.911822878889767e-06, 'learning_rate/action_model': 9.911822878889767e-05, 'epoch': 0.16}) +[2026-06-23 17:17:34,758][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.2732900381088257, 'train/grad_norm_pre_clip': 31.399314880371094, 'timing/data': 0.0007508900016546249, 'timing/model': 1.6989628011360765, 'learning_rate/qwen_vl_interface': 9.911237845772681e-06, 'learning_rate/action_model': 9.911237845772681e-05, 'epoch': 0.16}) +[2026-06-23 17:17:38,104][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.2538231611251831, 'train/grad_norm_pre_clip': 64.4997787475586, 'timing/data': 0.000599847175180912, 'timing/model': 1.6973029719665647, 'learning_rate/qwen_vl_interface': 9.910650896596787e-06, 'learning_rate/action_model': 9.910650896596786e-05, 'epoch': 0.16}) +[2026-06-23 17:17:41,691][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.2309129685163498, 'train/grad_norm_pre_clip': 9.753166198730469, 'timing/data': 0.0005929640028625727, 'timing/model': 1.7064195501152426, 'learning_rate/qwen_vl_interface': 9.910062031603356e-06, 'learning_rate/action_model': 9.910062031603356e-05, 'epoch': 0.16}) +[2026-06-23 17:17:45,066][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.22124943137168884, 'train/grad_norm_pre_clip': 22.887039184570312, 'timing/data': 0.0006147550884634256, 'timing/model': 1.7111957229208201, 'learning_rate/qwen_vl_interface': 9.909471251034451e-06, 'learning_rate/action_model': 9.90947125103445e-05, 'epoch': 0.16}) +[2026-06-23 17:17:48,421][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.3772878348827362, 'train/grad_norm_pre_clip': 27.412206649780273, 'timing/data': 0.0005910210311412811, 'timing/model': 1.7050843900069594, 'learning_rate/qwen_vl_interface': 9.908878555132918e-06, 'learning_rate/action_model': 9.908878555132918e-05, 'epoch': 0.16}) +[2026-06-23 17:17:51,782][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.3108711540699005, 'train/grad_norm_pre_clip': 13.634121894836426, 'timing/data': 0.0006929619703441858, 'timing/model': 1.7058341479860246, 'learning_rate/qwen_vl_interface': 9.908283944142391e-06, 'learning_rate/action_model': 9.908283944142391e-05, 'epoch': 0.16}) +[2026-06-23 17:17:55,149][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.4147343039512634, 'train/grad_norm_pre_clip': 117.17620086669922, 'timing/data': 0.0006054381374269724, 'timing/model': 1.7077620509080589, 'learning_rate/qwen_vl_interface': 9.907687418307297e-06, 'learning_rate/action_model': 9.907687418307296e-05, 'epoch': 0.16}) +[2026-06-23 17:17:58,519][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.4290129542350769, 'train/grad_norm_pre_clip': 155.60330200195312, 'timing/data': 0.0006005789618939161, 'timing/model': 1.7099767511244863, 'learning_rate/qwen_vl_interface': 9.907088977872837e-06, 'learning_rate/action_model': 9.907088977872837e-05, 'epoch': 0.16}) +[2026-06-23 17:18:01,870][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.36880332231521606, 'train/grad_norm_pre_clip': 61.30170440673828, 'timing/data': 0.0006017000414431095, 'timing/model': 1.7007607540581375, 'learning_rate/qwen_vl_interface': 9.906488623085015e-06, 'learning_rate/action_model': 9.906488623085015e-05, 'epoch': 0.16}) +[2026-06-23 17:18:05,224][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.19713321328163147, 'train/grad_norm_pre_clip': 13.63925552368164, 'timing/data': 0.0005970019847154617, 'timing/model': 1.699050429975614, 'learning_rate/qwen_vl_interface': 9.905886354190611e-06, 'learning_rate/action_model': 9.90588635419061e-05, 'epoch': 0.16}) +[2026-06-23 17:18:08,578][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.27022725343704224, 'train/grad_norm_pre_clip': 39.549068450927734, 'timing/data': 0.0005947579629719257, 'timing/model': 1.7021659940946847, 'learning_rate/qwen_vl_interface': 9.905282171437192e-06, 'learning_rate/action_model': 9.905282171437192e-05, 'epoch': 0.16}) +[2026-06-23 17:18:11,939][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.42576566338539124, 'train/grad_norm_pre_clip': 21.85907554626465, 'timing/data': 0.0005937849637120962, 'timing/model': 1.7057976960204542, 'learning_rate/qwen_vl_interface': 9.904676075073121e-06, 'learning_rate/action_model': 9.904676075073121e-05, 'epoch': 0.16}) +[2026-06-23 17:18:15,297][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.3334692418575287, 'train/grad_norm_pre_clip': 41.733699798583984, 'timing/data': 0.0005981330759823322, 'timing/model': 1.7031564130447805, 'learning_rate/qwen_vl_interface': 9.904068065347536e-06, 'learning_rate/action_model': 9.904068065347536e-05, 'epoch': 0.16}) +[2026-06-23 17:18:18,657][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.2562237083911896, 'train/grad_norm_pre_clip': 86.58379364013672, 'timing/data': 0.0005910408217459917, 'timing/model': 1.7041360829025507, 'learning_rate/qwen_vl_interface': 9.90345814251037e-06, 'learning_rate/action_model': 9.90345814251037e-05, 'epoch': 0.16}) +[2026-06-23 17:18:22,008][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.4523438513278961, 'train/grad_norm_pre_clip': 147.37762451171875, 'timing/data': 0.000606179004535079, 'timing/model': 1.7033560769632459, 'learning_rate/qwen_vl_interface': 9.902846306812339e-06, 'learning_rate/action_model': 9.902846306812338e-05, 'epoch': 0.16}) +[2026-06-23 17:18:25,381][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.35647982358932495, 'train/grad_norm_pre_clip': 116.37924194335938, 'timing/data': 0.0006070300005376339, 'timing/model': 1.7155127678997815, 'learning_rate/qwen_vl_interface': 9.902232558504942e-06, 'learning_rate/action_model': 9.902232558504941e-05, 'epoch': 0.16}) +[2026-06-23 17:18:28,929][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.35620933771133423, 'train/grad_norm_pre_clip': 27.220306396484375, 'timing/data': 0.0005994259845465422, 'timing/model': 1.9000367019325495, 'learning_rate/qwen_vl_interface': 9.901616897840471e-06, 'learning_rate/action_model': 9.90161689784047e-05, 'epoch': 0.16}) +[2026-06-23 17:18:32,295][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.4879816770553589, 'train/grad_norm_pre_clip': 91.04881286621094, 'timing/data': 0.0006150060798972845, 'timing/model': 1.7100735979620367, 'learning_rate/qwen_vl_interface': 9.900999325072e-06, 'learning_rate/action_model': 9.900999325072001e-05, 'epoch': 0.16}) +[2026-06-23 17:18:35,660][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.5960503816604614, 'train/grad_norm_pre_clip': 111.4842300415039, 'timing/data': 0.0005809708964079618, 'timing/model': 1.7095773038454354, 'learning_rate/qwen_vl_interface': 9.900379840453392e-06, 'learning_rate/action_model': 9.900379840453392e-05, 'epoch': 0.16}) +[2026-06-23 17:18:39,030][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.4143749177455902, 'train/grad_norm_pre_clip': 102.23890686035156, 'timing/data': 0.0006576760206371546, 'timing/model': 1.710666720988229, 'learning_rate/qwen_vl_interface': 9.89975844423929e-06, 'learning_rate/action_model': 9.899758444239289e-05, 'epoch': 0.16}) +[2026-06-23 17:18:42,385][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.2570452094078064, 'train/grad_norm_pre_clip': 24.356586456298828, 'timing/data': 0.0006032439414411783, 'timing/model': 1.7028253071475774, 'learning_rate/qwen_vl_interface': 9.899135136685131e-06, 'learning_rate/action_model': 9.89913513668513e-05, 'epoch': 0.16}) +[2026-06-23 17:18:45,751][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.34218326210975647, 'train/grad_norm_pre_clip': 58.096744537353516, 'timing/data': 0.000604214845225215, 'timing/model': 1.7085674859117717, 'learning_rate/qwen_vl_interface': 9.898509918047127e-06, 'learning_rate/action_model': 9.898509918047126e-05, 'epoch': 0.16}) +[2026-06-23 17:18:49,113][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.3902508318424225, 'train/grad_norm_pre_clip': 104.24832916259766, 'timing/data': 0.0006053599063307047, 'timing/model': 1.7043091980740428, 'learning_rate/qwen_vl_interface': 9.897882788582286e-06, 'learning_rate/action_model': 9.897882788582287e-05, 'epoch': 0.16}) +[2026-06-23 17:18:52,476][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.4505801796913147, 'train/grad_norm_pre_clip': 66.66011047363281, 'timing/data': 0.000817940104752779, 'timing/model': 1.7036482580006123, 'learning_rate/qwen_vl_interface': 9.897253748548398e-06, 'learning_rate/action_model': 9.897253748548399e-05, 'epoch': 0.16}) +[2026-06-23 17:18:55,838][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.2666085660457611, 'train/grad_norm_pre_clip': 26.717769622802734, 'timing/data': 0.000812961021438241, 'timing/model': 1.7093912111595273, 'learning_rate/qwen_vl_interface': 9.896622798204035e-06, 'learning_rate/action_model': 9.896622798204035e-05, 'epoch': 0.16}) +[2026-06-23 17:18:59,199][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.44218823313713074, 'train/grad_norm_pre_clip': 60.65703201293945, 'timing/data': 0.0006070139352232218, 'timing/model': 1.7107510780915618, 'learning_rate/qwen_vl_interface': 9.89598993780856e-06, 'learning_rate/action_model': 9.89598993780856e-05, 'epoch': 0.17}) +[2026-06-23 17:19:02,563][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.3167746961116791, 'train/grad_norm_pre_clip': 71.44535827636719, 'timing/data': 0.0006001300644129515, 'timing/model': 1.714401006931439, 'learning_rate/qwen_vl_interface': 9.895355167622113e-06, 'learning_rate/action_model': 9.895355167622113e-05, 'epoch': 0.17}) +[2026-06-23 17:19:05,926][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.37722474336624146, 'train/grad_norm_pre_clip': 65.18465423583984, 'timing/data': 0.0007041261997073889, 'timing/model': 1.7047320960555226, 'learning_rate/qwen_vl_interface': 9.894718487905629e-06, 'learning_rate/action_model': 9.894718487905629e-05, 'epoch': 0.17}) +[2026-06-23 17:19:09,291][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.3376769423484802, 'train/grad_norm_pre_clip': 40.590206146240234, 'timing/data': 0.0006053501274436712, 'timing/model': 1.7099957158789039, 'learning_rate/qwen_vl_interface': 9.89407989892082e-06, 'learning_rate/action_model': 9.894079898920821e-05, 'epoch': 0.17}) +[2026-06-23 17:19:12,647][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.2808517813682556, 'train/grad_norm_pre_clip': 18.984020233154297, 'timing/data': 0.0005924461875110865, 'timing/model': 1.7044762189034373, 'learning_rate/qwen_vl_interface': 9.893439400930187e-06, 'learning_rate/action_model': 9.893439400930187e-05, 'epoch': 0.17}) +[2026-06-23 17:19:16,022][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.3582211434841156, 'train/grad_norm_pre_clip': 74.15484619140625, 'timing/data': 0.0006049589719623327, 'timing/model': 1.7137708910740912, 'learning_rate/qwen_vl_interface': 9.892796994197014e-06, 'learning_rate/action_model': 9.892796994197014e-05, 'epoch': 0.17}) +[2026-06-23 17:19:19,559][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.3950856924057007, 'train/grad_norm_pre_clip': 19.2675838470459, 'timing/data': 0.0005891388282179832, 'timing/model': 1.7044808198697865, 'learning_rate/qwen_vl_interface': 9.892152678985372e-06, 'learning_rate/action_model': 9.892152678985372e-05, 'epoch': 0.17}) +[2026-06-23 17:19:22,919][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.29732343554496765, 'train/grad_norm_pre_clip': 31.17203712463379, 'timing/data': 0.0006127739325165749, 'timing/model': 1.7082328849937767, 'learning_rate/qwen_vl_interface': 9.891506455560111e-06, 'learning_rate/action_model': 9.891506455560111e-05, 'epoch': 0.17}) +[2026-06-23 17:19:26,278][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.3813709020614624, 'train/grad_norm_pre_clip': 59.9197883605957, 'timing/data': 0.0006029051728546619, 'timing/model': 1.7063075781334192, 'learning_rate/qwen_vl_interface': 9.890858324186872e-06, 'learning_rate/action_model': 9.890858324186872e-05, 'epoch': 0.17}) +[2026-06-23 17:19:29,651][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.32503437995910645, 'train/grad_norm_pre_clip': 31.74932289123535, 'timing/data': 0.0008580051362514496, 'timing/model': 1.7121951321605593, 'learning_rate/qwen_vl_interface': 9.890208285132077e-06, 'learning_rate/action_model': 9.890208285132077e-05, 'epoch': 0.17}) +[2026-06-23 17:19:33,028][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.24094662070274353, 'train/grad_norm_pre_clip': 30.115102767944336, 'timing/data': 0.0007026528473943472, 'timing/model': 1.7149821911007166, 'learning_rate/qwen_vl_interface': 9.889556338662933e-06, 'learning_rate/action_model': 9.889556338662932e-05, 'epoch': 0.17}) +[2026-06-23 17:19:36,401][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.3764127492904663, 'train/grad_norm_pre_clip': 50.363014221191406, 'timing/data': 0.0006961508188396692, 'timing/model': 1.709382175002247, 'learning_rate/qwen_vl_interface': 9.888902485047427e-06, 'learning_rate/action_model': 9.888902485047427e-05, 'epoch': 0.17}) +[2026-06-23 17:19:39,780][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.32513338327407837, 'train/grad_norm_pre_clip': 76.08602142333984, 'timing/data': 0.0006824349984526634, 'timing/model': 1.7163566669914871, 'learning_rate/qwen_vl_interface': 9.888246724554337e-06, 'learning_rate/action_model': 9.888246724554337e-05, 'epoch': 0.17}) +[2026-06-23 17:19:43,152][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.2221822738647461, 'train/grad_norm_pre_clip': 40.4078483581543, 'timing/data': 0.0006886471528559923, 'timing/model': 1.7097939089871943, 'learning_rate/qwen_vl_interface': 9.887589057453222e-06, 'learning_rate/action_model': 9.88758905745322e-05, 'epoch': 0.17}) +[2026-06-23 17:19:46,523][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.2609810531139374, 'train/grad_norm_pre_clip': 57.40428161621094, 'timing/data': 0.0006064618937671185, 'timing/model': 1.7130581252276897, 'learning_rate/qwen_vl_interface': 9.88692948401442e-06, 'learning_rate/action_model': 9.88692948401442e-05, 'epoch': 0.17}) +[2026-06-23 17:19:49,897][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.2648264169692993, 'train/grad_norm_pre_clip': 56.52412033081055, 'timing/data': 0.0006723660044372082, 'timing/model': 1.7087640699464828, 'learning_rate/qwen_vl_interface': 9.88626800450906e-06, 'learning_rate/action_model': 9.88626800450906e-05, 'epoch': 0.17}) +[2026-06-23 17:19:53,271][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.416705846786499, 'train/grad_norm_pre_clip': 93.40370178222656, 'timing/data': 0.000694196904078126, 'timing/model': 1.7142379970755428, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.17}) +[2026-06-23 17:19:56,643][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.35621464252471924, 'train/grad_norm_pre_clip': 61.25831604003906, 'timing/data': 0.0008568039629608393, 'timing/model': 1.7122783979866654, 'learning_rate/qwen_vl_interface': 9.884939328387085e-06, 'learning_rate/action_model': 9.884939328387084e-05, 'epoch': 0.17}) +[2026-06-23 17:20:00,007][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.28365209698677063, 'train/grad_norm_pre_clip': 63.822296142578125, 'timing/data': 0.000607434194535017, 'timing/model': 1.7081900318153203, 'learning_rate/qwen_vl_interface': 9.884272132316638e-06, 'learning_rate/action_model': 9.884272132316637e-05, 'epoch': 0.17}) +[2026-06-23 17:20:03,376][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.2425352931022644, 'train/grad_norm_pre_clip': 23.516979217529297, 'timing/data': 0.0006046479102224112, 'timing/model': 1.703909554053098, 'learning_rate/qwen_vl_interface': 9.883603031271968e-06, 'learning_rate/action_model': 9.883603031271969e-05, 'epoch': 0.17}) +[2026-06-23 17:20:06,886][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.3270495533943176, 'train/grad_norm_pre_clip': 47.35828399658203, 'timing/data': 0.0006049089133739471, 'timing/model': 1.8622147249989212, 'learning_rate/qwen_vl_interface': 9.882932025528122e-06, 'learning_rate/action_model': 9.882932025528121e-05, 'epoch': 0.17}) +[2026-06-23 17:20:10,243][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.28753146529197693, 'train/grad_norm_pre_clip': 66.39427185058594, 'timing/data': 0.0005972839426249266, 'timing/model': 1.7077613780274987, 'learning_rate/qwen_vl_interface': 9.882259115360918e-06, 'learning_rate/action_model': 9.882259115360917e-05, 'epoch': 0.17}) +[2026-06-23 17:20:13,616][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.26633501052856445, 'train/grad_norm_pre_clip': 59.848167419433594, 'timing/data': 0.000600811094045639, 'timing/model': 1.7092511649243534, 'learning_rate/qwen_vl_interface': 9.881584301046968e-06, 'learning_rate/action_model': 9.881584301046968e-05, 'epoch': 0.17}) +[2026-06-23 17:20:16,984][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.36278417706489563, 'train/grad_norm_pre_clip': 36.9636344909668, 'timing/data': 0.0006096179131418467, 'timing/model': 1.705836890032515, 'learning_rate/qwen_vl_interface': 9.880907582863662e-06, 'learning_rate/action_model': 9.880907582863662e-05, 'epoch': 0.17}) +[2026-06-23 17:20:20,347][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.38879188895225525, 'train/grad_norm_pre_clip': 49.38894271850586, 'timing/data': 0.0006114908028393984, 'timing/model': 1.7063218848779798, 'learning_rate/qwen_vl_interface': 9.880228961089175e-06, 'learning_rate/action_model': 9.880228961089175e-05, 'epoch': 0.17}) +[2026-06-23 17:20:23,710][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.21028538048267365, 'train/grad_norm_pre_clip': 7.639817714691162, 'timing/data': 0.000608846778050065, 'timing/model': 1.7015302958898246, 'learning_rate/qwen_vl_interface': 9.87954843600246e-06, 'learning_rate/action_model': 9.879548436002459e-05, 'epoch': 0.17}) +[2026-06-23 17:20:27,070][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.23405368626117706, 'train/grad_norm_pre_clip': 48.97783660888672, 'timing/data': 0.0006029750220477581, 'timing/model': 1.7051645431201905, 'learning_rate/qwen_vl_interface': 9.878866007883256e-06, 'learning_rate/action_model': 9.878866007883257e-05, 'epoch': 0.18}) +[2026-06-23 17:20:30,431][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.3639237582683563, 'train/grad_norm_pre_clip': 53.97774887084961, 'timing/data': 0.0005945090670138597, 'timing/model': 1.701069155940786, 'learning_rate/qwen_vl_interface': 9.878181677012084e-06, 'learning_rate/action_model': 9.878181677012085e-05, 'epoch': 0.18}) +[2026-06-23 17:20:33,810][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.240456223487854, 'train/grad_norm_pre_clip': 48.95188522338867, 'timing/data': 0.00074760802090168, 'timing/model': 1.71177020506002, 'learning_rate/qwen_vl_interface': 9.877495443670247e-06, 'learning_rate/action_model': 9.877495443670246e-05, 'epoch': 0.18}) +[2026-06-23 17:20:37,183][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.25905004143714905, 'train/grad_norm_pre_clip': 88.81428527832031, 'timing/data': 0.0005817750934511423, 'timing/model': 1.707562267081812, 'learning_rate/qwen_vl_interface': 9.876807308139828e-06, 'learning_rate/action_model': 9.876807308139829e-05, 'epoch': 0.18}) +[2026-06-23 17:20:40,552][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.30477479100227356, 'train/grad_norm_pre_clip': 121.7742691040039, 'timing/data': 0.0006455450784415007, 'timing/model': 1.7103272459935397, 'learning_rate/qwen_vl_interface': 9.876117270703695e-06, 'learning_rate/action_model': 9.876117270703694e-05, 'epoch': 0.18}) +[2026-06-23 17:20:43,927][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.1823592334985733, 'train/grad_norm_pre_clip': 86.60823822021484, 'timing/data': 0.0006156689487397671, 'timing/model': 1.7151826210319996, 'learning_rate/qwen_vl_interface': 9.875425331645495e-06, 'learning_rate/action_model': 9.875425331645495e-05, 'epoch': 0.18}) +[2026-06-23 17:20:47,305][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.2882191836833954, 'train/grad_norm_pre_clip': 35.35050964355469, 'timing/data': 0.0005910429172217846, 'timing/model': 1.7165808780118823, 'learning_rate/qwen_vl_interface': 9.874731491249658e-06, 'learning_rate/action_model': 9.874731491249657e-05, 'epoch': 0.18}) +[2026-06-23 17:20:50,676][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.3288089632987976, 'train/grad_norm_pre_clip': 37.65586471557617, 'timing/data': 0.0005941679701209068, 'timing/model': 1.7083397479727864, 'learning_rate/qwen_vl_interface': 9.874035749801398e-06, 'learning_rate/action_model': 9.874035749801399e-05, 'epoch': 0.18}) +[2026-06-23 17:20:54,054][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.3680344820022583, 'train/grad_norm_pre_clip': 61.39771270751953, 'timing/data': 0.0006502640899270773, 'timing/model': 1.7172446369659156, 'learning_rate/qwen_vl_interface': 9.873338107586704e-06, 'learning_rate/action_model': 9.873338107586705e-05, 'epoch': 0.18}) +[2026-06-23 17:20:57,548][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.32595235109329224, 'train/grad_norm_pre_clip': 56.97343063354492, 'timing/data': 0.0006116610020399094, 'timing/model': 1.7068945670034736, 'learning_rate/qwen_vl_interface': 9.872638564892358e-06, 'learning_rate/action_model': 9.872638564892357e-05, 'epoch': 0.18}) +[2026-06-23 17:21:00,925][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.2971305847167969, 'train/grad_norm_pre_clip': 31.815505981445312, 'timing/data': 0.000702952966094017, 'timing/model': 1.7168725761584938, 'learning_rate/qwen_vl_interface': 9.871937122005905e-06, 'learning_rate/action_model': 9.871937122005905e-05, 'epoch': 0.18}) +[2026-06-23 17:21:04,293][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.31302088499069214, 'train/grad_norm_pre_clip': 23.89602279663086, 'timing/data': 0.0006024839822202921, 'timing/model': 1.7035914198495448, 'learning_rate/qwen_vl_interface': 9.87123377921569e-06, 'learning_rate/action_model': 9.871233779215689e-05, 'epoch': 0.18}) +[2026-06-23 17:21:07,667][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.40516918897628784, 'train/grad_norm_pre_clip': 20.564085006713867, 'timing/data': 0.0006027240306138992, 'timing/model': 1.7150875160004944, 'learning_rate/qwen_vl_interface': 9.870528536810825e-06, 'learning_rate/action_model': 9.870528536810825e-05, 'epoch': 0.18}) +[2026-06-23 17:21:11,028][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.1771543174982071, 'train/grad_norm_pre_clip': 29.592267990112305, 'timing/data': 0.00060466886498034, 'timing/model': 1.7071682659443468, 'learning_rate/qwen_vl_interface': 9.869821395081214e-06, 'learning_rate/action_model': 9.869821395081214e-05, 'epoch': 0.18}) +[2026-06-23 17:21:14,393][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.27885159850120544, 'train/grad_norm_pre_clip': 15.212302207946777, 'timing/data': 0.0005946201272308826, 'timing/model': 1.7128420108929276, 'learning_rate/qwen_vl_interface': 9.869112354317532e-06, 'learning_rate/action_model': 9.869112354317532e-05, 'epoch': 0.18}) +[2026-06-23 17:21:17,767][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.38099735975265503, 'train/grad_norm_pre_clip': 103.6348648071289, 'timing/data': 0.0006234429311007261, 'timing/model': 1.7123988228850067, 'learning_rate/qwen_vl_interface': 9.86840141481124e-06, 'learning_rate/action_model': 9.868401414811239e-05, 'epoch': 0.18}) +[2026-06-23 17:21:21,135][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.2937644422054291, 'train/grad_norm_pre_clip': 54.50283432006836, 'timing/data': 0.0005948489997535944, 'timing/model': 1.7106192400678992, 'learning_rate/qwen_vl_interface': 9.867688576854578e-06, 'learning_rate/action_model': 9.867688576854578e-05, 'epoch': 0.18}) +[2026-06-23 17:21:24,491][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.41012656688690186, 'train/grad_norm_pre_clip': 91.54227447509766, 'timing/data': 0.0007740170694887638, 'timing/model': 1.7009007048327476, 'learning_rate/qwen_vl_interface': 9.86697384074057e-06, 'learning_rate/action_model': 9.866973840740568e-05, 'epoch': 0.18}) +[2026-06-23 17:21:27,864][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.4230409860610962, 'train/grad_norm_pre_clip': 10.019691467285156, 'timing/data': 0.000609417213127017, 'timing/model': 1.7098764970432967, 'learning_rate/qwen_vl_interface': 9.86625720676301e-06, 'learning_rate/action_model': 9.86625720676301e-05, 'epoch': 0.18}) +[2026-06-23 17:21:31,219][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.2971397340297699, 'train/grad_norm_pre_clip': 42.4332160949707, 'timing/data': 0.000608926173299551, 'timing/model': 1.7022925498895347, 'learning_rate/qwen_vl_interface': 9.865538675216488e-06, 'learning_rate/action_model': 9.865538675216486e-05, 'epoch': 0.18}) +[2026-06-23 17:21:34,586][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.26793769001960754, 'train/grad_norm_pre_clip': 20.961610794067383, 'timing/data': 0.0006013219244778156, 'timing/model': 1.708723812829703, 'learning_rate/qwen_vl_interface': 9.864818246396358e-06, 'learning_rate/action_model': 9.864818246396358e-05, 'epoch': 0.18}) +[2026-06-23 17:21:37,950][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.33679383993148804, 'train/grad_norm_pre_clip': 12.845641136169434, 'timing/data': 0.0006131848786026239, 'timing/model': 1.7021938499528915, 'learning_rate/qwen_vl_interface': 9.864095920598764e-06, 'learning_rate/action_model': 9.864095920598763e-05, 'epoch': 0.18}) +[2026-06-23 17:21:41,321][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.2532620429992676, 'train/grad_norm_pre_clip': 32.34479522705078, 'timing/data': 0.0005905909929424524, 'timing/model': 1.712430024985224, 'learning_rate/qwen_vl_interface': 9.86337169812063e-06, 'learning_rate/action_model': 9.863371698120629e-05, 'epoch': 0.18}) +[2026-06-23 17:21:44,686][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.29271385073661804, 'train/grad_norm_pre_clip': 35.427337646484375, 'timing/data': 0.0005880061071366072, 'timing/model': 1.705575771164149, 'learning_rate/qwen_vl_interface': 9.862645579259648e-06, 'learning_rate/action_model': 9.862645579259648e-05, 'epoch': 0.18}) +[2026-06-23 17:21:48,336][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.2792089581489563, 'train/grad_norm_pre_clip': 11.932374954223633, 'timing/data': 0.000646206084638834, 'timing/model': 1.7057905620895326, 'learning_rate/qwen_vl_interface': 9.861917564314308e-06, 'learning_rate/action_model': 9.861917564314308e-05, 'epoch': 0.18}) +[2026-06-23 17:21:51,711][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.32638993859291077, 'train/grad_norm_pre_clip': 45.906982421875, 'timing/data': 0.0006332620978355408, 'timing/model': 1.7143871090374887, 'learning_rate/qwen_vl_interface': 9.861187653583862e-06, 'learning_rate/action_model': 9.861187653583862e-05, 'epoch': 0.18}) +[2026-06-23 17:21:55,076][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.2559114098548889, 'train/grad_norm_pre_clip': 38.25004577636719, 'timing/data': 0.0006997969467192888, 'timing/model': 1.7040588310919702, 'learning_rate/qwen_vl_interface': 9.860455847368352e-06, 'learning_rate/action_model': 9.860455847368351e-05, 'epoch': 0.19}) +[2026-06-23 17:21:58,440][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.23554305732250214, 'train/grad_norm_pre_clip': 46.122703552246094, 'timing/data': 0.0006013719830662012, 'timing/model': 1.7014260140713304, 'learning_rate/qwen_vl_interface': 9.859722145968595e-06, 'learning_rate/action_model': 9.859722145968595e-05, 'epoch': 0.19}) +[2026-06-23 17:22:01,799][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.3343399167060852, 'train/grad_norm_pre_clip': 28.701465606689453, 'timing/data': 0.0005992671940475702, 'timing/model': 1.7023816420696676, 'learning_rate/qwen_vl_interface': 9.858986549686188e-06, 'learning_rate/action_model': 9.858986549686188e-05, 'epoch': 0.19}) +[2026-06-23 17:22:05,166][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.23926973342895508, 'train/grad_norm_pre_clip': 28.571208953857422, 'timing/data': 0.0006187048275023699, 'timing/model': 1.7152378289029002, 'learning_rate/qwen_vl_interface': 9.85824905882351e-06, 'learning_rate/action_model': 9.858249058823509e-05, 'epoch': 0.19}) +[2026-06-23 17:22:08,533][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.2745562195777893, 'train/grad_norm_pre_clip': 20.54108238220215, 'timing/data': 0.0005995491519570351, 'timing/model': 1.7136579540092498, 'learning_rate/qwen_vl_interface': 9.857509673683709e-06, 'learning_rate/action_model': 9.857509673683708e-05, 'epoch': 0.19}) +[2026-06-23 17:22:11,898][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.28253209590911865, 'train/grad_norm_pre_clip': 12.344919204711914, 'timing/data': 0.0006014520768076181, 'timing/model': 1.7050989270210266, 'learning_rate/qwen_vl_interface': 9.856768394570725e-06, 'learning_rate/action_model': 9.856768394570723e-05, 'epoch': 0.19}) +[2026-06-23 17:22:15,271][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.33900368213653564, 'train/grad_norm_pre_clip': 74.7366714477539, 'timing/data': 0.0008260151371359825, 'timing/model': 1.7119118890259415, 'learning_rate/qwen_vl_interface': 9.856025221789264e-06, 'learning_rate/action_model': 9.856025221789264e-05, 'epoch': 0.19}) +[2026-06-23 17:22:18,632][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.3659592270851135, 'train/grad_norm_pre_clip': 41.01642990112305, 'timing/data': 0.0007056379690766335, 'timing/model': 1.7055823260452598, 'learning_rate/qwen_vl_interface': 9.85528015564482e-06, 'learning_rate/action_model': 9.85528015564482e-05, 'epoch': 0.19}) +[2026-06-23 17:22:21,992][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.24474848806858063, 'train/grad_norm_pre_clip': 56.41007995605469, 'timing/data': 0.0006861409638077021, 'timing/model': 1.7066205621231347, 'learning_rate/qwen_vl_interface': 9.854533196443661e-06, 'learning_rate/action_model': 9.854533196443661e-05, 'epoch': 0.19}) +[2026-06-23 17:22:25,357][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.2138856202363968, 'train/grad_norm_pre_clip': 32.12245178222656, 'timing/data': 0.0007172499317675829, 'timing/model': 1.7090196360368282, 'learning_rate/qwen_vl_interface': 9.853784344492833e-06, 'learning_rate/action_model': 9.853784344492833e-05, 'epoch': 0.19}) +[2026-06-23 17:22:28,735][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.30039286613464355, 'train/grad_norm_pre_clip': 64.74297332763672, 'timing/data': 0.0006474889814853668, 'timing/model': 1.7149833200965077, 'learning_rate/qwen_vl_interface': 9.853033600100162e-06, 'learning_rate/action_model': 9.853033600100162e-05, 'epoch': 0.19}) +[2026-06-23 17:22:32,110][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.5162758827209473, 'train/grad_norm_pre_clip': 197.65135192871094, 'timing/data': 0.0006130130495876074, 'timing/model': 1.712032068055123, 'learning_rate/qwen_vl_interface': 9.85228096357425e-06, 'learning_rate/action_model': 9.852280963574249e-05, 'epoch': 0.19}) +[2026-06-23 17:22:35,568][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.3006523847579956, 'train/grad_norm_pre_clip': 148.7488250732422, 'timing/data': 0.0006073431577533484, 'timing/model': 1.7085042390972376, 'learning_rate/qwen_vl_interface': 9.851526435224475e-06, 'learning_rate/action_model': 9.851526435224474e-05, 'epoch': 0.19}) +[2026-06-23 17:22:38,938][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.27659937739372253, 'train/grad_norm_pre_clip': 74.23126220703125, 'timing/data': 0.0005960010457783937, 'timing/model': 1.7091726390644908, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.19}) +[2026-06-23 17:22:42,304][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.22982941567897797, 'train/grad_norm_pre_clip': 37.36162567138672, 'timing/data': 0.0006179630290716887, 'timing/model': 1.711602252908051, 'learning_rate/qwen_vl_interface': 9.850011704294755e-06, 'learning_rate/action_model': 9.850011704294755e-05, 'epoch': 0.19}) +[2026-06-23 17:22:45,676][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.2332279086112976, 'train/grad_norm_pre_clip': 9.488557815551758, 'timing/data': 0.000609125941991806, 'timing/model': 1.7122138640843332, 'learning_rate/qwen_vl_interface': 9.849251502337458e-06, 'learning_rate/action_model': 9.849251502337458e-05, 'epoch': 0.19}) +[2026-06-23 17:22:49,044][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.25168678164482117, 'train/grad_norm_pre_clip': 15.63892936706543, 'timing/data': 0.0005973440129309893, 'timing/model': 1.7089268420822918, 'learning_rate/qwen_vl_interface': 9.848489409801598e-06, 'learning_rate/action_model': 9.848489409801598e-05, 'epoch': 0.19}) +[2026-06-23 17:22:52,415][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.2749202847480774, 'train/grad_norm_pre_clip': 10.5626802444458, 'timing/data': 0.000600229948759079, 'timing/model': 1.711843391880393, 'learning_rate/qwen_vl_interface': 9.84772542700044e-06, 'learning_rate/action_model': 9.84772542700044e-05, 'epoch': 0.19}) +[2026-06-23 17:22:55,790][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.26309916377067566, 'train/grad_norm_pre_clip': 19.43912696838379, 'timing/data': 0.0005967328324913979, 'timing/model': 1.7083584759384394, 'learning_rate/qwen_vl_interface': 9.846959554248032e-06, 'learning_rate/action_model': 9.84695955424803e-05, 'epoch': 0.19}) +[2026-06-23 17:22:59,159][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.23470443487167358, 'train/grad_norm_pre_clip': 14.169997215270996, 'timing/data': 0.0006071520037949085, 'timing/model': 1.7075563101097941, 'learning_rate/qwen_vl_interface': 9.846191791859192e-06, 'learning_rate/action_model': 9.846191791859192e-05, 'epoch': 0.19}) +[2026-06-23 17:23:02,520][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.28471243381500244, 'train/grad_norm_pre_clip': 43.1070671081543, 'timing/data': 0.000608435133472085, 'timing/model': 1.6996592220384628, 'learning_rate/qwen_vl_interface': 9.845422140149518e-06, 'learning_rate/action_model': 9.845422140149519e-05, 'epoch': 0.19}) +[2026-06-23 17:23:05,880][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 500 +[2026-06-23 17:24:40,414][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 500 +[2026-06-23 17:24:40,415][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 500 +[2026-06-23 17:25:53,705][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 500 +[2026-06-23 17:26:02,995][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500 +[2026-06-23 17:26:02,995][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.2837664484977722, 'train/grad_norm_pre_clip': 23.455690383911133, 'eval/action_loss/seconds': 94.53185321111232, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.14791882038116455, 'eval/latency_2/loss': 0.14791882038116455, 'eval/flappy/loss': 0.14791882038116455, 'eval/flappy/latency_2/loss': 0.14791882038116455, 'eval/flap_precision': 0.8722627737226277, 'eval/flap_recall': 0.8063427800269906, 'eval/flap_f1': 0.8380084151472651, 'eval/flap_rate_model': 0.19027777777777777, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.838008415147265, 'eval/latency_2/flap_precision': 0.8722627737226277, 'eval/latency_2/flap_recall': 0.8063427800269906, 'eval/latency_2/flap_f1': 0.8380084151472651, 'eval/latency_2/flap_rate_model': 0.19027777777777777, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.838008415147265, 'eval/action_classification/data_wait_seconds': 0.8673052191734314, 'eval/action_classification/predict_seconds': 72.283203125, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 4.799999999999995, 'rl_games_eval/mid_train/mean_length': 57.2, 'rl_games_eval/mid_train/std_reward': 1.7999999999999952, 'rl_games_eval/mid_train/std_length': 14.4, 'rl_games_eval/mid_train/macro_mean_reward': 4.799999999999995, 'rl_games_eval/mid_train/macro_mean_length': 57.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 4.799999999999995, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 57.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 1.7999999999999952, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 14.4, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 4.799999999999995, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006133140996098518, 'timing/model': 1.7051050630398095, 'learning_rate/qwen_vl_interface': 9.84465059943539e-06, 'learning_rate/action_model': 9.84465059943539e-05, 'epoch': 0.19}) +[2026-06-23 17:26:02,996][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_500_state +[2026-06-23 17:26:02,999][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 17:26:05,839][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_500_state/model.safetensors +[2026-06-23 17:26:06,998][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_500_state/optimizer.bin +[2026-06-23 17:26:06,998][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_500_state/sampler.bin +[2026-06-23 17:26:06,998][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_500_state/sampler_1.bin +[2026-06-23 17:26:06,999][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_500_state/random_states_0.pkl +[2026-06-23 17:26:07,000][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500 +[2026-06-23 17:26:07,000][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 17:26:07,010][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 17:26:10,306][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.2929084300994873, 'train/grad_norm_pre_clip': 32.509883880615234, 'timing/data': 0.000773996114730835, 'timing/model': 1.6762976788450032, 'learning_rate/qwen_vl_interface': 9.843877170033956e-06, 'learning_rate/action_model': 9.843877170033954e-05, 'epoch': 0.19}) +[2026-06-23 17:26:13,602][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.434052973985672, 'train/grad_norm_pre_clip': 117.49625396728516, 'timing/data': 0.0006007298361510038, 'timing/model': 1.6757654480170459, 'learning_rate/qwen_vl_interface': 9.84310185226314e-06, 'learning_rate/action_model': 9.84310185226314e-05, 'epoch': 0.19}) +[2026-06-23 17:26:16,911][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.3239220380783081, 'train/grad_norm_pre_clip': 73.25457763671875, 'timing/data': 0.0007959469221532345, 'timing/model': 1.681260477984324, 'learning_rate/qwen_vl_interface': 9.842324646441652e-06, 'learning_rate/action_model': 9.842324646441652e-05, 'epoch': 0.19}) +[2026-06-23 17:26:20,219][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.26822227239608765, 'train/grad_norm_pre_clip': 47.08143615722656, 'timing/data': 0.0006104190833866596, 'timing/model': 1.6828038431704044, 'learning_rate/qwen_vl_interface': 9.841545552888968e-06, 'learning_rate/action_model': 9.841545552888967e-05, 'epoch': 0.2}) +[2026-06-23 17:26:23,540][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.24609017372131348, 'train/grad_norm_pre_clip': 53.907798767089844, 'timing/data': 0.0007404619827866554, 'timing/model': 1.6857817259151489, 'learning_rate/qwen_vl_interface': 9.840764571925346e-06, 'learning_rate/action_model': 9.840764571925345e-05, 'epoch': 0.2}) +[2026-06-23 17:26:26,855][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.2853347063064575, 'train/grad_norm_pre_clip': 44.31238555908203, 'timing/data': 0.0008360929787158966, 'timing/model': 1.6859632798004895, 'learning_rate/qwen_vl_interface': 9.839981703871815e-06, 'learning_rate/action_model': 9.839981703871814e-05, 'epoch': 0.2}) +[2026-06-23 17:26:30,173][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.4275084137916565, 'train/grad_norm_pre_clip': 88.72140502929688, 'timing/data': 0.0006048069335520267, 'timing/model': 1.6800056658685207, 'learning_rate/qwen_vl_interface': 9.839196949050185e-06, 'learning_rate/action_model': 9.839196949050185e-05, 'epoch': 0.2}) +[2026-06-23 17:26:33,487][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.22385668754577637, 'train/grad_norm_pre_clip': 21.58433723449707, 'timing/data': 0.0006071119569242001, 'timing/model': 1.6862841378897429, 'learning_rate/qwen_vl_interface': 9.838410307783038e-06, 'learning_rate/action_model': 9.838410307783038e-05, 'epoch': 0.2}) +[2026-06-23 17:26:36,799][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.2564844787120819, 'train/grad_norm_pre_clip': 25.193538665771484, 'timing/data': 0.0006124519277364016, 'timing/model': 1.680256913881749, 'learning_rate/qwen_vl_interface': 9.837621780393734e-06, 'learning_rate/action_model': 9.837621780393733e-05, 'epoch': 0.2}) +[2026-06-23 17:26:40,130][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.17573927342891693, 'train/grad_norm_pre_clip': 34.471275329589844, 'timing/data': 0.0006052679382264614, 'timing/model': 1.6899439378175884, 'learning_rate/qwen_vl_interface': 9.836831367206406e-06, 'learning_rate/action_model': 9.836831367206406e-05, 'epoch': 0.2}) +[2026-06-23 17:26:43,466][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.22816871106624603, 'train/grad_norm_pre_clip': 15.091856002807617, 'timing/data': 0.0006207569967955351, 'timing/model': 1.6893989038653672, 'learning_rate/qwen_vl_interface': 9.836039068545962e-06, 'learning_rate/action_model': 9.836039068545961e-05, 'epoch': 0.2}) +[2026-06-23 17:26:46,787][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.35069897770881653, 'train/grad_norm_pre_clip': 79.45344543457031, 'timing/data': 0.0006055890116840601, 'timing/model': 1.689586099004373, 'learning_rate/qwen_vl_interface': 9.835244884738086e-06, 'learning_rate/action_model': 9.835244884738085e-05, 'epoch': 0.2}) +[2026-06-23 17:26:50,200][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.21314051747322083, 'train/grad_norm_pre_clip': 47.92470169067383, 'timing/data': 0.0006006001494824886, 'timing/model': 1.6828143149614334, 'learning_rate/qwen_vl_interface': 9.83444881610924e-06, 'learning_rate/action_model': 9.834448816109238e-05, 'epoch': 0.2}) +[2026-06-23 17:26:53,520][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.34220457077026367, 'train/grad_norm_pre_clip': 71.15129852294922, 'timing/data': 0.0007060780189931393, 'timing/model': 1.689005594002083, 'learning_rate/qwen_vl_interface': 9.833650862986652e-06, 'learning_rate/action_model': 9.833650862986653e-05, 'epoch': 0.2}) +[2026-06-23 17:26:56,839][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.30156317353248596, 'train/grad_norm_pre_clip': 30.540916442871094, 'timing/data': 0.0008169061038643122, 'timing/model': 1.6853417749516666, 'learning_rate/qwen_vl_interface': 9.832851025698335e-06, 'learning_rate/action_model': 9.832851025698335e-05, 'epoch': 0.2}) +[2026-06-23 17:27:00,163][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.18482577800750732, 'train/grad_norm_pre_clip': 50.94317626953125, 'timing/data': 0.0006088248919695616, 'timing/model': 1.6901394228916615, 'learning_rate/qwen_vl_interface': 9.832049304573072e-06, 'learning_rate/action_model': 9.832049304573072e-05, 'epoch': 0.2}) +[2026-06-23 17:27:03,493][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.2991928458213806, 'train/grad_norm_pre_clip': 24.293569564819336, 'timing/data': 0.0005961209535598755, 'timing/model': 1.6901648938655853, 'learning_rate/qwen_vl_interface': 9.831245699940419e-06, 'learning_rate/action_model': 9.831245699940418e-05, 'epoch': 0.2}) +[2026-06-23 17:27:06,823][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.3055378794670105, 'train/grad_norm_pre_clip': 16.80253028869629, 'timing/data': 0.0005918329115957022, 'timing/model': 1.691498865839094, 'learning_rate/qwen_vl_interface': 9.830440212130707e-06, 'learning_rate/action_model': 9.830440212130708e-05, 'epoch': 0.2}) +[2026-06-23 17:27:10,155][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.3036034405231476, 'train/grad_norm_pre_clip': 16.640029907226562, 'timing/data': 0.0006101781036704779, 'timing/model': 1.6974559449590743, 'learning_rate/qwen_vl_interface': 9.829632841475041e-06, 'learning_rate/action_model': 9.829632841475041e-05, 'epoch': 0.2}) +[2026-06-23 17:27:13,486][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.300216943025589, 'train/grad_norm_pre_clip': 66.70109558105469, 'timing/data': 0.0006034651305526495, 'timing/model': 1.6923014628700912, 'learning_rate/qwen_vl_interface': 9.828823588305305e-06, 'learning_rate/action_model': 9.828823588305303e-05, 'epoch': 0.2}) +[2026-06-23 17:27:16,821][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.24826446175575256, 'train/grad_norm_pre_clip': 23.750263214111328, 'timing/data': 0.0006014921236783266, 'timing/model': 1.6955601959489286, 'learning_rate/qwen_vl_interface': 9.828012452954146e-06, 'learning_rate/action_model': 9.828012452954145e-05, 'epoch': 0.2}) +[2026-06-23 17:27:20,156][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.30629029870033264, 'train/grad_norm_pre_clip': 11.580890655517578, 'timing/data': 0.0007341110613197088, 'timing/model': 1.6916863040532917, 'learning_rate/qwen_vl_interface': 9.827199435754995e-06, 'learning_rate/action_model': 9.827199435754995e-05, 'epoch': 0.2}) +[2026-06-23 17:27:23,500][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.23547358810901642, 'train/grad_norm_pre_clip': 13.256994247436523, 'timing/data': 0.0005922438576817513, 'timing/model': 1.7012575690168887, 'learning_rate/qwen_vl_interface': 9.826384537042052e-06, 'learning_rate/action_model': 9.826384537042052e-05, 'epoch': 0.2}) +[2026-06-23 17:27:26,839][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.1975681185722351, 'train/grad_norm_pre_clip': 46.20705032348633, 'timing/data': 0.0005939980037510395, 'timing/model': 1.6936519490554929, 'learning_rate/qwen_vl_interface': 9.825567757150291e-06, 'learning_rate/action_model': 9.82556775715029e-05, 'epoch': 0.2}) +[2026-06-23 17:27:30,179][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.3027908504009247, 'train/grad_norm_pre_clip': 63.56972885131836, 'timing/data': 0.0006743390113115311, 'timing/model': 1.6956005280371755, 'learning_rate/qwen_vl_interface': 9.82474909641546e-06, 'learning_rate/action_model': 9.82474909641546e-05, 'epoch': 0.2}) +[2026-06-23 17:27:33,516][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.2672153115272522, 'train/grad_norm_pre_clip': 67.60602569580078, 'timing/data': 0.0006011698860675097, 'timing/model': 1.6939531329553574, 'learning_rate/qwen_vl_interface': 9.823928555174078e-06, 'learning_rate/action_model': 9.823928555174077e-05, 'epoch': 0.2}) +[2026-06-23 17:27:36,946][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.3301118016242981, 'train/grad_norm_pre_clip': 32.43972396850586, 'timing/data': 0.0005802009254693985, 'timing/model': 1.7927167189773172, 'learning_rate/qwen_vl_interface': 9.82310613376344e-06, 'learning_rate/action_model': 9.82310613376344e-05, 'epoch': 0.2}) +[2026-06-23 17:27:40,282][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.20665118098258972, 'train/grad_norm_pre_clip': 9.27099323272705, 'timing/data': 0.0005872650071978569, 'timing/model': 1.6905694480519742, 'learning_rate/qwen_vl_interface': 9.822281832521613e-06, 'learning_rate/action_model': 9.822281832521613e-05, 'epoch': 0.2}) +[2026-06-23 17:27:43,609][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.22781093418598175, 'train/grad_norm_pre_clip': 18.368762969970703, 'timing/data': 0.0005960208363831043, 'timing/model': 1.686651353025809, 'learning_rate/qwen_vl_interface': 9.821455651787432e-06, 'learning_rate/action_model': 9.821455651787432e-05, 'epoch': 0.2}) +[2026-06-23 17:27:46,935][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.3094737231731415, 'train/grad_norm_pre_clip': 53.226444244384766, 'timing/data': 0.0005927048623561859, 'timing/model': 1.6854580538347363, 'learning_rate/qwen_vl_interface': 9.820627591900514e-06, 'learning_rate/action_model': 9.820627591900513e-05, 'epoch': 0.21}) +[2026-06-23 17:27:50,272][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.40902501344680786, 'train/grad_norm_pre_clip': 75.88439178466797, 'timing/data': 0.0005850000306963921, 'timing/model': 1.6946376040577888, 'learning_rate/qwen_vl_interface': 9.819797653201238e-06, 'learning_rate/action_model': 9.819797653201238e-05, 'epoch': 0.21}) +[2026-06-23 17:27:53,609][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.3693120777606964, 'train/grad_norm_pre_clip': 47.63759994506836, 'timing/data': 0.0008137209806591272, 'timing/model': 1.6968541869428009, 'learning_rate/qwen_vl_interface': 9.818965836030765e-06, 'learning_rate/action_model': 9.818965836030764e-05, 'epoch': 0.21}) +[2026-06-23 17:27:56,944][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.33839645981788635, 'train/grad_norm_pre_clip': 59.17898178100586, 'timing/data': 0.0006008001510053873, 'timing/model': 1.697461114032194, 'learning_rate/qwen_vl_interface': 9.81813214073102e-06, 'learning_rate/action_model': 9.81813214073102e-05, 'epoch': 0.21}) +[2026-06-23 17:28:00,279][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.432003915309906, 'train/grad_norm_pre_clip': 114.32295227050781, 'timing/data': 0.0005967519246041775, 'timing/model': 1.6958297609817237, 'learning_rate/qwen_vl_interface': 9.817296567644709e-06, 'learning_rate/action_model': 9.817296567644709e-05, 'epoch': 0.21}) +[2026-06-23 17:28:03,608][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.28422096371650696, 'train/grad_norm_pre_clip': 14.96388053894043, 'timing/data': 0.0005965919699519873, 'timing/model': 1.691137561108917, 'learning_rate/qwen_vl_interface': 9.816459117115296e-06, 'learning_rate/action_model': 9.816459117115297e-05, 'epoch': 0.21}) +[2026-06-23 17:28:06,931][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.27913734316825867, 'train/grad_norm_pre_clip': 20.546287536621094, 'timing/data': 0.0006024530157446861, 'timing/model': 1.692737685982138, 'learning_rate/qwen_vl_interface': 9.815619789487035e-06, 'learning_rate/action_model': 9.815619789487034e-05, 'epoch': 0.21}) +[2026-06-23 17:28:10,255][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.17288838326931, 'train/grad_norm_pre_clip': 31.253326416015625, 'timing/data': 0.0006032651290297508, 'timing/model': 1.685205786023289, 'learning_rate/qwen_vl_interface': 9.814778585104933e-06, 'learning_rate/action_model': 9.814778585104933e-05, 'epoch': 0.21}) +[2026-06-23 17:28:13,591][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.26599928736686707, 'train/grad_norm_pre_clip': 26.298259735107422, 'timing/data': 0.0006852888036519289, 'timing/model': 1.6913655160460621, 'learning_rate/qwen_vl_interface': 9.813935504314784e-06, 'learning_rate/action_model': 9.813935504314784e-05, 'epoch': 0.21}) +[2026-06-23 17:28:16,920][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.36160707473754883, 'train/grad_norm_pre_clip': 78.80451965332031, 'timing/data': 0.0007375068962574005, 'timing/model': 1.6867541389074177, 'learning_rate/qwen_vl_interface': 9.813090547463144e-06, 'learning_rate/action_model': 9.813090547463143e-05, 'epoch': 0.21}) +[2026-06-23 17:28:20,283][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.3797368109226227, 'train/grad_norm_pre_clip': 109.6338882446289, 'timing/data': 0.0006405250169336796, 'timing/model': 1.6901979818940163, 'learning_rate/qwen_vl_interface': 9.812243714897342e-06, 'learning_rate/action_model': 9.812243714897343e-05, 'epoch': 0.21}) +[2026-06-23 17:28:23,633][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.26171165704727173, 'train/grad_norm_pre_clip': 57.715885162353516, 'timing/data': 0.000686099985614419, 'timing/model': 1.6942523210309446, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.21}) +[2026-06-23 17:28:27,115][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.1234547346830368, 'train/grad_norm_pre_clip': 11.358086585998535, 'timing/data': 0.0005893579218536615, 'timing/model': 1.8379228301346302, 'learning_rate/qwen_vl_interface': 9.81054442401643e-06, 'learning_rate/action_model': 9.81054442401643e-05, 'epoch': 0.21}) +[2026-06-23 17:28:30,449][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.33070430159568787, 'train/grad_norm_pre_clip': 32.473934173583984, 'timing/data': 0.0005898489616811275, 'timing/model': 1.6917015160433948, 'learning_rate/qwen_vl_interface': 9.809691966399834e-06, 'learning_rate/action_model': 9.809691966399835e-05, 'epoch': 0.21}) +[2026-06-23 17:28:33,783][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.281854510307312, 'train/grad_norm_pre_clip': 18.31739044189453, 'timing/data': 0.00059127202257514, 'timing/model': 1.6957980650477111, 'learning_rate/qwen_vl_interface': 9.808837634466104e-06, 'learning_rate/action_model': 9.808837634466104e-05, 'epoch': 0.21}) +[2026-06-23 17:28:37,119][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.21865616738796234, 'train/grad_norm_pre_clip': 27.123394012451172, 'timing/data': 0.0006968099623918533, 'timing/model': 1.6973541979677975, 'learning_rate/qwen_vl_interface': 9.807981428566428e-06, 'learning_rate/action_model': 9.807981428566427e-05, 'epoch': 0.21}) +[2026-06-23 17:28:40,466][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.3734855651855469, 'train/grad_norm_pre_clip': 59.74913024902344, 'timing/data': 0.0009921560995280743, 'timing/model': 1.6979066841304302, 'learning_rate/qwen_vl_interface': 9.807123349052754e-06, 'learning_rate/action_model': 9.807123349052753e-05, 'epoch': 0.21}) +[2026-06-23 17:28:43,794][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.20976629853248596, 'train/grad_norm_pre_clip': 39.60070037841797, 'timing/data': 0.0006807709578424692, 'timing/model': 1.6898462111130357, 'learning_rate/qwen_vl_interface': 9.80626339627781e-06, 'learning_rate/action_model': 9.806263396277809e-05, 'epoch': 0.21}) +[2026-06-23 17:28:47,133][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.20354239642620087, 'train/grad_norm_pre_clip': 10.090332984924316, 'timing/data': 0.0009133180137723684, 'timing/model': 1.6955139371566474, 'learning_rate/qwen_vl_interface': 9.805401570595087e-06, 'learning_rate/action_model': 9.805401570595087e-05, 'epoch': 0.21}) +[2026-06-23 17:28:50,471][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.2845207154750824, 'train/grad_norm_pre_clip': 9.590728759765625, 'timing/data': 0.0005959109403192997, 'timing/model': 1.6947175948880613, 'learning_rate/qwen_vl_interface': 9.804537872358853e-06, 'learning_rate/action_model': 9.804537872358853e-05, 'epoch': 0.21}) +[2026-06-23 17:28:53,797][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.16806191205978394, 'train/grad_norm_pre_clip': 40.50376510620117, 'timing/data': 0.0010060619097203016, 'timing/model': 1.6875029241200536, 'learning_rate/qwen_vl_interface': 9.80367230192414e-06, 'learning_rate/action_model': 9.803672301924138e-05, 'epoch': 0.21}) +[2026-06-23 17:28:57,121][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.257193386554718, 'train/grad_norm_pre_clip': 55.11146545410156, 'timing/data': 0.0006961789913475513, 'timing/model': 1.693010828923434, 'learning_rate/qwen_vl_interface': 9.80280485964675e-06, 'learning_rate/action_model': 9.80280485964675e-05, 'epoch': 0.21}) +[2026-06-23 17:29:00,450][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.2261040210723877, 'train/grad_norm_pre_clip': 48.160186767578125, 'timing/data': 0.0006063699256628752, 'timing/model': 1.692233253037557, 'learning_rate/qwen_vl_interface': 9.801935545883255e-06, 'learning_rate/action_model': 9.801935545883255e-05, 'epoch': 0.21}) +[2026-06-23 17:29:03,791][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.18041247129440308, 'train/grad_norm_pre_clip': 25.130327224731445, 'timing/data': 0.0008293988648802042, 'timing/model': 1.6920262707863003, 'learning_rate/qwen_vl_interface': 9.801064360991e-06, 'learning_rate/action_model': 9.801064360991e-05, 'epoch': 0.21}) +[2026-06-23 17:29:07,131][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.3385697603225708, 'train/grad_norm_pre_clip': 67.63404846191406, 'timing/data': 0.0006005889736115932, 'timing/model': 1.69608332705684, 'learning_rate/qwen_vl_interface': 9.800191305328095e-06, 'learning_rate/action_model': 9.800191305328094e-05, 'epoch': 0.21}) +[2026-06-23 17:29:10,466][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.31344422698020935, 'train/grad_norm_pre_clip': 41.427921295166016, 'timing/data': 0.0006455641705542803, 'timing/model': 1.691767678130418, 'learning_rate/qwen_vl_interface': 9.79931637925342e-06, 'learning_rate/action_model': 9.79931637925342e-05, 'epoch': 0.21}) +[2026-06-23 17:29:13,797][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.26956701278686523, 'train/grad_norm_pre_clip': 42.4048957824707, 'timing/data': 0.0008386068511754274, 'timing/model': 1.690136277815327, 'learning_rate/qwen_vl_interface': 9.798439583126625e-06, 'learning_rate/action_model': 9.798439583126625e-05, 'epoch': 0.22}) +[2026-06-23 17:29:17,261][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.2361532300710678, 'train/grad_norm_pre_clip': 50.17700958251953, 'timing/data': 0.0006070721428841352, 'timing/model': 1.6892931289039552, 'learning_rate/qwen_vl_interface': 9.797560917308126e-06, 'learning_rate/action_model': 9.797560917308126e-05, 'epoch': 0.22}) +[2026-06-23 17:29:20,595][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.23810724914073944, 'train/grad_norm_pre_clip': 14.509211540222168, 'timing/data': 0.000649942085146904, 'timing/model': 1.6954211990814656, 'learning_rate/qwen_vl_interface': 9.796680382159112e-06, 'learning_rate/action_model': 9.79668038215911e-05, 'epoch': 0.22}) +[2026-06-23 17:29:23,930][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.2887234389781952, 'train/grad_norm_pre_clip': 45.51133346557617, 'timing/data': 0.0008792639710009098, 'timing/model': 1.6968733288813382, 'learning_rate/qwen_vl_interface': 9.795797978041533e-06, 'learning_rate/action_model': 9.795797978041533e-05, 'epoch': 0.22}) +[2026-06-23 17:29:27,256][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.23211780190467834, 'train/grad_norm_pre_clip': 35.5739631652832, 'timing/data': 0.0006547921802848577, 'timing/model': 1.695230766898021, 'learning_rate/qwen_vl_interface': 9.794913705318117e-06, 'learning_rate/action_model': 9.794913705318118e-05, 'epoch': 0.22}) +[2026-06-23 17:29:30,588][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.3261752128601074, 'train/grad_norm_pre_clip': 21.192655563354492, 'timing/data': 0.0006505630444735289, 'timing/model': 1.6908313711173832, 'learning_rate/qwen_vl_interface': 9.794027564352352e-06, 'learning_rate/action_model': 9.794027564352352e-05, 'epoch': 0.22}) +[2026-06-23 17:29:33,930][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.2492486983537674, 'train/grad_norm_pre_clip': 21.19281578063965, 'timing/data': 0.0007216969970613718, 'timing/model': 1.7013447389472276, 'learning_rate/qwen_vl_interface': 9.793139555508501e-06, 'learning_rate/action_model': 9.7931395555085e-05, 'epoch': 0.22}) +[2026-06-23 17:29:37,257][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.3678732216358185, 'train/grad_norm_pre_clip': 123.28522491455078, 'timing/data': 0.0005972539074718952, 'timing/model': 1.6914190910756588, 'learning_rate/qwen_vl_interface': 9.792249679151585e-06, 'learning_rate/action_model': 9.792249679151585e-05, 'epoch': 0.22}) +[2026-06-23 17:29:40,586][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.304507315158844, 'train/grad_norm_pre_clip': 61.034358978271484, 'timing/data': 0.0005945179145783186, 'timing/model': 1.6934338780120015, 'learning_rate/qwen_vl_interface': 9.791357935647403e-06, 'learning_rate/action_model': 9.791357935647403e-05, 'epoch': 0.22}) +[2026-06-23 17:29:43,915][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.23490256071090698, 'train/grad_norm_pre_clip': 8.316655158996582, 'timing/data': 0.0006093361880630255, 'timing/model': 1.6948108349461108, 'learning_rate/qwen_vl_interface': 9.790464325362516e-06, 'learning_rate/action_model': 9.790464325362516e-05, 'epoch': 0.22}) +[2026-06-23 17:29:47,247][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.20031996071338654, 'train/grad_norm_pre_clip': 8.454647064208984, 'timing/data': 0.0006064909975975752, 'timing/model': 1.6963286269456148, 'learning_rate/qwen_vl_interface': 9.789568848664254e-06, 'learning_rate/action_model': 9.789568848664253e-05, 'epoch': 0.22}) +[2026-06-23 17:29:50,583][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.18732377886772156, 'train/grad_norm_pre_clip': 43.574188232421875, 'timing/data': 0.0006007801275700331, 'timing/model': 1.6925801469478756, 'learning_rate/qwen_vl_interface': 9.788671505920712e-06, 'learning_rate/action_model': 9.788671505920711e-05, 'epoch': 0.22}) +[2026-06-23 17:29:53,920][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.2545814514160156, 'train/grad_norm_pre_clip': 23.42268943786621, 'timing/data': 0.0006099168676882982, 'timing/model': 1.6953212590888143, 'learning_rate/qwen_vl_interface': 9.787772297500754e-06, 'learning_rate/action_model': 9.787772297500754e-05, 'epoch': 0.22}) +[2026-06-23 17:29:57,257][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.20641577243804932, 'train/grad_norm_pre_clip': 27.9992733001709, 'timing/data': 0.0005867430008947849, 'timing/model': 1.6897565831895918, 'learning_rate/qwen_vl_interface': 9.78687122377401e-06, 'learning_rate/action_model': 9.78687122377401e-05, 'epoch': 0.22}) +[2026-06-23 17:30:00,601][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.2986855208873749, 'train/grad_norm_pre_clip': 25.437013626098633, 'timing/data': 0.000819189939647913, 'timing/model': 1.7013526759110391, 'learning_rate/qwen_vl_interface': 9.78596828511088e-06, 'learning_rate/action_model': 9.78596828511088e-05, 'epoch': 0.22}) +[2026-06-23 17:30:03,945][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.2718792259693146, 'train/grad_norm_pre_clip': 21.870586395263672, 'timing/data': 0.0006155879236757755, 'timing/model': 1.6976008999627084, 'learning_rate/qwen_vl_interface': 9.785063481882527e-06, 'learning_rate/action_model': 9.785063481882527e-05, 'epoch': 0.22}) +[2026-06-23 17:30:07,551][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.33260348439216614, 'train/grad_norm_pre_clip': 23.99315071105957, 'timing/data': 0.0006511348765343428, 'timing/model': 1.694554579211399, 'learning_rate/qwen_vl_interface': 9.78415681446088e-06, 'learning_rate/action_model': 9.784156814460879e-05, 'epoch': 0.22}) +[2026-06-23 17:30:10,892][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.22691267728805542, 'train/grad_norm_pre_clip': 55.133121490478516, 'timing/data': 0.0006016709376126528, 'timing/model': 1.6956089709419757, 'learning_rate/qwen_vl_interface': 9.783248283218638e-06, 'learning_rate/action_model': 9.783248283218638e-05, 'epoch': 0.22}) +[2026-06-23 17:30:14,223][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.15166763961315155, 'train/grad_norm_pre_clip': 33.664371490478516, 'timing/data': 0.0006143248174339533, 'timing/model': 1.6921897190622985, 'learning_rate/qwen_vl_interface': 9.782337888529261e-06, 'learning_rate/action_model': 9.782337888529261e-05, 'epoch': 0.22}) +[2026-06-23 17:30:17,554][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.2938157320022583, 'train/grad_norm_pre_clip': 26.746103286743164, 'timing/data': 0.0006795178633183241, 'timing/model': 1.6929293589200824, 'learning_rate/qwen_vl_interface': 9.781425630766978e-06, 'learning_rate/action_model': 9.781425630766979e-05, 'epoch': 0.22}) +[2026-06-23 17:30:20,885][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.2344050407409668, 'train/grad_norm_pre_clip': 10.10951042175293, 'timing/data': 0.0008346298709511757, 'timing/model': 1.697546259034425, 'learning_rate/qwen_vl_interface': 9.780511510306789e-06, 'learning_rate/action_model': 9.780511510306788e-05, 'epoch': 0.22}) +[2026-06-23 17:30:24,213][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.12623706459999084, 'train/grad_norm_pre_clip': 15.270267486572266, 'timing/data': 0.0006655619945377111, 'timing/model': 1.6856435660738498, 'learning_rate/qwen_vl_interface': 9.779595527524447e-06, 'learning_rate/action_model': 9.779595527524447e-05, 'epoch': 0.22}) +[2026-06-23 17:30:27,547][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.2678530514240265, 'train/grad_norm_pre_clip': 29.733299255371094, 'timing/data': 0.0006222398951649666, 'timing/model': 1.6939291139133275, 'learning_rate/qwen_vl_interface': 9.778677682796483e-06, 'learning_rate/action_model': 9.778677682796483e-05, 'epoch': 0.22}) +[2026-06-23 17:30:30,890][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.22637467086315155, 'train/grad_norm_pre_clip': 54.7159309387207, 'timing/data': 0.0005951188504695892, 'timing/model': 1.6984670970123261, 'learning_rate/qwen_vl_interface': 9.777757976500185e-06, 'learning_rate/action_model': 9.777757976500186e-05, 'epoch': 0.22}) +[2026-06-23 17:30:34,214][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.2952177822589874, 'train/grad_norm_pre_clip': 40.706146240234375, 'timing/data': 0.0006832950748503208, 'timing/model': 1.689561891136691, 'learning_rate/qwen_vl_interface': 9.776836409013614e-06, 'learning_rate/action_model': 9.776836409013615e-05, 'epoch': 0.22}) +[2026-06-23 17:30:37,544][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.2889315187931061, 'train/grad_norm_pre_clip': 30.2763729095459, 'timing/data': 0.0006577260792255402, 'timing/model': 1.684839660068974, 'learning_rate/qwen_vl_interface': 9.775912980715589e-06, 'learning_rate/action_model': 9.775912980715588e-05, 'epoch': 0.22}) +[2026-06-23 17:30:40,874][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.4138261675834656, 'train/grad_norm_pre_clip': 106.90151977539062, 'timing/data': 0.0006667140405625105, 'timing/model': 1.6883126799948514, 'learning_rate/qwen_vl_interface': 9.774987691985694e-06, 'learning_rate/action_model': 9.774987691985694e-05, 'epoch': 0.23}) +[2026-06-23 17:30:44,204][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.4825633764266968, 'train/grad_norm_pre_clip': 77.25257110595703, 'timing/data': 0.0006611540447920561, 'timing/model': 1.6892120300326496, 'learning_rate/qwen_vl_interface': 9.774060543204283e-06, 'learning_rate/action_model': 9.774060543204283e-05, 'epoch': 0.23}) +[2026-06-23 17:30:47,532][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.31443867087364197, 'train/grad_norm_pre_clip': 71.34246063232422, 'timing/data': 0.0006598900072276592, 'timing/model': 1.6875426799524575, 'learning_rate/qwen_vl_interface': 9.773131534752472e-06, 'learning_rate/action_model': 9.773131534752472e-05, 'epoch': 0.23}) +[2026-06-23 17:30:50,864][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.2696683406829834, 'train/grad_norm_pre_clip': 11.009355545043945, 'timing/data': 0.0006519658491015434, 'timing/model': 1.6852250420488417, 'learning_rate/qwen_vl_interface': 9.772200667012142e-06, 'learning_rate/action_model': 9.772200667012142e-05, 'epoch': 0.23}) +[2026-06-23 17:30:54,284][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.17163455486297607, 'train/grad_norm_pre_clip': 16.280527114868164, 'timing/data': 0.0005971330683678389, 'timing/model': 1.6861444199457765, 'learning_rate/qwen_vl_interface': 9.771267940365934e-06, 'learning_rate/action_model': 9.771267940365935e-05, 'epoch': 0.23}) +[2026-06-23 17:30:57,618][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.3334481716156006, 'train/grad_norm_pre_clip': 153.75021362304688, 'timing/data': 0.0006882839370518923, 'timing/model': 1.6952124100644141, 'learning_rate/qwen_vl_interface': 9.770333355197262e-06, 'learning_rate/action_model': 9.77033335519726e-05, 'epoch': 0.23}) +[2026-06-23 17:31:00,952][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.3630859851837158, 'train/grad_norm_pre_clip': 115.4881591796875, 'timing/data': 0.0006075720302760601, 'timing/model': 1.69116222509183, 'learning_rate/qwen_vl_interface': 9.769396911890291e-06, 'learning_rate/action_model': 9.769396911890291e-05, 'epoch': 0.23}) +[2026-06-23 17:31:04,279][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.5099048614501953, 'train/grad_norm_pre_clip': 108.3923568725586, 'timing/data': 0.0006030439399182796, 'timing/model': 1.6877844019327313, 'learning_rate/qwen_vl_interface': 9.768458610829965e-06, 'learning_rate/action_model': 9.768458610829965e-05, 'epoch': 0.23}) +[2026-06-23 17:31:07,612][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.44366222620010376, 'train/grad_norm_pre_clip': 100.48828887939453, 'timing/data': 0.0005983349401503801, 'timing/model': 1.6907148221507668, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.23}) +[2026-06-23 17:31:10,957][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.2915360927581787, 'train/grad_norm_pre_clip': 79.62843322753906, 'timing/data': 0.0005953200161457062, 'timing/model': 1.6982017990667373, 'learning_rate/qwen_vl_interface': 9.766576436992802e-06, 'learning_rate/action_model': 9.766576436992801e-05, 'epoch': 0.23}) +[2026-06-23 17:31:14,291][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.3572314381599426, 'train/grad_norm_pre_clip': 51.49434280395508, 'timing/data': 0.0006083438638597727, 'timing/model': 1.69328223913908, 'learning_rate/qwen_vl_interface': 9.765632564989657e-06, 'learning_rate/action_model': 9.765632564989656e-05, 'epoch': 0.23}) +[2026-06-23 17:31:17,633][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.22493989765644073, 'train/grad_norm_pre_clip': 62.727359771728516, 'timing/data': 0.000586171867325902, 'timing/model': 1.6979590731207281, 'learning_rate/qwen_vl_interface': 9.764686836780533e-06, 'learning_rate/action_model': 9.764686836780532e-05, 'epoch': 0.23}) +[2026-06-23 17:31:20,974][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.2612033188343048, 'train/grad_norm_pre_clip': 36.98051834106445, 'timing/data': 0.0008816181216388941, 'timing/model': 1.6964828460477293, 'learning_rate/qwen_vl_interface': 9.763739252754187e-06, 'learning_rate/action_model': 9.763739252754186e-05, 'epoch': 0.23}) +[2026-06-23 17:31:24,311][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.19594116508960724, 'train/grad_norm_pre_clip': 18.955808639526367, 'timing/data': 0.0006492410320788622, 'timing/model': 1.692567033926025, 'learning_rate/qwen_vl_interface': 9.76278981330013e-06, 'learning_rate/action_model': 9.762789813300131e-05, 'epoch': 0.23}) +[2026-06-23 17:31:27,641][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.2287057787179947, 'train/grad_norm_pre_clip': 36.27191925048828, 'timing/data': 0.000646855914965272, 'timing/model': 1.6911934309173375, 'learning_rate/qwen_vl_interface': 9.761838518808647e-06, 'learning_rate/action_model': 9.761838518808646e-05, 'epoch': 0.23}) +[2026-06-23 17:31:30,974][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.18205884099006653, 'train/grad_norm_pre_clip': 8.146184921264648, 'timing/data': 0.0006728051230311394, 'timing/model': 1.6894481498748064, 'learning_rate/qwen_vl_interface': 9.760885369670776e-06, 'learning_rate/action_model': 9.760885369670776e-05, 'epoch': 0.23}) +[2026-06-23 17:31:34,315][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.24381476640701294, 'train/grad_norm_pre_clip': 12.738250732421875, 'timing/data': 0.0006864310707896948, 'timing/model': 1.69832162396051, 'learning_rate/qwen_vl_interface': 9.759930366278319e-06, 'learning_rate/action_model': 9.759930366278318e-05, 'epoch': 0.23}) +[2026-06-23 17:31:37,660][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.16824029386043549, 'train/grad_norm_pre_clip': 16.163408279418945, 'timing/data': 0.0006071820389479399, 'timing/model': 1.6955233430489898, 'learning_rate/qwen_vl_interface': 9.758973509023842e-06, 'learning_rate/action_model': 9.758973509023843e-05, 'epoch': 0.23}) +[2026-06-23 17:31:41,006][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 600 +[2026-06-23 17:33:13,970][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 600 +[2026-06-23 17:33:13,970][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 600 +[2026-06-23 17:34:24,852][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 600 +[2026-06-23 17:34:24,853][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.1960286945104599, 'train/grad_norm_pre_clip': 11.814169883728027, 'eval/action_loss/seconds': 92.96115161594935, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.14444051682949066, 'eval/latency_2/loss': 0.14444051682949066, 'eval/flappy/loss': 0.14444051682949066, 'eval/flappy/latency_2/loss': 0.14444051682949066, 'eval/flap_precision': 0.904498816101026, 'eval/flap_recall': 0.7732793522267206, 'eval/flap_f1': 0.8337577300836667, 'eval/flap_rate_model': 0.17597222222222222, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8337577300836667, 'eval/latency_2/flap_precision': 0.904498816101026, 'eval/latency_2/flap_recall': 0.7732793522267206, 'eval/latency_2/flap_f1': 0.8337577300836667, 'eval/latency_2/flap_rate_model': 0.17597222222222222, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8337577300836667, 'eval/action_classification/data_wait_seconds': 0.9160074591636658, 'eval/action_classification/predict_seconds': 69.83580017089844, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006145648658275604, 'timing/model': 1.6984973850194365, 'learning_rate/qwen_vl_interface': 9.758014798300677e-06, 'learning_rate/action_model': 9.758014798300676e-05, 'epoch': 0.23}) +[2026-06-23 17:34:28,179][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.2515973150730133, 'train/grad_norm_pre_clip': 76.26310729980469, 'timing/data': 0.0008278160821646452, 'timing/model': 1.6868428911548108, 'learning_rate/qwen_vl_interface': 9.757054234502909e-06, 'learning_rate/action_model': 9.757054234502909e-05, 'epoch': 0.23}) +[2026-06-23 17:34:31,501][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.24188804626464844, 'train/grad_norm_pre_clip': 96.02682495117188, 'timing/data': 0.0006033540703356266, 'timing/model': 1.6913110530003905, 'learning_rate/qwen_vl_interface': 9.756091818025394e-06, 'learning_rate/action_model': 9.756091818025394e-05, 'epoch': 0.23}) +[2026-06-23 17:34:34,820][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.2737480103969574, 'train/grad_norm_pre_clip': 11.043970108032227, 'timing/data': 0.0005979139823466539, 'timing/model': 1.6900127460248768, 'learning_rate/qwen_vl_interface': 9.75512754926374e-06, 'learning_rate/action_model': 9.755127549263741e-05, 'epoch': 0.23}) +[2026-06-23 17:34:38,134][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.24675865471363068, 'train/grad_norm_pre_clip': 16.441993713378906, 'timing/data': 0.0005965810269117355, 'timing/model': 1.6820442890748382, 'learning_rate/qwen_vl_interface': 9.754161428614325e-06, 'learning_rate/action_model': 9.754161428614324e-05, 'epoch': 0.23}) +[2026-06-23 17:34:41,462][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.38123831152915955, 'train/grad_norm_pre_clip': 65.34088897705078, 'timing/data': 0.0006143751088529825, 'timing/model': 1.6930842271540314, 'learning_rate/qwen_vl_interface': 9.753193456474284e-06, 'learning_rate/action_model': 9.753193456474284e-05, 'epoch': 0.23}) +[2026-06-23 17:34:44,791][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.20438626408576965, 'train/grad_norm_pre_clip': 35.8604736328125, 'timing/data': 0.0006069710943847895, 'timing/model': 1.6911420579999685, 'learning_rate/qwen_vl_interface': 9.752223633241514e-06, 'learning_rate/action_model': 9.752223633241514e-05, 'epoch': 0.23}) +[2026-06-23 17:34:48,125][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.16791363060474396, 'train/grad_norm_pre_clip': 35.349308013916016, 'timing/data': 0.0005946678575128317, 'timing/model': 1.690419714199379, 'learning_rate/qwen_vl_interface': 9.751251959314671e-06, 'learning_rate/action_model': 9.751251959314672e-05, 'epoch': 0.23}) +[2026-06-23 17:34:51,507][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.23266920447349548, 'train/grad_norm_pre_clip': 17.913021087646484, 'timing/data': 0.0006097070872783661, 'timing/model': 1.7515357530210167, 'learning_rate/qwen_vl_interface': 9.750278435093176e-06, 'learning_rate/action_model': 9.750278435093177e-05, 'epoch': 0.24}) +[2026-06-23 17:34:54,840][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.17362339794635773, 'train/grad_norm_pre_clip': 21.44915199279785, 'timing/data': 0.0006118700839579105, 'timing/model': 1.6897091649007052, 'learning_rate/qwen_vl_interface': 9.749303060977205e-06, 'learning_rate/action_model': 9.749303060977205e-05, 'epoch': 0.24}) +[2026-06-23 17:34:58,164][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.14812041819095612, 'train/grad_norm_pre_clip': 14.114258766174316, 'timing/data': 0.0006139939650893211, 'timing/model': 1.6859062591101974, 'learning_rate/qwen_vl_interface': 9.748325837367701e-06, 'learning_rate/action_model': 9.748325837367702e-05, 'epoch': 0.24}) +[2026-06-23 17:35:01,498][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.2323218733072281, 'train/grad_norm_pre_clip': 62.40727233886719, 'timing/data': 0.000761992996558547, 'timing/model': 1.6938131269998848, 'learning_rate/qwen_vl_interface': 9.747346764666362e-06, 'learning_rate/action_model': 9.747346764666361e-05, 'epoch': 0.24}) +[2026-06-23 17:35:04,840][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.1526508778333664, 'train/grad_norm_pre_clip': 17.917287826538086, 'timing/data': 0.0006743480917066336, 'timing/model': 1.6926931850612164, 'learning_rate/qwen_vl_interface': 9.746365843275648e-06, 'learning_rate/action_model': 9.746365843275646e-05, 'epoch': 0.24}) +[2026-06-23 17:35:08,181][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.23846983909606934, 'train/grad_norm_pre_clip': 24.304454803466797, 'timing/data': 0.0005979840643703938, 'timing/model': 1.696705561131239, 'learning_rate/qwen_vl_interface': 9.745383073598776e-06, 'learning_rate/action_model': 9.745383073598775e-05, 'epoch': 0.24}) +[2026-06-23 17:35:11,518][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.30657967925071716, 'train/grad_norm_pre_clip': 73.82829284667969, 'timing/data': 0.000663186889141798, 'timing/model': 1.6917818279471248, 'learning_rate/qwen_vl_interface': 9.744398456039729e-06, 'learning_rate/action_model': 9.744398456039729e-05, 'epoch': 0.24}) +[2026-06-23 17:35:14,845][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.322049081325531, 'train/grad_norm_pre_clip': 51.550804138183594, 'timing/data': 0.0006688970606774092, 'timing/model': 1.6916823650244623, 'learning_rate/qwen_vl_interface': 9.743411991003246e-06, 'learning_rate/action_model': 9.743411991003247e-05, 'epoch': 0.24}) +[2026-06-23 17:35:18,174][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.3333354890346527, 'train/grad_norm_pre_clip': 16.12258529663086, 'timing/data': 0.0006013098172843456, 'timing/model': 1.6873043179512024, 'learning_rate/qwen_vl_interface': 9.742423678894823e-06, 'learning_rate/action_model': 9.742423678894823e-05, 'epoch': 0.24}) +[2026-06-23 17:35:21,516][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.2094915211200714, 'train/grad_norm_pre_clip': 98.53215789794922, 'timing/data': 0.0006690679583698511, 'timing/model': 1.6970969920512289, 'learning_rate/qwen_vl_interface': 9.741433520120722e-06, 'learning_rate/action_model': 9.74143352012072e-05, 'epoch': 0.24}) +[2026-06-23 17:35:24,852][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.2221292406320572, 'train/grad_norm_pre_clip': 20.70049476623535, 'timing/data': 0.0005905809812247753, 'timing/model': 1.6920487249735743, 'learning_rate/qwen_vl_interface': 9.740441515087953e-06, 'learning_rate/action_model': 9.740441515087953e-05, 'epoch': 0.24}) +[2026-06-23 17:35:28,178][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.24378564953804016, 'train/grad_norm_pre_clip': 15.745842933654785, 'timing/data': 0.0006074320990592241, 'timing/model': 1.687646342907101, 'learning_rate/qwen_vl_interface': 9.739447664204298e-06, 'learning_rate/action_model': 9.739447664204297e-05, 'epoch': 0.24}) +[2026-06-23 17:35:31,514][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.3067602515220642, 'train/grad_norm_pre_clip': 31.891887664794922, 'timing/data': 0.0006505930796265602, 'timing/model': 1.6891959900967777, 'learning_rate/qwen_vl_interface': 9.73845196787829e-06, 'learning_rate/action_model': 9.738451967878289e-05, 'epoch': 0.24}) +[2026-06-23 17:35:34,852][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.22659967839717865, 'train/grad_norm_pre_clip': 17.22186851501465, 'timing/data': 0.0005953491199761629, 'timing/model': 1.6944966781884432, 'learning_rate/qwen_vl_interface': 9.737454426519221e-06, 'learning_rate/action_model': 9.737454426519221e-05, 'epoch': 0.24}) +[2026-06-23 17:35:38,183][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.4304725229740143, 'train/grad_norm_pre_clip': 19.166030883789062, 'timing/data': 0.0006852878723293543, 'timing/model': 1.69043730199337, 'learning_rate/qwen_vl_interface': 9.736455040537145e-06, 'learning_rate/action_model': 9.736455040537144e-05, 'epoch': 0.24}) +[2026-06-23 17:35:41,597][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.15531392395496368, 'train/grad_norm_pre_clip': 32.97993850708008, 'timing/data': 0.0007031520362943411, 'timing/model': 1.6941088871099055, 'learning_rate/qwen_vl_interface': 9.735453810342868e-06, 'learning_rate/action_model': 9.735453810342867e-05, 'epoch': 0.24}) +[2026-06-23 17:35:44,928][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.33458054065704346, 'train/grad_norm_pre_clip': 101.16029357910156, 'timing/data': 0.0005986250471323729, 'timing/model': 1.6959783360362053, 'learning_rate/qwen_vl_interface': 9.734450736347958e-06, 'learning_rate/action_model': 9.73445073634796e-05, 'epoch': 0.24}) +[2026-06-23 17:35:48,261][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.3673076033592224, 'train/grad_norm_pre_clip': 73.89812469482422, 'timing/data': 0.0006967899389564991, 'timing/model': 1.6870516689959913, 'learning_rate/qwen_vl_interface': 9.733445818964746e-06, 'learning_rate/action_model': 9.733445818964746e-05, 'epoch': 0.24}) +[2026-06-23 17:35:51,593][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.29233258962631226, 'train/grad_norm_pre_clip': 27.121471405029297, 'timing/data': 0.0007080910727381706, 'timing/model': 1.6907800810877234, 'learning_rate/qwen_vl_interface': 9.732439058606311e-06, 'learning_rate/action_model': 9.732439058606312e-05, 'epoch': 0.24}) +[2026-06-23 17:35:54,931][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.22839532792568207, 'train/grad_norm_pre_clip': 12.650405883789062, 'timing/data': 0.0006027529016137123, 'timing/model': 1.6946287308819592, 'learning_rate/qwen_vl_interface': 9.731430455686499e-06, 'learning_rate/action_model': 9.731430455686498e-05, 'epoch': 0.24}) +[2026-06-23 17:35:58,264][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.24650856852531433, 'train/grad_norm_pre_clip': 27.277721405029297, 'timing/data': 0.0006078728474676609, 'timing/model': 1.6923714322037995, 'learning_rate/qwen_vl_interface': 9.730420010619901e-06, 'learning_rate/action_model': 9.7304200106199e-05, 'epoch': 0.24}) +[2026-06-23 17:36:01,593][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.21876394748687744, 'train/grad_norm_pre_clip': 21.067964553833008, 'timing/data': 0.0005961710121482611, 'timing/model': 1.6918999899644405, 'learning_rate/qwen_vl_interface': 9.72940772382188e-06, 'learning_rate/action_model': 9.729407723821878e-05, 'epoch': 0.24}) +[2026-06-23 17:36:04,921][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.29360532760620117, 'train/grad_norm_pre_clip': 44.6309814453125, 'timing/data': 0.0006082840263843536, 'timing/model': 1.6866722260601819, 'learning_rate/qwen_vl_interface': 9.728393595708542e-06, 'learning_rate/action_model': 9.728393595708542e-05, 'epoch': 0.24}) +[2026-06-23 17:36:08,256][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.23957185447216034, 'train/grad_norm_pre_clip': 16.447046279907227, 'timing/data': 0.0006037449929863214, 'timing/model': 1.6946746530011296, 'learning_rate/qwen_vl_interface': 9.727377626696762e-06, 'learning_rate/action_model': 9.727377626696762e-05, 'epoch': 0.24}) +[2026-06-23 17:36:11,595][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.24343685805797577, 'train/grad_norm_pre_clip': 54.707340240478516, 'timing/data': 0.000661172904074192, 'timing/model': 1.696409820113331, 'learning_rate/qwen_vl_interface': 9.726359817204165e-06, 'learning_rate/action_model': 9.726359817204166e-05, 'epoch': 0.24}) +[2026-06-23 17:36:14,926][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.13679149746894836, 'train/grad_norm_pre_clip': 12.824528694152832, 'timing/data': 0.000681981910020113, 'timing/model': 1.6919377068988979, 'learning_rate/qwen_vl_interface': 9.725340167649136e-06, 'learning_rate/action_model': 9.725340167649136e-05, 'epoch': 0.24}) +[2026-06-23 17:36:18,251][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.2662983238697052, 'train/grad_norm_pre_clip': 39.867698669433594, 'timing/data': 0.0005963209550827742, 'timing/model': 1.6860644309781492, 'learning_rate/qwen_vl_interface': 9.72431867845081e-06, 'learning_rate/action_model': 9.72431867845081e-05, 'epoch': 0.25}) +[2026-06-23 17:36:21,584][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.2781122922897339, 'train/grad_norm_pre_clip': 33.49360656738281, 'timing/data': 0.000605498906224966, 'timing/model': 1.6940631221514195, 'learning_rate/qwen_vl_interface': 9.723295350029086e-06, 'learning_rate/action_model': 9.723295350029086e-05, 'epoch': 0.25}) +[2026-06-23 17:36:24,923][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.14963853359222412, 'train/grad_norm_pre_clip': 22.524883270263672, 'timing/data': 0.0006206969264894724, 'timing/model': 1.692281886935234, 'learning_rate/qwen_vl_interface': 9.722270182804616e-06, 'learning_rate/action_model': 9.722270182804614e-05, 'epoch': 0.25}) +[2026-06-23 17:36:28,256][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.20445770025253296, 'train/grad_norm_pre_clip': 81.94581604003906, 'timing/data': 0.0006045859772711992, 'timing/model': 1.6931805510539562, 'learning_rate/qwen_vl_interface': 9.721243177198805e-06, 'learning_rate/action_model': 9.721243177198804e-05, 'epoch': 0.25}) +[2026-06-23 17:36:31,887][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.2715264558792114, 'train/grad_norm_pre_clip': 139.30172729492188, 'timing/data': 0.000608684029430151, 'timing/model': 1.6894284321460873, 'learning_rate/qwen_vl_interface': 9.720214333633817e-06, 'learning_rate/action_model': 9.720214333633817e-05, 'epoch': 0.25}) +[2026-06-23 17:36:35,216][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.35326266288757324, 'train/grad_norm_pre_clip': 163.07276916503906, 'timing/data': 0.0005991670768707991, 'timing/model': 1.6895473299082369, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.25}) +[2026-06-23 17:36:38,545][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.16799551248550415, 'train/grad_norm_pre_clip': 11.017184257507324, 'timing/data': 0.0005998380947858095, 'timing/model': 1.6905614009592682, 'learning_rate/qwen_vl_interface': 9.718151134318743e-06, 'learning_rate/action_model': 9.718151134318743e-05, 'epoch': 0.25}) +[2026-06-23 17:36:41,876][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.4357689321041107, 'train/grad_norm_pre_clip': 114.08940887451172, 'timing/data': 0.0006050171796232462, 'timing/model': 1.6860370610374957, 'learning_rate/qwen_vl_interface': 9.71711677941676e-06, 'learning_rate/action_model': 9.717116779416758e-05, 'epoch': 0.25}) +[2026-06-23 17:36:45,219][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.13005879521369934, 'train/grad_norm_pre_clip': 35.95862579345703, 'timing/data': 0.0005956001114100218, 'timing/model': 1.6951166139915586, 'learning_rate/qwen_vl_interface': 9.716080588251805e-06, 'learning_rate/action_model': 9.716080588251805e-05, 'epoch': 0.25}) +[2026-06-23 17:36:48,546][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.1754826307296753, 'train/grad_norm_pre_clip': 64.18547821044922, 'timing/data': 0.0006969098467379808, 'timing/model': 1.686383320018649, 'learning_rate/qwen_vl_interface': 9.71504256124982e-06, 'learning_rate/action_model': 9.71504256124982e-05, 'epoch': 0.25}) +[2026-06-23 17:36:51,875][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.17668496072292328, 'train/grad_norm_pre_clip': 15.022073745727539, 'timing/data': 0.0006132731214165688, 'timing/model': 1.6896749129518867, 'learning_rate/qwen_vl_interface': 9.714002698837499e-06, 'learning_rate/action_model': 9.714002698837498e-05, 'epoch': 0.25}) +[2026-06-23 17:36:55,211][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.40496671199798584, 'train/grad_norm_pre_clip': 56.46922302246094, 'timing/data': 0.0006355650257319212, 'timing/model': 1.6930593301076442, 'learning_rate/qwen_vl_interface': 9.712961001442285e-06, 'learning_rate/action_model': 9.712961001442286e-05, 'epoch': 0.25}) +[2026-06-23 17:36:58,551][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.2219691425561905, 'train/grad_norm_pre_clip': 22.144662857055664, 'timing/data': 0.0007005471270531416, 'timing/model': 1.6932650420349091, 'learning_rate/qwen_vl_interface': 9.711917469492388e-06, 'learning_rate/action_model': 9.711917469492388e-05, 'epoch': 0.25}) +[2026-06-23 17:37:01,888][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.24307814240455627, 'train/grad_norm_pre_clip': 13.466058731079102, 'timing/data': 0.0006791469641029835, 'timing/model': 1.6956104659475386, 'learning_rate/qwen_vl_interface': 9.71087210341676e-06, 'learning_rate/action_model': 9.71087210341676e-05, 'epoch': 0.25}) +[2026-06-23 17:37:05,227][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.24976301193237305, 'train/grad_norm_pre_clip': 24.403841018676758, 'timing/data': 0.0007099651265889406, 'timing/model': 1.6939449899364263, 'learning_rate/qwen_vl_interface': 9.70982490364511e-06, 'learning_rate/action_model': 9.70982490364511e-05, 'epoch': 0.25}) +[2026-06-23 17:37:08,568][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.40742161870002747, 'train/grad_norm_pre_clip': 59.65050506591797, 'timing/data': 0.000665029976516962, 'timing/model': 1.6916483868844807, 'learning_rate/qwen_vl_interface': 9.708775870607905e-06, 'learning_rate/action_model': 9.708775870607904e-05, 'epoch': 0.25}) +[2026-06-23 17:37:11,904][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.1730850338935852, 'train/grad_norm_pre_clip': 20.0224552154541, 'timing/data': 0.0005950289778411388, 'timing/model': 1.6923986130859703, 'learning_rate/qwen_vl_interface': 9.707725004736363e-06, 'learning_rate/action_model': 9.707725004736363e-05, 'epoch': 0.25}) +[2026-06-23 17:37:15,234][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.3023534119129181, 'train/grad_norm_pre_clip': 10.234100341796875, 'timing/data': 0.0006614839658141136, 'timing/model': 1.6958617370110005, 'learning_rate/qwen_vl_interface': 9.706672306462454e-06, 'learning_rate/action_model': 9.706672306462454e-05, 'epoch': 0.25}) +[2026-06-23 17:37:18,571][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.23095303773880005, 'train/grad_norm_pre_clip': 39.19683074951172, 'timing/data': 0.0005962599534541368, 'timing/model': 1.6940325228497386, 'learning_rate/qwen_vl_interface': 9.705617776218905e-06, 'learning_rate/action_model': 9.705617776218904e-05, 'epoch': 0.25}) +[2026-06-23 17:37:22,142][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.2612660825252533, 'train/grad_norm_pre_clip': 46.18599319458008, 'timing/data': 0.0006226110272109509, 'timing/model': 1.6900071220006794, 'learning_rate/qwen_vl_interface': 9.70456141443919e-06, 'learning_rate/action_model': 9.704561414439188e-05, 'epoch': 0.25}) +[2026-06-23 17:37:25,486][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.17869041860103607, 'train/grad_norm_pre_clip': 7.841964244842529, 'timing/data': 0.0006007091142237186, 'timing/model': 1.699057289166376, 'learning_rate/qwen_vl_interface': 9.703503221557538e-06, 'learning_rate/action_model': 9.703503221557538e-05, 'epoch': 0.25}) +[2026-06-23 17:37:28,822][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.24625438451766968, 'train/grad_norm_pre_clip': 10.092735290527344, 'timing/data': 0.0006081541068851948, 'timing/model': 1.6949065709486604, 'learning_rate/qwen_vl_interface': 9.70244319800894e-06, 'learning_rate/action_model': 9.70244319800894e-05, 'epoch': 0.25}) +[2026-06-23 17:37:32,155][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.41635662317276, 'train/grad_norm_pre_clip': 11.659213066101074, 'timing/data': 0.0006707510910928249, 'timing/model': 1.6910711470991373, 'learning_rate/qwen_vl_interface': 9.701381344229124e-06, 'learning_rate/action_model': 9.701381344229123e-05, 'epoch': 0.25}) +[2026-06-23 17:37:35,501][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.20730091631412506, 'train/grad_norm_pre_clip': 41.778465270996094, 'timing/data': 0.000740633113309741, 'timing/model': 1.6977376940194517, 'learning_rate/qwen_vl_interface': 9.700317660654581e-06, 'learning_rate/action_model': 9.700317660654581e-05, 'epoch': 0.25}) +[2026-06-23 17:37:38,837][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.2628835141658783, 'train/grad_norm_pre_clip': 45.4039306640625, 'timing/data': 0.0006030139047652483, 'timing/model': 1.6901805230882019, 'learning_rate/qwen_vl_interface': 9.699252147722554e-06, 'learning_rate/action_model': 9.699252147722553e-05, 'epoch': 0.25}) +[2026-06-23 17:37:42,174][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.27762627601623535, 'train/grad_norm_pre_clip': 23.57503890991211, 'timing/data': 0.0006099359598010778, 'timing/model': 1.6935626871418208, 'learning_rate/qwen_vl_interface': 9.698184805871028e-06, 'learning_rate/action_model': 9.698184805871028e-05, 'epoch': 0.26}) +[2026-06-23 17:37:45,508][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.22940784692764282, 'train/grad_norm_pre_clip': 62.28782653808594, 'timing/data': 0.0006071608513593674, 'timing/model': 1.6923667441587895, 'learning_rate/qwen_vl_interface': 9.697115635538753e-06, 'learning_rate/action_model': 9.697115635538753e-05, 'epoch': 0.26}) +[2026-06-23 17:37:48,843][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.30517032742500305, 'train/grad_norm_pre_clip': 48.72445297241211, 'timing/data': 0.0006619151681661606, 'timing/model': 1.6842207158915699, 'learning_rate/qwen_vl_interface': 9.69604463716522e-06, 'learning_rate/action_model': 9.69604463716522e-05, 'epoch': 0.26}) +[2026-06-23 17:37:52,183][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.21534833312034607, 'train/grad_norm_pre_clip': 42.63652420043945, 'timing/data': 0.0006591288838535547, 'timing/model': 1.6931508381385356, 'learning_rate/qwen_vl_interface': 9.69497181119068e-06, 'learning_rate/action_model': 9.694971811190681e-05, 'epoch': 0.26}) +[2026-06-23 17:37:55,513][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.4290631413459778, 'train/grad_norm_pre_clip': 90.75431823730469, 'timing/data': 0.0006537889130413532, 'timing/model': 1.6973949309904128, 'learning_rate/qwen_vl_interface': 9.693897158056132e-06, 'learning_rate/action_model': 9.693897158056132e-05, 'epoch': 0.26}) +[2026-06-23 17:37:58,850][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.40662533044815063, 'train/grad_norm_pre_clip': 93.62549591064453, 'timing/data': 0.0006733459886163473, 'timing/model': 1.6963293028529733, 'learning_rate/qwen_vl_interface': 9.692820678203322e-06, 'learning_rate/action_model': 9.692820678203322e-05, 'epoch': 0.26}) +[2026-06-23 17:38:02,188][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.3981848359107971, 'train/grad_norm_pre_clip': 121.72682189941406, 'timing/data': 0.0006065608467906713, 'timing/model': 1.694865306140855, 'learning_rate/qwen_vl_interface': 9.691742372074749e-06, 'learning_rate/action_model': 9.691742372074749e-05, 'epoch': 0.26}) +[2026-06-23 17:38:05,538][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.23396793007850647, 'train/grad_norm_pre_clip': 55.751556396484375, 'timing/data': 0.0007756880950182676, 'timing/model': 1.7026655650697649, 'learning_rate/qwen_vl_interface': 9.690662240113669e-06, 'learning_rate/action_model': 9.690662240113668e-05, 'epoch': 0.26}) +[2026-06-23 17:38:09,015][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.20385263860225677, 'train/grad_norm_pre_clip': 37.10930252075195, 'timing/data': 0.0006723939441144466, 'timing/model': 1.6898299420718104, 'learning_rate/qwen_vl_interface': 9.689580282764077e-06, 'learning_rate/action_model': 9.689580282764077e-05, 'epoch': 0.26}) +[2026-06-23 17:38:12,349][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.2638733386993408, 'train/grad_norm_pre_clip': 62.1284294128418, 'timing/data': 0.0006061990279704332, 'timing/model': 1.6958010620437562, 'learning_rate/qwen_vl_interface': 9.688496500470731e-06, 'learning_rate/action_model': 9.68849650047073e-05, 'epoch': 0.26}) +[2026-06-23 17:38:15,687][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.28253892064094543, 'train/grad_norm_pre_clip': 63.43735885620117, 'timing/data': 0.0007013489957898855, 'timing/model': 1.6966993960086256, 'learning_rate/qwen_vl_interface': 9.68741089367913e-06, 'learning_rate/action_model': 9.687410893679129e-05, 'epoch': 0.26}) +[2026-06-23 17:38:19,018][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.3144475817680359, 'train/grad_norm_pre_clip': 66.12962341308594, 'timing/data': 0.000635514035820961, 'timing/model': 1.6959812419954687, 'learning_rate/qwen_vl_interface': 9.686323462835527e-06, 'learning_rate/action_model': 9.686323462835526e-05, 'epoch': 0.26}) +[2026-06-23 17:38:22,351][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.27648797631263733, 'train/grad_norm_pre_clip': 12.730224609375, 'timing/data': 0.0005953891668468714, 'timing/model': 1.6921443450264633, 'learning_rate/qwen_vl_interface': 9.685234208386922e-06, 'learning_rate/action_model': 9.685234208386921e-05, 'epoch': 0.26}) +[2026-06-23 17:38:25,686][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.4306827187538147, 'train/grad_norm_pre_clip': 86.6072006225586, 'timing/data': 0.0005798297934234142, 'timing/model': 1.6928384848870337, 'learning_rate/qwen_vl_interface': 9.684143130781066e-06, 'learning_rate/action_model': 9.684143130781066e-05, 'epoch': 0.26}) +[2026-06-23 17:38:29,027][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.3322181701660156, 'train/grad_norm_pre_clip': 34.22093200683594, 'timing/data': 0.0005858109798282385, 'timing/model': 1.6952393939718604, 'learning_rate/qwen_vl_interface': 9.683050230466462e-06, 'learning_rate/action_model': 9.683050230466462e-05, 'epoch': 0.26}) +[2026-06-23 17:38:32,364][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.3003736436367035, 'train/grad_norm_pre_clip': 10.025832176208496, 'timing/data': 0.0005903400015085936, 'timing/model': 1.6938903231639415, 'learning_rate/qwen_vl_interface': 9.681955507892361e-06, 'learning_rate/action_model': 9.68195550789236e-05, 'epoch': 0.26}) +[2026-06-23 17:38:35,690][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.28141626715660095, 'train/grad_norm_pre_clip': 27.822742462158203, 'timing/data': 0.0006068910006433725, 'timing/model': 1.6913005111273378, 'learning_rate/qwen_vl_interface': 9.680858963508759e-06, 'learning_rate/action_model': 9.68085896350876e-05, 'epoch': 0.26}) +[2026-06-23 17:38:39,024][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.23157957196235657, 'train/grad_norm_pre_clip': 16.855865478515625, 'timing/data': 0.0006705299019813538, 'timing/model': 1.6956866709515452, 'learning_rate/qwen_vl_interface': 9.679760597766405e-06, 'learning_rate/action_model': 9.679760597766405e-05, 'epoch': 0.26}) +[2026-06-23 17:38:42,362][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.25174903869628906, 'train/grad_norm_pre_clip': 59.06988525390625, 'timing/data': 0.0006201360374689102, 'timing/model': 1.6946904549840838, 'learning_rate/qwen_vl_interface': 9.678660411116796e-06, 'learning_rate/action_model': 9.678660411116796e-05, 'epoch': 0.26}) +[2026-06-23 17:38:45,696][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.19060201942920685, 'train/grad_norm_pre_clip': 48.110965728759766, 'timing/data': 0.0006005691830068827, 'timing/model': 1.6882059208583087, 'learning_rate/qwen_vl_interface': 9.677558404012178e-06, 'learning_rate/action_model': 9.677558404012177e-05, 'epoch': 0.26}) +[2026-06-23 17:38:49,032][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.179653599858284, 'train/grad_norm_pre_clip': 20.12313461303711, 'timing/data': 0.0006136039737612009, 'timing/model': 1.6942506600171328, 'learning_rate/qwen_vl_interface': 9.676454576905542e-06, 'learning_rate/action_model': 9.676454576905542e-05, 'epoch': 0.26}) +[2026-06-23 17:38:52,372][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.2850407063961029, 'train/grad_norm_pre_clip': 42.26376724243164, 'timing/data': 0.0005949579644948244, 'timing/model': 1.694083219859749, 'learning_rate/qwen_vl_interface': 9.675348930250633e-06, 'learning_rate/action_model': 9.675348930250632e-05, 'epoch': 0.26}) +[2026-06-23 17:38:55,841][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.32549381256103516, 'train/grad_norm_pre_clip': 89.5211410522461, 'timing/data': 0.0005978739354759455, 'timing/model': 1.83000790909864, 'learning_rate/qwen_vl_interface': 9.674241464501938e-06, 'learning_rate/action_model': 9.674241464501937e-05, 'epoch': 0.26}) +[2026-06-23 17:38:59,174][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.2096869796514511, 'train/grad_norm_pre_clip': 94.31221771240234, 'timing/data': 0.0005971519276499748, 'timing/model': 1.6892052330076694, 'learning_rate/qwen_vl_interface': 9.673132180114695e-06, 'learning_rate/action_model': 9.673132180114695e-05, 'epoch': 0.26}) +[2026-06-23 17:39:02,503][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.22832362353801727, 'train/grad_norm_pre_clip': 11.083212852478027, 'timing/data': 0.0005967821925878525, 'timing/model': 1.6925446558743715, 'learning_rate/qwen_vl_interface': 9.67202107754489e-06, 'learning_rate/action_model': 9.672021077544889e-05, 'epoch': 0.26}) +[2026-06-23 17:39:05,837][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.4523487389087677, 'train/grad_norm_pre_clip': 68.32688903808594, 'timing/data': 0.0009166328236460686, 'timing/model': 1.6987185580655932, 'learning_rate/qwen_vl_interface': 9.670908157249254e-06, 'learning_rate/action_model': 9.670908157249253e-05, 'epoch': 0.26}) +[2026-06-23 17:39:09,171][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.3487531542778015, 'train/grad_norm_pre_clip': 76.21760559082031, 'timing/data': 0.0008155228570103645, 'timing/model': 1.694010853767395, 'learning_rate/qwen_vl_interface': 9.669793419685267e-06, 'learning_rate/action_model': 9.669793419685266e-05, 'epoch': 0.27}) +[2026-06-23 17:39:12,512][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.28263425827026367, 'train/grad_norm_pre_clip': 57.31999969482422, 'timing/data': 0.0008336668834090233, 'timing/model': 1.6987156339455396, 'learning_rate/qwen_vl_interface': 9.668676865311155e-06, 'learning_rate/action_model': 9.668676865311154e-05, 'epoch': 0.27}) +[2026-06-23 17:39:15,854][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.23989449441432953, 'train/grad_norm_pre_clip': 92.5878677368164, 'timing/data': 0.000823308015242219, 'timing/model': 1.7012543121818453, 'learning_rate/qwen_vl_interface': 9.667558494585891e-06, 'learning_rate/action_model': 9.66755849458589e-05, 'epoch': 0.27}) +[2026-06-23 17:39:19,193][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.16850492358207703, 'train/grad_norm_pre_clip': 36.59069061279297, 'timing/data': 0.000603965949267149, 'timing/model': 1.6953261939343065, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.27}) +[2026-06-23 17:39:22,522][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.21988677978515625, 'train/grad_norm_pre_clip': 58.692161560058594, 'timing/data': 0.0006030029617249966, 'timing/model': 1.6902956748381257, 'learning_rate/qwen_vl_interface': 9.665316305921535e-06, 'learning_rate/action_model': 9.665316305921535e-05, 'epoch': 0.27}) +[2026-06-23 17:39:25,854][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.22590579092502594, 'train/grad_norm_pre_clip': 65.20758056640625, 'timing/data': 0.0006004590541124344, 'timing/model': 1.6894884840585291, 'learning_rate/qwen_vl_interface': 9.664192488904123e-06, 'learning_rate/action_model': 9.664192488904122e-05, 'epoch': 0.27}) +[2026-06-23 17:39:29,195][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.16005241870880127, 'train/grad_norm_pre_clip': 24.677059173583984, 'timing/data': 0.0006079019512981176, 'timing/model': 1.696702980902046, 'learning_rate/qwen_vl_interface': 9.663066857378917e-06, 'learning_rate/action_model': 9.663066857378917e-05, 'epoch': 0.27}) +[2026-06-23 17:39:32,521][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.3688635230064392, 'train/grad_norm_pre_clip': 76.16184997558594, 'timing/data': 0.0006078521255403757, 'timing/model': 1.6875906388740987, 'learning_rate/qwen_vl_interface': 9.661939411808623e-06, 'learning_rate/action_model': 9.661939411808622e-05, 'epoch': 0.27}) +[2026-06-23 17:39:35,858][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.4329800307750702, 'train/grad_norm_pre_clip': 101.63319396972656, 'timing/data': 0.0006172000430524349, 'timing/model': 1.6938005480915308, 'learning_rate/qwen_vl_interface': 9.66081015265669e-06, 'learning_rate/action_model': 9.660810152656689e-05, 'epoch': 0.27}) +[2026-06-23 17:39:39,202][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.15034006536006927, 'train/grad_norm_pre_clip': 94.30683135986328, 'timing/data': 0.0006039759609848261, 'timing/model': 1.6954993670806289, 'learning_rate/qwen_vl_interface': 9.659679080387314e-06, 'learning_rate/action_model': 9.659679080387313e-05, 'epoch': 0.27}) +[2026-06-23 17:39:42,543][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.23322106897830963, 'train/grad_norm_pre_clip': 34.94524002075195, 'timing/data': 0.000596189871430397, 'timing/model': 1.700954345986247, 'learning_rate/qwen_vl_interface': 9.658546195465436e-06, 'learning_rate/action_model': 9.658546195465435e-05, 'epoch': 0.27}) +[2026-06-23 17:39:45,969][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.22146350145339966, 'train/grad_norm_pre_clip': 50.199432373046875, 'timing/data': 0.0005969021003693342, 'timing/model': 1.7864978790748864, 'learning_rate/qwen_vl_interface': 9.65741149835674e-06, 'learning_rate/action_model': 9.657411498356741e-05, 'epoch': 0.27}) +[2026-06-23 17:39:49,311][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.15910261869430542, 'train/grad_norm_pre_clip': 85.35821533203125, 'timing/data': 0.0006493909750133753, 'timing/model': 1.6982553259003907, 'learning_rate/qwen_vl_interface': 9.656274989527663e-06, 'learning_rate/action_model': 9.656274989527664e-05, 'epoch': 0.27}) +[2026-06-23 17:39:52,647][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.16490744054317474, 'train/grad_norm_pre_clip': 75.2925033569336, 'timing/data': 0.000600307947024703, 'timing/model': 1.695255773840472, 'learning_rate/qwen_vl_interface': 9.655136669445378e-06, 'learning_rate/action_model': 9.655136669445378e-05, 'epoch': 0.27}) +[2026-06-23 17:39:55,985][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.2331351786851883, 'train/grad_norm_pre_clip': 28.418804168701172, 'timing/data': 0.0006667429115623236, 'timing/model': 1.69575858511962, 'learning_rate/qwen_vl_interface': 9.653996538577802e-06, 'learning_rate/action_model': 9.653996538577802e-05, 'epoch': 0.27}) +[2026-06-23 17:39:59,319][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 700 +[2026-06-23 17:41:32,187][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 700 +[2026-06-23 17:41:32,188][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 700 +[2026-06-23 17:42:43,359][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 700 +[2026-06-23 17:42:43,360][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.22901949286460876, 'train/grad_norm_pre_clip': 31.117185592651367, 'eval/action_loss/seconds': 92.8664910371881, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.13629919290542603, 'eval/latency_2/loss': 0.13629919290542603, 'eval/flappy/loss': 0.13629919290542603, 'eval/flappy/latency_2/loss': 0.13629919290542603, 'eval/flap_precision': 0.7814569536423841, 'eval/flap_recall': 0.9554655870445344, 'eval/flap_f1': 0.8597449908925319, 'eval/flap_rate_model': 0.25166666666666665, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8597449908925318, 'eval/latency_2/flap_precision': 0.7814569536423841, 'eval/latency_2/flap_recall': 0.9554655870445344, 'eval/latency_2/flap_f1': 0.8597449908925319, 'eval/latency_2/flap_rate_model': 0.25166666666666665, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8597449908925318, 'eval/action_classification/data_wait_seconds': 0.9153331518173218, 'eval/action_classification/predict_seconds': 70.10014343261719, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006745480932295322, 'timing/model': 1.6915859489236027, 'learning_rate/qwen_vl_interface': 9.652854597393604e-06, 'learning_rate/action_model': 9.652854597393603e-05, 'epoch': 0.27}) +[2026-06-23 17:42:46,671][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.2482939064502716, 'train/grad_norm_pre_clip': 49.23087692260742, 'timing/data': 0.0005832160823047161, 'timing/model': 1.6808106689713895, 'learning_rate/qwen_vl_interface': 9.651710846362192e-06, 'learning_rate/action_model': 9.651710846362192e-05, 'epoch': 0.27}) +[2026-06-23 17:42:49,978][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.2037278562784195, 'train/grad_norm_pre_clip': 40.58707809448242, 'timing/data': 0.0006027319468557835, 'timing/model': 1.6794024901464581, 'learning_rate/qwen_vl_interface': 9.650565285953716e-06, 'learning_rate/action_model': 9.650565285953715e-05, 'epoch': 0.27}) +[2026-06-23 17:42:53,291][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.26418402791023254, 'train/grad_norm_pre_clip': 27.537017822265625, 'timing/data': 0.0006000280845910311, 'timing/model': 1.6863908229861408, 'learning_rate/qwen_vl_interface': 9.649417916639077e-06, 'learning_rate/action_model': 9.649417916639076e-05, 'epoch': 0.27}) +[2026-06-23 17:42:56,615][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.2730291485786438, 'train/grad_norm_pre_clip': 70.20277404785156, 'timing/data': 0.0006085340864956379, 'timing/model': 1.688122133957222, 'learning_rate/qwen_vl_interface': 9.648268738889911e-06, 'learning_rate/action_model': 9.648268738889911e-05, 'epoch': 0.27}) +[2026-06-23 17:43:00,040][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.33792123198509216, 'train/grad_norm_pre_clip': 96.12322235107422, 'timing/data': 0.0005962310824543238, 'timing/model': 1.689270048867911, 'learning_rate/qwen_vl_interface': 9.647117753178604e-06, 'learning_rate/action_model': 9.647117753178604e-05, 'epoch': 0.27}) +[2026-06-23 17:43:03,358][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.40277913212776184, 'train/grad_norm_pre_clip': 109.79093933105469, 'timing/data': 0.0006508231163024902, 'timing/model': 1.6821888969279826, 'learning_rate/qwen_vl_interface': 9.645964959978284e-06, 'learning_rate/action_model': 9.645964959978284e-05, 'epoch': 0.27}) +[2026-06-23 17:43:06,681][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.1397608369588852, 'train/grad_norm_pre_clip': 21.78580665588379, 'timing/data': 0.0005924631841480732, 'timing/model': 1.690131212119013, 'learning_rate/qwen_vl_interface': 9.644810359762818e-06, 'learning_rate/action_model': 9.644810359762817e-05, 'epoch': 0.27}) +[2026-06-23 17:43:10,003][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.17327068746089935, 'train/grad_norm_pre_clip': 21.555356979370117, 'timing/data': 0.0006007789634168148, 'timing/model': 1.6880197520367801, 'learning_rate/qwen_vl_interface': 9.64365395300682e-06, 'learning_rate/action_model': 9.643653953006819e-05, 'epoch': 0.27}) +[2026-06-23 17:43:13,331][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.16191521286964417, 'train/grad_norm_pre_clip': 47.24591827392578, 'timing/data': 0.0007372461259365082, 'timing/model': 1.6898210749495775, 'learning_rate/qwen_vl_interface': 9.642495740185644e-06, 'learning_rate/action_model': 9.642495740185643e-05, 'epoch': 0.27}) +[2026-06-23 17:43:16,660][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.20115582644939423, 'train/grad_norm_pre_clip': 68.99817657470703, 'timing/data': 0.0005836568307131529, 'timing/model': 1.6916875499300659, 'learning_rate/qwen_vl_interface': 9.641335721775388e-06, 'learning_rate/action_model': 9.641335721775387e-05, 'epoch': 0.27}) +[2026-06-23 17:43:19,989][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.31122368574142456, 'train/grad_norm_pre_clip': 11.738932609558105, 'timing/data': 0.0005963509902358055, 'timing/model': 1.689163584029302, 'learning_rate/qwen_vl_interface': 9.640173898252893e-06, 'learning_rate/action_model': 9.640173898252892e-05, 'epoch': 0.28}) +[2026-06-23 17:43:23,309][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.2463512122631073, 'train/grad_norm_pre_clip': 18.209026336669922, 'timing/data': 0.0006450631190091372, 'timing/model': 1.6811699168756604, 'learning_rate/qwen_vl_interface': 9.639010270095738e-06, 'learning_rate/action_model': 9.639010270095739e-05, 'epoch': 0.28}) +[2026-06-23 17:43:26,631][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.24065963923931122, 'train/grad_norm_pre_clip': 48.36061096191406, 'timing/data': 0.0006097659934312105, 'timing/model': 1.6888897339813411, 'learning_rate/qwen_vl_interface': 9.637844837782251e-06, 'learning_rate/action_model': 9.63784483778225e-05, 'epoch': 0.28}) +[2026-06-23 17:43:29,959][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.17182521522045135, 'train/grad_norm_pre_clip': 39.6563720703125, 'timing/data': 0.0006006290204823017, 'timing/model': 1.6897088321857154, 'learning_rate/qwen_vl_interface': 9.636677601791494e-06, 'learning_rate/action_model': 9.636677601791493e-05, 'epoch': 0.28}) +[2026-06-23 17:43:33,294][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.15915577113628387, 'train/grad_norm_pre_clip': 17.257532119750977, 'timing/data': 0.0005880750250071287, 'timing/model': 1.693176596891135, 'learning_rate/qwen_vl_interface': 9.635508562603273e-06, 'learning_rate/action_model': 9.635508562603272e-05, 'epoch': 0.28}) +[2026-06-23 17:43:36,615][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.19730478525161743, 'train/grad_norm_pre_clip': 13.52623176574707, 'timing/data': 0.0006102370098233223, 'timing/model': 1.6874862690456212, 'learning_rate/qwen_vl_interface': 9.634337720698139e-06, 'learning_rate/action_model': 9.634337720698137e-05, 'epoch': 0.28}) +[2026-06-23 17:43:39,947][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.17172957956790924, 'train/grad_norm_pre_clip': 10.754497528076172, 'timing/data': 0.0006047459319233894, 'timing/model': 1.6948396381922066, 'learning_rate/qwen_vl_interface': 9.633165076557376e-06, 'learning_rate/action_model': 9.633165076557375e-05, 'epoch': 0.28}) +[2026-06-23 17:43:43,277][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.19938653707504272, 'train/grad_norm_pre_clip': 44.250057220458984, 'timing/data': 0.0005948080215603113, 'timing/model': 1.6887885020114481, 'learning_rate/qwen_vl_interface': 9.631990630663019e-06, 'learning_rate/action_model': 9.631990630663018e-05, 'epoch': 0.28}) +[2026-06-23 17:43:46,613][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.209010511636734, 'train/grad_norm_pre_clip': 12.510675430297852, 'timing/data': 0.0006083541084080935, 'timing/model': 1.6933927240315825, 'learning_rate/qwen_vl_interface': 9.630814383497836e-06, 'learning_rate/action_model': 9.630814383497835e-05, 'epoch': 0.28}) +[2026-06-23 17:43:50,244][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.2614479064941406, 'train/grad_norm_pre_clip': 73.24919128417969, 'timing/data': 0.0006144349463284016, 'timing/model': 1.6926449448801577, 'learning_rate/qwen_vl_interface': 9.629636335545336e-06, 'learning_rate/action_model': 9.629636335545335e-05, 'epoch': 0.28}) +[2026-06-23 17:43:53,570][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.33100640773773193, 'train/grad_norm_pre_clip': 17.06214141845703, 'timing/data': 0.000591641990467906, 'timing/model': 1.6873176000081003, 'learning_rate/qwen_vl_interface': 9.628456487289773e-06, 'learning_rate/action_model': 9.628456487289773e-05, 'epoch': 0.28}) +[2026-06-23 17:43:56,900][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.3480892479419708, 'train/grad_norm_pre_clip': 22.339767456054688, 'timing/data': 0.000609787181019783, 'timing/model': 1.6906051668338478, 'learning_rate/qwen_vl_interface': 9.627274839216135e-06, 'learning_rate/action_model': 9.627274839216136e-05, 'epoch': 0.28}) +[2026-06-23 17:44:00,228][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.2742901146411896, 'train/grad_norm_pre_clip': 61.22562026977539, 'timing/data': 0.0005921430420130491, 'timing/model': 1.6905520958825946, 'learning_rate/qwen_vl_interface': 9.626091391810158e-06, 'learning_rate/action_model': 9.626091391810157e-05, 'epoch': 0.28}) +[2026-06-23 17:44:03,551][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.16934964060783386, 'train/grad_norm_pre_clip': 23.323890686035156, 'timing/data': 0.0009767059236764908, 'timing/model': 1.6841571528930217, 'learning_rate/qwen_vl_interface': 9.624906145558309e-06, 'learning_rate/action_model': 9.624906145558307e-05, 'epoch': 0.28}) +[2026-06-23 17:44:06,885][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.18140029907226562, 'train/grad_norm_pre_clip': 82.00306701660156, 'timing/data': 0.0005897581577301025, 'timing/model': 1.6911736601032317, 'learning_rate/qwen_vl_interface': 9.623719100947795e-06, 'learning_rate/action_model': 9.623719100947795e-05, 'epoch': 0.28}) +[2026-06-23 17:44:10,226][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.2338898479938507, 'train/grad_norm_pre_clip': 21.395639419555664, 'timing/data': 0.0005961700808256865, 'timing/model': 1.6969553830567747, 'learning_rate/qwen_vl_interface': 9.622530258466571e-06, 'learning_rate/action_model': 9.62253025846657e-05, 'epoch': 0.28}) +[2026-06-23 17:44:13,560][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.27272576093673706, 'train/grad_norm_pre_clip': 54.700782775878906, 'timing/data': 0.0006090949755162001, 'timing/model': 1.691792296944186, 'learning_rate/qwen_vl_interface': 9.62133961860332e-06, 'learning_rate/action_model': 9.62133961860332e-05, 'epoch': 0.28}) +[2026-06-23 17:44:16,914][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.29484468698501587, 'train/grad_norm_pre_clip': 129.93106079101562, 'timing/data': 0.0005964608862996101, 'timing/model': 1.6895571639761329, 'learning_rate/qwen_vl_interface': 9.620147181847474e-06, 'learning_rate/action_model': 9.620147181847475e-05, 'epoch': 0.28}) +[2026-06-23 17:44:20,256][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.4387910068035126, 'train/grad_norm_pre_clip': 119.07422637939453, 'timing/data': 0.0006034639663994312, 'timing/model': 1.6914479760453105, 'learning_rate/qwen_vl_interface': 9.618952948689196e-06, 'learning_rate/action_model': 9.618952948689196e-05, 'epoch': 0.28}) +[2026-06-23 17:44:23,596][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.2734215259552002, 'train/grad_norm_pre_clip': 64.72954559326172, 'timing/data': 0.0006029328797012568, 'timing/model': 1.6921094770077616, 'learning_rate/qwen_vl_interface': 9.617756919619393e-06, 'learning_rate/action_model': 9.617756919619392e-05, 'epoch': 0.28}) +[2026-06-23 17:44:26,930][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.24420249462127686, 'train/grad_norm_pre_clip': 73.4769287109375, 'timing/data': 0.0005918119568377733, 'timing/model': 1.6926844660192728, 'learning_rate/qwen_vl_interface': 9.616559095129701e-06, 'learning_rate/action_model': 9.6165590951297e-05, 'epoch': 0.28}) +[2026-06-23 17:44:30,258][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.3656526803970337, 'train/grad_norm_pre_clip': 41.122737884521484, 'timing/data': 0.0006013601087033749, 'timing/model': 1.6867555200587958, 'learning_rate/qwen_vl_interface': 9.615359475712506e-06, 'learning_rate/action_model': 9.615359475712506e-05, 'epoch': 0.28}) +[2026-06-23 17:44:33,586][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.17996220290660858, 'train/grad_norm_pre_clip': 40.57463455200195, 'timing/data': 0.0005902091506868601, 'timing/model': 1.689680946059525, 'learning_rate/qwen_vl_interface': 9.614158061860924e-06, 'learning_rate/action_model': 9.614158061860924e-05, 'epoch': 0.28}) +[2026-06-23 17:44:36,919][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.20545509457588196, 'train/grad_norm_pre_clip': 15.436736106872559, 'timing/data': 0.0005918119568377733, 'timing/model': 1.692386840004474, 'learning_rate/qwen_vl_interface': 9.612954854068813e-06, 'learning_rate/action_model': 9.612954854068812e-05, 'epoch': 0.28}) +[2026-06-23 17:44:40,477][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.26292863488197327, 'train/grad_norm_pre_clip': 50.30833053588867, 'timing/data': 0.0005992359947413206, 'timing/model': 1.690384423127398, 'learning_rate/qwen_vl_interface': 9.611749852830763e-06, 'learning_rate/action_model': 9.611749852830763e-05, 'epoch': 0.28}) +[2026-06-23 17:44:43,806][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.23178227245807648, 'train/grad_norm_pre_clip': 53.23656463623047, 'timing/data': 0.0006574960425496101, 'timing/model': 1.6896931729279459, 'learning_rate/qwen_vl_interface': 9.610543058642107e-06, 'learning_rate/action_model': 9.610543058642107e-05, 'epoch': 0.28}) +[2026-06-23 17:44:47,146][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.18058791756629944, 'train/grad_norm_pre_clip': 12.587577819824219, 'timing/data': 0.0005886668805032969, 'timing/model': 1.694499374134466, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.29}) +[2026-06-23 17:44:50,484][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.26877516508102417, 'train/grad_norm_pre_clip': 13.711737632751465, 'timing/data': 0.0006135930307209492, 'timing/model': 1.695559644838795, 'learning_rate/qwen_vl_interface': 9.60812409339798e-06, 'learning_rate/action_model': 9.60812409339798e-05, 'epoch': 0.29}) +[2026-06-23 17:44:53,819][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.22266167402267456, 'train/grad_norm_pre_clip': 79.57888793945312, 'timing/data': 0.0005917618982493877, 'timing/model': 1.6916820439510047, 'learning_rate/qwen_vl_interface': 9.606911923336856e-06, 'learning_rate/action_model': 9.606911923336856e-05, 'epoch': 0.29}) +[2026-06-23 17:44:57,152][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.45499739050865173, 'train/grad_norm_pre_clip': 51.919677734375, 'timing/data': 0.0005852391477674246, 'timing/model': 1.6919959620572627, 'learning_rate/qwen_vl_interface': 9.605697962313817e-06, 'learning_rate/action_model': 9.605697962313816e-05, 'epoch': 0.29}) +[2026-06-23 17:45:00,481][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.2686999440193176, 'train/grad_norm_pre_clip': 27.789466857910156, 'timing/data': 0.0005949782207608223, 'timing/model': 1.690084557980299, 'learning_rate/qwen_vl_interface': 9.604482210827871e-06, 'learning_rate/action_model': 9.604482210827871e-05, 'epoch': 0.29}) +[2026-06-23 17:45:03,821][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.27281326055526733, 'train/grad_norm_pre_clip': 67.7083969116211, 'timing/data': 0.000601000152528286, 'timing/model': 1.6953937150537968, 'learning_rate/qwen_vl_interface': 9.603264669378772e-06, 'learning_rate/action_model': 9.603264669378772e-05, 'epoch': 0.29}) +[2026-06-23 17:45:07,148][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.4193251430988312, 'train/grad_norm_pre_clip': 92.85057067871094, 'timing/data': 0.0008310619741678238, 'timing/model': 1.689403654076159, 'learning_rate/qwen_vl_interface': 9.602045338467007e-06, 'learning_rate/action_model': 9.602045338467006e-05, 'epoch': 0.29}) +[2026-06-23 17:45:10,481][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.2121160924434662, 'train/grad_norm_pre_clip': 24.452320098876953, 'timing/data': 0.0005954490043222904, 'timing/model': 1.6940062229987234, 'learning_rate/qwen_vl_interface': 9.600824218593791e-06, 'learning_rate/action_model': 9.600824218593791e-05, 'epoch': 0.29}) +[2026-06-23 17:45:13,821][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.25078439712524414, 'train/grad_norm_pre_clip': 13.616620063781738, 'timing/data': 0.00058577093295753, 'timing/model': 1.6936641070060432, 'learning_rate/qwen_vl_interface': 9.599601310261087e-06, 'learning_rate/action_model': 9.599601310261085e-05, 'epoch': 0.29}) +[2026-06-23 17:45:17,160][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.25967463850975037, 'train/grad_norm_pre_clip': 10.1332426071167, 'timing/data': 0.0006182622164487839, 'timing/model': 1.6927565201185644, 'learning_rate/qwen_vl_interface': 9.598376613971582e-06, 'learning_rate/action_model': 9.598376613971582e-05, 'epoch': 0.29}) +[2026-06-23 17:45:20,503][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.3174426853656769, 'train/grad_norm_pre_clip': 15.317654609680176, 'timing/data': 0.0007350619416683912, 'timing/model': 1.6975150820799172, 'learning_rate/qwen_vl_interface': 9.597150130228703e-06, 'learning_rate/action_model': 9.597150130228703e-05, 'epoch': 0.29}) +[2026-06-23 17:45:23,832][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.2773726284503937, 'train/grad_norm_pre_clip': 44.753482818603516, 'timing/data': 0.0006384099833667278, 'timing/model': 1.6897559161297977, 'learning_rate/qwen_vl_interface': 9.595921859536612e-06, 'learning_rate/action_model': 9.595921859536611e-05, 'epoch': 0.29}) +[2026-06-23 17:45:27,256][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.18938766419887543, 'train/grad_norm_pre_clip': 18.84123992919922, 'timing/data': 0.0006516049616038799, 'timing/model': 1.7810687301680446, 'learning_rate/qwen_vl_interface': 9.594691802400203e-06, 'learning_rate/action_model': 9.594691802400202e-05, 'epoch': 0.29}) +[2026-06-23 17:45:46,996][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 750 +[2026-06-23 17:45:46,997][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.21248598396778107, 'train/grad_norm_pre_clip': 31.400794982910156, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 10.299999999999992, 'rl_games_eval/mid_train/mean_length': 101.4, 'rl_games_eval/mid_train/std_reward': 4.747630988187691, 'rl_games_eval/mid_train/std_length': 38.30195817448502, 'rl_games_eval/mid_train/macro_mean_reward': 10.299999999999992, 'rl_games_eval/mid_train/macro_mean_length': 101.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 10.299999999999992, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 101.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 4.747630988187691, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 38.30195817448502, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 10.299999999999992, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005956790409982204, 'timing/model': 1.690450657857582, 'learning_rate/qwen_vl_interface': 9.59345995932511e-06, 'learning_rate/action_model': 9.59345995932511e-05, 'epoch': 0.29}) +[2026-06-23 17:45:50,303][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.19467473030090332, 'train/grad_norm_pre_clip': 32.6825065612793, 'timing/data': 0.0006972700357437134, 'timing/model': 1.6758639609906822, 'learning_rate/qwen_vl_interface': 9.592226330817693e-06, 'learning_rate/action_model': 9.592226330817693e-05, 'epoch': 0.29}) +[2026-06-23 17:45:53,612][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.24763496220111847, 'train/grad_norm_pre_clip': 26.60786247253418, 'timing/data': 0.0005827648565173149, 'timing/model': 1.6848735131788999, 'learning_rate/qwen_vl_interface': 9.590990917385052e-06, 'learning_rate/action_model': 9.590990917385052e-05, 'epoch': 0.29}) +[2026-06-23 17:45:56,929][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.18580810725688934, 'train/grad_norm_pre_clip': 22.226125717163086, 'timing/data': 0.0005969421472400427, 'timing/model': 1.6811340230051428, 'learning_rate/qwen_vl_interface': 9.58975371953502e-06, 'learning_rate/action_model': 9.589753719535018e-05, 'epoch': 0.29}) +[2026-06-23 17:46:00,253][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.2352878600358963, 'train/grad_norm_pre_clip': 31.360488891601562, 'timing/data': 0.0005853299517184496, 'timing/model': 1.685154136037454, 'learning_rate/qwen_vl_interface': 9.588514737776158e-06, 'learning_rate/action_model': 9.588514737776157e-05, 'epoch': 0.29}) +[2026-06-23 17:46:03,586][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.2372240424156189, 'train/grad_norm_pre_clip': 16.84416961669922, 'timing/data': 0.0005855211056768894, 'timing/model': 1.697167705046013, 'learning_rate/qwen_vl_interface': 9.587273972617769e-06, 'learning_rate/action_model': 9.587273972617768e-05, 'epoch': 0.29}) +[2026-06-23 17:46:06,910][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.3070695400238037, 'train/grad_norm_pre_clip': 92.9539794921875, 'timing/data': 0.0005967419128865004, 'timing/model': 1.688538984861225, 'learning_rate/qwen_vl_interface': 9.586031424569882e-06, 'learning_rate/action_model': 9.58603142456988e-05, 'epoch': 0.29}) +[2026-06-23 17:46:10,234][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.2256583422422409, 'train/grad_norm_pre_clip': 20.641517639160156, 'timing/data': 0.0005949379410594702, 'timing/model': 1.685706279007718, 'learning_rate/qwen_vl_interface': 9.584787094143265e-06, 'learning_rate/action_model': 9.584787094143264e-05, 'epoch': 0.29}) +[2026-06-23 17:46:13,563][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.21391919255256653, 'train/grad_norm_pre_clip': 38.457122802734375, 'timing/data': 0.0006129820831120014, 'timing/model': 1.6893913140520453, 'learning_rate/qwen_vl_interface': 9.58354098184941e-06, 'learning_rate/action_model': 9.583540981849409e-05, 'epoch': 0.29}) +[2026-06-23 17:46:16,895][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.30069252848625183, 'train/grad_norm_pre_clip': 15.728763580322266, 'timing/data': 0.0007099050562828779, 'timing/model': 1.6897041599731892, 'learning_rate/qwen_vl_interface': 9.582293088200549e-06, 'learning_rate/action_model': 9.582293088200549e-05, 'epoch': 0.29}) +[2026-06-23 17:46:20,220][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.3169413208961487, 'train/grad_norm_pre_clip': 50.464385986328125, 'timing/data': 0.0005959898699074984, 'timing/model': 1.6860855270642787, 'learning_rate/qwen_vl_interface': 9.581043413709646e-06, 'learning_rate/action_model': 9.581043413709646e-05, 'epoch': 0.29}) +[2026-06-23 17:46:23,554][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.17640995979309082, 'train/grad_norm_pre_clip': 17.815309524536133, 'timing/data': 0.000700406963005662, 'timing/model': 1.6902106679044664, 'learning_rate/qwen_vl_interface': 9.579791958890393e-06, 'learning_rate/action_model': 9.579791958890394e-05, 'epoch': 0.29}) +[2026-06-23 17:46:26,890][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.266364723443985, 'train/grad_norm_pre_clip': 20.382436752319336, 'timing/data': 0.0006487600039690733, 'timing/model': 1.698258062126115, 'learning_rate/qwen_vl_interface': 9.578538724257218e-06, 'learning_rate/action_model': 9.578538724257216e-05, 'epoch': 0.29}) +[2026-06-23 17:46:30,526][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.23059618473052979, 'train/grad_norm_pre_clip': 48.14793395996094, 'timing/data': 0.0006415748503059149, 'timing/model': 1.9933926491066813, 'learning_rate/qwen_vl_interface': 9.577283710325274e-06, 'learning_rate/action_model': 9.577283710325274e-05, 'epoch': 0.3}) +[2026-06-23 17:46:33,853][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.21409456431865692, 'train/grad_norm_pre_clip': 11.979962348937988, 'timing/data': 0.0008119160775095224, 'timing/model': 1.6895121298730373, 'learning_rate/qwen_vl_interface': 9.576026917610455e-06, 'learning_rate/action_model': 9.576026917610455e-05, 'epoch': 0.3}) +[2026-06-23 17:46:37,187][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.2496582567691803, 'train/grad_norm_pre_clip': 54.17401123046875, 'timing/data': 0.0005838179495185614, 'timing/model': 1.6937473479192704, 'learning_rate/qwen_vl_interface': 9.574768346629376e-06, 'learning_rate/action_model': 9.574768346629376e-05, 'epoch': 0.3}) +[2026-06-23 17:46:40,523][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.1266295164823532, 'train/grad_norm_pre_clip': 51.95635223388672, 'timing/data': 0.0005962098948657513, 'timing/model': 1.6961512849666178, 'learning_rate/qwen_vl_interface': 9.573507997899391e-06, 'learning_rate/action_model': 9.573507997899392e-05, 'epoch': 0.3}) +[2026-06-23 17:46:43,849][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.2645239531993866, 'train/grad_norm_pre_clip': 45.196571350097656, 'timing/data': 0.0005903400015085936, 'timing/model': 1.6907900769729167, 'learning_rate/qwen_vl_interface': 9.572245871938582e-06, 'learning_rate/action_model': 9.572245871938582e-05, 'epoch': 0.3}) +[2026-06-23 17:46:47,173][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.28369876742362976, 'train/grad_norm_pre_clip': 51.04393768310547, 'timing/data': 0.0006824829615652561, 'timing/model': 1.6869723701383919, 'learning_rate/qwen_vl_interface': 9.570981969265761e-06, 'learning_rate/action_model': 9.570981969265761e-05, 'epoch': 0.3}) +[2026-06-23 17:46:50,505][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.40031298995018005, 'train/grad_norm_pre_clip': 126.82193756103516, 'timing/data': 0.0006606020033359528, 'timing/model': 1.692493875976652, 'learning_rate/qwen_vl_interface': 9.56971629040047e-06, 'learning_rate/action_model': 9.56971629040047e-05, 'epoch': 0.3}) +[2026-06-23 17:46:53,833][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.2897682785987854, 'train/grad_norm_pre_clip': 62.551055908203125, 'timing/data': 0.0006610320415347815, 'timing/model': 1.6892236380372196, 'learning_rate/qwen_vl_interface': 9.56844883586298e-06, 'learning_rate/action_model': 9.568448835862981e-05, 'epoch': 0.3}) +[2026-06-23 17:46:57,164][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.24408385157585144, 'train/grad_norm_pre_clip': 18.26831817626953, 'timing/data': 0.0007758880965411663, 'timing/model': 1.6880429279990494, 'learning_rate/qwen_vl_interface': 9.567179606174299e-06, 'learning_rate/action_model': 9.567179606174299e-05, 'epoch': 0.3}) +[2026-06-23 17:47:00,502][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.38935449719429016, 'train/grad_norm_pre_clip': 63.43185043334961, 'timing/data': 0.0008723700884729624, 'timing/model': 1.694296859903261, 'learning_rate/qwen_vl_interface': 9.565908601856155e-06, 'learning_rate/action_model': 9.565908601856154e-05, 'epoch': 0.3}) +[2026-06-23 17:47:03,837][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.20726576447486877, 'train/grad_norm_pre_clip': 51.786800384521484, 'timing/data': 0.0009134069550782442, 'timing/model': 1.6921435110270977, 'learning_rate/qwen_vl_interface': 9.56463582343101e-06, 'learning_rate/action_model': 9.56463582343101e-05, 'epoch': 0.3}) +[2026-06-23 17:47:07,171][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.1812690943479538, 'train/grad_norm_pre_clip': 34.95912551879883, 'timing/data': 0.0008402890525758266, 'timing/model': 1.69083619187586, 'learning_rate/qwen_vl_interface': 9.563361271422056e-06, 'learning_rate/action_model': 9.563361271422056e-05, 'epoch': 0.3}) +[2026-06-23 17:47:10,508][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.19953618943691254, 'train/grad_norm_pre_clip': 24.356908798217773, 'timing/data': 0.0006620439235121012, 'timing/model': 1.6938320610206574, 'learning_rate/qwen_vl_interface': 9.562084946353212e-06, 'learning_rate/action_model': 9.562084946353213e-05, 'epoch': 0.3}) +[2026-06-23 17:47:13,847][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.25443893671035767, 'train/grad_norm_pre_clip': 10.47225570678711, 'timing/data': 0.0006181220524013042, 'timing/model': 1.6944217379204929, 'learning_rate/qwen_vl_interface': 9.560806848749132e-06, 'learning_rate/action_model': 9.560806848749132e-05, 'epoch': 0.3}) +[2026-06-23 17:47:17,186][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.2743387520313263, 'train/grad_norm_pre_clip': 16.015785217285156, 'timing/data': 0.000599676975980401, 'timing/model': 1.6950639940332621, 'learning_rate/qwen_vl_interface': 9.559526979135186e-06, 'learning_rate/action_model': 9.559526979135187e-05, 'epoch': 0.3}) +[2026-06-23 17:47:20,615][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.20727430284023285, 'train/grad_norm_pre_clip': 8.434625625610352, 'timing/data': 0.0006140440236777067, 'timing/model': 1.693412919063121, 'learning_rate/qwen_vl_interface': 9.558245338037486e-06, 'learning_rate/action_model': 9.558245338037486e-05, 'epoch': 0.3}) +[2026-06-23 17:47:23,946][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.2706771194934845, 'train/grad_norm_pre_clip': 32.47169494628906, 'timing/data': 0.0006033740937709808, 'timing/model': 1.69465619395487, 'learning_rate/qwen_vl_interface': 9.556961925982864e-06, 'learning_rate/action_model': 9.556961925982863e-05, 'epoch': 0.3}) +[2026-06-23 17:47:27,279][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.31439781188964844, 'train/grad_norm_pre_clip': 39.90757369995117, 'timing/data': 0.0006128819659352303, 'timing/model': 1.6939055621623993, 'learning_rate/qwen_vl_interface': 9.555676743498882e-06, 'learning_rate/action_model': 9.555676743498881e-05, 'epoch': 0.3}) +[2026-06-23 17:47:30,617][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.170514315366745, 'train/grad_norm_pre_clip': 7.881543159484863, 'timing/data': 0.0006091049872338772, 'timing/model': 1.6952227060683072, 'learning_rate/qwen_vl_interface': 9.55438979111383e-06, 'learning_rate/action_model': 9.554389791113831e-05, 'epoch': 0.3}) +[2026-06-23 17:47:33,959][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.26209884881973267, 'train/grad_norm_pre_clip': 31.606260299682617, 'timing/data': 0.0005817939527332783, 'timing/model': 1.697485167067498, 'learning_rate/qwen_vl_interface': 9.553101069356728e-06, 'learning_rate/action_model': 9.553101069356727e-05, 'epoch': 0.3}) +[2026-06-23 17:47:37,295][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.13889168202877045, 'train/grad_norm_pre_clip': 27.391918182373047, 'timing/data': 0.000596621073782444, 'timing/model': 1.6934735251124948, 'learning_rate/qwen_vl_interface': 9.551810578757316e-06, 'learning_rate/action_model': 9.551810578757315e-05, 'epoch': 0.3}) +[2026-06-23 17:47:40,634][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.1946377456188202, 'train/grad_norm_pre_clip': 32.42021942138672, 'timing/data': 0.0005948080215603113, 'timing/model': 1.6960071851499379, 'learning_rate/qwen_vl_interface': 9.550518319846069e-06, 'learning_rate/action_model': 9.550518319846069e-05, 'epoch': 0.3}) +[2026-06-23 17:47:43,976][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.2289569228887558, 'train/grad_norm_pre_clip': 39.37372589111328, 'timing/data': 0.0005923928692936897, 'timing/model': 1.6953882509842515, 'learning_rate/qwen_vl_interface': 9.549224293154187e-06, 'learning_rate/action_model': 9.549224293154187e-05, 'epoch': 0.3}) +[2026-06-23 17:47:47,313][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.2522023022174835, 'train/grad_norm_pre_clip': 28.20582389831543, 'timing/data': 0.0006018811836838722, 'timing/model': 1.694411874981597, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.3}) +[2026-06-23 17:47:50,642][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.20266249775886536, 'train/grad_norm_pre_clip': 23.47051239013672, 'timing/data': 0.0005982839502394199, 'timing/model': 1.688194357091561, 'learning_rate/qwen_vl_interface': 9.54663093855694e-06, 'learning_rate/action_model': 9.54663093855694e-05, 'epoch': 0.3}) +[2026-06-23 17:47:53,961][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.20351730287075043, 'train/grad_norm_pre_clip': 55.679256439208984, 'timing/data': 0.0005904398858547211, 'timing/model': 1.6863139912020415, 'learning_rate/qwen_vl_interface': 9.545331611717607e-06, 'learning_rate/action_model': 9.545331611717608e-05, 'epoch': 0.3}) +[2026-06-23 17:47:57,304][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.24282144010066986, 'train/grad_norm_pre_clip': 59.77931213378906, 'timing/data': 0.0005927740130573511, 'timing/model': 1.7009636838920414, 'learning_rate/qwen_vl_interface': 9.544030519229697e-06, 'learning_rate/action_model': 9.544030519229697e-05, 'epoch': 0.31}) +[2026-06-23 17:48:00,639][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.31104686856269836, 'train/grad_norm_pre_clip': 12.365315437316895, 'timing/data': 0.0006088749505579472, 'timing/model': 1.6913509259466082, 'learning_rate/qwen_vl_interface': 9.54272766162804e-06, 'learning_rate/action_model': 9.54272766162804e-05, 'epoch': 0.31}) +[2026-06-23 17:48:03,977][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.18542014062404633, 'train/grad_norm_pre_clip': 22.73174285888672, 'timing/data': 0.0006009999196976423, 'timing/model': 1.6925622508861125, 'learning_rate/qwen_vl_interface': 9.54142303944819e-06, 'learning_rate/action_model': 9.541423039448191e-05, 'epoch': 0.31}) +[2026-06-23 17:48:07,397][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.27792519330978394, 'train/grad_norm_pre_clip': 70.0365219116211, 'timing/data': 0.0006020818836987019, 'timing/model': 1.780311814043671, 'learning_rate/qwen_vl_interface': 9.540116653226433e-06, 'learning_rate/action_model': 9.540116653226431e-05, 'epoch': 0.31}) +[2026-06-23 17:48:10,738][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.1722891926765442, 'train/grad_norm_pre_clip': 27.96068572998047, 'timing/data': 0.0006228298880159855, 'timing/model': 1.6972466830629855, 'learning_rate/qwen_vl_interface': 9.53880850349977e-06, 'learning_rate/action_model': 9.538808503499768e-05, 'epoch': 0.31}) +[2026-06-23 17:48:14,085][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.16857744753360748, 'train/grad_norm_pre_clip': 59.084068298339844, 'timing/data': 0.0006065699271857738, 'timing/model': 1.7008220800198615, 'learning_rate/qwen_vl_interface': 9.537498590805934e-06, 'learning_rate/action_model': 9.537498590805933e-05, 'epoch': 0.31}) +[2026-06-23 17:48:17,420][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.33278709650039673, 'train/grad_norm_pre_clip': 90.19962310791016, 'timing/data': 0.0006556231528520584, 'timing/model': 1.6941292909905314, 'learning_rate/qwen_vl_interface': 9.536186915683379e-06, 'learning_rate/action_model': 9.53618691568338e-05, 'epoch': 0.31}) +[2026-06-23 17:48:20,753][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.16993090510368347, 'train/grad_norm_pre_clip': 25.896739959716797, 'timing/data': 0.0006531369872391224, 'timing/model': 1.6907834608573467, 'learning_rate/qwen_vl_interface': 9.534873478671289e-06, 'learning_rate/action_model': 9.534873478671288e-05, 'epoch': 0.31}) +[2026-06-23 17:48:24,101][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.16179004311561584, 'train/grad_norm_pre_clip': 51.00761413574219, 'timing/data': 0.0005903199780732393, 'timing/model': 1.6983570139855146, 'learning_rate/qwen_vl_interface': 9.533558280309563e-06, 'learning_rate/action_model': 9.533558280309563e-05, 'epoch': 0.31}) +[2026-06-23 17:48:27,439][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.14855767786502838, 'train/grad_norm_pre_clip': 17.269752502441406, 'timing/data': 0.0006919198203831911, 'timing/model': 1.6956110820174217, 'learning_rate/qwen_vl_interface': 9.532241321138832e-06, 'learning_rate/action_model': 9.532241321138833e-05, 'epoch': 0.31}) +[2026-06-23 17:48:30,777][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.1962081491947174, 'train/grad_norm_pre_clip': 35.473934173583984, 'timing/data': 0.00065944017842412, 'timing/model': 1.6956935150083154, 'learning_rate/qwen_vl_interface': 9.530922601700451e-06, 'learning_rate/action_model': 9.530922601700452e-05, 'epoch': 0.31}) +[2026-06-23 17:48:34,109][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 800 +[2026-06-23 17:50:07,580][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 800 +[2026-06-23 17:50:07,581][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 800 +[2026-06-23 17:51:18,787][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 800 +[2026-06-23 17:51:18,787][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.16917981207370758, 'train/grad_norm_pre_clip': 78.57182312011719, 'eval/action_loss/seconds': 93.46883713686839, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.11687126755714417, 'eval/latency_2/loss': 0.11687126755714417, 'eval/flappy/loss': 0.11687126755714417, 'eval/flappy/latency_2/loss': 0.11687126755714417, 'eval/flap_precision': 0.8766278272789582, 'eval/flap_recall': 0.8630229419703104, 'eval/flap_f1': 0.8697721863311799, 'eval/flap_rate_model': 0.2026388888888889, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8697721863311799, 'eval/latency_2/flap_precision': 0.8766278272789582, 'eval/latency_2/flap_recall': 0.8630229419703104, 'eval/latency_2/flap_f1': 0.8697721863311799, 'eval/latency_2/flap_rate_model': 0.2026388888888889, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8697721863311799, 'eval/action_classification/data_wait_seconds': 0.9778464436531067, 'eval/action_classification/predict_seconds': 70.09473419189453, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.000628890935331583, 'timing/model': 1.6880193199031055, 'learning_rate/qwen_vl_interface': 9.52960212253649e-06, 'learning_rate/action_model': 9.529602122536491e-05, 'epoch': 0.31}) +[2026-06-23 17:51:22,210][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.2740599513053894, 'train/grad_norm_pre_clip': 30.336605072021484, 'timing/data': 0.0006027629133313894, 'timing/model': 1.6865797380451113, 'learning_rate/qwen_vl_interface': 9.528279884189754e-06, 'learning_rate/action_model': 9.528279884189754e-05, 'epoch': 0.31}) +[2026-06-23 17:51:25,539][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.28889742493629456, 'train/grad_norm_pre_clip': 26.02754020690918, 'timing/data': 0.0009303579572588205, 'timing/model': 1.6900976980105042, 'learning_rate/qwen_vl_interface': 9.526955887203762e-06, 'learning_rate/action_model': 9.526955887203761e-05, 'epoch': 0.31}) +[2026-06-23 17:51:28,866][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.1935095191001892, 'train/grad_norm_pre_clip': 27.563669204711914, 'timing/data': 0.000747766112908721, 'timing/model': 1.6873728160280734, 'learning_rate/qwen_vl_interface': 9.525630132122758e-06, 'learning_rate/action_model': 9.525630132122758e-05, 'epoch': 0.31}) +[2026-06-23 17:51:32,193][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.19169560074806213, 'train/grad_norm_pre_clip': 33.542415618896484, 'timing/data': 0.0005906498990952969, 'timing/model': 1.6895733107812703, 'learning_rate/qwen_vl_interface': 9.524302619491712e-06, 'learning_rate/action_model': 9.524302619491712e-05, 'epoch': 0.31}) +[2026-06-23 17:51:35,526][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.3106968402862549, 'train/grad_norm_pre_clip': 53.21586608886719, 'timing/data': 0.0005950380582362413, 'timing/model': 1.6926754810847342, 'learning_rate/qwen_vl_interface': 9.522973349856315e-06, 'learning_rate/action_model': 9.522973349856315e-05, 'epoch': 0.31}) +[2026-06-23 17:51:38,856][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.27976369857788086, 'train/grad_norm_pre_clip': 17.383602142333984, 'timing/data': 0.0006019920110702515, 'timing/model': 1.6925137960352004, 'learning_rate/qwen_vl_interface': 9.521642323762976e-06, 'learning_rate/action_model': 9.521642323762976e-05, 'epoch': 0.31}) +[2026-06-23 17:51:42,189][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.1526993364095688, 'train/grad_norm_pre_clip': 21.99629783630371, 'timing/data': 0.0005970718339085579, 'timing/model': 1.6933544951025397, 'learning_rate/qwen_vl_interface': 9.520309541758833e-06, 'learning_rate/action_model': 9.520309541758833e-05, 'epoch': 0.31}) +[2026-06-23 17:51:45,528][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.12184888124465942, 'train/grad_norm_pre_clip': 16.132761001586914, 'timing/data': 0.0005986848846077919, 'timing/model': 1.6981980169657618, 'learning_rate/qwen_vl_interface': 9.51897500439174e-06, 'learning_rate/action_model': 9.51897500439174e-05, 'epoch': 0.31}) +[2026-06-23 17:51:48,864][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.1808752864599228, 'train/grad_norm_pre_clip': 10.750253677368164, 'timing/data': 0.0006039650179445744, 'timing/model': 1.6980135198682547, 'learning_rate/qwen_vl_interface': 9.517638712210277e-06, 'learning_rate/action_model': 9.517638712210276e-05, 'epoch': 0.31}) +[2026-06-23 17:51:52,204][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.2200268805027008, 'train/grad_norm_pre_clip': 12.698219299316406, 'timing/data': 0.0006088241934776306, 'timing/model': 1.698951160069555, 'learning_rate/qwen_vl_interface': 9.51630066576374e-06, 'learning_rate/action_model': 9.51630066576374e-05, 'epoch': 0.31}) +[2026-06-23 17:51:55,540][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.32897770404815674, 'train/grad_norm_pre_clip': 110.88097381591797, 'timing/data': 0.0006057980936020613, 'timing/model': 1.694453045958653, 'learning_rate/qwen_vl_interface': 9.514960865602154e-06, 'learning_rate/action_model': 9.514960865602153e-05, 'epoch': 0.31}) +[2026-06-23 17:51:58,870][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.27388668060302734, 'train/grad_norm_pre_clip': 87.59092712402344, 'timing/data': 0.0006568250246345997, 'timing/model': 1.6905214660800993, 'learning_rate/qwen_vl_interface': 9.513619312276259e-06, 'learning_rate/action_model': 9.513619312276258e-05, 'epoch': 0.31}) +[2026-06-23 17:52:02,212][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.400144100189209, 'train/grad_norm_pre_clip': 101.0188980102539, 'timing/data': 0.0006341519765555859, 'timing/model': 1.6974065348040313, 'learning_rate/qwen_vl_interface': 9.512276006337514e-06, 'learning_rate/action_model': 9.512276006337513e-05, 'epoch': 0.31}) +[2026-06-23 17:52:05,549][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.28206777572631836, 'train/grad_norm_pre_clip': 48.674827575683594, 'timing/data': 0.0006854080129414797, 'timing/model': 1.695027682930231, 'learning_rate/qwen_vl_interface': 9.510930948338102e-06, 'learning_rate/action_model': 9.510930948338102e-05, 'epoch': 0.32}) +[2026-06-23 17:52:08,889][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.2598041296005249, 'train/grad_norm_pre_clip': 67.92772674560547, 'timing/data': 0.0006921009626239538, 'timing/model': 1.6974232748616487, 'learning_rate/qwen_vl_interface': 9.509584138830929e-06, 'learning_rate/action_model': 9.509584138830928e-05, 'epoch': 0.32}) +[2026-06-23 17:52:12,524][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.22082102298736572, 'train/grad_norm_pre_clip': 93.16625213623047, 'timing/data': 0.0006637871265411377, 'timing/model': 1.6929766559042037, 'learning_rate/qwen_vl_interface': 9.508235578369613e-06, 'learning_rate/action_model': 9.508235578369612e-05, 'epoch': 0.32}) +[2026-06-23 17:52:15,865][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.17703506350517273, 'train/grad_norm_pre_clip': 109.0671157836914, 'timing/data': 0.0006521160248667002, 'timing/model': 1.69836216699332, 'learning_rate/qwen_vl_interface': 9.506885267508498e-06, 'learning_rate/action_model': 9.506885267508498e-05, 'epoch': 0.32}) +[2026-06-23 17:52:19,201][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.21075838804244995, 'train/grad_norm_pre_clip': 97.23604583740234, 'timing/data': 0.000661343103274703, 'timing/model': 1.6954901788849384, 'learning_rate/qwen_vl_interface': 9.505533206802648e-06, 'learning_rate/action_model': 9.505533206802647e-05, 'epoch': 0.32}) +[2026-06-23 17:52:22,541][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.2880805730819702, 'train/grad_norm_pre_clip': 52.92049789428711, 'timing/data': 0.0006431480869650841, 'timing/model': 1.6959592690691352, 'learning_rate/qwen_vl_interface': 9.50417939680784e-06, 'learning_rate/action_model': 9.504179396807839e-05, 'epoch': 0.32}) +[2026-06-23 17:52:25,878][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.3066999912261963, 'train/grad_norm_pre_clip': 98.20279693603516, 'timing/data': 0.0005921828560531139, 'timing/model': 1.6950823499355465, 'learning_rate/qwen_vl_interface': 9.502823838080578e-06, 'learning_rate/action_model': 9.502823838080578e-05, 'epoch': 0.32}) +[2026-06-23 17:52:29,222][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.3437517583370209, 'train/grad_norm_pre_clip': 142.61715698242188, 'timing/data': 0.0005919227842241526, 'timing/model': 1.69880176987499, 'learning_rate/qwen_vl_interface': 9.501466531178079e-06, 'learning_rate/action_model': 9.501466531178079e-05, 'epoch': 0.32}) +[2026-06-23 17:52:32,567][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.42888113856315613, 'train/grad_norm_pre_clip': 176.70199584960938, 'timing/data': 0.0006492200773209333, 'timing/model': 1.7035793790128082, 'learning_rate/qwen_vl_interface': 9.50010747665828e-06, 'learning_rate/action_model': 9.500107476658281e-05, 'epoch': 0.32}) +[2026-06-23 17:52:35,916][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.24364128708839417, 'train/grad_norm_pre_clip': 129.5491943359375, 'timing/data': 0.0006013510283082724, 'timing/model': 1.7086901429574937, 'learning_rate/qwen_vl_interface': 9.498746675079842e-06, 'learning_rate/action_model': 9.49874667507984e-05, 'epoch': 0.32}) +[2026-06-23 17:52:39,252][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.20947323739528656, 'train/grad_norm_pre_clip': 45.535701751708984, 'timing/data': 0.0006104169879108667, 'timing/model': 1.6954000950790942, 'learning_rate/qwen_vl_interface': 9.49738412700213e-06, 'learning_rate/action_model': 9.49738412700213e-05, 'epoch': 0.32}) +[2026-06-23 17:52:42,584][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.28937941789627075, 'train/grad_norm_pre_clip': 44.47818374633789, 'timing/data': 0.0006060490850359201, 'timing/model': 1.6931512570008636, 'learning_rate/qwen_vl_interface': 9.496019832985243e-06, 'learning_rate/action_model': 9.496019832985243e-05, 'epoch': 0.32}) +[2026-06-23 17:52:45,922][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.2618197798728943, 'train/grad_norm_pre_clip': 77.74312591552734, 'timing/data': 0.0009543031919747591, 'timing/model': 1.6936272499151528, 'learning_rate/qwen_vl_interface': 9.49465379358999e-06, 'learning_rate/action_model': 9.49465379358999e-05, 'epoch': 0.32}) +[2026-06-23 17:52:49,258][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.1709917187690735, 'train/grad_norm_pre_clip': 67.81155395507812, 'timing/data': 0.0006173411384224892, 'timing/model': 1.6937719921115786, 'learning_rate/qwen_vl_interface': 9.493286009377896e-06, 'learning_rate/action_model': 9.493286009377896e-05, 'epoch': 0.32}) +[2026-06-23 17:52:52,596][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.19996386766433716, 'train/grad_norm_pre_clip': 20.009220123291016, 'timing/data': 0.0006732749752700329, 'timing/model': 1.6953791221603751, 'learning_rate/qwen_vl_interface': 9.491916480911208e-06, 'learning_rate/action_model': 9.491916480911207e-05, 'epoch': 0.32}) +[2026-06-23 17:52:55,940][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.19347435235977173, 'train/grad_norm_pre_clip': 33.2064208984375, 'timing/data': 0.000777160981670022, 'timing/model': 1.6988831469789147, 'learning_rate/qwen_vl_interface': 9.490545208752884e-06, 'learning_rate/action_model': 9.490545208752883e-05, 'epoch': 0.32}) +[2026-06-23 17:52:59,277][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.4478599429130554, 'train/grad_norm_pre_clip': 118.48204040527344, 'timing/data': 0.0006571749690920115, 'timing/model': 1.6916949769947678, 'learning_rate/qwen_vl_interface': 9.489172193466603e-06, 'learning_rate/action_model': 9.489172193466603e-05, 'epoch': 0.32}) +[2026-06-23 17:53:02,854][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.14339374005794525, 'train/grad_norm_pre_clip': 92.1922378540039, 'timing/data': 0.0005988148041069508, 'timing/model': 1.6968970349989831, 'learning_rate/qwen_vl_interface': 9.487797435616762e-06, 'learning_rate/action_model': 9.487797435616761e-05, 'epoch': 0.32}) +[2026-06-23 17:53:06,198][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.26355013251304626, 'train/grad_norm_pre_clip': 96.60707092285156, 'timing/data': 0.0009209208656102419, 'timing/model': 1.6950357230380177, 'learning_rate/qwen_vl_interface': 9.486420935768472e-06, 'learning_rate/action_model': 9.486420935768472e-05, 'epoch': 0.32}) +[2026-06-23 17:53:09,534][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.4310687780380249, 'train/grad_norm_pre_clip': 103.75700378417969, 'timing/data': 0.0005994760431349277, 'timing/model': 1.69059297000058, 'learning_rate/qwen_vl_interface': 9.485042694487558e-06, 'learning_rate/action_model': 9.485042694487559e-05, 'epoch': 0.32}) +[2026-06-23 17:53:12,872][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.24641062319278717, 'train/grad_norm_pre_clip': 20.33856201171875, 'timing/data': 0.0006956369616091251, 'timing/model': 1.694082081085071, 'learning_rate/qwen_vl_interface': 9.483662712340565e-06, 'learning_rate/action_model': 9.483662712340564e-05, 'epoch': 0.32}) +[2026-06-23 17:53:16,211][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.18001876771450043, 'train/grad_norm_pre_clip': 23.42308235168457, 'timing/data': 0.0005897269584238529, 'timing/model': 1.6960580579470843, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.32}) +[2026-06-23 17:53:19,559][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.18922735750675201, 'train/grad_norm_pre_clip': 75.45948791503906, 'timing/data': 0.0006541190668940544, 'timing/model': 1.697877231054008, 'learning_rate/qwen_vl_interface': 9.480897527718083e-06, 'learning_rate/action_model': 9.480897527718084e-05, 'epoch': 0.32}) +[2026-06-23 17:53:22,907][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.15813010931015015, 'train/grad_norm_pre_clip': 23.30299949645996, 'timing/data': 0.0006201351061463356, 'timing/model': 1.699112375965342, 'learning_rate/qwen_vl_interface': 9.479512326379263e-06, 'learning_rate/action_model': 9.479512326379264e-05, 'epoch': 0.32}) +[2026-06-23 17:53:26,254][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.18289409577846527, 'train/grad_norm_pre_clip': 18.93598747253418, 'timing/data': 0.0005946368910372257, 'timing/model': 1.6999572378117591, 'learning_rate/qwen_vl_interface': 9.478125386447686e-06, 'learning_rate/action_model': 9.478125386447686e-05, 'epoch': 0.32}) +[2026-06-23 17:53:29,603][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.2462170422077179, 'train/grad_norm_pre_clip': 7.890890121459961, 'timing/data': 0.0006073210388422012, 'timing/model': 1.6988055789843202, 'learning_rate/qwen_vl_interface': 9.476736708493472e-06, 'learning_rate/action_model': 9.476736708493472e-05, 'epoch': 0.32}) +[2026-06-23 17:53:32,940][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.27258679270744324, 'train/grad_norm_pre_clip': 55.61789321899414, 'timing/data': 0.0005914911162108183, 'timing/model': 1.6965570689644665, 'learning_rate/qwen_vl_interface': 9.475346293087457e-06, 'learning_rate/action_model': 9.475346293087458e-05, 'epoch': 0.33}) +[2026-06-23 17:53:36,277][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.2423301339149475, 'train/grad_norm_pre_clip': 48.80621337890625, 'timing/data': 0.000594025943428278, 'timing/model': 1.6965746260248125, 'learning_rate/qwen_vl_interface': 9.473954140801184e-06, 'learning_rate/action_model': 9.473954140801185e-05, 'epoch': 0.33}) +[2026-06-23 17:53:39,624][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.25638842582702637, 'train/grad_norm_pre_clip': 81.45389556884766, 'timing/data': 0.0005973519291728735, 'timing/model': 1.7016309211030602, 'learning_rate/qwen_vl_interface': 9.472560252206918e-06, 'learning_rate/action_model': 9.472560252206918e-05, 'epoch': 0.33}) +[2026-06-23 17:53:42,966][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.12772519886493683, 'train/grad_norm_pre_clip': 8.75564193725586, 'timing/data': 0.0006058989092707634, 'timing/model': 1.6982484699692577, 'learning_rate/qwen_vl_interface': 9.471164627877633e-06, 'learning_rate/action_model': 9.471164627877633e-05, 'epoch': 0.33}) +[2026-06-23 17:53:46,301][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.20892448723316193, 'train/grad_norm_pre_clip': 16.294281005859375, 'timing/data': 0.000614294083788991, 'timing/model': 1.6953684431500733, 'learning_rate/qwen_vl_interface': 9.469767268387016e-06, 'learning_rate/action_model': 9.469767268387016e-05, 'epoch': 0.33}) +[2026-06-23 17:53:49,799][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.36298924684524536, 'train/grad_norm_pre_clip': 61.316165924072266, 'timing/data': 0.0006075920537114143, 'timing/model': 1.6992535190656781, 'learning_rate/qwen_vl_interface': 9.468368174309468e-06, 'learning_rate/action_model': 9.468368174309469e-05, 'epoch': 0.33}) +[2026-06-23 17:53:53,137][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.17198430001735687, 'train/grad_norm_pre_clip': 62.76435470581055, 'timing/data': 0.0005811720620840788, 'timing/model': 1.6925049819983542, 'learning_rate/qwen_vl_interface': 9.466967346220107e-06, 'learning_rate/action_model': 9.466967346220107e-05, 'epoch': 0.33}) +[2026-06-23 17:53:56,480][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.2056492418050766, 'train/grad_norm_pre_clip': 18.332155227661133, 'timing/data': 0.000620625913143158, 'timing/model': 1.6959316148422658, 'learning_rate/qwen_vl_interface': 9.465564784694759e-06, 'learning_rate/action_model': 9.465564784694759e-05, 'epoch': 0.33}) +[2026-06-23 17:53:59,812][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.2403203845024109, 'train/grad_norm_pre_clip': 65.80179595947266, 'timing/data': 0.0005925830919295549, 'timing/model': 1.6920712469145656, 'learning_rate/qwen_vl_interface': 9.464160490309962e-06, 'learning_rate/action_model': 9.464160490309962e-05, 'epoch': 0.33}) +[2026-06-23 17:54:03,151][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.14817284047603607, 'train/grad_norm_pre_clip': 32.96056365966797, 'timing/data': 0.0006110891699790955, 'timing/model': 1.69453265494667, 'learning_rate/qwen_vl_interface': 9.462754463642972e-06, 'learning_rate/action_model': 9.46275446364297e-05, 'epoch': 0.33}) +[2026-06-23 17:54:06,493][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.15838585793972015, 'train/grad_norm_pre_clip': 16.33565330505371, 'timing/data': 0.0006078018341213465, 'timing/model': 1.696090767160058, 'learning_rate/qwen_vl_interface': 9.461346705271753e-06, 'learning_rate/action_model': 9.461346705271752e-05, 'epoch': 0.33}) +[2026-06-23 17:54:09,839][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.23227658867835999, 'train/grad_norm_pre_clip': 22.977096557617188, 'timing/data': 0.0005838470533490181, 'timing/model': 1.6973761760164052, 'learning_rate/qwen_vl_interface': 9.459937215774976e-06, 'learning_rate/action_model': 9.459937215774976e-05, 'epoch': 0.33}) +[2026-06-23 17:54:13,175][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.19950711727142334, 'train/grad_norm_pre_clip': 51.08951950073242, 'timing/data': 0.0005821138620376587, 'timing/model': 1.6967744398862123, 'learning_rate/qwen_vl_interface': 9.458525995732035e-06, 'learning_rate/action_model': 9.458525995732035e-05, 'epoch': 0.33}) +[2026-06-23 17:54:16,525][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.23511788249015808, 'train/grad_norm_pre_clip': 115.59920501708984, 'timing/data': 0.0009185150265693665, 'timing/model': 1.7029139299411327, 'learning_rate/qwen_vl_interface': 9.457113045723027e-06, 'learning_rate/action_model': 9.457113045723026e-05, 'epoch': 0.33}) +[2026-06-23 17:54:19,876][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.38203442096710205, 'train/grad_norm_pre_clip': 36.9187126159668, 'timing/data': 0.0006117289885878563, 'timing/model': 1.703783739125356, 'learning_rate/qwen_vl_interface': 9.455698366328762e-06, 'learning_rate/action_model': 9.455698366328762e-05, 'epoch': 0.33}) +[2026-06-23 17:54:23,222][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.3253776431083679, 'train/grad_norm_pre_clip': 16.395883560180664, 'timing/data': 0.0005955090746283531, 'timing/model': 1.69930764217861, 'learning_rate/qwen_vl_interface': 9.454281958130761e-06, 'learning_rate/action_model': 9.454281958130761e-05, 'epoch': 0.33}) +[2026-06-23 17:54:26,562][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.33908137679100037, 'train/grad_norm_pre_clip': 39.368526458740234, 'timing/data': 0.0006144349463284016, 'timing/model': 1.6942429200280458, 'learning_rate/qwen_vl_interface': 9.452863821711257e-06, 'learning_rate/action_model': 9.452863821711258e-05, 'epoch': 0.33}) +[2026-06-23 17:54:29,907][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.35345661640167236, 'train/grad_norm_pre_clip': 30.250730514526367, 'timing/data': 0.0005899679381400347, 'timing/model': 1.7001040659379214, 'learning_rate/qwen_vl_interface': 9.451443957653195e-06, 'learning_rate/action_model': 9.451443957653195e-05, 'epoch': 0.33}) +[2026-06-23 17:54:33,256][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.1993560791015625, 'train/grad_norm_pre_clip': 23.24360466003418, 'timing/data': 0.0005935849621891975, 'timing/model': 1.7018166079651564, 'learning_rate/qwen_vl_interface': 9.450022366540221e-06, 'learning_rate/action_model': 9.45002236654022e-05, 'epoch': 0.33}) +[2026-06-23 17:54:36,737][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.23102650046348572, 'train/grad_norm_pre_clip': 76.85818481445312, 'timing/data': 0.0006602010689675808, 'timing/model': 1.8282191019970924, 'learning_rate/qwen_vl_interface': 9.448599048956702e-06, 'learning_rate/action_model': 9.448599048956702e-05, 'epoch': 0.33}) +[2026-06-23 17:54:40,086][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.27707749605178833, 'train/grad_norm_pre_clip': 30.60680389404297, 'timing/data': 0.0006032730452716351, 'timing/model': 1.7065317160449922, 'learning_rate/qwen_vl_interface': 9.447174005487708e-06, 'learning_rate/action_model': 9.447174005487708e-05, 'epoch': 0.33}) +[2026-06-23 17:54:43,435][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.26899608969688416, 'train/grad_norm_pre_clip': 15.805970191955566, 'timing/data': 0.000600537983700633, 'timing/model': 1.6962959580123425, 'learning_rate/qwen_vl_interface': 9.445747236719022e-06, 'learning_rate/action_model': 9.445747236719022e-05, 'epoch': 0.33}) +[2026-06-23 17:54:46,784][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.4339488744735718, 'train/grad_norm_pre_clip': 172.4916229248047, 'timing/data': 0.0008332259021699429, 'timing/model': 1.6991383319254965, 'learning_rate/qwen_vl_interface': 9.444318743237135e-06, 'learning_rate/action_model': 9.444318743237134e-05, 'epoch': 0.33}) +[2026-06-23 17:54:50,136][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.5028581619262695, 'train/grad_norm_pre_clip': 176.78668212890625, 'timing/data': 0.0008240379393100739, 'timing/model': 1.7014358430169523, 'learning_rate/qwen_vl_interface': 9.442888525629244e-06, 'learning_rate/action_model': 9.442888525629245e-05, 'epoch': 0.33}) +[2026-06-23 17:54:53,480][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.28188085556030273, 'train/grad_norm_pre_clip': 149.00497436523438, 'timing/data': 0.0006237321067601442, 'timing/model': 1.7033028351143003, 'learning_rate/qwen_vl_interface': 9.441456584483261e-06, 'learning_rate/action_model': 9.441456584483261e-05, 'epoch': 0.33}) +[2026-06-23 17:54:56,828][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.350960373878479, 'train/grad_norm_pre_clip': 130.28253173828125, 'timing/data': 0.0005999181885272264, 'timing/model': 1.6984888538718224, 'learning_rate/qwen_vl_interface': 9.440022920387801e-06, 'learning_rate/action_model': 9.440022920387801e-05, 'epoch': 0.33}) +[2026-06-23 17:55:00,186][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.19637568295001984, 'train/grad_norm_pre_clip': 51.749725341796875, 'timing/data': 0.0006922709289938211, 'timing/model': 1.7078201910480857, 'learning_rate/qwen_vl_interface': 9.438587533932191e-06, 'learning_rate/action_model': 9.43858753393219e-05, 'epoch': 0.34}) +[2026-06-23 17:55:03,543][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.1587933897972107, 'train/grad_norm_pre_clip': 60.0893669128418, 'timing/data': 0.000664678867906332, 'timing/model': 1.7018761960789561, 'learning_rate/qwen_vl_interface': 9.43715042570646e-06, 'learning_rate/action_model': 9.43715042570646e-05, 'epoch': 0.34}) +[2026-06-23 17:55:06,880][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.3592855930328369, 'train/grad_norm_pre_clip': 75.20557403564453, 'timing/data': 0.0005979437846690416, 'timing/model': 1.6952860769815743, 'learning_rate/qwen_vl_interface': 9.435711596301353e-06, 'learning_rate/action_model': 9.435711596301352e-05, 'epoch': 0.34}) +[2026-06-23 17:55:10,225][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.33900147676467896, 'train/grad_norm_pre_clip': 74.93563842773438, 'timing/data': 0.0005987049080431461, 'timing/model': 1.700405941111967, 'learning_rate/qwen_vl_interface': 9.434271046308317e-06, 'learning_rate/action_model': 9.434271046308317e-05, 'epoch': 0.34}) +[2026-06-23 17:55:13,570][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.30677226185798645, 'train/grad_norm_pre_clip': 83.2889633178711, 'timing/data': 0.0006090740207582712, 'timing/model': 1.697642863029614, 'learning_rate/qwen_vl_interface': 9.432828776319507e-06, 'learning_rate/action_model': 9.432828776319505e-05, 'epoch': 0.34}) +[2026-06-23 17:55:16,919][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.16622956097126007, 'train/grad_norm_pre_clip': 44.83325958251953, 'timing/data': 0.0005917220842093229, 'timing/model': 1.6982988801319152, 'learning_rate/qwen_vl_interface': 9.431384786927785e-06, 'learning_rate/action_model': 9.431384786927786e-05, 'epoch': 0.34}) +[2026-06-23 17:55:20,261][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.23923148214817047, 'train/grad_norm_pre_clip': 37.06254959106445, 'timing/data': 0.0006057280115783215, 'timing/model': 1.6995211199391633, 'learning_rate/qwen_vl_interface': 9.429939078726723e-06, 'learning_rate/action_model': 9.429939078726722e-05, 'epoch': 0.34}) +[2026-06-23 17:55:23,604][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.3379865884780884, 'train/grad_norm_pre_clip': 27.854934692382812, 'timing/data': 0.0006076209247112274, 'timing/model': 1.7012273790314794, 'learning_rate/qwen_vl_interface': 9.428491652310598e-06, 'learning_rate/action_model': 9.428491652310597e-05, 'epoch': 0.34}) +[2026-06-23 17:55:27,227][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.23882536590099335, 'train/grad_norm_pre_clip': 54.853939056396484, 'timing/data': 0.0005938259419053793, 'timing/model': 1.9728270152118057, 'learning_rate/qwen_vl_interface': 9.427042508274386e-06, 'learning_rate/action_model': 9.427042508274386e-05, 'epoch': 0.34}) +[2026-06-23 17:55:30,569][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.37257498502731323, 'train/grad_norm_pre_clip': 54.100643157958984, 'timing/data': 0.0006870811339467764, 'timing/model': 1.6990489040035754, 'learning_rate/qwen_vl_interface': 9.425591647213783e-06, 'learning_rate/action_model': 9.425591647213782e-05, 'epoch': 0.34}) +[2026-06-23 17:55:33,911][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.20618145167827606, 'train/grad_norm_pre_clip': 10.1049222946167, 'timing/data': 0.0006029829382896423, 'timing/model': 1.697526958072558, 'learning_rate/qwen_vl_interface': 9.424139069725177e-06, 'learning_rate/action_model': 9.424139069725178e-05, 'epoch': 0.34}) +[2026-06-23 17:55:37,260][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.3826453387737274, 'train/grad_norm_pre_clip': 87.0844955444336, 'timing/data': 0.0006712910253554583, 'timing/model': 1.6994953900575638, 'learning_rate/qwen_vl_interface': 9.422684776405673e-06, 'learning_rate/action_model': 9.422684776405672e-05, 'epoch': 0.34}) +[2026-06-23 17:55:40,615][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.4445469379425049, 'train/grad_norm_pre_clip': 139.39581298828125, 'timing/data': 0.0006143348291516304, 'timing/model': 1.7041112559381872, 'learning_rate/qwen_vl_interface': 9.421228767853072e-06, 'learning_rate/action_model': 9.421228767853072e-05, 'epoch': 0.34}) +[2026-06-23 17:55:43,957][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.1867693066596985, 'train/grad_norm_pre_clip': 80.30593872070312, 'timing/data': 0.0005963610019534826, 'timing/model': 1.7013905479107052, 'learning_rate/qwen_vl_interface': 9.419771044665887e-06, 'learning_rate/action_model': 9.419771044665887e-05, 'epoch': 0.34}) +[2026-06-23 17:55:47,300][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.2579490840435028, 'train/grad_norm_pre_clip': 91.64398193359375, 'timing/data': 0.0006116789299994707, 'timing/model': 1.6953252640087157, 'learning_rate/qwen_vl_interface': 9.41831160744333e-06, 'learning_rate/action_model': 9.41831160744333e-05, 'epoch': 0.34}) +[2026-06-23 17:55:50,645][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.16374880075454712, 'train/grad_norm_pre_clip': 15.625699996948242, 'timing/data': 0.0005953290965408087, 'timing/model': 1.6966722288634628, 'learning_rate/qwen_vl_interface': 9.416850456785322e-06, 'learning_rate/action_model': 9.416850456785321e-05, 'epoch': 0.34}) +[2026-06-23 17:55:53,992][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.18595995008945465, 'train/grad_norm_pre_clip': 61.22185134887695, 'timing/data': 0.0005985850002616644, 'timing/model': 1.6987684110645205, 'learning_rate/qwen_vl_interface': 9.415387593292489e-06, 'learning_rate/action_model': 9.415387593292488e-05, 'epoch': 0.34}) +[2026-06-23 17:55:57,342][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.29618942737579346, 'train/grad_norm_pre_clip': 116.04952239990234, 'timing/data': 0.0005943069700151682, 'timing/model': 1.702615783084184, 'learning_rate/qwen_vl_interface': 9.413923017566155e-06, 'learning_rate/action_model': 9.413923017566155e-05, 'epoch': 0.34}) +[2026-06-23 17:56:00,688][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.4194701910018921, 'train/grad_norm_pre_clip': 109.26509857177734, 'timing/data': 0.0005840680096298456, 'timing/model': 1.6980559087824076, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.34}) +[2026-06-23 17:56:04,033][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.3557177484035492, 'train/grad_norm_pre_clip': 99.84632110595703, 'timing/data': 0.0006739259697496891, 'timing/model': 1.6968324440531433, 'learning_rate/qwen_vl_interface': 9.41098873182182e-06, 'learning_rate/action_model': 9.41098873182182e-05, 'epoch': 0.34}) +[2026-06-23 17:56:07,373][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.28310659527778625, 'train/grad_norm_pre_clip': 40.130558013916016, 'timing/data': 0.0005978038534522057, 'timing/model': 1.6964452499523759, 'learning_rate/qwen_vl_interface': 9.409519023009993e-06, 'learning_rate/action_model': 9.409519023009993e-05, 'epoch': 0.34}) +[2026-06-23 17:56:10,724][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.16086143255233765, 'train/grad_norm_pre_clip': 10.957207679748535, 'timing/data': 0.0006487080827355385, 'timing/model': 1.702871847897768, 'learning_rate/qwen_vl_interface': 9.408047604377014e-06, 'learning_rate/action_model': 9.408047604377014e-05, 'epoch': 0.34}) +[2026-06-23 17:56:14,165][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.2898566424846649, 'train/grad_norm_pre_clip': 62.53397750854492, 'timing/data': 0.0006527760997414589, 'timing/model': 1.7876976570114493, 'learning_rate/qwen_vl_interface': 9.40657447652773e-06, 'learning_rate/action_model': 9.406574476527731e-05, 'epoch': 0.34}) +[2026-06-23 17:56:17,521][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.2198011875152588, 'train/grad_norm_pre_clip': 82.31111145019531, 'timing/data': 0.0006145050283521414, 'timing/model': 1.7057175119407475, 'learning_rate/qwen_vl_interface': 9.405099640067685e-06, 'learning_rate/action_model': 9.405099640067685e-05, 'epoch': 0.34}) +[2026-06-23 17:56:20,867][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.34935858845710754, 'train/grad_norm_pre_clip': 78.8346939086914, 'timing/data': 0.0006156470626592636, 'timing/model': 1.6980442381463945, 'learning_rate/qwen_vl_interface': 9.40362309560313e-06, 'learning_rate/action_model': 9.40362309560313e-05, 'epoch': 0.34}) +[2026-06-23 17:56:24,214][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.1817222237586975, 'train/grad_norm_pre_clip': 32.83877182006836, 'timing/data': 0.0005913709755986929, 'timing/model': 1.7015936088282615, 'learning_rate/qwen_vl_interface': 9.402144843741016e-06, 'learning_rate/action_model': 9.402144843741017e-05, 'epoch': 0.34}) +[2026-06-23 17:56:27,562][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.1897251456975937, 'train/grad_norm_pre_clip': 23.293350219726562, 'timing/data': 0.000903337961062789, 'timing/model': 1.694823302095756, 'learning_rate/qwen_vl_interface': 9.400664885088999e-06, 'learning_rate/action_model': 9.400664885088999e-05, 'epoch': 0.35}) +[2026-06-23 17:56:30,911][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.3381686508655548, 'train/grad_norm_pre_clip': 61.31461715698242, 'timing/data': 0.0008529729675501585, 'timing/model': 1.6987244358751923, 'learning_rate/qwen_vl_interface': 9.39918322025543e-06, 'learning_rate/action_model': 9.39918322025543e-05, 'epoch': 0.35}) +[2026-06-23 17:56:34,248][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.24520243704319, 'train/grad_norm_pre_clip': 46.44223403930664, 'timing/data': 0.0008207920473068953, 'timing/model': 1.692043726099655, 'learning_rate/qwen_vl_interface': 9.397699849849368e-06, 'learning_rate/action_model': 9.397699849849367e-05, 'epoch': 0.35}) +[2026-06-23 17:56:37,596][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.3010540008544922, 'train/grad_norm_pre_clip': 24.46830940246582, 'timing/data': 0.000837232917547226, 'timing/model': 1.7055166650097817, 'learning_rate/qwen_vl_interface': 9.39621477448057e-06, 'learning_rate/action_model': 9.39621477448057e-05, 'epoch': 0.35}) +[2026-06-23 17:56:40,948][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.2860971987247467, 'train/grad_norm_pre_clip': 17.76014518737793, 'timing/data': 0.0008348789997398853, 'timing/model': 1.7005969230085611, 'learning_rate/qwen_vl_interface': 9.394727994759495e-06, 'learning_rate/action_model': 9.394727994759495e-05, 'epoch': 0.35}) +[2026-06-23 17:56:44,299][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.24706386029720306, 'train/grad_norm_pre_clip': 59.627811431884766, 'timing/data': 0.0005982141010463238, 'timing/model': 1.7013627048581839, 'learning_rate/qwen_vl_interface': 9.3932395112973e-06, 'learning_rate/action_model': 9.3932395112973e-05, 'epoch': 0.35}) +[2026-06-23 17:56:47,645][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.21654807031154633, 'train/grad_norm_pre_clip': 15.662550926208496, 'timing/data': 0.0005987449549138546, 'timing/model': 1.7000662519130856, 'learning_rate/qwen_vl_interface': 9.391749324705846e-06, 'learning_rate/action_model': 9.391749324705846e-05, 'epoch': 0.35}) +[2026-06-23 17:56:50,989][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.2212940752506256, 'train/grad_norm_pre_clip': 28.450023651123047, 'timing/data': 0.0006064900662750006, 'timing/model': 1.6989757160190493, 'learning_rate/qwen_vl_interface': 9.390257435597692e-06, 'learning_rate/action_model': 9.390257435597692e-05, 'epoch': 0.35}) +[2026-06-23 17:56:54,331][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 900 +[2026-06-23 17:58:28,001][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 900 +[2026-06-23 17:58:28,001][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 900 +[2026-06-23 17:59:39,473][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 900 +[2026-06-23 17:59:39,474][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.2129509449005127, 'train/grad_norm_pre_clip': 52.01813888549805, 'eval/action_loss/seconds': 93.66741435485892, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.13309583067893982, 'eval/latency_2/loss': 0.13309583067893982, 'eval/flappy/loss': 0.13309583067893982, 'eval/flappy/latency_2/loss': 0.13309583067893982, 'eval/flap_precision': 0.9426910299003323, 'eval/flap_recall': 0.7658569500674763, 'eval/flap_f1': 0.8451228592702904, 'eval/flap_rate_model': 0.16722222222222222, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8451228592702904, 'eval/latency_2/flap_precision': 0.9426910299003323, 'eval/latency_2/flap_recall': 0.7658569500674763, 'eval/latency_2/flap_f1': 0.8451228592702904, 'eval/latency_2/flap_rate_model': 0.16722222222222222, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8451228592702904, 'eval/action_classification/data_wait_seconds': 1.0662951469421387, 'eval/action_classification/predict_seconds': 70.28166961669922, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005931048654019833, 'timing/model': 1.6977702141739428, 'learning_rate/qwen_vl_interface': 9.3887638445861e-06, 'learning_rate/action_model': 9.3887638445861e-05, 'epoch': 0.35}) +[2026-06-23 17:59:42,786][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.13699708878993988, 'train/grad_norm_pre_clip': 78.08470916748047, 'timing/data': 0.0006094749551266432, 'timing/model': 1.6813316368497908, 'learning_rate/qwen_vl_interface': 9.387268552285024e-06, 'learning_rate/action_model': 9.387268552285025e-05, 'epoch': 0.35}) +[2026-06-23 17:59:46,092][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.21992728114128113, 'train/grad_norm_pre_clip': 59.38677215576172, 'timing/data': 0.0006036539562046528, 'timing/model': 1.6820552179124206, 'learning_rate/qwen_vl_interface': 9.385771559309125e-06, 'learning_rate/action_model': 9.385771559309125e-05, 'epoch': 0.35}) +[2026-06-23 17:59:49,401][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.34697723388671875, 'train/grad_norm_pre_clip': 35.50373840332031, 'timing/data': 0.000595219200477004, 'timing/model': 1.6839306140318513, 'learning_rate/qwen_vl_interface': 9.38427286627376e-06, 'learning_rate/action_model': 9.38427286627376e-05, 'epoch': 0.35}) +[2026-06-23 17:59:52,723][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.24079489707946777, 'train/grad_norm_pre_clip': 34.06501007080078, 'timing/data': 0.0007182098925113678, 'timing/model': 1.6901774851139635, 'learning_rate/qwen_vl_interface': 9.382772473794985e-06, 'learning_rate/action_model': 9.382772473794985e-05, 'epoch': 0.35}) +[2026-06-23 17:59:56,044][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.2461767941713333, 'train/grad_norm_pre_clip': 63.7710075378418, 'timing/data': 0.0006098160520195961, 'timing/model': 1.6835021830629557, 'learning_rate/qwen_vl_interface': 9.381270382489554e-06, 'learning_rate/action_model': 9.381270382489554e-05, 'epoch': 0.35}) +[2026-06-23 17:59:59,377][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.27207478880882263, 'train/grad_norm_pre_clip': 71.59082794189453, 'timing/data': 0.000585279893130064, 'timing/model': 1.6928866289090365, 'learning_rate/qwen_vl_interface': 9.37976659297492e-06, 'learning_rate/action_model': 9.379766592974921e-05, 'epoch': 0.35}) +[2026-06-23 18:00:02,707][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.20040307939052582, 'train/grad_norm_pre_clip': 36.79790496826172, 'timing/data': 0.0009486430790275335, 'timing/model': 1.6860529510304332, 'learning_rate/qwen_vl_interface': 9.378261105869237e-06, 'learning_rate/action_model': 9.378261105869237e-05, 'epoch': 0.35}) +[2026-06-23 18:00:06,045][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.17447645962238312, 'train/grad_norm_pre_clip': 11.251888275146484, 'timing/data': 0.0007713991217315197, 'timing/model': 1.6961940280161798, 'learning_rate/qwen_vl_interface': 9.376753921791348e-06, 'learning_rate/action_model': 9.376753921791348e-05, 'epoch': 0.35}) +[2026-06-23 18:00:09,375][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.12466634064912796, 'train/grad_norm_pre_clip': 28.633081436157227, 'timing/data': 0.000610036076977849, 'timing/model': 1.6932572030927986, 'learning_rate/qwen_vl_interface': 9.375245041360805e-06, 'learning_rate/action_model': 9.375245041360804e-05, 'epoch': 0.35}) +[2026-06-23 18:00:12,735][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.21658408641815186, 'train/grad_norm_pre_clip': 44.37571334838867, 'timing/data': 0.0006097261793911457, 'timing/model': 1.6947564519941807, 'learning_rate/qwen_vl_interface': 9.37373446519785e-06, 'learning_rate/action_model': 9.373734465197849e-05, 'epoch': 0.35}) +[2026-06-23 18:00:16,168][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.26624271273612976, 'train/grad_norm_pre_clip': 10.5484619140625, 'timing/data': 0.0007007571402937174, 'timing/model': 1.7827665628865361, 'learning_rate/qwen_vl_interface': 9.372222193923421e-06, 'learning_rate/action_model': 9.372222193923421e-05, 'epoch': 0.35}) +[2026-06-23 18:00:19,498][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.38808465003967285, 'train/grad_norm_pre_clip': 22.92547607421875, 'timing/data': 0.0008145601022988558, 'timing/model': 1.69102828996256, 'learning_rate/qwen_vl_interface': 9.370708228159158e-06, 'learning_rate/action_model': 9.370708228159159e-05, 'epoch': 0.35}) +[2026-06-23 18:00:22,832][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.2106907218694687, 'train/grad_norm_pre_clip': 27.066286087036133, 'timing/data': 0.0006658008787781, 'timing/model': 1.6907994030043483, 'learning_rate/qwen_vl_interface': 9.369192568527398e-06, 'learning_rate/action_model': 9.369192568527398e-05, 'epoch': 0.35}) +[2026-06-23 18:00:26,157][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.18392646312713623, 'train/grad_norm_pre_clip': 25.26030921936035, 'timing/data': 0.0005838780198246241, 'timing/model': 1.6896886369213462, 'learning_rate/qwen_vl_interface': 9.367675215651168e-06, 'learning_rate/action_model': 9.367675215651168e-05, 'epoch': 0.35}) +[2026-06-23 18:00:29,483][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.18643702566623688, 'train/grad_norm_pre_clip': 26.814416885375977, 'timing/data': 0.000649159075692296, 'timing/model': 1.6873658739496022, 'learning_rate/qwen_vl_interface': 9.366156170154194e-06, 'learning_rate/action_model': 9.366156170154194e-05, 'epoch': 0.35}) +[2026-06-23 18:00:32,800][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.1856376826763153, 'train/grad_norm_pre_clip': 37.58681869506836, 'timing/data': 0.0006597898900508881, 'timing/model': 1.6864272030070424, 'learning_rate/qwen_vl_interface': 9.3646354326609e-06, 'learning_rate/action_model': 9.364635432660899e-05, 'epoch': 0.35}) +[2026-06-23 18:00:36,134][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.19230395555496216, 'train/grad_norm_pre_clip': 36.470767974853516, 'timing/data': 0.0007589159067720175, 'timing/model': 1.6932284040376544, 'learning_rate/qwen_vl_interface': 9.3631130037964e-06, 'learning_rate/action_model': 9.363113003796401e-05, 'epoch': 0.35}) +[2026-06-23 18:00:39,471][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.11026913672685623, 'train/grad_norm_pre_clip': 20.947582244873047, 'timing/data': 0.0006714209448546171, 'timing/model': 1.6944937340449542, 'learning_rate/qwen_vl_interface': 9.361588884186513e-06, 'learning_rate/action_model': 9.361588884186513e-05, 'epoch': 0.36}) +[2026-06-23 18:00:42,807][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.21491974592208862, 'train/grad_norm_pre_clip': 16.86621856689453, 'timing/data': 0.0005973130464553833, 'timing/model': 1.694113200996071, 'learning_rate/qwen_vl_interface': 9.360063074457744e-06, 'learning_rate/action_model': 9.360063074457743e-05, 'epoch': 0.36}) +[2026-06-23 18:00:46,141][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.2647094428539276, 'train/grad_norm_pre_clip': 22.061120986938477, 'timing/data': 0.0008377840276807547, 'timing/model': 1.6903387198690325, 'learning_rate/qwen_vl_interface': 9.358535575237296e-06, 'learning_rate/action_model': 9.358535575237295e-05, 'epoch': 0.36}) +[2026-06-23 18:00:49,484][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.2509576678276062, 'train/grad_norm_pre_clip': 12.538508415222168, 'timing/data': 0.0006008988711982965, 'timing/model': 1.6985274988692254, 'learning_rate/qwen_vl_interface': 9.357006387153065e-06, 'learning_rate/action_model': 9.357006387153065e-05, 'epoch': 0.36}) +[2026-06-23 18:00:52,816][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.29398685693740845, 'train/grad_norm_pre_clip': 76.31513977050781, 'timing/data': 0.000685347942635417, 'timing/model': 1.6883477831725031, 'learning_rate/qwen_vl_interface': 9.355475510833643e-06, 'learning_rate/action_model': 9.355475510833643e-05, 'epoch': 0.36}) +[2026-06-23 18:00:56,151][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.2920229136943817, 'train/grad_norm_pre_clip': 39.87266540527344, 'timing/data': 0.0005852498579770327, 'timing/model': 1.6903777089901268, 'learning_rate/qwen_vl_interface': 9.353942946908318e-06, 'learning_rate/action_model': 9.353942946908317e-05, 'epoch': 0.36}) +[2026-06-23 18:00:59,477][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.14014829695224762, 'train/grad_norm_pre_clip': 27.413957595825195, 'timing/data': 0.0006479979492723942, 'timing/model': 1.6844556229189038, 'learning_rate/qwen_vl_interface': 9.352408696007065e-06, 'learning_rate/action_model': 9.352408696007065e-05, 'epoch': 0.36}) +[2026-06-23 18:01:02,798][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.2957407236099243, 'train/grad_norm_pre_clip': 21.107744216918945, 'timing/data': 0.0005857411306351423, 'timing/model': 1.6864995849318802, 'learning_rate/qwen_vl_interface': 9.350872758760558e-06, 'learning_rate/action_model': 9.350872758760559e-05, 'epoch': 0.36}) +[2026-06-23 18:01:06,227][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.25008559226989746, 'train/grad_norm_pre_clip': 37.27618408203125, 'timing/data': 0.0005971419159322977, 'timing/model': 1.6922055399045348, 'learning_rate/qwen_vl_interface': 9.349335135800164e-06, 'learning_rate/action_model': 9.349335135800163e-05, 'epoch': 0.36}) +[2026-06-23 18:01:09,561][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.0943930372595787, 'train/grad_norm_pre_clip': 14.191441535949707, 'timing/data': 0.000602812971919775, 'timing/model': 1.6961676271166652, 'learning_rate/qwen_vl_interface': 9.347795827757942e-06, 'learning_rate/action_model': 9.347795827757941e-05, 'epoch': 0.36}) +[2026-06-23 18:01:12,903][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.27042320370674133, 'train/grad_norm_pre_clip': 45.5870246887207, 'timing/data': 0.0005835359916090965, 'timing/model': 1.6975285860244185, 'learning_rate/qwen_vl_interface': 9.346254835266643e-06, 'learning_rate/action_model': 9.346254835266642e-05, 'epoch': 0.36}) +[2026-06-23 18:01:16,244][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.1216653436422348, 'train/grad_norm_pre_clip': 24.62147331237793, 'timing/data': 0.0006060679443180561, 'timing/model': 1.7001284130383283, 'learning_rate/qwen_vl_interface': 9.344712158959709e-06, 'learning_rate/action_model': 9.344712158959709e-05, 'epoch': 0.36}) +[2026-06-23 18:01:19,582][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.18454687297344208, 'train/grad_norm_pre_clip': 31.73365592956543, 'timing/data': 0.0006042160093784332, 'timing/model': 1.6965021609794348, 'learning_rate/qwen_vl_interface': 9.343167799471278e-06, 'learning_rate/action_model': 9.343167799471277e-05, 'epoch': 0.36}) +[2026-06-23 18:01:22,923][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.3690546154975891, 'train/grad_norm_pre_clip': 99.32965087890625, 'timing/data': 0.0006549209356307983, 'timing/model': 1.696969199925661, 'learning_rate/qwen_vl_interface': 9.341621757436177e-06, 'learning_rate/action_model': 9.341621757436177e-05, 'epoch': 0.36}) +[2026-06-23 18:01:26,264][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.21313606202602386, 'train/grad_norm_pre_clip': 18.468605041503906, 'timing/data': 0.0010119720827788115, 'timing/model': 1.6984418290667236, 'learning_rate/qwen_vl_interface': 9.340074033489924e-06, 'learning_rate/action_model': 9.340074033489923e-05, 'epoch': 0.36}) +[2026-06-23 18:01:29,597][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.1373807042837143, 'train/grad_norm_pre_clip': 75.82113647460938, 'timing/data': 0.0006048770155757666, 'timing/model': 1.6958433990366757, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.36}) +[2026-06-23 18:01:32,922][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.11772475391626358, 'train/grad_norm_pre_clip': 41.40081024169922, 'timing/data': 0.0006027829367667437, 'timing/model': 1.6826265540439636, 'learning_rate/qwen_vl_interface': 9.336973542409505e-06, 'learning_rate/action_model': 9.336973542409504e-05, 'epoch': 0.36}) +[2026-06-23 18:01:36,257][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.2167511135339737, 'train/grad_norm_pre_clip': 35.249881744384766, 'timing/data': 0.0006548808887600899, 'timing/model': 1.6997508250642568, 'learning_rate/qwen_vl_interface': 9.33542077654983e-06, 'learning_rate/action_model': 9.33542077654983e-05, 'epoch': 0.36}) +[2026-06-23 18:01:39,596][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.25716763734817505, 'train/grad_norm_pre_clip': 32.1138801574707, 'timing/data': 0.0006647489499300718, 'timing/model': 1.6985226389952004, 'learning_rate/qwen_vl_interface': 9.333866331327995e-06, 'learning_rate/action_model': 9.333866331327995e-05, 'epoch': 0.36}) +[2026-06-23 18:01:42,930][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.1693609207868576, 'train/grad_norm_pre_clip': 97.07989501953125, 'timing/data': 0.0006693180184811354, 'timing/model': 1.6913480679504573, 'learning_rate/qwen_vl_interface': 9.332310207382972e-06, 'learning_rate/action_model': 9.332310207382972e-05, 'epoch': 0.36}) +[2026-06-23 18:01:46,266][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.5067924857139587, 'train/grad_norm_pre_clip': 110.78267669677734, 'timing/data': 0.0005992560181766748, 'timing/model': 1.6926010940223932, 'learning_rate/qwen_vl_interface': 9.330752405354426e-06, 'learning_rate/action_model': 9.330752405354426e-05, 'epoch': 0.36}) +[2026-06-23 18:01:49,608][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.2065446674823761, 'train/grad_norm_pre_clip': 67.51349639892578, 'timing/data': 0.0008549571502953768, 'timing/model': 1.6968673029914498, 'learning_rate/qwen_vl_interface': 9.32919292588271e-06, 'learning_rate/action_model': 9.32919292588271e-05, 'epoch': 0.36}) +[2026-06-23 18:01:52,945][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.4061198830604553, 'train/grad_norm_pre_clip': 108.99617767333984, 'timing/data': 0.0006547798402607441, 'timing/model': 1.6967624670360237, 'learning_rate/qwen_vl_interface': 9.327631769608864e-06, 'learning_rate/action_model': 9.327631769608864e-05, 'epoch': 0.36}) +[2026-06-23 18:01:56,419][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.2877708077430725, 'train/grad_norm_pre_clip': 44.394447326660156, 'timing/data': 0.0006652600131928921, 'timing/model': 1.6891842689365149, 'learning_rate/qwen_vl_interface': 9.326068937174627e-06, 'learning_rate/action_model': 9.326068937174626e-05, 'epoch': 0.36}) +[2026-06-23 18:01:59,765][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.32356441020965576, 'train/grad_norm_pre_clip': 98.19560241699219, 'timing/data': 0.0006304951384663582, 'timing/model': 1.697946626925841, 'learning_rate/qwen_vl_interface': 9.324504429222412e-06, 'learning_rate/action_model': 9.324504429222412e-05, 'epoch': 0.36}) +[2026-06-23 18:02:03,104][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.35643649101257324, 'train/grad_norm_pre_clip': 148.33883666992188, 'timing/data': 0.0005924229044467211, 'timing/model': 1.696218178840354, 'learning_rate/qwen_vl_interface': 9.322938246395336e-06, 'learning_rate/action_model': 9.322938246395337e-05, 'epoch': 0.36}) +[2026-06-23 18:02:06,448][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.3096889555454254, 'train/grad_norm_pre_clip': 155.4764404296875, 'timing/data': 0.000700606033205986, 'timing/model': 1.6982371751219034, 'learning_rate/qwen_vl_interface': 9.321370389337196e-06, 'learning_rate/action_model': 9.321370389337196e-05, 'epoch': 0.37}) +[2026-06-23 18:02:09,795][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.42503276467323303, 'train/grad_norm_pre_clip': 107.12483215332031, 'timing/data': 0.0006895558908581734, 'timing/model': 1.700079069007188, 'learning_rate/qwen_vl_interface': 9.319800858692478e-06, 'learning_rate/action_model': 9.319800858692476e-05, 'epoch': 0.37}) +[2026-06-23 18:02:13,142][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.22909919917583466, 'train/grad_norm_pre_clip': 74.7346420288086, 'timing/data': 0.0006944860797375441, 'timing/model': 1.700384994968772, 'learning_rate/qwen_vl_interface': 9.318229655106357e-06, 'learning_rate/action_model': 9.318229655106356e-05, 'epoch': 0.37}) +[2026-06-23 18:02:16,483][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.20290005207061768, 'train/grad_norm_pre_clip': 18.970256805419922, 'timing/data': 0.0006807888858020306, 'timing/model': 1.7003204349894077, 'learning_rate/qwen_vl_interface': 9.316656779224693e-06, 'learning_rate/action_model': 9.316656779224693e-05, 'epoch': 0.37}) +[2026-06-23 18:02:19,822][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.21050262451171875, 'train/grad_norm_pre_clip': 80.06480407714844, 'timing/data': 0.0006370770279318094, 'timing/model': 1.693391637178138, 'learning_rate/qwen_vl_interface': 9.31508223169404e-06, 'learning_rate/action_model': 9.31508223169404e-05, 'epoch': 0.37}) +[2026-06-23 18:02:23,169][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.23709408938884735, 'train/grad_norm_pre_clip': 98.83761596679688, 'timing/data': 0.000607210909947753, 'timing/model': 1.697398868156597, 'learning_rate/qwen_vl_interface': 9.313506013161633e-06, 'learning_rate/action_model': 9.313506013161633e-05, 'epoch': 0.37}) +[2026-06-23 18:02:26,511][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.2635282278060913, 'train/grad_norm_pre_clip': 112.77143859863281, 'timing/data': 0.000669518020004034, 'timing/model': 1.696802909951657, 'learning_rate/qwen_vl_interface': 9.311928124275396e-06, 'learning_rate/action_model': 9.311928124275395e-05, 'epoch': 0.37}) +[2026-06-23 18:02:29,850][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.19710028171539307, 'train/grad_norm_pre_clip': 22.810068130493164, 'timing/data': 0.0006557621527463198, 'timing/model': 1.6952945268712938, 'learning_rate/qwen_vl_interface': 9.310348565683938e-06, 'learning_rate/action_model': 9.310348565683939e-05, 'epoch': 0.37}) +[2026-06-23 18:02:33,188][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.41360044479370117, 'train/grad_norm_pre_clip': 72.11874389648438, 'timing/data': 0.0006540699396282434, 'timing/model': 1.7030222089961171, 'learning_rate/qwen_vl_interface': 9.30876733803656e-06, 'learning_rate/action_model': 9.308767338036559e-05, 'epoch': 0.37}) +[2026-06-23 18:02:36,533][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.29900842905044556, 'train/grad_norm_pre_clip': 73.14657592773438, 'timing/data': 0.000692661153152585, 'timing/model': 1.6957338631618768, 'learning_rate/qwen_vl_interface': 9.30718444198324e-06, 'learning_rate/action_model': 9.307184441983239e-05, 'epoch': 0.37}) +[2026-06-23 18:02:39,876][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.30876028537750244, 'train/grad_norm_pre_clip': 73.9805908203125, 'timing/data': 0.0006710710003972054, 'timing/model': 1.6967417211271822, 'learning_rate/qwen_vl_interface': 9.30559987817465e-06, 'learning_rate/action_model': 9.30559987817465e-05, 'epoch': 0.37}) +[2026-06-23 18:02:43,350][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.22358500957489014, 'train/grad_norm_pre_clip': 20.654727935791016, 'timing/data': 0.0008261119946837425, 'timing/model': 1.8290923880413175, 'learning_rate/qwen_vl_interface': 9.304013647262143e-06, 'learning_rate/action_model': 9.304013647262143e-05, 'epoch': 0.37}) +[2026-06-23 18:02:46,693][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.10441382229328156, 'train/grad_norm_pre_clip': 15.925748825073242, 'timing/data': 0.0007548488210886717, 'timing/model': 1.6999068628065288, 'learning_rate/qwen_vl_interface': 9.302425749897761e-06, 'learning_rate/action_model': 9.302425749897761e-05, 'epoch': 0.37}) +[2026-06-23 18:02:50,031][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.22313232719898224, 'train/grad_norm_pre_clip': 68.23184204101562, 'timing/data': 0.0006641182117164135, 'timing/model': 1.693465008866042, 'learning_rate/qwen_vl_interface': 9.300836186734225e-06, 'learning_rate/action_model': 9.300836186734226e-05, 'epoch': 0.37}) +[2026-06-23 18:02:53,375][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.27600961923599243, 'train/grad_norm_pre_clip': 105.14986419677734, 'timing/data': 0.0005906701553612947, 'timing/model': 1.6989873580168933, 'learning_rate/qwen_vl_interface': 9.299244958424947e-06, 'learning_rate/action_model': 9.299244958424946e-05, 'epoch': 0.37}) +[2026-06-23 18:02:56,715][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.2501792311668396, 'train/grad_norm_pre_clip': 102.91152954101562, 'timing/data': 0.0005880449898540974, 'timing/model': 1.6956523419357836, 'learning_rate/qwen_vl_interface': 9.29765206562402e-06, 'learning_rate/action_model': 9.297652065624018e-05, 'epoch': 0.37}) +[2026-06-23 18:03:00,057][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.22121107578277588, 'train/grad_norm_pre_clip': 79.66252136230469, 'timing/data': 0.0006523560732603073, 'timing/model': 1.6964400340802968, 'learning_rate/qwen_vl_interface': 9.296057508986221e-06, 'learning_rate/action_model': 9.296057508986221e-05, 'epoch': 0.37}) +[2026-06-23 18:03:03,399][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.2549900710582733, 'train/grad_norm_pre_clip': 66.86886596679688, 'timing/data': 0.0006518450099974871, 'timing/model': 1.692729122005403, 'learning_rate/qwen_vl_interface': 9.294461289167013e-06, 'learning_rate/action_model': 9.294461289167014e-05, 'epoch': 0.37}) +[2026-06-23 18:03:06,748][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.11496077477931976, 'train/grad_norm_pre_clip': 22.66091537475586, 'timing/data': 0.0006183618679642677, 'timing/model': 1.695770027115941, 'learning_rate/qwen_vl_interface': 9.292863406822543e-06, 'learning_rate/action_model': 9.292863406822542e-05, 'epoch': 0.37}) +[2026-06-23 18:03:10,082][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.37335655093193054, 'train/grad_norm_pre_clip': 105.26297760009766, 'timing/data': 0.000609035138040781, 'timing/model': 1.6912112219724804, 'learning_rate/qwen_vl_interface': 9.291263862609638e-06, 'learning_rate/action_model': 9.291263862609638e-05, 'epoch': 0.37}) +[2026-06-23 18:03:13,418][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.35900118947029114, 'train/grad_norm_pre_clip': 104.07341766357422, 'timing/data': 0.0005986150354146957, 'timing/model': 1.6912114541046321, 'learning_rate/qwen_vl_interface': 9.28966265718581e-06, 'learning_rate/action_model': 9.28966265718581e-05, 'epoch': 0.37}) +[2026-06-23 18:03:16,762][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.3428690731525421, 'train/grad_norm_pre_clip': 81.64876556396484, 'timing/data': 0.0005968920886516571, 'timing/model': 1.6921808370389044, 'learning_rate/qwen_vl_interface': 9.288059791209256e-06, 'learning_rate/action_model': 9.288059791209256e-05, 'epoch': 0.37}) +[2026-06-23 18:03:20,114][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.22435621917247772, 'train/grad_norm_pre_clip': 67.01517486572266, 'timing/data': 0.0006007188931107521, 'timing/model': 1.7042746250517666, 'learning_rate/qwen_vl_interface': 9.286455265338853e-06, 'learning_rate/action_model': 9.286455265338853e-05, 'epoch': 0.37}) +[2026-06-23 18:03:23,457][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.20600907504558563, 'train/grad_norm_pre_clip': 41.487937927246094, 'timing/data': 0.0008387758862227201, 'timing/model': 1.695478165987879, 'learning_rate/qwen_vl_interface': 9.284849080234161e-06, 'learning_rate/action_model': 9.28484908023416e-05, 'epoch': 0.37}) +[2026-06-23 18:03:26,794][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.2064318060874939, 'train/grad_norm_pre_clip': 88.92247009277344, 'timing/data': 0.0006986618973314762, 'timing/model': 1.6956827118992805, 'learning_rate/qwen_vl_interface': 9.283241236555421e-06, 'learning_rate/action_model': 9.283241236555421e-05, 'epoch': 0.37}) +[2026-06-23 18:03:30,130][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.3110907971858978, 'train/grad_norm_pre_clip': 82.02146911621094, 'timing/data': 0.0006969890091568232, 'timing/model': 1.6919221470598131, 'learning_rate/qwen_vl_interface': 9.281631734963557e-06, 'learning_rate/action_model': 9.281631734963557e-05, 'epoch': 0.38}) +[2026-06-23 18:03:33,559][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.2367224544286728, 'train/grad_norm_pre_clip': 32.548583984375, 'timing/data': 0.0005956289824098349, 'timing/model': 1.7842655051499605, 'learning_rate/qwen_vl_interface': 9.280020576120176e-06, 'learning_rate/action_model': 9.280020576120176e-05, 'epoch': 0.38}) +[2026-06-23 18:03:36,901][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.16244851052761078, 'train/grad_norm_pre_clip': 30.058805465698242, 'timing/data': 0.0005882461555302143, 'timing/model': 1.6972333600278944, 'learning_rate/qwen_vl_interface': 9.278407760687562e-06, 'learning_rate/action_model': 9.278407760687562e-05, 'epoch': 0.38}) +[2026-06-23 18:03:40,236][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.27886247634887695, 'train/grad_norm_pre_clip': 65.31661224365234, 'timing/data': 0.0007099939975887537, 'timing/model': 1.692063125083223, 'learning_rate/qwen_vl_interface': 9.276793289328686e-06, 'learning_rate/action_model': 9.276793289328685e-05, 'epoch': 0.38}) +[2026-06-23 18:03:43,573][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.33688339591026306, 'train/grad_norm_pre_clip': 100.22877502441406, 'timing/data': 0.0005929339677095413, 'timing/model': 1.6898742751218379, 'learning_rate/qwen_vl_interface': 9.27517716270719e-06, 'learning_rate/action_model': 9.275177162707191e-05, 'epoch': 0.38}) +[2026-06-23 18:03:46,917][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.25777921080589294, 'train/grad_norm_pre_clip': 66.27654266357422, 'timing/data': 0.0006012897938489914, 'timing/model': 1.6973841178696603, 'learning_rate/qwen_vl_interface': 9.273559381487409e-06, 'learning_rate/action_model': 9.273559381487409e-05, 'epoch': 0.38}) +[2026-06-23 18:03:50,262][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.16043978929519653, 'train/grad_norm_pre_clip': 78.30170440673828, 'timing/data': 0.0006002280861139297, 'timing/model': 1.700615060981363, 'learning_rate/qwen_vl_interface': 9.27193994633435e-06, 'learning_rate/action_model': 9.27193994633435e-05, 'epoch': 0.38}) +[2026-06-23 18:03:53,598][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.28385627269744873, 'train/grad_norm_pre_clip': 46.97634506225586, 'timing/data': 0.0008235480636358261, 'timing/model': 1.697422823868692, 'learning_rate/qwen_vl_interface': 9.2703188579137e-06, 'learning_rate/action_model': 9.2703188579137e-05, 'epoch': 0.38}) +[2026-06-23 18:03:56,936][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.23263202607631683, 'train/grad_norm_pre_clip': 44.69579315185547, 'timing/data': 0.0006437900010496378, 'timing/model': 1.6911856951192021, 'learning_rate/qwen_vl_interface': 9.26869611689183e-06, 'learning_rate/action_model': 9.26869611689183e-05, 'epoch': 0.38}) +[2026-06-23 18:04:00,273][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.2266506552696228, 'train/grad_norm_pre_clip': 46.03805160522461, 'timing/data': 0.0005988550838083029, 'timing/model': 1.6951255791354924, 'learning_rate/qwen_vl_interface': 9.267071723935784e-06, 'learning_rate/action_model': 9.267071723935784e-05, 'epoch': 0.38}) +[2026-06-23 18:04:03,609][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.29708632826805115, 'train/grad_norm_pre_clip': 88.37623596191406, 'timing/data': 0.0005986148025840521, 'timing/model': 1.69756468385458, 'learning_rate/qwen_vl_interface': 9.26544567971329e-06, 'learning_rate/action_model': 9.26544567971329e-05, 'epoch': 0.38}) +[2026-06-23 18:04:06,954][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.21706688404083252, 'train/grad_norm_pre_clip': 57.52606201171875, 'timing/data': 0.000841842032968998, 'timing/model': 1.6995845208875835, 'learning_rate/qwen_vl_interface': 9.263817984892756e-06, 'learning_rate/action_model': 9.263817984892757e-05, 'epoch': 0.38}) +[2026-06-23 18:04:10,298][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.25414177775382996, 'train/grad_norm_pre_clip': 12.5008544921875, 'timing/data': 0.0006114491261541843, 'timing/model': 1.6934436920564622, 'learning_rate/qwen_vl_interface': 9.262188640143262e-06, 'learning_rate/action_model': 9.262188640143262e-05, 'epoch': 0.38}) +[2026-06-23 18:04:13,646][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.12541238963603973, 'train/grad_norm_pre_clip': 15.730246543884277, 'timing/data': 0.0005989761557430029, 'timing/model': 1.7017722739838064, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.38}) +[2026-06-23 18:04:16,984][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.14808733761310577, 'train/grad_norm_pre_clip': 55.80206298828125, 'timing/data': 0.0006018509157001972, 'timing/model': 1.6977073568850756, 'learning_rate/qwen_vl_interface': 9.258925003537127e-06, 'learning_rate/action_model': 9.258925003537126e-05, 'epoch': 0.38}) +[2026-06-23 18:04:20,326][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.16613730788230896, 'train/grad_norm_pre_clip': 30.6210880279541, 'timing/data': 0.0005950781051069498, 'timing/model': 1.696445677895099, 'learning_rate/qwen_vl_interface': 9.257290713022041e-06, 'learning_rate/action_model': 9.257290713022041e-05, 'epoch': 0.38}) +[2026-06-23 18:04:23,760][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.2630074918270111, 'train/grad_norm_pre_clip': 48.93386459350586, 'timing/data': 0.0006138740573078394, 'timing/model': 1.6970716160722077, 'learning_rate/qwen_vl_interface': 9.255654775261112e-06, 'learning_rate/action_model': 9.255654775261112e-05, 'epoch': 0.38}) +[2026-06-23 18:04:27,105][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.14560647308826447, 'train/grad_norm_pre_clip': 15.885668754577637, 'timing/data': 0.0005849390290677547, 'timing/model': 1.7024013819172978, 'learning_rate/qwen_vl_interface': 9.254017190926813e-06, 'learning_rate/action_model': 9.254017190926813e-05, 'epoch': 0.38}) +[2026-06-23 18:04:30,447][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.19454801082611084, 'train/grad_norm_pre_clip': 64.95608520507812, 'timing/data': 0.0006032739765942097, 'timing/model': 1.6958835148252547, 'learning_rate/qwen_vl_interface': 9.25237796069229e-06, 'learning_rate/action_model': 9.25237796069229e-05, 'epoch': 0.38}) +[2026-06-23 18:04:33,781][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.594497799873352, 'train/grad_norm_pre_clip': 194.01739501953125, 'timing/data': 0.0005925828590989113, 'timing/model': 1.6971863738726825, 'learning_rate/qwen_vl_interface': 9.250737085231371e-06, 'learning_rate/action_model': 9.250737085231372e-05, 'epoch': 0.38}) +[2026-06-23 18:04:37,110][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.4060255289077759, 'train/grad_norm_pre_clip': 99.89413452148438, 'timing/data': 0.0005867029540240765, 'timing/model': 1.6890712960157543, 'learning_rate/qwen_vl_interface': 9.24909456521856e-06, 'learning_rate/action_model': 9.249094565218559e-05, 'epoch': 0.38}) +[2026-06-23 18:04:40,451][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.2418714314699173, 'train/grad_norm_pre_clip': 98.55211639404297, 'timing/data': 0.0006075119599699974, 'timing/model': 1.6960410049650818, 'learning_rate/qwen_vl_interface': 9.247450401329027e-06, 'learning_rate/action_model': 9.247450401329027e-05, 'epoch': 0.38}) +[2026-06-23 18:04:43,788][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.2100573480129242, 'train/grad_norm_pre_clip': 64.73013305664062, 'timing/data': 0.0005867129657417536, 'timing/model': 1.6991219560150057, 'learning_rate/qwen_vl_interface': 9.245804594238637e-06, 'learning_rate/action_model': 9.245804594238636e-05, 'epoch': 0.38}) +[2026-06-23 18:04:47,136][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.16445401310920715, 'train/grad_norm_pre_clip': 8.927122116088867, 'timing/data': 0.0006579668261110783, 'timing/model': 1.697282230015844, 'learning_rate/qwen_vl_interface': 9.244157144623909e-06, 'learning_rate/action_model': 9.244157144623909e-05, 'epoch': 0.38}) +[2026-06-23 18:04:50,483][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.24086466431617737, 'train/grad_norm_pre_clip': 74.02192687988281, 'timing/data': 0.0005945870652794838, 'timing/model': 1.6963819488883018, 'learning_rate/qwen_vl_interface': 9.242508053162051e-06, 'learning_rate/action_model': 9.242508053162051e-05, 'epoch': 0.38}) +[2026-06-23 18:04:53,825][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.2216922640800476, 'train/grad_norm_pre_clip': 44.04719924926758, 'timing/data': 0.0005960999988019466, 'timing/model': 1.6955627910792828, 'learning_rate/qwen_vl_interface': 9.240857320530944e-06, 'learning_rate/action_model': 9.240857320530943e-05, 'epoch': 0.38}) +[2026-06-23 18:04:57,172][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.21089015901088715, 'train/grad_norm_pre_clip': 54.28589630126953, 'timing/data': 0.0006023619789630175, 'timing/model': 1.7038459880277514, 'learning_rate/qwen_vl_interface': 9.23920494740914e-06, 'learning_rate/action_model': 9.23920494740914e-05, 'epoch': 0.39}) +[2026-06-23 18:05:00,505][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.4084387421607971, 'train/grad_norm_pre_clip': 87.24849700927734, 'timing/data': 0.0006071110256016254, 'timing/model': 1.693551684031263, 'learning_rate/qwen_vl_interface': 9.237550934475868e-06, 'learning_rate/action_model': 9.237550934475867e-05, 'epoch': 0.39}) +[2026-06-23 18:05:03,843][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.1974654644727707, 'train/grad_norm_pre_clip': 10.94320297241211, 'timing/data': 0.0006034739781171083, 'timing/model': 1.693550464231521, 'learning_rate/qwen_vl_interface': 9.235895282411029e-06, 'learning_rate/action_model': 9.235895282411029e-05, 'epoch': 0.39}) +[2026-06-23 18:05:07,188][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.34088656306266785, 'train/grad_norm_pre_clip': 72.22093200683594, 'timing/data': 0.0005934750661253929, 'timing/model': 1.698502419050783, 'learning_rate/qwen_vl_interface': 9.2342379918952e-06, 'learning_rate/action_model': 9.2342379918952e-05, 'epoch': 0.39}) +[2026-06-23 18:05:10,516][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.2693367004394531, 'train/grad_norm_pre_clip': 85.71101379394531, 'timing/data': 0.0006045359186828136, 'timing/model': 1.6853775070048869, 'learning_rate/qwen_vl_interface': 9.232579063609631e-06, 'learning_rate/action_model': 9.232579063609631e-05, 'epoch': 0.39}) +[2026-06-23 18:05:13,956][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1000 +[2026-06-23 18:06:47,381][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1000 +[2026-06-23 18:06:47,381][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1000 +[2026-06-23 18:07:58,593][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1000 +[2026-06-23 18:08:05,500][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000 +[2026-06-23 18:08:05,501][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.14775845408439636, 'train/grad_norm_pre_clip': 56.400814056396484, 'eval/action_loss/seconds': 93.4227817452047, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.12358386814594269, 'eval/latency_2/loss': 0.12358386814594269, 'eval/flappy/loss': 0.12358386814594269, 'eval/flappy/latency_2/loss': 0.12358386814594269, 'eval/flap_precision': 0.9528455284552846, 'eval/flap_recall': 0.7908232118758435, 'eval/flap_f1': 0.864306784660767, 'eval/flap_rate_model': 0.17083333333333334, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.864306784660767, 'eval/latency_2/flap_precision': 0.9528455284552846, 'eval/latency_2/flap_recall': 0.7908232118758435, 'eval/latency_2/flap_f1': 0.864306784660767, 'eval/latency_2/flap_rate_model': 0.17083333333333334, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.864306784660767, 'eval/action_classification/data_wait_seconds': 0.9977372288703918, 'eval/action_classification/predict_seconds': 70.08353424072266, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 3.1399999999999997, 'rl_games_eval/mid_train/mean_length': 42.4, 'rl_games_eval/mid_train/std_reward': 0.9308061022576063, 'rl_games_eval/mid_train/std_length': 9.308061022576076, 'rl_games_eval/mid_train/macro_mean_reward': 3.1399999999999997, 'rl_games_eval/mid_train/macro_mean_length': 42.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 3.1399999999999997, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 42.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 0.9308061022576063, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 9.308061022576076, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 3.1399999999999997, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006074020639061928, 'timing/model': 1.6952872201800346, 'learning_rate/qwen_vl_interface': 9.230918498236243e-06, 'learning_rate/action_model': 9.230918498236244e-05, 'epoch': 0.39}) +[2026-06-23 18:08:05,501][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_1000_state +[2026-06-23 18:08:05,504][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 18:08:08,192][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_1000_state/model.safetensors +[2026-06-23 18:08:09,344][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_1000_state/optimizer.bin +[2026-06-23 18:08:09,344][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_1000_state/sampler.bin +[2026-06-23 18:08:09,345][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_1000_state/sampler_1.bin +[2026-06-23 18:08:09,345][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_1000_state/random_states_0.pkl +[2026-06-23 18:08:09,370][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000 +[2026-06-23 18:08:09,370][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 18:08:09,381][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 18:08:12,694][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.27103182673454285, 'train/grad_norm_pre_clip': 79.00690460205078, 'timing/data': 0.0006885139737278223, 'timing/model': 1.6834100731648505, 'learning_rate/qwen_vl_interface': 9.229256296457635e-06, 'learning_rate/action_model': 9.229256296457634e-05, 'epoch': 0.39}) +[2026-06-23 18:08:16,007][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.22415360808372498, 'train/grad_norm_pre_clip': 45.72490310668945, 'timing/data': 0.000674587907269597, 'timing/model': 1.6829965359065682, 'learning_rate/qwen_vl_interface': 9.22759245895707e-06, 'learning_rate/action_model': 9.227592458957071e-05, 'epoch': 0.39}) +[2026-06-23 18:08:19,326][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.17675122618675232, 'train/grad_norm_pre_clip': 41.20058059692383, 'timing/data': 0.0006044157780706882, 'timing/model': 1.6842564039397985, 'learning_rate/qwen_vl_interface': 9.225926986418494e-06, 'learning_rate/action_model': 9.225926986418494e-05, 'epoch': 0.39}) +[2026-06-23 18:08:22,643][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.25056201219558716, 'train/grad_norm_pre_clip': 117.9715805053711, 'timing/data': 0.0008182169403880835, 'timing/model': 1.6850974259432405, 'learning_rate/qwen_vl_interface': 9.224259879526519e-06, 'learning_rate/action_model': 9.224259879526519e-05, 'epoch': 0.39}) +[2026-06-23 18:08:25,965][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.28727030754089355, 'train/grad_norm_pre_clip': 105.07565307617188, 'timing/data': 0.0005997170228511095, 'timing/model': 1.6886099269613624, 'learning_rate/qwen_vl_interface': 9.222591138966427e-06, 'learning_rate/action_model': 9.222591138966426e-05, 'epoch': 0.39}) +[2026-06-23 18:08:29,291][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.29272541403770447, 'train/grad_norm_pre_clip': 101.40538024902344, 'timing/data': 0.0007458110339939594, 'timing/model': 1.688919979846105, 'learning_rate/qwen_vl_interface': 9.220920765424175e-06, 'learning_rate/action_model': 9.220920765424174e-05, 'epoch': 0.39}) +[2026-06-23 18:08:32,716][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.299435019493103, 'train/grad_norm_pre_clip': 77.12188720703125, 'timing/data': 0.0006433690432459116, 'timing/model': 1.790809583151713, 'learning_rate/qwen_vl_interface': 9.219248759586392e-06, 'learning_rate/action_model': 9.219248759586391e-05, 'epoch': 0.39}) +[2026-06-23 18:08:36,038][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.25306376814842224, 'train/grad_norm_pre_clip': 18.985342025756836, 'timing/data': 0.0006510838866233826, 'timing/model': 1.68692820915021, 'learning_rate/qwen_vl_interface': 9.217575122140375e-06, 'learning_rate/action_model': 9.217575122140375e-05, 'epoch': 0.39}) +[2026-06-23 18:08:39,372][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.1681666523218155, 'train/grad_norm_pre_clip': 36.33350372314453, 'timing/data': 0.0005942760035395622, 'timing/model': 1.6936624539084733, 'learning_rate/qwen_vl_interface': 9.215899853774094e-06, 'learning_rate/action_model': 9.215899853774093e-05, 'epoch': 0.39}) +[2026-06-23 18:08:42,705][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.24334484338760376, 'train/grad_norm_pre_clip': 62.796016693115234, 'timing/data': 0.0005949179176241159, 'timing/model': 1.6923316449392587, 'learning_rate/qwen_vl_interface': 9.214222955176189e-06, 'learning_rate/action_model': 9.214222955176189e-05, 'epoch': 0.39}) +[2026-06-23 18:08:46,036][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.1835581660270691, 'train/grad_norm_pre_clip': 33.80104064941406, 'timing/data': 0.0006086640059947968, 'timing/model': 1.6938639648724347, 'learning_rate/qwen_vl_interface': 9.212544427035967e-06, 'learning_rate/action_model': 9.212544427035967e-05, 'epoch': 0.39}) +[2026-06-23 18:08:49,363][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.2065877765417099, 'train/grad_norm_pre_clip': 11.645783424377441, 'timing/data': 0.0006034139078110456, 'timing/model': 1.6916469598654658, 'learning_rate/qwen_vl_interface': 9.21086427004341e-06, 'learning_rate/action_model': 9.210864270043411e-05, 'epoch': 0.39}) +[2026-06-23 18:08:52,698][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.25288745760917664, 'train/grad_norm_pre_clip': 50.20351791381836, 'timing/data': 0.0005904489662498236, 'timing/model': 1.6933241919614375, 'learning_rate/qwen_vl_interface': 9.209182484889168e-06, 'learning_rate/action_model': 9.209182484889167e-05, 'epoch': 0.39}) +[2026-06-23 18:08:56,039][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.14859513938426971, 'train/grad_norm_pre_clip': 13.080256462097168, 'timing/data': 0.0007069988641887903, 'timing/model': 1.6974828781094402, 'learning_rate/qwen_vl_interface': 9.207499072264557e-06, 'learning_rate/action_model': 9.207499072264558e-05, 'epoch': 0.39}) +[2026-06-23 18:08:59,378][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.2300776243209839, 'train/grad_norm_pre_clip': 11.248162269592285, 'timing/data': 0.0005971919745206833, 'timing/model': 1.70019631809555, 'learning_rate/qwen_vl_interface': 9.205814032861566e-06, 'learning_rate/action_model': 9.205814032861566e-05, 'epoch': 0.39}) +[2026-06-23 18:09:02,714][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.23333734273910522, 'train/grad_norm_pre_clip': 14.198264122009277, 'timing/data': 0.0008542451541870832, 'timing/model': 1.696220476878807, 'learning_rate/qwen_vl_interface': 9.204127367372849e-06, 'learning_rate/action_model': 9.204127367372849e-05, 'epoch': 0.39}) +[2026-06-23 18:09:06,059][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.18359898030757904, 'train/grad_norm_pre_clip': 16.79412078857422, 'timing/data': 0.0006009999196976423, 'timing/model': 1.6992797469720244, 'learning_rate/qwen_vl_interface': 9.202439076491732e-06, 'learning_rate/action_model': 9.202439076491732e-05, 'epoch': 0.39}) +[2026-06-23 18:09:09,406][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.19411925971508026, 'train/grad_norm_pre_clip': 12.31601619720459, 'timing/data': 0.0006021310109645128, 'timing/model': 1.699669630965218, 'learning_rate/qwen_vl_interface': 9.200749160912208e-06, 'learning_rate/action_model': 9.200749160912208e-05, 'epoch': 0.39}) +[2026-06-23 18:09:12,752][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.1630057543516159, 'train/grad_norm_pre_clip': 22.291194915771484, 'timing/data': 0.000587955117225647, 'timing/model': 1.6998824609909207, 'learning_rate/qwen_vl_interface': 9.199057621328937e-06, 'learning_rate/action_model': 9.199057621328935e-05, 'epoch': 0.39}) +[2026-06-23 18:09:16,095][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.10259317606687546, 'train/grad_norm_pre_clip': 23.61553955078125, 'timing/data': 0.0005803098902106285, 'timing/model': 1.695795139996335, 'learning_rate/qwen_vl_interface': 9.197364458437244e-06, 'learning_rate/action_model': 9.197364458437245e-05, 'epoch': 0.39}) +[2026-06-23 18:09:19,440][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.3138309121131897, 'train/grad_norm_pre_clip': 15.751444816589355, 'timing/data': 0.0007177488878369331, 'timing/model': 1.6979292519390583, 'learning_rate/qwen_vl_interface': 9.19566967293313e-06, 'learning_rate/action_model': 9.19566967293313e-05, 'epoch': 0.4}) +[2026-06-23 18:09:22,879][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.17322511970996857, 'train/grad_norm_pre_clip': 47.737369537353516, 'timing/data': 0.0006056590937077999, 'timing/model': 1.6999168379697949, 'learning_rate/qwen_vl_interface': 9.193973265513253e-06, 'learning_rate/action_model': 9.193973265513253e-05, 'epoch': 0.4}) +[2026-06-23 18:09:26,224][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.22577248513698578, 'train/grad_norm_pre_clip': 20.773826599121094, 'timing/data': 0.0005849499721080065, 'timing/model': 1.6988975629210472, 'learning_rate/qwen_vl_interface': 9.192275236874946e-06, 'learning_rate/action_model': 9.192275236874946e-05, 'epoch': 0.4}) +[2026-06-23 18:09:29,567][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.14176009595394135, 'train/grad_norm_pre_clip': 23.598093032836914, 'timing/data': 0.0005906091537326574, 'timing/model': 1.6961731240153313, 'learning_rate/qwen_vl_interface': 9.190575587716199e-06, 'learning_rate/action_model': 9.190575587716199e-05, 'epoch': 0.4}) +[2026-06-23 18:09:32,915][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.1919696182012558, 'train/grad_norm_pre_clip': 68.00059509277344, 'timing/data': 0.0005853599868714809, 'timing/model': 1.7022486119531095, 'learning_rate/qwen_vl_interface': 9.188874318735677e-06, 'learning_rate/action_model': 9.188874318735676e-05, 'epoch': 0.4}) +[2026-06-23 18:09:36,261][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.18037360906600952, 'train/grad_norm_pre_clip': 17.639802932739258, 'timing/data': 0.0005903190467506647, 'timing/model': 1.698233837960288, 'learning_rate/qwen_vl_interface': 9.187171430632707e-06, 'learning_rate/action_model': 9.187171430632707e-05, 'epoch': 0.4}) +[2026-06-23 18:09:39,609][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.10984928905963898, 'train/grad_norm_pre_clip': 35.50041198730469, 'timing/data': 0.0006145050283521414, 'timing/model': 1.6986258861143142, 'learning_rate/qwen_vl_interface': 9.185466924107283e-06, 'learning_rate/action_model': 9.185466924107282e-05, 'epoch': 0.4}) +[2026-06-23 18:09:42,953][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.23451732099056244, 'train/grad_norm_pre_clip': 30.490875244140625, 'timing/data': 0.0005998669657856226, 'timing/model': 1.6962088239379227, 'learning_rate/qwen_vl_interface': 9.18376079986006e-06, 'learning_rate/action_model': 9.183760799860061e-05, 'epoch': 0.4}) +[2026-06-23 18:09:46,303][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.3863220512866974, 'train/grad_norm_pre_clip': 18.510601043701172, 'timing/data': 0.000598234124481678, 'timing/model': 1.6992040139157325, 'learning_rate/qwen_vl_interface': 9.182053058592365e-06, 'learning_rate/action_model': 9.182053058592365e-05, 'epoch': 0.4}) +[2026-06-23 18:09:49,643][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.20824846625328064, 'train/grad_norm_pre_clip': 63.2168083190918, 'timing/data': 0.0005972818471491337, 'timing/model': 1.6965620389673859, 'learning_rate/qwen_vl_interface': 9.180343701006183e-06, 'learning_rate/action_model': 9.180343701006183e-05, 'epoch': 0.4}) +[2026-06-23 18:09:52,986][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.25108614563941956, 'train/grad_norm_pre_clip': 28.341835021972656, 'timing/data': 0.0007192010525614023, 'timing/model': 1.7016397009138018, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.4}) +[2026-06-23 18:09:56,331][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.19091567397117615, 'train/grad_norm_pre_clip': 18.489904403686523, 'timing/data': 0.0006080630701035261, 'timing/model': 1.6996159199625254, 'learning_rate/qwen_vl_interface': 9.176920139689638e-06, 'learning_rate/action_model': 9.176920139689637e-05, 'epoch': 0.4}) +[2026-06-23 18:09:59,675][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.35329940915107727, 'train/grad_norm_pre_clip': 64.16532897949219, 'timing/data': 0.0006144240032881498, 'timing/model': 1.6967457060236484, 'learning_rate/qwen_vl_interface': 9.175205937366568e-06, 'learning_rate/action_model': 9.175205937366569e-05, 'epoch': 0.4}) +[2026-06-23 18:10:03,023][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.1770620346069336, 'train/grad_norm_pre_clip': 41.08979415893555, 'timing/data': 0.0005971018690615892, 'timing/model': 1.6990919560194016, 'learning_rate/qwen_vl_interface': 9.173490121539608e-06, 'learning_rate/action_model': 9.173490121539608e-05, 'epoch': 0.4}) +[2026-06-23 18:10:06,371][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.12667319178581238, 'train/grad_norm_pre_clip': 12.068547248840332, 'timing/data': 0.0005982751026749611, 'timing/model': 1.701737207127735, 'learning_rate/qwen_vl_interface': 9.171772692914062e-06, 'learning_rate/action_model': 9.171772692914061e-05, 'epoch': 0.4}) +[2026-06-23 18:10:09,719][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.19614753127098083, 'train/grad_norm_pre_clip': 61.82358932495117, 'timing/data': 0.0006020108703523874, 'timing/model': 1.698453907854855, 'learning_rate/qwen_vl_interface': 9.1700536521959e-06, 'learning_rate/action_model': 9.170053652195899e-05, 'epoch': 0.4}) +[2026-06-23 18:10:13,174][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.2744935154914856, 'train/grad_norm_pre_clip': 73.56716918945312, 'timing/data': 0.0005924040451645851, 'timing/model': 1.7000333371106535, 'learning_rate/qwen_vl_interface': 9.168333000091754e-06, 'learning_rate/action_model': 9.168333000091754e-05, 'epoch': 0.4}) +[2026-06-23 18:10:16,530][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.2406824678182602, 'train/grad_norm_pre_clip': 52.495094299316406, 'timing/data': 0.0005904501304030418, 'timing/model': 1.7060146878939122, 'learning_rate/qwen_vl_interface': 9.16661073730892e-06, 'learning_rate/action_model': 9.166610737308919e-05, 'epoch': 0.4}) +[2026-06-23 18:10:19,880][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.22810201346874237, 'train/grad_norm_pre_clip': 31.684371948242188, 'timing/data': 0.0005931949708610773, 'timing/model': 1.700637985020876, 'learning_rate/qwen_vl_interface': 9.164886864555355e-06, 'learning_rate/action_model': 9.164886864555355e-05, 'epoch': 0.4}) +[2026-06-23 18:10:23,238][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.1133824810385704, 'train/grad_norm_pre_clip': 18.252031326293945, 'timing/data': 0.0005962499417364597, 'timing/model': 1.7045504269190133, 'learning_rate/qwen_vl_interface': 9.163161382539678e-06, 'learning_rate/action_model': 9.163161382539678e-05, 'epoch': 0.4}) +[2026-06-23 18:10:26,590][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.12549564242362976, 'train/grad_norm_pre_clip': 58.11549377441406, 'timing/data': 0.0005896880757063627, 'timing/model': 1.7025644360110164, 'learning_rate/qwen_vl_interface': 9.16143429197117e-06, 'learning_rate/action_model': 9.16143429197117e-05, 'epoch': 0.4}) +[2026-06-23 18:10:29,941][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.20222218334674835, 'train/grad_norm_pre_clip': 28.081443786621094, 'timing/data': 0.0005978439003229141, 'timing/model': 1.6977783129550517, 'learning_rate/qwen_vl_interface': 9.159705593559773e-06, 'learning_rate/action_model': 9.159705593559773e-05, 'epoch': 0.4}) +[2026-06-23 18:10:33,291][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.1892198920249939, 'train/grad_norm_pre_clip': 17.1452693939209, 'timing/data': 0.0006010590586811304, 'timing/model': 1.7009820130188018, 'learning_rate/qwen_vl_interface': 9.157975288016089e-06, 'learning_rate/action_model': 9.157975288016088e-05, 'epoch': 0.4}) +[2026-06-23 18:10:36,636][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.18645921349525452, 'train/grad_norm_pre_clip': 31.90919303894043, 'timing/data': 0.0006689371075481176, 'timing/model': 1.6938908989541233, 'learning_rate/qwen_vl_interface': 9.156243376051381e-06, 'learning_rate/action_model': 9.156243376051381e-05, 'epoch': 0.4}) +[2026-06-23 18:10:39,983][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.2198750376701355, 'train/grad_norm_pre_clip': 61.62116622924805, 'timing/data': 0.0007452710997313261, 'timing/model': 1.7028366199228913, 'learning_rate/qwen_vl_interface': 9.154509858377574e-06, 'learning_rate/action_model': 9.154509858377574e-05, 'epoch': 0.4}) +[2026-06-23 18:10:43,336][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.14388227462768555, 'train/grad_norm_pre_clip': 32.29182052612305, 'timing/data': 0.0006474461406469345, 'timing/model': 1.699342945124954, 'learning_rate/qwen_vl_interface': 9.152774735707248e-06, 'learning_rate/action_model': 9.152774735707247e-05, 'epoch': 0.4}) +[2026-06-23 18:10:46,679][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.22151818871498108, 'train/grad_norm_pre_clip': 17.00250816345215, 'timing/data': 0.0006891749799251556, 'timing/model': 1.6963385499548167, 'learning_rate/qwen_vl_interface': 9.15103800875365e-06, 'learning_rate/action_model': 9.15103800875365e-05, 'epoch': 0.41}) +[2026-06-23 18:10:50,032][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.24439279735088348, 'train/grad_norm_pre_clip': 43.472198486328125, 'timing/data': 0.0008299101609736681, 'timing/model': 1.701062788022682, 'learning_rate/qwen_vl_interface': 9.149299678230681e-06, 'learning_rate/action_model': 9.149299678230681e-05, 'epoch': 0.41}) +[2026-06-23 18:10:53,376][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.289590060710907, 'train/grad_norm_pre_clip': 42.850284576416016, 'timing/data': 0.0006477769929915667, 'timing/model': 1.6971160031389445, 'learning_rate/qwen_vl_interface': 9.147559744852905e-06, 'learning_rate/action_model': 9.147559744852905e-05, 'epoch': 0.41}) +[2026-06-23 18:10:56,722][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.12276268750429153, 'train/grad_norm_pre_clip': 19.759464263916016, 'timing/data': 0.0005973130464553833, 'timing/model': 1.6986494739539921, 'learning_rate/qwen_vl_interface': 9.14581820933554e-06, 'learning_rate/action_model': 9.14581820933554e-05, 'epoch': 0.41}) +[2026-06-23 18:11:00,156][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.2091800421476364, 'train/grad_norm_pre_clip': 31.371417999267578, 'timing/data': 0.0005960410926491022, 'timing/model': 1.7857150409836322, 'learning_rate/qwen_vl_interface': 9.14407507239447e-06, 'learning_rate/action_model': 9.14407507239447e-05, 'epoch': 0.41}) +[2026-06-23 18:11:03,507][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.20938223600387573, 'train/grad_norm_pre_clip': 58.33910369873047, 'timing/data': 0.0006889351643621922, 'timing/model': 1.6969004990532994, 'learning_rate/qwen_vl_interface': 9.142330334746227e-06, 'learning_rate/action_model': 9.142330334746227e-05, 'epoch': 0.41}) +[2026-06-23 18:11:06,862][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.17455990612506866, 'train/grad_norm_pre_clip': 57.20537567138672, 'timing/data': 0.0006993738934397697, 'timing/model': 1.7050453750416636, 'learning_rate/qwen_vl_interface': 9.14058399710801e-06, 'learning_rate/action_model': 9.14058399710801e-05, 'epoch': 0.41}) +[2026-06-23 18:11:10,212][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.13659749925136566, 'train/grad_norm_pre_clip': 85.36344909667969, 'timing/data': 0.0006007188931107521, 'timing/model': 1.7005047029815614, 'learning_rate/qwen_vl_interface': 9.138836060197673e-06, 'learning_rate/action_model': 9.138836060197672e-05, 'epoch': 0.41}) +[2026-06-23 18:11:13,566][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.23374439775943756, 'train/grad_norm_pre_clip': 74.22456359863281, 'timing/data': 0.0009015148971229792, 'timing/model': 1.7007433520630002, 'learning_rate/qwen_vl_interface': 9.137086524733725e-06, 'learning_rate/action_model': 9.137086524733724e-05, 'epoch': 0.41}) +[2026-06-23 18:11:16,910][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.18886302411556244, 'train/grad_norm_pre_clip': 55.048011779785156, 'timing/data': 0.0006543099880218506, 'timing/model': 1.6956359650939703, 'learning_rate/qwen_vl_interface': 9.135335391435336e-06, 'learning_rate/action_model': 9.135335391435335e-05, 'epoch': 0.41}) +[2026-06-23 18:11:20,260][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.1190798357129097, 'train/grad_norm_pre_clip': 38.955787658691406, 'timing/data': 0.0006613130681216717, 'timing/model': 1.6993557561654598, 'learning_rate/qwen_vl_interface': 9.13358266102233e-06, 'learning_rate/action_model': 9.133582661022328e-05, 'epoch': 0.41}) +[2026-06-23 18:11:23,608][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.20589111745357513, 'train/grad_norm_pre_clip': 33.64970397949219, 'timing/data': 0.000634332187473774, 'timing/model': 1.7001996519975364, 'learning_rate/qwen_vl_interface': 9.131828334215185e-06, 'learning_rate/action_model': 9.131828334215185e-05, 'epoch': 0.41}) +[2026-06-23 18:11:26,956][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.2000076025724411, 'train/grad_norm_pre_clip': 43.310184478759766, 'timing/data': 0.0006939440499991179, 'timing/model': 1.6957319870125502, 'learning_rate/qwen_vl_interface': 9.130072411735043e-06, 'learning_rate/action_model': 9.130072411735043e-05, 'epoch': 0.41}) +[2026-06-23 18:11:30,310][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.2722184360027313, 'train/grad_norm_pre_clip': 21.775611877441406, 'timing/data': 0.0005937549285590649, 'timing/model': 1.7033828250132501, 'learning_rate/qwen_vl_interface': 9.128314894303693e-06, 'learning_rate/action_model': 9.128314894303693e-05, 'epoch': 0.41}) +[2026-06-23 18:11:33,665][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.27887123823165894, 'train/grad_norm_pre_clip': 89.50347137451172, 'timing/data': 0.0006080730818212032, 'timing/model': 1.705786813981831, 'learning_rate/qwen_vl_interface': 9.126555782643592e-06, 'learning_rate/action_model': 9.126555782643591e-05, 'epoch': 0.41}) +[2026-06-23 18:11:37,018][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.3309817314147949, 'train/grad_norm_pre_clip': 118.02223205566406, 'timing/data': 0.0006035638507455587, 'timing/model': 1.7034008412156254, 'learning_rate/qwen_vl_interface': 9.124795077477835e-06, 'learning_rate/action_model': 9.124795077477834e-05, 'epoch': 0.41}) +[2026-06-23 18:11:40,381][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.4009588062763214, 'train/grad_norm_pre_clip': 86.62806701660156, 'timing/data': 0.0006029428914189339, 'timing/model': 1.7055508010089397, 'learning_rate/qwen_vl_interface': 9.123032779530186e-06, 'learning_rate/action_model': 9.123032779530187e-05, 'epoch': 0.41}) +[2026-06-23 18:11:43,724][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.22107736766338348, 'train/grad_norm_pre_clip': 77.99404907226562, 'timing/data': 0.0005990448407828808, 'timing/model': 1.699476094916463, 'learning_rate/qwen_vl_interface': 9.12126888952506e-06, 'learning_rate/action_model': 9.12126888952506e-05, 'epoch': 0.41}) +[2026-06-23 18:11:47,076][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.1716735064983368, 'train/grad_norm_pre_clip': 36.92935562133789, 'timing/data': 0.0006071510724723339, 'timing/model': 1.7014028569683433, 'learning_rate/qwen_vl_interface': 9.119503408187521e-06, 'learning_rate/action_model': 9.119503408187522e-05, 'epoch': 0.41}) +[2026-06-23 18:11:50,726][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.286434143781662, 'train/grad_norm_pre_clip': 67.59444427490234, 'timing/data': 0.0005930939223617315, 'timing/model': 2.002314280020073, 'learning_rate/qwen_vl_interface': 9.117736336243299e-06, 'learning_rate/action_model': 9.117736336243298e-05, 'epoch': 0.41}) +[2026-06-23 18:11:54,076][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.1976400911808014, 'train/grad_norm_pre_clip': 71.24251556396484, 'timing/data': 0.0005893080960959196, 'timing/model': 1.6982434811070561, 'learning_rate/qwen_vl_interface': 9.115967674418765e-06, 'learning_rate/action_model': 9.115967674418765e-05, 'epoch': 0.41}) +[2026-06-23 18:11:57,420][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.2082289755344391, 'train/grad_norm_pre_clip': 65.56720733642578, 'timing/data': 0.0006460640579462051, 'timing/model': 1.698449739953503, 'learning_rate/qwen_vl_interface': 9.114197423440949e-06, 'learning_rate/action_model': 9.11419742344095e-05, 'epoch': 0.41}) +[2026-06-23 18:12:00,767][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.3148840069770813, 'train/grad_norm_pre_clip': 53.46928024291992, 'timing/data': 0.0006519951857626438, 'timing/model': 1.698356755077839, 'learning_rate/qwen_vl_interface': 9.112425584037536e-06, 'learning_rate/action_model': 9.112425584037536e-05, 'epoch': 0.41}) +[2026-06-23 18:12:04,118][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.2621496915817261, 'train/grad_norm_pre_clip': 26.99077606201172, 'timing/data': 0.0006041049491614103, 'timing/model': 1.699487331090495, 'learning_rate/qwen_vl_interface': 9.110652156936863e-06, 'learning_rate/action_model': 9.110652156936863e-05, 'epoch': 0.41}) +[2026-06-23 18:12:07,470][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.2753710150718689, 'train/grad_norm_pre_clip': 51.706485748291016, 'timing/data': 0.0008213440887629986, 'timing/model': 1.707445174921304, 'learning_rate/qwen_vl_interface': 9.108877142867918e-06, 'learning_rate/action_model': 9.108877142867918e-05, 'epoch': 0.41}) +[2026-06-23 18:12:10,810][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.19635716080665588, 'train/grad_norm_pre_clip': 46.08741760253906, 'timing/data': 0.0006660211365669966, 'timing/model': 1.6964990759734064, 'learning_rate/qwen_vl_interface': 9.107100542560342e-06, 'learning_rate/action_model': 9.107100542560342e-05, 'epoch': 0.41}) +[2026-06-23 18:12:14,162][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.1971483826637268, 'train/grad_norm_pre_clip': 38.61199188232422, 'timing/data': 0.0006955070421099663, 'timing/model': 1.702018678886816, 'learning_rate/qwen_vl_interface': 9.105322356744429e-06, 'learning_rate/action_model': 9.105322356744429e-05, 'epoch': 0.42}) +[2026-06-23 18:12:17,517][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.1830853372812271, 'train/grad_norm_pre_clip': 45.32378387451172, 'timing/data': 0.0006681161466985941, 'timing/model': 1.70206024707295, 'learning_rate/qwen_vl_interface': 9.103542586151121e-06, 'learning_rate/action_model': 9.10354258615112e-05, 'epoch': 0.42}) +[2026-06-23 18:12:20,869][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.18913975358009338, 'train/grad_norm_pre_clip': 24.844154357910156, 'timing/data': 0.0008726899977773428, 'timing/model': 1.7042745088692755, 'learning_rate/qwen_vl_interface': 9.101761231512017e-06, 'learning_rate/action_model': 9.101761231512017e-05, 'epoch': 0.42}) +[2026-06-23 18:12:24,214][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.1544792205095291, 'train/grad_norm_pre_clip': 62.079872131347656, 'timing/data': 0.0006017410196363926, 'timing/model': 1.6995582259260118, 'learning_rate/qwen_vl_interface': 9.099978293559365e-06, 'learning_rate/action_model': 9.099978293559366e-05, 'epoch': 0.42}) +[2026-06-23 18:12:27,560][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.21840661764144897, 'train/grad_norm_pre_clip': 72.5185546875, 'timing/data': 0.000587784918025136, 'timing/model': 1.6979969029780477, 'learning_rate/qwen_vl_interface': 9.098193773026065e-06, 'learning_rate/action_model': 9.098193773026064e-05, 'epoch': 0.42}) +[2026-06-23 18:12:30,909][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.2217615395784378, 'train/grad_norm_pre_clip': 77.40311431884766, 'timing/data': 0.0006070609670132399, 'timing/model': 1.7011551389005035, 'learning_rate/qwen_vl_interface': 9.096407670645663e-06, 'learning_rate/action_model': 9.096407670645662e-05, 'epoch': 0.42}) +[2026-06-23 18:12:34,257][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.31647756695747375, 'train/grad_norm_pre_clip': 93.14833068847656, 'timing/data': 0.0006130121182650328, 'timing/model': 1.7009465771261603, 'learning_rate/qwen_vl_interface': 9.094619987152357e-06, 'learning_rate/action_model': 9.094619987152358e-05, 'epoch': 0.42}) +[2026-06-23 18:12:37,602][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.23858608305454254, 'train/grad_norm_pre_clip': 34.24367904663086, 'timing/data': 0.0005967919714748859, 'timing/model': 1.6956307778600603, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.42}) +[2026-06-23 18:12:41,260][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.12476087361574173, 'train/grad_norm_pre_clip': 54.190494537353516, 'timing/data': 0.0006301740650087595, 'timing/model': 2.001796825788915, 'learning_rate/qwen_vl_interface': 9.09103987976709e-06, 'learning_rate/action_model': 9.09103987976709e-05, 'epoch': 0.42}) +[2026-06-23 18:12:44,606][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.16332480311393738, 'train/grad_norm_pre_clip': 27.381650924682617, 'timing/data': 0.0005971828941255808, 'timing/model': 1.6993092310149223, 'learning_rate/qwen_vl_interface': 9.089247457346774e-06, 'learning_rate/action_model': 9.089247457346773e-05, 'epoch': 0.42}) +[2026-06-23 18:12:47,957][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.2758031487464905, 'train/grad_norm_pre_clip': 70.91291046142578, 'timing/data': 0.0006948451045900583, 'timing/model': 1.703856857959181, 'learning_rate/qwen_vl_interface': 9.087453456756849e-06, 'learning_rate/action_model': 9.087453456756849e-05, 'epoch': 0.42}) +[2026-06-23 18:12:51,297][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.2458823025226593, 'train/grad_norm_pre_clip': 41.13259506225586, 'timing/data': 0.0006057280115783215, 'timing/model': 1.6944512729533017, 'learning_rate/qwen_vl_interface': 9.085657878734759e-06, 'learning_rate/action_model': 9.085657878734758e-05, 'epoch': 0.42}) +[2026-06-23 18:12:54,643][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.15928000211715698, 'train/grad_norm_pre_clip': 16.83504295349121, 'timing/data': 0.0006024721078574657, 'timing/model': 1.6970432049129158, 'learning_rate/qwen_vl_interface': 9.083860724018603e-06, 'learning_rate/action_model': 9.083860724018602e-05, 'epoch': 0.42}) +[2026-06-23 18:12:57,990][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.16648545861244202, 'train/grad_norm_pre_clip': 19.782943725585938, 'timing/data': 0.0005941670387983322, 'timing/model': 1.697304085129872, 'learning_rate/qwen_vl_interface': 9.08206199334712e-06, 'learning_rate/action_model': 9.082061993347119e-05, 'epoch': 0.42}) +[2026-06-23 18:13:01,340][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.41881364583969116, 'train/grad_norm_pre_clip': 139.13742065429688, 'timing/data': 0.0006220489740371704, 'timing/model': 1.7051795949228108, 'learning_rate/qwen_vl_interface': 9.0802616874597e-06, 'learning_rate/action_model': 9.0802616874597e-05, 'epoch': 0.42}) +[2026-06-23 18:13:04,681][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.29957592487335205, 'train/grad_norm_pre_clip': 106.6224594116211, 'timing/data': 0.000616980018094182, 'timing/model': 1.6964104669168591, 'learning_rate/qwen_vl_interface': 9.078459807096385e-06, 'learning_rate/action_model': 9.078459807096385e-05, 'epoch': 0.42}) +[2026-06-23 18:13:08,030][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.3296798765659332, 'train/grad_norm_pre_clip': 95.37274932861328, 'timing/data': 0.0005989058408886194, 'timing/model': 1.6978951280470937, 'learning_rate/qwen_vl_interface': 9.076656352997854e-06, 'learning_rate/action_model': 9.076656352997853e-05, 'epoch': 0.42}) +[2026-06-23 18:13:11,385][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.10888029634952545, 'train/grad_norm_pre_clip': 15.031476974487305, 'timing/data': 0.0005981340073049068, 'timing/model': 1.6991919551510364, 'learning_rate/qwen_vl_interface': 9.074851325905443e-06, 'learning_rate/action_model': 9.074851325905442e-05, 'epoch': 0.42}) +[2026-06-23 18:13:14,735][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.07656820118427277, 'train/grad_norm_pre_clip': 19.460634231567383, 'timing/data': 0.0006091950926929712, 'timing/model': 1.6999489380978048, 'learning_rate/qwen_vl_interface': 9.073044726561128e-06, 'learning_rate/action_model': 9.073044726561129e-05, 'epoch': 0.42}) +[2026-06-23 18:13:18,075][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.14095579087734222, 'train/grad_norm_pre_clip': 62.15625762939453, 'timing/data': 0.0006743480917066336, 'timing/model': 1.6923891920596361, 'learning_rate/qwen_vl_interface': 9.071236555707536e-06, 'learning_rate/action_model': 9.071236555707537e-05, 'epoch': 0.42}) +[2026-06-23 18:13:21,459][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.32766449451446533, 'train/grad_norm_pre_clip': 97.13726043701172, 'timing/data': 0.0006213269662111998, 'timing/model': 1.7032966502010822, 'learning_rate/qwen_vl_interface': 9.069426814087937e-06, 'learning_rate/action_model': 9.069426814087937e-05, 'epoch': 0.42}) +[2026-06-23 18:13:24,820][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.12049531191587448, 'train/grad_norm_pre_clip': 39.38274383544922, 'timing/data': 0.0006076919380575418, 'timing/model': 1.7031669770367444, 'learning_rate/qwen_vl_interface': 9.067615502446247e-06, 'learning_rate/action_model': 9.067615502446246e-05, 'epoch': 0.42}) +[2026-06-23 18:13:28,325][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.3462159037590027, 'train/grad_norm_pre_clip': 74.1498794555664, 'timing/data': 0.0006051070522516966, 'timing/model': 1.8564421979244798, 'learning_rate/qwen_vl_interface': 9.06580262152703e-06, 'learning_rate/action_model': 9.065802621527028e-05, 'epoch': 0.42}) +[2026-06-23 18:13:31,677][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.24704936146736145, 'train/grad_norm_pre_clip': 21.40284538269043, 'timing/data': 0.0005982241127640009, 'timing/model': 1.7012531200889498, 'learning_rate/qwen_vl_interface': 9.063988172075486e-06, 'learning_rate/action_model': 9.063988172075486e-05, 'epoch': 0.42}) +[2026-06-23 18:13:35,029][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.23408901691436768, 'train/grad_norm_pre_clip': 70.88882446289062, 'timing/data': 0.0006076619029045105, 'timing/model': 1.7012240060139447, 'learning_rate/qwen_vl_interface': 9.062172154837475e-06, 'learning_rate/action_model': 9.062172154837475e-05, 'epoch': 0.42}) +[2026-06-23 18:13:38,378][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.307046115398407, 'train/grad_norm_pre_clip': 77.77776336669922, 'timing/data': 0.0005958699621260166, 'timing/model': 1.698826501145959, 'learning_rate/qwen_vl_interface': 9.060354570559487e-06, 'learning_rate/action_model': 9.060354570559487e-05, 'epoch': 0.42}) +[2026-06-23 18:13:41,730][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.2648155093193054, 'train/grad_norm_pre_clip': 70.91481018066406, 'timing/data': 0.0005914908833801746, 'timing/model': 1.7046766160055995, 'learning_rate/qwen_vl_interface': 9.058535419988667e-06, 'learning_rate/action_model': 9.058535419988667e-05, 'epoch': 0.43}) +[2026-06-23 18:13:45,070][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1100 +[2026-06-23 18:15:18,674][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1100 +[2026-06-23 18:15:18,675][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1100 +[2026-06-23 18:16:29,760][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1100 +[2026-06-23 18:16:29,761][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.3046308755874634, 'train/grad_norm_pre_clip': 20.394268035888672, 'eval/action_loss/seconds': 93.60180282313377, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.11172904074192047, 'eval/latency_2/loss': 0.11172904074192047, 'eval/flappy/loss': 0.11172904074192047, 'eval/flappy/latency_2/loss': 0.11172904074192047, 'eval/flap_precision': 0.8422315536892622, 'eval/flap_recall': 0.9473684210526315, 'eval/flap_f1': 0.8917116544934899, 'eval/flap_rate_model': 0.23152777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8917116544934899, 'eval/latency_2/flap_precision': 0.8422315536892622, 'eval/latency_2/flap_recall': 0.9473684210526315, 'eval/latency_2/flap_f1': 0.8917116544934899, 'eval/latency_2/flap_rate_model': 0.23152777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8917116544934899, 'eval/action_classification/data_wait_seconds': 1.0139199495315552, 'eval/action_classification/predict_seconds': 69.92601013183594, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006764219142496586, 'timing/model': 1.6982207640539855, 'learning_rate/qwen_vl_interface': 9.056714703872794e-06, 'learning_rate/action_model': 9.056714703872794e-05, 'epoch': 0.43}) +[2026-06-23 18:16:33,079][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.17082473635673523, 'train/grad_norm_pre_clip': 35.42194747924805, 'timing/data': 0.000593886012211442, 'timing/model': 1.6851298711262643, 'learning_rate/qwen_vl_interface': 9.0548924229603e-06, 'learning_rate/action_model': 9.054892422960301e-05, 'epoch': 0.43}) +[2026-06-23 18:16:36,396][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.21202322840690613, 'train/grad_norm_pre_clip': 59.02579116821289, 'timing/data': 0.0005893069319427013, 'timing/model': 1.6844716139603406, 'learning_rate/qwen_vl_interface': 9.053068578000256e-06, 'learning_rate/action_model': 9.053068578000256e-05, 'epoch': 0.43}) +[2026-06-23 18:16:40,011][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.21638664603233337, 'train/grad_norm_pre_clip': 39.72100830078125, 'timing/data': 0.000601339852437377, 'timing/model': 1.9793935469351709, 'learning_rate/qwen_vl_interface': 9.051243169742372e-06, 'learning_rate/action_model': 9.051243169742372e-05, 'epoch': 0.43}) +[2026-06-23 18:16:43,326][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.12490400671958923, 'train/grad_norm_pre_clip': 45.0449333190918, 'timing/data': 0.0006036839913576841, 'timing/model': 1.6807736260816455, 'learning_rate/qwen_vl_interface': 9.049416198937008e-06, 'learning_rate/action_model': 9.049416198937007e-05, 'epoch': 0.43}) +[2026-06-23 18:16:46,651][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.07862058281898499, 'train/grad_norm_pre_clip': 12.01920223236084, 'timing/data': 0.0005982238799333572, 'timing/model': 1.6877546650357544, 'learning_rate/qwen_vl_interface': 9.047587666335159e-06, 'learning_rate/action_model': 9.047587666335159e-05, 'epoch': 0.43}) +[2026-06-23 18:16:49,980][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.14690294861793518, 'train/grad_norm_pre_clip': 28.221641540527344, 'timing/data': 0.0006053680554032326, 'timing/model': 1.689121323870495, 'learning_rate/qwen_vl_interface': 9.045757572688469e-06, 'learning_rate/action_model': 9.04575757268847e-05, 'epoch': 0.43}) +[2026-06-23 18:16:53,301][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.16233906149864197, 'train/grad_norm_pre_clip': 15.037567138671875, 'timing/data': 0.0006011901423335075, 'timing/model': 1.6849185759201646, 'learning_rate/qwen_vl_interface': 9.043925918749218e-06, 'learning_rate/action_model': 9.043925918749218e-05, 'epoch': 0.43}) +[2026-06-23 18:16:56,623][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.1682252734899521, 'train/grad_norm_pre_clip': 17.85198974609375, 'timing/data': 0.0005967209581285715, 'timing/model': 1.6852464319672436, 'learning_rate/qwen_vl_interface': 9.042092705270333e-06, 'learning_rate/action_model': 9.042092705270333e-05, 'epoch': 0.43}) +[2026-06-23 18:16:59,960][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.16770830750465393, 'train/grad_norm_pre_clip': 11.940325736999512, 'timing/data': 0.001052668085321784, 'timing/model': 1.7007033100817353, 'learning_rate/qwen_vl_interface': 9.040257933005375e-06, 'learning_rate/action_model': 9.040257933005374e-05, 'epoch': 0.43}) +[2026-06-23 18:17:03,296][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.11881629377603531, 'train/grad_norm_pre_clip': 40.498573303222656, 'timing/data': 0.0006584569346159697, 'timing/model': 1.6908529999200255, 'learning_rate/qwen_vl_interface': 9.038421602708552e-06, 'learning_rate/action_model': 9.038421602708552e-05, 'epoch': 0.43}) +[2026-06-23 18:17:06,628][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.16496030986309052, 'train/grad_norm_pre_clip': 20.1356143951416, 'timing/data': 0.0006722630932927132, 'timing/model': 1.6919085329864174, 'learning_rate/qwen_vl_interface': 9.036583715134709e-06, 'learning_rate/action_model': 9.036583715134708e-05, 'epoch': 0.43}) +[2026-06-23 18:17:09,960][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.39854171872138977, 'train/grad_norm_pre_clip': 21.21768569946289, 'timing/data': 0.0005905500147491693, 'timing/model': 1.691836629062891, 'learning_rate/qwen_vl_interface': 9.034744271039332e-06, 'learning_rate/action_model': 9.034744271039332e-05, 'epoch': 0.43}) +[2026-06-23 18:17:13,292][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.15323606133460999, 'train/grad_norm_pre_clip': 98.92082977294922, 'timing/data': 0.000613492913544178, 'timing/model': 1.6943327391054481, 'learning_rate/qwen_vl_interface': 9.03290327117855e-06, 'learning_rate/action_model': 9.032903271178549e-05, 'epoch': 0.43}) +[2026-06-23 18:17:16,629][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.2973698377609253, 'train/grad_norm_pre_clip': 81.33558654785156, 'timing/data': 0.0005862421821802855, 'timing/model': 1.6908181349281222, 'learning_rate/qwen_vl_interface': 9.031060716309122e-06, 'learning_rate/action_model': 9.031060716309121e-05, 'epoch': 0.43}) +[2026-06-23 18:17:19,963][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.25937825441360474, 'train/grad_norm_pre_clip': 93.72724151611328, 'timing/data': 0.0007163060363382101, 'timing/model': 1.6905965788755566, 'learning_rate/qwen_vl_interface': 9.02921660718846e-06, 'learning_rate/action_model': 9.029216607188459e-05, 'epoch': 0.43}) +[2026-06-23 18:17:23,294][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.14314106106758118, 'train/grad_norm_pre_clip': 27.826828002929688, 'timing/data': 0.0008359318599104881, 'timing/model': 1.6905284309759736, 'learning_rate/qwen_vl_interface': 9.027370944574604e-06, 'learning_rate/action_model': 9.027370944574604e-05, 'epoch': 0.43}) +[2026-06-23 18:17:26,632][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.1720234453678131, 'train/grad_norm_pre_clip': 35.773441314697266, 'timing/data': 0.0006029331125319004, 'timing/model': 1.696373184910044, 'learning_rate/qwen_vl_interface': 9.025523729226236e-06, 'learning_rate/action_model': 9.025523729226236e-05, 'epoch': 0.43}) +[2026-06-23 18:17:30,194][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.11867132782936096, 'train/grad_norm_pre_clip': 10.08015251159668, 'timing/data': 0.0006049070507287979, 'timing/model': 1.9196916371583939, 'learning_rate/qwen_vl_interface': 9.023674961902677e-06, 'learning_rate/action_model': 9.023674961902676e-05, 'epoch': 0.43}) +[2026-06-23 18:17:33,529][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.21105994284152985, 'train/grad_norm_pre_clip': 18.570270538330078, 'timing/data': 0.0006058379076421261, 'timing/model': 1.694297594949603, 'learning_rate/qwen_vl_interface': 9.021824643363883e-06, 'learning_rate/action_model': 9.021824643363883e-05, 'epoch': 0.43}) +[2026-06-23 18:17:36,862][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.14351557195186615, 'train/grad_norm_pre_clip': 21.890859603881836, 'timing/data': 0.0006784650031477213, 'timing/model': 1.6962097990326583, 'learning_rate/qwen_vl_interface': 9.019972774370456e-06, 'learning_rate/action_model': 9.019972774370455e-05, 'epoch': 0.43}) +[2026-06-23 18:17:40,201][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.25797638297080994, 'train/grad_norm_pre_clip': 17.441478729248047, 'timing/data': 0.0006019810680299997, 'timing/model': 1.6954486691392958, 'learning_rate/qwen_vl_interface': 9.018119355683623e-06, 'learning_rate/action_model': 9.018119355683623e-05, 'epoch': 0.43}) +[2026-06-23 18:17:43,543][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.17799267172813416, 'train/grad_norm_pre_clip': 21.706031799316406, 'timing/data': 0.0006291719619184732, 'timing/model': 1.6984555800445378, 'learning_rate/qwen_vl_interface': 9.016264388065258e-06, 'learning_rate/action_model': 9.016264388065258e-05, 'epoch': 0.43}) +[2026-06-23 18:17:46,883][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.30882972478866577, 'train/grad_norm_pre_clip': 51.568790435791016, 'timing/data': 0.0007023499347269535, 'timing/model': 1.6952421211171895, 'learning_rate/qwen_vl_interface': 9.014407872277869e-06, 'learning_rate/action_model': 9.014407872277869e-05, 'epoch': 0.43}) +[2026-06-23 18:17:50,216][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.1628076434135437, 'train/grad_norm_pre_clip': 70.63705444335938, 'timing/data': 0.0005883460398763418, 'timing/model': 1.6963786170817912, 'learning_rate/qwen_vl_interface': 9.012549809084595e-06, 'learning_rate/action_model': 9.012549809084595e-05, 'epoch': 0.44}) +[2026-06-23 18:17:53,554][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.12246176600456238, 'train/grad_norm_pre_clip': 14.608576774597168, 'timing/data': 0.0007712591905146837, 'timing/model': 1.6976175149902701, 'learning_rate/qwen_vl_interface': 9.01069019924922e-06, 'learning_rate/action_model': 9.010690199249219e-05, 'epoch': 0.44}) +[2026-06-23 18:17:56,883][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.19660909473896027, 'train/grad_norm_pre_clip': 36.75523376464844, 'timing/data': 0.0007648479659110308, 'timing/model': 1.695247741881758, 'learning_rate/qwen_vl_interface': 9.008829043536157e-06, 'learning_rate/action_model': 9.008829043536156e-05, 'epoch': 0.44}) +[2026-06-23 18:18:00,214][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.3136886954307556, 'train/grad_norm_pre_clip': 44.542781829833984, 'timing/data': 0.00060851383022964, 'timing/model': 1.6940005389042199, 'learning_rate/qwen_vl_interface': 9.006966342710458e-06, 'learning_rate/action_model': 9.006966342710457e-05, 'epoch': 0.44}) +[2026-06-23 18:18:03,547][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.09943973273038864, 'train/grad_norm_pre_clip': 19.25712776184082, 'timing/data': 0.0006141441408544779, 'timing/model': 1.688392821000889, 'learning_rate/qwen_vl_interface': 9.005102097537809e-06, 'learning_rate/action_model': 9.005102097537809e-05, 'epoch': 0.44}) +[2026-06-23 18:18:06,883][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.22037245333194733, 'train/grad_norm_pre_clip': 31.765668869018555, 'timing/data': 0.0006057380232959986, 'timing/model': 1.6942708268761635, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.44}) +[2026-06-23 18:18:10,223][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.19065961241722107, 'train/grad_norm_pre_clip': 16.805892944335938, 'timing/data': 0.0006043759640306234, 'timing/model': 1.6980577539652586, 'learning_rate/qwen_vl_interface': 9.001368977217577e-06, 'learning_rate/action_model': 9.001368977217577e-05, 'epoch': 0.44}) +[2026-06-23 18:18:13,564][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.1922609806060791, 'train/grad_norm_pre_clip': 29.365238189697266, 'timing/data': 0.000609175069257617, 'timing/model': 1.6949566770344973, 'learning_rate/qwen_vl_interface': 8.999500103604538e-06, 'learning_rate/action_model': 8.999500103604536e-05, 'epoch': 0.44}) +[2026-06-23 18:18:17,044][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.29099541902542114, 'train/grad_norm_pre_clip': 49.58649826049805, 'timing/data': 0.0005918429233133793, 'timing/model': 1.841286877868697, 'learning_rate/qwen_vl_interface': 8.997629688713639e-06, 'learning_rate/action_model': 8.997629688713638e-05, 'epoch': 0.44}) +[2026-06-23 18:18:20,378][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.1687587946653366, 'train/grad_norm_pre_clip': 34.49908447265625, 'timing/data': 0.0006007589399814606, 'timing/model': 1.6975184010807425, 'learning_rate/qwen_vl_interface': 8.995757733313733e-06, 'learning_rate/action_model': 8.995757733313732e-05, 'epoch': 0.44}) +[2026-06-23 18:18:23,715][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.21182650327682495, 'train/grad_norm_pre_clip': 15.07426929473877, 'timing/data': 0.0006045761983841658, 'timing/model': 1.6938795228488743, 'learning_rate/qwen_vl_interface': 8.993884238174314e-06, 'learning_rate/action_model': 8.993884238174314e-05, 'epoch': 0.44}) +[2026-06-23 18:18:27,041][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.19602176547050476, 'train/grad_norm_pre_clip': 45.704742431640625, 'timing/data': 0.0005871241446584463, 'timing/model': 1.6911846189759672, 'learning_rate/qwen_vl_interface': 8.992009204065505e-06, 'learning_rate/action_model': 8.992009204065505e-05, 'epoch': 0.44}) +[2026-06-23 18:18:30,377][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.26189902424812317, 'train/grad_norm_pre_clip': 14.683820724487305, 'timing/data': 0.0007895040325820446, 'timing/model': 1.694586382014677, 'learning_rate/qwen_vl_interface': 8.990132631758058e-06, 'learning_rate/action_model': 8.990132631758059e-05, 'epoch': 0.44}) +[2026-06-23 18:18:33,722][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.13334976136684418, 'train/grad_norm_pre_clip': 18.789499282836914, 'timing/data': 0.0008235878776758909, 'timing/model': 1.701864920090884, 'learning_rate/qwen_vl_interface': 8.988254522023364e-06, 'learning_rate/action_model': 8.988254522023363e-05, 'epoch': 0.44}) +[2026-06-23 18:18:37,064][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.2267242819070816, 'train/grad_norm_pre_clip': 23.008968353271484, 'timing/data': 0.0005977030377835035, 'timing/model': 1.7045552260242403, 'learning_rate/qwen_vl_interface': 8.986374875633441e-06, 'learning_rate/action_model': 8.98637487563344e-05, 'epoch': 0.44}) +[2026-06-23 18:18:40,401][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.34228894114494324, 'train/grad_norm_pre_clip': 105.1556625366211, 'timing/data': 0.0006166279781609774, 'timing/model': 1.6932110181078315, 'learning_rate/qwen_vl_interface': 8.984493693360944e-06, 'learning_rate/action_model': 8.984493693360943e-05, 'epoch': 0.44}) +[2026-06-23 18:18:43,746][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.24764302372932434, 'train/grad_norm_pre_clip': 73.53416442871094, 'timing/data': 0.0005949779879301786, 'timing/model': 1.6996676120907068, 'learning_rate/qwen_vl_interface': 8.982610975979153e-06, 'learning_rate/action_model': 8.982610975979152e-05, 'epoch': 0.44}) +[2026-06-23 18:18:47,073][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.35857918858528137, 'train/grad_norm_pre_clip': 118.10189819335938, 'timing/data': 0.0006056381389498711, 'timing/model': 1.688511786982417, 'learning_rate/qwen_vl_interface': 8.98072672426198e-06, 'learning_rate/action_model': 8.980726724261981e-05, 'epoch': 0.44}) +[2026-06-23 18:18:50,406][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.25219184160232544, 'train/grad_norm_pre_clip': 105.17052459716797, 'timing/data': 0.0006002380978316069, 'timing/model': 1.6900624511763453, 'learning_rate/qwen_vl_interface': 8.978840938983977e-06, 'learning_rate/action_model': 8.978840938983977e-05, 'epoch': 0.44}) +[2026-06-23 18:18:53,745][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.27127596735954285, 'train/grad_norm_pre_clip': 29.221843719482422, 'timing/data': 0.0006630159914493561, 'timing/model': 1.6930159130133688, 'learning_rate/qwen_vl_interface': 8.976953620920313e-06, 'learning_rate/action_model': 8.976953620920312e-05, 'epoch': 0.44}) +[2026-06-23 18:18:57,077][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.17604686319828033, 'train/grad_norm_pre_clip': 45.08445358276367, 'timing/data': 0.0005882349796593189, 'timing/model': 1.6957080820575356, 'learning_rate/qwen_vl_interface': 8.975064770846794e-06, 'learning_rate/action_model': 8.975064770846794e-05, 'epoch': 0.44}) +[2026-06-23 18:19:00,414][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.258947491645813, 'train/grad_norm_pre_clip': 84.2529296875, 'timing/data': 0.0006248240824788809, 'timing/model': 1.693385697901249, 'learning_rate/qwen_vl_interface': 8.973174389539856e-06, 'learning_rate/action_model': 8.973174389539855e-05, 'epoch': 0.44}) +[2026-06-23 18:19:03,743][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.2510196566581726, 'train/grad_norm_pre_clip': 78.75465393066406, 'timing/data': 0.0006597600877285004, 'timing/model': 1.6875788448378444, 'learning_rate/qwen_vl_interface': 8.971282477776563e-06, 'learning_rate/action_model': 8.971282477776563e-05, 'epoch': 0.44}) +[2026-06-23 18:19:07,208][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.22744140028953552, 'train/grad_norm_pre_clip': 52.40263366699219, 'timing/data': 0.0006713520269840956, 'timing/model': 1.691746728029102, 'learning_rate/qwen_vl_interface': 8.969389036334609e-06, 'learning_rate/action_model': 8.969389036334609e-05, 'epoch': 0.44}) +[2026-06-23 18:19:10,547][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.12515263259410858, 'train/grad_norm_pre_clip': 21.86553382873535, 'timing/data': 0.0006116400472819805, 'timing/model': 1.6942958789877594, 'learning_rate/qwen_vl_interface': 8.967494065992315e-06, 'learning_rate/action_model': 8.967494065992315e-05, 'epoch': 0.44}) +[2026-06-23 18:19:13,932][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.11158354580402374, 'train/grad_norm_pre_clip': 36.62691879272461, 'timing/data': 0.0006055778358131647, 'timing/model': 1.6917514170054346, 'learning_rate/qwen_vl_interface': 8.965597567528634e-06, 'learning_rate/action_model': 8.965597567528633e-05, 'epoch': 0.44}) +[2026-06-23 18:19:17,267][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.28909042477607727, 'train/grad_norm_pre_clip': 105.73878479003906, 'timing/data': 0.0006949359085410833, 'timing/model': 1.6887395181693137, 'learning_rate/qwen_vl_interface': 8.963699541723143e-06, 'learning_rate/action_model': 8.963699541723142e-05, 'epoch': 0.45}) +[2026-06-23 18:19:20,600][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.2698307931423187, 'train/grad_norm_pre_clip': 64.9996566772461, 'timing/data': 0.0006043061148375273, 'timing/model': 1.6913063309621066, 'learning_rate/qwen_vl_interface': 8.961799989356048e-06, 'learning_rate/action_model': 8.961799989356048e-05, 'epoch': 0.45}) +[2026-06-23 18:19:23,929][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.24474941194057465, 'train/grad_norm_pre_clip': 40.6744270324707, 'timing/data': 0.0006114088464528322, 'timing/model': 1.6919183819554746, 'learning_rate/qwen_vl_interface': 8.959898911208185e-06, 'learning_rate/action_model': 8.959898911208185e-05, 'epoch': 0.45}) +[2026-06-23 18:19:27,264][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.1648591160774231, 'train/grad_norm_pre_clip': 11.277294158935547, 'timing/data': 0.0006212170701473951, 'timing/model': 1.694618416018784, 'learning_rate/qwen_vl_interface': 8.957996308061015e-06, 'learning_rate/action_model': 8.957996308061014e-05, 'epoch': 0.45}) +[2026-06-23 18:19:30,595][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.173698291182518, 'train/grad_norm_pre_clip': 13.493762016296387, 'timing/data': 0.0005991058424115181, 'timing/model': 1.6865248268004507, 'learning_rate/qwen_vl_interface': 8.956092180696626e-06, 'learning_rate/action_model': 8.956092180696625e-05, 'epoch': 0.45}) +[2026-06-23 18:19:33,926][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.2841552495956421, 'train/grad_norm_pre_clip': 39.87632751464844, 'timing/data': 0.0006071911193430424, 'timing/model': 1.6911820080131292, 'learning_rate/qwen_vl_interface': 8.95418652989773e-06, 'learning_rate/action_model': 8.95418652989773e-05, 'epoch': 0.45}) +[2026-06-23 18:19:37,256][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.2421651929616928, 'train/grad_norm_pre_clip': 10.89932632446289, 'timing/data': 0.0006148258689790964, 'timing/model': 1.688756010029465, 'learning_rate/qwen_vl_interface': 8.952279356447675e-06, 'learning_rate/action_model': 8.952279356447675e-05, 'epoch': 0.45}) +[2026-06-23 18:19:40,596][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.20489570498466492, 'train/grad_norm_pre_clip': 30.61802864074707, 'timing/data': 0.0006062199827283621, 'timing/model': 1.6979924880433828, 'learning_rate/qwen_vl_interface': 8.950370661130421e-06, 'learning_rate/action_model': 8.950370661130422e-05, 'epoch': 0.45}) +[2026-06-23 18:19:43,939][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.12958331406116486, 'train/grad_norm_pre_clip': 25.19474220275879, 'timing/data': 0.0005924929864704609, 'timing/model': 1.6974979978986084, 'learning_rate/qwen_vl_interface': 8.948460444730566e-06, 'learning_rate/action_model': 8.948460444730566e-05, 'epoch': 0.45}) +[2026-06-23 18:19:47,273][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.155772402882576, 'train/grad_norm_pre_clip': 32.93374252319336, 'timing/data': 0.0007033520378172398, 'timing/model': 1.6917571199592203, 'learning_rate/qwen_vl_interface': 8.946548708033324e-06, 'learning_rate/action_model': 8.946548708033323e-05, 'epoch': 0.45}) +[2026-06-23 18:19:50,608][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.15446420013904572, 'train/grad_norm_pre_clip': 9.177031517028809, 'timing/data': 0.00065183499827981, 'timing/model': 1.6931020179763436, 'learning_rate/qwen_vl_interface': 8.944635451824538e-06, 'learning_rate/action_model': 8.944635451824538e-05, 'epoch': 0.45}) +[2026-06-23 18:19:53,954][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.21611063182353973, 'train/grad_norm_pre_clip': 42.096927642822266, 'timing/data': 0.0006627151742577553, 'timing/model': 1.6971620789263397, 'learning_rate/qwen_vl_interface': 8.942720676890677e-06, 'learning_rate/action_model': 8.942720676890677e-05, 'epoch': 0.45}) +[2026-06-23 18:19:57,571][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.21449871361255646, 'train/grad_norm_pre_clip': 23.716419219970703, 'timing/data': 0.0006037240382283926, 'timing/model': 1.6882579948287457, 'learning_rate/qwen_vl_interface': 8.940804384018833e-06, 'learning_rate/action_model': 8.940804384018832e-05, 'epoch': 0.45}) +[2026-06-23 18:20:00,909][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.2965646982192993, 'train/grad_norm_pre_clip': 12.968724250793457, 'timing/data': 0.0006036439444869757, 'timing/model': 1.6975696738809347, 'learning_rate/qwen_vl_interface': 8.93888657399672e-06, 'learning_rate/action_model': 8.93888657399672e-05, 'epoch': 0.45}) +[2026-06-23 18:20:04,253][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.23063333332538605, 'train/grad_norm_pre_clip': 16.03987693786621, 'timing/data': 0.0005952480714768171, 'timing/model': 1.6963309871498495, 'learning_rate/qwen_vl_interface': 8.936967247612678e-06, 'learning_rate/action_model': 8.936967247612678e-05, 'epoch': 0.45}) +[2026-06-23 18:20:07,601][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.13780692219734192, 'train/grad_norm_pre_clip': 42.17923355102539, 'timing/data': 0.0006438500713557005, 'timing/model': 1.6997550120577216, 'learning_rate/qwen_vl_interface': 8.93504640565567e-06, 'learning_rate/action_model': 8.93504640565567e-05, 'epoch': 0.45}) +[2026-06-23 18:20:10,948][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.1961945742368698, 'train/grad_norm_pre_clip': 34.15719985961914, 'timing/data': 0.0006732349283993244, 'timing/model': 1.700811668066308, 'learning_rate/qwen_vl_interface': 8.933124048915278e-06, 'learning_rate/action_model': 8.933124048915278e-05, 'epoch': 0.45}) +[2026-06-23 18:20:14,292][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.13808773458003998, 'train/grad_norm_pre_clip': 13.293587684631348, 'timing/data': 0.0005903490819036961, 'timing/model': 1.7018759679049253, 'learning_rate/qwen_vl_interface': 8.931200178181715e-06, 'learning_rate/action_model': 8.931200178181715e-05, 'epoch': 0.45}) +[2026-06-23 18:20:17,646][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.1767161786556244, 'train/grad_norm_pre_clip': 39.85264205932617, 'timing/data': 0.0006095550488680601, 'timing/model': 1.699658471159637, 'learning_rate/qwen_vl_interface': 8.929274794245812e-06, 'learning_rate/action_model': 8.929274794245812e-05, 'epoch': 0.45}) +[2026-06-23 18:20:20,995][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.2516704499721527, 'train/grad_norm_pre_clip': 16.748817443847656, 'timing/data': 0.0006049768999218941, 'timing/model': 1.7008198029361665, 'learning_rate/qwen_vl_interface': 8.927347897899018e-06, 'learning_rate/action_model': 8.927347897899016e-05, 'epoch': 0.45}) +[2026-06-23 18:20:24,347][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.16872334480285645, 'train/grad_norm_pre_clip': 33.45763397216797, 'timing/data': 0.0006085340864956379, 'timing/model': 1.7049626389052719, 'learning_rate/qwen_vl_interface': 8.925419489933407e-06, 'learning_rate/action_model': 8.925419489933408e-05, 'epoch': 0.45}) +[2026-06-23 18:20:27,696][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.22432547807693481, 'train/grad_norm_pre_clip': 30.232656478881836, 'timing/data': 0.000614415155723691, 'timing/model': 1.7029794619884342, 'learning_rate/qwen_vl_interface': 8.923489571141679e-06, 'learning_rate/action_model': 8.923489571141679e-05, 'epoch': 0.45}) +[2026-06-23 18:20:31,055][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.1994999498128891, 'train/grad_norm_pre_clip': 37.274356842041016, 'timing/data': 0.0005932049825787544, 'timing/model': 1.7043126379139721, 'learning_rate/qwen_vl_interface': 8.921558142317147e-06, 'learning_rate/action_model': 8.921558142317147e-05, 'epoch': 0.45}) +[2026-06-23 18:20:34,426][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.41204988956451416, 'train/grad_norm_pre_clip': 34.67645263671875, 'timing/data': 0.0006012700032442808, 'timing/model': 1.7136517490725964, 'learning_rate/qwen_vl_interface': 8.91962520425375e-06, 'learning_rate/action_model': 8.91962520425375e-05, 'epoch': 0.45}) +[2026-06-23 18:20:37,791][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.1645354926586151, 'train/grad_norm_pre_clip': 35.39255905151367, 'timing/data': 0.0006103769410401583, 'timing/model': 1.7064547939226031, 'learning_rate/qwen_vl_interface': 8.917690757746044e-06, 'learning_rate/action_model': 8.917690757746044e-05, 'epoch': 0.45}) +[2026-06-23 18:20:41,154][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.22124692797660828, 'train/grad_norm_pre_clip': 69.40530395507812, 'timing/data': 0.0006068199872970581, 'timing/model': 1.708088434068486, 'learning_rate/qwen_vl_interface': 8.915754803589208e-06, 'learning_rate/action_model': 8.915754803589209e-05, 'epoch': 0.45}) +[2026-06-23 18:20:44,593][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.22275277972221375, 'train/grad_norm_pre_clip': 42.101783752441406, 'timing/data': 0.000609725946560502, 'timing/model': 1.7022268089931458, 'learning_rate/qwen_vl_interface': 8.913817342579042e-06, 'learning_rate/action_model': 8.913817342579042e-05, 'epoch': 0.46}) +[2026-06-23 18:20:47,954][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.12453766167163849, 'train/grad_norm_pre_clip': 48.00631332397461, 'timing/data': 0.0006779748946428299, 'timing/model': 1.70776814292185, 'learning_rate/qwen_vl_interface': 8.91187837551196e-06, 'learning_rate/action_model': 8.911878375511961e-05, 'epoch': 0.46}) +[2026-06-23 18:20:51,311][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.22350740432739258, 'train/grad_norm_pre_clip': 15.451359748840332, 'timing/data': 0.0006024718750268221, 'timing/model': 1.7021217609290034, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.46}) +[2026-06-23 18:20:54,662][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.1612008661031723, 'train/grad_norm_pre_clip': 38.979488372802734, 'timing/data': 0.000607241177931428, 'timing/model': 1.7035596440546215, 'learning_rate/qwen_vl_interface': 8.907995926395814e-06, 'learning_rate/action_model': 8.907995926395815e-05, 'epoch': 0.46}) +[2026-06-23 18:20:58,023][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.11512553691864014, 'train/grad_norm_pre_clip': 27.52886199951172, 'timing/data': 0.0008421130478382111, 'timing/model': 1.7060205189045519, 'learning_rate/qwen_vl_interface': 8.90605244594268e-06, 'learning_rate/action_model': 8.906052445942679e-05, 'epoch': 0.46}) +[2026-06-23 18:21:01,383][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.2269323766231537, 'train/grad_norm_pre_clip': 25.430065155029297, 'timing/data': 0.0006025820039212704, 'timing/model': 1.7047717729583383, 'learning_rate/qwen_vl_interface': 8.904107462624484e-06, 'learning_rate/action_model': 8.904107462624483e-05, 'epoch': 0.46}) +[2026-06-23 18:21:04,743][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.24063514173030853, 'train/grad_norm_pre_clip': 54.023014068603516, 'timing/data': 0.0006158968899399042, 'timing/model': 1.7065619670320302, 'learning_rate/qwen_vl_interface': 8.902160977240739e-06, 'learning_rate/action_model': 8.902160977240739e-05, 'epoch': 0.46}) +[2026-06-23 18:21:08,084][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.13486412167549133, 'train/grad_norm_pre_clip': 24.755952835083008, 'timing/data': 0.0006108079105615616, 'timing/model': 1.698094064136967, 'learning_rate/qwen_vl_interface': 8.90021299059157e-06, 'learning_rate/action_model': 8.90021299059157e-05, 'epoch': 0.46}) +[2026-06-23 18:21:11,435][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.16664691269397736, 'train/grad_norm_pre_clip': 14.172304153442383, 'timing/data': 0.0005994669627398252, 'timing/model': 1.700329194078222, 'learning_rate/qwen_vl_interface': 8.898263503477722e-06, 'learning_rate/action_model': 8.898263503477721e-05, 'epoch': 0.46}) +[2026-06-23 18:21:14,789][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.11626012623310089, 'train/grad_norm_pre_clip': 13.015243530273438, 'timing/data': 0.0006067201029509306, 'timing/model': 1.7012209801468998, 'learning_rate/qwen_vl_interface': 8.896312516700553e-06, 'learning_rate/action_model': 8.896312516700554e-05, 'epoch': 0.46}) +[2026-06-23 18:21:18,144][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.22176487743854523, 'train/grad_norm_pre_clip': 54.8740348815918, 'timing/data': 0.0006131730042397976, 'timing/model': 1.7028580671176314, 'learning_rate/qwen_vl_interface': 8.894360031062043e-06, 'learning_rate/action_model': 8.894360031062043e-05, 'epoch': 0.46}) +[2026-06-23 18:21:21,485][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.18616458773612976, 'train/grad_norm_pre_clip': 34.353755950927734, 'timing/data': 0.0006159069016575813, 'timing/model': 1.6982387360185385, 'learning_rate/qwen_vl_interface': 8.892406047364783e-06, 'learning_rate/action_model': 8.892406047364784e-05, 'epoch': 0.46}) +[2026-06-23 18:21:24,829][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.20634718239307404, 'train/grad_norm_pre_clip': 44.283748626708984, 'timing/data': 0.0006042458117008209, 'timing/model': 1.6984067319426686, 'learning_rate/qwen_vl_interface': 8.890450566411984e-06, 'learning_rate/action_model': 8.890450566411984e-05, 'epoch': 0.46}) +[2026-06-23 18:21:28,181][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.18808521330356598, 'train/grad_norm_pre_clip': 61.794620513916016, 'timing/data': 0.0008400790393352509, 'timing/model': 1.6996868879068643, 'learning_rate/qwen_vl_interface': 8.888493589007467e-06, 'learning_rate/action_model': 8.888493589007468e-05, 'epoch': 0.46}) +[2026-06-23 18:21:31,533][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.21105431020259857, 'train/grad_norm_pre_clip': 44.76942825317383, 'timing/data': 0.000706337857991457, 'timing/model': 1.7013024650514126, 'learning_rate/qwen_vl_interface': 8.886535115955676e-06, 'learning_rate/action_model': 8.886535115955675e-05, 'epoch': 0.46}) +[2026-06-23 18:21:35,175][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.13737303018569946, 'train/grad_norm_pre_clip': 18.015396118164062, 'timing/data': 0.0006113590206950903, 'timing/model': 1.7016214940231293, 'learning_rate/qwen_vl_interface': 8.884575148061662e-06, 'learning_rate/action_model': 8.88457514806166e-05, 'epoch': 0.46}) +[2026-06-23 18:21:38,513][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.22360782325267792, 'train/grad_norm_pre_clip': 58.40961456298828, 'timing/data': 0.0006000869907438755, 'timing/model': 1.6939539550803602, 'learning_rate/qwen_vl_interface': 8.882613686131095e-06, 'learning_rate/action_model': 8.882613686131093e-05, 'epoch': 0.46}) +[2026-06-23 18:21:41,855][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.2304113656282425, 'train/grad_norm_pre_clip': 75.2614974975586, 'timing/data': 0.0006013601087033749, 'timing/model': 1.6931739288847893, 'learning_rate/qwen_vl_interface': 8.880650730970254e-06, 'learning_rate/action_model': 8.880650730970254e-05, 'epoch': 0.46}) +[2026-06-23 18:21:45,210][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.18951144814491272, 'train/grad_norm_pre_clip': 54.470088958740234, 'timing/data': 0.0008160939905792475, 'timing/model': 1.70074347499758, 'learning_rate/qwen_vl_interface': 8.878686283386042e-06, 'learning_rate/action_model': 8.878686283386042e-05, 'epoch': 0.46}) +[2026-06-23 18:21:48,553][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.12546920776367188, 'train/grad_norm_pre_clip': 10.139461517333984, 'timing/data': 0.0006625358946621418, 'timing/model': 1.6969211020041257, 'learning_rate/qwen_vl_interface': 8.876720344185967e-06, 'learning_rate/action_model': 8.876720344185967e-05, 'epoch': 0.46}) +[2026-06-23 18:21:51,902][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.28421130776405334, 'train/grad_norm_pre_clip': 49.58763122558594, 'timing/data': 0.000609806040301919, 'timing/model': 1.7021488861646503, 'learning_rate/qwen_vl_interface': 8.874752914178149e-06, 'learning_rate/action_model': 8.87475291417815e-05, 'epoch': 0.46}) +[2026-06-23 18:21:55,255][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.16015471518039703, 'train/grad_norm_pre_clip': 20.826143264770508, 'timing/data': 0.000604966888204217, 'timing/model': 1.699062957195565, 'learning_rate/qwen_vl_interface': 8.87278399417133e-06, 'learning_rate/action_model': 8.872783994171329e-05, 'epoch': 0.46}) +[2026-06-23 18:21:58,608][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.1724875420331955, 'train/grad_norm_pre_clip': 14.813721656799316, 'timing/data': 0.0006084740161895752, 'timing/model': 1.702101106988266, 'learning_rate/qwen_vl_interface': 8.870813584974852e-06, 'learning_rate/action_model': 8.870813584974852e-05, 'epoch': 0.46}) +[2026-06-23 18:22:01,959][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.23591656982898712, 'train/grad_norm_pre_clip': 67.62328338623047, 'timing/data': 0.0006081729661673307, 'timing/model': 1.6997793449554592, 'learning_rate/qwen_vl_interface': 8.868841687398683e-06, 'learning_rate/action_model': 8.868841687398683e-05, 'epoch': 0.46}) +[2026-06-23 18:22:05,300][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1200 +[2026-06-23 18:23:37,885][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1200 +[2026-06-23 18:23:37,886][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1200 +[2026-06-23 18:24:49,213][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1200 +[2026-06-23 18:24:49,214][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.26857730746269226, 'train/grad_norm_pre_clip': 17.297611236572266, 'eval/action_loss/seconds': 92.58221142599359, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.09471018612384796, 'eval/latency_2/loss': 0.09471018612384796, 'eval/flappy/loss': 0.09471018612384796, 'eval/flappy/latency_2/loss': 0.09471018612384796, 'eval/flap_precision': 0.8787684413085312, 'eval/flap_recall': 0.9244264507422402, 'eval/flap_f1': 0.9010194015126602, 'eval/flap_rate_model': 0.2165277777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9010194015126602, 'eval/latency_2/flap_precision': 0.8787684413085312, 'eval/latency_2/flap_recall': 0.9244264507422402, 'eval/latency_2/flap_f1': 0.9010194015126602, 'eval/latency_2/flap_rate_model': 0.2165277777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9010194015126602, 'eval/action_classification/data_wait_seconds': 0.8086502552032471, 'eval/action_classification/predict_seconds': 70.37298583984375, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005964708980172873, 'timing/model': 1.69465761189349, 'learning_rate/qwen_vl_interface': 8.86686830225339e-06, 'learning_rate/action_model': 8.866868302253389e-05, 'epoch': 0.46}) +[2026-06-23 18:24:52,547][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.18794183433055878, 'train/grad_norm_pre_clip': 24.065298080444336, 'timing/data': 0.0006202049553394318, 'timing/model': 1.6951731650624424, 'learning_rate/qwen_vl_interface': 8.864893430350159e-06, 'learning_rate/action_model': 8.864893430350159e-05, 'epoch': 0.46}) +[2026-06-23 18:24:55,871][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.16559118032455444, 'train/grad_norm_pre_clip': 12.35129451751709, 'timing/data': 0.0006046860944479704, 'timing/model': 1.6892099380493164, 'learning_rate/qwen_vl_interface': 8.862917072500785e-06, 'learning_rate/action_model': 8.862917072500784e-05, 'epoch': 0.47}) +[2026-06-23 18:24:59,193][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.11456023901700974, 'train/grad_norm_pre_clip': 33.14202880859375, 'timing/data': 0.0006253449246287346, 'timing/model': 1.6930039080325514, 'learning_rate/qwen_vl_interface': 8.860939229517674e-06, 'learning_rate/action_model': 8.860939229517674e-05, 'epoch': 0.47}) +[2026-06-23 18:25:02,512][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.1300540566444397, 'train/grad_norm_pre_clip': 39.30165100097656, 'timing/data': 0.0006190929561853409, 'timing/model': 1.6849716030992568, 'learning_rate/qwen_vl_interface': 8.858959902213844e-06, 'learning_rate/action_model': 8.858959902213845e-05, 'epoch': 0.47}) +[2026-06-23 18:25:05,836][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.1311422437429428, 'train/grad_norm_pre_clip': 16.31360626220703, 'timing/data': 0.0006170999258756638, 'timing/model': 1.6891524600796402, 'learning_rate/qwen_vl_interface': 8.856979091402921e-06, 'learning_rate/action_model': 8.85697909140292e-05, 'epoch': 0.47}) +[2026-06-23 18:25:09,168][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.2408561408519745, 'train/grad_norm_pre_clip': 35.41147994995117, 'timing/data': 0.0006055079866200686, 'timing/model': 1.6940207390580326, 'learning_rate/qwen_vl_interface': 8.854996797899139e-06, 'learning_rate/action_model': 8.854996797899139e-05, 'epoch': 0.47}) +[2026-06-23 18:25:12,492][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.23916974663734436, 'train/grad_norm_pre_clip': 19.761125564575195, 'timing/data': 0.0006092449184507132, 'timing/model': 1.6884245921391994, 'learning_rate/qwen_vl_interface': 8.853013022517348e-06, 'learning_rate/action_model': 8.853013022517348e-05, 'epoch': 0.47}) +[2026-06-23 18:25:15,813][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.1389513611793518, 'train/grad_norm_pre_clip': 37.51153564453125, 'timing/data': 0.0007132701575756073, 'timing/model': 1.684491230873391, 'learning_rate/qwen_vl_interface': 8.851027766073002e-06, 'learning_rate/action_model': 8.851027766073001e-05, 'epoch': 0.47}) +[2026-06-23 18:25:19,141][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.17509961128234863, 'train/grad_norm_pre_clip': 47.52159881591797, 'timing/data': 0.0006025428883731365, 'timing/model': 1.6924971060361713, 'learning_rate/qwen_vl_interface': 8.84904102938216e-06, 'learning_rate/action_model': 8.84904102938216e-05, 'epoch': 0.47}) +[2026-06-23 18:25:22,465][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.13260528445243835, 'train/grad_norm_pre_clip': 17.313989639282227, 'timing/data': 0.0006944348569959402, 'timing/model': 1.6851894040592015, 'learning_rate/qwen_vl_interface': 8.8470528132615e-06, 'learning_rate/action_model': 8.8470528132615e-05, 'epoch': 0.47}) +[2026-06-23 18:25:25,785][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.19712801277637482, 'train/grad_norm_pre_clip': 24.555143356323242, 'timing/data': 0.0006203658413141966, 'timing/model': 1.6850678748451173, 'learning_rate/qwen_vl_interface': 8.845063118528303e-06, 'learning_rate/action_model': 8.845063118528304e-05, 'epoch': 0.47}) +[2026-06-23 18:25:29,111][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.2421155571937561, 'train/grad_norm_pre_clip': 18.080820083618164, 'timing/data': 0.0006008290220052004, 'timing/model': 1.688160747056827, 'learning_rate/qwen_vl_interface': 8.843071946000455e-06, 'learning_rate/action_model': 8.843071946000456e-05, 'epoch': 0.47}) +[2026-06-23 18:25:32,447][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.1362208127975464, 'train/grad_norm_pre_clip': 74.98580932617188, 'timing/data': 0.0007047648541629314, 'timing/model': 1.6935313201975077, 'learning_rate/qwen_vl_interface': 8.841079296496451e-06, 'learning_rate/action_model': 8.84107929649645e-05, 'epoch': 0.47}) +[2026-06-23 18:25:35,871][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.4710841178894043, 'train/grad_norm_pre_clip': 71.15061950683594, 'timing/data': 0.0007930400315672159, 'timing/model': 1.692022274946794, 'learning_rate/qwen_vl_interface': 8.839085170835396e-06, 'learning_rate/action_model': 8.839085170835395e-05, 'epoch': 0.47}) +[2026-06-23 18:25:39,194][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.20216470956802368, 'train/grad_norm_pre_clip': 52.38676452636719, 'timing/data': 0.0006102169863879681, 'timing/model': 1.685105917043984, 'learning_rate/qwen_vl_interface': 8.837089569837e-06, 'learning_rate/action_model': 8.837089569836998e-05, 'epoch': 0.47}) +[2026-06-23 18:25:42,523][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.20669908821582794, 'train/grad_norm_pre_clip': 23.196828842163086, 'timing/data': 0.0007664100266993046, 'timing/model': 1.6876572109758854, 'learning_rate/qwen_vl_interface': 8.835092494321575e-06, 'learning_rate/action_model': 8.835092494321575e-05, 'epoch': 0.47}) +[2026-06-23 18:25:45,858][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.2346753031015396, 'train/grad_norm_pre_clip': 46.70823669433594, 'timing/data': 0.000624402891844511, 'timing/model': 1.6946426189970225, 'learning_rate/qwen_vl_interface': 8.833093945110048e-06, 'learning_rate/action_model': 8.833093945110047e-05, 'epoch': 0.47}) +[2026-06-23 18:25:49,184][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.18436561524868011, 'train/grad_norm_pre_clip': 69.48001861572266, 'timing/data': 0.0005998869892209768, 'timing/model': 1.6927614230662584, 'learning_rate/qwen_vl_interface': 8.831093923023946e-06, 'learning_rate/action_model': 8.831093923023945e-05, 'epoch': 0.47}) +[2026-06-23 18:25:52,517][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.15660472214221954, 'train/grad_norm_pre_clip': 38.7891845703125, 'timing/data': 0.000802747905254364, 'timing/model': 1.6850560430902988, 'learning_rate/qwen_vl_interface': 8.8290924288854e-06, 'learning_rate/action_model': 8.8290924288854e-05, 'epoch': 0.47}) +[2026-06-23 18:25:55,844][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.11781226843595505, 'train/grad_norm_pre_clip': 20.494117736816406, 'timing/data': 0.0006193241570144892, 'timing/model': 1.6900110449641943, 'learning_rate/qwen_vl_interface': 8.827089463517153e-06, 'learning_rate/action_model': 8.827089463517153e-05, 'epoch': 0.47}) +[2026-06-23 18:25:59,180][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.20941638946533203, 'train/grad_norm_pre_clip': 24.36775779724121, 'timing/data': 0.0006213078740984201, 'timing/model': 1.694997096201405, 'learning_rate/qwen_vl_interface': 8.825085027742545e-06, 'learning_rate/action_model': 8.825085027742544e-05, 'epoch': 0.47}) +[2026-06-23 18:26:02,521][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.273171067237854, 'train/grad_norm_pre_clip': 22.77899742126465, 'timing/data': 0.0006175900343805552, 'timing/model': 1.6941882250830531, 'learning_rate/qwen_vl_interface': 8.823079122385523e-06, 'learning_rate/action_model': 8.823079122385523e-05, 'epoch': 0.47}) +[2026-06-23 18:26:05,863][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.11030586808919907, 'train/grad_norm_pre_clip': 10.275677680969238, 'timing/data': 0.0006140030454844236, 'timing/model': 1.698035083943978, 'learning_rate/qwen_vl_interface': 8.821071748270644e-06, 'learning_rate/action_model': 8.821071748270644e-05, 'epoch': 0.47}) +[2026-06-23 18:26:09,196][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.1511385589838028, 'train/grad_norm_pre_clip': 10.919818878173828, 'timing/data': 0.0005983239971101284, 'timing/model': 1.6943960960488766, 'learning_rate/qwen_vl_interface': 8.81906290622306e-06, 'learning_rate/action_model': 8.81906290622306e-05, 'epoch': 0.47}) +[2026-06-23 18:26:12,528][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.2739621102809906, 'train/grad_norm_pre_clip': 26.65003776550293, 'timing/data': 0.0006199351046234369, 'timing/model': 1.6895633940584958, 'learning_rate/qwen_vl_interface': 8.81705259706853e-06, 'learning_rate/action_model': 8.817052597068529e-05, 'epoch': 0.47}) +[2026-06-23 18:26:15,867][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.07231325656175613, 'train/grad_norm_pre_clip': 11.504427909851074, 'timing/data': 0.0005959700793027878, 'timing/model': 1.6975221410393715, 'learning_rate/qwen_vl_interface': 8.815040821633416e-06, 'learning_rate/action_model': 8.815040821633416e-05, 'epoch': 0.47}) +[2026-06-23 18:26:19,201][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.0644276961684227, 'train/grad_norm_pre_clip': 31.03984260559082, 'timing/data': 0.000605317996814847, 'timing/model': 1.6970371301285923, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.47}) +[2026-06-23 18:26:22,619][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.24237360060214996, 'train/grad_norm_pre_clip': 81.26124572753906, 'timing/data': 0.0006081731989979744, 'timing/model': 1.7761252778582275, 'learning_rate/qwen_vl_interface': 8.8110128752299e-06, 'learning_rate/action_model': 8.8110128752299e-05, 'epoch': 0.48}) +[2026-06-23 18:26:25,963][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.16958144307136536, 'train/grad_norm_pre_clip': 35.23102951049805, 'timing/data': 0.0006153869908303022, 'timing/model': 1.6987805059179664, 'learning_rate/qwen_vl_interface': 8.808996705917235e-06, 'learning_rate/action_model': 8.808996705917235e-05, 'epoch': 0.48}) +[2026-06-23 18:26:29,296][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.22701656818389893, 'train/grad_norm_pre_clip': 24.809497833251953, 'timing/data': 0.000616719014942646, 'timing/model': 1.6869935248978436, 'learning_rate/qwen_vl_interface': 8.80697907363546e-06, 'learning_rate/action_model': 8.80697907363546e-05, 'epoch': 0.48}) +[2026-06-23 18:26:32,627][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.2644055187702179, 'train/grad_norm_pre_clip': 15.668859481811523, 'timing/data': 0.000604746863245964, 'timing/model': 1.689463728107512, 'learning_rate/qwen_vl_interface': 8.804959979213945e-06, 'learning_rate/action_model': 8.804959979213945e-05, 'epoch': 0.48}) +[2026-06-23 18:26:35,948][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.22342558205127716, 'train/grad_norm_pre_clip': 40.41975021362305, 'timing/data': 0.000679075950756669, 'timing/model': 1.6883211198728532, 'learning_rate/qwen_vl_interface': 8.802939423482667e-06, 'learning_rate/action_model': 8.802939423482666e-05, 'epoch': 0.48}) +[2026-06-23 18:26:39,271][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.14433863759040833, 'train/grad_norm_pre_clip': 69.18836975097656, 'timing/data': 0.0008297201711684465, 'timing/model': 1.68929178500548, 'learning_rate/qwen_vl_interface': 8.800917407272198e-06, 'learning_rate/action_model': 8.800917407272198e-05, 'epoch': 0.48}) +[2026-06-23 18:26:42,597][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.09590476006269455, 'train/grad_norm_pre_clip': 26.564306259155273, 'timing/data': 0.000905181048437953, 'timing/model': 1.6844880178105086, 'learning_rate/qwen_vl_interface': 8.798893931413711e-06, 'learning_rate/action_model': 8.79889393141371e-05, 'epoch': 0.48}) +[2026-06-23 18:26:45,929][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.30420947074890137, 'train/grad_norm_pre_clip': 78.27346801757812, 'timing/data': 0.0009172728750854731, 'timing/model': 1.6919859577901661, 'learning_rate/qwen_vl_interface': 8.796868996738985e-06, 'learning_rate/action_model': 8.796868996738984e-05, 'epoch': 0.48}) +[2026-06-23 18:26:49,261][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.09102730453014374, 'train/grad_norm_pre_clip': 29.75697135925293, 'timing/data': 0.0005946268793195486, 'timing/model': 1.6892441450618207, 'learning_rate/qwen_vl_interface': 8.794842604080389e-06, 'learning_rate/action_model': 8.794842604080389e-05, 'epoch': 0.48}) +[2026-06-23 18:26:52,604][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.13593095541000366, 'train/grad_norm_pre_clip': 37.60404586791992, 'timing/data': 0.0006821220740675926, 'timing/model': 1.7003572799731046, 'learning_rate/qwen_vl_interface': 8.7928147542709e-06, 'learning_rate/action_model': 8.7928147542709e-05, 'epoch': 0.48}) +[2026-06-23 18:26:55,938][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.1801699846982956, 'train/grad_norm_pre_clip': 21.00803565979004, 'timing/data': 0.0006037151906639338, 'timing/model': 1.6920518528204411, 'learning_rate/qwen_vl_interface': 8.790785448144088e-06, 'learning_rate/action_model': 8.790785448144088e-05, 'epoch': 0.48}) +[2026-06-23 18:26:59,270][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.21283024549484253, 'train/grad_norm_pre_clip': 29.8482608795166, 'timing/data': 0.0006086139474064112, 'timing/model': 1.692779980134219, 'learning_rate/qwen_vl_interface': 8.788754686534126e-06, 'learning_rate/action_model': 8.788754686534127e-05, 'epoch': 0.48}) +[2026-06-23 18:27:02,601][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.1834143102169037, 'train/grad_norm_pre_clip': 66.5860824584961, 'timing/data': 0.0006123809143900871, 'timing/model': 1.6972526160534471, 'learning_rate/qwen_vl_interface': 8.786722470275784e-06, 'learning_rate/action_model': 8.786722470275784e-05, 'epoch': 0.48}) +[2026-06-23 18:27:05,927][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.2836948335170746, 'train/grad_norm_pre_clip': 73.88521575927734, 'timing/data': 0.0007014109287410975, 'timing/model': 1.6864947008434683, 'learning_rate/qwen_vl_interface': 8.784688800204427e-06, 'learning_rate/action_model': 8.784688800204426e-05, 'epoch': 0.48}) +[2026-06-23 18:27:09,259][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.1753019392490387, 'train/grad_norm_pre_clip': 32.26492691040039, 'timing/data': 0.0006881069857627153, 'timing/model': 1.6914954660460353, 'learning_rate/qwen_vl_interface': 8.782653677156022e-06, 'learning_rate/action_model': 8.782653677156022e-05, 'epoch': 0.48}) +[2026-06-23 18:27:12,675][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.0867580696940422, 'train/grad_norm_pre_clip': 29.77113151550293, 'timing/data': 0.0006071138195693493, 'timing/model': 1.6924177869223058, 'learning_rate/qwen_vl_interface': 8.780617101967129e-06, 'learning_rate/action_model': 8.780617101967129e-05, 'epoch': 0.48}) +[2026-06-23 18:27:16,012][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.2129812240600586, 'train/grad_norm_pre_clip': 17.498430252075195, 'timing/data': 0.0005979461129754782, 'timing/model': 1.6977492060977966, 'learning_rate/qwen_vl_interface': 8.77857907547491e-06, 'learning_rate/action_model': 8.77857907547491e-05, 'epoch': 0.48}) +[2026-06-23 18:27:19,348][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.1274583339691162, 'train/grad_norm_pre_clip': 10.649163246154785, 'timing/data': 0.0007127320859581232, 'timing/model': 1.6886746759992093, 'learning_rate/qwen_vl_interface': 8.776539598517119e-06, 'learning_rate/action_model': 8.776539598517118e-05, 'epoch': 0.48}) +[2026-06-23 18:27:22,691][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.150921031832695, 'train/grad_norm_pre_clip': 11.073580741882324, 'timing/data': 0.0006626378744840622, 'timing/model': 1.6986053679138422, 'learning_rate/qwen_vl_interface': 8.774498671932108e-06, 'learning_rate/action_model': 8.774498671932107e-05, 'epoch': 0.48}) +[2026-06-23 18:27:26,028][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.11630532145500183, 'train/grad_norm_pre_clip': 17.32577896118164, 'timing/data': 0.0005990879144519567, 'timing/model': 1.694636387983337, 'learning_rate/qwen_vl_interface': 8.772456296558824e-06, 'learning_rate/action_model': 8.772456296558824e-05, 'epoch': 0.48}) +[2026-06-23 18:27:29,373][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.12877404689788818, 'train/grad_norm_pre_clip': 15.108266830444336, 'timing/data': 0.0005954219959676266, 'timing/model': 1.6975687169469893, 'learning_rate/qwen_vl_interface': 8.77041247323681e-06, 'learning_rate/action_model': 8.770412473236811e-05, 'epoch': 0.48}) +[2026-06-23 18:27:32,708][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.10094393789768219, 'train/grad_norm_pre_clip': 34.70924758911133, 'timing/data': 0.00061180186457932, 'timing/model': 1.690846238983795, 'learning_rate/qwen_vl_interface': 8.76836720280621e-06, 'learning_rate/action_model': 8.768367202806209e-05, 'epoch': 0.48}) +[2026-06-23 18:28:09,379][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1250 +[2026-06-23 18:28:09,380][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.20566464960575104, 'train/grad_norm_pre_clip': 28.258935928344727, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 23.40000000000013, 'rl_games_eval/mid_train/mean_length': 207.2, 'rl_games_eval/mid_train/std_reward': 20.32486162314536, 'rl_games_eval/mid_train/std_length': 164.06023284147807, 'rl_games_eval/mid_train/macro_mean_reward': 23.40000000000013, 'rl_games_eval/mid_train/macro_mean_length': 207.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 23.40000000000013, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 207.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 20.32486162314536, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 164.06023284147807, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 23.40000000000013, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006626180838793516, 'timing/model': 1.694537791190669, 'learning_rate/qwen_vl_interface': 8.766320486107749e-06, 'learning_rate/action_model': 8.766320486107749e-05, 'epoch': 0.48}) +[2026-06-23 18:28:12,668][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.09637722373008728, 'train/grad_norm_pre_clip': 64.18651580810547, 'timing/data': 0.0005887490697205067, 'timing/model': 1.6740427070762962, 'learning_rate/qwen_vl_interface': 8.76427232398276e-06, 'learning_rate/action_model': 8.76427232398276e-05, 'epoch': 0.48}) +[2026-06-23 18:28:15,972][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.1783657819032669, 'train/grad_norm_pre_clip': 38.5776481628418, 'timing/data': 0.0006093080155551434, 'timing/model': 1.6765876119025052, 'learning_rate/qwen_vl_interface': 8.762222717273167e-06, 'learning_rate/action_model': 8.762222717273167e-05, 'epoch': 0.48}) +[2026-06-23 18:28:19,281][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.1275675892829895, 'train/grad_norm_pre_clip': 22.85401725769043, 'timing/data': 0.0006194370798766613, 'timing/model': 1.6793405979406089, 'learning_rate/qwen_vl_interface': 8.76017166682148e-06, 'learning_rate/action_model': 8.760171666821481e-05, 'epoch': 0.48}) +[2026-06-23 18:28:22,593][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.1614370048046112, 'train/grad_norm_pre_clip': 17.035715103149414, 'timing/data': 0.0005990678910166025, 'timing/model': 1.6859477520920336, 'learning_rate/qwen_vl_interface': 8.758119173470815e-06, 'learning_rate/action_model': 8.758119173470815e-05, 'epoch': 0.49}) +[2026-06-23 18:28:25,914][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.20718905329704285, 'train/grad_norm_pre_clip': 34.31474304199219, 'timing/data': 0.0006169120315462351, 'timing/model': 1.6882013778667897, 'learning_rate/qwen_vl_interface': 8.75606523806487e-06, 'learning_rate/action_model': 8.756065238064868e-05, 'epoch': 0.49}) +[2026-06-23 18:28:29,353][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.19344425201416016, 'train/grad_norm_pre_clip': 17.136899948120117, 'timing/data': 0.0009703980758786201, 'timing/model': 1.6772551818285137, 'learning_rate/qwen_vl_interface': 8.75400986144794e-06, 'learning_rate/action_model': 8.75400986144794e-05, 'epoch': 0.49}) +[2026-06-23 18:28:32,674][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.12670722603797913, 'train/grad_norm_pre_clip': 47.58987045288086, 'timing/data': 0.0006174531299620867, 'timing/model': 1.6853074980899692, 'learning_rate/qwen_vl_interface': 8.751953044464913e-06, 'learning_rate/action_model': 8.751953044464913e-05, 'epoch': 0.49}) +[2026-06-23 18:28:35,995][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.2523939907550812, 'train/grad_norm_pre_clip': 18.820356369018555, 'timing/data': 0.0007008898537606001, 'timing/model': 1.6865575050469488, 'learning_rate/qwen_vl_interface': 8.749894787961272e-06, 'learning_rate/action_model': 8.749894787961271e-05, 'epoch': 0.49}) +[2026-06-23 18:28:39,308][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.14592835307121277, 'train/grad_norm_pre_clip': 23.45819664001465, 'timing/data': 0.0006075841374695301, 'timing/model': 1.6872135309968144, 'learning_rate/qwen_vl_interface': 8.747835092783086e-06, 'learning_rate/action_model': 8.747835092783085e-05, 'epoch': 0.49}) +[2026-06-23 18:28:42,632][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.08376382291316986, 'train/grad_norm_pre_clip': 20.1951961517334, 'timing/data': 0.0007752999663352966, 'timing/model': 1.685976062901318, 'learning_rate/qwen_vl_interface': 8.745773959777018e-06, 'learning_rate/action_model': 8.745773959777017e-05, 'epoch': 0.49}) +[2026-06-23 18:28:45,959][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.16254931688308716, 'train/grad_norm_pre_clip': 76.0770034790039, 'timing/data': 0.00061215297318995, 'timing/model': 1.692801009165123, 'learning_rate/qwen_vl_interface': 8.743711389790321e-06, 'learning_rate/action_model': 8.74371138979032e-05, 'epoch': 0.49}) +[2026-06-23 18:28:49,292][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.3837455213069916, 'train/grad_norm_pre_clip': 130.88453674316406, 'timing/data': 0.0006239749491214752, 'timing/model': 1.6942970589734614, 'learning_rate/qwen_vl_interface': 8.741647383670839e-06, 'learning_rate/action_model': 8.741647383670839e-05, 'epoch': 0.49}) +[2026-06-23 18:28:52,623][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.3288218677043915, 'train/grad_norm_pre_clip': 40.445308685302734, 'timing/data': 0.0006025049369782209, 'timing/model': 1.692547305021435, 'learning_rate/qwen_vl_interface': 8.739581942267011e-06, 'learning_rate/action_model': 8.739581942267011e-05, 'epoch': 0.49}) +[2026-06-23 18:28:55,965][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.1773306280374527, 'train/grad_norm_pre_clip': 31.851470947265625, 'timing/data': 0.000626410124823451, 'timing/model': 1.6967093970160931, 'learning_rate/qwen_vl_interface': 8.737515066427861e-06, 'learning_rate/action_model': 8.73751506642786e-05, 'epoch': 0.49}) +[2026-06-23 18:28:59,311][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.13736864924430847, 'train/grad_norm_pre_clip': 17.60292625427246, 'timing/data': 0.000604389002546668, 'timing/model': 1.7038951211143285, 'learning_rate/qwen_vl_interface': 8.735446757003e-06, 'learning_rate/action_model': 8.735446757003e-05, 'epoch': 0.49}) +[2026-06-23 18:29:02,649][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.24588069319725037, 'train/grad_norm_pre_clip': 16.347227096557617, 'timing/data': 0.0005892498884350061, 'timing/model': 1.7050734318327159, 'learning_rate/qwen_vl_interface': 8.733377014842635e-06, 'learning_rate/action_model': 8.733377014842635e-05, 'epoch': 0.49}) +[2026-06-23 18:29:05,985][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.12190349400043488, 'train/grad_norm_pre_clip': 19.7104434967041, 'timing/data': 0.0006069119554013014, 'timing/model': 1.6943428399972618, 'learning_rate/qwen_vl_interface': 8.731305840797557e-06, 'learning_rate/action_model': 8.731305840797558e-05, 'epoch': 0.49}) +[2026-06-23 18:29:09,316][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.15929730236530304, 'train/grad_norm_pre_clip': 13.66027545928955, 'timing/data': 0.0006131338886916637, 'timing/model': 1.6899559628218412, 'learning_rate/qwen_vl_interface': 8.72923323571915e-06, 'learning_rate/action_model': 8.72923323571915e-05, 'epoch': 0.49}) +[2026-06-23 18:29:12,653][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.16829796135425568, 'train/grad_norm_pre_clip': 12.720922470092773, 'timing/data': 0.0006001710426062346, 'timing/model': 1.6914135799743235, 'learning_rate/qwen_vl_interface': 8.727159200459381e-06, 'learning_rate/action_model': 8.727159200459381e-05, 'epoch': 0.49}) +[2026-06-23 18:29:15,985][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.22512491047382355, 'train/grad_norm_pre_clip': 25.77815055847168, 'timing/data': 0.0006001999136060476, 'timing/model': 1.6906499878969043, 'learning_rate/qwen_vl_interface': 8.725083735870807e-06, 'learning_rate/action_model': 8.725083735870807e-05, 'epoch': 0.49}) +[2026-06-23 18:29:19,410][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.2670564651489258, 'train/grad_norm_pre_clip': 29.8482723236084, 'timing/data': 0.0006182750221341848, 'timing/model': 1.6917270990088582, 'learning_rate/qwen_vl_interface': 8.723006842806575e-06, 'learning_rate/action_model': 8.723006842806576e-05, 'epoch': 0.49}) +[2026-06-23 18:29:22,741][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.20940500497817993, 'train/grad_norm_pre_clip': 57.97315979003906, 'timing/data': 0.0006114610005170107, 'timing/model': 1.6934573319740593, 'learning_rate/qwen_vl_interface': 8.720928522120419e-06, 'learning_rate/action_model': 8.720928522120418e-05, 'epoch': 0.49}) +[2026-06-23 18:29:26,082][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.11224372684955597, 'train/grad_norm_pre_clip': 50.38676071166992, 'timing/data': 0.0006166510283946991, 'timing/model': 1.695422095945105, 'learning_rate/qwen_vl_interface': 8.718848774666655e-06, 'learning_rate/action_model': 8.718848774666655e-05, 'epoch': 0.49}) +[2026-06-23 18:29:29,424][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.1770080327987671, 'train/grad_norm_pre_clip': 23.671772003173828, 'timing/data': 0.0006092868279665709, 'timing/model': 1.6989279419649392, 'learning_rate/qwen_vl_interface': 8.716767601300189e-06, 'learning_rate/action_model': 8.716767601300188e-05, 'epoch': 0.49}) +[2026-06-23 18:29:32,784][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.21794600784778595, 'train/grad_norm_pre_clip': 10.583357810974121, 'timing/data': 0.0007455239538103342, 'timing/model': 1.6928488609846681, 'learning_rate/qwen_vl_interface': 8.714685002876514e-06, 'learning_rate/action_model': 8.714685002876513e-05, 'epoch': 0.49}) +[2026-06-23 18:29:36,121][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.0889471247792244, 'train/grad_norm_pre_clip': 17.24711036682129, 'timing/data': 0.0006227230187505484, 'timing/model': 1.6886335799936205, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.49}) +[2026-06-23 18:29:39,450][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.19158940017223358, 'train/grad_norm_pre_clip': 46.48765563964844, 'timing/data': 0.0006164100486785173, 'timing/model': 1.686871426878497, 'learning_rate/qwen_vl_interface': 8.710515534282428e-06, 'learning_rate/action_model': 8.710515534282428e-05, 'epoch': 0.49}) +[2026-06-23 18:29:42,788][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.25352486968040466, 'train/grad_norm_pre_clip': 28.983917236328125, 'timing/data': 0.0005977961700409651, 'timing/model': 1.6958906459622085, 'learning_rate/qwen_vl_interface': 8.708428665825931e-06, 'learning_rate/action_model': 8.708428665825931e-05, 'epoch': 0.49}) +[2026-06-23 18:29:46,123][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.19632075726985931, 'train/grad_norm_pre_clip': 20.48369026184082, 'timing/data': 0.0005945500452071428, 'timing/model': 1.6926594891119748, 'learning_rate/qwen_vl_interface': 8.706340375740045e-06, 'learning_rate/action_model': 8.706340375740045e-05, 'epoch': 0.49}) +[2026-06-23 18:29:49,459][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.11324529349803925, 'train/grad_norm_pre_clip': 42.44377517700195, 'timing/data': 0.0006053100805729628, 'timing/model': 1.6939938471186906, 'learning_rate/qwen_vl_interface': 8.704250664883188e-06, 'learning_rate/action_model': 8.704250664883188e-05, 'epoch': 0.5}) +[2026-06-23 18:29:52,795][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.19677701592445374, 'train/grad_norm_pre_clip': 41.861854553222656, 'timing/data': 0.0006205278914421797, 'timing/model': 1.6906367719639093, 'learning_rate/qwen_vl_interface': 8.70215953411436e-06, 'learning_rate/action_model': 8.70215953411436e-05, 'epoch': 0.5}) +[2026-06-23 18:29:56,129][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.18625721335411072, 'train/grad_norm_pre_clip': 24.728239059448242, 'timing/data': 0.0006184340454638004, 'timing/model': 1.6927437169943005, 'learning_rate/qwen_vl_interface': 8.70006698429315e-06, 'learning_rate/action_model': 8.700066984293149e-05, 'epoch': 0.5}) +[2026-06-23 18:29:59,456][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.263068288564682, 'train/grad_norm_pre_clip': 18.607263565063477, 'timing/data': 0.0006025549955666065, 'timing/model': 1.6861587660387158, 'learning_rate/qwen_vl_interface': 8.697973016279722e-06, 'learning_rate/action_model': 8.697973016279722e-05, 'epoch': 0.5}) +[2026-06-23 18:30:02,779][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.11630479246377945, 'train/grad_norm_pre_clip': 38.70344543457031, 'timing/data': 0.0006034860853105783, 'timing/model': 1.6835636370815337, 'learning_rate/qwen_vl_interface': 8.695877630934827e-06, 'learning_rate/action_model': 8.695877630934826e-05, 'epoch': 0.5}) +[2026-06-23 18:30:06,215][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.24942843616008759, 'train/grad_norm_pre_clip': 47.45169448852539, 'timing/data': 0.0006258378271013498, 'timing/model': 1.7968416269868612, 'learning_rate/qwen_vl_interface': 8.693780829119801e-06, 'learning_rate/action_model': 8.693780829119801e-05, 'epoch': 0.5}) +[2026-06-23 18:30:09,547][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.15024802088737488, 'train/grad_norm_pre_clip': 22.653743743896484, 'timing/data': 0.0006003910675644875, 'timing/model': 1.6910331149119884, 'learning_rate/qwen_vl_interface': 8.69168261169656e-06, 'learning_rate/action_model': 8.691682611696559e-05, 'epoch': 0.5}) +[2026-06-23 18:30:12,881][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.23388902842998505, 'train/grad_norm_pre_clip': 43.460174560546875, 'timing/data': 0.0008611211087554693, 'timing/model': 1.6947247139178216, 'learning_rate/qwen_vl_interface': 8.689582979527599e-06, 'learning_rate/action_model': 8.689582979527599e-05, 'epoch': 0.5}) +[2026-06-23 18:30:16,221][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.10383670777082443, 'train/grad_norm_pre_clip': 11.073301315307617, 'timing/data': 0.0009193501900881529, 'timing/model': 1.6969441501423717, 'learning_rate/qwen_vl_interface': 8.687481933476e-06, 'learning_rate/action_model': 8.687481933476e-05, 'epoch': 0.5}) +[2026-06-23 18:30:19,565][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.19199837744235992, 'train/grad_norm_pre_clip': 35.35857391357422, 'timing/data': 0.0006248271092772484, 'timing/model': 1.698873600922525, 'learning_rate/qwen_vl_interface': 8.685379474405423e-06, 'learning_rate/action_model': 8.685379474405423e-05, 'epoch': 0.5}) +[2026-06-23 18:30:22,923][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.12712061405181885, 'train/grad_norm_pre_clip': 16.752546310424805, 'timing/data': 0.000651275971904397, 'timing/model': 1.703093925025314, 'learning_rate/qwen_vl_interface': 8.68327560318011e-06, 'learning_rate/action_model': 8.68327560318011e-05, 'epoch': 0.5}) +[2026-06-23 18:30:26,266][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.1730336844921112, 'train/grad_norm_pre_clip': 39.193912506103516, 'timing/data': 0.0008350210264325142, 'timing/model': 1.6959247859194875, 'learning_rate/qwen_vl_interface': 8.681170320664883e-06, 'learning_rate/action_model': 8.681170320664883e-05, 'epoch': 0.5}) +[2026-06-23 18:30:29,598][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.31137770414352417, 'train/grad_norm_pre_clip': 94.8140869140625, 'timing/data': 0.0006143359933048487, 'timing/model': 1.690213963855058, 'learning_rate/qwen_vl_interface': 8.679063627725143e-06, 'learning_rate/action_model': 8.679063627725144e-05, 'epoch': 0.5}) +[2026-06-23 18:30:32,937][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.3054567575454712, 'train/grad_norm_pre_clip': 73.0459213256836, 'timing/data': 0.0006051990203559399, 'timing/model': 1.6973837921395898, 'learning_rate/qwen_vl_interface': 8.676955525226876e-06, 'learning_rate/action_model': 8.676955525226875e-05, 'epoch': 0.5}) +[2026-06-23 18:30:36,274][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.1715996265411377, 'train/grad_norm_pre_clip': 31.561080932617188, 'timing/data': 0.0006108800880610943, 'timing/model': 1.6909118660259992, 'learning_rate/qwen_vl_interface': 8.674846014036636e-06, 'learning_rate/action_model': 8.674846014036636e-05, 'epoch': 0.5}) +[2026-06-23 18:30:39,615][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.3083937168121338, 'train/grad_norm_pre_clip': 60.720184326171875, 'timing/data': 0.0005995589308440685, 'timing/model': 1.6946616650093347, 'learning_rate/qwen_vl_interface': 8.672735095021569e-06, 'learning_rate/action_model': 8.672735095021569e-05, 'epoch': 0.5}) +[2026-06-23 18:30:42,956][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.22771137952804565, 'train/grad_norm_pre_clip': 79.82923126220703, 'timing/data': 0.000880679115653038, 'timing/model': 1.69793098885566, 'learning_rate/qwen_vl_interface': 8.670622769049393e-06, 'learning_rate/action_model': 8.670622769049392e-05, 'epoch': 0.5}) +[2026-06-23 18:30:46,294][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.15964002907276154, 'train/grad_norm_pre_clip': 64.19322204589844, 'timing/data': 0.0006093368865549564, 'timing/model': 1.695870130090043, 'learning_rate/qwen_vl_interface': 8.668509036988406e-06, 'learning_rate/action_model': 8.668509036988405e-05, 'epoch': 0.5}) +[2026-06-23 18:30:49,622][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.15332825481891632, 'train/grad_norm_pre_clip': 84.04974365234375, 'timing/data': 0.0006097080186009407, 'timing/model': 1.687486354028806, 'learning_rate/qwen_vl_interface': 8.66639389970748e-06, 'learning_rate/action_model': 8.666393899707479e-05, 'epoch': 0.5}) +[2026-06-23 18:30:52,951][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.25848913192749023, 'train/grad_norm_pre_clip': 20.103151321411133, 'timing/data': 0.0006210389547049999, 'timing/model': 1.6874837139621377, 'learning_rate/qwen_vl_interface': 8.66427735807607e-06, 'learning_rate/action_model': 8.664277358076069e-05, 'epoch': 0.5}) +[2026-06-23 18:30:56,387][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1300 +[2026-06-23 18:32:28,984][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1300 +[2026-06-23 18:32:28,984][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1300 +[2026-06-23 18:33:40,247][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1300 +[2026-06-23 18:33:40,248][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.23144641518592834, 'train/grad_norm_pre_clip': 26.37131690979004, 'eval/action_loss/seconds': 92.59459412097931, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.09036531299352646, 'eval/latency_2/loss': 0.09036531299352646, 'eval/flappy/loss': 0.09036531299352646, 'eval/flappy/latency_2/loss': 0.09036531299352646, 'eval/flap_precision': 0.899009900990099, 'eval/flap_recall': 0.9190283400809717, 'eval/flap_f1': 0.908908908908909, 'eval/flap_rate_model': 0.21041666666666667, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.908908908908909, 'eval/latency_2/flap_precision': 0.899009900990099, 'eval/latency_2/flap_recall': 0.9190283400809717, 'eval/latency_2/flap_f1': 0.908908908908909, 'eval/latency_2/flap_rate_model': 0.21041666666666667, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.908908908908909, 'eval/action_classification/data_wait_seconds': 0.9583540558815002, 'eval/action_classification/predict_seconds': 70.1650390625, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006128440145403147, 'timing/model': 1.7933027930557728, 'learning_rate/qwen_vl_interface': 8.662159412964208e-06, 'learning_rate/action_model': 8.662159412964207e-05, 'epoch': 0.5}) +[2026-06-23 18:33:43,567][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.2841181755065918, 'train/grad_norm_pre_clip': 41.361995697021484, 'timing/data': 0.0006189849227666855, 'timing/model': 1.6852951950859278, 'learning_rate/qwen_vl_interface': 8.660040065242499e-06, 'learning_rate/action_model': 8.660040065242498e-05, 'epoch': 0.5}) +[2026-06-23 18:33:46,899][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.12583908438682556, 'train/grad_norm_pre_clip': 38.14454650878906, 'timing/data': 0.0006932138931006193, 'timing/model': 1.6903465390205383, 'learning_rate/qwen_vl_interface': 8.657919315782127e-06, 'learning_rate/action_model': 8.657919315782127e-05, 'epoch': 0.5}) +[2026-06-23 18:33:50,223][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.16781334578990936, 'train/grad_norm_pre_clip': 11.447400093078613, 'timing/data': 0.00060688192024827, 'timing/model': 1.6884538480080664, 'learning_rate/qwen_vl_interface': 8.655797165454853e-06, 'learning_rate/action_model': 8.655797165454853e-05, 'epoch': 0.5}) +[2026-06-23 18:33:53,545][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.1716531664133072, 'train/grad_norm_pre_clip': 22.906696319580078, 'timing/data': 0.0006069331429898739, 'timing/model': 1.6916028519626707, 'learning_rate/qwen_vl_interface': 8.653673615133013e-06, 'learning_rate/action_model': 8.653673615133013e-05, 'epoch': 0.5}) +[2026-06-23 18:33:56,872][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.16187019646167755, 'train/grad_norm_pre_clip': 19.616222381591797, 'timing/data': 0.0006040071602910757, 'timing/model': 1.6903991759754717, 'learning_rate/qwen_vl_interface': 8.651548665689519e-06, 'learning_rate/action_model': 8.651548665689518e-05, 'epoch': 0.51}) +[2026-06-23 18:34:00,205][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.2828098237514496, 'train/grad_norm_pre_clip': 16.351688385009766, 'timing/data': 0.0007112680468708277, 'timing/model': 1.6926797670312226, 'learning_rate/qwen_vl_interface': 8.649422317997855e-06, 'learning_rate/action_model': 8.649422317997856e-05, 'epoch': 0.51}) +[2026-06-23 18:34:03,532][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.3192211389541626, 'train/grad_norm_pre_clip': 21.497695922851562, 'timing/data': 0.0005916031077504158, 'timing/model': 1.6903464270289987, 'learning_rate/qwen_vl_interface': 8.647294572932082e-06, 'learning_rate/action_model': 8.647294572932082e-05, 'epoch': 0.51}) +[2026-06-23 18:34:07,024][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.27454817295074463, 'train/grad_norm_pre_clip': 77.7624740600586, 'timing/data': 0.0008949839975684881, 'timing/model': 1.8485602459404618, 'learning_rate/qwen_vl_interface': 8.64516543136684e-06, 'learning_rate/action_model': 8.64516543136684e-05, 'epoch': 0.51}) +[2026-06-23 18:34:10,353][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.22940298914909363, 'train/grad_norm_pre_clip': 63.533512115478516, 'timing/data': 0.0006201670039445162, 'timing/model': 1.6890874260570854, 'learning_rate/qwen_vl_interface': 8.643034894177335e-06, 'learning_rate/action_model': 8.643034894177335e-05, 'epoch': 0.51}) +[2026-06-23 18:34:13,684][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.09901656210422516, 'train/grad_norm_pre_clip': 19.780359268188477, 'timing/data': 0.0007380279712378979, 'timing/model': 1.6909333891235292, 'learning_rate/qwen_vl_interface': 8.64090296223935e-06, 'learning_rate/action_model': 8.64090296223935e-05, 'epoch': 0.51}) +[2026-06-23 18:34:17,021][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.19128884375095367, 'train/grad_norm_pre_clip': 23.875886917114258, 'timing/data': 0.0006179031915962696, 'timing/model': 1.6976923369802535, 'learning_rate/qwen_vl_interface': 8.638769636429245e-06, 'learning_rate/action_model': 8.638769636429245e-05, 'epoch': 0.51}) +[2026-06-23 18:34:20,351][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.15069487690925598, 'train/grad_norm_pre_clip': 18.450071334838867, 'timing/data': 0.0006086048670113087, 'timing/model': 1.689822891028598, 'learning_rate/qwen_vl_interface': 8.636634917623946e-06, 'learning_rate/action_model': 8.636634917623944e-05, 'epoch': 0.51}) +[2026-06-23 18:34:23,678][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.13822054862976074, 'train/grad_norm_pre_clip': 13.306058883666992, 'timing/data': 0.0006124828942120075, 'timing/model': 1.6879102918319404, 'learning_rate/qwen_vl_interface': 8.634498806700955e-06, 'learning_rate/action_model': 8.634498806700954e-05, 'epoch': 0.51}) +[2026-06-23 18:34:27,013][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.36273473501205444, 'train/grad_norm_pre_clip': 32.570960998535156, 'timing/data': 0.000610017916187644, 'timing/model': 1.6896351091563702, 'learning_rate/qwen_vl_interface': 8.63236130453835e-06, 'learning_rate/action_model': 8.63236130453835e-05, 'epoch': 0.51}) +[2026-06-23 18:34:30,351][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.2815319001674652, 'train/grad_norm_pre_clip': 27.2970027923584, 'timing/data': 0.0008286598604172468, 'timing/model': 1.69629119313322, 'learning_rate/qwen_vl_interface': 8.630222412014777e-06, 'learning_rate/action_model': 8.630222412014776e-05, 'epoch': 0.51}) +[2026-06-23 18:34:33,683][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.10398095101118088, 'train/grad_norm_pre_clip': 49.01442337036133, 'timing/data': 0.000596201978623867, 'timing/model': 1.6912278961390257, 'learning_rate/qwen_vl_interface': 8.628082130009451e-06, 'learning_rate/action_model': 8.628082130009451e-05, 'epoch': 0.51}) +[2026-06-23 18:34:37,018][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.11252182722091675, 'train/grad_norm_pre_clip': 43.4052848815918, 'timing/data': 0.0006068020593374968, 'timing/model': 1.69550201902166, 'learning_rate/qwen_vl_interface': 8.62594045940216e-06, 'learning_rate/action_model': 8.62594045940216e-05, 'epoch': 0.51}) +[2026-06-23 18:34:40,356][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.2911481559276581, 'train/grad_norm_pre_clip': 127.03626251220703, 'timing/data': 0.0006091860122978687, 'timing/model': 1.6963583920150995, 'learning_rate/qwen_vl_interface': 8.623797401073272e-06, 'learning_rate/action_model': 8.623797401073271e-05, 'epoch': 0.51}) +[2026-06-23 18:34:43,693][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.20569515228271484, 'train/grad_norm_pre_clip': 18.375825881958008, 'timing/data': 0.0006145259831100702, 'timing/model': 1.6944936141371727, 'learning_rate/qwen_vl_interface': 8.62165295590371e-06, 'learning_rate/action_model': 8.621652955903709e-05, 'epoch': 0.51}) +[2026-06-23 18:34:47,027][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.16514605283737183, 'train/grad_norm_pre_clip': 19.39834976196289, 'timing/data': 0.0007201449479907751, 'timing/model': 1.6945798581000417, 'learning_rate/qwen_vl_interface': 8.619507124774975e-06, 'learning_rate/action_model': 8.619507124774975e-05, 'epoch': 0.51}) +[2026-06-23 18:34:50,364][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.07866452634334564, 'train/grad_norm_pre_clip': 35.53487777709961, 'timing/data': 0.000616519944742322, 'timing/model': 1.69610974797979, 'learning_rate/qwen_vl_interface': 8.61735990856914e-06, 'learning_rate/action_model': 8.617359908569139e-05, 'epoch': 0.51}) +[2026-06-23 18:34:53,688][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.17637193202972412, 'train/grad_norm_pre_clip': 42.815216064453125, 'timing/data': 0.0005967230536043644, 'timing/model': 1.6907298769801855, 'learning_rate/qwen_vl_interface': 8.615211308168842e-06, 'learning_rate/action_model': 8.61521130816884e-05, 'epoch': 0.51}) +[2026-06-23 18:34:57,158][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.1978597342967987, 'train/grad_norm_pre_clip': 44.30400466918945, 'timing/data': 0.0006108989473432302, 'timing/model': 1.693754558917135, 'learning_rate/qwen_vl_interface': 8.613061324457289e-06, 'learning_rate/action_model': 8.613061324457288e-05, 'epoch': 0.51}) +[2026-06-23 18:35:00,493][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.19185014069080353, 'train/grad_norm_pre_clip': 11.627906799316406, 'timing/data': 0.0008195110131055117, 'timing/model': 1.694801189005375, 'learning_rate/qwen_vl_interface': 8.610909958318262e-06, 'learning_rate/action_model': 8.610909958318261e-05, 'epoch': 0.51}) +[2026-06-23 18:35:03,831][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.2196163684129715, 'train/grad_norm_pre_clip': 85.29198455810547, 'timing/data': 0.0007855680305510759, 'timing/model': 1.6976896859705448, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.51}) +[2026-06-23 18:35:07,155][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.18012720346450806, 'train/grad_norm_pre_clip': 17.47801399230957, 'timing/data': 0.000702601857483387, 'timing/model': 1.692192506045103, 'learning_rate/qwen_vl_interface': 8.606603082295724e-06, 'learning_rate/action_model': 8.606603082295724e-05, 'epoch': 0.51}) +[2026-06-23 18:35:10,499][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.1633068174123764, 'train/grad_norm_pre_clip': 24.643085479736328, 'timing/data': 0.0006121618207544088, 'timing/model': 1.6933298769872636, 'learning_rate/qwen_vl_interface': 8.60444757418261e-06, 'learning_rate/action_model': 8.604447574182609e-05, 'epoch': 0.51}) +[2026-06-23 18:35:13,846][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.13199301064014435, 'train/grad_norm_pre_clip': 19.411596298217773, 'timing/data': 0.0009922070894390345, 'timing/model': 1.7003313519526273, 'learning_rate/qwen_vl_interface': 8.602290687182805e-06, 'learning_rate/action_model': 8.602290687182806e-05, 'epoch': 0.51}) +[2026-06-23 18:35:17,187][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.13412629067897797, 'train/grad_norm_pre_clip': 11.004595756530762, 'timing/data': 0.0006058600265532732, 'timing/model': 1.6950942219700664, 'learning_rate/qwen_vl_interface': 8.600132422182929e-06, 'learning_rate/action_model': 8.600132422182928e-05, 'epoch': 0.51}) +[2026-06-23 18:35:20,537][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.09707711637020111, 'train/grad_norm_pre_clip': 21.494192123413086, 'timing/data': 0.0006133639253675938, 'timing/model': 1.704290535999462, 'learning_rate/qwen_vl_interface': 8.597972780070156e-06, 'learning_rate/action_model': 8.597972780070155e-05, 'epoch': 0.51}) +[2026-06-23 18:35:23,884][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.3320164382457733, 'train/grad_norm_pre_clip': 33.87372970581055, 'timing/data': 0.0006643189117312431, 'timing/model': 1.6975723979994655, 'learning_rate/qwen_vl_interface': 8.595811761732239e-06, 'learning_rate/action_model': 8.595811761732238e-05, 'epoch': 0.52}) +[2026-06-23 18:35:27,222][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.12716573476791382, 'train/grad_norm_pre_clip': 15.079987525939941, 'timing/data': 0.0005998089909553528, 'timing/model': 1.6927686841227114, 'learning_rate/qwen_vl_interface': 8.593649368057489e-06, 'learning_rate/action_model': 8.59364936805749e-05, 'epoch': 0.52}) +[2026-06-23 18:35:30,559][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.2584902048110962, 'train/grad_norm_pre_clip': 34.5240364074707, 'timing/data': 0.0005885469727218151, 'timing/model': 1.693404451943934, 'learning_rate/qwen_vl_interface': 8.591485599934786e-06, 'learning_rate/action_model': 8.591485599934785e-05, 'epoch': 0.52}) +[2026-06-23 18:35:33,893][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.18396224081516266, 'train/grad_norm_pre_clip': 13.530448913574219, 'timing/data': 0.0006143860518932343, 'timing/model': 1.6919680428691208, 'learning_rate/qwen_vl_interface': 8.58932045825357e-06, 'learning_rate/action_model': 8.589320458253569e-05, 'epoch': 0.52}) +[2026-06-23 18:35:37,228][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.13897699117660522, 'train/grad_norm_pre_clip': 28.26982307434082, 'timing/data': 0.0006091860122978687, 'timing/model': 1.6977301670704037, 'learning_rate/qwen_vl_interface': 8.58715394390385e-06, 'learning_rate/action_model': 8.58715394390385e-05, 'epoch': 0.52}) +[2026-06-23 18:35:40,565][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.11732325702905655, 'train/grad_norm_pre_clip': 32.90233612060547, 'timing/data': 0.0006220100913196802, 'timing/model': 1.6917257830500603, 'learning_rate/qwen_vl_interface': 8.5849860577762e-06, 'learning_rate/action_model': 8.5849860577762e-05, 'epoch': 0.52}) +[2026-06-23 18:35:43,900][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.2683654725551605, 'train/grad_norm_pre_clip': 73.9677734375, 'timing/data': 0.0006139050237834454, 'timing/model': 1.6965763019397855, 'learning_rate/qwen_vl_interface': 8.582816800761756e-06, 'learning_rate/action_model': 8.582816800761755e-05, 'epoch': 0.52}) +[2026-06-23 18:35:47,330][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.18838366866111755, 'train/grad_norm_pre_clip': 27.027063369750977, 'timing/data': 0.0006065010093152523, 'timing/model': 1.691785081056878, 'learning_rate/qwen_vl_interface': 8.580646173752214e-06, 'learning_rate/action_model': 8.580646173752213e-05, 'epoch': 0.52}) +[2026-06-23 18:35:50,659][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.10821261256933212, 'train/grad_norm_pre_clip': 71.37596893310547, 'timing/data': 0.0006083638872951269, 'timing/model': 1.686396726872772, 'learning_rate/qwen_vl_interface': 8.578474177639841e-06, 'learning_rate/action_model': 8.578474177639842e-05, 'epoch': 0.52}) +[2026-06-23 18:35:54,001][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.16261087357997894, 'train/grad_norm_pre_clip': 43.908809661865234, 'timing/data': 0.0008801061194390059, 'timing/model': 1.702876424184069, 'learning_rate/qwen_vl_interface': 8.576300813317463e-06, 'learning_rate/action_model': 8.576300813317462e-05, 'epoch': 0.52}) +[2026-06-23 18:35:57,338][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.2735987901687622, 'train/grad_norm_pre_clip': 83.98375701904297, 'timing/data': 0.0006014220416545868, 'timing/model': 1.696224784012884, 'learning_rate/qwen_vl_interface': 8.574126081678463e-06, 'learning_rate/action_model': 8.574126081678462e-05, 'epoch': 0.52}) +[2026-06-23 18:36:00,673][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.20041541755199432, 'train/grad_norm_pre_clip': 12.3158540725708, 'timing/data': 0.0005948389880359173, 'timing/model': 1.6922754880506545, 'learning_rate/qwen_vl_interface': 8.571949983616796e-06, 'learning_rate/action_model': 8.571949983616795e-05, 'epoch': 0.52}) +[2026-06-23 18:36:04,012][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.1935046762228012, 'train/grad_norm_pre_clip': 13.02892780303955, 'timing/data': 0.000595940975472331, 'timing/model': 1.6927689800504595, 'learning_rate/qwen_vl_interface': 8.569772520026971e-06, 'learning_rate/action_model': 8.569772520026971e-05, 'epoch': 0.52}) +[2026-06-23 18:36:07,345][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.11719729006290436, 'train/grad_norm_pre_clip': 12.672191619873047, 'timing/data': 0.0006711420137435198, 'timing/model': 1.694348325021565, 'learning_rate/qwen_vl_interface': 8.567593691804061e-06, 'learning_rate/action_model': 8.567593691804061e-05, 'epoch': 0.52}) +[2026-06-23 18:36:10,689][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.18841780722141266, 'train/grad_norm_pre_clip': 13.261938095092773, 'timing/data': 0.0006528780795633793, 'timing/model': 1.699007524875924, 'learning_rate/qwen_vl_interface': 8.565413499843703e-06, 'learning_rate/action_model': 8.565413499843703e-05, 'epoch': 0.52}) +[2026-06-23 18:36:14,029][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.2884252369403839, 'train/grad_norm_pre_clip': 53.183162689208984, 'timing/data': 0.000596882775425911, 'timing/model': 1.697734243934974, 'learning_rate/qwen_vl_interface': 8.563231945042087e-06, 'learning_rate/action_model': 8.563231945042088e-05, 'epoch': 0.52}) +[2026-06-23 18:36:17,364][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.16283684968948364, 'train/grad_norm_pre_clip': 29.11345672607422, 'timing/data': 0.000594868790358305, 'timing/model': 1.692615107866004, 'learning_rate/qwen_vl_interface': 8.56104902829597e-06, 'learning_rate/action_model': 8.56104902829597e-05, 'epoch': 0.52}) +[2026-06-23 18:36:20,699][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.18399357795715332, 'train/grad_norm_pre_clip': 30.0952205657959, 'timing/data': 0.0006697899661958218, 'timing/model': 1.6980241970159113, 'learning_rate/qwen_vl_interface': 8.558864750502669e-06, 'learning_rate/action_model': 8.558864750502668e-05, 'epoch': 0.52}) +[2026-06-23 18:36:24,039][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.19452005624771118, 'train/grad_norm_pre_clip': 46.246925354003906, 'timing/data': 0.000605930108577013, 'timing/model': 1.6974529668223113, 'learning_rate/qwen_vl_interface': 8.556679112560055e-06, 'learning_rate/action_model': 8.556679112560054e-05, 'epoch': 0.52}) +[2026-06-23 18:36:27,378][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.14770455658435822, 'train/grad_norm_pre_clip': 31.483705520629883, 'timing/data': 0.0006072628311812878, 'timing/model': 1.6957446879241616, 'learning_rate/qwen_vl_interface': 8.554492115366562e-06, 'learning_rate/action_model': 8.554492115366561e-05, 'epoch': 0.52}) +[2026-06-23 18:36:30,716][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.19462822377681732, 'train/grad_norm_pre_clip': 21.190990447998047, 'timing/data': 0.0006023128516972065, 'timing/model': 1.6950661959126592, 'learning_rate/qwen_vl_interface': 8.55230375982118e-06, 'learning_rate/action_model': 8.55230375982118e-05, 'epoch': 0.52}) +[2026-06-23 18:36:34,151][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.21181946992874146, 'train/grad_norm_pre_clip': 60.12019729614258, 'timing/data': 0.0005985160823911428, 'timing/model': 1.7926494830753654, 'learning_rate/qwen_vl_interface': 8.550114046823463e-06, 'learning_rate/action_model': 8.550114046823463e-05, 'epoch': 0.52}) +[2026-06-23 18:36:37,491][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.14542056620121002, 'train/grad_norm_pre_clip': 29.57219123840332, 'timing/data': 0.0006160291377454996, 'timing/model': 1.69739849306643, 'learning_rate/qwen_vl_interface': 8.547922977273518e-06, 'learning_rate/action_model': 8.547922977273518e-05, 'epoch': 0.52}) +[2026-06-23 18:36:40,826][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.19463172554969788, 'train/grad_norm_pre_clip': 68.5636215209961, 'timing/data': 0.0006622658111155033, 'timing/model': 1.6935367139521986, 'learning_rate/qwen_vl_interface': 8.54573055207201e-06, 'learning_rate/action_model': 8.545730552072008e-05, 'epoch': 0.52}) +[2026-06-23 18:36:44,162][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.1997215300798416, 'train/grad_norm_pre_clip': 62.78712844848633, 'timing/data': 0.0006231130100786686, 'timing/model': 1.6965009549167007, 'learning_rate/qwen_vl_interface': 8.543536772120161e-06, 'learning_rate/action_model': 8.54353677212016e-05, 'epoch': 0.52}) +[2026-06-23 18:36:47,499][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.25080162286758423, 'train/grad_norm_pre_clip': 12.039645195007324, 'timing/data': 0.0006146160885691643, 'timing/model': 1.6948503248859197, 'learning_rate/qwen_vl_interface': 8.541341638319754e-06, 'learning_rate/action_model': 8.541341638319754e-05, 'epoch': 0.52}) +[2026-06-23 18:36:50,830][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.22100473940372467, 'train/grad_norm_pre_clip': 20.797582626342773, 'timing/data': 0.0007047960534691811, 'timing/model': 1.690294360043481, 'learning_rate/qwen_vl_interface': 8.539145151573126e-06, 'learning_rate/action_model': 8.539145151573125e-05, 'epoch': 0.53}) +[2026-06-23 18:36:54,160][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.14464159309864044, 'train/grad_norm_pre_clip': 60.776607513427734, 'timing/data': 0.0006053789984434843, 'timing/model': 1.6931302701123059, 'learning_rate/qwen_vl_interface': 8.536947312783167e-06, 'learning_rate/action_model': 8.536947312783166e-05, 'epoch': 0.53}) +[2026-06-23 18:36:57,493][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.11702246963977814, 'train/grad_norm_pre_clip': 36.3719596862793, 'timing/data': 0.0006083138287067413, 'timing/model': 1.693947282852605, 'learning_rate/qwen_vl_interface': 8.534748122853326e-06, 'learning_rate/action_model': 8.534748122853325e-05, 'epoch': 0.53}) +[2026-06-23 18:37:00,833][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.14518257975578308, 'train/grad_norm_pre_clip': 31.60923194885254, 'timing/data': 0.0005956299137324095, 'timing/model': 1.6951826410368085, 'learning_rate/qwen_vl_interface': 8.532547582687609e-06, 'learning_rate/action_model': 8.532547582687608e-05, 'epoch': 0.53}) +[2026-06-23 18:37:04,161][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.37724000215530396, 'train/grad_norm_pre_clip': 72.66293334960938, 'timing/data': 0.0006257470231503248, 'timing/model': 1.6904637708794326, 'learning_rate/qwen_vl_interface': 8.530345693190573e-06, 'learning_rate/action_model': 8.530345693190572e-05, 'epoch': 0.53}) +[2026-06-23 18:37:07,503][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.11617010086774826, 'train/grad_norm_pre_clip': 26.898344039916992, 'timing/data': 0.0006134239956736565, 'timing/model': 1.697722079930827, 'learning_rate/qwen_vl_interface': 8.528142455267328e-06, 'learning_rate/action_model': 8.528142455267329e-05, 'epoch': 0.53}) +[2026-06-23 18:37:10,839][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.09922641515731812, 'train/grad_norm_pre_clip': 12.149861335754395, 'timing/data': 0.0006010800134390593, 'timing/model': 1.6955077301245183, 'learning_rate/qwen_vl_interface': 8.52593786982355e-06, 'learning_rate/action_model': 8.525937869823551e-05, 'epoch': 0.53}) +[2026-06-23 18:37:14,169][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.14881263673305511, 'train/grad_norm_pre_clip': 13.116857528686523, 'timing/data': 0.000623784027993679, 'timing/model': 1.6957218491006643, 'learning_rate/qwen_vl_interface': 8.523731937765456e-06, 'learning_rate/action_model': 8.523731937765456e-05, 'epoch': 0.53}) +[2026-06-23 18:37:17,501][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.19796423614025116, 'train/grad_norm_pre_clip': 50.754310607910156, 'timing/data': 0.0006103580817580223, 'timing/model': 1.688300214940682, 'learning_rate/qwen_vl_interface': 8.521524659999823e-06, 'learning_rate/action_model': 8.521524659999822e-05, 'epoch': 0.53}) +[2026-06-23 18:37:20,839][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.1901100128889084, 'train/grad_norm_pre_clip': 46.86310958862305, 'timing/data': 0.0006104479543864727, 'timing/model': 1.697255315957591, 'learning_rate/qwen_vl_interface': 8.519316037433976e-06, 'learning_rate/action_model': 8.519316037433977e-05, 'epoch': 0.53}) +[2026-06-23 18:37:24,279][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.08656977117061615, 'train/grad_norm_pre_clip': 9.400238990783691, 'timing/data': 0.0005896089132875204, 'timing/model': 1.796247721184045, 'learning_rate/qwen_vl_interface': 8.517106070975801e-06, 'learning_rate/action_model': 8.5171060709758e-05, 'epoch': 0.53}) +[2026-06-23 18:37:27,614][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.2141677439212799, 'train/grad_norm_pre_clip': 63.485355377197266, 'timing/data': 0.000597833888605237, 'timing/model': 1.693409533938393, 'learning_rate/qwen_vl_interface': 8.514894761533729e-06, 'learning_rate/action_model': 8.514894761533728e-05, 'epoch': 0.53}) +[2026-06-23 18:37:30,951][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.30024898052215576, 'train/grad_norm_pre_clip': 91.80253601074219, 'timing/data': 0.0006041259039193392, 'timing/model': 1.696534247836098, 'learning_rate/qwen_vl_interface': 8.512682110016747e-06, 'learning_rate/action_model': 8.512682110016747e-05, 'epoch': 0.53}) +[2026-06-23 18:37:34,285][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.15220291912555695, 'train/grad_norm_pre_clip': 37.497581481933594, 'timing/data': 0.000604596920311451, 'timing/model': 1.6919300640001893, 'learning_rate/qwen_vl_interface': 8.51046811733439e-06, 'learning_rate/action_model': 8.51046811733439e-05, 'epoch': 0.53}) +[2026-06-23 18:37:37,623][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.28633880615234375, 'train/grad_norm_pre_clip': 22.450122833251953, 'timing/data': 0.0006145359948277473, 'timing/model': 1.6933949859812856, 'learning_rate/qwen_vl_interface': 8.50825278439675e-06, 'learning_rate/action_model': 8.508252784396749e-05, 'epoch': 0.53}) +[2026-06-23 18:37:40,951][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.2475290447473526, 'train/grad_norm_pre_clip': 46.537574768066406, 'timing/data': 0.0006317680235952139, 'timing/model': 1.690674235112965, 'learning_rate/qwen_vl_interface': 8.506036112114465e-06, 'learning_rate/action_model': 8.506036112114464e-05, 'epoch': 0.53}) +[2026-06-23 18:37:44,280][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.29230669140815735, 'train/grad_norm_pre_clip': 41.653594970703125, 'timing/data': 0.0006062409374862909, 'timing/model': 1.6913583979476243, 'learning_rate/qwen_vl_interface': 8.503818101398725e-06, 'learning_rate/action_model': 8.503818101398725e-05, 'epoch': 0.53}) +[2026-06-23 18:37:47,618][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.16262906789779663, 'train/grad_norm_pre_clip': 45.61030960083008, 'timing/data': 0.0006055489648133516, 'timing/model': 1.6932107291650027, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.53}) +[2026-06-23 18:37:50,953][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.06195446848869324, 'train/grad_norm_pre_clip': 27.062713623046875, 'timing/data': 0.000603965949267149, 'timing/model': 1.6923998210113496, 'learning_rate/qwen_vl_interface': 8.499378068314395e-06, 'learning_rate/action_model': 8.499378068314395e-05, 'epoch': 0.53}) +[2026-06-23 18:37:54,287][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.34141039848327637, 'train/grad_norm_pre_clip': 86.11254119873047, 'timing/data': 0.0008157950360327959, 'timing/model': 1.6917744618840516, 'learning_rate/qwen_vl_interface': 8.497156047770932e-06, 'learning_rate/action_model': 8.497156047770933e-05, 'epoch': 0.53}) +[2026-06-23 18:37:57,619][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.16560447216033936, 'train/grad_norm_pre_clip': 27.923402786254883, 'timing/data': 0.0006193651352077723, 'timing/model': 1.697481414070353, 'learning_rate/qwen_vl_interface': 8.494932692444276e-06, 'learning_rate/action_model': 8.494932692444276e-05, 'epoch': 0.53}) +[2026-06-23 18:38:00,957][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.1317347139120102, 'train/grad_norm_pre_clip': 30.14231300354004, 'timing/data': 0.0006045070476830006, 'timing/model': 1.6948593489360064, 'learning_rate/qwen_vl_interface': 8.492708003248362e-06, 'learning_rate/action_model': 8.492708003248362e-05, 'epoch': 0.53}) +[2026-06-23 18:38:04,287][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.19535097479820251, 'train/grad_norm_pre_clip': 55.61418151855469, 'timing/data': 0.0005973528604954481, 'timing/model': 1.6893635680899024, 'learning_rate/qwen_vl_interface': 8.490481981097674e-06, 'learning_rate/action_model': 8.490481981097674e-05, 'epoch': 0.53}) +[2026-06-23 18:38:07,628][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.3752973973751068, 'train/grad_norm_pre_clip': 81.71849060058594, 'timing/data': 0.0006099469028413296, 'timing/model': 1.6949384519830346, 'learning_rate/qwen_vl_interface': 8.488254626907249e-06, 'learning_rate/action_model': 8.488254626907249e-05, 'epoch': 0.53}) +[2026-06-23 18:38:10,954][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.2783489227294922, 'train/grad_norm_pre_clip': 65.85682678222656, 'timing/data': 0.0005946781020611525, 'timing/model': 1.691329832188785, 'learning_rate/qwen_vl_interface': 8.486025941592664e-06, 'learning_rate/action_model': 8.486025941592664e-05, 'epoch': 0.53}) +[2026-06-23 18:38:14,380][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.30211156606674194, 'train/grad_norm_pre_clip': 28.20279312133789, 'timing/data': 0.0006096370052546263, 'timing/model': 1.691533001838252, 'learning_rate/qwen_vl_interface': 8.483795926070054e-06, 'learning_rate/action_model': 8.483795926070052e-05, 'epoch': 0.53}) +[2026-06-23 18:38:17,704][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.1794876605272293, 'train/grad_norm_pre_clip': 34.6848030090332, 'timing/data': 0.0005959209520369768, 'timing/model': 1.6858142018318176, 'learning_rate/qwen_vl_interface': 8.481564581256087e-06, 'learning_rate/action_model': 8.481564581256087e-05, 'epoch': 0.54}) +[2026-06-23 18:38:21,034][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.1920665204524994, 'train/grad_norm_pre_clip': 16.643491744995117, 'timing/data': 0.0006038860883563757, 'timing/model': 1.6890093400143087, 'learning_rate/qwen_vl_interface': 8.479331908067987e-06, 'learning_rate/action_model': 8.479331908067987e-05, 'epoch': 0.54}) +[2026-06-23 18:38:24,365][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.09800797700881958, 'train/grad_norm_pre_clip': 27.85803985595703, 'timing/data': 0.0006798079703003168, 'timing/model': 1.6912158580962569, 'learning_rate/qwen_vl_interface': 8.477097907423524e-06, 'learning_rate/action_model': 8.477097907423524e-05, 'epoch': 0.54}) +[2026-06-23 18:38:27,698][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.21208754181861877, 'train/grad_norm_pre_clip': 60.75945281982422, 'timing/data': 0.000685699051246047, 'timing/model': 1.6941034870687872, 'learning_rate/qwen_vl_interface': 8.474862580241006e-06, 'learning_rate/action_model': 8.474862580241006e-05, 'epoch': 0.54}) +[2026-06-23 18:38:31,037][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.16462045907974243, 'train/grad_norm_pre_clip': 18.284061431884766, 'timing/data': 0.0006145560182631016, 'timing/model': 1.6970420819707215, 'learning_rate/qwen_vl_interface': 8.472625927439299e-06, 'learning_rate/action_model': 8.472625927439299e-05, 'epoch': 0.54}) +[2026-06-23 18:38:34,374][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.1573062390089035, 'train/grad_norm_pre_clip': 53.00318145751953, 'timing/data': 0.0006025230977684259, 'timing/model': 1.6933809339534491, 'learning_rate/qwen_vl_interface': 8.4703879499378e-06, 'learning_rate/action_model': 8.4703879499378e-05, 'epoch': 0.54}) +[2026-06-23 18:38:37,698][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.2536042332649231, 'train/grad_norm_pre_clip': 46.848716735839844, 'timing/data': 0.0006065811030566692, 'timing/model': 1.6878255719784647, 'learning_rate/qwen_vl_interface': 8.46814864865646e-06, 'learning_rate/action_model': 8.46814864865646e-05, 'epoch': 0.54}) +[2026-06-23 18:38:41,033][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.17171110212802887, 'train/grad_norm_pre_clip': 22.196063995361328, 'timing/data': 0.0008357120677828789, 'timing/model': 1.6911018940154463, 'learning_rate/qwen_vl_interface': 8.465908024515771e-06, 'learning_rate/action_model': 8.46590802451577e-05, 'epoch': 0.54}) +[2026-06-23 18:38:44,376][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.17114698886871338, 'train/grad_norm_pre_clip': 33.156375885009766, 'timing/data': 0.0007984419353306293, 'timing/model': 1.70016859495081, 'learning_rate/qwen_vl_interface': 8.463666078436769e-06, 'learning_rate/action_model': 8.463666078436769e-05, 'epoch': 0.54}) +[2026-06-23 18:38:47,718][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.1581517457962036, 'train/grad_norm_pre_clip': 74.06702423095703, 'timing/data': 0.0008251219987869263, 'timing/model': 1.7002287060022354, 'learning_rate/qwen_vl_interface': 8.46142281134103e-06, 'learning_rate/action_model': 8.461422811341031e-05, 'epoch': 0.54}) +[2026-06-23 18:38:51,043][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.27457067370414734, 'train/grad_norm_pre_clip': 99.17438507080078, 'timing/data': 0.0006149772088974714, 'timing/model': 1.6835295469500124, 'learning_rate/qwen_vl_interface': 8.459178224150683e-06, 'learning_rate/action_model': 8.459178224150683e-05, 'epoch': 0.54}) +[2026-06-23 18:38:54,372][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.1949334293603897, 'train/grad_norm_pre_clip': 85.72431182861328, 'timing/data': 0.0006064509507268667, 'timing/model': 1.6839118341449648, 'learning_rate/qwen_vl_interface': 8.456932317788389e-06, 'learning_rate/action_model': 8.456932317788389e-05, 'epoch': 0.54}) +[2026-06-23 18:38:57,712][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.1391092985868454, 'train/grad_norm_pre_clip': 22.738117218017578, 'timing/data': 0.0005944480653852224, 'timing/model': 1.7004761470016092, 'learning_rate/qwen_vl_interface': 8.454685093177356e-06, 'learning_rate/action_model': 8.454685093177354e-05, 'epoch': 0.54}) +[2026-06-23 18:39:01,051][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.22767873108386993, 'train/grad_norm_pre_clip': 83.7760238647461, 'timing/data': 0.0006183029618114233, 'timing/model': 1.6962993601337075, 'learning_rate/qwen_vl_interface': 8.452436551241328e-06, 'learning_rate/action_model': 8.452436551241328e-05, 'epoch': 0.54}) +[2026-06-23 18:39:04,697][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.3768078684806824, 'train/grad_norm_pre_clip': 73.10441589355469, 'timing/data': 0.0006243039388209581, 'timing/model': 1.6951650939881802, 'learning_rate/qwen_vl_interface': 8.450186692904605e-06, 'learning_rate/action_model': 8.450186692904606e-05, 'epoch': 0.54}) +[2026-06-23 18:39:08,029][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.26344677805900574, 'train/grad_norm_pre_clip': 80.01499938964844, 'timing/data': 0.0006205369718372822, 'timing/model': 1.6919713059905916, 'learning_rate/qwen_vl_interface': 8.447935519092014e-06, 'learning_rate/action_model': 8.447935519092014e-05, 'epoch': 0.54}) +[2026-06-23 18:39:11,361][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.28858402371406555, 'train/grad_norm_pre_clip': 63.59648132324219, 'timing/data': 0.000613001873716712, 'timing/model': 1.6954308710992336, 'learning_rate/qwen_vl_interface': 8.445683030728925e-06, 'learning_rate/action_model': 8.445683030728925e-05, 'epoch': 0.54}) +[2026-06-23 18:39:14,701][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1400 +[2026-06-23 18:40:47,858][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1400 +[2026-06-23 18:40:47,859][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1400 +[2026-06-23 18:41:59,101][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1400 +[2026-06-23 18:41:59,102][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.17797215282917023, 'train/grad_norm_pre_clip': 35.321868896484375, 'eval/action_loss/seconds': 93.15570886107162, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.09186499565839767, 'eval/latency_2/loss': 0.09186499565839767, 'eval/flappy/loss': 0.09186499565839767, 'eval/flappy/latency_2/loss': 0.09186499565839767, 'eval/flap_precision': 0.9195876288659793, 'eval/flap_recall': 0.902834008097166, 'eval/flap_f1': 0.9111338100102145, 'eval/flap_rate_model': 0.20208333333333334, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9111338100102145, 'eval/latency_2/flap_precision': 0.9195876288659793, 'eval/latency_2/flap_recall': 0.902834008097166, 'eval/latency_2/flap_f1': 0.9111338100102145, 'eval/latency_2/flap_rate_model': 0.20208333333333334, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9111338100102145, 'eval/action_classification/data_wait_seconds': 0.9678757190704346, 'eval/action_classification/predict_seconds': 70.14808654785156, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006443921010941267, 'timing/model': 1.6960039481054991, 'learning_rate/qwen_vl_interface': 8.443429228741256e-06, 'learning_rate/action_model': 8.443429228741255e-05, 'epoch': 0.54}) +[2026-06-23 18:42:02,423][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.10194121301174164, 'train/grad_norm_pre_clip': 28.558040618896484, 'timing/data': 0.0005889080930501223, 'timing/model': 1.687353976070881, 'learning_rate/qwen_vl_interface': 8.441174114055457e-06, 'learning_rate/action_model': 8.441174114055456e-05, 'epoch': 0.54}) +[2026-06-23 18:42:05,747][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.16205266118049622, 'train/grad_norm_pre_clip': 60.50637435913086, 'timing/data': 0.0007080610375851393, 'timing/model': 1.6866926639340818, 'learning_rate/qwen_vl_interface': 8.43891768759852e-06, 'learning_rate/action_model': 8.43891768759852e-05, 'epoch': 0.54}) +[2026-06-23 18:42:09,068][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.26854223012924194, 'train/grad_norm_pre_clip': 37.47091293334961, 'timing/data': 0.0006762619595974684, 'timing/model': 1.687748071970418, 'learning_rate/qwen_vl_interface': 8.436659950297978e-06, 'learning_rate/action_model': 8.436659950297977e-05, 'epoch': 0.54}) +[2026-06-23 18:42:12,390][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.17117848992347717, 'train/grad_norm_pre_clip': 23.17477798461914, 'timing/data': 0.0007962279487401247, 'timing/model': 1.685911379987374, 'learning_rate/qwen_vl_interface': 8.4344009030819e-06, 'learning_rate/action_model': 8.4344009030819e-05, 'epoch': 0.54}) +[2026-06-23 18:42:15,820][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.28446367383003235, 'train/grad_norm_pre_clip': 46.33016586303711, 'timing/data': 0.0006944760680198669, 'timing/model': 1.6891461240593344, 'learning_rate/qwen_vl_interface': 8.432140546878899e-06, 'learning_rate/action_model': 8.432140546878898e-05, 'epoch': 0.54}) +[2026-06-23 18:42:19,145][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.1101478710770607, 'train/grad_norm_pre_clip': 13.67768383026123, 'timing/data': 0.0006556629668921232, 'timing/model': 1.6885244080331177, 'learning_rate/qwen_vl_interface': 8.429878882618116e-06, 'learning_rate/action_model': 8.429878882618116e-05, 'epoch': 0.54}) +[2026-06-23 18:42:22,470][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.3328278660774231, 'train/grad_norm_pre_clip': 41.98638153076172, 'timing/data': 0.0008303220383822918, 'timing/model': 1.6874158659484237, 'learning_rate/qwen_vl_interface': 8.42761591122924e-06, 'learning_rate/action_model': 8.427615911229241e-05, 'epoch': 0.54}) +[2026-06-23 18:42:25,803][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.15006372332572937, 'train/grad_norm_pre_clip': 28.510509490966797, 'timing/data': 0.0006268690340220928, 'timing/model': 1.6921388690825552, 'learning_rate/qwen_vl_interface': 8.425351633642491e-06, 'learning_rate/action_model': 8.425351633642492e-05, 'epoch': 0.54}) +[2026-06-23 18:42:29,125][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.19151580333709717, 'train/grad_norm_pre_clip': 14.913657188415527, 'timing/data': 0.0006543910130858421, 'timing/model': 1.6852329759858549, 'learning_rate/qwen_vl_interface': 8.423086050788632e-06, 'learning_rate/action_model': 8.423086050788631e-05, 'epoch': 0.55}) +[2026-06-23 18:42:32,450][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.14034900069236755, 'train/grad_norm_pre_clip': 51.52505874633789, 'timing/data': 0.0006156580056995153, 'timing/model': 1.6898757391609251, 'learning_rate/qwen_vl_interface': 8.420819163598953e-06, 'learning_rate/action_model': 8.420819163598952e-05, 'epoch': 0.55}) +[2026-06-23 18:42:35,774][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.13768254220485687, 'train/grad_norm_pre_clip': 27.764583587646484, 'timing/data': 0.000822555972263217, 'timing/model': 1.688466700958088, 'learning_rate/qwen_vl_interface': 8.418550973005289e-06, 'learning_rate/action_model': 8.418550973005289e-05, 'epoch': 0.55}) +[2026-06-23 18:42:39,097][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.2591957151889801, 'train/grad_norm_pre_clip': 19.383899688720703, 'timing/data': 0.0006096470169723034, 'timing/model': 1.6904880250804126, 'learning_rate/qwen_vl_interface': 8.416281479940005e-06, 'learning_rate/action_model': 8.416281479940005e-05, 'epoch': 0.55}) +[2026-06-23 18:42:42,431][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.14824558794498444, 'train/grad_norm_pre_clip': 18.266300201416016, 'timing/data': 0.0006225111428648233, 'timing/model': 1.6926060609985143, 'learning_rate/qwen_vl_interface': 8.414010685336008e-06, 'learning_rate/action_model': 8.414010685336008e-05, 'epoch': 0.55}) +[2026-06-23 18:42:45,755][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.11507640033960342, 'train/grad_norm_pre_clip': 36.239803314208984, 'timing/data': 0.0006063601467758417, 'timing/model': 1.6935889211017638, 'learning_rate/qwen_vl_interface': 8.411738590126733e-06, 'learning_rate/action_model': 8.411738590126732e-05, 'epoch': 0.55}) +[2026-06-23 18:42:49,085][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.20879948139190674, 'train/grad_norm_pre_clip': 78.71125793457031, 'timing/data': 0.0008472739718854427, 'timing/model': 1.6915091038681567, 'learning_rate/qwen_vl_interface': 8.409465195246152e-06, 'learning_rate/action_model': 8.409465195246152e-05, 'epoch': 0.55}) +[2026-06-23 18:42:52,405][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.21010704338550568, 'train/grad_norm_pre_clip': 28.342649459838867, 'timing/data': 0.0006128428503870964, 'timing/model': 1.6844908290077, 'learning_rate/qwen_vl_interface': 8.407190501628772e-06, 'learning_rate/action_model': 8.407190501628772e-05, 'epoch': 0.55}) +[2026-06-23 18:42:55,734][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.30324864387512207, 'train/grad_norm_pre_clip': 22.94708251953125, 'timing/data': 0.0006107089575380087, 'timing/model': 1.6884117948357016, 'learning_rate/qwen_vl_interface': 8.404914510209635e-06, 'learning_rate/action_model': 8.404914510209635e-05, 'epoch': 0.55}) +[2026-06-23 18:42:59,068][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.1631225198507309, 'train/grad_norm_pre_clip': 27.42049217224121, 'timing/data': 0.0006868219934403896, 'timing/model': 1.6943268610630184, 'learning_rate/qwen_vl_interface': 8.402637221924315e-06, 'learning_rate/action_model': 8.402637221924314e-05, 'epoch': 0.55}) +[2026-06-23 18:43:02,492][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.12317909300327301, 'train/grad_norm_pre_clip': 42.15828323364258, 'timing/data': 0.0005995980463922024, 'timing/model': 1.78307369700633, 'learning_rate/qwen_vl_interface': 8.400358637708919e-06, 'learning_rate/action_model': 8.400358637708918e-05, 'epoch': 0.55}) +[2026-06-23 18:43:05,830][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.22752892971038818, 'train/grad_norm_pre_clip': 17.471736907958984, 'timing/data': 0.0006147760432213545, 'timing/model': 1.6955544541124254, 'learning_rate/qwen_vl_interface': 8.398078758500086e-06, 'learning_rate/action_model': 8.398078758500086e-05, 'epoch': 0.55}) +[2026-06-23 18:43:09,166][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.08513371646404266, 'train/grad_norm_pre_clip': 12.279303550720215, 'timing/data': 0.000642809085547924, 'timing/model': 1.6944732239935547, 'learning_rate/qwen_vl_interface': 8.395797585234987e-06, 'learning_rate/action_model': 8.395797585234988e-05, 'epoch': 0.55}) +[2026-06-23 18:43:12,493][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.18036159873008728, 'train/grad_norm_pre_clip': 64.79830169677734, 'timing/data': 0.0006661629304289818, 'timing/model': 1.6880849848967046, 'learning_rate/qwen_vl_interface': 8.39351511885133e-06, 'learning_rate/action_model': 8.393515118851329e-05, 'epoch': 0.55}) +[2026-06-23 18:43:15,817][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.19057808816432953, 'train/grad_norm_pre_clip': 31.25100326538086, 'timing/data': 0.0006247248966246843, 'timing/model': 1.690134541830048, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.55}) +[2026-06-23 18:43:19,143][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.2272823303937912, 'train/grad_norm_pre_clip': 67.99545288085938, 'timing/data': 0.000593976816162467, 'timing/model': 1.689051918918267, 'learning_rate/qwen_vl_interface': 8.388946310481807e-06, 'learning_rate/action_model': 8.388946310481807e-05, 'epoch': 0.55}) +[2026-06-23 18:43:22,469][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.20292232930660248, 'train/grad_norm_pre_clip': 57.69239807128906, 'timing/data': 0.0005995579995214939, 'timing/model': 1.686924872919917, 'learning_rate/qwen_vl_interface': 8.386659970374007e-06, 'learning_rate/action_model': 8.386659970374006e-05, 'epoch': 0.55}) +[2026-06-23 18:43:25,798][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.12872864305973053, 'train/grad_norm_pre_clip': 39.197792053222656, 'timing/data': 0.000608192989602685, 'timing/model': 1.6913997698575258, 'learning_rate/qwen_vl_interface': 8.384372340903775e-06, 'learning_rate/action_model': 8.384372340903775e-05, 'epoch': 0.55}) +[2026-06-23 18:43:29,130][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.21628344058990479, 'train/grad_norm_pre_clip': 66.44706726074219, 'timing/data': 0.000604826956987381, 'timing/model': 1.6930088500957936, 'learning_rate/qwen_vl_interface': 8.382083423011472e-06, 'learning_rate/action_model': 8.382083423011471e-05, 'epoch': 0.55}) +[2026-06-23 18:43:32,460][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.21883854269981384, 'train/grad_norm_pre_clip': 14.504188537597656, 'timing/data': 0.0005996669642627239, 'timing/model': 1.6929564541205764, 'learning_rate/qwen_vl_interface': 8.379793217637979e-06, 'learning_rate/action_model': 8.379793217637979e-05, 'epoch': 0.55}) +[2026-06-23 18:43:35,796][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.1567610651254654, 'train/grad_norm_pre_clip': 25.100290298461914, 'timing/data': 0.0006187728140503168, 'timing/model': 1.695995552930981, 'learning_rate/qwen_vl_interface': 8.377501725724719e-06, 'learning_rate/action_model': 8.377501725724719e-05, 'epoch': 0.55}) +[2026-06-23 18:43:39,128][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.17567862570285797, 'train/grad_norm_pre_clip': 16.275920867919922, 'timing/data': 0.0006905379705131054, 'timing/model': 1.6925873369909823, 'learning_rate/qwen_vl_interface': 8.375208948213636e-06, 'learning_rate/action_model': 8.375208948213635e-05, 'epoch': 0.55}) +[2026-06-23 18:43:42,465][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.13041412830352783, 'train/grad_norm_pre_clip': 18.4560604095459, 'timing/data': 0.0006131529808044434, 'timing/model': 1.6940782240126282, 'learning_rate/qwen_vl_interface': 8.372914886047203e-06, 'learning_rate/action_model': 8.372914886047203e-05, 'epoch': 0.55}) +[2026-06-23 18:43:45,798][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.12735575437545776, 'train/grad_norm_pre_clip': 15.760881423950195, 'timing/data': 0.0006562839262187481, 'timing/model': 1.6879088960122317, 'learning_rate/qwen_vl_interface': 8.370619540168422e-06, 'learning_rate/action_model': 8.370619540168423e-05, 'epoch': 0.55}) +[2026-06-23 18:43:49,136][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.15928317606449127, 'train/grad_norm_pre_clip': 16.968637466430664, 'timing/data': 0.0006881330627948046, 'timing/model': 1.6968951581511647, 'learning_rate/qwen_vl_interface': 8.368322911520826e-06, 'learning_rate/action_model': 8.368322911520826e-05, 'epoch': 0.55}) +[2026-06-23 18:43:52,776][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.1855500340461731, 'train/grad_norm_pre_clip': 13.063762664794922, 'timing/data': 0.0006170698907226324, 'timing/model': 1.999936544103548, 'learning_rate/qwen_vl_interface': 8.36602500104847e-06, 'learning_rate/action_model': 8.36602500104847e-05, 'epoch': 0.55}) +[2026-06-23 18:43:56,110][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.1386338472366333, 'train/grad_norm_pre_clip': 32.35920333862305, 'timing/data': 0.0006959391757845879, 'timing/model': 1.6890672349836677, 'learning_rate/qwen_vl_interface': 8.36372580969594e-06, 'learning_rate/action_model': 8.36372580969594e-05, 'epoch': 0.56}) +[2026-06-23 18:43:59,446][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.26960331201553345, 'train/grad_norm_pre_clip': 14.714105606079102, 'timing/data': 0.0006310869939625263, 'timing/model': 1.6896768249571323, 'learning_rate/qwen_vl_interface': 8.361425338408344e-06, 'learning_rate/action_model': 8.361425338408343e-05, 'epoch': 0.56}) +[2026-06-23 18:44:02,770][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.2474091649055481, 'train/grad_norm_pre_clip': 23.931913375854492, 'timing/data': 0.0006046770140528679, 'timing/model': 1.6868958799168468, 'learning_rate/qwen_vl_interface': 8.359123588131319e-06, 'learning_rate/action_model': 8.359123588131319e-05, 'epoch': 0.56}) +[2026-06-23 18:44:06,106][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.26148372888565063, 'train/grad_norm_pre_clip': 55.77412796020508, 'timing/data': 0.0006104581989347935, 'timing/model': 1.6921235430054367, 'learning_rate/qwen_vl_interface': 8.35682055981103e-06, 'learning_rate/action_model': 8.35682055981103e-05, 'epoch': 0.56}) +[2026-06-23 18:44:09,445][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.11714015901088715, 'train/grad_norm_pre_clip': 63.410858154296875, 'timing/data': 0.0006044870242476463, 'timing/model': 1.6944459429942071, 'learning_rate/qwen_vl_interface': 8.354516254394163e-06, 'learning_rate/action_model': 8.354516254394162e-05, 'epoch': 0.56}) +[2026-06-23 18:44:12,780][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.2163863182067871, 'train/grad_norm_pre_clip': 52.17024230957031, 'timing/data': 0.000628290930762887, 'timing/model': 1.693723789183423, 'learning_rate/qwen_vl_interface': 8.352210672827932e-06, 'learning_rate/action_model': 8.352210672827931e-05, 'epoch': 0.56}) +[2026-06-23 18:44:16,114][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.18467575311660767, 'train/grad_norm_pre_clip': 14.11190128326416, 'timing/data': 0.0006120610050857067, 'timing/model': 1.6939493981190026, 'learning_rate/qwen_vl_interface': 8.349903816060074e-06, 'learning_rate/action_model': 8.349903816060075e-05, 'epoch': 0.56}) +[2026-06-23 18:44:19,448][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.19674868881702423, 'train/grad_norm_pre_clip': 37.49287796020508, 'timing/data': 0.000655611976981163, 'timing/model': 1.692100544925779, 'learning_rate/qwen_vl_interface': 8.34759568503885e-06, 'learning_rate/action_model': 8.347595685038851e-05, 'epoch': 0.56}) +[2026-06-23 18:44:22,789][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.19741810858249664, 'train/grad_norm_pre_clip': 28.46489143371582, 'timing/data': 0.0007058170158416033, 'timing/model': 1.7009829010348767, 'learning_rate/qwen_vl_interface': 8.345286280713049e-06, 'learning_rate/action_model': 8.345286280713048e-05, 'epoch': 0.56}) +[2026-06-23 18:44:26,125][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.05795306712388992, 'train/grad_norm_pre_clip': 35.200401306152344, 'timing/data': 0.0006743678823113441, 'timing/model': 1.6949972189031541, 'learning_rate/qwen_vl_interface': 8.342975604031974e-06, 'learning_rate/action_model': 8.342975604031974e-05, 'epoch': 0.56}) +[2026-06-23 18:44:29,461][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.27301180362701416, 'train/grad_norm_pre_clip': 34.91859436035156, 'timing/data': 0.0006029030773788691, 'timing/model': 1.6933311400935054, 'learning_rate/qwen_vl_interface': 8.340663655945462e-06, 'learning_rate/action_model': 8.340663655945462e-05, 'epoch': 0.56}) +[2026-06-23 18:44:32,793][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.150506392121315, 'train/grad_norm_pre_clip': 29.806119918823242, 'timing/data': 0.0006721641402691603, 'timing/model': 1.6901372589636594, 'learning_rate/qwen_vl_interface': 8.338350437403866e-06, 'learning_rate/action_model': 8.338350437403865e-05, 'epoch': 0.56}) +[2026-06-23 18:44:36,120][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.10468340665102005, 'train/grad_norm_pre_clip': 19.404970169067383, 'timing/data': 0.000608373899012804, 'timing/model': 1.6940457429736853, 'learning_rate/qwen_vl_interface': 8.336035949358063e-06, 'learning_rate/action_model': 8.336035949358062e-05, 'epoch': 0.56}) +[2026-06-23 18:44:39,455][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.12613414227962494, 'train/grad_norm_pre_clip': 20.412364959716797, 'timing/data': 0.0005966820754110813, 'timing/model': 1.6929509579204023, 'learning_rate/qwen_vl_interface': 8.333720192759448e-06, 'learning_rate/action_model': 8.333720192759448e-05, 'epoch': 0.56}) +[2026-06-23 18:44:43,018][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.14815407991409302, 'train/grad_norm_pre_clip': 19.928056716918945, 'timing/data': 0.0006143650971353054, 'timing/model': 1.925009628990665, 'learning_rate/qwen_vl_interface': 8.331403168559946e-06, 'learning_rate/action_model': 8.331403168559945e-05, 'epoch': 0.56}) +[2026-06-23 18:44:46,341][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.1765771061182022, 'train/grad_norm_pre_clip': 20.52215003967285, 'timing/data': 0.0006093960255384445, 'timing/model': 1.689863139996305, 'learning_rate/qwen_vl_interface': 8.329084877711995e-06, 'learning_rate/action_model': 8.329084877711996e-05, 'epoch': 0.56}) +[2026-06-23 18:44:49,694][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.14248570799827576, 'train/grad_norm_pre_clip': 62.757137298583984, 'timing/data': 0.0005849101580679417, 'timing/model': 1.7164991688914597, 'learning_rate/qwen_vl_interface': 8.32676532116856e-06, 'learning_rate/action_model': 8.326765321168559e-05, 'epoch': 0.56}) +[2026-06-23 18:44:53,028][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.20314815640449524, 'train/grad_norm_pre_clip': 46.5781364440918, 'timing/data': 0.0009056329727172852, 'timing/model': 1.692000085953623, 'learning_rate/qwen_vl_interface': 8.324444499883119e-06, 'learning_rate/action_model': 8.324444499883119e-05, 'epoch': 0.56}) +[2026-06-23 18:44:56,358][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.26448845863342285, 'train/grad_norm_pre_clip': 43.964447021484375, 'timing/data': 0.0006106281653046608, 'timing/model': 1.6889576700050384, 'learning_rate/qwen_vl_interface': 8.322122414809678e-06, 'learning_rate/action_model': 8.322122414809677e-05, 'epoch': 0.56}) +[2026-06-23 18:44:59,691][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.13997943699359894, 'train/grad_norm_pre_clip': 19.38069725036621, 'timing/data': 0.0006050881929695606, 'timing/model': 1.692996789002791, 'learning_rate/qwen_vl_interface': 8.319799066902755e-06, 'learning_rate/action_model': 8.319799066902755e-05, 'epoch': 0.56}) +[2026-06-23 18:45:03,021][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.1918024867773056, 'train/grad_norm_pre_clip': 59.46002197265625, 'timing/data': 0.000606070039793849, 'timing/model': 1.692207789979875, 'learning_rate/qwen_vl_interface': 8.317474457117393e-06, 'learning_rate/action_model': 8.317474457117392e-05, 'epoch': 0.56}) +[2026-06-23 18:45:06,352][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.18598783016204834, 'train/grad_norm_pre_clip': 37.04132843017578, 'timing/data': 0.0006704709958285093, 'timing/model': 1.6907175430096686, 'learning_rate/qwen_vl_interface': 8.315148586409148e-06, 'learning_rate/action_model': 8.315148586409148e-05, 'epoch': 0.56}) +[2026-06-23 18:45:09,689][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.17615419626235962, 'train/grad_norm_pre_clip': 28.42894172668457, 'timing/data': 0.0006144249346107244, 'timing/model': 1.6952373748645186, 'learning_rate/qwen_vl_interface': 8.312821455734104e-06, 'learning_rate/action_model': 8.312821455734103e-05, 'epoch': 0.56}) +[2026-06-23 18:45:13,022][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.23199672996997833, 'train/grad_norm_pre_clip': 48.48561096191406, 'timing/data': 0.0005974229425191879, 'timing/model': 1.6951645240187645, 'learning_rate/qwen_vl_interface': 8.31049306604885e-06, 'learning_rate/action_model': 8.310493066048849e-05, 'epoch': 0.56}) +[2026-06-23 18:45:16,356][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.23273666203022003, 'train/grad_norm_pre_clip': 67.23677062988281, 'timing/data': 0.0006110789254307747, 'timing/model': 1.694249046035111, 'learning_rate/qwen_vl_interface': 8.3081634183105e-06, 'learning_rate/action_model': 8.308163418310501e-05, 'epoch': 0.56}) +[2026-06-23 18:45:19,694][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.10993525385856628, 'train/grad_norm_pre_clip': 18.009769439697266, 'timing/data': 0.000616920180618763, 'timing/model': 1.6979858370032161, 'learning_rate/qwen_vl_interface': 8.305832513476689e-06, 'learning_rate/action_model': 8.305832513476688e-05, 'epoch': 0.57}) +[2026-06-23 18:45:23,030][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.1486583948135376, 'train/grad_norm_pre_clip': 14.703202247619629, 'timing/data': 0.0005982050206512213, 'timing/model': 1.6952177570201457, 'learning_rate/qwen_vl_interface': 8.303500352505557e-06, 'learning_rate/action_model': 8.303500352505557e-05, 'epoch': 0.57}) +[2026-06-23 18:45:26,378][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.1347528100013733, 'train/grad_norm_pre_clip': 12.713019371032715, 'timing/data': 0.0006003088783472776, 'timing/model': 1.6986920540221035, 'learning_rate/qwen_vl_interface': 8.301166936355773e-06, 'learning_rate/action_model': 8.301166936355773e-05, 'epoch': 0.57}) +[2026-06-23 18:45:29,904][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.21715518832206726, 'train/grad_norm_pre_clip': 57.39472198486328, 'timing/data': 0.0006104079075157642, 'timing/model': 1.8491807500831783, 'learning_rate/qwen_vl_interface': 8.298832265986514e-06, 'learning_rate/action_model': 8.298832265986513e-05, 'epoch': 0.57}) +[2026-06-23 18:45:33,250][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.2498537302017212, 'train/grad_norm_pre_clip': 58.158260345458984, 'timing/data': 0.0009011540096253157, 'timing/model': 1.696688855998218, 'learning_rate/qwen_vl_interface': 8.296496342357475e-06, 'learning_rate/action_model': 8.296496342357474e-05, 'epoch': 0.57}) +[2026-06-23 18:45:36,587][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.17997971177101135, 'train/grad_norm_pre_clip': 15.042487144470215, 'timing/data': 0.0006126619409769773, 'timing/model': 1.6947445760015398, 'learning_rate/qwen_vl_interface': 8.294159166428864e-06, 'learning_rate/action_model': 8.294159166428864e-05, 'epoch': 0.57}) +[2026-06-23 18:45:39,916][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.17743316292762756, 'train/grad_norm_pre_clip': 16.641742706298828, 'timing/data': 0.0005828258581459522, 'timing/model': 1.6897970768623054, 'learning_rate/qwen_vl_interface': 8.29182073916141e-06, 'learning_rate/action_model': 8.29182073916141e-05, 'epoch': 0.57}) +[2026-06-23 18:45:43,251][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.27604180574417114, 'train/grad_norm_pre_clip': 17.491703033447266, 'timing/data': 0.0007343008182942867, 'timing/model': 1.695280633168295, 'learning_rate/qwen_vl_interface': 8.28948106151635e-06, 'learning_rate/action_model': 8.28948106151635e-05, 'epoch': 0.57}) +[2026-06-23 18:45:46,580][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.2591804265975952, 'train/grad_norm_pre_clip': 51.9624137878418, 'timing/data': 0.0005838479846715927, 'timing/model': 1.688494968926534, 'learning_rate/qwen_vl_interface': 8.287140134455435e-06, 'learning_rate/action_model': 8.287140134455435e-05, 'epoch': 0.57}) +[2026-06-23 18:45:49,910][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.19146226346492767, 'train/grad_norm_pre_clip': 57.901519775390625, 'timing/data': 0.0006088849622756243, 'timing/model': 1.6900030390825123, 'learning_rate/qwen_vl_interface': 8.284797958940934e-06, 'learning_rate/action_model': 8.284797958940935e-05, 'epoch': 0.57}) +[2026-06-23 18:45:53,245][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.19904257357120514, 'train/grad_norm_pre_clip': 12.560612678527832, 'timing/data': 0.0005993470549583435, 'timing/model': 1.6948758559301496, 'learning_rate/qwen_vl_interface': 8.282454535935626e-06, 'learning_rate/action_model': 8.282454535935626e-05, 'epoch': 0.57}) +[2026-06-23 18:45:56,573][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.21425047516822815, 'train/grad_norm_pre_clip': 26.32895278930664, 'timing/data': 0.0008486660663038492, 'timing/model': 1.6888232110068202, 'learning_rate/qwen_vl_interface': 8.280109866402805e-06, 'learning_rate/action_model': 8.280109866402806e-05, 'epoch': 0.57}) +[2026-06-23 18:45:59,904][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.16780546307563782, 'train/grad_norm_pre_clip': 45.29116439819336, 'timing/data': 0.000601812032982707, 'timing/model': 1.6920372401364148, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.57}) +[2026-06-23 18:46:03,237][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.19285324215888977, 'train/grad_norm_pre_clip': 36.47686767578125, 'timing/data': 0.0006336511578410864, 'timing/model': 1.6951985789928585, 'learning_rate/qwen_vl_interface': 8.275416791610356e-06, 'learning_rate/action_model': 8.275416791610355e-05, 'epoch': 0.57}) +[2026-06-23 18:46:06,566][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.24950969219207764, 'train/grad_norm_pre_clip': 45.204219818115234, 'timing/data': 0.0006053480319678783, 'timing/model': 1.6927118180319667, 'learning_rate/qwen_vl_interface': 8.27306838827987e-06, 'learning_rate/action_model': 8.273068388279871e-05, 'epoch': 0.57}) +[2026-06-23 18:46:09,899][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.14175955951213837, 'train/grad_norm_pre_clip': 49.663726806640625, 'timing/data': 0.0006130519323050976, 'timing/model': 1.6921267230063677, 'learning_rate/qwen_vl_interface': 8.270718742280162e-06, 'learning_rate/action_model': 8.270718742280162e-05, 'epoch': 0.57}) +[2026-06-23 18:46:13,230][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.16866087913513184, 'train/grad_norm_pre_clip': 12.54663372039795, 'timing/data': 0.0008293990977108479, 'timing/model': 1.6905814020428807, 'learning_rate/qwen_vl_interface': 8.268367854577084e-06, 'learning_rate/action_model': 8.268367854577082e-05, 'epoch': 0.57}) +[2026-06-23 18:46:16,560][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.17227621376514435, 'train/grad_norm_pre_clip': 35.772735595703125, 'timing/data': 0.0005824652034789324, 'timing/model': 1.6878265580162406, 'learning_rate/qwen_vl_interface': 8.266015726136991e-06, 'learning_rate/action_model': 8.266015726136991e-05, 'epoch': 0.57}) +[2026-06-23 18:46:20,035][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.2250853031873703, 'train/grad_norm_pre_clip': 27.611093521118164, 'timing/data': 0.0006006499752402306, 'timing/model': 1.6923326549585909, 'learning_rate/qwen_vl_interface': 8.26366235792676e-06, 'learning_rate/action_model': 8.26366235792676e-05, 'epoch': 0.57}) +[2026-06-23 18:46:23,362][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.1509629338979721, 'train/grad_norm_pre_clip': 14.317357063293457, 'timing/data': 0.0006082740146666765, 'timing/model': 1.6816424501594156, 'learning_rate/qwen_vl_interface': 8.261307750913769e-06, 'learning_rate/action_model': 8.261307750913769e-05, 'epoch': 0.57}) +[2026-06-23 18:46:26,689][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.1881597489118576, 'train/grad_norm_pre_clip': 34.631385803222656, 'timing/data': 0.0006141040939837694, 'timing/model': 1.6914882780984044, 'learning_rate/qwen_vl_interface': 8.258951906065908e-06, 'learning_rate/action_model': 8.258951906065908e-05, 'epoch': 0.57}) +[2026-06-23 18:46:30,020][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.1117786318063736, 'train/grad_norm_pre_clip': 8.49523639678955, 'timing/data': 0.0006737669464200735, 'timing/model': 1.691773439059034, 'learning_rate/qwen_vl_interface': 8.256594824351577e-06, 'learning_rate/action_model': 8.256594824351576e-05, 'epoch': 0.57}) +[2026-06-23 18:46:33,348][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.21232475340366364, 'train/grad_norm_pre_clip': 43.76485824584961, 'timing/data': 0.0008305308874696493, 'timing/model': 1.6913514209445566, 'learning_rate/qwen_vl_interface': 8.254236506739683e-06, 'learning_rate/action_model': 8.254236506739681e-05, 'epoch': 0.57}) +[2026-06-23 18:46:36,675][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.1188914105296135, 'train/grad_norm_pre_clip': 28.9919490814209, 'timing/data': 0.000599568011239171, 'timing/model': 1.685541212093085, 'learning_rate/qwen_vl_interface': 8.25187695419964e-06, 'learning_rate/action_model': 8.25187695419964e-05, 'epoch': 0.57}) +[2026-06-23 18:46:40,005][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.10973570495843887, 'train/grad_norm_pre_clip': 19.594486236572266, 'timing/data': 0.0008306719828397036, 'timing/model': 1.6898561811540276, 'learning_rate/qwen_vl_interface': 8.249516167701373e-06, 'learning_rate/action_model': 8.249516167701374e-05, 'epoch': 0.57}) +[2026-06-23 18:46:43,335][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.1418183445930481, 'train/grad_norm_pre_clip': 23.094724655151367, 'timing/data': 0.0006152070127427578, 'timing/model': 1.692241481039673, 'learning_rate/qwen_vl_interface': 8.247154148215311e-06, 'learning_rate/action_model': 8.24715414821531e-05, 'epoch': 0.57}) +[2026-06-23 18:46:46,666][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.12655772268772125, 'train/grad_norm_pre_clip': 37.72119140625, 'timing/data': 0.0005954490043222904, 'timing/model': 1.6927770331967622, 'learning_rate/qwen_vl_interface': 8.24479089671239e-06, 'learning_rate/action_model': 8.24479089671239e-05, 'epoch': 0.58}) +[2026-06-23 18:46:49,997][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.17775021493434906, 'train/grad_norm_pre_clip': 111.15675354003906, 'timing/data': 0.0005899488460272551, 'timing/model': 1.691151682054624, 'learning_rate/qwen_vl_interface': 8.242426414164056e-06, 'learning_rate/action_model': 8.242426414164056e-05, 'epoch': 0.58}) +[2026-06-23 18:46:53,324][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.14430055022239685, 'train/grad_norm_pre_clip': 20.48688507080078, 'timing/data': 0.0006234119646251202, 'timing/model': 1.6904862660448998, 'learning_rate/qwen_vl_interface': 8.240060701542257e-06, 'learning_rate/action_model': 8.240060701542257e-05, 'epoch': 0.58}) +[2026-06-23 18:46:56,656][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.16914910078048706, 'train/grad_norm_pre_clip': 67.13086700439453, 'timing/data': 0.000676862895488739, 'timing/model': 1.6932238491717726, 'learning_rate/qwen_vl_interface': 8.237693759819447e-06, 'learning_rate/action_model': 8.237693759819448e-05, 'epoch': 0.58}) +[2026-06-23 18:46:59,987][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.04988561198115349, 'train/grad_norm_pre_clip': 33.055538177490234, 'timing/data': 0.0006719529628753662, 'timing/model': 1.6924584361258894, 'learning_rate/qwen_vl_interface': 8.235325589968587e-06, 'learning_rate/action_model': 8.235325589968587e-05, 'epoch': 0.58}) +[2026-06-23 18:47:03,312][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.24922749400138855, 'train/grad_norm_pre_clip': 15.071738243103027, 'timing/data': 0.0006091359537094831, 'timing/model': 1.6892191010992974, 'learning_rate/qwen_vl_interface': 8.232956192963143e-06, 'learning_rate/action_model': 8.232956192963144e-05, 'epoch': 0.58}) +[2026-06-23 18:47:06,647][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.08722318708896637, 'train/grad_norm_pre_clip': 43.45369338989258, 'timing/data': 0.0006054290570318699, 'timing/model': 1.6927125151269138, 'learning_rate/qwen_vl_interface': 8.230585569777086e-06, 'learning_rate/action_model': 8.230585569777085e-05, 'epoch': 0.58}) +[2026-06-23 18:47:10,267][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.23778700828552246, 'train/grad_norm_pre_clip': 52.14543533325195, 'timing/data': 0.0006165090017020702, 'timing/model': 1.6844802810810506, 'learning_rate/qwen_vl_interface': 8.228213721384885e-06, 'learning_rate/action_model': 8.228213721384885e-05, 'epoch': 0.58}) +[2026-06-23 18:47:13,602][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.1533236801624298, 'train/grad_norm_pre_clip': 30.445392608642578, 'timing/data': 0.0006116600707173347, 'timing/model': 1.6946869778912514, 'learning_rate/qwen_vl_interface': 8.225840648761523e-06, 'learning_rate/action_model': 8.225840648761523e-05, 'epoch': 0.58}) +[2026-06-23 18:47:16,939][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.077060267329216, 'train/grad_norm_pre_clip': 14.597467422485352, 'timing/data': 0.000600077910348773, 'timing/model': 1.6935704939533025, 'learning_rate/qwen_vl_interface': 8.223466352882476e-06, 'learning_rate/action_model': 8.223466352882475e-05, 'epoch': 0.58}) +[2026-06-23 18:47:20,275][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.1354246586561203, 'train/grad_norm_pre_clip': 43.64910125732422, 'timing/data': 0.0006088749505579472, 'timing/model': 1.6937894020229578, 'learning_rate/qwen_vl_interface': 8.22109083472373e-06, 'learning_rate/action_model': 8.22109083472373e-05, 'epoch': 0.58}) +[2026-06-23 18:47:23,604][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.1266471892595291, 'train/grad_norm_pre_clip': 50.009178161621094, 'timing/data': 0.0006075720302760601, 'timing/model': 1.6927635679021478, 'learning_rate/qwen_vl_interface': 8.218714095261769e-06, 'learning_rate/action_model': 8.218714095261768e-05, 'epoch': 0.58}) +[2026-06-23 18:47:26,940][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.11037560552358627, 'train/grad_norm_pre_clip': 63.22458267211914, 'timing/data': 0.0006684870459139347, 'timing/model': 1.6951342821121216, 'learning_rate/qwen_vl_interface': 8.21633613547358e-06, 'learning_rate/action_model': 8.21633613547358e-05, 'epoch': 0.58}) +[2026-06-23 18:47:30,278][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.13852903246879578, 'train/grad_norm_pre_clip': 24.88720703125, 'timing/data': 0.0006608730182051659, 'timing/model': 1.6949048358947039, 'learning_rate/qwen_vl_interface': 8.213956956336656e-06, 'learning_rate/action_model': 8.213956956336656e-05, 'epoch': 0.58}) +[2026-06-23 18:47:33,604][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1500 +[2026-06-23 18:49:06,770][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1500 +[2026-06-23 18:49:06,771][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1500 +[2026-06-23 18:50:18,614][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1500 +[2026-06-23 18:50:40,767][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500 +[2026-06-23 18:50:40,767][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.11410384625196457, 'train/grad_norm_pre_clip': 10.198287010192871, 'eval/action_loss/seconds': 93.16400029906072, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.09083063155412674, 'eval/latency_2/loss': 0.09083063155412674, 'eval/flappy/loss': 0.09083063155412674, 'eval/flappy/latency_2/loss': 0.09083063155412674, 'eval/flap_precision': 0.9334748761500354, 'eval/flap_recall': 0.8900134952766532, 'eval/flap_f1': 0.9112262521588946, 'eval/flap_rate_model': 0.19625, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9112262521588946, 'eval/latency_2/flap_precision': 0.9334748761500354, 'eval/latency_2/flap_recall': 0.8900134952766532, 'eval/latency_2/flap_f1': 0.9112262521588946, 'eval/latency_2/flap_rate_model': 0.19625, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9112262521588946, 'eval/action_classification/data_wait_seconds': 0.9107963442802429, 'eval/action_classification/predict_seconds': 70.79410552978516, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 14.900000000000025, 'rl_games_eval/mid_train/mean_length': 138.4, 'rl_games_eval/mid_train/std_reward': 8.573214099741168, 'rl_games_eval/mid_train/std_length': 69.04085746860333, 'rl_games_eval/mid_train/macro_mean_reward': 14.900000000000025, 'rl_games_eval/mid_train/macro_mean_length': 138.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 14.900000000000025, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 138.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 8.573214099741168, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 69.04085746860333, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 14.900000000000025, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0008020778186619282, 'timing/model': 1.6916296239942312, 'learning_rate/qwen_vl_interface': 8.211576558828985e-06, 'learning_rate/action_model': 8.211576558828985e-05, 'epoch': 0.58}) +[2026-06-23 18:50:40,768][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_1500_state +[2026-06-23 18:50:40,771][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 18:50:43,419][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_1500_state/model.safetensors +[2026-06-23 18:50:44,575][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_1500_state/optimizer.bin +[2026-06-23 18:50:44,575][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_1500_state/sampler.bin +[2026-06-23 18:50:44,575][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_1500_state/sampler_1.bin +[2026-06-23 18:50:44,576][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_1500_state/random_states_0.pkl +[2026-06-23 18:50:44,601][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500 +[2026-06-23 18:50:44,601][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 18:50:44,611][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 18:50:47,912][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.2680146396160126, 'train/grad_norm_pre_clip': 30.923988342285156, 'timing/data': 0.000601600855588913, 'timing/model': 1.675187790999189, 'learning_rate/qwen_vl_interface': 8.209194943929059e-06, 'learning_rate/action_model': 8.20919494392906e-05, 'epoch': 0.58}) +[2026-06-23 18:50:51,206][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.141452819108963, 'train/grad_norm_pre_clip': 27.13105010986328, 'timing/data': 0.0007031820714473724, 'timing/model': 1.6754826919641346, 'learning_rate/qwen_vl_interface': 8.20681211261587e-06, 'learning_rate/action_model': 8.20681211261587e-05, 'epoch': 0.58}) +[2026-06-23 18:50:54,506][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.1124492660164833, 'train/grad_norm_pre_clip': 27.71666717529297, 'timing/data': 0.0007975390180945396, 'timing/model': 1.6776451850309968, 'learning_rate/qwen_vl_interface': 8.204428065868911e-06, 'learning_rate/action_model': 8.204428065868911e-05, 'epoch': 0.58}) +[2026-06-23 18:50:57,813][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.21271473169326782, 'train/grad_norm_pre_clip': 25.707475662231445, 'timing/data': 0.0006551810074597597, 'timing/model': 1.6801807191222906, 'learning_rate/qwen_vl_interface': 8.202042804668173e-06, 'learning_rate/action_model': 8.202042804668172e-05, 'epoch': 0.58}) +[2026-06-23 18:51:01,125][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.1823933869600296, 'train/grad_norm_pre_clip': 71.83769226074219, 'timing/data': 0.0009064949117600918, 'timing/model': 1.6836844941135496, 'learning_rate/qwen_vl_interface': 8.199656329994145e-06, 'learning_rate/action_model': 8.199656329994145e-05, 'epoch': 0.58}) +[2026-06-23 18:51:04,437][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.10410548746585846, 'train/grad_norm_pre_clip': 70.37672424316406, 'timing/data': 0.0007007280364632607, 'timing/model': 1.6818228110205382, 'learning_rate/qwen_vl_interface': 8.19726864282782e-06, 'learning_rate/action_model': 8.197268642827818e-05, 'epoch': 0.58}) +[2026-06-23 18:51:07,753][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.18886996805667877, 'train/grad_norm_pre_clip': 54.4235725402832, 'timing/data': 0.0006751788314431906, 'timing/model': 1.6840887789148837, 'learning_rate/qwen_vl_interface': 8.194879744150682e-06, 'learning_rate/action_model': 8.194879744150681e-05, 'epoch': 0.58}) +[2026-06-23 18:51:11,077][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.41723138093948364, 'train/grad_norm_pre_clip': 24.61347007751465, 'timing/data': 0.0006236820481717587, 'timing/model': 1.6896593789570034, 'learning_rate/qwen_vl_interface': 8.192489634944718e-06, 'learning_rate/action_model': 8.192489634944718e-05, 'epoch': 0.58}) +[2026-06-23 18:51:14,392][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.20827513933181763, 'train/grad_norm_pre_clip': 77.00430297851562, 'timing/data': 0.0006797478999942541, 'timing/model': 1.6813402748666704, 'learning_rate/qwen_vl_interface': 8.190098316192416e-06, 'learning_rate/action_model': 8.190098316192416e-05, 'epoch': 0.58}) +[2026-06-23 18:51:17,712][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.17905208468437195, 'train/grad_norm_pre_clip': 79.80130767822266, 'timing/data': 0.0007111669983714819, 'timing/model': 1.6836762039456517, 'learning_rate/qwen_vl_interface': 8.18770578887675e-06, 'learning_rate/action_model': 8.18770578887675e-05, 'epoch': 0.58}) +[2026-06-23 18:51:21,036][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.15297450125217438, 'train/grad_norm_pre_clip': 54.21612548828125, 'timing/data': 0.0006699101068079472, 'timing/model': 1.6861264989711344, 'learning_rate/qwen_vl_interface': 8.185312053981204e-06, 'learning_rate/action_model': 8.185312053981204e-05, 'epoch': 0.58}) +[2026-06-23 18:51:24,358][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.08805392682552338, 'train/grad_norm_pre_clip': 45.12770080566406, 'timing/data': 0.0006281910464167595, 'timing/model': 1.6848243679851294, 'learning_rate/qwen_vl_interface': 8.182917112489747e-06, 'learning_rate/action_model': 8.182917112489746e-05, 'epoch': 0.59}) +[2026-06-23 18:51:27,686][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.18556079268455505, 'train/grad_norm_pre_clip': 13.669029235839844, 'timing/data': 0.0007558311335742474, 'timing/model': 1.6894163580145687, 'learning_rate/qwen_vl_interface': 8.18052096538685e-06, 'learning_rate/action_model': 8.18052096538685e-05, 'epoch': 0.59}) +[2026-06-23 18:51:31,113][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.1454005241394043, 'train/grad_norm_pre_clip': 75.00626373291016, 'timing/data': 0.0006932930555194616, 'timing/model': 1.6893985718488693, 'learning_rate/qwen_vl_interface': 8.178123613657483e-06, 'learning_rate/action_model': 8.178123613657482e-05, 'epoch': 0.59}) +[2026-06-23 18:51:34,439][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.13585180044174194, 'train/grad_norm_pre_clip': 38.210445404052734, 'timing/data': 0.0006937950383871794, 'timing/model': 1.6836490998975933, 'learning_rate/qwen_vl_interface': 8.175725058287102e-06, 'learning_rate/action_model': 8.175725058287102e-05, 'epoch': 0.59}) +[2026-06-23 18:51:37,771][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.2100282460451126, 'train/grad_norm_pre_clip': 60.627197265625, 'timing/data': 0.0006579270120710135, 'timing/model': 1.6930511039681733, 'learning_rate/qwen_vl_interface': 8.173325300261664e-06, 'learning_rate/action_model': 8.173325300261663e-05, 'epoch': 0.59}) +[2026-06-23 18:51:41,095][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.08362463116645813, 'train/grad_norm_pre_clip': 15.439934730529785, 'timing/data': 0.0007035320159047842, 'timing/model': 1.6822534708771855, 'learning_rate/qwen_vl_interface': 8.170924340567617e-06, 'learning_rate/action_model': 8.170924340567616e-05, 'epoch': 0.59}) +[2026-06-23 18:51:44,424][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.08950906246900558, 'train/grad_norm_pre_clip': 20.861358642578125, 'timing/data': 0.0006137241143733263, 'timing/model': 1.691614928888157, 'learning_rate/qwen_vl_interface': 8.168522180191909e-06, 'learning_rate/action_model': 8.168522180191909e-05, 'epoch': 0.59}) +[2026-06-23 18:51:47,764][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.22300417721271515, 'train/grad_norm_pre_clip': 85.81884765625, 'timing/data': 0.0006211879663169384, 'timing/model': 1.6885340139269829, 'learning_rate/qwen_vl_interface': 8.166118820121972e-06, 'learning_rate/action_model': 8.166118820121973e-05, 'epoch': 0.59}) +[2026-06-23 18:51:51,094][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.17111694812774658, 'train/grad_norm_pre_clip': 54.06967544555664, 'timing/data': 0.0006076118443161249, 'timing/model': 1.6953091898467392, 'learning_rate/qwen_vl_interface': 8.163714261345744e-06, 'learning_rate/action_model': 8.163714261345742e-05, 'epoch': 0.59}) +[2026-06-23 18:51:54,428][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.3012545704841614, 'train/grad_norm_pre_clip': 41.99447250366211, 'timing/data': 0.0006005191244184971, 'timing/model': 1.6950635248795152, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.59}) +[2026-06-23 18:51:57,763][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.1499340832233429, 'train/grad_norm_pre_clip': 28.31502914428711, 'timing/data': 0.0007813491392880678, 'timing/model': 1.6925919449422508, 'learning_rate/qwen_vl_interface': 8.158901551628579e-06, 'learning_rate/action_model': 8.158901551628579e-05, 'epoch': 0.59}) +[2026-06-23 18:52:01,100][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.11837919056415558, 'train/grad_norm_pre_clip': 35.328792572021484, 'timing/data': 0.0006055280100554228, 'timing/model': 1.6948348588775843, 'learning_rate/qwen_vl_interface': 8.156493402665971e-06, 'learning_rate/action_model': 8.15649340266597e-05, 'epoch': 0.59}) +[2026-06-23 18:52:04,441][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.19444963335990906, 'train/grad_norm_pre_clip': 38.29113006591797, 'timing/data': 0.0006139641627669334, 'timing/model': 1.6943688609171659, 'learning_rate/qwen_vl_interface': 8.154084058953713e-06, 'learning_rate/action_model': 8.154084058953713e-05, 'epoch': 0.59}) +[2026-06-23 18:52:07,771][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.2374792844057083, 'train/grad_norm_pre_clip': 37.65034103393555, 'timing/data': 0.0006168701220303774, 'timing/model': 1.69251555390656, 'learning_rate/qwen_vl_interface': 8.151673521482193e-06, 'learning_rate/action_model': 8.151673521482194e-05, 'epoch': 0.59}) +[2026-06-23 18:52:11,110][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.16969701647758484, 'train/grad_norm_pre_clip': 12.737442970275879, 'timing/data': 0.0006047268398106098, 'timing/model': 1.6948248140979558, 'learning_rate/qwen_vl_interface': 8.149261791242297e-06, 'learning_rate/action_model': 8.149261791242297e-05, 'epoch': 0.59}) +[2026-06-23 18:52:14,443][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.14147602021694183, 'train/grad_norm_pre_clip': 42.893585205078125, 'timing/data': 0.0006005088798701763, 'timing/model': 1.692301396979019, 'learning_rate/qwen_vl_interface': 8.146848869225391e-06, 'learning_rate/action_model': 8.146848869225391e-05, 'epoch': 0.59}) +[2026-06-23 18:52:17,873][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.18395835161209106, 'train/grad_norm_pre_clip': 59.98493957519531, 'timing/data': 0.0006017610430717468, 'timing/model': 1.7884763251058757, 'learning_rate/qwen_vl_interface': 8.14443475642334e-06, 'learning_rate/action_model': 8.144434756423339e-05, 'epoch': 0.59}) +[2026-06-23 18:52:21,207][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.18487600982189178, 'train/grad_norm_pre_clip': 41.49809265136719, 'timing/data': 0.0006136840675026178, 'timing/model': 1.6962811939883977, 'learning_rate/qwen_vl_interface': 8.142019453828493e-06, 'learning_rate/action_model': 8.142019453828493e-05, 'epoch': 0.59}) +[2026-06-23 18:52:24,544][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.17623910307884216, 'train/grad_norm_pre_clip': 16.253509521484375, 'timing/data': 0.0005994369275867939, 'timing/model': 1.6892079229000956, 'learning_rate/qwen_vl_interface': 8.139602962433688e-06, 'learning_rate/action_model': 8.139602962433688e-05, 'epoch': 0.59}) +[2026-06-23 18:52:27,880][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.101686030626297, 'train/grad_norm_pre_clip': 33.8680534362793, 'timing/data': 0.0006633768789470196, 'timing/model': 1.6944837011396885, 'learning_rate/qwen_vl_interface': 8.137185283232255e-06, 'learning_rate/action_model': 8.137185283232256e-05, 'epoch': 0.59}) +[2026-06-23 18:52:31,218][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.13733309507369995, 'train/grad_norm_pre_clip': 47.59716033935547, 'timing/data': 0.0006612930446863174, 'timing/model': 1.6950894971378148, 'learning_rate/qwen_vl_interface': 8.134766417218014e-06, 'learning_rate/action_model': 8.134766417218012e-05, 'epoch': 0.59}) +[2026-06-23 18:52:34,552][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.2124207466840744, 'train/grad_norm_pre_clip': 23.223997116088867, 'timing/data': 0.000621297862380743, 'timing/model': 1.69214246282354, 'learning_rate/qwen_vl_interface': 8.132346365385262e-06, 'learning_rate/action_model': 8.132346365385261e-05, 'epoch': 0.59}) +[2026-06-23 18:52:37,888][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.09771984815597534, 'train/grad_norm_pre_clip': 14.499031066894531, 'timing/data': 0.0006233819294720888, 'timing/model': 1.6913933330215514, 'learning_rate/qwen_vl_interface': 8.129925128728796e-06, 'learning_rate/action_model': 8.129925128728796e-05, 'epoch': 0.59}) +[2026-06-23 18:52:41,229][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.14628131687641144, 'train/grad_norm_pre_clip': 15.838179588317871, 'timing/data': 0.000615506898611784, 'timing/model': 1.6960256001912057, 'learning_rate/qwen_vl_interface': 8.127502708243895e-06, 'learning_rate/action_model': 8.127502708243895e-05, 'epoch': 0.59}) +[2026-06-23 18:52:44,562][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.11043287068605423, 'train/grad_norm_pre_clip': 15.102163314819336, 'timing/data': 0.000617471057921648, 'timing/model': 1.6929215909913182, 'learning_rate/qwen_vl_interface': 8.125079104926324e-06, 'learning_rate/action_model': 8.125079104926323e-05, 'epoch': 0.59}) +[2026-06-23 18:52:47,902][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.14229021966457367, 'train/grad_norm_pre_clip': 19.825847625732422, 'timing/data': 0.0006051380187273026, 'timing/model': 1.6956959159579128, 'learning_rate/qwen_vl_interface': 8.122654319772332e-06, 'learning_rate/action_model': 8.122654319772332e-05, 'epoch': 0.59}) +[2026-06-23 18:52:51,245][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.14081084728240967, 'train/grad_norm_pre_clip': 14.94082260131836, 'timing/data': 0.0006456130649894476, 'timing/model': 1.6980027391109616, 'learning_rate/qwen_vl_interface': 8.120228353778661e-06, 'learning_rate/action_model': 8.12022835377866e-05, 'epoch': 0.6}) +[2026-06-23 18:52:54,590][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.18270881474018097, 'train/grad_norm_pre_clip': 17.852336883544922, 'timing/data': 0.0005871839821338654, 'timing/model': 1.6987502730917186, 'learning_rate/qwen_vl_interface': 8.117801207942528e-06, 'learning_rate/action_model': 8.117801207942528e-05, 'epoch': 0.6}) +[2026-06-23 18:52:57,938][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.2370651513338089, 'train/grad_norm_pre_clip': 25.09639549255371, 'timing/data': 0.0006054979749023914, 'timing/model': 1.7015575889963657, 'learning_rate/qwen_vl_interface': 8.115372883261648e-06, 'learning_rate/action_model': 8.115372883261647e-05, 'epoch': 0.6}) +[2026-06-23 18:53:01,275][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.06327975541353226, 'train/grad_norm_pre_clip': 23.88884925842285, 'timing/data': 0.0006350339390337467, 'timing/model': 1.6923971951473504, 'learning_rate/qwen_vl_interface': 8.11294338073421e-06, 'learning_rate/action_model': 8.11294338073421e-05, 'epoch': 0.6}) +[2026-06-23 18:53:04,614][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.48026829957962036, 'train/grad_norm_pre_clip': 50.163612365722656, 'timing/data': 0.0006110689137130976, 'timing/model': 1.695144196972251, 'learning_rate/qwen_vl_interface': 8.110512701358892e-06, 'learning_rate/action_model': 8.110512701358891e-05, 'epoch': 0.6}) +[2026-06-23 18:53:08,091][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.18893063068389893, 'train/grad_norm_pre_clip': 84.5499038696289, 'timing/data': 0.0006182021461427212, 'timing/model': 1.837207644013688, 'learning_rate/qwen_vl_interface': 8.108080846134852e-06, 'learning_rate/action_model': 8.108080846134852e-05, 'epoch': 0.6}) +[2026-06-23 18:53:11,433][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.14511118829250336, 'train/grad_norm_pre_clip': 46.0335578918457, 'timing/data': 0.0006115289870649576, 'timing/model': 1.7002504719421268, 'learning_rate/qwen_vl_interface': 8.105647816061734e-06, 'learning_rate/action_model': 8.105647816061735e-05, 'epoch': 0.6}) +[2026-06-23 18:53:14,770][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.21879462897777557, 'train/grad_norm_pre_clip': 39.168148040771484, 'timing/data': 0.0006527469959110022, 'timing/model': 1.6914444849826396, 'learning_rate/qwen_vl_interface': 8.103213612139667e-06, 'learning_rate/action_model': 8.103213612139667e-05, 'epoch': 0.6}) +[2026-06-23 18:53:18,111][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.10229496657848358, 'train/grad_norm_pre_clip': 81.19358825683594, 'timing/data': 0.0007839140016585588, 'timing/model': 1.6935895660426468, 'learning_rate/qwen_vl_interface': 8.100778235369262e-06, 'learning_rate/action_model': 8.10077823536926e-05, 'epoch': 0.6}) +[2026-06-23 18:53:21,448][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.16998977959156036, 'train/grad_norm_pre_clip': 75.19132995605469, 'timing/data': 0.0007089029531925917, 'timing/model': 1.69460270809941, 'learning_rate/qwen_vl_interface': 8.098341686751605e-06, 'learning_rate/action_model': 8.098341686751605e-05, 'epoch': 0.6}) +[2026-06-23 18:53:24,788][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.1477362960577011, 'train/grad_norm_pre_clip': 67.56121826171875, 'timing/data': 0.0006457839626818895, 'timing/model': 1.6963301529176533, 'learning_rate/qwen_vl_interface': 8.095903967288273e-06, 'learning_rate/action_model': 8.095903967288272e-05, 'epoch': 0.6}) +[2026-06-23 18:53:28,136][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.18171745538711548, 'train/grad_norm_pre_clip': 48.771583557128906, 'timing/data': 0.0006686670240014791, 'timing/model': 1.6981735171284527, 'learning_rate/qwen_vl_interface': 8.093465077981318e-06, 'learning_rate/action_model': 8.093465077981318e-05, 'epoch': 0.6}) +[2026-06-23 18:53:31,478][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.09159979969263077, 'train/grad_norm_pre_clip': 16.725753784179688, 'timing/data': 0.0006632967852056026, 'timing/model': 1.69775929604657, 'learning_rate/qwen_vl_interface': 8.091025019833276e-06, 'learning_rate/action_model': 8.091025019833276e-05, 'epoch': 0.6}) +[2026-06-23 18:53:34,815][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.14519721269607544, 'train/grad_norm_pre_clip': 26.280975341796875, 'timing/data': 0.000594958895817399, 'timing/model': 1.6958856009878218, 'learning_rate/qwen_vl_interface': 8.088583793847165e-06, 'learning_rate/action_model': 8.088583793847164e-05, 'epoch': 0.6}) +[2026-06-23 18:53:38,157][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.13482733070850372, 'train/grad_norm_pre_clip': 53.382259368896484, 'timing/data': 0.0006752901244908571, 'timing/model': 1.6991008140612394, 'learning_rate/qwen_vl_interface': 8.086141401026478e-06, 'learning_rate/action_model': 8.086141401026478e-05, 'epoch': 0.6}) +[2026-06-23 18:53:41,495][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.139752596616745, 'train/grad_norm_pre_clip': 92.42357635498047, 'timing/data': 0.0006904080510139465, 'timing/model': 1.6930348679888994, 'learning_rate/qwen_vl_interface': 8.083697842375193e-06, 'learning_rate/action_model': 8.083697842375193e-05, 'epoch': 0.6}) +[2026-06-23 18:53:44,840][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.13466505706310272, 'train/grad_norm_pre_clip': 44.172027587890625, 'timing/data': 0.0007151039317250252, 'timing/model': 1.6995758530683815, 'learning_rate/qwen_vl_interface': 8.081253118897762e-06, 'learning_rate/action_model': 8.081253118897761e-05, 'epoch': 0.6}) +[2026-06-23 18:53:48,180][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.2765686511993408, 'train/grad_norm_pre_clip': 29.94049835205078, 'timing/data': 0.0005954699590802193, 'timing/model': 1.6928438250906765, 'learning_rate/qwen_vl_interface': 8.07880723159912e-06, 'learning_rate/action_model': 8.07880723159912e-05, 'epoch': 0.6}) +[2026-06-23 18:53:51,523][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.15284353494644165, 'train/grad_norm_pre_clip': 71.88162994384766, 'timing/data': 0.0007101150695234537, 'timing/model': 1.7012591240927577, 'learning_rate/qwen_vl_interface': 8.076360181484676e-06, 'learning_rate/action_model': 8.076360181484676e-05, 'epoch': 0.6}) +[2026-06-23 18:53:54,861][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.06135861948132515, 'train/grad_norm_pre_clip': 26.754138946533203, 'timing/data': 0.0007798459846526384, 'timing/model': 1.6922862029168755, 'learning_rate/qwen_vl_interface': 8.073911969560326e-06, 'learning_rate/action_model': 8.073911969560326e-05, 'epoch': 0.6}) +[2026-06-23 18:53:58,330][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.19685737788677216, 'train/grad_norm_pre_clip': 66.22039031982422, 'timing/data': 0.0007319459691643715, 'timing/model': 1.6918199949432164, 'learning_rate/qwen_vl_interface': 8.071462596832432e-06, 'learning_rate/action_model': 8.071462596832432e-05, 'epoch': 0.6}) +[2026-06-23 18:54:01,665][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.19429147243499756, 'train/grad_norm_pre_clip': 44.210418701171875, 'timing/data': 0.000708320876583457, 'timing/model': 1.6932930808980018, 'learning_rate/qwen_vl_interface': 8.06901206430784e-06, 'learning_rate/action_model': 8.06901206430784e-05, 'epoch': 0.6}) +[2026-06-23 18:54:04,999][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.09333590418100357, 'train/grad_norm_pre_clip': 8.792957305908203, 'timing/data': 0.0006845667958259583, 'timing/model': 1.6946573939640075, 'learning_rate/qwen_vl_interface': 8.066560372993873e-06, 'learning_rate/action_model': 8.066560372993874e-05, 'epoch': 0.6}) +[2026-06-23 18:54:08,331][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.1876826286315918, 'train/grad_norm_pre_clip': 14.703424453735352, 'timing/data': 0.000676190946251154, 'timing/model': 1.6914244038052857, 'learning_rate/qwen_vl_interface': 8.064107523898327e-06, 'learning_rate/action_model': 8.064107523898326e-05, 'epoch': 0.6}) +[2026-06-23 18:54:11,669][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.10888227820396423, 'train/grad_norm_pre_clip': 21.06229591369629, 'timing/data': 0.00065756612457335, 'timing/model': 1.6926592839881778, 'learning_rate/qwen_vl_interface': 8.061653518029476e-06, 'learning_rate/action_model': 8.061653518029476e-05, 'epoch': 0.6}) +[2026-06-23 18:54:15,006][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.27993306517601013, 'train/grad_norm_pre_clip': 73.17049407958984, 'timing/data': 0.0006753590423613787, 'timing/model': 1.6960096301045269, 'learning_rate/qwen_vl_interface': 8.059198356396069e-06, 'learning_rate/action_model': 8.059198356396068e-05, 'epoch': 0.6}) +[2026-06-23 18:54:18,348][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.15076076984405518, 'train/grad_norm_pre_clip': 19.609474182128906, 'timing/data': 0.0006709920708090067, 'timing/model': 1.6981393520254642, 'learning_rate/qwen_vl_interface': 8.05674204000733e-06, 'learning_rate/action_model': 8.05674204000733e-05, 'epoch': 0.61}) +[2026-06-23 18:54:21,684][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.271602600812912, 'train/grad_norm_pre_clip': 13.583707809448242, 'timing/data': 0.0007427260279655457, 'timing/model': 1.6920148369390517, 'learning_rate/qwen_vl_interface': 8.05428456987296e-06, 'learning_rate/action_model': 8.05428456987296e-05, 'epoch': 0.61}) +[2026-06-23 18:54:25,018][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.26555150747299194, 'train/grad_norm_pre_clip': 19.70496940612793, 'timing/data': 0.0006712020840495825, 'timing/model': 1.6946291599888355, 'learning_rate/qwen_vl_interface': 8.05182594700313e-06, 'learning_rate/action_model': 8.05182594700313e-05, 'epoch': 0.61}) +[2026-06-23 18:54:28,359][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.1536233127117157, 'train/grad_norm_pre_clip': 28.3830509185791, 'timing/data': 0.0006620138883590698, 'timing/model': 1.6944078519009054, 'learning_rate/qwen_vl_interface': 8.049366172408484e-06, 'learning_rate/action_model': 8.049366172408485e-05, 'epoch': 0.61}) +[2026-06-23 18:54:31,702][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.20450276136398315, 'train/grad_norm_pre_clip': 22.850862503051758, 'timing/data': 0.0006528568919748068, 'timing/model': 1.6983309760689735, 'learning_rate/qwen_vl_interface': 8.04690524710015e-06, 'learning_rate/action_model': 8.046905247100151e-05, 'epoch': 0.61}) +[2026-06-23 18:54:35,042][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.20332521200180054, 'train/grad_norm_pre_clip': 21.718944549560547, 'timing/data': 0.0005919330287724733, 'timing/model': 1.6966457250528038, 'learning_rate/qwen_vl_interface': 8.044443172089715e-06, 'learning_rate/action_model': 8.044443172089714e-05, 'epoch': 0.61}) +[2026-06-23 18:54:38,381][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.18393217027187347, 'train/grad_norm_pre_clip': 11.041496276855469, 'timing/data': 0.0006328399758785963, 'timing/model': 1.695852833101526, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.61}) +[2026-06-23 18:54:41,731][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.2686671316623688, 'train/grad_norm_pre_clip': 16.497539520263672, 'timing/data': 0.0006412358488887548, 'timing/model': 1.6969505350571126, 'learning_rate/qwen_vl_interface': 8.039515577011287e-06, 'learning_rate/action_model': 8.039515577011288e-05, 'epoch': 0.61}) +[2026-06-23 18:54:45,076][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.22074878215789795, 'train/grad_norm_pre_clip': 54.89733123779297, 'timing/data': 0.0007229899056255817, 'timing/model': 1.6994135731365532, 'learning_rate/qwen_vl_interface': 8.037050058968841e-06, 'learning_rate/action_model': 8.03705005896884e-05, 'epoch': 0.61}) +[2026-06-23 18:54:48,517][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.17795784771442413, 'train/grad_norm_pre_clip': 53.39863967895508, 'timing/data': 0.0007833128329366446, 'timing/model': 1.6948765118140727, 'learning_rate/qwen_vl_interface': 8.034583395275392e-06, 'learning_rate/action_model': 8.034583395275392e-05, 'epoch': 0.61}) +[2026-06-23 18:54:51,853][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.13210277259349823, 'train/grad_norm_pre_clip': 13.011008262634277, 'timing/data': 0.0006116500589996576, 'timing/model': 1.689657328883186, 'learning_rate/qwen_vl_interface': 8.032115586944893e-06, 'learning_rate/action_model': 8.032115586944892e-05, 'epoch': 0.61}) +[2026-06-23 18:54:55,193][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.22874140739440918, 'train/grad_norm_pre_clip': 56.665992736816406, 'timing/data': 0.0006049280054867268, 'timing/model': 1.696245462168008, 'learning_rate/qwen_vl_interface': 8.029646634991762e-06, 'learning_rate/action_model': 8.029646634991762e-05, 'epoch': 0.61}) +[2026-06-23 18:54:58,533][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.14992640912532806, 'train/grad_norm_pre_clip': 51.87295150756836, 'timing/data': 0.0006601808127015829, 'timing/model': 1.6967088009696454, 'learning_rate/qwen_vl_interface': 8.0271765404309e-06, 'learning_rate/action_model': 8.0271765404309e-05, 'epoch': 0.61}) +[2026-06-23 18:55:01,872][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.132610023021698, 'train/grad_norm_pre_clip': 26.09199333190918, 'timing/data': 0.0006760410033166409, 'timing/model': 1.695413143839687, 'learning_rate/qwen_vl_interface': 8.024705304277664e-06, 'learning_rate/action_model': 8.024705304277664e-05, 'epoch': 0.61}) +[2026-06-23 18:55:05,211][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.19988121092319489, 'train/grad_norm_pre_clip': 39.947017669677734, 'timing/data': 0.0006772938650101423, 'timing/model': 1.697702526114881, 'learning_rate/qwen_vl_interface': 8.022232927547888e-06, 'learning_rate/action_model': 8.022232927547886e-05, 'epoch': 0.61}) +[2026-06-23 18:55:08,556][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.23267896473407745, 'train/grad_norm_pre_clip': 20.05622100830078, 'timing/data': 0.0005989158526062965, 'timing/model': 1.6998341518919915, 'learning_rate/qwen_vl_interface': 8.01975941125787e-06, 'learning_rate/action_model': 8.01975941125787e-05, 'epoch': 0.61}) +[2026-06-23 18:55:11,903][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.18392443656921387, 'train/grad_norm_pre_clip': 10.60641098022461, 'timing/data': 0.0006075119599699974, 'timing/model': 1.6989792129024863, 'learning_rate/qwen_vl_interface': 8.017284756424381e-06, 'learning_rate/action_model': 8.017284756424381e-05, 'epoch': 0.61}) +[2026-06-23 18:55:15,253][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.21471434831619263, 'train/grad_norm_pre_clip': 56.17201614379883, 'timing/data': 0.0006633470766246319, 'timing/model': 1.702539674937725, 'learning_rate/qwen_vl_interface': 8.01480896406466e-06, 'learning_rate/action_model': 8.01480896406466e-05, 'epoch': 0.61}) +[2026-06-23 18:55:18,606][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.33926987648010254, 'train/grad_norm_pre_clip': 96.52545928955078, 'timing/data': 0.0006619540508836508, 'timing/model': 1.7029008220415562, 'learning_rate/qwen_vl_interface': 8.01233203519641e-06, 'learning_rate/action_model': 8.012332035196409e-05, 'epoch': 0.61}) +[2026-06-23 18:55:21,956][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.17309816181659698, 'train/grad_norm_pre_clip': 61.165958404541016, 'timing/data': 0.0006645489484071732, 'timing/model': 1.6999753301497549, 'learning_rate/qwen_vl_interface': 8.009853970837801e-06, 'learning_rate/action_model': 8.009853970837801e-05, 'epoch': 0.61}) +[2026-06-23 18:55:25,319][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.2497897893190384, 'train/grad_norm_pre_clip': 113.3685531616211, 'timing/data': 0.0006542399059981108, 'timing/model': 1.7080271758604795, 'learning_rate/qwen_vl_interface': 8.007374772007474e-06, 'learning_rate/action_model': 8.007374772007474e-05, 'epoch': 0.61}) +[2026-06-23 18:55:28,661][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.18189555406570435, 'train/grad_norm_pre_clip': 28.338647842407227, 'timing/data': 0.0006103678606450558, 'timing/model': 1.6971887180116028, 'learning_rate/qwen_vl_interface': 8.004894439724533e-06, 'learning_rate/action_model': 8.004894439724533e-05, 'epoch': 0.61}) +[2026-06-23 18:55:32,018][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.11245270073413849, 'train/grad_norm_pre_clip': 23.99199867248535, 'timing/data': 0.0006668330170214176, 'timing/model': 1.7081276110839099, 'learning_rate/qwen_vl_interface': 8.002412975008552e-06, 'learning_rate/action_model': 8.002412975008552e-05, 'epoch': 0.61}) +[2026-06-23 18:55:35,456][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.24924498796463013, 'train/grad_norm_pre_clip': 77.75202178955078, 'timing/data': 0.0006678060162812471, 'timing/model': 1.7924304041080177, 'learning_rate/qwen_vl_interface': 7.999930378879562e-06, 'learning_rate/action_model': 7.999930378879561e-05, 'epoch': 0.61}) +[2026-06-23 18:55:38,808][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.17187264561653137, 'train/grad_norm_pre_clip': 61.13814163208008, 'timing/data': 0.0006572951097041368, 'timing/model': 1.7034045080654323, 'learning_rate/qwen_vl_interface': 7.997446652358067e-06, 'learning_rate/action_model': 7.997446652358068e-05, 'epoch': 0.61}) +[2026-06-23 18:55:42,154][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.12891152501106262, 'train/grad_norm_pre_clip': 39.59181594848633, 'timing/data': 0.0006089350208640099, 'timing/model': 1.6979992950800806, 'learning_rate/qwen_vl_interface': 7.994961796465035e-06, 'learning_rate/action_model': 7.994961796465035e-05, 'epoch': 0.61}) +[2026-06-23 18:55:45,502][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.3917788565158844, 'train/grad_norm_pre_clip': 33.843772888183594, 'timing/data': 0.0006163390353322029, 'timing/model': 1.7012073250953108, 'learning_rate/qwen_vl_interface': 7.992475812221894e-06, 'learning_rate/action_model': 7.992475812221894e-05, 'epoch': 0.62}) +[2026-06-23 18:55:48,846][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.0804278701543808, 'train/grad_norm_pre_clip': 19.831499099731445, 'timing/data': 0.0006095049902796745, 'timing/model': 1.6961577499751002, 'learning_rate/qwen_vl_interface': 7.989988700650538e-06, 'learning_rate/action_model': 7.989988700650538e-05, 'epoch': 0.62}) +[2026-06-23 18:55:52,196][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.2790831923484802, 'train/grad_norm_pre_clip': 110.43067932128906, 'timing/data': 0.0006144850049167871, 'timing/model': 1.7036218501161784, 'learning_rate/qwen_vl_interface': 7.987500462773326e-06, 'learning_rate/action_model': 7.987500462773326e-05, 'epoch': 0.62}) +[2026-06-23 18:55:55,551][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.1992599368095398, 'train/grad_norm_pre_clip': 96.61832427978516, 'timing/data': 0.0006130619440227747, 'timing/model': 1.6985256080515683, 'learning_rate/qwen_vl_interface': 7.985011099613078e-06, 'learning_rate/action_model': 7.985011099613077e-05, 'epoch': 0.62}) +[2026-06-23 18:55:58,900][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.2264857292175293, 'train/grad_norm_pre_clip': 108.05943298339844, 'timing/data': 0.0006064700428396463, 'timing/model': 1.7010510750114918, 'learning_rate/qwen_vl_interface': 7.98252061219308e-06, 'learning_rate/action_model': 7.982520612193078e-05, 'epoch': 0.62}) +[2026-06-23 18:56:02,244][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.3547167479991913, 'train/grad_norm_pre_clip': 163.03829956054688, 'timing/data': 0.000613924115896225, 'timing/model': 1.6975985730532557, 'learning_rate/qwen_vl_interface': 7.980029001537073e-06, 'learning_rate/action_model': 7.980029001537072e-05, 'epoch': 0.62}) +[2026-06-23 18:56:05,591][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.26366814970970154, 'train/grad_norm_pre_clip': 84.8575210571289, 'timing/data': 0.0006767320446670055, 'timing/model': 1.6988200671039522, 'learning_rate/qwen_vl_interface': 7.977536268669264e-06, 'learning_rate/action_model': 7.977536268669264e-05, 'epoch': 0.62}) +[2026-06-23 18:56:08,937][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.21402157843112946, 'train/grad_norm_pre_clip': 27.276472091674805, 'timing/data': 0.0006624949164688587, 'timing/model': 1.6984385380055755, 'learning_rate/qwen_vl_interface': 7.975042414614324e-06, 'learning_rate/action_model': 7.975042414614324e-05, 'epoch': 0.62}) +[2026-06-23 18:56:12,294][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.1608884185552597, 'train/grad_norm_pre_clip': 22.151901245117188, 'timing/data': 0.0007169281598180532, 'timing/model': 1.70605582697317, 'learning_rate/qwen_vl_interface': 7.972547440397382e-06, 'learning_rate/action_model': 7.972547440397383e-05, 'epoch': 0.62}) +[2026-06-23 18:56:15,642][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.18453240394592285, 'train/grad_norm_pre_clip': 75.66004180908203, 'timing/data': 0.000606701010838151, 'timing/model': 1.701139312935993, 'learning_rate/qwen_vl_interface': 7.970051347044026e-06, 'learning_rate/action_model': 7.970051347044025e-05, 'epoch': 0.62}) +[2026-06-23 18:56:18,988][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1600 +[2026-06-23 18:57:51,876][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1600 +[2026-06-23 18:57:51,877][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1600 +[2026-06-23 18:59:03,242][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1600 +[2026-06-23 18:59:03,243][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.14930234849452972, 'train/grad_norm_pre_clip': 50.61878204345703, 'eval/action_loss/seconds': 92.88669235818088, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.11389551311731339, 'eval/latency_2/loss': 0.11389551311731339, 'eval/flappy/loss': 0.11389551311731339, 'eval/flappy/latency_2/loss': 0.11389551311731339, 'eval/flap_precision': 0.8186440677966101, 'eval/flap_recall': 0.9777327935222672, 'eval/flap_f1': 0.8911439114391143, 'eval/flap_rate_model': 0.24583333333333332, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8911439114391142, 'eval/latency_2/flap_precision': 0.8186440677966101, 'eval/latency_2/flap_recall': 0.9777327935222672, 'eval/latency_2/flap_f1': 0.8911439114391143, 'eval/latency_2/flap_rate_model': 0.24583333333333332, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8911439114391142, 'eval/action_classification/data_wait_seconds': 0.9000844359397888, 'eval/action_classification/predict_seconds': 70.31707000732422, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006030239164829254, 'timing/model': 1.700174859026447, 'learning_rate/qwen_vl_interface': 7.967554135580307e-06, 'learning_rate/action_model': 7.967554135580307e-05, 'epoch': 0.62}) +[2026-06-23 18:59:06,570][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.0869709923863411, 'train/grad_norm_pre_clip': 28.6319637298584, 'timing/data': 0.0006096058059483767, 'timing/model': 1.6893922339659184, 'learning_rate/qwen_vl_interface': 7.965055807032734e-06, 'learning_rate/action_model': 7.965055807032735e-05, 'epoch': 0.62}) +[2026-06-23 18:59:09,892][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.1120578870177269, 'train/grad_norm_pre_clip': 37.32186508178711, 'timing/data': 0.0006052779499441385, 'timing/model': 1.6837371319998056, 'learning_rate/qwen_vl_interface': 7.962556362428275e-06, 'learning_rate/action_model': 7.962556362428274e-05, 'epoch': 0.62}) +[2026-06-23 18:59:13,223][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.17849420011043549, 'train/grad_norm_pre_clip': 29.295717239379883, 'timing/data': 0.0006113199051469564, 'timing/model': 1.6920754259917885, 'learning_rate/qwen_vl_interface': 7.960055802794358e-06, 'learning_rate/action_model': 7.960055802794357e-05, 'epoch': 0.62}) +[2026-06-23 18:59:16,552][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.20479464530944824, 'train/grad_norm_pre_clip': 92.33708190917969, 'timing/data': 0.0006134130526334047, 'timing/model': 1.6913414341397583, 'learning_rate/qwen_vl_interface': 7.957554129158866e-06, 'learning_rate/action_model': 7.957554129158866e-05, 'epoch': 0.62}) +[2026-06-23 18:59:19,883][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.2168237417936325, 'train/grad_norm_pre_clip': 83.20191955566406, 'timing/data': 0.0006268380675464869, 'timing/model': 1.692972029093653, 'learning_rate/qwen_vl_interface': 7.955051342550148e-06, 'learning_rate/action_model': 7.955051342550148e-05, 'epoch': 0.62}) +[2026-06-23 18:59:23,219][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.11105142533779144, 'train/grad_norm_pre_clip': 92.191162109375, 'timing/data': 0.0009618580807000399, 'timing/model': 1.69190408103168, 'learning_rate/qwen_vl_interface': 7.952547443997e-06, 'learning_rate/action_model': 7.952547443997e-05, 'epoch': 0.62}) +[2026-06-23 18:59:26,556][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.13679490983486176, 'train/grad_norm_pre_clip': 50.48795700073242, 'timing/data': 0.000628461129963398, 'timing/model': 1.6940084758680314, 'learning_rate/qwen_vl_interface': 7.950042434528681e-06, 'learning_rate/action_model': 7.950042434528682e-05, 'epoch': 0.62}) +[2026-06-23 18:59:29,886][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.2104121446609497, 'train/grad_norm_pre_clip': 61.48198699951172, 'timing/data': 0.0006086241919547319, 'timing/model': 1.6914587409701198, 'learning_rate/qwen_vl_interface': 7.94753631517491e-06, 'learning_rate/action_model': 7.947536315174909e-05, 'epoch': 0.62}) +[2026-06-23 18:59:33,223][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.14488355815410614, 'train/grad_norm_pre_clip': 26.102291107177734, 'timing/data': 0.0006115089636296034, 'timing/model': 1.6949372289236635, 'learning_rate/qwen_vl_interface': 7.945029086965854e-06, 'learning_rate/action_model': 7.945029086965853e-05, 'epoch': 0.62}) +[2026-06-23 18:59:36,649][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.19622819125652313, 'train/grad_norm_pre_clip': 33.9280891418457, 'timing/data': 0.0006654909811913967, 'timing/model': 1.6908918740227818, 'learning_rate/qwen_vl_interface': 7.942520750932138e-06, 'learning_rate/action_model': 7.942520750932138e-05, 'epoch': 0.62}) +[2026-06-23 18:59:39,986][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.18477113544940948, 'train/grad_norm_pre_clip': 39.64196014404297, 'timing/data': 0.000673685921356082, 'timing/model': 1.6947743808850646, 'learning_rate/qwen_vl_interface': 7.940011308104848e-06, 'learning_rate/action_model': 7.940011308104848e-05, 'epoch': 0.62}) +[2026-06-23 18:59:43,321][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.15723493695259094, 'train/grad_norm_pre_clip': 57.987274169921875, 'timing/data': 0.000685699051246047, 'timing/model': 1.6932039980310947, 'learning_rate/qwen_vl_interface': 7.93750075951552e-06, 'learning_rate/action_model': 7.93750075951552e-05, 'epoch': 0.62}) +[2026-06-23 18:59:46,657][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.20869919657707214, 'train/grad_norm_pre_clip': 26.525440216064453, 'timing/data': 0.0006679759826511145, 'timing/model': 1.6949065269436687, 'learning_rate/qwen_vl_interface': 7.934989106196145e-06, 'learning_rate/action_model': 7.934989106196146e-05, 'epoch': 0.62}) +[2026-06-23 18:59:49,994][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.21438437700271606, 'train/grad_norm_pre_clip': 25.712726593017578, 'timing/data': 0.0006609128322452307, 'timing/model': 1.6908511228393763, 'learning_rate/qwen_vl_interface': 7.93247634917917e-06, 'learning_rate/action_model': 7.93247634917917e-05, 'epoch': 0.62}) +[2026-06-23 18:59:53,330][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.11313404887914658, 'train/grad_norm_pre_clip': 63.73495864868164, 'timing/data': 0.0006555721629410982, 'timing/model': 1.6948723320383579, 'learning_rate/qwen_vl_interface': 7.929962489497494e-06, 'learning_rate/action_model': 7.929962489497493e-05, 'epoch': 0.63}) +[2026-06-23 18:59:56,673][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.1659175008535385, 'train/grad_norm_pre_clip': 61.49560546875, 'timing/data': 0.0006102868355810642, 'timing/model': 1.7001190141309053, 'learning_rate/qwen_vl_interface': 7.927447528184469e-06, 'learning_rate/action_model': 7.927447528184468e-05, 'epoch': 0.63}) +[2026-06-23 19:00:00,006][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.14901103079319, 'train/grad_norm_pre_clip': 48.67607879638672, 'timing/data': 0.0005984650924801826, 'timing/model': 1.693292910931632, 'learning_rate/qwen_vl_interface': 7.924931466273901e-06, 'learning_rate/action_model': 7.924931466273902e-05, 'epoch': 0.63}) +[2026-06-23 19:00:03,349][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.10778157413005829, 'train/grad_norm_pre_clip': 14.818041801452637, 'timing/data': 0.0006365568842738867, 'timing/model': 1.7000295240432024, 'learning_rate/qwen_vl_interface': 7.92241430480005e-06, 'learning_rate/action_model': 7.922414304800049e-05, 'epoch': 0.63}) +[2026-06-23 19:00:06,689][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.22259843349456787, 'train/grad_norm_pre_clip': 23.49890899658203, 'timing/data': 0.0005896890070289373, 'timing/model': 1.6958829669747502, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.63}) +[2026-06-23 19:00:10,026][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.17819826304912567, 'train/grad_norm_pre_clip': 67.18035125732422, 'timing/data': 0.0008111251518130302, 'timing/model': 1.6916164548601955, 'learning_rate/qwen_vl_interface': 7.91737668730178e-06, 'learning_rate/action_model': 7.91737668730178e-05, 'epoch': 0.63}) +[2026-06-23 19:00:13,362][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.24718958139419556, 'train/grad_norm_pre_clip': 50.71849060058594, 'timing/data': 0.0006900569424033165, 'timing/model': 1.6972283250652254, 'learning_rate/qwen_vl_interface': 7.91485623334814e-06, 'learning_rate/action_model': 7.91485623334814e-05, 'epoch': 0.63}) +[2026-06-23 19:00:16,704][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.14299435913562775, 'train/grad_norm_pre_clip': 57.11122512817383, 'timing/data': 0.0006123608909547329, 'timing/model': 1.697983152931556, 'learning_rate/qwen_vl_interface': 7.91233468397276e-06, 'learning_rate/action_model': 7.912334683972761e-05, 'epoch': 0.63}) +[2026-06-23 19:00:20,047][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.25954943895339966, 'train/grad_norm_pre_clip': 56.121219635009766, 'timing/data': 0.0006034949328750372, 'timing/model': 1.6952248059678823, 'learning_rate/qwen_vl_interface': 7.90981204021216e-06, 'learning_rate/action_model': 7.90981204021216e-05, 'epoch': 0.63}) +[2026-06-23 19:00:23,388][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.19157381355762482, 'train/grad_norm_pre_clip': 18.681724548339844, 'timing/data': 0.0006343219429254532, 'timing/model': 1.6960940801072866, 'learning_rate/qwen_vl_interface': 7.907288303103298e-06, 'learning_rate/action_model': 7.907288303103298e-05, 'epoch': 0.63}) +[2026-06-23 19:00:27,012][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.15660862624645233, 'train/grad_norm_pre_clip': 56.40798568725586, 'timing/data': 0.0006008101627230644, 'timing/model': 1.6994212928693742, 'learning_rate/qwen_vl_interface': 7.90476347368359e-06, 'learning_rate/action_model': 7.90476347368359e-05, 'epoch': 0.63}) +[2026-06-23 19:00:30,354][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.1854366958141327, 'train/grad_norm_pre_clip': 86.65925598144531, 'timing/data': 0.0005861418321728706, 'timing/model': 1.6948361431714147, 'learning_rate/qwen_vl_interface': 7.902237552990897e-06, 'learning_rate/action_model': 7.902237552990897e-05, 'epoch': 0.63}) +[2026-06-23 19:00:33,694][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.1593717634677887, 'train/grad_norm_pre_clip': 42.63607406616211, 'timing/data': 0.0006024329923093319, 'timing/model': 1.6960776560008526, 'learning_rate/qwen_vl_interface': 7.89971054206353e-06, 'learning_rate/action_model': 7.89971054206353e-05, 'epoch': 0.63}) +[2026-06-23 19:00:37,038][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.1418255865573883, 'train/grad_norm_pre_clip': 41.82349395751953, 'timing/data': 0.0006149560213088989, 'timing/model': 1.6978942600544542, 'learning_rate/qwen_vl_interface': 7.897182441940247e-06, 'learning_rate/action_model': 7.897182441940247e-05, 'epoch': 0.63}) +[2026-06-23 19:00:40,376][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.10466226935386658, 'train/grad_norm_pre_clip': 29.006263732910156, 'timing/data': 0.0006195651367306709, 'timing/model': 1.6925548119470477, 'learning_rate/qwen_vl_interface': 7.894653253660256e-06, 'learning_rate/action_model': 7.894653253660256e-05, 'epoch': 0.63}) +[2026-06-23 19:00:43,714][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.057098258286714554, 'train/grad_norm_pre_clip': 23.94674301147461, 'timing/data': 0.0006008390337228775, 'timing/model': 1.6956331480760127, 'learning_rate/qwen_vl_interface': 7.89212297826321e-06, 'learning_rate/action_model': 7.892122978263209e-05, 'epoch': 0.63}) +[2026-06-23 19:00:47,053][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.16798904538154602, 'train/grad_norm_pre_clip': 29.31500244140625, 'timing/data': 0.000706828897818923, 'timing/model': 1.6945742459502071, 'learning_rate/qwen_vl_interface': 7.889591616789206e-06, 'learning_rate/action_model': 7.889591616789206e-05, 'epoch': 0.63}) +[2026-06-23 19:00:50,390][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.11003152281045914, 'train/grad_norm_pre_clip': 10.991294860839844, 'timing/data': 0.0006174498703330755, 'timing/model': 1.691885479958728, 'learning_rate/qwen_vl_interface': 7.887059170278799e-06, 'learning_rate/action_model': 7.887059170278797e-05, 'epoch': 0.63}) +[2026-06-23 19:00:53,732][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.2441807985305786, 'train/grad_norm_pre_clip': 15.508068084716797, 'timing/data': 0.0006243239622563124, 'timing/model': 1.6959384689107537, 'learning_rate/qwen_vl_interface': 7.884525639772974e-06, 'learning_rate/action_model': 7.884525639772974e-05, 'epoch': 0.63}) +[2026-06-23 19:00:57,080][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.15647611021995544, 'train/grad_norm_pre_clip': 20.99489402770996, 'timing/data': 0.0006068411748856306, 'timing/model': 1.7001994280144572, 'learning_rate/qwen_vl_interface': 7.881991026313176e-06, 'learning_rate/action_model': 7.881991026313175e-05, 'epoch': 0.63}) +[2026-06-23 19:01:00,425][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.17342476546764374, 'train/grad_norm_pre_clip': 17.372886657714844, 'timing/data': 0.0005880361422896385, 'timing/model': 1.6975764359813184, 'learning_rate/qwen_vl_interface': 7.879455330941283e-06, 'learning_rate/action_model': 7.879455330941282e-05, 'epoch': 0.63}) +[2026-06-23 19:01:03,773][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.20647242665290833, 'train/grad_norm_pre_clip': 15.713815689086914, 'timing/data': 0.0006002588197588921, 'timing/model': 1.699376589851454, 'learning_rate/qwen_vl_interface': 7.876918554699627e-06, 'learning_rate/action_model': 7.876918554699627e-05, 'epoch': 0.63}) +[2026-06-23 19:01:07,112][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.1482173502445221, 'train/grad_norm_pre_clip': 21.265483856201172, 'timing/data': 0.0006072318647056818, 'timing/model': 1.6995964220259339, 'learning_rate/qwen_vl_interface': 7.874380698630982e-06, 'learning_rate/action_model': 7.874380698630982e-05, 'epoch': 0.63}) +[2026-06-23 19:01:10,462][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.2211369276046753, 'train/grad_norm_pre_clip': 75.1578140258789, 'timing/data': 0.00061611901037395, 'timing/model': 1.702060162089765, 'learning_rate/qwen_vl_interface': 7.871841763778562e-06, 'learning_rate/action_model': 7.871841763778563e-05, 'epoch': 0.63}) +[2026-06-23 19:01:13,901][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.14185118675231934, 'train/grad_norm_pre_clip': 82.90792846679688, 'timing/data': 0.000609495909884572, 'timing/model': 1.699911420000717, 'learning_rate/qwen_vl_interface': 7.86930175118603e-06, 'learning_rate/action_model': 7.86930175118603e-05, 'epoch': 0.63}) +[2026-06-23 19:01:17,242][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.2883296608924866, 'train/grad_norm_pre_clip': 22.472986221313477, 'timing/data': 0.0006028229836374521, 'timing/model': 1.6955955501180142, 'learning_rate/qwen_vl_interface': 7.866760661897487e-06, 'learning_rate/action_model': 7.866760661897487e-05, 'epoch': 0.63}) +[2026-06-23 19:01:20,582][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.17418862879276276, 'train/grad_norm_pre_clip': 43.62870788574219, 'timing/data': 0.0005987950135022402, 'timing/model': 1.6967049669474363, 'learning_rate/qwen_vl_interface': 7.86421849695748e-06, 'learning_rate/action_model': 7.864218496957481e-05, 'epoch': 0.64}) +[2026-06-23 19:01:23,922][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.09564193338155746, 'train/grad_norm_pre_clip': 58.7794075012207, 'timing/data': 0.0005946678575128317, 'timing/model': 1.6991002617869526, 'learning_rate/qwen_vl_interface': 7.861675257410998e-06, 'learning_rate/action_model': 7.861675257410997e-05, 'epoch': 0.64}) +[2026-06-23 19:01:27,289][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.2029738575220108, 'train/grad_norm_pre_clip': 55.95201110839844, 'timing/data': 0.0006066500209271908, 'timing/model': 1.6982831181958318, 'learning_rate/qwen_vl_interface': 7.859130944303468e-06, 'learning_rate/action_model': 7.859130944303467e-05, 'epoch': 0.64}) +[2026-06-23 19:01:30,644][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.07375762611627579, 'train/grad_norm_pre_clip': 15.430256843566895, 'timing/data': 0.0006080628372728825, 'timing/model': 1.6965615239460021, 'learning_rate/qwen_vl_interface': 7.856585558680761e-06, 'learning_rate/action_model': 7.856585558680761e-05, 'epoch': 0.64}) +[2026-06-23 19:01:33,983][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.12710733711719513, 'train/grad_norm_pre_clip': 27.54227638244629, 'timing/data': 0.0006035640835762024, 'timing/model': 1.6941958791576326, 'learning_rate/qwen_vl_interface': 7.854039101589191e-06, 'learning_rate/action_model': 7.854039101589191e-05, 'epoch': 0.64}) +[2026-06-23 19:01:37,332][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.15043456852436066, 'train/grad_norm_pre_clip': 55.331172943115234, 'timing/data': 0.0006006790790706873, 'timing/model': 1.699978570919484, 'learning_rate/qwen_vl_interface': 7.851491574075509e-06, 'learning_rate/action_model': 7.851491574075509e-05, 'epoch': 0.64}) +[2026-06-23 19:01:40,687][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.11447854340076447, 'train/grad_norm_pre_clip': 18.764530181884766, 'timing/data': 0.0006051880773156881, 'timing/model': 1.7067441300023347, 'learning_rate/qwen_vl_interface': 7.848942977186909e-06, 'learning_rate/action_model': 7.848942977186908e-05, 'epoch': 0.64}) +[2026-06-23 19:01:44,036][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.07278240472078323, 'train/grad_norm_pre_clip': 20.450342178344727, 'timing/data': 0.0006152060814201832, 'timing/model': 1.7016689928714186, 'learning_rate/qwen_vl_interface': 7.84639331197102e-06, 'learning_rate/action_model': 7.84639331197102e-05, 'epoch': 0.64}) +[2026-06-23 19:01:47,382][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.2059909701347351, 'train/grad_norm_pre_clip': 14.000960350036621, 'timing/data': 0.0006113590206950903, 'timing/model': 1.6978850120212883, 'learning_rate/qwen_vl_interface': 7.843842579475918e-06, 'learning_rate/action_model': 7.843842579475918e-05, 'epoch': 0.64}) +[2026-06-23 19:01:50,724][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.16038218140602112, 'train/grad_norm_pre_clip': 23.02151107788086, 'timing/data': 0.0006102570332586765, 'timing/model': 1.6974736009724438, 'learning_rate/qwen_vl_interface': 7.841290780750107e-06, 'learning_rate/action_model': 7.841290780750107e-05, 'epoch': 0.64}) +[2026-06-23 19:01:54,061][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.1074000746011734, 'train/grad_norm_pre_clip': 17.225854873657227, 'timing/data': 0.0005940061528235674, 'timing/model': 1.6962181329727173, 'learning_rate/qwen_vl_interface': 7.838737916842538e-06, 'learning_rate/action_model': 7.838737916842538e-05, 'epoch': 0.64}) +[2026-06-23 19:01:57,408][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.1544564664363861, 'train/grad_norm_pre_clip': 17.4289493560791, 'timing/data': 0.0005935761146247387, 'timing/model': 1.700799976941198, 'learning_rate/qwen_vl_interface': 7.836183988802596e-06, 'learning_rate/action_model': 7.836183988802595e-05, 'epoch': 0.64}) +[2026-06-23 19:02:00,746][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.19470012187957764, 'train/grad_norm_pre_clip': 19.617765426635742, 'timing/data': 0.0005913819186389446, 'timing/model': 1.6919213410001248, 'learning_rate/qwen_vl_interface': 7.833628997680107e-06, 'learning_rate/action_model': 7.833628997680107e-05, 'epoch': 0.64}) +[2026-06-23 19:02:04,385][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.2760893702507019, 'train/grad_norm_pre_clip': 24.336801528930664, 'timing/data': 0.0006265880074352026, 'timing/model': 1.6955073329154402, 'learning_rate/qwen_vl_interface': 7.831072944525329e-06, 'learning_rate/action_model': 7.831072944525328e-05, 'epoch': 0.64}) +[2026-06-23 19:02:07,725][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.11441261321306229, 'train/grad_norm_pre_clip': 23.7036190032959, 'timing/data': 0.0006185630336403847, 'timing/model': 1.6966968609485775, 'learning_rate/qwen_vl_interface': 7.828515830388956e-06, 'learning_rate/action_model': 7.828515830388956e-05, 'epoch': 0.64}) +[2026-06-23 19:02:11,074][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.11943940073251724, 'train/grad_norm_pre_clip': 32.22972869873047, 'timing/data': 0.0007150638848543167, 'timing/model': 1.7015786380507052, 'learning_rate/qwen_vl_interface': 7.825957656322129e-06, 'learning_rate/action_model': 7.825957656322128e-05, 'epoch': 0.64}) +[2026-06-23 19:02:14,420][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.1814742088317871, 'train/grad_norm_pre_clip': 42.2054443359375, 'timing/data': 0.0006089149974286556, 'timing/model': 1.698773848824203, 'learning_rate/qwen_vl_interface': 7.823398423376406e-06, 'learning_rate/action_model': 7.823398423376406e-05, 'epoch': 0.64}) +[2026-06-23 19:02:17,760][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.15815423429012299, 'train/grad_norm_pre_clip': 31.38895606994629, 'timing/data': 0.0006071710959076881, 'timing/model': 1.6982650351710618, 'learning_rate/qwen_vl_interface': 7.820838132603801e-06, 'learning_rate/action_model': 7.8208381326038e-05, 'epoch': 0.64}) +[2026-06-23 19:02:21,103][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.12794966995716095, 'train/grad_norm_pre_clip': 38.34281921386719, 'timing/data': 0.0007420140318572521, 'timing/model': 1.697066503111273, 'learning_rate/qwen_vl_interface': 7.818276785056748e-06, 'learning_rate/action_model': 7.818276785056749e-05, 'epoch': 0.64}) +[2026-06-23 19:02:24,451][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.14500510692596436, 'train/grad_norm_pre_clip': 61.794795989990234, 'timing/data': 0.0006052269600331783, 'timing/model': 1.7007242911495268, 'learning_rate/qwen_vl_interface': 7.815714381788123e-06, 'learning_rate/action_model': 7.815714381788122e-05, 'epoch': 0.64}) +[2026-06-23 19:02:27,785][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.17071093618869781, 'train/grad_norm_pre_clip': 39.951881408691406, 'timing/data': 0.000597432954236865, 'timing/model': 1.691882352810353, 'learning_rate/qwen_vl_interface': 7.813150923851227e-06, 'learning_rate/action_model': 7.813150923851227e-05, 'epoch': 0.64}) +[2026-06-23 19:02:31,128][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.17131297290325165, 'train/grad_norm_pre_clip': 37.64748001098633, 'timing/data': 0.0006188640836626291, 'timing/model': 1.6978878830559552, 'learning_rate/qwen_vl_interface': 7.810586412299807e-06, 'learning_rate/action_model': 7.810586412299806e-05, 'epoch': 0.64}) +[2026-06-23 19:02:34,464][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.09371764212846756, 'train/grad_norm_pre_clip': 29.293304443359375, 'timing/data': 0.0006030930671840906, 'timing/model': 1.694047847064212, 'learning_rate/qwen_vl_interface': 7.808020848188034e-06, 'learning_rate/action_model': 7.808020848188034e-05, 'epoch': 0.64}) +[2026-06-23 19:02:37,799][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.24714362621307373, 'train/grad_norm_pre_clip': 14.984249114990234, 'timing/data': 0.0006025631446391344, 'timing/model': 1.6947193308733404, 'learning_rate/qwen_vl_interface': 7.805454232570517e-06, 'learning_rate/action_model': 7.805454232570517e-05, 'epoch': 0.64}) +[2026-06-23 19:02:41,134][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.175787091255188, 'train/grad_norm_pre_clip': 24.790719985961914, 'timing/data': 0.0006200261414051056, 'timing/model': 1.6930624479427934, 'learning_rate/qwen_vl_interface': 7.80288656650229e-06, 'learning_rate/action_model': 7.80288656650229e-05, 'epoch': 0.64}) +[2026-06-23 19:02:44,483][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.1958080679178238, 'train/grad_norm_pre_clip': 56.15794372558594, 'timing/data': 0.0006183919031172991, 'timing/model': 1.7009728390257806, 'learning_rate/qwen_vl_interface': 7.800317851038829e-06, 'learning_rate/action_model': 7.800317851038828e-05, 'epoch': 0.64}) +[2026-06-23 19:02:47,823][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.1806706339120865, 'train/grad_norm_pre_clip': 23.05672264099121, 'timing/data': 0.0005970820784568787, 'timing/model': 1.6972279909532517, 'learning_rate/qwen_vl_interface': 7.797748087236033e-06, 'learning_rate/action_model': 7.797748087236033e-05, 'epoch': 0.65}) +[2026-06-23 19:02:51,169][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.2594491243362427, 'train/grad_norm_pre_clip': 22.828887939453125, 'timing/data': 0.0006411848589777946, 'timing/model': 1.7020945409312844, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.65}) +[2026-06-23 19:02:54,804][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.09830270707607269, 'train/grad_norm_pre_clip': 16.123950958251953, 'timing/data': 0.000605317996814847, 'timing/model': 1.6944034069310874, 'learning_rate/qwen_vl_interface': 7.792605418838196e-06, 'learning_rate/action_model': 7.792605418838197e-05, 'epoch': 0.65}) +[2026-06-23 19:02:58,147][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.17394012212753296, 'train/grad_norm_pre_clip': 27.19209861755371, 'timing/data': 0.0006186829414218664, 'timing/model': 1.6959799812175333, 'learning_rate/qwen_vl_interface': 7.790032516357116e-06, 'learning_rate/action_model': 7.790032516357117e-05, 'epoch': 0.65}) +[2026-06-23 19:03:01,489][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.24501685798168182, 'train/grad_norm_pre_clip': 49.434295654296875, 'timing/data': 0.0005958399269729853, 'timing/model': 1.6955301680136472, 'learning_rate/qwen_vl_interface': 7.787458569764613e-06, 'learning_rate/action_model': 7.787458569764612e-05, 'epoch': 0.65}) +[2026-06-23 19:03:04,831][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.10129072517156601, 'train/grad_norm_pre_clip': 36.063926696777344, 'timing/data': 0.0007489481940865517, 'timing/model': 1.6962712821550667, 'learning_rate/qwen_vl_interface': 7.784883580118738e-06, 'learning_rate/action_model': 7.784883580118738e-05, 'epoch': 0.65}) +[2026-06-23 19:03:08,172][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.052752863615751266, 'train/grad_norm_pre_clip': 27.119659423828125, 'timing/data': 0.0006035140249878168, 'timing/model': 1.6974801470059901, 'learning_rate/qwen_vl_interface': 7.782307548477975e-06, 'learning_rate/action_model': 7.782307548477975e-05, 'epoch': 0.65}) +[2026-06-23 19:03:11,513][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.14457230269908905, 'train/grad_norm_pre_clip': 45.14684295654297, 'timing/data': 0.0006007598713040352, 'timing/model': 1.6978815321344882, 'learning_rate/qwen_vl_interface': 7.779730475901233e-06, 'learning_rate/action_model': 7.779730475901232e-05, 'epoch': 0.65}) +[2026-06-23 19:03:14,847][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.21543803811073303, 'train/grad_norm_pre_clip': 55.6038818359375, 'timing/data': 0.0006041761953383684, 'timing/model': 1.692543189972639, 'learning_rate/qwen_vl_interface': 7.777152363447849e-06, 'learning_rate/action_model': 7.77715236344785e-05, 'epoch': 0.65}) +[2026-06-23 19:03:18,188][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.17959751188755035, 'train/grad_norm_pre_clip': 30.757108688354492, 'timing/data': 0.0006007589399814606, 'timing/model': 1.693700728006661, 'learning_rate/qwen_vl_interface': 7.774573212177587e-06, 'learning_rate/action_model': 7.774573212177588e-05, 'epoch': 0.65}) +[2026-06-23 19:03:21,523][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.12492003291845322, 'train/grad_norm_pre_clip': 48.120277404785156, 'timing/data': 0.0006037838757038116, 'timing/model': 1.690981418127194, 'learning_rate/qwen_vl_interface': 7.77199302315064e-06, 'learning_rate/action_model': 7.77199302315064e-05, 'epoch': 0.65}) +[2026-06-23 19:03:24,859][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.1313711702823639, 'train/grad_norm_pre_clip': 53.01197052001953, 'timing/data': 0.0006110889371484518, 'timing/model': 1.6921788901090622, 'learning_rate/qwen_vl_interface': 7.769411797427622e-06, 'learning_rate/action_model': 7.769411797427623e-05, 'epoch': 0.65}) +[2026-06-23 19:03:28,194][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.17346173524856567, 'train/grad_norm_pre_clip': 50.10542297363281, 'timing/data': 0.0006166689563542604, 'timing/model': 1.6942462981678545, 'learning_rate/qwen_vl_interface': 7.766829536069584e-06, 'learning_rate/action_model': 7.766829536069584e-05, 'epoch': 0.65}) +[2026-06-23 19:03:31,540][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.0895886942744255, 'train/grad_norm_pre_clip': 38.2577018737793, 'timing/data': 0.0006665131077170372, 'timing/model': 1.699472222942859, 'learning_rate/qwen_vl_interface': 7.764246240137991e-06, 'learning_rate/action_model': 7.764246240137991e-05, 'epoch': 0.65}) +[2026-06-23 19:03:34,877][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.18056152760982513, 'train/grad_norm_pre_clip': 16.81877326965332, 'timing/data': 0.0005952089559286833, 'timing/model': 1.6952658479567617, 'learning_rate/qwen_vl_interface': 7.761661910694741e-06, 'learning_rate/action_model': 7.761661910694741e-05, 'epoch': 0.65}) +[2026-06-23 19:03:38,218][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.09133671224117279, 'train/grad_norm_pre_clip': 28.239805221557617, 'timing/data': 0.0006117799784988165, 'timing/model': 1.6960366680286825, 'learning_rate/qwen_vl_interface': 7.759076548802153e-06, 'learning_rate/action_model': 7.759076548802152e-05, 'epoch': 0.65}) +[2026-06-23 19:03:41,650][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.2147127389907837, 'train/grad_norm_pre_clip': 61.58323669433594, 'timing/data': 0.0006077820435166359, 'timing/model': 1.7910980300512165, 'learning_rate/qwen_vl_interface': 7.75649015552297e-06, 'learning_rate/action_model': 7.75649015552297e-05, 'epoch': 0.65}) +[2026-06-23 19:03:44,992][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.1418413370847702, 'train/grad_norm_pre_clip': 45.696327209472656, 'timing/data': 0.0006160379853099585, 'timing/model': 1.697631207993254, 'learning_rate/qwen_vl_interface': 7.753902731920364e-06, 'learning_rate/action_model': 7.753902731920364e-05, 'epoch': 0.65}) +[2026-06-23 19:03:48,328][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.08467243611812592, 'train/grad_norm_pre_clip': 12.160359382629395, 'timing/data': 0.0008464308921247721, 'timing/model': 1.6929903470445424, 'learning_rate/qwen_vl_interface': 7.751314279057925e-06, 'learning_rate/action_model': 7.751314279057923e-05, 'epoch': 0.65}) +[2026-06-23 19:03:51,669][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.2759178876876831, 'train/grad_norm_pre_clip': 30.321346282958984, 'timing/data': 0.0008362021762877703, 'timing/model': 1.7000571619719267, 'learning_rate/qwen_vl_interface': 7.74872479799967e-06, 'learning_rate/action_model': 7.74872479799967e-05, 'epoch': 0.65}) +[2026-06-23 19:03:55,009][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.17741459608078003, 'train/grad_norm_pre_clip': 34.784889221191406, 'timing/data': 0.0007509118877351284, 'timing/model': 1.6992041000630707, 'learning_rate/qwen_vl_interface': 7.746134289810034e-06, 'learning_rate/action_model': 7.746134289810033e-05, 'epoch': 0.65}) +[2026-06-23 19:03:58,354][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.25606781244277954, 'train/grad_norm_pre_clip': 39.773136138916016, 'timing/data': 0.0006044560577720404, 'timing/model': 1.698750559007749, 'learning_rate/qwen_vl_interface': 7.74354275555388e-06, 'learning_rate/action_model': 7.74354275555388e-05, 'epoch': 0.65}) +[2026-06-23 19:04:01,689][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.13091573119163513, 'train/grad_norm_pre_clip': 26.599882125854492, 'timing/data': 0.0005909011233597994, 'timing/model': 1.6913566829171032, 'learning_rate/qwen_vl_interface': 7.740950196296488e-06, 'learning_rate/action_model': 7.740950196296488e-05, 'epoch': 0.65}) +[2026-06-23 19:04:05,028][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.09104331582784653, 'train/grad_norm_pre_clip': 17.00574493408203, 'timing/data': 0.0006263270042836666, 'timing/model': 1.6979255320038646, 'learning_rate/qwen_vl_interface': 7.738356613103562e-06, 'learning_rate/action_model': 7.73835661310356e-05, 'epoch': 0.65}) +[2026-06-23 19:04:08,368][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.19966691732406616, 'train/grad_norm_pre_clip': 15.723430633544922, 'timing/data': 0.000609945971518755, 'timing/model': 1.6971697029657662, 'learning_rate/qwen_vl_interface': 7.735762007041226e-06, 'learning_rate/action_model': 7.735762007041226e-05, 'epoch': 0.65}) +[2026-06-23 19:04:11,720][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.2780281901359558, 'train/grad_norm_pre_clip': 56.442718505859375, 'timing/data': 0.0006130728870630264, 'timing/model': 1.697850292781368, 'learning_rate/qwen_vl_interface': 7.733166379176023e-06, 'learning_rate/action_model': 7.733166379176023e-05, 'epoch': 0.65}) +[2026-06-23 19:04:15,064][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.24369463324546814, 'train/grad_norm_pre_clip': 61.085716247558594, 'timing/data': 0.0006700300145894289, 'timing/model': 1.6989715620875359, 'learning_rate/qwen_vl_interface': 7.730569730574921e-06, 'learning_rate/action_model': 7.73056973057492e-05, 'epoch': 0.66}) +[2026-06-23 19:04:18,420][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.17001062631607056, 'train/grad_norm_pre_clip': 24.802034378051758, 'timing/data': 0.0006123511120676994, 'timing/model': 1.7019277000799775, 'learning_rate/qwen_vl_interface': 7.727972062305299e-06, 'learning_rate/action_model': 7.727972062305299e-05, 'epoch': 0.66}) +[2026-06-23 19:04:21,761][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.18618881702423096, 'train/grad_norm_pre_clip': 29.668575286865234, 'timing/data': 0.0007585349958389997, 'timing/model': 1.696939375018701, 'learning_rate/qwen_vl_interface': 7.725373375434965e-06, 'learning_rate/action_model': 7.725373375434965e-05, 'epoch': 0.66}) +[2026-06-23 19:04:25,104][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.09953781962394714, 'train/grad_norm_pre_clip': 73.06523895263672, 'timing/data': 0.0006217989139258862, 'timing/model': 1.6961963400244713, 'learning_rate/qwen_vl_interface': 7.722773671032141e-06, 'learning_rate/action_model': 7.722773671032142e-05, 'epoch': 0.66}) +[2026-06-23 19:04:28,445][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.20843423902988434, 'train/grad_norm_pre_clip': 68.85201263427734, 'timing/data': 0.0006061699241399765, 'timing/model': 1.695723843993619, 'learning_rate/qwen_vl_interface': 7.720172950165466e-06, 'learning_rate/action_model': 7.720172950165465e-05, 'epoch': 0.66}) +[2026-06-23 19:04:31,877][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.10111084580421448, 'train/grad_norm_pre_clip': 46.42878723144531, 'timing/data': 0.000601370120421052, 'timing/model': 1.6986988671123981, 'learning_rate/qwen_vl_interface': 7.717571213903996e-06, 'learning_rate/action_model': 7.717571213903996e-05, 'epoch': 0.66}) +[2026-06-23 19:04:35,214][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.14617355167865753, 'train/grad_norm_pre_clip': 23.643020629882812, 'timing/data': 0.0006025130860507488, 'timing/model': 1.6946806760970503, 'learning_rate/qwen_vl_interface': 7.714968463317207e-06, 'learning_rate/action_model': 7.714968463317207e-05, 'epoch': 0.66}) +[2026-06-23 19:04:38,553][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1700 +[2026-06-23 19:06:12,170][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1700 +[2026-06-23 19:06:12,171][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1700 +[2026-06-23 19:07:23,297][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1700 +[2026-06-23 19:07:23,298][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.24157680571079254, 'train/grad_norm_pre_clip': 15.57193660736084, 'eval/action_loss/seconds': 93.61471540015191, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.08418277651071548, 'eval/latency_2/loss': 0.08418277651071548, 'eval/flappy/loss': 0.08418277651071548, 'eval/flappy/latency_2/loss': 0.08418277651071548, 'eval/flap_precision': 0.9328719723183391, 'eval/flap_recall': 0.9095816464237517, 'eval/flap_f1': 0.9210796036897848, 'eval/flap_rate_model': 0.20069444444444445, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9210796036897849, 'eval/latency_2/flap_precision': 0.9328719723183391, 'eval/latency_2/flap_recall': 0.9095816464237517, 'eval/latency_2/flap_f1': 0.9210796036897848, 'eval/latency_2/flap_rate_model': 0.20069444444444445, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9210796036897849, 'eval/action_classification/data_wait_seconds': 0.8386251926422119, 'eval/action_classification/predict_seconds': 70.14799499511719, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006519749294966459, 'timing/model': 1.6966118379496038, 'learning_rate/qwen_vl_interface': 7.712364699474996e-06, 'learning_rate/action_model': 7.712364699474996e-05, 'epoch': 0.66}) +[2026-06-23 19:07:26,610][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.12193664908409119, 'train/grad_norm_pre_clip': 26.503551483154297, 'timing/data': 0.0006075319834053516, 'timing/model': 1.6781030090060085, 'learning_rate/qwen_vl_interface': 7.709759923447667e-06, 'learning_rate/action_model': 7.709759923447665e-05, 'epoch': 0.66}) +[2026-06-23 19:07:29,930][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.15605847537517548, 'train/grad_norm_pre_clip': 17.08344078063965, 'timing/data': 0.0006218289490789175, 'timing/model': 1.6872839271090925, 'learning_rate/qwen_vl_interface': 7.707154136305945e-06, 'learning_rate/action_model': 7.707154136305945e-05, 'epoch': 0.66}) +[2026-06-23 19:07:33,256][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.10455239564180374, 'train/grad_norm_pre_clip': 45.98463439941406, 'timing/data': 0.0006620739586651325, 'timing/model': 1.6900705960579216, 'learning_rate/qwen_vl_interface': 7.704547339120974e-06, 'learning_rate/action_model': 7.704547339120974e-05, 'epoch': 0.66}) +[2026-06-23 19:07:36,576][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.14682744443416595, 'train/grad_norm_pre_clip': 33.25550842285156, 'timing/data': 0.0006104782223701477, 'timing/model': 1.6874707699753344, 'learning_rate/qwen_vl_interface': 7.70193953296431e-06, 'learning_rate/action_model': 7.70193953296431e-05, 'epoch': 0.66}) +[2026-06-23 19:07:39,889][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.16056758165359497, 'train/grad_norm_pre_clip': 13.702906608581543, 'timing/data': 0.0006028229836374521, 'timing/model': 1.6857906428631395, 'learning_rate/qwen_vl_interface': 7.69933071890792e-06, 'learning_rate/action_model': 7.699330718907919e-05, 'epoch': 0.66}) +[2026-06-23 19:07:43,314][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.1716560423374176, 'train/grad_norm_pre_clip': 41.4652214050293, 'timing/data': 0.0006074421107769012, 'timing/model': 1.686713658971712, 'learning_rate/qwen_vl_interface': 7.696720898024188e-06, 'learning_rate/action_model': 7.69672089802419e-05, 'epoch': 0.66}) +[2026-06-23 19:07:46,644][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.10849275439977646, 'train/grad_norm_pre_clip': 13.405014991760254, 'timing/data': 0.0006891048979014158, 'timing/model': 1.6929059498943388, 'learning_rate/qwen_vl_interface': 7.69411007138592e-06, 'learning_rate/action_model': 7.694110071385919e-05, 'epoch': 0.66}) +[2026-06-23 19:07:49,970][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.07359566539525986, 'train/grad_norm_pre_clip': 11.877581596374512, 'timing/data': 0.0006239430513232946, 'timing/model': 1.6882351208478212, 'learning_rate/qwen_vl_interface': 7.691498240066322e-06, 'learning_rate/action_model': 7.691498240066322e-05, 'epoch': 0.66}) +[2026-06-23 19:07:53,289][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.1373017281293869, 'train/grad_norm_pre_clip': 21.153711318969727, 'timing/data': 0.0006185730453580618, 'timing/model': 1.6888463711366057, 'learning_rate/qwen_vl_interface': 7.688885405139022e-06, 'learning_rate/action_model': 7.688885405139021e-05, 'epoch': 0.66}) +[2026-06-23 19:07:56,619][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.12373998761177063, 'train/grad_norm_pre_clip': 15.171225547790527, 'timing/data': 0.0005976930260658264, 'timing/model': 1.6892016199417412, 'learning_rate/qwen_vl_interface': 7.686271567678056e-06, 'learning_rate/action_model': 7.686271567678054e-05, 'epoch': 0.66}) +[2026-06-23 19:07:59,945][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.0796099379658699, 'train/grad_norm_pre_clip': 42.65359878540039, 'timing/data': 0.0006080931052565575, 'timing/model': 1.6916525780688971, 'learning_rate/qwen_vl_interface': 7.683656728757873e-06, 'learning_rate/action_model': 7.683656728757872e-05, 'epoch': 0.66}) +[2026-06-23 19:08:03,301][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.14295274019241333, 'train/grad_norm_pre_clip': 23.392303466796875, 'timing/data': 0.0006176910828799009, 'timing/model': 1.7140955340582877, 'learning_rate/qwen_vl_interface': 7.681040889453337e-06, 'learning_rate/action_model': 7.681040889453336e-05, 'epoch': 0.66}) +[2026-06-23 19:08:06,629][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.15545235574245453, 'train/grad_norm_pre_clip': 15.536858558654785, 'timing/data': 0.0006143350619822741, 'timing/model': 1.6911832080222666, 'learning_rate/qwen_vl_interface': 7.678424050839722e-06, 'learning_rate/action_model': 7.67842405083972e-05, 'epoch': 0.66}) +[2026-06-23 19:08:09,964][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.1820027083158493, 'train/grad_norm_pre_clip': 44.94675064086914, 'timing/data': 0.0006200061179697514, 'timing/model': 1.6913842959329486, 'learning_rate/qwen_vl_interface': 7.675806213992706e-06, 'learning_rate/action_model': 7.675806213992707e-05, 'epoch': 0.66}) +[2026-06-23 19:08:13,292][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.22813883423805237, 'train/grad_norm_pre_clip': 35.59434509277344, 'timing/data': 0.0006056178826838732, 'timing/model': 1.6895514822099358, 'learning_rate/qwen_vl_interface': 7.673187379988388e-06, 'learning_rate/action_model': 7.673187379988388e-05, 'epoch': 0.66}) +[2026-06-23 19:08:16,626][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.09911323338747025, 'train/grad_norm_pre_clip': 17.440391540527344, 'timing/data': 0.0006257060449570417, 'timing/model': 1.6909492199774832, 'learning_rate/qwen_vl_interface': 7.67056754990327e-06, 'learning_rate/action_model': 7.670567549903269e-05, 'epoch': 0.66}) +[2026-06-23 19:08:19,955][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.17765799164772034, 'train/grad_norm_pre_clip': 61.58728790283203, 'timing/data': 0.0006053878460079432, 'timing/model': 1.6874822238460183, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.66}) +[2026-06-23 19:08:23,282][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.17372974753379822, 'train/grad_norm_pre_clip': 26.941688537597656, 'timing/data': 0.0006208771374076605, 'timing/model': 1.6898062040563673, 'learning_rate/qwen_vl_interface': 7.665324905798694e-06, 'learning_rate/action_model': 7.665324905798695e-05, 'epoch': 0.66}) +[2026-06-23 19:08:26,616][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.09038417786359787, 'train/grad_norm_pre_clip': 22.590560913085938, 'timing/data': 0.0006032940000295639, 'timing/model': 1.6944187548942864, 'learning_rate/qwen_vl_interface': 7.66270209393429e-06, 'learning_rate/action_model': 7.66270209393429e-05, 'epoch': 0.67}) +[2026-06-23 19:08:30,049][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.23236849904060364, 'train/grad_norm_pre_clip': 25.039674758911133, 'timing/data': 0.0006769129540771246, 'timing/model': 1.7891648691147566, 'learning_rate/qwen_vl_interface': 7.66007829029919e-06, 'learning_rate/action_model': 7.660078290299191e-05, 'epoch': 0.67}) +[2026-06-23 19:08:33,383][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.044483281672000885, 'train/grad_norm_pre_clip': 26.244888305664062, 'timing/data': 0.0007079809438437223, 'timing/model': 1.6939208810217679, 'learning_rate/qwen_vl_interface': 7.657453495971942e-06, 'learning_rate/action_model': 7.657453495971943e-05, 'epoch': 0.67}) +[2026-06-23 19:08:36,704][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.13914695382118225, 'train/grad_norm_pre_clip': 36.97007751464844, 'timing/data': 0.0006954269483685493, 'timing/model': 1.6874367340933532, 'learning_rate/qwen_vl_interface': 7.6548277120315e-06, 'learning_rate/action_model': 7.6548277120315e-05, 'epoch': 0.67}) +[2026-06-23 19:08:40,022][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.27716660499572754, 'train/grad_norm_pre_clip': 50.071434020996094, 'timing/data': 0.000591290881857276, 'timing/model': 1.6859746461268514, 'learning_rate/qwen_vl_interface': 7.652200939557225e-06, 'learning_rate/action_model': 7.652200939557225e-05, 'epoch': 0.67}) +[2026-06-23 19:08:43,355][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.10878801345825195, 'train/grad_norm_pre_clip': 45.868778228759766, 'timing/data': 0.0006588790565729141, 'timing/model': 1.692664392059669, 'learning_rate/qwen_vl_interface': 7.649573179628881e-06, 'learning_rate/action_model': 7.649573179628881e-05, 'epoch': 0.67}) +[2026-06-23 19:08:46,682][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.20601411163806915, 'train/grad_norm_pre_clip': 26.881479263305664, 'timing/data': 0.0007597480434924364, 'timing/model': 1.6879781938623637, 'learning_rate/qwen_vl_interface': 7.646944433326645e-06, 'learning_rate/action_model': 7.646944433326643e-05, 'epoch': 0.67}) +[2026-06-23 19:08:50,001][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.16884085536003113, 'train/grad_norm_pre_clip': 24.552431106567383, 'timing/data': 0.0006670639850199223, 'timing/model': 1.6825723841320723, 'learning_rate/qwen_vl_interface': 7.64431470173109e-06, 'learning_rate/action_model': 7.64431470173109e-05, 'epoch': 0.67}) +[2026-06-23 19:08:53,326][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.15330560505390167, 'train/grad_norm_pre_clip': 77.9960708618164, 'timing/data': 0.0007732040248811245, 'timing/model': 1.6897440049797297, 'learning_rate/qwen_vl_interface': 7.641683985923203e-06, 'learning_rate/action_model': 7.641683985923203e-05, 'epoch': 0.67}) +[2026-06-23 19:08:56,655][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.2246648222208023, 'train/grad_norm_pre_clip': 133.23416137695312, 'timing/data': 0.0007067390251904726, 'timing/model': 1.689420608105138, 'learning_rate/qwen_vl_interface': 7.63905228698437e-06, 'learning_rate/action_model': 7.63905228698437e-05, 'epoch': 0.67}) +[2026-06-23 19:08:59,985][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.2879065275192261, 'train/grad_norm_pre_clip': 113.38062286376953, 'timing/data': 0.0006309659220278263, 'timing/model': 1.693258403101936, 'learning_rate/qwen_vl_interface': 7.636419605996385e-06, 'learning_rate/action_model': 7.636419605996384e-05, 'epoch': 0.67}) +[2026-06-23 19:09:03,318][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.11292565613985062, 'train/grad_norm_pre_clip': 30.83081817626953, 'timing/data': 0.001000471180304885, 'timing/model': 1.6939814530778676, 'learning_rate/qwen_vl_interface': 7.633785944041444e-06, 'learning_rate/action_model': 7.633785944041444e-05, 'epoch': 0.67}) +[2026-06-23 19:09:06,641][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.211263507604599, 'train/grad_norm_pre_clip': 28.160587310791016, 'timing/data': 0.0006136531010270119, 'timing/model': 1.6899797508958727, 'learning_rate/qwen_vl_interface': 7.631151302202145e-06, 'learning_rate/action_model': 7.631151302202144e-05, 'epoch': 0.67}) +[2026-06-23 19:09:09,967][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.1650775671005249, 'train/grad_norm_pre_clip': 44.750205993652344, 'timing/data': 0.0007443090435117483, 'timing/model': 1.6889916751533747, 'learning_rate/qwen_vl_interface': 7.628515681561488e-06, 'learning_rate/action_model': 7.628515681561487e-05, 'epoch': 0.67}) +[2026-06-23 19:09:13,301][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.15971827507019043, 'train/grad_norm_pre_clip': 55.7406120300293, 'timing/data': 0.0006107480730861425, 'timing/model': 1.6891835960559547, 'learning_rate/qwen_vl_interface': 7.62587908320288e-06, 'learning_rate/action_model': 7.62587908320288e-05, 'epoch': 0.67}) +[2026-06-23 19:09:16,632][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.18324536085128784, 'train/grad_norm_pre_clip': 34.2843132019043, 'timing/data': 0.0006233621388673782, 'timing/model': 1.6866805818863213, 'learning_rate/qwen_vl_interface': 7.6232415082101255e-06, 'learning_rate/action_model': 7.623241508210124e-05, 'epoch': 0.67}) +[2026-06-23 19:09:20,245][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.13396179676055908, 'train/grad_norm_pre_clip': 14.264554023742676, 'timing/data': 0.0005891779437661171, 'timing/model': 1.973096810048446, 'learning_rate/qwen_vl_interface': 7.620602957667432e-06, 'learning_rate/action_model': 7.620602957667432e-05, 'epoch': 0.67}) +[2026-06-23 19:09:23,583][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.17650333046913147, 'train/grad_norm_pre_clip': 45.78153610229492, 'timing/data': 0.0006164088845252991, 'timing/model': 1.6915081180632114, 'learning_rate/qwen_vl_interface': 7.617963432659408e-06, 'learning_rate/action_model': 7.617963432659408e-05, 'epoch': 0.67}) +[2026-06-23 19:09:26,911][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.11977427452802658, 'train/grad_norm_pre_clip': 13.448909759521484, 'timing/data': 0.000603764783591032, 'timing/model': 1.6935207529459149, 'learning_rate/qwen_vl_interface': 7.615322934271063e-06, 'learning_rate/action_model': 7.615322934271064e-05, 'epoch': 0.67}) +[2026-06-23 19:09:30,245][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.07386326044797897, 'train/grad_norm_pre_clip': 35.88164520263672, 'timing/data': 0.0006011798977851868, 'timing/model': 1.6926826080307364, 'learning_rate/qwen_vl_interface': 7.6126814635878055e-06, 'learning_rate/action_model': 7.612681463587805e-05, 'epoch': 0.67}) +[2026-06-23 19:09:33,579][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.09560836106538773, 'train/grad_norm_pre_clip': 25.254514694213867, 'timing/data': 0.0005996869876980782, 'timing/model': 1.6899146458599716, 'learning_rate/qwen_vl_interface': 7.6100390216954445e-06, 'learning_rate/action_model': 7.610039021695445e-05, 'epoch': 0.67}) +[2026-06-23 19:09:36,920][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.154771089553833, 'train/grad_norm_pre_clip': 19.754947662353516, 'timing/data': 0.0006123911589384079, 'timing/model': 1.6959516240749508, 'learning_rate/qwen_vl_interface': 7.607395609680189e-06, 'learning_rate/action_model': 7.60739560968019e-05, 'epoch': 0.67}) +[2026-06-23 19:09:40,261][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.1486024260520935, 'train/grad_norm_pre_clip': 61.76459503173828, 'timing/data': 0.0006129620596766472, 'timing/model': 1.7002370099071413, 'learning_rate/qwen_vl_interface': 7.604751228628644e-06, 'learning_rate/action_model': 7.604751228628644e-05, 'epoch': 0.67}) +[2026-06-23 19:09:43,604][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.0986645296216011, 'train/grad_norm_pre_clip': 32.426334381103516, 'timing/data': 0.0005908298771828413, 'timing/model': 1.698511536931619, 'learning_rate/qwen_vl_interface': 7.602105879627814e-06, 'learning_rate/action_model': 7.602105879627814e-05, 'epoch': 0.67}) +[2026-06-23 19:09:46,947][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.219985231757164, 'train/grad_norm_pre_clip': 51.33903884887695, 'timing/data': 0.0006126409862190485, 'timing/model': 1.6996539460960776, 'learning_rate/qwen_vl_interface': 7.599459563765107e-06, 'learning_rate/action_model': 7.599459563765106e-05, 'epoch': 0.67}) +[2026-06-23 19:09:50,285][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.12701092660427094, 'train/grad_norm_pre_clip': 23.001401901245117, 'timing/data': 0.0006122910417616367, 'timing/model': 1.6977206219453365, 'learning_rate/qwen_vl_interface': 7.596812282128319e-06, 'learning_rate/action_model': 7.596812282128319e-05, 'epoch': 0.67}) +[2026-06-23 19:09:53,627][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.23121261596679688, 'train/grad_norm_pre_clip': 71.70146179199219, 'timing/data': 0.0006233109161257744, 'timing/model': 1.6986696780659258, 'learning_rate/qwen_vl_interface': 7.594164035805648e-06, 'learning_rate/action_model': 7.594164035805648e-05, 'epoch': 0.68}) +[2026-06-23 19:09:56,970][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.10424824804067612, 'train/grad_norm_pre_clip': 56.062313079833984, 'timing/data': 0.000619885977357626, 'timing/model': 1.6973325838334858, 'learning_rate/qwen_vl_interface': 7.59151482588569e-06, 'learning_rate/action_model': 7.59151482588569e-05, 'epoch': 0.68}) +[2026-06-23 19:10:00,294][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.1437579244375229, 'train/grad_norm_pre_clip': 58.60237503051758, 'timing/data': 0.0006148950196802616, 'timing/model': 1.6874515018425882, 'learning_rate/qwen_vl_interface': 7.588864653457433e-06, 'learning_rate/action_model': 7.588864653457432e-05, 'epoch': 0.68}) +[2026-06-23 19:10:03,624][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.07967621088027954, 'train/grad_norm_pre_clip': 20.08049964904785, 'timing/data': 0.00071759894490242, 'timing/model': 1.694840301061049, 'learning_rate/qwen_vl_interface': 7.5862135196102634e-06, 'learning_rate/action_model': 7.586213519610263e-05, 'epoch': 0.68}) +[2026-06-23 19:10:06,956][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.129476398229599, 'train/grad_norm_pre_clip': 13.821518898010254, 'timing/data': 0.0006115802098065615, 'timing/model': 1.6915101630147547, 'learning_rate/qwen_vl_interface': 7.583561425433962e-06, 'learning_rate/action_model': 7.583561425433962e-05, 'epoch': 0.68}) +[2026-06-23 19:10:35,643][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1750 +[2026-06-23 19:10:35,644][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.04849224165081978, 'train/grad_norm_pre_clip': 16.879253387451172, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 16.64000000000008, 'rl_games_eval/mid_train/mean_length': 152.2, 'rl_games_eval/mid_train/std_reward': 17.710177864719626, 'rl_games_eval/mid_train/std_length': 143.28209937043778, 'rl_games_eval/mid_train/macro_mean_reward': 16.64000000000008, 'rl_games_eval/mid_train/macro_mean_length': 152.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 16.64000000000008, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 152.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 17.710177864719626, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 143.28209937043778, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 16.64000000000008, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0011699190363287926, 'timing/model': 1.9981155910063535, 'learning_rate/qwen_vl_interface': 7.580908372018706e-06, 'learning_rate/action_model': 7.580908372018705e-05, 'epoch': 0.68}) +[2026-06-23 19:10:38,939][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.1301886886358261, 'train/grad_norm_pre_clip': 32.44518280029297, 'timing/data': 0.001048531150445342, 'timing/model': 1.6782143369782716, 'learning_rate/qwen_vl_interface': 7.5782543604550655e-06, 'learning_rate/action_model': 7.578254360455064e-05, 'epoch': 0.68}) +[2026-06-23 19:10:42,257][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.15030747652053833, 'train/grad_norm_pre_clip': 60.92503356933594, 'timing/data': 0.000682152109220624, 'timing/model': 1.6828719631303102, 'learning_rate/qwen_vl_interface': 7.575599391834e-06, 'learning_rate/action_model': 7.575599391834e-05, 'epoch': 0.68}) +[2026-06-23 19:10:45,567][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.21341371536254883, 'train/grad_norm_pre_clip': 22.879518508911133, 'timing/data': 0.0006056991405785084, 'timing/model': 1.679164073895663, 'learning_rate/qwen_vl_interface': 7.572943467246871e-06, 'learning_rate/action_model': 7.57294346724687e-05, 'epoch': 0.68}) +[2026-06-23 19:10:48,881][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.18603724241256714, 'train/grad_norm_pre_clip': 52.51789093017578, 'timing/data': 0.0008771189022809267, 'timing/model': 1.6885374041739851, 'learning_rate/qwen_vl_interface': 7.570286587785428e-06, 'learning_rate/action_model': 7.570286587785428e-05, 'epoch': 0.68}) +[2026-06-23 19:10:52,200][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.16598014533519745, 'train/grad_norm_pre_clip': 38.82851791381836, 'timing/data': 0.0006049079820513725, 'timing/model': 1.687403011834249, 'learning_rate/qwen_vl_interface': 7.567628754541815e-06, 'learning_rate/action_model': 7.567628754541814e-05, 'epoch': 0.68}) +[2026-06-23 19:10:55,519][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.211599662899971, 'train/grad_norm_pre_clip': 35.91217803955078, 'timing/data': 0.0007074100431054831, 'timing/model': 1.686486260034144, 'learning_rate/qwen_vl_interface': 7.564969968608565e-06, 'learning_rate/action_model': 7.564969968608565e-05, 'epoch': 0.68}) +[2026-06-23 19:10:58,832][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.17752380669116974, 'train/grad_norm_pre_clip': 17.375333786010742, 'timing/data': 0.000914999982342124, 'timing/model': 1.6797940109390765, 'learning_rate/qwen_vl_interface': 7.562310231078605e-06, 'learning_rate/action_model': 7.562310231078604e-05, 'epoch': 0.68}) +[2026-06-23 19:11:02,157][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.08714263886213303, 'train/grad_norm_pre_clip': 22.227237701416016, 'timing/data': 0.0006911589298397303, 'timing/model': 1.6881342679262161, 'learning_rate/qwen_vl_interface': 7.559649543045252e-06, 'learning_rate/action_model': 7.559649543045252e-05, 'epoch': 0.68}) +[2026-06-23 19:11:05,489][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.2252025455236435, 'train/grad_norm_pre_clip': 49.712825775146484, 'timing/data': 0.0008056238293647766, 'timing/model': 1.6959830629639328, 'learning_rate/qwen_vl_interface': 7.556987905602215e-06, 'learning_rate/action_model': 7.556987905602215e-05, 'epoch': 0.68}) +[2026-06-23 19:11:08,821][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.2159135639667511, 'train/grad_norm_pre_clip': 63.22003936767578, 'timing/data': 0.0006046469788998365, 'timing/model': 1.6924179818015546, 'learning_rate/qwen_vl_interface': 7.554325319843593e-06, 'learning_rate/action_model': 7.554325319843592e-05, 'epoch': 0.68}) +[2026-06-23 19:11:12,146][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.21905377507209778, 'train/grad_norm_pre_clip': 63.74388885498047, 'timing/data': 0.0006283309776335955, 'timing/model': 1.685663236072287, 'learning_rate/qwen_vl_interface': 7.551661786863874e-06, 'learning_rate/action_model': 7.551661786863874e-05, 'epoch': 0.68}) +[2026-06-23 19:11:15,477][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.3250797688961029, 'train/grad_norm_pre_clip': 64.10874938964844, 'timing/data': 0.0006469457875937223, 'timing/model': 1.6926366749685258, 'learning_rate/qwen_vl_interface': 7.548997307757936e-06, 'learning_rate/action_model': 7.548997307757935e-05, 'epoch': 0.68}) +[2026-06-23 19:11:19,111][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.18228355050086975, 'train/grad_norm_pre_clip': 53.15264129638672, 'timing/data': 0.00062360311858356, 'timing/model': 1.690476904856041, 'learning_rate/qwen_vl_interface': 7.546331883621044e-06, 'learning_rate/action_model': 7.546331883621044e-05, 'epoch': 0.68}) +[2026-06-23 19:11:22,437][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.15903149545192719, 'train/grad_norm_pre_clip': 44.04191207885742, 'timing/data': 0.0006151061970740557, 'timing/model': 1.6886343129444867, 'learning_rate/qwen_vl_interface': 7.543665515548857e-06, 'learning_rate/action_model': 7.543665515548856e-05, 'epoch': 0.68}) +[2026-06-23 19:11:25,765][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.10432582348585129, 'train/grad_norm_pre_clip': 29.615530014038086, 'timing/data': 0.0006031638476997614, 'timing/model': 1.688212261069566, 'learning_rate/qwen_vl_interface': 7.540998204637413e-06, 'learning_rate/action_model': 7.540998204637413e-05, 'epoch': 0.68}) +[2026-06-23 19:11:29,091][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.1522737741470337, 'train/grad_norm_pre_clip': 71.07511138916016, 'timing/data': 0.0006118800956755877, 'timing/model': 1.6873653889633715, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.68}) +[2026-06-23 19:11:32,426][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.15537133812904358, 'train/grad_norm_pre_clip': 60.54435729980469, 'timing/data': 0.0006100570317357779, 'timing/model': 1.690508628031239, 'learning_rate/qwen_vl_interface': 7.535660758682877e-06, 'learning_rate/action_model': 7.535660758682876e-05, 'epoch': 0.68}) +[2026-06-23 19:11:35,759][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.17780110239982605, 'train/grad_norm_pre_clip': 67.64810943603516, 'timing/data': 0.0007043040823191404, 'timing/model': 1.692808171035722, 'learning_rate/qwen_vl_interface': 7.532990625833806e-06, 'learning_rate/action_model': 7.532990625833805e-05, 'epoch': 0.68}) +[2026-06-23 19:11:39,088][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.15763084590435028, 'train/grad_norm_pre_clip': 49.35941696166992, 'timing/data': 0.0007219468243420124, 'timing/model': 1.6885411099065095, 'learning_rate/qwen_vl_interface': 7.5303195545335225e-06, 'learning_rate/action_model': 7.530319554533522e-05, 'epoch': 0.68}) +[2026-06-23 19:11:42,416][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.32358577847480774, 'train/grad_norm_pre_clip': 91.67957305908203, 'timing/data': 0.0006108691450208426, 'timing/model': 1.6944657678250223, 'learning_rate/qwen_vl_interface': 7.527647545880007e-06, 'learning_rate/action_model': 7.527647545880007e-05, 'epoch': 0.69}) +[2026-06-23 19:11:45,742][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.12712904810905457, 'train/grad_norm_pre_clip': 36.01435852050781, 'timing/data': 0.0006734160706400871, 'timing/model': 1.692762266146019, 'learning_rate/qwen_vl_interface': 7.52497460097162e-06, 'learning_rate/action_model': 7.52497460097162e-05, 'epoch': 0.69}) +[2026-06-23 19:11:49,065][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.17446960508823395, 'train/grad_norm_pre_clip': 30.73128890991211, 'timing/data': 0.000717769144102931, 'timing/model': 1.6883621260058135, 'learning_rate/qwen_vl_interface': 7.522300720907105e-06, 'learning_rate/action_model': 7.522300720907105e-05, 'epoch': 0.69}) +[2026-06-23 19:11:52,395][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.2283964902162552, 'train/grad_norm_pre_clip': 14.147595405578613, 'timing/data': 0.0007933310698717833, 'timing/model': 1.6949664610438049, 'learning_rate/qwen_vl_interface': 7.519625906785601e-06, 'learning_rate/action_model': 7.5196259067856e-05, 'epoch': 0.69}) +[2026-06-23 19:11:55,730][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.10108107328414917, 'train/grad_norm_pre_clip': 13.128485679626465, 'timing/data': 0.0007017189636826515, 'timing/model': 1.6942180851474404, 'learning_rate/qwen_vl_interface': 7.516950159706614e-06, 'learning_rate/action_model': 7.516950159706614e-05, 'epoch': 0.69}) +[2026-06-23 19:11:59,067][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.22144263982772827, 'train/grad_norm_pre_clip': 16.825803756713867, 'timing/data': 0.0006694679614156485, 'timing/model': 1.695528909098357, 'learning_rate/qwen_vl_interface': 7.514273480770047e-06, 'learning_rate/action_model': 7.514273480770047e-05, 'epoch': 0.69}) +[2026-06-23 19:12:02,408][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.13996091485023499, 'train/grad_norm_pre_clip': 17.59073829650879, 'timing/data': 0.000650492962449789, 'timing/model': 1.695449901977554, 'learning_rate/qwen_vl_interface': 7.511595871076181e-06, 'learning_rate/action_model': 7.511595871076181e-05, 'epoch': 0.69}) +[2026-06-23 19:12:05,738][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.185845285654068, 'train/grad_norm_pre_clip': 28.67477035522461, 'timing/data': 0.0008080890402197838, 'timing/model': 1.688741222023964, 'learning_rate/qwen_vl_interface': 7.50891733172568e-06, 'learning_rate/action_model': 7.50891733172568e-05, 'epoch': 0.69}) +[2026-06-23 19:12:09,312][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.16000132262706757, 'train/grad_norm_pre_clip': 21.54069709777832, 'timing/data': 0.0006866010371595621, 'timing/model': 1.6950209569185972, 'learning_rate/qwen_vl_interface': 7.506237863819593e-06, 'learning_rate/action_model': 7.506237863819593e-05, 'epoch': 0.69}) +[2026-06-23 19:12:12,647][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.17108137905597687, 'train/grad_norm_pre_clip': 15.467697143554688, 'timing/data': 0.0007166170980781317, 'timing/model': 1.693572954973206, 'learning_rate/qwen_vl_interface': 7.503557468459345e-06, 'learning_rate/action_model': 7.503557468459344e-05, 'epoch': 0.69}) +[2026-06-23 19:12:15,978][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.10383080691099167, 'train/grad_norm_pre_clip': 33.87103271484375, 'timing/data': 0.0006925109773874283, 'timing/model': 1.6902594771236181, 'learning_rate/qwen_vl_interface': 7.5008761467467435e-06, 'learning_rate/action_model': 7.500876146746743e-05, 'epoch': 0.69}) +[2026-06-23 19:12:19,307][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.11341872811317444, 'train/grad_norm_pre_clip': 35.388397216796875, 'timing/data': 0.0006317379884421825, 'timing/model': 1.6873027300462127, 'learning_rate/qwen_vl_interface': 7.498193899783983e-06, 'learning_rate/action_model': 7.498193899783983e-05, 'epoch': 0.69}) +[2026-06-23 19:12:22,647][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.06129782646894455, 'train/grad_norm_pre_clip': 46.72447967529297, 'timing/data': 0.0006032041274011135, 'timing/model': 1.6935662550386041, 'learning_rate/qwen_vl_interface': 7.495510728673631e-06, 'learning_rate/action_model': 7.495510728673632e-05, 'epoch': 0.69}) +[2026-06-23 19:12:25,986][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.18209540843963623, 'train/grad_norm_pre_clip': 35.97675704956055, 'timing/data': 0.0008416329510509968, 'timing/model': 1.6966600101441145, 'learning_rate/qwen_vl_interface': 7.492826634518642e-06, 'learning_rate/action_model': 7.492826634518641e-05, 'epoch': 0.69}) +[2026-06-23 19:12:29,314][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.12965495884418488, 'train/grad_norm_pre_clip': 58.46003341674805, 'timing/data': 0.000666863052174449, 'timing/model': 1.6914162850007415, 'learning_rate/qwen_vl_interface': 7.4901416184223404e-06, 'learning_rate/action_model': 7.49014161842234e-05, 'epoch': 0.69}) +[2026-06-23 19:12:32,651][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.1848563551902771, 'train/grad_norm_pre_clip': 51.45810317993164, 'timing/data': 0.0006137939635664225, 'timing/model': 1.6957011490594596, 'learning_rate/qwen_vl_interface': 7.487455681488437e-06, 'learning_rate/action_model': 7.487455681488436e-05, 'epoch': 0.69}) +[2026-06-23 19:12:35,985][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.23941576480865479, 'train/grad_norm_pre_clip': 81.66808319091797, 'timing/data': 0.0006221500225365162, 'timing/model': 1.6936331030447036, 'learning_rate/qwen_vl_interface': 7.484768824821018e-06, 'learning_rate/action_model': 7.484768824821018e-05, 'epoch': 0.69}) +[2026-06-23 19:12:39,313][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.15398113429546356, 'train/grad_norm_pre_clip': 20.80402946472168, 'timing/data': 0.0006115101277828217, 'timing/model': 1.6877076050732285, 'learning_rate/qwen_vl_interface': 7.482081049524551e-06, 'learning_rate/action_model': 7.482081049524551e-05, 'epoch': 0.69}) +[2026-06-23 19:12:42,647][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.17704695463180542, 'train/grad_norm_pre_clip': 14.037513732910156, 'timing/data': 0.0006153869908303022, 'timing/model': 1.6938514539506286, 'learning_rate/qwen_vl_interface': 7.479392356703878e-06, 'learning_rate/action_model': 7.479392356703878e-05, 'epoch': 0.69}) +[2026-06-23 19:12:45,983][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.22456413507461548, 'train/grad_norm_pre_clip': 40.85854721069336, 'timing/data': 0.0005975628737360239, 'timing/model': 1.6957807708531618, 'learning_rate/qwen_vl_interface': 7.476702747464219e-06, 'learning_rate/action_model': 7.476702747464219e-05, 'epoch': 0.69}) +[2026-06-23 19:12:49,326][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.13570287823677063, 'train/grad_norm_pre_clip': 13.082566261291504, 'timing/data': 0.0006116798613220453, 'timing/model': 1.697972360998392, 'learning_rate/qwen_vl_interface': 7.474012222911169e-06, 'learning_rate/action_model': 7.474012222911169e-05, 'epoch': 0.69}) +[2026-06-23 19:12:52,671][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.18291911482810974, 'train/grad_norm_pre_clip': 29.042943954467773, 'timing/data': 0.0006051280070096254, 'timing/model': 1.6974922800436616, 'learning_rate/qwen_vl_interface': 7.471320784150703e-06, 'learning_rate/action_model': 7.471320784150703e-05, 'epoch': 0.69}) +[2026-06-23 19:12:56,155][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.20431987941265106, 'train/grad_norm_pre_clip': 21.10089874267578, 'timing/data': 0.0006979918107390404, 'timing/model': 1.6923475798685104, 'learning_rate/qwen_vl_interface': 7.468628432289169e-06, 'learning_rate/action_model': 7.468628432289168e-05, 'epoch': 0.69}) +[2026-06-23 19:12:59,483][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.09726708382368088, 'train/grad_norm_pre_clip': 23.807552337646484, 'timing/data': 0.0005935949739068747, 'timing/model': 1.6942112429533154, 'learning_rate/qwen_vl_interface': 7.465935168433292e-06, 'learning_rate/action_model': 7.465935168433292e-05, 'epoch': 0.69}) +[2026-06-23 19:13:02,808][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.12367244064807892, 'train/grad_norm_pre_clip': 18.890474319458008, 'timing/data': 0.000682904152199626, 'timing/model': 1.6923525200691074, 'learning_rate/qwen_vl_interface': 7.463240993690168e-06, 'learning_rate/action_model': 7.463240993690168e-05, 'epoch': 0.69}) +[2026-06-23 19:13:06,143][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.13224956393241882, 'train/grad_norm_pre_clip': 26.00756072998047, 'timing/data': 0.0006146559026092291, 'timing/model': 1.6960672431159765, 'learning_rate/qwen_vl_interface': 7.460545909167276e-06, 'learning_rate/action_model': 7.460545909167275e-05, 'epoch': 0.69}) +[2026-06-23 19:13:09,476][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.15363454818725586, 'train/grad_norm_pre_clip': 33.425323486328125, 'timing/data': 0.0006122400518506765, 'timing/model': 1.6924999682232738, 'learning_rate/qwen_vl_interface': 7.457849915972459e-06, 'learning_rate/action_model': 7.457849915972458e-05, 'epoch': 0.7}) +[2026-06-23 19:13:12,800][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.209406316280365, 'train/grad_norm_pre_clip': 43.15019607543945, 'timing/data': 0.0007122389506548643, 'timing/model': 1.68626644089818, 'learning_rate/qwen_vl_interface': 7.455153015213941e-06, 'learning_rate/action_model': 7.45515301521394e-05, 'epoch': 0.7}) +[2026-06-23 19:13:16,122][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.08707068115472794, 'train/grad_norm_pre_clip': 10.043779373168945, 'timing/data': 0.0006650001741945744, 'timing/model': 1.68653927417472, 'learning_rate/qwen_vl_interface': 7.452455208000311e-06, 'learning_rate/action_model': 7.452455208000311e-05, 'epoch': 0.7}) +[2026-06-23 19:13:19,448][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.29197296500205994, 'train/grad_norm_pre_clip': 13.60958480834961, 'timing/data': 0.0006201150827109814, 'timing/model': 1.6888625619467348, 'learning_rate/qwen_vl_interface': 7.449756495440541e-06, 'learning_rate/action_model': 7.44975649544054e-05, 'epoch': 0.7}) +[2026-06-23 19:13:22,779][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1800 +[2026-06-23 19:14:55,586][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1800 +[2026-06-23 19:14:55,586][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1800 +[2026-06-23 19:16:07,681][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1800 +[2026-06-23 19:16:07,682][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.13197356462478638, 'train/grad_norm_pre_clip': 57.46784591674805, 'eval/action_loss/seconds': 92.80479093478061, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.08328328281641006, 'eval/latency_2/loss': 0.08328328281641006, 'eval/flappy/loss': 0.08328328281641006, 'eval/flappy/latency_2/loss': 0.08328328281641006, 'eval/flap_precision': 0.9270620313565099, 'eval/flap_recall': 0.9176788124156545, 'eval/flap_f1': 0.9223465581553069, 'eval/flap_rate_model': 0.20375, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9223465581553069, 'eval/latency_2/flap_precision': 0.9270620313565099, 'eval/latency_2/flap_recall': 0.9176788124156545, 'eval/latency_2/flap_f1': 0.9223465581553069, 'eval/latency_2/flap_rate_model': 0.20375, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9223465581553069, 'eval/action_classification/data_wait_seconds': 0.9150279760360718, 'eval/action_classification/predict_seconds': 71.04178619384766, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0007295310497283936, 'timing/model': 1.692590950988233, 'learning_rate/qwen_vl_interface': 7.447056878643969e-06, 'learning_rate/action_model': 7.447056878643968e-05, 'epoch': 0.7}) +[2026-06-23 19:16:11,008][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.1502126008272171, 'train/grad_norm_pre_clip': 17.24942970275879, 'timing/data': 0.0007114480249583721, 'timing/model': 1.6919482450466603, 'learning_rate/qwen_vl_interface': 7.444356358720303e-06, 'learning_rate/action_model': 7.444356358720302e-05, 'epoch': 0.7}) +[2026-06-23 19:16:14,346][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.17534227669239044, 'train/grad_norm_pre_clip': 31.095422744750977, 'timing/data': 0.000852873083204031, 'timing/model': 1.7007277889642864, 'learning_rate/qwen_vl_interface': 7.4416549367796255e-06, 'learning_rate/action_model': 7.441654936779625e-05, 'epoch': 0.7}) +[2026-06-23 19:16:17,674][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.24815762042999268, 'train/grad_norm_pre_clip': 30.53800392150879, 'timing/data': 0.0006142950151115656, 'timing/model': 1.6886324828956276, 'learning_rate/qwen_vl_interface': 7.4389526139323895e-06, 'learning_rate/action_model': 7.438952613932389e-05, 'epoch': 0.7}) +[2026-06-23 19:16:21,001][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.25295647978782654, 'train/grad_norm_pre_clip': 31.866512298583984, 'timing/data': 0.0006144149228930473, 'timing/model': 1.6919037019833922, 'learning_rate/qwen_vl_interface': 7.436249391289419e-06, 'learning_rate/action_model': 7.436249391289418e-05, 'epoch': 0.7}) +[2026-06-23 19:16:24,330][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.13970474898815155, 'train/grad_norm_pre_clip': 15.487015724182129, 'timing/data': 0.0006267388816922903, 'timing/model': 1.694298294140026, 'learning_rate/qwen_vl_interface': 7.433545269961906e-06, 'learning_rate/action_model': 7.433545269961906e-05, 'epoch': 0.7}) +[2026-06-23 19:16:27,661][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.16582056879997253, 'train/grad_norm_pre_clip': 23.670297622680664, 'timing/data': 0.0006076928693801165, 'timing/model': 1.690530912950635, 'learning_rate/qwen_vl_interface': 7.430840251061411e-06, 'learning_rate/action_model': 7.430840251061411e-05, 'epoch': 0.7}) +[2026-06-23 19:16:30,987][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.1763041615486145, 'train/grad_norm_pre_clip': 24.382766723632812, 'timing/data': 0.0007152839098125696, 'timing/model': 1.689590675989166, 'learning_rate/qwen_vl_interface': 7.428134335699866e-06, 'learning_rate/action_model': 7.428134335699865e-05, 'epoch': 0.7}) +[2026-06-23 19:16:34,315][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.12255730479955673, 'train/grad_norm_pre_clip': 29.5448055267334, 'timing/data': 0.0006312460172921419, 'timing/model': 1.6894293129444122, 'learning_rate/qwen_vl_interface': 7.425427524989572e-06, 'learning_rate/action_model': 7.425427524989572e-05, 'epoch': 0.7}) +[2026-06-23 19:16:37,655][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.1146128699183464, 'train/grad_norm_pre_clip': 22.08319091796875, 'timing/data': 0.0006015310063958168, 'timing/model': 1.7019969769753516, 'learning_rate/qwen_vl_interface': 7.422719820043194e-06, 'learning_rate/action_model': 7.422719820043193e-05, 'epoch': 0.7}) +[2026-06-23 19:16:40,989][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.13630056381225586, 'train/grad_norm_pre_clip': 30.74420928955078, 'timing/data': 0.0006059890147298574, 'timing/model': 1.6919389611575752, 'learning_rate/qwen_vl_interface': 7.420011221973771e-06, 'learning_rate/action_model': 7.420011221973771e-05, 'epoch': 0.7}) +[2026-06-23 19:16:44,320][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.12922243773937225, 'train/grad_norm_pre_clip': 24.344053268432617, 'timing/data': 0.0006297731306403875, 'timing/model': 1.6899016320239753, 'learning_rate/qwen_vl_interface': 7.417301731894702e-06, 'learning_rate/action_model': 7.417301731894701e-05, 'epoch': 0.7}) +[2026-06-23 19:16:47,660][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.14736615121364594, 'train/grad_norm_pre_clip': 45.78441619873047, 'timing/data': 0.0006140240002423525, 'timing/model': 1.6994597190059721, 'learning_rate/qwen_vl_interface': 7.414591350919756e-06, 'learning_rate/action_model': 7.414591350919756e-05, 'epoch': 0.7}) +[2026-06-23 19:16:50,997][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.24830204248428345, 'train/grad_norm_pre_clip': 21.325536727905273, 'timing/data': 0.0006114291027188301, 'timing/model': 1.6928510069847107, 'learning_rate/qwen_vl_interface': 7.4118800801630715e-06, 'learning_rate/action_model': 7.411880080163072e-05, 'epoch': 0.7}) +[2026-06-23 19:16:54,434][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.2278508096933365, 'train/grad_norm_pre_clip': 45.48293685913086, 'timing/data': 0.0006108481902629137, 'timing/model': 1.694293909939006, 'learning_rate/qwen_vl_interface': 7.409167920739146e-06, 'learning_rate/action_model': 7.409167920739147e-05, 'epoch': 0.7}) +[2026-06-23 19:16:57,779][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.09007836878299713, 'train/grad_norm_pre_clip': 44.65335464477539, 'timing/data': 0.0006147259846329689, 'timing/model': 1.6990603390149772, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.7}) +[2026-06-23 19:17:01,118][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.18447649478912354, 'train/grad_norm_pre_clip': 68.02336120605469, 'timing/data': 0.0006286611314862967, 'timing/model': 1.696502752136439, 'learning_rate/qwen_vl_interface': 7.403740940349408e-06, 'learning_rate/action_model': 7.403740940349408e-05, 'epoch': 0.7}) +[2026-06-23 19:17:04,458][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.0767354965209961, 'train/grad_norm_pre_clip': 16.135272979736328, 'timing/data': 0.0006031030789017677, 'timing/model': 1.6939070329535753, 'learning_rate/qwen_vl_interface': 7.401026121614421e-06, 'learning_rate/action_model': 7.40102612161442e-05, 'epoch': 0.7}) +[2026-06-23 19:17:07,796][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.20200905203819275, 'train/grad_norm_pre_clip': 26.251296997070312, 'timing/data': 0.0006034339312463999, 'timing/model': 1.6993914130143821, 'learning_rate/qwen_vl_interface': 7.398310418673846e-06, 'learning_rate/action_model': 7.398310418673846e-05, 'epoch': 0.7}) +[2026-06-23 19:17:11,149][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.25793635845184326, 'train/grad_norm_pre_clip': 96.27562713623047, 'timing/data': 0.0006861400324851274, 'timing/model': 1.7036464919801801, 'learning_rate/qwen_vl_interface': 7.395593832644007e-06, 'learning_rate/action_model': 7.395593832644007e-05, 'epoch': 0.7}) +[2026-06-23 19:17:14,491][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.3946335017681122, 'train/grad_norm_pre_clip': 73.66233825683594, 'timing/data': 0.0006218089256435633, 'timing/model': 1.6938502669800073, 'learning_rate/qwen_vl_interface': 7.39287636464159e-06, 'learning_rate/action_model': 7.39287636464159e-05, 'epoch': 0.7}) +[2026-06-23 19:17:17,826][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.11062219738960266, 'train/grad_norm_pre_clip': 62.3284912109375, 'timing/data': 0.000616889912635088, 'timing/model': 1.6937755371909589, 'learning_rate/qwen_vl_interface': 7.3901580157836425e-06, 'learning_rate/action_model': 7.390158015783643e-05, 'epoch': 0.7}) +[2026-06-23 19:17:21,164][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.14286509156227112, 'train/grad_norm_pre_clip': 28.222393035888672, 'timing/data': 0.000605457928031683, 'timing/model': 1.693046536995098, 'learning_rate/qwen_vl_interface': 7.387438787187576e-06, 'learning_rate/action_model': 7.387438787187576e-05, 'epoch': 0.71}) +[2026-06-23 19:17:24,503][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.15920603275299072, 'train/grad_norm_pre_clip': 28.145851135253906, 'timing/data': 0.0005991770885884762, 'timing/model': 1.6936248149722815, 'learning_rate/qwen_vl_interface': 7.3847186799711625e-06, 'learning_rate/action_model': 7.384718679971162e-05, 'epoch': 0.71}) +[2026-06-23 19:17:27,841][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.25629907846450806, 'train/grad_norm_pre_clip': 102.9127426147461, 'timing/data': 0.0006145851220935583, 'timing/model': 1.695377049036324, 'learning_rate/qwen_vl_interface': 7.381997695252535e-06, 'learning_rate/action_model': 7.381997695252535e-05, 'epoch': 0.71}) +[2026-06-23 19:17:31,219][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.1489744633436203, 'train/grad_norm_pre_clip': 115.5963134765625, 'timing/data': 0.0005980138666927814, 'timing/model': 1.7318758370820433, 'learning_rate/qwen_vl_interface': 7.379275834150186e-06, 'learning_rate/action_model': 7.379275834150186e-05, 'epoch': 0.71}) +[2026-06-23 19:17:34,561][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.16094668209552765, 'train/grad_norm_pre_clip': 97.79817199707031, 'timing/data': 0.0006021310109645128, 'timing/model': 1.6953089910093695, 'learning_rate/qwen_vl_interface': 7.376553097782972e-06, 'learning_rate/action_model': 7.376553097782972e-05, 'epoch': 0.71}) +[2026-06-23 19:17:37,902][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.19221724569797516, 'train/grad_norm_pre_clip': 56.002201080322266, 'timing/data': 0.0006077231373637915, 'timing/model': 1.6972147040069103, 'learning_rate/qwen_vl_interface': 7.373829487270109e-06, 'learning_rate/action_model': 7.373829487270109e-05, 'epoch': 0.71}) +[2026-06-23 19:17:41,249][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.17770028114318848, 'train/grad_norm_pre_clip': 66.62029266357422, 'timing/data': 0.0006166689563542604, 'timing/model': 1.6991567239165306, 'learning_rate/qwen_vl_interface': 7.371105003731165e-06, 'learning_rate/action_model': 7.371105003731164e-05, 'epoch': 0.71}) +[2026-06-23 19:17:44,890][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.26184824109077454, 'train/grad_norm_pre_clip': 33.416873931884766, 'timing/data': 0.0005945980083197355, 'timing/model': 1.6969093310181051, 'learning_rate/qwen_vl_interface': 7.3683796482860755e-06, 'learning_rate/action_model': 7.368379648286075e-05, 'epoch': 0.71}) +[2026-06-23 19:17:48,234][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.17473316192626953, 'train/grad_norm_pre_clip': 21.919260025024414, 'timing/data': 0.0006058590952306986, 'timing/model': 1.697915341006592, 'learning_rate/qwen_vl_interface': 7.365653422055131e-06, 'learning_rate/action_model': 7.365653422055131e-05, 'epoch': 0.71}) +[2026-06-23 19:17:51,575][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.2280377745628357, 'train/grad_norm_pre_clip': 74.49481201171875, 'timing/data': 0.00071099610067904, 'timing/model': 1.6968018491752446, 'learning_rate/qwen_vl_interface': 7.3629263261589816e-06, 'learning_rate/action_model': 7.362926326158981e-05, 'epoch': 0.71}) +[2026-06-23 19:17:54,923][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.1472717821598053, 'train/grad_norm_pre_clip': 101.62010955810547, 'timing/data': 0.000728449085727334, 'timing/model': 1.6981702821794897, 'learning_rate/qwen_vl_interface': 7.36019836171863e-06, 'learning_rate/action_model': 7.360198361718629e-05, 'epoch': 0.71}) +[2026-06-23 19:17:58,261][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.3823288083076477, 'train/grad_norm_pre_clip': 134.50677490234375, 'timing/data': 0.0006908690556883812, 'timing/model': 1.696540147997439, 'learning_rate/qwen_vl_interface': 7.357469529855441e-06, 'learning_rate/action_model': 7.35746952985544e-05, 'epoch': 0.71}) +[2026-06-23 19:18:01,603][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.18266341090202332, 'train/grad_norm_pre_clip': 73.42057800292969, 'timing/data': 0.0007028609979897738, 'timing/model': 1.6976964918430895, 'learning_rate/qwen_vl_interface': 7.354739831691134e-06, 'learning_rate/action_model': 7.354739831691135e-05, 'epoch': 0.71}) +[2026-06-23 19:18:04,949][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.19021092355251312, 'train/grad_norm_pre_clip': 41.2737922668457, 'timing/data': 0.0006140938494354486, 'timing/model': 1.6995554459281266, 'learning_rate/qwen_vl_interface': 7.352009268347786e-06, 'learning_rate/action_model': 7.352009268347786e-05, 'epoch': 0.71}) +[2026-06-23 19:18:08,298][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.09440114349126816, 'train/grad_norm_pre_clip': 13.728182792663574, 'timing/data': 0.0006052569951862097, 'timing/model': 1.700387611053884, 'learning_rate/qwen_vl_interface': 7.349277840947828e-06, 'learning_rate/action_model': 7.349277840947828e-05, 'epoch': 0.71}) +[2026-06-23 19:18:11,632][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.19082574546337128, 'train/grad_norm_pre_clip': 103.28266143798828, 'timing/data': 0.0006885440088808537, 'timing/model': 1.6951899009291083, 'learning_rate/qwen_vl_interface': 7.346545550614046e-06, 'learning_rate/action_model': 7.346545550614046e-05, 'epoch': 0.71}) +[2026-06-23 19:18:14,971][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.17327411472797394, 'train/grad_norm_pre_clip': 105.28136444091797, 'timing/data': 0.0007140529341995716, 'timing/model': 1.6954464330337942, 'learning_rate/qwen_vl_interface': 7.343812398469582e-06, 'learning_rate/action_model': 7.343812398469581e-05, 'epoch': 0.71}) +[2026-06-23 19:18:18,315][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.44878092408180237, 'train/grad_norm_pre_clip': 137.38807678222656, 'timing/data': 0.0006763809360563755, 'timing/model': 1.6950451000593603, 'learning_rate/qwen_vl_interface': 7.34107838563793e-06, 'learning_rate/action_model': 7.341078385637929e-05, 'epoch': 0.71}) +[2026-06-23 19:18:21,655][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.24588508903980255, 'train/grad_norm_pre_clip': 102.45138549804688, 'timing/data': 0.0005981039721518755, 'timing/model': 1.6949217051733285, 'learning_rate/qwen_vl_interface': 7.338343513242941e-06, 'learning_rate/action_model': 7.338343513242941e-05, 'epoch': 0.71}) +[2026-06-23 19:18:24,996][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.2083129584789276, 'train/grad_norm_pre_clip': 84.37738037109375, 'timing/data': 0.0006332509219646454, 'timing/model': 1.6953444930259138, 'learning_rate/qwen_vl_interface': 7.335607782408818e-06, 'learning_rate/action_model': 7.335607782408818e-05, 'epoch': 0.71}) +[2026-06-23 19:18:28,340][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.3275855481624603, 'train/grad_norm_pre_clip': 79.66732025146484, 'timing/data': 0.0005905600264668465, 'timing/model': 1.6995076178573072, 'learning_rate/qwen_vl_interface': 7.3328711942601145e-06, 'learning_rate/action_model': 7.332871194260115e-05, 'epoch': 0.71}) +[2026-06-23 19:18:31,687][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.16274328529834747, 'train/grad_norm_pre_clip': 29.356576919555664, 'timing/data': 0.0008360019419342279, 'timing/model': 1.69244140596129, 'learning_rate/qwen_vl_interface': 7.33013374992174e-06, 'learning_rate/action_model': 7.33013374992174e-05, 'epoch': 0.71}) +[2026-06-23 19:18:35,264][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.27411818504333496, 'train/grad_norm_pre_clip': 25.85735321044922, 'timing/data': 0.0006664530374109745, 'timing/model': 1.699843280017376, 'learning_rate/qwen_vl_interface': 7.327395450518956e-06, 'learning_rate/action_model': 7.327395450518956e-05, 'epoch': 0.71}) +[2026-06-23 19:18:38,611][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.14743515849113464, 'train/grad_norm_pre_clip': 26.484241485595703, 'timing/data': 0.0006610429845750332, 'timing/model': 1.7007184349931777, 'learning_rate/qwen_vl_interface': 7.324656297177372e-06, 'learning_rate/action_model': 7.324656297177371e-05, 'epoch': 0.71}) +[2026-06-23 19:18:41,956][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.23223382234573364, 'train/grad_norm_pre_clip': 56.98413848876953, 'timing/data': 0.0005937661044299603, 'timing/model': 1.7016105321235955, 'learning_rate/qwen_vl_interface': 7.32191629102295e-06, 'learning_rate/action_model': 7.32191629102295e-05, 'epoch': 0.71}) +[2026-06-23 19:18:45,303][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.11087287962436676, 'train/grad_norm_pre_clip': 24.332260131835938, 'timing/data': 0.0008581029251217842, 'timing/model': 1.701731220120564, 'learning_rate/qwen_vl_interface': 7.319175433182004e-06, 'learning_rate/action_model': 7.319175433182004e-05, 'epoch': 0.71}) +[2026-06-23 19:18:48,646][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.15120601654052734, 'train/grad_norm_pre_clip': 23.267915725708008, 'timing/data': 0.0006342120468616486, 'timing/model': 1.697234848048538, 'learning_rate/qwen_vl_interface': 7.316433724781198e-06, 'learning_rate/action_model': 7.316433724781197e-05, 'epoch': 0.72}) +[2026-06-23 19:18:51,986][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.20730413496494293, 'train/grad_norm_pre_clip': 32.14458465576172, 'timing/data': 0.0008397689089179039, 'timing/model': 1.6956305329222232, 'learning_rate/qwen_vl_interface': 7.313691166947543e-06, 'learning_rate/action_model': 7.313691166947543e-05, 'epoch': 0.72}) +[2026-06-23 19:18:55,323][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.12018442898988724, 'train/grad_norm_pre_clip': 28.590003967285156, 'timing/data': 0.0008276260923594236, 'timing/model': 1.6914478498511016, 'learning_rate/qwen_vl_interface': 7.310947760808401e-06, 'learning_rate/action_model': 7.3109477608084e-05, 'epoch': 0.72}) +[2026-06-23 19:18:58,663][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.1439526081085205, 'train/grad_norm_pre_clip': 19.75416374206543, 'timing/data': 0.0006071219686418772, 'timing/model': 1.694799912162125, 'learning_rate/qwen_vl_interface': 7.308203507491485e-06, 'learning_rate/action_model': 7.308203507491484e-05, 'epoch': 0.72}) +[2026-06-23 19:19:02,003][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.1507176160812378, 'train/grad_norm_pre_clip': 16.51753044128418, 'timing/data': 0.0006035040132701397, 'timing/model': 1.6949423488695174, 'learning_rate/qwen_vl_interface': 7.3054584081248515e-06, 'learning_rate/action_model': 7.305458408124852e-05, 'epoch': 0.72}) +[2026-06-23 19:19:05,343][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.20224228501319885, 'train/grad_norm_pre_clip': 20.204883575439453, 'timing/data': 0.0006169290281832218, 'timing/model': 1.696549171116203, 'learning_rate/qwen_vl_interface': 7.302712463836907e-06, 'learning_rate/action_model': 7.302712463836907e-05, 'epoch': 0.72}) +[2026-06-23 19:19:08,688][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.10521340370178223, 'train/grad_norm_pre_clip': 30.12137794494629, 'timing/data': 0.0006176210008561611, 'timing/model': 1.7005079220980406, 'learning_rate/qwen_vl_interface': 7.2999656757564086e-06, 'learning_rate/action_model': 7.299965675756409e-05, 'epoch': 0.72}) +[2026-06-23 19:19:12,032][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.11079156398773193, 'train/grad_norm_pre_clip': 11.636667251586914, 'timing/data': 0.000866900198161602, 'timing/model': 1.6952884709462523, 'learning_rate/qwen_vl_interface': 7.297218045012453e-06, 'learning_rate/action_model': 7.297218045012453e-05, 'epoch': 0.72}) +[2026-06-23 19:19:15,382][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.1436353474855423, 'train/grad_norm_pre_clip': 26.961429595947266, 'timing/data': 0.0006038849242031574, 'timing/model': 1.7012240889016539, 'learning_rate/qwen_vl_interface': 7.29446957273449e-06, 'learning_rate/action_model': 7.294469572734489e-05, 'epoch': 0.72}) +[2026-06-23 19:19:18,731][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.21977104246616364, 'train/grad_norm_pre_clip': 38.154144287109375, 'timing/data': 0.0006613540463149548, 'timing/model': 1.7019418380223215, 'learning_rate/qwen_vl_interface': 7.2917202600523104e-06, 'learning_rate/action_model': 7.29172026005231e-05, 'epoch': 0.72}) +[2026-06-23 19:19:22,171][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.1738036870956421, 'train/grad_norm_pre_clip': 72.8222427368164, 'timing/data': 0.0005995158571749926, 'timing/model': 1.7944016200490296, 'learning_rate/qwen_vl_interface': 7.2889701080960565e-06, 'learning_rate/action_model': 7.288970108096056e-05, 'epoch': 0.72}) +[2026-06-23 19:19:25,518][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.10249745100736618, 'train/grad_norm_pre_clip': 36.45817184448242, 'timing/data': 0.0006145150400698185, 'timing/model': 1.6963329161517322, 'learning_rate/qwen_vl_interface': 7.286219117996209e-06, 'learning_rate/action_model': 7.286219117996207e-05, 'epoch': 0.72}) +[2026-06-23 19:19:28,861][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.16619254648685455, 'train/grad_norm_pre_clip': 14.12938404083252, 'timing/data': 0.0006621249485760927, 'timing/model': 1.6968617488164455, 'learning_rate/qwen_vl_interface': 7.283467290883598e-06, 'learning_rate/action_model': 7.283467290883598e-05, 'epoch': 0.72}) +[2026-06-23 19:19:32,197][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.11338406801223755, 'train/grad_norm_pre_clip': 13.192996978759766, 'timing/data': 0.0006012502126395702, 'timing/model': 1.6911911899223924, 'learning_rate/qwen_vl_interface': 7.2807146278893925e-06, 'learning_rate/action_model': 7.280714627889393e-05, 'epoch': 0.72}) +[2026-06-23 19:19:35,533][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.13387411832809448, 'train/grad_norm_pre_clip': 20.35445213317871, 'timing/data': 0.0005853299517184496, 'timing/model': 1.693587054964155, 'learning_rate/qwen_vl_interface': 7.277961130145111e-06, 'learning_rate/action_model': 7.277961130145111e-05, 'epoch': 0.72}) +[2026-06-23 19:19:38,869][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.18095457553863525, 'train/grad_norm_pre_clip': 37.89432144165039, 'timing/data': 0.0006014700047671795, 'timing/model': 1.6905187668744475, 'learning_rate/qwen_vl_interface': 7.2752067987826116e-06, 'learning_rate/action_model': 7.275206798782611e-05, 'epoch': 0.72}) +[2026-06-23 19:19:42,208][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.1910477727651596, 'train/grad_norm_pre_clip': 42.89637756347656, 'timing/data': 0.0008202509488910437, 'timing/model': 1.6948950339574367, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.72}) +[2026-06-23 19:19:45,549][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.21347357332706451, 'train/grad_norm_pre_clip': 42.152915954589844, 'timing/data': 0.0006870219949632883, 'timing/model': 1.696216729003936, 'learning_rate/qwen_vl_interface': 7.269695639732108e-06, 'learning_rate/action_model': 7.269695639732108e-05, 'epoch': 0.72}) +[2026-06-23 19:19:48,891][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.1561433970928192, 'train/grad_norm_pre_clip': 39.2992057800293, 'timing/data': 0.000755850225687027, 'timing/model': 1.697159610921517, 'learning_rate/qwen_vl_interface': 7.266938814309533e-06, 'learning_rate/action_model': 7.266938814309533e-05, 'epoch': 0.72}) +[2026-06-23 19:19:52,236][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.07968880236148834, 'train/grad_norm_pre_clip': 14.637443542480469, 'timing/data': 0.000620485981926322, 'timing/model': 1.7006641700863838, 'learning_rate/qwen_vl_interface': 7.264181159799598e-06, 'learning_rate/action_model': 7.264181159799597e-05, 'epoch': 0.72}) +[2026-06-23 19:19:55,584][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.180819571018219, 'train/grad_norm_pre_clip': 27.648426055908203, 'timing/data': 0.0008564400486648083, 'timing/model': 1.7001324300654233, 'learning_rate/qwen_vl_interface': 7.261422677335871e-06, 'learning_rate/action_model': 7.261422677335871e-05, 'epoch': 0.72}) +[2026-06-23 19:19:58,931][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.1500147134065628, 'train/grad_norm_pre_clip': 84.95317077636719, 'timing/data': 0.0006102870684117079, 'timing/model': 1.7000380319077522, 'learning_rate/qwen_vl_interface': 7.258663368052259e-06, 'learning_rate/action_model': 7.258663368052259e-05, 'epoch': 0.72}) +[2026-06-23 19:20:02,281][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.10467871278524399, 'train/grad_norm_pre_clip': 22.93926239013672, 'timing/data': 0.0008509301114827394, 'timing/model': 1.7005860330536962, 'learning_rate/qwen_vl_interface': 7.2559032330830105e-06, 'learning_rate/action_model': 7.25590323308301e-05, 'epoch': 0.72}) +[2026-06-23 19:20:05,622][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.12357500195503235, 'train/grad_norm_pre_clip': 18.58351707458496, 'timing/data': 0.0006210480351001024, 'timing/model': 1.6959606891032308, 'learning_rate/qwen_vl_interface': 7.253142273562712e-06, 'learning_rate/action_model': 7.253142273562712e-05, 'epoch': 0.72}) +[2026-06-23 19:20:08,967][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.07388360798358917, 'train/grad_norm_pre_clip': 52.31303024291992, 'timing/data': 0.0006026229821145535, 'timing/model': 1.6996426000259817, 'learning_rate/qwen_vl_interface': 7.25038049062629e-06, 'learning_rate/action_model': 7.250380490626289e-05, 'epoch': 0.72}) +[2026-06-23 19:20:12,408][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.2201460301876068, 'train/grad_norm_pre_clip': 23.109615325927734, 'timing/data': 0.0008558379486203194, 'timing/model': 1.6996575391385704, 'learning_rate/qwen_vl_interface': 7.24761788540901e-06, 'learning_rate/action_model': 7.24761788540901e-05, 'epoch': 0.72}) +[2026-06-23 19:20:15,745][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.09889163821935654, 'train/grad_norm_pre_clip': 18.1331729888916, 'timing/data': 0.0006073620170354843, 'timing/model': 1.696378534892574, 'learning_rate/qwen_vl_interface': 7.244854459046473e-06, 'learning_rate/action_model': 7.244854459046473e-05, 'epoch': 0.73}) +[2026-06-23 19:20:19,082][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.0804256796836853, 'train/grad_norm_pre_clip': 9.529520034790039, 'timing/data': 0.0006853288505226374, 'timing/model': 1.6930330758914351, 'learning_rate/qwen_vl_interface': 7.24209021267462e-06, 'learning_rate/action_model': 7.24209021267462e-05, 'epoch': 0.73}) +[2026-06-23 19:20:22,428][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.11930855363607407, 'train/grad_norm_pre_clip': 43.666011810302734, 'timing/data': 0.0005979840643703938, 'timing/model': 1.697597797960043, 'learning_rate/qwen_vl_interface': 7.23932514742973e-06, 'learning_rate/action_model': 7.23932514742973e-05, 'epoch': 0.73}) +[2026-06-23 19:20:25,768][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.15124601125717163, 'train/grad_norm_pre_clip': 55.73512268066406, 'timing/data': 0.0006793569773435593, 'timing/model': 1.6972921930719167, 'learning_rate/qwen_vl_interface': 7.236559264448413e-06, 'learning_rate/action_model': 7.236559264448412e-05, 'epoch': 0.73}) +[2026-06-23 19:20:29,106][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.1990014910697937, 'train/grad_norm_pre_clip': 106.94346618652344, 'timing/data': 0.0007128498982638121, 'timing/model': 1.696379688102752, 'learning_rate/qwen_vl_interface': 7.233792564867623e-06, 'learning_rate/action_model': 7.233792564867623e-05, 'epoch': 0.73}) +[2026-06-23 19:20:32,444][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.17757384479045868, 'train/grad_norm_pre_clip': 28.551515579223633, 'timing/data': 0.0006931929383426905, 'timing/model': 1.6923942458815873, 'learning_rate/qwen_vl_interface': 7.231025049824644e-06, 'learning_rate/action_model': 7.231025049824643e-05, 'epoch': 0.73}) +[2026-06-23 19:20:35,794][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.18204325437545776, 'train/grad_norm_pre_clip': 57.85771560668945, 'timing/data': 0.0006460449658334255, 'timing/model': 1.6995197210926563, 'learning_rate/qwen_vl_interface': 7.228256720457097e-06, 'learning_rate/action_model': 7.228256720457096e-05, 'epoch': 0.73}) +[2026-06-23 19:20:39,139][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.31378787755966187, 'train/grad_norm_pre_clip': 93.50970458984375, 'timing/data': 0.0006116090808063745, 'timing/model': 1.6992338828276843, 'learning_rate/qwen_vl_interface': 7.225487577902937e-06, 'learning_rate/action_model': 7.225487577902936e-05, 'epoch': 0.73}) +[2026-06-23 19:20:42,481][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.2047290802001953, 'train/grad_norm_pre_clip': 96.8353042602539, 'timing/data': 0.0006346330046653748, 'timing/model': 1.6980146521236748, 'learning_rate/qwen_vl_interface': 7.222717623300455e-06, 'learning_rate/action_model': 7.222717623300455e-05, 'epoch': 0.73}) +[2026-06-23 19:20:45,815][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.20305587351322174, 'train/grad_norm_pre_clip': 59.51459884643555, 'timing/data': 0.0006024730391800404, 'timing/model': 1.6975323560182005, 'learning_rate/qwen_vl_interface': 7.219946857788273e-06, 'learning_rate/action_model': 7.219946857788272e-05, 'epoch': 0.73}) +[2026-06-23 19:20:49,162][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.1818591058254242, 'train/grad_norm_pre_clip': 27.506101608276367, 'timing/data': 0.0006972001865506172, 'timing/model': 1.7006212719716132, 'learning_rate/qwen_vl_interface': 7.21717528250535e-06, 'learning_rate/action_model': 7.21717528250535e-05, 'epoch': 0.73}) +[2026-06-23 19:20:52,505][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.13583000004291534, 'train/grad_norm_pre_clip': 11.737451553344727, 'timing/data': 0.0006159879267215729, 'timing/model': 1.7013981528580189, 'learning_rate/qwen_vl_interface': 7.214402898590974e-06, 'learning_rate/action_model': 7.214402898590973e-05, 'epoch': 0.73}) +[2026-06-23 19:20:55,850][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.18808859586715698, 'train/grad_norm_pre_clip': 95.35444641113281, 'timing/data': 0.0006046060007065535, 'timing/model': 1.6983131740707904, 'learning_rate/qwen_vl_interface': 7.211629707184768e-06, 'learning_rate/action_model': 7.211629707184768e-05, 'epoch': 0.73}) +[2026-06-23 19:20:59,289][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.17189978063106537, 'train/grad_norm_pre_clip': 96.79084014892578, 'timing/data': 0.0006252860184758902, 'timing/model': 1.7933555110357702, 'learning_rate/qwen_vl_interface': 7.2088557094266885e-06, 'learning_rate/action_model': 7.208855709426688e-05, 'epoch': 0.73}) +[2026-06-23 19:21:02,628][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.17303138971328735, 'train/grad_norm_pre_clip': 91.73477172851562, 'timing/data': 0.0006133930291980505, 'timing/model': 1.6971851538401097, 'learning_rate/qwen_vl_interface': 7.2060809064570195e-06, 'learning_rate/action_model': 7.20608090645702e-05, 'epoch': 0.73}) +[2026-06-23 19:21:05,979][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.13133442401885986, 'train/grad_norm_pre_clip': 51.89181900024414, 'timing/data': 0.0006281200330704451, 'timing/model': 1.7007026760838926, 'learning_rate/qwen_vl_interface': 7.203305299416375e-06, 'learning_rate/action_model': 7.203305299416375e-05, 'epoch': 0.73}) +[2026-06-23 19:21:09,329][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.13865643739700317, 'train/grad_norm_pre_clip': 46.2678108215332, 'timing/data': 0.0006662518717348576, 'timing/model': 1.7009293218143284, 'learning_rate/qwen_vl_interface': 7.200528889445707e-06, 'learning_rate/action_model': 7.200528889445706e-05, 'epoch': 0.73}) +[2026-06-23 19:21:12,678][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.09775159507989883, 'train/grad_norm_pre_clip': 17.618515014648438, 'timing/data': 0.0006212880834937096, 'timing/model': 1.700048286933452, 'learning_rate/qwen_vl_interface': 7.19775167768629e-06, 'learning_rate/action_model': 7.19775167768629e-05, 'epoch': 0.73}) +[2026-06-23 19:21:16,026][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.1833590269088745, 'train/grad_norm_pre_clip': 61.806053161621094, 'timing/data': 0.0006875519175082445, 'timing/model': 1.6999072921462357, 'learning_rate/qwen_vl_interface': 7.194973665279731e-06, 'learning_rate/action_model': 7.19497366527973e-05, 'epoch': 0.73}) +[2026-06-23 19:21:19,370][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.1450916826725006, 'train/grad_norm_pre_clip': 57.61328125, 'timing/data': 0.0007080109789967537, 'timing/model': 1.6934851298574358, 'learning_rate/qwen_vl_interface': 7.1921948533679666e-06, 'learning_rate/action_model': 7.192194853367967e-05, 'epoch': 0.73}) +[2026-06-23 19:21:22,721][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.20085641741752625, 'train/grad_norm_pre_clip': 104.09261322021484, 'timing/data': 0.0007620919495820999, 'timing/model': 1.7020075500477105, 'learning_rate/qwen_vl_interface': 7.1894152430932605e-06, 'learning_rate/action_model': 7.189415243093261e-05, 'epoch': 0.73}) +[2026-06-23 19:21:26,066][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.14918234944343567, 'train/grad_norm_pre_clip': 64.1115493774414, 'timing/data': 0.0006302341353148222, 'timing/model': 1.696058007190004, 'learning_rate/qwen_vl_interface': 7.186634835598205e-06, 'learning_rate/action_model': 7.186634835598205e-05, 'epoch': 0.73}) +[2026-06-23 19:21:29,418][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.11483659595251083, 'train/grad_norm_pre_clip': 24.707014083862305, 'timing/data': 0.0005966310855001211, 'timing/model': 1.7022837700787932, 'learning_rate/qwen_vl_interface': 7.183853632025724e-06, 'learning_rate/action_model': 7.183853632025723e-05, 'epoch': 0.73}) +[2026-06-23 19:21:32,771][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.13411307334899902, 'train/grad_norm_pre_clip': 22.233076095581055, 'timing/data': 0.0006105981301516294, 'timing/model': 1.7012378650251776, 'learning_rate/qwen_vl_interface': 7.181071633519062e-06, 'learning_rate/action_model': 7.181071633519062e-05, 'epoch': 0.73}) +[2026-06-23 19:21:36,119][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.16932222247123718, 'train/grad_norm_pre_clip': 29.577646255493164, 'timing/data': 0.0006223199889063835, 'timing/model': 1.6957261541392654, 'learning_rate/qwen_vl_interface': 7.178288841221795e-06, 'learning_rate/action_model': 7.178288841221794e-05, 'epoch': 0.73}) +[2026-06-23 19:21:39,472][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.2062782496213913, 'train/grad_norm_pre_clip': 48.94100570678711, 'timing/data': 0.0007281478028744459, 'timing/model': 1.7061845830176026, 'learning_rate/qwen_vl_interface': 7.1755052562778215e-06, 'learning_rate/action_model': 7.175505256277822e-05, 'epoch': 0.73}) +[2026-06-23 19:21:42,823][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1900 +[2026-06-23 19:23:17,092][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1900 +[2026-06-23 19:23:17,092][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 1900 +[2026-06-23 19:24:28,330][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 1900 +[2026-06-23 19:24:28,331][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.22359970211982727, 'train/grad_norm_pre_clip': 65.87474060058594, 'eval/action_loss/seconds': 94.26680863602087, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.08520537614822388, 'eval/latency_2/loss': 0.08520537614822388, 'eval/flappy/loss': 0.08520537614822388, 'eval/flappy/latency_2/loss': 0.08520537614822388, 'eval/flap_precision': 0.9330007127583749, 'eval/flap_recall': 0.8832658569500674, 'eval/flap_f1': 0.9074523396880416, 'eval/flap_rate_model': 0.19486111111111112, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9074523396880417, 'eval/latency_2/flap_precision': 0.9330007127583749, 'eval/latency_2/flap_recall': 0.8832658569500674, 'eval/latency_2/flap_f1': 0.9074523396880416, 'eval/latency_2/flap_rate_model': 0.19486111111111112, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9074523396880417, 'eval/action_classification/data_wait_seconds': 0.9200523495674133, 'eval/action_classification/predict_seconds': 70.19178009033203, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006259270012378693, 'timing/model': 1.7023039089981467, 'learning_rate/qwen_vl_interface': 7.1727208798313704e-06, 'learning_rate/action_model': 7.172720879831371e-05, 'epoch': 0.74}) +[2026-06-23 19:24:31,662][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.18192565441131592, 'train/grad_norm_pre_clip': 12.515934944152832, 'timing/data': 0.0005962199065834284, 'timing/model': 1.688409297959879, 'learning_rate/qwen_vl_interface': 7.169935713026994e-06, 'learning_rate/action_model': 7.169935713026994e-05, 'epoch': 0.74}) +[2026-06-23 19:24:34,994][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.06420209258794785, 'train/grad_norm_pre_clip': 12.456694602966309, 'timing/data': 0.0007196131628006697, 'timing/model': 1.688007143093273, 'learning_rate/qwen_vl_interface': 7.1671497570095684e-06, 'learning_rate/action_model': 7.167149757009568e-05, 'epoch': 0.74}) +[2026-06-23 19:24:38,321][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.20479536056518555, 'train/grad_norm_pre_clip': 16.74859046936035, 'timing/data': 0.0006034141406416893, 'timing/model': 1.6874305189121515, 'learning_rate/qwen_vl_interface': 7.164363012924295e-06, 'learning_rate/action_model': 7.164363012924294e-05, 'epoch': 0.74}) +[2026-06-23 19:24:41,660][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.14186589419841766, 'train/grad_norm_pre_clip': 12.238262176513672, 'timing/data': 0.0006238119676709175, 'timing/model': 1.6962594860233366, 'learning_rate/qwen_vl_interface': 7.161575481916699e-06, 'learning_rate/action_model': 7.1615754819167e-05, 'epoch': 0.74}) +[2026-06-23 19:24:44,988][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.11022620648145676, 'train/grad_norm_pre_clip': 22.230104446411133, 'timing/data': 0.0006069710943847895, 'timing/model': 1.689452189952135, 'learning_rate/qwen_vl_interface': 7.158787165132627e-06, 'learning_rate/action_model': 7.158787165132626e-05, 'epoch': 0.74}) +[2026-06-23 19:24:48,323][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.19471019506454468, 'train/grad_norm_pre_clip': 24.258272171020508, 'timing/data': 0.0005940571427345276, 'timing/model': 1.6923292479477823, 'learning_rate/qwen_vl_interface': 7.1559980637182535e-06, 'learning_rate/action_model': 7.155998063718254e-05, 'epoch': 0.74}) +[2026-06-23 19:24:51,666][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.10269252955913544, 'train/grad_norm_pre_clip': 22.820846557617188, 'timing/data': 0.000603795051574707, 'timing/model': 1.6981139311101288, 'learning_rate/qwen_vl_interface': 7.153208178820071e-06, 'learning_rate/action_model': 7.153208178820072e-05, 'epoch': 0.74}) +[2026-06-23 19:24:55,016][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.1395532190799713, 'train/grad_norm_pre_clip': 38.914058685302734, 'timing/data': 0.0006100370083004236, 'timing/model': 1.7021164859179407, 'learning_rate/qwen_vl_interface': 7.150417511584898e-06, 'learning_rate/action_model': 7.150417511584898e-05, 'epoch': 0.74}) +[2026-06-23 19:24:58,359][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.1549329310655594, 'train/grad_norm_pre_clip': 75.08467864990234, 'timing/data': 0.0006135830190032721, 'timing/model': 1.697014005156234, 'learning_rate/qwen_vl_interface': 7.147626063159872e-06, 'learning_rate/action_model': 7.147626063159872e-05, 'epoch': 0.74}) +[2026-06-23 19:25:01,845][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.09181293100118637, 'train/grad_norm_pre_clip': 49.8328742980957, 'timing/data': 0.0006168500985950232, 'timing/model': 1.6987851338926703, 'learning_rate/qwen_vl_interface': 7.14483383469245e-06, 'learning_rate/action_model': 7.144833834692449e-05, 'epoch': 0.74}) +[2026-06-23 19:25:05,187][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.13980260491371155, 'train/grad_norm_pre_clip': 19.187114715576172, 'timing/data': 0.0006047068163752556, 'timing/model': 1.6944683098699898, 'learning_rate/qwen_vl_interface': 7.142040827330411e-06, 'learning_rate/action_model': 7.142040827330411e-05, 'epoch': 0.74}) +[2026-06-23 19:25:08,529][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.07070575654506683, 'train/grad_norm_pre_clip': 49.43608093261719, 'timing/data': 0.0006212780717760324, 'timing/model': 1.6999626769684255, 'learning_rate/qwen_vl_interface': 7.139247042221859e-06, 'learning_rate/action_model': 7.139247042221858e-05, 'epoch': 0.74}) +[2026-06-23 19:25:11,878][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.1750687211751938, 'train/grad_norm_pre_clip': 40.720947265625, 'timing/data': 0.0006052681710571051, 'timing/model': 1.6948320230003446, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.74}) +[2026-06-23 19:25:15,221][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.06578711420297623, 'train/grad_norm_pre_clip': 28.471879959106445, 'timing/data': 0.0006058691069483757, 'timing/model': 1.6962148330640048, 'learning_rate/qwen_vl_interface': 7.133657143359203e-06, 'learning_rate/action_model': 7.133657143359202e-05, 'epoch': 0.74}) +[2026-06-23 19:25:18,561][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.14613831043243408, 'train/grad_norm_pre_clip': 37.64408874511719, 'timing/data': 0.0006233418826013803, 'timing/model': 1.695808979915455, 'learning_rate/qwen_vl_interface': 7.130861031902897e-06, 'learning_rate/action_model': 7.130861031902897e-05, 'epoch': 0.74}) +[2026-06-23 19:25:21,907][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.24548831582069397, 'train/grad_norm_pre_clip': 35.46684265136719, 'timing/data': 0.0006182719953358173, 'timing/model': 1.699604423949495, 'learning_rate/qwen_vl_interface': 7.128064147295667e-06, 'learning_rate/action_model': 7.128064147295667e-05, 'epoch': 0.74}) +[2026-06-23 19:25:25,247][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.1940479874610901, 'train/grad_norm_pre_clip': 71.37052917480469, 'timing/data': 0.0006139741744846106, 'timing/model': 1.6949284349102527, 'learning_rate/qwen_vl_interface': 7.125266490687206e-06, 'learning_rate/action_model': 7.125266490687206e-05, 'epoch': 0.74}) +[2026-06-23 19:25:28,605][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.11615398526191711, 'train/grad_norm_pre_clip': 33.30348587036133, 'timing/data': 0.0006161381024867296, 'timing/model': 1.7113360778894275, 'learning_rate/qwen_vl_interface': 7.1224680632275265e-06, 'learning_rate/action_model': 7.122468063227527e-05, 'epoch': 0.74}) +[2026-06-23 19:25:31,951][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.1377108097076416, 'train/grad_norm_pre_clip': 14.198895454406738, 'timing/data': 0.0006171402055770159, 'timing/model': 1.6986171989701688, 'learning_rate/qwen_vl_interface': 7.1196688660669555e-06, 'learning_rate/action_model': 7.119668866066955e-05, 'epoch': 0.74}) +[2026-06-23 19:25:35,302][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.2204091101884842, 'train/grad_norm_pre_clip': 40.92136001586914, 'timing/data': 0.0006170300766825676, 'timing/model': 1.7046347020659596, 'learning_rate/qwen_vl_interface': 7.116868900356138e-06, 'learning_rate/action_model': 7.116868900356138e-05, 'epoch': 0.74}) +[2026-06-23 19:25:38,654][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.09337327629327774, 'train/grad_norm_pre_clip': 42.40986633300781, 'timing/data': 0.0006115699652582407, 'timing/model': 1.7033557570539415, 'learning_rate/qwen_vl_interface': 7.114068167246033e-06, 'learning_rate/action_model': 7.114068167246033e-05, 'epoch': 0.74}) +[2026-06-23 19:25:42,000][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.13488389551639557, 'train/grad_norm_pre_clip': 49.99468231201172, 'timing/data': 0.0006846368778496981, 'timing/model': 1.6992836631834507, 'learning_rate/qwen_vl_interface': 7.111266667887916e-06, 'learning_rate/action_model': 7.111266667887915e-05, 'epoch': 0.74}) +[2026-06-23 19:25:45,351][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.15091781318187714, 'train/grad_norm_pre_clip': 14.800217628479004, 'timing/data': 0.0008591949008405209, 'timing/model': 1.703424836974591, 'learning_rate/qwen_vl_interface': 7.108464403433378e-06, 'learning_rate/action_model': 7.108464403433378e-05, 'epoch': 0.74}) +[2026-06-23 19:25:48,806][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.16404291987419128, 'train/grad_norm_pre_clip': 34.941200256347656, 'timing/data': 0.0006231619045138359, 'timing/model': 1.8142500731628388, 'learning_rate/qwen_vl_interface': 7.105661375034324e-06, 'learning_rate/action_model': 7.105661375034324e-05, 'epoch': 0.74}) +[2026-06-23 19:25:52,143][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.17726346850395203, 'train/grad_norm_pre_clip': 34.62255096435547, 'timing/data': 0.0005996571853756905, 'timing/model': 1.6956657019909471, 'learning_rate/qwen_vl_interface': 7.102857583842973e-06, 'learning_rate/action_model': 7.102857583842974e-05, 'epoch': 0.75}) +[2026-06-23 19:25:55,482][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.1718330830335617, 'train/grad_norm_pre_clip': 19.19925880432129, 'timing/data': 0.000609495909884572, 'timing/model': 1.696316255023703, 'learning_rate/qwen_vl_interface': 7.100053031011859e-06, 'learning_rate/action_model': 7.100053031011859e-05, 'epoch': 0.75}) +[2026-06-23 19:25:58,822][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.22135759890079498, 'train/grad_norm_pre_clip': 38.24595260620117, 'timing/data': 0.0006111480761319399, 'timing/model': 1.6961789370980114, 'learning_rate/qwen_vl_interface': 7.097247717693827e-06, 'learning_rate/action_model': 7.097247717693827e-05, 'epoch': 0.75}) +[2026-06-23 19:26:02,171][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.20127689838409424, 'train/grad_norm_pre_clip': 101.91551208496094, 'timing/data': 0.0006844662129878998, 'timing/model': 1.7001304849982262, 'learning_rate/qwen_vl_interface': 7.094441645042033e-06, 'learning_rate/action_model': 7.094441645042032e-05, 'epoch': 0.75}) +[2026-06-23 19:26:05,515][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.21241921186447144, 'train/grad_norm_pre_clip': 59.02775192260742, 'timing/data': 0.0006043459288775921, 'timing/model': 1.6969774579629302, 'learning_rate/qwen_vl_interface': 7.09163481420995e-06, 'learning_rate/action_model': 7.09163481420995e-05, 'epoch': 0.75}) +[2026-06-23 19:26:08,860][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.26232099533081055, 'train/grad_norm_pre_clip': 17.077285766601562, 'timing/data': 0.0006142039783298969, 'timing/model': 1.6979677800554782, 'learning_rate/qwen_vl_interface': 7.0888272263513595e-06, 'learning_rate/action_model': 7.088827226351359e-05, 'epoch': 0.75}) +[2026-06-23 19:26:12,201][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.16783949732780457, 'train/grad_norm_pre_clip': 40.80256271362305, 'timing/data': 0.0006049580406397581, 'timing/model': 1.6948850450571626, 'learning_rate/qwen_vl_interface': 7.086018882620354e-06, 'learning_rate/action_model': 7.086018882620354e-05, 'epoch': 0.75}) +[2026-06-23 19:26:15,540][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.15926963090896606, 'train/grad_norm_pre_clip': 47.38175964355469, 'timing/data': 0.0006923309992998838, 'timing/model': 1.6946084250230342, 'learning_rate/qwen_vl_interface': 7.083209784171341e-06, 'learning_rate/action_model': 7.083209784171341e-05, 'epoch': 0.75}) +[2026-06-23 19:26:18,878][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.13480766117572784, 'train/grad_norm_pre_clip': 22.744028091430664, 'timing/data': 0.0006184920202940702, 'timing/model': 1.6946344140451401, 'learning_rate/qwen_vl_interface': 7.080399932159029e-06, 'learning_rate/action_model': 7.080399932159029e-05, 'epoch': 0.75}) +[2026-06-23 19:26:22,220][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.16337521374225616, 'train/grad_norm_pre_clip': 64.34965515136719, 'timing/data': 0.0006028029602020979, 'timing/model': 1.6974003731738776, 'learning_rate/qwen_vl_interface': 7.077589327738448e-06, 'learning_rate/action_model': 7.077589327738447e-05, 'epoch': 0.75}) +[2026-06-23 19:26:25,561][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.13870960474014282, 'train/grad_norm_pre_clip': 28.589906692504883, 'timing/data': 0.0006174310110509396, 'timing/model': 1.6964587038382888, 'learning_rate/qwen_vl_interface': 7.0747779720649265e-06, 'learning_rate/action_model': 7.074777972064927e-05, 'epoch': 0.75}) +[2026-06-23 19:26:28,912][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.12604451179504395, 'train/grad_norm_pre_clip': 15.307595252990723, 'timing/data': 0.0006142549682408571, 'timing/model': 1.7017285299953073, 'learning_rate/qwen_vl_interface': 7.071965866294108e-06, 'learning_rate/action_model': 7.071965866294108e-05, 'epoch': 0.75}) +[2026-06-23 19:26:32,256][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.18382449448108673, 'train/grad_norm_pre_clip': 27.93405532836914, 'timing/data': 0.0006143050268292427, 'timing/model': 1.7027203340549022, 'learning_rate/qwen_vl_interface': 7.0691530115819454e-06, 'learning_rate/action_model': 7.069153011581945e-05, 'epoch': 0.75}) +[2026-06-23 19:26:35,602][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.14644397795200348, 'train/grad_norm_pre_clip': 89.2166976928711, 'timing/data': 0.000613663112744689, 'timing/model': 1.6981260480824858, 'learning_rate/qwen_vl_interface': 7.066339409084694e-06, 'learning_rate/action_model': 7.066339409084694e-05, 'epoch': 0.75}) +[2026-06-23 19:26:39,022][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.2417772114276886, 'train/grad_norm_pre_clip': 89.71903991699219, 'timing/data': 0.0006160780321806669, 'timing/model': 1.7779281379189342, 'learning_rate/qwen_vl_interface': 7.0635250599589215e-06, 'learning_rate/action_model': 7.06352505995892e-05, 'epoch': 0.75}) +[2026-06-23 19:26:42,363][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.17314013838768005, 'train/grad_norm_pre_clip': 76.86725616455078, 'timing/data': 0.0006157669704407454, 'timing/model': 1.6969004550483078, 'learning_rate/qwen_vl_interface': 7.060709965361499e-06, 'learning_rate/action_model': 7.060709965361499e-05, 'epoch': 0.75}) +[2026-06-23 19:26:45,700][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.16225722432136536, 'train/grad_norm_pre_clip': 46.08650207519531, 'timing/data': 0.0005991461221128702, 'timing/model': 1.6975909520406276, 'learning_rate/qwen_vl_interface': 7.057894126449606e-06, 'learning_rate/action_model': 7.057894126449606e-05, 'epoch': 0.75}) +[2026-06-23 19:26:49,043][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.18508586287498474, 'train/grad_norm_pre_clip': 16.979263305664062, 'timing/data': 0.0006013310048729181, 'timing/model': 1.6973497988656163, 'learning_rate/qwen_vl_interface': 7.055077544380728e-06, 'learning_rate/action_model': 7.055077544380728e-05, 'epoch': 0.75}) +[2026-06-23 19:26:52,376][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.13889682292938232, 'train/grad_norm_pre_clip': 23.61064910888672, 'timing/data': 0.0007042230572551489, 'timing/model': 1.6934264139272273, 'learning_rate/qwen_vl_interface': 7.0522602203126554e-06, 'learning_rate/action_model': 7.052260220312656e-05, 'epoch': 0.75}) +[2026-06-23 19:26:55,722][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.207800030708313, 'train/grad_norm_pre_clip': 95.80501556396484, 'timing/data': 0.0006109080277383327, 'timing/model': 1.6988117177970707, 'learning_rate/qwen_vl_interface': 7.0494421554034845e-06, 'learning_rate/action_model': 7.049442155403484e-05, 'epoch': 0.75}) +[2026-06-23 19:26:59,064][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.22657831013202667, 'train/grad_norm_pre_clip': 89.71292114257812, 'timing/data': 0.0007186601869761944, 'timing/model': 1.6969587749335915, 'learning_rate/qwen_vl_interface': 7.046623350811612e-06, 'learning_rate/action_model': 7.046623350811612e-05, 'epoch': 0.75}) +[2026-06-23 19:27:02,410][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.240553081035614, 'train/grad_norm_pre_clip': 78.10272979736328, 'timing/data': 0.0005990560166537762, 'timing/model': 1.6984607290942222, 'learning_rate/qwen_vl_interface': 7.043803807695747e-06, 'learning_rate/action_model': 7.043803807695747e-05, 'epoch': 0.75}) +[2026-06-23 19:27:05,754][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.18745771050453186, 'train/grad_norm_pre_clip': 81.14483642578125, 'timing/data': 0.000814409926533699, 'timing/model': 1.6962957391515374, 'learning_rate/qwen_vl_interface': 7.040983527214893e-06, 'learning_rate/action_model': 7.040983527214893e-05, 'epoch': 0.75}) +[2026-06-23 19:27:09,099][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.06782613694667816, 'train/grad_norm_pre_clip': 38.81398391723633, 'timing/data': 0.0006134530995041132, 'timing/model': 1.697102215141058, 'learning_rate/qwen_vl_interface': 7.038162510528363e-06, 'learning_rate/action_model': 7.038162510528362e-05, 'epoch': 0.75}) +[2026-06-23 19:27:12,437][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.08638456463813782, 'train/grad_norm_pre_clip': 32.96275329589844, 'timing/data': 0.0007031320128589869, 'timing/model': 1.694626672193408, 'learning_rate/qwen_vl_interface': 7.03534075879577e-06, 'learning_rate/action_model': 7.03534075879577e-05, 'epoch': 0.75}) +[2026-06-23 19:27:15,788][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.16169631481170654, 'train/grad_norm_pre_clip': 44.50674819946289, 'timing/data': 0.0008292777929455042, 'timing/model': 1.7014192598871887, 'learning_rate/qwen_vl_interface': 7.032518273177028e-06, 'learning_rate/action_model': 7.032518273177028e-05, 'epoch': 0.75}) +[2026-06-23 19:27:19,129][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.24916242063045502, 'train/grad_norm_pre_clip': 48.08043670654297, 'timing/data': 0.0006891849916428328, 'timing/model': 1.6937095171306282, 'learning_rate/qwen_vl_interface': 7.029695054832357e-06, 'learning_rate/action_model': 7.029695054832357e-05, 'epoch': 0.76}) +[2026-06-23 19:27:22,475][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.2024346888065338, 'train/grad_norm_pre_clip': 74.95714569091797, 'timing/data': 0.0006143050268292427, 'timing/model': 1.7011074449401349, 'learning_rate/qwen_vl_interface': 7.026871104922273e-06, 'learning_rate/action_model': 7.026871104922273e-05, 'epoch': 0.76}) +[2026-06-23 19:27:25,825][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.16330334544181824, 'train/grad_norm_pre_clip': 18.578054428100586, 'timing/data': 0.0009338550735265017, 'timing/model': 1.6995312611106783, 'learning_rate/qwen_vl_interface': 7.024046424607596e-06, 'learning_rate/action_model': 7.024046424607595e-05, 'epoch': 0.76}) +[2026-06-23 19:27:29,267][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.23040716350078583, 'train/grad_norm_pre_clip': 54.51994323730469, 'timing/data': 0.0007032309658825397, 'timing/model': 1.7029373238328844, 'learning_rate/qwen_vl_interface': 7.021221015049445e-06, 'learning_rate/action_model': 7.021221015049446e-05, 'epoch': 0.76}) +[2026-06-23 19:27:32,615][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.1392592340707779, 'train/grad_norm_pre_clip': 56.549259185791016, 'timing/data': 0.0006905680056661367, 'timing/model': 1.6983157168142498, 'learning_rate/qwen_vl_interface': 7.01839487740924e-06, 'learning_rate/action_model': 7.01839487740924e-05, 'epoch': 0.76}) +[2026-06-23 19:27:35,959][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.1499153971672058, 'train/grad_norm_pre_clip': 60.561866760253906, 'timing/data': 0.0006178808398544788, 'timing/model': 1.702177566010505, 'learning_rate/qwen_vl_interface': 7.015568012848699e-06, 'learning_rate/action_model': 7.015568012848699e-05, 'epoch': 0.76}) +[2026-06-23 19:27:39,313][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.1889980584383011, 'train/grad_norm_pre_clip': 13.701214790344238, 'timing/data': 0.0006057689897716045, 'timing/model': 1.7044418528676033, 'learning_rate/qwen_vl_interface': 7.01274042252984e-06, 'learning_rate/action_model': 7.01274042252984e-05, 'epoch': 0.76}) +[2026-06-23 19:27:42,660][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.21359795331954956, 'train/grad_norm_pre_clip': 8.86925220489502, 'timing/data': 0.0006078931037336588, 'timing/model': 1.6998404529877007, 'learning_rate/qwen_vl_interface': 7.009912107614978e-06, 'learning_rate/action_model': 7.009912107614978e-05, 'epoch': 0.76}) +[2026-06-23 19:27:46,011][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.10342215746641159, 'train/grad_norm_pre_clip': 12.00719928741455, 'timing/data': 0.0005975628737360239, 'timing/model': 1.7010690129827708, 'learning_rate/qwen_vl_interface': 7.0070830692667244e-06, 'learning_rate/action_model': 7.007083069266724e-05, 'epoch': 0.76}) +[2026-06-23 19:27:49,370][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.11329653114080429, 'train/grad_norm_pre_clip': 31.407941818237305, 'timing/data': 0.0006176100578159094, 'timing/model': 1.7073371950536966, 'learning_rate/qwen_vl_interface': 7.004253308647993e-06, 'learning_rate/action_model': 7.004253308647994e-05, 'epoch': 0.76}) +[2026-06-23 19:27:52,724][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.0865917056798935, 'train/grad_norm_pre_clip': 24.77004623413086, 'timing/data': 0.0006644891109317541, 'timing/model': 1.7049041620921344, 'learning_rate/qwen_vl_interface': 7.001422826921991e-06, 'learning_rate/action_model': 7.00142282692199e-05, 'epoch': 0.76}) +[2026-06-23 19:27:56,076][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.18992669880390167, 'train/grad_norm_pre_clip': 40.55241394042969, 'timing/data': 0.0006521660834550858, 'timing/model': 1.6999031610321254, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.76}) +[2026-06-23 19:27:59,428][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.14541107416152954, 'train/grad_norm_pre_clip': 27.254058837890625, 'timing/data': 0.0006665028631687164, 'timing/model': 1.7018639880698174, 'learning_rate/qwen_vl_interface': 6.995759704802485e-06, 'learning_rate/action_model': 6.995759704802485e-05, 'epoch': 0.76}) +[2026-06-23 19:28:02,790][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.05826391279697418, 'train/grad_norm_pre_clip': 10.078161239624023, 'timing/data': 0.0010264490265399218, 'timing/model': 1.7046293849125504, 'learning_rate/qwen_vl_interface': 6.992927066736876e-06, 'learning_rate/action_model': 6.992927066736875e-05, 'epoch': 0.76}) +[2026-06-23 19:28:06,138][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.1882132887840271, 'train/grad_norm_pre_clip': 53.188636779785156, 'timing/data': 0.0007027711253613234, 'timing/model': 1.6959364148788154, 'learning_rate/qwen_vl_interface': 6.990093712219787e-06, 'learning_rate/action_model': 6.990093712219787e-05, 'epoch': 0.76}) +[2026-06-23 19:28:09,486][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.11308867484331131, 'train/grad_norm_pre_clip': 26.81495475769043, 'timing/data': 0.0006649999413639307, 'timing/model': 1.7033154862001538, 'learning_rate/qwen_vl_interface': 6.9872596424159e-06, 'learning_rate/action_model': 6.9872596424159e-05, 'epoch': 0.76}) +[2026-06-23 19:28:12,840][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.1309361308813095, 'train/grad_norm_pre_clip': 46.643863677978516, 'timing/data': 0.0006649799179285765, 'timing/model': 1.705202953889966, 'learning_rate/qwen_vl_interface': 6.9844248584902e-06, 'learning_rate/action_model': 6.984424858490199e-05, 'epoch': 0.76}) +[2026-06-23 19:28:16,195][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.13009941577911377, 'train/grad_norm_pre_clip': 12.461606979370117, 'timing/data': 0.0006190428975969553, 'timing/model': 1.7040269228164107, 'learning_rate/qwen_vl_interface': 6.981589361607952e-06, 'learning_rate/action_model': 6.981589361607953e-05, 'epoch': 0.76}) +[2026-06-23 19:28:19,647][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.28574904799461365, 'train/grad_norm_pre_clip': 82.35903930664062, 'timing/data': 0.0008293089922517538, 'timing/model': 1.701793004060164, 'learning_rate/qwen_vl_interface': 6.9787531529347274e-06, 'learning_rate/action_model': 6.978753152934727e-05, 'epoch': 0.76}) +[2026-06-23 19:28:22,995][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.2882411777973175, 'train/grad_norm_pre_clip': 115.49659729003906, 'timing/data': 0.0006601421628147364, 'timing/model': 1.7001948000397533, 'learning_rate/qwen_vl_interface': 6.97591623363638e-06, 'learning_rate/action_model': 6.97591623363638e-05, 'epoch': 0.76}) +[2026-06-23 19:28:26,337][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.3393135964870453, 'train/grad_norm_pre_clip': 139.25286865234375, 'timing/data': 0.0006060488522052765, 'timing/model': 1.697442688047886, 'learning_rate/qwen_vl_interface': 6.9730786048790644e-06, 'learning_rate/action_model': 6.973078604879064e-05, 'epoch': 0.76}) +[2026-06-23 19:28:29,680][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.2458626925945282, 'train/grad_norm_pre_clip': 61.54499053955078, 'timing/data': 0.0007209950126707554, 'timing/model': 1.7015208429656923, 'learning_rate/qwen_vl_interface': 6.970240267829221e-06, 'learning_rate/action_model': 6.970240267829222e-05, 'epoch': 0.76}) +[2026-06-23 19:28:33,023][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.0721999853849411, 'train/grad_norm_pre_clip': 24.262052536010742, 'timing/data': 0.0009026071056723595, 'timing/model': 1.6986855240538716, 'learning_rate/qwen_vl_interface': 6.967401223653582e-06, 'learning_rate/action_model': 6.967401223653581e-05, 'epoch': 0.76}) +[2026-06-23 19:28:36,372][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.10942280292510986, 'train/grad_norm_pre_clip': 21.623334884643555, 'timing/data': 0.0007658100221306086, 'timing/model': 1.7017359279561788, 'learning_rate/qwen_vl_interface': 6.9645614735191724e-06, 'learning_rate/action_model': 6.964561473519172e-05, 'epoch': 0.76}) +[2026-06-23 19:28:39,720][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.20137116312980652, 'train/grad_norm_pre_clip': 20.906675338745117, 'timing/data': 0.0006608220282942057, 'timing/model': 1.697664835024625, 'learning_rate/qwen_vl_interface': 6.961721018593305e-06, 'learning_rate/action_model': 6.961721018593305e-05, 'epoch': 0.76}) +[2026-06-23 19:28:43,064][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.17072860896587372, 'train/grad_norm_pre_clip': 34.05712127685547, 'timing/data': 0.000611680094152689, 'timing/model': 1.6978239840827882, 'learning_rate/qwen_vl_interface': 6.958879860043585e-06, 'learning_rate/action_model': 6.958879860043585e-05, 'epoch': 0.76}) +[2026-06-23 19:28:46,411][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.04911592975258827, 'train/grad_norm_pre_clip': 20.510658264160156, 'timing/data': 0.0005942969582974911, 'timing/model': 1.6994909581262618, 'learning_rate/qwen_vl_interface': 6.9560379990379025e-06, 'learning_rate/action_model': 6.956037999037902e-05, 'epoch': 0.77}) +[2026-06-23 19:28:49,762][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.18614599108695984, 'train/grad_norm_pre_clip': 58.1086311340332, 'timing/data': 0.0007617820519953966, 'timing/model': 1.7027855520136654, 'learning_rate/qwen_vl_interface': 6.953195436744442e-06, 'learning_rate/action_model': 6.953195436744442e-05, 'epoch': 0.77}) +[2026-06-23 19:28:53,111][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.22850239276885986, 'train/grad_norm_pre_clip': 16.923250198364258, 'timing/data': 0.0006812300998717546, 'timing/model': 1.7013867311179638, 'learning_rate/qwen_vl_interface': 6.950352174331676e-06, 'learning_rate/action_model': 6.950352174331675e-05, 'epoch': 0.77}) +[2026-06-23 19:28:56,457][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.13794615864753723, 'train/grad_norm_pre_clip': 13.028966903686523, 'timing/data': 0.000609725946560502, 'timing/model': 1.6963599319569767, 'learning_rate/qwen_vl_interface': 6.947508212968356e-06, 'learning_rate/action_model': 6.947508212968356e-05, 'epoch': 0.77}) +[2026-06-23 19:28:59,803][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.17695783078670502, 'train/grad_norm_pre_clip': 83.66429901123047, 'timing/data': 0.0007077909540385008, 'timing/model': 1.6980569430161268, 'learning_rate/qwen_vl_interface': 6.9446635538235295e-06, 'learning_rate/action_model': 6.944663553823529e-05, 'epoch': 0.77}) +[2026-06-23 19:29:03,152][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.1782141476869583, 'train/grad_norm_pre_clip': 53.955230712890625, 'timing/data': 0.0010931550059467554, 'timing/model': 1.7023483400698751, 'learning_rate/qwen_vl_interface': 6.94181819806653e-06, 'learning_rate/action_model': 6.94181819806653e-05, 'epoch': 0.77}) +[2026-06-23 19:29:06,580][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.06032194942235947, 'train/grad_norm_pre_clip': 43.605281829833984, 'timing/data': 0.0006605219095945358, 'timing/model': 1.7796823501121253, 'learning_rate/qwen_vl_interface': 6.938972146866975e-06, 'learning_rate/action_model': 6.938972146866974e-05, 'epoch': 0.77}) +[2026-06-23 19:29:09,933][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.1969386339187622, 'train/grad_norm_pre_clip': 61.85057067871094, 'timing/data': 0.0006928720977157354, 'timing/model': 1.7018536189571023, 'learning_rate/qwen_vl_interface': 6.936125401394769e-06, 'learning_rate/action_model': 6.936125401394768e-05, 'epoch': 0.77}) +[2026-06-23 19:29:13,280][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.12415560334920883, 'train/grad_norm_pre_clip': 22.593427658081055, 'timing/data': 0.000720473937690258, 'timing/model': 1.701407641870901, 'learning_rate/qwen_vl_interface': 6.933277962820101e-06, 'learning_rate/action_model': 6.9332779628201e-05, 'epoch': 0.77}) +[2026-06-23 19:29:16,621][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.05409551411867142, 'train/grad_norm_pre_clip': 38.591041564941406, 'timing/data': 0.0006099171005189419, 'timing/model': 1.6960344419348985, 'learning_rate/qwen_vl_interface': 6.930429832313444e-06, 'learning_rate/action_model': 6.930429832313444e-05, 'epoch': 0.77}) +[2026-06-23 19:29:19,966][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.19657643139362335, 'train/grad_norm_pre_clip': 79.95184326171875, 'timing/data': 0.0006995250005275011, 'timing/model': 1.6977132470346987, 'learning_rate/qwen_vl_interface': 6.92758101104556e-06, 'learning_rate/action_model': 6.92758101104556e-05, 'epoch': 0.77}) +[2026-06-23 19:29:23,308][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.09694399684667587, 'train/grad_norm_pre_clip': 102.6312484741211, 'timing/data': 0.0006752798799425364, 'timing/model': 1.6958367901388556, 'learning_rate/qwen_vl_interface': 6.9247315001874885e-06, 'learning_rate/action_model': 6.924731500187488e-05, 'epoch': 0.77}) +[2026-06-23 19:29:26,645][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.28245609998703003, 'train/grad_norm_pre_clip': 145.03134155273438, 'timing/data': 0.0006094560958445072, 'timing/model': 1.6965158169623464, 'learning_rate/qwen_vl_interface': 6.9218813009105605e-06, 'learning_rate/action_model': 6.92188130091056e-05, 'epoch': 0.77}) +[2026-06-23 19:29:29,991][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.16104930639266968, 'train/grad_norm_pre_clip': 24.193347930908203, 'timing/data': 0.0006073520053178072, 'timing/model': 1.7002433729358017, 'learning_rate/qwen_vl_interface': 6.919030414386378e-06, 'learning_rate/action_model': 6.919030414386379e-05, 'epoch': 0.77}) +[2026-06-23 19:29:33,334][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.0885004773736, 'train/grad_norm_pre_clip': 41.8172607421875, 'timing/data': 0.0006648688577115536, 'timing/model': 1.6962114139460027, 'learning_rate/qwen_vl_interface': 6.916178841786839e-06, 'learning_rate/action_model': 6.916178841786838e-05, 'epoch': 0.77}) +[2026-06-23 19:29:36,682][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.103719562292099, 'train/grad_norm_pre_clip': 30.280284881591797, 'timing/data': 0.0006586981471627951, 'timing/model': 1.7018131949007511, 'learning_rate/qwen_vl_interface': 6.913326584284115e-06, 'learning_rate/action_model': 6.913326584284115e-05, 'epoch': 0.77}) +[2026-06-23 19:29:40,021][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.11513764411211014, 'train/grad_norm_pre_clip': 48.3011360168457, 'timing/data': 0.0006793967913836241, 'timing/model': 1.6944120319094509, 'learning_rate/qwen_vl_interface': 6.910473643050662e-06, 'learning_rate/action_model': 6.910473643050662e-05, 'epoch': 0.77}) +[2026-06-23 19:29:43,358][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.1342145949602127, 'train/grad_norm_pre_clip': 39.85450744628906, 'timing/data': 0.0006061689928174019, 'timing/model': 1.6992577109485865, 'learning_rate/qwen_vl_interface': 6.907620019259215e-06, 'learning_rate/action_model': 6.907620019259216e-05, 'epoch': 0.77}) +[2026-06-23 19:29:46,694][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.13810022175312042, 'train/grad_norm_pre_clip': 62.36275863647461, 'timing/data': 0.0007089730352163315, 'timing/model': 1.6951705878600478, 'learning_rate/qwen_vl_interface': 6.9047657140827935e-06, 'learning_rate/action_model': 6.904765714082794e-05, 'epoch': 0.77}) +[2026-06-23 19:29:50,037][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.09468454122543335, 'train/grad_norm_pre_clip': 18.464689254760742, 'timing/data': 0.0006670739967375994, 'timing/model': 1.6983510870486498, 'learning_rate/qwen_vl_interface': 6.901910728694694e-06, 'learning_rate/action_model': 6.901910728694694e-05, 'epoch': 0.77}) +[2026-06-23 19:29:53,384][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.07309714704751968, 'train/grad_norm_pre_clip': 32.46222686767578, 'timing/data': 0.0006249649450182915, 'timing/model': 1.6991779911331832, 'learning_rate/qwen_vl_interface': 6.899055064268489e-06, 'learning_rate/action_model': 6.899055064268489e-05, 'epoch': 0.77}) +[2026-06-23 19:29:57,012][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.09706965833902359, 'train/grad_norm_pre_clip': 31.788963317871094, 'timing/data': 0.0006225199904292822, 'timing/model': 1.9858628220390528, 'learning_rate/qwen_vl_interface': 6.896198721978038e-06, 'learning_rate/action_model': 6.896198721978038e-05, 'epoch': 0.77}) +[2026-06-23 19:30:00,357][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.14880849421024323, 'train/grad_norm_pre_clip': 28.43255615234375, 'timing/data': 0.000683103920891881, 'timing/model': 1.6998744211159647, 'learning_rate/qwen_vl_interface': 6.893341702997475e-06, 'learning_rate/action_model': 6.893341702997475e-05, 'epoch': 0.77}) +[2026-06-23 19:30:03,713][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2000 +[2026-06-23 19:31:36,977][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2000 +[2026-06-23 19:31:36,978][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2000 +[2026-06-23 19:32:48,266][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2000 +[2026-06-23 19:33:18,356][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000 +[2026-06-23 19:33:18,357][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.07732357084751129, 'train/grad_norm_pre_clip': 21.798608779907227, 'eval/action_loss/seconds': 93.26225141691975, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.0768205001950264, 'eval/latency_2/loss': 0.0768205001950264, 'eval/flappy/loss': 0.0768205001950264, 'eval/flappy/latency_2/loss': 0.0768205001950264, 'eval/flap_precision': 0.9267624914442163, 'eval/flap_recall': 0.9136302294197031, 'eval/flap_f1': 0.9201495073054706, 'eval/flap_rate_model': 0.20291666666666666, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9201495073054706, 'eval/latency_2/flap_precision': 0.9267624914442163, 'eval/latency_2/flap_recall': 0.9136302294197031, 'eval/latency_2/flap_f1': 0.9201495073054706, 'eval/latency_2/flap_rate_model': 0.20291666666666666, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9201495073054706, 'eval/action_classification/data_wait_seconds': 0.9479663968086243, 'eval/action_classification/predict_seconds': 70.19444274902344, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 20.600000000000115, 'rl_games_eval/mid_train/mean_length': 184.6, 'rl_games_eval/mid_train/std_reward': 19.205728312147027, 'rl_games_eval/mid_train/std_length': 154.90590692417123, 'rl_games_eval/mid_train/macro_mean_reward': 20.600000000000115, 'rl_games_eval/mid_train/macro_mean_length': 184.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 20.600000000000115, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 184.6, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 19.205728312147027, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 154.90590692417123, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 20.600000000000115, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0008348401170223951, 'timing/model': 1.7075961350928992, 'learning_rate/qwen_vl_interface': 6.890484008501211e-06, 'learning_rate/action_model': 6.890484008501211e-05, 'epoch': 0.77}) +[2026-06-23 19:33:18,357][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_2000_state +[2026-06-23 19:33:18,360][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 19:33:21,001][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_2000_state/model.safetensors +[2026-06-23 19:33:22,185][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_2000_state/optimizer.bin +[2026-06-23 19:33:22,186][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_2000_state/sampler.bin +[2026-06-23 19:33:22,186][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_2000_state/sampler_1.bin +[2026-06-23 19:33:22,186][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_2000_state/random_states_0.pkl +[2026-06-23 19:33:22,212][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000 +[2026-06-23 19:33:22,212][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 19:33:22,222][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 19:33:25,514][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.20286352932453156, 'train/grad_norm_pre_clip': 17.245922088623047, 'timing/data': 0.0006264680996537209, 'timing/model': 1.6649745909962803, 'learning_rate/qwen_vl_interface': 6.887625639663941e-06, 'learning_rate/action_model': 6.88762563966394e-05, 'epoch': 0.77}) +[2026-06-23 19:33:28,808][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.1730547398328781, 'train/grad_norm_pre_clip': 20.795225143432617, 'timing/data': 0.0007081008516252041, 'timing/model': 1.675478745950386, 'learning_rate/qwen_vl_interface': 6.884766597660626e-06, 'learning_rate/action_model': 6.884766597660626e-05, 'epoch': 0.77}) +[2026-06-23 19:33:32,116][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.13491201400756836, 'train/grad_norm_pre_clip': 24.116899490356445, 'timing/data': 0.0006183418445289135, 'timing/model': 1.6748336439486593, 'learning_rate/qwen_vl_interface': 6.881906883666515e-06, 'learning_rate/action_model': 6.881906883666515e-05, 'epoch': 0.78}) +[2026-06-23 19:33:35,454][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.13368083536624908, 'train/grad_norm_pre_clip': 22.667818069458008, 'timing/data': 0.0006880229339003563, 'timing/model': 1.6841697529889643, 'learning_rate/qwen_vl_interface': 6.879046498857125e-06, 'learning_rate/action_model': 6.879046498857124e-05, 'epoch': 0.78}) +[2026-06-23 19:33:38,875][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.14271101355552673, 'train/grad_norm_pre_clip': 17.681842803955078, 'timing/data': 0.0007852970156818628, 'timing/model': 1.682419473072514, 'learning_rate/qwen_vl_interface': 6.876185444408255e-06, 'learning_rate/action_model': 6.876185444408255e-05, 'epoch': 0.78}) +[2026-06-23 19:33:42,188][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.11610078811645508, 'train/grad_norm_pre_clip': 19.838003158569336, 'timing/data': 0.0006069110240787268, 'timing/model': 1.6838714818004519, 'learning_rate/qwen_vl_interface': 6.873323721495976e-06, 'learning_rate/action_model': 6.873323721495976e-05, 'epoch': 0.78}) +[2026-06-23 19:33:45,497][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.057871002703905106, 'train/grad_norm_pre_clip': 15.871553421020508, 'timing/data': 0.0006859400309622288, 'timing/model': 1.681607676902786, 'learning_rate/qwen_vl_interface': 6.870461331296632e-06, 'learning_rate/action_model': 6.870461331296632e-05, 'epoch': 0.78}) +[2026-06-23 19:33:48,810][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.15309855341911316, 'train/grad_norm_pre_clip': 59.53564453125, 'timing/data': 0.00059840502217412, 'timing/model': 1.6804467150941491, 'learning_rate/qwen_vl_interface': 6.8675982749868455e-06, 'learning_rate/action_model': 6.867598274986845e-05, 'epoch': 0.78}) +[2026-06-23 19:33:52,129][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.13012288510799408, 'train/grad_norm_pre_clip': 30.362071990966797, 'timing/data': 0.0006393820513039827, 'timing/model': 1.6869639360811561, 'learning_rate/qwen_vl_interface': 6.864734553743509e-06, 'learning_rate/action_model': 6.86473455374351e-05, 'epoch': 0.78}) +[2026-06-23 19:33:55,451][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.10516843944787979, 'train/grad_norm_pre_clip': 23.087955474853516, 'timing/data': 0.0007271971553564072, 'timing/model': 1.687123856972903, 'learning_rate/qwen_vl_interface': 6.861870168743793e-06, 'learning_rate/action_model': 6.861870168743793e-05, 'epoch': 0.78}) +[2026-06-23 19:33:58,770][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.10827276855707169, 'train/grad_norm_pre_clip': 34.684078216552734, 'timing/data': 0.0006511039100587368, 'timing/model': 1.6858612031210214, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.78}) +[2026-06-23 19:34:02,092][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.17135758697986603, 'train/grad_norm_pre_clip': 43.671165466308594, 'timing/data': 0.0006639480125159025, 'timing/model': 1.690898803062737, 'learning_rate/qwen_vl_interface': 6.85613941218525e-06, 'learning_rate/action_model': 6.85613941218525e-05, 'epoch': 0.78}) +[2026-06-23 19:34:05,413][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.0730343833565712, 'train/grad_norm_pre_clip': 42.57518005371094, 'timing/data': 0.0006083028856664896, 'timing/model': 1.6833866720553488, 'learning_rate/qwen_vl_interface': 6.8532730429821215e-06, 'learning_rate/action_model': 6.853273042982122e-05, 'epoch': 0.78}) +[2026-06-23 19:34:08,738][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.08298209309577942, 'train/grad_norm_pre_clip': 14.5785551071167, 'timing/data': 0.0006229311693459749, 'timing/model': 1.6884432369843125, 'learning_rate/qwen_vl_interface': 6.850406014734005e-06, 'learning_rate/action_model': 6.850406014734005e-05, 'epoch': 0.78}) +[2026-06-23 19:34:12,063][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.11471322923898697, 'train/grad_norm_pre_clip': 31.98552894592285, 'timing/data': 0.0007893240544945002, 'timing/model': 1.684004592942074, 'learning_rate/qwen_vl_interface': 6.8475383286194295e-06, 'learning_rate/action_model': 6.84753832861943e-05, 'epoch': 0.78}) +[2026-06-23 19:34:15,389][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.10284624993801117, 'train/grad_norm_pre_clip': 28.245929718017578, 'timing/data': 0.0006094060372561216, 'timing/model': 1.6888390611857176, 'learning_rate/qwen_vl_interface': 6.844669985817192e-06, 'learning_rate/action_model': 6.844669985817192e-05, 'epoch': 0.78}) +[2026-06-23 19:34:18,722][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.07989509403705597, 'train/grad_norm_pre_clip': 11.99203872680664, 'timing/data': 0.0006690279114991426, 'timing/model': 1.6928664022125304, 'learning_rate/qwen_vl_interface': 6.8418009875063594e-06, 'learning_rate/action_model': 6.84180098750636e-05, 'epoch': 0.78}) +[2026-06-23 19:34:22,050][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.1910841017961502, 'train/grad_norm_pre_clip': 49.513633728027344, 'timing/data': 0.0008873881306499243, 'timing/model': 1.6916150702163577, 'learning_rate/qwen_vl_interface': 6.838931334866267e-06, 'learning_rate/action_model': 6.838931334866266e-05, 'epoch': 0.78}) +[2026-06-23 19:34:25,471][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.1437753438949585, 'train/grad_norm_pre_clip': 16.27073097229004, 'timing/data': 0.0007274779491126537, 'timing/model': 1.777083460940048, 'learning_rate/qwen_vl_interface': 6.836061029076523e-06, 'learning_rate/action_model': 6.836061029076522e-05, 'epoch': 0.78}) +[2026-06-23 19:34:28,799][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.17606186866760254, 'train/grad_norm_pre_clip': 23.276020050048828, 'timing/data': 0.0006210468709468842, 'timing/model': 1.6890303411055356, 'learning_rate/qwen_vl_interface': 6.833190071317e-06, 'learning_rate/action_model': 6.833190071316999e-05, 'epoch': 0.78}) +[2026-06-23 19:34:32,128][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.10149972140789032, 'train/grad_norm_pre_clip': 30.948719024658203, 'timing/data': 0.0009197690524160862, 'timing/model': 1.6909079097677022, 'learning_rate/qwen_vl_interface': 6.830318462767841e-06, 'learning_rate/action_model': 6.830318462767841e-05, 'epoch': 0.78}) +[2026-06-23 19:34:35,448][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.182386577129364, 'train/grad_norm_pre_clip': 71.43306732177734, 'timing/data': 0.0006029030773788691, 'timing/model': 1.6823680470697582, 'learning_rate/qwen_vl_interface': 6.827446204609456e-06, 'learning_rate/action_model': 6.827446204609456e-05, 'epoch': 0.78}) +[2026-06-23 19:34:38,770][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.1135711595416069, 'train/grad_norm_pre_clip': 20.166702270507812, 'timing/data': 0.000608373899012804, 'timing/model': 1.683627574937418, 'learning_rate/qwen_vl_interface': 6.824573298022522e-06, 'learning_rate/action_model': 6.824573298022522e-05, 'epoch': 0.78}) +[2026-06-23 19:34:42,105][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.11840910464525223, 'train/grad_norm_pre_clip': 45.588871002197266, 'timing/data': 0.0008260419126600027, 'timing/model': 1.6940538939088583, 'learning_rate/qwen_vl_interface': 6.8216997441879805e-06, 'learning_rate/action_model': 6.821699744187981e-05, 'epoch': 0.78}) +[2026-06-23 19:34:45,433][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.18406084179878235, 'train/grad_norm_pre_clip': 64.72659301757812, 'timing/data': 0.0006222398951649666, 'timing/model': 1.6860876090358943, 'learning_rate/qwen_vl_interface': 6.818825544287045e-06, 'learning_rate/action_model': 6.818825544287045e-05, 'epoch': 0.78}) +[2026-06-23 19:34:48,752][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.2637609839439392, 'train/grad_norm_pre_clip': 39.008697509765625, 'timing/data': 0.0006627258844673634, 'timing/model': 1.685526164015755, 'learning_rate/qwen_vl_interface': 6.815950699501186e-06, 'learning_rate/action_model': 6.815950699501186e-05, 'epoch': 0.78}) +[2026-06-23 19:34:52,075][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.060693394392728806, 'train/grad_norm_pre_clip': 15.981035232543945, 'timing/data': 0.0006474070250988007, 'timing/model': 1.6865426748991013, 'learning_rate/qwen_vl_interface': 6.813075211012148e-06, 'learning_rate/action_model': 6.813075211012147e-05, 'epoch': 0.78}) +[2026-06-23 19:34:55,395][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.1243433952331543, 'train/grad_norm_pre_clip': 42.908538818359375, 'timing/data': 0.0006736661307513714, 'timing/model': 1.6851421019528061, 'learning_rate/qwen_vl_interface': 6.81019908000193e-06, 'learning_rate/action_model': 6.81019908000193e-05, 'epoch': 0.78}) +[2026-06-23 19:34:58,725][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.06693822145462036, 'train/grad_norm_pre_clip': 42.287105560302734, 'timing/data': 0.0006221688818186522, 'timing/model': 1.6872173240408301, 'learning_rate/qwen_vl_interface': 6.807322307652806e-06, 'learning_rate/action_model': 6.807322307652806e-05, 'epoch': 0.79}) +[2026-06-23 19:35:02,054][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.2308182567358017, 'train/grad_norm_pre_clip': 36.061973571777344, 'timing/data': 0.0005996068939566612, 'timing/model': 1.6889301631599665, 'learning_rate/qwen_vl_interface': 6.804444895147306e-06, 'learning_rate/action_model': 6.804444895147306e-05, 'epoch': 0.79}) +[2026-06-23 19:35:05,388][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.13129159808158875, 'train/grad_norm_pre_clip': 36.88295364379883, 'timing/data': 0.0008167850319296122, 'timing/model': 1.692550488980487, 'learning_rate/qwen_vl_interface': 6.8015668436682255e-06, 'learning_rate/action_model': 6.801566843668226e-05, 'epoch': 0.79}) +[2026-06-23 19:35:08,716][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.15273375809192657, 'train/grad_norm_pre_clip': 33.53312301635742, 'timing/data': 0.0006370581686496735, 'timing/model': 1.6887649928685278, 'learning_rate/qwen_vl_interface': 6.7986881543986245e-06, 'learning_rate/action_model': 6.798688154398624e-05, 'epoch': 0.79}) +[2026-06-23 19:35:12,044][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.07408184558153152, 'train/grad_norm_pre_clip': 41.86908721923828, 'timing/data': 0.0005995570681989193, 'timing/model': 1.68800303991884, 'learning_rate/qwen_vl_interface': 6.795808828521822e-06, 'learning_rate/action_model': 6.795808828521821e-05, 'epoch': 0.79}) +[2026-06-23 19:35:15,470][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.3079332709312439, 'train/grad_norm_pre_clip': 106.95878601074219, 'timing/data': 0.0005992068909108639, 'timing/model': 1.694070405093953, 'learning_rate/qwen_vl_interface': 6.792928867221401e-06, 'learning_rate/action_model': 6.792928867221401e-05, 'epoch': 0.79}) +[2026-06-23 19:35:18,804][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.23443949222564697, 'train/grad_norm_pre_clip': 80.54605865478516, 'timing/data': 0.0006214978639036417, 'timing/model': 1.6921753229107708, 'learning_rate/qwen_vl_interface': 6.790048271681202e-06, 'learning_rate/action_model': 6.790048271681202e-05, 'epoch': 0.79}) +[2026-06-23 19:35:22,133][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.06904397904872894, 'train/grad_norm_pre_clip': 98.31547546386719, 'timing/data': 0.0006153190042823553, 'timing/model': 1.6899348779115826, 'learning_rate/qwen_vl_interface': 6.7871670430853335e-06, 'learning_rate/action_model': 6.787167043085333e-05, 'epoch': 0.79}) +[2026-06-23 19:35:25,470][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.15475283563137054, 'train/grad_norm_pre_clip': 13.795825004577637, 'timing/data': 0.0006108100060373545, 'timing/model': 1.6936677349731326, 'learning_rate/qwen_vl_interface': 6.784285182618156e-06, 'learning_rate/action_model': 6.784285182618156e-05, 'epoch': 0.79}) +[2026-06-23 19:35:28,815][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.156626895070076, 'train/grad_norm_pre_clip': 51.1397705078125, 'timing/data': 0.0006615060847252607, 'timing/model': 1.6995098029728979, 'learning_rate/qwen_vl_interface': 6.781402691464297e-06, 'learning_rate/action_model': 6.781402691464297e-05, 'epoch': 0.79}) +[2026-06-23 19:35:32,141][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.1555168777704239, 'train/grad_norm_pre_clip': 84.2265625, 'timing/data': 0.000692443922162056, 'timing/model': 1.6858608471229672, 'learning_rate/qwen_vl_interface': 6.778519570808636e-06, 'learning_rate/action_model': 6.778519570808636e-05, 'epoch': 0.79}) +[2026-06-23 19:35:35,478][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.18876126408576965, 'train/grad_norm_pre_clip': 53.78763961791992, 'timing/data': 0.0006059810984879732, 'timing/model': 1.69587944005616, 'learning_rate/qwen_vl_interface': 6.775635821836317e-06, 'learning_rate/action_model': 6.775635821836317e-05, 'epoch': 0.79}) +[2026-06-23 19:35:38,817][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.10009195655584335, 'train/grad_norm_pre_clip': 16.947641372680664, 'timing/data': 0.0005984471645206213, 'timing/model': 1.6965186418965459, 'learning_rate/qwen_vl_interface': 6.77275144573274e-06, 'learning_rate/action_model': 6.77275144573274e-05, 'epoch': 0.79}) +[2026-06-23 19:35:42,152][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.14208297431468964, 'train/grad_norm_pre_clip': 21.625513076782227, 'timing/data': 0.0006126640364527702, 'timing/model': 1.6959707878995687, 'learning_rate/qwen_vl_interface': 6.769866443683564e-06, 'learning_rate/action_model': 6.769866443683563e-05, 'epoch': 0.79}) +[2026-06-23 19:35:45,486][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.09766250848770142, 'train/grad_norm_pre_clip': 18.789260864257812, 'timing/data': 0.0006116321310400963, 'timing/model': 1.6945770611055195, 'learning_rate/qwen_vl_interface': 6.766980816874701e-06, 'learning_rate/action_model': 6.7669808168747e-05, 'epoch': 0.79}) +[2026-06-23 19:35:48,826][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.11658944189548492, 'train/grad_norm_pre_clip': 45.945987701416016, 'timing/data': 0.0006092970725148916, 'timing/model': 1.6937379580922425, 'learning_rate/qwen_vl_interface': 6.764094566492327e-06, 'learning_rate/action_model': 6.764094566492328e-05, 'epoch': 0.79}) +[2026-06-23 19:35:52,165][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.10124288499355316, 'train/grad_norm_pre_clip': 16.36595916748047, 'timing/data': 0.0006296359933912754, 'timing/model': 1.697297960985452, 'learning_rate/qwen_vl_interface': 6.7612076937228685e-06, 'learning_rate/action_model': 6.761207693722869e-05, 'epoch': 0.79}) +[2026-06-23 19:35:55,499][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.10343710333108902, 'train/grad_norm_pre_clip': 13.018038749694824, 'timing/data': 0.0006053601391613483, 'timing/model': 1.6919590360485017, 'learning_rate/qwen_vl_interface': 6.7583201997530126e-06, 'learning_rate/action_model': 6.758320199753012e-05, 'epoch': 0.79}) +[2026-06-23 19:35:58,822][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.24703262746334076, 'train/grad_norm_pre_clip': 17.047006607055664, 'timing/data': 0.0008060978725552559, 'timing/model': 1.6880524780135602, 'learning_rate/qwen_vl_interface': 6.755432085769693e-06, 'learning_rate/action_model': 6.755432085769694e-05, 'epoch': 0.79}) +[2026-06-23 19:36:02,149][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.22952763736248016, 'train/grad_norm_pre_clip': 73.88887786865234, 'timing/data': 0.000821626977995038, 'timing/model': 1.6926517179235816, 'learning_rate/qwen_vl_interface': 6.75254335296011e-06, 'learning_rate/action_model': 6.75254335296011e-05, 'epoch': 0.79}) +[2026-06-23 19:36:05,638][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.05056339502334595, 'train/grad_norm_pre_clip': 52.85069274902344, 'timing/data': 0.0006090758834034204, 'timing/model': 1.695262640947476, 'learning_rate/qwen_vl_interface': 6.749654002511712e-06, 'learning_rate/action_model': 6.749654002511712e-05, 'epoch': 0.79}) +[2026-06-23 19:36:08,969][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.16798128187656403, 'train/grad_norm_pre_clip': 33.92658233642578, 'timing/data': 0.0006142968777567148, 'timing/model': 1.69180580182001, 'learning_rate/qwen_vl_interface': 6.7467640356122e-06, 'learning_rate/action_model': 6.746764035612199e-05, 'epoch': 0.79}) +[2026-06-23 19:36:12,291][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.24823766946792603, 'train/grad_norm_pre_clip': 17.648439407348633, 'timing/data': 0.0006131441332399845, 'timing/model': 1.688022157875821, 'learning_rate/qwen_vl_interface': 6.74387345344953e-06, 'learning_rate/action_model': 6.74387345344953e-05, 'epoch': 0.79}) +[2026-06-23 19:36:15,617][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.19617442786693573, 'train/grad_norm_pre_clip': 29.76495933532715, 'timing/data': 0.0006065219640731812, 'timing/model': 1.689415160100907, 'learning_rate/qwen_vl_interface': 6.740982257211911e-06, 'learning_rate/action_model': 6.740982257211911e-05, 'epoch': 0.79}) +[2026-06-23 19:36:18,959][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.1367582082748413, 'train/grad_norm_pre_clip': 62.157745361328125, 'timing/data': 0.0006700018420815468, 'timing/model': 1.6996150761842728, 'learning_rate/qwen_vl_interface': 6.738090448087806e-06, 'learning_rate/action_model': 6.738090448087806e-05, 'epoch': 0.79}) +[2026-06-23 19:36:22,292][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.12651637196540833, 'train/grad_norm_pre_clip': 33.09436798095703, 'timing/data': 0.0007163188420236111, 'timing/model': 1.6932141531724483, 'learning_rate/qwen_vl_interface': 6.735198027265928e-06, 'learning_rate/action_model': 6.735198027265928e-05, 'epoch': 0.79}) +[2026-06-23 19:36:25,624][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.17799238860607147, 'train/grad_norm_pre_clip': 11.608689308166504, 'timing/data': 0.0006711040623486042, 'timing/model': 1.694074132014066, 'learning_rate/qwen_vl_interface': 6.7323049959352424e-06, 'learning_rate/action_model': 6.732304995935243e-05, 'epoch': 0.8}) +[2026-06-23 19:36:28,953][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.13129504024982452, 'train/grad_norm_pre_clip': 18.407150268554688, 'timing/data': 0.0005960229318588972, 'timing/model': 1.6943585791159421, 'learning_rate/qwen_vl_interface': 6.729411355284962e-06, 'learning_rate/action_model': 6.729411355284963e-05, 'epoch': 0.8}) +[2026-06-23 19:36:32,292][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.16654546558856964, 'train/grad_norm_pre_clip': 28.999103546142578, 'timing/data': 0.000608034897595644, 'timing/model': 1.693852897034958, 'learning_rate/qwen_vl_interface': 6.726517106504559e-06, 'learning_rate/action_model': 6.726517106504558e-05, 'epoch': 0.8}) +[2026-06-23 19:36:35,634][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.18939530849456787, 'train/grad_norm_pre_clip': 21.081872940063477, 'timing/data': 0.0006270110607147217, 'timing/model': 1.6989535610191524, 'learning_rate/qwen_vl_interface': 6.723622250783745e-06, 'learning_rate/action_model': 6.723622250783744e-05, 'epoch': 0.8}) +[2026-06-23 19:36:38,974][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.20776543021202087, 'train/grad_norm_pre_clip': 66.89269256591797, 'timing/data': 0.0006630178540945053, 'timing/model': 1.6892009200528264, 'learning_rate/qwen_vl_interface': 6.720726789312487e-06, 'learning_rate/action_model': 6.720726789312487e-05, 'epoch': 0.8}) +[2026-06-23 19:36:42,319][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.11103236675262451, 'train/grad_norm_pre_clip': 21.886075973510742, 'timing/data': 0.0006848301272839308, 'timing/model': 1.6997407649178058, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.8}) +[2026-06-23 19:36:45,661][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.1511794626712799, 'train/grad_norm_pre_clip': 49.039825439453125, 'timing/data': 0.0007001680787652731, 'timing/model': 1.69754817802459, 'learning_rate/qwen_vl_interface': 6.714934053879749e-06, 'learning_rate/action_model': 6.71493405387975e-05, 'epoch': 0.8}) +[2026-06-23 19:36:49,003][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.13033823668956757, 'train/grad_norm_pre_clip': 14.26913070678711, 'timing/data': 0.0006250371225178242, 'timing/model': 1.6981628579087555, 'learning_rate/qwen_vl_interface': 6.712036782299442e-06, 'learning_rate/action_model': 6.712036782299443e-05, 'epoch': 0.8}) +[2026-06-23 19:36:52,475][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.19320577383041382, 'train/grad_norm_pre_clip': 45.69242477416992, 'timing/data': 0.0007122410461306572, 'timing/model': 1.8268292199354619, 'learning_rate/qwen_vl_interface': 6.709138909731041e-06, 'learning_rate/action_model': 6.709138909731041e-05, 'epoch': 0.8}) +[2026-06-23 19:36:55,804][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.1576065719127655, 'train/grad_norm_pre_clip': 29.881839752197266, 'timing/data': 0.0006203781813383102, 'timing/model': 1.6954494290985167, 'learning_rate/qwen_vl_interface': 6.706240437365751e-06, 'learning_rate/action_model': 6.70624043736575e-05, 'epoch': 0.8}) +[2026-06-23 19:36:59,132][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.08205612003803253, 'train/grad_norm_pre_clip': 44.7177848815918, 'timing/data': 0.000608926871791482, 'timing/model': 1.6834950770717114, 'learning_rate/qwen_vl_interface': 6.703341366395023e-06, 'learning_rate/action_model': 6.703341366395023e-05, 'epoch': 0.8}) +[2026-06-23 19:37:02,470][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.10951153188943863, 'train/grad_norm_pre_clip': 35.97692108154297, 'timing/data': 0.0006134449504315853, 'timing/model': 1.6959195809904486, 'learning_rate/qwen_vl_interface': 6.700441698010558e-06, 'learning_rate/action_model': 6.700441698010558e-05, 'epoch': 0.8}) +[2026-06-23 19:37:05,811][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.16404588520526886, 'train/grad_norm_pre_clip': 43.57570266723633, 'timing/data': 0.0006072330288589001, 'timing/model': 1.6989020770415664, 'learning_rate/qwen_vl_interface': 6.697541433404299e-06, 'learning_rate/action_model': 6.697541433404299e-05, 'epoch': 0.8}) +[2026-06-23 19:37:09,144][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.05422685295343399, 'train/grad_norm_pre_clip': 9.607114791870117, 'timing/data': 0.0006081648170948029, 'timing/model': 1.6945485949981958, 'learning_rate/qwen_vl_interface': 6.6946405737684325e-06, 'learning_rate/action_model': 6.694640573768433e-05, 'epoch': 0.8}) +[2026-06-23 19:37:12,475][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.09556421637535095, 'train/grad_norm_pre_clip': 58.076534271240234, 'timing/data': 0.0008084808941930532, 'timing/model': 1.6935068159364164, 'learning_rate/qwen_vl_interface': 6.691739120295397e-06, 'learning_rate/action_model': 6.691739120295397e-05, 'epoch': 0.8}) +[2026-06-23 19:37:15,819][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.1021851897239685, 'train/grad_norm_pre_clip': 55.95854187011719, 'timing/data': 0.0006130239926278591, 'timing/model': 1.6998991630971432, 'learning_rate/qwen_vl_interface': 6.688837074177868e-06, 'learning_rate/action_model': 6.688837074177868e-05, 'epoch': 0.8}) +[2026-06-23 19:37:19,159][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.2954379916191101, 'train/grad_norm_pre_clip': 94.71033477783203, 'timing/data': 0.000659391051158309, 'timing/model': 1.6956952530890703, 'learning_rate/qwen_vl_interface': 6.68593443660877e-06, 'learning_rate/action_model': 6.685934436608769e-05, 'epoch': 0.8}) +[2026-06-23 19:37:22,489][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.06543092429637909, 'train/grad_norm_pre_clip': 10.134166717529297, 'timing/data': 0.0005974951200187206, 'timing/model': 1.6920234928838909, 'learning_rate/qwen_vl_interface': 6.6830312087812625e-06, 'learning_rate/action_model': 6.683031208781262e-05, 'epoch': 0.8}) +[2026-06-23 19:37:25,826][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.13622763752937317, 'train/grad_norm_pre_clip': 50.007232666015625, 'timing/data': 0.0006174331065267324, 'timing/model': 1.6965587048325688, 'learning_rate/qwen_vl_interface': 6.680127391888754e-06, 'learning_rate/action_model': 6.680127391888754e-05, 'epoch': 0.8}) +[2026-06-23 19:37:29,164][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.12135138362646103, 'train/grad_norm_pre_clip': 21.745208740234375, 'timing/data': 0.0006186640821397305, 'timing/model': 1.6907850198913366, 'learning_rate/qwen_vl_interface': 6.677222987124897e-06, 'learning_rate/action_model': 6.677222987124897e-05, 'epoch': 0.8}) +[2026-06-23 19:37:32,496][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.059341758489608765, 'train/grad_norm_pre_clip': 13.951281547546387, 'timing/data': 0.0006086761131882668, 'timing/model': 1.6941446580458432, 'learning_rate/qwen_vl_interface': 6.67431799568358e-06, 'learning_rate/action_model': 6.67431799568358e-05, 'epoch': 0.8}) +[2026-06-23 19:37:35,837][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.2286863923072815, 'train/grad_norm_pre_clip': 44.69472122192383, 'timing/data': 0.0006089559756219387, 'timing/model': 1.692960652988404, 'learning_rate/qwen_vl_interface': 6.671412418758939e-06, 'learning_rate/action_model': 6.671412418758938e-05, 'epoch': 0.8}) +[2026-06-23 19:37:39,179][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.20362691581249237, 'train/grad_norm_pre_clip': 24.740121841430664, 'timing/data': 0.0006605039816349745, 'timing/model': 1.6976322720292956, 'learning_rate/qwen_vl_interface': 6.668506257545341e-06, 'learning_rate/action_model': 6.668506257545341e-05, 'epoch': 0.8}) +[2026-06-23 19:37:42,807][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.14237305521965027, 'train/grad_norm_pre_clip': 30.524539947509766, 'timing/data': 0.0005984469316899776, 'timing/model': 1.9828254911117256, 'learning_rate/qwen_vl_interface': 6.665599513237403e-06, 'learning_rate/action_model': 6.665599513237402e-05, 'epoch': 0.8}) +[2026-06-23 19:37:46,147][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.13917015492916107, 'train/grad_norm_pre_clip': 31.075374603271484, 'timing/data': 0.00061448710039258, 'timing/model': 1.6971667020116001, 'learning_rate/qwen_vl_interface': 6.662692187029978e-06, 'learning_rate/action_model': 6.662692187029978e-05, 'epoch': 0.8}) +[2026-06-23 19:37:49,487][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.19721068441867828, 'train/grad_norm_pre_clip': 18.892805099487305, 'timing/data': 0.0006022630259394646, 'timing/model': 1.6969476230442524, 'learning_rate/qwen_vl_interface': 6.659784280118156e-06, 'learning_rate/action_model': 6.659784280118156e-05, 'epoch': 0.8}) +[2026-06-23 19:37:52,826][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.30560365319252014, 'train/grad_norm_pre_clip': 29.626609802246094, 'timing/data': 0.000609156908467412, 'timing/model': 1.6972068469040096, 'learning_rate/qwen_vl_interface': 6.65687579369727e-06, 'learning_rate/action_model': 6.656875793697271e-05, 'epoch': 0.81}) +[2026-06-23 19:37:56,170][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.15365605056285858, 'train/grad_norm_pre_clip': 62.16006851196289, 'timing/data': 0.0006225020624697208, 'timing/model': 1.697930432856083, 'learning_rate/qwen_vl_interface': 6.653966728962885e-06, 'learning_rate/action_model': 6.653966728962885e-05, 'epoch': 0.81}) +[2026-06-23 19:37:59,513][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.11337875574827194, 'train/grad_norm_pre_clip': 21.56889533996582, 'timing/data': 0.0006684481631964445, 'timing/model': 1.7014748509973288, 'learning_rate/qwen_vl_interface': 6.651057087110813e-06, 'learning_rate/action_model': 6.651057087110813e-05, 'epoch': 0.81}) +[2026-06-23 19:38:02,855][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.10036914050579071, 'train/grad_norm_pre_clip': 38.14523696899414, 'timing/data': 0.0006700109224766493, 'timing/model': 1.6975411910098046, 'learning_rate/qwen_vl_interface': 6.6481468693370924e-06, 'learning_rate/action_model': 6.648146869337091e-05, 'epoch': 0.81}) +[2026-06-23 19:38:06,187][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.08223100006580353, 'train/grad_norm_pre_clip': 13.419384002685547, 'timing/data': 0.0006259181536734104, 'timing/model': 1.6929604848846793, 'learning_rate/qwen_vl_interface': 6.645236076838009e-06, 'learning_rate/action_model': 6.645236076838009e-05, 'epoch': 0.81}) +[2026-06-23 19:38:09,535][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.23483528196811676, 'train/grad_norm_pre_clip': 34.48908996582031, 'timing/data': 0.0007133630570024252, 'timing/model': 1.6982892300002277, 'learning_rate/qwen_vl_interface': 6.642324710810075e-06, 'learning_rate/action_model': 6.642324710810076e-05, 'epoch': 0.81}) +[2026-06-23 19:38:12,873][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.06581667810678482, 'train/grad_norm_pre_clip': 13.126439094543457, 'timing/data': 0.0006232939194887877, 'timing/model': 1.6961407121270895, 'learning_rate/qwen_vl_interface': 6.639412772450047e-06, 'learning_rate/action_model': 6.639412772450047e-05, 'epoch': 0.81}) +[2026-06-23 19:38:16,209][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.10368166118860245, 'train/grad_norm_pre_clip': 45.19327926635742, 'timing/data': 0.0006175930611789227, 'timing/model': 1.6957254770677537, 'learning_rate/qwen_vl_interface': 6.6365002629549104e-06, 'learning_rate/action_model': 6.63650026295491e-05, 'epoch': 0.81}) +[2026-06-23 19:38:19,551][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.18043024837970734, 'train/grad_norm_pre_clip': 44.4842643737793, 'timing/data': 0.000658129109069705, 'timing/model': 1.6980462209321558, 'learning_rate/qwen_vl_interface': 6.633587183521887e-06, 'learning_rate/action_model': 6.633587183521887e-05, 'epoch': 0.81}) +[2026-06-23 19:38:22,887][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.1572919487953186, 'train/grad_norm_pre_clip': 34.3633918762207, 'timing/data': 0.0006018630228936672, 'timing/model': 1.6939432341605425, 'learning_rate/qwen_vl_interface': 6.6306735353484365e-06, 'learning_rate/action_model': 6.630673535348436e-05, 'epoch': 0.81}) +[2026-06-23 19:38:26,233][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.20857694745063782, 'train/grad_norm_pre_clip': 67.91268157958984, 'timing/data': 0.0006592911668121815, 'timing/model': 1.7004853740800172, 'learning_rate/qwen_vl_interface': 6.627759319632246e-06, 'learning_rate/action_model': 6.627759319632247e-05, 'epoch': 0.81}) +[2026-06-23 19:38:29,679][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.1823759377002716, 'train/grad_norm_pre_clip': 24.659709930419922, 'timing/data': 0.0006161800120025873, 'timing/model': 1.8012589630670846, 'learning_rate/qwen_vl_interface': 6.624844537571245e-06, 'learning_rate/action_model': 6.624844537571245e-05, 'epoch': 0.81}) +[2026-06-23 19:38:33,020][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.152879998087883, 'train/grad_norm_pre_clip': 17.72287368774414, 'timing/data': 0.0005995389074087143, 'timing/model': 1.695513429120183, 'learning_rate/qwen_vl_interface': 6.621929190363585e-06, 'learning_rate/action_model': 6.621929190363585e-05, 'epoch': 0.81}) +[2026-06-23 19:38:36,354][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.1565120816230774, 'train/grad_norm_pre_clip': 15.750247955322266, 'timing/data': 0.0006202568765729666, 'timing/model': 1.6910432509612292, 'learning_rate/qwen_vl_interface': 6.619013279207657e-06, 'learning_rate/action_model': 6.619013279207656e-05, 'epoch': 0.81}) +[2026-06-23 19:38:39,696][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.21541395783424377, 'train/grad_norm_pre_clip': 30.53196907043457, 'timing/data': 0.0006854101084172726, 'timing/model': 1.6977380900643766, 'learning_rate/qwen_vl_interface': 6.616096805302083e-06, 'learning_rate/action_model': 6.616096805302082e-05, 'epoch': 0.81}) +[2026-06-23 19:38:43,035][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.10112756490707397, 'train/grad_norm_pre_clip': 14.359405517578125, 'timing/data': 0.0006159490440040827, 'timing/model': 1.6954554610420018, 'learning_rate/qwen_vl_interface': 6.6131797698457135e-06, 'learning_rate/action_model': 6.613179769845713e-05, 'epoch': 0.81}) +[2026-06-23 19:38:46,375][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.07079602032899857, 'train/grad_norm_pre_clip': 14.023174285888672, 'timing/data': 0.0006980439648032188, 'timing/model': 1.6936379470862448, 'learning_rate/qwen_vl_interface': 6.610262174037633e-06, 'learning_rate/action_model': 6.610262174037633e-05, 'epoch': 0.81}) +[2026-06-23 19:38:49,716][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.11055091768503189, 'train/grad_norm_pre_clip': 22.744916915893555, 'timing/data': 0.0006167010869830847, 'timing/model': 1.69685611105524, 'learning_rate/qwen_vl_interface': 6.607344019077155e-06, 'learning_rate/action_model': 6.607344019077156e-05, 'epoch': 0.81}) +[2026-06-23 19:38:53,052][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.1008741483092308, 'train/grad_norm_pre_clip': 27.349658966064453, 'timing/data': 0.0006037969142198563, 'timing/model': 1.6927323578856885, 'learning_rate/qwen_vl_interface': 6.604425306163822e-06, 'learning_rate/action_model': 6.604425306163822e-05, 'epoch': 0.81}) +[2026-06-23 19:38:56,395][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2100 +[2026-06-23 19:40:29,374][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2100 +[2026-06-23 19:40:29,375][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2100 +[2026-06-23 19:41:40,919][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2100 +[2026-06-23 19:41:40,920][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.24027177691459656, 'train/grad_norm_pre_clip': 22.396465301513672, 'eval/action_loss/seconds': 92.97689916798845, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07780817151069641, 'eval/latency_2/loss': 0.07780817151069641, 'eval/flappy/loss': 0.07780817151069641, 'eval/flappy/latency_2/loss': 0.07780817151069641, 'eval/flap_precision': 0.9110085695451549, 'eval/flap_recall': 0.9325236167341431, 'eval/flap_f1': 0.9216405468489497, 'eval/flap_rate_model': 0.21069444444444443, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9216405468489496, 'eval/latency_2/flap_precision': 0.9110085695451549, 'eval/latency_2/flap_recall': 0.9325236167341431, 'eval/latency_2/flap_f1': 0.9216405468489497, 'eval/latency_2/flap_rate_model': 0.21069444444444443, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9216405468489496, 'eval/action_classification/data_wait_seconds': 0.9097819328308105, 'eval/action_classification/predict_seconds': 70.4920425415039, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006057298742234707, 'timing/model': 1.697051805909723, 'learning_rate/qwen_vl_interface': 6.6015060364974045e-06, 'learning_rate/action_model': 6.601506036497404e-05, 'epoch': 0.81}) +[2026-06-23 19:41:44,474][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.12156373262405396, 'train/grad_norm_pre_clip': 15.400840759277344, 'timing/data': 0.0006087459623813629, 'timing/model': 1.6861367370001972, 'learning_rate/qwen_vl_interface': 6.598586211277907e-06, 'learning_rate/action_model': 6.598586211277907e-05, 'epoch': 0.81}) +[2026-06-23 19:41:47,804][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.18180550634860992, 'train/grad_norm_pre_clip': 16.247175216674805, 'timing/data': 0.0006130931433290243, 'timing/model': 1.6886384589597583, 'learning_rate/qwen_vl_interface': 6.59566583170556e-06, 'learning_rate/action_model': 6.59566583170556e-05, 'epoch': 0.81}) +[2026-06-23 19:41:51,139][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.09723012149333954, 'train/grad_norm_pre_clip': 23.90923500061035, 'timing/data': 0.0006082248874008656, 'timing/model': 1.6948304139077663, 'learning_rate/qwen_vl_interface': 6.59274489898082e-06, 'learning_rate/action_model': 6.592744898980819e-05, 'epoch': 0.81}) +[2026-06-23 19:41:54,459][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.22138699889183044, 'train/grad_norm_pre_clip': 31.015710830688477, 'timing/data': 0.0008192220702767372, 'timing/model': 1.685719555011019, 'learning_rate/qwen_vl_interface': 6.5898234143043716e-06, 'learning_rate/action_model': 6.589823414304371e-05, 'epoch': 0.81}) +[2026-06-23 19:41:57,790][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.07175914943218231, 'train/grad_norm_pre_clip': 15.360445022583008, 'timing/data': 0.0006416379474103451, 'timing/model': 1.6915442801546305, 'learning_rate/qwen_vl_interface': 6.5869013788771264e-06, 'learning_rate/action_model': 6.586901378877126e-05, 'epoch': 0.81}) +[2026-06-23 19:42:01,125][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.14239917695522308, 'train/grad_norm_pre_clip': 35.963748931884766, 'timing/data': 0.0006667948327958584, 'timing/model': 1.69397779321298, 'learning_rate/qwen_vl_interface': 6.583978793900223e-06, 'learning_rate/action_model': 6.583978793900223e-05, 'epoch': 0.82}) +[2026-06-23 19:42:04,456][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.09897416830062866, 'train/grad_norm_pre_clip': 22.005203247070312, 'timing/data': 0.0006965098436921835, 'timing/model': 1.6926480161491781, 'learning_rate/qwen_vl_interface': 6.581055660575028e-06, 'learning_rate/action_model': 6.581055660575027e-05, 'epoch': 0.82}) +[2026-06-23 19:42:07,790][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.1585814505815506, 'train/grad_norm_pre_clip': 37.42885971069336, 'timing/data': 0.0006125429645180702, 'timing/model': 1.6920569080393761, 'learning_rate/qwen_vl_interface': 6.578131980103128e-06, 'learning_rate/action_model': 6.578131980103128e-05, 'epoch': 0.82}) +[2026-06-23 19:42:11,126][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.1876821666955948, 'train/grad_norm_pre_clip': 13.727001190185547, 'timing/data': 0.000765389995649457, 'timing/model': 1.6908398130908608, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.82}) +[2026-06-23 19:42:14,464][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.11786628514528275, 'train/grad_norm_pre_clip': 23.244789123535156, 'timing/data': 0.0005996278487145901, 'timing/model': 1.6948160589672625, 'learning_rate/qwen_vl_interface': 6.572282982526702e-06, 'learning_rate/action_model': 6.572282982526702e-05, 'epoch': 0.82}) +[2026-06-23 19:42:17,801][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.13267172873020172, 'train/grad_norm_pre_clip': 31.27680015563965, 'timing/data': 0.0005975938402116299, 'timing/model': 1.69530682801269, 'learning_rate/qwen_vl_interface': 6.5693576678264755e-06, 'learning_rate/action_model': 6.569357667826476e-05, 'epoch': 0.82}) +[2026-06-23 19:42:21,141][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.16070297360420227, 'train/grad_norm_pre_clip': 53.041141510009766, 'timing/data': 0.0005922440905123949, 'timing/model': 1.6977833120618016, 'learning_rate/qwen_vl_interface': 6.566431810788148e-06, 'learning_rate/action_model': 6.566431810788148e-05, 'epoch': 0.82}) +[2026-06-23 19:42:24,483][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.18994192779064178, 'train/grad_norm_pre_clip': 16.833921432495117, 'timing/data': 0.0006174610462039709, 'timing/model': 1.6970348989125341, 'learning_rate/qwen_vl_interface': 6.5635054126144306e-06, 'learning_rate/action_model': 6.56350541261443e-05, 'epoch': 0.82}) +[2026-06-23 19:42:27,819][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.24748951196670532, 'train/grad_norm_pre_clip': 48.17110061645508, 'timing/data': 0.0005998089909553528, 'timing/model': 1.6965929130092263, 'learning_rate/qwen_vl_interface': 6.560578474508254e-06, 'learning_rate/action_model': 6.560578474508253e-05, 'epoch': 0.82}) +[2026-06-23 19:42:31,354][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.20115821063518524, 'train/grad_norm_pre_clip': 41.58103561401367, 'timing/data': 0.0006045771297067404, 'timing/model': 1.8928177789784968, 'learning_rate/qwen_vl_interface': 6.5576509976727685e-06, 'learning_rate/action_model': 6.557650997672769e-05, 'epoch': 0.82}) +[2026-06-23 19:42:34,686][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.10908854752779007, 'train/grad_norm_pre_clip': 45.42377853393555, 'timing/data': 0.0006219600327312946, 'timing/model': 1.6915008230134845, 'learning_rate/qwen_vl_interface': 6.554722983311354e-06, 'learning_rate/action_model': 6.554722983311355e-05, 'epoch': 0.82}) +[2026-06-23 19:42:38,025][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.2014697790145874, 'train/grad_norm_pre_clip': 20.724445343017578, 'timing/data': 0.0006161290220916271, 'timing/model': 1.6935797701589763, 'learning_rate/qwen_vl_interface': 6.551794432627605e-06, 'learning_rate/action_model': 6.551794432627605e-05, 'epoch': 0.82}) +[2026-06-23 19:42:41,361][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.22989870607852936, 'train/grad_norm_pre_clip': 38.529457092285156, 'timing/data': 0.000600059051066637, 'timing/model': 1.6927593720611185, 'learning_rate/qwen_vl_interface': 6.548865346825338e-06, 'learning_rate/action_model': 6.548865346825338e-05, 'epoch': 0.82}) +[2026-06-23 19:42:44,696][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.134430930018425, 'train/grad_norm_pre_clip': 19.55051612854004, 'timing/data': 0.0006138051394373178, 'timing/model': 1.6904441909864545, 'learning_rate/qwen_vl_interface': 6.545935727108589e-06, 'learning_rate/action_model': 6.545935727108588e-05, 'epoch': 0.82}) +[2026-06-23 19:42:48,031][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.16494639217853546, 'train/grad_norm_pre_clip': 52.26099395751953, 'timing/data': 0.0006017419509589672, 'timing/model': 1.6975521109998226, 'learning_rate/qwen_vl_interface': 6.543005574681613e-06, 'learning_rate/action_model': 6.543005574681613e-05, 'epoch': 0.82}) +[2026-06-23 19:42:51,375][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.24520672857761383, 'train/grad_norm_pre_clip': 56.7384147644043, 'timing/data': 0.000627289991825819, 'timing/model': 1.6986909301485866, 'learning_rate/qwen_vl_interface': 6.5400748907488886e-06, 'learning_rate/action_model': 6.540074890748888e-05, 'epoch': 0.82}) +[2026-06-23 19:42:54,723][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.14095774292945862, 'train/grad_norm_pre_clip': 12.536474227905273, 'timing/data': 0.0006062309257686138, 'timing/model': 1.7013507189694792, 'learning_rate/qwen_vl_interface': 6.537143676515105e-06, 'learning_rate/action_model': 6.537143676515105e-05, 'epoch': 0.82}) +[2026-06-23 19:42:58,065][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.13406799733638763, 'train/grad_norm_pre_clip': 12.799057006835938, 'timing/data': 0.0006177821196615696, 'timing/model': 1.695433123037219, 'learning_rate/qwen_vl_interface': 6.534211933185176e-06, 'learning_rate/action_model': 6.534211933185176e-05, 'epoch': 0.82}) +[2026-06-23 19:43:01,415][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.227002814412117, 'train/grad_norm_pre_clip': 56.13302993774414, 'timing/data': 0.0006125820800662041, 'timing/model': 1.7024261360056698, 'learning_rate/qwen_vl_interface': 6.531279661964231e-06, 'learning_rate/action_model': 6.531279661964231e-05, 'epoch': 0.82}) +[2026-06-23 19:43:04,770][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.13869492709636688, 'train/grad_norm_pre_clip': 31.109174728393555, 'timing/data': 0.0006288830190896988, 'timing/model': 1.7018562480807304, 'learning_rate/qwen_vl_interface': 6.528346864057616e-06, 'learning_rate/action_model': 6.528346864057615e-05, 'epoch': 0.82}) +[2026-06-23 19:43:08,117][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.17848165333271027, 'train/grad_norm_pre_clip': 54.43674850463867, 'timing/data': 0.0006131839472800493, 'timing/model': 1.7021632429677993, 'learning_rate/qwen_vl_interface': 6.5254135406708906e-06, 'learning_rate/action_model': 6.52541354067089e-05, 'epoch': 0.82}) +[2026-06-23 19:43:11,468][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.15878668427467346, 'train/grad_norm_pre_clip': 20.60839080810547, 'timing/data': 0.0006157578900456429, 'timing/model': 1.7028158579487354, 'learning_rate/qwen_vl_interface': 6.522479693009836e-06, 'learning_rate/action_model': 6.522479693009837e-05, 'epoch': 0.82}) +[2026-06-23 19:43:14,809][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.180888831615448, 'train/grad_norm_pre_clip': 29.085529327392578, 'timing/data': 0.0006073920521885157, 'timing/model': 1.6943889290560037, 'learning_rate/qwen_vl_interface': 6.519545322280446e-06, 'learning_rate/action_model': 6.519545322280446e-05, 'epoch': 0.82}) +[2026-06-23 19:43:18,153][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.0779312327504158, 'train/grad_norm_pre_clip': 30.093856811523438, 'timing/data': 0.0007065390236675739, 'timing/model': 1.6986216688528657, 'learning_rate/qwen_vl_interface': 6.5166104296889275e-06, 'learning_rate/action_model': 6.516610429688927e-05, 'epoch': 0.82}) +[2026-06-23 19:43:21,679][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.13178391754627228, 'train/grad_norm_pre_clip': 57.068843841552734, 'timing/data': 0.0006792980711907148, 'timing/model': 1.6975577441044152, 'learning_rate/qwen_vl_interface': 6.513675016441707e-06, 'learning_rate/action_model': 6.513675016441707e-05, 'epoch': 0.82}) +[2026-06-23 19:43:25,024][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.09465000033378601, 'train/grad_norm_pre_clip': 15.858015060424805, 'timing/data': 0.0006530890241265297, 'timing/model': 1.6961856801062822, 'learning_rate/qwen_vl_interface': 6.510739083745422e-06, 'learning_rate/action_model': 6.510739083745422e-05, 'epoch': 0.82}) +[2026-06-23 19:43:28,377][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.10892720520496368, 'train/grad_norm_pre_clip': 10.2368803024292, 'timing/data': 0.0006102181505411863, 'timing/model': 1.702261817874387, 'learning_rate/qwen_vl_interface': 6.507802632806919e-06, 'learning_rate/action_model': 6.507802632806919e-05, 'epoch': 0.83}) +[2026-06-23 19:43:31,725][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.09786154329776764, 'train/grad_norm_pre_clip': 11.19179916381836, 'timing/data': 0.0006114200223237276, 'timing/model': 1.6982650051359087, 'learning_rate/qwen_vl_interface': 6.5048656648332646e-06, 'learning_rate/action_model': 6.504865664833264e-05, 'epoch': 0.83}) +[2026-06-23 19:43:35,073][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.167796790599823, 'train/grad_norm_pre_clip': 26.251108169555664, 'timing/data': 0.0006166498642414808, 'timing/model': 1.700120863970369, 'learning_rate/qwen_vl_interface': 6.501928181031737e-06, 'learning_rate/action_model': 6.501928181031737e-05, 'epoch': 0.83}) +[2026-06-23 19:43:38,420][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.2264467179775238, 'train/grad_norm_pre_clip': 31.18205451965332, 'timing/data': 0.0006039869040250778, 'timing/model': 1.6970586099196225, 'learning_rate/qwen_vl_interface': 6.498990182609823e-06, 'learning_rate/action_model': 6.498990182609823e-05, 'epoch': 0.83}) +[2026-06-23 19:43:41,776][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.11326365172863007, 'train/grad_norm_pre_clip': 16.445646286010742, 'timing/data': 0.0005955197848379612, 'timing/model': 1.7016064329072833, 'learning_rate/qwen_vl_interface': 6.4960516707752255e-06, 'learning_rate/action_model': 6.496051670775226e-05, 'epoch': 0.83}) +[2026-06-23 19:43:45,124][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.2062157392501831, 'train/grad_norm_pre_clip': 14.786640167236328, 'timing/data': 0.0006143059581518173, 'timing/model': 1.702043142169714, 'learning_rate/qwen_vl_interface': 6.493112646735851e-06, 'learning_rate/action_model': 6.493112646735851e-05, 'epoch': 0.83}) +[2026-06-23 19:43:48,470][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.07880368083715439, 'train/grad_norm_pre_clip': 23.07830047607422, 'timing/data': 0.0009146200027316809, 'timing/model': 1.6968062568921596, 'learning_rate/qwen_vl_interface': 6.490173111699824e-06, 'learning_rate/action_model': 6.490173111699825e-05, 'epoch': 0.83}) +[2026-06-23 19:43:51,816][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.26969537138938904, 'train/grad_norm_pre_clip': 20.93761444091797, 'timing/data': 0.0006810310296714306, 'timing/model': 1.699331094045192, 'learning_rate/qwen_vl_interface': 6.487233066875479e-06, 'learning_rate/action_model': 6.487233066875479e-05, 'epoch': 0.83}) +[2026-06-23 19:43:55,169][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.1794504076242447, 'train/grad_norm_pre_clip': 20.97416877746582, 'timing/data': 0.0006145760416984558, 'timing/model': 1.7042861038353294, 'learning_rate/qwen_vl_interface': 6.484292513471352e-06, 'learning_rate/action_model': 6.484292513471352e-05, 'epoch': 0.83}) +[2026-06-23 19:43:58,517][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.1285388469696045, 'train/grad_norm_pre_clip': 28.287273406982422, 'timing/data': 0.0006259770598262548, 'timing/model': 1.6968674049712718, 'learning_rate/qwen_vl_interface': 6.4813514526961975e-06, 'learning_rate/action_model': 6.481351452696198e-05, 'epoch': 0.83}) +[2026-06-23 19:44:01,873][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.09898053854703903, 'train/grad_norm_pre_clip': 10.847081184387207, 'timing/data': 0.000763376010581851, 'timing/model': 1.7031151158735156, 'learning_rate/qwen_vl_interface': 6.478409885758972e-06, 'learning_rate/action_model': 6.478409885758971e-05, 'epoch': 0.83}) +[2026-06-23 19:44:05,220][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.14449134469032288, 'train/grad_norm_pre_clip': 68.12107849121094, 'timing/data': 0.0006175108719617128, 'timing/model': 1.7016733309719712, 'learning_rate/qwen_vl_interface': 6.475467813868844e-06, 'learning_rate/action_model': 6.475467813868844e-05, 'epoch': 0.83}) +[2026-06-23 19:44:08,574][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.12028271704912186, 'train/grad_norm_pre_clip': 38.88024139404297, 'timing/data': 0.0006028739735484123, 'timing/model': 1.7058658171445131, 'learning_rate/qwen_vl_interface': 6.472525238235187e-06, 'learning_rate/action_model': 6.472525238235186e-05, 'epoch': 0.83}) +[2026-06-23 19:44:12,219][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.0709175243973732, 'train/grad_norm_pre_clip': 38.23738479614258, 'timing/data': 0.0010703729931265116, 'timing/model': 1.699758955044672, 'learning_rate/qwen_vl_interface': 6.4695821600675834e-06, 'learning_rate/action_model': 6.469582160067583e-05, 'epoch': 0.83}) +[2026-06-23 19:44:15,573][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.15029895305633545, 'train/grad_norm_pre_clip': 18.0033016204834, 'timing/data': 0.0006736079230904579, 'timing/model': 1.7032565830741078, 'learning_rate/qwen_vl_interface': 6.466638580575824e-06, 'learning_rate/action_model': 6.466638580575824e-05, 'epoch': 0.83}) +[2026-06-23 19:44:18,921][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.17569176852703094, 'train/grad_norm_pre_clip': 15.997425079345703, 'timing/data': 0.0007217780221253633, 'timing/model': 1.6987022729590535, 'learning_rate/qwen_vl_interface': 6.463694500969898e-06, 'learning_rate/action_model': 6.463694500969898e-05, 'epoch': 0.83}) +[2026-06-23 19:44:22,273][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.11281296610832214, 'train/grad_norm_pre_clip': 34.07804870605469, 'timing/data': 0.0006005300674587488, 'timing/model': 1.7025067589711398, 'learning_rate/qwen_vl_interface': 6.460749922460008e-06, 'learning_rate/action_model': 6.460749922460008e-05, 'epoch': 0.83}) +[2026-06-23 19:44:25,627][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.15787744522094727, 'train/grad_norm_pre_clip': 20.471647262573242, 'timing/data': 0.0006136340089142323, 'timing/model': 1.7030403099488467, 'learning_rate/qwen_vl_interface': 6.457804846256559e-06, 'learning_rate/action_model': 6.457804846256559e-05, 'epoch': 0.83}) +[2026-06-23 19:44:28,989][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.13867919147014618, 'train/grad_norm_pre_clip': 15.221216201782227, 'timing/data': 0.0006318790838122368, 'timing/model': 1.707855517975986, 'learning_rate/qwen_vl_interface': 6.45485927357016e-06, 'learning_rate/action_model': 6.454859273570159e-05, 'epoch': 0.83}) +[2026-06-23 19:44:32,335][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.11846425384283066, 'train/grad_norm_pre_clip': 15.601765632629395, 'timing/data': 0.0006242350209504366, 'timing/model': 1.6960660370532423, 'learning_rate/qwen_vl_interface': 6.451913205611626e-06, 'learning_rate/action_model': 6.451913205611626e-05, 'epoch': 0.83}) +[2026-06-23 19:44:35,689][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.13611732423305511, 'train/grad_norm_pre_clip': 26.032222747802734, 'timing/data': 0.0005896189250051975, 'timing/model': 1.7065504880156368, 'learning_rate/qwen_vl_interface': 6.448966643591974e-06, 'learning_rate/action_model': 6.448966643591974e-05, 'epoch': 0.83}) +[2026-06-23 19:44:39,044][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.079440638422966, 'train/grad_norm_pre_clip': 40.05197525024414, 'timing/data': 0.0006187239196151495, 'timing/model': 1.703059340128675, 'learning_rate/qwen_vl_interface': 6.446019588722423e-06, 'learning_rate/action_model': 6.446019588722422e-05, 'epoch': 0.83}) +[2026-06-23 19:44:42,398][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.07978769391775131, 'train/grad_norm_pre_clip': 32.816165924072266, 'timing/data': 0.0006114300340414047, 'timing/model': 1.7032046208623797, 'learning_rate/qwen_vl_interface': 6.443072042214398e-06, 'learning_rate/action_model': 6.443072042214397e-05, 'epoch': 0.83}) +[2026-06-23 19:44:45,750][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.1269046813249588, 'train/grad_norm_pre_clip': 31.750843048095703, 'timing/data': 0.0006221600342541933, 'timing/model': 1.699477889109403, 'learning_rate/qwen_vl_interface': 6.440124005279521e-06, 'learning_rate/action_model': 6.440124005279521e-05, 'epoch': 0.83}) +[2026-06-23 19:44:49,104][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.12798148393630981, 'train/grad_norm_pre_clip': 11.754822731018066, 'timing/data': 0.0006191239226609468, 'timing/model': 1.7040994020644575, 'learning_rate/qwen_vl_interface': 6.43717547912962e-06, 'learning_rate/action_model': 6.437175479129621e-05, 'epoch': 0.83}) +[2026-06-23 19:44:52,461][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.1093810647726059, 'train/grad_norm_pre_clip': 18.069007873535156, 'timing/data': 0.0006182331126183271, 'timing/model': 1.704399747075513, 'learning_rate/qwen_vl_interface': 6.434226464976725e-06, 'learning_rate/action_model': 6.434226464976724e-05, 'epoch': 0.83}) +[2026-06-23 19:44:55,821][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.1617821604013443, 'train/grad_norm_pre_clip': 38.50584411621094, 'timing/data': 0.0008355120662599802, 'timing/model': 1.7113894540816545, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.84}) +[2026-06-23 19:44:59,306][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.12948405742645264, 'train/grad_norm_pre_clip': 36.607269287109375, 'timing/data': 0.0006103080231696367, 'timing/model': 1.70288160815835, 'learning_rate/qwen_vl_interface': 6.428326977511059e-06, 'learning_rate/action_model': 6.428326977511058e-05, 'epoch': 0.84}) +[2026-06-23 19:45:02,662][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.12318065017461777, 'train/grad_norm_pre_clip': 45.98176956176758, 'timing/data': 0.0006125019863247871, 'timing/model': 1.7037276478949934, 'learning_rate/qwen_vl_interface': 6.425376506623344e-06, 'learning_rate/action_model': 6.425376506623344e-05, 'epoch': 0.84}) +[2026-06-23 19:45:06,018][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.08803235739469528, 'train/grad_norm_pre_clip': 23.930471420288086, 'timing/data': 0.0007694370578974485, 'timing/model': 1.7056423500180244, 'learning_rate/qwen_vl_interface': 6.422425552582747e-06, 'learning_rate/action_model': 6.422425552582747e-05, 'epoch': 0.84}) +[2026-06-23 19:45:09,365][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.12034821510314941, 'train/grad_norm_pre_clip': 54.62788391113281, 'timing/data': 0.0006090160459280014, 'timing/model': 1.7008301259484142, 'learning_rate/qwen_vl_interface': 6.41947411660229e-06, 'learning_rate/action_model': 6.41947411660229e-05, 'epoch': 0.84}) +[2026-06-23 19:45:12,711][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.104535311460495, 'train/grad_norm_pre_clip': 64.16058349609375, 'timing/data': 0.0006008401978760958, 'timing/model': 1.6990160630084574, 'learning_rate/qwen_vl_interface': 6.4165221998951995e-06, 'learning_rate/action_model': 6.4165221998952e-05, 'epoch': 0.84}) +[2026-06-23 19:45:16,064][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.09147544205188751, 'train/grad_norm_pre_clip': 64.03411102294922, 'timing/data': 0.0007893750444054604, 'timing/model': 1.7030380391515791, 'learning_rate/qwen_vl_interface': 6.413569803674896e-06, 'learning_rate/action_model': 6.413569803674896e-05, 'epoch': 0.84}) +[2026-06-23 19:45:19,418][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.16048312187194824, 'train/grad_norm_pre_clip': 11.770523071289062, 'timing/data': 0.0006035950500518084, 'timing/model': 1.7020688641350716, 'learning_rate/qwen_vl_interface': 6.410616929154997e-06, 'learning_rate/action_model': 6.410616929154996e-05, 'epoch': 0.84}) +[2026-06-23 19:45:22,780][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.07946900278329849, 'train/grad_norm_pre_clip': 27.315616607666016, 'timing/data': 0.000597954960539937, 'timing/model': 1.7057587848976254, 'learning_rate/qwen_vl_interface': 6.40766357754932e-06, 'learning_rate/action_model': 6.407663577549319e-05, 'epoch': 0.84}) +[2026-06-23 19:45:26,133][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.179582878947258, 'train/grad_norm_pre_clip': 19.194238662719727, 'timing/data': 0.0006163688376545906, 'timing/model': 1.703176436945796, 'learning_rate/qwen_vl_interface': 6.404709750071875e-06, 'learning_rate/action_model': 6.404709750071874e-05, 'epoch': 0.84}) +[2026-06-23 19:45:29,493][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.1735984832048416, 'train/grad_norm_pre_clip': 25.976600646972656, 'timing/data': 0.0006946069188416004, 'timing/model': 1.703545852098614, 'learning_rate/qwen_vl_interface': 6.401755447936869e-06, 'learning_rate/action_model': 6.401755447936869e-05, 'epoch': 0.84}) +[2026-06-23 19:45:32,848][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.1826200634241104, 'train/grad_norm_pre_clip': 55.453861236572266, 'timing/data': 0.0007123099640011787, 'timing/model': 1.7056730529293418, 'learning_rate/qwen_vl_interface': 6.3988006723587036e-06, 'learning_rate/action_model': 6.398800672358703e-05, 'epoch': 0.84}) +[2026-06-23 19:45:36,201][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.22859448194503784, 'train/grad_norm_pre_clip': 97.74189758300781, 'timing/data': 0.0008512111380696297, 'timing/model': 1.706397966016084, 'learning_rate/qwen_vl_interface': 6.395845424551975e-06, 'learning_rate/action_model': 6.395845424551975e-05, 'epoch': 0.84}) +[2026-06-23 19:45:39,556][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.1952665001153946, 'train/grad_norm_pre_clip': 85.5710220336914, 'timing/data': 0.000619836151599884, 'timing/model': 1.7028456330299377, 'learning_rate/qwen_vl_interface': 6.392889705731475e-06, 'learning_rate/action_model': 6.392889705731476e-05, 'epoch': 0.84}) +[2026-06-23 19:45:42,913][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.11933664232492447, 'train/grad_norm_pre_clip': 85.03648376464844, 'timing/data': 0.0007600998505949974, 'timing/model': 1.705253091873601, 'learning_rate/qwen_vl_interface': 6.389933517112188e-06, 'learning_rate/action_model': 6.389933517112187e-05, 'epoch': 0.84}) +[2026-06-23 19:45:46,280][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.13816583156585693, 'train/grad_norm_pre_clip': 15.0599946975708, 'timing/data': 0.0010662258137017488, 'timing/model': 1.7133114500902593, 'learning_rate/qwen_vl_interface': 6.386976859909293e-06, 'learning_rate/action_model': 6.386976859909292e-05, 'epoch': 0.84}) +[2026-06-23 19:45:49,712][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.1797187179327011, 'train/grad_norm_pre_clip': 38.81877517700195, 'timing/data': 0.0006103881169110537, 'timing/model': 1.7032322301529348, 'learning_rate/qwen_vl_interface': 6.384019735338156e-06, 'learning_rate/action_model': 6.384019735338156e-05, 'epoch': 0.84}) +[2026-06-23 19:45:53,073][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.19234688580036163, 'train/grad_norm_pre_clip': 16.127981185913086, 'timing/data': 0.0007845649961382151, 'timing/model': 1.7082589049823582, 'learning_rate/qwen_vl_interface': 6.381062144614342e-06, 'learning_rate/action_model': 6.381062144614342e-05, 'epoch': 0.84}) +[2026-06-23 19:45:56,428][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.1029602587223053, 'train/grad_norm_pre_clip': 10.113958358764648, 'timing/data': 0.0006252860184758902, 'timing/model': 1.7037581498734653, 'learning_rate/qwen_vl_interface': 6.378104088953605e-06, 'learning_rate/action_model': 6.378104088953605e-05, 'epoch': 0.84}) +[2026-06-23 19:45:59,787][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.26835575699806213, 'train/grad_norm_pre_clip': 30.9400577545166, 'timing/data': 0.0006038951687514782, 'timing/model': 1.7079708341043442, 'learning_rate/qwen_vl_interface': 6.375145569571889e-06, 'learning_rate/action_model': 6.375145569571889e-05, 'epoch': 0.84}) +[2026-06-23 19:46:03,142][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.1627076119184494, 'train/grad_norm_pre_clip': 57.248592376708984, 'timing/data': 0.0006098970770835876, 'timing/model': 1.7019414228852838, 'learning_rate/qwen_vl_interface': 6.372186587685329e-06, 'learning_rate/action_model': 6.372186587685329e-05, 'epoch': 0.84}) +[2026-06-23 19:46:06,500][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.14866577088832855, 'train/grad_norm_pre_clip': 45.06874465942383, 'timing/data': 0.0006401739083230495, 'timing/model': 1.7032295861281455, 'learning_rate/qwen_vl_interface': 6.369227144510252e-06, 'learning_rate/action_model': 6.369227144510252e-05, 'epoch': 0.84}) +[2026-06-23 19:46:09,858][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.11786484718322754, 'train/grad_norm_pre_clip': 26.762432098388672, 'timing/data': 0.0008556700777262449, 'timing/model': 1.7045538781676441, 'learning_rate/qwen_vl_interface': 6.3662672412631736e-06, 'learning_rate/action_model': 6.366267241263173e-05, 'epoch': 0.84}) +[2026-06-23 19:46:13,211][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.1469966471195221, 'train/grad_norm_pre_clip': 10.613325119018555, 'timing/data': 0.000688524916768074, 'timing/model': 1.7047339770942926, 'learning_rate/qwen_vl_interface': 6.3633068791607954e-06, 'learning_rate/action_model': 6.363306879160795e-05, 'epoch': 0.84}) +[2026-06-23 19:46:16,567][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.17376038432121277, 'train/grad_norm_pre_clip': 13.932245254516602, 'timing/data': 0.0005953100044280291, 'timing/model': 1.7041632540058345, 'learning_rate/qwen_vl_interface': 6.360346059420014e-06, 'learning_rate/action_model': 6.360346059420014e-05, 'epoch': 0.84}) +[2026-06-23 19:46:19,930][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.08267560601234436, 'train/grad_norm_pre_clip': 29.061141967773438, 'timing/data': 0.0006103480700403452, 'timing/model': 1.7086529391817749, 'learning_rate/qwen_vl_interface': 6.357384783257909e-06, 'learning_rate/action_model': 6.357384783257908e-05, 'epoch': 0.84}) +[2026-06-23 19:46:23,280][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.22635847330093384, 'train/grad_norm_pre_clip': 81.85059356689453, 'timing/data': 0.0006148559041321278, 'timing/model': 1.7000006150919944, 'learning_rate/qwen_vl_interface': 6.354423051891748e-06, 'learning_rate/action_model': 6.354423051891749e-05, 'epoch': 0.85}) +[2026-06-23 19:46:26,628][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.16251790523529053, 'train/grad_norm_pre_clip': 68.15650939941406, 'timing/data': 0.0006058700382709503, 'timing/model': 1.7035054878797382, 'learning_rate/qwen_vl_interface': 6.35146086653899e-06, 'learning_rate/action_model': 6.35146086653899e-05, 'epoch': 0.85}) +[2026-06-23 19:46:29,972][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.1954561322927475, 'train/grad_norm_pre_clip': 87.28501892089844, 'timing/data': 0.0006609230767935514, 'timing/model': 1.7011410819832236, 'learning_rate/qwen_vl_interface': 6.348498228417273e-06, 'learning_rate/action_model': 6.348498228417273e-05, 'epoch': 0.85}) +[2026-06-23 19:46:33,324][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.20582419633865356, 'train/grad_norm_pre_clip': 105.44404602050781, 'timing/data': 0.0007039448246359825, 'timing/model': 1.7037517428398132, 'learning_rate/qwen_vl_interface': 6.34553513874443e-06, 'learning_rate/action_model': 6.34553513874443e-05, 'epoch': 0.85}) +[2026-06-23 19:46:36,764][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.10215511173009872, 'train/grad_norm_pre_clip': 37.453826904296875, 'timing/data': 0.0009095710702240467, 'timing/model': 1.7941213969606906, 'learning_rate/qwen_vl_interface': 6.342571598738473e-06, 'learning_rate/action_model': 6.342571598738472e-05, 'epoch': 0.85}) +[2026-06-23 19:46:40,110][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.12584413588047028, 'train/grad_norm_pre_clip': 45.3384895324707, 'timing/data': 0.001045677112415433, 'timing/model': 1.6998209210578352, 'learning_rate/qwen_vl_interface': 6.339607609617603e-06, 'learning_rate/action_model': 6.339607609617603e-05, 'epoch': 0.85}) +[2026-06-23 19:46:43,454][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.13265153765678406, 'train/grad_norm_pre_clip': 38.76327133178711, 'timing/data': 0.0007259459234774113, 'timing/model': 1.7006607421208173, 'learning_rate/qwen_vl_interface': 6.336643172600202e-06, 'learning_rate/action_model': 6.336643172600202e-05, 'epoch': 0.85}) +[2026-06-23 19:46:46,809][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.21855778992176056, 'train/grad_norm_pre_clip': 52.171024322509766, 'timing/data': 0.0005900000687688589, 'timing/model': 1.7027664040215313, 'learning_rate/qwen_vl_interface': 6.333678288904838e-06, 'learning_rate/action_model': 6.333678288904837e-05, 'epoch': 0.85}) +[2026-06-23 19:46:50,155][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.09978701174259186, 'train/grad_norm_pre_clip': 36.34783172607422, 'timing/data': 0.0006046369671821594, 'timing/model': 1.6997678950428963, 'learning_rate/qwen_vl_interface': 6.330712959750264e-06, 'learning_rate/action_model': 6.330712959750263e-05, 'epoch': 0.85}) +[2026-06-23 19:46:53,509][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.07981671392917633, 'train/grad_norm_pre_clip': 17.2893123626709, 'timing/data': 0.0006185430102050304, 'timing/model': 1.7017403759527951, 'learning_rate/qwen_vl_interface': 6.3277471863554145e-06, 'learning_rate/action_model': 6.327747186355414e-05, 'epoch': 0.85}) +[2026-06-23 19:46:56,863][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.23917706310749054, 'train/grad_norm_pre_clip': 37.94620895385742, 'timing/data': 0.0006004900205880404, 'timing/model': 1.7058981279842556, 'learning_rate/qwen_vl_interface': 6.324780969939406e-06, 'learning_rate/action_model': 6.324780969939405e-05, 'epoch': 0.85}) +[2026-06-23 19:47:00,209][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.07234472781419754, 'train/grad_norm_pre_clip': 18.612327575683594, 'timing/data': 0.0006091659888625145, 'timing/model': 1.698406559182331, 'learning_rate/qwen_vl_interface': 6.32181431172154e-06, 'learning_rate/action_model': 6.32181431172154e-05, 'epoch': 0.85}) +[2026-06-23 19:47:03,595][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.1040581613779068, 'train/grad_norm_pre_clip': 27.105070114135742, 'timing/data': 0.0006491108797490597, 'timing/model': 1.7036452030297369, 'learning_rate/qwen_vl_interface': 6.318847212921295e-06, 'learning_rate/action_model': 6.318847212921295e-05, 'epoch': 0.85}) +[2026-06-23 19:47:06,942][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.19530396163463593, 'train/grad_norm_pre_clip': 16.023666381835938, 'timing/data': 0.0006071520037949085, 'timing/model': 1.6980078511405736, 'learning_rate/qwen_vl_interface': 6.3158796747583364e-06, 'learning_rate/action_model': 6.315879674758336e-05, 'epoch': 0.85}) +[2026-06-23 19:47:10,288][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.13788935542106628, 'train/grad_norm_pre_clip': 25.295059204101562, 'timing/data': 0.0010583100374788046, 'timing/model': 1.7005259688012302, 'learning_rate/qwen_vl_interface': 6.312911698452503e-06, 'learning_rate/action_model': 6.312911698452502e-05, 'epoch': 0.85}) +[2026-06-23 19:47:13,635][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.12811648845672607, 'train/grad_norm_pre_clip': 20.417882919311523, 'timing/data': 0.0006022530142217875, 'timing/model': 1.6991475878749043, 'learning_rate/qwen_vl_interface': 6.309943285223824e-06, 'learning_rate/action_model': 6.309943285223824e-05, 'epoch': 0.85}) +[2026-06-23 19:47:16,979][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2200 +[2026-06-23 19:48:50,989][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2200 +[2026-06-23 19:48:50,990][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2200 +[2026-06-23 19:50:02,276][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2200 +[2026-06-23 19:50:02,277][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.17091473937034607, 'train/grad_norm_pre_clip': 18.694843292236328, 'eval/action_loss/seconds': 94.008147056913, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07568588852882385, 'eval/latency_2/loss': 0.07568588852882385, 'eval/flappy/loss': 0.07568588852882385, 'eval/flappy/latency_2/loss': 0.07568588852882385, 'eval/flap_precision': 0.9272237196765498, 'eval/flap_recall': 0.9284750337381916, 'eval/flap_f1': 0.927848954821308, 'eval/flap_rate_model': 0.2061111111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9278489548213081, 'eval/latency_2/flap_precision': 0.9272237196765498, 'eval/latency_2/flap_recall': 0.9284750337381916, 'eval/latency_2/flap_f1': 0.927848954821308, 'eval/latency_2/flap_rate_model': 0.2061111111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9278489548213081, 'eval/action_classification/data_wait_seconds': 0.9243382215499878, 'eval/action_classification/predict_seconds': 70.22804260253906, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0007293110247701406, 'timing/model': 1.695453210035339, 'learning_rate/qwen_vl_interface': 6.306974436292493e-06, 'learning_rate/action_model': 6.306974436292494e-05, 'epoch': 0.85}) +[2026-06-23 19:50:05,593][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.06183173507452011, 'train/grad_norm_pre_clip': 31.045148849487305, 'timing/data': 0.0006056991405785084, 'timing/model': 1.6834909250028431, 'learning_rate/qwen_vl_interface': 6.304005152878902e-06, 'learning_rate/action_model': 6.304005152878901e-05, 'epoch': 0.85}) +[2026-06-23 19:50:08,920][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.16365951299667358, 'train/grad_norm_pre_clip': 29.39765167236328, 'timing/data': 0.0008493470959365368, 'timing/model': 1.6880028250161558, 'learning_rate/qwen_vl_interface': 6.301035436203602e-06, 'learning_rate/action_model': 6.301035436203602e-05, 'epoch': 0.85}) +[2026-06-23 19:50:12,246][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.1740843802690506, 'train/grad_norm_pre_clip': 67.01055908203125, 'timing/data': 0.0006214079912751913, 'timing/model': 1.6906863860785961, 'learning_rate/qwen_vl_interface': 6.298065287487337e-06, 'learning_rate/action_model': 6.298065287487337e-05, 'epoch': 0.85}) +[2026-06-23 19:50:15,575][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.177851602435112, 'train/grad_norm_pre_clip': 115.74794006347656, 'timing/data': 0.0006724451668560505, 'timing/model': 1.6913814670406282, 'learning_rate/qwen_vl_interface': 6.295094707951019e-06, 'learning_rate/action_model': 6.295094707951019e-05, 'epoch': 0.85}) +[2026-06-23 19:50:18,909][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.16811062395572662, 'train/grad_norm_pre_clip': 11.586556434631348, 'timing/data': 0.0008466229774057865, 'timing/model': 1.694434944074601, 'learning_rate/qwen_vl_interface': 6.292123698815745e-06, 'learning_rate/action_model': 6.292123698815745e-05, 'epoch': 0.85}) +[2026-06-23 19:50:22,243][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.26843589544296265, 'train/grad_norm_pre_clip': 42.53206253051758, 'timing/data': 0.0006231020670384169, 'timing/model': 1.6912205929402262, 'learning_rate/qwen_vl_interface': 6.289152261302783e-06, 'learning_rate/action_model': 6.289152261302784e-05, 'epoch': 0.85}) +[2026-06-23 19:50:25,572][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.12537139654159546, 'train/grad_norm_pre_clip': 24.799772262573242, 'timing/data': 0.00067130196839571, 'timing/model': 1.6873587539885193, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.85}) +[2026-06-23 19:50:28,898][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.11352232098579407, 'train/grad_norm_pre_clip': 15.511346817016602, 'timing/data': 0.0007118789944797754, 'timing/model': 1.6902251387946308, 'learning_rate/qwen_vl_interface': 6.283208106029753e-06, 'learning_rate/action_model': 6.283208106029752e-05, 'epoch': 0.85}) +[2026-06-23 19:50:32,218][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.09794850647449493, 'train/grad_norm_pre_clip': 15.127545356750488, 'timing/data': 0.0006120309699326754, 'timing/model': 1.6823057199362665, 'learning_rate/qwen_vl_interface': 6.280235390713102e-06, 'learning_rate/action_model': 6.280235390713102e-05, 'epoch': 0.85}) +[2026-06-23 19:50:35,857][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.06245984137058258, 'train/grad_norm_pre_clip': 17.29146385192871, 'timing/data': 0.0006243449170142412, 'timing/model': 1.9996798587962985, 'learning_rate/qwen_vl_interface': 6.277262251905597e-06, 'learning_rate/action_model': 6.277262251905597e-05, 'epoch': 0.86}) +[2026-06-23 19:50:39,193][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.06992237269878387, 'train/grad_norm_pre_clip': 16.79460906982422, 'timing/data': 0.0006168498657643795, 'timing/model': 1.69348197709769, 'learning_rate/qwen_vl_interface': 6.274288690829385e-06, 'learning_rate/action_model': 6.274288690829384e-05, 'epoch': 0.86}) +[2026-06-23 19:50:42,521][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.0688709244132042, 'train/grad_norm_pre_clip': 19.555509567260742, 'timing/data': 0.0006053780671209097, 'timing/model': 1.6859319030772895, 'learning_rate/qwen_vl_interface': 6.271314708706781e-06, 'learning_rate/action_model': 6.271314708706782e-05, 'epoch': 0.86}) +[2026-06-23 19:50:45,860][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.08354348689317703, 'train/grad_norm_pre_clip': 29.453102111816406, 'timing/data': 0.0006180619820952415, 'timing/model': 1.6950169019401073, 'learning_rate/qwen_vl_interface': 6.26834030676028e-06, 'learning_rate/action_model': 6.26834030676028e-05, 'epoch': 0.86}) +[2026-06-23 19:50:49,193][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.08169815689325333, 'train/grad_norm_pre_clip': 33.06544494628906, 'timing/data': 0.0006262180395424366, 'timing/model': 1.6903603598475456, 'learning_rate/qwen_vl_interface': 6.265365486212545e-06, 'learning_rate/action_model': 6.265365486212545e-05, 'epoch': 0.86}) +[2026-06-23 19:50:52,522][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.08610312640666962, 'train/grad_norm_pre_clip': 41.56523895263672, 'timing/data': 0.0006292629987001419, 'timing/model': 1.6881039610598236, 'learning_rate/qwen_vl_interface': 6.26239024828641e-06, 'learning_rate/action_model': 6.26239024828641e-05, 'epoch': 0.86}) +[2026-06-23 19:50:55,847][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.16836994886398315, 'train/grad_norm_pre_clip': 21.048315048217773, 'timing/data': 0.0006105590146034956, 'timing/model': 1.683793599018827, 'learning_rate/qwen_vl_interface': 6.259414594204888e-06, 'learning_rate/action_model': 6.259414594204888e-05, 'epoch': 0.86}) +[2026-06-23 19:50:59,170][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.09646439552307129, 'train/grad_norm_pre_clip': 16.915637969970703, 'timing/data': 0.0007579850498586893, 'timing/model': 1.6836050099227577, 'learning_rate/qwen_vl_interface': 6.256438525191155e-06, 'learning_rate/action_model': 6.256438525191156e-05, 'epoch': 0.86}) +[2026-06-23 19:51:02,502][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.2686617970466614, 'train/grad_norm_pre_clip': 42.679412841796875, 'timing/data': 0.0006185038946568966, 'timing/model': 1.6934898539911956, 'learning_rate/qwen_vl_interface': 6.253462042468561e-06, 'learning_rate/action_model': 6.25346204246856e-05, 'epoch': 0.86}) +[2026-06-23 19:51:05,845][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.19967021048069, 'train/grad_norm_pre_clip': 59.32110595703125, 'timing/data': 0.0006671550218015909, 'timing/model': 1.6992592890746891, 'learning_rate/qwen_vl_interface': 6.2504851472606246e-06, 'learning_rate/action_model': 6.250485147260625e-05, 'epoch': 0.86}) +[2026-06-23 19:51:09,177][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.262372761964798, 'train/grad_norm_pre_clip': 21.450788497924805, 'timing/data': 0.0006164391525089741, 'timing/model': 1.6910201490391046, 'learning_rate/qwen_vl_interface': 6.247507840791035e-06, 'learning_rate/action_model': 6.247507840791035e-05, 'epoch': 0.86}) +[2026-06-23 19:51:12,514][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.19229169189929962, 'train/grad_norm_pre_clip': 17.34938621520996, 'timing/data': 0.0006247649434953928, 'timing/model': 1.6965611439663917, 'learning_rate/qwen_vl_interface': 6.244530124283651e-06, 'learning_rate/action_model': 6.24453012428365e-05, 'epoch': 0.86}) +[2026-06-23 19:51:15,854][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.11766278743743896, 'train/grad_norm_pre_clip': 50.23994445800781, 'timing/data': 0.0006611531134694815, 'timing/model': 1.6944110749755055, 'learning_rate/qwen_vl_interface': 6.2415519989625016e-06, 'learning_rate/action_model': 6.241551998962501e-05, 'epoch': 0.86}) +[2026-06-23 19:51:19,199][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.1331959366798401, 'train/grad_norm_pre_clip': 22.212739944458008, 'timing/data': 0.0006201060023158789, 'timing/model': 1.6987115470692515, 'learning_rate/qwen_vl_interface': 6.238573466051781e-06, 'learning_rate/action_model': 6.23857346605178e-05, 'epoch': 0.86}) +[2026-06-23 19:51:22,524][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.16718782484531403, 'train/grad_norm_pre_clip': 23.065387725830078, 'timing/data': 0.0006886760238558054, 'timing/model': 1.6910433610901237, 'learning_rate/qwen_vl_interface': 6.2355945267758465e-06, 'learning_rate/action_model': 6.235594526775846e-05, 'epoch': 0.86}) +[2026-06-23 19:51:26,091][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.10392546653747559, 'train/grad_norm_pre_clip': 62.076271057128906, 'timing/data': 0.0007112170569598675, 'timing/model': 1.9301077139098197, 'learning_rate/qwen_vl_interface': 6.232615182359231e-06, 'learning_rate/action_model': 6.232615182359231e-05, 'epoch': 0.86}) +[2026-06-23 19:51:29,424][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.1986652910709381, 'train/grad_norm_pre_clip': 53.52042770385742, 'timing/data': 0.00059875613078475, 'timing/model': 1.6952716729138047, 'learning_rate/qwen_vl_interface': 6.229635434026634e-06, 'learning_rate/action_model': 6.229635434026634e-05, 'epoch': 0.86}) +[2026-06-23 19:51:32,758][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.1118050068616867, 'train/grad_norm_pre_clip': 58.55506896972656, 'timing/data': 0.0006193749140948057, 'timing/model': 1.6934977918863297, 'learning_rate/qwen_vl_interface': 6.226655283002912e-06, 'learning_rate/action_model': 6.226655283002911e-05, 'epoch': 0.86}) +[2026-06-23 19:51:36,089][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.1131119579076767, 'train/grad_norm_pre_clip': 61.673980712890625, 'timing/data': 0.0006188640836626291, 'timing/model': 1.6928074781317264, 'learning_rate/qwen_vl_interface': 6.223674730513096e-06, 'learning_rate/action_model': 6.223674730513096e-05, 'epoch': 0.86}) +[2026-06-23 19:51:39,426][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.14564745128154755, 'train/grad_norm_pre_clip': 23.423297882080078, 'timing/data': 0.0006656618788838387, 'timing/model': 1.6973444358445704, 'learning_rate/qwen_vl_interface': 6.220693777782375e-06, 'learning_rate/action_model': 6.220693777782375e-05, 'epoch': 0.86}) +[2026-06-23 19:51:42,758][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.11972395330667496, 'train/grad_norm_pre_clip': 17.043203353881836, 'timing/data': 0.000922375125810504, 'timing/model': 1.6883122250437737, 'learning_rate/qwen_vl_interface': 6.21771242603611e-06, 'learning_rate/action_model': 6.21771242603611e-05, 'epoch': 0.86}) +[2026-06-23 19:51:46,094][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.11010494828224182, 'train/grad_norm_pre_clip': 35.23529815673828, 'timing/data': 0.000609265873208642, 'timing/model': 1.6924490300007164, 'learning_rate/qwen_vl_interface': 6.21473067649982e-06, 'learning_rate/action_model': 6.21473067649982e-05, 'epoch': 0.86}) +[2026-06-23 19:51:49,432][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.25337183475494385, 'train/grad_norm_pre_clip': 53.41905975341797, 'timing/data': 0.0007116589695215225, 'timing/model': 1.695871361065656, 'learning_rate/qwen_vl_interface': 6.211748530399189e-06, 'learning_rate/action_model': 6.211748530399189e-05, 'epoch': 0.86}) +[2026-06-23 19:51:52,758][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.14339026808738708, 'train/grad_norm_pre_clip': 51.220008850097656, 'timing/data': 0.0006061401218175888, 'timing/model': 1.6888800279702991, 'learning_rate/qwen_vl_interface': 6.208765988960067e-06, 'learning_rate/action_model': 6.208765988960066e-05, 'epoch': 0.86}) +[2026-06-23 19:51:56,095][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.1197865903377533, 'train/grad_norm_pre_clip': 34.197696685791016, 'timing/data': 0.000605529174208641, 'timing/model': 1.6938444799743593, 'learning_rate/qwen_vl_interface': 6.2057830534084615e-06, 'learning_rate/action_model': 6.205783053408461e-05, 'epoch': 0.86}) +[2026-06-23 19:51:59,434][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.21560367941856384, 'train/grad_norm_pre_clip': 18.766752243041992, 'timing/data': 0.0006370379123836756, 'timing/model': 1.691306680906564, 'learning_rate/qwen_vl_interface': 6.202799724970546e-06, 'learning_rate/action_model': 6.202799724970546e-05, 'epoch': 0.86}) +[2026-06-23 19:52:02,769][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.25315019488334656, 'train/grad_norm_pre_clip': 57.456214904785156, 'timing/data': 0.0006192540749907494, 'timing/model': 1.6970125469379127, 'learning_rate/qwen_vl_interface': 6.199816004872655e-06, 'learning_rate/action_model': 6.199816004872655e-05, 'epoch': 0.87}) +[2026-06-23 19:52:06,111][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.1866448074579239, 'train/grad_norm_pre_clip': 132.3986053466797, 'timing/data': 0.0005998180713504553, 'timing/model': 1.7048624448943883, 'learning_rate/qwen_vl_interface': 6.196831894341281e-06, 'learning_rate/action_model': 6.19683189434128e-05, 'epoch': 0.87}) +[2026-06-23 19:52:09,453][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.1865139752626419, 'train/grad_norm_pre_clip': 59.84102249145508, 'timing/data': 0.0008077290840446949, 'timing/model': 1.6996818480547518, 'learning_rate/qwen_vl_interface': 6.1938473946030795e-06, 'learning_rate/action_model': 6.193847394603079e-05, 'epoch': 0.87}) +[2026-06-23 19:52:12,794][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.15024380385875702, 'train/grad_norm_pre_clip': 63.8054084777832, 'timing/data': 0.0006164589431136847, 'timing/model': 1.6988904250320047, 'learning_rate/qwen_vl_interface': 6.1908625068848695e-06, 'learning_rate/action_model': 6.19086250688487e-05, 'epoch': 0.87}) +[2026-06-23 19:52:16,219][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.10981304198503494, 'train/grad_norm_pre_clip': 38.093746185302734, 'timing/data': 0.0008280968759208918, 'timing/model': 1.7005719819571823, 'learning_rate/qwen_vl_interface': 6.187877232413619e-06, 'learning_rate/action_model': 6.18787723241362e-05, 'epoch': 0.87}) +[2026-06-23 19:52:19,561][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.21611490845680237, 'train/grad_norm_pre_clip': 39.63582992553711, 'timing/data': 0.0006179220508784056, 'timing/model': 1.6979540849570185, 'learning_rate/qwen_vl_interface': 6.184891572416467e-06, 'learning_rate/action_model': 6.184891572416466e-05, 'epoch': 0.87}) +[2026-06-23 19:52:22,905][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.15547938644886017, 'train/grad_norm_pre_clip': 19.839988708496094, 'timing/data': 0.0007067089900374413, 'timing/model': 1.694600532995537, 'learning_rate/qwen_vl_interface': 6.181905528120703e-06, 'learning_rate/action_model': 6.181905528120703e-05, 'epoch': 0.87}) +[2026-06-23 19:52:26,240][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.10996578633785248, 'train/grad_norm_pre_clip': 32.573814392089844, 'timing/data': 0.0006856098771095276, 'timing/model': 1.692768142791465, 'learning_rate/qwen_vl_interface': 6.178919100753778e-06, 'learning_rate/action_model': 6.178919100753778e-05, 'epoch': 0.87}) +[2026-06-23 19:52:29,580][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.10569511353969574, 'train/grad_norm_pre_clip': 24.633237838745117, 'timing/data': 0.0006968800444155931, 'timing/model': 1.6963304670061916, 'learning_rate/qwen_vl_interface': 6.175932291543299e-06, 'learning_rate/action_model': 6.175932291543298e-05, 'epoch': 0.87}) +[2026-06-23 19:52:32,924][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.2340298593044281, 'train/grad_norm_pre_clip': 51.31344985961914, 'timing/data': 0.0006999769248068333, 'timing/model': 1.6989638658706099, 'learning_rate/qwen_vl_interface': 6.172945101717029e-06, 'learning_rate/action_model': 6.17294510171703e-05, 'epoch': 0.87}) +[2026-06-23 19:52:36,264][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.07307077944278717, 'train/grad_norm_pre_clip': 25.14228057861328, 'timing/data': 0.0007015201263129711, 'timing/model': 1.6926959538832307, 'learning_rate/qwen_vl_interface': 6.169957532502892e-06, 'learning_rate/action_model': 6.169957532502892e-05, 'epoch': 0.87}) +[2026-06-23 19:52:39,606][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.1995871663093567, 'train/grad_norm_pre_clip': 27.10055923461914, 'timing/data': 0.0006166200619190931, 'timing/model': 1.6974418240133673, 'learning_rate/qwen_vl_interface': 6.166969585128962e-06, 'learning_rate/action_model': 6.166969585128962e-05, 'epoch': 0.87}) +[2026-06-23 19:52:42,946][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.14624646306037903, 'train/grad_norm_pre_clip': 28.862197875976562, 'timing/data': 0.000610909191891551, 'timing/model': 1.6948464079760015, 'learning_rate/qwen_vl_interface': 6.163981260823473e-06, 'learning_rate/action_model': 6.163981260823473e-05, 'epoch': 0.87}) +[2026-06-23 19:52:46,285][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.09103233367204666, 'train/grad_norm_pre_clip': 48.791290283203125, 'timing/data': 0.0006175711750984192, 'timing/model': 1.693611121037975, 'learning_rate/qwen_vl_interface': 6.16099256081481e-06, 'learning_rate/action_model': 6.16099256081481e-05, 'epoch': 0.87}) +[2026-06-23 19:53:34,465][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2250 +[2026-06-23 19:53:34,466][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.2657443881034851, 'train/grad_norm_pre_clip': 19.591615676879883, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 31.799999999999994, 'rl_games_eval/mid_train/mean_length': 275.0, 'rl_games_eval/mid_train/std_reward': 30.682894257224078, 'rl_games_eval/mid_train/std_length': 247.40169764979382, 'rl_games_eval/mid_train/macro_mean_reward': 31.799999999999994, 'rl_games_eval/mid_train/macro_mean_length': 275.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 31.799999999999994, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 275.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 30.682894257224078, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 247.40169764979382, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 31.799999999999994, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006145858205854893, 'timing/model': 1.6954719659406692, 'learning_rate/qwen_vl_interface': 6.15800348633152e-06, 'learning_rate/action_model': 6.15800348633152e-05, 'epoch': 0.87}) +[2026-06-23 19:53:37,760][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.04642794653773308, 'train/grad_norm_pre_clip': 38.175872802734375, 'timing/data': 0.0006153769791126251, 'timing/model': 1.6726412430871278, 'learning_rate/qwen_vl_interface': 6.155014038602293e-06, 'learning_rate/action_model': 6.155014038602292e-05, 'epoch': 0.87}) +[2026-06-23 19:53:41,159][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.19351741671562195, 'train/grad_norm_pre_clip': 34.42917251586914, 'timing/data': 0.0006816918030381203, 'timing/model': 1.6727354770991951, 'learning_rate/qwen_vl_interface': 6.152024218855979e-06, 'learning_rate/action_model': 6.15202421885598e-05, 'epoch': 0.87}) +[2026-06-23 19:53:44,463][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.07772558182477951, 'train/grad_norm_pre_clip': 21.684606552124023, 'timing/data': 0.0006283819675445557, 'timing/model': 1.6765619090292603, 'learning_rate/qwen_vl_interface': 6.149034028321583e-06, 'learning_rate/action_model': 6.149034028321583e-05, 'epoch': 0.87}) +[2026-06-23 19:53:47,764][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.18146979808807373, 'train/grad_norm_pre_clip': 76.16258239746094, 'timing/data': 0.0006092358380556107, 'timing/model': 1.674197027925402, 'learning_rate/qwen_vl_interface': 6.1460434682282566e-06, 'learning_rate/action_model': 6.146043468228256e-05, 'epoch': 0.87}) +[2026-06-23 19:53:51,079][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.12173721939325333, 'train/grad_norm_pre_clip': 58.076019287109375, 'timing/data': 0.0006708111613988876, 'timing/model': 1.6867321860045195, 'learning_rate/qwen_vl_interface': 6.143052539805308e-06, 'learning_rate/action_model': 6.143052539805308e-05, 'epoch': 0.87}) +[2026-06-23 19:53:54,399][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.12838564813137054, 'train/grad_norm_pre_clip': 49.32187271118164, 'timing/data': 0.0006053580436855555, 'timing/model': 1.6881964360363781, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.87}) +[2026-06-23 19:53:57,717][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.13168275356292725, 'train/grad_norm_pre_clip': 55.83956527709961, 'timing/data': 0.0006353249773383141, 'timing/model': 1.6847066190093756, 'learning_rate/qwen_vl_interface': 6.137069582888521e-06, 'learning_rate/action_model': 6.137069582888521e-05, 'epoch': 0.87}) +[2026-06-23 19:54:01,030][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.14065898954868317, 'train/grad_norm_pre_clip': 37.3238525390625, 'timing/data': 0.0006172009743750095, 'timing/model': 1.6826931699179113, 'learning_rate/qwen_vl_interface': 6.13407755685405e-06, 'learning_rate/action_model': 6.13407755685405e-05, 'epoch': 0.87}) +[2026-06-23 19:54:04,351][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.07929012924432755, 'train/grad_norm_pre_clip': 33.373374938964844, 'timing/data': 0.0009423429146409035, 'timing/model': 1.688171047018841, 'learning_rate/qwen_vl_interface': 6.131085167408692e-06, 'learning_rate/action_model': 6.131085167408692e-05, 'epoch': 0.87}) +[2026-06-23 19:54:07,674][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.10934659093618393, 'train/grad_norm_pre_clip': 25.559165954589844, 'timing/data': 0.0006954781711101532, 'timing/model': 1.6868342589586973, 'learning_rate/qwen_vl_interface': 6.128092415782502e-06, 'learning_rate/action_model': 6.128092415782501e-05, 'epoch': 0.87}) +[2026-06-23 19:54:10,999][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.07546261698007584, 'train/grad_norm_pre_clip': 21.207561492919922, 'timing/data': 0.0008374850731343031, 'timing/model': 1.6910965200513601, 'learning_rate/qwen_vl_interface': 6.125099303205688e-06, 'learning_rate/action_model': 6.125099303205688e-05, 'epoch': 0.88}) +[2026-06-23 19:54:14,323][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.08629026263952255, 'train/grad_norm_pre_clip': 18.92876434326172, 'timing/data': 0.000681963050737977, 'timing/model': 1.6906175760086626, 'learning_rate/qwen_vl_interface': 6.122105830908607e-06, 'learning_rate/action_model': 6.122105830908606e-05, 'epoch': 0.88}) +[2026-06-23 19:54:17,645][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.11507780104875565, 'train/grad_norm_pre_clip': 26.422945022583008, 'timing/data': 0.0007038728799670935, 'timing/model': 1.6849161309655756, 'learning_rate/qwen_vl_interface': 6.11911200012176e-06, 'learning_rate/action_model': 6.11911200012176e-05, 'epoch': 0.88}) +[2026-06-23 19:54:20,968][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.06268547475337982, 'train/grad_norm_pre_clip': 14.557616233825684, 'timing/data': 0.0007176189683377743, 'timing/model': 1.6865595181006938, 'learning_rate/qwen_vl_interface': 6.116117812075799e-06, 'learning_rate/action_model': 6.116117812075799e-05, 'epoch': 0.88}) +[2026-06-23 19:54:24,292][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.14137567579746246, 'train/grad_norm_pre_clip': 25.894426345825195, 'timing/data': 0.0006079538725316525, 'timing/model': 1.690075194928795, 'learning_rate/qwen_vl_interface': 6.113123268001524e-06, 'learning_rate/action_model': 6.113123268001524e-05, 'epoch': 0.88}) +[2026-06-23 19:54:27,617][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.16781024634838104, 'train/grad_norm_pre_clip': 38.15615463256836, 'timing/data': 0.0006137241143733263, 'timing/model': 1.6858165839221328, 'learning_rate/qwen_vl_interface': 6.110128369129873e-06, 'learning_rate/action_model': 6.110128369129873e-05, 'epoch': 0.88}) +[2026-06-23 19:54:31,248][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.16089019179344177, 'train/grad_norm_pre_clip': 50.89623260498047, 'timing/data': 0.0007030819542706013, 'timing/model': 1.690152764087543, 'learning_rate/qwen_vl_interface': 6.107133116691943e-06, 'learning_rate/action_model': 6.107133116691942e-05, 'epoch': 0.88}) +[2026-06-23 19:54:34,569][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.11334732919931412, 'train/grad_norm_pre_clip': 26.500808715820312, 'timing/data': 0.000617241021245718, 'timing/model': 1.6838040919974446, 'learning_rate/qwen_vl_interface': 6.104137511918963e-06, 'learning_rate/action_model': 6.104137511918963e-05, 'epoch': 0.88}) +[2026-06-23 19:54:37,907][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.14193253219127655, 'train/grad_norm_pre_clip': 20.358928680419922, 'timing/data': 0.000627520028501749, 'timing/model': 1.7014264569152147, 'learning_rate/qwen_vl_interface': 6.1011415560423155e-06, 'learning_rate/action_model': 6.101141556042316e-05, 'epoch': 0.88}) +[2026-06-23 19:54:41,238][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.18468095362186432, 'train/grad_norm_pre_clip': 22.164783477783203, 'timing/data': 0.0005995368119329214, 'timing/model': 1.6911822399124503, 'learning_rate/qwen_vl_interface': 6.098145250293525e-06, 'learning_rate/action_model': 6.0981452502935246e-05, 'epoch': 0.88}) +[2026-06-23 19:54:44,581][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.16999712586402893, 'train/grad_norm_pre_clip': 47.434391021728516, 'timing/data': 0.0006082439795136452, 'timing/model': 1.6985981909092516, 'learning_rate/qwen_vl_interface': 6.09514859590426e-06, 'learning_rate/action_model': 6.09514859590426e-05, 'epoch': 0.88}) +[2026-06-23 19:54:47,919][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.12223286926746368, 'train/grad_norm_pre_clip': 26.961305618286133, 'timing/data': 0.0006591500714421272, 'timing/model': 1.6938984910957515, 'learning_rate/qwen_vl_interface': 6.09215159410633e-06, 'learning_rate/action_model': 6.09215159410633e-05, 'epoch': 0.88}) +[2026-06-23 19:54:51,254][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.26671040058135986, 'train/grad_norm_pre_clip': 51.927467346191406, 'timing/data': 0.0007133819162845612, 'timing/model': 1.695778112160042, 'learning_rate/qwen_vl_interface': 6.089154246131692e-06, 'learning_rate/action_model': 6.089154246131692e-05, 'epoch': 0.88}) +[2026-06-23 19:54:54,582][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.09731562435626984, 'train/grad_norm_pre_clip': 23.3985595703125, 'timing/data': 0.0006034551188349724, 'timing/model': 1.6891647819429636, 'learning_rate/qwen_vl_interface': 6.08615655321244e-06, 'learning_rate/action_model': 6.08615655321244e-05, 'epoch': 0.88}) +[2026-06-23 19:54:57,918][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.11941450089216232, 'train/grad_norm_pre_clip': 58.70680618286133, 'timing/data': 0.0007059581112116575, 'timing/model': 1.69322249898687, 'learning_rate/qwen_vl_interface': 6.083158516580814e-06, 'learning_rate/action_model': 6.083158516580814e-05, 'epoch': 0.88}) +[2026-06-23 19:55:01,249][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.24795225262641907, 'train/grad_norm_pre_clip': 59.44681930541992, 'timing/data': 0.0006978719029575586, 'timing/model': 1.6913844689261168, 'learning_rate/qwen_vl_interface': 6.080160137469192e-06, 'learning_rate/action_model': 6.0801601374691916e-05, 'epoch': 0.88}) +[2026-06-23 19:55:04,587][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.2014225870370865, 'train/grad_norm_pre_clip': 47.35640335083008, 'timing/data': 0.0006600909400731325, 'timing/model': 1.690857654903084, 'learning_rate/qwen_vl_interface': 6.077161417110097e-06, 'learning_rate/action_model': 6.0771614171100965e-05, 'epoch': 0.88}) +[2026-06-23 19:55:07,910][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.14128190279006958, 'train/grad_norm_pre_clip': 13.15817642211914, 'timing/data': 0.0006250750739127398, 'timing/model': 1.6856955050025135, 'learning_rate/qwen_vl_interface': 6.0741623567361864e-06, 'learning_rate/action_model': 6.074162356736186e-05, 'epoch': 0.88}) +[2026-06-23 19:55:11,242][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.10215859860181808, 'train/grad_norm_pre_clip': 11.618548393249512, 'timing/data': 0.000609837006777525, 'timing/model': 1.6914811509195715, 'learning_rate/qwen_vl_interface': 6.071162957580262e-06, 'learning_rate/action_model': 6.071162957580262e-05, 'epoch': 0.88}) +[2026-06-23 19:55:14,571][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.07874350249767303, 'train/grad_norm_pre_clip': 17.736196517944336, 'timing/data': 0.0007050859276205301, 'timing/model': 1.6890017259866, 'learning_rate/qwen_vl_interface': 6.068163220875266e-06, 'learning_rate/action_model': 6.068163220875266e-05, 'epoch': 0.88}) +[2026-06-23 19:55:17,912][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.07318046689033508, 'train/grad_norm_pre_clip': 14.288949966430664, 'timing/data': 0.000964432954788208, 'timing/model': 1.698269851040095, 'learning_rate/qwen_vl_interface': 6.065163147854273e-06, 'learning_rate/action_model': 6.065163147854273e-05, 'epoch': 0.88}) +[2026-06-23 19:55:21,479][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.11474566161632538, 'train/grad_norm_pre_clip': 31.029399871826172, 'timing/data': 0.0008430348243564367, 'timing/model': 1.692532157059759, 'learning_rate/qwen_vl_interface': 6.062162739750502e-06, 'learning_rate/action_model': 6.062162739750502e-05, 'epoch': 0.88}) +[2026-06-23 19:55:24,817][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.100010946393013, 'train/grad_norm_pre_clip': 12.544440269470215, 'timing/data': 0.000608945032581687, 'timing/model': 1.6959451229777187, 'learning_rate/qwen_vl_interface': 6.059161997797306e-06, 'learning_rate/action_model': 6.0591619977973056e-05, 'epoch': 0.88}) +[2026-06-23 19:55:28,150][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.15065456926822662, 'train/grad_norm_pre_clip': 19.393844604492188, 'timing/data': 0.0006249351426959038, 'timing/model': 1.6945296870544553, 'learning_rate/qwen_vl_interface': 6.056160923228178e-06, 'learning_rate/action_model': 6.056160923228178e-05, 'epoch': 0.88}) +[2026-06-23 19:55:31,491][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.0789828971028328, 'train/grad_norm_pre_clip': 28.458826065063477, 'timing/data': 0.0008743740618228912, 'timing/model': 1.6952485560905188, 'learning_rate/qwen_vl_interface': 6.0531595172767455e-06, 'learning_rate/action_model': 6.0531595172767455e-05, 'epoch': 0.88}) +[2026-06-23 19:55:34,825][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.05169948562979698, 'train/grad_norm_pre_clip': 44.65019607543945, 'timing/data': 0.0006116100121289492, 'timing/model': 1.6905592549592257, 'learning_rate/qwen_vl_interface': 6.050157781176774e-06, 'learning_rate/action_model': 6.050157781176774e-05, 'epoch': 0.88}) +[2026-06-23 19:55:38,165][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.31380656361579895, 'train/grad_norm_pre_clip': 34.1656379699707, 'timing/data': 0.0006195150781422853, 'timing/model': 1.6963292418513447, 'learning_rate/qwen_vl_interface': 6.047155716162164e-06, 'learning_rate/action_model': 6.047155716162164e-05, 'epoch': 0.89}) +[2026-06-23 19:55:41,500][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.1614445447921753, 'train/grad_norm_pre_clip': 43.59832000732422, 'timing/data': 0.0006116409786045551, 'timing/model': 1.6911835630889982, 'learning_rate/qwen_vl_interface': 6.044153323466948e-06, 'learning_rate/action_model': 6.044153323466948e-05, 'epoch': 0.89}) +[2026-06-23 19:55:44,844][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.1009325459599495, 'train/grad_norm_pre_clip': 13.598036766052246, 'timing/data': 0.0006103981286287308, 'timing/model': 1.7002501680981368, 'learning_rate/qwen_vl_interface': 6.0411506043253e-06, 'learning_rate/action_model': 6.0411506043253e-05, 'epoch': 0.89}) +[2026-06-23 19:55:48,181][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.08921966701745987, 'train/grad_norm_pre_clip': 17.609018325805664, 'timing/data': 0.0005981349386274815, 'timing/model': 1.6946322321891785, 'learning_rate/qwen_vl_interface': 6.038147559971523e-06, 'learning_rate/action_model': 6.038147559971523e-05, 'epoch': 0.89}) +[2026-06-23 19:55:51,514][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.1710890531539917, 'train/grad_norm_pre_clip': 26.433135986328125, 'timing/data': 0.0006092959083616734, 'timing/model': 1.6922703369054943, 'learning_rate/qwen_vl_interface': 6.035144191640055e-06, 'learning_rate/action_model': 6.035144191640055e-05, 'epoch': 0.89}) +[2026-06-23 19:55:54,860][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.12327893078327179, 'train/grad_norm_pre_clip': 17.033252716064453, 'timing/data': 0.0006374679505825043, 'timing/model': 1.6966228580567986, 'learning_rate/qwen_vl_interface': 6.032140500565467e-06, 'learning_rate/action_model': 6.0321405005654665e-05, 'epoch': 0.89}) +[2026-06-23 19:55:58,199][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.13011434674263, 'train/grad_norm_pre_clip': 36.98871994018555, 'timing/data': 0.000617380952462554, 'timing/model': 1.6913950548041612, 'learning_rate/qwen_vl_interface': 6.029136487982463e-06, 'learning_rate/action_model': 6.029136487982463e-05, 'epoch': 0.89}) +[2026-06-23 19:56:01,537][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.15507002174854279, 'train/grad_norm_pre_clip': 38.17540740966797, 'timing/data': 0.0006690078880637884, 'timing/model': 1.6903787611518055, 'learning_rate/qwen_vl_interface': 6.02613215512588e-06, 'learning_rate/action_model': 6.0261321551258796e-05, 'epoch': 0.89}) +[2026-06-23 19:56:04,870][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.07158339023590088, 'train/grad_norm_pre_clip': 32.84907531738281, 'timing/data': 0.0006635382305830717, 'timing/model': 1.694383188849315, 'learning_rate/qwen_vl_interface': 6.023127503230684e-06, 'learning_rate/action_model': 6.023127503230683e-05, 'epoch': 0.89}) +[2026-06-23 19:56:08,294][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.1516188532114029, 'train/grad_norm_pre_clip': 41.95930480957031, 'timing/data': 0.0006069510709494352, 'timing/model': 1.7880271149333566, 'learning_rate/qwen_vl_interface': 6.020122533531976e-06, 'learning_rate/action_model': 6.0201225335319755e-05, 'epoch': 0.89}) +[2026-06-23 19:56:11,628][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.15172459185123444, 'train/grad_norm_pre_clip': 16.599811553955078, 'timing/data': 0.0006677759811282158, 'timing/model': 1.693108073901385, 'learning_rate/qwen_vl_interface': 6.017117247264983e-06, 'learning_rate/action_model': 6.017117247264983e-05, 'epoch': 0.89}) +[2026-06-23 19:56:14,962][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.08221663534641266, 'train/grad_norm_pre_clip': 20.292224884033203, 'timing/data': 0.0006087250076234341, 'timing/model': 1.6940570259466767, 'learning_rate/qwen_vl_interface': 6.014111645665069e-06, 'learning_rate/action_model': 6.0141116456650684e-05, 'epoch': 0.89}) +[2026-06-23 19:56:18,299][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.1117141917347908, 'train/grad_norm_pre_clip': 32.58024215698242, 'timing/data': 0.000612541101872921, 'timing/model': 1.6921784500591457, 'learning_rate/qwen_vl_interface': 6.011105729967719e-06, 'learning_rate/action_model': 6.0111057299677184e-05, 'epoch': 0.89}) +[2026-06-23 19:56:21,635][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2300 +[2026-06-23 19:57:54,782][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2300 +[2026-06-23 19:57:54,782][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2300 +[2026-06-23 19:59:06,676][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2300 +[2026-06-23 19:59:06,677][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.11585145443677902, 'train/grad_norm_pre_clip': 55.4166259765625, 'eval/action_loss/seconds': 93.14390364405699, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.0735531821846962, 'eval/latency_2/loss': 0.0735531821846962, 'eval/flappy/loss': 0.0735531821846962, 'eval/flappy/latency_2/loss': 0.0735531821846962, 'eval/flap_precision': 0.9256530475552579, 'eval/flap_recall': 0.9325236167341431, 'eval/flap_f1': 0.9290756302521008, 'eval/flap_rate_model': 0.2073611111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9290756302521008, 'eval/latency_2/flap_precision': 0.9256530475552579, 'eval/latency_2/flap_recall': 0.9325236167341431, 'eval/latency_2/flap_f1': 0.9290756302521008, 'eval/latency_2/flap_rate_model': 0.2073611111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9290756302521008, 'eval/action_classification/data_wait_seconds': 0.8924435377120972, 'eval/action_classification/predict_seconds': 70.84159851074219, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006901179440319538, 'timing/model': 1.6916976140346378, 'learning_rate/qwen_vl_interface': 6.008099501408552e-06, 'learning_rate/action_model': 6.008099501408552e-05, 'epoch': 0.89}) +[2026-06-23 19:59:10,018][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.13226328790187836, 'train/grad_norm_pre_clip': 21.53870964050293, 'timing/data': 0.0010563060641288757, 'timing/model': 1.6883142988663167, 'learning_rate/qwen_vl_interface': 6.0050929612233165e-06, 'learning_rate/action_model': 6.005092961223316e-05, 'epoch': 0.89}) +[2026-06-23 19:59:13,341][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.16629573702812195, 'train/grad_norm_pre_clip': 19.103078842163086, 'timing/data': 0.000597864156588912, 'timing/model': 1.6870799760799855, 'learning_rate/qwen_vl_interface': 6.002086110647888e-06, 'learning_rate/action_model': 6.002086110647888e-05, 'epoch': 0.89}) +[2026-06-23 19:59:16,662][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.2406739890575409, 'train/grad_norm_pre_clip': 82.02140045166016, 'timing/data': 0.0006435699760913849, 'timing/model': 1.6872502740006894, 'learning_rate/qwen_vl_interface': 5.999078950918267e-06, 'learning_rate/action_model': 5.999078950918267e-05, 'epoch': 0.89}) +[2026-06-23 19:59:20,288][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.17616386711597443, 'train/grad_norm_pre_clip': 65.24671173095703, 'timing/data': 0.0007286909967660904, 'timing/model': 1.690466018859297, 'learning_rate/qwen_vl_interface': 5.996071483270585e-06, 'learning_rate/action_model': 5.996071483270585e-05, 'epoch': 0.89}) +[2026-06-23 19:59:23,612][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.1548050493001938, 'train/grad_norm_pre_clip': 43.40781021118164, 'timing/data': 0.0006182119250297546, 'timing/model': 1.6893946318887174, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.89}) +[2026-06-23 19:59:26,937][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.10277897119522095, 'train/grad_norm_pre_clip': 15.817422866821289, 'timing/data': 0.0006163481157273054, 'timing/model': 1.687947589205578, 'learning_rate/qwen_vl_interface': 5.990055629166187e-06, 'learning_rate/action_model': 5.990055629166187e-05, 'epoch': 0.89}) +[2026-06-23 19:59:30,273][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.15660622715950012, 'train/grad_norm_pre_clip': 12.7003173828125, 'timing/data': 0.0006875419057905674, 'timing/model': 1.690824736841023, 'learning_rate/qwen_vl_interface': 5.987047245182363e-06, 'learning_rate/action_model': 5.987047245182362e-05, 'epoch': 0.89}) +[2026-06-23 19:59:33,598][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.20937579870224, 'train/grad_norm_pre_clip': 35.92755126953125, 'timing/data': 0.0006365668959915638, 'timing/model': 1.6860290069598705, 'learning_rate/qwen_vl_interface': 5.984038558226255e-06, 'learning_rate/action_model': 5.9840385582262546e-05, 'epoch': 0.89}) +[2026-06-23 19:59:36,928][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.12570291757583618, 'train/grad_norm_pre_clip': 24.18589973449707, 'timing/data': 0.000614505959674716, 'timing/model': 1.6927484800107777, 'learning_rate/qwen_vl_interface': 5.981029569534625e-06, 'learning_rate/action_model': 5.981029569534625e-05, 'epoch': 0.89}) +[2026-06-23 19:59:40,258][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.12342503666877747, 'train/grad_norm_pre_clip': 15.640961647033691, 'timing/data': 0.0006568548269569874, 'timing/model': 1.689512754091993, 'learning_rate/qwen_vl_interface': 5.978020280344351e-06, 'learning_rate/action_model': 5.9780202803443506e-05, 'epoch': 0.89}) +[2026-06-23 19:59:43,604][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.14618657529354095, 'train/grad_norm_pre_clip': 49.27679443359375, 'timing/data': 0.0006659019272774458, 'timing/model': 1.6971241040155292, 'learning_rate/qwen_vl_interface': 5.97501069189244e-06, 'learning_rate/action_model': 5.97501069189244e-05, 'epoch': 0.89}) +[2026-06-23 19:59:46,946][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.32879921793937683, 'train/grad_norm_pre_clip': 51.537784576416016, 'timing/data': 0.0006685371045023203, 'timing/model': 1.6948535130359232, 'learning_rate/qwen_vl_interface': 5.97200080541602e-06, 'learning_rate/action_model': 5.97200080541602e-05, 'epoch': 0.89}) +[2026-06-23 19:59:50,286][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.18206816911697388, 'train/grad_norm_pre_clip': 37.80232620239258, 'timing/data': 0.0006116698496043682, 'timing/model': 1.6957820071838796, 'learning_rate/qwen_vl_interface': 5.968990622152341e-06, 'learning_rate/action_model': 5.9689906221523404e-05, 'epoch': 0.9}) +[2026-06-23 19:59:53,620][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.09787552058696747, 'train/grad_norm_pre_clip': 20.635406494140625, 'timing/data': 0.0006166789680719376, 'timing/model': 1.6929992120712996, 'learning_rate/qwen_vl_interface': 5.965980143338774e-06, 'learning_rate/action_model': 5.965980143338774e-05, 'epoch': 0.9}) +[2026-06-23 19:59:56,964][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.15379202365875244, 'train/grad_norm_pre_clip': 41.95795822143555, 'timing/data': 0.0005983349401503801, 'timing/model': 1.6968066298868507, 'learning_rate/qwen_vl_interface': 5.962969370212819e-06, 'learning_rate/action_model': 5.9629693702128185e-05, 'epoch': 0.9}) +[2026-06-23 20:00:00,311][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.14547845721244812, 'train/grad_norm_pre_clip': 35.482173919677734, 'timing/data': 0.0007145330309867859, 'timing/model': 1.6992455029394478, 'learning_rate/qwen_vl_interface': 5.959958304012085e-06, 'learning_rate/action_model': 5.9599583040120845e-05, 'epoch': 0.9}) +[2026-06-23 20:00:03,651][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.17878535389900208, 'train/grad_norm_pre_clip': 38.366397857666016, 'timing/data': 0.0008300598710775375, 'timing/model': 1.6962382271885872, 'learning_rate/qwen_vl_interface': 5.956946945974309e-06, 'learning_rate/action_model': 5.9569469459743085e-05, 'epoch': 0.9}) +[2026-06-23 20:00:06,988][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.20551735162734985, 'train/grad_norm_pre_clip': 14.942044258117676, 'timing/data': 0.0006580369081348181, 'timing/model': 1.6929869519080967, 'learning_rate/qwen_vl_interface': 5.953935297337347e-06, 'learning_rate/action_model': 5.953935297337347e-05, 'epoch': 0.9}) +[2026-06-23 20:00:10,563][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.10681917518377304, 'train/grad_norm_pre_clip': 18.28056526184082, 'timing/data': 0.0007753891404718161, 'timing/model': 1.6994879129342735, 'learning_rate/qwen_vl_interface': 5.950923359339175e-06, 'learning_rate/action_model': 5.950923359339174e-05, 'epoch': 0.9}) +[2026-06-23 20:00:13,908][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.14952170848846436, 'train/grad_norm_pre_clip': 18.222213745117188, 'timing/data': 0.0006992148701101542, 'timing/model': 1.6964607201516628, 'learning_rate/qwen_vl_interface': 5.947911133217886e-06, 'learning_rate/action_model': 5.947911133217886e-05, 'epoch': 0.9}) +[2026-06-23 20:00:17,250][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.1627170890569687, 'train/grad_norm_pre_clip': 17.688030242919922, 'timing/data': 0.0007253040093928576, 'timing/model': 1.695425079902634, 'learning_rate/qwen_vl_interface': 5.944898620211693e-06, 'learning_rate/action_model': 5.944898620211693e-05, 'epoch': 0.9}) +[2026-06-23 20:00:20,598][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.186370849609375, 'train/grad_norm_pre_clip': 50.88053894042969, 'timing/data': 0.0006978518795222044, 'timing/model': 1.7022304448764771, 'learning_rate/qwen_vl_interface': 5.941885821558925e-06, 'learning_rate/action_model': 5.9418858215589245e-05, 'epoch': 0.9}) +[2026-06-23 20:00:23,951][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.15704840421676636, 'train/grad_norm_pre_clip': 18.961105346679688, 'timing/data': 0.0007168478332459927, 'timing/model': 1.7011997560039163, 'learning_rate/qwen_vl_interface': 5.938872738498031e-06, 'learning_rate/action_model': 5.93887273849803e-05, 'epoch': 0.9}) +[2026-06-23 20:00:27,299][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.0988861694931984, 'train/grad_norm_pre_clip': 10.226420402526855, 'timing/data': 0.0007296109106391668, 'timing/model': 1.6977746731135994, 'learning_rate/qwen_vl_interface': 5.935859372267575e-06, 'learning_rate/action_model': 5.935859372267575e-05, 'epoch': 0.9}) +[2026-06-23 20:00:30,653][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.09475535154342651, 'train/grad_norm_pre_clip': 25.40407371520996, 'timing/data': 0.0006634281016886234, 'timing/model': 1.7034879738930613, 'learning_rate/qwen_vl_interface': 5.9328457241062396e-06, 'learning_rate/action_model': 5.9328457241062394e-05, 'epoch': 0.9}) +[2026-06-23 20:00:33,999][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.1726236343383789, 'train/grad_norm_pre_clip': 32.40015411376953, 'timing/data': 0.0007027420215308666, 'timing/model': 1.6999629640486091, 'learning_rate/qwen_vl_interface': 5.929831795252819e-06, 'learning_rate/action_model': 5.9298317952528184e-05, 'epoch': 0.9}) +[2026-06-23 20:00:37,352][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.09517142176628113, 'train/grad_norm_pre_clip': 20.792333602905273, 'timing/data': 0.000877860002219677, 'timing/model': 1.7042673260439187, 'learning_rate/qwen_vl_interface': 5.926817586946227e-06, 'learning_rate/action_model': 5.926817586946227e-05, 'epoch': 0.9}) +[2026-06-23 20:00:40,695][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.15082286298274994, 'train/grad_norm_pre_clip': 12.968677520751953, 'timing/data': 0.0006241030059754848, 'timing/model': 1.6994904722087085, 'learning_rate/qwen_vl_interface': 5.923803100425491e-06, 'learning_rate/action_model': 5.92380310042549e-05, 'epoch': 0.9}) +[2026-06-23 20:00:44,049][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.13527007400989532, 'train/grad_norm_pre_clip': 37.75594711303711, 'timing/data': 0.0006131930276751518, 'timing/model': 1.700139501132071, 'learning_rate/qwen_vl_interface': 5.92078833692975e-06, 'learning_rate/action_model': 5.92078833692975e-05, 'epoch': 0.9}) +[2026-06-23 20:00:47,398][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.18478398025035858, 'train/grad_norm_pre_clip': 30.66372299194336, 'timing/data': 0.0006183329969644547, 'timing/model': 1.7020242440048605, 'learning_rate/qwen_vl_interface': 5.917773297698264e-06, 'learning_rate/action_model': 5.9177732976982636e-05, 'epoch': 0.9}) +[2026-06-23 20:00:50,750][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.1503172665834427, 'train/grad_norm_pre_clip': 21.992082595825195, 'timing/data': 0.0007723618764430285, 'timing/model': 1.7025730938185006, 'learning_rate/qwen_vl_interface': 5.9147579839704e-06, 'learning_rate/action_model': 5.9147579839704e-05, 'epoch': 0.9}) +[2026-06-23 20:00:54,096][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.13471904397010803, 'train/grad_norm_pre_clip': 13.122445106506348, 'timing/data': 0.0005967931356281042, 'timing/model': 1.7026005929801613, 'learning_rate/qwen_vl_interface': 5.911742396985637e-06, 'learning_rate/action_model': 5.9117423969856364e-05, 'epoch': 0.9}) +[2026-06-23 20:00:57,529][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.12289905548095703, 'train/grad_norm_pre_clip': 16.422592163085938, 'timing/data': 0.0006892660167068243, 'timing/model': 1.785760898841545, 'learning_rate/qwen_vl_interface': 5.908726537983571e-06, 'learning_rate/action_model': 5.908726537983571e-05, 'epoch': 0.9}) +[2026-06-23 20:01:00,879][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.1307094693183899, 'train/grad_norm_pre_clip': 12.896760940551758, 'timing/data': 0.0006139250472187996, 'timing/model': 1.7025593519210815, 'learning_rate/qwen_vl_interface': 5.9057104082039085e-06, 'learning_rate/action_model': 5.905710408203908e-05, 'epoch': 0.9}) +[2026-06-23 20:01:04,221][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.09427421540021896, 'train/grad_norm_pre_clip': 32.885189056396484, 'timing/data': 0.0007332991808652878, 'timing/model': 1.6938359478954226, 'learning_rate/qwen_vl_interface': 5.902694008886466e-06, 'learning_rate/action_model': 5.902694008886466e-05, 'epoch': 0.9}) +[2026-06-23 20:01:07,562][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.05578547343611717, 'train/grad_norm_pre_clip': 12.150276184082031, 'timing/data': 0.0006102980114519596, 'timing/model': 1.698198680067435, 'learning_rate/qwen_vl_interface': 5.8996773412711694e-06, 'learning_rate/action_model': 5.899677341271169e-05, 'epoch': 0.9}) +[2026-06-23 20:01:10,910][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.10736458003520966, 'train/grad_norm_pre_clip': 12.38711929321289, 'timing/data': 0.0006082339677959681, 'timing/model': 1.6971646340098232, 'learning_rate/qwen_vl_interface': 5.896660406598059e-06, 'learning_rate/action_model': 5.896660406598059e-05, 'epoch': 0.9}) +[2026-06-23 20:01:14,265][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.028930893167853355, 'train/grad_norm_pre_clip': 38.81965637207031, 'timing/data': 0.0006074318662285805, 'timing/model': 1.7090765808243304, 'learning_rate/qwen_vl_interface': 5.893643206107283e-06, 'learning_rate/action_model': 5.893643206107282e-05, 'epoch': 0.9}) +[2026-06-23 20:01:17,612][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.13508710265159607, 'train/grad_norm_pre_clip': 22.23442268371582, 'timing/data': 0.0006576371379196644, 'timing/model': 1.6980102409143, 'learning_rate/qwen_vl_interface': 5.890625741039096e-06, 'learning_rate/action_model': 5.890625741039095e-05, 'epoch': 0.91}) +[2026-06-23 20:01:20,951][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.126820906996727, 'train/grad_norm_pre_clip': 30.53037452697754, 'timing/data': 0.000595390098169446, 'timing/model': 1.694740581093356, 'learning_rate/qwen_vl_interface': 5.887608012633864e-06, 'learning_rate/action_model': 5.887608012633864e-05, 'epoch': 0.91}) +[2026-06-23 20:01:24,296][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.1012115553021431, 'train/grad_norm_pre_clip': 39.37940216064453, 'timing/data': 0.00068901595659554, 'timing/model': 1.701668456196785, 'learning_rate/qwen_vl_interface': 5.884590022132064e-06, 'learning_rate/action_model': 5.884590022132064e-05, 'epoch': 0.91}) +[2026-06-23 20:01:27,637][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.15515604615211487, 'train/grad_norm_pre_clip': 20.309776306152344, 'timing/data': 0.0006559328176081181, 'timing/model': 1.6972334450110793, 'learning_rate/qwen_vl_interface': 5.881571770774278e-06, 'learning_rate/action_model': 5.881571770774278e-05, 'epoch': 0.91}) +[2026-06-23 20:01:30,981][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.04036686569452286, 'train/grad_norm_pre_clip': 17.870285034179688, 'timing/data': 0.0006043459288775921, 'timing/model': 1.7013452860992402, 'learning_rate/qwen_vl_interface': 5.878553259801192e-06, 'learning_rate/action_model': 5.878553259801192e-05, 'epoch': 0.91}) +[2026-06-23 20:01:34,323][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.11524509638547897, 'train/grad_norm_pre_clip': 22.0053768157959, 'timing/data': 0.0006052779499441385, 'timing/model': 1.6969800551887602, 'learning_rate/qwen_vl_interface': 5.8755344904536025e-06, 'learning_rate/action_model': 5.8755344904536025e-05, 'epoch': 0.91}) +[2026-06-23 20:01:37,663][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.13076534867286682, 'train/grad_norm_pre_clip': 18.98198127746582, 'timing/data': 0.0006103178020566702, 'timing/model': 1.6946279818657786, 'learning_rate/qwen_vl_interface': 5.872515463972415e-06, 'learning_rate/action_model': 5.8725154639724146e-05, 'epoch': 0.91}) +[2026-06-23 20:01:41,007][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.19042541086673737, 'train/grad_norm_pre_clip': 44.733245849609375, 'timing/data': 0.00061447499319911, 'timing/model': 1.6992843048647046, 'learning_rate/qwen_vl_interface': 5.869496181598633e-06, 'learning_rate/action_model': 5.869496181598633e-05, 'epoch': 0.91}) +[2026-06-23 20:01:44,354][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.24418121576309204, 'train/grad_norm_pre_clip': 28.667184829711914, 'timing/data': 0.0006275400519371033, 'timing/model': 1.7033213740214705, 'learning_rate/qwen_vl_interface': 5.866476644573371e-06, 'learning_rate/action_model': 5.866476644573371e-05, 'epoch': 0.91}) +[2026-06-23 20:01:47,788][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.11188013851642609, 'train/grad_norm_pre_clip': 20.514345169067383, 'timing/data': 0.0006271989550441504, 'timing/model': 1.695262118941173, 'learning_rate/qwen_vl_interface': 5.863456854137848e-06, 'learning_rate/action_model': 5.8634568541378474e-05, 'epoch': 0.91}) +[2026-06-23 20:01:51,140][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.14553707838058472, 'train/grad_norm_pre_clip': 73.89453125, 'timing/data': 0.0006173611618578434, 'timing/model': 1.7054673270322382, 'learning_rate/qwen_vl_interface': 5.860436811533383e-06, 'learning_rate/action_model': 5.860436811533383e-05, 'epoch': 0.91}) +[2026-06-23 20:01:54,479][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.10403743386268616, 'train/grad_norm_pre_clip': 84.76752471923828, 'timing/data': 0.0006108579691499472, 'timing/model': 1.6949968170374632, 'learning_rate/qwen_vl_interface': 5.857416518001404e-06, 'learning_rate/action_model': 5.857416518001404e-05, 'epoch': 0.91}) +[2026-06-23 20:01:57,821][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.2570693790912628, 'train/grad_norm_pre_clip': 27.780078887939453, 'timing/data': 0.0006868720520287752, 'timing/model': 1.6979605490341783, 'learning_rate/qwen_vl_interface': 5.854395974783439e-06, 'learning_rate/action_model': 5.8543959747834386e-05, 'epoch': 0.91}) +[2026-06-23 20:02:01,155][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.08751548826694489, 'train/grad_norm_pre_clip': 51.44892120361328, 'timing/data': 0.0006654108874499798, 'timing/model': 1.6944894909393042, 'learning_rate/qwen_vl_interface': 5.851375183121119e-06, 'learning_rate/action_model': 5.8513751831211186e-05, 'epoch': 0.91}) +[2026-06-23 20:02:04,498][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.16500897705554962, 'train/grad_norm_pre_clip': 27.784887313842773, 'timing/data': 0.0007210460025817156, 'timing/model': 1.7016445929184556, 'learning_rate/qwen_vl_interface': 5.848354144256175e-06, 'learning_rate/action_model': 5.8483541442561754e-05, 'epoch': 0.91}) +[2026-06-23 20:02:07,840][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.09125497192144394, 'train/grad_norm_pre_clip': 15.21790599822998, 'timing/data': 0.0006626159884035587, 'timing/model': 1.6987544759176672, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.91}) +[2026-06-23 20:02:11,188][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.03551231324672699, 'train/grad_norm_pre_clip': 11.741617202758789, 'timing/data': 0.0008407910354435444, 'timing/model': 1.7026410328689963, 'learning_rate/qwen_vl_interface': 5.842311329885866e-06, 'learning_rate/action_model': 5.842311329885866e-05, 'epoch': 0.91}) +[2026-06-23 20:02:14,535][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.10441690683364868, 'train/grad_norm_pre_clip': 64.35240173339844, 'timing/data': 0.0006092749536037445, 'timing/model': 1.6988697398919612, 'learning_rate/qwen_vl_interface': 5.839289556864472e-06, 'learning_rate/action_model': 5.8392895568644714e-05, 'epoch': 0.91}) +[2026-06-23 20:02:17,879][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.07223475724458694, 'train/grad_norm_pre_clip': 12.893043518066406, 'timing/data': 0.0006650509312748909, 'timing/model': 1.699073983123526, 'learning_rate/qwen_vl_interface': 5.836267541608398e-06, 'learning_rate/action_model': 5.836267541608398e-05, 'epoch': 0.91}) +[2026-06-23 20:02:21,223][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.09636398404836655, 'train/grad_norm_pre_clip': 12.777701377868652, 'timing/data': 0.0006166489329189062, 'timing/model': 1.700852746842429, 'learning_rate/qwen_vl_interface': 5.833245285359887e-06, 'learning_rate/action_model': 5.833245285359886e-05, 'epoch': 0.91}) +[2026-06-23 20:02:24,562][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.08263514935970306, 'train/grad_norm_pre_clip': 37.054969787597656, 'timing/data': 0.000626478111371398, 'timing/model': 1.6933465108741075, 'learning_rate/qwen_vl_interface': 5.830222789361268e-06, 'learning_rate/action_model': 5.8302227893612684e-05, 'epoch': 0.91}) +[2026-06-23 20:02:27,915][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.03589537739753723, 'train/grad_norm_pre_clip': 17.900020599365234, 'timing/data': 0.0006120810285210609, 'timing/model': 1.7023334370460361, 'learning_rate/qwen_vl_interface': 5.8272000548549776e-06, 'learning_rate/action_model': 5.827200054854978e-05, 'epoch': 0.91}) +[2026-06-23 20:02:31,265][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.18453988432884216, 'train/grad_norm_pre_clip': 30.236351013183594, 'timing/data': 0.0008334871381521225, 'timing/model': 1.701989219058305, 'learning_rate/qwen_vl_interface': 5.824177083083547e-06, 'learning_rate/action_model': 5.8241770830835475e-05, 'epoch': 0.91}) +[2026-06-23 20:02:34,697][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.12805184721946716, 'train/grad_norm_pre_clip': 12.646767616271973, 'timing/data': 0.0006722740363329649, 'timing/model': 1.7829878870397806, 'learning_rate/qwen_vl_interface': 5.8211538752896056e-06, 'learning_rate/action_model': 5.821153875289606e-05, 'epoch': 0.91}) +[2026-06-23 20:02:38,041][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.09427184611558914, 'train/grad_norm_pre_clip': 16.2503662109375, 'timing/data': 0.0006110391113907099, 'timing/model': 1.6996006949339062, 'learning_rate/qwen_vl_interface': 5.8181304327158814e-06, 'learning_rate/action_model': 5.8181304327158814e-05, 'epoch': 0.91}) +[2026-06-23 20:02:41,390][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.11049264669418335, 'train/grad_norm_pre_clip': 14.340943336486816, 'timing/data': 0.0006162989884614944, 'timing/model': 1.7022014639806002, 'learning_rate/qwen_vl_interface': 5.8151067566051965e-06, 'learning_rate/action_model': 5.815106756605196e-05, 'epoch': 0.91}) +[2026-06-23 20:02:44,738][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.2123228758573532, 'train/grad_norm_pre_clip': 41.879364013671875, 'timing/data': 0.0006031040102243423, 'timing/model': 1.7016535731963813, 'learning_rate/qwen_vl_interface': 5.81208284820047e-06, 'learning_rate/action_model': 5.81208284820047e-05, 'epoch': 0.92}) +[2026-06-23 20:02:48,088][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.14220930635929108, 'train/grad_norm_pre_clip': 41.789405822753906, 'timing/data': 0.0006157378666102886, 'timing/model': 1.698837336152792, 'learning_rate/qwen_vl_interface': 5.809058708744716e-06, 'learning_rate/action_model': 5.809058708744716e-05, 'epoch': 0.92}) +[2026-06-23 20:02:51,439][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.19344376027584076, 'train/grad_norm_pre_clip': 19.5190486907959, 'timing/data': 0.0006220899522304535, 'timing/model': 1.7027771819848567, 'learning_rate/qwen_vl_interface': 5.8060343394810445e-06, 'learning_rate/action_model': 5.806034339481045e-05, 'epoch': 0.92}) +[2026-06-23 20:02:54,783][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.07961918413639069, 'train/grad_norm_pre_clip': 17.119081497192383, 'timing/data': 0.0005992259830236435, 'timing/model': 1.6994843089487404, 'learning_rate/qwen_vl_interface': 5.803009741652659e-06, 'learning_rate/action_model': 5.8030097416526585e-05, 'epoch': 0.92}) +[2026-06-23 20:02:58,133][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.12662558257579803, 'train/grad_norm_pre_clip': 73.39846801757812, 'timing/data': 0.0006164589431136847, 'timing/model': 1.7037494480609894, 'learning_rate/qwen_vl_interface': 5.799984916502859e-06, 'learning_rate/action_model': 5.799984916502859e-05, 'epoch': 0.92}) +[2026-06-23 20:03:01,485][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.14146725833415985, 'train/grad_norm_pre_clip': 78.14722442626953, 'timing/data': 0.000604547094553709, 'timing/model': 1.70420909114182, 'learning_rate/qwen_vl_interface': 5.7969598652750334e-06, 'learning_rate/action_model': 5.796959865275033e-05, 'epoch': 0.92}) +[2026-06-23 20:03:04,835][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.17227803170681, 'train/grad_norm_pre_clip': 37.936038970947266, 'timing/data': 0.0006080931052565575, 'timing/model': 1.7013031060341746, 'learning_rate/qwen_vl_interface': 5.793934589212666e-06, 'learning_rate/action_model': 5.793934589212666e-05, 'epoch': 0.92}) +[2026-06-23 20:03:08,186][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.04632383584976196, 'train/grad_norm_pre_clip': 13.80055046081543, 'timing/data': 0.000623342115432024, 'timing/model': 1.7031919760629535, 'learning_rate/qwen_vl_interface': 5.7909090895593375e-06, 'learning_rate/action_model': 5.790909089559338e-05, 'epoch': 0.92}) +[2026-06-23 20:03:11,537][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.13783356547355652, 'train/grad_norm_pre_clip': 23.777328491210938, 'timing/data': 0.0006046169437468052, 'timing/model': 1.701084311120212, 'learning_rate/qwen_vl_interface': 5.7878833675587135e-06, 'learning_rate/action_model': 5.787883367558713e-05, 'epoch': 0.92}) +[2026-06-23 20:03:14,882][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.17103929817676544, 'train/grad_norm_pre_clip': 26.7851619720459, 'timing/data': 0.000615136930719018, 'timing/model': 1.6997564141638577, 'learning_rate/qwen_vl_interface': 5.784857424454556e-06, 'learning_rate/action_model': 5.7848574244545563e-05, 'epoch': 0.92}) +[2026-06-23 20:03:18,230][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.09651585668325424, 'train/grad_norm_pre_clip': 46.33668518066406, 'timing/data': 0.0006016409024596214, 'timing/model': 1.7004323720466346, 'learning_rate/qwen_vl_interface': 5.7818312614907125e-06, 'learning_rate/action_model': 5.781831261490712e-05, 'epoch': 0.92}) +[2026-06-23 20:03:21,621][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.3148041069507599, 'train/grad_norm_pre_clip': 37.9078254699707, 'timing/data': 0.0008302319329231977, 'timing/model': 1.742990893078968, 'learning_rate/qwen_vl_interface': 5.778804879911129e-06, 'learning_rate/action_model': 5.778804879911128e-05, 'epoch': 0.92}) +[2026-06-23 20:03:25,141][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.035460010170936584, 'train/grad_norm_pre_clip': 12.099513053894043, 'timing/data': 0.0007879820186644793, 'timing/model': 1.8473298270255327, 'learning_rate/qwen_vl_interface': 5.775778280959832e-06, 'learning_rate/action_model': 5.7757782809598315e-05, 'epoch': 0.92}) +[2026-06-23 20:03:28,493][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.16980960965156555, 'train/grad_norm_pre_clip': 18.484786987304688, 'timing/data': 0.0006113790441304445, 'timing/model': 1.6948236259631813, 'learning_rate/qwen_vl_interface': 5.7727514658809435e-06, 'learning_rate/action_model': 5.772751465880944e-05, 'epoch': 0.92}) +[2026-06-23 20:03:31,838][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.15452606976032257, 'train/grad_norm_pre_clip': 11.221878051757812, 'timing/data': 0.0006148261018097401, 'timing/model': 1.699489447986707, 'learning_rate/qwen_vl_interface': 5.769724435918676e-06, 'learning_rate/action_model': 5.7697244359186755e-05, 'epoch': 0.92}) +[2026-06-23 20:03:35,188][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.10863832384347916, 'train/grad_norm_pre_clip': 11.105545997619629, 'timing/data': 0.0006152561400085688, 'timing/model': 1.7002316711004823, 'learning_rate/qwen_vl_interface': 5.766697192317323e-06, 'learning_rate/action_model': 5.766697192317323e-05, 'epoch': 0.92}) +[2026-06-23 20:03:38,536][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.11817539483308792, 'train/grad_norm_pre_clip': 14.303092002868652, 'timing/data': 0.0006006199400871992, 'timing/model': 1.7012620479799807, 'learning_rate/qwen_vl_interface': 5.763669736321271e-06, 'learning_rate/action_model': 5.763669736321271e-05, 'epoch': 0.92}) +[2026-06-23 20:03:41,885][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.15029996633529663, 'train/grad_norm_pre_clip': 35.49319076538086, 'timing/data': 0.0006143951322883368, 'timing/model': 1.699719242984429, 'learning_rate/qwen_vl_interface': 5.760642069174994e-06, 'learning_rate/action_model': 5.760642069174994e-05, 'epoch': 0.92}) +[2026-06-23 20:03:45,229][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.06565406918525696, 'train/grad_norm_pre_clip': 47.72725296020508, 'timing/data': 0.0006048069335520267, 'timing/model': 1.6960562251042575, 'learning_rate/qwen_vl_interface': 5.757614192123051e-06, 'learning_rate/action_model': 5.757614192123051e-05, 'epoch': 0.92}) +[2026-06-23 20:03:48,575][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.18808923661708832, 'train/grad_norm_pre_clip': 53.38563537597656, 'timing/data': 0.000751884188503027, 'timing/model': 1.696636774810031, 'learning_rate/qwen_vl_interface': 5.754586106410089e-06, 'learning_rate/action_model': 5.754586106410088e-05, 'epoch': 0.92}) +[2026-06-23 20:03:51,916][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.23995688557624817, 'train/grad_norm_pre_clip': 37.29158020019531, 'timing/data': 0.0006361159030348063, 'timing/model': 1.6957673190627247, 'learning_rate/qwen_vl_interface': 5.751557813280838e-06, 'learning_rate/action_model': 5.751557813280838e-05, 'epoch': 0.92}) +[2026-06-23 20:03:55,269][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.05605556070804596, 'train/grad_norm_pre_clip': 20.42665672302246, 'timing/data': 0.000620787963271141, 'timing/model': 1.7011288988869637, 'learning_rate/qwen_vl_interface': 5.7485293139801146e-06, 'learning_rate/action_model': 5.7485293139801146e-05, 'epoch': 0.92}) +[2026-06-23 20:03:58,616][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.08102080225944519, 'train/grad_norm_pre_clip': 58.148048400878906, 'timing/data': 0.0006175409071147442, 'timing/model': 1.7023158399388194, 'learning_rate/qwen_vl_interface': 5.745500609752821e-06, 'learning_rate/action_model': 5.745500609752821e-05, 'epoch': 0.92}) +[2026-06-23 20:04:01,966][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.22991850972175598, 'train/grad_norm_pre_clip': 77.53899383544922, 'timing/data': 0.0008508299943059683, 'timing/model': 1.7000306718982756, 'learning_rate/qwen_vl_interface': 5.742471701843943e-06, 'learning_rate/action_model': 5.742471701843943e-05, 'epoch': 0.92}) +[2026-06-23 20:04:05,314][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.2146362066268921, 'train/grad_norm_pre_clip': 70.8191909790039, 'timing/data': 0.0006062001921236515, 'timing/model': 1.700190271018073, 'learning_rate/qwen_vl_interface': 5.739442591498553e-06, 'learning_rate/action_model': 5.739442591498553e-05, 'epoch': 0.92}) +[2026-06-23 20:04:08,671][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.11285875737667084, 'train/grad_norm_pre_clip': 49.32614517211914, 'timing/data': 0.0006220089271664619, 'timing/model': 1.707018107175827, 'learning_rate/qwen_vl_interface': 5.7364132799617985e-06, 'learning_rate/action_model': 5.7364132799617984e-05, 'epoch': 0.92}) +[2026-06-23 20:04:12,018][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.1891811490058899, 'train/grad_norm_pre_clip': 36.851104736328125, 'timing/data': 0.0006278811488300562, 'timing/model': 1.700260850833729, 'learning_rate/qwen_vl_interface': 5.733383768478923e-06, 'learning_rate/action_model': 5.7333837684789225e-05, 'epoch': 0.93}) +[2026-06-23 20:04:15,493][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.1299477219581604, 'train/grad_norm_pre_clip': 60.612953186035156, 'timing/data': 0.0006744891870766878, 'timing/model': 1.6959315189160407, 'learning_rate/qwen_vl_interface': 5.730354058295234e-06, 'learning_rate/action_model': 5.730354058295234e-05, 'epoch': 0.93}) +[2026-06-23 20:04:18,841][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.2186402976512909, 'train/grad_norm_pre_clip': 73.3335952758789, 'timing/data': 0.0006032539531588554, 'timing/model': 1.701718098949641, 'learning_rate/qwen_vl_interface': 5.72732415065614e-06, 'learning_rate/action_model': 5.727324150656139e-05, 'epoch': 0.93}) +[2026-06-23 20:04:22,193][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.12591548264026642, 'train/grad_norm_pre_clip': 92.20220947265625, 'timing/data': 0.0006868510972708464, 'timing/model': 1.7037344588898122, 'learning_rate/qwen_vl_interface': 5.724294046807115e-06, 'learning_rate/action_model': 5.724294046807115e-05, 'epoch': 0.93}) +[2026-06-23 20:04:25,541][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.19029828906059265, 'train/grad_norm_pre_clip': 96.79569244384766, 'timing/data': 0.0007010179106146097, 'timing/model': 1.7004059487953782, 'learning_rate/qwen_vl_interface': 5.7212637479937234e-06, 'learning_rate/action_model': 5.721263747993724e-05, 'epoch': 0.93}) +[2026-06-23 20:04:28,885][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.3515659272670746, 'train/grad_norm_pre_clip': 61.185821533203125, 'timing/data': 0.0007010479457676411, 'timing/model': 1.696491336915642, 'learning_rate/qwen_vl_interface': 5.718233255461607e-06, 'learning_rate/action_model': 5.718233255461607e-05, 'epoch': 0.93}) +[2026-06-23 20:04:32,232][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.15887032449245453, 'train/grad_norm_pre_clip': 87.185791015625, 'timing/data': 0.0007171980105340481, 'timing/model': 1.6980396481230855, 'learning_rate/qwen_vl_interface': 5.715202570456486e-06, 'learning_rate/action_model': 5.715202570456486e-05, 'epoch': 0.93}) +[2026-06-23 20:04:35,574][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.09326649457216263, 'train/grad_norm_pre_clip': 19.402097702026367, 'timing/data': 0.0006988239474594593, 'timing/model': 1.6989412622060627, 'learning_rate/qwen_vl_interface': 5.7121716942241615e-06, 'learning_rate/action_model': 5.712171694224161e-05, 'epoch': 0.93}) +[2026-06-23 20:04:38,921][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.08723905682563782, 'train/grad_norm_pre_clip': 20.876771926879883, 'timing/data': 0.0006171709392219782, 'timing/model': 1.7002834139857441, 'learning_rate/qwen_vl_interface': 5.709140628010509e-06, 'learning_rate/action_model': 5.7091406280105085e-05, 'epoch': 0.93}) +[2026-06-23 20:04:42,267][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2400 +[2026-06-23 20:06:15,492][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2400 +[2026-06-23 20:06:15,493][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2400 +[2026-06-23 20:07:27,067][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2400 +[2026-06-23 20:07:27,068][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.11141914874315262, 'train/grad_norm_pre_clip': 55.8487548828125, 'eval/action_loss/seconds': 93.22309930017218, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.10734595358371735, 'eval/latency_2/loss': 0.10734595358371735, 'eval/flappy/loss': 0.10734595358371735, 'eval/flappy/latency_2/loss': 0.10734595358371735, 'eval/flap_precision': 0.9782608695652174, 'eval/flap_recall': 0.819838056680162, 'eval/flap_f1': 0.8920704845814977, 'eval/flap_rate_model': 0.1725, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.8920704845814977, 'eval/latency_2/flap_precision': 0.9782608695652174, 'eval/latency_2/flap_recall': 0.819838056680162, 'eval/latency_2/flap_f1': 0.8920704845814977, 'eval/latency_2/flap_rate_model': 0.1725, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.8920704845814977, 'eval/action_classification/data_wait_seconds': 0.7808969020843506, 'eval/action_classification/predict_seconds': 70.67652130126953, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0007628239691257477, 'timing/model': 1.700137719977647, 'learning_rate/qwen_vl_interface': 5.706109373061489e-06, 'learning_rate/action_model': 5.706109373061489e-05, 'epoch': 0.93}) +[2026-06-23 20:07:30,400][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.1570279598236084, 'train/grad_norm_pre_clip': 162.32376098632812, 'timing/data': 0.0005993971135467291, 'timing/model': 1.6919329459778965, 'learning_rate/qwen_vl_interface': 5.703077930623134e-06, 'learning_rate/action_model': 5.7030779306231344e-05, 'epoch': 0.93}) +[2026-06-23 20:07:33,727][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.20179414749145508, 'train/grad_norm_pre_clip': 80.19086456298828, 'timing/data': 0.000617241021245718, 'timing/model': 1.6879230558406562, 'learning_rate/qwen_vl_interface': 5.7000463019415565e-06, 'learning_rate/action_model': 5.7000463019415563e-05, 'epoch': 0.93}) +[2026-06-23 20:07:37,054][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.14743672311306, 'train/grad_norm_pre_clip': 95.95279693603516, 'timing/data': 0.0006208969280123711, 'timing/model': 1.690578417852521, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.93}) +[2026-06-23 20:07:40,385][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.15681138634681702, 'train/grad_norm_pre_clip': 65.49034118652344, 'timing/data': 0.0006079529412090778, 'timing/model': 1.692835820838809, 'learning_rate/qwen_vl_interface': 5.693982490833561e-06, 'learning_rate/action_model': 5.693982490833561e-05, 'epoch': 0.93}) +[2026-06-23 20:07:43,717][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.12251817435026169, 'train/grad_norm_pre_clip': 29.040714263916016, 'timing/data': 0.0006127920933067799, 'timing/model': 1.6931663418654352, 'learning_rate/qwen_vl_interface': 5.690950310899745e-06, 'learning_rate/action_model': 5.6909503108997456e-05, 'epoch': 0.93}) +[2026-06-23 20:07:47,052][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.09901194274425507, 'train/grad_norm_pre_clip': 11.386555671691895, 'timing/data': 0.0006775038782507181, 'timing/model': 1.6953168148174882, 'learning_rate/qwen_vl_interface': 5.687917949707913e-06, 'learning_rate/action_model': 5.687917949707913e-05, 'epoch': 0.93}) +[2026-06-23 20:07:50,389][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.1912619024515152, 'train/grad_norm_pre_clip': 30.603710174560547, 'timing/data': 0.0006075520068407059, 'timing/model': 1.6936191918794066, 'learning_rate/qwen_vl_interface': 5.684885408504555e-06, 'learning_rate/action_model': 5.684885408504554e-05, 'epoch': 0.93}) +[2026-06-23 20:07:53,725][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.21002568304538727, 'train/grad_norm_pre_clip': 49.508522033691406, 'timing/data': 0.0006220790091902018, 'timing/model': 1.694911491125822, 'learning_rate/qwen_vl_interface': 5.681852688536231e-06, 'learning_rate/action_model': 5.681852688536231e-05, 'epoch': 0.93}) +[2026-06-23 20:07:57,061][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.22802983224391937, 'train/grad_norm_pre_clip': 127.54486846923828, 'timing/data': 0.0006062500178813934, 'timing/model': 1.6938439740333706, 'learning_rate/qwen_vl_interface': 5.67881979104958e-06, 'learning_rate/action_model': 5.67881979104958e-05, 'epoch': 0.93}) +[2026-06-23 20:08:00,397][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.1749730259180069, 'train/grad_norm_pre_clip': 57.168617248535156, 'timing/data': 0.0006130828987807035, 'timing/model': 1.6950466739945114, 'learning_rate/qwen_vl_interface': 5.67578671729131e-06, 'learning_rate/action_model': 5.675786717291309e-05, 'epoch': 0.93}) +[2026-06-23 20:08:03,732][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.08024808019399643, 'train/grad_norm_pre_clip': 44.876373291015625, 'timing/data': 0.0006076321005821228, 'timing/model': 1.6980651679914445, 'learning_rate/qwen_vl_interface': 5.672753468508204e-06, 'learning_rate/action_model': 5.672753468508204e-05, 'epoch': 0.93}) +[2026-06-23 20:08:07,065][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.1003490462899208, 'train/grad_norm_pre_clip': 12.400888442993164, 'timing/data': 0.0006136537995189428, 'timing/model': 1.6916044598910958, 'learning_rate/qwen_vl_interface': 5.6697200459471165e-06, 'learning_rate/action_model': 5.669720045947117e-05, 'epoch': 0.93}) +[2026-06-23 20:08:10,400][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.12710286676883698, 'train/grad_norm_pre_clip': 18.9543514251709, 'timing/data': 0.0008711479604244232, 'timing/model': 1.695732390973717, 'learning_rate/qwen_vl_interface': 5.666686450854973e-06, 'learning_rate/action_model': 5.666686450854973e-05, 'epoch': 0.93}) +[2026-06-23 20:08:13,740][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.15057040750980377, 'train/grad_norm_pre_clip': 56.749942779541016, 'timing/data': 0.000603384105488658, 'timing/model': 1.6969632050022483, 'learning_rate/qwen_vl_interface': 5.663652684478768e-06, 'learning_rate/action_model': 5.6636526844787676e-05, 'epoch': 0.93}) +[2026-06-23 20:08:17,169][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.09583910554647446, 'train/grad_norm_pre_clip': 37.4039421081543, 'timing/data': 0.0006173700094223022, 'timing/model': 1.6946813431568444, 'learning_rate/qwen_vl_interface': 5.660618748065569e-06, 'learning_rate/action_model': 5.6606187480655695e-05, 'epoch': 0.93}) +[2026-06-23 20:08:20,513][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.15474721789360046, 'train/grad_norm_pre_clip': 57.91612243652344, 'timing/data': 0.0006208270788192749, 'timing/model': 1.697320254985243, 'learning_rate/qwen_vl_interface': 5.657584642862514e-06, 'learning_rate/action_model': 5.657584642862513e-05, 'epoch': 0.94}) +[2026-06-23 20:08:23,853][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.09186127036809921, 'train/grad_norm_pre_clip': 16.8859806060791, 'timing/data': 0.0006242739036679268, 'timing/model': 1.697718142066151, 'learning_rate/qwen_vl_interface': 5.654550370116808e-06, 'learning_rate/action_model': 5.654550370116808e-05, 'epoch': 0.94}) +[2026-06-23 20:08:27,201][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.1353434920310974, 'train/grad_norm_pre_clip': 23.703096389770508, 'timing/data': 0.0008614799007773399, 'timing/model': 1.7029794440604746, 'learning_rate/qwen_vl_interface': 5.651515931075726e-06, 'learning_rate/action_model': 5.651515931075726e-05, 'epoch': 0.94}) +[2026-06-23 20:08:30,546][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.1489982157945633, 'train/grad_norm_pre_clip': 51.61604309082031, 'timing/data': 0.0006227109115570784, 'timing/model': 1.699066726025194, 'learning_rate/qwen_vl_interface': 5.6484813269866126e-06, 'learning_rate/action_model': 5.648481326986612e-05, 'epoch': 0.94}) +[2026-06-23 20:08:33,891][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.16135719418525696, 'train/grad_norm_pre_clip': 45.3243293762207, 'timing/data': 0.000616488978266716, 'timing/model': 1.6971742969471961, 'learning_rate/qwen_vl_interface': 5.645446559096875e-06, 'learning_rate/action_model': 5.645446559096875e-05, 'epoch': 0.94}) +[2026-06-23 20:08:37,239][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.10881824046373367, 'train/grad_norm_pre_clip': 35.72079086303711, 'timing/data': 0.0006241139490157366, 'timing/model': 1.701089738169685, 'learning_rate/qwen_vl_interface': 5.6424116286539965e-06, 'learning_rate/action_model': 5.642411628653996e-05, 'epoch': 0.94}) +[2026-06-23 20:08:40,597][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.15618084371089935, 'train/grad_norm_pre_clip': 26.86005210876465, 'timing/data': 0.0006012599915266037, 'timing/model': 1.692752321017906, 'learning_rate/qwen_vl_interface': 5.639376536905517e-06, 'learning_rate/action_model': 5.6393765369055175e-05, 'epoch': 0.94}) +[2026-06-23 20:08:43,942][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.10293461382389069, 'train/grad_norm_pre_clip': 13.365544319152832, 'timing/data': 0.0005938669200986624, 'timing/model': 1.6994190560653806, 'learning_rate/qwen_vl_interface': 5.636341285099055e-06, 'learning_rate/action_model': 5.636341285099055e-05, 'epoch': 0.94}) +[2026-06-23 20:08:47,287][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.19952906668186188, 'train/grad_norm_pre_clip': 51.32857131958008, 'timing/data': 0.000616719014942646, 'timing/model': 1.6984176519326866, 'learning_rate/qwen_vl_interface': 5.6333058744822835e-06, 'learning_rate/action_model': 5.633305874482283e-05, 'epoch': 0.94}) +[2026-06-23 20:08:50,621][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.1403036266565323, 'train/grad_norm_pre_clip': 23.1205997467041, 'timing/data': 0.0006141841877251863, 'timing/model': 1.6920240600593388, 'learning_rate/qwen_vl_interface': 5.630270306302944e-06, 'learning_rate/action_model': 5.6302703063029436e-05, 'epoch': 0.94}) +[2026-06-23 20:08:53,974][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.171291321516037, 'train/grad_norm_pre_clip': 29.584959030151367, 'timing/data': 0.0006112901028245687, 'timing/model': 1.7101044340524822, 'learning_rate/qwen_vl_interface': 5.627234581808848e-06, 'learning_rate/action_model': 5.627234581808848e-05, 'epoch': 0.94}) +[2026-06-23 20:08:57,322][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.104981429874897, 'train/grad_norm_pre_clip': 22.92134666442871, 'timing/data': 0.0006332609336823225, 'timing/model': 1.696569716092199, 'learning_rate/qwen_vl_interface': 5.624198702247863e-06, 'learning_rate/action_model': 5.624198702247863e-05, 'epoch': 0.94}) +[2026-06-23 20:09:00,674][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.1127823144197464, 'train/grad_norm_pre_clip': 24.89087677001953, 'timing/data': 0.0008424639236181974, 'timing/model': 1.7026888409163803, 'learning_rate/qwen_vl_interface': 5.621162668867925e-06, 'learning_rate/action_model': 5.621162668867925e-05, 'epoch': 0.94}) +[2026-06-23 20:09:04,112][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.11020395159721375, 'train/grad_norm_pre_clip': 32.54998016357422, 'timing/data': 0.0006183020304888487, 'timing/model': 1.7907615250442177, 'learning_rate/qwen_vl_interface': 5.618126482917037e-06, 'learning_rate/action_model': 5.618126482917036e-05, 'epoch': 0.94}) +[2026-06-23 20:09:07,461][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.109378382563591, 'train/grad_norm_pre_clip': 20.662046432495117, 'timing/data': 0.0008875890634953976, 'timing/model': 1.7015757670160383, 'learning_rate/qwen_vl_interface': 5.6150901456432546e-06, 'learning_rate/action_model': 5.615090145643255e-05, 'epoch': 0.94}) +[2026-06-23 20:09:10,808][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.0836087092757225, 'train/grad_norm_pre_clip': 40.057456970214844, 'timing/data': 0.0006092861294746399, 'timing/model': 1.7023866609670222, 'learning_rate/qwen_vl_interface': 5.612053658294705e-06, 'learning_rate/action_model': 5.6120536582947045e-05, 'epoch': 0.94}) +[2026-06-23 20:09:14,158][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.09978057444095612, 'train/grad_norm_pre_clip': 57.482730865478516, 'timing/data': 0.0006946159992367029, 'timing/model': 1.6971332218963653, 'learning_rate/qwen_vl_interface': 5.609017022119572e-06, 'learning_rate/action_model': 5.609017022119573e-05, 'epoch': 0.94}) +[2026-06-23 20:09:17,507][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.03879569470882416, 'train/grad_norm_pre_clip': 12.285836219787598, 'timing/data': 0.0006052178796380758, 'timing/model': 1.7011011010035872, 'learning_rate/qwen_vl_interface': 5.605980238366104e-06, 'learning_rate/action_model': 5.605980238366103e-05, 'epoch': 0.94}) +[2026-06-23 20:09:20,846][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.09730426967144012, 'train/grad_norm_pre_clip': 18.6088924407959, 'timing/data': 0.0006665629334747791, 'timing/model': 1.6957480260170996, 'learning_rate/qwen_vl_interface': 5.602943308282606e-06, 'learning_rate/action_model': 5.602943308282605e-05, 'epoch': 0.94}) +[2026-06-23 20:09:24,186][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.1660158485174179, 'train/grad_norm_pre_clip': 32.89042282104492, 'timing/data': 0.000617732061073184, 'timing/model': 1.698790717171505, 'learning_rate/qwen_vl_interface': 5.599906233117445e-06, 'learning_rate/action_model': 5.5999062331174446e-05, 'epoch': 0.94}) +[2026-06-23 20:09:27,521][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.08865945786237717, 'train/grad_norm_pre_clip': 11.367202758789062, 'timing/data': 0.000618683872744441, 'timing/model': 1.694080889923498, 'learning_rate/qwen_vl_interface': 5.596869014119048e-06, 'learning_rate/action_model': 5.596869014119048e-05, 'epoch': 0.94}) +[2026-06-23 20:09:30,861][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.15521644055843353, 'train/grad_norm_pre_clip': 72.11548614501953, 'timing/data': 0.0008203380275517702, 'timing/model': 1.697392089990899, 'learning_rate/qwen_vl_interface': 5.593831652535902e-06, 'learning_rate/action_model': 5.593831652535902e-05, 'epoch': 0.94}) +[2026-06-23 20:09:34,198][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.16970504820346832, 'train/grad_norm_pre_clip': 92.30465698242188, 'timing/data': 0.0006021999288350344, 'timing/model': 1.6943980681244284, 'learning_rate/qwen_vl_interface': 5.590794149616552e-06, 'learning_rate/action_model': 5.590794149616552e-05, 'epoch': 0.94}) +[2026-06-23 20:09:37,539][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.09925732016563416, 'train/grad_norm_pre_clip': 17.595726013183594, 'timing/data': 0.0006138221360743046, 'timing/model': 1.6966031179763377, 'learning_rate/qwen_vl_interface': 5.587756506609601e-06, 'learning_rate/action_model': 5.5877565066096006e-05, 'epoch': 0.94}) +[2026-06-23 20:09:40,882][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.26141688227653503, 'train/grad_norm_pre_clip': 34.74982833862305, 'timing/data': 0.0006265351548790932, 'timing/model': 1.697839530883357, 'learning_rate/qwen_vl_interface': 5.584718724763707e-06, 'learning_rate/action_model': 5.584718724763707e-05, 'epoch': 0.94}) +[2026-06-23 20:09:44,228][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.11473681777715683, 'train/grad_norm_pre_clip': 20.154306411743164, 'timing/data': 0.0006122889462858438, 'timing/model': 1.6975968489423394, 'learning_rate/qwen_vl_interface': 5.581680805327588e-06, 'learning_rate/action_model': 5.5816808053275885e-05, 'epoch': 0.94}) +[2026-06-23 20:09:47,563][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.113763228058815, 'train/grad_norm_pre_clip': 17.00766372680664, 'timing/data': 0.0006287989672273397, 'timing/model': 1.69381517986767, 'learning_rate/qwen_vl_interface': 5.578642749550019e-06, 'learning_rate/action_model': 5.578642749550019e-05, 'epoch': 0.95}) +[2026-06-23 20:09:50,905][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.10754822194576263, 'train/grad_norm_pre_clip': 14.970296859741211, 'timing/data': 0.000651773065328598, 'timing/model': 1.6962097641080618, 'learning_rate/qwen_vl_interface': 5.575604558679831e-06, 'learning_rate/action_model': 5.57560455867983e-05, 'epoch': 0.95}) +[2026-06-23 20:09:54,338][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.0842113345861435, 'train/grad_norm_pre_clip': 14.068349838256836, 'timing/data': 0.0006194119341671467, 'timing/model': 1.6970424798782915, 'learning_rate/qwen_vl_interface': 5.5725662339659085e-06, 'learning_rate/action_model': 5.572566233965908e-05, 'epoch': 0.95}) +[2026-06-23 20:09:57,685][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.10897308588027954, 'train/grad_norm_pre_clip': 61.16147232055664, 'timing/data': 0.0006257041823118925, 'timing/model': 1.69678315683268, 'learning_rate/qwen_vl_interface': 5.56952777665719e-06, 'learning_rate/action_model': 5.56952777665719e-05, 'epoch': 0.95}) +[2026-06-23 20:10:01,024][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.0634259507060051, 'train/grad_norm_pre_clip': 48.79075241088867, 'timing/data': 0.0006015079561620951, 'timing/model': 1.694936202839017, 'learning_rate/qwen_vl_interface': 5.566489188002671e-06, 'learning_rate/action_model': 5.56648918800267e-05, 'epoch': 0.95}) +[2026-06-23 20:10:04,366][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.17392566800117493, 'train/grad_norm_pre_clip': 23.183305740356445, 'timing/data': 0.0007471418939530849, 'timing/model': 1.6949135840404779, 'learning_rate/qwen_vl_interface': 5.563450469251403e-06, 'learning_rate/action_model': 5.563450469251402e-05, 'epoch': 0.95}) +[2026-06-23 20:10:07,706][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.14524926245212555, 'train/grad_norm_pre_clip': 17.33597755432129, 'timing/data': 0.0006613300647586584, 'timing/model': 1.6949969020206481, 'learning_rate/qwen_vl_interface': 5.560411621652488e-06, 'learning_rate/action_model': 5.560411621652488e-05, 'epoch': 0.95}) +[2026-06-23 20:10:11,048][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.2045074701309204, 'train/grad_norm_pre_clip': 31.793533325195312, 'timing/data': 0.000939943129196763, 'timing/model': 1.6951034320518374, 'learning_rate/qwen_vl_interface': 5.557372646455081e-06, 'learning_rate/action_model': 5.55737264645508e-05, 'epoch': 0.95}) +[2026-06-23 20:10:14,386][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.1859298050403595, 'train/grad_norm_pre_clip': 14.992171287536621, 'timing/data': 0.0006039028521627188, 'timing/model': 1.6978925860021263, 'learning_rate/qwen_vl_interface': 5.554333544908388e-06, 'learning_rate/action_model': 5.5543335449083875e-05, 'epoch': 0.95}) +[2026-06-23 20:10:17,726][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.2206941395998001, 'train/grad_norm_pre_clip': 27.36467742919922, 'timing/data': 0.000606197165325284, 'timing/model': 1.697037602076307, 'learning_rate/qwen_vl_interface': 5.551294318261671e-06, 'learning_rate/action_model': 5.5512943182616705e-05, 'epoch': 0.95}) +[2026-06-23 20:10:21,067][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.29827040433883667, 'train/grad_norm_pre_clip': 25.928447723388672, 'timing/data': 0.000617949990555644, 'timing/model': 1.697179437847808, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.95}) +[2026-06-23 20:10:24,413][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.051129646599292755, 'train/grad_norm_pre_clip': 49.512367248535156, 'timing/data': 0.0006176279857754707, 'timing/model': 1.6984312769491225, 'learning_rate/qwen_vl_interface': 5.545215494665458e-06, 'learning_rate/action_model': 5.545215494665458e-05, 'epoch': 0.95}) +[2026-06-23 20:10:27,760][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.16076993942260742, 'train/grad_norm_pre_clip': 23.709251403808594, 'timing/data': 0.0006059869192540646, 'timing/model': 1.7008412689901888, 'learning_rate/qwen_vl_interface': 5.5421759002147376e-06, 'learning_rate/action_model': 5.542175900214738e-05, 'epoch': 0.95}) +[2026-06-23 20:10:31,110][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.16298907995224, 'train/grad_norm_pre_clip': 25.136241912841797, 'timing/data': 0.0006134707946330309, 'timing/model': 1.7023910949938, 'learning_rate/qwen_vl_interface': 5.53913618566154e-06, 'learning_rate/action_model': 5.53913618566154e-05, 'epoch': 0.95}) +[2026-06-23 20:10:34,463][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.24434512853622437, 'train/grad_norm_pre_clip': 31.194494247436523, 'timing/data': 0.0008762138895690441, 'timing/model': 1.701367445057258, 'learning_rate/qwen_vl_interface': 5.536096352255381e-06, 'learning_rate/action_model': 5.536096352255381e-05, 'epoch': 0.95}) +[2026-06-23 20:10:37,809][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.11357128620147705, 'train/grad_norm_pre_clip': 48.54438781738281, 'timing/data': 0.0006138819735497236, 'timing/model': 1.6995599039364606, 'learning_rate/qwen_vl_interface': 5.533056401245817e-06, 'learning_rate/action_model': 5.5330564012458164e-05, 'epoch': 0.95}) +[2026-06-23 20:10:41,155][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.12256836891174316, 'train/grad_norm_pre_clip': 63.659664154052734, 'timing/data': 0.0006066979840397835, 'timing/model': 1.6988451411016285, 'learning_rate/qwen_vl_interface': 5.530016333882456e-06, 'learning_rate/action_model': 5.5300163338824565e-05, 'epoch': 0.95}) +[2026-06-23 20:10:44,657][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.29505664110183716, 'train/grad_norm_pre_clip': 49.76553726196289, 'timing/data': 0.0008411980234086514, 'timing/model': 1.6972517240792513, 'learning_rate/qwen_vl_interface': 5.52697615141496e-06, 'learning_rate/action_model': 5.52697615141496e-05, 'epoch': 0.95}) +[2026-06-23 20:10:48,003][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.19003354012966156, 'train/grad_norm_pre_clip': 66.44686126708984, 'timing/data': 0.0006079711019992828, 'timing/model': 1.6969111419748515, 'learning_rate/qwen_vl_interface': 5.52393585509303e-06, 'learning_rate/action_model': 5.52393585509303e-05, 'epoch': 0.95}) +[2026-06-23 20:10:51,342][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.149341881275177, 'train/grad_norm_pre_clip': 19.044967651367188, 'timing/data': 0.0006082011386752129, 'timing/model': 1.6959817770402879, 'learning_rate/qwen_vl_interface': 5.520895446166418e-06, 'learning_rate/action_model': 5.5208954461664175e-05, 'epoch': 0.95}) +[2026-06-23 20:10:54,675][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.15984207391738892, 'train/grad_norm_pre_clip': 17.777753829956055, 'timing/data': 0.0006082509644329548, 'timing/model': 1.6916564621496946, 'learning_rate/qwen_vl_interface': 5.517854925884919e-06, 'learning_rate/action_model': 5.517854925884919e-05, 'epoch': 0.95}) +[2026-06-23 20:10:58,025][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.20691655576229095, 'train/grad_norm_pre_clip': 13.860527992248535, 'timing/data': 0.0005953270010650158, 'timing/model': 1.7007160941138864, 'learning_rate/qwen_vl_interface': 5.514814295498379e-06, 'learning_rate/action_model': 5.514814295498378e-05, 'epoch': 0.95}) +[2026-06-23 20:11:01,375][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.13466288149356842, 'train/grad_norm_pre_clip': 35.70231246948242, 'timing/data': 0.0006063180044293404, 'timing/model': 1.7012819440569729, 'learning_rate/qwen_vl_interface': 5.5117735562566846e-06, 'learning_rate/action_model': 5.5117735562566844e-05, 'epoch': 0.95}) +[2026-06-23 20:11:04,719][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.10628495365381241, 'train/grad_norm_pre_clip': 14.476548194885254, 'timing/data': 0.0007248700130730867, 'timing/model': 1.699574199039489, 'learning_rate/qwen_vl_interface': 5.5087327094097685e-06, 'learning_rate/action_model': 5.508732709409769e-05, 'epoch': 0.95}) +[2026-06-23 20:11:08,063][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.11558359116315842, 'train/grad_norm_pre_clip': 19.88298225402832, 'timing/data': 0.0006211448926478624, 'timing/model': 1.6976032969541848, 'learning_rate/qwen_vl_interface': 5.505691756207609e-06, 'learning_rate/action_model': 5.505691756207609e-05, 'epoch': 0.95}) +[2026-06-23 20:11:11,415][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.07406846433877945, 'train/grad_norm_pre_clip': 16.147994995117188, 'timing/data': 0.0008136369287967682, 'timing/model': 1.7033400170039386, 'learning_rate/qwen_vl_interface': 5.502650697900228e-06, 'learning_rate/action_model': 5.502650697900228e-05, 'epoch': 0.95}) +[2026-06-23 20:11:14,766][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.08919492363929749, 'train/grad_norm_pre_clip': 13.199044227600098, 'timing/data': 0.0007484741508960724, 'timing/model': 1.7001940910704434, 'learning_rate/qwen_vl_interface': 5.499609535737688e-06, 'learning_rate/action_model': 5.4996095357376875e-05, 'epoch': 0.96}) +[2026-06-23 20:11:18,105][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.19614656269550323, 'train/grad_norm_pre_clip': 45.6016845703125, 'timing/data': 0.0006091229151934385, 'timing/model': 1.697246507043019, 'learning_rate/qwen_vl_interface': 5.496568270970096e-06, 'learning_rate/action_model': 5.496568270970096e-05, 'epoch': 0.96}) +[2026-06-23 20:11:21,451][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.16895553469657898, 'train/grad_norm_pre_clip': 16.854984283447266, 'timing/data': 0.0006226291880011559, 'timing/model': 1.7017539208754897, 'learning_rate/qwen_vl_interface': 5.4935269048476015e-06, 'learning_rate/action_model': 5.493526904847601e-05, 'epoch': 0.96}) +[2026-06-23 20:11:24,795][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.2283439338207245, 'train/grad_norm_pre_clip': 18.167699813842773, 'timing/data': 0.0006078709848225117, 'timing/model': 1.6980408700183034, 'learning_rate/qwen_vl_interface': 5.490485438620397e-06, 'learning_rate/action_model': 5.4904854386203965e-05, 'epoch': 0.96}) +[2026-06-23 20:11:28,141][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.18131054937839508, 'train/grad_norm_pre_clip': 62.28154754638672, 'timing/data': 0.000597361009567976, 'timing/model': 1.6987174649257213, 'learning_rate/qwen_vl_interface': 5.487443873538712e-06, 'learning_rate/action_model': 5.4874438735387115e-05, 'epoch': 0.96}) +[2026-06-23 20:11:31,617][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.22765865921974182, 'train/grad_norm_pre_clip': 53.676265716552734, 'timing/data': 0.0008537508547306061, 'timing/model': 1.8277487731538713, 'learning_rate/qwen_vl_interface': 5.484402210852819e-06, 'learning_rate/action_model': 5.484402210852818e-05, 'epoch': 0.96}) +[2026-06-23 20:11:34,965][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.13079890608787537, 'train/grad_norm_pre_clip': 49.41604995727539, 'timing/data': 0.0006248729769140482, 'timing/model': 1.6977341619785875, 'learning_rate/qwen_vl_interface': 5.48136045181303e-06, 'learning_rate/action_model': 5.48136045181303e-05, 'epoch': 0.96}) +[2026-06-23 20:11:38,306][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.15207995474338531, 'train/grad_norm_pre_clip': 35.00947952270508, 'timing/data': 0.0006282490212470293, 'timing/model': 1.6935870668385178, 'learning_rate/qwen_vl_interface': 5.478318597669699e-06, 'learning_rate/action_model': 5.4783185976696995e-05, 'epoch': 0.96}) +[2026-06-23 20:11:41,647][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.1260797083377838, 'train/grad_norm_pre_clip': 32.29153060913086, 'timing/data': 0.0006128000095486641, 'timing/model': 1.6940397510770708, 'learning_rate/qwen_vl_interface': 5.475276649673218e-06, 'learning_rate/action_model': 5.4752766496732174e-05, 'epoch': 0.96}) +[2026-06-23 20:11:44,984][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.1377115249633789, 'train/grad_norm_pre_clip': 30.581968307495117, 'timing/data': 0.0006233900785446167, 'timing/model': 1.695314907003194, 'learning_rate/qwen_vl_interface': 5.472234609074014e-06, 'learning_rate/action_model': 5.472234609074014e-05, 'epoch': 0.96}) +[2026-06-23 20:11:48,331][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.16376693546772003, 'train/grad_norm_pre_clip': 18.29889488220215, 'timing/data': 0.0008332640863955021, 'timing/model': 1.6978104838635772, 'learning_rate/qwen_vl_interface': 5.46919247712256e-06, 'learning_rate/action_model': 5.46919247712256e-05, 'epoch': 0.96}) +[2026-06-23 20:11:51,680][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.09436661750078201, 'train/grad_norm_pre_clip': 40.399620056152344, 'timing/data': 0.0006138219032436609, 'timing/model': 1.7006214528810233, 'learning_rate/qwen_vl_interface': 5.466150255069356e-06, 'learning_rate/action_model': 5.4661502550693555e-05, 'epoch': 0.96}) +[2026-06-23 20:11:55,027][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.2049773931503296, 'train/grad_norm_pre_clip': 66.45590209960938, 'timing/data': 0.0006227891426533461, 'timing/model': 1.7002231180667877, 'learning_rate/qwen_vl_interface': 5.463107944164947e-06, 'learning_rate/action_model': 5.4631079441649466e-05, 'epoch': 0.96}) +[2026-06-23 20:11:58,370][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.15782667696475983, 'train/grad_norm_pre_clip': 14.250722885131836, 'timing/data': 0.0006614311132580042, 'timing/model': 1.6947591181378812, 'learning_rate/qwen_vl_interface': 5.46006554565991e-06, 'learning_rate/action_model': 5.4600655456599105e-05, 'epoch': 0.96}) +[2026-06-23 20:12:01,712][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.09211380779743195, 'train/grad_norm_pre_clip': 12.398896217346191, 'timing/data': 0.000659167068079114, 'timing/model': 1.693746262928471, 'learning_rate/qwen_vl_interface': 5.457023060804865e-06, 'learning_rate/action_model': 5.457023060804864e-05, 'epoch': 0.96}) +[2026-06-23 20:12:05,055][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.17424209415912628, 'train/grad_norm_pre_clip': 12.368478775024414, 'timing/data': 0.0008247180376201868, 'timing/model': 1.6978128559421748, 'learning_rate/qwen_vl_interface': 5.453980490850459e-06, 'learning_rate/action_model': 5.453980490850459e-05, 'epoch': 0.96}) +[2026-06-23 20:12:08,396][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.09102728217840195, 'train/grad_norm_pre_clip': 18.906776428222656, 'timing/data': 0.0007537549827247858, 'timing/model': 1.6926355948671699, 'learning_rate/qwen_vl_interface': 5.4509378370473785e-06, 'learning_rate/action_model': 5.450937837047378e-05, 'epoch': 0.96}) +[2026-06-23 20:12:11,736][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.08727575838565826, 'train/grad_norm_pre_clip': 33.00619125366211, 'timing/data': 0.0007297089323401451, 'timing/model': 1.6953996641095728, 'learning_rate/qwen_vl_interface': 5.4478951006463424e-06, 'learning_rate/action_model': 5.447895100646342e-05, 'epoch': 0.96}) +[2026-06-23 20:12:15,081][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.2816751003265381, 'train/grad_norm_pre_clip': 78.94329071044922, 'timing/data': 0.0006309838499873877, 'timing/model': 1.6989348891656846, 'learning_rate/qwen_vl_interface': 5.444852282898106e-06, 'learning_rate/action_model': 5.444852282898106e-05, 'epoch': 0.96}) +[2026-06-23 20:12:18,425][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.10257841646671295, 'train/grad_norm_pre_clip': 36.158164978027344, 'timing/data': 0.0006117480807006359, 'timing/model': 1.70077088708058, 'learning_rate/qwen_vl_interface': 5.441809385053456e-06, 'learning_rate/action_model': 5.4418093850534555e-05, 'epoch': 0.96}) +[2026-06-23 20:12:21,862][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.09862121194601059, 'train/grad_norm_pre_clip': 14.941667556762695, 'timing/data': 0.0005935737863183022, 'timing/model': 1.792725034058094, 'learning_rate/qwen_vl_interface': 5.438766408363214e-06, 'learning_rate/action_model': 5.438766408363214e-05, 'epoch': 0.96}) +[2026-06-23 20:12:25,210][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.11819540709257126, 'train/grad_norm_pre_clip': 22.24146270751953, 'timing/data': 0.000619472935795784, 'timing/model': 1.698673966806382, 'learning_rate/qwen_vl_interface': 5.43572335407823e-06, 'learning_rate/action_model': 5.43572335407823e-05, 'epoch': 0.96}) +[2026-06-23 20:12:28,556][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.10489910840988159, 'train/grad_norm_pre_clip': 26.131420135498047, 'timing/data': 0.0006123990751802921, 'timing/model': 1.7010495071299374, 'learning_rate/qwen_vl_interface': 5.432680223449392e-06, 'learning_rate/action_model': 5.4326802234493924e-05, 'epoch': 0.96}) +[2026-06-23 20:12:31,906][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.15473076701164246, 'train/grad_norm_pre_clip': 14.842156410217285, 'timing/data': 0.0006173481233417988, 'timing/model': 1.6997683849185705, 'learning_rate/qwen_vl_interface': 5.429637017727614e-06, 'learning_rate/action_model': 5.4296370177276144e-05, 'epoch': 0.96}) +[2026-06-23 20:12:35,244][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.0656031146645546, 'train/grad_norm_pre_clip': 14.061171531677246, 'timing/data': 0.0006095040589570999, 'timing/model': 1.6936195869930089, 'learning_rate/qwen_vl_interface': 5.426593738163845e-06, 'learning_rate/action_model': 5.4265937381638446e-05, 'epoch': 0.96}) +[2026-06-23 20:12:38,581][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.11246883124113083, 'train/grad_norm_pre_clip': 28.063377380371094, 'timing/data': 0.0006210058927536011, 'timing/model': 1.6940396970603615, 'learning_rate/qwen_vl_interface': 5.42355038600906e-06, 'learning_rate/action_model': 5.42355038600906e-05, 'epoch': 0.96}) +[2026-06-23 20:12:41,924][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.18135832250118256, 'train/grad_norm_pre_clip': 25.586923599243164, 'timing/data': 0.0006268960423767567, 'timing/model': 1.696719966828823, 'learning_rate/qwen_vl_interface': 5.420506962514267e-06, 'learning_rate/action_model': 5.420506962514267e-05, 'epoch': 0.97}) +[2026-06-23 20:12:45,263][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.09839614480733871, 'train/grad_norm_pre_clip': 22.52423095703125, 'timing/data': 0.0006110670510679483, 'timing/model': 1.6931748148053885, 'learning_rate/qwen_vl_interface': 5.4174634689305015e-06, 'learning_rate/action_model': 5.417463468930501e-05, 'epoch': 0.97}) +[2026-06-23 20:12:48,611][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.10709626972675323, 'train/grad_norm_pre_clip': 17.843847274780273, 'timing/data': 0.0006159760523587465, 'timing/model': 1.7066817940212786, 'learning_rate/qwen_vl_interface': 5.41441990650883e-06, 'learning_rate/action_model': 5.4144199065088295e-05, 'epoch': 0.97}) +[2026-06-23 20:12:51,957][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.15810257196426392, 'train/grad_norm_pre_clip': 58.247554779052734, 'timing/data': 0.0006220180075615644, 'timing/model': 1.7009507871698588, 'learning_rate/qwen_vl_interface': 5.411376276500345e-06, 'learning_rate/action_model': 5.411376276500345e-05, 'epoch': 0.97}) +[2026-06-23 20:12:55,302][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.07942074537277222, 'train/grad_norm_pre_clip': 29.554035186767578, 'timing/data': 0.0006117781158536673, 'timing/model': 1.700241015991196, 'learning_rate/qwen_vl_interface': 5.408332580156168e-06, 'learning_rate/action_model': 5.408332580156168e-05, 'epoch': 0.97}) +[2026-06-23 20:12:58,650][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.12743137776851654, 'train/grad_norm_pre_clip': 16.423099517822266, 'timing/data': 0.000611176947131753, 'timing/model': 1.699958726996556, 'learning_rate/qwen_vl_interface': 5.40528881872745e-06, 'learning_rate/action_model': 5.40528881872745e-05, 'epoch': 0.97}) +[2026-06-23 20:13:01,996][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2500 +[2026-06-23 20:14:35,262][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2500 +[2026-06-23 20:14:35,263][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2500 +[2026-06-23 20:15:47,101][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2500 +[2026-06-23 20:16:50,017][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500 +[2026-06-23 20:16:50,017][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.17154650390148163, 'train/grad_norm_pre_clip': 13.54828929901123, 'eval/action_loss/seconds': 93.26469749794342, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07915365695953369, 'eval/latency_2/loss': 0.07915365695953369, 'eval/flappy/loss': 0.07915365695953369, 'eval/flappy/latency_2/loss': 0.07915365695953369, 'eval/flap_precision': 0.8915360501567398, 'eval/flap_recall': 0.9595141700404858, 'eval/flap_f1': 0.9242768930776729, 'eval/flap_rate_model': 0.22152777777777777, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.924276893077673, 'eval/latency_2/flap_precision': 0.8915360501567398, 'eval/latency_2/flap_recall': 0.9595141700404858, 'eval/latency_2/flap_f1': 0.9242768930776729, 'eval/latency_2/flap_rate_model': 0.22152777777777777, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.924276893077673, 'eval/action_classification/data_wait_seconds': 0.7933933138847351, 'eval/action_classification/predict_seconds': 70.89595794677734, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 45.79999999999991, 'rl_games_eval/mid_train/mean_length': 388.0, 'rl_games_eval/mid_train/std_reward': 33.14423026712149, 'rl_games_eval/mid_train/std_length': 267.54887403986584, 'rl_games_eval/mid_train/macro_mean_reward': 45.79999999999991, 'rl_games_eval/mid_train/macro_mean_length': 388.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 45.79999999999991, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 388.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 33.14423026712149, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 267.54887403986584, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 45.79999999999991, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000628218986093998, 'timing/model': 1.6992561069782823, 'learning_rate/qwen_vl_interface': 5.402244993465364e-06, 'learning_rate/action_model': 5.402244993465364e-05, 'epoch': 0.97}) +[2026-06-23 20:16:50,018][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_2500_state +[2026-06-23 20:16:50,020][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 20:16:52,658][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_2500_state/model.safetensors +[2026-06-23 20:16:53,841][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_2500_state/optimizer.bin +[2026-06-23 20:16:53,841][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_2500_state/sampler.bin +[2026-06-23 20:16:53,841][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_2500_state/sampler_1.bin +[2026-06-23 20:16:53,842][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_2500_state/random_states_0.pkl +[2026-06-23 20:16:53,879][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500 +[2026-06-23 20:16:53,879][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 20:16:53,890][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 20:16:57,193][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.1027316078543663, 'train/grad_norm_pre_clip': 22.27585220336914, 'timing/data': 0.0006070199888199568, 'timing/model': 1.6728170709684491, 'learning_rate/qwen_vl_interface': 5.399201105621111e-06, 'learning_rate/action_model': 5.39920110562111e-05, 'epoch': 0.97}) +[2026-06-23 20:17:00,486][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.1093004047870636, 'train/grad_norm_pre_clip': 38.94613265991211, 'timing/data': 0.0006153958383947611, 'timing/model': 1.6723922919481993, 'learning_rate/qwen_vl_interface': 5.396157156445918e-06, 'learning_rate/action_model': 5.396157156445918e-05, 'epoch': 0.97}) +[2026-06-23 20:17:03,797][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.13690008223056793, 'train/grad_norm_pre_clip': 60.86284255981445, 'timing/data': 0.000625835033133626, 'timing/model': 1.6783493049442768, 'learning_rate/qwen_vl_interface': 5.39311314719104e-06, 'learning_rate/action_model': 5.39311314719104e-05, 'epoch': 0.97}) +[2026-06-23 20:17:07,100][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.20098674297332764, 'train/grad_norm_pre_clip': 43.59029006958008, 'timing/data': 0.0006122798658907413, 'timing/model': 1.6807929398491979, 'learning_rate/qwen_vl_interface': 5.390069079107755e-06, 'learning_rate/action_model': 5.390069079107755e-05, 'epoch': 0.97}) +[2026-06-23 20:17:10,405][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.17444902658462524, 'train/grad_norm_pre_clip': 60.7826042175293, 'timing/data': 0.0006002271547913551, 'timing/model': 1.6809752210974693, 'learning_rate/qwen_vl_interface': 5.387024953447361e-06, 'learning_rate/action_model': 5.387024953447361e-05, 'epoch': 0.97}) +[2026-06-23 20:17:13,719][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.050639692693948746, 'train/grad_norm_pre_clip': 26.294658660888672, 'timing/data': 0.0006802969146519899, 'timing/model': 1.6827740969602019, 'learning_rate/qwen_vl_interface': 5.383980771461184e-06, 'learning_rate/action_model': 5.3839807714611846e-05, 'epoch': 0.97}) +[2026-06-23 20:17:17,042][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.04511403292417526, 'train/grad_norm_pre_clip': 27.321598052978516, 'timing/data': 0.000615074997767806, 'timing/model': 1.6900630521122366, 'learning_rate/qwen_vl_interface': 5.380936534400572e-06, 'learning_rate/action_model': 5.3809365344005726e-05, 'epoch': 0.97}) +[2026-06-23 20:17:20,689][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.29061123728752136, 'train/grad_norm_pre_clip': 24.609134674072266, 'timing/data': 0.0006615619640797377, 'timing/model': 2.0154912180732936, 'learning_rate/qwen_vl_interface': 5.377892243516899e-06, 'learning_rate/action_model': 5.3778922435168985e-05, 'epoch': 0.97}) +[2026-06-23 20:17:24,050][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.20944640040397644, 'train/grad_norm_pre_clip': 92.32030487060547, 'timing/data': 0.0006141338963061571, 'timing/model': 1.726052108919248, 'learning_rate/qwen_vl_interface': 5.3748479000615525e-06, 'learning_rate/action_model': 5.374847900061553e-05, 'epoch': 0.97}) +[2026-06-23 20:17:27,378][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.14789535105228424, 'train/grad_norm_pre_clip': 64.65473175048828, 'timing/data': 0.0006239220965653658, 'timing/model': 1.6931240539997816, 'learning_rate/qwen_vl_interface': 5.371803505285954e-06, 'learning_rate/action_model': 5.371803505285954e-05, 'epoch': 0.97}) +[2026-06-23 20:17:30,695][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.31482797861099243, 'train/grad_norm_pre_clip': 111.63629913330078, 'timing/data': 0.0006117289885878563, 'timing/model': 1.6830128659494221, 'learning_rate/qwen_vl_interface': 5.368759060441534e-06, 'learning_rate/action_model': 5.368759060441534e-05, 'epoch': 0.97}) +[2026-06-23 20:17:34,017][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.10151398926973343, 'train/grad_norm_pre_clip': 16.9891414642334, 'timing/data': 0.0006047149654477835, 'timing/model': 1.6842515268363059, 'learning_rate/qwen_vl_interface': 5.365714566779749e-06, 'learning_rate/action_model': 5.3657145667797484e-05, 'epoch': 0.97}) +[2026-06-23 20:17:37,343][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.08984783291816711, 'train/grad_norm_pre_clip': 44.24818801879883, 'timing/data': 0.0006801472045481205, 'timing/model': 1.6897949050180614, 'learning_rate/qwen_vl_interface': 5.362670025552077e-06, 'learning_rate/action_model': 5.362670025552077e-05, 'epoch': 0.97}) +[2026-06-23 20:17:40,673][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.07912780344486237, 'train/grad_norm_pre_clip': 43.693668365478516, 'timing/data': 0.0006329179741442204, 'timing/model': 1.692690348951146, 'learning_rate/qwen_vl_interface': 5.359625438010013e-06, 'learning_rate/action_model': 5.359625438010013e-05, 'epoch': 0.97}) +[2026-06-23 20:17:44,002][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.14597134292125702, 'train/grad_norm_pre_clip': 21.343778610229492, 'timing/data': 0.0006158370524644852, 'timing/model': 1.6905378031078726, 'learning_rate/qwen_vl_interface': 5.356580805405075e-06, 'learning_rate/action_model': 5.356580805405075e-05, 'epoch': 0.97}) +[2026-06-23 20:17:47,328][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.14876511693000793, 'train/grad_norm_pre_clip': 73.42377471923828, 'timing/data': 0.0006083720363676548, 'timing/model': 1.6885697548277676, 'learning_rate/qwen_vl_interface': 5.353536128988791e-06, 'learning_rate/action_model': 5.353536128988791e-05, 'epoch': 0.97}) +[2026-06-23 20:17:50,666][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.14773739874362946, 'train/grad_norm_pre_clip': 86.56005859375, 'timing/data': 0.0006162570789456367, 'timing/model': 1.6988783900160342, 'learning_rate/qwen_vl_interface': 5.350491410012717e-06, 'learning_rate/action_model': 5.350491410012717e-05, 'epoch': 0.97}) +[2026-06-23 20:17:54,008][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.16695427894592285, 'train/grad_norm_pre_clip': 82.75932312011719, 'timing/data': 0.0008821368683129549, 'timing/model': 1.6955434242263436, 'learning_rate/qwen_vl_interface': 5.347446649728419e-06, 'learning_rate/action_model': 5.347446649728418e-05, 'epoch': 0.97}) +[2026-06-23 20:17:57,343][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.1278838813304901, 'train/grad_norm_pre_clip': 51.69712829589844, 'timing/data': 0.0006116479635238647, 'timing/model': 1.6961433768738061, 'learning_rate/qwen_vl_interface': 5.344401849387484e-06, 'learning_rate/action_model': 5.344401849387484e-05, 'epoch': 0.97}) +[2026-06-23 20:18:00,672][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.22890141606330872, 'train/grad_norm_pre_clip': 56.41332244873047, 'timing/data': 0.0006212771404534578, 'timing/model': 1.6887812479399145, 'learning_rate/qwen_vl_interface': 5.3413570102415144e-06, 'learning_rate/action_model': 5.3413570102415144e-05, 'epoch': 0.98}) +[2026-06-23 20:18:04,004][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.08538645505905151, 'train/grad_norm_pre_clip': 15.136552810668945, 'timing/data': 0.0006040539592504501, 'timing/model': 1.6951876250095665, 'learning_rate/qwen_vl_interface': 5.338312133542131e-06, 'learning_rate/action_model': 5.3383121335421305e-05, 'epoch': 0.98}) +[2026-06-23 20:18:07,464][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.10883327573537827, 'train/grad_norm_pre_clip': 53.88139343261719, 'timing/data': 0.0005983829032629728, 'timing/model': 1.820995711022988, 'learning_rate/qwen_vl_interface': 5.335267220540963e-06, 'learning_rate/action_model': 5.335267220540963e-05, 'epoch': 0.98}) +[2026-06-23 20:18:10,802][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.15183214843273163, 'train/grad_norm_pre_clip': 26.336774826049805, 'timing/data': 0.0006102859042584896, 'timing/model': 1.694053779123351, 'learning_rate/qwen_vl_interface': 5.332222272489662e-06, 'learning_rate/action_model': 5.332222272489662e-05, 'epoch': 0.98}) +[2026-06-23 20:18:14,136][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.05139540508389473, 'train/grad_norm_pre_clip': 10.066755294799805, 'timing/data': 0.0006015500985085964, 'timing/model': 1.6940439059399068, 'learning_rate/qwen_vl_interface': 5.3291772906398895e-06, 'learning_rate/action_model': 5.32917729063989e-05, 'epoch': 0.98}) +[2026-06-23 20:18:17,469][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.08673493564128876, 'train/grad_norm_pre_clip': 24.54651641845703, 'timing/data': 0.0007784408517181873, 'timing/model': 1.695399718126282, 'learning_rate/qwen_vl_interface': 5.326132276243327e-06, 'learning_rate/action_model': 5.3261322762433266e-05, 'epoch': 0.98}) +[2026-06-23 20:18:20,803][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.1781003326177597, 'train/grad_norm_pre_clip': 20.797500610351562, 'timing/data': 0.0006961370818316936, 'timing/model': 1.6924397258553654, 'learning_rate/qwen_vl_interface': 5.323087230551659e-06, 'learning_rate/action_model': 5.323087230551659e-05, 'epoch': 0.98}) +[2026-06-23 20:18:24,139][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.1093304380774498, 'train/grad_norm_pre_clip': 20.16790008544922, 'timing/data': 0.0006624539382755756, 'timing/model': 1.6917408080771565, 'learning_rate/qwen_vl_interface': 5.3200421548165916e-06, 'learning_rate/action_model': 5.3200421548165914e-05, 'epoch': 0.98}) +[2026-06-23 20:18:27,476][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.11710815876722336, 'train/grad_norm_pre_clip': 26.3374080657959, 'timing/data': 0.0006131120026111603, 'timing/model': 1.6964012570679188, 'learning_rate/qwen_vl_interface': 5.3169970502898405e-06, 'learning_rate/action_model': 5.31699705028984e-05, 'epoch': 0.98}) +[2026-06-23 20:18:30,811][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.1281561702489853, 'train/grad_norm_pre_clip': 24.595346450805664, 'timing/data': 0.0006298229563981295, 'timing/model': 1.6932128751650453, 'learning_rate/qwen_vl_interface': 5.31395191822313e-06, 'learning_rate/action_model': 5.31395191822313e-05, 'epoch': 0.98}) +[2026-06-23 20:18:34,152][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.06591296941041946, 'train/grad_norm_pre_clip': 25.483407974243164, 'timing/data': 0.0007285380270332098, 'timing/model': 1.6969364830292761, 'learning_rate/qwen_vl_interface': 5.310906759868205e-06, 'learning_rate/action_model': 5.310906759868205e-05, 'epoch': 0.98}) +[2026-06-23 20:18:37,489][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.12222231179475784, 'train/grad_norm_pre_clip': 27.63103675842285, 'timing/data': 0.0006832629442214966, 'timing/model': 1.6933406661264598, 'learning_rate/qwen_vl_interface': 5.30786157647681e-06, 'learning_rate/action_model': 5.30786157647681e-05, 'epoch': 0.98}) +[2026-06-23 20:18:40,820][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.061085134744644165, 'train/grad_norm_pre_clip': 27.500905990600586, 'timing/data': 0.000691819004714489, 'timing/model': 1.6914122691377997, 'learning_rate/qwen_vl_interface': 5.304816369300709e-06, 'learning_rate/action_model': 5.304816369300709e-05, 'epoch': 0.98}) +[2026-06-23 20:18:44,156][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.15629567205905914, 'train/grad_norm_pre_clip': 39.79671096801758, 'timing/data': 0.000945043982937932, 'timing/model': 1.6953925690613687, 'learning_rate/qwen_vl_interface': 5.301771139591666e-06, 'learning_rate/action_model': 5.301771139591666e-05, 'epoch': 0.98}) +[2026-06-23 20:18:47,491][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.04905379191040993, 'train/grad_norm_pre_clip': 27.217933654785156, 'timing/data': 0.0006734149064868689, 'timing/model': 1.688465225044638, 'learning_rate/qwen_vl_interface': 5.298725888601465e-06, 'learning_rate/action_model': 5.298725888601465e-05, 'epoch': 0.98}) +[2026-06-23 20:18:50,827][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.15456397831439972, 'train/grad_norm_pre_clip': 56.256141662597656, 'timing/data': 0.0006765800062566996, 'timing/model': 1.6946602631360292, 'learning_rate/qwen_vl_interface': 5.295680617581891e-06, 'learning_rate/action_model': 5.2956806175818904e-05, 'epoch': 0.98}) +[2026-06-23 20:18:54,162][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.09366849064826965, 'train/grad_norm_pre_clip': 13.481876373291016, 'timing/data': 0.0006015291437506676, 'timing/model': 1.692712158896029, 'learning_rate/qwen_vl_interface': 5.2926353277847395e-06, 'learning_rate/action_model': 5.292635327784739e-05, 'epoch': 0.98}) +[2026-06-23 20:18:57,777][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.181110218167305, 'train/grad_norm_pre_clip': 60.4825325012207, 'timing/data': 0.0006924408953636885, 'timing/model': 1.9715587988030165, 'learning_rate/qwen_vl_interface': 5.2895900204618175e-06, 'learning_rate/action_model': 5.289590020461818e-05, 'epoch': 0.98}) +[2026-06-23 20:19:01,117][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.07674829661846161, 'train/grad_norm_pre_clip': 24.405550003051758, 'timing/data': 0.0006616329774260521, 'timing/model': 1.697085744002834, 'learning_rate/qwen_vl_interface': 5.286544696864933e-06, 'learning_rate/action_model': 5.2865446968649324e-05, 'epoch': 0.98}) +[2026-06-23 20:19:04,460][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.10930654406547546, 'train/grad_norm_pre_clip': 30.030324935913086, 'timing/data': 0.0006213868036866188, 'timing/model': 1.6950339858885854, 'learning_rate/qwen_vl_interface': 5.283499358245904e-06, 'learning_rate/action_model': 5.2834993582459034e-05, 'epoch': 0.98}) +[2026-06-23 20:19:07,808][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.2190839797258377, 'train/grad_norm_pre_clip': 44.248748779296875, 'timing/data': 0.0006021610461175442, 'timing/model': 1.7004169099964201, 'learning_rate/qwen_vl_interface': 5.280454005856553e-06, 'learning_rate/action_model': 5.2804540058565534e-05, 'epoch': 0.98}) +[2026-06-23 20:19:11,147][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.1690627485513687, 'train/grad_norm_pre_clip': 27.009048461914062, 'timing/data': 0.0006199749186635017, 'timing/model': 1.6985196410678327, 'learning_rate/qwen_vl_interface': 5.277408640948711e-06, 'learning_rate/action_model': 5.277408640948711e-05, 'epoch': 0.98}) +[2026-06-23 20:19:14,489][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.14836087822914124, 'train/grad_norm_pre_clip': 45.298343658447266, 'timing/data': 0.0006066400092095137, 'timing/model': 1.695739784045145, 'learning_rate/qwen_vl_interface': 5.274363264774213e-06, 'learning_rate/action_model': 5.274363264774212e-05, 'epoch': 0.98}) +[2026-06-23 20:19:17,835][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.09932846575975418, 'train/grad_norm_pre_clip': 27.403270721435547, 'timing/data': 0.0006146950181573629, 'timing/model': 1.695026308996603, 'learning_rate/qwen_vl_interface': 5.271317878584897e-06, 'learning_rate/action_model': 5.2713178785848974e-05, 'epoch': 0.98}) +[2026-06-23 20:19:21,177][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.09326006472110748, 'train/grad_norm_pre_clip': 20.393301010131836, 'timing/data': 0.0006191420834511518, 'timing/model': 1.6957597399596125, 'learning_rate/qwen_vl_interface': 5.268272483632606e-06, 'learning_rate/action_model': 5.268272483632606e-05, 'epoch': 0.98}) +[2026-06-23 20:19:24,520][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.14853867888450623, 'train/grad_norm_pre_clip': 15.713217735290527, 'timing/data': 0.0006221088115125895, 'timing/model': 1.6955120088532567, 'learning_rate/qwen_vl_interface': 5.265227081169187e-06, 'learning_rate/action_model': 5.265227081169186e-05, 'epoch': 0.98}) +[2026-06-23 20:19:27,864][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.16103294491767883, 'train/grad_norm_pre_clip': 94.7653579711914, 'timing/data': 0.0005991850048303604, 'timing/model': 1.6987882459070534, 'learning_rate/qwen_vl_interface': 5.2621816724464895e-06, 'learning_rate/action_model': 5.262181672446489e-05, 'epoch': 0.99}) +[2026-06-23 20:19:31,212][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.09899982810020447, 'train/grad_norm_pre_clip': 24.838062286376953, 'timing/data': 0.0009058320429176092, 'timing/model': 1.699649698100984, 'learning_rate/qwen_vl_interface': 5.259136258716367e-06, 'learning_rate/action_model': 5.2591362587163675e-05, 'epoch': 0.99}) +[2026-06-23 20:19:34,548][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.19591239094734192, 'train/grad_norm_pre_clip': 51.282501220703125, 'timing/data': 0.000620134873315692, 'timing/model': 1.6942210260313004, 'learning_rate/qwen_vl_interface': 5.256090841230675e-06, 'learning_rate/action_model': 5.256090841230674e-05, 'epoch': 0.99}) +[2026-06-23 20:19:37,888][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.06332416832447052, 'train/grad_norm_pre_clip': 17.27983283996582, 'timing/data': 0.0006461238954216242, 'timing/model': 1.6954558589495718, 'learning_rate/qwen_vl_interface': 5.253045421241266e-06, 'learning_rate/action_model': 5.253045421241266e-05, 'epoch': 0.99}) +[2026-06-23 20:19:41,237][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.13178527355194092, 'train/grad_norm_pre_clip': 44.665157318115234, 'timing/data': 0.0006234508473426104, 'timing/model': 1.6989697718527168, 'learning_rate/qwen_vl_interface': 5.2500000000000006e-06, 'learning_rate/action_model': 5.25e-05, 'epoch': 0.99}) +[2026-06-23 20:19:44,683][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.09007015824317932, 'train/grad_norm_pre_clip': 9.369803428649902, 'timing/data': 0.0006155660375952721, 'timing/model': 1.7923146449029446, 'learning_rate/qwen_vl_interface': 5.246954578758735e-06, 'learning_rate/action_model': 5.246954578758735e-05, 'epoch': 0.99}) +[2026-06-23 20:19:48,030][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.01883063092827797, 'train/grad_norm_pre_clip': 9.980971336364746, 'timing/data': 0.000615556025877595, 'timing/model': 1.6984940459951758, 'learning_rate/qwen_vl_interface': 5.243909158769327e-06, 'learning_rate/action_model': 5.2439091587693264e-05, 'epoch': 0.99}) +[2026-06-23 20:19:51,373][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.1375347524881363, 'train/grad_norm_pre_clip': 19.40742301940918, 'timing/data': 0.0006221090443432331, 'timing/model': 1.6987432271707803, 'learning_rate/qwen_vl_interface': 5.240863741283634e-06, 'learning_rate/action_model': 5.2408637412836336e-05, 'epoch': 0.99}) +[2026-06-23 20:19:54,722][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.14992250502109528, 'train/grad_norm_pre_clip': 17.318605422973633, 'timing/data': 0.0006561018526554108, 'timing/model': 1.6994683619122952, 'learning_rate/qwen_vl_interface': 5.2378183275535125e-06, 'learning_rate/action_model': 5.237818327553512e-05, 'epoch': 0.99}) +[2026-06-23 20:19:58,064][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.11543700844049454, 'train/grad_norm_pre_clip': 16.668006896972656, 'timing/data': 0.0007146119605749846, 'timing/model': 1.699474676977843, 'learning_rate/qwen_vl_interface': 5.234772918830815e-06, 'learning_rate/action_model': 5.234772918830815e-05, 'epoch': 0.99}) +[2026-06-23 20:20:01,435][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.10966084152460098, 'train/grad_norm_pre_clip': 24.44685173034668, 'timing/data': 0.000600217143073678, 'timing/model': 1.7011267829220742, 'learning_rate/qwen_vl_interface': 5.231727516367395e-06, 'learning_rate/action_model': 5.231727516367395e-05, 'epoch': 0.99}) +[2026-06-23 20:20:04,779][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.15491166710853577, 'train/grad_norm_pre_clip': 30.714893341064453, 'timing/data': 0.0007315138354897499, 'timing/model': 1.691516777034849, 'learning_rate/qwen_vl_interface': 5.228682121415104e-06, 'learning_rate/action_model': 5.228682121415104e-05, 'epoch': 0.99}) +[2026-06-23 20:20:08,131][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.06669387966394424, 'train/grad_norm_pre_clip': 40.096588134765625, 'timing/data': 0.0006196729373186827, 'timing/model': 1.703072353033349, 'learning_rate/qwen_vl_interface': 5.225636735225789e-06, 'learning_rate/action_model': 5.225636735225788e-05, 'epoch': 0.99}) +[2026-06-23 20:20:11,473][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.06268999725580215, 'train/grad_norm_pre_clip': 69.68462371826172, 'timing/data': 0.0006092539988458157, 'timing/model': 1.6981533281505108, 'learning_rate/qwen_vl_interface': 5.22259135905129e-06, 'learning_rate/action_model': 5.2225913590512895e-05, 'epoch': 0.99}) +[2026-06-23 20:20:14,812][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.14105647802352905, 'train/grad_norm_pre_clip': 17.440975189208984, 'timing/data': 0.0006143040955066681, 'timing/model': 1.6938761658966541, 'learning_rate/qwen_vl_interface': 5.219545994143448e-06, 'learning_rate/action_model': 5.219545994143448e-05, 'epoch': 0.99}) +[2026-06-23 20:20:18,145][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.1342543512582779, 'train/grad_norm_pre_clip': 26.75275993347168, 'timing/data': 0.0006072199903428555, 'timing/model': 1.6927194269374013, 'learning_rate/qwen_vl_interface': 5.216500641754097e-06, 'learning_rate/action_model': 5.216500641754097e-05, 'epoch': 0.99}) +[2026-06-23 20:20:21,488][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.13118359446525574, 'train/grad_norm_pre_clip': 18.37687110900879, 'timing/data': 0.0007352011743932962, 'timing/model': 1.6941260409075767, 'learning_rate/qwen_vl_interface': 5.213455303135067e-06, 'learning_rate/action_model': 5.2134553031350673e-05, 'epoch': 0.99}) +[2026-06-23 20:20:24,833][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.11658688634634018, 'train/grad_norm_pre_clip': 34.598167419433594, 'timing/data': 0.0006058579310774803, 'timing/model': 1.6994241951033473, 'learning_rate/qwen_vl_interface': 5.2104099795381836e-06, 'learning_rate/action_model': 5.210409979538183e-05, 'epoch': 0.99}) +[2026-06-23 20:20:28,170][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.07901277393102646, 'train/grad_norm_pre_clip': 55.892391204833984, 'timing/data': 0.0006124109495431185, 'timing/model': 1.6930465979967266, 'learning_rate/qwen_vl_interface': 5.207364672215261e-06, 'learning_rate/action_model': 5.2073646722152604e-05, 'epoch': 0.99}) +[2026-06-23 20:20:31,507][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.23554429411888123, 'train/grad_norm_pre_clip': 26.829097747802734, 'timing/data': 0.0006114391144365072, 'timing/model': 1.693601642968133, 'learning_rate/qwen_vl_interface': 5.2043193824181095e-06, 'learning_rate/action_model': 5.2043193824181094e-05, 'epoch': 0.99}) +[2026-06-23 20:20:35,171][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.1853063404560089, 'train/grad_norm_pre_clip': 27.190534591674805, 'timing/data': 0.0006976700387895107, 'timing/model': 2.007364648161456, 'learning_rate/qwen_vl_interface': 5.201274111398535e-06, 'learning_rate/action_model': 5.2012741113985356e-05, 'epoch': 0.99}) +[2026-06-23 20:20:38,521][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.17352041602134705, 'train/grad_norm_pre_clip': 12.882328987121582, 'timing/data': 0.0007076289039105177, 'timing/model': 1.6995764411985874, 'learning_rate/qwen_vl_interface': 5.198228860408334e-06, 'learning_rate/action_model': 5.198228860408334e-05, 'epoch': 0.99}) +[2026-06-23 20:20:41,848][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.1509055197238922, 'train/grad_norm_pre_clip': 25.877111434936523, 'timing/data': 0.000713930930942297, 'timing/model': 1.678842858877033, 'learning_rate/qwen_vl_interface': 5.195183630699292e-06, 'learning_rate/action_model': 5.195183630699292e-05, 'epoch': 0.99}) +[2026-06-23 20:20:45,152][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.14474855363368988, 'train/grad_norm_pre_clip': 40.485652923583984, 'timing/data': 0.0007190699689090252, 'timing/model': 1.67927012196742, 'learning_rate/qwen_vl_interface': 5.192138423523189e-06, 'learning_rate/action_model': 5.1921384235231895e-05, 'epoch': 0.99}) +[2026-06-23 20:20:48,460][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.2562806010246277, 'train/grad_norm_pre_clip': 24.021528244018555, 'timing/data': 0.0006964881904423237, 'timing/model': 1.6797242669854313, 'learning_rate/qwen_vl_interface': 5.189093240131795e-06, 'learning_rate/action_model': 5.189093240131795e-05, 'epoch': 0.99}) +[2026-06-23 20:20:51,769][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.1943071037530899, 'train/grad_norm_pre_clip': 22.965129852294922, 'timing/data': 0.0006122901104390621, 'timing/model': 1.680698538897559, 'learning_rate/qwen_vl_interface': 5.186048081776869e-06, 'learning_rate/action_model': 5.1860480817768685e-05, 'epoch': 1.0}) +[2026-06-23 20:20:55,077][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.11658337712287903, 'train/grad_norm_pre_clip': 18.402048110961914, 'timing/data': 0.0006151951383799314, 'timing/model': 1.679469722090289, 'learning_rate/qwen_vl_interface': 5.183002949710161e-06, 'learning_rate/action_model': 5.18300294971016e-05, 'epoch': 1.0}) +[2026-06-23 20:20:58,398][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.08566969633102417, 'train/grad_norm_pre_clip': 35.93827438354492, 'timing/data': 0.0006274180486798286, 'timing/model': 1.683999452041462, 'learning_rate/qwen_vl_interface': 5.179957845183411e-06, 'learning_rate/action_model': 5.179957845183411e-05, 'epoch': 1.0}) +[2026-06-23 20:21:01,710][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.0962267518043518, 'train/grad_norm_pre_clip': 13.055201530456543, 'timing/data': 0.000603182939812541, 'timing/model': 1.682007993804291, 'learning_rate/qwen_vl_interface': 5.176912769448343e-06, 'learning_rate/action_model': 5.176912769448343e-05, 'epoch': 1.0}) +[2026-06-23 20:21:05,021][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.15219782292842865, 'train/grad_norm_pre_clip': 16.894271850585938, 'timing/data': 0.0006058469880372286, 'timing/model': 1.6822483059950173, 'learning_rate/qwen_vl_interface': 5.173867723756676e-06, 'learning_rate/action_model': 5.173867723756676e-05, 'epoch': 1.0}) +[2026-06-23 20:21:08,334][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.15916581451892853, 'train/grad_norm_pre_clip': 29.09092140197754, 'timing/data': 0.0006183912046253681, 'timing/model': 1.6819773179013282, 'learning_rate/qwen_vl_interface': 5.170822709360112e-06, 'learning_rate/action_model': 5.170822709360111e-05, 'epoch': 1.0}) +[2026-06-23 20:21:11,645][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.07317068427801132, 'train/grad_norm_pre_clip': 13.220819473266602, 'timing/data': 0.0006103070918470621, 'timing/model': 1.6816881159320474, 'learning_rate/qwen_vl_interface': 5.16777772751034e-06, 'learning_rate/action_model': 5.16777772751034e-05, 'epoch': 1.0}) +[2026-06-23 20:21:14,955][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.2761496305465698, 'train/grad_norm_pre_clip': 36.437225341796875, 'timing/data': 0.0006074910052120686, 'timing/model': 1.6803141359705478, 'learning_rate/qwen_vl_interface': 5.16473277945904e-06, 'learning_rate/action_model': 5.1647327794590396e-05, 'epoch': 1.0}) +[2026-06-23 20:21:18,265][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.08785812556743622, 'train/grad_norm_pre_clip': 43.725162506103516, 'timing/data': 0.0006093650590628386, 'timing/model': 1.6814275491051376, 'learning_rate/qwen_vl_interface': 5.161687866457872e-06, 'learning_rate/action_model': 5.161687866457871e-05, 'epoch': 1.0}) +[2026-06-23 20:21:21,574][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.13082954287528992, 'train/grad_norm_pre_clip': 36.63820266723633, 'timing/data': 0.0006142128258943558, 'timing/model': 1.6803280939348042, 'learning_rate/qwen_vl_interface': 5.158642989758487e-06, 'learning_rate/action_model': 5.158642989758487e-05, 'epoch': 1.0}) +[2026-06-23 20:21:24,888][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.14075756072998047, 'train/grad_norm_pre_clip': 12.924711227416992, 'timing/data': 0.0008640231098979712, 'timing/model': 1.6839658189564943, 'learning_rate/qwen_vl_interface': 5.155598150612519e-06, 'learning_rate/action_model': 5.1555981506125185e-05, 'epoch': 1.0}) +[2026-06-23 20:21:28,379][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.060307007282972336, 'train/grad_norm_pre_clip': 21.861879348754883, 'timing/data': 0.0006298029329627752, 'timing/model': 1.6810291439760476, 'learning_rate/qwen_vl_interface': 5.152553350271584e-06, 'learning_rate/action_model': 5.152553350271584e-05, 'epoch': 1.0}) +[2026-06-23 20:21:31,689][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.15121454000473022, 'train/grad_norm_pre_clip': 20.01078224182129, 'timing/data': 0.000610897084698081, 'timing/model': 1.6800885580014437, 'learning_rate/qwen_vl_interface': 5.149508589987286e-06, 'learning_rate/action_model': 5.1495085899872855e-05, 'epoch': 1.0}) +[2026-06-23 20:21:34,463][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.24298548698425293, 'train/grad_norm_pre_clip': 30.706953048706055, 'timing/data': 0.0004896498285233974, 'timing/model': 1.1469557990785688, 'learning_rate/qwen_vl_interface': 5.14646387101121e-06, 'learning_rate/action_model': 5.14646387101121e-05, 'epoch': 1.0}) +[2026-06-23 20:21:42,114][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.1060713604092598, 'train/grad_norm_pre_clip': 13.66024112701416, 'timing/data': 0.0015104769263416529, 'timing/model': 1.8401544138323516, 'learning_rate/qwen_vl_interface': 5.143419194594928e-06, 'learning_rate/action_model': 5.143419194594927e-05, 'epoch': 1.0}) +[2026-06-23 20:21:45,867][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.09274522215127945, 'train/grad_norm_pre_clip': 11.048803329467773, 'timing/data': 0.0012681619264185429, 'timing/model': 1.8367084939964116, 'learning_rate/qwen_vl_interface': 5.140374561989988e-06, 'learning_rate/action_model': 5.140374561989988e-05, 'epoch': 1.0}) +[2026-06-23 20:21:49,470][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.18373332917690277, 'train/grad_norm_pre_clip': 16.91815948486328, 'timing/data': 0.001096328953281045, 'timing/model': 1.8154903589747846, 'learning_rate/qwen_vl_interface': 5.137329974447924e-06, 'learning_rate/action_model': 5.137329974447924e-05, 'epoch': 1.0}) +[2026-06-23 20:21:53,081][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.0748816430568695, 'train/grad_norm_pre_clip': 21.856252670288086, 'timing/data': 0.0016801259480416775, 'timing/model': 1.819680262124166, 'learning_rate/qwen_vl_interface': 5.1342854332202516e-06, 'learning_rate/action_model': 5.1342854332202514e-05, 'epoch': 1.0}) +[2026-06-23 20:21:56,588][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.14525450766086578, 'train/grad_norm_pre_clip': 27.15455436706543, 'timing/data': 0.0010608730372041464, 'timing/model': 1.7277103520464152, 'learning_rate/qwen_vl_interface': 5.131240939558467e-06, 'learning_rate/action_model': 5.131240939558467e-05, 'epoch': 1.0}) +[2026-06-23 20:21:59,918][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.11875132471323013, 'train/grad_norm_pre_clip': 21.564260482788086, 'timing/data': 0.0005796090699732304, 'timing/model': 1.6918665210250765, 'learning_rate/qwen_vl_interface': 5.128196494714048e-06, 'learning_rate/action_model': 5.128196494714047e-05, 'epoch': 1.0}) +[2026-06-23 20:22:03,244][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.11777909100055695, 'train/grad_norm_pre_clip': 15.840145111083984, 'timing/data': 0.0005634091794490814, 'timing/model': 1.688850655918941, 'learning_rate/qwen_vl_interface': 5.125152099938448e-06, 'learning_rate/action_model': 5.125152099938447e-05, 'epoch': 1.0}) +[2026-06-23 20:22:06,580][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.2283971756696701, 'train/grad_norm_pre_clip': 20.741718292236328, 'timing/data': 0.0005795590113848448, 'timing/model': 1.6960575359407812, 'learning_rate/qwen_vl_interface': 5.122107756483104e-06, 'learning_rate/action_model': 5.122107756483103e-05, 'epoch': 1.0}) +[2026-06-23 20:22:09,916][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.16683152318000793, 'train/grad_norm_pre_clip': 16.1657772064209, 'timing/data': 0.0005850889720022678, 'timing/model': 1.6920113328378648, 'learning_rate/qwen_vl_interface': 5.119063465599429e-06, 'learning_rate/action_model': 5.1190634655994285e-05, 'epoch': 1.0}) +[2026-06-23 20:22:13,252][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.08132093399763107, 'train/grad_norm_pre_clip': 15.843695640563965, 'timing/data': 0.000595147954300046, 'timing/model': 1.6948825370054692, 'learning_rate/qwen_vl_interface': 5.116019228538818e-06, 'learning_rate/action_model': 5.116019228538818e-05, 'epoch': 1.0}) +[2026-06-23 20:22:16,593][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.04615241661667824, 'train/grad_norm_pre_clip': 25.432109832763672, 'timing/data': 0.0007968458812683821, 'timing/model': 1.6979791240300983, 'learning_rate/qwen_vl_interface': 5.112975046552641e-06, 'learning_rate/action_model': 5.1129750465526404e-05, 'epoch': 1.0}) +[2026-06-23 20:22:20,034][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.12096124142408371, 'train/grad_norm_pre_clip': 25.684139251708984, 'timing/data': 0.0007930388674139977, 'timing/model': 1.6998436958529055, 'learning_rate/qwen_vl_interface': 5.109930920892246e-06, 'learning_rate/action_model': 5.1099309208922464e-05, 'epoch': 1.0}) +[2026-06-23 20:22:23,379][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.17835892736911774, 'train/grad_norm_pre_clip': 35.90288543701172, 'timing/data': 0.0005681670736521482, 'timing/model': 1.7004451961256564, 'learning_rate/qwen_vl_interface': 5.10688685280896e-06, 'learning_rate/action_model': 5.10688685280896e-05, 'epoch': 1.01}) +[2026-06-23 20:22:26,718][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.08032215386629105, 'train/grad_norm_pre_clip': 31.23629379272461, 'timing/data': 0.000571572920307517, 'timing/model': 1.6974088829010725, 'learning_rate/qwen_vl_interface': 5.103842843554082e-06, 'learning_rate/action_model': 5.103842843554082e-05, 'epoch': 1.01}) +[2026-06-23 20:22:30,065][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.1799791008234024, 'train/grad_norm_pre_clip': 15.183155059814453, 'timing/data': 0.0005669249221682549, 'timing/model': 1.7017321439925581, 'learning_rate/qwen_vl_interface': 5.100798894378891e-06, 'learning_rate/action_model': 5.100798894378891e-05, 'epoch': 1.01}) +[2026-06-23 20:22:33,414][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2600 +[2026-06-23 20:24:06,990][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2600 +[2026-06-23 20:24:06,990][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2600 +[2026-06-23 20:25:17,797][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2600 +[2026-06-23 20:25:17,798][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.1364424228668213, 'train/grad_norm_pre_clip': 18.153196334838867, 'eval/action_loss/seconds': 93.57305049407296, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07235188037157059, 'eval/latency_2/loss': 0.07235188037157059, 'eval/flappy/loss': 0.07235188037157059, 'eval/flappy/latency_2/loss': 0.07235188037157059, 'eval/flap_precision': 0.9092079845460399, 'eval/flap_recall': 0.9527665317139001, 'eval/flap_f1': 0.9304777594728171, 'eval/flap_rate_model': 0.21569444444444444, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9304777594728171, 'eval/latency_2/flap_precision': 0.9092079845460399, 'eval/latency_2/flap_recall': 0.9527665317139001, 'eval/latency_2/flap_f1': 0.9304777594728171, 'eval/latency_2/flap_rate_model': 0.21569444444444444, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9304777594728171, 'eval/action_classification/data_wait_seconds': 0.9632074236869812, 'eval/action_classification/predict_seconds': 69.70864868164062, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005805909167975187, 'timing/model': 1.7027286859229207, 'learning_rate/qwen_vl_interface': 5.097755006534638e-06, 'learning_rate/action_model': 5.097755006534638e-05, 'epoch': 1.01}) +[2026-06-23 20:25:21,122][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.1036427766084671, 'train/grad_norm_pre_clip': 15.557354927062988, 'timing/data': 0.0006252441089600325, 'timing/model': 1.687856798991561, 'learning_rate/qwen_vl_interface': 5.094711181272551e-06, 'learning_rate/action_model': 5.094711181272551e-05, 'epoch': 1.01}) +[2026-06-23 20:25:24,434][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.13902220129966736, 'train/grad_norm_pre_clip': 30.31882667541504, 'timing/data': 0.0005887260194867849, 'timing/model': 1.6791717989835888, 'learning_rate/qwen_vl_interface': 5.091667419843833e-06, 'learning_rate/action_model': 5.091667419843833e-05, 'epoch': 1.01}) +[2026-06-23 20:25:27,752][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.17933733761310577, 'train/grad_norm_pre_clip': 25.93682289123535, 'timing/data': 0.0005680569447577, 'timing/model': 1.6865466120652854, 'learning_rate/qwen_vl_interface': 5.088623723499657e-06, 'learning_rate/action_model': 5.088623723499657e-05, 'epoch': 1.01}) +[2026-06-23 20:25:31,206][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.12322844564914703, 'train/grad_norm_pre_clip': 20.922584533691406, 'timing/data': 0.0005774549208581448, 'timing/model': 1.8225259948521852, 'learning_rate/qwen_vl_interface': 5.085580093491171e-06, 'learning_rate/action_model': 5.085580093491171e-05, 'epoch': 1.01}) +[2026-06-23 20:25:34,523][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.16296717524528503, 'train/grad_norm_pre_clip': 67.99604034423828, 'timing/data': 0.0005600219592452049, 'timing/model': 1.6851369170472026, 'learning_rate/qwen_vl_interface': 5.0825365310695e-06, 'learning_rate/action_model': 5.082536531069499e-05, 'epoch': 1.01}) +[2026-06-23 20:25:37,846][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.12551249563694, 'train/grad_norm_pre_clip': 24.009037017822266, 'timing/data': 0.0006768719758838415, 'timing/model': 1.6868720839265734, 'learning_rate/qwen_vl_interface': 5.079493037485735e-06, 'learning_rate/action_model': 5.0794930374857344e-05, 'epoch': 1.01}) +[2026-06-23 20:25:41,170][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.07918291538953781, 'train/grad_norm_pre_clip': 23.331462860107422, 'timing/data': 0.0005842281971126795, 'timing/model': 1.6843079428654164, 'learning_rate/qwen_vl_interface': 5.076449613990941e-06, 'learning_rate/action_model': 5.076449613990941e-05, 'epoch': 1.01}) +[2026-06-23 20:25:44,492][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.09602507203817368, 'train/grad_norm_pre_clip': 45.32752990722656, 'timing/data': 0.0005727859679609537, 'timing/model': 1.6868885969743133, 'learning_rate/qwen_vl_interface': 5.073406261836157e-06, 'learning_rate/action_model': 5.0734062618361565e-05, 'epoch': 1.01}) +[2026-06-23 20:25:47,813][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.1381620168685913, 'train/grad_norm_pre_clip': 52.446956634521484, 'timing/data': 0.0006443611346185207, 'timing/model': 1.6875437449198216, 'learning_rate/qwen_vl_interface': 5.070362982272386e-06, 'learning_rate/action_model': 5.070362982272385e-05, 'epoch': 1.01}) +[2026-06-23 20:25:51,133][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.14686420559883118, 'train/grad_norm_pre_clip': 89.88666534423828, 'timing/data': 0.0005853499751538038, 'timing/model': 1.689341840101406, 'learning_rate/qwen_vl_interface': 5.067319776550607e-06, 'learning_rate/action_model': 5.067319776550607e-05, 'epoch': 1.01}) +[2026-06-23 20:25:54,463][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.1331377923488617, 'train/grad_norm_pre_clip': 31.19693374633789, 'timing/data': 0.0006113890558481216, 'timing/model': 1.6928840409964323, 'learning_rate/qwen_vl_interface': 5.064276645921768e-06, 'learning_rate/action_model': 5.064276645921768e-05, 'epoch': 1.01}) +[2026-06-23 20:25:57,789][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.1283929944038391, 'train/grad_norm_pre_clip': 40.98199462890625, 'timing/data': 0.0007825200445950031, 'timing/model': 1.6875504471827298, 'learning_rate/qwen_vl_interface': 5.061233591636787e-06, 'learning_rate/action_model': 5.061233591636787e-05, 'epoch': 1.01}) +[2026-06-23 20:26:01,114][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.19634215533733368, 'train/grad_norm_pre_clip': 39.722347259521484, 'timing/data': 0.0006872608792036772, 'timing/model': 1.6887163638602942, 'learning_rate/qwen_vl_interface': 5.058190614946544e-06, 'learning_rate/action_model': 5.058190614946544e-05, 'epoch': 1.01}) +[2026-06-23 20:26:04,444][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.22369669377803802, 'train/grad_norm_pre_clip': 42.79667282104492, 'timing/data': 0.0005744390655308962, 'timing/model': 1.693493747152388, 'learning_rate/qwen_vl_interface': 5.055147717101895e-06, 'learning_rate/action_model': 5.055147717101894e-05, 'epoch': 1.01}) +[2026-06-23 20:26:07,768][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.1028410792350769, 'train/grad_norm_pre_clip': 36.4760856628418, 'timing/data': 0.0005945570301264524, 'timing/model': 1.685998820932582, 'learning_rate/qwen_vl_interface': 5.052104899353658e-06, 'learning_rate/action_model': 5.0521048993536577e-05, 'epoch': 1.01}) +[2026-06-23 20:26:11,099][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.20892687141895294, 'train/grad_norm_pre_clip': 79.773681640625, 'timing/data': 0.0005838770885020494, 'timing/model': 1.6929886699654162, 'learning_rate/qwen_vl_interface': 5.049062162952622e-06, 'learning_rate/action_model': 5.0490621629526214e-05, 'epoch': 1.01}) +[2026-06-23 20:26:14,423][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.07053614407777786, 'train/grad_norm_pre_clip': 21.202619552612305, 'timing/data': 0.000572756165638566, 'timing/model': 1.689341964898631, 'learning_rate/qwen_vl_interface': 5.046019509149541e-06, 'learning_rate/action_model': 5.046019509149541e-05, 'epoch': 1.01}) +[2026-06-23 20:26:17,746][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.09831482172012329, 'train/grad_norm_pre_clip': 14.300924301147461, 'timing/data': 0.0005860610399395227, 'timing/model': 1.6877419541124254, 'learning_rate/qwen_vl_interface': 5.042976939195135e-06, 'learning_rate/action_model': 5.0429769391951355e-05, 'epoch': 1.01}) +[2026-06-23 20:26:21,193][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.24351456761360168, 'train/grad_norm_pre_clip': 33.867591857910156, 'timing/data': 0.0005976129323244095, 'timing/model': 1.6918006970081478, 'learning_rate/qwen_vl_interface': 5.039934454340091e-06, 'learning_rate/action_model': 5.0399344543400907e-05, 'epoch': 1.01}) +[2026-06-23 20:26:24,519][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.16089025139808655, 'train/grad_norm_pre_clip': 50.8582878112793, 'timing/data': 0.000614384887740016, 'timing/model': 1.6867653320077807, 'learning_rate/qwen_vl_interface': 5.036892055835056e-06, 'learning_rate/action_model': 5.0368920558350566e-05, 'epoch': 1.01}) +[2026-06-23 20:26:27,846][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.15932032465934753, 'train/grad_norm_pre_clip': 24.97748565673828, 'timing/data': 0.0005659139715135098, 'timing/model': 1.6884903688915074, 'learning_rate/qwen_vl_interface': 5.033849744930647e-06, 'learning_rate/action_model': 5.033849744930646e-05, 'epoch': 1.01}) +[2026-06-23 20:26:31,168][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.10009834170341492, 'train/grad_norm_pre_clip': 52.493492126464844, 'timing/data': 0.0005702809430658817, 'timing/model': 1.6881003219168633, 'learning_rate/qwen_vl_interface': 5.030807522877443e-06, 'learning_rate/action_model': 5.030807522877443e-05, 'epoch': 1.01}) +[2026-06-23 20:26:34,489][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.13836835324764252, 'train/grad_norm_pre_clip': 27.84388542175293, 'timing/data': 0.0005601330194622278, 'timing/model': 1.6868900249246508, 'learning_rate/qwen_vl_interface': 5.027765390925987e-06, 'learning_rate/action_model': 5.027765390925987e-05, 'epoch': 1.02}) +[2026-06-23 20:26:37,818][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.07667327672243118, 'train/grad_norm_pre_clip': 12.895242691040039, 'timing/data': 0.0005748500116169453, 'timing/model': 1.6908139628358185, 'learning_rate/qwen_vl_interface': 5.024723350326783e-06, 'learning_rate/action_model': 5.024723350326783e-05, 'epoch': 1.02}) +[2026-06-23 20:26:41,143][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.1245323121547699, 'train/grad_norm_pre_clip': 32.83480453491211, 'timing/data': 0.0005916010122746229, 'timing/model': 1.6918576799798757, 'learning_rate/qwen_vl_interface': 5.021681402330303e-06, 'learning_rate/action_model': 5.021681402330303e-05, 'epoch': 1.02}) +[2026-06-23 20:26:44,470][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.2212546020746231, 'train/grad_norm_pre_clip': 57.22987747192383, 'timing/data': 0.000569589901715517, 'timing/model': 1.6860525291413069, 'learning_rate/qwen_vl_interface': 5.018639548186972e-06, 'learning_rate/action_model': 5.0186395481869716e-05, 'epoch': 1.02}) +[2026-06-23 20:26:47,799][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.07349210232496262, 'train/grad_norm_pre_clip': 14.108332633972168, 'timing/data': 0.0005728360265493393, 'timing/model': 1.6916488900315017, 'learning_rate/qwen_vl_interface': 5.015597789147184e-06, 'learning_rate/action_model': 5.015597789147184e-05, 'epoch': 1.02}) +[2026-06-23 20:26:51,137][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.10499982535839081, 'train/grad_norm_pre_clip': 11.992805480957031, 'timing/data': 0.0006437590345740318, 'timing/model': 1.695836853934452, 'learning_rate/qwen_vl_interface': 5.012556126461291e-06, 'learning_rate/action_model': 5.012556126461291e-05, 'epoch': 1.02}) +[2026-06-23 20:26:54,460][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.20048080384731293, 'train/grad_norm_pre_clip': 14.745885848999023, 'timing/data': 0.0005794079042971134, 'timing/model': 1.685800091829151, 'learning_rate/qwen_vl_interface': 5.0095145613796044e-06, 'learning_rate/action_model': 5.0095145613796046e-05, 'epoch': 1.02}) +[2026-06-23 20:26:57,792][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.17776286602020264, 'train/grad_norm_pre_clip': 35.72126388549805, 'timing/data': 0.0007434671279042959, 'timing/model': 1.689008346060291, 'learning_rate/qwen_vl_interface': 5.0064730951524e-06, 'learning_rate/action_model': 5.0064730951524e-05, 'epoch': 1.02}) +[2026-06-23 20:27:01,121][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.06448277831077576, 'train/grad_norm_pre_clip': 15.80125904083252, 'timing/data': 0.000572865828871727, 'timing/model': 1.689625300001353, 'learning_rate/qwen_vl_interface': 5.003431729029905e-06, 'learning_rate/action_model': 5.003431729029905e-05, 'epoch': 1.02}) +[2026-06-23 20:27:04,450][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.14528149366378784, 'train/grad_norm_pre_clip': 26.61811065673828, 'timing/data': 0.0006816599052399397, 'timing/model': 1.6885240571573377, 'learning_rate/qwen_vl_interface': 5.000390464262314e-06, 'learning_rate/action_model': 5.0003904642623136e-05, 'epoch': 1.02}) +[2026-06-23 20:27:07,778][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.07890509068965912, 'train/grad_norm_pre_clip': 14.203540802001953, 'timing/data': 0.0006756288930773735, 'timing/model': 1.691500835120678, 'learning_rate/qwen_vl_interface': 4.997349302099774e-06, 'learning_rate/action_model': 4.997349302099773e-05, 'epoch': 1.02}) +[2026-06-23 20:27:11,380][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.08357080072164536, 'train/grad_norm_pre_clip': 33.230751037597656, 'timing/data': 0.000574460020288825, 'timing/model': 1.6847151420079172, 'learning_rate/qwen_vl_interface': 4.994308243792391e-06, 'learning_rate/action_model': 4.994308243792391e-05, 'epoch': 1.02}) +[2026-06-23 20:27:14,709][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.14033326506614685, 'train/grad_norm_pre_clip': 40.0566291809082, 'timing/data': 0.0005783559754490852, 'timing/model': 1.6890070820227265, 'learning_rate/qwen_vl_interface': 4.991267290590233e-06, 'learning_rate/action_model': 4.991267290590233e-05, 'epoch': 1.02}) +[2026-06-23 20:27:18,039][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.1385568082332611, 'train/grad_norm_pre_clip': 47.098636627197266, 'timing/data': 0.0005726658273488283, 'timing/model': 1.691941540921107, 'learning_rate/qwen_vl_interface': 4.988226443743317e-06, 'learning_rate/action_model': 4.9882264437433174e-05, 'epoch': 1.02}) +[2026-06-23 20:27:21,366][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.16244520246982574, 'train/grad_norm_pre_clip': 24.832992553710938, 'timing/data': 0.0009227038826793432, 'timing/model': 1.6912467810325325, 'learning_rate/qwen_vl_interface': 4.985185704501623e-06, 'learning_rate/action_model': 4.985185704501623e-05, 'epoch': 1.02}) +[2026-06-23 20:27:24,695][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.1390502005815506, 'train/grad_norm_pre_clip': 22.523229598999023, 'timing/data': 0.0005793990567326546, 'timing/model': 1.6901699129957706, 'learning_rate/qwen_vl_interface': 4.982145074115082e-06, 'learning_rate/action_model': 4.982145074115082e-05, 'epoch': 1.02}) +[2026-06-23 20:27:28,024][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.12409428507089615, 'train/grad_norm_pre_clip': 18.696144104003906, 'timing/data': 0.0006718619260936975, 'timing/model': 1.691255268175155, 'learning_rate/qwen_vl_interface': 4.979104553833584e-06, 'learning_rate/action_model': 4.979104553833584e-05, 'epoch': 1.02}) +[2026-06-23 20:27:31,362][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.10970421880483627, 'train/grad_norm_pre_clip': 18.727033615112305, 'timing/data': 0.0005698997993022203, 'timing/model': 1.6964900579769164, 'learning_rate/qwen_vl_interface': 4.9760641449069724e-06, 'learning_rate/action_model': 4.976064144906972e-05, 'epoch': 1.02}) +[2026-06-23 20:27:34,701][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.0666356086730957, 'train/grad_norm_pre_clip': 15.99338436126709, 'timing/data': 0.0005845578853040934, 'timing/model': 1.6966356879565865, 'learning_rate/qwen_vl_interface': 4.973023848585042e-06, 'learning_rate/action_model': 4.973023848585042e-05, 'epoch': 1.02}) +[2026-06-23 20:27:38,037][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.08833702653646469, 'train/grad_norm_pre_clip': 17.61125946044922, 'timing/data': 0.0006287908181548119, 'timing/model': 1.694398710038513, 'learning_rate/qwen_vl_interface': 4.969983666117545e-06, 'learning_rate/action_model': 4.9699836661175446e-05, 'epoch': 1.02}) +[2026-06-23 20:27:41,369][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.11151613295078278, 'train/grad_norm_pre_clip': 13.318204879760742, 'timing/data': 0.000730954110622406, 'timing/model': 1.6944717329461128, 'learning_rate/qwen_vl_interface': 4.966943598754185e-06, 'learning_rate/action_model': 4.9669435987541854e-05, 'epoch': 1.02}) +[2026-06-23 20:27:44,700][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.0428101010620594, 'train/grad_norm_pre_clip': 13.356358528137207, 'timing/data': 0.0005656529683619738, 'timing/model': 1.6908863917924464, 'learning_rate/qwen_vl_interface': 4.963903647744621e-06, 'learning_rate/action_model': 4.96390364774462e-05, 'epoch': 1.02}) +[2026-06-23 20:27:48,031][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.16745832562446594, 'train/grad_norm_pre_clip': 20.427249908447266, 'timing/data': 0.0009081361349672079, 'timing/model': 1.6939271201845258, 'learning_rate/qwen_vl_interface': 4.9608638143384604e-06, 'learning_rate/action_model': 4.960863814338461e-05, 'epoch': 1.02}) +[2026-06-23 20:27:51,361][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.15880149602890015, 'train/grad_norm_pre_clip': 32.936988830566406, 'timing/data': 0.0005755620077252388, 'timing/model': 1.6892450170125812, 'learning_rate/qwen_vl_interface': 4.9578240997852636e-06, 'learning_rate/action_model': 4.957824099785264e-05, 'epoch': 1.02}) +[2026-06-23 20:27:54,688][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.15749864280223846, 'train/grad_norm_pre_clip': 40.26133346557617, 'timing/data': 0.0005838170181959867, 'timing/model': 1.689079260919243, 'learning_rate/qwen_vl_interface': 4.954784505334544e-06, 'learning_rate/action_model': 4.9547845053345434e-05, 'epoch': 1.02}) +[2026-06-23 20:27:58,105][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.10109147429466248, 'train/grad_norm_pre_clip': 17.431581497192383, 'timing/data': 0.0005715538281947374, 'timing/model': 1.694233660819009, 'learning_rate/qwen_vl_interface': 4.951745032235762e-06, 'learning_rate/action_model': 4.951745032235762e-05, 'epoch': 1.02}) +[2026-06-23 20:28:01,437][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.17355172336101532, 'train/grad_norm_pre_clip': 16.027996063232422, 'timing/data': 0.0005806009285151958, 'timing/model': 1.6949420380406082, 'learning_rate/qwen_vl_interface': 4.94870568173833e-06, 'learning_rate/action_model': 4.9487056817383307e-05, 'epoch': 1.03}) +[2026-06-23 20:28:04,774][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.08190672844648361, 'train/grad_norm_pre_clip': 37.00273513793945, 'timing/data': 0.0006476268172264099, 'timing/model': 1.6916455891914666, 'learning_rate/qwen_vl_interface': 4.945666455091614e-06, 'learning_rate/action_model': 4.9456664550916136e-05, 'epoch': 1.03}) +[2026-06-23 20:28:08,115][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.0956815555691719, 'train/grad_norm_pre_clip': 28.61551284790039, 'timing/data': 0.0005833159666508436, 'timing/model': 1.6956550171598792, 'learning_rate/qwen_vl_interface': 4.942627353544921e-06, 'learning_rate/action_model': 4.942627353544921e-05, 'epoch': 1.03}) +[2026-06-23 20:28:11,444][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.22888615727424622, 'train/grad_norm_pre_clip': 101.72760772705078, 'timing/data': 0.0007680931594222784, 'timing/model': 1.685598897980526, 'learning_rate/qwen_vl_interface': 4.939588378347513e-06, 'learning_rate/action_model': 4.9395883783475125e-05, 'epoch': 1.03}) +[2026-06-23 20:28:14,783][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.17850717902183533, 'train/grad_norm_pre_clip': 87.21500396728516, 'timing/data': 0.0007561901584267616, 'timing/model': 1.6943112721201032, 'learning_rate/qwen_vl_interface': 4.936549530748597e-06, 'learning_rate/action_model': 4.9365495307485963e-05, 'epoch': 1.03}) +[2026-06-23 20:28:18,136][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.06742952764034271, 'train/grad_norm_pre_clip': 23.12015724182129, 'timing/data': 0.0005707519594579935, 'timing/model': 1.7092882981523871, 'learning_rate/qwen_vl_interface': 4.9335108119973286e-06, 'learning_rate/action_model': 4.933510811997328e-05, 'epoch': 1.03}) +[2026-06-23 20:28:21,473][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.26928848028182983, 'train/grad_norm_pre_clip': 31.39331817626953, 'timing/data': 0.00059734215028584, 'timing/model': 1.6965199320111424, 'learning_rate/qwen_vl_interface': 4.930472223342812e-06, 'learning_rate/action_model': 4.930472223342812e-05, 'epoch': 1.03}) +[2026-06-23 20:28:24,815][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.0695103108882904, 'train/grad_norm_pre_clip': 13.993212699890137, 'timing/data': 0.0008167251944541931, 'timing/model': 1.696387797826901, 'learning_rate/qwen_vl_interface': 4.927433766034093e-06, 'learning_rate/action_model': 4.9274337660340934e-05, 'epoch': 1.03}) +[2026-06-23 20:28:28,146][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.10696757584810257, 'train/grad_norm_pre_clip': 14.700264930725098, 'timing/data': 0.0005836368072777987, 'timing/model': 1.692814849782735, 'learning_rate/qwen_vl_interface': 4.92439544132017e-06, 'learning_rate/action_model': 4.92439544132017e-05, 'epoch': 1.03}) +[2026-06-23 20:28:31,485][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.1473652422428131, 'train/grad_norm_pre_clip': 20.292587280273438, 'timing/data': 0.0005794689059257507, 'timing/model': 1.6980353938415647, 'learning_rate/qwen_vl_interface': 4.92135725044998e-06, 'learning_rate/action_model': 4.92135725044998e-05, 'epoch': 1.03}) +[2026-06-23 20:28:34,829][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.193265900015831, 'train/grad_norm_pre_clip': 69.52842712402344, 'timing/data': 0.0005862321704626083, 'timing/model': 1.69062065705657, 'learning_rate/qwen_vl_interface': 4.918319194672412e-06, 'learning_rate/action_model': 4.918319194672412e-05, 'epoch': 1.03}) +[2026-06-23 20:28:38,163][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.15579472482204437, 'train/grad_norm_pre_clip': 28.684358596801758, 'timing/data': 0.0005759319756180048, 'timing/model': 1.6924482858739793, 'learning_rate/qwen_vl_interface': 4.9152812752362935e-06, 'learning_rate/action_model': 4.915281275236293e-05, 'epoch': 1.03}) +[2026-06-23 20:28:41,484][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.11406797915697098, 'train/grad_norm_pre_clip': 18.851749420166016, 'timing/data': 0.0005782470107078552, 'timing/model': 1.6915176620241255, 'learning_rate/qwen_vl_interface': 4.9122434933904e-06, 'learning_rate/action_model': 4.9122434933904e-05, 'epoch': 1.03}) +[2026-06-23 20:28:44,824][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.23145534098148346, 'train/grad_norm_pre_clip': 21.490732192993164, 'timing/data': 0.0006082931067794561, 'timing/model': 1.6959493330214173, 'learning_rate/qwen_vl_interface': 4.909205850383449e-06, 'learning_rate/action_model': 4.9092058503834485e-05, 'epoch': 1.03}) +[2026-06-23 20:28:48,455][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.12120706588029861, 'train/grad_norm_pre_clip': 17.56791877746582, 'timing/data': 0.0005924440920352936, 'timing/model': 1.691825059009716, 'learning_rate/qwen_vl_interface': 4.906168347464098e-06, 'learning_rate/action_model': 4.906168347464098e-05, 'epoch': 1.03}) +[2026-06-23 20:28:51,794][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.18929129838943481, 'train/grad_norm_pre_clip': 44.87337875366211, 'timing/data': 0.0005810819566249847, 'timing/model': 1.6944869151338935, 'learning_rate/qwen_vl_interface': 4.903130985880955e-06, 'learning_rate/action_model': 4.903130985880954e-05, 'epoch': 1.03}) +[2026-06-23 20:28:55,127][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.06836407631635666, 'train/grad_norm_pre_clip': 22.792911529541016, 'timing/data': 0.0006446409970521927, 'timing/model': 1.6928416681475937, 'learning_rate/qwen_vl_interface': 4.900093766882558e-06, 'learning_rate/action_model': 4.900093766882558e-05, 'epoch': 1.03}) +[2026-06-23 20:28:58,463][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.1344510018825531, 'train/grad_norm_pre_clip': 16.02394676208496, 'timing/data': 0.0006050369702279568, 'timing/model': 1.6940888159442693, 'learning_rate/qwen_vl_interface': 4.897056691717397e-06, 'learning_rate/action_model': 4.8970566917173966e-05, 'epoch': 1.03}) +[2026-06-23 20:29:01,787][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.08463000506162643, 'train/grad_norm_pre_clip': 13.547057151794434, 'timing/data': 0.000583436107262969, 'timing/model': 1.6838834648951888, 'learning_rate/qwen_vl_interface': 4.894019761633898e-06, 'learning_rate/action_model': 4.894019761633898e-05, 'epoch': 1.03}) +[2026-06-23 20:29:05,116][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.12594002485275269, 'train/grad_norm_pre_clip': 23.60013198852539, 'timing/data': 0.0006916590500622988, 'timing/model': 1.6902427580207586, 'learning_rate/qwen_vl_interface': 4.8909829778804304e-06, 'learning_rate/action_model': 4.89098297788043e-05, 'epoch': 1.03}) +[2026-06-23 20:29:08,453][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.1527065932750702, 'train/grad_norm_pre_clip': 21.167089462280273, 'timing/data': 0.0006302848923951387, 'timing/model': 1.6959287570789456, 'learning_rate/qwen_vl_interface': 4.887946341705298e-06, 'learning_rate/action_model': 4.887946341705297e-05, 'epoch': 1.03}) +[2026-06-23 20:29:11,782][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.1627330332994461, 'train/grad_norm_pre_clip': 22.130876541137695, 'timing/data': 0.0006551810074597597, 'timing/model': 1.690800806041807, 'learning_rate/qwen_vl_interface': 4.8849098543567465e-06, 'learning_rate/action_model': 4.884909854356747e-05, 'epoch': 1.03}) +[2026-06-23 20:29:15,116][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.18353424966335297, 'train/grad_norm_pre_clip': 37.984832763671875, 'timing/data': 0.0007680230773985386, 'timing/model': 1.6917513159569353, 'learning_rate/qwen_vl_interface': 4.881873517082965e-06, 'learning_rate/action_model': 4.8818735170829655e-05, 'epoch': 1.03}) +[2026-06-23 20:29:18,458][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.09321810305118561, 'train/grad_norm_pre_clip': 20.406869888305664, 'timing/data': 0.000634301919490099, 'timing/model': 1.6979714180342853, 'learning_rate/qwen_vl_interface': 4.878837331132076e-06, 'learning_rate/action_model': 4.8788373311320754e-05, 'epoch': 1.03}) +[2026-06-23 20:29:21,789][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.1536455601453781, 'train/grad_norm_pre_clip': 15.498883247375488, 'timing/data': 0.000791175989434123, 'timing/model': 1.6951674839947373, 'learning_rate/qwen_vl_interface': 4.87580129775214e-06, 'learning_rate/action_model': 4.87580129775214e-05, 'epoch': 1.03}) +[2026-06-23 20:29:25,127][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.21406936645507812, 'train/grad_norm_pre_clip': 27.454532623291016, 'timing/data': 0.0008272640407085419, 'timing/model': 1.694384496891871, 'learning_rate/qwen_vl_interface': 4.872765418191156e-06, 'learning_rate/action_model': 4.872765418191155e-05, 'epoch': 1.03}) +[2026-06-23 20:29:28,467][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.19112150371074677, 'train/grad_norm_pre_clip': 27.666261672973633, 'timing/data': 0.0005787878762930632, 'timing/model': 1.6955870310775936, 'learning_rate/qwen_vl_interface': 4.8697296936970575e-06, 'learning_rate/action_model': 4.869729693697057e-05, 'epoch': 1.04}) +[2026-06-23 20:29:31,801][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.18236348032951355, 'train/grad_norm_pre_clip': 81.84752655029297, 'timing/data': 0.0006485090125352144, 'timing/model': 1.6954936599358916, 'learning_rate/qwen_vl_interface': 4.8666941255177185e-06, 'learning_rate/action_model': 4.8666941255177185e-05, 'epoch': 1.04}) +[2026-06-23 20:29:35,137][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.08135565370321274, 'train/grad_norm_pre_clip': 50.08403015136719, 'timing/data': 0.00059465691447258, 'timing/model': 1.693970430875197, 'learning_rate/qwen_vl_interface': 4.863658714900945e-06, 'learning_rate/action_model': 4.863658714900945e-05, 'epoch': 1.04}) +[2026-06-23 20:29:38,696][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.18160736560821533, 'train/grad_norm_pre_clip': 53.024776458740234, 'timing/data': 0.0005884659476578236, 'timing/model': 1.6853798150550574, 'learning_rate/qwen_vl_interface': 4.860623463094483e-06, 'learning_rate/action_model': 4.860623463094483e-05, 'epoch': 1.04}) +[2026-06-23 20:29:42,031][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.06309948861598969, 'train/grad_norm_pre_clip': 26.55144500732422, 'timing/data': 0.0006344621069729328, 'timing/model': 1.693708969047293, 'learning_rate/qwen_vl_interface': 4.857588371346006e-06, 'learning_rate/action_model': 4.8575883713460057e-05, 'epoch': 1.04}) +[2026-06-23 20:29:45,364][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.13840791583061218, 'train/grad_norm_pre_clip': 45.51729202270508, 'timing/data': 0.0005792679730802774, 'timing/model': 1.6930531109683216, 'learning_rate/qwen_vl_interface': 4.854553440903127e-06, 'learning_rate/action_model': 4.854553440903126e-05, 'epoch': 1.04}) +[2026-06-23 20:29:48,699][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.19780021905899048, 'train/grad_norm_pre_clip': 22.809066772460938, 'timing/data': 0.0007460920605808496, 'timing/model': 1.6951962618622929, 'learning_rate/qwen_vl_interface': 4.85151867301339e-06, 'learning_rate/action_model': 4.85151867301339e-05, 'epoch': 1.04}) +[2026-06-23 20:29:52,035][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.12773853540420532, 'train/grad_norm_pre_clip': 17.004247665405273, 'timing/data': 0.0005807608831673861, 'timing/model': 1.6959676737897098, 'learning_rate/qwen_vl_interface': 4.848484068924274e-06, 'learning_rate/action_model': 4.848484068924275e-05, 'epoch': 1.04}) +[2026-06-23 20:29:55,368][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.2299419641494751, 'train/grad_norm_pre_clip': 36.39615249633789, 'timing/data': 0.0006232610903680325, 'timing/model': 1.697096522897482, 'learning_rate/qwen_vl_interface': 4.845449629883193e-06, 'learning_rate/action_model': 4.8454496298831934e-05, 'epoch': 1.04}) +[2026-06-23 20:29:58,713][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.1878821700811386, 'train/grad_norm_pre_clip': 46.053714752197266, 'timing/data': 0.0005799499340355396, 'timing/model': 1.6996014120522887, 'learning_rate/qwen_vl_interface': 4.842415357137487e-06, 'learning_rate/action_model': 4.842415357137487e-05, 'epoch': 1.04}) +[2026-06-23 20:30:02,053][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.1234147846698761, 'train/grad_norm_pre_clip': 48.47269058227539, 'timing/data': 0.0006450118962675333, 'timing/model': 1.693493759026751, 'learning_rate/qwen_vl_interface': 4.839381251934432e-06, 'learning_rate/action_model': 4.8393812519344316e-05, 'epoch': 1.04}) +[2026-06-23 20:30:05,392][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.04677026718854904, 'train/grad_norm_pre_clip': 12.269822120666504, 'timing/data': 0.0007024600636214018, 'timing/model': 1.696033796062693, 'learning_rate/qwen_vl_interface': 4.836347315521233e-06, 'learning_rate/action_model': 4.836347315521233e-05, 'epoch': 1.04}) +[2026-06-23 20:30:08,724][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.17285053431987762, 'train/grad_norm_pre_clip': 37.607460021972656, 'timing/data': 0.0006021710578352213, 'timing/model': 1.6895586329046637, 'learning_rate/qwen_vl_interface': 4.833313549145029e-06, 'learning_rate/action_model': 4.833313549145029e-05, 'epoch': 1.04}) +[2026-06-23 20:30:12,059][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.09623300284147263, 'train/grad_norm_pre_clip': 15.164965629577637, 'timing/data': 0.0005838568322360516, 'timing/model': 1.6922186049632728, 'learning_rate/qwen_vl_interface': 4.830279954052884e-06, 'learning_rate/action_model': 4.8302799540528836e-05, 'epoch': 1.04}) +[2026-06-23 20:30:15,393][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.168985053896904, 'train/grad_norm_pre_clip': 26.734825134277344, 'timing/data': 0.0005721950437873602, 'timing/model': 1.6883389891590923, 'learning_rate/qwen_vl_interface': 4.827246531491797e-06, 'learning_rate/action_model': 4.827246531491797e-05, 'epoch': 1.04}) +[2026-06-23 20:30:18,728][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.09724703431129456, 'train/grad_norm_pre_clip': 28.44582176208496, 'timing/data': 0.0005786770489066839, 'timing/model': 1.6935188360512257, 'learning_rate/qwen_vl_interface': 4.824213282708691e-06, 'learning_rate/action_model': 4.824213282708691e-05, 'epoch': 1.04}) +[2026-06-23 20:30:22,065][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.18246416747570038, 'train/grad_norm_pre_clip': 21.720396041870117, 'timing/data': 0.0005926829762756824, 'timing/model': 1.6936405978631228, 'learning_rate/qwen_vl_interface': 4.821180208950422e-06, 'learning_rate/action_model': 4.821180208950421e-05, 'epoch': 1.04}) +[2026-06-23 20:30:25,493][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.03877304494380951, 'train/grad_norm_pre_clip': 14.6489839553833, 'timing/data': 0.0005893278867006302, 'timing/model': 1.781534189824015, 'learning_rate/qwen_vl_interface': 4.81814731146377e-06, 'learning_rate/action_model': 4.8181473114637695e-05, 'epoch': 1.04}) +[2026-06-23 20:30:28,832][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.2987653613090515, 'train/grad_norm_pre_clip': 20.92682647705078, 'timing/data': 0.0006047459319233894, 'timing/model': 1.6951166610233486, 'learning_rate/qwen_vl_interface': 4.8151145914954465e-06, 'learning_rate/action_model': 4.815114591495446e-05, 'epoch': 1.04}) +[2026-06-23 20:30:32,163][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.22930143773555756, 'train/grad_norm_pre_clip': 30.663509368896484, 'timing/data': 0.0005973631050437689, 'timing/model': 1.6914941768627614, 'learning_rate/qwen_vl_interface': 4.812082050292088e-06, 'learning_rate/action_model': 4.8120820502920876e-05, 'epoch': 1.04}) +[2026-06-23 20:30:35,504][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.08700201660394669, 'train/grad_norm_pre_clip': 24.259262084960938, 'timing/data': 0.0006080418825149536, 'timing/model': 1.7002481098752469, 'learning_rate/qwen_vl_interface': 4.809049689100256e-06, 'learning_rate/action_model': 4.8090496891002555e-05, 'epoch': 1.04}) +[2026-06-23 20:30:38,841][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.06677263230085373, 'train/grad_norm_pre_clip': 19.638565063476562, 'timing/data': 0.0005911309272050858, 'timing/model': 1.6938260160386562, 'learning_rate/qwen_vl_interface': 4.80601750916644e-06, 'learning_rate/action_model': 4.8060175091664404e-05, 'epoch': 1.04}) +[2026-06-23 20:30:42,177][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.08918830007314682, 'train/grad_norm_pre_clip': 47.381370544433594, 'timing/data': 0.0005865329876542091, 'timing/model': 1.6938418000936508, 'learning_rate/qwen_vl_interface': 4.802985511737057e-06, 'learning_rate/action_model': 4.802985511737057e-05, 'epoch': 1.04}) +[2026-06-23 20:30:45,511][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.08605794608592987, 'train/grad_norm_pre_clip': 28.840673446655273, 'timing/data': 0.0005817629862576723, 'timing/model': 1.6925660870037973, 'learning_rate/qwen_vl_interface': 4.799953698058444e-06, 'learning_rate/action_model': 4.7999536980584434e-05, 'epoch': 1.04}) +[2026-06-23 20:30:48,843][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.11421924829483032, 'train/grad_norm_pre_clip': 21.39519500732422, 'timing/data': 0.0005959600675851107, 'timing/model': 1.6926945510786027, 'learning_rate/qwen_vl_interface': 4.796922069376867e-06, 'learning_rate/action_model': 4.796922069376867e-05, 'epoch': 1.04}) +[2026-06-23 20:30:52,182][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2700 +[2026-06-23 20:32:25,297][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2700 +[2026-06-23 20:32:25,297][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2700 +[2026-06-23 20:33:36,952][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2700 +[2026-06-23 20:33:36,952][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.2047947198152542, 'train/grad_norm_pre_clip': 71.03958892822266, 'eval/action_loss/seconds': 93.11285197199322, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.08296409249305725, 'eval/latency_2/loss': 0.08296409249305725, 'eval/flappy/loss': 0.08296409249305725, 'eval/flappy/latency_2/loss': 0.08296409249305725, 'eval/flap_precision': 0.8685782556750299, 'eval/flap_recall': 0.9811066126855601, 'eval/flap_f1': 0.9214195183776933, 'eval/flap_rate_model': 0.2325, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9214195183776933, 'eval/latency_2/flap_precision': 0.8685782556750299, 'eval/latency_2/flap_recall': 0.9811066126855601, 'eval/latency_2/flap_f1': 0.9214195183776933, 'eval/latency_2/flap_rate_model': 0.2325, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9214195183776933, 'eval/action_classification/data_wait_seconds': 1.020983338356018, 'eval/action_classification/predict_seconds': 70.48956298828125, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005838568322360516, 'timing/model': 1.6949886160437018, 'learning_rate/qwen_vl_interface': 4.793890626938511e-06, 'learning_rate/action_model': 4.793890626938511e-05, 'epoch': 1.04}) +[2026-06-23 20:33:40,278][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.16048875451087952, 'train/grad_norm_pre_clip': 73.28594970703125, 'timing/data': 0.0005915819201618433, 'timing/model': 1.6885808201041073, 'learning_rate/qwen_vl_interface': 4.7908593719894914e-06, 'learning_rate/action_model': 4.790859371989491e-05, 'epoch': 1.05}) +[2026-06-23 20:33:43,604][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.04269305616617203, 'train/grad_norm_pre_clip': 43.31455612182617, 'timing/data': 0.0005714839790016413, 'timing/model': 1.6891671339981258, 'learning_rate/qwen_vl_interface': 4.78782830577584e-06, 'learning_rate/action_model': 4.7878283057758396e-05, 'epoch': 1.05}) +[2026-06-23 20:33:46,937][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.10821357369422913, 'train/grad_norm_pre_clip': 42.01861572265625, 'timing/data': 0.0008590950164943933, 'timing/model': 1.6951273679733276, 'learning_rate/qwen_vl_interface': 4.784797429543513e-06, 'learning_rate/action_model': 4.784797429543513e-05, 'epoch': 1.05}) +[2026-06-23 20:33:50,265][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.08327654749155045, 'train/grad_norm_pre_clip': 17.356353759765625, 'timing/data': 0.0006404730957001448, 'timing/model': 1.6908726340625435, 'learning_rate/qwen_vl_interface': 4.781766744538393e-06, 'learning_rate/action_model': 4.781766744538393e-05, 'epoch': 1.05}) +[2026-06-23 20:33:53,596][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.08496560901403427, 'train/grad_norm_pre_clip': 11.018411636352539, 'timing/data': 0.0005851390305906534, 'timing/model': 1.691316190874204, 'learning_rate/qwen_vl_interface': 4.778736252006277e-06, 'learning_rate/action_model': 4.7787362520062765e-05, 'epoch': 1.05}) +[2026-06-23 20:33:56,925][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.20372654497623444, 'train/grad_norm_pre_clip': 40.822853088378906, 'timing/data': 0.0005947379395365715, 'timing/model': 1.6905718450434506, 'learning_rate/qwen_vl_interface': 4.775705953192885e-06, 'learning_rate/action_model': 4.7757059531928856e-05, 'epoch': 1.05}) +[2026-06-23 20:34:00,257][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.2285110354423523, 'train/grad_norm_pre_clip': 15.884806632995605, 'timing/data': 0.0005799899809062481, 'timing/model': 1.6953852220904082, 'learning_rate/qwen_vl_interface': 4.7726758493438615e-06, 'learning_rate/action_model': 4.7726758493438613e-05, 'epoch': 1.05}) +[2026-06-23 20:34:03,588][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.12496177852153778, 'train/grad_norm_pre_clip': 69.90715026855469, 'timing/data': 0.0005820139776915312, 'timing/model': 1.6920398890506476, 'learning_rate/qwen_vl_interface': 4.769645941704766e-06, 'learning_rate/action_model': 4.769645941704766e-05, 'epoch': 1.05}) +[2026-06-23 20:34:06,941][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.1391664743423462, 'train/grad_norm_pre_clip': 32.210113525390625, 'timing/data': 0.0005943370051681995, 'timing/model': 1.6954243909567595, 'learning_rate/qwen_vl_interface': 4.766616231521081e-06, 'learning_rate/action_model': 4.7666162315210806e-05, 'epoch': 1.05}) +[2026-06-23 20:34:10,272][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.14915023744106293, 'train/grad_norm_pre_clip': 30.18094253540039, 'timing/data': 0.0005902789998799562, 'timing/model': 1.6940670730546117, 'learning_rate/qwen_vl_interface': 4.763586720038202e-06, 'learning_rate/action_model': 4.7635867200382014e-05, 'epoch': 1.05}) +[2026-06-23 20:34:13,608][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.06396596878767014, 'train/grad_norm_pre_clip': 18.865787506103516, 'timing/data': 0.0005809119902551174, 'timing/model': 1.695085918996483, 'learning_rate/qwen_vl_interface': 4.76055740850145e-06, 'learning_rate/action_model': 4.76055740850145e-05, 'epoch': 1.05}) +[2026-06-23 20:34:16,950][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.1705368161201477, 'train/grad_norm_pre_clip': 16.236772537231445, 'timing/data': 0.0007607301231473684, 'timing/model': 1.6920477629173547, 'learning_rate/qwen_vl_interface': 4.757528298156058e-06, 'learning_rate/action_model': 4.757528298156058e-05, 'epoch': 1.05}) +[2026-06-23 20:34:20,283][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.18251849710941315, 'train/grad_norm_pre_clip': 25.065723419189453, 'timing/data': 0.0006075210403650999, 'timing/model': 1.691027858061716, 'learning_rate/qwen_vl_interface': 4.754499390247181e-06, 'learning_rate/action_model': 4.754499390247181e-05, 'epoch': 1.05}) +[2026-06-23 20:34:23,617][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.05040767043828964, 'train/grad_norm_pre_clip': 18.571208953857422, 'timing/data': 0.0005836170166730881, 'timing/model': 1.6919341848697513, 'learning_rate/qwen_vl_interface': 4.751470686019887e-06, 'learning_rate/action_model': 4.751470686019888e-05, 'epoch': 1.05}) +[2026-06-23 20:34:27,051][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.16358645260334015, 'train/grad_norm_pre_clip': 54.10945129394531, 'timing/data': 0.0006876119878143072, 'timing/model': 1.6933027079794556, 'learning_rate/qwen_vl_interface': 4.7484421867191645e-06, 'learning_rate/action_model': 4.748442186719165e-05, 'epoch': 1.05}) +[2026-06-23 20:34:30,387][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.13300660252571106, 'train/grad_norm_pre_clip': 105.54557800292969, 'timing/data': 0.0006727341096848249, 'timing/model': 1.6919691448565573, 'learning_rate/qwen_vl_interface': 4.745413893589913e-06, 'learning_rate/action_model': 4.7454138935899125e-05, 'epoch': 1.05}) +[2026-06-23 20:34:33,730][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.16163645684719086, 'train/grad_norm_pre_clip': 68.12693786621094, 'timing/data': 0.0005849988665431738, 'timing/model': 1.6997678319457918, 'learning_rate/qwen_vl_interface': 4.7423858078769506e-06, 'learning_rate/action_model': 4.742385807876951e-05, 'epoch': 1.05}) +[2026-06-23 20:34:37,070][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.1340503692626953, 'train/grad_norm_pre_clip': 42.01642990112305, 'timing/data': 0.0006202959921211004, 'timing/model': 1.6989085299428552, 'learning_rate/qwen_vl_interface': 4.739357930825008e-06, 'learning_rate/action_model': 4.739357930825008e-05, 'epoch': 1.05}) +[2026-06-23 20:34:40,399][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.15664376318454742, 'train/grad_norm_pre_clip': 73.20355987548828, 'timing/data': 0.0005789080169051886, 'timing/model': 1.6901463170070201, 'learning_rate/qwen_vl_interface': 4.736330263678731e-06, 'learning_rate/action_model': 4.736330263678731e-05, 'epoch': 1.05}) +[2026-06-23 20:34:43,743][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.058494072407484055, 'train/grad_norm_pre_clip': 32.06680679321289, 'timing/data': 0.0006936541758477688, 'timing/model': 1.6952425360213965, 'learning_rate/qwen_vl_interface': 4.733302807682679e-06, 'learning_rate/action_model': 4.733302807682679e-05, 'epoch': 1.05}) +[2026-06-23 20:34:47,087][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.14596542716026306, 'train/grad_norm_pre_clip': 62.57273483276367, 'timing/data': 0.0007982200477272272, 'timing/model': 1.704938359092921, 'learning_rate/qwen_vl_interface': 4.7302755640813255e-06, 'learning_rate/action_model': 4.7302755640813256e-05, 'epoch': 1.05}) +[2026-06-23 20:34:50,429][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.0813005119562149, 'train/grad_norm_pre_clip': 32.02793502807617, 'timing/data': 0.0005942170973867178, 'timing/model': 1.6989080188795924, 'learning_rate/qwen_vl_interface': 4.727248534119058e-06, 'learning_rate/action_model': 4.7272485341190574e-05, 'epoch': 1.05}) +[2026-06-23 20:34:53,777][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.17833682894706726, 'train/grad_norm_pre_clip': 15.764238357543945, 'timing/data': 0.0005708220414817333, 'timing/model': 1.7010348469484597, 'learning_rate/qwen_vl_interface': 4.72422171904017e-06, 'learning_rate/action_model': 4.7242217190401697e-05, 'epoch': 1.05}) +[2026-06-23 20:34:57,128][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.12628786265850067, 'train/grad_norm_pre_clip': 65.73113250732422, 'timing/data': 0.0005883350968360901, 'timing/model': 1.6972193599212915, 'learning_rate/qwen_vl_interface': 4.721195120088873e-06, 'learning_rate/action_model': 4.721195120088873e-05, 'epoch': 1.05}) +[2026-06-23 20:35:00,469][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.21619603037834167, 'train/grad_norm_pre_clip': 54.790348052978516, 'timing/data': 0.0005886259023100138, 'timing/model': 1.7013048501685262, 'learning_rate/qwen_vl_interface': 4.718168738509288e-06, 'learning_rate/action_model': 4.718168738509288e-05, 'epoch': 1.05}) +[2026-06-23 20:35:03,811][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.09298969805240631, 'train/grad_norm_pre_clip': 19.178489685058594, 'timing/data': 0.0005948671605437994, 'timing/model': 1.6976394429802895, 'learning_rate/qwen_vl_interface': 4.715142575545445e-06, 'learning_rate/action_model': 4.7151425755454454e-05, 'epoch': 1.06}) +[2026-06-23 20:35:07,155][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.1942058652639389, 'train/grad_norm_pre_clip': 39.259307861328125, 'timing/data': 0.0005816030316054821, 'timing/model': 1.6977836468722671, 'learning_rate/qwen_vl_interface': 4.712116632441287e-06, 'learning_rate/action_model': 4.7121166324412866e-05, 'epoch': 1.06}) +[2026-06-23 20:35:10,499][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.1938447505235672, 'train/grad_norm_pre_clip': 21.058549880981445, 'timing/data': 0.0006001479923725128, 'timing/model': 1.7003885752055794, 'learning_rate/qwen_vl_interface': 4.709090910440664e-06, 'learning_rate/action_model': 4.7090909104406634e-05, 'epoch': 1.06}) +[2026-06-23 20:35:13,944][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.19144046306610107, 'train/grad_norm_pre_clip': 24.67723274230957, 'timing/data': 0.0005928941536694765, 'timing/model': 1.8001651370432228, 'learning_rate/qwen_vl_interface': 4.7060654107873335e-06, 'learning_rate/action_model': 4.7060654107873335e-05, 'epoch': 1.06}) +[2026-06-23 20:35:17,282][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.08217490464448929, 'train/grad_norm_pre_clip': 16.42463493347168, 'timing/data': 0.0006015910767018795, 'timing/model': 1.6949211140163243, 'learning_rate/qwen_vl_interface': 4.7030401347249685e-06, 'learning_rate/action_model': 4.703040134724968e-05, 'epoch': 1.06}) +[2026-06-23 20:35:20,622][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.13044001162052155, 'train/grad_norm_pre_clip': 47.06324768066406, 'timing/data': 0.0006625649984925985, 'timing/model': 1.6962496631313115, 'learning_rate/qwen_vl_interface': 4.700015083497143e-06, 'learning_rate/action_model': 4.700015083497142e-05, 'epoch': 1.06}) +[2026-06-23 20:35:23,968][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.20515763759613037, 'train/grad_norm_pre_clip': 17.96688461303711, 'timing/data': 0.0005652620457112789, 'timing/model': 1.7003180698957294, 'learning_rate/qwen_vl_interface': 4.696990258347342e-06, 'learning_rate/action_model': 4.696990258347342e-05, 'epoch': 1.06}) +[2026-06-23 20:35:27,312][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.13698440790176392, 'train/grad_norm_pre_clip': 27.66927146911621, 'timing/data': 0.0005868228618055582, 'timing/model': 1.6995581551454961, 'learning_rate/qwen_vl_interface': 4.693965660518957e-06, 'learning_rate/action_model': 4.693965660518957e-05, 'epoch': 1.06}) +[2026-06-23 20:35:30,654][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.1083354651927948, 'train/grad_norm_pre_clip': 15.450630187988281, 'timing/data': 0.0005859609227627516, 'timing/model': 1.6967210650909692, 'learning_rate/qwen_vl_interface': 4.690941291255285e-06, 'learning_rate/action_model': 4.6909412912552844e-05, 'epoch': 1.06}) +[2026-06-23 20:35:34,000][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.1091318279504776, 'train/grad_norm_pre_clip': 33.57244110107422, 'timing/data': 0.0005782360676676035, 'timing/model': 1.6984290268737823, 'learning_rate/qwen_vl_interface': 4.68791715179953e-06, 'learning_rate/action_model': 4.68791715179953e-05, 'epoch': 1.06}) +[2026-06-23 20:35:37,346][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.13694603741168976, 'train/grad_norm_pre_clip': 43.4735107421875, 'timing/data': 0.000587032875046134, 'timing/model': 1.6983717910479754, 'learning_rate/qwen_vl_interface': 4.684893243394804e-06, 'learning_rate/action_model': 4.6848932433948036e-05, 'epoch': 1.06}) +[2026-06-23 20:35:40,682][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.233537495136261, 'train/grad_norm_pre_clip': 23.247947692871094, 'timing/data': 0.0005873739719390869, 'timing/model': 1.6942724101245403, 'learning_rate/qwen_vl_interface': 4.68186956728412e-06, 'learning_rate/action_model': 4.68186956728412e-05, 'epoch': 1.06}) +[2026-06-23 20:35:44,016][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.09487637132406235, 'train/grad_norm_pre_clip': 42.28915023803711, 'timing/data': 0.0005915309302508831, 'timing/model': 1.6922957068309188, 'learning_rate/qwen_vl_interface': 4.678846124710395e-06, 'learning_rate/action_model': 4.678846124710394e-05, 'epoch': 1.06}) +[2026-06-23 20:35:47,362][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.11867000162601471, 'train/grad_norm_pre_clip': 20.953685760498047, 'timing/data': 0.0006277901120483875, 'timing/model': 1.7001657790970057, 'learning_rate/qwen_vl_interface': 4.675822916916454e-06, 'learning_rate/action_model': 4.6758229169164536e-05, 'epoch': 1.06}) +[2026-06-23 20:35:50,707][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.15049806237220764, 'train/grad_norm_pre_clip': 44.76565170288086, 'timing/data': 0.0005700620822608471, 'timing/model': 1.6992898650933057, 'learning_rate/qwen_vl_interface': 4.6727999451450236e-06, 'learning_rate/action_model': 4.672799945145023e-05, 'epoch': 1.06}) +[2026-06-23 20:35:54,047][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.17358717322349548, 'train/grad_norm_pre_clip': 58.56324768066406, 'timing/data': 0.0005928638856858015, 'timing/model': 1.6959335999563336, 'learning_rate/qwen_vl_interface': 4.669777210638733e-06, 'learning_rate/action_model': 4.669777210638733e-05, 'epoch': 1.06}) +[2026-06-23 20:35:57,392][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.13139839470386505, 'train/grad_norm_pre_clip': 21.52408218383789, 'timing/data': 0.0005912110209465027, 'timing/model': 1.6971706270705909, 'learning_rate/qwen_vl_interface': 4.666754714640115e-06, 'learning_rate/action_model': 4.666754714640115e-05, 'epoch': 1.06}) +[2026-06-23 20:36:00,736][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.07490422576665878, 'train/grad_norm_pre_clip': 16.43394660949707, 'timing/data': 0.0005904200952500105, 'timing/model': 1.6982946610078216, 'learning_rate/qwen_vl_interface': 4.663732458391602e-06, 'learning_rate/action_model': 4.663732458391601e-05, 'epoch': 1.06}) +[2026-06-23 20:36:04,242][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.0762254148721695, 'train/grad_norm_pre_clip': 13.050873756408691, 'timing/data': 0.0007971581071615219, 'timing/model': 1.857765463879332, 'learning_rate/qwen_vl_interface': 4.6607104431355295e-06, 'learning_rate/action_model': 4.66071044313553e-05, 'epoch': 1.06}) +[2026-06-23 20:36:07,618][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.1677754521369934, 'train/grad_norm_pre_clip': 16.489288330078125, 'timing/data': 0.0005913209170103073, 'timing/model': 1.695959737058729, 'learning_rate/qwen_vl_interface': 4.657688670114135e-06, 'learning_rate/action_model': 4.657688670114135e-05, 'epoch': 1.06}) +[2026-06-23 20:36:10,960][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.12482818961143494, 'train/grad_norm_pre_clip': 18.0114803314209, 'timing/data': 0.0005882459226995707, 'timing/model': 1.6977790880482644, 'learning_rate/qwen_vl_interface': 4.654667140569555e-06, 'learning_rate/action_model': 4.6546671405695544e-05, 'epoch': 1.06}) +[2026-06-23 20:36:14,299][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.1274772584438324, 'train/grad_norm_pre_clip': 47.6586799621582, 'timing/data': 0.0006109080277383327, 'timing/model': 1.695600335020572, 'learning_rate/qwen_vl_interface': 4.651645855743825e-06, 'learning_rate/action_model': 4.651645855743825e-05, 'epoch': 1.06}) +[2026-06-23 20:36:17,637][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.09947016090154648, 'train/grad_norm_pre_clip': 31.364465713500977, 'timing/data': 0.0005782160442322493, 'timing/model': 1.698353541083634, 'learning_rate/qwen_vl_interface': 4.648624816878883e-06, 'learning_rate/action_model': 4.6486248168788825e-05, 'epoch': 1.06}) +[2026-06-23 20:36:20,977][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.09371204674243927, 'train/grad_norm_pre_clip': 23.94754981994629, 'timing/data': 0.0005842479877173901, 'timing/model': 1.6959253100212663, 'learning_rate/qwen_vl_interface': 4.6456040252165615e-06, 'learning_rate/action_model': 4.645604025216562e-05, 'epoch': 1.06}) +[2026-06-23 20:37:15,558][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2750 +[2026-06-23 20:37:15,559][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.12434335798025131, 'train/grad_norm_pre_clip': 24.796096801757812, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 36.49999999999981, 'rl_games_eval/mid_train/mean_length': 313.0, 'rl_games_eval/mid_train/std_reward': 34.30655913961595, 'rl_games_eval/mid_train/std_length': 276.92959394040935, 'rl_games_eval/mid_train/macro_mean_reward': 36.49999999999981, 'rl_games_eval/mid_train/macro_mean_length': 313.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 36.49999999999981, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 313.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 34.30655913961595, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 276.92959394040935, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 36.49999999999981, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005781659856438637, 'timing/model': 1.6949954140000045, 'learning_rate/qwen_vl_interface': 4.642583481998596e-06, 'learning_rate/action_model': 4.642583481998596e-05, 'epoch': 1.06}) +[2026-06-23 20:37:18,862][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.15829290449619293, 'train/grad_norm_pre_clip': 52.3491096496582, 'timing/data': 0.000592764001339674, 'timing/model': 1.6736505820881575, 'learning_rate/qwen_vl_interface': 4.639563188466616e-06, 'learning_rate/action_model': 4.6395631884666165e-05, 'epoch': 1.06}) +[2026-06-23 20:37:22,164][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.12775732576847076, 'train/grad_norm_pre_clip': 52.743629455566406, 'timing/data': 0.0006316069047898054, 'timing/model': 1.679764192784205, 'learning_rate/qwen_vl_interface': 4.636543145862152e-06, 'learning_rate/action_model': 4.6365431458621523e-05, 'epoch': 1.07}) +[2026-06-23 20:37:25,471][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.1499098688364029, 'train/grad_norm_pre_clip': 82.12541961669922, 'timing/data': 0.0006795269437134266, 'timing/model': 1.6830558800138533, 'learning_rate/qwen_vl_interface': 4.633523355426629e-06, 'learning_rate/action_model': 4.6335233554266295e-05, 'epoch': 1.07}) +[2026-06-23 20:37:28,777][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.09625811874866486, 'train/grad_norm_pre_clip': 49.121337890625, 'timing/data': 0.0005926438607275486, 'timing/model': 1.6776726958341897, 'learning_rate/qwen_vl_interface': 4.6305038184013675e-06, 'learning_rate/action_model': 4.6305038184013675e-05, 'epoch': 1.07}) +[2026-06-23 20:37:32,091][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.22321492433547974, 'train/grad_norm_pre_clip': 55.55046463012695, 'timing/data': 0.0005768542177975178, 'timing/model': 1.6851607989519835, 'learning_rate/qwen_vl_interface': 4.627484536027589e-06, 'learning_rate/action_model': 4.627484536027588e-05, 'epoch': 1.07}) +[2026-06-23 20:37:35,410][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.05250529572367668, 'train/grad_norm_pre_clip': 19.810029983520508, 'timing/data': 0.0006358749233186245, 'timing/model': 1.6871928591281176, 'learning_rate/qwen_vl_interface': 4.624465509546399e-06, 'learning_rate/action_model': 4.6244655095463986e-05, 'epoch': 1.07}) +[2026-06-23 20:37:39,035][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.1360381841659546, 'train/grad_norm_pre_clip': 30.750696182250977, 'timing/data': 0.0006064600311219692, 'timing/model': 1.6894503941293806, 'learning_rate/qwen_vl_interface': 4.6214467401988106e-06, 'learning_rate/action_model': 4.621446740198811e-05, 'epoch': 1.07}) +[2026-06-23 20:37:42,350][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.13329140841960907, 'train/grad_norm_pre_clip': 17.786602020263672, 'timing/data': 0.0009703431278467178, 'timing/model': 1.6815049031283706, 'learning_rate/qwen_vl_interface': 4.6184282292257245e-06, 'learning_rate/action_model': 4.618428229225725e-05, 'epoch': 1.07}) +[2026-06-23 20:37:45,670][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.07962650060653687, 'train/grad_norm_pre_clip': 44.27765655517578, 'timing/data': 0.0010346241760998964, 'timing/model': 1.6891163990367204, 'learning_rate/qwen_vl_interface': 4.615409977867938e-06, 'learning_rate/action_model': 4.615409977867937e-05, 'epoch': 1.07}) +[2026-06-23 20:37:48,998][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.05960964038968086, 'train/grad_norm_pre_clip': 36.65352249145508, 'timing/data': 0.0005773648153990507, 'timing/model': 1.6890875769313425, 'learning_rate/qwen_vl_interface': 4.612391987366138e-06, 'learning_rate/action_model': 4.612391987366138e-05, 'epoch': 1.07}) +[2026-06-23 20:37:52,316][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.1250167042016983, 'train/grad_norm_pre_clip': 19.373380661010742, 'timing/data': 0.0005819541402161121, 'timing/model': 1.68928029993549, 'learning_rate/qwen_vl_interface': 4.609374258960908e-06, 'learning_rate/action_model': 4.6093742589609074e-05, 'epoch': 1.07}) +[2026-06-23 20:37:55,640][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.18963858485221863, 'train/grad_norm_pre_clip': 22.354978561401367, 'timing/data': 0.0006742270197719336, 'timing/model': 1.6883204989135265, 'learning_rate/qwen_vl_interface': 4.60635679389272e-06, 'learning_rate/action_model': 4.60635679389272e-05, 'epoch': 1.07}) +[2026-06-23 20:37:58,963][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.21208345890045166, 'train/grad_norm_pre_clip': 23.1344051361084, 'timing/data': 0.0005972019862383604, 'timing/model': 1.6872696550562978, 'learning_rate/qwen_vl_interface': 4.603339593401943e-06, 'learning_rate/action_model': 4.603339593401942e-05, 'epoch': 1.07}) +[2026-06-23 20:38:02,290][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.14985699951648712, 'train/grad_norm_pre_clip': 15.393226623535156, 'timing/data': 0.0006849570199847221, 'timing/model': 1.6926748051773757, 'learning_rate/qwen_vl_interface': 4.600322658728832e-06, 'learning_rate/action_model': 4.600322658728832e-05, 'epoch': 1.07}) +[2026-06-23 20:38:05,620][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.13326133787631989, 'train/grad_norm_pre_clip': 14.667786598205566, 'timing/data': 0.0005874440539628267, 'timing/model': 1.6913703540340066, 'learning_rate/qwen_vl_interface': 4.597305991113537e-06, 'learning_rate/action_model': 4.597305991113537e-05, 'epoch': 1.07}) +[2026-06-23 20:38:08,960][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.14383065700531006, 'train/grad_norm_pre_clip': 47.940956115722656, 'timing/data': 0.0005874938797205687, 'timing/model': 1.6989783539902419, 'learning_rate/qwen_vl_interface': 4.5942895917960935e-06, 'learning_rate/action_model': 4.5942895917960935e-05, 'epoch': 1.07}) +[2026-06-23 20:38:12,293][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.16624312102794647, 'train/grad_norm_pre_clip': 32.73395919799805, 'timing/data': 0.0008087500464171171, 'timing/model': 1.695525827119127, 'learning_rate/qwen_vl_interface': 4.59127346201643e-06, 'learning_rate/action_model': 4.59127346201643e-05, 'epoch': 1.07}) +[2026-06-23 20:38:15,624][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.0987350195646286, 'train/grad_norm_pre_clip': 18.000165939331055, 'timing/data': 0.0006294630002230406, 'timing/model': 1.6900325750466436, 'learning_rate/qwen_vl_interface': 4.588257603014364e-06, 'learning_rate/action_model': 4.588257603014364e-05, 'epoch': 1.07}) +[2026-06-23 20:38:18,959][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.08786818385124207, 'train/grad_norm_pre_clip': 14.532655715942383, 'timing/data': 0.0006008991040289402, 'timing/model': 1.6933679939247668, 'learning_rate/qwen_vl_interface': 4.585242016029602e-06, 'learning_rate/action_model': 4.5852420160296015e-05, 'epoch': 1.07}) +[2026-06-23 20:38:22,290][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.1540890634059906, 'train/grad_norm_pre_clip': 31.74442481994629, 'timing/data': 0.0005803711246699095, 'timing/model': 1.6905160769820213, 'learning_rate/qwen_vl_interface': 4.582226702301737e-06, 'learning_rate/action_model': 4.5822267023017375e-05, 'epoch': 1.07}) +[2026-06-23 20:38:25,722][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.10054042190313339, 'train/grad_norm_pre_clip': 11.69548511505127, 'timing/data': 0.0007068889681249857, 'timing/model': 1.797777330968529, 'learning_rate/qwen_vl_interface': 4.57921166307025e-06, 'learning_rate/action_model': 4.57921166307025e-05, 'epoch': 1.07}) +[2026-06-23 20:38:29,048][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.0378321148455143, 'train/grad_norm_pre_clip': 34.801719665527344, 'timing/data': 0.0008131179492920637, 'timing/model': 1.6901949769817293, 'learning_rate/qwen_vl_interface': 4.576196899574511e-06, 'learning_rate/action_model': 4.576196899574511e-05, 'epoch': 1.07}) +[2026-06-23 20:38:32,381][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.11936876177787781, 'train/grad_norm_pre_clip': 42.227455139160156, 'timing/data': 0.0005909400060772896, 'timing/model': 1.6943991878069937, 'learning_rate/qwen_vl_interface': 4.573182413053775e-06, 'learning_rate/action_model': 4.5731824130537746e-05, 'epoch': 1.07}) +[2026-06-23 20:38:35,711][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.1346617043018341, 'train/grad_norm_pre_clip': 61.910194396972656, 'timing/data': 0.0005796591285616159, 'timing/model': 1.689040589146316, 'learning_rate/qwen_vl_interface': 4.570168204747183e-06, 'learning_rate/action_model': 4.570168204747183e-05, 'epoch': 1.07}) +[2026-06-23 20:38:39,048][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.06320895254611969, 'train/grad_norm_pre_clip': 15.116988182067871, 'timing/data': 0.0005885458085685968, 'timing/model': 1.6954264119267464, 'learning_rate/qwen_vl_interface': 4.567154275893762e-06, 'learning_rate/action_model': 4.5671542758937624e-05, 'epoch': 1.07}) +[2026-06-23 20:38:42,379][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.16229335963726044, 'train/grad_norm_pre_clip': 18.407066345214844, 'timing/data': 0.0005874040070921183, 'timing/model': 1.6920067057944834, 'learning_rate/qwen_vl_interface': 4.564140627732426e-06, 'learning_rate/action_model': 4.564140627732426e-05, 'epoch': 1.07}) +[2026-06-23 20:38:45,714][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.07253733277320862, 'train/grad_norm_pre_clip': 9.3081693649292, 'timing/data': 0.0005906498990952969, 'timing/model': 1.695982659002766, 'learning_rate/qwen_vl_interface': 4.56112726150197e-06, 'learning_rate/action_model': 4.56112726150197e-05, 'epoch': 1.07}) +[2026-06-23 20:38:49,046][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.20590922236442566, 'train/grad_norm_pre_clip': 37.822776794433594, 'timing/data': 0.0005808810237795115, 'timing/model': 1.6917439349927008, 'learning_rate/qwen_vl_interface': 4.558114178441076e-06, 'learning_rate/action_model': 4.558114178441076e-05, 'epoch': 1.08}) +[2026-06-23 20:38:52,375][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.14053680002689362, 'train/grad_norm_pre_clip': 30.40304183959961, 'timing/data': 0.0005809119902551174, 'timing/model': 1.6908822879195213, 'learning_rate/qwen_vl_interface': 4.555101379788308e-06, 'learning_rate/action_model': 4.555101379788308e-05, 'epoch': 1.08}) +[2026-06-23 20:38:55,707][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.1309346705675125, 'train/grad_norm_pre_clip': 18.253963470458984, 'timing/data': 0.0005811720620840788, 'timing/model': 1.689048483967781, 'learning_rate/qwen_vl_interface': 4.552088866782115e-06, 'learning_rate/action_model': 4.552088866782115e-05, 'epoch': 1.08}) +[2026-06-23 20:38:59,038][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.08342054486274719, 'train/grad_norm_pre_clip': 23.54044532775879, 'timing/data': 0.0006036041304469109, 'timing/model': 1.6912241040263325, 'learning_rate/qwen_vl_interface': 4.5490766406608265e-06, 'learning_rate/action_model': 4.549076640660827e-05, 'epoch': 1.08}) +[2026-06-23 20:39:02,370][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.13113805651664734, 'train/grad_norm_pre_clip': 34.7166633605957, 'timing/data': 0.0008116147946566343, 'timing/model': 1.6920544551685452, 'learning_rate/qwen_vl_interface': 4.546064702662655e-06, 'learning_rate/action_model': 4.546064702662655e-05, 'epoch': 1.08}) +[2026-06-23 20:39:05,702][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.0760134682059288, 'train/grad_norm_pre_clip': 41.281028747558594, 'timing/data': 0.0006312669720500708, 'timing/model': 1.6919990919996053, 'learning_rate/qwen_vl_interface': 4.543053054025693e-06, 'learning_rate/action_model': 4.5430530540256926e-05, 'epoch': 1.08}) +[2026-06-23 20:39:09,036][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.15926603972911835, 'train/grad_norm_pre_clip': 61.06229019165039, 'timing/data': 0.0007324160542339087, 'timing/model': 1.6917518391273916, 'learning_rate/qwen_vl_interface': 4.540041695987917e-06, 'learning_rate/action_model': 4.540041695987917e-05, 'epoch': 1.08}) +[2026-06-23 20:39:12,364][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.09634005278348923, 'train/grad_norm_pre_clip': 14.399130821228027, 'timing/data': 0.0006279509980231524, 'timing/model': 1.6916962759569287, 'learning_rate/qwen_vl_interface': 4.537030629787182e-06, 'learning_rate/action_model': 4.5370306297871826e-05, 'epoch': 1.08}) +[2026-06-23 20:39:15,781][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.13448457419872284, 'train/grad_norm_pre_clip': 31.308137893676758, 'timing/data': 0.0007296809926629066, 'timing/model': 1.6845716692041606, 'learning_rate/qwen_vl_interface': 4.534019856661226e-06, 'learning_rate/action_model': 4.534019856661226e-05, 'epoch': 1.08}) +[2026-06-23 20:39:19,113][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.10630755871534348, 'train/grad_norm_pre_clip': 26.371084213256836, 'timing/data': 0.0005939260590821505, 'timing/model': 1.6944353559520096, 'learning_rate/qwen_vl_interface': 4.531009377847661e-06, 'learning_rate/action_model': 4.531009377847661e-05, 'epoch': 1.08}) +[2026-06-23 20:39:22,447][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.1392417550086975, 'train/grad_norm_pre_clip': 23.395864486694336, 'timing/data': 0.000574049074202776, 'timing/model': 1.692196981050074, 'learning_rate/qwen_vl_interface': 4.527999194583982e-06, 'learning_rate/action_model': 4.527999194583981e-05, 'epoch': 1.08}) +[2026-06-23 20:39:25,775][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.07496250420808792, 'train/grad_norm_pre_clip': 17.613428115844727, 'timing/data': 0.0005703021306544542, 'timing/model': 1.6895782898645848, 'learning_rate/qwen_vl_interface': 4.52498930810756e-06, 'learning_rate/action_model': 4.52498930810756e-05, 'epoch': 1.08}) +[2026-06-23 20:39:29,105][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.06110701337456703, 'train/grad_norm_pre_clip': 33.07551956176758, 'timing/data': 0.0005733580328524113, 'timing/model': 1.6925733010284603, 'learning_rate/qwen_vl_interface': 4.521979719655648e-06, 'learning_rate/action_model': 4.5219797196556484e-05, 'epoch': 1.08}) +[2026-06-23 20:39:32,436][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.3100825250148773, 'train/grad_norm_pre_clip': 91.64262390136719, 'timing/data': 0.0005752409342676401, 'timing/model': 1.6926277941092849, 'learning_rate/qwen_vl_interface': 4.518970430465377e-06, 'learning_rate/action_model': 4.518970430465376e-05, 'epoch': 1.08}) +[2026-06-23 20:39:35,764][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.0718279778957367, 'train/grad_norm_pre_clip': 47.29914474487305, 'timing/data': 0.0005997270345687866, 'timing/model': 1.6880870319437236, 'learning_rate/qwen_vl_interface': 4.515961441773745e-06, 'learning_rate/action_model': 4.515961441773745e-05, 'epoch': 1.08}) +[2026-06-23 20:39:39,087][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.08911225944757462, 'train/grad_norm_pre_clip': 73.46576690673828, 'timing/data': 0.0005810409784317017, 'timing/model': 1.6896742458920926, 'learning_rate/qwen_vl_interface': 4.5129527548176385e-06, 'learning_rate/action_model': 4.512952754817638e-05, 'epoch': 1.08}) +[2026-06-23 20:39:42,416][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.13314707577228546, 'train/grad_norm_pre_clip': 24.080097198486328, 'timing/data': 0.00061950390227139, 'timing/model': 1.695794879924506, 'learning_rate/qwen_vl_interface': 4.509944370833813e-06, 'learning_rate/action_model': 4.509944370833813e-05, 'epoch': 1.08}) +[2026-06-23 20:39:45,749][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.08243392407894135, 'train/grad_norm_pre_clip': 19.813247680664062, 'timing/data': 0.0006402940489351749, 'timing/model': 1.6960068801417947, 'learning_rate/qwen_vl_interface': 4.506936291058902e-06, 'learning_rate/action_model': 4.506936291058902e-05, 'epoch': 1.08}) +[2026-06-23 20:39:49,084][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.18417538702487946, 'train/grad_norm_pre_clip': 38.479671478271484, 'timing/data': 0.000600920058786869, 'timing/model': 1.6926541801076382, 'learning_rate/qwen_vl_interface': 4.503928516729416e-06, 'learning_rate/action_model': 4.503928516729416e-05, 'epoch': 1.08}) +[2026-06-23 20:39:52,411][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.09775335341691971, 'train/grad_norm_pre_clip': 14.321935653686523, 'timing/data': 0.0006823630537837744, 'timing/model': 1.691103236982599, 'learning_rate/qwen_vl_interface': 4.500921049081734e-06, 'learning_rate/action_model': 4.500921049081734e-05, 'epoch': 1.08}) +[2026-06-23 20:39:55,735][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.20531274378299713, 'train/grad_norm_pre_clip': 17.377349853515625, 'timing/data': 0.0006682060193270445, 'timing/model': 1.6880703091155738, 'learning_rate/qwen_vl_interface': 4.497913889352113e-06, 'learning_rate/action_model': 4.4979138893521126e-05, 'epoch': 1.08}) +[2026-06-23 20:39:59,071][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.21721866726875305, 'train/grad_norm_pre_clip': 51.406742095947266, 'timing/data': 0.0006005880422890186, 'timing/model': 1.6938943949062377, 'learning_rate/qwen_vl_interface': 4.494907038776684e-06, 'learning_rate/action_model': 4.4949070387766836e-05, 'epoch': 1.08}) +[2026-06-23 20:40:02,509][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2800 +[2026-06-23 20:41:36,060][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2800 +[2026-06-23 20:41:36,061][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2800 +[2026-06-23 20:42:47,116][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2800 +[2026-06-23 20:42:47,117][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.22774985432624817, 'train/grad_norm_pre_clip': 18.701215744018555, 'eval/action_loss/seconds': 93.54900268092752, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07061553746461868, 'eval/latency_2/loss': 0.07061553746461868, 'eval/flappy/loss': 0.07061553746461868, 'eval/flappy/latency_2/loss': 0.07061553746461868, 'eval/flap_precision': 0.9146735617323852, 'eval/flap_recall': 0.9547908232118758, 'eval/flap_f1': 0.9343017497523934, 'eval/flap_rate_model': 0.2148611111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9343017497523933, 'eval/latency_2/flap_precision': 0.9146735617323852, 'eval/latency_2/flap_recall': 0.9547908232118758, 'eval/latency_2/flap_f1': 0.9343017497523934, 'eval/latency_2/flap_rate_model': 0.2148611111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9343017497523933, 'eval/action_classification/data_wait_seconds': 0.8375265598297119, 'eval/action_classification/predict_seconds': 70.06946563720703, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005916818045079708, 'timing/model': 1.8007436560001224, 'learning_rate/qwen_vl_interface': 4.49190049859145e-06, 'learning_rate/action_model': 4.49190049859145e-05, 'epoch': 1.08}) +[2026-06-23 20:42:50,430][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.06349273025989532, 'train/grad_norm_pre_clip': 22.045358657836914, 'timing/data': 0.0005919418763369322, 'timing/model': 1.6840319740585983, 'learning_rate/qwen_vl_interface': 4.488894270032284e-06, 'learning_rate/action_model': 4.4888942700322834e-05, 'epoch': 1.08}) +[2026-06-23 20:42:53,746][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.15156139433383942, 'train/grad_norm_pre_clip': 28.440467834472656, 'timing/data': 0.0005914811044931412, 'timing/model': 1.686140799196437, 'learning_rate/qwen_vl_interface': 4.485888354334933e-06, 'learning_rate/action_model': 4.4858883543349334e-05, 'epoch': 1.08}) +[2026-06-23 20:42:57,060][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.17166876792907715, 'train/grad_norm_pre_clip': 22.21075439453125, 'timing/data': 0.0006074218545109034, 'timing/model': 1.682976092910394, 'learning_rate/qwen_vl_interface': 4.482882752735017e-06, 'learning_rate/action_model': 4.482882752735017e-05, 'epoch': 1.08}) +[2026-06-23 20:43:00,381][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.18238487839698792, 'train/grad_norm_pre_clip': 26.379539489746094, 'timing/data': 0.0005890070460736752, 'timing/model': 1.6846026701387018, 'learning_rate/qwen_vl_interface': 4.479877466468027e-06, 'learning_rate/action_model': 4.479877466468026e-05, 'epoch': 1.09}) +[2026-06-23 20:43:03,700][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.02186562679708004, 'train/grad_norm_pre_clip': 12.400192260742188, 'timing/data': 0.0005861509125679731, 'timing/model': 1.6855082020629197, 'learning_rate/qwen_vl_interface': 4.476872496769318e-06, 'learning_rate/action_model': 4.476872496769318e-05, 'epoch': 1.09}) +[2026-06-23 20:43:07,025][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.09035497158765793, 'train/grad_norm_pre_clip': 21.922346115112305, 'timing/data': 0.0005879448726773262, 'timing/model': 1.6885242308489978, 'learning_rate/qwen_vl_interface': 4.473867844874122e-06, 'learning_rate/action_model': 4.473867844874122e-05, 'epoch': 1.09}) +[2026-06-23 20:43:10,341][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.11239080876111984, 'train/grad_norm_pre_clip': 38.916229248046875, 'timing/data': 0.0005939661059528589, 'timing/model': 1.6840141259599477, 'learning_rate/qwen_vl_interface': 4.470863512017539e-06, 'learning_rate/action_model': 4.470863512017539e-05, 'epoch': 1.09}) +[2026-06-23 20:43:13,752][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.29007887840270996, 'train/grad_norm_pre_clip': 79.22608184814453, 'timing/data': 0.0005950480699539185, 'timing/model': 1.776993105886504, 'learning_rate/qwen_vl_interface': 4.4678594994345344e-06, 'learning_rate/action_model': 4.467859499434534e-05, 'epoch': 1.09}) +[2026-06-23 20:43:17,078][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.24042581021785736, 'train/grad_norm_pre_clip': 35.30811309814453, 'timing/data': 0.0005928138270974159, 'timing/model': 1.6898570088669658, 'learning_rate/qwen_vl_interface': 4.4648558083599475e-06, 'learning_rate/action_model': 4.464855808359947e-05, 'epoch': 1.09}) +[2026-06-23 20:43:20,400][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.17888562381267548, 'train/grad_norm_pre_clip': 36.58561325073242, 'timing/data': 0.0008360510692000389, 'timing/model': 1.687633692054078, 'learning_rate/qwen_vl_interface': 4.461852440028479e-06, 'learning_rate/action_model': 4.461852440028479e-05, 'epoch': 1.09}) +[2026-06-23 20:43:23,722][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.11989689618349075, 'train/grad_norm_pre_clip': 20.69643211364746, 'timing/data': 0.0005977130495011806, 'timing/model': 1.686936290934682, 'learning_rate/qwen_vl_interface': 4.458849395674702e-06, 'learning_rate/action_model': 4.458849395674701e-05, 'epoch': 1.09}) +[2026-06-23 20:43:27,051][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.08474363386631012, 'train/grad_norm_pre_clip': 9.85241985321045, 'timing/data': 0.000576153164729476, 'timing/model': 1.6896622749045491, 'learning_rate/qwen_vl_interface': 4.455846676533052e-06, 'learning_rate/action_model': 4.4558466765330525e-05, 'epoch': 1.09}) +[2026-06-23 20:43:30,375][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.1295030266046524, 'train/grad_norm_pre_clip': 58.190574645996094, 'timing/data': 0.0007245619781315327, 'timing/model': 1.6900804520118982, 'learning_rate/qwen_vl_interface': 4.452844283837838e-06, 'learning_rate/action_model': 4.452844283837838e-05, 'epoch': 1.09}) +[2026-06-23 20:43:33,700][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.22777973115444183, 'train/grad_norm_pre_clip': 30.469863891601562, 'timing/data': 0.000738547882065177, 'timing/model': 1.6863253561314195, 'learning_rate/qwen_vl_interface': 4.449842218823228e-06, 'learning_rate/action_model': 4.449842218823228e-05, 'epoch': 1.09}) +[2026-06-23 20:43:37,033][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.19792330265045166, 'train/grad_norm_pre_clip': 41.308143615722656, 'timing/data': 0.0006781150586903095, 'timing/model': 1.6937744009774178, 'learning_rate/qwen_vl_interface': 4.446840482723256e-06, 'learning_rate/action_model': 4.446840482723255e-05, 'epoch': 1.09}) +[2026-06-23 20:43:40,359][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.032649438828229904, 'train/grad_norm_pre_clip': 9.682157516479492, 'timing/data': 0.0009121738839894533, 'timing/model': 1.6890450958162546, 'learning_rate/qwen_vl_interface': 4.443839076771823e-06, 'learning_rate/action_model': 4.443839076771823e-05, 'epoch': 1.09}) +[2026-06-23 20:43:43,686][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.21503640711307526, 'train/grad_norm_pre_clip': 16.742700576782227, 'timing/data': 0.0005862119141966105, 'timing/model': 1.6898444560356438, 'learning_rate/qwen_vl_interface': 4.4408380022026945e-06, 'learning_rate/action_model': 4.440838002202694e-05, 'epoch': 1.09}) +[2026-06-23 20:43:47,015][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.10701743513345718, 'train/grad_norm_pre_clip': 36.16504669189453, 'timing/data': 0.0006009798962622881, 'timing/model': 1.6886901431716979, 'learning_rate/qwen_vl_interface': 4.437837260249499e-06, 'learning_rate/action_model': 4.437837260249499e-05, 'epoch': 1.09}) +[2026-06-23 20:43:50,342][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.15187641978263855, 'train/grad_norm_pre_clip': 15.184378623962402, 'timing/data': 0.0006565640214830637, 'timing/model': 1.693412346066907, 'learning_rate/qwen_vl_interface': 4.434836852145728e-06, 'learning_rate/action_model': 4.434836852145728e-05, 'epoch': 1.09}) +[2026-06-23 20:43:53,669][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.15820980072021484, 'train/grad_norm_pre_clip': 61.88965606689453, 'timing/data': 0.0005691300611943007, 'timing/model': 1.688766550971195, 'learning_rate/qwen_vl_interface': 4.431836779124736e-06, 'learning_rate/action_model': 4.431836779124735e-05, 'epoch': 1.09}) +[2026-06-23 20:43:56,997][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.1948830783367157, 'train/grad_norm_pre_clip': 62.96653747558594, 'timing/data': 0.0005696110893040895, 'timing/model': 1.691278369864449, 'learning_rate/qwen_vl_interface': 4.428837042419738e-06, 'learning_rate/action_model': 4.428837042419738e-05, 'epoch': 1.09}) +[2026-06-23 20:44:00,332][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.09000055491924286, 'train/grad_norm_pre_clip': 14.84289264678955, 'timing/data': 0.0006449820939451456, 'timing/model': 1.6940739310812205, 'learning_rate/qwen_vl_interface': 4.425837643263815e-06, 'learning_rate/action_model': 4.425837643263814e-05, 'epoch': 1.09}) +[2026-06-23 20:44:03,756][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.14708714187145233, 'train/grad_norm_pre_clip': 39.176727294921875, 'timing/data': 0.0005897791124880314, 'timing/model': 1.6943466409575194, 'learning_rate/qwen_vl_interface': 4.4228385828899045e-06, 'learning_rate/action_model': 4.4228385828899046e-05, 'epoch': 1.09}) +[2026-06-23 20:44:07,089][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.14441893994808197, 'train/grad_norm_pre_clip': 16.30844497680664, 'timing/data': 0.0005779361817985773, 'timing/model': 1.6938970829360187, 'learning_rate/qwen_vl_interface': 4.419839862530809e-06, 'learning_rate/action_model': 4.419839862530809e-05, 'epoch': 1.09}) +[2026-06-23 20:44:10,429][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.1453571617603302, 'train/grad_norm_pre_clip': 30.315776824951172, 'timing/data': 0.0005857208743691444, 'timing/model': 1.697103878017515, 'learning_rate/qwen_vl_interface': 4.416841483419188e-06, 'learning_rate/action_model': 4.416841483419188e-05, 'epoch': 1.09}) +[2026-06-23 20:44:13,766][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.14980410039424896, 'train/grad_norm_pre_clip': 27.447036743164062, 'timing/data': 0.0005746900569647551, 'timing/model': 1.6963702561333776, 'learning_rate/qwen_vl_interface': 4.413843446787561e-06, 'learning_rate/action_model': 4.413843446787561e-05, 'epoch': 1.09}) +[2026-06-23 20:44:17,093][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.12325660139322281, 'train/grad_norm_pre_clip': 14.885869979858398, 'timing/data': 0.0005886859726160765, 'timing/model': 1.6912726920563728, 'learning_rate/qwen_vl_interface': 4.41084575386831e-06, 'learning_rate/action_model': 4.4108457538683093e-05, 'epoch': 1.09}) +[2026-06-23 20:44:20,427][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.05095391720533371, 'train/grad_norm_pre_clip': 20.885028839111328, 'timing/data': 0.0005832358729094267, 'timing/model': 1.6947952089831233, 'learning_rate/qwen_vl_interface': 4.40784840589367e-06, 'learning_rate/action_model': 4.4078484058936706e-05, 'epoch': 1.09}) +[2026-06-23 20:44:23,766][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.20456933975219727, 'train/grad_norm_pre_clip': 28.231870651245117, 'timing/data': 0.0005920829717069864, 'timing/model': 1.697834260063246, 'learning_rate/qwen_vl_interface': 4.404851404095742e-06, 'learning_rate/action_model': 4.404851404095742e-05, 'epoch': 1.09}) +[2026-06-23 20:44:27,102][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.12584300339221954, 'train/grad_norm_pre_clip': 15.965763092041016, 'timing/data': 0.0006350339390337467, 'timing/model': 1.6923644351772964, 'learning_rate/qwen_vl_interface': 4.4018547497064765e-06, 'learning_rate/action_model': 4.401854749706476e-05, 'epoch': 1.1}) +[2026-06-23 20:44:30,436][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.046934112906455994, 'train/grad_norm_pre_clip': 16.995380401611328, 'timing/data': 0.00063744792714715, 'timing/model': 1.6947440248914063, 'learning_rate/qwen_vl_interface': 4.398858443957686e-06, 'learning_rate/action_model': 4.398858443957686e-05, 'epoch': 1.1}) +[2026-06-23 20:44:33,775][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.19792155921459198, 'train/grad_norm_pre_clip': 18.436065673828125, 'timing/data': 0.0006138340104371309, 'timing/model': 1.6967840059660375, 'learning_rate/qwen_vl_interface': 4.395862488081038e-06, 'learning_rate/action_model': 4.395862488081038e-05, 'epoch': 1.1}) +[2026-06-23 20:44:37,119][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.19516152143478394, 'train/grad_norm_pre_clip': 32.653594970703125, 'timing/data': 0.0006426379550248384, 'timing/model': 1.6952823619358242, 'learning_rate/qwen_vl_interface': 4.392866883308059e-06, 'learning_rate/action_model': 4.392866883308059e-05, 'epoch': 1.1}) +[2026-06-23 20:44:40,454][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.06982233375310898, 'train/grad_norm_pre_clip': 28.049230575561523, 'timing/data': 0.0006512950640171766, 'timing/model': 1.692294331965968, 'learning_rate/qwen_vl_interface': 4.389871630870127e-06, 'learning_rate/action_model': 4.389871630870127e-05, 'epoch': 1.1}) +[2026-06-23 20:44:43,787][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.17777888476848602, 'train/grad_norm_pre_clip': 16.845699310302734, 'timing/data': 0.0005917719099670649, 'timing/model': 1.692726592067629, 'learning_rate/qwen_vl_interface': 4.386876731998478e-06, 'learning_rate/action_model': 4.386876731998478e-05, 'epoch': 1.1}) +[2026-06-23 20:44:47,121][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.16308613121509552, 'train/grad_norm_pre_clip': 87.58893585205078, 'timing/data': 0.0007039830088615417, 'timing/model': 1.6949180390220135, 'learning_rate/qwen_vl_interface': 4.3838821879242006e-06, 'learning_rate/action_model': 4.3838821879242e-05, 'epoch': 1.1}) +[2026-06-23 20:44:50,460][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.06064305827021599, 'train/grad_norm_pre_clip': 16.561420440673828, 'timing/data': 0.0006449031643569469, 'timing/model': 1.6925756430719048, 'learning_rate/qwen_vl_interface': 4.38088799987824e-06, 'learning_rate/action_model': 4.38088799987824e-05, 'epoch': 1.1}) +[2026-06-23 20:44:54,087][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.06969862431287766, 'train/grad_norm_pre_clip': 19.003822326660156, 'timing/data': 0.0005803201347589493, 'timing/model': 1.6871990740764886, 'learning_rate/qwen_vl_interface': 4.3778941690913935e-06, 'learning_rate/action_model': 4.3778941690913934e-05, 'epoch': 1.1}) +[2026-06-23 20:44:57,418][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.13653109967708588, 'train/grad_norm_pre_clip': 59.40755844116211, 'timing/data': 0.000571665121242404, 'timing/model': 1.6904939541127533, 'learning_rate/qwen_vl_interface': 4.374900696794311e-06, 'learning_rate/action_model': 4.374900696794311e-05, 'epoch': 1.1}) +[2026-06-23 20:45:00,753][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.14880429208278656, 'train/grad_norm_pre_clip': 22.162155151367188, 'timing/data': 0.0005825948901474476, 'timing/model': 1.6931532102171332, 'learning_rate/qwen_vl_interface': 4.371907584217499e-06, 'learning_rate/action_model': 4.371907584217498e-05, 'epoch': 1.1}) +[2026-06-23 20:45:04,084][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.18357986211776733, 'train/grad_norm_pre_clip': 25.522315979003906, 'timing/data': 0.000836231978610158, 'timing/model': 1.692773373099044, 'learning_rate/qwen_vl_interface': 4.368914832591309e-06, 'learning_rate/action_model': 4.368914832591309e-05, 'epoch': 1.1}) +[2026-06-23 20:45:07,411][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.24389943480491638, 'train/grad_norm_pre_clip': 67.74323272705078, 'timing/data': 0.0008024889975786209, 'timing/model': 1.6890363169368356, 'learning_rate/qwen_vl_interface': 4.365922443145949e-06, 'learning_rate/action_model': 4.365922443145949e-05, 'epoch': 1.1}) +[2026-06-23 20:45:10,737][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.12785519659519196, 'train/grad_norm_pre_clip': 14.278609275817871, 'timing/data': 0.000584157882258296, 'timing/model': 1.688010956859216, 'learning_rate/qwen_vl_interface': 4.362930417111479e-06, 'learning_rate/action_model': 4.362930417111479e-05, 'epoch': 1.1}) +[2026-06-23 20:45:14,072][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.054667409509420395, 'train/grad_norm_pre_clip': 16.27579689025879, 'timing/data': 0.0005702308844774961, 'timing/model': 1.6939425440505147, 'learning_rate/qwen_vl_interface': 4.359938755717807e-06, 'learning_rate/action_model': 4.359938755717807e-05, 'epoch': 1.1}) +[2026-06-23 20:45:17,404][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.08900968730449677, 'train/grad_norm_pre_clip': 45.63513946533203, 'timing/data': 0.0005813729949295521, 'timing/model': 1.6940663040149957, 'learning_rate/qwen_vl_interface': 4.3569474601946934e-06, 'learning_rate/action_model': 4.356947460194694e-05, 'epoch': 1.1}) +[2026-06-23 20:45:20,740][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.13116015493869781, 'train/grad_norm_pre_clip': 38.85664367675781, 'timing/data': 0.0005811229348182678, 'timing/model': 1.6930947750806808, 'learning_rate/qwen_vl_interface': 4.3539565317717446e-06, 'learning_rate/action_model': 4.353956531771745e-05, 'epoch': 1.1}) +[2026-06-23 20:45:24,077][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.05993780493736267, 'train/grad_norm_pre_clip': 15.829275131225586, 'timing/data': 0.0005740700289607048, 'timing/model': 1.693385107908398, 'learning_rate/qwen_vl_interface': 4.350965971678419e-06, 'learning_rate/action_model': 4.3509659716784185e-05, 'epoch': 1.1}) +[2026-06-23 20:45:27,406][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.04836428537964821, 'train/grad_norm_pre_clip': 38.0305061340332, 'timing/data': 0.0005882761906832457, 'timing/model': 1.6928236701060086, 'learning_rate/qwen_vl_interface': 4.3479757811440225e-06, 'learning_rate/action_model': 4.347975781144022e-05, 'epoch': 1.1}) +[2026-06-23 20:45:30,740][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.06517645716667175, 'train/grad_norm_pre_clip': 12.695109367370605, 'timing/data': 0.0008406098932027817, 'timing/model': 1.6956557009834796, 'learning_rate/qwen_vl_interface': 4.34498596139771e-06, 'learning_rate/action_model': 4.34498596139771e-05, 'epoch': 1.1}) +[2026-06-23 20:45:34,071][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.06981314718723297, 'train/grad_norm_pre_clip': 29.342193603515625, 'timing/data': 0.0005897190421819687, 'timing/model': 1.6895447729621083, 'learning_rate/qwen_vl_interface': 4.341996513668483e-06, 'learning_rate/action_model': 4.3419965136684824e-05, 'epoch': 1.1}) +[2026-06-23 20:45:37,409][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.063428595662117, 'train/grad_norm_pre_clip': 66.95791625976562, 'timing/data': 0.000583976972848177, 'timing/model': 1.69309396087192, 'learning_rate/qwen_vl_interface': 4.33900743918519e-06, 'learning_rate/action_model': 4.33900743918519e-05, 'epoch': 1.1}) +[2026-06-23 20:45:40,740][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.04263922944664955, 'train/grad_norm_pre_clip': 23.50169563293457, 'timing/data': 0.0006033740937709808, 'timing/model': 1.6926222830079496, 'learning_rate/qwen_vl_interface': 4.336018739176529e-06, 'learning_rate/action_model': 4.33601873917653e-05, 'epoch': 1.1}) +[2026-06-23 20:45:44,303][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.10822651535272598, 'train/grad_norm_pre_clip': 13.8502836227417, 'timing/data': 0.0006410551723092794, 'timing/model': 1.6934798120055348, 'learning_rate/qwen_vl_interface': 4.333030414871041e-06, 'learning_rate/action_model': 4.33303041487104e-05, 'epoch': 1.1}) +[2026-06-23 20:45:47,627][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.21479730308055878, 'train/grad_norm_pre_clip': 50.46721649169922, 'timing/data': 0.0005758330225944519, 'timing/model': 1.6868382850661874, 'learning_rate/qwen_vl_interface': 4.330042467497111e-06, 'learning_rate/action_model': 4.330042467497111e-05, 'epoch': 1.1}) +[2026-06-23 20:45:50,952][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.11911563575267792, 'train/grad_norm_pre_clip': 56.536094665527344, 'timing/data': 0.0005840270314365625, 'timing/model': 1.6847789329476655, 'learning_rate/qwen_vl_interface': 4.327054898282973e-06, 'learning_rate/action_model': 4.3270548982829727e-05, 'epoch': 1.1}) +[2026-06-23 20:45:54,279][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.11211442202329636, 'train/grad_norm_pre_clip': 47.4833869934082, 'timing/data': 0.0006420658901333809, 'timing/model': 1.6909156749024987, 'learning_rate/qwen_vl_interface': 4.324067708456703e-06, 'learning_rate/action_model': 4.3240677084567036e-05, 'epoch': 1.11}) +[2026-06-23 20:45:57,615][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.15359428524971008, 'train/grad_norm_pre_clip': 28.785764694213867, 'timing/data': 0.0006570860277861357, 'timing/model': 1.6931848879903555, 'learning_rate/qwen_vl_interface': 4.321080899246225e-06, 'learning_rate/action_model': 4.321080899246225e-05, 'epoch': 1.11}) +[2026-06-23 20:46:00,945][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.13048963248729706, 'train/grad_norm_pre_clip': 13.861627578735352, 'timing/data': 0.0005856209900230169, 'timing/model': 1.6915531589183956, 'learning_rate/qwen_vl_interface': 4.318094471879299e-06, 'learning_rate/action_model': 4.3180944718792985e-05, 'epoch': 1.11}) +[2026-06-23 20:46:04,277][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.2729721963405609, 'train/grad_norm_pre_clip': 37.267372131347656, 'timing/data': 0.0005852500908076763, 'timing/model': 1.6918662840034813, 'learning_rate/qwen_vl_interface': 4.315108427583535e-06, 'learning_rate/action_model': 4.315108427583535e-05, 'epoch': 1.11}) +[2026-06-23 20:46:07,606][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.21581126749515533, 'train/grad_norm_pre_clip': 77.2280502319336, 'timing/data': 0.0005849900189787149, 'timing/model': 1.6911176049616188, 'learning_rate/qwen_vl_interface': 4.312122767586381e-06, 'learning_rate/action_model': 4.3121227675863815e-05, 'epoch': 1.11}) +[2026-06-23 20:46:10,943][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.0930275246500969, 'train/grad_norm_pre_clip': 31.100353240966797, 'timing/data': 0.0005936350207775831, 'timing/model': 1.6970825549215078, 'learning_rate/qwen_vl_interface': 4.3091374931151324e-06, 'learning_rate/action_model': 4.309137493115132e-05, 'epoch': 1.11}) +[2026-06-23 20:46:14,283][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.1059780940413475, 'train/grad_norm_pre_clip': 23.713998794555664, 'timing/data': 0.0005955488886684179, 'timing/model': 1.6937908050604165, 'learning_rate/qwen_vl_interface': 4.306152605396921e-06, 'learning_rate/action_model': 4.306152605396921e-05, 'epoch': 1.11}) +[2026-06-23 20:46:17,614][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.11410538852214813, 'train/grad_norm_pre_clip': 29.624013900756836, 'timing/data': 0.0007896451279520988, 'timing/model': 1.6908176438882947, 'learning_rate/qwen_vl_interface': 4.30316810565872e-06, 'learning_rate/action_model': 4.3031681056587206e-05, 'epoch': 1.11}) +[2026-06-23 20:46:20,930][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.08730705827474594, 'train/grad_norm_pre_clip': 40.859737396240234, 'timing/data': 0.0005766740068793297, 'timing/model': 1.6882262211292982, 'learning_rate/qwen_vl_interface': 4.300183995127347e-06, 'learning_rate/action_model': 4.300183995127346e-05, 'epoch': 1.11}) +[2026-06-23 20:46:24,263][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.16691854596138, 'train/grad_norm_pre_clip': 72.69135284423828, 'timing/data': 0.0006142950151115656, 'timing/model': 1.6932497420348227, 'learning_rate/qwen_vl_interface': 4.297200275029454e-06, 'learning_rate/action_model': 4.297200275029454e-05, 'epoch': 1.11}) +[2026-06-23 20:46:27,595][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.1697635054588318, 'train/grad_norm_pre_clip': 23.00385093688965, 'timing/data': 0.0006026930641382933, 'timing/model': 1.6914999110158533, 'learning_rate/qwen_vl_interface': 4.29421694659154e-06, 'learning_rate/action_model': 4.294216946591539e-05, 'epoch': 1.11}) +[2026-06-23 20:46:31,082][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.17428813874721527, 'train/grad_norm_pre_clip': 33.06705093383789, 'timing/data': 0.000593516044318676, 'timing/model': 1.6942259219940752, 'learning_rate/qwen_vl_interface': 4.2912340110399345e-06, 'learning_rate/action_model': 4.291234011039934e-05, 'epoch': 1.11}) +[2026-06-23 20:46:34,416][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.11900623142719269, 'train/grad_norm_pre_clip': 14.984113693237305, 'timing/data': 0.0005863518454134464, 'timing/model': 1.6899651209823787, 'learning_rate/qwen_vl_interface': 4.2882514696008125e-06, 'learning_rate/action_model': 4.2882514696008125e-05, 'epoch': 1.11}) +[2026-06-23 20:46:37,748][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.06295650452375412, 'train/grad_norm_pre_clip': 34.86698532104492, 'timing/data': 0.0005814118776470423, 'timing/model': 1.694390025921166, 'learning_rate/qwen_vl_interface': 4.2852693235001815e-06, 'learning_rate/action_model': 4.2852693235001815e-05, 'epoch': 1.11}) +[2026-06-23 20:46:41,074][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.1891627013683319, 'train/grad_norm_pre_clip': 16.923669815063477, 'timing/data': 0.000820081913843751, 'timing/model': 1.6921358571853489, 'learning_rate/qwen_vl_interface': 4.2822875739638915e-06, 'learning_rate/action_model': 4.282287573963891e-05, 'epoch': 1.11}) +[2026-06-23 20:46:44,403][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.09140624850988388, 'train/grad_norm_pre_clip': 13.382329940795898, 'timing/data': 0.0005897791124880314, 'timing/model': 1.6905509461648762, 'learning_rate/qwen_vl_interface': 4.279306222217625e-06, 'learning_rate/action_model': 4.279306222217626e-05, 'epoch': 1.11}) +[2026-06-23 20:46:47,740][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.2342025637626648, 'train/grad_norm_pre_clip': 80.93169403076172, 'timing/data': 0.0005795389879494905, 'timing/model': 1.6944397839251906, 'learning_rate/qwen_vl_interface': 4.276325269486906e-06, 'learning_rate/action_model': 4.276325269486906e-05, 'epoch': 1.11}) +[2026-06-23 20:46:51,064][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.14795933663845062, 'train/grad_norm_pre_clip': 19.675851821899414, 'timing/data': 0.0006540489848703146, 'timing/model': 1.689842419931665, 'learning_rate/qwen_vl_interface': 4.273344716997089e-06, 'learning_rate/action_model': 4.273344716997089e-05, 'epoch': 1.11}) +[2026-06-23 20:46:54,397][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.1336999386548996, 'train/grad_norm_pre_clip': 31.17780876159668, 'timing/data': 0.0005868829321116209, 'timing/model': 1.6923047460149974, 'learning_rate/qwen_vl_interface': 4.270364565973368e-06, 'learning_rate/action_model': 4.270364565973367e-05, 'epoch': 1.11}) +[2026-06-23 20:46:57,720][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.07321329414844513, 'train/grad_norm_pre_clip': 43.047027587890625, 'timing/data': 0.0006850478239357471, 'timing/model': 1.6909304009750485, 'learning_rate/qwen_vl_interface': 4.267384817640768e-06, 'learning_rate/action_model': 4.2673848176407685e-05, 'epoch': 1.11}) +[2026-06-23 20:47:01,051][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.18493154644966125, 'train/grad_norm_pre_clip': 20.09722900390625, 'timing/data': 0.0006498410366475582, 'timing/model': 1.6925252371001989, 'learning_rate/qwen_vl_interface': 4.264405473224155e-06, 'learning_rate/action_model': 4.2644054732241546e-05, 'epoch': 1.11}) +[2026-06-23 20:47:04,374][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.1054457426071167, 'train/grad_norm_pre_clip': 17.957687377929688, 'timing/data': 0.0005839068908244371, 'timing/model': 1.6870617661625147, 'learning_rate/qwen_vl_interface': 4.261426533948221e-06, 'learning_rate/action_model': 4.2614265339482213e-05, 'epoch': 1.11}) +[2026-06-23 20:47:07,707][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.1172216534614563, 'train/grad_norm_pre_clip': 14.639606475830078, 'timing/data': 0.0005824251566082239, 'timing/model': 1.695176986977458, 'learning_rate/qwen_vl_interface': 4.258448001037499e-06, 'learning_rate/action_model': 4.2584480010374985e-05, 'epoch': 1.11}) +[2026-06-23 20:47:11,037][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.09938910603523254, 'train/grad_norm_pre_clip': 16.728662490844727, 'timing/data': 0.0006013899110257626, 'timing/model': 1.6946529690176249, 'learning_rate/qwen_vl_interface': 4.255469875716348e-06, 'learning_rate/action_model': 4.255469875716348e-05, 'epoch': 1.11}) +[2026-06-23 20:47:14,372][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.1848386526107788, 'train/grad_norm_pre_clip': 27.32833480834961, 'timing/data': 0.0005816540215164423, 'timing/model': 1.6944316420704126, 'learning_rate/qwen_vl_interface': 4.252492159208966e-06, 'learning_rate/action_model': 4.252492159208966e-05, 'epoch': 1.11}) +[2026-06-23 20:47:17,836][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.24440482258796692, 'train/grad_norm_pre_clip': 56.82750701904297, 'timing/data': 0.0006856080144643784, 'timing/model': 1.8271528540644795, 'learning_rate/qwen_vl_interface': 4.2495148527393766e-06, 'learning_rate/action_model': 4.2495148527393766e-05, 'epoch': 1.11}) +[2026-06-23 20:47:21,174][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.17304863035678864, 'train/grad_norm_pre_clip': 18.322999954223633, 'timing/data': 0.0007216858211904764, 'timing/model': 1.696015584981069, 'learning_rate/qwen_vl_interface': 4.246537957531441e-06, 'learning_rate/action_model': 4.2465379575314404e-05, 'epoch': 1.12}) +[2026-06-23 20:47:24,508][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.1606653481721878, 'train/grad_norm_pre_clip': 46.360755920410156, 'timing/data': 0.0006091950926929712, 'timing/model': 1.6943602610845119, 'learning_rate/qwen_vl_interface': 4.243561474808846e-06, 'learning_rate/action_model': 4.2435614748088456e-05, 'epoch': 1.12}) +[2026-06-23 20:47:27,849][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.3814641237258911, 'train/grad_norm_pre_clip': 22.64785385131836, 'timing/data': 0.0006142740603536367, 'timing/model': 1.6949039311148226, 'learning_rate/qwen_vl_interface': 4.240585405795112e-06, 'learning_rate/action_model': 4.240585405795112e-05, 'epoch': 1.12}) +[2026-06-23 20:47:31,185][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.10966380685567856, 'train/grad_norm_pre_clip': 17.296466827392578, 'timing/data': 0.0005931139457970858, 'timing/model': 1.696232316084206, 'learning_rate/qwen_vl_interface': 4.237609751713589e-06, 'learning_rate/action_model': 4.2376097517135896e-05, 'epoch': 1.12}) +[2026-06-23 20:47:34,528][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.04621599242091179, 'train/grad_norm_pre_clip': 20.932533264160156, 'timing/data': 0.0005771440919488668, 'timing/model': 1.6966942318249494, 'learning_rate/qwen_vl_interface': 4.234634513787455e-06, 'learning_rate/action_model': 4.2346345137874556e-05, 'epoch': 1.12}) +[2026-06-23 20:47:37,866][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.11437470465898514, 'train/grad_norm_pre_clip': 22.99658203125, 'timing/data': 0.0005957998801022768, 'timing/model': 1.698547865031287, 'learning_rate/qwen_vl_interface': 4.2316596932397195e-06, 'learning_rate/action_model': 4.2316596932397195e-05, 'epoch': 1.12}) +[2026-06-23 20:47:41,206][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.1642294079065323, 'train/grad_norm_pre_clip': 54.65012741088867, 'timing/data': 0.0005859711673110723, 'timing/model': 1.69566116691567, 'learning_rate/qwen_vl_interface': 4.22868529129322e-06, 'learning_rate/action_model': 4.22868529129322e-05, 'epoch': 1.12}) +[2026-06-23 20:47:44,552][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.07441513240337372, 'train/grad_norm_pre_clip': 30.94672393798828, 'timing/data': 0.000577947124838829, 'timing/model': 1.6988312141038477, 'learning_rate/qwen_vl_interface': 4.2257113091706165e-06, 'learning_rate/action_model': 4.2257113091706165e-05, 'epoch': 1.12}) +[2026-06-23 20:47:47,894][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.20252399146556854, 'train/grad_norm_pre_clip': 39.348323822021484, 'timing/data': 0.0005877239163964987, 'timing/model': 1.696927173063159, 'learning_rate/qwen_vl_interface': 4.222737748094404e-06, 'learning_rate/action_model': 4.222737748094404e-05, 'epoch': 1.12}) +[2026-06-23 20:47:51,235][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.07637283951044083, 'train/grad_norm_pre_clip': 12.013782501220703, 'timing/data': 0.0006420558784157038, 'timing/model': 1.7008208888582885, 'learning_rate/qwen_vl_interface': 4.219764609286901e-06, 'learning_rate/action_model': 4.219764609286901e-05, 'epoch': 1.12}) +[2026-06-23 20:47:54,583][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.08612324297428131, 'train/grad_norm_pre_clip': 13.86566162109375, 'timing/data': 0.0005857809446752071, 'timing/model': 1.7032504139933735, 'learning_rate/qwen_vl_interface': 4.21679189397025e-06, 'learning_rate/action_model': 4.21679189397025e-05, 'epoch': 1.12}) +[2026-06-23 20:47:57,921][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.18347914516925812, 'train/grad_norm_pre_clip': 55.189002990722656, 'timing/data': 0.0005818440113216639, 'timing/model': 1.697314766002819, 'learning_rate/qwen_vl_interface': 4.213819603366424e-06, 'learning_rate/action_model': 4.213819603366424e-05, 'epoch': 1.12}) +[2026-06-23 20:48:01,271][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.04935194179415703, 'train/grad_norm_pre_clip': 27.151893615722656, 'timing/data': 0.0007686850149184465, 'timing/model': 1.7034022458828986, 'learning_rate/qwen_vl_interface': 4.210847738697219e-06, 'learning_rate/action_model': 4.210847738697218e-05, 'epoch': 1.12}) +[2026-06-23 20:48:04,618][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.0873502716422081, 'train/grad_norm_pre_clip': 43.695770263671875, 'timing/data': 0.0005872331093996763, 'timing/model': 1.6991780640091747, 'learning_rate/qwen_vl_interface': 4.207876301184256e-06, 'learning_rate/action_model': 4.207876301184256e-05, 'epoch': 1.12}) +[2026-06-23 20:48:08,241][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.08895684033632278, 'train/grad_norm_pre_clip': 53.98052215576172, 'timing/data': 0.0005964711308479309, 'timing/model': 1.9754540999419987, 'learning_rate/qwen_vl_interface': 4.2049052920489825e-06, 'learning_rate/action_model': 4.2049052920489825e-05, 'epoch': 1.12}) +[2026-06-23 20:48:11,583][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.23080050945281982, 'train/grad_norm_pre_clip': 36.45699691772461, 'timing/data': 0.0005764130037277937, 'timing/model': 1.697461755014956, 'learning_rate/qwen_vl_interface': 4.201934712512666e-06, 'learning_rate/action_model': 4.2019347125126656e-05, 'epoch': 1.12}) +[2026-06-23 20:48:14,923][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.08219509571790695, 'train/grad_norm_pre_clip': 14.970307350158691, 'timing/data': 0.0006006490439176559, 'timing/model': 1.695630189962685, 'learning_rate/qwen_vl_interface': 4.1989645637964e-06, 'learning_rate/action_model': 4.1989645637964e-05, 'epoch': 1.12}) +[2026-06-23 20:48:18,267][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.05815335363149643, 'train/grad_norm_pre_clip': 14.261510848999023, 'timing/data': 0.0005881460383534431, 'timing/model': 1.6948571850080043, 'learning_rate/qwen_vl_interface': 4.195994847121101e-06, 'learning_rate/action_model': 4.195994847121101e-05, 'epoch': 1.12}) +[2026-06-23 20:48:21,607][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2900 +[2026-06-23 20:49:54,980][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2900 +[2026-06-23 20:49:54,982][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 2900 +[2026-06-23 20:51:05,767][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 2900 +[2026-06-23 20:51:05,768][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.15749172866344452, 'train/grad_norm_pre_clip': 22.846309661865234, 'eval/action_loss/seconds': 93.37004317203537, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.0693730041384697, 'eval/latency_2/loss': 0.0693730041384697, 'eval/flappy/loss': 0.0693730041384697, 'eval/flappy/latency_2/loss': 0.0693730041384697, 'eval/flap_precision': 0.9450171821305842, 'eval/flap_recall': 0.9278002699055331, 'eval/flap_f1': 0.9363295880149812, 'eval/flap_rate_model': 0.20208333333333334, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9363295880149812, 'eval/latency_2/flap_precision': 0.9450171821305842, 'eval/latency_2/flap_recall': 0.9278002699055331, 'eval/latency_2/flap_f1': 0.9363295880149812, 'eval/latency_2/flap_rate_model': 0.20208333333333334, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9363295880149812, 'eval/action_classification/data_wait_seconds': 0.871126115322113, 'eval/action_classification/predict_seconds': 69.77032470703125, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006077119614928961, 'timing/model': 1.6962086139246821, 'learning_rate/qwen_vl_interface': 4.193025563707507e-06, 'learning_rate/action_model': 4.1930255637075066e-05, 'epoch': 1.12}) +[2026-06-23 20:51:09,071][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.1865665316581726, 'train/grad_norm_pre_clip': 55.82101058959961, 'timing/data': 0.0005862119141966105, 'timing/model': 1.676350103924051, 'learning_rate/qwen_vl_interface': 4.19005671477618e-06, 'learning_rate/action_model': 4.1900567147761794e-05, 'epoch': 1.12}) +[2026-06-23 20:51:12,384][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.11220438033342361, 'train/grad_norm_pre_clip': 15.624008178710938, 'timing/data': 0.0005857909563928843, 'timing/model': 1.6833265649620444, 'learning_rate/qwen_vl_interface': 4.1870883015474975e-06, 'learning_rate/action_model': 4.187088301547497e-05, 'epoch': 1.12}) +[2026-06-23 20:51:15,693][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.12380317598581314, 'train/grad_norm_pre_clip': 27.973604202270508, 'timing/data': 0.0008258828893303871, 'timing/model': 1.6764477151446044, 'learning_rate/qwen_vl_interface': 4.1841203252416655e-06, 'learning_rate/action_model': 4.1841203252416655e-05, 'epoch': 1.12}) +[2026-06-23 20:51:19,111][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.04511084780097008, 'train/grad_norm_pre_clip': 25.201435089111328, 'timing/data': 0.0006723231635987759, 'timing/model': 1.7857844170648605, 'learning_rate/qwen_vl_interface': 4.181152787078705e-06, 'learning_rate/action_model': 4.1811527870787056e-05, 'epoch': 1.12}) +[2026-06-23 20:51:22,436][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.1840091347694397, 'train/grad_norm_pre_clip': 43.564884185791016, 'timing/data': 0.0005857609212398529, 'timing/model': 1.6965486330445856, 'learning_rate/qwen_vl_interface': 4.178185688278462e-06, 'learning_rate/action_model': 4.1781856882784616e-05, 'epoch': 1.12}) +[2026-06-23 20:51:25,753][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.15067079663276672, 'train/grad_norm_pre_clip': 32.468631744384766, 'timing/data': 0.0005881860852241516, 'timing/model': 1.6839561730157584, 'learning_rate/qwen_vl_interface': 4.175219030060595e-06, 'learning_rate/action_model': 4.175219030060595e-05, 'epoch': 1.12}) +[2026-06-23 20:51:29,080][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.20253732800483704, 'train/grad_norm_pre_clip': 22.09330940246582, 'timing/data': 0.0005859020166099072, 'timing/model': 1.6873942040838301, 'learning_rate/qwen_vl_interface': 4.1722528136445875e-06, 'learning_rate/action_model': 4.172252813644588e-05, 'epoch': 1.13}) +[2026-06-23 20:51:32,444][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.058504924178123474, 'train/grad_norm_pre_clip': 12.515329360961914, 'timing/data': 0.0006721429526805878, 'timing/model': 1.7234549110289663, 'learning_rate/qwen_vl_interface': 4.169287040249738e-06, 'learning_rate/action_model': 4.169287040249738e-05, 'epoch': 1.13}) +[2026-06-23 20:51:35,760][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.019596198573708534, 'train/grad_norm_pre_clip': 12.664902687072754, 'timing/data': 0.0005866521969437599, 'timing/model': 1.6821302380412817, 'learning_rate/qwen_vl_interface': 4.166321711095164e-06, 'learning_rate/action_model': 4.166321711095164e-05, 'epoch': 1.13}) +[2026-06-23 20:51:39,080][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.192593514919281, 'train/grad_norm_pre_clip': 13.634549140930176, 'timing/data': 0.0005921530537307262, 'timing/model': 1.6857603690586984, 'learning_rate/qwen_vl_interface': 4.1633568273998e-06, 'learning_rate/action_model': 4.1633568273998e-05, 'epoch': 1.13}) +[2026-06-23 20:51:42,408][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.06463900208473206, 'train/grad_norm_pre_clip': 17.846046447753906, 'timing/data': 0.0005879150703549385, 'timing/model': 1.690173263195902, 'learning_rate/qwen_vl_interface': 4.1603923903823995e-06, 'learning_rate/action_model': 4.160392390382399e-05, 'epoch': 1.13}) +[2026-06-23 20:51:45,734][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.06859566271305084, 'train/grad_norm_pre_clip': 79.68055725097656, 'timing/data': 0.0006420470308512449, 'timing/model': 1.6904783151112497, 'learning_rate/qwen_vl_interface': 4.1574284012615276e-06, 'learning_rate/action_model': 4.157428401261528e-05, 'epoch': 1.13}) +[2026-06-23 20:51:49,062][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.07419320940971375, 'train/grad_norm_pre_clip': 45.88997268676758, 'timing/data': 0.0006701000966131687, 'timing/model': 1.69152223598212, 'learning_rate/qwen_vl_interface': 4.154464861255571e-06, 'learning_rate/action_model': 4.154464861255571e-05, 'epoch': 1.13}) +[2026-06-23 20:51:52,391][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.1381727010011673, 'train/grad_norm_pre_clip': 14.804762840270996, 'timing/data': 0.0005802200175821781, 'timing/model': 1.693693506065756, 'learning_rate/qwen_vl_interface': 4.151501771582728e-06, 'learning_rate/action_model': 4.151501771582728e-05, 'epoch': 1.13}) +[2026-06-23 20:51:55,712][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.049755971878767014, 'train/grad_norm_pre_clip': 35.464839935302734, 'timing/data': 0.000643719919025898, 'timing/model': 1.68434944213368, 'learning_rate/qwen_vl_interface': 4.148539133461012e-06, 'learning_rate/action_model': 4.1485391334610116e-05, 'epoch': 1.13}) +[2026-06-23 20:51:59,034][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.11465086787939072, 'train/grad_norm_pre_clip': 18.151411056518555, 'timing/data': 0.0005951791536062956, 'timing/model': 1.6956828420516104, 'learning_rate/qwen_vl_interface': 4.145576948108254e-06, 'learning_rate/action_model': 4.145576948108253e-05, 'epoch': 1.13}) +[2026-06-23 20:52:02,364][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.17948690056800842, 'train/grad_norm_pre_clip': 16.25210189819336, 'timing/data': 0.0007665108423680067, 'timing/model': 1.6912985709495842, 'learning_rate/qwen_vl_interface': 4.142615216742093e-06, 'learning_rate/action_model': 4.142615216742093e-05, 'epoch': 1.13}) +[2026-06-23 20:52:05,691][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.10317833721637726, 'train/grad_norm_pre_clip': 69.47456359863281, 'timing/data': 0.0005910799372941256, 'timing/model': 1.688999657984823, 'learning_rate/qwen_vl_interface': 4.139653940579988e-06, 'learning_rate/action_model': 4.1396539405799874e-05, 'epoch': 1.13}) +[2026-06-23 20:52:09,110][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.20731769502162933, 'train/grad_norm_pre_clip': 68.46765899658203, 'timing/data': 0.0008227957878261805, 'timing/model': 1.6932317020837218, 'learning_rate/qwen_vl_interface': 4.136693120839206e-06, 'learning_rate/action_model': 4.136693120839205e-05, 'epoch': 1.13}) +[2026-06-23 20:52:12,441][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.1371772438287735, 'train/grad_norm_pre_clip': 42.765045166015625, 'timing/data': 0.0005940569099038839, 'timing/model': 1.6899788070004433, 'learning_rate/qwen_vl_interface': 4.133732758736828e-06, 'learning_rate/action_model': 4.133732758736828e-05, 'epoch': 1.13}) +[2026-06-23 20:52:15,769][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.07853516936302185, 'train/grad_norm_pre_clip': 24.475496292114258, 'timing/data': 0.0006715022027492523, 'timing/model': 1.6919123998377472, 'learning_rate/qwen_vl_interface': 4.13077285548975e-06, 'learning_rate/action_model': 4.13077285548975e-05, 'epoch': 1.13}) +[2026-06-23 20:52:19,094][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.10476469248533249, 'train/grad_norm_pre_clip': 15.965690612792969, 'timing/data': 0.0005645109340548515, 'timing/model': 1.6847333260811865, 'learning_rate/qwen_vl_interface': 4.127813412314672e-06, 'learning_rate/action_model': 4.1278134123146725e-05, 'epoch': 1.13}) +[2026-06-23 20:52:22,410][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.19458799064159393, 'train/grad_norm_pre_clip': 44.60520935058594, 'timing/data': 0.0006204061210155487, 'timing/model': 1.6824999370146543, 'learning_rate/qwen_vl_interface': 4.124854430428113e-06, 'learning_rate/action_model': 4.1248544304281126e-05, 'epoch': 1.13}) +[2026-06-23 20:52:25,733][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.129184752702713, 'train/grad_norm_pre_clip': 19.174951553344727, 'timing/data': 0.00085305399261415, 'timing/model': 1.6861857229378074, 'learning_rate/qwen_vl_interface': 4.121895911046397e-06, 'learning_rate/action_model': 4.121895911046396e-05, 'epoch': 1.13}) +[2026-06-23 20:52:29,054][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.2336304783821106, 'train/grad_norm_pre_clip': 54.214298248291016, 'timing/data': 0.0006396421231329441, 'timing/model': 1.685976719018072, 'learning_rate/qwen_vl_interface': 4.118937855385659e-06, 'learning_rate/action_model': 4.1189378553856584e-05, 'epoch': 1.13}) +[2026-06-23 20:52:32,407][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.06602110713720322, 'train/grad_norm_pre_clip': 12.383832931518555, 'timing/data': 0.0006617030594497919, 'timing/model': 1.6919541128445417, 'learning_rate/qwen_vl_interface': 4.115980264661844e-06, 'learning_rate/action_model': 4.115980264661844e-05, 'epoch': 1.13}) +[2026-06-23 20:52:35,742][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.1482296586036682, 'train/grad_norm_pre_clip': 21.794387817382812, 'timing/data': 0.0006857090629637241, 'timing/model': 1.6901614931412041, 'learning_rate/qwen_vl_interface': 4.113023140090709e-06, 'learning_rate/action_model': 4.113023140090709e-05, 'epoch': 1.13}) +[2026-06-23 20:52:39,078][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.07658933848142624, 'train/grad_norm_pre_clip': 13.174421310424805, 'timing/data': 0.0005986250471323729, 'timing/model': 1.6952885591890663, 'learning_rate/qwen_vl_interface': 4.110066482887812e-06, 'learning_rate/action_model': 4.110066482887812e-05, 'epoch': 1.13}) +[2026-06-23 20:52:42,420][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.03712358698248863, 'train/grad_norm_pre_clip': 34.38834762573242, 'timing/data': 0.0005839979276061058, 'timing/model': 1.696986508090049, 'learning_rate/qwen_vl_interface': 4.107110294268525e-06, 'learning_rate/action_model': 4.107110294268525e-05, 'epoch': 1.13}) +[2026-06-23 20:52:45,752][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.05741678923368454, 'train/grad_norm_pre_clip': 15.556848526000977, 'timing/data': 0.0005839471705257893, 'timing/model': 1.693244795082137, 'learning_rate/qwen_vl_interface': 4.104154575448026e-06, 'learning_rate/action_model': 4.104154575448025e-05, 'epoch': 1.13}) +[2026-06-23 20:52:49,081][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.07495677471160889, 'train/grad_norm_pre_clip': 75.57760620117188, 'timing/data': 0.0005958601832389832, 'timing/model': 1.6905499189160764, 'learning_rate/qwen_vl_interface': 4.101199327641297e-06, 'learning_rate/action_model': 4.101199327641297e-05, 'epoch': 1.13}) +[2026-06-23 20:52:52,418][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.2944864332675934, 'train/grad_norm_pre_clip': 20.74886703491211, 'timing/data': 0.0006760600954294205, 'timing/model': 1.6924116569571197, 'learning_rate/qwen_vl_interface': 4.098244552063133e-06, 'learning_rate/action_model': 4.098244552063132e-05, 'epoch': 1.13}) +[2026-06-23 20:52:55,753][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.1848187893629074, 'train/grad_norm_pre_clip': 70.31817626953125, 'timing/data': 0.0006549910176545382, 'timing/model': 1.6976185929961503, 'learning_rate/qwen_vl_interface': 4.0952902499281254e-06, 'learning_rate/action_model': 4.095290249928125e-05, 'epoch': 1.14}) +[2026-06-23 20:52:59,233][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.2130996584892273, 'train/grad_norm_pre_clip': 58.92372131347656, 'timing/data': 0.0006202859804034233, 'timing/model': 1.6890609469264746, 'learning_rate/qwen_vl_interface': 4.092336422450681e-06, 'learning_rate/action_model': 4.0923364224506804e-05, 'epoch': 1.14}) +[2026-06-23 20:53:02,564][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.08308971673250198, 'train/grad_norm_pre_clip': 27.495956420898438, 'timing/data': 0.0005963409785181284, 'timing/model': 1.6864850940182805, 'learning_rate/qwen_vl_interface': 4.089383070845003e-06, 'learning_rate/action_model': 4.089383070845003e-05, 'epoch': 1.14}) +[2026-06-23 20:53:05,899][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.18066634237766266, 'train/grad_norm_pre_clip': 54.12306594848633, 'timing/data': 0.0005952590145170689, 'timing/model': 1.6921803781297058, 'learning_rate/qwen_vl_interface': 4.086430196325106e-06, 'learning_rate/action_model': 4.086430196325106e-05, 'epoch': 1.14}) +[2026-06-23 20:53:09,237][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.341122031211853, 'train/grad_norm_pre_clip': 43.533447265625, 'timing/data': 0.0005971330683678389, 'timing/model': 1.6955974490847439, 'learning_rate/qwen_vl_interface': 4.083477800104802e-06, 'learning_rate/action_model': 4.083477800104802e-05, 'epoch': 1.14}) +[2026-06-23 20:53:12,586][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.11294271796941757, 'train/grad_norm_pre_clip': 20.574853897094727, 'timing/data': 0.0006453630048781633, 'timing/model': 1.7006454260554165, 'learning_rate/qwen_vl_interface': 4.080525883397711e-06, 'learning_rate/action_model': 4.080525883397711e-05, 'epoch': 1.14}) +[2026-06-23 20:53:15,925][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.14018464088439941, 'train/grad_norm_pre_clip': 23.958417892456055, 'timing/data': 0.0005634389817714691, 'timing/model': 1.6940869679674506, 'learning_rate/qwen_vl_interface': 4.077574447417256e-06, 'learning_rate/action_model': 4.077574447417256e-05, 'epoch': 1.14}) +[2026-06-23 20:53:19,255][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.21099898219108582, 'train/grad_norm_pre_clip': 31.049318313598633, 'timing/data': 0.0006142340134829283, 'timing/model': 1.6877916851080954, 'learning_rate/qwen_vl_interface': 4.074623493376657e-06, 'learning_rate/action_model': 4.074623493376657e-05, 'epoch': 1.14}) +[2026-06-23 20:53:22,582][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.060799889266490936, 'train/grad_norm_pre_clip': 34.11895751953125, 'timing/data': 0.000696518924087286, 'timing/model': 1.6885436880402267, 'learning_rate/qwen_vl_interface': 4.071673022488944e-06, 'learning_rate/action_model': 4.071673022488944e-05, 'epoch': 1.14}) +[2026-06-23 20:53:25,919][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.0765623152256012, 'train/grad_norm_pre_clip': 36.37397003173828, 'timing/data': 0.0005786679685115814, 'timing/model': 1.6917695989832282, 'learning_rate/qwen_vl_interface': 4.06872303596694e-06, 'learning_rate/action_model': 4.06872303596694e-05, 'epoch': 1.14}) +[2026-06-23 20:53:29,246][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.21475473046302795, 'train/grad_norm_pre_clip': 37.74678039550781, 'timing/data': 0.0005786078982055187, 'timing/model': 1.6948525230400264, 'learning_rate/qwen_vl_interface': 4.065773535023277e-06, 'learning_rate/action_model': 4.0657735350232766e-05, 'epoch': 1.14}) +[2026-06-23 20:53:32,572][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.08550158143043518, 'train/grad_norm_pre_clip': 50.508487701416016, 'timing/data': 0.0005919530522078276, 'timing/model': 1.6910901609808207, 'learning_rate/qwen_vl_interface': 4.062824520870382e-06, 'learning_rate/action_model': 4.062824520870381e-05, 'epoch': 1.14}) +[2026-06-23 20:53:35,908][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.1438237726688385, 'train/grad_norm_pre_clip': 34.559024810791016, 'timing/data': 0.0005854701157659292, 'timing/model': 1.6923798739444464, 'learning_rate/qwen_vl_interface': 4.059875994720481e-06, 'learning_rate/action_model': 4.059875994720481e-05, 'epoch': 1.14}) +[2026-06-23 20:53:39,244][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.039294686168432236, 'train/grad_norm_pre_clip': 16.78643035888672, 'timing/data': 0.0005889269523322582, 'timing/model': 1.6959351550322026, 'learning_rate/qwen_vl_interface': 4.056927957785605e-06, 'learning_rate/action_model': 4.056927957785605e-05, 'epoch': 1.14}) +[2026-06-23 20:53:42,577][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.053556375205516815, 'train/grad_norm_pre_clip': 14.86051082611084, 'timing/data': 0.0005758528131991625, 'timing/model': 1.693163393996656, 'learning_rate/qwen_vl_interface': 4.0539804112775785e-06, 'learning_rate/action_model': 4.053980411277578e-05, 'epoch': 1.14}) +[2026-06-23 20:53:46,043][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.12041795998811722, 'train/grad_norm_pre_clip': 37.237632751464844, 'timing/data': 0.000599527033045888, 'timing/model': 1.8230451790150255, 'learning_rate/qwen_vl_interface': 4.051033356408027e-06, 'learning_rate/action_model': 4.0510333564080264e-05, 'epoch': 1.14}) +[2026-06-23 20:53:49,380][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.08445093035697937, 'train/grad_norm_pre_clip': 43.54741668701172, 'timing/data': 0.0006235120818018913, 'timing/model': 1.6914285318925977, 'learning_rate/qwen_vl_interface': 4.048086794388375e-06, 'learning_rate/action_model': 4.048086794388375e-05, 'epoch': 1.14}) +[2026-06-23 20:53:52,708][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.11735322326421738, 'train/grad_norm_pre_clip': 43.28000259399414, 'timing/data': 0.0005923029966652393, 'timing/model': 1.6872652359306812, 'learning_rate/qwen_vl_interface': 4.045140726429841e-06, 'learning_rate/action_model': 4.0451407264298406e-05, 'epoch': 1.14}) +[2026-06-23 20:53:56,047][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.12427866458892822, 'train/grad_norm_pre_clip': 30.972442626953125, 'timing/data': 0.0005985349416732788, 'timing/model': 1.6949696759693325, 'learning_rate/qwen_vl_interface': 4.042195153743443e-06, 'learning_rate/action_model': 4.042195153743443e-05, 'epoch': 1.14}) +[2026-06-23 20:53:59,382][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.13379044830799103, 'train/grad_norm_pre_clip': 51.240108489990234, 'timing/data': 0.0006618138868361712, 'timing/model': 1.6902427999302745, 'learning_rate/qwen_vl_interface': 4.039250077539993e-06, 'learning_rate/action_model': 4.039250077539993e-05, 'epoch': 1.14}) +[2026-06-23 20:54:02,720][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.166493758559227, 'train/grad_norm_pre_clip': 25.939470291137695, 'timing/data': 0.000654750969260931, 'timing/model': 1.6955989869311452, 'learning_rate/qwen_vl_interface': 4.036305499030104e-06, 'learning_rate/action_model': 4.0363054990301036e-05, 'epoch': 1.14}) +[2026-06-23 20:54:06,064][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.148185133934021, 'train/grad_norm_pre_clip': 21.870988845825195, 'timing/data': 0.0005904790014028549, 'timing/model': 1.6940865830983967, 'learning_rate/qwen_vl_interface': 4.033361419424178e-06, 'learning_rate/action_model': 4.033361419424178e-05, 'epoch': 1.14}) +[2026-06-23 20:54:09,395][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.07969296723604202, 'train/grad_norm_pre_clip': 14.279136657714844, 'timing/data': 0.0005880251992493868, 'timing/model': 1.6916796390432864, 'learning_rate/qwen_vl_interface': 4.030417839932418e-06, 'learning_rate/action_model': 4.030417839932417e-05, 'epoch': 1.14}) +[2026-06-23 20:54:12,730][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.0506151057779789, 'train/grad_norm_pre_clip': 25.338857650756836, 'timing/data': 0.0005904599092900753, 'timing/model': 1.6922093031462282, 'learning_rate/qwen_vl_interface': 4.027474761764815e-06, 'learning_rate/action_model': 4.027474761764814e-05, 'epoch': 1.14}) +[2026-06-23 20:54:16,061][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.09163982421159744, 'train/grad_norm_pre_clip': 25.090879440307617, 'timing/data': 0.0005878950469195843, 'timing/model': 1.6908765281550586, 'learning_rate/qwen_vl_interface': 4.024532186131158e-06, 'learning_rate/action_model': 4.0245321861311575e-05, 'epoch': 1.14}) +[2026-06-23 20:54:19,393][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.09293672442436218, 'train/grad_norm_pre_clip': 19.012292861938477, 'timing/data': 0.0005953491199761629, 'timing/model': 1.6928139079827815, 'learning_rate/qwen_vl_interface': 4.021590114241029e-06, 'learning_rate/action_model': 4.021590114241029e-05, 'epoch': 1.14}) +[2026-06-23 20:54:22,718][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.14652277529239655, 'train/grad_norm_pre_clip': 15.529274940490723, 'timing/data': 0.0006062299944460392, 'timing/model': 1.6910896259360015, 'learning_rate/qwen_vl_interface': 4.018648547303804e-06, 'learning_rate/action_model': 4.0186485473038036e-05, 'epoch': 1.15}) +[2026-06-23 20:54:26,049][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.11477142572402954, 'train/grad_norm_pre_clip': 19.270355224609375, 'timing/data': 0.0005893779452890158, 'timing/model': 1.6904127630405128, 'learning_rate/qwen_vl_interface': 4.015707486528649e-06, 'learning_rate/action_model': 4.015707486528649e-05, 'epoch': 1.15}) +[2026-06-23 20:54:29,383][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.14345644414424896, 'train/grad_norm_pre_clip': 23.55867576599121, 'timing/data': 0.0005952389910817146, 'timing/model': 1.6917862130794674, 'learning_rate/qwen_vl_interface': 4.012766933124523e-06, 'learning_rate/action_model': 4.012766933124523e-05, 'epoch': 1.15}) +[2026-06-23 20:54:32,718][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.21804285049438477, 'train/grad_norm_pre_clip': 18.68311309814453, 'timing/data': 0.0005964210722595453, 'timing/model': 1.6967913808766752, 'learning_rate/qwen_vl_interface': 4.009826888300176e-06, 'learning_rate/action_model': 4.0098268883001755e-05, 'epoch': 1.15}) +[2026-06-23 20:54:36,144][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.18470536172389984, 'train/grad_norm_pre_clip': 48.36778259277344, 'timing/data': 0.000599185936152935, 'timing/model': 1.7879974318202585, 'learning_rate/qwen_vl_interface': 4.00688735326415e-06, 'learning_rate/action_model': 4.0068873532641495e-05, 'epoch': 1.15}) +[2026-06-23 20:54:39,469][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.21492207050323486, 'train/grad_norm_pre_clip': 35.35584259033203, 'timing/data': 0.0005944471340626478, 'timing/model': 1.6841470629442483, 'learning_rate/qwen_vl_interface': 4.0039483292247765e-06, 'learning_rate/action_model': 4.003948329224776e-05, 'epoch': 1.15}) +[2026-06-23 20:54:42,800][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.13085226714611053, 'train/grad_norm_pre_clip': 19.621532440185547, 'timing/data': 0.0005988259799778461, 'timing/model': 1.6904216960538179, 'learning_rate/qwen_vl_interface': 4.001009817390178e-06, 'learning_rate/action_model': 4.001009817390178e-05, 'epoch': 1.15}) +[2026-06-23 20:54:46,130][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.12092215567827225, 'train/grad_norm_pre_clip': 30.68148422241211, 'timing/data': 0.000594147015362978, 'timing/model': 1.687747318064794, 'learning_rate/qwen_vl_interface': 3.998071818968264e-06, 'learning_rate/action_model': 3.998071818968264e-05, 'epoch': 1.15}) +[2026-06-23 20:54:49,463][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.1285657286643982, 'train/grad_norm_pre_clip': 17.217466354370117, 'timing/data': 0.0008195000700652599, 'timing/model': 1.692305936012417, 'learning_rate/qwen_vl_interface': 3.995134335166736e-06, 'learning_rate/action_model': 3.995134335166736e-05, 'epoch': 1.15}) +[2026-06-23 20:54:52,782][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.17418617010116577, 'train/grad_norm_pre_clip': 37.047943115234375, 'timing/data': 0.0006079219747334719, 'timing/model': 1.684010692872107, 'learning_rate/qwen_vl_interface': 3.992197367193083e-06, 'learning_rate/action_model': 3.992197367193082e-05, 'epoch': 1.15}) +[2026-06-23 20:54:56,119][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.060210052877664566, 'train/grad_norm_pre_clip': 19.261411666870117, 'timing/data': 0.0006033640820533037, 'timing/model': 1.697135740891099, 'learning_rate/qwen_vl_interface': 3.98926091625458e-06, 'learning_rate/action_model': 3.9892609162545795e-05, 'epoch': 1.15}) +[2026-06-23 20:54:59,455][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.16265682876110077, 'train/grad_norm_pre_clip': 63.49681854248047, 'timing/data': 0.0008166150655597448, 'timing/model': 1.6921568629331887, 'learning_rate/qwen_vl_interface': 3.986324983558294e-06, 'learning_rate/action_model': 3.9863249835582934e-05, 'epoch': 1.15}) +[2026-06-23 20:55:02,796][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.08598851412534714, 'train/grad_norm_pre_clip': 26.696640014648438, 'timing/data': 0.0005912811029702425, 'timing/model': 1.7016531999688596, 'learning_rate/qwen_vl_interface': 3.983389570311073e-06, 'learning_rate/action_model': 3.9833895703110724e-05, 'epoch': 1.15}) +[2026-06-23 20:55:06,129][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.1710764616727829, 'train/grad_norm_pre_clip': 30.089916229248047, 'timing/data': 0.0006034548860043287, 'timing/model': 1.6973004590254277, 'learning_rate/qwen_vl_interface': 3.9804546777195546e-06, 'learning_rate/action_model': 3.980454677719554e-05, 'epoch': 1.15}) +[2026-06-23 20:55:09,465][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.08076386898756027, 'train/grad_norm_pre_clip': 13.56770133972168, 'timing/data': 0.0005957200191915035, 'timing/model': 1.6951509178616107, 'learning_rate/qwen_vl_interface': 3.977520306990164e-06, 'learning_rate/action_model': 3.977520306990163e-05, 'epoch': 1.15}) +[2026-06-23 20:55:12,800][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.09324241429567337, 'train/grad_norm_pre_clip': 40.541996002197266, 'timing/data': 0.0006235020700842142, 'timing/model': 1.6949687239248306, 'learning_rate/qwen_vl_interface': 3.974586459329109e-06, 'learning_rate/action_model': 3.9745864593291085e-05, 'epoch': 1.15}) +[2026-06-23 20:55:16,144][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.12140011042356491, 'train/grad_norm_pre_clip': 29.65293312072754, 'timing/data': 0.0007391090039163828, 'timing/model': 1.7000966419000179, 'learning_rate/qwen_vl_interface': 3.9716531359423855e-06, 'learning_rate/action_model': 3.971653135942385e-05, 'epoch': 1.15}) +[2026-06-23 20:55:19,475][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.12015186250209808, 'train/grad_norm_pre_clip': 44.09768295288086, 'timing/data': 0.0005984250456094742, 'timing/model': 1.6885119581129402, 'learning_rate/qwen_vl_interface': 3.968720338035769e-06, 'learning_rate/action_model': 3.968720338035769e-05, 'epoch': 1.15}) +[2026-06-23 20:55:22,807][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.13248054683208466, 'train/grad_norm_pre_clip': 18.131654739379883, 'timing/data': 0.0006000478751957417, 'timing/model': 1.695941851940006, 'learning_rate/qwen_vl_interface': 3.965788066814824e-06, 'learning_rate/action_model': 3.965788066814823e-05, 'epoch': 1.15}) +[2026-06-23 20:55:26,242][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.1433415412902832, 'train/grad_norm_pre_clip': 27.2324275970459, 'timing/data': 0.0005848000291734934, 'timing/model': 1.6876469100825489, 'learning_rate/qwen_vl_interface': 3.962856323484895e-06, 'learning_rate/action_model': 3.9628563234848945e-05, 'epoch': 1.15}) +[2026-06-23 20:55:29,577][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.15875115990638733, 'train/grad_norm_pre_clip': 27.029645919799805, 'timing/data': 0.0005852000322192907, 'timing/model': 1.6888580860104412, 'learning_rate/qwen_vl_interface': 3.959925109251112e-06, 'learning_rate/action_model': 3.959925109251112e-05, 'epoch': 1.15}) +[2026-06-23 20:55:32,917][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.14605416357517242, 'train/grad_norm_pre_clip': 15.302277565002441, 'timing/data': 0.0006251749582588673, 'timing/model': 1.6983903208747506, 'learning_rate/qwen_vl_interface': 3.956994425318387e-06, 'learning_rate/action_model': 3.956994425318387e-05, 'epoch': 1.15}) +[2026-06-23 20:55:36,243][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.1687612235546112, 'train/grad_norm_pre_clip': 15.703604698181152, 'timing/data': 0.0006898569408804178, 'timing/model': 1.6922287098132074, 'learning_rate/qwen_vl_interface': 3.9540642728914115e-06, 'learning_rate/action_model': 3.9540642728914115e-05, 'epoch': 1.15}) +[2026-06-23 20:55:39,578][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.09720562398433685, 'train/grad_norm_pre_clip': 34.440284729003906, 'timing/data': 0.0005984350573271513, 'timing/model': 1.696880271891132, 'learning_rate/qwen_vl_interface': 3.951134653174664e-06, 'learning_rate/action_model': 3.951134653174663e-05, 'epoch': 1.15}) +[2026-06-23 20:55:42,914][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.17463290691375732, 'train/grad_norm_pre_clip': 26.54337501525879, 'timing/data': 0.0005895979702472687, 'timing/model': 1.6930300039239228, 'learning_rate/qwen_vl_interface': 3.948205567372397e-06, 'learning_rate/action_model': 3.948205567372397e-05, 'epoch': 1.15}) +[2026-06-23 20:55:46,245][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.08877259492874146, 'train/grad_norm_pre_clip': 21.734975814819336, 'timing/data': 0.0005908207967877388, 'timing/model': 1.6913645029999316, 'learning_rate/qwen_vl_interface': 3.945277016688647e-06, 'learning_rate/action_model': 3.945277016688647e-05, 'epoch': 1.15}) +[2026-06-23 20:55:49,582][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.1913008838891983, 'train/grad_norm_pre_clip': 33.90518569946289, 'timing/data': 0.000609945971518755, 'timing/model': 1.696201752871275, 'learning_rate/qwen_vl_interface': 3.9423490023272326e-06, 'learning_rate/action_model': 3.9423490023272324e-05, 'epoch': 1.16}) +[2026-06-23 20:55:52,935][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.10273653268814087, 'train/grad_norm_pre_clip': 69.18567657470703, 'timing/data': 0.0006517148576676846, 'timing/model': 1.7027849359437823, 'learning_rate/qwen_vl_interface': 3.939421525491748e-06, 'learning_rate/action_model': 3.9394215254917483e-05, 'epoch': 1.16}) +[2026-06-23 20:55:56,279][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.09968014806509018, 'train/grad_norm_pre_clip': 26.325504302978516, 'timing/data': 0.0005967509932816029, 'timing/model': 1.699407656909898, 'learning_rate/qwen_vl_interface': 3.9364945873855705e-06, 'learning_rate/action_model': 3.93649458738557e-05, 'epoch': 1.16}) +[2026-06-23 20:55:59,620][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.16557520627975464, 'train/grad_norm_pre_clip': 25.91187286376953, 'timing/data': 0.0007901550270617008, 'timing/model': 1.698785156942904, 'learning_rate/qwen_vl_interface': 3.9335681892118536e-06, 'learning_rate/action_model': 3.933568189211854e-05, 'epoch': 1.16}) +[2026-06-23 20:56:02,959][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.09238961338996887, 'train/grad_norm_pre_clip': 17.357990264892578, 'timing/data': 0.000591641990467906, 'timing/model': 1.6982115081045777, 'learning_rate/qwen_vl_interface': 3.930642332173527e-06, 'learning_rate/action_model': 3.930642332173527e-05, 'epoch': 1.16}) +[2026-06-23 20:56:06,292][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.11722058057785034, 'train/grad_norm_pre_clip': 38.36425018310547, 'timing/data': 0.0005909909959882498, 'timing/model': 1.692781592020765, 'learning_rate/qwen_vl_interface': 3.927717017473301e-06, 'learning_rate/action_model': 3.9277170174733006e-05, 'epoch': 1.16}) +[2026-06-23 20:56:09,630][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.16655313968658447, 'train/grad_norm_pre_clip': 23.927064895629883, 'timing/data': 0.0006057179998606443, 'timing/model': 1.6926101290155202, 'learning_rate/qwen_vl_interface': 3.9247922463136615e-06, 'learning_rate/action_model': 3.924792246313662e-05, 'epoch': 1.16}) +[2026-06-23 20:56:12,960][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.15829512476921082, 'train/grad_norm_pre_clip': 55.86243438720703, 'timing/data': 0.0005846389103680849, 'timing/model': 1.6915537030436099, 'learning_rate/qwen_vl_interface': 3.921868019896873e-06, 'learning_rate/action_model': 3.921868019896873e-05, 'epoch': 1.16}) +[2026-06-23 20:56:16,399][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.09846784919500351, 'train/grad_norm_pre_clip': 25.913827896118164, 'timing/data': 0.0005868831649422646, 'timing/model': 1.6994986250065267, 'learning_rate/qwen_vl_interface': 3.918944339424974e-06, 'learning_rate/action_model': 3.9189443394249736e-05, 'epoch': 1.16}) +[2026-06-23 20:56:19,730][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.12853895127773285, 'train/grad_norm_pre_clip': 35.59791946411133, 'timing/data': 0.0005995470564812422, 'timing/model': 1.691557327983901, 'learning_rate/qwen_vl_interface': 3.9160212060997774e-06, 'learning_rate/action_model': 3.9160212060997774e-05, 'epoch': 1.16}) +[2026-06-23 20:56:23,073][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.04581617936491966, 'train/grad_norm_pre_clip': 10.648911476135254, 'timing/data': 0.0005919919349253178, 'timing/model': 1.7005508199799806, 'learning_rate/qwen_vl_interface': 3.913098621122875e-06, 'learning_rate/action_model': 3.913098621122875e-05, 'epoch': 1.16}) +[2026-06-23 20:56:26,410][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.08514338731765747, 'train/grad_norm_pre_clip': 16.97308349609375, 'timing/data': 0.000601460924372077, 'timing/model': 1.6964631569571793, 'learning_rate/qwen_vl_interface': 3.91017658569563e-06, 'learning_rate/action_model': 3.91017658569563e-05, 'epoch': 1.16}) +[2026-06-23 20:56:29,742][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.19303275644779205, 'train/grad_norm_pre_clip': 37.10825729370117, 'timing/data': 0.0005930340848863125, 'timing/model': 1.6944406051188707, 'learning_rate/qwen_vl_interface': 3.907255101019181e-06, 'learning_rate/action_model': 3.907255101019181e-05, 'epoch': 1.16}) +[2026-06-23 20:56:33,081][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.14948269724845886, 'train/grad_norm_pre_clip': 15.90888786315918, 'timing/data': 0.0005884459242224693, 'timing/model': 1.6975964060984552, 'learning_rate/qwen_vl_interface': 3.904334168294442e-06, 'learning_rate/action_model': 3.9043341682944415e-05, 'epoch': 1.16}) +[2026-06-23 20:56:36,414][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.06942456215620041, 'train/grad_norm_pre_clip': 19.079607009887695, 'timing/data': 0.0006070111412554979, 'timing/model': 1.6929307458922267, 'learning_rate/qwen_vl_interface': 3.901413788722093e-06, 'learning_rate/action_model': 3.901413788722093e-05, 'epoch': 1.16}) +[2026-06-23 20:56:39,759][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3000 +[2026-06-23 20:58:12,875][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3000 +[2026-06-23 20:58:12,876][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3000 +[2026-06-23 20:59:23,999][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3000 +[2026-06-23 20:59:37,453][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3000 +[2026-06-23 20:59:37,454][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.15128560364246368, 'train/grad_norm_pre_clip': 27.7713680267334, 'eval/action_loss/seconds': 93.1143518241588, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06859205663204193, 'eval/latency_2/loss': 0.06859205663204193, 'eval/flappy/loss': 0.06859205663204193, 'eval/flappy/latency_2/loss': 0.06859205663204193, 'eval/flap_precision': 0.9256578947368421, 'eval/flap_recall': 0.9493927125506073, 'eval/flap_f1': 0.9373750832778147, 'eval/flap_rate_model': 0.2111111111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9373750832778148, 'eval/latency_2/flap_precision': 0.9256578947368421, 'eval/latency_2/flap_recall': 0.9493927125506073, 'eval/latency_2/flap_f1': 0.9373750832778147, 'eval/latency_2/flap_rate_model': 0.2111111111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9373750832778148, 'eval/action_classification/data_wait_seconds': 0.8451777100563049, 'eval/action_classification/predict_seconds': 70.1408462524414, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 8.119999999999997, 'rl_games_eval/mid_train/mean_length': 81.4, 'rl_games_eval/mid_train/std_reward': 4.977308509626468, 'rl_games_eval/mid_train/std_length': 41.219412902175115, 'rl_games_eval/mid_train/macro_mean_reward': 8.119999999999997, 'rl_games_eval/mid_train/macro_mean_length': 81.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 8.119999999999997, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 81.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 4.977308509626468, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 41.219412902175115, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 8.119999999999997, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000616088043898344, 'timing/model': 1.7011256280820817, 'learning_rate/qwen_vl_interface': 3.8984939635025966e-06, 'learning_rate/action_model': 3.898493963502596e-05, 'epoch': 1.16}) +[2026-06-23 20:59:37,454][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_3000_state +[2026-06-23 20:59:37,457][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 20:59:40,103][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_3000_state/model.safetensors +[2026-06-23 20:59:41,244][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_3000_state/optimizer.bin +[2026-06-23 20:59:41,244][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_3000_state/sampler.bin +[2026-06-23 20:59:41,245][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_3000_state/sampler_1.bin +[2026-06-23 20:59:41,245][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_3000_state/random_states_0.pkl +[2026-06-23 20:59:41,271][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000 +[2026-06-23 20:59:41,271][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 20:59:41,281][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 20:59:44,573][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.11178456991910934, 'train/grad_norm_pre_clip': 31.87874412536621, 'timing/data': 0.0005905600264668465, 'timing/model': 1.667689804919064, 'learning_rate/qwen_vl_interface': 3.895574693836181e-06, 'learning_rate/action_model': 3.8955746938361804e-05, 'epoch': 1.16}) +[2026-06-23 20:59:47,872][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.12498509883880615, 'train/grad_norm_pre_clip': 22.196269989013672, 'timing/data': 0.000620837090536952, 'timing/model': 1.6776099789422005, 'learning_rate/qwen_vl_interface': 3.892655980922847e-06, 'learning_rate/action_model': 3.8926559809228466e-05, 'epoch': 1.16}) +[2026-06-23 20:59:51,168][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.1666945219039917, 'train/grad_norm_pre_clip': 81.4642105102539, 'timing/data': 0.0005833969917148352, 'timing/model': 1.674257627921179, 'learning_rate/qwen_vl_interface': 3.889737825962368e-06, 'learning_rate/action_model': 3.889737825962368e-05, 'epoch': 1.16}) +[2026-06-23 20:59:54,479][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.24981583654880524, 'train/grad_norm_pre_clip': 94.62310028076172, 'timing/data': 0.000585139961913228, 'timing/model': 1.6812914169859141, 'learning_rate/qwen_vl_interface': 3.886820230154288e-06, 'learning_rate/action_model': 3.886820230154287e-05, 'epoch': 1.16}) +[2026-06-23 20:59:57,787][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.20620591938495636, 'train/grad_norm_pre_clip': 85.78119659423828, 'timing/data': 0.0006068809889256954, 'timing/model': 1.67971606599167, 'learning_rate/qwen_vl_interface': 3.8839031946979184e-06, 'learning_rate/action_model': 3.883903194697918e-05, 'epoch': 1.16}) +[2026-06-23 21:00:01,097][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.11908812820911407, 'train/grad_norm_pre_clip': 40.80738830566406, 'timing/data': 0.0006044360343366861, 'timing/model': 1.6809286039788276, 'learning_rate/qwen_vl_interface': 3.880986720792344e-06, 'learning_rate/action_model': 3.880986720792344e-05, 'epoch': 1.16}) +[2026-06-23 21:00:04,408][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.15366590023040771, 'train/grad_norm_pre_clip': 32.16633224487305, 'timing/data': 0.0005971021018922329, 'timing/model': 1.6805788360070437, 'learning_rate/qwen_vl_interface': 3.878070809636416e-06, 'learning_rate/action_model': 3.878070809636416e-05, 'epoch': 1.16}) +[2026-06-23 21:00:07,724][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.10550538450479507, 'train/grad_norm_pre_clip': 40.87873458862305, 'timing/data': 0.0006021209992468357, 'timing/model': 1.6824809908866882, 'learning_rate/qwen_vl_interface': 3.875155462428758e-06, 'learning_rate/action_model': 3.875155462428758e-05, 'epoch': 1.16}) +[2026-06-23 21:00:11,044][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.1721300333738327, 'train/grad_norm_pre_clip': 30.013309478759766, 'timing/data': 0.0005943570286035538, 'timing/model': 1.6869576149620116, 'learning_rate/qwen_vl_interface': 3.872240680367754e-06, 'learning_rate/action_model': 3.872240680367754e-05, 'epoch': 1.16}) +[2026-06-23 21:00:14,361][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.07891729474067688, 'train/grad_norm_pre_clip': 15.422433853149414, 'timing/data': 0.0005895679350942373, 'timing/model': 1.6835097840521485, 'learning_rate/qwen_vl_interface': 3.869326464651566e-06, 'learning_rate/action_model': 3.869326464651566e-05, 'epoch': 1.16}) +[2026-06-23 21:00:17,678][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.1870635747909546, 'train/grad_norm_pre_clip': 21.381256103515625, 'timing/data': 0.0005753510631620884, 'timing/model': 1.686572700040415, 'learning_rate/qwen_vl_interface': 3.866412816478114e-06, 'learning_rate/action_model': 3.8664128164781144e-05, 'epoch': 1.17}) +[2026-06-23 21:00:20,995][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.11903072893619537, 'train/grad_norm_pre_clip': 12.26395320892334, 'timing/data': 0.0005749200936406851, 'timing/model': 1.6832208200357854, 'learning_rate/qwen_vl_interface': 3.8634997370450915e-06, 'learning_rate/action_model': 3.8634997370450914e-05, 'epoch': 1.17}) +[2026-06-23 21:00:24,311][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.15866392850875854, 'train/grad_norm_pre_clip': 28.480276107788086, 'timing/data': 0.0005779359489679337, 'timing/model': 1.6817914329003543, 'learning_rate/qwen_vl_interface': 3.860587227549956e-06, 'learning_rate/action_model': 3.860587227549955e-05, 'epoch': 1.17}) +[2026-06-23 21:00:27,790][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.07830005139112473, 'train/grad_norm_pre_clip': 20.048885345458984, 'timing/data': 0.0006105368956923485, 'timing/model': 1.8439923950936645, 'learning_rate/qwen_vl_interface': 3.857675289189926e-06, 'learning_rate/action_model': 3.857675289189926e-05, 'epoch': 1.17}) +[2026-06-23 21:00:31,109][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.07359064370393753, 'train/grad_norm_pre_clip': 25.501665115356445, 'timing/data': 0.0008134390227496624, 'timing/model': 1.6865040690172464, 'learning_rate/qwen_vl_interface': 3.854763923161993e-06, 'learning_rate/action_model': 3.854763923161992e-05, 'epoch': 1.17}) +[2026-06-23 21:00:34,438][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.1254817247390747, 'train/grad_norm_pre_clip': 73.89623260498047, 'timing/data': 0.0006331298500299454, 'timing/model': 1.6944976528175175, 'learning_rate/qwen_vl_interface': 3.851853130662908e-06, 'learning_rate/action_model': 3.8518531306629077e-05, 'epoch': 1.17}) +[2026-06-23 21:00:37,753][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.2334383875131607, 'train/grad_norm_pre_clip': 19.120908737182617, 'timing/data': 0.0006619039922952652, 'timing/model': 1.6822392628528178, 'learning_rate/qwen_vl_interface': 3.8489429128891884e-06, 'learning_rate/action_model': 3.848942912889188e-05, 'epoch': 1.17}) +[2026-06-23 21:00:41,073][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.19615481793880463, 'train/grad_norm_pre_clip': 21.17021942138672, 'timing/data': 0.0005929751787334681, 'timing/model': 1.6852918500080705, 'learning_rate/qwen_vl_interface': 3.846033271037114e-06, 'learning_rate/action_model': 3.8460332710371136e-05, 'epoch': 1.17}) +[2026-06-23 21:00:44,391][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.10195284336805344, 'train/grad_norm_pre_clip': 17.673328399658203, 'timing/data': 0.000595899997279048, 'timing/model': 1.685962239978835, 'learning_rate/qwen_vl_interface': 3.843124206302732e-06, 'learning_rate/action_model': 3.8431242063027316e-05, 'epoch': 1.17}) +[2026-06-23 21:00:47,710][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.13297368586063385, 'train/grad_norm_pre_clip': 35.499576568603516, 'timing/data': 0.0006737459916621447, 'timing/model': 1.6900655019562691, 'learning_rate/qwen_vl_interface': 3.840215719881844e-06, 'learning_rate/action_model': 3.8402157198818445e-05, 'epoch': 1.17}) +[2026-06-23 21:00:51,030][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.13179069757461548, 'train/grad_norm_pre_clip': 58.064796447753906, 'timing/data': 0.0006039149593561888, 'timing/model': 1.6868593150284141, 'learning_rate/qwen_vl_interface': 3.837307812970022e-06, 'learning_rate/action_model': 3.837307812970022e-05, 'epoch': 1.17}) +[2026-06-23 21:00:54,355][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.2008533924818039, 'train/grad_norm_pre_clip': 81.74884033203125, 'timing/data': 0.0006396719254553318, 'timing/model': 1.6891350008081645, 'learning_rate/qwen_vl_interface': 3.834400486762596e-06, 'learning_rate/action_model': 3.834400486762596e-05, 'epoch': 1.17}) +[2026-06-23 21:00:57,674][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.12024451047182083, 'train/grad_norm_pre_clip': 45.397743225097656, 'timing/data': 0.0006473059765994549, 'timing/model': 1.6854281541891396, 'learning_rate/qwen_vl_interface': 3.8314937424546585e-06, 'learning_rate/action_model': 3.8314937424546584e-05, 'epoch': 1.17}) +[2026-06-23 21:01:00,990][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.056772440671920776, 'train/grad_norm_pre_clip': 43.99647903442383, 'timing/data': 0.0005882550030946732, 'timing/model': 1.6832678329665214, 'learning_rate/qwen_vl_interface': 3.828587581241064e-06, 'learning_rate/action_model': 3.8285875812410635e-05, 'epoch': 1.17}) +[2026-06-23 21:01:04,307][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.05029216781258583, 'train/grad_norm_pre_clip': 36.38404083251953, 'timing/data': 0.0006933929398655891, 'timing/model': 1.682973851216957, 'learning_rate/qwen_vl_interface': 3.825682004316419e-06, 'learning_rate/action_model': 3.825682004316419e-05, 'epoch': 1.17}) +[2026-06-23 21:01:07,623][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.13099771738052368, 'train/grad_norm_pre_clip': 19.99771499633789, 'timing/data': 0.0005813220050185919, 'timing/model': 1.6869487911462784, 'learning_rate/qwen_vl_interface': 3.822777012875104e-06, 'learning_rate/action_model': 3.822777012875103e-05, 'epoch': 1.17}) +[2026-06-23 21:01:10,946][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.0944579541683197, 'train/grad_norm_pre_clip': 17.206953048706055, 'timing/data': 0.0005980541463941336, 'timing/model': 1.6892058730591089, 'learning_rate/qwen_vl_interface': 3.819872608111248e-06, 'learning_rate/action_model': 3.819872608111248e-05, 'epoch': 1.17}) +[2026-06-23 21:01:14,266][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.15344491600990295, 'train/grad_norm_pre_clip': 57.31028747558594, 'timing/data': 0.0005865930579602718, 'timing/model': 1.6858750320971012, 'learning_rate/qwen_vl_interface': 3.816968791218741e-06, 'learning_rate/action_model': 3.816968791218741e-05, 'epoch': 1.17}) +[2026-06-23 21:01:17,728][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.1495758593082428, 'train/grad_norm_pre_clip': 24.2579402923584, 'timing/data': 0.0005887469742447138, 'timing/model': 1.6881383450236171, 'learning_rate/qwen_vl_interface': 3.8140655633912337e-06, 'learning_rate/action_model': 3.814065563391234e-05, 'epoch': 1.17}) +[2026-06-23 21:01:21,059][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.09926298260688782, 'train/grad_norm_pre_clip': 10.554821014404297, 'timing/data': 0.0006883630994707346, 'timing/model': 1.6900595461484045, 'learning_rate/qwen_vl_interface': 3.811162925822133e-06, 'learning_rate/action_model': 3.811162925822132e-05, 'epoch': 1.17}) +[2026-06-23 21:01:24,384][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.08169643580913544, 'train/grad_norm_pre_clip': 12.88901138305664, 'timing/data': 0.0005721559282392263, 'timing/model': 1.6899986311327666, 'learning_rate/qwen_vl_interface': 3.8082608797046045e-06, 'learning_rate/action_model': 3.8082608797046045e-05, 'epoch': 1.17}) +[2026-06-23 21:01:27,714][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.11087856441736221, 'train/grad_norm_pre_clip': 16.355031967163086, 'timing/data': 0.000592994038015604, 'timing/model': 1.6869165080133826, 'learning_rate/qwen_vl_interface': 3.8053594262315687e-06, 'learning_rate/action_model': 3.805359426231568e-05, 'epoch': 1.17}) +[2026-06-23 21:01:31,038][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.15339456498622894, 'train/grad_norm_pre_clip': 18.41694450378418, 'timing/data': 0.0005959200207144022, 'timing/model': 1.6907848280388862, 'learning_rate/qwen_vl_interface': 3.802458566595704e-06, 'learning_rate/action_model': 3.802458566595704e-05, 'epoch': 1.17}) +[2026-06-23 21:01:34,366][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.17406214773654938, 'train/grad_norm_pre_clip': 17.917564392089844, 'timing/data': 0.0005899991374462843, 'timing/model': 1.6905682110227644, 'learning_rate/qwen_vl_interface': 3.7995583019894444e-06, 'learning_rate/action_model': 3.7995583019894445e-05, 'epoch': 1.17}) +[2026-06-23 21:01:37,686][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.08464089035987854, 'train/grad_norm_pre_clip': 28.438703536987305, 'timing/data': 0.0005766239482909441, 'timing/model': 1.686674322001636, 'learning_rate/qwen_vl_interface': 3.7966586336049782e-06, 'learning_rate/action_model': 3.796658633604978e-05, 'epoch': 1.17}) +[2026-06-23 21:01:41,011][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.1854988932609558, 'train/grad_norm_pre_clip': 55.0252571105957, 'timing/data': 0.0006061189342290163, 'timing/model': 1.6857989458367229, 'learning_rate/qwen_vl_interface': 3.7937595626342516e-06, 'learning_rate/action_model': 3.793759562634252e-05, 'epoch': 1.17}) +[2026-06-23 21:01:44,342][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.09415910392999649, 'train/grad_norm_pre_clip': 18.398679733276367, 'timing/data': 0.00081049301661551, 'timing/model': 1.6927713099867105, 'learning_rate/qwen_vl_interface': 3.79086109026896e-06, 'learning_rate/action_model': 3.7908610902689604e-05, 'epoch': 1.18}) +[2026-06-23 21:01:47,671][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.1687445193529129, 'train/grad_norm_pre_clip': 72.96865844726562, 'timing/data': 0.0009215120226144791, 'timing/model': 1.6904520499520004, 'learning_rate/qwen_vl_interface': 3.7879632177005584e-06, 'learning_rate/action_model': 3.787963217700558e-05, 'epoch': 1.18}) +[2026-06-23 21:01:50,999][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.16196735203266144, 'train/grad_norm_pre_clip': 57.34691619873047, 'timing/data': 0.0006679748184978962, 'timing/model': 1.6900969929993153, 'learning_rate/qwen_vl_interface': 3.7850659461202523e-06, 'learning_rate/action_model': 3.785065946120252e-05, 'epoch': 1.18}) +[2026-06-23 21:01:54,338][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.0655602440237999, 'train/grad_norm_pre_clip': 33.99347686767578, 'timing/data': 0.0008077281527221203, 'timing/model': 1.6963009629398584, 'learning_rate/qwen_vl_interface': 3.782169276719e-06, 'learning_rate/action_model': 3.782169276719e-05, 'epoch': 1.18}) +[2026-06-23 21:01:57,676][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.10641265660524368, 'train/grad_norm_pre_clip': 58.83988952636719, 'timing/data': 0.0006111189723014832, 'timing/model': 1.6969400129746646, 'learning_rate/qwen_vl_interface': 3.779273210687514e-06, 'learning_rate/action_model': 3.779273210687514e-05, 'epoch': 1.18}) +[2026-06-23 21:02:01,007][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.1025654673576355, 'train/grad_norm_pre_clip': 70.22003173828125, 'timing/data': 0.0008376240730285645, 'timing/model': 1.6911674339789897, 'learning_rate/qwen_vl_interface': 3.776377749216257e-06, 'learning_rate/action_model': 3.776377749216257e-05, 'epoch': 1.18}) +[2026-06-23 21:02:04,340][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.10453802347183228, 'train/grad_norm_pre_clip': 40.58052444458008, 'timing/data': 0.0005922329146414995, 'timing/model': 1.6937644351273775, 'learning_rate/qwen_vl_interface': 3.773482893495442e-06, 'learning_rate/action_model': 3.773482893495442e-05, 'epoch': 1.18}) +[2026-06-23 21:02:07,937][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.08970829099416733, 'train/grad_norm_pre_clip': 27.690580368041992, 'timing/data': 0.0005929339677095413, 'timing/model': 1.6933795199729502, 'learning_rate/qwen_vl_interface': 3.7705886447150375e-06, 'learning_rate/action_model': 3.7705886447150376e-05, 'epoch': 1.18}) +[2026-06-23 21:02:11,272][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.08150911331176758, 'train/grad_norm_pre_clip': 20.355411529541016, 'timing/data': 0.0006415850948542356, 'timing/model': 1.6964559420011938, 'learning_rate/qwen_vl_interface': 3.767695004064759e-06, 'learning_rate/action_model': 3.767695004064759e-05, 'epoch': 1.18}) +[2026-06-23 21:02:14,611][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.13343243300914764, 'train/grad_norm_pre_clip': 44.990699768066406, 'timing/data': 0.000817315885797143, 'timing/model': 1.6956687229685485, 'learning_rate/qwen_vl_interface': 3.7648019727340725e-06, 'learning_rate/action_model': 3.764801972734072e-05, 'epoch': 1.18}) +[2026-06-23 21:02:17,941][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.08349661529064178, 'train/grad_norm_pre_clip': 66.89987182617188, 'timing/data': 0.0006744880229234695, 'timing/model': 1.6937340460717678, 'learning_rate/qwen_vl_interface': 3.761909551912196e-06, 'learning_rate/action_model': 3.761909551912196e-05, 'epoch': 1.18}) +[2026-06-23 21:02:21,276][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.16516529023647308, 'train/grad_norm_pre_clip': 101.47271728515625, 'timing/data': 0.0005931449122726917, 'timing/model': 1.6967847598716617, 'learning_rate/qwen_vl_interface': 3.75901774278809e-06, 'learning_rate/action_model': 3.7590177427880904e-05, 'epoch': 1.18}) +[2026-06-23 21:02:24,612][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.2556675374507904, 'train/grad_norm_pre_clip': 125.83504486083984, 'timing/data': 0.0005923439748585224, 'timing/model': 1.6928603928536177, 'learning_rate/qwen_vl_interface': 3.756126546550472e-06, 'learning_rate/action_model': 3.756126546550472e-05, 'epoch': 1.18}) +[2026-06-23 21:02:27,944][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.08147512376308441, 'train/grad_norm_pre_clip': 67.83345794677734, 'timing/data': 0.000588505994528532, 'timing/model': 1.6941046910360456, 'learning_rate/qwen_vl_interface': 3.753235964387802e-06, 'learning_rate/action_model': 3.753235964387802e-05, 'epoch': 1.18}) +[2026-06-23 21:02:31,277][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.06444705277681351, 'train/grad_norm_pre_clip': 15.131540298461914, 'timing/data': 0.0006580972112715244, 'timing/model': 1.694191584130749, 'learning_rate/qwen_vl_interface': 3.750345997488289e-06, 'learning_rate/action_model': 3.7503459974882886e-05, 'epoch': 1.18}) +[2026-06-23 21:02:34,611][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.14670974016189575, 'train/grad_norm_pre_clip': 29.861371994018555, 'timing/data': 0.0006549209356307983, 'timing/model': 1.6892655859701335, 'learning_rate/qwen_vl_interface': 3.7474566470398903e-06, 'learning_rate/action_model': 3.74745664703989e-05, 'epoch': 1.18}) +[2026-06-23 21:02:37,945][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.1806359440088272, 'train/grad_norm_pre_clip': 27.21457862854004, 'timing/data': 0.0007659699767827988, 'timing/model': 1.691288388101384, 'learning_rate/qwen_vl_interface': 3.744567914230307e-06, 'learning_rate/action_model': 3.744567914230307e-05, 'epoch': 1.18}) +[2026-06-23 21:02:41,291][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.09658033400774002, 'train/grad_norm_pre_clip': 29.21802520751953, 'timing/data': 0.000980502925813198, 'timing/model': 1.705061140935868, 'learning_rate/qwen_vl_interface': 3.7416798002469902e-06, 'learning_rate/action_model': 3.74167980024699e-05, 'epoch': 1.18}) +[2026-06-23 21:02:44,625][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.19112707674503326, 'train/grad_norm_pre_clip': 66.53732299804688, 'timing/data': 0.000656074145808816, 'timing/model': 1.6937638490926474, 'learning_rate/qwen_vl_interface': 3.7387923062771313e-06, 'learning_rate/action_model': 3.738792306277131e-05, 'epoch': 1.18}) +[2026-06-23 21:02:47,959][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.12307127565145493, 'train/grad_norm_pre_clip': 69.22237396240234, 'timing/data': 0.0006852578371763229, 'timing/model': 1.6941184350289404, 'learning_rate/qwen_vl_interface': 3.7359054335076732e-06, 'learning_rate/action_model': 3.7359054335076734e-05, 'epoch': 1.18}) +[2026-06-23 21:02:51,300][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.19135811924934387, 'train/grad_norm_pre_clip': 28.975011825561523, 'timing/data': 0.0006582369096577168, 'timing/model': 1.6971149172168225, 'learning_rate/qwen_vl_interface': 3.7330191831253006e-06, 'learning_rate/action_model': 3.7330191831253e-05, 'epoch': 1.18}) +[2026-06-23 21:02:54,735][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.08249612152576447, 'train/grad_norm_pre_clip': 11.257487297058105, 'timing/data': 0.0010509651619940996, 'timing/model': 1.6995665438007563, 'learning_rate/qwen_vl_interface': 3.7301335563164382e-06, 'learning_rate/action_model': 3.730133556316438e-05, 'epoch': 1.18}) +[2026-06-23 21:02:58,075][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.10282914340496063, 'train/grad_norm_pre_clip': 14.907642364501953, 'timing/data': 0.0006458139978349209, 'timing/model': 1.6958464819472283, 'learning_rate/qwen_vl_interface': 3.727248554267261e-06, 'learning_rate/action_model': 3.727248554267261e-05, 'epoch': 1.18}) +[2026-06-23 21:03:01,410][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.10684065520763397, 'train/grad_norm_pre_clip': 22.573894500732422, 'timing/data': 0.0006455231923609972, 'timing/model': 1.6952467248775065, 'learning_rate/qwen_vl_interface': 3.7243641781636838e-06, 'learning_rate/action_model': 3.7243641781636836e-05, 'epoch': 1.18}) +[2026-06-23 21:03:04,749][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.16503113508224487, 'train/grad_norm_pre_clip': 23.379179000854492, 'timing/data': 0.0008243289776146412, 'timing/model': 1.696712207980454, 'learning_rate/qwen_vl_interface': 3.721480429191364e-06, 'learning_rate/action_model': 3.721480429191364e-05, 'epoch': 1.18}) +[2026-06-23 21:03:08,091][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.07741917669773102, 'train/grad_norm_pre_clip': 16.033910751342773, 'timing/data': 0.0006408339831978083, 'timing/model': 1.699447268852964, 'learning_rate/qwen_vl_interface': 3.7185973085357045e-06, 'learning_rate/action_model': 3.718597308535705e-05, 'epoch': 1.19}) +[2026-06-23 21:03:11,425][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.21781212091445923, 'train/grad_norm_pre_clip': 15.48218822479248, 'timing/data': 0.0005834768526256084, 'timing/model': 1.6970176321920007, 'learning_rate/qwen_vl_interface': 3.7157148173818443e-06, 'learning_rate/action_model': 3.715714817381844e-05, 'epoch': 1.19}) +[2026-06-23 21:03:14,758][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.03773834928870201, 'train/grad_norm_pre_clip': 10.654858589172363, 'timing/data': 0.0005853800103068352, 'timing/model': 1.6960127840284258, 'learning_rate/qwen_vl_interface': 3.7128329569146672e-06, 'learning_rate/action_model': 3.7128329569146666e-05, 'epoch': 1.19}) +[2026-06-23 21:03:18,100][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.2026786059141159, 'train/grad_norm_pre_clip': 19.45481300354004, 'timing/data': 0.000579178100451827, 'timing/model': 1.696008465019986, 'learning_rate/qwen_vl_interface': 3.7099517283187977e-06, 'learning_rate/action_model': 3.7099517283187976e-05, 'epoch': 1.19}) +[2026-06-23 21:03:21,443][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.17833997309207916, 'train/grad_norm_pre_clip': 38.66590118408203, 'timing/data': 0.0006591789424419403, 'timing/model': 1.6982330961618572, 'learning_rate/qwen_vl_interface': 3.7070711327786e-06, 'learning_rate/action_model': 3.7070711327785997e-05, 'epoch': 1.19}) +[2026-06-23 21:03:24,785][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.08528434485197067, 'train/grad_norm_pre_clip': 26.231658935546875, 'timing/data': 0.0006652709562331438, 'timing/model': 1.6987788509577513, 'learning_rate/qwen_vl_interface': 3.704191171478179e-06, 'learning_rate/action_model': 3.704191171478179e-05, 'epoch': 1.19}) +[2026-06-23 21:03:28,131][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.19571739435195923, 'train/grad_norm_pre_clip': 28.962860107421875, 'timing/data': 0.0006625349633395672, 'timing/model': 1.696900940965861, 'learning_rate/qwen_vl_interface': 3.701311845601376e-06, 'learning_rate/action_model': 3.701311845601376e-05, 'epoch': 1.19}) +[2026-06-23 21:03:31,474][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.06430178880691528, 'train/grad_norm_pre_clip': 23.793811798095703, 'timing/data': 0.0007059669587761164, 'timing/model': 1.6969128728378564, 'learning_rate/qwen_vl_interface': 3.6984331563317744e-06, 'learning_rate/action_model': 3.698433156331774e-05, 'epoch': 1.19}) +[2026-06-23 21:03:34,808][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.1635647416114807, 'train/grad_norm_pre_clip': 48.469783782958984, 'timing/data': 0.0005978939589112997, 'timing/model': 1.6936490589287132, 'learning_rate/qwen_vl_interface': 3.695555104852695e-06, 'learning_rate/action_model': 3.6955551048526946e-05, 'epoch': 1.19}) +[2026-06-23 21:03:38,146][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.07972371578216553, 'train/grad_norm_pre_clip': 72.92814636230469, 'timing/data': 0.0007064871024340391, 'timing/model': 1.6945385599974543, 'learning_rate/qwen_vl_interface': 3.6926776923471948e-06, 'learning_rate/action_model': 3.6926776923471946e-05, 'epoch': 1.19}) +[2026-06-23 21:03:41,486][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.06982553750276566, 'train/grad_norm_pre_clip': 54.13116455078125, 'timing/data': 0.0006161881610751152, 'timing/model': 1.693495298968628, 'learning_rate/qwen_vl_interface': 3.6898009199980715e-06, 'learning_rate/action_model': 3.6898009199980714e-05, 'epoch': 1.19}) +[2026-06-23 21:03:45,123][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.11979692429304123, 'train/grad_norm_pre_clip': 26.677471160888672, 'timing/data': 0.0006025920156389475, 'timing/model': 1.690923485904932, 'learning_rate/qwen_vl_interface': 3.686924788987855e-06, 'learning_rate/action_model': 3.686924788987855e-05, 'epoch': 1.19}) +[2026-06-23 21:03:48,456][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.13704776763916016, 'train/grad_norm_pre_clip': 41.9217643737793, 'timing/data': 0.0005985749885439873, 'timing/model': 1.6916387488599867, 'learning_rate/qwen_vl_interface': 3.6840493004988155e-06, 'learning_rate/action_model': 3.684049300498815e-05, 'epoch': 1.19}) +[2026-06-23 21:03:51,792][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.04823242872953415, 'train/grad_norm_pre_clip': 12.01193618774414, 'timing/data': 0.0005909199826419353, 'timing/model': 1.6940293519292027, 'learning_rate/qwen_vl_interface': 3.681174455712958e-06, 'learning_rate/action_model': 3.681174455712958e-05, 'epoch': 1.19}) +[2026-06-23 21:03:55,133][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.09081671386957169, 'train/grad_norm_pre_clip': 18.63751983642578, 'timing/data': 0.0005796800833195448, 'timing/model': 1.6959090649615973, 'learning_rate/qwen_vl_interface': 3.6783002558120206e-06, 'learning_rate/action_model': 3.67830025581202e-05, 'epoch': 1.19}) +[2026-06-23 21:03:58,469][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.06814762204885483, 'train/grad_norm_pre_clip': 42.75634765625, 'timing/data': 0.0005736381281167269, 'timing/model': 1.6962901700753719, 'learning_rate/qwen_vl_interface': 3.6754267019774804e-06, 'learning_rate/action_model': 3.67542670197748e-05, 'epoch': 1.19}) +[2026-06-23 21:04:01,807][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.09478534758090973, 'train/grad_norm_pre_clip': 24.295520782470703, 'timing/data': 0.0005890571046620607, 'timing/model': 1.6964993129950017, 'learning_rate/qwen_vl_interface': 3.6725537953905453e-06, 'learning_rate/action_model': 3.672553795390545e-05, 'epoch': 1.19}) +[2026-06-23 21:04:05,151][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.06841988861560822, 'train/grad_norm_pre_clip': 25.404218673706055, 'timing/data': 0.0005954790394753218, 'timing/model': 1.7028161461930722, 'learning_rate/qwen_vl_interface': 3.6696815372321603e-06, 'learning_rate/action_model': 3.6696815372321604e-05, 'epoch': 1.19}) +[2026-06-23 21:04:08,487][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.1824045479297638, 'train/grad_norm_pre_clip': 43.260650634765625, 'timing/data': 0.0006011200603097677, 'timing/model': 1.697122213896364, 'learning_rate/qwen_vl_interface': 3.6668099286830023e-06, 'learning_rate/action_model': 3.666809928683002e-05, 'epoch': 1.19}) +[2026-06-23 21:04:11,825][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.12086059153079987, 'train/grad_norm_pre_clip': 26.01084327697754, 'timing/data': 0.0006090649403631687, 'timing/model': 1.6975554679520428, 'learning_rate/qwen_vl_interface': 3.66393897092348e-06, 'learning_rate/action_model': 3.66393897092348e-05, 'epoch': 1.19}) +[2026-06-23 21:04:15,171][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.11032630503177643, 'train/grad_norm_pre_clip': 29.553434371948242, 'timing/data': 0.0005870540626347065, 'timing/model': 1.6982871119398624, 'learning_rate/qwen_vl_interface': 3.661068665133735e-06, 'learning_rate/action_model': 3.661068665133735e-05, 'epoch': 1.19}) +[2026-06-23 21:04:18,512][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.11935409903526306, 'train/grad_norm_pre_clip': 32.51701736450195, 'timing/data': 0.0006422270089387894, 'timing/model': 1.695966110099107, 'learning_rate/qwen_vl_interface': 3.6581990124936434e-06, 'learning_rate/action_model': 3.6581990124936434e-05, 'epoch': 1.19}) +[2026-06-23 21:04:21,855][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.1583244949579239, 'train/grad_norm_pre_clip': 26.9202938079834, 'timing/data': 0.0005850899033248425, 'timing/model': 1.699521993054077, 'learning_rate/qwen_vl_interface': 3.65533001418281e-06, 'learning_rate/action_model': 3.655330014182809e-05, 'epoch': 1.19}) +[2026-06-23 21:04:25,195][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.11494670063257217, 'train/grad_norm_pre_clip': 21.75098419189453, 'timing/data': 0.0007897440809756517, 'timing/model': 1.6947895179037005, 'learning_rate/qwen_vl_interface': 3.652461671380572e-06, 'learning_rate/action_model': 3.652461671380572e-05, 'epoch': 1.19}) +[2026-06-23 21:04:28,531][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.23282811045646667, 'train/grad_norm_pre_clip': 76.92529296875, 'timing/data': 0.0005898079834878445, 'timing/model': 1.6934524439275265, 'learning_rate/qwen_vl_interface': 3.6495939852659966e-06, 'learning_rate/action_model': 3.649593985265997e-05, 'epoch': 1.19}) +[2026-06-23 21:04:31,875][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.13049474358558655, 'train/grad_norm_pre_clip': 14.016169548034668, 'timing/data': 0.000598554965108633, 'timing/model': 1.6982604209333658, 'learning_rate/qwen_vl_interface': 3.646726957017881e-06, 'learning_rate/action_model': 3.646726957017881e-05, 'epoch': 1.19}) +[2026-06-23 21:04:35,450][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.09604325890541077, 'train/grad_norm_pre_clip': 39.557552337646484, 'timing/data': 0.000599356833845377, 'timing/model': 1.6955005237832665, 'learning_rate/qwen_vl_interface': 3.643860587814752e-06, 'learning_rate/action_model': 3.643860587814752e-05, 'epoch': 1.2}) +[2026-06-23 21:04:38,796][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.23554302752017975, 'train/grad_norm_pre_clip': 21.419034957885742, 'timing/data': 0.0008424830157309771, 'timing/model': 1.6988119969610125, 'learning_rate/qwen_vl_interface': 3.640994878834866e-06, 'learning_rate/action_model': 3.640994878834866e-05, 'epoch': 1.2}) +[2026-06-23 21:04:42,134][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.14486630260944366, 'train/grad_norm_pre_clip': 33.758628845214844, 'timing/data': 0.0005956799723207951, 'timing/model': 1.6947103380225599, 'learning_rate/qwen_vl_interface': 3.6381298312562095e-06, 'learning_rate/action_model': 3.6381298312562094e-05, 'epoch': 1.2}) +[2026-06-23 21:04:45,477][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.17464087903499603, 'train/grad_norm_pre_clip': 14.605504989624023, 'timing/data': 0.00060073914937675, 'timing/model': 1.6990017229691148, 'learning_rate/qwen_vl_interface': 3.635265446256493e-06, 'learning_rate/action_model': 3.6352654462564926e-05, 'epoch': 1.2}) +[2026-06-23 21:04:48,825][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.10710520297288895, 'train/grad_norm_pre_clip': 20.930940628051758, 'timing/data': 0.0005891670007258654, 'timing/model': 1.7031665612012148, 'learning_rate/qwen_vl_interface': 3.6324017250131565e-06, 'learning_rate/action_model': 3.6324017250131567e-05, 'epoch': 1.2}) +[2026-06-23 21:04:52,173][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.043417610228061676, 'train/grad_norm_pre_clip': 11.240057945251465, 'timing/data': 0.0006054581608623266, 'timing/model': 1.7020407740492374, 'learning_rate/qwen_vl_interface': 3.6295386687033696e-06, 'learning_rate/action_model': 3.6295386687033695e-05, 'epoch': 1.2}) +[2026-06-23 21:04:55,509][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.09819820523262024, 'train/grad_norm_pre_clip': 20.009185791015625, 'timing/data': 0.0005894480273127556, 'timing/model': 1.6973234678152949, 'learning_rate/qwen_vl_interface': 3.6266762785040253e-06, 'learning_rate/action_model': 3.626676278504025e-05, 'epoch': 1.2}) +[2026-06-23 21:04:58,856][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.09468826651573181, 'train/grad_norm_pre_clip': 25.561511993408203, 'timing/data': 0.0006204769015312195, 'timing/model': 1.7037193088326603, 'learning_rate/qwen_vl_interface': 3.6238145555917464e-06, 'learning_rate/action_model': 3.6238145555917465e-05, 'epoch': 1.2}) +[2026-06-23 21:05:02,205][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.18844769895076752, 'train/grad_norm_pre_clip': 16.539724349975586, 'timing/data': 0.0005907900631427765, 'timing/model': 1.6998711959458888, 'learning_rate/qwen_vl_interface': 3.6209535011428757e-06, 'learning_rate/action_model': 3.620953501142876e-05, 'epoch': 1.2}) +[2026-06-23 21:05:05,550][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.0477057546377182, 'train/grad_norm_pre_clip': 9.807546615600586, 'timing/data': 0.0006060490850359201, 'timing/model': 1.6968155428767204, 'learning_rate/qwen_vl_interface': 3.6180931163334863e-06, 'learning_rate/action_model': 3.6180931163334863e-05, 'epoch': 1.2}) +[2026-06-23 21:05:08,895][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.19172465801239014, 'train/grad_norm_pre_clip': 66.50653076171875, 'timing/data': 0.0007160159293562174, 'timing/model': 1.700362078147009, 'learning_rate/qwen_vl_interface': 3.6152334023393747e-06, 'learning_rate/action_model': 3.6152334023393745e-05, 'epoch': 1.2}) +[2026-06-23 21:05:12,244][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.08701009303331375, 'train/grad_norm_pre_clip': 16.168720245361328, 'timing/data': 0.0005720749031752348, 'timing/model': 1.6967191209550947, 'learning_rate/qwen_vl_interface': 3.6123743603360613e-06, 'learning_rate/action_model': 3.612374360336061e-05, 'epoch': 1.2}) +[2026-06-23 21:05:15,586][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3100 +[2026-06-23 21:06:48,883][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3100 +[2026-06-23 21:06:48,883][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3100 +[2026-06-23 21:08:00,286][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3100 +[2026-06-23 21:08:00,287][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.0831790342926979, 'train/grad_norm_pre_clip': 24.896089553833008, 'eval/action_loss/seconds': 93.29416952305473, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.0699504092335701, 'eval/latency_2/loss': 0.0699504092335701, 'eval/flappy/loss': 0.0699504092335701, 'eval/flappy/latency_2/loss': 0.0699504092335701, 'eval/flap_precision': 0.9424657534246575, 'eval/flap_recall': 0.9284750337381916, 'eval/flap_f1': 0.9354180829367776, 'eval/flap_rate_model': 0.20277777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9354180829367776, 'eval/latency_2/flap_precision': 0.9424657534246575, 'eval/latency_2/flap_recall': 0.9284750337381916, 'eval/latency_2/flap_f1': 0.9354180829367776, 'eval/latency_2/flap_rate_model': 0.20277777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9354180829367776, 'eval/action_classification/data_wait_seconds': 0.8176094889640808, 'eval/action_classification/predict_seconds': 70.45585632324219, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005962809082120657, 'timing/model': 1.69709053914994, 'learning_rate/qwen_vl_interface': 3.6095159914987893e-06, 'learning_rate/action_model': 3.6095159914987894e-05, 'epoch': 1.2}) +[2026-06-23 21:08:03,610][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.09863728284835815, 'train/grad_norm_pre_clip': 63.20642852783203, 'timing/data': 0.0005903700366616249, 'timing/model': 1.6868386310525239, 'learning_rate/qwen_vl_interface': 3.6066582970025266e-06, 'learning_rate/action_model': 3.6066582970025264e-05, 'epoch': 1.2}) +[2026-06-23 21:08:06,928][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.11153337359428406, 'train/grad_norm_pre_clip': 14.598977088928223, 'timing/data': 0.0005952180363237858, 'timing/model': 1.6869348911568522, 'learning_rate/qwen_vl_interface': 3.603801278021964e-06, 'learning_rate/action_model': 3.6038012780219635e-05, 'epoch': 1.2}) +[2026-06-23 21:08:10,260][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.09566663205623627, 'train/grad_norm_pre_clip': 20.207752227783203, 'timing/data': 0.0006854189559817314, 'timing/model': 1.690242206910625, 'learning_rate/qwen_vl_interface': 3.6009449357315122e-06, 'learning_rate/action_model': 3.6009449357315125e-05, 'epoch': 1.2}) +[2026-06-23 21:08:13,590][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.16801002621650696, 'train/grad_norm_pre_clip': 35.50138854980469, 'timing/data': 0.0008290188852697611, 'timing/model': 1.6916021530050784, 'learning_rate/qwen_vl_interface': 3.5980892713053084e-06, 'learning_rate/action_model': 3.598089271305308e-05, 'epoch': 1.2}) +[2026-06-23 21:08:16,918][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.08752752840518951, 'train/grad_norm_pre_clip': 34.980716705322266, 'timing/data': 0.0005931449122726917, 'timing/model': 1.692133171018213, 'learning_rate/qwen_vl_interface': 3.5952342859172063e-06, 'learning_rate/action_model': 3.595234285917206e-05, 'epoch': 1.2}) +[2026-06-23 21:08:20,277][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.1876254677772522, 'train/grad_norm_pre_clip': 19.68558120727539, 'timing/data': 0.0005899490788578987, 'timing/model': 1.6955030041281134, 'learning_rate/qwen_vl_interface': 3.592379980740785e-06, 'learning_rate/action_model': 3.5923799807407845e-05, 'epoch': 1.2}) +[2026-06-23 21:08:23,621][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.052350811660289764, 'train/grad_norm_pre_clip': 16.575851440429688, 'timing/data': 0.0005865518469363451, 'timing/model': 1.693124223034829, 'learning_rate/qwen_vl_interface': 3.5895263569493384e-06, 'learning_rate/action_model': 3.589526356949338e-05, 'epoch': 1.2}) +[2026-06-23 21:08:26,955][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.04457150772213936, 'train/grad_norm_pre_clip': 17.321197509765625, 'timing/data': 0.0005888361483812332, 'timing/model': 1.6954243660438806, 'learning_rate/qwen_vl_interface': 3.586673415715885e-06, 'learning_rate/action_model': 3.586673415715885e-05, 'epoch': 1.2}) +[2026-06-23 21:08:30,293][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.1357160359621048, 'train/grad_norm_pre_clip': 39.334754943847656, 'timing/data': 0.000623973086476326, 'timing/model': 1.6944911209866405, 'learning_rate/qwen_vl_interface': 3.5838211582131603e-06, 'learning_rate/action_model': 3.58382115821316e-05, 'epoch': 1.2}) +[2026-06-23 21:08:33,700][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.17273402214050293, 'train/grad_norm_pre_clip': 36.481719970703125, 'timing/data': 0.0006796070374548435, 'timing/model': 1.6945177919697016, 'learning_rate/qwen_vl_interface': 3.580969585613622e-06, 'learning_rate/action_model': 3.580969585613622e-05, 'epoch': 1.2}) +[2026-06-23 21:08:37,034][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.11869588494300842, 'train/grad_norm_pre_clip': 24.594711303710938, 'timing/data': 0.0006000280845910311, 'timing/model': 1.6922515220940113, 'learning_rate/qwen_vl_interface': 3.5781186990894423e-06, 'learning_rate/action_model': 3.578118699089442e-05, 'epoch': 1.2}) +[2026-06-23 21:08:40,368][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.04400959983468056, 'train/grad_norm_pre_clip': 39.36825942993164, 'timing/data': 0.0006980218458920717, 'timing/model': 1.694804570171982, 'learning_rate/qwen_vl_interface': 3.5752684998125113e-06, 'learning_rate/action_model': 3.5752684998125116e-05, 'epoch': 1.2}) +[2026-06-23 21:08:43,704][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.11716702580451965, 'train/grad_norm_pre_clip': 20.11440086364746, 'timing/data': 0.0006057480350136757, 'timing/model': 1.6969381000380963, 'learning_rate/qwen_vl_interface': 3.5724189889544407e-06, 'learning_rate/action_model': 3.572418988954441e-05, 'epoch': 1.2}) +[2026-06-23 21:08:47,041][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.14000998437404633, 'train/grad_norm_pre_clip': 20.565988540649414, 'timing/data': 0.000599126098677516, 'timing/model': 1.6939642808865756, 'learning_rate/qwen_vl_interface': 3.569570167686556e-06, 'learning_rate/action_model': 3.569570167686556e-05, 'epoch': 1.21}) +[2026-06-23 21:08:50,379][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.1612289547920227, 'train/grad_norm_pre_clip': 45.50837326049805, 'timing/data': 0.0005909309256821871, 'timing/model': 1.6968683691229671, 'learning_rate/qwen_vl_interface': 3.5667220371798995e-06, 'learning_rate/action_model': 3.566722037179899e-05, 'epoch': 1.21}) +[2026-06-23 21:08:53,708][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.1204063892364502, 'train/grad_norm_pre_clip': 58.389034271240234, 'timing/data': 0.0007816390134394169, 'timing/model': 1.6918489451054484, 'learning_rate/qwen_vl_interface': 3.563874598605232e-06, 'learning_rate/action_model': 3.5638745986052316e-05, 'epoch': 1.21}) +[2026-06-23 21:08:57,044][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.2343660145998001, 'train/grad_norm_pre_clip': 52.34673309326172, 'timing/data': 0.0005953791551291943, 'timing/model': 1.6938331839628518, 'learning_rate/qwen_vl_interface': 3.561027853133025e-06, 'learning_rate/action_model': 3.561027853133025e-05, 'epoch': 1.21}) +[2026-06-23 21:09:00,386][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.08802872896194458, 'train/grad_norm_pre_clip': 14.831721305847168, 'timing/data': 0.0006620241329073906, 'timing/model': 1.6963834671769291, 'learning_rate/qwen_vl_interface': 3.558181801933472e-06, 'learning_rate/action_model': 3.5581818019334714e-05, 'epoch': 1.21}) +[2026-06-23 21:09:03,730][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.13737080991268158, 'train/grad_norm_pre_clip': 15.061615943908691, 'timing/data': 0.0005923430435359478, 'timing/model': 1.7029529029969126, 'learning_rate/qwen_vl_interface': 3.555336446176473e-06, 'learning_rate/action_model': 3.5553364461764726e-05, 'epoch': 1.21}) +[2026-06-23 21:09:07,069][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.20781728625297546, 'train/grad_norm_pre_clip': 51.00795364379883, 'timing/data': 0.0007216059602797031, 'timing/model': 1.6958998469635844, 'learning_rate/qwen_vl_interface': 3.552491787031647e-06, 'learning_rate/action_model': 3.552491787031647e-05, 'epoch': 1.21}) +[2026-06-23 21:09:10,406][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.07779012620449066, 'train/grad_norm_pre_clip': 12.189152717590332, 'timing/data': 0.0006129730027168989, 'timing/model': 1.6957928959745914, 'learning_rate/qwen_vl_interface': 3.5496478256683277e-06, 'learning_rate/action_model': 3.549647825668327e-05, 'epoch': 1.21}) +[2026-06-23 21:09:13,742][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.048409655690193176, 'train/grad_norm_pre_clip': 6.813482761383057, 'timing/data': 0.0006787758320569992, 'timing/model': 1.6941100130788982, 'learning_rate/qwen_vl_interface': 3.5468045632555585e-06, 'learning_rate/action_model': 3.546804563255559e-05, 'epoch': 1.21}) +[2026-06-23 21:09:17,077][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.13581900298595428, 'train/grad_norm_pre_clip': 37.953529357910156, 'timing/data': 0.0006383799482136965, 'timing/model': 1.6935193229001015, 'learning_rate/qwen_vl_interface': 3.5439620009620995e-06, 'learning_rate/action_model': 3.543962000962099e-05, 'epoch': 1.21}) +[2026-06-23 21:09:20,505][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.11704888194799423, 'train/grad_norm_pre_clip': 67.35595703125, 'timing/data': 0.0006465448532253504, 'timing/model': 1.789714327082038, 'learning_rate/qwen_vl_interface': 3.541120139956419e-06, 'learning_rate/action_model': 3.5411201399564184e-05, 'epoch': 1.21}) +[2026-06-23 21:09:23,845][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.1371614933013916, 'train/grad_norm_pre_clip': 31.70297622680664, 'timing/data': 0.0005984539166092873, 'timing/model': 1.6964823710732162, 'learning_rate/qwen_vl_interface': 3.538278981406697e-06, 'learning_rate/action_model': 3.5382789814066966e-05, 'epoch': 1.21}) +[2026-06-23 21:09:27,190][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.040244754403829575, 'train/grad_norm_pre_clip': 10.313444137573242, 'timing/data': 0.0005855800118297338, 'timing/model': 1.7050942068453878, 'learning_rate/qwen_vl_interface': 3.5354385264808295e-06, 'learning_rate/action_model': 3.535438526480829e-05, 'epoch': 1.21}) +[2026-06-23 21:09:30,534][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.25502029061317444, 'train/grad_norm_pre_clip': 24.0952091217041, 'timing/data': 0.0008236679714173079, 'timing/model': 1.7009958839043975, 'learning_rate/qwen_vl_interface': 3.5325987763464196e-06, 'learning_rate/action_model': 3.5325987763464194e-05, 'epoch': 1.21}) +[2026-06-23 21:09:33,874][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.05631781369447708, 'train/grad_norm_pre_clip': 48.96247863769531, 'timing/data': 0.0006806498859077692, 'timing/model': 1.6939812200143933, 'learning_rate/qwen_vl_interface': 3.529759732170781e-06, 'learning_rate/action_model': 3.529759732170781e-05, 'epoch': 1.21}) +[2026-06-23 21:09:37,220][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.072681725025177, 'train/grad_norm_pre_clip': 13.587767601013184, 'timing/data': 0.0006053480319678783, 'timing/model': 1.7008615320082754, 'learning_rate/qwen_vl_interface': 3.5269213951209375e-06, 'learning_rate/action_model': 3.5269213951209374e-05, 'epoch': 1.21}) +[2026-06-23 21:09:40,567][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.20651021599769592, 'train/grad_norm_pre_clip': 74.50686645507812, 'timing/data': 0.0006228310521692038, 'timing/model': 1.699311258038506, 'learning_rate/qwen_vl_interface': 3.5240837663636207e-06, 'learning_rate/action_model': 3.5240837663636204e-05, 'epoch': 1.21}) +[2026-06-23 21:09:43,909][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.1203036904335022, 'train/grad_norm_pre_clip': 28.92325782775879, 'timing/data': 0.0005999880377203226, 'timing/model': 1.7016264689154923, 'learning_rate/qwen_vl_interface': 3.5212468470652745e-06, 'learning_rate/action_model': 3.521246847065274e-05, 'epoch': 1.21}) +[2026-06-23 21:09:47,254][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.03158041089773178, 'train/grad_norm_pre_clip': 30.48419189453125, 'timing/data': 0.0006427678745239973, 'timing/model': 1.6991298869252205, 'learning_rate/qwen_vl_interface': 3.5184106383920487e-06, 'learning_rate/action_model': 3.5184106383920486e-05, 'epoch': 1.21}) +[2026-06-23 21:09:50,590][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.17938564717769623, 'train/grad_norm_pre_clip': 54.55825424194336, 'timing/data': 0.00060424511320889, 'timing/model': 1.694317582063377, 'learning_rate/qwen_vl_interface': 3.5155751415098022e-06, 'learning_rate/action_model': 3.515575141509802e-05, 'epoch': 1.21}) +[2026-06-23 21:09:53,922][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.06532930582761765, 'train/grad_norm_pre_clip': 20.23950958251953, 'timing/data': 0.0006042360328137875, 'timing/model': 1.6917805538978428, 'learning_rate/qwen_vl_interface': 3.5127403575841e-06, 'learning_rate/action_model': 3.5127403575841e-05, 'epoch': 1.21}) +[2026-06-23 21:09:57,268][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.1851939857006073, 'train/grad_norm_pre_clip': 22.544666290283203, 'timing/data': 0.0006467259954661131, 'timing/model': 1.6949645881541073, 'learning_rate/qwen_vl_interface': 3.509906287780215e-06, 'learning_rate/action_model': 3.5099062877802144e-05, 'epoch': 1.21}) +[2026-06-23 21:10:00,608][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.06529434770345688, 'train/grad_norm_pre_clip': 26.705400466918945, 'timing/data': 0.0008182469755411148, 'timing/model': 1.6950523618143052, 'learning_rate/qwen_vl_interface': 3.5070729332631247e-06, 'learning_rate/action_model': 3.507072933263125e-05, 'epoch': 1.21}) +[2026-06-23 21:10:03,948][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.11740217357873917, 'train/grad_norm_pre_clip': 21.905927658081055, 'timing/data': 0.0007064780220389366, 'timing/model': 1.695488492026925, 'learning_rate/qwen_vl_interface': 3.5042402951975167e-06, 'learning_rate/action_model': 3.5042402951975167e-05, 'epoch': 1.21}) +[2026-06-23 21:10:07,293][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.20764116942882538, 'train/grad_norm_pre_clip': 114.44141387939453, 'timing/data': 0.0006173909641802311, 'timing/model': 1.7000074149109423, 'learning_rate/qwen_vl_interface': 3.50140837474778e-06, 'learning_rate/action_model': 3.50140837474778e-05, 'epoch': 1.21}) +[2026-06-23 21:10:10,796][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.10549081116914749, 'train/grad_norm_pre_clip': 111.82977294921875, 'timing/data': 0.0005967309698462486, 'timing/model': 1.8568998349364847, 'learning_rate/qwen_vl_interface': 3.498577173078011e-06, 'learning_rate/action_model': 3.498577173078011e-05, 'epoch': 1.21}) +[2026-06-23 21:10:14,132][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.14171232283115387, 'train/grad_norm_pre_clip': 89.72376251220703, 'timing/data': 0.000636545941233635, 'timing/model': 1.6937993369065225, 'learning_rate/qwen_vl_interface': 3.495746691352008e-06, 'learning_rate/action_model': 3.495746691352008e-05, 'epoch': 1.22}) +[2026-06-23 21:10:17,478][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.11280250549316406, 'train/grad_norm_pre_clip': 59.65163803100586, 'timing/data': 0.0005915309302508831, 'timing/model': 1.6991165899671614, 'learning_rate/qwen_vl_interface': 3.4929169307332763e-06, 'learning_rate/action_model': 3.492916930733276e-05, 'epoch': 1.22}) +[2026-06-23 21:10:20,826][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.18994580209255219, 'train/grad_norm_pre_clip': 55.04207229614258, 'timing/data': 0.0005896880757063627, 'timing/model': 1.700583586934954, 'learning_rate/qwen_vl_interface': 3.4900878923850234e-06, 'learning_rate/action_model': 3.490087892385023e-05, 'epoch': 1.22}) +[2026-06-23 21:10:24,173][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.2637244760990143, 'train/grad_norm_pre_clip': 73.96846008300781, 'timing/data': 0.000594958895817399, 'timing/model': 1.7007774219382554, 'learning_rate/qwen_vl_interface': 3.4872595774701607e-06, 'learning_rate/action_model': 3.487259577470161e-05, 'epoch': 1.22}) +[2026-06-23 21:10:27,520][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.14776058495044708, 'train/grad_norm_pre_clip': 18.765531539916992, 'timing/data': 0.000838044798001647, 'timing/model': 1.705136233009398, 'learning_rate/qwen_vl_interface': 3.4844319871513014e-06, 'learning_rate/action_model': 3.484431987151301e-05, 'epoch': 1.22}) +[2026-06-23 21:10:30,867][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.19469669461250305, 'train/grad_norm_pre_clip': 38.065711975097656, 'timing/data': 0.0005916520021855831, 'timing/model': 1.700231194961816, 'learning_rate/qwen_vl_interface': 3.4816051225907606e-06, 'learning_rate/action_model': 3.48160512259076e-05, 'epoch': 1.22}) +[2026-06-23 21:10:34,207][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.18041063845157623, 'train/grad_norm_pre_clip': 20.932872772216797, 'timing/data': 0.0005945870652794838, 'timing/model': 1.6950320729520172, 'learning_rate/qwen_vl_interface': 3.478778984950556e-06, 'learning_rate/action_model': 3.478778984950556e-05, 'epoch': 1.22}) +[2026-06-23 21:10:37,549][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.18589991331100464, 'train/grad_norm_pre_clip': 95.88418579101562, 'timing/data': 0.0005816728807985783, 'timing/model': 1.6985518878791481, 'learning_rate/qwen_vl_interface': 3.475953575392405e-06, 'learning_rate/action_model': 3.475953575392405e-05, 'epoch': 1.22}) +[2026-06-23 21:10:40,890][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.1628217250108719, 'train/grad_norm_pre_clip': 88.46864318847656, 'timing/data': 0.0005908110179007053, 'timing/model': 1.698555744951591, 'learning_rate/qwen_vl_interface': 3.473128895077728e-06, 'learning_rate/action_model': 3.473128895077728e-05, 'epoch': 1.22}) +[2026-06-23 21:10:44,232][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.14525021612644196, 'train/grad_norm_pre_clip': 45.48729705810547, 'timing/data': 0.0008054140489548445, 'timing/model': 1.6969601961318403, 'learning_rate/qwen_vl_interface': 3.4703049451676453e-06, 'learning_rate/action_model': 3.4703049451676454e-05, 'epoch': 1.22}) +[2026-06-23 21:10:47,578][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.15815405547618866, 'train/grad_norm_pre_clip': 87.62528228759766, 'timing/data': 0.0005828449502587318, 'timing/model': 1.7014225330203772, 'learning_rate/qwen_vl_interface': 3.4674817268229732e-06, 'learning_rate/action_model': 3.467481726822973e-05, 'epoch': 1.22}) +[2026-06-23 21:10:50,926][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.07512687891721725, 'train/grad_norm_pre_clip': 42.92790603637695, 'timing/data': 0.0007470338605344296, 'timing/model': 1.6992881908081472, 'learning_rate/qwen_vl_interface': 3.464659241204232e-06, 'learning_rate/action_model': 3.4646592412042314e-05, 'epoch': 1.22}) +[2026-06-23 21:10:54,269][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.07543346285820007, 'train/grad_norm_pre_clip': 62.4847526550293, 'timing/data': 0.0006507830694317818, 'timing/model': 1.6978388570714742, 'learning_rate/qwen_vl_interface': 3.4618374894716377e-06, 'learning_rate/action_model': 3.461837489471638e-05, 'epoch': 1.22}) +[2026-06-23 21:10:57,621][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.0849752277135849, 'train/grad_norm_pre_clip': 10.586091041564941, 'timing/data': 0.0005803098902106285, 'timing/model': 1.6982753281481564, 'learning_rate/qwen_vl_interface': 3.459016472785107e-06, 'learning_rate/action_model': 3.459016472785107e-05, 'epoch': 1.22}) +[2026-06-23 21:11:01,106][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.06999882310628891, 'train/grad_norm_pre_clip': 12.795402526855469, 'timing/data': 0.0006070809904485941, 'timing/model': 1.7009191759862006, 'learning_rate/qwen_vl_interface': 3.456196192304254e-06, 'learning_rate/action_model': 3.456196192304254e-05, 'epoch': 1.22}) +[2026-06-23 21:11:04,451][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.08210233598947525, 'train/grad_norm_pre_clip': 28.63420295715332, 'timing/data': 0.0005995659157633781, 'timing/model': 1.6979151598643512, 'learning_rate/qwen_vl_interface': 3.4533766491883887e-06, 'learning_rate/action_model': 3.453376649188388e-05, 'epoch': 1.22}) +[2026-06-23 21:11:07,793][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.18935814499855042, 'train/grad_norm_pre_clip': 33.58720779418945, 'timing/data': 0.0006038041319698095, 'timing/model': 1.6976444909814745, 'learning_rate/qwen_vl_interface': 3.450557844596517e-06, 'learning_rate/action_model': 3.450557844596517e-05, 'epoch': 1.22}) +[2026-06-23 21:11:11,125][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.10840242356061935, 'train/grad_norm_pre_clip': 25.285066604614258, 'timing/data': 0.0005893181078135967, 'timing/model': 1.6916859429329634, 'learning_rate/qwen_vl_interface': 3.447739779687345e-06, 'learning_rate/action_model': 3.4477397796873446e-05, 'epoch': 1.22}) +[2026-06-23 21:11:14,469][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.09172902256250381, 'train/grad_norm_pre_clip': 23.752519607543945, 'timing/data': 0.0005945879966020584, 'timing/model': 1.697161684045568, 'learning_rate/qwen_vl_interface': 3.4449224556192722e-06, 'learning_rate/action_model': 3.444922455619272e-05, 'epoch': 1.22}) +[2026-06-23 21:11:17,814][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.15987929701805115, 'train/grad_norm_pre_clip': 19.792865753173828, 'timing/data': 0.0006086041685193777, 'timing/model': 1.6980418479070067, 'learning_rate/qwen_vl_interface': 3.442105873550395e-06, 'learning_rate/action_model': 3.442105873550395e-05, 'epoch': 1.22}) +[2026-06-23 21:11:21,146][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.11734025180339813, 'train/grad_norm_pre_clip': 39.776512145996094, 'timing/data': 0.0005876040086150169, 'timing/model': 1.6940819048322737, 'learning_rate/qwen_vl_interface': 3.4392900346385022e-06, 'learning_rate/action_model': 3.4392900346385024e-05, 'epoch': 1.22}) +[2026-06-23 21:11:24,487][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.014690927229821682, 'train/grad_norm_pre_clip': 21.924938201904297, 'timing/data': 0.0005934261716902256, 'timing/model': 1.6978217239957303, 'learning_rate/qwen_vl_interface': 3.4364749400410796e-06, 'learning_rate/action_model': 3.43647494004108e-05, 'epoch': 1.22}) +[2026-06-23 21:11:27,827][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.1389547735452652, 'train/grad_norm_pre_clip': 18.41777992248535, 'timing/data': 0.0005871530156582594, 'timing/model': 1.6938103949651122, 'learning_rate/qwen_vl_interface': 3.4336605909153066e-06, 'learning_rate/action_model': 3.433660590915306e-05, 'epoch': 1.22}) +[2026-06-23 21:11:31,168][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.0629129484295845, 'train/grad_norm_pre_clip': 9.96967601776123, 'timing/data': 0.000583296874538064, 'timing/model': 1.7018523691222072, 'learning_rate/qwen_vl_interface': 3.430846988418057e-06, 'learning_rate/action_model': 3.430846988418057e-05, 'epoch': 1.22}) +[2026-06-23 21:11:34,513][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.09968827664852142, 'train/grad_norm_pre_clip': 37.731937408447266, 'timing/data': 0.00059500802308321, 'timing/model': 1.6979280540253967, 'learning_rate/qwen_vl_interface': 3.428034133705893e-06, 'learning_rate/action_model': 3.428034133705893e-05, 'epoch': 1.22}) +[2026-06-23 21:11:37,863][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.13671183586120605, 'train/grad_norm_pre_clip': 46.87773895263672, 'timing/data': 0.0005849790759384632, 'timing/model': 1.6992850739043206, 'learning_rate/qwen_vl_interface': 3.4252220279350763e-06, 'learning_rate/action_model': 3.425222027935076e-05, 'epoch': 1.22}) +[2026-06-23 21:11:41,210][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.13190466165542603, 'train/grad_norm_pre_clip': 22.81729507446289, 'timing/data': 0.0006684060208499432, 'timing/model': 1.7003375620115548, 'learning_rate/qwen_vl_interface': 3.4224106722615547e-06, 'learning_rate/action_model': 3.422410672261554e-05, 'epoch': 1.23}) +[2026-06-23 21:11:44,562][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.14683474600315094, 'train/grad_norm_pre_clip': 43.60987854003906, 'timing/data': 0.0006832540966570377, 'timing/model': 1.7045161160640419, 'learning_rate/qwen_vl_interface': 3.419600067840973e-06, 'learning_rate/action_model': 3.419600067840973e-05, 'epoch': 1.23}) +[2026-06-23 21:11:47,908][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.18123023211956024, 'train/grad_norm_pre_clip': 27.071813583374023, 'timing/data': 0.000650141853839159, 'timing/model': 1.7000134221743792, 'learning_rate/qwen_vl_interface': 3.4167902158286627e-06, 'learning_rate/action_model': 3.4167902158286626e-05, 'epoch': 1.23}) +[2026-06-23 21:11:51,524][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.18569977581501007, 'train/grad_norm_pre_clip': 13.988276481628418, 'timing/data': 0.0006984029896557331, 'timing/model': 1.7018660937901586, 'learning_rate/qwen_vl_interface': 3.4139811173796468e-06, 'learning_rate/action_model': 3.413981117379647e-05, 'epoch': 1.23}) +[2026-06-23 21:11:54,873][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.11010416597127914, 'train/grad_norm_pre_clip': 51.80720138549805, 'timing/data': 0.0008991200011223555, 'timing/model': 1.702423221198842, 'learning_rate/qwen_vl_interface': 3.411172773648642e-06, 'learning_rate/action_model': 3.411172773648642e-05, 'epoch': 1.23}) +[2026-06-23 21:11:58,216][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.048605870455503464, 'train/grad_norm_pre_clip': 9.158061981201172, 'timing/data': 0.0008176260162144899, 'timing/model': 1.6959832469001412, 'learning_rate/qwen_vl_interface': 3.408365185790052e-06, 'learning_rate/action_model': 3.4083651857900515e-05, 'epoch': 1.23}) +[2026-06-23 21:12:01,552][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.07325000315904617, 'train/grad_norm_pre_clip': 26.108566284179688, 'timing/data': 0.0006767318118363619, 'timing/model': 1.6923850649036467, 'learning_rate/qwen_vl_interface': 3.40555835495797e-06, 'learning_rate/action_model': 3.40555835495797e-05, 'epoch': 1.23}) +[2026-06-23 21:12:04,891][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.2769372761249542, 'train/grad_norm_pre_clip': 12.764156341552734, 'timing/data': 0.000589447095990181, 'timing/model': 1.6947633551899344, 'learning_rate/qwen_vl_interface': 3.4027522823061763e-06, 'learning_rate/action_model': 3.4027522823061765e-05, 'epoch': 1.23}) +[2026-06-23 21:12:08,230][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.19513918459415436, 'train/grad_norm_pre_clip': 27.53299331665039, 'timing/data': 0.0006351741030812263, 'timing/model': 1.693998778006062, 'learning_rate/qwen_vl_interface': 3.399946968988143e-06, 'learning_rate/action_model': 3.399946968988143e-05, 'epoch': 1.23}) +[2026-06-23 21:12:11,571][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.12009228765964508, 'train/grad_norm_pre_clip': 21.06184959411621, 'timing/data': 0.0005926741287112236, 'timing/model': 1.695552096934989, 'learning_rate/qwen_vl_interface': 3.397142416157028e-06, 'learning_rate/action_model': 3.397142416157028e-05, 'epoch': 1.23}) +[2026-06-23 21:12:14,902][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.1378873586654663, 'train/grad_norm_pre_clip': 20.53094482421875, 'timing/data': 0.0006028430070728064, 'timing/model': 1.691171354148537, 'learning_rate/qwen_vl_interface': 3.3943386249656766e-06, 'learning_rate/action_model': 3.394338624965677e-05, 'epoch': 1.23}) +[2026-06-23 21:12:18,238][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.10554418712854385, 'train/grad_norm_pre_clip': 15.593612670898438, 'timing/data': 0.0006009589415043592, 'timing/model': 1.6937702568247914, 'learning_rate/qwen_vl_interface': 3.391535596566624e-06, 'learning_rate/action_model': 3.391535596566624e-05, 'epoch': 1.23}) +[2026-06-23 21:12:21,573][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.09293416887521744, 'train/grad_norm_pre_clip': 66.65962219238281, 'timing/data': 0.0006067801732569933, 'timing/model': 1.6941556630190462, 'learning_rate/qwen_vl_interface': 3.388733332112086e-06, 'learning_rate/action_model': 3.3887333321120856e-05, 'epoch': 1.23}) +[2026-06-23 21:12:24,912][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.1348041296005249, 'train/grad_norm_pre_clip': 13.078756332397461, 'timing/data': 0.0006909889634698629, 'timing/model': 1.6976057360880077, 'learning_rate/qwen_vl_interface': 3.3859318327539693e-06, 'learning_rate/action_model': 3.385931832753969e-05, 'epoch': 1.23}) +[2026-06-23 21:12:28,255][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.022798841819167137, 'train/grad_norm_pre_clip': 18.161331176757812, 'timing/data': 0.0007298120763152838, 'timing/model': 1.6976527639199048, 'learning_rate/qwen_vl_interface': 3.3831310996438633e-06, 'learning_rate/action_model': 3.383131099643863e-05, 'epoch': 1.23}) +[2026-06-23 21:12:31,595][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.05349249392747879, 'train/grad_norm_pre_clip': 20.477632522583008, 'timing/data': 0.0005741091445088387, 'timing/model': 1.6947405100800097, 'learning_rate/qwen_vl_interface': 3.3803311339330447e-06, 'learning_rate/action_model': 3.3803311339330444e-05, 'epoch': 1.23}) +[2026-06-23 21:12:34,939][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.1128215566277504, 'train/grad_norm_pre_clip': 25.401626586914062, 'timing/data': 0.0005971719510853291, 'timing/model': 1.6996911049354821, 'learning_rate/qwen_vl_interface': 3.3775319367724746e-06, 'learning_rate/action_model': 3.3775319367724745e-05, 'epoch': 1.23}) +[2026-06-23 21:12:38,372][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.06899002939462662, 'train/grad_norm_pre_clip': 19.777748107910156, 'timing/data': 0.0005992569494992495, 'timing/model': 1.6982135081198066, 'learning_rate/qwen_vl_interface': 3.374733509312795e-06, 'learning_rate/action_model': 3.374733509312795e-05, 'epoch': 1.23}) +[2026-06-23 21:12:41,715][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.16570639610290527, 'train/grad_norm_pre_clip': 23.893951416015625, 'timing/data': 0.0005806009285151958, 'timing/model': 1.698688460048288, 'learning_rate/qwen_vl_interface': 3.3719358527043345e-06, 'learning_rate/action_model': 3.371935852704334e-05, 'epoch': 1.23}) +[2026-06-23 21:12:45,063][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.05298482999205589, 'train/grad_norm_pre_clip': 27.4176025390625, 'timing/data': 0.0006583770737051964, 'timing/model': 1.7000020591076463, 'learning_rate/qwen_vl_interface': 3.3691389680971047e-06, 'learning_rate/action_model': 3.3691389680971044e-05, 'epoch': 1.23}) +[2026-06-23 21:12:48,409][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.08496107161045074, 'train/grad_norm_pre_clip': 33.121524810791016, 'timing/data': 0.0005934450309723616, 'timing/model': 1.701288416981697, 'learning_rate/qwen_vl_interface': 3.3663428566407985e-06, 'learning_rate/action_model': 3.366342856640798e-05, 'epoch': 1.23}) +[2026-06-23 21:12:51,758][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.09417267143726349, 'train/grad_norm_pre_clip': 37.3555793762207, 'timing/data': 0.0006053380202502012, 'timing/model': 1.7000947338528931, 'learning_rate/qwen_vl_interface': 3.3635475194847927e-06, 'learning_rate/action_model': 3.363547519484793e-05, 'epoch': 1.23}) +[2026-06-23 21:12:55,102][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.20112290978431702, 'train/grad_norm_pre_clip': 18.03097152709961, 'timing/data': 0.0005968220066279173, 'timing/model': 1.6988114509731531, 'learning_rate/qwen_vl_interface': 3.3607529577781433e-06, 'learning_rate/action_model': 3.360752957778143e-05, 'epoch': 1.23}) +[2026-06-23 21:12:58,451][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.16100049018859863, 'train/grad_norm_pre_clip': 33.254066467285156, 'timing/data': 0.0007213258650153875, 'timing/model': 1.6985690149012953, 'learning_rate/qwen_vl_interface': 3.3579591726695906e-06, 'learning_rate/action_model': 3.35795917266959e-05, 'epoch': 1.23}) +[2026-06-23 21:13:01,799][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.0536298006772995, 'train/grad_norm_pre_clip': 44.09953689575195, 'timing/data': 0.0009315710049122572, 'timing/model': 1.7018315470777452, 'learning_rate/qwen_vl_interface': 3.3551661653075522e-06, 'learning_rate/action_model': 3.355166165307552e-05, 'epoch': 1.23}) +[2026-06-23 21:13:05,149][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.05673676356673241, 'train/grad_norm_pre_clip': 18.10547637939453, 'timing/data': 0.0005930447950959206, 'timing/model': 1.7035440860781819, 'learning_rate/qwen_vl_interface': 3.35237393684013e-06, 'learning_rate/action_model': 3.35237393684013e-05, 'epoch': 1.23}) +[2026-06-23 21:13:08,498][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.12043839693069458, 'train/grad_norm_pre_clip': 24.037548065185547, 'timing/data': 0.0005924131255596876, 'timing/model': 1.7005401090718806, 'learning_rate/qwen_vl_interface': 3.349582488415103e-06, 'learning_rate/action_model': 3.3495824884151026e-05, 'epoch': 1.24}) +[2026-06-23 21:13:11,845][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.08579425513744354, 'train/grad_norm_pre_clip': 39.754608154296875, 'timing/data': 0.0005944869481027126, 'timing/model': 1.699403292965144, 'learning_rate/qwen_vl_interface': 3.3467918211799294e-06, 'learning_rate/action_model': 3.3467918211799295e-05, 'epoch': 1.24}) +[2026-06-23 21:13:15,187][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.1385553628206253, 'train/grad_norm_pre_clip': 48.54887771606445, 'timing/data': 0.0006006590556353331, 'timing/model': 1.6993833461310714, 'learning_rate/qwen_vl_interface': 3.3440019362817476e-06, 'learning_rate/action_model': 3.344001936281747e-05, 'epoch': 1.24}) +[2026-06-23 21:13:18,522][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.11636077612638474, 'train/grad_norm_pre_clip': 15.809579849243164, 'timing/data': 0.0006399920675903559, 'timing/model': 1.6932939309626818, 'learning_rate/qwen_vl_interface': 3.3412128348673745e-06, 'learning_rate/action_model': 3.341212834867374e-05, 'epoch': 1.24}) +[2026-06-23 21:13:21,864][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.054075825959444046, 'train/grad_norm_pre_clip': 9.27043628692627, 'timing/data': 0.0005822440143674612, 'timing/model': 1.6975678829476237, 'learning_rate/qwen_vl_interface': 3.338424518083303e-06, 'learning_rate/action_model': 3.338424518083303e-05, 'epoch': 1.24}) +[2026-06-23 21:13:25,206][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.09673672914505005, 'train/grad_norm_pre_clip': 37.05747604370117, 'timing/data': 0.0005891069304198027, 'timing/model': 1.7001689819153398, 'learning_rate/qwen_vl_interface': 3.3356369870757056e-06, 'learning_rate/action_model': 3.335636987075705e-05, 'epoch': 1.24}) +[2026-06-23 21:13:28,861][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.0722651258111, 'train/grad_norm_pre_clip': 13.41092586517334, 'timing/data': 0.0005851390305906534, 'timing/model': 1.6998843369074166, 'learning_rate/qwen_vl_interface': 3.3328502429904327e-06, 'learning_rate/action_model': 3.3328502429904325e-05, 'epoch': 1.24}) +[2026-06-23 21:13:32,208][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.15896017849445343, 'train/grad_norm_pre_clip': 58.125911712646484, 'timing/data': 0.0005991461221128702, 'timing/model': 1.6997194869909436, 'learning_rate/qwen_vl_interface': 3.3300642869730066e-06, 'learning_rate/action_model': 3.330064286973006e-05, 'epoch': 1.24}) +[2026-06-23 21:13:35,551][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3200 +[2026-06-23 21:15:08,807][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3200 +[2026-06-23 21:15:08,808][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3200 +[2026-06-23 21:16:19,826][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3200 +[2026-06-23 21:16:19,827][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.15561342239379883, 'train/grad_norm_pre_clip': 12.493876457214355, 'eval/action_loss/seconds': 93.2541430159472, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.0675433799624443, 'eval/latency_2/loss': 0.0675433799624443, 'eval/flappy/loss': 0.0675433799624443, 'eval/flappy/latency_2/loss': 0.0675433799624443, 'eval/flap_precision': 0.9182879377431906, 'eval/flap_recall': 0.9554655870445344, 'eval/flap_f1': 0.9365079365079365, 'eval/flap_rate_model': 0.21416666666666667, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9365079365079365, 'eval/latency_2/flap_precision': 0.9182879377431906, 'eval/latency_2/flap_recall': 0.9554655870445344, 'eval/latency_2/flap_f1': 0.9365079365079365, 'eval/latency_2/flap_rate_model': 0.21416666666666667, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9365079365079365, 'eval/action_classification/data_wait_seconds': 0.9515487551689148, 'eval/action_classification/predict_seconds': 69.9285659790039, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005875739734619856, 'timing/model': 1.699053023941815, 'learning_rate/qwen_vl_interface': 3.3272791201686294e-06, 'learning_rate/action_model': 3.327279120168629e-05, 'epoch': 1.24}) +[2026-06-23 21:16:23,140][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.08477865904569626, 'train/grad_norm_pre_clip': 46.94358444213867, 'timing/data': 0.0005932650528848171, 'timing/model': 1.6817276279907674, 'learning_rate/qwen_vl_interface': 3.324494743722179e-06, 'learning_rate/action_model': 3.3244947437221785e-05, 'epoch': 1.24}) +[2026-06-23 21:16:26,450][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.0968736857175827, 'train/grad_norm_pre_clip': 11.905380249023438, 'timing/data': 0.0006050069350749254, 'timing/model': 1.6795317609794438, 'learning_rate/qwen_vl_interface': 3.321711158778205e-06, 'learning_rate/action_model': 3.321711158778205e-05, 'epoch': 1.24}) +[2026-06-23 21:16:29,763][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.08222673088312149, 'train/grad_norm_pre_clip': 23.719514846801758, 'timing/data': 0.0005926941521465778, 'timing/model': 1.6839785990305245, 'learning_rate/qwen_vl_interface': 3.3189283664809393e-06, 'learning_rate/action_model': 3.318928366480939e-05, 'epoch': 1.24}) +[2026-06-23 21:16:33,078][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.11513639986515045, 'train/grad_norm_pre_clip': 15.878012657165527, 'timing/data': 0.0005962608847767115, 'timing/model': 1.6857756259851158, 'learning_rate/qwen_vl_interface': 3.316146367974276e-06, 'learning_rate/action_model': 3.316146367974276e-05, 'epoch': 1.24}) +[2026-06-23 21:16:36,398][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.09664221853017807, 'train/grad_norm_pre_clip': 40.80419158935547, 'timing/data': 0.0005998581182211637, 'timing/model': 1.6873192870989442, 'learning_rate/qwen_vl_interface': 3.3133651644017945e-06, 'learning_rate/action_model': 3.3133651644017944e-05, 'epoch': 1.24}) +[2026-06-23 21:16:39,824][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.11509799212217331, 'train/grad_norm_pre_clip': 21.063941955566406, 'timing/data': 0.0006156070157885551, 'timing/model': 1.6869643500540406, 'learning_rate/qwen_vl_interface': 3.31058475690674e-06, 'learning_rate/action_model': 3.31058475690674e-05, 'epoch': 1.24}) +[2026-06-23 21:16:43,143][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.11202528327703476, 'train/grad_norm_pre_clip': 71.84870910644531, 'timing/data': 0.000586140900850296, 'timing/model': 1.6849390929564834, 'learning_rate/qwen_vl_interface': 3.3078051466320337e-06, 'learning_rate/action_model': 3.307805146632034e-05, 'epoch': 1.24}) +[2026-06-23 21:16:46,472][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.0699731633067131, 'train/grad_norm_pre_clip': 14.870774269104004, 'timing/data': 0.000596269965171814, 'timing/model': 1.6924785340670496, 'learning_rate/qwen_vl_interface': 3.3050263347202712e-06, 'learning_rate/action_model': 3.305026334720271e-05, 'epoch': 1.24}) +[2026-06-23 21:16:49,798][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.14151319861412048, 'train/grad_norm_pre_clip': 26.678695678710938, 'timing/data': 0.0005930748302489519, 'timing/model': 1.690947274910286, 'learning_rate/qwen_vl_interface': 3.302248322313712e-06, 'learning_rate/action_model': 3.302248322313712e-05, 'epoch': 1.24}) +[2026-06-23 21:16:53,121][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.1534033864736557, 'train/grad_norm_pre_clip': 62.54945373535156, 'timing/data': 0.000614524818956852, 'timing/model': 1.688781163888052, 'learning_rate/qwen_vl_interface': 3.299471110554295e-06, 'learning_rate/action_model': 3.299471110554295e-05, 'epoch': 1.24}) +[2026-06-23 21:16:56,455][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.08171349763870239, 'train/grad_norm_pre_clip': 28.78413200378418, 'timing/data': 0.0005983640439808369, 'timing/model': 1.6899214580189437, 'learning_rate/qwen_vl_interface': 3.296694700583626e-06, 'learning_rate/action_model': 3.296694700583626e-05, 'epoch': 1.24}) +[2026-06-23 21:16:59,782][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.07657045125961304, 'train/grad_norm_pre_clip': 13.003250122070312, 'timing/data': 0.0007397800218313932, 'timing/model': 1.6914476358797401, 'learning_rate/qwen_vl_interface': 3.2939190935429833e-06, 'learning_rate/action_model': 3.293919093542983e-05, 'epoch': 1.24}) +[2026-06-23 21:17:03,105][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.19314436614513397, 'train/grad_norm_pre_clip': 26.6583194732666, 'timing/data': 0.0006000280845910311, 'timing/model': 1.686014574021101, 'learning_rate/qwen_vl_interface': 3.2911442905733126e-06, 'learning_rate/action_model': 3.291144290573312e-05, 'epoch': 1.24}) +[2026-06-23 21:17:06,427][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.13318072259426117, 'train/grad_norm_pre_clip': 25.746196746826172, 'timing/data': 0.0006120200268924236, 'timing/model': 1.6898574267979711, 'learning_rate/qwen_vl_interface': 3.2883702928152323e-06, 'learning_rate/action_model': 3.288370292815232e-05, 'epoch': 1.24}) +[2026-06-23 21:17:09,754][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.16528792679309845, 'train/grad_norm_pre_clip': 14.651163101196289, 'timing/data': 0.0005857609212398529, 'timing/model': 1.689683370059356, 'learning_rate/qwen_vl_interface': 3.2855971014090275e-06, 'learning_rate/action_model': 3.285597101409028e-05, 'epoch': 1.24}) +[2026-06-23 21:17:13,091][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.18199008703231812, 'train/grad_norm_pre_clip': 30.92445945739746, 'timing/data': 0.000659929821267724, 'timing/model': 1.693852094002068, 'learning_rate/qwen_vl_interface': 3.2828247174946528e-06, 'learning_rate/action_model': 3.2828247174946524e-05, 'epoch': 1.24}) +[2026-06-23 21:17:16,431][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.03721385821700096, 'train/grad_norm_pre_clip': 15.540486335754395, 'timing/data': 0.0005914620123803616, 'timing/model': 1.6953554400242865, 'learning_rate/qwen_vl_interface': 3.280053142211729e-06, 'learning_rate/action_model': 3.280053142211729e-05, 'epoch': 1.25}) +[2026-06-23 21:17:19,760][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.07507950812578201, 'train/grad_norm_pre_clip': 17.006601333618164, 'timing/data': 0.0006219090428203344, 'timing/model': 1.693276273086667, 'learning_rate/qwen_vl_interface': 3.2772823766995475e-06, 'learning_rate/action_model': 3.277282376699548e-05, 'epoch': 1.25}) +[2026-06-23 21:17:23,094][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.12419447302818298, 'train/grad_norm_pre_clip': 15.644363403320312, 'timing/data': 0.0005899290554225445, 'timing/model': 1.6929498380050063, 'learning_rate/qwen_vl_interface': 3.2745124220970646e-06, 'learning_rate/action_model': 3.274512422097065e-05, 'epoch': 1.25}) +[2026-06-23 21:17:26,426][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.07590937614440918, 'train/grad_norm_pre_clip': 22.817134857177734, 'timing/data': 0.0006053380202502012, 'timing/model': 1.6904234800022095, 'learning_rate/qwen_vl_interface': 3.2717432795429057e-06, 'learning_rate/action_model': 3.271743279542905e-05, 'epoch': 1.25}) +[2026-06-23 21:17:30,052][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.05578906834125519, 'train/grad_norm_pre_clip': 11.618302345275879, 'timing/data': 0.0006501120515167713, 'timing/model': 1.6851402909960598, 'learning_rate/qwen_vl_interface': 3.2689749501753578e-06, 'learning_rate/action_model': 3.2689749501753574e-05, 'epoch': 1.25}) +[2026-06-23 21:17:33,384][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.10039731115102768, 'train/grad_norm_pre_clip': 19.464529037475586, 'timing/data': 0.0005819129291921854, 'timing/model': 1.6896998800802976, 'learning_rate/qwen_vl_interface': 3.2662074351323774e-06, 'learning_rate/action_model': 3.266207435132378e-05, 'epoch': 1.25}) +[2026-06-23 21:17:36,713][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.11991114169359207, 'train/grad_norm_pre_clip': 38.6882209777832, 'timing/data': 0.0006666029803454876, 'timing/model': 1.6883778050541878, 'learning_rate/qwen_vl_interface': 3.2634407355515875e-06, 'learning_rate/action_model': 3.2634407355515875e-05, 'epoch': 1.25}) +[2026-06-23 21:17:40,040][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.10171674191951752, 'train/grad_norm_pre_clip': 24.99525260925293, 'timing/data': 0.0006695978809148073, 'timing/model': 1.692407889990136, 'learning_rate/qwen_vl_interface': 3.2606748525702715e-06, 'learning_rate/action_model': 3.260674852570271e-05, 'epoch': 1.25}) +[2026-06-23 21:17:43,369][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.16728553175926208, 'train/grad_norm_pre_clip': 53.015567779541016, 'timing/data': 0.0006630760617554188, 'timing/model': 1.6896986369974911, 'learning_rate/qwen_vl_interface': 3.2579097873253803e-06, 'learning_rate/action_model': 3.25790978732538e-05, 'epoch': 1.25}) +[2026-06-23 21:17:46,707][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.04351760447025299, 'train/grad_norm_pre_clip': 19.722864151000977, 'timing/data': 0.0006042469758540392, 'timing/model': 1.6949901960324496, 'learning_rate/qwen_vl_interface': 3.2551455409535274e-06, 'learning_rate/action_model': 3.255145540953528e-05, 'epoch': 1.25}) +[2026-06-23 21:17:50,046][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.05865650251507759, 'train/grad_norm_pre_clip': 31.943296432495117, 'timing/data': 0.0008596279658377171, 'timing/model': 1.693092942936346, 'learning_rate/qwen_vl_interface': 3.2523821145909907e-06, 'learning_rate/action_model': 3.252382114590991e-05, 'epoch': 1.25}) +[2026-06-23 21:17:53,378][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.07586240023374557, 'train/grad_norm_pre_clip': 14.522439002990723, 'timing/data': 0.0005901600234210491, 'timing/model': 1.692247328814119, 'learning_rate/qwen_vl_interface': 3.249619509373712e-06, 'learning_rate/action_model': 3.249619509373712e-05, 'epoch': 1.25}) +[2026-06-23 21:17:56,718][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.15157929062843323, 'train/grad_norm_pre_clip': 20.56837272644043, 'timing/data': 0.0006493520922958851, 'timing/model': 1.6952750131022185, 'learning_rate/qwen_vl_interface': 3.2468577264372905e-06, 'learning_rate/action_model': 3.24685772643729e-05, 'epoch': 1.25}) +[2026-06-23 21:18:00,051][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.15093834698200226, 'train/grad_norm_pre_clip': 43.37275314331055, 'timing/data': 0.0005995479878038168, 'timing/model': 1.690721333026886, 'learning_rate/qwen_vl_interface': 3.244096766916992e-06, 'learning_rate/action_model': 3.244096766916992e-05, 'epoch': 1.25}) +[2026-06-23 21:18:03,383][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.04856090247631073, 'train/grad_norm_pre_clip': 30.396146774291992, 'timing/data': 0.0005929050967097282, 'timing/model': 1.6946998168714345, 'learning_rate/qwen_vl_interface': 3.241336631947743e-06, 'learning_rate/action_model': 3.2413366319477427e-05, 'epoch': 1.25}) +[2026-06-23 21:18:06,711][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.08792443573474884, 'train/grad_norm_pre_clip': 20.20419692993164, 'timing/data': 0.0005939069669693708, 'timing/model': 1.6919873710721731, 'learning_rate/qwen_vl_interface': 3.238577322664131e-06, 'learning_rate/action_model': 3.238577322664131e-05, 'epoch': 1.25}) +[2026-06-23 21:18:10,039][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.2578478157520294, 'train/grad_norm_pre_clip': 49.62973403930664, 'timing/data': 0.0005829969886690378, 'timing/model': 1.687214698875323, 'learning_rate/qwen_vl_interface': 3.2358188402004026e-06, 'learning_rate/action_model': 3.2358188402004024e-05, 'epoch': 1.25}) +[2026-06-23 21:18:13,373][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.16227804124355316, 'train/grad_norm_pre_clip': 48.490299224853516, 'timing/data': 0.000594638055190444, 'timing/model': 1.6930344868451357, 'learning_rate/qwen_vl_interface': 3.2330611856904675e-06, 'learning_rate/action_model': 3.233061185690468e-05, 'epoch': 1.25}) +[2026-06-23 21:18:16,710][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.09784180670976639, 'train/grad_norm_pre_clip': 30.902019500732422, 'timing/data': 0.0005917830858379602, 'timing/model': 1.7004038619343191, 'learning_rate/qwen_vl_interface': 3.2303043602678928e-06, 'learning_rate/action_model': 3.230304360267893e-05, 'epoch': 1.25}) +[2026-06-23 21:18:20,351][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.10709010809659958, 'train/grad_norm_pre_clip': 47.18149185180664, 'timing/data': 0.000585292000323534, 'timing/model': 1.694820763077587, 'learning_rate/qwen_vl_interface': 3.227548365065905e-06, 'learning_rate/action_model': 3.227548365065905e-05, 'epoch': 1.25}) +[2026-06-23 21:18:23,684][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.13305702805519104, 'train/grad_norm_pre_clip': 36.1021614074707, 'timing/data': 0.0006965810898691416, 'timing/model': 1.6920895630028099, 'learning_rate/qwen_vl_interface': 3.2247932012173887e-06, 'learning_rate/action_model': 3.2247932012173884e-05, 'epoch': 1.25}) +[2026-06-23 21:18:27,021][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.0939313992857933, 'train/grad_norm_pre_clip': 50.40224075317383, 'timing/data': 0.0007778028957545757, 'timing/model': 1.6943176540080458, 'learning_rate/qwen_vl_interface': 3.2220388698548897e-06, 'learning_rate/action_model': 3.22203886985489e-05, 'epoch': 1.25}) +[2026-06-23 21:18:30,359][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.09910117834806442, 'train/grad_norm_pre_clip': 73.08954620361328, 'timing/data': 0.0006803900469094515, 'timing/model': 1.6944215891417116, 'learning_rate/qwen_vl_interface': 3.21928537211061e-06, 'learning_rate/action_model': 3.21928537211061e-05, 'epoch': 1.25}) +[2026-06-23 21:18:33,691][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.29809409379959106, 'train/grad_norm_pre_clip': 74.60978698730469, 'timing/data': 0.0006036160048097372, 'timing/model': 1.6935192190576345, 'learning_rate/qwen_vl_interface': 3.216532709116405e-06, 'learning_rate/action_model': 3.216532709116405e-05, 'epoch': 1.25}) +[2026-06-23 21:18:37,020][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.055968280881643295, 'train/grad_norm_pre_clip': 40.92743682861328, 'timing/data': 0.0006238229107111692, 'timing/model': 1.6879328668583184, 'learning_rate/qwen_vl_interface': 3.2137808820037925e-06, 'learning_rate/action_model': 3.2137808820037924e-05, 'epoch': 1.25}) +[2026-06-23 21:18:40,355][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.09186659008264542, 'train/grad_norm_pre_clip': 42.8496208190918, 'timing/data': 0.0005964720621705055, 'timing/model': 1.693293523043394, 'learning_rate/qwen_vl_interface': 3.211029891903945e-06, 'learning_rate/action_model': 3.211029891903945e-05, 'epoch': 1.25}) +[2026-06-23 21:18:43,688][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.12932921946048737, 'train/grad_norm_pre_clip': 40.470909118652344, 'timing/data': 0.0006036350969225168, 'timing/model': 1.6936774039641023, 'learning_rate/qwen_vl_interface': 3.2082797399476894e-06, 'learning_rate/action_model': 3.2082797399476895e-05, 'epoch': 1.26}) +[2026-06-23 21:18:47,025][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.19586506485939026, 'train/grad_norm_pre_clip': 16.034568786621094, 'timing/data': 0.0005985959433019161, 'timing/model': 1.6974410461261868, 'learning_rate/qwen_vl_interface': 3.2055304272655114e-06, 'learning_rate/action_model': 3.205530427265511e-05, 'epoch': 1.26}) +[2026-06-23 21:18:50,356][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.1769254356622696, 'train/grad_norm_pre_clip': 39.6727180480957, 'timing/data': 0.0006002490408718586, 'timing/model': 1.6879816281143576, 'learning_rate/qwen_vl_interface': 3.2027819549875477e-06, 'learning_rate/action_model': 3.202781954987548e-05, 'epoch': 1.26}) +[2026-06-23 21:18:53,696][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.07876273989677429, 'train/grad_norm_pre_clip': 51.66143798828125, 'timing/data': 0.0005903700366616249, 'timing/model': 1.7007822510786355, 'learning_rate/qwen_vl_interface': 3.2000343242435926e-06, 'learning_rate/action_model': 3.2000343242435924e-05, 'epoch': 1.26}) +[2026-06-23 21:18:57,036][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.13305149972438812, 'train/grad_norm_pre_clip': 19.503753662109375, 'timing/data': 0.0009384150616824627, 'timing/model': 1.6922978279180825, 'learning_rate/qwen_vl_interface': 3.1972875361630923e-06, 'learning_rate/action_model': 3.1972875361630923e-05, 'epoch': 1.26}) +[2026-06-23 21:19:00,374][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.14963114261627197, 'train/grad_norm_pre_clip': 31.667648315429688, 'timing/data': 0.0005893290508538485, 'timing/model': 1.6953328491654247, 'learning_rate/qwen_vl_interface': 3.1945415918751484e-06, 'learning_rate/action_model': 3.194541591875148e-05, 'epoch': 1.26}) +[2026-06-23 21:19:03,704][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.04859133064746857, 'train/grad_norm_pre_clip': 21.650644302368164, 'timing/data': 0.0005946590099483728, 'timing/model': 1.6909863110631704, 'learning_rate/qwen_vl_interface': 3.191796492508516e-06, 'learning_rate/action_model': 3.191796492508516e-05, 'epoch': 1.26}) +[2026-06-23 21:20:22,669][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3250 +[2026-06-23 21:20:22,670][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.06978872418403625, 'train/grad_norm_pre_clip': 14.519410133361816, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 55.09999999999926, 'rl_games_eval/mid_train/mean_length': 463.0, 'rl_games_eval/mid_train/std_reward': 51.20605432954091, 'rl_games_eval/mid_train/std_length': 413.1425903970686, 'rl_games_eval/mid_train/macro_mean_reward': 55.09999999999926, 'rl_games_eval/mid_train/macro_mean_length': 463.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 55.09999999999926, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 463.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 51.20605432954091, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 413.1425903970686, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 55.09999999999926, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006003491580486298, 'timing/model': 1.6939603260252625, 'learning_rate/qwen_vl_interface': 3.189052239191599e-06, 'learning_rate/action_model': 3.189052239191599e-05, 'epoch': 1.26}) +[2026-06-23 21:20:25,958][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.08423744142055511, 'train/grad_norm_pre_clip': 18.142505645751953, 'timing/data': 0.0007038540206849575, 'timing/model': 1.6734333410859108, 'learning_rate/qwen_vl_interface': 3.1863088330524577e-06, 'learning_rate/action_model': 3.186308833052458e-05, 'epoch': 1.26}) +[2026-06-23 21:20:29,246][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.08434537798166275, 'train/grad_norm_pre_clip': 36.11726379394531, 'timing/data': 0.0005929060280323029, 'timing/model': 1.6692676779348403, 'learning_rate/qwen_vl_interface': 3.183566275218803e-06, 'learning_rate/action_model': 3.1835662752188026e-05, 'epoch': 1.26}) +[2026-06-23 21:20:32,543][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.11223795264959335, 'train/grad_norm_pre_clip': 25.329349517822266, 'timing/data': 0.0006090558599680662, 'timing/model': 1.674662551144138, 'learning_rate/qwen_vl_interface': 3.180824566817996e-06, 'learning_rate/action_model': 3.1808245668179956e-05, 'epoch': 1.26}) +[2026-06-23 21:20:35,832][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.13564348220825195, 'train/grad_norm_pre_clip': 27.787517547607422, 'timing/data': 0.00069439597427845, 'timing/model': 1.6723349860403687, 'learning_rate/qwen_vl_interface': 3.1780837089770524e-06, 'learning_rate/action_model': 3.1780837089770524e-05, 'epoch': 1.26}) +[2026-06-23 21:20:39,135][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.2348368763923645, 'train/grad_norm_pre_clip': 21.64394187927246, 'timing/data': 0.0006067210342735052, 'timing/model': 1.6780111079569906, 'learning_rate/qwen_vl_interface': 3.17534370282263e-06, 'learning_rate/action_model': 3.1753437028226303e-05, 'epoch': 1.26}) +[2026-06-23 21:20:42,440][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.17733809351921082, 'train/grad_norm_pre_clip': 17.579078674316406, 'timing/data': 0.000662266043946147, 'timing/model': 1.6776924130972475, 'learning_rate/qwen_vl_interface': 3.172604549481046e-06, 'learning_rate/action_model': 3.1726045494810455e-05, 'epoch': 1.26}) +[2026-06-23 21:20:45,755][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.19461458921432495, 'train/grad_norm_pre_clip': 22.463178634643555, 'timing/data': 0.0006484591867774725, 'timing/model': 1.686556010041386, 'learning_rate/qwen_vl_interface': 3.169866250078261e-06, 'learning_rate/action_model': 3.169866250078261e-05, 'epoch': 1.26}) +[2026-06-23 21:20:49,068][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.159688800573349, 'train/grad_norm_pre_clip': 19.726364135742188, 'timing/data': 0.0007897061295807362, 'timing/model': 1.6858219939749688, 'learning_rate/qwen_vl_interface': 3.167128805739887e-06, 'learning_rate/action_model': 3.1671288057398866e-05, 'epoch': 1.26}) +[2026-06-23 21:20:52,377][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.13618971407413483, 'train/grad_norm_pre_clip': 15.239566802978516, 'timing/data': 0.00058951904065907, 'timing/model': 1.6823059930466115, 'learning_rate/qwen_vl_interface': 3.1643922175911847e-06, 'learning_rate/action_model': 3.164392217591185e-05, 'epoch': 1.26}) +[2026-06-23 21:20:55,706][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.0901784673333168, 'train/grad_norm_pre_clip': 16.774280548095703, 'timing/data': 0.0006055000703781843, 'timing/model': 1.6890542418695986, 'learning_rate/qwen_vl_interface': 3.1616564867570603e-06, 'learning_rate/action_model': 3.16165648675706e-05, 'epoch': 1.26}) +[2026-06-23 21:20:59,225][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.17593689262866974, 'train/grad_norm_pre_clip': 36.552310943603516, 'timing/data': 0.0006644600071012974, 'timing/model': 1.68600640213117, 'learning_rate/qwen_vl_interface': 3.1589216143620733e-06, 'learning_rate/action_model': 3.1589216143620734e-05, 'epoch': 1.26}) +[2026-06-23 21:21:02,547][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.05467529594898224, 'train/grad_norm_pre_clip': 8.13341236114502, 'timing/data': 0.0006174009758979082, 'timing/model': 1.6881022790912539, 'learning_rate/qwen_vl_interface': 3.1561876015304215e-06, 'learning_rate/action_model': 3.156187601530421e-05, 'epoch': 1.26}) +[2026-06-23 21:21:05,873][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.06128188595175743, 'train/grad_norm_pre_clip': 22.011327743530273, 'timing/data': 0.0008021090179681778, 'timing/model': 1.6906548461411148, 'learning_rate/qwen_vl_interface': 3.1534544493859563e-06, 'learning_rate/action_model': 3.1534544493859566e-05, 'epoch': 1.26}) +[2026-06-23 21:21:09,194][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.09343574941158295, 'train/grad_norm_pre_clip': 20.569564819335938, 'timing/data': 0.0005974939558655024, 'timing/model': 1.6876421829219908, 'learning_rate/qwen_vl_interface': 3.150722159052174e-06, 'learning_rate/action_model': 3.150722159052174e-05, 'epoch': 1.26}) +[2026-06-23 21:21:12,521][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.10359479486942291, 'train/grad_norm_pre_clip': 17.90332794189453, 'timing/data': 0.0007786240894347429, 'timing/model': 1.6911785688716918, 'learning_rate/qwen_vl_interface': 3.147990731652215e-06, 'learning_rate/action_model': 3.147990731652215e-05, 'epoch': 1.26}) +[2026-06-23 21:21:15,844][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.10711225122213364, 'train/grad_norm_pre_clip': 24.958499908447266, 'timing/data': 0.0005887378938496113, 'timing/model': 1.6879940240178257, 'learning_rate/qwen_vl_interface': 3.1452601683088673e-06, 'learning_rate/action_model': 3.145260168308867e-05, 'epoch': 1.26}) +[2026-06-23 21:21:19,170][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.21832914650440216, 'train/grad_norm_pre_clip': 34.312801361083984, 'timing/data': 0.0005918531678617001, 'timing/model': 1.689768907148391, 'learning_rate/qwen_vl_interface': 3.1425304701445608e-06, 'learning_rate/action_model': 3.1425304701445605e-05, 'epoch': 1.26}) +[2026-06-23 21:21:22,496][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.2927229702472687, 'train/grad_norm_pre_clip': 29.38325309753418, 'timing/data': 0.0005963218864053488, 'timing/model': 1.6893291051965207, 'learning_rate/qwen_vl_interface': 3.139801638281372e-06, 'learning_rate/action_model': 3.139801638281372e-05, 'epoch': 1.26}) +[2026-06-23 21:21:25,823][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.13871055841445923, 'train/grad_norm_pre_clip': 24.982011795043945, 'timing/data': 0.0006435310933738947, 'timing/model': 1.6892386500257999, 'learning_rate/qwen_vl_interface': 3.13707367384102e-06, 'learning_rate/action_model': 3.13707367384102e-05, 'epoch': 1.27}) +[2026-06-23 21:21:29,148][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.07734562456607819, 'train/grad_norm_pre_clip': 10.817845344543457, 'timing/data': 0.0006221199873834848, 'timing/model': 1.691086812177673, 'learning_rate/qwen_vl_interface': 3.134346577944869e-06, 'learning_rate/action_model': 3.134346577944869e-05, 'epoch': 1.27}) +[2026-06-23 21:21:32,474][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.10509326308965683, 'train/grad_norm_pre_clip': 59.44778823852539, 'timing/data': 0.0006121508777141571, 'timing/model': 1.6892564638983458, 'learning_rate/qwen_vl_interface': 3.1316203517139243e-06, 'learning_rate/action_model': 3.131620351713924e-05, 'epoch': 1.27}) +[2026-06-23 21:21:35,797][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.09660699218511581, 'train/grad_norm_pre_clip': 20.514963150024414, 'timing/data': 0.0005903909914195538, 'timing/model': 1.6840977659448981, 'learning_rate/qwen_vl_interface': 3.128894996268837e-06, 'learning_rate/action_model': 3.128894996268837e-05, 'epoch': 1.27}) +[2026-06-23 21:21:39,134][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.1030673235654831, 'train/grad_norm_pre_clip': 21.90696907043457, 'timing/data': 0.000592394033446908, 'timing/model': 1.6923799910582602, 'learning_rate/qwen_vl_interface': 3.1261705127298934e-06, 'learning_rate/action_model': 3.1261705127298935e-05, 'epoch': 1.27}) +[2026-06-23 21:21:42,458][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.06552387028932571, 'train/grad_norm_pre_clip': 17.273712158203125, 'timing/data': 0.0006030439399182796, 'timing/model': 1.6837156182155013, 'learning_rate/qwen_vl_interface': 3.123446902217028e-06, 'learning_rate/action_model': 3.1234469022170283e-05, 'epoch': 1.27}) +[2026-06-23 21:21:45,790][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.14484044909477234, 'train/grad_norm_pre_clip': 24.927505493164062, 'timing/data': 0.0006827041506767273, 'timing/model': 1.695960345910862, 'learning_rate/qwen_vl_interface': 3.1207241658498145e-06, 'learning_rate/action_model': 3.1207241658498145e-05, 'epoch': 1.27}) +[2026-06-23 21:21:49,427][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.18437916040420532, 'train/grad_norm_pre_clip': 49.852054595947266, 'timing/data': 0.0008288491517305374, 'timing/model': 1.6932080029509962, 'learning_rate/qwen_vl_interface': 3.1180023047474667e-06, 'learning_rate/action_model': 3.1180023047474664e-05, 'epoch': 1.27}) +[2026-06-23 21:21:52,753][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.13498559594154358, 'train/grad_norm_pre_clip': 64.63475036621094, 'timing/data': 0.0006451529916375875, 'timing/model': 1.6911453290376812, 'learning_rate/qwen_vl_interface': 3.1152813200288386e-06, 'learning_rate/action_model': 3.115281320028838e-05, 'epoch': 1.27}) +[2026-06-23 21:21:56,075][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.030521811917424202, 'train/grad_norm_pre_clip': 28.98148536682129, 'timing/data': 0.0006524671334773302, 'timing/model': 1.6865975609980524, 'learning_rate/qwen_vl_interface': 3.1125612128124244e-06, 'learning_rate/action_model': 3.112561212812424e-05, 'epoch': 1.27}) +[2026-06-23 21:21:59,394][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.13033339381217957, 'train/grad_norm_pre_clip': 50.81602096557617, 'timing/data': 0.0006041368469595909, 'timing/model': 1.6880425550043583, 'learning_rate/qwen_vl_interface': 3.1098419842163573e-06, 'learning_rate/action_model': 3.109841984216357e-05, 'epoch': 1.27}) +[2026-06-23 21:22:02,715][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.08503808826208115, 'train/grad_norm_pre_clip': 24.582042694091797, 'timing/data': 0.000577226048335433, 'timing/model': 1.6863720989786088, 'learning_rate/qwen_vl_interface': 3.10712363535841e-06, 'learning_rate/action_model': 3.10712363535841e-05, 'epoch': 1.27}) +[2026-06-23 21:22:06,048][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.06067953631281853, 'train/grad_norm_pre_clip': 31.420520782470703, 'timing/data': 0.0006016620900481939, 'timing/model': 1.6941878970246762, 'learning_rate/qwen_vl_interface': 3.1044061673559926e-06, 'learning_rate/action_model': 3.1044061673559925e-05, 'epoch': 1.27}) +[2026-06-23 21:22:09,382][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.05027031525969505, 'train/grad_norm_pre_clip': 26.156415939331055, 'timing/data': 0.0006918208673596382, 'timing/model': 1.6943830091040581, 'learning_rate/qwen_vl_interface': 3.1016895813261556e-06, 'learning_rate/action_model': 3.101689581326155e-05, 'epoch': 1.27}) +[2026-06-23 21:22:12,713][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.10099875926971436, 'train/grad_norm_pre_clip': 45.4334716796875, 'timing/data': 0.0007235209923237562, 'timing/model': 1.6910093200858682, 'learning_rate/qwen_vl_interface': 3.098973878385581e-06, 'learning_rate/action_model': 3.098973878385581e-05, 'epoch': 1.27}) +[2026-06-23 21:22:16,051][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.1221531480550766, 'train/grad_norm_pre_clip': 35.08056640625, 'timing/data': 0.0005808928981423378, 'timing/model': 1.6956044500693679, 'learning_rate/qwen_vl_interface': 3.096259059650594e-06, 'learning_rate/action_model': 3.096259059650594e-05, 'epoch': 1.27}) +[2026-06-23 21:22:19,384][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.09935466945171356, 'train/grad_norm_pre_clip': 11.951403617858887, 'timing/data': 0.000573648838326335, 'timing/model': 1.6924272698815912, 'learning_rate/qwen_vl_interface': 3.093545126237153e-06, 'learning_rate/action_model': 3.0935451262371527e-05, 'epoch': 1.27}) +[2026-06-23 21:22:22,707][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.06940227001905441, 'train/grad_norm_pre_clip': 18.351268768310547, 'timing/data': 0.0009117349982261658, 'timing/model': 1.691756886197254, 'learning_rate/qwen_vl_interface': 3.090832079260855e-06, 'learning_rate/action_model': 3.0908320792608544e-05, 'epoch': 1.27}) +[2026-06-23 21:22:26,042][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.12584903836250305, 'train/grad_norm_pre_clip': 25.69228744506836, 'timing/data': 0.0005944781005382538, 'timing/model': 1.6936285130213946, 'learning_rate/qwen_vl_interface': 3.088119919836929e-06, 'learning_rate/action_model': 3.088119919836929e-05, 'epoch': 1.27}) +[2026-06-23 21:22:29,373][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.10003452003002167, 'train/grad_norm_pre_clip': 24.075096130371094, 'timing/data': 0.0007160969544202089, 'timing/model': 1.6905009269248694, 'learning_rate/qwen_vl_interface': 3.0854086490802433e-06, 'learning_rate/action_model': 3.085408649080243e-05, 'epoch': 1.27}) +[2026-06-23 21:22:32,704][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.12213697284460068, 'train/grad_norm_pre_clip': 31.674131393432617, 'timing/data': 0.000601199921220541, 'timing/model': 1.6913152560591698, 'learning_rate/qwen_vl_interface': 3.082698268105299e-06, 'learning_rate/action_model': 3.082698268105299e-05, 'epoch': 1.27}) +[2026-06-23 21:22:36,189][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.1651032418012619, 'train/grad_norm_pre_clip': 21.57218360900879, 'timing/data': 0.0006750489119440317, 'timing/model': 1.6900615938939154, 'learning_rate/qwen_vl_interface': 3.0799887780262295e-06, 'learning_rate/action_model': 3.0799887780262296e-05, 'epoch': 1.27}) +[2026-06-23 21:22:39,545][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.04006665572524071, 'train/grad_norm_pre_clip': 29.023038864135742, 'timing/data': 0.0009225150570273399, 'timing/model': 1.7021340178325772, 'learning_rate/qwen_vl_interface': 3.0772801799568054e-06, 'learning_rate/action_model': 3.0772801799568054e-05, 'epoch': 1.27}) +[2026-06-23 21:22:42,878][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.1089477688074112, 'train/grad_norm_pre_clip': 20.896240234375, 'timing/data': 0.0008363830856978893, 'timing/model': 1.694446356035769, 'learning_rate/qwen_vl_interface': 3.07457247501043e-06, 'learning_rate/action_model': 3.0745724750104296e-05, 'epoch': 1.27}) +[2026-06-23 21:22:46,209][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.09457946568727493, 'train/grad_norm_pre_clip': 30.541744232177734, 'timing/data': 0.000773505074903369, 'timing/model': 1.6918392188381404, 'learning_rate/qwen_vl_interface': 3.0718656643001353e-06, 'learning_rate/action_model': 3.071865664300135e-05, 'epoch': 1.27}) +[2026-06-23 21:22:49,543][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.11730063706636429, 'train/grad_norm_pre_clip': 40.70737075805664, 'timing/data': 0.0005845599807798862, 'timing/model': 1.6998579481150955, 'learning_rate/qwen_vl_interface': 3.069159748938591e-06, 'learning_rate/action_model': 3.069159748938591e-05, 'epoch': 1.27}) +[2026-06-23 21:22:52,883][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.2492758184671402, 'train/grad_norm_pre_clip': 77.33429718017578, 'timing/data': 0.0005986460018903017, 'timing/model': 1.693449212005362, 'learning_rate/qwen_vl_interface': 3.066454730038096e-06, 'learning_rate/action_model': 3.0664547300380956e-05, 'epoch': 1.28}) +[2026-06-23 21:22:56,220][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.1289438158273697, 'train/grad_norm_pre_clip': 26.34910774230957, 'timing/data': 0.0005909418687224388, 'timing/model': 1.6935783619992435, 'learning_rate/qwen_vl_interface': 3.0637506087105816e-06, 'learning_rate/action_model': 3.0637506087105815e-05, 'epoch': 1.28}) +[2026-06-23 21:22:59,546][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.10683303326368332, 'train/grad_norm_pre_clip': 16.100231170654297, 'timing/data': 0.0005904901772737503, 'timing/model': 1.6918108949903399, 'learning_rate/qwen_vl_interface': 3.0610473860676103e-06, 'learning_rate/action_model': 3.06104738606761e-05, 'epoch': 1.28}) +[2026-06-23 21:23:02,878][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.07529382407665253, 'train/grad_norm_pre_clip': 18.711454391479492, 'timing/data': 0.000598787097260356, 'timing/model': 1.690945945912972, 'learning_rate/qwen_vl_interface': 3.0583450632203748e-06, 'learning_rate/action_model': 3.058345063220375e-05, 'epoch': 1.28}) +[2026-06-23 21:23:06,204][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.10381133109331131, 'train/grad_norm_pre_clip': 18.54932403564453, 'timing/data': 0.000604826956987381, 'timing/model': 1.6880116551183164, 'learning_rate/qwen_vl_interface': 3.055643641279699e-06, 'learning_rate/action_model': 3.055643641279699e-05, 'epoch': 1.28}) +[2026-06-23 21:23:09,539][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3300 +[2026-06-23 21:24:42,584][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3300 +[2026-06-23 21:24:42,584][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3300 +[2026-06-23 21:25:54,368][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3300 +[2026-06-23 21:25:54,369][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.11321227252483368, 'train/grad_norm_pre_clip': 14.785456657409668, 'eval/action_loss/seconds': 93.04215317289345, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06660762429237366, 'eval/latency_2/loss': 0.06660762429237366, 'eval/flappy/loss': 0.06660762429237366, 'eval/flappy/latency_2/loss': 0.06660762429237366, 'eval/flap_precision': 0.9363270777479893, 'eval/flap_recall': 0.9426450742240216, 'eval/flap_f1': 0.9394754539340954, 'eval/flap_rate_model': 0.20722222222222222, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9394754539340953, 'eval/latency_2/flap_precision': 0.9363270777479893, 'eval/latency_2/flap_recall': 0.9426450742240216, 'eval/latency_2/flap_f1': 0.9394754539340954, 'eval/latency_2/flap_rate_model': 0.20722222222222222, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9394754539340953, 'eval/action_classification/data_wait_seconds': 0.876843273639679, 'eval/action_classification/predict_seconds': 70.7754898071289, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006051480304449797, 'timing/model': 1.6943033060524613, 'learning_rate/qwen_vl_interface': 3.0529431213560334e-06, 'learning_rate/action_model': 3.0529431213560334e-05, 'epoch': 1.28}) +[2026-06-23 21:25:57,693][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.11923836916685104, 'train/grad_norm_pre_clip': 13.510711669921875, 'timing/data': 0.0005948278121650219, 'timing/model': 1.6899550748057663, 'learning_rate/qwen_vl_interface': 3.0502435045594596e-06, 'learning_rate/action_model': 3.0502435045594595e-05, 'epoch': 1.28}) +[2026-06-23 21:26:01,016][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.1606026589870453, 'train/grad_norm_pre_clip': 24.724117279052734, 'timing/data': 0.0008360119536519051, 'timing/model': 1.69052800280042, 'learning_rate/qwen_vl_interface': 3.047544791999689e-06, 'learning_rate/action_model': 3.0475447919996892e-05, 'epoch': 1.28}) +[2026-06-23 21:26:04,345][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.07149790972471237, 'train/grad_norm_pre_clip': 10.838298797607422, 'timing/data': 0.0009105531498789787, 'timing/model': 1.6895998700056225, 'learning_rate/qwen_vl_interface': 3.044846984786061e-06, 'learning_rate/action_model': 3.044846984786061e-05, 'epoch': 1.28}) +[2026-06-23 21:26:07,670][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.0925191193819046, 'train/grad_norm_pre_clip': 31.62137222290039, 'timing/data': 0.0005984748713672161, 'timing/model': 1.6901702429167926, 'learning_rate/qwen_vl_interface': 3.0421500840275413e-06, 'learning_rate/action_model': 3.0421500840275413e-05, 'epoch': 1.28}) +[2026-06-23 21:26:10,997][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.13922829926013947, 'train/grad_norm_pre_clip': 32.79067611694336, 'timing/data': 0.0005867541767656803, 'timing/model': 1.6893565249629319, 'learning_rate/qwen_vl_interface': 3.039454090832727e-06, 'learning_rate/action_model': 3.0394540908327268e-05, 'epoch': 1.28}) +[2026-06-23 21:26:14,324][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.11382105201482773, 'train/grad_norm_pre_clip': 35.1000862121582, 'timing/data': 0.0008375251200050116, 'timing/model': 1.690052435034886, 'learning_rate/qwen_vl_interface': 3.0367590063098333e-06, 'learning_rate/action_model': 3.0367590063098333e-05, 'epoch': 1.28}) +[2026-06-23 21:26:17,658][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.16496218740940094, 'train/grad_norm_pre_clip': 41.173763275146484, 'timing/data': 0.0006375389639288187, 'timing/model': 1.6937649080064148, 'learning_rate/qwen_vl_interface': 3.0340648315667115e-06, 'learning_rate/action_model': 3.0340648315667113e-05, 'epoch': 1.28}) +[2026-06-23 21:26:20,987][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.032111626118421555, 'train/grad_norm_pre_clip': 8.78260326385498, 'timing/data': 0.000645314110442996, 'timing/model': 1.689013073919341, 'learning_rate/qwen_vl_interface': 3.0313715677108336e-06, 'learning_rate/action_model': 3.0313715677108335e-05, 'epoch': 1.28}) +[2026-06-23 21:26:24,317][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.043443553149700165, 'train/grad_norm_pre_clip': 28.66139030456543, 'timing/data': 0.0005856510251760483, 'timing/model': 1.692211019108072, 'learning_rate/qwen_vl_interface': 3.0286792158492996e-06, 'learning_rate/action_model': 3.0286792158492994e-05, 'epoch': 1.28}) +[2026-06-23 21:26:27,647][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.1688661277294159, 'train/grad_norm_pre_clip': 28.973722457885742, 'timing/data': 0.0006029938813298941, 'timing/model': 1.6892571148928255, 'learning_rate/qwen_vl_interface': 3.0259877770888334e-06, 'learning_rate/action_model': 3.0259877770888333e-05, 'epoch': 1.28}) +[2026-06-23 21:26:30,983][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.11901739239692688, 'train/grad_norm_pre_clip': 95.9170150756836, 'timing/data': 0.0008837617933750153, 'timing/model': 1.69251787988469, 'learning_rate/qwen_vl_interface': 3.0232972525357836e-06, 'learning_rate/action_model': 3.0232972525357834e-05, 'epoch': 1.28}) +[2026-06-23 21:26:34,435][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.13516148924827576, 'train/grad_norm_pre_clip': 45.481224060058594, 'timing/data': 0.0005888768937438726, 'timing/model': 1.8108830088749528, 'learning_rate/qwen_vl_interface': 3.020607643296123e-06, 'learning_rate/action_model': 3.0206076432961226e-05, 'epoch': 1.28}) +[2026-06-23 21:26:37,774][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.14135558903217316, 'train/grad_norm_pre_clip': 15.224740982055664, 'timing/data': 0.0005982050206512213, 'timing/model': 1.6941653010435402, 'learning_rate/qwen_vl_interface': 3.017918950475449e-06, 'learning_rate/action_model': 3.017918950475449e-05, 'epoch': 1.28}) +[2026-06-23 21:26:41,107][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.08962429314851761, 'train/grad_norm_pre_clip': 17.779117584228516, 'timing/data': 0.0006954770069569349, 'timing/model': 1.69254185911268, 'learning_rate/qwen_vl_interface': 3.0152311751789815e-06, 'learning_rate/action_model': 3.0152311751789814e-05, 'epoch': 1.28}) +[2026-06-23 21:26:44,446][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.12248986214399338, 'train/grad_norm_pre_clip': 40.039066314697266, 'timing/data': 0.0008532539941370487, 'timing/model': 1.6972516977693886, 'learning_rate/qwen_vl_interface': 3.0125443185115655e-06, 'learning_rate/action_model': 3.0125443185115655e-05, 'epoch': 1.28}) +[2026-06-23 21:26:47,779][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.22798430919647217, 'train/grad_norm_pre_clip': 31.77834701538086, 'timing/data': 0.0005912010092288256, 'timing/model': 1.6861658939160407, 'learning_rate/qwen_vl_interface': 3.0098583815776615e-06, 'learning_rate/action_model': 3.0098583815776615e-05, 'epoch': 1.28}) +[2026-06-23 21:26:51,117][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.056110840290784836, 'train/grad_norm_pre_clip': 42.11000442504883, 'timing/data': 0.000683656195178628, 'timing/model': 1.6938928859308362, 'learning_rate/qwen_vl_interface': 3.0071733654813595e-06, 'learning_rate/action_model': 3.0071733654813595e-05, 'epoch': 1.28}) +[2026-06-23 21:26:54,454][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.11533655971288681, 'train/grad_norm_pre_clip': 28.560882568359375, 'timing/data': 0.0006515251006931067, 'timing/model': 1.6960672670975327, 'learning_rate/qwen_vl_interface': 3.0044892713263685e-06, 'learning_rate/action_model': 3.0044892713263682e-05, 'epoch': 1.28}) +[2026-06-23 21:26:57,796][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.13915760815143585, 'train/grad_norm_pre_clip': 26.819551467895508, 'timing/data': 0.0006095659919083118, 'timing/model': 1.6993785458616912, 'learning_rate/qwen_vl_interface': 3.001806100216017e-06, 'learning_rate/action_model': 3.001806100216017e-05, 'epoch': 1.28}) +[2026-06-23 21:27:01,134][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.09206519275903702, 'train/grad_norm_pre_clip': 20.850709915161133, 'timing/data': 0.0006097359582781792, 'timing/model': 1.6951488240156323, 'learning_rate/qwen_vl_interface': 2.9991238532532567e-06, 'learning_rate/action_model': 2.9991238532532567e-05, 'epoch': 1.28}) +[2026-06-23 21:27:04,508][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.08378130197525024, 'train/grad_norm_pre_clip': 15.38054084777832, 'timing/data': 0.0006093659903854132, 'timing/model': 1.7307435988914222, 'learning_rate/qwen_vl_interface': 2.996442531540657e-06, 'learning_rate/action_model': 2.9964425315406568e-05, 'epoch': 1.29}) +[2026-06-23 21:27:07,848][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.13163158297538757, 'train/grad_norm_pre_clip': 46.63578414916992, 'timing/data': 0.000574049074202776, 'timing/model': 1.6988877900876105, 'learning_rate/qwen_vl_interface': 2.993762136180408e-06, 'learning_rate/action_model': 2.993762136180408e-05, 'epoch': 1.29}) +[2026-06-23 21:27:11,184][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.05979519709944725, 'train/grad_norm_pre_clip': 12.97889518737793, 'timing/data': 0.000598324928432703, 'timing/model': 1.695391105953604, 'learning_rate/qwen_vl_interface': 2.9910826682743195e-06, 'learning_rate/action_model': 2.9910826682743192e-05, 'epoch': 1.29}) +[2026-06-23 21:27:14,520][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.30883240699768066, 'train/grad_norm_pre_clip': 62.07303237915039, 'timing/data': 0.0005863129626959562, 'timing/model': 1.6949778320267797, 'learning_rate/qwen_vl_interface': 2.9884041289238192e-06, 'learning_rate/action_model': 2.988404128923819e-05, 'epoch': 1.29}) +[2026-06-23 21:27:17,865][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.12949399650096893, 'train/grad_norm_pre_clip': 17.619518280029297, 'timing/data': 0.0007577049545943737, 'timing/model': 1.701947971014306, 'learning_rate/qwen_vl_interface': 2.9857265192299557e-06, 'learning_rate/action_model': 2.9857265192299556e-05, 'epoch': 1.29}) +[2026-06-23 21:27:21,201][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.08958622068166733, 'train/grad_norm_pre_clip': 29.27118492126465, 'timing/data': 0.0006473669782280922, 'timing/model': 1.6978728608228266, 'learning_rate/qwen_vl_interface': 2.9830498402933894e-06, 'learning_rate/action_model': 2.9830498402933894e-05, 'epoch': 1.29}) +[2026-06-23 21:27:24,825][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.1163896843791008, 'train/grad_norm_pre_clip': 47.295780181884766, 'timing/data': 0.0007642868440598249, 'timing/model': 1.981766166165471, 'learning_rate/qwen_vl_interface': 2.9803740932144025e-06, 'learning_rate/action_model': 2.9803740932144024e-05, 'epoch': 1.29}) +[2026-06-23 21:27:28,152][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.13668256998062134, 'train/grad_norm_pre_clip': 54.35285949707031, 'timing/data': 0.0006116998847573996, 'timing/model': 1.6914670700207353, 'learning_rate/qwen_vl_interface': 2.9776992790928953e-06, 'learning_rate/action_model': 2.9776992790928952e-05, 'epoch': 1.29}) +[2026-06-23 21:27:31,499][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.15363189578056335, 'train/grad_norm_pre_clip': 33.3349494934082, 'timing/data': 0.0007993530016392469, 'timing/model': 1.7026119788642973, 'learning_rate/qwen_vl_interface': 2.975025399028382e-06, 'learning_rate/action_model': 2.975025399028382e-05, 'epoch': 1.29}) +[2026-06-23 21:27:34,835][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.044985461980104446, 'train/grad_norm_pre_clip': 46.87656784057617, 'timing/data': 0.0008170059882104397, 'timing/model': 1.6945941848680377, 'learning_rate/qwen_vl_interface': 2.9723524541199943e-06, 'learning_rate/action_model': 2.972352454119994e-05, 'epoch': 1.29}) +[2026-06-23 21:27:38,174][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.0633387416601181, 'train/grad_norm_pre_clip': 24.53875732421875, 'timing/data': 0.0006357249803841114, 'timing/model': 1.6974275060929358, 'learning_rate/qwen_vl_interface': 2.969680445466478e-06, 'learning_rate/action_model': 2.969680445466478e-05, 'epoch': 1.29}) +[2026-06-23 21:27:41,509][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.1462208330631256, 'train/grad_norm_pre_clip': 40.91690444946289, 'timing/data': 0.0005811629816889763, 'timing/model': 1.6918836538679898, 'learning_rate/qwen_vl_interface': 2.9670093741661955e-06, 'learning_rate/action_model': 2.9670093741661954e-05, 'epoch': 1.29}) +[2026-06-23 21:27:44,849][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.10006582736968994, 'train/grad_norm_pre_clip': 19.628175735473633, 'timing/data': 0.0005946080200374126, 'timing/model': 1.6939492819365114, 'learning_rate/qwen_vl_interface': 2.964339241317123e-06, 'learning_rate/action_model': 2.964339241317123e-05, 'epoch': 1.29}) +[2026-06-23 21:27:48,194][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.12380941212177277, 'train/grad_norm_pre_clip': 23.508506774902344, 'timing/data': 0.0005920331459492445, 'timing/model': 1.6998387179337442, 'learning_rate/qwen_vl_interface': 2.961670048016852e-06, 'learning_rate/action_model': 2.9616700480168518e-05, 'epoch': 1.29}) +[2026-06-23 21:27:51,530][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.21706734597682953, 'train/grad_norm_pre_clip': 28.291582107543945, 'timing/data': 0.000594268087297678, 'timing/model': 1.6955743730068207, 'learning_rate/qwen_vl_interface': 2.9590017953625854e-06, 'learning_rate/action_model': 2.9590017953625853e-05, 'epoch': 1.29}) +[2026-06-23 21:27:54,868][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.14768663048744202, 'train/grad_norm_pre_clip': 22.4478816986084, 'timing/data': 0.0007746568880975246, 'timing/model': 1.695626774104312, 'learning_rate/qwen_vl_interface': 2.9563344844511455e-06, 'learning_rate/action_model': 2.9563344844511455e-05, 'epoch': 1.29}) +[2026-06-23 21:27:58,205][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.12916290760040283, 'train/grad_norm_pre_clip': 58.87310791015625, 'timing/data': 0.0005935861263424158, 'timing/model': 1.6951286210678518, 'learning_rate/qwen_vl_interface': 2.9536681163789568e-06, 'learning_rate/action_model': 2.9536681163789564e-05, 'epoch': 1.29}) +[2026-06-23 21:28:01,544][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.08100055903196335, 'train/grad_norm_pre_clip': 35.0118408203125, 'timing/data': 0.0006039149593561888, 'timing/model': 1.696640546899289, 'learning_rate/qwen_vl_interface': 2.9510026922420653e-06, 'learning_rate/action_model': 2.9510026922420653e-05, 'epoch': 1.29}) +[2026-06-23 21:28:04,883][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.10830022394657135, 'train/grad_norm_pre_clip': 23.238821029663086, 'timing/data': 0.0008442678954452276, 'timing/model': 1.6952732889913023, 'learning_rate/qwen_vl_interface': 2.9483382131361273e-06, 'learning_rate/action_model': 2.948338213136127e-05, 'epoch': 1.29}) +[2026-06-23 21:28:08,215][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.2062631994485855, 'train/grad_norm_pre_clip': 43.24272537231445, 'timing/data': 0.0006073920521885157, 'timing/model': 1.6918583009392023, 'learning_rate/qwen_vl_interface': 2.9456746801564072e-06, 'learning_rate/action_model': 2.945674680156407e-05, 'epoch': 1.29}) +[2026-06-23 21:28:11,649][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.10344380885362625, 'train/grad_norm_pre_clip': 13.33565616607666, 'timing/data': 0.0005891178734600544, 'timing/model': 1.7945340359583497, 'learning_rate/qwen_vl_interface': 2.9430120943977854e-06, 'learning_rate/action_model': 2.9430120943977853e-05, 'epoch': 1.29}) +[2026-06-23 21:28:14,993][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.13048669695854187, 'train/grad_norm_pre_clip': 32.982784271240234, 'timing/data': 0.000600690022110939, 'timing/model': 1.6971761158201844, 'learning_rate/qwen_vl_interface': 2.940350456954748e-06, 'learning_rate/action_model': 2.940350456954748e-05, 'epoch': 1.29}) +[2026-06-23 21:28:18,331][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.16582828760147095, 'train/grad_norm_pre_clip': 70.37964630126953, 'timing/data': 0.000621238024905324, 'timing/model': 1.6952823279425502, 'learning_rate/qwen_vl_interface': 2.9376897689213952e-06, 'learning_rate/action_model': 2.937689768921395e-05, 'epoch': 1.29}) +[2026-06-23 21:28:21,670][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.09133189916610718, 'train/grad_norm_pre_clip': 54.81558609008789, 'timing/data': 0.0005919630639255047, 'timing/model': 1.6973893730901182, 'learning_rate/qwen_vl_interface': 2.9350300313914364e-06, 'learning_rate/action_model': 2.9350300313914363e-05, 'epoch': 1.29}) +[2026-06-23 21:28:25,007][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.0807197168469429, 'train/grad_norm_pre_clip': 19.48939323425293, 'timing/data': 0.0006907288916409016, 'timing/model': 1.6960746350232512, 'learning_rate/qwen_vl_interface': 2.9323712454581867e-06, 'learning_rate/action_model': 2.9323712454581866e-05, 'epoch': 1.29}) +[2026-06-23 21:28:28,346][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.14053219556808472, 'train/grad_norm_pre_clip': 28.34834861755371, 'timing/data': 0.0005933060310781002, 'timing/model': 1.695891864830628, 'learning_rate/qwen_vl_interface': 2.929713412214572e-06, 'learning_rate/action_model': 2.929713412214572e-05, 'epoch': 1.29}) +[2026-06-23 21:28:31,687][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.06752830743789673, 'train/grad_norm_pre_clip': 15.944045066833496, 'timing/data': 0.0007969681173563004, 'timing/model': 1.6952597999479622, 'learning_rate/qwen_vl_interface': 2.9270565327531294e-06, 'learning_rate/action_model': 2.927056532753129e-05, 'epoch': 1.3}) +[2026-06-23 21:28:35,031][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.08161204308271408, 'train/grad_norm_pre_clip': 20.250076293945312, 'timing/data': 0.0005926638841629028, 'timing/model': 1.6985692859161645, 'learning_rate/qwen_vl_interface': 2.924400608166003e-06, 'learning_rate/action_model': 2.924400608166003e-05, 'epoch': 1.3}) +[2026-06-23 21:28:38,377][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.09625913202762604, 'train/grad_norm_pre_clip': 14.040946960449219, 'timing/data': 0.0006489399820566177, 'timing/model': 1.700310543179512, 'learning_rate/qwen_vl_interface': 2.921745639544939e-06, 'learning_rate/action_model': 2.921745639544939e-05, 'epoch': 1.3}) +[2026-06-23 21:28:41,715][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.023481130599975586, 'train/grad_norm_pre_clip': 67.228515625, 'timing/data': 0.0006416561082005501, 'timing/model': 1.6951811129692942, 'learning_rate/qwen_vl_interface': 2.919091627981297e-06, 'learning_rate/action_model': 2.919091627981297e-05, 'epoch': 1.3}) +[2026-06-23 21:28:45,065][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.04646608233451843, 'train/grad_norm_pre_clip': 48.92653274536133, 'timing/data': 0.0007449400145560503, 'timing/model': 1.7024704010691494, 'learning_rate/qwen_vl_interface': 2.9164385745660396e-06, 'learning_rate/action_model': 2.9164385745660394e-05, 'epoch': 1.3}) +[2026-06-23 21:28:48,398][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.09384844452142715, 'train/grad_norm_pre_clip': 85.42417907714844, 'timing/data': 0.0006941650062799454, 'timing/model': 1.6943854028359056, 'learning_rate/qwen_vl_interface': 2.9137864803897385e-06, 'learning_rate/action_model': 2.9137864803897387e-05, 'epoch': 1.3}) +[2026-06-23 21:28:51,739][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.08301134407520294, 'train/grad_norm_pre_clip': 19.281757354736328, 'timing/data': 0.0006540999747812748, 'timing/model': 1.6965628911275417, 'learning_rate/qwen_vl_interface': 2.9111353465425696e-06, 'learning_rate/action_model': 2.9111353465425694e-05, 'epoch': 1.3}) +[2026-06-23 21:28:55,082][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.1631956249475479, 'train/grad_norm_pre_clip': 25.30096435546875, 'timing/data': 0.0007482168730348349, 'timing/model': 1.6963158210273832, 'learning_rate/qwen_vl_interface': 2.908485174114312e-06, 'learning_rate/action_model': 2.908485174114312e-05, 'epoch': 1.3}) +[2026-06-23 21:28:58,430][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.11751759052276611, 'train/grad_norm_pre_clip': 37.66745376586914, 'timing/data': 0.0006555828731507063, 'timing/model': 1.695973550900817, 'learning_rate/qwen_vl_interface': 2.9058359641943523e-06, 'learning_rate/action_model': 2.9058359641943523e-05, 'epoch': 1.3}) +[2026-06-23 21:29:02,042][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.16699069738388062, 'train/grad_norm_pre_clip': 23.603788375854492, 'timing/data': 0.0006570650730282068, 'timing/model': 1.9722908299881965, 'learning_rate/qwen_vl_interface': 2.9031877178716823e-06, 'learning_rate/action_model': 2.903187717871682e-05, 'epoch': 1.3}) +[2026-06-23 21:29:05,401][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.10029684007167816, 'train/grad_norm_pre_clip': 46.6727409362793, 'timing/data': 0.0008275960572063923, 'timing/model': 1.7020071018487215, 'learning_rate/qwen_vl_interface': 2.900540436234894e-06, 'learning_rate/action_model': 2.9005404362348937e-05, 'epoch': 1.3}) +[2026-06-23 21:29:08,740][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.046228229999542236, 'train/grad_norm_pre_clip': 18.564393997192383, 'timing/data': 0.0006713629700243473, 'timing/model': 1.6953348910901695, 'learning_rate/qwen_vl_interface': 2.897894120372185e-06, 'learning_rate/action_model': 2.897894120372185e-05, 'epoch': 1.3}) +[2026-06-23 21:29:12,088][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.19127748906612396, 'train/grad_norm_pre_clip': 32.64638900756836, 'timing/data': 0.0006100770551711321, 'timing/model': 1.6982849219348282, 'learning_rate/qwen_vl_interface': 2.8952487713713585e-06, 'learning_rate/action_model': 2.8952487713713582e-05, 'epoch': 1.3}) +[2026-06-23 21:29:15,430][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.12332814931869507, 'train/grad_norm_pre_clip': 21.353878021240234, 'timing/data': 0.0007787640206515789, 'timing/model': 1.699254539096728, 'learning_rate/qwen_vl_interface': 2.8926043903198135e-06, 'learning_rate/action_model': 2.8926043903198135e-05, 'epoch': 1.3}) +[2026-06-23 21:29:18,771][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.11807996779680252, 'train/grad_norm_pre_clip': 17.21095085144043, 'timing/data': 0.0006518049631267786, 'timing/model': 1.695480797905475, 'learning_rate/qwen_vl_interface': 2.8899609783045574e-06, 'learning_rate/action_model': 2.8899609783045573e-05, 'epoch': 1.3}) +[2026-06-23 21:29:22,105][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.11521003395318985, 'train/grad_norm_pre_clip': 25.292390823364258, 'timing/data': 0.0006503730546683073, 'timing/model': 1.6949085651431233, 'learning_rate/qwen_vl_interface': 2.8873185364121965e-06, 'learning_rate/action_model': 2.8873185364121963e-05, 'epoch': 1.3}) +[2026-06-23 21:29:25,441][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.09985660016536713, 'train/grad_norm_pre_clip': 14.033268928527832, 'timing/data': 0.0006783457938581705, 'timing/model': 1.6957419170066714, 'learning_rate/qwen_vl_interface': 2.8846770657289386e-06, 'learning_rate/action_model': 2.8846770657289386e-05, 'epoch': 1.3}) +[2026-06-23 21:29:28,773][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.18345054984092712, 'train/grad_norm_pre_clip': 20.246381759643555, 'timing/data': 0.0006374591030180454, 'timing/model': 1.7029290238860995, 'learning_rate/qwen_vl_interface': 2.882036567340593e-06, 'learning_rate/action_model': 2.8820365673405932e-05, 'epoch': 1.3}) +[2026-06-23 21:29:32,106][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.05519841983914375, 'train/grad_norm_pre_clip': 19.83233642578125, 'timing/data': 0.0008080289699137211, 'timing/model': 1.69124072114937, 'learning_rate/qwen_vl_interface': 2.8793970423325684e-06, 'learning_rate/action_model': 2.8793970423325685e-05, 'epoch': 1.3}) +[2026-06-23 21:29:35,440][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.07869038730859756, 'train/grad_norm_pre_clip': 19.8212890625, 'timing/data': 0.0006480079609900713, 'timing/model': 1.6967112720012665, 'learning_rate/qwen_vl_interface': 2.8767584917898756e-06, 'learning_rate/action_model': 2.8767584917898754e-05, 'epoch': 1.3}) +[2026-06-23 21:29:38,783][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.05186547338962555, 'train/grad_norm_pre_clip': 29.400367736816406, 'timing/data': 0.0009098099544644356, 'timing/model': 1.6933462088927627, 'learning_rate/qwen_vl_interface': 2.874120916797121e-06, 'learning_rate/action_model': 2.8741209167971207e-05, 'epoch': 1.3}) +[2026-06-23 21:29:42,122][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.1111610010266304, 'train/grad_norm_pre_clip': 26.73686408996582, 'timing/data': 0.0007960970979183912, 'timing/model': 1.696506085107103, 'learning_rate/qwen_vl_interface': 2.8714843184385124e-06, 'learning_rate/action_model': 2.871484318438512e-05, 'epoch': 1.3}) +[2026-06-23 21:29:45,460][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.11283114552497864, 'train/grad_norm_pre_clip': 41.04865264892578, 'timing/data': 0.000589658971875906, 'timing/model': 1.6940206161234528, 'learning_rate/qwen_vl_interface': 2.8688486977978575e-06, 'learning_rate/action_model': 2.8688486977978572e-05, 'epoch': 1.3}) +[2026-06-23 21:29:48,799][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.07279571890830994, 'train/grad_norm_pre_clip': 12.463409423828125, 'timing/data': 0.0005956699606031179, 'timing/model': 1.6946841180324554, 'learning_rate/qwen_vl_interface': 2.8662140559585584e-06, 'learning_rate/action_model': 2.866214055958558e-05, 'epoch': 1.3}) +[2026-06-23 21:29:52,444][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.06921461224555969, 'train/grad_norm_pre_clip': 12.855087280273438, 'timing/data': 0.0009076967835426331, 'timing/model': 2.0048082468565553, 'learning_rate/qwen_vl_interface': 2.8635803940036162e-06, 'learning_rate/action_model': 2.863580394003616e-05, 'epoch': 1.3}) +[2026-06-23 21:29:55,779][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.08953270316123962, 'train/grad_norm_pre_clip': 21.279050827026367, 'timing/data': 0.0006273288745433092, 'timing/model': 1.6939341959077865, 'learning_rate/qwen_vl_interface': 2.8609477130156317e-06, 'learning_rate/action_model': 2.8609477130156314e-05, 'epoch': 1.31}) +[2026-06-23 21:29:59,120][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.14129871129989624, 'train/grad_norm_pre_clip': 26.3673038482666, 'timing/data': 0.0006021619774401188, 'timing/model': 1.698467673966661, 'learning_rate/qwen_vl_interface': 2.8583160140767997e-06, 'learning_rate/action_model': 2.8583160140767994e-05, 'epoch': 1.31}) +[2026-06-23 21:30:02,462][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.13134850561618805, 'train/grad_norm_pre_clip': 25.625595092773438, 'timing/data': 0.0007640470284968615, 'timing/model': 1.6948063850868493, 'learning_rate/qwen_vl_interface': 2.8556852982689125e-06, 'learning_rate/action_model': 2.8556852982689126e-05, 'epoch': 1.31}) +[2026-06-23 21:30:05,798][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.07162082195281982, 'train/grad_norm_pre_clip': 27.890644073486328, 'timing/data': 0.0006053280085325241, 'timing/model': 1.692264901008457, 'learning_rate/qwen_vl_interface': 2.8530555666733584e-06, 'learning_rate/action_model': 2.8530555666733583e-05, 'epoch': 1.31}) +[2026-06-23 21:30:09,132][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.0836261734366417, 'train/grad_norm_pre_clip': 28.243009567260742, 'timing/data': 0.0006023629102855921, 'timing/model': 1.691511538811028, 'learning_rate/qwen_vl_interface': 2.85042682037112e-06, 'learning_rate/action_model': 2.85042682037112e-05, 'epoch': 1.31}) +[2026-06-23 21:30:12,472][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.13761095702648163, 'train/grad_norm_pre_clip': 21.586320877075195, 'timing/data': 0.0006021829321980476, 'timing/model': 1.6980775690171868, 'learning_rate/qwen_vl_interface': 2.8477990604427764e-06, 'learning_rate/action_model': 2.8477990604427763e-05, 'epoch': 1.31}) +[2026-06-23 21:30:15,805][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.05780353024601936, 'train/grad_norm_pre_clip': 11.365341186523438, 'timing/data': 0.0007969879079610109, 'timing/model': 1.6942227359395474, 'learning_rate/qwen_vl_interface': 2.8451722879684994e-06, 'learning_rate/action_model': 2.8451722879684994e-05, 'epoch': 1.31}) +[2026-06-23 21:30:19,139][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.09739485383033752, 'train/grad_norm_pre_clip': 22.343751907348633, 'timing/data': 0.0006667140405625105, 'timing/model': 1.6941142969299108, 'learning_rate/qwen_vl_interface': 2.842546504028059e-06, 'learning_rate/action_model': 2.8425465040280586e-05, 'epoch': 1.31}) +[2026-06-23 21:30:22,468][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.15424951910972595, 'train/grad_norm_pre_clip': 28.348918914794922, 'timing/data': 0.0005924140568822622, 'timing/model': 1.6901325851213187, 'learning_rate/qwen_vl_interface': 2.83992170970081e-06, 'learning_rate/action_model': 2.8399217097008102e-05, 'epoch': 1.31}) +[2026-06-23 21:30:25,797][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.11214225739240646, 'train/grad_norm_pre_clip': 30.153610229492188, 'timing/data': 0.0007294910028576851, 'timing/model': 1.690415922086686, 'learning_rate/qwen_vl_interface': 2.837297906065711e-06, 'learning_rate/action_model': 2.8372979060657107e-05, 'epoch': 1.31}) +[2026-06-23 21:30:29,136][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.09003596752882004, 'train/grad_norm_pre_clip': 12.941479682922363, 'timing/data': 0.0005995170213282108, 'timing/model': 1.6921985789667815, 'learning_rate/qwen_vl_interface': 2.834675094201306e-06, 'learning_rate/action_model': 2.8346750942013063e-05, 'epoch': 1.31}) +[2026-06-23 21:30:32,478][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.10592491179704666, 'train/grad_norm_pre_clip': 25.207780838012695, 'timing/data': 0.0005904100835323334, 'timing/model': 1.6981449651066214, 'learning_rate/qwen_vl_interface': 2.8320532751857365e-06, 'learning_rate/action_model': 2.8320532751857366e-05, 'epoch': 1.31}) +[2026-06-23 21:30:35,820][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.05587457865476608, 'train/grad_norm_pre_clip': 11.229948043823242, 'timing/data': 0.0005967919714748859, 'timing/model': 1.6944544098805636, 'learning_rate/qwen_vl_interface': 2.829432450096731e-06, 'learning_rate/action_model': 2.8294324500967306e-05, 'epoch': 1.31}) +[2026-06-23 21:30:39,313][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.1509401947259903, 'train/grad_norm_pre_clip': 50.15735626220703, 'timing/data': 0.0008507389575242996, 'timing/model': 1.8455670198891312, 'learning_rate/qwen_vl_interface': 2.8268126200116118e-06, 'learning_rate/action_model': 2.8268126200116117e-05, 'epoch': 1.31}) +[2026-06-23 21:30:42,656][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.12817810475826263, 'train/grad_norm_pre_clip': 48.80119323730469, 'timing/data': 0.0006936339195817709, 'timing/model': 1.6965248929336667, 'learning_rate/qwen_vl_interface': 2.824193786007294e-06, 'learning_rate/action_model': 2.8241937860072937e-05, 'epoch': 1.31}) +[2026-06-23 21:30:45,998][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.09477042406797409, 'train/grad_norm_pre_clip': 32.47795867919922, 'timing/data': 0.0005959710106253624, 'timing/model': 1.696917346213013, 'learning_rate/qwen_vl_interface': 2.821575949160279e-06, 'learning_rate/action_model': 2.8215759491602787e-05, 'epoch': 1.31}) +[2026-06-23 21:30:49,340][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.13680529594421387, 'train/grad_norm_pre_clip': 51.530033111572266, 'timing/data': 0.0005961710121482611, 'timing/model': 1.6959222198929638, 'learning_rate/qwen_vl_interface': 2.8189591105466625e-06, 'learning_rate/action_model': 2.8189591105466624e-05, 'epoch': 1.31}) +[2026-06-23 21:30:52,679][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.06311880052089691, 'train/grad_norm_pre_clip': 42.54863739013672, 'timing/data': 0.0005862221587449312, 'timing/model': 1.6923909650649875, 'learning_rate/qwen_vl_interface': 2.816343271242126e-06, 'learning_rate/action_model': 2.816343271242126e-05, 'epoch': 1.31}) +[2026-06-23 21:30:56,026][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.23878847062587738, 'train/grad_norm_pre_clip': 25.632169723510742, 'timing/data': 0.0006902781315147877, 'timing/model': 1.699945363914594, 'learning_rate/qwen_vl_interface': 2.8137284323219467e-06, 'learning_rate/action_model': 2.8137284323219464e-05, 'epoch': 1.31}) +[2026-06-23 21:30:59,369][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.0980229526758194, 'train/grad_norm_pre_clip': 26.32048225402832, 'timing/data': 0.0005986259784549475, 'timing/model': 1.6973426651675254, 'learning_rate/qwen_vl_interface': 2.81111459486098e-06, 'learning_rate/action_model': 2.81111459486098e-05, 'epoch': 1.31}) +[2026-06-23 21:31:02,713][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.04803852364420891, 'train/grad_norm_pre_clip': 12.079316139221191, 'timing/data': 0.000607331981882453, 'timing/model': 1.6992735911626369, 'learning_rate/qwen_vl_interface': 2.8085017599336806e-06, 'learning_rate/action_model': 2.8085017599336806e-05, 'epoch': 1.31}) +[2026-06-23 21:31:06,059][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.062265794724226, 'train/grad_norm_pre_clip': 15.582797050476074, 'timing/data': 0.0006153869908303022, 'timing/model': 1.7018422458786517, 'learning_rate/qwen_vl_interface': 2.8058899286140828e-06, 'learning_rate/action_model': 2.805889928614083e-05, 'epoch': 1.31}) +[2026-06-23 21:31:09,403][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.19516977667808533, 'train/grad_norm_pre_clip': 62.959754943847656, 'timing/data': 0.0005906298756599426, 'timing/model': 1.697414797032252, 'learning_rate/qwen_vl_interface': 2.8032791019758127e-06, 'learning_rate/action_model': 2.8032791019758127e-05, 'epoch': 1.31}) +[2026-06-23 21:31:12,744][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.12017293274402618, 'train/grad_norm_pre_clip': 56.78200149536133, 'timing/data': 0.0006038350984454155, 'timing/model': 1.7014290690422058, 'learning_rate/qwen_vl_interface': 2.800669281092084e-06, 'learning_rate/action_model': 2.8006692810920838e-05, 'epoch': 1.31}) +[2026-06-23 21:31:16,080][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.030737420544028282, 'train/grad_norm_pre_clip': 42.490234375, 'timing/data': 0.00059477798640728, 'timing/model': 1.6932272540871054, 'learning_rate/qwen_vl_interface': 2.7980604670356938e-06, 'learning_rate/action_model': 2.7980604670356937e-05, 'epoch': 1.31}) +[2026-06-23 21:31:19,420][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.23606379330158234, 'train/grad_norm_pre_clip': 30.50107192993164, 'timing/data': 0.0005941670387983322, 'timing/model': 1.6963796359486878, 'learning_rate/qwen_vl_interface': 2.795452660879027e-06, 'learning_rate/action_model': 2.7954526608790273e-05, 'epoch': 1.31}) +[2026-06-23 21:31:22,761][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.20873641967773438, 'train/grad_norm_pre_clip': 38.369998931884766, 'timing/data': 0.0005915919318795204, 'timing/model': 1.6963379730004817, 'learning_rate/qwen_vl_interface': 2.7928458636940565e-06, 'learning_rate/action_model': 2.7928458636940564e-05, 'epoch': 1.32}) +[2026-06-23 21:31:26,107][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.1821390688419342, 'train/grad_norm_pre_clip': 32.91093063354492, 'timing/data': 0.0006004779133945704, 'timing/model': 1.6991541928146034, 'learning_rate/qwen_vl_interface': 2.790240076552336e-06, 'learning_rate/action_model': 2.7902400765523355e-05, 'epoch': 1.32}) +[2026-06-23 21:31:29,567][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3400 +[2026-06-23 21:33:03,367][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3400 +[2026-06-23 21:33:03,368][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3400 +[2026-06-23 21:34:14,245][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3400 +[2026-06-23 21:34:14,246][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.0732085183262825, 'train/grad_norm_pre_clip': 32.588218688964844, 'eval/action_loss/seconds': 93.79774437611923, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06782340258359909, 'eval/latency_2/loss': 0.06782340258359909, 'eval/flappy/loss': 0.06782340258359909, 'eval/flappy/latency_2/loss': 0.06782340258359909, 'eval/flap_precision': 0.9500346981263011, 'eval/flap_recall': 0.9237516869095816, 'eval/flap_f1': 0.9367088607594938, 'eval/flap_rate_model': 0.2001388888888889, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9367088607594937, 'eval/latency_2/flap_precision': 0.9500346981263011, 'eval/latency_2/flap_recall': 0.9237516869095816, 'eval/latency_2/flap_f1': 0.9367088607594938, 'eval/latency_2/flap_rate_model': 0.2001388888888889, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9367088607594937, 'eval/action_classification/data_wait_seconds': 0.8862079381942749, 'eval/action_classification/predict_seconds': 69.8530044555664, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006610630080103874, 'timing/model': 1.6938990789931267, 'learning_rate/qwen_vl_interface': 2.787635300525006e-06, 'learning_rate/action_model': 2.787635300525006e-05, 'epoch': 1.32}) +[2026-06-23 21:34:17,568][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.03441121429204941, 'train/grad_norm_pre_clip': 16.061365127563477, 'timing/data': 0.0008099130354821682, 'timing/model': 1.6854929281398654, 'learning_rate/qwen_vl_interface': 2.7850315366827933e-06, 'learning_rate/action_model': 2.7850315366827934e-05, 'epoch': 1.32}) +[2026-06-23 21:34:20,886][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.09061621874570847, 'train/grad_norm_pre_clip': 30.44788932800293, 'timing/data': 0.000583296176046133, 'timing/model': 1.6875277389772236, 'learning_rate/qwen_vl_interface': 2.7824287860960076e-06, 'learning_rate/action_model': 2.7824287860960075e-05, 'epoch': 1.32}) +[2026-06-23 21:34:24,211][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.13035325706005096, 'train/grad_norm_pre_clip': 25.104154586791992, 'timing/data': 0.0005822740495204926, 'timing/model': 1.6903891859110445, 'learning_rate/qwen_vl_interface': 2.7798270498345377e-06, 'learning_rate/action_model': 2.7798270498345376e-05, 'epoch': 1.32}) +[2026-06-23 21:34:27,525][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.09303918480873108, 'train/grad_norm_pre_clip': 13.539728164672852, 'timing/data': 0.0005927139427512884, 'timing/model': 1.6774035000707954, 'learning_rate/qwen_vl_interface': 2.777226328967861e-06, 'learning_rate/action_model': 2.7772263289678607e-05, 'epoch': 1.32}) +[2026-06-23 21:34:30,845][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.12304335832595825, 'train/grad_norm_pre_clip': 14.617997169494629, 'timing/data': 0.0005904198624193668, 'timing/model': 1.6853572300169617, 'learning_rate/qwen_vl_interface': 2.774626624565035e-06, 'learning_rate/action_model': 2.774626624565035e-05, 'epoch': 1.32}) +[2026-06-23 21:34:34,166][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.08235307782888412, 'train/grad_norm_pre_clip': 21.783096313476562, 'timing/data': 0.0006409250199794769, 'timing/model': 1.689152010018006, 'learning_rate/qwen_vl_interface': 2.772027937694701e-06, 'learning_rate/action_model': 2.772027937694701e-05, 'epoch': 1.32}) +[2026-06-23 21:34:37,783][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.07430429756641388, 'train/grad_norm_pre_clip': 31.414264678955078, 'timing/data': 0.000706138089299202, 'timing/model': 1.9852693730499595, 'learning_rate/qwen_vl_interface': 2.7694302694250806e-06, 'learning_rate/action_model': 2.7694302694250808e-05, 'epoch': 1.32}) +[2026-06-23 21:34:41,109][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.10506626218557358, 'train/grad_norm_pre_clip': 10.88129997253418, 'timing/data': 0.0006094861309975386, 'timing/model': 1.6892168261110783, 'learning_rate/qwen_vl_interface': 2.7668336208239775e-06, 'learning_rate/action_model': 2.7668336208239776e-05, 'epoch': 1.32}) +[2026-06-23 21:34:44,422][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.04099705442786217, 'train/grad_norm_pre_clip': 12.988472938537598, 'timing/data': 0.0005880361422896385, 'timing/model': 1.6785446088761091, 'learning_rate/qwen_vl_interface': 2.7642379929587755e-06, 'learning_rate/action_model': 2.7642379929587754e-05, 'epoch': 1.32}) +[2026-06-23 21:34:47,738][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.1696501225233078, 'train/grad_norm_pre_clip': 21.487140655517578, 'timing/data': 0.0005970129277557135, 'timing/model': 1.6878369750920683, 'learning_rate/qwen_vl_interface': 2.7616433868964387e-06, 'learning_rate/action_model': 2.7616433868964388e-05, 'epoch': 1.32}) +[2026-06-23 21:34:51,059][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.06643583625555038, 'train/grad_norm_pre_clip': 20.303285598754883, 'timing/data': 0.0005945370066910982, 'timing/model': 1.6872826698236167, 'learning_rate/qwen_vl_interface': 2.7590498037035122e-06, 'learning_rate/action_model': 2.7590498037035124e-05, 'epoch': 1.32}) +[2026-06-23 21:34:54,385][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.0528780072927475, 'train/grad_norm_pre_clip': 11.042990684509277, 'timing/data': 0.0005977139808237553, 'timing/model': 1.688193439040333, 'learning_rate/qwen_vl_interface': 2.7564572444461218e-06, 'learning_rate/action_model': 2.7564572444461217e-05, 'epoch': 1.32}) +[2026-06-23 21:34:57,707][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.14876927435398102, 'train/grad_norm_pre_clip': 77.81346130371094, 'timing/data': 0.0005932650528848171, 'timing/model': 1.6880065340083092, 'learning_rate/qwen_vl_interface': 2.7538657101899674e-06, 'learning_rate/action_model': 2.7538657101899673e-05, 'epoch': 1.32}) +[2026-06-23 21:35:01,033][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.20612958073616028, 'train/grad_norm_pre_clip': 12.9816255569458, 'timing/data': 0.0006003580056130886, 'timing/model': 1.6898173671215773, 'learning_rate/qwen_vl_interface': 2.751275202000332e-06, 'learning_rate/action_model': 2.7512752020003317e-05, 'epoch': 1.32}) +[2026-06-23 21:35:04,376][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.06366663426160812, 'train/grad_norm_pre_clip': 24.345449447631836, 'timing/data': 0.0008275760337710381, 'timing/model': 1.7050390092190355, 'learning_rate/qwen_vl_interface': 2.7486857209420763e-06, 'learning_rate/action_model': 2.7486857209420763e-05, 'epoch': 1.32}) +[2026-06-23 21:35:07,707][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.11562969535589218, 'train/grad_norm_pre_clip': 42.17665481567383, 'timing/data': 0.0006043261382728815, 'timing/model': 1.6898981919512153, 'learning_rate/qwen_vl_interface': 2.7460972680796377e-06, 'learning_rate/action_model': 2.7460972680796375e-05, 'epoch': 1.32}) +[2026-06-23 21:35:11,038][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.22831964492797852, 'train/grad_norm_pre_clip': 32.17906951904297, 'timing/data': 0.0006488999351859093, 'timing/model': 1.6918314329814166, 'learning_rate/qwen_vl_interface': 2.7435098444770314e-06, 'learning_rate/action_model': 2.7435098444770315e-05, 'epoch': 1.32}) +[2026-06-23 21:35:14,367][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.14952987432479858, 'train/grad_norm_pre_clip': 35.8045654296875, 'timing/data': 0.0005877150688320398, 'timing/model': 1.6897287878673524, 'learning_rate/qwen_vl_interface': 2.7409234511978495e-06, 'learning_rate/action_model': 2.7409234511978492e-05, 'epoch': 1.32}) +[2026-06-23 21:35:17,696][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.043327391147613525, 'train/grad_norm_pre_clip': 18.76616096496582, 'timing/data': 0.0005787981208413839, 'timing/model': 1.690429852111265, 'learning_rate/qwen_vl_interface': 2.7383380893052606e-06, 'learning_rate/action_model': 2.7383380893052608e-05, 'epoch': 1.32}) +[2026-06-23 21:35:21,031][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.07519908994436264, 'train/grad_norm_pre_clip': 27.465927124023438, 'timing/data': 0.0005985649768263102, 'timing/model': 1.6966755620669574, 'learning_rate/qwen_vl_interface': 2.73575375986201e-06, 'learning_rate/action_model': 2.7357537598620096e-05, 'epoch': 1.32}) +[2026-06-23 21:35:24,361][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.08166284114122391, 'train/grad_norm_pre_clip': 24.913328170776367, 'timing/data': 0.0005811329465359449, 'timing/model': 1.6915596411563456, 'learning_rate/qwen_vl_interface': 2.733170463930417e-06, 'learning_rate/action_model': 2.7331704639304167e-05, 'epoch': 1.32}) +[2026-06-23 21:35:27,989][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.11716092377901077, 'train/grad_norm_pre_clip': 37.189048767089844, 'timing/data': 0.0005875839851796627, 'timing/model': 1.9904272460844368, 'learning_rate/qwen_vl_interface': 2.7305882025723774e-06, 'learning_rate/action_model': 2.7305882025723774e-05, 'epoch': 1.32}) +[2026-06-23 21:35:31,315][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.09290215373039246, 'train/grad_norm_pre_clip': 8.213872909545898, 'timing/data': 0.0006000578869134188, 'timing/model': 1.6841481011360884, 'learning_rate/qwen_vl_interface': 2.7280069768493626e-06, 'learning_rate/action_model': 2.7280069768493628e-05, 'epoch': 1.32}) +[2026-06-23 21:35:34,638][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.10915842652320862, 'train/grad_norm_pre_clip': 18.816511154174805, 'timing/data': 0.000607130816206336, 'timing/model': 1.691053920891136, 'learning_rate/qwen_vl_interface': 2.7254267878224144e-06, 'learning_rate/action_model': 2.7254267878224143e-05, 'epoch': 1.33}) +[2026-06-23 21:35:37,971][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.08502410352230072, 'train/grad_norm_pre_clip': 19.496187210083008, 'timing/data': 0.0005938571412116289, 'timing/model': 1.6936607288662344, 'learning_rate/qwen_vl_interface': 2.7228476365521524e-06, 'learning_rate/action_model': 2.722847636552152e-05, 'epoch': 1.33}) +[2026-06-23 21:35:41,306][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.11317139863967896, 'train/grad_norm_pre_clip': 66.60321807861328, 'timing/data': 0.0007029608823359013, 'timing/model': 1.696257347939536, 'learning_rate/qwen_vl_interface': 2.7202695240987674e-06, 'learning_rate/action_model': 2.7202695240987676e-05, 'epoch': 1.33}) +[2026-06-23 21:35:44,638][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.08043824881315231, 'train/grad_norm_pre_clip': 22.71194839477539, 'timing/data': 0.0006128419190645218, 'timing/model': 1.6910375619772822, 'learning_rate/qwen_vl_interface': 2.717692451522025e-06, 'learning_rate/action_model': 2.7176924515220248e-05, 'epoch': 1.33}) +[2026-06-23 21:35:47,973][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.08033150434494019, 'train/grad_norm_pre_clip': 14.179500579833984, 'timing/data': 0.000584278954192996, 'timing/model': 1.6938822269439697, 'learning_rate/qwen_vl_interface': 2.7151164198812624e-06, 'learning_rate/action_model': 2.7151164198812622e-05, 'epoch': 1.33}) +[2026-06-23 21:35:51,300][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.21637417376041412, 'train/grad_norm_pre_clip': 42.54732894897461, 'timing/data': 0.000610437011346221, 'timing/model': 1.6930656600743532, 'learning_rate/qwen_vl_interface': 2.712541430235388e-06, 'learning_rate/action_model': 2.712541430235388e-05, 'epoch': 1.33}) +[2026-06-23 21:35:54,627][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.0519571453332901, 'train/grad_norm_pre_clip': 42.31631088256836, 'timing/data': 0.00065183499827981, 'timing/model': 1.6897176620550454, 'learning_rate/qwen_vl_interface': 2.7099674836428842e-06, 'learning_rate/action_model': 2.709967483642884e-05, 'epoch': 1.33}) +[2026-06-23 21:35:57,967][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.05239042639732361, 'train/grad_norm_pre_clip': 9.028943061828613, 'timing/data': 0.0007067278493195772, 'timing/model': 1.694439475890249, 'learning_rate/qwen_vl_interface': 2.7073945811618018e-06, 'learning_rate/action_model': 2.7073945811618017e-05, 'epoch': 1.33}) +[2026-06-23 21:36:01,297][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.06400914490222931, 'train/grad_norm_pre_clip': 34.826087951660156, 'timing/data': 0.000613063108175993, 'timing/model': 1.690386253874749, 'learning_rate/qwen_vl_interface': 2.704822723849765e-06, 'learning_rate/action_model': 2.7048227238497648e-05, 'epoch': 1.33}) +[2026-06-23 21:36:04,631][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.050659727305173874, 'train/grad_norm_pre_clip': 13.160863876342773, 'timing/data': 0.0006044059991836548, 'timing/model': 1.6956949208397418, 'learning_rate/qwen_vl_interface': 2.7022519127639687e-06, 'learning_rate/action_model': 2.7022519127639682e-05, 'epoch': 1.33}) +[2026-06-23 21:36:07,961][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.06917766481637955, 'train/grad_norm_pre_clip': 21.91994857788086, 'timing/data': 0.0007017690222710371, 'timing/model': 1.689468934899196, 'learning_rate/qwen_vl_interface': 2.6996821489611713e-06, 'learning_rate/action_model': 2.699682148961171e-05, 'epoch': 1.33}) +[2026-06-23 21:36:11,296][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.07210566103458405, 'train/grad_norm_pre_clip': 15.667524337768555, 'timing/data': 0.000603304011747241, 'timing/model': 1.6962924299295992, 'learning_rate/qwen_vl_interface': 2.697113433497712e-06, 'learning_rate/action_model': 2.6971134334977115e-05, 'epoch': 1.33}) +[2026-06-23 21:36:14,631][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.07227900624275208, 'train/grad_norm_pre_clip': 35.035423278808594, 'timing/data': 0.0006472968962043524, 'timing/model': 1.6914412470068783, 'learning_rate/qwen_vl_interface': 2.6945457674294863e-06, 'learning_rate/action_model': 2.694545767429486e-05, 'epoch': 1.33}) +[2026-06-23 21:36:18,054][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.0734236091375351, 'train/grad_norm_pre_clip': 75.22328186035156, 'timing/data': 0.0005908808670938015, 'timing/model': 1.6990226071793586, 'learning_rate/qwen_vl_interface': 2.6919791518119676e-06, 'learning_rate/action_model': 2.6919791518119674e-05, 'epoch': 1.33}) +[2026-06-23 21:36:21,392][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.09023823589086533, 'train/grad_norm_pre_clip': 17.38474464416504, 'timing/data': 0.000648068031296134, 'timing/model': 1.69818658195436, 'learning_rate/qwen_vl_interface': 2.6894135877001946e-06, 'learning_rate/action_model': 2.689413587700195e-05, 'epoch': 1.33}) +[2026-06-23 21:36:24,729][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.21357418596744537, 'train/grad_norm_pre_clip': 15.109232902526855, 'timing/data': 0.0005895679350942373, 'timing/model': 1.6976998450700194, 'learning_rate/qwen_vl_interface': 2.686849076148775e-06, 'learning_rate/action_model': 2.6868490761487747e-05, 'epoch': 1.33}) +[2026-06-23 21:36:28,066][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.20764420926570892, 'train/grad_norm_pre_clip': 23.504648208618164, 'timing/data': 0.0007577049545943737, 'timing/model': 1.6949186029378325, 'learning_rate/qwen_vl_interface': 2.68428561821188e-06, 'learning_rate/action_model': 2.6842856182118802e-05, 'epoch': 1.33}) +[2026-06-23 21:36:31,382][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.23210608959197998, 'train/grad_norm_pre_clip': 23.50290870666504, 'timing/data': 0.0006044360343366861, 'timing/model': 1.6840202319435775, 'learning_rate/qwen_vl_interface': 2.681723214943253e-06, 'learning_rate/action_model': 2.681723214943253e-05, 'epoch': 1.33}) +[2026-06-23 21:36:34,706][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.047199178487062454, 'train/grad_norm_pre_clip': 26.713825225830078, 'timing/data': 0.0006639792118221521, 'timing/model': 1.6891762430313975, 'learning_rate/qwen_vl_interface': 2.6791618673961996e-06, 'learning_rate/action_model': 2.6791618673961994e-05, 'epoch': 1.33}) +[2026-06-23 21:36:38,045][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.11486158519983292, 'train/grad_norm_pre_clip': 19.757787704467773, 'timing/data': 0.0005978839471936226, 'timing/model': 1.6957187801599503, 'learning_rate/qwen_vl_interface': 2.676601576623594e-06, 'learning_rate/action_model': 2.6766015766235937e-05, 'epoch': 1.33}) +[2026-06-23 21:36:41,371][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.1859818398952484, 'train/grad_norm_pre_clip': 18.48447036743164, 'timing/data': 0.000600338913500309, 'timing/model': 1.6885689869523048, 'learning_rate/qwen_vl_interface': 2.674042343677874e-06, 'learning_rate/action_model': 2.6740423436778738e-05, 'epoch': 1.33}) +[2026-06-23 21:36:44,696][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.18363189697265625, 'train/grad_norm_pre_clip': 53.31834411621094, 'timing/data': 0.0006549209356307983, 'timing/model': 1.6873460649512708, 'learning_rate/qwen_vl_interface': 2.6714841696110456e-06, 'learning_rate/action_model': 2.6714841696110456e-05, 'epoch': 1.33}) +[2026-06-23 21:36:48,024][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.1427117884159088, 'train/grad_norm_pre_clip': 28.582027435302734, 'timing/data': 0.0006705299019813538, 'timing/model': 1.6924869678914547, 'learning_rate/qwen_vl_interface': 2.6689270554746743e-06, 'learning_rate/action_model': 2.6689270554746743e-05, 'epoch': 1.33}) +[2026-06-23 21:36:51,358][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.22387084364891052, 'train/grad_norm_pre_clip': 31.917743682861328, 'timing/data': 0.0006616241298615932, 'timing/model': 1.6919842099305242, 'learning_rate/qwen_vl_interface': 2.6663710023198956e-06, 'learning_rate/action_model': 2.6663710023198956e-05, 'epoch': 1.33}) +[2026-06-23 21:36:54,689][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.0621853843331337, 'train/grad_norm_pre_clip': 14.373590469360352, 'timing/data': 0.0006538489833474159, 'timing/model': 1.6891341088339686, 'learning_rate/qwen_vl_interface': 2.663816011197405e-06, 'learning_rate/action_model': 2.6638160111974048e-05, 'epoch': 1.33}) +[2026-06-23 21:36:58,020][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.13056568801403046, 'train/grad_norm_pre_clip': 25.331958770751953, 'timing/data': 0.0006913188844919205, 'timing/model': 1.696040390059352, 'learning_rate/qwen_vl_interface': 2.661262083157463e-06, 'learning_rate/action_model': 2.661262083157463e-05, 'epoch': 1.33}) +[2026-06-23 21:37:01,344][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.11352100968360901, 'train/grad_norm_pre_clip': 25.12144660949707, 'timing/data': 0.0006135341245681047, 'timing/model': 1.6866354870144278, 'learning_rate/qwen_vl_interface': 2.6587092192498943e-06, 'learning_rate/action_model': 2.658709219249894e-05, 'epoch': 1.34}) +[2026-06-23 21:37:04,781][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.06521319597959518, 'train/grad_norm_pre_clip': 26.910045623779297, 'timing/data': 0.0006103271152824163, 'timing/model': 1.7916062499862164, 'learning_rate/qwen_vl_interface': 2.656157420524084e-06, 'learning_rate/action_model': 2.6561574205240842e-05, 'epoch': 1.34}) +[2026-06-23 21:37:08,131][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.18809644877910614, 'train/grad_norm_pre_clip': 22.3248348236084, 'timing/data': 0.000857681967318058, 'timing/model': 1.6965510330628604, 'learning_rate/qwen_vl_interface': 2.6536066880289795e-06, 'learning_rate/action_model': 2.6536066880289794e-05, 'epoch': 1.34}) +[2026-06-23 21:37:11,465][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.09533483535051346, 'train/grad_norm_pre_clip': 19.549997329711914, 'timing/data': 0.0005981149151921272, 'timing/model': 1.691400898154825, 'learning_rate/qwen_vl_interface': 2.6510570228130916e-06, 'learning_rate/action_model': 2.6510570228130914e-05, 'epoch': 1.34}) +[2026-06-23 21:37:14,798][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.1529308408498764, 'train/grad_norm_pre_clip': 27.20994758605957, 'timing/data': 0.0005845190025866032, 'timing/model': 1.6885774079710245, 'learning_rate/qwen_vl_interface': 2.648508425924491e-06, 'learning_rate/action_model': 2.6485084259244907e-05, 'epoch': 1.34}) +[2026-06-23 21:37:18,132][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.050030723214149475, 'train/grad_norm_pre_clip': 13.739775657653809, 'timing/data': 0.0006063398905098438, 'timing/model': 1.6928462411742657, 'learning_rate/qwen_vl_interface': 2.6459608984108093e-06, 'learning_rate/action_model': 2.645960898410809e-05, 'epoch': 1.34}) +[2026-06-23 21:37:21,469][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.08766569942235947, 'train/grad_norm_pre_clip': 25.595041275024414, 'timing/data': 0.0007295911200344563, 'timing/model': 1.695705302990973, 'learning_rate/qwen_vl_interface': 2.6434144413192414e-06, 'learning_rate/action_model': 2.6434144413192414e-05, 'epoch': 1.34}) +[2026-06-23 21:37:24,806][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.20376557111740112, 'train/grad_norm_pre_clip': 31.99052619934082, 'timing/data': 0.0006139238830655813, 'timing/model': 1.6998648098669946, 'learning_rate/qwen_vl_interface': 2.6408690556965347e-06, 'learning_rate/action_model': 2.640869055696535e-05, 'epoch': 1.34}) +[2026-06-23 21:37:28,137][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.14779961109161377, 'train/grad_norm_pre_clip': 25.115873336791992, 'timing/data': 0.0006521260365843773, 'timing/model': 1.686327600153163, 'learning_rate/qwen_vl_interface': 2.6383247425890047e-06, 'learning_rate/action_model': 2.6383247425890047e-05, 'epoch': 1.34}) +[2026-06-23 21:37:31,474][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.16391882300376892, 'train/grad_norm_pre_clip': 17.975852966308594, 'timing/data': 0.0006727040745317936, 'timing/model': 1.6910187860485166, 'learning_rate/qwen_vl_interface': 2.6357815030425204e-06, 'learning_rate/action_model': 2.6357815030425205e-05, 'epoch': 1.34}) +[2026-06-23 21:37:34,800][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.10712143778800964, 'train/grad_norm_pre_clip': 14.285249710083008, 'timing/data': 0.0006475269328802824, 'timing/model': 1.6905596610158682, 'learning_rate/qwen_vl_interface': 2.6332393381025132e-06, 'learning_rate/action_model': 2.6332393381025133e-05, 'epoch': 1.34}) +[2026-06-23 21:37:38,140][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.12447797507047653, 'train/grad_norm_pre_clip': 29.797178268432617, 'timing/data': 0.0006764610297977924, 'timing/model': 1.6922794010024518, 'learning_rate/qwen_vl_interface': 2.6306982488139704e-06, 'learning_rate/action_model': 2.6306982488139702e-05, 'epoch': 1.34}) +[2026-06-23 21:37:41,477][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.10511694848537445, 'train/grad_norm_pre_clip': 37.1583366394043, 'timing/data': 0.0006957680452615023, 'timing/model': 1.6927021141164005, 'learning_rate/qwen_vl_interface': 2.628158236221438e-06, 'learning_rate/action_model': 2.628158236221438e-05, 'epoch': 1.34}) +[2026-06-23 21:37:44,819][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.06843388825654984, 'train/grad_norm_pre_clip': 47.37158203125, 'timing/data': 0.0006060590967535973, 'timing/model': 1.6970198890194297, 'learning_rate/qwen_vl_interface': 2.6256193013690183e-06, 'learning_rate/action_model': 2.6256193013690182e-05, 'epoch': 1.34}) +[2026-06-23 21:37:48,153][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.14877399802207947, 'train/grad_norm_pre_clip': 19.37152862548828, 'timing/data': 0.0005870838649570942, 'timing/model': 1.6906654010526836, 'learning_rate/qwen_vl_interface': 2.623081445300373e-06, 'learning_rate/action_model': 2.6230814453003728e-05, 'epoch': 1.34}) +[2026-06-23 21:37:51,495][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.06199115887284279, 'train/grad_norm_pre_clip': 10.610944747924805, 'timing/data': 0.0005843080580234528, 'timing/model': 1.6913515790365636, 'learning_rate/qwen_vl_interface': 2.6205446690587183e-06, 'learning_rate/action_model': 2.6205446690587182e-05, 'epoch': 1.34}) +[2026-06-23 21:37:54,919][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.061121609061956406, 'train/grad_norm_pre_clip': 31.50823402404785, 'timing/data': 0.000595729798078537, 'timing/model': 1.6926690128166229, 'learning_rate/qwen_vl_interface': 2.618008973686828e-06, 'learning_rate/action_model': 2.618008973686828e-05, 'epoch': 1.34}) +[2026-06-23 21:37:58,257][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.074612557888031, 'train/grad_norm_pre_clip': 11.492240905761719, 'timing/data': 0.0006057790014892817, 'timing/model': 1.6973176191095263, 'learning_rate/qwen_vl_interface': 2.615474360227028e-06, 'learning_rate/action_model': 2.615474360227028e-05, 'epoch': 1.34}) +[2026-06-23 21:38:01,598][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.15472492575645447, 'train/grad_norm_pre_clip': 48.81889724731445, 'timing/data': 0.0005990869831293821, 'timing/model': 1.6954081980511546, 'learning_rate/qwen_vl_interface': 2.612940829721204e-06, 'learning_rate/action_model': 2.612940829721204e-05, 'epoch': 1.34}) +[2026-06-23 21:38:04,921][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.11012054979801178, 'train/grad_norm_pre_clip': 17.662845611572266, 'timing/data': 0.0005980241112411022, 'timing/model': 1.6895983940921724, 'learning_rate/qwen_vl_interface': 2.6104083832107948e-06, 'learning_rate/action_model': 2.6104083832107945e-05, 'epoch': 1.34}) +[2026-06-23 21:38:08,264][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.11480119079351425, 'train/grad_norm_pre_clip': 29.635656356811523, 'timing/data': 0.0005879050586372614, 'timing/model': 1.6995270571205765, 'learning_rate/qwen_vl_interface': 2.607877021736793e-06, 'learning_rate/action_model': 2.6078770217367926e-05, 'epoch': 1.34}) +[2026-06-23 21:38:11,603][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.04694259539246559, 'train/grad_norm_pre_clip': 50.21635055541992, 'timing/data': 0.0006044670008122921, 'timing/model': 1.6949673630297184, 'learning_rate/qwen_vl_interface': 2.605346746339745e-06, 'learning_rate/action_model': 2.605346746339745e-05, 'epoch': 1.34}) +[2026-06-23 21:38:14,943][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.2088288813829422, 'train/grad_norm_pre_clip': 44.086097717285156, 'timing/data': 0.0005994769744575024, 'timing/model': 1.6949092021677643, 'learning_rate/qwen_vl_interface': 2.602817558059753e-06, 'learning_rate/action_model': 2.6028175580597532e-05, 'epoch': 1.34}) +[2026-06-23 21:38:18,284][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.0708492249250412, 'train/grad_norm_pre_clip': 20.87818145751953, 'timing/data': 0.0006313370540738106, 'timing/model': 1.700951267965138, 'learning_rate/qwen_vl_interface': 2.60028945793647e-06, 'learning_rate/action_model': 2.60028945793647e-05, 'epoch': 1.34}) +[2026-06-23 21:38:21,623][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.15130895376205444, 'train/grad_norm_pre_clip': 45.41253662109375, 'timing/data': 0.0006097161676734686, 'timing/model': 1.6948506680782884, 'learning_rate/qwen_vl_interface': 2.5977624470091024e-06, 'learning_rate/action_model': 2.5977624470091023e-05, 'epoch': 1.34}) +[2026-06-23 21:38:24,962][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.08034859597682953, 'train/grad_norm_pre_clip': 10.282413482666016, 'timing/data': 0.0006559430621564388, 'timing/model': 1.6988101489841938, 'learning_rate/qwen_vl_interface': 2.59523652631641e-06, 'learning_rate/action_model': 2.5952365263164097e-05, 'epoch': 1.34}) +[2026-06-23 21:38:28,294][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.16763973236083984, 'train/grad_norm_pre_clip': 33.252017974853516, 'timing/data': 0.0006023820023983717, 'timing/model': 1.686792213935405, 'learning_rate/qwen_vl_interface': 2.5927116968967024e-06, 'learning_rate/action_model': 2.5927116968967025e-05, 'epoch': 1.35}) +[2026-06-23 21:38:31,635][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.29645463824272156, 'train/grad_norm_pre_clip': 21.90684700012207, 'timing/data': 0.0005992769729346037, 'timing/model': 1.699066166067496, 'learning_rate/qwen_vl_interface': 2.5901879597878417e-06, 'learning_rate/action_model': 2.5901879597878415e-05, 'epoch': 1.35}) +[2026-06-23 21:38:34,999][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.1413070261478424, 'train/grad_norm_pre_clip': 40.836238861083984, 'timing/data': 0.0006154770962893963, 'timing/model': 1.6942343201953918, 'learning_rate/qwen_vl_interface': 2.5876653160272395e-06, 'learning_rate/action_model': 2.5876653160272397e-05, 'epoch': 1.35}) +[2026-06-23 21:38:38,348][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.33574581146240234, 'train/grad_norm_pre_clip': 71.0676040649414, 'timing/data': 0.0008232479449361563, 'timing/model': 1.695987194078043, 'learning_rate/qwen_vl_interface': 2.5851437666518607e-06, 'learning_rate/action_model': 2.5851437666518607e-05, 'epoch': 1.35}) +[2026-06-23 21:38:41,682][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.12748309969902039, 'train/grad_norm_pre_clip': 17.69931411743164, 'timing/data': 0.0006688870489597321, 'timing/model': 1.6954608319792897, 'learning_rate/qwen_vl_interface': 2.5826233126982198e-06, 'learning_rate/action_model': 2.5826233126982196e-05, 'epoch': 1.35}) +[2026-06-23 21:38:45,166][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.16337749361991882, 'train/grad_norm_pre_clip': 78.88715362548828, 'timing/data': 0.0005881760735064745, 'timing/model': 1.6925879190675914, 'learning_rate/qwen_vl_interface': 2.580103955202381e-06, 'learning_rate/action_model': 2.580103955202381e-05, 'epoch': 1.35}) +[2026-06-23 21:38:48,507][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.09353954344987869, 'train/grad_norm_pre_clip': 52.47753143310547, 'timing/data': 0.0007276181131601334, 'timing/model': 1.6959702158346772, 'learning_rate/qwen_vl_interface': 2.577585695199954e-06, 'learning_rate/action_model': 2.5775856951999538e-05, 'epoch': 1.35}) +[2026-06-23 21:38:51,836][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.1662897914648056, 'train/grad_norm_pre_clip': 43.37989807128906, 'timing/data': 0.0008207031060010195, 'timing/model': 1.688273805892095, 'learning_rate/qwen_vl_interface': 2.575068533726101e-06, 'learning_rate/action_model': 2.575068533726101e-05, 'epoch': 1.35}) +[2026-06-23 21:38:55,176][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.08823399990797043, 'train/grad_norm_pre_clip': 19.606653213500977, 'timing/data': 0.000831283163279295, 'timing/model': 1.6952107439283282, 'learning_rate/qwen_vl_interface': 2.5725524718155333e-06, 'learning_rate/action_model': 2.572552471815533e-05, 'epoch': 1.35}) +[2026-06-23 21:38:58,519][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.03396190330386162, 'train/grad_norm_pre_clip': 43.611183166503906, 'timing/data': 0.0005846992135047913, 'timing/model': 1.6908324209507555, 'learning_rate/qwen_vl_interface': 2.570037510502508e-06, 'learning_rate/action_model': 2.5700375105025083e-05, 'epoch': 1.35}) +[2026-06-23 21:39:01,855][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.0868413895368576, 'train/grad_norm_pre_clip': 25.059253692626953, 'timing/data': 0.0006108689121901989, 'timing/model': 1.6940677422098815, 'learning_rate/qwen_vl_interface': 2.5675236508208313e-06, 'learning_rate/action_model': 2.5675236508208313e-05, 'epoch': 1.35}) +[2026-06-23 21:39:05,191][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.2500919699668884, 'train/grad_norm_pre_clip': 69.56835174560547, 'timing/data': 0.0005950888153165579, 'timing/model': 1.6926300069317222, 'learning_rate/qwen_vl_interface': 2.565010893803856e-06, 'learning_rate/action_model': 2.565010893803856e-05, 'epoch': 1.35}) +[2026-06-23 21:39:08,526][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.15959633886814117, 'train/grad_norm_pre_clip': 39.94553756713867, 'timing/data': 0.0005821450613439083, 'timing/model': 1.6924776390660554, 'learning_rate/qwen_vl_interface': 2.562499240484482e-06, 'learning_rate/action_model': 2.5624992404844816e-05, 'epoch': 1.35}) +[2026-06-23 21:39:11,858][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.1833769679069519, 'train/grad_norm_pre_clip': 20.29344367980957, 'timing/data': 0.0005982948932796717, 'timing/model': 1.6907151129562408, 'learning_rate/qwen_vl_interface': 2.559988691895155e-06, 'learning_rate/action_model': 2.559988691895155e-05, 'epoch': 1.35}) +[2026-06-23 21:39:15,197][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.15859028697013855, 'train/grad_norm_pre_clip': 30.255645751953125, 'timing/data': 0.0006046660710126162, 'timing/model': 1.6947225551120937, 'learning_rate/qwen_vl_interface': 2.5574792490678648e-06, 'learning_rate/action_model': 2.5574792490678644e-05, 'epoch': 1.35}) +[2026-06-23 21:39:18,524][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.19482043385505676, 'train/grad_norm_pre_clip': 50.29634094238281, 'timing/data': 0.0005854710470885038, 'timing/model': 1.6888913121074438, 'learning_rate/qwen_vl_interface': 2.5549709130341493e-06, 'learning_rate/action_model': 2.5549709130341493e-05, 'epoch': 1.35}) +[2026-06-23 21:39:21,859][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.09675834327936172, 'train/grad_norm_pre_clip': 22.580360412597656, 'timing/data': 0.000691368943080306, 'timing/model': 1.6904499791562557, 'learning_rate/qwen_vl_interface': 2.5524636848250924e-06, 'learning_rate/action_model': 2.552463684825092e-05, 'epoch': 1.35}) +[2026-06-23 21:39:25,195][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.10892251878976822, 'train/grad_norm_pre_clip': 20.517351150512695, 'timing/data': 0.0006130421534180641, 'timing/model': 1.698193916119635, 'learning_rate/qwen_vl_interface': 2.5499575654713186e-06, 'learning_rate/action_model': 2.5499575654713185e-05, 'epoch': 1.35}) +[2026-06-23 21:39:28,538][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.07652781903743744, 'train/grad_norm_pre_clip': 26.707860946655273, 'timing/data': 0.000904581043869257, 'timing/model': 1.699301365064457, 'learning_rate/qwen_vl_interface': 2.547452556003001e-06, 'learning_rate/action_model': 2.547452556003001e-05, 'epoch': 1.35}) +[2026-06-23 21:39:31,879][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.118583545088768, 'train/grad_norm_pre_clip': 13.090888023376465, 'timing/data': 0.0006947759538888931, 'timing/model': 1.696761348983273, 'learning_rate/qwen_vl_interface': 2.5449486574498535e-06, 'learning_rate/action_model': 2.5449486574498534e-05, 'epoch': 1.35}) +[2026-06-23 21:39:35,500][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.22886793315410614, 'train/grad_norm_pre_clip': 47.180877685546875, 'timing/data': 0.000604195985943079, 'timing/model': 1.6925946541596204, 'learning_rate/qwen_vl_interface': 2.5424458708411335e-06, 'learning_rate/action_model': 2.5424458708411335e-05, 'epoch': 1.35}) +[2026-06-23 21:39:38,828][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.09259400516748428, 'train/grad_norm_pre_clip': 34.75300598144531, 'timing/data': 0.0006032639648765326, 'timing/model': 1.6932946289889514, 'learning_rate/qwen_vl_interface': 2.5399441972056437e-06, 'learning_rate/action_model': 2.5399441972056438e-05, 'epoch': 1.35}) +[2026-06-23 21:39:42,163][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.06693321466445923, 'train/grad_norm_pre_clip': 42.41135025024414, 'timing/data': 0.0006133129354566336, 'timing/model': 1.6948094239924103, 'learning_rate/qwen_vl_interface': 2.537443637571726e-06, 'learning_rate/action_model': 2.537443637571726e-05, 'epoch': 1.35}) +[2026-06-23 21:39:45,493][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.08974609524011612, 'train/grad_norm_pre_clip': 79.22187805175781, 'timing/data': 0.0006032639648765326, 'timing/model': 1.690520429983735, 'learning_rate/qwen_vl_interface': 2.534944192967269e-06, 'learning_rate/action_model': 2.534944192967269e-05, 'epoch': 1.35}) +[2026-06-23 21:39:48,826][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3500 +[2026-06-23 21:41:22,869][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3500 +[2026-06-23 21:41:22,870][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3500 +[2026-06-23 21:42:33,744][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3500 +[2026-06-23 21:43:30,804][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3500 +[2026-06-23 21:43:30,805][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.040635839104652405, 'train/grad_norm_pre_clip': 12.430926322937012, 'eval/action_loss/seconds': 94.04022755799815, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06832483410835266, 'eval/latency_2/loss': 0.06832483410835266, 'eval/flappy/loss': 0.06832483410835266, 'eval/flappy/latency_2/loss': 0.06832483410835266, 'eval/flap_precision': 0.9141575912876362, 'eval/flap_recall': 0.9628879892037787, 'eval/flap_f1': 0.9378902398948408, 'eval/flap_rate_model': 0.21680555555555556, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9378902398948408, 'eval/latency_2/flap_precision': 0.9141575912876362, 'eval/latency_2/flap_recall': 0.9628879892037787, 'eval/latency_2/flap_f1': 0.9378902398948408, 'eval/latency_2/flap_rate_model': 0.21680555555555556, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9378902398948408, 'eval/action_classification/data_wait_seconds': 0.7922989726066589, 'eval/action_classification/predict_seconds': 69.94784545898438, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 40.73999999999998, 'rl_games_eval/mid_train/mean_length': 346.4, 'rl_games_eval/mid_train/std_reward': 32.796682759083815, 'rl_games_eval/mid_train/std_length': 265.3696290082948, 'rl_games_eval/mid_train/macro_mean_reward': 40.73999999999998, 'rl_games_eval/mid_train/macro_mean_length': 346.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 40.73999999999998, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 346.4, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 32.796682759083815, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 265.3696290082948, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 40.73999999999998, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005895178765058517, 'timing/model': 1.6922234501689672, 'learning_rate/qwen_vl_interface': 2.532445864419695e-06, 'learning_rate/action_model': 2.5324458644196952e-05, 'epoch': 1.35}) +[2026-06-23 21:43:30,805][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_3500_state +[2026-06-23 21:43:30,808][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 21:43:33,426][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_3500_state/model.safetensors +[2026-06-23 21:43:34,804][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_3500_state/optimizer.bin +[2026-06-23 21:43:34,805][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_3500_state/sampler.bin +[2026-06-23 21:43:34,805][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_3500_state/sampler_1.bin +[2026-06-23 21:43:34,805][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_3500_state/random_states_0.pkl +[2026-06-23 21:43:34,840][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500 +[2026-06-23 21:43:34,841][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 21:43:34,851][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 21:43:38,133][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.13599959015846252, 'train/grad_norm_pre_clip': 18.514633178710938, 'timing/data': 0.0005950289778411388, 'timing/model': 1.6650611138902605, 'learning_rate/qwen_vl_interface': 2.5299486529559758e-06, 'learning_rate/action_model': 2.5299486529559756e-05, 'epoch': 1.35}) +[2026-06-23 21:43:41,415][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.05875764414668083, 'train/grad_norm_pre_clip': 21.276363372802734, 'timing/data': 0.0006016308907419443, 'timing/model': 1.6659677871502936, 'learning_rate/qwen_vl_interface': 2.5274525596026202e-06, 'learning_rate/action_model': 2.52745255960262e-05, 'epoch': 1.36}) +[2026-06-23 21:43:44,706][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.10835622251033783, 'train/grad_norm_pre_clip': 16.91672706604004, 'timing/data': 0.0006005491595715284, 'timing/model': 1.6703176728915423, 'learning_rate/qwen_vl_interface': 2.524957585385677e-06, 'learning_rate/action_model': 2.524957585385677e-05, 'epoch': 1.36}) +[2026-06-23 21:43:47,996][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.14832161366939545, 'train/grad_norm_pre_clip': 14.65406322479248, 'timing/data': 0.0006057790014892817, 'timing/model': 1.6720193710643798, 'learning_rate/qwen_vl_interface': 2.522463731330738e-06, 'learning_rate/action_model': 2.5224637313307374e-05, 'epoch': 1.36}) +[2026-06-23 21:43:51,303][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.09913890063762665, 'train/grad_norm_pre_clip': 27.840618133544922, 'timing/data': 0.0006017307750880718, 'timing/model': 1.684634924866259, 'learning_rate/qwen_vl_interface': 2.5199709984629294e-06, 'learning_rate/action_model': 2.5199709984629292e-05, 'epoch': 1.36}) +[2026-06-23 21:43:54,606][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.08897975087165833, 'train/grad_norm_pre_clip': 30.993167877197266, 'timing/data': 0.0008910649921745062, 'timing/model': 1.677289224928245, 'learning_rate/qwen_vl_interface': 2.517479387806922e-06, 'learning_rate/action_model': 2.517479387806922e-05, 'epoch': 1.36}) +[2026-06-23 21:43:57,909][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.17774507403373718, 'train/grad_norm_pre_clip': 23.279390335083008, 'timing/data': 0.000584878958761692, 'timing/model': 1.6795234638266265, 'learning_rate/qwen_vl_interface': 2.5149889003869217e-06, 'learning_rate/action_model': 2.5149889003869214e-05, 'epoch': 1.36}) +[2026-06-23 21:44:01,219][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.05312733352184296, 'train/grad_norm_pre_clip': 43.965328216552734, 'timing/data': 0.0007366749923676252, 'timing/model': 1.680538770975545, 'learning_rate/qwen_vl_interface': 2.5124995372266736e-06, 'learning_rate/action_model': 2.5124995372266735e-05, 'epoch': 1.36}) +[2026-06-23 21:44:04,529][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.17067866027355194, 'train/grad_norm_pre_clip': 92.94027709960938, 'timing/data': 0.0007469039410352707, 'timing/model': 1.679124271031469, 'learning_rate/qwen_vl_interface': 2.5100112993494626e-06, 'learning_rate/action_model': 2.5100112993494624e-05, 'epoch': 1.36}) +[2026-06-23 21:44:07,837][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.14382410049438477, 'train/grad_norm_pre_clip': 58.1721305847168, 'timing/data': 0.0005984448362141848, 'timing/model': 1.6786024200264364, 'learning_rate/qwen_vl_interface': 2.5075241877781085e-06, 'learning_rate/action_model': 2.5075241877781086e-05, 'epoch': 1.36}) +[2026-06-23 21:44:11,144][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.02914937399327755, 'train/grad_norm_pre_clip': 33.818241119384766, 'timing/data': 0.000588095048442483, 'timing/model': 1.6859972281381488, 'learning_rate/qwen_vl_interface': 2.5050382035349673e-06, 'learning_rate/action_model': 2.5050382035349674e-05, 'epoch': 1.36}) +[2026-06-23 21:44:14,456][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.22614789009094238, 'train/grad_norm_pre_clip': 23.79105567932129, 'timing/data': 0.0006102169863879681, 'timing/model': 1.68392768711783, 'learning_rate/qwen_vl_interface': 2.5025533476419343e-06, 'learning_rate/action_model': 2.5025533476419343e-05, 'epoch': 1.36}) +[2026-06-23 21:44:17,769][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.0647670328617096, 'train/grad_norm_pre_clip': 20.322792053222656, 'timing/data': 0.0006051978562027216, 'timing/model': 1.687801379011944, 'learning_rate/qwen_vl_interface': 2.5000696211204394e-06, 'learning_rate/action_model': 2.5000696211204394e-05, 'epoch': 1.36}) +[2026-06-23 21:44:21,079][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.19499899446964264, 'train/grad_norm_pre_clip': 17.542987823486328, 'timing/data': 0.0005883050616830587, 'timing/model': 1.6791883260011673, 'learning_rate/qwen_vl_interface': 2.4975870249914495e-06, 'learning_rate/action_model': 2.4975870249914497e-05, 'epoch': 1.36}) +[2026-06-23 21:44:24,494][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.2081080824136734, 'train/grad_norm_pre_clip': 37.3673210144043, 'timing/data': 0.0006639179773628712, 'timing/model': 1.684842383954674, 'learning_rate/qwen_vl_interface': 2.4951055602754664e-06, 'learning_rate/action_model': 2.4951055602754665e-05, 'epoch': 1.36}) +[2026-06-23 21:44:27,810][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.07422597706317902, 'train/grad_norm_pre_clip': 26.924470901489258, 'timing/data': 0.000707058934494853, 'timing/model': 1.6796434500720352, 'learning_rate/qwen_vl_interface': 2.4926252279925255e-06, 'learning_rate/action_model': 2.4926252279925254e-05, 'epoch': 1.36}) +[2026-06-23 21:44:31,130][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.05109064280986786, 'train/grad_norm_pre_clip': 22.335826873779297, 'timing/data': 0.0005994969978928566, 'timing/model': 1.6878881470765918, 'learning_rate/qwen_vl_interface': 2.490146029162199e-06, 'learning_rate/action_model': 2.4901460291621988e-05, 'epoch': 1.36}) +[2026-06-23 21:44:34,447][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.12513409554958344, 'train/grad_norm_pre_clip': 32.03950881958008, 'timing/data': 0.0005866531282663345, 'timing/model': 1.684535123873502, 'learning_rate/qwen_vl_interface': 2.4876679648035905e-06, 'learning_rate/action_model': 2.4876679648035903e-05, 'epoch': 1.36}) +[2026-06-23 21:44:37,776][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.08017073571681976, 'train/grad_norm_pre_clip': 15.230900764465332, 'timing/data': 0.000613694079220295, 'timing/model': 1.6888144069816917, 'learning_rate/qwen_vl_interface': 2.4851910359353392e-06, 'learning_rate/action_model': 2.485191035935339e-05, 'epoch': 1.36}) +[2026-06-23 21:44:41,091][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.11916925013065338, 'train/grad_norm_pre_clip': 43.48700714111328, 'timing/data': 0.0005857509095221758, 'timing/model': 1.684001648798585, 'learning_rate/qwen_vl_interface': 2.4827152435756185e-06, 'learning_rate/action_model': 2.4827152435756185e-05, 'epoch': 1.36}) +[2026-06-23 21:44:44,405][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.09558748453855515, 'train/grad_norm_pre_clip': 34.80775833129883, 'timing/data': 0.0005779161583632231, 'timing/model': 1.6859744160901755, 'learning_rate/qwen_vl_interface': 2.4802405887421307e-06, 'learning_rate/action_model': 2.4802405887421306e-05, 'epoch': 1.36}) +[2026-06-23 21:44:47,723][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.2229263037443161, 'train/grad_norm_pre_clip': 25.214969635009766, 'timing/data': 0.0005890780594199896, 'timing/model': 1.6878106568474323, 'learning_rate/qwen_vl_interface': 2.4777670724521134e-06, 'learning_rate/action_model': 2.4777670724521134e-05, 'epoch': 1.36}) +[2026-06-23 21:44:51,065][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.08664606511592865, 'train/grad_norm_pre_clip': 35.03337478637695, 'timing/data': 0.000588645925745368, 'timing/model': 1.7068764860741794, 'learning_rate/qwen_vl_interface': 2.475294695722337e-06, 'learning_rate/action_model': 2.475294695722337e-05, 'epoch': 1.36}) +[2026-06-23 21:44:54,392][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.28646889328956604, 'train/grad_norm_pre_clip': 85.9751968383789, 'timing/data': 0.0007676640525460243, 'timing/model': 1.688926501898095, 'learning_rate/qwen_vl_interface': 2.4728234595691006e-06, 'learning_rate/action_model': 2.4728234595691005e-05, 'epoch': 1.36}) +[2026-06-23 21:44:57,719][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.24995461106300354, 'train/grad_norm_pre_clip': 72.36773681640625, 'timing/data': 0.0008486961014568806, 'timing/model': 1.6904952300246805, 'learning_rate/qwen_vl_interface': 2.4703533650082368e-06, 'learning_rate/action_model': 2.4703533650082366e-05, 'epoch': 1.36}) +[2026-06-23 21:45:01,043][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.15272323787212372, 'train/grad_norm_pre_clip': 57.52609634399414, 'timing/data': 0.0005937658715993166, 'timing/model': 1.6866708102170378, 'learning_rate/qwen_vl_interface': 2.4678844130551103e-06, 'learning_rate/action_model': 2.4678844130551103e-05, 'epoch': 1.36}) +[2026-06-23 21:45:04,366][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.08979254215955734, 'train/grad_norm_pre_clip': 37.21062088012695, 'timing/data': 0.0006087850779294968, 'timing/model': 1.690628970041871, 'learning_rate/qwen_vl_interface': 2.4654166047246102e-06, 'learning_rate/action_model': 2.46541660472461e-05, 'epoch': 1.37}) +[2026-06-23 21:45:07,691][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.11244475096464157, 'train/grad_norm_pre_clip': 32.397342681884766, 'timing/data': 0.000901954947039485, 'timing/model': 1.68772456003353, 'learning_rate/qwen_vl_interface': 2.4629499410311606e-06, 'learning_rate/action_model': 2.462949941031161e-05, 'epoch': 1.37}) +[2026-06-23 21:45:11,009][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.2776237428188324, 'train/grad_norm_pre_clip': 19.121095657348633, 'timing/data': 0.0007771309465169907, 'timing/model': 1.6774150589480996, 'learning_rate/qwen_vl_interface': 2.460484422988715e-06, 'learning_rate/action_model': 2.4604844229887146e-05, 'epoch': 1.37}) +[2026-06-23 21:45:14,445][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.05577126517891884, 'train/grad_norm_pre_clip': 16.970308303833008, 'timing/data': 0.000677524134516716, 'timing/model': 1.6892683699261397, 'learning_rate/qwen_vl_interface': 2.458020051610753e-06, 'learning_rate/action_model': 2.4580200516107528e-05, 'epoch': 1.37}) +[2026-06-23 21:45:17,767][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.10074436664581299, 'train/grad_norm_pre_clip': 22.475961685180664, 'timing/data': 0.0005858710501343012, 'timing/model': 1.6859346020501107, 'learning_rate/qwen_vl_interface': 2.455556827910286e-06, 'learning_rate/action_model': 2.455556827910286e-05, 'epoch': 1.37}) +[2026-06-23 21:45:21,092][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.11000943183898926, 'train/grad_norm_pre_clip': 41.568946838378906, 'timing/data': 0.0006688470020890236, 'timing/model': 1.6904580739792436, 'learning_rate/qwen_vl_interface': 2.453094752899853e-06, 'learning_rate/action_model': 2.453094752899853e-05, 'epoch': 1.37}) +[2026-06-23 21:45:24,419][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.12109845876693726, 'train/grad_norm_pre_clip': 28.399930953979492, 'timing/data': 0.0006448018830269575, 'timing/model': 1.6878485481720418, 'learning_rate/qwen_vl_interface': 2.4506338275915173e-06, 'learning_rate/action_model': 2.4506338275915172e-05, 'epoch': 1.37}) +[2026-06-23 21:45:27,746][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.07211685180664062, 'train/grad_norm_pre_clip': 24.924705505371094, 'timing/data': 0.0006698791403323412, 'timing/model': 1.6856146580539644, 'learning_rate/qwen_vl_interface': 2.4481740529968735e-06, 'learning_rate/action_model': 2.4481740529968734e-05, 'epoch': 1.37}) +[2026-06-23 21:45:31,076][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.09795079380273819, 'train/grad_norm_pre_clip': 27.95465087890625, 'timing/data': 0.0006049168296158314, 'timing/model': 1.6916373788844794, 'learning_rate/qwen_vl_interface': 2.4457154301270427e-06, 'learning_rate/action_model': 2.4457154301270427e-05, 'epoch': 1.37}) +[2026-06-23 21:45:34,407][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.1973501294851303, 'train/grad_norm_pre_clip': 99.32413482666016, 'timing/data': 0.0005959810223430395, 'timing/model': 1.6910564647987485, 'learning_rate/qwen_vl_interface': 2.443257959992671e-06, 'learning_rate/action_model': 2.443257959992671e-05, 'epoch': 1.37}) +[2026-06-23 21:45:37,734][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.1709766834974289, 'train/grad_norm_pre_clip': 53.87592315673828, 'timing/data': 0.0006027431227266788, 'timing/model': 1.6895934441126883, 'learning_rate/qwen_vl_interface': 2.440801643603933e-06, 'learning_rate/action_model': 2.440801643603933e-05, 'epoch': 1.37}) +[2026-06-23 21:45:41,070][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.1461866945028305, 'train/grad_norm_pre_clip': 40.225006103515625, 'timing/data': 0.000581854023039341, 'timing/model': 1.6929796091280878, 'learning_rate/qwen_vl_interface': 2.4383464819705252e-06, 'learning_rate/action_model': 2.438346481970525e-05, 'epoch': 1.37}) +[2026-06-23 21:45:44,396][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.1476520299911499, 'train/grad_norm_pre_clip': 30.26905059814453, 'timing/data': 0.0006304248236119747, 'timing/model': 1.6848835789132863, 'learning_rate/qwen_vl_interface': 2.435892476101674e-06, 'learning_rate/action_model': 2.435892476101674e-05, 'epoch': 1.37}) +[2026-06-23 21:45:47,725][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.04908077418804169, 'train/grad_norm_pre_clip': 31.975894927978516, 'timing/data': 0.0007448401302099228, 'timing/model': 1.6934376179706305, 'learning_rate/qwen_vl_interface': 2.433439627006128e-06, 'learning_rate/action_model': 2.4334396270061277e-05, 'epoch': 1.37}) +[2026-06-23 21:45:51,058][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.061292801052331924, 'train/grad_norm_pre_clip': 70.62999725341797, 'timing/data': 0.0006871509831398726, 'timing/model': 1.6927009518258274, 'learning_rate/qwen_vl_interface': 2.4309879356921602e-06, 'learning_rate/action_model': 2.43098793569216e-05, 'epoch': 1.37}) +[2026-06-23 21:45:54,382][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.2070826143026352, 'train/grad_norm_pre_clip': 50.91204833984375, 'timing/data': 0.0006159481126815081, 'timing/model': 1.6873671649955213, 'learning_rate/qwen_vl_interface': 2.428537403167569e-06, 'learning_rate/action_model': 2.4285374031675687e-05, 'epoch': 1.37}) +[2026-06-23 21:45:57,720][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.06457384675741196, 'train/grad_norm_pre_clip': 15.668339729309082, 'timing/data': 0.0007518241181969643, 'timing/model': 1.6965766861103475, 'learning_rate/qwen_vl_interface': 2.4260880304396765e-06, 'learning_rate/action_model': 2.4260880304396765e-05, 'epoch': 1.37}) +[2026-06-23 21:46:01,150][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.08050116896629333, 'train/grad_norm_pre_clip': 12.012225151062012, 'timing/data': 0.0005923840217292309, 'timing/model': 1.78839849983342, 'learning_rate/qwen_vl_interface': 2.4236398185153245e-06, 'learning_rate/action_model': 2.4236398185153245e-05, 'epoch': 1.37}) +[2026-06-23 21:46:04,477][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.14529407024383545, 'train/grad_norm_pre_clip': 65.95470428466797, 'timing/data': 0.000598495127633214, 'timing/model': 1.6939951519016176, 'learning_rate/qwen_vl_interface': 2.4211927684008823e-06, 'learning_rate/action_model': 2.4211927684008824e-05, 'epoch': 1.37}) +[2026-06-23 21:46:07,810][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.0908050462603569, 'train/grad_norm_pre_clip': 62.89104080200195, 'timing/data': 0.0006089850794523954, 'timing/model': 1.6929767560213804, 'learning_rate/qwen_vl_interface': 2.4187468811022398e-06, 'learning_rate/action_model': 2.4187468811022396e-05, 'epoch': 1.37}) +[2026-06-23 21:46:11,135][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.09241755306720734, 'train/grad_norm_pre_clip': 23.20384407043457, 'timing/data': 0.0005890671163797379, 'timing/model': 1.6891661728732288, 'learning_rate/qwen_vl_interface': 2.416302157624809e-06, 'learning_rate/action_model': 2.4163021576248088e-05, 'epoch': 1.37}) +[2026-06-23 21:46:14,463][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.09550070762634277, 'train/grad_norm_pre_clip': 19.53407859802246, 'timing/data': 0.0005909111350774765, 'timing/model': 1.6898023379035294, 'learning_rate/qwen_vl_interface': 2.4138585989735227e-06, 'learning_rate/action_model': 2.4138585989735226e-05, 'epoch': 1.37}) +[2026-06-23 21:46:17,797][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.15357078611850739, 'train/grad_norm_pre_clip': 64.62898254394531, 'timing/data': 0.0006814808584749699, 'timing/model': 1.693827283103019, 'learning_rate/qwen_vl_interface': 2.4114162061528357e-06, 'learning_rate/action_model': 2.4114162061528354e-05, 'epoch': 1.37}) +[2026-06-23 21:46:21,123][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.08167712390422821, 'train/grad_norm_pre_clip': 12.15070915222168, 'timing/data': 0.0008149121422320604, 'timing/model': 1.6871008679736406, 'learning_rate/qwen_vl_interface': 2.408974980166724e-06, 'learning_rate/action_model': 2.408974980166724e-05, 'epoch': 1.37}) +[2026-06-23 21:46:24,492][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.12949182093143463, 'train/grad_norm_pre_clip': 42.61381530761719, 'timing/data': 0.0005860719829797745, 'timing/model': 1.6959374479483813, 'learning_rate/qwen_vl_interface': 2.406534922018683e-06, 'learning_rate/action_model': 2.4065349220186827e-05, 'epoch': 1.37}) +[2026-06-23 21:46:27,820][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.1288052797317505, 'train/grad_norm_pre_clip': 29.77737045288086, 'timing/data': 0.0006254150066524744, 'timing/model': 1.6884881937876344, 'learning_rate/qwen_vl_interface': 2.4040960327117287e-06, 'learning_rate/action_model': 2.4040960327117283e-05, 'epoch': 1.37}) +[2026-06-23 21:46:31,146][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.22754639387130737, 'train/grad_norm_pre_clip': 20.341812133789062, 'timing/data': 0.0005908110179007053, 'timing/model': 1.6925730421207845, 'learning_rate/qwen_vl_interface': 2.4016583132483976e-06, 'learning_rate/action_model': 2.4016583132483976e-05, 'epoch': 1.38}) +[2026-06-23 21:46:34,468][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.13970690965652466, 'train/grad_norm_pre_clip': 18.210500717163086, 'timing/data': 0.000585139961913228, 'timing/model': 1.6829986840020865, 'learning_rate/qwen_vl_interface': 2.399221764630741e-06, 'learning_rate/action_model': 2.3992217646307406e-05, 'epoch': 1.38}) +[2026-06-23 21:46:37,792][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.05307653546333313, 'train/grad_norm_pre_clip': 17.84888458251953, 'timing/data': 0.0005967919714748859, 'timing/model': 1.6872913520783186, 'learning_rate/qwen_vl_interface': 2.3967863878603334e-06, 'learning_rate/action_model': 2.3967863878603332e-05, 'epoch': 1.38}) +[2026-06-23 21:46:41,122][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.15705662965774536, 'train/grad_norm_pre_clip': 44.39897155761719, 'timing/data': 0.0005787878762930632, 'timing/model': 1.690534858033061, 'learning_rate/qwen_vl_interface': 2.394352183938266e-06, 'learning_rate/action_model': 2.3943521839382664e-05, 'epoch': 1.38}) +[2026-06-23 21:46:44,450][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.1716468632221222, 'train/grad_norm_pre_clip': 79.46923828125, 'timing/data': 0.0006931428797543049, 'timing/model': 1.6857580479700118, 'learning_rate/qwen_vl_interface': 2.39191915386515e-06, 'learning_rate/action_model': 2.39191915386515e-05, 'epoch': 1.38}) +[2026-06-23 21:46:47,779][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.08011824637651443, 'train/grad_norm_pre_clip': 26.3599853515625, 'timing/data': 0.0005906301084905863, 'timing/model': 1.6942163279745728, 'learning_rate/qwen_vl_interface': 2.3894872986411095e-06, 'learning_rate/action_model': 2.3894872986411094e-05, 'epoch': 1.38}) +[2026-06-23 21:46:51,418][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.04674640670418739, 'train/grad_norm_pre_clip': 45.48541259765625, 'timing/data': 0.0005919830873608589, 'timing/model': 2.0055707490537316, 'learning_rate/qwen_vl_interface': 2.3870566192657903e-06, 'learning_rate/action_model': 2.38705661926579e-05, 'epoch': 1.38}) +[2026-06-23 21:46:54,755][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.23975273966789246, 'train/grad_norm_pre_clip': 20.40447425842285, 'timing/data': 0.0006801888812333345, 'timing/model': 1.6948901328723878, 'learning_rate/qwen_vl_interface': 2.3846271167383515e-06, 'learning_rate/action_model': 2.3846271167383515e-05, 'epoch': 1.38}) +[2026-06-23 21:46:58,088][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.07381270080804825, 'train/grad_norm_pre_clip': 26.25928497314453, 'timing/data': 0.0005923139397054911, 'timing/model': 1.6986001848708838, 'learning_rate/qwen_vl_interface': 2.382198792057471e-06, 'learning_rate/action_model': 2.3821987920574707e-05, 'epoch': 1.38}) +[2026-06-23 21:47:01,413][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.17960751056671143, 'train/grad_norm_pre_clip': 20.626203536987305, 'timing/data': 0.0008465209975838661, 'timing/model': 1.6898781589698046, 'learning_rate/qwen_vl_interface': 2.3797716462213402e-06, 'learning_rate/action_model': 2.3797716462213402e-05, 'epoch': 1.38}) +[2026-06-23 21:47:04,748][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.11046972125768661, 'train/grad_norm_pre_clip': 46.907936096191406, 'timing/data': 0.0006152358837425709, 'timing/model': 1.6931420429609716, 'learning_rate/qwen_vl_interface': 2.37734568022767e-06, 'learning_rate/action_model': 2.3773456802276697e-05, 'epoch': 1.38}) +[2026-06-23 21:47:08,078][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.09052907675504684, 'train/grad_norm_pre_clip': 28.09237289428711, 'timing/data': 0.0005990858189761639, 'timing/model': 1.6895140639971942, 'learning_rate/qwen_vl_interface': 2.3749208950736783e-06, 'learning_rate/action_model': 2.374920895073678e-05, 'epoch': 1.38}) +[2026-06-23 21:47:11,413][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.11222367733716965, 'train/grad_norm_pre_clip': 18.03158950805664, 'timing/data': 0.0006819718983024359, 'timing/model': 1.6935648769140244, 'learning_rate/qwen_vl_interface': 2.372497291756106e-06, 'learning_rate/action_model': 2.3724972917561058e-05, 'epoch': 1.38}) +[2026-06-23 21:47:14,741][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.06486283987760544, 'train/grad_norm_pre_clip': 49.951576232910156, 'timing/data': 0.0005907700397074223, 'timing/model': 1.6896046621259302, 'learning_rate/qwen_vl_interface': 2.3700748712712043e-06, 'learning_rate/action_model': 2.370074871271204e-05, 'epoch': 1.38}) +[2026-06-23 21:47:18,066][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.10496728867292404, 'train/grad_norm_pre_clip': 19.40267562866211, 'timing/data': 0.0006923810578882694, 'timing/model': 1.6927381411660463, 'learning_rate/qwen_vl_interface': 2.3676536346147388e-06, 'learning_rate/action_model': 2.3676536346147388e-05, 'epoch': 1.38}) +[2026-06-23 21:47:21,389][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.09713520109653473, 'train/grad_norm_pre_clip': 56.202449798583984, 'timing/data': 0.0007136608473956585, 'timing/model': 1.6836058150511235, 'learning_rate/qwen_vl_interface': 2.365233582781988e-06, 'learning_rate/action_model': 2.365233582781988e-05, 'epoch': 1.38}) +[2026-06-23 21:47:24,723][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.1108885183930397, 'train/grad_norm_pre_clip': 30.991689682006836, 'timing/data': 0.0005997668486088514, 'timing/model': 1.6989312509540468, 'learning_rate/qwen_vl_interface': 2.3628147167677444e-06, 'learning_rate/action_model': 2.3628147167677443e-05, 'epoch': 1.38}) +[2026-06-23 21:47:28,056][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.05511856824159622, 'train/grad_norm_pre_clip': 21.51907730102539, 'timing/data': 0.0005927241872996092, 'timing/model': 1.6937095359899104, 'learning_rate/qwen_vl_interface': 2.3603970375663123e-06, 'learning_rate/action_model': 2.3603970375663125e-05, 'epoch': 1.38}) +[2026-06-23 21:47:31,384][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.08936218172311783, 'train/grad_norm_pre_clip': 19.839698791503906, 'timing/data': 0.0005979849956929684, 'timing/model': 1.6877501630224288, 'learning_rate/qwen_vl_interface': 2.35798054617151e-06, 'learning_rate/action_model': 2.3579805461715096e-05, 'epoch': 1.38}) +[2026-06-23 21:47:34,719][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.17793133854866028, 'train/grad_norm_pre_clip': 79.46568298339844, 'timing/data': 0.000694134971126914, 'timing/model': 1.6954893989022821, 'learning_rate/qwen_vl_interface': 2.355565243576662e-06, 'learning_rate/action_model': 2.3555652435766617e-05, 'epoch': 1.38}) +[2026-06-23 21:47:38,046][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.20478254556655884, 'train/grad_norm_pre_clip': 16.22136688232422, 'timing/data': 0.0006320280954241753, 'timing/model': 1.6910701119340956, 'learning_rate/qwen_vl_interface': 2.35315113077461e-06, 'learning_rate/action_model': 2.3531511307746098e-05, 'epoch': 1.38}) +[2026-06-23 21:47:41,674][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.07071589678525925, 'train/grad_norm_pre_clip': 18.437057495117188, 'timing/data': 0.0006471159867942333, 'timing/model': 1.9895728940609843, 'learning_rate/qwen_vl_interface': 2.3507382087577047e-06, 'learning_rate/action_model': 2.350738208757705e-05, 'epoch': 1.38}) +[2026-06-23 21:47:44,998][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.09553778916597366, 'train/grad_norm_pre_clip': 16.933382034301758, 'timing/data': 0.0005994569510221481, 'timing/model': 1.6904285051859915, 'learning_rate/qwen_vl_interface': 2.348326478517807e-06, 'learning_rate/action_model': 2.3483264785178072e-05, 'epoch': 1.38}) +[2026-06-23 21:47:48,329][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.057264260947704315, 'train/grad_norm_pre_clip': 11.70136833190918, 'timing/data': 0.0006052579265087843, 'timing/model': 1.69146195682697, 'learning_rate/qwen_vl_interface': 2.3459159410462907e-06, 'learning_rate/action_model': 2.3459159410462905e-05, 'epoch': 1.38}) +[2026-06-23 21:47:51,660][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.06376904994249344, 'train/grad_norm_pre_clip': 39.993465423583984, 'timing/data': 0.0006054979749023914, 'timing/model': 1.6917777189519256, 'learning_rate/qwen_vl_interface': 2.343506597334031e-06, 'learning_rate/action_model': 2.343506597334031e-05, 'epoch': 1.38}) +[2026-06-23 21:47:54,980][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.17419680953025818, 'train/grad_norm_pre_clip': 40.78464126586914, 'timing/data': 0.0006252250168472528, 'timing/model': 1.6854622368700802, 'learning_rate/qwen_vl_interface': 2.341098448371423e-06, 'learning_rate/action_model': 2.3410984483714227e-05, 'epoch': 1.38}) +[2026-06-23 21:47:58,297][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.08742506057024002, 'train/grad_norm_pre_clip': 21.780492782592773, 'timing/data': 0.0005956289824098349, 'timing/model': 1.6817162449005991, 'learning_rate/qwen_vl_interface': 2.338691495148363e-06, 'learning_rate/action_model': 2.3386914951483626e-05, 'epoch': 1.39}) +[2026-06-23 21:48:01,627][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.19326898455619812, 'train/grad_norm_pre_clip': 51.254310607910156, 'timing/data': 0.0006161381024867296, 'timing/model': 1.6939425258897245, 'learning_rate/qwen_vl_interface': 2.33628573865426e-06, 'learning_rate/action_model': 2.3362857386542596e-05, 'epoch': 1.39}) +[2026-06-23 21:48:04,955][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.10733947157859802, 'train/grad_norm_pre_clip': 15.367819786071777, 'timing/data': 0.0006924220360815525, 'timing/model': 1.6909870801027864, 'learning_rate/qwen_vl_interface': 2.3338811798780282e-06, 'learning_rate/action_model': 2.333881179878028e-05, 'epoch': 1.39}) +[2026-06-23 21:48:08,277][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.152843177318573, 'train/grad_norm_pre_clip': 18.904876708984375, 'timing/data': 0.0005927549209445715, 'timing/model': 1.6891323029994965, 'learning_rate/qwen_vl_interface': 2.3314778198080933e-06, 'learning_rate/action_model': 2.331477819808093e-05, 'epoch': 1.39}) +[2026-06-23 21:48:11,603][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.1297820508480072, 'train/grad_norm_pre_clip': 29.745681762695312, 'timing/data': 0.0006072919350117445, 'timing/model': 1.691836325917393, 'learning_rate/qwen_vl_interface': 2.3290756594323834e-06, 'learning_rate/action_model': 2.3290756594323836e-05, 'epoch': 1.39}) +[2026-06-23 21:48:14,928][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.07415139675140381, 'train/grad_norm_pre_clip': 29.047056198120117, 'timing/data': 0.0006846671458333731, 'timing/model': 1.6914048839826137, 'learning_rate/qwen_vl_interface': 2.3266746997383378e-06, 'learning_rate/action_model': 2.3266746997383377e-05, 'epoch': 1.39}) +[2026-06-23 21:48:18,260][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.1305234283208847, 'train/grad_norm_pre_clip': 56.19530487060547, 'timing/data': 0.0006481790915131569, 'timing/model': 1.6933932609390467, 'learning_rate/qwen_vl_interface': 2.324274941712899e-06, 'learning_rate/action_model': 2.324274941712899e-05, 'epoch': 1.39}) +[2026-06-23 21:48:21,585][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.09057687222957611, 'train/grad_norm_pre_clip': 42.045440673828125, 'timing/data': 0.0009127059020102024, 'timing/model': 1.6930521398317069, 'learning_rate/qwen_vl_interface': 2.321876386342519e-06, 'learning_rate/action_model': 2.3218763863425187e-05, 'epoch': 1.39}) +[2026-06-23 21:48:24,920][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.12153179943561554, 'train/grad_norm_pre_clip': 26.298952102661133, 'timing/data': 0.0005926028825342655, 'timing/model': 1.6946035139262676, 'learning_rate/qwen_vl_interface': 2.31947903461315e-06, 'learning_rate/action_model': 2.3194790346131502e-05, 'epoch': 1.39}) +[2026-06-23 21:48:28,258][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.13384626805782318, 'train/grad_norm_pre_clip': 26.152233123779297, 'timing/data': 0.0006083641201257706, 'timing/model': 1.6993421909864992, 'learning_rate/qwen_vl_interface': 2.3170828875102544e-06, 'learning_rate/action_model': 2.3170828875102543e-05, 'epoch': 1.39}) +[2026-06-23 21:48:31,674][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.06715835630893707, 'train/grad_norm_pre_clip': 20.344701766967773, 'timing/data': 0.0005983549635857344, 'timing/model': 1.6879910740535706, 'learning_rate/qwen_vl_interface': 2.3146879460187984e-06, 'learning_rate/action_model': 2.3146879460187982e-05, 'epoch': 1.39}) +[2026-06-23 21:48:34,997][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.07740108668804169, 'train/grad_norm_pre_clip': 13.037406921386719, 'timing/data': 0.0005927348975092173, 'timing/model': 1.6887791249901056, 'learning_rate/qwen_vl_interface': 2.3122942111232506e-06, 'learning_rate/action_model': 2.3122942111232506e-05, 'epoch': 1.39}) +[2026-06-23 21:48:38,326][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.17400237917900085, 'train/grad_norm_pre_clip': 26.371049880981445, 'timing/data': 0.0005758919287472963, 'timing/model': 1.688321724999696, 'learning_rate/qwen_vl_interface': 2.3099016838075857e-06, 'learning_rate/action_model': 2.3099016838075856e-05, 'epoch': 1.39}) +[2026-06-23 21:48:41,644][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.12650862336158752, 'train/grad_norm_pre_clip': 14.112495422363281, 'timing/data': 0.0006306848954409361, 'timing/model': 1.6861588768661022, 'learning_rate/qwen_vl_interface': 2.307510365055282e-06, 'learning_rate/action_model': 2.307510365055282e-05, 'epoch': 1.39}) +[2026-06-23 21:48:44,980][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.09247662127017975, 'train/grad_norm_pre_clip': 23.907873153686523, 'timing/data': 0.0005871839821338654, 'timing/model': 1.6996905039995909, 'learning_rate/qwen_vl_interface': 2.3051202558493196e-06, 'learning_rate/action_model': 2.3051202558493192e-05, 'epoch': 1.39}) +[2026-06-23 21:48:48,304][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.13715122640132904, 'train/grad_norm_pre_clip': 19.04229164123535, 'timing/data': 0.0005851299501955509, 'timing/model': 1.6895940338727087, 'learning_rate/qwen_vl_interface': 2.302731357172182e-06, 'learning_rate/action_model': 2.302731357172182e-05, 'epoch': 1.39}) +[2026-06-23 21:48:51,630][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.13977381587028503, 'train/grad_norm_pre_clip': 31.639802932739258, 'timing/data': 0.0005970529746264219, 'timing/model': 1.6898040880914778, 'learning_rate/qwen_vl_interface': 2.3003436700058556e-06, 'learning_rate/action_model': 2.3003436700058554e-05, 'epoch': 1.39}) +[2026-06-23 21:48:54,953][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.1076018214225769, 'train/grad_norm_pre_clip': 14.917695045471191, 'timing/data': 0.0007078100461512804, 'timing/model': 1.6894011599943042, 'learning_rate/qwen_vl_interface': 2.297957195331828e-06, 'learning_rate/action_model': 2.2979571953318276e-05, 'epoch': 1.39}) +[2026-06-23 21:48:58,278][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.1529756784439087, 'train/grad_norm_pre_clip': 40.142948150634766, 'timing/data': 0.0006016208790242672, 'timing/model': 1.6896560878958553, 'learning_rate/qwen_vl_interface': 2.2955719341310903e-06, 'learning_rate/action_model': 2.2955719341310903e-05, 'epoch': 1.39}) +[2026-06-23 21:49:01,595][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.07217654585838318, 'train/grad_norm_pre_clip': 16.645416259765625, 'timing/data': 0.0006061990279704332, 'timing/model': 1.681438370142132, 'learning_rate/qwen_vl_interface': 2.2931878873841304e-06, 'learning_rate/action_model': 2.2931878873841302e-05, 'epoch': 1.39}) +[2026-06-23 21:49:04,921][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.09869984537363052, 'train/grad_norm_pre_clip': 29.736679077148438, 'timing/data': 0.0005881150718778372, 'timing/model': 1.6886822159867734, 'learning_rate/qwen_vl_interface': 2.290805056070942e-06, 'learning_rate/action_model': 2.290805056070942e-05, 'epoch': 1.39}) +[2026-06-23 21:49:08,234][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3600 +[2026-06-23 21:50:41,277][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3600 +[2026-06-23 21:50:41,278][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3600 +[2026-06-23 21:51:53,188][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3600 +[2026-06-23 21:51:53,188][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.09412969648838043, 'train/grad_norm_pre_clip': 15.45102310180664, 'eval/action_loss/seconds': 93.04042657115497, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06626207381486893, 'eval/latency_2/loss': 0.06626207381486893, 'eval/flappy/loss': 0.06626207381486893, 'eval/flappy/latency_2/loss': 0.06626207381486893, 'eval/flap_precision': 0.9269736842105263, 'eval/flap_recall': 0.9507422402159245, 'eval/flap_f1': 0.9387075283144571, 'eval/flap_rate_model': 0.2111111111111111, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9387075283144573, 'eval/latency_2/flap_precision': 0.9269736842105263, 'eval/latency_2/flap_recall': 0.9507422402159245, 'eval/latency_2/flap_f1': 0.9387075283144571, 'eval/latency_2/flap_rate_model': 0.2111111111111111, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9387075283144573, 'eval/action_classification/data_wait_seconds': 0.8555967807769775, 'eval/action_classification/predict_seconds': 70.90900421142578, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006442510057240725, 'timing/model': 1.6816741228103638, 'learning_rate/qwen_vl_interface': 2.288423441171016e-06, 'learning_rate/action_model': 2.288423441171016e-05, 'epoch': 1.39}) +[2026-06-23 21:51:56,514][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.12094125896692276, 'train/grad_norm_pre_clip': 18.282663345336914, 'timing/data': 0.0007967578712850809, 'timing/model': 1.6885803020559251, 'learning_rate/qwen_vl_interface': 2.2860430436633447e-06, 'learning_rate/action_model': 2.2860430436633447e-05, 'epoch': 1.39}) +[2026-06-23 21:51:59,837][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.09481367468833923, 'train/grad_norm_pre_clip': 9.569594383239746, 'timing/data': 0.0005897390656173229, 'timing/model': 1.688305919058621, 'learning_rate/qwen_vl_interface': 2.28366386452642e-06, 'learning_rate/action_model': 2.28366386452642e-05, 'epoch': 1.39}) +[2026-06-23 21:52:03,166][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.10499922931194305, 'train/grad_norm_pre_clip': 20.234100341796875, 'timing/data': 0.0006449420470744371, 'timing/model': 1.6896076989360154, 'learning_rate/qwen_vl_interface': 2.2812859047382326e-06, 'learning_rate/action_model': 2.2812859047382323e-05, 'epoch': 1.39}) +[2026-06-23 21:52:06,493][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.11247853934764862, 'train/grad_norm_pre_clip': 35.53672409057617, 'timing/data': 0.0006107471417635679, 'timing/model': 1.6875921729952097, 'learning_rate/qwen_vl_interface': 2.2789091652762712e-06, 'learning_rate/action_model': 2.2789091652762712e-05, 'epoch': 1.39}) +[2026-06-23 21:52:09,816][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.16198450326919556, 'train/grad_norm_pre_clip': 28.77012062072754, 'timing/data': 0.0006520450115203857, 'timing/model': 1.6901113400235772, 'learning_rate/qwen_vl_interface': 2.276533647117524e-06, 'learning_rate/action_model': 2.2765336471175242e-05, 'epoch': 1.4}) +[2026-06-23 21:52:13,143][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.07019238919019699, 'train/grad_norm_pre_clip': 29.681640625, 'timing/data': 0.0005940471310168505, 'timing/model': 1.6906343449372798, 'learning_rate/qwen_vl_interface': 2.2741593512384773e-06, 'learning_rate/action_model': 2.2741593512384772e-05, 'epoch': 1.4}) +[2026-06-23 21:52:16,473][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.13650652766227722, 'train/grad_norm_pre_clip': 17.602371215820312, 'timing/data': 0.0006750680040568113, 'timing/model': 1.6907678090501577, 'learning_rate/qwen_vl_interface': 2.2717862786151153e-06, 'learning_rate/action_model': 2.271786278615115e-05, 'epoch': 1.4}) +[2026-06-23 21:52:19,800][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.10923389345407486, 'train/grad_norm_pre_clip': 16.209684371948242, 'timing/data': 0.0006966001819819212, 'timing/model': 1.6907309421803802, 'learning_rate/qwen_vl_interface': 2.2694144302229152e-06, 'learning_rate/action_model': 2.2694144302229152e-05, 'epoch': 1.4}) +[2026-06-23 21:52:23,137][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.14183947443962097, 'train/grad_norm_pre_clip': 24.381444931030273, 'timing/data': 0.0006971701513975859, 'timing/model': 1.6967750638723373, 'learning_rate/qwen_vl_interface': 2.2670438070368568e-06, 'learning_rate/action_model': 2.2670438070368566e-05, 'epoch': 1.4}) +[2026-06-23 21:52:26,474][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.06907473504543304, 'train/grad_norm_pre_clip': 24.00330352783203, 'timing/data': 0.000610788119956851, 'timing/model': 1.6943651270121336, 'learning_rate/qwen_vl_interface': 2.264674410031413e-06, 'learning_rate/action_model': 2.2646744100314127e-05, 'epoch': 1.4}) +[2026-06-23 21:52:29,896][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.06959721446037292, 'train/grad_norm_pre_clip': 14.444687843322754, 'timing/data': 0.0006072409451007843, 'timing/model': 1.7802908620797098, 'learning_rate/qwen_vl_interface': 2.2623062401805537e-06, 'learning_rate/action_model': 2.2623062401805535e-05, 'epoch': 1.4}) +[2026-06-23 21:52:33,230][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.21017268300056458, 'train/grad_norm_pre_clip': 110.44304656982422, 'timing/data': 0.000847402960062027, 'timing/model': 1.691956183873117, 'learning_rate/qwen_vl_interface': 2.2599392984577436e-06, 'learning_rate/action_model': 2.2599392984577438e-05, 'epoch': 1.4}) +[2026-06-23 21:52:36,562][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.2214585840702057, 'train/grad_norm_pre_clip': 98.8135757446289, 'timing/data': 0.0006069601513445377, 'timing/model': 1.6883331770077348, 'learning_rate/qwen_vl_interface': 2.2575735858359436e-06, 'learning_rate/action_model': 2.257573585835944e-05, 'epoch': 1.4}) +[2026-06-23 21:52:39,899][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.057799383997917175, 'train/grad_norm_pre_clip': 21.07708168029785, 'timing/data': 0.0007074391469359398, 'timing/model': 1.6952412729151547, 'learning_rate/qwen_vl_interface': 2.255209103287609e-06, 'learning_rate/action_model': 2.255209103287609e-05, 'epoch': 1.4}) +[2026-06-23 21:52:43,238][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.23270218074321747, 'train/grad_norm_pre_clip': 17.114112854003906, 'timing/data': 0.0005978730041533709, 'timing/model': 1.6961522868368775, 'learning_rate/qwen_vl_interface': 2.2528458517846886e-06, 'learning_rate/action_model': 2.252845851784689e-05, 'epoch': 1.4}) +[2026-06-23 21:52:46,581][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.08562464267015457, 'train/grad_norm_pre_clip': 10.747664451599121, 'timing/data': 0.0005999470595270395, 'timing/model': 1.6959623619914055, 'learning_rate/qwen_vl_interface': 2.250483832298627e-06, 'learning_rate/action_model': 2.250483832298627e-05, 'epoch': 1.4}) +[2026-06-23 21:52:49,923][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.2207394540309906, 'train/grad_norm_pre_clip': 48.96573257446289, 'timing/data': 0.0006018111016601324, 'timing/model': 1.6987933560740203, 'learning_rate/qwen_vl_interface': 2.2481230458003608e-06, 'learning_rate/action_model': 2.2481230458003606e-05, 'epoch': 1.4}) +[2026-06-23 21:52:53,266][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.09529905766248703, 'train/grad_norm_pre_clip': 18.14603042602539, 'timing/data': 0.0006041750311851501, 'timing/model': 1.6983031469862908, 'learning_rate/qwen_vl_interface': 2.2457634932603184e-06, 'learning_rate/action_model': 2.2457634932603183e-05, 'epoch': 1.4}) +[2026-06-23 21:52:56,604][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.047521304339170456, 'train/grad_norm_pre_clip': 15.276793479919434, 'timing/data': 0.0005951090715825558, 'timing/model': 1.6940492840949446, 'learning_rate/qwen_vl_interface': 2.2434051756484255e-06, 'learning_rate/action_model': 2.2434051756484253e-05, 'epoch': 1.4}) +[2026-06-23 21:52:59,954][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.06399542093276978, 'train/grad_norm_pre_clip': 16.193925857543945, 'timing/data': 0.0010302860755473375, 'timing/model': 1.6998294298537076, 'learning_rate/qwen_vl_interface': 2.2410480939340945e-06, 'learning_rate/action_model': 2.2410480939340945e-05, 'epoch': 1.4}) +[2026-06-23 21:53:03,300][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.0937911793589592, 'train/grad_norm_pre_clip': 57.639949798583984, 'timing/data': 0.0005983640439808369, 'timing/model': 1.6992717438843101, 'learning_rate/qwen_vl_interface': 2.238692249086234e-06, 'learning_rate/action_model': 2.2386922490862336e-05, 'epoch': 1.4}) +[2026-06-23 21:53:06,643][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.1002923846244812, 'train/grad_norm_pre_clip': 22.44370460510254, 'timing/data': 0.0005870538298040628, 'timing/model': 1.6968338540755212, 'learning_rate/qwen_vl_interface': 2.2363376420732426e-06, 'learning_rate/action_model': 2.2363376420732423e-05, 'epoch': 1.4}) +[2026-06-23 21:53:09,983][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.17674876749515533, 'train/grad_norm_pre_clip': 46.74126052856445, 'timing/data': 0.0006636469624936581, 'timing/model': 1.6957883809227496, 'learning_rate/qwen_vl_interface': 2.233984273863011e-06, 'learning_rate/action_model': 2.233984273863011e-05, 'epoch': 1.4}) +[2026-06-23 21:53:13,322][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.03737042099237442, 'train/grad_norm_pre_clip': 15.637894630432129, 'timing/data': 0.0005824640393257141, 'timing/model': 1.6957691870629787, 'learning_rate/qwen_vl_interface': 2.231632145422919e-06, 'learning_rate/action_model': 2.231632145422919e-05, 'epoch': 1.4}) +[2026-06-23 21:53:16,671][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.03634975850582123, 'train/grad_norm_pre_clip': 8.777754783630371, 'timing/data': 0.0005996269173920155, 'timing/model': 1.7034297781065106, 'learning_rate/qwen_vl_interface': 2.2292812577198383e-06, 'learning_rate/action_model': 2.229281257719838e-05, 'epoch': 1.4}) +[2026-06-23 21:53:20,110][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.12699271738529205, 'train/grad_norm_pre_clip': 43.58397674560547, 'timing/data': 0.0006006089970469475, 'timing/model': 1.7929484280757606, 'learning_rate/qwen_vl_interface': 2.2269316117201307e-06, 'learning_rate/action_model': 2.2269316117201307e-05, 'epoch': 1.4}) +[2026-06-23 21:53:23,451][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.056764524430036545, 'train/grad_norm_pre_clip': 16.389970779418945, 'timing/data': 0.0005970718339085579, 'timing/model': 1.6985228890553117, 'learning_rate/qwen_vl_interface': 2.224583208389646e-06, 'learning_rate/action_model': 2.224583208389646e-05, 'epoch': 1.4}) +[2026-06-23 21:53:26,798][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.11667940020561218, 'train/grad_norm_pre_clip': 33.89575958251953, 'timing/data': 0.0005820139776915312, 'timing/model': 1.697081923019141, 'learning_rate/qwen_vl_interface': 2.2222360486937238e-06, 'learning_rate/action_model': 2.2222360486937235e-05, 'epoch': 1.4}) +[2026-06-23 21:53:30,143][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.08750627189874649, 'train/grad_norm_pre_clip': 24.79844093322754, 'timing/data': 0.0006466361228376627, 'timing/model': 1.698140364838764, 'learning_rate/qwen_vl_interface': 2.219890133597194e-06, 'learning_rate/action_model': 2.219890133597194e-05, 'epoch': 1.4}) +[2026-06-23 21:53:33,485][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.08438734710216522, 'train/grad_norm_pre_clip': 28.83987045288086, 'timing/data': 0.0007650679908692837, 'timing/model': 1.6996838280465454, 'learning_rate/qwen_vl_interface': 2.217545464064374e-06, 'learning_rate/action_model': 2.217545464064374e-05, 'epoch': 1.4}) +[2026-06-23 21:53:36,825][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.12206295877695084, 'train/grad_norm_pre_clip': 26.309545516967773, 'timing/data': 0.0005983649753034115, 'timing/model': 1.6965739459265023, 'learning_rate/qwen_vl_interface': 2.215202041059067e-06, 'learning_rate/action_model': 2.215202041059067e-05, 'epoch': 1.41}) +[2026-06-23 21:53:40,168][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.16614767909049988, 'train/grad_norm_pre_clip': 50.369667053222656, 'timing/data': 0.0007448699325323105, 'timing/model': 1.6951565130148083, 'learning_rate/qwen_vl_interface': 2.212859865544566e-06, 'learning_rate/action_model': 2.212859865544566e-05, 'epoch': 1.41}) +[2026-06-23 21:53:43,509][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.21447601914405823, 'train/grad_norm_pre_clip': 26.259416580200195, 'timing/data': 0.0006554620340466499, 'timing/model': 1.6937571151647717, 'learning_rate/qwen_vl_interface': 2.210518938483652e-06, 'learning_rate/action_model': 2.210518938483652e-05, 'epoch': 1.41}) +[2026-06-23 21:53:46,850][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.05880131945014, 'train/grad_norm_pre_clip': 23.06735610961914, 'timing/data': 0.000607962952926755, 'timing/model': 1.6973918101284653, 'learning_rate/qwen_vl_interface': 2.20817926083859e-06, 'learning_rate/action_model': 2.20817926083859e-05, 'epoch': 1.41}) +[2026-06-23 21:53:50,194][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.07121291011571884, 'train/grad_norm_pre_clip': 36.28335189819336, 'timing/data': 0.0007375171408057213, 'timing/model': 1.6962141778785735, 'learning_rate/qwen_vl_interface': 2.205840833571135e-06, 'learning_rate/action_model': 2.205840833571135e-05, 'epoch': 1.41}) +[2026-06-23 21:53:53,543][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.10411053150892258, 'train/grad_norm_pre_clip': 23.18983268737793, 'timing/data': 0.0006510838866233826, 'timing/model': 1.7007036269642413, 'learning_rate/qwen_vl_interface': 2.2035036576425264e-06, 'learning_rate/action_model': 2.203503657642526e-05, 'epoch': 1.41}) +[2026-06-23 21:53:56,890][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.1335078924894333, 'train/grad_norm_pre_clip': 26.579383850097656, 'timing/data': 0.000599816907197237, 'timing/model': 1.701491710031405, 'learning_rate/qwen_vl_interface': 2.2011677340134867e-06, 'learning_rate/action_model': 2.2011677340134867e-05, 'epoch': 1.41}) +[2026-06-23 21:54:00,231][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.05645741522312164, 'train/grad_norm_pre_clip': 42.670833587646484, 'timing/data': 0.0005978429690003395, 'timing/model': 1.6969001330435276, 'learning_rate/qwen_vl_interface': 2.198833063644227e-06, 'learning_rate/action_model': 2.198833063644227e-05, 'epoch': 1.41}) +[2026-06-23 21:54:03,573][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.1729092299938202, 'train/grad_norm_pre_clip': 58.72043228149414, 'timing/data': 0.0006960080936551094, 'timing/model': 1.692771164001897, 'learning_rate/qwen_vl_interface': 2.196499647494443e-06, 'learning_rate/action_model': 2.196499647494443e-05, 'epoch': 1.41}) +[2026-06-23 21:54:06,914][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.14968134462833405, 'train/grad_norm_pre_clip': 20.64566993713379, 'timing/data': 0.0007609000895172358, 'timing/model': 1.699215428205207, 'learning_rate/qwen_vl_interface': 2.1941674865233144e-06, 'learning_rate/action_model': 2.1941674865233143e-05, 'epoch': 1.41}) +[2026-06-23 21:54:10,355][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.12028630822896957, 'train/grad_norm_pre_clip': 25.698394775390625, 'timing/data': 0.0006945950444787741, 'timing/model': 1.7009558209683746, 'learning_rate/qwen_vl_interface': 2.1918365816895013e-06, 'learning_rate/action_model': 2.191836581689501e-05, 'epoch': 1.41}) +[2026-06-23 21:54:13,697][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.10073879361152649, 'train/grad_norm_pre_clip': 43.09922409057617, 'timing/data': 0.0009933060500770807, 'timing/model': 1.6987556379754096, 'learning_rate/qwen_vl_interface': 2.189506933951153e-06, 'learning_rate/action_model': 2.1895069339511525e-05, 'epoch': 1.41}) +[2026-06-23 21:54:17,040][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.06495855003595352, 'train/grad_norm_pre_clip': 25.008771896362305, 'timing/data': 0.0007544378750026226, 'timing/model': 1.6978761868085712, 'learning_rate/qwen_vl_interface': 2.187178544265898e-06, 'learning_rate/action_model': 2.1871785442658983e-05, 'epoch': 1.41}) +[2026-06-23 21:54:20,387][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.17418761551380157, 'train/grad_norm_pre_clip': 53.971824645996094, 'timing/data': 0.0005993160884827375, 'timing/model': 1.700671962928027, 'learning_rate/qwen_vl_interface': 2.184851413590852e-06, 'learning_rate/action_model': 2.1848514135908518e-05, 'epoch': 1.41}) +[2026-06-23 21:54:23,733][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.13553160429000854, 'train/grad_norm_pre_clip': 38.05967330932617, 'timing/data': 0.000597883015871048, 'timing/model': 1.6970976570155472, 'learning_rate/qwen_vl_interface': 2.1825255428826087e-06, 'learning_rate/action_model': 2.182525542882609e-05, 'epoch': 1.41}) +[2026-06-23 21:54:27,079][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.08728274703025818, 'train/grad_norm_pre_clip': 31.973247528076172, 'timing/data': 0.0005890668835490942, 'timing/model': 1.6981966050807387, 'learning_rate/qwen_vl_interface': 2.180200933097246e-06, 'learning_rate/action_model': 2.180200933097246e-05, 'epoch': 1.41}) +[2026-06-23 21:54:30,423][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.15533971786499023, 'train/grad_norm_pre_clip': 54.95405197143555, 'timing/data': 0.000588095048442483, 'timing/model': 1.6968171030748636, 'learning_rate/qwen_vl_interface': 2.1778775851903235e-06, 'learning_rate/action_model': 2.177877585190323e-05, 'epoch': 1.41}) +[2026-06-23 21:54:33,766][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.07869691401720047, 'train/grad_norm_pre_clip': 15.825018882751465, 'timing/data': 0.0005998779088258743, 'timing/model': 1.6967689248267561, 'learning_rate/qwen_vl_interface': 2.175555500116881e-06, 'learning_rate/action_model': 2.175555500116881e-05, 'epoch': 1.41}) +[2026-06-23 21:54:37,109][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.10969964414834976, 'train/grad_norm_pre_clip': 29.202547073364258, 'timing/data': 0.0005981340073049068, 'timing/model': 1.6986414468847215, 'learning_rate/qwen_vl_interface': 2.1732346788314404e-06, 'learning_rate/action_model': 2.1732346788314402e-05, 'epoch': 1.41}) +[2026-06-23 21:54:40,450][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.17990027368068695, 'train/grad_norm_pre_clip': 25.234241485595703, 'timing/data': 0.000600127037614584, 'timing/model': 1.6947481199167669, 'learning_rate/qwen_vl_interface': 2.170915122288006e-06, 'learning_rate/action_model': 2.170915122288006e-05, 'epoch': 1.41}) +[2026-06-23 21:54:43,794][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.05072658509016037, 'train/grad_norm_pre_clip': 25.033479690551758, 'timing/data': 0.0005978140980005264, 'timing/model': 1.694174183998257, 'learning_rate/qwen_vl_interface': 2.168596831440055e-06, 'learning_rate/action_model': 2.1685968314400552e-05, 'epoch': 1.41}) +[2026-06-23 21:54:47,141][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.0950804278254509, 'train/grad_norm_pre_clip': 27.174495697021484, 'timing/data': 0.0006717531941831112, 'timing/model': 1.6954877339303493, 'learning_rate/qwen_vl_interface': 2.166279807240553e-06, 'learning_rate/action_model': 2.1662798072405528e-05, 'epoch': 1.41}) +[2026-06-23 21:54:50,487][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.14221808314323425, 'train/grad_norm_pre_clip': 20.4237003326416, 'timing/data': 0.0008265229407697916, 'timing/model': 1.6979787580203265, 'learning_rate/qwen_vl_interface': 2.1639640506419394e-06, 'learning_rate/action_model': 2.163964050641939e-05, 'epoch': 1.41}) +[2026-06-23 21:54:53,836][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.1477787345647812, 'train/grad_norm_pre_clip': 23.583017349243164, 'timing/data': 0.000602601794525981, 'timing/model': 1.6996165988966823, 'learning_rate/qwen_vl_interface': 2.161649562596135e-06, 'learning_rate/action_model': 2.161649562596135e-05, 'epoch': 1.41}) +[2026-06-23 21:54:57,182][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.1271073967218399, 'train/grad_norm_pre_clip': 15.816243171691895, 'timing/data': 0.0005883448757231236, 'timing/model': 1.6975037478841841, 'learning_rate/qwen_vl_interface': 2.1593363440545383e-06, 'learning_rate/action_model': 2.1593363440545384e-05, 'epoch': 1.41}) +[2026-06-23 21:55:00,630][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.09808428585529327, 'train/grad_norm_pre_clip': 45.76405334472656, 'timing/data': 0.0005989361088722944, 'timing/model': 1.6937640719115734, 'learning_rate/qwen_vl_interface': 2.157024395968025e-06, 'learning_rate/action_model': 2.1570243959680252e-05, 'epoch': 1.41}) +[2026-06-23 21:55:03,976][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.09652054309844971, 'train/grad_norm_pre_clip': 37.8050537109375, 'timing/data': 0.0007738040294498205, 'timing/model': 1.698805107967928, 'learning_rate/qwen_vl_interface': 2.154713719286952e-06, 'learning_rate/action_model': 2.154713719286952e-05, 'epoch': 1.42}) +[2026-06-23 21:55:07,362][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.1138548031449318, 'train/grad_norm_pre_clip': 17.538816452026367, 'timing/data': 0.000597864156588912, 'timing/model': 1.7014219688717276, 'learning_rate/qwen_vl_interface': 2.152404314961149e-06, 'learning_rate/action_model': 2.152404314961149e-05, 'epoch': 1.42}) +[2026-06-23 21:55:10,715][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.08456441760063171, 'train/grad_norm_pre_clip': 17.84701156616211, 'timing/data': 0.0007334889378398657, 'timing/model': 1.6974278229754418, 'learning_rate/qwen_vl_interface': 2.1500961839399264e-06, 'learning_rate/action_model': 2.1500961839399262e-05, 'epoch': 1.42}) +[2026-06-23 21:55:14,082][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.05775853246450424, 'train/grad_norm_pre_clip': 23.26305389404297, 'timing/data': 0.0005867029540240765, 'timing/model': 1.7162944450974464, 'learning_rate/qwen_vl_interface': 2.1477893271720683e-06, 'learning_rate/action_model': 2.1477893271720682e-05, 'epoch': 1.42}) +[2026-06-23 21:55:17,428][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.11541782319545746, 'train/grad_norm_pre_clip': 17.650135040283203, 'timing/data': 0.0006896061822772026, 'timing/model': 1.6972902179695666, 'learning_rate/qwen_vl_interface': 2.1454837456058385e-06, 'learning_rate/action_model': 2.1454837456058386e-05, 'epoch': 1.42}) +[2026-06-23 21:55:20,776][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.05103759095072746, 'train/grad_norm_pre_clip': 25.69538116455078, 'timing/data': 0.0006998861208558083, 'timing/model': 1.699467201018706, 'learning_rate/qwen_vl_interface': 2.1431794401889722e-06, 'learning_rate/action_model': 2.143179440188972e-05, 'epoch': 1.42}) +[2026-06-23 21:55:24,126][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.16398461163043976, 'train/grad_norm_pre_clip': 27.627748489379883, 'timing/data': 0.0006031740922480822, 'timing/model': 1.6989513530861586, 'learning_rate/qwen_vl_interface': 2.1408764118686837e-06, 'learning_rate/action_model': 2.1408764118686838e-05, 'epoch': 1.42}) +[2026-06-23 21:55:27,472][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.15631034970283508, 'train/grad_norm_pre_clip': 20.60443115234375, 'timing/data': 0.0007501300424337387, 'timing/model': 1.697323075029999, 'learning_rate/qwen_vl_interface': 2.1385746615916598e-06, 'learning_rate/action_model': 2.1385746615916596e-05, 'epoch': 1.42}) +[2026-06-23 21:55:30,822][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.049406714737415314, 'train/grad_norm_pre_clip': 14.279974937438965, 'timing/data': 0.0006132530979812145, 'timing/model': 1.7030129111371934, 'learning_rate/qwen_vl_interface': 2.1362741903040636e-06, 'learning_rate/action_model': 2.1362741903040633e-05, 'epoch': 1.42}) +[2026-06-23 21:55:34,185][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.16640031337738037, 'train/grad_norm_pre_clip': 94.28040313720703, 'timing/data': 0.0009330138564109802, 'timing/model': 1.704396473011002, 'learning_rate/qwen_vl_interface': 2.1339749989515318e-06, 'learning_rate/action_model': 2.1339749989515316e-05, 'epoch': 1.42}) +[2026-06-23 21:55:37,541][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.25578874349594116, 'train/grad_norm_pre_clip': 51.80487823486328, 'timing/data': 0.0007143719121813774, 'timing/model': 1.7071227577980608, 'learning_rate/qwen_vl_interface': 2.131677088479175e-06, 'learning_rate/action_model': 2.1316770884791752e-05, 'epoch': 1.42}) +[2026-06-23 21:55:40,888][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.12435664236545563, 'train/grad_norm_pre_clip': 42.82122802734375, 'timing/data': 0.000642398139461875, 'timing/model': 1.6983809899538755, 'learning_rate/qwen_vl_interface': 2.1293804598315787e-06, 'learning_rate/action_model': 2.1293804598315787e-05, 'epoch': 1.42}) +[2026-06-23 21:55:44,240][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.050679370760917664, 'train/grad_norm_pre_clip': 49.05284118652344, 'timing/data': 0.0006634867750108242, 'timing/model': 1.702877172967419, 'learning_rate/qwen_vl_interface': 2.1270851139527986e-06, 'learning_rate/action_model': 2.1270851139527983e-05, 'epoch': 1.42}) +[2026-06-23 21:55:47,684][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.07489502429962158, 'train/grad_norm_pre_clip': 11.817956924438477, 'timing/data': 0.0005895779468119144, 'timing/model': 1.7969290069304407, 'learning_rate/qwen_vl_interface': 2.1247910517863653e-06, 'learning_rate/action_model': 2.124791051786365e-05, 'epoch': 1.42}) +[2026-06-23 21:55:51,027][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.3305872082710266, 'train/grad_norm_pre_clip': 55.27070999145508, 'timing/data': 0.0007272670045495033, 'timing/model': 1.695639411918819, 'learning_rate/qwen_vl_interface': 2.1224982742752814e-06, 'learning_rate/action_model': 2.1224982742752812e-05, 'epoch': 1.42}) +[2026-06-23 21:55:54,370][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.10380283743143082, 'train/grad_norm_pre_clip': 57.17641067504883, 'timing/data': 0.0006045668851584196, 'timing/model': 1.6975467780139297, 'learning_rate/qwen_vl_interface': 2.1202067823620213e-06, 'learning_rate/action_model': 2.1202067823620212e-05, 'epoch': 1.42}) +[2026-06-23 21:55:57,707][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.14436765015125275, 'train/grad_norm_pre_clip': 38.99695587158203, 'timing/data': 0.0005962599534541368, 'timing/model': 1.694443317130208, 'learning_rate/qwen_vl_interface': 2.117916576988531e-06, 'learning_rate/action_model': 2.117916576988531e-05, 'epoch': 1.42}) +[2026-06-23 21:56:01,049][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.056435927748680115, 'train/grad_norm_pre_clip': 18.998186111450195, 'timing/data': 0.0005933649372309446, 'timing/model': 1.69703758507967, 'learning_rate/qwen_vl_interface': 2.1156276590962257e-06, 'learning_rate/action_model': 2.1156276590962257e-05, 'epoch': 1.42}) +[2026-06-23 21:56:04,392][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.07348513603210449, 'train/grad_norm_pre_clip': 14.35720443725586, 'timing/data': 0.0006436097901314497, 'timing/model': 1.698527887929231, 'learning_rate/qwen_vl_interface': 2.113340029625994e-06, 'learning_rate/action_model': 2.1133400296259938e-05, 'epoch': 1.42}) +[2026-06-23 21:56:07,738][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.1130913645029068, 'train/grad_norm_pre_clip': 29.134363174438477, 'timing/data': 0.0006017209962010384, 'timing/model': 1.7014761799946427, 'learning_rate/qwen_vl_interface': 2.111053689518194e-06, 'learning_rate/action_model': 2.111053689518194e-05, 'epoch': 1.42}) +[2026-06-23 21:56:11,073][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.08324550092220306, 'train/grad_norm_pre_clip': 32.39763641357422, 'timing/data': 0.0005954590160399675, 'timing/model': 1.694655061000958, 'learning_rate/qwen_vl_interface': 2.1087686397126544e-06, 'learning_rate/action_model': 2.1087686397126545e-05, 'epoch': 1.42}) +[2026-06-23 21:56:14,422][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.06048224866390228, 'train/grad_norm_pre_clip': 25.8182430267334, 'timing/data': 0.0005988760385662317, 'timing/model': 1.699967317050323, 'learning_rate/qwen_vl_interface': 2.106484881148672e-06, 'learning_rate/action_model': 2.1064848811486718e-05, 'epoch': 1.42}) +[2026-06-23 21:56:17,778][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.13455721735954285, 'train/grad_norm_pre_clip': 84.78024291992188, 'timing/data': 0.0006163180805742741, 'timing/model': 1.7066758470609784, 'learning_rate/qwen_vl_interface': 2.1042024147650137e-06, 'learning_rate/action_model': 2.1042024147650137e-05, 'epoch': 1.42}) +[2026-06-23 21:56:21,135][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.23333679139614105, 'train/grad_norm_pre_clip': 62.17351531982422, 'timing/data': 0.0005913909990340471, 'timing/model': 1.707549351034686, 'learning_rate/qwen_vl_interface': 2.1019212414999163e-06, 'learning_rate/action_model': 2.101921241499916e-05, 'epoch': 1.42}) +[2026-06-23 21:56:24,482][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.0597572922706604, 'train/grad_norm_pre_clip': 22.748369216918945, 'timing/data': 0.0006095049902796745, 'timing/model': 1.7010656259953976, 'learning_rate/qwen_vl_interface': 2.099641362291082e-06, 'learning_rate/action_model': 2.099641362291082e-05, 'epoch': 1.42}) +[2026-06-23 21:56:27,838][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.15986832976341248, 'train/grad_norm_pre_clip': 50.70623779296875, 'timing/data': 0.0008415118791162968, 'timing/model': 1.7007106710225344, 'learning_rate/qwen_vl_interface': 2.097362778075685e-06, 'learning_rate/action_model': 2.097362778075685e-05, 'epoch': 1.43}) +[2026-06-23 21:56:31,182][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.09568707644939423, 'train/grad_norm_pre_clip': 40.11315155029297, 'timing/data': 0.0005967309698462486, 'timing/model': 1.7000119991134852, 'learning_rate/qwen_vl_interface': 2.0950854897903663e-06, 'learning_rate/action_model': 2.095085489790366e-05, 'epoch': 1.43}) +[2026-06-23 21:56:34,525][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.18859605491161346, 'train/grad_norm_pre_clip': 39.17511749267578, 'timing/data': 0.0006907181814312935, 'timing/model': 1.6998669079039246, 'learning_rate/qwen_vl_interface': 2.092809498371229e-06, 'learning_rate/action_model': 2.0928094983712292e-05, 'epoch': 1.43}) +[2026-06-23 21:56:38,179][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.12693800032138824, 'train/grad_norm_pre_clip': 31.17275619506836, 'timing/data': 0.0006436100229620934, 'timing/model': 2.0073289449792355, 'learning_rate/qwen_vl_interface': 2.0905348047538498e-06, 'learning_rate/action_model': 2.0905348047538497e-05, 'epoch': 1.43}) +[2026-06-23 21:56:41,532][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.14818985760211945, 'train/grad_norm_pre_clip': 15.703669548034668, 'timing/data': 0.0005979740526527166, 'timing/model': 1.7034690720029175, 'learning_rate/qwen_vl_interface': 2.0882614098732686e-06, 'learning_rate/action_model': 2.0882614098732685e-05, 'epoch': 1.43}) +[2026-06-23 21:56:44,873][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.1257343292236328, 'train/grad_norm_pre_clip': 41.0655403137207, 'timing/data': 0.0006028939969837666, 'timing/model': 1.6962309279479086, 'learning_rate/qwen_vl_interface': 2.085989314663993e-06, 'learning_rate/action_model': 2.0859893146639927e-05, 'epoch': 1.43}) +[2026-06-23 21:56:48,211][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.08960223942995071, 'train/grad_norm_pre_clip': 14.826285362243652, 'timing/data': 0.000597743783146143, 'timing/model': 1.6920241629704833, 'learning_rate/qwen_vl_interface': 2.083718520059995e-06, 'learning_rate/action_model': 2.0837185200599946e-05, 'epoch': 1.43}) +[2026-06-23 21:56:51,554][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.06517477333545685, 'train/grad_norm_pre_clip': 51.59471893310547, 'timing/data': 0.0006686069536954165, 'timing/model': 1.6975533470977098, 'learning_rate/qwen_vl_interface': 2.0814490269947117e-06, 'learning_rate/action_model': 2.0814490269947118e-05, 'epoch': 1.43}) +[2026-06-23 21:56:54,901][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.07966197282075882, 'train/grad_norm_pre_clip': 54.7094612121582, 'timing/data': 0.000603014137595892, 'timing/model': 1.6989683371502906, 'learning_rate/qwen_vl_interface': 2.079180836401047e-06, 'learning_rate/action_model': 2.079180836401047e-05, 'epoch': 1.43}) +[2026-06-23 21:56:58,246][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.14531472325325012, 'train/grad_norm_pre_clip': 40.04571533203125, 'timing/data': 0.0005858619697391987, 'timing/model': 1.6986887331586331, 'learning_rate/qwen_vl_interface': 2.0769139492113687e-06, 'learning_rate/action_model': 2.076913949211369e-05, 'epoch': 1.43}) +[2026-06-23 21:57:01,584][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.10127630829811096, 'train/grad_norm_pre_clip': 19.89362907409668, 'timing/data': 0.0005906200967729092, 'timing/model': 1.6977195339277387, 'learning_rate/qwen_vl_interface': 2.0746483663575075e-06, 'learning_rate/action_model': 2.0746483663575076e-05, 'epoch': 1.43}) +[2026-06-23 21:57:04,940][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.11719835549592972, 'train/grad_norm_pre_clip': 37.54124069213867, 'timing/data': 0.0006010090000927448, 'timing/model': 1.7082085891161114, 'learning_rate/qwen_vl_interface': 2.0723840887707595e-06, 'learning_rate/action_model': 2.0723840887707594e-05, 'epoch': 1.43}) +[2026-06-23 21:57:08,280][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.08754761517047882, 'train/grad_norm_pre_clip': 44.97873306274414, 'timing/data': 0.0005949989426881075, 'timing/model': 1.697787013137713, 'learning_rate/qwen_vl_interface': 2.0701211173818854e-06, 'learning_rate/action_model': 2.0701211173818853e-05, 'epoch': 1.43}) +[2026-06-23 21:57:11,624][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.08870784193277359, 'train/grad_norm_pre_clip': 37.695068359375, 'timing/data': 0.0005991158541291952, 'timing/model': 1.701777715003118, 'learning_rate/qwen_vl_interface': 2.0678594531211036e-06, 'learning_rate/action_model': 2.0678594531211033e-05, 'epoch': 1.43}) +[2026-06-23 21:57:14,963][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.23805153369903564, 'train/grad_norm_pre_clip': 47.65687942504883, 'timing/data': 0.0006007491610944271, 'timing/model': 1.696258872980252, 'learning_rate/qwen_vl_interface': 2.0655990969181e-06, 'learning_rate/action_model': 2.0655990969181003e-05, 'epoch': 1.43}) +[2026-06-23 21:57:18,312][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.1026846319437027, 'train/grad_norm_pre_clip': 14.125824928283691, 'timing/data': 0.0006877928972244263, 'timing/model': 1.6994465060997754, 'learning_rate/qwen_vl_interface': 2.0633400497020236e-06, 'learning_rate/action_model': 2.0633400497020232e-05, 'epoch': 1.43}) +[2026-06-23 21:57:21,666][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.12898969650268555, 'train/grad_norm_pre_clip': 36.7559700012207, 'timing/data': 0.0006583069916814566, 'timing/model': 1.7002256920095533, 'learning_rate/qwen_vl_interface': 2.061082312401481e-06, 'learning_rate/action_model': 2.061082312401481e-05, 'epoch': 1.43}) +[2026-06-23 21:57:25,016][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.23344136774539948, 'train/grad_norm_pre_clip': 69.56010437011719, 'timing/data': 0.0006976709701120853, 'timing/model': 1.7011362749617547, 'learning_rate/qwen_vl_interface': 2.058825885944544e-06, 'learning_rate/action_model': 2.0588258859445442e-05, 'epoch': 1.43}) +[2026-06-23 21:57:28,672][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3700 +[2026-06-23 21:59:02,790][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3700 +[2026-06-23 21:59:02,791][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3700 +[2026-06-23 22:00:14,452][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3700 +[2026-06-23 22:00:14,453][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.0719832256436348, 'train/grad_norm_pre_clip': 28.428075790405273, 'eval/action_loss/seconds': 94.11603509495035, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.07269567996263504, 'eval/latency_2/loss': 0.07269567996263504, 'eval/flappy/loss': 0.07269567996263504, 'eval/flappy/latency_2/loss': 0.07269567996263504, 'eval/flap_precision': 0.9015673981191222, 'eval/flap_recall': 0.970310391363023, 'eval/flap_f1': 0.9346766330841729, 'eval/flap_rate_model': 0.22152777777777777, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9346766330841729, 'eval/latency_2/flap_precision': 0.9015673981191222, 'eval/latency_2/flap_recall': 0.970310391363023, 'eval/latency_2/flap_f1': 0.9346766330841729, 'eval/latency_2/flap_rate_model': 0.22152777777777777, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9346766330841729, 'eval/action_classification/data_wait_seconds': 1.0259373188018799, 'eval/action_classification/predict_seconds': 70.47058868408203, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0008608580101281404, 'timing/model': 2.00822193804197, 'learning_rate/qwen_vl_interface': 2.0565707712587447e-06, 'learning_rate/action_model': 2.0565707712587444e-05, 'epoch': 1.43}) +[2026-06-23 22:00:17,773][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.11875590682029724, 'train/grad_norm_pre_clip': 51.57381820678711, 'timing/data': 0.0007631238549947739, 'timing/model': 1.6862079850398004, 'learning_rate/qwen_vl_interface': 2.0543169692710747e-06, 'learning_rate/action_model': 2.054316969271075e-05, 'epoch': 1.43}) +[2026-06-23 22:00:21,095][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.026996653527021408, 'train/grad_norm_pre_clip': 32.651851654052734, 'timing/data': 0.0006630669813603163, 'timing/model': 1.686907229013741, 'learning_rate/qwen_vl_interface': 2.0520644809079873e-06, 'learning_rate/action_model': 2.0520644809079873e-05, 'epoch': 1.43}) +[2026-06-23 22:00:24,418][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.17975124716758728, 'train/grad_norm_pre_clip': 56.92326354980469, 'timing/data': 0.0006551318801939487, 'timing/model': 1.690481117926538, 'learning_rate/qwen_vl_interface': 2.049813307095395e-06, 'learning_rate/action_model': 2.0498133070953947e-05, 'epoch': 1.43}) +[2026-06-23 22:00:27,743][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.03781038522720337, 'train/grad_norm_pre_clip': 26.50094223022461, 'timing/data': 0.0005942978896200657, 'timing/model': 1.6898102180566639, 'learning_rate/qwen_vl_interface': 2.047563448758672e-06, 'learning_rate/action_model': 2.0475634487586717e-05, 'epoch': 1.43}) +[2026-06-23 22:00:31,073][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.08264371752738953, 'train/grad_norm_pre_clip': 32.25498580932617, 'timing/data': 0.0006397019606083632, 'timing/model': 1.6941567191388458, 'learning_rate/qwen_vl_interface': 2.045314906822647e-06, 'learning_rate/action_model': 2.0453149068226467e-05, 'epoch': 1.43}) +[2026-06-23 22:00:34,397][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.06767357885837555, 'train/grad_norm_pre_clip': 15.41801643371582, 'timing/data': 0.0005876549985259771, 'timing/model': 1.6902182861231267, 'learning_rate/qwen_vl_interface': 2.0430676822116124e-06, 'learning_rate/action_model': 2.0430676822116123e-05, 'epoch': 1.43}) +[2026-06-23 22:00:37,717][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.25094515085220337, 'train/grad_norm_pre_clip': 28.94192886352539, 'timing/data': 0.0006064102053642273, 'timing/model': 1.6885951068252325, 'learning_rate/qwen_vl_interface': 2.040821775849319e-06, 'learning_rate/action_model': 2.0408217758493188e-05, 'epoch': 1.43}) +[2026-06-23 22:00:41,130][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.0546758659183979, 'train/grad_norm_pre_clip': 12.772658348083496, 'timing/data': 0.0006926821079105139, 'timing/model': 1.6887238388881087, 'learning_rate/qwen_vl_interface': 2.03857718865897e-06, 'learning_rate/action_model': 2.0385771886589702e-05, 'epoch': 1.44}) +[2026-06-23 22:00:44,461][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.053622350096702576, 'train/grad_norm_pre_clip': 31.755971908569336, 'timing/data': 0.0008259019814431667, 'timing/model': 1.6936909661162645, 'learning_rate/qwen_vl_interface': 2.0363339215632337e-06, 'learning_rate/action_model': 2.0363339215632334e-05, 'epoch': 1.44}) +[2026-06-23 22:00:47,785][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.06476259976625443, 'train/grad_norm_pre_clip': 36.27961349487305, 'timing/data': 0.0006090239621698856, 'timing/model': 1.6880905509460717, 'learning_rate/qwen_vl_interface': 2.034091975484231e-06, 'learning_rate/action_model': 2.0340919754842307e-05, 'epoch': 1.44}) +[2026-06-23 22:00:51,110][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.1543690413236618, 'train/grad_norm_pre_clip': 51.98048782348633, 'timing/data': 0.0005891569890081882, 'timing/model': 1.6885547849815339, 'learning_rate/qwen_vl_interface': 2.0318513513435415e-06, 'learning_rate/action_model': 2.0318513513435415e-05, 'epoch': 1.44}) +[2026-06-23 22:00:54,434][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.05593525245785713, 'train/grad_norm_pre_clip': 12.875115394592285, 'timing/data': 0.0006115389987826347, 'timing/model': 1.6878678139764816, 'learning_rate/qwen_vl_interface': 2.0296120500622018e-06, 'learning_rate/action_model': 2.0296120500622017e-05, 'epoch': 1.44}) +[2026-06-23 22:00:57,756][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.08185871690511703, 'train/grad_norm_pre_clip': 16.054121017456055, 'timing/data': 0.0005977838300168514, 'timing/model': 1.6855401298962533, 'learning_rate/qwen_vl_interface': 2.0273740725607034e-06, 'learning_rate/action_model': 2.0273740725607032e-05, 'epoch': 1.44}) +[2026-06-23 22:01:01,075][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.07984474301338196, 'train/grad_norm_pre_clip': 28.16472625732422, 'timing/data': 0.0006570150144398212, 'timing/model': 1.6859309449791908, 'learning_rate/qwen_vl_interface': 2.0251374197589942e-06, 'learning_rate/action_model': 2.0251374197589943e-05, 'epoch': 1.44}) +[2026-06-23 22:01:04,396][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.23984141647815704, 'train/grad_norm_pre_clip': 34.99618911743164, 'timing/data': 0.0008265639189630747, 'timing/model': 1.686669104034081, 'learning_rate/qwen_vl_interface': 2.0229020925764785e-06, 'learning_rate/action_model': 2.0229020925764785e-05, 'epoch': 1.44}) +[2026-06-23 22:01:07,724][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.13171616196632385, 'train/grad_norm_pre_clip': 32.91958999633789, 'timing/data': 0.0006175409071147442, 'timing/model': 1.6899622539058328, 'learning_rate/qwen_vl_interface': 2.0206680919320145e-06, 'learning_rate/action_model': 2.020668091932014e-05, 'epoch': 1.44}) +[2026-06-23 22:01:11,049][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.1172744557261467, 'train/grad_norm_pre_clip': 17.292076110839844, 'timing/data': 0.0006189839914441109, 'timing/model': 1.6901518309023231, 'learning_rate/qwen_vl_interface': 2.0184354187439164e-06, 'learning_rate/action_model': 2.0184354187439162e-05, 'epoch': 1.44}) +[2026-06-23 22:01:14,378][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.09234966337680817, 'train/grad_norm_pre_clip': 30.43029022216797, 'timing/data': 0.0005912119522690773, 'timing/model': 1.6918468079529703, 'learning_rate/qwen_vl_interface': 2.016204073929949e-06, 'learning_rate/action_model': 2.016204073929949e-05, 'epoch': 1.44}) +[2026-06-23 22:01:17,716][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.11118467897176743, 'train/grad_norm_pre_clip': 22.69049072265625, 'timing/data': 0.0009241469670087099, 'timing/model': 1.6935626540798694, 'learning_rate/qwen_vl_interface': 2.0139740584073363e-06, 'learning_rate/action_model': 2.0139740584073364e-05, 'epoch': 1.44}) +[2026-06-23 22:01:21,046][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.17672684788703918, 'train/grad_norm_pre_clip': 36.56354904174805, 'timing/data': 0.000603043008595705, 'timing/model': 1.6897003650665283, 'learning_rate/qwen_vl_interface': 2.011745373092753e-06, 'learning_rate/action_model': 2.011745373092753e-05, 'epoch': 1.44}) +[2026-06-23 22:01:24,375][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.04457535967230797, 'train/grad_norm_pre_clip': 15.36903190612793, 'timing/data': 0.0006422169972211123, 'timing/model': 1.6929796088952571, 'learning_rate/qwen_vl_interface': 2.0095180189023267e-06, 'learning_rate/action_model': 2.0095180189023265e-05, 'epoch': 1.44}) +[2026-06-23 22:01:27,801][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.160737082362175, 'train/grad_norm_pre_clip': 46.06181716918945, 'timing/data': 0.0005922031123191118, 'timing/model': 1.7915834761224687, 'learning_rate/qwen_vl_interface': 2.0072919967516393e-06, 'learning_rate/action_model': 2.0072919967516392e-05, 'epoch': 1.44}) +[2026-06-23 22:01:31,129][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.1197737380862236, 'train/grad_norm_pre_clip': 30.097576141357422, 'timing/data': 0.0005873129703104496, 'timing/model': 1.691630044952035, 'learning_rate/qwen_vl_interface': 2.0050673075557245e-06, 'learning_rate/action_model': 2.0050673075557243e-05, 'epoch': 1.44}) +[2026-06-23 22:01:34,461][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.061050958931446075, 'train/grad_norm_pre_clip': 17.219327926635742, 'timing/data': 0.000654981005936861, 'timing/model': 1.6918441478628665, 'learning_rate/qwen_vl_interface': 2.0028439522290677e-06, 'learning_rate/action_model': 2.0028439522290676e-05, 'epoch': 1.44}) +[2026-06-23 22:01:37,788][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.1505298912525177, 'train/grad_norm_pre_clip': 19.688344955444336, 'timing/data': 0.0006734358612447977, 'timing/model': 1.690247057005763, 'learning_rate/qwen_vl_interface': 2.000621931685606e-06, 'learning_rate/action_model': 2.000621931685606e-05, 'epoch': 1.44}) +[2026-06-23 22:01:41,118][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.10887765884399414, 'train/grad_norm_pre_clip': 69.25499725341797, 'timing/data': 0.0006018318235874176, 'timing/model': 1.691630687797442, 'learning_rate/qwen_vl_interface': 1.9984012468387285e-06, 'learning_rate/action_model': 1.9984012468387285e-05, 'epoch': 1.44}) +[2026-06-23 22:01:44,455][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.15413160622119904, 'train/grad_norm_pre_clip': 21.39615249633789, 'timing/data': 0.0006275491323322058, 'timing/model': 1.692616161890328, 'learning_rate/qwen_vl_interface': 1.9961818986012767e-06, 'learning_rate/action_model': 1.9961818986012763e-05, 'epoch': 1.44}) +[2026-06-23 22:01:47,786][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.14936892688274384, 'train/grad_norm_pre_clip': 74.60305786132812, 'timing/data': 0.0007028910331428051, 'timing/model': 1.6938901271205395, 'learning_rate/qwen_vl_interface': 1.9939638878855372e-06, 'learning_rate/action_model': 1.993963887885537e-05, 'epoch': 1.44}) +[2026-06-23 22:01:51,124][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.09208612889051437, 'train/grad_norm_pre_clip': 33.204410552978516, 'timing/data': 0.0006010490469634533, 'timing/model': 1.6954834619536996, 'learning_rate/qwen_vl_interface': 1.991747215603252e-06, 'learning_rate/action_model': 1.991747215603252e-05, 'epoch': 1.44}) +[2026-06-23 22:01:54,459][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.16311918199062347, 'train/grad_norm_pre_clip': 31.789833068847656, 'timing/data': 0.0006011698860675097, 'timing/model': 1.6930899140425026, 'learning_rate/qwen_vl_interface': 1.9895318826656116e-06, 'learning_rate/action_model': 1.9895318826656114e-05, 'epoch': 1.44}) +[2026-06-23 22:01:57,800][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.2303706556558609, 'train/grad_norm_pre_clip': 33.883087158203125, 'timing/data': 0.0008342878427356482, 'timing/model': 1.6969872198533267, 'learning_rate/qwen_vl_interface': 1.9873178899832553e-06, 'learning_rate/action_model': 1.9873178899832554e-05, 'epoch': 1.44}) +[2026-06-23 22:02:01,140][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.06557597219944, 'train/grad_norm_pre_clip': 44.08939743041992, 'timing/data': 0.0006104269996285439, 'timing/model': 1.699401818914339, 'learning_rate/qwen_vl_interface': 1.985105238466272e-06, 'learning_rate/action_model': 1.9851052384662722e-05, 'epoch': 1.44}) +[2026-06-23 22:02:04,479][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.04666871950030327, 'train/grad_norm_pre_clip': 20.85789680480957, 'timing/data': 0.0006072418764233589, 'timing/model': 1.6923315410967916, 'learning_rate/qwen_vl_interface': 1.9828939290242e-06, 'learning_rate/action_model': 1.9828939290241998e-05, 'epoch': 1.44}) +[2026-06-23 22:02:07,812][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.12803128361701965, 'train/grad_norm_pre_clip': 17.962749481201172, 'timing/data': 0.0007023611105978489, 'timing/model': 1.6977101310621947, 'learning_rate/qwen_vl_interface': 1.980683962566024e-06, 'learning_rate/action_model': 1.980683962566024e-05, 'epoch': 1.45}) +[2026-06-23 22:02:11,151][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.15723203122615814, 'train/grad_norm_pre_clip': 35.99197006225586, 'timing/data': 0.0005958289839327335, 'timing/model': 1.6950744879432023, 'learning_rate/qwen_vl_interface': 1.9784753400001784e-06, 'learning_rate/action_model': 1.978475340000178e-05, 'epoch': 1.45}) +[2026-06-23 22:02:14,487][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.20175422728061676, 'train/grad_norm_pre_clip': 58.694358825683594, 'timing/data': 0.000681782141327858, 'timing/model': 1.6942666009999812, 'learning_rate/qwen_vl_interface': 1.9762680622345443e-06, 'learning_rate/action_model': 1.9762680622345442e-05, 'epoch': 1.45}) +[2026-06-23 22:02:17,912][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.07837812602519989, 'train/grad_norm_pre_clip': 45.502410888671875, 'timing/data': 0.0007938419003039598, 'timing/model': 1.6962386080995202, 'learning_rate/qwen_vl_interface': 1.9740621301764513e-06, 'learning_rate/action_model': 1.9740621301764513e-05, 'epoch': 1.45}) +[2026-06-23 22:02:21,244][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.07905696332454681, 'train/grad_norm_pre_clip': 36.74699020385742, 'timing/data': 0.0005824549589306116, 'timing/model': 1.693086224142462, 'learning_rate/qwen_vl_interface': 1.971857544732672e-06, 'learning_rate/action_model': 1.971857544732672e-05, 'epoch': 1.45}) +[2026-06-23 22:02:24,578][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.20851242542266846, 'train/grad_norm_pre_clip': 90.40367889404297, 'timing/data': 0.0005896079819649458, 'timing/model': 1.6901508669834584, 'learning_rate/qwen_vl_interface': 1.96965430680943e-06, 'learning_rate/action_model': 1.96965430680943e-05, 'epoch': 1.45}) +[2026-06-23 22:02:27,920][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.17769242823123932, 'train/grad_norm_pre_clip': 20.91550636291504, 'timing/data': 0.0005833359900861979, 'timing/model': 1.6968414329458028, 'learning_rate/qwen_vl_interface': 1.967452417312393e-06, 'learning_rate/action_model': 1.9674524173123927e-05, 'epoch': 1.45}) +[2026-06-23 22:02:31,262][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.11962810903787613, 'train/grad_norm_pre_clip': 69.38998413085938, 'timing/data': 0.0008197110146284103, 'timing/model': 1.6988241199869663, 'learning_rate/qwen_vl_interface': 1.9652518771466746e-06, 'learning_rate/action_model': 1.9652518771466747e-05, 'epoch': 1.45}) +[2026-06-23 22:02:34,603][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.11576284468173981, 'train/grad_norm_pre_clip': 16.453554153442383, 'timing/data': 0.0006137529853731394, 'timing/model': 1.6975223149638623, 'learning_rate/qwen_vl_interface': 1.963052687216834e-06, 'learning_rate/action_model': 1.963052687216834e-05, 'epoch': 1.45}) +[2026-06-23 22:02:37,944][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.09562921524047852, 'train/grad_norm_pre_clip': 19.99584197998047, 'timing/data': 0.0007276269607245922, 'timing/model': 1.6984028001315892, 'learning_rate/qwen_vl_interface': 1.960854848426874e-06, 'learning_rate/action_model': 1.960854848426874e-05, 'epoch': 1.45}) +[2026-06-23 22:02:41,286][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.05056391656398773, 'train/grad_norm_pre_clip': 10.220209121704102, 'timing/data': 0.0005963009316474199, 'timing/model': 1.6982328719459474, 'learning_rate/qwen_vl_interface': 1.9586583616802446e-06, 'learning_rate/action_model': 1.9586583616802446e-05, 'epoch': 1.45}) +[2026-06-23 22:02:44,625][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.06439808011054993, 'train/grad_norm_pre_clip': 34.3559684753418, 'timing/data': 0.0006053580436855555, 'timing/model': 1.6984315051231533, 'learning_rate/qwen_vl_interface': 1.956463227879838e-06, 'learning_rate/action_model': 1.9564632278798377e-05, 'epoch': 1.45}) +[2026-06-23 22:02:47,966][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.12318406999111176, 'train/grad_norm_pre_clip': 25.064289093017578, 'timing/data': 0.0006046269554644823, 'timing/model': 1.6970326870214194, 'learning_rate/qwen_vl_interface': 1.95426944792799e-06, 'learning_rate/action_model': 1.95426944792799e-05, 'epoch': 1.45}) +[2026-06-23 22:02:51,311][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.09749563783407211, 'train/grad_norm_pre_clip': 16.63336753845215, 'timing/data': 0.0006004080642014742, 'timing/model': 1.6997689199633896, 'learning_rate/qwen_vl_interface': 1.9520770227264817e-06, 'learning_rate/action_model': 1.9520770227264817e-05, 'epoch': 1.45}) +[2026-06-23 22:02:54,655][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.2303280234336853, 'train/grad_norm_pre_clip': 23.09741973876953, 'timing/data': 0.0008283171337097883, 'timing/model': 1.700779580976814, 'learning_rate/qwen_vl_interface': 1.9498859531765373e-06, 'learning_rate/action_model': 1.9498859531765372e-05, 'epoch': 1.45}) +[2026-06-23 22:02:58,005][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.039799533784389496, 'train/grad_norm_pre_clip': 30.22431755065918, 'timing/data': 0.0008049230091273785, 'timing/model': 1.7022969431709498, 'learning_rate/qwen_vl_interface': 1.9476962401788204e-06, 'learning_rate/action_model': 1.9476962401788204e-05, 'epoch': 1.45}) +[2026-06-23 22:03:32,709][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3750 +[2026-06-23 22:03:32,710][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.14817315340042114, 'train/grad_norm_pre_clip': 42.90458297729492, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 21.90000000000008, 'rl_games_eval/mid_train/mean_length': 194.0, 'rl_games_eval/mid_train/std_reward': 12.267844146385391, 'rl_games_eval/mid_train/std_length': 99.29149006838401, 'rl_games_eval/mid_train/macro_mean_reward': 21.90000000000008, 'rl_games_eval/mid_train/macro_mean_length': 194.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 21.90000000000008, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 194.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 12.267844146385391, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 99.29149006838401, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 21.90000000000008, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006543600466102362, 'timing/model': 1.700827423017472, 'learning_rate/qwen_vl_interface': 1.9455078846334397e-06, 'learning_rate/action_model': 1.9455078846334395e-05, 'epoch': 1.45}) +[2026-06-23 22:03:36,003][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.06574317812919617, 'train/grad_norm_pre_clip': 29.089706420898438, 'timing/data': 0.0006767520681023598, 'timing/model': 1.6725361929275095, 'learning_rate/qwen_vl_interface': 1.9433208874399465e-06, 'learning_rate/action_model': 1.9433208874399464e-05, 'epoch': 1.45}) +[2026-06-23 22:03:39,306][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.06647239625453949, 'train/grad_norm_pre_clip': 36.246944427490234, 'timing/data': 0.0005867830477654934, 'timing/model': 1.676659022923559, 'learning_rate/qwen_vl_interface': 1.9411352494973314e-06, 'learning_rate/action_model': 1.9411352494973315e-05, 'epoch': 1.45}) +[2026-06-23 22:03:42,618][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.2650359272956848, 'train/grad_norm_pre_clip': 77.299072265625, 'timing/data': 0.0007051960565149784, 'timing/model': 1.6811001910828054, 'learning_rate/qwen_vl_interface': 1.938950971704031e-06, 'learning_rate/action_model': 1.9389509717040308e-05, 'epoch': 1.45}) +[2026-06-23 22:03:45,934][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.086510568857193, 'train/grad_norm_pre_clip': 41.65842819213867, 'timing/data': 0.0006029629148542881, 'timing/model': 1.683983552036807, 'learning_rate/qwen_vl_interface': 1.936768054957915e-06, 'learning_rate/action_model': 1.9367680549579147e-05, 'epoch': 1.45}) +[2026-06-23 22:03:49,251][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.24977833032608032, 'train/grad_norm_pre_clip': 44.605918884277344, 'timing/data': 0.0006011801306158304, 'timing/model': 1.6847484700847417, 'learning_rate/qwen_vl_interface': 1.9345865001562994e-06, 'learning_rate/action_model': 1.9345865001562994e-05, 'epoch': 1.45}) +[2026-06-23 22:03:52,577][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.11098986864089966, 'train/grad_norm_pre_clip': 24.835695266723633, 'timing/data': 0.0006111990660429001, 'timing/model': 1.6948525719344616, 'learning_rate/qwen_vl_interface': 1.9324063081959394e-06, 'learning_rate/action_model': 1.9324063081959395e-05, 'epoch': 1.45}) +[2026-06-23 22:03:55,900][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.2277904599905014, 'train/grad_norm_pre_clip': 18.967519760131836, 'timing/data': 0.0006041449960321188, 'timing/model': 1.6862680269405246, 'learning_rate/qwen_vl_interface': 1.9302274799730304e-06, 'learning_rate/action_model': 1.9302274799730306e-05, 'epoch': 1.45}) +[2026-06-23 22:03:59,222][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.1586492359638214, 'train/grad_norm_pre_clip': 40.92327880859375, 'timing/data': 0.0006794368382543325, 'timing/model': 1.687165274983272, 'learning_rate/qwen_vl_interface': 1.9280500163832055e-06, 'learning_rate/action_model': 1.9280500163832055e-05, 'epoch': 1.45}) +[2026-06-23 22:04:02,544][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.11991459131240845, 'train/grad_norm_pre_clip': 27.511417388916016, 'timing/data': 0.0005835171323269606, 'timing/model': 1.6869968990795314, 'learning_rate/qwen_vl_interface': 1.925873918321537e-06, 'learning_rate/action_model': 1.925873918321537e-05, 'epoch': 1.45}) +[2026-06-23 22:04:05,866][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.061726659536361694, 'train/grad_norm_pre_clip': 42.73944091796875, 'timing/data': 0.0008868868462741375, 'timing/model': 1.6864939441438764, 'learning_rate/qwen_vl_interface': 1.92369918668254e-06, 'learning_rate/action_model': 1.92369918668254e-05, 'epoch': 1.46}) +[2026-06-23 22:04:09,181][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.07810639590024948, 'train/grad_norm_pre_clip': 49.71634292602539, 'timing/data': 0.0006627060938626528, 'timing/model': 1.6823724580463022, 'learning_rate/qwen_vl_interface': 1.9215258223601605e-06, 'learning_rate/action_model': 1.9215258223601605e-05, 'epoch': 1.46}) +[2026-06-23 22:04:12,507][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.2176276594400406, 'train/grad_norm_pre_clip': 63.27267837524414, 'timing/data': 0.0006711611058562994, 'timing/model': 1.690111184027046, 'learning_rate/qwen_vl_interface': 1.9193538262477874e-06, 'learning_rate/action_model': 1.9193538262477874e-05, 'epoch': 1.46}) +[2026-06-23 22:04:15,830][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.07223831117153168, 'train/grad_norm_pre_clip': 13.46176815032959, 'timing/data': 0.0005839369259774685, 'timing/model': 1.68581139203161, 'learning_rate/qwen_vl_interface': 1.917183199238247e-06, 'learning_rate/action_model': 1.917183199238247e-05, 'epoch': 1.46}) +[2026-06-23 22:04:19,155][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.05662897229194641, 'train/grad_norm_pre_clip': 24.253183364868164, 'timing/data': 0.0006008089985698462, 'timing/model': 1.6888298569247127, 'learning_rate/qwen_vl_interface': 1.9150139422238016e-06, 'learning_rate/action_model': 1.9150139422238018e-05, 'epoch': 1.46}) +[2026-06-23 22:04:22,700][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.16705100238323212, 'train/grad_norm_pre_clip': 13.962717056274414, 'timing/data': 0.0006428880151361227, 'timing/model': 1.6913858919870108, 'learning_rate/qwen_vl_interface': 1.9128460560961513e-06, 'learning_rate/action_model': 1.9128460560961514e-05, 'epoch': 1.46}) +[2026-06-23 22:04:26,025][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.15271872282028198, 'train/grad_norm_pre_clip': 77.0314712524414, 'timing/data': 0.0006008499767631292, 'timing/model': 1.6895580959971994, 'learning_rate/qwen_vl_interface': 1.910679541746432e-06, 'learning_rate/action_model': 1.9106795417464323e-05, 'epoch': 1.46}) +[2026-06-23 22:04:29,346][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.10710363835096359, 'train/grad_norm_pre_clip': 37.68840408325195, 'timing/data': 0.0006612329743802547, 'timing/model': 1.6857721209526062, 'learning_rate/qwen_vl_interface': 1.9085144000652166e-06, 'learning_rate/action_model': 1.9085144000652166e-05, 'epoch': 1.46}) +[2026-06-23 22:04:32,674][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.12532281875610352, 'train/grad_norm_pre_clip': 23.946176528930664, 'timing/data': 0.0006573449354618788, 'timing/model': 1.6879218721296638, 'learning_rate/qwen_vl_interface': 1.9063506319425116e-06, 'learning_rate/action_model': 1.9063506319425114e-05, 'epoch': 1.46}) +[2026-06-23 22:04:35,997][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.08362987637519836, 'train/grad_norm_pre_clip': 45.63315963745117, 'timing/data': 0.0005933060310781002, 'timing/model': 1.689520528074354, 'learning_rate/qwen_vl_interface': 1.9041882382677613e-06, 'learning_rate/action_model': 1.9041882382677612e-05, 'epoch': 1.46}) +[2026-06-23 22:04:39,327][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.12833577394485474, 'train/grad_norm_pre_clip': 24.521814346313477, 'timing/data': 0.0006142840720713139, 'timing/model': 1.6916966990102082, 'learning_rate/qwen_vl_interface': 1.9020272199298452e-06, 'learning_rate/action_model': 1.902027219929845e-05, 'epoch': 1.46}) +[2026-06-23 22:04:42,654][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.10077927261590958, 'train/grad_norm_pre_clip': 18.23257827758789, 'timing/data': 0.0006492000538855791, 'timing/model': 1.6917298811022192, 'learning_rate/qwen_vl_interface': 1.8998675778170736e-06, 'learning_rate/action_model': 1.8998675778170734e-05, 'epoch': 1.46}) +[2026-06-23 22:04:45,979][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.12566106021404266, 'train/grad_norm_pre_clip': 20.331649780273438, 'timing/data': 0.0008083491120487452, 'timing/model': 1.6920688499230891, 'learning_rate/qwen_vl_interface': 1.8977093128171955e-06, 'learning_rate/action_model': 1.8977093128171954e-05, 'epoch': 1.46}) +[2026-06-23 22:04:49,301][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.10516129434108734, 'train/grad_norm_pre_clip': 15.739238739013672, 'timing/data': 0.0008105330634862185, 'timing/model': 1.6895952001214027, 'learning_rate/qwen_vl_interface': 1.8955524258173919e-06, 'learning_rate/action_model': 1.895552425817392e-05, 'epoch': 1.46}) +[2026-06-23 22:04:52,629][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.13215185701847076, 'train/grad_norm_pre_clip': 30.63399887084961, 'timing/data': 0.0006028730422258377, 'timing/model': 1.6918070579413325, 'learning_rate/qwen_vl_interface': 1.8933969177042769e-06, 'learning_rate/action_model': 1.8933969177042768e-05, 'epoch': 1.46}) +[2026-06-23 22:04:55,967][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.05034147575497627, 'train/grad_norm_pre_clip': 10.056256294250488, 'timing/data': 0.0006716628558933735, 'timing/model': 1.6902502330485731, 'learning_rate/qwen_vl_interface': 1.8912427893638996e-06, 'learning_rate/action_model': 1.8912427893638996e-05, 'epoch': 1.46}) +[2026-06-23 22:04:59,292][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.09162714332342148, 'train/grad_norm_pre_clip': 21.516185760498047, 'timing/data': 0.0005927840247750282, 'timing/model': 1.6891943889204413, 'learning_rate/qwen_vl_interface': 1.8890900416817398e-06, 'learning_rate/action_model': 1.8890900416817397e-05, 'epoch': 1.46}) +[2026-06-23 22:05:02,622][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.187413290143013, 'train/grad_norm_pre_clip': 28.98047637939453, 'timing/data': 0.0005925039295107126, 'timing/model': 1.6905502080917358, 'learning_rate/qwen_vl_interface': 1.8869386755427117e-06, 'learning_rate/action_model': 1.8869386755427114e-05, 'epoch': 1.46}) +[2026-06-23 22:05:05,954][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.10955747216939926, 'train/grad_norm_pre_clip': 31.154998779296875, 'timing/data': 0.0008407300338149071, 'timing/model': 1.6953267061617225, 'learning_rate/qwen_vl_interface': 1.8847886918311597e-06, 'learning_rate/action_model': 1.8847886918311598e-05, 'epoch': 1.46}) +[2026-06-23 22:05:09,491][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.14080040156841278, 'train/grad_norm_pre_clip': 90.25801086425781, 'timing/data': 0.0006562941707670689, 'timing/model': 1.8861167191062123, 'learning_rate/qwen_vl_interface': 1.8826400914308616e-06, 'learning_rate/action_model': 1.8826400914308613e-05, 'epoch': 1.46}) +[2026-06-23 22:05:12,812][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.11278308182954788, 'train/grad_norm_pre_clip': 23.65241813659668, 'timing/data': 0.0006945149507373571, 'timing/model': 1.6862662669736892, 'learning_rate/qwen_vl_interface': 1.880492875225025e-06, 'learning_rate/action_model': 1.8804928752250252e-05, 'epoch': 1.46}) +[2026-06-23 22:05:16,134][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.07448121905326843, 'train/grad_norm_pre_clip': 27.960859298706055, 'timing/data': 0.0007007168605923653, 'timing/model': 1.6881720388773829, 'learning_rate/qwen_vl_interface': 1.8783470440962922e-06, 'learning_rate/action_model': 1.878347044096292e-05, 'epoch': 1.46}) +[2026-06-23 22:05:19,460][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.08073212206363678, 'train/grad_norm_pre_clip': 29.2303466796875, 'timing/data': 0.0006510138045996428, 'timing/model': 1.6881112458650023, 'learning_rate/qwen_vl_interface': 1.8762025989267293e-06, 'learning_rate/action_model': 1.876202598926729e-05, 'epoch': 1.46}) +[2026-06-23 22:05:22,785][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.11952722072601318, 'train/grad_norm_pre_clip': 46.67496871948242, 'timing/data': 0.0005917518865317106, 'timing/model': 1.6872117358725518, 'learning_rate/qwen_vl_interface': 1.874059540597839e-06, 'learning_rate/action_model': 1.874059540597839e-05, 'epoch': 1.46}) +[2026-06-23 22:05:26,111][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.20700660347938538, 'train/grad_norm_pre_clip': 33.280357360839844, 'timing/data': 0.0007258541882038116, 'timing/model': 1.6909143149387091, 'learning_rate/qwen_vl_interface': 1.8719178699905508e-06, 'learning_rate/action_model': 1.8719178699905508e-05, 'epoch': 1.46}) +[2026-06-23 22:05:29,439][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.12362772226333618, 'train/grad_norm_pre_clip': 16.52880859375, 'timing/data': 0.0006620250642299652, 'timing/model': 1.6899901791475713, 'learning_rate/qwen_vl_interface': 1.8697775879852242e-06, 'learning_rate/action_model': 1.8697775879852242e-05, 'epoch': 1.46}) +[2026-06-23 22:05:32,762][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.07859938591718674, 'train/grad_norm_pre_clip': 18.85558319091797, 'timing/data': 0.0005887469742447138, 'timing/model': 1.6876783850602806, 'learning_rate/qwen_vl_interface': 1.86763869546165e-06, 'learning_rate/action_model': 1.86763869546165e-05, 'epoch': 1.47}) +[2026-06-23 22:05:36,099][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.07109688967466354, 'train/grad_norm_pre_clip': 12.38011646270752, 'timing/data': 0.0006037640850991011, 'timing/model': 1.6972358191851526, 'learning_rate/qwen_vl_interface': 1.865501193299045e-06, 'learning_rate/action_model': 1.8655011932990447e-05, 'epoch': 1.47}) +[2026-06-23 22:05:39,427][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.08126906305551529, 'train/grad_norm_pre_clip': 33.56873321533203, 'timing/data': 0.0005818339996039867, 'timing/model': 1.6875624579843134, 'learning_rate/qwen_vl_interface': 1.8633650823760555e-06, 'learning_rate/action_model': 1.8633650823760555e-05, 'epoch': 1.47}) +[2026-06-23 22:05:42,750][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.07277525216341019, 'train/grad_norm_pre_clip': 9.62540340423584, 'timing/data': 0.0006976509466767311, 'timing/model': 1.6888571409508586, 'learning_rate/qwen_vl_interface': 1.8612303635707562e-06, 'learning_rate/action_model': 1.861230363570756e-05, 'epoch': 1.47}) +[2026-06-23 22:05:46,079][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.1492890864610672, 'train/grad_norm_pre_clip': 18.035194396972656, 'timing/data': 0.0005994059611111879, 'timing/model': 1.6921097349841148, 'learning_rate/qwen_vl_interface': 1.8590970377606494e-06, 'learning_rate/action_model': 1.859097037760649e-05, 'epoch': 1.47}) +[2026-06-23 22:05:49,404][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.1374664306640625, 'train/grad_norm_pre_clip': 42.95612716674805, 'timing/data': 0.0006941040046513081, 'timing/model': 1.685134825995192, 'learning_rate/qwen_vl_interface': 1.8569651058226667e-06, 'learning_rate/action_model': 1.8569651058226666e-05, 'epoch': 1.47}) +[2026-06-23 22:05:52,733][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.12452564388513565, 'train/grad_norm_pre_clip': 50.94783401489258, 'timing/data': 0.0006550820544362068, 'timing/model': 1.6854582440573722, 'learning_rate/qwen_vl_interface': 1.854834568633161e-06, 'learning_rate/action_model': 1.854834568633161e-05, 'epoch': 1.47}) +[2026-06-23 22:05:56,067][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.1034107655286789, 'train/grad_norm_pre_clip': 47.65422439575195, 'timing/data': 0.0008791121654212475, 'timing/model': 1.688693004893139, 'learning_rate/qwen_vl_interface': 1.8527054270679174e-06, 'learning_rate/action_model': 1.8527054270679172e-05, 'epoch': 1.47}) +[2026-06-23 22:05:59,575][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.07513070851564407, 'train/grad_norm_pre_clip': 13.290499687194824, 'timing/data': 0.0005907800514250994, 'timing/model': 1.6907993690110743, 'learning_rate/qwen_vl_interface': 1.850577682002146e-06, 'learning_rate/action_model': 1.850577682002146e-05, 'epoch': 1.47}) +[2026-06-23 22:06:02,906][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.11041911691427231, 'train/grad_norm_pre_clip': 17.44248390197754, 'timing/data': 0.0007073190063238144, 'timing/model': 1.6904887331184, 'learning_rate/qwen_vl_interface': 1.848451334310482e-06, 'learning_rate/action_model': 1.848451334310482e-05, 'epoch': 1.47}) +[2026-06-23 22:06:06,239][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.1574169248342514, 'train/grad_norm_pre_clip': 40.96858596801758, 'timing/data': 0.0009979151654988527, 'timing/model': 1.691073050023988, 'learning_rate/qwen_vl_interface': 1.8463263848669865e-06, 'learning_rate/action_model': 1.8463263848669864e-05, 'epoch': 1.47}) +[2026-06-23 22:06:09,565][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.03520837426185608, 'train/grad_norm_pre_clip': 26.869457244873047, 'timing/data': 0.0006989529356360435, 'timing/model': 1.687703215982765, 'learning_rate/qwen_vl_interface': 1.8442028345451468e-06, 'learning_rate/action_model': 1.8442028345451468e-05, 'epoch': 1.47}) +[2026-06-23 22:06:12,892][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.28942397236824036, 'train/grad_norm_pre_clip': 25.4309139251709, 'timing/data': 0.000812286976724863, 'timing/model': 1.6897119728382677, 'learning_rate/qwen_vl_interface': 1.8420806842178746e-06, 'learning_rate/action_model': 1.8420806842178745e-05, 'epoch': 1.47}) +[2026-06-23 22:06:16,228][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.11868105828762054, 'train/grad_norm_pre_clip': 25.443628311157227, 'timing/data': 0.0006794070359319448, 'timing/model': 1.6909013078548014, 'learning_rate/qwen_vl_interface': 1.8399599347575034e-06, 'learning_rate/action_model': 1.8399599347575033e-05, 'epoch': 1.47}) +[2026-06-23 22:06:19,562][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3800 +[2026-06-23 22:07:52,968][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3800 +[2026-06-23 22:07:52,968][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3800 +[2026-06-23 22:09:04,339][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3800 +[2026-06-23 22:09:04,340][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.1341521292924881, 'train/grad_norm_pre_clip': 23.311426162719727, 'eval/action_loss/seconds': 93.40383928501979, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06533362716436386, 'eval/latency_2/loss': 0.06533362716436386, 'eval/flappy/loss': 0.06533362716436386, 'eval/flappy/latency_2/loss': 0.06533362716436386, 'eval/flap_precision': 0.9287598944591029, 'eval/flap_recall': 0.9500674763832658, 'eval/flap_f1': 0.9392928619079386, 'eval/flap_rate_model': 0.21055555555555555, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9392928619079386, 'eval/latency_2/flap_precision': 0.9287598944591029, 'eval/latency_2/flap_recall': 0.9500674763832658, 'eval/latency_2/flap_f1': 0.9392928619079386, 'eval/latency_2/flap_rate_model': 0.21055555555555555, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9392928619079386, 'eval/action_classification/data_wait_seconds': 0.9290518164634705, 'eval/action_classification/predict_seconds': 70.30436706542969, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006281400565057993, 'timing/model': 1.6915471428073943, 'learning_rate/qwen_vl_interface': 1.8378405870357943e-06, 'learning_rate/action_model': 1.837840587035794e-05, 'epoch': 1.47}) +[2026-06-23 22:09:07,646][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.12629097700119019, 'train/grad_norm_pre_clip': 15.015389442443848, 'timing/data': 0.0007022100035101175, 'timing/model': 1.6754373519215733, 'learning_rate/qwen_vl_interface': 1.8357226419239316e-06, 'learning_rate/action_model': 1.8357226419239315e-05, 'epoch': 1.47}) +[2026-06-23 22:09:11,210][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.07871118932962418, 'train/grad_norm_pre_clip': 31.546777725219727, 'timing/data': 0.0006721029058098793, 'timing/model': 1.6816084019374102, 'learning_rate/qwen_vl_interface': 1.8336061002925224e-06, 'learning_rate/action_model': 1.8336061002925224e-05, 'epoch': 1.47}) +[2026-06-23 22:09:14,523][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.059949614107608795, 'train/grad_norm_pre_clip': 26.750333786010742, 'timing/data': 0.00080148596316576, 'timing/model': 1.6829334839712828, 'learning_rate/qwen_vl_interface': 1.8314909630115985e-06, 'learning_rate/action_model': 1.8314909630115983e-05, 'epoch': 1.47}) +[2026-06-23 22:09:17,832][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.042318135499954224, 'train/grad_norm_pre_clip': 13.69953441619873, 'timing/data': 0.0005924340803176165, 'timing/model': 1.6789482720196247, 'learning_rate/qwen_vl_interface': 1.8293772309506092e-06, 'learning_rate/action_model': 1.829377230950609e-05, 'epoch': 1.47}) +[2026-06-23 22:09:21,149][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.22712762653827667, 'train/grad_norm_pre_clip': 22.638538360595703, 'timing/data': 0.0007184199057519436, 'timing/model': 1.6848839388694614, 'learning_rate/qwen_vl_interface': 1.8272649049784326e-06, 'learning_rate/action_model': 1.8272649049784323e-05, 'epoch': 1.47}) +[2026-06-23 22:09:24,459][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.2563231885433197, 'train/grad_norm_pre_clip': 18.730886459350586, 'timing/data': 0.0006369571201503277, 'timing/model': 1.6793134298641235, 'learning_rate/qwen_vl_interface': 1.8251539859633653e-06, 'learning_rate/action_model': 1.825153985963365e-05, 'epoch': 1.47}) +[2026-06-23 22:09:27,770][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.10441049933433533, 'train/grad_norm_pre_clip': 19.064945220947266, 'timing/data': 0.0006842359434813261, 'timing/model': 1.6808160739019513, 'learning_rate/qwen_vl_interface': 1.8230444747731265e-06, 'learning_rate/action_model': 1.8230444747731264e-05, 'epoch': 1.47}) +[2026-06-23 22:09:31,089][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.05262990668416023, 'train/grad_norm_pre_clip': 15.184450149536133, 'timing/data': 0.0006140540353953838, 'timing/model': 1.686721193837002, 'learning_rate/qwen_vl_interface': 1.820936372274857e-06, 'learning_rate/action_model': 1.820936372274857e-05, 'epoch': 1.47}) +[2026-06-23 22:09:34,401][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.19541960954666138, 'train/grad_norm_pre_clip': 25.27779197692871, 'timing/data': 0.0005956401582807302, 'timing/model': 1.6825041479896754, 'learning_rate/qwen_vl_interface': 1.8188296793351178e-06, 'learning_rate/action_model': 1.818829679335118e-05, 'epoch': 1.47}) +[2026-06-23 22:09:37,720][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.027068166062235832, 'train/grad_norm_pre_clip': 27.351816177368164, 'timing/data': 0.0005953090731054544, 'timing/model': 1.6859708949923515, 'learning_rate/qwen_vl_interface': 1.8167243968198905e-06, 'learning_rate/action_model': 1.8167243968198905e-05, 'epoch': 1.47}) +[2026-06-23 22:09:41,033][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.05156451463699341, 'train/grad_norm_pre_clip': 16.461936950683594, 'timing/data': 0.000589307164773345, 'timing/model': 1.6849480820819736, 'learning_rate/qwen_vl_interface': 1.8146205255945775e-06, 'learning_rate/action_model': 1.8146205255945777e-05, 'epoch': 1.47}) +[2026-06-23 22:09:44,358][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.09195850789546967, 'train/grad_norm_pre_clip': 13.113137245178223, 'timing/data': 0.0005924138240516186, 'timing/model': 1.6917049328330904, 'learning_rate/qwen_vl_interface': 1.8125180665240007e-06, 'learning_rate/action_model': 1.8125180665240008e-05, 'epoch': 1.48}) +[2026-06-23 22:09:47,680][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.11856207996606827, 'train/grad_norm_pre_clip': 27.033926010131836, 'timing/data': 0.0008056438528001308, 'timing/model': 1.6900935338344425, 'learning_rate/qwen_vl_interface': 1.8104170204724025e-06, 'learning_rate/action_model': 1.8104170204724024e-05, 'epoch': 1.48}) +[2026-06-23 22:09:51,002][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.109381765127182, 'train/grad_norm_pre_clip': 32.90177917480469, 'timing/data': 0.0005989058408886194, 'timing/model': 1.686288031982258, 'learning_rate/qwen_vl_interface': 1.8083173883034435e-06, 'learning_rate/action_model': 1.8083173883034435e-05, 'epoch': 1.48}) +[2026-06-23 22:09:54,317][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.20116837322711945, 'train/grad_norm_pre_clip': 30.18569564819336, 'timing/data': 0.0006065100897103548, 'timing/model': 1.685328037943691, 'learning_rate/qwen_vl_interface': 1.8062191708802013e-06, 'learning_rate/action_model': 1.8062191708802012e-05, 'epoch': 1.48}) +[2026-06-23 22:09:57,945][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.07989519834518433, 'train/grad_norm_pre_clip': 31.707529067993164, 'timing/data': 0.000681441044434905, 'timing/model': 1.9932783739641309, 'learning_rate/qwen_vl_interface': 1.804122369065175e-06, 'learning_rate/action_model': 1.8041223690651748e-05, 'epoch': 1.48}) +[2026-06-23 22:10:01,266][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.12914355099201202, 'train/grad_norm_pre_clip': 15.068151473999023, 'timing/data': 0.000601690961048007, 'timing/model': 1.685999290086329, 'learning_rate/qwen_vl_interface': 1.8020269837202806e-06, 'learning_rate/action_model': 1.8020269837202804e-05, 'epoch': 1.48}) +[2026-06-23 22:10:04,590][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.06451336294412613, 'train/grad_norm_pre_clip': 19.779495239257812, 'timing/data': 0.0006673950701951981, 'timing/model': 1.6871275410521775, 'learning_rate/qwen_vl_interface': 1.7999330157068517e-06, 'learning_rate/action_model': 1.7999330157068515e-05, 'epoch': 1.48}) +[2026-06-23 22:10:07,911][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.2557249665260315, 'train/grad_norm_pre_clip': 43.98124694824219, 'timing/data': 0.0005885458085685968, 'timing/model': 1.687129745958373, 'learning_rate/qwen_vl_interface': 1.7978404658856393e-06, 'learning_rate/action_model': 1.7978404658856394e-05, 'epoch': 1.48}) +[2026-06-23 22:10:11,238][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.045259781181812286, 'train/grad_norm_pre_clip': 9.465692520141602, 'timing/data': 0.0006481180898845196, 'timing/model': 1.6851346651092172, 'learning_rate/qwen_vl_interface': 1.7957493351168126e-06, 'learning_rate/action_model': 1.7957493351168126e-05, 'epoch': 1.48}) +[2026-06-23 22:10:14,564][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.08440364897251129, 'train/grad_norm_pre_clip': 23.991830825805664, 'timing/data': 0.0006993238348513842, 'timing/model': 1.689430289901793, 'learning_rate/qwen_vl_interface': 1.7936596242599552e-06, 'learning_rate/action_model': 1.7936596242599553e-05, 'epoch': 1.48}) +[2026-06-23 22:10:17,893][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.09085670113563538, 'train/grad_norm_pre_clip': 22.800609588623047, 'timing/data': 0.0006897759158164263, 'timing/model': 1.689474143087864, 'learning_rate/qwen_vl_interface': 1.7915713341740692e-06, 'learning_rate/action_model': 1.791571334174069e-05, 'epoch': 1.48}) +[2026-06-23 22:10:21,216][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.15679959952831268, 'train/grad_norm_pre_clip': 17.801288604736328, 'timing/data': 0.0009216121397912502, 'timing/model': 1.684179601026699, 'learning_rate/qwen_vl_interface': 1.7894844657175716e-06, 'learning_rate/action_model': 1.7894844657175717e-05, 'epoch': 1.48}) +[2026-06-23 22:10:24,541][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.12112044543027878, 'train/grad_norm_pre_clip': 17.774829864501953, 'timing/data': 0.0006852180231362581, 'timing/model': 1.688456960953772, 'learning_rate/qwen_vl_interface': 1.7873990197482962e-06, 'learning_rate/action_model': 1.787399019748296e-05, 'epoch': 1.48}) +[2026-06-23 22:10:27,868][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.06777132302522659, 'train/grad_norm_pre_clip': 37.17208480834961, 'timing/data': 0.000587725080549717, 'timing/model': 1.6880631519015878, 'learning_rate/qwen_vl_interface': 1.7853149971234883e-06, 'learning_rate/action_model': 1.785314997123488e-05, 'epoch': 1.48}) +[2026-06-23 22:10:31,192][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.08193527907133102, 'train/grad_norm_pre_clip': 24.524730682373047, 'timing/data': 0.0010538299102336168, 'timing/model': 1.6895557679235935, 'learning_rate/qwen_vl_interface': 1.7832323986998128e-06, 'learning_rate/action_model': 1.783232398699813e-05, 'epoch': 1.48}) +[2026-06-23 22:10:34,512][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.044561099261045456, 'train/grad_norm_pre_clip': 13.015166282653809, 'timing/data': 0.0006020418368279934, 'timing/model': 1.6837539130356163, 'learning_rate/qwen_vl_interface': 1.7811512253333468e-06, 'learning_rate/action_model': 1.7811512253333467e-05, 'epoch': 1.48}) +[2026-06-23 22:10:37,840][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.1090545505285263, 'train/grad_norm_pre_clip': 16.01885223388672, 'timing/data': 0.0006497118156403303, 'timing/model': 1.689282194012776, 'learning_rate/qwen_vl_interface': 1.7790714778795823e-06, 'learning_rate/action_model': 1.779071477879582e-05, 'epoch': 1.48}) +[2026-06-23 22:10:41,173][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.06127489358186722, 'train/grad_norm_pre_clip': 13.464902877807617, 'timing/data': 0.0006097869481891394, 'timing/model': 1.6939679619390517, 'learning_rate/qwen_vl_interface': 1.7769931571934246e-06, 'learning_rate/action_model': 1.7769931571934246e-05, 'epoch': 1.48}) +[2026-06-23 22:10:44,507][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.12431572377681732, 'train/grad_norm_pre_clip': 49.51933288574219, 'timing/data': 0.000595990102738142, 'timing/model': 1.6970985969528556, 'learning_rate/qwen_vl_interface': 1.7749162641291933e-06, 'learning_rate/action_model': 1.7749162641291933e-05, 'epoch': 1.48}) +[2026-06-23 22:10:48,136][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.09213433414697647, 'train/grad_norm_pre_clip': 26.354312896728516, 'timing/data': 0.0005945079028606415, 'timing/model': 1.9929739159997553, 'learning_rate/qwen_vl_interface': 1.7728407995406206e-06, 'learning_rate/action_model': 1.7728407995406205e-05, 'epoch': 1.48}) +[2026-06-23 22:10:51,465][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.06882202625274658, 'train/grad_norm_pre_clip': 45.37580108642578, 'timing/data': 0.0006768219172954559, 'timing/model': 1.689763039117679, 'learning_rate/qwen_vl_interface': 1.7707667642808513e-06, 'learning_rate/action_model': 1.7707667642808512e-05, 'epoch': 1.48}) +[2026-06-23 22:10:54,793][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.0579979382455349, 'train/grad_norm_pre_clip': 8.744733810424805, 'timing/data': 0.0006672339513897896, 'timing/model': 1.690571599174291, 'learning_rate/qwen_vl_interface': 1.7686941592024423e-06, 'learning_rate/action_model': 1.7686941592024423e-05, 'epoch': 1.48}) +[2026-06-23 22:10:58,124][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.097885861992836, 'train/grad_norm_pre_clip': 15.719590187072754, 'timing/data': 0.0007037429604679346, 'timing/model': 1.691303035011515, 'learning_rate/qwen_vl_interface': 1.7666229851573662e-06, 'learning_rate/action_model': 1.7666229851573664e-05, 'epoch': 1.48}) +[2026-06-23 22:11:01,454][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.029207881540060043, 'train/grad_norm_pre_clip': 30.271142959594727, 'timing/data': 0.0007006768137216568, 'timing/model': 1.6911267130635679, 'learning_rate/qwen_vl_interface': 1.7645532429970006e-06, 'learning_rate/action_model': 1.7645532429970004e-05, 'epoch': 1.48}) +[2026-06-23 22:11:04,787][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.14230556786060333, 'train/grad_norm_pre_clip': 31.488162994384766, 'timing/data': 0.0006945459172129631, 'timing/model': 1.695122845005244, 'learning_rate/qwen_vl_interface': 1.7624849335721402e-06, 'learning_rate/action_model': 1.7624849335721402e-05, 'epoch': 1.48}) +[2026-06-23 22:11:08,122][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.048855993896722794, 'train/grad_norm_pre_clip': 14.33806324005127, 'timing/data': 0.0005987761542201042, 'timing/model': 1.6943926829844713, 'learning_rate/qwen_vl_interface': 1.7604180577329883e-06, 'learning_rate/action_model': 1.7604180577329883e-05, 'epoch': 1.48}) +[2026-06-23 22:11:11,458][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.08813990652561188, 'train/grad_norm_pre_clip': 18.989805221557617, 'timing/data': 0.0006003992166370153, 'timing/model': 1.6928866719827056, 'learning_rate/qwen_vl_interface': 1.7583526163291604e-06, 'learning_rate/action_model': 1.7583526163291604e-05, 'epoch': 1.49}) +[2026-06-23 22:11:14,816][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.15391041338443756, 'train/grad_norm_pre_clip': 17.252927780151367, 'timing/data': 0.0006039650179445744, 'timing/model': 1.6908981439191848, 'learning_rate/qwen_vl_interface': 1.7562886102096807e-06, 'learning_rate/action_model': 1.7562886102096807e-05, 'epoch': 1.49}) +[2026-06-23 22:11:18,157][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.14124281704425812, 'train/grad_norm_pre_clip': 20.89781951904297, 'timing/data': 0.0008261329494416714, 'timing/model': 1.6958348341286182, 'learning_rate/qwen_vl_interface': 1.7542260402229838e-06, 'learning_rate/action_model': 1.7542260402229838e-05, 'epoch': 1.49}) +[2026-06-23 22:11:21,495][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.09272120147943497, 'train/grad_norm_pre_clip': 33.700801849365234, 'timing/data': 0.0006076321005821228, 'timing/model': 1.6929946150630713, 'learning_rate/qwen_vl_interface': 1.7521649072169152e-06, 'learning_rate/action_model': 1.752164907216915e-05, 'epoch': 1.49}) +[2026-06-23 22:11:24,826][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.16240538656711578, 'train/grad_norm_pre_clip': 51.480255126953125, 'timing/data': 0.000595468794927001, 'timing/model': 1.690388707909733, 'learning_rate/qwen_vl_interface': 1.7501052120387285e-06, 'learning_rate/action_model': 1.7501052120387285e-05, 'epoch': 1.49}) +[2026-06-23 22:11:28,160][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.08693267405033112, 'train/grad_norm_pre_clip': 24.17119789123535, 'timing/data': 0.0005997270345687866, 'timing/model': 1.692330809077248, 'learning_rate/qwen_vl_interface': 1.7480469555350866e-06, 'learning_rate/action_model': 1.7480469555350865e-05, 'epoch': 1.49}) +[2026-06-23 22:11:31,495][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.06752920895814896, 'train/grad_norm_pre_clip': 50.4503173828125, 'timing/data': 0.0010538010392338037, 'timing/model': 1.6968381709884852, 'learning_rate/qwen_vl_interface': 1.7459901385520619e-06, 'learning_rate/action_model': 1.745990138552062e-05, 'epoch': 1.49}) +[2026-06-23 22:11:34,830][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.08533893525600433, 'train/grad_norm_pre_clip': 13.036601066589355, 'timing/data': 0.0005818640347570181, 'timing/model': 1.6915240918751806, 'learning_rate/qwen_vl_interface': 1.7439347619351324e-06, 'learning_rate/action_model': 1.7439347619351324e-05, 'epoch': 1.49}) +[2026-06-23 22:11:38,257][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.19770945608615875, 'train/grad_norm_pre_clip': 26.984453201293945, 'timing/data': 0.0005994969978928566, 'timing/model': 1.6961683931294829, 'learning_rate/qwen_vl_interface': 1.741880826529186e-06, 'learning_rate/action_model': 1.741880826529186e-05, 'epoch': 1.49}) +[2026-06-23 22:11:41,594][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.1992359161376953, 'train/grad_norm_pre_clip': 18.96652603149414, 'timing/data': 0.0006481178570538759, 'timing/model': 1.6940510228741914, 'learning_rate/qwen_vl_interface': 1.7398283331785205e-06, 'learning_rate/action_model': 1.7398283331785203e-05, 'epoch': 1.49}) +[2026-06-23 22:11:44,952][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.0884857177734375, 'train/grad_norm_pre_clip': 23.23736572265625, 'timing/data': 0.00069087790325284, 'timing/model': 1.6955843050964177, 'learning_rate/qwen_vl_interface': 1.737777282726835e-06, 'learning_rate/action_model': 1.737777282726835e-05, 'epoch': 1.49}) +[2026-06-23 22:11:48,293][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.08401688188314438, 'train/grad_norm_pre_clip': 17.772417068481445, 'timing/data': 0.0008197308052331209, 'timing/model': 1.6994788961019367, 'learning_rate/qwen_vl_interface': 1.7357276760172403e-06, 'learning_rate/action_model': 1.7357276760172404e-05, 'epoch': 1.49}) +[2026-06-23 22:11:51,619][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.2268088012933731, 'train/grad_norm_pre_clip': 32.86957550048828, 'timing/data': 0.0006011498626321554, 'timing/model': 1.6924264191184193, 'learning_rate/qwen_vl_interface': 1.733679513892252e-06, 'learning_rate/action_model': 1.7336795138922517e-05, 'epoch': 1.49}) +[2026-06-23 22:11:54,959][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.05372984707355499, 'train/grad_norm_pre_clip': 19.730289459228516, 'timing/data': 0.0006876529660075903, 'timing/model': 1.6965104669798166, 'learning_rate/qwen_vl_interface': 1.7316327971937926e-06, 'learning_rate/action_model': 1.7316327971937926e-05, 'epoch': 1.49}) +[2026-06-23 22:11:58,289][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.11141033470630646, 'train/grad_norm_pre_clip': 24.279741287231445, 'timing/data': 0.000950958114117384, 'timing/model': 1.694199152989313, 'learning_rate/qwen_vl_interface': 1.7295875267631904e-06, 'learning_rate/action_model': 1.7295875267631904e-05, 'epoch': 1.49}) +[2026-06-23 22:12:01,621][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.13090218603610992, 'train/grad_norm_pre_clip': 24.00994110107422, 'timing/data': 0.0005997179541736841, 'timing/model': 1.6918136088643223, 'learning_rate/qwen_vl_interface': 1.7275437034411778e-06, 'learning_rate/action_model': 1.7275437034411776e-05, 'epoch': 1.49}) +[2026-06-23 22:12:04,955][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.20845025777816772, 'train/grad_norm_pre_clip': 32.09195327758789, 'timing/data': 0.0006786659359931946, 'timing/model': 1.6955394828692079, 'learning_rate/qwen_vl_interface': 1.7255013280678941e-06, 'learning_rate/action_model': 1.725501328067894e-05, 'epoch': 1.49}) +[2026-06-23 22:12:08,290][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.1813834309577942, 'train/grad_norm_pre_clip': 63.68231201171875, 'timing/data': 0.0006006890907883644, 'timing/model': 1.6939023749437183, 'learning_rate/qwen_vl_interface': 1.7234604014828824e-06, 'learning_rate/action_model': 1.7234604014828823e-05, 'epoch': 1.49}) +[2026-06-23 22:12:11,629][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.12967614829540253, 'train/grad_norm_pre_clip': 33.09688186645508, 'timing/data': 0.0005983349401503801, 'timing/model': 1.696865334874019, 'learning_rate/qwen_vl_interface': 1.7214209245250906e-06, 'learning_rate/action_model': 1.7214209245250905e-05, 'epoch': 1.49}) +[2026-06-23 22:12:14,962][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.06823588907718658, 'train/grad_norm_pre_clip': 29.495859146118164, 'timing/data': 0.0006845570169389248, 'timing/model': 1.6919453169684857, 'learning_rate/qwen_vl_interface': 1.7193828980328723e-06, 'learning_rate/action_model': 1.719382898032872e-05, 'epoch': 1.49}) +[2026-06-23 22:12:18,294][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.08530694991350174, 'train/grad_norm_pre_clip': 18.82370948791504, 'timing/data': 0.0006438200362026691, 'timing/model': 1.6909396070986986, 'learning_rate/qwen_vl_interface': 1.71734632284398e-06, 'learning_rate/action_model': 1.7173463228439798e-05, 'epoch': 1.49}) +[2026-06-23 22:12:21,629][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.15386615693569183, 'train/grad_norm_pre_clip': 32.576271057128906, 'timing/data': 0.0005904200952500105, 'timing/model': 1.696449854876846, 'learning_rate/qwen_vl_interface': 1.7153111997955744e-06, 'learning_rate/action_model': 1.7153111997955743e-05, 'epoch': 1.49}) +[2026-06-23 22:12:25,063][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.04837631806731224, 'train/grad_norm_pre_clip': 34.298927307128906, 'timing/data': 0.0005916119553148746, 'timing/model': 1.7904038350097835, 'learning_rate/qwen_vl_interface': 1.7132775297242175e-06, 'learning_rate/action_model': 1.7132775297242172e-05, 'epoch': 1.49}) +[2026-06-23 22:12:28,396][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.12232619524002075, 'train/grad_norm_pre_clip': 16.403934478759766, 'timing/data': 0.0005985950119793415, 'timing/model': 1.6935845399275422, 'learning_rate/qwen_vl_interface': 1.7112453134658747e-06, 'learning_rate/action_model': 1.7112453134658747e-05, 'epoch': 1.49}) +[2026-06-23 22:12:31,730][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.19471369683742523, 'train/grad_norm_pre_clip': 65.4933853149414, 'timing/data': 0.000602753134444356, 'timing/model': 1.6941173209343106, 'learning_rate/qwen_vl_interface': 1.7092145518559129e-06, 'learning_rate/action_model': 1.709214551855913e-05, 'epoch': 1.49}) +[2026-06-23 22:12:35,061][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.17024581134319305, 'train/grad_norm_pre_clip': 29.9291934967041, 'timing/data': 0.0006074721459299326, 'timing/model': 1.6921652420423925, 'learning_rate/qwen_vl_interface': 1.7071852457291013e-06, 'learning_rate/action_model': 1.707185245729101e-05, 'epoch': 1.5}) +[2026-06-23 22:12:38,384][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.03360066935420036, 'train/grad_norm_pre_clip': 51.45120620727539, 'timing/data': 0.0005899190437048674, 'timing/model': 1.6928234400693327, 'learning_rate/qwen_vl_interface': 1.7051573959196122e-06, 'learning_rate/action_model': 1.7051573959196122e-05, 'epoch': 1.5}) +[2026-06-23 22:12:41,718][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.07772820442914963, 'train/grad_norm_pre_clip': 13.625540733337402, 'timing/data': 0.0006013009697198868, 'timing/model': 1.6910988881718367, 'learning_rate/qwen_vl_interface': 1.7031310032610166e-06, 'learning_rate/action_model': 1.7031310032610163e-05, 'epoch': 1.5}) +[2026-06-23 22:12:45,054][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.11631877720355988, 'train/grad_norm_pre_clip': 35.40624237060547, 'timing/data': 0.0005852100439369678, 'timing/model': 1.6922519251238555, 'learning_rate/qwen_vl_interface': 1.701106068586289e-06, 'learning_rate/action_model': 1.701106068586289e-05, 'epoch': 1.5}) +[2026-06-23 22:12:48,391][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.11717133969068527, 'train/grad_norm_pre_clip': 28.282716751098633, 'timing/data': 0.0007771099917590618, 'timing/model': 1.6952357538975775, 'learning_rate/qwen_vl_interface': 1.6990825927278032e-06, 'learning_rate/action_model': 1.699082592727803e-05, 'epoch': 1.5}) +[2026-06-23 22:12:51,726][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.05924515798687935, 'train/grad_norm_pre_clip': 16.588409423828125, 'timing/data': 0.0006545300129801035, 'timing/model': 1.6964795929379761, 'learning_rate/qwen_vl_interface': 1.6970605765173344e-06, 'learning_rate/action_model': 1.6970605765173346e-05, 'epoch': 1.5}) +[2026-06-23 22:12:55,057][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.17142602801322937, 'train/grad_norm_pre_clip': 55.65690994262695, 'timing/data': 0.0006195639725774527, 'timing/model': 1.6885139101650566, 'learning_rate/qwen_vl_interface': 1.6950400207860557e-06, 'learning_rate/action_model': 1.6950400207860557e-05, 'epoch': 1.5}) +[2026-06-23 22:12:58,388][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.06789469718933105, 'train/grad_norm_pre_clip': 12.484408378601074, 'timing/data': 0.0006786750163882971, 'timing/model': 1.6912229121662676, 'learning_rate/qwen_vl_interface': 1.6930209263645412e-06, 'learning_rate/action_model': 1.6930209263645412e-05, 'epoch': 1.5}) +[2026-06-23 22:13:01,723][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.1581241935491562, 'train/grad_norm_pre_clip': 42.7647590637207, 'timing/data': 0.0006543397903442383, 'timing/model': 1.6929016378708184, 'learning_rate/qwen_vl_interface': 1.6910032940827656e-06, 'learning_rate/action_model': 1.6910032940827657e-05, 'epoch': 1.5}) +[2026-06-23 22:13:05,052][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.061502501368522644, 'train/grad_norm_pre_clip': 20.301511764526367, 'timing/data': 0.0006659610662609339, 'timing/model': 1.6901801719795913, 'learning_rate/qwen_vl_interface': 1.6889871247701003e-06, 'learning_rate/action_model': 1.6889871247701002e-05, 'epoch': 1.5}) +[2026-06-23 22:13:08,384][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.1295352429151535, 'train/grad_norm_pre_clip': 21.749589920043945, 'timing/data': 0.0006716831121593714, 'timing/model': 1.688504823949188, 'learning_rate/qwen_vl_interface': 1.6869724192553168e-06, 'learning_rate/action_model': 1.686972419255317e-05, 'epoch': 1.5}) +[2026-06-23 22:13:11,720][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.09278757125139236, 'train/grad_norm_pre_clip': 21.7260799407959, 'timing/data': 0.0005869031883776188, 'timing/model': 1.694546981016174, 'learning_rate/qwen_vl_interface': 1.6849591783665847e-06, 'learning_rate/action_model': 1.6849591783665847e-05, 'epoch': 1.5}) +[2026-06-23 22:13:15,140][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.11336854100227356, 'train/grad_norm_pre_clip': 25.15010643005371, 'timing/data': 0.0006005589384585619, 'timing/model': 1.6884688271675259, 'learning_rate/qwen_vl_interface': 1.6829474029314713e-06, 'learning_rate/action_model': 1.6829474029314713e-05, 'epoch': 1.5}) +[2026-06-23 22:13:18,474][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.16959039866924286, 'train/grad_norm_pre_clip': 20.477754592895508, 'timing/data': 0.000589537899941206, 'timing/model': 1.6899965689517558, 'learning_rate/qwen_vl_interface': 1.6809370937769408e-06, 'learning_rate/action_model': 1.680937093776941e-05, 'epoch': 1.5}) +[2026-06-23 22:13:21,805][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.16658730804920197, 'train/grad_norm_pre_clip': 27.760854721069336, 'timing/data': 0.000619363971054554, 'timing/model': 1.6900188198778778, 'learning_rate/qwen_vl_interface': 1.678928251729356e-06, 'learning_rate/action_model': 1.6789282517293557e-05, 'epoch': 1.5}) +[2026-06-23 22:13:25,143][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.09871789813041687, 'train/grad_norm_pre_clip': 27.74959945678711, 'timing/data': 0.0007116768974810839, 'timing/model': 1.69279629108496, 'learning_rate/qwen_vl_interface': 1.6769208776144767e-06, 'learning_rate/action_model': 1.6769208776144764e-05, 'epoch': 1.5}) +[2026-06-23 22:13:28,478][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.09635499119758606, 'train/grad_norm_pre_clip': 22.20939826965332, 'timing/data': 0.0006009801290929317, 'timing/model': 1.6927851310465485, 'learning_rate/qwen_vl_interface': 1.6749149722574564e-06, 'learning_rate/action_model': 1.6749149722574564e-05, 'epoch': 1.5}) +[2026-06-23 22:13:31,812][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.08227929472923279, 'train/grad_norm_pre_clip': 14.333551406860352, 'timing/data': 0.0005901290569454432, 'timing/model': 1.6887224749661982, 'learning_rate/qwen_vl_interface': 1.6729105364828486e-06, 'learning_rate/action_model': 1.6729105364828484e-05, 'epoch': 1.5}) +[2026-06-23 22:13:35,138][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.08643593639135361, 'train/grad_norm_pre_clip': 25.109420776367188, 'timing/data': 0.0006982521153986454, 'timing/model': 1.6861790039110929, 'learning_rate/qwen_vl_interface': 1.6709075711146e-06, 'learning_rate/action_model': 1.6709075711146e-05, 'epoch': 1.5}) +[2026-06-23 22:13:38,466][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.07056073099374771, 'train/grad_norm_pre_clip': 33.18528747558594, 'timing/data': 0.0006641980726271868, 'timing/model': 1.6951520210132003, 'learning_rate/qwen_vl_interface': 1.6689060769760556e-06, 'learning_rate/action_model': 1.6689060769760557e-05, 'epoch': 1.5}) +[2026-06-23 22:13:41,800][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.1592756062746048, 'train/grad_norm_pre_clip': 18.526002883911133, 'timing/data': 0.0006513639818876982, 'timing/model': 1.6929040751419961, 'learning_rate/qwen_vl_interface': 1.6669060548899528e-06, 'learning_rate/action_model': 1.6669060548899525e-05, 'epoch': 1.5}) +[2026-06-23 22:13:45,132][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.10020509362220764, 'train/grad_norm_pre_clip': 19.56795883178711, 'timing/data': 0.0006437299307435751, 'timing/model': 1.6964070298708975, 'learning_rate/qwen_vl_interface': 1.6649075056784253e-06, 'learning_rate/action_model': 1.6649075056784253e-05, 'epoch': 1.5}) +[2026-06-23 22:13:48,464][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.13213343918323517, 'train/grad_norm_pre_clip': 15.390621185302734, 'timing/data': 0.0007053359877318144, 'timing/model': 1.6927542029879987, 'learning_rate/qwen_vl_interface': 1.6629104301630019e-06, 'learning_rate/action_model': 1.662910430163002e-05, 'epoch': 1.5}) +[2026-06-23 22:13:51,792][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.05403880402445793, 'train/grad_norm_pre_clip': 22.733293533325195, 'timing/data': 0.0006568341050297022, 'timing/model': 1.6910456831101328, 'learning_rate/qwen_vl_interface': 1.6609148291646043e-06, 'learning_rate/action_model': 1.6609148291646042e-05, 'epoch': 1.5}) +[2026-06-23 22:13:55,127][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.09801309555768967, 'train/grad_norm_pre_clip': 15.904184341430664, 'timing/data': 0.0006923009641468525, 'timing/model': 1.6958113610744476, 'learning_rate/qwen_vl_interface': 1.658920703503549e-06, 'learning_rate/action_model': 1.6589207035035488e-05, 'epoch': 1.5}) +[2026-06-23 22:13:58,460][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.0940466970205307, 'train/grad_norm_pre_clip': 20.469079971313477, 'timing/data': 0.0007548590656369925, 'timing/model': 1.694045932032168, 'learning_rate/qwen_vl_interface': 1.6569280539995467e-06, 'learning_rate/action_model': 1.6569280539995467e-05, 'epoch': 1.5}) +[2026-06-23 22:14:01,793][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.1154722198843956, 'train/grad_norm_pre_clip': 21.127639770507812, 'timing/data': 0.0006920900195837021, 'timing/model': 1.6910736148711294, 'learning_rate/qwen_vl_interface': 1.6549368814716977e-06, 'learning_rate/action_model': 1.6549368814716977e-05, 'epoch': 1.51}) +[2026-06-23 22:14:05,284][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.1399013102054596, 'train/grad_norm_pre_clip': 18.2760066986084, 'timing/data': 0.000916613033041358, 'timing/model': 1.6940933018922806, 'learning_rate/qwen_vl_interface': 1.6529471867385009e-06, 'learning_rate/action_model': 1.6529471867385007e-05, 'epoch': 1.51}) +[2026-06-23 22:14:08,611][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.254459023475647, 'train/grad_norm_pre_clip': 54.883636474609375, 'timing/data': 0.0006857588887214661, 'timing/model': 1.6914807208813727, 'learning_rate/qwen_vl_interface': 1.6509589706178417e-06, 'learning_rate/action_model': 1.6509589706178417e-05, 'epoch': 1.51}) +[2026-06-23 22:14:11,944][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.12476397305727005, 'train/grad_norm_pre_clip': 27.36410903930664, 'timing/data': 0.0006839949637651443, 'timing/model': 1.6926039119716734, 'learning_rate/qwen_vl_interface': 1.648972233927002e-06, 'learning_rate/action_model': 1.648972233927002e-05, 'epoch': 1.51}) +[2026-06-23 22:14:15,273][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.07909917086362839, 'train/grad_norm_pre_clip': 22.522371292114258, 'timing/data': 0.0005896980874240398, 'timing/model': 1.6876296629197896, 'learning_rate/qwen_vl_interface': 1.646986977482654e-06, 'learning_rate/action_model': 1.6469869774826538e-05, 'epoch': 1.51}) +[2026-06-23 22:14:18,600][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.060914214700460434, 'train/grad_norm_pre_clip': 26.984201431274414, 'timing/data': 0.0006025428883731365, 'timing/model': 1.690596739994362, 'learning_rate/qwen_vl_interface': 1.6450032021008618e-06, 'learning_rate/action_model': 1.6450032021008617e-05, 'epoch': 1.51}) +[2026-06-23 22:14:21,924][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.14330081641674042, 'train/grad_norm_pre_clip': 24.75341033935547, 'timing/data': 0.0006355850491672754, 'timing/model': 1.6901566530577838, 'learning_rate/qwen_vl_interface': 1.6430209085970812e-06, 'learning_rate/action_model': 1.643020908597081e-05, 'epoch': 1.51}) +[2026-06-23 22:14:25,254][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.0707181766629219, 'train/grad_norm_pre_clip': 28.83819007873535, 'timing/data': 0.0005935057997703552, 'timing/model': 1.6908350989688188, 'learning_rate/qwen_vl_interface': 1.641040097786157e-06, 'learning_rate/action_model': 1.641040097786157e-05, 'epoch': 1.51}) +[2026-06-23 22:14:28,584][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.10862255841493607, 'train/grad_norm_pre_clip': 33.63606262207031, 'timing/data': 0.000598265090957284, 'timing/model': 1.6925486689433455, 'learning_rate/qwen_vl_interface': 1.6390607704823267e-06, 'learning_rate/action_model': 1.6390607704823265e-05, 'epoch': 1.51}) +[2026-06-23 22:14:31,927][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.15600423514842987, 'train/grad_norm_pre_clip': 23.310237884521484, 'timing/data': 0.0007192008197307587, 'timing/model': 1.689199403859675, 'learning_rate/qwen_vl_interface': 1.6370829274992162e-06, 'learning_rate/action_model': 1.6370829274992162e-05, 'epoch': 1.51}) +[2026-06-23 22:14:35,254][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.21432846784591675, 'train/grad_norm_pre_clip': 28.130821228027344, 'timing/data': 0.0006102870684117079, 'timing/model': 1.6901175898965448, 'learning_rate/qwen_vl_interface': 1.6351065696498422e-06, 'learning_rate/action_model': 1.6351065696498422e-05, 'epoch': 1.51}) +[2026-06-23 22:14:38,582][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3900 +[2026-06-23 22:16:12,568][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3900 +[2026-06-23 22:16:12,569][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 3900 +[2026-06-23 22:17:24,190][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 3900 +[2026-06-23 22:17:24,191][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.09951210767030716, 'train/grad_norm_pre_clip': 20.163230895996094, 'eval/action_loss/seconds': 93.98398566106334, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06592433899641037, 'eval/latency_2/loss': 0.06592433899641037, 'eval/flappy/loss': 0.06592433899641037, 'eval/flappy/latency_2/loss': 0.06592433899641037, 'eval/flap_precision': 0.9312623925974884, 'eval/flap_recall': 0.9507422402159245, 'eval/flap_f1': 0.9409015025041736, 'eval/flap_rate_model': 0.21013888888888888, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9409015025041736, 'eval/latency_2/flap_precision': 0.9312623925974884, 'eval/latency_2/flap_recall': 0.9507422402159245, 'eval/latency_2/flap_f1': 0.9409015025041736, 'eval/latency_2/flap_rate_model': 0.21013888888888888, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9409015025041736, 'eval/action_classification/data_wait_seconds': 0.7933874726295471, 'eval/action_classification/predict_seconds': 70.67594909667969, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0006699890363961458, 'timing/model': 1.6910226908512414, 'learning_rate/qwen_vl_interface': 1.6331316977466117e-06, 'learning_rate/action_model': 1.6331316977466115e-05, 'epoch': 1.51}) +[2026-06-23 22:17:27,507][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.11293748766183853, 'train/grad_norm_pre_clip': 22.74237060546875, 'timing/data': 0.0006101371254771948, 'timing/model': 1.685574047965929, 'learning_rate/qwen_vl_interface': 1.6311583126013198e-06, 'learning_rate/action_model': 1.6311583126013198e-05, 'epoch': 1.51}) +[2026-06-23 22:17:30,822][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.09859378635883331, 'train/grad_norm_pre_clip': 32.388427734375, 'timing/data': 0.0007314039394259453, 'timing/model': 1.6841235209722072, 'learning_rate/qwen_vl_interface': 1.6291864150251488e-06, 'learning_rate/action_model': 1.629186415025149e-05, 'epoch': 1.51}) +[2026-06-23 22:17:34,141][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.15212894976139069, 'train/grad_norm_pre_clip': 26.142353057861328, 'timing/data': 0.0005995370447635651, 'timing/model': 1.6847044839523733, 'learning_rate/qwen_vl_interface': 1.6272160058286727e-06, 'learning_rate/action_model': 1.6272160058286726e-05, 'epoch': 1.51}) +[2026-06-23 22:17:37,460][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.10880114883184433, 'train/grad_norm_pre_clip': 26.080556869506836, 'timing/data': 0.0008365618996322155, 'timing/model': 1.6863116340246052, 'learning_rate/qwen_vl_interface': 1.6252470858218519e-06, 'learning_rate/action_model': 1.6252470858218517e-05, 'epoch': 1.51}) +[2026-06-23 22:17:40,770][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.032641250640153885, 'train/grad_norm_pre_clip': 17.27389144897461, 'timing/data': 0.0005829860456287861, 'timing/model': 1.6824765959754586, 'learning_rate/qwen_vl_interface': 1.6232796558140357e-06, 'learning_rate/action_model': 1.6232796558140357e-05, 'epoch': 1.51}) +[2026-06-23 22:17:44,083][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.05652186647057533, 'train/grad_norm_pre_clip': 16.137592315673828, 'timing/data': 0.0005806419067084789, 'timing/model': 1.6833100030198693, 'learning_rate/qwen_vl_interface': 1.6213137166139592e-06, 'learning_rate/action_model': 1.621313716613959e-05, 'epoch': 1.51}) +[2026-06-23 22:17:47,393][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.15106084942817688, 'train/grad_norm_pre_clip': 23.891298294067383, 'timing/data': 0.0006335708312690258, 'timing/model': 1.6804515400435776, 'learning_rate/qwen_vl_interface': 1.6193492690297463e-06, 'learning_rate/action_model': 1.6193492690297464e-05, 'epoch': 1.51}) +[2026-06-23 22:17:50,710][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.14345665276050568, 'train/grad_norm_pre_clip': 19.478355407714844, 'timing/data': 0.0005956699606031179, 'timing/model': 1.6845827051438391, 'learning_rate/qwen_vl_interface': 1.6173863138689078e-06, 'learning_rate/action_model': 1.617386313868908e-05, 'epoch': 1.51}) +[2026-06-23 22:17:54,025][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.10507730394601822, 'train/grad_norm_pre_clip': 25.763504028320312, 'timing/data': 0.0006001479923725128, 'timing/model': 1.686147697037086, 'learning_rate/qwen_vl_interface': 1.6154248519383396e-06, 'learning_rate/action_model': 1.6154248519383395e-05, 'epoch': 1.51}) +[2026-06-23 22:17:57,341][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.1464988887310028, 'train/grad_norm_pre_clip': 65.94684600830078, 'timing/data': 0.0006047768983989954, 'timing/model': 1.6862302818335593, 'learning_rate/qwen_vl_interface': 1.6134648840443248e-06, 'learning_rate/action_model': 1.613464884044325e-05, 'epoch': 1.51}) +[2026-06-23 22:18:00,658][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.09162544459104538, 'train/grad_norm_pre_clip': 24.21114730834961, 'timing/data': 0.0008445479907095432, 'timing/model': 1.6838816781528294, 'learning_rate/qwen_vl_interface': 1.6115064109925334e-06, 'learning_rate/action_model': 1.6115064109925336e-05, 'epoch': 1.51}) +[2026-06-23 22:18:04,075][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.08723652362823486, 'train/grad_norm_pre_clip': 30.912866592407227, 'timing/data': 0.0007553999312222004, 'timing/model': 1.6907714379485697, 'learning_rate/qwen_vl_interface': 1.6095494335880173e-06, 'learning_rate/action_model': 1.6095494335880174e-05, 'epoch': 1.51}) +[2026-06-23 22:18:07,397][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.0525810681283474, 'train/grad_norm_pre_clip': 15.992510795593262, 'timing/data': 0.0005927740130573511, 'timing/model': 1.688736513024196, 'learning_rate/qwen_vl_interface': 1.6075939526352175e-06, 'learning_rate/action_model': 1.6075939526352175e-05, 'epoch': 1.51}) +[2026-06-23 22:18:10,723][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.07413571327924728, 'train/grad_norm_pre_clip': 41.06515121459961, 'timing/data': 0.0009233248420059681, 'timing/model': 1.6929225311614573, 'learning_rate/qwen_vl_interface': 1.6056399689379576e-06, 'learning_rate/action_model': 1.6056399689379575e-05, 'epoch': 1.51}) +[2026-06-23 22:18:14,041][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.28260907530784607, 'train/grad_norm_pre_clip': 20.639564514160156, 'timing/data': 0.0006450731307268143, 'timing/model': 1.6830452859867364, 'learning_rate/qwen_vl_interface': 1.6036874832994473e-06, 'learning_rate/action_model': 1.603687483299447e-05, 'epoch': 1.52}) +[2026-06-23 22:18:17,359][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.10280860215425491, 'train/grad_norm_pre_clip': 41.196678161621094, 'timing/data': 0.0005998369306325912, 'timing/model': 1.6875065120402724, 'learning_rate/qwen_vl_interface': 1.6017364965222796e-06, 'learning_rate/action_model': 1.6017364965222793e-05, 'epoch': 1.52}) +[2026-06-23 22:18:20,684][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.21585173904895782, 'train/grad_norm_pre_clip': 22.684526443481445, 'timing/data': 0.0005941360723227262, 'timing/model': 1.688273573992774, 'learning_rate/qwen_vl_interface': 1.5997870094084305e-06, 'learning_rate/action_model': 1.5997870094084306e-05, 'epoch': 1.52}) +[2026-06-23 22:18:24,000][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.11023568361997604, 'train/grad_norm_pre_clip': 14.887625694274902, 'timing/data': 0.0005943181458860636, 'timing/model': 1.682515850989148, 'learning_rate/qwen_vl_interface': 1.5978390227592616e-06, 'learning_rate/action_model': 1.5978390227592615e-05, 'epoch': 1.52}) +[2026-06-23 22:18:27,326][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.10277237743139267, 'train/grad_norm_pre_clip': 27.366296768188477, 'timing/data': 0.0006133730057626963, 'timing/model': 1.6879198800306767, 'learning_rate/qwen_vl_interface': 1.5958925373755164e-06, 'learning_rate/action_model': 1.5958925373755164e-05, 'epoch': 1.52}) +[2026-06-23 22:18:30,657][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.06981436163187027, 'train/grad_norm_pre_clip': 41.411251068115234, 'timing/data': 0.0006022718735039234, 'timing/model': 1.692360705928877, 'learning_rate/qwen_vl_interface': 1.5939475540573213e-06, 'learning_rate/action_model': 1.5939475540573212e-05, 'epoch': 1.52}) +[2026-06-23 22:18:33,987][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.15976697206497192, 'train/grad_norm_pre_clip': 55.581748962402344, 'timing/data': 0.0006144139915704727, 'timing/model': 1.6914666860830039, 'learning_rate/qwen_vl_interface': 1.5920040736041867e-06, 'learning_rate/action_model': 1.5920040736041866e-05, 'epoch': 1.52}) +[2026-06-23 22:18:37,313][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.091709204018116, 'train/grad_norm_pre_clip': 23.4747314453125, 'timing/data': 0.0005963919684290886, 'timing/model': 1.6876257760450244, 'learning_rate/qwen_vl_interface': 1.5900620968150015e-06, 'learning_rate/action_model': 1.5900620968150015e-05, 'epoch': 1.52}) +[2026-06-23 22:18:40,646][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.06420446187257767, 'train/grad_norm_pre_clip': 24.77165412902832, 'timing/data': 0.0005842780228704214, 'timing/model': 1.6935277110897005, 'learning_rate/qwen_vl_interface': 1.5881216244880407e-06, 'learning_rate/action_model': 1.5881216244880407e-05, 'epoch': 1.52}) +[2026-06-23 22:18:43,970][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.059846505522727966, 'train/grad_norm_pre_clip': 18.13539695739746, 'timing/data': 0.0006081829778850079, 'timing/model': 1.6888508500996977, 'learning_rate/qwen_vl_interface': 1.5861826574209587e-06, 'learning_rate/action_model': 1.5861826574209587e-05, 'epoch': 1.52}) +[2026-06-23 22:18:47,304][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.19047553837299347, 'train/grad_norm_pre_clip': 17.435205459594727, 'timing/data': 0.0005844889674335718, 'timing/model': 1.6919443639926612, 'learning_rate/qwen_vl_interface': 1.5842451964107916e-06, 'learning_rate/action_model': 1.5842451964107916e-05, 'epoch': 1.52}) +[2026-06-23 22:18:50,632][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.13626866042613983, 'train/grad_norm_pre_clip': 19.71507453918457, 'timing/data': 0.0005875539500266314, 'timing/model': 1.6914741611108184, 'learning_rate/qwen_vl_interface': 1.5823092422539567e-06, 'learning_rate/action_model': 1.582309242253957e-05, 'epoch': 1.52}) +[2026-06-23 22:18:54,105][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.11051614582538605, 'train/grad_norm_pre_clip': 16.226390838623047, 'timing/data': 0.0005937360692769289, 'timing/model': 1.6971130098681897, 'learning_rate/qwen_vl_interface': 1.5803747957462514e-06, 'learning_rate/action_model': 1.5803747957462514e-05, 'epoch': 1.52}) +[2026-06-23 22:18:57,435][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.08076971024274826, 'train/grad_norm_pre_clip': 23.097116470336914, 'timing/data': 0.000596161000430584, 'timing/model': 1.6907764379866421, 'learning_rate/qwen_vl_interface': 1.5784418576828541e-06, 'learning_rate/action_model': 1.5784418576828543e-05, 'epoch': 1.52}) +[2026-06-23 22:19:00,766][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.20735995471477509, 'train/grad_norm_pre_clip': 34.78782653808594, 'timing/data': 0.0005895579233765602, 'timing/model': 1.6932221169117838, 'learning_rate/qwen_vl_interface': 1.576510428858322e-06, 'learning_rate/action_model': 1.5765104288583217e-05, 'epoch': 1.52}) +[2026-06-23 22:19:04,097][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.16774922609329224, 'train/grad_norm_pre_clip': 20.419435501098633, 'timing/data': 0.0005849890876561403, 'timing/model': 1.691323908045888, 'learning_rate/qwen_vl_interface': 1.5745805100665922e-06, 'learning_rate/action_model': 1.5745805100665923e-05, 'epoch': 1.52}) +[2026-06-23 22:19:07,432][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.15736517310142517, 'train/grad_norm_pre_clip': 29.456520080566406, 'timing/data': 0.0005883760750293732, 'timing/model': 1.6972437659278512, 'learning_rate/qwen_vl_interface': 1.572652102100983e-06, 'learning_rate/action_model': 1.572652102100983e-05, 'epoch': 1.52}) +[2026-06-23 22:19:10,767][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.0875743106007576, 'train/grad_norm_pre_clip': 22.437000274658203, 'timing/data': 0.0005853499751538038, 'timing/model': 1.6960529279895127, 'learning_rate/qwen_vl_interface': 1.5707252057541898e-06, 'learning_rate/action_model': 1.5707252057541896e-05, 'epoch': 1.52}) +[2026-06-23 22:19:14,094][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.08680559694766998, 'train/grad_norm_pre_clip': 19.15498924255371, 'timing/data': 0.0005921428091824055, 'timing/model': 1.6896819230169058, 'learning_rate/qwen_vl_interface': 1.5687998218182849e-06, 'learning_rate/action_model': 1.5687998218182847e-05, 'epoch': 1.52}) +[2026-06-23 22:19:17,436][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.04789794236421585, 'train/grad_norm_pre_clip': 14.558812141418457, 'timing/data': 0.0006193439476191998, 'timing/model': 1.6973119648173451, 'learning_rate/qwen_vl_interface': 1.5668759510847235e-06, 'learning_rate/action_model': 1.5668759510847233e-05, 'epoch': 1.52}) +[2026-06-23 22:19:20,772][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.13240298628807068, 'train/grad_norm_pre_clip': 29.08687400817871, 'timing/data': 0.0005945670418441296, 'timing/model': 1.6957397190853953, 'learning_rate/qwen_vl_interface': 1.5649535943443331e-06, 'learning_rate/action_model': 1.564953594344333e-05, 'epoch': 1.52}) +[2026-06-23 22:19:24,100][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.18102625012397766, 'train/grad_norm_pre_clip': 98.02352142333984, 'timing/data': 0.0005965919699519873, 'timing/model': 1.6911873130593449, 'learning_rate/qwen_vl_interface': 1.5630327523873244e-06, 'learning_rate/action_model': 1.563032752387324e-05, 'epoch': 1.52}) +[2026-06-23 22:19:27,439][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.07922676205635071, 'train/grad_norm_pre_clip': 50.36883544921875, 'timing/data': 0.0006589081604033709, 'timing/model': 1.6956021001096815, 'learning_rate/qwen_vl_interface': 1.561113426003282e-06, 'learning_rate/action_model': 1.561113426003282e-05, 'epoch': 1.52}) +[2026-06-23 22:19:30,781][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.09301131218671799, 'train/grad_norm_pre_clip': 24.12983512878418, 'timing/data': 0.0005964010488241911, 'timing/model': 1.698879512026906, 'learning_rate/qwen_vl_interface': 1.5591956159811686e-06, 'learning_rate/action_model': 1.5591956159811687e-05, 'epoch': 1.52}) +[2026-06-23 22:19:34,127][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.08386455476284027, 'train/grad_norm_pre_clip': 24.167091369628906, 'timing/data': 0.000598234822973609, 'timing/model': 1.6979002309963107, 'learning_rate/qwen_vl_interface': 1.5572793231093235e-06, 'learning_rate/action_model': 1.5572793231093236e-05, 'epoch': 1.52}) +[2026-06-23 22:19:37,469][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.19951671361923218, 'train/grad_norm_pre_clip': 42.241607666015625, 'timing/data': 0.0006037750281393528, 'timing/model': 1.6982570320833474, 'learning_rate/qwen_vl_interface': 1.5553645481754626e-06, 'learning_rate/action_model': 1.5553645481754626e-05, 'epoch': 1.52}) +[2026-06-23 22:19:40,945][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.17373022437095642, 'train/grad_norm_pre_clip': 69.6923828125, 'timing/data': 0.0006005577743053436, 'timing/model': 1.828318462939933, 'learning_rate/qwen_vl_interface': 1.5534512919666775e-06, 'learning_rate/action_model': 1.5534512919666776e-05, 'epoch': 1.53}) +[2026-06-23 22:19:44,288][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.08930860459804535, 'train/grad_norm_pre_clip': 23.61189079284668, 'timing/data': 0.0005866528954356909, 'timing/model': 1.6986216271761805, 'learning_rate/qwen_vl_interface': 1.551539555269435e-06, 'learning_rate/action_model': 1.5515395552694348e-05, 'epoch': 1.53}) +[2026-06-23 22:19:47,633][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.09202969074249268, 'train/grad_norm_pre_clip': 37.087974548339844, 'timing/data': 0.0006031841039657593, 'timing/model': 1.6970090460963547, 'learning_rate/qwen_vl_interface': 1.5496293388695787e-06, 'learning_rate/action_model': 1.5496293388695786e-05, 'epoch': 1.53}) +[2026-06-23 22:19:50,970][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.0828181654214859, 'train/grad_norm_pre_clip': 13.603856086730957, 'timing/data': 0.0005890270695090294, 'timing/model': 1.6924684511031955, 'learning_rate/qwen_vl_interface': 1.547720643552327e-06, 'learning_rate/action_model': 1.547720643552327e-05, 'epoch': 1.53}) +[2026-06-23 22:19:54,311][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.2844204604625702, 'train/grad_norm_pre_clip': 42.641109466552734, 'timing/data': 0.0008118860423564911, 'timing/model': 1.6958512600976974, 'learning_rate/qwen_vl_interface': 1.5458134701022706e-06, 'learning_rate/action_model': 1.5458134701022705e-05, 'epoch': 1.53}) +[2026-06-23 22:19:57,658][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.07510711997747421, 'train/grad_norm_pre_clip': 10.552460670471191, 'timing/data': 0.0005943980067968369, 'timing/model': 1.6984245269559324, 'learning_rate/qwen_vl_interface': 1.5439078193033765e-06, 'learning_rate/action_model': 1.5439078193033766e-05, 'epoch': 1.53}) +[2026-06-23 22:20:01,001][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.048784635961055756, 'train/grad_norm_pre_clip': 26.760282516479492, 'timing/data': 0.0005867630243301392, 'timing/model': 1.6973664481192827, 'learning_rate/qwen_vl_interface': 1.5420036919389872e-06, 'learning_rate/action_model': 1.5420036919389873e-05, 'epoch': 1.53}) +[2026-06-23 22:20:04,342][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.08920873701572418, 'train/grad_norm_pre_clip': 14.813095092773438, 'timing/data': 0.0006053680554032326, 'timing/model': 1.700541776837781, 'learning_rate/qwen_vl_interface': 1.540101088791817e-06, 'learning_rate/action_model': 1.5401010887918168e-05, 'epoch': 1.53}) +[2026-06-23 22:20:07,684][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.0864308625459671, 'train/grad_norm_pre_clip': 24.995880126953125, 'timing/data': 0.0006025331094861031, 'timing/model': 1.6974593331106007, 'learning_rate/qwen_vl_interface': 1.5382000106439532e-06, 'learning_rate/action_model': 1.5382000106439532e-05, 'epoch': 1.53}) +[2026-06-23 22:20:11,028][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.132057785987854, 'train/grad_norm_pre_clip': 56.28723907470703, 'timing/data': 0.0006009889766573906, 'timing/model': 1.6975201878231019, 'learning_rate/qwen_vl_interface': 1.5363004582768584e-06, 'learning_rate/action_model': 1.5363004582768584e-05, 'epoch': 1.53}) +[2026-06-23 22:20:14,370][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.05200609192252159, 'train/grad_norm_pre_clip': 20.487512588500977, 'timing/data': 0.0006001279689371586, 'timing/model': 1.699993205955252, 'learning_rate/qwen_vl_interface': 1.534402432471367e-06, 'learning_rate/action_model': 1.534402432471367e-05, 'epoch': 1.53}) +[2026-06-23 22:20:17,722][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.1366381049156189, 'train/grad_norm_pre_clip': 18.343605041503906, 'timing/data': 0.0006170801352709532, 'timing/model': 1.7019497940782458, 'learning_rate/qwen_vl_interface': 1.532505934007685e-06, 'learning_rate/action_model': 1.532505934007685e-05, 'epoch': 1.53}) +[2026-06-23 22:20:21,072][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.13433504104614258, 'train/grad_norm_pre_clip': 19.7127628326416, 'timing/data': 0.0005989749915897846, 'timing/model': 1.6999328231904656, 'learning_rate/qwen_vl_interface': 1.5306109636653915e-06, 'learning_rate/action_model': 1.5306109636653914e-05, 'epoch': 1.53}) +[2026-06-23 22:20:24,428][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.06681859493255615, 'train/grad_norm_pre_clip': 20.35370635986328, 'timing/data': 0.000660681864246726, 'timing/model': 1.7032133089378476, 'learning_rate/qwen_vl_interface': 1.5287175222234383e-06, 'learning_rate/action_model': 1.528717522223438e-05, 'epoch': 1.53}) +[2026-06-23 22:20:27,776][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.15932735800743103, 'train/grad_norm_pre_clip': 15.020113945007324, 'timing/data': 0.0005938969552516937, 'timing/model': 1.6987700571771711, 'learning_rate/qwen_vl_interface': 1.5268256104601457e-06, 'learning_rate/action_model': 1.5268256104601454e-05, 'epoch': 1.53}) +[2026-06-23 22:20:31,211][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.12866614758968353, 'train/grad_norm_pre_clip': 22.65770149230957, 'timing/data': 0.0005908210296183825, 'timing/model': 1.7901188221294433, 'learning_rate/qwen_vl_interface': 1.5249352291532075e-06, 'learning_rate/action_model': 1.5249352291532073e-05, 'epoch': 1.53}) +[2026-06-23 22:20:34,547][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.09786456823348999, 'train/grad_norm_pre_clip': 13.22103500366211, 'timing/data': 0.0005939060356467962, 'timing/model': 1.694430341012776, 'learning_rate/qwen_vl_interface': 1.523046379079689e-06, 'learning_rate/action_model': 1.523046379079689e-05, 'epoch': 1.53}) +[2026-06-23 22:20:37,879][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.0861770510673523, 'train/grad_norm_pre_clip': 26.512495040893555, 'timing/data': 0.0006151169072836637, 'timing/model': 1.6937506331596524, 'learning_rate/qwen_vl_interface': 1.5211590610160244e-06, 'learning_rate/action_model': 1.5211590610160242e-05, 'epoch': 1.53}) +[2026-06-23 22:20:41,230][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.07353582233190536, 'train/grad_norm_pre_clip': 22.376928329467773, 'timing/data': 0.0006123010534793139, 'timing/model': 1.711472102906555, 'learning_rate/qwen_vl_interface': 1.5192732757380193e-06, 'learning_rate/action_model': 1.5192732757380194e-05, 'epoch': 1.53}) +[2026-06-23 22:20:44,571][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.20926949381828308, 'train/grad_norm_pre_clip': 68.29739379882812, 'timing/data': 0.0005897181108593941, 'timing/model': 1.6951504240278155, 'learning_rate/qwen_vl_interface': 1.517389024020849e-06, 'learning_rate/action_model': 1.5173890240208488e-05, 'epoch': 1.53}) +[2026-06-23 22:20:47,914][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.07719492167234421, 'train/grad_norm_pre_clip': 29.26492691040039, 'timing/data': 0.0006371678318828344, 'timing/model': 1.6990625478792936, 'learning_rate/qwen_vl_interface': 1.515506306639057e-06, 'learning_rate/action_model': 1.515506306639057e-05, 'epoch': 1.53}) +[2026-06-23 22:20:51,268][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.07720857858657837, 'train/grad_norm_pre_clip': 59.59312438964844, 'timing/data': 0.0005972930230200291, 'timing/model': 1.7052998528815806, 'learning_rate/qwen_vl_interface': 1.5136251243665589e-06, 'learning_rate/action_model': 1.5136251243665589e-05, 'epoch': 1.53}) +[2026-06-23 22:20:54,610][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.12025745958089828, 'train/grad_norm_pre_clip': 21.027162551879883, 'timing/data': 0.0005758728366345167, 'timing/model': 1.6961794788949192, 'learning_rate/qwen_vl_interface': 1.5117454779766372e-06, 'learning_rate/action_model': 1.5117454779766371e-05, 'epoch': 1.53}) +[2026-06-23 22:20:57,960][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.2702018618583679, 'train/grad_norm_pre_clip': 26.24781036376953, 'timing/data': 0.0005927749443799257, 'timing/model': 1.7005516071803868, 'learning_rate/qwen_vl_interface': 1.5098673682419428e-06, 'learning_rate/action_model': 1.5098673682419429e-05, 'epoch': 1.53}) +[2026-06-23 22:21:01,304][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.13945358991622925, 'train/grad_norm_pre_clip': 27.553077697753906, 'timing/data': 0.0005964718293398619, 'timing/model': 1.7035593499895185, 'learning_rate/qwen_vl_interface': 1.5079907959344977e-06, 'learning_rate/action_model': 1.5079907959344978e-05, 'epoch': 1.53}) +[2026-06-23 22:21:04,653][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.09833820164203644, 'train/grad_norm_pre_clip': 17.88838005065918, 'timing/data': 0.0005838980432599783, 'timing/model': 1.703172354027629, 'learning_rate/qwen_vl_interface': 1.5061157618256873e-06, 'learning_rate/action_model': 1.5061157618256871e-05, 'epoch': 1.53}) +[2026-06-23 22:21:07,997][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.11373315751552582, 'train/grad_norm_pre_clip': 21.742965698242188, 'timing/data': 0.00075916713103652, 'timing/model': 1.6976083349436522, 'learning_rate/qwen_vl_interface': 1.5042422666862675e-06, 'learning_rate/action_model': 1.5042422666862673e-05, 'epoch': 1.54}) +[2026-06-23 22:21:11,344][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.1658564656972885, 'train/grad_norm_pre_clip': 32.542747497558594, 'timing/data': 0.0005997170228511095, 'timing/model': 1.6992766610346735, 'learning_rate/qwen_vl_interface': 1.5023703112863628e-06, 'learning_rate/action_model': 1.5023703112863626e-05, 'epoch': 1.54}) +[2026-06-23 22:21:14,693][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.04758825525641441, 'train/grad_norm_pre_clip': 36.559112548828125, 'timing/data': 0.0005962408613413572, 'timing/model': 1.7019234949257225, 'learning_rate/qwen_vl_interface': 1.5004998963954629e-06, 'learning_rate/action_model': 1.5004998963954628e-05, 'epoch': 1.54}) +[2026-06-23 22:21:18,037][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.12089046090841293, 'train/grad_norm_pre_clip': 62.225032806396484, 'timing/data': 0.0006065899506211281, 'timing/model': 1.7003645261283964, 'learning_rate/qwen_vl_interface': 1.498631022782424e-06, 'learning_rate/action_model': 1.4986310227824241e-05, 'epoch': 1.54}) +[2026-06-23 22:21:21,460][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.109676294028759, 'train/grad_norm_pre_clip': 18.456872940063477, 'timing/data': 0.0005988560151308775, 'timing/model': 1.695929649984464, 'learning_rate/qwen_vl_interface': 1.4967636912154707e-06, 'learning_rate/action_model': 1.4967636912154706e-05, 'epoch': 1.54}) +[2026-06-23 22:21:24,806][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.12637877464294434, 'train/grad_norm_pre_clip': 48.86743927001953, 'timing/data': 0.0006434200331568718, 'timing/model': 1.7005998778622597, 'learning_rate/qwen_vl_interface': 1.4948979024621915e-06, 'learning_rate/action_model': 1.4948979024621914e-05, 'epoch': 1.54}) +[2026-06-23 22:21:28,154][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.1626296490430832, 'train/grad_norm_pre_clip': 54.423667907714844, 'timing/data': 0.0005950790364295244, 'timing/model': 1.7026394619606435, 'learning_rate/qwen_vl_interface': 1.493033657289542e-06, 'learning_rate/action_model': 1.493033657289542e-05, 'epoch': 1.54}) +[2026-06-23 22:21:31,495][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.15740156173706055, 'train/grad_norm_pre_clip': 48.16073226928711, 'timing/data': 0.0005953300278633833, 'timing/model': 1.6949004051275551, 'learning_rate/qwen_vl_interface': 1.4911709564638428e-06, 'learning_rate/action_model': 1.4911709564638427e-05, 'epoch': 1.54}) +[2026-06-23 22:21:34,842][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.08434836566448212, 'train/grad_norm_pre_clip': 44.70667266845703, 'timing/data': 0.0005951989442110062, 'timing/model': 1.7002623539883643, 'learning_rate/qwen_vl_interface': 1.4893098007507817e-06, 'learning_rate/action_model': 1.4893098007507816e-05, 'epoch': 1.54}) +[2026-06-23 22:21:38,191][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.07204341888427734, 'train/grad_norm_pre_clip': 12.744667053222656, 'timing/data': 0.000660671154037118, 'timing/model': 1.7019201300572604, 'learning_rate/qwen_vl_interface': 1.487450190915406e-06, 'learning_rate/action_model': 1.487450190915406e-05, 'epoch': 1.54}) +[2026-06-23 22:21:41,531][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.15560781955718994, 'train/grad_norm_pre_clip': 16.315080642700195, 'timing/data': 0.0006491399835795164, 'timing/model': 1.69877248397097, 'learning_rate/qwen_vl_interface': 1.4855921277221336e-06, 'learning_rate/action_model': 1.4855921277221336e-05, 'epoch': 1.54}) +[2026-06-23 22:21:44,874][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.12376497685909271, 'train/grad_norm_pre_clip': 39.8350830078125, 'timing/data': 0.0005879860837012529, 'timing/model': 1.6975118739064783, 'learning_rate/qwen_vl_interface': 1.4837356119347423e-06, 'learning_rate/action_model': 1.4837356119347423e-05, 'epoch': 1.54}) +[2026-06-23 22:21:48,212][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.07576359808444977, 'train/grad_norm_pre_clip': 14.771843910217285, 'timing/data': 0.0005919928662478924, 'timing/model': 1.696530639892444, 'learning_rate/qwen_vl_interface': 1.4818806443163774e-06, 'learning_rate/action_model': 1.4818806443163775e-05, 'epoch': 1.54}) +[2026-06-23 22:21:51,551][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.15340439975261688, 'train/grad_norm_pre_clip': 28.144004821777344, 'timing/data': 0.0005958098918199539, 'timing/model': 1.696365109179169, 'learning_rate/qwen_vl_interface': 1.480027225629547e-06, 'learning_rate/action_model': 1.480027225629547e-05, 'epoch': 1.54}) +[2026-06-23 22:21:54,892][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.2714868187904358, 'train/grad_norm_pre_clip': 28.347152709960938, 'timing/data': 0.0005745701491832733, 'timing/model': 1.6956729488447309, 'learning_rate/qwen_vl_interface': 1.4781753566361182e-06, 'learning_rate/action_model': 1.4781753566361182e-05, 'epoch': 1.54}) +[2026-06-23 22:21:58,250][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.11206431686878204, 'train/grad_norm_pre_clip': 22.6072940826416, 'timing/data': 0.000709964893758297, 'timing/model': 1.7122798331547529, 'learning_rate/qwen_vl_interface': 1.4763250380973255e-06, 'learning_rate/action_model': 1.4763250380973254e-05, 'epoch': 1.54}) +[2026-06-23 22:22:01,588][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.15234322845935822, 'train/grad_norm_pre_clip': 59.23408508300781, 'timing/data': 0.0006462649907916784, 'timing/model': 1.693060838850215, 'learning_rate/qwen_vl_interface': 1.4744762707737662e-06, 'learning_rate/action_model': 1.4744762707737662e-05, 'epoch': 1.54}) +[2026-06-23 22:22:04,931][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.09704699367284775, 'train/grad_norm_pre_clip': 20.517541885375977, 'timing/data': 0.0006023519672453403, 'timing/model': 1.695107646053657, 'learning_rate/qwen_vl_interface': 1.4726290554253977e-06, 'learning_rate/action_model': 1.4726290554253977e-05, 'epoch': 1.54}) +[2026-06-23 22:22:08,271][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.05434219166636467, 'train/grad_norm_pre_clip': 11.104354858398438, 'timing/data': 0.0006436200346797705, 'timing/model': 1.697151516797021, 'learning_rate/qwen_vl_interface': 1.4707833928115406e-06, 'learning_rate/action_model': 1.4707833928115405e-05, 'epoch': 1.54}) +[2026-06-23 22:22:11,714][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.06860534101724625, 'train/grad_norm_pre_clip': 23.34326934814453, 'timing/data': 0.0005921428091824055, 'timing/model': 1.7009602659381926, 'learning_rate/qwen_vl_interface': 1.4689392836908775e-06, 'learning_rate/action_model': 1.4689392836908774e-05, 'epoch': 1.54}) +[2026-06-23 22:22:15,058][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.16867119073867798, 'train/grad_norm_pre_clip': 26.977306365966797, 'timing/data': 0.0008210830856114626, 'timing/model': 1.697470507118851, 'learning_rate/qwen_vl_interface': 1.4670967288214514e-06, 'learning_rate/action_model': 1.4670967288214513e-05, 'epoch': 1.54}) +[2026-06-23 22:22:18,400][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.08318386971950531, 'train/grad_norm_pre_clip': 11.828008651733398, 'timing/data': 0.000732676824554801, 'timing/model': 1.6973313558846712, 'learning_rate/qwen_vl_interface': 1.4652557289606687e-06, 'learning_rate/action_model': 1.4652557289606685e-05, 'epoch': 1.54}) +[2026-06-23 22:22:21,747][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.2157866656780243, 'train/grad_norm_pre_clip': 68.03337097167969, 'timing/data': 0.0006058290600776672, 'timing/model': 1.7016054899431765, 'learning_rate/qwen_vl_interface': 1.4634162848652918e-06, 'learning_rate/action_model': 1.4634162848652919e-05, 'epoch': 1.54}) +[2026-06-23 22:22:25,093][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.04673028364777565, 'train/grad_norm_pre_clip': 7.788409233093262, 'timing/data': 0.0006007992196828127, 'timing/model': 1.6994556391146034, 'learning_rate/qwen_vl_interface': 1.4615783972914494e-06, 'learning_rate/action_model': 1.4615783972914493e-05, 'epoch': 1.54}) +[2026-06-23 22:22:28,434][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.256973534822464, 'train/grad_norm_pre_clip': 59.70066452026367, 'timing/data': 0.0005852899048477411, 'timing/model': 1.6980635221116245, 'learning_rate/qwen_vl_interface': 1.459742066994626e-06, 'learning_rate/action_model': 1.4597420669946261e-05, 'epoch': 1.54}) +[2026-06-23 22:22:31,783][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.10559607297182083, 'train/grad_norm_pre_clip': 16.721132278442383, 'timing/data': 0.0006513341795653105, 'timing/model': 1.6974532059393823, 'learning_rate/qwen_vl_interface': 1.4579072947296687e-06, 'learning_rate/action_model': 1.4579072947296687e-05, 'epoch': 1.54}) +[2026-06-23 22:22:35,132][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.10113639384508133, 'train/grad_norm_pre_clip': 73.40007781982422, 'timing/data': 0.0006614141166210175, 'timing/model': 1.7001332729123533, 'learning_rate/qwen_vl_interface': 1.4560740812507823e-06, 'learning_rate/action_model': 1.4560740812507822e-05, 'epoch': 1.55}) +[2026-06-23 22:22:38,481][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.0932639092206955, 'train/grad_norm_pre_clip': 13.442140579223633, 'timing/data': 0.0005878249648958445, 'timing/model': 1.7010946909431368, 'learning_rate/qwen_vl_interface': 1.4542424273115322e-06, 'learning_rate/action_model': 1.454242427311532e-05, 'epoch': 1.55}) +[2026-06-23 22:22:41,833][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.10713113844394684, 'train/grad_norm_pre_clip': 15.225471496582031, 'timing/data': 0.00061985501088202, 'timing/model': 1.7036083040293306, 'learning_rate/qwen_vl_interface': 1.4524123336648419e-06, 'learning_rate/action_model': 1.4524123336648418e-05, 'epoch': 1.55}) +[2026-06-23 22:22:45,174][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.1960286796092987, 'train/grad_norm_pre_clip': 20.468116760253906, 'timing/data': 0.0006225411780178547, 'timing/model': 1.6944137520622462, 'learning_rate/qwen_vl_interface': 1.4505838010629942e-06, 'learning_rate/action_model': 1.4505838010629941e-05, 'epoch': 1.55}) +[2026-06-23 22:22:48,521][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.09252142161130905, 'train/grad_norm_pre_clip': 39.75093460083008, 'timing/data': 0.0005990760400891304, 'timing/model': 1.6998935830779374, 'learning_rate/qwen_vl_interface': 1.4487568302576284e-06, 'learning_rate/action_model': 1.4487568302576282e-05, 'epoch': 1.55}) +[2026-06-23 22:22:51,869][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.06630603969097137, 'train/grad_norm_pre_clip': 13.582204818725586, 'timing/data': 0.0005719349719583988, 'timing/model': 1.7011468580458313, 'learning_rate/qwen_vl_interface': 1.446931421999746e-06, 'learning_rate/action_model': 1.4469314219997459e-05, 'epoch': 1.55}) +[2026-06-23 22:22:55,213][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.08814119547605515, 'train/grad_norm_pre_clip': 44.87563705444336, 'timing/data': 0.0006055978592485189, 'timing/model': 1.6984614210668951, 'learning_rate/qwen_vl_interface': 1.4451075770397004e-06, 'learning_rate/action_model': 1.4451075770397004e-05, 'epoch': 1.55}) +[2026-06-23 22:22:58,640][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4000 +[2026-06-23 22:24:32,056][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4000 +[2026-06-23 22:24:32,057][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4000 +[2026-06-23 22:25:43,131][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4000 +[2026-06-23 22:26:33,240][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000 +[2026-06-23 22:26:33,241][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.1167680099606514, 'train/grad_norm_pre_clip': 30.318777084350586, 'eval/action_loss/seconds': 93.41386396600865, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06504729390144348, 'eval/latency_2/loss': 0.06504729390144348, 'eval/flappy/loss': 0.06504729390144348, 'eval/flappy/latency_2/loss': 0.06504729390144348, 'eval/flap_precision': 0.9448979591836735, 'eval/flap_recall': 0.937246963562753, 'eval/flap_f1': 0.9410569105691056, 'eval/flap_rate_model': 0.20416666666666666, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9410569105691057, 'eval/latency_2/flap_precision': 0.9448979591836735, 'eval/latency_2/flap_recall': 0.937246963562753, 'eval/latency_2/flap_f1': 0.9410569105691056, 'eval/latency_2/flap_rate_model': 0.20416666666666666, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9410569105691057, 'eval/action_classification/data_wait_seconds': 0.7998072504997253, 'eval/action_classification/predict_seconds': 70.14533233642578, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 35.979999999999954, 'rl_games_eval/mid_train/mean_length': 307.8, 'rl_games_eval/mid_train/std_reward': 30.149056369975977, 'rl_games_eval/mid_train/std_length': 244.03884936624334, 'rl_games_eval/mid_train/macro_mean_reward': 35.979999999999954, 'rl_games_eval/mid_train/macro_mean_length': 307.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 35.979999999999954, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 307.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 30.149056369975977, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 244.03884936624334, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 35.979999999999954, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005847390275448561, 'timing/model': 1.7866843901574612, 'learning_rate/qwen_vl_interface': 1.4432852961272068e-06, 'learning_rate/action_model': 1.4432852961272066e-05, 'epoch': 1.55}) +[2026-06-23 22:26:33,241][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_4000_state +[2026-06-23 22:26:33,245][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 22:26:35,853][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_4000_state/model.safetensors +[2026-06-23 22:26:36,997][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_4000_state/optimizer.bin +[2026-06-23 22:26:36,997][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_4000_state/sampler.bin +[2026-06-23 22:26:36,997][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_4000_state/sampler_1.bin +[2026-06-23 22:26:36,998][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_4000_state/random_states_0.pkl +[2026-06-23 22:26:37,021][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000 +[2026-06-23 22:26:37,022][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 22:26:37,032][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 22:26:40,301][starVLA.training.train_starvla][INFO] - Step 4001, Loss: {'train/loss': 0.22463996708393097, 'train/grad_norm_pre_clip': 43.93478012084961, 'timing/data': 0.0005867439322173595, 'timing/model': 1.6610451759770513, 'learning_rate/qwen_vl_interface': 1.4414645800113353e-06, 'learning_rate/action_model': 1.4414645800113352e-05, 'epoch': 1.55}) +[2026-06-23 22:26:43,588][starVLA.training.train_starvla][INFO] - Step 4002, Loss: {'train/loss': 0.06791745871305466, 'train/grad_norm_pre_clip': 21.218647003173828, 'timing/data': 0.0006850589998066425, 'timing/model': 1.680068165063858, 'learning_rate/qwen_vl_interface': 1.4396454294405138e-06, 'learning_rate/action_model': 1.4396454294405136e-05, 'epoch': 1.55}) +[2026-06-23 22:26:46,868][starVLA.training.train_starvla][INFO] - Step 4003, Loss: {'train/loss': 0.07033572345972061, 'train/grad_norm_pre_clip': 26.267932891845703, 'timing/data': 0.0005953710060566664, 'timing/model': 1.6660962828900665, 'learning_rate/qwen_vl_interface': 1.4378278451625271e-06, 'learning_rate/action_model': 1.4378278451625271e-05, 'epoch': 1.55}) +[2026-06-23 22:26:50,149][starVLA.training.train_starvla][INFO] - Step 4004, Loss: {'train/loss': 0.10213690996170044, 'train/grad_norm_pre_clip': 17.773488998413086, 'timing/data': 0.0006758710369467735, 'timing/model': 1.668293367838487, 'learning_rate/qwen_vl_interface': 1.4360118279245143e-06, 'learning_rate/action_model': 1.4360118279245144e-05, 'epoch': 1.55}) +[2026-06-23 22:26:53,435][starVLA.training.train_starvla][INFO] - Step 4005, Loss: {'train/loss': 0.06047414243221283, 'train/grad_norm_pre_clip': 24.874380111694336, 'timing/data': 0.0006561849731951952, 'timing/model': 1.6683945588301867, 'learning_rate/qwen_vl_interface': 1.434197378472973e-06, 'learning_rate/action_model': 1.4341973784729729e-05, 'epoch': 1.55}) +[2026-06-23 22:26:56,833][starVLA.training.train_starvla][INFO] - Step 4006, Loss: {'train/loss': 0.16413843631744385, 'train/grad_norm_pre_clip': 29.198150634765625, 'timing/data': 0.0006337719969451427, 'timing/model': 1.6708473251201212, 'learning_rate/qwen_vl_interface': 1.4323844975537535e-06, 'learning_rate/action_model': 1.4323844975537534e-05, 'epoch': 1.55}) +[2026-06-23 22:27:00,128][starVLA.training.train_starvla][INFO] - Step 4007, Loss: {'train/loss': 0.10661014914512634, 'train/grad_norm_pre_clip': 26.201807022094727, 'timing/data': 0.0006071219686418772, 'timing/model': 1.6727672908455133, 'learning_rate/qwen_vl_interface': 1.4305731859120628e-06, 'learning_rate/action_model': 1.4305731859120628e-05, 'epoch': 1.55}) +[2026-06-23 22:27:03,429][starVLA.training.train_starvla][INFO] - Step 4008, Loss: {'train/loss': 0.14424794912338257, 'train/grad_norm_pre_clip': 46.474937438964844, 'timing/data': 0.0006622560322284698, 'timing/model': 1.6767096491530538, 'learning_rate/qwen_vl_interface': 1.4287634442924647e-06, 'learning_rate/action_model': 1.4287634442924646e-05, 'epoch': 1.55}) +[2026-06-23 22:27:06,742][starVLA.training.train_starvla][INFO] - Step 4009, Loss: {'train/loss': 0.09806375205516815, 'train/grad_norm_pre_clip': 33.381107330322266, 'timing/data': 0.0006407559849321842, 'timing/model': 1.6824360359460115, 'learning_rate/qwen_vl_interface': 1.4269552734388725e-06, 'learning_rate/action_model': 1.4269552734388724e-05, 'epoch': 1.55}) +[2026-06-23 22:27:10,050][starVLA.training.train_starvla][INFO] - Step 4010, Loss: {'train/loss': 0.09414849430322647, 'train/grad_norm_pre_clip': 22.585655212402344, 'timing/data': 0.0005994681268930435, 'timing/model': 1.6807140370365232, 'learning_rate/qwen_vl_interface': 1.4251486740945584e-06, 'learning_rate/action_model': 1.4251486740945583e-05, 'epoch': 1.55}) +[2026-06-23 22:27:13,363][starVLA.training.train_starvla][INFO] - Step 4011, Loss: {'train/loss': 0.14050689339637756, 'train/grad_norm_pre_clip': 20.43529510498047, 'timing/data': 0.0006556839216500521, 'timing/model': 1.6853987050708383, 'learning_rate/qwen_vl_interface': 1.423343647002147e-06, 'learning_rate/action_model': 1.423343647002147e-05, 'epoch': 1.55}) +[2026-06-23 22:27:16,678][starVLA.training.train_starvla][INFO] - Step 4012, Loss: {'train/loss': 0.13860006630420685, 'train/grad_norm_pre_clip': 21.420076370239258, 'timing/data': 0.0006044779438525438, 'timing/model': 1.6842638100497425, 'learning_rate/qwen_vl_interface': 1.421540192903617e-06, 'learning_rate/action_model': 1.421540192903617e-05, 'epoch': 1.55}) +[2026-06-23 22:27:20,002][starVLA.training.train_starvla][INFO] - Step 4013, Loss: {'train/loss': 0.06984617561101913, 'train/grad_norm_pre_clip': 13.447129249572754, 'timing/data': 0.000591654097661376, 'timing/model': 1.6896970961242914, 'learning_rate/qwen_vl_interface': 1.4197383125402997e-06, 'learning_rate/action_model': 1.4197383125402997e-05, 'epoch': 1.55}) +[2026-06-23 22:27:23,314][starVLA.training.train_starvla][INFO] - Step 4014, Loss: {'train/loss': 0.11701108515262604, 'train/grad_norm_pre_clip': 30.52741241455078, 'timing/data': 0.0005830370355397463, 'timing/model': 1.6848019161261618, 'learning_rate/qwen_vl_interface': 1.4179380066528807e-06, 'learning_rate/action_model': 1.4179380066528807e-05, 'epoch': 1.55}) +[2026-06-23 22:27:26,630][starVLA.training.train_starvla][INFO] - Step 4015, Loss: {'train/loss': 0.0829526036977768, 'train/grad_norm_pre_clip': 17.961597442626953, 'timing/data': 0.0006458351854234934, 'timing/model': 1.6820179289206862, 'learning_rate/qwen_vl_interface': 1.4161392759813984e-06, 'learning_rate/action_model': 1.4161392759813983e-05, 'epoch': 1.55}) +[2026-06-23 22:27:29,941][starVLA.training.train_starvla][INFO] - Step 4016, Loss: {'train/loss': 0.09502842277288437, 'train/grad_norm_pre_clip': 13.52375316619873, 'timing/data': 0.0006017119158059359, 'timing/model': 1.6835334419738501, 'learning_rate/qwen_vl_interface': 1.4143421212652411e-06, 'learning_rate/action_model': 1.414342121265241e-05, 'epoch': 1.55}) +[2026-06-23 22:27:33,262][starVLA.training.train_starvla][INFO] - Step 4017, Loss: {'train/loss': 0.07061291486024857, 'train/grad_norm_pre_clip': 32.70703887939453, 'timing/data': 0.0006385908927768469, 'timing/model': 1.6871697881724685, 'learning_rate/qwen_vl_interface': 1.4125465432431529e-06, 'learning_rate/action_model': 1.4125465432431529e-05, 'epoch': 1.55}) +[2026-06-23 22:27:36,577][starVLA.training.train_starvla][INFO] - Step 4018, Loss: {'train/loss': 0.12711073458194733, 'train/grad_norm_pre_clip': 16.971115112304688, 'timing/data': 0.0005998690612614155, 'timing/model': 1.6852121958509088, 'learning_rate/qwen_vl_interface': 1.4107525426532273e-06, 'learning_rate/action_model': 1.4107525426532272e-05, 'epoch': 1.56}) +[2026-06-23 22:27:39,941][starVLA.training.train_starvla][INFO] - Step 4019, Loss: {'train/loss': 0.05422128364443779, 'train/grad_norm_pre_clip': 11.627652168273926, 'timing/data': 0.0006708120927214622, 'timing/model': 1.6999192219227552, 'learning_rate/qwen_vl_interface': 1.4089601202329116e-06, 'learning_rate/action_model': 1.4089601202329114e-05, 'epoch': 1.56}) +[2026-06-23 22:27:43,357][starVLA.training.train_starvla][INFO] - Step 4020, Loss: {'train/loss': 0.17143170535564423, 'train/grad_norm_pre_clip': 18.944244384765625, 'timing/data': 0.0006115108262747526, 'timing/model': 1.779520944925025, 'learning_rate/qwen_vl_interface': 1.407169276719e-06, 'learning_rate/action_model': 1.4071692767190001e-05, 'epoch': 1.56}) +[2026-06-23 22:27:46,682][starVLA.training.train_starvla][INFO] - Step 4021, Loss: {'train/loss': 0.189162015914917, 'train/grad_norm_pre_clip': 44.65397644042969, 'timing/data': 0.0008704380597919226, 'timing/model': 1.6895865630358458, 'learning_rate/qwen_vl_interface': 1.4053800128476426e-06, 'learning_rate/action_model': 1.4053800128476427e-05, 'epoch': 1.56}) +[2026-06-23 22:27:50,004][starVLA.training.train_starvla][INFO] - Step 4022, Loss: {'train/loss': 0.20110327005386353, 'train/grad_norm_pre_clip': 21.376916885375977, 'timing/data': 0.0006093960255384445, 'timing/model': 1.6881215490866452, 'learning_rate/qwen_vl_interface': 1.403592329354338e-06, 'learning_rate/action_model': 1.403592329354338e-05, 'epoch': 1.56}) +[2026-06-23 22:27:53,326][starVLA.training.train_starvla][INFO] - Step 4023, Loss: {'train/loss': 0.07598920166492462, 'train/grad_norm_pre_clip': 12.308414459228516, 'timing/data': 0.0006510750390589237, 'timing/model': 1.6875175349414349, 'learning_rate/qwen_vl_interface': 1.4018062269739354e-06, 'learning_rate/action_model': 1.4018062269739355e-05, 'epoch': 1.56}) +[2026-06-23 22:27:56,649][starVLA.training.train_starvla][INFO] - Step 4024, Loss: {'train/loss': 0.13551375269889832, 'train/grad_norm_pre_clip': 40.27723693847656, 'timing/data': 0.0006887160707265139, 'timing/model': 1.685522312996909, 'learning_rate/qwen_vl_interface': 1.4000217064406339e-06, 'learning_rate/action_model': 1.4000217064406337e-05, 'epoch': 1.56}) +[2026-06-23 22:27:59,981][starVLA.training.train_starvla][INFO] - Step 4025, Loss: {'train/loss': 0.13448520004749298, 'train/grad_norm_pre_clip': 22.73443603515625, 'timing/data': 0.0005984960589557886, 'timing/model': 1.691642252029851, 'learning_rate/qwen_vl_interface': 1.3982387684879836e-06, 'learning_rate/action_model': 1.3982387684879836e-05, 'epoch': 1.56}) +[2026-06-23 22:28:03,303][starVLA.training.train_starvla][INFO] - Step 4026, Loss: {'train/loss': 0.08682948350906372, 'train/grad_norm_pre_clip': 14.50098991394043, 'timing/data': 0.0005938271060585976, 'timing/model': 1.687712610932067, 'learning_rate/qwen_vl_interface': 1.3964574138488804e-06, 'learning_rate/action_model': 1.3964574138488803e-05, 'epoch': 1.56}) +[2026-06-23 22:28:06,628][starVLA.training.train_starvla][INFO] - Step 4027, Loss: {'train/loss': 0.09442630410194397, 'train/grad_norm_pre_clip': 29.91018295288086, 'timing/data': 0.0006037249695509672, 'timing/model': 1.6906816000118852, 'learning_rate/qwen_vl_interface': 1.3946776432555743e-06, 'learning_rate/action_model': 1.3946776432555741e-05, 'epoch': 1.56}) +[2026-06-23 22:28:09,959][starVLA.training.train_starvla][INFO] - Step 4028, Loss: {'train/loss': 0.05060841888189316, 'train/grad_norm_pre_clip': 31.231447219848633, 'timing/data': 0.0006993659771978855, 'timing/model': 1.6901643199380487, 'learning_rate/qwen_vl_interface': 1.3928994574396595e-06, 'learning_rate/action_model': 1.3928994574396594e-05, 'epoch': 1.56}) +[2026-06-23 22:28:13,286][starVLA.training.train_starvla][INFO] - Step 4029, Loss: {'train/loss': 0.18168547749519348, 'train/grad_norm_pre_clip': 17.132728576660156, 'timing/data': 0.0005804621614515781, 'timing/model': 1.6882015399169177, 'learning_rate/qwen_vl_interface': 1.391122857132083e-06, 'learning_rate/action_model': 1.3911228571320831e-05, 'epoch': 1.56}) +[2026-06-23 22:28:16,604][starVLA.training.train_starvla][INFO] - Step 4030, Loss: {'train/loss': 0.08828406035900116, 'train/grad_norm_pre_clip': 39.528194427490234, 'timing/data': 0.0006117210723459721, 'timing/model': 1.6900729748886079, 'learning_rate/qwen_vl_interface': 1.3893478430631376e-06, 'learning_rate/action_model': 1.3893478430631376e-05, 'epoch': 1.56}) +[2026-06-23 22:28:19,943][starVLA.training.train_starvla][INFO] - Step 4031, Loss: {'train/loss': 0.052759312093257904, 'train/grad_norm_pre_clip': 31.38217544555664, 'timing/data': 0.000593637116253376, 'timing/model': 1.6994278610218316, 'learning_rate/qwen_vl_interface': 1.3875744159624652e-06, 'learning_rate/action_model': 1.3875744159624652e-05, 'epoch': 1.56}) +[2026-06-23 22:28:23,276][starVLA.training.train_starvla][INFO] - Step 4032, Loss: {'train/loss': 0.060835424810647964, 'train/grad_norm_pre_clip': 10.830671310424805, 'timing/data': 0.0008013679180294275, 'timing/model': 1.6928025980014354, 'learning_rate/qwen_vl_interface': 1.3858025765590525e-06, 'learning_rate/action_model': 1.3858025765590525e-05, 'epoch': 1.56}) +[2026-06-23 22:28:26,607][starVLA.training.train_starvla][INFO] - Step 4033, Loss: {'train/loss': 0.09549947082996368, 'train/grad_norm_pre_clip': 20.0107364654541, 'timing/data': 0.0006930639501661062, 'timing/model': 1.692281189840287, 'learning_rate/qwen_vl_interface': 1.384032325581237e-06, 'learning_rate/action_model': 1.3840323255812371e-05, 'epoch': 1.56}) +[2026-06-23 22:28:29,934][starVLA.training.train_starvla][INFO] - Step 4034, Loss: {'train/loss': 0.1376500278711319, 'train/grad_norm_pre_clip': 42.23332595825195, 'timing/data': 0.0006026639603078365, 'timing/model': 1.6902007600292563, 'learning_rate/qwen_vl_interface': 1.3822636637567025e-06, 'learning_rate/action_model': 1.3822636637567024e-05, 'epoch': 1.56}) +[2026-06-23 22:28:33,347][starVLA.training.train_starvla][INFO] - Step 4035, Loss: {'train/loss': 0.051116034388542175, 'train/grad_norm_pre_clip': 14.441001892089844, 'timing/data': 0.0006015121471136808, 'timing/model': 1.6887821948621422, 'learning_rate/qwen_vl_interface': 1.3804965918124788e-06, 'learning_rate/action_model': 1.3804965918124787e-05, 'epoch': 1.56}) +[2026-06-23 22:28:36,674][starVLA.training.train_starvla][INFO] - Step 4036, Loss: {'train/loss': 0.11072532087564468, 'train/grad_norm_pre_clip': 22.325698852539062, 'timing/data': 0.0007831430993974209, 'timing/model': 1.6919606989249587, 'learning_rate/qwen_vl_interface': 1.3787311104749418e-06, 'learning_rate/action_model': 1.3787311104749418e-05, 'epoch': 1.56}) +[2026-06-23 22:28:40,010][starVLA.training.train_starvla][INFO] - Step 4037, Loss: {'train/loss': 0.1819276511669159, 'train/grad_norm_pre_clip': 17.774641036987305, 'timing/data': 0.0005981950089335442, 'timing/model': 1.6934167728759348, 'learning_rate/qwen_vl_interface': 1.3769672204698149e-06, 'learning_rate/action_model': 1.3769672204698147e-05, 'epoch': 1.56}) +[2026-06-23 22:28:43,349][starVLA.training.train_starvla][INFO] - Step 4038, Loss: {'train/loss': 0.07702109217643738, 'train/grad_norm_pre_clip': 16.96674346923828, 'timing/data': 0.0005862531252205372, 'timing/model': 1.699664952000603, 'learning_rate/qwen_vl_interface': 1.3752049225221664e-06, 'learning_rate/action_model': 1.3752049225221663e-05, 'epoch': 1.56}) +[2026-06-23 22:28:46,685][starVLA.training.train_starvla][INFO] - Step 4039, Loss: {'train/loss': 0.12026531249284744, 'train/grad_norm_pre_clip': 19.172697067260742, 'timing/data': 0.0006164098158478737, 'timing/model': 1.689729003002867, 'learning_rate/qwen_vl_interface': 1.3734442173564103e-06, 'learning_rate/action_model': 1.3734442173564102e-05, 'epoch': 1.56}) +[2026-06-23 22:28:50,022][starVLA.training.train_starvla][INFO] - Step 4040, Loss: {'train/loss': 0.16896194219589233, 'train/grad_norm_pre_clip': 26.587913513183594, 'timing/data': 0.0006077138241380453, 'timing/model': 1.6935159729328007, 'learning_rate/qwen_vl_interface': 1.3716851056963062e-06, 'learning_rate/action_model': 1.3716851056963062e-05, 'epoch': 1.56}) +[2026-06-23 22:28:53,352][starVLA.training.train_starvla][INFO] - Step 4041, Loss: {'train/loss': 0.08198337256908417, 'train/grad_norm_pre_clip': 10.160730361938477, 'timing/data': 0.0005865239072591066, 'timing/model': 1.6919023010414094, 'learning_rate/qwen_vl_interface': 1.3699275882649597e-06, 'learning_rate/action_model': 1.3699275882649596e-05, 'epoch': 1.56}) +[2026-06-23 22:28:56,681][starVLA.training.train_starvla][INFO] - Step 4042, Loss: {'train/loss': 0.1782253533601761, 'train/grad_norm_pre_clip': 20.01380157470703, 'timing/data': 0.0006042569875717163, 'timing/model': 1.6943429070524871, 'learning_rate/qwen_vl_interface': 1.368171665784817e-06, 'learning_rate/action_model': 1.3681716657848171e-05, 'epoch': 1.56}) +[2026-06-23 22:29:00,005][starVLA.training.train_starvla][INFO] - Step 4043, Loss: {'train/loss': 0.1434091180562973, 'train/grad_norm_pre_clip': 45.36277389526367, 'timing/data': 0.0005932159256190062, 'timing/model': 1.690561929019168, 'learning_rate/qwen_vl_interface': 1.3664173389776735e-06, 'learning_rate/action_model': 1.3664173389776735e-05, 'epoch': 1.56}) +[2026-06-23 22:29:03,341][starVLA.training.train_starvla][INFO] - Step 4044, Loss: {'train/loss': 0.06073477491736412, 'train/grad_norm_pre_clip': 19.673704147338867, 'timing/data': 0.0005797108169645071, 'timing/model': 1.6904904989060014, 'learning_rate/qwen_vl_interface': 1.3646646085646658e-06, 'learning_rate/action_model': 1.3646646085646659e-05, 'epoch': 1.57}) +[2026-06-23 22:29:06,672][starVLA.training.train_starvla][INFO] - Step 4045, Loss: {'train/loss': 0.10549822449684143, 'train/grad_norm_pre_clip': 20.55779457092285, 'timing/data': 0.0006045370828360319, 'timing/model': 1.6916395551525056, 'learning_rate/qwen_vl_interface': 1.3629134752662756e-06, 'learning_rate/action_model': 1.3629134752662756e-05, 'epoch': 1.57}) +[2026-06-23 22:29:09,994][starVLA.training.train_starvla][INFO] - Step 4046, Loss: {'train/loss': 0.06743193417787552, 'train/grad_norm_pre_clip': 17.95026206970215, 'timing/data': 0.0005859520751982927, 'timing/model': 1.6877571970690042, 'learning_rate/qwen_vl_interface': 1.3611639398023282e-06, 'learning_rate/action_model': 1.3611639398023283e-05, 'epoch': 1.57}) +[2026-06-23 22:29:13,321][starVLA.training.train_starvla][INFO] - Step 4047, Loss: {'train/loss': 0.1507469117641449, 'train/grad_norm_pre_clip': 37.15028381347656, 'timing/data': 0.0005973440129309893, 'timing/model': 1.6939240540377796, 'learning_rate/qwen_vl_interface': 1.35941600289199e-06, 'learning_rate/action_model': 1.35941600289199e-05, 'epoch': 1.57}) +[2026-06-23 22:29:16,654][starVLA.training.train_starvla][INFO] - Step 4048, Loss: {'train/loss': 0.1466638594865799, 'train/grad_norm_pre_clip': 25.571266174316406, 'timing/data': 0.0005950601771473885, 'timing/model': 1.694686682894826, 'learning_rate/qwen_vl_interface': 1.3576696652537735e-06, 'learning_rate/action_model': 1.3576696652537736e-05, 'epoch': 1.57}) +[2026-06-23 22:29:19,978][starVLA.training.train_starvla][INFO] - Step 4049, Loss: {'train/loss': 0.1765335202217102, 'train/grad_norm_pre_clip': 40.96285629272461, 'timing/data': 0.0005939980037510395, 'timing/model': 1.687021397985518, 'learning_rate/qwen_vl_interface': 1.355924927605531e-06, 'learning_rate/action_model': 1.3559249276055311e-05, 'epoch': 1.57}) +[2026-06-23 22:29:23,469][starVLA.training.train_starvla][INFO] - Step 4050, Loss: {'train/loss': 0.20155014097690582, 'train/grad_norm_pre_clip': 19.58281135559082, 'timing/data': 0.0005987859331071377, 'timing/model': 1.6963351948652416, 'learning_rate/qwen_vl_interface': 1.3541817906644591e-06, 'learning_rate/action_model': 1.3541817906644591e-05, 'epoch': 1.57}) +[2026-06-23 22:29:26,799][starVLA.training.train_starvla][INFO] - Step 4051, Loss: {'train/loss': 0.1696833223104477, 'train/grad_norm_pre_clip': 45.305397033691406, 'timing/data': 0.0005944990552961826, 'timing/model': 1.6899078488349915, 'learning_rate/qwen_vl_interface': 1.3524402551470955e-06, 'learning_rate/action_model': 1.3524402551470956e-05, 'epoch': 1.57}) +[2026-06-23 22:29:30,140][starVLA.training.train_starvla][INFO] - Step 4052, Loss: {'train/loss': 0.11470437794923782, 'train/grad_norm_pre_clip': 18.391027450561523, 'timing/data': 0.0005969929043203592, 'timing/model': 1.695188078097999, 'learning_rate/qwen_vl_interface': 1.3507003217693196e-06, 'learning_rate/action_model': 1.3507003217693196e-05, 'epoch': 1.57}) +[2026-06-23 22:29:33,482][starVLA.training.train_starvla][INFO] - Step 4053, Loss: {'train/loss': 0.1468687802553177, 'train/grad_norm_pre_clip': 52.10504913330078, 'timing/data': 0.0006061398889869452, 'timing/model': 1.6969673109706491, 'learning_rate/qwen_vl_interface': 1.3489619912463515e-06, 'learning_rate/action_model': 1.3489619912463514e-05, 'epoch': 1.57}) +[2026-06-23 22:29:36,829][starVLA.training.train_starvla][INFO] - Step 4054, Loss: {'train/loss': 0.16882020235061646, 'train/grad_norm_pre_clip': 46.45729064941406, 'timing/data': 0.0006112600676715374, 'timing/model': 1.7017236140090972, 'learning_rate/qwen_vl_interface': 1.3472252642927534e-06, 'learning_rate/action_model': 1.3472252642927533e-05, 'epoch': 1.57}) +[2026-06-23 22:29:40,168][starVLA.training.train_starvla][INFO] - Step 4055, Loss: {'train/loss': 0.1732703298330307, 'train/grad_norm_pre_clip': 23.25210189819336, 'timing/data': 0.0005962420254945755, 'timing/model': 1.6893960889428854, 'learning_rate/qwen_vl_interface': 1.3454901416224284e-06, 'learning_rate/action_model': 1.3454901416224284e-05, 'epoch': 1.57}) +[2026-06-23 22:29:43,506][starVLA.training.train_starvla][INFO] - Step 4056, Loss: {'train/loss': 0.09774833917617798, 'train/grad_norm_pre_clip': 37.294090270996094, 'timing/data': 0.0006031249649822712, 'timing/model': 1.690473390975967, 'learning_rate/qwen_vl_interface': 1.34375662394862e-06, 'learning_rate/action_model': 1.34375662394862e-05, 'epoch': 1.57}) +[2026-06-23 22:29:46,847][starVLA.training.train_starvla][INFO] - Step 4057, Loss: {'train/loss': 0.1719098687171936, 'train/grad_norm_pre_clip': 22.04265022277832, 'timing/data': 0.0006176219321787357, 'timing/model': 1.698699533008039, 'learning_rate/qwen_vl_interface': 1.3420247119839116e-06, 'learning_rate/action_model': 1.3420247119839117e-05, 'epoch': 1.57}) +[2026-06-23 22:29:50,180][starVLA.training.train_starvla][INFO] - Step 4058, Loss: {'train/loss': 0.21647292375564575, 'train/grad_norm_pre_clip': 48.80430603027344, 'timing/data': 0.0006020220462232828, 'timing/model': 1.6942231878638268, 'learning_rate/qwen_vl_interface': 1.340294406440227e-06, 'learning_rate/action_model': 1.340294406440227e-05, 'epoch': 1.57}) +[2026-06-23 22:29:53,519][starVLA.training.train_starvla][INFO] - Step 4059, Loss: {'train/loss': 0.11911619454622269, 'train/grad_norm_pre_clip': 47.56470489501953, 'timing/data': 0.0007274590898305178, 'timing/model': 1.6951713680755347, 'learning_rate/qwen_vl_interface': 1.3385657080288297e-06, 'learning_rate/action_model': 1.3385657080288297e-05, 'epoch': 1.57}) +[2026-06-23 22:29:56,854][starVLA.training.train_starvla][INFO] - Step 4060, Loss: {'train/loss': 0.11232890188694, 'train/grad_norm_pre_clip': 26.15401268005371, 'timing/data': 0.0006479090079665184, 'timing/model': 1.6952800990547985, 'learning_rate/qwen_vl_interface': 1.336838617460322e-06, 'learning_rate/action_model': 1.336838617460322e-05, 'epoch': 1.57}) +[2026-06-23 22:30:00,199][starVLA.training.train_starvla][INFO] - Step 4061, Loss: {'train/loss': 0.09052329510450363, 'train/grad_norm_pre_clip': 18.619548797607422, 'timing/data': 0.0005962410941720009, 'timing/model': 1.7005923381075263, 'learning_rate/qwen_vl_interface': 1.335113135444646e-06, 'learning_rate/action_model': 1.335113135444646e-05, 'epoch': 1.57}) +[2026-06-23 22:30:03,540][starVLA.training.train_starvla][INFO] - Step 4062, Loss: {'train/loss': 0.12957027554512024, 'train/grad_norm_pre_clip': 17.347341537475586, 'timing/data': 0.0005977649707347155, 'timing/model': 1.6983119749929756, 'learning_rate/qwen_vl_interface': 1.3333892626910823e-06, 'learning_rate/action_model': 1.3333892626910823e-05, 'epoch': 1.57}) +[2026-06-23 22:30:06,884][starVLA.training.train_starvla][INFO] - Step 4063, Loss: {'train/loss': 0.06470318883657455, 'train/grad_norm_pre_clip': 12.459220886230469, 'timing/data': 0.0005922638811171055, 'timing/model': 1.6973139240872115, 'learning_rate/qwen_vl_interface': 1.3316669999082486e-06, 'learning_rate/action_model': 1.3316669999082484e-05, 'epoch': 1.57}) +[2026-06-23 22:30:10,348][starVLA.training.train_starvla][INFO] - Step 4064, Loss: {'train/loss': 0.09396462142467499, 'train/grad_norm_pre_clip': 15.68787670135498, 'timing/data': 0.0005942480638623238, 'timing/model': 1.8208427070640028, 'learning_rate/qwen_vl_interface': 1.3299463478041024e-06, 'learning_rate/action_model': 1.3299463478041024e-05, 'epoch': 1.57}) +[2026-06-23 22:30:13,690][starVLA.training.train_starvla][INFO] - Step 4065, Loss: {'train/loss': 0.0654936134815216, 'train/grad_norm_pre_clip': 22.63469886779785, 'timing/data': 0.00060454779304564, 'timing/model': 1.693451460218057, 'learning_rate/qwen_vl_interface': 1.3282273070859396e-06, 'learning_rate/action_model': 1.3282273070859396e-05, 'epoch': 1.57}) +[2026-06-23 22:30:17,026][starVLA.training.train_starvla][INFO] - Step 4066, Loss: {'train/loss': 0.1857660859823227, 'train/grad_norm_pre_clip': 46.2398681640625, 'timing/data': 0.0005989070050418377, 'timing/model': 1.6926861281972378, 'learning_rate/qwen_vl_interface': 1.3265098784603924e-06, 'learning_rate/action_model': 1.3265098784603924e-05, 'epoch': 1.57}) +[2026-06-23 22:30:20,364][starVLA.training.train_starvla][INFO] - Step 4067, Loss: {'train/loss': 0.11210882663726807, 'train/grad_norm_pre_clip': 18.56040382385254, 'timing/data': 0.0007356840651482344, 'timing/model': 1.694892565952614, 'learning_rate/qwen_vl_interface': 1.324794062633431e-06, 'learning_rate/action_model': 1.324794062633431e-05, 'epoch': 1.57}) +[2026-06-23 22:30:23,697][starVLA.training.train_starvla][INFO] - Step 4068, Loss: {'train/loss': 0.09048208594322205, 'train/grad_norm_pre_clip': 46.42323303222656, 'timing/data': 0.000655272975564003, 'timing/model': 1.6957010328769684, 'learning_rate/qwen_vl_interface': 1.323079860310363e-06, 'learning_rate/action_model': 1.3230798603103628e-05, 'epoch': 1.57}) +[2026-06-23 22:30:27,034][starVLA.training.train_starvla][INFO] - Step 4069, Loss: {'train/loss': 0.06887877732515335, 'train/grad_norm_pre_clip': 10.430889129638672, 'timing/data': 0.0006696600466966629, 'timing/model': 1.6943270959891379, 'learning_rate/qwen_vl_interface': 1.3213672721958309e-06, 'learning_rate/action_model': 1.321367272195831e-05, 'epoch': 1.57}) +[2026-06-23 22:30:30,368][starVLA.training.train_starvla][INFO] - Step 4070, Loss: {'train/loss': 0.07262134552001953, 'train/grad_norm_pre_clip': 15.34787368774414, 'timing/data': 0.000596832949668169, 'timing/model': 1.6927823070436716, 'learning_rate/qwen_vl_interface': 1.3196562989938175e-06, 'learning_rate/action_model': 1.3196562989938174e-05, 'epoch': 1.58}) +[2026-06-23 22:30:33,702][starVLA.training.train_starvla][INFO] - Step 4071, Loss: {'train/loss': 0.1498870998620987, 'train/grad_norm_pre_clip': 59.05705642700195, 'timing/data': 0.0005972229409962893, 'timing/model': 1.6923682349734008, 'learning_rate/qwen_vl_interface': 1.3179469414076357e-06, 'learning_rate/action_model': 1.3179469414076356e-05, 'epoch': 1.58}) +[2026-06-23 22:30:37,046][starVLA.training.train_starvla][INFO] - Step 4072, Loss: {'train/loss': 0.08655570447444916, 'train/grad_norm_pre_clip': 34.112266540527344, 'timing/data': 0.0009432937949895859, 'timing/model': 1.6989244760479778, 'learning_rate/qwen_vl_interface': 1.3162392001399401e-06, 'learning_rate/action_model': 1.3162392001399401e-05, 'epoch': 1.58}) +[2026-06-23 22:30:40,383][starVLA.training.train_starvla][INFO] - Step 4073, Loss: {'train/loss': 0.12783250212669373, 'train/grad_norm_pre_clip': 42.32171630859375, 'timing/data': 0.0006777648814022541, 'timing/model': 1.6915218881331384, 'learning_rate/qwen_vl_interface': 1.3145330758927182e-06, 'learning_rate/action_model': 1.3145330758927182e-05, 'epoch': 1.58}) +[2026-06-23 22:30:43,718][starVLA.training.train_starvla][INFO] - Step 4074, Loss: {'train/loss': 0.028605472296476364, 'train/grad_norm_pre_clip': 17.546222686767578, 'timing/data': 0.0006607428658753633, 'timing/model': 1.696739090140909, 'learning_rate/qwen_vl_interface': 1.3128285693672935e-06, 'learning_rate/action_model': 1.3128285693672934e-05, 'epoch': 1.58}) +[2026-06-23 22:30:47,046][starVLA.training.train_starvla][INFO] - Step 4075, Loss: {'train/loss': 0.15757699310779572, 'train/grad_norm_pre_clip': 17.283889770507812, 'timing/data': 0.0006958788726478815, 'timing/model': 1.6922965100966394, 'learning_rate/qwen_vl_interface': 1.311125681264325e-06, 'learning_rate/action_model': 1.311125681264325e-05, 'epoch': 1.58}) +[2026-06-23 22:30:50,383][starVLA.training.train_starvla][INFO] - Step 4076, Loss: {'train/loss': 0.041212134063243866, 'train/grad_norm_pre_clip': 12.050932884216309, 'timing/data': 0.0005898999515920877, 'timing/model': 1.6916038349736482, 'learning_rate/qwen_vl_interface': 1.309424412283803e-06, 'learning_rate/action_model': 1.3094244122838029e-05, 'epoch': 1.58}) +[2026-06-23 22:30:53,718][starVLA.training.train_starvla][INFO] - Step 4077, Loss: {'train/loss': 0.15650759637355804, 'train/grad_norm_pre_clip': 27.05176544189453, 'timing/data': 0.0006399441044777632, 'timing/model': 1.6950225329492241, 'learning_rate/qwen_vl_interface': 1.307724763125057e-06, 'learning_rate/action_model': 1.307724763125057e-05, 'epoch': 1.58}) +[2026-06-23 22:30:57,048][starVLA.training.train_starvla][INFO] - Step 4078, Loss: {'train/loss': 0.12840375304222107, 'train/grad_norm_pre_clip': 21.20367431640625, 'timing/data': 0.000661924947053194, 'timing/model': 1.6863425639458, 'learning_rate/qwen_vl_interface': 1.3060267344867478e-06, 'learning_rate/action_model': 1.3060267344867478e-05, 'epoch': 1.58}) +[2026-06-23 22:31:00,661][starVLA.training.train_starvla][INFO] - Step 4079, Loss: {'train/loss': 0.05077871307730675, 'train/grad_norm_pre_clip': 18.6762638092041, 'timing/data': 0.0005955910310149193, 'timing/model': 1.9697553559672087, 'learning_rate/qwen_vl_interface': 1.3043303270668715e-06, 'learning_rate/action_model': 1.3043303270668713e-05, 'epoch': 1.58}) +[2026-06-23 22:31:03,999][starVLA.training.train_starvla][INFO] - Step 4080, Loss: {'train/loss': 0.08919425308704376, 'train/grad_norm_pre_clip': 23.138185501098633, 'timing/data': 0.0006047580391168594, 'timing/model': 1.6961501881014556, 'learning_rate/qwen_vl_interface': 1.3026355415627565e-06, 'learning_rate/action_model': 1.3026355415627564e-05, 'epoch': 1.58}) +[2026-06-23 22:31:07,331][starVLA.training.train_starvla][INFO] - Step 4081, Loss: {'train/loss': 0.08840563148260117, 'train/grad_norm_pre_clip': 49.181209564208984, 'timing/data': 0.0006458039861172438, 'timing/model': 1.6901526430156082, 'learning_rate/qwen_vl_interface': 1.3009423786710653e-06, 'learning_rate/action_model': 1.3009423786710653e-05, 'epoch': 1.58}) +[2026-06-23 22:31:10,664][starVLA.training.train_starvla][INFO] - Step 4082, Loss: {'train/loss': 0.04302005097270012, 'train/grad_norm_pre_clip': 18.295080184936523, 'timing/data': 0.0006790270563215017, 'timing/model': 1.6955752121284604, 'learning_rate/qwen_vl_interface': 1.2992508390877933e-06, 'learning_rate/action_model': 1.2992508390877933e-05, 'epoch': 1.58}) +[2026-06-23 22:31:14,013][starVLA.training.train_starvla][INFO] - Step 4083, Loss: {'train/loss': 0.1311858743429184, 'train/grad_norm_pre_clip': 59.30284881591797, 'timing/data': 0.0007335601840168238, 'timing/model': 1.6993238818831742, 'learning_rate/qwen_vl_interface': 1.297560923508269e-06, 'learning_rate/action_model': 1.297560923508269e-05, 'epoch': 1.58}) +[2026-06-23 22:31:17,355][starVLA.training.train_starvla][INFO] - Step 4084, Loss: {'train/loss': 0.06524717062711716, 'train/grad_norm_pre_clip': 31.786161422729492, 'timing/data': 0.0008280379697680473, 'timing/model': 1.6973561390768737, 'learning_rate/qwen_vl_interface': 1.295872632627152e-06, 'learning_rate/action_model': 1.2958726326271519e-05, 'epoch': 1.58}) +[2026-06-23 22:31:20,688][starVLA.training.train_starvla][INFO] - Step 4085, Loss: {'train/loss': 0.2581898272037506, 'train/grad_norm_pre_clip': 41.83467483520508, 'timing/data': 0.0005856421776115894, 'timing/model': 1.6969436809886247, 'learning_rate/qwen_vl_interface': 1.2941859671384364e-06, 'learning_rate/action_model': 1.2941859671384362e-05, 'epoch': 1.58}) +[2026-06-23 22:31:24,029][starVLA.training.train_starvla][INFO] - Step 4086, Loss: {'train/loss': 0.1121203824877739, 'train/grad_norm_pre_clip': 21.73116683959961, 'timing/data': 0.0006002089940011501, 'timing/model': 1.6928275069221854, 'learning_rate/qwen_vl_interface': 1.2925009277354444e-06, 'learning_rate/action_model': 1.2925009277354444e-05, 'epoch': 1.58}) +[2026-06-23 22:31:27,368][starVLA.training.train_starvla][INFO] - Step 4087, Loss: {'train/loss': 0.08489231020212173, 'train/grad_norm_pre_clip': 22.671300888061523, 'timing/data': 0.0005931858904659748, 'timing/model': 1.6964082398917526, 'learning_rate/qwen_vl_interface': 1.2908175151108335e-06, 'learning_rate/action_model': 1.2908175151108334e-05, 'epoch': 1.58}) +[2026-06-23 22:31:30,705][starVLA.training.train_starvla][INFO] - Step 4088, Loss: {'train/loss': 0.15577921271324158, 'train/grad_norm_pre_clip': 33.4138069152832, 'timing/data': 0.0007157458458095789, 'timing/model': 1.6962306038476527, 'learning_rate/qwen_vl_interface': 1.2891357299565899e-06, 'learning_rate/action_model': 1.2891357299565898e-05, 'epoch': 1.58}) +[2026-06-23 22:31:34,045][starVLA.training.train_starvla][INFO] - Step 4089, Loss: {'train/loss': 0.11899325996637344, 'train/grad_norm_pre_clip': 20.526086807250977, 'timing/data': 0.0008216260466724634, 'timing/model': 1.6963275158777833, 'learning_rate/qwen_vl_interface': 1.2874555729640333e-06, 'learning_rate/action_model': 1.2874555729640334e-05, 'epoch': 1.58}) +[2026-06-23 22:31:37,382][starVLA.training.train_starvla][INFO] - Step 4090, Loss: {'train/loss': 0.08107844740152359, 'train/grad_norm_pre_clip': 21.814210891723633, 'timing/data': 0.0006910099182277918, 'timing/model': 1.696740121114999, 'learning_rate/qwen_vl_interface': 1.2857770448238124e-06, 'learning_rate/action_model': 1.2857770448238123e-05, 'epoch': 1.58}) +[2026-06-23 22:31:40,729][starVLA.training.train_starvla][INFO] - Step 4091, Loss: {'train/loss': 0.1538754105567932, 'train/grad_norm_pre_clip': 13.540708541870117, 'timing/data': 0.0007052370347082615, 'timing/model': 1.6942336650099605, 'learning_rate/qwen_vl_interface': 1.2841001462259067e-06, 'learning_rate/action_model': 1.2841001462259067e-05, 'epoch': 1.58}) +[2026-06-23 22:31:44,076][starVLA.training.train_starvla][INFO] - Step 4092, Loss: {'train/loss': 0.3269965350627899, 'train/grad_norm_pre_clip': 24.306196212768555, 'timing/data': 0.0006363771390169859, 'timing/model': 1.6876821708865464, 'learning_rate/qwen_vl_interface': 1.282424877859626e-06, 'learning_rate/action_model': 1.282424877859626e-05, 'epoch': 1.58}) +[2026-06-23 22:31:47,479][starVLA.training.train_starvla][INFO] - Step 4093, Loss: {'train/loss': 0.08807893842458725, 'train/grad_norm_pre_clip': 14.304141998291016, 'timing/data': 0.0005812239833176136, 'timing/model': 1.767873550998047, 'learning_rate/qwen_vl_interface': 1.2807512404136088e-06, 'learning_rate/action_model': 1.2807512404136087e-05, 'epoch': 1.58}) +[2026-06-23 22:31:50,816][starVLA.training.train_starvla][INFO] - Step 4094, Loss: {'train/loss': 0.18700598180294037, 'train/grad_norm_pre_clip': 24.671274185180664, 'timing/data': 0.0006839670240879059, 'timing/model': 1.6968666738830507, 'learning_rate/qwen_vl_interface': 1.2790792345758255e-06, 'learning_rate/action_model': 1.2790792345758254e-05, 'epoch': 1.58}) +[2026-06-23 22:31:54,143][starVLA.training.train_starvla][INFO] - Step 4095, Loss: {'train/loss': 0.06688446551561356, 'train/grad_norm_pre_clip': 27.054330825805664, 'timing/data': 0.000650262925773859, 'timing/model': 1.6893637711182237, 'learning_rate/qwen_vl_interface': 1.2774088610335752e-06, 'learning_rate/action_model': 1.2774088610335752e-05, 'epoch': 1.58}) +[2026-06-23 22:31:57,461][starVLA.training.train_starvla][INFO] - Step 4096, Loss: {'train/loss': 0.04006850719451904, 'train/grad_norm_pre_clip': 27.178709030151367, 'timing/data': 0.0006734458729624748, 'timing/model': 1.6844676800537854, 'learning_rate/qwen_vl_interface': 1.2757401204734827e-06, 'learning_rate/action_model': 1.2757401204734826e-05, 'epoch': 1.59}) +[2026-06-23 22:32:00,788][starVLA.training.train_starvla][INFO] - Step 4097, Loss: {'train/loss': 0.17395979166030884, 'train/grad_norm_pre_clip': 23.128196716308594, 'timing/data': 0.0005504048895090818, 'timing/model': 1.6916150241158903, 'learning_rate/qwen_vl_interface': 1.2740730135815064e-06, 'learning_rate/action_model': 1.2740730135815062e-05, 'epoch': 1.59}) +[2026-06-23 22:32:04,115][starVLA.training.train_starvla][INFO] - Step 4098, Loss: {'train/loss': 0.11706554889678955, 'train/grad_norm_pre_clip': 17.893627166748047, 'timing/data': 0.0005572990048676729, 'timing/model': 1.6881115029100329, 'learning_rate/qwen_vl_interface': 1.2724075410429298e-06, 'learning_rate/action_model': 1.2724075410429299e-05, 'epoch': 1.59}) +[2026-06-23 22:32:07,441][starVLA.training.train_starvla][INFO] - Step 4099, Loss: {'train/loss': 0.0388215072453022, 'train/grad_norm_pre_clip': 15.57386589050293, 'timing/data': 0.0005823851097375154, 'timing/model': 1.6876755519770086, 'learning_rate/qwen_vl_interface': 1.2707437035423664e-06, 'learning_rate/action_model': 1.2707437035423664e-05, 'epoch': 1.59}) +[2026-06-23 22:32:10,770][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4100 +[2026-06-23 22:33:42,761][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4100 +[2026-06-23 22:33:42,762][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4100 +[2026-06-23 22:34:53,716][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4100 +[2026-06-23 22:34:53,716][starVLA.training.train_starvla][INFO] - Step 4100, Loss: {'train/loss': 0.128099724650383, 'train/grad_norm_pre_clip': 28.829004287719727, 'eval/action_loss/seconds': 91.98890518001281, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.0661616250872612, 'eval/latency_2/loss': 0.0661616250872612, 'eval/flappy/loss': 0.0661616250872612, 'eval/flappy/latency_2/loss': 0.0661616250872612, 'eval/flap_precision': 0.9467213114754098, 'eval/flap_recall': 0.9352226720647774, 'eval/flap_f1': 0.9409368635437882, 'eval/flap_rate_model': 0.20333333333333334, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9409368635437881, 'eval/latency_2/flap_precision': 0.9467213114754098, 'eval/latency_2/flap_recall': 0.9352226720647774, 'eval/latency_2/flap_f1': 0.9409368635437882, 'eval/latency_2/flap_rate_model': 0.20333333333333334, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9409368635437881, 'eval/action_classification/data_wait_seconds': 0.9617764353752136, 'eval/action_classification/predict_seconds': 69.87118530273438, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005484709981828928, 'timing/model': 1.6899964071344584, 'learning_rate/qwen_vl_interface': 1.2690815017637573e-06, 'learning_rate/action_model': 1.2690815017637572e-05, 'epoch': 1.59}) +[2026-06-23 22:34:57,127][starVLA.training.train_starvla][INFO] - Step 4101, Loss: {'train/loss': 0.05995533987879753, 'train/grad_norm_pre_clip': 23.717967987060547, 'timing/data': 0.0006122908089309931, 'timing/model': 1.7833417870569974, 'learning_rate/qwen_vl_interface': 1.2674209363903694e-06, 'learning_rate/action_model': 1.2674209363903693e-05, 'epoch': 1.59}) +[2026-06-23 22:35:00,439][starVLA.training.train_starvla][INFO] - Step 4102, Loss: {'train/loss': 0.04077250137925148, 'train/grad_norm_pre_clip': 20.270524978637695, 'timing/data': 0.0006606329698115587, 'timing/model': 1.6793006770312786, 'learning_rate/qwen_vl_interface': 1.2657620081048005e-06, 'learning_rate/action_model': 1.2657620081048006e-05, 'epoch': 1.59}) +[2026-06-23 22:35:03,766][starVLA.training.train_starvla][INFO] - Step 4103, Loss: {'train/loss': 0.15993933379650116, 'train/grad_norm_pre_clip': 50.965091705322266, 'timing/data': 0.0008563208393752575, 'timing/model': 1.694379820022732, 'learning_rate/qwen_vl_interface': 1.2641047175889714e-06, 'learning_rate/action_model': 1.2641047175889714e-05, 'epoch': 1.59}) +[2026-06-23 22:35:07,084][starVLA.training.train_starvla][INFO] - Step 4104, Loss: {'train/loss': 0.1452513188123703, 'train/grad_norm_pre_clip': 22.42953109741211, 'timing/data': 0.0005698809400200844, 'timing/model': 1.6846672049723566, 'learning_rate/qwen_vl_interface': 1.2624490655241332e-06, 'learning_rate/action_model': 1.262449065524133e-05, 'epoch': 1.59}) +[2026-06-23 22:35:10,406][starVLA.training.train_starvla][INFO] - Step 4105, Loss: {'train/loss': 0.23880814015865326, 'train/grad_norm_pre_clip': 48.197086334228516, 'timing/data': 0.0005663749761879444, 'timing/model': 1.6843789061531425, 'learning_rate/qwen_vl_interface': 1.2607950525908602e-06, 'learning_rate/action_model': 1.2607950525908601e-05, 'epoch': 1.59}) +[2026-06-23 22:35:13,719][starVLA.training.train_starvla][INFO] - Step 4106, Loss: {'train/loss': 0.13530778884887695, 'train/grad_norm_pre_clip': 22.060226440429688, 'timing/data': 0.0005589418578892946, 'timing/model': 1.681956933112815, 'learning_rate/qwen_vl_interface': 1.2591426794690575e-06, 'learning_rate/action_model': 1.2591426794690576e-05, 'epoch': 1.59}) +[2026-06-23 22:35:17,030][starVLA.training.train_starvla][INFO] - Step 4107, Loss: {'train/loss': 0.07221159338951111, 'train/grad_norm_pre_clip': 16.24174690246582, 'timing/data': 0.0005834170151501894, 'timing/model': 1.683269294211641, 'learning_rate/qwen_vl_interface': 1.25749194683795e-06, 'learning_rate/action_model': 1.25749194683795e-05, 'epoch': 1.59}) +[2026-06-23 22:35:20,345][starVLA.training.train_starvla][INFO] - Step 4108, Loss: {'train/loss': 0.11594300717115402, 'train/grad_norm_pre_clip': 15.818978309631348, 'timing/data': 0.0005496640224009752, 'timing/model': 1.6847530480008572, 'learning_rate/qwen_vl_interface': 1.2558428553760933e-06, 'learning_rate/action_model': 1.2558428553760932e-05, 'epoch': 1.59}) +[2026-06-23 22:35:23,661][starVLA.training.train_starvla][INFO] - Step 4109, Loss: {'train/loss': 0.17752806842327118, 'train/grad_norm_pre_clip': 17.76166534423828, 'timing/data': 0.0005386031698435545, 'timing/model': 1.6868194669950753, 'learning_rate/qwen_vl_interface': 1.2541954057613659e-06, 'learning_rate/action_model': 1.2541954057613658e-05, 'epoch': 1.59}) +[2026-06-23 22:35:26,976][starVLA.training.train_starvla][INFO] - Step 4110, Loss: {'train/loss': 0.07163034379482269, 'train/grad_norm_pre_clip': 12.734057426452637, 'timing/data': 0.0005793888121843338, 'timing/model': 1.6828761028591543, 'learning_rate/qwen_vl_interface': 1.2525495986709727e-06, 'learning_rate/action_model': 1.2525495986709726e-05, 'epoch': 1.59}) +[2026-06-23 22:35:30,291][starVLA.training.train_starvla][INFO] - Step 4111, Loss: {'train/loss': 0.11590085178613663, 'train/grad_norm_pre_clip': 27.4788818359375, 'timing/data': 0.0005771759897470474, 'timing/model': 1.6841037150006741, 'learning_rate/qwen_vl_interface': 1.2509054347814423e-06, 'learning_rate/action_model': 1.2509054347814423e-05, 'epoch': 1.59}) +[2026-06-23 22:35:33,609][starVLA.training.train_starvla][INFO] - Step 4112, Loss: {'train/loss': 0.1370447278022766, 'train/grad_norm_pre_clip': 17.426795959472656, 'timing/data': 0.0006748491432517767, 'timing/model': 1.6851198489312083, 'learning_rate/qwen_vl_interface': 1.2492629147686288e-06, 'learning_rate/action_model': 1.2492629147686289e-05, 'epoch': 1.59}) +[2026-06-23 22:35:36,922][starVLA.training.train_starvla][INFO] - Step 4113, Loss: {'train/loss': 0.10954336822032928, 'train/grad_norm_pre_clip': 17.588333129882812, 'timing/data': 0.0005766451358795166, 'timing/model': 1.6835369758773595, 'learning_rate/qwen_vl_interface': 1.2476220393077098e-06, 'learning_rate/action_model': 1.2476220393077098e-05, 'epoch': 1.59}) +[2026-06-23 22:35:40,237][starVLA.training.train_starvla][INFO] - Step 4114, Loss: {'train/loss': 0.1314684897661209, 'train/grad_norm_pre_clip': 17.297422409057617, 'timing/data': 0.0005814938340336084, 'timing/model': 1.685308659914881, 'learning_rate/qwen_vl_interface': 1.2459828090731877e-06, 'learning_rate/action_model': 1.2459828090731876e-05, 'epoch': 1.59}) +[2026-06-23 22:35:43,553][starVLA.training.train_starvla][INFO] - Step 4115, Loss: {'train/loss': 0.05189267918467522, 'train/grad_norm_pre_clip': 25.862409591674805, 'timing/data': 0.0005804211832582951, 'timing/model': 1.6853166080545634, 'learning_rate/qwen_vl_interface': 1.2443452247388876e-06, 'learning_rate/action_model': 1.2443452247388877e-05, 'epoch': 1.59}) +[2026-06-23 22:35:47,174][starVLA.training.train_starvla][INFO] - Step 4116, Loss: {'train/loss': 0.09469486773014069, 'train/grad_norm_pre_clip': 37.64415740966797, 'timing/data': 0.0005681691691279411, 'timing/model': 1.9866976260673255, 'learning_rate/qwen_vl_interface': 1.2427092869779603e-06, 'learning_rate/action_model': 1.2427092869779602e-05, 'epoch': 1.59}) +[2026-06-23 22:35:50,488][starVLA.training.train_starvla][INFO] - Step 4117, Loss: {'train/loss': 0.15544641017913818, 'train/grad_norm_pre_clip': 26.62335205078125, 'timing/data': 0.0005680280737578869, 'timing/model': 1.6826514010317624, 'learning_rate/qwen_vl_interface': 1.2410749964628751e-06, 'learning_rate/action_model': 1.2410749964628751e-05, 'epoch': 1.59}) +[2026-06-23 22:35:53,803][starVLA.training.train_starvla][INFO] - Step 4118, Loss: {'train/loss': 0.07261335104703903, 'train/grad_norm_pre_clip': 30.58224868774414, 'timing/data': 0.000560784013941884, 'timing/model': 1.6858927111607045, 'learning_rate/qwen_vl_interface': 1.2394423538654296e-06, 'learning_rate/action_model': 1.2394423538654296e-05, 'epoch': 1.59}) +[2026-06-23 22:35:57,117][starVLA.training.train_starvla][INFO] - Step 4119, Loss: {'train/loss': 0.093783438205719, 'train/grad_norm_pre_clip': 14.847965240478516, 'timing/data': 0.0005627088248729706, 'timing/model': 1.6795529809314758, 'learning_rate/qwen_vl_interface': 1.2378113598567394e-06, 'learning_rate/action_model': 1.2378113598567392e-05, 'epoch': 1.59}) +[2026-06-23 22:36:00,451][starVLA.training.train_starvla][INFO] - Step 4120, Loss: {'train/loss': 0.07892091572284698, 'train/grad_norm_pre_clip': 13.641613006591797, 'timing/data': 0.0006286920979619026, 'timing/model': 1.6963882560376078, 'learning_rate/qwen_vl_interface': 1.2361820151072454e-06, 'learning_rate/action_model': 1.2361820151072453e-05, 'epoch': 1.59}) +[2026-06-23 22:36:03,770][starVLA.training.train_starvla][INFO] - Step 4121, Loss: {'train/loss': 0.06064816564321518, 'train/grad_norm_pre_clip': 19.90686798095703, 'timing/data': 0.0005853008478879929, 'timing/model': 1.6849525719881058, 'learning_rate/qwen_vl_interface': 1.23455432028671e-06, 'learning_rate/action_model': 1.23455432028671e-05, 'epoch': 1.59}) +[2026-06-23 22:36:07,088][starVLA.training.train_starvla][INFO] - Step 4122, Loss: {'train/loss': 0.07974795997142792, 'train/grad_norm_pre_clip': 30.766098022460938, 'timing/data': 0.0006348041351884604, 'timing/model': 1.685965842101723, 'learning_rate/qwen_vl_interface': 1.2329282760642167e-06, 'learning_rate/action_model': 1.2329282760642166e-05, 'epoch': 1.6}) +[2026-06-23 22:36:10,412][starVLA.training.train_starvla][INFO] - Step 4123, Loss: {'train/loss': 0.10637902468442917, 'train/grad_norm_pre_clip': 30.857643127441406, 'timing/data': 0.0005838680081069469, 'timing/model': 1.6884899821598083, 'learning_rate/qwen_vl_interface': 1.2313038831081716e-06, 'learning_rate/action_model': 1.2313038831081717e-05, 'epoch': 1.6}) +[2026-06-23 22:36:13,733][starVLA.training.train_starvla][INFO] - Step 4124, Loss: {'train/loss': 0.17323064804077148, 'train/grad_norm_pre_clip': 26.741634368896484, 'timing/data': 0.0005797599442303181, 'timing/model': 1.6854540759231895, 'learning_rate/qwen_vl_interface': 1.2296811420863002e-06, 'learning_rate/action_model': 1.2296811420863003e-05, 'epoch': 1.6}) +[2026-06-23 22:36:17,052][starVLA.training.train_starvla][INFO] - Step 4125, Loss: {'train/loss': 0.06007838994264603, 'train/grad_norm_pre_clip': 25.97695541381836, 'timing/data': 0.0005622571334242821, 'timing/model': 1.6871177689172328, 'learning_rate/qwen_vl_interface': 1.2280600536656504e-06, 'learning_rate/action_model': 1.2280600536656503e-05, 'epoch': 1.6}) +[2026-06-23 22:36:20,369][starVLA.training.train_starvla][INFO] - Step 4126, Loss: {'train/loss': 0.3070284426212311, 'train/grad_norm_pre_clip': 22.247060775756836, 'timing/data': 0.0005433419719338417, 'timing/model': 1.6854650538880378, 'learning_rate/qwen_vl_interface': 1.2264406185125907e-06, 'learning_rate/action_model': 1.2264406185125907e-05, 'epoch': 1.6}) +[2026-06-23 22:36:23,688][starVLA.training.train_starvla][INFO] - Step 4127, Loss: {'train/loss': 0.1153089627623558, 'train/grad_norm_pre_clip': 31.881982803344727, 'timing/data': 0.0005467280279844999, 'timing/model': 1.6869273481424898, 'learning_rate/qwen_vl_interface': 1.2248228372928098e-06, 'learning_rate/action_model': 1.2248228372928097e-05, 'epoch': 1.6}) +[2026-06-23 22:36:27,013][starVLA.training.train_starvla][INFO] - Step 4128, Loss: {'train/loss': 0.0497327484190464, 'train/grad_norm_pre_clip': 16.334545135498047, 'timing/data': 0.0006925419438630342, 'timing/model': 1.6858532920014113, 'learning_rate/qwen_vl_interface': 1.2232067106713174e-06, 'learning_rate/action_model': 1.2232067106713174e-05, 'epoch': 1.6}) +[2026-06-23 22:36:30,331][starVLA.training.train_starvla][INFO] - Step 4129, Loss: {'train/loss': 0.17340165376663208, 'train/grad_norm_pre_clip': 31.632368087768555, 'timing/data': 0.0005753820296376944, 'timing/model': 1.6825144661124796, 'learning_rate/qwen_vl_interface': 1.22159223931244e-06, 'learning_rate/action_model': 1.2215922393124399e-05, 'epoch': 1.6}) +[2026-06-23 22:36:33,648][starVLA.training.train_starvla][INFO] - Step 4130, Loss: {'train/loss': 0.03372177481651306, 'train/grad_norm_pre_clip': 34.2581672668457, 'timing/data': 0.0006684870459139347, 'timing/model': 1.6853373681660742, 'learning_rate/qwen_vl_interface': 1.2199794238798266e-06, 'learning_rate/action_model': 1.2199794238798264e-05, 'epoch': 1.6}) +[2026-06-23 22:36:37,276][starVLA.training.train_starvla][INFO] - Step 4131, Loss: {'train/loss': 0.1122693344950676, 'train/grad_norm_pre_clip': 15.032608985900879, 'timing/data': 0.0005687589291483164, 'timing/model': 1.991079119965434, 'learning_rate/qwen_vl_interface': 1.2183682650364447e-06, 'learning_rate/action_model': 1.2183682650364447e-05, 'epoch': 1.6}) +[2026-06-23 22:36:40,595][starVLA.training.train_starvla][INFO] - Step 4132, Loss: {'train/loss': 0.10112817585468292, 'train/grad_norm_pre_clip': 15.057962417602539, 'timing/data': 0.0005641609895974398, 'timing/model': 1.6846064520068467, 'learning_rate/qwen_vl_interface': 1.216758763444581e-06, 'learning_rate/action_model': 1.2167587634445809e-05, 'epoch': 1.6}) +[2026-06-23 22:36:43,910][starVLA.training.train_starvla][INFO] - Step 4133, Loss: {'train/loss': 0.0484621599316597, 'train/grad_norm_pre_clip': 12.619375228881836, 'timing/data': 0.0005552140064537525, 'timing/model': 1.6841398179531097, 'learning_rate/qwen_vl_interface': 1.215150919765841e-06, 'learning_rate/action_model': 1.215150919765841e-05, 'epoch': 1.6}) +[2026-06-23 22:36:47,229][starVLA.training.train_starvla][INFO] - Step 4134, Loss: {'train/loss': 0.06711174547672272, 'train/grad_norm_pre_clip': 18.744827270507812, 'timing/data': 0.0005413778126239777, 'timing/model': 1.6856739269569516, 'learning_rate/qwen_vl_interface': 1.213544734661148e-06, 'learning_rate/action_model': 1.2135447346611478e-05, 'epoch': 1.6}) +[2026-06-23 22:36:50,545][starVLA.training.train_starvla][INFO] - Step 4135, Loss: {'train/loss': 0.09673488140106201, 'train/grad_norm_pre_clip': 28.604516983032227, 'timing/data': 0.0006160980556160212, 'timing/model': 1.6840627109631896, 'learning_rate/qwen_vl_interface': 1.2119402087907448e-06, 'learning_rate/action_model': 1.2119402087907448e-05, 'epoch': 1.6}) +[2026-06-23 22:36:53,874][starVLA.training.train_starvla][INFO] - Step 4136, Loss: {'train/loss': 0.10971268266439438, 'train/grad_norm_pre_clip': 16.21051025390625, 'timing/data': 0.0005848899018019438, 'timing/model': 1.6915043690241873, 'learning_rate/qwen_vl_interface': 1.2103373428141903e-06, 'learning_rate/action_model': 1.2103373428141903e-05, 'epoch': 1.6}) +[2026-06-23 22:36:57,193][starVLA.training.train_starvla][INFO] - Step 4137, Loss: {'train/loss': 0.06206325441598892, 'train/grad_norm_pre_clip': 34.910614013671875, 'timing/data': 0.0005802011583000422, 'timing/model': 1.683000260964036, 'learning_rate/qwen_vl_interface': 1.2087361373903633e-06, 'learning_rate/action_model': 1.2087361373903633e-05, 'epoch': 1.6}) +[2026-06-23 22:37:00,519][starVLA.training.train_starvla][INFO] - Step 4138, Loss: {'train/loss': 0.07218501716852188, 'train/grad_norm_pre_clip': 17.927621841430664, 'timing/data': 0.0005729368422180414, 'timing/model': 1.6898349770344794, 'learning_rate/qwen_vl_interface': 1.2071365931774584e-06, 'learning_rate/action_model': 1.2071365931774584e-05, 'epoch': 1.6}) +[2026-06-23 22:37:03,840][starVLA.training.train_starvla][INFO] - Step 4139, Loss: {'train/loss': 0.21055291593074799, 'train/grad_norm_pre_clip': 23.053024291992188, 'timing/data': 0.000553821912035346, 'timing/model': 1.6862328879069537, 'learning_rate/qwen_vl_interface': 1.2055387108329885e-06, 'learning_rate/action_model': 1.2055387108329886e-05, 'epoch': 1.6}) +[2026-06-23 22:37:07,177][starVLA.training.train_starvla][INFO] - Step 4140, Loss: {'train/loss': 0.11146272718906403, 'train/grad_norm_pre_clip': 14.4325590133667, 'timing/data': 0.0005528000183403492, 'timing/model': 1.6860765679739416, 'learning_rate/qwen_vl_interface': 1.2039424910137807e-06, 'learning_rate/action_model': 1.2039424910137806e-05, 'epoch': 1.6}) +[2026-06-23 22:37:10,504][starVLA.training.train_starvla][INFO] - Step 4141, Loss: {'train/loss': 0.04355432465672493, 'train/grad_norm_pre_clip': 42.81526184082031, 'timing/data': 0.000626969151198864, 'timing/model': 1.691792911849916, 'learning_rate/qwen_vl_interface': 1.2023479343759822e-06, 'learning_rate/action_model': 1.2023479343759823e-05, 'epoch': 1.6}) +[2026-06-23 22:37:13,831][starVLA.training.train_starvla][INFO] - Step 4142, Loss: {'train/loss': 0.12970998883247375, 'train/grad_norm_pre_clip': 22.044939041137695, 'timing/data': 0.0005607740022242069, 'timing/model': 1.688674241071567, 'learning_rate/qwen_vl_interface': 1.2007550415750546e-06, 'learning_rate/action_model': 1.2007550415750545e-05, 'epoch': 1.6}) +[2026-06-23 22:37:17,149][starVLA.training.train_starvla][INFO] - Step 4143, Loss: {'train/loss': 0.16533473134040833, 'train/grad_norm_pre_clip': 30.949159622192383, 'timing/data': 0.0006366770248860121, 'timing/model': 1.6854510640259832, 'learning_rate/qwen_vl_interface': 1.1991638132657759e-06, 'learning_rate/action_model': 1.199163813265776e-05, 'epoch': 1.6}) +[2026-06-23 22:37:20,471][starVLA.training.train_starvla][INFO] - Step 4144, Loss: {'train/loss': 0.07836762815713882, 'train/grad_norm_pre_clip': 20.50981330871582, 'timing/data': 0.0005622070748358965, 'timing/model': 1.6875374221708626, 'learning_rate/qwen_vl_interface': 1.1975742501022406e-06, 'learning_rate/action_model': 1.1975742501022405e-05, 'epoch': 1.6}) +[2026-06-23 22:37:23,792][starVLA.training.train_starvla][INFO] - Step 4145, Loss: {'train/loss': 0.1303604394197464, 'train/grad_norm_pre_clip': 19.201614379882812, 'timing/data': 0.0005852810572832823, 'timing/model': 1.6875160459894687, 'learning_rate/qwen_vl_interface': 1.1959863527378569e-06, 'learning_rate/action_model': 1.1959863527378567e-05, 'epoch': 1.6}) +[2026-06-23 22:37:27,205][starVLA.training.train_starvla][INFO] - Step 4146, Loss: {'train/loss': 0.05979965254664421, 'train/grad_norm_pre_clip': 31.23829460144043, 'timing/data': 0.0005826258566230536, 'timing/model': 1.6842162248212844, 'learning_rate/qwen_vl_interface': 1.1944001218253505e-06, 'learning_rate/action_model': 1.1944001218253503e-05, 'epoch': 1.6}) +[2026-06-23 22:37:30,526][starVLA.training.train_starvla][INFO] - Step 4147, Loss: {'train/loss': 0.2268134355545044, 'train/grad_norm_pre_clip': 35.12753677368164, 'timing/data': 0.0006587491370737553, 'timing/model': 1.689140269998461, 'learning_rate/qwen_vl_interface': 1.1928155580167608e-06, 'learning_rate/action_model': 1.1928155580167608e-05, 'epoch': 1.6}) +[2026-06-23 22:37:33,846][starVLA.training.train_starvla][INFO] - Step 4148, Loss: {'train/loss': 0.12560847401618958, 'train/grad_norm_pre_clip': 46.46011734008789, 'timing/data': 0.0005508861504495144, 'timing/model': 1.6870774650014937, 'learning_rate/qwen_vl_interface': 1.1912326619634412e-06, 'learning_rate/action_model': 1.1912326619634412e-05, 'epoch': 1.61}) +[2026-06-23 22:37:37,166][starVLA.training.train_starvla][INFO] - Step 4149, Loss: {'train/loss': 0.19642981886863708, 'train/grad_norm_pre_clip': 38.93024826049805, 'timing/data': 0.000550334807485342, 'timing/model': 1.6859828608576208, 'learning_rate/qwen_vl_interface': 1.1896514343160632e-06, 'learning_rate/action_model': 1.189651434316063e-05, 'epoch': 1.61}) +[2026-06-23 22:37:40,483][starVLA.training.train_starvla][INFO] - Step 4150, Loss: {'train/loss': 0.20600226521492004, 'train/grad_norm_pre_clip': 37.63136672973633, 'timing/data': 0.0005666459910571575, 'timing/model': 1.6844787860754877, 'learning_rate/qwen_vl_interface': 1.1880718757246062e-06, 'learning_rate/action_model': 1.1880718757246063e-05, 'epoch': 1.61}) +[2026-06-23 22:37:43,807][starVLA.training.train_starvla][INFO] - Step 4151, Loss: {'train/loss': 0.06347949802875519, 'train/grad_norm_pre_clip': 21.90552520751953, 'timing/data': 0.0005584000609815121, 'timing/model': 1.6909669321030378, 'learning_rate/qwen_vl_interface': 1.186493986838369e-06, 'learning_rate/action_model': 1.186493986838369e-05, 'epoch': 1.61}) +[2026-06-23 22:37:47,127][starVLA.training.train_starvla][INFO] - Step 4152, Loss: {'train/loss': 0.14173835515975952, 'train/grad_norm_pre_clip': 18.788728713989258, 'timing/data': 0.0006045869085937738, 'timing/model': 1.6854568049311638, 'learning_rate/qwen_vl_interface': 1.1849177683059612e-06, 'learning_rate/action_model': 1.1849177683059612e-05, 'epoch': 1.61}) +[2026-06-23 22:37:50,448][starVLA.training.train_starvla][INFO] - Step 4153, Loss: {'train/loss': 0.15141308307647705, 'train/grad_norm_pre_clip': 18.338607788085938, 'timing/data': 0.0006013510283082724, 'timing/model': 1.6858798449393362, 'learning_rate/qwen_vl_interface': 1.1833432207753075e-06, 'learning_rate/action_model': 1.1833432207753076e-05, 'epoch': 1.61}) +[2026-06-23 22:37:53,766][starVLA.training.train_starvla][INFO] - Step 4154, Loss: {'train/loss': 0.09853474795818329, 'train/grad_norm_pre_clip': 12.685240745544434, 'timing/data': 0.0006229011341929436, 'timing/model': 1.685474868165329, 'learning_rate/qwen_vl_interface': 1.1817703448936448e-06, 'learning_rate/action_model': 1.181770344893645e-05, 'epoch': 1.61}) +[2026-06-23 22:37:57,089][starVLA.training.train_starvla][INFO] - Step 4155, Loss: {'train/loss': 0.16861329972743988, 'train/grad_norm_pre_clip': 20.874156951904297, 'timing/data': 0.0006340020336210728, 'timing/model': 1.6838116960134357, 'learning_rate/qwen_vl_interface': 1.1801991413075228e-06, 'learning_rate/action_model': 1.1801991413075227e-05, 'epoch': 1.61}) +[2026-06-23 22:38:00,407][starVLA.training.train_starvla][INFO] - Step 4156, Loss: {'train/loss': 0.13866126537322998, 'train/grad_norm_pre_clip': 19.759750366210938, 'timing/data': 0.0006437310948967934, 'timing/model': 1.686870361911133, 'learning_rate/qwen_vl_interface': 1.1786296106628038e-06, 'learning_rate/action_model': 1.1786296106628036e-05, 'epoch': 1.61}) +[2026-06-23 22:38:03,727][starVLA.training.train_starvla][INFO] - Step 4157, Loss: {'train/loss': 0.055502284318208694, 'train/grad_norm_pre_clip': 11.809212684631348, 'timing/data': 0.0007211959455162287, 'timing/model': 1.687547761015594, 'learning_rate/qwen_vl_interface': 1.1770617536046633e-06, 'learning_rate/action_model': 1.1770617536046632e-05, 'epoch': 1.61}) +[2026-06-23 22:38:07,047][starVLA.training.train_starvla][INFO] - Step 4158, Loss: {'train/loss': 0.06542173773050308, 'train/grad_norm_pre_clip': 29.348913192749023, 'timing/data': 0.0005701319314539433, 'timing/model': 1.6862624979112297, 'learning_rate/qwen_vl_interface': 1.1754955707775873e-06, 'learning_rate/action_model': 1.1754955707775872e-05, 'epoch': 1.61}) +[2026-06-23 22:38:10,369][starVLA.training.train_starvla][INFO] - Step 4159, Loss: {'train/loss': 0.13988831639289856, 'train/grad_norm_pre_clip': 59.065391540527344, 'timing/data': 0.0006652111187577248, 'timing/model': 1.6878950230311602, 'learning_rate/qwen_vl_interface': 1.1739310628253753e-06, 'learning_rate/action_model': 1.1739310628253753e-05, 'epoch': 1.61}) +[2026-06-23 22:38:13,792][starVLA.training.train_starvla][INFO] - Step 4160, Loss: {'train/loss': 0.11615344882011414, 'train/grad_norm_pre_clip': 19.737730026245117, 'timing/data': 0.0005861220415681601, 'timing/model': 1.7862265999428928, 'learning_rate/qwen_vl_interface': 1.172368230391136e-06, 'learning_rate/action_model': 1.1723682303911359e-05, 'epoch': 1.61}) +[2026-06-23 22:38:17,112][starVLA.training.train_starvla][INFO] - Step 4161, Loss: {'train/loss': 0.08644036948680878, 'train/grad_norm_pre_clip': 38.4942512512207, 'timing/data': 0.0005624569021165371, 'timing/model': 1.6853285259567201, 'learning_rate/qwen_vl_interface': 1.1708070741172928e-06, 'learning_rate/action_model': 1.1708070741172928e-05, 'epoch': 1.61}) +[2026-06-23 22:38:20,435][starVLA.training.train_starvla][INFO] - Step 4162, Loss: {'train/loss': 0.13603544235229492, 'train/grad_norm_pre_clip': 27.971712112426758, 'timing/data': 0.000570683041587472, 'timing/model': 1.6879638568498194, 'learning_rate/qwen_vl_interface': 1.1692475946455755e-06, 'learning_rate/action_model': 1.1692475946455755e-05, 'epoch': 1.61}) +[2026-06-23 22:38:23,756][starVLA.training.train_starvla][INFO] - Step 4163, Loss: {'train/loss': 0.1293618530035019, 'train/grad_norm_pre_clip': 24.472129821777344, 'timing/data': 0.0006457439158111811, 'timing/model': 1.6865693100262433, 'learning_rate/qwen_vl_interface': 1.1676897926170289e-06, 'learning_rate/action_model': 1.1676897926170289e-05, 'epoch': 1.61}) +[2026-06-23 22:38:27,073][starVLA.training.train_starvla][INFO] - Step 4164, Loss: {'train/loss': 0.13038699328899384, 'train/grad_norm_pre_clip': 33.0879020690918, 'timing/data': 0.0005691200494766235, 'timing/model': 1.6824021991342306, 'learning_rate/qwen_vl_interface': 1.166133668672006e-06, 'learning_rate/action_model': 1.166133668672006e-05, 'epoch': 1.61}) +[2026-06-23 22:38:30,397][starVLA.training.train_starvla][INFO] - Step 4165, Loss: {'train/loss': 0.1798247992992401, 'train/grad_norm_pre_clip': 16.66342544555664, 'timing/data': 0.0006180929485708475, 'timing/model': 1.6881166589446366, 'learning_rate/qwen_vl_interface': 1.1645792234501706e-06, 'learning_rate/action_model': 1.1645792234501706e-05, 'epoch': 1.61}) +[2026-06-23 22:38:33,717][starVLA.training.train_starvla][INFO] - Step 4166, Loss: {'train/loss': 0.22735758125782013, 'train/grad_norm_pre_clip': 29.550495147705078, 'timing/data': 0.0005746008828282356, 'timing/model': 1.6856605450157076, 'learning_rate/qwen_vl_interface': 1.1630264575904972e-06, 'learning_rate/action_model': 1.1630264575904973e-05, 'epoch': 1.61}) +[2026-06-23 22:38:37,035][starVLA.training.train_starvla][INFO] - Step 4167, Loss: {'train/loss': 0.08507044613361359, 'train/grad_norm_pre_clip': 47.49848175048828, 'timing/data': 0.0007416338194161654, 'timing/model': 1.6836296608671546, 'learning_rate/qwen_vl_interface': 1.1614753717312683e-06, 'learning_rate/action_model': 1.1614753717312681e-05, 'epoch': 1.61}) +[2026-06-23 22:38:40,360][starVLA.training.train_starvla][INFO] - Step 4168, Loss: {'train/loss': 0.1457405835390091, 'train/grad_norm_pre_clip': 33.66580581665039, 'timing/data': 0.0005713941063731909, 'timing/model': 1.6914345838595182, 'learning_rate/qwen_vl_interface': 1.1599259665100767e-06, 'learning_rate/action_model': 1.1599259665100767e-05, 'epoch': 1.61}) +[2026-06-23 22:38:43,683][starVLA.training.train_starvla][INFO] - Step 4169, Loss: {'train/loss': 0.09233785420656204, 'train/grad_norm_pre_clip': 17.151195526123047, 'timing/data': 0.000636586919426918, 'timing/model': 1.6875313590280712, 'learning_rate/qwen_vl_interface': 1.1583782425638252e-06, 'learning_rate/action_model': 1.158378242563825e-05, 'epoch': 1.61}) +[2026-06-23 22:38:47,005][starVLA.training.train_starvla][INFO] - Step 4170, Loss: {'train/loss': 0.08593853563070297, 'train/grad_norm_pre_clip': 23.134033203125, 'timing/data': 0.0006592301651835442, 'timing/model': 1.6869453319814056, 'learning_rate/qwen_vl_interface': 1.1568322005287237e-06, 'learning_rate/action_model': 1.1568322005287235e-05, 'epoch': 1.61}) +[2026-06-23 22:38:50,329][starVLA.training.train_starvla][INFO] - Step 4171, Loss: {'train/loss': 0.12198473513126373, 'train/grad_norm_pre_clip': 23.645709991455078, 'timing/data': 0.0005767450202256441, 'timing/model': 1.6898966170847416, 'learning_rate/qwen_vl_interface': 1.1552878410402923e-06, 'learning_rate/action_model': 1.1552878410402924e-05, 'epoch': 1.61}) +[2026-06-23 22:38:53,651][starVLA.training.train_starvla][INFO] - Step 4172, Loss: {'train/loss': 0.10921023041009903, 'train/grad_norm_pre_clip': 18.606430053710938, 'timing/data': 0.0005683989729732275, 'timing/model': 1.6870581340044737, 'learning_rate/qwen_vl_interface': 1.1537451647333594e-06, 'learning_rate/action_model': 1.1537451647333593e-05, 'epoch': 1.61}) +[2026-06-23 22:38:56,971][starVLA.training.train_starvla][INFO] - Step 4173, Loss: {'train/loss': 0.0952303409576416, 'train/grad_norm_pre_clip': 28.504568099975586, 'timing/data': 0.0005919928662478924, 'timing/model': 1.6848617899231613, 'learning_rate/qwen_vl_interface': 1.1522041722420594e-06, 'learning_rate/action_model': 1.1522041722420594e-05, 'epoch': 1.62}) +[2026-06-23 22:39:00,290][starVLA.training.train_starvla][INFO] - Step 4174, Loss: {'train/loss': 0.2878496050834656, 'train/grad_norm_pre_clip': 25.490158081054688, 'timing/data': 0.0005704329814761877, 'timing/model': 1.6847893288359046, 'learning_rate/qwen_vl_interface': 1.1506648641998365e-06, 'learning_rate/action_model': 1.1506648641998365e-05, 'epoch': 1.62}) +[2026-06-23 22:39:03,709][starVLA.training.train_starvla][INFO] - Step 4175, Loss: {'train/loss': 0.06227400526404381, 'train/grad_norm_pre_clip': 45.81285858154297, 'timing/data': 0.0005861921235918999, 'timing/model': 1.6872674440965056, 'learning_rate/qwen_vl_interface': 1.1491272412394433e-06, 'learning_rate/action_model': 1.1491272412394433e-05, 'epoch': 1.62}) +[2026-06-23 22:39:07,031][starVLA.training.train_starvla][INFO] - Step 4176, Loss: {'train/loss': 0.2084556370973587, 'train/grad_norm_pre_clip': 23.744657516479492, 'timing/data': 0.0005989661440253258, 'timing/model': 1.6886132201179862, 'learning_rate/qwen_vl_interface': 1.1475913039929373e-06, 'learning_rate/action_model': 1.1475913039929372e-05, 'epoch': 1.62}) +[2026-06-23 22:39:10,351][starVLA.training.train_starvla][INFO] - Step 4177, Loss: {'train/loss': 0.07820942997932434, 'train/grad_norm_pre_clip': 50.59053039550781, 'timing/data': 0.000598184997215867, 'timing/model': 1.6874727208632976, 'learning_rate/qwen_vl_interface': 1.1460570530916844e-06, 'learning_rate/action_model': 1.1460570530916844e-05, 'epoch': 1.62}) +[2026-06-23 22:39:13,675][starVLA.training.train_starvla][INFO] - Step 4178, Loss: {'train/loss': 0.17950557172298431, 'train/grad_norm_pre_clip': 38.08330535888672, 'timing/data': 0.0006176619790494442, 'timing/model': 1.687349263811484, 'learning_rate/qwen_vl_interface': 1.1445244891663581e-06, 'learning_rate/action_model': 1.1445244891663581e-05, 'epoch': 1.62}) +[2026-06-23 22:39:16,995][starVLA.training.train_starvla][INFO] - Step 4179, Loss: {'train/loss': 0.2307431548833847, 'train/grad_norm_pre_clip': 28.678672790527344, 'timing/data': 0.0005621570162475109, 'timing/model': 1.6854722399730235, 'learning_rate/qwen_vl_interface': 1.1429936128469366e-06, 'learning_rate/action_model': 1.1429936128469366e-05, 'epoch': 1.62}) +[2026-06-23 22:39:20,317][starVLA.training.train_starvla][INFO] - Step 4180, Loss: {'train/loss': 0.25282737612724304, 'train/grad_norm_pre_clip': 75.72583770751953, 'timing/data': 0.0005526591558009386, 'timing/model': 1.685884951846674, 'learning_rate/qwen_vl_interface': 1.1414644247627056e-06, 'learning_rate/action_model': 1.1414644247627058e-05, 'epoch': 1.62}) +[2026-06-23 22:39:23,640][starVLA.training.train_starvla][INFO] - Step 4181, Loss: {'train/loss': 0.1429644376039505, 'train/grad_norm_pre_clip': 20.86893081665039, 'timing/data': 0.0007842648774385452, 'timing/model': 1.6930223479866982, 'learning_rate/qwen_vl_interface': 1.1399369255422568e-06, 'learning_rate/action_model': 1.1399369255422568e-05, 'epoch': 1.62}) +[2026-06-23 22:39:26,965][starVLA.training.train_starvla][INFO] - Step 4182, Loss: {'train/loss': 0.12270510196685791, 'train/grad_norm_pre_clip': 27.10480499267578, 'timing/data': 0.0005773759912699461, 'timing/model': 1.6898957749363035, 'learning_rate/qwen_vl_interface': 1.138411115813488e-06, 'learning_rate/action_model': 1.138411115813488e-05, 'epoch': 1.62}) +[2026-06-23 22:39:30,281][starVLA.training.train_starvla][INFO] - Step 4183, Loss: {'train/loss': 0.06638418883085251, 'train/grad_norm_pre_clip': 9.990524291992188, 'timing/data': 0.000596341909840703, 'timing/model': 1.6855448591522872, 'learning_rate/qwen_vl_interface': 1.1368869962036006e-06, 'learning_rate/action_model': 1.1368869962036005e-05, 'epoch': 1.62}) +[2026-06-23 22:39:33,608][starVLA.training.train_starvla][INFO] - Step 4184, Loss: {'train/loss': 0.14782196283340454, 'train/grad_norm_pre_clip': 32.447086334228516, 'timing/data': 0.0005611060187220573, 'timing/model': 1.6878775099758059, 'learning_rate/qwen_vl_interface': 1.1353645673391025e-06, 'learning_rate/action_model': 1.1353645673391026e-05, 'epoch': 1.62}) +[2026-06-23 22:39:36,929][starVLA.training.train_starvla][INFO] - Step 4185, Loss: {'train/loss': 0.11169708520174026, 'train/grad_norm_pre_clip': 19.545726776123047, 'timing/data': 0.0006775639485567808, 'timing/model': 1.6867260008584708, 'learning_rate/qwen_vl_interface': 1.1338438298458078e-06, 'learning_rate/action_model': 1.1338438298458078e-05, 'epoch': 1.62}) +[2026-06-23 22:39:40,253][starVLA.training.train_starvla][INFO] - Step 4186, Loss: {'train/loss': 0.043266769498586655, 'train/grad_norm_pre_clip': 37.30823516845703, 'timing/data': 0.0005685391370207071, 'timing/model': 1.6884883800521493, 'learning_rate/qwen_vl_interface': 1.1323247843488345e-06, 'learning_rate/action_model': 1.1323247843488344e-05, 'epoch': 1.62}) +[2026-06-23 22:39:43,572][starVLA.training.train_starvla][INFO] - Step 4187, Loss: {'train/loss': 0.1201123595237732, 'train/grad_norm_pre_clip': 26.267271041870117, 'timing/data': 0.0005533909425139427, 'timing/model': 1.6838646021205932, 'learning_rate/qwen_vl_interface': 1.130807431472603e-06, 'learning_rate/action_model': 1.130807431472603e-05, 'epoch': 1.62}) +[2026-06-23 22:39:46,889][starVLA.training.train_starvla][INFO] - Step 4188, Loss: {'train/loss': 0.09001168608665466, 'train/grad_norm_pre_clip': 17.37085723876953, 'timing/data': 0.000549752963706851, 'timing/model': 1.681135836057365, 'learning_rate/qwen_vl_interface': 1.1292917718408416e-06, 'learning_rate/action_model': 1.1292917718408416e-05, 'epoch': 1.62}) +[2026-06-23 22:39:50,208][starVLA.training.train_starvla][INFO] - Step 4189, Loss: {'train/loss': 0.0925075113773346, 'train/grad_norm_pre_clip': 20.977828979492188, 'timing/data': 0.0005464069545269012, 'timing/model': 1.6876173799391836, 'learning_rate/qwen_vl_interface': 1.12777780607658e-06, 'learning_rate/action_model': 1.12777780607658e-05, 'epoch': 1.62}) +[2026-06-23 22:39:53,680][starVLA.training.train_starvla][INFO] - Step 4190, Loss: {'train/loss': 0.061788253486156464, 'train/grad_norm_pre_clip': 10.916698455810547, 'timing/data': 0.0005821739323437214, 'timing/model': 1.6861218840349466, 'learning_rate/qwen_vl_interface': 1.1262655348021523e-06, 'learning_rate/action_model': 1.1262655348021523e-05, 'epoch': 1.62}) +[2026-06-23 22:39:57,001][starVLA.training.train_starvla][INFO] - Step 4191, Loss: {'train/loss': 0.04971115663647652, 'train/grad_norm_pre_clip': 7.535142421722412, 'timing/data': 0.000578348059207201, 'timing/model': 1.6865528889466077, 'learning_rate/qwen_vl_interface': 1.1247549586391966e-06, 'learning_rate/action_model': 1.1247549586391964e-05, 'epoch': 1.62}) +[2026-06-23 22:40:00,319][starVLA.training.train_starvla][INFO] - Step 4192, Loss: {'train/loss': 0.1111568808555603, 'train/grad_norm_pre_clip': 21.60517692565918, 'timing/data': 0.0006218489725142717, 'timing/model': 1.684861675137654, 'learning_rate/qwen_vl_interface': 1.1232460782086534e-06, 'learning_rate/action_model': 1.1232460782086534e-05, 'epoch': 1.62}) +[2026-06-23 22:40:03,641][starVLA.training.train_starvla][INFO] - Step 4193, Loss: {'train/loss': 0.06088707968592644, 'train/grad_norm_pre_clip': 29.026546478271484, 'timing/data': 0.0006482088938355446, 'timing/model': 1.6887237641494721, 'learning_rate/qwen_vl_interface': 1.1217388941307658e-06, 'learning_rate/action_model': 1.1217388941307658e-05, 'epoch': 1.62}) +[2026-06-23 22:40:06,968][starVLA.training.train_starvla][INFO] - Step 4194, Loss: {'train/loss': 0.16364194452762604, 'train/grad_norm_pre_clip': 16.300260543823242, 'timing/data': 0.0006569058168679476, 'timing/model': 1.6903047522064298, 'learning_rate/qwen_vl_interface': 1.1202334070250807e-06, 'learning_rate/action_model': 1.1202334070250806e-05, 'epoch': 1.62}) +[2026-06-23 22:40:10,294][starVLA.training.train_starvla][INFO] - Step 4195, Loss: {'train/loss': 0.13565143942832947, 'train/grad_norm_pre_clip': 31.775720596313477, 'timing/data': 0.0005657239817082882, 'timing/model': 1.6882355369161814, 'learning_rate/qwen_vl_interface': 1.1187296175104472e-06, 'learning_rate/action_model': 1.1187296175104471e-05, 'epoch': 1.62}) +[2026-06-23 22:40:13,641][starVLA.training.train_starvla][INFO] - Step 4196, Loss: {'train/loss': 0.11691436916589737, 'train/grad_norm_pre_clip': 30.673961639404297, 'timing/data': 0.0005555250681936741, 'timing/model': 1.7101682350039482, 'learning_rate/qwen_vl_interface': 1.1172275262050158e-06, 'learning_rate/action_model': 1.1172275262050159e-05, 'epoch': 1.62}) +[2026-06-23 22:40:16,962][starVLA.training.train_starvla][INFO] - Step 4197, Loss: {'train/loss': 0.13304664194583893, 'train/grad_norm_pre_clip': 22.737646102905273, 'timing/data': 0.0005597020499408245, 'timing/model': 1.6859625750221312, 'learning_rate/qwen_vl_interface': 1.1157271337262401e-06, 'learning_rate/action_model': 1.11572713372624e-05, 'epoch': 1.62}) +[2026-06-23 22:40:20,284][starVLA.training.train_starvla][INFO] - Step 4198, Loss: {'train/loss': 0.19858185946941376, 'train/grad_norm_pre_clip': 42.337852478027344, 'timing/data': 0.000558098778128624, 'timing/model': 1.6887422378640622, 'learning_rate/qwen_vl_interface': 1.1142284406908751e-06, 'learning_rate/action_model': 1.114228440690875e-05, 'epoch': 1.62}) +[2026-06-23 22:40:23,607][starVLA.training.train_starvla][INFO] - Step 4199, Loss: {'train/loss': 0.03265714272856712, 'train/grad_norm_pre_clip': 25.96323585510254, 'timing/data': 0.0005701431073248386, 'timing/model': 1.68828791892156, 'learning_rate/qwen_vl_interface': 1.1127314477149758e-06, 'learning_rate/action_model': 1.1127314477149759e-05, 'epoch': 1.63}) +[2026-06-23 22:40:26,930][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4200 +[2026-06-23 22:41:58,570][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4200 +[2026-06-23 22:41:58,570][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4200 +[2026-06-23 22:43:09,733][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4200 +[2026-06-23 22:43:09,734][starVLA.training.train_starvla][INFO] - Step 4200, Loss: {'train/loss': 0.07642439752817154, 'train/grad_norm_pre_clip': 10.987058639526367, 'eval/action_loss/seconds': 91.6380731430836, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06564455479383469, 'eval/latency_2/loss': 0.06564455479383469, 'eval/flappy/loss': 0.06564455479383469, 'eval/flappy/latency_2/loss': 0.06564455479383469, 'eval/flap_precision': 0.9435757987763427, 'eval/flap_recall': 0.9365721997300944, 'eval/flap_f1': 0.9400609549610565, 'eval/flap_rate_model': 0.20430555555555555, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9400609549610565, 'eval/latency_2/flap_precision': 0.9435757987763427, 'eval/latency_2/flap_recall': 0.9365721997300944, 'eval/latency_2/flap_f1': 0.9400609549610565, 'eval/latency_2/flap_rate_model': 0.20430555555555555, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9400609549610565, 'eval/action_classification/data_wait_seconds': 0.8552919030189514, 'eval/action_classification/predict_seconds': 70.17616271972656, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.000578777864575386, 'timing/model': 1.6885044530499727, 'learning_rate/qwen_vl_interface': 1.1112361554139003e-06, 'learning_rate/action_model': 1.1112361554139004e-05, 'epoch': 1.63}) +[2026-06-23 22:43:13,046][starVLA.training.train_starvla][INFO] - Step 4201, Loss: {'train/loss': 0.11102999001741409, 'train/grad_norm_pre_clip': 50.79072189331055, 'timing/data': 0.0005522090941667557, 'timing/model': 1.684083711123094, 'learning_rate/qwen_vl_interface': 1.1097425644023065e-06, 'learning_rate/action_model': 1.1097425644023065e-05, 'epoch': 1.63}) +[2026-06-23 22:43:16,355][starVLA.training.train_starvla][INFO] - Step 4202, Loss: {'train/loss': 0.09292104840278625, 'train/grad_norm_pre_clip': 21.495866775512695, 'timing/data': 0.0005501441191881895, 'timing/model': 1.6803531129844487, 'learning_rate/qwen_vl_interface': 1.1082506752941529e-06, 'learning_rate/action_model': 1.1082506752941528e-05, 'epoch': 1.63}) +[2026-06-23 22:43:19,664][starVLA.training.train_starvla][INFO] - Step 4203, Loss: {'train/loss': 0.21004290878772736, 'train/grad_norm_pre_clip': 26.201763153076172, 'timing/data': 0.0005461669061332941, 'timing/model': 1.6802594151813537, 'learning_rate/qwen_vl_interface': 1.1067604887027004e-06, 'learning_rate/action_model': 1.1067604887027004e-05, 'epoch': 1.63}) +[2026-06-23 22:43:22,982][starVLA.training.train_starvla][INFO] - Step 4204, Loss: {'train/loss': 0.09045270085334778, 'train/grad_norm_pre_clip': 18.204647064208984, 'timing/data': 0.0005692301783710718, 'timing/model': 1.6823448231443763, 'learning_rate/qwen_vl_interface': 1.1052720052405066e-06, 'learning_rate/action_model': 1.1052720052405064e-05, 'epoch': 1.63}) +[2026-06-23 22:43:26,293][starVLA.training.train_starvla][INFO] - Step 4205, Loss: {'train/loss': 0.1262197494506836, 'train/grad_norm_pre_clip': 56.49967956542969, 'timing/data': 0.0006053689867258072, 'timing/model': 1.6813033521175385, 'learning_rate/qwen_vl_interface': 1.1037852255194309e-06, 'learning_rate/action_model': 1.1037852255194308e-05, 'epoch': 1.63}) +[2026-06-23 22:43:29,606][starVLA.training.train_starvla][INFO] - Step 4206, Loss: {'train/loss': 0.056853167712688446, 'train/grad_norm_pre_clip': 15.084524154663086, 'timing/data': 0.0005419699009507895, 'timing/model': 1.6823024780023843, 'learning_rate/qwen_vl_interface': 1.1023001501506329e-06, 'learning_rate/action_model': 1.102300150150633e-05, 'epoch': 1.63}) +[2026-06-23 22:43:32,953][starVLA.training.train_starvla][INFO] - Step 4207, Loss: {'train/loss': 0.133614644408226, 'train/grad_norm_pre_clip': 21.292936325073242, 'timing/data': 0.0006101669277995825, 'timing/model': 1.6847569500096142, 'learning_rate/qwen_vl_interface': 1.1008167797445717e-06, 'learning_rate/action_model': 1.1008167797445718e-05, 'epoch': 1.63}) +[2026-06-23 22:43:36,269][starVLA.training.train_starvla][INFO] - Step 4208, Loss: {'train/loss': 0.12825538218021393, 'train/grad_norm_pre_clip': 17.54201889038086, 'timing/data': 0.0005838179495185614, 'timing/model': 1.683483012020588, 'learning_rate/qwen_vl_interface': 1.099335114911003e-06, 'learning_rate/action_model': 1.0993351149110031e-05, 'epoch': 1.63}) +[2026-06-23 22:43:39,584][starVLA.training.train_starvla][INFO] - Step 4209, Loss: {'train/loss': 0.05876791477203369, 'train/grad_norm_pre_clip': 12.218022346496582, 'timing/data': 0.0006386509630829096, 'timing/model': 1.6855135881341994, 'learning_rate/qwen_vl_interface': 1.0978551562589846e-06, 'learning_rate/action_model': 1.0978551562589845e-05, 'epoch': 1.63}) +[2026-06-23 22:43:42,894][starVLA.training.train_starvla][INFO] - Step 4210, Loss: {'train/loss': 0.09275849163532257, 'train/grad_norm_pre_clip': 16.48904800415039, 'timing/data': 0.000577996950596571, 'timing/model': 1.6787460900377482, 'learning_rate/qwen_vl_interface': 1.0963769043968708e-06, 'learning_rate/action_model': 1.0963769043968708e-05, 'epoch': 1.63}) +[2026-06-23 22:43:46,211][starVLA.training.train_starvla][INFO] - Step 4211, Loss: {'train/loss': 0.07361243665218353, 'train/grad_norm_pre_clip': 16.56304359436035, 'timing/data': 0.0006273097824305296, 'timing/model': 1.6837638020515442, 'learning_rate/qwen_vl_interface': 1.0949003599323158e-06, 'learning_rate/action_model': 1.0949003599323157e-05, 'epoch': 1.63}) +[2026-06-23 22:43:49,820][starVLA.training.train_starvla][INFO] - Step 4212, Loss: {'train/loss': 0.09350790083408356, 'train/grad_norm_pre_clip': 35.84054183959961, 'timing/data': 0.0006421469151973724, 'timing/model': 1.9824511739425361, 'learning_rate/qwen_vl_interface': 1.0934255234722708e-06, 'learning_rate/action_model': 1.0934255234722706e-05, 'epoch': 1.63}) +[2026-06-23 22:43:53,134][starVLA.training.train_starvla][INFO] - Step 4213, Loss: {'train/loss': 0.08165863156318665, 'train/grad_norm_pre_clip': 38.06764221191406, 'timing/data': 0.0005663849879056215, 'timing/model': 1.6837016998324543, 'learning_rate/qwen_vl_interface': 1.0919523956229858e-06, 'learning_rate/action_model': 1.0919523956229857e-05, 'epoch': 1.63}) +[2026-06-23 22:43:56,433][starVLA.training.train_starvla][INFO] - Step 4214, Loss: {'train/loss': 0.03499430790543556, 'train/grad_norm_pre_clip': 28.54324722290039, 'timing/data': 0.0006159578915685415, 'timing/model': 1.6758274019230157, 'learning_rate/qwen_vl_interface': 1.0904809769900075e-06, 'learning_rate/action_model': 1.0904809769900073e-05, 'epoch': 1.63}) +[2026-06-23 22:43:59,753][starVLA.training.train_starvla][INFO] - Step 4215, Loss: {'train/loss': 0.07534424215555191, 'train/grad_norm_pre_clip': 19.17618179321289, 'timing/data': 0.0005658939480781555, 'timing/model': 1.6866368539631367, 'learning_rate/qwen_vl_interface': 1.0890112681781821e-06, 'learning_rate/action_model': 1.089011268178182e-05, 'epoch': 1.63}) +[2026-06-23 22:44:03,065][starVLA.training.train_starvla][INFO] - Step 4216, Loss: {'train/loss': 0.07032959163188934, 'train/grad_norm_pre_clip': 19.53627586364746, 'timing/data': 0.0005771850701421499, 'timing/model': 1.6848776331171393, 'learning_rate/qwen_vl_interface': 1.0875432697916492e-06, 'learning_rate/action_model': 1.087543269791649e-05, 'epoch': 1.63}) +[2026-06-23 22:44:06,382][starVLA.training.train_starvla][INFO] - Step 4217, Loss: {'train/loss': 0.23464235663414001, 'train/grad_norm_pre_clip': 28.0050106048584, 'timing/data': 0.0005748309195041656, 'timing/model': 1.6838654670864344, 'learning_rate/qwen_vl_interface': 1.0860769824338477e-06, 'learning_rate/action_model': 1.0860769824338478e-05, 'epoch': 1.63}) +[2026-06-23 22:44:09,687][starVLA.training.train_starvla][INFO] - Step 4218, Loss: {'train/loss': 0.0948759987950325, 'train/grad_norm_pre_clip': 16.673912048339844, 'timing/data': 0.0006558529566973448, 'timing/model': 1.6781396688893437, 'learning_rate/qwen_vl_interface': 1.084612406707514e-06, 'learning_rate/action_model': 1.084612406707514e-05, 'epoch': 1.63}) +[2026-06-23 22:44:13,004][starVLA.training.train_starvla][INFO] - Step 4219, Loss: {'train/loss': 0.14661122858524323, 'train/grad_norm_pre_clip': 23.120378494262695, 'timing/data': 0.0007591578178107738, 'timing/model': 1.681957625085488, 'learning_rate/qwen_vl_interface': 1.0831495432146795e-06, 'learning_rate/action_model': 1.0831495432146795e-05, 'epoch': 1.63}) +[2026-06-23 22:44:16,317][starVLA.training.train_starvla][INFO] - Step 4220, Loss: {'train/loss': 0.1417856216430664, 'train/grad_norm_pre_clip': 41.28369140625, 'timing/data': 0.0006066909991204739, 'timing/model': 1.6841437630355358, 'learning_rate/qwen_vl_interface': 1.0816883925566717e-06, 'learning_rate/action_model': 1.0816883925566716e-05, 'epoch': 1.63}) +[2026-06-23 22:44:19,632][starVLA.training.train_starvla][INFO] - Step 4221, Loss: {'train/loss': 0.14584195613861084, 'train/grad_norm_pre_clip': 51.25965118408203, 'timing/data': 0.0006555020809173584, 'timing/model': 1.684490814106539, 'learning_rate/qwen_vl_interface': 1.0802289553341153e-06, 'learning_rate/action_model': 1.0802289553341152e-05, 'epoch': 1.63}) +[2026-06-23 22:44:22,945][starVLA.training.train_starvla][INFO] - Step 4222, Loss: {'train/loss': 0.1320684850215912, 'train/grad_norm_pre_clip': 16.632875442504883, 'timing/data': 0.0005721759516745806, 'timing/model': 1.6847242701333016, 'learning_rate/qwen_vl_interface': 1.0787712321469289e-06, 'learning_rate/action_model': 1.0787712321469288e-05, 'epoch': 1.63}) +[2026-06-23 22:44:26,262][starVLA.training.train_starvla][INFO] - Step 4223, Loss: {'train/loss': 0.07668603956699371, 'train/grad_norm_pre_clip': 51.631690979003906, 'timing/data': 0.0005919630639255047, 'timing/model': 1.6840203530155122, 'learning_rate/qwen_vl_interface': 1.077315223594328e-06, 'learning_rate/action_model': 1.077315223594328e-05, 'epoch': 1.63}) +[2026-06-23 22:44:29,582][starVLA.training.train_starvla][INFO] - Step 4224, Loss: {'train/loss': 0.21268531680107117, 'train/grad_norm_pre_clip': 27.525474548339844, 'timing/data': 0.0006067410577088594, 'timing/model': 1.6890230469871312, 'learning_rate/qwen_vl_interface': 1.0758609302748227e-06, 'learning_rate/action_model': 1.0758609302748227e-05, 'epoch': 1.63}) +[2026-06-23 22:44:32,901][starVLA.training.train_starvla][INFO] - Step 4225, Loss: {'train/loss': 0.06955087184906006, 'train/grad_norm_pre_clip': 17.58647918701172, 'timing/data': 0.0006557030137628317, 'timing/model': 1.6887730651069432, 'learning_rate/qwen_vl_interface': 1.0744083527862182e-06, 'learning_rate/action_model': 1.074408352786218e-05, 'epoch': 1.64}) +[2026-06-23 22:44:36,221][starVLA.training.train_starvla][INFO] - Step 4226, Loss: {'train/loss': 0.036818914115428925, 'train/grad_norm_pre_clip': 38.6522331237793, 'timing/data': 0.0006032739765942097, 'timing/model': 1.6868122259620577, 'learning_rate/qwen_vl_interface': 1.0729574917256144e-06, 'learning_rate/action_model': 1.0729574917256145e-05, 'epoch': 1.64}) +[2026-06-23 22:44:39,768][starVLA.training.train_starvla][INFO] - Step 4227, Loss: {'train/loss': 0.15360304713249207, 'train/grad_norm_pre_clip': 19.38189697265625, 'timing/data': 0.0006777949165552855, 'timing/model': 1.9109961660578847, 'learning_rate/qwen_vl_interface': 1.0715083476894047e-06, 'learning_rate/action_model': 1.0715083476894047e-05, 'epoch': 1.64}) +[2026-06-23 22:44:43,090][starVLA.training.train_starvla][INFO] - Step 4228, Loss: {'train/loss': 0.04974609613418579, 'train/grad_norm_pre_clip': 31.197025299072266, 'timing/data': 0.0005560349673032761, 'timing/model': 1.6866151001304388, 'learning_rate/qwen_vl_interface': 1.0700609212732773e-06, 'learning_rate/action_model': 1.0700609212732774e-05, 'epoch': 1.64}) +[2026-06-23 22:44:46,401][starVLA.training.train_starvla][INFO] - Step 4229, Loss: {'train/loss': 0.056973423808813095, 'train/grad_norm_pre_clip': 22.690860748291016, 'timing/data': 0.0005815329495817423, 'timing/model': 1.682474602945149, 'learning_rate/qwen_vl_interface': 1.0686152130722155e-06, 'learning_rate/action_model': 1.0686152130722154e-05, 'epoch': 1.64}) +[2026-06-23 22:44:49,717][starVLA.training.train_starvla][INFO] - Step 4230, Loss: {'train/loss': 0.08479654043912888, 'train/grad_norm_pre_clip': 34.46316909790039, 'timing/data': 0.0005759929772466421, 'timing/model': 1.6842210239265114, 'learning_rate/qwen_vl_interface': 1.0671712236804947e-06, 'learning_rate/action_model': 1.0671712236804946e-05, 'epoch': 1.64}) +[2026-06-23 22:44:53,028][starVLA.training.train_starvla][INFO] - Step 4231, Loss: {'train/loss': 0.09182213246822357, 'train/grad_norm_pre_clip': 34.96193313598633, 'timing/data': 0.0006491509266197681, 'timing/model': 1.6810114171821624, 'learning_rate/qwen_vl_interface': 1.0657289536916845e-06, 'learning_rate/action_model': 1.0657289536916845e-05, 'epoch': 1.64}) +[2026-06-23 22:44:56,352][starVLA.training.train_starvla][INFO] - Step 4232, Loss: {'train/loss': 0.06607043743133545, 'train/grad_norm_pre_clip': 26.164091110229492, 'timing/data': 0.0007588870357722044, 'timing/model': 1.691158734029159, 'learning_rate/qwen_vl_interface': 1.0642884036986481e-06, 'learning_rate/action_model': 1.064288403698648e-05, 'epoch': 1.64}) +[2026-06-23 22:44:59,665][starVLA.training.train_starvla][INFO] - Step 4233, Loss: {'train/loss': 0.2059807926416397, 'train/grad_norm_pre_clip': 47.065025329589844, 'timing/data': 0.0005414481274783611, 'timing/model': 1.6835496239364147, 'learning_rate/qwen_vl_interface': 1.0628495742935404e-06, 'learning_rate/action_model': 1.0628495742935405e-05, 'epoch': 1.64}) +[2026-06-23 22:45:02,992][starVLA.training.train_starvla][INFO] - Step 4234, Loss: {'train/loss': 0.1776653528213501, 'train/grad_norm_pre_clip': 29.14071273803711, 'timing/data': 0.0005753010045737028, 'timing/model': 1.6936573770362884, 'learning_rate/qwen_vl_interface': 1.0614124660678103e-06, 'learning_rate/action_model': 1.0614124660678102e-05, 'epoch': 1.64}) +[2026-06-23 22:45:06,312][starVLA.training.train_starvla][INFO] - Step 4235, Loss: {'train/loss': 0.07093124836683273, 'train/grad_norm_pre_clip': 19.973722457885742, 'timing/data': 0.0005729671102017164, 'timing/model': 1.691924547078088, 'learning_rate/qwen_vl_interface': 1.0599770796121988e-06, 'learning_rate/action_model': 1.0599770796121987e-05, 'epoch': 1.64}) +[2026-06-23 22:45:09,628][starVLA.training.train_starvla][INFO] - Step 4236, Loss: {'train/loss': 0.0825221836566925, 'train/grad_norm_pre_clip': 19.645233154296875, 'timing/data': 0.0007862979546189308, 'timing/model': 1.6846805550158024, 'learning_rate/qwen_vl_interface': 1.0585434155167394e-06, 'learning_rate/action_model': 1.0585434155167393e-05, 'epoch': 1.64}) +[2026-06-23 22:45:12,943][starVLA.training.train_starvla][INFO] - Step 4237, Loss: {'train/loss': 0.11298099160194397, 'train/grad_norm_pre_clip': 25.95359992980957, 'timing/data': 0.0006589088588953018, 'timing/model': 1.6831889068707824, 'learning_rate/qwen_vl_interface': 1.0571114743707556e-06, 'learning_rate/action_model': 1.0571114743707555e-05, 'epoch': 1.64}) +[2026-06-23 22:45:16,259][starVLA.training.train_starvla][INFO] - Step 4238, Loss: {'train/loss': 0.24818721413612366, 'train/grad_norm_pre_clip': 14.080126762390137, 'timing/data': 0.0005296559538692236, 'timing/model': 1.6841496611014009, 'learning_rate/qwen_vl_interface': 1.0556812567628661e-06, 'learning_rate/action_model': 1.055681256762866e-05, 'epoch': 1.64}) +[2026-06-23 22:45:19,571][starVLA.training.train_starvla][INFO] - Step 4239, Loss: {'train/loss': 0.05419347435235977, 'train/grad_norm_pre_clip': 19.487449645996094, 'timing/data': 0.0005406059790402651, 'timing/model': 1.680104612838477, 'learning_rate/qwen_vl_interface': 1.0542527632809783e-06, 'learning_rate/action_model': 1.0542527632809784e-05, 'epoch': 1.64}) +[2026-06-23 22:45:22,883][starVLA.training.train_starvla][INFO] - Step 4240, Loss: {'train/loss': 0.07976502180099487, 'train/grad_norm_pre_clip': 27.262720108032227, 'timing/data': 0.000540396198630333, 'timing/model': 1.6831072370987386, 'learning_rate/qwen_vl_interface': 1.0528259945122925e-06, 'learning_rate/action_model': 1.0528259945122925e-05, 'epoch': 1.64}) +[2026-06-23 22:45:26,349][starVLA.training.train_starvla][INFO] - Step 4241, Loss: {'train/loss': 0.11889121681451797, 'train/grad_norm_pre_clip': 20.445871353149414, 'timing/data': 0.0006307661533355713, 'timing/model': 1.8356358639430255, 'learning_rate/qwen_vl_interface': 1.0514009510432988e-06, 'learning_rate/action_model': 1.0514009510432988e-05, 'epoch': 1.64}) +[2026-06-23 22:45:29,661][starVLA.training.train_starvla][INFO] - Step 4242, Loss: {'train/loss': 0.1551281064748764, 'train/grad_norm_pre_clip': 23.8475399017334, 'timing/data': 0.0005562261212617159, 'timing/model': 1.6809394559822977, 'learning_rate/qwen_vl_interface': 1.0499776334597796e-06, 'learning_rate/action_model': 1.0499776334597797e-05, 'epoch': 1.64}) +[2026-06-23 22:45:32,981][starVLA.training.train_starvla][INFO] - Step 4243, Loss: {'train/loss': 0.04727088660001755, 'train/grad_norm_pre_clip': 21.909015655517578, 'timing/data': 0.0005905299913138151, 'timing/model': 1.6886559328995645, 'learning_rate/qwen_vl_interface': 1.0485560423468064e-06, 'learning_rate/action_model': 1.0485560423468063e-05, 'epoch': 1.64}) +[2026-06-23 22:45:36,294][starVLA.training.train_starvla][INFO] - Step 4244, Loss: {'train/loss': 0.0586300864815712, 'train/grad_norm_pre_clip': 33.21141815185547, 'timing/data': 0.000581823056563735, 'timing/model': 1.6845578150823712, 'learning_rate/qwen_vl_interface': 1.0471361782887421e-06, 'learning_rate/action_model': 1.0471361782887421e-05, 'epoch': 1.64}) +[2026-06-23 22:45:39,612][starVLA.training.train_starvla][INFO] - Step 4245, Loss: {'train/loss': 0.08877592533826828, 'train/grad_norm_pre_clip': 20.031408309936523, 'timing/data': 0.0005763028748333454, 'timing/model': 1.6856504979077727, 'learning_rate/qwen_vl_interface': 1.0457180418692385e-06, 'learning_rate/action_model': 1.0457180418692383e-05, 'epoch': 1.64}) +[2026-06-23 22:45:42,928][starVLA.training.train_starvla][INFO] - Step 4246, Loss: {'train/loss': 0.10411512851715088, 'train/grad_norm_pre_clip': 19.446334838867188, 'timing/data': 0.0005520479753613472, 'timing/model': 1.6861169829498976, 'learning_rate/qwen_vl_interface': 1.0443016336712395e-06, 'learning_rate/action_model': 1.0443016336712393e-05, 'epoch': 1.64}) +[2026-06-23 22:45:46,245][starVLA.training.train_starvla][INFO] - Step 4247, Loss: {'train/loss': 0.05928843840956688, 'train/grad_norm_pre_clip': 10.573064804077148, 'timing/data': 0.0005503748543560505, 'timing/model': 1.6843269439414144, 'learning_rate/qwen_vl_interface': 1.0428869542769744e-06, 'learning_rate/action_model': 1.0428869542769744e-05, 'epoch': 1.64}) +[2026-06-23 22:45:49,563][starVLA.training.train_starvla][INFO] - Step 4248, Loss: {'train/loss': 0.19945096969604492, 'train/grad_norm_pre_clip': 21.125974655151367, 'timing/data': 0.0005469180177897215, 'timing/model': 1.6866403119638562, 'learning_rate/qwen_vl_interface': 1.0414740042679663e-06, 'learning_rate/action_model': 1.0414740042679662e-05, 'epoch': 1.64}) +[2026-06-23 22:45:52,879][starVLA.training.train_starvla][INFO] - Step 4249, Loss: {'train/loss': 0.12976109981536865, 'train/grad_norm_pre_clip': 20.091825485229492, 'timing/data': 0.0007190320175141096, 'timing/model': 1.6815549989696592, 'learning_rate/qwen_vl_interface': 1.040062784225025e-06, 'learning_rate/action_model': 1.0400627842250249e-05, 'epoch': 1.64}) +[2026-06-23 22:46:38,010][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4250 +[2026-06-23 22:46:38,010][starVLA.training.train_starvla][INFO] - Step 4250, Loss: {'train/loss': 0.17315354943275452, 'train/grad_norm_pre_clip': 15.038045883178711, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 31.11999999999994, 'rl_games_eval/mid_train/mean_length': 268.2, 'rl_games_eval/mid_train/std_reward': 33.25449744019579, 'rl_games_eval/mid_train/std_length': 269.435261240989, 'rl_games_eval/mid_train/macro_mean_reward': 31.11999999999994, 'rl_games_eval/mid_train/macro_mean_length': 268.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 31.11999999999994, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 268.2, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 33.25449744019579, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 269.435261240989, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 31.11999999999994, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005711440462619066, 'timing/model': 1.6931096620392054, 'learning_rate/qwen_vl_interface': 1.0386532947282496e-06, 'learning_rate/action_model': 1.0386532947282495e-05, 'epoch': 1.64}) +[2026-06-23 22:46:41,290][starVLA.training.train_starvla][INFO] - Step 4251, Loss: {'train/loss': 0.11262916028499603, 'train/grad_norm_pre_clip': 22.765178680419922, 'timing/data': 0.0005559760611504316, 'timing/model': 1.6677565360441804, 'learning_rate/qwen_vl_interface': 1.0372455363570282e-06, 'learning_rate/action_model': 1.0372455363570283e-05, 'epoch': 1.65}) +[2026-06-23 22:46:44,578][starVLA.training.train_starvla][INFO] - Step 4252, Loss: {'train/loss': 0.18183213472366333, 'train/grad_norm_pre_clip': 36.22489547729492, 'timing/data': 0.0005600329022854567, 'timing/model': 1.672291160095483, 'learning_rate/qwen_vl_interface': 1.035839509690039e-06, 'learning_rate/action_model': 1.0358395096900388e-05, 'epoch': 1.65}) +[2026-06-23 22:46:47,968][starVLA.training.train_starvla][INFO] - Step 4253, Loss: {'train/loss': 0.057237692177295685, 'train/grad_norm_pre_clip': 13.9740571975708, 'timing/data': 0.0006470768712460995, 'timing/model': 1.7657403519842774, 'learning_rate/qwen_vl_interface': 1.0344352153052429e-06, 'learning_rate/action_model': 1.0344352153052427e-05, 'epoch': 1.65}) +[2026-06-23 22:46:51,265][starVLA.training.train_starvla][INFO] - Step 4254, Loss: {'train/loss': 0.14053821563720703, 'train/grad_norm_pre_clip': 22.358600616455078, 'timing/data': 0.0005705629009753466, 'timing/model': 1.6761815110221505, 'learning_rate/qwen_vl_interface': 1.033032653779894e-06, 'learning_rate/action_model': 1.033032653779894e-05, 'epoch': 1.65}) +[2026-06-23 22:46:54,569][starVLA.training.train_starvla][INFO] - Step 4255, Loss: {'train/loss': 0.13473065197467804, 'train/grad_norm_pre_clip': 29.590105056762695, 'timing/data': 0.0005478900857269764, 'timing/model': 1.6804837279487401, 'learning_rate/qwen_vl_interface': 1.0316318256905325e-06, 'learning_rate/action_model': 1.0316318256905324e-05, 'epoch': 1.65}) +[2026-06-23 22:46:57,877][starVLA.training.train_starvla][INFO] - Step 4256, Loss: {'train/loss': 0.06791453808546066, 'train/grad_norm_pre_clip': 20.3670711517334, 'timing/data': 0.0006752898916602135, 'timing/model': 1.678529652999714, 'learning_rate/qwen_vl_interface': 1.0302327316129856e-06, 'learning_rate/action_model': 1.0302327316129855e-05, 'epoch': 1.65}) +[2026-06-23 22:47:01,180][starVLA.training.train_starvla][INFO] - Step 4257, Loss: {'train/loss': 0.03640566021203995, 'train/grad_norm_pre_clip': 9.146332740783691, 'timing/data': 0.0005758830811828375, 'timing/model': 1.6787702930159867, 'learning_rate/qwen_vl_interface': 1.028835372122368e-06, 'learning_rate/action_model': 1.0288353721223682e-05, 'epoch': 1.65}) +[2026-06-23 22:47:04,484][starVLA.training.train_starvla][INFO] - Step 4258, Loss: {'train/loss': 0.1065240278840065, 'train/grad_norm_pre_clip': 18.44496726989746, 'timing/data': 0.0005774849560111761, 'timing/model': 1.6797986140009016, 'learning_rate/qwen_vl_interface': 1.0274397477930822e-06, 'learning_rate/action_model': 1.0274397477930822e-05, 'epoch': 1.65}) +[2026-06-23 22:47:07,787][starVLA.training.train_starvla][INFO] - Step 4259, Loss: {'train/loss': 0.09872489422559738, 'train/grad_norm_pre_clip': 44.85873031616211, 'timing/data': 0.0005550140049308538, 'timing/model': 1.6773981899023056, 'learning_rate/qwen_vl_interface': 1.0260458591988164e-06, 'learning_rate/action_model': 1.0260458591988165e-05, 'epoch': 1.65}) +[2026-06-23 22:47:11,092][starVLA.training.train_starvla][INFO] - Step 4260, Loss: {'train/loss': 0.11834730207920074, 'train/grad_norm_pre_clip': 44.00344467163086, 'timing/data': 0.0005539320409297943, 'timing/model': 1.6798304510302842, 'learning_rate/qwen_vl_interface': 1.0246537069125445e-06, 'learning_rate/action_model': 1.0246537069125444e-05, 'epoch': 1.65}) +[2026-06-23 22:47:14,399][starVLA.training.train_starvla][INFO] - Step 4261, Loss: {'train/loss': 0.11349322646856308, 'train/grad_norm_pre_clip': 19.970882415771484, 'timing/data': 0.0005659551825374365, 'timing/model': 1.6820560740306973, 'learning_rate/qwen_vl_interface': 1.0232632915065286e-06, 'learning_rate/action_model': 1.0232632915065285e-05, 'epoch': 1.65}) +[2026-06-23 22:47:17,710][starVLA.training.train_starvla][INFO] - Step 4262, Loss: {'train/loss': 0.09653984755277634, 'train/grad_norm_pre_clip': 25.878740310668945, 'timing/data': 0.0005575190298259258, 'timing/model': 1.6811470629181713, 'learning_rate/qwen_vl_interface': 1.021874613552316e-06, 'learning_rate/action_model': 1.0218746135523157e-05, 'epoch': 1.65}) +[2026-06-23 22:47:21,022][starVLA.training.train_starvla][INFO] - Step 4263, Loss: {'train/loss': 0.0992661640048027, 'train/grad_norm_pre_clip': 14.400978088378906, 'timing/data': 0.000551536912098527, 'timing/model': 1.6822555449325591, 'learning_rate/qwen_vl_interface': 1.0204876736207384e-06, 'learning_rate/action_model': 1.0204876736207384e-05, 'epoch': 1.65}) +[2026-06-23 22:47:24,332][starVLA.training.train_starvla][INFO] - Step 4264, Loss: {'train/loss': 0.05925615504384041, 'train/grad_norm_pre_clip': 21.297639846801758, 'timing/data': 0.0006311170291155577, 'timing/model': 1.6823055879212916, 'learning_rate/qwen_vl_interface': 1.0191024722819163e-06, 'learning_rate/action_model': 1.0191024722819162e-05, 'epoch': 1.65}) +[2026-06-23 22:47:27,636][starVLA.training.train_starvla][INFO] - Step 4265, Loss: {'train/loss': 0.14918522536754608, 'train/grad_norm_pre_clip': 27.665796279907227, 'timing/data': 0.0005468379240483046, 'timing/model': 1.6828209548257291, 'learning_rate/qwen_vl_interface': 1.017719010105253e-06, 'learning_rate/action_model': 1.017719010105253e-05, 'epoch': 1.65}) +[2026-06-23 22:47:30,954][starVLA.training.train_starvla][INFO] - Step 4266, Loss: {'train/loss': 0.05027301609516144, 'train/grad_norm_pre_clip': 26.791725158691406, 'timing/data': 0.0005691300611943007, 'timing/model': 1.6861264680046588, 'learning_rate/qwen_vl_interface': 1.016337287659437e-06, 'learning_rate/action_model': 1.0163372876594368e-05, 'epoch': 1.65}) +[2026-06-23 22:47:34,267][starVLA.training.train_starvla][INFO] - Step 4267, Loss: {'train/loss': 0.16871975362300873, 'train/grad_norm_pre_clip': 20.10643768310547, 'timing/data': 0.0005832859314978123, 'timing/model': 1.6811688558664173, 'learning_rate/qwen_vl_interface': 1.0149573055124427e-06, 'learning_rate/action_model': 1.0149573055124426e-05, 'epoch': 1.65}) +[2026-06-23 22:47:37,872][starVLA.training.train_starvla][INFO] - Step 4268, Loss: {'train/loss': 0.1105353981256485, 'train/grad_norm_pre_clip': 21.134504318237305, 'timing/data': 0.0005707328673452139, 'timing/model': 1.9803483269643039, 'learning_rate/qwen_vl_interface': 1.013579064231529e-06, 'learning_rate/action_model': 1.013579064231529e-05, 'epoch': 1.65}) +[2026-06-23 22:47:41,181][starVLA.training.train_starvla][INFO] - Step 4269, Loss: {'train/loss': 0.04902710020542145, 'train/grad_norm_pre_clip': 28.512006759643555, 'timing/data': 0.0005680178292095661, 'timing/model': 1.6813263578806072, 'learning_rate/qwen_vl_interface': 1.012202564383239e-06, 'learning_rate/action_model': 1.0122025643832388e-05, 'epoch': 1.65}) +[2026-06-23 22:47:44,491][starVLA.training.train_starvla][INFO] - Step 4270, Loss: {'train/loss': 0.13686048984527588, 'train/grad_norm_pre_clip': 26.63829803466797, 'timing/data': 0.0006875519175082445, 'timing/model': 1.6826967119704932, 'learning_rate/qwen_vl_interface': 1.0108278065333978e-06, 'learning_rate/action_model': 1.0108278065333978e-05, 'epoch': 1.65}) +[2026-06-23 22:47:47,801][starVLA.training.train_starvla][INFO] - Step 4271, Loss: {'train/loss': 0.14993403851985931, 'train/grad_norm_pre_clip': 26.246601104736328, 'timing/data': 0.0005593919195234776, 'timing/model': 1.681270519969985, 'learning_rate/qwen_vl_interface': 1.0094547912471183e-06, 'learning_rate/action_model': 1.0094547912471184e-05, 'epoch': 1.65}) +[2026-06-23 22:47:51,110][starVLA.training.train_starvla][INFO] - Step 4272, Loss: {'train/loss': 0.19309689104557037, 'train/grad_norm_pre_clip': 22.09541893005371, 'timing/data': 0.0005487320013344288, 'timing/model': 1.6841358351521194, 'learning_rate/qwen_vl_interface': 1.0080835190887949e-06, 'learning_rate/action_model': 1.0080835190887947e-05, 'epoch': 1.65}) +[2026-06-23 22:47:54,425][starVLA.training.train_starvla][INFO] - Step 4273, Loss: {'train/loss': 0.08131293207406998, 'train/grad_norm_pre_clip': 15.049540519714355, 'timing/data': 0.0006435101386159658, 'timing/model': 1.6850354950875044, 'learning_rate/qwen_vl_interface': 1.0067139906221052e-06, 'learning_rate/action_model': 1.0067139906221052e-05, 'epoch': 1.65}) +[2026-06-23 22:47:57,742][starVLA.training.train_starvla][INFO] - Step 4274, Loss: {'train/loss': 0.04925598204135895, 'train/grad_norm_pre_clip': 30.63395118713379, 'timing/data': 0.0005648320075124502, 'timing/model': 1.6857300088740885, 'learning_rate/qwen_vl_interface': 1.0053462064100112e-06, 'learning_rate/action_model': 1.0053462064100112e-05, 'epoch': 1.65}) +[2026-06-23 22:48:01,050][starVLA.training.train_starvla][INFO] - Step 4275, Loss: {'train/loss': 0.1614590734243393, 'train/grad_norm_pre_clip': 47.016578674316406, 'timing/data': 0.0005737789906561375, 'timing/model': 1.6816140690352768, 'learning_rate/qwen_vl_interface': 1.003980167014757e-06, 'learning_rate/action_model': 1.003980167014757e-05, 'epoch': 1.65}) +[2026-06-23 22:48:04,362][starVLA.training.train_starvla][INFO] - Step 4276, Loss: {'train/loss': 0.19599974155426025, 'train/grad_norm_pre_clip': 25.6046142578125, 'timing/data': 0.0005757829640060663, 'timing/model': 1.6830072470474988, 'learning_rate/qwen_vl_interface': 1.0026158729978702e-06, 'learning_rate/action_model': 1.0026158729978702e-05, 'epoch': 1.65}) +[2026-06-23 22:48:07,677][starVLA.training.train_starvla][INFO] - Step 4277, Loss: {'train/loss': 0.182033970952034, 'train/grad_norm_pre_clip': 20.71097755432129, 'timing/data': 0.0005821939557790756, 'timing/model': 1.6837705201469362, 'learning_rate/qwen_vl_interface': 1.0012533249201604e-06, 'learning_rate/action_model': 1.0012533249201602e-05, 'epoch': 1.66}) +[2026-06-23 22:48:10,992][starVLA.training.train_starvla][INFO] - Step 4278, Loss: {'train/loss': 0.1207590103149414, 'train/grad_norm_pre_clip': 19.854534149169922, 'timing/data': 0.0006012211088091135, 'timing/model': 1.6833443469367921, 'learning_rate/qwen_vl_interface': 9.99892523341719e-07, 'learning_rate/action_model': 9.99892523341719e-06, 'epoch': 1.66}) +[2026-06-23 22:48:14,303][starVLA.training.train_starvla][INFO] - Step 4279, Loss: {'train/loss': 0.21001945436000824, 'train/grad_norm_pre_clip': 23.19849395751953, 'timing/data': 0.0005974429659545422, 'timing/model': 1.6846316859591752, 'learning_rate/qwen_vl_interface': 9.98533468821922e-07, 'learning_rate/action_model': 9.98533468821922e-06, 'epoch': 1.66}) +[2026-06-23 22:48:17,617][starVLA.training.train_starvla][INFO] - Step 4280, Loss: {'train/loss': 0.04633104056119919, 'train/grad_norm_pre_clip': 14.965865135192871, 'timing/data': 0.0006138640455901623, 'timing/model': 1.6840264068450779, 'learning_rate/qwen_vl_interface': 9.971761619194231e-07, 'learning_rate/action_model': 9.97176161919423e-06, 'epoch': 1.66}) +[2026-06-23 22:48:20,931][starVLA.training.train_starvla][INFO] - Step 4281, Loss: {'train/loss': 0.11305591464042664, 'train/grad_norm_pre_clip': 24.344194412231445, 'timing/data': 0.0005720260087400675, 'timing/model': 1.683717564912513, 'learning_rate/qwen_vl_interface': 9.958206031921606e-07, 'learning_rate/action_model': 9.958206031921604e-06, 'epoch': 1.66}) +[2026-06-23 22:48:24,246][starVLA.training.train_starvla][INFO] - Step 4282, Loss: {'train/loss': 0.0612248033285141, 'train/grad_norm_pre_clip': 10.356902122497559, 'timing/data': 0.0005494130309671164, 'timing/model': 1.6838029529899359, 'learning_rate/qwen_vl_interface': 9.944667931973537e-07, 'learning_rate/action_model': 9.944667931973536e-06, 'epoch': 1.66}) +[2026-06-23 22:48:27,797][starVLA.training.train_starvla][INFO] - Step 4283, Loss: {'train/loss': 0.13131140172481537, 'train/grad_norm_pre_clip': 54.943328857421875, 'timing/data': 0.0005508461035788059, 'timing/model': 1.9199349209666252, 'learning_rate/qwen_vl_interface': 9.93114732491503e-07, 'learning_rate/action_model': 9.93114732491503e-06, 'epoch': 1.66}) +[2026-06-23 22:48:31,114][starVLA.training.train_starvla][INFO] - Step 4284, Loss: {'train/loss': 0.1982147991657257, 'train/grad_norm_pre_clip': 22.527812957763672, 'timing/data': 0.0005742090288549662, 'timing/model': 1.6820561750791967, 'learning_rate/qwen_vl_interface': 9.91764421630389e-07, 'learning_rate/action_model': 9.91764421630389e-06, 'epoch': 1.66}) +[2026-06-23 22:48:34,426][starVLA.training.train_starvla][INFO] - Step 4285, Loss: {'train/loss': 0.07763120532035828, 'train/grad_norm_pre_clip': 22.638315200805664, 'timing/data': 0.0005721959751099348, 'timing/model': 1.6823769968468696, 'learning_rate/qwen_vl_interface': 9.904158611690732e-07, 'learning_rate/action_model': 9.90415861169073e-06, 'epoch': 1.66}) +[2026-06-23 22:48:37,738][starVLA.training.train_starvla][INFO] - Step 4286, Loss: {'train/loss': 0.12338439375162125, 'train/grad_norm_pre_clip': 16.906959533691406, 'timing/data': 0.0005543420556932688, 'timing/model': 1.6819530380889773, 'learning_rate/qwen_vl_interface': 9.890690516618987e-07, 'learning_rate/action_model': 9.890690516618985e-06, 'epoch': 1.66}) +[2026-06-23 22:48:41,052][starVLA.training.train_starvla][INFO] - Step 4287, Loss: {'train/loss': 0.08836404979228973, 'train/grad_norm_pre_clip': 38.42443084716797, 'timing/data': 0.0005540819838643074, 'timing/model': 1.683641562005505, 'learning_rate/qwen_vl_interface': 9.877239936624875e-07, 'learning_rate/action_model': 9.877239936624875e-06, 'epoch': 1.66}) +[2026-06-23 22:48:44,370][starVLA.training.train_starvla][INFO] - Step 4288, Loss: {'train/loss': 0.10140177607536316, 'train/grad_norm_pre_clip': 21.85098648071289, 'timing/data': 0.0005655130371451378, 'timing/model': 1.6852792799472809, 'learning_rate/qwen_vl_interface': 9.863806877237421e-07, 'learning_rate/action_model': 9.863806877237422e-06, 'epoch': 1.66}) +[2026-06-23 22:48:47,686][starVLA.training.train_starvla][INFO] - Step 4289, Loss: {'train/loss': 0.12357921153306961, 'train/grad_norm_pre_clip': 22.215879440307617, 'timing/data': 0.000549783930182457, 'timing/model': 1.685464246198535, 'learning_rate/qwen_vl_interface': 9.850391343978454e-07, 'learning_rate/action_model': 9.850391343978454e-06, 'epoch': 1.66}) +[2026-06-23 22:48:51,000][starVLA.training.train_starvla][INFO] - Step 4290, Loss: {'train/loss': 0.1019628643989563, 'train/grad_norm_pre_clip': 17.082122802734375, 'timing/data': 0.0005454360507428646, 'timing/model': 1.6862662320490927, 'learning_rate/qwen_vl_interface': 9.836993342362596e-07, 'learning_rate/action_model': 9.836993342362596e-06, 'epoch': 1.66}) +[2026-06-23 22:48:54,318][starVLA.training.train_starvla][INFO] - Step 4291, Loss: {'train/loss': 0.06083774194121361, 'train/grad_norm_pre_clip': 32.37126541137695, 'timing/data': 0.0005520479753613472, 'timing/model': 1.6836216789670289, 'learning_rate/qwen_vl_interface': 9.823612877897246e-07, 'learning_rate/action_model': 9.823612877897245e-06, 'epoch': 1.66}) +[2026-06-23 22:48:57,634][starVLA.training.train_starvla][INFO] - Step 4292, Loss: {'train/loss': 0.053807903081178665, 'train/grad_norm_pre_clip': 31.760276794433594, 'timing/data': 0.0005608340725302696, 'timing/model': 1.6836108069401234, 'learning_rate/qwen_vl_interface': 9.81024995608261e-07, 'learning_rate/action_model': 9.810249956082609e-06, 'epoch': 1.66}) +[2026-06-23 22:49:00,952][starVLA.training.train_starvla][INFO] - Step 4293, Loss: {'train/loss': 0.10427399724721909, 'train/grad_norm_pre_clip': 23.562320709228516, 'timing/data': 0.000575491925701499, 'timing/model': 1.6862459101248533, 'learning_rate/qwen_vl_interface': 9.796904582411685e-07, 'learning_rate/action_model': 9.796904582411684e-06, 'epoch': 1.66}) +[2026-06-23 22:49:04,269][starVLA.training.train_starvla][INFO] - Step 4294, Loss: {'train/loss': 0.085843525826931, 'train/grad_norm_pre_clip': 21.992212295532227, 'timing/data': 0.000597895123064518, 'timing/model': 1.6865538370329887, 'learning_rate/qwen_vl_interface': 9.78357676237025e-07, 'learning_rate/action_model': 9.78357676237025e-06, 'epoch': 1.66}) +[2026-06-23 22:49:07,588][starVLA.training.train_starvla][INFO] - Step 4295, Loss: {'train/loss': 0.024975964799523354, 'train/grad_norm_pre_clip': 16.149154663085938, 'timing/data': 0.0005836181808263063, 'timing/model': 1.6828376590274274, 'learning_rate/qwen_vl_interface': 9.77026650143687e-07, 'learning_rate/action_model': 9.77026650143687e-06, 'epoch': 1.66}) +[2026-06-23 22:49:10,905][starVLA.training.train_starvla][INFO] - Step 4296, Loss: {'train/loss': 0.07096575945615768, 'train/grad_norm_pre_clip': 17.658740997314453, 'timing/data': 0.0005504949949681759, 'timing/model': 1.6845321939326823, 'learning_rate/qwen_vl_interface': 9.756973805082882e-07, 'learning_rate/action_model': 9.756973805082883e-06, 'epoch': 1.66}) +[2026-06-23 22:49:14,349][starVLA.training.train_starvla][INFO] - Step 4297, Loss: {'train/loss': 0.14034675061702728, 'train/grad_norm_pre_clip': 53.38147735595703, 'timing/data': 0.0006142850033938885, 'timing/model': 1.8173805379774421, 'learning_rate/qwen_vl_interface': 9.743698678772434e-07, 'learning_rate/action_model': 9.743698678772433e-06, 'epoch': 1.66}) +[2026-06-23 22:49:17,669][starVLA.training.train_starvla][INFO] - Step 4298, Loss: {'train/loss': 0.1656913459300995, 'train/grad_norm_pre_clip': 64.43021392822266, 'timing/data': 0.0005501741543412209, 'timing/model': 1.6839597839862108, 'learning_rate/qwen_vl_interface': 9.730441127962395e-07, 'learning_rate/action_model': 9.730441127962395e-06, 'epoch': 1.66}) +[2026-06-23 22:49:20,982][starVLA.training.train_starvla][INFO] - Step 4299, Loss: {'train/loss': 0.16239959001541138, 'train/grad_norm_pre_clip': 26.096649169921875, 'timing/data': 0.0005596422124654055, 'timing/model': 1.683198771905154, 'learning_rate/qwen_vl_interface': 9.717201158102465e-07, 'learning_rate/action_model': 9.717201158102466e-06, 'epoch': 1.66}) +[2026-06-23 22:49:24,301][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4300 +[2026-06-23 22:50:56,367][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4300 +[2026-06-23 22:50:56,368][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4300 +[2026-06-23 22:52:06,401][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4300 +[2026-06-23 22:52:06,401][starVLA.training.train_starvla][INFO] - Step 4300, Loss: {'train/loss': 0.06520895659923553, 'train/grad_norm_pre_clip': 51.528587341308594, 'eval/action_loss/seconds': 92.0639146368485, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06613437086343765, 'eval/latency_2/loss': 0.06613437086343765, 'eval/flappy/loss': 0.06613437086343765, 'eval/flappy/latency_2/loss': 0.06613437086343765, 'eval/flap_precision': 0.9267015706806283, 'eval/flap_recall': 0.9554655870445344, 'eval/flap_f1': 0.9408637873754153, 'eval/flap_rate_model': 0.21222222222222223, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9408637873754153, 'eval/latency_2/flap_precision': 0.9267015706806283, 'eval/latency_2/flap_recall': 0.9554655870445344, 'eval/latency_2/flap_f1': 0.9408637873754153, 'eval/latency_2/flap_rate_model': 0.21222222222222223, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9408637873754153, 'eval/action_classification/data_wait_seconds': 0.9099829196929932, 'eval/action_classification/predict_seconds': 68.99047088623047, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005632189568132162, 'timing/model': 1.6864454948808998, 'learning_rate/qwen_vl_interface': 9.703978774635094e-07, 'learning_rate/action_model': 9.703978774635095e-06, 'epoch': 1.66}) +[2026-06-23 22:52:09,698][starVLA.training.train_starvla][INFO] - Step 4301, Loss: {'train/loss': 0.15859965980052948, 'train/grad_norm_pre_clip': 89.59797668457031, 'timing/data': 0.0005919830873608589, 'timing/model': 1.6743506269995123, 'learning_rate/qwen_vl_interface': 9.6907739829955e-07, 'learning_rate/action_model': 9.6907739829955e-06, 'epoch': 1.66}) +[2026-06-23 22:52:12,994][starVLA.training.train_starvla][INFO] - Step 4302, Loss: {'train/loss': 0.11660860478878021, 'train/grad_norm_pre_clip': 48.36653518676758, 'timing/data': 0.0005644508637487888, 'timing/model': 1.6753829941153526, 'learning_rate/qwen_vl_interface': 9.677586788611681e-07, 'learning_rate/action_model': 9.677586788611681e-06, 'epoch': 1.66}) +[2026-06-23 22:52:16,288][starVLA.training.train_starvla][INFO] - Step 4303, Loss: {'train/loss': 0.05099703371524811, 'train/grad_norm_pre_clip': 19.3167667388916, 'timing/data': 0.0005681391339749098, 'timing/model': 1.6764547219499946, 'learning_rate/qwen_vl_interface': 9.664417196904382e-07, 'learning_rate/action_model': 9.664417196904382e-06, 'epoch': 1.67}) +[2026-06-23 22:52:19,587][starVLA.training.train_starvla][INFO] - Step 4304, Loss: {'train/loss': 0.11567308008670807, 'train/grad_norm_pre_clip': 20.00015640258789, 'timing/data': 0.000559832900762558, 'timing/model': 1.6736427308060229, 'learning_rate/qwen_vl_interface': 9.651265213287134e-07, 'learning_rate/action_model': 9.651265213287134e-06, 'epoch': 1.67}) +[2026-06-23 22:52:23,138][starVLA.training.train_starvla][INFO] - Step 4305, Loss: {'train/loss': 0.03354726731777191, 'train/grad_norm_pre_clip': 26.781192779541016, 'timing/data': 0.0005533599760383368, 'timing/model': 1.9301107497885823, 'learning_rate/qwen_vl_interface': 9.638130843166218e-07, 'learning_rate/action_model': 9.638130843166219e-06, 'epoch': 1.67}) +[2026-06-23 22:52:26,437][starVLA.training.train_starvla][INFO] - Step 4306, Loss: {'train/loss': 0.078947052359581, 'train/grad_norm_pre_clip': 23.17439079284668, 'timing/data': 0.0005780360661447048, 'timing/model': 1.6777681671082973, 'learning_rate/qwen_vl_interface': 9.62501409194068e-07, 'learning_rate/action_model': 9.625014091940678e-06, 'epoch': 1.67}) +[2026-06-23 22:52:29,738][starVLA.training.train_starvla][INFO] - Step 4307, Loss: {'train/loss': 0.2401738166809082, 'train/grad_norm_pre_clip': 17.00811767578125, 'timing/data': 0.0005707631353288889, 'timing/model': 1.6780879178550094, 'learning_rate/qwen_vl_interface': 9.611914965002317e-07, 'learning_rate/action_model': 9.611914965002316e-06, 'epoch': 1.67}) +[2026-06-23 22:52:33,041][starVLA.training.train_starvla][INFO] - Step 4308, Loss: {'train/loss': 0.08394946157932281, 'train/grad_norm_pre_clip': 17.587440490722656, 'timing/data': 0.0005591609515249729, 'timing/model': 1.6768650270532817, 'learning_rate/qwen_vl_interface': 9.598833467735686e-07, 'learning_rate/action_model': 9.598833467735685e-06, 'epoch': 1.67}) +[2026-06-23 22:52:36,340][starVLA.training.train_starvla][INFO] - Step 4309, Loss: {'train/loss': 0.17280656099319458, 'train/grad_norm_pre_clip': 28.46580696105957, 'timing/data': 0.0005836370401084423, 'timing/model': 1.6731606510002166, 'learning_rate/qwen_vl_interface': 9.585769605518097e-07, 'learning_rate/action_model': 9.585769605518097e-06, 'epoch': 1.67}) +[2026-06-23 22:52:39,640][starVLA.training.train_starvla][INFO] - Step 4310, Loss: {'train/loss': 0.1376255452632904, 'train/grad_norm_pre_clip': 28.85895347595215, 'timing/data': 0.0005687491502612829, 'timing/model': 1.6778880490455776, 'learning_rate/qwen_vl_interface': 9.572723383719611e-07, 'learning_rate/action_model': 9.572723383719612e-06, 'epoch': 1.67}) +[2026-06-23 22:52:42,938][starVLA.training.train_starvla][INFO] - Step 4311, Loss: {'train/loss': 0.049880582839250565, 'train/grad_norm_pre_clip': 20.74662971496582, 'timing/data': 0.0006134631112217903, 'timing/model': 1.678283441113308, 'learning_rate/qwen_vl_interface': 9.559694807703041e-07, 'learning_rate/action_model': 9.55969480770304e-06, 'epoch': 1.67}) +[2026-06-23 22:52:46,241][starVLA.training.train_starvla][INFO] - Step 4312, Loss: {'train/loss': 0.07232880592346191, 'train/grad_norm_pre_clip': 30.57586669921875, 'timing/data': 0.000603014137595892, 'timing/model': 1.6780344620347023, 'learning_rate/qwen_vl_interface': 9.546683882823945e-07, 'learning_rate/action_model': 9.546683882823945e-06, 'epoch': 1.67}) +[2026-06-23 22:52:49,540][starVLA.training.train_starvla][INFO] - Step 4313, Loss: {'train/loss': 0.18085287511348724, 'train/grad_norm_pre_clip': 39.64153289794922, 'timing/data': 0.0005587500054389238, 'timing/model': 1.6794494709465653, 'learning_rate/qwen_vl_interface': 9.533690614430608e-07, 'learning_rate/action_model': 9.533690614430608e-06, 'epoch': 1.67}) +[2026-06-23 22:52:52,841][starVLA.training.train_starvla][INFO] - Step 4314, Loss: {'train/loss': 0.06472276896238327, 'train/grad_norm_pre_clip': 31.308053970336914, 'timing/data': 0.0005592021625488997, 'timing/model': 1.6803759629838169, 'learning_rate/qwen_vl_interface': 9.520715007864082e-07, 'learning_rate/action_model': 9.52071500786408e-06, 'epoch': 1.67}) +[2026-06-23 22:52:56,142][starVLA.training.train_starvla][INFO] - Step 4315, Loss: {'train/loss': 0.23044253885746002, 'train/grad_norm_pre_clip': 22.547101974487305, 'timing/data': 0.000577166210860014, 'timing/model': 1.674957183888182, 'learning_rate/qwen_vl_interface': 9.50775706845814e-07, 'learning_rate/action_model': 9.50775706845814e-06, 'epoch': 1.67}) +[2026-06-23 22:52:59,449][starVLA.training.train_starvla][INFO] - Step 4316, Loss: {'train/loss': 0.11808869242668152, 'train/grad_norm_pre_clip': 22.480411529541016, 'timing/data': 0.0005964410956948996, 'timing/model': 1.6795938201248646, 'learning_rate/qwen_vl_interface': 9.49481680153931e-07, 'learning_rate/action_model': 9.494816801539309e-06, 'epoch': 1.67}) +[2026-06-23 22:53:02,752][starVLA.training.train_starvla][INFO] - Step 4317, Loss: {'train/loss': 0.06252292543649673, 'train/grad_norm_pre_clip': 22.714567184448242, 'timing/data': 0.0005714651197195053, 'timing/model': 1.6790221950504929, 'learning_rate/qwen_vl_interface': 9.481894212426848e-07, 'learning_rate/action_model': 9.481894212426847e-06, 'epoch': 1.67}) +[2026-06-23 22:53:06,055][starVLA.training.train_starvla][INFO] - Step 4318, Loss: {'train/loss': 0.1389664113521576, 'train/grad_norm_pre_clip': 11.804939270019531, 'timing/data': 0.0005749010015279055, 'timing/model': 1.6780338149983436, 'learning_rate/qwen_vl_interface': 9.468989306432738e-07, 'learning_rate/action_model': 9.468989306432738e-06, 'epoch': 1.67}) +[2026-06-23 22:53:09,367][starVLA.training.train_starvla][INFO] - Step 4319, Loss: {'train/loss': 0.07604768127202988, 'train/grad_norm_pre_clip': 47.77524948120117, 'timing/data': 0.0006112491246312857, 'timing/model': 1.6821634990628809, 'learning_rate/qwen_vl_interface': 9.456102088861702e-07, 'learning_rate/action_model': 9.456102088861701e-06, 'epoch': 1.67}) +[2026-06-23 22:53:12,958][starVLA.training.train_starvla][INFO] - Step 4320, Loss: {'train/loss': 0.09384743869304657, 'train/grad_norm_pre_clip': 17.053150177001953, 'timing/data': 0.0006253358442336321, 'timing/model': 1.675502595026046, 'learning_rate/qwen_vl_interface': 9.443232565011188e-07, 'learning_rate/action_model': 9.443232565011187e-06, 'epoch': 1.67}) +[2026-06-23 22:53:16,265][starVLA.training.train_starvla][INFO] - Step 4321, Loss: {'train/loss': 0.06096363812685013, 'train/grad_norm_pre_clip': 13.627250671386719, 'timing/data': 0.0005576589610427618, 'timing/model': 1.679769376059994, 'learning_rate/qwen_vl_interface': 9.430380740171369e-07, 'learning_rate/action_model': 9.430380740171368e-06, 'epoch': 1.67}) +[2026-06-23 22:53:19,575][starVLA.training.train_starvla][INFO] - Step 4322, Loss: {'train/loss': 0.04618155583739281, 'train/grad_norm_pre_clip': 19.81612777709961, 'timing/data': 0.000542470021173358, 'timing/model': 1.6810784661211073, 'learning_rate/qwen_vl_interface': 9.417546619625143e-07, 'learning_rate/action_model': 9.417546619625143e-06, 'epoch': 1.67}) +[2026-06-23 22:53:22,879][starVLA.training.train_starvla][INFO] - Step 4323, Loss: {'train/loss': 0.06527117639780045, 'train/grad_norm_pre_clip': 17.136537551879883, 'timing/data': 0.0006221088115125895, 'timing/model': 1.6818054819013923, 'learning_rate/qwen_vl_interface': 9.404730208648144e-07, 'learning_rate/action_model': 9.404730208648144e-06, 'epoch': 1.67}) +[2026-06-23 22:53:26,184][starVLA.training.train_starvla][INFO] - Step 4324, Loss: {'train/loss': 0.11141499131917953, 'train/grad_norm_pre_clip': 19.84815788269043, 'timing/data': 0.0006217688787728548, 'timing/model': 1.677919636014849, 'learning_rate/qwen_vl_interface': 9.3919315125087e-07, 'learning_rate/action_model': 9.3919315125087e-06, 'epoch': 1.67}) +[2026-06-23 22:53:29,491][starVLA.training.train_starvla][INFO] - Step 4325, Loss: {'train/loss': 0.13168759644031525, 'train/grad_norm_pre_clip': 45.60135269165039, 'timing/data': 0.0006115599535405636, 'timing/model': 1.6778699809219688, 'learning_rate/qwen_vl_interface': 9.379150536467879e-07, 'learning_rate/action_model': 9.379150536467878e-06, 'epoch': 1.67}) +[2026-06-23 22:53:32,796][starVLA.training.train_starvla][INFO] - Step 4326, Loss: {'train/loss': 0.0467032864689827, 'train/grad_norm_pre_clip': 24.217906951904297, 'timing/data': 0.0005689200479537249, 'timing/model': 1.6794967788737267, 'learning_rate/qwen_vl_interface': 9.366387285779453e-07, 'learning_rate/action_model': 9.366387285779453e-06, 'epoch': 1.67}) +[2026-06-23 22:53:36,101][starVLA.training.train_starvla][INFO] - Step 4327, Loss: {'train/loss': 0.15260517597198486, 'train/grad_norm_pre_clip': 24.41788673400879, 'timing/data': 0.0005602031014859676, 'timing/model': 1.681080776033923, 'learning_rate/qwen_vl_interface': 9.353641765689915e-07, 'learning_rate/action_model': 9.353641765689915e-06, 'epoch': 1.67}) +[2026-06-23 22:53:39,405][starVLA.training.train_starvla][INFO] - Step 4328, Loss: {'train/loss': 0.11877237260341644, 'train/grad_norm_pre_clip': 17.127405166625977, 'timing/data': 0.0006337210070341825, 'timing/model': 1.6784525939729065, 'learning_rate/qwen_vl_interface': 9.340913981438462e-07, 'learning_rate/action_model': 9.340913981438463e-06, 'epoch': 1.68}) +[2026-06-23 22:53:42,709][starVLA.training.train_starvla][INFO] - Step 4329, Loss: {'train/loss': 0.18640470504760742, 'train/grad_norm_pre_clip': 26.97972869873047, 'timing/data': 0.0005732078570872545, 'timing/model': 1.6774893819820136, 'learning_rate/qwen_vl_interface': 9.328203938257019e-07, 'learning_rate/action_model': 9.328203938257018e-06, 'epoch': 1.68}) +[2026-06-23 22:53:46,018][starVLA.training.train_starvla][INFO] - Step 4330, Loss: {'train/loss': 0.0749816745519638, 'train/grad_norm_pre_clip': 15.875572204589844, 'timing/data': 0.0005579290445894003, 'timing/model': 1.6822799281217158, 'learning_rate/qwen_vl_interface': 9.315511641370191e-07, 'learning_rate/action_model': 9.315511641370191e-06, 'epoch': 1.68}) +[2026-06-23 22:53:49,320][starVLA.training.train_starvla][INFO] - Step 4331, Loss: {'train/loss': 0.07615193724632263, 'train/grad_norm_pre_clip': 18.749706268310547, 'timing/data': 0.0005676180589944124, 'timing/model': 1.6821547721046954, 'learning_rate/qwen_vl_interface': 9.302837095995306e-07, 'learning_rate/action_model': 9.302837095995306e-06, 'epoch': 1.68}) +[2026-06-23 22:53:52,628][starVLA.training.train_starvla][INFO] - Step 4332, Loss: {'train/loss': 0.04521966353058815, 'train/grad_norm_pre_clip': 15.62057113647461, 'timing/data': 0.0005815941840410233, 'timing/model': 1.682086813962087, 'learning_rate/qwen_vl_interface': 9.290180307342392e-07, 'learning_rate/action_model': 9.290180307342391e-06, 'epoch': 1.68}) +[2026-06-23 22:53:55,937][starVLA.training.train_starvla][INFO] - Step 4333, Loss: {'train/loss': 0.06358112394809723, 'train/grad_norm_pre_clip': 21.252870559692383, 'timing/data': 0.0005836680065840483, 'timing/model': 1.6821875721216202, 'learning_rate/qwen_vl_interface': 9.277541280614183e-07, 'learning_rate/action_model': 9.277541280614183e-06, 'epoch': 1.68}) +[2026-06-23 22:53:59,243][starVLA.training.train_starvla][INFO] - Step 4334, Loss: {'train/loss': 0.034962400794029236, 'train/grad_norm_pre_clip': 24.393362045288086, 'timing/data': 0.0005813429597765207, 'timing/model': 1.680518823908642, 'learning_rate/qwen_vl_interface': 9.264920021006093e-07, 'learning_rate/action_model': 9.264920021006093e-06, 'epoch': 1.68}) +[2026-06-23 22:54:02,771][starVLA.training.train_starvla][INFO] - Step 4335, Loss: {'train/loss': 0.08191116899251938, 'train/grad_norm_pre_clip': 14.916450500488281, 'timing/data': 0.0005757729522883892, 'timing/model': 1.6800071811303496, 'learning_rate/qwen_vl_interface': 9.252316533706246e-07, 'learning_rate/action_model': 9.252316533706246e-06, 'epoch': 1.68}) +[2026-06-23 22:54:06,080][starVLA.training.train_starvla][INFO] - Step 4336, Loss: {'train/loss': 0.06783811748027802, 'train/grad_norm_pre_clip': 22.343034744262695, 'timing/data': 0.0005810419097542763, 'timing/model': 1.679016298148781, 'learning_rate/qwen_vl_interface': 9.239730823895469e-07, 'learning_rate/action_model': 9.23973082389547e-06, 'epoch': 1.68}) +[2026-06-23 22:54:09,389][starVLA.training.train_starvla][INFO] - Step 4337, Loss: {'train/loss': 0.15063180029392242, 'train/grad_norm_pre_clip': 41.82415008544922, 'timing/data': 0.0005558650009334087, 'timing/model': 1.682140571065247, 'learning_rate/qwen_vl_interface': 9.227162896747264e-07, 'learning_rate/action_model': 9.227162896747264e-06, 'epoch': 1.68}) +[2026-06-23 22:54:12,696][starVLA.training.train_starvla][INFO] - Step 4338, Loss: {'train/loss': 0.11543908715248108, 'train/grad_norm_pre_clip': 22.8087158203125, 'timing/data': 0.0006180631462484598, 'timing/model': 1.680644826963544, 'learning_rate/qwen_vl_interface': 9.214612757427833e-07, 'learning_rate/action_model': 9.214612757427833e-06, 'epoch': 1.68}) +[2026-06-23 22:54:16,002][starVLA.training.train_starvla][INFO] - Step 4339, Loss: {'train/loss': 0.10430054366588593, 'train/grad_norm_pre_clip': 13.091720581054688, 'timing/data': 0.0006106880027800798, 'timing/model': 1.6787446639500558, 'learning_rate/qwen_vl_interface': 9.202080411096071e-07, 'learning_rate/action_model': 9.20208041109607e-06, 'epoch': 1.68}) +[2026-06-23 22:54:19,307][starVLA.training.train_starvla][INFO] - Step 4340, Loss: {'train/loss': 0.12745164334774017, 'train/grad_norm_pre_clip': 28.127819061279297, 'timing/data': 0.0005677379667758942, 'timing/model': 1.678350382950157, 'learning_rate/qwen_vl_interface': 9.189565862903543e-07, 'learning_rate/action_model': 9.189565862903543e-06, 'epoch': 1.68}) +[2026-06-23 22:54:22,606][starVLA.training.train_starvla][INFO] - Step 4341, Loss: {'train/loss': 0.12292773276567459, 'train/grad_norm_pre_clip': 17.278881072998047, 'timing/data': 0.0005478709936141968, 'timing/model': 1.6769850240089, 'learning_rate/qwen_vl_interface': 9.17706911799451e-07, 'learning_rate/action_model': 9.17706911799451e-06, 'epoch': 1.68}) +[2026-06-23 22:54:25,913][starVLA.training.train_starvla][INFO] - Step 4342, Loss: {'train/loss': 0.07070557773113251, 'train/grad_norm_pre_clip': 50.334815979003906, 'timing/data': 0.0005513268988579512, 'timing/model': 1.681154884165153, 'learning_rate/qwen_vl_interface': 9.164590181505912e-07, 'learning_rate/action_model': 9.164590181505911e-06, 'epoch': 1.68}) +[2026-06-23 22:54:29,221][starVLA.training.train_starvla][INFO] - Step 4343, Loss: {'train/loss': 0.08850319683551788, 'train/grad_norm_pre_clip': 23.181642532348633, 'timing/data': 0.0006281309761106968, 'timing/model': 1.678591012954712, 'learning_rate/qwen_vl_interface': 9.15212905856738e-07, 'learning_rate/action_model': 9.15212905856738e-06, 'epoch': 1.68}) +[2026-06-23 22:54:32,565][starVLA.training.train_starvla][INFO] - Step 4344, Loss: {'train/loss': 0.16541816294193268, 'train/grad_norm_pre_clip': 32.23554992675781, 'timing/data': 0.0005780570209026337, 'timing/model': 1.711568061960861, 'learning_rate/qwen_vl_interface': 9.139685754301189e-07, 'learning_rate/action_model': 9.139685754301189e-06, 'epoch': 1.68}) +[2026-06-23 22:54:35,873][starVLA.training.train_starvla][INFO] - Step 4345, Loss: {'train/loss': 0.09570451825857162, 'train/grad_norm_pre_clip': 27.486732482910156, 'timing/data': 0.00067107193171978, 'timing/model': 1.6819533030502498, 'learning_rate/qwen_vl_interface': 9.127260273822324e-07, 'learning_rate/action_model': 9.127260273822325e-06, 'epoch': 1.68}) +[2026-06-23 22:54:39,178][starVLA.training.train_starvla][INFO] - Step 4346, Loss: {'train/loss': 0.32379311323165894, 'train/grad_norm_pre_clip': 21.496732711791992, 'timing/data': 0.0006192750297486782, 'timing/model': 1.6795276110060513, 'learning_rate/qwen_vl_interface': 9.114852622238435e-07, 'learning_rate/action_model': 9.114852622238435e-06, 'epoch': 1.68}) +[2026-06-23 22:54:42,497][starVLA.training.train_starvla][INFO] - Step 4347, Loss: {'train/loss': 0.20925083756446838, 'train/grad_norm_pre_clip': 33.50865173339844, 'timing/data': 0.0005710541736334562, 'timing/model': 1.6865997959394008, 'learning_rate/qwen_vl_interface': 9.102462804649827e-07, 'learning_rate/action_model': 9.102462804649827e-06, 'epoch': 1.68}) +[2026-06-23 22:54:45,806][starVLA.training.train_starvla][INFO] - Step 4348, Loss: {'train/loss': 0.121015265583992, 'train/grad_norm_pre_clip': 18.181955337524414, 'timing/data': 0.0005709540564566851, 'timing/model': 1.6821136609651148, 'learning_rate/qwen_vl_interface': 9.090090826149489e-07, 'learning_rate/action_model': 9.09009082614949e-06, 'epoch': 1.68}) +[2026-06-23 22:54:49,213][starVLA.training.train_starvla][INFO] - Step 4349, Loss: {'train/loss': 0.10313602536916733, 'train/grad_norm_pre_clip': 21.954090118408203, 'timing/data': 0.0005586710758507252, 'timing/model': 1.7784972321242094, 'learning_rate/qwen_vl_interface': 9.077736691823081e-07, 'learning_rate/action_model': 9.07773669182308e-06, 'epoch': 1.68}) +[2026-06-23 22:54:52,523][starVLA.training.train_starvla][INFO] - Step 4350, Loss: {'train/loss': 0.06838609278202057, 'train/grad_norm_pre_clip': 17.880136489868164, 'timing/data': 0.0006621452048420906, 'timing/model': 1.6823586700484157, 'learning_rate/qwen_vl_interface': 9.065400406748912e-07, 'learning_rate/action_model': 9.06540040674891e-06, 'epoch': 1.68}) +[2026-06-23 22:54:55,833][starVLA.training.train_starvla][INFO] - Step 4351, Loss: {'train/loss': 0.10295688360929489, 'train/grad_norm_pre_clip': 18.365520477294922, 'timing/data': 0.0006643489468842745, 'timing/model': 1.6793781372252852, 'learning_rate/qwen_vl_interface': 9.053081975997969e-07, 'learning_rate/action_model': 9.053081975997968e-06, 'epoch': 1.68}) +[2026-06-23 22:54:59,145][starVLA.training.train_starvla][INFO] - Step 4352, Loss: {'train/loss': 0.11651162058115005, 'train/grad_norm_pre_clip': 46.289432525634766, 'timing/data': 0.0005754721350967884, 'timing/model': 1.6815978789236397, 'learning_rate/qwen_vl_interface': 9.040781404633894e-07, 'learning_rate/action_model': 9.040781404633895e-06, 'epoch': 1.68}) +[2026-06-23 22:55:02,454][starVLA.training.train_starvla][INFO] - Step 4353, Loss: {'train/loss': 0.11423980444669724, 'train/grad_norm_pre_clip': 47.49538040161133, 'timing/data': 0.0005793999880552292, 'timing/model': 1.6804530560038984, 'learning_rate/qwen_vl_interface': 9.028498697712982e-07, 'learning_rate/action_model': 9.028498697712981e-06, 'epoch': 1.68}) +[2026-06-23 22:55:05,764][starVLA.training.train_starvla][INFO] - Step 4354, Loss: {'train/loss': 0.16048942506313324, 'train/grad_norm_pre_clip': 49.38529586791992, 'timing/data': 0.0005689191166311502, 'timing/model': 1.6810350590385497, 'learning_rate/qwen_vl_interface': 9.016233860284195e-07, 'learning_rate/action_model': 9.016233860284195e-06, 'epoch': 1.69}) +[2026-06-23 22:55:09,074][starVLA.training.train_starvla][INFO] - Step 4355, Loss: {'train/loss': 0.09379265457391739, 'train/grad_norm_pre_clip': 14.689848899841309, 'timing/data': 0.0005486519075930119, 'timing/model': 1.6816563468892127, 'learning_rate/qwen_vl_interface': 9.003986897389142e-07, 'learning_rate/action_model': 9.003986897389141e-06, 'epoch': 1.69}) +[2026-06-23 22:55:12,384][starVLA.training.train_starvla][INFO] - Step 4356, Loss: {'train/loss': 0.09357239305973053, 'train/grad_norm_pre_clip': 22.14614486694336, 'timing/data': 0.0005636801943182945, 'timing/model': 1.6821696509141475, 'learning_rate/qwen_vl_interface': 8.991757814062094e-07, 'learning_rate/action_model': 8.991757814062093e-06, 'epoch': 1.69}) +[2026-06-23 22:55:15,693][starVLA.training.train_starvla][INFO] - Step 4357, Loss: {'train/loss': 0.18624930083751678, 'train/grad_norm_pre_clip': 16.351911544799805, 'timing/data': 0.000556637067347765, 'timing/model': 1.6803188510239124, 'learning_rate/qwen_vl_interface': 8.979546615329954e-07, 'learning_rate/action_model': 8.979546615329953e-06, 'epoch': 1.69}) +[2026-06-23 22:55:19,009][starVLA.training.train_starvla][INFO] - Step 4358, Loss: {'train/loss': 0.0587276853621006, 'train/grad_norm_pre_clip': 14.4557466506958, 'timing/data': 0.0005472290795296431, 'timing/model': 1.6836541299708188, 'learning_rate/qwen_vl_interface': 8.967353306212286e-07, 'learning_rate/action_model': 8.967353306212286e-06, 'epoch': 1.69}) +[2026-06-23 22:55:22,320][starVLA.training.train_starvla][INFO] - Step 4359, Loss: {'train/loss': 0.11740431934595108, 'train/grad_norm_pre_clip': 17.274694442749023, 'timing/data': 0.0005463368725031614, 'timing/model': 1.6827930009458214, 'learning_rate/qwen_vl_interface': 8.955177891721303e-07, 'learning_rate/action_model': 8.955177891721303e-06, 'epoch': 1.69}) +[2026-06-23 22:55:25,629][starVLA.training.train_starvla][INFO] - Step 4360, Loss: {'train/loss': 0.10618630051612854, 'train/grad_norm_pre_clip': 33.87382125854492, 'timing/data': 0.0005632389802485704, 'timing/model': 1.681216245982796, 'learning_rate/qwen_vl_interface': 8.943020376861858e-07, 'learning_rate/action_model': 8.943020376861857e-06, 'epoch': 1.69}) +[2026-06-23 22:55:28,943][starVLA.training.train_starvla][INFO] - Step 4361, Loss: {'train/loss': 0.16922156512737274, 'train/grad_norm_pre_clip': 65.79277038574219, 'timing/data': 0.0006418868433684111, 'timing/model': 1.6842351399827749, 'learning_rate/qwen_vl_interface': 8.930880766631444e-07, 'learning_rate/action_model': 8.930880766631445e-06, 'epoch': 1.69}) +[2026-06-23 22:55:32,242][starVLA.training.train_starvla][INFO] - Step 4362, Loss: {'train/loss': 0.08947880566120148, 'train/grad_norm_pre_clip': 23.959585189819336, 'timing/data': 0.0006411161739379168, 'timing/model': 1.6731128301471472, 'learning_rate/qwen_vl_interface': 8.918759066020201e-07, 'learning_rate/action_model': 8.9187590660202e-06, 'epoch': 1.69}) +[2026-06-23 22:55:35,555][starVLA.training.train_starvla][INFO] - Step 4363, Loss: {'train/loss': 0.13823607563972473, 'train/grad_norm_pre_clip': 33.48459243774414, 'timing/data': 0.0006400931160897017, 'timing/model': 1.6850347130093724, 'learning_rate/qwen_vl_interface': 8.906655280010895e-07, 'learning_rate/action_model': 8.906655280010896e-06, 'epoch': 1.69}) +[2026-06-23 22:55:38,949][starVLA.training.train_starvla][INFO] - Step 4364, Loss: {'train/loss': 0.10298673808574677, 'train/grad_norm_pre_clip': 23.84792709350586, 'timing/data': 0.0005641109310090542, 'timing/model': 1.6802459091413766, 'learning_rate/qwen_vl_interface': 8.894569413578938e-07, 'learning_rate/action_model': 8.894569413578937e-06, 'epoch': 1.69}) +[2026-06-23 22:55:42,251][starVLA.training.train_starvla][INFO] - Step 4365, Loss: {'train/loss': 0.09734964370727539, 'train/grad_norm_pre_clip': 32.73111343383789, 'timing/data': 0.0005645907949656248, 'timing/model': 1.6778625778388232, 'learning_rate/qwen_vl_interface': 8.882501471692379e-07, 'learning_rate/action_model': 8.882501471692378e-06, 'epoch': 1.69}) +[2026-06-23 22:55:45,559][starVLA.training.train_starvla][INFO] - Step 4366, Loss: {'train/loss': 0.03905097395181656, 'train/grad_norm_pre_clip': 44.67987823486328, 'timing/data': 0.0005560158751904964, 'timing/model': 1.6792641521897167, 'learning_rate/qwen_vl_interface': 8.870451459311894e-07, 'learning_rate/action_model': 8.870451459311893e-06, 'epoch': 1.69}) +[2026-06-23 22:55:48,867][starVLA.training.train_starvla][INFO] - Step 4367, Loss: {'train/loss': 0.14833296835422516, 'train/grad_norm_pre_clip': 58.246212005615234, 'timing/data': 0.0005657130386680365, 'timing/model': 1.6767366100102663, 'learning_rate/qwen_vl_interface': 8.85841938139077e-07, 'learning_rate/action_model': 8.85841938139077e-06, 'epoch': 1.69}) +[2026-06-23 22:55:52,182][starVLA.training.train_starvla][INFO] - Step 4368, Loss: {'train/loss': 0.056630972772836685, 'train/grad_norm_pre_clip': 15.122937202453613, 'timing/data': 0.0005970720667392015, 'timing/model': 1.6849034121260047, 'learning_rate/qwen_vl_interface': 8.846405242874955e-07, 'learning_rate/action_model': 8.846405242874954e-06, 'epoch': 1.69}) +[2026-06-23 22:55:55,485][starVLA.training.train_starvla][INFO] - Step 4369, Loss: {'train/loss': 0.07374609261751175, 'train/grad_norm_pre_clip': 30.97709846496582, 'timing/data': 0.0005621768068522215, 'timing/model': 1.676407408900559, 'learning_rate/qwen_vl_interface': 8.834409048703004e-07, 'learning_rate/action_model': 8.834409048703005e-06, 'epoch': 1.69}) +[2026-06-23 22:55:58,790][starVLA.training.train_starvla][INFO] - Step 4370, Loss: {'train/loss': 0.1326885223388672, 'train/grad_norm_pre_clip': 25.673255920410156, 'timing/data': 0.0005453049670904875, 'timing/model': 1.6753550600260496, 'learning_rate/qwen_vl_interface': 8.822430803806098e-07, 'learning_rate/action_model': 8.822430803806097e-06, 'epoch': 1.69}) +[2026-06-23 22:56:02,104][starVLA.training.train_starvla][INFO] - Step 4371, Loss: {'train/loss': 0.09259200096130371, 'train/grad_norm_pre_clip': 48.207489013671875, 'timing/data': 0.0005401759408414364, 'timing/model': 1.6871841689571738, 'learning_rate/qwen_vl_interface': 8.810470513108041e-07, 'learning_rate/action_model': 8.81047051310804e-06, 'epoch': 1.69}) +[2026-06-23 22:56:05,413][starVLA.training.train_starvla][INFO] - Step 4372, Loss: {'train/loss': 0.14474213123321533, 'train/grad_norm_pre_clip': 18.072641372680664, 'timing/data': 0.0005811119917780161, 'timing/model': 1.6824701949954033, 'learning_rate/qwen_vl_interface': 8.798528181525262e-07, 'learning_rate/action_model': 8.798528181525261e-06, 'epoch': 1.69}) +[2026-06-23 22:56:08,721][starVLA.training.train_starvla][INFO] - Step 4373, Loss: {'train/loss': 0.16726461052894592, 'train/grad_norm_pre_clip': 33.3615837097168, 'timing/data': 0.0005777059122920036, 'timing/model': 1.6857599900104105, 'learning_rate/qwen_vl_interface': 8.786603813966798e-07, 'learning_rate/action_model': 8.786603813966798e-06, 'epoch': 1.69}) +[2026-06-23 22:56:12,029][starVLA.training.train_starvla][INFO] - Step 4374, Loss: {'train/loss': 0.058170560747385025, 'train/grad_norm_pre_clip': 14.326647758483887, 'timing/data': 0.0005802710074931383, 'timing/model': 1.681786860106513, 'learning_rate/qwen_vl_interface': 8.774697415334304e-07, 'learning_rate/action_model': 8.774697415334304e-06, 'epoch': 1.69}) +[2026-06-23 22:56:15,341][starVLA.training.train_starvla][INFO] - Step 4375, Loss: {'train/loss': 0.16481639444828033, 'train/grad_norm_pre_clip': 36.60916519165039, 'timing/data': 0.0005705929361283779, 'timing/model': 1.6827840839978307, 'learning_rate/qwen_vl_interface': 8.762808990522056e-07, 'learning_rate/action_model': 8.762808990522057e-06, 'epoch': 1.69}) +[2026-06-23 22:56:18,652][starVLA.training.train_starvla][INFO] - Step 4376, Loss: {'train/loss': 0.13972815871238708, 'train/grad_norm_pre_clip': 29.407779693603516, 'timing/data': 0.0005980839487165213, 'timing/model': 1.68351165112108, 'learning_rate/qwen_vl_interface': 8.750938544416934e-07, 'learning_rate/action_model': 8.750938544416934e-06, 'epoch': 1.69}) +[2026-06-23 22:56:21,966][starVLA.training.train_starvla][INFO] - Step 4377, Loss: {'train/loss': 0.05684319883584976, 'train/grad_norm_pre_clip': 20.10645294189453, 'timing/data': 0.0005665351636707783, 'timing/model': 1.6835647199768573, 'learning_rate/qwen_vl_interface': 8.739086081898437e-07, 'learning_rate/action_model': 8.739086081898437e-06, 'epoch': 1.69}) +[2026-06-23 22:56:25,270][starVLA.training.train_starvla][INFO] - Step 4378, Loss: {'train/loss': 0.08604821562767029, 'train/grad_norm_pre_clip': 26.051332473754883, 'timing/data': 0.0005736988969147205, 'timing/model': 1.6771481158211827, 'learning_rate/qwen_vl_interface': 8.727251607838647e-07, 'learning_rate/action_model': 8.727251607838647e-06, 'epoch': 1.69}) +[2026-06-23 22:56:28,736][starVLA.training.train_starvla][INFO] - Step 4379, Loss: {'train/loss': 0.08043140918016434, 'train/grad_norm_pre_clip': 35.72679901123047, 'timing/data': 0.0005525290034711361, 'timing/model': 1.6809943669941276, 'learning_rate/qwen_vl_interface': 8.715435127102281e-07, 'learning_rate/action_model': 8.715435127102281e-06, 'epoch': 1.69}) +[2026-06-23 22:56:32,049][starVLA.training.train_starvla][INFO] - Step 4380, Loss: {'train/loss': 0.08568896353244781, 'train/grad_norm_pre_clip': 62.69400405883789, 'timing/data': 0.0005436618812382221, 'timing/model': 1.6823715148493648, 'learning_rate/qwen_vl_interface': 8.703636644546651e-07, 'learning_rate/action_model': 8.703636644546651e-06, 'epoch': 1.7}) +[2026-06-23 22:56:35,360][starVLA.training.train_starvla][INFO] - Step 4381, Loss: {'train/loss': 0.18444764614105225, 'train/grad_norm_pre_clip': 47.217342376708984, 'timing/data': 0.0005916620139032602, 'timing/model': 1.684384779073298, 'learning_rate/qwen_vl_interface': 8.69185616502166e-07, 'learning_rate/action_model': 8.69185616502166e-06, 'epoch': 1.7}) +[2026-06-23 22:56:38,674][starVLA.training.train_starvla][INFO] - Step 4382, Loss: {'train/loss': 0.06854027509689331, 'train/grad_norm_pre_clip': 18.512590408325195, 'timing/data': 0.000580651918426156, 'timing/model': 1.6814510028343648, 'learning_rate/qwen_vl_interface': 8.680093693369821e-07, 'learning_rate/action_model': 8.68009369336982e-06, 'epoch': 1.7}) +[2026-06-23 22:56:41,984][starVLA.training.train_starvla][INFO] - Step 4383, Loss: {'train/loss': 0.2891640365123749, 'train/grad_norm_pre_clip': 48.970855712890625, 'timing/data': 0.000570863950997591, 'timing/model': 1.6836905130185187, 'learning_rate/qwen_vl_interface': 8.66834923442624e-07, 'learning_rate/action_model': 8.66834923442624e-06, 'epoch': 1.7}) +[2026-06-23 22:56:45,295][starVLA.training.train_starvla][INFO] - Step 4384, Loss: {'train/loss': 0.07839780300855637, 'train/grad_norm_pre_clip': 21.6664981842041, 'timing/data': 0.0005641910247504711, 'timing/model': 1.6800828301347792, 'learning_rate/qwen_vl_interface': 8.65662279301863e-07, 'learning_rate/action_model': 8.656622793018628e-06, 'epoch': 1.7}) +[2026-06-23 22:56:48,606][starVLA.training.train_starvla][INFO] - Step 4385, Loss: {'train/loss': 0.1068180575966835, 'train/grad_norm_pre_clip': 13.552648544311523, 'timing/data': 0.0005729470867663622, 'timing/model': 1.6810508789494634, 'learning_rate/qwen_vl_interface': 8.644914373967274e-07, 'learning_rate/action_model': 8.644914373967273e-06, 'epoch': 1.7}) +[2026-06-23 22:56:51,917][starVLA.training.train_starvla][INFO] - Step 4386, Loss: {'train/loss': 0.10738428682088852, 'train/grad_norm_pre_clip': 28.53254508972168, 'timing/data': 0.0005804209504276514, 'timing/model': 1.6840035000350326, 'learning_rate/qwen_vl_interface': 8.633223982085068e-07, 'learning_rate/action_model': 8.633223982085068e-06, 'epoch': 1.7}) +[2026-06-23 22:56:55,241][starVLA.training.train_starvla][INFO] - Step 4387, Loss: {'train/loss': 0.17478018999099731, 'train/grad_norm_pre_clip': 29.469547271728516, 'timing/data': 0.0005811131559312344, 'timing/model': 1.6883167689666152, 'learning_rate/qwen_vl_interface': 8.621551622177499e-07, 'learning_rate/action_model': 8.621551622177499e-06, 'epoch': 1.7}) +[2026-06-23 22:56:58,541][starVLA.training.train_starvla][INFO] - Step 4388, Loss: {'train/loss': 0.23181916773319244, 'train/grad_norm_pre_clip': 25.487947463989258, 'timing/data': 0.0005773650482296944, 'timing/model': 1.6763645759783685, 'learning_rate/qwen_vl_interface': 8.609897299042615e-07, 'learning_rate/action_model': 8.609897299042614e-06, 'epoch': 1.7}) +[2026-06-23 22:57:01,856][starVLA.training.train_starvla][INFO] - Step 4389, Loss: {'train/loss': 0.09066812694072723, 'train/grad_norm_pre_clip': 24.965173721313477, 'timing/data': 0.0005533911753445864, 'timing/model': 1.682752155000344, 'learning_rate/qwen_vl_interface': 8.598261017471088e-07, 'learning_rate/action_model': 8.598261017471088e-06, 'epoch': 1.7}) +[2026-06-23 22:57:05,172][starVLA.training.train_starvla][INFO] - Step 4390, Loss: {'train/loss': 0.12531036138534546, 'train/grad_norm_pre_clip': 23.174081802368164, 'timing/data': 0.0005342040676623583, 'timing/model': 1.683958473149687, 'learning_rate/qwen_vl_interface': 8.586642782246133e-07, 'learning_rate/action_model': 8.586642782246131e-06, 'epoch': 1.7}) +[2026-06-23 22:57:08,479][starVLA.training.train_starvla][INFO] - Step 4391, Loss: {'train/loss': 0.16914847493171692, 'train/grad_norm_pre_clip': 30.938753128051758, 'timing/data': 0.0005765540990978479, 'timing/model': 1.6833787530194968, 'learning_rate/qwen_vl_interface': 8.57504259814358e-07, 'learning_rate/action_model': 8.57504259814358e-06, 'epoch': 1.7}) +[2026-06-23 22:57:11,819][starVLA.training.train_starvla][INFO] - Step 4392, Loss: {'train/loss': 0.09087924659252167, 'train/grad_norm_pre_clip': 14.209284782409668, 'timing/data': 0.000578136183321476, 'timing/model': 1.6795626569073647, 'learning_rate/qwen_vl_interface': 8.563460469931823e-07, 'learning_rate/action_model': 8.563460469931822e-06, 'epoch': 1.7}) +[2026-06-23 22:57:15,260][starVLA.training.train_starvla][INFO] - Step 4393, Loss: {'train/loss': 0.09308458119630814, 'train/grad_norm_pre_clip': 31.969497680664062, 'timing/data': 0.0005767240654677153, 'timing/model': 1.8035312378779054, 'learning_rate/qwen_vl_interface': 8.551896402371834e-07, 'learning_rate/action_model': 8.551896402371834e-06, 'epoch': 1.7}) +[2026-06-23 22:57:18,574][starVLA.training.train_starvla][INFO] - Step 4394, Loss: {'train/loss': 0.12961958348751068, 'train/grad_norm_pre_clip': 21.754898071289062, 'timing/data': 0.0005518780089914799, 'timing/model': 1.684942580992356, 'learning_rate/qwen_vl_interface': 8.540350400217177e-07, 'learning_rate/action_model': 8.540350400217176e-06, 'epoch': 1.7}) +[2026-06-23 22:57:21,882][starVLA.training.train_starvla][INFO] - Step 4395, Loss: {'train/loss': 0.09036785364151001, 'train/grad_norm_pre_clip': 15.99201488494873, 'timing/data': 0.0005759429186582565, 'timing/model': 1.680394349154085, 'learning_rate/qwen_vl_interface': 8.528822468213964e-07, 'learning_rate/action_model': 8.528822468213964e-06, 'epoch': 1.7}) +[2026-06-23 22:57:25,199][starVLA.training.train_starvla][INFO] - Step 4396, Loss: {'train/loss': 0.0910106897354126, 'train/grad_norm_pre_clip': 29.826351165771484, 'timing/data': 0.0006420069839805365, 'timing/model': 1.6837294690776616, 'learning_rate/qwen_vl_interface': 8.517312611100898e-07, 'learning_rate/action_model': 8.517312611100897e-06, 'epoch': 1.7}) +[2026-06-23 22:57:28,510][starVLA.training.train_starvla][INFO] - Step 4397, Loss: {'train/loss': 0.2257625013589859, 'train/grad_norm_pre_clip': 23.332382202148438, 'timing/data': 0.0005557551048696041, 'timing/model': 1.681082131108269, 'learning_rate/qwen_vl_interface': 8.505820833609249e-07, 'learning_rate/action_model': 8.50582083360925e-06, 'epoch': 1.7}) +[2026-06-23 22:57:31,824][starVLA.training.train_starvla][INFO] - Step 4398, Loss: {'train/loss': 0.14740200340747833, 'train/grad_norm_pre_clip': 43.54542922973633, 'timing/data': 0.0005392329767346382, 'timing/model': 1.684238474816084, 'learning_rate/qwen_vl_interface': 8.494347140462847e-07, 'learning_rate/action_model': 8.494347140462846e-06, 'epoch': 1.7}) +[2026-06-23 22:57:35,138][starVLA.training.train_starvla][INFO] - Step 4399, Loss: {'train/loss': 0.030531663447618484, 'train/grad_norm_pre_clip': 16.576461791992188, 'timing/data': 0.0005740090273320675, 'timing/model': 1.6845642270054668, 'learning_rate/qwen_vl_interface': 8.482891536378105e-07, 'learning_rate/action_model': 8.482891536378105e-06, 'epoch': 1.7}) +[2026-06-23 22:57:38,448][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4400 +[2026-06-23 22:59:11,683][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4400 +[2026-06-23 22:59:11,684][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4400 +[2026-06-23 23:00:22,186][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4400 +[2026-06-23 23:00:22,187][starVLA.training.train_starvla][INFO] - Step 4400, Loss: {'train/loss': 0.08119608461856842, 'train/grad_norm_pre_clip': 22.949235916137695, 'eval/action_loss/seconds': 93.2332350590732, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06561631709337234, 'eval/latency_2/loss': 0.06561631709337234, 'eval/flappy/loss': 0.06561631709337234, 'eval/flappy/latency_2/loss': 0.06561631709337234, 'eval/flap_precision': 0.9441797140912185, 'eval/flap_recall': 0.9358974358974359, 'eval/flap_f1': 0.9400203320908167, 'eval/flap_rate_model': 0.20402777777777778, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9400203320908168, 'eval/latency_2/flap_precision': 0.9441797140912185, 'eval/latency_2/flap_recall': 0.9358974358974359, 'eval/latency_2/flap_f1': 0.9400203320908167, 'eval/latency_2/flap_rate_model': 0.20402777777777778, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9400203320908168, 'eval/action_classification/data_wait_seconds': 0.8942083120346069, 'eval/action_classification/predict_seconds': 69.4806137084961, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005971919745206833, 'timing/model': 1.6812662861775607, 'learning_rate/qwen_vl_interface': 8.471454026063972e-07, 'learning_rate/action_model': 8.47145402606397e-06, 'epoch': 1.7}) +[2026-06-23 23:00:25,631][starVLA.training.train_starvla][INFO] - Step 4401, Loss: {'train/loss': 0.06359616667032242, 'train/grad_norm_pre_clip': 26.49436378479004, 'timing/data': 0.0005613111425191164, 'timing/model': 1.8131069629453123, 'learning_rate/qwen_vl_interface': 8.460034614221988e-07, 'learning_rate/action_model': 8.460034614221988e-06, 'epoch': 1.7}) +[2026-06-23 23:00:28,941][starVLA.training.train_starvla][INFO] - Step 4402, Loss: {'train/loss': 0.16076576709747314, 'train/grad_norm_pre_clip': 24.4417781829834, 'timing/data': 0.0006017261184751987, 'timing/model': 1.681734343059361, 'learning_rate/qwen_vl_interface': 8.448633305546236e-07, 'learning_rate/action_model': 8.448633305546235e-06, 'epoch': 1.7}) +[2026-06-23 23:00:32,255][starVLA.training.train_starvla][INFO] - Step 4403, Loss: {'train/loss': 0.1266775131225586, 'train/grad_norm_pre_clip': 42.18434143066406, 'timing/data': 0.000575728015974164, 'timing/model': 1.6837363571394235, 'learning_rate/qwen_vl_interface': 8.437250104723369e-07, 'learning_rate/action_model': 8.437250104723369e-06, 'epoch': 1.7}) +[2026-06-23 23:00:35,567][starVLA.training.train_starvla][INFO] - Step 4404, Loss: {'train/loss': 0.2719173729419708, 'train/grad_norm_pre_clip': 34.115692138671875, 'timing/data': 0.0005897649098187685, 'timing/model': 1.6812475030310452, 'learning_rate/qwen_vl_interface': 8.42588501643259e-07, 'learning_rate/action_model': 8.42588501643259e-06, 'epoch': 1.7}) +[2026-06-23 23:00:38,878][starVLA.training.train_starvla][INFO] - Step 4405, Loss: {'train/loss': 0.02894878014922142, 'train/grad_norm_pre_clip': 16.829315185546875, 'timing/data': 0.0005695861764252186, 'timing/model': 1.6805795149412006, 'learning_rate/qwen_vl_interface': 8.414538045345655e-07, 'learning_rate/action_model': 8.414538045345654e-06, 'epoch': 1.7}) +[2026-06-23 23:00:42,188][starVLA.training.train_starvla][INFO] - Step 4406, Loss: {'train/loss': 0.08373058587312698, 'train/grad_norm_pre_clip': 20.48309326171875, 'timing/data': 0.0005657400470227003, 'timing/model': 1.6822146100457758, 'learning_rate/qwen_vl_interface': 8.403209196126871e-07, 'learning_rate/action_model': 8.403209196126871e-06, 'epoch': 1.71}) +[2026-06-23 23:00:45,498][starVLA.training.train_starvla][INFO] - Step 4407, Loss: {'train/loss': 0.11253312975168228, 'train/grad_norm_pre_clip': 41.259090423583984, 'timing/data': 0.0005747759714722633, 'timing/model': 1.6826211151201278, 'learning_rate/qwen_vl_interface': 8.391898473433111e-07, 'learning_rate/action_model': 8.39189847343311e-06, 'epoch': 1.71}) +[2026-06-23 23:00:48,811][starVLA.training.train_starvla][INFO] - Step 4408, Loss: {'train/loss': 0.14770938456058502, 'train/grad_norm_pre_clip': 24.458398818969727, 'timing/data': 0.0005586161278188229, 'timing/model': 1.6831863760016859, 'learning_rate/qwen_vl_interface': 8.380605881913778e-07, 'learning_rate/action_model': 8.380605881913779e-06, 'epoch': 1.71}) +[2026-06-23 23:00:52,123][starVLA.training.train_starvla][INFO] - Step 4409, Loss: {'train/loss': 0.03649308905005455, 'train/grad_norm_pre_clip': 30.486892700195312, 'timing/data': 0.0005714599974453449, 'timing/model': 1.6834553701337427, 'learning_rate/qwen_vl_interface': 8.369331426210832e-07, 'learning_rate/action_model': 8.369331426210832e-06, 'epoch': 1.71}) +[2026-06-23 23:00:55,435][starVLA.training.train_starvla][INFO] - Step 4410, Loss: {'train/loss': 0.022743768990039825, 'train/grad_norm_pre_clip': 7.898670673370361, 'timing/data': 0.0005510118789970875, 'timing/model': 1.6827121910173446, 'learning_rate/qwen_vl_interface': 8.358075110958785e-07, 'learning_rate/action_model': 8.358075110958784e-06, 'epoch': 1.71}) +[2026-06-23 23:00:58,744][starVLA.training.train_starvla][INFO] - Step 4411, Loss: {'train/loss': 0.09684618562459946, 'train/grad_norm_pre_clip': 20.31641960144043, 'timing/data': 0.0005751370918005705, 'timing/model': 1.6810196531005204, 'learning_rate/qwen_vl_interface': 8.346836940784661e-07, 'learning_rate/action_model': 8.346836940784661e-06, 'epoch': 1.71}) +[2026-06-23 23:01:02,058][starVLA.training.train_starvla][INFO] - Step 4412, Loss: {'train/loss': 0.10905253887176514, 'train/grad_norm_pre_clip': 33.37965774536133, 'timing/data': 0.0005353230517357588, 'timing/model': 1.6842683660797775, 'learning_rate/qwen_vl_interface': 8.335616920308061e-07, 'learning_rate/action_model': 8.33561692030806e-06, 'epoch': 1.71}) +[2026-06-23 23:01:05,378][starVLA.training.train_starvla][INFO] - Step 4413, Loss: {'train/loss': 0.2691546082496643, 'train/grad_norm_pre_clip': 35.42692947387695, 'timing/data': 0.000578031875193119, 'timing/model': 1.6829954939894378, 'learning_rate/qwen_vl_interface': 8.324415054141111e-07, 'learning_rate/action_model': 8.32441505414111e-06, 'epoch': 1.71}) +[2026-06-23 23:01:08,686][starVLA.training.train_starvla][INFO] - Step 4414, Loss: {'train/loss': 0.12482117861509323, 'train/grad_norm_pre_clip': 30.128894805908203, 'timing/data': 0.0005737349856644869, 'timing/model': 1.6840185809414834, 'learning_rate/qwen_vl_interface': 8.313231346888476e-07, 'learning_rate/action_model': 8.313231346888476e-06, 'epoch': 1.71}) +[2026-06-23 23:01:12,000][starVLA.training.train_starvla][INFO] - Step 4415, Loss: {'train/loss': 0.140733003616333, 'train/grad_norm_pre_clip': 23.5682373046875, 'timing/data': 0.0006322639528661966, 'timing/model': 1.6838219100609422, 'learning_rate/qwen_vl_interface': 8.302065803147353e-07, 'learning_rate/action_model': 8.302065803147353e-06, 'epoch': 1.71}) +[2026-06-23 23:01:15,412][starVLA.training.train_starvla][INFO] - Step 4416, Loss: {'train/loss': 0.03239259868860245, 'train/grad_norm_pre_clip': 21.400724411010742, 'timing/data': 0.0005790239665657282, 'timing/model': 1.7816092809662223, 'learning_rate/qwen_vl_interface': 8.29091842750748e-07, 'learning_rate/action_model': 8.29091842750748e-06, 'epoch': 1.71}) +[2026-06-23 23:01:18,725][starVLA.training.train_starvla][INFO] - Step 4417, Loss: {'train/loss': 0.1758977770805359, 'train/grad_norm_pre_clip': 23.32684898376465, 'timing/data': 0.0005494691431522369, 'timing/model': 1.682750194100663, 'learning_rate/qwen_vl_interface': 8.279789224551116e-07, 'learning_rate/action_model': 8.279789224551116e-06, 'epoch': 1.71}) +[2026-06-23 23:01:22,036][starVLA.training.train_starvla][INFO] - Step 4418, Loss: {'train/loss': 0.07135733962059021, 'train/grad_norm_pre_clip': 13.59604549407959, 'timing/data': 0.0005883621051907539, 'timing/model': 1.6819712270516902, 'learning_rate/qwen_vl_interface': 8.268678198853058e-07, 'learning_rate/action_model': 8.268678198853059e-06, 'epoch': 1.71}) +[2026-06-23 23:01:25,345][starVLA.training.train_starvla][INFO] - Step 4419, Loss: {'train/loss': 0.1706390082836151, 'train/grad_norm_pre_clip': 19.456789016723633, 'timing/data': 0.0006311419419944286, 'timing/model': 1.6826941941399127, 'learning_rate/qwen_vl_interface': 8.257585354980628e-07, 'learning_rate/action_model': 8.257585354980628e-06, 'epoch': 1.71}) +[2026-06-23 23:01:28,659][starVLA.training.train_starvla][INFO] - Step 4420, Loss: {'train/loss': 0.059335943311452866, 'train/grad_norm_pre_clip': 18.642431259155273, 'timing/data': 0.0005675030406564474, 'timing/model': 1.685535574099049, 'learning_rate/qwen_vl_interface': 8.246510697493681e-07, 'learning_rate/action_model': 8.246510697493682e-06, 'epoch': 1.71}) +[2026-06-23 23:01:31,967][starVLA.training.train_starvla][INFO] - Step 4421, Loss: {'train/loss': 0.1183559000492096, 'train/grad_norm_pre_clip': 22.140518188476562, 'timing/data': 0.0005507611203938723, 'timing/model': 1.6832831799983978, 'learning_rate/qwen_vl_interface': 8.235454230944584e-07, 'learning_rate/action_model': 8.235454230944584e-06, 'epoch': 1.71}) +[2026-06-23 23:01:35,276][starVLA.training.train_starvla][INFO] - Step 4422, Loss: {'train/loss': 0.07126664370298386, 'train/grad_norm_pre_clip': 27.01522445678711, 'timing/data': 0.0006015070248395205, 'timing/model': 1.6782396698836237, 'learning_rate/qwen_vl_interface': 8.224415959878234e-07, 'learning_rate/action_model': 8.224415959878234e-06, 'epoch': 1.71}) +[2026-06-23 23:01:38,591][starVLA.training.train_starvla][INFO] - Step 4423, Loss: {'train/loss': 0.15996553003787994, 'train/grad_norm_pre_clip': 19.674766540527344, 'timing/data': 0.000580166932195425, 'timing/model': 1.6845788210630417, 'learning_rate/qwen_vl_interface': 8.213395888832052e-07, 'learning_rate/action_model': 8.213395888832051e-06, 'epoch': 1.71}) +[2026-06-23 23:01:41,910][starVLA.training.train_starvla][INFO] - Step 4424, Loss: {'train/loss': 0.05721050128340721, 'train/grad_norm_pre_clip': 13.928248405456543, 'timing/data': 0.0005756979808211327, 'timing/model': 1.6855278729926795, 'learning_rate/qwen_vl_interface': 8.202394022335958e-07, 'learning_rate/action_model': 8.202394022335957e-06, 'epoch': 1.71}) +[2026-06-23 23:01:45,221][starVLA.training.train_starvla][INFO] - Step 4425, Loss: {'train/loss': 0.12213806807994843, 'train/grad_norm_pre_clip': 23.146522521972656, 'timing/data': 0.0005819799844175577, 'timing/model': 1.6853989560622722, 'learning_rate/qwen_vl_interface': 8.191410364912424e-07, 'learning_rate/action_model': 8.191410364912424e-06, 'epoch': 1.71}) +[2026-06-23 23:01:48,536][starVLA.training.train_starvla][INFO] - Step 4426, Loss: {'train/loss': 0.2047208696603775, 'train/grad_norm_pre_clip': 17.04965591430664, 'timing/data': 0.0005787238478660583, 'timing/model': 1.6851670891046524, 'learning_rate/qwen_vl_interface': 8.1804449210764e-07, 'learning_rate/action_model': 8.1804449210764e-06, 'epoch': 1.71}) +[2026-06-23 23:01:51,848][starVLA.training.train_starvla][INFO] - Step 4427, Loss: {'train/loss': 0.10203566402196884, 'train/grad_norm_pre_clip': 46.244110107421875, 'timing/data': 0.000563295092433691, 'timing/model': 1.6817853769753128, 'learning_rate/qwen_vl_interface': 8.169497695335377e-07, 'learning_rate/action_model': 8.169497695335376e-06, 'epoch': 1.71}) +[2026-06-23 23:01:55,164][starVLA.training.train_starvla][INFO] - Step 4428, Loss: {'train/loss': 0.24632059037685394, 'train/grad_norm_pre_clip': 26.29897689819336, 'timing/data': 0.0005590571090579033, 'timing/model': 1.6850992180407047, 'learning_rate/qwen_vl_interface': 8.158568692189342e-07, 'learning_rate/action_model': 8.158568692189342e-06, 'epoch': 1.71}) +[2026-06-23 23:01:58,483][starVLA.training.train_starvla][INFO] - Step 4429, Loss: {'train/loss': 0.11061617732048035, 'train/grad_norm_pre_clip': 15.482904434204102, 'timing/data': 0.0005790549330413342, 'timing/model': 1.6874275340232998, 'learning_rate/qwen_vl_interface': 8.147657916130797e-07, 'learning_rate/action_model': 8.147657916130796e-06, 'epoch': 1.71}) +[2026-06-23 23:02:01,799][starVLA.training.train_starvla][INFO] - Step 4430, Loss: {'train/loss': 0.20759332180023193, 'train/grad_norm_pre_clip': 61.080657958984375, 'timing/data': 0.0005885129794478416, 'timing/model': 1.6849600351415575, 'learning_rate/qwen_vl_interface': 8.136765371644753e-07, 'learning_rate/action_model': 8.136765371644753e-06, 'epoch': 1.71}) +[2026-06-23 23:02:05,210][starVLA.training.train_starvla][INFO] - Step 4431, Loss: {'train/loss': 0.052591484040021896, 'train/grad_norm_pre_clip': 14.940834045410156, 'timing/data': 0.0005769911222159863, 'timing/model': 1.683095765998587, 'learning_rate/qwen_vl_interface': 8.125891063208713e-07, 'learning_rate/action_model': 8.125891063208712e-06, 'epoch': 1.71}) +[2026-06-23 23:02:08,529][starVLA.training.train_starvla][INFO] - Step 4432, Loss: {'train/loss': 0.12230734527111053, 'train/grad_norm_pre_clip': 13.743363380432129, 'timing/data': 0.0005687549710273743, 'timing/model': 1.684974017785862, 'learning_rate/qwen_vl_interface': 8.115034995292698e-07, 'learning_rate/action_model': 8.115034995292698e-06, 'epoch': 1.72}) +[2026-06-23 23:02:11,842][starVLA.training.train_starvla][INFO] - Step 4433, Loss: {'train/loss': 0.19342179596424103, 'train/grad_norm_pre_clip': 25.207599639892578, 'timing/data': 0.0005871700122952461, 'timing/model': 1.6826343561988324, 'learning_rate/qwen_vl_interface': 8.104197172359235e-07, 'learning_rate/action_model': 8.104197172359234e-06, 'epoch': 1.72}) +[2026-06-23 23:02:15,156][starVLA.training.train_starvla][INFO] - Step 4434, Loss: {'train/loss': 0.057676319032907486, 'train/grad_norm_pre_clip': 18.553422927856445, 'timing/data': 0.0005977100227028131, 'timing/model': 1.6803807648830116, 'learning_rate/qwen_vl_interface': 8.093377598863334e-07, 'learning_rate/action_model': 8.093377598863334e-06, 'epoch': 1.72}) +[2026-06-23 23:02:18,467][starVLA.training.train_starvla][INFO] - Step 4435, Loss: {'train/loss': 0.14517293870449066, 'train/grad_norm_pre_clip': 50.26926040649414, 'timing/data': 0.0005666320212185383, 'timing/model': 1.6832906880881637, 'learning_rate/qwen_vl_interface': 8.082576279252519e-07, 'learning_rate/action_model': 8.082576279252518e-06, 'epoch': 1.72}) +[2026-06-23 23:02:21,783][starVLA.training.train_starvla][INFO] - Step 4436, Loss: {'train/loss': 0.09719619899988174, 'train/grad_norm_pre_clip': 18.212696075439453, 'timing/data': 0.000566571019589901, 'timing/model': 1.682427341118455, 'learning_rate/qwen_vl_interface': 8.071793217966803e-07, 'learning_rate/action_model': 8.071793217966802e-06, 'epoch': 1.72}) +[2026-06-23 23:02:25,099][starVLA.training.train_starvla][INFO] - Step 4437, Loss: {'train/loss': 0.12615342438220978, 'train/grad_norm_pre_clip': 21.814516067504883, 'timing/data': 0.0005699777975678444, 'timing/model': 1.6849145300220698, 'learning_rate/qwen_vl_interface': 8.061028419438691e-07, 'learning_rate/action_model': 8.06102841943869e-06, 'epoch': 1.72}) +[2026-06-23 23:02:28,422][starVLA.training.train_starvla][INFO] - Step 4438, Loss: {'train/loss': 0.06706546992063522, 'train/grad_norm_pre_clip': 32.78997802734375, 'timing/data': 0.0005731238052248955, 'timing/model': 1.6879355318378657, 'learning_rate/qwen_vl_interface': 8.050281888093193e-07, 'learning_rate/action_model': 8.050281888093192e-06, 'epoch': 1.72}) +[2026-06-23 23:02:31,739][starVLA.training.train_starvla][INFO] - Step 4439, Loss: {'train/loss': 0.09792225062847137, 'train/grad_norm_pre_clip': 22.015783309936523, 'timing/data': 0.0005700381007045507, 'timing/model': 1.6879406219813973, 'learning_rate/qwen_vl_interface': 8.039553628347794e-07, 'learning_rate/action_model': 8.039553628347793e-06, 'epoch': 1.72}) +[2026-06-23 23:02:35,054][starVLA.training.train_starvla][INFO] - Step 4440, Loss: {'train/loss': 0.07556931674480438, 'train/grad_norm_pre_clip': 26.905662536621094, 'timing/data': 0.0005436490755528212, 'timing/model': 1.683708469150588, 'learning_rate/qwen_vl_interface': 8.028843644612483e-07, 'learning_rate/action_model': 8.028843644612482e-06, 'epoch': 1.72}) +[2026-06-23 23:02:38,373][starVLA.training.train_starvla][INFO] - Step 4441, Loss: {'train/loss': 0.156263068318367, 'train/grad_norm_pre_clip': 22.570022583007812, 'timing/data': 0.0005486779846251011, 'timing/model': 1.6868446150328964, 'learning_rate/qwen_vl_interface': 8.01815194128973e-07, 'learning_rate/action_model': 8.01815194128973e-06, 'epoch': 1.72}) +[2026-06-23 23:02:41,697][starVLA.training.train_starvla][INFO] - Step 4442, Loss: {'train/loss': 0.08288519829511642, 'train/grad_norm_pre_clip': 67.7348861694336, 'timing/data': 0.0007707139011472464, 'timing/model': 1.692382091190666, 'learning_rate/qwen_vl_interface': 8.007478522774481e-07, 'learning_rate/action_model': 8.00747852277448e-06, 'epoch': 1.72}) +[2026-06-23 23:02:45,018][starVLA.training.train_starvla][INFO] - Step 4443, Loss: {'train/loss': 0.05882255733013153, 'train/grad_norm_pre_clip': 35.73057556152344, 'timing/data': 0.0005863180849701166, 'timing/model': 1.6860751789063215, 'learning_rate/qwen_vl_interface': 7.996823393454197e-07, 'learning_rate/action_model': 7.996823393454196e-06, 'epoch': 1.72}) +[2026-06-23 23:02:48,338][starVLA.training.train_starvla][INFO] - Step 4444, Loss: {'train/loss': 0.1420522779226303, 'train/grad_norm_pre_clip': 17.035789489746094, 'timing/data': 0.000577092170715332, 'timing/model': 1.6841161069460213, 'learning_rate/qwen_vl_interface': 7.986186557708773e-07, 'learning_rate/action_model': 7.986186557708771e-06, 'epoch': 1.72}) +[2026-06-23 23:02:51,663][starVLA.training.train_starvla][INFO] - Step 4445, Loss: {'train/loss': 0.09171196818351746, 'train/grad_norm_pre_clip': 27.540761947631836, 'timing/data': 0.0005788749549537897, 'timing/model': 1.689146411139518, 'learning_rate/qwen_vl_interface': 7.975568019910617e-07, 'learning_rate/action_model': 7.975568019910617e-06, 'epoch': 1.72}) +[2026-06-23 23:02:55,090][starVLA.training.train_starvla][INFO] - Step 4446, Loss: {'train/loss': 0.111016646027565, 'train/grad_norm_pre_clip': 46.78312683105469, 'timing/data': 0.000560790067538619, 'timing/model': 1.6887802421115339, 'learning_rate/qwen_vl_interface': 7.964967784424618e-07, 'learning_rate/action_model': 7.964967784424618e-06, 'epoch': 1.72}) +[2026-06-23 23:02:58,411][starVLA.training.train_starvla][INFO] - Step 4447, Loss: {'train/loss': 0.17660841345787048, 'train/grad_norm_pre_clip': 45.091331481933594, 'timing/data': 0.0005922699347138405, 'timing/model': 1.687887285836041, 'learning_rate/qwen_vl_interface': 7.954385855608129e-07, 'learning_rate/action_model': 7.95438585560813e-06, 'epoch': 1.72}) +[2026-06-23 23:03:01,725][starVLA.training.train_starvla][INFO] - Step 4448, Loss: {'train/loss': 0.16171099245548248, 'train/grad_norm_pre_clip': 32.52357482910156, 'timing/data': 0.0005477459635585546, 'timing/model': 1.6830952330492437, 'learning_rate/qwen_vl_interface': 7.943822237810974e-07, 'learning_rate/action_model': 7.943822237810975e-06, 'epoch': 1.72}) +[2026-06-23 23:03:05,047][starVLA.training.train_starvla][INFO] - Step 4449, Loss: {'train/loss': 0.15590913593769073, 'train/grad_norm_pre_clip': 51.52560806274414, 'timing/data': 0.00061807781457901, 'timing/model': 1.6876764120534062, 'learning_rate/qwen_vl_interface': 7.93327693537547e-07, 'learning_rate/action_model': 7.93327693537547e-06, 'epoch': 1.72}) +[2026-06-23 23:03:08,364][starVLA.training.train_starvla][INFO] - Step 4450, Loss: {'train/loss': 0.06359042972326279, 'train/grad_norm_pre_clip': 13.692846298217773, 'timing/data': 0.0005487781018018723, 'timing/model': 1.6841140917968005, 'learning_rate/qwen_vl_interface': 7.92274995263638e-07, 'learning_rate/action_model': 7.92274995263638e-06, 'epoch': 1.72}) +[2026-06-23 23:03:11,688][starVLA.training.train_starvla][INFO] - Step 4451, Loss: {'train/loss': 0.057579342275857925, 'train/grad_norm_pre_clip': 20.486495971679688, 'timing/data': 0.0006630721036344767, 'timing/model': 1.6912899371236563, 'learning_rate/qwen_vl_interface': 7.912241293920958e-07, 'learning_rate/action_model': 7.912241293920959e-06, 'epoch': 1.72}) +[2026-06-23 23:03:15,003][starVLA.training.train_starvla][INFO] - Step 4452, Loss: {'train/loss': 0.09002377837896347, 'train/grad_norm_pre_clip': 25.321456909179688, 'timing/data': 0.0005860079545527697, 'timing/model': 1.6839368790388107, 'learning_rate/qwen_vl_interface': 7.90175096354891e-07, 'learning_rate/action_model': 7.90175096354891e-06, 'epoch': 1.72}) +[2026-06-23 23:03:18,323][starVLA.training.train_starvla][INFO] - Step 4453, Loss: {'train/loss': 0.0729767456650734, 'train/grad_norm_pre_clip': 45.2230224609375, 'timing/data': 0.0005678541492670774, 'timing/model': 1.6822601018939167, 'learning_rate/qwen_vl_interface': 7.891278965832431e-07, 'learning_rate/action_model': 7.891278965832432e-06, 'epoch': 1.72}) +[2026-06-23 23:03:21,644][starVLA.training.train_starvla][INFO] - Step 4454, Loss: {'train/loss': 0.07540664076805115, 'train/grad_norm_pre_clip': 16.457416534423828, 'timing/data': 0.000587069895118475, 'timing/model': 1.6827418189495802, 'learning_rate/qwen_vl_interface': 7.880825305076138e-07, 'learning_rate/action_model': 7.880825305076139e-06, 'epoch': 1.72}) +[2026-06-23 23:03:24,956][starVLA.training.train_starvla][INFO] - Step 4455, Loss: {'train/loss': 0.08566173911094666, 'train/grad_norm_pre_clip': 25.022418975830078, 'timing/data': 0.0005698779132217169, 'timing/model': 1.6819852739572525, 'learning_rate/qwen_vl_interface': 7.870389985577148e-07, 'learning_rate/action_model': 7.870389985577148e-06, 'epoch': 1.72}) +[2026-06-23 23:03:28,270][starVLA.training.train_starvla][INFO] - Step 4456, Loss: {'train/loss': 0.033541347831487656, 'train/grad_norm_pre_clip': 35.09712219238281, 'timing/data': 0.0005442698020488024, 'timing/model': 1.6821177159436047, 'learning_rate/qwen_vl_interface': 7.859973011625028e-07, 'learning_rate/action_model': 7.859973011625026e-06, 'epoch': 1.72}) +[2026-06-23 23:03:31,595][starVLA.training.train_starvla][INFO] - Step 4457, Loss: {'train/loss': 0.16108688712120056, 'train/grad_norm_pre_clip': 20.867372512817383, 'timing/data': 0.000548758078366518, 'timing/model': 1.6895546279847622, 'learning_rate/qwen_vl_interface': 7.849574387501801e-07, 'learning_rate/action_model': 7.8495743875018e-06, 'epoch': 1.72}) +[2026-06-23 23:03:34,918][starVLA.training.train_starvla][INFO] - Step 4458, Loss: {'train/loss': 0.09079276025295258, 'train/grad_norm_pre_clip': 14.8768310546875, 'timing/data': 0.0005964178126305342, 'timing/model': 1.6924288168083876, 'learning_rate/qwen_vl_interface': 7.839194117481952e-07, 'learning_rate/action_model': 7.839194117481952e-06, 'epoch': 1.73}) +[2026-06-23 23:03:38,229][starVLA.training.train_starvla][INFO] - Step 4459, Loss: {'train/loss': 0.04252816364169121, 'train/grad_norm_pre_clip': 10.151355743408203, 'timing/data': 0.0005516938399523497, 'timing/model': 1.680635335855186, 'learning_rate/qwen_vl_interface': 7.828832205832418e-07, 'learning_rate/action_model': 7.828832205832418e-06, 'epoch': 1.73}) +[2026-06-23 23:03:41,642][starVLA.training.train_starvla][INFO] - Step 4460, Loss: {'train/loss': 0.05787504464387894, 'train/grad_norm_pre_clip': 17.600297927856445, 'timing/data': 0.000642373925074935, 'timing/model': 1.7819126721005887, 'learning_rate/qwen_vl_interface': 7.818488656812581e-07, 'learning_rate/action_model': 7.818488656812582e-06, 'epoch': 1.73}) +[2026-06-23 23:03:44,956][starVLA.training.train_starvla][INFO] - Step 4461, Loss: {'train/loss': 0.11680777370929718, 'train/grad_norm_pre_clip': 24.189546585083008, 'timing/data': 0.0005811690352857113, 'timing/model': 1.681379133136943, 'learning_rate/qwen_vl_interface': 7.808163474674291e-07, 'learning_rate/action_model': 7.80816347467429e-06, 'epoch': 1.73}) +[2026-06-23 23:03:48,273][starVLA.training.train_starvla][INFO] - Step 4462, Loss: {'train/loss': 0.10956131666898727, 'train/grad_norm_pre_clip': 87.57572937011719, 'timing/data': 0.0005826619453728199, 'timing/model': 1.6859762219246477, 'learning_rate/qwen_vl_interface': 7.797856663661836e-07, 'learning_rate/action_model': 7.797856663661835e-06, 'epoch': 1.73}) +[2026-06-23 23:03:51,590][starVLA.training.train_starvla][INFO] - Step 4463, Loss: {'train/loss': 0.13415023684501648, 'train/grad_norm_pre_clip': 29.499038696289062, 'timing/data': 0.0005730940029025078, 'timing/model': 1.684068859089166, 'learning_rate/qwen_vl_interface': 7.787568228011959e-07, 'learning_rate/action_model': 7.78756822801196e-06, 'epoch': 1.73}) +[2026-06-23 23:03:54,909][starVLA.training.train_starvla][INFO] - Step 4464, Loss: {'train/loss': 0.09944853186607361, 'train/grad_norm_pre_clip': 16.228071212768555, 'timing/data': 0.0005633160471916199, 'timing/model': 1.6849062179680914, 'learning_rate/qwen_vl_interface': 7.777298171953858e-07, 'learning_rate/action_model': 7.777298171953857e-06, 'epoch': 1.73}) +[2026-06-23 23:03:58,226][starVLA.training.train_starvla][INFO] - Step 4465, Loss: {'train/loss': 0.11696270108222961, 'train/grad_norm_pre_clip': 22.362993240356445, 'timing/data': 0.0006177080795168877, 'timing/model': 1.6833709399215877, 'learning_rate/qwen_vl_interface': 7.767046499709144e-07, 'learning_rate/action_model': 7.767046499709143e-06, 'epoch': 1.73}) +[2026-06-23 23:04:01,554][starVLA.training.train_starvla][INFO] - Step 4466, Loss: {'train/loss': 0.13825303316116333, 'train/grad_norm_pre_clip': 18.824466705322266, 'timing/data': 0.0005774530582129955, 'timing/model': 1.6903557241894305, 'learning_rate/qwen_vl_interface': 7.756813215491904e-07, 'learning_rate/action_model': 7.756813215491904e-06, 'epoch': 1.73}) +[2026-06-23 23:04:04,875][starVLA.training.train_starvla][INFO] - Step 4467, Loss: {'train/loss': 0.11419831216335297, 'train/grad_norm_pre_clip': 33.746917724609375, 'timing/data': 0.0005580461584031582, 'timing/model': 1.691085611935705, 'learning_rate/qwen_vl_interface': 7.746598323508655e-07, 'learning_rate/action_model': 7.746598323508655e-06, 'epoch': 1.73}) +[2026-06-23 23:04:08,194][starVLA.training.train_starvla][INFO] - Step 4468, Loss: {'train/loss': 0.167552188038826, 'train/grad_norm_pre_clip': 40.826927185058594, 'timing/data': 0.000554319005459547, 'timing/model': 1.6862477120012045, 'learning_rate/qwen_vl_interface': 7.736401827958351e-07, 'learning_rate/action_model': 7.73640182795835e-06, 'epoch': 1.73}) +[2026-06-23 23:04:11,516][starVLA.training.train_starvla][INFO] - Step 4469, Loss: {'train/loss': 0.11831274628639221, 'train/grad_norm_pre_clip': 13.075369834899902, 'timing/data': 0.0005616829730570316, 'timing/model': 1.6857919569592923, 'learning_rate/qwen_vl_interface': 7.726223733032382e-07, 'learning_rate/action_model': 7.72622373303238e-06, 'epoch': 1.73}) +[2026-06-23 23:04:14,842][starVLA.training.train_starvla][INFO] - Step 4470, Loss: {'train/loss': 0.11304090172052383, 'train/grad_norm_pre_clip': 21.320499420166016, 'timing/data': 0.0005845960695296526, 'timing/model': 1.6910684879403561, 'learning_rate/qwen_vl_interface': 7.716064042914584e-07, 'learning_rate/action_model': 7.716064042914584e-06, 'epoch': 1.73}) +[2026-06-23 23:04:18,160][starVLA.training.train_starvla][INFO] - Step 4471, Loss: {'train/loss': 0.05950005725026131, 'train/grad_norm_pre_clip': 33.58362579345703, 'timing/data': 0.0005703300703316927, 'timing/model': 1.6859190920367837, 'learning_rate/qwen_vl_interface': 7.705922761781226e-07, 'learning_rate/action_model': 7.705922761781226e-06, 'epoch': 1.73}) +[2026-06-23 23:04:21,485][starVLA.training.train_starvla][INFO] - Step 4472, Loss: {'train/loss': 0.0716162770986557, 'train/grad_norm_pre_clip': 18.535863876342773, 'timing/data': 0.0005741959903389215, 'timing/model': 1.6879555680789053, 'learning_rate/qwen_vl_interface': 7.695799893800996e-07, 'learning_rate/action_model': 7.695799893800996e-06, 'epoch': 1.73}) +[2026-06-23 23:04:24,808][starVLA.training.train_starvla][INFO] - Step 4473, Loss: {'train/loss': 0.06149211898446083, 'train/grad_norm_pre_clip': 20.872358322143555, 'timing/data': 0.0005440199747681618, 'timing/model': 1.6882656780071557, 'learning_rate/qwen_vl_interface': 7.685695443135028e-07, 'learning_rate/action_model': 7.685695443135027e-06, 'epoch': 1.73}) +[2026-06-23 23:04:28,132][starVLA.training.train_starvla][INFO] - Step 4474, Loss: {'train/loss': 0.1211000457406044, 'train/grad_norm_pre_clip': 15.737244606018066, 'timing/data': 0.0005794358439743519, 'timing/model': 1.6907340330071747, 'learning_rate/qwen_vl_interface': 7.675609413936888e-07, 'learning_rate/action_model': 7.675609413936888e-06, 'epoch': 1.73}) +[2026-06-23 23:04:31,762][starVLA.training.train_starvla][INFO] - Step 4475, Loss: {'train/loss': 0.10042256116867065, 'train/grad_norm_pre_clip': 25.301939010620117, 'timing/data': 0.0005748381372541189, 'timing/model': 1.9938887099269778, 'learning_rate/qwen_vl_interface': 7.665541810352541e-07, 'learning_rate/action_model': 7.66554181035254e-06, 'epoch': 1.73}) +[2026-06-23 23:04:35,090][starVLA.training.train_starvla][INFO] - Step 4476, Loss: {'train/loss': 0.08928132802248001, 'train/grad_norm_pre_clip': 21.745145797729492, 'timing/data': 0.0006520829629153013, 'timing/model': 1.6919515240006149, 'learning_rate/qwen_vl_interface': 7.655492636520415e-07, 'learning_rate/action_model': 7.655492636520414e-06, 'epoch': 1.73}) +[2026-06-23 23:04:38,412][starVLA.training.train_starvla][INFO] - Step 4477, Loss: {'train/loss': 0.09921012818813324, 'train/grad_norm_pre_clip': 17.225584030151367, 'timing/data': 0.0006871181540191174, 'timing/model': 1.686798769980669, 'learning_rate/qwen_vl_interface': 7.64546189657134e-07, 'learning_rate/action_model': 7.645461896571339e-06, 'epoch': 1.73}) +[2026-06-23 23:04:41,731][starVLA.training.train_starvla][INFO] - Step 4478, Loss: {'train/loss': 0.08424990624189377, 'train/grad_norm_pre_clip': 35.22650146484375, 'timing/data': 0.0005754691082984209, 'timing/model': 1.6863255898933858, 'learning_rate/qwen_vl_interface': 7.635449594628571e-07, 'learning_rate/action_model': 7.635449594628571e-06, 'epoch': 1.73}) +[2026-06-23 23:04:45,050][starVLA.training.train_starvla][INFO] - Step 4479, Loss: {'train/loss': 0.08173566311597824, 'train/grad_norm_pre_clip': 19.71263885498047, 'timing/data': 0.0008657439611852169, 'timing/model': 1.6832646050024778, 'learning_rate/qwen_vl_interface': 7.625455734807803e-07, 'learning_rate/action_model': 7.625455734807803e-06, 'epoch': 1.73}) +[2026-06-23 23:04:48,366][starVLA.training.train_starvla][INFO] - Step 4480, Loss: {'train/loss': 0.08134668320417404, 'train/grad_norm_pre_clip': 48.71025848388672, 'timing/data': 0.0005604210309684277, 'timing/model': 1.68502003303729, 'learning_rate/qwen_vl_interface': 7.615480321217116e-07, 'learning_rate/action_model': 7.615480321217117e-06, 'epoch': 1.73}) +[2026-06-23 23:04:51,690][starVLA.training.train_starvla][INFO] - Step 4481, Loss: {'train/loss': 0.1036255732178688, 'train/grad_norm_pre_clip': 16.042146682739258, 'timing/data': 0.0005715109873563051, 'timing/model': 1.6891562938690186, 'learning_rate/qwen_vl_interface': 7.605523357957028e-07, 'learning_rate/action_model': 7.605523357957028e-06, 'epoch': 1.73}) +[2026-06-23 23:04:55,012][starVLA.training.train_starvla][INFO] - Step 4482, Loss: {'train/loss': 0.13692030310630798, 'train/grad_norm_pre_clip': 15.39006519317627, 'timing/data': 0.0005711808335036039, 'timing/model': 1.6879510378930718, 'learning_rate/qwen_vl_interface': 7.59558484912048e-07, 'learning_rate/action_model': 7.59558484912048e-06, 'epoch': 1.73}) +[2026-06-23 23:04:58,332][starVLA.training.train_starvla][INFO] - Step 4483, Loss: {'train/loss': 0.16852757334709167, 'train/grad_norm_pre_clip': 16.62684440612793, 'timing/data': 0.0005568638443946838, 'timing/model': 1.6819135330151767, 'learning_rate/qwen_vl_interface': 7.585664798792807e-07, 'learning_rate/action_model': 7.585664798792807e-06, 'epoch': 1.74}) +[2026-06-23 23:05:01,659][starVLA.training.train_starvla][INFO] - Step 4484, Loss: {'train/loss': 0.15772628784179688, 'train/grad_norm_pre_clip': 13.325132369995117, 'timing/data': 0.0005666830111294985, 'timing/model': 1.6914906159508973, 'learning_rate/qwen_vl_interface': 7.575763211051779e-07, 'learning_rate/action_model': 7.575763211051778e-06, 'epoch': 1.74}) +[2026-06-23 23:05:04,986][starVLA.training.train_starvla][INFO] - Step 4485, Loss: {'train/loss': 0.132389098405838, 'train/grad_norm_pre_clip': 23.36025047302246, 'timing/data': 0.000540673965588212, 'timing/model': 1.688616099068895, 'learning_rate/qwen_vl_interface': 7.565880089967553e-07, 'learning_rate/action_model': 7.565880089967553e-06, 'epoch': 1.74}) +[2026-06-23 23:05:08,312][starVLA.training.train_starvla][INFO] - Step 4486, Loss: {'train/loss': 0.030111847445368767, 'train/grad_norm_pre_clip': 16.38805389404297, 'timing/data': 0.0005653100088238716, 'timing/model': 1.6902238389011472, 'learning_rate/qwen_vl_interface': 7.556015439602725e-07, 'learning_rate/action_model': 7.556015439602725e-06, 'epoch': 1.74}) +[2026-06-23 23:05:11,638][starVLA.training.train_starvla][INFO] - Step 4487, Loss: {'train/loss': 0.13252080976963043, 'train/grad_norm_pre_clip': 13.131174087524414, 'timing/data': 0.000720650888979435, 'timing/model': 1.6863974679727107, 'learning_rate/qwen_vl_interface': 7.546169264012252e-07, 'learning_rate/action_model': 7.5461692640122525e-06, 'epoch': 1.74}) +[2026-06-23 23:05:14,960][starVLA.training.train_starvla][INFO] - Step 4488, Loss: {'train/loss': 0.06602779775857925, 'train/grad_norm_pre_clip': 13.899469375610352, 'timing/data': 0.0005815608892589808, 'timing/model': 1.6872786418534815, 'learning_rate/qwen_vl_interface': 7.536341567243544e-07, 'learning_rate/action_model': 7.536341567243543e-06, 'epoch': 1.74}) +[2026-06-23 23:05:18,286][starVLA.training.train_starvla][INFO] - Step 4489, Loss: {'train/loss': 0.06433135271072388, 'train/grad_norm_pre_clip': 25.957433700561523, 'timing/data': 0.0007937091868370771, 'timing/model': 1.6889640958979726, 'learning_rate/qwen_vl_interface': 7.526532353336395e-07, 'learning_rate/action_model': 7.526532353336395e-06, 'epoch': 1.74}) +[2026-06-23 23:05:21,830][starVLA.training.train_starvla][INFO] - Step 4490, Loss: {'train/loss': 0.12993040680885315, 'train/grad_norm_pre_clip': 16.610502243041992, 'timing/data': 0.0005602010060101748, 'timing/model': 1.9106621099635959, 'learning_rate/qwen_vl_interface': 7.516741626322994e-07, 'learning_rate/action_model': 7.516741626322994e-06, 'epoch': 1.74}) +[2026-06-23 23:05:25,157][starVLA.training.train_starvla][INFO] - Step 4491, Loss: {'train/loss': 0.09862552583217621, 'train/grad_norm_pre_clip': 25.5960693359375, 'timing/data': 0.0006804659496992826, 'timing/model': 1.68820880795829, 'learning_rate/qwen_vl_interface': 7.506969390227945e-07, 'learning_rate/action_model': 7.506969390227945e-06, 'epoch': 1.74}) +[2026-06-23 23:05:28,479][starVLA.training.train_starvla][INFO] - Step 4492, Loss: {'train/loss': 0.0968702957034111, 'train/grad_norm_pre_clip': 17.232891082763672, 'timing/data': 0.0006043729372322559, 'timing/model': 1.6863136070314795, 'learning_rate/qwen_vl_interface': 7.497215649068244e-07, 'learning_rate/action_model': 7.497215649068244e-06, 'epoch': 1.74}) +[2026-06-23 23:05:31,802][starVLA.training.train_starvla][INFO] - Step 4493, Loss: {'train/loss': 0.09904220700263977, 'train/grad_norm_pre_clip': 69.85369873046875, 'timing/data': 0.000595086021348834, 'timing/model': 1.6858917181380093, 'learning_rate/qwen_vl_interface': 7.48748040685329e-07, 'learning_rate/action_model': 7.487480406853289e-06, 'epoch': 1.74}) +[2026-06-23 23:05:35,124][starVLA.training.train_starvla][INFO] - Step 4494, Loss: {'train/loss': 0.05662092566490173, 'train/grad_norm_pre_clip': 41.6857795715332, 'timing/data': 0.0008437321521341801, 'timing/model': 1.6904397478792816, 'learning_rate/qwen_vl_interface': 7.477763667584867e-07, 'learning_rate/action_model': 7.477763667584866e-06, 'epoch': 1.74}) +[2026-06-23 23:05:38,452][starVLA.training.train_starvla][INFO] - Step 4495, Loss: {'train/loss': 0.1320364773273468, 'train/grad_norm_pre_clip': 26.495437622070312, 'timing/data': 0.0006413229275494814, 'timing/model': 1.691306215012446, 'learning_rate/qwen_vl_interface': 7.468065435257159e-07, 'learning_rate/action_model': 7.468065435257159e-06, 'epoch': 1.74}) +[2026-06-23 23:05:41,776][starVLA.training.train_starvla][INFO] - Step 4496, Loss: {'train/loss': 0.051342692226171494, 'train/grad_norm_pre_clip': 25.490720748901367, 'timing/data': 0.0006697559729218483, 'timing/model': 1.687908705091104, 'learning_rate/qwen_vl_interface': 7.458385713856745e-07, 'learning_rate/action_model': 7.458385713856744e-06, 'epoch': 1.74}) +[2026-06-23 23:05:45,098][starVLA.training.train_starvla][INFO] - Step 4497, Loss: {'train/loss': 0.11629504710435867, 'train/grad_norm_pre_clip': 15.846205711364746, 'timing/data': 0.0006401510909199715, 'timing/model': 1.6865865760482848, 'learning_rate/qwen_vl_interface': 7.448724507362605e-07, 'learning_rate/action_model': 7.448724507362605e-06, 'epoch': 1.74}) +[2026-06-23 23:05:48,420][starVLA.training.train_starvla][INFO] - Step 4498, Loss: {'train/loss': 0.055647991597652435, 'train/grad_norm_pre_clip': 28.554174423217773, 'timing/data': 0.000645000021904707, 'timing/model': 1.6859221609774977, 'learning_rate/qwen_vl_interface': 7.439081819746078e-07, 'learning_rate/action_model': 7.4390818197460775e-06, 'epoch': 1.74}) +[2026-06-23 23:05:51,748][starVLA.training.train_starvla][INFO] - Step 4499, Loss: {'train/loss': 0.09030212461948395, 'train/grad_norm_pre_clip': 18.580036163330078, 'timing/data': 0.0006172480061650276, 'timing/model': 1.6923835100606084, 'learning_rate/qwen_vl_interface': 7.42945765497091e-07, 'learning_rate/action_model': 7.42945765497091e-06, 'epoch': 1.74}) +[2026-06-23 23:05:55,076][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4500 +[2026-06-23 23:07:26,661][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4500 +[2026-06-23 23:07:26,662][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4500 +[2026-06-23 23:08:37,554][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4500 +[2026-06-23 23:09:37,820][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4500 +[2026-06-23 23:09:37,821][starVLA.training.train_starvla][INFO] - Step 4500, Loss: {'train/loss': 0.16881856322288513, 'train/grad_norm_pre_clip': 24.866607666015625, 'eval/action_loss/seconds': 91.58323521283455, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06473440676927567, 'eval/latency_2/loss': 0.06473440676927567, 'eval/flappy/loss': 0.06473440676927567, 'eval/flappy/latency_2/loss': 0.06473440676927567, 'eval/flap_precision': 0.9394347240915208, 'eval/flap_recall': 0.941970310391363, 'eval/flap_f1': 0.940700808625337, 'eval/flap_rate_model': 0.2063888888888889, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.940700808625337, 'eval/latency_2/flap_precision': 0.9394347240915208, 'eval/latency_2/flap_recall': 0.941970310391363, 'eval/latency_2/flap_f1': 0.940700808625337, 'eval/latency_2/flap_rate_model': 0.2063888888888889, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.940700808625337, 'eval/action_classification/data_wait_seconds': 0.807358980178833, 'eval/action_classification/predict_seconds': 69.95413970947266, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 45.799999999999315, 'rl_games_eval/mid_train/mean_length': 388.0, 'rl_games_eval/mid_train/std_reward': 58.619450696844915, 'rl_games_eval/mid_train/std_length': 473.2098900065382, 'rl_games_eval/mid_train/macro_mean_reward': 45.799999999999315, 'rl_games_eval/mid_train/macro_mean_length': 388.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 45.799999999999315, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 388.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 58.619450696844915, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 473.2098900065382, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 45.799999999999315, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005648098886013031, 'timing/model': 1.6901752969715744, 'learning_rate/qwen_vl_interface': 7.41985201699324e-07, 'learning_rate/action_model': 7.41985201699324e-06, 'epoch': 1.74}) +[2026-06-23 23:09:37,821][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_4500_state +[2026-06-23 23:09:37,824][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 23:09:40,586][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_4500_state/model.safetensors +[2026-06-23 23:09:41,746][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_4500_state/optimizer.bin +[2026-06-23 23:09:41,746][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_4500_state/sampler.bin +[2026-06-23 23:09:41,747][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_4500_state/sampler_1.bin +[2026-06-23 23:09:41,747][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_4500_state/random_states_0.pkl +[2026-06-23 23:09:41,772][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4500 +[2026-06-23 23:09:41,772][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 23:09:41,782][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 23:09:45,069][starVLA.training.train_starvla][INFO] - Step 4501, Loss: {'train/loss': 0.14373327791690826, 'train/grad_norm_pre_clip': 38.75566482543945, 'timing/data': 0.0005780148785561323, 'timing/model': 1.6615001729223877, 'learning_rate/qwen_vl_interface': 7.410264909761581e-07, 'learning_rate/action_model': 7.41026490976158e-06, 'epoch': 1.74}) +[2026-06-23 23:09:48,349][starVLA.training.train_starvla][INFO] - Step 4502, Loss: {'train/loss': 0.09374593198299408, 'train/grad_norm_pre_clip': 15.496793746948242, 'timing/data': 0.0006986910011619329, 'timing/model': 1.6660853340290487, 'learning_rate/qwen_vl_interface': 7.40069633721683e-07, 'learning_rate/action_model': 7.400696337216829e-06, 'epoch': 1.74}) +[2026-06-23 23:09:51,635][starVLA.training.train_starvla][INFO] - Step 4503, Loss: {'train/loss': 0.09663183987140656, 'train/grad_norm_pre_clip': 26.760696411132812, 'timing/data': 0.00066590984351933, 'timing/model': 1.6696769900154322, 'learning_rate/qwen_vl_interface': 7.391146303292262e-07, 'learning_rate/action_model': 7.391146303292262e-06, 'epoch': 1.74}) +[2026-06-23 23:09:54,929][starVLA.training.train_starvla][INFO] - Step 4504, Loss: {'train/loss': 0.14249616861343384, 'train/grad_norm_pre_clip': 14.257808685302734, 'timing/data': 0.0006320660468190908, 'timing/model': 1.6724560791626573, 'learning_rate/qwen_vl_interface': 7.381614811913538e-07, 'learning_rate/action_model': 7.381614811913537e-06, 'epoch': 1.74}) +[2026-06-23 23:09:58,229][starVLA.training.train_starvla][INFO] - Step 4505, Loss: {'train/loss': 0.06198018044233322, 'train/grad_norm_pre_clip': 21.476673126220703, 'timing/data': 0.0005772439762949944, 'timing/model': 1.676998530048877, 'learning_rate/qwen_vl_interface': 7.372101866998693e-07, 'learning_rate/action_model': 7.372101866998693e-06, 'epoch': 1.74}) +[2026-06-23 23:10:01,533][starVLA.training.train_starvla][INFO] - Step 4506, Loss: {'train/loss': 0.050212156027555466, 'train/grad_norm_pre_clip': 9.621697425842285, 'timing/data': 0.0006060472223907709, 'timing/model': 1.6795449540950358, 'learning_rate/qwen_vl_interface': 7.362607472458147e-07, 'learning_rate/action_model': 7.362607472458147e-06, 'epoch': 1.74}) +[2026-06-23 23:10:04,840][starVLA.training.train_starvla][INFO] - Step 4507, Loss: {'train/loss': 0.17456412315368652, 'train/grad_norm_pre_clip': 46.97911834716797, 'timing/data': 0.0006291901227086782, 'timing/model': 1.6760144717991352, 'learning_rate/qwen_vl_interface': 7.353131632194682e-07, 'learning_rate/action_model': 7.3531316321946816e-06, 'epoch': 1.74}) +[2026-06-23 23:10:08,247][starVLA.training.train_starvla][INFO] - Step 4508, Loss: {'train/loss': 0.12738938629627228, 'train/grad_norm_pre_clip': 25.677541732788086, 'timing/data': 0.0006340399850159883, 'timing/model': 1.778120411094278, 'learning_rate/qwen_vl_interface': 7.343674350103455e-07, 'learning_rate/action_model': 7.343674350103455e-06, 'epoch': 1.74}) +[2026-06-23 23:10:11,547][starVLA.training.train_starvla][INFO] - Step 4509, Loss: {'train/loss': 0.13468407094478607, 'train/grad_norm_pre_clip': 29.271169662475586, 'timing/data': 0.0006339298561215401, 'timing/model': 1.6756079278420657, 'learning_rate/qwen_vl_interface': 7.334235630071995e-07, 'learning_rate/action_model': 7.334235630071995e-06, 'epoch': 1.75}) +[2026-06-23 23:10:14,857][starVLA.training.train_starvla][INFO] - Step 4510, Loss: {'train/loss': 0.1451193392276764, 'train/grad_norm_pre_clip': 48.206451416015625, 'timing/data': 0.0005799680948257446, 'timing/model': 1.6799290829803795, 'learning_rate/qwen_vl_interface': 7.324815475980207e-07, 'learning_rate/action_model': 7.324815475980207e-06, 'epoch': 1.75}) +[2026-06-23 23:10:18,164][starVLA.training.train_starvla][INFO] - Step 4511, Loss: {'train/loss': 0.1669047623872757, 'train/grad_norm_pre_clip': 28.370962142944336, 'timing/data': 0.0005758509505540133, 'timing/model': 1.6803320930339396, 'learning_rate/qwen_vl_interface': 7.315413891700363e-07, 'learning_rate/action_model': 7.315413891700362e-06, 'epoch': 1.75}) +[2026-06-23 23:10:21,471][starVLA.training.train_starvla][INFO] - Step 4512, Loss: {'train/loss': 0.17678822576999664, 'train/grad_norm_pre_clip': 32.872745513916016, 'timing/data': 0.0008341760840266943, 'timing/model': 1.680733168963343, 'learning_rate/qwen_vl_interface': 7.306030881097092e-07, 'learning_rate/action_model': 7.3060308810970925e-06, 'epoch': 1.75}) +[2026-06-23 23:10:24,783][starVLA.training.train_starvla][INFO] - Step 4513, Loss: {'train/loss': 0.15820714831352234, 'train/grad_norm_pre_clip': 23.36239242553711, 'timing/data': 0.0006151348352432251, 'timing/model': 1.6833292548544705, 'learning_rate/qwen_vl_interface': 7.296666448027406e-07, 'learning_rate/action_model': 7.296666448027406e-06, 'epoch': 1.75}) +[2026-06-23 23:10:28,094][starVLA.training.train_starvla][INFO] - Step 4514, Loss: {'train/loss': 0.230451762676239, 'train/grad_norm_pre_clip': 31.24138069152832, 'timing/data': 0.0006387489847838879, 'timing/model': 1.6817442830651999, 'learning_rate/qwen_vl_interface': 7.287320596340664e-07, 'learning_rate/action_model': 7.287320596340663e-06, 'epoch': 1.75}) +[2026-06-23 23:10:31,405][starVLA.training.train_starvla][INFO] - Step 4515, Loss: {'train/loss': 0.0771777555346489, 'train/grad_norm_pre_clip': 16.779560089111328, 'timing/data': 0.000633730087429285, 'timing/model': 1.6820345418527722, 'learning_rate/qwen_vl_interface': 7.277993329878593e-07, 'learning_rate/action_model': 7.277993329878594e-06, 'epoch': 1.75}) +[2026-06-23 23:10:34,724][starVLA.training.train_starvla][INFO] - Step 4516, Loss: {'train/loss': 0.10487902164459229, 'train/grad_norm_pre_clip': 61.96746063232422, 'timing/data': 0.0006542580667883158, 'timing/model': 1.6891739768907428, 'learning_rate/qwen_vl_interface': 7.268684652475277e-07, 'learning_rate/action_model': 7.268684652475277e-06, 'epoch': 1.75}) +[2026-06-23 23:10:38,041][starVLA.training.train_starvla][INFO] - Step 4517, Loss: {'train/loss': 0.09682713449001312, 'train/grad_norm_pre_clip': 25.530988693237305, 'timing/data': 0.0007162739057093859, 'timing/model': 1.6866009989753366, 'learning_rate/qwen_vl_interface': 7.259394567957169e-07, 'learning_rate/action_model': 7.259394567957169e-06, 'epoch': 1.75}) +[2026-06-23 23:10:41,363][starVLA.training.train_starvla][INFO] - Step 4518, Loss: {'train/loss': 0.08415717631578445, 'train/grad_norm_pre_clip': 21.901399612426758, 'timing/data': 0.0006447599735110998, 'timing/model': 1.6924667719285935, 'learning_rate/qwen_vl_interface': 7.250123080143066e-07, 'learning_rate/action_model': 7.2501230801430665e-06, 'epoch': 1.75}) +[2026-06-23 23:10:44,679][starVLA.training.train_starvla][INFO] - Step 4519, Loss: {'train/loss': 0.1544000506401062, 'train/grad_norm_pre_clip': 41.67699432373047, 'timing/data': 0.0006787439342588186, 'timing/model': 1.6839764788746834, 'learning_rate/qwen_vl_interface': 7.240870192844127e-07, 'learning_rate/action_model': 7.240870192844127e-06, 'epoch': 1.75}) +[2026-06-23 23:10:47,995][starVLA.training.train_starvla][INFO] - Step 4520, Loss: {'train/loss': 0.09332035481929779, 'train/grad_norm_pre_clip': 26.181991577148438, 'timing/data': 0.0006140030454844236, 'timing/model': 1.6860907189548016, 'learning_rate/qwen_vl_interface': 7.231635909863858e-07, 'learning_rate/action_model': 7.2316359098638585e-06, 'epoch': 1.75}) +[2026-06-23 23:10:51,308][starVLA.training.train_starvla][INFO] - Step 4521, Loss: {'train/loss': 0.1261468231678009, 'train/grad_norm_pre_clip': 46.47977828979492, 'timing/data': 0.0005591390654444695, 'timing/model': 1.6823689660523087, 'learning_rate/qwen_vl_interface': 7.222420234998138e-07, 'learning_rate/action_model': 7.2224202349981385e-06, 'epoch': 1.75}) +[2026-06-23 23:10:54,623][starVLA.training.train_starvla][INFO] - Step 4522, Loss: {'train/loss': 0.0729975774884224, 'train/grad_norm_pre_clip': 24.512453079223633, 'timing/data': 0.0005695389118045568, 'timing/model': 1.6831132660154253, 'learning_rate/qwen_vl_interface': 7.213223172035175e-07, 'learning_rate/action_model': 7.213223172035174e-06, 'epoch': 1.75}) +[2026-06-23 23:10:58,239][starVLA.training.train_starvla][INFO] - Step 4523, Loss: {'train/loss': 0.19056355953216553, 'train/grad_norm_pre_clip': 26.925268173217773, 'timing/data': 0.0005539301782846451, 'timing/model': 1.9866093969903886, 'learning_rate/qwen_vl_interface': 7.204044724755528e-07, 'learning_rate/action_model': 7.204044724755528e-06, 'epoch': 1.75}) +[2026-06-23 23:11:01,569][starVLA.training.train_starvla][INFO] - Step 4524, Loss: {'train/loss': 0.06881216168403625, 'train/grad_norm_pre_clip': 10.151228904724121, 'timing/data': 0.0006972090341150761, 'timing/model': 1.6964789109770209, 'learning_rate/qwen_vl_interface': 7.194884896932132e-07, 'learning_rate/action_model': 7.194884896932132e-06, 'epoch': 1.75}) +[2026-06-23 23:11:04,885][starVLA.training.train_starvla][INFO] - Step 4525, Loss: {'train/loss': 0.13272204995155334, 'train/grad_norm_pre_clip': 28.082599639892578, 'timing/data': 0.0005630769301205873, 'timing/model': 1.6852809221018106, 'learning_rate/qwen_vl_interface': 7.185743692330221e-07, 'learning_rate/action_model': 7.185743692330221e-06, 'epoch': 1.75}) +[2026-06-23 23:11:08,205][starVLA.training.train_starvla][INFO] - Step 4526, Loss: {'train/loss': 0.09467696398496628, 'train/grad_norm_pre_clip': 40.129066467285156, 'timing/data': 0.0006343210116028786, 'timing/model': 1.686534307198599, 'learning_rate/qwen_vl_interface': 7.176621114707408e-07, 'learning_rate/action_model': 7.176621114707407e-06, 'epoch': 1.75}) +[2026-06-23 23:11:11,521][starVLA.training.train_starvla][INFO] - Step 4527, Loss: {'train/loss': 0.12159538269042969, 'train/grad_norm_pre_clip': 26.748334884643555, 'timing/data': 0.0005739571060985327, 'timing/model': 1.682651188923046, 'learning_rate/qwen_vl_interface': 7.167517167813642e-07, 'learning_rate/action_model': 7.167517167813642e-06, 'epoch': 1.75}) +[2026-06-23 23:11:14,838][starVLA.training.train_starvla][INFO] - Step 4528, Loss: {'train/loss': 0.12032394856214523, 'train/grad_norm_pre_clip': 22.02321434020996, 'timing/data': 0.0005625961348414421, 'timing/model': 1.6854129950515926, 'learning_rate/qwen_vl_interface': 7.15843185539121e-07, 'learning_rate/action_model': 7.15843185539121e-06, 'epoch': 1.75}) +[2026-06-23 23:11:18,152][starVLA.training.train_starvla][INFO] - Step 4529, Loss: {'train/loss': 0.08117020130157471, 'train/grad_norm_pre_clip': 15.238597869873047, 'timing/data': 0.0005620759911835194, 'timing/model': 1.6850087570492178, 'learning_rate/qwen_vl_interface': 7.149365181174739e-07, 'learning_rate/action_model': 7.149365181174739e-06, 'epoch': 1.75}) +[2026-06-23 23:11:21,475][starVLA.training.train_starvla][INFO] - Step 4530, Loss: {'train/loss': 0.10345947742462158, 'train/grad_norm_pre_clip': 19.878503799438477, 'timing/data': 0.0005524670705199242, 'timing/model': 1.6873593819327652, 'learning_rate/qwen_vl_interface': 7.14031714889121e-07, 'learning_rate/action_model': 7.14031714889121e-06, 'epoch': 1.75}) +[2026-06-23 23:11:24,790][starVLA.training.train_starvla][INFO] - Step 4531, Loss: {'train/loss': 0.11867893487215042, 'train/grad_norm_pre_clip': 34.959381103515625, 'timing/data': 0.0005700800102204084, 'timing/model': 1.6861165089067072, 'learning_rate/qwen_vl_interface': 7.131287762259907e-07, 'learning_rate/action_model': 7.131287762259907e-06, 'epoch': 1.75}) +[2026-06-23 23:11:28,104][starVLA.training.train_starvla][INFO] - Step 4532, Loss: {'train/loss': 0.05871326103806496, 'train/grad_norm_pre_clip': 20.4764461517334, 'timing/data': 0.0005735459271818399, 'timing/model': 1.6828707661479712, 'learning_rate/qwen_vl_interface': 7.12227702499248e-07, 'learning_rate/action_model': 7.12227702499248e-06, 'epoch': 1.75}) +[2026-06-23 23:11:31,421][starVLA.training.train_starvla][INFO] - Step 4533, Loss: {'train/loss': 0.05088742449879646, 'train/grad_norm_pre_clip': 21.133098602294922, 'timing/data': 0.0005637879949063063, 'timing/model': 1.6870393219869584, 'learning_rate/qwen_vl_interface': 7.113284940792903e-07, 'learning_rate/action_model': 7.113284940792903e-06, 'epoch': 1.75}) +[2026-06-23 23:11:34,750][starVLA.training.train_starvla][INFO] - Step 4534, Loss: {'train/loss': 0.10437899082899094, 'train/grad_norm_pre_clip': 24.927143096923828, 'timing/data': 0.0006663510575890541, 'timing/model': 1.691067645791918, 'learning_rate/qwen_vl_interface': 7.104311513357476e-07, 'learning_rate/action_model': 7.1043115133574755e-06, 'epoch': 1.75}) +[2026-06-23 23:11:38,077][starVLA.training.train_starvla][INFO] - Step 4535, Loss: {'train/loss': 0.12591539323329926, 'train/grad_norm_pre_clip': 24.47463035583496, 'timing/data': 0.0006338399834930897, 'timing/model': 1.6912721199914813, 'learning_rate/qwen_vl_interface': 7.095356746374845e-07, 'learning_rate/action_model': 7.095356746374844e-06, 'epoch': 1.76}) +[2026-06-23 23:11:41,394][starVLA.training.train_starvla][INFO] - Step 4536, Loss: {'train/loss': 0.16750012338161469, 'train/grad_norm_pre_clip': 71.43716430664062, 'timing/data': 0.0005835348274558783, 'timing/model': 1.686340562067926, 'learning_rate/qwen_vl_interface': 7.086420643525973e-07, 'learning_rate/action_model': 7.086420643525972e-06, 'epoch': 1.76}) +[2026-06-23 23:11:44,711][starVLA.training.train_starvla][INFO] - Step 4537, Loss: {'train/loss': 0.15622563660144806, 'train/grad_norm_pre_clip': 45.76068878173828, 'timing/data': 0.000579147832468152, 'timing/model': 1.6872641569934785, 'learning_rate/qwen_vl_interface': 7.077503208484151e-07, 'learning_rate/action_model': 7.0775032084841515e-06, 'epoch': 1.76}) +[2026-06-23 23:11:48,337][starVLA.training.train_starvla][INFO] - Step 4538, Loss: {'train/loss': 0.16445292532444, 'train/grad_norm_pre_clip': 33.7940788269043, 'timing/data': 0.0005420679226517677, 'timing/model': 1.990312362788245, 'learning_rate/qwen_vl_interface': 7.068604444915003e-07, 'learning_rate/action_model': 7.068604444915003e-06, 'epoch': 1.76}) +[2026-06-23 23:11:51,658][starVLA.training.train_starvla][INFO] - Step 4539, Loss: {'train/loss': 0.06155338138341904, 'train/grad_norm_pre_clip': 27.37860679626465, 'timing/data': 0.0008307190146297216, 'timing/model': 1.6871516169048846, 'learning_rate/qwen_vl_interface': 7.059724356476475e-07, 'learning_rate/action_model': 7.059724356476474e-06, 'epoch': 1.76}) +[2026-06-23 23:11:54,970][starVLA.training.train_starvla][INFO] - Step 4540, Loss: {'train/loss': 0.07319849729537964, 'train/grad_norm_pre_clip': 24.45195770263672, 'timing/data': 0.0006639959756284952, 'timing/model': 1.6826357699465007, 'learning_rate/qwen_vl_interface': 7.050862946818836e-07, 'learning_rate/action_model': 7.050862946818836e-06, 'epoch': 1.76}) +[2026-06-23 23:11:58,291][starVLA.training.train_starvla][INFO] - Step 4541, Loss: {'train/loss': 0.18152281641960144, 'train/grad_norm_pre_clip': 41.376434326171875, 'timing/data': 0.0005421880632638931, 'timing/model': 1.6856362919788808, 'learning_rate/qwen_vl_interface': 7.042020219584673e-07, 'learning_rate/action_model': 7.042020219584673e-06, 'epoch': 1.76}) +[2026-06-23 23:12:01,613][starVLA.training.train_starvla][INFO] - Step 4542, Loss: {'train/loss': 0.21222133934497833, 'train/grad_norm_pre_clip': 21.73111915588379, 'timing/data': 0.0005858899094164371, 'timing/model': 1.686669260961935, 'learning_rate/qwen_vl_interface': 7.0331961784089e-07, 'learning_rate/action_model': 7.033196178408899e-06, 'epoch': 1.76}) +[2026-06-23 23:12:04,934][starVLA.training.train_starvla][INFO] - Step 4543, Loss: {'train/loss': 0.06247762590646744, 'train/grad_norm_pre_clip': 17.70055389404297, 'timing/data': 0.0005659819580614567, 'timing/model': 1.6865605770144612, 'learning_rate/qwen_vl_interface': 7.024390826918748e-07, 'learning_rate/action_model': 7.024390826918747e-06, 'epoch': 1.76}) +[2026-06-23 23:12:08,248][starVLA.training.train_starvla][INFO] - Step 4544, Loss: {'train/loss': 0.05865531414747238, 'train/grad_norm_pre_clip': 9.930499076843262, 'timing/data': 0.0005716730374842882, 'timing/model': 1.681265203980729, 'learning_rate/qwen_vl_interface': 7.01560416873376e-07, 'learning_rate/action_model': 7.015604168733759e-06, 'epoch': 1.76}) +[2026-06-23 23:12:11,570][starVLA.training.train_starvla][INFO] - Step 4545, Loss: {'train/loss': 0.08822271227836609, 'train/grad_norm_pre_clip': 25.379167556762695, 'timing/data': 0.0005673549603670835, 'timing/model': 1.6875868521165103, 'learning_rate/qwen_vl_interface': 7.006836207465803e-07, 'learning_rate/action_model': 7.006836207465803e-06, 'epoch': 1.76}) +[2026-06-23 23:12:14,887][starVLA.training.train_starvla][INFO] - Step 4546, Loss: {'train/loss': 0.17718403041362762, 'train/grad_norm_pre_clip': 27.74561882019043, 'timing/data': 0.0005830449517816305, 'timing/model': 1.6838452860247344, 'learning_rate/qwen_vl_interface': 6.998086946719058e-07, 'learning_rate/action_model': 6.998086946719057e-06, 'epoch': 1.76}) +[2026-06-23 23:12:18,205][starVLA.training.train_starvla][INFO] - Step 4547, Loss: {'train/loss': 0.09338243305683136, 'train/grad_norm_pre_clip': 18.14983558654785, 'timing/data': 0.0005451939068734646, 'timing/model': 1.6867485961411148, 'learning_rate/qwen_vl_interface': 6.989356390090008e-07, 'learning_rate/action_model': 6.989356390090007e-06, 'epoch': 1.76}) +[2026-06-23 23:12:21,525][starVLA.training.train_starvla][INFO] - Step 4548, Loss: {'train/loss': 0.031492944806814194, 'train/grad_norm_pre_clip': 30.464359283447266, 'timing/data': 0.000571653014048934, 'timing/model': 1.6867020069621503, 'learning_rate/qwen_vl_interface': 6.980644541167459e-07, 'learning_rate/action_model': 6.980644541167458e-06, 'epoch': 1.76}) +[2026-06-23 23:12:24,845][starVLA.training.train_starvla][INFO] - Step 4549, Loss: {'train/loss': 0.10501697659492493, 'train/grad_norm_pre_clip': 23.792402267456055, 'timing/data': 0.0005603018216788769, 'timing/model': 1.6900154468603432, 'learning_rate/qwen_vl_interface': 6.971951403532524e-07, 'learning_rate/action_model': 6.971951403532523e-06, 'epoch': 1.76}) +[2026-06-23 23:12:28,169][starVLA.training.train_starvla][INFO] - Step 4550, Loss: {'train/loss': 0.08836431056261063, 'train/grad_norm_pre_clip': 25.598342895507812, 'timing/data': 0.0006475050467997789, 'timing/model': 1.6907180149573833, 'learning_rate/qwen_vl_interface': 6.963276980758622e-07, 'learning_rate/action_model': 6.963276980758622e-06, 'epoch': 1.76}) +[2026-06-23 23:12:31,489][starVLA.training.train_starvla][INFO] - Step 4551, Loss: {'train/loss': 0.09517814964056015, 'train/grad_norm_pre_clip': 31.550012588500977, 'timing/data': 0.0005769741255789995, 'timing/model': 1.6864442189689726, 'learning_rate/qwen_vl_interface': 6.954621276411485e-07, 'learning_rate/action_model': 6.954621276411484e-06, 'epoch': 1.76}) +[2026-06-23 23:12:34,813][starVLA.training.train_starvla][INFO] - Step 4552, Loss: {'train/loss': 0.08530115336179733, 'train/grad_norm_pre_clip': 17.513273239135742, 'timing/data': 0.0005830749869346619, 'timing/model': 1.6923808939754963, 'learning_rate/qwen_vl_interface': 6.945984294049136e-07, 'learning_rate/action_model': 6.9459842940491364e-06, 'epoch': 1.76}) +[2026-06-23 23:12:38,234][starVLA.training.train_starvla][INFO] - Step 4553, Loss: {'train/loss': 0.18349000811576843, 'train/grad_norm_pre_clip': 19.793167114257812, 'timing/data': 0.0005747491959482431, 'timing/model': 1.6889965180307627, 'learning_rate/qwen_vl_interface': 6.937366037221919e-07, 'learning_rate/action_model': 6.937366037221918e-06, 'epoch': 1.76}) +[2026-06-23 23:12:41,549][starVLA.training.train_starvla][INFO] - Step 4554, Loss: {'train/loss': 0.11662299931049347, 'train/grad_norm_pre_clip': 29.98152732849121, 'timing/data': 0.000572896096855402, 'timing/model': 1.6841936891432852, 'learning_rate/qwen_vl_interface': 6.928766509472471e-07, 'learning_rate/action_model': 6.92876650947247e-06, 'epoch': 1.76}) +[2026-06-23 23:12:44,871][starVLA.training.train_starvla][INFO] - Step 4555, Loss: {'train/loss': 0.1323288083076477, 'train/grad_norm_pre_clip': 23.405481338500977, 'timing/data': 0.0005504328291863203, 'timing/model': 1.6869464749470353, 'learning_rate/qwen_vl_interface': 6.920185714335737e-07, 'learning_rate/action_model': 6.9201857143357375e-06, 'epoch': 1.76}) +[2026-06-23 23:12:48,189][starVLA.training.train_starvla][INFO] - Step 4556, Loss: {'train/loss': 0.16132886707782745, 'train/grad_norm_pre_clip': 19.566438674926758, 'timing/data': 0.0007045629899948835, 'timing/model': 1.6856213451828808, 'learning_rate/qwen_vl_interface': 6.911623655338959e-07, 'learning_rate/action_model': 6.911623655338959e-06, 'epoch': 1.76}) +[2026-06-23 23:12:51,513][starVLA.training.train_starvla][INFO] - Step 4557, Loss: {'train/loss': 0.06270027160644531, 'train/grad_norm_pre_clip': 14.701421737670898, 'timing/data': 0.0006053862161934376, 'timing/model': 1.6877584950998425, 'learning_rate/qwen_vl_interface': 6.903080336001666e-07, 'learning_rate/action_model': 6.903080336001666e-06, 'epoch': 1.76}) +[2026-06-23 23:12:54,829][starVLA.training.train_starvla][INFO] - Step 4558, Loss: {'train/loss': 0.0918445885181427, 'train/grad_norm_pre_clip': 18.630126953125, 'timing/data': 0.0005450230091810226, 'timing/model': 1.6861628170590848, 'learning_rate/qwen_vl_interface': 6.894555759835707e-07, 'learning_rate/action_model': 6.894555759835707e-06, 'epoch': 1.76}) +[2026-06-23 23:12:58,150][starVLA.training.train_starvla][INFO] - Step 4559, Loss: {'train/loss': 0.12874166667461395, 'train/grad_norm_pre_clip': 27.791494369506836, 'timing/data': 0.0005436909850686789, 'timing/model': 1.686163425212726, 'learning_rate/qwen_vl_interface': 6.886049930345204e-07, 'learning_rate/action_model': 6.8860499303452035e-06, 'epoch': 1.76}) +[2026-06-23 23:13:01,473][starVLA.training.train_starvla][INFO] - Step 4560, Loss: {'train/loss': 0.1657305210828781, 'train/grad_norm_pre_clip': 51.276885986328125, 'timing/data': 0.0005998769775032997, 'timing/model': 1.6875279450323433, 'learning_rate/qwen_vl_interface': 6.877562851026584e-07, 'learning_rate/action_model': 6.877562851026584e-06, 'epoch': 1.76}) +[2026-06-23 23:13:04,790][starVLA.training.train_starvla][INFO] - Step 4561, Loss: {'train/loss': 0.04476549103856087, 'train/grad_norm_pre_clip': 27.230743408203125, 'timing/data': 0.0005848980508744717, 'timing/model': 1.6860183079261333, 'learning_rate/qwen_vl_interface': 6.869094525368572e-07, 'learning_rate/action_model': 6.8690945253685715e-06, 'epoch': 1.77}) +[2026-06-23 23:13:08,120][starVLA.training.train_starvla][INFO] - Step 4562, Loss: {'train/loss': 0.1315046101808548, 'train/grad_norm_pre_clip': 18.691553115844727, 'timing/data': 0.0005803799722343683, 'timing/model': 1.6890143910422921, 'learning_rate/qwen_vl_interface': 6.860644956852168e-07, 'learning_rate/action_model': 6.860644956852168e-06, 'epoch': 1.77}) +[2026-06-23 23:13:11,444][starVLA.training.train_starvla][INFO] - Step 4563, Loss: {'train/loss': 0.14208842813968658, 'train/grad_norm_pre_clip': 22.6707706451416, 'timing/data': 0.0005810610018670559, 'timing/model': 1.6899901200085878, 'learning_rate/qwen_vl_interface': 6.852214148950667e-07, 'learning_rate/action_model': 6.852214148950667e-06, 'epoch': 1.77}) +[2026-06-23 23:13:14,762][starVLA.training.train_starvla][INFO] - Step 4564, Loss: {'train/loss': 0.1201813593506813, 'train/grad_norm_pre_clip': 43.99493408203125, 'timing/data': 0.0005496111698448658, 'timing/model': 1.683934235945344, 'learning_rate/qwen_vl_interface': 6.843802105129667e-07, 'learning_rate/action_model': 6.843802105129666e-06, 'epoch': 1.77}) +[2026-06-23 23:13:18,114][starVLA.training.train_starvla][INFO] - Step 4565, Loss: {'train/loss': 0.09999427944421768, 'train/grad_norm_pre_clip': 37.88290786743164, 'timing/data': 0.000561645021662116, 'timing/model': 1.715496954973787, 'learning_rate/qwen_vl_interface': 6.835408828847035e-07, 'learning_rate/action_model': 6.8354088288470356e-06, 'epoch': 1.77}) +[2026-06-23 23:13:21,433][starVLA.training.train_starvla][INFO] - Step 4566, Loss: {'train/loss': 0.14048925042152405, 'train/grad_norm_pre_clip': 17.62897300720215, 'timing/data': 0.0005615148693323135, 'timing/model': 1.6865428281016648, 'learning_rate/qwen_vl_interface': 6.827034323552929e-07, 'learning_rate/action_model': 6.827034323552929e-06, 'epoch': 1.77}) +[2026-06-23 23:13:24,849][starVLA.training.train_starvla][INFO] - Step 4567, Loss: {'train/loss': 0.04267319664359093, 'train/grad_norm_pre_clip': 17.657062530517578, 'timing/data': 0.0005530789494514465, 'timing/model': 1.7825101760681719, 'learning_rate/qwen_vl_interface': 6.818678592689794e-07, 'learning_rate/action_model': 6.818678592689794e-06, 'epoch': 1.77}) +[2026-06-23 23:13:28,170][starVLA.training.train_starvla][INFO] - Step 4568, Loss: {'train/loss': 0.17874731123447418, 'train/grad_norm_pre_clip': 20.79177474975586, 'timing/data': 0.0005423780530691147, 'timing/model': 1.6869070348329842, 'learning_rate/qwen_vl_interface': 6.810341639692357e-07, 'learning_rate/action_model': 6.810341639692357e-06, 'epoch': 1.77}) +[2026-06-23 23:13:31,491][starVLA.training.train_starvla][INFO] - Step 4569, Loss: {'train/loss': 0.08658905327320099, 'train/grad_norm_pre_clip': 29.072729110717773, 'timing/data': 0.0005810910370200872, 'timing/model': 1.6899179150350392, 'learning_rate/qwen_vl_interface': 6.802023467987631e-07, 'learning_rate/action_model': 6.802023467987631e-06, 'epoch': 1.77}) +[2026-06-23 23:13:34,812][starVLA.training.train_starvla][INFO] - Step 4570, Loss: {'train/loss': 0.0428985171020031, 'train/grad_norm_pre_clip': 16.377628326416016, 'timing/data': 0.0005837259814143181, 'timing/model': 1.6870159429963678, 'learning_rate/qwen_vl_interface': 6.793724080994884e-07, 'learning_rate/action_model': 6.793724080994884e-06, 'epoch': 1.77}) +[2026-06-23 23:13:38,167][starVLA.training.train_starvla][INFO] - Step 4571, Loss: {'train/loss': 0.15947465598583221, 'train/grad_norm_pre_clip': 24.121305465698242, 'timing/data': 0.0005907588638365269, 'timing/model': 1.6914840079843998, 'learning_rate/qwen_vl_interface': 6.785443482125693e-07, 'learning_rate/action_model': 6.7854434821256915e-06, 'epoch': 1.77}) +[2026-06-23 23:13:41,503][starVLA.training.train_starvla][INFO] - Step 4572, Loss: {'train/loss': 0.034914642572402954, 'train/grad_norm_pre_clip': 25.477365493774414, 'timing/data': 0.0005567851476371288, 'timing/model': 1.6890111109241843, 'learning_rate/qwen_vl_interface': 6.777181674783895e-07, 'learning_rate/action_model': 6.777181674783895e-06, 'epoch': 1.77}) +[2026-06-23 23:13:44,824][starVLA.training.train_starvla][INFO] - Step 4573, Loss: {'train/loss': 0.1005847230553627, 'train/grad_norm_pre_clip': 82.05660247802734, 'timing/data': 0.0005689579993486404, 'timing/model': 1.6855299009475857, 'learning_rate/qwen_vl_interface': 6.76893866236561e-07, 'learning_rate/action_model': 6.7689386623656094e-06, 'epoch': 1.77}) +[2026-06-23 23:13:48,143][starVLA.training.train_starvla][INFO] - Step 4574, Loss: {'train/loss': 0.1260266900062561, 'train/grad_norm_pre_clip': 32.5700798034668, 'timing/data': 0.0005759720224887133, 'timing/model': 1.6872888791840523, 'learning_rate/qwen_vl_interface': 6.76071444825923e-07, 'learning_rate/action_model': 6.76071444825923e-06, 'epoch': 1.77}) +[2026-06-23 23:13:51,467][starVLA.training.train_starvla][INFO] - Step 4575, Loss: {'train/loss': 0.0690789595246315, 'train/grad_norm_pre_clip': 49.246341705322266, 'timing/data': 0.0005629470106214285, 'timing/model': 1.6892723399214447, 'learning_rate/qwen_vl_interface': 6.752509035845413e-07, 'learning_rate/action_model': 6.7525090358454126e-06, 'epoch': 1.77}) +[2026-06-23 23:13:54,788][starVLA.training.train_starvla][INFO] - Step 4576, Loss: {'train/loss': 0.07418651133775711, 'train/grad_norm_pre_clip': 38.19924545288086, 'timing/data': 0.0005550819914788008, 'timing/model': 1.6860685700085014, 'learning_rate/qwen_vl_interface': 6.744322428497095e-07, 'learning_rate/action_model': 6.744322428497095e-06, 'epoch': 1.77}) +[2026-06-23 23:13:58,117][starVLA.training.train_starvla][INFO] - Step 4577, Loss: {'train/loss': 0.08578372746706009, 'train/grad_norm_pre_clip': 28.961254119873047, 'timing/data': 0.0005956690292805433, 'timing/model': 1.6908355860505253, 'learning_rate/qwen_vl_interface': 6.736154629579487e-07, 'learning_rate/action_model': 6.736154629579487e-06, 'epoch': 1.77}) +[2026-06-23 23:14:01,438][starVLA.training.train_starvla][INFO] - Step 4578, Loss: {'train/loss': 0.07590913772583008, 'train/grad_norm_pre_clip': 17.297311782836914, 'timing/data': 0.0005794479511678219, 'timing/model': 1.68532641697675, 'learning_rate/qwen_vl_interface': 6.728005642450055e-07, 'learning_rate/action_model': 6.7280056424500554e-06, 'epoch': 1.77}) +[2026-06-23 23:14:04,756][starVLA.training.train_starvla][INFO] - Step 4579, Loss: {'train/loss': 0.09428703039884567, 'train/grad_norm_pre_clip': 19.254844665527344, 'timing/data': 0.000586590962484479, 'timing/model': 1.6812399001792073, 'learning_rate/qwen_vl_interface': 6.719875470458553e-07, 'learning_rate/action_model': 6.719875470458553e-06, 'epoch': 1.77}) +[2026-06-23 23:14:08,080][starVLA.training.train_starvla][INFO] - Step 4580, Loss: {'train/loss': 0.07790505141019821, 'train/grad_norm_pre_clip': 24.452850341796875, 'timing/data': 0.0005688080564141273, 'timing/model': 1.687618331052363, 'learning_rate/qwen_vl_interface': 6.711764116946972e-07, 'learning_rate/action_model': 6.7117641169469715e-06, 'epoch': 1.77}) +[2026-06-23 23:14:11,394][starVLA.training.train_starvla][INFO] - Step 4581, Loss: {'train/loss': 0.05410149693489075, 'train/grad_norm_pre_clip': 37.83258819580078, 'timing/data': 0.00055354880169034, 'timing/model': 1.6796537118498236, 'learning_rate/qwen_vl_interface': 6.703671585249587e-07, 'learning_rate/action_model': 6.703671585249586e-06, 'epoch': 1.77}) +[2026-06-23 23:14:14,801][starVLA.training.train_starvla][INFO] - Step 4582, Loss: {'train/loss': 0.08300291746854782, 'train/grad_norm_pre_clip': 26.264509201049805, 'timing/data': 0.0005946171004325151, 'timing/model': 1.6858332201372832, 'learning_rate/qwen_vl_interface': 6.69559787869294e-07, 'learning_rate/action_model': 6.69559787869294e-06, 'epoch': 1.77}) +[2026-06-23 23:14:18,124][starVLA.training.train_starvla][INFO] - Step 4583, Loss: {'train/loss': 0.11890172213315964, 'train/grad_norm_pre_clip': 39.10643005371094, 'timing/data': 0.0005516959354281425, 'timing/model': 1.689147116849199, 'learning_rate/qwen_vl_interface': 6.68754300059582e-07, 'learning_rate/action_model': 6.687543000595819e-06, 'epoch': 1.77}) +[2026-06-23 23:14:21,443][starVLA.training.train_starvla][INFO] - Step 4584, Loss: {'train/loss': 0.05868842080235481, 'train/grad_norm_pre_clip': 14.922727584838867, 'timing/data': 0.0005672951228916645, 'timing/model': 1.6880785760004073, 'learning_rate/qwen_vl_interface': 6.67950695426929e-07, 'learning_rate/action_model': 6.679506954269291e-06, 'epoch': 1.77}) +[2026-06-23 23:14:24,762][starVLA.training.train_starvla][INFO] - Step 4585, Loss: {'train/loss': 0.17228466272354126, 'train/grad_norm_pre_clip': 32.331642150878906, 'timing/data': 0.0005603320896625519, 'timing/model': 1.6860955271404237, 'learning_rate/qwen_vl_interface': 6.671489743016651e-07, 'learning_rate/action_model': 6.671489743016651e-06, 'epoch': 1.77}) +[2026-06-23 23:14:28,081][starVLA.training.train_starvla][INFO] - Step 4586, Loss: {'train/loss': 0.14178848266601562, 'train/grad_norm_pre_clip': 14.15504264831543, 'timing/data': 0.0005541110876947641, 'timing/model': 1.685668810037896, 'learning_rate/qwen_vl_interface': 6.663491370133487e-07, 'learning_rate/action_model': 6.663491370133487e-06, 'epoch': 1.77}) +[2026-06-23 23:14:31,409][starVLA.training.train_starvla][INFO] - Step 4587, Loss: {'train/loss': 0.07955886423587799, 'train/grad_norm_pre_clip': 34.15338897705078, 'timing/data': 0.0005921320989727974, 'timing/model': 1.6871842620894313, 'learning_rate/qwen_vl_interface': 6.655511838907629e-07, 'learning_rate/action_model': 6.655511838907629e-06, 'epoch': 1.78}) +[2026-06-23 23:14:34,728][starVLA.training.train_starvla][INFO] - Step 4588, Loss: {'train/loss': 0.12429683655500412, 'train/grad_norm_pre_clip': 19.686729431152344, 'timing/data': 0.0005872028414160013, 'timing/model': 1.6856897789984941, 'learning_rate/qwen_vl_interface': 6.647551152619149e-07, 'learning_rate/action_model': 6.647551152619148e-06, 'epoch': 1.78}) +[2026-06-23 23:14:38,049][starVLA.training.train_starvla][INFO] - Step 4589, Loss: {'train/loss': 0.12843111157417297, 'train/grad_norm_pre_clip': 26.227264404296875, 'timing/data': 0.0005744288209825754, 'timing/model': 1.6849908609874547, 'learning_rate/qwen_vl_interface': 6.639609314540396e-07, 'learning_rate/action_model': 6.639609314540396e-06, 'epoch': 1.78}) +[2026-06-23 23:14:41,365][starVLA.training.train_starvla][INFO] - Step 4590, Loss: {'train/loss': 0.07186667621135712, 'train/grad_norm_pre_clip': 15.165873527526855, 'timing/data': 0.0005421680398285389, 'timing/model': 1.6873399671167135, 'learning_rate/qwen_vl_interface': 6.631686327935954e-07, 'learning_rate/action_model': 6.631686327935954e-06, 'epoch': 1.78}) +[2026-06-23 23:14:44,690][starVLA.training.train_starvla][INFO] - Step 4591, Loss: {'train/loss': 0.04001254960894585, 'train/grad_norm_pre_clip': 17.45502471923828, 'timing/data': 0.0005489608738571405, 'timing/model': 1.6906188891734928, 'learning_rate/qwen_vl_interface': 6.623782196062665e-07, 'learning_rate/action_model': 6.623782196062664e-06, 'epoch': 1.78}) +[2026-06-23 23:14:48,019][starVLA.training.train_starvla][INFO] - Step 4592, Loss: {'train/loss': 0.09208543598651886, 'train/grad_norm_pre_clip': 49.84103775024414, 'timing/data': 0.0005700599867850542, 'timing/model': 1.6913203799631447, 'learning_rate/qwen_vl_interface': 6.615896922169622e-07, 'learning_rate/action_model': 6.615896922169622e-06, 'epoch': 1.78}) +[2026-06-23 23:14:51,342][starVLA.training.train_starvla][INFO] - Step 4593, Loss: {'train/loss': 0.08773135393857956, 'train/grad_norm_pre_clip': 21.987606048583984, 'timing/data': 0.0005728860851377249, 'timing/model': 1.685094793094322, 'learning_rate/qwen_vl_interface': 6.608030509498156e-07, 'learning_rate/action_model': 6.608030509498156e-06, 'epoch': 1.78}) +[2026-06-23 23:14:54,668][starVLA.training.train_starvla][INFO] - Step 4594, Loss: {'train/loss': 0.06525000929832458, 'train/grad_norm_pre_clip': 19.48552131652832, 'timing/data': 0.0005452539771795273, 'timing/model': 1.6907410151325166, 'learning_rate/qwen_vl_interface': 6.600182961281865e-07, 'learning_rate/action_model': 6.600182961281864e-06, 'epoch': 1.78}) +[2026-06-23 23:14:57,991][starVLA.training.train_starvla][INFO] - Step 4595, Loss: {'train/loss': 0.0561622716486454, 'train/grad_norm_pre_clip': 13.11662483215332, 'timing/data': 0.000667863991111517, 'timing/model': 1.6865320999640971, 'learning_rate/qwen_vl_interface': 6.592354280746561e-07, 'learning_rate/action_model': 6.5923542807465595e-06, 'epoch': 1.78}) +[2026-06-23 23:15:01,320][starVLA.training.train_starvla][INFO] - Step 4596, Loss: {'train/loss': 0.09875769913196564, 'train/grad_norm_pre_clip': 28.25117301940918, 'timing/data': 0.0005799289792776108, 'timing/model': 1.6903633680194616, 'learning_rate/qwen_vl_interface': 6.584544471110335e-07, 'learning_rate/action_model': 6.584544471110335e-06, 'epoch': 1.78}) +[2026-06-23 23:15:04,799][starVLA.training.train_starvla][INFO] - Step 4597, Loss: {'train/loss': 0.04219534248113632, 'train/grad_norm_pre_clip': 13.224226951599121, 'timing/data': 0.000582883832976222, 'timing/model': 1.685822464991361, 'learning_rate/qwen_vl_interface': 6.576753535583495e-07, 'learning_rate/action_model': 6.576753535583495e-06, 'epoch': 1.78}) +[2026-06-23 23:15:08,119][starVLA.training.train_starvla][INFO] - Step 4598, Loss: {'train/loss': 0.06741131097078323, 'train/grad_norm_pre_clip': 13.04037094116211, 'timing/data': 0.0005755710881203413, 'timing/model': 1.6847141981124878, 'learning_rate/qwen_vl_interface': 6.5689814773686e-07, 'learning_rate/action_model': 6.568981477368599e-06, 'epoch': 1.78}) +[2026-06-23 23:15:11,445][starVLA.training.train_starvla][INFO] - Step 4599, Loss: {'train/loss': 0.08865653723478317, 'train/grad_norm_pre_clip': 14.553756713867188, 'timing/data': 0.0005429799202829599, 'timing/model': 1.6878568730317056, 'learning_rate/qwen_vl_interface': 6.56122829966046e-07, 'learning_rate/action_model': 6.56122829966046e-06, 'epoch': 1.78}) +[2026-06-23 23:15:14,763][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4600 +[2026-06-23 23:16:46,855][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4600 +[2026-06-23 23:16:46,855][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4600 +[2026-06-23 23:17:57,289][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4600 +[2026-06-23 23:17:57,290][starVLA.training.train_starvla][INFO] - Step 4600, Loss: {'train/loss': 0.08993463963270187, 'train/grad_norm_pre_clip': 13.00318717956543, 'eval/action_loss/seconds': 92.08965525985695, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06501748412847519, 'eval/latency_2/loss': 0.06501748412847519, 'eval/flappy/loss': 0.06501748412847519, 'eval/flappy/latency_2/loss': 0.06501748412847519, 'eval/flap_precision': 0.9348556077904634, 'eval/flap_recall': 0.9392712550607287, 'eval/flap_f1': 0.9370582295523392, 'eval/flap_rate_model': 0.20680555555555555, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9370582295523393, 'eval/latency_2/flap_precision': 0.9348556077904634, 'eval/latency_2/flap_recall': 0.9392712550607287, 'eval/latency_2/flap_f1': 0.9370582295523392, 'eval/latency_2/flap_rate_model': 0.20680555555555555, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9370582295523393, 'eval/action_classification/data_wait_seconds': 0.9383082985877991, 'eval/action_classification/predict_seconds': 69.36094665527344, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005550917703658342, 'timing/model': 1.686845256946981, 'learning_rate/qwen_vl_interface': 6.553494005646101e-07, 'learning_rate/action_model': 6.553494005646101e-06, 'epoch': 1.78}) +[2026-06-23 23:18:00,606][starVLA.training.train_starvla][INFO] - Step 4601, Loss: {'train/loss': 0.046435240656137466, 'train/grad_norm_pre_clip': 17.235231399536133, 'timing/data': 0.0005463662091642618, 'timing/model': 1.6855237858835608, 'learning_rate/qwen_vl_interface': 6.545778598504807e-07, 'learning_rate/action_model': 6.545778598504807e-06, 'epoch': 1.78}) +[2026-06-23 23:18:03,914][starVLA.training.train_starvla][INFO] - Step 4602, Loss: {'train/loss': 0.1462073177099228, 'train/grad_norm_pre_clip': 31.352291107177734, 'timing/data': 0.0005730860866606236, 'timing/model': 1.6809789589606225, 'learning_rate/qwen_vl_interface': 6.538082081408093e-07, 'learning_rate/action_model': 6.538082081408093e-06, 'epoch': 1.78}) +[2026-06-23 23:18:07,217][starVLA.training.train_starvla][INFO] - Step 4603, Loss: {'train/loss': 0.0654233992099762, 'train/grad_norm_pre_clip': 15.775812149047852, 'timing/data': 0.0005591700319200754, 'timing/model': 1.679472206858918, 'learning_rate/qwen_vl_interface': 6.530404457519697e-07, 'learning_rate/action_model': 6.530404457519697e-06, 'epoch': 1.78}) +[2026-06-23 23:18:10,523][starVLA.training.train_starvla][INFO] - Step 4604, Loss: {'train/loss': 0.14038816094398499, 'train/grad_norm_pre_clip': 24.375226974487305, 'timing/data': 0.0005631579551845789, 'timing/model': 1.6759476480074227, 'learning_rate/qwen_vl_interface': 6.522745729995614e-07, 'learning_rate/action_model': 6.522745729995613e-06, 'epoch': 1.78}) +[2026-06-23 23:18:13,941][starVLA.training.train_starvla][INFO] - Step 4605, Loss: {'train/loss': 0.10055413097143173, 'train/grad_norm_pre_clip': 32.15478515625, 'timing/data': 0.0005700811743736267, 'timing/model': 1.6822129550855607, 'learning_rate/qwen_vl_interface': 6.515105901984037e-07, 'learning_rate/action_model': 6.515105901984037e-06, 'epoch': 1.78}) +[2026-06-23 23:18:17,253][starVLA.training.train_starvla][INFO] - Step 4606, Loss: {'train/loss': 0.11608891189098358, 'train/grad_norm_pre_clip': 51.849422454833984, 'timing/data': 0.0006352229975163937, 'timing/model': 1.6816881590057164, 'learning_rate/qwen_vl_interface': 6.50748497662543e-07, 'learning_rate/action_model': 6.507484976625429e-06, 'epoch': 1.78}) +[2026-06-23 23:18:20,565][starVLA.training.train_starvla][INFO] - Step 4607, Loss: {'train/loss': 0.07226685434579849, 'train/grad_norm_pre_clip': 13.174633979797363, 'timing/data': 0.0006309850141406059, 'timing/model': 1.6830810338724405, 'learning_rate/qwen_vl_interface': 6.499882957052459e-07, 'learning_rate/action_model': 6.4998829570524586e-06, 'epoch': 1.78}) +[2026-06-23 23:18:23,876][starVLA.training.train_starvla][INFO] - Step 4608, Loss: {'train/loss': 0.1002873107790947, 'train/grad_norm_pre_clip': 33.64353561401367, 'timing/data': 0.000568547984585166, 'timing/model': 1.6850058978889138, 'learning_rate/qwen_vl_interface': 6.492299846390024e-07, 'learning_rate/action_model': 6.492299846390023e-06, 'epoch': 1.78}) +[2026-06-23 23:18:27,192][starVLA.training.train_starvla][INFO] - Step 4609, Loss: {'train/loss': 0.10960204154253006, 'train/grad_norm_pre_clip': 32.48153305053711, 'timing/data': 0.000559621024876833, 'timing/model': 1.682945327134803, 'learning_rate/qwen_vl_interface': 6.484735647755259e-07, 'learning_rate/action_model': 6.48473564775526e-06, 'epoch': 1.78}) +[2026-06-23 23:18:30,505][starVLA.training.train_starvla][INFO] - Step 4610, Loss: {'train/loss': 0.14194709062576294, 'train/grad_norm_pre_clip': 15.850604057312012, 'timing/data': 0.0005629970692098141, 'timing/model': 1.683560332050547, 'learning_rate/qwen_vl_interface': 6.477190364257524e-07, 'learning_rate/action_model': 6.477190364257524e-06, 'epoch': 1.78}) +[2026-06-23 23:18:33,822][starVLA.training.train_starvla][INFO] - Step 4611, Loss: {'train/loss': 0.05502593144774437, 'train/grad_norm_pre_clip': 41.40830612182617, 'timing/data': 0.0005858708173036575, 'timing/model': 1.6841400999110192, 'learning_rate/qwen_vl_interface': 6.469663998998393e-07, 'learning_rate/action_model': 6.469663998998393e-06, 'epoch': 1.78}) +[2026-06-23 23:18:37,142][starVLA.training.train_starvla][INFO] - Step 4612, Loss: {'train/loss': 0.10444643348455429, 'train/grad_norm_pre_clip': 21.408817291259766, 'timing/data': 0.0006930509116500616, 'timing/model': 1.6854954140726477, 'learning_rate/qwen_vl_interface': 6.462156555071681e-07, 'learning_rate/action_model': 6.46215655507168e-06, 'epoch': 1.78}) +[2026-06-23 23:18:40,458][starVLA.training.train_starvla][INFO] - Step 4613, Loss: {'train/loss': 0.07228139787912369, 'train/grad_norm_pre_clip': 49.03090286254883, 'timing/data': 0.0005799389909952879, 'timing/model': 1.684486169135198, 'learning_rate/qwen_vl_interface': 6.454668035563401e-07, 'learning_rate/action_model': 6.454668035563401e-06, 'epoch': 1.79}) +[2026-06-23 23:18:43,781][starVLA.training.train_starvla][INFO] - Step 4614, Loss: {'train/loss': 0.13167999684810638, 'train/grad_norm_pre_clip': 22.014694213867188, 'timing/data': 0.000624573091045022, 'timing/model': 1.6899992329999804, 'learning_rate/qwen_vl_interface': 6.447198443551809e-07, 'learning_rate/action_model': 6.4471984435518075e-06, 'epoch': 1.79}) +[2026-06-23 23:18:47,100][starVLA.training.train_starvla][INFO] - Step 4615, Loss: {'train/loss': 0.0820157527923584, 'train/grad_norm_pre_clip': 31.663389205932617, 'timing/data': 0.0008908228483051062, 'timing/model': 1.6849094710778445, 'learning_rate/qwen_vl_interface': 6.439747782107373e-07, 'learning_rate/action_model': 6.4397477821073735e-06, 'epoch': 1.79}) +[2026-06-23 23:18:50,415][starVLA.training.train_starvla][INFO] - Step 4616, Loss: {'train/loss': 0.14897708594799042, 'train/grad_norm_pre_clip': 61.529319763183594, 'timing/data': 0.0006896548438817263, 'timing/model': 1.6833912378642708, 'learning_rate/qwen_vl_interface': 6.432316054292779e-07, 'learning_rate/action_model': 6.43231605429278e-06, 'epoch': 1.79}) +[2026-06-23 23:18:53,735][starVLA.training.train_starvla][INFO] - Step 4617, Loss: {'train/loss': 0.03746934235095978, 'train/grad_norm_pre_clip': 19.135995864868164, 'timing/data': 0.0006383289583027363, 'timing/model': 1.686420856975019, 'learning_rate/qwen_vl_interface': 6.424903263162924e-07, 'learning_rate/action_model': 6.424903263162923e-06, 'epoch': 1.79}) +[2026-06-23 23:18:57,053][starVLA.training.train_starvla][INFO] - Step 4618, Loss: {'train/loss': 0.12931165099143982, 'train/grad_norm_pre_clip': 43.854766845703125, 'timing/data': 0.0006228398997336626, 'timing/model': 1.6859480068087578, 'learning_rate/qwen_vl_interface': 6.417509411764924e-07, 'learning_rate/action_model': 6.4175094117649244e-06, 'epoch': 1.79}) +[2026-06-23 23:19:00,374][starVLA.training.train_starvla][INFO] - Step 4619, Loss: {'train/loss': 0.14025327563285828, 'train/grad_norm_pre_clip': 25.771495819091797, 'timing/data': 0.000593886012211442, 'timing/model': 1.687510111136362, 'learning_rate/qwen_vl_interface': 6.410134503138122e-07, 'learning_rate/action_model': 6.4101345031381224e-06, 'epoch': 1.79}) +[2026-06-23 23:19:03,843][starVLA.training.train_starvla][INFO] - Step 4620, Loss: {'train/loss': 0.034106165170669556, 'train/grad_norm_pre_clip': 54.00157928466797, 'timing/data': 0.000679215881973505, 'timing/model': 1.6839924121741205, 'learning_rate/qwen_vl_interface': 6.402778540314054e-07, 'learning_rate/action_model': 6.402778540314054e-06, 'epoch': 1.79}) +[2026-06-23 23:19:07,162][starVLA.training.train_starvla][INFO] - Step 4621, Loss: {'train/loss': 0.11812029778957367, 'train/grad_norm_pre_clip': 13.55309009552002, 'timing/data': 0.0006446309853345156, 'timing/model': 1.6862816100474447, 'learning_rate/qwen_vl_interface': 6.39544152631649e-07, 'learning_rate/action_model': 6.3954415263164895e-06, 'epoch': 1.79}) +[2026-06-23 23:19:10,481][starVLA.training.train_starvla][INFO] - Step 4622, Loss: {'train/loss': 0.1131213903427124, 'train/grad_norm_pre_clip': 18.056777954101562, 'timing/data': 0.0006288220174610615, 'timing/model': 1.6856440310366452, 'learning_rate/qwen_vl_interface': 6.388123464161388e-07, 'learning_rate/action_model': 6.388123464161389e-06, 'epoch': 1.79}) +[2026-06-23 23:19:13,795][starVLA.training.train_starvla][INFO] - Step 4623, Loss: {'train/loss': 0.10048256069421768, 'train/grad_norm_pre_clip': 74.96434020996094, 'timing/data': 0.0006131620611995459, 'timing/model': 1.6854269420728087, 'learning_rate/qwen_vl_interface': 6.380824356856936e-07, 'learning_rate/action_model': 6.380824356856936e-06, 'epoch': 1.79}) +[2026-06-23 23:19:17,111][starVLA.training.train_starvla][INFO] - Step 4624, Loss: {'train/loss': 0.13514144718647003, 'train/grad_norm_pre_clip': 63.28374481201172, 'timing/data': 0.0006234520114958286, 'timing/model': 1.6836203669663519, 'learning_rate/qwen_vl_interface': 6.373544207403513e-07, 'learning_rate/action_model': 6.373544207403513e-06, 'epoch': 1.79}) +[2026-06-23 23:19:20,427][starVLA.training.train_starvla][INFO] - Step 4625, Loss: {'train/loss': 0.08524090051651001, 'train/grad_norm_pre_clip': 17.801034927368164, 'timing/data': 0.000623681116849184, 'timing/model': 1.6843922359403223, 'learning_rate/qwen_vl_interface': 6.366283018793723e-07, 'learning_rate/action_model': 6.366283018793723e-06, 'epoch': 1.79}) +[2026-06-23 23:19:23,743][starVLA.training.train_starvla][INFO] - Step 4626, Loss: {'train/loss': 0.1220962330698967, 'train/grad_norm_pre_clip': 31.036231994628906, 'timing/data': 0.0005693191196769476, 'timing/model': 1.6850031812209636, 'learning_rate/qwen_vl_interface': 6.359040794012358e-07, 'learning_rate/action_model': 6.359040794012358e-06, 'epoch': 1.79}) +[2026-06-23 23:19:27,059][starVLA.training.train_starvla][INFO] - Step 4627, Loss: {'train/loss': 0.0785236656665802, 'train/grad_norm_pre_clip': 23.111330032348633, 'timing/data': 0.000599816907197237, 'timing/model': 1.6857943651266396, 'learning_rate/qwen_vl_interface': 6.351817536036431e-07, 'learning_rate/action_model': 6.351817536036431e-06, 'epoch': 1.79}) +[2026-06-23 23:19:30,375][starVLA.training.train_starvla][INFO] - Step 4628, Loss: {'train/loss': 0.07284795492887497, 'train/grad_norm_pre_clip': 22.32637596130371, 'timing/data': 0.0005639791488647461, 'timing/model': 1.685410653008148, 'learning_rate/qwen_vl_interface': 6.344613247835135e-07, 'learning_rate/action_model': 6.344613247835135e-06, 'epoch': 1.79}) +[2026-06-23 23:19:33,699][starVLA.training.train_starvla][INFO] - Step 4629, Loss: {'train/loss': 0.05767611414194107, 'train/grad_norm_pre_clip': 10.032222747802734, 'timing/data': 0.0006155171431601048, 'timing/model': 1.685964070027694, 'learning_rate/qwen_vl_interface': 6.337427932369898e-07, 'learning_rate/action_model': 6.337427932369898e-06, 'epoch': 1.79}) +[2026-06-23 23:19:37,017][starVLA.training.train_starvla][INFO] - Step 4630, Loss: {'train/loss': 0.0677502304315567, 'train/grad_norm_pre_clip': 17.858001708984375, 'timing/data': 0.0005517259705811739, 'timing/model': 1.68545638397336, 'learning_rate/qwen_vl_interface': 6.330261592594328e-07, 'learning_rate/action_model': 6.330261592594327e-06, 'epoch': 1.79}) +[2026-06-23 23:19:40,333][starVLA.training.train_starvla][INFO] - Step 4631, Loss: {'train/loss': 0.2465009242296219, 'train/grad_norm_pre_clip': 49.618408203125, 'timing/data': 0.0007297599222511053, 'timing/model': 1.685546562075615, 'learning_rate/qwen_vl_interface': 6.323114231454229e-07, 'learning_rate/action_model': 6.323114231454229e-06, 'epoch': 1.79}) +[2026-06-23 23:19:43,658][starVLA.training.train_starvla][INFO] - Step 4632, Loss: {'train/loss': 0.1434098780155182, 'train/grad_norm_pre_clip': 16.790334701538086, 'timing/data': 0.0005380101501941681, 'timing/model': 1.6874579400755465, 'learning_rate/qwen_vl_interface': 6.315985851887618e-07, 'learning_rate/action_model': 6.315985851887618e-06, 'epoch': 1.79}) +[2026-06-23 23:19:46,982][starVLA.training.train_starvla][INFO] - Step 4633, Loss: {'train/loss': 0.024241022765636444, 'train/grad_norm_pre_clip': 16.1171932220459, 'timing/data': 0.0005560040008276701, 'timing/model': 1.687474749982357, 'learning_rate/qwen_vl_interface': 6.308876456824697e-07, 'learning_rate/action_model': 6.3088764568246975e-06, 'epoch': 1.79}) +[2026-06-23 23:19:50,432][starVLA.training.train_starvla][INFO] - Step 4634, Loss: {'train/loss': 0.13753513991832733, 'train/grad_norm_pre_clip': 14.194538116455078, 'timing/data': 0.000607580877840519, 'timing/model': 1.819735649973154, 'learning_rate/qwen_vl_interface': 6.301786049187875e-07, 'learning_rate/action_model': 6.3017860491878745e-06, 'epoch': 1.79}) +[2026-06-23 23:19:53,749][starVLA.training.train_starvla][INFO] - Step 4635, Loss: {'train/loss': 0.20974338054656982, 'train/grad_norm_pre_clip': 18.758892059326172, 'timing/data': 0.0006202149670571089, 'timing/model': 1.6866922359913588, 'learning_rate/qwen_vl_interface': 6.294714631891755e-07, 'learning_rate/action_model': 6.294714631891754e-06, 'epoch': 1.79}) +[2026-06-23 23:19:57,069][starVLA.training.train_starvla][INFO] - Step 4636, Loss: {'train/loss': 0.2047528624534607, 'train/grad_norm_pre_clip': 38.373592376708984, 'timing/data': 0.0005433100741356611, 'timing/model': 1.688601149013266, 'learning_rate/qwen_vl_interface': 6.287662207843115e-07, 'learning_rate/action_model': 6.287662207843114e-06, 'epoch': 1.79}) +[2026-06-23 23:20:00,395][starVLA.training.train_starvla][INFO] - Step 4637, Loss: {'train/loss': 0.08021136373281479, 'train/grad_norm_pre_clip': 42.83698654174805, 'timing/data': 0.0005504637956619263, 'timing/model': 1.6870314788538963, 'learning_rate/qwen_vl_interface': 6.280628779940956e-07, 'learning_rate/action_model': 6.280628779940956e-06, 'epoch': 1.79}) +[2026-06-23 23:20:03,715][starVLA.training.train_starvla][INFO] - Step 4638, Loss: {'train/loss': 0.08860905468463898, 'train/grad_norm_pre_clip': 19.5460147857666, 'timing/data': 0.0007422140333801508, 'timing/model': 1.6877473201602697, 'learning_rate/qwen_vl_interface': 6.273614351076445e-07, 'learning_rate/action_model': 6.273614351076445e-06, 'epoch': 1.79}) +[2026-06-23 23:20:07,032][starVLA.training.train_starvla][INFO] - Step 4639, Loss: {'train/loss': 0.11999962478876114, 'train/grad_norm_pre_clip': 15.193964958190918, 'timing/data': 0.0005706117954105139, 'timing/model': 1.686067852890119, 'learning_rate/qwen_vl_interface': 6.266618924132951e-07, 'learning_rate/action_model': 6.266618924132951e-06, 'epoch': 1.8}) +[2026-06-23 23:20:10,350][starVLA.training.train_starvla][INFO] - Step 4640, Loss: {'train/loss': 0.11291643977165222, 'train/grad_norm_pre_clip': 20.24484634399414, 'timing/data': 0.0005677668377757072, 'timing/model': 1.6880580289289355, 'learning_rate/qwen_vl_interface': 6.25964250198603e-07, 'learning_rate/action_model': 6.2596425019860294e-06, 'epoch': 1.8}) +[2026-06-23 23:20:13,674][starVLA.training.train_starvla][INFO] - Step 4641, Loss: {'train/loss': 0.1649729311466217, 'train/grad_norm_pre_clip': 39.410823822021484, 'timing/data': 0.0005517660174518824, 'timing/model': 1.6907128437887877, 'learning_rate/qwen_vl_interface': 6.25268508750342e-07, 'learning_rate/action_model': 6.25268508750342e-06, 'epoch': 1.8}) +[2026-06-23 23:20:16,996][starVLA.training.train_starvla][INFO] - Step 4642, Loss: {'train/loss': 0.1656721532344818, 'train/grad_norm_pre_clip': 30.375083923339844, 'timing/data': 0.0005588599015027285, 'timing/model': 1.6869119808543473, 'learning_rate/qwen_vl_interface': 6.245746683545062e-07, 'learning_rate/action_model': 6.245746683545062e-06, 'epoch': 1.8}) +[2026-06-23 23:20:20,319][starVLA.training.train_starvla][INFO] - Step 4643, Loss: {'train/loss': 0.14502547681331635, 'train/grad_norm_pre_clip': 16.910734176635742, 'timing/data': 0.0007169761229306459, 'timing/model': 1.6870323051698506, 'learning_rate/qwen_vl_interface': 6.238827292963063e-07, 'learning_rate/action_model': 6.238827292963063e-06, 'epoch': 1.8}) +[2026-06-23 23:20:23,637][starVLA.training.train_starvla][INFO] - Step 4644, Loss: {'train/loss': 0.12804129719734192, 'train/grad_norm_pre_clip': 30.661497116088867, 'timing/data': 0.0005790379364043474, 'timing/model': 1.6860517610330135, 'learning_rate/qwen_vl_interface': 6.231926918601731e-07, 'learning_rate/action_model': 6.2319269186017305e-06, 'epoch': 1.8}) +[2026-06-23 23:20:26,957][starVLA.training.train_starvla][INFO] - Step 4645, Loss: {'train/loss': 0.1454712450504303, 'train/grad_norm_pre_clip': 22.736616134643555, 'timing/data': 0.0006448519416153431, 'timing/model': 1.6864659460261464, 'learning_rate/qwen_vl_interface': 6.22504556329754e-07, 'learning_rate/action_model': 6.22504556329754e-06, 'epoch': 1.8}) +[2026-06-23 23:20:30,277][starVLA.training.train_starvla][INFO] - Step 4646, Loss: {'train/loss': 0.08265481889247894, 'train/grad_norm_pre_clip': 41.31633758544922, 'timing/data': 0.0006645980756729841, 'timing/model': 1.683964563999325, 'learning_rate/qwen_vl_interface': 6.218183229879164e-07, 'learning_rate/action_model': 6.218183229879164e-06, 'epoch': 1.8}) +[2026-06-23 23:20:33,598][starVLA.training.train_starvla][INFO] - Step 4647, Loss: {'train/loss': 0.1849558800458908, 'train/grad_norm_pre_clip': 42.137332916259766, 'timing/data': 0.0005627770442515612, 'timing/model': 1.6883906859438866, 'learning_rate/qwen_vl_interface': 6.211339921167447e-07, 'learning_rate/action_model': 6.211339921167446e-06, 'epoch': 1.8}) +[2026-06-23 23:20:36,923][starVLA.training.train_starvla][INFO] - Step 4648, Loss: {'train/loss': 0.11679526418447495, 'train/grad_norm_pre_clip': 14.219254493713379, 'timing/data': 0.0005838670767843723, 'timing/model': 1.6876888149417937, 'learning_rate/qwen_vl_interface': 6.204515639975415e-07, 'learning_rate/action_model': 6.204515639975414e-06, 'epoch': 1.8}) +[2026-06-23 23:20:40,333][starVLA.training.train_starvla][INFO] - Step 4649, Loss: {'train/loss': 0.048928748816251755, 'train/grad_norm_pre_clip': 16.61815643310547, 'timing/data': 0.0005675759166479111, 'timing/model': 1.772612038999796, 'learning_rate/qwen_vl_interface': 6.197710389108262e-07, 'learning_rate/action_model': 6.197710389108262e-06, 'epoch': 1.8}) +[2026-06-23 23:20:43,656][starVLA.training.train_starvla][INFO] - Step 4650, Loss: {'train/loss': 0.07652683556079865, 'train/grad_norm_pre_clip': 11.410754203796387, 'timing/data': 0.0008100720588117838, 'timing/model': 1.6871127670165151, 'learning_rate/qwen_vl_interface': 6.190924171363378e-07, 'learning_rate/action_model': 6.190924171363378e-06, 'epoch': 1.8}) +[2026-06-23 23:20:46,982][starVLA.training.train_starvla][INFO] - Step 4651, Loss: {'train/loss': 0.10585975646972656, 'train/grad_norm_pre_clip': 16.3182373046875, 'timing/data': 0.0005926541052758694, 'timing/model': 1.688030747929588, 'learning_rate/qwen_vl_interface': 6.184156989530326e-07, 'learning_rate/action_model': 6.184156989530325e-06, 'epoch': 1.8}) +[2026-06-23 23:20:50,317][starVLA.training.train_starvla][INFO] - Step 4652, Loss: {'train/loss': 0.03271631523966789, 'train/grad_norm_pre_clip': 16.77789878845215, 'timing/data': 0.0005850892048329115, 'timing/model': 1.7036702840123326, 'learning_rate/qwen_vl_interface': 6.17740884639083e-07, 'learning_rate/action_model': 6.177408846390829e-06, 'epoch': 1.8}) +[2026-06-23 23:20:53,637][starVLA.training.train_starvla][INFO] - Step 4653, Loss: {'train/loss': 0.0998530387878418, 'train/grad_norm_pre_clip': 29.075565338134766, 'timing/data': 0.000555993989109993, 'timing/model': 1.6844954809639603, 'learning_rate/qwen_vl_interface': 6.170679744718798e-07, 'learning_rate/action_model': 6.1706797447187975e-06, 'epoch': 1.8}) +[2026-06-23 23:20:56,958][starVLA.training.train_starvla][INFO] - Step 4654, Loss: {'train/loss': 0.05849863961338997, 'train/grad_norm_pre_clip': 31.708805084228516, 'timing/data': 0.0005615651607513428, 'timing/model': 1.6873475259635597, 'learning_rate/qwen_vl_interface': 6.163969687280311e-07, 'learning_rate/action_model': 6.163969687280311e-06, 'epoch': 1.8}) +[2026-06-23 23:21:00,276][starVLA.training.train_starvla][INFO] - Step 4655, Loss: {'train/loss': 0.20943132042884827, 'train/grad_norm_pre_clip': 94.40782928466797, 'timing/data': 0.0006840950809419155, 'timing/model': 1.6830068880226463, 'learning_rate/qwen_vl_interface': 6.157278676833627e-07, 'learning_rate/action_model': 6.157278676833627e-06, 'epoch': 1.8}) +[2026-06-23 23:21:03,594][starVLA.training.train_starvla][INFO] - Step 4656, Loss: {'train/loss': 0.04864609241485596, 'train/grad_norm_pre_clip': 12.41706657409668, 'timing/data': 0.0006179008632898331, 'timing/model': 1.683216511970386, 'learning_rate/qwen_vl_interface': 6.150606716129155e-07, 'learning_rate/action_model': 6.150606716129154e-06, 'epoch': 1.8}) +[2026-06-23 23:21:06,917][starVLA.training.train_starvla][INFO] - Step 4657, Loss: {'train/loss': 0.10711237788200378, 'train/grad_norm_pre_clip': 48.35973358154297, 'timing/data': 0.0005459550302475691, 'timing/model': 1.6886434010230005, 'learning_rate/qwen_vl_interface': 6.143953807909497e-07, 'learning_rate/action_model': 6.143953807909496e-06, 'epoch': 1.8}) +[2026-06-23 23:21:10,241][starVLA.training.train_starvla][INFO] - Step 4658, Loss: {'train/loss': 0.054445430636405945, 'train/grad_norm_pre_clip': 29.022396087646484, 'timing/data': 0.0006506720092147589, 'timing/model': 1.6881340900436044, 'learning_rate/qwen_vl_interface': 6.137319954909407e-07, 'learning_rate/action_model': 6.137319954909406e-06, 'epoch': 1.8}) +[2026-06-23 23:21:13,556][starVLA.training.train_starvla][INFO] - Step 4659, Loss: {'train/loss': 0.041604164987802505, 'train/grad_norm_pre_clip': 21.400102615356445, 'timing/data': 0.00064481096342206, 'timing/model': 1.6857549969572574, 'learning_rate/qwen_vl_interface': 6.130705159855815e-07, 'learning_rate/action_model': 6.130705159855814e-06, 'epoch': 1.8}) +[2026-06-23 23:21:16,876][starVLA.training.train_starvla][INFO] - Step 4660, Loss: {'train/loss': 0.11437143385410309, 'train/grad_norm_pre_clip': 33.65677261352539, 'timing/data': 0.0007724009919911623, 'timing/model': 1.6858873628079891, 'learning_rate/qwen_vl_interface': 6.124109425467805e-07, 'learning_rate/action_model': 6.124109425467805e-06, 'epoch': 1.8}) +[2026-06-23 23:21:20,208][starVLA.training.train_starvla][INFO] - Step 4661, Loss: {'train/loss': 0.1812973916530609, 'train/grad_norm_pre_clip': 35.785400390625, 'timing/data': 0.0006732649635523558, 'timing/model': 1.6865761729422957, 'learning_rate/qwen_vl_interface': 6.117532754456642e-07, 'learning_rate/action_model': 6.117532754456642e-06, 'epoch': 1.8}) +[2026-06-23 23:21:23,529][starVLA.training.train_starvla][INFO] - Step 4662, Loss: {'train/loss': 0.11182129383087158, 'train/grad_norm_pre_clip': 31.812864303588867, 'timing/data': 0.000572986202314496, 'timing/model': 1.688137382036075, 'learning_rate/qwen_vl_interface': 6.110975149525746e-07, 'learning_rate/action_model': 6.110975149525746e-06, 'epoch': 1.8}) +[2026-06-23 23:21:26,847][starVLA.training.train_starvla][INFO] - Step 4663, Loss: {'train/loss': 0.0957145169377327, 'train/grad_norm_pre_clip': 25.914548873901367, 'timing/data': 0.0005701810587197542, 'timing/model': 1.685010866029188, 'learning_rate/qwen_vl_interface': 6.104436613370692e-07, 'learning_rate/action_model': 6.104436613370692e-06, 'epoch': 1.8}) +[2026-06-23 23:21:30,265][starVLA.training.train_starvla][INFO] - Step 4664, Loss: {'train/loss': 0.16815704107284546, 'train/grad_norm_pre_clip': 26.012380599975586, 'timing/data': 0.0005535001400858164, 'timing/model': 1.6855568380560726, 'learning_rate/qwen_vl_interface': 6.09791714867924e-07, 'learning_rate/action_model': 6.0979171486792396e-06, 'epoch': 1.81}) +[2026-06-23 23:21:33,593][starVLA.training.train_starvla][INFO] - Step 4665, Loss: {'train/loss': 0.19494612514972687, 'train/grad_norm_pre_clip': 23.160839080810547, 'timing/data': 0.0005554039962589741, 'timing/model': 1.6909167128615081, 'learning_rate/qwen_vl_interface': 6.091416758131287e-07, 'learning_rate/action_model': 6.091416758131286e-06, 'epoch': 1.81}) +[2026-06-23 23:21:36,911][starVLA.training.train_starvla][INFO] - Step 4666, Loss: {'train/loss': 0.09967724978923798, 'train/grad_norm_pre_clip': 34.5745735168457, 'timing/data': 0.000634301919490099, 'timing/model': 1.6867340728640556, 'learning_rate/qwen_vl_interface': 6.084935444398895e-07, 'learning_rate/action_model': 6.084935444398894e-06, 'epoch': 1.81}) +[2026-06-23 23:21:40,228][starVLA.training.train_starvla][INFO] - Step 4667, Loss: {'train/loss': 0.07279003411531448, 'train/grad_norm_pre_clip': 22.3585147857666, 'timing/data': 0.0005794980097562075, 'timing/model': 1.6857545380480587, 'learning_rate/qwen_vl_interface': 6.078473210146294e-07, 'learning_rate/action_model': 6.078473210146294e-06, 'epoch': 1.81}) +[2026-06-23 23:21:43,549][starVLA.training.train_starvla][INFO] - Step 4668, Loss: {'train/loss': 0.07080330699682236, 'train/grad_norm_pre_clip': 19.03723907470703, 'timing/data': 0.0005671849939972162, 'timing/model': 1.6884052238892764, 'learning_rate/qwen_vl_interface': 6.072030058029862e-07, 'learning_rate/action_model': 6.072030058029861e-06, 'epoch': 1.81}) +[2026-06-23 23:21:46,867][starVLA.training.train_starvla][INFO] - Step 4669, Loss: {'train/loss': 0.07122670114040375, 'train/grad_norm_pre_clip': 19.746231079101562, 'timing/data': 0.0005802900996059179, 'timing/model': 1.6869362709112465, 'learning_rate/qwen_vl_interface': 6.065605990698136e-07, 'learning_rate/action_model': 6.065605990698136e-06, 'epoch': 1.81}) +[2026-06-23 23:21:50,193][starVLA.training.train_starvla][INFO] - Step 4670, Loss: {'train/loss': 0.21179847419261932, 'train/grad_norm_pre_clip': 28.678388595581055, 'timing/data': 0.0005835159681737423, 'timing/model': 1.6938319311011583, 'learning_rate/qwen_vl_interface': 6.059201010791811e-07, 'learning_rate/action_model': 6.059201010791811e-06, 'epoch': 1.81}) +[2026-06-23 23:21:53,514][starVLA.training.train_starvla][INFO] - Step 4671, Loss: {'train/loss': 0.06119377538561821, 'train/grad_norm_pre_clip': 11.57860279083252, 'timing/data': 0.0005672159604728222, 'timing/model': 1.6871164119802415, 'learning_rate/qwen_vl_interface': 6.052815120943721e-07, 'learning_rate/action_model': 6.052815120943721e-06, 'epoch': 1.81}) +[2026-06-23 23:21:56,832][starVLA.training.train_starvla][INFO] - Step 4672, Loss: {'train/loss': 0.09021033346652985, 'train/grad_norm_pre_clip': 53.597347259521484, 'timing/data': 0.000567516079172492, 'timing/model': 1.6857546290848404, 'learning_rate/qwen_vl_interface': 6.046448323778874e-07, 'learning_rate/action_model': 6.046448323778874e-06, 'epoch': 1.81}) +[2026-06-23 23:22:00,152][starVLA.training.train_starvla][INFO] - Step 4673, Loss: {'train/loss': 0.161148339509964, 'train/grad_norm_pre_clip': 19.69276237487793, 'timing/data': 0.0005477790255099535, 'timing/model': 1.6870585458818823, 'learning_rate/qwen_vl_interface': 6.040100621914419e-07, 'learning_rate/action_model': 6.040100621914418e-06, 'epoch': 1.81}) +[2026-06-23 23:22:03,473][starVLA.training.train_starvla][INFO] - Step 4674, Loss: {'train/loss': 0.10703390836715698, 'train/grad_norm_pre_clip': 39.10014724731445, 'timing/data': 0.0005591001827269793, 'timing/model': 1.6868921781424433, 'learning_rate/qwen_vl_interface': 6.033772017959652e-07, 'learning_rate/action_model': 6.033772017959652e-06, 'epoch': 1.81}) +[2026-06-23 23:22:06,794][starVLA.training.train_starvla][INFO] - Step 4675, Loss: {'train/loss': 0.171910360455513, 'train/grad_norm_pre_clip': 22.44011878967285, 'timing/data': 0.0005683780182152987, 'timing/model': 1.6869510051328689, 'learning_rate/qwen_vl_interface': 6.027462514516022e-07, 'learning_rate/action_model': 6.0274625145160214e-06, 'epoch': 1.81}) +[2026-06-23 23:22:10,112][starVLA.training.train_starvla][INFO] - Step 4676, Loss: {'train/loss': 0.19155417382717133, 'train/grad_norm_pre_clip': 66.80723571777344, 'timing/data': 0.0005723051726818085, 'timing/model': 1.6870868860278279, 'learning_rate/qwen_vl_interface': 6.021172114177139e-07, 'learning_rate/action_model': 6.021172114177139e-06, 'epoch': 1.81}) +[2026-06-23 23:22:13,434][starVLA.training.train_starvla][INFO] - Step 4677, Loss: {'train/loss': 0.08800879120826721, 'train/grad_norm_pre_clip': 25.108747482299805, 'timing/data': 0.0005581679288297892, 'timing/model': 1.691590952919796, 'learning_rate/qwen_vl_interface': 6.014900819528734e-07, 'learning_rate/action_model': 6.0149008195287335e-06, 'epoch': 1.81}) +[2026-06-23 23:22:16,759][starVLA.training.train_starvla][INFO] - Step 4678, Loss: {'train/loss': 0.10774026066064835, 'train/grad_norm_pre_clip': 24.726930618286133, 'timing/data': 0.0005578380078077316, 'timing/model': 1.6886595729738474, 'learning_rate/qwen_vl_interface': 6.008648633148707e-07, 'learning_rate/action_model': 6.008648633148707e-06, 'epoch': 1.81}) +[2026-06-23 23:22:20,192][starVLA.training.train_starvla][INFO] - Step 4679, Loss: {'train/loss': 0.17400549352169037, 'train/grad_norm_pre_clip': 25.78154754638672, 'timing/data': 0.0005765729583799839, 'timing/model': 1.6880442430265248, 'learning_rate/qwen_vl_interface': 6.002415557607097e-07, 'learning_rate/action_model': 6.002415557607096e-06, 'epoch': 1.81}) +[2026-06-23 23:22:23,510][starVLA.training.train_starvla][INFO] - Step 4680, Loss: {'train/loss': 0.07029472291469574, 'train/grad_norm_pre_clip': 21.342674255371094, 'timing/data': 0.0005676860455423594, 'timing/model': 1.686508493963629, 'learning_rate/qwen_vl_interface': 5.996201595466088e-07, 'learning_rate/action_model': 5.996201595466088e-06, 'epoch': 1.81}) +[2026-06-23 23:22:26,829][starVLA.training.train_starvla][INFO] - Step 4681, Loss: {'train/loss': 0.02758190594613552, 'train/grad_norm_pre_clip': 18.81682777404785, 'timing/data': 0.0005716939922422171, 'timing/model': 1.6875783691648394, 'learning_rate/qwen_vl_interface': 5.990006749279998e-07, 'learning_rate/action_model': 5.990006749279998e-06, 'epoch': 1.81}) +[2026-06-23 23:22:30,154][starVLA.training.train_starvla][INFO] - Step 4682, Loss: {'train/loss': 0.10666706413030624, 'train/grad_norm_pre_clip': 20.116260528564453, 'timing/data': 0.0007807270158082247, 'timing/model': 1.6889905980788171, 'learning_rate/qwen_vl_interface': 5.983831021595296e-07, 'learning_rate/action_model': 5.9838310215952956e-06, 'epoch': 1.81}) +[2026-06-23 23:22:33,479][starVLA.training.train_starvla][INFO] - Step 4683, Loss: {'train/loss': 0.06113952398300171, 'train/grad_norm_pre_clip': 22.762258529663086, 'timing/data': 0.0005614438559859991, 'timing/model': 1.6898372038267553, 'learning_rate/qwen_vl_interface': 5.977674414950595e-07, 'learning_rate/action_model': 5.9776744149505955e-06, 'epoch': 1.81}) +[2026-06-23 23:22:36,794][starVLA.training.train_starvla][INFO] - Step 4684, Loss: {'train/loss': 0.21284528076648712, 'train/grad_norm_pre_clip': 28.22197723388672, 'timing/data': 0.0005428299773484468, 'timing/model': 1.6873397279996425, 'learning_rate/qwen_vl_interface': 5.971536931876635e-07, 'learning_rate/action_model': 5.971536931876634e-06, 'epoch': 1.81}) +[2026-06-23 23:22:40,114][starVLA.training.train_starvla][INFO] - Step 4685, Loss: {'train/loss': 0.09025420993566513, 'train/grad_norm_pre_clip': 20.248186111450195, 'timing/data': 0.0005963000003248453, 'timing/model': 1.6888850568793714, 'learning_rate/qwen_vl_interface': 5.965418574896315e-07, 'learning_rate/action_model': 5.9654185748963145e-06, 'epoch': 1.81}) +[2026-06-23 23:22:43,439][starVLA.training.train_starvla][INFO] - Step 4686, Loss: {'train/loss': 0.17496779561042786, 'train/grad_norm_pre_clip': 21.77298355102539, 'timing/data': 0.0005481299012899399, 'timing/model': 1.688344356836751, 'learning_rate/qwen_vl_interface': 5.959319346524646e-07, 'learning_rate/action_model': 5.959319346524646e-06, 'epoch': 1.81}) +[2026-06-23 23:22:46,760][starVLA.training.train_starvla][INFO] - Step 4687, Loss: {'train/loss': 0.106387659907341, 'train/grad_norm_pre_clip': 18.56180191040039, 'timing/data': 0.0006615931633859873, 'timing/model': 1.684580750996247, 'learning_rate/qwen_vl_interface': 5.953239249268804e-07, 'learning_rate/action_model': 5.953239249268804e-06, 'epoch': 1.81}) +[2026-06-23 23:22:50,080][starVLA.training.train_starvla][INFO] - Step 4688, Loss: {'train/loss': 0.10372716188430786, 'train/grad_norm_pre_clip': 28.976722717285156, 'timing/data': 0.0006333901546895504, 'timing/model': 1.6858201958239079, 'learning_rate/qwen_vl_interface': 5.947178285628081e-07, 'learning_rate/action_model': 5.94717828562808e-06, 'epoch': 1.81}) +[2026-06-23 23:22:53,395][starVLA.training.train_starvla][INFO] - Step 4689, Loss: {'train/loss': 0.06552047282457352, 'train/grad_norm_pre_clip': 13.09703540802002, 'timing/data': 0.0007671411149203777, 'timing/model': 1.6867326851934195, 'learning_rate/qwen_vl_interface': 5.941136458093907e-07, 'learning_rate/action_model': 5.941136458093907e-06, 'epoch': 1.81}) +[2026-06-23 23:22:56,713][starVLA.training.train_starvla][INFO] - Step 4690, Loss: {'train/loss': 0.13491462171077728, 'train/grad_norm_pre_clip': 34.19086456298828, 'timing/data': 0.0005738581530749798, 'timing/model': 1.6854286808520555, 'learning_rate/qwen_vl_interface': 5.935113769149855e-07, 'learning_rate/action_model': 5.935113769149856e-06, 'epoch': 1.82}) +[2026-06-23 23:23:00,033][starVLA.training.train_starvla][INFO] - Step 4691, Loss: {'train/loss': 0.05316364765167236, 'train/grad_norm_pre_clip': 25.826738357543945, 'timing/data': 0.000552718061953783, 'timing/model': 1.6882658961694688, 'learning_rate/qwen_vl_interface': 5.92911022127163e-07, 'learning_rate/action_model': 5.929110221271629e-06, 'epoch': 1.82}) +[2026-06-23 23:23:03,353][starVLA.training.train_starvla][INFO] - Step 4692, Loss: {'train/loss': 0.10598160326480865, 'train/grad_norm_pre_clip': 18.510963439941406, 'timing/data': 0.0005813019815832376, 'timing/model': 1.6870330658275634, 'learning_rate/qwen_vl_interface': 5.92312581692705e-07, 'learning_rate/action_model': 5.923125816927049e-06, 'epoch': 1.82}) +[2026-06-23 23:23:06,757][starVLA.training.train_starvla][INFO] - Step 4693, Loss: {'train/loss': 0.14137601852416992, 'train/grad_norm_pre_clip': 54.36721420288086, 'timing/data': 0.0005750099662691355, 'timing/model': 1.7717885880265385, 'learning_rate/qwen_vl_interface': 5.917160558576088e-07, 'learning_rate/action_model': 5.917160558576088e-06, 'epoch': 1.82}) +[2026-06-23 23:23:10,079][starVLA.training.train_starvla][INFO] - Step 4694, Loss: {'train/loss': 0.07253730297088623, 'train/grad_norm_pre_clip': 14.19543170928955, 'timing/data': 0.0005737680476158857, 'timing/model': 1.6872344850562513, 'learning_rate/qwen_vl_interface': 5.911214448670829e-07, 'learning_rate/action_model': 5.911214448670828e-06, 'epoch': 1.82}) +[2026-06-23 23:23:13,400][starVLA.training.train_starvla][INFO] - Step 4695, Loss: {'train/loss': 0.08775857836008072, 'train/grad_norm_pre_clip': 20.40269660949707, 'timing/data': 0.0005476290825754404, 'timing/model': 1.684628915041685, 'learning_rate/qwen_vl_interface': 5.905287489655499e-07, 'learning_rate/action_model': 5.905287489655499e-06, 'epoch': 1.82}) +[2026-06-23 23:23:16,728][starVLA.training.train_starvla][INFO] - Step 4696, Loss: {'train/loss': 0.13981734216213226, 'train/grad_norm_pre_clip': 46.986995697021484, 'timing/data': 0.0005833359900861979, 'timing/model': 1.6961017039138824, 'learning_rate/qwen_vl_interface': 5.899379683966445e-07, 'learning_rate/action_model': 5.899379683966445e-06, 'epoch': 1.82}) +[2026-06-23 23:23:20,046][starVLA.training.train_starvla][INFO] - Step 4697, Loss: {'train/loss': 0.13946859538555145, 'train/grad_norm_pre_clip': 19.385950088500977, 'timing/data': 0.000583115965127945, 'timing/model': 1.6851003021001816, 'learning_rate/qwen_vl_interface': 5.893491034032142e-07, 'learning_rate/action_model': 5.893491034032142e-06, 'epoch': 1.82}) +[2026-06-23 23:23:23,365][starVLA.training.train_starvla][INFO] - Step 4698, Loss: {'train/loss': 0.24145235121250153, 'train/grad_norm_pre_clip': 19.698530197143555, 'timing/data': 0.0005766940303146839, 'timing/model': 1.6878339019604027, 'learning_rate/qwen_vl_interface': 5.887621542273196e-07, 'learning_rate/action_model': 5.887621542273196e-06, 'epoch': 1.82}) +[2026-06-23 23:23:26,684][starVLA.training.train_starvla][INFO] - Step 4699, Loss: {'train/loss': 0.21223770081996918, 'train/grad_norm_pre_clip': 26.0142822265625, 'timing/data': 0.0005738078616559505, 'timing/model': 1.68558022310026, 'learning_rate/qwen_vl_interface': 5.881771211102327e-07, 'learning_rate/action_model': 5.881771211102327e-06, 'epoch': 1.82}) +[2026-06-23 23:23:30,005][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4700 +[2026-06-23 23:25:01,369][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4700 +[2026-06-23 23:25:01,369][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4700 +[2026-06-23 23:26:12,261][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4700 +[2026-06-23 23:26:12,262][starVLA.training.train_starvla][INFO] - Step 4700, Loss: {'train/loss': 0.04369558393955231, 'train/grad_norm_pre_clip': 15.418947219848633, 'eval/action_loss/seconds': 91.36097247386351, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06522886455059052, 'eval/latency_2/loss': 0.06522886455059052, 'eval/flappy/loss': 0.06522886455059052, 'eval/flappy/latency_2/loss': 0.06522886455059052, 'eval/flap_precision': 0.9321808510638298, 'eval/flap_recall': 0.9460188933873145, 'eval/flap_f1': 0.9390488948425988, 'eval/flap_rate_model': 0.2088888888888889, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9390488948425988, 'eval/latency_2/flap_precision': 0.9321808510638298, 'eval/latency_2/flap_recall': 0.9460188933873145, 'eval/latency_2/flap_f1': 0.9390488948425988, 'eval/latency_2/flap_rate_model': 0.2088888888888889, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9390488948425988, 'eval/action_classification/data_wait_seconds': 1.039236307144165, 'eval/action_classification/predict_seconds': 69.70941925048828, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005489520262926817, 'timing/model': 1.6881444959435612, 'learning_rate/qwen_vl_interface': 5.875940042924391e-07, 'learning_rate/action_model': 5.875940042924391e-06, 'epoch': 1.82}) +[2026-06-23 23:26:15,742][starVLA.training.train_starvla][INFO] - Step 4701, Loss: {'train/loss': 0.2164950966835022, 'train/grad_norm_pre_clip': 13.811018943786621, 'timing/data': 0.0006427471525967121, 'timing/model': 1.8561603610869497, 'learning_rate/qwen_vl_interface': 5.870128040136368e-07, 'learning_rate/action_model': 5.870128040136367e-06, 'epoch': 1.82}) +[2026-06-23 23:26:19,053][starVLA.training.train_starvla][INFO] - Step 4702, Loss: {'train/loss': 0.2136278748512268, 'train/grad_norm_pre_clip': 18.962295532226562, 'timing/data': 0.0005871939938515425, 'timing/model': 1.682980784913525, 'learning_rate/qwen_vl_interface': 5.864335205127332e-07, 'learning_rate/action_model': 5.864335205127332e-06, 'epoch': 1.82}) +[2026-06-23 23:26:22,361][starVLA.training.train_starvla][INFO] - Step 4703, Loss: {'train/loss': 0.17966888844966888, 'train/grad_norm_pre_clip': 37.29389572143555, 'timing/data': 0.0005677971057593822, 'timing/model': 1.681528931017965, 'learning_rate/qwen_vl_interface': 5.858561540278513e-07, 'learning_rate/action_model': 5.858561540278513e-06, 'epoch': 1.82}) +[2026-06-23 23:26:25,674][starVLA.training.train_starvla][INFO] - Step 4704, Loss: {'train/loss': 0.10255078226327896, 'train/grad_norm_pre_clip': 12.667470932006836, 'timing/data': 0.0005555939860641956, 'timing/model': 1.6817507310770452, 'learning_rate/qwen_vl_interface': 5.852807047963244e-07, 'learning_rate/action_model': 5.852807047963243e-06, 'epoch': 1.82}) +[2026-06-23 23:26:29,005][starVLA.training.train_starvla][INFO] - Step 4705, Loss: {'train/loss': 0.15118670463562012, 'train/grad_norm_pre_clip': 37.852298736572266, 'timing/data': 0.0005531290080398321, 'timing/model': 1.7039665740448982, 'learning_rate/qwen_vl_interface': 5.847071730546974e-07, 'learning_rate/action_model': 5.847071730546974e-06, 'epoch': 1.82}) +[2026-06-23 23:26:32,320][starVLA.training.train_starvla][INFO] - Step 4706, Loss: {'train/loss': 0.07429376244544983, 'train/grad_norm_pre_clip': 20.009170532226562, 'timing/data': 0.000584387918934226, 'timing/model': 1.6841411888599396, 'learning_rate/qwen_vl_interface': 5.841355590387291e-07, 'learning_rate/action_model': 5.841355590387291e-06, 'epoch': 1.82}) +[2026-06-23 23:26:35,631][starVLA.training.train_starvla][INFO] - Step 4707, Loss: {'train/loss': 0.15727317333221436, 'train/grad_norm_pre_clip': 30.780542373657227, 'timing/data': 0.0006596690509468317, 'timing/model': 1.682436027796939, 'learning_rate/qwen_vl_interface': 5.835658629833866e-07, 'learning_rate/action_model': 5.8356586298338654e-06, 'epoch': 1.82}) +[2026-06-23 23:26:38,946][starVLA.training.train_starvla][INFO] - Step 4708, Loss: {'train/loss': 0.23758815228939056, 'train/grad_norm_pre_clip': 17.966691970825195, 'timing/data': 0.0005724960938096046, 'timing/model': 1.682584190974012, 'learning_rate/qwen_vl_interface': 5.829980851228512e-07, 'learning_rate/action_model': 5.829980851228512e-06, 'epoch': 1.82}) +[2026-06-23 23:26:42,258][starVLA.training.train_starvla][INFO] - Step 4709, Loss: {'train/loss': 0.15963585674762726, 'train/grad_norm_pre_clip': 34.11533737182617, 'timing/data': 0.0005860410165041685, 'timing/model': 1.6805045371875167, 'learning_rate/qwen_vl_interface': 5.824322256905155e-07, 'learning_rate/action_model': 5.824322256905154e-06, 'epoch': 1.82}) +[2026-06-23 23:26:45,578][starVLA.training.train_starvla][INFO] - Step 4710, Loss: {'train/loss': 0.10769118368625641, 'train/grad_norm_pre_clip': 38.14039611816406, 'timing/data': 0.0005871329922229052, 'timing/model': 1.6880648150108755, 'learning_rate/qwen_vl_interface': 5.818682849189818e-07, 'learning_rate/action_model': 5.818682849189817e-06, 'epoch': 1.82}) +[2026-06-23 23:26:48,902][starVLA.training.train_starvla][INFO] - Step 4711, Loss: {'train/loss': 0.12661966681480408, 'train/grad_norm_pre_clip': 17.91309356689453, 'timing/data': 0.0005784959066659212, 'timing/model': 1.6859838981181383, 'learning_rate/qwen_vl_interface': 5.813062630400653e-07, 'learning_rate/action_model': 5.813062630400653e-06, 'epoch': 1.82}) +[2026-06-23 23:26:52,216][starVLA.training.train_starvla][INFO] - Step 4712, Loss: {'train/loss': 0.11546270549297333, 'train/grad_norm_pre_clip': 20.0234375, 'timing/data': 0.0005697810556739569, 'timing/model': 1.684742759913206, 'learning_rate/qwen_vl_interface': 5.807461602847922e-07, 'learning_rate/action_model': 5.807461602847922e-06, 'epoch': 1.82}) +[2026-06-23 23:26:55,527][starVLA.training.train_starvla][INFO] - Step 4713, Loss: {'train/loss': 0.045132435858249664, 'train/grad_norm_pre_clip': 12.21380615234375, 'timing/data': 0.0005712341517210007, 'timing/model': 1.6804129790980369, 'learning_rate/qwen_vl_interface': 5.801879768833998e-07, 'learning_rate/action_model': 5.801879768833998e-06, 'epoch': 1.82}) +[2026-06-23 23:26:58,846][starVLA.training.train_starvla][INFO] - Step 4714, Loss: {'train/loss': 0.10128483921289444, 'train/grad_norm_pre_clip': 17.470895767211914, 'timing/data': 0.0005678969901055098, 'timing/model': 1.685862981947139, 'learning_rate/qwen_vl_interface': 5.796317130653363e-07, 'learning_rate/action_model': 5.796317130653363e-06, 'epoch': 1.82}) +[2026-06-23 23:27:02,202][starVLA.training.train_starvla][INFO] - Step 4715, Loss: {'train/loss': 0.10670332610607147, 'train/grad_norm_pre_clip': 33.586341857910156, 'timing/data': 0.0005628881044685841, 'timing/model': 1.6854276868980378, 'learning_rate/qwen_vl_interface': 5.790773690592592e-07, 'learning_rate/action_model': 5.7907736905925925e-06, 'epoch': 1.82}) +[2026-06-23 23:27:05,677][starVLA.training.train_starvla][INFO] - Step 4716, Loss: {'train/loss': 0.08892502635717392, 'train/grad_norm_pre_clip': 8.031096458435059, 'timing/data': 0.0005797799676656723, 'timing/model': 1.687551731010899, 'learning_rate/qwen_vl_interface': 5.7852494509304e-07, 'learning_rate/action_model': 5.7852494509304e-06, 'epoch': 1.83}) +[2026-06-23 23:27:08,996][starVLA.training.train_starvla][INFO] - Step 4717, Loss: {'train/loss': 0.0953083410859108, 'train/grad_norm_pre_clip': 12.995129585266113, 'timing/data': 0.0006572150159627199, 'timing/model': 1.6834534350782633, 'learning_rate/qwen_vl_interface': 5.779744413937592e-07, 'learning_rate/action_model': 5.779744413937592e-06, 'epoch': 1.83}) +[2026-06-23 23:27:12,318][starVLA.training.train_starvla][INFO] - Step 4718, Loss: {'train/loss': 0.10840623825788498, 'train/grad_norm_pre_clip': 23.03187370300293, 'timing/data': 0.0006286408752202988, 'timing/model': 1.6855908741708845, 'learning_rate/qwen_vl_interface': 5.774258581877073e-07, 'learning_rate/action_model': 5.774258581877073e-06, 'epoch': 1.83}) +[2026-06-23 23:27:15,637][starVLA.training.train_starvla][INFO] - Step 4719, Loss: {'train/loss': 0.1774323284626007, 'train/grad_norm_pre_clip': 17.73509979248047, 'timing/data': 0.0005867930594831705, 'timing/model': 1.6855259949807078, 'learning_rate/qwen_vl_interface': 5.768791957003866e-07, 'learning_rate/action_model': 5.768791957003865e-06, 'epoch': 1.83}) +[2026-06-23 23:27:18,960][starVLA.training.train_starvla][INFO] - Step 4720, Loss: {'train/loss': 0.07509198784828186, 'train/grad_norm_pre_clip': 13.584145545959473, 'timing/data': 0.0005862920079380274, 'timing/model': 1.6887642329093069, 'learning_rate/qwen_vl_interface': 5.763344541565091e-07, 'learning_rate/action_model': 5.7633445415650905e-06, 'epoch': 1.83}) +[2026-06-23 23:27:22,283][starVLA.training.train_starvla][INFO] - Step 4721, Loss: {'train/loss': 0.16703951358795166, 'train/grad_norm_pre_clip': 17.343473434448242, 'timing/data': 0.0005632478278130293, 'timing/model': 1.6863253430929035, 'learning_rate/qwen_vl_interface': 5.75791633779998e-07, 'learning_rate/action_model': 5.75791633779998e-06, 'epoch': 1.83}) +[2026-06-23 23:27:25,604][starVLA.training.train_starvla][INFO] - Step 4722, Loss: {'train/loss': 0.04182109981775284, 'train/grad_norm_pre_clip': 9.855237007141113, 'timing/data': 0.0005809820722788572, 'timing/model': 1.6865968171041459, 'learning_rate/qwen_vl_interface': 5.752507347939857e-07, 'learning_rate/action_model': 5.752507347939858e-06, 'epoch': 1.83}) +[2026-06-23 23:27:28,926][starVLA.training.train_starvla][INFO] - Step 4723, Loss: {'train/loss': 0.23999954760074615, 'train/grad_norm_pre_clip': 25.69905662536621, 'timing/data': 0.0005705209914594889, 'timing/model': 1.687900949968025, 'learning_rate/qwen_vl_interface': 5.74711757420816e-07, 'learning_rate/action_model': 5.747117574208159e-06, 'epoch': 1.83}) +[2026-06-23 23:27:32,246][starVLA.training.train_starvla][INFO] - Step 4724, Loss: {'train/loss': 0.13211211562156677, 'train/grad_norm_pre_clip': 28.1229248046875, 'timing/data': 0.0005589600186794996, 'timing/model': 1.6876902170479298, 'learning_rate/qwen_vl_interface': 5.74174701882042e-07, 'learning_rate/action_model': 5.74174701882042e-06, 'epoch': 1.83}) +[2026-06-23 23:27:35,569][starVLA.training.train_starvla][INFO] - Step 4725, Loss: {'train/loss': 0.10317326337099075, 'train/grad_norm_pre_clip': 26.428077697753906, 'timing/data': 0.0005477790255099535, 'timing/model': 1.6883170378860086, 'learning_rate/qwen_vl_interface': 5.736395683984262e-07, 'learning_rate/action_model': 5.736395683984261e-06, 'epoch': 1.83}) +[2026-06-23 23:27:38,897][starVLA.training.train_starvla][INFO] - Step 4726, Loss: {'train/loss': 0.06069627404212952, 'train/grad_norm_pre_clip': 37.40501403808594, 'timing/data': 0.0006550408434122801, 'timing/model': 1.6921197460033, 'learning_rate/qwen_vl_interface': 5.731063571899421e-07, 'learning_rate/action_model': 5.731063571899421e-06, 'epoch': 1.83}) +[2026-06-23 23:27:42,227][starVLA.training.train_starvla][INFO] - Step 4727, Loss: {'train/loss': 0.06654252856969833, 'train/grad_norm_pre_clip': 40.00923156738281, 'timing/data': 0.00064282794483006, 'timing/model': 1.68825087021105, 'learning_rate/qwen_vl_interface': 5.72575068475773e-07, 'learning_rate/action_model': 5.72575068475773e-06, 'epoch': 1.83}) +[2026-06-23 23:27:45,547][starVLA.training.train_starvla][INFO] - Step 4728, Loss: {'train/loss': 0.12738066911697388, 'train/grad_norm_pre_clip': 19.83553695678711, 'timing/data': 0.0005869229789823294, 'timing/model': 1.6866849220823497, 'learning_rate/qwen_vl_interface': 5.720457024743115e-07, 'learning_rate/action_model': 5.720457024743115e-06, 'epoch': 1.83}) +[2026-06-23 23:27:48,874][starVLA.training.train_starvla][INFO] - Step 4729, Loss: {'train/loss': 0.05066343769431114, 'train/grad_norm_pre_clip': 44.13233184814453, 'timing/data': 0.0005736378952860832, 'timing/model': 1.6920414480846375, 'learning_rate/qwen_vl_interface': 5.715182594031601e-07, 'learning_rate/action_model': 5.7151825940316e-06, 'epoch': 1.83}) +[2026-06-23 23:27:52,324][starVLA.training.train_starvla][INFO] - Step 4730, Loss: {'train/loss': 0.1202927827835083, 'train/grad_norm_pre_clip': 29.200925827026367, 'timing/data': 0.0005631579551845789, 'timing/model': 1.8142550238408148, 'learning_rate/qwen_vl_interface': 5.709927394791305e-07, 'learning_rate/action_model': 5.709927394791305e-06, 'epoch': 1.83}) +[2026-06-23 23:27:55,650][starVLA.training.train_starvla][INFO] - Step 4731, Loss: {'train/loss': 0.155814528465271, 'train/grad_norm_pre_clip': 15.665444374084473, 'timing/data': 0.0005539099220186472, 'timing/model': 1.6889929580502212, 'learning_rate/qwen_vl_interface': 5.70469142918244e-07, 'learning_rate/action_model': 5.704691429182441e-06, 'epoch': 1.83}) +[2026-06-23 23:27:58,969][starVLA.training.train_starvla][INFO] - Step 4732, Loss: {'train/loss': 0.08847744762897491, 'train/grad_norm_pre_clip': 16.183603286743164, 'timing/data': 0.0006504119373857975, 'timing/model': 1.68738334486261, 'learning_rate/qwen_vl_interface': 5.699474699357317e-07, 'learning_rate/action_model': 5.699474699357317e-06, 'epoch': 1.83}) +[2026-06-23 23:28:02,290][starVLA.training.train_starvla][INFO] - Step 4733, Loss: {'train/loss': 0.09663543105125427, 'train/grad_norm_pre_clip': 11.554698944091797, 'timing/data': 0.0006207369733601809, 'timing/model': 1.6870915009640157, 'learning_rate/qwen_vl_interface': 5.694277207460331e-07, 'learning_rate/action_model': 5.694277207460332e-06, 'epoch': 1.83}) +[2026-06-23 23:28:05,610][starVLA.training.train_starvla][INFO] - Step 4734, Loss: {'train/loss': 0.09535472095012665, 'train/grad_norm_pre_clip': 17.343942642211914, 'timing/data': 0.0006715620402246714, 'timing/model': 1.6853699090424925, 'learning_rate/qwen_vl_interface': 5.689098955627982e-07, 'learning_rate/action_model': 5.689098955627982e-06, 'epoch': 1.83}) +[2026-06-23 23:28:08,935][starVLA.training.train_starvla][INFO] - Step 4735, Loss: {'train/loss': 0.09690681844949722, 'train/grad_norm_pre_clip': 39.141666412353516, 'timing/data': 0.0006884639151394367, 'timing/model': 1.6887999367900193, 'learning_rate/qwen_vl_interface': 5.683939945988853e-07, 'learning_rate/action_model': 5.683939945988853e-06, 'epoch': 1.83}) +[2026-06-23 23:28:12,256][starVLA.training.train_starvla][INFO] - Step 4736, Loss: {'train/loss': 0.13717207312583923, 'train/grad_norm_pre_clip': 32.04903030395508, 'timing/data': 0.0006688770372420549, 'timing/model': 1.6877894911449403, 'learning_rate/qwen_vl_interface': 5.678800180663608e-07, 'learning_rate/action_model': 5.678800180663609e-06, 'epoch': 1.83}) +[2026-06-23 23:28:15,586][starVLA.training.train_starvla][INFO] - Step 4737, Loss: {'train/loss': 0.04277709871530533, 'train/grad_norm_pre_clip': 17.29039192199707, 'timing/data': 0.0005815720651298761, 'timing/model': 1.6907595451921225, 'learning_rate/qwen_vl_interface': 5.673679661765018e-07, 'learning_rate/action_model': 5.6736796617650175e-06, 'epoch': 1.83}) +[2026-06-23 23:28:18,909][starVLA.training.train_starvla][INFO] - Step 4738, Loss: {'train/loss': 0.055547378957271576, 'train/grad_norm_pre_clip': 9.55043888092041, 'timing/data': 0.0006399431731551886, 'timing/model': 1.6916509319562465, 'learning_rate/qwen_vl_interface': 5.668578391397938e-07, 'learning_rate/action_model': 5.668578391397937e-06, 'epoch': 1.83}) +[2026-06-23 23:28:22,225][starVLA.training.train_starvla][INFO] - Step 4739, Loss: {'train/loss': 0.09862584620714188, 'train/grad_norm_pre_clip': 20.768362045288086, 'timing/data': 0.0007940120995044708, 'timing/model': 1.6801713958848268, 'learning_rate/qwen_vl_interface': 5.663496371659303e-07, 'learning_rate/action_model': 5.663496371659303e-06, 'epoch': 1.83}) +[2026-06-23 23:28:25,542][starVLA.training.train_starvla][INFO] - Step 4740, Loss: {'train/loss': 0.08684805780649185, 'train/grad_norm_pre_clip': 23.784189224243164, 'timing/data': 0.0005901188123971224, 'timing/model': 1.6860097378958017, 'learning_rate/qwen_vl_interface': 5.658433604638141e-07, 'learning_rate/action_model': 5.65843360463814e-06, 'epoch': 1.83}) +[2026-06-23 23:28:28,862][starVLA.training.train_starvla][INFO] - Step 4741, Loss: {'train/loss': 0.08114149421453476, 'train/grad_norm_pre_clip': 12.043164253234863, 'timing/data': 0.0005665749777108431, 'timing/model': 1.682081128936261, 'learning_rate/qwen_vl_interface': 5.653390092415559e-07, 'learning_rate/action_model': 5.653390092415559e-06, 'epoch': 1.83}) +[2026-06-23 23:28:32,191][starVLA.training.train_starvla][INFO] - Step 4742, Loss: {'train/loss': 0.07531324774026871, 'train/grad_norm_pre_clip': 18.22842788696289, 'timing/data': 0.0005988350603729486, 'timing/model': 1.6917159052100033, 'learning_rate/qwen_vl_interface': 5.648365837064759e-07, 'learning_rate/action_model': 5.648365837064759e-06, 'epoch': 1.84}) +[2026-06-23 23:28:35,515][starVLA.training.train_starvla][INFO] - Step 4743, Loss: {'train/loss': 0.07029999047517776, 'train/grad_norm_pre_clip': 17.58108901977539, 'timing/data': 0.0005623968318104744, 'timing/model': 1.6888861409388483, 'learning_rate/qwen_vl_interface': 5.643360840651025e-07, 'learning_rate/action_model': 5.643360840651025e-06, 'epoch': 1.84}) +[2026-06-23 23:28:38,847][starVLA.training.train_starvla][INFO] - Step 4744, Loss: {'train/loss': 0.09137535095214844, 'train/grad_norm_pre_clip': 18.12328338623047, 'timing/data': 0.0006605619564652443, 'timing/model': 1.690633927937597, 'learning_rate/qwen_vl_interface': 5.638375105231719e-07, 'learning_rate/action_model': 5.638375105231719e-06, 'epoch': 1.84}) +[2026-06-23 23:28:42,267][starVLA.training.train_starvla][INFO] - Step 4745, Loss: {'train/loss': 0.08620612323284149, 'train/grad_norm_pre_clip': 19.601806640625, 'timing/data': 0.0006342409178614616, 'timing/model': 1.789761247811839, 'learning_rate/qwen_vl_interface': 5.633408632856294e-07, 'learning_rate/action_model': 5.6334086328562945e-06, 'epoch': 1.84}) +[2026-06-23 23:28:45,599][starVLA.training.train_starvla][INFO] - Step 4746, Loss: {'train/loss': 0.08691218495368958, 'train/grad_norm_pre_clip': 25.351760864257812, 'timing/data': 0.0005746888928115368, 'timing/model': 1.6893630460835993, 'learning_rate/qwen_vl_interface': 5.628461425566268e-07, 'learning_rate/action_model': 5.6284614255662675e-06, 'epoch': 1.84}) +[2026-06-23 23:28:48,927][starVLA.training.train_starvla][INFO] - Step 4747, Loss: {'train/loss': 0.07711430639028549, 'train/grad_norm_pre_clip': 25.321802139282227, 'timing/data': 0.0005596920382231474, 'timing/model': 1.6901968349702656, 'learning_rate/qwen_vl_interface': 5.62353348539526e-07, 'learning_rate/action_model': 5.623533485395259e-06, 'epoch': 1.84}) +[2026-06-23 23:28:52,250][starVLA.training.train_starvla][INFO] - Step 4748, Loss: {'train/loss': 0.07368820160627365, 'train/grad_norm_pre_clip': 26.353748321533203, 'timing/data': 0.0005967109464108944, 'timing/model': 1.6849762587808073, 'learning_rate/qwen_vl_interface': 5.618624814368956e-07, 'learning_rate/action_model': 5.618624814368956e-06, 'epoch': 1.84}) +[2026-06-23 23:28:55,580][starVLA.training.train_starvla][INFO] - Step 4749, Loss: {'train/loss': 0.16680142283439636, 'train/grad_norm_pre_clip': 23.500713348388672, 'timing/data': 0.0006783250719308853, 'timing/model': 1.688949627103284, 'learning_rate/qwen_vl_interface': 5.613735414505126e-07, 'learning_rate/action_model': 5.613735414505125e-06, 'epoch': 1.84}) +[2026-06-23 23:30:29,062][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4750 +[2026-06-23 23:30:29,063][starVLA.training.train_starvla][INFO] - Step 4750, Loss: {'train/loss': 0.0727052167057991, 'train/grad_norm_pre_clip': 14.865706443786621, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 69.19999999999878, 'rl_games_eval/mid_train/mean_length': 577.0, 'rl_games_eval/mid_train/std_reward': 67.3198336302144, 'rl_games_eval/mid_train/std_length': 543.3985645914056, 'rl_games_eval/mid_train/macro_mean_reward': 69.19999999999878, 'rl_games_eval/mid_train/macro_mean_length': 577.0, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 69.19999999999878, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 577.0, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 67.3198336302144, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 543.3985645914056, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 69.19999999999878, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0006548198871314526, 'timing/model': 1.6893355250358582, 'learning_rate/qwen_vl_interface': 5.608865287813617e-07, 'learning_rate/action_model': 5.608865287813616e-06, 'epoch': 1.84}) +[2026-06-23 23:30:32,346][starVLA.training.train_starvla][INFO] - Step 4751, Loss: {'train/loss': 0.1495775729417801, 'train/grad_norm_pre_clip': 24.04956817626953, 'timing/data': 0.0006267179269343615, 'timing/model': 1.6689932320732623, 'learning_rate/qwen_vl_interface': 5.604014436296362e-07, 'learning_rate/action_model': 5.604014436296362e-06, 'epoch': 1.84}) +[2026-06-23 23:30:35,660][starVLA.training.train_starvla][INFO] - Step 4752, Loss: {'train/loss': 0.14341557025909424, 'train/grad_norm_pre_clip': 42.39606857299805, 'timing/data': 0.0006290520541369915, 'timing/model': 1.6708411159925163, 'learning_rate/qwen_vl_interface': 5.599182861947348e-07, 'learning_rate/action_model': 5.599182861947347e-06, 'epoch': 1.84}) +[2026-06-23 23:30:38,955][starVLA.training.train_starvla][INFO] - Step 4753, Loss: {'train/loss': 0.1380188912153244, 'train/grad_norm_pre_clip': 15.077099800109863, 'timing/data': 0.0006643789820373058, 'timing/model': 1.6721625609789044, 'learning_rate/qwen_vl_interface': 5.594370566752665e-07, 'learning_rate/action_model': 5.5943705667526645e-06, 'epoch': 1.84}) +[2026-06-23 23:30:42,256][starVLA.training.train_starvla][INFO] - Step 4754, Loss: {'train/loss': 0.11102713644504547, 'train/grad_norm_pre_clip': 52.66040802001953, 'timing/data': 0.0005781571380794048, 'timing/model': 1.6828972329385579, 'learning_rate/qwen_vl_interface': 5.589577552690456e-07, 'learning_rate/action_model': 5.589577552690456e-06, 'epoch': 1.84}) +[2026-06-23 23:30:45,787][starVLA.training.train_starvla][INFO] - Step 4755, Loss: {'train/loss': 0.10666727274656296, 'train/grad_norm_pre_clip': 48.24868392944336, 'timing/data': 0.0005626371130347252, 'timing/model': 1.906224224017933, 'learning_rate/qwen_vl_interface': 5.584803821730961e-07, 'learning_rate/action_model': 5.58480382173096e-06, 'epoch': 1.84}) +[2026-06-23 23:30:49,085][starVLA.training.train_starvla][INFO] - Step 4756, Loss: {'train/loss': 0.12984617054462433, 'train/grad_norm_pre_clip': 29.13039207458496, 'timing/data': 0.00054356106556952, 'timing/model': 1.6760006190743297, 'learning_rate/qwen_vl_interface': 5.580049375836466e-07, 'learning_rate/action_model': 5.580049375836466e-06, 'epoch': 1.84}) +[2026-06-23 23:30:52,385][starVLA.training.train_starvla][INFO] - Step 4757, Loss: {'train/loss': 0.12746797502040863, 'train/grad_norm_pre_clip': 22.454038619995117, 'timing/data': 0.0005511960480362177, 'timing/model': 1.6752291219308972, 'learning_rate/qwen_vl_interface': 5.575314216961356e-07, 'learning_rate/action_model': 5.575314216961356e-06, 'epoch': 1.84}) +[2026-06-23 23:30:55,693][starVLA.training.train_starvla][INFO] - Step 4758, Loss: {'train/loss': 0.19991779327392578, 'train/grad_norm_pre_clip': 33.19924545288086, 'timing/data': 0.0007572828326374292, 'timing/model': 1.6817523010540754, 'learning_rate/qwen_vl_interface': 5.570598347052073e-07, 'learning_rate/action_model': 5.570598347052072e-06, 'epoch': 1.84}) +[2026-06-23 23:30:58,998][starVLA.training.train_starvla][INFO] - Step 4759, Loss: {'train/loss': 0.1556544005870819, 'train/grad_norm_pre_clip': 28.118688583374023, 'timing/data': 0.0005905001889914274, 'timing/model': 1.6784098129719496, 'learning_rate/qwen_vl_interface': 5.565901768047126e-07, 'learning_rate/action_model': 5.5659017680471255e-06, 'epoch': 1.84}) +[2026-06-23 23:31:02,311][starVLA.training.train_starvla][INFO] - Step 4760, Loss: {'train/loss': 0.08481413871049881, 'train/grad_norm_pre_clip': 16.435199737548828, 'timing/data': 0.0005777559708803892, 'timing/model': 1.6823773188516498, 'learning_rate/qwen_vl_interface': 5.561224481877105e-07, 'learning_rate/action_model': 5.561224481877105e-06, 'epoch': 1.84}) +[2026-06-23 23:31:05,608][starVLA.training.train_starvla][INFO] - Step 4761, Loss: {'train/loss': 0.2699602246284485, 'train/grad_norm_pre_clip': 45.7982063293457, 'timing/data': 0.0005855299532413483, 'timing/model': 1.6743958950974047, 'learning_rate/qwen_vl_interface': 5.55656649046467e-07, 'learning_rate/action_model': 5.556566490464669e-06, 'epoch': 1.84}) +[2026-06-23 23:31:08,920][starVLA.training.train_starvla][INFO] - Step 4762, Loss: {'train/loss': 0.12966126203536987, 'train/grad_norm_pre_clip': 29.79806137084961, 'timing/data': 0.0005611439701169729, 'timing/model': 1.683004762046039, 'learning_rate/qwen_vl_interface': 5.551927795724544e-07, 'learning_rate/action_model': 5.551927795724543e-06, 'epoch': 1.84}) +[2026-06-23 23:31:12,240][starVLA.training.train_starvla][INFO] - Step 4763, Loss: {'train/loss': 0.059460461139678955, 'train/grad_norm_pre_clip': 14.764610290527344, 'timing/data': 0.0005665142089128494, 'timing/model': 1.6846267920918763, 'learning_rate/qwen_vl_interface': 5.547308399563515e-07, 'learning_rate/action_model': 5.547308399563514e-06, 'epoch': 1.84}) +[2026-06-23 23:31:15,556][starVLA.training.train_starvla][INFO] - Step 4764, Loss: {'train/loss': 0.06140044704079628, 'train/grad_norm_pre_clip': 31.890871047973633, 'timing/data': 0.0005693398416042328, 'timing/model': 1.6863325089216232, 'learning_rate/qwen_vl_interface': 5.542708303880442e-07, 'learning_rate/action_model': 5.542708303880442e-06, 'epoch': 1.84}) +[2026-06-23 23:31:18,862][starVLA.training.train_starvla][INFO] - Step 4765, Loss: {'train/loss': 0.10714221000671387, 'train/grad_norm_pre_clip': 17.52384376525879, 'timing/data': 0.000592272961512208, 'timing/model': 1.6824026070535183, 'learning_rate/qwen_vl_interface': 5.538127510566258e-07, 'learning_rate/action_model': 5.538127510566258e-06, 'epoch': 1.84}) +[2026-06-23 23:31:22,173][starVLA.training.train_starvla][INFO] - Step 4766, Loss: {'train/loss': 0.1481565237045288, 'train/grad_norm_pre_clip': 17.854782104492188, 'timing/data': 0.0005582890007644892, 'timing/model': 1.6829251511953771, 'learning_rate/qwen_vl_interface': 5.533566021503944e-07, 'learning_rate/action_model': 5.533566021503943e-06, 'epoch': 1.84}) +[2026-06-23 23:31:25,486][starVLA.training.train_starvla][INFO] - Step 4767, Loss: {'train/loss': 0.08248700946569443, 'train/grad_norm_pre_clip': 51.59553909301758, 'timing/data': 0.0005603029858320951, 'timing/model': 1.6832776251249015, 'learning_rate/qwen_vl_interface': 5.529023838568566e-07, 'learning_rate/action_model': 5.529023838568566e-06, 'epoch': 1.84}) +[2026-06-23 23:31:28,797][starVLA.training.train_starvla][INFO] - Step 4768, Loss: {'train/loss': 0.13708734512329102, 'train/grad_norm_pre_clip': 22.076234817504883, 'timing/data': 0.0005618860013782978, 'timing/model': 1.6809994468931109, 'learning_rate/qwen_vl_interface': 5.524500963627239e-07, 'learning_rate/action_model': 5.524500963627239e-06, 'epoch': 1.85}) +[2026-06-23 23:31:32,112][starVLA.training.train_starvla][INFO] - Step 4769, Loss: {'train/loss': 0.07892259955406189, 'train/grad_norm_pre_clip': 9.759037017822266, 'timing/data': 0.0006321379914879799, 'timing/model': 1.6875715169589967, 'learning_rate/qwen_vl_interface': 5.519997398539145e-07, 'learning_rate/action_model': 5.519997398539145e-06, 'epoch': 1.85}) +[2026-06-23 23:31:35,622][starVLA.training.train_starvla][INFO] - Step 4770, Loss: {'train/loss': 0.10635187476873398, 'train/grad_norm_pre_clip': 11.46031665802002, 'timing/data': 0.0006361049599945545, 'timing/model': 1.680135197006166, 'learning_rate/qwen_vl_interface': 5.515513145155531e-07, 'learning_rate/action_model': 5.515513145155531e-06, 'epoch': 1.85}) +[2026-06-23 23:31:38,931][starVLA.training.train_starvla][INFO] - Step 4771, Loss: {'train/loss': 0.13084830343723297, 'train/grad_norm_pre_clip': 39.210845947265625, 'timing/data': 0.0006253351457417011, 'timing/model': 1.6806040999945253, 'learning_rate/qwen_vl_interface': 5.511048205319702e-07, 'learning_rate/action_model': 5.511048205319701e-06, 'epoch': 1.85}) +[2026-06-23 23:31:42,240][starVLA.training.train_starvla][INFO] - Step 4772, Loss: {'train/loss': 0.0779363214969635, 'train/grad_norm_pre_clip': 17.318485260009766, 'timing/data': 0.0006364160217344761, 'timing/model': 1.6821393859572709, 'learning_rate/qwen_vl_interface': 5.506602580867034e-07, 'learning_rate/action_model': 5.506602580867034e-06, 'epoch': 1.85}) +[2026-06-23 23:31:45,549][starVLA.training.train_starvla][INFO] - Step 4773, Loss: {'train/loss': 0.09781321883201599, 'train/grad_norm_pre_clip': 48.97718048095703, 'timing/data': 0.0005535408854484558, 'timing/model': 1.6829303340055048, 'learning_rate/qwen_vl_interface': 5.502176273624942e-07, 'learning_rate/action_model': 5.502176273624942e-06, 'epoch': 1.85}) +[2026-06-23 23:31:48,862][starVLA.training.train_starvla][INFO] - Step 4774, Loss: {'train/loss': 0.07579823583364487, 'train/grad_norm_pre_clip': 19.85407257080078, 'timing/data': 0.0005618061404675245, 'timing/model': 1.6831422040704638, 'learning_rate/qwen_vl_interface': 5.497769285412921e-07, 'learning_rate/action_model': 5.497769285412921e-06, 'epoch': 1.85}) +[2026-06-23 23:31:52,179][starVLA.training.train_starvla][INFO] - Step 4775, Loss: {'train/loss': 0.17984794080257416, 'train/grad_norm_pre_clip': 35.57109451293945, 'timing/data': 0.0006360250990837812, 'timing/model': 1.6861343879718333, 'learning_rate/qwen_vl_interface': 5.493381618042519e-07, 'learning_rate/action_model': 5.493381618042518e-06, 'epoch': 1.85}) +[2026-06-23 23:31:55,490][starVLA.training.train_starvla][INFO] - Step 4776, Loss: {'train/loss': 0.07979439944028854, 'train/grad_norm_pre_clip': 17.032976150512695, 'timing/data': 0.0005797888152301311, 'timing/model': 1.6806446458213031, 'learning_rate/qwen_vl_interface': 5.489013273317343e-07, 'learning_rate/action_model': 5.489013273317343e-06, 'epoch': 1.85}) +[2026-06-23 23:31:58,806][starVLA.training.train_starvla][INFO] - Step 4777, Loss: {'train/loss': 0.0389944352209568, 'train/grad_norm_pre_clip': 12.320565223693848, 'timing/data': 0.0005731771234422922, 'timing/model': 1.684506693854928, 'learning_rate/qwen_vl_interface': 5.484664253033042e-07, 'learning_rate/action_model': 5.484664253033042e-06, 'epoch': 1.85}) +[2026-06-23 23:32:02,123][starVLA.training.train_starvla][INFO] - Step 4778, Loss: {'train/loss': 0.2218538373708725, 'train/grad_norm_pre_clip': 55.54701232910156, 'timing/data': 0.0006330490577965975, 'timing/model': 1.686465275939554, 'learning_rate/qwen_vl_interface': 5.480334558977351e-07, 'learning_rate/action_model': 5.480334558977351e-06, 'epoch': 1.85}) +[2026-06-23 23:32:05,433][starVLA.training.train_starvla][INFO] - Step 4779, Loss: {'train/loss': 0.05117190256714821, 'train/grad_norm_pre_clip': 19.899503707885742, 'timing/data': 0.0005759920459240675, 'timing/model': 1.683152011828497, 'learning_rate/qwen_vl_interface': 5.476024192930029e-07, 'learning_rate/action_model': 5.47602419293003e-06, 'epoch': 1.85}) +[2026-06-23 23:32:08,747][starVLA.training.train_starvla][INFO] - Step 4780, Loss: {'train/loss': 0.08794216811656952, 'train/grad_norm_pre_clip': 17.49308204650879, 'timing/data': 0.0005817229393869638, 'timing/model': 1.6813203531783074, 'learning_rate/qwen_vl_interface': 5.471733156662914e-07, 'learning_rate/action_model': 5.4717331566629135e-06, 'epoch': 1.85}) +[2026-06-23 23:32:12,061][starVLA.training.train_starvla][INFO] - Step 4781, Loss: {'train/loss': 0.15133769810199738, 'train/grad_norm_pre_clip': 15.317242622375488, 'timing/data': 0.0005429598968476057, 'timing/model': 1.6850630268454552, 'learning_rate/qwen_vl_interface': 5.467461451939882e-07, 'learning_rate/action_model': 5.467461451939882e-06, 'epoch': 1.85}) +[2026-06-23 23:32:15,379][starVLA.training.train_starvla][INFO] - Step 4782, Loss: {'train/loss': 0.12312480062246323, 'train/grad_norm_pre_clip': 21.46522331237793, 'timing/data': 0.0005611439701169729, 'timing/model': 1.6850898088887334, 'learning_rate/qwen_vl_interface': 5.463209080516873e-07, 'learning_rate/action_model': 5.463209080516873e-06, 'epoch': 1.85}) +[2026-06-23 23:32:18,695][starVLA.training.train_starvla][INFO] - Step 4783, Loss: {'train/loss': 0.09074121713638306, 'train/grad_norm_pre_clip': 12.814186096191406, 'timing/data': 0.0006158971227705479, 'timing/model': 1.6835957830771804, 'learning_rate/qwen_vl_interface': 5.458976044141875e-07, 'learning_rate/action_model': 5.458976044141875e-06, 'epoch': 1.85}) +[2026-06-23 23:32:22,191][starVLA.training.train_starvla][INFO] - Step 4784, Loss: {'train/loss': 0.14508074522018433, 'train/grad_norm_pre_clip': 19.435020446777344, 'timing/data': 0.0005659530870616436, 'timing/model': 1.8653814459685236, 'learning_rate/qwen_vl_interface': 5.454762344554931e-07, 'learning_rate/action_model': 5.4547623445549316e-06, 'epoch': 1.85}) +[2026-06-23 23:32:25,505][starVLA.training.train_starvla][INFO] - Step 4785, Loss: {'train/loss': 0.058516595512628555, 'train/grad_norm_pre_clip': 45.11918640136719, 'timing/data': 0.0006599898915737867, 'timing/model': 1.685546002117917, 'learning_rate/qwen_vl_interface': 5.450567983488137e-07, 'learning_rate/action_model': 5.450567983488137e-06, 'epoch': 1.85}) +[2026-06-23 23:32:28,812][starVLA.training.train_starvla][INFO] - Step 4786, Loss: {'train/loss': 0.1128394827246666, 'train/grad_norm_pre_clip': 17.758771896362305, 'timing/data': 0.0006822319701313972, 'timing/model': 1.682010950986296, 'learning_rate/qwen_vl_interface': 5.446392962665629e-07, 'learning_rate/action_model': 5.4463929626656296e-06, 'epoch': 1.85}) +[2026-06-23 23:32:32,125][starVLA.training.train_starvla][INFO] - Step 4787, Loss: {'train/loss': 0.13492541015148163, 'train/grad_norm_pre_clip': 28.17966651916504, 'timing/data': 0.0006497208960354328, 'timing/model': 1.6826256981585175, 'learning_rate/qwen_vl_interface': 5.442237283803605e-07, 'learning_rate/action_model': 5.442237283803604e-06, 'epoch': 1.85}) +[2026-06-23 23:32:35,438][starVLA.training.train_starvla][INFO] - Step 4788, Loss: {'train/loss': 0.109965980052948, 'train/grad_norm_pre_clip': 81.99909973144531, 'timing/data': 0.0006377180106937885, 'timing/model': 1.6815021990332752, 'learning_rate/qwen_vl_interface': 5.43810094861031e-07, 'learning_rate/action_model': 5.43810094861031e-06, 'epoch': 1.85}) +[2026-06-23 23:32:38,756][starVLA.training.train_starvla][INFO] - Step 4789, Loss: {'train/loss': 0.10286535322666168, 'train/grad_norm_pre_clip': 12.433951377868652, 'timing/data': 0.000636636046692729, 'timing/model': 1.6872585560195148, 'learning_rate/qwen_vl_interface': 5.433983958786027e-07, 'learning_rate/action_model': 5.433983958786027e-06, 'epoch': 1.85}) +[2026-06-23 23:32:42,076][starVLA.training.train_starvla][INFO] - Step 4790, Loss: {'train/loss': 0.22765837609767914, 'train/grad_norm_pre_clip': 49.57173156738281, 'timing/data': 0.0006270490121096373, 'timing/model': 1.6866633400786668, 'learning_rate/qwen_vl_interface': 5.429886316023102e-07, 'learning_rate/action_model': 5.429886316023101e-06, 'epoch': 1.85}) +[2026-06-23 23:32:45,389][starVLA.training.train_starvla][INFO] - Step 4791, Loss: {'train/loss': 0.0903034508228302, 'train/grad_norm_pre_clip': 12.913949012756348, 'timing/data': 0.0006290420424193144, 'timing/model': 1.6849633171223104, 'learning_rate/qwen_vl_interface': 5.425808022005916e-07, 'learning_rate/action_model': 5.425808022005916e-06, 'epoch': 1.85}) +[2026-06-23 23:32:48,696][starVLA.training.train_starvla][INFO] - Step 4792, Loss: {'train/loss': 0.0933377593755722, 'train/grad_norm_pre_clip': 15.142346382141113, 'timing/data': 0.0007480049971491098, 'timing/model': 1.6756546380929649, 'learning_rate/qwen_vl_interface': 5.421749078410903e-07, 'learning_rate/action_model': 5.421749078410903e-06, 'epoch': 1.85}) +[2026-06-23 23:32:52,014][starVLA.training.train_starvla][INFO] - Step 4793, Loss: {'train/loss': 0.09136935323476791, 'train/grad_norm_pre_clip': 26.89604949951172, 'timing/data': 0.0005571770016103983, 'timing/model': 1.6858902019448578, 'learning_rate/qwen_vl_interface': 5.417709486906547e-07, 'learning_rate/action_model': 5.417709486906547e-06, 'epoch': 1.85}) +[2026-06-23 23:32:55,331][starVLA.training.train_starvla][INFO] - Step 4794, Loss: {'train/loss': 0.05442722886800766, 'train/grad_norm_pre_clip': 18.48802947998047, 'timing/data': 0.000577154103666544, 'timing/model': 1.6849228660576046, 'learning_rate/qwen_vl_interface': 5.413689249153367e-07, 'learning_rate/action_model': 5.413689249153366e-06, 'epoch': 1.86}) +[2026-06-23 23:32:58,652][starVLA.training.train_starvla][INFO] - Step 4795, Loss: {'train/loss': 0.06284652650356293, 'train/grad_norm_pre_clip': 17.020950317382812, 'timing/data': 0.000558098778128624, 'timing/model': 1.686722599901259, 'learning_rate/qwen_vl_interface': 5.409688366803931e-07, 'learning_rate/action_model': 5.409688366803931e-06, 'epoch': 1.86}) +[2026-06-23 23:33:01,969][starVLA.training.train_starvla][INFO] - Step 4796, Loss: {'train/loss': 0.04778685420751572, 'train/grad_norm_pre_clip': 18.4111270904541, 'timing/data': 0.0005861211102455854, 'timing/model': 1.6817761061247438, 'learning_rate/qwen_vl_interface': 5.405706841502853e-07, 'learning_rate/action_model': 5.4057068415028525e-06, 'epoch': 1.86}) +[2026-06-23 23:33:05,286][starVLA.training.train_starvla][INFO] - Step 4797, Loss: {'train/loss': 0.06247759610414505, 'train/grad_norm_pre_clip': 22.652362823486328, 'timing/data': 0.0005658329464495182, 'timing/model': 1.6874490370973945, 'learning_rate/qwen_vl_interface': 5.401744674886781e-07, 'learning_rate/action_model': 5.40174467488678e-06, 'epoch': 1.86}) +[2026-06-23 23:33:08,605][starVLA.training.train_starvla][INFO] - Step 4798, Loss: {'train/loss': 0.07224179804325104, 'train/grad_norm_pre_clip': 35.48018264770508, 'timing/data': 0.0005575870163738728, 'timing/model': 1.6886028270237148, 'learning_rate/qwen_vl_interface': 5.397801868584418e-07, 'learning_rate/action_model': 5.397801868584417e-06, 'epoch': 1.86}) +[2026-06-23 23:33:12,185][starVLA.training.train_starvla][INFO] - Step 4799, Loss: {'train/loss': 0.08800061047077179, 'train/grad_norm_pre_clip': 19.6861629486084, 'timing/data': 0.0006515639834105968, 'timing/model': 1.9484791150316596, 'learning_rate/qwen_vl_interface': 5.393878424216502e-07, 'learning_rate/action_model': 5.393878424216501e-06, 'epoch': 1.86}) +[2026-06-23 23:33:15,494][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4800 +[2026-06-23 23:34:47,558][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4800 +[2026-06-23 23:34:47,559][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4800 +[2026-06-23 23:35:58,253][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4800 +[2026-06-23 23:35:58,254][starVLA.training.train_starvla][INFO] - Step 4800, Loss: {'train/loss': 0.13594630360603333, 'train/grad_norm_pre_clip': 29.100906372070312, 'eval/action_loss/seconds': 92.06181157799438, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06477173417806625, 'eval/latency_2/loss': 0.06477173417806625, 'eval/flappy/loss': 0.06477173417806625, 'eval/flappy/latency_2/loss': 0.06477173417806625, 'eval/flap_precision': 0.9370816599732262, 'eval/flap_recall': 0.9446693657219973, 'eval/flap_f1': 0.9408602150537634, 'eval/flap_rate_model': 0.2075, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9408602150537634, 'eval/latency_2/flap_precision': 0.9370816599732262, 'eval/latency_2/flap_recall': 0.9446693657219973, 'eval/latency_2/flap_f1': 0.9408602150537634, 'eval/latency_2/flap_rate_model': 0.2075, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9408602150537634, 'eval/action_classification/data_wait_seconds': 0.7914840579032898, 'eval/action_classification/predict_seconds': 69.7696533203125, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005614038091152906, 'timing/model': 1.6819152650423348, 'learning_rate/qwen_vl_interface': 5.389974343395813e-07, 'learning_rate/action_model': 5.389974343395812e-06, 'epoch': 1.86}) +[2026-06-23 23:36:01,563][starVLA.training.train_starvla][INFO] - Step 4801, Loss: {'train/loss': 0.078071728348732, 'train/grad_norm_pre_clip': 16.204212188720703, 'timing/data': 0.0005645898636430502, 'timing/model': 1.6841731609310955, 'learning_rate/qwen_vl_interface': 5.38608962772717e-07, 'learning_rate/action_model': 5.38608962772717e-06, 'epoch': 1.86}) +[2026-06-23 23:36:04,867][starVLA.training.train_starvla][INFO] - Step 4802, Loss: {'train/loss': 0.20586806535720825, 'train/grad_norm_pre_clip': 25.63711929321289, 'timing/data': 0.0006399110425263643, 'timing/model': 1.6766115301288664, 'learning_rate/qwen_vl_interface': 5.382224278807429e-07, 'learning_rate/action_model': 5.382224278807429e-06, 'epoch': 1.86}) +[2026-06-23 23:36:08,177][starVLA.training.train_starvla][INFO] - Step 4803, Loss: {'train/loss': 0.15591955184936523, 'train/grad_norm_pre_clip': 66.65635681152344, 'timing/data': 0.0005506530869752169, 'timing/model': 1.6811563619412482, 'learning_rate/qwen_vl_interface': 5.378378298225496e-07, 'learning_rate/action_model': 5.378378298225496e-06, 'epoch': 1.86}) +[2026-06-23 23:36:11,485][starVLA.training.train_starvla][INFO] - Step 4804, Loss: {'train/loss': 0.17126652598381042, 'train/grad_norm_pre_clip': 24.76357650756836, 'timing/data': 0.0005694290157407522, 'timing/model': 1.6805380531586707, 'learning_rate/qwen_vl_interface': 5.374551687562306e-07, 'learning_rate/action_model': 5.374551687562306e-06, 'epoch': 1.86}) +[2026-06-23 23:36:14,802][starVLA.training.train_starvla][INFO] - Step 4805, Loss: {'train/loss': 0.16677653789520264, 'train/grad_norm_pre_clip': 38.77375030517578, 'timing/data': 0.0005622049793601036, 'timing/model': 1.685995265841484, 'learning_rate/qwen_vl_interface': 5.37074444839083e-07, 'learning_rate/action_model': 5.370744448390829e-06, 'epoch': 1.86}) +[2026-06-23 23:36:18,245][starVLA.training.train_starvla][INFO] - Step 4806, Loss: {'train/loss': 0.09056519716978073, 'train/grad_norm_pre_clip': 33.73933792114258, 'timing/data': 0.0007346388883888721, 'timing/model': 1.8137507010251284, 'learning_rate/qwen_vl_interface': 5.366956582276085e-07, 'learning_rate/action_model': 5.366956582276084e-06, 'epoch': 1.86}) +[2026-06-23 23:36:21,558][starVLA.training.train_starvla][INFO] - Step 4807, Loss: {'train/loss': 0.06593641638755798, 'train/grad_norm_pre_clip': 26.48750877380371, 'timing/data': 0.0005768928676843643, 'timing/model': 1.6851001000031829, 'learning_rate/qwen_vl_interface': 5.363188090775122e-07, 'learning_rate/action_model': 5.363188090775122e-06, 'epoch': 1.86}) +[2026-06-23 23:36:24,874][starVLA.training.train_starvla][INFO] - Step 4808, Loss: {'train/loss': 0.11820489168167114, 'train/grad_norm_pre_clip': 32.33219528198242, 'timing/data': 0.0005793077871203423, 'timing/model': 1.6843575169332325, 'learning_rate/qwen_vl_interface': 5.359438975437018e-07, 'learning_rate/action_model': 5.359438975437017e-06, 'epoch': 1.86}) +[2026-06-23 23:36:28,188][starVLA.training.train_starvla][INFO] - Step 4809, Loss: {'train/loss': 0.06628552824258804, 'train/grad_norm_pre_clip': 18.185211181640625, 'timing/data': 0.0005707510281354189, 'timing/model': 1.682484552031383, 'learning_rate/qwen_vl_interface': 5.355709237802895e-07, 'learning_rate/action_model': 5.3557092378028955e-06, 'epoch': 1.86}) +[2026-06-23 23:36:31,503][starVLA.training.train_starvla][INFO] - Step 4810, Loss: {'train/loss': 0.17162759602069855, 'train/grad_norm_pre_clip': 33.30585861206055, 'timing/data': 0.0005516749806702137, 'timing/model': 1.6851870261598378, 'learning_rate/qwen_vl_interface': 5.351998879405911e-07, 'learning_rate/action_model': 5.351998879405911e-06, 'epoch': 1.86}) +[2026-06-23 23:36:34,821][starVLA.training.train_starvla][INFO] - Step 4811, Loss: {'train/loss': 0.08410481363534927, 'train/grad_norm_pre_clip': 14.12121295928955, 'timing/data': 0.0005665430799126625, 'timing/model': 1.6845038230530918, 'learning_rate/qwen_vl_interface': 5.348307901771256e-07, 'learning_rate/action_model': 5.348307901771256e-06, 'epoch': 1.86}) +[2026-06-23 23:36:38,134][starVLA.training.train_starvla][INFO] - Step 4812, Loss: {'train/loss': 0.04269695281982422, 'train/grad_norm_pre_clip': 25.068622589111328, 'timing/data': 0.0005612641107290983, 'timing/model': 1.6834369720891118, 'learning_rate/qwen_vl_interface': 5.344636306416147e-07, 'learning_rate/action_model': 5.344636306416146e-06, 'epoch': 1.86}) +[2026-06-23 23:36:41,445][starVLA.training.train_starvla][INFO] - Step 4813, Loss: {'train/loss': 0.13292422890663147, 'train/grad_norm_pre_clip': 47.072452545166016, 'timing/data': 0.0005740479100495577, 'timing/model': 1.6825635428540409, 'learning_rate/qwen_vl_interface': 5.340984094849836e-07, 'learning_rate/action_model': 5.340984094849836e-06, 'epoch': 1.86}) +[2026-06-23 23:36:44,790][starVLA.training.train_starvla][INFO] - Step 4814, Loss: {'train/loss': 0.07388506084680557, 'train/grad_norm_pre_clip': 26.064651489257812, 'timing/data': 0.0006113681010901928, 'timing/model': 1.7099806040059775, 'learning_rate/qwen_vl_interface': 5.337351268573606e-07, 'learning_rate/action_model': 5.337351268573606e-06, 'epoch': 1.86}) +[2026-06-23 23:36:48,109][starVLA.training.train_starvla][INFO] - Step 4815, Loss: {'train/loss': 0.0747964158654213, 'train/grad_norm_pre_clip': 17.067113876342773, 'timing/data': 0.0006314951460808516, 'timing/model': 1.6856688410043716, 'learning_rate/qwen_vl_interface': 5.33373782908079e-07, 'learning_rate/action_model': 5.333737829080789e-06, 'epoch': 1.86}) +[2026-06-23 23:36:51,421][starVLA.training.train_starvla][INFO] - Step 4816, Loss: {'train/loss': 0.22804585099220276, 'train/grad_norm_pre_clip': 19.745439529418945, 'timing/data': 0.0005801289808005095, 'timing/model': 1.6857532400172204, 'learning_rate/qwen_vl_interface': 5.330143777856718e-07, 'learning_rate/action_model': 5.330143777856718e-06, 'epoch': 1.86}) +[2026-06-23 23:36:54,734][starVLA.training.train_starvla][INFO] - Step 4817, Loss: {'train/loss': 0.07261449098587036, 'train/grad_norm_pre_clip': 16.637908935546875, 'timing/data': 0.0005828749854117632, 'timing/model': 1.683899660827592, 'learning_rate/qwen_vl_interface': 5.326569116378786e-07, 'learning_rate/action_model': 5.326569116378786e-06, 'epoch': 1.86}) +[2026-06-23 23:36:58,050][starVLA.training.train_starvla][INFO] - Step 4818, Loss: {'train/loss': 0.10563663393259048, 'train/grad_norm_pre_clip': 22.047847747802734, 'timing/data': 0.0008904819842427969, 'timing/model': 1.6833592848852277, 'learning_rate/qwen_vl_interface': 5.323013846116387e-07, 'learning_rate/action_model': 5.323013846116386e-06, 'epoch': 1.86}) +[2026-06-23 23:37:01,371][starVLA.training.train_starvla][INFO] - Step 4819, Loss: {'train/loss': 0.1943139284849167, 'train/grad_norm_pre_clip': 53.02862548828125, 'timing/data': 0.0006347219459712505, 'timing/model': 1.6851451678667217, 'learning_rate/qwen_vl_interface': 5.31947796853097e-07, 'learning_rate/action_model': 5.3194779685309696e-06, 'epoch': 1.87}) +[2026-06-23 23:37:04,688][starVLA.training.train_starvla][INFO] - Step 4820, Loss: {'train/loss': 0.2175101786851883, 'train/grad_norm_pre_clip': 30.52669906616211, 'timing/data': 0.0006583859212696552, 'timing/model': 1.6849276490975171, 'learning_rate/qwen_vl_interface': 5.315961485075992e-07, 'learning_rate/action_model': 5.315961485075993e-06, 'epoch': 1.87}) +[2026-06-23 23:37:08,275][starVLA.training.train_starvla][INFO] - Step 4821, Loss: {'train/loss': 0.09416960924863815, 'train/grad_norm_pre_clip': 54.833492279052734, 'timing/data': 0.0005823930259793997, 'timing/model': 1.9554523730184883, 'learning_rate/qwen_vl_interface': 5.312464397196958e-07, 'learning_rate/action_model': 5.312464397196958e-06, 'epoch': 1.87}) +[2026-06-23 23:37:11,591][starVLA.training.train_starvla][INFO] - Step 4822, Loss: {'train/loss': 0.14121492207050323, 'train/grad_norm_pre_clip': 34.0997200012207, 'timing/data': 0.0005619050934910774, 'timing/model': 1.6818797381129116, 'learning_rate/qwen_vl_interface': 5.308986706331377e-07, 'learning_rate/action_model': 5.308986706331376e-06, 'epoch': 1.87}) +[2026-06-23 23:37:14,911][starVLA.training.train_starvla][INFO] - Step 4823, Loss: {'train/loss': 0.07605389505624771, 'train/grad_norm_pre_clip': 17.188711166381836, 'timing/data': 0.0005830938462167978, 'timing/model': 1.686437712982297, 'learning_rate/qwen_vl_interface': 5.305528413908797e-07, 'learning_rate/action_model': 5.305528413908798e-06, 'epoch': 1.87}) +[2026-06-23 23:37:18,234][starVLA.training.train_starvla][INFO] - Step 4824, Loss: {'train/loss': 0.07168086618185043, 'train/grad_norm_pre_clip': 18.78696632385254, 'timing/data': 0.0005776949692517519, 'timing/model': 1.689322094898671, 'learning_rate/qwen_vl_interface': 5.302089521350799e-07, 'learning_rate/action_model': 5.3020895213507996e-06, 'epoch': 1.87}) +[2026-06-23 23:37:21,550][starVLA.training.train_starvla][INFO] - Step 4825, Loss: {'train/loss': 0.19941341876983643, 'train/grad_norm_pre_clip': 29.71452522277832, 'timing/data': 0.0005937051028013229, 'timing/model': 1.6838948670774698, 'learning_rate/qwen_vl_interface': 5.298670030070977e-07, 'learning_rate/action_model': 5.298670030070976e-06, 'epoch': 1.87}) +[2026-06-23 23:37:24,868][starVLA.training.train_starvla][INFO] - Step 4826, Loss: {'train/loss': 0.15738849341869354, 'train/grad_norm_pre_clip': 16.247512817382812, 'timing/data': 0.0005922319833189249, 'timing/model': 1.6851638578809798, 'learning_rate/qwen_vl_interface': 5.295269941474957e-07, 'learning_rate/action_model': 5.295269941474957e-06, 'epoch': 1.87}) +[2026-06-23 23:37:28,186][starVLA.training.train_starvla][INFO] - Step 4827, Loss: {'train/loss': 0.06611905246973038, 'train/grad_norm_pre_clip': 13.308258056640625, 'timing/data': 0.0005650999955832958, 'timing/model': 1.6847629430703819, 'learning_rate/qwen_vl_interface': 5.291889256960384e-07, 'learning_rate/action_model': 5.291889256960384e-06, 'epoch': 1.87}) +[2026-06-23 23:37:31,505][starVLA.training.train_starvla][INFO] - Step 4828, Loss: {'train/loss': 0.0803147479891777, 'train/grad_norm_pre_clip': 14.269990921020508, 'timing/data': 0.0005533089861273766, 'timing/model': 1.6887690729927272, 'learning_rate/qwen_vl_interface': 5.288527977916929e-07, 'learning_rate/action_model': 5.288527977916929e-06, 'epoch': 1.87}) +[2026-06-23 23:37:34,820][starVLA.training.train_starvla][INFO] - Step 4829, Loss: {'train/loss': 0.2052697390317917, 'train/grad_norm_pre_clip': 46.472412109375, 'timing/data': 0.0005711228586733341, 'timing/model': 1.6845740862190723, 'learning_rate/qwen_vl_interface': 5.28518610572629e-07, 'learning_rate/action_model': 5.28518610572629e-06, 'epoch': 1.87}) +[2026-06-23 23:37:38,143][starVLA.training.train_starvla][INFO] - Step 4830, Loss: {'train/loss': 0.03374288231134415, 'train/grad_norm_pre_clip': 18.417694091796875, 'timing/data': 0.0005340829957276583, 'timing/model': 1.6857089751865715, 'learning_rate/qwen_vl_interface': 5.281863641762184e-07, 'learning_rate/action_model': 5.281863641762184e-06, 'epoch': 1.87}) +[2026-06-23 23:37:41,467][starVLA.training.train_starvla][INFO] - Step 4831, Loss: {'train/loss': 0.1689501702785492, 'train/grad_norm_pre_clip': 12.1732816696167, 'timing/data': 0.0005705610383301973, 'timing/model': 1.68741959403269, 'learning_rate/qwen_vl_interface': 5.278560587390355e-07, 'learning_rate/action_model': 5.278560587390355e-06, 'epoch': 1.87}) +[2026-06-23 23:37:44,783][starVLA.training.train_starvla][INFO] - Step 4832, Loss: {'train/loss': 0.11450595408678055, 'train/grad_norm_pre_clip': 20.055007934570312, 'timing/data': 0.0005763119552284479, 'timing/model': 1.6856007270980626, 'learning_rate/qwen_vl_interface': 5.275276943968557e-07, 'learning_rate/action_model': 5.275276943968557e-06, 'epoch': 1.87}) +[2026-06-23 23:37:48,106][starVLA.training.train_starvla][INFO] - Step 4833, Loss: {'train/loss': 0.060561709105968475, 'train/grad_norm_pre_clip': 18.619243621826172, 'timing/data': 0.000589447095990181, 'timing/model': 1.6884721859823912, 'learning_rate/qwen_vl_interface': 5.272012712846566e-07, 'learning_rate/action_model': 5.2720127128465654e-06, 'epoch': 1.87}) +[2026-06-23 23:37:51,426][starVLA.training.train_starvla][INFO] - Step 4834, Loss: {'train/loss': 0.2629888951778412, 'train/grad_norm_pre_clip': 25.513301849365234, 'timing/data': 0.0006270380690693855, 'timing/model': 1.6857522681821138, 'learning_rate/qwen_vl_interface': 5.268767895366198e-07, 'learning_rate/action_model': 5.268767895366198e-06, 'epoch': 1.87}) +[2026-06-23 23:37:54,848][starVLA.training.train_starvla][INFO] - Step 4835, Loss: {'train/loss': 0.02116900123655796, 'train/grad_norm_pre_clip': 48.148338317871094, 'timing/data': 0.0005696790758520365, 'timing/model': 1.790036922087893, 'learning_rate/qwen_vl_interface': 5.265542492861268e-07, 'learning_rate/action_model': 5.2655424928612676e-06, 'epoch': 1.87}) +[2026-06-23 23:37:58,174][starVLA.training.train_starvla][INFO] - Step 4836, Loss: {'train/loss': 0.19973093271255493, 'train/grad_norm_pre_clip': 22.540624618530273, 'timing/data': 0.0005606429185718298, 'timing/model': 1.6894336740951985, 'learning_rate/qwen_vl_interface': 5.262336506657617e-07, 'learning_rate/action_model': 5.2623365066576165e-06, 'epoch': 1.87}) +[2026-06-23 23:38:01,492][starVLA.training.train_starvla][INFO] - Step 4837, Loss: {'train/loss': 0.09998562932014465, 'train/grad_norm_pre_clip': 78.7704086303711, 'timing/data': 0.0005618850700557232, 'timing/model': 1.6896900639403611, 'learning_rate/qwen_vl_interface': 5.259149938073103e-07, 'learning_rate/action_model': 5.259149938073102e-06, 'epoch': 1.87}) +[2026-06-23 23:38:04,814][starVLA.training.train_starvla][INFO] - Step 4838, Loss: {'train/loss': 0.07364704459905624, 'train/grad_norm_pre_clip': 38.9884033203125, 'timing/data': 0.0005611630622297525, 'timing/model': 1.6882436049636453, 'learning_rate/qwen_vl_interface': 5.255982788417609e-07, 'learning_rate/action_model': 5.255982788417608e-06, 'epoch': 1.87}) +[2026-06-23 23:38:08,134][starVLA.training.train_starvla][INFO] - Step 4839, Loss: {'train/loss': 0.09009808301925659, 'train/grad_norm_pre_clip': 19.915788650512695, 'timing/data': 0.0005739680491387844, 'timing/model': 1.6852707408834249, 'learning_rate/qwen_vl_interface': 5.252835058993029e-07, 'learning_rate/action_model': 5.252835058993029e-06, 'epoch': 1.87}) +[2026-06-23 23:38:11,459][starVLA.training.train_starvla][INFO] - Step 4840, Loss: {'train/loss': 0.09383443742990494, 'train/grad_norm_pre_clip': 14.879644393920898, 'timing/data': 0.0005529879126697779, 'timing/model': 1.6919973359908909, 'learning_rate/qwen_vl_interface': 5.249706751093274e-07, 'learning_rate/action_model': 5.249706751093273e-06, 'epoch': 1.87}) +[2026-06-23 23:38:14,782][starVLA.training.train_starvla][INFO] - Step 4841, Loss: {'train/loss': 0.10450306534767151, 'train/grad_norm_pre_clip': 18.88660430908203, 'timing/data': 0.0006369049660861492, 'timing/model': 1.6893252569716424, 'learning_rate/qwen_vl_interface': 5.246597866004273e-07, 'learning_rate/action_model': 5.246597866004273e-06, 'epoch': 1.87}) +[2026-06-23 23:38:18,101][starVLA.training.train_starvla][INFO] - Step 4842, Loss: {'train/loss': 0.13444341719150543, 'train/grad_norm_pre_clip': 47.20711135864258, 'timing/data': 0.0005880340468138456, 'timing/model': 1.6863556990865618, 'learning_rate/qwen_vl_interface': 5.243508405003975e-07, 'learning_rate/action_model': 5.243508405003975e-06, 'epoch': 1.87}) +[2026-06-23 23:38:21,422][starVLA.training.train_starvla][INFO] - Step 4843, Loss: {'train/loss': 0.05640694126486778, 'train/grad_norm_pre_clip': 19.027057647705078, 'timing/data': 0.0007121372036635876, 'timing/model': 1.686293409904465, 'learning_rate/qwen_vl_interface': 5.240438369362336e-07, 'learning_rate/action_model': 5.2404383693623366e-06, 'epoch': 1.87}) +[2026-06-23 23:38:24,744][starVLA.training.train_starvla][INFO] - Step 4844, Loss: {'train/loss': 0.08380433171987534, 'train/grad_norm_pre_clip': 39.23844909667969, 'timing/data': 0.0006474859546869993, 'timing/model': 1.687580994097516, 'learning_rate/qwen_vl_interface': 5.237387760341335e-07, 'learning_rate/action_model': 5.237387760341335e-06, 'epoch': 1.87}) +[2026-06-23 23:38:28,068][starVLA.training.train_starvla][INFO] - Step 4845, Loss: {'train/loss': 0.06664276123046875, 'train/grad_norm_pre_clip': 15.321671485900879, 'timing/data': 0.0005523769650608301, 'timing/model': 1.6912971758283675, 'learning_rate/qwen_vl_interface': 5.234356579194963e-07, 'learning_rate/action_model': 5.234356579194963e-06, 'epoch': 1.88}) +[2026-06-23 23:38:31,393][starVLA.training.train_starvla][INFO] - Step 4846, Loss: {'train/loss': 0.06298935413360596, 'train/grad_norm_pre_clip': 13.127761840820312, 'timing/data': 0.0006388791371136904, 'timing/model': 1.6885934970341623, 'learning_rate/qwen_vl_interface': 5.231344827169218e-07, 'learning_rate/action_model': 5.231344827169217e-06, 'epoch': 1.88}) +[2026-06-23 23:38:34,715][starVLA.training.train_starvla][INFO] - Step 4847, Loss: {'train/loss': 0.13103386759757996, 'train/grad_norm_pre_clip': 28.394161224365234, 'timing/data': 0.0005559038836508989, 'timing/model': 1.6857494339346886, 'learning_rate/qwen_vl_interface': 5.228352505502119e-07, 'learning_rate/action_model': 5.228352505502119e-06, 'epoch': 1.88}) +[2026-06-23 23:38:38,042][starVLA.training.train_starvla][INFO] - Step 4848, Loss: {'train/loss': 0.06465979665517807, 'train/grad_norm_pre_clip': 22.700490951538086, 'timing/data': 0.0005717640742659569, 'timing/model': 1.684284269111231, 'learning_rate/qwen_vl_interface': 5.225379615423701e-07, 'learning_rate/action_model': 5.2253796154237e-06, 'epoch': 1.88}) +[2026-06-23 23:38:41,367][starVLA.training.train_starvla][INFO] - Step 4849, Loss: {'train/loss': 0.10718879103660583, 'train/grad_norm_pre_clip': 25.549001693725586, 'timing/data': 0.0006804370786994696, 'timing/model': 1.6911739110946655, 'learning_rate/qwen_vl_interface': 5.222426158156008e-07, 'learning_rate/action_model': 5.222426158156008e-06, 'epoch': 1.88}) +[2026-06-23 23:38:44,994][starVLA.training.train_starvla][INFO] - Step 4850, Loss: {'train/loss': 0.13850831985473633, 'train/grad_norm_pre_clip': 22.7856388092041, 'timing/data': 0.0005783960223197937, 'timing/model': 1.9911704731639475, 'learning_rate/qwen_vl_interface': 5.219492134913084e-07, 'learning_rate/action_model': 5.219492134913084e-06, 'epoch': 1.88}) +[2026-06-23 23:38:48,318][starVLA.training.train_starvla][INFO] - Step 4851, Loss: {'train/loss': 0.12909206748008728, 'train/grad_norm_pre_clip': 27.299800872802734, 'timing/data': 0.0005842270329594612, 'timing/model': 1.6882166820578277, 'learning_rate/qwen_vl_interface': 5.216577546901008e-07, 'learning_rate/action_model': 5.216577546901007e-06, 'epoch': 1.88}) +[2026-06-23 23:38:51,641][starVLA.training.train_starvla][INFO] - Step 4852, Loss: {'train/loss': 0.07169955968856812, 'train/grad_norm_pre_clip': 19.219432830810547, 'timing/data': 0.0005675461143255234, 'timing/model': 1.6861583539284766, 'learning_rate/qwen_vl_interface': 5.213682395317852e-07, 'learning_rate/action_model': 5.213682395317851e-06, 'epoch': 1.88}) +[2026-06-23 23:38:54,962][starVLA.training.train_starvla][INFO] - Step 4853, Loss: {'train/loss': 0.11073725670576096, 'train/grad_norm_pre_clip': 21.49325180053711, 'timing/data': 0.0005800281651318073, 'timing/model': 1.6874556040856987, 'learning_rate/qwen_vl_interface': 5.2108066813537e-07, 'learning_rate/action_model': 5.2108066813537e-06, 'epoch': 1.88}) +[2026-06-23 23:38:58,287][starVLA.training.train_starvla][INFO] - Step 4854, Loss: {'train/loss': 0.1288936287164688, 'train/grad_norm_pre_clip': 37.47483825683594, 'timing/data': 0.000554871978238225, 'timing/model': 1.6880646930076182, 'learning_rate/qwen_vl_interface': 5.207950406190654e-07, 'learning_rate/action_model': 5.207950406190653e-06, 'epoch': 1.88}) +[2026-06-23 23:39:01,612][starVLA.training.train_starvla][INFO] - Step 4855, Loss: {'train/loss': 0.16007083654403687, 'train/grad_norm_pre_clip': 26.962535858154297, 'timing/data': 0.0005730250850319862, 'timing/model': 1.686981586040929, 'learning_rate/qwen_vl_interface': 5.205113571002818e-07, 'learning_rate/action_model': 5.205113571002818e-06, 'epoch': 1.88}) +[2026-06-23 23:39:04,936][starVLA.training.train_starvla][INFO] - Step 4856, Loss: {'train/loss': 0.06789728254079819, 'train/grad_norm_pre_clip': 12.385903358459473, 'timing/data': 0.000558558851480484, 'timing/model': 1.6883572139777243, 'learning_rate/qwen_vl_interface': 5.202296176956308e-07, 'learning_rate/action_model': 5.202296176956308e-06, 'epoch': 1.88}) +[2026-06-23 23:39:08,256][starVLA.training.train_starvla][INFO] - Step 4857, Loss: {'train/loss': 0.25495845079421997, 'train/grad_norm_pre_clip': 36.73972702026367, 'timing/data': 0.0005622059106826782, 'timing/model': 1.6880466560833156, 'learning_rate/qwen_vl_interface': 5.19949822520925e-07, 'learning_rate/action_model': 5.19949822520925e-06, 'epoch': 1.88}) +[2026-06-23 23:39:11,581][starVLA.training.train_starvla][INFO] - Step 4858, Loss: {'train/loss': 0.05601367726922035, 'train/grad_norm_pre_clip': 22.02996253967285, 'timing/data': 0.0005875229835510254, 'timing/model': 1.6932783129159361, 'learning_rate/qwen_vl_interface': 5.196719716911777e-07, 'learning_rate/action_model': 5.1967197169117764e-06, 'epoch': 1.88}) +[2026-06-23 23:39:14,904][starVLA.training.train_starvla][INFO] - Step 4859, Loss: {'train/loss': 0.0778944343328476, 'train/grad_norm_pre_clip': 30.55670738220215, 'timing/data': 0.0005883551202714443, 'timing/model': 1.6888787690550089, 'learning_rate/qwen_vl_interface': 5.193960653206029e-07, 'learning_rate/action_model': 5.193960653206029e-06, 'epoch': 1.88}) +[2026-06-23 23:39:18,229][starVLA.training.train_starvla][INFO] - Step 4860, Loss: {'train/loss': 0.22426435351371765, 'train/grad_norm_pre_clip': 22.115358352661133, 'timing/data': 0.000581761822104454, 'timing/model': 1.6879380769096315, 'learning_rate/qwen_vl_interface': 5.191221035226144e-07, 'learning_rate/action_model': 5.191221035226144e-06, 'epoch': 1.88}) +[2026-06-23 23:39:21,546][starVLA.training.train_starvla][INFO] - Step 4861, Loss: {'train/loss': 0.18678544461727142, 'train/grad_norm_pre_clip': 29.11663246154785, 'timing/data': 0.0005975121166557074, 'timing/model': 1.6813055698294193, 'learning_rate/qwen_vl_interface': 5.188500864098292e-07, 'learning_rate/action_model': 5.1885008640982915e-06, 'epoch': 1.88}) +[2026-06-23 23:39:24,877][starVLA.training.train_starvla][INFO] - Step 4862, Loss: {'train/loss': 0.10376933962106705, 'train/grad_norm_pre_clip': 25.08704948425293, 'timing/data': 0.0005478691309690475, 'timing/model': 1.6950103959534317, 'learning_rate/qwen_vl_interface': 5.185800140940613e-07, 'learning_rate/action_model': 5.185800140940613e-06, 'epoch': 1.88}) +[2026-06-23 23:39:28,200][starVLA.training.train_starvla][INFO] - Step 4863, Loss: {'train/loss': 0.09176523983478546, 'train/grad_norm_pre_clip': 16.71356773376465, 'timing/data': 0.0006528159137815237, 'timing/model': 1.6888327919878066, 'learning_rate/qwen_vl_interface': 5.183118866863288e-07, 'learning_rate/action_model': 5.1831188668632875e-06, 'epoch': 1.88}) +[2026-06-23 23:39:31,513][starVLA.training.train_starvla][INFO] - Step 4864, Loss: {'train/loss': 0.06062902882695198, 'train/grad_norm_pre_clip': 14.00034236907959, 'timing/data': 0.00054519297555089, 'timing/model': 1.6903088681865484, 'learning_rate/qwen_vl_interface': 5.180457042968484e-07, 'learning_rate/action_model': 5.1804570429684835e-06, 'epoch': 1.88}) +[2026-06-23 23:39:35,052][starVLA.training.train_starvla][INFO] - Step 4865, Loss: {'train/loss': 0.04706013202667236, 'train/grad_norm_pre_clip': 10.305659294128418, 'timing/data': 0.0005609530489891768, 'timing/model': 1.9087224369868636, 'learning_rate/qwen_vl_interface': 5.17781467035037e-07, 'learning_rate/action_model': 5.177814670350371e-06, 'epoch': 1.88}) +[2026-06-23 23:39:38,380][starVLA.training.train_starvla][INFO] - Step 4866, Loss: {'train/loss': 0.0960737019777298, 'train/grad_norm_pre_clip': 25.32887840270996, 'timing/data': 0.0006792859639972448, 'timing/model': 1.6880245089996606, 'learning_rate/qwen_vl_interface': 5.175191750095133e-07, 'learning_rate/action_model': 5.175191750095132e-06, 'epoch': 1.88}) +[2026-06-23 23:39:41,701][starVLA.training.train_starvla][INFO] - Step 4867, Loss: {'train/loss': 0.08769704401493073, 'train/grad_norm_pre_clip': 12.87654972076416, 'timing/data': 0.0005714530125260353, 'timing/model': 1.6845648880116642, 'learning_rate/qwen_vl_interface': 5.172588283280953e-07, 'learning_rate/action_model': 5.172588283280953e-06, 'epoch': 1.88}) +[2026-06-23 23:39:45,027][starVLA.training.train_starvla][INFO] - Step 4868, Loss: {'train/loss': 0.11143121868371964, 'train/grad_norm_pre_clip': 35.82535171508789, 'timing/data': 0.0007081488147377968, 'timing/model': 1.6917027728632092, 'learning_rate/qwen_vl_interface': 5.170004270978019e-07, 'learning_rate/action_model': 5.170004270978018e-06, 'epoch': 1.88}) +[2026-06-23 23:39:48,347][starVLA.training.train_starvla][INFO] - Step 4869, Loss: {'train/loss': 0.06609859317541122, 'train/grad_norm_pre_clip': 31.079648971557617, 'timing/data': 0.0005642201285809278, 'timing/model': 1.6855755241122097, 'learning_rate/qwen_vl_interface': 5.167439714248512e-07, 'learning_rate/action_model': 5.1674397142485116e-06, 'epoch': 1.88}) +[2026-06-23 23:39:51,670][starVLA.training.train_starvla][INFO] - Step 4870, Loss: {'train/loss': 0.04911163076758385, 'train/grad_norm_pre_clip': 26.489990234375, 'timing/data': 0.0005640389863401651, 'timing/model': 1.690115581965074, 'learning_rate/qwen_vl_interface': 5.164894614146633e-07, 'learning_rate/action_model': 5.164894614146634e-06, 'epoch': 1.88}) +[2026-06-23 23:39:54,997][starVLA.training.train_starvla][INFO] - Step 4871, Loss: {'train/loss': 0.2518325448036194, 'train/grad_norm_pre_clip': 20.538490295410156, 'timing/data': 0.0005709519609808922, 'timing/model': 1.6894065248779953, 'learning_rate/qwen_vl_interface': 5.16236897171858e-07, 'learning_rate/action_model': 5.16236897171858e-06, 'epoch': 1.89}) +[2026-06-23 23:39:58,319][starVLA.training.train_starvla][INFO] - Step 4872, Loss: {'train/loss': 0.10120311379432678, 'train/grad_norm_pre_clip': 30.5593204498291, 'timing/data': 0.0005721149500459433, 'timing/model': 1.6869029658846557, 'learning_rate/qwen_vl_interface': 5.15986278800254e-07, 'learning_rate/action_model': 5.15986278800254e-06, 'epoch': 1.89}) +[2026-06-23 23:40:01,642][starVLA.training.train_starvla][INFO] - Step 4873, Loss: {'train/loss': 0.06549625843763351, 'train/grad_norm_pre_clip': 23.99659538269043, 'timing/data': 0.0006233199965208769, 'timing/model': 1.6891222482081503, 'learning_rate/qwen_vl_interface': 5.157376064028719e-07, 'learning_rate/action_model': 5.157376064028718e-06, 'epoch': 1.89}) +[2026-06-23 23:40:04,967][starVLA.training.train_starvla][INFO] - Step 4874, Loss: {'train/loss': 0.06308306008577347, 'train/grad_norm_pre_clip': 15.361666679382324, 'timing/data': 0.0005863311234861612, 'timing/model': 1.6861666329205036, 'learning_rate/qwen_vl_interface': 5.154908800819307e-07, 'learning_rate/action_model': 5.154908800819306e-06, 'epoch': 1.89}) +[2026-06-23 23:40:08,292][starVLA.training.train_starvla][INFO] - Step 4875, Loss: {'train/loss': 0.17553628981113434, 'train/grad_norm_pre_clip': 51.79621887207031, 'timing/data': 0.0006035028491169214, 'timing/model': 1.6893192960415035, 'learning_rate/qwen_vl_interface': 5.152460999388504e-07, 'learning_rate/action_model': 5.152460999388503e-06, 'epoch': 1.89}) +[2026-06-23 23:40:11,622][starVLA.training.train_starvla][INFO] - Step 4876, Loss: {'train/loss': 0.07319480925798416, 'train/grad_norm_pre_clip': 28.233976364135742, 'timing/data': 0.0005802300292998552, 'timing/model': 1.6890921229496598, 'learning_rate/qwen_vl_interface': 5.150032660742517e-07, 'learning_rate/action_model': 5.150032660742517e-06, 'epoch': 1.89}) +[2026-06-23 23:40:14,944][starVLA.training.train_starvla][INFO] - Step 4877, Loss: {'train/loss': 0.08377289772033691, 'train/grad_norm_pre_clip': 19.325897216796875, 'timing/data': 0.0005730860866606236, 'timing/model': 1.6867164950817823, 'learning_rate/qwen_vl_interface': 5.147623785879535e-07, 'learning_rate/action_model': 5.147623785879534e-06, 'epoch': 1.89}) +[2026-06-23 23:40:18,269][starVLA.training.train_starvla][INFO] - Step 4878, Loss: {'train/loss': 0.09214913100004196, 'train/grad_norm_pre_clip': 15.465099334716797, 'timing/data': 0.0005727659445255995, 'timing/model': 1.6894022019114345, 'learning_rate/qwen_vl_interface': 5.145234375789758e-07, 'learning_rate/action_model': 5.145234375789758e-06, 'epoch': 1.89}) +[2026-06-23 23:40:21,747][starVLA.training.train_starvla][INFO] - Step 4879, Loss: {'train/loss': 0.03987719118595123, 'train/grad_norm_pre_clip': 16.448755264282227, 'timing/data': 0.0005555129610002041, 'timing/model': 1.8444003039039671, 'learning_rate/qwen_vl_interface': 5.142864431455387e-07, 'learning_rate/action_model': 5.1428644314553875e-06, 'epoch': 1.89}) +[2026-06-23 23:40:25,083][starVLA.training.train_starvla][INFO] - Step 4880, Loss: {'train/loss': 0.15836189687252045, 'train/grad_norm_pre_clip': 36.615821838378906, 'timing/data': 0.0007281170692294836, 'timing/model': 1.6975368929561228, 'learning_rate/qwen_vl_interface': 5.140513953850612e-07, 'learning_rate/action_model': 5.140513953850612e-06, 'epoch': 1.89}) +[2026-06-23 23:40:28,409][starVLA.training.train_starvla][INFO] - Step 4881, Loss: {'train/loss': 0.15522079169750214, 'train/grad_norm_pre_clip': 18.930397033691406, 'timing/data': 0.0005459149833768606, 'timing/model': 1.6884220619685948, 'learning_rate/qwen_vl_interface': 5.138182943941623e-07, 'learning_rate/action_model': 5.138182943941623e-06, 'epoch': 1.89}) +[2026-06-23 23:40:31,731][starVLA.training.train_starvla][INFO] - Step 4882, Loss: {'train/loss': 0.0502813458442688, 'train/grad_norm_pre_clip': 11.02620792388916, 'timing/data': 0.0005630869418382645, 'timing/model': 1.687609049025923, 'learning_rate/qwen_vl_interface': 5.135871402686615e-07, 'learning_rate/action_model': 5.135871402686615e-06, 'epoch': 1.89}) +[2026-06-23 23:40:35,058][starVLA.training.train_starvla][INFO] - Step 4883, Loss: {'train/loss': 0.09874673932790756, 'train/grad_norm_pre_clip': 34.69135284423828, 'timing/data': 0.0005754411686211824, 'timing/model': 1.6905604128260165, 'learning_rate/qwen_vl_interface': 5.133579331035777e-07, 'learning_rate/action_model': 5.133579331035777e-06, 'epoch': 1.89}) +[2026-06-23 23:40:38,389][starVLA.training.train_starvla][INFO] - Step 4884, Loss: {'train/loss': 0.0947704017162323, 'train/grad_norm_pre_clip': 54.17145919799805, 'timing/data': 0.0005952978972345591, 'timing/model': 1.6928994178306311, 'learning_rate/qwen_vl_interface': 5.131306729931284e-07, 'learning_rate/action_model': 5.1313067299312836e-06, 'epoch': 1.89}) +[2026-06-23 23:40:41,713][starVLA.training.train_starvla][INFO] - Step 4885, Loss: {'train/loss': 0.150074303150177, 'train/grad_norm_pre_clip': 49.27698516845703, 'timing/data': 0.0005813108291476965, 'timing/model': 1.6853878260590136, 'learning_rate/qwen_vl_interface': 5.129053600307326e-07, 'learning_rate/action_model': 5.129053600307326e-06, 'epoch': 1.89}) +[2026-06-23 23:40:45,039][starVLA.training.train_starvla][INFO] - Step 4886, Loss: {'train/loss': 0.10783989727497101, 'train/grad_norm_pre_clip': 36.23161315917969, 'timing/data': 0.0005722651258111, 'timing/model': 1.6876298889983445, 'learning_rate/qwen_vl_interface': 5.126819943090078e-07, 'learning_rate/action_model': 5.1268199430900775e-06, 'epoch': 1.89}) +[2026-06-23 23:40:48,362][starVLA.training.train_starvla][INFO] - Step 4887, Loss: {'train/loss': 0.10908947885036469, 'train/grad_norm_pre_clip': 38.28927993774414, 'timing/data': 0.0005586089100688696, 'timing/model': 1.6910579889081419, 'learning_rate/qwen_vl_interface': 5.124605759197706e-07, 'learning_rate/action_model': 5.124605759197706e-06, 'epoch': 1.89}) +[2026-06-23 23:40:51,697][starVLA.training.train_starvla][INFO] - Step 4888, Loss: {'train/loss': 0.11821465194225311, 'train/grad_norm_pre_clip': 19.687881469726562, 'timing/data': 0.0005569660570472479, 'timing/model': 1.694794937968254, 'learning_rate/qwen_vl_interface': 5.122411049540383e-07, 'learning_rate/action_model': 5.1224110495403825e-06, 'epoch': 1.89}) +[2026-06-23 23:40:55,022][starVLA.training.train_starvla][INFO] - Step 4889, Loss: {'train/loss': 0.11180296540260315, 'train/grad_norm_pre_clip': 28.881542205810547, 'timing/data': 0.0005764130037277937, 'timing/model': 1.6877523460425436, 'learning_rate/qwen_vl_interface': 5.120235815020269e-07, 'learning_rate/action_model': 5.120235815020269e-06, 'epoch': 1.89}) +[2026-06-23 23:40:58,346][starVLA.training.train_starvla][INFO] - Step 4890, Loss: {'train/loss': 0.08723117411136627, 'train/grad_norm_pre_clip': 15.317776679992676, 'timing/data': 0.0006008381024003029, 'timing/model': 1.687841054983437, 'learning_rate/qwen_vl_interface': 5.118080056531524e-07, 'learning_rate/action_model': 5.118080056531524e-06, 'epoch': 1.89}) +[2026-06-23 23:41:01,670][starVLA.training.train_starvla][INFO] - Step 4891, Loss: {'train/loss': 0.09852977097034454, 'train/grad_norm_pre_clip': 12.690513610839844, 'timing/data': 0.000549721997231245, 'timing/model': 1.6877136090770364, 'learning_rate/qwen_vl_interface': 5.115943774960295e-07, 'learning_rate/action_model': 5.115943774960295e-06, 'epoch': 1.89}) +[2026-06-23 23:41:04,996][starVLA.training.train_starvla][INFO] - Step 4892, Loss: {'train/loss': 0.20856191217899323, 'train/grad_norm_pre_clip': 16.990978240966797, 'timing/data': 0.0006708989385515451, 'timing/model': 1.6879049290437251, 'learning_rate/qwen_vl_interface': 5.113826971184727e-07, 'learning_rate/action_model': 5.113826971184727e-06, 'epoch': 1.89}) +[2026-06-23 23:41:08,321][starVLA.training.train_starvla][INFO] - Step 4893, Loss: {'train/loss': 0.0747743770480156, 'train/grad_norm_pre_clip': 13.432419776916504, 'timing/data': 0.0005852088797837496, 'timing/model': 1.6897496948949993, 'learning_rate/qwen_vl_interface': 5.111729646074954e-07, 'learning_rate/action_model': 5.111729646074954e-06, 'epoch': 1.89}) +[2026-06-23 23:41:11,781][starVLA.training.train_starvla][INFO] - Step 4894, Loss: {'train/loss': 0.09823310375213623, 'train/grad_norm_pre_clip': 19.21611213684082, 'timing/data': 0.0006674439646303654, 'timing/model': 1.6889487151056528, 'learning_rate/qwen_vl_interface': 5.10965180049312e-07, 'learning_rate/action_model': 5.10965180049312e-06, 'epoch': 1.89}) +[2026-06-23 23:41:15,103][starVLA.training.train_starvla][INFO] - Step 4895, Loss: {'train/loss': 0.0777641087770462, 'train/grad_norm_pre_clip': 17.509784698486328, 'timing/data': 0.000567043898627162, 'timing/model': 1.6858697340358049, 'learning_rate/qwen_vl_interface': 5.107593435293336e-07, 'learning_rate/action_model': 5.107593435293336e-06, 'epoch': 1.89}) +[2026-06-23 23:41:18,435][starVLA.training.train_starvla][INFO] - Step 4896, Loss: {'train/loss': 0.3525516986846924, 'train/grad_norm_pre_clip': 78.6976547241211, 'timing/data': 0.0005616950802505016, 'timing/model': 1.6924370690248907, 'learning_rate/qwen_vl_interface': 5.105554551321724e-07, 'learning_rate/action_model': 5.105554551321723e-06, 'epoch': 1.89}) +[2026-06-23 23:41:21,763][starVLA.training.train_starvla][INFO] - Step 4897, Loss: {'train/loss': 0.1191377192735672, 'train/grad_norm_pre_clip': 40.60206985473633, 'timing/data': 0.0005735468585044146, 'timing/model': 1.6918509302195162, 'learning_rate/qwen_vl_interface': 5.103535149416389e-07, 'learning_rate/action_model': 5.1035351494163885e-06, 'epoch': 1.9}) +[2026-06-23 23:41:25,089][starVLA.training.train_starvla][INFO] - Step 4898, Loss: {'train/loss': 0.11474505066871643, 'train/grad_norm_pre_clip': 23.77772331237793, 'timing/data': 0.0005640687886625528, 'timing/model': 1.68536851904355, 'learning_rate/qwen_vl_interface': 5.101535230407436e-07, 'learning_rate/action_model': 5.101535230407436e-06, 'epoch': 1.9}) +[2026-06-23 23:41:28,413][starVLA.training.train_starvla][INFO] - Step 4899, Loss: {'train/loss': 0.13578496873378754, 'train/grad_norm_pre_clip': 38.195281982421875, 'timing/data': 0.0005547220353037119, 'timing/model': 1.687157701002434, 'learning_rate/qwen_vl_interface': 5.099554795116956e-07, 'learning_rate/action_model': 5.0995547951169565e-06, 'epoch': 1.9}) +[2026-06-23 23:41:31,740][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4900 +[2026-06-23 23:43:03,735][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4900 +[2026-06-23 23:43:03,736][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 4900 +[2026-06-23 23:44:14,038][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 4900 +[2026-06-23 23:44:14,039][starVLA.training.train_starvla][INFO] - Step 4900, Loss: {'train/loss': 0.11356398463249207, 'train/grad_norm_pre_clip': 27.563270568847656, 'eval/action_loss/seconds': 91.99287419207394, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06477682292461395, 'eval/latency_2/loss': 0.06477682292461395, 'eval/flappy/loss': 0.06477682292461395, 'eval/flappy/latency_2/loss': 0.06477682292461395, 'eval/flap_precision': 0.9341755319148937, 'eval/flap_recall': 0.9480431848852902, 'eval/flap_f1': 0.9410582719357, 'eval/flap_rate_model': 0.2088888888888889, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9410582719357, 'eval/latency_2/flap_precision': 0.9341755319148937, 'eval/latency_2/flap_recall': 0.9480431848852902, 'eval/latency_2/flap_f1': 0.9410582719357, 'eval/latency_2/flap_rate_model': 0.2088888888888889, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9410582719357, 'eval/action_classification/data_wait_seconds': 1.0128816366195679, 'eval/action_classification/predict_seconds': 69.14168548583984, 'eval/action_classification/frames': 7200.0, 'timing/data': 0.0005482500419020653, 'timing/model': 1.689898878801614, 'learning_rate/qwen_vl_interface': 5.097593844359024e-07, 'learning_rate/action_model': 5.0975938443590245e-06, 'epoch': 1.9}) +[2026-06-23 23:44:17,348][starVLA.training.train_starvla][INFO] - Step 4901, Loss: {'train/loss': 0.13682864606380463, 'train/grad_norm_pre_clip': 50.857521057128906, 'timing/data': 0.000561113003641367, 'timing/model': 1.6803369431290776, 'learning_rate/qwen_vl_interface': 5.095652378939718e-07, 'learning_rate/action_model': 5.095652378939717e-06, 'epoch': 1.9}) +[2026-06-23 23:44:20,905][starVLA.training.train_starvla][INFO] - Step 4902, Loss: {'train/loss': 0.1862221211194992, 'train/grad_norm_pre_clip': 15.893533706665039, 'timing/data': 0.0005634580738842487, 'timing/model': 1.9171508371364325, 'learning_rate/qwen_vl_interface': 5.093730399657101e-07, 'learning_rate/action_model': 5.0937303996571e-06, 'epoch': 1.9}) +[2026-06-23 23:44:24,220][starVLA.training.train_starvla][INFO] - Step 4903, Loss: {'train/loss': 0.06721000373363495, 'train/grad_norm_pre_clip': 14.314949989318848, 'timing/data': 0.0007066270336508751, 'timing/model': 1.6854350951034576, 'learning_rate/qwen_vl_interface': 5.091827907301227e-07, 'learning_rate/action_model': 5.091827907301227e-06, 'epoch': 1.9}) +[2026-06-23 23:44:27,533][starVLA.training.train_starvla][INFO] - Step 4904, Loss: {'train/loss': 0.12374278903007507, 'train/grad_norm_pre_clip': 33.1682243347168, 'timing/data': 0.0006311249453574419, 'timing/model': 1.6830100980587304, 'learning_rate/qwen_vl_interface': 5.089944902654134e-07, 'learning_rate/action_model': 5.089944902654134e-06, 'epoch': 1.9}) +[2026-06-23 23:44:30,847][starVLA.training.train_starvla][INFO] - Step 4905, Loss: {'train/loss': 0.14797979593276978, 'train/grad_norm_pre_clip': 18.540546417236328, 'timing/data': 0.0006399019621312618, 'timing/model': 1.6850967360660434, 'learning_rate/qwen_vl_interface': 5.088081386489856e-07, 'learning_rate/action_model': 5.088081386489857e-06, 'epoch': 1.9}) +[2026-06-23 23:44:34,163][starVLA.training.train_starvla][INFO] - Step 4906, Loss: {'train/loss': 0.1653585284948349, 'train/grad_norm_pre_clip': 47.16431427001953, 'timing/data': 0.0006216380279511213, 'timing/model': 1.6850969849620014, 'learning_rate/qwen_vl_interface': 5.086237359574419e-07, 'learning_rate/action_model': 5.086237359574419e-06, 'epoch': 1.9}) +[2026-06-23 23:44:37,477][starVLA.training.train_starvla][INFO] - Step 4907, Loss: {'train/loss': 0.11179989576339722, 'train/grad_norm_pre_clip': 20.885961532592773, 'timing/data': 0.0006599689368158579, 'timing/model': 1.6812254809774458, 'learning_rate/qwen_vl_interface': 5.084412822665825e-07, 'learning_rate/action_model': 5.084412822665825e-06, 'epoch': 1.9}) +[2026-06-23 23:44:40,797][starVLA.training.train_starvla][INFO] - Step 4908, Loss: {'train/loss': 0.08402428030967712, 'train/grad_norm_pre_clip': 20.326444625854492, 'timing/data': 0.000654448987916112, 'timing/model': 1.6825590240769088, 'learning_rate/qwen_vl_interface': 5.082607776514081e-07, 'learning_rate/action_model': 5.082607776514081e-06, 'epoch': 1.9}) +[2026-06-23 23:44:44,112][starVLA.training.train_starvla][INFO] - Step 4909, Loss: {'train/loss': 0.11806698143482208, 'train/grad_norm_pre_clip': 36.36162185668945, 'timing/data': 0.000571743817999959, 'timing/model': 1.684253494022414, 'learning_rate/qwen_vl_interface': 5.080822221861167e-07, 'learning_rate/action_model': 5.080822221861166e-06, 'epoch': 1.9}) +[2026-06-23 23:44:47,424][starVLA.training.train_starvla][INFO] - Step 4910, Loss: {'train/loss': 0.14551569521427155, 'train/grad_norm_pre_clip': 16.02788543701172, 'timing/data': 0.0005505641456693411, 'timing/model': 1.6850322850514203, 'learning_rate/qwen_vl_interface': 5.07905615944106e-07, 'learning_rate/action_model': 5.0790561594410595e-06, 'epoch': 1.9}) +[2026-06-23 23:44:50,741][starVLA.training.train_starvla][INFO] - Step 4911, Loss: {'train/loss': 0.05077143758535385, 'train/grad_norm_pre_clip': 19.114194869995117, 'timing/data': 0.0006108870729804039, 'timing/model': 1.6807059771381319, 'learning_rate/qwen_vl_interface': 5.077309589979716e-07, 'learning_rate/action_model': 5.077309589979715e-06, 'epoch': 1.9}) +[2026-06-23 23:44:54,057][starVLA.training.train_starvla][INFO] - Step 4912, Loss: {'train/loss': 0.11764131486415863, 'train/grad_norm_pre_clip': 14.492796897888184, 'timing/data': 0.0006548301316797733, 'timing/model': 1.6828041749540716, 'learning_rate/qwen_vl_interface': 5.075582514195089e-07, 'learning_rate/action_model': 5.0755825141950895e-06, 'epoch': 1.9}) +[2026-06-23 23:44:57,369][starVLA.training.train_starvla][INFO] - Step 4913, Loss: {'train/loss': 0.08988545089960098, 'train/grad_norm_pre_clip': 16.46828842163086, 'timing/data': 0.000632639043033123, 'timing/model': 1.6830022351350635, 'learning_rate/qwen_vl_interface': 5.073874932797115e-07, 'learning_rate/action_model': 5.0738749327971155e-06, 'epoch': 1.9}) +[2026-06-23 23:45:00,684][starVLA.training.train_starvla][INFO] - Step 4914, Loss: {'train/loss': 0.14647597074508667, 'train/grad_norm_pre_clip': 22.06899070739746, 'timing/data': 0.0006420558784157038, 'timing/model': 1.6848086949903518, 'learning_rate/qwen_vl_interface': 5.072186846487713e-07, 'learning_rate/action_model': 5.072186846487713e-06, 'epoch': 1.9}) +[2026-06-23 23:45:03,999][starVLA.training.train_starvla][INFO] - Step 4915, Loss: {'train/loss': 0.14258135855197906, 'train/grad_norm_pre_clip': 17.930862426757812, 'timing/data': 0.0006398109253495932, 'timing/model': 1.6836649349424988, 'learning_rate/qwen_vl_interface': 5.070518255960796e-07, 'learning_rate/action_model': 5.070518255960795e-06, 'epoch': 1.9}) +[2026-06-23 23:45:07,315][starVLA.training.train_starvla][INFO] - Step 4916, Loss: {'train/loss': 0.13840173184871674, 'train/grad_norm_pre_clip': 16.52012062072754, 'timing/data': 0.000558007974177599, 'timing/model': 1.6841779560782015, 'learning_rate/qwen_vl_interface': 5.068869161902256e-07, 'learning_rate/action_model': 5.068869161902256e-06, 'epoch': 1.9}) +[2026-06-23 23:45:10,723][starVLA.training.train_starvla][INFO] - Step 4917, Loss: {'train/loss': 0.09974327683448792, 'train/grad_norm_pre_clip': 13.504478454589844, 'timing/data': 0.000606409041211009, 'timing/model': 1.6813400839455426, 'learning_rate/qwen_vl_interface': 5.067239564989968e-07, 'learning_rate/action_model': 5.067239564989968e-06, 'epoch': 1.9}) +[2026-06-23 23:45:14,041][starVLA.training.train_starvla][INFO] - Step 4918, Loss: {'train/loss': 0.12217427790164948, 'train/grad_norm_pre_clip': 24.160703659057617, 'timing/data': 0.0005538600962609053, 'timing/model': 1.6895967479795218, 'learning_rate/qwen_vl_interface': 5.065629465893808e-07, 'learning_rate/action_model': 5.065629465893807e-06, 'epoch': 1.9}) +[2026-06-23 23:45:17,366][starVLA.training.train_starvla][INFO] - Step 4919, Loss: {'train/loss': 0.08019831776618958, 'train/grad_norm_pre_clip': 25.62944984436035, 'timing/data': 0.0005547411274164915, 'timing/model': 1.690403622109443, 'learning_rate/qwen_vl_interface': 5.064038865275625e-07, 'learning_rate/action_model': 5.0640388652756245e-06, 'epoch': 1.9}) +[2026-06-23 23:45:20,692][starVLA.training.train_starvla][INFO] - Step 4920, Loss: {'train/loss': 0.045701414346694946, 'train/grad_norm_pre_clip': 12.132026672363281, 'timing/data': 0.0006284201517701149, 'timing/model': 1.6897448929958045, 'learning_rate/qwen_vl_interface': 5.062467763789242e-07, 'learning_rate/action_model': 5.062467763789242e-06, 'epoch': 1.9}) +[2026-06-23 23:45:24,019][starVLA.training.train_starvla][INFO] - Step 4921, Loss: {'train/loss': 0.07950045168399811, 'train/grad_norm_pre_clip': 7.71885871887207, 'timing/data': 0.0005845380946993828, 'timing/model': 1.6898592440411448, 'learning_rate/qwen_vl_interface': 5.060916162080496e-07, 'learning_rate/action_model': 5.060916162080496e-06, 'epoch': 1.9}) +[2026-06-23 23:45:27,343][starVLA.training.train_starvla][INFO] - Step 4922, Loss: {'train/loss': 0.026238437741994858, 'train/grad_norm_pre_clip': 13.133833885192871, 'timing/data': 0.0006064889021217823, 'timing/model': 1.6836305430624634, 'learning_rate/qwen_vl_interface': 5.059384060787182e-07, 'learning_rate/action_model': 5.059384060787183e-06, 'epoch': 1.9}) +[2026-06-23 23:45:30,667][starVLA.training.train_starvla][INFO] - Step 4923, Loss: {'train/loss': 0.14013627171516418, 'train/grad_norm_pre_clip': 26.114971160888672, 'timing/data': 0.0006261158268898726, 'timing/model': 1.68887869711034, 'learning_rate/qwen_vl_interface': 5.05787146053909e-07, 'learning_rate/action_model': 5.0578714605390904e-06, 'epoch': 1.91}) +[2026-06-23 23:45:33,999][starVLA.training.train_starvla][INFO] - Step 4924, Loss: {'train/loss': 0.10229337215423584, 'train/grad_norm_pre_clip': 20.06255531311035, 'timing/data': 0.0005548319313675165, 'timing/model': 1.6943539881613106, 'learning_rate/qwen_vl_interface': 5.056378361957998e-07, 'learning_rate/action_model': 5.056378361957998e-06, 'epoch': 1.91}) +[2026-06-23 23:45:37,326][starVLA.training.train_starvla][INFO] - Step 4925, Loss: {'train/loss': 0.0757659450173378, 'train/grad_norm_pre_clip': 14.671674728393555, 'timing/data': 0.0005473780911415815, 'timing/model': 1.6922496929764748, 'learning_rate/qwen_vl_interface': 5.05490476565766e-07, 'learning_rate/action_model': 5.054904765657659e-06, 'epoch': 1.91}) +[2026-06-23 23:45:40,644][starVLA.training.train_starvla][INFO] - Step 4926, Loss: {'train/loss': 0.12617166340351105, 'train/grad_norm_pre_clip': 16.124116897583008, 'timing/data': 0.0006054770201444626, 'timing/model': 1.6848539300262928, 'learning_rate/qwen_vl_interface': 5.053450672243811e-07, 'learning_rate/action_model': 5.0534506722438114e-06, 'epoch': 1.91}) +[2026-06-23 23:45:43,969][starVLA.training.train_starvla][INFO] - Step 4927, Loss: {'train/loss': 0.04011482372879982, 'train/grad_norm_pre_clip': 29.840749740600586, 'timing/data': 0.0006345519796013832, 'timing/model': 1.6880082511343062, 'learning_rate/qwen_vl_interface': 5.052016082314184e-07, 'learning_rate/action_model': 5.052016082314184e-06, 'epoch': 1.91}) +[2026-06-23 23:45:47,289][starVLA.training.train_starvla][INFO] - Step 4928, Loss: {'train/loss': 0.10228807479143143, 'train/grad_norm_pre_clip': 24.40380859375, 'timing/data': 0.000551996985450387, 'timing/model': 1.6888971370644867, 'learning_rate/qwen_vl_interface': 5.050600996458473e-07, 'learning_rate/action_model': 5.050600996458473e-06, 'epoch': 1.91}) +[2026-06-23 23:45:50,608][starVLA.training.train_starvla][INFO] - Step 4929, Loss: {'train/loss': 0.12022804468870163, 'train/grad_norm_pre_clip': 22.510976791381836, 'timing/data': 0.0005482800770550966, 'timing/model': 1.6855874080210924, 'learning_rate/qwen_vl_interface': 5.049205415258373e-07, 'learning_rate/action_model': 5.049205415258373e-06, 'epoch': 1.91}) +[2026-06-23 23:45:53,933][starVLA.training.train_starvla][INFO] - Step 4930, Loss: {'train/loss': 0.14875951409339905, 'train/grad_norm_pre_clip': 43.341796875, 'timing/data': 0.0005828649736940861, 'timing/model': 1.6836741708684713, 'learning_rate/qwen_vl_interface': 5.047829339287557e-07, 'learning_rate/action_model': 5.047829339287557e-06, 'epoch': 1.91}) +[2026-06-23 23:45:57,351][starVLA.training.train_starvla][INFO] - Step 4931, Loss: {'train/loss': 0.19521456956863403, 'train/grad_norm_pre_clip': 39.79086685180664, 'timing/data': 0.0005984040908515453, 'timing/model': 1.7844117749482393, 'learning_rate/qwen_vl_interface': 5.046472769111668e-07, 'learning_rate/action_model': 5.046472769111667e-06, 'epoch': 1.91}) +[2026-06-23 23:46:00,678][starVLA.training.train_starvla][INFO] - Step 4932, Loss: {'train/loss': 0.07751575112342834, 'train/grad_norm_pre_clip': 18.822322845458984, 'timing/data': 0.000774736050516367, 'timing/model': 1.6910945330746472, 'learning_rate/qwen_vl_interface': 5.045135705288353e-07, 'learning_rate/action_model': 5.0451357052883525e-06, 'epoch': 1.91}) +[2026-06-23 23:46:04,001][starVLA.training.train_starvla][INFO] - Step 4933, Loss: {'train/loss': 0.13597136735916138, 'train/grad_norm_pre_clip': 27.518604278564453, 'timing/data': 0.0006120291072875261, 'timing/model': 1.6862353719770908, 'learning_rate/qwen_vl_interface': 5.043818148367221e-07, 'learning_rate/action_model': 5.043818148367221e-06, 'epoch': 1.91}) +[2026-06-23 23:46:07,326][starVLA.training.train_starvla][INFO] - Step 4934, Loss: {'train/loss': 0.11289152503013611, 'train/grad_norm_pre_clip': 64.46708679199219, 'timing/data': 0.0006212671287357807, 'timing/model': 1.689008136978373, 'learning_rate/qwen_vl_interface': 5.042520098889869e-07, 'learning_rate/action_model': 5.0425200988898685e-06, 'epoch': 1.91}) +[2026-06-23 23:46:10,656][starVLA.training.train_starvla][INFO] - Step 4935, Loss: {'train/loss': 0.20672453939914703, 'train/grad_norm_pre_clip': 44.40918731689453, 'timing/data': 0.0005596920382231474, 'timing/model': 1.696244414895773, 'learning_rate/qwen_vl_interface': 5.041241557389879e-07, 'learning_rate/action_model': 5.041241557389879e-06, 'epoch': 1.91}) +[2026-06-23 23:46:13,976][starVLA.training.train_starvla][INFO] - Step 4936, Loss: {'train/loss': 0.06299014389514923, 'train/grad_norm_pre_clip': 38.775779724121094, 'timing/data': 0.000566994771361351, 'timing/model': 1.6871700489427894, 'learning_rate/qwen_vl_interface': 5.039982524392814e-07, 'learning_rate/action_model': 5.039982524392814e-06, 'epoch': 1.91}) +[2026-06-23 23:46:17,304][starVLA.training.train_starvla][INFO] - Step 4937, Loss: {'train/loss': 0.14939655363559723, 'train/grad_norm_pre_clip': 27.462528228759766, 'timing/data': 0.0005578780546784401, 'timing/model': 1.6864823729265481, 'learning_rate/qwen_vl_interface': 5.038743000416207e-07, 'learning_rate/action_model': 5.0387430004162064e-06, 'epoch': 1.91}) +[2026-06-23 23:46:20,629][starVLA.training.train_starvla][INFO] - Step 4938, Loss: {'train/loss': 0.05352923646569252, 'train/grad_norm_pre_clip': 14.071488380432129, 'timing/data': 0.0005463161505758762, 'timing/model': 1.6909243629779667, 'learning_rate/qwen_vl_interface': 5.037522985969588e-07, 'learning_rate/action_model': 5.037522985969587e-06, 'epoch': 1.91}) +[2026-06-23 23:46:23,956][starVLA.training.train_starvla][INFO] - Step 4939, Loss: {'train/loss': 0.025859400629997253, 'train/grad_norm_pre_clip': 20.84172248840332, 'timing/data': 0.0005902990233153105, 'timing/model': 1.688595976214856, 'learning_rate/qwen_vl_interface': 5.036322481554451e-07, 'learning_rate/action_model': 5.036322481554451e-06, 'epoch': 1.91}) +[2026-06-23 23:46:27,277][starVLA.training.train_starvla][INFO] - Step 4940, Loss: {'train/loss': 0.052417974919080734, 'train/grad_norm_pre_clip': 40.92375946044922, 'timing/data': 0.0006019000429660082, 'timing/model': 1.6869525159709156, 'learning_rate/qwen_vl_interface': 5.035141487664285e-07, 'learning_rate/action_model': 5.035141487664285e-06, 'epoch': 1.91}) +[2026-06-23 23:46:30,645][starVLA.training.train_starvla][INFO] - Step 4941, Loss: {'train/loss': 0.03378893435001373, 'train/grad_norm_pre_clip': 27.522140502929688, 'timing/data': 0.0005706320516765118, 'timing/model': 1.690195312956348, 'learning_rate/qwen_vl_interface': 5.03398000478455e-07, 'learning_rate/action_model': 5.03398000478455e-06, 'epoch': 1.91}) +[2026-06-23 23:46:33,972][starVLA.training.train_starvla][INFO] - Step 4942, Loss: {'train/loss': 0.06710158288478851, 'train/grad_norm_pre_clip': 15.32149887084961, 'timing/data': 0.0005787969566881657, 'timing/model': 1.6892070029862225, 'learning_rate/qwen_vl_interface': 5.032838033392686e-07, 'learning_rate/action_model': 5.032838033392685e-06, 'epoch': 1.91}) +[2026-06-23 23:46:37,296][starVLA.training.train_starvla][INFO] - Step 4943, Loss: {'train/loss': 0.057066213339567184, 'train/grad_norm_pre_clip': 22.37825584411621, 'timing/data': 0.0005704809445887804, 'timing/model': 1.6870990910101682, 'learning_rate/qwen_vl_interface': 5.031715573958111e-07, 'learning_rate/action_model': 5.031715573958111e-06, 'epoch': 1.91}) +[2026-06-23 23:46:40,622][starVLA.training.train_starvla][INFO] - Step 4944, Loss: {'train/loss': 0.021342135965824127, 'train/grad_norm_pre_clip': 18.216129302978516, 'timing/data': 0.0005599719006568193, 'timing/model': 1.690901359077543, 'learning_rate/qwen_vl_interface': 5.030612626942235e-07, 'learning_rate/action_model': 5.030612626942235e-06, 'epoch': 1.91}) +[2026-06-23 23:46:43,947][starVLA.training.train_starvla][INFO] - Step 4945, Loss: {'train/loss': 0.14179284870624542, 'train/grad_norm_pre_clip': 18.855215072631836, 'timing/data': 0.0009305889252573252, 'timing/model': 1.6901278020814061, 'learning_rate/qwen_vl_interface': 5.029529192798433e-07, 'learning_rate/action_model': 5.029529192798433e-06, 'epoch': 1.91}) +[2026-06-23 23:46:47,374][starVLA.training.train_starvla][INFO] - Step 4946, Loss: {'train/loss': 0.08489544689655304, 'train/grad_norm_pre_clip': 19.141204833984375, 'timing/data': 0.0005768740084022284, 'timing/model': 1.6914252149872482, 'learning_rate/qwen_vl_interface': 5.028465271972058e-07, 'learning_rate/action_model': 5.028465271972058e-06, 'epoch': 1.91}) +[2026-06-23 23:46:50,701][starVLA.training.train_starvla][INFO] - Step 4947, Loss: {'train/loss': 0.09050692617893219, 'train/grad_norm_pre_clip': 62.6237678527832, 'timing/data': 0.000611279159784317, 'timing/model': 1.6898696899879724, 'learning_rate/qwen_vl_interface': 5.027420864900457e-07, 'learning_rate/action_model': 5.027420864900456e-06, 'epoch': 1.91}) +[2026-06-23 23:46:54,026][starVLA.training.train_starvla][INFO] - Step 4948, Loss: {'train/loss': 0.13865028321743011, 'train/grad_norm_pre_clip': 43.41585922241211, 'timing/data': 0.0005765731912106276, 'timing/model': 1.687528618145734, 'learning_rate/qwen_vl_interface': 5.026395972012943e-07, 'learning_rate/action_model': 5.0263959720129436e-06, 'epoch': 1.91}) +[2026-06-23 23:46:57,348][starVLA.training.train_starvla][INFO] - Step 4949, Loss: {'train/loss': 0.18053753674030304, 'train/grad_norm_pre_clip': 20.059280395507812, 'timing/data': 0.0005655819550156593, 'timing/model': 1.685398104134947, 'learning_rate/qwen_vl_interface': 5.025390593730809e-07, 'learning_rate/action_model': 5.025390593730809e-06, 'epoch': 1.92}) +[2026-06-23 23:47:00,677][starVLA.training.train_starvla][INFO] - Step 4950, Loss: {'train/loss': 0.10913906991481781, 'train/grad_norm_pre_clip': 15.749425888061523, 'timing/data': 0.000637267017737031, 'timing/model': 1.691805134061724, 'learning_rate/qwen_vl_interface': 5.024404730467325e-07, 'learning_rate/action_model': 5.024404730467325e-06, 'epoch': 1.92}) +[2026-06-23 23:47:04,002][starVLA.training.train_starvla][INFO] - Step 4951, Loss: {'train/loss': 0.16550090909004211, 'train/grad_norm_pre_clip': 26.411230087280273, 'timing/data': 0.0006312460172921419, 'timing/model': 1.6894956619944423, 'learning_rate/qwen_vl_interface': 5.023438382627749e-07, 'learning_rate/action_model': 5.0234383826277486e-06, 'epoch': 1.92}) +[2026-06-23 23:47:07,326][starVLA.training.train_starvla][INFO] - Step 4952, Loss: {'train/loss': 0.061662402004003525, 'train/grad_norm_pre_clip': 9.768619537353516, 'timing/data': 0.0006382688879966736, 'timing/model': 1.6886903562117368, 'learning_rate/qwen_vl_interface': 5.022491550609311e-07, 'learning_rate/action_model': 5.022491550609311e-06, 'epoch': 1.92}) +[2026-06-23 23:47:10,652][starVLA.training.train_starvla][INFO] - Step 4953, Loss: {'train/loss': 0.06263939291238785, 'train/grad_norm_pre_clip': 41.791133880615234, 'timing/data': 0.0005767738912254572, 'timing/model': 1.6896397939417511, 'learning_rate/qwen_vl_interface': 5.021564234801213e-07, 'learning_rate/action_model': 5.021564234801213e-06, 'epoch': 1.92}) +[2026-06-23 23:47:13,976][starVLA.training.train_starvla][INFO] - Step 4954, Loss: {'train/loss': 0.10499109327793121, 'train/grad_norm_pre_clip': 40.888404846191406, 'timing/data': 0.0005619151052087545, 'timing/model': 1.689498103922233, 'learning_rate/qwen_vl_interface': 5.020656435584636e-07, 'learning_rate/action_model': 5.020656435584635e-06, 'epoch': 1.92}) +[2026-06-23 23:47:17,300][starVLA.training.train_starvla][INFO] - Step 4955, Loss: {'train/loss': 0.13210193812847137, 'train/grad_norm_pre_clip': 13.244431495666504, 'timing/data': 0.0005608031060546637, 'timing/model': 1.689289758913219, 'learning_rate/qwen_vl_interface': 5.019768153332752e-07, 'learning_rate/action_model': 5.019768153332752e-06, 'epoch': 1.92}) +[2026-06-23 23:47:20,628][starVLA.training.train_starvla][INFO] - Step 4956, Loss: {'train/loss': 0.03407891094684601, 'train/grad_norm_pre_clip': 14.47396469116211, 'timing/data': 0.0005456048529595137, 'timing/model': 1.6930573307909071, 'learning_rate/qwen_vl_interface': 5.018899388410693e-07, 'learning_rate/action_model': 5.0188993884106935e-06, 'epoch': 1.92}) +[2026-06-23 23:47:23,955][starVLA.training.train_starvla][INFO] - Step 4957, Loss: {'train/loss': 0.0952843725681305, 'train/grad_norm_pre_clip': 22.341157913208008, 'timing/data': 0.0005812819581478834, 'timing/model': 1.692192783113569, 'learning_rate/qwen_vl_interface': 5.018050141175576e-07, 'learning_rate/action_model': 5.018050141175575e-06, 'epoch': 1.92}) +[2026-06-23 23:47:27,281][starVLA.training.train_starvla][INFO] - Step 4958, Loss: {'train/loss': 0.16990379989147186, 'train/grad_norm_pre_clip': 28.52495765686035, 'timing/data': 0.0005790782161056995, 'timing/model': 1.6881354078650475, 'learning_rate/qwen_vl_interface': 5.017220411976496e-07, 'learning_rate/action_model': 5.0172204119764965e-06, 'epoch': 1.92}) +[2026-06-23 23:47:30,609][starVLA.training.train_starvla][INFO] - Step 4959, Loss: {'train/loss': 0.04672401398420334, 'train/grad_norm_pre_clip': 31.379213333129883, 'timing/data': 0.000568808987736702, 'timing/model': 1.6889794690068811, 'learning_rate/qwen_vl_interface': 5.016410201154523e-07, 'learning_rate/action_model': 5.016410201154523e-06, 'epoch': 1.92}) +[2026-06-23 23:47:33,928][starVLA.training.train_starvla][INFO] - Step 4960, Loss: {'train/loss': 0.25267142057418823, 'train/grad_norm_pre_clip': 20.236927032470703, 'timing/data': 0.000559080159291625, 'timing/model': 1.6848622120451182, 'learning_rate/qwen_vl_interface': 5.015619509042709e-07, 'learning_rate/action_model': 5.015619509042709e-06, 'epoch': 1.92}) +[2026-06-23 23:47:37,406][starVLA.training.train_starvla][INFO] - Step 4961, Loss: {'train/loss': 0.02411072328686714, 'train/grad_norm_pre_clip': 18.076419830322266, 'timing/data': 0.00056618289090693, 'timing/model': 1.691297615878284, 'learning_rate/qwen_vl_interface': 5.014848335966067e-07, 'learning_rate/action_model': 5.014848335966067e-06, 'epoch': 1.92}) +[2026-06-23 23:47:40,733][starVLA.training.train_starvla][INFO] - Step 4962, Loss: {'train/loss': 0.10464932769536972, 'train/grad_norm_pre_clip': 28.469270706176758, 'timing/data': 0.0006085729692131281, 'timing/model': 1.6899001561105251, 'learning_rate/qwen_vl_interface': 5.014096682241608e-07, 'learning_rate/action_model': 5.014096682241608e-06, 'epoch': 1.92}) +[2026-06-23 23:47:44,063][starVLA.training.train_starvla][INFO] - Step 4963, Loss: {'train/loss': 0.13485126197338104, 'train/grad_norm_pre_clip': 16.873647689819336, 'timing/data': 0.0005747789982706308, 'timing/model': 1.6916557680815458, 'learning_rate/qwen_vl_interface': 5.013364548178299e-07, 'learning_rate/action_model': 5.013364548178299e-06, 'epoch': 1.92}) +[2026-06-23 23:47:47,384][starVLA.training.train_starvla][INFO] - Step 4964, Loss: {'train/loss': 0.10593748092651367, 'train/grad_norm_pre_clip': 12.763915061950684, 'timing/data': 0.0005462560802698135, 'timing/model': 1.6887859040871263, 'learning_rate/qwen_vl_interface': 5.012651934077101e-07, 'learning_rate/action_model': 5.0126519340771e-06, 'epoch': 1.92}) +[2026-06-23 23:47:50,709][starVLA.training.train_starvla][INFO] - Step 4965, Loss: {'train/loss': 0.13482089340686798, 'train/grad_norm_pre_clip': 30.338354110717773, 'timing/data': 0.0005927130114287138, 'timing/model': 1.6893083611503243, 'learning_rate/qwen_vl_interface': 5.01195884023094e-07, 'learning_rate/action_model': 5.0119588402309395e-06, 'epoch': 1.92}) +[2026-06-23 23:47:54,043][starVLA.training.train_starvla][INFO] - Step 4966, Loss: {'train/loss': 0.08697793632745743, 'train/grad_norm_pre_clip': 15.610225677490234, 'timing/data': 0.000566474162042141, 'timing/model': 1.6939645619131625, 'learning_rate/qwen_vl_interface': 5.011285266924714e-07, 'learning_rate/action_model': 5.011285266924714e-06, 'epoch': 1.92}) +[2026-06-23 23:47:57,376][starVLA.training.train_starvla][INFO] - Step 4967, Loss: {'train/loss': 0.037046875804662704, 'train/grad_norm_pre_clip': 24.255033493041992, 'timing/data': 0.0006271579768508673, 'timing/model': 1.6941561580169946, 'learning_rate/qwen_vl_interface': 5.010631214435312e-07, 'learning_rate/action_model': 5.010631214435312e-06, 'epoch': 1.92}) +[2026-06-23 23:48:00,702][starVLA.training.train_starvla][INFO] - Step 4968, Loss: {'train/loss': 0.13317178189754486, 'train/grad_norm_pre_clip': 37.58884811401367, 'timing/data': 0.0005464961286634207, 'timing/model': 1.6898127850145102, 'learning_rate/qwen_vl_interface': 5.009996683031586e-07, 'learning_rate/action_model': 5.009996683031586e-06, 'epoch': 1.92}) +[2026-06-23 23:48:04,029][starVLA.training.train_starvla][INFO] - Step 4969, Loss: {'train/loss': 0.2142910212278366, 'train/grad_norm_pre_clip': 72.46605682373047, 'timing/data': 0.0005534400697797537, 'timing/model': 1.6904069858137518, 'learning_rate/qwen_vl_interface': 5.009381672974375e-07, 'learning_rate/action_model': 5.009381672974375e-06, 'epoch': 1.92}) +[2026-06-23 23:48:07,362][starVLA.training.train_starvla][INFO] - Step 4970, Loss: {'train/loss': 0.04084402322769165, 'train/grad_norm_pre_clip': 14.56131362915039, 'timing/data': 0.0005787871778011322, 'timing/model': 1.6909931830596179, 'learning_rate/qwen_vl_interface': 5.008786184516475e-07, 'learning_rate/action_model': 5.008786184516474e-06, 'epoch': 1.92}) +[2026-06-23 23:48:10,695][starVLA.training.train_starvla][INFO] - Step 4971, Loss: {'train/loss': 0.04950374364852905, 'train/grad_norm_pre_clip': 24.915864944458008, 'timing/data': 0.0005596810951828957, 'timing/model': 1.6900171251036227, 'learning_rate/qwen_vl_interface': 5.008210217902676e-07, 'learning_rate/action_model': 5.008210217902676e-06, 'epoch': 1.92}) +[2026-06-23 23:48:14,024][starVLA.training.train_starvla][INFO] - Step 4972, Loss: {'train/loss': 0.12544254958629608, 'train/grad_norm_pre_clip': 31.2641658782959, 'timing/data': 0.0005613840185105801, 'timing/model': 1.6924609909765422, 'learning_rate/qwen_vl_interface': 5.007653773369737e-07, 'learning_rate/action_model': 5.007653773369736e-06, 'epoch': 1.92}) +[2026-06-23 23:48:17,347][starVLA.training.train_starvla][INFO] - Step 4973, Loss: {'train/loss': 0.19959834218025208, 'train/grad_norm_pre_clip': 21.89016342163086, 'timing/data': 0.0005630080122500658, 'timing/model': 1.6857320358976722, 'learning_rate/qwen_vl_interface': 5.007116851146386e-07, 'learning_rate/action_model': 5.007116851146386e-06, 'epoch': 1.92}) +[2026-06-23 23:48:20,680][starVLA.training.train_starvla][INFO] - Step 4974, Loss: {'train/loss': 0.08505547791719437, 'train/grad_norm_pre_clip': 37.40668487548828, 'timing/data': 0.0005870219320058823, 'timing/model': 1.6903152910526842, 'learning_rate/qwen_vl_interface': 5.006599451453334e-07, 'learning_rate/action_model': 5.006599451453334e-06, 'epoch': 1.93}) +[2026-06-23 23:48:24,145][starVLA.training.train_starvla][INFO] - Step 4975, Loss: {'train/loss': 0.1456400454044342, 'train/grad_norm_pre_clip': 43.95859909057617, 'timing/data': 0.0005958101246505976, 'timing/model': 1.826296087121591, 'learning_rate/qwen_vl_interface': 5.006101574503268e-07, 'learning_rate/action_model': 5.006101574503267e-06, 'epoch': 1.93}) +[2026-06-23 23:48:27,477][starVLA.training.train_starvla][INFO] - Step 4976, Loss: {'train/loss': 0.1268051713705063, 'train/grad_norm_pre_clip': 43.67766571044922, 'timing/data': 0.0005659239832311869, 'timing/model': 1.6910898259375244, 'learning_rate/qwen_vl_interface': 5.005623220500839e-07, 'learning_rate/action_model': 5.005623220500839e-06, 'epoch': 1.93}) +[2026-06-23 23:48:30,804][starVLA.training.train_starvla][INFO] - Step 4977, Loss: {'train/loss': 0.07571598887443542, 'train/grad_norm_pre_clip': 23.877601623535156, 'timing/data': 0.0005785261746495962, 'timing/model': 1.6889249139931053, 'learning_rate/qwen_vl_interface': 5.005164389642687e-07, 'learning_rate/action_model': 5.005164389642687e-06, 'epoch': 1.93}) +[2026-06-23 23:48:34,134][starVLA.training.train_starvla][INFO] - Step 4978, Loss: {'train/loss': 0.04994870349764824, 'train/grad_norm_pre_clip': 18.05096435546875, 'timing/data': 0.0006171490531414747, 'timing/model': 1.6912619390059263, 'learning_rate/qwen_vl_interface': 5.004725082117423e-07, 'learning_rate/action_model': 5.004725082117422e-06, 'epoch': 1.93}) +[2026-06-23 23:48:37,463][starVLA.training.train_starvla][INFO] - Step 4979, Loss: {'train/loss': 0.12164494395256042, 'train/grad_norm_pre_clip': 15.727014541625977, 'timing/data': 0.0005782460793852806, 'timing/model': 1.6908257950562984, 'learning_rate/qwen_vl_interface': 5.004305298105616e-07, 'learning_rate/action_model': 5.004305298105616e-06, 'epoch': 1.93}) +[2026-06-23 23:48:40,794][starVLA.training.train_starvla][INFO] - Step 4980, Loss: {'train/loss': 0.2506694793701172, 'train/grad_norm_pre_clip': 21.24483299255371, 'timing/data': 0.0005524971056729555, 'timing/model': 1.6929579849820584, 'learning_rate/qwen_vl_interface': 5.003905037779834e-07, 'learning_rate/action_model': 5.0039050377798345e-06, 'epoch': 1.93}) +[2026-06-23 23:48:44,117][starVLA.training.train_starvla][INFO] - Step 4981, Loss: {'train/loss': 0.09967690706253052, 'train/grad_norm_pre_clip': 62.55168151855469, 'timing/data': 0.0006141441408544779, 'timing/model': 1.6884253970347345, 'learning_rate/qwen_vl_interface': 5.003524301304608e-07, 'learning_rate/action_model': 5.003524301304608e-06, 'epoch': 1.93}) +[2026-06-23 23:48:47,444][starVLA.training.train_starvla][INFO] - Step 4982, Loss: {'train/loss': 0.052821867167949677, 'train/grad_norm_pre_clip': 15.205957412719727, 'timing/data': 0.0005788868293166161, 'timing/model': 1.6864225990138948, 'learning_rate/qwen_vl_interface': 5.003163088836445e-07, 'learning_rate/action_model': 5.003163088836446e-06, 'epoch': 1.93}) +[2026-06-23 23:48:50,772][starVLA.training.train_starvla][INFO] - Step 4983, Loss: {'train/loss': 0.05530953034758568, 'train/grad_norm_pre_clip': 37.64792251586914, 'timing/data': 0.0005636080168187618, 'timing/model': 1.6903768309857696, 'learning_rate/qwen_vl_interface': 5.002821400523818e-07, 'learning_rate/action_model': 5.002821400523818e-06, 'epoch': 1.93}) +[2026-06-23 23:48:54,107][starVLA.training.train_starvla][INFO] - Step 4984, Loss: {'train/loss': 0.09399624168872833, 'train/grad_norm_pre_clip': 50.66041946411133, 'timing/data': 0.000642476137727499, 'timing/model': 1.6938792050350457, 'learning_rate/qwen_vl_interface': 5.002499236507192e-07, 'learning_rate/action_model': 5.0024992365071925e-06, 'epoch': 1.93}) +[2026-06-23 23:48:57,433][starVLA.training.train_starvla][INFO] - Step 4985, Loss: {'train/loss': 0.06414972245693207, 'train/grad_norm_pre_clip': 46.89641571044922, 'timing/data': 0.0005511660128831863, 'timing/model': 1.6888354860711843, 'learning_rate/qwen_vl_interface': 5.002196596918993e-07, 'learning_rate/action_model': 5.002196596918993e-06, 'epoch': 1.93}) +[2026-06-23 23:49:00,763][starVLA.training.train_starvla][INFO] - Step 4986, Loss: {'train/loss': 0.09294456988573074, 'train/grad_norm_pre_clip': 34.56462860107422, 'timing/data': 0.0005554028321057558, 'timing/model': 1.6898931481409818, 'learning_rate/qwen_vl_interface': 5.001913481883623e-07, 'learning_rate/action_model': 5.001913481883623e-06, 'epoch': 1.93}) +[2026-06-23 23:49:04,090][starVLA.training.train_starvla][INFO] - Step 4987, Loss: {'train/loss': 0.06098958104848862, 'train/grad_norm_pre_clip': 16.936386108398438, 'timing/data': 0.000640813959762454, 'timing/model': 1.693932820810005, 'learning_rate/qwen_vl_interface': 5.001649891517459e-07, 'learning_rate/action_model': 5.001649891517459e-06, 'epoch': 1.93}) +[2026-06-23 23:49:07,420][starVLA.training.train_starvla][INFO] - Step 4988, Loss: {'train/loss': 0.10460857301950455, 'train/grad_norm_pre_clip': 22.934913635253906, 'timing/data': 0.0005581080913543701, 'timing/model': 1.6945973539259285, 'learning_rate/qwen_vl_interface': 5.001405825928858e-07, 'learning_rate/action_model': 5.0014058259288575e-06, 'epoch': 1.93}) +[2026-06-23 23:49:10,757][starVLA.training.train_starvla][INFO] - Step 4989, Loss: {'train/loss': 0.08872866630554199, 'train/grad_norm_pre_clip': 19.902070999145508, 'timing/data': 0.0005926629528403282, 'timing/model': 1.6912043949123472, 'learning_rate/qwen_vl_interface': 5.001181285218143e-07, 'learning_rate/action_model': 5.0011812852181434e-06, 'epoch': 1.93}) +[2026-06-23 23:49:14,363][starVLA.training.train_starvla][INFO] - Step 4990, Loss: {'train/loss': 0.08805752545595169, 'train/grad_norm_pre_clip': 13.301843643188477, 'timing/data': 0.0005514861550182104, 'timing/model': 1.967637844150886, 'learning_rate/qwen_vl_interface': 5.000976269477613e-07, 'learning_rate/action_model': 5.0009762694776125e-06, 'epoch': 1.93}) +[2026-06-23 23:49:17,696][starVLA.training.train_starvla][INFO] - Step 4991, Loss: {'train/loss': 0.14406898617744446, 'train/grad_norm_pre_clip': 19.317174911499023, 'timing/data': 0.0005756320897489786, 'timing/model': 1.6908993020188063, 'learning_rate/qwen_vl_interface': 5.000790778791542e-07, 'learning_rate/action_model': 5.000790778791543e-06, 'epoch': 1.93}) +[2026-06-23 23:49:21,025][starVLA.training.train_starvla][INFO] - Step 4992, Loss: {'train/loss': 0.1863030344247818, 'train/grad_norm_pre_clip': 30.079055786132812, 'timing/data': 0.0005858808290213346, 'timing/model': 1.6892222261521965, 'learning_rate/qwen_vl_interface': 5.000624813236185e-07, 'learning_rate/action_model': 5.000624813236185e-06, 'epoch': 1.93}) +[2026-06-23 23:49:24,357][starVLA.training.train_starvla][INFO] - Step 4993, Loss: {'train/loss': 0.15128067135810852, 'train/grad_norm_pre_clip': 47.68583297729492, 'timing/data': 0.0005493711214512587, 'timing/model': 1.692468968918547, 'learning_rate/qwen_vl_interface': 5.000478372879752e-07, 'learning_rate/action_model': 5.000478372879752e-06, 'epoch': 1.93}) +[2026-06-23 23:49:27,685][starVLA.training.train_starvla][INFO] - Step 4994, Loss: {'train/loss': 0.12081100791692734, 'train/grad_norm_pre_clip': 32.736167907714844, 'timing/data': 0.0005541809368878603, 'timing/model': 1.6917029051110148, 'learning_rate/qwen_vl_interface': 5.000351457782452e-07, 'learning_rate/action_model': 5.000351457782451e-06, 'epoch': 1.93}) +[2026-06-23 23:49:31,014][starVLA.training.train_starvla][INFO] - Step 4995, Loss: {'train/loss': 0.07482956349849701, 'train/grad_norm_pre_clip': 15.332528114318848, 'timing/data': 0.000565732829272747, 'timing/model': 1.6937249777838588, 'learning_rate/qwen_vl_interface': 5.000244067996445e-07, 'learning_rate/action_model': 5.000244067996445e-06, 'epoch': 1.93}) +[2026-06-23 23:49:34,340][starVLA.training.train_starvla][INFO] - Step 4996, Loss: {'train/loss': 0.040900468826293945, 'train/grad_norm_pre_clip': 34.1961669921875, 'timing/data': 0.0005594808608293533, 'timing/model': 1.689206664916128, 'learning_rate/qwen_vl_interface': 5.000156203565884e-07, 'learning_rate/action_model': 5.000156203565883e-06, 'epoch': 1.93}) +[2026-06-23 23:49:37,668][starVLA.training.train_starvla][INFO] - Step 4997, Loss: {'train/loss': 0.10918847471475601, 'train/grad_norm_pre_clip': 19.31138038635254, 'timing/data': 0.0005506840534508228, 'timing/model': 1.6946272510103881, 'learning_rate/qwen_vl_interface': 5.000087864526878e-07, 'learning_rate/action_model': 5.000087864526878e-06, 'epoch': 1.93}) +[2026-06-23 23:49:40,995][starVLA.training.train_starvla][INFO] - Step 4998, Loss: {'train/loss': 0.08856413513422012, 'train/grad_norm_pre_clip': 38.655494689941406, 'timing/data': 0.0005733771249651909, 'timing/model': 1.6862205350771546, 'learning_rate/qwen_vl_interface': 5.000039050907523e-07, 'learning_rate/action_model': 5.000039050907523e-06, 'epoch': 1.93}) +[2026-06-23 23:49:44,319][starVLA.training.train_starvla][INFO] - Step 4999, Loss: {'train/loss': 0.07334768027067184, 'train/grad_norm_pre_clip': 15.831018447875977, 'timing/data': 0.0005800388753414154, 'timing/model': 1.6874850050080568, 'learning_rate/qwen_vl_interface': 5.000009762727885e-07, 'learning_rate/action_model': 5.000009762727884e-06, 'epoch': 1.93}) +[2026-06-23 23:49:47,652][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 5000 +[2026-06-23 23:51:19,656][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 5000 +[2026-06-23 23:51:19,657][starVLA.training.train_starvla][INFO] - Starting held-out action classification eval at step 5000 +[2026-06-23 23:52:30,932][starVLA.training.train_starvla][INFO] - Finished held-out action classification eval at step 5000 +[2026-06-23 23:53:18,989][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 5000 +[2026-06-23 23:53:18,989][starVLA.training.train_starvla][INFO] - Step 5000, Loss: {'train/loss': 0.1065579354763031, 'train/grad_norm_pre_clip': 19.516754150390625, 'eval/action_loss/seconds': 92.00275186891668, 'eval/action_loss/samples': 6400.0, 'eval/loss': 0.06451020389795303, 'eval/latency_2/loss': 0.06451020389795303, 'eval/flappy/loss': 0.06451020389795303, 'eval/flappy/latency_2/loss': 0.06451020389795303, 'eval/flap_precision': 0.9324055666003976, 'eval/flap_recall': 0.9493927125506073, 'eval/flap_f1': 0.9408224674022068, 'eval/flap_rate_model': 0.20958333333333334, 'eval/flap_rate_teacher': 0.20583333333333334, 'eval/flap_cc_f1': 0.9408224674022069, 'eval/latency_2/flap_precision': 0.9324055666003976, 'eval/latency_2/flap_recall': 0.9493927125506073, 'eval/latency_2/flap_f1': 0.9408224674022068, 'eval/latency_2/flap_rate_model': 0.20958333333333334, 'eval/latency_2/flap_rate_teacher': 0.20583333333333334, 'eval/latency_2/flap_cc_f1': 0.9408224674022069, 'eval/action_classification/data_wait_seconds': 0.9620541930198669, 'eval/action_classification/predict_seconds': 70.18016815185547, 'eval/action_classification/frames': 7200.0, 'rl_games_eval/mid_train/total_episodes': 5.0, 'rl_games_eval/mid_train/mean_reward': 35.979999999999954, 'rl_games_eval/mid_train/mean_length': 307.8, 'rl_games_eval/mid_train/std_reward': 30.149056369975977, 'rl_games_eval/mid_train/std_length': 244.03884936624334, 'rl_games_eval/mid_train/macro_mean_reward': 35.979999999999954, 'rl_games_eval/mid_train/macro_mean_length': 307.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_2/mean_reward': 35.979999999999954, 'rl_games_eval/mid_train/flappy__latency_2/mean_length': 307.8, 'rl_games_eval/mid_train/flappy__latency_2/std_reward': 30.149056369975977, 'rl_games_eval/mid_train/flappy__latency_2/std_length': 244.03884936624334, 'rl_games_eval/mid_train/flappy__latency_2/num_episodes': 5.0, 'checkpoint/current_eval_score': 35.979999999999954, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0005880340468138456, 'timing/model': 1.6932119969278574, 'learning_rate/qwen_vl_interface': 5e-07, 'learning_rate/action_model': 4.9999999999999996e-06, 'epoch': 1.94}) +[2026-06-23 23:53:18,990][accelerate.accelerator][INFO] - Saving current state to /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_5000_state +[2026-06-23 23:53:18,992][accelerate.utils.other][WARNING] - Removed shared tensor {'qwen_vl_interface.model.lm_head.weight'} while saving. This should be OK, but check by verifying that you don't receive any warning while reloading +[2026-06-23 23:53:21,735][accelerate.checkpointing][INFO] - Model weights saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_5000_state/model.safetensors +[2026-06-23 23:53:22,848][accelerate.checkpointing][INFO] - Optimizer state saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_5000_state/optimizer.bin +[2026-06-23 23:53:22,848][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_5000_state/sampler.bin +[2026-06-23 23:53:22,849][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_5000_state/sampler_1.bin +[2026-06-23 23:53:22,849][accelerate.checkpointing][INFO] - Random states saved in /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints/steps_5000_state/random_states_0.pkl +[2026-06-23 23:53:22,875][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 5000 +[2026-06-23 23:53:22,875][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration... +[2026-06-23 23:53:22,885][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved +[2026-06-23 23:53:22,886][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /workspace/latency-sensitive-bench/starVLA/results/Checkpoints/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/checkpoints +[2026-06-23 23:53:24,825][starVLA.training.train_starvla][INFO] - ... and that's all, folks! diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/summary.jsonl b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/summary.jsonl new file mode 100644 index 0000000000000000000000000000000000000000..a0e21985e6b7ed5b88fddcc5f74f6ec45063b603 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy_fix_latency_2_200ep_last_8_layers_corrected_v2/summary.jsonl @@ -0,0 +1,10 @@ +{"steps": 500} +{"steps": 1000} +{"steps": 1500} +{"steps": 2000} +{"steps": 2500} +{"steps": 3000} +{"steps": 3500} +{"steps": 4000} +{"steps": 4500} +{"steps": 5000} diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..0c3655c15eb902341c015dad2ed84bfae5d6d3fe --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/config.full.yaml @@ -0,0 +1,121 @@ +rl_games: + task: flappy + model_alias: openvla + env_eval: + enabled: false + interval_steps: 100 + num_episodes: 5 + max_episode_steps: 2000 + frameskip: 1 + image_size: 84 + latency: + mode: single + values: + - 2 + prompt_map_path: /root/latency-sensitive-bench/playground/Datasets/rl_games/flappy_fixed_l2_fs4/latency_prompt_map.json + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 2000 + post_train: + enabled: false + latencies: + - 2 + num_episodes: 50 + max_steps_per_episode: 2000 +framework: + name: QwenOFT + qwenvl: + base_vlm: /root/latency-sensitive-bench/playground/Pretrained_models/Qwen3-VL-4B-Instruct-Action + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + lora: + enabled: false + r: 32 + alpha: 64 + dropout: 0.05 + target_modules: all-linear + bias: none + action_model: + action_model_type: MLP + action_dim: 2 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: ce + ce_label_smoothing: 0.0 + action_env_dim: 2 + action_layout: flappy_categorical_2 + state_dim: 1 + action_horizon: 1 +trainer: + learning_rate: + base: 2.0e-05 + action_model: 0.0001 + qwen_vl_interface: 2.0e-05 + max_train_steps: 2000 + num_warmup_steps: 100 + save_interval: 2001 + eval_interval: 100 + eval_num_batches: 20 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 2 + distributed_backend: deepspeed + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +datasets: + vla_data: + data_mix: flappy_fixed_l2_fs4 + dataset_py: lerobot_datasets + include_state: false + data_root_dir: /root/latency-sensitive-bench/playground/Datasets/rl_games + eval_data_mix: flappy_fixed_l2_fs4__val + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 32 + load_all_data_for_training: true + obs_image_size: + - 224 + - 224 + video_backend: torchvision_av +checkpoint: + local: + keep_last_n: 3 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_flappy_l2_full +run_root_dir: /root/latency-sensitive-bench/results/Checkpoints +output_dir: /root/latency-sensitive-bench/results/Checkpoints/openvla_flappy_l2_full +config_yaml: null +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..c2a7e19ac8a7ce72169aca92472769f878877da2 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/config.yaml @@ -0,0 +1,55 @@ +checkpoint: + local: + keep_last_n: 3 + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_fixed_l2_fs4 + dataset_py: lerobot_datasets + eval_data_mix: flappy_fixed_l2_fs4__val + include_state: false + obs_image_size: + - 224 + - 224 + per_device_batch_size: 32 +framework: + action_model: + action_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: StarVLA/Qwen3-VL-4B-Instruct-Action + enable_gradient_checkpointing: true + lora: + enabled: false +output_dir: /root/latency-sensitive-bench/results/Checkpoints/openvla_flappy_l2_full +rl_games: + env_eval: + enabled: false +run_id: openvla_flappy_l2_full +run_root_dir: /root/latency-sensitive-bench/results/Checkpoints +seed: 42 +trainer: + freeze_modules: '' + learning_rate: + action_model: 0.0001 + base: 2.0e-05 + qwen_vl_interface: 2.0e-05 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 2000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + scheduler_specific_kwargs: + min_lr: 1.0e-06 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..6ef7516fb940012861f81b9aa87fbec8d4ac19ff --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/dataset_statistics.json @@ -0,0 +1,56 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7675621509552002, + 0.2324378490447998 + ], + "std": [ + 0.4213072657585144, + 0.4213072657585144 + ], + "max": [ + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0 + ], + "mask": [ + true, + true + ] + }, + "state": { + "mean": [ + 0.0 + ], + "std": [ + 0.0 + ], + "max": [ + 0.0 + ], + "min": [ + 0.0 + ], + "q01": [ + 0.0 + ], + "q99": [ + 0.0 + ] + }, + "num_transitions": 948645, + "num_trajectories": 990 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..a2f006673c3802e01782881afa1c8f51834d448f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/dataset_statistics_eval.json @@ -0,0 +1,56 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7666000127792358, + 0.23340000212192535 + ], + "std": [ + 0.42298394441604614, + 0.42298394441604614 + ], + "max": [ + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0 + ], + "mask": [ + true, + true + ] + }, + "state": { + "mean": [ + 0.0 + ], + "std": [ + 0.0 + ], + "max": [ + 0.0 + ], + "min": [ + 0.0 + ], + "q01": [ + 0.0 + ], + "q99": [ + 0.0 + ] + }, + "num_transitions": 10000, + "num_trajectories": 10 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/final_model/pytorch_model.pt b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/final_model/pytorch_model.pt new file mode 100644 index 0000000000000000000000000000000000000000..9cc52d54a8b6850db9187024eba21ad94d7fea9e --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_full/final_model/pytorch_model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f6be353e457ac9cf6dbda6cd2e93692d797016100b69ecea4bcab0ececb9a076 +size 9805972271 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/config.full.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/config.full.yaml new file mode 100644 index 0000000000000000000000000000000000000000..08880fed320e37143a43ec37535106812b565455 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/config.full.yaml @@ -0,0 +1,121 @@ +rl_games: + task: flappy + model_alias: openvla + env_eval: + enabled: false + interval_steps: 100 + num_episodes: 5 + max_episode_steps: 2000 + frameskip: 1 + image_size: 84 + latency: + mode: single + values: + - 2 + prompt_map_path: /mnt/data/latency_data/playground/Datasets/rl_games/flappy_fixed_l2_fs4/latency_prompt_map.json + mid_train: + enabled: false + interval_steps: 100 + latencies: + - 2 + num_episodes: 5 + max_steps_per_episode: 2000 + post_train: + enabled: false + latencies: + - 2 + num_episodes: 50 + max_steps_per_episode: 2000 +framework: + name: QwenOFT + qwenvl: + base_vlm: /mnt/data/latency_data/playground/Pretrained_models/Qwen3-VL-4B-Instruct-Action + attn_implementation: flash_attention_2 + enable_gradient_checkpointing: true + lora: + enabled: true + r: 32 + alpha: 64 + dropout: 0.05 + target_modules: all-linear + bias: none + action_model: + action_model_type: MLP + action_dim: 2 + action_hidden_dim: 2560 + future_action_window_size: 0 + past_action_window_size: 0 + loss_type: ce + ce_label_smoothing: 0.0 + action_env_dim: 2 + action_layout: flappy_categorical_2 + state_dim: 1 + action_horizon: 1 +trainer: + learning_rate: + base: 0.0002 + action_model: 0.0001 + qwen_vl_interface: 0.0002 + max_train_steps: 2000 + num_warmup_steps: 100 + save_interval: 100 + eval_interval: 100 + eval_num_batches: 20 + lr_scheduler_type: cosine_with_min_lr + scheduler_specific_kwargs: + min_lr: 1.0e-06 + freeze_modules: '' + loss_scale: + vla: 1.0 + vlm: 0.1 + max_grad_norm: 1.0 + weight_decay: 0.0 + logging_frequency: 10 + gradient_clipping: 1.0 + gradient_accumulation_steps: 4 + distributed_backend: deepspeed + is_resume: false + pretrained_checkpoint: null + resume_step: 0 + reload_modules: null + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + weight_decay: 1.0e-08 + fused: false + save_format: pt +datasets: + vla_data: + data_mix: flappy_fixed_l2_fs4 + dataset_py: lerobot_datasets + include_state: false + data_root_dir: /mnt/data/latency_data/playground/Datasets/rl_games + eval_data_mix: flappy_fixed_l2_fs4__val + action_type: discrete + sequential_step_sampling: false + per_device_batch_size: 16 + load_all_data_for_training: true + obs_image_size: + - 224 + - 224 + video_backend: torchvision_av +checkpoint: + local: + keep_last_n: 3 + sync: + enabled: false + repo_id: null + keep_last_n: 0 + sync_every_n_checkpoints: 1 + resume_policy: local_latest +run_id: openvla_flappy_l2_lora +run_root_dir: /mnt/data/latency_data/results/Checkpoints +output_dir: /mnt/data/latency_data/results/Checkpoints/openvla_flappy_l2_lora +config_yaml: null +seed: 42 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games +is_debug: false +version_id: '0.21' diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/config.yaml b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/config.yaml new file mode 100644 index 0000000000000000000000000000000000000000..0d57203b12edbc326db87df7cef15ede58a31b09 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/config.yaml @@ -0,0 +1,71 @@ +checkpoint: + local: + keep_last_n: 3 + sync: + enabled: false + keep_last_n: 0 + repo_id: null +datasets: + vla_data: + data_mix: flappy_fixed_l2_fs4 + dataset_py: lerobot_datasets + eval_data_mix: flappy_fixed_l2_fs4__val + include_state: false + obs_image_size: + - 224 + - 224 + per_device_batch_size: 16 +framework: + action_model: + action_dim: 2 + action_hidden_dim: 2560 + action_horizon: 1 + action_model_type: MLP + ce_label_smoothing: 0.0 + loss_type: ce + name: QwenOFT + qwenvl: + attn_implementation: flash_attention_2 + base_vlm: /mnt/data/latency_data/playground/Pretrained_models/Qwen3-VL-4B-Instruct-Action + enable_gradient_checkpointing: true + lora: + alpha: 64 + bias: none + dropout: 0.05 + enabled: true + r: 32 + target_modules: all-linear +output_dir: /mnt/data/latency_data/results/Checkpoints/openvla_flappy_l2_lora +rl_games: + env_eval: + enabled: false +run_id: openvla_flappy_l2_lora +run_root_dir: /mnt/data/latency_data/results/Checkpoints +seed: 42 +trainer: + eval_interval: 100 + eval_num_batches: 20 + freeze_modules: '' + is_resume: false + learning_rate: + action_model: 0.0001 + base: 0.0002 + qwen_vl_interface: 0.0002 + logging_frequency: 10 + lr_scheduler_type: cosine_with_min_lr + max_train_steps: 2000 + num_warmup_steps: 100 + optimizer: + betas: + - 0.9 + - 0.95 + eps: 1.0e-08 + fused: false + weight_decay: 1.0e-08 + pretrained_checkpoint: null + save_format: pt + save_interval: 100 + scheduler_specific_kwargs: + min_lr: 1.0e-06 +wandb_entity: saberrr-zju +wandb_project: starVLA_rl_games diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/dataset_statistics.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/dataset_statistics.json new file mode 100644 index 0000000000000000000000000000000000000000..6ef7516fb940012861f81b9aa87fbec8d4ac19ff --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/dataset_statistics.json @@ -0,0 +1,56 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7675621509552002, + 0.2324378490447998 + ], + "std": [ + 0.4213072657585144, + 0.4213072657585144 + ], + "max": [ + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0 + ], + "mask": [ + true, + true + ] + }, + "state": { + "mean": [ + 0.0 + ], + "std": [ + 0.0 + ], + "max": [ + 0.0 + ], + "min": [ + 0.0 + ], + "q01": [ + 0.0 + ], + "q99": [ + 0.0 + ] + }, + "num_transitions": 948645, + "num_trajectories": 990 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/dataset_statistics_eval.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/dataset_statistics_eval.json new file mode 100644 index 0000000000000000000000000000000000000000..a2f006673c3802e01782881afa1c8f51834d448f --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/dataset_statistics_eval.json @@ -0,0 +1,56 @@ +{ + "new_embodiment": { + "action": { + "mean": [ + 0.7666000127792358, + 0.23340000212192535 + ], + "std": [ + 0.42298394441604614, + 0.42298394441604614 + ], + "max": [ + 1.0, + 1.0 + ], + "min": [ + 0.0, + 0.0 + ], + "q01": [ + 0.0, + 0.0 + ], + "q99": [ + 1.0, + 1.0 + ], + "mask": [ + true, + true + ] + }, + "state": { + "mean": [ + 0.0 + ], + "std": [ + 0.0 + ], + "max": [ + 0.0 + ], + "min": [ + 0.0 + ], + "q01": [ + 0.0 + ], + "q99": [ + 0.0 + ] + }, + "num_transitions": 10000, + "num_trajectories": 10 + } +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/final_model/README.md b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/final_model/README.md new file mode 100644 index 0000000000000000000000000000000000000000..a8e8f9df2d944c1f2decd3a0520d1b7c1c2548fd --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/final_model/README.md @@ -0,0 +1,206 @@ +--- +base_model: /mnt/data/latency_data/playground/Pretrained_models/Qwen3-VL-4B-Instruct-Action +library_name: peft +tags: +- base_model:adapter:/mnt/data/latency_data/playground/Pretrained_models/Qwen3-VL-4B-Instruct-Action +- lora +- transformers +--- + +# Model Card for Model ID + + + + + +## Model Details + +### Model Description + + + + + +- **Developed by:** [More Information Needed] +- **Funded by [optional]:** [More Information Needed] +- **Shared by [optional]:** [More Information Needed] +- **Model type:** [More Information Needed] +- **Language(s) (NLP):** [More Information Needed] +- **License:** [More Information Needed] +- **Finetuned from model [optional]:** [More Information Needed] + +### Model Sources [optional] + + + +- **Repository:** [More Information Needed] +- **Paper [optional]:** [More Information Needed] +- **Demo [optional]:** [More Information Needed] + +## Uses + + + +### Direct Use + + + +[More Information Needed] + +### Downstream Use [optional] + + + +[More Information Needed] + +### Out-of-Scope Use + + + +[More Information Needed] + +## Bias, Risks, and Limitations + + + +[More Information Needed] + +### Recommendations + + + +Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. + +## How to Get Started with the Model + +Use the code below to get started with the model. + +[More Information Needed] + +## Training Details + +### Training Data + + + +[More Information Needed] + +### Training Procedure + + + +#### Preprocessing [optional] + +[More Information Needed] + + +#### Training Hyperparameters + +- **Training regime:** [More Information Needed] + +#### Speeds, Sizes, Times [optional] + + + +[More Information Needed] + +## Evaluation + + + +### Testing Data, Factors & Metrics + +#### Testing Data + + + +[More Information Needed] + +#### Factors + + + +[More Information Needed] + +#### Metrics + + + +[More Information Needed] + +### Results + +[More Information Needed] + +#### Summary + + + +## Model Examination [optional] + + + +[More Information Needed] + +## Environmental Impact + + + +Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). + +- **Hardware Type:** [More Information Needed] +- **Hours used:** [More Information Needed] +- **Cloud Provider:** [More Information Needed] +- **Compute Region:** [More Information Needed] +- **Carbon Emitted:** [More Information Needed] + +## Technical Specifications [optional] + +### Model Architecture and Objective + +[More Information Needed] + +### Compute Infrastructure + +[More Information Needed] + +#### Hardware + +[More Information Needed] + +#### Software + +[More Information Needed] + +## Citation [optional] + + + +**BibTeX:** + +[More Information Needed] + +**APA:** + +[More Information Needed] + +## Glossary [optional] + + + +[More Information Needed] + +## More Information [optional] + +[More Information Needed] + +## Model Card Authors [optional] + +[More Information Needed] + +## Model Card Contact + +[More Information Needed] +### Framework versions + +- PEFT 0.18.1 \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/final_model/action_model.pt b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/final_model/action_model.pt new file mode 100644 index 0000000000000000000000000000000000000000..661881ca12ca9d350b9bc3bb657a6aa6416d2ff7 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/final_model/action_model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e8bc83aaf046d0b6c2045cd38c06875ccc60c808701d6eba10fc491e2b016ffb +size 131200330 diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/final_model/adapter_config.json b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/final_model/adapter_config.json new file mode 100644 index 0000000000000000000000000000000000000000..01cbc20d3949a6b089fc5721a4f277afe369dc7b --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/final_model/adapter_config.json @@ -0,0 +1,53 @@ +{ + "alora_invocation_tokens": null, + "alpha_pattern": {}, + "arrow_config": null, + "auto_mapping": { + "base_model_class": "Qwen3VLForConditionalGeneration", + "parent_library": "transformers.models.qwen3_vl.modeling_qwen3_vl" + }, + "base_model_name_or_path": "/mnt/data/latency_data/playground/Pretrained_models/Qwen3-VL-4B-Instruct-Action", + "bias": "none", + "corda_config": null, + "ensure_weight_tying": false, + "eva_config": null, + "exclude_modules": null, + "fan_in_fan_out": false, + "inference_mode": true, + "init_lora_weights": true, + "layer_replication": null, + "layers_pattern": null, + "layers_to_transform": null, + "loftq_config": {}, + "lora_alpha": 64, + "lora_bias": false, + "lora_dropout": 0.05, + "megatron_config": null, + "megatron_core": "megatron.core", + "modules_to_save": null, + "peft_type": "LORA", + "peft_version": "0.18.1", + "qalora_group_size": 16, + "r": 32, + "rank_pattern": {}, + "revision": null, + "target_modules": [ + "gate_proj", + "o_proj", + "qkv", + "attn.proj", + "linear_fc2", + "up_proj", + "linear_fc1", + "k_proj", + "q_proj", + "v_proj", + "down_proj" + ], + "target_parameters": null, + "task_type": null, + "trainable_token_indices": null, + "use_dora": false, + "use_qalora": false, + "use_rslora": false +} \ No newline at end of file diff --git a/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/final_model/adapter_model.safetensors b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/final_model/adapter_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..76ef65bf4e3fdfb1a48157dd0f49b94db9379400 --- /dev/null +++ b/latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_l2_lora/final_model/adapter_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:551bd2362cc9548b9d97f3a4f77fcfbe2b28f2725b2897ab42b996485962b8bb +size 161192096